Cancer Epidemiology and Prevention

Cancer Epidemiology and Prevention

Associate Editors Graham A. Colditz, M.D., Dr.P.H. Department of Medicine Harvard Medical School Department of Epidemiology Harvard School of Public Health Boston, Massachusetts Jonathan M. Samet, M.D. Department of Epidemiology Johns Hopkins Bloomberg School of Public Health Baltimore, Maryland Alice S. Whittemore, Ph.D. Department of Health Research and Policy Stanford University School of Medicine Stanford, California

Cancer Epidemiology and Prevention Third Edition Edited by

DAVID SCHOTTENFELD, M.D. John G. Searle Professor Emeritus of Epidemiology University of Michigan School of Public Health Ann Arbor, Michigan

JOSEPH F. FRAUMENI, JR., M.D. Director, Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

1 2006

3 Oxford University Press, Inc., publishes works that further Oxford University’s objective of excellence in research, scholarship, and education. Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi New Delhi Shanghai Taipei Toronto With offices in Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Thailand Turkey Ukraine Vietnam

Copyright © 2006 by Oxford University Press, Inc. Published by Oxford University Press, Inc. 198 Madison Avenue, New York, New York 10016 www.oup.com Oxford is a registered trademark of Oxford University Press All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of Oxford University Press. Library of Congress Cataloging-in-Publication Data Cancer epidemiology and prevention / edited by David Schottenfeld, Joseph F. Fraumeni Jr.—3rd ed. p. ; cm. Includes bibliographical references and index. ISBN-13: 978-0-19-514961-6 ISBN-10: 0-19-514961-0 1. Cancer—Prevention. 2. Cancer–Epidemiology. I. Schottenfeld, David. II. Fraumeni, Joseph F. [DNLM: 1. Neoplasms–epidemiology. 2. Neoplasms—prevention & control. QZ 220.1 C215 2006] RC268.C354 2006 616.99¢4—dc22 2005051838 The science of medicine is a rapidly changing field. As new research and clinical experience broaden our knowledge, changes in treatment and drug therapy do occur. The authors and the publisher of this work have checked with sources believed to be reliable in their efforts to provide information that is accurate and complete, and in accordance with the standards accepted at the time of publication. However, in light of the possibility of human error or changes in the practice of medicine. neither the authors, nor the publisher, nor any other party who has been involved in the preparation or publication of this work warrants that the information contained herein is in every respect accurate or complete. Readers are encouraged to confirm the information contained herein with other reliable sources, and are strongly advised to check the product information sheet provided by the pharmaceutical company for each drug they plan to administer.

9 8 7 6 5 4 3 2 1 Printed in the United States of America on acid-free paper

To Rosalie and Tricia

This page intentionally left blank

Preface

The third edition of Cancer Epidemiology and Prevention represents a comprehensive update of information that has increased dramatically since the publication of the second edition in 1996. Once again the volume addresses the global burden of cancer, the complex interrelationship of environmental and genetic factors involved in the induction and progression of a broad spectrum of malignancies, and the current priorities and challenges in cancer epidemiology and prevention research. Cancer incidence and mortality is universal, but the burden of cancer classified by morphology and organ site is distributed unevenly in different populations around the world. In developing countries the populations are disproportionately affected by cancers related to infectious agents, while in industrialized countries a large percentage of cancers are associated with lifestyle factors including smoking, obesity, and physical inactivity. The heavy worldwide toll of cancer can be gauged in a survey of the International Agency for Research on Cancer, which estimated that in 2002 there were 10.9 million cancers diagnosed, 6.7 million cancer deaths, and 24.6 million persons living with cancers detected within the previous 5 years. In the United States, cancer mortality in men and women younger than 85 years has surpassed heart disease as the leading cause of death since 1999, although cancer mortality is still less common than heart disease after age 85. This volume maintains the structure of previous editions, with chapters grouped into five major sections: Basic Concepts, The Magnitude of Cancer, The Causes of Cancer, Cancer by Tissue of Origin, and Cancer Prevention and Control. We have undertaken an extensive revision, adding new chapters in each section to expand the scope of coverage and keep pace with the striking progress in our understanding of cancer biology and etiology over the past decade. The introductory chapters under Basic Concepts have been amended to highlight the advances in genomic and molecular sciences that are increasingly incorporated into epidemiologic research designed to uncover the environmental and heritable determinants of cancer development and progression. The section now includes chapters on the molecular as well as morphologic classification of cancer, the critical genetic events that provoke normal cells to malignant transformation and tumor invasion, the origins and natural history of cancer precursors, and the epidemiologic application of emerging molecular and biochemical biomarkers that sharpen our measures of carcinogenic exposures, susceptibility genes, and intermediate outcomes. The section on the magnitude of cancer highlights the enormous variation in cancer incidence that occurs around the world, and the major shifts in incidence within one or two generations among migrant populations that have provided important leads into the lifestyle and cultural determinants of cancer risk. The patterns of cancer incidence, mortality and survival in the United States are reviewed for various racial and ethnic groups, while a complementary chapter dealing with the substantial socioeconomic disparities in cancer incidence and mortality is added. Also covered is the economic impact of cancer in a chapter that estimates the direct medical

viii

Preface

costs of treating specific types of cancer, costs to the family and community in providing rehabilitative services or palliative care, and costs to society as a result of premature death or lost productivity. The section on the causes of cancer provides the foundation for viewing the multifactorial origins of human cancer. Whereas the neoplastic process at the cellular level results from cumulative genetic perturbations in the mechanisms that regulate proliferation, differentiation, apoptosis and DNA repair, these events appear to be triggered or propagated by lifestyle and other environmental exposures that are critically evaluated in this section. To keep abreast of the rapid expansion of epidemiologic literature, separate chapters are devoted to the cancer risks associated with obesity and with physical activity in addition to dietary and nutritional factors, while a new chapter on electromagnetic fields and radiofrequencies evaluates the current epidemiologic evidence for this widely studied exposure that remains a speculative risk factor for cancer. To highlight the exciting recent progress made in identifying mechanisms of genetic susceptibility, chapters are devoted to the uncommon but important hereditary neoplastic syndromes, and to the extensive ongoing search for common susceptibility or modifier genes that may play a major role in cancer development through interactions with environmental exposures. The section on cancer by tissue of origin provides a comprehensive epidemiologic survey covering a wide variety of cancers, including childhood tumors and multiple primary cancers. New chapters on pleural and peritoneal malignancies, as well as anal cancer, are included. Each chapter systematically addresses the demographic, environmental and host factors that influence cancer risk, and efforts are made to integrate developments from clinical and laboratory sciences into concepts of carcinogenesis, current strategies aimed at cancer prevention, and future directions in illuminating causal mechanisms and assessing the benefits of preventive interventions. The concluding section on cancer prevention and control addresses the methods and applications for translating the results of epidemiologic, clinical, and laboratory research into preventive interventions that will contribute to the eventual goal of eliminating suffering and death due to cancer. It is generally understood that preventing cancer is far preferable to experiencing aggressive treatment, and many strategies are currently available to accomplish this objective. Emphasis is given to approaches aimed at identifying, quantifying, and ultimately reducing the prevalence of cancer risk factors in diverse populations. Special attention is placed on measuring the impact of lifestyle and behavioral interventions, health-promoting practices, as well as governmental policies that regulate environmental carcinogens. The benefits and challenges of cancer screening practices and the use of chemopreventive agents are also reviewed. Although further understanding of causal factors and pathways through epidemiologic and other research will contribute to improvements in preventive strategies, it is important to capitalize on the panoply of preventive measures that are already at hand. Through the shared efforts of innovative research programs, a supportive health care system, a cooperative network of national and international agencies, and an informed public, we can expect to see a rapid acceleration of evidence-based public health policies and clinical interventions that are designed to keep cancer from developing and escaping control.

Contents

I BASIC CONCEPTS 1. Cause and Cancer Epidemiology Steven N. Goodman and Jonathan M. Samet

3

2. Morphologic and Molecular Classification of Human Cancer Thomas J. Giordano

10

3. Cancer Precursors Thomas E. Rohan, Donald E. Henson, Eduardo L. Franco, and Jorge Albores-Saavedra

21

4. Molecular and Genetic Events in Neoplastic Transformation Ayse E. Erson and Elizabeth M. Petty

47

5. Risk Assessment of Carcinogenic Hazards Leslie T. Stayner, Paolo Boffetta, and Harri Vainio

65

6. Application of Biomarkers in Cancer Epidemiology Montserrat García-Closas, Roel Vermeulen, Mark E. Sherman, Lee E. Moore, Martyn T. Smith, and Nathaniel Rothman

70

7. Genetic Concepts and Methods in Epidemiologic Research Neil J. Risch and Alice S. Whittemore

89

II THE MAGNITUDE OF CANCER 8. International Patterns of Cancer Incidence and Mortality D. Maxwell Parkin and Freddie I. Bray

101

9. Cancer Incidence, Mortality, and Patient Survival in the United States Lynn A.G. Ries and Susan S. Devesa

139

10. Socioeconomic Disparities in Cancer Incidence and Mortality Ichiro Kawachi and Candyce Kroenke

174

11. Migrant Studies Laurence N. Kolonel and Lynne R. Wilkens

189

12. Economic Impact of Cancer in the United States Martin L. Brown and K. Robin Yabroff

202

x

Contents

III THE CAUSES OF CANCER 13. Tobacco Michael J. Thun and S. Jane Henley

217

14. Alcohol James R. Marshall and Jo Freudenheim

243

15. Ionizing Radiation John D. Boice, Jr.

259

16. Solar Radiation Adele C. Green and David C. Whiteman

294

17. Electromagnetic Fields and Radiofrequency Radiation David A. Savitz and Anders Ahlbom

306

18. Occupation Jack Siemiatycki, Lesley Richardson, and Paolo Boffetta

322

19. Air Pollution Jonathan M. Samet and Aaron J. Cohen

355

20. Water Contaminants Kenneth P. Cantor, Mary H. Ward, Lee E. Moore, and Jay H. Lubin

382

21. Diet and Nutrition Walter C. Willett

405

22. Obesity and Body Composition Rachel Ballard-Barbash, Christine Friedenreich, Martha Slattery, and Inger Thune

422

23. Physical Activity I-Min Lee and Yuko Oguma

449

24. Exogenous Hormones James V. Lacey, Jr., Graham A. Colditz, and David Schottenfeld

468

25. Pharmaceuticals Other Than Hormones Laurel A. Habel and Gary D. Friedman

489

26. Infectious Agents Nancy E. Mueller, Brenda M. Birmann, Julie Parsonnet, Mark H. Schiffman, and Sherri O. Stuver

507

27. Immunologic Factors Gareth J. Morgan, Martha S. Linet, and Charles S. Rabkin

549

28. Hereditary Neoplastic Syndromes Noralane M. Lindor, Carl J. Lindor, and Mark H. Greene

562

29. Genetic Modifiers of Cancer Risk Neil E. Caporaso

577

IV CANCER BY TISSUE OF ORIGIN 30. Cancers of the Nasal Cavity and Paranasal Sinuses Alyson J. Littman and Thomas L. Vaughan

603

31. Nasopharyngeal Cancer Mimi C. Yu and Jian-Min Yuan

620

32. Cancer of the Larynx Andrew F. Olshan

627

33. Cancer of the Lung Margaret R. Spitz, Xifeng Wu, Anna Wilkinson, and Qingyi Wei

638

xi

Contents

34. Pleural and Peritoneal Neoplasms Paolo Boffetta and Leslie T. Stayner

659

35. Cancers of the Oral Cavity and Pharynx Susan T. Mayne, Douglas E. Morse, and Deborah M. Winn

674

36. Esophageal Cancer William J. Blot, Joseph K. McLaughlin, and Joseph F. Fraumeni, Jr.

697

37. Stomach Cancer Atsuko Shibata and Julie Parsonnet

707

38. Cancer of the Pancreas Kristin E. Anderson, Thomas M. Mack, and Debra T. Silverman

721

39. Liver Cancer W. Thomas London and Katherine A. McGlynn

763

40. Biliary Tract Cancer Ann W. Hsing, Asif Rashid, Susan S. Devesa, and Joseph F. Fraumeni, Jr.

787

41. Cancers of the Small Intestine Jennifer L. Beebe-Dimmer and David Schottenfeld

801

42. Cancers of the Colon and Rectum Edward Giovannucci and Kana Wu

809

43. Anal Cancer Morten Frisch and Mads Melbye

830

44. The Leukemias Martha S. Linet, Susan S. Devesa, and Gareth J. Morgan

841

45. Hodgkin Lymphoma Nancy E. Mueller and Seymour Grufferman

872

46. Non-Hodgkin Lymphoma Patricia Hartge, Sophia S. Wang, Paige M. Bracci, Susan S. Devesa, and Elizabeth A. Holly

898

47. Multiple Myeloma Anneclaire J. De Roos, Dalsu Baris, Noel S. Weiss, and Lisa J. Herrinton

919

48. Bone Cancer Robert W. Miller, John D. Boice, Jr., and Rochelle E. Curtis

946

49. Soft Tissue Sarcoma Marianne Berwick

959

50. Thyroid Cancer Elaine Ron and Arthur B. Schneider

975

51. Breast Cancer Graham A. Colditz, Heather J. Baer, and Rulla M. Tamimi

995

52. Ovarian Cancer Susan E. Hankinson and Kim N. Danforth

1013

53. Endometrial Cancer Linda S. Cook, Noel S. Weiss, Jennifer A. Doherty, and Chu Chen

1027

54. Cervical Cancer Mark H. Schiffman and Allan Hildesheim

1044

55. Cancers of the Vulva and Vagina Margaret M. Madeleine and Janet R. Daling

1068

56. Choriocarcinoma Julie R. Palmer and Colleen M. Feltmate

1075

xii

Contents

57. Renal Cancer Joseph K. McLaughlin, Loren Lipworth, Robert E. Tarone, and William J. Blot

1087

58. Bladder Cancer Debra T. Silverman, Susan S. Devesa, Lee E. Moore, and Nathaniel Rothman

1101

59. Prostate Cancer Elizabeth A. Platz and Edward Giovannucci

1128

60. Testicular Cancer Aruna V. Sarma, Julie C. McLaughlin, and David Schottenfeld

1151

61. Penile Cancer Louise Wideroff and David Schottenfeld

1166

62. Nervous System Susan Preston-Martin, Reema Munir, and Indro Chakrabarti

1173

63. Cutaneous and Ocular Melanoma Stephen B. Gruber and Bruce K. Armstrong

1196

64. Keratinocyte Carcinomas (Basal and Squamous Cell Carcinomas of the Skin) Margaret R. Karagas, Martin A. Weinstock, and Heather H. Nelson

1230

65. Cancers in Children Julie A. Ross and Logan G. Spector

1251

66. Multiple Primary Cancers David Schottenfeld and Jennifer L. Beebe-Dimmer

1269

V CANCER PREVENTION AND CONTROL 67. Principles and Applications of Cancer Prevention and Control Interventions Robert A. Hiatt and Barbara K. Rimer

1283

68. Increasing the Contribution of Epidemiology to the Primary Prevention of Cancer Beverly Rockhill and Douglas Weed

1292

69. Cancer Risk Communication and Comprehension Karen M. Emmons, Cara Cuite, and Erika Waters

1303

70. Principles of Screening Bernard Levin and Philip C. Prorok

1310

71. Cancer Chemoprevention Jaye L. Viner, Ernest Hawk, and Scott M. Lippman

1318

72. Regulating Carcinogens Jonathan M. Samet, Thomas A. Burke, and Lynn Goldman

1341

Index

1355

Contributors

Anders Ahlbom, PhD

Marianne Berwick, PhD, MPH

Division of Epidemiology National Institute of Environmental Medicine Karolinska University Stockholm, Sweden

Division of Epidemiology and Cancer Prevention University of New Mexico Albuquerque, New Mexico

Jorge Albores-Saavedra, MD

Brenda M. Birmann, ScD

Department of Pathology Louisiana State University Shreveport, Louisiana

Department of Epidemiology Harvard School of Public Health Boston, Massachusetts

Kristin E. Anderson, PhD, MPH

William J. Blot, PhD

Division of Epidemiology University of Minnesota School of Public Health Minneapolis, Minnesota

International Epidemiology Institute Rockville, Maryland

Bruce K. Armstrong, MBBS, DPhil

International Agency for Research on Cancer Lyon, France

School of Public Health University of Sydney Sydney, Australia

Heather J. Baer, ScD Department of Medicine Brigham and Women’s Hospital Harvard Medical School Channing Laboratory Boston, Massachusetts

Paolo Boffetta, MD, MPH

John D. Boice, Jr., ScD International Epidemiology Institute Rockville, Maryland

Paige M. Bracci, MSc, MPH Department of Epidemiology and Biostatistics University of California San Francisco San Francisco, California

Rachel Ballard-Barbash, MD, MPH Division of Cancer Control and Population Sciences National Cancer Institute Bethesda, Maryland

Dalsu Baris, MD, PhD Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Freddie I. Bray, BSc, MSc The Cancer Registry of Norway Oslo, Norway

Martin L. Brown, PhD Division of Cancer Control and Population Sciences National Cancer Institute Bethesda, Maryland

Jennifer L. Beebe-Dimmer, MPH, PhD Departments of Epidemiology and Urology University of Michigan School of Public Health and University of Michigan Medical School Ann Arbor, Michigan

Thomas A. Burke, PhD, MPH Department of Health Policy and Management Johns Hopkins Bloomberg School of Public Health Baltimore, Maryland

xiii

xiv

Contributors

Kenneth P. Cantor, PhD, MPH

Jennifer A. Doherty, PhD

Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Division of Public Health Sciences Fred Hutchinson Cancer Research Center Seattle, Washington

Neil E. Caporaso, MD

Karen M. Emmons, PhD

Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Division of Community-Based Research Dana Farber Cancer Institute Boston, Massachusetts

Indro Chakrabarti, MD, MPH

Ayse E. Erson, PhD

Department of Neurosurgery Keck School of Medicine University of Southern California Los Angeles, California

Colleen M. Feltmate, MD

Chu Chen, PhD Division of Public Health Sciences Fred Hutchinson Cancer Research Center Seattle, Washington

Aaron J. Cohen, MPH, DSc

Department of Internal Medicine University of Michigan Medical School Ann Arbor, Michigan Department of Obstetrics and Gynecology Brigham and Women’s Hospital Harvard Medical School Boston, Massachusetts

Eduardo L. Franco, DrPH, MPH

Health Effects Institute Boston, Massachusetts

Departments of Epidemiology and Oncology McGill University Montreal, Quebec, Canada

Graham A. Colditz, MD, DrPH

Joseph F. Fraumeni, Jr., MD, MSc

Department of Medicine Harvard Medical School Department of Epidemiology Harvard School of Public Health Boston, Massachusetts

Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Linda S. Cook, PhD Department of Community Health Sciences University of Calgary Calgary, Alberta, Canada

Cara Cuite, PhD Food Policy Institute Rutgers University New Brunswick, New Jersey

Rochelle E. Curtis, MA Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Janet R. Daling, PhD Division of Public Health Sciences Fred Hutchinson Cancer Research Center Seattle, Washington

Kim N. Danforth, MPH Department of Epidemiology Harvard School of Public Health Boston, Massachusetts

Anneclaire J. De Roos, MPH, PhD Division of Public Health Sciences Fred Hutchinson Cancer Research Center Seattle, Washington

Jo Freudenheim, PhD School of Medicine and Biomedical Sciences State University of New York at Buffalo Buffalo, New York

Christine Friedenreich, PhD Alberta Cancer Board Calgary, Alberta, Canada

Gary D. Friedman, MD, MS Department of Health Research and Policy Stanford University School of Medicine Stanford, California

Morten Frisch, MD, PhD Danish Epidemiology Science Center State Serum Institute Copenhagen, Denmark

Montserrat García-Closas, MD, DrPH Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Thomas J. Giordano, MD, PhD Department of Pathology University of Michigan Medical School Ann Arbor, Michigan

Edward Giovannucci, MD, ScD Departments of Nutrition and Epidemiology Harvard School of Public Health Boston, Massachusetts

Susan S. Devesa, PhD

Lynn Goldman, MD, MPH, MS

Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Department of Environmental Health Sciences Johns Hopkins Bloomberg School of Public Health Baltimore, Maryland

Contributors

Steven N. Goodman, MD, PhD

Elizabeth A. Holly, PhD, MPH

Department of Oncology Johns Hopkins University School of Medicine Baltimore, Maryland

Division of Cancer Epidemiology University of California San Francisco School of Medicine San Francisco, California

Adele C. Green, MBBS, PhD

Ann W. Hsing, PhD

Population Studies and Human Genetics Division Queensland Institute of Medical Research Brisbane, Queensland, Australia

Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Mark H. Greene, MD

Margaret R. Karagas, PhD

Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Department of Community and Family Medicine Norris Cotton Cancer Center Dartmouth Medical School Lebanon, New Hampshire

Stephen B. Gruber, MD, PhD, MPH Division of Molecular Medicine & Genetics University of Michigan Medical School Ann Arbor, Michigan

Seymour Grufferman, MD, DrPH Department of Pathology University of New Mexico School of Medicine Albuquerque, New Mexico

Laurel A. Habel, PhD Division of Research Kaiser Permanente Oakland, California

Susan E. Hankinson, ScD Department of Medicine Brigham and Women’s Hospital Harvard Medical School Boston, Massachusetts

Patricia Hartge, ScD

Ichiro Kawachi, MD, PhD Department of Society, Human Development, and Health Harvard School of Public Health Boston, Massachusetts

Laurence N. Kolonel, MD, PhD Cancer Research Center of Hawaii University of Hawaii Honolulu, Hawaii

Candyce Kroenke, ScD, MPH Department of Medicine Harvard Medical School Boston, Massachusetts

James V. Lacey, Jr., PhD, MPH Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

I-Min Lee, MBBS, ScD

Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Department of Medicine Brigham and Women’s Hospital Harvard Medical School Boston, Massachusetts

Ernest Hawk, MD, MPH

Bernard Levin, MD

National Cancer Institute Bethesda, Maryland

S. Jane Henley, MSPH American Cancer Society Atlanta, Georgia

Donald E. Henson, MD Department of Pathology George Washington University Medical Center Washington, District of Columbia

Lisa J. Herrinton, PhD Division of Research Kaiser Permanente Oakland, California

Robert A. Hiatt, MD, PhD University of California San Francisco Comprehensive Cancer Center San Francisco, California

Allan Hildesheim, PhD Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Division of Cancer Prevention MD Anderson Cancer Center Houston, Texas

Carl J. Lindor, BA University of St. Thomas St. Paul, Minnesota

Noralane M. Lindor, MD Department of Medical Genetics Mayo Clinic College of Medicine Rochester, Minnesota

Martha S. Linet, MD, MPH Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Scott M. Lippman, MD Department of Clinical Cancer Prevention MD Anderson Cancer Center Houston, Texas

Loren Lipworth, ScD International Epidemiology Institute Rockville, Maryland

xv

xvi

Contributors

Alyson J. Littman, PhD

Douglas E. Morse, DDS, PhD

Department of Epidemiology School of Public Health and Community Medicine University of Washington Seattle, Washington

Department of Epidemiology and Health Promotion New York University College of Dentistry New York, New York

W. Thomas London, MD Fox Chase Cancer Center Philadelphia, Pennsylvania

Department of Epidemiology Harvard School of Public Health Boston, Massachusetts

Jay H. Lubin, PhD

Reema Munir, MD

Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Thomas M. Mack, MD, MPH Department of Preventive Medicine Keck School of Medicine University of Southern California Los Angeles, California

Margaret M. Madeleine, PhD, MPH Division of Public Health Sciences Fred Hutchinson Cancer Research Center Seattle, Washington

James R. Marshall, PhD Cancer Prevention and Population Sciences Roswell Park Cancer Institute Buffalo, New York

Susan T. Mayne, PhD Department of Epidemiology and Public Health Yale University School of Medicine New Haven, Connecticut

Katherine A. McGlynn, PhD Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Nancy E. Mueller, ScD

Department of Preventive Medicine Keck School of Medicine University of Southern California Los Angeles, California

Heather H. Nelson, PhD, MPH Department of Environmental Health Harvard School of Public Health Boston, Massachusetts

Yuko Oguma, MD Department of Epidemiology Harvard School of Public Health Boston, Massachusetts

Andrew F. Olshan, PhD Department of Epidemiology School of Public Health University of North Carolina Chapel Hill, North Carolina

Julie R. Palmer, ScD Slone Epidemiology Center Boston University Boston, Massachusetts

D. Maxwell Parkin, MD

International Epidemiology Institute Rockville, Maryland

Clinical Trials Service Unit, and Epidemiological Studies Unit Nuffield Department of Clinical Medicine University of Oxford Oxford, United Kingdom

Julie C. McLaughlin, MPH, MS

Julie Parsonnet, MD

Department of Urology University of Michigan Ann Arbor, Michigan

Departments of Medicine, and Health Research and Policy Stanford University School of Medicine Stanford, California

Mads Melbye, MD, PhD

Elizabeth M. Petty, MD

Danish Epidemiology Science Center State Serum Institute Copenhagen, Denmark

Departments of Internal Medicine and Human Genetics University of Michigan Medical School Ann Arbor, Michigan

Robert W. Miller, MD, DrPH*

Elizabeth A. Platz, ScD

Joseph K. McLaughlin, PhD

Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Lee E. Moore, PhD, MPH Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Gareth J. Morgan, MD, PhD Haemato-Oncology Unit Royal Marsden Hospital Sutton, Surrey, United Kingdom * deceased

Department of Epidemiology Johns Hopkins Bloomberg School of Public Health Baltimore, Maryland

Susan Preston-Martin, PhD Department of Preventive Medicine Keck School of Medicine University of Southern California Los Angeles, California

Philip C. Prorok, PhD Division of Cancer Prevention National Cancer Institute Bethesda, Maryland

Contributors

Charles S. Rabkin, MD, MSc

Mark H. Schiffman, MD, MPH

Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Asif Rashid, MD, PhD

Arthur B. Schneider, MD, PhD

Department of Pathology MD Anderson Cancer Center Houston, Texas

Department of Medicine University of Illinois at Chicago Chicago, Illinois

Lesley Richardson, MSc

David Schottenfeld, MD, MSc

International Agency for Research on Cancer Lyon, France

Departments of Epidemiology and Internal Medicine University of Michigan School of Public Health and University of Michigan Medical School Ann Arbor, Michigan

Lynn A.G. Ries, MS Division of Cancer Control and Population Sciences National Cancer Institute Bethesda, Maryland

Barbara K. Rimer, DrPH School of Public Health University of North Carolina Chapel Hill, North Carolina

Neil J. Risch, PhD

Mark E. Sherman, MD Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Atsuko Shibata, MD, PhD Amgen Inc. Thousand Oaks, California

Center for Human Genetics University of California San Francisco San Francisco, California

Jack Siemiatycki, PhD

Beverly Rockhill, PhD

Debra T. Silverman, ScD

Department of Epidemiology School of Public Health University of North Carolina Chapel Hill, North Carolina

Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Thomas E. Rohan, MD, PhD

University of Utah Health Research Center Salt Lake City, Utah

Department of Epidemiology and Population Health Albert Einstein College of Medicine Bronx, New York

CRCHUM—Population Health Montreal, Quebec, Canada

Martha Slattery, PhD, MPH

Martyn T. Smith, PhD

Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Environmental Health Sciences Division School of Public Health University of California at Berkeley Berkeley, California

Julie A. Ross, PhD

Logan G. Spector, PhD

Department of Pediatrics University of Minnesota Minneapolis, Minnesota

Department of Pediatrics University of Minnesota Minneapolis, Minnesota

Nathaniel Rothman, MD, MPH

Margaret R. Spitz, MD, MPH

Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Department of Epidemiology MD Anderson Cancer Center Houston, Texas

Jonathan M. Samet, MD

Leslie T. Stayner, PhD, MSc

Department of Epidemiology Johns Hopkins Bloomberg School of Public Health Baltimore, Maryland

University of Illinois at Chicago School of Public Health Chicago, Illinois

Aruna V. Sarma, PhD, MHA

Sherri O. Stuver, ScD

Departments of Urology and Epidemiology University of Michigan School of Public Health and University of Michigan Medical School Ann Arbor, Michigan

Department of Epidemiology Boston University School of Public Health Boston, Massachusetts

David A. Savitz, PhD

Department of Medicine Harvard Medical School Channing Laboratory Boston, Massachusetts

Elaine Ron, PhD, MPH

Department of Community and Preventive Medicine Mount Sinai School of Medicine New York, New York

Rulla M. Tamimi, ScD

xvii

xviii

Contributors

Robert E. Tarone, PhD

David C. Whiteman, BMedSc, MB, PhD

International Epidemiology Institute Rockville, Maryland

Population Studies and Human Genetics Division Queensland Institute of Medical Research Brisbane, Queensland, Australia

Michael J. Thun, MD American Cancer Society Atlanta, Georgia

Inger Thune, MD, PhD Norwegian Cancer Society University of Tromsø Tromsø, Norway

Harri Vainio, PhD International Agency for Research on Cancer Lyon, France

Thomas L. Vaughan, MD, MPH Department of Epidemiology School of Public Health and Community Medicine University of Washington Seattle, Washington

Roel Vermeulen, PhD Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Jaye L. Viner, MD, MPH National Cancer Institute Bethesda, Maryland

Sophia S. Wang, PhD Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, Maryland

Mary H. Ward, PhD Division of Cancer Epidemiology and Genetics National Cancer Institute Bethesda, MD

Erika Waters, MS Rutgers University New Brunswick, New Jersey

Douglas Weed, MD, PhD Division of Cancer Prevention National Cancer Institute Bethesda, Maryland

Qingyi Wei, MD, PhD Department of Epidemiology MD Anderson Cancer Center Houston, Texas

Martin A. Weinstock, MD, PhD Brown University School of Medicine Dermatoepidemiology Unit Providence, Rhode Island

Noel S. Weiss, MD, DrPH Department of Epidemiology School of Public Health and Community Medicine University of Washington Seattle, Washington

Alice S. Whittemore, PhD Department of Health Research and Policy Stanford University School of Medicine Stanford, California

Louise Wideroff, PhD Division of Cancer Control and Population Sciences National Cancer Institute Bethesda, Maryland

Lynne R. Wilkens, DrPH Cancer Research Center of Hawaii University of Hawaii Honolulu, Hawaii

Anna Wilkinson, PhD Department of Epidemiology MD Anderson Cancer Center Houston, Texas

Walter C. Willett, MD, DrPH Department of Nutrition Harvard School of Public Health Boston, Massachusetts

Deborah M. Winn, PhD Division of Cancer Control and Population Sciences National Cancer Institute Bethesda, Maryland

Kana Wu, MD, PhD Department of Nutrition Harvard School of Public Health Boston, Massachusetts

Xifeng Wu, MD, PhD Department of Epidemiology MD Anderson Cancer Center Houston, Texas

K. Robin Yabroff, PhD Division of Cancer Control and Population Sciences National Cancer Institute Bethesda, Maryland

Mimi C. Yu, PhD Cancer Center University of Minnesota Minneapolis, Minnesota

Jian-Min Yuan, MD, PhD Cancer Center University of Minnesota Minneapolis, Minnesota

I BASIC CONCEPTS

This page intentionally left blank

1

Cause and Cancer Epidemiology STEVEN N. GOODMAN AND JONATHAN M. SAMET

T

he prevention of disease has long been based implicitly on taking action on the assumption that a disease is caused by a factor that can be controlled. Early examples include the experimental evidence generated by Lind, showing that consumption of oranges and lemons prevented scurvy, and Snow’s observations on cholera occurrence in London, showing a disease pattern consistent with water-borne transmission (Rosen, 1993). In these examples, preventive steps followed: After Lind’s experiment, the diets of the British navy were supplemented with citrus fruits; and after Snow’s observational study, steps were taken to ensure that the source of water was changed in the affected areas of London. Over the ensuing centuries, infectious agents were causally linked to specific diseases, and prevention was accomplished by interrupting transmission and by vaccines. During the twentieth century, public health was threatened by parallel epidemics of chronic diseases, including cancer; and as the causal agents were identified, a broad range of preventive strategies were implemented. The concept of causation has long had a central role in the application of epidemiologic evidence for controlling cancer. The designation of a risk factor as “causal” has been the starting point for initiating cancer prevention programs based on reducing exposure to the risk factor. Although the concept of causation itself remains a matter of continuing discussion among philosophers and others, use of the term in public health implies that the evidence supporting causality of association has reached a critical threshold of certainty and that reduced exposure can be expected to be followed by reduced disease occurrence. Over the last 50 years, identification of the causes of cancer has been the primary focus of most epidemiologic research on cancer; only recently has attention shifted toward identifying genetic determinants of susceptibility and markers of the early stages of carcinogenesis. There are numerous examples of how identifying a cause of cancer has led to intervention and reduction of cancer occurrence. Tobacco use and cancer of the lung is a notable example for its historical precedence and for the framework applied to the scientific evidence as the causality of the association was evaluated (US Department of Health Education and Welfare—DHEW, 1964; White, 1990). The range of causal risk factors for cancer is broad, including infectious agents (e.g., human papillomavirus and cervical cancer), physical agents (e.g., ionizing radiation and leukemia), inhaled agents (e.g., radon and lung cancer), pharmaceutical agents and hormones (e.g., diethylstilbestrol and adenocarcinoma of the vagina), food contaminants (e.g., aflatoxin and liver cancer), workplace exposures (e.g., asbestos), life stylerelated exposures (e.g., alcohol consumption), and genetic mutations (e.g., Li-Fraumeni syndrome). These and other factors considered to be causes of cancer have been given this label only after the accumulation of sufficient evidence, in most instances derived from both epidemiologic and laboratory research. This chapter provides an overview of causal inference with a focus on the interpretation of epidemiologic data on cancer risk. It begins with an introduction to the centuries-old discussion on cause and causation and next considers the epidemiologic concept of causation, setting the discussion in the context of current understanding of carcinogenesis as a multistep process. The criteria for causation, often attributed to the British medical statistician Sir Austin Bradford Hill (Hill, 1965) or to the 1964 Report of the U.S. Surgeon General on tobacco (US Department of Health Education and Welfare—DHEW, 1964), have provided a framework for evaluating evidence to judge the causality of associations.

These criteria are addressed in depth, and their application is illustrated with the example of smoking, both active and passive, and lung cancer. The chapter concludes with a consideration of emerging issues concerned with causation, including the interpretation of data coming from the new technologies of contemporary “molecular epidemiology” and new approaches to evaluating causation.

CONCEPTS OF CAUSATION At its foundation, “cause” is not knowable with certainty. This fact underlies much of the methodologic and conceptual confusion that often swirls around claims of causation based on scientific data. The fundamental intuition underlying the causal concept is that event “A” somehow produces another event, “B.” However, the “production” of “B” is not observable. The philosopher Bryan Magee summarized this conundrum eloquently. It seems to be impossible for us to form any conception of an ordered world at all without the idea of there being causal connections between events. But when we pursue this idea seriously we find that causal connection is not anything we ever actually observe, nor ever can observe. We may say that Event A causes Event B, but when we examine the situation we find that what we actually observed is Event A followed by Event B. There is not some third entity between them, a casual link, which we also observe. . . . So we have this indispensable notion of cause at the very heart of our conception of the world, and of our understanding of our own experience, which we find ourselves quite unable to validate by observation or experience . . . It actually purports to tell us how specific material events are related to each other in the real world, yet it is not derived from, nor can it be validated by, observation of that world. This is deeply mysterious. (Magee, 2001) The fact that causation is not directly observable means that scientists and philosophers have had to develop a set of constructs and heuristics by which to define a “cause” operationally. These constructs typically have two components: a predictive or associational one, determined empirically, and an explanatory one, based on a proposed underlying mechanism. All causal claims rest on these twin pillars; an association with no plausible mechanistic basis is typically not accepted as causal, and a proposed mechanism, however well founded, cannot be accepted as the basis for a causal claim without empirical demonstration that the effect occurs more often in the presence of the purported cause than in its absence. However, these components need not contribute equally, and various causal claims may rest on quite different balances of contributions of empirical and mechanistic information. Underlying any operational definition of causality must be an ontologic one: that is, how a cause is defined in principle. A particularly useful, widely accepted definition in both philosophy and epidemiology is the “counterfactual” notion of causation. This concept had its origins at least as far back as the English philosopher David Hume (1711–1776) (Hume, 1739). During the twentieth century, this concept was further developed and applied by statisticians, philosophers, and epidemiologists (Bunge, 1959; Lewis, 1973; Rubin, 1974; Robins, 1986, 1987; Greenland, 1990; Neyman, 1990; Greenland et al., 1999;

3

4

PART I: BASIC CONCEPTS

Pearl, 2000). The counterfactual definition holds that something is a cause of a given outcome if, when the same individual is observed with and without a purported cause and without changing any other characteristic of that individual, a different outcome would be observed. For example, the counterfactual state for a smoker is the same individual never having smoked. The state that cannot be observed is called the counterfactual state: literally, counter to the observed facts. The impossibility of observing the counterfactual state is what makes all causal claims subject to uncertainty. The above definition is deterministic; that is, the outcome always occurs in the presence of the cause and never occurs without it. However, health research rarely deals with either a cause that inevitably produces certain outcomes, or outcomes that cannot occur absent specific causes. For example, smokers do not always get lung cancer, and never-smokers do develop this malignancy. Therefore, the counterfactual definition must be expanded to encompass the notion of a probabilistic outcome. That is, the formal definition of a cause in epidemiology requires that a factor X be associated with a difference in the probability of an outcome. For example, if X may take on two different values, y or z: Condition 1: observed association Pr(outcome | X = y) π Pr(outcome | X = z) Properly designed studies provide a scientific basis for inferring what the outcome of the counterfactual state would be and permit the related uncertainty to be quantified. In a laboratory, scientists are able to predict the outcome in this counterfactual state, generally with a high degree of confidence, by repeating an experimental procedure with every factor tightly controlled, varying only the factor of interest. In observational studies of humans, however, researchers must try to infer what the outcome would be in a counterfactual state by studying another group of persons who, at least on average, are substantively different from the exposed group in only one variable: the exposure under study. The outcome of this second group is used to represent what would have occurred in the original group if it were observed with an exposure different from that which actually existed (Greenland, 1990). In the case of smoking and disease, this comparison is between disease risk in smokers and nonsmokers. Simply observing a difference in the probability of an outcome between two groups that differ on X is not sufficient condition for causation because it does not distinguish between causation and spurious or indirect association, produced by “confounders,” or ancillary causes. The notion of “causation” requires that the cause somehow actively “produce” its effect, which is captured operationally by the requirement that active manipulation of the cause should produce a change in the probability of the outcome. For example, if one saw that students with poor visual acuity typically sat closer to the front of a classroom, one would not call the seating arrangement a “cause” of their poor eyesight unless it could be shown that seating them farther back improved it. The notation that captures this idea is one that introduces an operator, not part of traditional statistical notation “Set (X = x),” which corresponds to actively setting a risk factor X equal to some value x, rather than simply observing that the factor is equal to x. Thus the counterfactual notion of probabilistic causation for a risk factor X requires condition 2. Condition 2: no confounding Pr(outcome | set[X = x)] = Pr(outcome | X = x) If we put together condition 1—that there is an observed association between cause and effect—with condition 2—that there is no other indirect cause responsible for the observed effect—we have the counterfactual condition for probabilistic causation, expressed as follows. Condition 1 + Condition 2 = Causality condition Pr[outcome | set(X = y)] π Pr[outcome | set(X = z)] This condition states that if the probability of an outcome changes when risk factor X is actively changed from z to y, then X is regarded as a cause of the outcome.

In the randomized controlled trial, a risk factor is actively manipulated. Understanding the role of randomization can deepen insights into the interpretation of nonrandomized designs used in epidemiology. Randomization has two critical consequences: (1) it makes exposure to a proposed causal factor independent of potentially confounding factors; and (2) it provides a known probability distribution for the potential outcomes in each group under a given mathematical hypothesis (i.e., the null) (Greenland, 1990). Randomization does not necessarily free the inference from an individual randomized study from unmeasured confounding (it does so only on average). Randomization does imply that measures of uncertainty about causal estimates from randomized studies have an experimental foundation. In the absence of randomization, uncertainty about causal effects depends in part on the confidence that all substantive confounding has been eliminated or controlled by either the study design or the analysis. The level of confidence is ultimately based on scientific judgment and consequently is subject to uncertainty and questioning. One way to increase that confidence is to repeat the study. Similar results in a series of randomized studies make it increasingly unlikely that unmeasured confounding is accounting for the findings, as the process of randomization makes the mathematical probability of such confounding progressively smaller as the sample size or number of studies increase. In observational studies, however, increasing the number of studies may reduce the random component of uncertainty, but not necessarily the systematic component attributable to confounding. Without randomization, there is no mathematical basis for assuming that an imbalance of unknown confounders decreases with an increase in the number of studies. However, if observational studies are repeated in different settings with different persons, different eligibility criteria, and/or different exposure opportunities, each of which might eliminate another source of confounding from consideration, the confidence that unmeasured confounders are not producing the findings is increased. How many studies need to be done, how diverse they need to be, and how relevant they are to the question at hand are matters of scientific judgment, and explicit criteria cannot be offered. Confidence that unmeasured confounding is not producing the observed results is further increased by understanding the biologic process by which the exposure might affect the outcome. This understanding allows better identification and measurement of relevant confounders, making it more unlikely that unmeasured factors are of concern. Biologic understanding can also serve as the basis for a judgment that the observed difference in outcome frequency could be produced only by an implausible degree of confounder imbalance between exposed and unexposed groups. Thus, causal conclusions from observational studies typically require more and stronger biologic evidence to support plausibility and to exclude confounding than is needed for causal inferences based on randomized studies.

COMPONENT CAUSE MODEL Causes defined in the manner described above can be viewed as working together in many ways. In 1976 Rothman proposed a useful framework for considering multiple-cause diseases (Rothman, 1976) that has ready extension to the causation of cancer, as most cancers have several causes. Rothman proposed that a disease may have several sufficient causes, each accounting for some proportion of the cases in the population, and that each cause may have several components (Fig. 1–1) (Rothman, 1976). With this model, each component of the three complete causes must be present for disease to develop. For example, sufficient cause I would be incomplete if A were not present; and because A is a component of each of the three complete causes, it is a necessary cause. Rothman’s model is useful for considering causation of malignancy, particularly for most of the cancers for which multiple genetic and environmental risk factors may play a role in a multistage process that transforms a normal cell into a malignant cell. There may be multiple ways to complete this sequence of changes, involving the actions of different complexes of factors, analogous to complete causes I, II, and III in Figure 1–1. The sufficient causes might include multiple envi-

Cause and Cancer Epidemiology Sufficient Cause I

Sufficient Cause II

Sufficient Cause III

5

Table 1–1. Henle-Koch Postulates 1. The parasite occurs in every case of the disease in question and under circumstances that can account for the pathologic changes and clinical course of the disease. 2. It occurs in no other disease as a fortuitous and nonpathogenic parasite. 3. After being fully isolated from the body and repeatedly grown in pure culture, it can induce the disease anew. Source: Evans (1993).

Figure 1–1. Conceptual scheme for the causes of a hypothetical disease. (Source: Rothman, 1976.)

ronmental and genetic risk factors, including environmental exposures and genes determining carcinogen metabolism and DNA repair. Individual cases would result from having the full complement of components for a complete cause. We know, for example, that cigarette smoking is a powerful cause of lung cancer, but not all smokers develop lung cancer, implying that this factor may need to act in combination with other factors, perhaps genetic, to complete one sufficient cause for lung cancer. Some sufficient causes for lung cancer do not include smoking, as some percentage of lung cancer cases occur in never-smokers (about 5%–10% of cases in the United States at present) (Alberg and Samet, 2003). Rothman’s model has one significant implication for considering the burden of cancer attributable to a particular risk factor, a calculation often made when assigning priorities to prevention initiatives. The presence of several components in the same complete causes (e.g., A and B in causes I and II) implies that the cases associated with these causes might be prevented by eliminating exposure to either A or B. The burden of disease to be prevented exceeds 100%, as the attributable risk estimates for A and B would include some of the same cases. In some past reviews of the burden of preventable cancer, the assumption was made incorrectly that the attributable risks associated with various causal factors should add up to 100% (Samet and Lerchen, 1984).

CRITERIA FOR CAUSALITY Epidemiologists and other public health researchers have needed pragmatic definitions of causation to support the translation of research evidence into interventions directed at reducing the exposure to causal risk factors (Susser, 1973, 1991; Olsen, 2003). The epidemiologic literature has consequently placed great emphasis on the approach to evaluating evidence to determine if a factor can be considered to cause disease. The approaches that have been developed for evaluating causality of associations also draw on multiple lines of scientific evidence; epidemiologic evidence alone is generally not regarded as sufficient for establishing causality (Last, 2000). Making causal inferences from observational data, in combination with other relevant forms of data, can be a challenging task that requires expert judgment regarding the likely sources and magnitude of confounding, together with judgment about how well the existing constellation of study designs, results, and analyses address this potential threat to inferential validity. This judgment also needs to incorporate a broader assessment of the evidence, evaluating whether a causal effect has support in the existing knowledge of the underlying biologic process. To aid this judgment, criteria for determining a cause have been proposed by many philosophers and scientists over the centuries. In biomedical research, the first criteria came following the discovery of bacteria during the nineteenth century. A method was then needed for judging if an organism caused a particular disease. The first criteria put forward for making this judgment are generally attributed to Robert Koch and his mentor Jacob Henle, although Koch also acknowledged Eugene Klebs. Evans (1993) provided a full accounting of the elaboration of these criteria, now referred to as the Henle-

Koch postulates (Table 1–1). The criteria proved valuable for linking infectious agents to infectious diseases, which often have specific clinical features and unique, specific causal agents (e.g., pulmonary tuberculosis and Mycobacterium tuberculosis). These criteria, however, proved unsuitable for establishing the causes of the epidemics of “chronic disease,” including coronary heart disease, chronic lung disease, and cancer, that became the dominant causes of death spanning the twentieth century, as infectious diseases were controlled. Unlike many infectious diseases, these diseases were often found to be associated with multiple factors, and many cases could not be clearly linked to any risk factors. The limitations of the Henle-Koch postulates were recognized as the results of the first wave of epidemiologic studies on the chronic diseases were reported. In 1959, Yerushalmy and Palmer proposed criteria for evaluating possible etiologic risk factors for chronic diseases that acknowledged the need for evidence of increased risk in exposed persons and for handling the nonspecific causation of these diseases. Lilienfeld (1959) and Sartwell (1960), discussing the article, added the consideration of dose-response, the strength of the association, its consistency, and its biologic plausibility. The most widely cited criteria in epidemiology and public health more generally were set forth by Sir Austin Bradford Hill in 1965 (Table 1–2) (Hill, 1965). Five of the nine criteria he listed were also put forward in the 1964 Surgeon General’s report (US Department of Health Education and Welfare—DHEW, 1964) as the criteria for causal judgment: consistency, strength, specificity, temporality, and coherence of an observed association. Hill also listed biologic gradient (dose-response), plausibility, experiment (or natural experiment), and analogy. Many of these criteria had been cited in earlier epidemiologic writings (Lilienfeld, 1959; Yerushalmy and Palmer, 1959; Sartwell, 1960); and Susser and others have refined them extensively by exploring their justification, merits, and interpretations (Susser, 1973, 1977; Kaufman and Poole, 2000). Hill (1965) clearly stated that these criteria were not intended to serve as a checklist. Here are then nine different viewpoints from all of which we should study association before we cry causation. What I do not believe . . . is that we can usefully lay down some hard-and-fast rules of evidence that must be obeyed before we accept cause and effect. None of my nine viewpoints can bring indisputable evidence for or against the cause-and-effect hypothesis, and none can be required as a sine qua non. What they can do, with greater or less strength, is to help us to make up our minds on the fundamental

Table 1–2. Sir Austin Bradford Hill’s Causal Criteria: Aspects of Association to Be Considered Before Deciding on Causation Strength Consistency Specificity Temporality Biologic gradient Plausibility Coherence Experiment Analogy Source: Hill (1965).

6

PART I: BASIC CONCEPTS question—is there any other way of explaining the facts before us, is there any other answer equally, or more, likely than cause and effect? (Hill 1965, p. 299)

All of these criteria were meant to be applied to evidence related to an already established statistical association; if no association has been observed, these criteria are not relevant. Hill explained how if a given criterion were satisfied it strengthened a causal claim. Each of these nine criteria served one of two purposes: as evidence against competing noncausal explanations or as positive support for causal ones. Noncausal explanations for associations include chance; residual or unmeasured confounding; model misspecification; selection bias; errors in measurement of exposure, confounders, or outcome; and issues regarding missing data (which can also include missing studies, such as publication bias). The criteria are discussed below.

Consistency The criterion consistency refers to the persistent finding of an association between exposure and outcome in multiple studies of adequate power and carried out by different investigators in studies involving different persons, places, circumstances, and times. Consistency can have two implications for causal inference. First, consistent findings make unmeasured confounding an unlikely alternative explanation to causation for an observed association. Such confounding would have to persist across diverse populations, exposure opportunities, and measurement methods. The confounding is still possible if the exposure of interest were strongly and universally tied to an alternative cause, as was claimed in the form of the “constitutional hypothesis” put forward in the early days of the smoking-disease debate (US Department of Health Education and Welfare—DHEW, 1964). This hypothesis held that there was a constitutional (i.e., genetic) factor that made people more likely to both smoke and develop cancer. Thus, consistency serves mainly to exclude the possibility that the association is produced by an ancillary factor that differs across studies but not one factor that is common to all or most of them (Rothman and Greenland, 1998). The second implication of the consistency criterion is to reduce the possibility of a chance effect by increasing the statistical strength of an association through the accumulation of a large body of data. Consistency does not include the qualitative strength of such studies, which Susser subsumed under his subsidiary concept of “survivability,” relating to the rigor and severity of tests of association (Susser, 1991).

Strength of Association Strength of association includes two dimensions: the magnitude of the association and its statistical strength. An association strong in both aspects makes the alternative explanations of chance and confounding unlikely. The larger the measured effect, the less likely it is that an unmeasured or poorly controlled confounder could account for it completely. Associations that have a small magnitude or weak statistical strength are more likely to reflect chance, a modest degree of bias, or unmeasured weak confounding. However, the magnitude of association is reflective of underlying biologic processes and should be consistent with understanding the role of the risk factor in these processes. Either a strong or a weak effect might be considered plausible based on knowledge of the underlying processes. In the example of active smoking and lung cancer, the relative risks listed in the first Surgeon General’s Report (US Department of Health Education and Welfare—DHEW, 1964) were notably elevated in men, reaching as high as 10 or more. At that time, other causes of lung cancer, including air pollution and occupational agents, had been identified. However, for the general population, the risks from these factors were far lower, making them unsatisfactory as potential confounders, leading to the observed association of active smoking with lung cancer. Passive smoking, by contrast, has a far smaller effect on lung cancer risk. Comparing persons with greater and lesser exposures (e.g., never-

smoking women married to smokers compared with never-smoking women married to never-smokers). The 1986 report of the U.S. Surgeon General (US Department of Health and Human Services— USDHHS, 1986) concluded that passive smoking does cause lung cancer. The magnitude of the effect was small in most of the studies, as anticipated on a biologic basis, but within a plausible range. The relative risk associated with marriage to a smoker has been estimated to be 1.2 in a recent meta-analysis (International Agency for Research on Cancer—IARC, 2002).

Specificity Specificity has been interpreted to mean both a single (or few) effect(s) of one cause or no more than one possible cause for one effect. In addition to specific infectious diseases caused by specific infectious agents, other examples include asbestos exposure and mesothelioma and thalidomide exposure during gestation and the resulting unusual constellation of birth defects. This criterion is rarely used as it was originally proposed, having been derived primarily from the HenleKoch postulates for infectious causes of disease (Susser, 1991). When specificity exists, it can strengthen a causal claim, but its absence does not weaken it (Sartwell, 1960). For example, most cancers are known to have multifactorial etiologies; many cancer-causing agents can cause several types of cancer, and these agents can also have noncancerous effects. When considering specificity in relation to the smoking–lung cancer association, the 1964 Surgeon General’s report (US Department of Health Education and Welfare—DHEW, 1964) provides a rich discussion of this criterion. The committee recognized the linkage between this criterion and strength of association and offered a symmetrical formulation of specificity in the relation between exposure and disease; that is, a particular exposure always results in a particular disease, and the disease always results from the exposure. The committee acknowledged that smoking does not always result in lung cancer and that lung cancer has other causes. The report noted the extremely high relative risk for lung cancer in smokers and the high attributable risk, and it concluded that the association between smoking and lung cancer has “a high degree of specificity.”

Temporality Temporality refers to the occurrence of a cause before its purported effect. Temporality is the sine qua non of causality, as a cause clearly cannot occur after its purported effect. Rothman (1986) emphasized that temporality is the only one of the criteria that must be fulfilled for an association to be considered causal. Any question about a temporal sequence seriously weakens a causal claim; but establishing temporal precedence is by itself not strong evidence in favor of causality.

Coherence, Plausibility, and Analogy Although the original definitions of coherence, plausibility, and analogy were subtly different, in practice they have been treated essentially as one idea: that a proposed causal relation should not violate known scientific principles, and that it be consistent with experimentally demonstrated biologic mechanisms and other relevant data, such as ecologic patterns of disease (Rothman and Greenland, 1998). In addition, if biologic understanding can be used to set aside explanations other than a causal association, it offers further support for causality. Together, these criteria can serve to both support a causal claim (by supporting the proposed mechanism) and refute it (by showing that the proposed mechanism is unlikely). Biologic understanding, of course, is always evolving as scientific advances make possible ever deeper exploration of disease pathogenesis. For example, in 1964 the Surgeon General’s committee found the causal association of smoking with lung cancer to be biologically plausible based on knowledge of the presence of carcinogens in tobacco smoke and animal experiments. Nearly 40 years later, this association remains biologically plausible, but that determination rests not only on the earlier evidence but on more recent findings that address the

Cause and Cancer Epidemiology genetic and molecular basis of carcinogenesis, providing a level of understanding that could not have been anticipated in 1964.

Biologic Gradient (Dose-Response) The finding of a graded increase in effect with an increase in the strength of the possible cause provides strong positive support in favor of a causal hypothesis. This is not just because such an observation is predicted by many cause-and-effect models and biologic processes but, more importantly, because it makes most noncausal explanations highly unlikely. If some factor other than that of interest explains the observed gradient, the unmeasured factor must change in the same manner as the exposure of interest. Except for confounders that are closely related to a causal factor, it is extremely difficult for such a pattern to be created by virtually any of the noncausal explanations for an association listed earlier. The finding of a dose-response relation has long been a mainstay of causal arguments in smoking investigations; virtually all health outcomes causally linked to smoking have shown an increase in risk and/or severity with an increase in the lifetime smoking history. This criterion is not based on any specific shape of the dose-response relation.

Experiment The criterion “experiment” refers to situations where natural conditions might plausibly be thought to imitate conditions of a randomized experiment, producing a “natural experiment” whose results might have the force of a true experiment. An experiment is typically a situation in which a scientist controls who is exposed in a way that does not depend on any of the subject’s characteristics. Sometimes nature produces similar exposure patterns. The reduced risk after smoking cessation serves as one such situation that approximates an experiment; an alternative noncausal explanation might posit that an unmeasured causal factor of that health outcome was more frequent among those who did not stop smoking than among those who did. The causal interpretation is further strengthened if risk continues to decline in former smokers with increasing time since quitting. Similar to the dose-response criteria, observations of risk reduction after quitting smoking have the dual effects of making most noncausal explanations unlikely and supporting the biologic model that underlies the causal claim.

APPLYING THE CAUSAL CRITERIA The greater the extent to which an association fulfills the previous criteria, the more difficult it is to offer a more compelling alternative explanation. Which of these criteria may be more important and whether some can be unfulfilled and still justify the causal claim is a matter of judgment. Temporality, however, cannot be violated. When there is a still incompletely understood pathogenic mechanism, the causal claim might still be justified by strong direct empirical evidence of higher lung cancer rates in smokers (i.e., strong, consistent associations). Moderate associations (e.g., relative risk of 1–2) in only a few studies, without adequate understanding of potential confounders or with weak designs, might result in a suspicion of causal linkage. The process of applying the criteria extends beyond simply lining up the evidence against each criterion, although there is evidence that epidemiologists tend to use the evidence in neither a consistent nor comprehensive manner (Weed and Gorelic, 1996). Rather, the criteria should be used to integrate multiple lines of evidence coming from chemical and toxicologic characterizations of tobacco smoke and its components, epidemiologic approaches, and clinical investigations. Those applying the criteria weigh the totality of the evidence in a decision-making process that synthesizes and, of necessity, involves a multidisciplinary judgment. The 1964 Surgeon General’s report still stands as one of the finest examples of the power of applying these criteria systematically and comprehensively. Starting with the criterion for consistency, the committee noted that all 29 retrospective studies (i.e., case-control) and 7

7

prospective studies (i.e., cohort) at the time reported highly significant smoking–lung cancer relations. They further noted that all of the studies comparing smokers to nonsmokers showed high relative risks for lung cancer (~10). Dose-response effects were also observed in every prospective study and in all retrospective studies where it could be calculated. The temporal sequence was reported to be not absolutely certain but seemed highly unlikely to be in the lung cancer–smoking direction, as cancer typically appears many years or decades after the onset of smoking. With regard to coherence of the association with known facts, the studies noted the ecologic increase in lung cancer rates with increased smoking in the population; the gender differential in lung cancer, which at the time was consistent with more smoking by men; an urban-rural difference, which air pollution could not completely explain; socioeconomic differentials in lung cancer, for which smoking seemed to be the strongest explanation; and the localization of cancer in the respiratory tract in relation to the type of smoking. The studies also cited the known reduction in risk among former smokers, with greater risk reductions correlated with more time spent not smoking. These observations, in combination with histopathologic evidence, basic biologic observations, and an in-depth discussion of each competing non-smoking-related explanation (e.g., occupation, constitutional hypothesis, infections, environmental factors such as pollution), produced a case for causation that proved irrefutable. The 1986 Report of the Surgeon General (US Department of Health and Human Services—USDHHS, 1986) concluded that passive smoking causes lung cancer, a conclusion that has proved momentous in its implications. This report also based its evaluation of the evidence on the causal criteria. A clear distinction was made between the evidence on active smoking and that expected from the much lower carcinogen doses arising from passive smoking. Biologic plausibility was emphasized, including the substantial evidence on lung cancer risk in active smokers. This causal conclusion has been reaffirmed in all subsequent reports (Samet and Wang, 2000; International Agency for Research on Cancer—IARC, 2002).

EMERGING ISSUES IN CAUSAL INFERENCE AND CANCER Perhaps the most challenging and exciting issue facing cancer scientists now and in the future is the prospect of understanding the processes of cancer development at the molecular level. However, with a richer understanding of basic mechanisms comes concomitant complexity in the concept and determination of cause. Biomarkers can serve as indicators of exposure, dose, susceptibility, or effect, each of which can elevate the cancer risk, albeit via quite different routes (Links et al., 1995). Similarly, the mechanisms by which various genes affect cancer risk are diverse, from genes that directly modulate tumor growth to others responsible for cellular homeostasis, DNA repair, genetic stability, or a host of interrelated functions that protect the cell against damage from somatic or environmental factors or affect its repair capacity when damage occurs (Vineis and Porta, 1996; Hussain and Harris, 1998). It is interesting to consider the implications of this kind of knowledge for causal inference in cancer. The most obvious change is that we now understand the biologic basis of action of long-established carcinogens, such as smoking, chemotherapy, and various chemical agents. Mechanistic explanations of how environmental exposures have a carcinogenic effect provide the basis for increased confidence that any given association between the exposure and cancer incidence is justifiably labeled causal. Molecular “signatures” of specific exposures (e.g., p53 CpG hotspots) (Greenblatt et al., 1994) are defining the relevant effects of certain exposures more precisely and are also making increasingly possible what could not be done before: establish causal connections between exposure and disease on the individual level. Second, by understanding better what mediates risk due to exposures, we are increasingly able to identify subpopulations of individuals at substantially different degrees of risk from an exposure (Shields

8

PART I: BASIC CONCEPTS

and Harris, 2000). This risk heterogeneity can be due to genetic or somatic variations that affect the absorption, metabolism, or cellular effect of an agent or the host’s susceptibility to those effects (Links et al., 1995). The smaller size of these subpopulations, however, poses greater challenges to epidemiologic approaches to risk assessment: the finer the risk stratification based on mechanistic criteria, the more difficult it is to confirm it with epidemiologic methods (Slattery, 2002; Moore et al., 2003). Third, with the opening of the cancer “black box” comes greater recognition of the extraordinary complexity of the carcinogenic process at the molecular level. This is seen not only with external agents dependent on genes for their effects but with environmental modification of gene expression (e.g., through DNA methylation) (Moore et al., 2003), gene functions interacting in myriad ways, and perhaps most intriguingly genetic determinants of exposure, as shown with developing understanding of the biologic bases of nicotine, alcohol, and drug addiction (Shields et al., 1993; Greenblatt et al., 1994; Hemby, 1999). This complexity poses significant problems for population-based risk assessments and causal inference because it raises, more acutely than with conventional risk factors, questions about which genes lie in the causal pathway (those that do should not be considered covariates), which genes or biomarkers are necessary for the effects or function of others (i.e., there may be many biologically based high level interaction terms), whether an observed association represents a direct or indirect causal effect, and whether it is meaningful to talk about binary exposure–outcome or gene–outcome relations when the components of these complex networks either cannot be changed individually or changed at all (Taioli and Garte, 2002). Fourth, although the functions of many genes and gene products are understood, many more are not. The current ability to perform mass screening of potential causal factors via high throughput, genomic and proteomic technologies is far outstripping our ability to explain what we find; and severe problems related to multiplicity (“data dredging”) arise in many of these investigations (Reiner et al., 2003). Epidemiologists have long faced this problem but never on this scale, with literally hundreds or even thousands of potential markers measured and analyzed, in all combinations, sometimes in only tens or hundreds of subjects. The weakness of findings subject to these problems is not often fully appreciated, and researchers are making claims about possibilities of screening, treatment, or prevention before findings are replicated in independent data sets. A final issue that is often neglected in this rush toward exploring new potential causes is that of measurement (Little et al., 2002). Many of the techniques and assays used to identify metabolic products, genes, and gene products are relatively new and not standardized. Understanding the reliability and validity of these techniques is an arduous process that often does not get sufficient attention, yet it is critical for distinguishing likely spurious claims from those that are well grounded. As we come closer to understanding cancer on the molecular level, many have raised the possibility of individual risk prediction (Vineis, 1997; Hussain et al., 2001). Some commentators have suggested that mechanistic understanding may ultimately render population studies irrelevant. However, it is instructive to consider the case of infectious diseases. Understanding the basic mechanisms of infectious disease has not eliminated the need to study disease patterns on a population level, and the same will likely be true for cancer (Nevins et al., 2003). There are many reasons to believe that population-based studies will be as important in the future as they are now, albeit perhaps focused on different kinds of questions. To understand a mechanism after a cancer occurs is not to predict it; early steps in the process, when the disease can be prevented, will necessarily be less than 100% predictive; and defining optimal groups for screening and early intervention will still require population-based data. Risk groups must be defined using far fewer factors than we know are operating at the molecular level, the latter being almost unique for an individual. The molecular revolution in cancer may ultimately force a merging of two “schools” of causal inference: the probabilistic, chronic disease model that the Hill criteria addressed and the more mechanistic, deter-

ministic models used for infectious disease, for which the Henle-Koch criteria were devised (Fredericks and Relman, 1996). Nowhere is this seen more clearly than with viral carcinogenesis. The Henle-Koch criteria were based on the nineteenth century understanding of bacterial disease causation and are poorly suited for viral disease mechanisms or even for infectious disease outcomes. Efforts to refashion the HenleKoch criteria for the new era of molecular medicine (Fredericks and Relman, 1996; Vineis and Porta, 1996) have shown that it is extraordinarily difficult to outline a set of experimental conditions that all known pathogens—not to mention new pathogens with different mechanisms—must satisfy to justify causal claims for new infectious diseases. In the case of viral carcinogenesis, the situation is even more complex because the final disease is not a direct manifestation of an infectious process. Numerous viral agents have been linked to cancer with varying degrees of certainty: Epstein-Barr virus and Burkitt’s lymphoma (Pagano, 1999), human papillomavirus and cervical cancer (Bosch and de Sanjose, 2003), SV40 and mesothelioma (Carbone et al., 1997; Klein et al., 2002). The evidential basis of these claims includes findings that might implicate the virus in individual cancer cases (e.g., finding viral genetic sequences or viral-specific proteins in tumor tissue) and traditional epidemiologic evidence (e.g., high incidence in persons with evidence of viral exposure). However, as both these and other examples have shown, there is no single molecular finding that definitively implicates a virus as a cause of cancer. The variety and complexity of mechanisms by which viruses can directly (by inducing oncogenic changes) or indirectly (by increasing host susceptibility to exogenous agents) raise cancer risk seems to defy a set of causal criteria designed specifically for viral agents or that are based on any specific mechanism. Therefore, from the standpoint of causal inference, it may be best to consider viral agents under the same umbrella as toxic exposures and other environmental causes of cancer. The fact that the pathways from viral infection to cancer appearance often share components (e.g., p53 inactivation) with those of noninfectious carcinogenic exposures supports this view. It seems unlikely that the need for population-based risk estimates and causal inference will disappear from cancer research, just as it has not in infectious disease. However, we are entering an era where the relative strength of the “twin pillars” of causal inference—knowledge derived from empirical, population-based patterns and that based on understanding of biologic mechanisms in individuals—will tilt farther toward the mechanistic end, requiring less proof from populations and more from the laboratory. One challenge for causal inference in the future will be how best to integrate these various forms of evidence and how to assemble groups with the sufficient interdisciplinary expertise to assess them. References Alberg AJ, Samet JM. 2003. Epidemiology of lung cancer. Chest 123:21S–49S. Bosch FX, de Sanjose S. 2003. Chapter 1: Human papillomavirus and cervical cancer—burden and assessment of causality. J Natl Cancer Inst Monogr 31:3–13. Bunge M. 1959. Causality: The Place of the Causal Principle in Modern Science. Cambridge, MA: Harvard University Press. Carbone M, Rizzo P, Pass HI. 1997. Simian virus 40, poliovaccines and human tumors: a review of recent developments. Oncogene 15:1877–1888. Evans AS. 1993. Causation and Disease: A Chronological Journey. New York: Plenum. Fredericks DN, Relman DA. 1996. Sequence-based identification of microbial pathogens: a reconsideration of Koch’s postulates. Clin Microbiol Rev 9:18–33. Greenblatt MS, Bennett WP, Hollstein M, Harris CC. 1994. Mutations in the p53 tumor suppressor gene: clues to cancer etiology and molecular pathogenesis. Cancer Res 54:4855–4878. Greenland S. 1990. Randomization, statistics, and causal inference. Epidemiology 1:421–429. Greenland S, Robins JM, Pearl J. 1999. Confounding and collapsibility in causal inference. Stat Sci 14:29–46. Hemby SE. 1999. Recent advances in the biology of addiction. Curr Psychiatry Rep 1:159–165. Hill AB. 1965. The environment and disease: association or causation? Proc R Soc Med 58:295–300.

Cause and Cancer Epidemiology Hume D. 1739. A Treatise of Human Nature. London: Oxford University Press. Hussain SP, Harris CC. 1998. Molecular epidemiology of human cancer: contribution of mutation spectra studies of tumor suppressor genes. Cancer Res 58:4023–4037. Hussain SP, Hofseth LJ, Harris CC. 2001. Tumor suppressor genes: at the crossroads of molecular carcinogenesis, molecular epidemiology and human risk assessment. Lung Cancer 34(Suppl 2):S7–S15. International Agency for Research on Cancer (IARC). 2002. Tobacco smoke and involuntary smoking. IARC Monograph 83. Lyon: IARC. Kaufman JS, Poole C. 2000. Looking back: causal thinking in the health sciences. Annu Rev Public Health 21:101–119. Klein G, Powers A, Croce C. 2002. Association of SV40 with human tumors. Oncogene 21:1141–1149. Last JM. 2000. A Dictionary of Epidemiology. New York: Oxford University Press. Lewis D. 1973. Counterfactuals. Cambridge, MA: Harvard University Press. Lilienfeld AM. 1959. On the methodology of investigations of etiologic factors in chronic diseases: some comments. J Chronic Dis 10:41–46. Links JM, Kensler TW, Groopman JD. 1995. Biomarkers and mechanistic approaches in environmental epidemiology. Annu Rev Public Health 16:83–103. Little J, Bradley L, Bray MS, Clyne M, Dorman J, Ellsworth DL, Hanson J, Khoury M, Lau J, O’Brien TR, Rothman N, Stroup D, Taioli E, Thomas D, Vainio H, Wacholder S, Weinberg C. 2002. Reporting, appraising, and integrating data on genotype prevalence and gene-disease associations. Am J Epidemiol 156:300–310. Magee B. 2001. The Great Philosophers: An Introduction to Western Philosophy. Oxford, UK: Oxford University Press. Moore LE, Huang WY, Chung J, Hayes RB. 2003. Epidemiologic considerations to assess altered DNA methylation from environmental exposures in cancer. Ann NY Acad Sci 983:181–196. Nevins JR, Huang ES, Dressman H, Pittman J, Huang AT, West M. 2003. Towards integrated clinico-genomic models for personalized medicine: combining gene expression signatures and clinical factors in breast cancer outcomes prediction. Hum Mol Genet 12(Spec. No. 2):R153–R157. Neyman J. 1990. On the application of probability theory to agricultural experiments: essay on principles (1923). Stat Sci 5:463–572. Olsen J. 2003. What characterises a useful concept of causation in epidemiology? J Epidemiol Community Health 57:86–88. Pagano JS. 1999. Epstein-Barr virus: the first human tumor virus and its role in cancer. Proc Assoc Am Physicians 111:573–580. Pearl J. 2000. Causality: Models, Reasoning and Inference. Cambridge, UK: Cambridge University Press. Reiner A, Yekutieli D, Benjamini Y. 2003. Identifying differentially expressed genes using false discovery rate controlling procedures. Bioinformatics 19:368–375. Robins J. 1986. A new approach to causal inference in mortality studies with sustained exposure periods: applications to control of the healthy worker survivor effect. Mathematical Modelling 7:1393–1512. Robins J. 1987. A graphical approach to the identification and estimation of causal parameters in mortality studies with sustained exposure periods. J Chronic Dis 40:139S–161S.

9

Rosen G. 1993. A History of Public Health. Baltimore, MD: The Johns Hopkins University Press. Rothman KJ. 1976. Causes. Am J Epidemiol 104:587–592. Rothman KJ. 1986. Interactions Between Causes. Modern Epidemiology. Boston: Little, Brown. Rothman KJ, Greenland S. 1998. Modern Epidemiology. Philadelphia: Lippincott-Raven. Rubin D. 1974. Estimating causal effects of treatments in randomized and nonrandomized studies. J Educ Psychol 66:688–701. Samet JM, Lerchen ML. 1984. Proportion of lung cancer caused by occupation: a critical review. In: Bernard J, Gee L, Keith W, Morgan C, Brooks SM, editors. Occupational Lung Disease. New York: Raven Press, pp. 55–67. Samet JM, Wang SS. 2000. Environmental tobacco smoke. In: Lippmann M, editor. Environmental Toxicants: Human Exposures and Their Health Effects. New York: Van Nostrand Reinhold, pp. 319–375. Sartwell PE. 1960. On the methodology of investigations of etiologic factors in chronic diseases: further comments. J Chronic Dis 11:61–63. Shields PG, Harris CC. 2000. Cancer risk and low-penetrance susceptibility genes in gene-environment interactions. J Clin Oncol 18:2309–2315. Shields PG, Caporaso NE, Falk RT, Sugimura H, Trivers GE, Trump BF, Hoover RN, Weston A, Harris CC. 1993. Lung cancer, race, and a CYP1A1 genetic polymorphism. Cancer Epidemiol Biomarkers Prev 2:481–485. Slattery ML. 2002. The science and art of molecular epidemiology. J Epidemiol Community Health 56:728–729. Susser M. 1973. Causal Thinking in the Health Sciences: Concepts and Strategies in Epidemiology. New York: Oxford University Press. Susser M. 1977. Judgement and causal inference: criteria in epidemiologic studies. Am J Epidemiol 105:1–15. Susser M. 1991. What is a cause and how do we know one? A grammar for pragmatic epidemiology. Am J Epidemiol 133:635–648. Taioli E, Garte S. 2002. Covariates and confounding in epidemiologic studies using metabolic gene polymorphisms. Int J Cancer 100:97–100. US Department of Health and Human Services (USDHHS). 1986. The Health Consequences of Involuntary Smoking: A Report of the Surgeon General. DHHS Publ. No. (CDC) 87-8398. Washington, DC: U.S. Government Printing Office. US Department of Health Education and Welfare (DHEW). 1964. Smoking and Health. Report of the Advisory Committee to the Surgeon General. DHEW Publ. No. (PHS) 1103. Washington, DC: U.S. Government Printing Office. Vineis P. 1997. Sources of variation in biomarkers. IARC Sci Publ 142: 59–71. Vineis P, Porta M. 1996. Causal thinking, biomarkers, and mechanisms of carcinogenesis. J Clin Epidemiol 490:951–956. Weed DL, Gorelic LS. 1996. The practice of causal inference in cancer epidemiology. Cancer Epidemiol Biomarkers Prev 5:303–311. White C. 1990. Research on smoking and lung cancer: a landmark in the history of chronic disease epidemiology. Yale J Biol Med 63:29–46. Yerushalmy J, Palmer C. 1959. On the methodology of investigations of etiologic factors in chronic diseases. J Chronic Dis 10:27–40.

2

Morphologic and Molecular Classification of Human Cancer THOMAS J. GIORDANO

T

umor morphology, assessed by light microscopic examination of stained tissue sections (the essence of surgical pathology practice), has been the foundation for the pathologic assessment of human cancer for more than 100 years. The durability of morphology and surgical pathology in medicine can be attributed to many factors, including its efficiency and reproducibility. However, the predominant reason for its persistence as a worthwhile diagnostic tool is its significant predictive power and its ability to dictate therapy. No other single medical test can provide as much clinically useful information as a well documented surgical pathology examination of a resected neoplasm. Using morphology alone, a tumor’s grade, stage (primary and nodal status), type of differentiation, and other informative morphologic features, can be determined with accuracy, efficiency, and reproducibility. Using this information, oncologists and other cancer therapists can design and implement therapy and predict a patient’s outcome with reasonable accuracy. Despite the power of morphologic assessment for diagnosis and prognosis, there is still a need to provide additional and better predictive information about a particular patient’s disease. For example, the standard of care for patients with stage 1 adenocarcinoma of the lung is surgery followed by careful follow-up. These patients generally do well, with most of them alive and free of disease 5 years following surgery. Although the morphology can be helpful for identifying the patients who will do poorly, it is not adequate. Similar arguments can be made for some patients with node-negative breast carcinoma. Thus, there is considerable enthusiasm about new molecular approaches that might be able to stratify these patients into groups with varying risks of poor outcome, thereby affording the opportunity to treat them differently. Furthermore, it is hoped that molecular approaches will lead to the discovery of new subclasses of tumors not appreciable by morphology. In fact, early work on breast cancer and lymphoma suggests that this is indeed the case (see Breast Carcinoma; Hematologic Malignancies). As molecularly targeted therapeutics become increasingly available, it will be important to evaluate specific therapeutic targets in tumors to determine the appropriateness of a particular therapy for a given patient. The most illustrative paradigm is breast cancer, for which every case of invasive carcinoma is evaluated for expression of estrogen receptors (ERs) and progesterone receptors (PRs) and overexpression of the erbB2/Her2/neu gene. This information is used to classify patients further and guide decisions regarding antiestrogenic (tamoxifen) and anti-erbB2 therapies (trastuzumab). In this instance, morphology is augmented by information regarding expression of specific target genes to select the most appropriate therapy. Determining the sequence of the human genome, together with the development of powerful new molecular biology techniques that permit comprehensive and parallel assessment of gene expression, has sparked a genomic revolution in the assessment of human cancer. Using this gene expression approach, it is anticipated that a new cancer classification based on gene expression will be developed that, together with traditional morphology, will lead to a more informative classification of human cancer. In this chapter, an overview of the morphologic assessment of cancer along with the progress made to date for select tumor types in the molecular classification revolution are presented. Detailed description of all the types of tumor and their morphology is clearly beyond the scope of this chapter. Rather, the goal here is to present a selected

10

overview of tumor morphology that also provides a foundation for the discussion of molecular classification approaches. Readers seeking additional information are directed to the many volumes dedicated to the morphologic description of tumors, such as the Atlas of Tumor Pathology published by the Armed Forces Institute of Pathology and the WHO Classification of Tumours series published by the World Health Organization.

HISTORICAL PERSPECTIVE The historical aspects of morphology and human cancer classification are rich and fascinating yet clearly beyond the scope of this chapter. Interested readers are referred to the writings of Juan Rosai (Rosai, 1997) and others (Gal, 2001; Acs et al., 2002).

MORPHOLOGIC CLASSIFICATION OF CANCER The morphologic approach to cancer assessment is based on the pathologist’s ability to analyze myriad histologic patterns and synthesize the information into a rational classification scheme. The neural network that is the human brain is amazingly well suited for this task of pattern recognition. The art and science of surgical pathology is the recognition of patterns common to tumors and the assembly of classification schemes based on those patterns. In the ideal setting, a classification scheme would group patients with similar diseases and thereby provide the basis for a rational approach to therapy. For example, the classification of lung cancer into two large groups, smallcell carcinoma and non-small-cell carcinoma, with four main subtypes—adenocarcinoma, squamous cell carcinoma, small-cell carcinoma, large-cell carcinoma—resulted from the histologic examination of many lung tumors and recognition of these common morphologic patterns. This classification scheme is clinically informative regarding the clinical course and response to therapy and is largely the determining factor in the choice of therapy. Applying this approach to all cancers, surgical pathologists over the years have assembled a comprehensive classification of human cancer. An abbreviated classification of human cancer is presented in Table 2–1.

OVERVIEW OF CURRENT NOMENCLATURE OF HUMAN CANCER Consistent and informative cancer nomenclature is vital to effective patient care and all types of cancer research. Thus, much effort over the years has been invested in developing systemized cancer nomenclatures, resulting in two that are commonly used. The Systematized Nomenclature of Human and Veterinary Medicine (SNOMED) (College of American Pathologists, 1993), also called SNOMED International, traces its beginnings to back the publication of the Systematized Nomenclature of Pathology (SNOP) (College of American Pathologists, 1965). SNOP was expanded into the Systematized Nomenclature of Medicine (SNOMED) (College of American Pathologists, 1979). SNOMED assigns terms to one of eleven independent systematized modules, including those for site of tumor origin (topography) and histologic type (morphology). Within each of the 11

Table 2–1. Simplified Organ-Based Classification of Primary Human Cancers 1. Skin A. Epidermis a. Basal cell carcinoma b. Squamous cell carcinoma B. Adnexae a. Malignant versions of numerous tumors with eccrine, apocrine and sebaceous differentiation, along with tumors of the hair follicle 2. Oral cavity and oropharynx A. Squamous cell carcinoma B. Malignant tumors of minor salivary glands 3. Mandible and maxilla A. Odontogenic tumors a. Ameloblastoma and ameloblastic carcinoma 4. Nasopharynx and sinuses A. Squamous cell carcinoma B. Nasopharyngeal carcinoma C. Olfactory neuroblastoma 5. Lung A. Squamous cell carcinoma B. Adenocarcinoma C. Large cell carcinoma D. Small cell carcinoma E. Carcinoid 6. Pleura A. Mesothelioma 7. Mediastinum A. Thymoma and thymic carcinoma B. Neuroendocrine neoplasms C. Malignant lymphoma D. Neurogenic tumors 8. Thyroid A. Papillary carcinoma B. Follicular carcinoma C. Hurthle cell carcinoma D. Poorly differentiated carcinoma E. Anaplastic (undifferentiated) carcinoma F. Medullary carcinoma 9. Parathyroid A. Parathyroid carcinoma 10. Esophagus A. Adenocarcinoma B. Squamous cell carcinoma 11. Stomach A. Adenocarcinoma a. Intestinal type b. Diffuse type B. Neuroendocrine tumors C. Stromal tumors D. Malignant lymphoma 12. Small intestine A. Adenocarcinoma B. Neuroendocrine tumors 13. Appendix A. Adenocarcinoma B. Carcinoid 14. Large intestine A. Adenocarcinoma B. Carcinoid and other neuroendocrine tumors 15. Anus A. Malignant melanoma B. Adenocarcinoma C. Squamous cell carcinoma 16. Liver A. Hepatocellular carcinoma B. Cholangiocarcinoma C. Angiosarcoma 17. Gallbladder A. Adenocarcinoma 18. Pancreas A. Ductal adenocarcinomas B. Anaplastic carcinoma C. Endocrine tumors

19. Salivary glands A. Mucoepidermoid carcinoma B. Acinic cell carcinoma C. Adenoid cystic carcinoma D. Ductal carcinoma E. Malignant lymphoma 20. Adrenal gland A. Adrenocortical carcinoma B. Pheochromocytoma C. Neuroblastoma 21. Kidney A. Renal cell carcinoma and related tumors B. Wilms’ tumor 22. Bladder and renal pelvis A. Urothelial carcinoma B. Neuroendocrine carcinoma 23. Prostate A. Adenocarcinoma 24. Testis A. Germ cell tumors a. Seminoma b. Mature and immature teratoma c. Yolk sac tumor d. Choriocarcinoma e. Teratocarcinoma 25. Sex cord-stromal tumors A. Leydig cell tumor 26. Penis A. Squamous cell carcinoma 27. Vulva A. Squamous cell carcinoma B. Extramammary Paget’s disease C. Malignant melanoma 28. Vagina A. Squamous cell carcinoma B. Adenocarcinoma C. Botryoid rhabdomyosarcoma 29. Cervix A. Squamous cell carcinoma B. Neuroendocrine carcinoma C. Adenocarcinoma 30. Uterus A. Endometrial adenocarcinoma B. Malignant mixed müllerian tumors C. Leiomyosarcoma 31. Fallopian tube A. Adenocarcinoma 32. Ovary A. Surface epithelial tumors a. Serous adenocarcinoma b. Endometrioid adenocarcinoma c. Mucinous adenocarcinoma d. Clear cell carcinoma B. Germ cell tumors a. Dysgerminoma b. Yolk sac tumor c. Choriocarcinoma d. Mature and immature teratoma C. Sex cord-stromal tumors a. Granulosa cell tumor b. Sertoli-Leydig cell tumor 33. Placenta A. Hydatidiform mole B. Placental site trophoblastic tumor C. Choriocarcinoma 34. Mammary gland A. In situ carcinoma a. Ductal carcinoma in situ b. Lobular carcinoma in situ B. Invasive carcinoma a. Invasive ductal carcinoma b. Invasive lobular carcinoma c. Tubular carcinoma d. Mucinous carcinoma e. Medullary carcinoma f. Metaplastic carcinoma

(continued)

11

12

PART I: BASIC CONCEPTS

Table 2–1. (cont.) C. Stromal sarcoma (phylloides tumors) D. Angiosarcoma 35. Lymph nodes A. Hodgkin’s lymphoma a. Classic Hodgkin’s lymphoma b. Nodular lymphocyte-predominant Hodgkin’s lymphoma c. Nodular sclerosis Hodgkin’s lymphoma d. Mixed cellularity Hodgkin’s lymphoma e. Lymphocyte depletion Hodgkin’s lymphoma B. Non-Hodgkin’s lymphoma a. Precursor B-cell leukemia/lymphoma b. Mature (peripheral) B-cell leukemia/lymphoma c. Precursor T-cell leukemia/lymphoma d. Mature (peripheral) T-cell leukemia/lymphoma 36. Spleen A. Malignant lymphoma B. Leukemia C. Systemic mastocytosis 37. Bone marrow A. Leukemia a. Chronic myeloid leukemia b. Chronic lymphoid leukemia c. Acute lymphoblastic leukemia d. Acute myeloid leukemia B. Malignant lymphoma C. Histiocytosis D. Plasma cell neoplasms 38. Bone A. Osteosarcoma B. Aggressive and malignant osteoblastoma C. Chondrosarcoma D. Ewing’s sarcoma E. Other sarcomas 39. Soft tissue cancers A. Fibrosarcoma B. Malignant fibrous histiocytoma C. Malignant schwannoma D. Liposarcoma E. Angiosarcoma F. Leiomyosarcoma G. Rhabdomyosarcoma H. Clear cell sarcoma I. Epithelioid sarcoma J. Alveolar soft part sarcoma 40. Heart A. Sarcoma a. Angiosarcoma b. Rhabdomyosarcoma 41. Pericardium and peritoneum A. Malignant mesothelioma 42. Central nervous system A. Glial tumors a. Astrocytoma and glioblastoma multiforme b. Oligodendroglioma c. Malignant ependymoma B. Primitive neuroepithelial tumors a. Medulloblastoma C. Meningothelial tumors a. Malignant meningioma D. Lymphoid tumors a. Malignant lymphoma 43. Pituitary gland A. Invasive adenoma and pituitary carcinoma

modules, terms are placed into hierarchies, and a five- or six-digit alphanumeric code is designated. The latest version of SNOMED incorporates nomenclature from the other commonly used cancer nomenclature, the International Classification of Diseases for Oncology (ICD-O). The ICD-O traces its origins back to the formation of the World Health Organization (WHO), when after World War II it assumed responsibility for coding diseases. The original classifications

of neoplasms were based entirely on topographic site and behavior (benign or malignant). Tumor morphology was first incorporated with publication in 1951 of the Manual of Tumor Nomenclature and Coding (MOTNAC) (American Cancer Society, 1968). The ICD-O was initially published in 1976 (WHO, 1976) and then revised in 1990 (WHO, 1990). ICD-O is a dual classification and coding scheme that incorporates information regarding tumor topography and morphology. The third and latest edition of ICD-O (WHO, 2000) is similar to the second edition, with revision of the classification of leukemias and lymphomas to reflect the WHO classification of these diseases.

MOLECULAR METHODS FOR TUMOR CLASSIFICATION Molecular stratification of tumors for classification purposes has recently developed into a burgeoning field. Yet this approach has been used for more than two decades in hematopathology, starting with the realization that many of the hematolymphoid malignancies contained specific cytogenetic abnormalities that were characteristic for a particular disease. Recent excitement about gene expression profiling (Chung et al., 2002) can be directly attributed to technologic advances (e.g., serial analysis of gene expression, or SAGE, and DNA microarrays) combined with novel computational approaches to biology. Certainly in this postgenome era, the sequencing of the human genome and the consequent identification of most of the genes greatly amplifies the power of the comparative gene expression approach provided by these new technologies. A short overview of them is presented.

Cytogenetic Techniques Traditional cytogenetics employs chromosomal banding techniques of chromosomes (so-called G banding for the Giemsa stain, which is commonly used) to identify gross abnormalities such as deletions, inversions, isochromosomes, and translocations. The sensitivity for detecting relatively small abnormalities such as small deletions is limited. However, this technique is well suited for uncovering and mapping balanced reciprocal translocations (Sandberg, 1991; Sozzi et al., 1999), such as those present in sarcomas (Sreekantaiah et al., 1994) and hematolymphoid malignancies (Kaneko, 1990; Ambinder and Griffin, 1991; Clare and Hansen, 1994; Rowley, 1999). Fine mapping of the breakpoints in these translocated chromosomes has led to a greater understanding of the neoplastic pathogenesis for many tumor types by leading to the discovery of several oncogenes, such as c-myc.

Fluorescence In Situ Hybridization Fluorescence in situ hybridization (FISH) and the more recently developed chromogenic in situ hybridization (CISH) are molecular cytogenetic techniques that permit identification of specific nucleic acid sequences in intact cells in either metaphase or interphase (Muhlmann, 2002). Hybridization of labeled nucleic acid probes results in a detectable fluorescent signal, which can be quantitated along with a distinctly colored chromosome-specific control probe. FISH offers several advantages over traditional cytogenetics. First and importantly, it works on cells in interphase as well as metaphase. Second, the results are visually striking and easy to comprehend. Third, unlike traditional cytogenetics, which requires fresh tissue to culture and arresting the cells in metaphase, FISH is fully compatible with archival paraffin-embedded tissue, thus permitting significant retrospective analyses. Finally, FISH is flexible and adaptable to other techniques, expanding its usefulness. One of the disadvantages of the FISH technique is its lack of a discovery component, as the probe sequence must be known in advance, in contrast to comparative genomic hybridization (see below). Using FISH, cytogenetic abnormalities have been discovered and defined for most hematologic malignancies (Martin-Subero et al., 2003) and many solid tumors (Poetsch et al., 2000). However, given the extreme aneuploidy present in some solid tumors, most

Morphologic and Molecular Classification of Human Cancer FISH-detectable abnormalities have not yet been clinically adapted and do not yet represent the means for useful classification. Two exceptions are FISH analysis for erbB2/Her2/neu amplification in breast cancer (Kallioniemi et al., 1992) and for loss of 1p and 19q in oligodendrogliomas (Gelpi et al., 2003) (see Breast Carcinoma; Gliomas). Despite this limited clinical acceptance to date, much effort is being expended to develop novel FISH-based diagnostic tests, including assays for detecting recurrent bladder carcinoma using cells in urine.

Comparative Genomic Hybridization Comparative genomic hybridization (CGH), like cDNA microarrays (see below), involves competitive hybridization of differentially labeled DNAs (test and reference samples) to normal metaphase chromosomes to measure chromosome imbalances across the entire genome (Kallioniemi et al., 1993). Thus, CGH is useful for identifying tumor-specific chromosomal gains and losses (regions of chromosomal amplifications or deletions). By analyzing numerous related tumors in parallel, it is possible to identify regions of gain or loss common to the set of tumors being studied, thereby identifying alterations that are likely related to tumorigenesis. Comparative genomic hybridization, although not a particularly high-throughput technology, has been used successfully to classify human cancers, including sarcomas (Chibon et al., 2003), lung carcinoma (Aliferis et al., 2002), and breast carcinoma (Wessels et al., 2002). Yet CGH is not likely to become a clinically accepted diagnostic test.

DNA Microarrays DNA microarray technology has become one of the most robust methods for the parallel measurement of expression of numerous genes, presenting an alternative approach to existing techniques, such as differential display (Liang, 2002) and SAGE (see below). DNA microarrays exist in two fundamental types, each with advantages and disadvantages (Ramsay, 1998; Cheung et al., 1999; Khan et al., 1999; Gershon, 2002). cDNA microarrays (also called spotted microarrays because DNA is spotted onto glass slides) contain thousands of cDNA clones precisely arrayed on a slide. RNAs from two related tissues or cell lines (experimental and reference) are competitively hybridized to the spotted DNAs. Labeling the RNAs with different fluorescent molecules allows measurement of the relative amounts of a specific RNA in the experimental sample compared to the reference; the results are expressed as a ratio. Unlike spotted cDNA microarrays, oligonucleotide microarrays are synthesized in situ by means of photolithography; the technique utilizes inherent probe redundancy to measure RNA transcript levels. Affymetrix (Santa Clara, CA) manufactures one of the most widely used oligonucleotide microarrays (GeneChips), although chips from other manufacturers are gaining in use. With GeneChips (Affymetrix), each probe set for a given gene consists of a series of short oligonucleotides that are complementary to that gene. Both perfectly matched (PM) oligonucleotides and mismatched oligonucleotides (MM), which differ by a central base, are incorporated into the microarrays. The intensity signal derived from the MM is subtracted from the PM intensity signal, and the net signal from the entire series of probes for an individual gene is an indication of its RNA level. The presence of hundreds of thousands of oligonucleotides on a single chip allows the parallel assessment of transcripts levels for more than 22,000 probe sets. Advantages of cDNA microarrays include a lower cost and their flexibility of design; their disadvantages include the need for a reference RNA and the quality control aspects of producing spotted microarrays. Advantages of oligonucleotide microarrays include the absence of a required reference RNA (which permits more robust cross-experiment and cross-laboratory comparisons), commercial standards and quality assurance for chip manufacturing, and the independence from establishing and maintaining a spotted microarray facility. The major disadvantage of oligonucleotide microarrays is their high cost. For thorough discussions of cDNA and oligonucleotide

13

microarrays, see several recent reviews (Holloway et al., 2002; Yang and Speed, 2002; Stears et al., 2003).

Serial Analysis of Gene Expression Serial analysis of gene expression (SAGE) permits quantitative and simultaneous assessment of expressed genes by sequencing small specific sequence tags (Velculescu et al., 1995). The relative abundance of a specific tag reflects the relative expression of the gene that corresponds to the tag. SAGE offers some benefits over DNA microarrays, such as not having to know the genes to be placed on a microarray and the ability to detect and measure low abundance transcripts. Using SAGE, expression profiles of many types of cancer have been elucidated (e.g., Zhang et al., 1997; Oue et al., 2004; Weeraratna et al., 2004), and these profiles have been used to identify novel diagnostic markers and to develop an expression-based classification map (Buckhaults et al., 2003). SAGE data have been effectively combined with DNA microarray data to focus the list of differentially expressed genes by identifying genes common to both techniques in pancreas cancer (Iacobuzio-Donahue and Hruban, 2003).

Single Marker Approaches to Molecular Classification Immunohistochemical assessment of protein expression in tissue sections represents the most elementary form of “molecular” testing for stratifying cancers. Although immunohistochemistry is generally not viewed as a molecular test, the immunologic detection of specific protein molecules in tissue sections does fit its broad definition. The best example and one of the most clinically accepted tests is the immunohistochemical assessment of ER expression in invasive breast carcinoma (Fig. 2–1a). With this single immunohistochemical assay, invasive carcinomas can be roughly divided into two broad categories (ER-positive and ER-negative) with profound clinical significance (Osborne et al., 1981; Berger et al., 1991). Studies of the erbB2/Her2/neu gene in invasive breast cancer have shown that assessment of this gene, for genomic amplification or protein expression (Fig. 2–1b), is significantly associated with the prognosis as well as the response to anti-Her2 therapy. This correlation was first observed during the late 1980s (McGuire, 1987; Slamon et al., 1987; Berger et al., 1988; Guerin et al., 1988; Tandon et al., 1989; Wright et al., 1989), and since then numerous studies have examined this relation from a variety of angles including various assay methods (DiLeo et al., 2002; Ross et al., 2003). Regardless of the method used, it is clear that erbB2/Her2/neu is informative for breast and probably ovarian carcinoma. Lobular and ductal types of in situ breast carcinoma are usually distinguished by morphology with little difficulty (see Breast, below). However, a small number of in situ lesions share morphologic features of both types. E-cadherin expression as assessed by immunohistochemistry, however, can assist with this distinction in a high percentage of cases, as loss of E-cadherin expression is an early event in lobular carcinoma (Moll et al., 1993; Vos et al., 1997; Jacobs et al., 2001; Wahed et al., 2002). The type-specific translocations observed in soft tissue sarcomas and many leukemias and lymphomas are prime examples of a single molecular abnormality functioning as the defining feature of a specific tumor type. Although translocations are rare in epithelial tumors, thyroid tumors of follicular cell origin do display some specific chromosomal rearrangements (Tallini, 2002). A high percentage of papillary thyroid carcinomas display rearrangements of the RET proto-oncogene (Tallini and Asa, 2001; Nikiforov, 2002), whereas some follicular carcinomas contain rearrangements of the PAX8peroxisome proliferator-activated receptor gamma genes (Kroll et al., 2000; Marques et al., 2002; Nikiforova et al., 2002; Cheung et al., 2003; Dwight et al., 2003). These rearrangements are thought to be crucial to the development of the neoplastic state (Kim et al., 2003; Puxeddu et al., 2003).

A

B

C

D

E

F

G

H

Figure 2–1. A, Invasive breast ductal carcinoma with strong nuclear immunoreactivity for estrogen receptors, seen by immunohistochemistry using formalin-fixed, paraffin-embedded tissue sections. B, Invasive breast ductal carcinoma with strong membranous immunoreactivity for cerbB2/Her2/neu, as seen by immunohistochemistry using formalin-fixed, paraffin-embedded tissue sections. C, Ordinary invasive ductal carcinoma of breast showing occasional duct or tubule formation. D, Invasive lobular

14

carcinoma of breast with classic growth pattern. E, Invasive colorectal adenocarcinoma with stroma and necrosis. F, Squamous cell carcinoma of lung with focus of clear-cut squamous differentiation. G, Biphasic synovial sarcoma with spindle cell and epithelioid components. H, High-grade leiomyosarcoma with fascicles of spindle cells running in different directions. C–H: H&E.

Morphologic and Molecular Classification of Human Cancer

Single Marker versus Multimarker Approaches to Molecular Classification With the arrival of technologies for the serial assessment of expression of numerous genes (e.g., SAGE and DNA microarrays), it is now possible to use gene expression patterns as a tumor classification tool. As expected, this multigene approach provides greater analytic power than single gene approaches. For example, the diagnosis of small-cell lung carcinoma is routinely made with light microscopy and immunohistochemistry (IHC) for neuroendocrine markers, such as chromogranin A and synaptophysin. Although this approach routinely works well, expression of these markers can be variable, so a panel of IHC markers is sometimes needed for the highest diagnostic accuracy. DNA microarray analysis of lung carcinomas has detected many additional marker genes associated with the neuroendocrine phenotype of smallcell carcinoma (Bhattacharjee et al., 2001). Similarly, DNA microarray analysis of papillary thyroid carcinoma has detected numerous potential marker genes that, when combined into a panel of marker genes, effectively assists in the diagnosis of this tumor (unpublished results). Given the morphologic complexity of some tumors (e.g., lung carcinoma), it is not unexpected that multigene marker approaches are more informative and powerful than single-marker gene approaches.

TUMOR-SPECIFIC SIGNATURES AND MULTITUMOR CLASSIFICATIONS Several groups have developed gene expression profiles of morphologically related tumors (carcinomas or adenocarcinomas) from different organ systems to develop tumor-specific signatures and have used those signatures as a classification tool. One of the first attempts profiled acute leukemia and successfully rediscovered the tumor classes of lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML) (class discovery). It further correctly predicted the classification of an independent set of cases (class prediction) using gene expression data alone (Golub et al., 1999). A similar approach using expression profiles of adenocarcinomas of colon, lung, and ovary correctly classified 152 tumors and detected two tumors that were misclassified (Giordano et al., 2001). Importantly, this approach yielded a list of marker genes for each class that contained markers already accepted into surgical pathology practice (e.g., cytokeratin 20 as a marker of colon carcinoma), providing some validation of the signature approach. Others (Ramaswamy et al., 2001; Su et al., 2001) have expanded this approach to include more tumor types, although with fewer tumors comprising each cohort. Collectively, these studies demonstrate the feasibility of constructing gene expression-based cancer classification maps and discovering new marker genes that can be implemented in the reverse transcription-polymerase chain reaction (RT-PCR) (Buckhaults et al., 2003) and IHC assays (Moskaluk et al., 2003). Data analysis is one of the most significant challenges presented by the voluminous gene expression data generated by DNA microarray analysis. There are numerous ways to approach the analysis for tumor classification, ranging from sophisticated computer learning programs such as artificial neural networks (Khan et al., 2001) and support vector machines (Furey et al., 2000; Ramaswamy et al., 2001; Lee and Lee, 2003) to more direct approaches such as decision trees (Shedden et al., 2003). Although many approaches successfully classify cancers, the simpler, more direct approaches offer the advantages of being transparent in terms of which genes drive the classification and of more closely mimicking the approach used by practicing surgical pathologists.

ORGAN-SPECIFIC MORPHOLOGIC AND MOLECULAR CLASSIFICATION OF SELECT TUMORS Breast Carcinoma In many respects, breast carcinoma serves as the paradigm for the approach to classifying many human cancers. Carcinomas of the breast

15

are divided into in situ and invasive types, depending on whether the neoplastic cells are confined to the ductal system of the breast (in situ) or have left the ductal system to involve the supporting stroma (invasive). The same approach is taken with many epithelial tumors, including prostatic and pancreatic adenocarcinomas. In situ breast carcinomas are divided into ductal and lobular types, depending on the morphology of the neoplastic cells in the ducts. Invasive breast carcinomas are also divided into two large categories based on the morphologic appearance of the invasive tumor cells. Invasive ductal carcinomas are tumors that show evidence of glandular differentiation (Fig. 2–1c) and comprise most breast carcinomas. Invasive lobular carcinomas lack glandular differentiation and show a striking uniformity of tumor cells (Fig. 2–1d). Many other types and variants of both ductal and lobular carcinoma, some with prognostic significance, have been described. For example, tubular carcinoma is a well differentiated carcinoma that consists exclusively of tubules and is associated with a favorable prognosis (Kader et al., 2001; Kitchen et al., 2001; Cabral et al., 2003). The current evaluation of all invasive carcinomas includes assessment of hormone (estrogen and progesterone) receptor status, assessment of erbB2/Her2/neu expression, and/or genomic amplification. In fact, the separation of ER-positive and ER-negative breast carcinomas represents one of the earliest informative molecular classifications. Breast carcinomas that express high levels of these hormonal receptors usually respond to antihormonal therapy and have a more favorable prognosis (Osborne et al., 1981; Berger et al., 1991). Conversely, tumors with erbB2/Her2/neu amplification and overexpression that lack hormonal receptor expression have a poorer prognosis (Tandon et al., 1989; Press et al., 1997). Much recent effort has been invested in developing new molecular classifications of invasive ductal carcinomas using gene expression profiling approaches (Perou et al., 2000; Sorlie et al., 2001; Sorlie et al., 2003). These efforts by several groups have largely been successful in stratifying breast carcinoma into subsets. Interestingly, this approach reinforces the power of the ER and erbB2/Her2/neu status as useful classifying factors. Using hierarchical clustering algorithms to analyze gene expression in a cohort of breast carcinomas, the tumors were divided into two broad categories—ERnegative and ER-positive—each further containing additional subsets. The subsets were named based on specific gene expression. The ERnegative group consisted of the basal-like, the erbB2-positive, and the normal breast-like subtypes, whereas the ER-positive group consisted of the luminal subtypes A, B, and C. The basal-like subtype was characterized by high expression of keratins 5 and 17, laminin, and fatty acid-binding protein 7. The erbB2-positive subtype was characterized by high expression of erbB2 and other associated genes in its amplicon at 17q22.24. The normal breast-like subtype expressed genes related to adipose tissue and other mesenchymal cells. The luminal subtypes showed expression of genes common to luminal-type epithelial cells and genes associated with ER activation, with slight variation among the three subtypes. Importantly, this molecular classification scheme has been reproduced across several data sets (Sorlie et al., 2003). Because molecular classification schemes are useful only if they provide incremental information above and beyond that provided by morphology, the survival of breast carcinoma patients was examined to determine if the molecular subtype correlated with the outcome. The basal-like and erbB2-positive subtypes were associated with the shortest survival times and relapse-free survival times compared with the other subtypes. The association of erbB2 amplification and poor prognosis is well documented, thereby reinforcing the gene expression profiling approach to class discovery. The basal-like subtype expresses keratins 5 and 17, and expression of these proteins was shown to have prognostic significance as assessed by IHC (van de Rijn et al., 2002). The major significance of this work related to breast carcinoma classification lies in the discovery of the basal-like subtype. Although it can be argued that this subtype could be deduced by elimination using standard IHC methods (ER- and erbB2-negative tumors), the clear delineation of this subtype based on gene expression and the

16

PART I: BASIC CONCEPTS

development of positive IHC markers (keratins 5 and 17) should permit routine identification of this subtype in pathology practice. Moreover, knowledge of gene expression in the basal-like subtype should allow new therapeutic approaches.

Colorectal Carcinoma Most primary colorectal tumors are adenocarcinomas. Much work in this field led to the development of the adenoma–carcinoma sequence with an understanding of the accumulation of mutations during this histologic progression (Kinzler and Vogelstein, 1996). In addition to adenocarcinoma, tumors with neuroendocrine differentiation (e.g., carcinoids) are common in the colon and rectum as well as throughout the gastrointestinal tract. Colorectal adenocarcinoma is relatively morphologically homogeneous, especially when compared with adenocarcinomas of other organs (e.g., lung), as most tumors are well to moderately differentiated and have other characteristic features, such as mucin production and a prominent desmoplastic response. However, histologic subtypes have been developed, such as mucinous carcinoma, signet-ring cell carcinoma, and medullary carcinoma. There is a correlation between morphologic type and the underlying pathogenesis, as it is now appreciated that colorectal carcinomas develop through at least two distinct molecular genetic pathways (Jass et al., 1999). The bulk of the tumors develop through the chromosomal instability or the wnt signaling pathway and exhibit the typical, or ordinary, phenotype. The remaining tumors are associated with microsatellite instability owing to underlying defects in DNA mismatch repair, and they have a different phenotype that includes the mucinous and medullary carcinomas and is characterized by being either well or poorly differentiated and having a right-sided location, a host response with tumorinfiltrating lymphocytes, and mucinous differentiation (Greenson et al., 2003). Microsatellite instability status, which has gained clinical acceptance, is often included in the workup of patients with colorectal carcinoma and is determined by examining a panel of microsatellite markers (Boland et al., 1998) or documenting loss of DNA mismatch repair enzymes in tumor tissue sections by IHC (Lindor et al., 2002). Microarray analysis of colon cancer has not yet yielded any significant insights. However, distinct transcriptional profiles of microsatellite stable and unstable tumors have been derived.

Prostate Carcinoma Prostate cancer is almost invariably adenocarcinoma and can be divided into two major categories: adenocarcinoma of peripheral ducts and acini and large duct adenocarcinoma. Rare primary urothelial carcinomas of the prostate have also been recognized, as have several morphologic variants of adenocarcinoma. Grading prostatic adenocarcinoma uses the clinically accepted and preferred Gleason grading scheme, developed in association with the Veterans Administration Cooperative Urology Research Group (Gleason and Mellinger, 1974). Gleason grading is based on the degree of glandular differentiation and is assigned a score of 1–5, with 1 being well differentiated and 5 being nearly undifferentiated. The predominant and secondary patterns in a given tumor are each graded, and the two grades are added to obtain the Gleason score. Studies have demonstrated high interobserver reproducibility of Gleason grading (Mills et al., 1990; Allsbrook et al., 2001a,b). The acceptance of Gleason grading can be attributed to its documented high correlation with a variety of clinical and pathologic parameters, most significantly with survival (Mills et al., 1990). Despite the success of clinical staging and pathologic grading for predicting the outcome of patients with prostatic adenocarcinoma, methods to improve outcome prediction are needed. Several studies have employed gene expression profiling as a tool to develop novel single gene prognostic biomarkers (Dhanasekaran et al., 2001) and using clusters of small numbers of genes to predict the clinical course of prostate carcinoma (Glinsky et al., 2004).

Lung Carcinoma Lung tumors are overwhelmingly epithelial (carcinomas) and are broadly divided into two types: small-cell and non-small-cell carcinomas. Small-cell carcinoma belongs to the larger family of neuroendocrine (NE) neoplasms, which also includes carcinoid, atypical carcinoid, and large-cell NE carcinoma (numerous classifications schemes have been proposed for the classification of NE tumors of the lung). Small-cell carcinoma represents the most undifferentiated form of NE carcinoma, yet it does retain some morphologic and/or IHC evidence of NE differentiation. Conversely, carcinoid represents the most differentiated NE tumor. Non-small-cell carcinomas include adenocarcinoma and squamous cell carcinoma and less common tumors such as large-cell carcinoma, which is believed to represent an undifferentiated form of either adenocarcinoma or squamous cell carcinoma. The molecular classification of lung carcinoma has been the focus of much recent work. Using a DNA microarray-based approach with hierarchical clustering, Bhattacharjee et al. (2001) recapitulated the recognized classification of lung carcinomas and further subdivided the adenocarcinoma cohort into four subclasses. Interestingly, using this approach, several presumed primary lung adenocarcinomas were discovered to be metastases from colon, breast, and liver. A study by Garber et al. (2001), using a similar approach with a cDNA microarray, yielded similar results and showed a correlation between adenocarcinoma subsets and patient survival. Work from our group (Beer et al., 2002) focused on subclassification of lung adenocarcinoma and identified three subtypes based on hierarchical clustering. Using a statistical risk index based on expression data for 16 genes, stage I adenocarcinomas with good and poor prognoses were identified. These results provide the opportunity to treat patients with early high risk disease further using adjuvant therapy after surgery. The collective lung carcinoma profiling work is currently being reproduced and expanded via a multiinstitutional National Cancer Institute (NCI)funded project. In preparation for this project, the NCI organized an interlaboratory comparability study of gene expression profiling using DNA microarrays to assess the feasibility of combining microarray data generated in various laboratories (K. Dobbin, personal communication).

Gliomas The basic classification of glial brain tumors (diffuse gliomas), the most common primary brain tumors, rests on morphology, with qualifiers used to describe patterns of differentiation (astrocytic, oligodendroglial, oligoastrocytic). Tumor grade is important for prognosis and therapy and is crudely based on the histologic degree of malignancy (WHO grades II–IV). Astrocytic neoplasms include the welldifferentiated diffuse astrocytoma (WHO grade II), the anaplastic astrocytoma (WHO grade III), and the specially designated glioblastoma multiforme (WHO grade IV). Oligodendroglial tumors include oligodendroglioma (WHO grade II) and anaplastic oligodendroglioma (WHO grade III). Although tumor grade and differentiation provide a useful framework for glioma classification, difficulties persist. For example, grading gliomas strives to create discrete categories when in fact the tumors represent a continuous spectrum of neoplastic evolution. Furthermore, this evolutionary process is essentially genetic and is driven by the underlying accumulation of molecular events. Thus, the potential role of alternative approaches is considerable, and significant progress has been made toward a molecular glioma classification (Louis et al., 2001). Glioblastomas have been molecularly subclassified using the genes encoding p53 (TP53), the epidermal growth factor receptor (EGFR) (von Deimling et al., 1993). Tumors with EGFR genomic amplification do not contain a mutation of TP53 or allelic loss of its chromosomal location (17p); the converse relation is also observed (Watanabe et al., 1996). Division of glioblastomas into these two molecular subtypes (EGFR-amplified, TP53 wild-type, EGFR wild-type, TP53 mutated) correlates with the clinical features. Patients with a TP53 mutation are younger, and their tumors are associated with lower grade

Morphologic and Molecular Classification of Human Cancer astrocytomas (Reifenberger et al., 1996). Patients with EGFR genomic amplification are older, and their tumors appear to arise de novo without an associated low grade glioma (Watanabe et al., 1996). Molecular subtyping by cytogenetic approaches of oligodendrogliomas has become part of the standard diagnostic workup of these tumors (Ino et al., 2000; Smith et al., 2000; Thiessen et al., 2003; van den Bent et al., 2003). Based on chromosomal loss of 1p and 19q, it is possible to identify patients who will respond to combined chemotherapy (procarbazine, lomustine, and vincristine). Thus, patients with tumors who have loss of 1p and 19q have a much longer survival than those with intact 1p and mutation of the TP53, PTEN, or CDKN2A genes. Clinical implementation of this molecular approach to a “clinical laboratory improvement amendment” (CLIA)certified environment has been readily accomplished by FISH testing for 1p and 19q (Gelpi et al., 2003; Perry et al., 2003). DNA microarray studies of gliomas have defined gene expression signatures that correlate with tumor location (Mueller et al., 2002) and that can be more informative for survival than morphologic classification alone (Nutt et al., 2003).

Sarcoma The morphologic classification of sarcoma, particularly the spindle cell type, represents one of the most significant challenges for surgical pathologists. Based purely on morphology, the distinction between the various sarcoma subtypes is fraught with subjectivity and interobserver variability despite the availability of useful IHC diagnostic markers. Fortunately, an understanding of the underlying pathogenesis of these tumors, specifically the presence of type-specific chromosomal translocations, has led to some significant advances in sarcoma classification. Sarcomas can be broadly divided into spindle cell and epithelioid types, yet some tumors, such as the biphasic synovial sarcoma, which displays a spindle cell and an epithelioid component, defy this distinction. Any evidence of cellular differentiation, either morphologic or immunohistochemical, can be used to classify these mesenchymal tumors. For example, leiomyosarcomas often display a characteristic fascicular pattern that resembles smooth muscle and are immunoreactive for actins and desmin. Thus, the usual approach to spindle cell sarcomas includes careful histologic evaluation for any evidence of cellular differentiation along with a battery of IHC stains directed against type-specific proteins. Cytogenetic analysis has revealed two molecular subtypes of sarcoma: those with complex karyotypic abnormalities resulting in aneuploidy and those with specific chromosomal translocations and the relative absence of aneuploidy. During the late 1970s and early 1980s, cytogenetic approaches were applied to sarcomas, and some of the first histologic-cytogenetic associations were made (Aurias et al., 1984; de Chadarevian et al., 1984; Trent et al., 1985; Limon et al., 1986; Turc-Carel et al., 1986a,b; Douglass et al., 1987; Griffin and Emanuel, 1987). By the late 1980s, the diagnostic significance of specific translocations for sarcoma classification was beginning to emerge (Karakousis et al., 1987). As detailed in Table 2–2, specific recurrent chromosomal translocations have been detected for many types of sarcoma. Current molecular pathology practice includes routine Table 2–2. Translocations in Sarcoma with Diagnostic Utility Tumor Type

Cytogenetics

Fusion Genes and Proteins

Ewing’s sarcoma Synovial sarcoma Myxoid/round cell liposarcoma Alveolar rhabdomyosarcoma Extraskeletal myxoid chondrosarcoma Ewing’s sarcoma Clear cell sarcoma Desmoplastic small round blue cell tumor

t(11;22) t(X;18) t(12;16) t(2;13) t(9;22)

EWS/ETS SYT/SSX1 or SSX2 FUS/CHOP FKHR/PAX3 or PAX7 EWS/TEC

t(11;22) t(12;22) t(11;22)

EWS/FLI1 EWS/ATF1 EWS/WT1

17

translocation identification for all cases of uncertain classification by RT-PCR-based and/or interphase FISH-based methods. One of the most fascinating aspects of sarcoma translocations is the opportunity they provide to understand their biology and pathogenesis. As these translocations create novel fusion proteins involved in tumorigenesis, it is possible to identify the individual genes and explore the role of the resulting fusion proteins. In many cases, the genes involved encode DNA-binding proteins with transcriptional regulation activity (transcription factors), often developmentally regulated (e.g., the PAX genes). The cytogenetics of some sarcomas, malignant fibrous histiocytoma, and pleomorphic liposarcoma, for example, show complex chromosomal rearrangements without specific recurrent abnormalities. Because of this karyotypic heterogeneity, the development of molecular pathology assays is not practicable for these tumors. However, some have attempted to understand and classify these tumors using gene expression profiling approaches (Schofield and Triche, 2002). Molecular classification by comparative gene expression profiling has confirmed many of the existing sarcoma categories with consistent, distinct, homogeneous expression profiles. Such tumors include (not surprisingly based on their cytogenetic data) synovial sarcoma, round cell/myxoid liposarcoma, clear-cell sarcoma, and gastrointestinal stromal tumors.

Hematologic Malignancies The classification of hematologic malignancies (myeloid neoplasms, lymphoid neoplasms, posttransplant lymphoproliferative disorders, myelodysplastic syndromes, mast cell diseases, histiocytic and dendritic cell neoplasms) is vast and complicated, and a complete discussion of their current classification is beyond the scope of this chapter. However, the current classification recently published by WHO (Jancar, 2000; Jaffe et al., 2001) incorporates existing immunologic, cytogenetic (Martin-Subero et al., 2003), and molecular genetic information, thereby providing a complete framework for the classification of these neoplasms. Efforts to validate the classification, especially for the myelodysplastic syndromes, have been successful (Germing et al., 2000). Despite the success of the WHO classification, there is still a need to define the classification of hematologic neoplasms more precisely; and several recent studies using DNA microarrays have been performed with this goal in mind. A landmark study (Alizadeh et al., 2000) examined gene expression in diffuse large B-cell lymphoma (DLBCL), follicular lymphoma (FL), and chronic lymphocytic leukemia/lymphoma (CLL), as well as normal lymphocytes and lymphoma cell lines using a DNA microarray with 17,856 cDNA clones representing genes selected from various lymphoid cDNA libraries. These three lymphoma types were distinguishable using this approach, with the lower grade types (CLL and FL) sharing similar gene expression patterns with resting B cells. Interestingly, there was significant expression heterogeneity in the DLBCL group, and it was possible to define subgroups, designated germinal center B-like DLBCL and activated B-like DLBCL. Importantly, these subgroups have prognostic significance. This study was one of the first to demonstrate that gene expression profiling could define new cancer subtypes not discernible by existing morphologic and IHC approaches. A similar study of gene expression in DLBCL also defined subgroups that predicted outcome (Shipp et al., 2002), although the informative genes differed, suggesting that many factors contribute to DLBCL’s response to therapy and that much work remains to be done before clinical implementation of gene expression profiling.

CONCLUSIONS Despite the development of better prognostic markers, which is a large component of the current cancer genomics revolution, there is a need to predict which patients will respond to therapy. The ability to use an effective second- or third-line therapy and sidestep an ineffective, potentially toxic first-line therapy would greatly improve the care of

18

PART I: BASIC CONCEPTS

cancer patients, especially those with advanced cancers and a predicted rapid course. Thus, it is entirely possible to envision in the near future a scenario in which therapeutic decisions are determined primarily not by the histologic assessment of the tumor but, rather, by its molecular signature. References Acs G, Baloch ZW, LiVolsi VA. 2002. Intraoperative consultation: an historical perspective. Semin Diagn Pathol 19:190–191. Aliferis CF, Hardin D, Massion PP. 2002. Machine learning models for lung cancer classification using array comparative genomic hybridization. In: Proceedings of the AMIA Annual Symposium, pp. 7–11. Alizadeh AA, Eisen MB, Davis RE, Ma C, Lossos IS, Rosenwald A, Boldrick JC, Sabet H, Tran T, Yu X, Powell JI, Yang L, Marti GE, Moore T, Hudson J Jr, Lu L, Lewis DB, Tibshirani R, Sherlock G, Chan WC, Greiner TC, Weisenburger DD, Armitage JO, Warnke R, Levy R, Wilson W, Grever MR, Byrd JC, Botstein D, Brown PO, Staudt LM. 2000. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 403:503–511. Allsbrook WC Jr, Mangold KA, Johnson MH, Lane RB, Lane CG, Amin MB, Bostwick DG, Humphrey PA, Jones EC, Reuter VE, Sakr W, Sesterhenn IA, Troncoso P, Wheeler TM, Epstein JI. 2001. Interobserver reproducibility of Gleason grading of prostatic carcinoma: urologic pathologists. Hum Pathol 32:74–80. Allsbrook WC Jr, Mangold KA, Johnson MH, Lane RB, Lane CG, Epstein JI. 2001. Interobserver reproducibility of Gleason grading of prostatic carcinoma: general pathologist. Hum Pathol 32:81–88. Ambinder RF, Griffin CA. 1991. Biology of the lymphomas: cytogenetics, molecular biology, and virology. Curr Opin Oncol 3:806–812. American Cancer Society. 1968. Manual of Tumor Nomenclature and Coding. New York: American Cancer Society. Aurias A, Rimbaut C, Buffe D, Zucker JM, Mazabraud A. 1984. Translocation involving chromosome 22 in Ewing’s sarcoma: a cytogenetic study of four fresh tumors. Cancer Genet Cytogenet 12:21–25. Bhattacharjee A, Richards WG, Staunton J, Li C, Monti S, Vasa P, Ladd C, Beheshti J, Bueno R, Gillette M, Loda M, Weber G, Mark EJ, Lander ES, Wong W, Johnson BE, Golub TR, Sugarbaker DJ, Meyerson M. 2001. Classification of human lung carcinomas by mRNA expression profiling reveals distinct adenocarcinoma subclasses. Proc Natl Acad Sci USA 98:13790–13795. Beer DG, Kardia SL, Huang CC, Giordano TJ, Levin AM, Misek DE, Lin L, Chen G, Gharib TG, Thomas DG, Lizyness ML, Kuick R, Hayasaka S, Taylor JM, Iannettoni MD, Orringer MB, Hanash S. 2002. Geneexpression profiles predict survival of patients with lung adenocarcinoma. Nat Med 8:816–824. Berger MS, Locher GW, Saurer S, Gullick WJ, Waterfield MD, Groner B, Hynes NE. 1988. Correlation of c-erbB-2 gene amplification and protein expression in human breast carcinoma with nodal status and nuclear grading. Cancer Res 48:1238–1243. Berger U, McClelland RA, Wilson P, Greene GL, Haussler MR, Pike JW, Colston K, Easton D, Coombes RC. 1991. Immunocytochemical determination of estrogen receptor, progesterone receptor, and 1,25-dihydroxyvitamin D3 receptor in breast cancer and relationship to prognosis. Cancer Res 51:239–244. Boland CR, Thibodeau SN, Hamilton SR, Sidransky D, Eshleman JR, Burt RW, Meltzer SJ, Rodriguez-Bigas MA, Fodde R, Ranzani GN, Srivastava S. 1998. A National Cancer Institute workshop on microsatellite instability for cancer detection and familial predisposition: development of international criteria for the determination of microsatellite instability in colorectal cancer. Cancer Res 58:5248–5257. Buckhaults P, Zhang Z, Chen YC, Wang TL, St. Croix B, Saha S, Bardelli A, Morin PJ, Polyak K, Hruban RH, Velculescu VE, Shih IeM. 2003. Identifying tumor origin using a gene expression-based classification map. Cancer Res 63:4144–4149. Cabral AH, Recine M, Paramo JC, McPhee MM, Poppiti R, Mesko TW. 2003. Tubular carcinoma of the breast: an institutional experience and review of the literature. Breast J 9:298–301. Cheung L, Messina M, Gill A, Clarkson A, Learoyd D, Delbridge L, Wentworth J, Philips J, Clifton-Bligh R, Robinson BG. 2003. Detection of the PAX8-PPAR gamma fusion oncogene in both follicular thyroid carcinomas and adenomas. J Clin Endocrinol Metab 88:354–357. Cheung VG, Morley M, Aguilar F, Massimi A, Kucherlapati R, Childs G. 1999. Making and reading microarrays. Nat Genet 21:15–19. Chibon F, Mariani O, Mairal A, Derre J, Coindre JM, Terrier P, Lagace R, Sastre X, Aurias A. 2003. The use of clustering software for the classification of comparative genomic hybridization data: an analysis of 109 malignant fibrous histiocytomas. Cancer Genet Cytogenet 141:75–78.

Chung CH, Bernard PS, Perou CM. 2002. Molecular portraits and the family tree of cancer. Nat Genet 32(Suppl):533–540. Clare N, Hansen K. 1994. Cytogenetics in the diagnosis of hematologic malignancies. Hematol Oncol Clin North Am 8:785–807. College of American Pathologists. 1965. Systematized Nomenclature of Pathology. Chicago: College of American Pathologists. College of American Pathologists. 1979. Systematized Nomenclature of Medicine. Skokie, IL: College of American Pathologists. College of American Pathologists. 1993. Systematized Nomenclature of Human and Veterinary Medicine. Northfield, IL: College of American Pathologists. De Chadarevian JP, Vekemans M, Seemayer TA. 1984. Reciprocal translocation in small-cell sarcomas. N Engl J Med 311:1702–1703. Dhanasekaran SM, Barrette TR, Ghosh D, Shah R, Varambally S, Kurachi K, Pienta KJ, Rubin MA, Chinnaiyan AM. 2001. Delineation of prognostic biomarkers in prostate cancer. Nature 412:822–826. Di Leo A, Dowsett M, Horten B, Penault-Llorca F. 2002. Current status of HER2 testing. Oncology 63(Suppl 1):25–32. Douglass EC, Valentine M, Etcubanas E, Parham D, Webber BL, Houghton PJ, Houghton JA, Green AA. 1987. A specific chromosomal abnormality in rhabdomyosarcoma. Cytogenet Cell Genet 45:148–155. Dwight T, Thoppe SR, Foukakis T, Lui WO, Wallin G, Hoog A, Frisk T, Larsson C, Zedenius J. 2003. Involvement of the PAX8/peroxisome proliferator-activated receptor gamma rearrangement in follicular thyroid tumors. J Clin Endocrinol Metab 88:4440–4445. Furey TS, Cristianini N, Duffy N, Bednarski DW, Schummer M, Haussler D. 2000. Support vector machine classification and validation of cancer tissue samples using microarray expression data. Bioinformatics 16:906–914. Gal AA. 2001. In search of the origins of modern surgical pathology. Adv Anat Pathol 8:1–13. Garber ME, Troyanskaya OG, Schluens K, Petersen S, Thaesler Z, PacynaGengelbach M, van de Rijn M, Rosen GD, Perou CM, Whyte RI, Altman RB, Brown PO, Botstein D, Petersen I. 2001. Diversity of gene expression in adenocarcinoma of the lung. Proc Natl Acad Sci USA 98:13784–13789. Gelpi E, Ambros IM, Birner P, Luegmayr A, Drlicek M, Fischer I, Kleinert R, Maier H, Huemer M, Gatterbauer B, Anton J, Rossler K, Budka H, Ambros PF, Hainfellner JA. 2003. Fluorescent in situ hybridization on isolated tumor cell nuclei: a sensitive method for 1p and 19q deletion analysis in paraffin-embedded oligodendroglial tumor specimens. Mod Pathol 16:708–715. Germing U, Gattermann N, Strupp C, Aivado M, Aul C. 2000. Validation of the WHO proposals for a new classification of primary myelodysplastic syndromes: a retrospective analysis of 1600 patients. Leuk Res 24:983–992. Gershon D. 2002. Microarray technology: an array of opportunities. Nature 416:885–891. Giordano TJ, Shedden KA, Schwartz DR, Kuick R, Taylor JM, Lee N, Misek DE, Greenson JK, Kardia SL, Beer DG, Rennert G, Cho KR, Gruber SB, Fearon ER, Hanash S. 2001. Organ-specific molecular classification of primary lung, colon, and ovarian adenocarcinomas using gene expression profiles. Am J Pathol 159:1231–1238. Gleason DF, Mellinger GT. 1974. Prediction of prognosis for prostatic adenocarcinoma by combined histological grading and clinical staging. J Urol 111:58–64. Glinsky GV, Glinskii AB, Stephenson AJ, Hoffman RM, Gerald WL. 2004. Gene expression profiling predicts clinical outcome of prostate cancer. J Clin Invest 113:913–923. Golub TR, Slonim DK, Tamayo P, Huard C, Gaasenbeek M, Mesirov JP, Coller H, Loh ML, Downing JR, Caligiuri MA, Bloomfield CD, Lander ES. 1999. Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. Science 286:531– 537. Greenson JK, Bonner JD, Ben-Yzhak O, Cohen HI, Miselevich I, Resnick MB, Trougouboff P, Tomsho LD, Kim E, Low M, Almog R, Rennert G, Gruber SB. 2003. Phenotype of microsatellite unstable colorectal carcinomas: well-differentiated and focally mucinous tumors and the absence of dirty necrosis correlate with microsatellite instability. Am J Surg Pathol 27:563–570. Griffin CA, Emanuel BS. 1987. Translocation (X;18) in a synovial sarcoma. Cancer Genet Cytogenet 26:181–183. Guerin M, Barrois M, Terrier MJ, Spielmann M, Riou G. 1988. Overexpression of either c-myc or c-erbB-2/neu proto-oncogenes in human breast carcinomas: correlation with poor prognosis. Oncogene Res 3:21–31. Holloway AJ, van Laar RK, Tothill RW, Bowtell DD. 2002. Options available—from start to finish—for obtaining data from DNA microarrays. II. Nat Genet 32(Suppl):481–489.

Morphologic and Molecular Classification of Human Cancer Iacobuzio-Donahue CA, Hruban RH. 2003. Gene expression in neoplasms of the pancreas: applications to diagnostic pathology. Adv Anat Pathol 10:125–134. Ino Y, Zlatescu MC, Sasaki H, Macdonald DR, Stemmer-Rachamimov AO, Jhung S, Ramsay DA, von Deimling A, Louis DN, Cairncross JG. 2000. Long survival and therapeutic responses in patients with histologically disparate high-grade gliomas demonstrating chromosome 1p loss. J Neurosurg 92:983–990. Jacobs TW, Pliss N, Kouria G, Schnitt SJ. 2001. Carcinomas in situ of the breast with indeterminate features: role of E-cadherin staining in categorization. Am J Surg Pathol 25:229–236. Jancar J. 2000. Progress in the classification of hematological neoplasms: from REAL to WHO concept. Adv Clin Pathol 4:87–97. Jass JR, Biden KG, Cummings MC, Simms LA, Walsh M, Schoch E, Meltzer SJ, Wright C, Searle J, Young J, Leggett BA. 1999. Characterisation of a subtype of colorectal cancer combining features of the suppressor and mild mutator pathways. J Clin Pathol 52:455–460. Kader HA, Jackson J, Mates D, Andersen S, Hayes M, Olivotto IA. 2001. Tubular carcinoma of the breast: a population-based study of nodal metastases at presentation and of patterns of relapse. Breast J 7:8–13. Kallioniemi OP, Kallioniemi A, Kurisu W, Thor A, Chen LC, Smith HS, Waldman FM, Pinkel D, Gray JW. 1992. ERBB2 amplification in breast cancer analyzed by fluorescence in situ hybridization. Proc Natl Acad Sci USA 89:5321–5325. Kallioniemi OP, Kallioniemi A, Sudar D, Rutovitz D, Gray JW, Waldman F, Pinkel D. 1993. Comparative genomic hybridization: a rapid new method for detecting and mapping DNA amplification in tumors. Semin Cancer Biol 4:41–46. Kaneko Y. 1990. Cytogenetics in malignant lymphomas. Curr Opin Oncol 2:816–821. Karakousis CP, Dal Cin P, Turc-Carel C, Limon J, Sandberg AA. 1987. Chromosomal changes in soft-tissue sarcomas: a new diagnostic parameter. Arch Surg 122:1257–1260. Khan J, Saal LH, Bittner ML, Chen Y, Trent JM, Meltzer PS. 1999. Expression profiling in cancer using cDNA microarrays. Electrophoresis 20:223–229. Khan J, Wei JS, Ringner M, Saal LH, Ladanyi M, Westermann F, Berthold F, Schwab M, Antonescu CR, Peterson C, Meltzer PS. 2001. Classification and diagnostic prediction of cancers using gene expression profiling and artificial neural networks. Nat Med 7:673–679. Kim DW, Hwang JH, Suh JM, Kim H, Song JH, Hwang ES, Hwang IY, Park KC, Chung HK, Kim JM, Park J, Hemmings BA, Shong M. 2003. RET/PTC (rearranged in transformation/papillary thyroid carcinomas) tyrosine kinase phosphorylates and activates phosphoinositide-dependent kinase 1 (PDK1): an alternative phosphatidylinositol 3-kinase-independent pathway to activate PDK1. Mol Endocrinol 17:1382–1394. Kinzler KW, Vogelstein B. 1996. Lessons from hereditary colorectal cancer. Cell 87:159–170. Kitchen PR, Smith TH, Henderson MA, Goldhirsch A, Castiglione-Gertsch M, Coates AS, Gusterson B, Brown RW, Gelber RD, Collins JP. 2001. Tubular carcinoma of the breast: prognosis and response to adjuvant systemic therapy. Aust N Z J Surg 71:27–31. Kroll TG, Sarraf P, Pecciarini L, Chen CJ, Mueller E, Spiegelman BM, Fletcher JA. 2000. PAX8-PPARgamma1 fusion oncogene in human thyroid carcinoma [corrected]. Science 289:1357–1360. Lee Y, Lee CK. 2003. Classification of multiple cancer types by multicategory support vector machines using gene expression data. Bioinformatics 19:1132–1139. Liang P. 2002. A decade of differential display. Biotechniques 33:338–344, 346. Limon J, Turc-Carel C, Dal Cin P, Rao U, Sandberg AA. 1986. Recurrent chromosome translocations in liposarcoma. Cancer Genet Cytogenet 22:93–94. Lindor NM, Burgart LJ, Leontovich O, Goldberg RM, Cunningham JM, Sargent DJ, Walsh-Vockley C, Petersen GM, Walsh MD, Leggett BA, Young JP, Barker MA, Jass JR, Hopper J, Gallinger S, Bapat B, Redston M, Thibodeau SN. 2002. Immunohistochemistry versus microsatellite instability testing in phenotyping colorectal tumors. J Clin Oncol 20:1043–1048. Louis DN, Holland EC, Cairncross JG. 2001. Glioma classification: a molecular reappraisal. Am J Pathol 159:779–786. Marques AR, Espadinha C, Catarino AL, Moniz S, Pereira T, Sobrinho LG, Leite V. 2002. Expression of PAX8-PPAR gamma 1 rearrangements in both follicular thyroid carcinomas and adenomas. J Clin Endocrinol Metab 87:3947–3952. Martin-Subero JI, Gesk S, Harder L, Grote W, Siebert R. 2003. Interphase cytogenetics of hematological neoplasms under the perspective of the novel WHO classification. Anticancer Res 23:1139–1148.

19

McGuire WL. 1987. Prognostic factors for recurrence and survival in human breast cancer. Breast Cancer Res Treat 10:5–9. Mills SE, Bostwick DG, Murphy WM, Weiss MA. 1990. A symposium on the surgical pathology of the prostate. Pathol Annu 25(Pt 2):109–158. Moll R, Mitze M, Frixen UH, Birchmeier W. 1993. Differential loss of Ecadherin expression in infiltrating ductal and lobular breast carcinomas. Am J Pathol 143:1731–1742. Moskaluk CA, Zhang H, Powell SM, Cerilli LA, Hampton GM, Frierson HF Jr. 2003. Cdx2 protein expression in normal and malignant human tissues: an immunohistochemical survey using tissue microarrays. Mod Pathol 16:913–919. Mueller W, Hartmann C, Hoffmann A, Lanksch W, Kiwit J, Tonn J, Veelken J, Schramm J, Weller M, Wiestler OD, Louis DN, von Deimling A. 2002. Genetic signature of oligoastrocytomas correlates with tumor location and denotes distinct molecular subsets. Am J Pathol 161:313–319. Muhlmann M. 2002. Molecular cytogenetics in metaphase and interphase cells for cancer and genetic research, diagnosis and prognosis: application in tissue sections and cell suspensions. Genet Mol Res 1:117–127. Nikiforov YE. 2002. RET/PTC rearrangement in thyroid tumors. Endocr Pathol 13:3–16. Nikiforova MN, Biddinger PW, Caudill CM, Kroll TG, Nikiforov YE. 2002. PAX8-PPARgamma rearrangement in thyroid tumors: RT-PCR and immunohistochemical analyses. Am J Surg Pathol 26:1016–1023. Nutt CL, Mani DR, Betensky RA, Tamayo P, Cairncross JG, Ladd C, Pohl U, Hartmann C, McLaughlin ME, Batchelor TT, Black PM, von Deimling A, Pomeroy SL, Golub TR, Louis DN. 2003. Gene expression-based classification of malignant gliomas correlates better with survival than histological classification. Cancer Res 63:1602–1607. Osborne CK, Fisher E, Redmond C, Knight WA, Yochmowitz MG, McGuire WL. 1981. Estrogen receptor, a marker for human breast cancer differentiation and patient prognosis. Adv Exp Med Biol 138:377–385. Oue N, Hamai Y, Mitani Y, Matsumura S, Oshimo Y, Aung PP, Kuraoka K, Nakayama H, Yasui W. 2004. Gene expression profile of gastric carcinoma: identification of genes and tags potentially involved in invasion, metastasis, and carcinogenesis by serial analysis of gene expression. Cancer Res 64:2397–2405. Perou CM, Sorlie T, Eisen MB, van de Rijn M, Jeffrey SS, Rees CA, Pollack JR, Ross DT, Johnsen H, Akslen LA, Fluge O, Pergamenschikov A, Williams C, Zhu SX, Lonning PE, Borresen-Dale AL, Brown PO, Botstein D. 2000. Molecular portraits of human breast tumours. Nature 406:747–752. Perry A, Fuller CE, Banerjee R, Brat DJ, Scheithauer BW. 2003. Ancillary FISH analysis for 1p and 19q status: preliminary observations in 287 gliomas and oligodendroglioma mimics. Front Biosci 8:a1–9. Poetsch M, Dittberner T, Woenckhaus C, Kleist B. 2000. Use of interphase cytogenetics in demonstrating specific chromosomal aberrations in solid tumors—new insights in the pathogenesis of malignant melanoma and head and neck squamous cell carcinoma. Histol Histopathol 15:1225–1231. Press MF, Bernstein L, Thomas PA, Meisner LF, Zhou JY, Ma Y, Hung G, Robinson RA, Harris C, El-Naggar A, Slamon DJ, Phillips RN, Ross JS, Wolman SR, Flom KJ. 1997. HER-2/neu gene amplification characterized by fluorescence in situ hybridization: poor prognosis in nodenegative breast carcinomas. J Clin Oncol 15:2894–2904. Puxeddu E, Mitsutake N, Knauf JA, Moretti S, Kim HW, Seta KA, Brockman D, Myatt L, Millhorn DE, Fagin JA. 2003. Microsomal prostaglandin E2 synthase is induced by conditional expression of RET/PTC in thyroid PCCL3 cells through the activation of the MEK-ERK pathway. J Biol Chem 278:52131–52138. Ramaswamy S, Tamayo P, Rifkin R, Mukherjee S, Yeang CH, Angelo M, Ladd C, Reich M, Latulippe E, Mesirov JP, Poggio T, Gerald W, Loda M, Lander ES, Golub TR. 2001. Multiclass cancer diagnosis using tumor gene expression signatures. Proc Natl Acad Sci USA 98:15149–15154. Ramsay G. 1998. DNA chips: state-of-the art. Nat Biotechnol 16:40–44. Reifenberger J, Ring GU, Gies U, Cobbers L, Oberstrass J, An HX, Niederacher D, Wechsler W, Reifenberger G. 1996. Analysis of p53 mutation and epidermal growth factor receptor amplification in recurrent gliomas with malignant progression. J Neuropathol Exp Neurol 55:822–831. Rosai J. 1997. Pathology: a historical opportunity. Am J Pathol 151:3–6. Rowley JD. 1999. The role of chromosome translocations in leukemogenesis. Semin Hematol 36:59–72. Ross JS, Fletcher JA, Linette GP, Stec J, Clark E, Ayers M, Symmans WF, Pusztai L, Bloom KJ. 2003. The Her-2/neu gene and protein in breast cancer 2003: biomarker and target of therapy. Oncologist 8:307–325. Sandberg AA. 1991. Cytogenetics of human neoplasias: advances and perspectives. Prog Clin Biol Res 368:113–144. Schofield D, Triche TJ. 2002. cDNA microarray analysis of global gene expression in sarcomas. Curr Opin Oncol 14:406–411.

20

PART I: BASIC CONCEPTS

Shedden KA, Taylor JM, Giordano TJ, Kuick R, Misek DE, Rennert G, Schwartz DR, Gruber SB, Logsdon C, Simeone D, Kardia SL, Greenson JK, Cho KR, Beer DG, Fearon ER, Hanash S. 2003. Accurate molecular classification of human cancers based on gene expression using a simple classifier with a pathological tree-based framework. Am J Pathol 163:1985–1995. Shipp MA, Ross KN, Tamayo P, Weng AP, Kutok JL, Aguiar RC, Gaasenbeek M, Angelo M, Reich M, Pinkus GS, Ray TS, Koval MA, Last KW, Norton A, Lister TA, Mesirov J, Neuberg DS, Lander ES, Aster JC, Golub TR. 2002. Diffuse large B-cell lymphoma outcome prediction by gene-expression profiling and supervised machine learning. Nat Med 8:68–74. Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A, McGuire WL. 1987. Human breast cancer: correlation of relapse and survival with amplification of the HER-2/neu oncogene. Science 235:177–182. Smith JS, Perry A, Borell TJ, Lee HK, O’Fallon J, Hosek SM, Kimmel D, Yates A, Burger PC, Scheithauer BW, Jenkins RB. 2000. Alterations of chromosome arms 1p and 19q as predictors of survival in oligodendrogliomas, astrocytomas, and mixed oligoastrocytomas. J Clin Oncol 18:636–645. Sorlie T, Perou CM, Tibshirani R, Aas T, Geisler S, Johnsen H, Hastie T, Eisen MB, van de Rijn M, Jeffrey SS, Thorsen T, Quist H, Matese JC, Brown PO, Botstein D, Eystein Lonning P, Borresen-Dale AL. 2001. Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci USA 98:10869–10874. Sorlie T, Tibshirani R, Parker J, Hastie T, Marron JS, Nobel A, Deng S, Johnsen H, Pesich R, Geisler S, Demeter J, Perou CM, Lonning PE, Brown PO, Borresen-Dale AL, Botstein D. 2003. Repeated observation of breast tumor subtypes in independent gene expression data sets. Proc Natl Acad Sci USA 100:8418–8423. Sozzi G, Testi MA, Croce CM. 1999. Advances in cancer cytogenetics. J Cell Biochem 32–33(Suppl):173–182. Sreekantaiah C, Ladanyi M, Rodriguez E, Chaganti RS. 1994. Chromosomal aberrations in soft tissue tumors: relevance to diagnosis, classification, and molecular mechanisms. Am J Pathol 144:1121–1134. Stears RL, Martinsky T, Schena M. 2003. Trends in microarray analysis. Nat Med 9:140–145. Su AI, Welsh JB, Sapinoso LM, Kern SG, Dimitrov P, Lapp H, Schultz PG, Powell SM, Moskaluk CA, Frierson HF Jr, Hampton GM. 2001. Molecular classification of human carcinomas by use of gene expression signatures. Cancer Res 61:7388–7393. Tallini G. 2002. Molecular pathobiology of thyroid neoplasms. Endocr Pathol 13:271–288. Tallini G, Asa SL. 2001. RET oncogene activation in papillary thyroid carcinoma. Adv Anat Pathol 8:345–354. Tandon AK, Clark GM, Chamness GC, Ullrich A, McGuire WL. 1989. HER2/neu oncogene protein and prognosis in breast cancer. J Clin Oncol 7:1120–1128. Thiessen B, Maguire JA, McNeil K, Huntsman D, Martin MA, Horsman D. 2003. Loss of heterozygosity for loci on chromosome arms 1p and 10q in oligodendroglial tumors: relationship to outcome and chemosensitivity. J Neurooncol 64:271–278. Trent J, Casper J, Meltzer P, Thompson F, Fogh J. 1985. Nonrandom chromosome alterations in rhabdomyosarcoma. Cancer Genet Cytogenet 16:189–197.

Turc-Carel C, Dal Cin P, Limon J, Li F, Sandberg AA. 1986. Translocation X;18 in synovial sarcoma. Cancer Genet Cytogenet 23:93. Turc-Carel C, Limon J, Dal Cin P, Rao U, Karakousis C, Sandberg AA. 1986. Cytogenetic studies of adipose tissue tumors. II. Recurrent reciprocal translocation t(12;16)(q13;p11) in myxoid liposarcomas. Cancer Genet Cytogenet 23:291–299. Van den Bent M, Chinot OL, Cairncross JG. 2003. Recent developments in the molecular characterization and treatment of oligodendroglial tumors. Neurooncology 5:128–138. Van de Rijn M, Perou CM, Tibshirani R, Haas P, Kallioniemi O, Kononen J, Torhorst J, Sauter G, Zuber M, Kochli OR, Mross F, Dieterich H, Seitz R, Ross D, Botstein D, Brown P. 2002. Expression of cytokeratins 17 and 5 identifies a group of breast carcinomas with poor clinical outcome. Am J Pathol 161:1991–1996. Velculescu VE, Zhang L, Vogelstein B, Kinzler KW. 1995. Serial analysis of gene expression. Science 270:484–487. Von Deimling A, von Ammon K, Schoenfeld D, Wiestler OD, Seizinger BR, Louis DN. 1993. Subsets of glioblastoma multiforme defined by molecular genetic analysis. Brain Pathol 3:19–26. Vos CB, Cleton-Jansen AM, Berx G, de Leeuw WJ, ter Haar NT, van Roy F, Cornelisse CJ, Peterse JL, van de Vijver MJ. 1997. E-cadherin inactivation in lobular carcinoma in situ of the breast: an early event in tumorigenesis. Br J Cancer 76:1131–1133. Wahed A, Connelly J, Reese T. 2002. E-cadherin expression in pleomorphic lobular carcinoma: an aid to differentiation from ductal carcinoma. Ann Diagn Pathol 6:349–351. Watanabe K, Tachibana O, Sata K, Yonekawa Y, Kleihues P, Ohgaki H. 1996. Overexpression of the EGF receptor and p53 mutations are mutually exclusive in the evolution of primary and secondary glioblastomas. Brain Pathol 6:217–224. Weeraratna AT, Becker D, Carr KM, Duray PH, Rosenblatt KP, Yang S, Chen Y, Bittner M, Strausberg RL, Riggins GJ, Wagner U, Kallioniemi OP, Trent JM, Morin PJ, Meltzer PS. 2004. Generation and analysis of melanoma SAGE libraries: SAGE advice on the melanoma transcriptome. Oncogene 23:2264–2274. Wessels LF, van Welsem T, Hart AA, van’t Veer LJ, Reinders MJ, Nederlof PM. 2002. Molecular classification of breast carcinomas by comparative genomic hybridization: a specific somatic genetic profile for BRCA1 tumors. Cancer Res 62:7110–7117. World Health Organization. 1976. International Histological Classification of Tumors. Geneva. World Health Organization. 1990. ICD-O Second Edition International Classification of Diseases for Oncology. Geneva, WHO. World Health Organization. 2000. ICD-O Third Edition. International Classification of Diseases for Oncology. Geneva, WHO. Wright C, Angus B, Nicholson S, Sainsbury JR, Cairns J, Gullick WJ, Kelly P, Harris AL, Horne CH. 1989. Expression of c-erbB-2 oncoprotein: a prognostic indicator in human breast cancer. Cancer Res 49:2087–2090. Yang YH, Speed T. 2002. Design issues for cDNA microarray experiments. Nat Rev Genet 3:579–588. Zhang L, Zhou W, Velculescu VE, Kern SE, Hruban RH, Hamilton SR, Vogelstein B, Kinzler KW. 1997. Gene expression profiles in normal and cancer cells. Science 276:1268–1272.

3

Cancer Precursors THOMAS E. ROHAN, DONALD E. HENSON, EDUARDO L. FRANCO, AND JORGE ALBORES-SAAVEDRA

C

ancer is thought to arise as a result of a lengthy (up to several decades), multistep process involving changes in a number of genes following the initial clonal expansion of a mutated stem cell (Nowell, 1976; Kern, 1993; Vogelstein and Kinzler, 1993; Boone et al., 1999). These genotypic changes are accompanied by changes in cell and tissue morphology, characterized by loss of cellular differentiation and increased cytologic atypia. As a result, the tissue assumes progressively more of the morphologic characteristics of cancer. Although it is usually not difficult to make a clinical diagnosis of cancer, from the perspective of tumor biology the demarcation between what is unmistakably cancer and what precedes cancer is not entirely clear (Pontén, 1998). Perhaps not surprisingly, therefore, various definitions of precursor lesions have been proposed. For example, early during the last century it was suggested that a useful definition of a cancer precursor is “a condition which may be associated with development of cancer” (Stout, 1932). Subsequently, the term “precancerous lesions” was used to describe “visible steps in a dynamic process of neoplasia” that “may or may not undergo progression to a more advanced stage of neoplasia” (Foulds, 1958). The apparent intent of these definitions is to suggest that, even if a tissue abnormality at a specific anatomic site is associated with cancer development at that site, cancers at that site do not necessarily develop from such lesions, and that such abnormalities do not necessarily progress to cancer. In this chapter, the term “cancer precursors” is used to refer to all morphologic lesions on the pathway from normal tissue to cancer, up to but not including invasive cancer itself. That is, essentially we have adopted a morphologic definition of cancer precursors, and we have included within this definition carcinomas in situ, which are lesions with morphologic hallmarks of invasive cancer but that are confined to the thickness of the epithelium and do not penetrate the basement membrane. (Our focus here is on precursors of epithelial malignancies because precursors of other types of malignancy [mesenchymal, hematopoietic, lymphoid] are not well defined.) Other terms used to denote similar conditions include “incipient neoplasia” (Henson and Albores-Saavedra, 2001), “precancer” (Pontén, 1998), and “intraepithelial neoplasia” (O’Shaughnessy et al., 2002). Less commonly, the term “precancerous states” has been used to denote all conditions up to but not including carcinoma in situ (Carter, 1984), the latter thought to represent preinvasive cancer rather than a precancerous lesion. Increasing attention is being devoted to the study of cancer precursors, as evidenced by a number of recent publications on the topic (Pontén, 1998; Srivastava et al., 1999; Henson and Albores-Saavedra, 2001; Franco and Rohan, 2002). As stated in one of those publications, “Until now there has been such a heavy emphasis on cancer that we are only at a beginning in understanding precancer” (Pontén, 1998). In part, this reflects the difficulty of defining the natural history of cancer and of establishing, therefore, whether a given lesion is a cancer precursor. Indeed, validation of cancer precursors as intermediate end points for invasive cancer is challenging, as tissue sampling interrupts or alters the natural history of the neoplastic process. Furthermore, it is often difficult, if not impossible, to undertake the repeated tissue sampling that is required to study progression (and indeed regression to earlier stages), although in some cases alternative

Portions of this chapter were adapted with permission from Cancer Precursors: Epidemiology, Detection, and Prevention. E.L. Franco and T.E. Rohan, editors. Copyright © 2002. Springer-Verlag New York Inc., www.springer-ny.com

sampling techniques (e.g., Papanicolaou smears) might provide the possibility of doing so without encountering such problems. However, cytopathologic and even histologic examinations are prone to error owing to inadequate sampling and/or to incorrect microscopic interpretation. Nevertheless, the study of cancer precursors is important for several reasons. First, elucidation of the etiology of precursors provides insight into the etiology of the corresponding cancer, because if the precursor represents an intermediate stage in the causal pathway between exposure and the development of invasive cancer, etiologic factors for the former must be a subset of those for the latter. Second, if etiologic investigation of cancer precursors identifies potentially modifiable risk factors, it can provide opportunities for the primary prevention of both the precursors and the corresponding cancer. Third, if cancer precursors are clearly defined, they can provide targets for screening and hence early detection of those at increased risk of cancer, with obvious implications for the clinical management of individuals identified with such lesions. Finally, study of the molecular and genetic changes that occur with cancer precursors can provide fundamental insights into the nature of the carcinogenic process and may have practical benefits with respect to the classification of lesions and their clinical management. Our knowledge of the etiology of cancer precursors varies considerably by anatomic site. In part, this might reflect the relative inaccessibility of a site to tissue sampling (e.g., ovary and pancreas) and hence the difficulty of detecting and diagnosing precursors at that site. For some cancers (e.g., ovary), it might also be indicative of a relatively short premalignant phase, which therefore eludes detection. Furthermore, for some anatomic sites, it might also reflect the fact that it is only recently that we have identified putative precursors for the corresponding invasive cancer, as in the case of precursors of prostatic and pancreatic cancer (i.e., prostatic and pancreatic intraepithelial neoplasia, respectively). Nevertheless, as implied earlier, clues to the etiology of such conditions come in part from knowledge of the etiology of the corresponding cancer, which often is easier to study. Given the uneven state of knowledge of the etiology of cancer precursors by anatomic site, in this chapter we describe in some detail a few of the more well studied sites, for which there is substantial knowledge of the histopathology, epidemiology, and natural history. To date, relatively few studies have been undertaken to test the effect of preventive interventions on the risk of cancer precursors. However, with increasing recognition of the advantages that intermediate (or surrogate) end points offer for the study of preventive strategies in terms of reductions in time, sample size, and cost compared with the corresponding requirements for trials involving cancer as an end point, we are witnessing a burgeoning of activity in this area. One important approach is cancer chemoprevention, a relatively new area of research that involves the use of “agents that prevent cancer by either preventing or treating premalignant lesions” (Lippman et al., 1998). Although screening is usually targeted to the detection of cancerous lesions at a relatively early stage, screening at some sites (in particular, the cervix and colon) results in detection of a substantial proportion of cancer precursors. In principle, treatment or removal of such lesions results in a reduced risk of subsequent invasive cancer. For example, detection and ablative treatment of cancer precursors in the uterine cervix have resulted in a marked reduction in the incidence

21

22

PART I: BASIC CONCEPTS

of invasive squamous cell carcinoma (Miller et al., 2000); and removal of colorectal polyps has been associated with reduced risk of subsequent colorectal cancer (Muller and Sonnenberg, 1995). With further development of or improvement in screening modalities, it can be anticipated that detection of lesions at relatively early stages of carcinogenesis will increase. In some cases, this might pose diagnostic and therapeutic dilemmas, the former because of the need to establish that newly identified conditions (e.g., at sites that were previously inaccessible) are indeed cancer precursors and the latter because of concern over unnecessary treatment of lesions that might never progress. Indeed, the advent of new, sensitive technologies for cervical cancer screening has led to a debate as to the possibility that highgrade cervical lesions detected by human papillomavirus testing, for example, are less likely to progress than those detected by the traditional Papanicolaou test. Ongoing randomized trials are presently attempting to resolve this issue (Franco, 2003). The advent of the genetic era has spawned many studies of the molecular changes that characterize histologically defined cancer precursors (Srivastava et al., 1999). Such studies should provide insight into the progressive accumulation of the fundamental molecular changes leading to cancer. In addition, the results of such studies should lead to the development of new, molecular-based classifications of cancer precursors. Furthermore, they might have clinical implications (Ahrendt and Sidransky, 1999) because for women who are identified as being at increased risk of progression to invasive cancer (based on their status with respect to one or more molecular markers) close follow-up and early intervention might be warranted (Rohan et al., 1998). Also, such studies may lead to the identification of chemopreventive agents that target cellular or molecular alterations in preinvasive lesions (Franco and Rohan, 2002; Kelloff et al., 2003). Although it is only relatively recently that cancer precursors have been subjected to systematic study, it is clear that considerable progress has been made already in our understanding of these lesions. Our purpose in this chapter is to describe some of that progress.

TERMINOLOGY As indicated earlier, the term “cancer precursor” refers to specific morphologic changes that precede the development of cancer. The term does not imply that cancer is inevitable; rather, it refers to histologic changes associated with an increased probability or risk for cancer. These histologic changes are designated by morphologic terms that convey a mixture of diagnostic, prognostic, and etiologic significance. Such terms have included “atypical hyperplasia,” “mild, moderate, or severe dysplasia,” “epithelial atypia,” “high grade or low grade intraepithelial lesion,” “in situ carcinoma,” “intramucosal carcinoma,” “borderline tumor,” “grade one-half carcinoma,” “intraepithelial neoplasia,” and “minimal cancer.” Terms such as “actinic keratosis” and “arsenical keratosis” reflect the etiology of these cutaneous in situ carcinomas (Salasche, 2000). The most common diagnostic terms applied to precursors are “dysplasia” and “intraepithelial neoplasia,” which are often used synonymously. Dysplasia literally means disorganized cell proliferation and is usually characterized by abnormal epithelial maturation. For some sites, “intraepithelial neoplasia” is standard (e.g., “prostatic intraepithelial neoplasia” and “cervical intraepithelial neoplasia”). Unless the diagnosis is qualified by “mild,” “moderate,” “severe,” or some other modifier such as high-grade or low-grade, these terms alone do not provide information about the risk of progression or how advanced the lesion is, which are important parameters useful to clinicians and investigators.

abnormality, the risk of cancer increases. It is this outcome that allows us to infer that agents that induce precursor lesions are carcinogens. Morphologically, precursor lesions exhibit a continuum of histologic abnormalities. Moreover, this morphologic variability accompanies biologic variability. Not only do precursor lesions vary according to the tissue of origin, they also vary in size, rate of progression, rate of regression, rate of development, clinical presentation, and molecular profile. They may arise sporadically or may be genetically determined through germline mutations. Lesions may be multifocal or diffuse, often arising over wide areas of an epithelial surface.

Age of Onset Precursors arise in younger age groups than do invasive cancers. For most sporadic tumors, the peak age for the precursor is, on average, 10 years before the peak age for the invasive cancer. Thus, progression is a slow process, although there is variation among patients and presumably among histologic types of cancer. Many exceptions exist; for instance, there is evidence that progression to invasive carcinoma in the pancreas can take as long as 29 years (Brockie et al., 1998). On the other hand, genetically determined cancers tend to occur at an even younger age, often during early adult life. With familial adenomatous polyposis (FAP), adenomas appear at a mean age of 25 years, cancer at 39 years, and death due to the cancer at 42 years.

Location Precursor lesions are found primarily along epithelial surfaces, such as in the breast, uterine cervix, and prostate; throughout the gastrointestinal tract; along the respiratory mucosa; and in the urinary bladder. In contrast to precursors in epithelial tissues, the concept of cancer precursors is not well defined for mesenchymal and lymphoid tissues primarily because these tissues lack a basement membrane and their tumors are likely to be invasive from inception. However, progress has been made in identifying some of the precursor lesions for these sites, such as follicular hyperplasia in the stomach that may progress to malignant lymphoma when associated with Helicobacter pylori infection (Isaacson, 1999). Lymphoid hyperplasia associated with autoimmune diseases, viral infections such as Epstein-Barr virus infection, or immunodeficiency syndromes may also progress to malignant lymphoma. Follicular lymphoma in situ has been defined as involvement of single or scattered follicles in an otherwise normal lymph node (Beaty and Jaffe, 2001).

Size In general, precursor epithelial lesions are relatively small, occupying only the mucosal surface and, by definition, not violating the basement membrane. They may be flat or papillary. Often they are microscopic in size and not recognized on gross examination. Some lesions, however, can reach an unusually large size by growing along the epithelial surface. In situ carcinomas in the breast can be 5 cm in diameter as they wind their way through the mammary ducts. Flat in situ carcinomas of the urinary bladder may extend into the prostate gland and the seminal vesicles. Adenomas in the colon may measure 5 cm or larger in diameter. Mucinous cystic neoplasms of the pancreas of low malignant potential (borderline) are quite large but show only dysplastic and in situ changes along the surface epithelium. Some lesions can involve wide areas, such as diffuse metaplasia of the stomach. Lesions associated with germline mutations, such as with FAP, can extend along an entire epithelial surface.

Frequency GENERAL PROPERTIES OF CANCER PRECURSORS As morphologic risk factors for cancer, precursors have their own biology and natural history. Progression to invasive cancer is unpredictable and cannot be prognosticated reliably based on morphology. However, as a precursor progresses to successively higher grades of

Precursor lesions are more common than their corresponding invasive cancers. For example, sporadic colonic adenomas are more prevalent than the corresponding adenocarcinomas. Postmortem studies have shown that 25%–50% of the population have single or multiple adenomas in the colon by age 70 (Rickert et al., 1979; Williams et al., 1982). With the FAP syndrome, the colon contains hundreds of

Cancer Precursors adenomas, but only one (at most several) evolves into invasive cancer during the lifetime of the patient (Compton, 2001). In the lung, bronchial dysplastic changes are more common than invasive cancer (Kennedy et al., 1996; Park et al., 1999). In the skin, dysplastic and congenital nevi are more common than malignant melanomas, and actinic keratoses are more prevalent than squamous cell carcinomas (Schwartz, 1997). Most sporadic precursor lesions therefore do not progress to invasive cancer; progression may be more common with lesions secondary to germline mutations.

Multicentricity Precursor lesions are often multicentric and may occupy wide areas of a mucosal surface. Even a single microscopic lesion may indicate the existence of others nearby (Albores-Saavedra et al., 2000). In the lungs of smokers, for example, lesions are often multiple and bilateral. When multiple, they are often seen in various stages of development, which suggests that they do not all arise at the same time or progress at the same rate. Multiple lesions usually reflect chronic carcinogenic exposure and often arise in broad fields of exposed epithelium (Slaughter et al., 1953; Smith et al., 1996). Lobular carcinoma in situ of the breast is usually multicentric and frequently bilateral (Frykberg, 1999). Familial C-cell hyperplasia (medullary carcinoma in situ) of the thyroid is nearly always bilateral (Albores-Saavedra and Krueger, 2001). Precursor lesions may also coexist with invasive tumors. Pancreatic intraepithelial neoplasia is a multicentric lesion that often coexists with ampullary carcinoma, suggesting a field effect similar to that documented for urothelial carcinomas of the urinary bladder (Agoff et al., 2001).

Genomic Instability Most likely, progression is driven by an increased rate of unrepaired DNA damage with continued formation of abnormal genomic variants (Minna et al., 1997; Park et al., 1999; Hittelman, 2001). There is evidence that genetic changes antedate morphologic changes. For example, genetic changes have been found in normal-appearing mucosa of the lung and in the epithelium of the head and neck of smokers (Lydiatt et al., 1998; Park et al., 1999; Boyle et al., 2001). In the lung, these genetic changes resemble those seen in squamous cell carcinomas (Boyle et al., 2001). Genetic alterations have been found in colonic mucosa that shows no morphologic evidence of neoplastic transformation (Fearon and Vogelstein, 1990). In morphologically normal sun-exposed skin, multiple patches of keratinocytes with p53 mutations have been found (Ren et al., 1997). Approximately 100,000 times more common than dysplasia, these patches have practically no malignant potential. Molecular alterations associated with neoplasia that are found in normal tissues can be referred to as submorphologic precursors (Ren et al., 1997). Mutations are found in early precursor lesions. K-ras mutations, which are found in more than 80% of invasive carcinomas of the pancreas, have been detected in normal, hyperplastic, metaplastic, and neoplastic ductal pancreatic epithelium (Moskaluk et al., 1997; Sugio et al., 1997; Luttges et al., 1999). Genetic changes, especially 16p loss, have been found in cases of atypical hyperplasia of the breast (Gong et al., 2001). Genetic alterations have also been found in precursor lesions of the lung, especially loss of heterozygosity on chromosome 3 and mutations in p53 (Sozzi et al., 1992; Sundaresan et al., 1995). As a rule, genetic changes found in precursor lesions are also present in the corresponding invasive cancers. Alterations in gene expression and chromosome structure increase as lesions progress to invasive cancer (Wistuba et al., 1999). The more atypical lesions histologically are usually associated with more alterations (Wistuba et al., 1999).

Independence Multiple lesions arising over a broad mucosal surface are often independent because they have different genetic alterations. Independence has been extensively studied in the lung, head and neck, and urinary bladder (Sozzi et al., 1995; Barrera et al., 2001; Boyle et al., 2001;

23

Cheng et al., 2002). This independence may complicate chemopreventive interventions because different genetic alterations may indicate multiple pathways for malignant transformation, and it may not be possible to block all pathways with a single agent.

Heterogeneity Heterogeneity can be viewed as distinct subpopulations with differing selective growth advantages (Jotwani et al., 2001). These subpopulations, which reflect clonal evolution within a precursor lesion, are the result of genetic instability. Heterogeneity is reflected in all properties of precursor lesions. It applies to morphology, growth rate, karyotype, molecular abnormalities, surface antigens, and other biologic properties. As a result of heterogeneity, similar morphologic lesions may show variation in genetic alterations, progression rates, regression rates, and other biologic attributes. Furthermore, genetic changes that occur during progression of the precursor lesions to invasive cancer may differ among patients even though they have morphologically similar lesions. For instance, in dysplastic lung lesions, not all patients have similar genetic changes.

Preexisting Conditions Many preexisting conditions serve as risk factors because they may give rise to precursor lesions. Chronic inflammatory conditions (e.g., reflux esophagitis, ulcerative colitis, primary sclerosing cholangitis) and infections (e.g., those caused by oncogenic human papillomavirus types) may induce dysplastic changes that can progress to invasive carcinoma. Although these inflammatory or infectious conditions do not constitute neoplasia, they are risk factors for subsequent cancer and for this reason are considered precursors. Benign tumors such as colonic adenomas, which by definition are accompanied by dysplastic changes, may progress to in situ and intramucosal carcinoma. Proliferative lesions such as chronic lymphedema can give rise to lymphangiosarcomas. Immunodeficiency states may be associated with lymphoproliferative disorders and smooth muscle neoplasms (Monforte-Munoz et al., 2003). Rarely, other conditions may give rise to cancer as well. Fibrous dysplasia and Paget’s disease of bone, for example, may lead to osteogenic sarcoma (Unni and Dahlin, 1979; Fechner and Mills, 1993). Thus, the concept of precursors is broad, with multiple diverse conditions occasionally serving as a risk for cancer.

Progression Precursor lesions usually follow a morphologic sequence from hyperplasia or metaplasia through dysplasia to carcinoma in situ and subsequent invasion. Lesions that are more advanced along the sequence are likely to progress more rapidly than less advanced lesions. However, not all lesions follow the sequence, and some may never progress or may even reverse direction. There may be considerable variation among patients with respect to the time required for progression to invasive cancer. For instance, reports suggest that progression of precursor lesions of the pancreas may require many years—29 years in one case (Brat et al., 1998; Brockie et al., 1998). Progression of colonic adenomas to carcinoma has been estimated to require 10–15 years (Muto et al., 1975; Day and Morson, 1978). Progression may be rapid if in situ lesions already exist in the epithelial field. The rate of progression varies with the extent to which the cellular and architectural changes in the precursor lesion resemble those of the corresponding invasive cancer; such host factors as age, sex, ethnicity, and hormonal status; and the extent of genetic changes (Table 3–1). For acquired precursor lesions, there is currently no known specific genetic change that signals invasion. Indeed, constellations of molecular changes are often required for invasion. Rates of progression are difficult to estimate. Histologic grading of dysplastic lesions, for example, is subjective and often has low reproducibility. Also, changes in diagnostic criteria and terminology may affect the analysis of progression.

24

PART I: BASIC CONCEPTS

Table 3–1. Changes Occurring with Progression from Normal Bronchial Mucosa to Invasive Carcinoma in Central Bronchial Carcinogenesis Parameter Hyperproliferation 3p LOH 9p LOH p53 overexpression Rb expression Cyclin D1 overexpression Telomerase overexpression Bcl-2 overexpression Aneuploidy p53 mutation p16 loss FHIT loss 13q and 17p LOH 5p and 5q LOH

Normal Epithelium

Squamous Metaplasia

Low Grade Dysplasia

High Grade Dysplasia

Carcinoma In Situ

Invasive Carcinoma

+ + + + +

++ + + + ++ + +

++ ++ ++ ++ ++ + + + +

++ ++ ++ ++ ++ ++ + + ++ +

+++ +++ +++ +++ ++ ++ + ++ ++

+++ +++ +++ +++ ++

+

+ +

++ ++ +++

+++ ++ +++ ++ +++ +++ +

Source: Kerr (2001). Reproduced with permission from the BMJ Publishing Group. LOH, loss of heterozygosity.

Dysplasia is the sine qua non precursor lesion. In most anatomic sites, high grade dysplasia is considered an irreversible change that is truly neoplastic and the morphologic forerunner of most invasive epithelial tumors. Clinically, a diagnosis of high grade, severe, or grade III dysplasia usually indicates persistence of the lesion or subsequent progression in a large proportion of untreated patients. In some sites, however, high grade dysplastic lesions may regress (e.g., in the cervix) (Holowaty et al., 1999). Often high grade dysplasia is difficult to distinguish morphologically from carcinoma in situ. For this reason, these two precursors are often grouped together. Carcinoma in situ represents a dysplastic change that extends through the full thickness of the epithelium without penetrating the basement membrane. For some anatomic sites, metaplasia presages cancer because it seems to make tissues more susceptible to malignant transformation. An acquired condition, metaplasia is the physiologic transformation of one type of differentiated tissue into another, usually in response to chronic irritation. Often considered a phenotypic reactive change, metaplasia always precedes cancer at these sites. In some sites, metaplasia is considered a premalignant lesion, as with Barrett’s esophagus. Mutations have been found in the nonneoplastic mucosa of stomach showing intestinal metaplasia (Ochiai et al., 1996). Metaplasia influences the type of invasive cancer. For instance, in the gallbladder intestinal-type metaplasia gives rise to intestinal-type carcinoma, and in the cervix squamous cell metaplasia gives rise to squamous cell carcinomas.

CANCER PRECURSORS AT SPECIFIC ANATOMIC SITES As indicated earlier, knowledge of the etiology of cancer precursors varies by anatomic site. In this section we describe in detail a few of the better-studied sites, for which there is substantial knowledge not only of the epidemiology but also of the histopathology and natural history.

Oral Cavity It has long been known that a group of intraepithelial lesions clinically defined as leukoplakia, erythroleukoplakia, erythroplakia, and oral submucous fibrosis generally precede the onset of squamous cell carcinoma of the oral cavity. The histologic assessment of such lesions frequently reveals areas of dysplasia and sometimes of hidden carcinoma (Sankaranarayanan and Somanathan, 2002; Silverman, 2003). Although the knowledge base on the pathology, natural history, and epidemiology of precursor lesions of malignancies of the oral cavity is considerable, much remains to be studied.

Pathology Cancer precursors in the oral mucosa are frequently multifocal. Microscopically, the most common lesions are hyperkeratosis, dysplasia, and

carcinoma in situ (Luna et al., 2001). The terms squamous intraepithelial neoplasia I, II, and III are occasionally used as synonyms for mild, moderate, and severe dysplasia, respectively. The clinical diagnosis of oral cancer precursors requires histologic assessment to exclude malignancy and to define the extent of epithelial involvement (Sankaranarayanan and Somanathan, 2002). Hyperkeratosis is most commonly seen in the buccal mucosa, alveolar ridge, hard palate, and dorsal surface of the tongue; it is characterized by increased surface keratin and thickening of the underlying epithelium. Clinically, these hyperkeratotic lesions are usually described as leukoplakia, although biopsy is required for a more precise diagnosis. The extent of surface keratin seen clinically does not reflect the underlying cellular changes. Recent guidelines define leukoplakia as predominantly whitish lesions of the oral mucosa that cannot be ascribed to any other definable lesion. There are two clinical and morphologic subtypes: homogeneous and nonhomogeneous (Axell et al., 1996). The former is predominantly white with a flat, thin appearance and may exhibit shallow cracks. Nonhomogeneous leukoplakia can be white only or white and red, in which case it is called erythroleukoplakia, a more clinically relevant lesion. It can be irregularly flat, nodular, or exophytic. The term erythroplakia is used to denote oral lesions that appear as red patches or plaques that cannot be ascribed to any clinically or pathologically defined condition. Dysplastic lesions are similar to those seen in other sites, such as the lung and uterine cervix (Lumerman et al., 1995). However, lesions in the oral mucosa are more likely to accumulate surface keratin in response to injury, which imparts the white color. As with most dysplastic lesions, there are major maturation epithelial abnormalities that become more pronounced with progression (Crissman and Zarbo, 1989).

Descriptive Epidemiology The prevalence and incidence of leukoplakia in the general population are not known. Limited information exists on the prevalence in specific populations based on informal surveys. Leukoplakia is uncommon in young persons. Fewer than 1% of men less than 30 years of age are found with these lesions, but the prevalence can reach 8% among those above age 70 (Bouquot et al., 1986). The leukoplakia prevalence in high risk populations varies between 0.2% and 12.0% (Gupta et al., 1980; Kleinman et al., 1991; Zain et al., 1997; Silverman, 2003). The most common sites affected are the buccal mucosa, alveolar mucosa, and lower lip; lesions in the floor of the mouth, lateral tongue, and lower lip tend to exhibit dysplastic or malignant changes more frequently (Neville and Day, 2002). In high risk areas in India, the age-specific prevalence peaks at around 8% at ages 55–60 and then declines to the same levels as those observed in young adults, at 4% (Mehta et al., 1969; Sankaranarayanan et al., 2000). True incidence

Cancer Precursors rates are known from a few large cohort studies that were started during the 1960s in India. The age-standardized (world population of 1960) annual incidence rate of leukoplakia was 3.3/1000 in men and 1.9/1000 in women among more than 20,000 villagers in Kerala, India (Mehta et al., 1976). The International Agency for Research on Cancer’s (IARC) ongoing oral cancer screening study in that area found annual incidence rates of leukoplakia to be 5.5/1000 in males and 3.6/1000 in females (Sankaranarayanan et al., 2000; Sankaranarayanan and Somanathan, 2002). The variability in study design and diagnostic criteria accounts for much of the observed variations in incidence rates and prevalence, although differences among populations with respect to risk factor prevalence clearly explain the major differences between high risk and low risk areas. Much less is known concerning the epidemiology of erythroplakia. It is found much less frequently than leukoplakia. Erythroleukoplakia is even less common, perhaps because of difficulties related to lack of standardization in recognizing such lesions. Oral submucous fibrosis (OSF) seems to be more common in high risk areas in India and Southeast Asia. The prevalence of OSF in India is in the range of 0.2%–1.2%; and the annual incidence rates are 8–21/100,000 in men and 29–46/100,000 in women (Sankaranarayanan and Somanathan, 2002). As with leukoplakia, variations in morbidity rates are largely attributable to variability in the prevalence of risk factors and in disease definition. The fact that these lesions are more common in women may be due to the greater prevalence of nutritional deficiencies among women in India. The presence of leukoplakia often indicates underlying areas of dysplasia or invasive lesions upon histologic assessment. The likelihood varies with the anatomic site and is highest for lesions in the floor of the mouth (43%) and lowest for those in the retromolar area (11%) (Waldron and Shafer, 1975). On the other hand, more than twothirds of erythroplakia lesions contain areas of histologically verifiable dysplasia or carcinoma (Sankaranarayanan and Somanathan, 2002), which underscores the clinical relevance of these lesions and the need to confirm all diagnoses via biopsy.

Etiology In general, risk factors for leukoplakia and its related lesions are the same as those documented for invasive cancers of the oral cavity. Tobacco smoking and chewing are well established as the major risk factors for leukoplakia. The association between oral leukoplakia and tobacco consumption satisfies all conventional criteria for causality, being strong, consistent, temporally verified, biologically plausible, and with a demonstrable dose-response relationship (IARC 1985, 1986). Moreover, lesion risk decreases significantly upon cessation of tobacco consumption (Mehta et al., 1982; Gupta et al., 1986). The various forms of smokeless and smoking tobacco products used in Southeast Asia [i.e., betel quid, bidis, and betel nut (arecanut)] have been shown to influence the risk of leukoplakia, with more intensive consumption habits being associated with a relative risk in the double digits (Sankaranarayanan and Somanathan, 2002). Arecanut and betel quid chewing have emerged as important risk factors for OSF in India (Maher et al., 1994). Tobacco smoking is an important risk factor for leukoplakia and oral epithelial dysplasia in Europe and North America (Morse et al., 1996; Jaber et al., 1999). Alcohol consumption appears to exert a moderate effect on the risk of oral leukoplakia, although study results have been inconsistent (Sankaranarayanan and Somanathan, 2002). It is still uncertain whether alcohol acts independently of tobacco or acts to potentiate the effects of tobacco (van der Waal et al., 1997). As with oral cancers, consumption of fruits and vegetables seems to protect against the risk of leukoplakia (Gupta et al., 1999). Serum levels of vitamins A, B12, and C and b-carotene were found to be inversely associated with the risk of leukoplakia in a case-control study in India (Ramaswamy et al., 1996). Body mass index has also been shown to be inversely associated with the risk of leukoplakia, but concerns about confounding effects due to low socioeconomic status and undernutrition must be resolved before this variable can be considered as being on the causal pathway to risk (Hashibe et al., 2000).

25

There is some epidemiologic evidence to support an etiologic role for human papillomavirus (HPV) infection in oral cancers, but the overall picture is less coherent than the role these viruses play in anogenital cancer (Herrero et al., 2003). Two comprehensive literature overviews found that HPV DNA is present in biopsy specimens from 15%–20% of patients with leukoplakia, a figure that is intermediate between levels observed for normal oral mucosa and that affected by squamous carcinoma (Franceschi et al., 1996; Miller and White, 1996). Potentially precancerous oral lesions are also found more frequently in human immunodeficiency virus (HIV)-positive individuals. Hairy leukoplakia is a common finding in immunosuppressed HIV patients and is associated with abundant replication of the Epstein-Barr virus (Grbic and Lamster, 1997; Ikediobi and Tyring, 2002). Oral warts in these individuals are caused by oral HPV infection, a condition that has increased in frequency in the era of highly active antiretroviral therapy (Hille et al., 2002). Despite these clinical associations, there is no evidence that the incidence of oral cancer is increased in patients with long-term follow-up following HIV infection.

Progression to Cancer The propensity for progression to invasive carcinoma is relatively high for lesions showing leukoplakia, but rates vary widely as a function of study design, patient population, diagnostic criteria, follow-up time, and case selection for follow-up. Malignant transformation occurs in 0.1%–18.0% of precursor lesions after periods of 1–40 years (Silverman, 2003). An actuarial analysis of cohort studies from high risk areas indicated annual rates of progression of 0.1%–17.0%, depending on the prevalence of smoking and the study site. Hospital-based series tend to yield higher rates of progression than community-based cohorts (Sankaranarayanan and Somanathan, 2002). Lesion regression has been measured at 20%–30% over periods ranging from 1 to 30 years (Pindborgh et al., 1968; Silverman et al., 1976; Banoczy, 1977). Erythroplakias and erythroleukoplakias have the highest potential to become invasive because they contain more extensive areas of dysplasia and carcinoma in situ (Scully et al., 2003). Invasive carcinomas are associated 17 times more frequently with these lesions than with leukoplakias (Shafer and Waldron, 1975). Recently, there has been considerable interest in identifying molecular markers for the risk of progression in oral precursor lesions (Sudbø et al., 2001; Epstein et al., 2002). Accumulation of genomic deletions or amplification at specific sites in chromosomes 3 and 9 has been found to predict the subsequent risk of malignancy (Mao et al., 1996). Loss of heterozygosity in the latter sites and in chromosomes 4, 8, 11, 13, and 17 has been equated with progression to invasive carcinoma in 50% of cases within 5 years versus only 2% in those lacking such molecular changes (Partridge et al., 2000; Zhang et al., 2001).

Esophagus There are two main histologic types of cancer of the esophagus: squamous cell carcinoma and adenocarcinoma. Precursors of the former are not particularly well defined, although there is some suggestion that there is an ordered sequence of events leading from esophagitis (which is thought to develop as a result of chronic gastroesophageal reflux disease (GERD) (Sonnenberg and El-Serag, 1999) ), to low and high grade dysplasia, and then to squamous cell carcinoma (Vaughan, 2002). Esophagitis is generally thought to be a relatively early lesion in the natural history of adenocarcinoma despite some evidence to the contrary (Cameron and Arora, 2002). According to one model of the natural history of adenocarcinoma, individuals with esophagitis may develop specialized intestinal metaplasia (Barrett’s esophagus) as a metaplastic reaction to GERD and then progress through stages of low and high grade dysplasia before developing esophageal adenocarcinoma (Vaughan, 2002). As with cancer precursors at other anatomic sites (e.g., those of the cervix (Holowaty et al., 1999) ), low and high grade dysplasia appear to be capable of regressing to earlier stages (Miros et al., 1991; O’Connor et al., 1999). There has been little etiologic study of putative precursors of squamous cell carcinoma. The remainder of this section is devoted to a discussion of the epidemiology of Barrett’s esophagus.

26

PART I: BASIC CONCEPTS

A

Figure 3–1. Barrett’s esophagus. A. Villiform structures are lined by tall columnar cells with mucin-containing cytoplasm. B. This villiform structure contains goblet cells intermixed with the tall columnar cells. No dysplasia is seen. C. High-grade dysplasia. The surface epithelium and the glands contain highly atypical cells with large hyperchromatic and overlapping nuclei.

B

C

Pathology

Descriptive Epidemiology

Histologically, Barrett’s esophagus consists of two cell types: columnar mucin-producing cells and goblet cells that line villiform structures. Glands in the lamina propria are also lined by intestinalized columnar and goblet cells. Paneth and endocrine cells may also be present. With low grade dysplasia, the glands present in the lamina propria are lined by pseudostratified columnar or elongated cells with large hyperchromatic nuclei. These abnormal glands often resemble those of tubular adenomas of the colon with mild dysplasia. The dysplastic cells usually extend to the surface epithelium. With high grade dysplasia, there is greater cytologic atypia in both the glands and the surface epithelium, as well as increased mitotic activity (Fig. 3–1). Pathological evaluation is the basis for patient stratification and management. The histologic classification of Barrett’s dysplasia is identical to the classification of dysplasia in inflammatory bowel disease (Table 3–2). Standard definitions for grades of dysplasia have been published (Reid et al., 1988).

Barrett’s esophagus is usually diagnosed by biopsy during the course of esophagogastroduodenoscopy (EGD) for investigation of GERD, the symptoms of which include heartburn, regurgitation, and dysphagia. Estimates of the prevalence of Barrett’s esophagus among those with GERD have ranged from 6% to 14% (Gruppo Operativo per lo Studio delle Precancerosi delle’Esofago, 1991; Lieberman et al., 1997; Bersentes et al., 1998; DeVault and Castell, 1999; Voutilainen et al., Table 3–2. Classification of Epithelial Dysplasia in Barrett’s Esophagus Negative Indefinite for dysplasia Positive Low grade (mild/moderate dysplasia) High grade (severe dysplasia and carcinoma in situ)

Cancer Precursors 2000; Conio et al., 2001; Eloubeidi and Provenzale, 2001; Gerson et al., 2001; Shaheen and Ransohoff, 2002; van Sandick et al., 2002). The prevalence of Barrett’s esophagus increases with age (Cameron and Lomboy, 1992), and it is higher among males than females (DeMas et al., 1999; Campos et al., 2001; Gerson et al., 2001) and among whites than other ethnic groups (Kim et al., 1997). On a population basis, the prevalence of diagnosed Barrett’s esophagus appears to have risen substantially during the past few years; for example, in Olmsted County, Minnesota it increased from 22.6/100,000 in 1987 to 82.6/100,000 in 1998 (Conio et al., 2001). Although some of this increase probably reflects increased detection because of greater use of upper gastrointestinal endoscopy in more recent years (Conio et al., 2001), the sharp rise in the esophageal adenocarcinoma incidence rates over a similar period (Blot and McLaughlin, 1999) suggests that at least some of the increase in Barrett’s esophagus diagnoses may be real (Prach et al., 1997). However, it seems that many cases of Barrett’s esophagus remain undiagnosed because in the same Minnesota population in which an increase in the prevalence of diagnosed Barrett’s esophagus was demonstrated the prevalence of Barrett’s esophagus at autopsy was shown to be 376/100,000 during the mid-1980s, which is much higher than the prevalence of clinically diagnosed Barrett’s esophagus (Cameron et al., 1990).

Etiology There is now considerable evidence that Barrett’s esophagus has an acquired rather than a congenital etiology (Kim et al., 1997). Much of the emphasis in etiologic studies of Barrett’s esophagus has been on factors that have been shown to be associated with a risk of esophageal adenocarcinoma. Among the latter, major risk factors include GERD (Lagergren et al., 1999b) and a relatively high body mass index (BMI) (Lagergren et al., 1999a). Other risk factors for adenocarcinoma include cigarette smoking and alcohol consumption, both of which are associated with moderately increased risk (Vaughan, 2002), and infection with the cagA+ strain of Helicobacter pylori, which is associated with reduced risk (Chow et al., 1998). A relatively high BMI [weight (kg)/height (m2)] is associated with a substantial increase in the risk of adenocarcinoma (Vaughan, 2002). Moreover, although the relation between the BMI and Barrett’s esophagus has received little attention to date, several studies have presented evidence in support of a positive association between obesity and GERD (Fisher et al., 1999; Locke et al., 1999; Wajed et al., 2001; Vaughan, 2002), although one did not (Lagergren et al., 2000). To date, most of the emphasis in etiologic studies of Barrett’s esophagus has been on the role of GERD (Kim et al., 1997). Avidan and colleagues (2001, 2002) studied consecutive patients undergoing endoscopy to compare subjects with Barrett’s esophagus with those found not to have this condition. They observed that hiatal hernia and the duration or severity of acid reflux predisposed to Barrett’s esophagus (Avidan et al., 2002), whereas gastric surgery did not (Avidan et al., 2001). Others have also shown that the duration of GERD and the severity of symptoms are factors that predispose to the development of Barrett’s esophagus. For example, Eloubeidi and Provenzale (2001) observed an increased risk with increased frequency of heartburn, and Conio et al. (2002) showed that GERD symptom frequency and the presence of hiatal hernia increased the risk of Barrett’s esophagus. Gerson et al. (2001) showed that heartburn, nocturnal pain, and odynophagia were associated with increased risk. Campos et al. (2001) found that increased frequency of reflux episodes that lasted more than 5 minutes was associated with a twofold increase in the risk of Barrett’s esophagus, as did GERD symptoms of more than 5 years’ duration. Eisen et al. (1997) compared 79 patients with Barrett’s esophagus to 180 age-, gender-, and race-matched controls. This study had two control groups: one had undergone endoscopy for GERD, and the other had undergone endoscopy for other indications. Earlier age at reflux symptom onset, longer duration of symptoms, esophagitis, stricture, and ulceration were all associated with increased risk of Barrett’s esophagus, regardless of the control group. The GORGE consortium is a community-based study of 2641 consecutive patients undergoing elective endoscopy for GERD. In a report based on 701 of these patients, 77 of whom had Barrett’s esophagus, Lieberman et

27

al. (1997) found that the presence of GERD symptoms for 1–5 years was associated with an odds ratio of 3.0 for risk of Barrett’s esophagus; the ratio increased to 6.4 for those who had had symptoms for more than 10 years. The length of the esophagus showing the specialized metaplastic changes of Barrett’s esophagus seems to be related to the risk of esophageal adenocarcinoma. Individuals with long-segment Barrett’s esophagus (LSBE) (≥3 cm in length) are at higher risk than those with the more common short-segment Barrett’s esophagus (SSBE) (Hirota et al., 1999). In relation to this, Avidan et al. (2002) studied 502 consecutive patients with GERD determined by pH monitoring, 174 of whom had Barrett’s esophagus (67 SSBE, 107 LSBE). They showed that hiatal hernia and frequent reflux episodes had positive associations with the Barrett’s esophagus segment length (Avidan et al., 2002). The length of the longest reflux episode was also associated with a much higher risk of LSBE than SSBE. Cameron (1999) compared 46 patients with SSBE to 103 controls and found that 96% of the patients and 42% of the controls had a hiatal hernia. Fass et al. (2001) found a correlation between the amount of esophageal acid exposure and the length of Barrett’s esophagus. In general, although the severity of GERD symptoms and the presumed extent of acid exposure also seem to correlate with increased length of Barrett’s esophagus, this has not been a consistent finding across studies. Although the relation between BMI and Barrett’s esophagus has received little attention to date, several studies have presented evidence to support a positive association between obesity and GERD (Fisher et al., 1999; Locke et al., 1999; Wajed et al., 2001), although one did not (Lagergren et al., 2000). Vaughan (2002) reported data from a cross-sectional study that showed a twofold increase in the risk of developing Barrett’s esophagus in patients in the highest BMI quartile compared to those in the lowest quartile. The study by Avidan et al. (2002) found that smoking and high alcohol intake were associated with increased risk of Barrett’s esophagus, especially LSBE; in the same study, the use of nonsteroidal antiinflammatory drugs (NSAIDs) was not associated with Barrett’s esophagus. Vaughan (2002) has also reported an increased risk of Barrett’s esophagus among cigarette smokers. There is some evidence that CagA+ strains of Helicobacter pylori might be protective against GERD and Barrett’s esophagus. This finding perhaps reflects reduced acid production as a consequence of gastritis and gastric atrophy (Vicari et al., 1998; Loffeld et al., 2000; Vaezi et al., 2000).

Progression to Cancer Individuals with Barrett’s esophagus are at high risk of developing esophageal adenocarcinoma (Vaughan, 2002). Indeed, the risk of patients with Barrett’s esophagus developing adenocarcinoma has been estimated to be about 0.5%–1.0% per year, a 30- to 125-fold higher risk than that of the general population (Kim et al., 1997; Heath et al., 2000). The estimates of risk are somewhat imprecise because most studies have followed relatively small numbers of study subjects for short periods of time and therefore have observed only a few incident cancers each. In addition, some cohorts have included patients referred to specialist centers, as a result of which individuals with high grade dysplasia may have been overrepresented (O’Connor et al., 1999). As indicated earlier, the risk of progression to adenocarcinoma appears to be directly related to the length of esophagus showing the specialized metaplastic changes of Barrett’s esophagus (Hirota et al., 1999). It is also directly related to the extent of high grade dysplasia found in Barrett’s esophagus (Buttar et al., 2001). Progression from Barrett’s esophagus to high grade dysplasia has been estimated to take 9–13 years and from high-grade dysplasia to adenocarcinoma 3–4 years (O’Shaughnessy et al., 2002).

Colorectum Most colorectal cancers are thought to arise from adenomatous tissue, a progression referred to as the adenoma–carcinoma sequence (Morson, 1974; Hill et al., 1978). Histologically, the latter reflects the

28

PART I: BASIC CONCEPTS

progression of normal epithelium through a stage of epithelial dysplasia—the distinguishing histologic feature of colorectal adenomas (Compton, 2000)—and then on to invasive cancer. Recent data suggest that the adenomas themselves might be preceded by lesions called aberrant crypt foci (Bird, 1987; Takayama et al., 1998). There is little in the way of direct evidence for an adenoma–carcinoma sequence, but the designation of colorectal adenomas as cancer precursors is supported by the frequent presence of carcinomas within adenomas (Morson, 1974; Muto, 1989) and by the fact that the spectrum of somatic genetic changes observed in adenomas places them in an intermediate position in the progression from normal mucosa to invasive carcinoma (Baron, 2002). Furthermore, the distribution of large (>1 cm) adenomas in the bowel is similar to that of colorectal cancer (Konishi and Morson, 1982; Matek et al., 1986). Other potential colorectal cancer precursor lesions include hyperplastic polyps (Hamilton, 2001), the commonest type of polyp detected in the colorectum (Jass, 1991), and the morphologically similar serrated adenomas (Jass, 1999) as well as chronic inflammatory bowel disease, hamartomatous polyps, mixed adenomatoushyperplastic polyps, flat adenomas, and dysplastic aberrant crypt foci (Takayama et al., 1998; Hamilton, 2001). Epidemiologically, however, these lesions have been less well studied than adenomatous polyps, on which the remainder of this section focuses.

A

Pathology The formation of aberrant crypt foci (ACF) is considered one of the earliest histologic events in the development of colorectal cancer (Takayama et al., 1998; Fenglio-Preiser and Noffsinger, 1999). ACF are larger than normal crypts, stain more intensely with methylene blue, and are seen to have either a bulging or concave surface on endoscopy. ACF are found throughout the colon and rectum, although they are more common in the distal colon. Adenomas, which are thought to arise from ACF, are benign glandular neoplasms that arise from the intestinal mucosa and contain dysplastic epithelium. They may be solitary or multiple and sporadic or hereditary. Histologically, adenomas are classified as tubular (most common type), villous, or mixed tubulovillous (Fig. 3–2). Adenomas can grow to be more than 5 cm in diameter.

B

Descriptive Epidemiology Given that colorectal adenomas are usually asymptomatic, it is difficult to obtain unbiased estimates of their frequency of occurrence (Peipins and Sandler, 1994). Indeed, most of the available estimates have come from autopsy and screening studies, both of which usually involve selected groups of individuals. Autopsy studies have generally shown positive associations between the prevalence of adenomas and the risk of colorectal cancer in the corresponding underlying population (Correa, 1978; Clark et al., 1985), with prevalence estimates ranging from zero among the Bantu in South Africa (Bremner and Ackerman, 1970), a population at low risk of colorectal cancer, to between 40% (Blatt, 1961) and 60% (Stemmermann and Yatani, 1973) in the United States, a population at relatively high risk. The prevalence of adenomas increases with age, and in high risk populations (e.g., the United States) it has been observed to exceed 50% after age 65 (Neugut et al., 1993). Also, the prevalence of adenomas is generally higher in males than in females, particularly among the middleaged (Neugut et al., 1993). There appear to be no estimates of the incidence of adenomas (Peipins and Sandler, 1994).

Figure 3–2. Tubular adenoma of the colon. A. Tiny tubular adenoma in a patient with familial adenomatous polyposis. Although the normal architecture of the mucosa is maintained, some superficial crypts show adenomatous changes that are easily distinguished from the adjacent normal glands. B. Tubular adenoma with intramucosal carcinoma. A cribriform structure is seen in the lamina propria underneath a tubular adenoma. C. Tubular adenoma and infiltrating adenocarcinoma. The adenocarcinoma is composed of closely packed neoplastic glands different from those of the overlying tubulovillous adenoma.

C

Cancer Precursors

Etiology Much of the emphasis in etiologic studies of colorectal adenomas has been on diet, although other factors have been studied including alcohol consumption, cigarette smoking, BMI, physical activity, and use of hormone replacement therapy. Most of the evidence concerning the roles of these factors has accrued from case-control studies. Although recall bias may not be a major issue in case-control studies of adenoma, given that adenomas are often asymptomatic, selecting an appropriate comparison group can be problematic, given the need to enroll adenoma-free subjects from the same source population as the cases (Potter, 1996). Failure to accomplish this can lead to misclassification of the controls and as a consequence can bias estimates of association conservatively. Therefore, in many etiologic studies of polyps, cases have been subjects whose polyps were identified through sigmoidoscopy or colonoscopy, and controls have been subjects found on screening to be polyp-free. Essentially, therefore, many of the etiologic studies to date have focused on polyp prevalence rather than polyp incidence. Despite these caveats, many of the risk factors for colorectal adenomas appear to be similar to those for colorectal cancer (Baron, 2002). The risk of developing colorectal adenomas has been shown to be associated inversely with physical activity (Kato et al., 1990; Giovannucci et al., 1995, 1996; Sandler et al., 1995; Enger et al., 1997; Kono et al., 1999; Terry et al., 2002) and positively with the BMI and waist-to-hip ratio (Neugut et al., 1991; Shinchi et al., 1994; Giovannucci et al., 1995; Giovannucci et al., 1996; Bird et al., 1998; Kono et al., 1999); the associations were independent of each other in one study (Giovannucci et al., 1995) but not in another (Giovannucci et al., 1996). The association with obesity might result from hyperinsulinemia or insulin resistance (Kono et al., 1999), whereas the association with physical activity might result from a decrease in secondary bile acids and/or bowel transit time or from altered prostaglandin levels, with consequent inhibitory effects on colonic cell proliferation (Enger et al., 1997). Hormone replacement therapy, which can also affect bile acid profiles, has been associated with a decreased risk of colorectal adenoma (Potter et al., 1996; Peipins et al., 1997; Chen et al., 1998; Grodstein et al., 1998). Cigarette smoking has been found consistently to be associated with increased risk of colorectal adenomas (Martinez et al., 1995; Giovannucci and Martinez, 1996; Nagata et al., 1999; Almendingen et al., 2000; BreuerKatschinski et al., 2000; Inoue et al., 2000; Ulrich et al., 2001; Erhardt et al., 2002). The relation may be causal, given that it is strong, dosedependent, and generally persists after controlling for potential confounding variables such as diet and alcohol consumption (Giovannucci and Martinez, 1996). The dietary factors studied most extensively in relation to adenoma risk include fat and fiber intake, as well as intake of fruits and vegetables. To date, the findings have been considerably inconsistent. Specifically, although some case-control studies (Giovannucci et al., 1992; Sandler et al., 1993; Martinez et al., 1996; Almendingen et al., 2002) and cohort studies (Giovannucci et al., 1992) have suggested that risk is increased in association with total fat intake, other casecontrol studies have not found an association (Macquart-Moulin et al., 1987; Benito et al., 1993; Little et al., 1993; Olsen et al., 1994; BreuerKatschinski et al., 2001; Nagata et al., 2001; Voskuil et al., 2002). Several case-control and prevalence studies have shown inverse associations between total fiber intake or intake of cereal, vegetable, and/or fruit fiber and the risk of developing colorectal adenoma (Benito et al., 1993; Martinez et al., 1996; Breuer-Katschinski et al., 2001; Almendingen et al., 2002; Peters et al., 2003), whereas others, both case-control studies (Little et al., 1993; Platz et al., 1997) and cohort studies (Fuchs et al., 1999), have shown either equivocal findings or no association. Findings for fruit and vegetable intake have also been variable, with some studies showing inverse associations with either or both factors (Kato et al., 1990; Benito et al., 1993; Sandler et al., 1993; Platz et al., 1997; Almendingen et al., 2001) or no association (Nagata et al., 2001; Senesse et al., 2002; Smith-Warner et al., 2002). Other dietary factors that have been examined in relation to adenoma risk include red meat and fish/poultry intake, for which there

29

is some evidence for positive and inverse associations, respectively (Yoon et al., 2000), and folate intake, for which there is some evidence for an inverse association, especially in conjunction with relatively high alcohol consumption (Giovannucci et al., 1993; Bird et al., 1995; Boutron-Ruault et al., 1996). Alcohol consumption itself has been associated with increased risk in some studies (Giovannucci et al., 1993; Martinez et al., 1995; Boutron-Ruault et al., 1996; Tiemersma et al., 2003), but other studies have shown null associations or only weak increases in the risk (Benito et al., 1993; Lubin et al., 1997; Nagata et al., 1999; Breuer-Katschinski et al., 2000). The potential mechanisms by which dietary factors might influence colorectal cancer risk have been discussed elsewhere (Potter, 1999).

Progression to Cancer Colorectal cancer is generally thought to develop over several decades (Giovannucci and Martinez, 1996). Although the time required for an adenoma to develop following an initiating event is unknown (Giovannucci and Martinez, 1996), it has been suggested that it is of the order of 5–20 years (O’Shaughnessy et al., 2002). With respect to the time required for progression from adenoma to cancer, the average age at diagnosis of colorectal adenoma patients is about 7–8 years less than that of colorectal cancer patients (Enterline, 1976). However, this is likely to underestimate the interval between the occurrence of a polyp and the subsequent occurrence of cancer given the inevitable uncertainty that exists regarding the time of onset of polyps (Muto et al., 1975). Indeed, evidence from studies of metachronous cancer rates and age distribution curves suggests that progression from adenoma to carcinoma requires 10–15 years on average (Muto et al., 1975; Day and Morson, 1978). Approximately 50% of patients with an adenomatous polyp develop a subsequent (recurrent) adenoma within 7.6 years (Yood et al., 2003). Furthermore, data collected prior to the routine use of sigmoidoscopy, colonoscopy, and polypectomy showed that the cumulative risk of colon cancer at intervals of 5, 10, and 20 years after discovery of an index polyp was 4%, 14%, and 35%, respectively (Stryker et al., 1987). The risk of patients with polyps developing colorectal cancer is two- to fourfold higher than that for the general population (Lotfi et al., 1986; Atkin et al., 1992; Otchy et al., 1996). The risk of progression is related to the size, histologic type, and degree of dysplasia in the index adenoma. It is higher for those with polyps that are large, have villous architecture, and exhibit severe dysplasia (Peipins and Sandler, 1994). Polypectomy is associated with a reduction in the incidence of colorectal carcinoma (Winawer et al., 1993; Zheng et al., 2002).

Breast Benign breast disease (BBD) is a heterogeneous condition consisting of many histologic entities (see below). The prevailing hypothesis concerning the natural history of breast cancer is that nonatypical proliferative forms of BBD, proliferative disease with atypia, and in situ cancer represent successive steps preceding the development of invasive breast carcinoma (Lakhani, 1999). This model is supported by experimental and epidemiologic evidence. Experimentally, xenografts of MCF10AneoT cells have been shown to progress from intraductal proliferative changes to lesions resembling atypical hyperplasia of the human breast and ultimately to lesions resembling carcinoma in situ (Miller et al., 1993). Also, a transgenic rat model (Davies et al., 1999) and a mouse model (Li et al., 2000) have demonstrated the stepwise development of breast cancer. In epidemiologic studies, the risk of subsequent breast cancer has been observed to be increased in women with proliferative epithelial disorders affecting the small ducts and the terminal ductal lobular units of the breast, particularly when epithelial proliferation is accompanied by evidence of atypia (Rohan and Kandel, 2002; Schnitt, 2003). The higher risk associated with atypia is consistent with the notion that it is more proximal to carcinoma than is proliferative disease without atypia. As a result of such findings, benign proliferative epithelial disorders (BPED) of the breast are thought to have malignant potential

30

PART I: BASIC CONCEPTS

(Wang and Fentiman, 1985). Our focus in the remainder of this section is on BPED of the breast.

Pathology Although many histologic entities are included in the rubric “benign breast disease,” the relevant lesions with respect to the risk of subsequent breast cancer are those of epithelial origin. In addition to hyperplasia with or without atypia, these lesions include sclerosing adenosis, solitary papilloma, and fibroadenoma (Bodian, 1993; Fitzgibbons et al., 1998). Ductal epithelial hyperplasias display a spectrum of changes ranging from mild to florid. They are classified further as proliferative disease without atypia or atypical ductal hyperplasia depending on the architectural patterns and the cytologic appearance of the cells. Atypical ductal hyperplasia (ADH) is considered to be the precursor of ductal carcinoma in situ (DCIS) (a spectrum of diseases characterized by noninvasive epithelial proliferation), and it has some morphologic features of the cribriform and micropapillary types of in situ carcinoma. It involves only a portion of a single duct or several ducts, of which the aggregate sectional diameter does not exceed 2 mm.

Descriptive Epidemiology An unknown proportion of women with benign breast disease come to clinical attention and proceed to biopsy (Rohan et al., 1998b). Therefore, it is difficult to measure the population prevalence of benign breast lesions overall or by histologic subtype. Nevertheless, estimates of the frequency of occurrence of BPEDs can be obtained from autopsy and epidemiologic studies. Autopsy studies of nonfatal conditions in which the included subjects represent an unselected series can provide estimates of the prevalence of the condition at death (Cook and Rohan, 1985). Data on the prevalence of BPEDs of the breast are available from several autopsy studies (Frantz et al., 1951; Sloss et al., 1957; Sandison, 1962; Humphrey and Swerdlow, 1966; Kramer and Rush, 1973; Sasano et al., 1978; Nielsen et al., 1984; Alpers and Wellings, 1985; Bhathal et al., 1985; Nielsen et al., 1987; Sarnelli et al., 1991). Although there were some differences between the studies in their use of histopathologic terminology, they do indicate that BPEDs of the breast are relatively common at death, with prevalence estimates ranging from around 5%–15% (Frantz et al., 1951; Sandison, 1962) to as high as 64% (Nielsen et al., 1987). Furthermore, in most of these series, the prevalence of proliferative epithelial disorders substantially exceeded that of occult carcinoma of the breast, suggesting that even if BPEDs of the breast are precursors of breast cancer they do not necessarily progress to cancer. There are no published estimates of the incidence rates of BPEDs of the breast. However, incidence rates of broader groupings of benign breast disease (e.g., fibrocystic disease or benign mammary dysplasia) have been shown with considerable consistency to increase rapidly with age until about 40–44 years, with peak incidence rates being somewhere between 200 and 400/100,000/annum, and to decrease rapidly thereafter (Ory et al., 1976; Cole et al., 1978; Brinton et al., 1981; Soini et al., 1981; Fleming et al, 1982). Nevertheless, the disease remains relatively common after menopause, with estimates of the annual incidence rate ranging from about 100/100,000 women during the early postmenopausal years to 20–30/100,000 women during the later postmenopausal years (Cook and Rohan, 1985).

Etiology Studies of the etiology of putative breast cancer precursors have focused largely on factors suspected to be involved in the etiology of breast cancer itself (e.g., menstrual and reproductive history, use of oral contraceptives and hormone replacement therapy, cigarette smoking, obesity, and more recently diet). This follows from the premise that if specific types of benign breast lesions are precursors of breast cancer, then factors related to the etiology of the former should be a subset of those related to the latter. There have now been several case-control studies (Lance, 1981; Soini et al., 1981; Parazzini et al., 1984, 1991; Berkowitz et al., 1985; Pastides et al., 1985; Bright et al., 1989; Ingram et al., 1989, 1991;

Rohan and Cook, 1989; Rohan et al., 1989a,; London et al., 1992; Minami et al., 1998) and cohort studies (Hsieh et al., 1984; Rohan et al., 1998b; Rohan, 1999; Friedenreich et al., 2000) on the etiology of BPEDs of the breast. These studies have reported on the risk of such lesions overall (Lance 1981; Soini et al., 1981; Parazzini et al., 1984, 1991; Pastides et al., 1985; Bright et al., 1989; Ingram et al., 1989, 1991; Rohan et al., 1989; London et al., 1992; Minami et al., 1998) or on the risk by the degree of epithelial proliferation or cytologic atypia displayed in the benign lesions (Hsieh et al., 1984; Berkowitz et al., 1985; Pastides et al., 1985; Rohan and Cook, 1989; Rohan et al., 1989). Of these studies, none of those that examined age at menarche showed an association with risk (Soini, 1981; Parazzini et al., 1984; Berkowitz et al., 1985; Pastides et al., 1985; Ingram et al., 1991; London et al., 1992; Minami et al., 1998; Rohan et al., 1998b); two (Lance 1981; Parazzini et al., 1984) observed positive associations with age at first pregnancy, whereas in the remainder there was no association (Soini et al., 1981; Hsieh et al., 1984; Berkowitz et al., 1985; Pastides et al., 1985; London et al., 1992); one (Minami et al., 1998) showed an inverse association with parity, whereas the reminder showed no association (Soini et al., 1981; Hsieh et al., 1984; Parazzini et al., 1984; Berkowitz et al., 1985; Pastides et al., 1985; Bright et al., 1989; Ingram et al., 1991; London et al., 1992; Rohan et al., 1998b; Minami et al., 1998); several (Rohan and Cook, 1989; Ingram et al., 1991; Rohan et al., 1998b) showed increased risk for those with a family history of breast cancer, whereas others (Berkowitz et al., 1985; Pastides et al., 1985; Bright et al., 1989; London et al., 1992) observed no association; three (Ingram et al., 1989; Rohan and Cook, 1989; Rohan et al., 1998b) yielded findings suggesting an inverse association with the BMI, whereas three others (Berkowitz et al., 1985; Pastides et al., 1985; Bright et al., 1989) did not observe such an association; and to date, no study has found as association with cigarette smoking (Berkowitz et al., 1985; Pastides et al., 1985; Rohan et al., 1989; Parazzini et al., 1991; Rohan, 1999). Other variables that have been studied include use of oral contraceptives (OCs), use of hormone replacement therapy (HRT), and diet. Two studies, one cohort study (Rohan and Miller, 1999b) and one case-control study (Rohan et al., 1992), have presented results for the association between OC use and risk of BPED, and several other studies (all case-control) have reported on the association between OC use and the risk of BBD by degree of histologic atypia (Li Volsi et al., 1978; Kampert et al., 1983; Pastides et al., 1983; Berkowitz et al., 1984; Hsieh et al., 1984; Rohan et al., 1992). With respect to the former, one study (Rohan and Miller, 1999a) showed that the risk of BPED was reduced in association with OC use; the other (Rohan et al., 1992) showed no association. Findings for the latter association have varied from those showing reduced risk of all grades of atypia (Kampert et al., 1983; Hsieh et al., 1984) to those showing no reduction in risk with any grade of atypia (Rohan et al., 1992). With respect to HRT use, one recent prospective study (Rohan and Miller, 1999b) observed increased risk of BPED of the breast in association with HRT use of more than 8 years, but an earlier case-control study, in which the association between HRT use and risk of BBD was examined by the degree of cytologic atypia, found no evidence for a linear relation between the degree of atypia and the risk (Berkowitz et al., 1984). The association between diet and risk of BPED has been examined in several case-control studies (Lubin et al., 1989; Hislop et al., 1990; Rohan et al., 1990; Ingram et al., 1991; London et al., 1993) and cohort studies (Rohan et al., 1998c). In two studies (Lubin et al., 1989; Hislop et al., 1990), there were positive associations between saturated fat intake (or indices thereof) and risk of atypical (Lubin et al., 1989) or proliferative (Hislop et al., 1990) forms of benign breast disease, whereas in the remainder (Rohan et al., 1990; Ingram et al., 1991; London et al., 1993; Rohan et al., 1998c) there was little support for an association with dietary fat. With respect to other nutrients, one study provided some evidence for inverse associations between retinol and b-carotene intake and risk (Rohan et al., 1990) and showed strong inverse associations with dietary fiber and its constituents (soluble and insoluble nonstarch polysaccharides and cellulose) (Baghurst and Rohan, 1995). Although these findings were supported to some extent

Cancer Precursors by those of another study (Ingram et al., 1991) in which risk of benign epithelial hyperplasia was reduced in association with consumption of fruit and leafy orange-red vegetables, in yet another study (London et al., 1992) carotene and retinol intake were not associated with the risk of atypical or nonatypical forms of BPED. One study (Rohan et al., 1998c) showed no association between dietary calcium intake and risk of BPED. Between-study differences in the results described above may have arisen from the fact that studies of biopsy-confirmed BBD (as with studies of other potentially premalignant conditions) are prone to selection bias because (as indicated earlier) an unknown proportion of women with BBD come to clinical attention. The differences, however, may also have resulted from differences in the distribution of case groups by histologic subcategory of BBD and also in the method of classifying BBD (Cook and Rohan, 1985). For these reasons, there is a need for studies in which the histologic classification is based on a standardized classification scheme and in which the problem of selection bias is minimized. The latter might be addressed by conducting studies in screened populations (Dubin and Pasternack, 1984).

Progression to Cancer Women with a history of BBD are at increased risk of developing breast cancer, and risk differs according to the histologic characteristics of BBD (Rohan and Kandel, 2002; Schnitt, 2003). Depending on the reference group used, the relative risk for subsequent development of breast cancer in women with proliferative disease without atypia have ranged from 1.3 (Palli et al., 1991; Dupont et al., 1993) to 7.3 (Minami et al., 1999) but mostly have been of the order of 2.0 (Rohan and Kandel, 2002). The relative risk for women with proliferative disease with atypia has ranged from 2.5 (Carter et al., 1988) to 16.0 (Minami et al., 1999), but most authors have reported values between 3.0 and 5.0 (Rohan and Kandel, 2002; Schnitt, 2003). Although there appears to be no information on the time required for the development of the early morphologic stages preceding breast cancer development (i.e., proliferative disease without or with atypia), estimates of the time interval between a diagnosis of proliferative disease with atypia and the development of invasive carcinoma are available. In a follow-up study of 150 women with atypical ductal hyperplasia (Page et al., 1985), 18 women developed invasive breast cancer during follow-up; the mean interval between the diagnosis of hyperplasia and diagnosis of the subsequent invasive breast cancer was 8.2 years (range 1.4–24.3 years), and the annual incidence rate of carcinoma in women with atypical ductal hyperplasia was 7.5/1000 women. The mean interval between DCIS and subsequent invasive breast cancer ranges from 6 to 10 years (Frykberg and Bland, 1993). However, as with other anatomic sites, there is some uncertainty in the foregoing estimates given the difficulty of pinpointing the time of onset of the various lesions (Frykberg and Bland, 1993).

Uterine Cervix Cervical cancer precursors have by far been the most studied of all preinvasive neoplastic conditions, thanks to the work of Papanicolaou more than 50 years ago that led to the widespread acceptance of cervical cytology, or the Pap test, as the most established medical screening test (Papanicolaou, 1954). Indeed, the designation “cancer precursors” evokes the notion of cervical cancer prevention, and in many respects the study of precursor lesions of the uterine cervix has been a paradigm for studying the etiology and natural history of cancer. Cervical cancer ultimately arises from two main histologic lineages depending on whether its precursors originate in squamous or glandular cervical epithelium. Squamous and glandular precursors are described separately.

Squamous Lesions of the Uterine Cervix About 80% of cervical cancers are squamous cell in origin (Platz and Benda, 1995). The natural history of cervical cancer begins as a slow

31

process of disruption of the normal maturation of the transformation zone epithelium of the uterine cervix. Historically, this preinvasive phase has been known variably as dysplasia or dyskariosis under the traditional Pap cytology nomenclature (Papanicolaou, 1954), as cervical intraepithelial neoplasia (CIN) according to the classification scheme of the World Health Organization (WHO) (Richart, 1968), or as squamous intraepithelial lesion (SIL) by the more recent Bethesda classification system (Solomon et al., 1989; Kurman et al., 1991). The latter underwent further revisions in 2001 and is essentially a twotiered system consisting of low grade SIL (LSIL) and high grade SIL (HSIL), supplemented by a dichotomous equivocal atypia category of atypical squamous cells (ASCs), which is qualified either as “of undetermined significance” (ASC-US) or “cannot exclude HSIL” (ASC-H) (Solomon et al., 2002). The Bethesda system has brought the old classification nomenclatures in line with the new knowledge concerning the role of HPV infection as the central causal agent in the genesis of cervical cancer. HPV-associated changes in the absence of other squamous abnormalities are classified as LSIL (which includes CIN 1). More advanced degrees of dysplasia (originally graded as moderate and severe, corresponding to CIN 2 and 3, respectively) and carcinoma in situ (CIS) (originally defined as a lesion encompassing the full thickness of the epithelium or, equivalently, CIN 3 in its most severe form) are included in the HSIL category as a single lesion grade. The Bethesda classification was developed exclusively for cytopathology use to serve as an analogue to the histopathologic classification based on CIN grades. However, owing to its widespread use in North America, the Bethesda classification has occasionally been used to denote histologically ascertained lesions. In Europe, many laboratories still favor the dysplasia/dyskariosis/CIS classification (i.e., mild, moderate, and severe dysplasia and CIS), which makes it difficult to compare the results of studies conducted in different geographic locales.

Pathology According to the extent of the cytologic atypia, cervical squamous dysplasia is categorized as mild (CIN I), moderate (CIN II), or severe (CIN III) (Fig. 3–3). Applying the Bethesda system to these lesions (Solomon et al., 1989; Kurman et al., 1991), mild dysplasia (CIN I) is classified as a low grade squamous intraepithelial lesion (LSIL), and moderate to severe dysplasia and carcinoma in situ (CIN II and CIN III) are classified as high grade squamous intraepithelial lesions (HSIL). The category of atypical cells of undetermined significance (ASCUS) is reserved for borderline cytologic changes. As indicated above, the Bethesda system is now widely used for reporting the results of screening. In biopsy specimens, increased mitotic activity, immature cell proliferation, incomplete or lack of maturation, nuclear pleomorphism, and clumping of the chromatin characterize dysplasia. The excess proliferation is first seen in the basal or reserve cells located along the basement membrane. Dysplasia is considered present when the abnormal cells occupy less than the full thickness of the cervical epithelium. Full-thickness involvement is considered to be carcinoma in situ (Kurman et al., 1992). Koilocytotic atypia, manifested by large cells with prominent cytoplasmic vacuoles and enlarged nuclei, is often a prevalent feature of dysplasia and usually indicates a cytopathic effect of HPV. As with other sites, squamous cell dysplasia is preceded by squamous metaplasia.

Descriptive Epidemiology In the United States, for each new case of invasive cancer found by Pap cytology screening there are approximately 50–100 other cases of smears consistent with ASC or SIL abnormalities (Franco and Ferenczy, 2002b). Triaging these women to the most appropriate management option has become a great problem, given the growing concerns with widespread malpractice litigation in the United States today. Epidemiologic surveillance of precursor lesions of cervical cancer has traditionally been based on monitoring incidence trends of histologically ascertained CIS using tumor registry data (Chow et al., 1986; Morrison et al., 1996; Bergstrom et al., 1999), which are routinely collected in some jurisdictions, and/or on prevalence data from

32

A

PART I: BASIC CONCEPTS

B

Figure 3–3. Cervical squamous dysplasia. A. Mild squamous dysplasia (CIN I) of the cervix. Some cells show large nuclei. B. Severe squamous dysplasia (CIN III) with koilocytotic cells. C. Squamous cell carcinoma in situ (CIN III).

cytopathology series (Bjorge et al., 1994; CDC, 1994; Lawson et al., 1998). The former has the advantage of being population-based and thus reflects time trends for an entire region, whereas the latter simply reflects the particular profile of the clientele attending a single screening facility or network of sites. Changes in lesion nomenclature over the years and the gradual loss of emphasis on CIS as the sentinel precursor stage have hampered surveillance activities. In the United States, the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) program routinely collected data on CIS and invasive cancer from 1973 onwards for at least nine population-based tumor registration jurisdictions in the states of Connecticut, Hawaii, Iowa, New Mexico, and Utah and in the metropolitan areas of Detroit, Atlanta, San Francisco-Oakland, and Seattle-Puget Sound. However,

C

as was expected, the SEER program stopped collecting data on CIS in 1996 because this histologically based definition was quickly abandoned following adoption of the Bethesda classification. Concerning cytopathology surveillance, the U.S. Centers for Disease Control and Prevention’s (CDC) National Breast and Cervical Cancer Early Detection Program (NBCCEDP), albeit not population-based, represents the largest cytopathology database in the United States. It is based entirely on the Bethesda classification system and meets high quality control standards (CDC, 1994). Given the variety of health care delivery settings providing information nationwide, the NBCCEDP provides a comprehensive portrait of the situation with respect to cervical cytology screening and has been used for surveillance activities (Lawson et al., 1998; Sawaya et al., 2000).

33

Cancer Precursors

tion plays a necessary causal role in cervical neoplasia has gained ground and has led to rethinking of the role of other risk factors as simply cofactors determining the risk of acquisition and persistence of HPV infection or in mediating lesion risk among HPV-infected women (Castelsague and Munoz, 2003; Wacholder, 2003; Tortolero-Luna and Franco, 2004). Environmental and life style characteristics currently supported by epidemiologic data as potential cofactors in HPV-related cervical carcinogenesis include smoking, high parity, use of OCs, dietary factors, and infection with other sexually transmitted agents such as Chlamydia and herpesvirus type 2 (Potischman and Brinton, 1996; Ho et al., 1998; Castellsague et al., 2002; Franco and Ferenczy, 2002b; Munoz et al., 2002; Castelsague and Munoz, 2003; Castle and Giuliano, 2003; Green et al., 2003). Host-related factors that seem to influence the early phases of cervical carcinogenesis include endogenous hormones, immunosuppressive conditions, and genetic susceptibility traits that affect antigen recognition and processing and DNA repair, such as specific HLA alleles and haplotypes and polymorphisms in the p53 tumor suppressor gene, respectively (Makni et al., 2000; Hildesheim and Wang, 2002; Wang and Hildesheim, 2003; Koushik et al., 2004). Factors related to HPV, such as HPV type and variant, viral load, and viral integration, have also been found to affect the natural history of precursor lesions (Xi et al., 1997; Villa et al., 2000; Ylitalo et al., 2000; Lorincz et al., 2002; Schlecht et al., 2003).

Cervical cytologic abnormalities are a common finding during opportunistic screening. Prevalence data from the NBCCEDP for the mid to late 1990s showed the following distribution of results: 5.2% ASC, 2.9% LSIL, 0.8% HSIL, and <0.1% invasive cancer. LSIL prevalence varied substantially by age: 6.8%, 3.1%, 1.6%, 0.9%, and 0.6%, for the age groups <30, 30–39, 40–49, 50–64, and ≥65 years, respectively. HSIL rates also decreased with age albeit less pronouncedly than those for LSIL (range 1.4%–0.3% as per the latter age groups) (Lawson et al., 1998). The contrast between peak age-specific incidence rates for CIS and invasive squamous cell carcinoma can be easily and more validly seen using SEER’s tumor registry data (Fig. 3–4). The incidence rate for CIS increased more steeply with age than that for invasive cancers, reaching a peak at age 25–29 years and then gradually declining at older ages. The incidence rate of invasive cancers leveled off after age 40–44, leaving an approximately 15-year gap between the peak incidence rates for CIS and for invasive cancer. Overall, the comparability among various populations with respect to the occurrence of CIN/SIL is affected by factors such as screening practices, proportion of the population covered by screening, quality of the cytopathology information, the cytomorphologic criteria used, and the target groups. In addition to age, the screening era is the single most important variable determining the relative prevalence of cervical cancer precursors. There seems to have been an increase in the detection rates of cervical dysplasia and CIS throughout the 1970s and 1980s, particularly among young women (Dietl et al., 1983; Learmonth et al., 1990; Utagawa et al., 1998; Sigurdsson, 1999). This increase probably can be attributed to two factors: a cohort effect caused by the increasingly larger proportion of women at high risk of lesion development in successive age cohorts (i.e., those with multiple sexual partners and an early sexual debut) and the intensification of screening efforts during the last 30 years in most Western countries, leading to more aggressive case finding of treatable precursors.

Progression to Cancer There has been much research on the propensity for progression and regression of preinvasive cervical lesions. However, the comparability of results across studies is hampered by methodologic problems such as small sample size, selection biases, insufficient follow-up time, use of biopsy to monitor lesions over time, and variable statistical methods for reporting rates of progression and regression. These drawbacks notwithstanding, it has been possible to learn from a number of meta-analysis overviews that most LSILs are transient, regressing to normal within a relatively short period of time. However, some progress to HSIL or to cancer over variable intervals. HSIL, on the other hand, carries a much greater probability of progressing to invasion, although most such lesions eventually regress. Early estimates indicated average probabilities of regression of 57% for CIN 1, 43% for CIN 2, and 32% for CIN 3 (Östor, 1993). The equivalent probabilities of progression to CIS were 11% for CIN 1 and 22% for CIN 2; the probabilities of progression to invasion were 1% for CIN 1, 5% for CIN 2, and 12% for CIN 3 (Östor, 1993). Other overviews indicated that the probabilities of regression, persistence, and progression to any higher grade lesion were in the ranges of 34%–45%, 31%–41%,

Etiology Epidemiologic research conducted during the past 30 years has been fairly consistent in demonstrating similar risk factors for cervical cancer and for its precursor CIN/SIL stages, although the strength of the epidemiologic associations seems to be somewhat weaker for precursor lesions than for invasive cancer. In fact, the similarity of risk factor profiles has lent credibility to the natural history model specifying that the abnormal changes seen in the cervical epithelium in reality follow a continuum leading to invasive cervical carcinoma. As mentioned, HPV infection plays a central causal role in the genesis of cervical carcinoma and its precursor squamous lesions LSIL and HSIL (Schiffman et al., 1993). More recently, the concept that HPV infec-

10

1

SCC CIS 4

+ 85

9

-8

80

4

-7

-7

75

9

70

65

4

Age at diagnosis

-6

9

-6 60

4

-5

55

9

-5

-4

50

4

45

9

-4

40

4

-3

35

9

-3

30

4

-2

-2

25

9

20

-1

-1

15

4

0.1 10

Age-specific incidence rate (x 100,000)

100

Figure 3–4. Age-specific incidence rates of invasive, squamous cell carcinomas (SCC) and preinvasive, carcinoma in situ (CIS) in all SEER registration areas during 1992–1999. Rates are standardized according to the age structure of the 2000 U.S. population and expressed per 100,000 women. (Source: Ries et al., 2002.)

34

PART I: BASIC CONCEPTS

and 23%–25%, respectively, for all grades of CIN combined depending on whether studies used cytology alone or cytology and biopsy to monitor lesion occurrence over time. They also showed that progression rates from CIS to invasive cancer ranged from 29% to 36% (McIndoe et al., 1984; Mitchell et al., 1994). By taking time into account, the average rates of progression to HSIL at 24 months according to baseline cytologic abnormalities have been estimated as 7.1% for ASC and 20.8% for LSIL, whereas the equivalent progression rates to invasive cancer were 0.3% for ASC, 0.2% for LSIL, and 1.4% for HSIL (Melnikow et al., 1998).

Glandular Lesions of the Uterine Cervix The original Bethesda classification included categories of glandular cytologic abnormalities indicative of endocervical or endometrial malignancy or its precursors (Solomon et al., 1998). In the 2001 revision, the Bethesda system incorporated specific categories for “atypical glandular cells” (AGCs) with specification of origin (endocervical, endometrial, or not otherwise specified); “AGC, favor neoplastic,” either endocervical or not otherwise specified; “endocervical adenocarcinoma in situ” (AIS); and “adenocarcinoma” (Solomon et al., 2002). AGC abnormalities originating from the endocervix can generally be recognized based on the larger nuclear size and more abundant cytoplasm compared to those whose origin was in endometrial cells. AIS appears cytomorphologically as sheets of packed glandular cells with clear pseudostratification, a high nuclear/cytoplasm ratio, nuclear hyperchromasia, feathering, and palisading borders (Meisels and Morin, 1997; Solomon et al., 1998). Endocervical cytology is difficult and is plagued by high falsenegative rates. Many glandular lesions are missed because their relative rarity precludes developing expertise in their recognition, and many AIS cases are discovered during histologic examination because of coexistent SIL, which is easy to recognize on cytology (Ferenczy 1997). Histologically, both AIS and adenocarcinoma may be classified as the endocervical type, endometrioid type, intestinal type, mixtures of the above, mucoepidermoid, clear-cell type, adenoid basal cell carcinoma, adenoid cystic carcinoma, adenoma malignum, and glassy cell carcinoma. There is no compelling evidence that the various histologic types can be reliably identified cytologically. Most AIS cases are not grossly visible, and only a small number are suspected at colposcopy. They produce no appreciable clinical symptoms or signs, such as vaginal bleeding. AIS involves the transformation zone in more than two-thirds of cases and is uncommonly multifocal (Ostor et al., 1984; Pacey, 1991; Ferenczy 1997).

Pathology Adenocarcinoma in situ is an uncommon but well defined preinvasive lesion of the endocervical glands. It is considered the preinvasive phase of the invasive endocervical adenocarcinoma. Four main histologic types have been described: endocervical, endometrioid, intestinal, and mixed or combined (Fig. 3–5). The neoplastic cells in the endocervical type are columnar and mucin-producing; in the endometrioid type the cells are similar to those seen in invasive endometrial adenocarcinoma; and the intestinal type consists of columnar cells, goblet cells, and occasionally Paneth and endocrine cells. The combined type has features of two or more of the previous types. Squamous cell dysplasia and carcinoma in situ of the cervix often coexist with adenocarcinoma in situ. This association suggests that the same etiologic agent may play a role in malignant transformation of the reserve cell, which is presumed to be the cell of origin for both types of cervical neoplasm.

Descriptive Epidemiology Adenocarcinoma in situ (AIS), a rare condition, is believed to be a precursor lesion that may progress to invasive adenocarcinoma. Only 0.7% of the 64,628 incident in situ cervical cancers registered in the SEER program between 1973 and 1987 were of adenocarcinoma (0.6%) or adenosquamous (0.1%) histology (Platz and Benda, 1995). They account for an average of about 30 new cases per year for the SEER areas or approximately 300 new cases per year if extrapolated to the entire U.S. population. Atypical glandular cell (AGC) diagnoses

A

B

Figure 3–5. A. Adenocarcinoma in situ of the cervix. B. Adenocarcinoma in situ of the cervix with partial involvement of the endocervical glands.

typically constitute less than 1% of a cytopathology laboratory workload (Solomon et al., 1998) but seem to be associated with greater risks of underlying lesions than their atypical squamous cell (ASC) counterparts. An AIS finding on cytology is confirmed on histology 48%–69% of the time and carries a positive predictive value of 38% for adenocarcinoma (Wright et al., 2002).

Etiology Risk factors for AIS and adenocarcinomas seem to be similar to those for cervical squamous carcinomas and include parity, early age at first intercourse, multiple sexual partners, and history of sexually transmitted diseases (Franco and Ferenczy, 2002b). There is increasing evidence supporting the hypothesis of a hormone–adenocarcinoma relationship. The evidence comes from the following findings: (1) AIS and adenocarcinomas are commonly encountered in pregnant women (Ferenczy, 1997); (2) adenocarcinoma seems to be associated with OC use (Chumas et al., 1985; Moreno et al., 2002) and estrogen replacement therapy (Lacey et al., 2000); and (3) adenocarcinomas often contain estrogen and progesterone receptors (Ford et al., 1983). The aforementioned variables are likely to act as cofactors of HPV infection, which (analogous to the situation with squamous lesions) is also the central etiologic agent in AIS and adenocarcinomas of the uterine cervix (Bosch and Sanjose, 2003). The prevalence of HPV DNA in AIS and adenocarcinomas ranges from 15% to 90% using Southern blot hybridization or the polymerase chain reaction. HPV 18 is the predominant type in most studies, in contrast to cervical squamous cell carcinomas, which contain HPV 16 as the prevailing type (Franco and Ferenczy, 2002b).

Cancer Precursors

Progression to Cancer Adenocarcinoma in situ is believed to be a precursor lesion that may progress to invasive adenocarcinoma. In most series, AIS is associated with HSIL in more than 60% of cases; and HPV DNA is found in up to 90% of coexistent lesions (Tase et al., 1989). It may be that both the squamous and glandular lesions develop through a process of bidirectional differentiation from subcolumnar reserve cells initially infected with HPV 18 or HPV 16 (Ferenczy, 1997). It is estimated that 13 years may elapse during the transition from AIS to adenocarcinoma, judging from the difference between the mean ages at diagnosis of patients with these two lesions as registered in the SEER program. This is somewhat shorter than the equivalent figure for the average transition between CIN and cervical squamous carcinoma (18 years) (Plaxe and Saltzstein, 1999).

Endometrium Endometrial carcinomas can be divided into two main types: endometrioid and nonendometrioid. The former are also called type 1 endometrial cancers and account for 80%–90% of all invasive malignancies of the uterine corpus. They seem to arise from unopposed estrogen stimulation and are typically responsive to the antiproliferative effects of progesterone (Ronnet et al., 2001). Type 2, or nonendometrioid, tumors are less frequent and have histologic features of serous, clear-cell, or mixed-type carcinomas; they are not linked to estrogenic effects. They tend to develop in the postmenopausal atrophic endometria of older women (Sherman, 2000; Franco and Ferenczy, 2002b). The serous types of carcinoma tend to be more aggressive than other histologic types. They are the most common nonendometrioid type of endometrial carcinoma. Endometrial hyperplasia (EH) with cytologic atypia is believed to represent the precursor lesion for type 1 endometrial carcinoma, whereas type 2 cancers are not associated with or preceded by EH; rather, they seem to arise from in situ carcinomas of the surface or glandular epithelium and are designated endometrial intraepithelial carcinoma (EIC) (Mutter, 2000; Mutter et al., 2000; Dietel, 2001; Inoue, 2001). Molecular genetic changes are also unique for each pathway (Matias-Guiu et al., 2001).

Pathology Endometrial hyperplasia is classified as typical or atypical, each of which in turn is further subclassified as simple or complex depending on glandular density, shape, and distribution. Recently, because of a lack of diagnostic reproducibility, there have been proposals to simplify this classification (Dietel et al., 2001). The designation “endometrioid neoplasia” has been frequently used for atypical and complex hyperplastic lesions, which are considered precancerous. Distinguishing atypical hyperplasia from well differentiated endometrioid carcinoma can be difficult. Typical EH is a benign lesion almost without risk (<3%) of progressing to cancer (Ronnnet et al., 2001). Indeed, these precursor lesions often regress. A new nomenclature has been proposed by the Endometrial Collaborative Group for type 1 lesions, which introduces the term endometrial intraepithelial neoplasia (EIN) to denote simple and complex EH with cytologic atypia (Franco and Ferenczy, 2002b). The EIN nomenclature makes a clear distinction between EH without atypia and EIN, the latter being considered a genuine precancerous state, whereas the former is not. It removes the emphasis from the distinction between simple and complex glandular architectures described above and establishes the importance of the relative fraction of the sectioned tissue that is occupied by stroma versus glands (Mutter, 2000; Mutter et al., 2000). Type 2, or nonendometrioid, tumors, particularly serous carcinomas, are preceded by EIC. They are characterized by glandular atrophy, anaplastic nuclei, abnormal mitotic figures, and apoptotic bodies.

Descriptive Epidemiology Screening for endometrial cancer and its precursors is not common medical practice. Therefore, detection of preinvasive endometrial

35

lesions results from incidental findings from histopathologic examination of specimens from women at high clinical risk or those presenting with symptoms (e.g., abnormal uterine bleeding) that warrant investigation. The SEER program classifies registered cases of preinvasive lesions as in situ cancers of the uterine corpus without specifying the histologic lineage as per the classification schemes described above. From 1973 to 1987 the SEER program registered 43,364 cases of invasive cancer and 2650 cases of in situ cancer of the uterine corpus (Platz and Benda, 1995). The annual incidence rate of invasive lesions peaked in 1975 at 32/100,000 women and that for in situ lesions in 1974 at 3.5/100,000 women (Ries et al., 2002; Franco and Ferenczy, 2002a). Thereafter, there was a pronounced decline in the incidence rate of in situ lesions, with estimates during the early 1990s that were 80% lower than when it peaked in 1974. Before peaking in the mid-1970s, the incidence rate of invasive cancer increased substantially during the late 1960s, presumably in response to the widespread use in North America of estrogen-only replacement therapy for postmenopausal symptoms. As soon as the association was recognized, the practice of prescribing estrogen without progestin declined dramatically during the late 1970s and the incidence rate of invasive cancer began to decrease shortly thereafter (Schottenfeld, 1995; Franco, 1997). The parallel decline in rates of in situ and invasive cancers supports the idea that the in situ lesion is a true precursor condition, logically sharing a common etiology with invasive adenocarcinomas. On the basis of SEER data, the incidence rate of endometrial neoplasms rises rapidly after the age of 40 years and then begins to decline after ages 70–74. During the above-mentioned highest risk period of 1973–1975, the incidence rate of invasive endometrial cancer peaked at ages 60–64 years, but the peak age-specific rate shifted to ages 65–69 during 1976–1985 and then to 70–74 years during 1986–1995. This shift was also seen for in situ lesions (Franco and Ferenczy, 2002a; Ries et al., 2002). The peak incidence rate of in situ lesions occurs about 5 years earlier than that for invasive cancers, which coincides with estimates of mean transit time from atypical hyperplasia (which presumably represents most of the in situ lesions registered in the SEER program) to endometrial carcinoma in natural history studies (Kurman et al., 1985; Ferenczy and Gelfand, 1989). Endometrial hyperplasia is a common finding in biopsies and curettage specimens from pre- or postmenopausal women with abnormal uterine bleeding. The prevalence of EH with or without atypia in such women is in the range of 5%–20%, whereas that of endometrial carcinoma ranges from < 1% to 18% (Feldman et al., 1995; Ben-Yehuda et al., 1998; Farquhar et al., 1999; Anastasiadis et al., 2000). Screening studies of asymptomatic women provide much lower figures. In a U.S. cohort study of 2586 asymptomatic postmenopausal women who were examined up to three times with endometrial sampling, the prevalence rate of endometrial cancers was 7 per 1000, and the annual incidence rate was 1.7 per 1000, with nearly identical rates of EH (Koss et al., 1984; Koss, 1995). Not surprisingly, because of the active screening, these incidence rates are much higher than the equivalent incidence rates in the SEER program, particularly for in situ lesions. Studies of cohorts in Sweden and Finland with intensive screening surveillance provided comparable figures for preinvasive and invasive lesions (Gredmark et al., 1999; Vuento et al., 1999). Clinical trials studying postmenopausal hormone therapy that measured the cumulative incidence of EH tended to find higher cumulative incidence rates for women on unopposed estrogens (range 15%–38% up to 1 year) than in those taking estrogen/progesterone combinations (range 0.5%–5.6% in up to 4 years) (Woodruff and Pickar, 1994; Archer et al., 1999; Bjarnason et al., 1999; Bergeron and Fox, 2000; Kurman et al., 2000).

Etiology The risk for developing endometrial cancer and its precursor lesions is influenced by reproductive and hormonal factors. In general, the risk factor profiles for invasive endometrial neoplasms and EH are similar (Baanders-van Halewyn et al., 1996; Sturgeon et al., 1998). Nulliparity (Weber et al., 1999), high socioeconomic status (Daly et al., 1995), obesity (Kreiger et al., 1986; Baanders-van Halewyn et al., 1996; Gredmark et al., 1999; Weber et al., 1999), and smoking (inverse

36

PART I: BASIC CONCEPTS

association) (Brinton et al., 1993; Weir et al., 1994) are prominent risk determinants. Case-control and cohort studies have demonstrated the effects of estrogen replacement therapy on the risk of developing both EH and endometrial cancer, with relative risks of 2–20 (Kreiger et al., 1986; Daly et al., 1995; Schottenfeld, 1995; Grady and Ernster, 1996; Weiderpass et al., 1999). Clinical trials of hormone replacement therapy have shown that unopposed estrogen use leads to a 5- to 15-fold higher incidence of EH and EH with atypia than when an estrogen/ progesterone combination is used (Woodruff and Pickar, 1994; The PEPI trial, 1996; Archer et al., 1999; Bjarnason et al., 1999; Bergeron and Fox, 2000; Kurman et al., 2000). Lesion risk in those treated with estrogen/progesterone combinations is no greater than that in those receiving placebo (The PEPI trial, 1996). Moreover, some studies show regression of EH with atypia following progesterone therapy (Ferenczy and Gelfand, 1989; The PEPI trial, 1996; Grimbizis et al., 1999). Long-term oral contraceptive use exerts a protective effect on the risk of EH (Kreiger et al., 1986). On the other hand, tamoxifen treatment leads to a three- to sevenfold increased risk of endometrial lesions (Fisher et al., 1994; Grady and Ernster, 1996; Mourits et al., 2001).

Progression to Cancer A histopathologic diagnosis of EH with atypia (or EIN under the new nomenclature) in an endometrial biopsy or curettage specimen is frequently coincident with an invasive neoplasm upon examination of the hysterectomy specimen (Franco and Ferenczy, 2002a). The percentage of occult endometrial carcinomas in women with EH with atypia ranges from 28% to 45%, whereas that in women with EH without atypia is considerably lower, ranging from 0% to 3% (Baak et al., 1992; Hunter et al., 1994; Janicek and Rosenshein, 1994; Ho et al., 1997). The risk of a subsequent endometrial carcinoma in women with EH on an initial biopsy has been analyzed in a few cohort studies with follow-up of as much as 20 years. Among those with EH without cytologic atypia the cumulative risk of cancer was 0%–8%, whereas progression to cancer among those with EH with atypia was much higher, at 25%–58% (Ferenczy and Gelfand, 1989; Lindahl and Willen, 1991; Ferenczy and Mutter, 2000; Kurman et al., 2000). Much of the variation across studies stems from the histologic heterogeneity of endometrial cancer precursors, which has made it difficult for pathologists to devise uniform classification protocols. Defined histologic criteria for classifying lesions have become available only recently (Mutter, 2000; Sherman, 2000; Dietel, 2001).

CANCER CONTROL THROUGH PREVENTION AND DETECTION OF CANCER PRECURSORS The long time interval required for cancer development, typically decades (Kelloff et al., 2000), provides abundant opportunity to screen for and thereby detect cancer early in its natural history. Typically, screening is undertaken for those anatomic sites with precursor lesions that are recognizable morphologically. The ultimate goal of recognizing these early manifestations of neoplasia is to intervene surgically or through chemoprevention to impede and hopefully reverse the process of carcinogenesis.

Chemoprevention Chemoprevention has been defined as the use of drugs or other agents to inhibit, delay, or reverse the progressive genetic damage and the associated tissue damage that accrue during carcinogenesis (Sporn, 1976; Kelloff and Sigman, 2002; O’Shaughnessy et al., 2002). In other words, cancer chemoprevention involves the treatment of carcinogenesis (Kelloff et al., 2000). Hong (2003) classified chemoprevention according to the setting in which it is used: primary, involving prevention of initial cancers in healthy individuals; secondary, involving prevention of cancer in individuals with premalignant lesions; and tertiary, involving the prevention of second primaries in individuals cured

of their initial cancer. Secondary chemoprevention might be viewed as an alternative to the standard approach to treating cancer precursors (i.e., surgery). Although surgical removal of precursors may reduce the risk of subsequent cancer, it is not necessarily the optimal approach because it may cause substantial morbidity (e.g., as with partial or total esophagectomy for dysplastic Barrett’s esophagus); and unlike chemoprevention, it does not treat the entire epithelial field at risk. The latter point is important and stems from the concept of “field cancerization,” which was proposed first by Slaughter et al. (1953) and is interpreted nowadays to mean that normal-appearing tissue may harbor genetic changes that render it susceptible to the development of cancer. Before a chemopreventive agent can be approved for use, its chemopreventive efficacy and relative safety from side effects must be demonstrated in clinical trials. However, the duration, size, and associated cost of a trial in which cancer itself is the proposed end point generally is prohibitive, thereby rendering cancer unsuitable as the outcome of interest for trials of chemopreventive agents (O’Shaughnessy et al., 2002)—hence the need for surrogate or intermediate end points for incident cancer to ensure that the evaluative process can be completed in a timely and cost-effective fashion. In the context of a clinical trial designed to assess the effect of an agent against incident cancer, a surrogate or intermediate end point is a biomarker or response variable that yields a valid test of the null hypothesis of no association between the intervention and cancer (Prentice, 1989; Schatzkin, 2002). This implies that the effect of the intervention on the surrogate is in accord with its effect on the incidence of cancer. Validation of a biomarker as an intermediate end point for cancer is important but difficult in practice. Ideally, it requires demonstration that the surrogate end point mediates the exposure–cancer relationship (Schatzkin et al., 1990; Schatzkin, 2002). There are many potential biomarkers of carcinogenesis. Such markers may be cellular (e.g., nuclear morphology, mitotic index, DNA ploidy), genotypic (e.g., loss of heterozygosity, gene amplification), molecular (e.g., cellular antigens such as proliferating cell nuclear antigen, growth factors such as insulin-like growth factor I, markers of apoptosis such as expression of bcl-2), tissue-related (e.g., expression of estrogen receptors in the breast), drug-related (e.g., prostaglandin biosynthesis), or phenotypic (e.g., intraepithelial neoplastic lesions such as colorectal adenomas and cervical intraepithelial neoplasia) (Kelloff et al., 2000). Among these biomarkers, the latter are perhaps the most appealing as surrogate end points for cancer because they are cancer precursors and therefore are on the causal pathway to cancer (O’Shaughnessy et al., 2002). However, their use for this purpose requires the development of standardized quantitative techniques for sampling and grading such lesions, a criterion that has been easier to fulfill for tissues that can be visualized directly (e.g., oral cavity, colon, cervix) than for those that cannot (e.g., breast, prostate). Furthermore, given that precursor lesions do not necessarily progress to cancer (e.g., low grade squamous intraepithelial lesions of the cervix are more common than high grade lesions, which in turn are more common than cervical cancer), establishing the effectiveness of agents in chemoprevention trials might require not only the demonstration of prevention or regression of lesions at the histologic level but also regression at the genetic or molecular level, the latter serving to demonstrate that the lesions that were prevented or reduced in frequency had the potential to progress to cancer. Table 3–3 lists potential phenotypic surrogate end points for cancer at a number of anatomic sites and associated molecular and cellular biomarkers. Many of the cancer precursors listed in Table 3–3 have been used as surrogate end points in randomized trials of chemopreventive agents. For example, leukoplakia is an oral premalignant lesion (with varying grades of dysplasia) that has been shown to be treated more effectively by low dose isotretinoin (13-cis-retionoic acid) than by b-carotene following induction therapy with high dose isotretinoin (Lippman et al., 1993). This trial was designed to address problems that arose in an earlier trial, which demonstrated that although high dose 13-cisretionoic acid induced a high response rate in patients with leukoplakia (compared to placebo), it was accompanied by considerable toxicity and a high relapse rate (Hong et al., 1986). Although chemoprevention of head and neck cancers has focused largely on the use of

37

Cancer Precursors Table 3–3. Potential Phenotypic Surrogate End Points for Cancer and Associated Molecular and Cellular Biomarkers, by Anatomic Site Target Organ Colon/rectum

Cohorts for Clinical Studies

Primary Endpoints

Previous adenomas

Prevention of adenomas

FAP

Regression and prevention of adenomas Prevention of adenomas

Adolescent FAP (prephenotype expression)

Early Associated Biomarkers ACF, proliferation antigens (PCNA, Ki-67), apoptosis, differentiation antigens (Lewisx, Lewisy, T, Tn, and sialyl-Tn antigens and apomucins), K-ras, DNA methylation

Head and neck

OPLs (atypical hyperplasia, hyperkeratosis, mild/severe dysplasia)

Regression of existing OPL and/or prevention of new OPL

LOH (3p, 9p), p53, cyclin D1, growth factors (e.g., EGFR), retinoid receptors (RARb), DNA content, proliferation antigens (PCNA, Ki-67), apoptosis

Esophagus

Barrett’s esophagus (mild to severe dysplasia)

Regression and/or slowed progression of Barrett’s dysplasia

DNA contents, prolilferation antigens (PCNA, Ki-67), ODC, growth factors (e.g., EGFR), cyclin D1, p16, p53, Cdx1 and Cdx2

Cervix

HGSIL (CIN 2/3)

Regression and prevention of CIN

Persistent HPV infection, viral integration, LOH (3p, 4p, 4q, 11q), proliferation antigens (PCNA), growth factors (e.g., EGFR), K-ras, differentiation antigens

Lung

Bronchial dysplasia (current or former smokers)

Regression and prevention of bronchial dysplasia

Proliferation antigens (Ki-67), LOH (3p, 8p, 9q), K-ras retinoid receptors, p53, FHIT, DNA methylation, apoptosis

Skin (nonmelanoma)

High risk AK (>10 AK within previous year)

Prevention of AK (and, as a secondary end point, of skin SCC and, as a tertiary end point regression of AK)

Breast

Hyperplasia without atypia

Prevention of atypical hyperplasia Regression of atypical hyperplasia

DNA methylation, LOH, growth factors (e.g., EGFR, erbB-2, IGF), proliferation antigens (Ki-67, PCNA), p53, apoptosis

Regression of HGPIN

DNA methylation, GST, pc-1 chromosomal loss or gain (8p, 9p, 16q), apoptosis, proliferation antigens (Ki-67), growth factors (IGF, TGFa TGFb)

Atypical hyperplasia Prostate

HGPIN without cancer

Source: Adapted with permission from O’Shaughessy et al. (2002). ACF, aberrant crypt foci; AK, actinic keratosis (or keratoses); CIN, cervical intraepithelial neoplasia; EGFR, epidermal growth factor receptor; FAP, familial adenomatous polyposis coli; FHIT, fragile histidine triad gene; GST, glutathione S-transferase; HGPIN, high grade prostatic intraepithelial neoplasia; HGSIL, high-grade squamous intraepithelial lesion; HPV, human papillomavirus; IEN, intraepithelial neoplasia; IGF, insulin-like growth factor; LOH, loss of heterozygosity; ODC, ornithine decarboxylase; OPL, oral premalignant lesion; PCNA, proliferating cell nuclear antigen; RAR, retinoic acid receptor; SCC, squamous cell carcinoma; TGF, transforming growth factor; TGFa, TGFb, transforming growth factors a and b.

retinoids, the chemopreventive properties of other agents, including protease inhibitors and NSAIDs, are also being investigated (Glover and Papadimitrakopolou, 2003). With respect to the colorectum, adenomatous polyps are precursors of colorectal cancer whose prevention is likely to result in the prevention of colorectal cancer as well (Jänne and Mayer, 2000). Recently, administration of the NSAID aspirin was shown to be associated with a moderate reduction in the risk of recurrence of colorectal adenomas (Baron et al., 2003). Earlier trials showed that the NSAIDs sulindac and celecoxib cause regression of colorectal adenomas in patients with familial adenomatous polyposis (Labayle et al., 1991; Giardiello et al., 1993; Nugent et al., 1993; Steinbach et al., 2000). Cancer chemoprevention as a discipline has a relatively short history. However, the foundations of the discipline have largely been laid, and we can expect to see a burgeoning of activity in this area over the next few years. Chemopreventive trials involving cancer precursors, supplemented by measurement of appropriate biomarkers, will play a central role in the field.

invasive cancers. Table 3–4 shows the screening tests that have been applied for targeting cancer precursors in clinical and research settings. Few of these tests have exhibited unequivocal effectiveness in terms of reducing mortality (Franco et al., 2002), although clinical benefit may exist in certain circumstances, such as in the surveillance of high risk individuals. By analogy to the chemoprevention of precursor lesions, the efficacy of an intervention based on a screening test is a function of how proximal or distal the target precursor lesion is with respect to its corresponding invasive cancer. Ultimately, screening for and treatment of a cancer precursor should avert the occurrence of the corresponding invasive lesion and, in consequence, of any clinical end points that derive from the latter, such as death. Evidence of benefit can be documented via end points such as a reduction in the incidence of invasive cancers, an increase in survival, down-staging, or increased detection of cancer precursors.

CONCLUSIONS Screening Much of the research on specific screening interventions has focused on precancerous or early cancerous lesions, such as high grade dysplasias of the uterine cervix, oral leukoplakias, and colonic adenomas. In fact, the most widely studied cancer screening intervention, the Pap test, is based on cytologic identification of precancerous lesions that can be treated or excised, with consequent arrest of neoplastic development in the cervix. For many cancer sites, however, screening interventions have been evaluated with respect to detection of early

During the past few years, much has been learned about cancer precursors. Nevertheless, there are many unresolved issues with respect to the definition of cancer precursors at some anatomic sites and their etiology, detection, and prevention. Although predicting future developments can be a somewhat hazardous undertaking, certain trends and opportunities are discernible with respect to the study of cancer precursors. Advances in the application of molecular methods and new cell sampling techniques for the analysis of human tissue are allowing

38

PART I: BASIC CONCEPTS

Table 3–4. Screening Tests that Can Detect Cancer Precursors or Biomarkers of Early Carcinogenesis Cancer Site

Screening Test

Colorectum

Sigmoidoscopy and colonoscopy to detect and remove colonic adenomas

Adequate evidence of cancer incidence reduction from observational studies

Accepted clinical practice as opportunistic screening modality

Endometrium

Endometrial sampling or transvaginal

Insufficient evidence

Pap cytology

No benefit

Use for diagnostic workup of high risk patients Ancillary technique for diagnostic workup

Lung

Spiral CT to detect small lesions

Promising technique but insufficient evidence of benefit

Research use only

Oral cavity

Oral examination or cytology

Insufficient evidence of mortality reduction as population screening strategy

Part of clinical examination of high risk patients

Skin

Physical examination for actinic keratoses and dysplastic nevi

Insufficient evidence

Clinical surveillance of high risk patients

Stomach

Double-contrast radiography followed by endoscopy

May have contributed to decline of mortality in Japan, although no evidence from randomized trials is available Promising technique but insufficient evidence of benefit

Widely used in population screening in Japan

Serum pepsinogen levels as markers of gastric atrophy Uterine cervix

Conventional, liquid-based, or automated Pap cytology Visual inspection with acetic acid and its variations with lugol and magnification HPV testing

Evidence of Benefit

Proven benefit for conventional cytology; technical improvements using automation and liquid-based techniques largely believed to be effective Promising technique, preliminary evidence that it may be more sensitive but somewhat less specific than the Pap test Proven as a secondary triage test for equivocal Pap smears; promising technique for population screening, where it is more sensitive but somewhat less specific than the Pap test

Current Use

Research use only Accepted clinical practice for opportunistic or organized screening Research use only, particularly in developing countries Standard clinical practice in triage; research use only for screening, particularly in Europe and North America

CT, computed tomography; Pap, Papanicolaou; US, ultrasonography.

identification of what are potentially the earliest molecular changes leading to carcinogenesis. For example, recent observations at some anatomic sites (e.g., upper respiratory tract, breast) suggest that molecular changes may precede histologic changes (Westra and Sidransky, 1998; Kandel et al., 2000). Currently, however, it is not clear whether tissue exhibiting these changes should be considered premalignant. Resolution of this question requires follow-up studies in which cancer incidence, or perhaps the incidence of cancer precursors, is compared in individuals with and without molecular changes in normal tissue and possibly studies in which the prevalence of molecular changes in histologically normal tissue is compared with that in adjacent tissue showing cancer precursors and invasive cancer. Molecular studies of this kind might lead us to revisit our current approaches to the classification of cancer or cancer-predisposing lesions. Indeed, given the increasingly molecular nature of a pathologic classification, it is conceivable that in time we will move away from terms such as “cancer precursor” and “invasive cancer” and move toward a system wherein lesions are classified on the basis of the genetic changes that they exhibit as “early” or “late” cancer, without further specification. Nevertheless, existing morphology-based definitions will continue to be used for the foreseeable future, given that to a large extent they form the cornerstone of the current approaches to the treatment of cancer. It is obvious that pathology has played a major role in the development of cancer epidemiology. Indeed, the current classification of cancers and their precursors, and hence case definition in epidemiologic studies, is based largely on histomorphologic criteria (Saxen, 1979; Eustis, 1989). Recently, however, the first tentative steps have been taken toward the creation of molecular classifications of cancers based on gene expression patterns (Perou et al., 2000). It is conceivable that similar classifications of cancer precursors might emerge in time, spurred on perhaps by recent technologic developments such as

laser capture microdissection, which can provide access to homogeneous samples of relevant tissue (Fend and Raffeld, 2000). Such molecular approaches to the classification of early lesions (or cancer precursors) will result in changes in our understanding of the etiology of such lesions (and their associated cancers). At the very least, the emerging era of molecular classification will spawn a new round of etiologic studies using these classification systems for case definition. These studies might change our understanding of the etiology of some conditions and might provide us with new opportunities to gain insight into the etiology of others. Molecular analysis of cancer precursor lesions might assist with the identification of lesions that are at particular risk of progression to invasive cancer. This will have implications for the clinical management of individuals harboring such lesions, for whom increased surveillance or early intervention might be warranted. Studies designed to identify markers of risk will require well characterized cohorts, and we anticipate that studies of this kind, which have already begun to be reported (Rohan et al., 1998; Gobbi et al., 1999), will increase. Existing archives of tumor tissue, used in conjunction with associated clinical and epidemiologic information, might represent a convenient starting point for such studies. Advances in approaches to the molecular characterization of tumors and their precursor lesions, with the resultant possibility of identifying early lesions on molecular grounds, leads inevitably to the possibility of molecular screening for such lesions, an approach to tumor detection that is already under investigation for invasive cancers (Ahrendt and Sidransky, 1999). If implemented, these developments will have clinical and ethical implications because practitioners will be confronted with an expanded array of lesions, of which the malignant potential and appropriate clinical management of at least some may not be clear. For example, with respect to clinical management,

Cancer Precursors adoption of molecular approaches to lesion detection will raise questions as to the appropriate margins for excising lesions given that similar molecular changes might be observed in adjacent, morphologically normal tissue (Westra and Sidransky, 1998; Kandel et al., 2000). Nevertheless, before molecular assays are incorporated into clinical practice, it will be necessary to validate them, demonstrate that they are superior to current approaches to tumor detection, and develop and refine the requisite high throughput automation of the assays (Ahrendt and Sidransky, 1999). Prevention of true cancer precursors necessarily results in a reduction in the subsequent incidence of the associated cancer at the same anatomic site. As such, cancer precursors serve dual roles: as potential targets for preventive agents and as intermediate end points for trials testing agents and interventions designed to prevent cancer occurrence. Therefore, approaches to the primary prevention of cancer precursors hold much promise for the prevention of cancer. Although relatively little is known about the prevention of cancer precursors at present, the current movement towards using intermediate end points in trials of preventive agents will result in a burgeoning of data on their prevention, especially with respect to the use of chemopreventive agents (Lippman et al., 1998). Mathematical models of the development of precursor lesions may assist in the development of optimal prevention (and screening) strategies by predicting the effect of an intervention given its proposed mechanism of action (Pinsky, 2000). Models for adenoma development (Pinsky, 2000) and the natural history of HPV infection (Myers et al., 2000) have been developed recently, and similar efforts can be anticipated with respect to precursor lesions at other sites, given the increasing interest in prevention. From an epidemiologic perspective, it is clear that researchers will have to contend with increasingly complex etiologic models when designing observational studies and intervention trials. The use of “softer” end points situated upstream from the earliest morphologic changes observable in the natural history of cancer will pose new challenges to molecular epidemiologists in terms of the safeguards needed to avoid bias resulting from errors in the measurement of molecular markers. Although the importance of using accurate, reproducible assays is well recognized, the literature on cancer precursors is not without examples of incoherent results due to misclassification of intermediate end points (Franco, 1991; Makni et al., 2000). Increasing reliance on prospective studies with repeated measurements of intermediate end points will help to assuage some of the concerns about misclassification bias. Statistical methods for the analysis of longitudinal data generated in such investigations are a relatively recent addition to the armamentarium of cancer epidemiologists, and they are yet to be fully appreciated as research tools. Use of these methods and judicious use of mediation analysis (Freedman et al., 1992; Buyse and Molenberghs, 1998) in studies of intermediate end points will play a valuable role in helping epidemiologists to decipher the sequence of events and associated molecular changes leading to cancer development. From the foregoing discussion it is evident that many disciplines, both individually and collectively, can contribute to the goal of understanding cancer precursors. Indeed, despite the reservations of occasional commentators (McMichael, 1994; zur Hausen, 2001), the blossoming field of molecular epidemiology, entailing incorporation of biologic measurements into epidemiologic research, attests to the power of a multidisciplinary approach to studying the etiology and pathogenesis of cancer. However, other, perhaps more unusual combinations of approaches might be mutually beneficial. For example, transgenic mice (i.e., mice with foreign DNA incorporated into their genome (Guha et al, 2001) ) with specific genetic lesions can develop histologic lesions similar to those that precede cancer development in humans (Wang et al., 1994) and therefore can provide insight into the molecular basis of carcinogenesis (Guha et al., 2001). Clues as to the appropriate transgenic mouse models that molecular biologists develop for this purpose might come, in part, from epidemiologic studies of the molecular pathogenesis of cancer. This is but one of many possible examples, and we end this chapter with a plea for novel cross-disciplinary approaches to investigations designed to enhance our knowledge of cancer precursors.

39

References Agoff SN, Crispin DA, Bronner MP, et al. 2001. Neoplasms of the ampulla of Vater with concurrent pancreatic intraepithelial neoplasia: a histological and molecular study. Mod Pathol 14:139–146. Ahrendt SA, Sidransky D. 1999. The potential of molecular screening. Surg Oncol Clin N Am 1999;8:641–656. Albores-Saavedra J, Krueger JE. 2001. C-cell hyperplasia and medullary microcarcinoma. Endocr Pathol 12:365–377. Albores-Saavedra J, Henson DE, Klimstra D. 2000. Tumors of the gallbladder, extrahepatic bile ducts and ampulla of Vater. In: Atlas of Tumor Pathology, Third Series, Fascicle 27. Washington, DC: Armed Forces Institute of Pathology. Almendingen K, Hofstad B, Trygg K, et al. 2000. Smoking and colorectal adenomas: a case-control study. Eur J Cancer Prev 9:193–203. Almendingen K, Hofstad B, Trygg K, et al. 2001. Current diet and colorectal adenomas: a case-control study including different sets of traditionally chosen control groups. Eur J Cancer Prev 10:395–406. Almendingen K, Hofstad B, Vatn MH. 2002. Lifestyle-related factors and colorectal polyps: preliminary results from a Norwegian follow-up and intervention study. Eur J Cancer Prev 11:153–158. Alpers CE, Wellings SR. 1985.The prevalence of carcinoma in situ in normal and cancer-associated breasts. Hum Pathol 16:796–807. Anastasiadis PG, Skaphida PG, Koutlaki NG, Galazios GC, Tsikouras PN, Liberis VA. 2000. Descriptive epidemiology of endometrial hyperplasia in patients with abnormal uterine bleeding. Eur J Gynaecol Oncol 21:131–134. Archer DF, Furst K, Tipping D, Dain MP, Vandepol C. 1999. A randomized comparison of continuous combined transdermal delivery of estradiolnorethindrone acetate and estradiol alone for menopause; CombiPatch Study Group. Obstet Gynecol 94:498–503. Atkin WS, Morson BC, Cuzick J. 1992. Long-term risk of colorectal cancer after excision of rectosigmoid adenomas. N Engl J Med 326:658–662. Avidan B, Sonnenberg A, Schnell TG, Sontag SJ. 2001. Gastric surgery is not a risk factor for Barrett’s esophagus or esophageal adenocarcinoma. Gastroenterology 121:1281–1285. Avidan B, Sonnenberg A, Schnell TG, Sontag SJ. 2002. Hiatal hernia and acid reflux frequency predict presence and length of Barrett’s esophagus. Dig Dis Sci 47:256–264. Axell T, Pindborg JJ, Smith CJ, et al. 1996. Oral white lesions with particular reference to precancerous and tobacco-related lesions: conclusions of an international symposium held in Uppsala, Sweden, May 18–21 1994. J Oral Pathol Med 25:49–54. Baak JP, Wisse-Brekelmans EC, Fleege JC, van der Putten HW, Bezemer PD. 1992. Assessment of the risk on endometrial cancer in hyperplasia, by means of morphological and morphometrical features. Pathol Res Pract 188:856–859. Baanders-van Halewyn EA, Blankenstein MA, Thijssen JH, de Ridder CM, de Waard F. 1996. A comparative study of risk factors for hyperplasia and cancer of the endometrium. Eur J Cancer Prev 5:105–112. Baghurst PA, Rohan TE. 1995. Dietary fibre and risk of benign proliferative epithelial disorders of the breast. Int J Cancer 63:481–485. Banoczy J. 1977. Follow-up studies in oral leukoplakia. J Maxillofacial Surg 5:69–75. Baron JA. 2002. Colon and rectum. In: Franco EL, Rohan TE, editors. Cancer Precursors: Epidemiology, Detection, and Prevention. New York: Springer-Verlag, pp. 127–150. Baron JA, Cole BF, Sandler RS, et al. 2003, A randomized trial of aspirin to prevent colorectal adenomas. N Engl J Med 348:891–899. Barrera AH, Meyers AD, Shroyer KR, et al. 2001. Chromosomal aneuploidy precedes morphological changes and supports multifocality in head and neck lesions. Laryngoscope 111:1853–1858. Beaty MW, Jaffe E. Lymphoid system. In: Henson DE, Albores-Saavedra J, editors. Pathology of Incipient Neoplasia, 3rd ed. New York: Oxford University Press, pp. 453–455. Benito E, Cabeza E, Moreno V, et al. 1993. Diet and colorectal adenomas: a case-control study in Majorca. Int J Cancer 55:213–219. Ben-Yehuda OM, Kim YB, Leuchter RS. 1998. Does hysteroscopy improve upon the sensitivity of dilatation and curettage in the diagnosis of endometrial hyperplasia or carcinoma? Gynecol Oncol 68:4–7. Bergeron C, Fox H. 2000. Low incidence of endometrial hyperplasia with acceptable bleeding patterns in women taking sequential hormone replacement therapy with dydrogesterone. Gynecol Endocrinol 14:275–281. Bergstrom R, Sparen P, Adami HO. 1999. Trends in cancer of the cervix uteri in Sweden following cytological screening. Br J Cancer 81:159–166. Berkowitz GS, Kelsey JL, Li Volsi VA, et al. 1984. Exogenous hormone use and fibrocystic disease by histopathologic component. Int J Cancer 34:443–449.

40

PART I: BASIC CONCEPTS

Berkowitz GS, Kelsey JL, LiVolsi VA, et al. 1985. Risk factors for fibrocystic disease and its histopathologic components. J Natl Cancer Inst 75:43–50. Bersentes K, Fass R, Padda S, et al. 1998. Prevalence of Barrett’s esophagus in Hispanics is similar to Caucasians. Dig Dis Sci 43:1038–1041. Bhathal PS, Brown RW, Lesueur GC, Russell IS. 1985. Frequency of benign and malignant breast lesions in 207 consecutive autopsies in Australian women. Br J Cancer 51:271–278. Bird CL, Frankl HD, Lee ER, Haile RW. 1998. Obesity, weight gain, large weight changes, and adenomatous polyps of the left colon and rectum. Am J Epidemiol 147:670–680. Bird CL, Swendseid ME, Witte JM, et al. 1995. Red cell and plasma folate, folate consumption, and the risk of colorectal adenomatous polyps. Cancer Epidemiol Biomarkers Prev 4:709–714. Bird RP. 1987. Observation and quantification of aberrant crypts in the murine colon treated with a colon carcinogen: preliminary findings. Cancer Lett 37:147–151. Bjarnason K, Cerin A, Lindgren R, Weber T. 1999. Adverse endometrial effects during long cycle hormone replacement therapy: Scandinavian Long Cycle Study Group. Maturitas 32:161–170. Bjorge T, Gunbjorud AB, Langmark F, Skare GB, Thoresen SO. 1994. Cervical mass screening in Norway—510,000 smears a year. Cancer Detect Prev 18:463–470. Blatt LJ. 1961. Polyps of the colon and rectum: incidence and distribution. Dis Colon Rectum 4:277–282. Blot WJ, McLaughlin JK. 1999. The changing epidemiology of esophageal cancer. Semin Oncol 26 (suppl 15):2–8 Bodian CA. 1993. Benign breast diseases, carcinoma in situ, and breast cancer risk. Epidemiol Rev 15:177–187. Boone CW, Henson DE, Kelloff GJ. 1999. Pathobiology of incipient neoplasia. In: Srivastava S, Henson DE, Gazdar A, editors. Molecular Pathology of Early Cancer. Amsterdam: IOS Press, pp. 3–11. Bosch FX, de Sanjose S. 2003. Human papillomavirus and cervical cancer— burden and assessment of causality. J Natl Cancer Inst Monogr 31: 3–13. Bouquot JE, Gorlin RJ. 1986. Leukoplakia, lichen planus, and other oral keratoses in 23,616 white Americans over the age of 35 years. Oral Surg Oral Med Oral Pathol 61:373–381. Boutron-Ruault M-C, Senesse P, Faivre J, et al. 1996. Folate and alcohol intakes: related or independent roles in the adenoma-carcinoma sequence. Nutr Cancer 26:337–346. Boyle JO, Lonardo F, Chang JH, et al. 2001. Multiple high-grade bronchial dysplasia and squamous cell carcinoma: concordant and discordant mutations. Clin Cancer Res 7:259–266. Brat DJ, Lillemoe KD, Yeo CJ, et al. 1998. Progression of intraductal neoplasia to infiltrating adenocarcinoma of the pancreas. Am J Surg Pathol 22:163–169. Bremner CG, Ackerman LV. 1970. Polyps and carcinoma of the large bowel in the South African Bantu. Cancer 26:991–999. Breuer-Katschinski B, Nemes K, Marr A, et al. 2000. Alcohol and cigarette smoking and the risk of colorectal adenomas. Dig Dis Sci 45:487– 493. Breuer-Katschinski B, Nemes K, Marr A, et al. 2001. Colorectal adenomas and diet: a case-control study. Dig Dis Sci 46:86–95. Bright RA, Morrison AS, Brisson J, et al. 1989. Histologic and mammographic specificity of risk factors for benign breast disease. Cancer 64:653– 657. Brinton LA, Barrett RJ, Berman ML, Mortel R, Twiggs LB, Willbanks GD. 1993. Cigarette smoking and the risk of endometrial cancer. Am J Epidemiol 137:281–291. Brinton LA, Vessey MP, Flavel R, Yeates D. 1981. Risk factors for benign breast disease. Am J Epidemiol 113:203–214. Brockie E, Anand A, Albores-Saavedra J. 1998. Progression of atypical ductal hyperplasia/carcinoma in situ of the pancreas to invasive adenocarcinoma. Ann Diagn Pathol 2:286–292. Buttar NS, Wang KK, Sebo TJ, et al. 2001. Extent of high-grade dysplasia in Barrett’s esophagus correlates with risk of adenocarcinoma. Gastroenterology 120:1630–1639. Buyse M, Molenberghs G. 1998. Criteria for the validation of surrogate endpoints in randomized experiments. Biometrics 54:1014–1029. Cameron AJ. 1999. Barrett’s esophagus: prevalence and size of hiatal hernia. Am J Gastroenterol 94:2054–2059. Cameron AJ, Arora AS. 2002. Barrett’s esophagus and reflux esophagitis: is there a missing link? Am J Gastroenterol 97:273–278. Cameron AJ, Lomboy CT. 1992. Barrett’s esophagus: age, prevalence, and extent of columnar epithelium. Gastroenterology 103:1241–1245. Cameron AJ, Zinsmeister AR, Ballard DJ, Carney JA. 1990. Prevalence of columnar-lined (Barrett’s) esophagus. Comparison of population-based clinical and autopsy findings. Gastroenterology 99:918–922.

Campos GM, DeMeester SP, Peters JH, et al. 2001. Predictive factors of Barrett’s esophagus: multivariate analysis of 502 patients with gastroesophageal reflux disease. Arch Surg 136:1267–1273. Carter RL, editor. 1984. Precancerous states. London: Oxford University Press. Carter CL, Corle DK, Micozzi MS, et al. 1988. A prospective study of the development of breast cancer in 16,692 women with benign breast disease. Am J Epidemiol 128:467–477. Castellsague X, Munoz N. 2003. Cofactors in human papillomavirus carcinogenesis: role of parity, oral contraceptives, and tobacco smoking. J Natl Cancer Inst Monogr 31:20–28. Castellsague X, Bosch FX, Munoz N. 2002. Environmental co-factors in HPV carcinogenesis. Virus Res 89:191–199. Castle PE, Giuliano AR. 2003. Genital tract infections, cervical inflammation, and antioxidant nutrients—assessing their roles as human papillomavirus cofactors. J Natl Cancer Inst Monogr 31:29–34. CDC. 1994. Results from the National Breast and Cervical Cancer Early Detection Program, October 31, 1991–September 30, 1993. MMWR Morb Mortal Wkly Rep 43:530–534. Chen M-J, Longnecker MP, Morgenstern H, et al. 1998. Recent use of hormone replacement therapy and the prevalence of colorectal adenomas. Cancer Epidemiol Biomark Prev 7:227–230. Cheng L, Gu J, Ulbright TM, et al. 2002. Precise microdissection of human bladder carcinomas reveals divergent tumor subclones in the same tumor. Cancer 94:104–110. Chow WH, Blaser MJ, Blot WJ, et al. 1998. An inverse relation between cagA+ strains of Helicobacter pylori infection and risk of esophageal and gastric cardia adenocarcinoma. Cancer Res 58:588–590. Chow WH, Greenberg RS, Liff JM. 1986. Decline in the incidence of carcinoma in situ of the cervix. Am J Public Health 76:1322–1324. Chumas JC, Nelson B, Mann WJ, Chalas E, Kaplan CG. 1985. Microglandular hyperplasia of the uterine cervix. Obstet Gynecol 66:406–409. Clark JC, Collan Y, Eide TJ, et al. 1985. Prevalence of polyps in an autopsy series from areas with varying incidence of large-bowel cancer. Int J Cancer 36:179–186. Cole P, Elwood JM, Kaplan SD. 1978. Incidence rates and risk factors of benign breast neoplasms. Am J Epidemiol 108:112–120. Compton C. 2001. Incipient neoplasia of the large intestine. In: Henson DE, Albores-Saavedra J, editors. The Pathology of Incipient Neoplasia. New York: Oxford University Press, pp. 174–206. Conio M, Cameron AJ, Romero Y, et al. 2001. Secular trends in the epidemiology and outcome of Barrett’s esophagus in Olmsted County, Minnesota. Gut 48:304–309. Conio M, Filiberti R, Blanchi S, et al. 2002. Risk factors for Barrett’s esophagus: a case-control study. Int J Cancer 97:225–229. Cook MG, Rohan TE. 1985. The patho-epidemiology of benign proliferative epithelial disorders of the female breast. J Pathol 146:1–15. Correa P. 1978. Epidemiology of polyps and cancer. In: Morson BC, editor. The Pathogenesis of Colorectal Cancer. Philadelphia: Saunders, pp. 126–152. Crissman JD, Zarbo RJ. 1989. Dysplasia, in situ carcinoma, and progression to invasive squamous cell carcinoma of the upper aerodigestive tract. Am J Surg Pathol 13:5S–16S. Daly MB, Bookman MA, Lerman CE. 1995. Female reproductive tract: cervix, endometrium, ovary. In: Greenwald P, Kramer BS, Weed DL, editors. Cancer Prevention and Control. New York: Marcel Dekker, pp. 585–610. Davies BR, Platt-Higgins AM, Schmidt G, Rudland PS. 1999. Development of hyperplasias, preneoplasias, and mammary tumors in MMTV-c-erbB-2 and MMTV-TFFa transgenic rats. Am J Pathol 155:303–314. Day DW, Morson BC. 1978. The adenoma-carcinoma sequence. In: Bennington JL, editor. The Pathogenesis of Colorectal Cancer. Philadelphia: Saunders, pp. 58–71. de Mas CR, Kramer M, Seifert G, et al. 1999. Short Barrett: prevalence and risk factors. Scand J Gastroenterol 34:1065–1070. DeVault KR, Castell DO. 1999. Updated guidelines for the diagnosis and treatment of gastroesophageal reflux disease. The Practice Parameters Committee of the American College of Gastroenterology. Am J Gastroenterol 94:1434–1442 Dietel M. 2001. The histological diagnosis of endometrial hyperplasia. Is there a need to simplify? Virchows Arch 439:604–608. Dietl J, Semm K, Hedderich J, Buchholz F. 1983. CIN and preclinical cervical carcinoma. A study of morbidity trends over a 10-year period. Int J Gynaecol Obstet 21:283–289. Dubin N, Pasternack BS. 1984. Breast cancer screening data in case-control studies. Am J Epidemiol 120:8–16. Dupont WD, Parl FF, Hartmann WH, et al. 1993. Breast cancer risk associated with proliferative breast disease and atypical hyperplasia. Cancer 71:128–165.

Cancer Precursors Eisen GM, Sandler RS, Murray S, Gottfried M. 1997. The relationship between gastroesophageal reflux disease and its complications with Barrett’s esophagus. Am J Gastroenterol 92:27–31. Eloubeidi M, Provenzale D. 2001. Clinical and demographic predictors of Barrett’s esophagus among patients with gastroesophageal reflux disease: a multivariable analysis in veterans. J Clin Gastroenterol 33:306–309. Enger SM, Longnecker MP, Lee ER, et al. 1997. Recent and past physical activity and prevalence of colorectal adenomas. Br J Cancer 75:740– 745. Enterline HT. 1976. Polyps and cancer of the large bowel. In: Morson BC, editor. Pathology of the gastro-intestinal tract. Berlin: Springer-Verlag, pp. 95–141. Epstein JB, Zhang L, Rosin M. 2002. Advances in the diagnosis of oral premalignant and malignant lesions. J Can Dent Assoc 68:617–621. Erhardt JG, Kreichgauer HP, Meisner C, et al. 2002. Alcohol, cigarette smoking, dietary factors and the risk of colorectal adenomas and hyperplastic polyps: a case control study. Eur J Nutr 41:35–43. Eustis SL. 1989. The sequential development of cancer: a morphological perspective. Toxicol Lett 49:267–81. Farquhar CM, Lethaby A, Sowter M, Verry J, Baranyai J. 1999. An evaluation of risk factors for endometrial hyperplasia in premenopausal women with abnormal menstrual bleeding. Am J Obstet Gynecol 181:525–529. Fass R, Hell RW, Garewal HS, et al. 2001. Correlation of oesophageal acid exposure with Barrett’s esophagus length. Gut 48:310–313. Fearon ER, Vogelstein B. 1990. A genetic model for colorectal tumorigenesis. Cell 61:759–767. Fechner RE, Mills SE. 1993. Tumors of bones and joints. In: Atlas of Tumor Pathology, Third Series, Fascicle 8. Washington, DC: Armed Forces Institute of Pathology. Feldman S, Cook EF, Harlow BL, Berkowitz RS. 1995. Predicting endometrial cancer among older women who present with abnormal vaginal bleeding. Gynecol Oncol 56:376–381. Fend F, Raffeld M. 2000. Laser capture microdissection in pathology. J Clin Pathol 53:666–672. Fenoglio-Preiser CM, Noffsinger A. 1999. Aberrant crypt foci: a review. Toxicol Pathol 27:632–642. Ferenczy A. 1997. Glandular lesions: an increasing problem. In: Franco EL, Monsonego J, editors. New Developments in Cervical Cancer Screening and Prevention. London: Blackwell, pp. 122–130. Ferenczy A, Gelfand M. 1989. The biologic significance of cytologic atypia in progestogen-treated endometrial hyperplasia. Am J Obstet Gynecol 160:126–131. Ferenczy A, Mutter G. 2000. Endometrial hyperplasia and neoplasia: definition, diagnosis and management principles. In: Sciarra J, editor. Gynecology and Obstetrics Looseleaf CD-ROM. Philadelphia: Lippincott, Williams & Wilkins. Fisher B, Costantino JP, Redmond CK. 1994. Endometrial cancer in tamoxifen-treated breast cancer patients: findings from the National Surgical Adjuvant Breast and Bowel Project (NSABP) B-14. J Natl Cancer Inst 86:527–537. Fisher BL, Pennathur A, Mutnick JL, et al. 1999. Obesity correlates with gastroesophageal reflux. Dig Dis Sci 44:2290–2294. Fitzgibbons PL, Henson DE, Hutter RVP. 1998. Benign breast changes and the risk for subsequent breast cancer: an update of the 1985 consensus statement. Cancer Committee of the College of American Pathologists. Arch Pathol Lab Med 122:1053–1055. Fleming NT, Armstrong BK, Sheiner HJ. 1982. The comparative epidemiology of benign breast lumps and breast cancer in western Australia. Int J Cancer 30:147–152. Ford LC, Berek JS, Lagasse LD, Hacker NF, Heins YL, DeLange RJ. 1983. Estrogen and progesterone receptor sites in malignancies of the uterine cervix, vagina, and vulva. Gynecol Oncol 15:27–31. Foulds L. 1958. The natural history of cancer. J Chronic Dis 8:2–37. Franceschi S, Munoz N, Bosch XF, Snijders PJ, Walboomers JM. 1996. Human papillomavirus and cancers of the upper aerodigestive tract: a review of epidemiological and experimental evidence. Cancer Epidemiol Biomarkers Prev 5:567–575. Franco EL. 1991. The sexually-transmitted disease model for cervical cancer: incoherent epidemiologic findings and the role of misclassification of human papillomavirus infection. Epidemiology 2:98–106. Franco EL. 1997. Epidemiology of uterine cancers. In: Meisels A, Morin C, editors. Cytopathology of the uterus, 2nd ed. Chicago: American Society of Clinical Pathologists, pp. 301–324. Franco EL. 2003. Primary screening of cervical cancer with human papillomavirus tests. J Natl Cancer Inst Monogr 31:89–96. Franco EL, Ferenczy A. 2002a. Endometrium. In: Franco EL, Rohan TE, editors. Cancer Precursors: Epidemiology, Detection, and Prevention. New York: Springer-Verlag, pp. 287–303.

41

Franco EL, Ferenczy A. 2002b. Cervix. In: Franco EL, Rohan TE, editors. Cancer Precursors: Epidemiology, Detection, and Prevention. New York: Springer-Verlag, pp. 249–286. Franco EL, Rohan TE, editors. 2002. Cancer Precursors: Epidemiology, Detection, and Prevention. New York: Springer-Verlag. Franco EL, Duarte-Franco E, Rohan TE. 2002. Evidence-based policy recommendations on cancer screening and prevention. Cancer Detect Prev 26:350–361. Frantz VK, Pickren JW, Melcher GW, Auchincloss H Jr. 1951. Incidence of chronic cystic disease in so-called normal breasts. Cancer 4:762– 783. Freedman LS, Graubard BI, Schatzkin A. 1992. Statistical validation of intermediate endpoints for chronic diseases. Stat Med 11:167–178. Friedenreich C, Bryant H, Alexander F, et al. 2000. Risk factors for benign proliferative breast disease. Int J Epidemiol 29:637–644. Frykberg ER. 1999. Lobular carcinoma in situ of the breast. Breast J 5:296–303. Frykberg ER, Bland KI. 1993. In situ breast carcinoma. Adv Surg 26:29–72. Fuchs CS, Giovannucci EL, Colditz GA, et al. 1999. Dietary fiber and the risk of colorectal cancer and adenoma in women. N Engl J Med 340: 169–176. Gerson LB, Edson R, Lvori PW, Triadafilapoulos G. 2001. Use of a simple symptom questionnaire to predict Barrett’s esophagus in patients with symptoms of gastroesophageal reflux. Am J Gastroenterol 96:2005–2012. Giardiello FM, Hamilton SR, Krush AJ, et al. 1993. Treatment of colonic and rectal adenomas with sulindac in familial adenomatous polyposis. N Engl J Med 328:1313–1316. Giovannucci E, Martinez ME. 1996. Tobacco, colorectal cancer, and adenomas: a review of the evidence. J Natl Cancer Inst 88:1717–1730. Giovannucci E, Ascherio A, Rimm EB, et al. 1995. Physical activity, obesity, and risk for colon cancer and adenoma in men. Ann Intern Med 122:327–334. Giovannucci E, Colditz GA, Stampfer MJ, Willett WC. 1996. Physical activity, obesity, and risk of colorectal adenoma in women (United States). Cancer Causes Control 7:253–263. Giovannucci E, Stampfer MJ, Colditz G, et al. 1992. Relationship of diet to risk of colorectal adenoma in men. J Natl Cancer Inst 84:91–98. Giovannucci E, Stampfer MJ, Colditz GA, et al. 1993. Folate, methionine, and alcohol intake and risk of colorectal adenoma. J Natl Cancer Inst 85:875–883. Glover KY, Papadimitrakopolou VA. 2003. Chemoprevention of head and neck cancer. Curr Oncol Rep 5:152–157. Gobbi H, Dupont WD, Simpson JF, et al. 1999. Transforming growth factorbeta and breast cancer risk in women with mammary epithelial hyperplasia. J Natl Cancer Inst 91:2096–2101. Gong G, DeVries S, Chew KL, et al. 2001. Genetic changes in paired atypical and usual ductal hyperplasia of the breast by comparative genomic hybridization. Clin Cancer Res 7:2410–2414. Grady D, Ernster VL. 1996. Endometrial cancer. In: Schottenfeld D, Fraumeni JF Jr, editors. Cancer epidemiology and prevention. New York: Oxford University Press, pp. 1058–1089. Grbic JT, Lamster IB. 1997. Oral manifestations of HIV infection. AIDS Patient Care STDS 1:18–24. Gredmark T, Kvint S, Havel G, Mattsson LA. 1999. Adipose tissue distribution in postmenopausal women with adenomatous hyperplasia of the endometrium. Gynecol Oncol 72:138–142. Green J, Berrington de Gonzalez A, Smith JS, et al. 2003. Human papillomavirus infection and use of oral contraceptives. Br J Cancer 88:1713–1720. Grimbizis G, Tsalikis T, Tzioufa V, Kasapis M, Mantalenakis S. 1999. Regression of endometrial hyperplasia after treatment with the gonadotrophinreleasing hormone analogue triptorelin: a prospective study. Hum Reprod 14:479–484. Grodstein F, Marinez E, Platz EA, et al. 1998. Postmenopausal hormone use and risk for colorectal cancer and adenoma. Ann Intern Med 128:705–712. Gruppo Operativo per lo Studio delle Precancerosi delle’Esofago (GOSPE). 1991. Barrett’s esophagus: epidemiological and clinical results of a multicentric survey. Int J Cancer 48:364–368. Guha U, Hulit J, Pestell RG. 2001. Transgenic mice in cancer research. In: Bertino J, editor. Encyclopedia of cancer, 2nd ed. San Diego: Academic Press, pp. 449–458. Gupta PC, Hebert JR, Bhonsle RB, et al. 1999. Influence of dietary factors on oral precancerous lesions in a population-based case-control study in Kerala, India. Cancer 85:1885–1893. Gupta PC, Mehta FS, Daftary DK, et al. 1980. Incidence of oral cancer and natural history of oral precancerous lesions in a 10-year follow-up study of Indian villagers. Comm Dent Oral Epidemiol 8:287–333.

42

PART I: BASIC CONCEPTS

Gupta PC, Pindborg JJ, Bhonsle RB, et al. 1986. Intervention study for primary prevention of oral cancer among 36,000 Indian tobacco users. Lancet 1:1235–1239. Hamilton SR. 2001. Origin of colorectal cancers in hyperplastic polyps and serrated adenomas: another truism bites the dust. J Natl Cancer Inst 93:1282–1283. Hashibe M, Sankaranarayanan R, Thomas G, et al. 2000. Alcohol drinking, body mass index and the risk of oral leukoplakia in an Indian population. Int J Cancer 88:129–134. Heath EI, Limburg PJ, Hawk ET, Forastiere AA. 2000. Adenocarcinoma of the esophagus: risk factors and prevention. Oncology (Huntingt) 14: 507–514. Henson DE, Albores-Saavedra J, editors. 2001. Pathology of Incipient Neoplasia, 3rd ed. New York: Oxford University Press. Herrero R, Castellsague X, Pawlita M, et al. 2003. Human papillomavirus and oral cancer: the International Agency for Research on Cancer multicenter study. J Natl Cancer Inst 95:1772–1783. Hildesheim A, Wang SS. 2002. Host and viral genetics and risk of cervical cancer: a review. Virus Res 89:229–240. Hill MJ, Morson BC, Bussey HJ. 1978. Aetiology of adenoma-carcinoma sequence in large bowel. Lancet 1:245–247. Hille JJ, Webster-Cyriaque J, Palefski JM, Raab-Traub N. 2002. Mechanisms of expression of HHV8, EBV and HPV in selected HIV-associated oral lesions. Oral Dis 8:161–168. Hirota WK, Lughney TM, Lazas DJ, et al. 1999. Specialized intestinal metaplasia, dysplasia, and cancer of the esophagus and esophagogastric junction: prevalence and clinical data. Gastroenterology 116:277–285. Hislop TG, Band PR, Deschamps M, et al. 1990. Diet and histologic types of benign breast disease defined by subsequent risk of breast cancer. Am J Epidemiol 131:263–270. Hittelman WN. 2001. Genetic instability in epithelial tissues at risk for cancer. Ann NY Acad Sci 952:1–12. Ho GY, Kadish AS, Burk RD, et al. 1998. HPV 16 and cigarette smoking as risk factors for high-grade cervical intra-epithelial neoplasia. Int J Cancer 78:281–285. Ho SP, Tan KT, Pang MW, Ho TH. 1997. Endometrial hyperplasia and the risk of endometrial carcinoma. Singapore Med J 38:11–15. Holowaty P, Miller AB, Rohan T, To T. 1999. The natural history of dysplasia of the uterine cervix. J Natl Cancer Inst 91:252–258. Hong WK. 2003. General keynote: the impact of cancer chemoprevention. Gynecol Oncol 88:S56–S58. Hong WK, Endicott J, Itri LM, et al. 1986. 13-cis-retinoic acid in the treatment of oral leukoplakia. N Engl J Med 315:1501–1505. Hsieh C, Walker AM, Trapido EJ, et al. 1984. Age at first birth and breast atypia. Int J Cancer 33:309–312. Humphrey LJ, Swerdlow M. 1966. Histologic changes in clinically normal breasts at postmortem examination. Arch Surg 92:192–193. Hunter JE, Tritz DE, Howell MG, et al. 1994. The prognostic and therapeutic implications of cytologic atypia in patients with endometrial hyperplasia. Gynecol Oncol 55:66–71. Ikediobi NI, Tyring SK. 2002. Cutaneous manifestations of Epstein-Barr virus infection. Dermatol Clin 20:283–289. Ingram D, Nottage E, Ng S, et al. 1989. Obesity and breast disease: the role of female sex hormones. Cancer 64:1049–1053. Ingram DM, Nottage E, Roberts T. 1991. The role of diet in the development of breast cancer: a case-control study of patients with breast cancer, benign epithelial hyperplasia and fibrocystic disease of the breast. Br J Cancer 64:187–191. Inoue H, Kiyohara C, Marugame T, et al. 2000. Cigarette smoking, CYP1A1 MspI and GSTM1 genotypes, and colorectal adenomas. Cancer Res 60:3749–3752. Inoue M. 2001. Current molecular aspects of the carcinogenesis of the uterine endometrium. Int J Gynecol Cancer 11:339–348. International Agency for Research on Cancer (IARC). 1985. Tobacco habits other than smoking; Betel-quid and areca-nut chewing; and some related nitrosamines. In: IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Vol. 37. Lyon: IARC. International Agency for Research on Cancer (IARC). 1986. Tobacco smoking. In: IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Vol. 38. Lyon: IARC. Isaacson PG. 1999. Mucosa-associated lymphoid tissue lymphoma. Semin Hematol 36:139–147. Jaber MA, Porter SR, Gilthorpe MS, et al. 1999. Risk factors for oral epithelial dysplasia—the role of smoking and alcohol. Eur J Cancer Oral Oncol 35B:151–156. Janicek MF, Rosenshein NB. 1994. Invasive endometrial cancer in uteri resected for atypical endometrial hyperplasia. Gynecol Oncol 52: 373–378.

Jänne PA, Mayer RJ. 2000. Chemoprevention of colorectal cancer. N Engl J Med 342:1960–1968. Jass JR. 1991. Nature and clinical significance of colorectal hyperplastic polyp. Semin Colon Rectal Surg 2:246–252. Jass JR. 1999. Serrated adenoma and colorectal cancer. J Pathol 187:499– 502. Jotwani G, Misra A, Chattopadhyay P, et al. 2001. Genetic heterogeneity and alterations in chromosome 9 loci in a localized region of a functional pituitary adenoma. Cancer Genet Cytogenet 125:41–45. Kampert JB, Wood DA, Paffenbarger RS. 1983. Oral contraceptives and breast disease. In: Feig SA, McLelland R, editors. Breast Carcinoma. Current Diagnosis and Treatment. Chicago: Masson, pp. 597–607. Kandel R, Li SQ, Ozcelik H, Rohan T. 2000. P53 protein accumulation and mutations in normal and benign breast tissue. Int J Cancer 87:73–8. Kato I, Tominaga S, Matsuura A, et al. 1990. A comparative case-control study of colorectal cancer and adenoma. Jpn J Cancer Res 81:1101– 1108. Kelloff GJ, Sigman CC. 2002. Chemoprevention. In: Franco EL, Rohan TE, editors. Cancer Precursors: Epidemiology, Detection, and Prevention. New York: Springer-Verlag, pp. 46–59. Kelloff GJ, O’Shaughnessy JA, Gordon GB, et al. 2003. Counterpoint: because some surrogate end point biomarkers measure the neoplastic process they will have high utility in the development of cancer chemopreventive agents against sporadic cancer. Cancer Epidemiol Biomark Prev 12: 593–596. Kelloff GJ, Sigman CC, Johnson KM, et al. 2000. Perspectives on surrogate end points in the development of drugs that reduce the risk of cancer. Cancer Epidemiol Biomark Prev 9:127–137. Kennedy TC, Proudfoot SP, Franklin WA, et al. 1996. Cytopathological analysis of sputum in patients with airflow distruction and significant smoking histories. Cancer Res 56:4673–4678. Kern SE. 1993. Clonality: more than just a tumor-progression model. J Natl Cancer Inst 85:1020–1021. Kerr KM. 2001. Pulmonary preinvasive neoplasia. J Clin Pathol 54:257–271. Kim R, Weissfeld JL, Reynolds JC, Kuller LH. 1997. Etiology of Barrett’s metaplasia and esophageal adenocarcinoma. Cancer Epidemiol Biomark Prev 6:369–377. Kleinman DV, Swango PA, Niesson LC. 1991. Epidemiologic studies of oral mucosal conditions—methodological issues. Community Dent Oral Epidemiol 19:129–140. Konishi F, Morson BC. 1982. Pathology of colorectal adenomas: a colonoscopic survey. J Clin Pathol 35:830–841. Kono S, Handa K, Hayabuchi H, et al. 1999. Obesity, weight gain and risk of colon adenomas in Japanese men. Jpn J Cancer Res 90:805–811. Koss LG. 1995. Detection of occult endometrial carcinoma. J Cell Biochem Suppl 23:165–173. Koss LG, Schreiber K, Oberlander SG, Moussouris HF, Lesser M. 1984. Detection of endometrial carcinoma and hyperplasia in asymptomatic women. Obstet Gynecol 64:1–11. Koushik A, Platt RW, Franco EL. 2004. P53 codon 72 polymorphism and cervical neoplasia. A meta-analysis review. Cancer Epidemiol Biomarkers Prev 13:11–22. Kramer WM, Rush BF Jr. 1973. Mammary duct proliferation in the elderly: a histopathologic study. Cancer 31:130–137. Kreiger N, Marrett LD, Clarke EA, Hilditch S, Woolever CA. 1986. Risk factors for adenomatous endometrial hyperplasia: a case-control study. Am J Epidemiol 123:291–301. Kurman RJ, Felix JC, Archer DF, Nanavati N, Arce J, Moyer DL. 2000. Norethindrone acetate and estradiol-induced endometrial hyperplasia. Obstet Gynecol 96:373–379. Kurman RJ, Kaminski PF, Norris HJ. 1985. The behavior of endometrial hyperplasia: a long-term study of “untreated” hyperplasia in 170 patients. Cancer 56:403–412. Kurman RJ, Malkasian GD Jr, Sedlis A, Solomon D. 1991. From Papanicolaou to Bethesda: the rationale for a new cervical cytologic classification. Obstet Gynecol 77:779–782. Kurman RJ, Norris HJ, Wilkinson E. 1992. Tumors of the cervix, vagina, and vulva. In: Atlas of Tumor Pathology, Third Series, Fascicle 4. Washington, DC: Armed Forces Institute of Pathology. Labayle D, Fischer D, Vielh P, et al. 1991. Sulindac causes regression of rectal polyps in familial adenomatous polyposis. Gastroenterology 101: 635–639. Lacey JV Jr, Brinton LA, Barnes WA, et al. 2000. Use of hormone replacement therapy and adenocarcinomas and squamous cell carcinomas of the uterine cervix. Gynecol Oncol 77: 149–154 Lagergren J, Bergstrom R, Lindgren A, Nyren O. 1999a. Symptomatic gastroesophageal reflux as a risk factor for esophageal adenocarcinoma. N Engl J Med 340:825–831.

Cancer Precursors Lagergren J, Bergstrom R, Nyren O. 1999b. Association between body mass and adenocarcinoma of the esophagus and gastric cardia. Ann Intern Med 130:883–890. Lagergren J, Bergstom R, Nyren O. 2000. No relation between body mass and gastro-esophageal reflux symptoms in a Swedish population-based study. Gut 47:26–29. Lakhani SR. 1999. The transition from hyperplasia to invasive carcinoma of the breast. J Pathol 187:272–278. Lance LL. 1981. Risk factors for benign breast disease. Am J Epidemiol 114:606–607. Lawson HW, Lee NC, Thames SF, Henson R, Miller DS. 1998. Cervical cancer screening among low-income women: results of a national screening program, 1991-1995. Obstet Gynecol 92:745–752. Learmonth GM, Durcan CM, Beck JD. 1990. The changing incidence of cervical intra-epithelial neoplasia. S Afr Med J 77:637–639. Li Y, Hively WP, Varmus HE. 2000. Use of MMTV-Wnt-1 transgenic mice for studying the genetic basis of breast cancer. Oncogene 19:1002–1009. Lieberman DA, Oehlke M, Helfand M, GORGE consortium. 1997. Risk factors for Barrett’s esophagus in community-based practice. Am J Gastroenterol 92:1293–1297. Lindahl B, Willen R. 1991. Endometrial hyperplasia following estrogen treatment without the addition of gestagen. A follow-up study after withdrawal of estrogens. Anticancer Res 11:2071–2073. Lippman SM, Batsakis JG, Toth BB, et al. 1993. Comparison of low-dose isotretinoin with beta carotene to prevent oral carcinogenesis. N Engl J Med 328:15–20. Lippman SM, Lee JJ, Sabichi AL. 1998. Cancer chemoprevention: progress and promise. J Natl Cancer Inst 90:1514–1528. Little J, Logan RFA, Hawtin PG, et al. 1993. Colorectal adenomas and diet: a case-control study of subjects participating in the Nottingham faecal occult blood screening programme. Br J Cancer 67:177–184. LiVolsi VA, Stadel BV, Kelsey JL, et al. 1978. Fibrocystic disease in oralcontraceptive users. N Engl J Med 299:381–385. Locke GR, Talley NJ, Fett SL. 1999. Risk factors associated with sysmptoms of gastroesophageal reflux. Am J Med 106:642–649. Loffeld RJLF, Werdmuller BFM, Kuster JG, et al. 2000. Colonization with cagA-positive Helicobacter pylori strains inversely associated with reflux esophagitis and Barrett’s esophagus. Digestion 62:95–99. London SJ, Sacks FM, Stampfer MJ, et al. 1993. Fatty acid composition of the subcutaneous adipose tissue and risk of proliferative benign breast disease and breast cancer. J Natl Cancer Inst 85:785–793. London SJ, Stein EA, Henderson IC, et al. 1992. Carotenoids, retinol, and vitamin E and risk of proliferative benign breast disease and breast cancer. Cancer Causes Control 3:503–512. Lorincz AT, Castle PE, Sherman ME, et al. 2002. Viral load of human papillomavirus and risk of CIN3 or cervical cancer. Lancet 360:228– 229. Lotfi AM, Spencer RJ, Ilstrup DM, Melton LJ III. 1986. Colorectal polyps and the risk of subsequent carcinoma. Mayo Clin Proc 1986;61:337– 343. Lubin F, Rozen P, Arieli B, et al. 1997. Nutritional and lifestyle habits and water-fiber intraction in colorectal adenoma etiology. Cancer Epidemiol Biomarkers Prev 6:79–85. Lubin F, Wax Y, Ron E, et al. 1989. Nutritional factors associated with benign breast disease etiology: a case-control study. Am J Clin Nutr 50:551–556. Lumerman H, Freedman P, Kerpel S. 1995. Oral epithelial dysplasia and the development of invasive squamous cell carcinoma. Oral Surg Oral Med Oral Pathol 79:321–329. Luna MA, Pineda-Daboin K. 2001. Upper aerodigestive tract. In: Henson DE, Albores-Saavedra J, editors. Pathology of Incipient Neoplasia, 3rd ed. New York: Oxford University Press, pp. 57–85. Luttges J, Schlehe B, Menke MA, et al. 1999. The K-ras mutation pattern in pancreatic ductal adenocarcinoma usually is identical to that in associated normal, hyperplastic, and metaplastic ductal epithelium. Cancer 85:1703–1710. Lydiatt WM, Anderson PE, Bazzana T, et al. 1998. Molecular support for field cancerization in the head and neck. Cancer 82:1376–1380. Macquart-Moulin G, Riboli E, Corne J, et al. 1987. Colorectal polyps and diet: a case-control study in Marseilles. Int J Cancer 40:179–188. Maher R, Lee AJ, Warnakulasuriya S, et al. 1994. Role of arecanut in the causation of oral submucous fibrosis: a case-control study in Pakistan. J Oral Pathol Med 23:65–69. Makni H, Franco E, Kaiano J, et al. 2000. P53 polymorphism in codon 72 and risk of HPV-induced cervical cancer: effect of inter-laboratory variation. Int J Cancer 87:528–33. Mao L, Lee JS, Fan YH, et al. 1996. Frequent microsatellite alterations at chromosomes 9p21 and 3p14 in oral premalignant lesions and their value in cancer risk assessment. Nat Med 2:682–685.

43

Martinez, ME, McPherson RS, Annegers JF, et al. 1995. Cigarette smoking and alcohol consumption as risk factors for colorectal adenomatous polyps. J Natl Cancer Inst 87:274–279. Martinez, ME, McPherson RS, Annegers JF, et al. 1996. Association of diet and colorectal adenomatous polyps: dietary fiber, calcium, and total fat. Epidemiology 7:264–268. Matek W, Hermanek P, Demling L. 1986. Is the adenoma-carcinoma sequence contradicted by the differing location of colorectal adenomas and carcinomas? Endoscopy 18:17–19. Matias-Guiu X, Catasus L, Bussaglia E, et al. 2001. Molecular pathology of endometrial hyperplasia and carcinoma. Hum Pathol 32:569–577. McIndoe WA, McLean MR, Jones RW, Mullins PR. 1984. The invasive potential of carcinoma in situ of the cervix. Obstet Gynecol 64:451–458. McMichael AJ. 1994. Invited commentary “molecular epidemiology”: new pathway or new traveling companion? Am J Epidemiol 140:1–11. Mehta FS, Aghi MB, Gupta PC, et al. 1982. An intervention study of oral cancer and precancer in rural Indian populations: a preliminary report. Bull WHO 60:441–446. Mehta FS, Pindborg JJ, Bhonsle RB, et al. 1976. Incidence of oral leukoplakias among 20,358 Indian villagers in a 7-year period. Br J Cancer 33: 549–554. Mehta FS, Pindborg JJ, Gupta PC, et al. 1969. Epidemiologic and histologic study of oral cancer and leukoplakia among 50,915 villagers in India. Cancer 24:832–849. Meisels A, Morin C. 1997. Cytopathology of the uterus, 2nd ed. Chicago: ASCP Press. Melnikow J, Nuovo J, Willan AR, et al. 1998. Natural history of cervical squamous intraepithelial lesions: a meta-analysis. Obstet Gynecol 92(Pt 2):727–735. Miller AB, Nazeer S, Fonn S, et al. 2000. Report on consensus conference on cervical cancer screening and management. Int J Cancer 86:440–447. Miller CS, White DK. 1996. Human papillomavirus expression in oral mucosa, premalignant conditions, and squamous cell carcinoma: a retrospective review of the literature. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 82:57–68. Miller FR, Soule HD, Tait L, et al. 1993. Xenograft model of progressive human proliferative breast disease. J Natl Cancer Inst 85:1725–1732. Minami Y, Ohuchi N, Taeda Y, et al. 1998. Risk factors for benign breast disease according to histopathological type: comparisons with risk factors for breast cancer. Jpn J Cancer Res 89:116–123. Minami Y, Ohuchi N, Taeda Y, et al. 1999. Risk of breast cancer in Japanese women with benign breast disease. Jpn J Cancer Res 90:600–606. Minna JD, Sekido Y, Fong KM, et al. 1997. Molecular biology of lung cancer. In: DeVita VT Jr, Hellman S, Rosenberg SA, editors. Cancer Principles and Practice of Oncology. Philadelphia: Lippincott-Raven. Miros M, Kerlin P, Walker N. 1991. Only patients with dysplasia progress to adenocarcinoma in Barrett’s esophagus. Gut 32:1441–1446. Mitchell MF, Hittelman WN, Hong WK, Lotan R, Schottenfeld D. 1994. The natural history of cervical intraepithelial neoplasia: an argument for intermediate endpoint biomarkers. Cancer Epidemiol Biomarkers Prev 3:619–626. Monforte-Munoz H, Kapoor N, Albores-Saavedra J. 2003. Epstein-Barr virusassociated leiomyomatosis and post-transplant lymphoproliferative disorder in a child with severe combined immunodeficiency: case report and review of the literature. Pediatr Dev Pathol 6:449–457. Moreno V, Bosch FX, Munoz N, et al. 2002. Effect of oral contraceptives on risk of cervical cancer in women with human papillomavirus infection: the IARC multicentric case-control study. Lancet 359:1085–1092. Morrison BJ, Coldman AJ, Boyes DA, Anderson GH. 1996. Forty years of repeated screening: the significance of carcinoma in situ. Br J Cancer 74:814–819. Morse DE, Katz RV, Pendrys DG, et al. 1996. Smoking and drinking in relation to oral epithelial dysplasia. Cancer Epidemiol Biomarkers Prev 5:769–777. Morson BC. 1974. The polyp-cancer sequence in the large bowel. Proc R Soc Med 67:451–457. Moskaluk CA, Hruban RH, Kern SE. 1997. P16 and K-ras gene mutations in the intraductal precursors of human pancreatic adenocarcinoma. Cancer Res 57:2140–2143. Mourits MJ, De Vries EG, Willemse PH, Ten Hoor KA, Hollema H, Van der Zee AG. 2001. Tamoxifen treatment and gynecologic side effects: a review. Obstet Gynecol 97:855–866. Muller AD, Sonnenberg A. 1995. Prevention of colorectal cancer by flexible endoscopy and polypectomy: a case-control study of 32,702 veterans. Ann Intern Med 123:904–910. Munoz N, Franceschi S, Bosetti C, et al. 2002. Role of parity and human papillomavirus in cervical cancer: the IARC multicentric case-control study. Lancet 359:1093–1101.

44

PART I: BASIC CONCEPTS

Muto T. 1989. Serial section of colonic adenomas with special reference to minute carcinoma. Jpn J Cancer Res 80:356–359. Muto T, Bussey HJR, Morson BC. 1975. The evolution of cancer of the colon and rectum. Cancer 36:2251–2270. Mutter GL. 2000. Endometrial intraepithelial neoplasia (EIN): will it bring order to chaos? The Endometrial Collaborative Group. Gynecol Oncol 76:287–290. Mutter GL, Baak JP, Crum CP, Richart RM, Ferenczy A, Faquin WC. 2000. Endometrial precancer diagnosis by histopathology, clonal analysis, and computerized morphometry. J Pathol 190:462–469. Myers ER, McCrory DC, Nanda K, et al. 2000. Mathematical model for the natural history of human papillomavirus infection and cervical carcinogenesis. Am J Epidemiol 151:1158–1171. Nagata C, Shimizu H, Kametani M, et al. 1999. Cigarette smoking, alcohol use, and colorectal adenoma in Japanese men and women. Dis Colon Rectum 42:337–342. Nagata CH, Shimizu H, Kametani M, et al. 2001. Diet and colorectal adenoma in Japanese males and females. Dis Colon Rectum 44:105–111. Neugut AI, Jacobson JS, DeVivo I. 1993. Epidemiology of colorectal adenomatous polyps. Cancer Epidemiol Biomarkers Prev 2:159–176. Neugut AI, Lee WC, Garbowski GC, et al. 1991. Obesity and colorectal adenomatous polyps. J Natl Cancer Inst 1991;83:359–361. Neville BW, Day TA. 2002. Oral cancer and precancerous lesions. CA Cancer J Clin 52:195–215. Nielsen M, Jensen J, Andersen J. 1984. Precancerous and cancerous breast lesions during lifetime and at autopsy. A study of 83 women. Cancer 54:612–615. Nielsen M, Thomsen JL, Primdahl S, Dyreborg U, Andersen JA. 1987. Breast cancer and atypia among young and middle aged women: a study of 110 medicolegal autopsies. Br J Cancer 56:814–819. Nowell PC. 1976. The clonal evolution of tumor cell populations. Science 194:23–28. Nugent KP, Farmer KC, Spigelman AD, et al. 1993. Randomized controlled trial of the effect of sulindac on duodenal and rectal polyposis and cell proliferation in patients with familial adenomatous polyposis. Br J Surg 80:1618–1619. Ochiai A, Yamauchi Y, Hirohashi S. 1996. P53 mutations in the nonneoplastic mucosa of the human stomach showing intestinal metaplasia. Int J Cancer 69:28–33. O’Connor JB, Falk GW, Richter JE. 1999. The incidence of adenocarcinoma and dysplasia in Barrett’s esophagus. Am J Gastroenterol 94:2037–2042. Olsen J, Kronborg O, Lynggaard J, Ewertz M. 1994. Dietary risk factors for cancer and adenomas of the large intestine. A case-control study within a screening trial in Denmark. Eur J Cancer 1994;30A:53–60. Ory H, Cole P, MacMahon B, Hoover R. 1976. Oral contraceptives and reduced risk of benign breast disease. N Engl J Med 249:419–422. O’Shaughnessy JA, Kelloff GJ, Gordon GB, et al. 2002. Treatment and prevention of intraepithelial neoplasia: an important target for accelerated new drug development; recommendations of the American Association for Cancer Research Task Force on the Treatment and Prevention of Intraepithelial Neoplasia. Clin Cancer Res 8:314–346. Ostor AG. 1993. Natural history of cervical intraepithelial neoplasia: a critical review. Int J Gynecol Pathol 12:186–192. Ostor AG, Pagano R, Davoren RA, Fortune DW, Chanen W, Rome R. 1984. Adenocarcinoma in situ of the cervix. Int J Gynecol Pathol 3:179–190. Otchy DP, Ransohoff DR, Wolff BG, et al. 1996. Metachronous colon cancer in persons who have had a large adenomatous polyp. Am J Gastroenterol 91:448–454. Pacey NF. 1991. Glandular neoplasms of the uterine cervix. In: Bibbo M, editor. Comprehensive Cytopathology. Philadelphia: Saunders, pp. 231–256. Page DL, Dupont WD, Rogers LW, Rados MS. 1985. Atypical hyperplastic lesions of the female breast. A long-term follow-up study. Cancer 55:2698–2708. Palli D, del Turco MR, Simoncini R, Bianchi S. 1991. Benign breast disease and breast cancer: a case-control study in a cohort in Italy. Int J Cancer 47:703–706. Papanicolaou GN. 1954. Atlas of Exfoliative Cytology. Cambridge, MA: Harvard University Press. Parazzini F, Ferraroni M, La Vecchia C, et al. 1991. Smoking habits and risk of benign breast disease. Int J Epidemiol 20:430–434. Parazzini F, La Vecchia C, Franceschi S, et al. 1984. Risk factors for pathologically confirmed benign breast disease. Am J Epidemiol 120: 115–122. Park I-W, Wistuba II, Maitra A, et al. 1999. Multiple clonal abnormalities in the bronchial epithelium of patients with lung cancer. J Natl Cancer Inst 91:1863–1868. Partridge M, Pateromichelakis S, Phillips E, Emilion GG, A’Hern RP, Langdon JD. 2000. A case control-study confirms that microsatellite assay can

identify patients at risk of developing oral squamous cell carcinoma within a field of cancerization. Cancer Res 60:3893–3898. Pastides H, Kelsey JL, Holford TR, LiVolsi VA. 1985. An epidemiologic study of fibrocystic breast disease with reference to ductal epithelial atypia. Am J Epidemiol 121:440–447. Pastides H, Kelsey JL, LiVolsi VA, Holford TR, Fischer DB, Goldenberg IS. 1983. Oral contraceptive use and fibrocystic breast disease with special reference to its histopathology. J Natl Cancer Inst 71:5–9. Peipins LA, Sandler RS. 1994. Epidemiology of colorectal adenomas. Epidemiol Rev 16:273–297. Peipins LA, Newman B, Sandler RS. 1997. Reproductive history, use of exogenous hormones, and risk of colorectal adenomas. Cancer Epidemiol Biomarkers Prev 6:671–675. Perou CM, Sørlie T, Eisen MB, et al. 2000. Molecular portraits of human breast tumors. Nature 406:747–757. Peters U, Sinha R, Chatterjee N, et al. 2003. Dietary fibre and colorectal adenoma in a colorectal cancer early detection programme. Lancet 361:1491–1495. Pindborg JJ, Jolst O, Renstrup G, et al. 1968. Studies in oral leukoplakia: a preliminary report on the period prevalence of malignant transformation in leukoplakia based on a follow-up study of 248 patients. J Am Dent Assoc 76:767–771. Pinsky PF. 2000. A multi-stage model of adenoma development. J Theor Biol 207:129–143. Platz CE, Benda JA. 1995. Female genital tract cancer. Cancer 75(Suppl 1):270–294 Platz EA, Giovannucci E, Rimm EB, et al. 1997. Dietary fiber and distal colorectal adenoma in men. Cancer Epidemiol Biomarkers Prev 6:661–670. Plaxe SC, Saltzstein SL. 1999. Estimation of the duration of the preclinical phase of cervical adenocarcinoma suggests that there is ample opportunity for screening. Gynecol Oncol 75:55–61. Pontén J, editor. 1998. Precancer: biology, importance and possible prevention. In: Cancer Surveys, Vol. 32. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. Potischman N, Brinton LA. 1996. Nutrition and cervical neoplasia. Cancer Causes Control 7:113–126. Potter JD. 1999. Colorectal cancer: molecules and populations. J Natl Cancer Inst 91:916–932. Potter JD, Bostick RM, Grandits GA, et al. 1996. Hormone replacement therapy is associated with lower risk of adenomatous polyps of the large bowel: the Minnesota cancer prevention research unit case-control study. Cancer Epidemiol Biomarkers Prev 5:779–784. Prach AT, MacDonald T, Hopwood D, Johnston DA. 1997. Increasing incidence of Barrett’s oesophagus: education, enthusiasm, or epidemiology? Lancet 350:933. Prentice RL. 1989. Surrogate endpoints in clinical trials: definition and operational criteria. Stat Med 8:431–440. Ramaswamy G, Rao VR, Kumaraswamy SV, Anantha N. 1996. Serum vitamins’ status in oral leucoplakias—a preliminary study. Eur J Cancer B Oral Oncol 32B:120–122. Reid BJ, Haggitt RC, Rubin CE, et al. 1988. Observer variation in the diagnosis of dysplasia in Barrett’s esophagus. Hum Pathol 19:166–178. Ren ZP, Ahmadian A, Ponten F, et al. 1997. Benign clonal keratinocyte patches with p53 mutations show no genetic link to synchronous squamous cell precancer or cancer in human skin. Am J Pathol 150:1791–1803. Richart RM. 1968. Natural history of cervical intraepithelial neoplasia. Clin Obstet Gynecol 5:748–784. Rickert RR, Auerbach O, Garfinkel L, et al. 1979. Adenomatous lesions of the large bowel: an autopsy survey. Cancer 43:1847–1857. Ries LAG, Eisner MP, Kosary CL, et al, editors. 2002. SEER Cancer Statistics Review, 1973–1999. Bethesda, MD: National Cancer Institute (http://seer.cancer.gov/csr/1973_1999/). Rohan TE. 1999. Cigarette smoking and risk of benign proliferative epithelial disorders of the female breast. Eur J Epidemiol 15:529–535. Rohan TE, Cook MG. 1989. Alcohol consumption and risk of benign proliferative epithelial disorders of the breast in women. Int J Cancer 43:631–636. Rohan TE, Kandel RA. Breast. In: Franco EL, Rohan TE, editors. Cancer Precursors: Epidemiology, Detection, and Prevention. New York: SpringerVerlag, pp. 232–248. Rohan TE, Miller AB. 1999a. A cohort study of oral contraceptives and risk of benign breast disease. Int J Cancer 82:191–196. Rohan TE, Miller AB. 1999b. Hormone replacement therapy and risk of benign proliferative epithelial disorders of the breast. Eur J Cancer Prev 8:123–130. Rohan TE, Cook MG, Baron JA. 1989. Cigarette smoking and benign proliferative epithelial disorders of the breast in women: a case-control study. J Epidemiol Community Health 43:362–368.

Cancer Precursors Rohan TE, Cook MG, Potter JD, McMichael AJ. 1990. A case-control study of diet and benign proliferative epithelial disorders of the breast. Cancer Res 50:3176–3181. Rohan TE, Hartwick W, Miller AB, Kandel RA. 1998. Immunohistochemical detection of c-erbB-2 and p53 proteins in benign breast disease and breast cancer risk. J Natl Cancer Inst 90:1262–1269. Rohan TE, Jain M, Miller AB. 1998a. Alcohol consumption and risk of benign proliferative epithelial disorders of the breast: a case-control study. Public Health Nutr 1:139–145. Rohan TE, Jain M, Miller AB. 1998b. A case-cohort study of diet and risk of benign proliferative epithelial disorders of the breast (Canada). Cancer Causes Control 9:19–27. Rohan TE, L’Abbé KA, Cook MG. 1992. Oral contraceptives and risk of benign proliferative epithelial disorders of the breast. Int J Cancer 50:891–894. Ronnet BM, Sherman ME, Kerman RJ. 2001. Endometrium. In: Henson DE, Albores-Saavedra, editors. Pathology of Incipient Neoplasia. New York: Oxford University Press, pp. 401–418. Salasche SJ. 2000. Epidemiology of actinic keratoses and squamous cell carcinoma. J Am Acad Dermatol 42:4–7. Sandison AT. 1962. An Autopsy Study of the Adult Human Breast. National Cancer Institute Monograph 8. Washington, DC: US Government Printing Office. Sandler RS, Lyles CM, Peipins LA, et al. 1993. Diet and risk of colorectal adenomas: macronutrients, cholesterol, and fiber. J Natl Cancer Inst 85:884–891. Sandler RS, Pritchard ML, Bangdiwala SI. 1995. Physical activity and the risk of colorectal adenomas. Epidemiology 6:602–606. Sankaranarayanan R, Mathew B, Jacob BJ, et al. 2000. Early findings from a community-based cluster-randomized controlled oral cancer screening trial in Kerala, India. Cancer 88:664–673. Sankaranarayanan S, Somanathan T. 2002. Upper aero-digestive tract. In: Franco EL, Rohan TE, editors. Cancer Precursors: Epidemiology, Detection, and Prevention. New York: Springer-Verlag, pp. 77–95. Sarnelli R, Squartini F. 1991. Fibrocystic condition and “at risk” lesions in asymptomatic breasts: a morphologic study of postmenopausal women. Clin Exp Obstet Gynecol 18:271–279. Sasano N, Tateno H, Stemmerman GN. 1978. Volume and hyperplastic lesions of breasts of Japanese women in Hawaii and Japan. Prev Med 7:196– 204. Sawaya GF, Kerlikowske K, Lee NC, Gildengorin G, Washington AE. 2000. Frequency of cervical smear abnormalities within 3 years of normal cytology. Obstet Gynecol 96:219–223. Saxen E. 1979. Histopathology in cancer epidemiology. In: Sommers SC, Rosen PR, editors. Pathology Annual, Pt. 1. New York: AppletonCentury-Crofts, pp. 203–217. Schatzkin A. 2002. Intermediate markers in cancer research: theoretical and practical issues in the use of surrogate endpoints. In: Franco EL, Rohan TE, editors. Cancer Precursors: Epidemiology, Detection, and Prevention. New York: Springer-Verlag, pp. 46–59. Schatzkin A, Freedman LS, Schiffman MH, Dawsey SM. 1990. Validation of intermediate end points in cancer research. J Natl Cancer Inst 82:1746–1752. Schnitt SJ. 2003. Benign breast disease and breast cancer risk. Am J Surg Pathol 27:836–841. Schwartz RA. 1997. The actinic keratosis: a perspective and update. Dermatol Surg 23:1009–1019. Schiffman MH, Bauer HM, Hoover RN, et al. 1993. Epidemiologic evidence showing that human papillomavirus infection causes most cervical intraepithelial neoplasia. J Natl Cancer Inst 85:958–964. Schlecht NF, Trevisan A, Duarte-Franco E, et al. 2003. Viral load as a predictor of the risk of cervical intraepithelial neoplasia. Int J Cancer 103:519–524. Schottenfeld D. 1995. Epidemiology of endometrial neoplasia. J Cell Biochem. 23(Suppl):151–159. Scully C, Sudbo J, Speight PM. 2003. Progress in determining the malignant potential of oral lesions. J Oral Pathol Med 32:251–256. Senesse P, Boutron-Ruault MC, Faivre J, et al. 2002. Foods as risk factors for colorectal adenomas: a case-control study in Burgundy (France). Nutr Cancer 44:7–15. Shafer WG, Waldron CA. 1975. Erythroplakia of the oral cavity. Cancer 36:1021–1028. Shaheen N, Ransohoff DF. 2002. Gastroesophageal reflux, Barrett’s esophagus, and esophageal cancer: scientific review. JAMA 287:1972–1981. Sherman ME. 2000. Theories of endometrial carcinogenesis: a multidisciplinary approach. Mod Pathol 13:295–308. Shinchi K, Kono S, Honjo S, et al. 1994. Obesity and adenomatous polyps of the sigmoid colon. Jpn J Cancer Res 85:479–484.

45

Sigurdsson K. 1999. Trends in cervical intra-epithelial neoplasia in Iceland through 1995: evaluation of targeted age groups and screening intervals. Acta Obstet Gynecol Scand 78:486–492. Silverman S. 2003. Leukoplakia and erythroplasia. In: Silverman S, editor. Oral Cancer, 5th ed. Hamilton, Ontario: BC Decker, pp. 29–47. Silverman S Jr, Bhargava K, Smith LW, et al. 1976. Malignant transformation and natural history of oral leukoplakia in 57,518 industrial workers of Gujarat, India. Cancer 38:1790–1795. Slaughter DP, Southwick HW, Smejkal W. 1953. “Field cancerization” in oral stratified squamous epithelium: clinical implications of multicentric origin. Cancer 6:963–968. Sloss PT, Bennett WA, Clagett OT. 1957. Incidence in normal breasts of features associated with chronic cystic mastitis. Am J Pathol 33:1181–1191. Smith AL, Hung J, Walker L, et al. 1996. Extensive areas of aneuploidy are present in the respiratory epithelium of lung cancer patients. Br J Cancer 73:203–209. Smith-Warner SA, Elmer PJ, Fosdick L, et al. 2002. Fruits, vegetables, and adenomatous polyps. Am J 155:1104–1113. Soini I, Aine R, Lauslahti K, Hakama M. 1981. Independent risk factors for benign and malignant breast lesions. Am J Epidemiol 114:507–514. Solomon D. 1989. The 1988 Bethesda system for reporting cervical/vaginal cytologic diagnoses: developed and approved at the National Cancer Institute Workshop, Bethesda, Maryland, USA, December, 12–13, 1988. J Clin Cytol Cytopathol 33:567–574. Solomon D, Davey D, Kurman R, et al. 2002. The 2001 Bethesda System: terminology for reporting results of cervical cytology. JAMA 287: 2114–2119. Solomon D, Frable WJ, Vooijs GP, et al. 1998. ASCUS and AGUS criteria; International Academy of Cytology Task Force summary. Diagnostic cytology towards the 21st century: an international expert conference and tutorial. Acta Cytol 42:16–24. Sonnenberg A, El-Serag HB. 1999. Clinical epidemiology and natural history of gastroesophageal reflux disease. Yale J Biol Med 72:81–92. Sozzi G, Miozzo M, Donghi R, et al. 1992. Deletions of 17p and p53 mutations in preneoplastic lesions of the lung. Cancer Res 52:6079–6082. Sozzi G, Miozzo M, Pastorino U, et al. 1995. Genetic evidence for an independent origin of multiple preneoplastic and neoplastic lung lesions. Cancer Res 55:135–140. Sporn MB. 1976. Approaches to prevention of epithelial cancer during the preneoplastic period. Cancer Res 36:2699–2702. Srivastava S, Henson DE, Gazdar A. 1999. Molecular Pathology of Early Cancer. Amsterdam: IOS Press. Steinbach G, Lynch PM, Phillips RKS, et al. 2000. The effect of celecoxib, a cyclooxygenase-2 inhibitor, in familial adenomatous polyposis. N Engl J Med 342:1946–1952. Stemmermann GN, Yatani R. 1973. Diverticulosis and polyps of the large intestine: a necropsy study of Hawaii Japanese. Cancer 31:1260–1270. Stout AP. 1932. Human cancer. Philadelphia: Lea & Febiger, p. 18. Stryker SJ, Wolff BG, Culp CE, et al. 1987. Natural history of untreated colonic polyps. Gastroenterology 93:1009–1013. Sturgeon SR, Sherman ME, Kurman RJ, et al. 1998. Analysis of histopathological features of endometrioid uterine carcinomas and epidemiologic risk factors. Cancer Epidemiol Biomarkers Prev 7:231–235. Sudbø J, Bryne M, Johannessen AC, et al. 2001. Comparison of histological grading and large-scale genomic status (DNA ploidy) as prognostic tools in oral dysplasia. J Pathol 194:303–310. Sugio K, Molberg K, Albores-Saavedra J, et al.1997. K-ras mutations and allelic loss at 5q and 18q in the development of human pancreatic cancers. Int J Pancreatol 21:205–217. Sundaresan V, Heppell-Parton A, Coleman N, et al. 1995. Somatic genetic changes in lung cancer and precancerous lesions Ann Oncol 6S:27– 32. Takayama T, Katsuki S, Takahashi Y, et al. 1998. Aberrant crypt foci of the colon as precursors of adenoma and cancer. N Engl J Med 339: 1277–1284. Tase T, Okagaki T, Clark BA, et al. 1989. Human papillomavirus DNA in adenocarcinoma in situ, microinvasive adenocarcinoma of the uterine cervix, and coexisting cervical squamous intraepithelial neoplasia. Int J Gynecol Pathol 8:8–17. Terry MB, Neugut AI, Bostick RM, et al. 2002. Risk factors for advanced colorectal adenomas: a pooled analysis. Cancer Epidemiol Biomarkers Prev 11:622–629. The PEPI Trial. 1996. Effects of hormone replacement therapy on endometrial histology in postmenopausal women: the Postmenopausal Estrogen/Progestin Interventions (PEPI) trial. JAMA 275:370–375. Tiemersma EW, Wark PA, Ocke MC, et al. 2003. Alcohol consumption, alcohol dehydrogenase 3 polymorphism, and colorectal adenomas. Cancer Epidemiol Biomarkers Prev 12:419–425.

46

PART I: BASIC CONCEPTS

Tortolero-Luna G, Franco EL. 2004. Epidemiology of cervical, vulvar, and vaginal cancer: unresolved issues and future research. In: Gershenson DM, Gore M, Quinn MJ, Thomas G, editors. Gynecologic cancer: Controversies in management. New York: Elsevier Science and Churchill Livingstone, pp. 3–30. Ulrich CM, Bigler J, Whitton JA, et al. 2001. Epoxide hydrolase Tyr113His polymorphism is associated with elevated risk of colorectal polyps in the presence of smoking and high meat intake. Cancer Epidemiol Biomarkers Prev 10:875–882. Unni KK, Dahlin DC. 1979. Premalignant tumors and conditions of bone. Am J Surg Pathol 3:47–60. Utagawa ML, Pereira SM, Cavaliere MJ, Maeda MY, Shih LW, Shirata NK. 1998. Cervical intraepithelial neoplasia in adolescents: study of cytological findings between 1987 and 1995 in Sao Paulo State-Brazil. Arch Gynecol Obstet 262:59–64. Vaezi MF, Falk GW, Peek RM et al. 2000. CagA-positive strains of Helicobacter pylori may protect against Barrett’s esophagus. Am J Gastroenterol 95:2206–2211. van der Waal I, Schepman KP, van der Meij EH, Smeele LE. 1997. Oral leukoplakia: a clinicopathological review. Oral Oncol 33:291–301. van Sandick JW, van Lanschot JB, van Felius L, et al. 2002. Intestinal metaplasia of the esophagus or esophagogastric junction: evidence of distinct clinical, pathologic, and histochemical staining features. Am J Clin Pathol 117:117–125. Vaughan T. 2002. Esophagus. In: Franco EL, Rohan TE, editors. Cancer Precursors: Epidemiology, Detection, and Prevention. New York: SpringerVerlag, pp. 96–116. Vicari JJ, Peek RM, Falk GW, et al. 1998. The seroprevalence of cagA positive Helicobacter pylori strains in the spectrum of gastroesophagaeal reflux disease. Gastroenterology 115:50–57. Villa LL, Sichero L, Rahal P, et al. 2000. Molecular variants of human papillomavirus types 16 and 18 preferentially associated with cervical neoplasia. J Gen Virol 81:2959–2968. Vogelstein B, Kinzler KW. 1993. The multistep nature of cancer. Trends Genet 9:138–141. Voskuil DW, Kampman E, Grubben MJAL, et al. 2002. Meat consumption and meat preparation in relation to colorectal adenomas among sporadic and HNPCC family patients in The Netherlands. Eur J Cancer 38:2300– 2308. Voutilainen M, Sipponen P, Mecklin JP, et al. 2000. Gastroesophageal reflux disease: prevalence, clinical, endoscopic, and histopathological findings in 1,128 consecutive patients referred for endoscopy due to dyspeptic and reflux symptoms. Digestion 61:6–13. Vuento MH, Pirhonen JP, Makinen JI, et al. 1999. Screening for endometrial cancer in asymptomatic postmenopausal women with conventional and colour Doppler sonography. Br J Obstet Gynaecol 106:14–20. Wacholder S. 2003. Statistical issues in the design and analysis of studies of human papillomavirus and cervical neoplasia. J Natl Cancer Inst Monogr 31:125–130. Wajed SA, Streets CG, Bremmer CG, et al. 2001. Elevated body mass disrupts the barrier to gastroesophageal reflux. Arch Surg 136:1014–1019. Waldron CA, Shafer WG. 1975. Leukoplakia revisited: a clinicopathologic study of 3256 oral leukoplakias. Cancer 36:1386–1392.

Wang DY, Fentiman IS. 1985. Epidemiology and endocrinology of benign breast disease. Breast Cancer Res Treat 6:5–36. Wang SS, Hildesheim A. 2003. Viral and host factors in human papillomavirus persistence and progression. J Natl Cancer Inst Monogr 31:35–40. Wang TC, Cardiff RD, Zukerberg L, et al. 1994. Mammary hyperplasia and carcinoma in MMTV-cyclin D1 transgenic mice. Nature 369:669–671. Weber AM, Belinson JL, Piedmonte MR. 1999. Risk factors for endometrial hyperplasia and cancer among women with abnormal bleeding. Obstet Gynecol 93:594–598. Weiderpass E, Baron JA, Adami HO, et al. 1999. Low-potency oestrogen and risk of endometrial cancer: a case-control study. Lancet 353:1824–1828. Weir HK, Sloan M, Kreiger N. 1994. The relationship between cigarette smoking and the risk of endometrial neoplasms. Int J Epidemiol 23:261–266. Westra WH, Sidransky D. 1998. Phenotypic and genotypic disparity in premalignant lesions: of calm water and crocodiles. J Natl Cancer Inst 90:1500–1501. Williams AR, Balasooriya BA, Day DW. 1982. Polyps and cancer of the large bowel: a necropsy study in Liverpool. Gut 23:835–842. Winawer SJ, Zauber AG, Ho MN, et al. 1993. Prevention of colorectal cancer by colonoscopic polypectomy. N Engl J Med 329:1977–1981. Wistuba II, Behrens C, Milchgrub S, et al. 1999. Sequential molecular abnormalities are involved in the multistage development of squamous cell lung carcinoma. Oncogene 18:643–650. Woodruff JD, Pickar JH. 1994. Incidence of endometrial hyperplasia in postmenopausal women taking conjugated estrogens (Premarin) with medroxyprogesterone acetate or conjugated estrogens alone. The Menopause Study Group. Am J Obstet Gynecol 170:1213–1223. Wright TC Jr, Cox JT, Massad LS, Twiggs LB, Wilkinson EJ. 2002. 2001 Consensus guidelines for the management of women with cervical cytological abnormalities. JAMA 287:2120–2129. Xi LF, Koutsky LA, Galloway DA, et al. 1997. Genomic variation of human papillomavirus type 16 and risk for high grade cervical intraepithelial neoplasia. J Natl Cancer Inst 89:796–802. Ylitalo N, Josefsson A, Melbye M, et al. 2000. A prospective study showing long-term infection with human papillomavirus 16 before the development of cervical carcinoma in situ. Cancer Res 60:6027–6032. Yood MU, Oliveria S, Boyer G, et al. 2003. Colon polyp recurrence in a managed care population. Arch Intern Med 2003;163:422–426. Yoon H, Benamouzig R, Little J, et al. 2000. Systematic review of epidemiological studies on meat, dairy products and egg consumption and risk of colorectal adenomas. Eur J Cancer Prev 9:151–164. Zain RB, Ikeda N, Razak IA, et al. 1997. A national epidemiological survey of oral mucosal lesions in Malaysia. Community Dent Oral Epidemiol 25:377–383. Zhang L, Poh CF, Lam WL, et al. 2001. Impact of localized treatment in reducing risk of progression of low-grade oral dysplasia: molecular evidence of incomplete resection. Oral Oncol 37:505–512. Zheng S, Liu X-Y, Ding K-F, et al. 2002. Reduction of the incidence and mortality of rectal cancer by polypectomy: a prospective cohort study in Haining county. World J Gastroenterol8:488–492. Zur Hausen H. 2001. Cervical carcinoma and human papillomavirus: on the road to preventing a major human cancer. J Natl Cancer Inst 93:252–253.

4

Molecular and Genetic Events in Neoplastic Transformation AYSE E. ERSON AND ELIZABETH M. PETTY

C

arcinogenesis is a multistep process involving the progressive accumulation of genetic and epigenetic alterations that ultimately transform normal cells into neoplastic cells. The word neoplasia literally means “new growth” (Fig. 4–1). New growths may be benign or malignant. Malignant growths have more aggressive characteristics than benign neoplasms. Common characteristics of malignant cells include self-sufficiency for growth stimuli, insensitivity to antigrowth signals, limitless replication potential, resistance to cell death, new and sustained angiogenesis, the ability to invade tissues, and a propensity to metastasize. Molecular changes that drive malignant progression most often involve mutations in genes that regulate cell proliferation. Genes that play positive roles in growth promotion are called protooncogenes, and genes that act as brakes to keep cell proliferation under control are called tumor suppressor genes. If genetic alterations are considered as the ON/OFF switches of carcinogenesis, activation of proto-oncogenes would be perpetual ON switches, and loss of function of tumor suppressors would be broken OFF switches. In general, mutations of proto-oncogenes are considered to be dominant gain of function changes because alteration of just one allele is enough to drive the transformation process. Conversely, mutations of tumor suppressors are considered to be recessive loss of function changes at the cellular level as both alleles have to be altered or lost to accelerate carcinogenesis. Alterations in both proto-oncogenes and tumor suppressor genes work together in a cell to drive malignant progression. The purpose of this chapter is to illustrate some of the basic molecular and genetic mechanisms that underlie the development of cancer. It should be recognized that our understanding of these complex mechanisms continues to evolve as new research in cancer genetics and biology emerges. According to a recent study, 291 cancer-related genes have been reported in the literature, excluding hundreds of other putative genes whose functions and significance during tumorigenesis are not yet well established (Futreal et al., 2004). Table 4–1 lists databases where information about specific cancer genes can be found, including those discussed in this chapter. It is a serious challenge to characterize the multidimensional functional roles of the increasing number of cancer-related genes. Although some important insights into the molecular pathogenesis of cancer are summarized in this chapter, it is important to keep in mind that the overview provided here is only a glimpse into the molecular complexity observed in malignant progression. The chapter is organized into four main sections. In the first section we examine perspectives on the roles of oncogenes, tumor suppressor genes, and genomic instability genes in the pathogenesis of malignancy. The second section presents the genetic and epigenetic alterations that disrupt normal function of cancer genes. In the third section we discuss the alteration of cell cycle progression and checkpoint mechanisms relative to their possible contribution to carcinogenesis. The last section summarizes the phenotypic characteristics of cancer cells, including proliferation, apoptosis, angiogenesis, and metastasis as related to specific molecular events.

TYPES OF CANCER GENES Proto-oncogenes Discovery of the oncogene hypothesis dates back to early 1900s when Ellermann observed the transmissibility of avian leukemia in

birds and Rous subsequently found that sarcoma in chickens was transplantable from one animal to another using a cell-free filtrate of the tumors (Rous, 1911). These studies suggested the presence of transmissible tumor-inducing factors. During the early 1900s, Boveri’s seminal observations in sea urchin embryos suggested that cancer cells contain defective chromosomes that can be passed on to progeny cells. He insightfully postulated that an increased number of growth “stimulatory chromosomes” and the loss of “growth inhibitory chromosomes” led to the unlimited growth of malignant tumor cells (Boveri, 1929; Balmain, 2001). Studies substantiating these findings in mammals started during the 1930s spearheaded by Shope, who studied cell free transmission of tumors in rabbits (Shope, 1933). The later discovery of murine leukemia viruses by Gross led to the important finding by Temin and Rubin that infection of cultured chicken fibroblasts with the Rous sarcoma virus (RSV) caused neoplastic transformation of cells (Rubin and Temin, 1958; Temin and Rubin, 1958). Later experiments revealed that the oncogenic portion of the RSV genome was v-SRC (the viral src gene). Taken together, these historically important findings suggested a transmissible mechanism for tumorigenesis. It was not until the early 1980s, however, that the molecular explanation underlying these observations became apparent, when Bishop and Varmus demonstrated that labeled v-SRC could hybridize to its complementary counterpart, c-SRC, in the normal avian genome. This observation led to their novel hypothesis that cancer-causing genes (oncogenes) carried by tumor-forming viruses have normal counterparts (protooncogenes) in the host vertebrate genomes. According to their hypothesis, retroviruses “captured” the cancer-causing genes from the host genomes, rather than the earlier hypothesis that viruses “introduced” cancer genes into host genomes (Bishop, 1989; Varmus, 1989). Specifically, evidence suggested that during infection an RNA virus could integrate its genome close to a proto-oncogene in the host genome, capture this sequence, and integrate it into its own genome during the viral replication cycle. They postulated that eventually proto-oncogenes would become mutated through many rounds of viral infection. Interestingly, RNA tumor viruses are common in some animals, such as chickens, mice, and cats, but are fortunately rare in humans. However, many of the homologous genes that are captured by viruses in other animals are also altered in nonvirus-induced human cancers. For example, c-MYC is the mammalian cellular homologue of the viral oncogene (v-MYC) of the avian myelocytomatosis retrovirus. Elevated or deregulated expression of the proto-oncogene c-MYC has been detected in a variety of human cancers such as breast, colon, cervical, and small-cell lung carcinomas, leukemias, osteosarcomas, melanomas, and glioblastomas. In more recent years, given the advances in molecular biology tools and methods, many proto-oncogenes that encode proteins involved in cell growth and regulation have been identified. Mutations or altered expression patterns of proto-oncogenes cause them to act as oncogenes and drive the pathogenesis of virtually all human cancers. Activated oncogenes such as c-MYC may play roles in cancer cells through stimulating growth and proliferation, disturbing normal cell cycle regulation, altering signal transduction pathways, or up-regulating transcription of growth-related genes or antiapoptotic mechanisms (Fig. 4–2).

47

48

PART I: BASIC CONCEPTS NORMAL CELLS

Genetic and epigenetic alterations

NEOPLASTIC CELLS

BENIGN TUMORS

Progressive accumulation of genetic and epigenetic alterations CANCER CELLS

Increased proliferation Limitless replication potential Decreased apoptosis Invasion Angiogenesis Metastasis

Figure 4–1. Carcinogenesis is a multistep process. As a result of genetic and/or epigenetic alterations, normal cells acquire new characteristics and become neoplastic. Neoplastic cells can be benign or malignant. Benign tumors have less aggressive characteristics than malignant tumors. They remain localized and do not spread to other tissues. Malignant tumors, on the other hand, acquire the ability to invade surrounding connective tissue, induce blood vessel growth (angiogenesis), and travel (metastasize) to distant organs. Cancer cells accumulate additional genetic and/or epigenetic alterations that drive tumor progression. These alterations may result in deregulated growth, replication, and programmed cell death (apoptosis)—pathways that ultimately lead to a progressive malignant phenotype.

Tumor Suppressor Genes Early somatic cell hybrid experiments showed that fusion of tumor cells with normal somatic cells could reverse the transformed phenotype back to normal (Harris et al., 1969). This correction of a malignant phenotype was due to the presence of certain chromosomes added during cell fusion that were originally lost in the tumor cells before cell fusion. This observation was later supported by experiments to correct the malignant phenotype of tumor cells upon transfer of specific normal chromosomes, suggesting that some genes on these specific chromosomes may be responsible for preventing or reversing the tumor phenotype (Harris et al., 1969; Stanbridge, 1976). Since then, many tumor suppressor genes have been identified. Tumor suppressor genes provide, in a sense, cancer-preventive effects in cells. The presence of a single copy of a normal tumor suppressor allele is generally sufficient for that growth-controlling purpose. Therefore, tumor suppressors usually act as recessive genes at the cellular level, as loss of both copies is required for the loss of function of a tumor suppressor gene. This important paradigm was first described in Knudson’s epidemiologic studies of an embryonal tumor model, retinoblastoma, which became widely known as Knudson’s “two-hit” hypothesis (Knudson et al., 1975) (Fig. 4–3). Retinoblastoma (RB) is a rare childhood cancer thought to arise in embryonic retinal epithelium. Knudson developed a mathematical system to examine the incidence and the age of onset of childhood retinoblastoma. The high incidence and early age of onset of familial retinoblastoma in children suggested that individuals with a germline mutation were highly susceptible to tumor formation after an additional somatic mutation. Specifically, individuals with a family history

of retinoblastoma would have inherited an altered RB allele from the affected parent and have this mutation in all of their cells. Knudson noted that the chance of getting a second mutation, or “second hit,” in one of these cells early in life, which would initiate tumor development, was extremely high. In contrast, with the sporadic forms of retinoblastoma, two independent RB mutations would have to occur in the same cell for the tumor to develop. Therefore he postulated that, among familial retinoblastomas, the sporadic form was much more unlikely to occur, resulting more often in single tumors at a later age of onset than in multiple tumors with an early age of onset. Based on Knudson’s and others’ observations, characteristics of “classic” tumor suppressors can be summarized as cellular recessive genes that undergo biallelic inactivation, giving rise to tumor formation. RB remains one of the most classic examples of a tumor suppressor gene. We now know that the RB protein plays many vital roles in cell cycle progression, apoptosis, and differentiation pathways (Jacks et al., 1992; Lee et al., 1992). Cell cycle regulatory proteins known as D-type cyclins and kinase complexes phosphorylate RB, which results in release of a transcription factor, E2F, to activate transcription of cell cycle progression genes (Fig. 4–4). RB can control both proliferation and apoptosis through repression of E2F-dependent promoters. In the absence of RB, defects or overexpression of members of these pathways, such as TP53 and MDM2, can cause hyperproliferation. It appears that many solid malignancies have a defect somewhere in this complex RB-related pathway resulting in the solid tumor phenotype and/or malignant progression.

Genomic Instability Genes Other cases of recessive germline mutations that increase the risk of developing cancer include alterations in genes involved in DNA damage-sensing pathways or DNA-repairing pathways. For example, mutations of DNA mismatch repair genes are seen in hereditary nonpolyposis colorectal cancer (HNPCC) patients. Sporadic colon cancers also often exhibit inactivation of mismatch repair genes or other genes affecting genome stability. Accumulation of errors during DNA replication due to malfunctioning of repair mechanisms may result in mutations of proto-oncogenes and tumor suppressor genes. DNA repair genes often resemble tumor suppressor genes in terms of their dominant inheritance pattern but their recessive nature at the cellular level. However, their effect is generally more indirect than classic tumor suppressors, such as RB described above, as they lead to a “mutator phenotype” in which the accumulation of additional mutations leads to genomic instability, thereby causing activation of oncogenes or loss of the classic tumor suppressor genes that drive malignant progression (Loeb et al., 1974, 2003; Lengauer et al., 1998). Neoplastic transformation results from the accumulation of a variety of alterations in cellular mechanisms through changes in genome. Many argue, however, that it is not possible to explain such accumulation of alterations with the random low mutation rate of DNA in somatic cells. The unstable genomes of cancers are thought to arise, in part, from a “mutator phenotype” through mutations in genes responsible for keeping the genome stable and successive clonal selection or growth advantage of these mutated cells (Nowell, 1976; Loeb, 1991). Genes involved in sensing and repairing DNA damage, checkpoint responses, and chromosome segregation can be grouped as the “caretaker” genes that are frequently altered in cancers (Levitt and Hickson, 2002).

Table 4–1. Databases at which Information About Cancer Genes Can Be Found Cancer.gov: http://www.cancer.gov/search/ Atlas of Genetics and Cytogenetics in Oncology and Haematology: http://www.infobiogen.fr/services/chromcancer/ OMIM (Online Mendelian Inheritance in Man): http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=OMIM The Cancer Gene Anatomy Project: http://cgap.nci.nih.gov/ GeneCards: http://bioinformatics.weizmann.ac.il/cards/

49

Molecular and Genetic Events in Neoplastic Transformation p27 MYC MAX

Cyclin D2 CDK4

p27 G1/S Progression Cyclin E/CDK2

MYC MIZ-1 MAX SP1

Cyclin E/CDK2

p15 p21

Cell Cyle Arrest

MYC HAT MAX TRRAP Histone Acetylation (

Gene Activation

)

MYC

Apoptosis

BAX

X

Deletion

Chromosome Loss

Figure 4–2. The proto-oncogene c-MYC has diverse roles during the cell cycle. c-MYC is a transcription factor that has roles in cell cycle progression, cell cycle arrest, transcriptional regulation and apoptosis. c-MYC forms heterodimeric complexes with MAX to transcriptionally activate target genes important in cell cycle progression such as the cyclin D2 gene, CCND2, and CDK4. Expression of cyclin D2, CDK4 sequesters the CDK inhibitor p27 from the cyclin E-CDK2 complex, which then drives G1/S progression. P27 is further degraded by the help of other myc target genes (CUL1, CKS) (Coller et al., 2000). As well as promoting G1/S progression, the c-MYC-MAX complex can induce cell cycle arrest by preventing the transactivation of CDK inhibitors (p15, p21) by binding to the transcription factors MIZ1 and SP1 (Wanzel et al., 2003). The c-MYC-MAX complexes can regulate gene transcription by chromatin remodeling. The amino-terminal of MYC binds to a protein known as TRRAP (transformation/transcription domain associated protein), which complexes with a histone acetyltransferase (HAT) for H4 acetylation of histones to alter the chromatin structure to allow ready access for the c-MYC-MAX complex. The c-MYC is involved in apoptosis regulation possibly through activation of BAX, a proapoptotic molecule that triggers cytochrome c release from mitochondria. The c-MYC also indirectly regulates TP53 through p14ARF (in humans) and p19ARF (in mice), leading to transcription of BAX. BAX is lost during some cases of colon tumorigenesis and possibly leads to the suppression of apoptosis. The presence of specific mitotic signals such as PKB/Akt kinase or RAS activation, may further suppress c-MYC-induced apoptosis (Source: Kauffmann-Zeh et al., 1997).

In solid tumors, genomic instability often manifests at the single nucleotide level as microsatellite instability (MIN) and at the chromosomal level as chromosomal instability (CIN) (Markowitz et al., 1995; Rampino et al., 1997). In general, MIN tumors exhibit a pattern of diploid karyotypes; however, they have increased alterations at the DNA sequence level. Problems during replication of microsatellite sequences between or within genes can cause expansion of these short repeats that might lead to inactivation of tumor suppressor genes. On the other hand, tumors with significant CIN generally have

X

X

RB P

P First Hit (e.g., inherited germline mutation)

P

XX

Mutation

Epigenetic Alteration

X Second Hit Figure 4–3. Inactivation of a classic tumor suppressor gene. Knudson, a cancer geneticist and a pediatrician, developed a mathematical model to study the occurrence of retinoblastoma in families. He suggested that children who have a family history of the disease are more prone to develop tumors. He postulated that all they needed was an additional somatic mutation to initiate retinoblastoma development, as they already had inherited one mutant copy of the gene in all of their cells. Thus, he noted that the chance of getting a second mutation (second hit) in the other allele of these cells early in life was high compared to individuals with sporadic forms of the disease, where two independent RB mutations would have to occur in the same cell. The second hit inactivating a tumor suppressor gene may be due to a deletion, chromosome loss, mutation, or epigenetic alteration.

E2F RB

Transcription OFF

G1 phase

Activation of S-phase cell cycle genes

E2F

Transcription ON

S Phase Cell Cycle Progression

Figure 4–4. The tumor suppressor RB plays an important role in cell cycle progression. RB is unphosphorylated or hypophosphorylated during the G1 (gap 1) phase of the cell cycle. In its hypophosphorylated state, RB can bind to E2F transcriptional regulatory proteins and represses transcriptional activation of genes involved in DNA synthesis. Cyclin–CDK complexes phosphorylated RB and cause its release from E2F, resulting in the availability of E2F to activate transcription of genes required for the initiation of the S (synthesis) phase of the cell cycle.

50

PART I: BASIC CONCEPTS

abnormal karyotypes with significant loss or gain of specific chromosomal regions or entire chromosomes. CIN is thought to be due to defects of DNA replication (Loeb and Loeb, 2000; Schar, 2001), rejoining of DNA double-strand breaks (Khanna and Jackson, 2001), and checkpoint mechanisms that monitor cell cycle progression to guarantee safety of the genome and faithful segregation of chromosomes into daughter cells (Cahill et al., 1998). Microsatellite and chromosomal instability can alter gene dosage or normal protein expression, leading to the cancer-promoting mutator phenotype. Large-scale studies to search for genome-wide alterations suggest that a strikingly large number of genes (as many as 30% of their genome) may have altered expression in cancer cells (Iwabuchi et al., 1995; Liotta and Petricoin, 2000). However, it should be emphasized that many of these changes may not be causative of tumorigenesis but rather may be artifacts of unstable genomes or reflections of increased proliferation. Thus, a major challenge is how to identify and characterize the relevant genes to understand the functional consequences of their potential contributions to carcinogenesis.

GENETIC ALTERATIONS CAUSING CANCER As noted at the start of this chapter, cancer is a multistep process whereby the progressive accumulation of genetic and epigenetic alterations leads to the transformed phenotype. These alterations may affect transmembrane signaling, cell adhesion, DNA repair, cell cycle control, genome stability, apoptosis, and angiogenesis mechanisms. There is an increasing body of evidence that identifies alterations in many novel proto-oncogenes and tumor suppressor genes involved in these pathways contributing to carcinogenesis. Various molecular alterations in these genes may occur to accelerate malignant progression.

Point Mutations Activation of oncogenes is a frequent mechanism of tumorigenesis due to point mutations, translocations, and gene amplifications. The first identified human oncogene activated by a point mutation was RAS (Tabin et al., 1982; Barbacid, 1987). RAS proteins are located at the inner plasma membrane, where they transmit extracellular mitogenic signals to a set of complex cellular signaling and proliferation pathways (Boguski and McCormick, 1993). Approximately 20% of all solid tumors have an activating mutation in one of a handful of various RAS genes (Bos, 1989). Upon an extracellular stimulus in normal cells, RAS becomes active by binding to guanosine 5¢-triphosphate (GTP), which in turn interacts with a complex network of downstream effectors. Guanine nucleotide exchange factors facilitate the return of RAS back to its inactive form coupled with guanosine 5¢-diphosphate (GDP). Oncogenic forms of RAS, however, often become resistant to GTP hydrolysis. Thus RAS becomes continuously active and sends proliferative signals to the cell even in the absence of any external mitotic stimulus. Point mutations in tumor suppressor genes may result in functional loss of activity when mutations occur in both alleles or when one allele is mutated and the other is lost.

Gene Amplification Another common mechanism of oncogene activation is through gene amplification. The term gene amplification refers to having an increased genomic copy number. Gene amplification may result in overexpression of one or more target genes in an amplicon, a region of genomic amplification, and confer a selective growth advantage to the cells. Such alterations in mammalian cells have been shown to contribute to tumorigenesis. Cytogenetic analysis of metaphase spreads of chromosomes in cancer cells reveal distinct chromosomal abnormalities associated with gene amplifications. These abnormalities manifest as double minutes (DMs) and homogeneously staining chromosomal regions (HSRs). DMs are spherical, small chromosome-like structures lacking cen-

tromeres, and they may contain circular DNA in chromatin form. HSRs are distinguished by the continuous homogeneous intensity of trypsin-giemsa staining patterns rather than normal chromosomes exhibiting alternating dark and light bands. Studies in model systems have shed some light on the possible mechanism of gene amplification. One possible explanation is the breakage-fusion-bridge (BFB) cycles mechanism that was originally defined by McClintock (1938, 1940) (Fig. 4–5). Studies suggest that BFB cycles are the primary mechanism for gene amplification in hamster cells (Smith et al., 1990; Ma et al., 1993). As a consequence of BFB cycles, large inverted repeats and accumulation of extra gene copies have been detected.

Translocations Effective double-stranded break (DSB) repair is essential to genomic integrity because defective repair of DSBs can lead to chromosome translocations. Translocations may result in overexpression of protooncogenes if fused to a constitutively active regulatory region or expression of a fusion gene with novel properties. In mammalian systems, DSBs are most often repaired by homologous recombination (HR) or nonhomologous end joining (NHEJ) (Pierce and Jasin, 2001). HR is a conservative process that depends on a homologous DNA template for the repair of DSBs to restore the original sequence. In contrast, NHEJ does not require homology but, rather, brings together the two free DNA ends created by a break. This process may result in deletions, duplications, insertions, and inversions. Specific chromosomal translocations have been detected in many hematologic and some solid tumors. In reciprocal translocation, partner chromosomes interchange segments, resulting in two derivative chromosomes (Elliott and Jasin, 2002). Translocations seen in leukemias involving several chromosomes have been described and often result in a novel gain of function fusion protein, as in the case with the BCR-ABL fusion gene. The best known example of reciprocal translocation is the Philadelphia chromosome t(9;22)(q34;q31), seen in chronic myeloid leukemia (CML) (Nowell, 1994). The powerful promoter of the BCR gene on chromosome 22 fuses with the tyrosine kinase gene, ABL, on chromosome 9 to form a novel BCR-ABL fusion gene (Tkachuk et al., 1990; Dobrovic et al., 1991; Shah et al., 1991). Identification and characterization of specific translocations have important clinical benefits for both diagnostic and prognostic purposes. For example, the Philadelphia chromosome is a reliable cytogenetic marker of CML. Characterization of this translocation led to the discovery of imatinib mesylate (Gleevec), a selective kinase inhibitor that is now used effectively for the treatment of CML (for a review on leukemias see Greaves and Wiemels, 2003).

Loss of Heterozygosity Loss of heterozygosity (LOH) can manifest at the molecular level as deletion of part of a gene or at the cytologic level as loss of a whole chromosome due to translocation, mitotic recombination, chromosome breakage and loss, telomeric end-to-end fusion, or chromosome fusion (Fig. 4–5). Such chromosomal abnormalities seem to be the main mechanism for introducing the “second hit” to a locus that already harbors a first genetic mutation, often a point mutation, leading to the observation of LOH at a specific locus. Deletion of specific chromosomal bands or LOH at a specific gene allele may indicate loss of function of the genes involved in cell proliferation control or differentiation pathways. Use of restriction fragment length polymorphisms (RFLPs) demonstrated LOH at chromosome 13q14, which harbors the RB gene, and the potential utility of using LOH studies to pinpoint the location of tumor suppressor genes (Cavenee et al., 1985). Similar studies also confirmed Knudson’s “two hit” hypothesis by revealing that the normal allele inherited from the unaffected parents is lost in retinoblastoma patients. Large-scale genome analyses of solid tumors have revealed numerous LOH regions that may harbor genes that play a role in carcinogenesis. For example, LOH at chromosomal band 17p13, which harbors the TP53 gene, and on the short arm of

51

Molecular and Genetic Events in Neoplastic Transformation

1 2 3 4 5

1 2 3 4 5

1 2 3

1 2 3

Double strand break 4 5

1 2 3

123321

1 2 3

Fusion between chromatids

Anaphase Bridges form as chromatids separate

123

3

21

A break may result during telophase

4 5

1 2 3 3

1 2

Segregation of chromosomes with addition and deletion of genetic material into daughter cells

Figure 4–5. Breakage-fusion-bridge (BFB) mechanism of gene amplification. A double-strand break can cause loss of a distal segment, resulting in fused chromatids and dicentric chromosomes. During anaphase, a bridge structure forms that may lead to breakage when the two centrosomes are pulled in opposite directions later in telophase. The BFB cycle

is then repeated in the next cell cycle. BFB cycles have been reported in solid tumors, resulting in dicentric chromosomes, ring chromosomes, or telomeric associations (or a combination). Chromosomal fragile sites have also been linked to BFB cycles.

chromosome 5, which harbors the adenomatous polyposis coli (APC) gene in familial adenomatous polyposis patients, are commonly detected in a variety of cancers.

causes aberrant expression of genes important for cell cycle regulation, which drives unchecked proliferation and malignant progression.

Epigenetic Alterations in Cancer Epigenetics is the study of gene expression differences that do not directly involve DNA nucleotide modifications. Instead, major steps of epigenetic regulation involve modification of histones, DNA methylation, genomic imprinting (differential expression of maternal and paternal genes), and X inactivation (repression of one of the two X chromosomes in the somatic cells of females as a method of dosage compensation). A significant distinction between epigenetic and genetic alterations is that epigenetic changes have been considered potentially more easily reversible with therapeutic applications, giving rise to promising treatment considerations. Methylation of cytosine residues in CpG islands of gene promoters causes strong transcriptional silencing. Initially, methylation was thought to regulate tissuespecific expression during mammalian development (Riggs, 1975). However, subsequent studies showed that in the absence of a DNA methylase, such as Dnma1, expression of tissue-specific genes was not altered, but lack of methylation caused biallelic expression of imprinted genes and activation of internal retroviral transcription (Bestor, 1999). Thus, methylation was shown to have a role in genomic imprinting and X inactivation and to be a host mechanism against transposons in the genome. Aberrant methylation has been widely studied in carcinogenesis and was hypothesized to play a role in regulating expression of cancer genes. Hypermethylation of CpG islands has been detected in the promoters of tumor suppressor genes to silence transcription. This hypermethylation serves as an alternative mechanism for inactivating normal gene expression. For example, aberrant methylation of normally unmethylated 5¢ CpG islands has been detected for the human adenomatous polyposis coli gene (APC) in early-onset familial adenomatous polyposis, the breast cancer susceptibility gene (BRCA1), the Ecadherin gene (CDH1), and the p16Ink4a tumor suppressor gene (CDKN2A). In contrast, hypomethylation has also been shown to contribute to carcinogenesis (Feinberg and Vogelstein, 1983; Feinberg et al., 1988; Feinberg and Tycko, 2004). For example, experiments with homozygous Dnmt1 embryonic mouse stem cells showed that an increased rate of rearrangements of repeats by hypomethylation can activate retrotransposons or retroviral-derived elements, which in return might increase the predisposition to genomic instability (Chen et al., 1998). Another example is that CpG demethylation of the cyclin D2 (CCND2) promoter leads to overexpression of cyclin D2, an important regulator of the G1/S phase of the cell cycle, as described in the next section (Oshimo et al., 2003). An increasing body of literature reports involvement of abnormal epigenetic regulation as an important mechanism in cancer, even both hyper- and hypomethylation of various genes in the same tumors (Feinberg and Tycko, 2004). Often abnormal methylation

CELL CYCLE DEFECTS Cell Cycle The somatic or replicative mitotic cell cycle is a tightly regulated process for faithful cell duplication. Genetic information must be passed to the next generation of cells without accumulating any defects. The cell cycle is generally divided into four phases (Fig. 4–6). Cells undergo replication of their genetic material during the S (synthesis) phase and divide their copied genome and other necessary cellular components between two daughter cells during the M (mitosis) phase. The G1 and G2 phases represent active gap periods between the S and M phases, where the cell prepares itself for these phases. When there is no proper mitogenic signal or other stimulus for proliferation, normal replicative cells exit the cycle and enter a quiescent state, referred to as G0. These events occur in the normal somatic human cell cycle in temporally distinct stages over a 20- to 24-hour period. These phases are highly regulated, in part through the activation and degradation of a group of serine threonine kinases called cyclin-dependent kinases (CDKs) (Morgan, 1995). Activation of CDKs is controlled by regulatory subunits called cyclins, named for their role in the cycling cell. Various levels of control are involved in progression through distinct stages of the cell cycle where cyclin–CDK complexes become active or inactive. For example, CDKs are negatively regulated by their association with cyclin-dependent kinase inhibitors. The Ink4 family of proteins (p15, p16, p18, p19) inhibit CDK4 and CDK6 (Hirai et al., 1995). Members of the Cip/Kip family of proteins (p21, p27, p57) inhibit cyclin E–CDK2 and cyclin A–CDK2 complexes (Lee et al., 1995). Alternatively, phosphorylation provides additional control (inhibition or activation) over these regulatory proteins. As cells enter prophase, the cyclin B–CDC2 complex, sometimes referred to as the mitosispromoting factor (MPF), is transported to the nucleus, where it is kept inactive by the inhibitory phosphorylation of CDC2 on tyrosine 15 and threonine 14 by the WEE1 and MYT1 kinases (Mueller et al., 1995). Activation of the cyclin B–CDC2 complex takes place only after dual phosphatase CDC25C dephosphorylates CDC2. Following activation of the cyclin B–CDC2 complex, chromosomes condense, the nuclear envelope breaks down, and mitotic spindles form in preparation for faithful segregation of genetic material during cell division.

Cell Cycle Checkpoints Quality control mechanisms of the cell cycle, known as “checkpoints,” monitor the accuracy of cell cycle progression as well as various other internal and external conditions (ribonucleotide and oxygen levels, status of DNA replication, mitotic spindle apparatus) that affect cell proliferation. Although checkpoints are essential for guaranteeing

52

PART I: BASIC CONCEPTS

S

G2

G1

M

Prophase

Metaphase

Anaphase

Telophase

Cytokinesis

Figure 4–6. Mitotic cell cycle. The cell cycle is divided into distinct stages. Cells undergo replication of their genetic material during the S (synthesis) phase and divide their copied genome and other necessary cellular components between two daughter cells during the M (mitosis) phase. The G1 (gap 1) and G2 (gap 2) phases represent active preparation periods for the S and M phases. M phase is further divided into four distinct phases for faithful segregation of chromosomes into daughter cells: prophase, metaphase, anaphase, and telophase. During prophase, the chromatin in the nucleus begins to condense, and the nucleolus disappears. Centrioles

begin moving to opposite ends of the cell, and fibers extend from the centromeres to form mitotic spindles. During metaphase, chromosomes attach to the microtubule spindles through their kinetochores and align along the “metaphase plate.” Chromosomes attached to the spindles start separating during anaphase and move toward opposite poles of the cell. During telophase, chromosomes arrive at opposite poles of the cell, and new membranes form around the two daughter nuclei followed by cytokinesis, division of the cell into two daughter cells.

genomic stability, they are not required for cell cycle progression. In response to such adverse conditions as DNA damage or stress, checkpoints halt the cell cycle. Mutations of various cell cycle checkpoint genes are detected in cancers. Malfunctioning of checkpoints that ensure genome integrity can help explain the common pattern of genomic instability seen during carcinogenesis. Specific checkpoints regulating normal cell cycle progression are discussed below with a particular focus on how alterations in them can contribute to cancer.

DNA synthesis (Nevins, 2001). Recruitment of chromatin-modifying enzymes, including histone deacetylases and protein complexes such as SWI–SNF, further represses transcription (Harbour and Dean, 2000). Inactivation of RB by phosphorylation is a vital step for cell proliferation. Phosphorylation of RB activates the transcription of genes that are essential for progression into S phase through activation of the E2F family of transcription factors. These transcription factors, in turn, activate the cyclins and CDKs critical for cell cycle progression, specifically cyclin E, cyclin A, and CDC25A. Cyclin E and its partner CDK2, which are required for S-phase initiation, collaborate with other cyclin D-dependent kinases to complete RB phosphorylation. Inhibition of the Cip/Kip and INK4 family of proteins that function as CDK inhibitors contributes to the irreversibility of this critical restriction point. Once the restriction point is passed, no further mitogenic stimulus is thought to be required. The cell proceeds to S phase, where the genome has to be precisely replicated. DNA lesions, stalled replication forks, and misincorporated nucleotide bases must be eliminated before they are transmitted to daughter cells. The signaling pathways that lead to cell cycle arrest due to DNA damage and other cellular stresses are frequently altered in various human cancers. Various kinds of DNA damage can activate specific kinases to trigger checkpoints in various cell cycle phases (G1, S, G2). The G1 phase checkpoint is mediated primarily by the tumor suppressor TP53, which in turn activates p21, a cyclin-dependent kinase inhibitor, also known as CDKN1A. The p21 (CDKN1A) acts as an inhibitor of certain cyclin–CDK complexes, especially CDK2 (Blow and

G1/S Transition Quiescent cells (G0) can be driven into mitosis upon a mitogenic signal. The presence of a continuous mitotic signal or stimulus is required in G1 phase until the cell cycle reaches a specific point, the “restriction point” (called “start” in unicellular eukaryotes). The decision of whether to proceed with DNA replication and cell division or exit from the cell cycle to go back to G0 is made at this restriction point. After passing this point, outside mitogenic signals are no longer required. Cell cycle progression moves into a self-sufficient, or “autopilot,” program where cells go through S phase and commit to completing the cell cycle even in the absence of any further mitogenic signals. The main target of G1/S transition control is the retinoblastoma tumor suppressor (RB) protein previously discussed. In response to an external mitogenic stimulus, cyclin D-dependent kinases accumulate and are activated to phosphorylate RB. Unphosphorylated RB exerts its growth-suppressive control by blocking the interaction between the transcription factor E2F and other transcription co-activators to repress expression of genes that are required for

Molecular and Genetic Events in Neoplastic Transformation Hodgson, 2002). However, if the damage cannot be repaired, TP53dependent or TP53-independent pathways may drive the cell toward apoptosis (Symonds et al., 1994) (Fig. 4–7). The incidence of a TP53 mutation in solid human tumors varies but may be as high as 70%, suggesting the importance of this gene as a tumor suppressor. In addition, germline mutations of TP53 cause the Li-Fraumeni syndrome, an autosomal dominant cancer predisposition condition characterized by malignant tumors of multiple tissues. Ataxia telangiectasia mutated protein (ATM) and atm-related protein (ATR) serve as initial DNA damage sensors and transmit this information to TP53 and MDM2. ATM and ATR are key members of the S phase as well as the G2/M checkpoint pathways (Abraham, 2001). Ionizing radiation causes activation of ATM, and ultraviolet light activates ATR-induced responses. Germline mutations of ATM cause ataxia telangiectasia, an autosomal recessive genetic disorder characterized by cancer predisposition and hypersensitivity to ionizing radiation (Rotman and Shiloh, 1999). Activation of CHK, which occurs normally downstream of the ATM pathway is also found to be mutated in patients with the Li-Fraumeni syndrome (Bell et al., 1999).

53

ATM and ATR phosphorylate and activate CHK2 and CHK1, respectively, to block CDK activity mediated by TP53 due to DNA damage or defects in replication. In addition to the pathways already discussed that facilitate the transition from the G1 through S phases of the cell cycle, other central processes are occurring concurrently in the cell and prepare it for normal division. Alterations in any of these pathways can contribute to malignant progression.

Centrosome Cycle Centrosomes are considered the primary microtubule-organizing centers in mammalian cells. Microtubules are critical to many processes in cell compartmentalization and division. Centrosomes consist of centriolar and pericentriolar domains. In most animal cells, the centriolar domain consists of a pair of centrioles and is thought to be critically involved in the structural organization of centrosomal components (Bobinnec et al., 1998). The pericentriolar material is connected to the cytoplasm and organelles of the cell through microtubules and microfilaments. Duplication of the centrosome is initiated at the G1/S

DNA Damage ATM ATR CHK2

BAX p21

P TP53

TP53

Apoptosis

MDM2

TP53 Degradation

CDK2 Cyclin E PCNA E2F RB

RB P P E2F S-phase genes P

ARF p14 Oncogene Activation

Nucleus Cytoplasm Figure 4–7. Simplified network illustrating well described TP53-related pathways. In normal cells the half-life of TP53 is short, and the level of TP53 is kept low under tight regulation by the ubiquitin ligase MDM2. Normally, MDM2, a negative regulator of p53, targets TP53 to be exported from the nucleus for degradation (Xirodimas et al., 2001). Stresses such as DNA damage are sensed by ATM and ATR kinases, which transmit this signal to TP53 and MDM2. ATM phosphorylates and activates CHK2, which further phosphorylates TP53, preventing the interaction of TP53 and MDM2. Phosphorylated TP53 prevents cell cycle progression by upregulating the CDK inhibitor p21. P21, a transcriptional target of TP53, binds to the proliferating cell nuclear antigen (PCNA) to prevent DNA replication without altering the function of PCNA in DNA repair (Levine, 1997) and thus prevents S phase progression. When the pathway is activated, MDM2 forms an autoregulatory loop with TP53. As TP53 levels increase in the cells, so do the MDM2 mRNA and protein levels. This ultimately may result in the formation of nonfunctional TP53 and MDM2 complexes, which can hinder transcriptional activity of TP53 and target TP53 for degradation (Wu et al., 1993; Haupt et al., 1997; Levine, 1997).

This autoregulatory loop involves other important components of cell cycle regulation, such as the tumor suppressor protein p14ARF (p19ARF in mice), which binds to MDM2 to inhibit ubiquitin ligase activity (Pomerantz et al., 1998; Honda and Yasuda, 1999). If the damage to the cell cannot be repaired, activated TP53 increases BAX transcription and blocks the antiapoptotic signals generated by bcl-2, (Miyashita and Reed, 1995), inducing the apoptotic pathways. Components of this or related pathways are altered in various cancer cells, resulting in increased proliferation or decreased apoptosis. For example, overexpression of MDM2 has been detected in soft tissue sarcomas and in tumors with wild-type TP53 (Momand et al., 1998). Activation of oncogenes also has a proliferative effect on this pathway. For example, the RAS oncogene can induce MDM2 transcription to counteract the activity of TP53 (Ries et al., 2000). In the event of oncogene activation in a cell, the tumor suppressor p14ARF can bind directly to the MDM2 protein, blocking MDM2-induced degradation of TP53. Loss of p14ARF expression has also been detected in some tumors, perhaps as a result of methylation of DNA regulatory sequences.

54

PART I: BASIC CONCEPTS

phase and must be completed before mitosis to ensure the polarity of microtubule spindles. During S phase, the split pair of centrioles is semiconservatively duplicated. When the cell reaches 4n DNA content or has effectively duplicated its DNA, centrosomes also have four centrioles and thus two centrosomes to be shared between daughter cells. Centrosome defects have been commonly reported for a variety of solid tumors (Lingle et al., 1998; Pihan et al., 1998; Weber et al., 1998; Sato et al., 1999; Ghadimi et al., 2000) including breast, lung, colon, and prostate cancers as well as hematologic malignancies (Kramer et al., 2003). It is not yet clear, however, whether centrosome defects are early inducers of genomic instability or they develop in parallel with other cellular abnormalities reflecting genomic instability. In both cases, aberrant centrosome defects may alter the polarity of cells and contribute to chromosomal segregation abnormalities and aneuploidy, which may further contribute to genomic instability and malignant progression.

G2/M Transition The mitosis-promoting factor (MPF) consists of the active cyclin B–CDC2 complex, which drives the cell toward mitosis. The cyclin B–CDC2 complex is monitored at several levels to regulate tight control of the mitotic entry time. Cyclin B availability increases toward the end of late S phase, and it accumulates in the cell until mitosis. In addition to accumulation, the specific subcellular localization of cyclin B is important in terms of how it exerts a regulatory force during cell proliferation. Cyclin B is continuously exported from the nucleus into the cytoplasm owing to its nuclear exclusion signal until the beginning of prophase (Gallant and Nigg, 1992; Hagting et al., 1998; Jin et al., 1998; Toyoshima et al., 1998). The activity of MPF is also regulated by phosphorylation. The cyclin B–CDC2 complex is maintained in an inactive state until the G2 phase of the cell cycle due to phosphorylation on tyrosine 15 and threonine 14 of CDC2 by protein kinases such as WEE1 and MYT1. WEE1 prevents premature progression of the cell cycle owing to a partially activated MPF in the nucleus (Heald et al., 1993). MYT1, in contrast, is localized in the Golgi complex and helps maintain an inactive form of CDC2 during interphase (Liu et al., 1997). Protein phosphatases CDC25B and CDC25C are responsible for activation of CDC2 by dephosphorylation. In addition to regulating itself by phosphorylating CDC25C, MPF-mediated phosphorylation induces changes in centrosomes, the nuclear lamina, transcription, the microtubule network, and actin microfilaments. Phosphorylation of lamin subunits results in nuclear lamina breakdown. Down-regulation of transcription during mitosis is thought to be mediated by cyclin B–CDC2, as the complex inhibits RNA polymerase III-mediated transcription (Gottesfeld et al., 1994). Cyclin B–CDC2 phosphorylation of the actin-binding protein caldesmon causes dissociation of microfilaments from calmodulin and thus contributes to cell rounding during mitosis (Yamashiro et al., 1990). Several mechanisms, some of which have been well characterized and others that are just being recognized, tightly regulate passage through the G2/M transition.

G2/M Checkpoints DNA Damage Checkpoint DNA damage induces G2/M arrest to allow repair of the damage or, if the damage cannot be repaired, death of the damaged cell before mitosis to ensure that the daughter cells receive an intact copy of the genome. The main driving force of mitotic entry is the activation of cyclin B–CDC2, the MPF complex, as described above. Activation of MPF depends on the phosphatase activity of CDC25C. ATR inhibits CDC25C through CHK1, resulting in the nuclear exclusion of CDC25C. ATM can also phosphorylate CDC25C through CHK2 in vitro (Matsuoka et al., 1998). However, it is still not clear how these complex pathways function in relation to S-phase DNA damage checkpoints (Qin and Li, 2003).

The ATR protein has also been linked to a “decatenation pathway” in G2/M (Deming et al., 2002). The ATR-dependent topoisomerase II checkpoint ensures decatenation of chromatids before chromosomes condense. A possible mechanism for this checkpoint response involves mitotic delay due to nuclear exclusion of the cyclin B–CDC2 complex (Deming et al., 2001, 2002).

CHFR-Associated Early G2/M Checkpoint The CHFR (checkpoint with FHA and ring finger) protein is a newly identified E3 ubiquitin ligase and part of an early mitotic checkpoint response (Scolnick and Halazonetis, 2000; Chaturvedi et al., 2002; Kang et al., 2002). The CHFR-associated early G2/M checkpoint responds to early microtubule-dependent defects. Cells halt cell cycle progression in prophase and prevent chromosome condensation in response to chemical disruption of the microtubules (e.g., nocodazole, a microtubule poison). Whether this checkpoint monitors microtubulerelated centrosome separation or other microtubule-dependent events early in the cell cycle is not yet clear. However, there is an increasing body of evidence to suggest that this early checkpoint may be defective in a variety of cancer cells and may contribute to the common genomic instability problem seen in cancer (Mariatos et al., 2003; Satoh et al., 2003; Shtivelman, 2003; Toyota et al., 2003; Erson and Petty, 2004). After progression through the G2/M transitions, the cells are ready to segregate their chromosomes and progress through anaphase, telophase, and ultimately cell division.

Progression Through Mitosis Mitosis is divided into several steps, including prophase, metaphase, anaphase, and telophase. After DNA is duplicated, sister chromatids are held together by a complex of proteins called cohesins (Nasmyth et al., 2000). During metaphase, sister chromatids attach via their kinetochores to microtubules of opposite poles in the cell. This process is regulated by the mitotic spindle checkpoint.

Mitotic Spindle Checkpoint Bipolar attachment of chromosomes to the mitotic spindle prior to anaphase is vital for the proper segregation of chromosomes. Therefore, the spindle checkpoint ensures that anaphase is not attempted until chromosomes are properly attached to and aligned on the spindle microtubules from opposite poles. The cell cycle is halted at metaphase if chromosomes are not properly attached. The spindle checkpoint is thought to monitor attachment of microtubules to the kinetochore of chromosomes, the mechanical tension that arises due to this bipolar attachment and sister chromatid cohesion. Even a single unattached kinetochore ablated by laser is enough to trigger the spindle checkpoint (Rieder et al., 1995). A group of mitotic arrest-deficient (Mad) and budding uninhibited by benomyl (Bub) proteins were first identified in budding yeast by mutants that failed to arrest in the cell cycle when treated with microtubule-depolymerizing drugs. Homologues have been identified for many of the checkpoint genes in higher organisms including humans. Mutations of genes encoding spindle checkpoint proteins have been detected in a variety of human cancers although they are relatively rare (Cahill et al., 1998; Lee et al., 1999; Takahashi et al., 1999). The checkpoint, once activated, inhibits activity of the anaphase-promoting complex or cyclosome (APC–C), which is a multisubunit complex responsible for the destruction of anaphaseonset inhibitors and mitotic cyclins and kinases to prevent anaphase transition (Fig. 4–8). Although mutations in the spindle checkpoint have been described in cancer, there is no one gene involved in this checkpoint that is altered in most human malignancies, perhaps reflecting significant redundancy in the pathway or complexities that are not yet understood.

Metaphase to Anaphase Transition At the onset of anaphase, cohesin complexes that hold sister chromatids together are disrupted by an enzyme called separase. Separase proteolytically cleaves a cohesin subunit called SCC1. Because separase is crucial for anaphase onset, the activity of separase is regulated at multiple levels. The most direct regulation of separase described to

55

Molecular and Genetic Events in Neoplastic Transformation

Separin APC/C CDC20 MAD/BUB Proteins

Metaphase

Separin/Securin Complex Anaphase

MAD/BUB Proteins

CDC20 Figure 4–8. Spindle checkpoint monitors proper alignment and attachment of chromosomes to microtubule spindles. During metaphase, chromosomes are held together by a cohesion complex. This connection has to be disturbed at the onset of anaphase through a caspase-related protein, separin. The timing of separin action is the key control point for the onset of anaphase. Separin is inhibited through direct association with a protein called securin. Securin destruction is achieved by an E3 ubiquitin ligase

known as the anaphase-promoting complex or cyclosome (APC–C). CDC20 and APC–C complex is required for the degradation of securin, leading to the loss of cohesin and chromosome separation. In the event of an unattached kinetochore, CDC20 is sequestered to the MAD–BUB complexes, preventing APC–C-dependent degradation of the separin–securin complex.

date is by the inhibitory action of a protein known as securin, which itself is regulated by ubiquitin-dependent proteolysis. Activation of the APC–C complex is achieved once it binds to CDC20 or CDH1, resulting in distinct APC forms such as APCCDC20 and APCCDH1. Polyubiquitin chains are added to securin by APCCDC20 to be degraded in the proteosome so separase is no more inhibited by securin. The additional control for securin degradation is provided by the phosphorylation of securin by a polo kinase. APCCDH1 is activated when degradation of B-type cyclins has been initiated by APCCDC20. In addition, CDH1 is kept phosphorylated during most of mitosis by CDKs and thus rendered incapable of binding to APC. APCCDH1 then ubiquitinates CDC20 to provide an exit from mitosis. Activation of APCCDH1 depends on prior activation of APCCDC20 and proper positioning of the mitotic nucleus between the mother cell and the future daughter cell (Bardin and Amon, 2001). This complex regulation is required to prevent cells from exiting mitosis before anaphase (Fig. 4–8).

CHARACTERISTICS OF NEOPLASTIC TRANSFORMATION Cellular Growth and Proliferation

Cytokinesis Cytokinesis, or cell division, is the final step after mitosis. In mammalian cells, cytokinesis is a combination of several events: cleavage plane specification, actin-myosin contractile ring formation, furrow ingression, and mid-body formation followed by separation. Failure of proper cytokinesis may result in a single cell with two sets of centrosomes and chromosomes, which may further induce formation of abnormal spindles, and finally aneuploidy. Cytokinesis defects have been observed in cancer cells (Shackney et al., 1989). In cells with abrogated TP53-p21, escaped G2 arrest after DNA damage may result in doubled DNA content at the end of mitosis owing to failed cytokinesis (Bunz et al., 1998). More recently, studies of several proteins involved in cytokinesis, such as aurora kinases and septins, have been implicated in human cancer (Warner et al., 2003), but, to date, no specific checkpoints regulating mammalian cytokinesis have been well characterized in relation to cancer.

All cells require growth factors that are generally produced by neighboring cells. Growth factors bind to and activate specific receptors, which in turn can induce various signaling pathways that result in cellular proliferation. Most cancer cells are able to synthesize their own growth factors (e.g., platelet-derived growth factor, transforming growth factor-a) that they need for proliferation. In this way cancer cells reduce their dependence on stimulation from the normal tissue microenvironment. Activation of oncogenes (e.g., RAS) in cancer cells can also cause overexpression of growth factor genes. Alternatively, signaling pathways can be activated by overactive growth factor receptors that are normally active only in response to their ligands. An example of proteins that are commonly altered or overexpressed in cancer is the protein family of tyrosine kinases (TKs). Receptor tyrosine kinases (RTKs) are a subclass of transmembrane spanning receptors with ligand-inducible TK activity. RTK activity is tightly controlled in resting cells, although RTKs can become extremely potent oncogenes if overexpressed or mutated (Craven et al., 2003). Overexpression of RTKs can cause continuous kinase activity because of an increase in the concentration of dimers that normally form in response to a ligand and result in autophosphorylation of receptor residues. Following activation of receptors, constitutive intracellular signaling promotes unregulated cell proliferation. One of the best studied examples of such receptors is the ERBB2 (HER2) receptor of the epidermal growth factor family of receptors. These receptors and related pathways have been implicated in cellular proliferation, apoptosis, differentiation, angiogenesis, migration, and invasion. The ERBB2 (HER2) gene is amplified in 20%–30% of breast and ovarian cancers as well as in lung, prostate, and gastric cancers (Slamon et al., 1989; Zhou and Hung, 2003). Tumors overexpressing ERBB2 (HER2) are highly sensitive to even minimal mitogenic effects

56

PART I: BASIC CONCEPTS

of growth factors and experience activation of a variety of downstream proliferative pathways. Consistent with constitutively active signaling pathways and proliferation, breast cancers overexpressing ERBB2 (HER2) have demonstrated poor prognostic characteristics (Nunes and Harris, 2002). Studies have shown that blockade of these growth factor receptors or pathways have therapeutic implications. In fact, development of an ERBB2 (HER2)-specific antibody (Herceptin) to block the extracellular domain of the receptor has proven to be clinically effective in some cases of breast cancer.

Limitless Replicative Potential and Telomerase Telomeres are the ends of linear chromosomes and consist of tandem GT-rich repeats in vertebrates. In humans, telomeres are 5–15 kb long double-stranded repeats of TTAGGG with a single-stranded 3¢ end overhang that forms a lariat structure by invading the duplex telomeric DNA to form a D (displacement) and a T (telomere) loop. During DNA replication catalyzed by DNA polymerase III, synthesis of the leading strand occurs by sequential addition of deoxyribonucleotides in the same direction (5¢ to 3¢) as movement of the growing fork. For the synthesis of the lagging strand, DNA polymerases need RNA primers for proceeding in the 5¢ to 3¢ direction. Therefore, the lagging strand is synthesized discontinuously as a series of short segments, called Okazaki fragments, initiated from multiple RNA primers. These fragments are ligated after removal of RNA primers. When the final RNA primer is removed from the segment corresponding to the end of the chromosome, there is no upstream strand onto which DNA polymerase can read to fill the resulting gap. Thus, a stretch of unreplicated region remains at the 3¢ end of chromosomes. Eventually this becomes a problem, as the cell loses DNA each time it undergoes replication. This phenomenon is also known as the “end of replication problem” (Fig. 4–9). After repeated cycles of unreplicated chromosome ends, cells with shortened telomeres reach the Hayflick limit, a permanent senescence period. Cells may temporarily escape this period if they lose an essential checkpoint or tumor suppressor gene. However, shortening of telomeres in these cells continues until they reach the second senescence, called “crisis.” Loss of telomeres causes the chromosomes to become unstable, possibly causing chromosome end-toend fusions and gene amplifications via the breakage-fusion-bridge mechanism, discussed earlier, that result in further genetic alterations (Blasco, 2003; Desmaze et al., 2003). Cells that survive “crisis” may

5'

3'

5'

Leading strand

3'

Lagging strand 5'

3'

5'

RNA primers

Okazaki fragments 3'

continue proliferating as they manage to maintain their telomeres by activating an enzyme called telomerase, an RNA-dependent DNA polymerase that synthesizes telomeric DNA. Thus, telomerase helps cells escape from crisis and is sufficient for the immortalization of many diverse cell types; it also provides unlimited proliferative capacity. In support of this idea, telomerase activity is found in more than 90% of cancerous and in vitro immortalized cells but not in most normal human somatic cells (Cong et al., 2002). In some cancers, however, heterogeneous and elongated telomeres are observed owing to one or more mechanisms that require no telomerase activity; these mechanisms are known as alternative lengthening of telomeres (ALT). ALT has been found in a subset of in vitro immortalized, tumor-derived cell lines and in spontaneous human tumors (Reddel et al., 2001). Recent studies suggest that excessive shortening of telomeres may contribute to genomic instability and that telomerase or ALT has key functions during immortalization of tumor cells. However, it is still not clear if telomerase has a role during neoplastic transformation other than maintaining the telomere length (Sharpless and DePinho, 2004). There are still many unknowns about the mechanism of telomerase activation and its role in tumorigenesis. Recent evidence links telomere dysfunction to aging and a variety of diseases, including hemopoietic disorders and cancer (Wong and Collins, 2003). A better understanding of how telomeres are maintained in the cell throughout the life span of an organism may allow the development of novel telomere-based approaches for treating aging-related diseases and cancer.

Apoptosis Cell death can be achieved by necrosis or apoptosis. Necrosis is a passive process that leads to the release of intracellular contents of cells into the extracellular matrix resulting in inflammation. Apoptosis, in contrast, is a highly regulated, active, energy-dependent process of cellular disintegration and death (Kerr et al., 1972) (Fig. 4–10). Apoptotic elimination of cells without an inflammatory response is achieved at specific time points during normal development of an organism. Improper regulation of apoptosis contributes to a variety of disorders, such as neurodegenerative disorders, autoimmune diseases, and cancer. Apoptotic cells are characterized by their loss of substrate attachment, cytoplasmic shrinkage, membrane blebbing, and formation of

3'

5'

5'

3'

3'

5'

5' 3'

RNA primers are removed Okazaki fragments are ligated Figure 4–9. End of replication problem DNA polymerase can add nucleotides only in the 5¢ to 3¢ direction. The newly synthesized leading strand is elongated continuously from a single primer. The other new strand, the lagging strand, is synthesized discontinuously as a series of short segments, called Okazaki fragments, initiated from multiple RNA primers. Adjacent Okazaki fragments are then joined by DNA ligase after

removing the RNA primers. Replication of chromosome ends poses a problem for the cells because the last RNA primer occupies a small portion of the DNA (gray arrow), which is not copied by the DNA polymerase. Therefore an unreplicated gap remains on the lagging strand. Eventually it becomes a problem, as the cell loses DNA each time it undergoes replication.

57

Molecular and Genetic Events in Neoplastic Transformation

INTRINSIC PATHWAY

EXTRINSIC PATHWAY FAS

Mitochondria

BAX

Procaspase 8 tBID

Caspase 8

BID

Granzyme B

Calpains

FADD

ER Stress Ca +2

AIF Endonuclease G

Smac/DIABLO

Cyctochrome c

ATP

Procaspase 9

APAF-1

Omi/HtrA2

XIAP Procaspase 3

Caspase 3 Apoptosome DNA Fragmentation Membrane Blebbing Loss of Substrate Attachment Cytoplasmic Shrinkage Nucleus

CAD

ICAD

CAD Cytoplasm

Figure 4–10. Apoptosis. Activation of apoptosis occurs through two main pathways: intrinsic and extrinsic. The intrinsic pathway involves the release of mitochondrial proteins. Mitochondrial membrane permeability increases owing to cell damage (DNA damaging agents, heat shock, cytotoxic drugs, hypoxia, growth factor withdrawal, irradiation), releasing cytochrome c and other proapoptotic mitochondrial proteins (e.g., Smac/DIABLO, Omi/HtrA2), which results with the activation of caspase 9. Once cytochrome c is released into the cytosol, it complexes with apoptosis protease activator factor 1 (APAF-1), dATP, and cytosolic caspase 9 to form a high-molecular-weight complex called the apoptosome, which in turn activates caspase 3. Activation of caspase-3 results in DNA fragmentation and apoptosis, as caspase 3 cleaves the inhibitor of CAD (caspase-activated DNase), ICAD. When ICAD is bound to CAD, the DNase activity of CAD is inhibited. CAD activation results in DNA cleavage. The extrinsic pathway is induced when death receptors (e.g., FAS)

trimerize upon ligand binding, which in turn facilitates the recruitment of specific adaptor proteins such as FADD (FADD–fas-associated protein with death domain) for activation of caspase 8. Caspase 8 and granzyme B cleave the proapoptotic BID (BCL-2 family member), which can induce the release of other proapoptotic factors, such as cytochrome c, Smac/DIABLO, and Omi/HtrA2. When Smac/DIABLO is released from the mitochondria, it inhibits the activity of an apoptosis inhibitor, XIAP (XIAP inhibits activation of caspase 3 and caspase 9). Omi/HtrA2 was also shown to interact with other IAP proteins via similar mechanisms. An apoptosis-inducing factor (AIF) and endonuclaese G, released from mitochondria, cause large-scale DNA fragmentation. In response to an endoplasmic reticulum stress, calcium-dependent cysteine proteases called calpains also trigger apoptotic pathways through activation of caspases. This highly regulated network of proteins results in apoptosis characterized by loss of substrate attachment, cytoplasmic shrinkage, membrane blebbing, formation of apoptotic bodies, and DNA fragmentation.

apoptotic bodies. Apoptotic nuclei become fragmented. DNA is degraded at internucleosomal linker sites, yielding several hundred basepair-long fragments. The effectors of apoptosis are cysteine-dependent aspartic acidspecific proteases known as caspases (Alnemri et al., 1996; Salvesen and Dixit, 1997; Thornberry and Lazebnik, 1998) that cause cell death by either degrading structural elements of the cell (e.g., lamins, gelsolin) or activating enzymes such as DNases. There are at least 14 caspases identified so far in mammalian cells. Caspases are kept in their inactive precursor forms (zymogen) and are converted to active enzymes by proteolytic cleavage. Some zymogen caspases (caspases 8 and 9) have regulatory amino-terminal prodomains that can interact with adaptor proteins in response to apoptotic stimuli. Formation of

such large oligomeric structures facilitates autocatalytic activation of zymogens, which further activate a proteolytic cascade consisting of a variety of targets in the cell. One mechanism of caspase activation is through death receptors. Death receptors can transmit apoptosis signals in response to external stimuli such as death ligands, or growth factor withdrawal. Death receptors belong to the tumor necrosis factor receptor family and have an extracellular domain and a cytoplasmic “death domain” that initiates apoptotic signaling inside the cell. Trimerization of death receptors leads to the activation of caspases through adaptor proteins (Ashkenazi and Dixit, 1998; Varfolomeev et al., 1998; Yeh et al., 1998). For example, the death ligand FasL binds and oligomerizes its receptor Fas (CD95/APO-1), which in turn facilitates recruitment of the adaptor protein FADD

58

PART I: BASIC CONCEPTS

(FADD/fas-associated protein with death domain). This mechanism of apoptosis regulation, which is dependent on activation through death receptors, is referred to as the extrinsic pathway. The other mechanism important for apoptosis involves the release of cytochrome c from mitochondria independent of caspase 8. Apoptosis regulated through release of mitochondrial enzymes is often referred to as the intrinsic pathway. Various cell damage pathways induced by DNA-damaging agents, heat shock, cytotoxic drugs, hypoxia (low oxygen pressure), growth factor withdrawal, and irradiation converge on mitochondria to permeabilize the membrane and release mitochondrial proteins (Ferri and Kroemer, 2001). Cytochrome c was the first identified protein to be released from mitochondrial intermembrane space because of an apoptotic signal (Budihardjo et al., 1999). Initial studies showed the requirement of cytochrome c for the proteolytic activation of caspase 3 and that microinjection of cytochrome c results in apoptosis (Liu et al., 1996; Zhivotovsky et al., 1998). Once cytochrome c is released from the mitochondrial intermembrane space into the cytosol, it complexes with apoptosis protease activator factor 1 (APAF-1), dATP, and cytosolic caspase 9 to form a high-molecular-weight complex called apoptosome (Li et al., 1997). APAF-1 is a cytosolic protein with an N-terminal caspase recruitment domain (CARD), a nucleotide binding domain, and multiple WD-40 (protein–protein interaction domain) repeats at the C-terminal. Cytochrome c binds to the WD-40 motif of APAF1. The weak interaction of APAF-1 and dATP is enhanced when cytochrome c released from mitochondria binds to the WD-40 domain of APAF-1. Thus the binding of cytochrome c and dATP exposes the CARD domain of APAF-1 to function as a docking region for procaspase 9 (Adrain et al., 1999; Hu et al., 1999; Benedict et al., 2000). Other mitochondrial proteins involved in apoptosis include apoptosis-inducing factor (AIF), endonuclease G, Smac/DIABLO, and Omi/HtrA2 (van Loo et al., 2002). AIF contains both mitochondrial and nuclear signal sequences. It is normally present in the mitochrondrial intermembrane but translocates to the nucleus owing to an apoptotic signal. AIF induces dissipation of the mitochondrial transmembrane potential and large-scale DNA fragmentation in the nucleus (50 kb). Similarly, endonuclaese G is thought to be involved in DNA breakdown (Li et al., 2001). Smac/DIABLO translocates to the cytosol to revert the effects of an apoptosis inhibitor, XIAP (a member of IAP, inhibitor of apoptosis). XIAP serves as a safety switch that inhibits apoptosis by sequestering apoptosome-activated caspase 3 and caspase 9 to block their activities. However, if the apoptotic signal is strong or persists long enough, competitive binding of mitochondrial proteins overcomes the inhibitory actions of XIAP. Omi/HtrA2 was also shown to interact with other IAP proteins via similar mechanisms. Members of the BCL family that have either proapoptotic (BAX, BAK, BID, BOK, BIM, BNIP3, BAD, BMF, NOXA, PUMA) or antiapoptotic (BCL-2, BCL-XL, BCL-W, BOO) functions are also involved in apoptotic signaling through mitochondrial release of cytochrome c. Caspases can also be activated through proteolysis by granzyme B, which is found in the granules of natural killer cells and cytotoxic T lymphocytes (Darmon et al., 1996; Harvey et al., 1996). Granzyme B or caspase 8 can cleave BID, a proapoptotic BCL family member, so the truncated BID (tBID) can translocate to the mitochondria. Activated tBID oligomerizes BAX into the pores of mitochondria, which results in the release of cytochrome c (Budihardjo et al., 1999; Heibein et al., 2000). Granzyme B can activate apoptosis independent of BID, resulting in mitochondrial depolarization and cell death. In addition, calcium-dependent calpain was shown to activate caspases in response to an endoplasmic reticulum stress mechanism (Nakagawa et al., 2000). The CDKs are also known to participate in apoptosis. Although CDKs do not appear to be part of the apoptotic machinery, they can activate pathways that may lead to apoptosis. For example, an increase in CDC2 kinase activity is followed by granzyme B-induced apoptosis among lymphoma cells. Similarly, inactivation of the kinase blocks the activity of the protease (Kasten and Giordano, 1998).

Decreased Apoptosis in Cancer Evasion of apoptosis in cancer can be achieved in a variety of ways. One of the most common pathways targeted during carcinogenesis is regulation of apoptosis by the previously discussed tumor suppressor gene TP53. TP53 is mutated in more than half of all cancers, eliminating an important DNA damage sensor, which normally induces cell cycle arrest or apoptosis. TP53 acts predominantly during the G1 phase of the cell cycle by affecting several pathways, discussed earlier, involving p16, cyclin D1, RB, CDK4, and p21. TP53 can also drive cells through apoptosis by up-regulating proapoptotic BAX due to DNA damage to ultimately induce cytochrome c release. Early experiments demonstrated this role of TP53 in apoptosis in TP53 knockout mice. Normal murine thymocytes and intestinal stem cells undergo apoptosis when exposed to radiation, whereas cells from the TP53 knockout mice do not undergo apoptosis (Lowe et al., 1993; Levine, 1997). Another example of altered apoptotic pathways is seen in follicular center cell lymphoma, where the antiapoptotic BCL-2 is overexpressed owing to the t(14;18) translocation that moves BCL-2 under the transcriptional control of an immunoglobulin light chain gene (Tsujimoto et al., 1984). In support of this, transgenic animals overexpressing BCL-2 have diminished apoptosis and are susceptible to lymphomas with the co-activation of MYC oncogene (Strasser et al., 1990). In some lung and colon cancer cell lines, nonfunctional receptors for FAS that block ligands were found to be up-regulated, resulting in failure to activate procaspases (Pitti et al., 1998). Interestingly, various mitochondrial enzymes have also been linked to inherited cancers, suggesting more unidentified apoptotic roles for mitochondria in addition to the powerhouse function of the cell. Heterozygous mutations of succinate dehydrogenase and fumarate hydratase (fumarase) enzymes that function in the Krebs cycle have been associated with neoplasms such as pheochromocytoma, paraganglioma, papillary renal carcinoma, and leiomyomatosis (Eng et al., 2003). Thus, the complicated role of apoptosis in cancer is still under intense investigation.

Cell–Cell and Cell–Matrix Interactions: Angiogenesis and Metastasis Metastasis is the major cause of death in cancer patients due to disruption of vital organs. Therefore, it is crucial to understand the molecular changes that allow a cell to invade surrounding tissues and metastasize to distant organs. Cells need oxygen and nutrients supplied by the vasculature to survive. During organogenesis, formation of new blood vessels (angiogenesis) is tightly regulated to maintain the needs of newly formed cells and tissues. Similarly, growth and spread of neoplasms to distant organs (metastasis) also require establishment of an adequate blood supply (Fig. 4–11). More than 100 years ago, Paget suggested the importance of the microenvironment in angiogenesis and metastasis by his seed (tumor) and soil (microenvironment) hypothesis. He proposed that metastasis occurs only when the seed and the soil are compatible. A more current version of this hypothesis suggests that neoplastic cells have angiogenic, invasive, and metastatic properties, and that metastasis is selective for the cells that promote angiogenesis, invasion, survival in the circulation, and extravasation into the parenchyma (Fidler, 1973). Angiogenesis, a hallmark of cancer, starts with formation of new blood vessels to promote neoplastic growth. Tumors activate angiogenesis by changing the balance between angiogenesis inducers and antiangiogenic molecules in the microenvironment (Hanahan et al., 1996). The microenvironment is composed of organ-specific cells, endothelial cells, pericytes, immune cells, fibroblasts, and extracellular matrix (ECM). Each component can communicate with one another and with the tumor cell. ECM breakdown allows cell migration and release of angiogenic molecules (Moses, 1997). Cells in the microenvironment can also induce angiogenic factor expression from the tumor cell by various cytokines and growth factors such as

59

Molecular and Genetic Events in Neoplastic Transformation

Genetic, epigenetic alterations in a single cell

Dysplasia

Carcinoma in situ

Invasion

Basal lamina Stroma

Metastasis

Angiogenesis

Circulation

Circulation

Figure 4–11. Angiogenesis and metastasis. Genetic or epigenetic alterations confer a growth advantage to cells. Highly proliferating cells have abnormal shapes but are not yet cancerous (dysplasia). Cells become more abnormal and cancerous as they continue to accumulate alterations but have not yet broken through any tissue boundaries (carcinoma in situ). Invasion takes place when cells acquire changes that let them pass through the basal lamina and move deeper into the connective tissue (stroma). Cancer cells secrete growth factors that stimulate proliferation of endothelial cells in the walls of capillaries in surrounding tissue, resulting in the

outgrowth of new capillaries into the tumor (angiogenesis). Angiogenesis is important not only for supporting tumor growth but also during metastasis. Tumor cells can penetrate the actively growing new capillaries to enter the circulatory system and begin the metastatic process. Alternatively, tumor cells can cross the wall of a lymphatic vessel that ultimately discharges its contents (lymph) into the bloodstream. As a result, secondary tumors arise in distant organs and may lead to death by disrupting a vital organ.

interleukin-6 (IL-6, IL-8, IL-1B, platelet-derived growth factor (PDGF), tumor necrosis factor-a (TNFa), and epidermal growth factor (EGF). Such signals can induce angiogenic factors, including vascular endothelial growth factor (VEGF) and can down-regulate endogenous inhibitors such as thrombospondin-1 which can modulate the angiogenic phenotype (Volpert et al., 1997). Expression of VEGF is up-regulated by lowered oxygen pressure (hypoxia) (Shweiki et al., 1992). VEGF, also known as the vascular permeability factor, stimulates the proliferation and migration of endothelial cells and induces metalloproteinase and plasminogen activities (Unemori et al., 1992; Kumar et al., 1998). Thus, cross-talk between the tumor cell and the endothelial cell is a vital component of angiogenesis. Perivascular cells and pericytes are other components of the angiogenesis procedure. Pericytes are associated with capillaries, provide support for endothelial cells, and are thought to be involved in the regulation of blood flow, phagocytosis, and modulation of new vessel growth (Hirschi and D’Amore, 1996). Breakdown and remodeling of the ECM is important for the angiogenesis process. Tumor cells interact with the ECM through transmembrane receptors called integrins, which form links between the cytoskeleton and the ECM. Cells have altered interactions with the ECM as cancer cells become invasive and metastatic as well as endothelial cells becoming angiogenic. Integrins and proteases degrade and remodel the ECM to encourage cancer cells to move into the connective tissue (stroma). Breakdown of the ECM also results in release of ECM-associated modulators (e.g., basic fibroblastic growth factor) that promote the migration of endothelial cells, tube formation, and vascular patterning. During tumorigenesis, cell–cell and cell–matrix interactions are also commonly altered in cells. Normal cells grown in culture respond to the signals of other cells and stop proliferating when they come into contact with each other. Neoplastically transformed cells, on the

other hand, are not inhibited by contact and, instead, are able to form clusters or foci of growing cells. Cellular adhesion molecules (CAMs) (such as cadherins, immunoglobulin-like molecules (IgCAMs), and integrins) regulate important cell adhesion and signaling pathways that may be disrupted in cancer cells. Cadherin molecules are usually located on adherens junctions and desmosomes, structures of cell–cell interactions. Cadherins on opposing cells interact with each other in a calcium-dependent manner to prevent the cells from entering the mitotic cycle (Christofori and Semb, 1999). Cytoplasmic domains of cadherins are linked to the actin cytoskeleton through various catenin interactions. In addition to the structural roles of cadherin–catenin complexes, they function in signaling pathways. For example, b-catenin also functions in the WNT signaling pathway, which is involved in the activation of genes that have well established links to cancer, such as MYC, cyclin D1, and MMP-7 (matrix metalloproteinase 7) (He et al., 1998; Brabletz et al., 1999; Crawford et al., 1999; Shtutman et al., 1999; Polakis, 2000) (Fig. 4–12). Moreover adenomatous polyposis coli, AXIN proteins, and glycogen synthase kinase 3b also regulate b-catenin function (Polakis, 2000). Expression of an epithelial cell cadherin known as E-cadherin is frequently lost in various cancers. E-cadherin has a vital role in cell polarity and organizing the epithelium. Therefore, loss of E-cadherin causes deregulation of cell-to-cell contact and an increase in cell motility and invasiveness in vitro. Introduction of E-cadherin into these cells suppresses the transformed phenotype (Vleminckx et al., 1991). Germline, somatic, and epigenetic alterations of E-cadherin have been detected in various cancers (Hajra and Fearon, 2002).One interesting observation associated with loss of E-cadherin is the de novo synthesis of mesenchymal cadherins (e.g., N-cadherin). E-cadherin is expressed in epithelial cells, whereas N-cadherins are found in stromal cells (e.g., fibroblasts). This “cadherin switch” mechanism is thought

60

PART I: BASIC CONCEPTS

WNT

Frizzled

LRP5/6

Frizzled

E-C ate nin

b-

b-

DSH AXIN AP C

E-C ate nin

AXIN

Ca ten in

Ca ten

in

b-Catenin

GSK3b b-Catenin

b-Catenin degradation TCF/LEF

LRP5/6

GSK3b

b-Catenin TCF/LEF

b-Catenin

ON

OFF

E.g., CCND1, MYC, MMP7 Figure 4–12. Canonical WNT signaling pathway. Glycogen synthase kinase 3b (GSK3b), AXIN, APC, and cytosolic free b-catenin form a complex in the absence of WNT signal. GSK3b phosphorylates b-catenin to target for degradation. The WNT signal causes Axin to bind to LRP5/6, with GSK3b inhibition by the signaling protein disheveled (DSH), resulting in decreased phosphorylation and increased stability of b-catenin.

Thus, free b-catenin forms complexes with the transcription factor Tcf/Lef for the activation of target genes [e.g., CCND1 (cyclin D1), myc, MMP7 (matrix metalloproteinase 7)]. Aberrant WNT signaling contributes to carcinogenesis possibly through the stabilization of cytoplasmic b-catenin and transcription of target genes involved in proliferation pathways. (Source: Polakis, 2000, 2001.)

to induce the motility of tumor cells, as they lose epithelial cell adhesion and acquire characteristics that would allow survival in a new environment (Cavallaro et al., 2002). Loss of E-cadherin is usually thought to be a marker for defining an invasive, malignant phenotype (Conacci-Sorrell et al., 2002). According to the classic definition of metastasis, cells actively invade the surrounding tissues, enter the circulation, and colonize in distant organs. However, metastasis of some tumors without E-cadherin loss has also been reported (Joo et al., 2003). One might argue that in such cases—where E-cadherin expression is intact but adhesion is defective—cells may passively enter the local lymph node and metastasize to distant organs without developing the highly invasive, malignant phenotype (Cavallaro and Christofori, 2004). The mechanisms of angiogenesis and metastasis are complex, and increased efforts to understand the molecular mechanisms underlying these complex processes may provide important insights to reduce the morbidity and mortality rates.

have been associated with cancer phenotypes. We have just begun to understand the complex interactions of various cellular pathways and the possible interrelated roles of the cancer-related genes in related specific molecular pathways. As our understanding of basic cellular pathways and molecular mechanisms evolves, we are realizing increasingly that there is significant cross-talk between many of these complex and vital processes. This may explain, in part, the great variety of genes that may be altered to generate cancer cells. One area that requires further exploration in cancer research that has not been discussed here is the important role of low-penetrance cancer susceptibility and modifier genes in the population. These genes may act in a dosage-dependent manner, determining variables such as the immune system response and cancer predisposition in individuals. A better understanding of these genes and the classic oncogenes, tumor suppressor genes, and genome instability genes will eventually provide clinical markers for early diagnosis and improved targeted treatment. Two of the best examples of how increased molecular understanding improves clinical management stemming from research are the discovery of Gleevec for the treatment of CML and Herceptin for ERBB2 (HER2)-overexpressing tumors and the clinical benefits patients receive from them. An improved understanding of molecular events should stimulate further benefits to improve the clinical care of cancer patients as well as increase opportunities for the development of cancer prevention strategies.

CONCLUSIONS Cancer is a heterogeneous, complex disease from both clinical and molecular perspectives (Fig. 4–13). With advances in molecular biology and molecular genetic techniques, many genetic alterations

Molecular and Genetic Events in Neoplastic Transformation Point mutations Gene amplification Translocation Loss of heterozygosity Epigenetic alterations

Dysregulated, inactivated DNA repair, and checkpoint genes

Oncogene activation Tumor suppressor loss

Genomic instability

CANCER Figure 4–13. Heterogeneous genetic and epigenetic alterations lead to the malignant cancer phenotype. Alterations may lead to activation of oncogenes, inactivation of tumor suppressor genes, or deregulation of genome stability caretaker genes (DNA damage sense and repair pathways, checkpoint genes), which may either directly alter characteristics of normal cells or induce a mutator phenotype followed by accumulation of other genetic changes. Ultimately, altered signaling, proliferation, the cell cycle, and apoptosis in cancer cells may allow them to invade the surrounding tissues and metastasize to distant organs.

References Abraham RT. 2001. Cell cycle checkpoint signaling through the ATM and ATR kinases. Genes Dev 15:2177–2196. Adrain C, Slee EA, Harte MT, Martin SJ. 1999. Regulation of apoptotic protease activating factor-1 oligomerization and apoptosis by the WD-40 repeat region. J Biol Chem 274:20855–20860. Alnemri ES, Livingston DJ, Nicholson DW, Salvesen G, Thornberry NA, Wong WW, Yuan J. 1996. Human ICE/CED-3 protease nomenclature. Cell 87:171. Ashkenazi A, Dixit VM. 1998. Death receptors: signaling and modulation. Science 281:1305–1308. Balmain A. 2001. Cancer genetics: from Boveri and Mendel to microarrays. Nat Rev Cancer 1:77–82. Barbacid M. 1987. Ras genes. Annu Rev Biochem 56:779–827. Bardin AJ, Amon A. 2001. Men and sin: what’s the difference? Nat Rev Mol Cell Biol 2:815–826. Bell DW, Varley JM, Szydlo TE, Kang DH, Wahrer DC, Shannon KE, Lubratovich M, Verselis SJ, Isselbacher KJ, Fraumeni JF, Birch JM, Li FP, Garber JE, Haber DA. 1999. Heterozygous germ line hCHK2 mutations in Li-Fraumeni syndrome. Science 286:2528–2531. Benedict MA, Hu Y, Inohara N, Nunez G. 2000. Expression and functional analysis of Apaf-1 isoforms: extra Wd-40 repeat is required for cytochrome c binding and regulated activation of procaspase-9. J Biol Chem 275:8461–8468. Bestor TH. 1999. Sex brings transposons and genomes into conflict. Genetica 107:289–295. Bishop JM. 1989. Retroviruses and Oncogenes I. Les Prix Nobel. Stockholm: Almqvist & Wiksell, pp. 220–238. Blasco MA. 2003. Telomeres and cancer: a tale with many endings. Curr Opin Genet Dev 13:70–76. Blow JJ, Hodgson B. 2002. Replication licensing—defining the proliferative state? Trends Cell Biol 12:72–78. Bobinnec Y, Khodjakov A, Mir LM, Rieder CL, Edde B, Bornens M. 1998. Centriole disassembly in vivo and its effect on centrosome structure and function in vertebrate cells. J Cell Biol 143:1575–1589. Boguski MS, McCormick F. 1993. Proteins regulating Ras and its relatives. Nature 366:643–654. Bos JL. 1989. Ras oncogenes in human cancer: a review. Cancer Res 49:4682–4689. Boveri T. 1929. The Origin of Malignant Tumors. Baltimore: Williams & Wilkins, pp. 1–119.

61

Brabletz T, Jung A, Dag S, Hlubek F, Kirchner T. 1999. Beta-catenin regulates the expression of the matrix metalloproteinase-7 in human colorectal cancer. Am J Pathol 155:1033–1038. Budihardjo I, Oliver H, Lutter M, Luo X, Wang X. 1999. Biochemical pathways of caspase activation during apoptosis. Annu Rev Cell Dev Biol 15:269–290. Bunz F, Dutriaux A, Lengauer C, Waldman T, Zhou S, Brown JP, Sedivy JM, Kinzler KW, Vogelstein B. 1998. Requirement for p53 and p21 to sustain G2 arrest after DNA damage. Science 282:1497–1501. Cahill DP, Lengauer C, Yu J, Riggins GJ, Willson JK, Markowitz SD, Kinzler KW, Vogelstein B. 1998. Mutations of mitotic checkpoint genes in human cancers. Nature 392:300–303. Cavallaro U, Christofori G. 2004. Cell adhesion and signalling by cadherins and Ig-CAMs in cancer. Nat Rev Cancer 4:118–132. Cavallaro U, Schaffhauser B, Christofori G. 2002. Cadherins and the tumour progression: is it all in a switch? Cancer Lett 176:123–128. Cavenee WK, Hansen MF, Nordenskjold M, Kock E, Maumenee I, Squire JA, Phillips RA, Gallie BL. 1985. Genetic origin of mutations predisposing to retinoblastoma. Science 228:501–503. Chaturvedi P, Sudakin V, Bobiak ML, Fisher PW, Mattern MR, Jablonski SA, Hurle MR, Zhu Y, Yen TJ, Zhou BB. 2002. Chfr regulates a mitotic stress pathway through its RING-finger domain with ubiquitin ligase activity. Cancer Res 62:1797–1801. Chen RZ, Pettersson U, Beard C, Jackson-Grusby L, Jaenisch R. 1998. DNA hypomethylation leads to elevated mutation rates. Nature 395:89–93. Christofori G, Semb H. 1999. The role of the cell-adhesion molecule Ecadherin as a tumour-suppressor gene. Trends Biochem Sci 24:73–76. Coller HA, Grandori C, Tamayo P, Colbert T, Lander ES, Eisenman RN, Golub TR. 2000. Expression analysis with oligonucleotide microarrays reveals that MYC regulates genes involved in growth, cell cycle, signaling, and adhesion. Proc Natl Acad Sci USA 97:3260–3265. Conacci-Sorrell M, Zhurinsky J, Ben-Ze’ev A. 2002. The cadherin-catenin adhesion system in signaling and cancer. J Clin Invest 109:987–991. Cong YS, Wright WE, Shay JW. 2002. Human telomerase and its regulation. Microbiol Mol Biol Rev 66:407–425. Craven RJ, Lightfoot H, Cance WG. 2003. A decade of tyrosine kinases: from gene discovery to therapeutics. Surg Oncol 12:39–49. Crawford HC, Fingleton BM, Rudolph-Owen LA, Goss KJ, Rubinfeld B, Polakis P, Matrisian LM. 1999. The metalloproteinase matrilysin is a target of beta-catenin transactivation in intestinal tumors. Oncogene 18:2883–2891. Darmon AJ, Ley TJ, Nicholson DW, Bleackley RC. 1996. Cleavage of CPP32 by granzyme B represents a critical role for granzyme B in the induction of target cell DNA fragmentation. J Biol Chem 271:21709–21712. Deming PB, Cistulli CA, Zhao H, Graves PR, Piwnica-Worms H, Paules RS, Downes CS, Kaufmann WK. 2001. The human decatenation checkpoint. Proc Natl Acad Sci USA 98:12044–12049. Deming PB, Flores KG, Downes CS, Paules RS, Kaufmann WK. 2002. ATR enforces the topoisomerase II-dependent G2 checkpoint through inhibition of Plk1 kinase. J Biol Chem 277:36832–36838. Desmaze C, Soria JC, Freulet-Marriere MA, Mathieu N, Sabatier L. 2003. Telomere-driven genomic instability in cancer cells. Cancer Lett 194:173–182. Dobrovic A, Peters GB, Ford JH. 1991. Molecular analysis of the Philadelphia chromosome. Chromosoma 100:479–486. Elliott B, Jasin M. 2002. Double-strand breaks and translocations in cancer. Cell Mol Life Sci 59:373–385. Eng C, Kiuru M, Fernandez MJ, Aaltonen LA. 2003. A role for mitochondrial enzymes in inherited neoplasia and beyond. Nat Rev Cancer 3:193–202. Erson AE, Petty EM. 2004. CHFR-associated early G2/M checkpoint defects in breast cancer cells. Mol Carcinog 39:26–33. Feinberg AP, Tycko B. 2004. The history of cancer epigenetics. Nat Rev Cancer 4:143–153. Feinberg AP, Vogelstein B. 1983. Hypomethylation of ras oncogenes in primary human cancers. Biochem Biophys Res Commun 111:47–54. Feinberg AP, Gehrke CW, Kuo KC, Ehrlich M. 1988. Reduced genomic 5-methylcytosine content in human colonic neoplasia. Cancer Res 48:1159–1161. Ferri KF, Kroemer G. 2001. Mitochondria—the suicide organelles. Bioessays 23:111–115. Fidler IJ. 1973. Selection of successive tumour lines for metastasis. Nature [New Biol] 242:148–149. Futreal PA, Coin L, Marshall M, Down T, Hubbard T, Wooster R, Rahman N, Stratton MR. 2004. A census of human cancer genes. Nat Rev Cancer 4:177–183. Gallant P, Nigg EA. 1992. Cyclin B2 undergoes cell cycle-dependent nuclear translocation and, when expressed as a non-destructible mutant, causes mitotic arrest in HeLa cells. J Cell Biol 117:213–224.

62

PART I: BASIC CONCEPTS

Ghadimi BM, Sackett DL, Difilippantonio MJ, Schrock E, Neumann T, Jauho A, Auer G, Ried T. 2000. Centrosome amplification and instability occurs exclusively in aneuploid, but not in diploid colorectal cancer cell lines, and correlates with numerical chromosomal aberrations. Genes Chromosomes Cancer 27:183–190. Gottesfeld JM, Wolf VJ, Dang T, Forbes DJ, Hartl P. 1994. Mitotic repression of RNA polymerase III transcription in vitro mediated by phosphorylation of a TFIIIB component. Science 263:81–84. Greaves MJ, Wiemels J. 2003. Origins of chromosome translocations in childhood leukaemia. Nat Rev Cancer 3:639–649. Hagting A, Karlsson C, Clute P, Jackman M, Pines J. 1998. MPF localization is controlled by nuclear export. EMBO J 17:4127–4138. Hajra KM, Fearon ER. 2002. Cadherin and catenin alterations in human cancer. Genes Chromosomes Cancer 34:255–268. Hanahan D, Christofori G, Naik P, Arbeit J. 1996. Transgenic mouse models of tumour angiogenesis: the angiogenic switch, its molecular controls, and prospects for preclinical therapeutic models. Eur J Cancer 32A:2386–2393. Harbour JW, Dean DC. 2000. Chromatin remodeling and Rb activity. Curr Opin Cell Biol 12:685–689. Harris H, Miller OJ, Klein G, Worst P, Tachibana T. 1969. Suppression of malignancy by cell fusion. Nature 223:363–368. Harvey NL, Trapani JA, Fernandes-Alnemri T, Litwack G, Alnemri ES, Kumar S. 1996. Processing of the Nedd2 precursor by ICE-like proteases and granzyme B. Genes Cells 1:673–685. Haupt Y, Maya R, Kazaz A, Oren M. 1997. Mdm2 promotes the rapid degradation of p53. Nature 387:296–299. He TC, Sparks AB, Rago C, Hermeking H, Zawel L, da Costa LT, Morin PJ, Vogelstein B, Kinzler KW. 1998. Identification of c-MYC as a target of the APC pathway. Science 281:1509–1512. Heald R, McLoughlin M, McKeon F. 1993. Human wee1 maintains mitotic timing by protecting the nucleus from cytoplasmically activated Cdc2 kinase. Cell 74:463–474. Heibein JA, Goping IS, Barry M, Pinkoski MJ, Shore GC, Green DR, Bleackley RC. 2000. Granzyme B-mediated cytochrome c release is regulated by the Bcl-2 family members bid and Bax. J Exp Med 192:1391–1402. Hirai H, Roussel MF, Kato JY, Ashmun RA, Sherr CJ. 1995. Novel INK4 proteins, p19 and p18, are specific inhibitors of the cyclin D-dependent kinases CDK4 and CDK6. Mol Cell Biol 15:2672–2681. Hirschi KK, D’Amore PA. 1996. Pericytes in the microvasculature. Cardiovasc Res 32:687–698. Honda R, Yasuda H. 1999. Association of p19(ARF) with Mdm2 inhibits ubiquitin ligase activity of Mdm2 for tumor suppressor p53. EMBO J 18:22–27. Hu Y, Benedict MA, Ding L, Nunez G. 1999. Role of cytochrome c and dATP/ATP hydrolysis in Apaf-1-mediated caspase-9 activation and apoptosis. EMBO J 18:3586–3595. Iwabuchi H, Sakamoto M, Sakunaga H, Ma YY, Carcangiu ML, Pinkel D, Yang-Feng TL, Gray JW. 1995. Genetic analysis of benign, low-grade, and high-grade ovarian tumors. Cancer Res 55:6172–6180. Jacks T, Fazeli A, Schmitt EM, Bronson RT, Goodell MA, Weinberg RA. 1992. Effects of an Rb mutation in the mouse. Nature 359:295–300. Jin P, Hardy S, Morgan DO. 1998. Nuclear localization of cyclin B1 controls mitotic entry after DNA damage. J Cell Biol 141:875–885. Joo M, Lee HK, Kang YK. 2003. Expression of E-cadherin, beta-catenin, CD44s and CD44v6 in gastric adenocarcinoma: relationship with lymph node metastasis. Anticancer Res 23:1581–1588. Kang D, Chen J, Wong J, Fang G. 2002. The checkpoint protein Chfr is a ligase that ubiquitinates Plk1 and inhibits Cdc2 at the G2 to M transition. J Cell Biol 156:249–259. Kasten MM, Giordano A. 1998. pRb and the cdks in apoptosis and the cell cycle. Cell Death Differ 5:132–140. Kauffmann-Zeh A, Rodriguez-Viciana P, Ulrich E, Gilbert C, Coffer P, Downward J, Evan G. 1997. Suppression of c-Myc-induced apoptosis by Ras signalling through PI(3)K and PKB. Nature 385:544–548. Kerr JF, Wyllie AH, Currie AR. 1972. Apoptosis: a basic biological phenomenon with wide-ranging implications in tissue kinetics. Br J Cancer 26:239–257. Khanna KK, Jackson SP. 2001. DNA double-strand breaks: signaling, repair and the cancer connection. Nat Genet 27:247–254. Kramer A, Schweizer S, Neben K, Giesecke C, Kalla J, Katzenberger T, Benner A, Muller-Hermelink HK, Ho AD, Ott G. 2003. Centrosome aberrations as a possible mechanism for chromosomal instability in non-Hodgkin’s lymphoma. Leukemia 17:2207–2213. Knudson AG Jr, Hethcote HW, Brown BW. 1975. Mutation and childhood cancer: a probabalistic model for the incidence of retinoblastoma. Proc Natl Acad Sci USA 72:5116–5120.

Kumar R, Yoneda J, Bucana CD, Fidler IJ. 1998. Regulation of distinct steps of angiogenesis by different angiogenic molecules. Int J Oncol 12:749–757. Lee EY, Chang CY, Hu N, Wang YC, Lai CC, Herrup K, Lee WH, Bradley A. 1992. Mice deficient for Rb are nonviable and show defects in neurogenesis and haematopoiesis. Nature 359:288–294. Lee H, Trainer AH, Friedman LS, Thistlethwaite FC, Evans MJ, Ponder BA, Venkitaraman AR. 1999. Mitotic checkpoint inactivation fosters transformation in cells lacking the breast cancer susceptibility gene, Brca2. Mol Cell 4:1–10. Lee MH, Reynisdottir I, Massague J. 1995. Cloning of p57KIP2, a cyclindependent kinase inhibitor with unique domain structure and tissue distribution. Genes Dev 9:639–649. Lengauer C, Kinzler KW, Vogelstein B. 1998. Genetic instabilities in human cancers. Nature 396:643–649. Levine AJ. 1997. P53, the cellular gatekeeper for growth and division. Cell 88:323–331. Levitt NC, Hickson ID. 2002. Caretaker tumour suppressor genes that defend genome integrity. Trends Mol Med 8:179–186. Li LY, Luo X, Wang X. 2001. Endonuclease G is an apoptotic DNase when released from mitochondria. Nature 412:95–99. Li P, Nijhawan D, Budihardjo I, Srinivasula SM, Ahmad M, Alnemri ES, Wang X. 1997. Cytochrome c and dATP-dependent formation of Apaf-1/caspase-9 complex initiates an apoptotic protease cascade. Cell 91:479–489. Lingle WL, Lutz WH, Ingle JN, Maihle NJ, Salisbury JL. 1998. Centrosome hypertrophy in human breast tumors: implications for genomic stability and cell polarity. Proc Natl Acad Sci USA 95:2950–2955. Liotta L, Petricoin E. 2000. Molecular profiling of human cancer. Nat Rev Genet 1:48–56. Liu F, Stanton JJ, Wu Z, Piwnica-Worms H. 1997. The human Myt1 kinase preferentially phosphorylates Cdc2 on threonine 14 and localizes to the endoplasmic reticulum and Golgi complex. Mol Cell Biol 17:571–583. Liu X, Kim CN, Yang J, Jemmerson R, Wang X. 1996. Induction of apoptotic program in cell-free extracts: requirement for dATP and cytochrome c. Cell 86:147–157. Loeb KR, Loeb LA. 2000. Significance of multiple mutations in cancer. Carcinogenesis 21:379–385. Loeb LA. 1991. Mutator phenotype may be required for multistage carcinogenesis. Cancer Res 51:3075–3079. Loeb LA, Loeb KR, Anderson JP. 2003. Multiple mutations and cancer. Proc Natl Acad Sci USA 100:776–781. Loeb LA, Springgate CF, Battula N. 1974. Errors in DNA replication as a basis of malignant changes. Cancer Res 34:2311–2321. Lowe SW, Schmitt EM, Smith SW, Osborne BA, Jacks T. 1993. P53 is required for radiation-induced apoptosis in mouse thymocytes. Nature 362:847–849. Ma C, Martin S, Trask B, Hamlin JL. 1993. Sister chromatid fusion initiates amplification of the dihydrofolate reductase gene in Chinese hamster cells. Genes Dev 7:605–620. Mariatos G, Bothos J, Zacharatos P, Summers MK, Scolnick DM, Kittas C, Halazonetis TD, Gorgoulis VG. 2003. Inactivating mutations targeting the chfr mitotic checkpoint gene in human lung cancer. Cancer Res 63:7185–7189. Markowitz S, Wang J, Myeroff L, Parsons R, Sun L, Lutterbaugh J, Fan RS, Zborowska E, Kinzler KW, Vogelstein B, et al. 1995. Inactivation of the type II TGF-beta receptor in colon cancer cells with microsatellite instability. Science 268:1336–1338. Matsuoka S, Huang M, Elledge SJ. 1998. Linkage of ATM to cell cycle regulation by the Chk2 protein kinase. Science 282:1893–1897. McClintock B. 1938. The fusion of broken ends of sister half-chromatids following chromatid breakage at meiotic anaphases. MO Agric Exp Stn Res Bull 290:1–48. McClintock B. 1940. The stability of broken ends of chromosomes in Zea Mays. Genetics 26:234–282. Miyashita T, Reed JC. 1995. Tumor suppressor p53 is a direct transcriptional activator of the human bax gene. Cell 80:293–299. Momand J, Jung D, Wilczynski S, Niland J. 1998. The MDM2 gene amplification database. Nucleic Acids Res 26:3453–3459. Morgan DO. 1995. Principles of CDK regulation. Nature 374:131–134. Moses MA. 1997. The regulation of neovascularization of matrix metalloproteinases and their inhibitors. Stem Cells 15:180–189. Mueller PR, Coleman TR, Dunphy WG. 1995. Cell cycle regulation of a Xenopus Wee1-like kinase. Mol Biol Cell. 6:119–134. Nakagawa T, Zhu H, Morishima N, Li E, Xu J, Yankner BA, Yuan J. 2000. Caspase-12 mediates endoplasmic-reticulum-specific apoptosis and cytotoxicity by amyloid-beta. Nature 403:98–103. Nasmyth K, Peters JM, Uhlmann F. 2000. Splitting the chromosome: cutting the ties that bind sister chromatids. Science 288:1379–1385.

Molecular and Genetic Events in Neoplastic Transformation Nevins JR. 2001. The Rb/E2F pathway and cancer. Hum Mol Genet 10:699–703. Nowell PC. 1976. The clonal evolution of tumor cell populations. Science 194:23–28. Nowell PC. 1994. Cytogenetic approaches to human cancer genes. FASEB J 8:408–413. Nunes RA, Harris LN. 2002. The HER2 extracellular domain as a prognostic and predictive factor in breast cancer. Clin Breast Cancer 3:125–137. Oshimo Y, Nakayama H, Ito R, Kitadai Y, Yoshida K, Chayama K, Yasui W. 2003. Promoter methylation of cyclin D2 gene in gastric carcinoma. Int J Oncol 23:1663–1670. Pierce AJ, Jasin M. 2001. NHEJ deficiency and disease. Mol Cell 8:1160–1161. Pihan GA, Purohit A, Wallace J, Knecht H, Woda B, Quesenberry P, Doxsey SJ. 1998. Centrosome defects and genetic instability in malignant tumors. Cancer Res 58:3974–3985. Pitti RM, Marsters SA, Lawrence DA, Roy M, Kischkel FC, Dowd P, Huang A, Donahue CJ, Sherwood SW, Baldwin DT, Godowski PJ, Wood WI, Gurney AL, Hillan KJ, Cohen RL, Goddard AD, Botstein D, Ashkenazi A. 1998. Genomic amplification of a decoy receptor for Fas ligand in lung and colon cancer. Nature 396:699–703. Polakis P. 2000. WNT signaling and cancer. Genes Dev 14:1837–1851. Polakis P. 2001. More than one way to skin a catenin. Cell 105:563–566. Pomerantz J, Schreiber-Agus N, Liegeois NJ, Silverman A, Alland L, Chin L, Potes J, Chen K, Orlow I, Lee HW, Cordon-Cardo C, DePinho RA. 1998. The Ink4a tumor suppressor gene product, p19Arf, interacts with MDM2 and neutralizes MDM2’s inhibition of p53. Cell 92:713–723. Qin J, Li L. 2003. Molecular anatomy of the DNA damage and replication checkpoints. Radiat Res 159:139–148. Rampino N, Yamamoto H, Ionov Y, Li Y, Sawa H, Reed JC, Perucho M. 1997. Somatic frameshift mutations in the BAX gene in colon cancers of the microsatellite mutator phenotype. Science 275:967–969. Reddel RR, Bryan TM, Colgin LM, Perrem KT, Yeager TR. 2001. Alternative lengthening of telomeres in human cells. Radiat Res 155:194–200. Rieder CL, Cole RW, Khodjakov A, Sluder G. 1995. The checkpoint delaying anaphase in response to chromosome monoorientation is mediated by an inhibitory signal produced by unattached kinetochores. J Cell Biol 130:941–948. Ries S, Biederer C, Woods D, Shifman O, Shirasawa S, Sasazuki T, McMahon M, Oren M, McCormick F. 2000. Opposing effects of Ras on p53: transcriptional activation of mdm2 and induction of p19ARF. Cell 103:321–330. Riggs AD. 1975. X inactivation, differentiation, and DNA methylation. Cytogenet Cell Genet 14:9–25. Rotman,G, Shiloh Y. 1999. ATM: a mediator of multiple responses to genotoxic stress. Oncogene 18:6135–6144. Rous P. 1911. A sarcoma of the fowl transmissible by an agent separable from the tumor cells. J Exp Med 13:397–411. Rubin H, Temin HM. 1958. Infection with the Rous sarcoma virus in vitro. Fed Proc 17:994–1003. Salvesen GS, Dixit VM. 1997. Caspases: intracellular signaling by proteolysis. Cell 91:443–446. Sato N, Mizumoto K, Nakamura M, Nakamura K, Kusumoto M, Niiyama H, Ogawa T, Tanaka M. 1999. Centrosome abnormalities in pancreatic ductal carcinoma. Clin Cancer Res 5:963–970. Satoh A, Toyota M, Itoh F, Sasaki Y, Suzuki H, Ogi K, Kikuchi T, Mita H, Yamashita T, Kojima T, Kusano M, Fujita M, Hosokawa M, Endo T, Tokino T, Imai K. 2003. Epigenetic inactivation of CHFR and sensitivity to microtubule inhibitors in gastric cancer. Cancer Res 63:8606–8613. Schar P. 2001. Spontaneous DNA damage, genome instability, and cancer: when DNA replication escapes control. Cell 104:329–332. Scolnick DM, Halazonetis TD. 2000. Chfr defines a mitotic stress checkpoint that delays entry into metaphase. Nature 406:430–435. Shackney SE, Smith CA, Miller BW, Burholt DR, Murtha K, Giles HR, Ketterer DM, Pollice AA. 1989. Model for the genetic evolution of human solid tumors. Cancer Res 49:3344–3354. Shah NP, Witte ON, Denny CT. 1991. Characterization of the BCR promoter in Philadelphia chromosome-positive and -negative cell lines. Mol Cell Biol 11:1854–1860. Sharpless NE, DePinho RA. 2004. Telomeres, stem cells, senescence, and cancer. J Clin Invest 113:160–168. Shope RE, Hurst EW. 1933. Infectious papillomatosis of rabbits, with a note on the histopathology. J Exp Med 58:607–624. Shtivelman E. 2003. Promotion of mitosis by activated protein kinase B after DNA damage involves polo-like kinase 1 and checkpoint protein CHFR. Mol Cancer Res 1:959–969. Shtutman M, Zhurinsky J, Simcha I, Albanese C, D’Amico M, Pestell R, Ben-Ze’ev A. 1999. The cyclin D1 gene is a target of the beta-catenin/ LEF-1 pathway. Proc Natl Acad Sci USA 96:5522–5527.

63

Shweiki D, Itin A, Soffer D, Keshet E. 1992. Vascular endothelial growth factor induced by hypoxia may mediate hypoxia-initiated angiogenesis. Nature 359:843–845. Slamon DJ, Godolphin W, Jones LA, Holt JA, Wong SG, Keith DE, Levin WJ, Stuart SG, Udove J, Ullrich A, et al. 1989. Studies of the HER-2/ neu proto-oncogene in human breast and ovarian cancer. Science 244:707–712. Smith KA, Gorman PA, Stark MB, Groves RP, Stark GR. 1990. Distinctive chromosomal structures are formed very early in the amplification of CAD genes in Syrian hamster cells. Cell 63:1219–1227. Stanbridge EJ. 1976. Suppression of malignancy in human cells. Nature 260:17–20. Strasser A, Harris AW, Bath ML, Cory S. 1990. Novel primitive lymphoid tumours induced in transgenic mice by cooperation between myc and bcl-2. Nature 348:331–333. Symonds H, Krall L, Remington L, Saenz Robles M, Jacks T, Van Dyke T. 1994. P53-dependent apoptosis in vivo: impact of p53 inactivation on tumorigenesis. Cold Spring Harb Symp Quant Biol 59:247–257. Tabin CJ, Bradley SM, Bargmann CI, Weinberg RA, Papageorge AG, Scolnick EM, Dhar R, Lowy DR, Chang EH. 1982. Mechanism of activation of a human oncogene. Nature 300:143–149. Takahashi T, Haruki N, Nomoto S, Masuda A, Saji S, Osada H. 1999. Identification of frequent impairment of the mitotic checkpoint and molecular analysis of the mitotic checkpoint genes, hsMAD2 and p55CDC, in human lung cancers. Oncogene 18:4295–4300. Temin HM, Rubin H. 1958. Characteristics of an assay for Rous sarcoma virus and Rous sarcoma cells in tissue culture. Virology 6:669–688. Thornberry NA, Lazebnik Y. 1998. Caspases: enemies within. Science 281:1312–1316. Tkachuk DC, Westbrook CA, Andreeff M, Donlon TA, Cleary ML, Suryanarayan K, Homge M, Redner A, Gray J, Pinkel D. 1990. Detection of bcr-abl fusion in chronic myelogeneous leukemia by in situ hybridization. Science 250:559–562. Toyoshima F, Moriguchi T, Wada A, Fukuda M, Nishida E. 1998. Nuclear export of cyclin B1 and its possible role in the DNA damage-induced G2 checkpoint. EMBO J 17:2728–2735. Toyota M, Sasaki Y, Satoh A, Ogi K, Kikuchi T, Suzuki H, Mita H, Tanaka N, Itoh F, Issa JP, Jair KW, SchuebelKE, Imai K, Tokino T. 2003. Epigenetic inactivation of CHFR in human tumors. Proc Natl Acad Sci USA 100:7818–7823. Tsujimoto Y, Yunis J, Onorato-Showe L, Erikson J, Nowell PC, Croce CM. 1984. Molecular cloning of the chromosomal breakpoint of B-cell lymphomas and leukemias with the t(11;14) chromosome translocation. Science 224:1403–1406. Unemori EN, Ferrara N, Bauer EA, Amento EP. 1992. Vascular endothelial growth factor induces interstitial collagenase expression in human endothelial cells. J Cell Physiol 153:557–562. Van Loo G, van Gurp M, Depuydt B, Srinivasula SM, Rodriguez I, Alnemri ES, Gevaert K, Vandekerckhove J, Declercq W, Vandenabeele P. 2002. The serine protease Omi/HtrA2 is released from mitochondria during apoptosis: Omi interacts with caspase-inhibitor XIAP and induces enhanced caspase activity. Cell Death Differ 9:20–26. Varfolomeev EE, Schuchmann M, Luria V, Chiannilkulchai N, Beckmann JS, Mett OL, Rebrikov D, Brodianski VM, Kemper OC, Kollet O, Lapidot T, Soffer D, Sobe T, Avraham KB, Goncharov T, Holtmann H, Lonai P, Wallach D. 1998. Targeted disruption of the mouse caspase 8 gene ablates cell death induction by the TNF receptors, Fas/Apo1, and DR3 and is lethal prenatally. Immunity 9:267–276. Varmus HE. 1989. Retroviruses and Oncogenes. I. Les Prix Nobel. Stockholm: Almqvist & Wiksell, pp. 194–212 Vleminckx K, Vakaet L Jr, Mareel M, Fiers W, van Roy F. 1991. Genetic manipulation of E-cadherin expression by epithelial tumor cells reveals an invasion suppressor role. Cell 66:107–119. Volpert OV, Dameron LM, Bouck N. 1997. Sequential development of an angiogenic phenotype by human fibroblasts progressing to tumorigenicity. Oncogene 14:1495–1502. Wanzel M, Herold S, Eilers M. 2003. Transcriptional repression by Myc. Trends Cell Biol 13:146–150. Warner SL, Bearss DJ, Han H, Von Hoff DD. 2003. Targeting aurora-2 kinase in cancer. Mol Cancer Ther 2:589–595. Weber RG, Bridger JM, Benner A, Weisenberger D, Ehemann V, Reifenberger G, Lichter P. 1998. Centrosome amplification as a possible mechanism for numerical chromosome aberrations in cerebral primitive neuroectodermal tumors with TP53 mutations. Cytogenet Cell Genet 83:266–269. Wong JM, Collins K. 2003. Telomere maintenance and disease. Lancet 362:983–988. Wu X, Bayle JH, Olson D, Levine AJ. 1993. The p53-mdm-2 autoregulatory feedback loop. Genes Dev 7:1126–1132.

64

PART I: BASIC CONCEPTS

Xirodimas DP, Stephen CW, Lane DP. 2001. Cocompartmentalization of p53 and Mdm2 is a major determinant for Mdm2-mediated degradation of p53. Exp Cell Res 270:66–77. Yamashiro S, Yamakita Y, Ishikawa R, Matsumura F. 1990. Mitosis-specific phosphorylation causes 83K non-muscle caldesmon to dissociate from microfilaments. Nature 344:675–678. Yeh WC, Pompa JL, McCurrach ME, Shu HB, Elia AJ, Shahinian A, Ng M, Wakeham A, Khoo W, Mitchell K, El-Deiry WS, Lowe SW, Goeddel

DV, Mak TW. 1998. FADD: essential for embryo development and signaling from some, but not all, inducers of apoptosis. Science 279: 1954–1958. Zhivotovsky B, Hanson KP, Orrenius S. 1998. Back to the future: the role of cytochrome c in cell death. Cell Death Differ 5:459–460. Zhou BP, Hung MC. 2003. Dysregulation of cellular signaling by HER2/neu in breast cancer. Semin Oncol 30:38–48.

5

Risk Assessment of Carcinogenic Hazards LESLIE T. STAYNER, PAOLO BOFFETTA, AND HARRI VAINIO

H

umans are faced every day with threats from their environment, and in a sense the evaluation of the severity of these risks is an activity that is as old as mankind. However, risk assessment has been recognized as a formal scientific discipline only during the past few decades. The growth of risk assessment has been fueled by societal concerns about the myriad potential environmental hazards that have been identified by epidemiologists, toxicologists, and other environmental scientists. Numerous reports of hazards related to exposures at work, pesticides in food, air and water pollution, and hazardous waste sites have created serious public concerns about possible threats to their health. Confronted with these hazards the public and relevant authorities have increasingly demanded that scientists not only identify hazards but also attempt to quantify the risk associated with them. In the United States, risk assessment has become a requirement for nearly all standards setting by regulatory agencies including the Occupational Safety and Health Administration (OSHA), the Food and Drug Administration (FDA), and the Environmental Protection Agency (EPA) as a result of court decisions and legislation. In a 1980 landmark case, the Supreme Court decision on Benzene (Industrial Union v American Petroleum, 100 S.C.T. 2844, 1980) made assessment of risk a requirement for setting occupational health standards. The Court remanded the OSHA standard for benzene on the basis that OSHA had inadequately demonstrated that lowering the standard would significantly reduce the risk of harm. In this decision the Court stated that Some risks are plainly acceptable and others are plainly unacceptable. If for example the odds are one in a billion that a person will die of cancer by taking a drink of chlorinated water, the risk clearly could not be considered significant. On the other hand, if the odds are one in a thousand that regular inhalation of gasoline vapors that are 2% benzene will be fatal a reasonable person might well consider the risk significant and take the appropriate steps to decrease or eliminate it. This statement has been interpreted by OSHA as indicating that lifetime cancer risks greater than 1 per 1000 are significant enough to warrant regulatory action (Infante 1995). Requirements for risk assessment have been mandated in enabling legislation for the FDA and EPA’s Clean Water and Clean Air Acts, which target hazards associated with cancer risks generally greater than one per million (Rodricks et al., 1987). In Europe, the German Maximum Workplace Concentration (MAK) Commission has a long tradition of identifying and classifying carcinogens (Greim and Reuter, 2001). Substances that have been shown to be carcinogenic in humans or animals are classified in categories 1 or 2. For the agents in these categories, no MAK or BAT (biologic tolerance value for occupational exposures) values are assigned because it is assumed that no safe level of exposure can be identified for a carcinogen. Suspected carcinogens are classified in category 3; and they are, in contrast to agents in categories 1 and 2, also assigned a MAK or BAT value. Furthermore, the MAK Commission placed in categories 4 and 5 substances with carcinogenic properties for which the available data are considered sufficient to define an occupational exposure level (MAK or BAT value) at which no significant contribution to the cancer risk of the exposed persons is to be expected. In response to societal and regulatory needs, methods for risk assessment have been rapidly evolving during the past few years. This

chapter presents a brief overview of the current concepts, methods, and controversies related to cancer risk assessment with a special emphasis on the role of epidemiologic studies. Several books and exhaustive reviews have been written on these issues, and interested readers are referred to these sources for additional information (NAS, 1983, 1996; OSTP, 1985; EPA, 1987; WHO, 1994; IARC, 1999).

DEFINITION OF THE RISK ASSESSMENT PROCESS In 1983 the National Academy of Sciences (NAS) published a landmark report that attempted to define the scope of risk assessment carried out by the U.S. government. According to the NAS report, risk assessment involves “the use of the factual base to define the health effects of exposure of individuals or populations to hazardous materials and situations,” whereas risk management involves “the process of weighing policy alternatives and selecting the most appropriate regulatory action, integrating the results of risk assessment with engineering data and social, economic, and political concerns to reach a decision.” The NAS report highlighted the need for a clear organizational separation between the functions of risk assessment and risk management activities. The NAS report divided the risk assessment process into four distinct elements: hazard identification, dose-response assessment, exposure assessment, and risk characterization, which are briefly defined below.

• Hazard identification: qualitative evaluation of the adverse health effects of a substance.

• Dose-response assessment: process of estimating the relation between the dose of a substance(s) and the incidence of an adverse health effect. • Exposure assessment: evaluation of the types (routes and media), magnitude, time, and duration of actual or anticipated exposures and of doses, when known; and, when appropriate, the number of persons who are likely to be exposed. • Risk characterization: process of estimating the probable incidence of an adverse health effect to humans under various conditions of exposure, including a description of the uncertainties involved. Risk characterization is dependent on information derived from hazard identification, exposure, and dose-response assessments. The first element, hazard identification is generally qualitative in nature but may include quantitative analyses (e.g., meta-analysis). The latter three elements combined are often referred to as quantitative risk assessment (QRA). There remain substantial uncertainties and controversies surrounding each of these steps in the process. These issues and approaches for addressing each of the components of the risk assessment process for carcinogenic hazards are described in the following sections.

HAZARD IDENTIFICATION Exposures to carcinogens were often higher in the past than they are today, and the risks were sufficiently large that they could be easily identified through direct observations of adverse events, such as scrotal cancer among chimney sweeps (Potts, 1775) and liver cancer

65

66

PART I: BASIC CONCEPTS

among workers exposed to vinyl chloride (Waxweiler et al., 1976). Currently, we are often dealing with weaker carcinogenic hazards; and the determination of whether exposure to a chemical, biologic, or physical agent poses a carcinogenic risk to humans has become a far more complex and controversial process than it once was. Several institutions—International Agency for Research on Cancer (IARC), National Toxicology Program (NTP), National Institute for Occupational Safety and Health (NIOSH), EPA—have developed processes for assessing evidence that a hazard has the potential to cause cancer in humans. One of the oldest, most comprehensive, most widely utilized system is the IARC’s monographs on the evaluation of carcinogenic risks to humans, which was initiated in 1969. This program developed formal criteria for determining whether, for humans, a hazard should be considered carcinogenic (group 1), probably carcinogenic (group 2a), possibly carcinogenic (group 2b), not classifiable (group 3), or probably not carcinogenic (group 4) (IARC, 1989, 1991, http://monographs.iarc.fr/).) To date, the IARC has classified 88 agents in group 1, 64 in group 2a, 236 in group 2b, 496 in group 3, and only one in group 4. A list of the agents that IARC has classified as known to be carcinogenic in humans (group 1) is presented in Table 5–1. In general, sufficient evidence* from epidemiologic studies has been necessary to achieve a group 1 (known carcinogen) classification. However, the IARC (1991) has modified its criteria to include the consideration of mechanistic data, which has thus far resulted in upgrading three agents from a lower degree of evidence to group 1 (known) that had lacked sufficient epidemiologic data (2,3,7,8-tetrachlorodibenzo-para-dioxin, neutrons and ethylene oxide), numerous (n = 38) agents† from group 2b (possible) to group 2a (probable), and 6 agents‡ from group 3 (unclassifiable) to 2b (possible). Mechanistic data have also been used to down-grade the classification of eight agents from group 2b (possible) to group 3 (unclassifiable).§ Despite their formal criteria, evaluation processes such as the IARC’s monograph program are obviously influenced by the subjective judgments of individuals performing the reviews. There has been increasing emphasis on using meta-analyses (Greenland 1994a,b) or pooled analyses (Gordon et al., 1998) for summarizing epidemiologic evidence. Meta-analyses are analyses in which the results from individual studies are used as the data for the statistical analysis; pooled analyses, in contrast, involve a combined analysis of the raw data from individual studies. Meta-analyses or pooled analyses are less prone to subjective bias than qualitative reviews of the literature. Although some have questioned the value of meta-analytic techniques for observational data (e.g., Shapiro, 1994), it appears that these methods are gaining acceptance, particularly when such analyses include proper assessment of the almost inevitable sources of heterogeneity in the various study findings (Greenland 1994a,b). It should be noted, however, that meta-analyses and pooled analyses may entail some degree of subjective judgment regarding the choice of studies to include and the methods used to summarize the evidence and to report the results.

DOSE–RESPONSE ASSESSMENT Perhaps the most controversial component in the QRA process is selection of a mathematical model that quantitatively describes the *The IARC defines sufficient evidence in humans as having studies that demonstrate a positive association between exposure and cancer in which chance, bias, and confounding could be ruled out with reasonable confidence. † Acrylamide, adriamycin, azacitidine, benz(a)anthracene, benzidine-based dyes, benzo(a)pyrene, captafol, chloramphenicol, 1-(2-chloroethyl)-3-cyclohexyl1-nitrosourea (ccnu), chlorozotocin, cisplatin, clonorchis sinensis (infection with), dibenz(a,h)anthracene, diethyl sulfate, dimethylcarbamoyl chloride, 1,2-dimethylhydrazine, dimethyl sulfate, epichlorohydrin, ethylene dibromide, n-ethyl-n-nitrosourea, etoposide, glycidol, 2-amino-3-methylimidazo(4,5f )quinoline, 5-methoxypsoralen, 4,4¢-methylene bis(2-chloroaniline) (moca), methyl methanesulfonate, n-methyl-n¢-nitro-n-nitrosoguanidine, n-methyl-nnitrosourea, n-nitrosodiethylamine. ‡ Aziridine, bleomycins, 1,2-epoxybutane, styrene, diesel fuel (marine), gasoline. § Amitrole, atrazine, di(2-ethylhexyl) phthalate, ethylenethiourea, d-limonene, melamine, saccharin, sulfamethazine.

Table 5–1. Agents, Radiation, Mixtures, and Exposure Circumstances Classified by IARC as Known Human Carcinogens (Group 1)

chemical and physical agents Aflatoxins, 4-aminobiphenyl, arsenic and arsenic compounds, asbestos, azathioprine, benzene, benzidine, beryllium and beryllium compounds, N,N-bis(2-chloroethyl)-2-naphthylamine (chlornaphazine), bis(chloromethyl)ether and chloromethyl methyl ether, 1,4-butanediol dimethanesulfonate (busulfan; Myleran), cadmium and cadmium compounds, chlorambucil, 1-(2-chloroethyl)-3-(4-methylcyclohexyl)-1nitrosourea (methyl-CCNU; semustine), chromium(VI) compounds, cyclosporin, cyclophosphamide, diethylstilbestrol, erionite, ethylene oxide, etoposide, herbal remedies containing plant species of the genus Aristolochia, melphalan, 8-methoxypsoralen (methoxsalen), MOPP and other combined chemotherapy including alkylating agents, mustard gas (sulfur mustard), 2-naphthylamine, neutrons, nickel compounds, estrogen therapy, postmenopausal, estrogens (steroidal and nonsteroidal), Opisthorchis viverrini, oral contraceptives (combine and sequential), silica (crystalline (inhaled in the form of quartz or cristobalite from occupational sources), talc containing asbestiform fibers, tamoxifen, 2,3,7,8tetrachlorodibenzo-p-dioxin, thiotepa, treosulfan, vinyl chloride

infectious agents Epstein-Barr virus, Helicobacter pylori (infection with), hepatitis B virus (chronic infection with), hepatitis C virus (chronic infection with), human immunodeficiency virus (HIV) type 1 (infection with), human papillomavirus type (HPV) types 16 and 18, human T-cell lymphotropic virus type I, Schistosoma haematobium (infection with)

radiation Phosphorus-32 (as phosphate), plutonium-239 and its decay products (may contain plutonium-240 and other isotopes), radioiodines (exposure during childhood to short-lived isotopes, including iodine-131, from atomic reactor accidents and nuclear weapons detonation), radionuclides (a- and b-particle-emitting, internally deposited), radium 222, 224, 226, or 228 (and their decay products), solar radiation, thorium-232 and its decay products (administered intravenously as a colloidal dispersion of thorium232 dioxide), X and g radiation

mixtures Alcoholic beverages, analgesic mixtures containing phenacetin, betel quid with tobacco, coal tar pitches, coal tars, mineral oils (untreated and mildly treated), salted fish (Chinese style), shale oils, soots, tobacco (smokeless, voluntary and involuntary smoking), wood dust

exposure circumstances Aluminium production, auramine (manufacture of), boot and shoe manufacture and repair, coal gasification, coke production, furniture and cabinet making, hematite mining (underground with exposure to radon), iron and steel founding, isopropanol manufacture (strong acid process), magenta (manufacture of), painter (occupational exposure), rubber industry, strong inorganic acid mists containing sulfuric acid

relation between the dose (or exposure) and the probability of an adverse health effect. This is because the choice of a model potentially has a significant impact on the resulting predictions of risk, particularly at low doses where few if any data may be available; the predictions thus involve a large extrapolation. For example, different models were applied to an analysis of exposure–response in a study of cadmium workers, which resulted in estimates of risk that varied nearly an order of magnitude (Stayner et al., 1995). In addition to the uncertainty related to model specification, there are obviously issues that must be considered when using epidemiologic data for QRA, such as the presence of biases or confounding, the quality of the exposure data, and the consistency of the exposure–response relation (HertzPicciotto, 1995). Dose–response models based on toxicologic data are often even more controversial than those based on epidemiologic data because of the potential interspecies differences in response to carcinogens (Ames and Gold, 1990). The relevance of animal data for predicting human risk is also frequently questioned on the basis that toxicologic studies are generally conducted at relatively high levels of exposure to achieve adequate statistical power for detecting an effect. For example, it has been suggested in the case of diesel exhaust particulates that rats developing lung tumors were exposed to excessive levels of particu-

Risk Assessment of Carcinogenic Hazards lates, thereby overloading normal lung clearance mechanisms; moreover, this response at high doses may be irrelevant for predicting human risk (Nikula et al., 1997). It has been suggested that the uncertainty concerning the choice of basic assumptions when extrapolating from animals to humans (e.g., scaling on body weight or surface area) may result in as much as a 10-fold variation in risk predictions, and that this uncertainty is generally at least as large as those related to potential uncertainties related to exposures or other issues involved when using epidemiologic data (Smith, 1988). It should also be recognized that, although ideally these analyses would be based on the dose to the target tissue, they have more often been based on exposure concentrations in the external environment (e.g., air or water). This may introduce substantial uncertainty in the risk assessment when toxicologic data are used and there are differences in interspecies metabolism of carcinogenic agents, or when the relation between exposure and dose in humans is nonlinear. Occasionally, it has been possible to estimate the dose based on a biologic sample (e.g., blood lead level) or using physiologically based pharmacokinetic models (PBPKs). The latter approach involves fitting compartmental models that reflect the deposition, clearance, and kinetics of metabolism in various organs of the body (Anderson et al., 1987). The parameters fitted in these PBPK models are often based on limited and poorly validated data, particularly for humans, and thus are also subject to a large degree of uncertainty. Historically, risk assessments for carcinogens and noncarcinogens were based on the concept of identifying a “safe” level of exposure by identifying a “no observed adverse effect level” (NOAEL) from an epidemiologic or toxicologic study and dividing the NOAEL by what has been referred to as a safety or uncertainty factor¶ (Dourson and Stara, 1983). It was subsequently recognized that based on the notion that if “one hit” of radiation or a chemical could cause a mutation resulting in a tumor then any level of exposure would be associated with some finite probability of cancer. Thus, a safe level of exposure could not be identified, and the focus shifted toward modeling the cancer risk to determine a “virtually” safe dose or, in other words, a dose that was associated with an insignificantly small risk. The European Union (Commission Directive 93/67/EEC, Article 3, paragraph 1), some countries (e.g., England), and the World Health Organization (1994) have continued to use the NOAEL/uncertainty factor approach for carcinogens, particularly for those believed to be nongenotoxic or are likely to have a “threshold effect.” The NOAEL/uncertainty factor approach has been criticized for failing to take into account the size of the study and the shape of the exposure–response curve; the benchmark dose approach has been suggested as an alternative (Crump, 1984). This approach essentially consists of modeling the data with any appropriately fitting method and then identifying from this model the statistical lower boundary (e.g., 95% lower boundary) on the dose corresponding to a risk of approximately 1%–10%, which is referred to as the benchmark dose. The benchmark dose is expected to be associated with some residual risk and uncertainty, so it should be treated essentially like a NOAEL and be divided by the appropriate uncertainty factors. Methods for dose–response modeling in cancer risk assessment may be broadly divided into biologically based and empirically (or statistically) based approaches. The most commonly used biologically based models have been based on the Armitage and Doll multistage theory of carcinogenesis (Armitage and Doll, 1954). The development of this theory was largely based on the observation that the incidence of most human cancers increases with age raised to a power. The theory suggests that for a cell to become cancerous it must progress through a number of irreversible changes (or stages), which must take place in a certain order. It is also assumed that the waiting time dis¶

In general, the NOAEL may be divided by a factor of 10 for human variability, a factor of 10 if animal data are used for interspecies variability, and a factor of 10 if the only studies available do not involve chronic exposures. The factors are multiplied to determine an overall uncertainty factor. For example, if a subchronic toxicologic study is used, the overall uncertainty factor would be 1000 (10 for human variability ¥ 10 for interspecies variability ¥ 10 for subchronic versus chronic effects).

67

tribution from any one stage to the next is exponential, and that the dose rate would be linearly related to the transition rates governing one or more of the stages. Several mathematical formulations of the multistage model have been developed for analyzing toxicologic data (Guess and Crump, 1976; Crump and Howe, 1984), but these methods are not appropriate for epidemiologic data that have a more complex structure. An approximate form of the model (Whittemore, 1977) has been used in a few analyses of epidemiologic studies for risk assessment purposes (e.g., Pearce, 1988; Stayner et al., 1995; Dawson and Alexeeff, 2001). This approach weights exposures in a manner consistent with the multistage theory: by age at first exposure, time since last exposure, the number of stages in the process, and the stage affected by the carcinogen. An exact form of the multistage model has been fitted to epidemiologic data in a few cases (e.g., Thomas, 1983). “Two-stage clonal expansion” models have been proposed as an extension to the multistage model for carcinogenic risk assessments (Moolgavkar and Knudson, 1981). These models attempt to incorporate information on the kinetics of cell growth and differentiation in addition to allowing for two mutational events. Exposure may be modeled as having an effect on the rate of the mutational events, on the rate of normal cell growth, and on the rate of proliferation of initiated cells. Although these models offer the potential for incorporating additional biologic data (and hence improve the risk assessment process), their use requires additional information that may not be available for human populations (e.g., growth rates of normal and initiated cells). Two-stage clonal expansion models have been found to provide a better representation of the age-related incidence of hormonally related tumors (e.g., breast) or tumors of tissues that have rapid growth during early life followed by little or no cell division in later life (e.g., lymphoid or brain) than the multistage model (Moolgavkar, 1986). Largely because of the complexity of fitting these models, they have primarily been used as a means of exploring the mechanisms involved in carcinogenesis and in a few cases for QRA based on epidemiologic data (Stayner et al., 1995; Moolgavkar et al., 1998, 1999). Statistical models have been used more frequently than biologically based models for risk assessments based on epidemiologic data. Although these models are empirical in nature and may lack a biologic basis, they are sufficiently flexible to allow a wide range of possible dose–response curves. The Cox proportional hazards (Cox, 1972) and Poisson regression (Frome, 1983) are statistical models that have frequently been used for the analysis of epidemiologic mortality studies. Although epidemiologists have commonly assumed a loglinear relation between disease and exposure when applying these models, the development of approaches and software for fitting generalized linear models (McCullagh and Nelder, 1985) and generalized additive models (Hastie and Tibshirani, 1990) make it now possible to fit almost any conceivable dose–response pattern. There has been a tradition in risk assessment to assume a linear relative rate model for the purposes of cancer risk assessment, which can be expressed mathematically as: RR = 1 + b(X), where RR is the relative rate, b is the slope, and X is cumulative exposure or another metric of exposure (e.g., Smith, 1988). The linear relative rate model has generally been used as the basis for risk assessments of the carcinogenic risk associated with radiation exposures (Committee on the Biological Effects, 1990). The generally offered basis for assuming a linear relative rate model is that under the multistage theory the effect of a carcinogen on cancer incidence is expected to be linear at low doses (e.g., OSTP, 1985) or that this is a conservative assumption intended to protect public health. In fact, this assumption may not always be conservative. There are examples of a supralinear exposure–response relation in toxicology and epidemiology (Stayner et al., 2003). The assumption of low dose linearity for all carcinogens has been under considerable attack in recent years. It has been increasingly argued that there is a threshold below which exposure is safe, particularly for carcinogens that act through a nongenotoxic mechanism and occasionally even for genotoxic carcinogens (Bolt, 2003). However, it has also been noted that even if a carcinogen has a threshold of effect it may linearly increase the risk of cancer at low doses if it is adding to a

68

PART I: BASIC CONCEPTS

background process involving the same mechanism (Peto, 1978; Hoel, 1980). It is also been suggested that low levels of exposure to carcinogens may have a protective effect (Calabrese and Baldwin, 2003). This argument is still controversial, and it is unclear to what extent it is generalizable, particularly for carcinogens. The U.S. EPA (1996) has proposed an alternative method for dose–response evaluation of carcinogens in which a best fitting empirical model is fitted to the data and is used to estimate the effect level corresponding to a 10% risk (ED10) or a lower level of risk if the data permit. A straight line would be drawn from the ED10 to the origin to estimate the probability of carcinogenic effects when the mode of action of a carcinogen is believed to lead to linearity at low doses. If nonlinearity is assumed, a “margin of exposure analysis” would be used rather than estimating the probability of effects at low doses. A margin of exposure analysis would involve estimating the ratio of the level of current exposures to the ED10 and contrasting this ratio with the standard uncertainty factors. Other approaches are used for estimating dose–response relations in countries such as Canada (based on the TD05, the dose that induces a 5% increase in the incidence of tumors in animals (Environmental Health Directorate, 1994) and the Netherlands (linear extrapolation from the lowest dose available in experimental studies) (Health Council of the Netherlands, 1994; for a review, see Zeise et al., 1999.)

EXPOSURE ASSESSMENT Estimation of risk is not only dependent on the development of an exposure–response model but also on estimation of exposures of the population at risk. Obviously, no risk exists in the absence of exposure. To characterize risk fully, information is needed on the distribution of exposures in the target population. Such information is unfortunately often lacking or inadequate for environmental and occupational hazards, and risk assessors have been forced to make assumptions about both the level and duration of exposure to a hazard. In the past, to err on the side of protecting the public’s health, government agencies have frequently used worst case scenarios to estimate exposures when assessing risk. For example, the EPA might estimate exposures to air pollution for an individual who resided for a lifetime at the fence of an industrial facility. OSHA frequently estimates risk for workers who are exposed for a “working lifetime” (e.g., 45 years) at the current or proposed regulatory standard. Increasingly, risk assessors have attempted to make more realistic assumptions about exposures and to evaluate the distribution of exposures in the population rather than relying on worst case exposure scenarios. Monte Carlo methods (e.g., Finley and Paustenbach, 1994) are being used for developing more realistic estimates of the distribution of exposures in environmental settings. However, these models are only as good or as realistic as the data on which they are based, which is still often limited.

RISK CHARACTERIZATION Risk characterization is the final and perhaps the most critical step of the risk assessment process. Risk characterization was the subject of a report produced by a panel convened by the National Research Council (NRC, 1996). This report offered the following new definition of risk characterization. Risk characterization is the synthesis and summary of information about a potentially hazardous situation that addresses the needs and interests of decision makers and of interested and affected parties. Risk characterization is a prelude to decision making and depends on an iterative, analytic-deliberative process. The report stressed that risk characterization is a decision-driven activity that should be tailored to fit the needs of the decision-makers and affected parties. Risk characterization generally involves the presentation of estimates of lifetime (e.g., up to age 75) risks for a particular exposure scenario. For occupational exposures the scenario generally assumes

45 years of exposure during a 40-hour week, whereas for environmental exposures it is commonly assumed that the exposure is continuous (i.e., 24 hours a day) over the entire lifetime. Life-table methods have been established for developing these risk estimates that take into account the influence of competing causes of death (Gail, 1975; Committee on the Biological Effects, 1990). The number of years of life lost is an alternative measure of the effect of a hazardous agent and better reflects the greater impact of a death that occurs at an early age (Park et al., 2002). It is important to recognize that a complete risk characterization should include, in addition to the risk estimates themselves, a clear description of the assumptions made during the analysis and the uncertainties underlying the risk estimates. Increasingly, there is a demand for risk assessment to include formal sensitivity and uncertainty analyses to quantify the potential impact of assumptions and uncertainties on the estimation of risk (Fayerweather, et al., 1999; Stayner et al., 1999).

CONCLUSIONS This chapter has briefly summarized current concepts and practices used for assessing carcinogenic hazards with an emphasis on the use of epidemiologic data. Risk assessment is a new and evolving science, and methods for conducting risk assessment are still at an embryonic stage, although undergoing rapid development. Controversies still abound concerning the appropriate methods and data to use and are likely to persist given the great uncertainties involved in extrapolating beyond the range of available data, the underlying biases and other limitations of observational data, and the political and societal implications of these analyses. Skeptics have argued that risk assessment, at least as it is currently practiced, has not been a useful tool for addressing societal concerns about exposures to environmental and occupational hazards (e.g., Silbergeld, 1993). Their primary concern is that the increasingly intense debates concerning risk assessments may come to be used as an excuse for delay in the development of appropriate regulatory and other responses to environmental and occupational hazards. For example, it has taken the U.S. EPA more than 20 years to finalize its risk assessment for exposure to diesel exhaust particulates (Stayner, 1999). A spirited debate has emerged over the use of the “precautionary principle” as an alternative basis for public health decision-making, and this approach has recently been embodied in some environmental legislation of the European Union (Commission of the European Communities, 2000). The precautionary principle has been defined as the need to take some precautionary measures to prevent threats to human health even when a cause-and-effect relation has not been fully established (Kriebel and Tickner, 2001) This is not a new principle for epidemiologists: We all know the story of how John Snow convinced the authorities to remove the Broad Street pump well before the cause of the cholera epidemic in London was understood. In our view, risk assessment and the precautionary principle should not be viewed as conflicting paradigms but, rather, as complementary approaches for developing appropriate policies to address risks posed by exposure to carcinogens and other hazards. Identification and quantification of risks is clearly a useful tool for informed decision-making. Risk assessments are inherently uncertain and should, as the NAS (1996) suggested, be viewed as an iterative process in need of continual improvements through research targeted to fill the gaps in our knowledge. However, our inability to characterize risks accurately at any one point in time should not be used as an excuse for not taking appropriate measures to prevent potential harm to the public. References Ames BN, Gold LS. 1990. Carcinogens and human health. Part 1. Science 251:1645–1646. Andersen ME, Clewell HJ, Smith FA, Reitz RH. 1987. Physiologically based pharmacokinetic modeling and the risk assessment process for methylene chloride. Toxicol Appl Pharmacol 87:185–205. Armitage P, Doll R. The age distribution of cancer and a muti-stage theory of carcinogenesis. Br J Cancer 8:1–12.

Risk Assessment of Carcinogenic Hazards Bolt HM. 2003. Genotoxicity-threshold or not? Introduction of cases of industrial chemicals. Toxicol Lett 140-141:43–51. Calabrese EJ, Baldwin LA. 2003. Toxicology rethinks its central belief: hormesis demands a reppraisal of the way risks are assessed. Nature 421:691–692. Commission of the European Communities. 2000. Communique on the Precautionary Principle. Brussels: European Commission. Committee on the Biological Effects of Ionizing Radiation, Board of Radiation Effects Research, Commission on Life Sciences, National Research Council. 1990. Biological Effects of Ionizing Radiation (BEIR). V. Health Effects of Exposure to Low Levels of Ionizing Radiation. Washington, DC: National Academy Press. Cox DR. 1972. Regression models and life tables. J R Stat Soc 34:187–202. Crump KS. 1984. A new method for determining allowable daily intakes. Fundam Appl Toxicol 4:854–871. Crump KS, Howe RB. 1984. The multistage model with a time-dependent dose pattern: applications to carcinogenic risk assessment. Risk Analysis 4:163–176. Dawson SV, Alexeeff GV. 2001. Multi-stage model estimates of lung cancer risk from exposure to diesel exhaust, based on a U.S. railroad worker cohort. Risk Analysis 21:1–18. Dourson ML, Stara JF. Regulatory history and experimental support of uncertainty (safety) factors. Regul Toxicol Pharmacol 3:224–238. Environmental Health Directorate. 1994. Health Canada. Human Health. Risk Assessment for Priority Substances. Ottawa: Environmental Health Directorate Canada. EPA. 1987. The Risk Assessment Guidelines of 1986. EPA/600/8-87/045. Washington, DC: EPA. EPA. 1996. Proposed guidelines for carcinogen risk assessment. Fed Register 61:17959–18101. Fayerweather WE, Collins JJ, Schnatter AR, Hearne FT, Menning RA, Reyner DP. 1999. Quantifying uncertainty in a risk assessment using human data. Risk Analysis 19:1077–1090. Frome EL. 1983. The analysis of rates using Poisson regression models. Biometics 39:665–674. Gail M. 1975. Measuring the benefit of reduced exposure to environmental carcinogens. J Chronic Dis 28:135–147. Gordon I, Boffetta P, Demers PA. 1998. A case study comparing a metaanalysis and a pooled analysis of studies of sinonasal cancer among wood workers. Epidemiology 9:518–524. Greenland S. 1994a. Can meta-analysis be salvaged? Am J Epidemiol 1:140:783–787. Greenland S. 1994b. Invited commentary: a critical look at some popular metaanalytic methods. Am J Epidemiol 140:290–296. Greim H, Reuter U. 2001. Classification of carcinogenic chemicals in the work area by the German MAK Commission: current examples for the new catageries. Toxicology 166:11–23. Guess H, Crump K. 1976. Low dose extrapolation of data from animal carcinogenicity experiments: analysis of a new statistical technique. Math Biosci 30:15–36. Hastie TJ, Tibshirani RJ. 1990. Generalized Additive Models. Monographs on Statistics and Applied Probability. London: Chapman & Hall. Health Council of the Netherlands. 1994. Carcinogenic risk assessment of carcinogenic chemicals in The Netherlands. Reg Toxicol Pharmacol 19:14–30. Hertz-Picciotto I. 1995. Epidemiology and quantitative risk assessment: a bridge from science to policy. Am J Public Health 85:484–491. Hoel DG. 1980. Incorporation of background in dose-response models. Fed Proc 39:73–75. Infante PF. 1995. Cancer and blue collar workers: who cares? New Solutions 5:52–57. International Agency for Research on Cancer (IARC). 1989. Report of an IARC Working Group to Review the Approaches and Processes Used to Evaluate the Carcinogenicity of Mixtures and Groups of Chemicals. IARC Internal Technical Report No. 89/004. Lyon: IARC. International Agency for Research on Cancer (IARC). 1991. A Consensus Report of an IARC Monographs Working Group on the Use of Mechanisms of Carcinogenesis in Risk Identification. IARC Internal Technical Report No. 91/002. Lyon: IARC. International Agency for Research on Cancer (IARC). 1999. Quantitative Estimation and Prediction of Human Cancer Risks. In: Moolgavkar S, Krewski D, Zeise L, Cardis E, Moller H, editors. IARC Scientific Publication No. 131. Lyon: IARC.

69

Kriebel D, Tickner J. 2001. Reenergizing public health through precaution. Am J Public Health 91:1351–1361. McCullagh P, Nelder JA. 1985. Generalized Linear Models. Monographs on Statistics and Probability. New York: Chapman & Hall. Moolgavkar SH. 1986. Carcinogenesis modeling: from molecular biology to epidemiology. Annu Rev Public Health 7:151–169. Moolgavkar SH, Knudson AG. 1981. Mutation and cancer: a model for human carcinogenesis. J Natl Cancer Inst 65:559–569. Moolgavkar SH, Luebeck EG, Anderson EL. 1998. Estimation of unit risk for coke oven emissions. Risk Analysis 18:813–825. Moolgavkar SH, Luebeck EG, Turim J, Hanna L. 1999. Quantitative assessment of the risk of lung cancer associated with occupational exposure to refractory ceramic fibers. Risk Analysis 19:599–611. NAS (National Academy of Science). 1983. Risk Assessment in the Federal Government: Managing the Process. Washington, DC: National Academy Press. NAS (National Academy of Science). 1996. Science and Judgement in Risk Assessment. Washington, DC: National Academy Press. Nikula KJ, Avila KJ, Griffith WC, Mauderly JL. 1997. Lung tissue responses and site of particle retention differ between rats and cynomolgus monkeys exposed chronically to diesel exhaust and coal dust. Fundam Applied Toxicol 37:37–53. NRC (National Research Council); Stern PC, Fineberg HV, editors. 1996. Understanding Risk. Informing Decisions in a Democratic Society. Washington, DC: National Academy Press. OSTP (Office of Science and Technology Policy). 1985. Chemical carcinogens: a review of the science and its associated principles. Fed Register 50:10372–10442. Park RM, Bailer AJ, Stayner LT, Halperin W, Gilbert SJ. 2002. An alternative characterization of hazard in occupational epidemiology: years of life lost per years worked. Am J Ind Med 42:1–10. Pearce N. 1988. Multistage modelling of lung cancer mortality in asbestos textile workers. Int J Epidemiol 17:747–752. Peto R. 1978. Carcinogenic effects of chronic exposure to very low levels of toxic substances. Environ Health Perspect 22:155–159. Potts P. 1775. Cancer scroti. In: Chirurgical Observations. London: Hawes, Clarke & Collins, pp. 63–68. Rodricks JV, Brett SM, Wrenn GC. 1987. Significant risk decisions in federal regulatory agencies. Regul Toxicol Pharmacol 7:307–320. Silbergeld EK. 1993. Risk assessment: the perspective and experience of U.S. environmentalists. Environ Health Perspect 101:100–104. Smith AH. 1988. Epidemiologic input to environmental risk assessment. Arch Environ Health 43:124–129. Shapiro S. 1994. Meta-analysis/shmeta-analysis. Am J Epidemiol 140:771–778. Stayner L. 1999. Protecting public health in the face of uncertain risks: the example of diesel exhaust. Am J Public Health 89:991–993. Stayner L, Bailer, AJ, Smith R, Gilbert S, Rice F, Kuempel E. 1999. Sources of uncertainty in dose-response modeling of epidemiologic data for cancer risk assessment. Ann NY Acad Sci 895:212–222. Stayner L, Smith R, Bailer J, Luebeck EG, Moolgavkar SH. 1995. Modeling epidemiologic studies of occupational cohorts for the quantitative assessment of carcinogenic hazards. Am J Ind Med 27:155–170. Stayner L, Steenland K, Dosemici M, Hertz-Piccioto I. 2003. Attenuation of exposure-response curves at high exposures. Scand J Work Environ Health 29:317–324. Thomas DC. 1983. Statistical methods for analyzing effects of temporal patterns of exposure on cancer risks. Scand J Work Environ Health 9:353. Waxweiler RJ, Stringer W, Wagoner JK, Jones J, Falk H, Carter C. 1976. Neoplastic risk among workers exposed to vinyl chloride. Ann NY Acad Sci 271:40–48. Whittemore AS. 1977. The age distribution of human cancer for carcinogenic exposures of varying intensity. Am J Epidemiol 106:418–432. World Health Organization (WHO). 1994. Assessing Human Health Risks of Chemicals: Derivation of Guidance Values for Health-Based Exposure Limits (Environmental Health Criteria 170). Geneva: WHO. Zeise L, Cardis E, Hemminki K, Schwarz M. 1999. Quantitative estimation and prediction of cancer risk: review of existing activities. In: Moolgavkar S, Krewski D, Zeise L, Cardis E, Moller H, editors. Quantitative Estimation and Prediction of Human Cancer Risk. IARC Scientific Publ. No. 131. Lyon: IARC, pp. 11–59.

6

Application of Biomarkers in Cancer Epidemiology MONTSERRAT GARCÍA-CLOSAS, ROEL VERMEULEN, MARK E. SHERMAN, LEE E. MOORE, MARTYN T. SMITH, AND NATHANIEL ROTHMAN

B

iomarkers are increasingly being incorporated into epidemiologic studies of cancer etiology, often referred to as molecular epidemiology (Perera and Weinstein, 1982; Schulte and Perera, 1993; Furberg and Ambrosone, 2001; Bonassi and Au, 2002; Vahakangas, 2003). Biomarkers can be used to enhance exposure assessment, identify key events along the pathway from exposure to disease, determine sources of genetic susceptibility, and categorize tumors into more homogeneous entities at the molecular level (National Research Council, 1987; Perera, 1987; Schulte, 1987; Rothman et al., 1995; Toniolo et al., 1997; Perera, 2000; Rothman et al., 2001). Furthermore, new discovery technologies, including whole genome analysis, mRNA expression arrays, proteomics, and metabolomics/metabonomics (Aardema and MacGregor, 2002; Baak et al., 2003; Hanash, 2003; Sellers and Yates, 2003; Staudt, 2003; Strausberg et al., 2003; Wang et al., 2003) should enable investigators to explore broadly the biologic responses to exogenous and endogenous exposures, evaluate potential modification of those responses by variants in essentially the entire genome, and define tumors at the chromosomal, DNA, mRNA, and protein levels. Clearly, biomarkers also have direct clinical applications for screening and early detection, optimizing and determining response to treatment, and helping predict prognosis. The emphasis of this chapter is the use of biomarkers in the context of etiologic research. We first discuss the development and characterization of biomarkers for use in epidemiologic studies, with a particular focus on understanding the components of variance and their impact on estimates of disease risk. We then discuss key issues in the use of exposure, intermediate end points, genetic susceptibility, and tumor biomarkers; and we review the main epidemiologic study designs that use biomarkers. We conclude with comments on collecting and processing biologic samples for use in molecular epidemiology studies.

BIOMARKER CHARACTERIZATION In this section we present some basic, generalizable concepts concerning the initial evaluation of newly developed biomarkers and sources of biomarker variability and then describe the implications for the design and interpretation of molecular epidemiologic studies.

Initial Evaluation of Biologic Markers for Use in Epidemiologic Studies Potential new biomarkers for epidemiologic research continuously arise owing to advances in understanding disease etiology and molecular laboratory techniques. The applicability of new biologic markers in epidemiologic studies depends on several factors, some of which are inherent to the analytic technique (e.g., sample medium, collection, processing, storage, volume, assay variability, throughput) and others which relate to the biomarker itself (e.g., latency, half-life), its inherent variability among the population of interest (e.g., interperson and intraperson variability), and its ultimate use in epidemiologic analyses (e.g., categoric, continuous). When a promising new biomarker emerges from the laboratory, some basic issues must be addressed before considering its application in human studies. These initial efforts to characterize biomarkers for use in epidemiologic studies have been called transitional studies by some investigators

70

(Hulka, 1991; Hulka and Margolin, 1992; Schulte et al., 1993; Rothman et al., 1995; Schulte and Perera, 1997), a term that serves to heighten awareness about the critical need to understand the determinants of biomarker levels and assays before they are used in molecular epidemiologic studies with precious, nonreplenishable biologic samples. The first issue that must be addressed about a new assay is its accuracy. Does it measure what it purports to measure? This point can be evaluated only when a gold standard is available against which to compare a new assay. For some well established assays there are certified reference materials. With time, gold standards often emerge for new assays, as samples are evaluated by a variety of techniques in one or more laboratories, and consensus is achieved (Gompertz, 1997). For example, the SNP500 Cancer project at the U.S. National Cancer Institute (NCI) is carrying out a resequencing effort of several thousand single nucleotide polymorphisms (SNPs) in genes important for molecular epidemiology research on 102 publicly available DNA samples from anonymous, ethnically diverse populations obtained from the Coriel Biorepository in Camden, NJ (Parker et al., 2004). After bidirectional sequencing of the flanking regions around the SNP in each sample, genotyping assays are developed on one or more platforms. When sequencing and genotyping assays concur, there is an exceptionally high probability that the genotype is accurate. The second issue when evaluating a marker for use in epidemiologic research is to determine its reliability in the laboratory. So long as an assay is reliable, the ordering of subjects by the measure is preserved. Assuming that the assay accuracy and reliability are acceptable, it is then important to define the optimal conditions for collecting, processing, and storing biologic specimens for eventual assay because variation in sample handling can introduce a large variation in assay results, making the measure unsuitable for research (Rothman et al., 1995).

Interindividual and Intraindividual Variability in Biomarker Response Variability in biomarker response for continuous, nonfixed biomarkers has two basic dimensions: the person and time. Variability over time with regard to biomarker response is an attribute at the individual level that is due to temporal changes in personal behavior and/or experiences. As epidemiologic analyses often assume that biomarker levels are a fixed attribute of an individual, rather than a timedependent attribute, temporal variability generally leads to classic measurement error and, as a result, to attenuation of the biomarkerdisease association. It is important to recognize that the temporal variability in biomarkers depends primarily on their half-life and secondarily on the temporal variation in exposure (e.g., continuous exposure at a constant level versus episodic exposure with variable intensity). Biomarkers with relatively short half-lives (e.g., insulin, hormones, water-soluble micronutrients), in general, display more temporal variability than biomarkers with relatively long half-lives (e.g., blood lead levels, protein adducts) as the temporal variance in exposure is dampened over time. It can be shown that for many biologic indicators with half-lives of more than 40 hours, less than 50% of the temporal variance in exposure is transmitted. In contrast, for biomarkers with half-lives of less than 5 hours the dampening of the temporal variance is negligible

71

Application of Biomarkers in Cancer Epidemiology

Quantification of Biomarker Variability The major components of biomarker variability (person and time) can be described by the following random effects model Xij = mx + bi + ei(j) where Xij is the normalized biomarker response for the ith person on the jth day, mx is the true population mean biomarker response, bi is the random deviation of the ith person’s biomarker response from the population mean, and ei(j) is the day-to-day random deviation from the ith person’s mean biomarker response, which is treated as random error. In this model it is assumed that bi and ei(j) are normally distributed and independent with means of zero and variances of s2inter and s2intra, representing the inter- and intraperson components of variance, respectively (Kromhout et al., 1993). The total variance in biomarker response, s2tot, is the sum of s2inter and s2intra. It should be noted that the intra-person component of variance results from two sources: the true temporal variation due to temporal changes in personal behavior and experiences and the sample variation due to collecting, processing, storing, and sometimes analyzing biologic specimens from the same individual on different days. However, it is normally assumed that the sample variance is small compared to the true temporal variance. This basic model can be expanded to estimate additional sources of variation (e.g., sample collection, interbatch and intrabatch laboratory variance) by assuming a hierarchical variance structure (e.g., Rappaport et al., 2002). Estimating the additional components of variance enables one to identify factors having the most effect on the observed analytic variance; then ways to minimize this variance can be sought. It is important, however, to realize that a repeated sampling design is necessary to calculate the major sources of variation and that repeated specimens of the same subject over time must be mutually independent (White, 1997).

Impact of Biomarker Variability on Biomarker–Disease Associations A general measure of the extent of laboratory variation for continuous measurements is the coefficient of variation (CV = standard deviation/mean, expressed as a percentage). Although often used in the evaluation of analytic variance in the laboratory and useful for identifying “bad” batches of samples that need to be reanalyzed or excluded from the data analysis, it does not provide an intrinsic feeling for the impact of the observed variance on the biomarker-disease association and cannot be used to correct measures of association to account for measurement error. A more intuitive and useful measure for evaluating the impact of the total measurement error (natural temporal plus analytic variance) with a continuous biomarker is the intraclass correlation coefficient (ICC), which can be defined as the s2inter divided by s2tot. The ICC is equal to one when all of the variance in the biomarker response can be attributed to differences between persons and no natural temporal or analytic variance is present. If we rewrite the above formula as (s2tot - s2intra)/(s2intra + s2inter), it becomes apparent that the larger the

temporal variance (s intra) is relative to the interperson variance, the lower the ICC is. The degree of acceptable measurement error depends on the magnitude of the true association and the availability and quality of alternative methodologies to estimate exposures (e.g., questionnaires, environmental monitoring, job exposure matrices, food composition tables). However, as a general rule, an ICC <0.5 (e.g., when there is equal variance between subjects as there is within subjects due to the total temporal and analytic variation) would be of limited value in many instances because it would lead to substantial attenuation of the risk estimates (Fig. 6–1) Additionally, the ICC can be used to correct measures of association to account for measurement error. However, as the assumptions used in these correction formulas cannot be known to be correct, the results of such adjustments should not be considered the “true association” but, rather, an indicator of the degree of bias in the observed odds ratio or other measure of association (White, 1997). Measurement error in categorical biomarkers (e.g., binary markers) is usually referred to as misclassification. In the instance of a genetic polymorphism, the biomarker is fixed and there is no temporal variation. Assuming that the accuracy of the assay has been previously established by comparing genotyping results in reference DNA samples with known genotypes, as described earlier, misclassification due to laboratory variability can be expressed by sensitivity and specificity when genotypes are dichotomized. The degree of attenuation due to misclassification in binary biomarkers depends, however, not only on the sensitivity and specificity of the biomarker test but also on the proportion of the nondisease or unexposed group who are true positives. In the case where the proportion of true positives is low, high specificity is more important; whereas when the proportion of true positives is high, high sensitivity is more important (White, 1997). The impact of misclassification can be substantially greater when interactions are modeled (Garcia-Closas et al., 1999; Rothman et al., 1999). For example, Table 6–1 shows that nondifferential misclassification of a binary exposure variable and genotype marker attenuates the estimate of the interaction parameter and increases the sample size requirements, often substantially.

Practical Implications Biomarker responses vary by person and time, and knowledge about these sources of variance is needed to understand their value and applicability in epidemiologic research. Most often, considerable attention is given to the analytic variance, but variation by person and time are often not quantified for continuous markers. Variability within a pop-

4.0

3.5

3.0

Observed OR

(Nieuwenhuijsen and Droz, 2003). Therefore the use of biomarkers with relatively long half-lives is generally more appropriate for epidemiologic purposes, especially when the biomarker can be measured only at a single point in time and not necessarily at the optimal time window related to the studied end point. An exception occurs where the exposure measured by the biomarker is relatively constant over time, in which case the biomarker half-life is less critical (Armstrong et al., 1992b). Another form of measurement error is associated with analytic variation, which can be due to differences in handling, processing, and storing specimens as well as laboratory variation. Preferably, the analytic variation is small relative to the natural temporal and personal variations. However, for some assays significant analytic variation (e.g., oxidative damage in DNA, 32P-postlabeling) has been reported (Phillips and Castegnaro, 1999; Huang et al., 2001), and the impact on the study results may be dramatic.

2

2.5

2.0

1.5

1.0

0.5 1.0

0.8

0.6

0.4

0.2

0.0

Intraclass Correlation Coefficient (ICC)

Figure 6–1. Impact of the intraclass correlation coefficient (ICC) on the observed odds ratio (OR) given true ORs for disease of 1.5, 2.0, 2.5, 3.0, and 3.5. The observed OR was obtained using the formula OR = exp(lnORt*ICC), where ORt is the true OR.

72

PART I: BASIC CONCEPTS

Table 6–1. Minimum Number of Cases (Case/Control Ratio = 1) Required to Detect a Two-fold Multiplicative Gene–Environment Interactiona for Different Levels of Accuracy of Environmental and Genetic Factors Sensitivityb Environmental Factor 1.0 0.8 1.0 0.8

Prevalence Genetic Factor

Environmental Factor

Genetic Factor

Interaction Parameter

No. of Cases

1.00 1.00 0.95 0.95

0.50 0.40 0.50 0.40

0.50 0.50 0.48 0.48

2.00 1.56 1.83 1.46

720 1600 900 2044

Source: Adapted from Garcia-Closas et al. (1999). a Odds ratios for environmental and genetic risk factors alone are 2.0, and the odds ratio for the joint effect of the environmental and genetic factor is 8.0. b Specificity for both genetic and environmental factor assessment is 1.0.

ulation can be evaluated in an analysis of unexposed individuals, and baseline biomarker prevalences can be estimated. Quantification of these sources of variance is, however, essential to quantifying the total measurement error (natural temporal and analytic variance) and placing the magnitude of the measurement error in perspective. For instance, a relatively large degree of measurement error can be tolerated if the interperson differences in the parameter to be measured are large. Alternatively, a biomarker with an extremely small degree of analytic error might be uninformative because of large within-person or small between-person variability. Many of these errors can be measured if one collects and analyzes two (or more) specimens from a group of subjects and inserts split samples (duplicates) in a pilot study to assess the ICC for the laboratory component of error. Care should be taken that samples are collected, stored, and handled in the same way and that, as always, laboratory analyses are blinded for exposure and/or disease status. By estimating the components of variance in a pilot setting the epidemiologist can assess if it is worthwhile to study a particular biomarker in the intended population. During the course of analysis of actual study samples, split and blinded duplicate samples randomly inserted across the sample batches allow an estimate of the ICC due to analytic variation. A rough rule of thumb used by many is that quality control samples should comprise about 10% of the study samples. During data analysis, the ICC allows an estimate of the total measurement error and can be used to quantify the effect on the observed association between the biomarker and disease. Alternatively, the ICC can be estimated by inserting identical quality control samples in duplicate in each batch. This allows one to estimate the between batch and within batch components of variance, whereas the variance observed in all measurements among controls can be regarded as the total variance (between subject, between batch and within batch). The ICC can be calculated using these measures, under the assumption that the observed variance in the quality control sample(s) is generalizable to the range of measures in the whole population. For categorical markers, multiple replicate samples from enough individuals so there are subjects with “positive” and “negative” values can be inserted blindly throughout the batches, and the sensitivity and specificity can be calculated, using the predominant value for each

subject’s samples as the standard. In the special case of genotype assays, where hundreds to thousands of SNPs may be analyzed, including those with low allele frequencies, it is useful to have DNA from a relatively large random sample of study subjects inserted as blind duplicate quality control samples. For example, if samples from 100 study subjects are analyzed a second time as blinded duplicates randomly dispersed throughout the batches, an estimate can be obtained for misclassification of even relatively low frequency SNPs. This is important to estimate, as specificity errors can substantially attenuate a risk estimate for low frequency variables, as discussed previously. Finally, data transmission errors sometimes affect a study if not detected. In this era of high throughput analysis, significant amounts of data can be generated. We have found that, even when assays are essentially perfect, data transmission errors, detected by discrepancies in quality control samples, can occur. The new generation of laboratory information management systems should help reduce such errors in the future.

BIOMARKER CATEGORIES Figure 6–2 presents a commonly used approach to categorize biomarkers that represent, directly or indirectly, the carcinogenic process derived from external exposure to disease (Perera and Weinstein, 1982; National Research Council, 1987). Biomarkers can be broadly classified into those that reflect exposure (internal and biologically effective dose), intermediate end points (early biologic effects and altered structure/function), and cancer. Susceptibility, acquired or inherited, has been shown to potentially modify the relation between each step in the progression from exposure to disease. In this section, we describe in more detail the various biomarkers and their application to epidemiologic studies (Table 6–2).

Exposure Biomarkers Exposure biomarkers measure the level of an external agent, its metabolic by-products in the free state or bound to macromolecules, or the

SUSCEPTIBILITY

Exposure

Internal Dose

Biologically Effective Dose

Early Biologic Effect

Altered Structure/ Function

Disease

Figure 6–2. A continuum of biomarker categories reflecting a carcinogenic process resulting from xenobiotic exposures. (Source: Adapted from National Research Council, 1987.)

73

Application of Biomarkers in Cancer Epidemiology Table 6–2. Types of Biomarkers That Can Be Incorporated into Epidemiologic Studies Type of Biomarkers Exposure

Intermediate end points

Susceptibility Tumor markers

Examples

Biologic Specimens

Endogenous/exogenous compounds (e.g., toxins, nutrients, metabolites, hormones, antibodies, infectious agents) Carcinogen macromolecular adducts (e.g., DNA, protein such as albumin or hemoglobin) DNA alterations (e.g., adducts, damage, somatic mutations, methylation) Cytogenetics (e.g., chromosomal aberrations) Gene expression (transcriptomics)

Serum/plasma, saliva, urine, normal tissue (e.g., skin, cervix, or colon biopsy) Nonneoplastic tissue or cellsa

Levels of specific proteins (e.g., hormones, growth factors, cytokines), proteomics, metabolomics Lymphocyte function (e.g., DNA repair, apoptosis, immune function) Cytotoxicity (e.g., number of lymphocytes) Genotype (i.e., genetic polymorphisms) Phenotype (e.g., DNA repair capacity) DNA alterations (e.g., somatic mutations, methylation) Cytogenetics (e.g., alteration in gene copy number) Gene expression (specific mRNA or transcriptomics) Levels of specific proteins (e.g., hormone receptors) and proteomics

Nonneoplastic tissue or cellsa Nonneoplastic tissue or cellsa Serum/plasma, saliva, urine, nonneoplastic tissue and cellsa Serum/plasma, saliva, urine, nonneoplastic tissue and cellsa Viable/cultured lymphocytes Nonneoplastic cells Genomic DNA Viable/cultured lymphocytes Tumor tissue DNA Tumor tissue DNA Tumor tissue mRNA Tumor tissue or cells

a

For example, white blood cells; nasopharyngeal, oropharyngeal, and bronchial cells in sputum or lavage fluid; urothelial cells in urine; colonic cells in feces; ductal epithelial cells in breast nipple aspirate or fine-needle aspiration.

specific immunologic response it elicits. In addition, exposure biomarkers measure endogenously produced compounds, which may be influenced directly or indirectly by external factors (e.g., hormones) and genetic factors. They are measured in tissues, body fluids, or any combination of these to evaluate internal exposure levels. In some instances, biomarkers complement or serve as an alternative to questionnaire and environmental data in an epidemiologic study, and in other instances they are the only approach available to assess exposure status (e.g., infectious agents, growth factors). A wide range of exposures can be measured biologically, including environmental factors (e.g., dioxins, polychlorobiphenyls, polycyclic aromatic hydrocarbons, aflatoxin, heavy metals), nutrients (e.g., b-carotene, phytoestrogens, folate), infectious agents (e.g., Epstein-Barr virus, human immunodeficiency virus, hepatitis B and C, Helicobacter pylori, SV40), and endogenous compounds (e.g., hormones, growth factors). These measurements have been successfully applied in cancer epidemiology (Coggon and Friesen, 1997; Kaaks et al., 1997a; Munoz and Bosch, 1997; Rothman et al., 1997; Wild and Pisani, 1997; Ketchum et al., 1999; Tang et al., 2001; Wild et al., 2001; Adlercreutz, 2002; Gammon et al., 2002; Krajcik et al., 2002; Riboli et al., 2002; Warner et al., 2002; Lamar et al., 2003; Pavuk et al., 2003; Starek, 2003). The first epidemiologic evaluation of potential biomarkers of exposure generally occurs in cross-sectional studies in the general population or in subgroups with specific, well characterized exposure and lifestyle patterns. Sometimes a biomarker of exposure can be used only in cross-sectional studies to determine if a population is exposed to an agent of concern, or it can be used as an independent marker of exposure in studies evaluating intermediate biomarker end points. Other times, a biologic marker of exposure may have utility only for validating external exposures estimated by questionnaires in casecontrol and cohort studies. In some instances, initial evaluation of an exposure biomarker indicates that it can be used to assess directly the exposure in case-control or (more commonly) cohort studies with prospectively collected biologic samples. The applicability of exposure biomarkers in case-control and cohort studies depends on certain intrinsic features related to the marker itself (e.g., half-life, variability, specificity) and the exposure pattern, as already noted. The first prerequisite for successful application of an exposure marker is that the assay is reliable and accurate, the marker is detectable in human populations, and important effect modifiers (e.g., nutrition and demographic variables) and kinetics are known (Rothman et al., 1995). Second, the timing of sample collection in combination with the biologic half-life of a biomarker of exposure is key, as it determines the exposure time window that a marker of expo-

sure reflects. The time of collection may be critical if, as is often the case in epidemiologic studies, only one sample per subject can be obtained on a given occasion or even during the course of the whole study and if the exposure is of brief duration, is highly variable in time, or has a distinct exposure pattern (e.g., diurnal variation in certain endogenous markers such as hormones) (Rejnmark et al., 2001). Again, chronic, near-constant exposures pose fewer problems. Ideally, the biomarker persists over time and is not affected by disease status in case-control studies. However, most biomarkers of an internal dose generally provide information about recent exposures (hours to days), with the exception of markers of persistent pesticides, dioxins, polychlorobiphenyls, certain metals, and serologic markers related to infectious agents, which may reflect exposures that took place many years before (Table 6–3). Although there are theoretical advantages to the use of biologic exposure markers in epidemiology, the information a biomarker provides must be compared to the availability and quality of other exposure assessment methods (e.g., food-frequency questionnaires, environmental monitoring, job exposure matrices, food composition tables). Essentially all exposure measures misclassify some subjects— it is the relative ability of different sources of information to classify

Table 6–3. Half-lives for Groups of Biomarkers Based on the Chemicals and Their Metabolites Measured in Blood and Urine in the National Health and Nutrition Examination Survey Biomarker

Half-life (hr),a Median and Range

Examples

Aromatic hydrocarbons Chlorinated hydrocarbons Dioxins/furans/coplanar PCBsb Metals Nonpersistent pesticides

2 (0.5–8.0) 4 (0.3–60.0) 22,000–166,500

Benzene, xylene, toluene TCE, PERC TCDD, TCDF

200 (3–87,600) 8 (0.3–33.0)

Lead, beryllium, cadmium 2,4-D, organophosphates, pyrethroid pesticides

PCBs (Noncoplanar)a Persistent pesticides Phytoestrogens Phthalates (Pro)-Vitamins

22,000–44,000 2,124 (100–70,000) 8 (3–10) 9 (6–12) 80 (0.4–1,140)

a

DDT, dieldrin, chlordane Daidzein, genistein Bisphenol A Carotene, folate, retinol

Half-lives extracted from the Hazardous Substances Data Bank (HSDB®) maintained by the National Library of Medicine’s (NLM) Toxicology Data Network (TOXNET®). b No specific half-lives have been determined for most of the dioxins, furans, and PCBs; therefore only the range in half-lives is presented.

74

PART I: BASIC CONCEPTS

people into exposure categories correctly that must be considered. For instance, subjects can generally report average smoking habits and smoking duration in an accurate manner, so cumulative exposure can be calculated. Because internal markers associated with tobacco smoking have relative short half-lives and thus reflect only recent exposure, they have limited utility by themselves to assess a smoking–cancer relation directly. In contrast, it is difficult to obtain accurate information on exposure to aflatoxin by questionnaire because exposure is sporadic and is presented in a spectrum of food items. In this instance, even short-term internal dose markers might be expected to classify the long-term exposure status more accurately than questionnaire data (Kaaks et al., 1997b). If the pattern of exposure being measured is relatively continuous, short-term markers may be applicable in cohort studies with prospectively collected biologic samples or in case-control studies of patients with early disease. In general, however, short-term markers have limited use in case-control studies unless they are to be used as recent phenotypic markers of nutritional or lifestyle behaviors. Generally, they are less likely to reflect usual patterns, and the disease or its treatment might influence absorption, metabolism, storage, and excretion. A special group of exposure markers are carcinogen macromolecular adducts. DNA adducts induced by carcinogenic chemicals reflect exposure and are directly related to tumor formation, whereas modification of protein provides relatively precise dosimetry for stable adducts of proteins with a known half-life (Poirier et al., 2000). The relatively long half-lives of albumin and especially hemoglobin make protein adducts highly promising exposure markers in epidemiologic research (serum albumin and hemoglobin have half-lives of 550 and 3000 hours, respectively). The half-lives of protein and DNA adducts depend on the carrier (e.g., cell type) and the stability of the adduct, and for DNA adducts, the DNA repair rate as well. The half-lives of white blood cells, which are often used to measure DNA adducts, vary considerably, from a few hours for neutrophils (±7 hours) to years for certain subtypes of T lymphocytes. In addition, the size of the cell populations may vary among subjects and can be influenced by a variety of immunologic stimuli. Given these limitations, it is generally difficult to estimate the persistence of DNA adducts. Smoking cessation studies, however, have shown that the overall half-life of bulky DNA adducts in lymphocytes is around 550 hours (Mooney et al., 1995; Godschalk et al., 2002). Therefore if no special effort is made to isolate lymphocyte DNA or even better, long-lived well-defined subsets of lymphocytes, the use of DNA adducts for exposure assessment has little benefit over protein adducts, as their half-lives are comparable or shorter. If DNA adducts can be measured in lymphocytes, however, they pose at least a theoretical advantage by virtue of the fact that they reflect metabolic activation and detoxification and DNA repair. As such, they may integrate external exposure and relevant metabolic processes and provide a measure of the “biologically effective dose” (Perera and Weinstein, 1982; National Research Council, 1987). Many of the environmental exposure markers based on metabolites are not specific to one exposure. Often the measured metabolites are a proxy for the parent or active compound, and the contribution of additional sources could lead to exposure misclassification. For example, trichloracetic acid in urine is a metabolite of 1,1,1trichloroethane, trichlorethylene (TCE), and tetrachloroethylene (PERC) and would not be a useful marker for evaluating the relation between TCE and kidney cancer without additional information about the use of certain chemicals. It is therefore important to evaluate all possible sources that could give rise to the biomarker of interest and to ensure that these factors can or have been assessed satisfactorily to enable some form of correction in the data analysis. In summary, the choice of a particular exposure biomarker requires careful consideration of the period of exposure the biomarker reflects, the expected temporal and personal variability in the biomarker response, the specificity of the biomarker for the studied exposure, and the overall study design. It is certainly not a given that biomarkers of exposure always provide the most accurate and precise estimates of that exposure; hence, classic alternatives should be carefully considered for exposure assessment (Saracci, 1997). Nevertheless, biomarkers of exposure that are collected and used correctly can be extremely

powerful tools in cancer epidemiology. As technology continues to evolve, the physical, chemical, immunologic, and molecular assays to determine exposure biomarkers are becoming more accurate, sensitive, and precise, and they can be performed more rapidly and efficiently (Barr and Needham, 2002). This trend should ultimately allow analysis of large panels of biomarkers using minimal amounts of biologic material.

Intermediate End Point Biomarkers Intermediate biomarkers directly or indirectly represent events on the continuum between exposure and disease. Intermediate biomarkers can provide important mechanistic insight into the pathogenesis of cancer, including early effects that occur proximate to the exposure and subsequent preneoplastic alterations. As such, they complement classic epidemiologic studies that use cancer end points. In addition, intermediate biomarkers can provide initial clues about the carcinogenic potential of new exposures years before cancer develops (National Research Council, 1987; Schatzkin et al., 1990; Schulte et al., 1993; Toniolo et al., 1997; Schatzkin and Gail, 2002). One group of intermediate biomarkers, those of early biologic effect (National Research Council, 1987), generally measure early biologic changes that reflect early, nonclonal, generally nonpersistent effects. Examples of early biologic effect biomarkers include measures of cellular toxicity; chromosomal alterations; DNA, RNA, and protein expression; and early nonneoplastic alterations in cell function (e.g., altered DNA repair, altered immune function). Generally, early biologic effect markers are measured in substances such as blood and blood components (e.g., red blood cells, white blood cells, DNA, RNA, plasma, serum) because they are easily accessible; and in some instances it is reasonable to assume that they can serve as surrogates for other organs. Early biologic effect markers also can be measured in other accessible tissues such as skin, cervical and colon biopsy specimens, epithelial cells from surface tissue scrapings or sputum samples, exfoliated urothelial cells in urine, colonic cells in feces, and epithelial cells in breast nipple aspirates. Other early effect markers include measures of circulating biologically active compounds in plasma that may have epigenetic effects on cancer development (e.g., hormones, growth factors, cytokines). Three new technologies, mRNA expression arrays, proteomics and metabolomics, have recently been used to measure early biologic effects associated with chemical exposures. Expression array analysis of mRNA (or cDNA) offers the possibility of broadly exploring differences in gene expression patterns associated with known or suggested disease risk factors. Although expression array technology should be a useful tool for identifying new intermediate biomarkers, the magnitude of the data produced from each experiment, data analysis, and interpretation remain major challenges. For example, different statistical software packages do not always identify the same sets of differentially expressed genes from the same array data (Irizarry et al., 2003; Smyth et al., 2003); therefore, it may not always be possible to form firm conclusions. Currently, expression arrays should be considered “discovery tools” rather than conventional assays familiar to most epidemiologists and laboratory scientists. It is also currently prohibitively expensive and labor-intensive to perform expression array analysis for every subject in large studies. Rather, one can select a small subset of matched pairs of exposed and unexposed subjects (or subjects with and without preneoplastic lesions) and discover differentially expressed genes. Once several target genes are identified, real-time polymerase chain reaction (PCR) analysis can be used to quantify expression of selected genes in all subjects (Forrest et al., 2005). The greater stability of proteins/peptides compared to RNA and the knowledge that the protein complement is perhaps the true phenotype makes proteomics and metabolomics, perhaps an attractive new technology for studying intermediate end points. Like microarrays, proteomic and metabolomic analysis also produces highly complex data sets, and statistical tools to perform such analyses are not yet standardized. An important challenge relating to proteomic data analysis is the large number of observations per sample (Petricoin et al., 2002b;

Application of Biomarkers in Cancer Epidemiology Qu et al., 2003). A key issue is to reduce the dimensionality of the data to make statistical analyses feasible and to reduce the likelihood of false-positive results (Qu et al., 2003). Sources of variation among samples and within the assay itself also need to be better understood. For maximum utility, an intermediate biomarker must be shown to be predictive of developing cancer, preferably in prospective cohort studies (Schatzkin et al., 1990) or potentially in carefully designed case-control studies of cases with low stage/low grade tumors. The criteria for validating intermediate biomarkers have been discussed by Schatzkin and colleagues (Schatzkin et al., 1990; Schatzkin and Gail, 2002) and focus on calculating the etiologic fraction of the intermediate end point, which varies from 0 to 1. The closer the etiologic fraction is to 1, the more the biologic marker reflects events, directly or indirectly, on the causal pathway to disease. Chromosomal aberrations in peripheral blood lymphocytes have been used extensively as the classic biomarker of early genotoxic effects in cross-sectional studies of populations exposed to a wide variety of potential carcinogens (Tucker et al., 1997; Zhang et al., 1999, 2002). Several small cohort studies have reported that the prevalence of chromosomal aberrations in peripheral lymphocytes can predict the subsequent risk of cancer (Hagmar et al., 1994; Bonassi et al., 1995; Liou et al., 1999; Smerhovsky et al., 2001). The predictive performance of this biomarker was shown to be similar irrespective of whether the subjects had been smokers or occupationally exposed to carcinogenic agents (Bonassi et al., 2000). In contrast, such associations were not observed for the sister chromatid exchange assay, another biomarker of genotoxicity that is also measured in peripheral lymphocytes (Hagmar et al., 1994; Bonassi et al., 1995; Liou et al., 1999). Larger studies are needed to evaluate the relation between chromosomal aberrations and specific tumor sites and to enable calculation of etiologic fractions with precision. The availability of numerous prospective cohort studies with stored blood specimens (Table 6–4) should enhance our ability to test rapidly the relation between a wide variety of early biologic effect markers and cancer risk using both standard and emerging technologies (Nicholson and Wilson, 2003; Tomer and Merrick, 2003). Such studies could ultimately produce a new generation of end points to evaluate the carcinogenic potential and mechanisms of action of various risk factors. In addition, this line of research may one day identify a panel of intermediate markers easily analyzed from blood samples that can be used to identify individuals at elevated risk of developing cancer in the future, who may then benefit from targeted primary and secondary preventive strategies. A second group of intermediate markers represents events farther down the continuum from exposure to disease, where early hyperplastic or preneoplastic alterations may have occurred, sometimes due to clonal expansion of a genetically or epigenetically altered cell. These markers have been referred to as biomarkers of altered structure and function (National Research Council, 1987). Some of these events can be identified by standard histologic techniques, at times enhanced through the use of special methods. More subtle and earlier preneoplastic changes may be detected through proliferation and apoptosis assays as well as molecular analyses that reflect early clonal events in cell cycle control. These markers are frequently analyzed in tissues from organ sites of interest. Testing the relation between intermediate end points in specific solid organ sites and the subsequent development of cancer or clear pathologically defined cancer precursors presents specific logistic challenges. One and preferably several samples must be collected over time from the organ site in an initially healthy study population, which is then followed up for an extended period of time. These studies are often referred to as natural history studies and have been carried out successfully for several cancer sites, particularly cervical cancer (Schiffman and Adrianza, 2000; Bosch et al., 2002; Cuzick, 2002). They are especially attractive studies in that the whole range of intermediate end points can be evaluated in the target organ site. Studies of morphologically “normal” cells and tissues may reveal how particular exposures and lifestyle factors induce cellular responses, such as adaptation with preservation of homeostasis, nonlethal damage resulting in increased cancer risk, nonlethal injury precluding future repli-

75

cation, and cell death via necrosis and apoptosis. Also, biologic markers of altered structure and function can be used to detect the earliest evidence of altered cell cycle control; and when ethically appropriate, these end points can be linked to the development of clinically relevant preneoplastic conditions.

Susceptibility Biomarkers Family history traditionally has been used in epidemiologic studies as a crude marker for inherited susceptibility to cancer; however, identification of specific susceptibility factors requires the use of biomarkers. Susceptibility biomarkers can be measured at the genotypic level (variations in DNA base sequences) or at the functional/phenotypic level (e.g., metabolic phenotypes, DNA repair capacity). Whereas phenotypic measures are closer to the disease process and can integrate the influences of multiple genetic and posttranscriptional influences on protein expression and function, genotypic measures are considerably easier to study because they are stable over time and less prone to measurement error (Ahsan and Rundle, 2003). In addition, the identification of most functional genes in the human genome and variations in the genetic sequences among individuals by the Human Genome Project, as well as the development of genotype technology, have enormously facilitated the study of genetic variants. Thus, from the logistical point of view, genotype assays are usually preferred to phenotypes. However, when phenotypic variation is determined by complex combinations of genetic variants and/or important posttranscriptional events, phenotypic assays become the optimal approach to capture variation in the population that might be relevant to cancer susceptibility. A discussion of rare genetic variants associated with large effects on uncommon hereditary forms of cancer (e.g., BRCA1/BRCA2 and hereditary breast and ovarian cancer) is outside the scope of this section. Here we focus on common genetic variants or polymorphisms (i.e., a minor allele frequency of more than 0.01), with smaller individual effects on more common, nonhereditary forms of cancer that are mainly driven by environmental exposures, broadly defined (Caporaso and Goldstein, 1995). The study of genetic polymorphism effects on cancer risk can enhance our understanding of the relation between environmental exposures and cancer by: (1) providing mechanistic insights into cancer etiology when the effects of established risk factors are evaluated among people with different genetic variants; (2) uncovering effects of environmental exposures on cancer risk when the effect of exposure is mainly or only present in small susceptible subgroups in a population; and (3) discovering new etiologic pathways to cancer that are mediated by alleles found to be associated with cancer (Rothman et al., 2001). The most common type of genetic variation are single nucleotide polymorphisms (SNPs), of which more than several million have already been identified (Altshuler et al., 2005). Other important but less common variations include deletions, insertions, variable tandem repeats, and gene amplifications. Initially epidemiologic studies of genetic polymorphisms only evaluated one or a few promising candidate genes. These efforts have led to identification of a few consistent associations, such as GSTM1 null genotype and NAT2 slow acetylation with bladder cancer risk (Garcia-Closas et al., 2005). Technologic advances are enabling researchers to move beyond evaluating only a few genetic variants to a more comprehensive evaluation of thousands of variants (Shen et al., 2005; Hardenbol et al., 2005) in important etiologic pathways (e.g., carcinogen activation and detoxification, DNA repair, inflammation, apoptosis) and to perform whole genome scans to screen for susceptibility markers (Lander, 1996; Risch and Merikangas, 1996; Hirschhorn and Daly, 2005; Marchini et al., 2005). A comprehensive evaluation of variants in groups of genes involved in etiologic pathways is a promising approach because functional alterations in many proteins encoded by different genes in a pathway are probably needed to have a substantial effect on carcinogenesis (Thomas et al., 2005). For instance, alterations in several key DNA repair proteins are likely needed to affect DNA repair capacity and through this mechanism affect cancer risk (Mohrenweiser et al., 2003).

76

PART I: BASIC CONCEPTS

Table 6–4. Selected Cohort Studies with Blood Sample Collections of Risk Factor Questionnaire Dataa Year Blood Year Blood Approximate Collection Began Collection Ended Study Size

Study

Location

Health Professional’s Follow-up Study

USA

1974

1976

18,000

Physician’s Health Study I

USA

1982

1984

15,000

NYU Women’s Health Study

USA

1985

1991

14,000

Northern Sweden Health and Diseases Study Shanghai Men’s Cohort

Sweden

1985

Ongoing

90,000

China

1986

1989

18,000

Italy

1987

1992

11,000

Hormones and Diet in the Etiology of Breast Tumors study (ORDET) Atherosclerosis Risk in Communities Study (ARIC) CLUE II Nurses’ Health Study I

USA

1987

1998

16,000

Maryland (USA) USA

1989 1989

1989 1990

33,000 33,000

Melbourne Collaborative Cohort Study

Australia

1990

1994

42,000

Japan Public Center-Based Prospective Study on Cancer and Cardiovascular Diseases (JPHC) Alpha-Tocopherol Beta-Carotene Cancer Prevention Study Group (ATBC) Women’s Health Study

Japan

1990

2000

60,000

Finland

1991

1992

20,000

USA

1992

1995

27,000

Women’s Health Initiative

USA

1993

1998

162,000

Korean Multicenter Cancer Cohort (KMCC) European Prospective Investigation into Cancer and Nutrition (EPIC) Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial (PLCO) Beta-Carotene and Retinol Efficacy Trial (CARET)

Korea

1993

Ongoing

17,000

Europe

1993

1999

389,000

USA

1993

2001

72,000

USA

1995

1996

12,000

Nurses’ Health Study II

USA

1996

1999

30,000

Shanghai Women’s Health Study

China

1997

2000

57,000

Cancer Prevention Study (CPS)-II LifeLink Cohort Proyecto Coyoacan

USA

1998

2001

39,000

Mexico City (Mexico) Hawaii (USA)

1999

2004

200,000

2001

Ongoing

39,000

Multiethnic Cohort Study

California (USA) Shanghai Husband’s Cohort Incheon Health Examinees Cohort Southern Communities Cohort UK Biobank CAPM/Oxford Kadoorie Study of Chronic Disease in China

China Korea Southeastern USA UK China

46,000 2001 2001 2002 2005 2004

Ongoing Ongoing Ongoing Ongoing Ongoing

45,000 18,000 42,000 500,000 500,000

References Walter Willett, personal communication; Giovannucci et al. (1999) Julie Buring, personal communication; Steering Committee of the Physicians’ Health Study Research Group (1989) Paolo Toniolo, personal communication; Toniolo et al. (1995); ZeleniuchJacquotte et al. (2003) Göran Hallmans, personal communication Mimi Yu and Jian-Min Yuan, personal communication; Ross et al. (1992) Franco Berrino, personal communication; Berrino et al. (1996) Gerardo Heiss, personal communication; ARIC Investigators (1989) Kathy Helzlsouer, personal communication Susan Hankinson, personal communication; Colditz (1995) Graham Giles, personal communication; Giles and English (2003) Shoichiro Tsugane, personal communication; Tsugane and Sobue (2001) Demetrius Albanes, personal communication; Malila et al. (2002); Woodson et al. (2003) Julie Buring, personal communication; Rexrode et al. (2000) Marcia Stefanick, personal communication; Anderson et al. (2003); Rossouw et al. (2003) Keun-Young Yoo, personal communication; Yoo et al. (2002) Elio Riboli, personal communication Riboli E et al. (2002) Richard Hayes, personal communication; Hayes et al. (2000) Lars Berglund, personal communication; Bowen et al. (2003); Neuhouser et al. (2003) Susan Hankinson, personal communication; Rich-Edwards et al. (2002) Wei Zheng, personal communication; Zheng et al. (2005) Carmen Rodriguez, personal communication; Calle et al. (2002) Sarah Clark and Rory Collins, personal communication Laurence Kolonel, personal communication; Kolonel et al. (2000) Laurence Kolonel, personal communication; Kolonel et al. (2000) Wei Zheng, personal communication Yun-Chul Hong, personal communication Wei Zheng, personal communication Ollier et al. (2005) Zhengming Chen, personal communication Chen et al. (2005)

a

Includes cohorts that had begun or were planning to begin collection in 2004 and that have at least 10,000 subjects with the basic risk factor questionnaire being followed up for cancer incidence. They were limited to broad cohorts that are being followed for most or all cancer end points, with blood samples that include white blood cells.

Several resequencing and assay validation efforts in candidate genes facilitate the study of common variation in genes implicated in candidate pathways. For instance, an important source of information is provided by a public website for the Cancer Genome Anatomy Project (CGAP) (http://cgap.nci.nih.gov) of the National Cancer Institute. The CGAP website provides researchers with access to genomic data, informatic tools to query and analyze the data, and information on methods and resources for reagents developed by the project. The SNP500Cancer project (Packer et al., 2004) noted earlier is an impor-

tant tool provided by CGAP for the identification and validation of known or newly discovered SNPs and other important classes of genetic variants of potential importance to studies of cancer and other diseases. Specifically, data on the SNP500Cancer web site (http://snp500cancer.nci.nih.gov) can be used by molecular epidemiologists to select SNPs for analysis, set up genotype assays using the validated sequence data, use selected assay conditions that have already been validated on one or more platforms, identify and obtain DNA samples with sequence-verified genotypes of interest for use as

Application of Biomarkers in Cancer Epidemiology quality control samples to establish the accuracy of genotyping platforms, and estimate common haplotypes in selected genes. Other important resources of information include the Environmental Genome Project (EGP) of the National Institute of Environmental Health Sciences which focuses on genes thought to play a role in susceptibility to environmental exposures (http://www.niehs.nih.gov/ envgenom/), and the SeattleSNPs of the National Heart Lung and Blood Institute which focuses on pathways important for inflammatory responses in humans (http://pga.mbt.washington.edu/). The number of haplotypes (i.e., distinct combinations of SNPs that co-occur in a chromosome) in the human genome represents only a small fraction of the theoretically possible combinations assuming random distribution of SNP alleles. This is because alleles at nearby sites can segregate together and thus appear on the same haplotype more often than expected by chance; that is, the alleles are in linkage disequilibrium (LD). Several investigators have proposed that the human genome is organized in regions of limited recombination, or haplotype blocks, separated by small regions with higher recombination rates (Gabriel et al., 2002; Wall and Pritchard, 2003). This haplotype-block model can facilitate the selection of SNPs for epidemiologic studies because the haplotypic variation in each block can be determined by a small set of SNPs or haplotypetagging SNPs (htSNPs), which can be an efficient method for screening for the association between candidate genes and cancer. Several groups of investigators are currently working on identifying htSNPs in cancer-related genes such as the Environmental Genome Project which focuses on environmental response genes (Livingston et al., 2004). In a large effort, the U.S. National Human Genome Research Institute has recently completed (Altshuler et al., 2005) the International HapMap Project (http://www.hapmap.org), which represents a major effort to create a public database of common variation in the human genome. However, the usefulness of HapMap for the discovery of susceptibility genes in complex diseases has been questioned because the haplotypes being mapped will include only common SNPs, so it may have limited use for diseases caused by combinations of rarer alleles (Couzin, 2002; Lai et al., 2002; Cardon and Abecasis, 2003; Zeggini et al., 2005). Several algorithms have been developed for selection of an optimal set of informative SNPs using publicly available databases or data from dense surveys of SNPs genotyped in a relatively small panel of subjects (Stram, 2004; Carlson et al., 2004; Wang et al., 2005). These methods are allowing investigators to efficiently screen for associations between common genetic variation and cancer risk in epidemiological studies. Genome-wide association studies have been proposed as a powerful tool for discovering disease-causing genes (Lander, 1996; Risch and Merikangas, 1996; Hirschhorn and Daly, 2005; Marchini et al., 2005). Such studies use a dense map of SNPs across the genome to account for the variation of large genomic regions or the whole genome (Collins et al., 1999; Kruglyak, 1999). As high throughput multiplexed SNP analysis technologies have become affordable, investigators are beginning to perform large-scale genome scans for cancer and other diseases in large epidemiologic studies. When available, functional/phenotypic assays can be used to clarify genotype–phenotype relations. For instance, in vivo metabolic phenotype determinations have been used to determine the genetic basis for an individual’s ability to metabolize carcinogens (e.g., acetylation phenotype and NAT2 genotype) (Pelkonen et al., 2003). When the genetic basis for some laborious metabolic phenotype analyses used in early epidemiologic studies was elucidated, these analyses were replaced by genotypic assays. However, for phenotypes reflecting posttranscriptional events such as induction/inhibition of enzymes, phenotypic assays are still the only means to capture the metabolic variability in the population (Pelkonen et al., 2003). A number of studies have assessed the role of DNA repair capacity (DRC) regarding cancer risk by using in vitro phenotypic assays mostly on circulating lymphocytes (e.g., mutagen sensitivity, Comet assay, and host cell reactivation assay). These studies have shown differences in DNA repair capacity between cases and controls; however, one should be cautious when interpreting these results because of common limitations such as small sample size, use of convenience

77

controls, use of lymphocytes to infer DRC in target tissues, the possible impact of disease status on assay results, and confounding by unmeasured or poorly measured risk factors that influence the assay (Berwick and Vineis, 2000; Spitz et al., 2003). DNA repair assays also are being used to study differences in repair capacity among cell lines with known genotypes and for determining which genetic variants or variant combinations contribute to interindividual variation in DNA repair capacity at the population level (Mohrenweiser et al., 2003). Knowledge of genotype-phenotype relations is extremely useful for guiding the selection of genetic variants (or combination of variants) that are most likely to affect the function of a pathway and disease risk. However, it is unclear whether even complex combinations of genotypes can explain the phenotypic variation observed in DNA repair; thus, genotypic assays might not completely replace DRC assays. The application of functional assays in large-scale epidemiologic studies requires the development of less costly and laborintensive assays, assessment of short- and long-term effects of cell cryopreservation on assay results, and improvement in the specificity and reliability of the assays.

Issues in Analysis and Interpretation of Results The effect of genetic variants on disease risk can be evaluated by estimating the main effect of a genetic variant, subgroup effects (e.g., resulting from an environmental exposure in subjects with a putative susceptibility genotype), and gene–gene or gene–environment interactions. Subgroup effects might help establish an association when it is only or mainly present in a relatively small subgroup, and it might allow better characterization of the dose–response relation. However, subgroup analyses can be problematic because of the increased chance of finding false-positive results, which is not always taken into account during analysis and interpretation of the data (Garcia-Closas et al., 2003; Wacholder et al., 2004). The assessment of differences in the magnitude of an association within population subgroups requires a formal test for statistical interaction. When considering interactions in the statistical sense, one must specify the scale of measurement (additive or multiplicative) or, equivalently, whether one refers to a modification of the absolute or relative effects on cancer risk. Most studies tend to evaluate multiplicative gene–environment interactions because of the common use of the relative risk. However, with the exception of situations where the effect on disease is limited to a particular subgroup or subgroups, and despite intense debate in the epidemiologic literature, there seems to be no clear etiologic or biologic implication that can be drawn from the presence or absence of a statistical interaction in complex diseases such as cancer. Additive interactions have clear public health implications because they imply that elimination or reduction of the risk factor has a different impact on the reduction of disease burden in population subgroups (Greenland and Rothman, 1998). There are no parallel direct public health implications for multiplicative interactions, except through the implication of supraadditive interaction by a multiplicative joint effect or supramultiplicative interaction. Sample size considerations are critical for the design of studies of genetic effects and gene–environment interactions (Garcia-Closas and Lubin, 1999). Generally, sample size requirements are large (hundreds to thousands of subjects) because the expected effects of individual genes are small and genetic variants of interest may be relatively uncommon. The study size to evaluate gene–environment interactions is also increased by the presence of errors measuring environmental and/or genetic exposures, even when the errors are small (GarciaClosas et al., 1999; Deitz et al., 2000) (Table 6–1). In addition, misclassification leads to biased estimates of risk (Armstrong et al., 1992a). Thus, high-quality exposure assessment and almost perfect genotype determinations are required for the evaluation of gene–environment interactions. This highlights the importance of validating genotype assays and including quality control samples during genotype determinations to assess the reproducibility of the assays. A strategy to reach the large sample sizes required to evaluate the main effects of less common SNPs and gene–environment interactions, especially when considering histologic subtypes of cancers, is to create consortiums of existent case-control or cohort studies to coordinate the analysis of pooled data from different studies. The InterLymph

78

PART I: BASIC CONCEPTS

Consortium to coordinate analyses of case-control studies of non-Hodgkin’s lymphoma (Rothman et al., 2006; Boffetta et al., 2004; Rothman et al., 2004), the International Project on Genetic Susceptibility to Environmental Carcinogens (Taioli, 1999), The Breast and Prostate Cancer Cohort Consortium (The National Cancer Institute Breast and Prostate Cancer Cohort Consortium, 2005) and other efforts (Ioannidis et al., 2005) are examples of such initiatives. Failure to account adequately for ethnicity in the study design or analysis can bias estimates of allele effects on disease risk when both the rate of disease and the allele frequency vary across ethnicities and correlate with each other. This special type of confounding bias has been referred to in the genetics field as “population stratification.” (Wacholder et al., 2002b, Thomas et al., 2002) Wacholder et al. (2002b) argued that in well conducted epidemiologic studies controlling for self-reported ethnicity the bias from population stratification when estimating the main effect of a candidate allele is likely to be small and even smaller when estimating interactions because the biases in each stratum are likely to be similar and to cancel each other out. They also pointed out that the bias tends to be smaller in a setting where there are many ethnic groups or intermarriage across groups. Others have advocated the use of genomic control correction methods in the presence of population stratification (Marchini et al., 2004; Pritchard et al., 2001; Freedman, 2004). An important problem in epidemiologic studies of genetic variants has been the large number of “statistically significant” findings that are not replicated in subsequent studies (Anonymous, 1999; Ioannidis et al., 2001; Hirschhorn et al., 2002). We (Wacholder et al., 2002b, 2004; Garcia-Closas et al., 2003) and others (Browner and Newman, 1987; Sterne and Davey, 2001; Colhoun et al., 2003) have noted that one of the most important determinants of frequent failure to replicate findings is the low probability that a single polymorphism in a complex pathway is truly related to disease, especially when its function is unknown. In fact, consistent findings in the current literature tend to be for genes known to be relevant to the etiology of disease and for variants with known functional importance (e.g., GSTM1 null genotype and NAT2 slow acetylation and bladder cancer risk GarciaClosas et al., 2005). Other determinants of false-positive findings are the power (Browner and Newman, 1987; Sterne and Davey, 2001; Colhoun et al., 2003; Garcia-Closas et al., 2003) and magnitude of the p value (Goodman, 1999; Cox, 2001; Sterne and Davey, 2001; Colhoun et al., 2003; Garcia-Closas et al., 2003; Wacholder et al., 2004). As investigators move from assessing a few variants with known functional significance to assessing a large number of variants with unknown functional significance, the problem of false-positive findings inevitably worsens. Wacholder et al. (2004) proposed a simple approach based on the probability of false-positive reports to decide whether a finding deserves attention, or is “noteworthy.” This strategy should facilitate the interpretation of reports from studies evaluating newly discovered genetic variants, as most of them have a low prior probability of being truly related to disease. There is continuing discussion and debate on alternative methods to identify robust associations when evaluating genetic variants (Thomas, 2004; Matullo, 2005). For instance, methods that correct p-values for multiple comparisons such as the false discovery rate defined as the expected ratio of erroneous rejections of the null hypothesis to the total number of rejected hypotheses (Benjamini et al., 2001) have also been proposed. Ultimately, replication of results in large, well conducted studies are critical for establishing associations. Current epidemiologic studies have applied standard analytic methods to estimate the effects of genetic variants and gene–environment interactions on disease risk. Although this is adequate when only a few variants with high prior probability of being associated with disease are evaluated, the analysis of larger amounts of information derived from the comprehensive evaluation of variants in etiologic pathways requires different strategies. For instance, one can use computational algorithms to predict the impact of combinations of variants on the function of a pathway; however, this approach requires information on kinetic parameters for each step in the pathway (Mohrenweiser et al., 2003). Hierarchical models can be used to improve estimates of risk when evaluating multiple genetic polymor-

phisms (Greenland, 1994; Steenland et al., 2000; De Roos et al., 2003) using information on characteristics of the polymorphisms being evaluated, such as etiologic pathway(s) where the gene is involved, enzyme kinetics, and the prior probability of being associated with disease.

Conclusions Rapid developments in genetic and molecular techniques are offering important opportunities to study susceptibility markers in epidemiologic studies. The success of this effort depends on large studies that are able to replicate results, pool data across studies when necessary, validate genotype assays, use efficient screening techniques such as highly multiplexed genotyping, establish haplotypes for genes involved in carcinogenic processes, and improve analytic tools. Whole genome scans are a promising approach to screening for genetic susceptibility markers in large areas in the genome; and are now being applied to epidemiologic studies. When phenotypic variation cannot be explained by combinations of genotypes, phenotypic/functional assays can be powerful tools for identifying susceptibility factors. However, their large-scale application in epidemiologic studies remains challenging.

Tumor Biomarkers Tumor markers can serve several purposes in cancer research, including early detection and diagnosis of cancer, determination of treatment strategies, enhancement of prognosis, and etiologic research. In addition, markers observed in normal, hyperproliferative, or preneoplastic tissues in healthy people can be used as measures of intermediate end points, as discussed earlier in the chapter. This section focuses on the use of tumor markers to identify etiologic heterogeneity in cancer research. Identification of etiologically distinct tumor types remains relatively unexplored in epidemiologic studies. Tumor marker studies have often been difficult to interpret because of weakness in the study design (e.g., limited statistical power, selected case series rather than population-based studies) and data analysis (e.g., failure to consider tumor stage and grade and possible confounding factors). This area of research is being facilitated by improvements in epidemiologic study designs with large collections of tissue samples and corresponding pathology and medical reports, technical advances in the detection of tumor markers using high throughput analyses, and development of innovative statistical methods to distinguish exposure effects among tumor subtypes (Chatterjee, 2004). Histopathologic classification of tumors by their appearance after hematoxylin and eosin staining under light microscopy provides a visual “gestalt” that encompasses many biologic events during tumor development that lead to distinct tumor morphologies. This has been a powerful tool for cancer diagnosis/treatment for more than 100 years and has more recently demonstrated its capacity to distinguish between etiologically distinct tumors in organ systems. Recognized differences in age-specific incidence rates and temporal trends for cancers stratified by histologic type support the concept that cancers arising in many organs are etiologically diverse (Anderson et al., 2004). In addition, a substantial body of evidence suggests that risk factors for many cancers differ by histologic type; for instance, smoking increases the risk of squamous cell carcinoma of the cervix and decreases the risk of adenocarcinoma (Lacey et al., 2001), and smoking is more strongly associated with squamous and small-cell carcinomas of the lung than with adenocarcinomas or other cell types (Lubin and Blot, 1984; Wynder and Hoffmann, 1994). One rationale for classifying tumors beyond their histopathologic appearance using tissue markers is to identify etiologically distinct subgroups of tumors that appear morphologically similar. Another application would be to subclassify tumors based on the presence or absence of a tumor marker, irrespective of histologic appearance. The use of tumor markers in epidemiologic studies can reduce misclassification of disease, thereby strengthening the causal inference. Detailed understanding of this heterogeneity might help identify associations between risk factors and specific subtypes of cancers that would be diluted or entirely masked without stratification by tumor markers. This area of research can also provide insights on the etio-

Application of Biomarkers in Cancer Epidemiology logic mechanisms leading to cancer. Protocols for collecting, preparing, and testing pathologic specimens are rapidly changing, but detailed descriptions are outside the scope of this section. Instead, we provide an overview of classes of tumor markers, emphasizing general principles and using specific examples for illustration. Tumor markers can be defined as any biologic product related to the development and progression of tumors measured at the DNA, messenger RNA (mRNA), or protein levels. Proteins, rather than DNA or RNA, are the molecules executing most cellular functions, so direct measures of protein levels and function might be the best representation of tumor and normal cellular function. However, protein and mRNA expression are subject to rapid biologic fluctuations that may reflect tumor biology at the time of sample collection rather than effects of exposures that occur before tumor initiation and/or development. Therefore, more stable genetic and epigenetic changes may prove to be better markers of etiologic heterogeneity. Both genetic changes (e.g., mutations and cytogenetic abnormalities) and epigenetic changes (e.g., altered transcription secondary to methylation and acetylation) in tumors that affect gene expression may provide important etiologic clues. It has been postulated that certain chemical carcinogens can cause specific somatic mutations in oncogenes and tumor suppressor genes, leaving a DNA “fingerprint” that can help identify tumors caused by this mechanism. Over the last decade, one of the most studied genes in epidemiology has been the TP53 tumor suppressor gene because of its multiple important roles in carcinogenicity and the high frequency and broad spectrum of mutations observed in most types of cancer. A database of all somatic TP53 mutations reported to date maintained at the International Agency for Research on Cancer (IARC) has been particularly useful for detecting relations between particular types of cancer, mutations, and exposures (http://www.iarc.fr/p53) (Hollstein et al., 1999; Hainaut and Pfeifer, 2001; Olivier et al., 2002). Studies of TP53 mutations as DNA fingerprints of exposures in several tumor types support the rationale of inferring environmental exposures from tumor gene mutations (Greenblatt et al., 1994). An example is the G : C Æ T : A transversion mutation in codon 249 of TP53 in hepatocellular carcinomas, which has been attributed to dietary aflatoxin B1 exposure in high risk populations (Staib et al., 2003). Another commonly cited example is the high frequency of G : C Æ T : A transversion mutations of TP53 found in lung tumors from smokers (Hainaut and Pfeifer, 2001; Pfeifer and Hainaut, 2003). Endogenous or exogenous exposures might increase the risk of cancer by causing epigenetic changes such as DNA hyperpmethylation of cytosine in CpG-rich promoter regions in tumor suppressor genes, leading to loss of function and tumorigenesis (Esteller, 2003; Jones, 2003; Moore et al., 2003). For instance, low dietary folate and high alcohol intake may be related to promoter hypermethylation of genes involved in sporadic colorectal cancer (van Engeland et al., 2003). Theoretically, methylation silencing of a gene should also be detectable by measuring mRNA transcripts or with protein assays. However, loss of gene expression may occur through several mechanisms, and gene methylation may be a marker associated with certain exposures. Therefore, understanding genetic and epigenetic mechanisms could reveal etiologic factors that would not be apparent using laboratory methods that assessed the transcriptome or the proteome. Techniques for DNA methylation profiling represent a promising area for future etiologic research because patterns tend to be stable and therefore more likely to represent the cumulative effects of chronic exposures (Laird, 2003). Some DNA methylation techniques have been adapted to DNA from formalin-fixed paraffin-embedded tissue samples and do not require DNA from fresh or freshly frozen tissues, greatly facilitating its application in epidemiology. Comparison of chromosomal alterations in tumor tissues from subjects with different exposures might provide insight into the molecular mechanisms by which exposures promote cancer. For instance, high levels of arsenic exposures have been found to be associated with an increased number of chromosomal gains and losses detected by comparative genomic hybridization (CGH) in bladder tumors (Moore et al., 2002). This could reflect the fact that arsenic-related tumors are less genetically stable than tumors unrelated to arsenic exposure.

79

Conventional cytogenetic techniques such as fluorescent in situ hybridization (FISH) have enabeled detection of chromosomal sites that might contain critical genes in hematologic malignancies and other types of cancer, such as specific translocation breakpoints and subtypes of leukemia and lymphoma (Rowley, 1999). This method has been rather successful in determining chromosomes frequently associated with particular types of cancer. Newer methods such as CGH (Kallioniemi et al., 1992; Kallioniemi et al., 1993) and spectral karyotyping (SKY) (Liyanage et al., 1996) can facilitate the assessment of numerical alterations in specific chromosomal regions or whole chromosomes (Tachdjian et al., 2000; Nath and Johnson, 2000; Swansbury, 2003). Tumor samples can also be characterized at the mRNA or gene expression level by detecting specific RNA transcripts (e.g., by quantitative reverse transcription PCR assays and FISH) or by simultaneously detecting transcripts of thousands of genes (e.g., using cDNA microarrays and oliognucleotide arrays) (Lobenhofer et al., 2001; Simon et al., 2002). Gene expression profiles have already revealed differences between tumor types and have been used to identify subtypes of morphologically homogeneous tumors (Alizadeh et al., 2000; Perou et al., 2000). Confirmation of key findings from expression arrays using other techniques, such as quantitative reverse transcription PCR assays designed for specific RNA transcripts, immunohistochemistry to detect protein products, and comparison with changes in gene copy number are desirable (Monni et al., 2001). Although expression profiles are finding important clinical uses (Lakhani and Ashworth, 2001), their ability to distinguish etiologically diverse tumors is only now being explored. Both logistic and theoretical considerations may limit their use in etiologic research. Tissue collection requires special processing protocols (e.g., freshly frozen or alcoholfixed tissue) for the preservation of mRNA, which are difficult to implement in population-based research. Also, mRNA expression represents only a snapshot of sometimes rapidly fluctuating biologic processes and thus may not directly reflect exposures that lead to tumor development. However, this approach should help identify genes that contribute to the cancer phenotype itself, and further analysis of somatic mutations and SNPs in key genes should provide important insights into both environmental exposures and inherited susceptibility. Detection of protein expression in cells and tissues by immunochemistry (IHC) has proven to have important implications for the treatment and prognosis of many cancers; however its etiologic relevance is less certain. For instance, whereas the expression of estrogen and progesterone receptors in breast tumors has clear implications for treatment, their use to distinguish etiologically diverse tumors has been suggested but not clearly established (Althuis et al., 2003). IHC methods can be used to detect proteins that have been altered by various biologic processes. For instance, wild-type p53 protein is often undetectable by IHC because of its short half-life, whereas mutated p53 protein often has increased stability, leading to intracellular accumulation and IHC detection. Although both mutation analysis and IHC can be used to assess p53 and many other proteins, these assays assess different outcomes and do not always yield concordant results. For instance, changes in expression of p53 protein detected by IHC may be caused by posttranscriptional events, and nonsense mutations may lead to an absence of protein product that would be detected as wildtype p53 using IHC (Greenblatt et al., 1994). Proteomic analysis can be used to map protein profiles in tumor cells (Graves and Haystead, 2002; Cutler, 2003; Sellers and Yates, 2003). Early detection studies of ovarian, prostate, breast, and other cancers have greatly heightened interest in proteomics technology as a potential screening method for cancer and other diseases in the future (Petricoin et al., 2002a; Wulfkuhle et al., 2003). In summary, tumor markers might be able to identify etiologically diverse tumors beyond histologic classification, including mutational spectra in tumor suppressor genes and oncogenes as well as gene expression arrays. The use of molecular/genetic profiling in etiologic research remains a promising, yet still largely unexplored approach. The relevance of tumor markers in etiologic research using currently available technologies and promising new technologies highlights the

80

PART I: BASIC CONCEPTS

importance of collecting tissue specimens in epidemiologic studies. Paraffin-embedded tissue blocks are often the most accessible source of tissue, as they are routinely prepared in pathology laboratories for clinical purposes. Other specialized tissue-processing protocols may be required to perform assays of interest (e.g., tissue fixatives that preserve RNA or snap-frozen tissue samples may be required to obtain high quality mRNA for gene expression arrays). However, it is unclear if processing beyond protocols commonly used in clinics is warranted at this time for etiologic cancer research.

INCORPORATION OF BIOMARKERS IN EPIDEMIOLOGIC STUDIES Cross-Sectional Studies with Biomarker End Points Cross-sectional or short-term longitudinal designs are investigations usually performed on healthy subjects exposed to particular exogenous or endogenous agents where the biomarker is treated as the outcome variable. These studies focus on exposure and intermediate end point biomarkers and often evaluate genetic and other modifiers of the exposure–end point relation. Cross-sectional studies are often used to answer questions about whether a given population has been exposed to a particular compound, the level of exposure, the range of the exposure, and the external and internal determinants of the exposure. Also, a repeat sample design is sometimes used to assess intraindividual variation in exposure levels and/or outcome variables. The design of these studies is generally straightforward in that protocols for collecting and processing biologic samples such as urine and blood are usually well established. Cross-sectional studies can be used to evaluate intermediate biologic effects from a wide range of exposures in the diet and environment as well as from lifestyle factors such as obesity and reproductive status. This design can be used to provide mechanistic insight into well established exposure–disease relations and to supplement suggestive but inconclusive evidence of the carcinogenicity of an exposure. Furthermore, it can evaluate whether there are early biologic perturbations caused by new exposures or recent changes in lifestyle factors that have not been present long enough to have been evaluated for their association with cancer (Rothman et al., 1995). A distinct advantage of the cross-sectional study is that detailed and accurate information can be collected on current exposure patterns, potential confounders, and effect modifiers. To take advantage of a wide range of potential analytic approaches, particularly those that require cell culturing, extensive processing often within a short period of time after collection is often needed. Interpretation of results from these studies is based on the assumption that the intermediate end points reflect biologic changes considered relevant to cancer development. This may be based on in vitro and animal models or on previous observations that the biomarker is altered in human populations exposed to known carcinogens. In and of themselves, however, these studies are not capable of directly establishing or refuting a causal relation between a given exposure or a given level of exposure and risk for developing diseases. Results of studies using most intermediate biomarkers as outcome measures are only suggestive; a biomarker may be overly sensitive (i.e., it may respond to low levels of chemical exposures that are below the disease threshold, if one exists) or insensitive; or it may reflect phenomena that are irrelevant to the disease process or fail to reflect important processes involved in the pathogenesis of disease. For maximum utility, as discussed previously, an intermediate biomarker must be shown to be predictive of developing cancer (Schatzkin et al., 1990). For intermediate end points with etiologic fractions that are close to 1.0, either positive or negative results in cross-sectional studies of an exposure–intermediate end point relation are particularly informative. For intermediate end points linked to the risk of developing cancer but with a substantially lower etiologic fraction, the interpretation must be more circumspect. Specifically, a positive association between an exposure and an intermediate biomarker is informative, but a null association does not

rule out that the exposure is carcinogenic, as the exposure may act through a mechanism not reflected by the particular end point under study. Cross-sectional biomarker studies can collect detailed information on exposure status, which should be exploited to the fullest as extremely accurate information can be obtained on the dose–response relation between external or internal exposures and intermediate end points. As most of these markers reflect exposures over several days to months, this information must be collected over the etiologically relevant time period. For example, in a study on hematologic, cytogenetic, and molecular end points among workers exposed to benzene, measurements were collected for more than a year prior to determining the biologic end points to assess individual exposures unequivocally (Vermeulen et al., 2004; Lan et al., 2004). To increase the accuracy of the biomarker assessment, repeated measures of the biologic end point might be necessary if there is relatively large intraindividual temporal variation. Selection of an appropriate control group can be challenging in a cross-sectional study. Apart from coming from the same base population as the exposed group, efficiency is increased by some level of matching based on age, sex, ethnicity, socioeconomic status, and perhaps smoking, as these studies are often relatively small. For biomarkers with large interindividual variation among controls, an alternative study design in which subjects are used as their own referent and followed from the start of the exposure up to a certain length of exposure (or vice versa) might be more efficient for biomarker end points with short half-lives.

Case-Control and Case Series Studies The advantages and disadvantages of case-control and cohort study designs have been discussed in detail elsewhere (Rothman and Greenland, 1998; Caporaso et al., 1999; Langholz et al., 1999; Clayton and McKeigue, 2001; Wacholder et al., 2002a). Here we focus on aspects relevant to the collection and use of biologic markers (Table 6–5). Important advantages of case-control studies compared to prospective cohort studies are their ability to enroll large numbers of cancer cases quickly and the potential to study uncommon tumors that do not occur in large enough numbers in cohort studies. Because only one disease is evaluated, questionnaires can be used for a more detailed and focused exposure assessment than in cohort studies, which address multiple outcomes. However, because case-control studies collect exposure information and biologic specimens after diagnosis and sometimes after treatment of the disease, they are vulnerable to differential misclassification and to uncertainties in the temporal relation between the cancer and the biomarker under study. Differential misclassification or recall bias from questionnaire information in case-control studies has been a concern, although its existence has been proven for only a few exposures. The influence of the disease process and treatment on biomarkers of interest is often unknown. Although this is not a concern for stable markers such as genetic polymorphisms, it can be an important problem for other types of biomarkers such as serum nutrients and white blood cell DNA repair capacity. The hospital-based case-control design has been popular in molecular epidemiology because it facilitates subject enrollment and intense collection and processing of biologic specimens. Personal contact with the study participants by doctors, nurses, or interviewers is made easy, which usually results in higher participation rates for interviews and collection of biologic specimens. In addition, the easier access to health professionals and laboratories allows collection of different types and larger quantities of biologic specimens and more elaborate processing protocols, such as cryopreservation of lymphocytes. Because study subjects generally are geographically less spread out than those in population-based or cohort studies, rapid shipment of specimens to central laboratories for more extensive processing protocols is facilitated. In addition, because cases are usually diagnosed in fewer hospitals than population-based studies, the collection of paraffin-embedded tumor samples is facilitated. Furthermore, it is easier to enroll cases before surgery and establish collaborations with pathology departments at the hospital(s) for more intense, specialized

81

Application of Biomarkers in Cancer Epidemiology

Table 6–5. Comparison of Advantages and Limitations Relevant to the Collection of Biologic Specimens and Data Interpretation in Molecular Epidemiology Study Designs Study Design Hospital-based case-control

Population-based case-control

Prospective cohort

Advantages

Limitations

Facilitates intense collection and processing of specimens (e.g., freshly frozen tumor samples, cryopreserved lymphocytes) Participation rates for biologic collections might be enhanced Facilitates follow-up of cases for treatment response and survival Less subject to biases (e.g., selection, exposure misclassification) than hospital-based studies

Allows study of multiple disease end points Allows study of transient biomarkers and biomarkers affected by disease process Selection bias and differential misclassification are avoided; nondifferential misclassification may be reduced for some exposures Nested case-control or case-cohort studies can be used to improve efficiency of the design

tissue collections (e.g., freshly frozen tissue, alcohol-fixed tissue). Another advantage of the hospital-based design is the easier followup of the patients to determine their response to treatment and their survival. Follow-up of cases is important in molecular epidemiology studies, as many of the biomarkers under study (e.g., tumor characteristics, genetic polymorphisms) may be especially relevant for clinical outcomes in addition to cancer etiology. There are established criteria for selecting controls in hospital-based studies to estimate the effect of a single factor without bias, but the criteria for assessing interactions between more than one factor (e.g., gene–environment interactions) are less clear. Wacholder et al. (2000a) discussed the appropriate exclusion criteria in control selection for estimating main and subgroup effects as well as additive and multiplicative gene–environment interactions. One of the main challenges of population-based case-control studies has been obtaining high participation, especially when participation involves collecting biologic specimens. Collection of biologic materials using noninvasive approaches (e.g., collecting genomic DNA from buccal cells) (GarciaClosas et al., 2001) might help increase participation rates. Lack of participation biases the study results when the reason for nonparticipation is directly or indirectly related to the factors under study. Because this condition is often difficult to prove, especially when many factors are under study, high participation rates become the only means to ensure comparability of participants and nonparticipants. When participation rates are lower than desirable, it is important to seek some basic risk factor information, and biospecimens when possible, from nonparticipants to identify differences between participants and nonparticipants. In the case–case, case series, or case-only design, only subjects with the disease of interest and no controls are enrolled in the study. This design has been proposed to evaluate etiologic heterogeneity using tumor markers. The degree of etiologic heterogeneity is quantified by the ratio of the odds ratio for the effect of exposure on markerpositive tumors to the odds ratio for marker-negative tumors. This parameter is equivalent to the odds ratio for the association between exposure and tumor marker in the cases (Begg and Zhang, 1994). However, case-only studies are limited to an estimation of the ratio of the odds ratios and cannot be used to obtain estimates of the odds ratios for different tumor types. It should be noted that the odds ratio from a case-only design underestimates the odds ratio derived in a casecontrol design when the exposure of interest is associated with more than one tumor type. The case-only study has also been proposed as a valid design to evaluate multiplicative gene–gene (Yang et al., 1999) and gene–envi-

More prone to selection and differential misclassification biases than other designs Some biomarkers might be affected by disease process or hospital stay

Some biomarkers might be affected by disease process May be more difficult to obtain high participation rates for biologic collections than in hospital-based designs Implementation of intense, specialized blood and tumor collection and processing protocols can be challenging Implementation of intense, specialized collection and processing protocols in entire cohort can be challenging Obtaining tissue samples and following cases for treatment response and survival can be challenging in many cohort studies

ronment (Khoury and Flanders, 1996) interactions. However, this design has important limitations; most notably it cannot be used to obtain estimates of risk for disease or additive interactions, is susceptible to misinterpretation of the interaction parameter (Schmidt and Schaid, 1999), and is highly dependent on the assumption of independence between the exposure and the genotype under study (Albert et al., 2001). Because of these limitations, case-control designs are preferable to case-series designs.

Prospective Cohort Studies Although establishing a cohort study is initially extremely costly and time-consuming, in the long run it becomes more cost-efficient because it can study multiple disease end points and provides a well defined population that can be easily sampled for efficiency (Potter, 1997). Biologic specimens are collected before disease diagnosis and, ideally, before the beginning of the disease process. Therefore, it may be the only method that enables researchers to study biomarkers that are directly or indirectly affected by the disease process (Hunter, 1997). Although cohort studies have the theoretical advantage of collecting serial biologic samples over time, many large studies have only been able to collect a single biologic sample at one point in time. Although this is not a concern for DNA-based assays of inherited susceptibility markers, it poses some limitations for several other categories of markers, particularly short-term exposure markers that may vary substantially from day to day. Given that most members of a cohort do not develop cancer, nested case-control and less commonly case-cohort studies are used to improve efficiency (Wacholder, 1991). With these designs, only samples from cases and a random subset of noncases are analyzed, reducing the laboratory requirements and cost considerably. The nested case-control design includes all cases identified in the cohort up to a particular point in time and a random sample of subjects free of disease at the time of the case diagnosis. On the other hand, a case-cohort design includes a random sample of the cohort population at the onset of the study and all cases identified in the cohort up to a particular point in time. The efficiency of nested case-control studies can be easily increased by increasing the case/control ratio to two or three controls per case. The case-cohort design is simpler and allows evaluation of several disease end points; however, because in case-cohort studies the same disease-free subjects are repeatedly used as “controls” for different disease end points, it is difficult to perform assays on matched samples from diseased and disease-free subjects; thus, depletion of samples from the disease-free group may be an issue.

82

PART I: BASIC CONCEPTS

Table 6–4 lists prospective cohort studies being followed up for cancer incidence with basic risk factor information from questionnaires and stored blood components, including white blood cells, which can be used as a source of DNA. At completion of the ongoing collections, the current studies will have stored DNA samples from more than two million individuals. These studies will provide information on large numbers of patients with cancer at the more common sites (e.g., breast, lung, prostate, colon), thereby allowing evaluation of genetic markers of susceptibility and biomarkers in serum or plasma, such as hormone levels, chemical carcinogen levels, and proteomic patterns. Most cohort studies do not have cryopreserved blood samples because the procedure is expensive and logistically challenging in large studies. Also, cohort studies often have limited ability to collect tumor samples from large numbers of subjects and to follow up cases for the purpose of carrying out survival studies. New cohort studies based on large institutions such as health maintenance organizations (HMOs) could enable access to tumor samples and easier follow-up of cases to study treatment response and survival.

BIOLOGIC SAMPLE COLLECTION, PROCESSING, AND STORAGE The proper collection, processing, and storage of biologic specimens for epidemiologic studies is critical for successful determination of biomarkers (Holland et al., 2003). Because molecular techniques often require special processing of biologic samples, it is important to maximize opportunities to apply future assays to the samples being collected. In this section, we discuss important aspects of the collection, processing, and storage of the most commonly collected specimens for epidemiologic studies.

Blood Samples Blood samples are a valuable source of specimens that can be used for determining a wide range of biomarkers. Leukocytes (white blood cells: granulocytes, lymphocytes, monocytes), erythrocytes (red blood cells), platelets, and plasma/serum can be obtained through appropriate separation of blood components. White blood cells are an excellent source of large amounts of high quality genomic DNA for the study of genetic polymorphisms. The most efficient blood-processing protocol for obtaining DNA is to freeze samples as whole blood. White blood cells in the buffy coat obtained after separating the blood components are also a good source of DNA and are probably the most commonly used blood fraction for this purpose. Although the yields from buffy coats tend to be more variable than those from whole blood samples, separation of the buffy coat has the advantage of allowing separation of blood components (i.e., plasma and red blood cells) that can be used to determine biomarkers of a dose or early biologic effect. Small amounts of DNA can be obtained from dried blood spots on filter paper using fingerpricks, avoiding the use of venipuncture. An advantage of blood spots is the lower cost for collection, shipping, and storage (Steinberg et al., 1997). Plasma/serum can be used to measure circulating levels of nutrients, hormones, lipid/lipoproteins, xenobiotic exposure materials, and other compounds. Serum is obtained from whole blood samples without anticoagulants after separation from the blood clot, whereas plasma is obtained from anticoagulated blood samples after separation from the white and red blood cells. Serum is preferred over plasma for measuring some compounds, such as antibodies, nutrients, lipids, and lipoproteins. The choice of anticoagulants can affect the usability of plasma or other blood components. For instance, EDTA is well suited for DNA-based assays, but it influences the Mg2+ concentration and presents problems for cytogenetic analyses (Landi and Caporaso, 1997). Another factor to take into consideration is that measurement of certain labile compounds requires preservation with stabilizing agents. For instance, EDTA and ascorbic acid are stabilizing agents for folate and must be added to the blood as soon as possible after collection (Holland et al., 2003).

Blood samples can also be a source of the viable lymphocytes needed to perform phenotypic assays. Although cryopreservation of lymphocytes results in a loss of viable cells, the use of viable lymphocytes isolated from fresh blood is often not feasible for epidemiologic studies. Studies have demonstrated the feasibility of obtaining viable cells after cryopreservation of separated lymphocytes (Kleeberger et al., 1999; Beck et al., 2001) and whole blood samples (Hayes et al., 2002). Cryopreservation of whole blood is simpler, faster, and more economical and requires less blood volume than isolation and cryopreservation of lymphocytes (Cheng et al., 2001). However, sample thawing and lymphocyte isolation from whole blood samples can be challenging. To assess the value of cryopreservation of blood in epidemiologic studies, it is important to determine the effects of isolation, freezing, storage time, and thawing on phenotypic assays of interest. For instance, cryopreserved lymphocytes have endogenous levels of DNA strand breaks and responses to oxidative challenge similar to those of freshly isolated lymphocytes, but they have reduced DNA repair capacity (Duthie et al., 2002). DNA repair capacity values have been found to be similar for frozen isolated lymphocytes and frozen whole blood (Cheng et al., 2001). When samples from cases and controls are treated similarly, the effects of cryopreservation are likely to be nondifferential, and thus tend to obscure associations rather than cause spurious associations. As for plasma/serum, the potential effects of disease processes on phenotypic assays make the collection of cryopreserved lymphocytes most valuable in prospective studies. However, given its cost and processing requirements, especially for isolated lymphocytes, most prospective studies are unable to cryopreserve cells. Despite the advantages of blood samples, the use of venipuncture in large-scale epidemiologic studies has two important limitations: high cost and, in some populations, relatively low acceptability. Thus, epidemiologic studies often require less expensive methods of collection with lower levels of discomfort to the study participant to increase participation rates. Indeed, methods suitable for self-collection are particularly advantageous in some instances.

Buccal Cells Exfoliated buccal epithelial cells are an attractive alternative to blood samples for the collection of genomic DNA because they can be selfcollected using noninvasive, relatively inexpensive techniques and sent to the researcher by mail. Moreover, they can be used as an alternative source of DNA for subjects unwilling to provide a blood sample. Buccal cytobrushes, cotton swabs, and mouthwashing are the most commonly used protocols for buccal cell collection. Table 6–6 shows the characteristics of blood and buccal cell collections as sources of DNA. The mouthwash protocol provides substantially larger amounts and better quality human genomic DNA than do cytobrush samples, although both methods seem to be adequate for a wide range of PCR-based assays (Garcia-Closas et al., 2001). Another potential advantage of mouthwash versus cytobrush samples is that it might be possible to use the small amount of saliva contained in the mouthwash sample to measure levels of chemicals that can serve as biomarkers of the internal dose. Despite the advantages of the mouth-

Table 6–6. Comparison of Phlebotomy Blood and Buccal Cell Samples as a Source of Genomic DNA in Epidemiologic Studies Buccal Cell Samples Characteristic

Phlebotomy Blood Sample

Self-collection, by mail Subject acceptability Cost for collection Human DNA yield

No Fair High High

DNA quality Other useful material

High Plasma, red blood cells

Mouthwash

Cytobrush

Yes Good Low Medium/ Low High Saliva

Yes Good Low Low/very low Medium None

Application of Biomarkers in Cancer Epidemiology wash samples, cytobrush samples might still be the method of choice for some studies, such as studies in young children. The DNA yield from buccal cells is substantially lower than the yield from blood, although it varies widely among individuals (GarciaClosas et al., 2001). As a consequence, only a limited number of genotype assays can be performed in subjects with low DNA yields, reducing the sample size and raising the possibility of bias if low DNA yields are related to the factors under study. Methods that amplify extremely small amounts of DNA and increase the number of genetic assays in subjects with low yields, such as whole genome amplification (Zheng et al., 2001), may ameliorate this potential limitation. In addition to blood and buccal cell collection, there are other sources of DNA for determining genetic polymorphisms in epidemiologic studies, including urine, plasma/serum, and paraffin-embedded tissue (Blomeke and Shields, 1999). However, these materials would generally not be the method of choice for de novo collections because they do not have any substantial advantage in terms of DNA yield, quality, subject acceptability, or cost.

Urine A wide variety of biologic markers of exposure and metabolic markers can be measured in urine samples (Gunter and McQuillan, 1990). Moreover, intermediate end points can be measured in exfoliated urothelial cells, which are relevant for studying bladder cancer. For most exposure markers, the gold standard is the 24-hour urine sample collection, followed, in general, by the 12-hour evening/overnight sample, the 8-hour overnight collection, the first morning voided sample, and the “spot” single urine sample. The utility of a single spot urine sample, relative to longer, timed collections, is highly specific to the kinetics associated with the pattern of exposure and the half-life of the biomarker. Of note, the National Health and Nutrition Examination Survey (NHANES 1999–2004) study is evaluating a wide range of exposure biomarkers on spot urine samples collected from several thousand people in the United States. The collection and processing of urine samples is often uncomplicated, with the sample being kept cold to maintain the stability of the analytes and to avoid bacterial overgrowth. Generally, urine is simply aliquoted and frozen. For some analytes, collection and storage containers have specific requirements, and preservatives may be needed. Urine can be filtered or centrifuged before freezing if assays are to be carried out on urothelial cells. Adding glycerol to urine prior to freezing keeps cell walls intact and allows filtration for the purpose of collecting cells in the future (Rothman et al., 1996).

Tumor Tissue The most common type of tissue specimen available for epidemiologic research is archived, paraffin-embedded tissue samples that have been used for clinical diagnosis. The successful use of these specimens requires acquiring information about when, why, and how specific lesions were removed and the methods that were used for processing and storing the specimens. These factors determine which assays can be performed and how the results should be analyzed and interpreted. For instance, neoadjuvant irradiation or chemotherapy performed prior to surgical removal may lead to extensive necrosis, leaving behind a biased sample of treatment-resistant tumor. The use of posttreatment specimens for etiologic research is suspect unless validated to address a particular question. There is considerable evidence that immunohistochemical assays performed on freshly cut sections prepared from archived paraffin blocks are satisfactory, even after years of storage (Camp et al., 2000). However, once sections are cut and mounted on glass slides, immunoreactivity may decline within weeks (Jacobs et al., 1996; Wester et al., 2000; Fergenbaum et al., 2004). Oxidation has been suggested as a possible cause of this loss of immunoreactivity, which has led to exploration of slide storage under gaseous nitrogen and other methods of preserving immunoreactivity in cut sections. Tissue microarray (TMA) technology represents a method for achieving high throughput immunohistochemical or in situ hybridiza-

83

tion analysis of many specimens (Kononen et al., 1998). With this method, small tissue cores of pathologic targets are removed from a large set of paraffin blocks (“donor” blocks) and transferred systematically into one or a few “recipient” blocks. Sections of single TMA blocks can provide representations of hundreds of cases suitable for testing in a single batch, thereby reducing cost, expense, and interbatch variability. Preparation of TMAs requires considerable effort and pathologic expertise, and interpretation is labor-intensive. Recent technologic advances include the development of automated instruments with increased capabilities, methods for imaging stained slides that permit pathology review on a computer monitor, automated image analysis, preparation of arrays from cytologic specimens and frozen tissues, and quantitative methods (Camp et al., 2002). Advances in immunohistochemistry and the development of new reagent antibodies and in situ probes should increase the power of this technology for use in large epidemiologic studies. Paraffin-embedded tissue blocks are generally fixed in 10% buffered formalin, which provides optimal morphology but is not necessarily best for studies of nucleic acids and other molecules (Srinivasan et al., 2002). Collecting tissue specimens by protocols not routinely used for diagnostic purposes are required to obtain high quality DNA and mRNA. Because of the need for close collaboration with surgeons and pathologists at the treating hospital, population-based studies involving only a few hospitals or hospital-based studies are best suited for accommodating to special tissue-processing protocols. Frozen specimens provide better sources of high quality DNA than fixed tissues, but fresh specimens cannot be microdissected and dilution of tumor DNA with adjacent normal tissue DNA can be problematic for certain assays. Preservation of RNA is more challenging than maintaining the integrity of DNA because RNAses are ubiquitous in the environment. Collecting specimens for RNA-based assays requires immediate handling of specimens after removal using clean gloves and instruments followed by deep freezing or immersion in special fixatives. Even with optimal handling and RNA extraction, RNA degradation in situ resulting from ischemia, surgical devascularization, or other factors may limit the value of the results. New methodologies allow stabilization of RNA in blood and tissues that can be stored refrigerated or at room temperature for extended periods, thereby facilitating RNA collection in epidemiologic field studies.

CONCLUSIONS The confluence of two major trends in cancer research should enable unprecedented exploration of the etiology of cancer over the next decade. First, many large prospective cohort and case-control studies with biologic samples have now been established that provide opportunities to evaluate environmental risks, intermediate end points, genetic factors, and tumor marker characteristics in an integrated manner for essentially all common and many less common tumors. Second, the mapping of the human genome and the rapidly evolving technologic advances in high throughput analytic platforms for genetic polymorphisms, gene transcripts, proteins, and low-molecular-weight compounds is allowing broad exploration of etiologic factors in biologic samples from subjects in these studies. Clearly, the statistical analysis and interpretation of study findings and their effective and ethically appropriate translation to the individual and general population will be challenging (Schulte et al., 1997; Guttmacher and Collins, 2003). We are optimistic, however, that methodologic answers to the former and appropriate responses to the latter will be successfully developed. Furthermore, adherence to the time-tested methods of classic epidemiology in the context of a strong public health perspective should continue to provide guidance as cancer epidemiology evolves during the postgenomic era. References Aardema MJ, MacGregor JT. 2002. Toxicology and genetic toxicology in the new era of “toxicogenomics”: impact of “-omics” technologies. Mutat Res 499:13–25. Adlercreutz H. 2002. Phytoestrogens and breast cancer. J Steroid Biochem Mol Biol 83:113–118.

84

PART I: BASIC CONCEPTS

Ahsan H, Rundle AG. 2003. Measures of genotype versus gene products: promise and pitfalls in cancer prevention. Carcinogenesis 24:1429–1434. Albert PS, Ratnasinghe D, Tangrea J, Wacholder S. 2001. Limitations of the case-only design for identifying gene-environment interactions. Am J Epidemiol 154:687–693. Alizadeh AA, Eisen MB, Davis RE, Ma C, Lossos IS, Rosenwald A, Boldrick JC, Sabet H, Tran T, Yu X, Powell JI, Yang L, Marti GE, Moore T, Hudson J Jr, Lu L, Lewis DB, Tibshirani R, Sherlock G, Chan WC, Greiner TC, Weisenburger DD, Armitage JO, Warnke R, Levy R, Wilson W, Grever MR, Byrd JC, Botstein D, Brown PO, Staudt LM. 2000. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 403:503–511. Althuis MD, Fergenbaum JH, Sherman ME, Garcia-Closas M, Madigan MP, Brinton LA. 2003. Breast cancer etiology by estrogen and progesterone receptors status: a review. Ann Epidemiol 13:559. Altshuler D, Brooks LD, Chakravarti A, Collins FS, Daly MJ, Donnelly P; International HapMap Consortium. 2005. A haplotype map of the human genome. Nature 437:1299–1320. Anderson GL, Manson J, Wallace R, Lund B, Hall D, Davis S, Shumaker S, Wang CY, Stein E, Prentice RL. 2003. Implementation of the Women’s Health Initiative study design. Ann Epidemiol 13:S5–S17. Anderson WF, Chu KC, Chang S, Sherman ME. 2004. Comparison of incidence rate patterns for different histopathologic types of breast carcinoma. J Clin Oncol 13:1128–1135. Anonymous. 1999. Freely associating. Nat Genet 22:1–2. ARIC Investigators. 1989. The Atherosclerosis Risk in Communities (ARIC) study: design and objectives. Am J Epidemiol 129:687–702. Armstrong B, White E, Saracci R. 1992a. Exposure measurement error and its effects. In: Armstrong BK, White E, Saracci R, editors. Principles of Exposure Measurements in Epidemiology. New York: Oxford University Press. Armstrong BK, White E, Saracci R. 1992b. Principles of Exposure Measurements in Epidemiology. Oxford: Oxford University Press. Baak JP, Path FR, Hermsen MA, Meijer G, Schmidt J, Janssen EA. 2003. Genomics and proteomics in cancer. Eur J Cancer 39:1199–1215. Barr DB, Needham LL. 2002. Analytical methods for biological monitoring of exposure to pesticides: a review. J Chromatogr B Analyt Technol Biomed Life Sci 778:5–29. Beck JC, Beiswanger CM, John EM, Satariano E, West D. 2001. Successful transformation of cryopreserved lymphocytes: a resource for epidemiological studies. Cancer Epidemiol Biomarkers Prev 10:551–554. Begg CB, Zhang ZF. 1994. Statistical-analysis of molecular epidemiology studies employing case-series. Cancer Epidemiol Biomarkers Prev 3:173–175. Benjamini Y, Drai D, Elmer G, Kafkafi N, Golani I. 2001. Controlling the false discovery rate in behavior genetics research. Behav Brain Res 125: 279–284. Berrino F, Muti P, Micheli A, Bolelli G, Krogh V, Sciajno R, Pisani P, Panico S, Secreto G. 1996. Serum sex hormone levels after menopause and subsequent breast cancer. J Natl Cancer Inst 88:291–296. Berwick M, Vineis P. 2000. Markers of DNA repair and susceptibility to cancer in humans: an epidemiologic review. J Natl Cancer Inst 92:874–897. Blomeke B, Shields PG. 1999. Laboratory methods for the determination of genetic polymorphisms in humans. IARC Sci Publ 148:133–147. Boffetta P, Linet M, Armstrong B. The InterLymph collaboration: a consortium of molecular epidemiological studies of non-Hodgkin’s lymphoma [abstract]. In: Proceedings of the American Association for Cancer Research. p. 44. Bonassi S, Au WW. 2002. Biomarkers in molecular epidemiology studies for health risk prediction. Mutat Res 511:73–86. Bonassi S, Abbondandolo A, Camurri L, Dal Pra L, De Ferrari M, Degrassi F, Forni A, Lamberti L, Lando C, Padovani P. 1995. Are chromosome aberrations in circulating lymphocytes predictive of future cancer onset in humans? Preliminary results of an Italian cohort study. Cancer Genet Cytogenet 79:133–135. Bonassi S, Hagmar L, Stromberg U, Montagud AH, Tinnerberg H, Forni A, Heikkila P, Wanders S, Wilhardt P, Hansteen IL, Knudsen LE, Norppa H. 2000. Chromosomal aberrations in lymphocytes predict human cancer independently of exposure to carcinogens; European Study Group on Cytogenetic Biomarkers and Health. Cancer Res 60:1619–1625. Bosch FX, Lorincz A, Munoz N, Meijer CJ, Shah KV. 2002. The causal relation between human papillomavirus and cervical cancer. J Clin Pathol 55:244–265. Bowen DJ, Thornquist M, Anderson K, Barnett M, Powell C, Goodman G, Omenn G. 2003. Stopping the active intervention: CARET. Control Clin Trials 24:39–50. Browner WS, Newman TB. 1987. Are all significant P values created equal? The analogy between diagnostic tests and clinical research. JAMA 257:2459–2463.

Calle EE, Rodriguez C, Jacobs EJ, Almon ML, Chao A, McCullough ML, Feigelson HS, Thun MJ. 2002. The American Cancer Society Cancer Prevention Study II Nutrition Cohort: rationale, study design, and baseline characteristics. Cancer 94:2490–2501. Camp RL, Charette LA, Rimm DL. 2000. Validation of tissue microarray technology in breast carcinoma. Lab Invest 80:1943–1949. Camp RL, Chung GG, Rimm DL. 2002. Automated subcellular localization and quantification of protein expression in tissue microarrays. Nat Med 8:1323–1327. Caporaso N, Goldstein A. 1995. Cancer genes: single and susceptibility: exposing the difference. Pharmacogenetics 5:59–63. Caporaso N, Rothman N, Wacholder S. 1999. Case-control of common alleles and environmental factors. J Natl Cancer Inst Monogr 26:25–30. Cardon LR, Abecasis GR. 2003. Using haplotype blocks to map human complex trait loci. Trends Genet 19:135–140. Carlson CS, Eberle MA, Rieder MJ, Yi Q, Kruglyak L, Nickerson DA. 2003. Selecting a maximally informative set of single-nucleotide polymorphisms for association analyses using linkage disequilibrium. 2004. Am J Hum Genet 74:106–120. Chatterjee N. 2004. A two-stage regression model for epidemiologic studies with multivariate disease classification data. J Am Stat Assoc 99:127–138. Chen Z, Lee L, Chen J, Collins R, Wu F, Guo Y, Linksted P, Peto R. 2005. Cohort profile: the Kadoorie Study of Chronic Disease in China (KSCDC). Int J Epidemiol 34:1243–1249. Cheng L, Wang LE, Spitz MR, Wei Q. 2001. Cryopreserving whole blood for functional assays using viable lymphocytes in molecular epidemiology studies. Cancer Lett 166:155–163. Clayton D, McKeigue PM. 2001. Epidemiological methods for studying genes and environmental factors in complex diseases. Lancet 358: 1356–1360. Coggon D, Friesen MD. 1997. Markers of internal dose: chemical agents. In: Toniolo P, Boffetta P, Shuker DEG, Rothman N, Hulka B, Pearce N, editors. Application of Biomarkers in Cancer Epidemiology. Lyon: IARC Scientific Publications, pp. 95–102. Colditz GA. 1995. The nurses’ health study: a cohort of US women followed since 1976. J Am Med Womens Assoc 50:40–44. Colhoun HM, McKeigue PM, Davey SG. 2003. Problems of reporting genetic associations with complex outcomes. Lancet 361:865–872. Collins A, Lonjou C, Morton NE. 1999. Genetic epidemiology of singlenucleotide polymorphisms. Proc Natl Acad Sci USA 96:15173–15177. Couzin J. 2002. Genomics: new mapping project splits the community. Science 296:1391–1393. Cox DR. 2001. Another comment on the role of statistical methods. BMJ 81:1469. Cutler P. 2003. Protein arrays: the current state-of-the-art. Proteomics 3:3–18. Cuzick J. 2002. Role of HPV testing in clinical practice. Virus Res 89:263–269. Deitz A, Garcia-Closas M, Rothman N, Hayes RB, Chow W-H, Hein D, Rebbeck TR. 2000. Impact of misclassification in genotype-disease association studies: example of N-acetyl 2 (NAT2) smoking and bladder cancer. Proc Am Assoc Cancer Res 41:559. De Roos AJ, Rothman N, Inskip PD, Linet MS, Shapiro WR, Selker RG, Fine HA, Black PM, Pittman GS, Bell DA. 2003. Genetic polymorphisms in GSTM1, -P1, -T1, and CYP2E1 and the risk of adult brain tumors. Cancer Epidemiol Biomarkers Prev 12:14–22. Duthie SJ, Pirie L, Jenkinson AM, Narayanan S. 2002. Cryopreserved versus freshly isolated lymphocytes in human biomonitoring: endogenous and induced DNA damage, antioxidant status and repair capability. Mutagenesis 17:211–214. Engel LS, Taioli E, Pfeiffer R, Garcia-Closas M, Marcus PM, Lan Q, Boffetta P, Vineis P, Autrup H, Bell DA, Branch RA, Brockmoller J, Daly AK, Heckbert SR, Kalina I, Kang D, Katoh T, Lafuente A, Lin HJ, Romkes M, Taylor JA, Rothman N. 2002. Pooled analysis and meta-analysis of glutathione S-transferase M1 and bladder cancer: a HuGE review. Am J Epidemiol 156:95–109. Esteller M. 2003. Cancer epigenetics: DNA methylation and chromatin alterations in human cancer. Adv Exp Med Biol 532:39–49. Fergenbaum JH, Garcia-Closas M, Hewitt SM, Lissowska J, Sakoda LC, Sherman ME. 2004. Loss of antigenicity in stored sections of breast cancer tissue microarrays. Cancer Epidemiol Biomarkers Prev 13:667–672. Forrest MS, Lan Q, Hubbard AE, Zhang L, Vermeulen R, Zhao X, Li G, Wu YY, Shen M, Yin S, Chanock SJ, Rothman N, Smith MT. 2005. Discovery of novel biomarkers by microarray analysis of peripheral blood mononuclear cell gene expression in benzene-exposed workers. Environ Health Perspect 113:801–807. Freedman ML, Reich D, Penney KL, McDonald GJ, Mignault AA, Patterson N, Gabriel SB, Topol EJ, Smoller JW, Pato CN, Pato MT, Petryshen TL, Kolonel LN, Lander ES, Sklar P, Henderson B, Hirschhorn JN, Altshuler

Application of Biomarkers in Cancer Epidemiology D. 2004. Assessing the impact of population stratification on genetic association studies. Nat Genet 36:383–393. Furberg AH, Ambrosone CB. 2001. Molecular epidemiology, biomarkers and cancer prevention. Trends Mol Med 7:517–521. Gabriel SB, Schaffner SF, Nguyen H, Moore JM, Roy J, Blumenstiel B, Higgins J, DeFelice M, Lochner A, Faggart M, Liu-Cordero SN, Rotimi C, Adeyemo A, Cooper R, Ward R, Lander ES, Daly MJ, Altshuler D. 2002. The structure of haplotype blocks in the human genome. Science 296:2225–2229. Gammon MD, Wolff MS, Neugut AI, Eng SM, Teitelbaum SL, Britton JA, Terry MB, Levin B, Stellman SD, Kabat GC, Hatch M, Senie R, Berkowitz G, Bradlow HL, Garbowski G, Maffeo C, Montalvan P, Kemeny M, Citron M, Schnabel F, Schuss A, Hajdu S, Vinceguerra V, Niguidula N, Ireland K, Santella RM. 2002. Environmental toxins and breast cancer on Long Island. II. Organochlorine compound levels in blood. Cancer Epidemiol Biomarkers Prev 11:686–697. Garcia-Closas M, Lubin JH. 1999. Power and sample size calculations in casecontrol studies of gene-environment interactions: comments on different approaches. Am J Epidemiol 149:689–692. Garcia-Closas M, Egan KM, Abruzzo J, Newcomb PA, Titus-Ernstoff L, Franklin T, Bender PK, Beck JC, Le Marchand L, Lum A, Alavanja M, Hayes RB, Rutter J, Buetow K, Brinton LA, Rothman N. 2001. Collection of genomic DNA from adults in epidemiological studies by buccal cytobrush and mouthwash. Cancer Epidemiol Biomarkers Prev 10:687–696. Garcia-Closas M, Malats N, Silverman D, Dosemeci M, Kogevinas M, Hein DW, Tardon A, Serra C, Carrato A, Garcia-Closas R, Lloreta J, CastanoVinyals G, Yeager M, Welch R, Chanock S, Chatterjee N, Wacholder S, Samanic C, Tora M, Fernandez F, Real FX, Rothman N. 2005. NAT2 slow acetylation, GSTM1 null genotype, and risk of bladder cancer: results from the Spanish Bladder Cancer Study and meta-analyses. Lancet 366:649–659. Garcia-Closas M, Rothman N, Lubin J. 1999. Misclassification in case-control studies of gene-environment interactions: assessment of bias and sample size. Cancer Epidemiol Biomarkers Prev 8:1043–1050. Garcia-Closas M, Wacholder S, Caporaso N, Rothman N. 2003. Inference issues in cohort and case-control studies of genetic effects and geneenvironment interactions. In: Khoury M, Little J, Burke W, editors. Human Genome Epidemiology: Scientific Foundation for Using Genetic Information to Improve Health and Prevent Disease. Oxford: Oxford University Press, pp. 127–144. Giles GG, English DR. 2003. The Melbourne Collaborative Study. In: Riboli E, Lambert R, editors. Nutrition and Lifestyle: Opportunities for Cancer Prevention. Lyon: International Agency for Research on Cancer, pp. 69–70. Giovannucci E, Rimm EB, Ascherio A, Colditz GA, Spiegelman D, Stampfer MJ, Willett WC. 1999. Smoking and risk of total and fatal prostate cancer in United States health professionals. Cancer Epidemiol Biomarkers Prev 8:277–282. Godschalk RW, Feldker DE, Borm PJ, Wouters EF, van Schooten FJ. 2002. Body mass index modulates aromatic DNA adduct levels and their persistence in smokers. Cancer Epidemiol Biomarkers Prev 11:790–793. Gompertz D. 1997. Quality control of biomarker measurement in epidemiology. IARC Sci Publ 142:215–222. Goodman SN. 1999. Toward evidence-based medical statistics. 2. The Bayes factor. Ann Intern Med 130:1005–1013. Graves PR, Haystead TA. 2002. Molecular biologist’s guide to proteomics. Microbiol Mol Biol Rev 66:39–63. Greenblatt MS, Bennett WP, Hollstein M, Harris CC. 1994. Mutations in the p53 tumor suppressor gene: clues to cancer etiology and molecular pathogenesis. Cancer Res 54:4855–4878. Greenland S. 1994. Hierarchical regression for epidemiologic analyses of multiple exposures. Environ Health Perspect 102(Suppl 8):33–39. Greenland S, Rothman KJ. 1998. Concepts of interaction. In: Modern Epidemiology, 2nd ed. Philadelphia: Lippincott-Raven, pp. 329–342. Gunter EW, McQuillan G. 1990. Quality control in planning and operating the laboratory component for the Third National Health and Nutrition Examination Survey. J Nutr 120(Suppl 11):1451–1454. Guttmacher AE, Collins FS. 2003. Welcome to the genomic era. N Engl J Med 349:996–998. Hagmar L, Brogger A, Hansteen IL, Heim S, Hogstedt B, Knudsen L, Lambert B, Linnainmaa K, Mitelman F, Nordenson I. 1994. Cancer risk in humans predicted by increased levels of chromosomal aberrations in lymphocytes: Nordic study group on the health risk of chromosome damage. Cancer Res 54:2919–2922. Hainaut P, Pfeifer GP. 2001. Patterns of p53 GÆT transversions in lung cancers reflect the primary mutagenic signature of DNA damage by tobacco smoke. Carcinogenesis 22:367–374.

85

Hanash S. 2003. Disease proteomics. Nature 422:226–232. Hardenbol P, Yu F, Belmont J, Mackenzie J, Bruckner C, Brundage T, Boudreau A, Chow S, Eberle J, Erbilgin A, Falkowski M, Fitzgerald R, Ghose S, Iartchouk O, Jain M, Karlin-Neumann G, Lu X, Miao X, Moore B, Moorhead M, Namsaraev E, Pasternak S, Prakash E, Tran K, Wang Z, Jones HB, Davis RW, Willis TD, Gibbs RA. 2005. Highly multiplexed molecular inversion probe genotyping: over 10,000 targeted SNPs genotyped in a single tube assay. Genome Res 15:269–275. Hayes RB, Reding D, Kopp W, Subar AF, Bhat N, Rothman N, Caporaso N, Ziegler RG, Johnson CC, Weissfeld JL, Hoover RN, Hartge P, Palace C, Gohagan JK. 2000. Etiologic and early marker studies in the prostate, lung, colorectal and ovarian (PLCO) cancer screening trial. Control Clin Trials 21:349S–355S. Hayes RB, Smith CO, Huang WY, Read Y, Kopp WC. 2002. Whole blood cryopreservation in epidemiological studies. Cancer Epidemiol Biomarkers Prev 11:1496–1498. Hirschhorn JN, Daly MJ. 2005. Genome-wide association studies for common diseases and complex traits. Nat Rev Genet 6:95–108. Hirschhorn JN, Lohmueller K, Byrne E, Hirschhorn K. 2002. A comprehensive review of genetic association studies. Genet Med 4:45–61. Holland NT, Smith MT, Eskenazi B, Bastaki M. 2003. Biological sample collection and processing for molecular epidemiological studies. Mutat Res 543:217–234. Hollstein M, Hergenhahn M, Yang Q, Bartsch H, Wang ZQ, Hainaut P. 1999. New approaches to understanding p53 gene tumor mutation spectra. Mutat Res 431:199–209. Huang X, Powell J, Mooney LA, Li C, Frenkel K. 2001. Importance of complete DNA digestion in minimizing variability of 8-oxo-dG analyses. Free Radic Biol Med 31:1341–1351. Hulka BS. 1991. ASPO Distinguished Achievement Award Lecture: epidemiological studies using biological markers: issues for epidemiologists. Cancer Epidemiol Biomarkers Prev 1:13–19. Hulka BS, Margolin BH. 1992. Methodological issues in epidemiologic studies using biologic markers. Am J Epidemiol 135:200–209. Hunter DJ. 1997. Methodological issues in the use of biological markers in cancer epidemiology: cohort studies. In: Toniolo P, Boffeta P, Shuker DEG, Rothman N, Hulka B, Pearce N, editors. Application of Biomarkers in Cancer Epidemiology. Lyon: IARC Scientific Publications, pp. 39–46. Ioannidis JP, Bernstein J, Boffetta P, Danesh J, Dolan S, Hartge P, Hunter D, Inskip P, Jarvelin MR, Little J, Maraganore DM, Bishop JA, O’Brien TR, Petersen G, Riboli E, Seminara D, Taioli E, Uitterlinden AG, Vineis P, Winn DM, Salanti G, Higgins JP, Khoury MJ. 2005. A network of investigator networks in human genome epidemiology. Am J Epidemiol 162:302–304. Epub 2005 Jul 13. Ioannidis JP, Ntzani EE, Trikalinos TA, Contopoulos-Ioannidis DG. 2001. Replication validity of genetic association studies. Nat Genet 29:306–309. Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U, Speed TP. 2003. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 4:249–264. Jacobs TW, Prioleau JE, Stillman IE, Schnitt SJ. 1996. Loss of tumor markerimmunostaining intensity on stored paraffin slides of breast cancer. J Natl Cancer Inst 88:1054–1059. Jones PA. 2003. Epigenetics in carcinogenesis and cancer prevention. Ann NY Acad Sci 983:213–219. Kaaks R, Riboli E, Sinha R. 1997a. Biochemical markers of dietary intake. IARC Sci Publ 142:103–126. Kaaks R, Riboli E, Sinha R. 1997b. Biochemical markers of dietary intake. In: Toniolo P, Boffetta P, Shuker D, Rothman N, Hulka B, Pearce N, editors Application of Biomarkers in Cancer Epidemiology. Lyon: IARC, pp. 103–126. Kallioniemi A, Kallioniemi OP, Sudar D, Rutovitz D, Gray JW, Waldman F, Pinkel D. 1992. Comparative genomic hybridization for molecular cytogenetic analysis of solid tumors. Science 258:818–821. Kallioniemi OP, Kallioniemi A, Sudar D, Rutovitz D, Gray JW, Waldman F, Pinkel D. 1993. Comparative genomic hybridization: a rapid new method for detecting and mapping DNA amplification in tumors. Semin Cancer Biol 4:41–46. Ketchum NS, Michalek JE, Burton JE. 1999. Serum dioxin and cancer in veterans of Operation Ranch Hand. Am J Epidemiol 149:630–639. Khoury MJ, Flanders WD. 1996. Nontraditional epidemiologic approaches in the analysis of gene-environment interaction: case-control studies with no controls. Am J Epidemiol 144:207–213. Kleeberger CA, Lyles RH, Margolick JB, Rinaldo CR, Phair JP, Giorgi JV. 1999. Viability and recovery of peripheral blood mononuclear cells cryopreserved for up to 12 years in a multicenter study. Clin Diagn Lab Immunol 6:14–19.

86

PART I: BASIC CONCEPTS

Kolonel LN, Henderson BE, Hankin JH, Nomura AM, Wilkens LR, Pike MC, Stram DO, Monroe KR, Earle ME, Nagamine FS. 2000. A multiethnic cohort in Hawaii and Los Angeles: baseline characteristics. Am J Epidemiol 151:346–357. Kononen J, Bubendorf L, Kallioniemi A, Barlund M, Schraml P, Leighton S, Torhorst J, Mihatsch MJ, Sauter G, Kallioniemi OP. 1998. Tissue microarrays for high-throughput molecular profiling of tumor specimens. Nat Med 4:844–847. Krajcik RA, Borofsky ND, Massardo S, Orentreich N. 2002. Insulin-like growth factor I (IGF-I), IGF-binding proteins, and breast cancer. Cancer Epidemiol Biomarkers Prev 11:1566–1573. Kromhout H, Symanski E, Rappaport SM. 1993. A comprehensive evaluation of within- and between-worker components of occupational exposure to chemical agents. Ann Occup Hyg 37:253–270. Kruglyak L. 1999. Prospects for whole-genome linkage disequilibrium mapping of common disease genes. Nat Genet 22:139–144. Lacey JV Jr, Frisch M, Brinton LA, Abbas FM, Barnes WA, Gravitt PE, Greenberg MD, Greene SM, Hadjimichael OC, McGowan L, Mortel R, Schwartz PE, Zaino RJ, Hildesheim A. 2001. Associations between smoking and adenocarcinomas and squamous cell carcinomas of the uterine cervix (United States). Cancer Causes Control 12:153–161. Lai E, Bowman C, Bansal A, Hughes A, Mosteller M, Roses AD. 2002. Medical applications of haplotype-based SNP maps: learning to walk before we run. Nat Genet 32:353. Laird PW. 2003. The power and the promise of DNA methylation markers. Nat Rev Cancer 3:253–266. Lakhani SR, Ashworth A. 2001. Microarray and histopathological analysis of tumours: the future and the past? Nat Rev Cancer 1:151–157. Lamar CA, Dorgan JF, Longcope C, Stanczyk FZ, Falk RT, Stephenson HE Jr. 2003. Serum sex hormones and breast cancer risk factors in postmenopausal women. Cancer Epidemiol Biomarkers Prev 12:380– 383. Lan Q, Zhang L, Li G, Vermeulen R, Weinberg RS, Dosemeci M, Rappaport SM, Shen M, Alter BP, Wu Y, Kopp W, Waidyanatha S, Rabkin C, Guo W, Chanock S, Hayes RB, Linet M, Kim S, Yin S, Rothman N, Smith MT. 2004. Hematotoxicity in workers exposed to low levels of benzene. Science 306:1774–1776. Lander ES. 1996. The new genomics: global views of biology. Science 274:536–539. Landi MT, Caporaso NE. 1997. Sample collection, processing and storage. In: Toniolo P, Boffeta P, Shuker DEG, Rothman N, Hulka B, Pearce N, editors. Application of Biomarkers in Cancer Epidemiology. Lyon: IARC Scientific Publications, pp. 223–236. Langholz B, Rothman N, Wacholder S, Thomas DC. 1999. Cohort studies for characterizing measured genes. J Natl Cancer Inst Monogr 26:39– 42. Liou SH, Lung JC, Chen YH, Yang T, Hsieh LL, Chen CJ, Wu TN. 1999. Increased chromosome-type chromosome aberration frequencies as biomarkers of cancer risk in a Blackfoot endemic area. Cancer Res 59:1481–1484. Livingston RJ, von Niederhausern A, Jegga AG, Crawford DC, Carlson CS, Rieder MJ, Gowrisankar S, Aronow BJ, Weiss RB, Nickerson DA. 2004. Pattern of sequence variation across 213 environmental response genes. Genome Res 14:1821–1831. Epub 2004 Sep 13. Liyanage M, Coleman A, du MS, Veldman T, McCormack S, Dickson RB, Barlow C, Wynshaw-Boris A, Janz S, Wienberg J, Ferguson-Smith MA, Schrock E, Ried T. 1996. Multicolour spectral karyotyping of mouse chromosomes. Nat Genet 14:312–315. Lobenhofer EK, Bushel PR, Afshari CA, Hamadeh HK. 2001. Progress in the application of DNA microarrays. Environ Health Perspect 109:881–891. Lubin JH, Blot WJ. 1984. Assessment of lung cancer risk factors by histologic category. J Natl Cancer Inst 73:383–389. Malila N, Taylor PR, Virtanen MJ, Korhonen P, Huttunen JK, Albanes D, Virtamo J. 2002. Effects of alpha-tocopherol and beta-carotene supplementation on gastric cancer incidence in male smokers (ATBC Study, Finland). Cancer Causes Control 13:617–623. Marchini J, Cardon LR, Phillips MS, Donnelly P. 2004. The effects of human population structure on large genetic association studies. Nat Genet 36:512–517. Marchini J, Donnelly P, Cardon LR. 2005. Genome-wide strategies for detecting multiple loci that influence complex diseases. Nat Genet 37:413–417. Epub 2005. Matullo G, Berwick M, Vineis P. 2005. Gene-environment interactions: how many false positives? J Natl Cancer Inst 97:550–551. Mohrenweiser HW, Wilson DM III, Jones IM. 2003. Challenges and complexities in estimating both the functional impact and the disease risk associated with the extensive genetic variation in human DNA repair genes. Mutat Res 526:93–125.

Monni O, Hyman E, Mousses S, Barlund M, Kallioniemi A, Kallioniemi OP. 2001. From chromosomal alterations to target genes for therapy: integrating cytogenetic and functional genomic views of the breast cancer genome. Semin Cancer Biol 11:395–401. Mooney LA, Santella RM, Covey L, Jeffrey AM, Bigbee W, Randall MC, Cooper TB, Ottman R, Tsai WY, Wazneh L. 1995. Decline of DNA damage and other biomarkers in peripheral blood following smoking cessation. Cancer Epidemiol Biomarkers Prev 4:627–634. Moore LE, Huang WY, Chung J, Hayes RB. 2003. Epidemiologic considerations to assess altered DNA methylation from environmental exposures in cancer. Ann NY Acad Sci 983:181–196. Moore LE, Smith AH, Eng C, Kalman D, DeVries S, Bhargava V, Chew K, Moore D, Ferreccio C, Rey OA, Waldman FM. 2002. Arsenic-related chromosomal alterations in bladder cancer. J Natl Cancer Inst 94:1688–1696. Munoz N, Bosch FX. 1997. Biomarkers of biological agents. In: Toniolo P, Boffetta P, Shuker N, Rothman N, Hulka B, Pearce N, editors. Application of Biomarkers in Cancer Epidemiology. Lyon: IARC, pp. 127–142. Nath J, Johnson KL. 2000. A review of fluorescence in situ hybridization (FISH): current status and future prospects. Biotech Histochem 75:54–78. The National Cancer Institute Breast and Prostate Cancer Cohort Consortium. 2005. A candidate gene approach to searching for low-penetrance breast and prostate cancer genes. Nature Rev 5:977–985. National Research Council. 1987. Biological markers in environmental health research; Committee on Biological Markers of the National Research Council. Environ Health Perspect 74:3–9. Neuhouser ML, Patterson RE, Thornquist MD, Omenn GS, King IB, Goodman GE. 2003. Fruits and vegetables are associated with lower lung cancer risk only in the placebo arm of the beta-carotene and retinol efficacy trial (CARET). Cancer Epidemiol Biomarkers Prev 12:350–358. Nicholson JK, Wilson ID. 2003. Opinion: understanding “global” systems biology: metabonomics and the continuum of metabolism. Nat Rev Drug Discov 2:668–676. Nieuwenhuijsen MJ, Droz P. 2003. Biological monitoring. In: Nieuwenhuijsen MJ, editor. Exposure Assessment in Occupational and Environmental Epidemiology. Oxford: Oxford University Press, pp 167–180. Olivier M, Eeles R, Hollstein M, Khan MA, Harris CC, Hainaut P. 2002. The IARC TP53 database: new online mutation analysis and recommendations to users. Hum Mutat 19:607–614. Ollier W, Sprosen T, Peakman T. 2005. UK Biobank: from concept to reality. Pharmacogenomics 6:639–646. Packer BR, Yeager M, Staats B, Welch R, Crenshaw A, Kiley M, Eckert A, Beerman M, Miller E, Bergen A, Rothman N, Strausberg R, Chanock SJ. 2004. SNP500Cancer: a public resource for sequence validation and assay development for genetic variation in candidate genes. Nucleic Acids Res 32:D528–D532. Pavuk M, Cerhan JR, Lynch CF, Kocan A, Petrik J, Chovancova J. 2003. Casecontrol study of PCBs, other organochlorines and breast cancer in Eastern Slovakia. J Expo Anal Environ Epidemiol 13:267–275. Pelkonen O, Raunio H, Rautio A, Lang M. 2003 Xenobiotic-metabolizing enzymes and cancer risk: correspondance between genotype and phenotype. In: Vineis P, Malats N, Lang M, D’Errico A, Caporaso NE, Cuzick J, Boffeta P, editors. Metabolic Polymorphisms and Susceptibility to Cancer, Lyon: IARC Scientific Publications, pp. 77–88. Perera FP. 1987. Molecular cancer epidemiology: a new tool in cancer prevention. J Natl Cancer Inst 78:887–898. Perera FP. 2000. Molecular epidemiology: on the path to prevention? J Natl Cancer Inst 92:602–612. Perera FP, Weinstein IB. 1982. Molecular epidemiology and carcinogen-DNA adduct detection: new approaches to studies of human cancer causation. J Chronic Dis 35:581–600. Perou CM, Sorlie T, Eisen MB, van de RM, Jeffrey SS, Rees CA, Pollack JR, Ross DT, Johnsen H, Akslen LA, Fluge O, Pergamenschikov A, Williams C, Zhu SX, Lonning PE, Borresen-Dale AL, Brown PO, Botstein D. 2000. Molecular portraits of human breast tumours. Nature 406:747–752. Petricoin EF, Ardekani AM, Hitt BA, Levine PJ, Fusaro VA, Steinberg SM, Mills GB, Simone C, Fishman DA, Kohn EC, Liotta LA. 2002a. Use of proteomic patterns in serum to identify ovarian cancer. Lancet 359:572–577. Pfeifer GP, Hainaut P. 2003. On the origin of GÆT transversions in lung cancer. Mutat Res 526:39–43. Phillips DH, Castegnaro M. 1999. Standardization and validation of DNA adduct postlabelling methods: report of interlaboratory trials and production of recommended protocols. Mutagenesis 14:301–315. Poirier MC, Santella RM, Weston A. 2000. Carcinogen macromolecular adducts and their measurement. Carcinogenesis 21:353–359. Potter JD. 1997. Logistics and design issues in the use of biological specimens in observational epidemiology. In: Toniolo P, Boffeta P, Shuker DEG,

Application of Biomarkers in Cancer Epidemiology Rothman N, Hulka B, Pearce N, editors. Application of Biomarkers in Cancer Epidemiology. Lyon: IARC Scientific Publications, pp. 31–37. Pritchard JK, Donnelly P. 2001. Case-control studies of association in structured or admixed populations. Theor Popul Biol 60:227–237. Rappaport SM, Waidyanatha S, Qu Q, Shore R, Jin X, Cohen B, Chen LC, Melikian AA, Li G, Yin S, Yan H, Xu B, Mu R, Li Y, Zhang X, Li K. 2002. Albumin adducts of benzene oxide and 1,4-benzoquinone as measures of human benzene metabolism. Cancer Res 62:1330–1337. Rejnmark L, Vestergaard P, Heickendorff L, Andreasen F, Mosekilde L. 2001. Loop diuretics alter the diurnal rhythm of endogenous parathyroid hormone secretion: a randomized-controlled study on the effects of loopand thiazide-diuretics on the diurnal rhythms of calcitropic hormones and biochemical bone markers in postmenopausal women. Eur J Clin Invest 31:764–772. Rexrode KM, Lee IM, Cook NR, Hennekens CH, Buring JE. 2000. Baseline characteristics of participants in the Women’s Health Study. J Womens Health Gend Based Med 9:19–27. Riboli E, Hunt KJ, Slimani N, Ferrari P, Norat T, Fahey M, Charrondiere UR, Hemon B, Casagrande C, Vignat J, Overvad K, Tjonneland A, Clavel-Chapelon F, Thiebaut A, Wahrendorf J, Boeing H, Trichopoulos D, Trichopoulou A, Vineis P, Palli D, Bueno-De-Mesquita HB, Peeters PH, Lund E, Engeset D, Gonzalez CA, Barricarte A, Berglund G, Hallmans G, Day NE, Key TJ, Kaaks R, Saracci R. 2002. European Prospective Investigation into Cancer and Nutrition (EPIC): study populations and data collection. Public Health Nutr 5:1113–1124. Rich-Edwards JW, Spiegelman D, Garland M, Hertzmark E, Hunter DJ, Colditz GA, Willett WC, Wand H, Manson JE. 2002. Physical activity, body mass index, and ovulatory disorder infertility. Epidemiology 13:184–190. Risch N, Merikangas K. 1996. The future of genetic studies of complex human diseases. Science 273:1516–1517. Rossouw JE, Anderson GL, Oberman A. 2003. Foreward. Ann Epidemiol 13:S1–S4. Rothman KJ, Greenland S. 1998. Modern Epidemiology. Philadelphia: Lippincott-Raven. Rothman N, Bhatnagar VK, Hayes RB, Zenser TV, Kashyap SK, Butler MA, Bell DA, Lakshmi V, Jaeger M, Kashyap R. 1996. The impact of interindividual variation in NAT2 activity on benzidine urinary metabolites and urothelial DNA adducts in exposed workers. Proc Natl Acad Sci USA 93:5084–5089. Rothman N, Cantor KP, Blair A, Bush D, Brock JW, Helzlsouer K, Zahm SH, Needham LL, Pearson GR, Hoover RN, Comstock GW, Strickland PT. 1997. A nested case-control study of non-Hodgkin lymphoma and serum organochlorine residues. Lancet 350:240–244. Rothman N, Garcia-Closas M, Stewart WT, Lubin J. 1999. The impact of misclassification in case-control studies of gene-environment interactions. IARC Sci Publ 148:89–96. Rothman, N, Skibola C, Morgan G, Boffeta P, Wacholder, S., Smith M, Yeager M, Chanock S, Nieters A. 2003. Evaluation of genetic susceptibility for Non-Hodgkin’s Lymphoma in the InterLymph Consortium [abstract]. Presented at the AACR Special Conference: SNPs, Haplotypes and Cancer. Rothman N, Skibola CF, Wang S, et al. 2006. Genetic variation in TNF and IL10 and risk of non-Hodgkin lymphoma: a report from the InterLymph consortium. Lancet Oncology 7:27–38. Rothman N, Stewart WF, Schulte PA. 1995. Incorporating biomarkers into cancer epidemiology: a matrix of biomarker and study design categories. Cancer Epidemiol Biomarkers Prev 4:301–311. Rothman N, Wacholder S, Caporaso NE, Garcia-Closas M, Buetow K, Fraumeni JF Jr. 2001. The use of common genetic polymorphisms to enhance the epidemiologic study of environmental carcinogens. Biochim Biophys Acta 1471:C1–C10. Rowley JD. 1999. The role of chromosome translocations in leukemogenesis. Semin Hematol 36:59–72. Saracci R. 1997. Comparing measurements of biomarkers with other measurements of exposure. In: Toniolo P, Boffetta P, Shuker DEG, Rothman N, Hulka B, Pearce N, editors. Application of Biomarkers in Cancer Epidemiology. IARC Scientific Publ. No. 142. Lyon: IARC, pp. 303–312. Schatzkin A, Gail M. 2002. The promise and peril of surrogate end points in cancer research. Nat Rev Cancer 2:19–27. Schatzkin A, Freedman LS, Schiffman MH, Dawsey SM. 1990. Validation of intermediate end points in cancer research. J Natl Cancer Inst 82:1746–1752. Schiffman M, Adrianza ME. 2000. ASCUS-LSIL triage study: design, methods and characteristics of trial participants. Acta Cytol 44:726–742. Schmidt S, Schaid DJ. 1999. Potential misinterpretation of the case-only study to assess gene-environment interaction. Am J Epidemiol 150:878–885. Schulte PA. 1987. Methodologic issues in the use of biologic markers in epidemiologic research. Am J Epidemiol 126:1006–1016.

87

Schulte PA, Perera FP. 1993. Molecular Epidemiology: Priciples and Practices. San Diego: Academic Press. Schulte PA, Perera FP. 1997. Transitional studies. In: Toniolo P, Boffeta P, Shuker DEG, Rothman N, Hulka B, Pearce N, editors. Application of Biomarkers in Cancer Epidemiology. Lyon: IARC Scientific Publications, pp. 19–29. Schulte PA, Hunter DJ, Rothman N. 1997. Ethical and social issues in the use of biomarkers in epidemiological studies. In: Toniolo P, Boffeta P, Shuker DEG, Rothman N, Hulka B, Pearce N, editors. Application of Biomarkers in Cancer Epidemiology. Lyon: IARC Scientific Publications, pp. 313–318. Schulte PA, Rothman N, Schottenfeld D. 1993. Design considerations in molecular epidemiology. In: Schulte PA, Perera FP, editors. Molecular Epidemiology: Principles and Practices. San Diego: Academic Press, pp. 159–198. Sellers TA, Yates JR. 2003. Review of proteomics with applications to genetic epidemiology. Genet Epidemiol 24:83–98. Sham P, Bader JS, Craig I, O’Donovan M, Owen M. 2002. DNA pooling: a tool for large-scale association studies. Nat Rev Genet 3:862–871. Shen R, Fan JB, Campbell D, Chang W, Chen J, Doucet D, Yeakley J, Bibikova M, Wickham Garcia E, McBride C, Steemers F, Garcia F, Kermani BG, Gunderson K, Oliphant A. 2005. High-throughput SNP genotyping on universal bead arrays. Mutat Res 573:70–82. Simon R, Radmacher MD, Dobbin K. 2002. Design of studies using DNA microarrays. Genet Epidemiol 23:21–36. Smerhovsky Z, Landa K, Rossner P, Brabec M, Zudova Z, Hola N, Pokorna Z, Mareckova J, Hurychova D. 2001. Risk of cancer in an occupationally exposed cohort with increased level of chromosomal aberrations. Environ Health Perspect 109:41–45. Smyth GK, Yang YH, Speed T. 2003. Statistical issues in cDNA microarray data analysis. Methods Mol Biol 224:111–136. Spitz MR, Wei Q, Dong Q, Amos CI, Wu X. 2003. Genetic susceptibility to lung cancer: the role of DNA damage and repair. Cancer Epidemiol Biomarkers Prev 12:689–698. Srinivasan M, Sedmak D, Jewell S. 2002. Effect of fixatives and tissue processing on the content and integrity of nucleic acids. Am J Pathol 161:1961–1971. Staib F, Hussain SP, Hofseth LJ, Wang XW, Harris CC. 2003. TP53 and liver carcinogenesis. Hum Mutat 21:201–216. Starek A. 2003. Estrogens and organochlorine xenoestrogens and breast cancer risk. Int J Occup Med Environ Health 16:113–124. Staudt LM. 2003. Molecular diagnosis of the hematologic cancers. N Engl J Med 348:1777–1785. Steenland K, Bray I, Greenland S, Boffetta P. 2000. Empirical Bayes adjustments for multiple results in hypothesis-generating or surveillance studies. Cancer Epidemiol Biomarkers Prev 9:895–903. Steering Committee of the Physicians’ Health Study Research Group. 1989. Final report on the aspirin component of the ongoing Physicians’ Health Study. N Engl J Med 321:129–135. Steinberg KK, Sanderlin KC, Ou CY, Hannon WH, McQuillan GM, Sampson EJ. 1997. DNA banking in epidemiologic studies. Epidemiol Rev 19:156–162. Sterne JA, Davey SG. 2001. Sifting the evidence: what’s wrong with significance tests? BMJ 322:226–231. Stram DO. 2004. Tag SNP selection for association studies. Genet Epidemiol 27:365–374. Strausberg RL, Simpson AJ, Wooster R. 2003. Sequence-based cancer genomics: progress, lessons and opportunities. Nat Rev Genet 4:409– 418. Swansbury J. 2003. Some difficult choices in cytogenetics. Methods Mol Biol 220:245–257. Tachdjian G, Aboura A, Lapierre JM, Viguie F. 2000. Cytogenetic analysis from DNA by comparative genomic hybridization. Ann Genet 43:147–154. Taioli E. 1999. International collaborative study on genetic susceptibility to environmental carcinogens. Cancer Epidemiol Biomarkers Prev 8:727–728. Tang D, Phillips DH, Stampfer M, Mooney LA, Hsu Y, Cho S, Tsai WY, Ma J, Cole KJ, She MN, Perera FP. 2001. Association between carcinogenDNA adducts in white blood cells and lung cancer risk in the physicians health study. Cancer Res 61:6708–6712. Thomas DC, Witte JS. 2002. Point: population stratification: a problem for case-control studies of candidate-gene associations? Cancer Epidemiol Biomarkers Prev 11:505–512. Thomas DC, Clayton DG. 2004. Betting odds and genetic associations. J Natl Cancer Inst 96:421–423. Thomas DC. 2005. The need for a systematic approach to complex pathways in molecular epidemiology. Cancer Epidemiol Biomarkers Prev 14:557–559.

88

PART I: BASIC CONCEPTS

Tomer KB, Merrick BA. 2003. Toxicoproteomics: a parallel approach to identifying biomarkers. Environ Health Perspect 111:A578–A579. Toniolo P, Boffetta P, Shuker DEG, Rothman N, Hulka B, Pearce N. 1997. Application of Biomarkers in Cancer Epidemiology. Lyon: IARC. Toniolo PG, Levitz M, Zeleniuch-Jacquotte A, Banerjee S, Koenig KL, Shore RE, Strax P, Pasternack BS. 1995. A prospective study of endogenous estrogens and breast cancer in postmenopausal women. J Natl Cancer Inst 87:190–197. Tsugane S, Sobue T. 2001. Baseline survey of JPHC study—design and participation rate; Japan Public Health Center-Based Prospective Study on Cancer and Cardiovascular Diseases. J Epidemiol 11:S24–S29. Tucker JD, Eastmond DA, Littlefield LG. 1997. Cytogenetic End-Points as Biological Dosimeters and Predictors of Risk in Epidemiological Studies. IARC Scientific Publication. Lyon: IARC, pp. 185–200. Vahakangas K. 2003. Molecular epidemiology of human cancer risk: geneenvironment interactions and p53 mutation spectrum in human lung cancer. Methods Mol Med 74:43–59. Van Engeland M, Weijenberg MP, Roemen GM, Brink M, de Bruine AP, Goldbohm RA, van den Brandt PA, Baylin SB, de Goeij AF, Herman JG. 2003. Effects of dietary folate and alcohol intake on promoter methylation in sporadic colorectal cancer: The Netherlands cohort study on diet and cancer. Cancer Res 63:3133–3137. Vermeulen R, Li G, Lan Q, Dosemeci M, Rappaport SM, Bohong X, Smith MT, Zhang L, Hayes RB, Linet M, Mu R, Wang L, Xu J, Yin S, Rothman N. 2004. Detailed exposure assessment for a molecular epidemiology study of benzene in two shoe factories in China. Ann Occup Hyg 48:105–116. Wacholder S. 1991. Practical considerations in choosing between the casecohort and nested case-control designs. Epidemiology 2:155–158. Wacholder S, Chanock S, Garcia-Closas M, El Ghormli L, Rothman N. 2004. Assessing the probability that a positive report is false: an approach for molecular epidemiology studies. J Natl Cancer Inst 17:434–442. Wacholder S, Garcia-Closas M, Rothman N. 2002a. Study of genes and environmental factors in complex diseases. Lancet 359:1155. Wacholder S, Rothman N, Caporaso N. 2002b. Counterpoint: bias from population stratification is not a major threat to the validity of conclusions from epidemiological studies of common polymorphisms and cancer. Cancer Epidemiol Biomarkers Prev 11:513–520. Wall JD, Pritchard JK. 2003. Haplotype blocks and linkage disequilibrium in the human genome. Nat Rev Genet 4:587–597. Wang L, Liu S, Niu T, Xu X. SNPHunter. 2005. A bioinformatic software for single nucleotide polymorphism data acquisition and management. BMC Bioinformatics 18:60. Wang W, Zhou H, Lin H, Roy S, Shaler TA, Hill LR, Norton S, Kumar P, Anderle M, Becker CH. 2003. Quantification of proteins and metabolites by mass spectrometry without isotopic labeling or spiked standards. Anal Chem 75:4818–4826. Warner M, Eskenazi B, Mocarelli P, Gerthoux PM, Samuels S, Needham L, Patterson D, Brambilla P. 2002. Serum dioxin concentrations and breast cancer risk in the Seveso Women’s Health Study. Environ Health Perspect 110:625–628.

Wester K, Wahlund E, Sundstrom C, Ranefall P, Bengtsson E, Russell PJ, Ow KT, Malmstrom PU, Busch C. 2000. Paraffin section storage and immunohistochemistry: effects of time, temperature, fixation, and retrieval protocol with emphasis on p53 protein and MIB1 antigen. Appl Immunohistochem Mol Morphol 8:61–70. White E. 1997. Effects of biomarker measurement error on epidemiological studies. In: Application of Biomarkers in Cancer Epidemiology. IARC Scientific Publ. No. 142. Lyon: IARC, pp. 73–93. Wild CP, Pisani P. 1997. Carcinogen-DNA and carcinogen-protein adducts in molecular epidemiology. In: Toniolo P, Boffetta P, Shuker DEG, Rothman N, Hulka B, Pearce N, editors. Application of Biomarkers in Cancer Epidemiology. Lyon: IARC, pp. 143–158. Wild CP, Andersson C, O’Brien NM, Wilson L, Woods JA. 2001. A critical evaluation of the application of biomarkers in epidemiological studies on diet and health. Br J Nutr 86(Suppl)1:S37–S53. Woodson K, Tangrea JA, Pollak M, Copeland TD, Taylor PR, Virtamo J, Albanes D. 2003. Serum insulin-like growth factor I: tumor marker or etiologic factor? A prospective study of prostate cancer among Finnish men. Cancer Res 63:3991–3994. Wulfkuhle JD, Liotta LA, Petricoin EF. 2003. Proteomic applications for the early detection of cancer. Nat Rev Cancer 3:267–275. Wynder EL, Hoffmann D. 1994. Smoking and lung cancer: scientific challenges and opportunities. Cancer Res 54:5284–5295. Yang Q, Khoury MJ, Sun F, Flanders WD. 1999. Case-only design to measure gene-gene interaction. Epidemiology 10:167–170. Yoo KY, Shin HR, Chang SH, Lee KS, Park SK, Kang D, Lee DH. 2002. Korean Multi-center Cancer Cohort Study including a Biological Materials Bank (KMCC-I). Asian Pac J Cancer Prev 3:85–92. Zeggini E, Rayner W, Morris AP, Hattersley AT, Walker M, Hitman GA, Deloukas P, Cardon LR, McCarthy MI. 2005. An evaluation of HapMap sample size and tagging SNP performance in large-scale empirical and simulated data sets. Nat Genet [Epub ahead of print]. Zeleniuch-Jacquotte A, Shore RE, Koenig KL, Akhmedkhanov A, Afanasyeva Y, Kato I, Kim M, Rnaldi S, Kaaks R, Toniolo P. 2004. Postmenopausal levels of oestrogen, androgen, and SHBG and breast cancer: long-term results of a prospective study. Br J Cancer 90:153–159. Zhang L, Eastmond DA, Smith MT. 2002. The nature of chromosomal aberrations detected in humans exposed to benzene. Crit Rev Toxicol 32:1–42. Zhang L, Rothman N, Wang Y, Hayes RB, Yin S, Titenko-Holland N, Dosemeci M, Wang YZ, Kolachana P, Lu W, Xi L, Li GL, Smith MT. 1999. Benzene increases aneuploidy in the lymphocytes of exposed workers: a comparison of data obtained by fluorescence in situ hybridization in interphase and metaphase cells. Environ Mol Mutagen 34:260–268. Zheng S, Ma X, Buffler PA, Smith MT, Wiencke JK. 2001. Whole genome amplification increases the efficiency and validity of buccal cell genotyping in pediatric populations. Cancer Epidemiol Biomarkers Prev 10:697–700. Zheng W, Chow WH, Yang G, Jin F, Rothman N, Blair A, Li HL, Wen W, Ji BT, Li Q, Shu XO, Gao YT. 2005. The Shanghai Women’s Health Study: Rationale, Study Design, and Baseline Characteristics. Am J Epidemiol [Epub ahead of print].

7

Genetic Concepts and Methods in Epidemiologic Research NEIL J. RISCH AND ALICE S. WHITTEMORE

A

s recently as the 1970s, the etiology of cancer was in a black box. Theories of carcinogenesis abounded; and in the absence of any hard evidence to confirm or refute them, there seemed little hope that cancer would soon be understood or that patients with cancer could be treated effectively. This situation has changed dramatically, largely because of advances in molecular genetics. Now we understand that cancer is in essence a genetic disease, one that arises from mutations in the chromosomes that comprise the genome. Although the causes of cancer are complex and nongenetic factors clearly play important roles in the neoplastic process, major advances have occurred recently because of the identification of genes that, when mutated, lead to cancer. We now understand that these mutations can be inherited in the germ cells that fuse at conception, or that they can occur somatically in stem cells whose clones become cancers. The advances achieved during the last 30 years have generated much work for cancer epidemiologists, who must work with geneticists to identify genes that alter cancer risk and then translate the genetic discoveries into risk estimates and cost/benefit analyses that lead to effective strategies for cancer prevention and control. In this chapter we describe this work as it relates to heritable germline mutations and their roles in cancer susceptibility. We begin by describing the evidence that such mutations are important in specific cancers. We then describe methods for identifying these mutations and discuss how to characterize their effects on risk. For a description of current progress in understanding how genetic mutations interact with each other and with personal attributes and environmental exposures, we refer the reader to Chapter 29. We conclude by considering the challenge for the future: how to use information on an individual’s genetic susceptibility to prevent cancer occurrence. For additional reading, we recommend reading Haiman and Hunter’s (2002) thoughts on the genetic epidemiology of cancer. These authors also offer a glossary of commonly used terms in human genetics and genetic epidemiology.

INHERITED SUSCEPTIBILITY TO CANCER Evidence of a Genetic Contribution Familial aggregation of a trait is a necessary but not sufficient condition to infer the importance of genetic susceptibility, as environmental and cultural influences can also aggregate in families, leading to family clustering and increased familial risk. Family aggregation is usually assessed by studying relatives of affected subjects and contrasting their rates of illness with those of a suitable control group, typically the relatives of unaffected subjects. Several approaches for disentangling genetic from environmental influences are also possible in studies of human disease, although practical difficulties often limit their use. The most powerful design compares risks for biologic relatives of affected versus unaffected adoptees, as adoption creates a separation between an individual’s biologic and environmental influences. Because it is often difficult to obtain access to information on the biologic relatives of adoptees, adoption studies typically focus only on common disease. Another study design often used to distinguish genetic and environmental influences involves twins. Identical (monozygotic, or MZ) twins derive from fission of a single fertilized egg and thus inherit identical genetic material. By contrast, fraternal (dizygotic, or DZ) twins derive from two distinct fertilized eggs, and thus have the same

genetic relationship as full siblings, although they may be more “biologically” related because of sharing the same prenatal intrauterine experience. Comparing the similarity of MZ twins with same-sex DZ twins is a common approach for gleaning the degree of genetic influence on a disease or trait, and it has been applied extensively to a broad range of disorders, including cancer. A standard measure of similarity employed in twin studies is the concordance rate. The “pairwise” concordance is calculated simply as the proportion of twin pairs with both twins affected among all ascertained twin pairs with at least one affected. In contrast, “probandwise” concordance allows double counting of doubly ascertained twin pairs and has the advantage of being interpretable as the recurrence risk in a co-twin of an affected individual (Khoury et al., 1993). Usually, the most critical assumption in twin studies is that MZ and DZ twins display a comparable degree of similarity owing to the sharing of environmental factors, so the difference in concordance rates between MZ and DZ twins is a reflection only of genetic factors.

Familial Aggregation and Age of Onset Numerous studies have addressed the degree to which site-specific cancers run in families (e.g., breast, colon, prostate, lung). However, few large-scale studies using a single standardized approach to many cancer sites have been reported. Such studies are useful for deriving a global view of the familiality of cancer. One such study from Utah examined the familial recurrence of cancer for 28 specific sites based on data from 35,228 individuals with cancer, called probands (Goldgar et al., 1994). These authors matched the Utah Genealogic Database, which provides names of all firstdegree relatives of these probands, with the Utah Cancer Registry to determine the frequency of cancer among these first-degree relatives. Comparable expected rates for each cancer site were obtained from the 399,786 first-degree relatives of a matched control group. Familiality (familial risk ratio, or FRR) was assessed as the ratio of the observed number of cancer cases among the first-degree relatives of the probands divided by the expected number derived from the control relatives based on the years of birth (cohort) of the case relatives. In essence, the FRR provides an age-adjusted risk ratio for the firstdegree relatives of cases compared to the general population (Risch, 1990). These authors also examined the FRR for a separate group of early-onset cancer probands—diagnosis prior to age 50 years for melanoma, breast, and brain/central nervous system cancers and prior to age 60 years for all other cancers. A second large population-based study of family recurrence for a variety of cancer sites has been reported (Dong and Hemminki, 2001; Hemminki et al., 2001). The authors linked the Swedish Cancer Registry to Swedish family records to study cancer incidence during the period 1958–1996 for a cohort of more than two million individuals born in Sweden since 1934. Incidence rates and standardized incidence ratios (SIRs) were calculated for the following subgroups of individuals determined by the cancers in their first-degree relatives: a parent but no sibling with cancer; a sibling but no parent with cancer; both a parent and a sibling with cancer. Incidence rates and SIRs also were calculated for the spouses of these individuals. Among 4,225,232 parents, 435,000 (10.3%) had a diagnosis of cancer. Among 5,520,756 offspring who were born after 1934 and followed until 1996, there

89

90

PART I: BASIC CONCEPTS

Table 7–1. Familial Recurrence Risk for First-Degree Relatives and Spouses of Cancer Probands by Site, in Decreasing Order of Prevalence Sweden Utah: All First-Degree Relatives Cancer Site Prostate Breast Colorectum Lung Endometrium Skin (melanoma) Bladder Non-Hodgkin lymphoma Brain/CNS Cervix Ovary Stomach Lip Pancreas Kidney Oral cavity Thyroid Multiple myeloma Granulocytic leukemia Lymphocytic leukemia Hodgkin disease Female genitalia Soft tissues (sarcoma) Testis Gallbladder Larynx Total (median)

Offspring

Spouses

All Probandsa

Early-Onset Probands

All Probandsb

Early-Onset Probands

Siblings: All Probands

All Probands

Early-Onset Probands

2.21 1.83 2.54 2.55 1.32 2.10 1.53 1.68 1.97 1.73 2.04 2.08 2.72 1.25 2.46 1.82 8.48 4.29 2.94 5.00 1.25 2.22 2.00 8.57 2.22 8.00 2.15

4.08 3.70 4.53 2.50 1.75 6.43 5.00 2.40

2.82 1.81 2.00 1.77 — 2.60 1.57 1.59 1.78 1.89 2.81 1.62

— 2.49 7.87 2.11 — 3.93 3.33 2.03 3.14 1.93 5.81 4.50

9.41 2.01 4.41 3.16 — 3.41 3.30 2.37 2.37 2.39 2.52 8.82

— 1.32 1.00 1.43

— 1.40 1.22 1.46

1.16 1.13 0.93 1.05 — — 1.26

1.81 0.82 0.96 0.94 — — 1.25

1.71 1.82

10.00 5.00

5.26

1.18 0.99

2.94 1.74

7.04 2.58 2.04

15.63 — 2.00

12.42 5.62 3.53

1.08 0.37 1.04

0.43 0.00 0.93

2.46

5.26

3.97

4.29

5.00

8.50

1.95

4.22

3.53

1.08

0.97

4.08

Source: Adapted with permission from Risch (2001). Data for Utah are from Goldgar et al. (1994) and data for Sweden are from Dong and Hemminki (2001) and Hemminki et al. (2001). a Cancer cases identified in Utah registry. b Individuals in cohort.

were 71,424 (1.3%) with a diagnosis of cancer. Among male offspring, the average age at diagnosis was 38 years; for female offspring, the mean age of diagnosis was 42 years. Results of reanalyses of both of the above data sets described previously (Risch, 2001) using FRRs and SIRs, respectively, are shown in Table 7–1. An important question is whether the rare cancer sites are less familial than the common sites. Thus, in Table 7–1 the various sites are listed in decreasing order of prevalence as reported in Utah. The FRRs are given for all and for early-onset probands in both studies. There are four important observations to be derived from Table 7–1. First, there is remarkable similarity of the FRR across cancer sites, with a median value of 2.15 for Utah (all probands) and a median of 1.95 for the Swedish offspring. There are a few notable exceptions, however. Thyroid, testicular, and laryngeal cancers and lymphocytic leukemia appear to have elevated recurrence risks in one or both studies. Also in both studies, all FRRs are greater than 1, and 18 of the 26 sites (Utah) and 16 of the 18 sites (Sweden) have an FRR between 1.5 and 3.0. Furthermore, there is overall consistency in the FRRs from Utah (all probands) and Sweden (offspring). For the 18 sites in common between the two studies, the correlation in FRRs is 0.83. However, this correlation is primarily due to the high values for thyroid cancer, melanoma, and testicular cancer in both studies. After removing the data for these three sites, the correlation becomes 0.01. This observation probably reflects a lack of true variation around the average FRR of 2 for the remaining sites, the observed variation being primarily random (i.e., statistical noise). Second, it is apparent from Table 7–1 that there is no decline in FRR with decreasing frequency of the cancer site. In fact, if anything, there is a trend toward an increasing FRR with decreasing frequency. For example, in Utah, for the first 13 cancers listed in Table 7–1, the median FRR is 2.04, and for the second 13 cancers it is 2.46. Simi-

larly, in Sweden, for the first nine sites listed, the median FRR for offspring is 1.81, and for the latter group of nine sites it is 2.85. Thus, when characterized by the FRR, the rare cancers are no less familial (and probably more familial) than the common cancers. They may appear “sporadic” because they are rare and most often occur in the absence of a family history (i.e., families with multiple cases are rare). However, when assessed systematically, relatives of cases with rare cancers have at least the same degree of increased risk (or more) compared with the relatives of cases with common cancers. Third, Table 7–1 reveals increased family recurrence associated with early age at diagnosis. In Utah, for the nine cancer sites listed, the median FRR for the early-onset probands was 4.08. This figure is nearly twofold higher than the FRR for the same nine sites for all probands (median 1.97). Eight of the nine sites listed showed an increased FRR with early onset (only lung cancer did not). In Sweden, the FRR is similarly elevated when we consider only the probands with early onset. In this study, the median FRR increased to 4.2. The increase in risk in both studies for relatives of early-onset versus lateonset probands appears, on average, to be about twofold. Thus, it appears to be a generalizable conclusion that increased familiality is associated with early age of diagnosis. In Sweden, the sibling recurrence ratios (median 3.53) are systematically higher than the offspring recurrence ratios (median 1.95). The average age at diagnosis of the index cases who were used to assess sibling risk was only 38–42 years on average, much younger than the average age of diagnosis of the affected parent index cases who were used in the estimation of risk to offspring (probably by 30 years or so). This was due to the fact that offspring were young at the time of the study (maximum age 61 years), whereas the parents were not. Thus, the elevated risks in siblings versus offspring of cancer probands, as observed in the Swedish data, is a reflection of increased familial risk associated with early age at diagnosis. Once age and age

91

Genetic Concepts and Methods in Epidemiologic Research at diagnosis are taken into account, it appears that offspring and sibling recurrence risks are similar. Fourth, the evidence regarding spouses in the Swedish data supports the familial aggregation in first-degree relatives as having a genetic basis. As seen in Table 7–1, the median FRR for spouses is 1.1 (range 0.4–1.4). Furthermore, there was no difference in the FRRs for spouses based on the age of onset in the proband, with a median of 1.0 for early-onset probands versus 1.1 for all probands. This observation also supports a genetic origin for the early-onset phenomenon observed for first-degree relatives.

Table 7–2. Twin Recurrence Risks: By Zygosity and Cancer Site

DISTINGUISHING GENES FROM ENVIRONMENT: ADOPTEES AND TWINS

female subjects

Because of the difficulty of conducting adoption studies, only one such study in cancer has been reported (Sorensen et al., 1988). This study examined cancer mortality in the 593 biologic (but adopted away) offspring of parents who died of cancer (all sites combined) by age 50 years. The authors observed a nonstatistically significant death rate ratio of 1.2 for cancer mortality (all sites) among the offspring. The number of subjects was far too small to examine individual cancer sites. The implication of pooling all cancer sites is discussed below. With respect to twin studies, because most cancers are rare and occur late in life, large twin cohorts are usually required to obtain sufficient cases. Thus, few studies of cancer in twins have been reported, and the ones that have focused primarily on the commonly occurring cancer sites or on all sites combined. For example, the National Academy of Sciences (NAS) Twin Cohort, containing nearly 16,000 male veteran twins, revealed no increased concordance in lung cancer mortality in MZ versus DZ twins (despite an observed increase in concordance for cigarette smoking in the MZ twins). This led the authors to conclude that genetic susceptibility has little influence on lung cancer mortality (Braun et al., 1995). The same cohort was also studied for death from all cancer sites combined. Here the MZ/DZ concordance ratio was 1.4, modestly suggestive of genetic influence. This cohort was also evaluated for prostate cancer risk (Page et al., 1997). In this case, the MZ concordance was estimated to be 27.1% compared to 7.1% for DZ twins, giving a concordance ratio of 3.8—strong evidence for the influence of genetic susceptibility. Instead of large population-based twin cohorts, an alternative strategy is to identify twins from a large sample of cancer cases and follow their co-twins for their cancer risk. This approach was used to study Hodgkin’s disease (Mack et al., 1995); 366 (179 MZ, 187 DZ) twins with the disease were identified, and their co-twins were followed. Of the 179 MZ co-twins, 10 similarly developed Hodgkin’s disease, compared with none of the 187 DZ co-twins, suggesting a strong heritable component to this form of cancer. The Nordic countries are an ideal setting for population-based twin studies because of the existence of population-based twin and cancer registries. For example, two Swedish twin cohorts, one of subjects born between 1886 and 1925 (10,503 pairs) and another of subjects born between 1926 and 1958 (12,883 pairs) were linked to that nation’s cancer registry (Ahlbom et al., 1997). The authors found increased concordance in MZ versus DZ twins for colorectal, breast, cervical, and prostate cancers, suggesting the importance of genetic factors for these sites. By contrast, MZ and DZ concordances were comparable for stomach and lung cancers, suggesting less of a genetic role in these cancers. Similarly, in Finland, 12,941 same-sex twin pairs were linked to that country’s cancer registry (Verkasalo et al., 1999). Examining all sites combined, the authors estimated a low overall influence of genetic factors and thus concluded that the environment plays the major role in cancer susceptibility. Most recently, the twin registry of Denmark was linked to that nation’s cancer registry and combined with similar analyses in Sweden and Finland to produce the largest population-based twin study of cancer to date (Lichtenstein et al., 2000). In total, 44,788 same-sex twins were followed for cancer prevalence.

Site

MZ twins (lM)

DZ twins (lD)

RMDa

2.4 1.7 1.5 1.0 1.0 0.1b

8.06 6.27 5.87 8.49 5.94 10.38 7.41

2.83 6.14 5.14 5.96 1.67 4.93 4.27

3.86 1.03 1.18 1.51 7.37 2.39 1.96

3.6 1.5 0.3b

4.09 10.46 6.27 5.29

2.51 3.95 3.32 2.84

2.05 3.21 2.27 2.33

6.14 7.61

3.35 4.02

2.19 2.19

Prevalence (%)

male subjects Prostate Lung Colon Stomach Bladder Other Total Breast Colon Other Total

male + female subjects Total Total without breast Source: Risch (2001). a RMD = (lM - 1)/(lD - 1). b Averaged over all other sites.

Because of its large size, the twin concordance data reported by Lichtenstein et al (2000) were recently reexamined (Risch, 2001). Estimates for the MZ recurrence risk ratio (lM) and DZ recurrence risk ratio (lD) were derived, along with the ratio RMD = (lM - 1)/(lD - 1). Most sites had cancer rates that are too low to allow reliable calculation of l values, so they were calculated individually only for sites with a prevalence of at least 1%. A weighted average estimate was derived for the remaining sites. Results are given in Table 7–2 individually for prostate, lung, colon, stomach, and bladder cancer among men, with a weighted average estimate for the remaining 20 sites, and for breast and colon cancer in women, with a weighted average estimate for the other 24 sites. The sites are listed in order of decreasing prevalence. As can be seen in Table 7–1, the values of lM and lD are reasonably constant across sites of varying prevalence, as are the values of RMD. However, statistical analysis showed that a model of identical l’s across cancer sites is rejected (Risch, 2001). The poor fit is due to the discrepancy between female breast cancer (the most common site), which has lower values of lM and lD than the other sites. Testing the model of a constant value of lM and lD for all sites other than female breast gave an excellent fit to the data with lM = 7.61 and lD = 4.02. The results of this analysis suggest that the values of M and lD are reasonably consistent across individual cancer sites. Most of the cancer sites listed in Table 7–1 have an FRR close to 2, and most sites also have stable values for lM and lD and for RMD, as given in Table 2, although for the rarer cancers reliable values of lM and lD are not possible. Based on the data in Table 7–1, a few sites have particularly low or high FRRs. For example, uterine and pancreatic cancer and Hodgkin’s disease all have an FRR of less than 1.32. Based on the twin data (Lichtenstein et al., 2000), uterine cancer has a lM of 2.2 and lD of 4.7. Both of these values are higher than the FRR of 1.32 (Table 7–2), but the higher lD than lM is also not consistent with a role of genetic susceptibility. For pancreatic cancer, for males and females combined, lM = 11.0 and lD = 1.7. The low value of lD is comparable to the observed FRR of 1.25, but the higher lM value is suggestive of genetic susceptibility. For Hodgkin’s disease (FRR = 1.25), too few cases were observed in the large populationbased twin study for meaningful analysis. However, another study (Mack et al., 1995) found 0 of 187 (0%) of DZ twins versus 10 of 179 (5.6%) of MZ twins to be concordant. The high observed lM in that study is again suggestive of genetic susceptibility. For the sites in Table 7–1 with high FRRs (thyroid, multiple myeloma, leukemia, larynx, testis), most were too rare to obtain individual lM and lD estimates from the large twin study (Lichtenstein

92

PART I: BASIC CONCEPTS

et al., 2000). However, combining across all five of these sites, weighted averages of lM = 17.1 and lD = 5.6 are obtained. These values are higher than the weighted average for all sites combined (even excluding breast) as given in Table 7–2, suggesting that for these rare sites genetic influence may be more prominent, consistent with the FRR results in Table 7–1.

INHERITED SUSCEPTIBILITY: SITE-SPECIFIC OR GENERALIZED? Another interesting observation from the study of Lichtenstein et al. (2000) is the MZ and DZ concordances when all cancer sites were grouped together and analyzed as a single entity. For example, such an analysis considers pairs concordant if one twin has lung cancer and the other colon cancer. For male MZ pairs, 262 were considered concordant versus 1252 discordant, giving lM = 2.40. For male DZ pairs, 356 were concordant and 2459 discordant, or lD = 1.95; also, RMD = 1.47. For female MZ pairs, 265 were concordant and 1487 discordant, or lM = 2.2; and for female DZ pairs 408 were concordant and 3023 discordant giving lD = 1.70, and RMD = 1.7. These values of l and RMD are considerably attenuated from the corresponding numbers calculated from site-specific analyses. This observation indicates that the inherited predisposition to cancer is likely to involve many genes that are primarily (but not entirely) site-specific. Goldgar et al. (1994) also examined all pairs of cancer sites in their probands and affected first-degree relatives to ascertain possible genetic relatedness of susceptibility to various cancer sites. It involved consideration of 1026 comparisons. Despite the large number of tests, many were deemed to be statistically significant. However, from a global perspective, considering all site pairs, the FRR was considerably reduced compared to “within-site” estimates, consistent with the observations from the twin data. Also, the analysis of the Swedish family study considered across-site comparisons (Dong et al., 2001). Although many of these comparisons were statistically significant, they also generally found that the highest risk ratios were associated with site-specific recurrence. The conclusion of site specificity is also consistent with molecular results, which to date have shown gene effects to be largely site-specific (e.g., for colon cancer, melanoma) and/or with limited range (e.g., breast/ovarian cancer).

GENETIC MODELS AND MODE OF INHERITANCE Understanding empirical evidence about genetic susceptibility to cancer requires consideration of models of genetic inheritance and their implications. The simplest way to measure genetic effects is through familial recurrence risks as described above. If genetic susceptibility is due to a single dominant or additive gene, it is easy to show (Risch, 1990) that lP = lO = lS = lD = (lM +1)/2, which implies that the MZ/DZ ratio defined by RMD = (lM - 1)/(lD 1) = 2 (here subscripts P, O, and S correspond to parents, offspring, and siblings, respectively). On the other hand, if susceptibility is due to a recessive gene, lP = lO < lS = lD, with the degree of difference between lS and lO depending on the frequency of the “at-risk” allele (lS/lO ranging from near 1 for a very common allele to infinity for a very rare allele). For a recessive model, RMD is usually higher than 2, again depending on the allele frequency. For a rare allele, RMD = 4 but diminishes toward 2 if the allele is very common. How are these expectations altered if there are also nongenetic cases mixed in or if more than one locus contributes to susceptibility? Nongenetic cases (phenocopies) do not influence the predictions given above. On the other hand, if more than one gene exists that influences susceptibility, the predictions may be altered, depending on whether interaction effects exist among the contributing genes—epistasis in genetics parlance. Specifically, if mutant alleles at different loci are individually rare so it is highly unlikely that an individual would carry more than one (a scenario typically termed genetic heterogeneity, locus heterogeneity, or nonallelic heterogeneity by geneticists), the same predictions as given above hold. Equivalently, for more common

alleles, if the risk associated with carrying multiple mutants is additive, the same predictions hold (Risch, 1990). By contrast, if the risk associated with carrying multiple “at-risk” alleles is not additive (e.g., multiplicative) a different pattern for the l values than described above occurs. Specifically, RMD is now higher than 2 and can achieve extremely high values depending on how many loci are involved and the degree of interaction. Another genetic model commonly employed in the analysis of family and twin data is the polygenic, or multifactorial, threshold (MFT) model. This model postulates a genetic basis consisting of numerous small, additive effects underlying a continuously distributed trait termed liability. The assumptions of the model imply a gaussian distribution for liability due to the central limit theorem. It is further assumed that risk, as a function of liability, increases sigmoidally (asymptotically to zero for liability equal to minus infinity and to 1 for liability equal to plus infinity). This sigmoid risk function is assumed to take the form of a cumulative normal distribution function. It can be shown that the latter assumption is mathematically equivalent to assuming an independent, additive random environmental component to liability, with a threshold imposed on the total liability scale determining affected status (e.g., a total liability value above a threshold T implies affected and below T unaffected). Thus, according to the MFT model, there are two additive, normally distributed components to liability: a polygenic component and a random environmental component. The proportion of the total variance of liability due to the polygenic component is usually termed “heritability,” where it is understood that this refers to the heritability of the latent liability trait. Because it is based on the underlying liability variable, heritability is independent of the threshold T. As was shown previously (Risch, 2001), for the MFT model RMD should range from about 2.5 for a common cancer (prevalence 3%) to around 4.0 (depending on H) for a rare cancer (prevalence 0.1%). Thus, the observed values of RMD in Table 7–2 conform poorly to the predictions of the MFT model but quite well to the single locus or additive genetic model described above, which predicts RMD = 2. Thus in general, it is more likely that genetic susceptibility to cancer entails dominant and/or additive gene effects across contributing loci than genetic interactions. It is also of interest to compare the numbers in Tables 7–1 and 7–2. For dominant gene effects, the FRR, as given in Table 7–1, should correspond to lD of Table 7–3. Because in Utah the FRR was based on all first-degree relatives, including parents and offspring as well as siblings, the FRR, in theory, might be less than lD if recessive genes are involved in cancer susceptibility. In fact, the average FRR for Utah estimated in Table 7–1 is around 2.1, compared with 3.4–4.0 for lD observed in Table 7–2. Similarly, the FRR for Sweden is lower for offspring (1.9) than for siblings (3.5). At first glance, these observations suggest the presence of recessive genes. However, it is more likely that these differences are a result of age and age-of-onset differences between studies. In contrast to the results of Goldgar et al. (1994), which were based on age-adjusted lifetime rates, the prevalence figures from which Table 7–2 was derived do not correspond to lifetime risks. This is because the cohorts of twins were surveyed for cancer risk only during a defined and limited period of time; that is, they were both left-censored and right-censored (e.g., Marenberg et al., 1994). Specifically, Swedish cohort I entered observation between ages 36 and 75 and was followed for 34 years (to ages 70–109 or death); 4490 of 21,006 subjects (21%) had a diagnosis of cancer. Similarly, the Danish cohort entered their study between ages 13 and 73 and were followed for 50 years to age 63–123 (or death); 3572 of 16,922 subjects (21%) had a cancer diagnosis. By contrast, Swedish cohort II entered observation at ages 14–46 and were followed for only 22 years (to ages 36–68); not surprisingly, only 1157 cancer diagnoses were made in this group of 25,716 subjects (i.e., 4.5%). Similarly, the Finnish cohort entered their study at ages 18–96 and were followed for only 20 years to age 38–116 or death. There were 1584 cancer diagnoses among the 25,882 subjects (6.1%). Swedish cohort II and the Finnish cohort represent more than half of all the twin pairs (25,824/44,788, or 58%). Most of the cancers in these two twin cohorts have yet to occur.

93

Genetic Concepts and Methods in Epidemiologic Research Table 7–3. Methods for Identifying Inherited Cancer-Susceptibility Genes Type

linkage studies

Nonparametric Parametric

association studies

Case-control Family-based

studies of chromosomal gain/loss CGH Differential gene expression

Strategy

Strengths

Genotype affected family members for a set of closely spaced markers of known location; look for allele-sharing among members that exceeds mendelian expectation Tests null hypothesis of mendelian allelesharing probabilities among affected family members Requires a model of disease inheritance and cancer risks associated with genotypes of putative cancer gene

Does not require knowledge of gene function; is robust with respect to how the multiple-case families are ascertained

Weaknesses

Does not require specification of a genetic model

Poor power to detect regions containing genes of moderate to low relative risk; decreased power in the presence of locus heterogeneity Lower power than parametric in mendelian case

Good power when genetic model is correctly specified

May lead to reduced power when genetic model is misspecified

Genotype affected and unaffected subjects for a set of closely spaced markers of known location; look for differences in specific genotypes between affected and unaffected Compares genotypes of cases and unrelated controls Compares alleles of cases to those not transmitted by their parents or to those of their unaffected siblings

Good power to detect common variants of low to moderate relative risk

Poor power when variant allele is rare, and poor power in the presence of weak linkage disequilibrium between disease locus and markers

Simple to conduct; results are easily interpreted and explained Less powerful than case-control comparisons; may require parental genotypes, which may be unavailable for late-onset cancers

May be biased by ethnic stratification Controls for ethnic stratification; allows separate evaluation of genes transmitted maternally or paternally

Comparison of attributes of cancer cells to normal cells in same tissue

Can study large numbers of genes with no advance knowledge of their functions Does not require advance knowledge of gene function Does not require advance knowledge of gene function; can evaluate gene–gene interactions

Abnormalities detected may be effect rather than cause of malignancy Does not distinguish cause from effect Requires RNA from snap-frozen tissue; does not distinguish cause from effect

Compares copy number of cancer cells to those of normal cells in same tissue Evaluates over- or underexpression of large numbers of genes in different cells types

The values of lM and lD given in Table 7–2 are likely to be strongly influenced by the age structure of the sample. As we showed above (Table 7–1), the familiality of many cancers is greater at an earlier age of diagnosis, and thus l values decrease with age. The numbers provided in Table 7–2 correspond to cancers diagnosed primarily in midlife. Indeed, the median FRRs for cancers occurring in midlife (before age 50 or 60), as in Table 7–1, is 3.8 for Utah and 4.2 in Sweden, close to the average value of lD in Table 7–2, as well as the sibling recurrence risk ratio from Sweden (Table 7–1). Thus, it appears most likely that the more modest values of FRR in Table 7–1 (all relatives in Utah, offspring risks in Sweden) versus values of lD in Table 7–2 reflects the different age structure and follow-up of the twin samples rather than the presence of recessive genes. One possible exception to the model of inheritance we described is for prostate cancer. For breast and colon (sex-averaged) cancers RMD was close to 2.0, whereas for prostate cancer RMD was estimated at 3.86 (Table 7–2). A previous, comparably sized twin study of prostate cancer (Page et al., 1997) found an MZ concordance of 27.1% and DZ concordance of 7.1% versus a prevalence of 3.17%. These rates translate into values of lM = 8.55, lD = 2.24, and RMD = 6.09. The lM and lD values for both studies are quite similar, and the RMD value appears to be significantly greater than 2.0. Thus, it may turn out for this cancer site that the genetic basis is not explained by independent, rare, autosomal dominant mutations but, rather, by multiple interacting loci. What about cancers that have an identified, major environmental component such as lung cancer and cigarette smoking? According to Table 7–1, lung cancer appears to be familial (FRR = 1.7 - 3.2), but the twin data provide nearly equal values of lM (6.27) and lD (6.14) in males. The latter suggests a strong environmental effect shared by twins (i.e., smoking behavior) rather than a genetic component. Ironically, twin studies have consistently shown greater concordance for smoking behavior in MZ twins than DZ twins. This clearly is an example of an environmental exposure being confounded with genetic influence in a twin study paradigm. Yet, paradoxically, this concordance difference in smoking behavior is not reflected in a concordance

difference for lung cancer. A comparable study of U.S. male twins (Braun et al., 1995) found the same thing—greater concordance in smoking for MZ versus DZ twins yet no difference in concordance for lung cancer. On the other hand, lung cancer in female twins (Lichtenstein et al., 2000), for whom the prevalence is much lower, does appear to follow a more genetic pattern (lM = 21.3 and lD = 1.76) although these figures are based on small numbers (Hoover, 2000). When a major environmental exposure is involved in cancer susceptibility, the question becomes: Are there specific genes that increase the risk of cancer in exposed individuals? In unexposed individuals? Are these genes the same? In theory, family and twin studies can address these questions. For example, if the genes are the same, the risk of cancer should be increased in family members who are both exposed and unexposed when the index subject is exposed or unexposed. Different genetic mechanisms would imply that only exposed family members of exposed probands are at increased risk. Although the numbers are small, the lung cancer twin data for females versus males is suggestive of more pronounced genetic influence on unexposed or less exposed individuals.

ESTIMATES OF GENETIC CONTRIBUTION Heritability, as defined above in the context of the MFT model, is often used as a measure of the importance of genetic effects. For example, in the study of Lichtenstein et al. (2000), because the heritabilities were generally estimated at less than 30% the authors concluded that genetic effects are minor relative to the environmental impact. To quantify the impact of risk factors, epidemiologists use several other measures, most notably relative risk (RR) (risk to exposed versus unexposed individuals) and the population-attributable fraction (PAF: proportion of disease prevented by eliminating the risk factor from the population). Risch (2001) showed that the observed twin relative risks for cancer are compatible with a broad range of RR and PAF values for individ-

94

PART I: BASIC CONCEPTS

ual gene effects, including extremely high ones. Indeed, he showed that the PAF can range from small values to 100% depending on the disease allele frequency. If susceptibility alleles are common, the reported twin data are consistent, with up to 100% of the cancers being attributable to inherited susceptibility. Of course, PAF values are not additive, and this would not exclude environmental factors as also important. Rather, it indicates that genetic predisposition is a necessary but not sufficient cause and depends on environmental interactions.

IDENTIFYING CANCER-SUSCEPTIBILITY GENES The preceding section has reviewed evidence for inherited genetic variation in cancer susceptibility based on the familial recurrence risk ratio as estimated from family and twin data. In this section we review methods to identify the genes responsible for this variation. It is now possible to localize and sequence an inherited cancer-susceptibility gene before its function is known. Such gene identification, known as positional cloning, is accomplished by scanning the chromosomes for regions containing “smoking gun” evidence for involvement in the carcinogenic process. An alternative approach to positional cloning for gene identification is evaluation of candidate genes (i.e., those encoding proteins whose functions are thought to be involved in the carcinogenic process). Investigations of candidate genes are based on a priori hypotheses that polymorphisms in a gene or family of genes play a role in the mechanisms generating the cancer of interest. By definition, the function of a candidate gene is at least partially understood. Examples of candidate genes include those involved in estrogen metabolism in relation to breast cancer; the synthesis of, metabolism of, and response to androgens in relation to prostate cancer; and the metabolism of heterocyclic amines from well cooked meat in relation to colon cancer. A resource for planning investigations of candidate genes is the Human Genome Epidemiology Network, or HuGE Net, offered by the Centers for Disease Control and Prevention (CDC) (Khoury and Dorman, 1998). HuGE Net provides online publishing of systematic reviews of genetic variants and their associations with the risk of specific diseases. Each HuGE Net review provides comprehensive background information on molecular/genetic techniques for allelic typing, the prevalence of allelic variants in various racial/ethnic populations, population-based disease risk information, evidence for gene–gene effects or gene–environment interaction, and implications for clinical practice and genetic counseling. All of the HuGE Net reviews are peer-reviewed and published in peer-reviewed journals, and each online review is updated as new information accumulates. The website address for this resource is http://www.cdc.gov/genetics/huge.htm. Studies to identify cancer-susceptibility genes through positional cloning or evaluation of candidate genes fall into several categories (Table 7–3): (1) linkage studies; (2) association studies using either unrelated or related individuals with and without the cancer of interest; (3) comparative genomic hybridization (CGH) studies evaluating chromosomal gain or loss in cancer cells. The choice of study design involves several factors, including the anticipated frequency and penetrance of alleles at the susceptibility locus, the phenotype under study (e.g., the cancer of interest or an intermediate biomarker), and the type and availability of biospecimens.

Linkage Studies One of the most fruitful approaches to searching for predisposition genes has been genetic linkage in families with multiple cases of the disease. This approach was successful in mapping and subsequently identifying, for example, the breast/ovarian cancer genes BRCA1 (Hall et al., 1990; Miki et al., 1994) and BRCA2 (Wooster et al., 1994); the colorectal cancer genes APC (Groden et al., 1991; Nishisho et al., 1991), MSH2 (Fishel et al., 1993; Leach et al., 1993), and MLH1 (Bronner et al., 1994; Papadopoulos et al., 1994); and the melanoma gene CDKN2A (Hussussian et al., 1994). It is an attractive approach

because it requires no prior knowledge of the type or location of the genes. However, because a linkage study is costly, is labor-intensive, and requires large numbers of multiple-case families, it typically is undertaken only with firm evidence that predisposition to the cancer of interest can be inherited. Such evidence may arise from the considerations discussed in the previous section or through segregation analyses, which we discuss in the following section. A linkage study exploits the availability of a large set of DNA polymorphisms called markers, each having a known location on a chromosome. A marker might be a tandemly repeated sequence of bases that varies in length from one chromosome to another in a detectable way (i.e., a “short tandem repeat” or “microsatellite” marker) or a change in a single nucleotide (single nucleotide polymorphism, or SNP). Testing a large number of markers for linkage in disease families is often referred to as a “genome scan.” The chromosomes of affected and unaffected family members are genotyped and labeled with respect to their observed alleles at each of the markers. Linkage studies work because markers having some alleles shared by the affected family members and other alleles shared by the unaffected members are likely to be proximal to a disease-susceptibility gene. When such disease-specific sharing of the marker alleles is unlikely to be due to chance, the chromosomal region around the marker(s) is searched in an attempt to identify a critical subregion containing the gene and indeed the gene itself. Parametric linkage analysis requires statistical models for both the cancer risks associated with alleles of the putative cancer-susceptibility gene and the distribution of family marker genotypes. The genetic model for the markers is typically well known. For the disease, however, parameters are often based on the results of previous segregation analyses that attempted to determine whether two copies (recessive model) or only one copy (dominant model) of the mutated variant allele confers increased risk and/or the age-specific risk associated with each disease locus genotype. These models also specify the locations of the markers, with an unknown parameter representing the location of the putative cancer-susceptibility gene. This parameter, which is of primary interest, is estimated from data on observed cancer statuses and marker genotypes in families by the method of maximum likelihood. In contrast, nonparametric, or “model-free,’’ linkage analysis, also called allele-sharing analysis, does not require specification of risks associated with genotypes at a putative disease locus. Instead, this analysis uses the family cancer data and marker genotypes to evaluate departures from the null hypothesis that the marker genotypes and disease status within a family are independently inherited. Whether parametric or nonparametric, linkage analysis is concerned not with association between disease phenotype and a particular marker allele but with patterns of marker allele sharing among affected and unaffected family members. Suppose, for example, families are typed at a marker (with alleles A and B) that is close to a disease gene. Then allele A may segregate with the cancer in some of the families, and allele B does so in others; the clues come from patterns of allelesharing among family members with and without the cancer, rather than an association of specific alleles with the cancer across families. In most cases of dominant diseases, especially those undergoing negative selection, there is significant mutational heterogeneity, meaning that most families carry different mutations (although they may be in the same gene). By chance, these mutations fall on chromosomes randomly distributed with respect to alleles A and B at the marker described above. For dominant diseases that are mild and/or of late onset, however, there can be persistence and spread of a single mutation in the population, especially in a population isolate. In this case, a single allele at a linked (but uninvolved) marker locus may be associated with the disease across families. This phenomenon is referred to as linkage disequilibrium. For example, this is the case for mutations 185delAG at BRCA1, 6174delT at BRCA2, and I1307K at APC in the Ashkenazi Jewish population. Linkage disequilibrium is more characteristic of recessive mutations, as they typically are associated with the normal phenotype in heterozygous carriers and therefore can easily spread and persist in a population without incurring negative selection (except in the rare homozygote). The most well known

95

Genetic Concepts and Methods in Epidemiologic Research

Association Studies Whereas linkage studies search for patterns of allele-sharing within families, association studies search for correlation between specific alleles of a polymorphism and cancer risk. Thus, association studies are similar to the classic epidemiologic studies that search for correlation between disease and personal attributes. The simplest association studies are case-control studies, which compare marker genotypes of cancer cases to those of cancer-free controls. Because of the invariance of genotypes with respect to age and disease status, such studies avoid some of the major limitations of case-control studies of other endogenous attributes (Clayton and McKeigue, 2001). However, investigators have worried that genotype–disease odds ratios estimated from such case-control comparisons may be biased by confounding due to ethnic stratification of the population. Suppose, for instance, that the locus under test is not near a cancer-susceptibility locus but that the population consists of two ethnic groups, with the first group having a higher prevalence of both the cancer and a specific marker allele than the second group. A random sample of cases would contain a higher fraction of the first ethnic group than would either the general population or a sample of controls and thus a higher total count of the allele. In this case, the test statistic leads to rejection of the null hypothesis more often than it should. Failure to account for such ethnic stratification could lead to invalid conclusions if the association is interpreted as causal. This potential bias is a classic example of confounding (by ethnic origin), a well known problem in observational epidemiology. The extent to which it actually is a problem in case-control studies of candidate genes is controversial because the confounding factor (ethnic ancestry) can produce substantial bias only if it is both strongly associated with cancer risk and strongly correlated with genotypes at the polymorphism of interest. Moreover, large case-control differences in the prevalence of suspect genotypes at the polymorphism are less likely to be due to such bias than are small differences (Wacholder et al., 2000). Nevertheless, potential confounding by ethnic stratification in casecontrol studies has prompted an interest in family-based designs, such as a comparison of genotypes of affected and unaffected siblings or a comparison of the alleles that parents transmit to affected offspring with those they do not transmit (Spielman et al., 1993; Gauderman et al., 1999). These designs automatically match the two comparison groups of genotypes on ethnic ancestry. For example, much has been written about the transmission disequilibrium test (TDT), which compares the marker alleles transmitted from heterozygous parents to affected offspring with those not transmitted (Spielman et al., 1993; see also Schaid and Rowland, 1998, for a review and commentary). Family-based designs can be less efficient than those based on unrelated controls (Risch and Teng, 1998; Gauderman et al., 1999). An alternative approach to such confounding by ethnicity is a set of statistical methods collectively called “genomic control of confounding.” There are two approaches to such “genomic control.” The first method considers a large collection of random markers and examines the distribution of allele frequency differences between cases and controls for these markers. The tested locus is then evaluated against this empirical “null” distribution (Devlin and Roeder, 1999). The other method uses each subject’s genotypes as a large collection of neutral markers to classify him or her probabilistically into several ancestral racial/ethnic groups (Pritchard et al., 2000; Satten et al., 2001). A classic example was provided by Williams et al. (2000), who studied a cohort of 7796 residents of the Gila Indian Reservation in Arizona for the risk of non-insulin-dependent diabetes mellitus (NIDDM). This cohort was ethnically admixed, descending largely from white Europeans and Pima Indians. To study the relation between NIDDM risk and ethnic admixture, the investigators genotyped cohort members at several markers whose alleles have different frequencies in the two populations. They used each subject’s genotypes at these markers to assign a pair of probabilities (p1, p2), with p1 + p2 = 1. These proba-

bilities correspond to the proportion of white European and Pima Indian ancestry, respectively. Specifically, p1 represents an estimate of the probability that the subject’s DNA at a randomly selected locus was inherited from a white European ancestor. Figure 7–1 is a graph of p1 values for the subjects in the study by Williams et al. (2000). The authors found a strong negative correlation between these values and NIDDM risk, suggesting that the more white European admixture an individual has, the lower is his or her risk. Thus, ethnic admixture appears to be a risk factor for NIDDM, and as such it is a potential confounding factor for the study of other risk factors. To control “genomically” for confounding, each individual’s p1 value would be included in regression models for disease in relation to genotypes of candidate genes and environmental characteristics. Ethnic admixture can often be determined reliably by selfreporting (Risch et al., 2002). However, one also can use an individual’s genotypes at neutral markers to classify him or her probabilistically into several groups without specifying in advance the race or ethnicity of the groups (Rosenberg et al., 2002). However, all of these methods require a large number of markers for accurate ancestry attribution and so are currently costly to implement.

Linkage Versus Association Studies An important question when searching for cancer-susceptibility genes concerns the relative merits of linkage and association studies. Risch and Merikangas (1996) argued that whereas linkage studies have more power for genes with rare disease-susceptibility alleles causing high risks, association studies are preferred for genes that have common variant alleles with lower risks. Their conclusions hold when the disease-causing and marker polymorphisms are identical. In situations where the disease and marker polymorphisms are not identical even in the same gene, association analyses can be less powerful than linkage analyses even for common variant alleles. The power of an association study of a particular polymorphism or set of polymorphisms depends on the extent of linkage disequilibrium between the alleles of the typed polymorphisms and the variant allele at the (typically untyped) disease-causing locus. The alleles of two loci are said to be in linkage disequilibrium (LD) in a given population if their cooccurrence within individuals is not random in the population. Although this may occur for unlinked loci in ethnically stratified populations that are not randomly mating, LD is more typical of closely linked markers because it decays at a rate of (1 - l) per generation, where l is the recombination fraction between the loci. LD is a characteristic of a population, and thus patterns of LD may differ across ancestrally differentiated groups (e.g., the major racial groups). The extent of disequilibrium between disease-causing and marker loci is a

6000 5500 5000 500

Frequency

example of this type is the cystic fibrosis F508 mutation, which is common and widespread among Caucasians.

400

300

200

100

0 0.00

0.25

0.50

0.75

1.00

Individual Admixture

Figure 7–1. Histogram showing distribution of European admixture indices p1 for 7796 residents of the Gila River Indian Reservation. (Source: Williams et al., 2000.)

96

PART I: BASIC CONCEPTS

critical determinant of the power of association studies (MullerMyhsok and Abel, 1997; Risch and Teng, 1998). When there is little disequilibrium between alleles at marker and disease loci, association studies have poor power compared to linkage studies (Tu and Whittemore, 1999). The choice of study design is thus complex because the extent of disequilibrium between two loci is not a simple function of physical distance—it can vary greatly over the genome. The choice of polymorphism to type in an LD study is also a critical issue. Currently, there are two approaches to this question. One advocates a judicious random distribution of SNPs based on observed patterns of LD in a collection of several racial groups (Collins et al., 1997). An alternative approach is to focus on SNPs found in functional regions of genes (e.g., the coding and promoter regions as well as intron/exon boundaries) (Risch, 2000; Botstein and Risch, 2003). The underlying assumptions and logic of these two approaches have been described (Botstein and Risch, 2003).

Array-Based CGH and Gene Expression Studies In contrast to linkage and association studies that evaluate genetic characteristics in individuals with and without cancer, studies of chromosomal gain and loss identify cancer-susceptibility genes by comparing genetic characteristics of cancer cells to those of normal cells from the same tissue. Allelic loss studies identify genes whose proteins prevent a cell from becoming malignant. These genes, the tumor suppressor genes, operate either by repairing damage to a cell’s DNA or, if repair is impossible, by initiating apoptosis. For example, a cell that has lost both functioning alleles of the tumor suppressor gene p53 (by mutation or deletion) is at high risk of malignant transformation. Therefore, evaluating allelic loss in cancer cells can point to new cancer-susceptibility genes. If the normal cells in the tissue contain two alleles at markers in the region, whereas the cancer cells have lost one or both alleles, the markers may be near a tumor suppressor gene whose loss is responsible for the cancer. A limitation of these allelic loss studies is that the loss in the cancer cells may merely be a consequence, rather than a cause, of the malignancy. Like allelic loss studies, differential gene expression studies compare cancer cells to normal cells. However the objective is not only to identify tumor suppressor genes but to identify oncogenes whose inappropriate expression is involved in malignant transformation. Genes detectable in the messenger RNA of cancer cells but absent in that of normal cells in the same tissue are good candidates for oncogenes. Array-based CGH is but one of many techniques recently developed to evaluate gain or loss of genetic material in cancer cells compared to normal ones. The Cancer Genome Anatomy Project (CGAP) of the U.S. National Cancer Institute (http://cgap.nci.nih.gov) aims to provide a public database of differential gene expression in cancer cells and normal cells. Further discussion of these methods can be found elsewhere (Pollack et al., 1999; Hedenfalk et al., 2002).

Barriers to Gene Identification Some issues in gene identification apply to both genome scans and evaluation of candidate genes. For example, there is a need to accommodate the late and variable age at onset of most cancers. Clearly, a man who dies in his forties should not be considered unaffected with respect to cancers of late onset, such as prostate cancer. A related issue concerns the need to adjust analyses for personal covariates (e.g., genotypes at other known predisposing genes) or lifestyle characteristics (e.g., diet, tobacco consumption). The choice of covariates included in regression models of genetic effects has implications for data collection procedures. The effort to identify cancer-susceptibility genes has been plagued by failure to replicate initial positive results. For example, Smith et al. (1996), studying 91 multiple-case families, found strong evidence of linkage of prostate cancer to markers on chromosome 1. However, attempts to replicate this result in other data have yielded mixed results, and a formal combined analysis of 772 families from nine groups obtained only modest evidence of linkage (Xu et al., 2000). Such failure may reflect the complexity of the underlying genetic

model. One type of genetic complexity is locus heterogeneity, whereby mutations in different genes can all cause increased risk of the cancer (as in the case of BRCA1 and BRCA2 for breast cancer). In this situation, different families are expected to show linkage to different loci, greatly reducing the power of the overall analysis. Such a model could explain the differing results among studies if mutations in different genes predominate in the families collected by different groups. Such effects have been observed, for example, with respect to hereditary nonpolyposis colon cancer (HNPCC) families in Finland due to a limited number of mutations that are restricted to that population. For association studies, inconsistent results may reflect one or more of several common study design flaws. The first investigations of a putative susceptibility gene may have been based on poorly defined, or “convenience,” controls, or they may have used allelic frequencies for controls that are based on the published literature but are inappropriate for the population studied. As a result, findings from early studies are variable, and resolution requires more definitive studies using appropriate population controls. A related concern is the failure to collect data or to adjust for differences in the racial/ethnic distributions of the case and control groups, which could lead to confounding by ethnic stratification, as discussed in the previous section. Some studies have used prevalent cases, which cannot discriminate between genetic variants that are associated with better survival (overrepresented among prevalent cases) and variants that alter the risk of cancer development. Characteristics of the polymorphisms in the gene of interest may also affect study feasibility. The population frequencies of the allelic variants may be extremely low, making it difficult to obtain a sample size large enough to obtain precise estimates of allele frequencies, estimate main effects, and estimate gene–gene and gene–environment interactions. Careful consideration must also be given to whether the allelic variants lie in the intron (noncoding) or exon (coding) regions of the gene. If a polymorphism lies in an exon (or a close-in promoter region), it is plausible that allelic variants could influence gene function and/or expression, and studies addressing the known functional effects of such variants should be consulted. If the allelic variants lie in an intron, they are less likely to have a direct functional effect but may be in linkage disequilibrium with the true disease-susceptibility locus. Finally, to avoid inflated type I error probabilities, linkage and association scans must accommodate the multiple testing involved when evaluating the large number of markers typically analyzed.

CHARACTERIZING CANCER-SUSCEPTIBILITY GENES Successful gene identification, whether by positional cloning or by implication of a candidate gene, introduces several new areas of inquiry. For example, we need to know the function of the protein encoded by the newly identified gene and how polymorphisms in the gene lead to changes in its protein that alter cancer risk. Answers to these questions may suggest gene–environment or gene–gene interactions and motivate preventive strategies. A question of great clinical importance concerns the age-specific risks associated with the various genotypes. Estimates of the age-specific and lifetime cancer risks in carriers of specific variant alleles are essential for informed clinical management of those with inherited susceptibility. These risks are collectively called the penetrance of the variant allele(s). A question of importance for public health planning and resource allocation concerns the population frequencies of genotypes associated with increased risk. This information is needed to estimate the fraction of the cancer burden that is attributable to the gene and to determine sample size requirements for future studies. For variant alleles with low frequencies (i.e., less than 10%), precise frequency estimates require large population samples or samples containing an overrepresentation of individuals who are likely to carry the variants. Furthermore, allele frequencies often vary according to racial or ethnic ancestry, making it important to collect detailed ancestry information from study subjects so race-specific allele frequencies can be estimated.

Genetic Concepts and Methods in Epidemiologic Research Genes with dominant expression pose a particular challenge. The reason is mutational heterogeneity, as described above. If different families carry distinct mutations, it is typically extremely difficult, time-consuming, and expensive to sequence the gene in each family to find its specific mutation. For the same reason, population surveys to obtain the overall mutational burden (the sum of all mutations) are also complicated. An additional challenge is the decision about whether a particular identified change does, in fact, cause disease. This is particularly the case for amino acid substitutions, as they can either seriously disable normal protein function or be totally benign. For these reasons, most reliable estimates of disease allele frequencies and penetrances in this case have been obtained in genetic isolates such as for BRCA1 and BRCA2 mutations in the Ashkenazi Jewish population. A related issue is estimation of that fraction of the cancer burden due to variant genotypes. When several genes have been identified, we need to know how these genes interact to affect disease risk. Perhaps the most important issue, from a preventive perspective, concerns how lifestyle characteristics modify risk in carriers of high risk genotypes. Specifically, what can such carriers do to reduce their risk? Why do some carriers remain disease-free well into old age? Studies aimed at investigating these areas of inquiry, which we call gene characterization studies, present challenging issues for genetic epidemiologists and will continue to do so well into the twenty-first century. Many of the approaches used to address these questions are based on a statistical method called segregation analysis. As noted in the previous section, segregation analysis also is used to motivate and guide linkage studies for gene identification. Accordingly, we begin with a brief description of segregation analysis and how it is used in gene characterization.

Segregation Analysis A statistical technique, segregation analysis, evaluates patterns of disease transmission in each of a set of pedigrees. It proceeds by fitting to the observed disease data statistical models of variant-allele transmission. These models have varying degrees of generality. Parameters in the models [e.g., the population frequency of the variant allele(s), the disease risks associated with the variant allele(s)] are estimated by the method of maximum likelihood. The models are tested for goodness of fit to the data. Included in the arsenal of candidate models is a “nongenetic model” that attributes the patterns of disease occurrence in families merely to chance or to shared environmental factors. Rejection of the nongenetic model in favor of a recessive model or a dominant model can guide subsequent linkage analyses. For example, segregation analyses of breast cancer have been reasonably consistent in providing evidence of a major dominant locus with particularly elevated penetrance at younger ages (Williams and Anderson, 1984; Bishop et al., 1988; Newman et al., 1988; Claus et al., 1991). This model has subsequently been confirmed by identification of the BRCA1 and BRCA2 loci. A segregation analysis of prostate cancer conducted by Carter et al. (1992) rejected a nongenetic model in favor of a dominant model that includes a high risk allele with a population frequency 0.003. This model has formed the basis of virtually all parametric linkage analyses of prostate cancer, although subsequent segregation analyses have supported a recessive model (Cui et al., 2001). After a cancer-susceptibility gene and its variant alleles have been identified, segregation analysis can be a useful tool for characterizing the gene. When used in this setting, the approach (sometimes called modified segregation analysis) is applied to multiple-case families containing members who have been genotyped for alleles of the known gene. At this point, the mode of inheritance of the known gene (dominant or recessive) usually is understood. Interest now focuses on estimating the age-specific risks associated with genotypes of the known gene and on evaluating the evidence for mendelian inheritance of additional cancer-susceptibility genes. Whatever its goals, such an analysis must accommodate the way families were ascertained for study. Ideally, families are recruited as the relatives of a well defined, population-based sample of individuals, called probands, with and without the cancer of interest. For rare mutations (e.g., those of BRCA1 or BRCA2), however, adequate power and precision may require

97

recruiting families that are enriched with individuals who carry the high risk genotypes. In this case, the analysis also must accommodate this special selection of families. Even when the selection criteria have been accounted for in the statistical analysis, the risk estimates obtained from the ascertained families may not be appropriate for the general population from which the families were selected. For example, the multiple-case families recruited for gene identification studies have been selected for their high cancer occurrence. They are apt to segregate deleterious alleles of unmeasured risk-modifying genes or to share unmeasured lifestyle characteristics that increase their genetic susceptibility. Therefore, their cancer experience may overestimate risk in the general population. In addition, because risk is likely to be influenced not only by the gene of interest but by environmental factors and subsequent interventions, it may be necessary to gather extensive data on these nongenetic risk factors. Further discussion of issues in the use of classic epidemiologic designs and segregation analysis for penetrance estimation can be found elsewhere (Struewing et al., 1997; Wacholder et al., 1998; Langholz et al., 1999; Begg et al., 2002; Antoniou et al., 2003; Gong and Whittemore, 2003).

CONCLUSIONS Progress in genetics during the last two decades has advanced our understanding of cancer mechanisms. The challenge now is to use this understanding to prevent and control cancer morbidity and mortality in human populations throughout the world. The possibility that modifiable lifestyle characteristics influence gene expression provides hope for the development of preventive strategies. A strategy of considerable appeal would offer specific lifestyle changes and preventive interventions to people at increased cancer risk because of inherited susceptibility. The hope for such a strategy stems in part from the variation in risk among carriers of cancer-predisposing mutations. For example, despite the high risks of cancers of the breast and ovary among BRCA1 and BRCA2 mutation carriers (Antoniou et al., 2003), an estimated 30% of these women reach age 70 years without developing either cancer. We need to know what protects these women in contrast to the carriers who develop these malignancies. Other than chance, possible explanations include variations in the type of mutation, the genotypes at other loci, or potentially modifiable lifestyle characteristics. As more genes with predisposing alleles are identified and as people become increasingly aware of genetic developments and interested in knowing about their own genes, there arises a need to offer them options for preventing the cancers to which they are particularly susceptible by inheritance. References Ahlbom A, Lichtenstein P, Malmstrom H, et al. 1997. Cancer in twins: genetic and nongenetic familial risk factors. J Natl Cancer Inst 89:287–293. Antoniou A, Pharoah PD, Narod S, et al. 2003. Average risks of breast and ovarian cancer associated with BRCA1 or BRCA2 mutations detected in case series unselected for family history: a combined analysis of 22 studies. Am J Hum Genet 72:1117–1130. Begg CB. 2002. On the use of familial aggregation in population-based case probands for calculating penetrance. J Natl Cancer Inst 94:1221–1226. Bishop DT, Cannon-Albright L, McLellan T, et al. 1988. Segregation and linkage analysis of nine Utah breast cancer pedigrees. Genet Epidemiol 5:151–169. Botstein D, Risch N. 2003. Discovering genotypes underlying human phenotypes: past successes for mendelian disease, future approaches for complex disease. Nat Genet 33(Suppl):228–237. Braun MM, Caporaso NE, Page WF, et al. 1995. A cohort study of twins and cancer. Cancer Epidemiol Biomarkers Prev 4:469–473. Bronner CE, Baker SM, Morrison PT, et al. 1994. Mutation in the DNA mismatch repair gene homologue hMLH1 is associated with hereditary nonpolyposis colon cancer. Nature 368:258–261. Carter BS, Beaty TH, Steinberg GD, et al. 1992. Mendelian inheritance of familial prostate cancer. Proc Natl Acad Sci USA 89:3367–3371. Claus EB, Risch N, Thompson WD. 1991. Genetic analysis of breast cancer in the cancer and steroid hormone study. Am J Hum Genet 48:232–242. Clayton D, McKeigue PM. 2001. Epidemiological methods for studying genes and environmental factors in complex diseases. Lancet 358:1356–1360.

98

PART I: BASIC CONCEPTS

Collins FS, Guyer MS, Chakravarti A. 1997. Variations on a theme: cataloging human DNA sequence variation. Science 278:1580–1581. Cui J, Staples MP, Hopper JL, et al. 2001. Segregation analysis of 1,476 population based Australian families affected by prostate cancer. Am J Hum Genet 68:1207–1218. Devlin B, Roeder K. 1999. Genomic control for association studies. Biometrics 55:997–1004. Dong C, Hemminki K. 2001. Modification of cancer risks in offspring by sibling and parental cancers from 2,112,616 nuclear families. Int J Cancer 92:144–150. Fishel R, Lescoe MK, Rao MR, et al. 1993. The human mutator gene homolog MSH2 and its association with hereditary nonpolyposis colon cancer. Cell 75:1027–1038. Gauderman WJ, Witte JS, Thomas DC. 1999. Family-based association studies. J Natl Cancer Inst Monogr 26:31–37. Goldgar DE, Easton DF, Cannon-Albright LA, et al. 1994. Systematic population-based assessment of cancer risk in first-degree relatives of cancer probands. J Natl Cancer Inst 86:1600–1608. Gong G, Whittemore AS. 2003. Optimal designs for estimating penetrance of rare mutations of a disease-susceptibility gene. Genet Epidemiol 24:173–180. Groden J, Thliveris A, Samowitz W, et al. 1991. Identification and characterization of the familial adenomatous polyposis coli gene. Cell 66:589–600. Haiman C, Hunter D. 2002. Genetic epidemiology of cancer. In:Adami H-O, Hunter D, Trichopolous D, editors. Textbook of Cancer Epidemiology. New York: Oxford University Press, pp. 54–72. Hall JM, Lee MK, Newman B, et al. 1990. Linkage of early-onset familial breast cancer to chromosome 17q21. Science 250:1684–1689. Hedenfalk IA, Ringner M, Trent JM et al. 2002. Gene expression in inherited breast cancer. Adv Cancer Res 84:1–34. Hemminki K, Dong C, Vaittinen P. 2001. Cancer risks to spouses and offspring in the family-cancer database. Genet Epidemiol 20:247–57. Hoover R. 2000. Cancer—nature, nurture or both. N Engl J Med 343:135–136. Hussussian CJ, Struewing JP, Goldstein AM, et al. 1994. Germline p16 mutations in familial melanoma. Nat Genet 8:15–21. Khoury MJ, Dorman JS. 1998. The human genome epidemiology network. Am J Epidemiol 148:1–3. Khoury MJ, Beaty TH, Cohen BH. 1993. Fundamentals of Genetic Epidemiology. New York: Oxford University Press. Langholz B, Rothman N, Wacholder S, et al. 1999. Cohort studies for characterizing measured genes. J Natl Cancer Inst Monogr 26:39–42. Leach FS, Nicolaides NC, Papadopoulos N, et al. 1993. Mutations of a mutS homolog in hereditary nonpolyposis colorectal cancer. Cell 75:1215–1225. Lichtenstein P, Holm NV, Verkasalo PK, et al. 2000. Environmental and heritable factors in the causation of cancer. N Engl J Med 343:78–85. Mack TM, Cozen W, Shibata DK, et al. 1995. Concordance for Hodgkin’s disease in identical twins suggesting genetic susceptibility to the youngadult form of the disease. N Engl J Med 332:413–418. Marenberg ME, Risch N, Berkman LF, et al. 1994. Genetic susceptibility to death from coronary heart disease in a study of twins. N Engl J Med 330:1041–1046. Miki Y, Swensen J, Shattuck-Eidens D, Futreal PA, Harshman K, Tavtigian S, Liu Q, Cochran C, Bennett LM, Ding W, et al. 1994. A strong candidate for the breast and ovarian cancer susceptibility gene BRCA1. Science 266:66–71. Muller-Myhsok B, Abel L. 1997. Genetic analysis of complex diseases. Science 275:1328–1329. Newman B, Austin MA, Lee M, et al. 1988. Inheritance of human breast cancer: evidence for autosomal dominant transmission in high-risk families. Proc Natl Acad Sci USA 85:3044–3048. Nishisho I, Nakamura Y, Miyoshi Y, et al. 1991. Mutations of chromosome 5q21 genes in FAP and colorectal cancer patients. Science 253:665–669. Page WF, Braun MM, Partin AW, et al. 1997. Heredity and prostate cancer: a study of World War II veteran twins. Prostate 33:240–245.

Papadopoulos N, Nicolaides NC, Wei YF, et al. 1994. Mutation of a mutL homolog in hereditary colon cancer. Science 263:1559–1560. Pollack JR, Perou CM, Alizadeh AA, et al. 1999. Genome-wide analysis of DNA copy-number changes using cDNA microarrays. Nat Genet 23:41–46. Pritchard JK, Stephens M, Rosenberg NA, et al. 2000. Association mapping in structured populations. Am J Hum Genet 67:170–181. Risch N. 1990. Linkage strategies for genetically complex traits. I. Multi-locus models. Am J Hum Genet 46:222–228. Risch NJ. 2000. Searching for genetic determinants in the new millennium. Nature 405:847–856. Risch N. 2001. The genetic epidemiology of cancer: interpreting family and twin studies and their implications for molecular genetic approaches. Cancer Epidemiol Biomarkers Prev 10:733–741. Risch N, Merikangas K. 1996. The future of genetic studies of complex human diseases. Science 273:1516–1517. Risch N, Teng J. 1998. The relative power of family-based and case-control designs for linkage disequilibrium studies of complex human diseases. I. DNA pooling. Genome Res 8:1273–1288. Risch N, Burchard E, Ziv E, et al. 2002. Categorization of humans in biomedical research: genes, race and disease. Genome Biol 3:comment2007. Rosenberg NA, Pritchard JK, Weber JL, et al. 2002. Genetic structure of human populations. Science 298:2381–2385. Satten GA, Flanders WD, Yang Q. 2001. Accounting for unmeasured population substructure in case-control studies of genetic association using a novel latent-class model. Am J Hum Genet 68:466–477. Schaid DJ, Rowland C. 1998. Use of parents, sibs, and unrelated controls for detection of associations between genetic markers and disease. Am J Hum Genet 63:1492–1506. Smith JR, Freije D, Carpten JD, et al. 1996. Major susceptibility locus for prostate cancer on chromosome 1 suggested by a genome-wide search. Science 274:1371–1374. Sorensen TIA, Nielsen GG, Andersen KA, et al. 1988. Genetic and environmental influences on premature death in adult adoptees. N Engl J Med 318:727–732. Spielman RS, McGinnis RE, Ewens WJ. 1993. Transmission test for linkage disequilibrium: the insulin gene region and insulin-dependent diabetes mellitus (IDDM). Am J Hum Genet 52:506–516. Struewing JP, Hartge P, Wacholder S, et al. 1997. The risk of cancer associated with specific mutations of BRCA1 and BRCA2 among Ashkenazi Jews. N Engl J Med 336:1401–1408. Tu IP, Whittemore AS. 1999. Power of association and linkage tests when the disease alleles are unobserved. Am J Hum Genet 64:641–649. Verkasalo PK, Kaprio J, Koskenvuo M, et al. 1999. Genetic predisposition, environment and cancer incidence: a nationwide twin study in Finland, 1976–1995. Int J Cancer 83:743–749. Wacholder S, Hartge P, Struewing JP, et al. 1998. The kin-cohort study for estimating penetrance. Am J Epidemiol 148:623–630. Wacholder S, Rothman N, Caporaso N. 2000. Population stratification in epidemiologic studies of common genetic variants and cancer: quantification of bias. J Natl Cancer Inst 92:1151–1158. Williams RC, Long JC, Hanson RL, et al. 2000. Individual estimates of European genetic admixture associated with lower body-mass index, plasma glucose, and prevalence of type 2 diabetes in Pima Indians. Am J Hum Genet 66:527–538. Williams WR, Anderson DE. 1984. Genetic epidemiology of breast cancer: segregation analysis of 200 Danish pedigrees. Genet Epidemiol 1: 7–20. Wooster R, Neuhausen SL, Mangion J, et al. 1994. Localization of a breast cancer susceptibility gene, BRCA2, to chromosome 13q12–13. Science 265:2088–2090. Xu J, International Consortium for Prostate Cancer Genetics. 2000. Combined analysis of hereditary prostate cancer linkage to 1q24–25: results from 772 hereditary prostate cancer families from the international consortium for prostate cancer genetics. Am J Hum Genet 66:945–957.

II THE MAGNITUDE OF CANCER

This page intentionally left blank

8

International Patterns of Cancer Incidence and Mortality D. MAXWELL PARKIN AND FREDDIE I. BRAY

E

pidemiology has been defined as the study of the distribution and determinants of health-related states or events in specified populations and the application of this study to the control of health problems (Last, 1988). This definition distinguishes the tasks of describing the distribution of “health-related states” (disease) by discovering the determinants (causes) and the process of acquiring such knowledge by putting it to some practical use. Ever since international data on cancer incidence and mortality have been available, it has been clear that there are large differences in the risk of specific cancers in different populations. Muir (1996) cited the early observations of Hoffman (1915), drawing attention to the 10-fold difference in mortality from cancer of the breast between Japanese and British women. The mere observation is interesting but only inasmuch as it helps when considering their possible causes. The geography of cancer can be useful for defining the relative contributions of “environment” and “genetics” in the etiology of specific cancers. This dichotomy, which was a familiar notion 20–30 years ago, expresses the idea that the inhabitants of different places may differ in their risk of certain cancers owing to a variation in exposure to carcinogens (in the external environment or through lifestyle choices) or in genetic susceptibility to them. In this chapter, the patterns of cancer are considered mainly with respect to what they tell us of differences in the risk of cancer among populations and the possible explanations for them. Information on the burden of disease has a different, albeit related, use during the planning and monitoring of programs of cancer control (WHO, 2002). The main role of cancer surveillance in this context lies in the assessment of the current magnitude of the cancer burden and its likely future evolution as well as monitoring the effects of early detection/screening, treatment, and palliative care. Some of the indices used to measure the risk of disease may also be of value with respect to evaluating the burden, although there are several indices specifically designed for the latter set of tasks.

MEASURING RISK OR BURDEN OF CANCER IN THE POPULATION Incidence is the number of new cases occurring. It can be expressed as an absolute number of cases or in relation to the size of the population at risk; it is also expressed as the time during which the cases occur, which is the incidence rate. Incidence requires definition of the moment at which the cancer “begins”—when an individual becomes a new “case.” Clearly, this is a somewhat arbitrary point in time for a biologic process that is a continuous spectrum between a set of mutations in critical genes of one cell and the death of the organism. By convention, incidence is counted from the time the cancer was diagnosed. However, the process may be difficult, arbitrary, and subject to differences between populations for reasons other than the true “risk” of disease. There are four areas where defining “incidence” may be problematic.

• Deciding when a tumor has become invasive and is therefore a “cancer.” This is a particular problem for cancers of the bladder, which often arise as invasive foci in papillomas, showing varying degrees of malignant change in the epithelium. A biopsy may reveal malignant cells, but at the spot where the biopsy specimen was obtained there was no invasion through the basement membrane of

the epithelium. Whether such tumors are included in “cancer” statistics can greatly influence the incidence rate (Saxen, 1982; Kiemeney et al., 1994). • Deciding whether two or more cancers in the same individual represent a new (incident) case, or an extension, recurrence, or metastasis of a first tumor. There is an international convention for making this decision for the purpose of comparing incidence rates between populations (Fritz et al., 2000), but the rules do not make any biologic sense, such as when studying similarities in etiology or genetic susceptibility to a first and second cancer in the same individual or the possibly carcinogenic results of therapy. • Asymptomatic cancers may be diagnosed during investigation for other conditions, during surgery, or at autopsy. The latter was a notable contributor to the apparent incidence of several cancers that can exist for long periods in latent form (e.g., thyroid, prostate) in populations in which autopsy was a common procedure (Saxen, 1982). • Screening programs advance the time of diagnosis by detecting cancer at an early, asymptomatic stage. When such programs are introduced, it results in a temporary increase in apparent incidence. In theory, as time goes by this should be compensated by the nonappearance of the same tumors at a later date, so the cumulative incidence in the population remains unchanged. However, for most if not all cancers, it is now accepted that screening brings to light cancers that, had they not been so detected, would never have been diagnosed during the subject’s lifetime—so-called overdiagnosis (Parkin and Moss, 2000). It is a particular problem for cancer of the prostate, where screening detects many latent cancers in elderly men who would have remained unaware of their presence for the rest of their lives had they not accepted the proposition of being screened (Legler et al., 1998; Hankey et al., 1999). In addition to these problems is the potential inaccuracy resulting from inadequacies of the data collection process itself and the classification and coding of cancers. The latter issue is not unique to incidence data and is certainly a much greater problem with respect to mortality statistics (see below). Precision of diagnosis has improved over time with the advent of noninvasive scanning methods, so the number of cases correctly allocated to certain sites (especially the central nervous system and internal organs) has increased over time at the expense of ill-specified cancers (or even noncancer diagnoses) (Modan et al., 1992). Mortality is the number of deaths occurring and the mortality rate the number of deaths per 100,000 persons per year. It is the product of the incidence and the fatality of a given cancer. Fatality, the complement of % survival, is the proportion of cancer patients who die. Mortality rates are the most useful measure of the impact, or burden, of cancer in a population. Mortality rates are probably equally used as a convenient proxy measure of the risk of acquiring the disease (incidence) when comparing groups, as they may be more generally available (as described below). However, when used in this way, an assumption of equal survival/fatality in the populations being compared is introduced. This may be reasonable for some cancers with a poor prognosis. However, for cancers for which early diagnosis and/or therapy can markedly influence survival between countries, population subgroups, or over time, mortality rates do not provide a good reflection of differences in risk (Fig. 8–1).

101

102

PART II: THE MAGNITUDE OF CANCER

Figure 8–1. Age standardized rates of incidence and mortality for two cancers of varying prognosis.

The survival time of a cancer patient is defined as the time that elapsed between diagnosis and death. Computation of survival depends on follow-up of diagnosed cancer patients and calculating the number surviving after various intervals of time. The usual method is the actuarial, or life-table, method; and there are various ways to allow for “normal,” or noncancer, mortality among the followed-up patients. The most familiar is relative survival, which computes the observed mortality rate in cancer patients as a ratio of that expected in the population from which they come (Ederer et al., 1961).

considered “cured” (before 5 years) (Pisani et al., 2002). Patients alive 5 years after diagnosis are usually considered cured because, for most cancers, the death rates for such patients are similar to those for the general population. There are exceptions, especially female breast cancer, for which the risk of death remains higher than that for the general population for many more years. The concept of person-years of life lost was introduced more than 50 years ago (Dempsey, 1947) to refine traditional mortality rates by providing a weighting for deaths occurring at different ages. These methods started to become more widely used during the late 1970s in health-services planning (Murray and Axtell, 1974). There are many variations in the calculations used, depending on the weights (the value of years of life at different ages), the “normal” lifespan against which to compare premature death (fixed upper limit, or life-table expectations of life), and the “discount rate” to apply to life-years that would have been lived in the future. Discounting gives decreasing weights to the life-years saved over time, admitting that life-years in the future are valued less highly than at present (Das Gupta et al., 1972; Layard and Glaister, 1972). Letting the value of life vary at different ages has been particularly appealing to economists, who are, implicitly at least, interested in the economic productivity of individuals, which of course varies by age. This approach has been taken a step further with the development of indices such as quality adjusted life-years (lost) (QALYs) or disability-adjusted life-years (lost) (DALYs) (Murray, 1994; Morrow and Bryant, 1995). Essentially, the measures admit that between the onset of a disease and death or recovery there is a spectrum of morbidity that can be quantified in terms of its duration and severity. Three elements are needed to calculate these indices: the incidence of the disease; its mean duration (or, equivalently, survival probability); and a measure of life “quality” between the onset and the end of disease. The problem encountered when using these indices lies in ascribing values to quality of life or level of disability, as both are subjective and vary with time since diagnosis and in different cultural and socioeconomic settings. Nevertheless, the estimation of DALYs for various conditions worldwide has been widely used by the World Health Organization (WHO) as a means of establishing priorities for health-care programs (WHO, 2000).

OTHER MEASURES OF CANCER BURDEN RATES OF DISEASE Several other indicators have been used to quantify the burden of cancer for the purpose of setting priorities for resource allocation within health services or within cancer control programs. They are not of relevance when investigating risk or cause. Prevalence is the proportion of a population that has the disease at a given point in time (Rothman and Greenland, 1998). For many diseases (e.g., hypertension, diabetes), prevalence usefully describes the number of individuals requiring care. Although prevalence has also been advanced as a useful measure of cancer burden (Hakama et al., 1975), many persons diagnosed in the past have been “cured”; that is, they no longer have an excess risk of death (although some residual disability may be present) such as following a resective operation. Estimates of cancer prevalence may consider all persons ever diagnosed (lifetime prevalence), although it is not clear that such statistics have much utility. They can be derived from cancer registries, which have long-term registration of cases and complete follow-up for vital status over many years (Feldman et al., 1986; Tulinius et al., 1992). Population surveys are another approach, although they underestimate true prevalence (Hewitt et al., 1999). In the absence of complete data, an estimate can be prepared using models that incorporate long-term series of incidence and survival (Capocaccia et al., 1990; Merrill et al., 2000). Other workers have attempted to define the proportion and timing of “cure” for various cancers, so only patients not cured are considered prevalent (Coldman et al., 1992). The data needed for such calculations are rarely available, however, and for international comparisons a simpler approach is generally necessary: estimation of the number of cases diagnosed within 1, 3, and 5 years (partial prevalence) to indicate the number of persons undergoing initial treatment (cases within 1 year of diagnosis), clinical follow-up (within 3 years), or not

Whereas numbers of new cases or deaths are essential for planning and prioritizing resources for cancer control, quantification and comparison of risk requires the computation of rates of cancer. The term rate is often used interchangeably with the risk of developing a cancer; but strictly speaking, risk is a proportion and describes the accumulation of the effect of rates over a given period of time. Ideally, we would estimate a rate by ascertaining, for every individual in the population, the risk of being diagnosed with cancer at a given age and specific point in time. This instantaneous rate requires that the designated period of time is infinitely small, approaching zero. As cancer is a relatively rare disease, however, we must estimate the average rate of occurrence of new cases of cancer in a sufficiently large population over a sufficiently long time period. In this formulation, the denominator is the underlying person-time at risk from which the cancer cases in the numerator arose.

POPULATION AT RISK In prospective cohort studies, the follow-up of individuals in a dynamic population is undertaken from their time of entry into the study until cancer diagnosis, loss to follow-up, or completion of follow-up. A summation of the varying lengths of individual followup accurately represents the person-time at risk of being diagnosed with cancer. However, information at the individual level is not normally available from vital registration systems; so, instead, personyears at risk are approximated using cross-sectional population data collated by national statistical organizations. The estimation of the

International Patterns of Cancer Incidence and Mortality denominator, a summation of the mid-year estimates for each of the years under consideration, thus depends on both the availability and the completeness of demographic information on the population under study. In most developing and developed countries, 10-year population censuses provide basic population estimates by age, sex, and census year; and statistics offices often produce estimates for intercensus years based on rates of birth, death, and migration. The approximation assumes that there is stability in the underlying population, as individuals traverse the age–time plane represented by the well known Lexis diagram. Given a steady state of demographic gains and losses, where the number of individuals during a designated period entering an age group equals the number who leave it, the method can be considered to provide adequate estimates of the person-time at risk in most circumstances and is not, for most cancers, unduly biased by the fact that the numerator is a subset of the denominator, given the rarity of cancer as a condition.

CRUDE, STRATUM-SPECIFIC, AND STANDARDIZED RATES Crude and Age-Specific Rates The term crude refers to a rate based on the frequency of cancer in the entire population, ignoring demographic subdivisions such as age (although rates are usually given separately for males and females because of the different disease patterns by sex). The measure can be useful for summarizing the extent of the cancer burden, but its utility for comparing risk is limited given the vastly different demographic structures in populations worldwide, both geographically (between populations) and within a given population over time. There are thus compelling grounds for adjusting for the effects of age when comparing cancer risk in populations, given that age is also a strong determinant of cancer risk. To obtain a more accurate picture of the true risk of cancer, rates are calculated for each age strata, usually grouped in 5-year intervals. Age- and sex-specific incidence and mortality rates are the foundation of an epidemiologic analysis of cancer frequency data.

Age Standardization To facilitate geographic and temporal comparisons among populations, a summary rate is required that absorbs each population’s schedule of age-specific rates. There are two commonly used techniques for age standardization. The direct method involves applying the agespecific rates in the observed population to the age-specific population counts (or weights) of a fixed reference population to obtain the age-standardized rate (ASR). The choice of standard population is a somewhat arbitrary one, but there are two widely used standards for international comparisons: the world standard of Doll (Doll and Cook, 1967) after Segi (1960), as used consecutively in the eight Cancer Incidence in Five Continents (CI5) volumes (Parkin et al., 2002), and the European standard (Doll and Cook, 1967), used for cancer incidence and mortality comparisons in Europe (Bray et al., 2002a). The alternative indirect method involves calculating the ratio of the total number of cancer events observed to the number of cases that would be expected if the age-specific rates of a designated reference population had applied. Standardized incidence and mortality ratios (SIR and SMR, respectively) are particularly useful in situations (rare cancers, small populations) where the age-specific rates in the population under study are associated with substantial random error. Although comparisons using either measure give similar results, the ASR is preferred to the SIR/SMR on the basis of statistical reasoning related to minimizing bias (Breslow and Day, 1987) and is used in this chapter to illustrate the international variations in cancer. As was mentioned earlier, selection of the standard is somewhat arbitrary, and various standards have been proposed usually because they resemble some “real” population, rather than an artificial one. The WHO, for example, introduces new “world standard” populations every so often (WHO, 1993; Ahmad, 2000), but they have no partic-

103

ular advantage when evaluating differences in risk between populations or over time (Bray et al., 2002b). An alternative method of direct standardization is cumulative risk— the probability that an individual will develop the cancer under study during a certain age span in the absence of other competing causes of death. The age ranges 0–64 and 0–74 years are commonly used to represent the lifetime risk of developing the disease. In other circumstances, different age ranges may be more appropriate (e.g., for childhood cancer). If the cumulative risk is less than 10%, as is the case for most tumors, it can be approximated well by the cumulative rate, the summation of the age-specific rates over each age strata from birth to a defined upper age limit. Mathematically, the cumulative risk is equal to 1 - exp(- cumulative rate). The cumulative rate is not, in fact, a rate but a dimensionless quantity. In other words, it is not expressed in units per annum but simply as a number. It is most conveniently expressed as a percentage. As well as avoiding the arbitrary choice of a standard population to some extent, it is an appealing summary measure, giving an idea of the lifetime risk of the disease.

INFORMATION SOURCES Incidence data are available from cancer registries. Disease registers are part of surveillance systems for various diseases, but they have been more important and successful for cancer than for any other condition. This is because of the serious nature of most cancers, which means that, except in a few societies without access to medical care and concepts, those affected almost always present for diagnosis (and treatment, if available). As a result, enumeration of incident cases of cancer is relatively easy compared to other diseases. It is this fact that has permitted the development and use of cancer registries, particularly population-based registries (PBCRs), which collect data on every person with cancer in a defined population (Jensen et al., 1991). Usually, the population is resident in a geographic region. The registry must therefore be able to distinguish between residents and nonresidents and should have sufficient information on each case to avoid multiple registrations. The cooperation of the medical profession and health care services is vital to the success of cancer registration. Because of the known population at risk, the PBCR allows calculation of incidence rates for different cancers, which may be defined by the traditional codes of the International Statistical Classification of Diseases and Health-Related Problems, 10th Revision (ICD-10) but may also be classified by histologic type or stage of disease. Incidence rates derived from cancer registries are considerably more restricted in availability than mortality. The establishment of cancer registration worldwide has been a haphazard process; in some countries there has been a more-or-less official policy to support and fund registries; elsewhere, individual initiative of research-orientated clinicians and pathologists has often been a major factor. Cancer registries may cover national populations or, more often, certain regions. In developing countries, in particular, coverage is often confined to the capital city and its environs. Incidence data from cancer registries worldwide are published at 5year intervals in the CI5 series. The data included in this series are considered to have met criteria of completeness and validity that allow their use in international comparative studies of incidence. The number of registries contributing data has increased progressively since the first volume of the series was published in 1962 (Table 8–1). The latest volume (the eighth) contains comparable incidence information from 186 registries in 57 countries, mainly over the period 1993–1997 (Parkin et al., 2002). In addition to CI5, individual cancer registries publish their own results in annual reports, compendia, or peer-reviewed articles. Some of the more recent data may be of a quality meriting their inclusion in future volumes of CI5. Statistics on cancer mortality derive from the information on death certificates collected by civil registration systems recording vital events (births, marriages, deaths). The responsible authority varies among countries, but usually the first level of data collection and processing is the municipality or province, with collation of national

104

PART II: THE MAGNITUDE OF CANCER

Table 8–1. Coverage of the Eight Volumes of Cancer Incidence in Five Continents Volume I II III IV V VI VII VIII

Year of Publication

Registries

Populations

Countries

Period (approx.)

1966 1970 1976 1982 1987 1992 1997 2002

32 47 61 79 105 138 150 186

35 58 79 103 137 166 183 214

29 24 29 32 36 49 50 57

1960–62 1963–67 1968–72 1973–77 1978–82 1983–87 1988–92 1993–97

statistics the responsibility of the Ministry of Health or the Interior Ministry. Death certificates record information on the person who died and the cause of death, as certified usually by a medical practitioner. The International Classification of Diseases (ICD) provides a uniform system of nomenclature and coding and a recommended format for the death certificate. Mortality statistics are produced according to the underlying cause of death; this may not equate with the presence of a particular tumor. Although the ICD contains a set of rules and guidelines that allow an underlying cause to be selected in a uniform manner, interpretation of the concept probably varies considerably; such as when death is due to pneumonia in a person previously diagnosed as having cancer. Comprehensive mortality statistics thus require that diagnostic data be available on the deceased and be transferred in a logical, standardized fashion to death certificates, which are then accurately and consistently coded, compiled, and analyzed. There have been a host of studies of the validity of cause-of-death statements in vital statistics data. They have compared cause of death entered on the death certificate with a reference diagnosis derived from autopsy reports (e.g., Heasman and Lipworth, 1966), detailed clinical records (Puffer and Wynne Griffith, 1967), or cancer registry data (Percy et al., 1981). Such studies reveal that the degree of accuracy of the stated cause of death declines as the degree of precision in the diagnosis increases. Thus, although the total number of deaths from cancer of all types may be only slightly underestimated, the distribution by cancer site may be incorrect. There is a tendency to overrecord nonspecific diagnoses instead of the correct location (e.g., large intestine instead of rectum), and accuracy is sometimes lower for those dying at older ages or at home. There are also marked differences among countries regarding the allocation of ICD codes to death certificate diagnoses (Percy and Dolman, 1978; Percy and Muir, 1989). Despite these problems, mortality data remain the most valuable source of information on cancer burden and a useful proxy for risk of disease in many circumstances. A major advantage is their widespread availability. About 30% of the world population is covered by national vital registration systems producing mortality statistics on cancer, including all of the developed countries and many of the developing countries. National-level statistics are collated and made available by the WHO (http://www-depdb.iarc.fr/who/menu.htm). Nevertheless, some knowledge of the likely accuracy of the data available is a prerequisite to their intelligent use. Thus, the fact of publication by national and international authorities is not a guarantee of data quality. For some countries or time periods, coverage of the population is manifestly incomplete, and the so-called mortality rates produced are implausibly low. In others, the quality of the cause of death information is poor. This can sometimes be predicted when a substantial proportion of certificates are completed by nonmedical practitioners [WHO published a useful table in World Health Statistics Annual (1996) giving (for a few countries at least) the relevant percentage]. Otherwise, quality of data must be judged from indicators such as the proportion of deaths coded to “senility and ill-defined conditions” and the proportion of cancer deaths without specification of the primary site or when the site is specified it is done so in only vague terms (ICD9 codes 195–199).

Survival statistics are also produced by cancer registries, and population-based figures have been published from many developed countries, such as the SEER program covering 10% of the U.S. population (Ries et al., 2001), and the EUROCARE II project, which included 17 European countries (Berrino et al., 1999). Survival data from populations in China, the Philippines, Thailand, India, and Cuba have been published by Sankaranarayanan et al. (1998).

ESTIMATION Because national cancer incidence and mortality data are available for only a small number of countries of the world, estimation procedures are required to obtain a comprehensive global picture of the cancer profile and its evolution over time. Estimation may be approached in various ways. By preparing estimates of the global pattern of mortality by groups of causes, the WHO/World Bank project “Global Burden of Disease” (Murray and Lopez, 1996) made use of regression models based on “all-cause” mortality for a country or region. Cancer mortality is estimated based on the observation that the proportion of deaths due to certain groups of diseases (e.g., infectious and parasitic diseases, maternal mortality, chronic diseases) correlate closely with the all-causes combined rate. The precise profile of various cancers within this global estimate is derived for any available data on relative frequencies of the various types of cancer. The International Agency for Research on Cancer (IARC), with its “GLOBOCAN” estimates, prepares national estimates of incidence, mortality, and prevalence of cancer that are based on all available sources of data in many countries. The level of accuracy depends on the extent and quality of the locally available data. The most recent country-level estimates have been provided for 24 cancers and five broad age groups. These estimates are available on CD-ROM (Ferlay et al., 2001) and, in a format allowing rather less flexibility in analysis and presentation, on the Internet (http://wwwdep.iarc.fr/globocan/globocan.html). The sources of data and the methods used to produce estimates of incidence, mortality, and prevalence are summarized in several reports (Parkin et al., 1999; Pisani et al., 1999, 2002).

VARIATION BY PLACE Geographic comparisons of cancer rates or risk frequently use national populations as the unit of study. The reason is that this dimension is the one for which statistics—especially mortality—are collected and published. Differences between countries may indeed be striking. National boundaries, however, have not always been based on levels of exposure to environmental risk factors of cancer or on the genetic homogeneity of the populations within them. Thus, studying populations within, and sometimes across, national boundaries has been particularly informative. In this context, the cancer atlas has proved to be a popular method of illustrating differences in risk between geographic subunits of national populations. It may show quite dramatic geographic patterns that are highly informative in suggesting likely etiologies. The atlas concept has been applied to wider geographic areas, too, if the disease data are truly comparable across national boundaries. Examples are the atlases of cancer mortality in Europe (Smans et al., 1992; WHO, 1997) (Fig. 8–2), and of cancer incidence in the Nordic countries/ northern Europe (Pukkala et al., 2001). Maps of disease rates according to geographic unit raise several technical issues with respect to how to illustrate gradients of risk (e.g., color scales) and the most appropriate indicator to plot. Use of disease rates per se means that if the population units are small or the disease is rare random variation may be responsible for many of the perceived differences. Various solutions have been proposed that incorporate the idea that the values for units that are geographically close should be more similar than those for more distant units (Wakefield et al., 2000).

International Patterns of Cancer Incidence and Mortality

6 . 00

105

2

3 . 93 2 . 81 1 . 32 0 . 84 0 . 60 0 . 42 0 . 00

INTERPRETATION OF GEOGRAPHIC VARIATION IN DISEASE RISK: PERSON OR PLACE? Artifact aside, the principal question posed by observed geographic differences in risk is how much is due to variation in exposure (to “carcinogens” or “risk factors”) and how much is the result of inherent differences in susceptibility to such exposures of the population resident in a particular place (and hence genetically determined). Of course, the major exposure to carcinogens is not through variations in the external environment (e.g., air, water, radiation) so much as in differences in lifestyle (e.g., reproduction, diet, tobacco use). These are culturally determined exposures and so are linked closely to sociocultural groups of the world population. If we wish to isolate the component of risk associated with genetic characteristics of a population (ethnicity), the first consideration is to eliminate the effect of these confounding variables associated with the risk of disease and differentially distributed by ethnic group. If the variable of primary interest is ethnicity, or racial group, the differences observed within the same geographic locality are more meaningful because at least some of the environmental differences present in international comparisons are reduced or eliminated. There are numerous examples of such studies from multiethnic populations in all parts of the world, such as the

Figure 8–2. European atlas of esophageal cancer mortality. Source: Smans M, Muir CS, Boyle P. 1992. Atlas of Cancer Mortality in the European Economic Community. IARC Scientific Publication No. 107.

white and black populations of Harare, Zimbabwe (Bassett et al., 1995), the Chinese, Indian, and Malay populations of Singapore (Lee et al., 1988), and, above all, the many ethnic populations in the United States (Miller et al., 1996) (see Fig. 8–22).

Ethnicity From an epidemiologic point of view, the variable ethnicity, or race, defines a constellation of genetic factors that relate to susceptibility to a given cancer. Of course, there is considerable variation within a given ethnic or racial group (however this is defined), but there are often sufficiently large differences between them to yield distinctive patterns of risk. Genetically determined risk may be mediated by several mechanisms (Easton, 1994; Ishibe and Kelsey, 1997; Norppa, 1997). 1. Germline mutations of genes that are normally concerned with the regulation of cell growth (oncogenes, tumor suppresser genes) 2. Variation in the genes (polymorphisms) that modulate the impact of environmental carcinogens a. Polymorphisms of carcinogen-metabolizing enzymes b. Inherited differences in DNA adduct formation c. Variation in the ability to repair DNA lesions induced by a carcinogen

106

PART II: THE MAGNITUDE OF CANCER

A striking illustration of the likely influence of genetic factors on risk is provided by certain childhood cancers. For several (e.g., Wilms’ tumor, Ewing’s sarcoma), there are marked differences in incidence between ethnic groups for which no plausible environmental “exposure” can be imagined. Any such exposure would have to be highly carcinogenic (to act so early in life), highly tissue-specific, and unevenly distributed among the ethnic group. In any case, some genetically determinant traits (e.g., skin pigmentation) are manifestly connected with susceptibility to carcinogens (ultraviolet light), so it is easy to imagine others that do not produce such clearly visible phenotypes.

Migrants The most fruitful approach to compiling statistics using routine data sources is to study migrants (Parkin and Khlat, 1996). The rationale is simple. The risk of various cancers in a given migrant population is compared with the risk of the host population (similar environment, different genetics) and with the population living in the place of origin (similar genetics, different environment). Ideally, such comparisons take into account the age at migration or the duration of residence in the two environments—original (origin) and new (host)—as a crude means of quantifying “exposure” to the new environment. This approach may also be able to compare the risk for the migrants with that for their offspring, who have lived in the new “environment” throughout their lives. In the context of classic migrant studies, relying on routinely collected information (descriptive studies), all environmental exposures are subsumed by a single variable (place of residence). Many specific exposures are associated with this, not only via the external environment (air, soil, water) but also through sociocultural factors (e.g., diet, fertility, smoking). Occasionally, there is information on other indirectly associated variables (e.g., socioeconomic status) that can be taken into account when estimating the ethnicityassociated component of risk.

VARIATION IN TIME Investigating trends of cancer incidence and mortality rates over time has important applications in both public health planning and epidemiologic research. In stable populations, a change in the incidence of a cancer should reflect changes in exposure to environmental risk factors, rather than differences in genetic susceptibility. Time trend studies therefore help when generating etiologic hypotheses or confirming suspected associations, when, for example, exposures to putative agents are known to be changing over time. Time trend studies are also widely used to evaluate cancer control programs and to study the effect of primary prevention interventions (planned or unplanned), programs of early detection, and the efficacy of treatment protocols.

WHAT TO STUDY? The strengths and weaknesses of incidence and mortality data as the basis for studying trends have been much debated. There are well documented complexities when interpreting trends of both incidence and mortality data (Saxén, 1982; Muir et al., 1994; Swerdlow et al., 2001). Some of the problems associated with defining incidence were described earlier; artifactual changes in rates, for example, may be caused by changing the completeness of the registration, improving diagnostic methods, and inaccurate population estimates at the subnational level. Although incidence data are generally of good diagnostic quality, there may be problems with the accuracy of coded “cause of death” in mortality data (mentioned earlier) and they may be subject to change with time. Bias is introduced by improvements in survival when mortality rates are being used as a proxy for incidence rates (as they often are, given their more extensive availability). Often a combined description of these indicators serves to confirm and clarify our understanding of the underlying disease processes, but it

is fundamental that the properties and interrelations are a priori understood. Although valuable information can be gleaned about temporal risk patterns from graphics depicting age-standardized rates over time (as used in this chapter), strictly speaking summary rates are accurate only as measures of risk in the absence of an interaction between age and time. Trends in age-standardized rates by calendar period may mask important changes in the age-specific rates, particularly in the presence of strong birth cohort effects (Day and Charnay, 1982).

AGE, PERIOD, AND COHORT EFFECTS Age is the most powerful determinant of cancer risk, as age parallels the cumulative exposure to carcinogens over time and the accumulation of the series of mutations necessary for the unregulated cell proliferation that is cancer (Peto et al., 1985). The effects of period and birth cohort, on the other hand, can be seen as weak proxies for events we cannot measure directly. The examination of rates by birth cohort is an essential part of a temporal analysis of diseases such as cancer, for which there is a long induction phase. Changes in lifestyle and environmental risk factors tend to affect particular generations of individuals and presumably influence the earlier stages of the carcinogenesis process. Period effects, on the other hand, may act as surrogate measures of events that quickly change incidence or mortality. These may be interventions on the later stages of carcinogenesis or artifactual influences on incidence (changes in coding practice or diagnostic methods); rapid alterations in mortality can be the result of an improvement in survival. A formal quantification of the separate age-adjusted contributions of the two time effects may thus give insight into the underlying nature of time trends. The classic analytic approach involves fitting age, period, and cohort as explanatory variables in a log-linear Poisson regression model of the number of disease events, offset by the corresponding person-years. Age-period-cohort modeling has been widely used in time trend studies, but there are inherent limits to the approach owing to nonidentifiability—the fact that knowledge of any two factors implies knowledge of the third, making one of the factors redundant (Barrett, 1978). Many “solutions” to estimating the joint effects of the parameters have been proposed; methods that do not impose arbitrary assumptions unsupported by the data are considered the most appropriate (Clayton and Schifflers, 1987a, 1987b).

GLOBAL BURDEN There were an estimated 10.1 million new cases of, 6.2 million deaths due to, and 22 million persons living with cancer (within 5 years of diagnosis) during the year 2000. The total “all cancers” category excludes nonmelanoma skin cancers because of the difficult measurement and consequent lack of data. The 2000 estimates represent an increase of around 22% in incidence and mortality for 1990 (Parkin et al., 1999; Pisani et al., 1999). The cancer profile is rather different, depending on whether incidence or mortality is the focus of interest, as shown in Figure 8–3. In terms of incidence, the most common cancers are in the lung (12.3%), breast (10.4%), and colorectum (9.4%). The most common causes of death due to cancer are cancers of the lung (17.8%), stomach (10.4%), and liver (8.8%). Figure 8–4 shows the 12 most common cancers for males and females (as number of new cases) in the developing and developed regions of the world. Developed countries comprise those of North America, Europe (including all of Russia), Australia/New Zealand, and Japan; developing countries comprise the remainder. Figure 8–5 shows the most prevalent cancers in men and women together with the annual number of new cases at the same site. In terms of prevalence, the most common cancers are those of the breast (17.2%), colorectum (10.6%), and prostate (6.9%). The ratio between prevalence and incidence is an indicator of prognosis; thus, breast cancer is the most prevalent cancer in the world despite the fact that

107

International Patterns of Cancer Incidence and Mortality

Figure 8–3. The number of new cases and deaths (thousands) for the 15 most common cancers worldwide, by sex, 2002. (Source: Ferley et al., 2001.)

there are fewer new cases than for stomach cancer (for females) or lung cancer, for which the outlook is considerably poorer. Table 8–2 shows incidence rates for all cancers (excluding skin) by world area and sex. Age-standardized rates in developed countries are about twice those in developing countries; the differential is less for the cumulative rate, which ignores disease rates in the 65+ age group. On average, worldwide there is about a 10% chance of getting a cancer before age 65. Incidence (and mortality) rates are highest in North America, Australia/New Zealand, and western Europe and lowest in parts of Africa. This overall risk, of course, depends on the contributions of various types of cancer. For example, in West Africa, the incidence of almost all cancers is low (except for cervical cancer in women and liver cancer in men). This contrasts with southern Africa, which has, in addition, high rates of lung and esophageal cancer, and East Africa, with high rates of acquired immunodeficiency disease syndrome (AIDS)-related tumors, notably Kaposi’s sarcoma.

CANCER PATTERNS AND TRENDS: BY SITE In this section, the frequency, geographic distribution, and recent trends in incidence and mortality are described for the major cancer sites. The age-standardized incidence rates are taken directly from the CI5 volumes (Doll et al., 1966, 1970; Waterhouse et al., 1976, 1982; Muir et al., 1987; Parkin et al., 1992, 1997, 2002), whereas the 3-year rolling average mortality rates are based on data extracted from the WHO mortality databank. U.S. incidence rates (as 3-year rolling averages, as above) by race are calculated from data from the Surveillance Epidemiology and End Results (SEER) registries from the U.S. National Cancer Institute (Ries et al., 2001); the corresponding U.S.

Table 8–2. Incidence Rates for all Cancers in 2000 (excluding nonmelanoma skin cancer) by World Region ASR (World) per 100,000

Eastern Africa Middle Africa Northern Africa Southern Arica Western Africa Caribbean Central America South America Northern America Eastern Asia South-Eastern Asia South Central Asia Western Asia Eastern Europe Northern Europe Southern Europe Western Europe Australia/New Zealand Melanesia Micronesia Polynesia More developed countries Less developed countries WORLD Source: Parkin (2001).

Cumulative Risk % (age 0–64)

Male

Female

Male

Female

177.7 141.8 124.5 217.5 81.2 187.9 178.5 201.4 357.4 205.3 131.1 106.6 151.1 290.0 263.4 275.4 318.7 358.6 149.8 175.5 200.7 301.0 153.8 201.9

176.4 121.6 106.8 153.7 94.1 177.2 213.8 201.8 281.5 126.6 120.1 112.0 111.3 197.2 235.1 194.3 230.6 283.2 178.3 149.6 216.3 218.3 127.9 157.8

9.4 7.7 6.8 9.4 4.8 7.6 7.5 9.3 16.2 10.5 7.0 6.2 8.0 16.2 10.9 13.3 14.9 15.6 7.4 9.4 9.9 14.4 8.2 10.0

11.3 7.9 7.2 8.7 6.6 9.7 11.9 11.2 15.3 7.3 7.8 7.8 6.9 12.4 13 11.1 13.2 15.8 10.9 9.1 13.1 12.5 8.0 9.2

472

Figure 8–4. The 12 most common cancers for males and females (new cases, in thousands), in the developing and developed regions of the world, 2002. (Source: Ferley et al., 2001.)

Figure 8–5. The most prevalent cancers and the number of new cases (thousands) at the same site 2002, by sex. (Source: Ferley et al., 2001.)

108

International Patterns of Cancer Incidence and Mortality death rates are based on mortality data encompassing the whole country obtained from the National Center for Health Statistics (NCHS), again by race. The incidence and mortality rates were ageadjusted using the weights from the world standard population. the following ICD codes were used for tabulating site-specific cancer data: esophagus (ICD7–9: 150); stomach (ICD7–9: 151); colon and rectum (ICD7–9: 153–154); liver (ICD9: 155.0, 155.1, 155.2); lung (ICD7: 162, 163; ICD8–9: 162); female breast (ICD7: 170; ICD8–9: 174); cervix uteri (ICD9: 171, ICD8–9: 180); prostate (ICD7: 177; ICD8–9: 185); bladder (ICD7: 181.0; ICD8–9: 188); and non-Hodgkin’s lymphoma (ICD7–9: 200 and 202).

Lung Cancer Lung cancer is the leading cancer in the world today (12.3% of all new cancers, 17.8% of cancer deaths). There were an estimated 1.2 million new cases and 1.1 million deaths in 2000, with the sex ratio (M : F) 2.7. The proportion of all cancer deaths due to lung cancer is substantially higher in developed (22% cancer deaths) than in developing (14.6% of deaths) countries. Geographic patterns are greatly influenced by past exposure to tobacco smoking (Doll and Peto, 1981a, 1981b). The areas with the highest incidence and mortality are Europe (especially eastern Europe), North America, Australia/New Zealand, and South America. The rates in China, Japan, and Southeast Asia are moderately high, and the lowest rates are found in southern Asia (India, Pakistan) and sub-Saharan Africa (excluding South Africa) (Fig. 8–6a). In certain population subgroups (e.g., U.S. African Americans and New Zealand Maoris), the incidence is even higher; and with current incidence rates, men in these two groups have about a 13% chance of developing lung cancer before age 75. In developing countries the highest rates are seen where the habit of tobacco smoking has been longest established: Middle East, China, Caribbean, South Africa, Zimbabwe, and the Pacific. In women, the geographic pattern reflects the historical patterns of smoking, which are different from those in men. Thus, the highest rates are observed in North America and northwestern Europe (United Kingdom, Iceland, Denmark), with moderate rates in Australia, New Zealand, and China (Fig. 8–6b). The proportion of lung cancer cases due to tobacco smoking has been estimated by comparing the incidence (or mortality) rates in various areas with the rates in nonsmokers observed in large cohort studies (Parkin et al., 1994; Peto et al., 1994). For the year 2000, an estimated 85% of lung cancer in men and 47% in women is the consequence of tobacco smoking. The percentage is 90%–95% of cases in men in Europe and North America. Only in the lowest incidence areas of east and west Africa are there no attributable cases. The fractions are lower for women, with several areas (where incidence rates are lower than in nonsmoking women in the United States and Japan), including south-central Asia, having no attributable cases. The highest fractions are in North America (85%), northern Europe (74%), and Australia/New Zealand (72%), where women have been smoking the longest. Almost all lung cancers are carcinomas (other histologies comprise well under 1%). In the combined data from the series published in the CI5 volumes (Parkin et al., 2002), small-cell carcinomas comprise about 20% of cases and large-cell/undifferentiated carcinomas about 9%. Except for the other histologic types, the proportions differ by sex: squamous cell carcinomas comprise 44% of lung cancers in men and 25% in women, and adenocarcinomas comprise 28% of cases in men and 42% in women. Figure 8–7 shows overall incidence rates and the estimated rates by histologic subtype for 30 populations for which a relatively high proportion of cases had a clear morphologic diagnosis (Parkin et al., 2002). Among men, with the exception of certain Asian populations (Chinese, Japanese), only in North America (United States, Canada) does the incidence of adenocarcinoma exceed that of squamous cell carcinoma. In women, however, adenocarcinoma is the dominant histologic type almost everywhere, except Poland, where squamous cell carcinomas predominate, and the United Kingdom (England, Scotland), where both squamous cell and small-cell carcinomas have higher rates. Adenocarcinomas are particularly predomi-

109

nant in Asian women (72% of cancers in Japan, 65% in Korea, 61% in Singapore Chinese). The differences in histologic profiles are strongly influenced by the evolution of the epidemic of smokingrelated lung cancer over time (see below).

Time Trends Trends in lung cancer incidence and mortality reflect population-level changes in smoking behavior, including dose, duration, and type of tobacco used (Gilliland and Samet, 1994; Lopez-Abente et al., 1995). Incidence or mortality by age group is closely related to birth cohort; in the United Kingdom and the United States, the cohort-specific incidence is related to the smoking habits of the same generation (Brown and Kessler, 1988; Lee et al., 1990). Thus, the men in countries where smoking was first established were first to see a diminution in smoking prevalence followed (in the same generations of men) by a decline in risk. Changes are therefore first seen among younger age groups (Muir et al., 1994); and as these generations of men reach an older age, when lung cancer is most common, a decline in overall incidence and mortality is seen. The United Kingdom was the first to show this decline (with the incidence/mortality falling since 1970–1974), followed by Finland, Australia, The Netherlands, New Zealand, the United States, Singapore, and, more recently Denmark, Germany, Italy, and Sweden (Bray et al, 2003) (Fig. 8–8). In most other countries there is a continuing rise in rates, which is most dramatic in the countries of eastern Europe (Borras et al., 2003). In women, the tobacco habit has usually been acquired recently or not at all. Thus, the most common picture in western populations is of rising rates, whereas in many developing countries (where female smoking generally remains rare) lung cancer rates remain low. A few countries where the prevalence of smoking among women is declining already show decreasing rates among younger women; and in the United Kingdom, where this trend is longest established, there has been a decline in overall incidence and mortality since about 1990. There are, however, clear differences in time trends by histologic type. In the United States (Devesa et al., 1991; Travis et al., 1996) squamous cell carcinoma reached a maximum incidence in men in 1981, but the incidence of adenocarcinoma continued to rise (until about 1987 in black men and around 1991 in white men). As a result, adenocarcinoma is now the most frequent form of lung cancer in men (Fig. 8–7), whereas it had constituted only a small number of cases (around 5%) during the 1950s (Wynder and Graham, 1950). In contrast, the incidence of both histologic types has continued to increase among women, though there is a suggestion that the incidence of squamous cell carcinomas had reached its maximum by 1990. These changes were related to specific birth cohorts, with the maximum incidence in men in the 1925–1929 cohort for squamous cell carcinomas and in the 1935–1939 cohort for adenocarcinomas; in women this occurred some 10–20 years later (Devesa et al., 1991; Zheng et al., 1994). Somewhat similar observations (increasing adenocarcinoma and decreasing squamous cell carcinoma) have been reported from The Netherlands (Janssen-Heijnen et al., 1995), Japan (Sobue et al., 1999), and the United Kingdom (Harkness et al., 2002). Part of this differential trend may be due to artifact (changes in classification and coding, improved diagnostic methods for peripheral tumors). In part, it may be due to an ever-increasing proportion of ex-smokers in the population, as the decline in the risk of lung cancer on smoking cessation is faster for squamous cell tumors than for small-cell carcinomas and adenocarcinomas (Lubin and Blot, 1984; Jedrychowski et al., 1992). It seems likely also that changes in cigarette composition (to low tar, low nicotine, filtered cigarettes) are partly responsible (Wynder and Muscat, 1995; Charloux et al., 1997).

Breast Cancer Breast cancer is the second most common cancer in the world today and by far the most common cancer in women. There are 1.05 million new cases each year (about 22% of cancers in women) and 373,000 deaths. More than half of the cases are in industrialized countries: about 346,000 in Europe (27% of cancers in women) and 202,000 in North America (31%). Breast cancer is relatively less common among

< 19.5 A

< 38.5

< 76.5

< 13.6

< 20.4

< 27.2

Age standardized (world) rate (per 100,000)

Figure 8–6. Incidence of lung cancer in the world. A, Males—all ages; B, Females—all ages.

110

< 95.5

Age standardized (world) rate (per 100,000)

< 6.8 B

<57.5

< 34.0

Men

0

10

20

30

40

50

60

70

80

USA, SEER : Black

90

Women

85.8

USA, SEER : Black

Croatia

73.6

*UK, Scotland

The Netherlands, Eindhoven

72.8

USA, SEER : White

*UK, Scotland *Slovakia Italy, Varese Province Canada East (3 registries) *Poland (3 registries) Slovenia

71.4 71.0 67.9

Canada East (3 registries)

67.8 67.7

*UK, England

65.9 63.6 58.1 55.8

Germany, Saarland Belgium (2 registries) Singapore: Chinese USA, SEER : White

54.3 53.9

France (6 registries)

Switzerland (3 registries) Australia (3 registries) Japan (2 registries)

41.4 36.3 35.1

Norway Israel: Non_Jews Israel: Jews

28.6 22.1 22.0

Brazil, Goiania Sweden USA, Puerto Rico India (2 registries)

19.5 7.6

* Percentage of unspecified cancers > 25%

Figure 8–7. Age standardized incidence rates of lung cancer by cell type, ASR per 100,000.

19.9 16.9 16.5 16.3

The Netherlands, Eindhoven

14.7 14.1

Switzerland (4 registries) Germany, Saarland

13.4 12.9 12.7

Austria (2 registries)

12.3 11.7

Japan (2 registries) *Croatia

10.6

Israel: Jews

10.3 10.1 10.0

Slovenia Belgium (2 registries) Italy,Varese

9.7

Brazil,Goiania

8.5 8.1

*Slovakia France (6 registries) USA, Puerto Rico Spain (2 registries) Israel: Non_Jews India (2 registries)

30

21.9

Australia

Korea (2 registries)

47.5 43.4 42.5

25

35

40

35.4 34.6 33.2 29.9 29.6

Norway

51.1

Canada West (3 registries)

20

36.7

Poland (3 registries)

Austria (2 registries)

Spain (2 registries)

15

Singapore: Chinese

Sweden

Korea (2 registries)

10

Denmark

51.1 48.9 48.7 47.9

5

Canada West (3 registries)

*UK, England Denmark

0

7.0 6.5 4.6 4.2 2.4 Squamous cell carcinoma Adenocarcinoma Small cell carcinoma Large and undifferentiated cell carcinoma Others and unspecified cancer

Europe

Americas

England, South Thames

Finland

Oceania/Asia

Canada

100

100

Australia, New South Wales

Colombia, Cali

100

100

100

50

50

China, Shanghai 100

IncM

50

IncM

IncM

50

50

IncM

50

IncM

IncF IncM

IncF

IncF

10

10

10

5

5

1 .5

IncF

10

10

10

5

5

5

5

1

1

1

1

1

.5

.5

.5

.5

.5

Spain, Zaragoza

Slovakia

IncF

Puerto Rico 100

100

United States of America

100

IncBM

IncM

India, Mumbai (Bombay) 100

100

50

50

50

50 IncM

10

10

5

5

100

10

IncF

IncM

50

IncBF IncWF

10

Japan, Miyagi Prefecture

IncWM IncM

50

IncF

10

IncM

IncF

10

IncF

5

5

5

IncF

5 IncF

1

1

1

1

1

1

.5

.5

.5

.5

.5

.5

1960 1970 1980 1990 2000

1960 1970 1980 1990 2000

Figure 8–8. Lung cancer incidence trends. (Source: C15/SEER/NCHS.)

1960 1970 1980 1990 2000

1960 1970 1980 1990 2000

1960 1970 1980 1990 2000

1960 1970 1980 1990 2000

113

International Patterns of Cancer Incidence and Mortality women in developing countries, although it still accounts for 18% of female cancers, and the incidence is increasing. The country with the highest incidence is The Netherlands (ASR of 91.6 per 100,000). The incidence in the United States is 91.4; and within the United States there are populations with age-adjusted rates of 100 or more (e.g., white women in California and Hawaiian women) (Parkin et al., 2002). High rates are also observed in Europe, Australia, New Zealand, Uruguay, and Argentina. In contrast, low rates are found in most African and Asian populations, although they are increasing; in some Asian populations they are already the same as in southern Europe, and in some cases (e.g., the Philippines) they are even higher (Fig. 8–9). The incidence in the Jewish population of Israel is especially high (87.1/100,000). Survival from breast cancer in Europe is 91% at 1 year and 65% at 5 years (Berrino et al., 1999). The stage of disease at diagnosis is the most important prognostic variable. Based on the SEER registries in the United States, 5-year survival for localized cases is about 97%, but for cases with metastases it is only 20% (Ries et al., 2001). Even in developing countries, the differences by stage at diagnosis are marked (Sankaranarayanan et al., 1998). Because of this relatively good prognosis, breast cancer is the most prevalent cancer in the world today; there are an estimated 3.7 million women alive who have had breast cancer diagnosed within the last 5 years (compared with just 1.3 million survivors—male or female—from lung cancer). It has been estimated that 1.5% of the U.S. female population are survivors of breast cancer (Hewitt et al., 1999). Trends in survival show clear improvement over time (Chu et al., 1996; Quinn et al., 1999). The risk of breast cancer increases with age, but the rate of increase slows after menopause, the lowering of risk coinciding with a decrease in circulating estrogens (Henderson et al., 1988). In low-incidence countries, the slope of the age–incidence curve after menopause may be flat or even negative (Fig. 8–10). This almost certainly reflects

< 19.3

< 26.1

an increasing risk in successive generations of women rather than a true decline in risk with age (Moolgavkar et al., 1979). The young age structure of the populations in many developing countries coupled with this rather flat age–incidence curve means that the average age at diagnosis is lower than in European and American populations. The international and interethnic differences in the incidence of breast cancer are for the most part the consequence of different environmental exposures. This is clear from studies of migrants, which clearly show that the incidence changes after migration; for example, a rise in the risk of breast cancer in populations from European countries at relatively low risk (Italy, Poland) occurs after migration to Australia, particularly if the subjects migrate as children (Geddes et al., 1993; Tyczynski et al., 1994). Furthermore, studies comparing the risks for migrants and their offspring (particularly among Asians migrating to the United States) demonstrate that there are major increases in risk between the first, second, and third generations (Ziegler et al., 1993). Genetic factors, including the major susceptibility genes (BRCA1, BRCA2), may account for up to 10% of breast cancer cases in developed countries (McPherson et al., 2000), but their prevalence in the population is too low to explain much of the international variation. The highest frequency of BRCA1 so far reported is in Ashkenazi Jews (around 1%) (Goldgar and Reilly, 1995). It is 10 times higher than in the general population of the United States or United Kingdom. Even with an associated relative risk of 50 (Ford et al., 1995), however, this would account for only a 40%–50% higher incidence. Nothing is known, so far, concerning differences between populations with respect to polymorphisms of genes concerned with the control of estrogen metabolism. Internationally, there is some association between the national incidence (or mortality) rates for breast cancer and population averages

< 36.0

< 54.1

Age standardized (world) rate (per 100,000) Figure 8–9. Incidence of female breast cancer.

< 91.6

114

PART II: THE MAGNITUDE OF CANCER Rate (per 100,000) 1000

100

10

1

0.1 0-

5- 10- 15- 20- 25- 30- 35- 40- 45- 50- 55- 60- 65- 70- 75- 80+ Age

San Francisco: NonHispanic White

Japan, Miyagi

Finland

India, Trivandrum

Singapore: Chinese

Thailand, Khon Kaen

recent to ascribe to screening. A more likely explanation is that it is the result of improved therapy, as shown by improved survival within stages. Improved treatment for node-positive disease (chemotherapy for premenopausal women, tamoxifen for those who are postmenopausal) became established practice during the mid-1980s and for node-negative patients by the end of that decade. In Japan, breast cancer remains relatively rare, but the incidence and mortality rates are increasing rapidly for successive generations. There are many possible explanations, including decreasing age at menarche, increasing age at menopause, decreasing fertility, increasing age at first birth, and increased height and weight (Wakai et al., 1995). Since 1975, the increase in incidence has been larger than the increase in mortality, indicating an improvement in survival. In the populations of developing countries, the breast cancer incidence and mortality are increasing, often more markedly among the younger generations of women (Parkin, 1994). For example, reported increases were 1% per year between 1964 and 1985 in Bombay (Yeole et al., 1990), 2.7% per year in Shanghai between 1972–1974 and 1992–1993 (Jin et al., 1999), 3.6% per year in Singapore between 1968 and 1992 (Seow et al., 1996), and 3.6% per year in Hong Kong during the period 1973–1999 (Leung et al., 2003). An exception may be the relatively high-risk populations in South American countries such as Uruguay and Chile, where the observed mortality rates for younger women have been more or less constant (Parkin, 1994).

Colon and Rectum

Figure 8–10. Age-specific incidence of breast cancer. (Source: Parkin et al., 2002.)

for the variables related to fertility (Parkin, 1989) or body weight (Bergström et al., 2001). However, such models can explain only a minor component of the variation in incidence. Similarly, in the United States the geographic variation in incidence is only partially explicable in terms of the prevalence of risk factors (Sturgeon et al., 1995; Laden et al., 1997). It has often been noted that the breast cancer incidence rates are associated with higher socioeconomic status (as estimated by such factors as income, education, and housing, among others). There are several reasons for this finding. First, when looking at mortality, one should take into account the fact that survival is clearly lower among the lower socioeconomic groups (Karjalainen and Pukkala, 1990; Schrijvers et al., 1995). Second, it seems that most of the gradient can be explained by the differing prevalence of known risk factors among social classes. In the United States, for example, the variation in risk by educational level or annual income is almost entirely explained by the differential distribution of factors such as parity, age at menstruation and menopause, obesity, height, and alcohol consumption (Heck and Pamuk, 1997).

Time Trends Early large-scale reviews of international trends in the incidence and mortality from breast cancer (e.g., Coleman et al., 1993; Ursin et al., 1994) showed, for the most part, that the incidence was increasing, with less marked, although similar, changes in mortality rates. Incidence continues to increase in most populations (Fig. 8–11), although a decrease in mortality rates due to breast cancer can be seen in several countries. This was first noted in the United States (Blot and Fraumeni, 1987), but it is also evident in Canada and some European countries (e.g., United Kingdom, The Netherlands, Denmark, Norway) (Hermon and Beral, 1996). Furthermore, this decline coincided, more or less, with the introduction of screening programs for breast cancer, accompanied by a brisk rise in incidence, as “prevalent” (undetected and asymptomatic) cancers were diagnosed by screening (Chu et al., 1996; Persson et al., 1998; Quinn et al., 1999). There has been considerable debate concerning the relative contributions to the observed trends of screening and of improved therapy (Blanks et al., 2000; Peto et al., 2000). The observed decreases in mortality are, in general, too

Cancers of the colon and rectum comprise the third most frequent form of malignancy worldwide in both sexes (about 945,000 new cases in 2000), with two-thirds occurring in developed countries, where colorectal cancer incidence ranks second only to lung cancer (Fig. 8–4). The 5-year survival ranges from 45% to 60%, and the variation between developed and developing countries is not particularly large. The relatively good prognosis means that the mortality rate is about half that of the incidence (about 492,000 deaths in 2000). The prevalence of colorectal cancer is second only to breast cancer worldwide, with an estimated 2.4 million persons alive with colorectal cancer diagnosed within the previous 5 years. There is at least a 25-fold variation in occurrence worldwide. The highest incidence rates are in Australia, New Zealand, North America, western and eastern Europe (rates in the Czech Republic and Hungary are among the highest worldwide), and Japan (Fig. 8–12). Rates from several Japanese registries are particularly elevated, notably in Hiroshima (Parkin et al., 2002). Incidence tends to be low in Africa and Asia and intermediate in southern parts of South America (Fig. 8–11). Whereas the geographic distribution of colon and rectal cancer is similar between the sexes, the variation between countries is more striking for colon cancer than for rectal cancer. Furthermore, cancer of the rectum is often 20%–50% more common in men than women in most populations. Thus, in high risk populations the colon/rectum ratio is 2 : 1 or more (rather more in females). In low risk countries, rates are generally of the same magnitude. The landmark studies of Japanese migrants to the United States (e.g., Haenszel, 1968) showed that when populations moved from low risk to high risk areas the incidence of colorectal cancer increased rapidly in the first generation, implying that dietary and other environmental factors constitute a major component of risk. Japanese born in the United States have higher rates than those of U.S. whites, and rates for Japanese living in Hawaii and Los Angeles are among the highest in the world (Parkin et al., 2002).

Time Trends Although there are distinct differences between the epidemiology of colon and rectal cancer (see Chapter 42), it is common to study time trends in incidence and mortality for large-bowel cancer as a whole to avoid the well known problems of varying subsite allocation of tumors at the rectosigmoid junction. Overall, the incidence of colorectal cancer is increasing rather rapidly in countries where overall risk was formerly low (Japan, Puerto Rico), whereas in high risk countries trends are either gradually increasing (England), stabilizing (New

Americas

Europe England, South Thames 100

Canada

Finland

IncF

Oceania/Asia

100

100

50

50

Australia, New South Wales

IncF

IncF

50

Colombia, Cali 100

100

50

50

China, Shanghai 100

IncF

50

IncF MortF

IncF

MortF

10

10

10

10

10

10

5

5

5

5

5

5

1

1

1

1

1

1

.5

.5

.5

.5

.5

.5

Slovakia

Spain, Zaragoza

Puerto Rico

100

100

United States of America

100

India, Mumbai (Bombay) 100

100

50

50

Japan, Miyagi Prefecture 100

IncWF

50

IncF

IncF

50

50

IncF

IncBF

50 IncF

IncF

MortBF MortWF

10

10

10

10

10

10

5

5

5

5

5

5

1

1

1

1

1

1

.5

.5

.5

.5

.5

.5

1960 1970 1980 1990 2000

1960 1970 1980 1990 2000

1960 1970 1980 1990 2000

Figure 8–11. Female breast cancer incidence and mortality trends. (Source: C15/SEER/NCHS.)

1960 1970 1980 1990 2000

1960 1970 1980 1990 2000

1960 1970 1980 1990 2000

116

PART II: THE MAGNITUDE OF CANCER

< 12.9

< 24.7

< 36.6

< 48.4

< 60.3

Age standardized (world) rate (per 100,000) Figure 8–12. Incidence of colorectal cancer, males all ages.

South Wales, Australia), or declining (North America) with time (Fig. 8–13). Such moderations with time have been noted particularly among the younger age groups (Coleman et al., 1993; McMichael and Giles, 1994). There are exceptions, however. Rates in Bombay, a population with low rates, have continued to fall since the 1960s (Fig. 8–13). The greatest increases in the incidence of colorectal cancer are in Asian countries (Japan, Hong Kong, Singapore) and Israel, several Eastern European countries (Hungary, Poland), the Americas, and Puerto Rico. In contrast to the recent attenuation of rates seen in some western and northern European countries, relatively large increases have been observed in Spain (Lopez-Abente et al., 1997). In the United States, the incidence rates in both white men and women began to decline during the mid-1980s (Fig. 8–12) (Chu, 1994; Troisi et al., 1999; Ries et al., 2001). A slightly later decline was also observed for black men, but no such trend has been apparent in the incidence rates in black women (Fig. 8–12). The changes in mortality (Fig. 8–12) may be a consequence of changes in incidence as well as a result of progress in therapy; alternatively, as in the United States (Troisi et al., 1999), they may include the effects of improved early detection probably due to more widespread screening, resulting in stage-specific shifts in incidence and a subsequent decrease in case fatality. Generally, in high-risk populations there have been increases in the incidence of proximal tumors (ascending colon) relative to distal tumors (descending and sigmoid colon) (Devesa and Chow, 1993; Thorn et al., 1998; Troisi et al., 1999; Svensson et al., 2002). In low risk populations (e.g., in Singapore), however, the contrary has been observed (Huang et al., 1999), whereas the growth in proximal and distal disease rates were similar in Shanghai (Ji et al., 1998). For rectal cancers, the countries with the most rapid upsurges in incidence and mortality tend to be in eastern Europe and Japan. In the United States,

there has been a decline in incidence and mortality for several decades in females of both races and in white men, although a recent increase in rectal cancer is apparent in black men (Troisi et al., 1999). The reasons for the geographic and temporal variations in risk are certainly multiple and interrelated. It is possible that the apparent declines in the incidence of distal cancers seen in some western populations are a consequence of detecting and treating premalignant polyps (Chu et al., 1994; Troisi et al., 1999; Hayne et al., 2001). However, the principal cause of the increased risk in Japan, Hungary, and Israel (Jews) is probably via modifications toward a more “western” lifestyle, particularly with respect to diet (WCRF, 1997). The effect in Asian populations may be augmented by genetic susceptibility (Le Marchand, 1999). The converse effect, with some improvements in the quality of diet in younger generations, may explain the observation (notably in the United States and Europe) of cohort-led declines in incidence rates pertaining to younger age groups (Chow et al., 1991; Coleman et al., 1993; Thorn et al., 1998; Svensson et al., 2002).

Stomach Cancer Until recently, stomach cancer was the second most common cancer worldwide, but with an estimated 876,000 over new cases per year in 2000 (8.7% of new cancer cases) it is in fourth place, behind cancers of the breast and large bowel. It is the second most common cause of death from cancer (647,000 deaths annually). Almost two-thirds of the cases occur in developing countries. The geographic distribution of stomach cancer is characterized by wide international variations; high risk areas (ASR in males >30 per 100,000) include East Asia (China, Japan, Korea, Mongolia), Central Asia, eastern Europe, and parts of Central and South America (Costa Rica, Colombia, Chile). Incidence

Europe

Americas

England, South Thames

Finland

Canada

100 50

100

100

50

50

10

10

5

5

1 .5

50

IncF IncM

IncM IncF

10

5

5

5

5

1

1

1

1

1

.5

.5

.5

.5

.5

100

100

50

50

United States of America

IncM

IncM

IncF

IncF

IncWM IncBM IncBF IncWF MortWM MortBF

IncHM

India, Mumbai (Bombay) 100

100

50

50

Japan, Miyagi Prefecture 100 IncM

50

IncF MortBM

10

10

5

5

5

1

1

1

1

1

.5

.5

.5

.5

.5

10

10

10

5

5

1 .5

1960 1970 1980 1990 2000

50

IncM IncF

MortF

Puerto Rico

IncM

1960 1970 1980 1990 2000

50

100

10

Spain, Zaragoza

IncF

50

China, Shanghai

10

MortF

100

100

MortM

MortM

Slovakia

100

IncF

IncM IncF

10

Australia, New South Wales

Colombia, Cali

IncM

IncM IncF

Oceania/Asia

1960 1970 1980 1990 2000

Figure 8–13. Colorectal cancer incidence and mortality trends. (Source: C15/SEER/NCHS.)

MortWF

1960 1970 1980 1990 2000

10 IncM IncF

1960 1970 1980 1990 2000

5

1960 1970 1980 1990 2000

118

PART II: THE MAGNITUDE OF CANCER

< 5.9

< 9.6

< 14.6

< 23.6

< 70.0

Age standardized (world) rate (per 100,000) Figure 8–14. Incidence of stomach cancer, males all ages.

rates are low (<10/100,000 among men in southern Asia, North and East Africa, North America, and Australia and New Zealand (Fig. 8–14). Contrary to popular belief, the incidence in Central Africa is not low (Parkin et al., 2003). The international differences conceal marked variations within some countries: There is a threefold difference in rates within Italy and even more marked differences within China, for example (Fig. 8–15). The incidence in men is twice that in women in both high and low risk countries, although it has been noted for some time (Griffith, 1968) that age-specific rates in women often exceed those in men in the youngest age groups (< age 40). This may be related to differences in the frequency of different subtypes of adenocarcinomas (intestinal and diffuse) (Lauren, 1965). Diffuse carcinoma tends to affect younger individuals and is relatively more common in females (Correa et al., 1973). Intestinal adenocarcinoma predominates in the high incidence areas (particularly in men and older age groups), and this subtype is responsible for much of the international variation (Muñoz, 1988). Migrant populations from high risk parts of the world show a marked diminution in risk when they move to a lower risk area, although this is gradual and seems to depend on the age at migration (McMichael et al., 1980). The data fit with observations concerning the importance of childhood environment in determining risk (Coggon et al., 1990). Italian migrants to Canada, for example, have a two- to threefold higher risk than native Canadians, but the excess had disappeared in their offspring born in Canada (Balzi et al., 1995). On the other hand, the incidence in some migrant groups (e.g., Japanese) with distinctive lifestyles (including dietary habits) may remain significantly elevated, such as among U.S.-born Japanese (Kamineni et al., 1999).

Stomach: ASR (World) (per 100,000) – Male (All ages)

CHINA

China, Changle China, Cixian China, Jiashan China, Wuhan China, Beijing China, Hong Kong 0

ITALY

50

100

150

Italy, Romagna Italy, Florence Italy, Varese Province

Italy, Ferrara Province Italy, Torino Italy, Ragusa Province Italy, Sassari 0

10

20

30

40

Figure 8–15. Regional variation of incidence of stomach cancer in China and Italy.

International Patterns of Cancer Incidence and Mortality

Time Trends The main epidemiologic feature of gastric cancer is the steady decline observed in most affluent countries over the last 50+ years (Correa and Chen, 1994; Aragones et al., 1997). Time trends in developing countries are less well documented, although stomach cancer risk is declining in most areas where surveillance through cancer registries or mortality statistics is possible (Fig. 8–16). Some studies have shown that declines in intestinal-type adenocarcinoma are responsible for the decreased rates, with rather little change in the incidence of diffusetype carcinomas (Muñoz and Asvall, 1971; Craanen et al., 1992; Pinheiro et al., 1999), whereas others (Ekstrom et al., 1999) found that the time trends were similar for the two types. In contrast to the overall decreasing trend, there has been an increase of cancers localized to the cardia, which is evident in several populations (Powell and McConkey, 1990; Devesa et al., 1998; Laheij et al., 1999). The reasons for this increase are not known. There has been a simultaneous increase in the prevalence of Barrett’s esophagus and adenocarcinoma of the lower third of the esophagus; and it is possible that some or all of the increase in the incidence of cardia cancers represents misclassification of cancers at the gastroesophageal junction (Ekstrom et al., 1999). The constant decline of stomach cancer in more affluent countries has been attributed to improved food preservation practices and better nutrition (more vitamins from fresh vegetables and fruits). The most important change may well be the invention of refrigeration for the transport and storage of food, making salting, smoking, and pickling obsolete, thereby leading to less consumption of salt. Campaigns to reduce salt intake (as a means of controlling hypertension) have probably also had a beneficial effect, such as in Japan. The availability of serum specimens from samples of the population over a lengthy time period allows comparison of infection rates with Helicobacter pylori in successive birth cohorts. There is evidence that, at least in western countries, there is a progressive decline between successive generations, presumably related to steady changes in childhood environment (Banatvala et al., 1993; Kosunen et al., 1997; Roosendaal et al., 1997).

Liver Cancer By 2000, liver cancer was the fifth most common cancer worldwide, responsible for around 551,000 new cases (399,000 in men and 153,000 in women). Because of the poor prognosis, the number of deaths (529,000) is not far short of the number of new cases, and it represents the third most common cause of death from cancer. The geographic distribution of liver cancer is very uneven, with 83% of cases reported from the developing countries. The highest incidence rates are in West and Central Africa (where it accounts for almost one-fourth of all cancers in men), eastern and southeastern Asia, and Melanesia (Fig. 8–17). China alone accounts for 54% of the total cases in the world. The incidence is low in developed countries (except Japan), with the highest rates found in southern Europe, especially in Greece. Quite large variations are observed in ethnic groups within the same country. In Singapore, the incidence in the Chinese population is 2.5 times that in Indians (Parkin et al., 2002), whereas in the United States the highest incidence is observed in Koreans and Chinese, with moderately high rates in persons of Japanese or Filipino ethnicity (Fig. 8–18). The rates are, however, considerably lower than in the countries of origin of these populations. Migrants to France from West and Central Africa have a high death rate from liver cancer (Bouchardy et al., 1995), as do migrants to England and Wales from West Africa, but not East African migrants, who were largely of southern Asian ethnicity (Grulich et al., 1992). Liver cancer comprises a variety of cancers, which show distinct epidemiologic features. The most frequent subtype in most areas is hepatocellular carcinoma (HCC), and much of the geographic variation seen in Figure 8–17 is due to this cancer. Although alcohol, aflatoxin, and other agents are important in its etiology (see Chapter 39), the geographic distribution corresponds closely to that of chronic infection with the hepatitis B virus (HBV), which was estimated to

119

account for 60% of all liver cancer cases worldwide in 1990 (Pisani et al., 1997). Cholangiocarcinoma (CCA), a tumor of the epithelium of the intrahepatic bile ducts, is generally less frequent, comprising around 10%–25% of liver cancers in men in Europe and North America and a rather larger proportion than this in women. This is because the incidence rates for CCA are similar in males and females, whereas HCC rates are about two- to threefold higher in males than females (the sex ratio is even greater in France and Switzerland (4 : 1 or 5 : 1), where alcohol is the major etiologic factor). The incidence of CCA shows little variation worldwide, with rates in males between 0.5 and 2.0 and somewhat lower in females (Parkin et al., 1993), although the incidence in northern Thailand, Korea, and parts of China is high (>4/100,000) and very high in Khon Kaen in northeastern Thailand (around 70/100,000 in men, 30/100,000 in women) (Parkin et al., 2002), due to endemic infection with liver flukes. Other types of liver cancer are much less common. Hepatoblastoma is a tumor of young children, with 80% of cases occurring during the first 5 years of life. There is little geographic variation in incidence. Malignant vascular tumors (hemangiosarcomas) are even more rare and affect principally adults.

Time Trends Time trend studies are particularly difficult for liver cancer. Mortality data may be unreliable because of the variable inclusion of metastatic liver cancers. In the seventh revision of the ICD (used until about 1965), the code for liver cancer included gallbladder cancers. Figure 8–19 shows the incidence trends based on data from CI5 Volumes III–VIII, for the three-digit ICD rubric 155. In the eighth revision it comprises only cancers of the liver and intrahepatic bile ducts specified as primary, whereas in the ICD ninth revision tumors that are unspecified (primary or secondary) are included. Some of the changes observed have been noted in previous reviews of incidence and mortality (La Vecchia et al., 2000; McGlynn et al., 2001); they include the increases in incidence in north America, central and western Europe, and Oceania, as well as Japan. Increasing mortality rates have also been noted in France (Deuffic et al., 1998). A decline in incidence has been seen in Shanghai and Singapore as well as some of the Nordic countries (Sweden, Finland). The declines in incidence in Singapore Chinese and Shanghai match the declining mortality rates in the urban population of China (Yang et al., 2003), and may reflect declines in the prevalence of infection with HBV. It is too soon for vaccination programs to have had an effect on incidence or mortality. In contrast, the rise in liver cancer incidence and mortality in Japan has been noted for some time; it has been ascribed to increasing alcohol consumption (in men) (Makimoto and Higuchi, 1999) and to increasing prevalence of hepatitis C virus (HCV) infection (Tanaka et al., 1991). Transmission of the virus by nonsterile transfusions and injections was maximum during the years after World War II, and the risk of liver cancer in Osaka (which has one of the highest rates in the world) has decreased in successive birth cohorts since around 1931–1935, along with the prevalence of infection with HCV (Tsukuma et al., 1999). In western countries, it is possible that some of the increase in mortality (and incidence) is due to improved detection of small cancers in patients with advanced cirrhosis. Increasing alcohol consumption may be part of the explanation. Although there have been declines in mortality from cirrhosis of the liver in many of the countries experiencing increasing liver cancer mortality, it could be that better management of cirrhotic patients increases the opportunity for them to develop an HCC (La Vecchia et al., 1994). More interest has focused on the possible role of HCV infection, which is likely to become increasingly important, as the generations infected (transfusion recipients and drug users) enter age groups at high risk for liver cancer (de Vos Irvine et al., 1998; El-Serag and Mason, 1999). The incidence of cholangiocarcinoma may also be increasing in some populations (McGlynn et al., 2001; Taylor-Robinson et al., 2001), although because it accounts for only a small proportion of liver cancers significant changes would have to be present to affect the trends for liver cancer as a whole.

Europe

Americas

England, South Thames

Finland

Oceania/Asia

Canada

100

Australia, New South Wales

Colombia, Cali

100 100

100

China, Shanghai

100

100 IncM

50

50

50

50

50

50

IncM IncF IncF IncM

IncM

10

10

10

5

5

IncM

10

10

5

5

10

IncM

IncF

5 IncF

5

IncF

IncF

1

1

1

1

1

1

.5

.5

.5

.5

.5

.5

Spain, Zaragoza

Slovakia

Puerto Rico

100

United States of America

100 100

India, Mumbai (Bombay) 100

Japan, Miyagi Prefecture

100

100 IncM

50

50

50

50

50

50 IncF

IncM IncM IncM

10

IncF

10

10

IncF

5

IncBM IncWM

10

10

IncBF

5

IncF

5

IncWF

10 IncM

5

5

5 IncF

1

1

1

1

1

1

.5

.5

.5

.5

.5

.5

1960 1970 1980 1990 2000

1960 1970 1980 1990 2000

1960 1970 1980 1990 2000

Figure 8–16. Stomach cancer incidence trends. (Source: C15/SEER/NCHS.)

1960 1970 1980 1990 2000

1960 1970 1980 1990 2000

1960 1970 1980 1990 2000

121

International Patterns of Cancer Incidence and Mortality

< 3.3

< 5.6

< 9.0

< 15.1

< 98.9

Age standardized (world) rate (per 100,000) Figure 8–17. Incidence of liver cancer worldwide.

Prostate Cancer Prostate cancer is now the sixth most common cancer in the world (in terms of the number of new cases) and third in importance in men (Parkin, 2001). The estimated number of cases was 513,000 during the year 2000. This represents 9.7% of cancers in men (15.3% in developed countries and 4.3% in developing countries). It is a less prominent cause of death from cancer, with 201,000 deaths (5.6% of cancer deaths in men, 3.2% of all cancer deaths). The low fatality rate means that many men are alive following a diagnosis of prostate cancer: an estimated 1.5 million at 5 years in 2000, making it the most prevalent form of cancer in men. The risk of prostate cancer rises steeply with

Figure 8–18. Incidence of liver cancer in Los Angeles.

age. The incidence of clinical disease is low until after age 45–50 and then increases at approximately the 9th to 10th power of age, compared with the 5th to 6th power for other epithelial cancers (Cook et al., 1969). Worldwide, about three-fourths of all cases occur in men aged 65 or older. Nowadays, incidence rates partly reflect the diagnosis of latent cancers by both screening asymptomatic individuals and detecting latent cancer in tissue removed during prostatectomy operations or at autopsy. Thus, especially where screening is widespread, the recorded “incidence” may be high (in the United States, for example, where it is now by far the most commonly diagnosed cancer in men). The incidence is high also in Australia and the Scandinavian countries (probably also due to screening); on the other hand, the incidence of prostate cancer remains low in Asian populations (Fig. 8–20). Mortality is less affected by the incidence of asymptomatic cancers but depends on survival as well as incidence of invasive cancer; survival at 5 years is significantly greater in high incidence countries (80% in the United States versus 40% in developing countries). However, this more favorable prognosis could well be due to more latent cancer being detected by screening procedures; this would also explain the absence of any change in mortality in the presence of the large increase in incidence (Brawley, 1997). Mortality rates are high in North America, northern and western Europe, Australia/New Zealand, parts of South America (Brazil) and the Caribbean, and in much of sub-Saharan Africa (Fig. 8–21). Mortality rates are low in Asian populations and in North Africa. The difference in mortality between China and the United States is 26-fold (it is almost 90-fold for incidence). These international differences are mirrored by ethnic variation in risk within the United States (Fig. 8–22), where the black population has the highest incidence (and mortality) rates—some 70% higher than in whites—who in turn have rates considerably higher than

Europe

Americas

England, South Thames

Finland

Oceania/Asia

Canada

100

Colombia, Cali

Australia, New South Wales

100 100

100

50

50

50

50

China, Shanghai

100

100 50

50 IncM

10 5

10

10

5

5

10 5 IncM

IncM

IncF

10

10

5

5

IncM IncM

IncM

IncF IncF

IncF

1

IncF

1 .5

.5

Slovakia

1

1

.5

.5

Spain, Zaragoza

.5

India, Mumbai (Bombay)

100 100

100

50

50

50

50

10

10

10

10

1

.5

United States of America

Puerto Rico

100

IncF

1

Japan, Miyagi Prefecture

100

100

50

50

IncM IncM

IncM

5

5

5

IncM

IncBM

5

IncF IncF

IncF

1

1

1

.5

.5

.5

1960 1970 1980 1990 2000

1960 1970 1980 1990 2000

Figure 8–19. Liver cancer incidence trends. (Source: C15/SEER/NCHS.)

IncF

5

IncM IncF

1 .5

1960 1970 1980 1990 2000

10

5

IncWM IncBF IncWF

1960 1970 1980 1990 2000

10

1

1

.5

.5

1960 1970 1980 1990 2000

1960 1970 1980 1990 2000

123

International Patterns of Cancer Incidence and Mortality USA, SEER: Black USA, SEER: White *France, Martinique Australia, New South Wales France, Isere The Netherlands *Uruguay, Montevideo *UK, England *Uganda, Kyadondo County *Zimbabwe, Harare: African Denmark *Ecuador, Quito Slovakia Spain, Granada *Janpan, Osaka Prefecture *China, Hong Kong *Korea, Seoul India, Chennai (Madras) *Thailand, Chiang mai *China, Shanghai 0

50

100

150

200

Figure 8–20. Incidence of prostate cancer worldwide.

populations of Asian origin (e.g., Chinese, Japanese, and Korean males). Similarly, in São Paulo, Brazil, the risk of prostate cancer in black males was 1.8 [95% confidence interval (CI) 1.4–2.3] times that of white men (Bouchardy et al., 1991). Many elderly men are found to harbor latent cancers in their prostate, the prevalence of which greatly exceeds the cumulative incidence in the same population. Two international studies have compared the prevalence of latent prostate cancer at autopsy and the incidence of clinical disease in various populations (Breslow et al., 1977; Yatani et al., 1982). The prevalence of latent cancer increases steeply with

< 4.0

< 8.3

age but shows much less geographic variation than clinical cancer, although the country/ethnic-specific ranking was much the same; the differences between populations was largely due to variations in the prevalence of the infiltrative type of latent cancer. The frequency of latent carcinoma of prostate in Japan is increasing (as for clinical prostate cancer) and approaching the prevalence among U.S. whites. Migrants from West Africa to England and Wales have mortality rates 3.5 times (95% CI 2.4–5.1) those of the local-born population, and mortality is significantly higher also among migrants from the Caribbean [relative risk (RR) 1.7; 95% CI 1.5–2.0]; in contrast, mortality among migrants from East Africa of predominantly Asian (Indian) ethnicity are not high (Grulich et al., 1992). Migrants from low risk countries to areas of higher risk show marked increases in incidence (e.g., Japanese living in the United States). Some of this change reflects elimination of the “diagnostic bias” influencing the international incidence rates. Shimizu et al. (1991) pointed out that localized prostate cancer forms a small proportion of cases in Japan (24%) compared with 66%–70% in the United States and that the incidence in Japan could be three to four times that actually recorded if, for example, all transurethral prostatectomy (TURP) sections were carefully examined. However, rates in Japanese migrants remain well below those in the U.S. white populations, even in Japanese born in the United States, suggesting that genetic factors are responsible for at least some of the differences between ethnic groups.

Time Trends Trends in incidence and mortality are shown in Figure 8–23. In the United States, prostate cancer incidence rates were increasing slowly until the 1980s, probably due to a genuine increase in risk coupled with increasing diagnosis of latent, asymptomatic cancers in prostatectomy specimens owing to the increasing use of TURP (Potosky et al.,

< 12.8

< 17.5

Age standardized (world) rate (per 100,000) Figure 8–21. Mortality rates from prostate cancer worldwide.

< 55.3

124

PART II: THE MAGNITUDE OF CANCER Australia) there has been considerable screening activity, but this is not the case in others, where the decreased mortality is just as marked (France, Germany, Italy, United Kingdom) (Oliver et al., 2001). There may be a contribution from improved treatment, but it is difficult to evaluate from survival data because of the lead-time bias introduced by the earlier diagnosis.

Cancer of the Cervix Uteri Figure 8–22. Prostate cancer incidence in SEER registries.

1990). Beginning in 1986 and accelerating after 1988, there was a rapid increase in incidence (Fig. 8–23), coinciding with the introduction of testing with prostate-specific antigen (PSA), allowing detection of preclinical (asymptomatic) disease (Potosky et al., 1995). The recorded incidence of prostate cancer doubled between 1984 and 1992, with the increase being mainly in younger men (under age 65) and confined to localized and regional disease; there was even a decline in late-stage cancer. The incidence rates began to fall again in 1992 (in 1993 in black men). This probably reflects the fact that, by this time, most of the PSA tests being carried out were repeat examinations and that the supply of prevalent latent cancers in the subset of the population reached by opportunistic screening has been largely exhausted (Brawley, 1997). Prostate cancer mortality rates in the United States had been increasing slowly since the 1970s (Fig. 8–23). With the introduction of PSA screening and the dramatic surge of incidence induced by it, there was an acceleration in the rate of increase in mortality, but this was much less marked than the change in incidence. More recently (since 1992 in white men and since 1994 in black men), mortality rates have decreased. The contribution of PSA screening and/or improved treatment to this decline has been the subject of considerable debate (Etzioni et al., 1999; Feuer et al., 1999; Hankey et al., 1999). The increased mortality is probably partly due to miscertification of the cause of death among the large number of men who had been diagnosed with latent prostate cancer during the late 1980s and early 1990s. The later decline may be partly due to a reversal of this effect; it seems unlikely that screening was entirely responsible. The lead time (between screening detection and usual clinical presentation) would have to be short if screening were to have such a rapid effect on mortality. Similar trends have been reported in Canada (Mercer et al., 1997), the United Kingdom (Chamberlain et al., 1997), France (Chirpaz et al., 2002), Australia (Threlfall et al., 1998), and The Netherlands (Post et al., 1998), although in general they are less marked, or occurred later, than in the United States. International trends in mortality have been reviewed by Oliver et al. (2001) and incidence and mortality by Hsing et al. (2000). The largest increases in incidence, especially among young men, have been seen in high risk countries, probably due partly to the effect of increasing detection following TURP, and, more recently, to use of PSA. There have been large increases also in low risk countries—3.5-fold in Shanghai, 3.0-fold in Singapore Chinese, 2.6-fold in Miyagi (Japan), 1.7-fold in Hong Kong—between 1975 and 1995. Only in India (Bombay) does there seem to have been little change (13%) in incidence (Fig. 8–23). Some of this increase may be due to greater awareness of the disease and the diagnosis of small and latent cancers; but it is also probable that there is a genuine increase in risk. This is confirmed by studying changes in mortality. The increased rates in “high risk” countries were much less than for incidence but were quite substantial nevertheless (15%–25%). In low risk countries, the increase in mortality rates is large and not much less than the changes observed in incidence. Although some of this change in low risk populations may relate to better detection and diagnosis, much of it probably relates to westernization of lifestyles, with increasing obesity, changes in diet (increased consumption of meat and fat), and decreased physical activity. As in the United States, there has been decreased mortality from prostate cancer since around 1988–1991 in several high risk populations, and it was rather more marked in older than in younger men. In some of the countries concerned (Canada,

Cancer of the cervix uteri is the second most common cancer among women worldwide, with an estimated 471,000 new cases and 233,000 deaths during the year 2000. Almost 80% of the cases occur in developing countries, where, in many regions it is the most common cancer in women. The highest incidence rates are observed in Latin America and the Caribbean, Sub-Saharan Africa, and southern and southeastern Asia (Fig. 8–24). Incidence rates are now generally low in developed countries, with age-standardized rates of less than 14 per 100,000. This pattern is relatively recent, however; before the introduction of screening programs during the 1960s and 1970s, the incidence in most of Europe, North America, and Australia/New Zealand was much as we see in developing countries today (Gustafsson et al., 1997a): it was 38.0 per 100,000 in the Second National Cancer Survey of the United States, for example (Dorn and Cutler, 1959). Low rates are also observed in China and western Asia (Fig. 8–24); the lowest recorded rate is 0.4 per 100,000 in Ardabil, in northwestern Iran (Sadjadi et al., 2003). Most cervical cancers are squamous cell carcinomas. Adenocarcinomas are rarer, but the proportion of cases with this histology is higher in low incidence areas than in high risk regions (Fig. 8–25). This is probably the result of screening programs, which are more effective in preventing squamous cell cancers than adenocarcinomas (see below). The incidence of cervical cancer begins to increase at ages 20–29, and the risk increases rapidly to reach a peak usually around age 45–49 in European populations but often later in developing countries. Incidence rates then decline somewhat, although the slope is much less steep than for the increase in young women (Fig. 8–26). This pattern is profoundly changed by screening programs (Gustafsson et al., 1997b). Mortality rates are substantially lower than the incidence. Worldwide, the mortality/incidence ratio is 49%. Survival rates vary among regions, with quite a good prognosis in low risk regions (69% in SEER and 59% in the European registries). Even in developing countries, however, where many cases present at a relatively advanced stage, survival rates are fair: 49% on average (Sankaranarayanan et al., 1998). The poorest survival rate is estimated for eastern Europe. Infection with human papillomavirus is now recognized to be the major determinant of cervical cancer (Walboomers et al., 1999; Bosch et al., 2002), so it is interesting to know whether the geographic patterns for risk of the disease reflect differences in the prevalence of infection. Incidence rates are a good indicator only of underlying risk in the absence of extensive population screening. Moreover, when comparisons are restricted to such populations, there appears to be a broad correlation.

Time Trends Time trends are of considerable interest, partly because of the light that may be shed on changes in exposure to etiologic factors (especially between women of different generations) and partly as a means of evaluating the success, or otherwise, of screening programs. Mortality data are often used in studies of time trends, although care is needed when they are used because the proportion of deaths recorded as “uterus, unspecified” varies greatly among populations and over time. Mortality is also influenced by changes in survival, which may be marked if long-term series are studied (Ponten et al., 1995). Overall, incidence and mortality have declined during the last 40 years in western Europe, the United States, Canada, Australia, New Zealand, and Japan (Fig. 8–27). In general, this has been ascribed to the combination of a reduction in risk in older generations of women (e.g., genital hygiene, parity) with, more recently, the beneficial effects

Europe

Americas

England, South Thames

Finland

Oceania/Asia

Canada

Colombia, Cali

Australia, New South Wales

200 200

200 100

100 100

IncM

China, Shanghai

200

200

100

100

50

50

50

200 100

IncM

IncM

50

50

IncM

50

IncM

MortM MortM

10

10

10

10

10

10

5

5

5

5

5

5 IncM

1

1

1

1

1

1

.5

.5

.5

.5

.5

.5

Slovakia

Spain, Zaragoza

Puerto Rico

200

200 200

100

100 100

50

50 IncM

United States of America

IncM

50

IncM

IncBM IncWM

India, Mumbai (Bombay) 200

200

200

100

100

100

50

50

50

10

10

MortBM MortWM

10

Japan, Miyagi Prefecture

10

10

IncM

10

IncM

5

5

5

5

5

5

1

1

1

1

1

1

.5

.5

.5

.5

.5

.5

1960 1970 1980 1990 2000

1960 1970 1980 1990 2000

1960 1970 1980 1990 2000

Figure 8–23. Prostate cancer incidence and mortality trends. (Source: C15/SEER/NCHS.)

1960 1970 1980 1990 2000

1960 1970 1980 1990 2000

1960 1970 1980 1990 2000

126

PART II: THE MAGNITUDE OF CANCER

< 9.3

< 16.1

< 23.8

< 35.8

< 93.9

Age standardized (world) rate (per 100,000) Figure 8–24. Incidence of cancer of the cervix uteri worldwide.

of population screening programs based on exfoliative cervical cytology. In the Nordic countries, national incidence data have been available for several decades, and trends in incidence in the various countries has been compared with national policies in relation to screening (Hakama, 1982; Hakulinen et al., 1986). The decline in inci-

0

10

20

30

40

50 44.5

Uganda, Kyadondo County

43.7

Brazil, Goiania 32.3

Viet Nam, Ho Chi Minh City 28.6

Thailand, Chiang Mai

28.0

Ecuador, Quito India, Bangalore

140.0 Rate [per 100,000]

24.9

USA, California, LA: Hispanic White

100.0

23.1

Slovakia

18.7

Singapore: Chinese

50.0

17.4

Poland, Lower Silesia

16.8

Estonia

16.0

Denmark

15.1

Algeria, Algiers

10.0

13.6 12.4

UK, Scotland USA, SEER: Black

Uganda, Kyadondo County India Algeria, Algiers Slovakia Spain Israel

5.0

11.5

Australia, New South Wales

10.3 9.0

Canada Switzerland, Zurich

1.0

8.7

USA, SEER: White

8.4

The Netherlands

0.5

7.9

Italy, Varese Province Spain, Navarra

60 56.9

Zimbabwe, Harare: African

dence is related to the coverage and extent of the organized programs (Sigurdsson, 1999) and is most marked in the age groups targeted by these programs. Nevertheless, even in some of the countries showing declines in overall (crude or age-adjusted) incidence and/or mortality, increases are seen among young women. This was first described in England and Wales, where successive generations of women born since about 1935 were at increasingly high risk (Hill and Adelstein, 1967; Cook and Draper, 1984; Parkin et al., 1985). Similar observations were made in Australia (Armstrong and Holman, 1981), New Zealand (Cox and Skegg, 1986), Belgium (Vyslouzilova et al., 1997), Slovenia

6.9 4.4

China, Beijing

2.7

Israel, non-Jews

2.5

0.1 Squamous

Adenocarcinoma

Other

Figure 8–25. Age standardized (world standard) incidence per 100,000, of cancer of the cervix, by histological type in 25 cancer registries. (Source: Perkins et al., 2002.)

0.05

15- 20- 25- 30- 35- 40- 45- 50- 55- 60- 65- 70- 75+Age

Figure 8–26. Age-specific incidence rates of cervical cancer.

Europe

Americas

England, South Thames

Finland

Oceania/Asia

Canada

100

Australia, New South Wales

Colombia, Cali

100 100

50

50

50

China, Shanghai

100

100

50

50

10

10

5

5

5

100 50

IncF

IncF

10

10

10 IncF

IncF

5

IncF

5

5

10

IncF

MortF MortF

1

1

1

1

1

1

.5

.5

.5

.5

.5

.5

Slovakia

Spain, Zaragoza

Puerto Rico

100 50

United States of America

India, Mumbai (Bombay)

Japan, Miyagi Prefecture

100 100

100

100

100

50

50

50

50

50 IncBF

IncF

10

10

10

5

5

1

1

.5

.5

IncF

5

19601970 1980 1990 2000

IncF

MortBF

19601970 1980 1990 2000

IncWF

10

10

5

5

1

1

1

1

.5

.5

.5

.5

IncF MortWF

19601970 1980 1990 2000

Figure 8–27. Cervical cancer incidence and mortality trends. (Source: C15/SEER/NCHS.)

19601970 1980 1990 2000

10 IncF

19601970 1980 1990 2000

19601970 1980 1990 2000

5

128

PART II: THE MAGNITUDE OF CANCER

(Pompe-Kirn et al., 1992), Slovakia (Vlasak et al., 1991), Spain (Llorca et al., 1999), and several countries of eastern Europe (Beral et al., 1994). Even Finland, with its remarkably successful screening program, which had reduced the incidence of cervical cancer in 1991 to 2.8 per 100,000, has observed marked increases in incidence among younger women (<55 years of age) since 1990 (Anttila et al., 1999). The consensus is that these trends are most likely due to changes in sexual habits and increased transmission of papillomavirus in younger generations of women, but that the magnitude of the effect will depend on the countervailing effects of screening. Thus in some countries (e.g. Sweden), there has been no increase in risk among young women (Bergstrom et al., 1998), and the upward trend in England and Wales has been successfully countered by a much improved screening program that was implemented in 1988 (Quinn et al., 1999). Analyses of time trends by histologic subtype show that trends for squamous cell carcinoma are more or less those observed for cervical cancer as a whole. The large international study of Vizcaino et al. (2000) of 25 countries found a decreased incidence for younger (25–49 years) and older (50–74 years) women in most countries. Exceptions were the increases in young women in the United Kingdom, Slovenia, Slovakia, and Israel. With respect to adenocarcinomas, several studies have shown rising incidence rates in populations where (presumably as a result of screening) incidence rates from squamous cell carcinomas are declining (Kjaer and Brinton, 1993; Bergstrom et al., 1998). The increasing risk of adenocarcinoma appears to affect relatively recent generations of women from many countries (Vizcaino et al., 1998). The cytologic detection of adenocarcinoma or precursor lesions is undoubtedly less efficient than for squamous cell tumors (Fu et al., 1987; Sigurdsson, 1995); and a case-control study (Mitchell et al., 1995) has shown that the risk of adenocarcinoma is not reduced by screening. The increasing incidence may reflect increases in exposure to human papillomavirus in recent generations, the effect of which on squamous cell tumors has been diminished by screening programs. The use of oral contraceptives has also been linked to an increased risk of cervical adenocarcinoma (Ursin et al., 1994a, 1994b). There is less information on time trends in cervical cancer in developing countries; as might be expected, the situation is varied. In general terms, incidence and mortality rates have been relatively stable or have shown only modest declines (Fig. 8–27). This probably reflects the absence of a systematic screening program or, where they have been introduced, low population coverage and poor quality cytology (Lazcano-Ponce et al., 1998). In Cuba (Fernandez Garrote et al., 1996) and Costa Rica (Herrero et al., 1992), for example, the screening programs seem to have had virtually no impact on the incidence of cancer. In contrast, there appear to have been dramatic declines in cervical cancer in China. The age-adjusted incidence of cervical cancer in Shanghai fell from 26.7 to 2.5 per 100,000 between 1972–1974 and 1993–1994 (Jin et al., 1999), and mortality rates have fallen dramatically, especially in urban populations, although the trend has reversed recently in younger women (Yang et al., 2003). The declines have been attributed to Papanicolaou smear screening, treatment programs, and improved genital hygiene; the increased rates among younger women may reflect changing economic circumstances and sexual mores, with a greater prevalence of infection with human papillomavirus and other agents (Li et al., 2000). The limited data available from Africa do not suggest any decrease in incidence of cervical cancer (Parkin et al., 2003).

Cancer of the Esophagus About 391,000 cases of cancer of the esophagus occurred in 2000, of which more than 80% were in developing countries. Because of the poor prognosis, the number of deaths (355,000 per year) is not much less; and geographic patterns and trends in occurrence have frequently been studied in terms of mortality as well as incidence. The geographic variability in risk is large—more than for almost any other cancer. The highest risk areas of the world are in the Asian “esophageal cancer belt” (stretching from northern Iran through the central Asian republics to north-central China). The incidence in Cixian, China, during 1993–1997 was 184 per 100,000 in men and 123 per 100,000 in

women. The rates recorded in Gonbad (northeastern Iran) during the 1960s were 109 per 100,000 in men and 175 per 100,000 in women (Mahboubi et al., 1973), and they remain high in this area today (Saidi et al., 2000). High rates are also present in parts of eastern and southeastern Africa (e.g., eastern Kenya, Zimbabwe, Transkei) (Parkin et al., 2003), parts of eastern South America (southern Brazil, Uruguay, Paraguay, northern Argentina), and certain parts of western Europe (especially France) (Fig. 8–28). For women, the pattern is much the same, with the Indian subcontinent added to the high ranking areas. Esophageal cancer is more common in males in most areas; the sex ratio is 6.5 : 1.0 in France, for example (Ferlay et al., 2001), although in the high risk areas of Asia the sex ratio is much closer to unity (e.g., 1.5 in Linxian County, Henan, China) (Lu et al., 1985). Even in the high risk areas, there are striking local variations in risk. For example, within the esophageal cancer belt the Chinese counties with the highest rates are located in the central/north provinces of Shanxi and Henan, and in central Asia the high risk areas are in parts of Turkmenistan (in particular) and Kazakhstan. In northern Iran, there is quite a dramatic difference as one passes east to west of the Caspian littoral (Muñoz and Day, 1996). Other workers have demonstrated large geographic variations within the high risk areas of South Africa (Rose and McGlashan, 1975) and northern France (Tuyns and Masse, 1973, 1975). Worldwide, most esophageal cancers are squamous cell carcinomas arising in the middle and lower third of the esophagus. Recently, in Western countries there appears to be an increase in relative and absolute numbers of adenocarcinomas of the lower third of the esophagus, associated with Barrett’s esophagus (see below). The profile of genetic changes (mutations) is different in these two histologic subtypes, implying different etiologies. It seems highly unlikely that the dramatic differences in risk observed between populations are related to genetic predisposition. For a start, there can be marked variation within limited geographic areas, as noted above. Migrant studies confirm that persons from high risk areas rapidly lose their elevated risk status after migration. For example, In Singapore Chinese, the rates for the population born in China tend to reflect their region of origin (highest in Teochew and Hokkien and lower in the Cantonese from southern China), whereas incidence rates in the locally born are much lower (Lee et al., 1992). Similar observations have been made with respect to migrants to Israel from Asian countries: The risk declined quite rapidly with duration of stay in Israel (Parkin et al., 1990). On the other hand, the relatively high risk associated with polymorphisms of two genes controlling the alcohol-metabolizing enzymes, alcohol dehydrogenase 2 (ADH2) and aldehyde dehydrogenase 2 (Yokoyama et al., 2002), which are notably frequent in populations in eastern and southeastern Asia (Goedde et al., 1992), may explain the rather high rates of esophageal cancer in Japan when compared with per-capita intake of alcohol, as well as the observation that the incidence among Japanese in the United States is higher than in the white population (Miller et al., 1996).

Time Trends Trends in esophageal carcinoma by geographic location are variable (Fig. 8–29). Recent studies have distinguished between trends in the two main histologic types, squamous cell carcinomas and adenocarcinomas, although the distinction between adenocarcinomas located at the junction of the esophagus and those of the gastric cardia is not always clear (Ekstrom et al., 1999). It is important to ensure that increases in one are not simply due to decreases in the other. The incidence of squamous cell tumors is generally stable or decreasing, although increases are observed in men in the Netherlands and Denmark and in women in Canada, Scotland, and Australia (Vizcaino et al., 2002). Increases in the incidence of adenocarcinomas are larger and more generalized in the United States (whites and blacks), Canada, Australia, Denmark, England, Finland, France, Norway, Scotland, Sweden. By the mid-1990s, the incidence of adenocarcinoma had surpassed that of squamous cell carcinoma in males in the United States (whites), Canada, Australia, New Zealand, and several countries of western Europe (Denmark, Ireland, The Netherlands, United Kingdom), but the rates in women remain considerably lower than in

129

International Patterns of Cancer Incidence and Mortality

< 7.7

< 15.5

< 23.2

< 30.9

< 38.7

Age standardized (world) rate (per 100,000) Figure 8–28. Incidence of esophageal cancer worldwide in males.

men, and adenocarcinoma in women is not more common than squamous cell carcinoma (Parkin et al., 2002). Because both squamous cell carcinomas and adenocarcinomas are related to alcohol and tobacco abuse (see Chapter 36), the differential trends must be due to a different cause. The most likely explanation for the increased incidence of adenocarcinoma seems to be the increasing prevalence of Barrett’s esophagus, as a consequence of gastroesophageal reflux (GERD), which is becoming more common with increasing levels of obesity (Chow et al., 1998a; Lagergren et al., 1999). It is possible that the declining prevalence of infection with Helicobacter pylori may also be increasing the risk of esophageal adenocarcinoma (Chow et al., 1998b). In China, where the risk is relatively high, substantial decreases in incidence have been reported, particularly in younger age groups (Yang et al., 2003). Among Chinese migrants in Singapore, the once high risk of esophageal cancer had dramatically fallen by the mid1980s, with cohort trends suggesting that the cancer would become increasingly rare (Lee, 1988). In the high risk area of Linxian, China, there appears to have been a decline in mortality among those less than 60 years of age since about 1970 (Lu et al., 1985). This pattern suggests sudden dietary improvement. In Japan, the patterns are quite different in men and women. Women show a decline in risk in most age groups, whereas in men there has been a rise in mortality in generations born since about 1920, which parallels changes in death from cirrhosis (Parrish et al., 1993), suggesting that rising alcohol consumption is important. Japanese (and possibly other Asian populations) may have increased susceptibility to alcohol carcinogenicity, as noted above. In Latin America and the Caribbean, where rates range from moderately high to high, declining trends in incidence have been reported in most countries (Coleman et al., 1993).

Bladder Cancer An estimated 340,000 bladder cancers occurred in 2000, when it was the ninth most common cause of cancer for both sexes combined. It is relatively common in developed countries, ranking sixth (210,000 cases), with high rates in North America and Europe, where 40% of all incident cases occur. Most (70%) bladder tumors occur in men. The utility of geographic comparisons of incidence and survival are limited by the well known variations in practices concerning cystoscopy, biopsy of lesions, the extent of the histologic examination of biopsy material, and the classification of papillomas and noninvasive tumors (Saxen, 1982; Kiemeney et al., 1994). The 10-fold variation in international incidence is not particularly striking, however, relative to other cancers. The number of deaths was around 130,000, with population-based 5-year survival rates ranging from 50% to 80% depending on whether noninvasive lesions are included in the computation. Bladder cancer incidence is high in many southern and eastern European countries, where smoking (in men) has been prevalent. It is also high in parts of Africa and the Middle East, where bladder cancer, particularly of the squamous cell type, is linked to chronic infection with Schistosoma haematobium. The highest recorded incidence rate is found in Egypt, where the estimated world-standardized rate of 45 is 60% higher than that of its nearest counterpart (Israel). In the United States, the incidence in whites is higher than in blacks: about double among men and 50% greater among women. It is unlikely that this is due to differences in exposure to environmental carcinogens, and explanations based on differential susceptibility have been proposed. Certainly, migrants to France from Algeria and West Africa (both relatively high-risk populations) appear to have rates higher than the local-born population of France (Bouchardy et al., 1995, 1996).

Europe

Americas

England, South Thames

Finland

Oceania/Asia

Canada

100 50

Australia, New South Wales

Colombia, Cali

100 100

100

100

50

50

50

50

China, Shanghai 100 50

IncM

IncF

10

IncM

5 IncF

10

10

5

5

IncM

IncM

10

10

5

5

10 5

IncM

IncM IncF

IncF

1 .5

Slovakia

1

1

.5

.5

Spain, Zaragoza

IncF

Puerto Rico

100 50

IncF

1

1

1

.5

.5

.5

United States of America

India, Mumbai (Bombay)

100 100

100

100

100

50

50

50

50

10

10

5

5

50

IncBM

10

10

IncM

10

IncM

IncM IncM

5

Japan, Miyagi Prefecture

5

5

IncWM

10

IncM IncF

5

IncBF IncF IncWF

1 .5

IncF

19601970 1980 1990 2000

IncF

19601970 1980 1990 2000

1

1

.5

.5

IncF

19601970 1980 1990 2000

Figure 8–29. Esophageal cancer incidence trends. (Source: C15/SEER/NCHS.)

19601970 1980 1990 2000

1

1

1

.5

.5

.5

19601970 1980 1990 2000

19601970 1980 1990 2000

Europe

Americas

England, South Thames

Finland

Oceania/Asia

Canada

100 50

Australia, New South Wales

Colombia, Cali

China, Shanghai

100 100

100

100

100

50

50

50

50

10

10

50

IncM

IncM

IncM

10

10

IncM

10

10 IncM

5

MortM

IncF

5

IncF MortM

5

IncM

5

5

5 IncF

IncF IncF

IncF

MortF MortF

1 .5

Slovakia

1

1

1

1

1

.5

.5

.5

.5

.5

Spain, Zaragoza

Puerto Rico

100 50

United States of America

India, Mumbai (Bombay)

Japan, Miyagi Prefecture

100 100

100

100

100

50

50

50

50

10

10

5

5

50

IncM IncWM IncM

10

10

10

5

5

5

IncF

IncM

IncBM IncWF MortWM MortBM

IncBF

IncM

10 5

IncM

IncF IncF

MortBF

IncF

MortWF

1

1

1

1

1

.5

.5

.5

.5

.5

19601970 1980 1990 2000

19601970 1980 1990 2000

19601970 1980 1990 2000

Figure 8–30. Bladder cancer incidence and mortality. (Source: C15/SEER/NCHS.)

19601970 1980 1990 2000

1

IncF

.5

19601970 1980 1990 2000

19601970 1980 1990 2000

Europe

Americas

England, South Thames

Finland

Canada

100 50

10

IncF MortM

IncF

5

MortF

Australia, New South Wales

Colombia, Cali

China, Shanghai

100 100

100

100

100

50

50

50

50

50

IncM

IncM

IncM

IncM

Oceania/Asia

10

10

IncF IncM IncF

MortM

5

5

MortF

10

10

5

5

IncF

10 5 IncM IncF

1

1

1

1

1

1

.5

.5

.5

.5

.5

.5

Slovakia

Spain, Zaragoza

Puerto Rico

100 50

10 5

IncM IncF IncM IncF

United States of America

India, Mumbai (Bombay)

Japan, Miyagi Prefecture

100 100

100

100

100

50

50

50

50

50

10

10

10

10

5

5

IncM IncF

IncWM IncWF IncBM MortWM MortBM MortWF

IncBF

10 IncM

5

5

IncM IncF

5 IncF

MortBF

1

1

1

1

1

1

.5

.5

.5

.5

.5

.5

19601970 1980 1990 2000

19601970 1980 1990 2000

19601970 1980 1990 2000

Figure 8–31. Non-Hodgkin lymphoma incidence and mortality trends. (Source: C15/SEER/NCHS.)

19601970 1980 1990 2000

19601970 1980 1990 2000

19601970 1980 1990 2000

International Patterns of Cancer Incidence and Mortality Genetic polymorphisms of metabolic enzymes such as N-acetyltransferase (NAT) and glutathione S-transferase 1 (GSTM1) may play a role (Yu et al., 1994, 1995). The most common histologic type of bladder cancer in industrialized countries is transitional cell carcinoma (TCC), which comprises, for example, 90% of cancers in England and Wales and 95% in the Netherlands and France (Parkin et al., 2002). In the United States, the differences in histology by race are not particularly striking. whites have about 95% TCCs and just over 1% squamous cell carcinomas (SCCs); the proportions are about 89% and 3%, respectively, in blacks.

Time Trends Incidence rates have increased in the United States, Japan, and European countries in recent decades for both sexes (Coleman et al., 1993; Thorn et al., 1997; Swerdlow et al., 1998), although there is evidence of some stabilization or even a decrease during the 1990s (Fig. 8–30). Although increased levels of diagnostic activity may partially explain the time trend, it is unlikely that the increases observed in many Western countries can be attributed simply to artifact. Mortality rates show rather different patterns: declining trends in men in northern and western Europe during the 1960s and 1970s, reaching a plateau in southern Europe but still increasing in northern Europe (Coleman et al., 1993). In the United States and Canada, trends have been downward since the 1960s (Coleman et al., 1993). In developing countries, the incidence trends are more disparate: gradual increases have been observed in India and recent relative stability or decreases seen in Puerto Rico and Colombia (Fig. 8–30). A decrease in incidence has been noted in Uganda (Wabinga et al., 2000). More than 40% of bladder cancers can be attributed to smoking, and indeed the trends are consistent with those of other tobacco-related neoplasms (lung, kidney, pancreas): mortality rates tend to be highest in birth cohorts with the maximum exposure to tobacco and certain occupational chemicals, with declines evident in more recent decades (La Vecchia and Airoldi, 1999).

Non-Hodgkin Lymphoma The 290,000 cases of non-Hodgkin lymphoma (NHL) that occurred in 2000 (3% of all cancers) comprise an extremely heterogeneous group of lymphoproliferative malignancies that display distinct behavioral, prognostic, and epidemiologic characteristics. Advances in molecular biology, genetics, and immunology have resulted in extensive changes in the classification of lymphoid tumors during the last few decades. The Revised European-American Lymphoma (REAL) classification system (Harris et al., 1994) and its successor, the WHO classification (Jaffe et al., 2001), are the result of a consensus that links two previous systems (the Kiel classification in Europe and the Working Formulation in the United States) that distinguished between lymphomas and leukemias. It acknowledges that some solid tumors also pass through circulating leukemic phases. The REAL classification seeks to classify tumors according to cell lineage defined by immunophenotype (Herrinton, 1998). Three broad categories are now recognized: B-cell neoplasms, T/NK-cell neoplasms, and Hodgkin lymphoma. Lymphocytic leukemias fall within the B-cell neoplasm group. NHL is slightly more common in developed countries (51% of the world’s total of cases), with rates highest in Australia and North America, intermediate in Europe (except eastern areas) and South America (elevated in Bolivia and Peru) and relatively low throughout Asia and eastern Europe (Ferlay et al., 2001). NHL is relatively rare in most African populations, but the relative frequency is above the world average in North and sub-Saharan Africa because of the high incidence of Burkitt lymphoma (BL) in children in the tropical zone of Africa. Approximately 5%–10% of human immunodeficiency virus (HIV)-infected persons develop a lymphoma, and NHL is the AIDSdefining illness in about 3% of HIV-infected patients (Remick, 1995). B-cell lymphomas tend to dominate in most parts of the world, although peripheral T-cell tumors comprise most NHLs in eastern Asia and the Caribbean (Melbye and Trichopoulos, 2002). In the United States, blacks have a higher incidence of T-cell lymphomas than

133

whites (Groves et al., 2000). In a recent study, adult T-cell leukemia/ lymphoma (ATL) accounted for more than half of the total NHL incidence in Nagasaki, Japan (Arisawa et al., 2000).

Time Trends There are clear increases in the incidence of NHL in many parts of the world (Fig. 8–31). Although this may be due in part to improved diagnostic procedures and changes in classification, there can be little doubt that much of the change is real (Hartge and Devesa, 1992), and the reasons for it have been the subject of much debate. The increase has been seen in both sexes across Europe since the 1960s (Coleman et al., 1993). More detailed analyses indicate the presence of birth cohort effects over and above the ubiquitous diagnostic-led period effects (Pollán et al., 1998). Increases in incidence rates of about 1%–2% per year in both sexes by period of diagnosis are seen in Australia and at a lower level in South America and Asia (Fig. 8–31). In the United States, the rapid rises, particularly among younger men, may be partially attributable to the onset of the AIDS epidemic in 1981; and the declines during the 1990s may be due in part to a decrease in the incidence of HIV infection and successful antiretroviral therapies (Eltom et al., 2002). AIDS cannot, however, account for all of the observed increase, as subtypes not associated with AIDS are continuing to increase. As in Europe, the rise in NHL in the United States (Connecticut) has been attributed to both secular and cohort influences for both sexes (Holford et al., 1992). Mortality rates have, in general, been increasing, albeit at a slower pace than the incidence (Coleman et al., 1993), probably because survival is improving. In some populations, mortality rates have reached a plateau or are beginning to decline: in the United States mortality has decreased in parallel with incidence (Fig. 8–31). Few studies have examined temporal variations in NHL according to subtype. Both nodal and extranodal disease (occurring at a ratio of 3 : 1) have increased in the United States, with the most rapid increases seen in extranodal disease (Devesa and Fears, 1992). Similarly, both diffuse and nodular tumors are increasing, particularly in the former, more common subtype. As noted, the recent stabilization of incidence in the United States is ascribed to diminishing rates of the AIDSassociated histologic subtypes, particularly immunoblastic NHL (Eltom et al., 2002). Groves et al. (2000) observed that high-grade NHL has been increasing rapidly among U.S. males, and the increase in follicular NHL was faster among blacks than among other races. More studies that address temporal and spatial analyses according to immunophenotype classification are needed. One such study linked the predominance of the ATL subtype to the prevalence of human T-cell lymphotropic virus type I in southwestern Japan (Arisawa et al., 2000). Why are the onset rates of NHL escalating at a pace unparalleled by most other cancers? Other than AIDS, established risk factors such as those related to disorders of the immune system (transplant patients, autoimmunity, congenital immunodeficiency) are not likely to explain more than a fraction of the observed incidence. The Epstein-Barr virus plays an important role in NHL etiology in persons with inherited or acquired immunosuppression (Mueller et al., 1992); but as a risk factor among immunocompetent persons, the evidence is not convincing (IARC, 1997). The biologic evidence that ultraviolet exposure from sunlight can result in immune modulation and increase NHL risk has been considered credible, but results from the epidemiologic studies have been conflicting (Melbye and Trichopoulos, 2002). Two recent studies have reported that sun exposure may in fact protect against NHL, with production of vitamin D hypothesised as a plausible mechanism for the inverse association (Hughes et al., 2004; Smedby et al., 2005). References Ahmad OE, Boschi-Pinto C, Lopez AD, Murray CJL, Lozano R, Inoue M. 2000. Age Standardization of Rates: A New WHO Standard. GPE Discussion Paper Series: No. 31. Geneva: World Health Organization. Anttila A, Pukkala E, Soderman B, Kallio M, Nieminen P, Hakama M. 1999. Effect of organised screening on cervical cancer incidence and mortality in Finland, 1963–1995: recent increase in cervical cancer incidence. Int J Cancer 83:59–65.

134

PART II: THE MAGNITUDE OF CANCER

Aragones N, Pollan M, Rodero I, Lopez-Abente G. 1997. Gastric cancer in the European Union (1968–1992): mortality trends and cohort effect. Ann Epidemiol 7:294–303. Arisawa K, Soda M, Endo S, et al. 2000. Evaluation of adult T-cell leukemia/lymphoma incidence and its impact on non-Hodgkin lymphoma incidence in southwestern Japan. Int J Cancer 85:319–324. Armstrong B, Holman D. 1981. Increasing mortality from cancer of the cervix in young Australian women. Med J Aust 1 1:460–462. Balzi D, Geddes M, Brancker A, Parkin DM. 1995. Cancer mortality in Italian migrants and their offspring in Canada. Cancer Causes Control 6:68–74. Banatvala N, Mayo K, Mégraud F, Jennings R, Deeks JJ, Feldman RA. 1993. The cohort effect and Helicobacter pylori. J Infect Dis 168:219–221. Barrett JC. 1978. The redundant factor method and bladder cancer mortality. J Epidemiol Community Health 32:314–316. Bassett MT, Chokunonga E, Mauchaza B, Levy L, Ferlay J, Parkin DM. 1995. Cancer in the African population of Harare, Zimbabwe, 1990–1992. Int J Cancer 63:29–36. Beral V, Hermon C, Muñoz N, Devesa SS. 1994. Cervical cancer. Cancer Surv 19–20:265–285. Bergström A, Pisani P, Wolk A, Adami HO. 2001. Overweight as an avoidable cause of cancer in Europe. Int J Cancer 91:421–430. Bergstrom R, Sparen P, Adami HO. 1998. Trends in cancer of the cervix uteri in Sweden following cytological screening. Br J Cancer 81:159–166. Berrino F, Capocaccia R, Estève J, et al. 1999. Survival of Cancer Patients in Europe: The EUROCARE-2 Study. IARC Scientific Publ. 151. Lyon: IARC. Blanks RG, Moss SM, McGahan CE, Quinn MJ, Babb PJ. 2000. Effect of NHS breast screening programme on mortality from breast cancer in England and Wales, 1990–8: comparison of observed with predicted mortality. BMJ 321:665–669. Blot WJ, Fraumeni JF Jr. 1987. Time-related factors in cancer epidemiology. J Chronic Dis 40:1S–8S. Borras JM, Fernandez E, Gonzalez JR, et al. 2003. Lung cancer mortality in European regions (1955–1997). Ann Oncol 14:159–161. Bosch FX, Lorincz A, Munoz N, Meijer CJ, Shah KV. 2002. The causal relation between human papillomavirus and cervical cancer. J Clin Pathol 55:244–265. Bouchardy C, Mirra AP, Khlat M, Parkin DM, de Souza JM, Gotlieb SL. 1991. Ethnicity and cancer risk in São Paulo, Brazil. Cancer Epidemiol Biomarkers Prev 1:21–27. Bouchardy C, Parkin DM, Wanner P, Khlat M. 1996. Cancer mortality among north African migrants in France. Int J Epidemiol 25:5–13. Bouchardy C, Wanner P, Parkin DM. 1995. Cancer mortality among subSaharan African migrants in France. Cancer Causes Control 6:539–544. Brawley OW. 1997. Prostate carcinoma incidence and patient mortality: the effects of screening and early detection. Cancer 80:1857–1863. Bray F, Guilloux A, Sankila R, Parkin DM. 2002a. Practical implications of imposing a new world standard population. Cancer Causes Control 13:175–182. Bray F, Sankila R, Ferlay J, Parkin DM. 2002b. Estimates of cancer incidence and mortality in Europe in 1995. Eur J Cancer 38:99–166. Bray F, Tyczynski JE, Parkin DM. 2004. Going up or coming down? The changing phases of the lung cancer epidemic from 1967 to 1999 in the 15 European Union countries. Eur J Cancer 40:96–125. Breslow N, Chan CW, Dhom G, et al. 1977. Latent carcinoma of prostate at autopsy in seven areas. Int J Cancer 20:680–688. Breslow NE, Day NE. 1987. Statistical methods in cancer research. Vol. II. The design and analysis of cohort studies. IARC Scientific Publ. No. 82, Lyon: IARC, pp. 1–406. Brown CC, Kessler LG. 1988. Projections of lung cancer mortality in the United States: 1985–2025. J Natl Cancer Inst 80:43–51. Capocaccia R, Verdecchia A, Micheli A, Sant M, Gatta G, Berrino F. 1990. Breast cancer incidence and prevalence estimated from survival and mortality. Cancer Causes Control 1:23–29. Chamberlain J, Melia J, Moss S, Brown J. 1997. The diagnosis, management, treatment and costs of prostate cancer in England and Wales. Health Technol Assess 1:1–53. Charloux A, Quoix E, Wolkove N, Small D, Pauli G, Kreisman H. 1997. The increasing incidence of lung adenocarcinoma: reality or artefact? A review of the epidemiology of lung adenocarcinoma. Int J Epidemiol 26:14–23. Chirpaz E, Colonna M, Ménégoz F, et al. 2002. Incidence and mortality trends for prostate cancer in 5 French areas from 1982 to 1996. Int J Cancer 97:372–376. Chow WH, Blaser MJ, Blot WJ, et al. 1998a. An inverse relation between Cag A+ strains of Helicobacter pylori and risk of esophageal and gastric cardia adenocarcinoma. Cancer Res 58:588–590.

Chow WH, Blot WJ, Vaughan TL, et al. 1998b. Body mass index and risk of adenocarcinomas of the esophagus and gastric cardia. J Natl Cancer Inst 90:150–155. Chow WH, Devesa SS, Blot WJ. 1991. Colon cancer incidence: recent trends in the United States. Cancer Causes Control 2:419–425. Chu KC, Tarone RE, Chow WH, Hankey BF, Gloeckler-Ries LA. 1994. Temporal patterns in colorectal cancer incidence, survival, and mortality from 1950 through 1990. J Natl Cancer Inst 86:997–1006. Chu KC, Tarone RE, Kessler LG, et al. 1996. Recent trends in United States breast cancer incidence, survival, and mortality rates. J Natl Cancer Inst 88:1571–1579. Clayton D, Schifflers E. 1987a. Models for temporal variation in cancer rates. I. Age-period and age-cohort models. Stat Med 6:449–467. Clayton D, Schifflers E. 1987b. Models for temporal variation in cancer rates. II. Age–period–cohort models. Stat Med 6:469–481. Coggon D, Osmond C, Barker DJ. 1990. Stomach cancer and migration within England and Wales. Br J Cancer 61:573–574. Coldman AJ, McBride M, Braun T. 1992. Calculating the prevalence of cancer. Stat Med 11:1579–1589. Coleman MP, Esteve J, Damiecki P, Arslan A, Renard H. 1993. Trends in cancer incidence and mortality. IARC Scientific publ. No. 121. Lyon: IARC. Cook GA, Draper GJ. 1984. Trends in cervical cancer and carcinoma in situ in Great Britain. Br J Cancer 50:367–375. Cook PJ, Doll R, Fellingham SA. 1969. A mathematical model for the age distribution of cancer in man. Int J Cancer 4:93–112. Correa P, Chen VW. 1994. Gastric cancer. Cancer Surv 19–20:55–76. Correa P, Sasano N, Stemmerman N, Haenszel W. 1973. Pathology of gastric carcinoma in Japanese populations: comparisons between Miyagi prefecture, Japan, and Hawaii. J Natl Cancer Inst 51:1449–1459. Cox B, Skegg DC. 1986. Trends in cervical cancer in New Zealand. N Z Med J 99:795–798. Craanen ME, Dekker W, Blok P, Ferwerda J, Tytgat GN. 1992. Time trends in gastric carcinoma: changing patterns of type and location. Am J Gastroenterol 87:572–579. Das Gupta, et al. 1972. Guidelines for Project Evaluation. New York: United Nations. Day NE, Charnay N. 1982. Time trends, cohort effects, and aging as influence on cancer incidence In: Magnus K, editor. Trends in Cancer Incidence: Causes and Practical Implications. Washington, DC: Hemisphere, pp. 51–66. Dempsey M. 1947. Decline in tuberculosis: the death rate fails to tell the entire story. Am Rev Tuberc 56:157–164. Deuffic S, Poynard T, Buffat L, Valleron AJ. 1998. Trends in primary liver cancer. Lancet 351:214–215. Devesa SS, Chow WH. 1993. Variation in colorectal cancer incidence in the United States by subsite of origin. Cancer 71:3819–3826. Devesa SS, Fears T. 1992. Non-Hodgkin’s lymphoma time trends: United States and international data. Cancer Res 5219(Suppl):5432s–5440s. Devesa SS, Blot WJ, Fraumeni JF Jr. 1998. Changing patterns in the incidence of esophageal and gastric carcinoma in the United States. Cancer 83:2049–2053. Devesa SS, Shaw GL, Blot WJ. 1991. Changing patterns of lung cancer incidence by histological type. Cancer Epidemiol Biomarkers Prev 1:29–34. De Vos Irvine H, Goldberg D, Hole DJ, McMenamin J. 1998. Trends in primary liver cancer. Lancet 351:215–216. Doll R, Cook P. 1967. Summarising indices for comparison of cancer incidence data. Int J Cancer 2:269–279. Doll R, Payne P, Waterhouse J, eds. 1966. Cancer Incidence in Five Continents: A Technical Report. New York: Springer. Doll R, Payne P, Waterhouse J, eds. 1970. Cancer Incidence in Five Continents, vol II. Berlin: Springer. Doll R, Peto R. 1981a. The Causes of Cancer. New York: Oxford University Press. Doll R, Peto R. 1981b. The causes of cancer: quantitative estimates of avoidable risks of cancer in the United States today. J Natl Cancer Inst 66:1191–1308. Dorn HF, Cutler SJ. 1959. Morbidity from cancer in the United States: Parts I and II. Public Health Monograph No. 56. Washington, DC: US Department of Health, Education, and Welfare. Dorn HF, Cutler SJ. Morbidity from cancer in the United States. Parts I and II. Public Health Monograph No. 56. Washington, DC: U.S. Department of Health, Education and Welfare. Easton DF. 1994. The inherited component of cancer. Br Med Bull 50:527–535. Ederer F, Axtell LM, Cutler SJ. 1961. The relative survival rate: a statistical methodology. Monogr Natl Cancer Inst 6:101–121. Ekstrom AM, Signorello LB, Hansson LE, Bergstrom R, Lindgren A, Nyren O. 1999. Evaluating gastric cancer misclassification: a potential explana-

International Patterns of Cancer Incidence and Mortality tion for the rise in cardia cancer incidence. J Natl Cancer Inst. 91: 786–790. El-Serag HB, Mason AC. 1999. Rising incidence of hepatocellular carcinoma in the United States. N Engl J Med 340:745–750. Eltom MA, Jemal A, Mbulaiteye SM, Devesa SS, Biggar RJ. 2002. Trends in Kaposi’s sarcoma and non-Hodgkin’s lymphoma incidence in the United States from 1973 through 1998. J Natl Cancer Inst 94:1204–1210. Etzioni R, Legler JM, Feuer EJ, Merrill RM, Cronin KA, Hankey BF. 1999. Cancer surveillance series: interpreting trends in prostate cancer. III. Quantifying the link between population prostate-specific antigen testing and recent declines in prostate cancer mortality. J Natl Cancer Inst 91:1033–1039. Feldman AR, Kessler L, Myers MH, Naughton MD. 1986. The prevalence of cancer: estimates based upon the Connecticut Tumor Registry. N Engl J Med 315:1394–1397. Ferlay J, Bray F, Pisani P, Parkin DM, Ferlay J. 2001. GLOBOCAN 2000: Cancer Incidence, Mortality and Prevalence Worldwide, Version 1.0. IARC Cancer Base No. 5. Lyon: IARC. Fernandez Garrote L, Lence Anta JJ, Cabezas Cruz E, Romero T, Camacho R. 1996. Evaluation of the cervical cancer control program in Cuba. Bull Pan Am Health Org 30:387–391. Feuer EJ, Merrill RM, Hankey BF. 1999. Cancer surveillance series: interpreting trends in prostate cancer. II. Cause of death misclassification and the recent rise and fall in prostate cancer mortality. J Natl Cancer Inst 91:1025–1032. Ford D, Easton DF, Peto J. 1995. Estimates of the gene frequency of BRCA1 and its contribution to breast and ovarian cancer incidence. Am J Hum Genet 57:1457–1462. Fritz A, Percy C, Parkin DM, et al., editors. 2000. International Classification of Diseases for Oncology, 3rd ed. Geneva: World Health Organization. Fu YS, Berek JS, Hilborne LH. 1987. Diagnostic problems of in situ and invasive adenocarcinomas of the uterine cervix. Appl Pathol 5:47–56. Geddes M, Parkin DM, Khlat M, Balzi D, Buiatti E, editors. 1993. Cancer in Italian Migrants Populations. IARC Scientific Publ. No. 123. Lyon: IARC. Gilliland FD, Samet JM. 1994. Lung cancer. Cancer Surv 19–20:175–195. Goedde HW, Agarwal DP, Fritze G, et al. 1992. Distribution of ADH2 and ALDH2 genotypes in different populations. Hum Genet 88:344–346. Goldcar DE, Reilly PR. 1995. A common BRCA1 mutation in the Ashkenazim. Nat Genet 11:113–114. Griffith GW. 1968. The sex ratio in gastric cancer and hypothetical considerations relative to aetiology. Br J Cancer 22:163–172. Groves FD, Linet MS, Travis LB, Devesa SS. 2000. Cancer surveillance series: non-Hodgkin’s lymphoma incidence by histologic subtype in the United States from 1978 through 1995. J Natl Cancer Inst 92:1240–1251. Grulich AE, Swerdlow AJ, Head J, Marmot MG. 1992. Cancer mortality in African and Caribbean migrants to England and Wales. Br J Cancer 66:905–911. Gustafsson L, Ponten J, Bergström R, Adami HO. 1997a International incidence rates of invasive cervical cancer before cytological screening. Int J Cancer 71:159–165. Gustafsson L, Ponten J, Zack M, Adami HO. 1997b. International incidence rates of invasive cervical cancer after introduction of cytological screening. Cancer Causes Control 8:755–763. Haenszel W, Kurihara M. 1968. Studies of Japanese migrants. I. Mortality from cancer and other diseases among Japanese in the United States. J Natl Cancer Inst 40:43–68. Hakama M. 1982. Trends in the incidence of cervical cancer in the Nordic countries. In: Magnus K, editor. Trends in Cancer Incidence. Causes and Practical Implications. Washington: Hemisphere Publishing Corporation, pp. 279–292. Hakama M, Hakulinen T, Teppo L, Saxen E. 1975. Incidence, mortality or prevalence as indicators of the cancer problem. Cancer 36:2227–2231. Hankey BF, Feuer EJ, Clegg LX, et al. 1999. Cancer surveillance series: interpreting trends in prostate cancer. I. evidence of the effects of screening in recent prostate cancer incidence, mortality and survival rates. J Natl Cancer Inst 91:1017–1024. Harkness EF, Brewster DH, Kerr KM, Fergusson RJ, MacFarlane GJ. 2002. Changing trends in incidence of lung cancer by histologic type in Scotland. Int J Cancer 102:179–183. Harris NL, Jaffe ES, Stein H, et al. 1994. A revised European-American classification of lymphoid neoplasms: a proposal from the International Lymphoma Study Group. Blood 84:1361–1392. Hartge P, Devesa SS. 1992. Quantification of the impact of known risk factors on time trends in non-Hodgkin’s lymphoma incidence. Cancer Res 52(Suppl 19):5566s–5569s. Hayne D, Brown RS, McCormack M, Quinn MJ, Payne HA, Babb P. 2001. Current trends in colorectal cancer: site, incidence, mortality and survival in England and Wales. Clin Oncol (R Coll Radiol). 13:448–452.

135

Heasman MA, Lipworth L. 1966. Accuracy of Certification of Cause of Death. Studies in Medical and Population Subjects 20. London: HMSO. Heck KE, Pamuk ER. 1997. Explaining the relation between education and postmenopausal breast cancer. Am J Epidemiol 145:366–372. Henderson BE, Ross R, Bernstein L. 1988. Estrogens as a cause of human cancer: the Richard and Hinda Rosenthal Foundation award lecture. Cancer Res 48:246–253. Hermon G, Beral V. 1996. Breast cancer mortality rates are levelling off or beginning to decline in many western countries: analysis of time trends, age-cohort and age-period models of breast cancer mortality in 20 countries. Br J Cancer 73:955–960. Herrero R, Brinton LA, Reeves WC, et al. 1992. Screening for cervical cancer in Latin America: a case-control study. Int J Epidemiol 21: 1050–1056. Herrinton LJ. 1998. Epidemiology of the Revised European-American Lymphoma Classification subtypes. Epidemiol Rev 20:187–203. Hewitt M, Breen N, Devesa S. 1999. Cancer prevalence and survivorship issues: analyses of the 1992 National Health Interview Survey. J Natl Cancer Inst 91:1480–1486. Hill GB, Adelstein AM. 1967. Cohort mortality from carcinoma of the cervix. Lancet 2:605–606. Hoffman FL. 1915. The Mortality from Cancer Throughout the World. Newark, NJ: Prudential Press. Holford TR, Zheng T, Mayne ST, McKay LA. 1992. Time trends of nonHodgkin’s lymphoma: are they real? What do they mean? Cancer Res 1;52(Suppl 19):5443s–5446s. Hsing AW, Tsao L, Devesa SS. 2000. International trends and patterns of prostate cancer incidence and mortality. Int J Cancer 85:60–67. Huang J, Seow A, Shi CY, Lee HP. 1999. Colorectal carcinoma among ethnic Chinese in Singapore: trends in incidence rate by anatomic subsite from 1968 to 1992. Cancer 85:2519–2525. Hughes AM, Armstrong BK, Vajdic CM, Turner J, Grulich AE, Fritschi L, Milliken S, Kaldor J, Benke G, Kricker A. 2004. Sun exposure may protect against non-Hodgkin lymphoma: a case-control study. Int J Cancer. 10;112:865–871. IARC. 1997. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. Vol. 70: Epstein-Barr Virus and Kaposi’s Sarcoma Herpesvirus/Human Herpesvirus 8. Lyon: IARC. IARC. 1986. Monographs on the Evaluation of Carcinogenic Risks to Humans. Vol. 38: Tobacco Smoking. Lyon: IARC. Ishibe N, Kelsey KT. 1997. Genetic susceptibility to environmental and occupational cancers. Cancer Causes Control 8:504–513. Jaffe ES, Harris NL, Stein H, Vardiman JW, editors. 2001. WHO Classification of Tumours. Pathology and Genetics of Tumours of the Haematopoietic and Lymphoid Tissues. Lyon: IARC. Janssen-Heijnen ML, Nab HW, van Reek J, van der Heijden LH, Schipper R, Coebergh JW. 1995. Striking changes in smoking behaviour and lung cancer incidence by histological type in south-east Netherlands, 1960–1991. Eur J Cancer 31A:949–952. Jedrychowski W, Becher H, Wahrendorf J, Basa-Cierpialek Z, Gomola K. 1992. Effect of tobacco smoking on various histological types of lung cancer. J Cancer Res Clin Oncol 118:276–282. Jensen OM, Parkin DM, MacLennan R, Muir CS, Skeet RG, editors. 1991. Cancer Registration Principles and Methods. Lyon: IARC. Ji BT, Devesa SS, Chow WH, Jin F, Gao YT. 1998. Colorectal cancer incidence trends by subsite in urban Shanghai, 1972–1994. Cancer Epidemiol Biomarkers Prev 7:661–666. Jin F, Devesa SS, Chow WH, et al. 1999. Cancer incidence trends in urban Shanghai, 1972–1994: an update. Int J Cancer 83:435–440. Kamineni A, Williams MA, Schwartz SM, Cook LS, Weiss NS. 1999. The incidence of gastric carcinoma in Asian migrants to the Untied States and their descendants. Cancer Causes Control 10:77–83. Karjalainen S, Pukkala E. 1990. Social class as a prognostic factor in breast cancer survival. Cancer 66:819–826. Kiemeney LA, Coebergh JW, Koper NP, et al. 1994. Bladder cancer incidence and survival in the south-eastern part of The Netherlands, 1975–1989. Eur J Cancer 30A:1134–1137. Kjaer SK, Brinton LA. 1993. Adenocarcinomas of the uterine cervix: the epidemiology of an increasing problem. Epidemiol Rev 15:486–498. Kosunen TU, Aromaa A, Knekt P, et al. 1997. Helicobacter antibodies in 1973 and 1994 in the adult population of Vammala, Finland. Epidemiol Infect 119:29–34. Laden F, Spiegelman D, Neas LM, et al. 1997. Geographic variation in breast cancer incidence rates in a cohort of U.S. women. J Natl Cancer Inst 89:1373–1378. Lagergren J, Bergstrom R, Nyren O. 1999. Association between body mass and adenocarcinoma of the esophagus and gastric cardia. Ann Intern Med 130: 883–890.

136

PART II: THE MAGNITUDE OF CANCER

Laheij RJ, Straatman H, Verbeek AL, Jansen JB. 1999. Mortality trend from cancer of the gastric cardia in The Netherlands, 1969–1994. Int J Epidemiol 28:391–395. Last JM editor. 1988. A Dictionary of Epidemiology. New York: Oxford University Press. Lauren P. 1965. The two histological main types of gastric carcinoma: diffuse and so-called intestinal-type carcinoma: an attempt at a histo-clinical classification. Acta Pathol Microbiol Scand 64:31–49. La Vecchia C, Airoldi L. 1999. Human bladder cancer: epidemiological, pathological and mechanistic aspects. IARC Sci Publ. 147:139–157. La Vecchia C, Levi F, Lucchini F, Franceschi S, Negri E. 1994. Worldwide patterns and trends in mortality from liver cirrhosis, 1955 to 1990. Ann Epidemiol 4:480–486. La Vecchia C, Lucchini F, Franceschi S, Negri E, Levi F. 2000. Trends in mortality from primary liver cancer in Europe. Eur J Cancer 36:909– 915. Layard R, Glaister S. 1994. Cost-Benefit Analysis. Cambridge, UK: Cambridge University Press. Lazcano-Ponce EC, Buiatti E, Najera-Aguilar P, Alonso-de-Ruiz P, HernandezAvila M. 1998. Evaluation model of the Mexican national program for early cervical cancer detection and proposals for a new approach. Cancer Causes Control 9:241–251. Lee HP, Chia KS, Shanmugaratnam K. 1992. Cancer incidence in Singapore 1983–87. Singapore: Singapore Cancer Registry. Lee HP, Day NE, Shanmugaratnam K. 1988. Trends in cancer incidence in Singapore 1968–82. IARC Scientific Publ. No. 91. Lyon: IARC. Lee PN, Fry JS, Forey BA. 1990. Trends in lung cancer, chronic obstructive lung disease, and emphysema death rates for England and Wales 1941–85 and their relation to trends in cigarette smoking. Thorax 45:657–665. Legler JM, Feuer EJ, Potosky AL et al. 1998. The role of prostate specific antigen (PSA) testing patterns in recent prostate cancer incidence decline in the United States. Cancer Causes Control 9:519–527. Le Marchand L. 1999. Combined influence of genetic and dietary factors on colorectal cancer incidence in Japanese Americans. Monogr Natl Cancer Inst 26:101–105. Leung GM, Thach TQ, Lam TH, et al. 2002. Trends in breast cancer incidence in Hong Kong between 1973 and 1999: an age-period-cohort analysis. Br J Cancer 87:982–988. Li HQ, Jin SQ, Xu HX, Thomas DB. 2000. The decline in the mortality rates of cervical cancer and a plausible explanation in Shandong, China. Int J Epidemiol 29:398–404. Llorca J, Prieto MD, Delgado-Rodriguez M. 1999. Increase in cervical cancer mortality in Spain, 1951–1991. J Epidemiol Community Health 53:408–411. Lopez-Abente G, Pollan M, de la Iglesia P, Ruiz M. 1995. Characterization of the lung cancer epidemic in the European Union 1970–1990. Cancer Epidemiol Biomarkers Prev 4:813–820. Lopez-Abente G, Pollan M, Vergara A, et al. 1997. Age-period-cohort modeling of colorectal cancer incidence and mortality in Spain. Cancer Epidemiol Biomarkers Prev 6:999–1005. Lu JB, Yang WX, Liu JM, Li YS, Qin YM. 1985. Trends in morbidity and mortality for oesophageal cancer in Linxian County, 1959–1983. Int J Cancer 36:643–645. Lubin JH, Blot WJ. 1984. Assessment of lung cancer risk factors by histologic category. J Natl Cancer Inst 73:383–389. Mahboubi E, Kmet J, Cook PJ, Day NE, Ghadirian P Salmasizadeh S. 1973. Oesophageal cancer studies in the Caspian Littoral of Iran: the Caspian Cancer Registry. Br J Cancer 28:197–214. Makimoto K, Higuchi S. 1999. Alcohol consumption as a major risk factor for the rise in liver cancer mortality rates in Japanese men. Int J Epidemiol 28:30–34. McGlynn KA, Tsao L, Hsing A, Devesa SS, Fraumeni JF Jr. 2001. International trends and patterns of primary liver cancer. Int J Cancer 94:290–296. McMichael AJ, Giles GG. 1994. Colorectal cancer. Cancer Surv 19–20: 77–98. McMichael AJ, McCall MG, Hartshorne JM, Woodings TL. 1980. Patterns of gastrointestinal cancer in European migrants to Australia: the role of dietary change. Int J Cancer 25:431–437. McPherson K, Steel CM, Dixon JM. 2000. Breast cancer: epidemiology, risk factors and genetics. BMJ 321:624–628. Melbye M, Trichopoulos D. 2002. Non-Hodgkin’s lymphomas. In: Adami HO, Hunter D, Trichoplous D, editors. Textbook of Cancer Epidemiology. Oxford: Oxford University Press. Mercer SL, Goel V, Levy IG, Ashbury FD, Iverson DC, Iscoe NA. 1997. Prostate cancer screening in the midst of controversy: Canadian men’s knowledge, beliefs, utilization, and future intentions. Can J Public Health 88:327–332.

Merrill RM, Capocaccia R, Feuer EJ, Mariotto A. 2000. Cancer prevalence estimates based on tumor registry data in the Surveillance, Epidemiology, and End Results (SEER) program. Int J Epidemiol 29:197–207. Miller BA, Kolonel LN, Bernstein L, et al, editors. 1996. Racial/Ethnic Patterns of Cancer in the United States 1988–1992. NIH Publ. No. 96-4104. Bethesda, MD: National Cancer Institute. Mitchell H, Medley G, Gordon I, Giles G. 1995. Cervical cytology reported as negative and risk of adenocarcinoma of the cervix: no strong evidence of benefit. Br J Cancer 71:894–897. Modan B, Wagener DK, Feldman JJ, Rosenberg HM, Feinleib M. 1992. Increased mortality from brain tumors: a combined outcome of diagnostic technology and change of attitude toward the elderly. Am J Epidemiol 135:1345–1357. Moolgavkar SH, Stevens RG, Lee JA. 1979. Effect of age on incidence of breast cancer in females. J Natl Cancer Inst 62:493–501. Morrow RH, Bryant JH. 1995. Health policy approaches to measuring and valuing human life: conceptual and ethical issues. Am J Public Health 85:1356–1360. Mueller NE, Mohar A, Evans A. 1992. Viruses other than HIV and nonHodgkin’s lymphoma. Cancer Res 52(19 Suppl):5479s–5481s. Muir C, Waterhouse J, Mack T, Powell J, Whelan S. 1987. Cancer Incidence in Five Continents, vol. V. IARC Scientific Publications No. 88. Lyon: IARC. Muir CS. 1996. Epidemiology of cancer in ethnic groups. Br J Cancer 2:(Suppl)12–16. Muir CS, Fraumeni JF Jr, Doll R. 1994. The interpretation of time trends. Cancer Surv 19/20:5–21. Muñoz N. 1988. Descriptive epidemiology of stomach cancer. In: Reed PI, Hill MJ, editors. Gastric carcinogenesis. Amsterdam: Elsevier Science, pp. 51–69. Muñoz N, Asvall J. 1971. Time trends of intestinal and diffuse types of gastric cancer in Norway. Int J Cancer 8:144–157. Muñoz N, Day NE. 1996. Esophageal cancer. In: Schottenfeld D, Fraumeni JF Jr, editors. Cancer Epidemiology and Prevention, 2nd edition. New York: Oxford University Press, pp. 681–706. Murray CJL. 1994. Quantifying the burden of disease: the technical basis for disability-adjusted life years. Bull WHO 72:429–445. Murray CJL, Lopez A. 1996. The Global Burden of Disease. Boston: Harvard University Press. Murray JL, Axtell LM. 1974. Impact of cancer: years of life lost due to cancer mortality. J Natl Cancer Inst 52:3–7. Norppa H. 1997. Cytogenetic markers of susceptibility: influence of polymorphic carcinogen-metabolizing enzymes. Environ Health Perspect 105:829–835. Oliver SE, May MT, Gunnell D. 2001. International trends in prostate-cancer mortality in the “PSA ERA.” Int J Cancer 92:893–898. Parkin DM. 1989. Cancers of the breast, endometrium and ovary: geographic correlations. Eur J Cancer Clin Oncol 25:1917–1925. Parkin DM. 2001. Global cancer statistics in the year 2000. Lancet Oncol 2:533–543. Parkin DM, Khlat M. 1996. Studies of cancer in migrants: rationale and methodology. Eur J Cancer 32A:761–771. Parkin DM, Moss SM. 2000. Lung cancer screening: improved survival but no reduction in deaths: the role of “overdiagnosis.” Cancer 89:2369–2376. Parkin DM, Ferlay J, Hamdi-Cherif M, et al. 2003. Cancer in Africa: Epidemiology and Prevention. IARC Scientific Publ. No. 153. Lyon: IARC. Parkin DM, Nguyen-Dinh X, Day NE. 1985. The impact of screening on the incidence of cervical cancer in England and Wales. Br J Obstet Gynaecol 92:150–157. Parkin DM, Muir CS, Whelan SL, Gao YT, Ferlay J, Powell J. 1992. Cancer Incidence in Five Continents, vol. VI. IARC Scientific Publications No. 120, Lyon: IARC. Parkin DM, Ohshima H, Srivatanakul P, Vatanasapt V. 1993. Cholangiocarcinoma: epidemiology, mechanisms of carcinogenesis and prevention. Cancer Epidemiol Biomarkers Prev 2:537–544. Parkin DM, Pisani P, Ferlay J. 1999. Estimates of the worldwide incidence of twenty-five major cancers in 1990. Int J Cancer 80:827–841. Parkin DM, Pisani P, Lopez AD, Masuyer E. 1994. At least one in seven cases of cancer is caused by smoking: global estimates for 1985. Int J Cancer 59:494–504. Parkin DM, Steinitz R, Khlat M, Kaldor J, Katz L, Young J. 1990. Cancer in Jewish migrants to Israel. Int J Cancer 45:614–621. Parkin DM, Whelan SL, Ferlay J, Raymond L, Young J, editors. 1997. Cancer Incidence in Five Continents. Vol VII. iarc Scientific Publ. No. 143. Lyon: IARC. Parkin DM, Whelan SL, Ferlay J, Teppo L, Thomas DB, editors. 2002. Cancer Incidence in Five Continents. Vol. VIII. IARC Scientific Publ. No. 155. Lyon: IARC.

International Patterns of Cancer Incidence and Mortality Parrish KM, Higuchi S, Lucas LJ. 1993. Increased alcohol-related oesophageal cancer mortality rates in Japanese men. Int J Epidemiol 22:600–605. Percy C, Dolman A. 1978. Comparison of the coding of death certificates related to cancer in seven countries. Public Health Rep 93:335–350. Percy C, Muir C. 1989. The international comparability of cancer mortality data: results of an international death certificate study. Am J Epidemiol 129:934–946. Percy C, Stanek E, Gloeckner L. 1981. Accuracy of cancer death certificates and its effect on cancer mortality statistics. Am J Public Health 71:242–250. Persson I, Bergstrom R, Barlow L, Adami HO. 1998. Recent trends in breast cancer incidence in Sweden. Br J Cancer 77: 167–169. Peto R, Boreham J, Clarke M, Davies C, Beral V. 2000. UK and USA breast cancer deaths down 25% in year 2000 at ages 20–69 years. Lancet 355:1822. Peto R, Lopez AD, Boreham J, Thun M, Heath C Jr, Doll R. 1994. Mortality from smoking in developed countries 1950–2000. New York: Oxford University Press. Peto R, Parish SE, Gray RG. 1985. There is no such thing as ageing, and cancer is not related to it. IARC Scientific Publ. No. 58. Lyon: IARC, pp. 43–53. Pinheiro PS, van der Heijden LH, Coebergh JW. 1999. Unchanged survival of gastric cancer in the southeastern Netherlands since 1982: result of differential trends in incidence according to Lauren type and subsite. Int J Cancer 84:28–32. Pisani P, Bray F, Parkin DM. 2002. Estimates of the worldwide prevalence of cancer for 25 in the adult population. Int J Cancer 97:72–81. Pisani P, Parkin DM, Bray FI, Ferlay J. 1999. Estimates of the worldwide mortality from twenty-five major cancers in 1990. Implications for prevention, and projections of future burden. Int J Cancer 83:18–29. Pisani P, Parkin DM, Munoz N, Ferlay J. 1997. Cancer and infection: estimates of the attributable fraction in 1990. Cancer Epidemiol Biomarkers Infect 6:387–400. Pollan M, Lopez-Abente G, Moreno C, et al. 1998. Rising incidence of nonHodgkin’s lymphoma in Spain: analysis of period of diagnosis and cohort effects. Cancer Epidemiol Biomarkers Prev 7:621–625. Pompe-Kirn V, Kovacic J, Primic Zakelj M. 1992. Epidemiological evaluation of cervical cancer screening in Slovenia up to 1986. Eur J Gynaecol Oncol 13:75–82. Ponten J, Adami HO, Bergstrom R, et al. 1995. Strategies for global control of cervical cancer. Int J Cancer 60:1–26. Post PN, Kil PJ, Crommelin MA, Schapers RF, Coebergh JW. 1998. Trends in incidence and mortality rates for prostate cancer before and after prostatespecific antigen introduction: a registry-based study in southeastern Netherlands, 1971–1995. Eur J Cancer 34:705–709. Potosky AL, Kessler L, Gridley G, Brown CC, Horm JW. 1990. Rise in prostatic cancer incidence associated with increased use of transurethral resection. J Natl Cancer Inst 82:1624–1628. Potosky AL, Miller BA, Albertsen PC, Kramer BS. 1995. The role of increasing detection in the rising incidence of prostate cancer. JAMA 273:548–552. Powell J, McConkey CC. 1990. Increasing incidence of adenocarcinoma of the gastric cardia and adjacent sites. Br J Cancer 62:440–443. Puffer RR, Wynne Griffith G. 1967. Patterns of Urban Mortality. Scientific Publ. No. 151. Washington, DC: Pan American Health Organization. Pukkala E, Soderman B, Okeanov A, et al. 2001. Cancer Atlas of northern Europe. Publ. No. 62. Helsinki: Cancer Society of Finland. Quinn M, Babb P, Jones J, Allen E. 1999. Effect of screening on incidence of and mortality from cancer of cervix in England: evaluation based on routinely collected statistics. BMJ 318:904–908. Remick SC. 1995. Acquired immunodeficiency syndrome-related nonHodgkin’s lymphoma. Cancer Control 2:97–103. Ries LAG, Eisner MP, Kosary CL, et al. 2001. SEER Cancer Statistics Review, 1973–1998. Bethesda: National Cancer Institute (http://seer.cancer.gov/ publications/csr/1973–1998/). Roosendaal R, Kuipers EJ, Buitenwerf J, et al. 1997. Helicobacter pylori and the birth cohort effect: evidence of a continuous decrease of infection rates in childhood. Am J Gastroenterol 92:1480–1482. Rose EF, McGlashan ND. 1975. The spatial distribution of oesophageal carcinoma in the Transkei, South Africa. Br J Cancer 31:197–206. Rothman KJ, Greenland S. 1998. Modern Epidemiology, 2nd ed. Philadelphia: Lippincot-Raven. Sadjadi A, Malekzadeh R, Derakhshan MH, et al. 2003. Cancer occurrence in Ardabil: results of a population-based cancer registry from Iran. Int J Cancer 107:113–118. Saidi F, Sepehr A, Fahimi S, et al. 2000. Oesophageal cancer among the Turkomans of northeastern Iran. Br J cancer 83:1249–1254. Sankaranarayanan R, Black RJ, Parkin DM, editors. 1998. Cancer Survival in Developing Countries. IARC Scientific Publ. No. 145. Lyon: IARC.

137

Saxen EA. 1982. Trends: facts or fallacy? In: Magnus K, editor. Trends in Cancer Incidence: Causes and Practical Implications. Washington, DC: Hemisphere, pp. 5–16. Schrijvers CTM, Mackenbach JP, Lutz JM, Quinn MJ, Coleman MP. 1995. Deprivation, stage at diagnosis and cancer survival. Int J Cancer 63:324–329. Segi M. 1960. Cancer Mortality for Selected Sites in 24 Countries (1950–57). Sendai: Tohoku University School of Public Health. Seow A, Duffy SW, McGee MA, Lee J, Lee HP. 1996. Breast cancer in Singapore: trends in incidence 1968–1992. Int J Epidemiol 25:40–45. Shimizu H, Ross RK, Bernstein L. 1991. Possible underestimation of the incidence rate of prostate cancer in Japan. Jpn J Cancer Res 82:483–485. Sigurdsson K. 1995. Quality assurance in cervical cancer screening: the Icelandic experience. Eur J Cancer 31A:728–734. Sigurdsson K. 1999. The Icelandic and Nordic cervical screening programs: trends in incidence and mortality rates through 1995. Acta Obstet Gynecol Scand 78:478–485. Smans M, Muir CS, Boyle P. 1992. Atlas of Cancer Mortality in the European Economic community. IARC Scientific Publ. No. 107. Lyon: IARC. Smedby KE, Hjalgrim H, Melbye M, Torrang A, Rostgaard K, Munksgaard L, Adami J, Hansen M, Porwit-MacDonald A, Jensen BA, Roos G, Pedersen BB, Sundstrom C, Glimelius B, Adami HO. 2005. Ultraviolet radiation exposure and risk of malignant lymphomas. J Natl Cancer Inst 97:199–209. Sobue T, Ajiki W, Tsukuma H, Oshima A, Hanai A, Fujimoto I. 1999. Trends of lung cancer incidence by histologic type: a population-based study in Osaka, Japan. Jpn J Cancer Res 90:6–15. Sturgeon SR, Schairer C, Gail M, McAdams M, Brinton LA, Hoover RN. 1995. Geographic variation in mortality from breast cancer among white women in the United States. J Natl Cancer Inst 87:1846–1853. Svensson E, Grotmol T, Hoff G, Langmark F, Norstein J, Tretli S. 2002. Trends in colorectal cancer incidence in Norway by gender and anatomic site: an age-period-cohort analysis. Eur J Cancer Prev 11:489–495. Swerdlow A, Dos Santos I, Doll R. 2001. Issues to consider in interpretation of trends. In: Swerdlow A, Dos Santos I, Doll R, editors. Cancer: Incidence and Mortality in England and Wales—Trends and Risk Factors. Oxford, UK: Oxford University Press, pp. 8–16. Swerdlow AJ, dos Santos Silva I, Reid A, Qiao Z, Brewster DH, Arrundale J. 1998. Trends in cancer incidence and mortality in Scotland: description and possible explanations. Br J Cancer 77 Suppl 3:1–54. Tanaka K, Hirohata T, Koga S, et al. 1991. Hepatitis C and hepatitis B in the etiology of hepatocellular carcinoma in the Japanese population. Cancer Res 51:2842–2847. Taylor-Robinson SD, Toledano MB, Arora S, Keegan TJ, Hargreaves S, Beck A, Khan SA, Elliott P, Thomas HC. 2001. Increase in mortality rates from intrahepatic cholangiocarcinoma in England and Wales 1968–1998. Gut 48:816–820. Thorn M, Bergstrom R, Johansson AM, Ramstrom L, Persson I, Malmstrom PU. 1997. Trends in urinary bladder cancer incidence in Sweden 1960–93 with special reference to histopathology, time period, birth cohort, and smoking. Cancer Causes Control 8:560–567. Thorn M, Bergstrom R, Kressner U, Sparen P, Zack M, Ekbom A. 1998. Trends in colorectal cancer incidence in Sweden 1959–93 by gender, localization, time, period, and birth cohort. Cancer Causes Control 9:145–152. Threlfall TJ, English DR, Rouse, IL. 1998. Prostate cancer in Western Australia: trends in incidence and mortality from 1985 to 1996. Med J Aust 169:21–24. Travis WD, Lubin J, Ries L, Devesa S. 1996. United States lung carcinoma incidence trends: declining for most histologic types among males, increasing among females. Cancer 77:2464–2470. Troisi RJ, Freedman AN, Devesa SS. 1999. Incidence of colorectal carcinoma in the U.S.: an update of trends by gender, race, age, subsite, and stage, 1975–1994. Cancer 85:1670–1676. Tsukuma H, Ajiki W, Oshima A. 1999. Cancer incidence and survival in Osaka. In: Cancer Epidemiology and Control in the Asia-Pacific Region. Kobe: International Center for Medical Research, Kobe University School of Medicine. Tulinius H, Storm HH, Pukkala E, Andersen A, Ericsson J. 1992. Cancer in the Nordic countries, 1981–86. A joint publication of the five Nordic Cancer Registries. APMIS Suppl. 31:1–194. Tuyns AJ, Masse LM. 1973. Mortality from cancer of the oesophagus in Brittany. Int J Epidemiol 2:241–245. Tuyns AJ, Masse G. 1975. Cancer of the oesophagus in Brittany: an incidence study in Ille-et-Vilaine. Int J Epidemiol 4:55–59. Tyczynski J, Tarkowski W, Parkin DM, Zatonski W. 1994. Cancer mortality among Polish migrants to Australia. Eur J Cancer 30A:478–484. Ursin G, Bernstein L, Pike MC. 1994a. Breast cancer. Cancer Surv 19–20:241–64.

138

PART II: THE MAGNITUDE OF CANCER

Ursin G, Peters RK, Henderson BE, d’Ablaing G III, Monroe KR, Pike MC. 1994b. Oral contraceptive use and adenocarcinoma of cervix. Lancet 344:1390–1394. Vizcaino AP, Moreno V, Bosch FX, Muñoz N, Barros-Dios XM, Parkin DM. 1998. International trends in the incidence of cervical cancer. I. Adenocarcinoma and adenosquamous cell carcinomas. Int J Cancer 75:536–545. Vizcaino AP, Moreno V, Lambert R, Parkin DM. 2002. Time trends in incidence of both major histologic (sic) types of esophageal carcinomas in selected countries, 1973–1995. Int J Cancer 99:860–868. Vlasak V, Plesko I, Dimitrova E, Hudakova G. 1991. Recent trends in uterine cervix cancer in Slovakia, 1968–1987. Neoplasma 38:533–540. Vyslouzilova S, Arbyn M, Van Oyen H, Drieskens S, Quataert P. 1997. Cervical cancer mortality in Belgium, 1955–1989: a descriptive study. Eur J Cancer 33:1841–1845. Wabinga HR, Parkin DM, Wabwire-Mangen F, Nambooze S. 2000. Trends in cancer incidence in Kyadondo county, Uganda, 1960–1997. Br J Cancer 82:1585–1592. Wakai K, Suzuki S, Ohno Y, Kawamura T, Tamakoshi A, Aoki R. 1995. Epidemiology of breast cancer in Japan. Int J Epidemiol 24:285–291. Wakefield JC, Best NG, Waller LA. 2000. Bayesian approaches to disease mapping. In: Elliot P, Wakefield JC, Best NG, Briggs DJ, editors. Spatial epidemiology: Methods and applications. Oxford, UK: Oxford University Press. Walboomers JM, Jacobs MV, Manos MM, et al. 1999. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol 189:12–19. Waterhouse J, Muir C, Shanmugaratnam K, Powell J, eds. 1982. Cancer Incidence in Five Continents, vol. IV. IARC Scientific Publications No. 42. Lyon: IARC. Waterhouse J, Muir C, Correa P, Powell J, eds. 1976. Cancer Incidence in Five Continents, vol. III. IARC Scientific Publications No. 15. Lyon: IARC. WHO. 1993. World Health Statistics Annual 1992. Geneva: WHO, p. xxii. WHO. 1997. Atlas of Mortality in Europe: Subnational Patterns 1980/1981 and 1990/1991. WHO Regional Publications; European Series 75. Geneva: WHO. WHO. 2000. The World Health Report 2000 Health Systems: Improving Performance. Geneva: WHO.

WHO. 2002. National Cancer Control Programmes: Policies and Managerial Guidelines. Geneva: WHO. World Cancer Research Fund (WCRF) Panel. 1997. Diet, Nutrition and the Prevention of Cancer: A Global Perspective. Washington, DC: World Cancer Research Fund. Wynder EL, Graham EA. 1950. Tobacco smoking as a possible etiologic factor in brochiogenus carcinoma: a study of six hundred and eighty four proved cases. JAMA 143:328–336. Wynder EL, Muscat JE. 1995. The changing epidemiology of smoking and lung cancer histology. Environ Health Perspect 103(Suppl 8):143– 148. Yang L, Li LD, Chen TD, Parkin DM. 2003. Time Trends in Cancer Mortality in China: 1987–1999. Int J Cancer 106:771–783. Yatani R, Chigusa I, Akazaki K, Stemmermann GN, Welsh RA, Correa P. 1982. Geographic pathology of latent prostatic carcinoma. Int J Cancer 29:611–616. Yeole BB, Jayant K, Jussawalla DJ. 1990. Trends in breast cancer incidence in greater Bombay: an epidemiological assessment. Bull WHO 68: 245–249. Yokoyama A, Kato H, Yokoyama T, et al. 2002. Genetic polymorphisms of alcohol and aldehyde dehydrogenases and glutathione S-transferase M1 and drinking, smoking, and diet in Japanese men with esophageal squamous cell carcinoma. Carcinogenesis 23:1851–1859. Yu MC, Ross RK, Chan KK, et al. 1995. Glutathione S-transferase m1 genotype affects aminobiphenyl-hemoglobin adduct levels in white, black, and Asian smokers and nonsmokers. Cancer Epidemiol Biomarkers Prev 4:861–864. Yu MC, Skipper PL, Taghizadeh K, et al. 1994. Acetylator phenotype, aminobiphenyl-hemoglobin adduct levels, and bladder cancer risk in white, black, and Asian men in Los Angeles, California. J Natl Cancer Inst 86:712–716. Zheng T, Holford TR, Boyle P, et al. 1994. Time trend and the ageperiod-cohort effect on the incidence of histologic types of lung cancer in Connecticut, 1960–1989. Cancer 74:1556–1567. Ziegler RG, Hoover RN, Pike MC, et al. 1993. Migration patterns and breast cancer risk in Asian-American women. J Natl Cancer Inst 85: 1819–1827.

9

Cancer Incidence, Mortality, and Patient Survival in the United States LYNN A.G. RIES AND SUSAN S. DEVESA

R

eliable assessment of the impact of cancer on the general population is predicated upon the availability of measurements related to cancer incidence, mortality, prevalence, and patient survival. Although population-based cancer mortality data have been available in the United States for nearly a century, cancer incidence data before 1973 were obtainable primarily from periodic surveys conducted in selected geographic areas of the United States during the periods 1937–1939, 1947–1948 (Dorn and Cutler, 1959), and 1969– 1971 (Cutler and Young, 1975). The National Cancer Institute (NCI) collected cancer patient survival data before 1973 mainly on hospitalbased cases through the End Results Program (Axtell et al., 1976). The Surveillance, Epidemiology, and End Results (SEER) program, the successor to the End Results Program and the periodic incidence surveys, has collected population-based data on newly diagnosed cancers since 1973. The SEER Program is an ongoing contractsupported program of the NCI that funds and coordinates the collection of cancer data in population-based cancer registries located throughout the United States. These registries, although not randomly selected, are thought to be reasonably representative of the U.S. population; a number of the registries cover diverse populations of particular epidemiologic interest. The SEER Program publishes incidence, mortality, and survival data annually. The SEER Cancer Statistics Review (CSR), 1975–2000 (Ries et al., 2003) contains data from 1975 to 2000 and is the primary source for much of the data in this chapter. Additional data are available on the SEER web site for the CSR and through Fast Stats and Cancer Query (http://www.seer.cancer.gov). A public-use file is also available for further epidemiologic analyses: see the same web site to access information.

MATERIAL AND METHODS The SEER Program collects incidence and survival data. Information on newly diagnosed cases includes the demographic characteristics of the patient, anatomic site of the malignancy, histologic cell type, extent (stage) of the disease at time of diagnosis, treatment, and follow-up including survival status and cause of death. The intent here is to provide the reader with an overview of the recent cancer data available by cancer, sex, and race/ethnicity. For this chapter, data from nine SEER registries (SEER9) (9.5% of the U.S. population) were used for long-term cancer incidence trends and survival rates: the states of Connecticut, Iowa, Utah, New Mexico, and Hawaii; and the metropolitan areas of Atlanta, Detroit, Seattle (Puget Sound), and San Francisco-Oakland. Additional data from Los Angeles, San Jose-Monterey, and the Alaska Native Cancer Registry were used for incidence rates for 1992–2000. These additional areas together with the SEER9 cover about 14% of the U.S. population and are labeled as SEER12 when the Alaska Native Registry is included and SEER11 areas otherwise. Calculations for Hispanic and white non-Hispanic populations excluded Detroit, Hawaii, and Alaska Natives in Alaska. Note that persons of Hispanic ethnicity may be of any race, and therefore Hispanic is not mutually exclusive from white, black, American Indian/Alaska Native (AI/AN), or Asian/Pacific Islander (API) groups. The white non-Hispanic group is a subset of total whites. Table 9–1 shows each of the SEER geographic catchment areas that were used in this chapter, along with percentages showing

the contributions of various subsets of the SEER data based on the number of geographic areas covered (SEER9, SEER11, and SEER12 and their contribution to the overall U.S. population) for all races and by selected race/ethnic groups in the United States. The SEER program recently has expanded to include four additional geographic areas (New Jersey, Kentucky, Louisiana, and the remaining geographic areas in California), which brings coverage of the United States to 26%. These data are not included in this chapter. Incidence rates and trends for 1992–2000 were based on 1,420,500 newly diagnosed malignancies among residents of the SEER12 areas. Long-term cancer incidence trends were based on 2,482,265 newly diagnosed cases during 1975–2000 for the SEER9 areas. Incidence data are presented for all malignant neoplasms classified according to the International Classification of Diseases for Oncology (ICD-O) (WHO, 1976, 1986, 1988, 1990). Although data for in situ cancers of all primary sites have been collected except for cervical cancer in situ since 1996, they are included only in the rates for bladder cancer because the distinction between in situ and early invasive bladder cancer has been inconsistent among pathologists and over time. In contrast with data from the CSR, borderline tumors of the ovary were excluded in this analysis because they were not reportable for the entire time period. Mortality data are obtained annually from the National Center for Health Statistics (NCHS), and the cancer is based on the underlying cause of death. Cancer death rates and recent trends for 1992–2000 were based on 4,846,984 cancer deaths in the total United States and 563,451 cancer deaths in the SEER11 areas plus an additional 1047 cancer deaths among Alaska Natives residing in Alaska. Long-term trends for 1975–2000 were based on 12,348,006 cancer deaths in the total United States. Population estimates were obtained from the Census Bureau, and now incorporate bridged single-race estimates for 2000 that are derived from the original multiple-race categories in the 2000 Census (as specified in the 1997 Office of Management and Budget standards for the collection of data on race and ethnicity). These bridged estimates are consistent with the four race groups enumerated in the 1990 Census (white; black; American Indian, Eskimo, and Aleut; Asian and Pacific Islander) and were produced under a collaborative arrangement between the NCHS and the Census Bureau. The bridged single-race estimates and a description of the methodology used to develop them appear on the NCHS website (http://www.cdc.gov/nchs/about/major/ dvs/popbridge/popbridge.htm). In addition, a revised set of 1990–1999 and new 2000 population estimates by county for ages <1 year, 1–4 years, 5–9 years, 10–14 years, . . . , 80–84 years, and 85+ years were recently obtained by the NCI from the Census Bureau through an interagency agreement. Populations are available by year, county, race, Hispanic origin, sex, and age. The race groups included in the census estimates are white, black, American Indian/Alaska Native, and Asian/Pacific Islander. The NCI makes a modification to the populations for Hawaii based on sample survey data collected by the Hawaii Department of Health. This effort grew out of a concern that the native Hawaiian population has been vastly undercounted in previous censuses. The “Hawaii adjustment” to the census estimates has the net result of reducing the estimated white population and increasing the Asian and Pacific Islander population for the state. The total population estimate for the United States reflects the substitution of the NCI estimates for Hawaii.

139

140

PART II: THE MAGNITUDE OF CANCER

Table 9–1. Total U.S. Populations and SEER Catchment Area Populations Used in this Chapter: 2000

Geographic Area Total US SEER 12: SEER11 + Alaska Native Alaskad SEER 11 SEER9 + SJ-M + LosAng San Jose-Monterey (SJ-M) Los Angeles (LosAng) SEER 9 Connecticut Atlanta Detroit Iowa Hawaii New Mexico Seattle-Puget Sound Utah San Francisco-Oakland

Blackb

American Indian/Alaska Native

Asian or Pacific Islanderb

211,460,626

34,658,190

2,475,956

13.8% 98,043

11.8% —

11.7% —

13.7% 2,393,183 9,519,338 9.5% 3,405,565 2,914,587 4,043,467 2,926,324 1,211,537 1,819,046 4,045,707 2,233,169 4,123,740

11.8% 1,356,968 4,637,062 9.0% 2,780,355 1,588,734 2,784,071 2,748,640 294,102 1,214,253 3,253,688 1,992,975 2,340,035

11.7% 65,282 930,957 8.8% 309,843 1,039,151 1,011,038 61,853 22,003 34,343 169,042 17,657 396,908

Total Populationa

Whiteb

281,421,906

Other Raceb

Two or More Races

Hispanicc

10,641,833

15,359,073

6,826,228

35,305,818

21.0% 98,043

35.7% —

24.7% —

23.0% —

21.8% —

17.0% 18,629 76,988 13.2% 9,639 7,030 13,375 8,989 3,535 173,483 57,340 29,684 22,635

35.7% 472,449 1,164,553 20.4% 83,679 124,467 102,365 37,644 617,407 20,758 310,052 52,253 817,906

24.7% 367,498 2,239,997 7.7% 147,201 100,263 45,190 37,420 15,147 309,882 94,541 93,405 338,983

23.0% 112,357 469,781 14.5% 74,848 54,942 87,428 31,778 259,343 66,327 161,044 47,195 207,273

21.8% 685,372 4,242,213 7.8% 320,323 229,427 118,641 82,473 87,699 765,386 213,215 201,559 733,249

Source: U.S. Bureau of Census, Census 2000, Summary File 1, Table DP-1. a Total population equals the sum of White, Black, American Indian/Alaska Native, Asian or Pacific Islander, Other Race, and two or more races. b Because each person could report multiple races in the 2000 Census, race-specific counts and percentages in these columns are based on persons self-reporting only one race. c Hispanic ethnicity is tabulated independently of race, so Hispanic persons may be of any race. d Only the American Indian/Alaska Native population in this state is covered by SEER.

Incidence and death rates are presented as the number of newly diagnosed cases or deaths per 100,000 people at risk per year (personyears). All incidence and death rates are age-adjusted by 5-year age groups to the 2000 U.S. standard million population. An age-adjusted rate is a weighted average of the age-specific rates, with each weight being the proportion of persons in the corresponding age group of a standard population that, in the present analysis, is the 2000 standard million population of the United States (Anderson and Rosenberg, 1998). Incidence rates for the SEER12 areas and death rates for the United States for 1992–2000 are presented in the tables in this chapter. For this short time period, the annual percent change (APC) was estimated by fitting a straight line through the natural logarithms of the yearly rates using standard least-squares procedures. Long-term rates and trends in incidence in the SEER9 areas and mortality in the United States during 1975–2000 are included in the CSR (Ries et al., 2003) and mentioned here where important. Increases or decreases in the APC are noted only when they are statistically significant, unless noted. Estimates of long-term trends between 1975 and 2000 are based on joinpoint analyses (Kim et al., 2000). For selected cancers, estimated incidence rates also were calculated to account for updates and additional reports that may not arrive at the cancer registry in a timely manner, especially for cancers diagnosed outside the hospital. In addition to delays in reporting, often pathology reports from a nonhospital setting do not have information on race, and further inquiry must be undertaken to update the race from “unknown race” to a known race and thereby increase the race-specific rate. A model was designed to take into account these delays and estimate what the rates will be when all of the data are finally reported (Clegg et al., 2002). Figures 9–1 to 9–8 (see later) present the long-term incidence and mortality trends for selected cancers using the same arithmetic x-axis and logarithmic y-axis, resulting in slopes that are comparable not only within but also between figures. The y : x-axis ratio has been chosen such that a slope of 10 degrees portrays a rate of change of 1% per year; as a result, one log cycle on the y-axis and 40 years on the x-axis are the same length (Devesa et al., 1995). For incidence, the observed data points or the delay-adjusted joinpoints, when they have been calculated, are presented. Analysis of longer-term trends (back to the 1950s) can be found in the report by Devesa et al. (1987) and in an updated table by Ries et al. (2003). The incidence of cancer varies enormously by age. The 5-year agespecific 1992–2000 incidence rates by race (whites, blacks) and sex

were calculated for the 11 SEER areas for 20 selected cancers. Figures are presented using an arithmetic x-axis for age and a logarithmic yaxis for the rate (see Fig. 9–9). Estimates of the number of new cases and deaths in 2003 were provided by the American Cancer Society (ACS) based on SEER and U.S. data for 1979–1999 (Jemal et al., 2003a). Relative survival rates are presented by historic stages of localized, regional, and distant disease for 765,470 patients diagnosed from 1992 to 1999 with follow-up through 2000. So a patient is not represented in multiple cohorts, survival time is computed from the time of diagnosis of the patient’s first primary cancer only; all cases identified by death certificate only or autopsy only and all cases with unknown survival time are excluded from the survival analysis. Relative survival rates estimate the likelihood that a patient will not die from causes directly related to the malignancy in question within a specific time period following diagnosis, the conventional survival interval being 5 years. The relative survival rate is obtained by correcting the observed survival rate for expected mortality using a procedure described by Ederer et al. (1961). Survival rates are presented for all races by stage (Appendix 9–C) or sex (Appendix 9–D). Additional survival rates by race (total, white, black), over time, or by age are found in the CSR (Ries et al., 2003). Prevalence estimates, when given, are the complete prevalence (Mariotto, 2002). Prevalence estimates were calculated for January 1, 2000 (Ries et al., 2003). The probability of developing cancer, when given, was based on SEER data from 1998–2000 (Ries et al., 2003). Before the SEER data can be offered as a measure of cancer incidence and patient survival for the entire United States, an assessment of how well the SEER areas represent the nation as a whole is essential. Because incidence data for the entire United States do not currently exist, cancer death rates for the SEER11 areas plus Alaska for the Native American rates were compared with total U.S. death rates for specific cancers among all races—white, black, API, AI/AN populations—and males and females (Table 9–2). For all races, both sexes, and all cancers combined, the U.S. rate is approximately 7% higher than the SEER rate. The rate is less than 1% different for Hispanics. The rates for the other racial groups are 4%–8% different, with the SEER rate being higher for some and lower for others. For all races, the rates are 9% lower for males and 4% lower for females in the SEER areas compared to those for the United States. By sex and racial/ethnic group, none of the SEER all cancers combined rates is

Table 9–2. Cancer Death Rates, United States vs. SEER by Selected Cancer Site/type, Race/Ethnicity, and Sex: 1992–2000 Rates for Total United States Cancer Site/Type

All Races White

All malignant cancers Total 206.5 202.9 Male 263.0 256.3 Female 170.7 169.3 Oral cavity and pharynx Total 3.0 2.8 Male 4.6 4.2 Female 1.8 1.7 Esophagus Total 4.3 4.0 Male 7.6 7.1 Female 1.8 1.6 Stomach Total 5.1 4.5 Male 7.3 6.5 Female 3.5 3.1 Colon and rectum Total 22.0 21.6 Male 26.9 26.5 Female 18.6 18.1 Liver and intrahepatic bile duct Total 4.4 4.0 Male 6.3 5.8 Female 2.9 2.7 Pancreas Total 10.6 10.3 Male 12.3 12.0 Female 9.2 8.9 Larynx Total 1.5 1.4 Male 2.8 2.5 Female 0.6 0.5 Lung and bronchus Total 57.6 57.4 Male 82.4 80.8 Female 40.2 40.9 Soft tissue including heart Total 1.5 1.4 Male 1.6 1.6 Female 1.4 1.3 Melanoma of the skin Total 2.7 3.0 Male 3.9 4.4 Female 1.8 2.0 Breast Female 29.2 28.7 Cervix uteri Female 3.2 2.8 Corpus and uterus, nos Female 4.1 3.9 Ovary Female 9.0 9.3 Prostate Male 35.3 32.5 Testis Male 0.3 0.3 Urinary bladder Total 4.4 4.5 Male 7.8 8.1 Female 2.3 2.3 Kidney and renal pelvis Total 4.2 4.3 Male 6.2 6.2 Female 2.8 2.9 Brain and other nervous system Total 4.7 5.0 Male 5.7 6.1 Female 3.9 4.1 Hodgkin lymphoma Total 0.5 0.6 Male 0.7 0.7 Female 0.4 0.5

Rates for SEER Areas

Black

AI/AN

API

Hispanic

White NonHispanic

Black

AI/AN

API

Hispanic

White NonHispanic

263.9 369.1 201.3

137.2 170.4 115.5

128.1 160.1 104.0

137.7 178.0 111.6

205.3 258.8 171.6

192.6 239.0 163.4

192.8 237.5 165.5

253.0 340.6 200.0

142.9 170.3 123.9

139.0 175.5 110.8

138.9 174.8 116.6

197.9 242.6 170.3

4.9 8.6 2.2

2.3 3.7 1.3

2.6 3.9 1.5

2.0 3.5 0.9

2.9 4.3 1.7

3.0 4.4 1.8

2.8 4.1 1.8

4.2 7.1 2.0

3.3 5.4 1.8

3.1 4.7 1.7

1.7 2.9 0.8

3.0 4.2 1.9

7.7 13.4 3.6

2.6 4.7 1.1

2.2 3.8 1.0

2.6 4.7 1.0

4.1 7.2 1.6

4.1 7.0 1.8

4.0 6.9 1.7

6.8 11.3 3.6

3.1 5.5 1.4

2.4 4.3 0.9

2.3 4.4 0.8

4.1 7.1 1.8

9.9 14.6 6.8

5.6 7.5 4.2

10.3 13.3 7.9

7.6 10.4 5.6

4.3 6.2 2.9

5.8 8.3 4.0

4.9 7.0 3.4

9.3 13.6 6.5

8.0 11.1 5.6

11.3 15.0 8.4

8.8 12.1 6.4

4.2 6.2 2.8

28.9 35.2 25.0

14.0 16.4 12.2

13.7 16.6 11.5

14.2 18.2 11.5

21.8 26.7 18.4

20.4 24.8 17.3

20.2 24.7 17.1

27.8 33.3 24.4

15.8 16.4 15.2

15.4 18.8 12.6

14.2 18.3 11.4

20.6 24.9 17.4

5.9 9.1 3.7

5.5 7.5 4.1

10.9 16.1 6.7

6.9 10.0 4.6

3.8 5.4 2.5

4.9 7.2 3.1

4.1 5.9 2.6

6.0 9.3 3.6

6.4 8.6 5.0

10.9 16.1 6.6

7.0 10.0 4.8

3.7 5.4 2.5

14.6 16.8 13.0

6.3 6.4 6.1

7.6 8.7 6.7

8.2 9.4 7.3

10.3 12.0 9.0

10.4 11.8 9.2

10.2 11.7 9.0

14.9 16.3 13.7

7.3 7.8 6.8

8.2 9.6 7.1

8.4 9.2 7.7

10.3 11.9 9.1

3.0 5.9 1.0

1.1 1.9 0.5

0.5 0.9 0.1

1.2 2.5 0.2

1.4 2.5 0.5

1.3 2.4 0.5

1.2 2.2 0.5

2.7 5.2 0.9

1.1 2.2 Ÿ

0.5 0.9 0.1

0.9 1.9 0.2

1.3 2.2 0.5

68.1 112.2 39.3

36.9 52.5 25.9

29.2 42.1 19.2

25.6 41.2 14.7

58.8 82.4 42.2

49.7 67.8 37.0

50.1 67.0 38.5

66.0 103.1 41.3

33.2 46.4 23.4

32.0 46.7 20.5

24.1 36.5 15.6

53.1 70.0 41.3

1.8 1.6 1.9

0.9 0.9 0.9

1.0 1.2 0.9

1.1 1.2 1.1

1.4 1.6 1.3

1.4 1.5 1.4

1.4 1.6 1.3

1.8 1.5 2.0

1.0 Ÿ Ÿ

1.2 1.3 1.1

1.2 1.2 1.2

1.4 1.6 1.3

0.5 0.5 0.5

0.6 0.8 0.5

0.4 0.5 0.3

0.8 1.1 0.6

3.2 4.6 2.1

2.4 3.5 1.6

2.9 4.2 2.0

0.4 0.5 0.4

Ÿ Ÿ Ÿ

0.5 0.6 0.4

0.8 1.0 0.7

3.2 4.7 2.1

36.7

14.9

12.9

18.1

29.2

28.2

28.7

37.1

16.6

15.0

17.9

29.9

6.5

3.3

3.1

4.0

2.7

2.7

2.5

4.9

2.7

2.8

4.0

2.3

7.0

2.4

2.2

3.2

3.9

4.1

4.0

6.4

2.7

2.5

3.2

4.1

7.6

4.8

4.7

6.1

9.5

9.0

9.6

7.0

5.4

4.9

6.5

10.1

75.9

22.9

15.2

25.5

32.7

33.2

32.3

69.0

17.9

16.2

25.4

33.1

0.1

0.2

0.1

0.2

0.3

0.3

0.3

0.1

Ÿ

0.1

0.3

0.3

4.1 6.0 3.0

1.5 2.4 1.0

1.8 2.8 1.1

2.4 4.2 1.2

4.6 8.2 2.3

4.1 7.1 2.2

4.3 7.7 2.3

4.0 6.1 3.0

1.4 2.0 Ÿ

1.9 2.9 1.1

2.3 3.6 1.4

4.5 7.9 2.3

4.2 6.2 2.8

4.8 6.7 3.4

1.9 2.8 1.2

3.6 5.2 2.4

4.3 6.2 2.9

3.9 5.6 2.6

4.0 5.9 2.7

3.9 5.6 2.8

5.5 7.5 4.0

2.2 3.3 1.3

3.7 5.4 2.5

4.1 5.9 2.7

2.8 3.3 2.3

2.0 2.5 1.6

1.9 2.2 1.7

2.9 3.5 2.4

5.2 6.3 4.3

4.6 5.6 3.7

5.0 6.2 4.1

2.9 3.5 2.5

2.0 2.5 1.5

2.2 2.6 1.9

3.0 3.6 2.6

5.4 6.7 4.3

0.5 0.7 0.4

0.2 0.3 0.2

0.2 0.3 0.1

0.6 0.8 0.4

0.6 0.7 0.5

0.5 0.7 0.4

0.6 0.7 0.4

0.5 0.6 0.4

Ÿ Ÿ Ÿ

0.2 0.2 0.1

0.6 0.9 0.5

0.6 0.7 0.4

All Races White

(continued)

141

142

PART II: THE MAGNITUDE OF CANCER

Table 9–2. (cont.) Rates for Total United States Cancer Site/Type

All Races White

Non-Hodgkin lymphoma Total 8.5 8.9 Male 10.6 11.0 Female 7.0 7.3 Myeloma Total 3.9 3.6 Male 4.8 4.5 Female 3.3 3.0 Leukemia Total 7.8 7.9 Male 10.5 10.7 Female 6.0 6.1 Miscellaneous malignant cancers Total 14.9 14.5 Male 18.5 18.0 Female 12.3 12.1

API

Rates for SEER Areas

Hispanic

White NonHispanic

Black

AI/AN

All Races White

5.8 7.5 4.5

4.5 5.3 3.9

5.4 6.9 4.2

6.6 8.2 5.3

8.9 11.0 7.3

8.3 10.3 6.8

7.6 9.3 6.6

3.0 3.5 2.6

1.9 2.3 1.6

3.1 3.7 2.7

3.6 4.5 3.0

7.1 9.4 5.5

4.2 5.4 3.4

4.4 5.5 3.5

5.3 6.6 4.3

19.5 25.3 15.5

13.0 14.3 12.1

8.6 10.2 7.3

9.8 12.1 8.2

API

Hispanic

White NonHispanic

Black

AI/AN

8.8 10.8 7.3

5.7 7.2 4.7

3.8 5.0 3.0

5.8 7.7 4.4

6.7 8.4 5.4

9.0 11.2 7.4

3.8 4.7 3.2

3.7 4.7 3.0

7.6 8.4 7.1

3.2 4.2 2.5

1.9 2.3 1.6

3.3 4.0 2.8

3.8 4.8 3.0

8.0 10.8 6.1

7.5 10.1 5.8

7.9 10.6 6.0

7.1 9.3 5.7

3.6 4.6 2.9

4.9 6.3 3.8

5.5 6.7 4.6

8.0 10.9 6.1

14.7 18.1 12.2

13.3 16.2 11.2

13.3 16.2 11.3

17.9 23.1 14.3

14.6 15.8 13.7

8.6 10.6 7.1

10.4 12.7 8.9

13.2 16.0 11.3

Source: Underlying mortality data provided by NCHS (www.cdc.gov/nchs). Rates are per 100,000 and age-adjusted to the 2000 U.S. standard million by 5-year age groups. Ÿ, Statistic not displayed due to less than 16 cases. United States data are for total United States except Hispanic and white non-Hispanic, which excludes NH, OK, LA, and CT. For SEER, all races, whites, blacks are based on SEER 11; American Indian/Alaska Native on SEER 12 areas and Hispanic and white non-Hispanic are based on SEER 11 excluding CT. AI/AN, American Indian/Alaska Native; API, Asian or Pacific Islander.

more than 10% different from the U.S. rates. SEER white males have 7% lower cancer mortality than U.S. white males, with much of this difference being due to lung and laryngeal cancer. SEER includes Utah, where abstinence from tobacco and alcohol use among the large Mormon population has resulted in low death rates. For all races combined, SEER/U.S. mortality rate ratios were within ± 10% for most cancers. Some notable exceptions were stomach and liver cancers (ratio >1.10) and laryngeal and cervical cancers (ratio <0.90). Among whites, white non-Hispanics, and blacks, death rates in the SEER areas were within 10% of the corresponding national death rates for virtually all cancers. Cervical cancer was the exception; rates were lower in the SEER areas than the United States for all racial/ethnic groups except Hispanics. Relative to the United States, rates were elevated in the SEER areas for oral cavity and pharyngeal cancers among AI/AN and API populations; for melanoma and breast and brain cancers among the API group; for stomach cancer among Hispanics; and for esophageal, stomach, liver, pancreas, corpus uterine, ovarian, and kidney cancers among the AI/AN group. Rates were notably lower in the SEER areas for oral cavity and pharyngeal cancers among black males and Hispanics, for laryngeal and lung cancers among white males, and for prostate cancer and leukemia among AI/ANs. The close approximation of the death rates for SEER to the total U.S. rates for most cancers provides some confidence that the cancer experience in the SEER areas is reasonably representative of the total U.S. population and that the temporal trends are similar. A comparison of mortality trends in the United States and a subset of the SEER areas revealed quite similar patterns (Devesa et al., 1987).

derline significance (Ries et al., 2003; Weir et al., 2003). For mortality, the rates increased 0.5% per year during 1975–1990, leveled off during 1990–1994, decreased -1.4% per year during 1994–1998, and leveled off again (1998–2000). Since 1974 the 5-year relative survival rate for all sites combined increased 13.4 percentage points, from 49.6% to 63.0%. Overall cancer incidence rates (1992–2000) were higher for males than for females for each racial/ethnic group. For males, the rates (per 100,000) were highest among blacks (724.4) followed by white nonHispanics (579.3) and whites (574.3). Rates were much lower for Hispanic, API, and AI/AN males. Rates for females ranged from 229.9 for AI/AN to 435.7 per 100,000 for white non-Hispanic females (Table 9–3). The lifetime probability of developing cancer is approximately 45% for males and 39% for females. Approximately 9.6 million people were alive as of January 1, 2001 who had been diagnosed with cancer at some time in their lifetime (Ries et al., 2003). Rates and trends for all cancers combined serve primarily as an indicator of the overall cancer burden on the population and by definition are the composite of many forms of cancer. Patterns of incidence, mortality, and survival for individual malignancies may vary because of differences in demographic characteristics, the influence of screening programs, lifestyle and environmental changes, and a myriad of other factors. Therefore, the remainder of this chapter focuses on individual cancers. Incidence and death rates by organ system are presented in Tables 9–3 through 9–15 (see below). Incidence trends for 1992–2000 are in Appendix 9–A, mortality trends for 1992–2000 are in Appendix 9–B, and survival rates are presented in Appendixes 9–C and 9–D.

RESULTS During 2003, it was expected that approximately 1,334,100 malignancies would be diagnosed and 556,500 persons would die due to cancer in the United States (Jemal et al., 2003a). From 1975 through 2000, the incidence rates and death rates did not follow one specific pattern of increase or decrease. Instead, incidence rates for all cancers combined increased 0.9% per year during 1975–1983, then increased more (1.8% per year), and leveled off most recently (0.1% per year during 1995–2000). If reporting delay is taken into account, the estimated trend of 0.6% per year for 1995–2000 shows an increase of bor-

ORAL CAVITY AND PHARYNX Cancers of the oral cavity and pharynx account for approximately 2% of all cancers diagnosed in the United States each year; they were diagnosed in approximately 27,700 people in 2003, nearly two-thirds of whom were male. Over the period 1975–2000, the incidence rate dropped -0.5% per year between 1975 and 1993 and then decreased more rapidly (-2.5%) between 1993 and 2000. These overall decreasing trends were mainly driven by male rates. In contrast, for females

143

Cancer Incidence, Mortality, and Patient Survival in the United States

Table 9–3. All Cancers Combined: Cancer Incidence and Death Rates and Counts by Sex and Race/Ethnicity in SEER 12 areas: incidence and U.S. mortality, 1992–2000 All Races Parameter

Rate

Count

White

Black

AI/AN

API

Hispanic

White NonHispanic

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

482.9 574.3 424.4

1,174,111 603,942 570,169

530.9 724.4 403.5

131,602 72,780 58,822

251.5 285.1 229.9

6,415 3,062 3,353

343.6 401.3 302.0

94,564 47,865 46,699

353.7 429.7 307.8

99,505 49,522 49,983

492.0 579.3 435.7

916,091 470,525 445,566

202.9 256.3 169.3

4,216,636 2,193,576 2,023,060

263.9 369.1 201.3

545,986 294,836 251,150

137.2 170.4 115.5

15,236 7,851 7,385

128.1 160.1 104.0

69,126 37,116 32,010

137.7 178.0 111.6

160,620 85,312 75,308

205.3 258.8 171.6

3,844,951 1,997,934 1,847,017

incidence All malignant cancers Total 476.3 1,420,500 Male 574.2 736,759 Female 412.0 683,741

mortality All malignant cancers Total 206.5 4,846,984 Male 263.0 2,533,379 Female 170.7 2,313,605

Rates are per 100,000 and age-adjusted to the 2000 U.S. standard million by 5-year age group. Ÿ, Statistic not displayed due to less than 16 cases. For incidence, all sites and ovary exclude borderline tumors of the ovary. Incidence data are from SEER 12 areas except for Hispanic and white not-Hispanic which exclude Alaska, Hawaii, and Detroit. Underlying mortality data provided by NCHS (www.cdc.gov/nchs). Mortality data based on total United States except Hispanic and white non-Hispanic exclude CT, NH, LA, OK. AI/AN, American Indian/Alaska Native; API, Asian or Pacific Islander.

the incidence rates increased 2.7% per year during 1975–1980 and did not begin to decline until 1980 (-1.0%). The trends for white males were similar to the overall trends for males, and those for white females were similar to the overall trends for females. For black males, the rates increased during 1975–1987 and then declined during 1987–2000; for black females, the rates decreased (-1.7%) during 1975–2000. Comparison of short-term trends (1992–2000) showed decreases for both males and females, white males, white females, black males, Hispanic males, and white non-Hispanic males and females. The rates decreased more among males than females (Appendix 9–A). Overall these cancers are rare; the incidence rate is 11.3 per 100,000 person-years. The incidence rate is much higher, more than double, in men than in women. The incidence rate is higher in black males (21.8) than in white males (16.7) and lower but similar among white females (6.7) and black females (6.6) (Table 9–4). For males, incidence rates were lowest for Hispanic males and highest for black males. Rates for black males were 31% higher than those for white males. For females, the incidence rates ranged from 3.7 for AI/AN females to 7.0 for white non-Hispanic females. The racial/gender/temporal patterns vary by histologic type and anatomic site, suggesting etiologic distinctions among squamous cell carcinomas (SCC) of the lip, SCC of the oral cavity, SCC of the pharynx, adenocarcinomas, and Kaposi sarcoma (Canto and Devesa, 2002). The American Cancer Society estimated that 7200 people would die from cancer of the oral cavity or pharynx during 2003. Over the period 1975–2000, death rates declined for white males and females, with a more rapid decline in recent years. For black males, the rates increased substantially before the decline began in 1980. For black females, the rates were stable until 1991, when they began declining. Overall, the death rate among blacks was 75% higher than that for the white population. Similar to the incidence, the highest death rates were seen for black males (8.6), and these rates were double those for white males (4.2) (Table 9–5). The short-term mortality trend for 1992–2000 showed decreases for both sexes (-2.7% per year), males (-3.0%), and females (-2.5%) (Appendix 9–B). Survival rates have increased only recently. The 5-year relative survival rate is 57.2%. Approximately one-third of the cases are localized at diagnosis and have an 82.1% five-year survival rate. Nearly onehalf are diagnosed as regional, with a survival rate of 47.9% (Appendix 9–C). The survival rates are slightly higher among females than males and are much higher among white patients (59.7%) than black patients (36.1%). Among black patients, black males have much lower survival (30.7%) than black females (50.6%).

Cancer of the oral cavity and pharynx is not one disease but a combination of several distinct cancers. Individually, these cancers have low incidence rates, the highest being 2.6 for cancer of the tongue; the incidence rates for the remaining subsites are extremely low, ranging from 0.3 to 1.9 (Table 9–4). Survival for patients with cancer of the oral cavity or pharynx varies by subsite. Malignancies of the lip (excluding the skin of the lip) have a more favorable extent of disease distribution at diagnosis and a far better 5-year relative survival rate (94%) than any of the other subsites (Appendix 9–C). The anatomic inaccessibility and histologic complexity of certain oral and pharyngeal cancers, coupled with their infrequent occurrence, contribute to a relatively low rate of early diagnosis as well as poor survival. Cancer of the base of the tongue, for example, tends toward early, silent, and deep infiltration. Because the base of the tongue can be visualized only by indirect mirror examination, early asymptomatic lesions are rarely diagnosed (DeVita et al., 1997). Another factor contributing to the poor prognosis for this group of tumors is the extreme vascularity of the entire area, which promotes rapid development of lymph node involvement and metastases. Although each of these cancers appears infrequently, they have had different trends, with most decreasing (lip, gum, floor of the mouth).

Lip Cancer of the lip is one of the more frequently diagnosed cancers of the oral cavity, with an overall incidence rate of 1.1 and an associated 5-year relative survival rate of 94%. The skin of the lip is excluded from this section, and the SEER program does not collect basal or squamous cell carcinomas of the skin of the lip. Among whites the incidence (per 100,000) was six times higher in males (2.6) than in females (0.4). This is one of the few sites for which the rates for black males (0.2) and black females (0.1) are much less than those for white males and females, respectively. The rates for white non-Hispanics are 15% higher than those for whites, and the rates for Hispanics are less than half those for whites.

Salivary Gland Cancer of the salivary gland is relatively rare, occurring in 1.2 per 100,000 persons per year. Nearly 50% of salivary gland cancers are diagnosed while the cancer is still confined to the salivary gland (localized stage); the 5-year relative survival was 95% for localized stage and 75% for all stages combined.

144

PART II: THE MAGNITUDE OF CANCER

Table 9–4. Oral Cavity and Pharynx: Cancer Incidence Rates and Counts by Primary Site, Sex, and Race/Ethnicity in SEER 12 Areas, 1992–2000 All Races Cancer Site

Rate

Count

Oral cavity and pharynx Total 11.3 33,861 Male 16.9 22,761 Female 6.7 11,100 Lip Total 1.1 3,306 Male 2.1 2,632 Female 0.4 674 Tongue Total 2.6 7,642 Male 3.7 5,036 Female 1.6 2,606 Salivary gland Total 1.2 3,601 Male 1.6 2,029 Female 0.9 1,572 Floor of mouth Total 0.9 2,761 Male 1.4 1,840 Female 0.6 921 Gum and other mouth Total 1.9 5,708 Male 2.4 3,205 Female 1.5 2,503 Nasopharynx Total 0.8 2,380 Male 1.1 1,645 Female 0.4 735 Tonsil Total 1.2 3,726 Male 2.0 2,808 Female 0.6 918 Oropharynx Total 0.3 976 Male 0.5 718 Female 0.2 258 Hypopharynx Total 0.9 2,773 Male 1.6 2,158 Female 0.4 615 Other oral cavity and pharynx Total 0.3 988 Male 0.5 690 Female 0.2 298

White

Black

AI/AN

API

Hispanic

White NonHispanic

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

11.2 16.7 6.7

27,173 18,138 9,035

13.2 21.8 6.6

3,525 2,536 989

7.8 13.0 3.7

214 158 56

9.2 13.2 5.8

2,680 1,767 913

6.9 10.6 4.0

2,004 1,368 636

11.7 17.2 7.0

21,551 14,320 7,231

1.3 2.4 0.4

3,146 2,521 625

0.2 0.2 0.1

39 20 19

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.1 0.1 0.1

32 16 16

0.6 1.2 0.2

153 130 23

1.5 2.7 0.5

2,726 2,180 546

2.6 3.8 1.6

6,318 4,154 2,164

2.8 4.8 1.3

755 555 200

1.2 1.8 Ÿ

35 25 Ÿ

1.7 2.3 1.3

481 278 203

1.4 2.0 0.9

407 260 147

2.7 3.9 1.7

5,030 3,307 1,723

1.2 1.7 1.0

3,029 1,739 1,290

1.0 1.2 0.8

264 136 128

0.6 Ÿ Ÿ

18 Ÿ Ÿ

0.8 0.9 0.8

253 127 126

0.8 0.9 0.8

259 119 140

1.3 1.7 1.0

2,386 1,400 986

1.0 1.4 0.6

2,284 1,492 792

1.3 2.3 0.6

357 267 90

0.7 Ÿ Ÿ

18 Ÿ Ÿ

0.3 0.5 0.2

90 60 30

0.6 1.0 0.3

157 120 37

1.0 1.3 0.6

1,787 1,135 652

1.9 2.4 1.5

4,682 2,576 2,106

2.3 3.5 1.5

626 411 215

0.8 1.3 Ÿ

23 18 Ÿ

1.2 1.5 1.0

325 176 149

1.2 1.4 1.0

352 205 147

2.0 2.4 1.6

3,709 2,026 1,683

0.4 0.6 0.3

1,030 688 342

0.7 1.1 0.4

203 146 57

2.1 3.7 Ÿ

61 47 Ÿ

3.4 5.1 1.9

1,076 761 315

0.4 0.6 0.2

144 97 47

0.4 0.6 0.2

727 490 237

1.2 2.0 0.6

3,005 2,260 745

1.9 3.4 0.8

514 394 120

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.6 1.0 0.3

177 131 46

0.9 1.5 0.3

248 196 52

1.3 2.0 0.6

2,322 1,736 586

0.3 0.5 0.2

761 542 219

0.7 1.2 0.2

179 147 32

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.1 0.2 Ÿ

30 25 Ÿ

0.2 0.4 Ÿ

62 50 Ÿ

0.3 0.5 0.2

604 424 180

0.9 1.5 0.4

2,118 1,619 499

1.7 3.1 0.6

435 348 87

1.0 1.8 Ÿ

23 17 Ÿ

0.7 1.4 0.1

189 168 21

0.6 1.3 0.2

173 148 25

0.9 1.5 0.4

1,651 1,228 423

0.3 0.5 0.2

800 547 253

0.6 1.0 0.3

153 112 41

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.1 0.2 Ÿ

27 25 Ÿ

0.2 0.3 Ÿ

49 43 Ÿ

0.3 0.5 0.2

609 394 215

Rates are per 100,000 and age-adjusted to the 2000 U.S. standard million by 5-year age group. Ÿ, Statistic not displayed due to less than 16 cases. Incidence data are from SEER 12 areas except for Hispanic and white not-Hispanic which exclude Alaska, Hawaii, and Detroit. AI/AN, American Indian/Alaska Native; API, Asian or Pacific Islander.

Tongue The tongue is the most frequent primary site of oral cavity cancer. The incidence (per 100,000) is more than twice as high among males (3.7) as among females (1.6). The highest occurrence is among black males, whose incidence rate of 4.8 is nearly four times that for black females (1.3). Rates for Hispanics, AI/ANs, and APIs are lower than those for whites or blacks. The mortality rate for all races combined is 0.7 per 100,000. Slightly less than 40% of tongue cancers are confined to the tongue at diagnosis (localized), and slightly more than 40% have regional spread at the time of diagnosis. The 5-year relative survival rate for all stages combined is 53%.

Nasopharynx Cancer of the nasopharynx is a rare tumor occurring at an annual rate of 0.8 per 100,000; it is more frequent among males than among

females (1.1 and 0.4, respectively). The incidence rates are higher among American Indians (2.1) and APIs (3.4). The rate for black males is higher than that for white males but much lower than the rates for AI/AN and API males. Nearly 70% of all nasopharyngeal malignancies have regional involvement at diagnosis, with an associated 5-year relative survival rate of 57%, which is the same as the 5-year relative survival rate for all stages combined.

Other Pharyngeal Sites The incidence of cancer of the hypopharynx is higher than that of the oropharynx. Pharyngeal cancers are diagnosed at later stages than oral cancers, mainly at the regional stage, resulting in poorer overall survival rates. Cancers of the hypopharynx, less than 10% of which are diagnosed while localized, have an overall 5-year survival rate of 31%.

145

Cancer Incidence, Mortality, and Patient Survival in the United States Table 9–5. Oral Cavity and Pharynx: Cancer Death Rates and Counts by Primary Site, Sex, and Race/Ethnicity: United States, 1992–2000 All Races Cancer Site

Rate

Count

Oral cavity and pharynx Total 3.0 71,011 Male 4.6 47,025 Female 1.8 23,986 Lip Total 0 742 Male 0.1 547 Female 0 195 Tongue Total 0.7 16,045 Male 1.0 10,412 Female 0.4 5,633 Salivary gland Total 0.3 5,981 Male 0.4 3,687 Female 0.2 2,294 Floor of mouth Total 0.1 2,083 Male 0.1 1,372 Female 0.1 711 Gum and other mouth Total 0.5 12,132 Male 0.7 6,770 Female 0.4 5,362 Nasopharynx Total 0.3 6,095 Male 0.4 4,060 Female 0.2 2,035 Tonsil Total 0.2 4,970 Male 0.3 3,604 Female 0.1 1,366 Oropharynx Total 0.2 4,954 Male 0.3 3,410 Female 0.1 1,544 Hypopharynx Total 0.2 3,763 Male 0.3 2,925 Female 0.1 838 Other oral cavity and pharynx Total 0.6 14,246 Male 1.0 10,238 Female 0.3 4,008

White

Black

AI/AN

API

Hispanic

White NonHispanic

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

2.8 4.2 1.7

58,327 37,673 20,654

4.9 8.6 2.2

10,796 8,014 2,782

2.3 3.7 1.3

283 197 86

2.6 3.9 1.5

1,605 1,141 464

2.0 3.5 0.9

2,400 1,838 562

2.9 4.3 1.7

52,966 33,961 19,005

0 0.1 0

718 537 181

0 Ÿ Ÿ

Ÿ Ÿ

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0 Ÿ Ÿ

17 Ÿ Ÿ

0 0.1 0

656 486 170

0.7 1.0 0.4

13,521 8,597 4,924

1.0 1.8 0.5

2,238 1,646 592

0.4 0.5 0.3

49 26 23

0.4 0.6 0.3

237 143 94

0.5 0.8 0.2

564 419 145

0.7 1.0 0.4

12,310 7,766 4,544

0.3 0.4 0.2

5,447 3,371 2,076

0.2 0.3 0.1

436 253 183

0.2 Ÿ Ÿ

18 Ÿ Ÿ

0.1 0.2 0.1

80 52 28

0.2 0.2 0.1

182 112 70

0.3 0.4 0.2

5,005 3,121 1,884

0.1 0.1 0.1

1,715 1,095 620

0.2 0.3 0.1

342 258 84

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0 0.1 Ÿ

22 16 Ÿ

0.1 0.1 Ÿ

74 60 Ÿ

0.1 0.1 0.1

1,552 986 566

0.5 0.6 0.4

10,388 5,603 4,785

0.7 1.2 0.4

1,561 1,059 502

0.4 0.8 Ÿ

50 37 Ÿ

0.3 0.3 0.2

133 71 62

0.3 0.4 0.2

319 215 104

0.5 0.7 0.4

9,539 5,110 4,429

0.2 0.3 0.1

4,355 2,792 1,563

0.3 0.6 0.2

806 568 238

0.5 0.9 0.2

65 47 18

1.2 2.0 0.6

869 653 216

0.2 0.3 0.1

254 185 69

0.2 0.3 0.1

3,881 2,459 1,422

0.2 0.3 0.1

3,950 2,807 1,143

0.4 0.8 0.2

955 747 208

0.1 Ÿ Ÿ

17 Ÿ Ÿ

0.1 0.2 Ÿ

48 39 Ÿ

0.1 0.3 0

170 149 21

0.2 0.3 0.1

3,605 2,530 1,075

0.2 0.3 0.1

3,878 2,587 1,291

0.5 0.8 0.2

1,020 783 237

0.1 Ÿ Ÿ

18 Ÿ Ÿ

0.1 0.1 Ÿ

38 26 Ÿ

0.1 0.3 0

173 147 26

0.2 0.3 0.1

3,504 2,319 1,185

0.1 0.3 0.1

3,047 2,334 713

0.3 0.6 0.1

661 545 116

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.1 0.1 Ÿ

43 35 Ÿ

0.1 0.3 0

154 138 16

0.1 0.3 0.1

2,744 2,088 656

0.5 0.9 0.3

11,308 7,950 3,358

1.3 2.4 0.5

2,759 2,148 611

0.4 0.7 Ÿ

49 37 Ÿ

0.2 0.4 0.1

130 103 27

0.4 0.8 0.2

493 398 95

0.5 0.9 0.3

10,170 7,096 3,074

18

Rates are per 100,000 and age-adjusted to the 2000 U.S. standard million by 5-year age group. Ÿ, Statistic not displayed due to less than 16 cases. Underlying mortality data provided by NCHS (www.cdc.gov/nchs). Mortality data based on total United States except Hispanic and white non-Hispanic exclude CT, NH, LA, OK. AI/AN, American Indian/Alaska Native; API, Asian or Pacific Islander.

DIGESTIVE SYSTEM Malignancies of the digestive system currently account for nearly 20% of all cancers diagnosed in the United States and 24% of the cancer deaths. Of these cases, 58% arise in the colon and rectum, 12% in the pancreas, and 9% in the stomach (Jemal et al., 2003a). Five-year relative survival rates vary from 4% for pancreatic cancer to more than 60% for colorectal cancer. Even when they are diagnosed while still localized, cancers of the esophagus, liver, and pancreas have 5-year relative survival rates of only 29%, 16%, and 17%, respectively.

Esophagus Cancer of the esophagus is a rare malignancy with an overall incidence rate of 4.5 per 100,000 persons, occurring three times more fre-

quently in males than females (Table 9–6). The rate for white males (7.3) is considerably lower than the rate for black males (12.9). Rates are lower among Hispanics and other racial/ethnic groups than among non-Hispanics. The incidence and death rates have been declining among blacks, are relatively stable among white females, and are increasing among white males (Brown and Devesa, 2002). Rates for squamous cell carcinomas, the most frequent cell type in the past, are highest among blacks and have been declining, related to decreases in cigarette smoking and hard liquor consumption and to increased intake of fresh fruits and vegetables. Adenocarcinoma rates have been rising, especially among white males, such that esophageal adenocarcinomas now are more frequent than squamous cell carcinomas; the increases appear related to increasing obesity and gastroesophageal reflux disease (Brown and Devesa, 2002). Survival is poor, with only 14% of patients surviving 5 years or more.

Table 9–6. Digestive System: Cancer Incidence Rates and Counts by Primary Site, Sex, and Race/Ethnicity in SEER 12 Areas, 1992–2000 All Races Cancer Site

Rate

Count

White Rate

Digestive system Total 92.0 269,843 88.7 Male 114.1 142,278 109.7 Female 75.4 127,565 72.9 Esophagus Total 4.5 13,220 4.3 Male 7.5 9,678 7.3 Female 2.1 3,542 2.0 Stomach Total 9.3 27,411 7.9 Male 13.5 16,733 11.7 Female 6.3 10,678 5.2 Small intestine Total 1.6 4,800 1.6 Male 2.0 2,531 1.9 Female 1.4 2,269 1.3 Colon and rectum Total 54.5 159,599 54.2 Male 65.0 80,289 64.9 Female 46.8 79,310 46.4 Colon excluding rectum Total 39.6 115,436 39.3 Male 45.6 55,607 45.6 Female 35.2 59,829 34.7 Cecum Total 9.2 26,800 9.4 Male 9.8 11,709 10.0 Female 8.8 15,091 8.9 Appendix Total 0.4 1,259 0.4 Male 0.4 608 0.5 Female 0.4 651 0.4 Ascending colon Total 6.0 17,403 6.0 Male 6.6 7,845 6.6 Female 5.6 9,558 5.6 Hepatic flexure Total 2.2 6,470 2.2 Male 2.6 3,090 2.6 Female 2.0 3,380 2.0 Transverse colon Total 3.6 10,562 3.6 Male 3.9 4,742 4.0 Female 3.4 5,820 3.4 Splenic flexure Total 1.5 4,484 1.5 Male 2.0 2,409 1.9 Female 1.2 2,075 1.2 Descending colon Total 2.4 7,103 2.3 Male 3.0 3,752 2.9 Female 2.0 3,351 1.9 Sigmoid colon Total 12.1 35,404 11.9 Male 14.9 18,667 14.8 Female 10.0 16,737 9.7 Large intestine, NOS Total 2.1 5,951 2.0 Male 2.4 2,785 2.3 Female 1.8 3,166 1.8 Rectum and rectosigmoid junction Total 15.0 44,163 14.9 Male 19.4 24,682 19.3 Female 11.6 19,481 11.6 Rectosigmoid junction Total 5.0 14,686 5.0 Male 6.4 8,124 6.4 Female 3.9 6,562 3.9 Rectum Total 10.0 29,477 10.0 Male 13.0 16,588 12.9 Female 7.7 12,919 7.7

146

Black

AI/AN

API

White NonHispanic

Hispanic

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

214,166 112,054 102,112

115.7 143.8 96.2

27,292 14,013 13,279

72.0 84.6 62.3

1,679 883 796

97.9 125.4 75.8

25,804 14,836 10,968

82.8 104.4 67.4

20,481 11,067 9,414

87.9 108.1 72.6

165,526 86,031 79,495

10,354 7,613 2,741

7.9 12.9 4.4

1,972 1,349 623

3.1 5.5 1.3

68 52 16

3.0 5.4 1.1

794 639 155

3.2 6.1 1.2

783 633 150

4.4 7.3 2.1

8,174 5,937 2,237

19,150 11,852 7,298

14.0 20.3 9.8

3,241 1,908 1,333

10.5 14.4 7.5

258 159 99

18.2 24.2 13.5

4,693 2,769 1,924

13.8 19.0 10.1

3,419 1,980 1,439

7.0 10.5 4.4

13,207 8,318 4,889

3,795 1,997 1,798

2.8 3.5 2.3

697 363 334

0.8 Ÿ Ÿ

22 Ÿ Ÿ

1.0 1.2 0.8

268 149 119

1.1 1.3 1.0

299 150 149

1.6 1.9 1.3

2,957 1,567 1,390

130,947 65,789 65,158

62.9 73.2 56.1

14,693 6,950 7,743

35.8 40.5 32.0

839 430 409

47.3 57.6 39.1

12,472 6,776 5,696

39.2 48.4 32.6

9,638 5,090 4,548

55.2 65.4 47.5

104,028 51,825 52,203

94,877 45,706 49,171

48.7 55.1 44.7

11,237 5,102 6,135

26.0 28.7 23.9

593 292 301

32.0 37.1 28.0

8,311 4,291 4,020

26.9 31.9 23.4

6,479 3,256 3,223

40.3 46.4 35.8

76,164 36,412 39,752

22,589 9,851 12,738

12.3 13.0 11.8

2,802 1,200 1,602

5.2 4.3 5.8

119 50 69

4.9 5.2 4.7

1,237 586 651

6.6 7.0 6.3

1,514 688 826

9.5 10.0 9.1

18,073 7,782 10,291

1,050 511 539

0.4 0.5 0.4

114 54 60

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.3 0.3 0.3

85 40 45

0.3 0.3 0.4

100 39 61

0.4 0.5 0.4

822 406 416

14,460 6,494 7,966

7.5 8.4 7.0

1,717 761 956

3.1 3.0 3.2

69 30 39

4.3 4.6 4.1

1,107 533 574

3.8 4.6 3.3

876 439 437

6.1 6.7 5.7

11,671 5,199 6,472

5,350 2,546 2,804

2.6 2.9 2.3

594 275 319

2.1 2.3 2.0

46 20 26

1.8 2.1 1.6

461 243 218

1.5 1.8 1.2

345 179 166

2.3 2.7 2.1

4,386 2,061 2,325

8,790 3,947 4,843

4.2 4.8 4.0

974 421 553

2.2 1.7 2.5

46 17 29

2.8 2.9 2.7

716 336 380

2.0 2.1 1.9

494 230 264

3.8 4.1 3.5

7,162 3,190 3,972

3,591 1,944 1,647

2.5 3.1 2.1

582 296 286

1.0 Ÿ Ÿ

22 Ÿ Ÿ

1.1 1.3 0.8

282 154 128

1.0 1.3 0.8

243 139 104

1.5 2.0 1.2

2,920 1,575 1,345

5,515 2,939 2,576

3.7 4.5 3.2

861 412 449

1.1 1.4 Ÿ

30 17 Ÿ

2.5 3.2 2.1

673 368 305

1.4 1.7 1.2

360 187 173

2.3 3.0 1.9

4,388 2,352 2,036

28,634 15,215 13,419

12.5 14.2 11.4

2,929 1,356 1,573

10.1 13.7 7.4

236 138 98

13.0 16.0 10.6

3,435 1,872 1,563

8.8 11.1 7.2

2,217 1,178 1,039

12.1 15.1 10.0

22,807 12,081 10,726

4,898 2,259 2,639

3.0 3.7 2.5

664 327 337

1.1 Ÿ Ÿ

22 Ÿ Ÿ

1.3 1.5 1.2

315 159 156

1.5 1.9 1.2

330 177 153

2.1 2.3 1.8

3,935 1,766 2,169

36,070 20,083 15,987

14.2 18.1 11.4

3,456 1,848 1,608

9.8 11.8 8.1

246 138 108

15.2 20.4 11.1

4,161 2,485 1,676

12.2 16.5 9.1

3,159 1,834 1,325

14.9 19.1 11.7

27,864 15,413 12,451

12,030 6,616 5,414

5.0 6.4 4.1

1,202 630 572

2.6 2.7 2.5

69 37 32

4.9 6.8 3.4

1,332 814 518

4.0 5.5 3.0

1,026 595 431

4.8 6.0 3.9

9,001 4,894 4,107

24,040 13,467 10,573

9.2 11.7 7.3

2,254 1,218 1,036

7.2 9.1 5.7

177 101 76

10.3 13.6 7.7

2,829 1,671 1,158

8.2 11.0 6.2

2,133 1,239 894

10.1 13.0 7.8

18,863 10,519 8,344

147

Cancer Incidence, Mortality, and Patient Survival in the United States Table 9–6. (cont.) All Races Cancer Site

Rate

Count

Anus, anal canal, anorectum Total 1.3 3,814 Male 1.2 1,611 Female 1.3 2,203 Liver and intrahepatic bile duct Total 5.5 16,170 Male 8.3 10,871 Female 3.2 5,299 Liver Total 4.6 13,646 Male 7.2 9,531 Female 2.5 4,115 Intrahepatic bile duct Total 0.9 2,524 Male 1.1 1,340 Female 0.7 1,184 Gallbladder Total 1.3 3,763 Male 0.9 1,049 Female 1.6 2,714 Other biliary Total 1.5 4,391 Male 1.9 2,255 Female 1.3 2,136 Pancreas Total 11.1 32,492 Male 12.7 15,726 Female 9.9 16,766 Retroperitoneum Total 0.4 1,342 Male 0.5 651 Female 0.4 691 Peritoneum, omentum, mesentery Total 0.5 1,626 Male 0.2 286 Female 0.8 1,340 Other digestive organs Total 0.4 1,215 Male 0.5 598 Female 0.4 617

White Rate

Black

Count

AI/AN

API

White NonHispanic

Hispanic

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

1.3 1.2 1.4

3,210 1,305 1,905

1.5 1.7 1.3

402 215 187

0.6 Ÿ Ÿ

16 Ÿ Ÿ

0.6 0.5 0.6

153 65 88

1.1 1.0 1.1

290 130 160

1.4 1.2 1.5

2,579 1,033 1,546

4.4 6.7 2.6

10,589 7,016 3,573

6.4 10.2 3.6

1,607 1,111 496

6.8 8.0 5.8

165 91 74

13.6 20.6 7.7

3,773 2,626 1,147

8.5 12.9 5.1

2,195 1,488 707

3.8 5.8 2.2

7,092 4,680 2,412

3.5 5.6 1.9

8,552 5,935 2,617

5.9 9.4 3.2

1,477 1,034 443

5.2 6.3 4.3

133 77 56

12.3 19.1 6.6

3,450 2,460 990

7.4 11.7 4.0

1,926 1,361 565

3.0 4.7 1.6

5,569 3,854 1,715

0.8 1.1 0.7

2,037 1,081 956

0.6 0.8 0.4

130 77 53

1.6 Ÿ 1.5

32 Ÿ 18

1.3 1.4 1.2

323 166 157

1.1 1.2 1.0

269 127 142

0.8 1.0 0.6

1,523 826 697

1.2 0.8 1.6

2,975 780 2,195

1.3 0.8 1.6

301 83 218

3.7 3.3 4.1

72 24 48

1.6 1.4 1.8

410 160 250

2.9 1.3 4.0

699 134 565

1.0 0.7 1.2

1,899 531 1,368

1.5 1.8 1.2

3,507 1,797 1,710

1.2 1.6 1.0

284 142 142

2.3 2.6 2.1

47 21 26

2.1 2.6 1.8

547 294 253

2.0 2.4 1.8

485 242 243

1.3 1.7 1.1

2,538 1,298 1,240

10.8 12.4 9.6

26,106 12,627 13,479

16.4 18.4 14.8

3,771 1,764 2,007

7.5 7.8 7.2

172 80 92

9.4 11.1 8.2

2,396 1,238 1,158

9.9 10.8 9.3

2,321 1,081 1,240

10.8 12.4 9.4

20,274 9,851 10,423

0.4 0.5 0.4

1,084 536 548

0.5 0.4 0.5

133 55 78

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.4 0.4 0.4

121 58 63

0.4 0.4 0.4

135 59 76

0.4 0.5 0.4

799 407 392

0.6 0.2 0.9

1,453 251 1,202

0.3 0.2 0.5

82 16 66

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.3 0.1 0.4

83 17 66

0.4 0.2 0.6

123 30 93

0.6 0.2 1.0

1,170 189 981

0.4 0.5 0.4

996 491 505

0.5 0.6 0.4

109 57 52

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.4 0.4 0.4

94 45 49

0.4 0.5 0.3

94 50 44

0.4 0.5 0.4

809 395 414

Rates are per 100,000 and age-adjusted to the 2000 U.S. standard million by 5-year age group. Ÿ, Statistic not displayed due to less than 16 cases. For incidence, all sites and ovary exclude borderline tumors of the ovary. Incidence data are from SEER 12 areas except for Hispanic and white not-Hispanic which exclude Alaska, Hawaii, and Detroit. AI/AN, American Indian/Alaska Native; API, Asian or Pacific Islander.

Stomach

Colon and Rectum

Stomach cancer occurs more than twice as often among males as among females (13.5 and 6.3 per 100,000, respectively). Rates among APIs are more than twice those among whites, and risks are also elevated among blacks and Hispanics compared to non-Hispanic whites. Incidence and death rates have been declining for many years, as shown for whites and blacks in Figure 9–1. During recent years, the overall incidence rate was 9.3 (Table 9–6), the mortality rate was 5.1 (Table 9–7), and the 5-year relative survival rate was 23%. About one-third of patients have distant disease at diagnosis, for whom the 5-year relative survival is less than 3%. Reduced smoking, improved diet, and reduced Helicobacter pylori prevalence probably have contributed to the consistent reductions observed for stomach cancer (Brown and Devesa, 2002). There is some evidence that tumors arising in the gastric cardia may be etiologically more similar to adenocarcinomas of the distal esophagus than to adenocarcinomas of the distal stomach.

Accounting for 147,500 new cases each year in the United States, cancers of the colon and rectum combined constitute the most frequently diagnosed malignancy after cancers of the prostate, breast, and lung. Colorectal cancer incidence among whites increased until the mid-1980s and then subsequently decreased; the death rates have been declining for years. Among blacks, incidence rates have not changed greatly, and death rates have declined since 1990. As a result, the racial relative risks have shifted from a white to a black excess (Fig. 9–2). The incidence rate for blacks is now about 16% higher (62.9) than that for whites (54.2), and the mortality rate is 34% higher based on rates of 28.9 and 21.6, respectively. Rates among the other racial/ethnic groups are 60%–70% those among whites. Colorectal cancer incidence is 39% higher among males (65.0) than females (46.8 per 100,000). More than 70% of colorectal cancers arise in the colon and fewer than 30% in the rectum. Of the colon cancers, more than half occur in

148

PART II: THE MAGNITUDE OF CANCER

Females

Rate per 100,000 person-years

Males 400

400

100

100

White incidence Black incidence White mortality Black mortality

10

10

1 1975 1980 1985 1990 1995 2000

Figure 9–1. Stomach cancer incidence (SEER9 areas) and mortality (United States) rates, ageadjusted using the 2000 U.S. standard, by gender and race, 1975–2000.

Year

1 1975 1980 1985 1990 1995 2000

Year

Symbols present observed rates Incidence lines estimated by delay-adjusted joinpoint regression Mortality lines estimated by joinpoint regression

the sigmoid (31%) or cecum (23%), and another 15% occur in the ascending colon. The hepatic flexure, transverse flexure, splenic flexure, and descending colon account for 6%, 9%, 4%, and 6%, respectively. Rectal cancers occur twice as frequently as those in the rectosigmoid. The black/white incidence ratios decreased from 1.31 for cecal cancer to 1.25, 1.17, 1.05, 1.00, and 0.92 for ascending, transverse, sigmoid, rectosigmoid, and rectal cancers, respectively. The API to white incidence rate ratios were virtually the reverse pattern, rising from 0.52 to 0.72, 0.78, 1.09, 0.98, and 1.03, respectively. Rates among Hispanics consistently were lower than among white non-Hispanics, with rate ratios of 0.6–0.7 in the colon and 0.8 in the rectum. The male/female incidence ratios increased from the proximal to the distal colon, with ratios less than 1.2 for cancers of the cecum and the ascending and transverse colon, about 1.5 for descending colon and sigmoid cancers, and more than 1.6 for rectosigmoid and rectal cancers. The temporal trends in incidence also have varied by subsite of the colon, with declines in rectal cancer, especially among whites, and increases in proximal colon cancer rates among blacks (Devesa and Chow, 1993; Troisi et al., 1999). Although screening and removal of precancerous lesions may be affecting the rates, it is not clear what other factors are operating,

especially in the emerging excess of proximal colon cancer among blacks. More than one-third of colorectal cancers are diagnosed while still localized, and about 20% have distant disease. Patients with colorectal cancer have an overall 5-year relative survival rate of about 62%, but it varies from about 90% to less than 10%, depending on the stage at diagnosis. Separating death rates for cancer of the colon from those of the rectum is inappropriate because death rates for rectal cancer are underestimated. During death certification, colon cancer is often designated as the underlying cause of death when the hospital diagnosis was rectal cancer (Percy et al., 1981). The impact of this misclassification has changed over time (Chow and Devesa, 1992).

Liver Primary cancers of the liver and intrahepatic bile duct are rare, with an overall incidence rate of only 5.5 per 100,000. The category includes hepatocellular carcinomas (HCC) (4.6) and, to a lesser extent, intrahepatic bile duct cancers (0.9). Liver and intrahepatic bile duct cancer incidence rates among APIs are more than threefold, and those

Table 9–7. Digestive System: Cancer Death Rates and Counts by Primary Site, Sex, and Race/Ethnicity: United States, 1992–2000 All Races Cancer Site

Rate

Count

Digestive system Total 48.8 1,143,084 Male 62.9 606,912 Female 38.6 536,172 Esophagus Total 4.3 100,893 Male 7.6 76,181 Female 1.8 24,712 Stomach Total 5.1 119,403 Male 7.3 70,257 Female 3.5 49,146 Small intestine Total 0.4 9,416 Male 0.5 4,967 Female 0.3 4,449 Colon and rectum Total 22.0 514,515 Male 26.9 254,235 Female 18.6 260,280 Anus, anal canal, anorectum Total 0.2 3,860 Male 0.1 1,474 Female 0.2 2,386 Liver and intrahepatic bile duct Total 4.4 102,432 Male 6.3 63,255 Female 2.9 39,177 Liver Total 3.5 82,403 Male 5.3 53,442 Female 2.1 28,961 Intrahepatic bile duct Total 0.9 20,029 Male 1.0 9,813 Female 0.7 10,216 Gallbladder Total 0.8 18,456 Male 0.5 5,154 Female 1.0 13,302 Other biliary Total 0.6 14,953 Male 0.8 7,035 Female 0.6 7,918 Pancreas Total 10.6 247,790 Male 12.3 119,799 Female 9.2 127,991 Retroperitoneum Total 0.1 2,391 Male 0.1 1,191 Female 0.1 1,200 Peritoneum, omentum, mesentery Total 0.2 3,753 Male 0.1 841 Female 0.2 2,912 Other digestive organs Total 0.2 5,222 Male 0.3 2,523 Female 0.2 2,699

White

Black

AI/AN

API

Hispanic

White NonHispanic

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

46.8 60.3 36.9

972,257 515,538 456,719

69.7 91.8 54.8

141,750 74,684 67,066

37.1 45.3 30.9

3,991 2,135 1,856

47.1 60.9 36.3

25,086 14,555 10,531

42.5 55.3 33.3

47,669 26,629 21,040

46.6 60.0 36.8

876,321 463,317 413,004

4.0 7.1 1.6

83,070 63,162 19,908

7.7 13.4 3.6

16,348 11,885 4,463

2.6 4.7 1.1

303 239 64

2.2 3.8 1.0

1,172 895 277

2.6 4.7 1.0

2,916 2,311 605

4.1 7.2 1.6

76,143 57,823 18,320

4.5 6.5 3.1

93,537 55,345 38,192

9.9 14.6 6.8

19,914 11,540 8,374

5.6 7.5 4.2

618 355 263

10.3 13.3 7.9

5,334 3,017 2,317

7.6 10.4 5.6

8,710 5,042 3,668

4.3 6.2 2.9

80,178 47,578 32,600

0.4 0.5 0.3

7,981 4,246 3,735

0.6 0.7 0.5

1,252 634 618

0.2 Ÿ Ÿ

25 Ÿ Ÿ

0.3 0.3 0.3

158 75 83

0.3 0.4 0.2

367 211 156

0.4 0.5 0.3

7,199 3,798 3,401

21.6 26.5 18.1

447,947 222,222 225,725

28.9 35.2 25.0

58,014 27,506 30,508

14.0 16.4 12.2

1,489 762 727

13.7 16.6 11.5

7,065 3,745 3,320

14.2 18.2 11.5

15,664 8,466 7,198

21.8 26.7 18.4

410,114 202,829 207,285

0.2 0.1 0.2

3,378 1,253 2,125

0.2 0.2 0.2

436 201 235

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.1 Ÿ 0.1

Ÿ

38 23

0.1 0.1 0.1

142 57 85

0.2 0.1 0.2

3,107 1,151 1,956

4.0 5.8 2.7

82,815 50,478 32,337

5.9 9.1 3.7

12,573 8,000 4,573

5.5 7.5 4.1

611 365 246

10.9 16.1 6.7

6,433 4,412 2,021

6.9 10.0 4.6

8,015 5,132 2,883

3.8 5.4 2.5

70,665 42,776 27,889

3.1 4.8 1.9

65,196 41,891 23,305

5.2 8.2 3.1

11,092 7,296 3,796

4.2 5.8 2.9

470 296 174

9.4 14.2 5.5

5,645 3,959 1,686

5.9 9.0 3.5

6,848 4,628 2,220

2.9 4.4 1.8

54,959 35,068 19,891

0.8 1.0 0.7

17,619 8,587 9,032

0.7 0.9 0.6

1,481 704 777

1.4 1.7 1.2

141 69 72

1.5 1.9 1.2

788 453 335

1.0 1.1 1.0

1,167 504 663

0.8 1.0 0.7

15,706 7,708 7,998

0.8 0.5 0.9

16,087 4,497 11,590

0.9 0.6 1.0

1,737 452 1,285

1.5 1.1 1.9

149 38 111

1.0 0.8 1.1

483 167 316

1.4 0.9 1.9

1,613 385 1,228

0.7 0.5 0.9

13,766 3,895 9,871

0.6 0.8 0.6

13,525 6,402 7,123

0.5 0.5 0.5

957 403 554

0.9 0.9 1.0

87 33 54

0.8 0.9 0.7

384 197 187

0.7 0.8 0.6

718 323 395

0.6 0.8 0.6

12,170 5,774 6,396

10.3 12.0 8.9

213,907 103,966 109,941

14.6 16.8 13.0

29,375 13,571 15,804

6.3 6.4 6.1

665 309 356

7.6 8.7 6.7

3,843 1,953 1,890

8.2 9.4 7.3

9,036 4,488 4,548

10.3 12.0 9.0

193,915 94,125 99,790

0.1 0.1 0.1

2,111 1,063 1,048

0.1 0.1 0.1

239 104 135

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.1 0.1 Ÿ

35 21

0.1 0.1 0.1

124 74 50

0.1 0.1 0.1

1,895 950 945

0.2 0.1 0.2

3,433 764 2,669

0.1 0.1 0.2

264 68 196

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.1 Ÿ 0.1

44 37

0.1 0 0.1

131 26 105

0.2 0.1 0.2

3,167 708 2,459

0.2 0.3 0.2

4,466 2,140 2,326

0.3 0.4 0.3

641 320 321

0.2 Ÿ Ÿ

18 Ÿ Ÿ

0.2 0.2 0.2

97 51 46

0.2 0.3 0.2

233 114 119

0.2 0.3 0.2

4,002 1,910 2,092

Ÿ Ÿ

Rates are per 100,000 and age-adjusted to the 2000 U.S. standard million by 5-year age group. Ÿ, Statistic not displayed due to less than 16 cases. Underlying mortality data provided by NCHS (www.cdc.gov/nchs). Mortality data based on total United States except Hispanic and white non-Hispanic exclude CT, NH, LA, OK. AI/AN, American Indian/Alaska Native; API, Asian or Pacific Islander.

149

150

PART II: THE MAGNITUDE OF CANCER

Females

Rate per 100,000 person-years

Males 400

400

100

100

White incidence Black incidence White mortality Black mortality

10

10

1 1975 1980 1985 1990 1995 2000

Figure 9–2. Colorectal cancer incidence (SEER9 areas) and mortality (United States) rates, ageadjusted using the 2000 U.S. standard, by gender and race, 1975–2000.

Year

1 1975 1980 1985 1990 1995 2000

Year

Symbols present observed rates Incidence lines estimated by delay-adjusted joinpoint regression Mortality lines estimated by joinpoint regression

among Hispanics twofold those among whites due particularly to elevated liver cancer rates, although bile duct cancer rates also are higher. Among APIs, rates have been notably elevated among Chinese and Japanese for many years (McGlynn et al., 2001). Rates among blacks are elevated relative to those of whites for liver cancer but are lower for intrahepatic bile duct cancer. The incidence rate is more than twice as high among males (8.3) as among females (3.2). The 5-year relative survival rate for all stages is only 7% but reaches 16% for localized tumors. Reported death rates for liver cancer overestimate the true rates because many deaths attributed to liver cancer on the death certificate may in reality be due to cancers that have metastasized to the liver (Percy et al., 1990b).

Pancreas Cancer of the pancreas accounts for about 30,000 deaths annually in the United States. The incidence rate among blacks (16.4 per 100,000) is more than 50% higher than that among whites (10.8). Rates among APIs and Hispanics are slightly lower than among whites. The malignancy occurs more frequently in males than in females: 12.7 and 9.9, respectively. The 5-year relative survival rate of 4%, which is the

poorest for any malignancy, has remained unchanged for years. Even when the disease is diagnosed while still localized, accounting for fewer than 8% of cases, the 5-year survival rate is only 17%. Because of such poor survival, the death rates are only slightly lower than the incidence rates; mortality figures are 12.3 for males and 9.2 per 100,000 for females.

RESPIRATORY SYSTEM Lung In 2003 lung cancer was diagnosed in approximately 171,900 people: 91,800 males and 80,100 females. For males, lung cancer incidence is second in frequency only to cancer of the prostate; and for females, it is second only to cancer of the breast. In 2003 there were 157,200 deaths (88,400 males, 68,800 females) attributed to lung cancer, which is more than to any other cancer. Lung cancer has been the leading cause of cancer death among U.S. men since the mid-1950s, whereas it surpassed breast cancer as the leading form among women during the late 1980s (Jemal et al., 2003b).

Cancer Incidence, Mortality, and Patient Survival in the United States

Females

Males

Rate per 100,000 person-years

151

400

400

100

100

White incidence Black incidence White mortality Black mortality

10

10

1 1975 1980 1985 1990 1995 2000

1 1975 1980 1985 1990 1995 2000

Year Symbols present observed rates Incidence lines estimated by delay-adjusted joinpoint regression Mortality lines estimated by joinpoint regression

After increasing for many years, the incidence of lung cancer among males peaked during the mid-1980s and subsequently has decreased more than 20% among both whites and blacks (Fig. 9–3). National death rates peaked somewhat later, around 1990. The incidence and mortality among white and black females more than doubled between 1975 and 2000, with increases more rapid during the 1970s and 1980s than during the 1990s; recent incidence rates among white females may be decreasing. Among all four race/sex groups, rates have declined among the young and middle-aged, in contrast to continuing increases among the elderly, with the changing trends most striking for squamous cell carcinoma (Devesa et al., 1989, 1991; Jemal et al., 2003b). Among males, lung cancer incidence and death rates among blacks have exceeded those among whites by 30%–50% for many years, in contrast to similar rates for white and black females. Rates for APIs, AI/ANs, and Hispanics are all 50%–70% of those for whites of the same gender. Declines in rates during the 1990s were suggested among males but not among females of each of these racial/ethnic groups. Male/female rate ratios therefore have been declining, but they remain about at 2.0 or more for all groups except whites, where the incidence

Year

Figure 9–3. Lung and bronchus cancer incidence (SEER9 areas) and mortality (United States) rates, age-adjusted using the 2000 U.S. standard, by gender and race, 1975–2000.

rates are about 60% higher among males than females. Among whites, the male/female mortality rate ratios were highest (9.3) for those aged 65–69 years during 1960–1964 (born around 1895) and have declined to 1.1 among those aged 25–29 years during 1995–1999 (born around 1970) (Jemal et al., 2003b). Pronounced geographic variation in lung cancer death rates in the United States has been described, with rates among white males shifting from excesses in the urban areas of the northeastern and north central states and in areas along the southeastern and gulf coast to recent clustering across the southeastern and south central areas (Devesa et al., 1999). These patterns were not apparent among white females, although consistently low rates were seen in the mountain and plains states for both sexes. Rates among blacks were consistently elevated in northern areas and low across the south. Survival following lung cancer is generally poor: only 15% at 5 years. For the 16% for whom the cancer has not spread beyond the lung (i.e., localized stage), the 5-year relative survival rate is about 50%. More than one-third of lung cancers are diagnosed at the regional and distant stages; survival rates drop to 16% and 2% for these two groups, respectively.

152

PART II: THE MAGNITUDE OF CANCER

Larynx Although less than 1/15th as frequent as lung cancer, cancer of the larynx is the second most frequently diagnosed malignancy in the respiratory system, with an overall incidence rate of 4.1 per 100,000. The incidence rate for males is more than four times that for females: 7.4 versus 1.6 (Table 9–8). The highest incidence, 13.0, is found among black males. Because 50% of laryngeal cancers are diagnosed while localized, the 5-year relative survival rate is 65% percent for all stages

combined, considerably better than that for lung cancer; for patients diagnosed with localized disease, the survival rate exceeds 80%.

Mesothelioma Incidence rates for all mesotheliomas in white males nearly doubled from 1977–1978 to 1991–1992 (from 1.3 to 2.5 per 100,000), whereas the rate for white females remained at about 0.4. Rates for males aged 75–84 years tripled from 6.3 to 18.2 per 100,000 between 1977–1978

Table 9–8. Respiratory System: Cancer Incidence and Death Rates and Counts by Primary Site, Sex, and Race/Ethnicity in SEER 12 Areas for Incidence and in U.S. for Mortality, 1992–2000 All Races Cancer Site

Rate

Count

White Rate

Black

AI/AN

API

White NonHispanic

Hispanic

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

170,419 97,335 73,084

91.8 141.6 57.6

22,300 14,177 8,123

40.1 56.5 27.5

922 574 348

46.6 68.2 29.5

12,379 8,050 4,329

38.9 57.0 26.3

9,468 5,804 3,664

72.0 94.3 56.0

134,916 76.510 58,406

1,710 970 740

0.8 1.1 0.6

206 114 92

1.0 Ÿ Ÿ

23 Ÿ Ÿ

0.6 0.9 0.4

185 116 69

0.7 0.9 0.5

199 114 85

0.7 0.9 0.5

1,264 716 548

9,920 7,859 2,061

7.0 13.0 2.7

1,758 1,374 384

1.3 2.1 Ÿ

33 25 Ÿ

1.9 3.6 0.5

506 434 72

2.8 5.6 0.8

756 633 123

4.1 7.2 1.6

7,601 5,994 1,607

155,692 86,087 69,605

83.1 126.0 53.9

20,113 12,532 7,581

37.0 51.7 25.8

849 522 327

43.5 62.7 28.4

11,518 7,370 4,148

34.2 48.3 24.5

8,168 4,780 3,388

66.0 83.9 53.4

123,672 67,934 55,738

2,519 2,041 478

0.6 1.2 0.3

152 112 40

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.4 0.7 0.2

105 81 24

1.0 1.9 0.4

230 180 50

1.1 2.1 0.3

2,003 1,631 372

578 378 200

0.2 0.3 0.2

71 45 26

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.2 0.3 0.1

65 49 16

0.2 0.4 0.1

115 97 18

0.2 0.3 0.1

376 235 141

1,237,788 743,889 493,899

71.5 118.9 40.6

149,896 99,522 50,374

38.3 54.6 26.6

4,232 2,564 1,668

29.9 43.4 19.5

15,690 9,968 5,722

27.1 44.2 15.1

30,061 20,525 9,536

60.7 85.6 43.1

1,143,593 684,441 459,152

3,592 2,034 1,558

0.2 0.4 0.2

535 328 207

0.2 Ÿ Ÿ

16 Ÿ Ÿ

0.1 0.2 0.1

83 60 23

0.1 0.2 0.1

175 104 71

0.2 0.2 0.1

3,238 1,813 1,425

28,236 22,117 6,119

3.0 5.9 1.0

6,435 5,220 1,215

1.1 1.9 0.5

118 88 30

0.5 0.9 0.1

246 206 40

1.2 2.5 0.2

1,320 1,174 146

1.4 2.5 0.5

25,390 19,761 5,629

1,200,254 715,715 484,539

68.1 112.2 39.3

142,406 93,632 48,774

36.9 52.5 25.9

4,087 2,463 1,624

29.2 42.1 19.2

15,287 9,651 5,636

25.6 41.2 14.7

28,314 19,052 9,262

58.8 82.4 42.2

1,109,794 659,233 450,561

3,423 2,664 759

0.1 0.2 0.0

177 126 51

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0 0.1 Ÿ

25 19

0.1 0.2 0.0

130 106 24

0.2 0.3 0.1

3,150 2,445 705

2,283 1,359 924

0.2 0.2 0.1

343 216 127

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.1 0.1 0.1

49 32 17

0.1 0.1 0.1

122 89 33

0.1 0.1 0.1

2,021 1,189 832

incidence Respiratory system Total 70.0 206,408 70.5 Male 95.0 120,373 93.6 Female 51.9 86,035 54.0 Nose, nasal cavity, middle ear Total 0.7 2,141 0.7 Male 0.9 1,225 0.9 Female 0.5 916 0.5 Larynx Total 4.1 12,269 4.1 Male 7.4 9,737 7.3 Female 1.6 2,532 1.6 Lung and bronchus Total 64.0 188,482 64.4 Male 84.5 106,684 83.1 Female 49.4 81,798 51.4 Pleura Total 1.0 2,798 1.0 Male 1.9 2,251 2.0 Female 0.3 547 0.3 Trachea, mediastinum, other respiratory Total 0.2 718 0.2 Male 0.3 476 0.3 Female 0.1 242 0.1

mortality Respiratory system Total 59.5 1,407,606 59.2 Male 85.9 855,943 84.0 Female 41.0 551,663 41.7 Nose, nasal cavity, middle ear Total 0.2 4,226 0.2 Male 0.2 2,430 0.2 Female 0.1 1,796 0.1 Larynx Total 1.5 35,035 1.4 Male 2.8 27,631 2.5 Female 0.6 7,404 0.5 Lung and bronchus Total 57.6 1,362,034 57.4 Male 82.4 821,461 80.8 Female 40.2 540,573 40.9 Pleura Total 0.2 3,633 0.2 Male 0.3 2,814 0.3 Female 0.1 819 0.1 Trachea, mediastinum, other respiratory Total 0.1 2,678 0.1 Male 0.2 1,607 0.2 Female 0.1 1,071 0.1

Rates are per 100,000 and age-adjusted to the 2000 U.S. standard million by 5-year age group. Ÿ, Statistic not displayed due to less than 16 cases. Incidence data are from SEER 12 areas except for Hispanic and white not-Hispanic which exclude Alaska, Hawaii, and Detroit. Underlying mortality data provided by NCHS (www.cdc.gov/nchs). Mortality data based on total United States except Hispanic and white non-Hispanic exclude CT, NH, LA, OK. AI/AN, American Indian/Alaska Native; API, Asian or Pacific Islander.

Ÿ

153

Cancer Incidence, Mortality, and Patient Survival in the United States and 1999–2000 (Ries et al., 2003). Although it is a rare malignancy, it has become the focus of intense interest because of its strong association with asbestos exposure. Rates during 1973–1984 were even higher in Seattle, Hawaii, and San Francisco-Oakland (2.1, 1.9, and 1.7, respectively), where asbestos exposures in the shipbuilding industry were prevalent during World War II (Connelly et al., 1987).

per 100,000 for white non-Hispanics. The current 5-year relative survival rate for all stages combined is 70%. The survival rate of 32% for distant disease is high compared to those for other sites. Approximately three-fourths of the cases are diagnosed while still confined to the bone or with regional spread.

Soft Tissues BONES AND JOINTS, SOFT TISSUES, SKIN Bones and Joints In 2003, approximately 2400 persons were diagnosed with cancer of the bones and joints, and 1300 died from the disease in the United States. Cancers of the bones and joints are very rare; the incidence rate is less than 1 per 100,000, with slightly higher rates among whites than blacks. The incidence ranges from 0.5 per 100,000 for AI/ANs to 1.0

Approximately 8300 people in the United States were diagnosed with cancers of the soft tissue in 2003. The cancers discussed in this section are those that occurred in the soft tissue (connective tissue), and most are sarcomas. There are, however, other sarcomas, such as bone or uterine sarcomas, that are reported with the specific site where they occurred. The overall incidence rate for soft tissue malignancies is 2.8 per 100,000, and it is lower among females than among males (Table 9–9). The 5-year relative survival rate for all ages combined was 68%.

Table 9–9. Bones, Soft Tissue, and Skin: Cancer Incidence and Death Rates and Counts by Sex and Race/Ethnicity in SEER 12 Areas for Incidence and U.S. for Mortality, 1992–2000 All Races Cancer Site/Type

Rate

Count

White Rate

Black

AI/AN

API

Hispanic

White NonHispanic

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

2,347 1,324 1,023

0.7 0.9 0.6

251 133 118

0.5 Ÿ Ÿ

21 Ÿ Ÿ

0.6 0.7 0.5

216 121 95

0.8 0.8 0.7

413 232 181

1.0 1.1 0.8

1,673 937 736

6,955 3,788 3,167

3.0 3.4 2.7

897 454 443

1.5 1.5 1.4

54 28 26

2.1 2.6 1.8

641 349 292

2.5 2.9 2.2

1,011 537 474

2.8 3.4 2.4

5,161 2,835 2,326

56,074 33,951 22,123

5.1 8.6 2.3

1,707 1,334 373

3.3 4.5 2.3

111 72 39

2.9 3.8 2.2

897 552 345

7.4 9.9 5.1

2,985 2,028 957

25.2 33.0 19.1

46,294 27,884 18,410

45,621 25,609 20,012

1.1 1.4 0.9

267 137 130

2.0 2.2 1.8

58 28 30

1.6 1.8 1.4

459 230 229

3.9 4.0 4.0

1,313 558 755

21.0 25.8 17.5

38,417 21,640 16,777

10,453 8,342 2,111

4.1 7.2 1.4

1,440 1,197 243

1.4 2.3 Ÿ

53 44 Ÿ

1.4 2.1 0.8

438 322 116

3.5 5.9 1.1

1,672 1,470 202

4.2 7.1 1.6

7,877 6,244 1,633

0.5 0.6 0.4

9,376 5,236 4,140

0.5 0.7 0.4

1,253 689 564

0.3 0.2 0.3

46 19 27

0.2 0.3 0.2

187 112 75

0.4 0.5 0.3

759 455 304

0.5 0.6 0.4

8,176 4,538 3,638

1.4 1.6 1.3

29,289 14,439 14,850

1.8 1.6 1.9

4,175 1,709 2,466

0.9 0.9 0.9

122 52 70

1.0 1.2 0.9

667 339 328

1.1 1.2 1.1

1,723 856 867

1.4 1.6 1.3

26,300 12,962 13,338

3.9 5.8 2.5

79,855 50,967 28,888

1.2 1.7 0.8

2,628 1,602 1,026

0.9 1.3 0.7

116 69 47

0.7 0.9 0.5

387 220 167

1.3 1.9 0.9

1,650 1,062 588

4.1 6.1 2.6

74,130 47,366 26,764

3.0 4.4 2.0

62,133 38,850 23,283

0.5 0.5 0.5

1,008 424 584

0.6 0.8 0.5

77 42 35

0.4 0.5 0.3

228 114 114

0.8 1.1 0.6

1,054 618 436

3.2 4.6 2.1

57,960 36,348 21,612

0.9 1.5 0.4

17,722 12,117 5,605

0.7 1.2 0.4

1,620 1,178 442

0.4 0.6 Ÿ

39 27 Ÿ

0.3 0.5 0.2

159 106 53

0.5 0.8 0.2

596 444 152

0.9 1.5 0.4

16,170 11,018 5,152

incidence Bones and joints Total 0.9 2,858 0.9 Male 1.0 1,599 1.1 Female 0.8 1,259 0.8 Soft tissue including heart Total 2.8 8,654 2.8 Male 3.3 4,676 3.4 Female 2.4 3,978 2.4 Skin excluding basal and squamous Total 19.6 61,336 22.5 Male 26.4 37,328 29.9 Female 14.3 24,008 16.7 Melanoma of the skin Total 15.7 48,552 18.4 Male 19.8 27,134 23.1 Female 12.8 21,418 15.2 Other nonepithelial skin Total 3.9 12,784 4.1 Male 6.6 10,194 6.8 Female 1.5 2,590 1.5

mortality Bones and Joints Total 0.5 10,862 Male 0.6 6,056 Female 0.4 4,806 Soft tissue including heart Total 1.5 34,253 Male 1.6 16,539 Female 1.4 17,714 Skin excluding basal and squamous Total 3.5 82,986 Male 5.3 52,858 Female 2.3 30,128 Melanoma of the skin Total 2.7 63,446 Male 3.9 39,430 Female 1.8 24,016 Other nonepithelial skin Total 0.8 19,540 Male 1.4 13,428 Female 0.4 6,112

Rates are per 100,000 and age-adjusted to the 2000 U.S. standard million by 5-year age group. Ÿ, Statistic not displayed due to less than 16 cases. Incidence data are from SEER 12 areas except for Hispanic and white not-Hispanic which exclude Alaska, Hawaii, and Detroit. Underlying mortality data provided by NCHS (www.cdc.gov/nchs). Mortality data based on total United States except Hispanic and white non-Hispanic exclude CT, NH, LA, OK. AI/AN, American Indian/Alaska Native; API, Asian or Pacific Islander.

154

PART II: THE MAGNITUDE OF CANCER

Melanoma of the Skin Melanoma of the skin accounts for slightly more than 4% of all malignancies, with more than 50,000 people diagnosed with melanoma each year in the United States. It is especially common among whites (18.4 per 100,000) and white non-Hispanics (21.0) but occurs rarely among blacks (1.1). The rates are also low for AI/ANs and APIs. Melanoma has been one of the most rapidly increasing cancers in the United States. It is also one of the cancers where reporting delay influences the trends. When reporting delay is taken into account, the incidence of melanoma among white males increased 7.5% per year between 1975 and 1980 and then increased 4.3% between 1980 and 2000. Among white females, the incidence increased 5.9% per year during 1975–1981, increased less rapidly during 1981–1993 (2.4%), and then increased 4.6% per year thereafter. If reporting delay is not considered, the increases are not as steep. The recent melanoma trends may reflect increased sun exposure more than increasing diagnosis (Jemal et al., 2001). Incidence probably will continue to increase, at least until most of the current middle-aged population ages during the next several decades. Incidence has generally been higher among males (19.8) than among females (12.8), except for Hispanics, where the rates were the same (4.0 per 100,000). The melanoma death rate is twice as high among males as females for all races, whites, and white non-Hispanics but is less than 1 per 100,000 and is similar for males and females among blacks, AI/ANs, and APIs. The overall 5year relative survival rate is 90%, ranging from a high of 97% for patients with localized disease to a low of 14% for those with distant disease. Melanomas on the skin of male and female genitalia, including the vagina, vulva, scrotum, and penis, are reported as cancers of the specific organ and are not included here.

et al., 2003). For females, the lifetime risk of being diagnosed with breast cancer is 14% (one in seven women) (Ries et al., 2003). The incidence rates (per 100,000 women) are more than double among white females (138.0) compared to AI/ANs (60.8) and are intermediate for blacks (121.1), APIs (92.6), and Hispanics (87.8). Female breast cancer incidence has varied somewhat since 1975 (Fig. 9–4). In 1974 the incidence spiked when women reacted to the highly publicized breast cancer diagnoses in two nationally prominent women and sought out diagnostic testing in larger than usual numbers. The rate subsequently fell. However, since 1980, the rate (per 100,000 women) increased rather dramatically: from 102.1 per 100,000 to a peak of 134.4 in 1987, but then it subsequently leveled off. The incidence rate has remained lower among blacks: 121.1 for black females versus 138.0 among white females during 1992–2000. These increases have been observed in both white and black females both <50 years of age and ≥50 years. The increases in incidence have been confined to women with early-stage disease and recently to women with regional lymph node involvement. Furthermore, the increase is confined to

400

100

The SEER program does not collect data on basal and squamous cell carcinomas of the skin. Although they are the most frequently diagnosed of all malignancies, they are about 99% curable and are usually diagnosed and treated in an ambulatory setting. A separate survey of nonmelanoma skin cancer conducted during the 1970s estimated an incidence rate among whites of 231.3 per 100,000 (Scotto et al., 1983).

Kaposi Sarcoma Prior to the 1980s, Kaposi sarcoma rarely occurred in the United States. The infrequent U.S. occurrences were seen among older men of Jewish or Mediterranean extraction (DeVita et al., 1997). With the emergence of acquired immunodeficiency syndrome (AIDS), Kaposi sarcoma became a frequent response to the compromised immune system; the incidence among high-risk populations in the United States increased accordingly. Among males aged 20–54 years from 1975 to 1980, the SEER Program counted a total of 19 cases (0.1 per 100,000 men). The rate in all SEER geographic areas combined increased from 0.1 per 100,000 during the latter 1970s to 17.5 during the late 1980s and then decreased to 2.2 during 1999–2000 for men aged 20–54 years (Ries et al., 2003). The rate appears to have decreased dramatically after 1989–1990. Men in this 20–54-year-old group in San Francisco County experienced the most dramatic increase of all SEER areas when, by 1989–1990, the incidence of Kaposi sarcoma reached 238.0 per 100,000 compared to a rate of zero during 1973–1979. In areas of the SEER Program other than San Francisco-Oakland, rates among males aged 20–54 years increased to 7.9 during 1989–1990 and then decreased dramatically to 1.6 (Ries et al., 2003). Kaposi sarcoma rates among AIDS patients tend to be underreported (Cote et al., 1995).

FEMALE BREAST There were approximately 212,600 diagnoses of invasive breast cancer in 2003, only 1300 of which were in males. Approximately 2.2 million women are alive who have had a history of breast cancer (Ries

Rate per 100,000 person-years

Skin (Basal and Squamous Cell) White incidence Black incidence White mortality Black mortality

10

1 1975 1980 1985 1990 1995 2000

Year Symbols present observed rates Incidence lines estimated by delay-adjusted joinpoint regression Mortality lines estimated by joinpoint regression

Figure 9–4. Female breast cancer incidence (SEER9 areas) and mortality (United States) rates, age-adjusted using the 2000 U.S. standard, by race, 1975–2000.

155

Cancer Incidence, Mortality, and Patient Survival in the United States

FEMALE GENITAL SYSTEM

400

100

Rate per 100,000 person-years

women with small tumors, implying that the increased utilization of mammography may be partially responsible (Miller et al., 1993). Survival for breast cancer patients increased gradually from 75% to 87% from the mid-1970s to the 1990s. The increased survival rate parallels the increased incidence of small tumors, which are associated with better survival. Therefore, part of the increase in survival may be due in part to increased screening and earlier diagnosis. For females diagnosed during 1992–1999, survival was poorer for blacks than for whites, with the 5-year relative rates being 73% and 88%, respectively. Blacks also had a prognostically less favorable stage distribution, with only 53% diagnosed while the lesion was still localized compared to 64% for whites and lower survival rates for each stage. Nearly 40,000 women died from breast cancer in 2003. The U.S. death rate for breast cancer increased 0.4% per year during 1975–1990 and then decreased (-2.3%) during 1990–2000. For whites, breast cancer mortality increased only 0.3% per year during 1975–1990 and then decreased (-2.3% per year) during 1990–2000. Death rates among black females increased 1.6% per year during 1975–1991 and then decreased (-1.0% per year) during 1991–2000. Death rates for both white and black women decreased more among younger women than older women. Even though the rates are decreasing for black females, they are not decreasing as rapidly as those for whites, and the racial disparity continues to increase between white and black women. Among white females under 50 years of age, breast cancer death rates decreased (-4.2% per year) compared to a decrease of only -2.3% per year among blacks for 1992–2000. For women aged 50 and older, death rates decreased -2.3% and -0.8% per year among whites and blacks, respectively, resulting in an excess risk among blacks during the 1990s. The risk of breast cancer increases rapidly with age during child-bearing years (Brinton et al., 2002; Lacey et al., 2002). After menopause, rates continue to increase but at a less rapid pace. The incidence rates are higher among blacks than whites during childbearing years, but rates are equal at age 40–44 years, with substantial excesses among whites of up to 29% apparent thereafter. Death rates show an excess among blacks compared to whites, which is apparent at all ages. Reasons for the higher death rates among blacks are not well understood.

White incidence Black incidence White mortality Black mortality

10

1 1975 1980 1985 1990 1995 2000

Year Approximately 83,700 females were diagnosed with some cancer of the female genital system in 2003. Cancers of the corpus uteri and uterus, not otherwise specified (NOS), account for nearly half of all malignancies of the female genital system. Ovarian cancer accounts for approximately 30% of such malignancies and cervical cancer 15%. Malignancies of the vagina, vulva, and other female genital organs are rare. In situ lesions are not included in the tabulations.

Symbols present observed rates Incidence lines estimated by delay-adjusted joinpoint regression Mortality lines estimated by joinpoint regression

Figure 9–5. Cervix uteri cancer incidence (SEER9 areas) and mortality (United States) rates, age-adjusted using the 2000 U.S. standard, by race, 1975–2000.

Cervix Uteri Both incidence and mortality for cervical cancer have decreased substantially over the past four decades due at least in part to widespread use of the Papanicolaou smear and advances in diagnostic techniques. From 1950 to 2000 the incidence for white females decreased -78% (-2.4% per year), as did mortality (-79%, or approximately -3.5% per year). The more recent trends for whites show that the decline is not as rapid (Fig. 9–5). The decreases have been largely due to earlier diagnosis and treatment of squamous cell carcinomas, the predominant form of cervical cancer (Wang et al., 2004). Adenocarcinomas are less frequent, more difficult to detect, and likely due to different causes. The incidence during 1992–2000 was highest among Hispanic females (17.9 per 100,000) followed by black females (13.0), APIs (11.1), and whites (9.5); it was lowest among AI/ANs (7.2) (Table 9–10). Death rates were more than twice as high among blacks as white non-Hispanics and were intermediary for the other racial/ethnic groups. Slightly more than one-half of all cases are diagnosed while the cancer is localized (i.e., still confined to the cervix uteri). Survival

rates range from 92% for women diagnosed with localized disease to 17% for those with distant disease, suggesting that mortality could be further reduced by improved early diagnosis. Five-year relative survival rates are high among young women (82.4%) and decrease as age increases, falling to 42.3% among women 75 years and older.

Corpus Uteri Shortly after the use of postmenopausal estrogens gained wide acceptance, the corpus uterine cancer incidence increased to 35.5 per 100,000 in 1975. When this association was identified, the rates fell sharply (-6.0% per year during 1975–1979) as the number of prescriptions decreased and dosages were modified. The rates declined -1.7% per year during 1979–1988. The rates hit a low of 23.6 in 1988, after which they increased 0.7% per year until 1998 and then stabilized. Cancer of the uterus, NOS is combined with that for cancer of the corpus uteri because most of these cancers probably originate in the corpus (Percy et al., 1990a). Accurate estimation of corpus uterine

156

PART II: THE MAGNITUDE OF CANCER

Table 9–10. Breast and Female Genital: Cancer Incidence and Death Rates and Counts by Primary Site, Sex, and Race/Ethnicity in SEER 12 Areas for Incidence and in U.S. for Mortality, 1992–2000 All Races Cancer Site

White

Black

AI/AN

API

Hispanic

White NonHispanic

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

72.7 1.2 132.5 52.4 10.0 24.5 24.0 0.4 14.3 0.7 2.2 0.7

217,520 1,456 216,064 86,195 16,728 39,912 39,189 723 23,437 1,217 3,680 1,221

75.3 1.1 138.0 54.5 9.5 26.0 25.7 0.4 15.1 0.7 2.3 0.8

181,969 1,189 180,780 71,867 12,431 34,266 33,719 547 20,012 938 3,222 998

69.9 1.8 121.1 44.9 13.0 17.7 16.9 0.7 10.5 1.1 1.8 0.7

18,335 171 18,164 6,745 2,125 2,530 2,427 103 1,539 162 275 114

33.0 Ÿ 60.8 28.7 7.2 10.1 9.7 Ÿ 9.0 Ÿ Ÿ Ÿ

962 Ÿ 960 458 128 159 154 Ÿ 138 Ÿ Ÿ Ÿ

50.5 0.6 92.6 40.2 11.1 16.7 16.3 0.4 10.4 0.6 0.9 0.5

14,981 75 14,906 6,504 1,831 2,687 2,627 60 1,678 87 133 88

48.5 0.7 87.8 49.0 17.9 16.3 15.9 0.3 11.5 0.9 1.8 0.5

14,676 69 14,607 8,705 3,642 2,642 2,590 52 1,914 140 248 119

78.6 1.2 144.7 54.2 7.9 27.0 26.6 0.4 15.5 0.6 2.4 0.8

144,607 962 143,645 54,069 7,436 27,161 26,732 429 15,552 663 2,494 763

16.6 0.3 29.2 17.2 3.2 4.1 2.1 2.0 9.0 0.3 0.5 0.2

387,142 3,253 383,889 229,990 40,111 56,533 28,485 28,048 120,637 3,600 6,381 2,728

16.3 0.3 28.7 16.9 2.8 3.9 2.0 1.9 9.3 0.2 0.5 0.2

334,172 2,722 331,450 197,966 30,241 47,132 24,115 23,017 109,292 3,044 5,930 2,327

21.8 0.6 36.7 22.0 6.5 7.0 3.2 3.7 7.6 0.4 0.3 0.3

47,304 504 46,800 27,736 8,517 8,551 3,965 4,586 9,416 501 405 346

8.4 Ÿ 14.9 11.1 3.3 2.4 1.2 1.2 4.8 Ÿ Ÿ Ÿ

1,045 Ÿ 1,041 755 250 155 82 73 319 Ÿ Ÿ Ÿ

7.1 0.1 12.9 10.4 3.1 2.2 1.0 1.2 4.7 0.1 0.1 0.1

4,621 23 4,598 3,533 1,103 695 323 372 1,610 45 32 48

10.2 0.2 18.1 14.0 4.0 3.2 1.4 1.7 6.1 0.2 0.4 0.1

13,322 103 13,219 9,983 3,275 2,113 962 1,151 4,133 140 221 101

16.6 0.3 29.2 17.0 2.7 3.9 2.0 1.9 9.5 0.2 0.5 0.2

304,439 2,466 301,973 178,926 25,561 42,922 22,063 20,859 100,145 2,751 5,446 2,101

incidence Breast Total Male Female Female genital system Cervix uteri Corpus and uterus, NOS Corpus uteri Uterus, NOS Ovary Vagina Vulva Other female genital organs

mortality Breast Total Male Female Female genital system Cervix uteri Corpus and uterus, NOS Corpus uteri Uterus, NOS Ovary Vagina Vulva Other female genital organs

Rates are per 100,000 and age-adjusted to the 2000 U.S. standard million by 5-year age group. Ÿ, Statistic not displayed due to less than 16 cases. For incidence, and ovary excludes borderline tumors of the ovary. Incidence data are from SEER 12 areas except for Hispanic and white not-Hispanic which exclude Alaska, Hawaii, and Detroit. Underlying mortality data provided by NCHS (www.cdc.gov/nchs). Mortality data based on total United States except Hispanic and white non-Hispanic exclude CT, NH, LA, OK. AI/AN, American Indian/Alaska Native; API, Asian or Pacific Islander.

cancer rates is complicated by including women who have had a hysterectomy in the population estimates. The prevalence of women with an intact uterus varies by geography, race, and time period, resulting in varying underestimations of the true risk of this cancer (Pokras and Hufnagel, 1987; CDC, 2002). This would not account, however, for the rapid changes in the incidence rates. Cancer of the corpus uteri has the most favorable prognosis of any gynecologic malignancy, with an 84% five-year relative survival rate for all stages combined. Nearly three-fourths are still localized at diagnosis. Trends in survival for this cancer seem to correspond to the increased incidence of early lesions diagnosed during the mid-1970s. The 5-year relative survival rate was 88% in 1975 and 83% in 1986. The survival rate for whites is more than 25 percentage points higher than that for blacks. The black–white survival differential is more than 10 percentage points for the <50 years age group and more than 30 percentage points for the ≥50 years age group (Ries et al., 2003). Much of the lower survival rates among blacks can be explained by stage at diagnosis (Hill et al., 1996). The incidence is nearly 50% higher among whites (26.0 per 100,000 women) than among blacks (17.7) (Table 9–10). Incidence rates are similar for Hispanics, APIs, and black females (16–18 per 100,000), and they are highest among white non-Hispanic females (27.0 per 100,000). Mortality, however, is nearly 80% higher among black females due to their much poorer survival rate than white females (Table 9–10; Appendix 9–C). The incidences for the various histologic types of corpus cancer vary by race; and less favorable outcomes for usual types of uterine adenocarcinomas and for rare aggressive tumors contribute equally to the relatively high mortality associated with corpus cancer among black females (Sherman and Devesa, 2003).

Mortality from cancer of the corpus uteri and uterus NOS decreased between 1975 and 1991 and then leveled off during 1991–2000; the pattern was similar for whites and showed a smaller but more consistent decrease for black females (-0.7% per year) between 1975 and 2000.

Ovary Borderline lesions of the ovary have been recorded only in recent years, so they must be excluded to evaluate the real trend (Fig. 9–6). When borderline lesions are excluded, the incidence rate decreased about -1% per year during 1990–2000. However, the incidence decreased among women <65 years of age, whereas they increased slightly for women ≥65 years (Ries et al., 2003). Ovarian cancer risk decreases with increasing parity and lengthening duration of oral contraceptive (OC) use, and OC use increased during the last quarter of the twentieth century. Comparison of observed rates with those predicted by the changes in risk factor prevalence revealed agreement in young women ages 30–49 years but less agreement among older women ages 50–64 years, suggesting that the protective effect of OC use declines with age (Gnagy et al., 2000). The incidence is more than 40% higher and death rates (per 100,000) are more than 20% higher among whites (15.1) than blacks (10.5) (Table 9–10). The reported rates for some specific histopathologic tumor types have changed over time, in part reflecting more specific pathologic classification (Mink et al., 2002). The possible effect of the shifting exposure prevalence on incidence patterns and the variations in racial relative risks by tumor type warrant further study. Rates per 100,000 are similar for blacks and APIs, slightly lower for AI/ANs

157

Cancer Incidence, Mortality, and Patient Survival in the United States

MALE GENITAL SYSTEM Genital cancers account for about 34% of all cancers diagnosed in males. Of all male genital cancers, cancer of the prostate accounts for more than 96%.

Prostate The incidence of prostate cancer is 63% higher among black men than among white men. Some of the largest differences among racial/ethnic groups were seen for prostate cancer, ranging from 62.9 per 100,000 among AI/AN men to 286.5 for black men. American blacks have the highest incidence of prostate cancer in the world (Hsing and Devesa, 2001; Parkin et al., 2003). The trend for prostate cancer has been one of the most volatile. Prostate cancer is one of the malignancies for which reporting delay is important (Fig. 9–7). When this delay was taken into account, it was estimated that the increase between 1995 and 2000 would be closer to

400

Rate per 100,000 person-years

100

White incidence Black incidence White mortality Black mortality

10

Figure 9–6. Ovary cancer incidence (SEER9 areas), excluding borderline tumors, and mortality (United States) rates, age-adjusted using the 2000 U.S. standard, by race, 1975–2000.

(9.0), and slightly higher for Hispanics (11.5) and white non-Hispanics (15.5). For APIs, ovarian cancer is diagnosed at younger ages and at an earlier stage than in women of other racial/ethnic groups (Goodman et al., 2003). Death rates for AI/ANs and APIs are lower than those for whites or blacks. Whereas ovarian cancer has the third highest incidence among all female genital sites, ovarian cancer has the highest mortality. Unlike cancers of the cervix and corpus uteri, death rates for ovarian cancer are higher among white females. Over a 9-year period (1992–2000), death rates decreased among white, black, AI/AN, and API females. The silence of this disease in its early stages has impeded early detection efforts, and survival rates have remained poor. Of all ovarian cancers diagnosed from 1992 to 1999 except borderline tumors, 67% were detected at an advanced stage, with a 5-year relative survival rate of only 28%. In contrast, the rate was 92% among those diagnosed with localized disease (20% of cases). The rate for all stages combined was 44%.

1 1975 1980 1985 1990 1995 2000

Year Symbols present observed rates Incidence lines estimated by delay-adjusted joinpoint regression Mortality lines estimated by joinpoint regression

Figure 9–7. Prostate cancer incidence (SEER9 areas) and mortality (United States) rates, age-adjusted using the 2000 U.S. standard, by race, 1975–2000.

158

PART II: THE MAGNITUDE OF CANCER increases were seen for both whites and blacks, but the 5-year survival rate is about 10 percentage points lower for black men. The mortality rate for all races combined decreased -10.4% per year during 1975–1981; the decrease slowed to -3.2% per year during 1981–1995 and then stabilized. During 1992–1999, about 69% of all cases were diagnosed while still in the localized stage, with a 99% five-year relative survival rate. Even with distant metastases, the survival rate was 73%. Because the survival rate is so high, the mortality rate is low (0.1–0.3 per 100,000) among all racial/ethnic groups.

a 2.3% annual increase. Rates for black men increased more than those for whites during the late 1980s to early 1990s; prostate cancer in whites decreased more rapidly during the mid-1990s, and then both increased between 1995 and 2000. There is indirect evidence that the increasing incidence over the period 1973–1986 may be due to increased detection of clinically asymptomatic cases associated with increasing rates of transurethral resection of the prostate (TURP) for benign prostatic hyperplasia (Potosky et al., 1990). The increases after 1987 do not seem to be associated with TURPs. During the late 1980s, increases in incidence were seen for all stages except distant stage. At the same time, there were increases in the use of tests to detect prostate cancer, such as transrectal ultrasound-guided needle biopsy and, in particular, blood testing for prostate-specific antigen (PSA), which suggests that the observed increases in incidence are surveillancerelated (Potosky et al., 1995). The lifetime risk of being diagnosed with prostate cancer is 17% (Ries et al., 2003). Death rates for prostate cancer increased 0.9% per year from 1975 through 1987, increased more rapidly (3.1% per year) during 1987–1991, stabilized, and then decreased -4.0% per year after 1994. Rates began to decrease earlier for younger men: -3.2% per year during 1990–1996 and -5.6% per year during 1996–2000. These figures can be compared to stable rates during 1991–1994 and -4.0% per year during 1994–2000 for men 65 years of age and older. Mortality among blacks was more than twice that among whites (Table 9–11). Five-year survival for patients with prostate cancer increased from 67% during 1974–1976 to 97% during 1992–1999. The 5-year relative survival rates for 1992–1999 were 93% among blacks and 98% among whites, reflecting a slightly more favorable stage distribution at diagnosis among whites. Among cases diagnosed during 1992–1999, survival was 90% for those with regional disease but fell to 34% for distant disease. The overall survival rate is very high, and few patients are diagnosed with distant disease.

Penis Cancer of the penis is rare among both blacks and whites in the United States, accounting for less than 0.5% of all male genital cancers, with an incidence rate of 0.8 per 100,000 men. Fifty-seven percent of all cases are diagnosed while still localized, and the overall 5-year relative survival rate is 75%.

URINARY SYSTEM Cancers of the urinary tract, of which bladder cancer is the most common, are more than three times as frequent in males as in females (52.6 and 17.2 per 100,000, respectively). They are about 32% more frequent among whites than among blacks (34.3 and 26.0 per 100,000, respectively).

Urinary Bladder Cancer of the bladder is seen predominantly in white males, with an incidence rate of 39.6 per 100,000, about twice that among black males (20.4), four times that among white females (9.9), and more than five times that among black females (7.6) (Table 9–12). In contrast to other cancers that include only invasive cancer, in situ cancer is combined with invasive cancer, as the distinction is particularly difficult. For stage comparisons, in situ bladder cancer is combined with localized cancer. Since 1988, the SEER Program has coded the tumor characteristics according to the American Joint Committee on Cancer (1988). Since that time, more than 70% of U.S. bladder cancer incidence was described as superficial (SEER*Stat, 2003). The incidence rates of bladder cancer have not changed greatly since the 1990s. The incidence has uniformly been highest in whites. The incidence for blacks is about 60%, Hispanics 50%, APIs 40%, and AI/ANs 20% that of the incidence for whites. The mortality patterns

Testis Cancer of the testis is a rare malignancy that mainly affects young adult men (median age 34 years). The incidence rose 1.6% per year during 1975–2000. This cancer was more than four times as common among whites (5.9 per 100,000 men) as among blacks (1.3) during 1992–2000 (Table 9–11). The increases have been due to rising rates for seminomas (McGlynn et al., 2003). Survival for patients with this malignancy is favorable, and improvement in survival has been dramatic, increasing from 79% for the years 1974–1976 to 95% for the years 1992–1999. Survival

Table 9–11. Male Genital: Cancer Incidence and Death Rates and Counts by Primary Site, Sex, and Race/Ethnicity in SEER 12 Areas for Incidence and in U.S. for Mortality, 1992–2000 All Races

White

Black

AI/AN

API

Hispanic

White NonHispanic

Cancer Site

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

incidence

186.8 180.6 5.0 0.8 0.4

237,746 227,232 8,996 1,055 463

182.5 175.4 5.9 0.9 0.3

191,489 182,018 8,235 887 349

289.1 286.5 1.3 1.0 0.3

27,712 27,358 229 93 32

66.3 62.9 2.5 Ÿ Ÿ

648 574 64 Ÿ Ÿ

109.2 106.1 2.1 0.4 0.6

12,499 11,993 386 49 71

146.7 141.8 3.4 1.2 0.3

15,255 13,862 1,200 157 36

182.0 174.3 6.6 0.8 0.3

148,688 141,750 6,063 614 261

35.8 35.3 0.3 0.2 0.0

305,772 300,499 3,139 1,814 320

33.0 32.5 0.3 0.2 0

252,247 247,514 2,897 1,560 276

76.4 75.9 0.1 0.3 0

50,152 49,702 187 224 39

23.3 22.9 0.2 Ÿ Ÿ

792 762 20 Ÿ Ÿ

15.4 15.2 0.1 0.1 Ÿ

2,581 2,521 35 21 Ÿ

26.0 25.5 0.2 0.3 Ÿ

9,489 9,016 316 144 Ÿ

33.2 32.7 0.3 0.2 0

231,074 227,041 2,459 1,331 243

Prostate Testis Penis Other male genital organs

mortality Prostate Testis Penis Other male genital organs

Rates are per 100,000 and age-adjusted to the 2000 U.S. standard million by 5-year age group. Ÿ, Statistic not displayed due to less than 16 cases. Incidence data are from SEER 12 areas except for Hispanic and white not-Hispanic which exclude Alaska, Hawaii, and Detroit. Underlying mortality data provided by NCHS (www.cdc.gov/nchs). Mortality data based on total United States except Hispanic and white non-Hispanic exclude CT, NH, LA, OK. AI/AN, American Indian/Alaska Native; API, Asian or Pacific Islander.

159

Cancer Incidence, Mortality, and Patient Survival in the United States

Table 9–12. Urinary System: Cancer Incidence and Death Rates and Counts by Primary Site, Sex, and Race/Ethnicity in SEER 12 Areas for Incidence and in U.S. for Mortality, 1992–2000 All Races Cancer Site

Rate

White

Black

AI/AN

API

White NonHispanic

Hispanic

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

94,461 65,588 28,873

34.3 56.5 18.1

82,838 57,979 24,859

26.0 39.3 17.1

6,271 3,859 2,412

16.3 24.2 10.4

401 256 145

16.6 26.2 9.0

4,377 3,062 1,315

21.5 33.5 13.0

5,511 3,596 1,915

34.7 57.3 18.1

65,005 45,798 19,207

59,663 44,113 15,550

22.3 39.6 9.9

53,776 39,987 13,789

12.6 20.4 7.6

2,846 1,832 1,014

4.9 8.4 2.1

108 83 25

9.8 16.5 4.4

2,499 1,865 634

10.6 18.7 5.1

2,495 1,814 681

23.0 40.8 10.2

43,333 32,289 11,044

32,279 19,852 12,427

11.1 15.5 7.7

26,951 16,593 10,358

12.6 17.7 8.9

3,216 1,912 1,304

11.2 15.4 8.0

287 170 117

6.2 8.9 4.0

1,703 1,104 599

10.3 14.0 7.6

2,887 1,699 1,188

10.9 15.2 7.4

20,073 12,461 7,612

1,544 959 585

0.6 0.9 0.4

1,357 857 500

0.3 0.4 0.2

56 33 23

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.5 0.6 0.4

126 66 60

0.4 0.5 0.2

80 50 30

0.6 0.9 0.4

1,080 674 406

975 664 311

0.3 0.5 0.2

754 542 212

0.6 0.9 0.5

153 82 71

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.2 0.2 0.1

49 27 22

0.2 0.3 0.1

49 33 16

0.3 0.5 0.1

519 374 145

9.0 14.6 5.3

187,029 121,047 65,982

8.4 12.4 6.0

17,017 9,640 7,377

6.5 9.3 4.5

709 436 273

3.9 5.7 2.4

1,898 1,260 638

6.1 9.5 3.7

6,650 4,333 2,317

9.1 14.7 5.4

171,006 110,793 60,213

4.5 8.1 2.3

93,866 64,298 29,568

4.1 6.0 3.0

7,799 4,189 3,610

1.5 2.4 1.0

151 95 56

1.8 2.8 1.1

806 542 264

2.4 4.2 1.2

2,324 1,613 711

4.6 8.2 2.3

86,958 59,597 27,361

4.3 6.2 2.9

88,941 54,472 34,469

4.2 6.2 2.8

8,788 5,298 3,490

4.8 6.7 3.4

546 334 212

1.9 2.8 1.2

1,031 686 345

3.6 5.2 2.4

4,216 2,663 1,553

4.3 6.2 2.9

80,140 49,078 31,062

0.1 0.2 0.1

2,725 1,527 1,198

0.1 0.1 0.1

115 44 71

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.1 0.1 0.1

47 27 20

0.1 0.1 0.1

63 34 29

0.1 0.2 0.1

2,529 1,419 1,110

0.1 0.1 0.1

1,497 750 747

0.2 0.1 0.2

315 109 206

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.0 0.0 0.0

47 23 24

0.1 0.1 0.1

1,379 699 680

incidence Urinary system Total 32.1 Male 52.6 Female 17.2 Urinary bladder Total 20.4 Male 36.1 Female 9.2 Kidney and renal pelvis Total 10.8 Male 15.1 Female 7.5 Ureter Total 0.5 Male 0.8 Female 0.3 Other urinary organs Total 0.3 Male 0.5 Female 0.2

mortality Urinary system Total 8.8 206,653 Male 14.2 132,383 Female 5.3 74,270 Urinary bladder Total 4.4 102,622 Male 7.8 69,124 Female 2.3 33,498 Kidney and renal pelvis Total 4.2 99,306 Male 6.2 60,790 Female 2.8 38,516 Ureter Total 0.1 2,895 Male 0.2 1,604 Female 0.1 1,291 Other urinary organs Total 0.1 1,830 Male 0.1 865 Female 0.1 965

Ÿ Ÿ Ÿ

Rates are per 100,000 and age-adjusted to the 2000 U.S. standard million by 5-year age group. Ÿ, Statistic not displayed due to less than 16 cases. Incidence data are from SEER 12 areas except for Hispanic and white not-Hispanic which exclude Alaska, Hawaii, and Detroit. Underlying mortality data provided by NCHS (www.cdc.gov/nchs). Mortality data based on total United States except Hispanic and white non-Hispanic exclude CT, NH, LA, OK. AI/AN, American Indian/Alaska Native; API, Asian or Pacific Islander.

generally are similar, except for higher rates (30%) among black compared to white females. During the 1990s, bladder cancer death rates declined in every race/sex group, although the decreases were significant only among white and black males. The male/female mortality ratio ranges from 3.5 among whites to 2.0 among blacks. The current 5-year survival rate for all stages combined is 82%; for cancer diagnosed while still localized (which accounts for 74% of the cases), the rate is 94%. The survival rate for whites (83%) is considerably higher than that for blacks (64%), partly reflecting the large difference in the proportion of cancers diagnosed while still localized: 75% for whites and only 57% for blacks (Ries et al., 2003). Factors associated with a poor prognosis and greater extent of disease were worse among black patients than whites (Prout, 2000). Even within each stage of the disease, survival differences by race persisted. Again, in situ cancers were grouped with localized disease for bladder cancer.

Kidney and Renal Pelvis Kidney cancer incidence and death rates are about twice as high for males as for females in every racial/ethnic group. Rates are lower for Hispanics and Asian/Pacific Islanders than for whites, and the rates are similar between American Indian/Alaska Natives and whites. Death rates are similar among whites and blacks of the same gender, but incidence rates have been increasing (more rapidly among blacks than whites such that recent rates are about 14% higher among blacks than whites). Increasing detection of presymptomatic tumors by imaging procedures such as ultrasonography, computed tomography, and magnetic resonance imaging does not fully explain the upward incidence trends of renal cell carcinoma, the dominant form of kidney cancer (Chow et al., 1999). Other factors may be contributing to the rapidly increasing incidence of renal cell cancer in the United States,

160

PART II: THE MAGNITUDE OF CANCER

particularly among blacks. The incidence of kidney cancer has increased substantially in many areas around the world; the trends may result partly from the increasing prevalence of risk factors, notably obesity, although variations in hypertension, cigarette smoking, and dietary habits may have played a role in some areas (Mathew et al., 2002). The 5-year relative survival rate for all stages is 63%; however, for localized cancer, which accounts for 50% of the cases, the 5-year relative survival rate is 90%.

LYMPHOMAS All lymphomas arising in both lymphatic and extralymphatic sites are included here. For example, a lymphoma arising in the stomach is tabulated with lymphomas, not with stomach cancers. These tumors are further subdivided into Hodgkin lymphoma and non-Hodgkin lymphoma.

Hodgkin Lymphoma EYE, NERVOUS SYSTEM, ENDOCRINE SYSTEM The combined incidence for eye, central nervous system (CNS), and endocrine system cancer in the general population was about 14 per 100,000. Ocular and orbital cancer occurred at a rate of 0.8; brain and CNS cancer, 6.4; and endocrine cancer, 7.1 (Table 9–13).

Eye More than 65% of all malignancies of the eye are melanomas; and another 14% are retinoblastomas, most of which occur during childhood. The skin of the eyelid is not included in this section. Incidence rates are higher among whites (0.9 per 100,000) than among blacks (0.2); death rates are low, less than 0.2 per 100,000 in each group. The 5-year relative survival rate is 82%, and more than three-fourths are diagnosed while the tumor is still confined to the eye.

Brain and Central Nervous System More than 90% of the malignancies in the brain and CNS occur in the brain. The incidence for cancer of the brain and CNS increased 1.4% per year during 1975–1988 and then decreased -0.5% per year during 1988–2000. The rates appeared to increase around the time of advances in the technology that allowed better detection of brain tumors (Legler et al., 1999). The 1992–2000 incidence rate was higher for whites (7.1 per 100,000) than for blacks (4.0) and higher among males than among females of each race. The 5-year relative survival rate for cancer of the brain is 33%. For cancers arising in the brain, stage at diagnosis has relatively little impact on survival; therefore the data are not shown by stage (Appendix 9–D). Although brain cancers rarely metastasize, as space-occupying lesions they compress and damage adjacent tissues.

Endocrine System Cancer of the thyroid accounts for about 90% of cancers of the endocrine system. Thyroid cancer is one of the few non-genderspecific malignancies that occur more often in females than males, with incidence rates of 9.2 and 3.5 per 100,000, respectively. It also occurs more frequently in whites (6.6) than in blacks (3.7). Rates are highest for API females (11.4) and lowest for black males (2.0). Rates are similar for blacks (3.7) and AI/ANs (3.8) and similar for Hispanics (6.0), whites (6.6), and white non-Hispanics (6.7). The median age at diagnosis, 46 years, is relatively young. The 5-year relative survival rate of 96% for all stages is one of the highest for any malignancy. About 56% of thyroid cancers are diagnosed while still localized, and these patients experience a 5-year relative survival rate close to 100%. Few cases (<6%) are metastatic at diagnosis. The overall death rate for thyroid cancer is only 0.5 per 100,000.

MISCELLANEOUS MALIGNANT TUMORS AND UNKNOWN PRIMARY SITE Miscellaneous malignant cancers include ill-defined and unknown primary sites plus reticuloendothelial neoplasms. If a site such as liver is biopsied and determined to be a metastatic site and the site of origin of the tumor cannot be determined, SEER classifies the primary site as “unknown.”

The current incidence rate for Hodgkin lymphoma is 2.7 per 100,000 (Table 9–14), with the rates having decreased somewhat over time. Incidence is higher among whites (3.0 per 100,000) than among blacks (2.4), Hispanics (2.3), or APIs (1.0) and is higher in males than in females (3.0 vs. 2.4, respectively). There is a bimodal age distribution, with peak risks occurring in the age groups 25–29 and 75–79 years (Ries et al., 2003). The 5-year survival rate is 84% (Appendix 9–C). Death rates for Hodgkin lymphoma have been decreasing since the late 1960s, reflecting substantial improvement in treatment and survival. Mortality decreased more than 60% from 1975 to 2000.

Non-Hodgkin Lymphoma The incidence of non-Hodgkin lymphoma (NHL) is higher among whites (20.1 per 100,000) than among Hispanics (16.0), blacks (14.4), or APIs (13.6) (Table 9–14). NHL rates have been rising for many years, with an acceleration during the 1980s due particularly to extranodal and high-grade tumors (Devesa and Fears, 1992; Groves et al., 2000). Changes in diagnosis and classification can account for only a small portion of the long-term increases (Hartge and Devesa, 1992). NHL incidence rose 78% during 1975–1995, one of the largest increases seen for any malignancy. These increases were more rapid among males than among females (Ries et al., 2003) (Fig. 9–8). The incidence has declined since 1995. During the 1980s, rates for NHL increased most rapidly among males under age 65. Part of the more rapid increase among this group was the effect of more diagnoses of NHL secondary to human immunodeficiency (HIV) infection. The incidence rates for NHL in males aged 20–54 were particularly high in the San Francisco-Oakland area. These dramatic increases lagged a couple of years behind the increases seen for Kaposi sarcoma (Eltom et al., 2002). The inclusion of San Francisco-Oakland in the incidence rates for all SEER areas combined exaggerates the overall increase in NHL incidence for men aged 20–54 (Ries et al., 2003). Mortality increased 59%, from 5.6 per 100,000 in 1975 to the peak of 8.9 in 1997. During the late 1990s, death rates declined in all four race/sex groups, most likely related to improved therapy for NHL. Recent 5-year relative survival is about 56% overall, ranging from 68% among patients diagnosed with localized disease to 45% among those with distant disease. Survival rates improved substantially during the 1960s and 1970s, particularly among children, resulting in declining death rates for those under 15 years of age (Ries et al., 2003).

MULTIPLE MYELOMA Multiple myeloma occurs predominantly among black males (13.3 per 100,000), a rate higher than that for black females (10.3), twice that for white males (6.7), and more than three times that for white females (4.2). American blacks have among the highest multiple myeloma rates in the world (Parkin et al., 2003). The incidence and death rates have not changed greatly over the last decade. The 5-year relative survival is poor, at only 32%.

LEUKEMIA The leukemias are a family of malignant diseases of the blood that are subclassified as acute leukemias and chronic leukemias; they are then further subdivided into categories based on cell type. The major cell

Table 9–13. Eye, Brain, Endocrine, and Miscellaneous: Cancer Incidence and Death Rates and Counts by Sex, and Race/Ethnicity in SEER 12 Areas for Incidence and in U.S. for Mortality, 1992–2000 All Races Cancer Site

Rate

Count

White Rate

Black

Count

AI/AN

Rate

Count

2,231 1,239 992

0.2 0.2 0.2

80 36 44

17,685 9,848 7,837

4.0 4.8 3.5

1,236 654 582

2.2 2.8 1.7

16,663 9,309 7,354

3.6 4.4 3.0

1,109 559 510

2.0 2.6 1.5

1,022 539 483

0.4 0.4 0.5

127 55 72

18,346 5,220 13,126

4.4 2.8 5.9

16,813 4,372 12,441

API

White NonHispanic

Hispanic

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.2 0.3 0.2

79 37 42

0.6 0.7 0.4

235 134 101

1.0 1.2 0.8

1,761 976 785

86 49 37

3.5 4.2 3.0

1,109 612 497

5.0 5.6 4.5

2,023 1,077 946

7.4 8.9 6.2

13,421 7,517 5,904

76 43 33

3.2 3.9 2.7

1,016 570 446

4.7 5.3 4.2

1,884 1,010 874

7.0 8.5 5.8

12,667 7,116 5,551

0.3 0.3 0.3

93 42 51

0.3 0.3 0.3

139 67 72

0.4 0.5 0.4

754 401 353

incidence Eye and orbit Total 0.8 2,467 0.9 Male 1.0 1,352 1.1 Female 0.7 1,115 0.7 Brain and other nervous system Total 6.4 20,194 7.1 Male 7.7 11,207 8.5 Female 5.4 8,987 5.9 Brain Total 6.0 18,933 6.7 Male 7.3 10,561 8.1 Female 5.0 8,372 5.5 Cranial nerves, other nervous system Total 0.4 1,261 0.4 Male 0.4 646 0.5 Female 0.4 615 0.4 Endocrine system Total 7.1 22,842 7.2 Male 4.2 6,341 4.3 Female 9.8 16,501 10.0 Thyroid Total 6.4 20,802 6.6 Male 3.5 5,226 3.6 Female 9.2 15,576 9.5 Other endocrine including thymus Total 0.6 2,040 0.6 Male 0.7 1,115 0.7 Female 0.6 925 0.5 Miscellaneous malignant cancer Total 12.1 35,392 12.0 Male 13.9 16,947 13.8 Female 10.8 18,445 10.7

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

1,350 359 991

4.3 2.8 5.8

160 39 121

8.6 4.7 12.1

2,792 678 2,114

6.5 3.2 9.6

2,776 618 2,158

7.3 4.5 10.1

13,273 3,919 9,354

3.7 2.0 5.2

1,122 252 870

3.8 2.1 5.6

144 27 117

7.8 3.8 11.4

2,543 535 2,008

6.0 2.6 9.2

2,537 481 2,056

6.7 3.8 9.6

12,166 3,307 8,859

1,533 848 685

0.7 0.8 0.7

228 107 121

0.4 Ÿ Ÿ

16 Ÿ Ÿ

0.8 0.9 0.6

249 143 106

0.5 0.6 0.4

239 137 102

0.6 0.7 0.5

1,107 612 495

29,038 13,830 15,208

15.6 18.3 13.7

3,663 1,772 1,891

9.6 9.6 9.6

206 90 116

8.8 9.9 7.9

2,236 1,125 1,111

11.8 13.2 10.9

2,826 1,321 1,505

12.1 14.0 10.7

22,807 10,933 11,874

2,228 1,124 1,104

0.0 0.0 0.0

91 39 52

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.0 Ÿ Ÿ

19 Ÿ Ÿ

0.0 0.0 0.0

76 36 40

0.1 0.1 0.1

2,065 1,042 1,023

103,142 56,357 46,785

2.8 3.3 2.3

6,524 3,426 3,098

2.0 2.5 1.6

292 163 129

1.9 2.2 1.7

1,286 694 592

2.9 3.5 2.4

4,411 2,433 1,978

5.2 6.3 4.3

93,816 51,237 42,579

16,012 7,178 8,834

0.8 0.7 0.9

1,854 742 1,112

0.4 0.3 0.5

54 21 33

0.9 0.8 1.0

513 204 309

0.8 0.7 0.9

1,114 461 653

0.8 0.8 0.7

14,152 6,370 7,782

9,298 3,706 5,592

0.4 0.3 0.5

896 273 623

0.3 Ÿ 0.4

33 Ÿ 24

0.7 0.5 0.9

353 107 246

0.6 0.4 0.7

677 214 463

0.4 0.4 0.4

8,184 3,306 4,878

6,714 3,472 3,242

0.4 0.4 0.3

958 469 489

0.1 Ÿ Ÿ

21 Ÿ Ÿ

0.2 0.3 0.2

160 97 63

0.2 0.3 0.2

437 247 190

0.3 0.4 0.3

5,968 3,064 2,904

301,964 154,195 147,769

19.5 25.3 15.5

40,130 20,977 19,153

13.0 14.3 12.1

1,393 661 732

8.6 10.2 7.3

4,506 2,350 2,156

9.8 12.1 8.2

10,987 5,808 5,179

14.7 18.1 12.2

274,865 140,034 134,831

mortality Eye and orbit Total 0.1 2,344 0.1 Male 0.1 1,175 0.1 Female 0.1 1,169 0.1 Brain and other nervous system Total 4.7 111,244 5.0 Male 5.7 60,640 6.1 Female 3.9 50,604 4.1 Endocrine system Total 0.8 18,433 0.8 Male 0.8 8,145 0.8 Female 0.8 10,288 0.8 Thyroid Total 0.5 10,580 0.4 Male 0.4 4,095 0.4 Female 0.5 6,485 0.5 Other endocrine including thymus Total 0.3 7,853 0.3 Male 0.4 4,050 0.4 Female 0.3 3,803 0.3 Miscellaneous malignant cancer Total 14.9 347,993 14.5 Male 18.5 178,183 18.0 Female 12.3 169,810 12.1

Rates are per 100,000 and age-adjusted to the 2000 U.S. standard million by 5-year age group. Ÿ, Statistic not displayed due to less than 16 cases. Incidence data are from SEER 12 areas except for Hispanic and white not-Hispanic which exclude Alaska, Hawaii, and Detroit. Underlying mortality data provided by NCHS (www.cdc.gov/nchs). Mortality data based on total United States except Hispanic and white non-Hispanic exclude CT, NH, LA, OK. AI/AN, American Indian/Alaska Native; API, Asian or Pacific Islander.

161

162

PART II: THE MAGNITUDE OF CANCER

Table 9–14. Lymphoma and Myeloma: Cancer Incidence and Death Rates and Counts by Primary Site/histology, Sex, and Race/Ethnicity in SEER 12 Areas for Incidence and in U.S. for Mortality, 1992–2000 All Races Cancer Site/Type

Rate

Count

White

Black

AI/AN

API

Hispanic

White NonHispanic

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

23.0 28.1 18.7

56,812 31,399 25,413

16.8 21.8 12.8

4,895 2,854 2,041

8.0 9.0 7.3

229 116 113

14.6 17.9 11.9

4,105 2,275 1,830

18.3 22.3 14.9

6,064 3,577 2,487

23.7 29.0 19.2

43,696 24,114 19,582

3.0 3.3 2.7

7,703 4,192 3,511

2.4 2.8 2.1

879 464 415

0.4 Ÿ Ÿ

17 Ÿ Ÿ

1.0 1.2 0.8

336 183 153

2.3 2.9 1.8

1,084 634 450

3.2 3.5 2.9

5,628 3,028 2,600

20.1 24.8 16.1

49,109 27,207 21,902

14.4 19.0 10.7

4,016 2,390 1,626

7.6 8.7 6.8

212 111 101

13.6 16.7 11.1

3,769 2,092 1,677

16.0 19.3 13.0

4,980 2,943 2,037

20.5 25.5 16.3

38,068 21,086 16,982

13.6 16.7 11.1

33,313 18,184 15,129

9.2 12.2 7.0

2,548 1,493 1,055

4.6 5.0 4.2

126 63 63

7.8 9.7 6.3

2,167 1,203 964

9.9 12.0 8.2

3,013 1,742 1,271

14.1 17.3 11.4

26,109 14,262 11,847

6.4 8.1 5.0

15,796 9,023 6,773

5.1 6.8 3.7

1,468 897 571

3.1 3.6 2.6

86 48 38

5.8 7.0 4.7

1,602 889 713

6.1 7.4 4.9

1,967 1,201 766

6.4 8.2 4.9

11,959 6,824 5,135

5.3 6.7 4.2

12,685 6,869 5,816

11.5 13.3 10.3

2,699 1,278 1,421

3.7 4.3 3.3

86 43 43

3.5 4.3 2.9

923 501 422

5.1 6.4 4.2

1,304 701 603

5.2 6.7 4.1

9,703 5,306 4,397

9.4 11.7 7.8

195,563 101,337 94,226

6.3 8.2 4.9

14,006 7,686 6,320

4.7 5.6 4.1

536 279 257

5.6 7.1 4.3

2,950 1,676 1,274

7.2 9.0 5.8

9,051 5,169 3,882

9.5 11.7 7.8

176,577 91,036 85,541

0.6 0.7 0.5

11,474 6,326 5,148

0.5 0.7 0.4

1,317 769 548

0.2 0.3 0.2

35 18 17

0.2 0.3 0.1

110 70 40

0.6 0.8 0.4

869 535 334

0.6 0.7 0.5

9,997 5,433 4,564

8.9 11.0 7.3

184,089 95,011 89,078

5.8 7.5 4.5

12,689 6,917 5,772

4.5 5.3 3.9

501 261 240

5.4 6.9 4.2

2,840 1,606 1,234

6.6 8.2 5.3

8,182 4,634 3,548

8.9 11.0 7.3

166,580 85,603 80,977

3.6 4.5 3.0

74,764 38,446 36,318

7.6 9.3 6.6

15,319 7,239 8,080

3.0 3.5 2.6

317 161 156

1.9 2.3 1.6

962 512 450

3.1 3.7 2.7

3,470 1,771 1,699

3.6 4.5 3.0

67,855 34,939 32,916

incidence Lymphoma Total 21.8 66,734 Male 26.8 37,012 Female 17.7 29,722 Hodgkin lymphoma Total 2.7 9,022 Male 3.0 4,895 Female 2.4 4,127 Non-Hodgkin lymphoma Total 19.1 57,712 Male 23.7 32,117 Female 15.3 25,595 NHL—nodal Total 12.8 38,499 Male 15.7 21,113 Female 10.4 17,386 NHL—extranodal Total 6.3 19,213 Male 8.0 11,004 Female 4.9 8,209 Myeloma Total 5.6 16,496 Male 7.0 8,751 Female 4.6 7,745

mortality Lymphoma Total 9.1 213,055 Male 11.3 110,978 Female 7.4 102,077 Hodgkin lymphoma Total 0.5 12,936 Male 0.7 7,183 Female 0.4 5,753 Non-Hodgkin lymphoma Total 8.5 200,119 Male 10.6 103,795 Female 7.0 96,324 Myeloma Total 3.9 91,362 Male 4.8 46,358 Female 3.3 45,004

Rates are per 100,000 and age-adjusted to the 2000 U.S. standard million by 5-year age group. Ÿ, Statistic not displayed due to less than 16 cases. Incidence data are from SEER 12 areas except for Hispanic and white not-Hispanic which exclude Alaska, Hawaii, and Detroit. Underlying mortality data provided by NCHS (www.cdc.gov/nchs). Mortality data based on total United States except Hispanic and white non-Hispanic exclude CT, NH, LA, OK. AI/AN, American Indian/Alaska Native; API, Asian or Pacific Islander.

types are lymphocytic and myeloid/granulocytic/monocytic; the less common forms include basophilic, eosinophilic, plasma cell, erythro, and hairy cell leukemias. Acute lymphocytic leukemia has a bimodal distribution, with the peak incidence in young children and again after age 85 years; but the second peak (1.5 per 100,000) is much lower than the peak of 7.0 per 100,000 for children ages 1–4 years. The incidence of total leukemias remained stable between 1975 and 1995 and then decreased -1.8% per year from 1995 through 2000. Variations in leukemia occurrence by subtype suggest that risk factors may not be identical for the various forms of leukemia; treatment and outcome also vary by subtype (Linet and Devesa, 2002). Whites have higher rates than blacks, even by sex. The rates are not different for most subtypes by race but are about twice as high for

whites as for blacks for acute lymphocytic leukemia. Rates are higher for males than females (Table 9–15). Death rates by subtype are not presented because death certificates frequently do not specify the type of leukemia. Death rates declined for whites, blacks, and white non-Hispanics, and they increased for Hispanics. Childhood leukemia mortality decreased dramatically following improvements in treatment and survival. Five-year relative survival rates among patients with myeloid/ monocytic leukemia are 24% and are less than half those for acute lymphocytic (63.5 per 100,000) or chronic lymphocytic (73.5 per 100,000) leukemia (Appendix 9–D). Survival rates improved over the last several decades, most notably for acute lymphocytic leukemia in children.

Cancer Incidence, Mortality, and Patient Survival in the United States

Females

Males

Rate per 100,000 person-years

163

400

400

100

100

White incidence Black incidence White mortality Black mortality

10

10

1 1975 1980 1985 1990 1995 2000

1 1975 1980 1985 1990 1995 2000

Year

INCIDENCE BY AGE Age-specific incidence rates for 20 major forms of cancer by gender and race are presented in Figure 9–9. Rates for most of the solid tumors rise exponentially with age, although leveling-off or peaking of the rates before the oldest ages is apparent for several cancers. For some, such as cancers of the oral cavity and pharynx, esophagus, and lung and bronchial cancer, these peaks reflect cohort effects, with certain age groups at higher risk than those born earlier (Devesa et al., 1989, Devesa et al., 1990; Jemal et al., 2003b); these patterns reflect the prevalence of cigarette smoking, which peaked among U.S. men born during 1925–1930 and women born during 1935–1940. For prostate cancer, the patterns most likely reflect the differential use of PSA screening by age group (Hsing and Devesa, 2001). The patterns for breast, corpus uteri and uterus NOS, and ovarian cancer probably reflect hormonal influences (Gnagy et al., 2000; Brinton et al., 2002, 2004; Lacey et al., 2002; Mink et al., 2002). The plateauing of cervical cancer rates has been observed for many years, with the peak occurring at younger ages for in situ cancers and deaths occurring at older ages (Devesa, 1984; Wang et al., 2004). The relationships by race and sex seen in the age-adjusted rates generally persist in the agespecific rates, although melanoma of the skin rates are higher among white females than males before age 40, breast cancer rates are higher

Year

Figure 9–8. Non-Hodgkin lymphoma incidence (SEER9 areas) and mortality (United States) rates, age-adjusted using the 2000 U.S. standard, by gender and race, 1975–2000. Symbols present observed rates. Incidence lines estimated by delayadjusted joinpoint regression. Mortality lines estimated by joinpoint regression.

among blacks than whites before age 40, and non-Hodgkin lymphoma rates are higher among males than females at middle ages and higher among whites than blacks at older ages. The curves for Hodgkin lymphoma are unique in that rates are elevated among young adults as well as among older persons. Rates are notably elevated during childhood for three cancers: kidney and renal pelvis cancers due to Wilms tumor; brain and other CNS cancers; and leukemia, particularly acute lymphoblastic leukemias.

CANCER IN CHILDREN UNDER 20 YEARS OF AGE Incidence rates for many childhood cancers continued to increase between 1975 and 2000. Among children aged 0–19 years, increases in incidence rates of more than 20% were seen for acute lymphocytic leukemia (58%) (total leukemia 35%), brain and CNS (44%), kidney (47%), NHL (33%), and soft tissue (50%) during the period 1975–2000. The only childhood cancer that declined in incidence was Hodgkin lymphoma (-25%). Although brain cancer increased 44% over the entire time period, the increase was primarily during the mid1980s, a time of technologic advances (Smith et al., 1998); there has actually been a decrease of -0.6% per year during 1987–2000 (Ries

164

PART II: THE MAGNITUDE OF CANCER

Table 9–15. Leukemia by Subtype: Cancer Incidence and Death Rates and Counts by Sex and Race/Ethnicity in SEER 12 Areas for Incidence and in U.S. for Mortality, 1992–2000 All Races Cancer/Type

Rate

Count

White

Black

AI/AN

API

White NonHispanic

Hispanic

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

Rate

Count

13.0 17.2 10.0

32,017 18,290 13,727

10.1 13.0 8.0

2,644 1,438 1,206

4.5 5.0 4.0

163 86 77

8.0 10.0 6.5

2,342 1,326 1,016

9.5 11.8 7.7

3,768 2,138 1,630

13.0 17.3 9.8

23,940 13,795 10,145

6.2 8.4 4.6

15,345 9,112 6,233

4.2 5.8 3.0

1,105 652 453

1.8 1.7 1.8

75 38 37

2.3 3.0 1.7

733 450 283

4.1 5.1 3.2

1,920 1,125 795

6.3 8.6 4.5

11,438 6,847 4,591

1.6 1.8 1.4

4,187 2,375 1,812

0.8 1.0 0.7

330 185 145

0.9 0.8 1.0

54 27 27

1.3 1.5 1.1

446 257 189

2.0 2.3 1.8

1,437 828 609

1.4 1.6 1.2

2,336 1,318 1,018

4.1 5.8 2.9

9,869 5,819 4,050

3.1 4.5 2.1

711 425 286

0.8 Ÿ Ÿ

18 Ÿ Ÿ

0.8 1.2 0.6

230 149 81

1.8 2.4 1.3

409 240 169

4.3 6.1 3.0

8,050 4,787 3,263

0.5 0.8 0.3

1,289 918 371

0.3 0.4 0.2

64 42 22

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.2 0.4 Ÿ

57 44 Ÿ

0.3 0.5 0.1

74 57 17

0.6 0.9 0.3

1,052 742 310

5.9 7.5 4.7

14,463 7,993 6,470

5.0 6.1 4.3

1,330 678 652

2.3 3.0 1.9

78 43 35

5.0 6.2 4.1

1,450 798 652

4.7 5.9 3.9

1,677 929 748

5.8 7.5 4.6

10,834 6,047 4,787

3.7 4.6 3.1

9,139 4,936 4,203

3.0 3.5 2.7

793 379 414

1.6 2.2 1.2

53 29 24

3.3 3.9 2.9

960 502 458

2.9 3.5 2.5

1,060 561 499

3.7 4.7 3.1

6,891 3,760 3,131

0.2 0.3 0.2

601 348 253

0.2 0.2 0.1

43 21 22

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.2 0.2 0.2

62 34 28

0.1 0.2 0.1

54 28 26

0.3 0.3 0.2

469 275 194

1.8 2.3 1.3

4,330 2,484 1,846

1.6 2.3 1.2

446 259 187

0.7 Ÿ Ÿ

23 Ÿ Ÿ

1.4 1.9 1.0

398 244 154

1.5 2.1 1.1

528 322 206

1.7 2.3 1.3

3,183 1,830 1,353

0.2 0.2 0.1

393 225 168

0.2 0.2 0.2

48 19 29

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.1 0.1 Ÿ

30 18 Ÿ

0.1 0.1 0.1

35 18 17

0.2 0.2 0.1

291 182 109

0.9 1.2 0.7

2,209 1,185 1,024

0.8 1.1 0.7

209 108 101

Ÿ Ÿ Ÿ

Ÿ Ÿ Ÿ

0.7 0.7 0.6

159 78 81

0.7 0.8 0.6

171 84 87

0.9 1.2 0.7

1,668 901 767

7.9 10.7 6.1

164,657 91,181 73,476

7.1 9.4 5.5

15,355 8,211 7,144

4.2 5.4 3.4

547 295 252

4.4 5.5 3.5

2,635 1,472 1,163

5.3 6.6 4.3

7,805 4,344 3,461

8.0 10.8 6.1

148,690 82,358 66,332

incidence Leukemias Total 12.4 37,493 Male 16.2 21,344 Female 9.5 16,149 Lymphocytic leukemia Total 5.7 17,509 Male 7.8 10,409 Female 4.2 7,100 Acute lymphocytic leukemia Total 1.5 5,042 Male 1.7 2,858 Female 1.3 2,184 Chronic lymphocytic leukemia Total 3.8 11,031 Male 5.3 6,528 Female 2.7 4,503 Other lymphocytic leukemia Total 0.5 1,436 Male 0.8 1,023 Female 0.2 413 Myeloid and monocytic leukemia Total 5.7 17,385 Male 7.3 9,550 Female 4.6 7,835 Acute myeloid leukemia Total 3.6 10,968 Male 4.5 5,862 Female 3.0 5,106 Acute monocytic leukemia Total 0.2 707 Male 0.3 403 Female 0.2 304 Chronic myeloid leukemia Total 1.7 5,237 Male 2.3 3,022 Female 1.3 2,215 Other myeloid/monocytic leukemia Total 0.2 473 Male 0.2 263 Female 0.1 210 Other leukemias Total 0.9 2,599 Male 1.2 1,385 Female 0.7 1,214

mortality Leukemia Total Male Female

7.8 10.5 6.0

183,194 101,159 82,035

Rates are per 100,000 and age-adjusted to the 2000 U.S. standard million by 5-year age group. Ÿ, Statistic not displayed due to less than 16 cases. Incidence data are from SEER 12 areas except for Hispanic and white not-Hispanic which exclude Alaska, Hawaii, and Detroit. Underlying mortality data provided by NCHS (www.cdc.gov/nchs). Mortality data based on total United States except Hispanic and white non-Hispanic exclude CT, NH, LA, OK. AI/AN, American Indian/Alaska Native; API, Asian or Pacific Islander.

et al., 2003). For the same period, 1975–2000, however, overall childhood cancer mortality decreased -45% owing to substantial declines in mortality for every major childhood cancer. Mortality decreased by more than 50% for Hodgkin lymphoma (-71%), kidney cancer (-53%), leukemia (-55%), and non-Hodgkin lymphoma (-68%). These decreases in mortality in the face of a rising incidence are attributable to large increases in the survival rates due to improved treatment regimens.

DISCUSSION Men develop and die from cancer with greater frequency than women. Tobacco use and alcohol consumption as well as occupational exposures contribute to the higher rates of mouth, lung, esophageal, laryngeal, and bladder cancers in men. Sex differences for other malignancies, such as stomach, colon, and rectal cancers, which predominate among men, and breast, gallbladder, and thyroid cancers,

1000 Oral cavity and pharynx

Esophagus

Stomach

Colon

Rectum

Pancreas

Lung and bronchus

Melanoma of skin

Breast

Cervix uteri

Corpus uteri and

100

10

1

0.1 1000

100

10

Rate per 100,000 person-years

1

0.1 1000

2

Ovary

1

uterus NOS

100

10

1

0.1 1000

Prostate gland

Bladder

Kidney and renal pelvis

Brain and other nervous system

Non-Hodgkin lymphoma

Multiple myeloma

Leukemia

3

100

10

1

0.1 1000

Hodgkin lymphoma

100

10

1

0.1 0

20 40 60 80 100

0

20 40 60 80 100

0

20 40 60 80 100

0

20 40 60 80 100

Age White males

White females

Black males

Black females

1

NOS = not otherwise specified 2 Ovary omitting borderline histologies 3 Prostate gland: Y-axis extends to 2000

Figure 9–9. Age-specific incidence rates for certain cancers by gender and race, SEER11 areas, 1992–2000.

165

166

PART II: THE MAGNITUDE OF CANCER

which predominate among women, have given rise to the investigation of other factors such as biologic differences, nutrition, parity, endogenous hormones, and sex steroid receptors, in attempts to explain such differences (Fraumeni et al., 1993). As the population continues to grow in size and shifts to an older age distribution, increasing numbers of people are affected by cancer. Because of the predicted population growth and aging in the U.S. population, it is estimated that the number of newly diagnosed cancer patients will double from 1.3 million to 2.6 million persons between 2000 and 2050 even if the cancer rate remains the same (Edwards et al., 2002). Progress against cancer continues on many fronts: prevention, early detection, diagnosis, treatment, and quality of life (U.S. Department of Health and Human Services, 2001). Primary prevention can progress with the identification of factors that influence risk, followed by modifications of exposures. As suggested by stage-specific survival rates, substantial declines in cancer mortality could also be achieved by shifting diagnoses to earlier, less advanced, more treatable stages of disease. When advances from laboratory research are translated into new and better treatments and management across the life course of the patient’s disease, death rates will decline as survival rates increase. The identification of incidence, mortality, and survival differentials by race, ethnicity, and age, among other factors, can be used to eliminate these disparities.

CONCLUSIONS The intent of this chapter has been to show the current estimates for cancer incidence, mortality, and survival in a large subgroup of the U.S. population. Trends in cancer incidence, mortality, and survival were presented when they appeared to be important. More information on cancer trends is provided in the most recent SEER Cancer Statistics Review (Ries et al., 2003). The SEER home page on the Internet (www.seer.cancer.gov) contains the most recent SEER publications and cancer statistics. Geographic variation in cancer mortality rates among whites and blacks is presented in the NCI’s recent Atlas (Devesa et al., 1999). More detailed and updated data can be accessed on the Internet (http://www3.cancer.gov/atlasplus/). Long-term incidence trends by geographic area are not yet possible. Cancer surveillance, however, is improving in the United States as more and more states collect high quality cancer incidence data for longer periods through the NCI’s SEER Program or the Centers for Disease Control and Prevention’s National Program of Cancer Registries. Data from these combined programs for 37 states, 6 metropolitan areas, and the District of Columbia were published for more than 1 million cancer cases for 1999 (U.S. Cancer Statistics Working Group, 2003). The reader is referred to other chapters in this book for additional information on specific cancers regarding changes in diagnostic procedures, new treatments, and effects of screening programs as well as discussions of known and suspected risk factors. Acknowledgment We thank Milton Eisner (NCI) for his helpful editorial comments and John Lahey (IMS) for the preparation of graphics.

References American Joint Committee on Cancer; Beahrs OH, Henson DE, Hutter RV, Myers MH, editors. 1988. Manual for Staging of Cancer, 3rd ed. Philadelphia: Lippincott. Anderson RN, Rosenberg HM. 1998, Age standardization of death rates: implementation of the year 2000 standard. Natl Vital Stat Rep 47:1–16. Axtell LM, Asire AJ, Myers MH. 1976. Cancer Patient Survival: Report Number 5. NIH Publ. 81-992. Bethesda: National Cancer Institute. Brinton LA, Lacey J, Devesa SS. 2002. Epidemiology of breast cancer. In: Donegan WL, Spratt JS, editors. Cancer of the breast, 5th ed. Orlando, FL: Saunders, pp. 111–132. Brinton LA, Lacey JV, Sherman ME, Devesa SS. 2004. Epidemiology of corpus uteri cancer. In: Gershenson DM, Gore M, McGuire WP, Quinn M, Thomas G, editors. Gynecologic Cancer: Controversies in Management. Amsterdam: Elsevier Science.

Brown LM, Devesa SS. 2002. Epidemiologic trends in esophageal and gastric cancer in the United States. Surg Oncol Clin N Am 11:235–256. Canto MT, Devesa SS. 2002. Oral cavity and pharynx cancer incidence rates in the United States, 1975–1998. Oral Oncol 38:610–617. CDC, National Center for Chronic Disease Prevention and Health Promotion. 2002. Behavioral Risk Factor Surveillance System, Nationwide. 2002 Women’s Health: Prevalence Data. http://www.cdc.gov/. Chow W-H, Devesa SS. 1992. Death certificate reporting of colon and rectal cancers [letter to the editor]. JAMA 267:3028. Chow WH, Devesa SS, Warren JL, Fraumeni JF Jr. 1999. Rising incidence of renal cell cancer in the United States. JAMA 281:1628–1631. Clegg LX, Feuer EJ, Midthune D, Fay MP, Hankey BF. 2002. Impact of reporting delay and reporting error on cancer incidence rates and trends. J Natl Cancer Inst 94:1537–1545. Connelly RR, Spirtas R, Myers MH, et al. 1987. Demographic patterns for mesothelioma in the United States. J Natl Cancer Inst 78:1053–1060. Cote TR, OíBrien TR, Ward JW, Wilson SE, Blattner WA. 1995. AIDS and cancer registry linkage: measurement and enhancement of registry completeness; the National AIDS/Cancer Match Study Group. Prev Med 24:375–377. Cutler SJ, Young JL, editors. 1975. Third National Cancer Survey: Incidence Data. National Cancer Institute Monograph 41. Washington, DC: U.S. Government Printing Office. Devesa SS. 1984. Descriptive epidemiology of cancer of the uterine cervix. Obstet Gynecol; 63:605–612. Devesa SS, Chow W-H. 1993. Variation in colorectal cancer incidence in the United States by subsite of origin. Cancer 71:3819–3826. Devesa SS, Fears T. 1992. Non-Hodgkin’s lymphoma time trends: United States and international data. Cancer Res 52(Suppl):5432s–5440s. Devesa SS, Blot WJ, FraumenI JF Jr. 1989. Declining lung cancer rates among young men and women in the United States: a cohort analysis. J Natl Cancer Inst 81:1568–1571. Devesa SS, Blot WJ, Fraumeni JF Jr. 1990. Cohort trends in mortality from oral, esophageal, and laryngeal cancer in the United States. Epidemiology 1:116–121. Devesa SS, Donaldson J, Fears T (1995). Graphic presentation of trends in rates. Am J Epidemiol 141:300–304. Devesa SS, Grauman DG, Blot WJ, Pennello G, Hoover RN, Fraumeni JF Jr. 1999. Atlas of Cancer Mortality in the United States, 1950–94. NIH Publ. No. (NIH)99_4564. Washington, DC: US Government Printing Office (http://www3.cancer.gov/atlasplus/). Devesa SS, Shaw GL, Blot WJ. 1991. Changing patterns of lung cancer incidence by histological type. Cancer Epidemiol Biomarkers Prev 1:29–34. Devesa SS, Silverman DT, McLaughlin JK, Brown CC, Connelly RR, Fraumeni JF Jr, 1990. Comparison of the descriptive epidemiology of urinary tract cancers. Cancer Causes Control 1:133–141. Devesa SS, Silverman DT, Young JL Jr, et al. 1987. Cancer incidence and mortality trends among whites in the United States, 1947–84. J Natl Cancer lnst 79:701–770. DeVita VT, Hellman S, Rosenberg SA. 1997. Cancer: Principles and Practice of Oncology, 5th ed. Philadelphia: Lippincott. Dorn HF, Cutler SJ. 1959. Morbidity from cancer in the United States: Parts I and H. Public Health Monograph 56-1-207. Bethesda: DHEW. Ederer F, Axtell LM, Cutler SJ. 1961. The relative survival rate: a statistical methodology. In: Cutler SJ, Ederer F, editors. End Results and Mortality Trends in Cancer. NCI Monograph 6. Bethesda: National Cancer Institute, pp. 101–121. Edwards BK, Howe HL, Ries LAG, Thun MJ, Rosenberg HM, Yancik R, Wingo PA, Jemal A, Feigal EG. 2002. Annual report to the nation on the status of cancer, 1973–1999, featuring implications of age and aging on U.S. cancer burden. Cancer 94:2767–2792 Eltom MA, Jemal A, Mbulaiteye SM, Devesa SS, Biggar RJ. 2002. Cancer surveillance series: trends in Kaposi’s sarcoma and non-Hodgkin’s lymphoma incidence in the United States from 1973 through 1998. J Natl Cancer Inst 94:1204–1210. Fraumeni JF Jr, Devesa SS, Hoover RN, Kinlen LJ. 1993. Epidemiology of cancer. In: DeVita VT, Hellman S, Rosenberg SA, editors. Cancer: Principles and Practice of Oncology, 4th ed. Philadelphia: Lippincott, pp. 150–181. Gnagy S, Ming EE, Devesa SS, Hartge P, Whittemore AS. 2000. Declining ovarian cancer rates in U.S. women: relation to parity and oral contraceptive use. Epidemiology 11:102–105. Goodman MT, Correa CN, Tung K-H, Roffers SD, Wu XC, Young JL, Wilkens LR, Carney ME, Howe HL. 2003. Stage at diagnosis of ovarian cancer in the United States, 1992–1997. Cancer 97(Suppl):2658–2659. Groves FD, Linet MS, Travis LB, Devesa SS. 2000. Cancer Surveillance Series: non-Hodgkin’s lymphoma incidence by histologic subtype in the United States, 1978–1995. J Natl Cancer Inst 92:1240–1251.

Cancer Incidence, Mortality, and Patient Survival in the United States Hartge P, Devesa SS. 1992. Quantification of the impact of known risk factors and time trends in non-Hodgkin’s lymphoma incidence. Cancer Res 52(Suppl):5566s–5569s. Hill HA, Eley JW, Harlan LC, Greenberg RS, Barrett RJ Jr, Chen VW. 1996. Racial differences in endometrial cancer survival: the black/white cancer survival study. Obstet Gynecol 88:919–926. Hsing A, Devesa SS. 2001. Trends and patterns of prostate cancer: what do they suggest? Epidemiol Rev 23:3–13. Jemal A, Devesa SS, Hartge P, Tucker MA. 2001. Cancer Surveillance Series: recent trends in cutaneous melanoma incidence among whites in the United States. J Natl Cancer Inst 93:678–683. Jemal A, Murray T, Samuels A, Ghafoor A, Ward E, Thun M. 2003a. Cancer statistics. CA Cancer J Clin 53:5–26. Jemal A, Travis WD, Tarone RE, Travis L, Devesa SS. 2003b. Lung cancer rates convergence in young men and women in the U.S.: an analysis by birth cohort and histologic type. Int J Cancer 105:101–107. Kim H-J, Fay MP, Feuer EJ, Midthune DN. 2000. Permutation tests for joinpoint regression with applications to cancer rates. Stat Med 19:335–351. Lacey JV, Devesa SS, Brinton LA. 2002. Recent trends in breast cancer incidence and mortality. Environ Mol Mutagen 39:82–88. Legler JM, Ries LAG, Smith MA, Warren JL, Heineman EF, Kaplan RS, Linet MS. 1999. Cancer surveillance series: brain and other central nervous system cancers: recent trends in incidence and mortality. J Natl Cancer Inst 91:1382–1390. Linet MS, Devesa SS. 2002. Epidemiology of leukemia: overview and patterns of occurrence. In: Henderson E, Lister TA, Greaves MF, editors. Leukemia, 7th ed. Philadelphia: Saunders, pp. 131–151. Mariotto A, Gigli A, Capocaccia R, Clegg L, Scoppa S, Ries LA, Tesauro GS, Rowland JS, Feuer EJ. 2002. Complete and Limited Duration Prevalence Estimates. SEER Cancer Statistics Review, 1973–1999, p. 19 (http://seer.cancer.gov/csr/1973_1999). Mathew A, Devesa SS, Fraumeni JF Jr, Chow W-H. 2002. Global increases in kidney cancer incidence, 1973–1992. Eur J Cancer Prev 11:171–178. McGlynn KA, Devesa SS, Sigurdson AJ, Brown LM, Tsao L, Tarone RE. 2003. Trends in the incidence of testicular germ cell tumors in the United States. Cancer 97:63–70. McGlynn KA, Tsao L, Hsing AW, Devesa SS, Fraumeni JF Jr. 2001. International trends and patterns of primary liver cancer. Int J Cancer 94:290– 296. Miller BA, Feuer EJ, Hankey BF. 1993. Recent incidence trends for breast cancer in women and the relevance of early detection: an update. CA Cancer J Clin 43:27–41. Mink PJ, Sherman ME, Devesa SS. 2002. Incidence patterns of invasive and borderline ovarian tumors among white and black women in the United States: results from the SEER program, 1978–1998. Cancer 95:2380– 2389. Parkin DM, Whelan SL, Ferlay J, Teppo L, Thomas DB, editors. 2003. Cancer Incidence in Five Continents, Vol. VIII. IARC Scientific Publ. No. 155. Lyon: IARC. Percy CL, Miller BA, Ries LAG. 1990a. Effect of changes in cancer classification and the accuracy of cancer death certificates on trends in cancer mortality. Ann NY Acad Sci 609:87–99. Percy CL, Ries LAG, Van Holten VD. 1990b. The accuracy of liver cancer as the underlying cause of death on death certificates. Public Health Rep 105:361–368.

167

Percy C, Stanek E III, Gloeckler L. 1981. Accuracy of cancer death certificates and its effect on cancer mortality statistics. Am J Public Health 71: 242–250. Percy C, Van Holten VV, Muir C, editors. 1990. International Classification of Diseases for Oncology, 2nd ed. Geneva: WHO. Pokras R, Hufnagel VG. 1987. Hysterectomies in the United States. DHHS Publ. No. (PHS) 88-1753. Hyattsville, MD: National Center for Health Statistics. Potosky AL, Kessler L, Gridley G, Brown CC, Horm JW. 1990. Rise in prostatic cancer incidence associated with increased use of transurethral resection. J Natl Cancer Inst 82:1624–1628. Potosky AL, Miller BA, Albertsen PC, Kramer BS. 1995. The role of increasing detection in the rising incidence of prostate cancer. JAMA 273:548– 552. Prout GR Jr, Wesley MN, Greenberg RS, Chen VW, Brown CC, Miller AW, Weinstein RS, Robboy SJ, Haynes MA, Blacklow RS, Edwards BK. 2000. Bladder cancer: race differences in extent of disease at diagnosis. Cancer 89:1349–1358. Ries LAG, Eisner MP, Kosary CL, Hankey BF, Miller BA, Clegg L, Mariotto A, Fay MP, Feuer EJ, Edwards BK, editors. 2003. SEER Cancer Statistics Review, 1975–2000. Bethesda: National Cancer Institute (http://seer.cancer.gov/csr/1975_2000). Scotto J, Fears TR, Fraumeni JF Jr. 1983. Incidence of Non-melanoma Skin Cancer in the United States. NIH Publ. 83-2433. Bethesda: National Cancer Institute. SEER*Stat. 2003. NCI software package to access SEER public-use file: 1973–2000 (http:seer.cancer.gov). Sherman ME, Devesa SS. 2003. Analysis of racial differences in incidence, survival, and mortality for malignant tumors of the uterine corpus. Cancer 98:176–186. Smith MA, Freidlin B, Ries LA, Simon R. 1998. Trends in reported incidence of primary malignant brain tumors in children in the United States. J Natl Cancer Inst 90:1269–1277. Troisi R, Freedman A, Devesa S. 1999. Incidence of colorectal carcinoma in the United States: an update of trends by gender, race, age, subsite, and stage, 1975–1994. Cancer 85:1670–1676. U.S. Cancer Statistics Working Group. 2003. United States Cancer Statistics: 2000 Incidence. Atlanta: Department of Health and Human Services, Centers for Disease Control and Prevention and National Cancer Institute. U.S. Department of Health and Human Services (DHHS). 2001. Cancer Progress Report: 2001. NIH Publ. No. 02-5045. Bethesda: National Cancer Institute. Wang SS, Sherman ME, Hildesheim A, Lacey JV, Devesa S. 2004. Cervical adenocarcinoma and squamous cell carcinoma incidence trends among white women and black women in the United States for 1976–2000. Cancer 100:1035–1044. Weir HK, Thun MJ, Hankey BF, Ries LAG, Howe HL, Wingo PA, Jemal A, Ward E, Anderson RN, Edwards BK. 2003. Annual report to the nation on the status of cancer, 1975–2000, featuring the uses of surveillance data for cancer prevention and control. J Natl Cancer Inst 95: 1276–1299. World Health Organization. 1976/1986/1988. International Classification of Diseases for Oncology, 1st ed. Geneva: WHO (field trials 1986 and 1988).

Appendix 9–A. Short-Term Trends (1992–2000) in Cancer Incidence Rates Expressed as Annual Percent Change: SEER Program Cancer Site/Type

All Races

White

Black

AI/AN

API

Hispanic

White NonHispanic

Appendix 9–A. (cont.) Cancer Site/Type

All Races

White

Black

1.6* 1.5 1.7*

3.6* 3.3* 4.0*

AI/AN

API

— — —

1.8 3.2 0.8

Hispanic

White NonHispanic

Ascending colon ALL MALIGNANT CANCERS Total Male Female

-0.7* -1.7* 0.2

-0.6* -1.8* 0.4

-1.0* -1.9* -0.1

-3.0* -4.8* -1.2

-0.5* -1.1* 0.4

-0.5 -1.2* 0.2

-0.5 -1.7* 0.5

-2.6* -3.4* -1.7

-1.9* -2.2* -1.8*

ORAL CAVITY AND PHARYNX Total Male Female

Lip Total Male Female

Tongue Total Male Female

-2.0* -2.3* -1.5*

-1.9* -2.2* -1.7*

-3.2* -3.2* -2.6

— — —

-0.7 -0.9 0.0

-4.5* -5.5* -1.9

-4.6* -5.6* -1.8

— — —

— — —

— — —

— — —

-5.0* -5.9* -2.4

-0.3 -0.5 -0.2

-0.2 -0.1 -1.0

-2.7 -4.0 0.5

— — —

3.0 0.0 —

0.5 -2.0 —

-0.1 0.2 -1.4

0.9 0.8 0.9

1.3 1.1 1.3

-1.5 — —

— — —

-0.3 — —

-4.3 — —

1.8 1.3 2.2

-5.0* -5.9* -3.1

-4.8* -5.8* -3.0

-5.4 -6.0 —

— — —

— — —

— — —

-4.6* -5.4* -3.3

-3.7* -4.6* -2.3*

-3.9* -4.8* -2.5*

-4.2* -5.0 —

— — —

-2.1 — —

-1.2 -1.7 -0.5

-1.4 -1.5 -2.1

-1.9 — —

— — —

-2.8 -3.2 -1.7

— — —

-1.6 -2.1 -1.4

0.5 2.0* -4.5*

0.8 2.3* -4.3*

-0.3 1.0 —

— — —

— — —

-2.3 -1.5 —

1.2 3.1* -5.0*

-1.6 -2.0 -0.3

-2.1 -3.0 0.0

— — —

— — —

— — —

— — —

-1.7 -3.1 —

-2.9* -3.0* -3.5

-2.4* -2.6 -2.8

-6.6* -5.7 —

— — —

-0.3 — —

— — —

-2.3 -2.5 -2.5

-5.7* -6.9* -3.5

— — —

— — —

— — —

— — —

-6.1* -6.9* -4.4*

-0.2 -0.5 -0.1

-0.6* -0.9* -0.2

-0.8 -3.0 1.4

-0.9* -0.9* -0.6

0.4 0.2 0.5

-0.2 -0.5 0.0

Salivary gland Total Male Female

Floor of mouth Total Male Female

Gum and other mouth Total Male Female

Nasopharynx Total Male Female

Tonsil Total Male Female

Oropharynx Total Male Female

Hypopharynx Total Male Female

-5.0* -3.2 —

Other oral cavity and pharynx Total Male Female

-5.8* -6.5* -4.7*

DIGESTIVE SYSTEM Total Male Female

-0.2 -0.5 -0.1

Esophagus Total Male Female

Stomach Total Male Female

-4.1* -5.2* -2.7*

0.3 0.4 -0.4

1.5* 1.8* 0.1

-5.3* -6.4* -3.1*

— — —

0.7 0.0 —

0.6 0.0 —

1.9* 2.1* 0.4

-1.3* -1.9* -0.5

-1.5* -2.1* -0.8

-1.0 -2.0 0.5

3.4 — —

-2.8* -2.9* -2.6*

-1.6* -2.1 -1.0

-1.9* -2.4* -1.4

0.8 -0.1 2.1

0.5 -0.2 1.3

— — —

1.0 — —

1.8 — —

0.7 -0.2 1.5

Small intestine Total Male Female

Colon and rectum Total Male Female

-0.6 -0.9* -0.3

-0.6 -1.0* -0.3

-0.2 -0.4 0.1

-1.9 -3.4 -0.2

-0.5* -0.3 -0.5

0.8 0.9 0.7

-0.5 -0.9* -0.3

-0.7* -1.1* -0.4*

-0.8* -1.2* -0.5

-0.1 -0.3 0.1

-3.3 -5.2 -1.3

-0.6 -0.6 -0.4

1.0 1.4 0.5

-0.7 -1.1* -0.3

-0.6 -0.9 -0.3

-0.4 -0.7 -0.2

-0.5 -0.9 -0.2

— — —

-0.1 -1.3 1.0

0.8 0.5 1.1

-0.3 -0.6 0.0

— — —

— — —

— — —

Colon excluding rectum Total Male Female

Cecum Total Male Female

3.3* 1.9 5.6*

Appendix Total Male Female

168

4.0* 3.3* 4.7*

4.8* 3.3 6.0*

— — —

3.7* 2.8 4.6*

Total Male Female

1.7* 1.7* 1.7*

Hepatic flexure Total Male Female

Splenic flexure Total Male Female

0.4 3.0 -1.6

— — —

-0.6 0.9 -1.9

2.0 — —

0.5 -1.0 1.8

-0.4 -0.4 -0.5

-0.7 -0.6 -0.9

-0.3 -0.3 -0.1

— — —

2.4 1.1 3.5

1.2 4.0 -1.5

-0.4 -0.3 -0.6

-0.9 -1.5 -0.5

-1.0 -1.7 -0.5

-0.1 -0.9 -0.1

— — —

-1.2 — —

0.2 — —

-1.3 -1.7 -0.9

-2.3* -3.3* -1.5*

-2.3* -3.0* -1.8*

-3.2 -6.0* -1.1

— — —

-2.3 -2.8 -1.5

1.1 — —

-1.8* -2.5* -1.3

-1.9* -2.1* -1.8*

-2.1* -2.4* -2.0*

-1.3* -1.0 -1.4*

— — —

-1.1 -1.0 -1.0

-0.3 0.8 -1.7

-2.2* -2.6* -1.9*

-1.8 -2.2 -1.6

1.3 0.8 1.3

— — —

-7.8* — —

2.5 — —

-1.3 -1.7 -1.3

-0.2 -0.5 0.1

-0.4 -0.7 0.0

1.7 — —

-0.4 0.1 -0.8

0.4 -0.1 1.0

-0.2 -0.6 -0.1

-2.7* -3.1* -2.4*

-2.8* -3.3* -2.4*

-2.8 -1.9 -3.2

— — —

-2.3 -2.1 -2.1

-1.8 -1.6 -2.5

-3.0* -3.6* -2.5*

1.1* 0.9* 1.3

1.2* 0.9 1.3

0.9 0.0 1.7

— — —

0.6 1.2* 0.0

1.5 0.7 2.7*

1.1* 0.9 1.2

0.1 -0.1 —

— — —

— — —

2.3 — —

3.2* 4.0* 2.8*

Descending colon Total Male Female

Sigmoid colon Total Male Female

Large intestine, NOS Total Male Female

-2.0 -2.6 -1.7

Rectum and rectosigmoid junction Total Male Female

-0.2 -0.4 0.0

Rectosigmoid junction Total Male Female

1.8* 1.9* 1.7*

0.3 -1.0 1.5*

0.3 -0.5 0.9

Transverse colon Total Male Female

2.6* 2.4 3.0

Rectum Total Male Female

Anus, anal canal, anorectum Total Male Female

1.8* 2.4 1.4

2.5* 3.2* 2.0

Liver and intrahepatic bile duct Total Male Female

3.9* 3.7* 3.7*

4.5* 4.1* 4.5*

3.5* 4.3* 2.3

— — —

0.1 0.7 -1.2

3.5* 2.5 5.0

4.0* 3.5* 4.0*

4.3* 4.0* 4.4*

5.1* 4.6* 5.2*

3.7* 4.4* 2.6

— — —

0.2 0.6 -0.6

4.5* 3.5 6.5*

4.1* 3.7* 4.1*

1.8 1.8 1.3

2.3* 1.8 2.4

— — —

— — —

-0.9 — —

-3.0* — —

3.3 2.4 3.7

-2.6* -2.6 -2.3*

-2.7* -2.0 -2.6*

-0.7 — 0.2

— — —

-5.1* — -3.4

-3.3 — -4.3

-3.5* -4.6 -2.5*

-0.6 -0.5 -0.7

-0.4 -0.4 -0.5

-1.2 — —

— — —

-1.8 -1.1 -2.3

-0.4 1.3 -1.5

-0.7 -0.9 -0.5

-0.4 -0.4 -0.5

-0.3 -0.2 -0.5

-1.8* -1.0* -2.7*

— — —

-0.1 -2.5 2.6*

-0.6 -1.1 -0.1

-0.1 0.3 -0.7

-1.3 1.0 -2.9*

-0.4 2.1* -2.5

— — —

— — —

— — —

— — —

-0.4 0.7 -1.1

8.6* -3.5 11.5*

— — —

— — —

— — —

— — —

10.1* -1.2 12.7*

0.7 -0.8 2.1

— — —

— — —

— — —

— — —

0.0 -1.4 1.3

Liver Total Male Female

Intrahepatic bile duct Total Male Female

Gallbladder Total Male Female

Other biliary Total Male Female

Pancreas Total Male Female

Retroperitoneum Total Male Female

Peritoneum, omentum, mesentery Total Male Female

8.6* -1.6 11.0*

Other digestive organs Total Male Female

1.0 -0.4 2.4

Appendix 9–A. (cont.) Cancer Site/Type

All Races

Appendix 9–A. (cont.)

White

RESPIRATORY SYSTEM -1.3* -2.4* 0.0

Total Male Female

-1.2* -2.3* 0.1

Nose, nasal cavity, middle ear Total Male Female

Larynx Total Male Female

Pleura Total Male Female

AI/AN

API

-1.5* -2.6* 0.6

-4.4* -4.6 -4.1*

-0.7 -0.8 -0.1

Hispanic

-1.6* -2.3* -0.9

-2.6 — —

— — —

— — —

-2.6* -3.2* -1.0

-2.4* -3.1* -0.7

-3.3* -3.4* -2.3

— — —

-1.7 -1.8 —

-1.0 -2.1* —

-2.5* -3.1* -1.2

-1.2* -2.3* 0.0

-1.1* -2.3* 0.1

-1.4* -2.5* 0.8

-4.9* -5.4 -4.1*

-0.7 -0.7 -0.2

-1.8* -2.4* -1.0

-1.0* -2.1* 0.2

-1.0 -1.6* 0.1

-0.8 -1.4 0.2

— — —

— — —

— — —

-3.3 — —

-0.9 -1.7 0.6

-3.1* -5.1* -0.1

BONES AND JOINTS -1.4 -1.6 -1.1

Total Male Female

-1.1* -2.2* 0.1

-0.3 -0.3 -0.8

— — —

Trachea, mediastinum, other respiratory Total Male Female

White NonHispanic

-0.5 -0.6 -1.1

Lung and bronchus Total Male Female

Black

-0.5 -0.9 -0.8

-3.5* -5.4* —

— — —

— — —

— — —

— — —

-6.3* -8.9* —

-1.0 -1.8* 0.0

-3.0 — —

— — —

-0.9 — —

-1.8 -1.3 -2.4

-0.6 -1.4 0.5

— — —

1.8 4.3* -1.0

2.2 -0.1 4.7

1.1 0.5 1.7

SOFT TISSUE INCLUDING HEART Total Male Female

1.3* 1.1 1.7

1.2* 0.7 1.9*

1.3 1.4 1.8

-0.5 -1.7* 2.1* 2.5* 2.8* 2.3*

Other nonepithelial skin Total Male Female

-6.4* -8.0* 0.9

— — —

-1.2 -4.0 4.1

-6.5* -10.6* 2.1

0.1 -1.3* 2.8*

3.0* 3.1* 2.9*

— — —

— — —

6.6 — —

3.9* 3.6 4.1*

2.9* 3.0* 2.8*

-8.5* -10.9* 2.2

— — —

-9.5* -11.8* —

-17.0* -19.2* —

-13.0* -15.8* 2.2

0.2 — 0.2

-3.9* — -3.7*

2.2* — 2.1*

1.2* — 1.3*

0.9* 3.1* 1.1*

-1.1 -2.7* 0.2

-2.0 — —

-1.0 -5.4* 1.4

-1.4* -3.0* 0.6

-0.7* -2.4* -0.1

0.6* 1.1 0.8*

0.7* 1.9 0.9*

BREAST

FEMALE GENITAL SYSTEM -0.8* -2.5* -0.1

-0.6* -2.1* 0.0

0.0 1.7

0.4 —

— —

1.5 —

0.6 —

-0.2 2.0

-1.1* -1.2 1.0 -1.5

-1.1* -0.7 1.7 -1.4

-1.0 — -1.8 —

— — — —

0.0 — — —

-1.9* — -1.5 —

-0.9* -2.5 1.8 -0.2

-3.0* -3.1* 1.3* -2.6 1.3

Urinary bladder Total Male Female

0.3* 0.2 0.3 -0.1 -0.1 -0.5

1.4* 1.3* 1.3*

1.8 1.0 3.1

-6.3* -5.3 —

1.6 0.9 2.8

1.8 0.3 3.7

Ureter Total Male Female

-1.0 -1.3 -0.3

— — —

— — —

— — —

— — —

-0.2 -0.1 -0.5

0.5 -0.3 2.1

— — —

— — —

— — —

— — —

0.4 -0.6 —

-1.3 -2.1 -0.9

— — —

— — —

— — —

-1.8 — —

-1.4 -2.3 -0.8

1.3* 1.1* 1.3* -1.2 -1.8 0.0

-3.3* -3.4* 1.3* -3.1 -0.4

-2.2* -2.3* — — —

1.5* 1.5* 1.1*

Other urinary organs Total Male Female

-0.3 -0.3 -0.1

EYE AND ORBIT Total Male Female

-1.5 -2.1 -1.3

BRAIN AND OTHER NERVOUS SYSTEM Total Male Female

Brain Total Male Female

-0.5 -0.6 -0.5

-0.3 -0.4 -0.2

0.0 1.2 -0.9

— — —

-2.4* -1.6 -3.0

0.3 0.2 0.0

0.0 -0.1 0.0

-0.4 -0.4 -0.5

-0.1 -0.2 -0.1

0.1 1.4 -1.2

— — —

-2.2* -1.8 -2.6

0.3 0.9 -0.5

0.1 0.1 0.1

— — —

— — —

— — —

— — —

-1.3 -2.0 -0.9

Cranial nerves, other nervous system Total Male Female

-1.9 -2.9 -0.9

-2.1 -3.2 -1.3

Total Male Female

3.1* 1.3* 3.9*

3.5* 1.9* 4.3*

2.4* -1.1 3.6*

— — —

0.7 -1.1 1.2

0.9 -2.6 2.0

3.9* 2.5* 4.6*

Total Male Female

3.4* 1.8* 4.0*

3.8* 2.3* 4.5*

3.1* -0.9 4.3*

— — —

0.8 -0.3 1.1

1.1 -2.3 2.1

4.2* 2.9* 4.8*

Total Male Female

0.3 -0.7 1.1

0.8 0.3 1.3

-1.4 — —

— — —

-1.2 — —

-1.8 — —

1.4 1.1 1.4

Total Male Female

0.0 -0.5 0.7

0.0 -0.5 0.7

0.2 -1.1 2.2

0.9 — —

0.6 0.3 0.9

-0.4 -0.8 0.3

-0.1 -0.6 0.7

-0.7 0.9 -2.5

— — —

1.3 — —

1.3 1.7 0.8

0.0 0.1 -0.1

HODGKIN LYMPHOMA

-0.1 0.6

URINARY SYSTEM Total Male Female

Total Male Female

Total Male Female

-9.2* -9.8* — — —

-2.2 -2.3 2.4 — —

-1.4* -1.4* 1.6 — —

0.5* 0.3 0.3

0.9 0.5 1.8

-2.1 -1.5 —

0.6 0.4 1.5

0.3 -0.3 1.4

0.0 0.0 -0.4*

0.3 0.2 0.8

— — —

0.1 0.3 0.4

-0.9 -0.6 -1.7

-3.2* -3.4* 1.8* -5.4* -1.2

0.6* 0.4 0.4 0.2 0.0 -0.1

-0.4 -0.1 -0.6

-0.1 -0.1 -0.1

NON-HODGKIN LYMPHOMA Total Male Female

NHL—nodal

MALE GENITAL SYSTEM Total Prostate Testis Penis Other male genital organs

API

LYMPHOMA -12.8* -15.3* 1.1

Total Cervix uteri Corpus and uterus, NOS Corpus uteri Uterus, NOS Ovary Vagina Vulva Other female genital organs

AI/AN

Other endocrine including thymus

-11.9* -14.4* 0.8

Total Male Female

Black

Kidney and renal pelvis

Hispanic

White NonHispanic

White

Thyroid

0.0 -1.3* 2.7*

Melanoma of the skin Total Male Female

All Races

ENDOCRINE SYSTEM

SKIN EXCLUDING BASAL AND SQUAMOUS Total Male Female

Cancer Site/Type

Total Male Female

0.0 -0.5 0.9*

0.0 -0.5 0.9*

0.4 -1.4 3.2*

1.0 — —

0.5 0.3 0.8

-0.7 -1.2 0.3

-0.1 -0.7 0.8

-0.1 -0.2 0.2

0.0 0.0 0.2

1.0 -0.5 3.1

— — —

-0.4 -1.1 0.3

-1.7* -1.7 -1.5*

0.0 -0.1 0.3

0.2 -1.3 2.5*

0.1 -1.5 2.5*

-0.7 -3.0 3.5

— — —

1.9 2.4 1.3

1.1 -0.4 3.4*

-0.3 -1.8 2.0*

-0.7 -0.8 -0.9

-0.6 -0.4 -1.2

-1.8 -2.1 -1.6

— — —

1.0 -1.6 4.5*

-1.2 -1.4 -1.2

-0.4 -0.2 -1.1

-1.3* -1.5* -1.0

-1.2* -1.6* -0.7

-1.0 -0.9 -0.9

— — —

-0.8 0.4 -1.9

-0.4 -0.1 -0.5

-1.3* -1.7* -1.0

-2.3* -3.0* -1.5

-2.6 -2.4 -2.5

— — —

-1.3 0.1 -3.6

0.1 0.5 0.0

-2.6* -3.2* -2.0

-0.6 0.1 —

— — —

-1.8 0.2 —

2.5 1.4 4.6*

0.7 -1.0 2.7

-3.4 -2.7 -3.5

— — —

— — —

NNL—extranodal Total Male Female

MYELOMA Total Male Female

LEUKEMIA Total Male Female

Lymphocytic leukemia Total Male Female

-2.5* -2.9* -2.0*

Acute lymphocytic leukemia Total Male Female

0.5 -0.8 2.1

1.0 -0.8 3.5*

Chronic lymphocytic leukemia Total Male Female

-3.8* -3.7* -4.3*

-3.8* -3.9* -4.2*

-2.4 1.4 —

-4.0* -4.0* -4.3*

(continued)

169

Appendix 9–A. (cont.) Cancer Site/Type

All Races

Appendix 9–B. (cont.) Black

AI/AN

API

Hispanic

White NonHispanic

— — —

— — —

— — —

— — —

-0.5 -1.6 2.2

0.5 0.6 0.3

1.7 2.4 1.0

— — —

0.7 1.0 0.5

-0.1 0.7 -0.9

0.5 0.7 0.1

1.7* 1.7* 1.5*

4.1* 6.5* 1.8*

— — —

0.4 0.7 0.4

1.8 3.1 0.6

1.4* 1.6* 1.0

— — —

— — —

— — —

— — —

3.0 3.6 3.2

-3.2 -4.1 —

— — —

1.0 1.6 —

-3.7 -4.3 -4.2

-1.8* -1.4 -2.3*

White

Nasopharynx

Other lymphocytic leukemia Total Male Female

-0.4 -1.5 2.4

-0.4 -1.6 2.7

Myeloid and monocytic leukemia Total Male Female

0.5 0.8 0.2

Acute myeloid leukemia Total Male Female

1.6* 1.8* 1.3*

Acute monocytic leukemia Total Male Female

2.8 3.9 1.7

2.9 3.4 3.0

Chronic myeloid leukemia Total Male Female

-1.9* -1.5 -2.5*

-2.0* -1.6 -2.6*

Other myeloid/monocytic leukemia Total Male Female

Other leukemia Total Male Female

-1.3 -2.5 0.7

-1.7 -1.7 —

— — —

— — —

— — —

— — —

-2.2 — —

-4.6* -6.0* -2.8

-4.1* -5.8* -1.9

— — —

— — —

— — —

— — —

-3.6* -5.4 -1.5

-2.9* -3.3* -2.6*

-3.9* -3.1* -4.5*

-2.1 — —

-3.2* -2.4 -3.8*

-2.2 -1.9 -2.5

-2.9* -3.5* -2.4*

MISCELLANEOUS Total Male Female

-3.1** -3.4** -2.9**

Cancer Site/Type

Results are the annual percent changes (APCs), which were calculated using weighted least-squares method. All sites and ovary exclude borderline tumors of the ovary. Data are from SEER 12 areas except Hispanic and white non-Hispanic exclude Alaska, Hawaii, and Detroit. AI/AN, American Indian/Alaska Native; API, Asian or Pacific Islander. —, Statistic could not be calculated. *The APC is significantly different from zero (p < 0.05).

Total Male Female

Tonsil Total Male Female

Oropharynx Total Male Female

Hypopharynx Total Male Female

All Races

White

-2.4* -2.5* -2.4*

-3.1* -2.8* -3.8*

0.1 -1.9 4.1

— — —

-4.5* -5.1* —

-2.3* -2.4* -2.9*

-1.4* -1.5 -2.5*

-5.2* -5.6* —

— — —

-0.5 -0.6 -0.5

-0.2 -0.5 0.2

-1.8 -1.4 -3.1

-4.9* -5.4* -3.8

-4.9* -5.4* -3.9*

Stomach Total Male Female

Small intestine Total Male Female

Colon and rectum Total Male Female

— — —

-1.2* -1.2 -2.4

— — —

— — —

— — —

0.2 -0.3 1.0

-5.6* -6.2* —

— — —

— — —

— — —

-4.9* -5.4* -4.0

-2.2* -2.0* -2.9*

-2.5* -2.3* -2.4

— — —

— — —

-4.0* -6.3* —

-1.9* -1.7 -2.9*

-0.9* -1.0* -1.0*

-0.9* -1.0* -1.0*

-1.1* -1.4* -0.9*

0.8 1.5 0.0

-1.8* -1.9* -1.4*

0.2 0.2 0.2

-0.8* -0.9* -0.9*

0.6* 0.6* 0.1

1.7* 1.7* 0.9*

-4.2* -4.6* -3.2*

2.5 0.9 —

-2.8 -3.2 0.0

0.5 -0.4 3.5

2.0* 2.0* 1.1*

-2.8* -3.2* -2.5*

-3.1* -3.5* -2.8*

-2.3* -2.5* -1.9*

-1.6 -2.3 -0.6

-3.2* -3.2* -3.0*

-1.6* -1.7* -1.6*

-3.2* -3.5* -2.9*

-1.0 -0.4 -1.5*

-1.0 -0.3 -1.6*

-1.2 -1.2 -1.1

— — —

— — —

1.4 — —

-1.0 -0.4 -1.5*

-1.7* -2.0* -1.7*

-1.8* -2.1* -1.8*

-0.6* -0.6* -0.7*

2.2 5.1 -0.2

-2.4* -2.0* -2.6*

0.3 0.3 0.3

-1.8* -2.1* -1.7*

-0.9 -1.4 -1.7

— — —

— — —

— — —

1.8* 2.0 2.0

2.8 1.7 3.3

-0.2 -0.4 0.3

3.1* 2.7* 3.3*

1.9* 2.1* 1.1*

5.9* 3.7 —

-0.4 -0.5 -0.2

2.8* 2.6* 2.5*

1.4* 1.8* -0.1

— — —

1.3 0.4 2.4

4.6* 3.3 5.8*

3.8* 3.2* 4.1*

-2.3* -2.1* -3.0*

Anus, anal canal, anorectum Total Male Female

0.7 0.8 0.8

1.0 1.2 1.2

Liver and intrahepatic bile duct

Appendix 9–B. Short-Term Trends (1992–2000) in Cancer Death Rates Expressed as Annual Percent Change: United States

Total Male Female

2.1* 2.2* 1.4*

2.1* 2.2* 1.4*

1.7* 1.9* 0.4

1.6* 1.9* 0.2

1.4* 1.6 0.8

Liver Cancer Site/Type

All Races

White

ALL MALIGNANT CANCERS Total Male Female

-1.0* -1.4* -0.7*

-0.9* -1.3* -0.6*

ORAL CAVITY AND PHARYNX Total Male Female

Lip Total Male Female

Tongue Total Male Female

Salivary gland Total Male Female

Floor of mouth Total Male Female

170

API

Hispanic

-1.3* -1.8* -0.7*

-0.2 -0.1 -0.4

-1.5* -1.8* -0.9*

-0.1 -0.4 0.0

-0.8* -1.2* -0.5*

-2.7* -3.0* -2.5*

-2.4* -2.7* -2.3*

-4.0* -4.2* -3.1*

-1.0 — —

-3.1* -2.9* -3.3

-3.2* -3.8* -1.7

-2.2* -2.5* -2.2*

-5.9* -6.7* —

-5.8* -6.7* —

— — —

— — —

— — —

— — —

-5.4* -6.2* —

-2.0* -2.5* -1.4*

-1.4* -1.9* -0.8

-5.8* -5.9* -5.6*

— — —

-2.3 — —

-1.8 -1.5 —

-1.1* -1.7* -0.6

-2.0* -2.1* -2.0*

-2.0* -2.1* -2.0

-1.0 -2.7 —

— — —

— — —

— — —

-1.8* -2.1* -1.7

-8.7* -9.0* -8.6*

-8.5* -8.8* -8.6*

-9.0* -9.1* —

— — —

— — —

— — —

-8.3* -8.6* -8.1*

-3.2* -4.1* -2.5*

-5.5* -5.5* -4.5

— — —

— — —

-2.6 — —

-3.1* -3.9* -2.5*

Gums and other mouth Total Male Female

Black AI/AN

White NonHispanic

-3.5* -4.3* -2.7*

Total Male Female

White NonHispanic

— — —

Esophagus Total Male Female

Hispanic

-2.9* -2.4* -3.8*

DIGESTIVE SYSTEM Total Male Female

API

-5.2 — —

Other oral cavity and pharynx Total Male Female

Black AI/AN

1.0 1.4 -0.3

Intrahepatic bile duct Total Male Female

Gallbladder Total Male Female

Other biliary Total Male Female

Pancreas Total Male Female

Retroperitoneum Total Male Female

3.9* 3.2* 4.5*

4.0* 3.3* 4.4*

4.8* 3.1 6.5*

-2.4* -1.7* -2.4*

-2.5* -1.9* -2.5*

-0.8 0.2 -1.1

— — —

-4.4 — -4.8

-1.9 -0.5 -2.2

-2.5* -1.8* -2.5*

-3.2* -3.4* -3.3*

-3.1* -3.3* -3.1*

-2.8* -0.8 -3.7*

— — —

-8.1* -8.2* —

-6.8* -7.0* -6.6

-2.7* -2.9* -2.7*

-0.1 -0.3 -0.1

0.0 -0.1 0.0

-0.9* -1.4* -0.7*

1.1 4.2 -0.6

-0.4 -1.8 1.2

0.1 -0.2 0.2

0.2 0.0 0.2

-2.8* -2.2 -3.1*

-2.8* -2.9 -2.4

-4.0 — —

— — —

— — —

— — —

-2.1 -2.8 -1.3

— — —

— — —

— — —

7.9* -3.0 11.8*

— — —

— — —

— — —

9.5* 9.2* 9.5*

Peritoneum, omentum, mesentery Total Male Female

7.7* -2.2 11.2*

7.8* -2.8 11.5*

6.9* — —

Other digestive organs Total Male Female

9.0* 9.3* 8.6*

9.3* 9.4* 8.9*

7.0 7.3 7.5

Appendix 9–B. (cont.) Cancer Site/Type

All Races

RESPIRATORY SYSTEM Total Male Female

-0.8* -1.9* 0.6*

Appendix 9–B. (cont.) White -0.7* -1.8* 0.7*

Nose, nasal cavity, middle ear Total Male Female

Larynx Total Male Female

Lung and bronchus Total Male Female

Pleura Total Male Female

Black AI/AN

API

Hispanic

White NonHispanic

-1.4* -2.4* 0.6*

-0.1 -0.5 0.7

-1.2* -1.6* -0.1

-0.4 -0.9* 0.5

-0.5* -1.6* 0.9*

-2.5* -2.2 -3.7*

-2.1* -1.5 -3.6*

-5.3* -7.4* —

— — —

— — —

— — —

-1.7 -1.3 -2.9*

-2.0* -2.4* -1.4*

-1.8* -2.3* -1.1

-2.5* -2.5* -2.5*

— — —

-3.2 — —

-1.0 -0.9 —

-1.7* -2.2* -0.8

-0.8* -1.8* 0.7*

-0.6* -1.7* 0.7*

-1.3* -2.4* 0.8*

-0.2 -0.6 0.7

-1.2* -1.6* -0.1

-0.3 -0.8* 0.6

-0.5* -1.6* 0.9*

-5.9* -5.6* -7.4*

-5.8* -5.5* -7.8*

— — —

— — —

— — —

— — —

-5.5* -5.1* -7.5*

Trachea, mediastinum, other respiratory Total Male Female

-4.7* -5.3* -3.9*

BONES AND JOINTS Total Male Female

-1.0* -1.0* -1.2*

-4.1* -5.1* -2.9

-8.5* -5.2 —

— — —

— — —

— — —

-4.4* -5.4* -3.3

-0.7* -0.8 -0.6

-4.0 -3.6 -5.5

— — —

— — —

1.1 -0.9 3.5

-0.8* -0.9 -0.9

— — —

-2.6 -2.6 -1.9

-0.9 -0.8 -1.0

-0.6 -0.4 -1.0

SOFT TISSUE INCLUDING HEART Total Male Female

-0.9 -0.5 -1.4

-0.6 -0.4 -1.0

-2.4 0.6 -4.0*

SKIN EXCLUDING BASAL AND SQUAMOUS Total Male Female

-0.4* -0.3 -0.6

Total Male Female

-3.8* -3.8* -3.4

— — —

-1.3 -0.9 —

0.0 0.6 -0.8

0.1 0.2 -0.1

0.0 0.3 -0.3

0.1 1.3 -0.5

— — —

-2.7 — —

0.7 3.1 -2.4

0.3 0.5 -0.1

-1.3 -1.4 -1.0

-0.8 -1.0 -0.4

-6.3* -5.8* -7.2*

— — —

— — —

-1.2 -2.7 —

-0.6 -0.7 -0.5

-2.5* 0.4 -2.4*

-2.7* 0.3 -2.6*

-1.0* 1.3 -1.1*

-0.6 — -0.6

-1.1 — -1.1

-1.2 — -1.2

-2.6* 0.3 -2.5*

-0.1 0.2 -0.5

Other nonepithelial skin Total Male Female

BREAST Total Male Female

FEMALE GENITAL SYSTEM Total Cervix uteri Corpus and uterus, NOS Corpus uteri Uterus, NOS Ovary Vagina Vulva Other female genital organs Total Prostate Testis Penis Other male genital organs

-1.9* -4.8* -0.1

-5.4* -0.6 -6.6* -2.3* — 1.4

-1.0 -2.5* 0.2

-0.7* -2.3* -0.3

-1.4* 0.8 -0.8* -0.9 -0.4 3.6

-1.5* 0.8 -0.7* -0.6 -0.1 3.6

-1.0 0.8 -1.2* -2.7 -3.9 3.6

— 2.2 — 0.5 -5.3* -0.8 — — — — — —

-0.7 1.0 -0.9 — 1.5 —

-1.3* 0.9* -0.5 -0.8 0.1 4.0

-3.5* -3.5* -1.1 -0.3 3.5

-2.1* -2.1* — -6.7* —

-3.6* -4.3* -3.8* -4.3* — — — — — —

-2.2* -2.2* -4.7 — —

-3.4* -3.5* -0.8 -0.6 3.2

-0.2 -0.3* -0.4

-0.7* -0.8* -0.7

-0.4 0.1 -2.0

0.2 0.3 -0.3

0.0 -0.2 -0.2

URINARY SYSTEM Total Male Female

-0.3* -0.4* -0.5

-0.4 0.2 -1.1

White

-0.3* -0.6* -0.5

-0.2 -0.4* -0.4

-1.5* -1.8* -1.1

— — —

-2.8* -1.6 -4.2

-0.2 -0.1 -1.0

0.0 -0.3 -0.2

-0.3 -0.2 -0.5

-0.2 -0.3 -0.5

0.2 0.4 -0.2

-0.6 -0.3 —

1.6 2.0 1.2

0.5 0.7 0.1

-0.2 -0.2 -0.4

-1.2 -1.9* -0.5

-1.0 -1.6 -0.5

— — —

— — —

— — —

— — —

-0.2 -0.9 0.6

5.7 9.3* 2.5

7.3* 11.0* 3.6

-3.8 — -3.2

— — —

— — —

— — —

7.8* 11.7* 3.6

-2.9* -3.1* -2.8

-3.0* -3.3* -2.9

— — —

— — —

— — —

— — —

-2.9* -3.4* -2.6

Kidney and renal pelvis Total Male Female

Ureter Total Male Female

Other urinary organs Total Male Female

EYE AND ORBIT Total Male Female

Black AI/AN

BRAIN AND OTHER NERVOUS SYSTEM Total Male Female

-0.7* -0.6* -0.8*

Total Male Female

Total Male Female Total Male Female

-0.7 1.2 -2.6

1.0 1.1 0.9

-0.5* -0.4* -0.6*

0.0 0.9 -0.6

-0.1 1.0 -0.8

0.8 0.3 1.2

— — —

-1.6 -0.4 -2.3

0.7 4.0 -0.8

0.0 0.9 -0.6

0.4 2.4* -0.4

0.2 2.3* -0.7

2.2 3.3 1.7

— — —

-2.2 — -2.7

0.3 — -1.1

0.3 2.3* -0.6

— — —

2.0 — —

-0.5 -0.7 -0.5

-0.7 -1.6 0.7

0.5 -0.1 1.3

0.0 0.1 -0.1

-0.6 -0.6 -0.8

-0.7 -0.5 -1.0

-0.9 -2.0 0.4

— — —

-0.2 -0.2 -0.2

-0.1 -0.1 -0.1

-0.5 -0.6 -0.3

5.9* 7.0* 5.4*

-3.6* -3.9* -3.0*

-5.0* -5.8* -4.2*

— — —

— — —

-0.9 -1.6 0.1

-3.3* -3.8* -2.7*

-3.7* -4.2* -3.0*

NON-HODGKIN LYMPHOMA Total Male Female Total Male Female

LEUKEMIA Total Male Female

White NonHispanic

2.7 — —

HODGKIN LYMPHOMA Total Male Female

Hispanic

-0.3 -0.6 -0.3

Thyroid Total Male Female

API

-0.6* -0.5* -0.7*

ENDOCRINE SYSTEM

MYELOMA

-0.8* -2.4* -0.4

-3.4* -3.4* -1.0 -1.1 3.7

Total Male Female

LYMPHOMA

-1.0* -2.9* -0.3

MALE GENITAL SYSTEM

Urinary bladder

All Races

Other endocrine including thymus

-0.2 -0.1 -0.3

Melanoma of the skin

Cancer Site/Type

0.0 0.1 0.0

0.1 0.1 0.0

-0.1 -0.1 0.0

5.6* 6.3* 5.0

-0.5 -1.0 0.3

0.6 0.0 1.5

0.2 0.4 0.1

-0.3 -0.5 -0.1

-0.3 -0.3 -0.3

-0.2 -0.8 0.3

1.3 — —

0.6 -0.9 2.3

1.1 -0.5 2.1

-0.2 -0.2 -0.2

-0.5* -0.7* -0.5*

-0.4* -0.6* -0.4

-1.0* -0.9 -0.8

-1.7 -2.2 -1.4

-1.2 -1.3 -0.9

0.8* 1.0 0.5

-0.3* -0.5* -0.2

0.2 1.0 -0.5

1.8 2.0 1.6

1.0 1.0 0.9

MISCELLANEOUS MALIGNANT CANCER Total Male Female

0.7 0.7 0.6

0.9 0.9 0.8

-0.5 -0.4 -0.4

0.3 1.4 -0.6

Underlying mortality data provided by NCHS (www.cdc.gov/nchs). Rates are per 100,000 and age-adjusted to the 2000 U.S. (19 age groups) standard. AI/AN, American Indian/Alaska Native; API, Asian or Pacific Islander. Based on total United States except Hispanic and white non-Hispanic exclude CT, NH, LA, OK. —, Statistic could not be calculated. *The APC is significantly different from zero (p < 0.05).

171

Appendix 9–C. 5-Year Relative Survival Rates and Stage Distribution for Selected Sites: SEER 9 Areas, 1992–1999 5-Year Relative Survival Rate (%) Cancer Site/Type

oral cavity and pharynx Lip Tongue Salivary gland Floor of mouth Gum and other mouth Nasopharynx Tonsil Oropharynx Hypopharynx Other oral cavity and pharynx

digestive system Esophagus Stomach Small intestine Colon and rectum Colon excluding rectum Cecum Appendix Ascending colon Hepatic flexure Transverse colon Splenic flexure Descending colon Sigmoid colon Large intestine, NOS Rectum and rectosigmoid junction Rectosigmoid junction Rectum Anus, anal canal, anorectum Liver and intrahepatic bile duct Liver Intrahepatic bile duct Gallbladder Other biliary Pancreas Retroperitoneum Peritoneum, omentum, mesentery Other digestive organs

respiratory system Nose, nasal cavity, middle ear Larynx Lung and bronchus Pleura Trachea, mediastinum, other respiratory organs

bones and joints soft tissue including heart skin excluding basal and squamous Melanoma of the skin Other nonepithelial skin

breast (female) female genital system Cervix uteri Corpus and uterus, NOS Corpus uteri Uterus, NOS Ovary Vagina Vulva Other female genital organs

male genital system Prostate Testis Penis Other male genital organs

urinary system Urinary bladder Kidney and renal pelvis

172

Total No.

Total Localized

Regional

Stage Distribution (%)

Distant Unstaged

Localized

Regional

Distant

Unstaged

17,845 1,916 3,900 1,889 1,505 2,897 1,266 1,986 477 1,484 525

57.2 94.4 53.1 74.7 51.7 54.8 57.2 54.0 37.3 30.9 32.3

82.1 98.0 72.1 94.9 71.2 82.4 84.2 66.0 58.3 44.4 59.7

47.9 80.7 45.9 59.3 39.3 45.9 57.2 55.9 38.5 33.1 32.8

26.1 0.0 26.4 30.5 22.1 27.2 38.1 34.6 12.7 12.7 9.0

42.8 86.9 41.1 52.8 49.2 32.0 49.9 47.1 23.2 30.9 20.6

34.1 78.3 37.6 46.6 38.7 33.7 10.8 12.6 13.8 9.8 15.4

48.2 13.2 44.1 38.5 50.3 42.5 69.6 69.9 62.9 69.5 59.0

9.2 0.7 12.1 7.6 4.3 6.4 12.4 12.4 13.4 16.0 12.4

8.5 7.9 6.3 7.3 6.7 17.4 7.2 5.1 9.9 4.7 13.1

139,265 7,012 13,540 2,364 83,374 59,530 13,943 669 8,894 3,240 5,078 2,238 3,599 19,061 2,808 23,844 7,685 16,159 1,870 7,772 6,462 1,310 1,946 2,167 17,116 716 826 562

43.8 14.0 22.5 52.6 62.3 62.3 60.9 59.6 63.7 62.1 61.7 58.8 63.5 67.6 27.2 62.4 62.1 62.5 64.4 6.9 7.5 3.9 15.1 18.9 4.4 48.2 31.0 7.7

77.8 29.1 59.0 74.0 90.1 91.6 93.3 90.7 90.4 90.1 90.4 90.9 85.9 92.6 86.7 87.1 89.8 86.0 82.1 16.3 17.1 10.4 43.9 36.2 16.6 68.6 64.2 0.0

49.3 13.1 21.7 60.1 65.5 67.9 67.3 58.6 69.1 65.0 66.9 65.9 70.0 69.6 57.3 59.2 61.9 57.5 54.8 6.0 6.7 2.6 13.3 22.2 6.8 52.7 47.0 9.9

6.2 2.2 2.5 28.7 9.2 9.4 10.9 23.3 9.0 6.6 6.8 9.4 8.5 10.5 3.0 8.5 9.7 7.6 15.3 1.9 2.1 1.1 0.8 2.1 1.6 24.4 21.2 4.0

17.1 11.6 13.2 27.5 35.5 31.4 32.8 0.0 26.2 21.2 36.3 30.5 34.6 40.2 26.0 41.9 41.4 41.9 53.3 2.5 2.3 3.5 2.6 8.2 4.1 29.4 28.4 14.3

30.4 25.4 22.5 27.9 37.8 35.6 32.4 39.3 34.6 33.2 32.3 29.5 38.3 43.3 10.7 43.2 37.8 45.8 48.2 27.0 28.4 20.0 21.1 18.1 7.6 30.7 9.8 0.4

34.0 29.5 31.9 36.0 37.6 38.7 41.0 26.0 43.6 45.6 45.4 44.5 39.6 35.0 12.9 34.9 40.4 32.3 31.8 24.5 25.4 20.2 39.5 44.7 24.0 34.1 18.6 6.6

25.6 25.8 32.1 28.3 19.3 21.0 23.4 32.0 17.8 17.5 19.1 22.9 18.1 18.6 42.5 15.2 18.0 13.8 8.1 22.2 21.4 25.8 33.8 15.9 51.3 25.8 64.0 66.0

9.9 19.2 13.5 7.9 5.3 4.8 3.2 2.7 4.0 3.7 3.2 3.1 3.9 3.0 33.9 6.7 3.8 8.1 11.9 26.3 24.8 34.0 5.5 21.3 17.1 9.4 7.5 27.0

110,228 1,123 6,746 100,433 1,525 401

18.4 55.9 64.7 14.9 7.1 44.2

55.0 80.8 82.6 48.7 18.2 65.2

18.8 49.6 47.9 16.0 8.7 39.4

2.4 30.7 20.0 2.1 3.2 32.8

10.4 55.1 54.0 8.3 7.3 34.8

17.8 26.6 50.5 15.6 13.0 26.2

36.0 48.3 41.3 35.7 20.4 33.7

36.1 13.3 3.8 38.3 52.3 19.2

10.1 11.8 4.4 10.4 14.2 20.9

1,581

69.6

85.1

68.6

32.3

64.2

39.8

35.6

14.3

10.3

4,571

67.6

84.4

58.2

22.8

57.9

55.0

21.5

14.1

9.5

32,360 25,746 6,614

80.2 89.6 45.5

96.6 96.7 95.2

66.0 60.1 86.8

14.9 13.8 32.2

36.8 79.9 26.6

69.6 81.8 22.1

10.1 10.1 10.5

3.0 3.5 0.9

17.3 4.6 66.5

116,587

86.6

97.0

78.7

23.3

56.0

63.2

28.7

5.7

2.5

49,308 8,628 22,437 22,055 382 12,719 482 1,937 625

71.4 71.3 84.4 85.3 29.4 43.7 46.1 76.3 64.9

94.6 92.2 96.2 96.3 71.6 91.7 56.1 89.1 88.9

59.6 50.9 64.7 65.5 27.6 68.2 44.6 54.7 72.0

29.4 16.5 26.0 27.1 12.3 27.7 33.1 18.9 45.2

46.5 56.5 53.9 56.6 30.9 25.2 43.5 61.5 59.1

54.8 54.4 72.7 73.7 13.6 19.7 34.6 63.8 27.5

15.6 31.5 14.6 14.5 21.5 6.4 28.2 25.7 18.6

24.0 7.6 8.2 7.7 36.4 67.1 21.2 3.4 41.3

5.5 6.5 4.4 4.0 28.5 6.7 16.0 7.1 12.6

138,235 132,087 5,359 557 232

97.2 97.5 95.5 74.9 78.5

100.0 100.0 99.1 90.1 88.7

100.0 100.0 95.0 62.0 72.1

37.0 34.0 73.1 0.0 0.0

87.6 88.1 90.4 49.8 0.0

68.5 68.5 69.4 57.5 66.8

17.3 17.2 18.5 26.2 19.0

6.1 5.9 10.6 3.4 4.7

8.2 8.4 1.6 12.9 9.5

48,855 31,273 16,561

74.4 81.8 62.6

92.9 94.4 89.9

53.1 48.2 60.0

8.5 5.8 9.1

47.0 59.4 29.4

65.1 74.0 50.2

20.2 18.7 21.7

9.7 3.2 21.8

5.1 4.1 6.2

Appendix 9–C. 5-Year Relative Survival Rates and Stage Distribution for Selected Sites: SEER 9 Areas, 1992–1999 5-Year Relative Survival Rate (%) Cancer Site/Type

Total No.

Ureter Other urinary organs

eye and orbit endocrine system Thyroid Other endocrine including thymus

lymphoma Hodgkin lymphoma Non-Hodgkin lymphoma NHL—nodal NHL—extranodal

Total Localized

Regional

Stage Distribution (%)

Distant Unstaged

Localized

Regional

Distant

Unstaged

625 396

55.6 63.7

75.3 84.9

55.6 50.7

5.9 22.5

37.9 69.5

34.9 27.3

45.1 26.8

10.7 9.6

9.3 36.4

1,350

82.0

85.0

61.8

65.0

76.4

76.9

4.9

2.5

15.7

12,582 11,462 1,120

92.6 95.8 60.2

98.5 99.3 77.0

93.3 95.5 70.3

50.9 59.9 35.7

79.1 85.1 52.8

53.3 55.7 29.0

35.5 35.8 33.2

7.8 5.5 30.7

3.3 3.0 7.1

36,482 5,368 31,114 21,032 10,082

60.7 84.1 56.1 54.1 60.4

71.5 90.8 68.3 73.4 63.7

71.3 88.6 61.6 61.1 62.9

48.1 73.9 44.5 44.9 42.6

64.0 78.2 63.0 51.6 74.1

31.1 26.5 31.8 22.3 51.7

16.8 37.7 13.2 14.2 11.0

43.3 31.9 45.3 56.3 22.3

8.9 3.8 9.7 7.2 15.0

Excludes borderline tumors of the ovary.

Appendix 9–D. 5-Year Relative Survival Rates by Sex for Selected Cancer Sites/Types and All Cancers Combined 5-Year Relative Survival Rate (%) Cancer Site/Type All cancers combined Brain and other nervous system Brain Cranial nerves, other nervous system Myeloma Leukemia Lymphocytic leukemia Acute lymphocytic leukemia Chronic lymphocytic leukemia Other lymphocytic leukemia Myeloid and monocytic leukemia Acute myeloid leukemia Acute monocytic leukemia Chronic myeloid leukemia Other myeloid/monocytic leukemia Other leukemia Miscellaneous

Sex Distribution (%)

No. of Cases

Total

Male

Female

Male

Female

765,470 11,484 10,808 676 8,949 19,891 9,641 2,602 6,206 833 8,983 5,526 361 2,847 249 1,267 17,642

62.9 32.8 30.2 73.6 31.5 46.3 70.6 63.5 73.5 80.0 23.8 18.7 19.2 34.9 23.1 15.1 13.8

62.6 33.1 30.5 76.0 33.5 47.1 70.3 61.6 72.9 83.5 23.4 17.5 21.5 35.0 23.8 15.9 15.8

63.2 32.5 29.8 71.1 29.4 45.3 71.1 66.1 74.5 70.6 24.2 20.0 15.7 34.9 22.3 14.2 12.0

52.5 56.2 56.5 51.5 52.5 57.3 59.7 57.6 58.9 72.0 55.2 54.0 58.2 57.1 54.6 54.1 47.7

47.5 43.8 43.5 48.5 47.5 42.7 40.3 42.4 41.1 28.0 44.8 46.0 41.8 42.9 45.4 45.9 52.3

Selected sites are sites for which information on historic stage does not exist for all or part of the site. All sites combined excludes borderline tumors of the ovary.

173

10

Socioeconomic Disparities in Cancer Incidence and Mortality ICHIRO KAWACHI AND CANDYCE KROENKE

T

he association between socioeconomic status (SES) and health status is so robust and consistent that epidemiologists routinely adjust for it as a potential confounding variable when evaluating the etiologic role of other risk factors for disease. This chapter turns this logic on its head, focusing on SES as a fundamental determinant of disease, specifically cancer incidence and mortality. The association between SES and health status has been recorded throughout history. William Farr, regarded as the intellectual founder of epidemiology, documented the existence of socioeconomic disparities in mortality in nineteenth century Britain and concluded that: “No variation in the health of the states of Europe is the result of chance; it is the direct result of the physical and political conditions in which nations live” (quoted in Beaglehole and Bonita, 1997, p. 93). The apparently persistent and pervasive association between lower socioeconomic position and worse health status accounts for the belief that SES may be a “fundamental” determinant of individual and population health; that is, no matter what the current health threats confronted in any given society, the disadvantaged groups are worse off in terms of their achieved health (Link and Phelan, 1995). However, as this chapter argues, there are important exceptions to these generalizations. For example, the direction of causation does not uniformly run from SES to health. Evidence suggests that socioeconomic attainment and health status exert reciprocal influences on each other. Second, the relation between SES and health is dynamic and may change (or even reverse) over time. Major risk factors for cancer, such as cigarette smoking, physical inactivity, and obesity, were each more prevalent among higher SES groups in industrialized societies during early periods of economic development, but the patterns reversed during more recent history so the same risk factors now exhibit higher prevalence among disadvantaged groups (Eckersley et al., 2001). Finally, when describing the relations between SES and cancer outcomes, it is important to be specific about the exact site and type of cancer, as well as the class of outcome measure (incidence versus survival). The incidence of some cancers, notably breast cancer and melanoma, is higher among more advantaged SES groups, presumably reflecting the underlying socioeconomic distribution of their risk factors. For breast cancer, the increased incidence among higher SES women is most likely explained by reproductive factors, including earlier age at menarche, later age at first birth, and lower fertility. On the other hand, survival following the diagnosis of breast cancer consistently favors higher SES women because, among other things, they are privy to earlier detection and better access to effective treatment (Lochner and Kawachi, 2000). The present chapter is organized into four sections. The first section defines the concept of SES and describes the various approaches to its measurement. The second section summarizes observations on the general nature of the association between SES and cancer morbidity, mortality, and survival. The third section outlines the general categories of explanations, both causal and noncausal, that have been put forward to account for the association between SES and cancer. The fourth and final section provides a survey of the specific causal mechanisms underlying the relation between SES and cancer. This section is organized into two subsections dealing, respectively, with early life and adult risk factors that potentially account for the observed disparities in cancer incidence.

174

WHAT IS “SES”? DEFINITIONS AND MEASUREMENT Sociologists and allied social scientists have studied social stratification and its consequences for more than 150 years (Grusky, 1994). Epidemiologists who study the effects of social stratification on health have, for the most part, followed the conceptual approach established by Max Weber (1864–1920) (Liberatos et al., 1988). According to Weber, socioeconomic status is defined as groups within society that are relatively homogeneous with respect to life chances and opportunities (Grusky, 1994). In turn, life opportunities (and hence SES groups) have been measured by socioeconomic indicators such as educational attainment (years of schooling, educational credentials gained), income and wealth, as well as occupational status or prestige (Lynch and Kaplan, 2000). SES is therefore a multidimensional concept that is intended to capture the manifold, and often subtle, differences between individuals with respect to their access to life opportunities. Education, occupation, and income are linked across the life course, so more schooling determines an individual’s ability to access higher status jobs, which in turn determines the rate of monetary compensation, and so on. It is not surprising, then, that the three indicators of SES are usually found to be correlated with one another. That said, epidemiologists often tend to use the individual indicators of SES in an interchangeable manner, even though they may in fact be linked to health outcomes through different mechanisms. For example, schooling and education is believed to influence health through the acquisition of knowledge in one or more forms (e.g., increased knowledge about health-promoting behaviors, enhanced ability to process and act on that information, and the ability to interact with and “navigate” medical care systems and options, i.e., boosting “health literacy”) (Yen and Moss, 1999). Income, on the other hand, enables individuals to purchase the various goods and services necessary for maintaining and promoting health. Occupational status often determines an individual’s exposure to hazards (including carcinogens) in the workplace (Pearce and Matos, 1994) as well as their access to psychosocial resources such as prestige, respect, and security. The example of cigarette smoking illustrates the potential pitfalls of the uncritical use of SES measures. Higher educational attainment is strongly inversely associated with cigarette smoking in most advanced industrialized societies, including the United States (Pamuk et al., 1998). The inverse association between smoking and education is thought to be mediated by knowledge about the hazards of cigarette smoking. In support of this hypothesis, an association between education and smoking among young adults did not exist prior to 1953, when the health hazards of smoking had yet to be established (Fuchs, 1983). Educational differentials in smoking emerged soon after the initial surgeon general’s report on smoking and health. Higher incomes are also correlated with a lower prevalence of smoking, although less strongly than is the case with education (Pamuk et al., 1998). However, the mechanism of the association with income is less obvious. Increased income should result in the enhanced ability to afford more goods, including cigarettes, all other things being equal. Historically, during the economic development of societies, as real personal income rises so does per capita cigarette consumption. Within certain subgroups in the population, such as teenagers, higher incomes (in the form of more pocket money) have been shown

Socioeconomic Disparities in Cancer Incidence and Mortality to be associated with more cigarette consumption (Scragg et al., 2002). Why, then, is higher income associated with lower cigarette consumption? The reason is unlikely to be due to a direct causal effect of income on smoking. Rather, higher incomes are most likely associated with other variables that reduce the probability of smoking. For example, teenagers who receive more pocket money are also more likely to come from higher socioeconomic backgrounds, so the effect of higher parental education (and hence stronger family socialization) may cancel out the effect of more pocket money on cigarette consumption. The point is that education and income are likely to be related to cigarette smoking through different mechanisms. Some care is therefore warranted when specifying the indicator of SES used for describing its association with health outcomes. For the remainder of the chapter, the term “socioeconomic disparities” is used as a broad descriptive term to refer to variations, or inequalities, in cancer outcomes by SES. The specific indicator of SES used is described whenever appropriate. Three additional points are worth noting with respect to measuring SES. First, researchers now increasingly recognize that SES can be conceptualized and measured at both the individual level and the area level (e.g., the neighborhoods in which individuals reside). An individual with a given level of income or educational attainment could experience different chances of health depending on the average SES level of her neighborhood (e.g., as measured by the median household income of a census tract). Capturing area-level SES is therefore of intrinsic interest over and above measuring individual SES. This is a different argument from using area SES as a proxy for individual SES when the latter measures are not available on data sets. The reason area SES is believed to contribute independently to health outcomes is based on the differential quality of neighborhood environments according to the average SES level of its residents, a phenomenon referred to as residential segregation (Kawachi and Berkman, 2003). Examples of such neighborhood-level patterning include differential access to services and amenities (e.g., mammography screening services, supermarkets that sell fresh fruits and vegetables), the physical environment (e.g., air pollution, the presence of parks and recreational playgrounds for physical activity, differential exposures to outdoor tobacco advertising), and the social context (e.g., social support among neighbors) (Kawachi and Berkman, 2003). In turn, the differential distribution of these protective and risk factors at the neighborhood level is thought to contribute to the disease risks of residents, including cancer risk. Second, in addition to the dimension of place, researchers have increasingly emphasized the importance of the dimension of time for conceptualizing and measuring SES and its effects on health. SES seldom remains static across the life course. Individuals exhibit upward (or downward) social mobility—when they quit jobs to return for more schooling, when they are involuntarily laid off, or when they are promoted to higher positions, among other reasons. Accordingly, measuring SES at any single point in time is unlikely to capture the dynamic as well as the cumulative effects of SES on health. Income dynamics, in the form of accumulated spells of poverty, have been shown to predict mortality and other health outcomes (McDonough et al., 1997). Childhood socioeconomic circumstances have been shown to predict health outcomes in later life, independent of the SES attained during adulthood (Gliksman et al., 1995; Davey-Smith et al., 2001). The third and final point worth mentioning about measuring socioeconomic disparities in health is in regard to its distinction from racial disparities in health. This point is particularly pertinent in the United States, where official statistics frequently conflate racial disparities in health with socioeconomic disparities (Williams, 1997). Race and SES are not synonymous or interchangeable. Race is emphatically not a proxy for SES, despite the fact that racial minorities in the United States are overrepresented among lower SES groups stemming from discrimination and the denial of opportunities throughout history. The finding that race is often associated with morbidity and mortality independent of SES emphasizes that race is more than SES. Accordingly, understanding the sources of racial disparities in health requires attention to a separate set of causes and mechanisms, including the poten-

175

tial influence of racial discrimination on health outcomes (Krieger, 2000).

SES GRADIENT IN CANCER INCIDENCE AND MORTALITY SES “Gradient” Individuals from low SES backgrounds—whether measured by educational attainment, income and wealth, or occupational status—generally experience worse health outcomes than those higher in the socioeconomic hierarchy (Adler et al., 1993). This widely observed pattern of the rise in health status with each level of SES has been referred to as the SES “gradient” in health. It is a “gradient” because there is no apparent threshold or cut-point in the relation between SES and health; that is, the excess morbidity and mortality risks of disadvantaged groups are not solely confined to those who are poor by officially defined criteria. Instead, at each level of the SES hierarchy, people experience better health than those immediately below them, even among groups considered middle class and above.

Socioeconomic Status and Cancer Socioeconomic disadvantage, whether measured by low income, low educational attainment, or low occupational status, has been linked with both higher overall cancer incidence and mortality (Lochner and Kawachi, 2000; Bradley et al., 2001; Coleman et al., 2001). However, in contrast to the SES gradient reported for other major health outcomes, such as cardiovascular disease or infectious diseases, the SES gradient across cancer incidence and mortality is often not as strong (Steenland et al., 2002). Two major reasons for the weaker associations between SES and overall cancer incidence and mortality are (1) the heterogeneity of associations between SES and specific cancer sites; and (2) the length of the induction period between exposure to low SES conditions and cancer onset. With regard to the first point, cancer (unlike cardiovascular disease) involves a highly heterogeneous mix of diseases caused by different sets of risk factors, many of which may be unrelated, or related in opposite directions, to SES. Thus, an association between overall cancer incidence and/or mortality with SES at any given point in time reflects the weighted contribution of the component sites, each of which may have a different relation with SES. For example, as the contribution of lung cancer deaths outgrows the contribution of breast cancer in overall cancer mortality rates for U.S. women, the SES gradient in total cancer deaths can be expected to grow stronger (Singh et al., 2002a). The analysis of SES gradients in overall cancer incidence and mortality is therefore of limited value in understanding etiologic relations, though it may serve a broader purpose when monitoring population trends in regard to health disparities. With regard to the second point, most cancers (unlike cardiovascular diseases) are associated with lengthy induction periods between exposure (in this case to the living conditions associated with low SES) and the onset of disease. Induction periods for cancers are typically of the order of decades, whereas they may be just a few years for heart disease, infectious diseases, or other health outcomes that exhibit strong SES gradients. The length of the induction period for most cancers produces measurement error in exposure status, as individuals move out of poverty, experience upward job mobility, and so forth, with the result that the association between SES and cancer outcomes is biased in the direction of the null. Alternatively, it may take decades for a socioeconomic gradient in cancer mortality (e.g., lung cancer) to emerge, even after SES gradients in health behaviors (e.g., smoking) have become well established. Such was the case for the educational gradient for lung cancer mortality among women in the United States (Steenland et al., 2002).

SES and Cancer Incidence With regard to incidence, the cancer sites that exhibit the strongest and most consistent associations with low SES are lung, stomach, and

176

PART II: THE MAGNITUDE OF CANCER

cervix (Lochner and Kawachi, 2000). Cancer sites showing a probable association with low SES include oral, esophageal, laryngeal, liver, and bladder cancer. Breast cancer and melanoma are two notable malignancies for which incidence rates appear to be higher among higher-SES individuals (Faggiano et al., 1997). There is a dearth of cancer incidence studies in the United States with available data on SES at the individual level. One reason for this is the paucity of socioeconomic information gathered by official sources of data, including cancer registries. Krieger and colleagues (1997) conducted a survey of all cancer registries in the United States to ascertain the availability of SES information. Of the 45 state cancer registries that responded to their mailed survey, 36 (80%) collected some information on occupational status, but only 4 routinely reported data broken down by occupation, reflecting the constraints posed by the accuracy and reliability of socioeconomic data available from patients’ medical records (the principal source of cancer registry data). Only two of the state cancer registries collected information on education, and none of the registries collected information on income. In the absence of individual-level SES information, some U.S. researchers have sought to examine area-level SES disparities in cancer incidence (Devesa and Diamond, 1980; Liu et al., 1998; Krieger et al., 1999). Krieger and colleagues (1999) examined the associations between cancer incidence and SES measured at the census block group level in California’s San Francisco Bay area. A census block group is a division of a census tract, typically contain-

ing 1000 residents, with boundaries drawn to maximize social homogeneity. All block groups were then categorized into three SES levels: “professional” areas, with a high concentration of employed persons who are in executive, professional, or supervisory occupations; “working class nonpoor” areas, with a high concentration of employed persons who are in nonsupervisory roles (e.g., clerical workers, laborers) but who are not below the official poverty threshold; and “working class poor” areas, with a high concentration of nonsupervisory workers and 20% or more of residents living below the poverty line. The associations between this area SES measure and the five causes of incident cancer (breast, cervix, colon, lung, prostate) examined in this study are shown in Table 10–1 (Krieger et al., 1999). The results of this study, which were also stratified by race/ ethnicity, indicated a strong area-SES gradient for lung cancer and cervical cancer, which parallels the strong individual-level associations that have been reported between SES and these cancer sites (Baquet et al., 1991; Kogevinas et al., 1997b). A notable exception in the case of lung cancer incidence was the apparent positive association between higher area SES and excess risk in Hispanic women and men (Table 10–1). The authors speculated that this Hispanic “paradox” may reflect the persistently higher smoking prevalence among professional classes in immigrant populations from Latin America (Krieger et al., 1999). Breast cancer incidence increased with area affluence only among Hispanic women (although there was a marginally statistically significant trend in the same direction among Asian American women).

Table 10–1. Age-Adjusted Invasive Cancer Incidence Rates (Per 100,000 Person-Years) and Relative Risks by Cancer Site, Gender, Race/Ethnicity, and Block Group Socioeconomic Position: San Francisco Bay Area, 1988–1992 Relative Risksa

Incidence Rate Cancer Site

Gender and Race

5-Year Case Count

Women Asian/PIb 1,328 Black 1,113 Hispanic 1,134 White 12,545 Cervix uteri Women Asian/PI 220 Black 143 Hispanic 193 White 712 Colon Women Asian/PI 394 Black 424 Hispanic 287 White 3,454 Men Asian/PI 427 Black 350 Hispanic 306 White 3,185 Lung Women Asian/PI 444 Black 561 Hispanic 364 White 4,976 Men Asian/PI 825 Black 986 Hispanic 521 White 5,924 Prostate Men Asian/PI 835 Black 1,342 Hispanic 931 White 10,738

Total

Professionals

Working Class, Nonpoor

Working Class, Poor

Working Class, Nonpoor

Working Class, Poor

p Trend

62.6 96.3 76.8 119.9

64 96.5 98.8 119.5

63.5 99 62.5 119.3

50.5 95.1 49.9 139.4

1.0 1.0 0.6 1.0

0.8 1.0 0.5 1.2

0.07 0.89 0.00 0.12

10.2 11 10.5 6.7

9.7 9.2 7.9 5.9

10.6 12.2 9.8 8

13.9 12.5 20.9 25.3

1.1 1.3 1.2 1.4

1.4 1.4 2.7 4.2

0.16 0.07 0.00 0.00

20.7 36.4 21.1 28.2

20.6 33.9 25.9 27.3

18.4 34.6 18.5 30.8

26.7 41.6 14.6 36.3

0.9 1.0 0.7 1.1

1.3 1.2 0.5 1.3

0.21 0.06 0 0

27.6 42.1 34.2 38.6

29 39.7 43.7 37.3

24.3 51.1 28.1 42.4

31.8 39.4 24.8 49.6

0.8 1.4 0.6 1.1

1.1 1.0 0.6 1.3

0.95 0.74 0 0

23.2 50.2 27.2 46.9

23.3 39.9 32.8 42.6

22.4 55.7 23.9 58.9

25.5 59.7 19.8 87

1.0 1.3 0.7 1.3

1.1 1.4 0.6 1.9

0.67 0 0 0

53.9 113 55.4 72

50.9 104.1 63.2 62.6

54.6 112 49.6 99.6

67.1 131.4 50.8 140.1

1.1 1.1 0.8 1.6

1.3 1.3 0.8 2.2

0.01 0 0.04 0

56.9 164.7 112.5 130.8

60 171.8 147.7 131.9

58.3 182.7 90.2 126.3

41.4 145.8 75.6 131.8

1.0 1.0 0.6 1.0

0.7 0.8 0.5 1.0

0 0.02 0 0.24

Breast

Source: Krieger et al. (1999). a Reference category: professionals. b Asian/PI, Asian and Pacific Islander.

Socioeconomic Disparities in Cancer Incidence and Mortality

177

Table 10–2. Lung Cancer Mortality Rates and Ratios by Educational Attainment: United States, 1959–1972 and 1982–1996 Education Level

Ratea

RRb (95% CI)

RRc (95% CI)

Education Gradientd (95% CI)

cancer prevention study i: men, 1959–1972, aged 45 years Grammar school Some high school High school graduate Some college College graduate

134.8 146.4 127.5 120.1 85.2

1.69 (1.54, 1.85) 1.85 (1.68, 2.04) 1.51 (1.36, 1.68) 1.45 (1.31, 1.61) 1.00

1.41 (1.29, 1.56) 1.49 (1.36, 1.64) 1.30 (1.17, 1.44) 1.25 (1.13, 1.39) 1.00

1.054 (1.044, 1.064)

cancer prevention study i: women, 1959–1972, aged 45 years Grammar school Some high school High school graduate Some college College graduate

18.9 20.9 23.0 24.2 18.9

1.08 (0.87, 1.33) 1.14 (0.92, 1.39) 1.20 (0.98, 1.47) 1.33 (1.09, 1.62) 1.00

1.27 (1.02, 1.58) 1.21 (0.98, 1.48) 1.22 (1.0, 1.5) 1.32 (1.08, 1.61) 1.00

0.997 (0.971, 1.023)

cancer prevention study ii: men, 1982–1996, aged 45 years Grammar school Some high school High school graduate Some college College graduate Graduate school

262.6 257.7 181.0 161.2 120.4 83.3

2.99 (2.76, 3.24) 3.00 (2.77, 3.24) 2.17 (2.02, 2.32) 1.94 (1.81, 2.09) 1.45 (1.34, 1.57) 1.00

1.95 (1.8, 2.12) 1.94 (1.79, 2.1) 1.61 (1.5, 1.73) 1.47 (1.37, 1.59) 1.25 (1.16, 1.36) 1.00

1.101 (1.094, 1.108)

cancer prevention study ii: women, 1982–1996, aged 45 years Grammar school Some high school High school graduate Some college College graduate Graduate school

84.7 94.1 71.7 73.3 61.1 50.5

1.34 (1.17, 1.53) 1.76 (1.57, 1.97) 1.39 (1.26, 1.54) 1.43 (1.29, 1.59) 1.20 (1.07, 1.35) 1.00

1.39 (1.21, 1.59) 1.52 (1.36, 1.71) 1.29 (1.17, 1.43) 1.24 (1.12, 1.38) 1.12 (1.0, 1.26) 1.00

1.035 (1.026, 1.046)

Source: Steenland et al. (2002). a Directly standardized rates per 100,000: 5-year age intervals. b RR, rate ratio; CI, confidence interval. Age-adjusted rate ratio from Cox regression (1-year age strata). c Multivariate-adjusted rate ratio from Cox regression adjusted for smoking, body mass index, diet, alcohol, prevalent hypertension, and menopausal status (women). d Rate ratio for each year less of education calculated via use of education as a continuous variable, age-adjusted only (5year intervals), using Poisson regression.

Although there appeared to be a slightly higher incidence (rate ratio 1.2) among white women residing in “working class poor” areas, this observation was based on extremely small numbers: only 3% of white women in the catchment area lived in poor block groups. The associations between area SES and the incidences of colon and prostate cancer were complex. Once again, both types of cancer were less common among Hispanics in more disadvantaged areas. Some studies in Europe, Canada, Asia, and New Zealand have documented higher incidences of colon cancer and prostate cancer among more affluent groups (Kogevinas et al., 1997b). However the area SES pattern for colon cancer ran in the opposite direction—with poorer areas exhibiting higher incidence rates—among white men and women (Table 10–1). In summary, the patterns of SES gradients can be quite complex and can vary according to the setting (i.e., region or country of the world), the time period, and the population. These patterns “defy easy generalization” (Krieger et al., 1999) and warrant careful attention to methodologic issues such as the level at which SES is being measured (area level or individual level) and stratifying the associations by race/ethnic group.

SES and Cancer Mortality Several large cohort studies have demonstrated an association between SES—whether measured at the individual or area level—and mortality rates from cancer (Bucher and Ragland, 1995; Sorlie et al., 1995; Smith et al., 1996; Howard et al., 2000). Among the largest such cohort studies to be conducted were the two American Cancer Society (ACS) cohorts: Cancer Prevention Study I (CPS-I) and Cancer Prevention Study II (CPS-II). Steenland and colleagues (2002) examined the association between educational attainment and cancer mortality rates in

the two cohorts. The CPS-I cohort comprised 1,051,038 men and women enlisted by ACS volunteers in 1959, whose vital status was followed until 1972. The CPS-II cohort consisted of 1,184,657 men and women enlisted by ACS volunteers in 1982 and followed until 1996. Death rates and rate ratios in the two cohorts, by level of educational attainment, are reproduced for two of the four cancer sites examined: lung cancer (Table 10–2) and breast cancer (Table 10–3). The lung cancer data (Table 10–2) show that the mortality rates increased between the two study periods for both men and women. Among men, there was a significant educational gradient during both periods, but the gradient grew stronger during the second period (CPSII). By contrast, among women there was no statistically significant educational gradient during the 1960s (CPS-I): only women with the highest educational attainment (college graduates) had a lower mortality rate from lung cancer compared to all other women. However, during the second period (1982–1996, CPS-II), a statistically significant educational gradient emerged. Adjusting the rate ratios for smoking, diet, and alcohol intake (among other things) attenuated but did not remove the excess risks among lower educational groups, suggesting that these factors alone did not explain the educational gradient. Age-adjusted breast cancer mortality rates were approximately 20% higher among the most educated groups for both time periods (Table 10–3). The CPS-II analyses included about 2.9% women who had prevalent breast cancer at the beginning of follow-up. A notable interaction was revealed when the analyses were repeated after stratifying by prevalent disease status at baseline. Among women without prevalent breast cancer at baseline, there was little evidence of lower mortality risk with less education, except for a weak effect in the lowest educational group: age-adjusted rate ratio of 0.89 with the 95% confidence interval (CI) 0.77–1.07. Among women with prevalent disease

178

PART II: THE MAGNITUDE OF CANCER Table 10–3. Breast Cancer Mortality Rates and Rate Ratios by Educational Attainmenta: United States, 1959–1972 and 1982–1996 Education Level

Rateb

RRc (95% CI)

RRd (95% CI)

Education Gradiente (95% CI)

cancer prevention study i: women, 1959–1972, aged ≥45 years Grammar school Some high school High school graduate Some college College graduate

71.3 76.1 81.1 88.4 88.7

0.81 (0.73, 0.89) 0.86 (0.78, 0.95) 0.93 (0.85, 1.03) 1.01 (0.92, 1.11) 1.00

0.91 (0.82, 1.01) 0.95 (0.86, 1.05) 0.99 (0.90, 1.09) 1.06 (0.97, 1.17) 1.00

0.977 (0.968, 0.986)

cancer prevention study ii: women, 1982–1996, aged ≥45 years Grammar school Some high school High school graduate Some college College graduate Graduate school

74.7 79.5 81.2 83.1 86.4 91.3

0.79 (0.7, 0.89) 0.84 (0.76, 0.93) 0.88 (0.81, 0.95) 0.90 (0.83, 0.98) 0.94 (0.86, 1.03) 1.00

0.84 (0.75, 0.96) 0.90 (0.80, 1.00) 0.93 (0.86, 1.00) 0.94 (0.87, 1.03) 0.98 (0.90, 1.07) 1.00

0.982 (0.975, 0.989)

Source: Steenland et al. (2002). a The education effect differed between those with and without prevalent disease at baseline, indicating effect Modification (see text). b Directly standardized rates per 100,000: 5-year age intervals. c Age-adjusted rate ratios from Cox regression (1-year age strata). d Multivariate-adjusted rate ratios from Cox regression adjusted for smoking, body mass index, diet, alcohol, prevalent hypertension, menopausal status (women), parity, and age at first birth. e Rate ratio for each year less of education calculated via use of education as a continuous variable, age-adjusted only (5year intervals), using Poisson regression.

at baseline, those with the least education had an increased risk of mortality (age-adjusted rate ratio 1.22, 95% CI 1.07–1.47), suggesting that the educational gradient in mortality masked an association between lower educational attainment and worse survival from breast cancer (Steenland et al., 2002). In the CPS cohorts, Steenland and colleagues (2002) also found no consistent educational gradient in colorectal cancer mortality for the earlier period, but 20%–30% lower mortality rates for the most educated men and women in the later cohort (CPS-II). A weak inverse gradient was also found for prostate cancer death rates (lower education, higher death rates), which was stronger among the 0.6% percent of men with prevalent disease at baseline. Although the ACS cohorts provide intriguing data for the 1960s, 1970s, and 1980s, some caution is warranted when interpreting the findings, as the CPS cohort members were not representative of the general American population. No reliable data are available on the longer-term trends regarding the association between individual SES and cancer mortality rates in the United States. The reason, as explained earlier, is because socioeconomic data have not been routinely collected or reported in official vital statistics or cancer registries within the United States (Krieger et al., 1997). In the absence of individual SES information, some investigators have sought to document the long-term trends regarding the relation between area-level SES measures and cancer mortality rates (Singh et al., 2002b). Here it is important to reiterate that area-level SES is not being used as a proxy for individual SES—doing so would result in a potential ecologic fallacy. Singh and colleagues (2002b) developed an area SES index at the level of the 3097 counties of the United States to track the evolution of the SES gradient in cancer mortality between 1950 and 1998. The SES index was composed of 11 variables available through the U.S. Census, including aggregate indicators that assessed the domains of education (percent of the county population with less than 9 years of education; percent with at least a high school education), occupation (percent of employed persons in white collar occupations; the unemployment rate), and income/wealth (median county-level family income; county-level income inequality; median home value; median gross rent; percent of families below the poverty level; percent occupying housing units without telephone access; and percent occupying housing units without complete plumbing). The authors found a dramatic change in the area SES pattern of cancer mortality during the 40-year study period (Figs. 10–1 and 10–2).

Throughout the 1950s and 1960s, there was a positive SES gradient (i.e., higher cancer mortality rates in areas of high SES than in areas of lower SES), which was true for both men and women. For example, during 1950–1952, cancer mortality was 49% (95% CI 41%–59%) higher in the highest SES areas than in the lowest SES areas. The positive SES gradient narrowed during the 1970s for men, and by the late 1980s the gradient began to reverse and then widen. By 1997–1998, male cancer mortality rates were 19% (95% CI 11%–28%) higher in the lowest SES areas compared to the highest SES areas. Among women, an interaction was found with age group. For women over 65 years of age, cancer mortality rates were higher in the high SES areas throughout the study period, although the gradient narrowed over time. In younger women (25–64 years of age), the SES gradient was reversed during the early 1990s. By 1998, the cancer mortality rate among younger women was 13% (95% CI 9%–16%) higher in the lowest SES areas than in the highest SES areas (Singh et al., 2002b). Similar reversals of the SES gradient in cancer mortality have been reported in Britain, Canada, and Australia (Singh et al., 2002a, 2002b). It is worth cautioning here that the SES–cancer relations at the individual level may look different over time; we currently lack data in the United States.

SES and Cancer Survival Even though lower SES is not universally associated with increased risks of cancer incidence across sites, lower SES is nonetheless consistently related to worse prognosis and survival following the diagnosis of cancer. Even for cancers with a positive SES gradient, people of lower SES have a greater risk of mortality following the diagnosis of cancer. The positive SES gradient for breast cancer incidence and the reversal of this gradient for breast cancer mortality has been frequently documented in the literature. Although women of higher SES have higher breast cancer rates, mortality rates decrease and survival times improve with increasing SES (Dayal et al., 1982; Bassett and Krieger, 1986; Gordon et al., 1992).

EXPLANATIONS FOR THE ASSOCIATION BETWEEN SES AND CANCER OUTCOMES Having described the general patterns of association between various indicators of SES and cancer outcomes, we turn now

Socioeconomic Disparities in Cancer Incidence and Mortality

179

Figure 10–1. Cancer mortality rates for U.S. men by age and socioeconomic status (SES) index, 1950–1998. (Source: Singh et al., 2002a.)

to consider the general categories of explanation for the observed gradients. Generally speaking, there are three explanations for the repeatedly observed correlation between SES and cancer. First, the association may arise through reverse causation; that is, the diagnosis of cancer may result in loss of employment, loss of income, or both. Alternatively, low socioeconomic position may be causally linked to cancer incidence and/or survival through mechanisms described in detail below. Lastly, the correlation between SES and cancer may be the spurious artifact of a third unobserved variable that affects both SES and cancer risk (e.g., inherited personality characteristics that cause people to strive harder and succeed at upward social mobility and at the same time invest in their personal future health including steps to prevent cancer). These three explanations are not necessarily mutually exclusive.

The reverse causation hypothesis is sometimes referred to as the “drift” hypothesis, meaning that the onset of illness (e.g., cancer) can act as a potent trigger for individuals to drift downward on the socioeconomic hierarchy. Conversely, healthy people are more likely to succeed in moving up the socioeconomic hierarchy. The implication is that health disparities are due to selective social mobility, which sorts healthy and unhealthy people into different socioeconomic positions (Wilkinson, 1996). So-called birth cohorts, in which individuals are followed from birth with repeated assessments of health status and SES, are ideally suited to test for these effects. The analysis of such birth cohorts, particularly in Britain, has documented that health status can (and does) affect later social mobility (Fox et al., 1985; Wadsworth, 1986; Power et al., 1990). However, the general magnitude of the drift effect is too small to explain away the observed SES gradient in health. That said, almost all of the studies of the

Figure 10–2. Cancer mortality rates for U.S. women by age and socioeconomic status (SES) index, 1950–1998. (Source: Singh et al., 2002a.)

180

PART II: THE MAGNITUDE OF CANCER

drift phenomenon have been carried out in studies of SES differentials in all-cause mortality. Whether reverse causation can account for some of the SES differentials in cancer outcomes remains an open issue. It is possible for investigators to minimize the influence of reverse causation by carefully selecting the indicator of SES. For example, educational attainment is less susceptible to drift effects than are other indicators of SES such as income or occupational status. The reason is because educational attainment is usually completed during early adulthood before individuals succumb to most forms of cancer. However, even educational attainment might be plausibly subject to reverse causation (i.e., effects running from health status to educational attainment). For example, the 1958 British National Child Development Study suggested that “health shocks” in the form of low birth weight resulted in a persistent deleterious effect on O-level examination performance (the O-levels are national examinations taken by British students at about age 16, the results of which determine whether one can proceed with further schooling) (Currie and Hyson, 1999). Low birth weight might thus predict both lower educational attainment and reduced risks of prostate and breast cancer (DaveySmith et al., 2001). According to this scenario, it is not low educational attainment that causes low risks of breast and prostate cancer; rather, low birth weight may be the factor underlying both. The argument that the correlation between SES and health may be due to (unobserved) third variable bias is often made by economists and frequently ignored by epidemiologists (Farrell and Fuchs, 1982; Fuchs, 1983). The argument is based on observations such as the timing of the onset of educational differentials in cigarette smoking. In a cohort of young persons with 12–18 years of completed schooling, Farrell and Fuchs (1982) found that the strong negative relation between schooling and smoking observed at age 24 was almost completely accounted for by differences in smoking behavior observed at age 17 when all subjects were still in the same grade. In other words, the amount of formal schooling persons eventually achieve predicts their smoking behavior before their schooling is completed. Achieving the additional years of schooling beyond age 17 had no additional effect on smoking behavior. It is possible that schooling is effective in preventing smoking uptake only up to the 12th grade and not thereafter. However, the authors rejected this unlikely explanation in favor of underlying “third variables” related to both schooling and health maintenance. The problem of omitted variable bias can be generalized to the entirety of empirical evidence linking SES to diverse health outcomes, including cancer. In other words, the challenge of research in this field is to move beyond documenting associations between SES and health and going on to test causal connections. The key issue is that the same factors that prompt people to seek more education and better jobs—such as higher abilities, higher family socioeconomic background, more patience and less impulsiveness (i.e., what economists refer to as time discount, or the willingness to incur current costs for future benefits)—are also likely to be determinants of health behaviors and future health. In an attempt to overcome the problem of unobserved heterogeneity and omitted variable bias, economists frequently resort to the use of instrumental variable analysis to identify the causal relation of SES and health outcomes (Greenland, 2000). These methods are not without their own problems, however, and the most convincing approach (short of experimentation) to establish the causal link between SES and health remains to measure and control for the array of potential “third variables” that are related to both socioeconomic position and health. Even after a causal relation between SES and cancer has been established, many unresolved questions remain about the exact mechanisms by which the link occurs. Broadly speaking, there are two pathways by which SES may improve cancer outcomes. One direct pathway is through the ability of higher SES individuals to gain access to various resources that help prevent cancer or improve outcomes following the onset of cancer. For example, more years of schooling may result in greater ability to translate health knowledge into health-promoting behavior (avoiding smoking, maintaining regular physical activity, presenting for regular screening checkups). Health literacy may also

enable individuals to “navigate” medical treatment options and maximize the effectiveness of cancer therapy. In the words of Grossman (1975), additional years of schooling make an individual a more efficient producer of their own health. Income may also operate directly to reduce an individual’s cancer risk by determining his or her level of access to material resources (e.g., the ability to afford nicotine replacement therapy or to afford a health plan that offers regular screening services for cancer detection). Higher incomes may also enable individuals to move to higher quality residential neighborhoods with access to safe recreational spaces (for maintaining exercise) or supermarkets and groceries with ready availability of fresh produce. Finally, the labor market tends to sort individuals into work places and jobs with differential exposures to carcinogens (e.g., chemical exposures, indoor tobacco smoke pollution). An alternative pathway from SES to health (including, potentially, cancer outcomes) is through differential exposure to psychosocial mediators, such as stress, a sense of control, social support, and so on (Ross and Wu, 1995). In turn, psychosocial factors such as stress and social support have been shown to influence health outcomes through health-related behaviors (Seeman and Seeman 1983) and through direct effects on physiologic mechanisms, such as altered immune and neuroendocrine function (McEwen, 1998). In the final section, we turn to survey and summarize the known risk (and protective) factors for cancer incidence and mortality thought to be shaped by the SES. We can hence offer potential clues about the mediating causal mechanisms linking the two.

MECHANISMS LINKING SOCIOECONOMIC STATUS TO CANCER Cigarette smoking is a risk factor par excellence that is linked to both low SES (in industrialized societies such as the United States) and cancer outcome. The relation between low SES and smoking-related cancers is by now well established (Kogevinas et al., 1997b; Lochner and Kawachi, 2000). However, as the previous section implied, socioeconomic position is also associated with a host of other risk (and protective) factors beyond cigarette smoking. Moreover, it is problematic from the point of view of prevention that many risk factors tend to cluster, or “crystallize,” around low SES living conditions. As emphasized at the outset of the chapter, SES is a multidimensional construct, with complex pathways linking it to cancer outcomes. Accordingly, reducing or eliminating socioeconomic disparities in cancer is unlikely to be accomplished by focusing on and acting to remove one risk factor at a time. The remainder of this section summarizes the known risk factors for cancer that also exhibit SES patterning. The section is subdivided into early-life risk factors and adult risk factors.

Early-Life Exposures Prenatal and Early Postnatal Exposures Relatively little is known about the effects of prenatal exposures linking SES to childhood cancer (Ross and Swensen, 2000). Some work has been done to explore the link between certain occupational exposures or lifestyle exposures (e.g., cigarette smoking or diet) in relation to specific childhood cancers (John et al., 1991; PrestonMartin et al., 1996; Holly et al., 1998). Current evidence is limited and equivocal. However, there is substantial evidence on several viruses that may be transmitted vertically from mother to child in utero or during childbirth or breast-feeding, including human lymphotropic virus-type I (HTLV-1) and hepatitis B virus (HBV), that are both patterned by SES and can lead to cancer during adulthood. These viruses are endemic in some countries but are far rarer in the United States. Hepatitis C virus (HCV) is more prevalent in the United States than HTLV-1 or HBV. Intravenous drug use, which is more prevalent among people of lower SES, accounts for 60% of HCV infections, a major cause of liver cancer. However, the risk of perinatal transmission of hepatitis C is relatively low.

Socioeconomic Disparities in Cancer Incidence and Mortality There has also been some exploration of birth weight and future cancer risk. Trichopoulos hypothesized that prenatal exposure to high concentrations of pregnancy estrogens related to fetal growth rate may influence the subsequent risk of breast cancer (Trichopoulos, 1990). Studies have found positive, or J-shaped, associations between birth weight/birth length/head circumference and breast cancer (Vatten et al., 2002; McCormack et al., 2003). Because low SES is associated with a high prevalence of low birth weight (Parker et al., 1994), this may be one mechanism through which high SES is related to an elevated risk of breast cancer. Fewer studies have examined birth size and other cancer outcomes. Sandhu and colleagues found a J-shaped association of birth weight and colorectal cancer (Sandhu et al., 2002), consistent with findings on birth weight and adult obesity (Leong et al., 2003). However, in a retrospective analysis of birth weight and prostate cancer in the Health Professionals Follow-up Study, Platz and colleagues found no association between birth weight and prostate cancer (Platz et al., 1998).

Postnatal and Later Exposures Major mechanisms for postnatal exposures include childhood infection, lifestyle behaviors, and second-hand exposure to tobacco smoke.

Childhood and Adolescent Infection

181

SES backgrounds (Goodman, 1999; Wang, 2001), and the SES gradient in overweight has been increasing (Moore et al., 2002). Maternal obesity, also predominant among adults from lower SES backgrounds, is strongly linked with childhood obesity (Strauss and Knight, 1999). Childhood overweight and obesity in turn may track into adulthood and are major causes of cancer in later life. Dietary habits, which begin during childhood, are predictive of future dietary patterns. Lack of local availability of fresh produce, cultural patterns of food consumption (Lee and Cubbin, 2002), weight norms (Becker et al., 1999; Crawford et al., 2001), and sedentary behaviors are each related to lower SES and overweight. In addition to habits that develop in the context of a lack of resources, current overweight predicts future overweight. Girls who are overweight have earlier maturational timing, which is associated with later overweight (Adair and Gordon-Larsen, 2001). Adolescents who are obese are far more likely to be obese as adults, with long-term, multiple consequences for morbidity (Dietz, 1998a, 1998b; Micic, 2001). Patterns established during childhood and adolescence therefore predispose people from lower SES backgrounds to worse health outcomes, including cancer, during later adulthood.

Smoking and Secondhand Exposure to Cigarette Smoke

There is substantial evidence that infections are likely causal factors of several cancers, including lymphoma and cancers of the liver, nasopharynx, cervix, and stomach, altogether accounting for up to 20% of cancers worldwide (Eckhart, 1998). Viruses linked to human tumors include Epstein-Barr virus (EBV) (B-cell lymphomas, Burkitt’s lymphoma, nasopharyngeal cancer, some Hodgkin’s disease and T-cell lymphomas, gastric cancer); HBV (hepatocellular carcinoma); papillomavirus types 16, 18, 31, 33, 35, 39, 45, 52, 56, and 58 plus a few others (cervical and anogenital cancer); and HTLV-1 (adult T-cell leukemia). The transmission and timing of infection have been linked to SES. The EBV is involved in about 35%–50% of the cases of Hodgkin’s disease. Most adults have had an EBV infection and are thus carriers of these viral genes (Evans and Mueller, 1997). However, first infection with EBV at a later age produces a much more severe clinical illness. Therefore infectious mononucleosis is largely a disease of individuals in upper socioeconomic groups who have escaped early infection. In the United States, Hodgkin’s disease in the younger age groups is largely a disease of high SES groups (Gutensohn, 1982). Recent findings strongly indicate that chronic Helicobacter pylori infection is involved in the development of gastric adenocarcinoma and some cases of gastric lymphoma (International Agency for Research on Cancer Ad Hoc Working Group, 1994; Nightingale and Gruber, 1994). About two-thirds of the world’s population and half of all adults over age 50 in the United States are infected with H. pylori. Since the 1930s, modernization leading to clean water, fewer children sharing a bed, smaller families, and possibly the increasing use of antibiotics in children (Blaser, 1999) have contributed to reductions in H. pylori transmission and prevalence (Parsonnet, 1995; Parsonnet et al., 1999). The infection is typically acquired during childhood, and there are fewer children in the United States today with H. pylori infection than during previous periods. H. pylori is found more often in low socioeconomic groups in the United States. Prior epidemiologic studies have identified several risk factors for H. pylori carriage, including a history of crowding during childhood, a mother who carries H pylori, a large number of siblings, the presence of older siblings (<4-year age difference), and unclean water sources. In turn, several of these factors are associated with lower SES (Goodman and Correa, 1995).

Children from lower SES backgrounds are more likely to be exposed to environmental tobacco smoke (ETS) in their homes (Emmons et al., 2001; Schuster et al., 2002). Although a meta-analysis found no increased association of parental smoking with lung cancer in the grown child (Boffetta et al., 2000), ETS has been associated with an increased risk of lung cancer in spouses (adult nonsmoking women) (Taylor et al., 2001; Kreuzer et al., 2002). Lower SES youths are subsequently more prone to take up smoking than are youths from more affluent backgrounds (Escobedo et al., 1990; Harrell et al., 1998). Factors that may be related to the higher initiation of smoking among low SES youths include parental smoking (especially for white adolescents) (Griesler and Kandel, 1998; Tyas and Pederson, 1998), greater exposure to peer norms of smoking (Tyas and Pederson, 1998; Alexander et al., 2001), and more intense targeting by the tobacco industry, a strong predictor of adolescent onset of smoking (Altman et al., 1996; Pierce et al., 1998). Lower SES youths are also more likely to start smoking at an earlier age, which has been linked to greater difficulty quitting later (Breslau and Peterson, 1996). Furthermore, it has been hypothesized that smoking during adolescence, a “critical period” in lung development, may be particularly hazardous, in that tobacco carcinogens may induce genetic alterations that make the early smoker more susceptible to the damaging effects of continued smoking (Wiencke and Kelsey, 2002).

Lifestyle Behaviors

Smoking continues to be among the most important factors responsible for the SES gradient in cancer. Cigarette smoking is the leading determinant of lung cancer: 90% of those with lung cancer were smokers. Reducing or eliminating smoking would prevent the bulk of lung cancer cases as well as drastically reduce socioeconomic disparities in lung cancer incidence. Unfortunately, in response to declining cigarette consumption in the United States (particularly among high

The U.S. Centers for Disease Control and Prevention (CDC) reported that the number of overweight children has more than doubled during the last three decades. Currently, 5.3 million, or 12.5%, of American children between the ages of 6 and 17 years are overweight or obese (Centers for Disease Control and Prevention, 1997). Overweight and obesity are more prevalent among children and adolescents from lower

Adult Exposures Lifestyle Major lifestyle factors in adult life linked to cancer incidence include smoking, poor diet, alcohol consumption, physical inactivity, and overweight or obesity. Many of these factors are associated with each other and tend to cluster together. Several health-related behaviors (e.g., smoking, heavy alcohol consumption, overeating) are used by individuals to cope with stress, a fact that has not escaped the notice of manufacturers and advertisers, who target vulnerable audiences. In combination with physical inactivity, poor diet can lead to overweight and obesity. The changing temporal socioeconomic patterns in U.S. lung and colorectal cancer mortality rates may be attributable in part to the changing socioeconomic patterns of these factors.

Smoking

182

PART II: THE MAGNITUDE OF CANCER

SES groups), the tobacco industry has substantially increased expenditures on marketing and advertising, particularly for tobacco promotional items (Federal Trade Commission, 1999). Tobacco advertising has been shown to be disproportionately targeted to people from lower SES backgrounds as well as youths (Parker-Pope, 2001).

Diet High SES is positively correlated with higher quality diet including greater consumption of fruits and vegetables (Centers for Disease Control and Prevention, 2000) and lower consumption of saturated fat (Erkkila et al., 1999). Factors such as fat intake, red meat consumption, inadequate vegetable consumption, high caloric intake, physical inactivity, and heavy alcohol consumption have been suggested to be important risk factors for colorectal cancer. Each of these factors has been related to lower SES, both at the individual level and the area (neighborhood) level (Diez-Roux et al., 1999; Kamimoto et al., 1999; Centers for Disease Control and Prevention, 2000; Li et al., 2000). However, the apparent correlation between low SES and an unhealthy diet has not always been direct. A study of dietary trends by Popkin and colleagues found large differences in dietary quality in 1965, with high SES whites eating the least healthful and low SES blacks eating the most healthful diet as measured by a diet quality index. By the 1989–1991 survey, the diets of all groups studied were relatively similar. Of particular note, grain and legume consumption declined among low SES blacks since the 1960s but increased among high SES whites (Popkin et al., 1996). Little or no research has examined the reasons underlying changing differentials in dietary trends by socioeconomic status. It is possible that changing patterns in diet may reflect a greater receptivity of high SES people to public health messages about eating healthy (Shi, 1998). However, the evidence (Popkin et al., 1996) suggests that people from both low and high SES backgrounds showed improvements in diet; that high SES groups initially had more improvements to make; and that diet quality was similar across SES groups by the early 1990s following several recommendations, including reducing fat (£30%), saturated fat (<10%), and cholesterol (<300 mg/day) intake; increasing consumption of fruits and vegetables (five or more servings per day) and complex starches (six or more servings per day); limiting sodium intake (£2400 mg/day); and maintaining proper calcium [recommended daily allowance (RDA)] and protein (more than twice the RDA) intake (Popkin et al., 1996). Nevertheless, the rise in obesity prevalence among lower SES groups has undoubtedly been influenced by increased consumption of low quality foods and overall calorie consumption (Guthrie and Morton, 2000; Jahns et al., 2001; Zizza et al., 2001; Nielsen et al., 2002; Nielsen and Popkin, 2003), even while improvements have been made toward meeting dietary recommendations (Heini and Weinsier, 1997; Harnack et al., 2000; Siega-Riz and Popkin, 2001). Trends toward increased consumption may be due to increases in availability and relative affordability of food items (Putnam and Allshouse, 1999; Blisard and Harris, 2001; Fried and Nestle, 2002), increased marketing of low quality food items (Taras and Gage, 1995; Gamble and Cotugna, 1999; Borzekowski and Robinson, 2001; Chestnutt and Ashraf, 2002; Tirodkar and Jain, 2003), and increased portion sizes (Nielsen and Popkin, 2003).

Alcohol Consumption Although wealthier people are more likely to engage in moderate alcohol consumption (Centers for Disease Control and Prevention, 2001), which has been linked to lower levels of insulin, body weight, and risk of diabetes (Colditz et al., 1991; Facchini et al., 1994; Conigrave et al., 2001; Davies et al., 2002; Kroenke et al., 2003, unpublished observations), heavy alcohol consumption is more predominant among lower SES groups (Centers for Disease Control and Prevention, 2001). Considerable evidence suggests a connection between heavy alcohol consumption and increased risk for cancer, with an estimated 2%–4% of all cancers thought to be directly or indirectly caused by alcohol (Rothman, 1980). Alcohol consumption is an established cause of cancers of the mouth, pharynx, larynx, esophagus,

liver, and breast. For each of these cancers, risk increases substantially with intake of more than two drinks per day, though regular consumption of even a few drinks per week has been associated with an increased risk of breast cancer in women (Hamajima et al., 2002; American Cancer Society, 2003). Other research suggests that other lifestyle factors associated with low SES in combination with heavy alcohol consumption, such as malnutrition (Su and Arab, 2001; Giskes et al., 2002) or smoking (American Cancer Society, 2003), may interact to further increase the risk of certain cancers. Alcohol consumption combined with low folate consumption appears to predict high levels of colon cancer; and combined with tobacco use, it increases the risk of cancers of the mouth, larynx, and esophagus more than the independent effect of either drinking or smoking.

Obesity and Physical Activity Overweight and obesity are associated with several cancers, including postmenopausal breast cancer, colon cancer, endometrial cancer, prostate cancer, renal cell carcinoma, and esophageal adenocarcinoma (Bianchini et al., 2002; Bray, 2002; Morimoto et al., 2002). In addition to changing the diet, the disproportionate prevalence of obesity among people from lower SES backgrounds may help account for the changing SES gradient for colorectal cancer. Obesity has increased dramatically over the past several decades (Kuczmarski et al., 1994; Flegal et al., 2002); and lower SES groups, especially women, are much more likely to be overweight than people from high SES backgrounds (Robbins et al., 2000). As the socioeconomic disparities in overweight widens, the SES gradient in colorectal cancer incidence and mortality may similarly increase (Steenland et al., 2002). High SES is positively correlated with recreational physical activity, which in turn has been associated with lower levels of obesity-related cancers. As an independent risk factor, the evidence for decreased risk with increased physical activity is classified as convincing for breast and colon cancers, probable for prostate cancer, and possible for lung and endometrial cancers (Friedenreich and Orenstein, 2002). The hypothesized biologic mechanisms for the association between physical activity and cancer include changes in endogenous sex hormone levels and growth factors, decreased obesity and central adiposity, and possibly changes in immune function. Central adiposity has been implicated in metabolic conditions that promote carcinogenesis. Evidence is also increasing that exercise influences other aspects of the cancer process, including cancer detection, coping, rehabilitation, and survival (Friedenreich and Orenstein, 2002). Based on existing evidence, the American Cancer Society has issued physical activity guidelines for cancer prevention (Byers et al., 2002). It generally recommends at least 30 minutes of moderate to vigorous intensity physical activity on 5 days or more per week.

Sexual Behavior Aspects of sexual behavior and sexually transmitted diseases (STDs) have been linked to reproductive cancers such as cervical and prostate cancers (Castellsague et al., 2002; Dennis and Dawson, 2002). Limited data are available for assessing rates of STDs by SES (Santelli et al., 2000). STD infection rates are higher among racial and ethnic minority groups, which is often attributed to poverty (Hofferth, 1987; Ellen et al., 1995). Cervical cancer is more common among women from lower SES backgrounds (Giuliano et al., 1999), who are less likely to undergo regular screening and early treatment. Other important factors related to the development of cervical cancer include a woman’s history of sexual exposure, her partners’ history of sexual exposure, use of barrier methods of contraception (protective), and the age at which a woman becomes sexually active. In particular, the risk of cervical cancer is increased when a woman is exposed to the virus before full maturation of the cervix. Several of these factors, in turn, have been shown to be related to SES, especially level of education. In a study by Hogan and colleagues, adolescent females whose parents were better educated (more than high school education) were 28% less likely to initiate sexual intercourse and 52% more likely to

Socioeconomic Disparities in Cancer Incidence and Mortality use a contraceptive at first intercourse (Hogan et al., 2000). Similar findings in other studies also suggest that females with lower income or lower parental educational attainment are more likely to be at an early age at first intercourse (Blum et al., 2000; Lammers et al., 2000; Santelli et al., 2000; Singh et al., 2001), subsequently have more sexual partners (Coker et al., 1994; O’Donnell et al., 2001), and be less likely to use barrier contraception (Bankole et al., 1999). Finally, women who smoke cigarettes are nearly 1.5 times more likely to develop cervical cancer (Sood, 1991). Chemicals in cigarette smoke may increase the risk by damaging cervical cells.

Reproductive Factors Reproductive factors associated with SES and with cancer incidence include age at menarche, parity, age at first birth, age at menopause, and use of exogenous hormones. High SES women have been shown consistently (though not always) to suffer higher rates of breast cancer. They are more likely to have first children at a much later age and to have fewer children. They have been more likely to take exogenous hormones after menopause. These factors have in turn been associated with an increased breast cancer risk. Early age at menarche is also associated with higher risk of breast cancer. High SES is associated with earlier age at menarche, which has been observed both within countries and in comparisons of developed versus less developed countries. On the other hand, with the obesity epidemic growing disproportionately among low SES children, early menarche may increasingly weigh against lower SES women in the United States, leading to a narrowing of the SES gap in breast cancer incidence in the future.

Occupational Exposures Blue collar and manual workers experience greater exposure than white collar workers to chemicals, diesel fumes, dyes, and other agents in the workplace. Such agents include inorganic gases, organic compounds, solvents, mineral and wood dusts, silica, metals, and bioaerosols, which are established carcinogens (Boffetta et al., 1997; Kauppinen et al., 1997; Kogevinas et al., 1997a; Weston et al., 2000; U.S. Department of Health and Human Services, 2002). Blue collar workers are also more likely to be exposed to environmental tobacco smoke (Curtin et al., 1998). Percent of white-collar workers was a predictor of more restrictive smoking policies in a national sample of worksites (Emmons et al., 2000). Furthermore, the interactive effects of smoking and chemical exposures may put workers of low SES at additional increased risk of cancer. The combination of asbestos exposure and cigarette smoking, both correlated with low SES, has been associated with a synergistically increased risk of lung cancer (Gustavsson et al., 2002). Although not all shift workers occupy lower SES positions (e.g., health care professionals), shift work is more common among persons from lower SES backgrounds (e.g., jobs in the service and transport industries, security work, manufacturing) (Boggild et al., 1999). Longer durations (>20 years) of a rotating night shift has been linked to an increased incidence of cardiovascular disease after controlling for other health behaviors and risk factors (Kawachi et al., 1995). More recently, a rotating night shift has also been linked to increased risks of breast and colon cancer (Schernhammer et al., 2001; E.S. Schernhammer et al., 2003, unpublished observations). In particular, working 30 years or more on the night shift was associated with a moderate elevation in the risk of breast cancer [relative risk (RR) 1.36; 95% confidence interval (CI) 1.04–1.78], and working 15 years or more on rotating night shifts was associated with an increase in colon cancer (RR 1.35, 95% CI 1.03–1.77). The mechanism is believed to be through suppression of melatonin production caused by prolonged exposure to light during nighttime hours (melatonin, in turn, is believed to have an oncostatic action).

Psychosocial Factors Several mechanisms have been hypothesized through which psychosocial factors (e.g., stress, depression, social support) may influence cancer outcomes. First, stress and depression are related to health behaviors that are linked to cancer. Cigarettes are a relatively afford-

183

able and accessible means of stress reduction particularly in the context of poverty (Emmons, 2000). Psychosocial factors have been also hypothesized to influence immune function directly (Spiegel et al., 1998). Acting via the hypothalamic-pituitary-adrenal axis in a complex feedback loop between the central nervous (CNS) and immune systems, stress increases cortisol levels (Mason, 1968; Makara et al., 1980; Rose, 1984), which adversely affects immune function (Kiecolt-Glaser et al., 1991; Dekaris et al., 1993; Herbert and Cohen, 1993; Esterling et al., 1994; Kiecolt-Glaser and Glaser, 1995; Kiecolt-Glaser et al., 1995, 1996; Cohen et al., 1999; Guidi et al., 1999; Lutgendorf et al., 1999; Song et al., 1999; Lacey et al., 2000). Frequent cortisol release that occurs with chronic stress may lead to persistently high cortisol levels (Kirschbaum et al., 1995), leading to immune suppression. Because the immune system is involved in the body’s immune surveillance mechanism (e.g., eliminating mutated cells), immune dysfunction may lead to more rapid development of cancer. Stress may also promote cancer through DNA damage, faulty DNA repair, inhibition of apoptosis, effects on endocrine parameters, or somatic mutation (Forlenza and Baum, 2000). These situations may be precursors to certain types of cancer such as hormonal cancers (Riley, 1981; Rowse et al., 1992) and lymphatic cancers (Fox, 1995; Levav et al., 2000). Despite plausible mechanisms linking psychosocial factors to cancer outcomes, the empirical evidence has been decidedly mixed. For example, certain types of job stress are hypothesized to be linked to increased disease risk, including cancer. According to Karasek (Karasek and Theorell, 1990), jobs that are both high in psychological demands and low in decision-making authority (or control) are hypothesized to result in job strain. High strain jobs are typically overrepresented in lower SES occupations, such as assembly-line jobs and certain kinds of jobs in the clerical and service sectors. In turn, job strain has been shown to increase the likelihood of deleterious behavioral coping responses (e.g., cigarette smoking, alcohol abuse), thereby increasing the risk of disease outcomes. Although most epidemiologic studies of job strain and cardiovascular disease have found support for such an association (Karasek et al., 1981; Karasek and Theorell, 1990; Schnall et al., 1990), the link to cancer has remained elusive. For example, in the Harvard Nurses’ Health Study, Achat and colleagues (2000) found no evidence of an association between job strain and breast cancer incidence in a cohort of nurses (Achat et al., 2000). Although the number of breast cancer cases was relatively small (n = 219), a recent follow-up study that included 1030 breast cancer cases found an inverse, rather than the expected positive, association between job strain and breast cancer (Schernhammer et al., 2004). A considerable amount of work has been devoted to exploring whether social connection (social ties and social support) is linked to breast cancer survival. The relation between low SES and lack of social support is not uniform or consistent, although social isolation may be a stronger risk factor for adverse health outcomes in disadvantaged women because of their high levels of background stress (i.e., there may be effect modification). Observational studies of women in different stages of breast cancer have typically found a slightly increased mortality rate (RR 1.4) following breast cancer among those with low levels of social support or social integration, adjusted for stage of disease and treatment factors (Reynolds et al., 1994; Maunsell et al., 1995; Goodwin et al., 1996). Promising results from early trials (Spiegel et al., 1989), however, have not been borne out by more recent studies. Social support interventions appear to have little effect on survival outcomes, at least among women with metastatic breast cancer (Goodwin et al., 2001). Little work has explored social connections in relation to other cancer outcomes. Nevertheless, social isolation (Reynolds et al., 1994; Reynolds et al., 2000) in combination with depression may be plausibly hypothesized to reduce the likelihood of adherence to medications or even the will to live.

Access to Health Care and Screening Socioeconomic status is inversely related to the likelihood of undergoing cancer screening (Katz and Hofer, 1994; Katz et al., 2000). In

184

PART II: THE MAGNITUDE OF CANCER

Healthy People 2010, women with high incomes were twice as likely to undergo mammography (80% during a recent 2-year period) than low-income women (<45%) (U.S. Department of Health and Human Services, 2000). As a consequence, the poor often present with cancer in later, less treatable stages. Bradley and colleagues found that persons under 65 years of age who were insured by Medicaid had a greater risk of late-stage diagnosis and death for breast, cervical, colon, lung, and prostate cancer than those not insured by Medicaid (Brewster et al., 2001). People from lower SES backgrounds generally present and are diagnosed at later stages, which explains, in part, their higher mortality rates after cancer (Farley and Flannery, 1989; Simon and Severson, 1997; Lannin et al., 1998; Marcella and Miller, 2001). Even holding the cancer stage constant, the poor have been shown to have worse survival (Bradley et al., 2001, 2002). They have more limited access to health care (Bindman et al., 1995; Hargraves, 2002) and often receive less aggressive treatment (VanEenwyk et al., 2002; Lim et al., 2003). A higher prevalence of co-morbid conditions may also complicate recovery. The socioeconomic pattern of colorectal cancer mortality began to cross over (from higher rates in lower SES areas up to the 1980s and then lower rates thereafter) prior to the dissemination of colorectal screening guidelines (Singh et al., 2002a), suggesting that factors in addition to screening are responsible for the reversal in the SES gradient for this disease. Screening for colorectal cancer has also not increased substantially in any socioeconomic group, although high SES people are slightly more likely to be screened.

Neighborhood Environments The bulk of empirical work on SES and cancer to date has focused on individual-level exposures and relations. Increasing attention has been paid more recently to the independent contribution of area-level socioeconomic factors on health outcomes (Kawachi and Berkman, 2003). A study by Merkin and colleagues found that living in areas with lower levels of education and income increased the odds of presenting with advanced-stage breast cancer by 50% for black women and by 75% for white women (Merkin et al., 2002). Community- and neighborhood-level factors, including the availability of green space (for physical activity), access to stores and grocers, and social norms and social cohesion, may have an influence on health above and beyond individual factors, including individual SES (Kawachi and Berkman, 2003). The identification of such contextual effects on health outcomes remains at a nascent stage. Special study designs and analytic techniques (multilevel analysis) are required to detect the presence of contextual effects. Obstacles abound in drawing causal inferences from such data. Nevertheless, by redirecting the focus of interventions from individuals toward improving the quality of the places where they reside, the potential existence of contextual influences on cancer risk offers a promising avenue for cancer prevention.

A NOTE ON SES VERSUS RACE Socioeconomic status is correlated with race. African Americans and Hispanics are disproportionately represented in low SES groups, whether measured by education, income, or occupation. A great deal of work has explored differential cancer outcomes by race, predominantly focusing on differences between blacks and whites. Unfortunately, in the absence of SES data in routine sources of data, there has been a tendency to use race as a proxy for SES. This has served only to obscure the understanding of how SES and race/ethnicity may be independently related to cancer outcomes. For some cancers, SES may entirely “explain” the association of race and cancer. In a study of Detroit women, Bradley and colleagues found, after accounting for SES, that race was not related to stage at presentation or survival after breast cancer diagnosis (Bradley et al., 2002). Also, after controlling for income, black women were as likely, or more likely, than white women to report having undergong recent mammography (U.S. Department of Health and Human Services,

2000). Black women from high SES backgrounds experienced breast cancer rates similar to those of affluent white women. In other cases, SES may explain part but not all of the racial difference in risk (Robbins et al., 2000). In a study of oral cancer, lower SES accounted for some, but not all, of the death risk from oral cancer among blacks. Tellingly, race and SES do not always vary in the same direction with respect to cancer incidence and mortality. Thus, when incidence rates were stratified by race and census block SES, Krieger et al. (1999) found marked heterogeneity in the relations of SES and race to individual cancer sites. Moreover, the strength of association between SES and cancer may vary by race. In a study by Yost et al., SES was positively related to breast cancer incidence, but the effect was stronger for Hispanic, Asian, and other women than it was for whites and blacks (Yost et al., 2001). The practice of conflating SES with race/ethnicity therefore poses a significant barrier to understanding the etiologic relations of each to cancer outcomes. As such, investigators are strongly urged to discontinue this practice.

CONCLUSIONS The relation of SES to cancer is both complex and dynamic. It is difficult, and potentially misleading, to make generalizations about the relation because of the heterogeneity of cancer types and outcome measures as well as the differences in approaches to measuring SES itself (different indicators, different levels of analysis). Whether the association is causal remains controversial in some cases, even though several plausible pathways have been identified by which SES could causally affect the onset and prognostic course of specific cancers. Further investigations of the link between SES and cancer would benefit from more sophisticated approaches to measuring SES as well as better specification and tests of etiologic mechanisms. Such an exercise is of more than passing academic interest, as any program or policy to reduce the burden of cancer must also pay attention to mitigating the existing socioeconomic disparities in cancer outcomes. References Achat H, Kawachi I, Byrne C, Hankinson S, Colditz G. 2000. A prospective study of job strain and risk of breast cancer. Int J Epidemiol 29:622–628. Adair LS, Gordon-Larsen P. 2001. Maturational timing and overweight prevalence in US adolescent girls. Am J Public Health 91:642–644. Adler NE, Boyce T, Chesney MA, Cohen S, Folkman S, Kahn RL, Syme SL. 1993. Socioeconomic status and health: the challenge of the gradient. Am Psychol 49:15–24. Alexander C, Piazza M, Mekos D, Valente T. 2001. Peers, schools, and adolescent cigarette smoking. J Adolesc Health 29:22–30. Altman DG, Levine DW, Coeytaux R, Slade J, Jaffe R. 1996. Tobacco promotion and susceptibility to tobacco use among adolescents aged 12 through 17 years in a nationally representative sample. Am J Public Health 86:1590–1593. American Cancer Society. 2003. Cancer Facts and Figures. Atlanta: American Cancer Society. Bankole A, Darroch JE, Singh S. 1999. Determinants of trends in condom use in the United States, 1988–1995. Fam Plann Perspect 31:264–271. Baquet CR, Horm JW, Gibbs T, Greenwald P. 1991. Socioeconomic factors and cancer incidence among blacks and whites. J Natl Cancer Inst 83:551– 557. Bassett MT, Krieger N. 1986. Social class and black-white differences in breast cancer survival. Am J Public Health 76:1400–1403. Beaglehole B, Bonita R. 1997. Public Health at the Crossroads. Cambridge, UK: Cambridge University Press. Becker DM, Yanek LR, Koffman DM, Bronner YC, 1999) Body image preferences among urban African Americans and whites from low income communities. Ethnic Dis 9:377–386. Bianchini F, Kaaks R, Vainio H. 2002. Overweight, obesity, and cancer risk. Lancet Oncol 3:565–574. Bindman AB, Grumbach K, Osmond D, Komaromy M, Vranizan K, Lurie N, Billings J, Stewart A. 1995. Preventable hospitalizations and access to health care. JAMA 274:305–311. Blaser MJ. 1999. Where does Helicobacter pylori come from and why is it going away? JAMA 282:2260–2262.

Socioeconomic Disparities in Cancer Incidence and Mortality Blisard N, Harris JM. 2001. Household Food Spending by Selected Demographics in the 1990s. ERS Agriculture Information Bulletin 773. Washington, DC: USDA. Blum RW, Beuhring T, Shew ML, Bearinger LH, Sieving RE, Resnick MD. 2000. The effects of race/ethnicity, income, and family structure on adolescent risk behaviors. Am J Public Health 90:1879–1884. Boffetta P, Kogevinas M, Westerholm P, Saracci R. 1997. Exposure to occupational carcinogens and social class differences in cancer occurrence. IARC Scientific Publ. 138. Lyon: IARC, pp. 331–341. Boffetta P, Tredaniel J, Greco A. 2000. Risk of childhood cancer and adult lung cancer after childhood exposure to passive smoke: a meta-analysis. Environ Health Perspect 108:73–82. Boggild H, Suadicani P, Hein HO, Gyntelberg F. 1999. Shift work, social class, and ischaemic heart disease in middle aged and elderly men; a 22 year follow up in the Copenhagen Male Study. Occup Environ Med 56: 640–645. Borzekowski DL, Robinson TN. 2001. The 30-second effect: an experiment revealing the impact of television commercials on food preferences of preschoolers. J Am Diet Assoc 101:42–46. Bradley CJ, Given CW, Roberts C. 2001. Disparities in cancer diagnosis and survival. Cancer 91:178–188. Bradley CJ, Given CW, Roberts C. 2002. Race, socioeconomic status, and breast cancer treatment and survival. J Natl Cancer Inst 94:490–496. Bray GA. 2002. The underlying basis for obesity: relationship to cancer. J Nutr 132(Suppl 11):3451S–3455S. Breslau N, Peterson EL. 1996. Smoking cessation in young adults: age at initiation of cigarette smoking and other suspected influences. Am J Public Health 86:214–220. Brewster DH, DThomson CS, Hole DJ, Black RJ, Stroner PL, Gillis CR. 2001. Relation between socioeconomic status and tumour stage in patients with breast, colorectal, ovarian, and lung cancer: results from four national, population based studies. BMJ 322:830–831. Bucher HC, Ragland DR. 1995. Socioeconomic indicators and mortality from coronary heart disease and cancer: a 22-year follow-up of middle-aged men. Am J Public Health 85:1231–1236. Byers T, Nestle M, McTiernan A, Doyle C, Currie-Williams A, Gansler T, Thun M. 2002. American Cancer Society guidelines on nutrition and physical activity for cancer prevention: reducing the risk of cancer with healthy food choices and physical activity. CA Cancer J Clin 52: 92–119. Castellsague X, Bosch FX, Munoz N, Meijer CJ, Shah KV, de Sanjose S, Eluf-Neto J, Ngelangel CA, Chichareon S, Smith JS, Herrero R, Moreno V, Franceschi S. 2002. Male circumcision, penile human papillomavirus infection, and cervical cancer in female partners. N Engl J Med 346: 1105–1112. Centers for Disease Control and Prevention. 1997. Update: prevalence of overweight among children, adolescents, and adults—United States, 1988–1994. MMWR Morbid Mortal Wkly Rep 46:199–202. Centers for Disease Control and Prevention 2000. Behavioral Risk Factor Surveillance System: Prevalence data: Nationwide—2000. nutrition. Atlanta: CDC. Centers for Disease Control and Prevention 2001. Behavioral Risk Factor Surveillance System: Prevalence data: Nationwide—2001. Alcohol Consumption. Atlanta: CDC. Chestnutt IG, Ashraf FJ. 2002. Television advertising of foodstuffs potentially detrimental to oral health: a content analysis and comparison of children’s and primetime broadcasts. Community Dent Health 19:86–89. Cohen S, Doyle WJ, Skoner DP. 1999. Psychological stress, cytokine production, and severity of upper respiratory illness. Psychosom Med 61: 175–180. Coker AL, Richter DL, Valois RF, McKeown RE, Garrison CZ, Vincent ML. 1994. Correlates and consequences of early initiation of sexual intercourse. J School Health 64:372–377. Colditz GA, Giovannucci E, Rimm EB, Stampfer MJ, Rosner B, Speizer FE, Gordis E, Willett WC. 1991. Alcohol intake in relation to diet and obesity in women and men. Am J Clin Nutr 54:49–55. Coleman MP, Babb P, Sloggett A, Quinn M, De Stavola B. 2001. Socioeconomic inequalities in cancer survival in England and Wales. Cancer 91(Suppl 1):208–216. Conigrave KM, Hu BF, Camargo CA Jr, Stampfer MJ, Willett WC, Rimm EB. 2001. A prospective study of drinking patterns in relation to risk of type 2 diabetes among men. Diabetes 50:2390–2395. Crawford PB, Story M, Wang MC, Ritchie LD, Sabry ZI. 2001. Ethnic issues in the epidemiology of childhood obesity. Pediatr Clin North Am 48: 855–878. Currie J, Hyson R. 1999. Is the impact of health shocks cushioned by socioeconomic status? The case of low birthweight. Am Econ Rev 89:245– 250.

185

Curtin F, Morabia A, Bernstein M. 1998. Lifetime exposure to environmental tobacco smoke among urban women: differences by socioeconomic class. Am J Epidemiol 148:1040–1047. Davey-Smith G, Gunnell D, Ben-Shlomo Y. 2001. Life-course approaches to socio-economic differentials in cause-specific adult mortality. In: Leon D, Walt G, editors. Poverty, Inequality and Health. An International Perspective. Oxford: Oxford University Press. Davies MJ, Baer DJ, Judd JT, Brown ED, Campbell WS, Taylor PR. 2002. Effects of moderate alcohol intake on fasting insulin and glucose concentrations and insulin sensitivity in postmenopausal women: a randomized controlled trial. Jama 287:2559–2562. Dayal HH, Power RN, Chiu C. 1982. Race and socio-economic status in survival from breast cancer. J Chronic Dis 35:675–683. Dekaris D, Sabioncello A, Mazuran R, Rabatic S, Svoboda-Beusan I, Racunica NL, Tomasic J. 1993. Multiple changes of immunologic parameters in prisoners of war: assessments after release from a camp in Manjaca, Bosnia. JAMA 270:595–599. Dennis LK, Dawson DV. 2002. Meta-analysis of measures of sexual activity and prostate cancer. Epidemiology 13:72–79. Devesa SS, Diamond EL. 1980. Association of breast cancer and cervical cancer incidence with income and education among whites and blacks. J Natl Cancer Inst 65:515–528. Dietz WH. 1998a. Childhood weight affects adult morbidity and mortality. J Nutr 128(Suppl 2):411S–414S. Dietz WH. 1998b. Health consequences of obesity in youth: childhood predictors of adult disease. Pediatrics 101:518–525. Diez-Roux AV, Nieto FJ, Caulfield L, L., Tyroler HA, Watson RL, Szklo M. 1999. Neighbourhood differences in diet: the Atherosclerosis Risk in Communities (ARIC) study. J Epidemiol Community Health 53:55–63. Eckersley R, Dixon J, Douglas B. 2001. The Social Origins of Health and WellBeing. Cambridge, UK: Cambridge University Press. Eckhart W. 1998. Viruses and human cancer. Sci Prog 81(Pt 4):315–328. Ellen JM, Kohn RP, GA, Shiboski S, Krieger N. 1995. Socioeconomic differences in sexually transmitted disease rates among black and white adolescents, San Francisco, 1990 to 1992. Am J Public Health 85:1546–1548. Emmons KM. 2000. Health Behaviors in a Social Context. In: Berkman LF, Kawachi I, editors. Social Epidemiology.New York: Oxford University Press. Emmons KM, Thompson B, McLerran D, Sorensen G, Linnan L, BasenEnquist K, Biener L. 2000. The relationship between organizational characteristics and the adoption of workplace smoking policies. Health Educ Behav 27:483–501. Emmons KM, Wong M, Hammond SK, Velicer WF, Fava JL, Monroe AD, Evans JL. 2001. Intervention and policy issues related to children’s exposure to environmental tobacco smoke. Prev Med 32:321–331. Erkkila AT, Sarkkinen ES, Lehto S, Pyorala K, Uusitupa MI. 1999. Diet in relation to socioeconomic status in patients with coronary heart disease. Eur J Clin Nutr 53:662–668. Escobedo LG, Anda RF, Smith PF, Remington PL, Mast EE. 1990. Sociodemographic characteristics of cigarette smoking initiation in the United States: implications for smoking prevention policy. Jama 264:1550–1555. Esterling BA, Kiecolt-Glaser JK, Bodnar JC, Glaser R. 1994. Chronic stress, social support, and persistent alterations in the natural killer cell response to cytokines in older adults. Health Psychol 13:291–298. Evans AS, Mueller NE. 1997. Malignant lymphomas. in: Evans AS, Kaslow R, editors. Viral Infections of Humans: Epidemiology and Control. New York: Plenum. Facchini F, Chen YD, Reaven GM. 1994. Light-to-moderate alcohol intake is associated with enhanced insulin sensitivity. Diabetes Care 17:115–119. Faggiano F, Partanen T, Kogevinas M, Boffetta P. 1997. Socioeconomic Differences in Cancer Incidence and Mortality. IARC Scientific Publ. 138. Lyon: IARC, pp. 65–176. Farley TA, Flannery JT. 1989. Late-stage diagnosis of breast cancer in women of lower socioeconomic status: public health implications. Am J Public Health 79:1508–1512. Farrell P, Fuchs VR. 1982. Schooling and health: the cigarette connection. J Health Econ 1:217–230. Federal Trade Commission. 1999. Report to Congress for 1997 Pursuant to the Federal Cigarette Labeling and Advertising Act. Washington, DC: FTC. Flegal KM, Carroll MD, Ogden CL, Johnson CL. 2002. Prevalence and trends in obesity among US adults, 1999–2000. Jama 288:1723–1727. Forlenza MJ, Baum A. 2000. Psychosocial influences on cancer progression: alternative cellular and molecular mechanisms Curr Opin Psychiatry 13:639–645. Fox AJ, Goldblatt PO, Jones DR. 1985. Social class mortality differentials: artefact, selection or life circumstances? J Epidemiol Community Health 39:1–8.

186

PART II: THE MAGNITUDE OF CANCER

Fox BH. 1995. The role of psychological factors in cancer incidence and prognosis. Oncology (Huntingt) 9:245–256. Fried EJ, Nestle M. 2002. The growing political movement against soft drinks in schools. Jama 288:2181. Friedenreich CM, Orenstein MR. 2002. Physical activity and cancer prevention: etiologic evidence and biological mechanisms, J Nutr 132(Suppl 11):3456S–3464S. Fuchs VR. 1983. How We Live: an Economic Perspective on Americans from Birth to Death. Cambridge, MA: Harvard University Press. Gamble M, Cotugna N. 1999. A quarter century of TV food advertising targeted at children. Am J Health Behav 23:261–267. Giskes K, Turrell G, Patterson C, Newman B. 2002. Socioeconomic differences among Australian adults in consumption of fruit and vegetables and intakes of vitamins A, C and folate. J Hum Nutr Diet 15:375–390. Giuliano AR, Papenfuss M, Schneider A, Nour M, Hatch K. 1999. Risk factors for high-risk type human papillomavirus infection among MexicanAmerican women. Cancer Epidemiol Biomarkers Prev 8:615–620. Gliksman MD, Kawachi I, Hunter D, Colditz GA, Manson JE, Stampfer MJ, Speizer FE, Willett WC, Hennekens CH. 1995. Childhood socioeconomic status and risk of cardiovascular disease in middle aged US women: a prospective study. J Epidemiol Community Health 49:10-15.Goodman E. 1999. The role of socioeconomic status gradients in explaining differences in US adolescents’ health. Am J Public Health 89:1522–1528. Goodman KJ, Correa P. 1995. The transmission of Helicobacter pylori: a critical review of the evidence. Int J Epidemiol 24:875–887. Goodwin PJ, Leszcz M, Ennis M, Koopmans J, Vincent L, Guther H, Drysdale E, Hundleby M, Chochinov HM, Navarro M, Speca M, Hunter J. 2001. The effect of group psychosocial support on survival in metastatic breast cancer. N Engl J Med 345:1719–1726. Goodwin PJ, Leszcz M, Koopmans J, Arnold A, Doll R, Chochinov H, Navarro M, Butler K, Pritchard KI. 1996. Randomized trial of group psychosocial support in metastatic breast cancer: the BEST study; BreastExpressive Supportive Therapy study. Cancer Treat Rev 22(Suppl A): 91–96. Gordon NH, Crowe JP, Brumberg DJ, Berger NA. 1992. Socioeconomic factors and race in breast cancer recurrence and survival. Am J Epidemiol 135:609–618. Greenland S. 2000. An introduction to instrumental variables for epidemiologists. Int J Epidemiol 29:722–729. Griesler PC, Kandel DB. 1998. Ethnic differences in correlates of adolescent cigarette smoking. J Adolesc Health 23:167–180. Grossman M. 1975. The correlation between health and schooling. In: Terleckyj NE, editor. Household Production and Consumption. New York: Columbia University Press. Grusky DB. 1994. Social Stratification in Sociological Perspective. Boulder, CO: Westview Press. Guidi L, Tricerri A, Vangeli M, Frasca D, Riccardo Errani A, Di Giovanni A, Antico L, Menini E, Sciamanna V, Magnavita N, Doria G, Bartoloni C. 1999. Neuropeptide Y plasma levels and immunological changes during academic stress. Neuropsychobiology 40:188–195. Gustavsson P, Nyberg F, Pershagen G, Scheele P, Jakobsson R, Plato N. 2002. Low-dose exposure to asbestos and lung cancer: dose-response relations and interaction with smoking in a population-based case-referent study in Stockholm, Sweden. Am J Epidemiol 155:1016–1022. Gutensohn NM. 1982. Social class and age at diagnosis of Hodgkin’s disease: new epidemiologic evidence for the “two-disease hypothesis.” Cancer Treat Rep 66:689–695. Guthrie JF, Morton JR. 2000. Food sources of added sweeteners in the diets of Americans. J Am Diet Assoc 100:43–51, quiz 49–50. Hamajima N, Hirose K, Tajima K, Rohan T, Calle EE, et al. 2002. Alcohol, tobacco and breast cancer—collaborative reanalysis of individual data from 53 epidemiological studies, including 58,515 women with breast cancer and 95,067 women without the disease. Br J Cancer 87:1234–1245. Hargraves JL. 2002. The Insurance Gap and Minority Health Care, 1997-2001. In: Center for Studying Health Systems Change Data Bulletin. Washington, DC: Center for Studying Health System Change. Harnack LJ, Jeffery RW, Boutelle KN. 2000.Temporal trends in energy intake in the United States: an ecologic perspective. Am J Clin Nutr 71:1478– 1484. Harrell JS, Bangdiwala SI, Deng S, Webb JP, Bradley C. 1998. Smoking initiation in youth: the roles of gender, race, socioeconomics, and developmental status. J Adolesc Health 23:271–279. Heini AF, Weinsier RL. 1997. Divergent trends in obesity and fat intake patterns: the American paradox. Am J Med 102:259–264. Herbert TB, Cohen S. 1993. Stress and immunity in humans: a meta-analytic review. Psychosom Med 55:364–379. Hofferth SL. 1987. Factors affecting initiation of sexual intercourse. in: Hofferth SL, Hayes CD, editors. Risking the Future: Adolescent

Sexuality, Pregnancy, and Childbearing, vol 2. Washington, DC: National Academy Press. Hogan DP, Sun R, Cornwell GT. 2000. Sexual and fertility behaviors of American females aged 15–19 years: 1985, 1990, and 1995. Am J Public Health 90:1421–1425. Holly EA, Bracci PM, Mueller BA, Preston-Martin S. 1998. Farm and animal exposures and pediatric brain tumors: results from the United States West Coast Childhood Brain Tumor Study. Cancer Epidemiol Biomarkers Prev 7:797–802. Howard G, Anderson RT, Russell G, Howard VJ, Burke GL. 2000. Race, socioeconomic status, and cause-specific mortality. Ann Epidemiol 10:214– 223. International Agency for Research on Cancer Ad Hoc Working Group. 1994. Schistosomes, Liver Flukes and Helicobacter pylori. Views and Expert Opinions of an IARC Working group on the Evaluation of Carcinogenic Risks to Humans. Lyon: IARC, p. 61. Jahns L, Siega-Riz AM, Popkin BM. 2001. The increasing prevalence of snacking among US children from 1977 to 1996 J Pediatr 138:493–498. John EM, Savitz DA, Sandler DP. 1991. Prenatal exposure to parents’ smoking and childhood cancer Am J Epidemiol 133:123–132. Kamimoto LA, Easton AN, Maurice E, Husten CG, Macera CA. 1999. Surveillance for five health risks among older adults—United States, 19931997. MMWR CDC Surveill Summ 48:89–130. Karasek R, Theorell T. 1990. Healthy Work. New York: Basic Books. Karasek R, Baker D, Marxer F, Ahlbom A, Theorell T. 1981. Job decision latitude, job demands, and cardiovascular disease: a prospective study of Swedish men. Am J Public Health 71:694–705. Katz SJ, Hofer TP. 1994. Socioeconomic disparities in preventive care persist despite universal coverage: breast and cervical cancer screening in Ontario and the United States. Jama 272:530–534. Katz SJ, Zemencuk JK, Hofer TP. 2000. Breast cancer screening in the United States and Canada, 1994: socioeconomic gradients persist. Am J Public Health 90:799–803. Kauppinen T, Teschke K, Savela A, Kogevinas M, Boffetta P. 1997. International data base of exposure measurements in the pulp, paper and paper product industries. Int Arch Occup Environ Health 70:119–127. Kawachi I, Berkman LF. 2003. Neighborhoods and Health. New York: Oxford University Press. Kawachi I, Colditz GA, Stampfer MJ, Willett WC, Manson JE, Speizer FE, Hennekens CH. 1995. Prospective study of shift work and risk of coronary heart disease in women. Circulation 92:3178–3182. Kiecolt-Glaser JK, Glaser R. 1995. Psychoneuroimmunology and health consequences: data and shared mechanisms. Psychosom Med 57:269–274. Kiecolt-Glaser JK, Dura JR, Speicher CE, Trask OJ, Glaser R. 1991. Spousal caregivers of dementia victims: longitudinal changes in immunity and health. Psychosom Med 53:345–362. Kiecolt-Glaser JK, Glaser R, Gravenstein S, Malarkey WB, Sheridan J. 1996. Chronic stress alters the immune response to influenza virus vaccine in older adults. Proc Natl Acad Sci USA 93:3043–3047. Kiecolt-Glaser JK, Marucha PT, Malarkey WB, Mercado AM, Glaser R. (1995) Slowing of wound healing by psychological stress. Lancet 346:1194– 1196. Kirschbaum C, Prussner JC, Stone AA, Federenko I, Gaab J, Lintz D, Schommer N, Hellhammer DH. 1995. Persistent high cortisol responses to repeated psychological stress in a subpopulation of healthy men. Psychosom Med 57:468–474. Kogevinas M, Becher H, Benn T, Bertazzi PA, Boffetta P, Bueno-deMesquita HB, Coggon D, Colin D, Flesch-Janys D, Fingerhut M, Green L, Kauppinen T, Littorin M, Lynge E, Mathews JD, Neuberger M, Pearce N, Saracci R. 1997a. Cancer mortality in workers exposed to phenoxy herbicides, chlorophenols, and dioxins: an expanded and updated international cohort study. Am J Epidemiol 145:1061–1075. Kogevinas M, Pearce N, Susser M, Boffetta P. 1997b. Social inequalities and cancer. IARC Scientific Publ. 138. Lyon: IARC. Kreuzer M, Heinrich J, Kreienbrock L, Rosario AS, Gerken M, Wichmann HE. 2002. Risk factors for lung cancer among nonsmoking women. Int J Cancer 100:706–713. Krieger N. 2000. Discrimination and health. In: Berkman LF, Kawachi I, editors. Social Epidemiology. New York: Oxford University Press. Krieger N, Chen JT, Ebel G. 1997. Can we monitor socioeconomic inequalities in health? A survey of U.S. health departments’ data collection and reporting practices. Public Health Rep 112:481–491. Krieger N, Quesenberry C Jr, Peng T, Horn-Ross P, Stewart S, Brown S, Swallen K, Guillermo T, Suh D, Alvarez-Martinez L, Ward F. 1999. Social class, race/ethnicity, and incidence of breast, cervix, colon, lung, and prostate cancer among Asian, Black, Hispanic, and White residents of the San Francisco Bay area, 1988–92 (United States). Cancer Causes Control 10:525–537.

Socioeconomic Disparities in Cancer Incidence and Mortality Kuczmarski RJ, Flegal KM, Campbell SM, Johnson CL. 1994. Increasing prevalence of overweight among US adults: the national health and nutrition examination surveys, 1960 to 1991. Jama 272:205–211. Lacey K, Zaharia MD, Griffiths J, Ravindran AV, Merali Z, Anisman H. 2000. A prospective study of neuroendocrine and immune alterations associated with the stress of an oral academic examination among graduate students. Psychoneuroendocrinology 25:339–356. Lammers C, Ireland M, Resnick M, Blum R. 2000. Influences on adolescents’ decision to postpone onset of sexual intercourse: a survival analysis of virginity among youths aged 13 to 18 years. J Adolesc Health 26:42–48. Lannin DR, Mathews HF, Mitchell J, Swanson MS, Swanson FH, Edwards MS. 1998. Influence of socioeconomic and cultural factors on racial differences in late-stage presentation of breast cancer. Jama 279:1801– 1807. Lee RE, Cubbin C. 2002. Neighborhood context and youth cardiovascular health behaviors. Am J Public Health 92:428–436. Leong NM, Mignone LI, Newcomb PA, Titus-Ernstoff L, Baron JA, TrenthamDietz A, Stampfer MJ, Willett WC, Egan KM. 2003. Early life risk factors in cancer: the relation of birth weight to adult obesity. Int J Cancer 103:789–791. Levav I, Kohn R, Iscovich J, Abramson JH, Tsai WY, Vigdorovich D. 2000. Cancer incidence and survival following bereavement. Am J Public Health 90:1601–1607. Li R, Serdula M, Bland S, Mokdad A, Bowman B, Nelson D. 2000. Trends in fruit and vegetable consumption among adults in 16 US states: Behavioral Risk Factor Surveillance System, 1990-1996. Am J Public Health 90:777–781. Liberatos P, Link BG, Kelsey JL. 1988. The measurement of social class in epidemiology. Epidemiol Rev 10:87–121. Lim JE, Chien MW, Earle CC. 2003. Prognostic factors following curative resection for pancreatic adenocarcinoma: a population-based, linked database analysis of 396 patients. Ann Surg 237:74–85. Link BG, Phelan J. 1995. Social conditions as fundamental causes of disease. J Health Soc Behav (Spec No):80–94. Liu L, Deapen D, Bernstein L. 1998. Socioeconomic status and cancers of the female breast and reproductive organs: a comparison across racial/ethnic populations in Los Angeles County, California (United States). Cancer Causes Control 9:369–380. Lochner K, Kawachi I. 2000. Socioeconomic status. In: Hunter D, Colditz GA, editors. Cancer Prevention: the Causes and Prevention of Cancer, Vol. 1. Dordrecht: Kluwer Academic. Lutgendorf SK, Garand L, Buckwalter KC, et al. 1999. Life stress, mood disturbance, and elevated interleukin-6 in healthy older women. J Gerontol 54:M434–M439. Lynch J, Kaplan GA. 2000. Socioeconomic position. In: Berkman LF, Kawachi I, editors. Social Epidemiology. New York: Oxford University Press. Makara GB, Palkovits M, Szentagoithai J. 1980. The endocrine hypothalamus and the hormonal response to stress. In: Selye H, editor. Selye’s Guide to Stress Research, vol 1. New York: Van Nostrand Reinhold. Marcella S, Miller JE. 2001. Racial differences in colorectal cancer mortality: the importance of stage and socioeconomic status. J Clin Epidemiol 54: 359–366. Mason JW. 1968. A review of the psychoendocrine research on the pituitaryadrenal cortisol system. Psychosom Med 30:576–607. Maunsell E, Brisson J, Deschenes L. 1995. Social support and survival among women with breast cancer. Cancer 76:631–637. McCormack VA, dos Santos Silva I, De Stavola BL, Mohsen R, Leon DA, Lithell HO. 2003. Fetal growth and subsequent risk of breast cancer: results from long term follow up of Swedish cohort. Bmj 326:248. McDonough P, Duncan GJ, Williams D, House J. 1997. Income dynamics and adult mortality in the United States, 1972 through 1989. Am J Public Health 87:1476–1483. McEwen BS. 1998. Protective and damaging effects of stress mediators. N Engl J Med 338:171–179. Merkin SS, Stevenson L, Powe N. 2002. Geographic socioeconomic status, race, and advanced-stage breast cancer in New York City. Am J Public Health 92:64–70. Micic D. 2001. Obesity in children and adolescents—a new epidemic? Consequences in adult life. J Pediatr Endocrinol Metab 14(Suppl 5):1345–1352; discussion 1365. Moore DB, Howell PB, Treiber FA. 2002. Changes in overweight in youth over a period of 7 years: impact of ethnicity, gender and socioeconomic status. Ethnic Dis 12(Suppl 1):83–86. Morimoto LM, White E, Chen Z, Chlebowski RT, Hays Kuller L, Lopez AM, Manson J, Margolis KL, Muti PC, Stefanick ML, McTiernan A. 2002. Obesity, body size, and risk of postmenopausal breast cancer: the Women’s Health Initiative (United States). Cancer Causes Control 13:741–751.

187

Nielsen SJ, Popkin BM. 2003. Patterns and trends in food portion sizes, 1977– 1998. Jama 289:450–453. Nielsen SJ, Siega-Riz AM, Popkin BM. 2002. Trends in energy intake in U.S. between 1977 and 1996: similar shifts seen across age groups. Obes Res 10:370–378. Nightingale TE, Gruber J. 1994. Helicobacter and human cancer. J Natl Cancer Inst 86:1505–1509. O’Donnell BL, O’Donnell CR, Stueve A. 2001. Early sexual initiation and subsequent sex-related risks among urban minority youth: the reach for health study. Fam Plann Perspect 33:268–275. Pamuk E, Makuc D, Heck K, Reuben C, Lochner K. 1998. Socioeconomic Status and Health Chartbook. Health, United States, 1998. Hyattsville, MD: National Center for Health Statistics. Parker JD, Schoendorf KC, Kiely JL. 1994. Associations between measures of socioeconomic status and low birth weight, small for gestational age, and premature delivery in the United States. Ann Epidemiol 4:271–278. Parker-Pope T. 2001. Cigarettes. Anatomy of an Industry from Seed to Smoke. New York: New Press. Parsonnet J. 1995. The incidence of Helicobacter pylori infection. Aliment Pharmacol Ther 9(Suppl 2):45–51. Parsonnet J, Shmuely H, Haggerty T. 1999. Fecal and oral shedding of Helicobacter pylori from healthy infected adults. Jama 282:2240–2245. Pearce N, Matos E. 1994. Occupational Cancer in Developing Countries: Introduction. IARC Scientific Publ. 129. Lyon: IARC, pp. 1–3. Pierce JP, Choi WS, Gilpin EA, Farkas AJ, Berry CC. 1998. Tobacco industry promotion of cigarettes and adolescent smoking. Jama 279:511–515. Platz EA, Giovannucci E, Rimm EB, Curhan GC, Spiegelman D, Colditz GA, Willett WC. 1998. Retrospective analysis of birth weight and prostate cancer in the Health Professionals Follow-up Study. Am J Epidemiol 147:1140–1144. Popkin BM, Siega-Riz AM, Haines PS. 1996. A comparison of dietary trends among racial and socioeconomic groups in the United States. N Engl J Med 335:716–720. Power C, Manor O, Fox AJ, Fogelman K. 1990. Health in childhood and social inequalities in young adults. J R Stat Soc B 153(series A):17–28. Preston-Martin S, Pogoda JM, Mueller BA, Holly EA, Lijinsky W, Davis RL. 1996. Maternal consumption of cured meats and vitamins in relation to pediatric brain tumors. Cancer Epidemiol Biomarkers Prev 5:599–605. Putnam JJ, Allshouse JE. 1999. Food Consumption, Prices, and Expenditures, 1970-97. Statistical Bulletin No. 965. Washington, DC: Food and Rural Economics Division, Economic Research Service, U.S. Department of Agriculture. Reynolds P, Boyd PT, Blacklow RS, Jackson JS, Greenberg RS, Austin DF, Chen VW, Edwards BK. 1994. The relationship between social ties and survival among black and white breast cancer patients; National Cancer Institute Black/White Cancer Survival Study Group. Cancer Epidemiol Biomarkers Prev 3:253–259. Reynolds P, Hurley S, Torres M, Jackson J, Boyd P, Chen VW. 2000. Use of coping strategies and breast cancer survival: results from the Black/White Cancer Survival Study. Am J Epidemiol 152:940–949. Riley V. 1981. Psychoneuroendocrine influences on immunocompetence and neoplasia. Science 212:1100–1109. Robbins JM, Vaccarino V, Zhang H, Kasl SV. 2000. Excess type 2 diabetes in African-American women and men aged 40–74 and socioeconomic status: evidence from the Third National Health and Nutrition Examination Survey. J Epidemiol Community Health 54:839–845. Rose RM. 1984. Overview of endocrinology of stress. In: Brown GM, Koslow SH, Reichlin S, editors. Neuroendocrinology and Psychiatric Disorder. New York: Raven Press. Ross CE, Wu C. 1995. The links between education and health. Am Sociol Rev 60:719–745. Ross JA, Swensen AR. 2000. Prenatal epidemiology of pediatric tumors. Curr Oncol Rep 2:234–241. Rothman KJ. 1980. The proportion of cancer attributable to alcohol consumption. Prev Med 9:174–179. Rowse GJ, Weinberg J, Bellward GD, Emerman JT. 1992. Endocrine mediation of psychosocial stressor effects on mouse mammary tumor growth. Cancer Lett 65:85–93. Sandhu MS, Luben R, Day NE, Khaw KT. 2002. Self-reported birth weight and subsequent risk of colorectal cancer. Cancer Epidemiol Biomarkers Prev 11:935–938. Santelli JS, Lowry R, Brener ND, Robin L. 2000. The association of sexual behaviors with socioeconomic status, family structure, and race/ethnicity among US adolescents. Am J Public Health 90:1582–1588. Schernhammer ES, Laden F, Speizer FE, Willett WC, Hunter DJ, Kawachi I, Colditz GA. 2001. Rotating night shifts and risk of breast cancer in women participating in the nurses’ health study. J Natl Cancer Inst 93:1563–1568.

188

PART II: THE MAGNITUDE OF CANCER

Schernhammer ES, Hankinson SE, Rosner B, Kroenke CH, Willett WC, Colditz GA, Kawachi I. 2004. Job stress and breast cancer risk: The Nurses’ Health Study. Am J Epidemiol 160:1079–1086. Schnall PL, Pieper C, Schwartz JE, Karasek RA, Schlussel Y, Devereux RB, Ganau A, Alderman M, Warren K, Pickering TG. 1990. The relationship between ‘job strain,’ workplace diastolic blood pressure, and left ventricular mass index: results of a case-control study. Jama 263: 1929–1935. Schuster MA, Franke T, Pham CB. 2002. Smoking patterns of household members and visitors in homes with children in the United States. Arch Pediatr Adolesc Med 156:1094–1100. Scragg R, Laugesen M, Robinson E. 2002. Cigarette smoking, pocket money and socioeconomic status: results from a national survey of 4th form students in 2000. N Z Med J 115:U108. Seeman M, Seeman TE. 1983. Health behavior and personal autonomy: a longitudinal study of the sense of control in illness. J Health Soc Behav 24:144–160. Shi L. 1998. Sociodemographic characteristics and individual health behaviors. South Med J 91:933–941. Siega-Riz AM, Popkin BM. 2001. Dietary trends among low socioeconomic status women of childbearing age in the United States from 1977 to 1996: a comparison among ethnic groups. J Am Med Womens Assoc 56(2):44– 48, 72. Simon MS, Severson RK. 1997. Racial differences in breast cancer survival: the interaction of socioeconomic status and tumor biology. Am J Obstet Gynecol 176:S233–S239. Singh GK, Miller BA, Hankey BF. 2002a. Changing area socioeconomic patterns in U.S. cancer mortality, 1950-1998. Part II. Lung and colorectal cancers. J Natl Cancer Inst 94:916–925. Singh GK, Miller BA, Hankey BF, Feuer EJ, Pickle LW. 2002b. Changing area socioeconomic patterns in U.S. cancer mortality, 1950–1998. Part I. All cancers among men. J Natl Cancer Inst 94:904–915. Singh S, Darroch JE, Frost JJ. 2001. Socioeconomic disadvantage and adolescent women’s sexual and reproductive behavior: the case of five developed countries. Fam Plann Perspect 33:251–258, 289. Smith GD, Neaton JD, Wentworth D, Stamler R, Stamler J. 1996. Socioeconomic differentials in mortality risk among men screened for the Multiple Risk Factor Intervention Trial. I. White men. Am J Public Health 86:486–496. Song C, Kenis G, van Gastel A, Bosmans E, Lin A, de Jong R, Neels H, Scharpe S, Janca A, Yasukawa K, Maes M. 1999. Influence of psychological stress on immune-inflammatory variables in normal humans. Part II. Altered serum concentrations of natural anti-inflammatory agents and soluble membrane antigens of monocytes and T lymphocytes. Psychiatry Res 85:293–303. Sood AK. 1991. Cigarette smoking and cervical cancer: meta-analysis and critical review of recent studies. Am J Prev Med 7:208–213. Sorlie PD, Backlund E, Keller JB. 1995. US mortality by economic, demographic, and social characteristics: the National Longitudinal Mortality Study. Am J Public Health 85:949–956. Spiegel D, Bloom JR, Kraemer HC, Gottheil E. 1989. Effect of psychosocial treatment on survival of patients with metastatic breast cancer. Lancet 2:888–891. Spiegel D, Sephton SE, Terr AI, Stites DP. 1998. Effects of psychosocial treatment in prolonging cancer survival may be mediated by neuroimmune pathways. Ann NY Acad Sci 840:674–683.

Steenland K, Henley J, Thun M. 2002. All-cause and cause-specific death rates by educational status for two million people in two American Cancer Society cohorts, 1959–1996. Am J Epidemiol 156:11–21. Strauss RS, Knight J. 1999. Influence of the home environment on the development of obesity in children. Pediatrics 103:e85. Su LJ, Arab L. 2001. Nutritional status of folate and colon cancer risk: evidence from NHANES I epidemiologic follow-up study. Ann Epidemiol 11:65–72. Taras HL, Gage M. 1995. Advertised foods on children’s television. Arch Pediatr Adolesc Med 149:649–652. Taylor R, Cumming R, Woodward A, Black M. 2001. Passive smoking and lung cancer: a cumulative meta-analysis. Aust N Z J Public Health 25:203–211. Tirodkar MA, Jain A. 2003. Food messages on African American television shows. Am J Public Health 93:439–441. Trichopoulos D. 1990. Hypothesis: does breast cancer originate in utero? Lancet 335:939–940. Tyas SL, Pederson LL. 1998. Psychosocial factors related to adolescent smoking: a critical review of the literature. Tob Control 7:409–420. U.S. Department of Health and Human Services (DHHS). 2000. Healthy People 2010: Understanding and Improving Health, 2nd ed. Washington, DC: U.S. Government Printing Office. U.S. Department of Health and Human Services (DHHS). 2002. Report on Carcinogens, 10th ed. Washington, DC: DHHHS, Public Health Service, National Toxicology Program. VanEenwyk J, Campo JS, Ossiander EM. 2002. Socioeconomic and demographic disparities in treatment for carcinomas of the colon and rectum. Cancer 95:39–46. Vatten LJ, Maehle BO, Lund Nilsen TI, Tretli S, Hsieh CC, Trichopoulos C, Stuver SO. 2002. Birth weight as a predictor of breast cancer: a casecontrol study in Norway. Br J Cancer 86:89–91. Wadsworth MEJ. 1986. Serious illness in childhood and its association with later-life achievement. In: Wilkinson RG, editor. Class and Health: Research and Longitudinal Data. London: Tavistock. Wang Y. 2001. Cross-national comparison of childhood obesity: the epidemic and the relationship between obesity and socioeconomic status. Int J Epidemiol 30:1129–1136. Weston TL, Aronson KJ, Siemiatycki J, Howe GR, Nadon L. 2000. Cancer mortality among males in relation to exposures assessed through a jobexposure matrix. Int J Occup Environ Health 6:194–202. Wiencke JK, Kelsey KT. 2002. Teen smoking, field cancerization, and a “critical period” hypothesis for lung cancer susceptibility. Environ Health Perspect 110:555–558. Wilkinson RG. 1996. Unhealthy Societies. The Afflictions of Inequality. London: Routledge. Williams DR. 1997. Commentary on monitoring socioeconomic status. Public Health Rep 112:492–494. Yen IH, Moss N. 1999. Unbundling education: a critical discussion of what education confers and how it lowers risk for disease and death. Ann NY Acad Sci 896:350–351. Yost K, Perkins C, Cohen R, Morris C, Wright W. 2001. Socioeconomic status and breast cancer incidence in California for different race/ethnic groups. Cancer Causes Control 12:703–711. Zizza C, Siega-Riz AM, Popkin BM. 2001. Significant increase in young adults’ snacking between 1977–1978 and 1994–1996 represents a cause for concern! Prev Med 32:303–310.

11

Migrant Studies LAURENCE N. KOLONEL AND LYNNE R. WILKENS

M

igrants are individuals who move from one location and resettle in another. Usually this entails transplantation from one country (referred to as the “home country” or “country of origin”) to another (“host country” or “adopted country”), though migrants are sometimes identified as persons who move from one region to another within the same country (“internal migration”) (e.g., Purde and Rahu, 1979; Coggon et al., 1990; Rosso et al., 1993; Borras et al., 1995; Fascioli et al., 1995). Most often, these uprooted individuals reestablish themselves in an environment that differs dramatically from their place of origin, not only physically (e.g., climate, level of air pollution, intensity of solar radiation) but also socially and culturally. When the numbers of migrants are sufficiently large, they provide a rich source of useful data for researchers in many fields, such as sociology, cultural anthropology, and epidemiology.

CONTRIBUTIONS OF MIGRANT STUDIES TO CANCER RESEARCH Migrant studies contribute to the advancement of knowledge about cancer in many ways (Table 11–1). Because migrants present unique research opportunities that would not have been created intentionally, epidemiologic studies of migrant groups are often referred to as natural experiments. Migrants may undergo extreme changes in exposures to risk factors for disease and may experience disease rates in the host country that are dramatically different from those in the home country. Such rapid and extreme changes in exposure and disease can help identify meaningful risk factors for cancer, especially those related to lifestyle (Prentice and Sheppard, 1989). Migrant studies may also identify populations particularly susceptible to certain types of cancer (Le Marchand, 1999). Whereas migrants experience considerable environmental disruption, their genetic makeup does not change. Thus, these individuals present a unique situation for epidemiologists to distinguish between the contributions of environmental and genetic factors to disease risk. A particular advantage of migrant studies is their potential to provide insights into the period of life when relevant exposures have their greatest impact on cancer risk. In some instances, cancer rates in first generation migrants have been used to infer rates in the home country when such statistics are not available (e.g., Iscovich and Howe, 1998); however, because cancer incidence rates can change dramatically even among first generation migrants (Kolonel et al., 1980), caution must be exercised when using migrant data for this purpose. Finally, the magnitude of change in cancer rates among migrants can suggest the extent of risk reduction achievable through successful public health interventions.

TYPES OF COMPARISONS IN STUDIES OF MIGRANTS Because cancer incidence and mortality rates between the home country and the migrant group in the host country can be dramatically different, this comparison forms the basis of most migrant studies. In this instance, the groups compared are genetically similar but are potentially exposed to different environmental risk factors. Comparisons may also be made between the migrants and the indigenous population (Kolonel et al., 1986; Shimizu et al., 1991; Grulich et al., 1995)

or other migrant groups in the host country (McCredie et al., 1990a; Harding and Rosato, 1999). In this instance, the groups compared are genetically different but are potentially exposed to similar environmental risk factors. Other refinements may include analyses by duration of residence in the host country (McMichael and Giles, 1988; Parkin et al., 1990; Whittemore et al., 1995), by age at migration (Shimizu et al., 1991; Ziegler et al., 1993), and between first generation migrants and their descendants in the host country (Kolonel et al., 1980; Shimizu et al., 1987; Herrinton et al., 1994; Flood et al., 2000). For certain cancer sites, analyses by anatomic region within the organ (Shimizu et al., 1987; Kamineni et al., 1999; Flood et al., 2000) or by histologic subtype (Akazaki and Stemmerman, 1973; Correa et al., 1973; Herrinton et al., 1996) may be possible. When changes in incidence rates are examined in relation to corresponding data on exposure variables, possible etiologic factors may be identified. One of the salient findings of migrant studies of cancer is that the patterns of change are site-specific. In general, incidence rates among migrants for any particular cancer shift in the direction of the prevailing rates in the host country.

CONSIDERATIONS IN THE USE AND INTERPRETATION OF MIGRANT DATA Factors Influencing Incidence Comparisons Taken at face value, changes in cancer incidence observed in a migrant group imply that exposure to risk factors has changed with translocation to the new environment. However, certain limitations of such comparisons must be recognized (Table 11–2). The completeness of case ascertainment, the definitions used to establish cases, the diagnostic coding schemes applied, and the rates of histologic confirmation of diagnoses may all differ between the home and host countries. Pathologists may apply different criteria for establishing diagnoses. Screening practices for cervical, breast, prostate, colorectal, and other cancers may differ, which can have a substantial effect not only on observed incidence rates but on the stage distribution of the tumors. Exposure of migrants to more intensive screening in the host country would lead initially to a higher incidence of disease because of the identification of previously undetected, more advanced cancers (detection bias); subsequent screening would detect proportionally more cancers at earlier stages. Utilization of screening by migrants may also differ from that of the native-born population in the host country (Kagawa-Singer and Pourat, 2000). The computation of incidence rates requires accurate denominator data, the quality of which may differ among countries. An important consideration is whether the sources of the numerator data (usually cancer registries) and the denominator data (usually national censuses) use consistent definitions for the migrant population of interest. Lastly, an important concern in migrant studies is selection bias; that is, migrants may not be representative of the general population of the home country on which comparisons using population-based registries are based. Migrants can differ from the home country’s general population in many ways. They often are more robust (healthy migrant effect) because of visa restrictions on immigrants with certain disease conditions or because individuals in poor health are less inclined to undergo the stresses of relocation and adjustment to a new homeland. They may derive disproportionately from a particular

189

190

PART II: THE MAGNITUDE OF CANCER

Table 11–1. Contributions of Migrant Studies 1. Constitute “natural experiments” in epidemiologic research 2. Yield more extreme variations in exposures than most other studies 3. Permit comparisons between genetically similar groups with widely different cancer incidence rates 4. Suggest possible risk factors for cancer 5. May identify populations genetically susceptible to etiologic factors for particular cancers 6. Offer insights into the contributions of genetics versus the environment to the risk of specific cancers 7. Suggest critical periods of life when risk factors are most influential 8. May be used to infer cancer incidence rates in the home country when they are not otherwise available 9. Indicate the extent of risk reduction achievable by public health programs designed to promote changes in environmental exposures

socioeconomic stratum; most often, migrants reflect a lower tier group seeking economic opportunity in the host country (McCredie et al., 1990a; Grulich et al., 1995), although sometimes they constitute a more educated group seeking better professional opportunities in the host country (Grulich et al., 1995). They may also reflect a religious minority or other persecuted group in the home country (Parkin et al., 1990). They often come from a particular region of a country that has different cancer rates than the country as a whole; this form of selection bias may be overcome if regional population-based cancer registries are available in the home country (Nomura and Hirohata, 1976; Minami et al., 1993). Finally, there may be undefined selection factors that influence the willingness to migrate as well as the levels of exposure to risk factors for cancer. It is sometimes possible to compare incidence rates between first generation migrants and their offspring in the host country. Such comparisons between generations in the host country obviate the issue of migrant selection bias and the problem of intercountry differences regarding diagnostic accuracy and completeness of reporting.

Factors Influencing Mortality Comparisons Most of the issues mentioned above for incidence studies of migrants also pertain to studies based on mortality. However, mortality data present additional limitations and concerns (Table 11–3). Because mortality reflects not only incidence but also treatment, survival, and competing causes of death, it is a poor measure of disease risk, except for highly fatal cancers. The effects of screening programs can also be problematic for mortality studies of migrants. The initial identification of more advanced cancers from screening in the host country could lead to apparently higher cause-specific mortality rates if many of these deaths would have been attributed to other causes in the home country. In addition, access to medical care, the quality of treatment for cancer, and available treatment modalities in the home and host countries may be quite different, thereby influencing disease-specific survival and mortality. Although health care in general might be better in the host country than the home country, migrants may experience cultural barriers (e.g., language) or other obstacles that limit their access, leading to higher disease-specific mortality rates compared

Table 11–2. Considerations for Incidence Comparisons Between Host and Home Countries 1. 2. 3. 4. 5.

Are case ascertainment rates high and comparable? Are case definitions and diagnostic practices comparable? Are pathologic criteria similar? Are rates of histologic confirmation high and comparable? Are screening rates similar? In the host country, do migrants utilize screening at the same rate as native-born residents? 6. Are denominator and numerator data comparable? Are migrants adequately identified in the host country? Is the definition of migrants the same in numerator and denominator sources in the host country? 7. Is there evidence of migrant selection bias?

Table 11–3. Additional Considerations for Mortality Comparisons Between Countries 1. Does the accuracy of cause-of-death assignment differ? Is there an effect of higher screening rates in the host country? 2. Are there differences in access to medical care and in quality and availability of treatment? 3. Do cancer survival rates differ? 4. What is the impact of competing causes of death?

with those for the indigenous population. For these reasons, mortality data can be quite misleading when used to reach conclusions regarding differences in exposure to risk factors in migrant and comparison groups. Because the focus of this chapter is on the contributions of migrant studies to epidemiologic research on cancer etiology, incidence rather than mortality data are emphasized in the sections below.

Ecologic Studies Most migrant studies are descriptive in nature. When correlations with exposure variables are made, the studies are typically ecologic in design. For example, cancer rates in the home and host countries have been examined in relation to per capita intakes based on national food disappearance data (McMichael et al., 1980). Such studies are subject to the ecologic fallacy (Kleinbaum et al., 1982); that is, the exposure information may not be representative of the individuals who migrate. Furthermore, control for potential confounding factors is not possible in such designs. Although case-control and cohort studies have the advantages of detailed information at the individual level and the ability to control for confounders, few case-control studies (Haenszel et al., 1973; Holman and Armstrong, 1984; Terracini et al., 1990; Whittemore et al., 1990; Ziegler et al., 1993) and even fewer cohort studies (Stemmermann et al., 1991; Swerdlow, 1991; Monroe et al., 2003) have been conducted among migrants.

ANALYTIC APPROACHES The prototype analysis for a migrant study involves computing cancer incidence rates for each comparison population for a particular cancer, age-adjusted by the direct method to a standard population. Rate ratios (RRs) of the adjusted incidence rates, using the rates of the host or home country as the reference group, can be computed as summary measures of the relative risk of cancer for migration; 95% confidence intervals (CIs) for the RRs are frequently displayed. Required data for this analysis are incident cancer counts and the size of the population at risk for each comparison population. Analyses are performed separately for men and women, as cancer profiles are distinct by gender. Mortality rates are computed similarly when death information is being used instead of incidence. Depending on the availability of data, the analytic approach in a particular study may vary somewhat from this prototype. Direct standardization has the advantage that rates directly adjusted to the same standard population can be compared across populations and studies. However, in practice, incidence rates across studies have often been adjusted to different standard populations. The direct standardization technique, which requires calculation of age-specific rates, may not be feasible when the number of cases in the migrant group is small. Indirect standardization is then used, wherein the number of observed cancer cases is compared to the number expected based on the age-specific cancer rates of the host (or home) country. Standardized incidence ratios (SIRs) are computed as the ratio of observed to expected cases, along with 95% CIs. Similarly, standardized mortality ratios (SMRs) are computed for death data. Generally, SIRs and SMRs cannot be compared across studies. For instance, an SIR for Italian migrants in the United States is weighted by the age-specific rates of U.S.-born whites, and an SIR for Italian migrants in Australia is weighted by the age-specific rates of Australian-born whites; this difference in weights renders the statistics not strictly comparable.

Migrant Studies On occasion, the population at risk cannot be estimated for the migrant population or the comparison group in the host or home country. This may occur when the numbers of incident cancers are obtained from a series of referral hospitals with an ill-defined catchment area. In this instance, the basis for comparison can be differences in the cancer site distributions in a case-only analysis. Proportional incidence ratios (PIRs) are computed as the ratio of the observed number of cases in the migrant group for the cancer of interest to the expected number based on the proportion of all cancers attributed to the cancer of interest in the reference (host or home) population (Kaldor et al., 1990). PIRs are valid estimates of the relative risk of a particular cancer provided the risk of cancers at other sites is not related to birthplace; this is probably unlikely in most instances. An extension of this approach uses logistic regression to model the incidence of the cancer of interest, taking patients with cancer at other sites as the control group. The resulting odds ratios are equivalent to a PIR but can be adjusted for age and other relevant factors. In instances when sufficient data are available for each case, such as information on risk factors beyond age and place of birth, a loglinear model of cancer rates can be created (Kaldor et al., 1990). Theoretically, all comparisons described for migrant studies could be made simultaneously in such a model. That is, separate rates could be estimated and compared in the model for the migrant population, categorized by length of residence in the host country, their descendants, other host country residents, and home country residents. Rate ratios with confidence limits can be computed from the model. Adjusted rate ratios can be estimated by adding age and other covariates to the model. These models are used infrequently because the required data are seldom available.

PATTERNS OF CHANGE IN CANCER RATES AMONG MIGRANTS When groups migrate from one country to another, their cancer incidence and mortality rates in the host country may increase, decrease, or stay the same as in the home country.

Increased Incidence in the Host Country An increased incidence of cancer in the host country implies that exposure to risk factors is higher in the host country than in the home country or, conversely, that exposure to protective factors is lower. Because generally the host country is more prosperous and has a higher standard of living than the home country (the most common reason for migration), changes in socioeconomic status should also be considered. An obvious area of dramatic change in the lives of migrants relates to diet. Although overall food availability (especially luxury items versus staples) is usually greater in the host country, certain ethnic foods or other traditional recipe ingredients may not be readily available in the new setting. More insights into changes in risk can be obtained by comparing generations of migrants in the host country and examining the effects of age at migration and duration of residence in the host country. Thus, for a cancer with a higher incidence in the host country, one would expect cancer rates in the second generation of migrants to be higher than in the first because the first generation would have experienced the exposure levels of the lower risk home country for a portion of their early lives (illustrated by breast cancer in Figure 11–1). Similarly, one would expect the rates to increase further between the second and third generations (although such data are seldom available) because the second generation would have been more influenced by the traditional cultural practices of the first generation. For any cancer site, if the rates in migrants eventually match those of the indigenous host country population, it suggests that differences in incidence between the home and host countries are related entirely to environmental variables. If a residual difference persists after several generations in the host country, despite evidence of complete acculturation, one might reasonably conclude that a hereditary component (perhaps reflecting differential susceptibility to the effects of environmental risk

191

Figure 11–1. Typical patterns of change in cancer incidence among migrant populations: breast and stomach cancer in Japanese migrants to Hawaii. Rates for the period 1973–1977 were age-adjusted to the world standard population. Japanese rates are from the Osaka Prefecture registry. (Source: Adapted from Kolonel et al., 1980 and Waterhouse et al., 1982.)

factors) contributes to the etiology of that cancer and to the intercountry difference in rates. Duration of residence for first generation migrants would be expected to show a similar relation; that is, the longer the time spent in the host country, the greater is the exposure to risk factors in the new environment and the higher the incidence (illustrated by malignant melanoma in Figure 11–2). Age at migration introduces another variable: the period of life when the influence of risk factors is greatest (illustrated by breast cancer in Figure 11–3). If the influence of a particular risk factor is greater during early life, migration during adulthood (most migrants are working-age adults) would have less effect in the migrating generation and would result in a greater difference between first and second generation migrants in the host country. Clearly, age at migration and duration of residence in the host country are related, and caution is needed if the two variables are examined simultaneously (Parkin, 1992).

Figure 11–2. Effect of the duration of residence on cancer incidence in migrants: cervical cancer and cutaneous malignant melanoma in European migrants to Israel. Israeli-born residents were the reference group. (Source: Adapted from Steinitz et al., 1989.)

192

PART II: THE MAGNITUDE OF CANCER higher rates of the cancer than the indigenous population of the host country.

Decreased Incidence in the Host Country

Figure 11–3. Effect of age at migration on cancer incidence in migrants: breast cancer risk in female migrants to the United States from Asia. U.S.born Asian women were the reference group. (Source: Adapted from Ziegler et al., 1993.)

Although the change in incidence between home and host countries usually shows a gradual increase over several generations, as explained above, in a few instances a more abrupt change can be seen, such that rates comparable to those prevailing in the host country are achieved in the first generation (illustrated by colon cancer in Figure 11–4). This could occur if the migrants rapidly adapt to their new environment, thereby experiencing greater exposure to major risk factors for the cancer and, most importantly, if these risk factors exert their effects primarily later in life (perhaps acting as tumor promoters rather than initiators) (WCRF-AICR, 1997). However, because this phenomenon is seen in migrant groups selectively (e.g., for colon cancer in Japanese migrants to Hawaii and California but not in Hispanic migrants to California) (Kolonel et al., 1980; Shimizu et al., 1987), an additional contributing factor could be greater susceptibility of a particular migrant population to the carcinogenic effects of certain risk factors. This genetic explanation implies that, after full acculturation, subsequent generations of the migrant group might very well exhibit

Figure 11–4. Unusual pattern of change in cancer incidence among migrant populations: colon cancer in Japanese migrants to Hawaii. Rates for the period 1973–1977 were age-adjusted to the world standard population. Japanese rates are from the Osaka Prefecture registry. (Source: Adapted from Kolonel et al., 1980 and Waterhouse et al., 1982.)

Decreased incidence of cancer among migrants in the host country implies that exposures to risk factors are lower in the host country than in the home country, although greater exposure to protective factors could also contribute to the lower rates. For example, exposure to foods that increase cancer risk may be lower in the host country, and exposure to foods that are protective may be higher. Lower cancer rates could also reflect more intensive screening in the host country if the screening program identifies precursor lesions that can be treated definitively (e.g., excision of colonic polyps, management of cervical dysplasia) (see Chapter 70). In contrast to cancers with increased incidence in the host country, for cancers with decreased incidence one would expect rates to be higher in the first generation of migrants than in the second generation (illustrated by stomach cancer in Figure 11–1), to continue to decrease between the second and third generations, to show progressive declines with duration of residence in the host country (illustrated by cervical cancer in Figure 11–2), and to be higher in individuals who migrated at later ages if childhood and/or adolescence was a critical period for the effects of exposure.

Similar Incidence in the Home and Host Countries Although similar incidence rates for a cancer in both countries implies that exposure to environmental risk factors is comparable in the two settings, it also suggests that hereditary factors play a dominant role in the etiology of the cancer. Unless comprehensive data on exposures in the home and host countries were available, one could not easily distinguish between these two explanations.

SPECIFIC CANCER SITES The changing patterns of cancer incidence have been examined in a wide variety of migrant groups and settings. They include migrants to the U.S. from Japan (Buell, 1973; Dunn, 1975; Kolonel et al., 1980; Shimizu et al., 1987; Shimizu et al., 1991; Stemmermann et al., 1991; Stanford et al., 1995; Flood et al., 2000), China (Thomas, 1979; Ziegler et al., 1993; Stanford et al., 1995; Kamineni et al., 1999; Flood et al., 2000), Korea (Gomez et al., 2003), the Philippines (Kolonel, 1985; Ziegler et al., 1993; Stanford et al., 1995; Kamineni et al., 1999; Flood et al., 2000), Vietnam (Le et al., 2002), Latin America (Thomas, 1979; Mack et al., 1985; Shimizu et al., 1987; Shimizu et al., 1991), Puerto Rico (Warshauer et al., 1986; Menendez-Bergad and Blum, 1989; Polednak, 1991), Scandinavia (Moradi et al., 1998), and Italy (Geddes et al., 1991); migrants to Australia from Europe (McMichael and Giles, 1988; McMichael et al., 1989; McCredie et al., 1990a; Geddes et al., 1991; Minami et al., 1993), Asia (McCredie et al., 1990a; Grulich et al., 1995), and the Middle East (McCredie et al., 1990a, 1994); migrants to Canada from Italy (Terracini et al., 1990; Geddes et al., 1993a, 1993b); migrants to the United Kingdom from India and Pakistan (Barker and Baker, 1990; Harding and Rosato, 1999; Warnakulasuriya et al., 1999; Winter et al., 1999), Scotland and Ireland (Harding and Rosato, 1999), Italy (Geddes et al., 1993a, 1993b), and the West Indies (Harding and Rosato, 1999); migrants to Brazil from Japan (Tsugane et al., 1990) and Europe (Bouchardy et al., 1993; Geddes et al., 1993b); migrants to Israel from Europe (Rozen et al., 1981; Katz et al., 1982; Steinitz et al., 1989; Parkin et al., 1990; Iscovich and Parkin, 1997), the former USSR (Iscovich and Howe, 1998), the United States (Rozen et al., 1981; Katz et al., 1982; Parkin et al., 1990), the Middle East/Asia (Parkin et al., 1990; Iscovich and Parkin, 1997) and Africa (Steinitz et al., 1989; Parkin et al., 1990; Iscovich et al., 1993; Iscovich and Parkin, 1997); migrants to Japan from Korea (Ubukata et al., 1987); and Inuit migrants to Denmark from Greenland (Prener et al., 1987). In addition, a few studies examined internal migrations within a country (Purde and Rahu, 1979;

Migrant Studies Hinds and Kolonel, 1980; Zemla, 1984; Rosso et al., 1993; Borras et al., 1995). Some of these studies included multiple cancer sites, whereas others focused on a single site or just a few sites. Other reports were based only on mortality data. They included migrants to the United States from Japan (Buell and Dunn, 1965; Haenszel and Kurihara, 1968; Hirohata, 1974; Nomura and Hirohata, 1976; Thomas and Karagas, 1987), China (Hirohata, 1974; King and Locke, 1980; King et al., 1985; Thomas and Karagas, 1987; Fang et al., 1996), Europe (Staszewski and Haenszel, 1965; Seidman, 1971; Nasca et al., 1981; Thomas and Karagas, 1987; Geddes et al., 1991), Mexico (Thomas and Karagas, 1987), Puerto Rico (Rosenwaike, 1984; Rosenwaike and Shai, 1986; Thomas and Karagas, 1987), and Canada (Nasca et al., 1981); migrants to Canada from Europe (Newman and Spengler, 1984; Balzi et al., 1995; Kliewer and Smith, 1995); Asia (Kliewer and Smith, 1995), the Middle East (Kliewer and Smith, 1995), and Africa (Kliewer and Smith, 1995); migrants to the United Kingdom from Europe (Adelstein et al., 1979; Geddes et al., 1991), Africa (Grulich et al., 1992), and the West Indies (Grulich et al., 1992); migrants to Australia from Europe (Staszewski et al., 1971; McMichael, 1979; McMichael et al., 1980; Dobson and Leeder, 1982; Geddes et al., 1991; Khlat et al., 1992; Kliewer and Smith, 1995), the Middle East (Khlat, 1995; Kliewer and Smith, 1995), Asia (Zhang et al., 1984; McMichael and Giles, 1988; Khlat et al., 1992; Khlat, 1995; Kliewer and Smith, 1995), and Africa (Khlat, 1995; Kliewer and Smith, 1995); migrants to New Zealand from Europe (Cooke and Fraser, 1985); migrants to France from Poland (Tyczynski et al., 1992), Italy (Khlat, 1995), Africa (Bouchardy et al., 1995; Khlat, 1995), and China (Bouchardy et al., 1994); migrants to Japan from Korea (Kim, 1984; Kono et al., 1987) and from China and the United States (Kono et al., 1987); migrants to Brazil from Europe (Bouchardy et al., 1993); migrants to Uruguay from Europe and South America (De Stefani et al., 1990); migrants to Argentina from Europe and South America (Matos et al., 1991); and internal migrations in Italy (Ceppi et al., 1995; Fascioli et al., 1995) and the United Kingdom (Coggon et al., 1990).

Breast Cancer Incidence Most comparisons of breast cancer incidence have entailed migrations from countries with lower rates to countries with higher rates. The results confirmed the generalization that under these circumstances cancer incidence increases in the migrants to levels intermediate between the home and host countries. For example, Japanese migrants to Hawaii and California showed incidence rates that more than doubled in the first (migrating) generation compared with rates in Japan for the same time period (Buell, 1973; Kolonel et al., 1980). Rates increased further in the second generation, although they were still considerably lower (at all ages) than those for white women (Buell, 1973; Dunn, 1975; Kolonel et al., 1980); a similar pattern was reported for Chinese migrants (Stanford et al., 1995). Although migrations from higher to lower risk countries for breast cancer were few, an example was migrants from the United Kingdom and Ireland, with higher incidence rates, to Australia, with lower incidence rates. These migrants experienced an incidence of breast cancer in Australia 20% higher than that of native-born women (McMichael and Giles, 1988). It appears that even internal migration in a country can generate sufficient lifestyle change to influence breast cancer incidence, as shown by Italian women who migrated from the low risk south to the higher risk north of Italy and experienced increased breast cancer incidence (Rosso et al., 1993). Finally, a few exceptions in which breast cancer incidence rates did not increase in migrants to higher risk host countries have been reported (Kolonel, 1985; Minami et al., 1993; Stanford et al., 1995), possibly reflecting slower acculturation to the new environment. Overall, these findings confirm the important influence of environmental, or exogenous, factors on breast cancer risk. The effect of age at migration on breast cancer risk was shown in Japanese migrants to California (Shimizu et al., 1991), who experienced a higher breast cancer incidence in the United States than in Japan. Among women who migrated as young adults, the incidence

193

was substantially higher than in those who migrated at later ages. Similarly, Asian (Japanese, Chinese, Filipino) women who migrated to the United States after age 36 had a lower risk of breast cancer than women who migrated before that age, suggesting that early life exposures have a lasting effect on breast cancer risk (Ziegler et al., 1993). An increase in risk with longer duration of residence in the adopted country was demonstrated in Jewish migrants to Israel from Europe and North Africa, although not from the Middle East/Asia (Parkin et al., 1990).

Mortality Many studies have examined mortality data, often because incidence data were not available for the group of interest. As discussed earlier, because of the impact of mammography screening, treatment, and competing causes of death, the mortality data present problems of interpretation. In a study of migrants to Australia and Canada from countries with higher breast cancer mortality rates than these two host countries, the investigators tended to find lower mortality rates than in the home country, whereas migrants from countries with lower mortality rates than Australia and Canada tended to show an opposite trend (Kliewer and Smith, 1995). Breast cancer mortality rates in Polish migrants to the United States (Staszewski and Haenszel, 1965), England and Wales (Adelstein et al., 1979), and Australia (Staszewski et al., 1971) were higher than those in Poland and similar to those of the native-born populations, whereas they were intermediate between the home and host countries in France (Tyczynski et al., 1992). In the latter study, the authors found that after adjusting for place of residence, age, and social status the mortality rate was similar to that in the home, rather than the host, country.

Explanatory Factors Few studies have assessed risk factors for breast cancer in relation to migration. Age at menarche and at first full-term pregnancy differ between the Japanese in Japan, first generation migrants to the United States, and second generation Japanese women born in the United States, accounting in part for the differences in breast cancer risk among these groups (Wu et al., 1996). Diet influences age at menarche and menopause and may alter breast cancer risk through other mechanisms as well (WCRF-AICR, 1997). Total fat, animal fat, and meat intake were higher among second generation versus first generation Japanese women in Hawaii, consistent with the higher incidence in the second generation women (Hankin et al., 1983); intake of these same dietary components was also higher in Japanese breast cancer patients compared with controls (Kolonel et al., 1986). In Australia, the breast cancer incidence rates among migrants from Hong Kong, Singapore, India, and Sri Lanka were high (similar to those for nativeborn women), which is attributed to a combination of higher socioeconomic status in the country of origin (selection bias) and reproductive behaviors in the adopted country associated with higher breast cancer risk (e.g., later age at first birth) (Grulich et al., 1995). In contrast, the incidence rates among women from Vietnam, China, and the Philippines were lower than in native-born women, consistent with their high fertility rates in Australia. Data on migrant Japanese suggest that environmental factors may also influence the characteristics of breast cancer tumors. In an autopsy study, the prevalence of ductal hyperplasia was lower among Japanese women in Japan and among first generation migrants in Hawaii compared with second generation women (Stemmermann, 1991). Survival of second generation Japanese women with breast cancer in Hawaii was higher than that of first generation women after adjusting for tumor stage, age, and year of diagnosis (Le Marchand et al., 1985), possibly indicating a difference in the biology of the tumors.

Prostate Cancer Incidence Migrations have primarily been to host countries with prostate cancer incidence rates that were higher than in the migrants’ home countries. Thus, the rates among migrants were nearly always increased and generally were intermediate between the home and host countries (Dunn,

194

PART II: THE MAGNITUDE OF CANCER

1975; Tsugane et al., 1990; Minami et al., 1993; Winter et al., 1999; Le et al., 2002; Gomez et al., 2003). Some interesting exceptions to this pattern have been reported. In California (Los Angeles), the incidence among Japanese and Spanishsurnamed migrants was much higher than in the home countries but was similar to that of U.S.-born residents, suggesting the importance of environmental factors acting relatively late in life (Shimizu et al., 1991). In Australia, the incidence of prostate cancer for migrants from certain Asian countries (Philippines, Indonesia, Hong Kong, Malaysia/Singapore) was already similar to that of native-born Australians within a couple of decades (Grulich et al., 1995); however, this was not seen for migrants from other Asian countries (Vietnam, India/Sri Lanka, China/Taiwan). This difference by country of origin could not be readily explained, although the migrant groups with the higher rates tended to be of higher socioeconomic standing, not generally a risk factor for prostate cancer, however (see Chapter 59). The risk of prostate cancer among migrants to Brazil from Italy, Spain, and Portugal was similar to that of the native-born white population of Brazil after adjusting for several variables (age, calendar period, social class, civil status, extent of histologic verification, proportion of deathcertificate-only cases) (Bouchardy et al., 1993). Lastly, Puerto Ricans in New York City had prostate cancer incidence rates much higher than those in Puerto Rico and almost equal to those of the United States overall (Polednak, 1991). Both selection and detection bias may play an important role in these exceptions. Prostate cancer incidence was compared between first and second generation Japanese migrants in Hawaii (Kolonel et al., 1980). The rate was more than 10-fold higher in the first generation migrants compared with that in Japan but was not further increased in the second generation; in both generations, the rates were much lower than among whites in Hawaii. The high rates in the first relative to the second generation may indicate detection bias, reflecting more intense screening in Hawaii than in Japan and the identification of previously undiagnosed cases. The effect of longer duration of residence in the host country in increasing the risk of prostate cancer was demonstrated in Asian migrants to the United States (Whittemore et al., 1995) and in Jewish migrants from Europe to Israel (Parkin et al., 1990). Unlike breast cancer, prostate cancer showed little effect for age at migration in a study of Japanese and Hispanic migrants to California (Shimizu et al., 1991). In a comparative autopsy study of latent cancer of the prostate among Japanese men in Japan and Hawaii (Akazaki and Stemmerman, 1973), the investigators found a similar overall ageadjusted prevalence of tumors in the two groups but a higher proportion of the proliferative (in contrast to nonproliferative) type of latent cancer in the migrants. This finding suggests that the relevant environmental exposures in Hawaii may act as tumor promoters.

Mortality Mortality data on prostate cancer present the same problems as breast cancer data, as this also is an organ site subject to screening. Although screening was primarily by digital rectal examination prior to the era of prostate-specific antigen (PSA) testing, the frequency and regularity of physician visits could have contributed to significant intercountry differences in stage at diagnosis and thus to mortality from the disease. Most migrants studied were from countries with lower prostate cancer mortality rates than the host countries, and for the most part their rates in the host countries were intermediate (e.g., Staszewski and Haenszel, 1965; Haenszel and Kurihara, 1968; De Stefani et al., 1990; Hanley et al., 1995). An exception to this general pattern was found in England and Wales, where migrants from West Africa and the Caribbean had increased prostate cancer mortality compared with native-born men (Grulich et al., 1992); this is discussed further below. The difficulty with drawing inferences about disease risk based on mortality data was illustrated by a survival study in Sweden for the period 1979–1985 (Nilsson et al., 1997). The investigators found that prostate cancer mortality among Estonian migrants in Sweden was similar to that of the total Swedish population and much lower than in Estonia. The better survival in Sweden was attributed to the delayed diagnosis and inferior treatment in Estonia, rather than to differences in exposure to risk factors.

Explanatory Factors Few studies have attempted to explain differences in prostate cancer rates between migrant populations in their host and home countries or between migrants and other groups in their adopted country. In a multicentered case-control study of Asian migrants to the United States, the risk of prostate cancer for foreign-born men who had resided in the United States for at least 25 years was more than twice that of men with a shorter duration of residence after adjusting for age, education, and dietary fat intake (Whittemore et al., 1995). The lower mortality due to prostate cancer among migrants to France from Italy and the Middle East compared with local-born men was speculated to be due to their Mediterranean diet, but no supportive data were available (Khlat, 1995). Finally, the observation of higher prostate cancer mortality in England and Wales among migrant men from West Africa compared with native-born men (Grulich et al., 1992) suggests a possible genetic susceptibility to this cancer in men of West African origin. This is supported by the fact that migrant men from the Caribbean, also of West African origin, showed the same high mortality in England and Wales, whereas migrant men from East Africa, who were mostly of Asian origin, did not (Grulich et al., 1992). Notably, African American men originated in West Africa as well and have the highest incidence of prostate cancer in the world (Parkin et al., 2002).

Colon and Rectal Cancer Incidence Findings on colon and rectal cancers as separate entities are presented first, followed by findings on colorectal cancer as a single site, as many studies did not distinguish between the two.

Colon. Most migrant populations moved from low risk to high risk areas for colon cancer and experienced substantial increases in their incidence rates. In contrast to other major cancer sites, the colon cancer rates of first generation migrants tended to be much closer to the rates of the adopted country and, in a few instances, exhibited a complete transition. As discussed earlier, because migration usually occurs during adulthood, it implies that an exposure later in life critically affects the risk of this cancer; it also implies that in migrant groups whose rates undergo a complete transition to the host country rates acculturation occurs rather rapidly and/or the migrant population has a greater genetic susceptibility to the effects of the pertinent exposure. One group displaying the complete transition—Japanese migrants to the United States—has been studied extensively. In a study of Japanese migrants to Hawaii, the incidence of colon cancer in the home country, Japan, was 25% that of whites in Hawaii, whereas the incidence for Japanese migrants (first generation) and their descendants was the same as, or slightly higher than, that of whites (see Figure 11–4) (Kolonel et al., 1980). The same pattern of convergence was seen in Japanese migrants to Los Angeles County, California; examination of the findings by anatomic subsite showed that the incidence of cancer of the upper colon was similar for Japanese and whites, whereas the incidence of cancer of the sigmoid colon was twofold higher in the Japanese (Shimizu et al., 1987). The increased rates in migrants were unlikely to be explained by a greater intensity of screening in the United States, as an autopsy study found that Japanese in Hawaii had a much higher prevalence of (asymptomatic) adenomatous polyps than Japanese in Japan (Stemmermann and Yatani, 1973). Confirming the importance of the host country rates on the experience of migrants, the incidence of colon cancer among Japanese migrants to Brazil, a low risk country (unlike the United States) remained low (Tsugane et al., 1990). Migrants to Israel from the Middle East/Asia showed a slight increase in colon cancer incidence with duration of residence, although their risk remained lower than that of the Israeli-born; in contrast, immigrants from Africa showed no convergence toward the host country rates (Parkin et al., 1990). Rectum. In general, the effect of migration on the risk for rectal cancer was similar to that for colon cancer. Most migrant groups

Migrant Studies

195

studied moved from low risk to higher risk areas for rectal cancer, and they demonstrated the usual convergence in incidence toward that of the host country. Among Japanese migrants to the United States, the rates of rectal cancer, like colon cancer, increased dramatically; but in this instance the rates for the first generation far exceeded those of U.S. whites. In a study of Japanese migrants to Hawaii, the incidence of rectal cancer among men in the home country, Japan, was 50% that of Hawaii whites, whereas the incidence in the migrants themselves and their Hawaii-born descendants was about 2.0 and 1.5 times that of whites, respectively; among females, the rates in the migrants converged toward, but did not exceed, those of whites (Kolonel et al., 1980). Male Japanese migrants to Los Angeles County showed the same pattern relative to whites as was seen in Hawaii (Shimizu et al., 1987). Similar to colon cancer, rectal cancer rates remained low among Japanese migrants to low risk Brazil (Tsugane et al., 1990). Migrants to Israel from Europe showed increasing rates of rectal cancer with longer duration of residence in the host country, exceeding the rates of the local-born Israelis after 30 years or more (Parkin et al., 1990).

Japanese migrants in Hawaii (Haenszel et al., 1973; Kolonel et al., 1980; McMichael and Giles, 1988; Minami et al., 1993), although exceptions were noted (Tsugane et al., 1990). Positive correlations with alcohol intake, especially beer, and inverse associations with fiber and vitamin C intake (McMichael et al., 1980; Rozen et al., 1981) were also reported. The association with meat intake in Japanese migrants, especially their preference for well done meat (Le Marchand et al., 2002b), may help explain their extraordinarily high colorectal cancer rates (higher than whites in later generations). Well done meat contains significant concentrations of potentially carcinogenic heterocyclic amines (HAAs) and has been associated with colorectal cancer risk (WCRF-AICR, 1997). Metabolism of HAAs is controlled in part by the polymorphic N-acetyltransferase-2 (NAT2) gene, and Japanese are more likely to have the rapid form of the enzyme product of this gene than are whites and other ethnic groups (Le Marchand et al., 2002b).

Colorectum. The studies of colorectal cancer as a single site supported the above findings. The general trend was for the rates of migrants to move to an intermediate level between the risks of the home and host countries, but several groups also exhibited a complete transition in risk to that of the host country. For example, first generation migrants to the United States and Australia from the low risk countries of Japan, Poland, and Scandinavia all showed colorectal cancer rates similar to those of the native-born populations (Kune et al., 1986; Moradi et al., 1998; Flood et al., 2000). An examination of descendants of migrants in the western United States showed that the rates for Japanese exceeded those of whites, whereas the rates for Chinese and Filipinos did not (Flood et al., 2000). In addition, among whites, Chinese, and Filipinos, the distribution by anatomic subsite in the large bowel was similar, whereas among Japanese the risks were higher for the proximal and distal colon and particularly for the rectum and rectosigmoid junction (Flood et al., 2000). An example of a migrant population experiencing a decrease in the incidence of large bowel cancer was provided by European and American migrants to Israel who lived on a kibbutz; the colorectal cancer incidence among these migrants was much lower than that in either their home country or in the general Israeli population (Rozen et al., 1981).

Most investigations of stomach cancer in migrants have involved populations moving from high risk areas to low risk areas. The general pattern was for the risk among migrants to decrease toward that of the host country but to remain elevated. Complete transition to the rates of the host country often took two or more generations to complete. The continued elevation in risk among first generation migrants could be accounted for by the lasting effect of an exposure early in life or by incomplete acculturation in the host country, although the former explanation has been favored. In a study of Japanese migrants to Hawaii, the investigators found that, compared to Hawaii whites, the risk of stomach cancer in the home country, Japan, was sevenfold higher, whereas the risk was fourfold higher in Japanese migrants and twofold higher in their descendants (see Fig. 11–1) (Kolonel et al., 1980). In a later calendar period, the elevation in rates among the descendants of high risk migrant groups persisted for Japanese but not for Chinese in the United States (Kamineni et al., 1999). Migration from Japan to Brazil, a medium risk country, also resulted in a decline in stomach cancer incidence to a level intermediate between the home and host countries, although the rates for Japanese migrants to Brazil were more than twofold higher than those for Japanese migrants to the United States, a low risk country (Tsugane et al., 1990). This example underscores the importance of the prevailing level of risk in the host country to the experience of the migrants. An exception to this pattern of convergence of risk was seen in migrants from Puerto Rico to New York City, among whom rates remained at the relatively high levels of the home country (Warshauer et al., 1986; Menendez-Bergad Bergad, 1989; Polednak, 1991). Because Puerto Rico is a U.S. commonwealth, there is free movement between the two, which was suggested as a reason for the lack of convergence in this instance. The effect of duration of residence was examined in migrants to Israel. Whereas those who migrated from Europe and the former U.S.S.R. exhibited a modest decline in risk with longer residence in Israel, those who migrated from the Middle East/Asia and Africa did not (Parkin et al., 1990; Iscovich and Howe, 1998). One study (Correa et al., 1973) found that the difference in risk for stomach cancer between Japanese in Japan and those in Hawaii was limited to the nondiffuse histologic types (intestinal, mixed, other). Another study found that the subsite distribution of cases differed between U.S.-born whites and migrants from Asia (Japan and China): Asian subjects were more likely to have cancer of the pylorus and corpus and less likely to have cancer of the cardia (Kamineni et al., 1999). These findings suggest that the etiology may differ among gastric cancer subtypes, with the nondiffuse and noncardia cancers more subject to change upon migration.

Mortality The findings from mortality studies of colon and rectal cancers largely agreed with those from the incidence studies. There was a shift in rates toward those prevailing in the host country, whether the migrants came from higher risk or lower risk countries. Similarly, duration of residence in the host country correlated with the extent of the changes (McMichael et al., 1980; Geddes et al., 1991). In a study of internal migration in Italy, place of residence was found to be a stronger predictor of colorectal cancer mortality than was place of birth (Fascioli et al., 1995). Interestingly, studies of Polish migrants to several countries were relatively uniform in showing complete transition in their colon or colorectal cancer mortality rates to those of the host countries (Staszewski and Haenszel, 1965; McMichael et al., 1980; Newman and Spengler, 1984; Matos et al., 1991). The consistent convergence in rates among Polish migrants suggests that they, along with the Japanese, may be particularly susceptible to the major risk factors for cancers of the large bowel. It seems unlikely that the mortality increases could be explained entirely by detection bias in the host countries or more accurate assignment of cause of death, as no increases were seen for rectal cancer. Furthermore, in one instance, the mortality results were confirmed with incidence data (Kune et al., 1986).

Stomach Cancer Incidence

Explanatory Factors Several studies have examined exposure data in an effort to explain the rates of colorectal cancer among migrants to the United States, Australia, and other countries. Fat and meat intake were positively correlated with risk in some reports, including a case-control study of

Mortality Most mortality studies have investigated migration from high to low risk areas and found that mortality rates of most migrants moved toward that of the host country, in agreement with studies of incidence.

196

PART II: THE MAGNITUDE OF CANCER

Because treatment of stomach cancer is not particularly effective, mortality rates are more useful indicators of risk for this cancer than for most others. Evidence of continued excess mortality was found among descendants of Japanese immigrants to the United States (Buell and Dunn, 1965; Haenszel and Kurihara, 1968) and of Italian immigrants to Canada (Balzi et al., 1995) but not of Chinese immigrants to the United States (Hanley et al., 1995). The complete transition to the prevailing host country rates in U.S.-born Chinese may reflect a larger number of generations in the United States and a greater degree of acculturation compared with the other migrant groups studied. The rates among migrants to Australia from higher risk European countries decreased with longer residence but remained higher than the nativeborn rates (McMichael et al., 1980). Italians moving to Australia prior to the age of 15 exhibited the low stomach cancer rates of those born in Australia (Balzi et al., 1995), and internal migrants in England and Wales reflected the stomach cancer rates of their county of birth rather than county of residence (Coggon et al., 1990), again pointing to the importance of early-life exposures. Mortality studies also showed that stomach cancer rates among low risk migrants can increase upon migration to a high risk area. The rates of immigrants to Argentina from Paraguay increased to an intermediate level (Matos et al., 1991), as did internal migrants in Italy from low risk to high risk regions (Fascioli et al., 1995).

to a disproportionately high frequency of adenocarcinomas) (Gao et al., 1987). In Hawaii, the incidence of lung cancer was compared between two generations of Japanese migrants. The rate in the first generation, for both males and females, was approximately double the rate in Japan; but the rate in the second generation was lower than in the first generation and not much higher than in Japan (Kolonel et al., 1980).

Explanatory Factors

Explanatory Factors

Several migrant studies included information on dietary exposures to generate hypotheses regarding the etiology of stomach cancer. In Hawaii, dried/salted fish and pickled vegetable consumption were positively correlated, and vitamin C intake was inversely correlated, with the incidence of stomach cancer in first generation Japanese migrants and their descendants (Haenszel et al., 1972; Kolonel et al., 1980). In a comparison of local-born residents of Hawaii with Japanese and white migrants (the latter from the U.S. mainland), higher consumption of rice, pickled vegetables, and carbohydrates was positively associated with stomach cancer risk (Kolonel et al., 1981). A positive correlation between carbohydrate consumption and stomach cancer was also found in a cross-country analysis based on food disappearance data (McMichael et al., 1980). Several observations suggested that an exposure early in life has a lasting effect on the risk of stomach cancer. This supports the etiologic role of Helicobacter pylori because infection with this bacterium is usually acquired during childhood (Malaty et al., 2002); moreover, H. pylori prevalence varies geographically (Graham et al., 1991; Tsugane et al., 1993; Zhang et al., 1996) and by ethnicity (Graham et al., 1988). As noted earlier, the nondiffuse histologic types of stomach cancer were subject to the influences of migration (Correa et al., 1973). This is consistent with the observation that blood type A was linked to the risk for diffuse tumors only (Correa et al., 1973) and led to the hypothesis that diffuse gastric cancer was associated with host factors more than was nondiffuse cancer.

Because lung cancer risk is so strongly related to tobacco exposure in most parts of the world (see Chapter 33), incidence rates in migrants for the most part can be assumed to reflect smoking behavior. The latency period between first exposure to tobacco smoke and the appearance of clinical disease is relatively long, so smoking habits prior to migration would be expected to have a major influence on the observed incidence of lung cancer among first generation migrants in the host country. Other factors, such as differences in the tar content of cigarettes in the home and host countries and increased smoking among migrants due to the stress of migration, could also contribute to differences in risk. For example, the higher incidence of lung cancer among migrants to Australia from the United Kingdom (McMichael and Giles, 1988; McMichael et al., 1989; McCredie et al., 1990a, 1990b) and to England/Wales from Scotland and Ireland (Harding and Rosato, 1999) could be accounted for by the higher prevalence of smoking in the migrants compared with the native-born populations (McMichael and Giles, 1988; Harding and Rosato, 1999). In some instances, the observed differences between lung cancer rates in migrants and local-born residents could not be explained by smoking differences in the host country, but data on smoking prior to migration were not available (McCredie et al., 1990a, 1994; Terracini et al., 1990; Grulich et al., 1995). Differences between migrant and nativeborn populations regarding susceptibility to the carcinogens in tobacco smoke might also contribute to such discrepancies, as ethnic differences in the distribution of variant forms of polymorphic genes involved in the metabolism of tobacco have been reported (Stephens et al., 1994; Le Marchand et al., 2002a). The high incidence of adenocarcinomas of the lung among female Chinese migrants to Australia was not accounted for by smoking (McCredie et al., 1990a; Stephens et al., 1994; Grulich et al., 1995) and was presumably a carryover from their unusually high rates of this cancer in southeast Asia, where exposure to fumes from open stoves used for cooking in poorly ventilated homes has been implicated (Liu et al., 1993).

Lung Cancer Incidence Lung cancer rates were generally higher in the host than the home country, and the incidence in the migrants was frequently intermediate (e.g., Polednak, 1991; Winter et al., 1999; Le et al., 2002; Gomez et al., 2003). Several exceptions, in which higher incidence rates were seen in the migrants than in the native population of the host country, were reported from Australia. For example, the incidence of lung cancer among both male and female migrants from the United Kingdom was higher than among native-born Australians (McMichael and Giles, 1988; McMichael et al., 1989; McCredie et al., 1990a, 1990b), which is not surprising as lung cancer incidence in the United Kingdom exceeded that of the white population of Australia at the time (Parkin et al., 1997). Similarly, lung cancer incidence rates among female Chinese migrants from Asia to Australia were much higher than among the native-born Australian women (McCredie et al., 1990a; Grulich et al., 1995), which was also not unexpected, as Chinese women in Asia have particularly high rates of lung cancer (largely due

Mortality Because routine screening for lung cancer is uncommon and the disease is highly fatal, mortality data on lung cancer show relations similar to those for incidence. For example, Chinese female migrants in Australia had higher lung cancer mortality than local-born Australian women (Zhang et al., 1984), paralleling the incidence findings noted above (McCredie et al., 1990a; Grulich et al., 1995). Similarly, Italian migrants to Canada had lower mortality from lung cancer than other Canadian-born residents (Balzi et al., 1995), as was found in another study based on incidence (Terracini et al., 1990). Finally, the observation of higher lung cancer incidence in first than in second generation Japanese migrants to Hawaii (Kolonel et al., 1980) was reflected in mortality data for first and second generation Japanese and Chinese migrants in the United States (Thomas and Karagas, 1987).

Malignant Melanoma Incidence The study of cutaneous malignant melanoma among migrants has been of particular interest to investigators in Australia and Israel, where the incidence of the disease is high, solar exposure is common, and the immigrant populations are large. In both countries, the incidence of melanoma was higher in the native-born population than in the migrants (Katz et al., 1982; McMichael and Giles, 1988; McCredie et al., 1990a); for example, the incidence in migrants from the United Kingdom to Australia was only half that of native-born Australians,

Migrant Studies though it was significantly higher than in the home country (McMichael and Giles, 1988; McCredie et al., 1990b). However, the incidence in migrants differed by country of origin. Migrants to Australia from the United Kingdom had much higher rates than migrants from Italy and Greece (McMichael and Giles, 1988; McMichael et al., 1989; McCredie et al., 1990a), the Middle East (McCredie et al., 1990a, 1994), or Asia (McCredie et al., 1990a; Grulich et al., 1995); and migrants to Israel from Europe and the United States had much higher rates than those from the Middle East/Asia and Africa (Katz et al., 1982). Differences related to country of origin may explain the observation in Hawaii that white migrants had higher incidence rates than local-born whites, because a large portion of the native-born whites were of Portuguese ancestry (Hinds and Kolonel, 1980). Duration of residence in high risk areas was positively associated with malignant melanoma risk (Katz et al., 1982; Holman and Armstrong, 1984). For example, in a case-control study in Australia (Holman and Armstrong, 1984), the cancer risk for persons who had resided in Australia for 60 years or more was nearly five times greater than for residents of less than 25 years; and the effect was even stronger for the nodular histologic subtype specifically. Age at migration was also examined in this study, and early age at migration (prior to age 10) was determined to be an important risk factor, with a stronger effect than the correlated variable duration of residence (Holman and Armstrong, 1984).

Mortality The results from analyses based on mortality data parallel the findings based on incidence. For example, migrants to Australia and New Zealand from Europe had mortality rates from malignant melanoma that were intermediate between those of the home country and those of native-born white residents in the host country (Dobson and Leeder, 1982; Cooke and Fraser, 1985; Khlat et al., 1992). Among migrants to Australia, mortality rates were lower in those from southern Europe and Asia than in those from the rest of Europe (Khlat et al., 1992). Longer duration of residence (Dobson and Leeder, 1982; Khlat et al., 1992) and earlier age at migration (Cooke and Fraser, 1985; Khlat et al., 1992) were also identified as risk factors. In one study (Khlat et al., 1992), the combination of a lighter complexion and migration by age 15 led to mortality rates similar to those of the native-born.

Explanatory Factors Data from migrant studies support the conclusion that the risk of cutaneous malignant melanoma is associated with the extent of solar exposure (see Chapter 63). Migrant studies also suggest that exposure during childhood and adolescence carries a greater risk than exposure at later ages. Pigmented nevi have been identified as risk factors for this cancer (Grulich et al., 1996; Whiteman et al., 2003), and a higher prevalence of melanocytic nevi was found in second-generation than in first-generation migrants to Israel (Pavlotsky et al., 1997). The finding of lower risk in Australia among southern European migrants compared with those from elsewhere in Europe, and in Israel among migrants from Asia and Africa compared with those from Europe and the United States, indicates that even small degrees of extra skin pigmentation afford relative protection against this form of cancer.

Other Sites A wide spectrum of cancers have been examined in studies of cancer incidence and mortality among migrant populations. Some additional sites of interest are discussed here.

Esophagus Studies of esophageal cancer has investigated migration from high risk to low risk areas and from low risk to high risk areas. Rates for this cancer among migrants have generally moved in the direction of those prevailing in the host country. However, the patterns of change in risk are complicated by geographic variation in the predominant risk factors. In countries with the highest rates for esophageal cancer, such as China and Iran, the etiology of the disease has been attributed to ingestion of pickled and preserved foods contaminated with

197

Aspergillus flavus and consumption of extremely hot drinks; in the West, where disease rates are lower, the major risk factors are cigarette smoking and alcohol use (see Chapter 36). First generation Chinese migrants to the United States exhibited mortality rates that were intermediate between the high rates of China and the low rates of the host country (King and Locke, 1980; Zhang et al., 1984; King et al., 1985). Because survival from this cancer is poor, incidence rates would be expected to show a similar relation. Among men, U.S.-born descendants of Chinese migrants had rates that were equal to the low rates of U.S. white men, and U.S.-born Chinese women had a lower risk than U.S. white women. This pattern was attributed to the reduced exposure to contaminated food among Chinese in the United States together with the low use of tobacco and alcohol by Chinese women (King et al., 1985; Fang et al., 1996). A similar phenomenon was observed for incidence rates among Japanese migrants to Hawaii (Kolonel et al., 1980; Stemmermann et al., 1987). In other groups, the change in the risk for cancer of the esophagus upon migration generally paralleled the use of tobacco and alcohol. For example, migrants to England from India, where chewing of betel quid with tobacco is common, had rates that were intermediate between those of the low risk host and the high risk home populations (Winter et al., 1999). The mostly Muslim migrants from the Middle East to Australia exhibited extremely low risk, corresponding to their avoidance of alcohol (McCredie et al., 1994). Although the underlying cause of the disease varied between the groups, migrants from Iran and India to Israel had declining rates with longer duration of residence in the host country (Parkin et al., 1990).

Liver Most migration has taken place from high risk regions to low risk regions for liver cancer and has resulted in a change in the risk of migrants toward that of the host country. Infection with hepatitis B and C viruses is the major cause of primary hepatocellular carcinoma (see Chapter 39), and indeed individuals migrating from areas where this virus is endemic (e.g., Asia, Middle East, Africa) to Western countries have had a high risk for this cancer. Thus, Chinese, Japanese, and Filipino migrants to the United States had liver cancer incidence rates that were much higher than those of U.S. whites but were less than half the rates in their home countries (Kolonel, 1985; Rosenblatt et al., 1996). The rates among U.S.-born descendants in these Asian groups declined substantially compared to those of the migrant groups, almost equaling those of U.S. whites (Rosenblatt et al., 1996). Similarly, Korean and Vietnamese migrants to the United States had higher rates than the host population (Le et al., 2002; Gomez et al., 2003), as did African migrants to France (Bouchardy et al., 1995), England and Wales (Grulich et al., 1992), and Israel (Iscovich et al., 1993). Other migration patterns corresponded to alcohol use behaviors. For example, women migrating from the Irish Republic to England and Wales had a substantially higher liver cancer rate and a higher prevalence of heavy drinking than did the host population (Harding and Rosato, 1999).

Pancreas Several investigators have studied the effect of migration on the risk of pancreatic cancer. Although the rates in migrants tended to move toward, and often match, those in the adopted country, the absolute changes in rates were not dramatic, as pancreatic cancer incidence is relatively low. In an example of migration from a low risk to a high risk area, the incidence rates of migrants from Mexico to the United States (California) increased to equal or exceed those of whites in the host country (Mack et al., 1985). Migrants from the higher risk British Isles to Australia retained somewhat higher rates than the local-born population, but the rates in the migrants declined with increasing duration in the host country (McMichael et al., 1980, 1989; McMichael and Bonett, 1981). The increased risk in these migrants was related to their prevalence of smoking, a known risk factor for pancreatic cancer (see Chapter 38).

Bladder In general, migration from low risk to high risk areas for bladder cancer resulted in migrant rates that increased toward those in the host

198

PART II: THE MAGNITUDE OF CANCER

country. The shifts in incidence among migrants most likely resulted from changes in smoking behavior, a major risk factor for transitional cell bladder cancer (see Chapter 58). An exception occurred among migrants from Africa to Argentina (Matos et al., 1991) and France (Bouchardy et al., 1995), who had higher mortality rates than the local populations. This was attributed to endemic schistosomiasis in the countries of origin, which has been causally related to the squamous cell histologic type of bladder cancer.

Thyroid Two general patterns have emerged for thyroid cancer upon migration: one where the rates for the home, migrant, and host populations were similar and the other where migrants had a higher risk than either the home or host populations. Because thyroid cancer is usually not fatal and the adopted countries generally had better health care systems than the countries of origin, the latter pattern may reflect in part better detection of undiagnosed cases in the migrants. However, Filipino migrants consistently exhibited an elevation in risk beyond what might normally be expected from detection bias: Migrants to the United States (Kolonel, 1985; Rossing et al., 1995) and to Australia (Grulich et al., 1995) had more than twice the risk of the host populations. In contrast, the risk of U.S.-born Filipinos declined to levels close to those of U.S. whites (Rossing et al., 1995). The high risk in Filipino migrants in Hawaii was associated with increased iodine intake (Kolonel, 1985). Deficiency or extreme excess in iodine intake can cause goiter (WCRF-AICR, 1997), which has been positively associated with thyroid cancer incidence and was present in the Philippines as recently as 1960 (Rossing et al., 1995). Furthermore, a rapid change from an iodine-deficient environment to one of sufficiency may also lead to thyroid malignancy (WCRF-AICR, 1997), which may help explain the increase in risk among Filipino migrants to the United States.

lomavirus as a result of changes in sexual practices in the host country. An exception to the usual trend was seen in migrants from the Indian subcontinent to England (Barker and Baker, 1990; Winter et al., 1999) and Australia (Grulich et al., 1995), who had lower incidence rates for cervical cancer than the host populations as well as their home populations. Selection bias may have accounted for this finding.

Corpus Uteri Migrants tended to move from countries at low risk for cancer of the corpus uteri to countries at high risk. The rates in migrants generally increased to a level intermediate between those of the host and home countries. As an example, the incidence rate among women in Japan was <1% that of whites in Hawaii; the risk increased in Japanese migrants to Hawaii and somewhat more in the second generation but remained considerably lower (less than half) than that of the host population (Kolonel et al., 1980; Stemmermann et al., 1987). Two major risk factors for corpus uteri cancer are obesity and the use of hormone replacement therapy (prior to the late 1970s), both of which are more common in Western countries (see Chapter 53).

Ovary Migration has traditionally occurred between low risk countries with higher parity and high risk countries with lower parity. The risk among migrants almost always increased to a level intermediate between the home and host countries. Mortality due to ovarian cancer among both first generation Chinese migrants and their U.S.-born descendants was 60% lower than in U.S. whites (King and Locke, 1980). In a more recent report, based on incidence data, the lower risk for ovarian cancer in U.S.-born Chinese (and Japanese) women was limited to those who were older (≥50 years); rates among the younger women were similar to those of U.S. whites (Herrinton et al., 1994). Among Japanese migrants to the United States, the increased incidence for this cancer was limited to epithelial tumors (Stemmermann et al., 1987).

Nasopharynx The most striking result in migrant studies of nasopharyngeal cancer (NPC) is the extremely high risk among Chinese immigrants to the United States, Canada, and Australia (Hanley et al., 1995). Mortality rates for Chinese migrants were 7–51 times those of the native-born white populations in their host countries. Mortality rates among U.S.born Chinese declined dramatically but were still twice as high as those of U.S. whites (King et al., 1985). Greenland Inuits who migrated to Denmark also had higher rates of NPC than the local population (Prener et al., 1987). Residents of southeastern China and Inuit populations have the highest rates of NPC in the world, and risk factors that have been identified in these and other populations include the consumption of salted fish (particularly at weaning), cigarette smoking, and infection with the Epstein-Barr virus (see Chapter 31). The continued elevated risk in migrant Chinese after several generations, as illustrated by the long-established Chinese population of Hawaii (Parkin et al., 2002), may reflect continued high exposure to certain NPC risk factors, although it also suggests that this group has a particular genetic susceptibility to the environmental causes of the disease.

Cervix Migration has generally occurred from countries at high risk for cervical cancer to those at lower risk. Almost universally, the risk among migrants has declined to levels intermediate between the home and host countries. Among migrants living in the low risk host country of Israel, longer duration of residence was associated with a lower incidence of cervical cancer (see Fig. 11–2) (Parkin et al., 1990; Iscovich and Howe, 1998). Similarly, younger Korean women in the United States (<50 years of age) had incidence rates comparable to those for whites, whereas older Korean women had higher rates, comparable to those in Korea (Gomez et al., 2003). The incidence in U.S.-born Japanese women was similar to that in U.S. whites and was less than half that of Japanese migrants (Kolonel et al., 1980). The general decrease in risk among migrants has been attributed to increased cytologic screening (Steinitz et al., 1989), leading to definitive treatment of precancerous lesions and possibly to reduced exposure to human papil-

CONCLUSIONS AND FUTURE RESEARCH OPPORTUNITIES Findings on migrant populations have contributed essential information to research on the etiology of cancer. Foremost, they have shown the dominant role of environmental factors in determining cancer risk. Because the direction and magnitude of change in incidence vary by cancer site for each migrant group, these studies support the view that the major risk factors for various cancer sites also differ. In some instances, migrant studies have provided information on critical periods of life when risk factors are most influential. They have also suggested useful etiologic hypotheses, particularly those related to lifestyle, especially diet. The observation that cancer rates in migrant populations move away from those prevailing in the home country toward those prevailing in the host country has become a hallmark finding that is widely cited by cancer researchers. Equally noteworthy is the observation that the patterns of change in migrant populations are site-specific. It is possible that migrant studies have already made their most important contributions to cancer research. Certainly, continued analysis of descriptive data on cancer incidence and mortality in migrant populations is unlikely to add substantially to knowledge about the causation of cancer. Furthermore, widespread screening for cancers of the cervix, breast, prostate, and colorectum in recent years in many developed countries has markedly changed the stage distributions of these cancers and has made intercountry migrant comparisons difficult to interpret without additional information that is often not available. However, novel approaches may yield useful information. For example, two-way migration studies, such as Americans moving to Mexico versus Mexicans moving to the United States could yield highly informative data, but such populations are not easy to identify given that countries tend to be either major sources of migrants or recipients of migrants—but not both. Also useful would be systematic comparisons for particular cancer sites of the effects of migration from low risk to high risk countries with migration from high risk to low

Migrant Studies risk countries. Studies to explore further some of the exceptions identified in the previous sections (e.g., variations in prostate cancer incidence between different Asian migrant groups to Australia or differences in colorectal cancer risk and anatomic location of the lesions between Japanese and Chinese migrant populations in the United States) should be pursued. Few cohort studies have provided data on migrants. Nested in a cohort study, a classic migrant study can be performed that compares incidence or mortality rates between migrant and native-born groups but with adjustment for risk factors and other covariates at the individual level. Finally, the study of gene–environment interactions in migrant studies could be highly productive. The combination of extreme dietary changes in migrants, other differences between migrants and host populations, and variations in the distributions of relevant polymorphic genes involved in the metabolism of exogenous exposures would add a unique element to research in this topical area. References Adelstein AM, Staszewski J, Muir CS. 1979. Cancer mortality in 1970–1972 among Polish-born migrants to England and Wales. Br J Cancer 40:464–475. Akazaki K, Stemmerman GN. 1973. Comparative study of latent carcinoma of the prostate among Japanese in Japan and Hawaii. J Natl Cancer Inst 50:1137–1144. Balzi D, Geddes M, Brancker A, et al. 1995. Cancer mortality in Italian migrants and their offspring in Canada. Cancer Causes Control 6:68–74. Barker RM, Baker MR. 1990. Incidence of cancer in Bradford Asians. J Epidemiol Community Health 44:125–129. Borras JM, Sanchez V, Moreno V, et al. 1995. Cervical cancer: incidence and survival in migrants within Spain. J Epidemiol Community Health 49:153–157. Bouchardy C, Khlat M, Mirra AP, et al. 1993. Cancer risks among European migrants in Sao Paulo, Brazil. Eur J Cancer 29A:1418–1423. Bouchardy C, Parkin DM, Khlat M. 1994. Cancer mortality among Chinese and South-East Asian migrants in France. Int J Cancer 58:638–643. Bouchardy C, Wanner P, Parkin DM. 1995. Cancer mortality among subSaharan African migrants in France. Cancer Causes Control 6:539–544. Buell P. 1973. Changing incidence of breast cancer in Japanese-American women. J Natl Cancer Inst 51:1479–1483. Buell P, Dunn JE. 1965. Cancer mortality among Japanese Issei and Nisei of California. Cancer 18:656–664. Ceppi M, Vercelli M, Decarli A, et al. 1995. The mortality rate of the province of birth as a risk indicator for lung and stomach cancer mortality among Genoa residents born in other Italian provinces. Eur J Cancer 31A:193–197. Coggon D, Osmond C, Barker DJ. 1990. Stomach cancer and migration within England and Wales. Br J Cancer 61:573–574. Cooke KR, Fraser J. 1985. Migration and death from malignant melanoma. Int J Cancer 36:175–178. Correa P, Sasano N, Stemmermann GN, et al. 1973. Pathology of gastric carcinoma in Japanese populations: comparisons between Miyagi Prefecture, Japan, and Hawaii. J Natl Cancer Inst 51:1449–1459. De Stefani E, Parkin DM, Khlat M, et al. 1990. Cancer in migrants to Uruguay. Int J Cancer 46:233–237. Dobson AJ, Leeder SR. 1982. Mortality from malignant melanoma in Australia: effects due to country of birth. Int J Epidemiol 11:207–211. Dunn JE. 1975. Cancer epidemiology in populations of the United States—with emphasis on Hawaii and California—and Japan. Cancer Res 35:3240–3245. Fang J, Madhavan S, Alderman MH. 1996. Cancer mortality of Chinese in New York City 1988–1992. Int J Epidemiol 25:907–912. Fascioli S, Capocaccia R, Mariotti S. 1995. Cancer mortality in migrant populations within Italy. Int J Epidemiol 24:8–18. Flood DM, Weiss NS, Cook LS, et al. 2000. Colorectal cancer incidence in Asian migrants to the United States and their descendants. Cancer Causes Control 11:403–411. Gao YT, Blot WJ, Zheng W, et al. 1987. Lung cancer among Chinese women. Int J Cancer 40:604–609. Geddes M, Balzi D, Buiatti E. 1993a. Nasopharynx cancer in Italian migrants. Cancer Causes Control 4:111–116. Geddes M, Balzi D, Buiatti E, et al. 1991. Cancer in Italian migrants. Cancer Causes Control 2:133–140. Geddes M, Parkin D, Khlat M, Balzi D, et al., editors. 1993b. Cancer in Italian Migrant Populations. Lyon: IARC, pp. 1–292. Gomez SL, Le GM, Clarke CA, et al. 2003. Cancer incidence patterns in Koreans in the US and in Kangwha, South Korea. Cancer Causes Control 14:167–174.

199

Graham DY, Klein PD, Opekun AR, et al. 1988. Epidemiology of Campylobacter pylori infection: ethnic considerations. Scand J Gastroenterol Suppl 142:9–13. Graham DY, Malaty HM, Evans DG, et al. 1991. Epidemiology of Helicobacter pylori in an asymptomatic population in the United States: effect of age, race, and socioeconomic status. Gastroenterology 100:1495–1501. Grulich AE, Bataille V, Swerdlow AJ, et al. 1996. Naevi and pigmentary characteristics as risk factors for melanoma in a high-risk population: a casecontrol study in New South Wales, Australia. Int J Cancer 67:485–491. Grulich AE, McCredie M, Coates M. 1995. Cancer incidence in Asian migrants to New South Wales, Australia. Br J Cancer 71:400–408. Grulich AE, Swerdlow AJ, Head J, et al. 1992. Cancer mortality in African and Caribbean migrants to England and Wales. Br J Cancer 66:905–911. Haenszel W, Berg JW, Segi M, et al. 1973. Large-bowel cancer in Hawaiian Japanese. J Natl Cancer Inst 51:1765–1779. Haenszel W, Kurihara M. 1968. Studies of Japanese migrants. I. Mortality from cancer and other diseases among Japanese in the United States. J Natl Cancer Inst 40:43–68. Haenszel W, Kurihara M, Segi M, et al. 1972. Stomach cancer among Japanese in Hawaii. J Natl Cancer Inst 49:969–988. Hankin JH, Kolonel LN, Yano K, et al. 1983. Epidemiology of diet-related diseases in the Japanese migrant population of Hawaii. Proc Nutr Soc Aust 8:22–40. Hanley AJ, Choi BC, Holowaty EJ. 1995. Cancer mortality among Chinese migrants: a review. Int J Epidemiol 24:255–265. Harding S, Rosato M. 1999. Cancer incidence among first generation Scottish, Irish, West Indian and South Asian migrants living in England and Wales. Ethn Health 4:83–92. Herrinton LJ, Goldoft M, Schwartz SM, et al. 1996. The incidence of nonHodgkin’s lymphoma and its histologic subtypes in Asian migrants to the United States and their descendants. Cancer Causes Control 7:224–230. Herrinton LJ, Stanford JL, Schwartz SM et al. 1994. Ovarian cancer incidence among Asian migrants to the United States and their descendants. J Natl Cancer Inst 86:1336–1339. Hinds MW, Kolonel LN. 1980. Malignant melanoma of the skin in Hawaii, 1960–1977. Cancer 45:811–817. Hirohata T. 1974. Mortality experience of various ethnic groups in Hawaii, with particular reference to the comparison with their mother populations. Jpn J Public Health 21:17–26. Holman CD, Armstrong BK. 1984. Cutaneous malignant melanoma and indicators of total accumulated exposure to the sun: an analysis separating histogenetic types. J Natl Cancer Inst 73:75–82. Iscovich J, Howe GR. 1998. Cancer incidence patterns (1972–91) among migrants from the Soviet Union to Israel. Cancer Causes Control 9:29–36. Iscovich J, Parkin DM. 1997. Risk of cancer in migrants and their descendants in Israel. I. Leukaemias and lymphomas. Int J Cancer 70:649–653. Iscovich J, Steinitz R, Andreev H. 1993. Descriptive epidemiology of malignant tumors in Ethiopian Jews immigrating to Israel, 1984–89. Isr J Med Sci 29:364–367. Kagawa-Singer M, Pourat N. 2000. Asian American and Pacific Islander breast and cervical carcinoma screening rates and healthy people 2000 objectives. Cancer 89:696–705. Kaldor J, Khlat M, Parkin DM, et al. 1990. Log-linear models for cancer risk among migrants. Int J Epidemiol 19:233–239. Kamineni A, Williams MA, Schwartz SM, et al. 1999. The incidence of gastric carcinoma in Asian migrants to the United States and their descendants. Cancer Causes Control 10:77–83. Katz L, Ben-Tuvia S, Steinitz R. 1982. Malignant melanoma of the skin in Israel: effect of migration. In: Magnus K, editor. Trends in Cancer Incidence: Causes and Practical Implications. New York: Hemisphere, pp. 419–426. Khlat M. 1995. Cancer in Mediterranean migrants—based on studies in France and Australia. Cancer Causes Control 6:525–531. Khlat M, Vail A, Parkin M, et al. 1992. Mortality from melanoma in migrants to Australia: variation by age at arrival and duration of stay. Am J Epidemiol 135:1103–1113. Kim YS. 1984. Cancer mortality of the Korean population in Japan 1968–1977. Int J Epidemiol 13:11–14. King H, Li JY, Locke FB, et al. 1985. Patterns of site-specific displacement in cancer mortality among migrants: the Chinese in the United States. Am J Public Health 75:237–242. King H, Locke FB. 1980. Cancer mortality among Chinese in the United States. J Natl Cancer Inst 65:1141–1148. Kleinbaum DG, Kupper LL, Morgenstern H. 1982. Epidemiologic Research: Principles and Quantitative Methods. Belmont, CA: Lifetime Learning Publications, p. 79. Kliewer EV, Smith KR. 1995. Breast cancer mortality among immigrants in Australia and Canada. J Natl Cancer Inst 87:1154–1161.

200

PART II: THE MAGNITUDE OF CANCER

Kolonel L, Hinds MW, Hankin JH. 1980. Cancer patterns among migrant and native-born Japanese in Hawaii in relation to smoking, drinking, and dietary habits. In: Gelboin HV, MacMahon B, Matsushima T, Sugimura T, Takayama S, Takebe H, editors. Genetic and Environmental Factors in Experimental and Human Cancer. Tokyo: Japan Scientific Societies Press, pp. 327–340. Kolonel LN. 1985. Cancer incidence among Filipinos in Hawaii and the Philippines. Natl Cancer Inst Monogr 69:93–98. Kolonel LN, Hankin JH, Nomura AM. 1986. Multiethnic studies of diet, nutrition, and cancer in Hawaii. In: Hayashi Y, Nagao M, Sugimura T, Takayama S, Tomatis L, Wattenberg LW, Wogan GN, editors. Diet, Nutrition and Cancer. Tokyo: Japan Scientific Societies Press, pp. 29–40. Kolonel LN, Nomura AM, Hirohata T, et al. 1981. Association of diet and place of birth with stomach cancer incidence in Hawaii Japanese and Caucasians. Am J Clin Nutr 34:2478–2485. Kono S, Isa AR, Ogimoto I, et al. 1987. Cause-specific mortality among Koreans, Chinese and Americans in Japan, 1973–1982. Int J Epidemiol 16:415–419. Kune S, Kune GA, Watson L. 1986. The Melbourne colorectal cancer study: incidence findings by age, sex, site, migrants and religion. Int J Epidemiol 15:483–493. Le GM, Gomez SL, Clarke CA et al. 2002. Cancer incidence patterns among Vietnamese in the United States and Ha Noi, Vietnam. Int J Cancer 102:412–417. Le Marchand L. 1999. Combined influence of genetic and dietary factors on colorectal cancer incidence in Japanese Americans. J Natl Cancer Inst Monogr 26:101–105. Le Marchand L, Donlon T, Lum-Jones A, et al. 2002a. Association of the hOGG1 Ser326Cys polymorphism with lung cancer risk. Cancer Epidemiol Biomarkers Prev 11:409–412. Le Marchand L, Hankin JH, Pierce LM, et al. 2002b. Well-done red meat, metabolic phenotypes and colorectal cancer in Hawaii. Mutat Res 506–507:205–214. Le Marchand L, Kolonel LN, Nomura AM. 1985. Breast cancer survival among Hawaii Japanese and Caucasian women: ten-year rates and survival by place of birth. Am J Epidemiol 122:571–578. Liu Q, Sasco AJ, Riboli E, et al. 1993. Indoor air pollution and lung cancer in Guangzhou, People’s Republic of China. Am J Epidemiol 137:145–154. Mack TM, Walker A, Mack W, et al. 1985. Cancer in Hispanics in Los Angeles County. Natl Cancer Inst Monogr 69:99–104. Malaty HM, El-Kasabany A, Graham DY, et al. 2002. Age at acquisition of Helicobacter pylori infection: a follow-up study from infancy to adulthood. Lancet 359:931–935. Matos EL, Khlat M, Loria DI, et al. 1991. Cancer in migrants to Argentina. Int J Cancer 49:805–811. McCredie M, Coates M, Grulich A. 1994. Cancer incidence in migrants to New South Wales (Australia) from the Middle East, 1972–91. Cancer Causes Control 5:414–421. McCredie M, Coates MS, Ford JM. 1990a. Cancer incidence in migrants to New South Wales. Int J Cancer 46:228–232. McCredie M, Coates MS, Ford JM. 1990b. Cancer incidence in migrants to New South Wales from England, Wales, Scotland and Ireland. Br J Cancer 62:992–995. McMichael AJ. 1979. Alimentary tract cancer in Australia in relation to diet and alcohol. Nutr Cancer 1:82–88. McMichael AJ, Bonett A. 1981. Cancer profiles of British and southernEuropean migrants. Exploring South Australia’s cancer registry data. Med J Aust 1:229–232. McMichael AJ, Giles GG. 1988. Cancer in migrants to Australia: extending the descriptive epidemiological data. Cancer Res 48:751–756. McMichael AJ, Bonett A, Roder D. 1989. Cancer incidence among migrant populations in South Australia. Med J Aust 150:417–420. McMichael AJ, McCall MG, Hartshorne JM, et al. 1980. Patterns of gastrointestinal cancer in European migrants to Australia: the role of dietary change. Int J Cancer 25:431–437. Menendez-Bergad B, Blum S. 1989. Stomach cancer in a native and migrant population in Puerto Rico and New York City, 1975 to 1979. Biol Assoc Med P R 81:95–98. Minami Y, Staples MP, Giles GG. 1993. The incidence of colon, breast and prostate cancer in Italian migrants to Victoria, Australia. Eur J Cancer 29A:1735–1740. Monroe KR, Hankin JH, Pike MC, et al. 2003. Correlation of dietary intake and colorectal cancer incidence among Mexican-American migrants: the multiethnic cohort study. Nutr Cancer 45:133–147. Moradi T, Delfino RJ, Bergstrom SR et al. 1998. Cancer risk among Scandinavian immigrants in the US and Scandinavian residents compared with US whites, 1973–89. Eur J Cancer Prev 7:117–125.

Nasca PC, Greenwald P, Burnett WS, et al. 1981. Cancer among the foreignborn in New York State. Cancer 48:2323–2328. Newman AM, Spengler RF. 1984. Cancer mortality among immigrant populations in Ontario, 1969 through 1973. Can Med Assoc J 130:399–405. Nilsson B, Gustavson-Kadaka E, Hakulinen T, et al. 1997. Cancer survival in Estonian migrants to Sweden. J Epidemiol Community Health 51: 418–423. Nomura A, Hirohata T. 1976. Cancer mortality among Japanese in Hawaii: comparison of observed and expected rates based on prefecture-of-origin in Japan. Hawaii Med J 35:293–297. Parkin DM, Whelan SL, Ferlay J, Raymond L, et al, editors. 1997. Cancer Incidence in Five Continents, Vol. VII. IARC Scientific Publ. 143. Lyon: IARC. Parkin D, Whelan SL., Ferlay J, et al., editors. 2002. Cancer Incidence in Five Continents, Vol. VIII. IARC Scientific Publ. 155. Lyon: IARC. Parkin DM. 1992. Studies of cancer in migrant populations: methods and interpretation. Rev Epidemiol Sante Publique 40:410–424. Parkin DM, Steinitz R, Khlat M, et al. 1990. Cancer in Jewish migrants to Israel. Int J Cancer 45:614–621. Pavlotsky F, Azizi E, Gurvich R, et al. 1997. Prevalence of melanocytic nevi and freckles in young Israeli males. Correlation with melanoma incidence in Jewish migrants: demographic and host factors. Am J Epidemiol 146:78–86. Polednak AP. 1991. Cancer incidence in the Puerto Rican-born population of Long Island, New York. Am J Public Health 81:1405–1407. Prener A, Nielsen NH, Hansen JP, et al. 1987. Cancer pattern among Greenlandic Inuit migrants in Denmark, 1968–1982. Br J Cancer 56: 679–684. Prentice RL, Sheppard L. 1989. Validity of international, time trend, and migrant studies of dietary factors and disease risk. Prev Med 18:167–179. Purde M, Rahu M. 1979. Cancer patterns in the oil shale area of the Estonian S.S.R. Environ Health Perspect 30:209–210. Rosenblatt KA, Weiss NS, Schwartz SM. 1996. Liver cancer in Asian migrants to the United States and their descendants. Cancer Causes Control 7:345–350. Rosenwaike I. 1984. Cancer mortality among Puerto Rican-born residents in New York City. Am J Epidemiol 119:177–185. Rosenwaike I, Shai D. 1986. Trends in cancer mortality among Puerto Rican-born migrants to New York City. Int J Epidemiol 15:30–35. Rossing MA, Schwartz SM, Weiss NS. 1995. Thyroid cancer incidence in Asian migrants to the United States and their descendants. Cancer Causes Control 6:439–444. Rosso S, Patriarca S, Vicari P, et al. 1993. Cancer incidence in Turin: the effect of migration. Tumori 79:304–310. Rozen P, Hellerstein SM, Horwitz C. 1981. The low incidence of colorectal cancer in a “high-risk” population: its correlation with dietary habits. Cancer 48:2692–2695. Seidman H. 1971. Cancer mortality in New York City for country-of-birth, religious, and socioeconomic groups. Environ Res 4:390–429. Shimizu H, Mack TM, Ross RK, et al. 1987. Cancer of the gastrointestinal tract among Japanese and white immigrants in Los Angeles County. J Natl Cancer Inst 78:223–228. Shimizu H, Ross RK, Bernstein L, et al. 1991. Cancers of the prostate and breast among Japanese and white immigrants in Los Angeles County. Br J Cancer 63:963–966. Stanford JL, Herrinton LJ, Schwartz SM, et al. 1995. Breast cancer incidence in Asian migrants to the United States and their descendants. Epidemiology 6:181–183. Staszewski J, Haenszel W. 1965. Cancer mortality among the Polish-born in the United States. J Natl Cancer Inst 35:291–297. Staszewski J, McCall MG, Stenhouse NS. 1971. Cancer mortality in 1962–66 among Polish migrants to Australia. Br J Cancer 25:599–610. Steinitz R, Parkin DM, Young JL, Bieber CA, et al, editors. 1989. Cancer Incidence in Jewish Migrants to Israel 1961–1981. IARC Scientific Publ. 98. Lyon: IARC. Stemmermann GN. 1991. The pathology of breast cancer in Japanese women compared to other ethnic groups: a review. Breast Cancer Res Treat 18(Suppl 1):S67–S72. Stemmermann GN, Yatani R. 1973. Diverticulosis and polyps of the large intestine: a necropsy study of Hawaii Japanese. Cancer 31:1260–1270. Stemmermann GN, Nomura AM, Chyou PH, et al. 1991. Cancer incidence in Hawaiian Japanese: migrants from Okinawa compared with those from other prefectures. Jpn J Cancer Res 82:1366–1370. Stemmermann G, Nomura AM, Kolonel LN. 1987. Cancer among JapaneseAmericans in Hawaii. Gann Monogr Cancer Res 33:99–108. Stephens EA, Taylor JA, Kaplan N, et al. 1994. Ethnic variation in the CYP2E1 gene: polymorphism analysis of 695 African-Americans, EuropeanAmericans and Taiwanese. Pharmacogenetics 4:185–192.

Migrant Studies Swerdlow A. 1991. Mortality and cancer incidence in Vietnamese refugees in England and Wales: a follow-up study. Int J Epidemiol 20: 13–19. Terracini B, Siemiatycki J, Richardson L. 1990. Cancer incidence and risk factors among Montreal residents of Italian origin. Int J Epidemiol 19:491–497. Thomas DB. 1979. Epidemiologic studies of cancer in minority groups in the western United States. Natl Cancer Inst Monogr 53:103–113. Thomas DB, Karagas MR. 1987. Cancer in first and second generation Americans. Cancer Res 47:5771–5776. Tsugane S, De Souza JM, Costa ML Jr, et al. 1990. Cancer incidence rates among Japanese immigrants in the city of Sao Paulo, Brazil, 1969–78. Cancer Causes Control 1:189–193. Tsugane S, Kabuto M, Imai H, et al. 1993. Helicobacter pylori, dietary factors, and atrophic gastritis in five Japanese populations with different gastric cancer mortality. Cancer Causes Control 4:297–305. Tyczynski J, Parkin D, Zatonski W, et al. 1992. Cancer mortality among Polish migrants to France. Bull Cancer 79:789–800. Ubukata T, Oshima A, Morinaga K, et al. 1987. Cancer patterns among Koreans in Japan, Koreans in Korea and Japanese in Japan in relation to life style factors. Jpn J Cancer Res 78:437–446. Warnakulasuriya KA, Johnson NW, Linklater KM, et al. 1999. Cancer of mouth, pharynx and nasopharynx in Asian and Chinese immigrants resident in Thames regions. Oral Oncol 35:471–475. Warshauer ME, Silverman DT, Schottenfeld D, et al. 1986. Stomach and colorectal cancers in Puerto Rican-born residents of New York City. J Natl Cancer Inst 76:591–595.

201

Waterhouse J, Muir C, Shanmugaratnam K, Powell J, editors. 1982. Cancer Incidence in Five Continents, Vol. IV. IARC Scientific Publ. 42. Lyon: IARC. WCRF-AICR. 1997. Food, Nutrition and the Prevention of Cancer: A Global Perspective. Washington, DC: American Institute for Cancer Research. Whiteman DC, Watt P, Purdie DM, et al. 2003. Melanocytic nevi, solar keratoses, and divergent pathways to cutaneous melanoma. J Natl Cancer Inst 95:806–812. Whittemore AS, Kolonel LN, Wu AH, et al. 1995. Prostate cancer in relation to diet, physical activity, and body size in blacks, whites, and Asians in the United States and Canada. J Natl Cancer Inst 87:652–661. Whittemore AS, Wu-Williams AH, Lee M, et al. 1990. Diet, physical activity, and colorectal cancer among Chinese in North America and China. J Natl Cancer Inst 82:915–926. Winter H, Cheng KK, Cummins C, et al. 1999. Cancer incidence in the south Asian population of England (1990–92). Br J Cancer 79:645–654. Wu AH, Ziegler RG, Pike MC, et al. 1996. Menstrual and reproductive factors and risk of breast cancer in Asian-Americans. Br J Cancer 73:680–686. Zemla B. 1984. The role of selected dietary elements in breast cancer risk among native and migrant populations in Poland. Nutr Cancer 6:187–195. Zhang L, Blot WJ, You WC, et al. 1996. Helicobacter pylori antibodies in relation to precancerous gastric lesions in a high-risk Chinese population. Cancer Epidemiol Biomarkers Prev 5:627–630. Zhang YQ, MacLennan R, Berry G. 1984. Mortality of Chinese in New South Wales, 1969–1978. Int J Epidemiol 13:188–192. Ziegler RG, Hoover RN, Pike MC, et al. 1993. Migration patterns and breast cancer risk in Asian-American women. J Natl Cancer Inst 85:1819–1827.

12

Economic Impact of Cancer in the United States MARTIN L. BROWN AND K. ROBIN YABROFF

A

n estimated 9.5 million Americans were alive in 2000 who have a history of cancer. Of this total, almost 5.9 million were diagnosed 5 years or more ago (SEER, 2003). Currently, more than 1 million people a year are diagnosed as having cancer; and based on an aging population and increases in population size this number is expected to double by the year 2050 (Edwards, 2002). Based on increased incidence alone, the expected economic burden of cancer is projected to increase. Improvements in early detection and treatment may lead to improvements in survival and reduced cancer mortality, but these developments will also result in increased cancer prevalence.

WHY MEASURE THE ECONOMIC BURDEN OF CANCER? Measuring the economic burden of cancer is important at many levels: for medical resource allocation, reimbursement decisions, and evaluation of specific programs throughout the course of cancer care, from prevention and early detection to treatment, surviviorship, and end of life. Estimates of the economic burden of cancer have been used to inform decision-making about the national allocation of research funding (NIH, 1997; Lichtenberg, 2001; Murphy, 2003). Descriptive studies of the various components of the economic burden of cancer also serve as an initial step in elucidating relevant domains in costeffectiveness analysis of cancer control interventions and identifying priorities and resource needs when planning cancer control strategies and programs.

WHAT IS THE ECONOMIC BURDEN OF CANCER? Illness and disease create an economic burden for the patient, family and friends, and society. Cancer and its treatment may result in pain and suffering, morbidity and reduced quality of life, premature mortality, and financial losses for the patient and family. Family and friends may also suffer emotional trauma, grief, and financial loss. In addition to the costs of medical care, all of these factors affect society as a whole. The economic burden of cancer includes the loss of economic resources and opportunities from the perspective of the individual, the family, and society—whether pecuniary or nonpecuniary—associated with the occurrence of cancer and its treatment. The economic burden of disease is measured by cost, the monetary valuation of resources used to treat disease, or the loss of economic opportunities related to disease occurence and treatment. Three categories of cost domains can be identified: direct costs, resulting from the use of resources for medical care; indirect costs, resulting from the loss of economic resources and opportunites associated with disease and treatment; and psychosocial costs, such as pain and suffering. Examples of these categories of costs are listed in Table 12–1 and described in greater detail in this section.

STUDIES OF THE NATIONAL COST OF CANCER Over the last few decades numerous studies have been conducted on the national cost of cancer to the United States. In appendix 12–B

202

we summarize these studies to provide information on the cancer site and cost domains included in the study and the nature of the population and data sources used to conduct the study.

Direct Costs Direct costs include the monetary value of resources used for medical care in the prevention, diagnosis, and treatment of disease and for continuing care or surveillance, rehabilitation, and terminal care. The distinguishing characteristic of a direct cost is that it represents resource utilization associated with a direct monetary payment. These payments are usually recorded in the financial records of a health care provider, a health care insurance system, or an individual household. Such costs includes medical resources provided under public or private insurance systems as well as individual out-of-pocket expenditures. These direct costs for medical care include hospitalization, outpatient clinical care, nursing home care, home health care, and services of primary physicians and specialists and other health practitioners. Drugs and radiation therapy used for treatment, rehabilitation counseling, and other rehabilitation costs (e.g., prostheses or speech devices) related to overcoming impairments resulting from disease are also included in the direct cost category. In addition to costs for the provision of medical care, other nonmedical direct costs borne by patients and other individuals include those associated with transportation to health providers and costs of relocating temporarily or permanently to improve access to specific treatments or facilities. Illness can force a family to incur expenses when caring and providing for the sick member of the family, including extra expenditures for routine household tasks; special diets or clothing; items for rehabilitation and comfort such as exercycles, vaporizers, humidifiers, and dehumidifiers; alterations of property, such as elevators and other special housing facilities; and vocational, social, and family counseling services. Expenditures for retraining or reeducation and interest lost on withdrawal of savings or interest charges on funds borrowed to pay illness-related expenses are also considered direct costs. As shown in Table 12–2, hospital care, estimated to be about $40 billion in 2002, is the largest component of direct medical expenditures for cancer patients. The next largest component is related to the professional services of physicians and other health care providers. Home health care, nursing home care, drugs, and medical durable goods each account for less than 5% of cancer-related medical care expenditures. Note that Table 12–2 does not include all the components of direct cost enumerated above. In addition, the method used to obtain the national estimates shown in Table 12–2 do not account for costs related to cancer prevention and screening or related diagnostic services. Based on the number of cancer screening procedures delivered nationally according the National Health Interview Survey (Swan, 2003) and Medicare payment rates, we estimate that screening for breast, cervical, colorectal, and prostate would add at least $5 billion to direct medical costs associated with cancer. Out-of-pocket costs are a component of direct costs that are of interest because of their often unpredictable impact on individuals. Several studies have separately measured out-of-pocket costs associated with cancer treatment, and most of them were based on small convenience samples (Lansky, 1979; Houts, 1984; Bloom, 1985; Stommel, 1993). More recently, Thorpe and Howard (2003) examined out-of-pocket

Economic Impact of Cancer in the United States Table 12–1. Cost Domains

direct costs Medical costs

Nonmedical costs

Hospitalizations Physician visits Home health care Hospice care Pharmaceutical agents Chemotherapy Radiation Rehabilitation Prostheses Transportation to hospital or physician’s office Housekeeping services Costs of moving Alterations to property

indirect costs Time spent seeking medical care Time lost from work/lost productivity Economic productivity lost due to premature death Caregiver time or changes in caregiver employment

intangible/psychosocial costs Pain Suffering Grief

costs for medical care among patients reporting cancer as a current condition; they used data from a nationally representative survey, the 1996–1999 wave of the Medical Expenditure Panel Survey. They reported average out-of-pocket medical expenditures to be about 7% of the total medical expenditures ($421 of $6115) over a 6-month period for all individuals and $576 of a total of $8252 for individuals under the age of 65.

Indirect Costs Indirect costs are the time and economic output lost or foregone by the patient, family, friends, and others from usual activities, including employment, housekeeping, volunteer activities, and leisure activities. These costs are not reflected by direct monetary transactions but do reflect the use of economic resources in response to disease occurrence and treatment, resources that could be used for other purposes in the absence of disease. The patient may suffer a cessation or reduction of usual activities because of morbidity, disability, or mortality associated with the illness. Family members and others may spend time caring for the patient rather than pursuing other activities, make unwanted job changes, or lose opportunities for promotion and education. Additional indirect costs include the time the patient and/or family spend visiting physicians and other health professionals. The impact of these costs may be substantial: Indirect costs are estimated

Table 12–2. Medical Care Expenditures for Neoplasms by Type of Medical Service: United States 2002 Source of Expenditure

Amount (millions)

% Distribution

All services Hospital care Physician and other professional services Home health care Prescription drugs Medical durables Nursing home care

$60,900 39,529 14,017

64.9 23.0

3,030 1,411 318 2,597

4.9 2.3 0.5 4.3

Sources: Hodgson (1999), Table 8, p. 139; NHLBI Fact Book (2002), p. 52. 1995 Estimates from Hodgson are updated to 2002 using methods described in the NHLBI Fact Book, p. 52.

203

to comprise more than two-thirds of the total costs in the United States (Brown et al., 1996). Because of data limitations and measurement problems, most “cost of cancer” studies that include indirect costs have focused on the major, but not exhaustive, components of indirect costs known as morbidity and mortality costs. Morbidity costs are the value of losses of economic output for people who are ill or disabled and unable to work or participate in their usual activities. Estimated morbidity costs of cancer typically include losses measured by the value of foregone earnings and the imputed value of lost housekeeping services (measured by wages that would have to be paid to replace those services). These are typically measured among the currently employed, persons not in institutions who are unable to work or participate in usual activities, and persons in institutions such as nursing homes or homes for the aged. Mortality costs are the present value of future output lost because of premature death. One method for estimating this loss, the human capital approach, uses gender- and age-specific average earnings and expected productivity trends to compute remaining lifetime earnings by age. Combined with illness-related, age-specific mortality, an estimate of potential lost earnings due to premature mortality can then be computed. These costs are equivalent to the magnitude of an investment that would have to be made today to make up for the lost productivity that would have accumulated over the remaining life expectancy had the person not died of cancer. This approach may also be used to estimate morbidity productivity losses resulting from absenteeism. The human capital approach to valuing lost productivity places no intrinsic value on human life itself or on noneconomic social relationships. It explicitly values the time of individuals with higher earning potential as greater than that for those with less earning potential. This “leads to unavoidable bias towards those diseases which affect white, middle-class males in employment” (Hartunian, 1981). Another approach—the willingness-to-pay approach—asks individuals how much they would be willing to pay to avoid illness or risks of death. The questioning can be direct, in the form of a lotterytype question (the “binary gamble”) or in terms of products the consumer would be willing to buy to avoid certain risks. Methods of revealed preference have also been developed that attempt to exploit information implicit in actual market behavior. For example, wage differentials between more and less hazardous occupations have been used to estimate the implicit value of life or the value of avoiding a nonfatal injury. Value of life estimates derived from these types of studies vary widely, from a few hundred thousand dollars to several million dollars. The values most often used in economic policy analyses, including cancer-related policy, are in the range of three to six million (1990) dollars, corresponding to a value per each addtional year of life in the range of $75,000 to $150,000 for middle-aged or elderly adults (Cutler and Richardson, 1997) It has recently been shown that the value of life, as estimated by these methods, increased over time in United States between 1940 and 1980. This can be explained, in part, by the increase in national wealth per capita, which makes investments in life-extending interventions more affordable and the increase in life expectancy itself, which increases the value of death averted at any given age (Costa, 2002). For example, as life expectancy has increased owing to decreasing mortality from heart disease, the value of preventing a cancer death is increased. Willingness-to-pay value of life estimates implicitly incorporate both economic losses due to illness and the intrinsic value of living; therefore this measure is not directly comparable to mortality cost as measured by the human capital approach. The following example illustrates the various estimates of mortality cost obtained using these methods. We estimated mortality cost for colorectal cancer in 2002 by the human capital method using mortality data from SEER and earning data from Max et al. (2002). The total estimate was $7667 million, with $2431 million attributed to female deaths and $5235 million to male deaths. We also estimated mortality costs with the willingness to pay approach using SEER data on agespecific average years of life lost, assuming that each life-year was valued at $150,000. Using the latter method, the mortality cost for colorectal cancer in 2002 was estimated to be $126,603 million, more

204

PART II: THE MAGNITUDE OF CANCER

than an order of magnitude greater than the human capital estimate. A discount rate of 3% was used with both methods. It is not surprising that the willingness-to-pay estimate is much larger than the human capital estimate. The former measurement incorporates the entire value that people place on life, and it values life equally for men and women, whereas the latter measures only one aspect of that value: wage earnings. Finally, some health economists have advocated productivity cost as an alternative concept to mortality and morbidity cost. The concept of productivity cost accounts for the net loss of economic productivity to the economy as a whole associated with the occurrence of a disease compared to what productivity would have been in the absence of that disease. From this perspective economic loss is equal to lost productivity during worker absenteeism plus any hiring and training costs that are needed to replace the absent or deceased worker temporarily or permanently (Koopmanschap et al., 1995). This is a difficult concept to operationalize and varies from one setting to another as a function of such factors as job requirements and the employment rate. However, the first step of measuring productivity cost has been implemented in studies of the employment experience of cancer survivors versus healthy individuals (Bradley et al., 2002a, 2002b). Table 12–3 shows estimates of direct and indirect costs of cancer for 2002, where indirect costs are composed of morbidity and mortality costs estimated using the human capital approach. Compared to direct costs, indirect costs, as measured by morbidity and mortality costs, are more than double in size, with most of them attributable to mortality costs. These estimates do not include all types of indirect costs, however, in part because most economic burden studies utilize existing data, which rarely capture this information. Few studies have systematically measured time associated with cancer care, lost wages, unpaid family labor, or changes in employment structure as a result of cancer and its treatment. Studies that have attempted to measure components of these costs have reported that these frequently unmeasured costs may be substantial. In a study conducted among active employees of a Fortune 100 company using health plan and disability data, the costs associated with work absenteeism and disability accounted for $3034, or 23% of the annual incremental costs for cancer patients compared to that for nonpatients ($12,982). (Barnett et al., 2000) Costs counted were employer payments to health care providers or to employees for missed workdays that were covered by disability benefits. In another study, Stommel and colleagues (1993) estimated the average value of unpaid family labor from a sample of 192 Michigan cancer patients currently undergoing outpatient treatment. The labor cost of family members was valued at the contemporary wage rate for home health aides of $7.82 per hour. The value of family caregiver labor averaged $3772 over a 3-month period, or about $14,000 on an annual basis. In one of the few studies to measure patient time costs in the United States, Secker-Walker and colleagues (1999) conducted interviews with 465 women in four Florida communities 4–6 months following the performance of a breast biopsy. They estimated that women expended a median time of 7 hours to participate in breast cancer detection and diagnosis; women diagnosed with breast cancer spent a median time of 72.5 hours receiving treatment and 9.6 hours for follow-up visits. Other small studies of patient time costs have been conducted around cancer screening (Shireman et al., 2001; Suter et al., 2002), although findings are rarely generalizable; and as mentioned

Table 12–3. Direct and Indirect Costs of Cancer: United States, 2002 Type of Cost

Amount (millions)

Percent of Total Cost

Total Direct Indirect Morbidity Mortality

$171.6 60.9 110.7 15.5 95.2

100.0 35.5 64.5 9.0 55.5

Source: NHLBI Fact Book (2001), p. 52.

above, the issue of valuation of patient time is complex, particularly for patients who are retired or not working outside the home. Because cancer is mainly a disease of the elderly, this issue is of particular importance. More recently, Bradley and colleagues (2002a, 2002b) examined the employment experience of long-term breast cancer survivors compared to a control group of individuals without cancer using data from the 1992 wave of the Health and Retirement Study. They found that women with breast cancer had a 10% lower probability of being employed than did women without breast cancer. Among women with breast cancer who were employed, however, the average number of hours worked per week was actually longer, by about 3 hours, than among women without breast cancer. In another study, Bradley and Bednarek (2003) conducted interviews with 253 breast, colon, lung, and prostate cancer patients who had survived 5–7 years. Among this selected group of long-term survivors there was little evidence that cancer had had a long-term impact on employment. Although the average work week for those still employed exceeded 40 hours, there was some evidence that work schedules were occasionally affected by cancer treatment or that individuals experienced work-related limitations due to cancer.

Psychosocial Costs Disease may affect the quality of life of the patient, family, and friends in ways that are not reflected in the categories of direct or indirect costs. These effects are referred to as psychosocial costs. As the result of cancer treatment, patients may suffer from disfigurement, loss of speech, disability, pain, or the threat of impending death. Patients and their families may make changes in life plans that induce anxiety, reduce self-esteem or feelings of well-being, and create resentment and family conflict. As a result of treatment, the patient may experience marked personality changes and reduced sexual function. Disrupted development and delinquency may occur among children. The combination of financial strain and psychosocial problems can be especially devastating. Psychosocial cost can be expressed in dollar terms through willingness-to-pay analysis, and this procedure is followed for purposes of cost-benefit analysis in which both economic and health state outcomes are expressed in monetary terms. Although it may be useful to measure psychosocial cost in monetary terms relative to the monetary value of direct and indirect costs for descriptive and comparative purposes, for the purpose of cost-effectiveness analysis— the mode of analysis most commonly used in the economic analysis of health interventions—psychosocial cost is conceptualized as a quality of life outcome and is measured in terms of decrements of quality-adjusted life-years or in terms of utility (Brown et al., 2001). Developing appropriate concepts and measures of the quality of life ramifications of cancer and cancer treatment is currently an active area of cancer research (Lipscomb et al., 2004).

MEASUREMENT OF HEALTH CARE COSTS In addition to the types of medical care costs, several other factors are critical when measuring cancer burden, including the perspective of analyses, consideration of current and future costs, measurement of resource utilization, and cancer-attributable costs. These issues are defined and discussed in the following sections.

Frame of Reference or Perspective The frame of reference of an economic study is the viewpoint from which the analyses are conducted. As shown in Table 12–4, examples of perspectives are the patient and family, employers, health care delivery organizations or payer (including federal and state programs such as Medicare and Medicaid), and society. In appendixes 12–A, 12–B, and 12–C we have summarized the cost of cancer studies by perspective. Depending on the perspective of the analysis, the types of costs included, sources of data, and study design vary, as reflected in the Appendix entries. The societal perspective is preferred for

Economic Impact of Cancer in the United States Table 12–4. Perspectives and Relevant Cost Categories Cost Category

Patient and Family

direct costs

X (health X (those with insurance employer-funded premiums, health plans) copayments) X

Medical Nonmedical

indirect costs

X

intangible costs

X

Employer

X (reduced productivity, disability payments, absenteeism, replacement costs)

Health Insurer/Payer

Society

X (minus copayments)

X

NA

X

NA

X

NA

X

205

applied to dollar expenditures during different years to account for the changing purchasing power of the currency. There is much controversy about what the appropriate discount rate should be from a social viewpoint. The most commonly used value in current health economics literature is 3% in real terms (i.e., adjusted for inflation). Recommended rates vary from as low as 2% to as high as 10% (Gillette and Hopkins, 1988). The choice of a discount rate reflects a value judgment: a low discount rate is indicative of a longterm horizon and a high discount rate is indicative of a short-term horizon. The use of a high discount rate, for instance, would tend to favor a treatment program over a prevention program, all things being equal. Although survey studies suggest that the average individual tends to use a high discount rate when making individual decisions (Cropper et al., 1992), it can be argued as well that intergenerational equity and even social “sustainability” require the use of a low social discount rate (Howarth, 1992). In other words, the social discount rate is the reflection, not the determinant, of social values.

Adjusting for Inflation

purposes of economic analysis for general policy analysis, whereas for a variety of policy analysis purposes, such as the impact of new health care legislation, it is often useful to define a frame of reference that is narrower than all of society (Gold et al., 1996). For example, whereas the societal perspective may be the relevant perspective for policy analysis involving approval of a new cancer control innovation, the perspective of Medicare or a health maintenance organization (HMO) may be relevant for assessing the financial impact of such a decision on the annual budget requirements or premium requirements required to fund these operations. Not all social costs are recognized as costs at various organizational levels. Time lost to paid work for unpaid caregiving by the relatives of the cancer patient is a cost to society in terms of lost economic productivity and cost to the family in terms of lost income, but it might constitute a savings to Medicare in terms of a shorter length of hospital stay or less provision of formal home care services. The time spent by cancer patients and family members related to cancer and its treatment is rarely fully accounted for in existing studies because this is not a cost that shows up in the usual accounting mechanisms of health care delivery or health insurance organizations. Expenditures that may appear as costs from the perspective of a particular individual or institution may not be true economic costs from a social perspective but, rather, may constitute a transfer of wealth from one sector to another. For example, lost productivity due to illness-related work disability is a true economic cost to society, but disability insurance payments to an employee from an employer or a social insurance trust fund is a transfer payment. In general, when accounting for direct and indirect costs due to illness it is important not to undercount, overcount, or double count by confusing costs and transfer payments. However, the magnitude and incidence of transfer payments (who pays, who recieves) may be germane to policymakers interested in issues of distribution and fairness. For example, a whole genre of literature has developed around the question of whether consumers of tobacco products pay society more in taxes than they cost society in increased health care expenditures (Hodgson, 1992; Barendregt et al., 1997).

It is often useful to express the results of cost-of-cancer studies conducted in past years in current-year dollars. Because the rate of inflation for the price of health care services has been consistently higher than for the economy as a whole, it is necessary to use a price index specific for this sector of the economy. The health care component of the Consumer Price Index (CPI) is often used for this purpose. This index, however, refers only to the components of health care expenditures paid for directly by consumers. Other price indices are available that are designed to reflect health care price inflation as experienced by the Medicare program. Over the period 1990–1996, the physician services component of the CPI indicated an inflation rate of 40%, but the Medical Economic Index (MEI), used to track price increases in the Medicare program indicated an inflation rate of only 14% (Sensenig et al., 1997). Note that adjustments for inflation and the discounting of future costs and benefits are distinct and unrelated issues.

Charges, Reimbursements, Expenditures, Costs The economic concept of cost relates to the use of specific resources in an efficient manner. Direct observation of resource use is rarely available, however. Reflections of cost are available, such as the amount billed, charged, or reimbursed for the provision of services. The amount billed or charged may not reflect the actual cost associated with providing services, however. In some cases these pecuniary entities may have little relation to underlying costs because they are based on historical, not current, determinants of cost or reflect financial strategies such as cross-subsidization or price discrimination (i.e., charging different prices for the same product to different groups of purchasers) (Finkler, 1982). For example, some hospitals or other providers may negotiate with health plans for reduced payments for specific patients. Here, the actual “cost” of the service may be unrelated to the amount charged to any of the purchasers. In addition, even the accounting of these pecuniary entities may be incomplete. For example, from a particular data source, reimbursements from insurance organizations may be available but not out-of-pocket payments from individual patients; or reimbursements for outpatient care may be available but not for oral medications.

Discounting

Measuring Costs Attributable to Cancer

As mentioned above, mortality costs are usually expressed in terms of present value: the sum of discounted annual costs. When costs are distributed over time and the purpose of analysis is to assess these costs relative to an initial investment, such as a cancer screening program to promote early detection, economists use an adjustment known as discounting. The purpose of discounting is to adjust the value of costs or savings incurred in the future because the same resources, if available and invested today, would yield a return if placed in a productive activity. This is unrelated to adjustments for inflation, which are

There are two general approaches to estimating the costs attributable to cancer: (1) development of treatment scenarios and estimating costs for specific events and (2) comparison of costs of all care for cancer patients compared to that for noncancer patients. With the first approach, researchers outline treatment scenarios as a series of probablistic events that could be applied for a well defined, hypothetical population of cancer patients. The cost of medical services and procedures defined by the scenario can then be estimated using a cost estimate from another data source.

206

PART II: THE MAGNITUDE OF CANCER

This treatment scenario approach has been used to compare coverage gaps by Medicare and other health insurance policies for selected conditions, including early stage breast cancer and lung cancer (Sofaer and Davidson, 1990; Sofaer et al., 1990). Findings from this approach indicated that Medicare covered 83% of the first-year cost of lung cancer treatment and 65% of the first-year cost of breast cancer treatment. More recently, the scenario approach has been used to develop disease-specific cost of treatment estimates by the Canadian Population Health Model (POHEM). POHEM generates a hypothetical cohort of people with demographic and labor force characteristics, risk factor exposures, and health histories typical of Canadians. To generate treatment cost for POHEM, treatment scenarios based on expert opinion and clinical trials are combined with unit costs and intensity of health care resources derived from administrative data (Wolfson, 1994; Evans et al., 1995). Because scenarios are typically constructed with expert opinion, rather than direct observation of treatment patterns and related care, this approach may underestimate the frequency of services cancer patients receive for seemingly unrelated care. Another approach to measuring attributable costs of cancer care is to compare the costs of all care for cancer patients to that for noncancer patients. This approach has been implemented, for example, in studies of direct cancer cost using Medicare and HMO administrative data linked to SEER cancer registries (Taplin et al., 1995; Fireman et al., 1997; Brown et al., 1999; Barlow et al., 2001; Warren et al., 2001). To implement this approach, the total medical cost experienced by cancer patients is compared to the medical care cost experienced by noncancer patients who are matched by relevant characteristics such as age, gender, geographic region, co-morbidity status, and time to death. This approach is particularly advantageous when using administrative data because it minimizes reliance on diagnostic and procedure codes for identifying services as cancer-related or for other conditions.

STUDY DESIGN AND SOURCES OF EXISTING COST, CHARGE, EXPENDITURE, AND REIMBURSEMENT DATA As in other areas of health care research, studies of the burden of cancer care can utilize a variety of study designs, including case series, cross-sectional studies, and case-control, cohort, and randomized controlled trials. As with studies with morbidity or mortality as the outcome measure, there are strengths and limits of these study designs for measuring the economic burden of cancer. To date, the design of most cost of cancer care studies have been driven more by existing data than by explicit consideration of design strengths and limits. This may be due, in part, to the complexities and expense of collecting detailed cost data for long periods of time. Data on the costs of medical care can be obtained from several existing sources including administrative claims or billing systems, surveys, or a combination of these records. In the next sections, these sources of existing data are briefly discussed.

Administrative Data Administrative data from health plans or state or federal plans such as Medicaid or Medicare contain information about payments for the provision of specific services. For the periods when patients maintain eligibility or remain members of the plan, these data can provide longitudinal information on service utilization and associated expenditures or reimbursements. By definition, these administrative data do not contain information on services that are not covered by the plan. Additionally, payments recorded in administrative data reflect only those reimbursed by the plan; they do not include patient co-payments or deductibles. Identifying specific services requires use of diagnostic or procedure codes. One of the major limitations when using administrative data to estimate costs associated with cancer care is that incident cases, stage of diagnosis, and recurrence cannot be reliably identified.

Hospital Discharge Data Information from billing systems from hospitals or hospital discharges contain information on all patients billed, rather than just indiviudals in a specifc health plan, making these data potentially more generalizable for hospitalizations. However, these data do not contain information on prior care or follow-up care; nor do they have information on care for patients discharged and readmitted or treated in physician offices or other outpatient settings. Additionally, these data are generally event-based and not linked by patient; for instance, four discharges may represent four patients or a single patient with multiple admissions. Discharge data also require reliance on diagnostic and procedure codes to identify cancer patients and cancer-related services. Despite these limitations, these data can provide useful descriptive information on trends in cancer-related hospital costs. For example, Seifeldin and Hantsch (1999) used hospital data from the Healthcare Cost and Utilization Project, a state-based hospital discharge database including hospital charges maintained by the Agency for Healthcare Research and Quality, to estimate total national hospital charges associated with admissions for the treatment of colorectal cancer. They noted that the costs increased from $3.90 billion in 1991 to $5.14 billion in 1994.

Cross-Sectional, Panel, and Follow-Back Surveys Surveys of cancer patients or their families have also been used to assess service utilization and the cost of care. Patients may be identified from specific sites, and population-based registries have been used as sample frames for patient surveys about cost. For example, as a supplemental study to the Third National Cancer Survey (TNCS), cancer patients diagnosed during 1969–1970 were followed for up to 2 years from the date of diagnosis, and information was collected on their use of hospital-based and outpatient health care services and the cost of these services (Scotto and Chiazze, 1977; Cromwell and Gertman, 1979). Other surveys have identified patients who have died from cancer and contacted surviving relatives for information about costs of care (Cancer Care Inc., 1973). These retrospective, follow-back studies may be particularly useful in terms of measuring the impact of cancer on the family, although response rates may be low and reporting incomplete. Additionally, depending on the source of the sample frame, the family or next of kin may be difficult to locate, the information may not be considered confidential, and samples are likely to be limited in some way. Limitations associated with survey-based estimates of costs include response rates, differential loss to followup (as patients become sicker their loss to follow-up is more likely), and issues with patient recall.

Nationally Representative Surveys with Record Review Other large nationally representative surveys have combined survey with medical record and administrative data review. Currently, the Medical Expenditure Panel Survey (MEPS) is co-sponsored by the Agency for Healthcare Research and Quality and the National Center for Health Statistics to provide national estimates on the costs of care. MEPS consists of overlapping 2-year panels of patients so it can be used to provide cross-sectional and limited longitudinal estimates. The advantage of MEPS is that it contains comprehensive data elements for all cost domains. For any given panel, however, the number of cancer patients available in MEPS is limited to a few thousand. Perhaps a greater weakness is that it is difficult to ascertain, for individuals in MEPS, how recently a cancer has been diagnosed (Cohen, 2000; Thorpe and Howard, 2003).

Population-Based Tumor Registries Linked to Administrative Data Some of the limitations associated with the use of administrative data can be eliminated through the linkage of administrative claims with tumor registry data. Linkages have been performed with the SEER

207

Economic Impact of Cancer in the United States tumor registry to Medicare, state-based registries and Medicaid, and specific SEER registries and managed care organizations. A major advantage to these linkages is that the date of cancer diagnosis and stage of disease at the time of diagnosis can be reliably ascertained. In addition, these data resources provide a longitudinal record of payments, procedures, and services so long as the beneficiaries remained enrolled in the plan. For Medicare, continuous enrollment approaches 100%. For a large nonprofit HMO the annual disenrollment of cancer patients is low, less than 5% (Riley et al., 1996). Although linkages are limited to individuals receiving coverage, the SEER–Medicare linkage includes almost all individuals aged 65, the age group where cancer incidence is highest. The limitation to these data sources is that there are no records for services not receiving coverage. For Medicare this includes orally administered prescription drugs, an important component of treatment for some cancers (Warren et al., 2002).

Table 12–5. Direct Costs of Prostate Cancer by Type of Health Care Service: California, 1998

Combination of Multiple Study Designs and Sources

are used in cost-effectiveness studies designed to ensure efficient use of increasingly constrained health care resources. For example, estimates of lifetime costs of cancer treatment have been incorporated into cost-effectiveness analyses of screening programs for cancer (Kerlikowske et al., 1999; Frazier et al., 2000), adjuvant therapy for early-stage breast cancer (Hillner and Smith, 1991), and cancer-related dietary prevention research (Urban, 1989).

Several cost of cancer care studies have combined data from multiple study designs and sources to estimate costs of care. The method is commonly used by simulation models that estimate the costeffectiveness of various treatments or procedures as well as the cost-of-illness approach to estimating cancer burden, described below. With this approach, cost parameters are identified and then costs are estimated for each parameter. For example, hospitalization costs might be measured by a cross-sectional sample of hospital discharges among individuals with cancer diagnosis codes, and costs associated with treatment might be estimated from a treatment algorithm, with the average costs of chemotherapy and radiotherapy and their delivery estimated from other sources. Indirect costs associated with caregiver time might be estimated from data published from a survey. Another, related approach measures service utilization among a cancer cohort and applies standard costs for each type of service. This approach has been used in estimates of costs associated with care in clinical trials and could also be applied to patient time costs (Wagner et al., 1999; Goldman et al., 2001). Although this approach to measuring the burden of cancer care increases the comprehensiveness of the estimate, the final estimate reflects the limitations of each of the underlying study designs and data sources.

APPROACHES TO ESTIMATING THE ECONOMIC BURDEN OF CANCER Two general approaches have been developed to measure the cost of cancer (Hodgson, 1988). One approach, which has its origin in national income accounting and is generally known as the cost-ofillness approach, uses available data to provide an estimate of the annual aggregate economic impact of disease. This approach tracks cost-generating events from national or regional health system survey sources and attributes a monetary value to each event. This process produces an estimate of the annual aggregate burden of illness measured by the value of goods and services diverted from other uses to provide medical care and that is lost because of idled labor. Morbidity and mortality from cancer are translated into use and expenditures for medical care, time lost from work and housekeeping, and foregone wages and salaries. The other main approach, the incidence approach, which is derived from the microeconomic field of project evaluation, describes the longitudinal pattern of costs incurred by the average patient from the date of diagnosis as well as the total lifetime costs of cancer treatment— cumulative costs from the date of diagnosis to death or death from other causes. Although the cost-of-illness approach provides policymakers with data on the economic order of magnitude of the disease problem and annual budgetary implications, data based on the incidence approach

Health Care Service

Total Costs (thousands of $)

%

Hospital inpatient facility Emergency department visits Hospital outpatient visits Office-based provider visits Prescription medications Home health care Nursing home care

180,303 105,070 872 25,135 23,010 3,827 5,692 8,799

100 58.3 0.5 13.9 12.8 2.1 3.2 4.9

mortality cost

180,198

total direct cost

Source: Max et al. (2002).

Cost-of-Illness Approach to Measuring Economic Burden The cost-of-illness approach has been operationalized in several cost-of-illness studies for all cancers by Rice and colleagues (1985). It has also been used to estimate the cost of colorectal and prostate cancer (Max et al., 2002; Sandler et al., 2002). As an example of this method, Max and colleagues recently estimated the cost of illness for prostate cancer in California (Max et al., 2002). Hospitalization charges were measured from the California Hospital Discharge Survey for individuals with primary or secondary prostate cancer diagnoses and then were adjusted to costs using a hospital-specific cost to charge ratio. Inpatient physician services, emergency department visits, outpatient visits, prescription drugs, and home health were estimated from the Medical Expenditure Panel Survey for all cancer patients. Nursing home costs were estimated from the number of individuals discharged to nursing homes from the California Hospital Discharge Data using average length of stays for people with cancer diagnoses multiplied by the per diem nursing home costs. Mortality costs were estimated from the California Mortality File, which contains death certificates with underlying cause of death and age at death. Earnings forgone due to premature mortality utilized age-specific earnings varying by different participation in the labor force (e.g., wage rates, housekeeping services). Estimates of outpatient physician visits, transportation, and patient and caregiver time costs were not included in the analysis. Table 12–5 shows the results of this analysis. Direct medical costs in 1998 for prostate cancer in California were estimated to be about $180 million, almost 60% due to hospital inpatient costs. Mortality costs accounted for an additional $181 million.

Incidence Approach to Measuring Economic Burden The incidence approach has been used by several studies of direct cancer costs that utilized administrative data linked to cancer registries (Riley et al., 1995; Taplin et al., 1995; Etzioni et al., 1996; Fireman et al., 1997; Brown et al., 1999; Barlow et al., 2001; Etzioni et al., 2001; Warren et al., 2001). For example, Brown and colleagues (1999) reported direct costs for colon and rectum cancer using SEER–Medicare data. Table 12–6 shows cancer-related costs as measured by Medicare payments for colorectal cancer by stage at diagnosis and phase of cancer treatment. The phases of treatment are defined as: initial phase, 6 months following diagnosis; terminal phase, 1 year prior to death; continuing phase, all other months. These

208

PART II: THE MAGNITUDE OF CANCER

Table 12–6. Estimates of Cancer-Related Medicare Payments for Colorectal Cancer Cases by Treatment Phase and Stage at Diagnosis Cancer Stage All stages In situ I II III IV Unknown

No.

Initial (6 months)

Continuing (per year)

Terminal (1 year)

16,527 1,378 4,229 5,635 3,515 1,155 615

$18,100 8,200 15,200 19,800 22,200 21,200 13,700

$1,500 1,200 1,200 1,200 2,200 5,200 2,500

$15,200 10,400 11,200 13,200 17,900 21,600 10,900

Source: Brown et al. (1999). Medicare payments for 1990–1994, expressed in 1999 dollars.

phase-specific cost estimates can be combined with SEER-based survival probabilities to obtain estimates of the long-term cost of cancer treatment: the accrued cost from the date of diagnosis to death averaged over the survival experience of all individuals diagnosed with cancer. For example, in this study, the long-term cost of treatment received by all individuals diagnosed with colon cancer was estimated to be $60,600, and the long-term cost related to the treatment of colon cancer was estimated to be $33,700 (both estimates undiscounted). These costs are not directly comparable to the direct component of cost-of-illness estimates. The incidence method measures all treatment costs that will ever be incurred by patients diagnosed during the current year, whereas the cost-of-illness method measures direct costs for all patients undergoing active cancer treatment during the current year. Treatment-phase-specific cost estimates, however, can be combined with SEER-derived treatment-phase prevalence estimates to yield an alternative estimate that is comparable to cost-of-illness estimates. Medicare-derived costs must be supplemented with information from HMO-based costs studies linked to SEER for patients under 65 years of age to produce national cost estimates for all cancer patients (Brown et al., 2002). In the case of colorectal cancer, this approach yields an estimate of $5.664 billion in 1998 dollars. This is in close agreement with the estimate of $5.754 billion obtained by Sandler and colleagues using the cost of illness approach (Sandler et al., 2002). Similarly, the SEER–Medicare based estimate of total direct cost for 1994–1997 is in close agreement with the cost-of-illness estimate for 1995 (Hodgson et al., 1999; Brown et al., 2002). Table 12–7 shows additional estimates of direct costs for the most prevalent cancer sites by treatment phase (Brown et al., 2002). Note that we have expressed these estimates in 2002 dollars, and they are only roughly comparable to the estimate of total direct cost shown in Table 12–2 because we have not taken the increase in cancer prevalence into account that occurred from our estimation period of 1994–1997 to 2002. We have examined this issue for colorectal cancer; and for that cancer site the increase preva-

Table 12–7. Estimates of National Expenditures for Medical Treatment of Cancer Cancer Site Breast Colorectal Lung Prostate

Total Cost

Initial Phase Cost

Continuing Phase Cost

Terminal Phase Cost

$6217 6324 5783 5335

$2709 3343 2053 2183

$1852 1412 665 2092

$1655 1568 3065 1059

Source: Updated from Brown (2002). Estimates based 1994–1997 SEER–Medicare data. Estimates are in millions of 2002 dollars.

lence would yield an estimate of direct cost 13% higher than the estimate shown in Table 12–7.

FUTURE DIRECTIONS Until recently, estimates of the economic burden of cancer were constrained by the limited availability of data resources. Estimates were based on either large national health surveys or small local convenience samples. The advantage of the former is that the data are nationally representative; the disadvantage is that these general purpose data resources contain only modest numbers of cancer cases; moreover, the tumor site and stage of cancer are not reliably identified, and the date of diagnosis is not reliably ascertained. Small local studies, sometimes linked to clinical trials, often provide more detailed information on important clinical variables but often address only one cancer site in a highly selected population. During the last decade a new generation of data resources has been developed that overcomes some of these limitations. These resources fall into two categories: health system administrative data linked to population-based tumor registries and large population-based prospective studies of the long-term experience of cancer patients. These resources make it possible to obtain reliable information on stage, diagnosis date, treatment, and elements of indirect costs for large cohorts of cancer patients and comparable noncancer patients. The first category includes the SEER–Medicare database (Warren et al., 2002; http://healthservices.cancer.gov/ seermedicare/). Currently there are efforts to evaluate the feasibility and utility of also linking SEER to Medicaid data (Bradley et al., 2002c) and to hospital discharge data (Brooks et al., 2000). This category also includes linkage of the administrative data of large HMOs to SEER or other cancer registry systems or to pathology records (Taplin et al., 1995; Fireman et al., 1997; Barlow et al., 2001). The National Cancer Institute also currently supports the Cancer Research Network, a system of research organizations affiliated with 11 large not-for-profit HMOs in the United States. One of the goals of this network is to increase the scope and depth of cancer-related economic data associated with these health care delivery systems (http://healthservices.cancer.gov/hmo/). The second category includes large prospective studies of national scope that are designed to obtain data on patient-oriented outcomes of cancer treatment, including economic burden. These studies include the Prostate Cancer Outcomes Study (Potosky et al., 1999) and the Cancer Care Outcomes Research & Surveillance Consortium (http://healthservices.cancer.gov/cancors/). These data resources should make it possible to obtain more detailed and more generalizable estimates of many of the elements of indirect cost than has been possible in the past. As these new data resources mature, it will be possible to explore several areas of the economic burden of cancer that have been elusive in the past. For example, the SEER–Medicare and similar data resources can be used to obtain generalizable estimates of patient time cost associated with cancer treatment, a cost element considered crucial for economic analysis (Gold et al., 1996) but one that hitherto has, by and large, eluded quantitative description. These estimates can be generated by comparing service use for cancer and noncancer patients and applying estimates of service and travel time by type of medical visit (Yabroff et al., 2003). These time estimates can be converted to time cost by applying a value to time measures, such as hourly wages. As the longtitudinal data from these resources accrues, it will also facilitate the analysis of important questions about trends in the components of economic burden: for example, the mix of economic costs between inpatient and outpatient care and between direct and indirect costs. Finally, combining these data with new methodologic approaches to estimating and projecting cancer prevalence will make it possible to provide meaningful long-term projects of the economic burden of cancer, including a consideration of the impact of the aging U.S. population on the economic burden of cancer.

Economic Impact of Cancer in the United States References Baker MS, Kessler LG, Urban N, Smucker RC. 1991. Estimating the treatment costs of breast and lung cancer. Med Care 29:40–49. Barendregt J J, Bonneux L, van der Maas P J. 1997. The health care costs of smoking. N Engl J Med 337:1052–1057. Barlow WE, Taplin SH, Yoshida CK, Buist DS, Seger D, Brown M. 2001. Cost comparison of mastectomy versus breast-conserving therapy for early stage breast cancer. J Natl Cancer Inst 93:447–455. Barnett A, Birnbaum H, Cremieux P-Y, Fendrick AM, Slavin M. 2000. The costs of cancer to major employer in the United States: a case-control analysis. Am J Manag Care 6:1243–1251. Bloom BS, Knorr RS, Evans AE. 1985. The epidemiology of disease expenses: the costs of caring for children with cancer. JAMA 253: 2393–2397. Bradley CJ, Bednarek HL. 2002. Employment patterns of long-term cancer survivors. Psychooncology 11:188–198. Bradley CJ, Bednark HL, Neumark D. 2002a. Breast cancer survival, work, and earnings. J Health Econ 21:757–779. Bradley CJ, Bednark HL, Neumark D. 2002b. Breast cancer and women’s labor supply. Health Services Res 37:1309–1328. Bradley CJ, Given CW, Roberts C. 2002c. Race, socioeconomic status, and breast cancer treatment and survival. J Natl Cancer Inst 94:490–496. Brooks JM, Chrischilles E, Scott S, Ritho J, Chen-Hardee S. 2000. Information gained from linking SEER cancer registry data to state-level hospital discharge abstracts. Med Care 38:1131–1140. Brown ML, Riley GF, Schussler N, Etzioni R. 2002. Estimating the health care costs related to cancer treatment from SEER-Medicare data. Med Care 40:IV-104–IV-117. Brown ML, Lipscomb J, Snyder C. 2001. The burden of illness of cancer: economic cost and quality of life. Annu Rev Public Health 22:91–113. Brown ML, Riley GF, Potosky AL, Etzioni RD. 1999. Obtaining long-term disease specific costs of care: application to Medicare enrollees diagnosed with colorectal cancer. Med Care 37:1249–1259. Brown ML, Hodgson TA, Rice DP. 1996. Economic impact of cancer in the United States. In: Schottenfeld D, Fraumeni JF Jr, editors. Cancer Epidemiology and Prevention, 2nd ed. New York: Oxford University Press. Calhoun EA, Chang C-H, Welshman EE, Fishman DA, Lurain JR, Bennett CL. 2001. Evaluating the total costs of chemotherapy-induced toxicity: results from a pilot study with ovarian cancer patients. Oncologist 6:441– 445. Cancer Care, Inc. 1973. The Impact, Costs and Consequences of Catastrophic Illness on Patients and Families. New York: Cancer Care and the National Cancer Foundation. Chirikos TN, Russell-Jacobs A, Cantor AB. 2002. Indirect economic effects of long-term breast cancer survival. Cancer Pract 10:248–255. Cohen S. 2000. Sample Design of the 1997 Medical Expenditure Panel Survey Household Component. Methodology Report no. 11, Publ. No. 01-0001. Rockville, Md: Agency for Healthcare Research and Quality. Costa DL. 2002. Changes in the value of life, 1940–1980. NBER Working Paper 9396 (http://www.nber.org/papers/w9396). Cambridge, MA: National Bureau of Economic Research. Cromwell J, Gertman P. 1979. The cost of cancer. Laryngoscope 89:393–409. Cropper ML, Aydede SK, Portney PR. Rates of time preference for saving lives. Am Econ Rev 82: 469–72, 1992 Cutler DM, Richardson E. 1997. Measuring the health of the U.S. population: Brookings papers on economic activity. Microeconomics 29:519–539. Edwards BK, Howe HL, Ries LA, Thun MJ, Rosenberg HM, Yancik R, Wingo PA, Jemal A, Feigal EG. 2002. Annual report to the nation on the status of cancer, 1973–1999, featuring implications of age and aging on U.S. cancer burden. Cancer 94:2766–2792. Etzioni R, Ramsey SD, Berry K, Brown M. 2001. The impact of including future medical care costs when estimating the costs attributable to disease: a colorectal cancer case study. Health Econ 10:245–256. Etzioni R, Urban N, Baker M. 1996. Estimating the costs attributable to a disease with application to ovarian cancer. J Clin Epidemiol 49:95– 103. Evans WK, Will BP, Berthelot JM, Wolfson MC. 1995. Diagnostic and therapeutic approaches to lung cancer in Canada and their costs. Br J Cancer 72:1270–1277. Finkler SA. The distinction between costs and charges. 1982. Ann Intern Med 96:102–109. Fireman BH, Quesenberry CP, Somkin CP, Jacobson AS, Baer D, West D, Potosky AL, Brown ML. 1997. Cost of care for cancer in a health maintenance organization. Health Care Financ Rev 1997;18:51–76. Fireman BH, Fehrenbacher L, Gruskin EP, Ray GT. 2000. Cost of care for patients in cancer clinical trials. J Natl Cancer Inst 92:136–142.

209

Frazier AL, Colditz GA, Fuchs CS, Kuntz KM. 2000. Cost-effectiveness of screening for colorectal cancer in the general population. JAMA 84:1954–1961. Gillette CP, Hopkins TD. 1988. Federal Agency Valuations of Human Life. Washington DC: The Administrative Conference of the United States. Gold MR, Siegel JE, Russell LB, Weinstein MC. 1996. Cost-Effectiveness In Health and Medicine. New York: Oxford University Press. Goldman DP, Schoenbaum ML, Potosky AL, Weeks JC, Berry SH, Escarce JJ, Weidmer BA, Kilgore ML, Wagle N, Adams JL, Figlin RA, Lewis JH, Cohen J, Kaplan R, McCabe M. 2001. Measuring the incremental cost of clinical cancer research. J Clin Oncol 19:105–110. Gulati SC, Bennett CL. 1992. Granulocyte-macrophage colony-stimulating factor (GM-CSF) as adjunct therapy in relapsed Hodgkin disease. Ann Intern Med 116:177–182. Hartunian NS, Smart CN, Thompson MS. 1981. The incidence and economic costs of major health impairments: a comparative analysis of cancer, motor vehicle injuries, coronary heart disease, and stroke. Lexington, MA: Lexington Books. Helms LJ, Melnikow J. 1999. Determining costs of health care services for cost-effectiveness analyses: the case of cervical cancer prevention and treatment. Med Care 37:652–661. Hillner BE, McDonald K, Desch CE, Smith TJ, Penberthy LT, Maddox P, Retchin SM. 1998. Costs of care associated with non-small-cell lung cancer in a commercially insured cohort. J Clin Oncol 16:1420–1424. Hillner BE, Smith TJ. 1991. Efficacy and cost effectiveness of adjuvant chemotherapy in women with node-negative breast cancer: a decisionanalysis model. N Engl J Med 324:160–168. Hodgson TA, Cohen AJ. 1999. Medical expenditures for major diseases, 1995. Health Care Financing Rev 21:119–164. Hodgson TA. 1988. Annual costs of illness versus lifetime costs of illness and implications of structural change. Drug Information J 22:323–341. Hodgson TA. 1992. Cigarette smoking and lifetime medical expenditures. Millbank Q 70:81–125. Houts PS, Lipton A, Harvey HA, Martin B, Simmonds MA, Dixon RH, Longo S, Andrews T, Gordon RA, Meloy J, Hoffman SL. 1984. Nonmedical costs to patients and their families associated with outpatient chemotherapy. Cancer 53:2388–2392. Howarth RB, Norgaard, RB. 1992. Environmental valuation under sustainable development Am Econ Rev 82:473–477. Kerlikowske K, Salzmann P, Phillips KA, Cauley JA, Cummings SR. 1999. Impact on life expectancy and cost-effectiveness. JAMA 282: 2156–2163. Koopmanschap MA, Rutten FF, van Ineveld BM, van Roijen L. 1995. The friction indirect costs of disease. J Health Econ 14:171–189. Lansky SB, Cairns NU, Clark GM, Lowman J, Miller L, Trueworthy R. 1979. Childhood cancer: nonmedical costs of illness. Cancer 43:403–408. Lawrence WF, Peshkin BN, Liang W, Isaacs C, Lerman C, Mandelblatt JS. 2001. Cost of genetic counseling and testing for BRCA1 and BRCA2 breast cancer susceptibility mutations. Cancer Epidemiol Biomarkers Prev 10:475–481. Legorreta AP, Brooks RJ, Leibowitz AN, Solin LJ. 1996. Cost of breast cancer treatment: a 4-year longitudinal study. Arch Intern Med 156:2197–2201. Lichtenberg FR. 2001. The allocation of publicly funded biomedical research. In: Berndt E, Cutler D, editors. Medical Care Output and Productivity, Studies in Income and Wealth, Vol. 62. Chicago: University of Chicago Press, pp. 565–589. Lipscomb J, Gotay CC, Snyder C. 2005. Outcomes Assessment in Cancer: Measures, Methods, and Applications. Cambridge, UK: Cambridge University Press. Max W, Rice DP, Sung H-Y, Michel M, Breuer W, Zhang X. 2002. The economic burden of prostate cancer, California, 1998. Cancer 94:2906–2913. Max W, Rice DP, Sung H-Y, Michel M, Breuer W, Zhang X. 2003. The economic burden of gynecologic cancers in California, 1998. Gynecol Oncol 88:96–103. Mor V, Kidder D. 1985. Cost savings in hospice: final results of the National Hospice Study. Health Serv Res 20:407–422. Murphy KM, Topel RH, editors. 2003. Measuring the Gains from Medical Research: An Economic Approach. Chicago: University of Chicago Press. National Heart Lung and Blood Institute. 2002. Fact Book Fiscal Year 2001. Bethesda, MD: National Institutes of Health. National Institutes of Health. 1997. Setting Research Priorities at the National Institutes of Health. NIH Publ. No. 97–4265. Bethesda, MD: NIH. Penberthy L, Retchin SM, McDonald MK, McClish DK, Desch CE, Riley GF, Smith TJ, Hillner BE, Newschaffer CJ. 1999. Predictors of Medicare costs in elderly beneficiaries with breast, colorectal, lung, or prostate cancer. Health Care Manag Sci 2:149–160.

210

PART II: THE MAGNITUDE OF CANCER

Potosky AL, Harlan LC, Stanford JL, Gilliland FD, Hamilton AS, Albertsen PC, Eley JW, Liff JM, Deapen D, Stephenson RA, Legler J, Ferrans CE, Talcott JA, Litwin MS. 1999. Prostate cancer practice patterns and quality of life: the Prostate Cancer Outcomes Study. J Natl Cancer Inst 91:1719–1724. Ramsey SD, Berry K, Etzioni R. 2002. Lifetime cancer-attributable cost of care for long term survivors of colorectal cancer. Am J Gastroenterol 97:440–445. Rice DP, Hodgson TA, Kopstein AN. 1985. The economic cost of illness: a replication and update. Health Care Financ Rev 7:61–80. Riley G, Lubitz J, Prihoda R, Rabey E. 1987. The use and costs of Medicare services by cause of death. Inquiry 24:233–244. Riley GF, Potosky AL, Lubitz JD, Kessler LG. 1995. Medicare payments from diagnosis to death for elderly cancer patients by stage and diagnosis. Med Care 33:828–841. Riley GF, Feuer EJ, Lubitz JD. 1996. Disenrollment of Medicare cancer patients from health maintenance organizations. Med Care 34:826–836. Sandler RS, Everhart JE, Donowitz M, Adams E, Cronin K, Goodman C, Gemmen E, Shah S, Avdic A, Rubin R. 2002. The burden of selected digestive diseases in the United States. Gastroenterology 122:1500–1511. Schulman KA, Dorsainvil D, Yabroff KR, DiCesare J, Felser JM, Glick HA, IL-3 BMT Study Team. 1998. Prospective economic evaluation of GMCSF/IL-3 in bone marrow transplant for leukemia and lymphoma. Bone Marrow Transplant 21:607–614. Scotto J, Chiazze L. 1977. Cancer prevalence and hospital payments. J Natl Cancer Inst 59:345–349. Secker-Walker RH, Vacek PM, Hooper GJ, Plante DA, Detsky AS. 1999. Screening for breast cancer: time, travel, and out-of-pocket expenses. J Natl Cancer Inst 91:702–708. Seifeldin R, Hantsch JJ. 1999. The economic burden associated with colon cancer in the United States. Clin Ther 21:1370–1379. Sensenig AL, Heffler SK, Donham CS. 1997. Hospital, employment, and price indicators for the health care industry: fourth quarter 19996 and annual data for 1988–96. Health Care Financ Rev 18:133–200. SEER. Prevalence. http://seer.cancer.gov/csr/1975_2000/results_merged/ topic_prevalence.pdf. Accessed May 1, 2003. Sherman EJ, Pfister DG, Ruchlin HS, Rubin DM, Radzyner MH, Kelleher GH, Slovin SF, Kelly WK, Scher HI. 2001. The collection of indirect and nonmedical direct costs (COIN) form. Cancer 91:841–853. Shireman TI, Tsevat J, Goldie SJ. 2001. Time costs associated with cervical cancer screening. Int J Technol Assess Health Care 17:146–152. Simon MS, Stano M, Severson RK, Hoff MS, Smith DW. 1996. Clinical surveillance for early stage breast cancer: an analysis of claims data. Breast Cancer Res Treat 40:119–128.

Sofaer S, Davidson BN. 1990. Illness-episode approach: costs and benefits of medigap insurance. Health Care Financ Rev 11:121–131. Sofaer S, Davidson BN, Goodman RD, Grier R, Kenney E, Monk N, et al. 1990. Helping Medicare beneficiaries choose health insurance: the illness episode approach. Gerontologist 30:308–315. Spector WD, Mor V. 1984. Utilization and charges for terminal cancer patients in Rhode Island. Inquiry 21:328–337. Stommel M, Given MW, Given BA. 1993. The cost of cancer home care to families. Cancer 71:1867–1874. Suter LG, Nakano CY, Elmore JG. 2002. The personal costs and convenience of screening mammography. J Womens Health Gend Based Med 11:667–672. Swan J, Breen N, Coates RJ, Rimer BK, Lee NC. 2003. Progress in cancer screening practices in the United States: results from the 2000 National Health Interview Survey. Cancer 15:1528–1540. Taplin SH, Barlow W, Urban N, Mandelson MT, Timlin DJ, Ichikawa L, Nefcy P. 1995. Stage, age, comorbidity, and direct costs of colon, prostate, and breast cancer care. J Natl Cancer Inst 87:417–426. Thorpe KE, Howard D. 2003. Health insurance and spending among cancer patients. Health Affairs 22:189–198. Tollestrup K, Frost FJ, Stidley CA, Bedrick E, McMillan G, Kunde T, Petersen HV. 2001. The excess costs of breast cancer health care in Hispanic and non-Hispanic female members of a managed care organization. Breast Cancer Res Treat 66:25–31. Urban N, Baker M. 1989. The women’s health trial as an investment. Med Decis Making 9:59–64. Wagner JL, Alberts SR, Sloan JA, Cha S, Killian J, O’Connell MJ, Van Grevenhof P, Lindman J, Chute CG. 1999. Incremental costs of enrolling cancer patients in clinical trials: a population-based study. J Natl Cancer Inst 91:847–853. Warren JL, Klabunde CN, Schrag D, Bach PB, Riley GF. 2002. Overview of the SEER-Medicare data: content, research applications, and generalizability to the United States elderly population. Med Care 40: IV-3–IV-18. Warren JL, Brown ML, Fay MP, Schussler N, Potosky AL, Riley GF. 2001. Costs of treatment for elderly women with early-stage breast cancer in fee-for-service settings. J Clin Oncol 20:307–316. Wolfson MC. 1994. POHEM—a framework for understanding and modelling the health of human populations. World Health Stat Q 47:157– 176. Yabroff KR, Warren JL, Knopf K, Davis WW, Brown ML. 2005. Estimating patient time costs associated with colorectal cancer care. Medical Care 43:640–648.

Appendix 12–A. Cost of Cancer Care Studies: Patient Perspective Study

Type of Costs Measured

Population/Data Source

Bradley et al. (2002a, 2002b)

Indirect medical costs: employment status, usual hours of weekly work, and earnings in cases and noncases

Breast

Calhoun et al. (2001)

Direct medical costs: out-of-pocket costs Indirect costs: time lost from work for patient and caregivers

Chirikos et al. (2002)

Indirect medical costs: days lost from work, earnings, income, and assets in cases compared to controls

Sherman et al. (2001)

Direct medical costs: inpatient and outpatient medical services, out-ofpocket expenditures, home physical/ occupational therapy Direct nonmedical costs: transportation and parking Indirect costs: lost wages of patient and caregiver Indirect costs: patient travel, waiting, and visiting time

Approximately 150 women aged 51–61 (or with spouses aged 51–61) reporting a breast cancer diagnosis and more than 5500 noncases reporting no breast cancer diagnosis, 1992 83 Patients receiving chemotherapy with neurotoxicity, neutropenia, or thrombocytopenia; completed resource utilization questionnaires Costs estimated from a variety of sources for each type of service 105 Cases identified from tumor registry of single cancer center 5 years or more after diagnosis and 105 “peer-nominated” controls completed an in-person interview 20 Patients with androgen-independent prostate cancer entering a Phase II clinical trial completed a questionnaire multiple times for at least 6 months (68% with complete data); trial started in 1995 Charges for inpatient and outpatient services collected from billing department 105 Women aged 18+ with a Pap smear at six Planned Parenthood clinics completed self-administered questionnaire about specific visit, 1998 (70% with complete data)

Cervix (screening)

Shireman et al. (2001)

Tumor Site

Ovarian

Breast; at least 5 years after diagnosis Prostate

Appendix 12–A. (cont.) Study Secker-Walker et al. (1999)

Stommel et al. (1993)

Type of Costs Measured Direct medical costs: out-of pocket expenses for medical diagnoses and treatment Indirect costs: patient time away from work or home on travel for service utilization Direct medical costs: out-of-pocket expenditures Indirect costs: caregiving time and lost earnings of patient and family members

Sofaer and Davidson (1990)

Direct medical costs: out-of-pocket costs faced by Medicare beneficiaries with and without Medigap policies

Bloom et al. (1985)

Direct medical costs: hospital, physician, pharmaceuticals, and out-of-pocket expenses Direct nonmedical costs: transportation, diets, tutors, and home renovations Indirect medical costs: lost wages Direct, nonmedical costs: food, lodging, family care (e.g., babysitting, housekeeping) Indirect costs: lost wages of the patient and family Direct nonmedical costs for pediatric cancer patients families: patient care, transportation, food, lodging while child hospitalized; clothing; family care (e.g., siblings) Indirect medical costs: hours of work missed and family wages

Houts et al. (1984)

Lansky et al. (1979)

Population/Data Source

Tumor Site

465 Women with a breast biopsy interviewed 11 months (±4 months) afterward about detection/ diagnosis, treatment, and follow-up visits, 1992–1995 (49% with complete data)

Breast; benign breast disease

192 Patients with solid tumors or lymphoma undergoing treatment and their primary caregiver completed self-administered questionnaires or telephone interview about the 3-month prior period

Breast, colon, rectum, lung, gastrointestinal tract, gynecologic system, prostate; lymphoma and other cancers Breast and lung

Profiles of diagnostic and treatment services developed by physicians. Charges estimated by type of service and compared to Medicare Part A and B benefit structure. Out-ofpocket costs estimated from deductibles, copayments, and excess charge over service limits; both with and without physician acceptance of assignment (i.e., Medicare-approved charge as payment in full) Families of 569 pediatric cancer patients receiving any care in 1981 at a referral center completed weekly logs. Hospital records were also obtained

139 Patients receiving outpatient chemotherapy completed 1-week self-administered diaries

70 Families of pediatric cancer patients in treatment at a University Medical Center reported weekly expenses

Leukemia, lymphoma, bone, soft-tissue sarcoma, brain, Wilms’ tumor, neuroblastoma, germ cell, retinoblastoma, others Breast, lung, colon, non-Hodgkin’s lymphoma, other cancers Leukemia, lymphoma, sarcoma, other pediatric solid tumors

Appendix 12–B. Cost of Cancer Care Studies: Payer Perspective Study Ramsey et al. (2002)

Warren et al. (2002)

Barlow et al. (2001)

Tollestrup et al. (2001) Fireman et al. (2000)

Type of Costs Measured Long-term (6–11 years after diagnosis) direct medical costs: hospital stays, skilled nursing facility stays, physician and laboratory services, hospital outpatient claims, home health and hospice care Direct medical costs: inpatient hospitalizations, skilled nursing facility (SNF), outpatient hospital services, physician and supplier services, hospice and home health care Total and cancer-specific costs: by phase of care Direct medical costs: inpatient care, primary care, specialty care, ambulatory surgery, radiology, pharmacy and other services. Costs of care: by type of treatment Direct medical costs: hospitalizations and outpatient procedures, categorized by type of service Direct medical costs: hospital services, chemotherapy outpatient clinic services including pharmacy, laboratory, imaging, and home health services. Patient co-payments also included

Population/Data Source

Tumor Site

Cases diagnosed at age 65+ identified from linked SEER–Medicare 1984–1994. Controls selected from 5% sample of Medicare enrollees in SEER areas, 1986–1994

Colorectal cancer patients surviving at least 5 years

More than 50,000 cases identified from linked SEER– Medicare claims and controls identified from 5% Medicare random sample without cancer Costs reported from 1990–1998 for cases identified 1983–1996

Breast, early stage only

1675 Women aged ≥35 identified from linked SEERmanaged care files, 1990–1997

Breast: early stage cancer

317 Cases identified from Lovelace Tumor Registry (participant in SEER) and 949 controls from any of Lovelace-managed care plans, 1989–1996. 135 Cancer patient cases in clinical trials at Kaiser Permanente and 135 matched cancer patient controls identified from tumor registry and Kaiser Permanente HMO, 1994–1996 Costs and utilization obtained from medical charts and Kaiser Permanente databases

Breast cancer Breast, melanoma, ovary, colon, stomach, brain, kidney, lung; lymphoma

(continued)

211

Appendix 12–B. (cont.) Study Brown et al. (1999)

Helms and Melnikow (1999)

Penberthy et al. (1999)

Type of Costs Measured Direct medical costs: inpatient hospitalizations, SNF, outpatient hospital services, physician and supplier services, hospice and home health care Cancer-specific costs: by phase of care and estimates of long-term Medicare payments Direct medical costs for cervical cancer prevention: medical staff time, supplies, and laboratory costs; specialized equipment; overhead Direct medical costs for cervical cancer treatment: hospitalizations and outpatient care, laboratory work, home services Direct medical costs: hospitalization, physician and other professional services

Wagner et al. (1999)

Direct medical costs: physician and hospital services. Incremental or attributable costs reported

Hillner et al. (1998)

Direct medical costs: insurance payment, patient co-payment, and deductibles for hospital and physician providers as well as medications, durable goods, hospice Direct medical costs: inpatient care, ambulatory surgery; outpatient visits to physicians and nonphysicians; referrals; SNFs; some durable medical equipment Estimates by phase of care, stage of disease at diagnosis, type of service Direct medical costs: inpatient and skilled nursing facility costs, home health and hospice services, physician services, outpatient services Direct medical costs: inpatient, specialist, outpatient costs

Fireman et al. (1997)

Etzioni et al. (1996)

Legorreta et al. (1996) Simon et al. (1996)

Direct medical costs: office visits, radiologic and laboratory tests for clinical surveillance

Riley et al. (1995)

Direct medical costs: inpatient hospitalization, physician/supplier; outpatient, home health, skilled nursing facility, hospice services in cancer patients Estimates by phase of care, stage at diagnosis, type of service Direct medical costs in cases vs. controls: inpatient, physician, personnel, and outpatient services. Estimates by phase of care, stage at diagnosis, age, level of co-morbidity Direct medical costs: hospital accommodation, laboratory tests, radiology and diagnostic tests, antimicrobial therapy, physician consultations, blood products; repiratory, physical, and occupational therapy; surgical and invasive procedures Lifetime direct medical expenses: inpatient hospital stays, home health care, SNF stays, physician services, outpatient and other services

Taplin et al. (1995)

Gulati and Bennett (1992)

Baker et al. (1991)

212

Population/Data Source

Tumor Site

More than 15,000 patients aged ≥65 identified from linked SEER–Medicare Claims and more than 20,000 controls identified from 5% Medicare random sample without cancer Costs reported during 1990–1994 for cases identified during 1983–1993

Colon, rectum

Clinical protocols for Pap smear, colposcopy, and cryotherapy used in a system of family planning clinics and time-and-motion study of clinic staff, payroll information, and cytology and pathology services and supplies for 98 cervical cancer patients and 133,058 controls treated in an HMO

Cervix: prevention and treatment services

More than 10,000 cancer patients aged ≥65 during the first year following diagnosis identified from state tumor registry linked to Medicare claims, 1985–1988 Cancer patients identified by county tumor registry— 61 cancer cases entered Phase II or III treatment trials and 61 cancer controls were eligible but did not enter trial, 1988–1994. Provider billing data for physician and hospital services was collected from a county-wide utilization database. Costs estimated based on standardized unit costs 336 Patients younger than 65 with lung cancer identified from linked Virginia state registry–Blue Cross and Blue Shield, 1989–1991

Breast, colon, rectum, lung, prostate

More than 20,000 patients identified from California Bay area SEER registry linked to Kaiser Permanente utilization files, 1987–1991. Controls selected from utilization files as well. Costs estimated from utilization and unit costs derived from accounting and cost reports 5012 Ovarian cancer patients identified from linked SEER–Medicare. Similar controls without cancer diagnoses, 1973–1989

Breast, colon, rectum, lung, ovary, prostate; non-Hodgkin’s lymphoma

200 Breast cancer cases identified with claims algorithm and review of breast cancer screening program in an HMO in 1989 222 Breast cancers diagnosed at a university-affiliated hospital identified through medical records, 1989–1991. Services collected from cost accounting department, outpatient billing department, and medical record and costs estimated with Medicare relative value units (RVUs) More than 250,000 cancer patients aged ≥65 identified from SEER–Medicare linked data, 1984–1990

Breast

Gastrointestinal tract, genitourinary tract, breast, lung, central nervous system, blood, head/lymphatics Non-small-cell lung cancer

Ovary

Breast (stage I or II cancer)

Bladder, breast, colon, rectum, lung, prostate

1489 Incident and 4640 prevalent cases identified from Puget Sound Tumor Registry (part of SEER) and matched to Group Health Cooperative enrollment files in 1990–1991. Noncancer enrollees served as controls. 24 Patients in Phase III clinical trial of high-dose chemotherapy with bone marrow transplantation at a single cancer center Costs estimated from service utilization from medical record and hospital bills

Breast, colon, prostate

Diagnosis codes from hospitalization claims used to identify more than 20,000 cases, Continuous Medicare History Sample File (CMHSF), 1974–1981. Controls selected after exclusion of individuals with hospitalization claims with cancer diagnosis

Breast, lung

Hodgkin’s disease

Appendix 12–B. (cont.) Study

Type of Costs Measured

Sofaer et al. (1990)

Direct medical costs: hospitalizations, medications, medical equipment, nursing homes, home health services, out-ofpocket costs

Riley et al. (1987)

Direct medical costs: physician, hospital outpatient, and home health services; selected information on inpatient hospital and skilled nursing facility stays

Mor and Kidder (1985)

Direct medical costs: hospice services and conventional oncologic care, excluding physician services

Spector and Mor (1984)

Direct medical costs: hospitalizations, nursing homes, home health, physician services, outpatient clinics, hospice

Scotto and Chiazze (1977)

Direct medical costs: inpatient hospitalization

Population/Data Source Profiles of diagnostic and treatment services developed by physicians Charges estimated by type of service and compared to Medicare Part A and B benefit structure. Total charges, Medicare share, and out-of-pocket costs estimated from deductibles, co-payments, and excess charge over service limits; both with and without physician acceptance of assignment (i.e., Medicareapproved charge as payment in full) Underlying cause of death identified death certificate data linked to Continuous Medicare History Sample (CMHS) file for more than 200,000 cancer deaths in 1979. Sample of persons alive at end of 1979 linked to CMHS were used as comparisons group More than 10,000 terminal cancer patients identified from hospices and conventional oncologic care 1980–1983; part of the National Hospice Study. Bill summary data from Medicare 2104 Terminal cancer patients identified from death certificates in 1980–1981 in a single state linked to Blue Cross/Blue Shield claims during 6 months preceding death Cancer cases identified from the Third National Cancer Survey conducted in 1969. Hospitalization data (>1.5 million admissions) were collected for approximately a 10% sample of cases

Tumor Site Breast, lung

Cancer patients during last year of life

Terminal cancer patients

All tumor sites

All sites

Appendix 12–C. Cost of Cancer Care Studies: Societala Perspective Study

Type of Costs Measured

Population/Data Source

Max et al. (2003)

Direct medical costs: hospitalizations, physician services, emergency department and outpatient visits, prescription medication, home health care services, nursing home costs Indirect mortality costs

Thorpe and Howard (2003)

Direct medical costs: inpatient hospital services, outpatient services, physician services, out-ofpocket spending Direct medical costs: hospitalizations, physician services, emergency department visits, outpatient, prescription medication, home health care services, nursing home costs. Indirect mortality costs

Cervical, ovarian, and uterine cancer diagnoses from 1998 California Hospital Discharge; 1997 National Nursing Home Survey; 1997 Medical Expenditure Panel Survey (MEPS). Cervical, ovarian, and uterine cancers as underlying cause of death from the 1998 California Mortality File combined with estimate of present value of lifetime earnings estimate National sample of 1383 cancer patients identified by cancer condition codes from the MEPS, 1996–1999

Max et al. (2002)

Sandler et al. (2002)

Lawrence et al. (2001)

Barnett et al. (2000)a

Direct medical costs: inpatient hospitalization, outpatient hospital, physician visits, emergency care, pharmaceuticals Direct nonmedical costs: patient time associated with receiving care Direct medical costs: genetic counseling including counselor time for counseling and preparation, office space, cost of BRCA1/ BRCA2 test, phlebotomy services Direct nonmedical costs: travel time to counseling and care for family members Direct medical costs: cancer treatment and other costs Indirect costs: for absenteeism and disability

Prostate cancer diagnoses from 1998 California Hospital Discharge; cancer diagnoses in men from 1997 MEPS (with prostate/all cancers ratio) 1997 National Nursing Home Survey, and prostate cancer as underlying cause of death from the 1998 California Mortality File combined with estimate of present value of lifetime earnings estimate Diagnostic codes used to estimate service utilization in the National Hospital Discharge Survey, National Ambulatory Medical Care Survey, National Hospital Ambulatory Medical Care Survey, and costs estimated from claims database 191 Patients undergoing genetic counseling for BRCA1/ BRCA2 testing. Genetic counselors completed time estimates for counseling, follow-up calls, and documentation time. Patients completed a written survey Cancer patients and controls identified from diagnosis codes of health care and disability data for active employees from a single corporation, 1995–1997

Tumor Site Cervix, ovary, uterus

Lung, skin, breast, cervix, prostate Prostate

Colon rectum, liver, pancreas

Individuals undergoing testing for BRCA1/BRCA2 cancer susceptibility mutations All cancers

(continued)

213

Appendix 12–C. (cont.) Study

Type of Costs Measured

Population/Data Source

Tumor Site

Hodgson and Cohen (1999)

Direct medical costs: hospital care, physician and other professional services, home health care, drugs, other medical nondurables, nursing home care, other personal health care

All malignant neoplasms

Schulman et al. (1998)

Direct medical costs: transplantation including hospitalization and physician payments, subsequent care including rehospitalizations, chemotherapy, radiation therapy, transfusions, provider visits, outpatient surgery and procedures Intangible costs measured with utility (EuroQol)

Physician visits to hospital inpatients, National Hospital Discharge Survey for inpatient stays, National Ambulatory Medical Care Survey for office-based physician visits, National Hospital Ambulatory Medical Care Survey for emergency room and hospital outpatient care, and National Nursing Home Survey for nursing home residents and discharges 115 Patients enrolled in a multicenter Phase III clinical trial of supportive care (IL-3 and GM-CSF) following high dose chemotherapy and bone marrow transplantation, 1993–1995

a

Studies included several components of costs from a societal perspective, but none included all types of costs from a societal perspective.

214

Hodgkin’s or nonHodgkin’s lymphoma

III THE CAUSES OF CANCER

This page intentionally left blank

13

Tobacco MICHAEL J. THUN AND S. JANE HENLEY

T

obacco use is the largest single recognized cause of human cancer in Western countries. Cigarette smoking alone accounts for about 30% of all cancer deaths in the United States (Doll and Peto, 1981; Centers for Disease Control and Prevention, 2002a) and an estimated 16% of all cancers worldwide (Parkin et al., 1999). In addition to cancer, cigarette smoking causes even more deaths from nonmalignant cardiovascular and respiratory diseases than from cancer (Forey et al., 2002; IARC, 2004). Tobacco smoking causes an estimated 4.9 million deaths annually worldwide (Ezzati and Lopez, 2003b). If current smoking patterns continue, the growing toll is expected to exceed 10 million deaths annually during the twenty-first century (Peto and Lopez, 2001). The extraordinary burden of cancer and other diseases caused by tobacco products is even more remarkable in that the pandemic is entirely manmade. Tobacco is the only major human carcinogen for which a combination of commercial marketing and physical addiction sustains the epidemic. Tobacco marketing creates the illusion that smoking is desirable and thus encourages nonsmokers to experiment with tobacco usage. Physical addiction to nicotine then obligates continued usage, making it extremely difficult for many smokers to quit. Chronic exposure to the numerous carcinogens and other toxic substances in tobacco is essentially a by-product of the quest to maintain nicotine intake. Epidemiologic studies have revealed much of what has been learned about the deleterious health effects of tobacco over the last half century (Thun et al., 2002). Humans, unlike other mammals, use tobacco voluntarily. Whereas tobacco smoke is irritating and highly toxic to other species exposed experimentally in laboratory studies, nonsmokers who experiment with tobacco persist though the initial noxious effects until they develop physical and psychological dependence on tobacco (Thun et al., 2002). Epidemiologic (nonrandomized) studies continue to be ethical and feasible, whereas randomized clinical trials of tobacco are not. Epidemiologic studies are better suited than clinical observations to surmount the long delay between the initiation of tobacco use and the onset of cancer and other chronic diseases. The association of tobacco use with various diseases is sufficiently strong that epidemiologic studies have been able to identify unequivocally many diseases caused by smoking despite imprecise quantification of lifetime tobacco exposure (Thun et al., 2002). This chapter considers the factors that transformed tobacco use from a ceremonial practice in pre-Columbian times to a global epidemic. It also discusses the role of nicotine addiction in sustaining and modifying exposure to the carcinogens in tobacco, the cancers caused by various forms of tobacco use, the global burden of tobaccoattributable disease, the extent to which design changes in cigarettes have altered their pathogenicity, the influence of genetic and other factors on susceptibility to addiction or carcinogenesis, and finally the immense opportunities for prevention.

HISTORY OF TOBACCO USE TO THE MID-TWENTIETH CENTURY Although tobacco leaves were burned in religious ceremonies and smoked and chewed for pleasure throughout precolonial North and South America (Doll, 1998a), these practices were unknown to Europeans prior to the voyages of Columbus. Tobacco was introduced

into Europe by Spanish explorers returning from the New World during the late fifteenth century. The main species of tobacco was named Nicotiana tobaccum after Jean Nicot, the French ambassador to Portugal who sent tobacco seeds to his queen, Catherine de Medici. Use of tobacco for medicinal purposes and as a curiosity was promoted first in Spain and later England (Doll, 1998a). Recreational pipe smoking subsequently spread from England to many countries in Europe and Asia. By the late nineteenth century, tobacco was widely used in Europe (Doll, 1998a) and the United States (Fig. 13–1) (U.S. Department of Agriculture, 2002) in the form of cigars, pipes, roll-your-own cigarettes, chewing tobacco, and snuff. Several new technologies converged at the beginning of the twentieth century to allow manufactured cigarettes to displace traditional tobacco products and to increase total tobacco consumption (Slade, 1989, 1993). Portable paper safety matches, patented in 1889 (Slade, 1989), made it possible to smoke tobacco frequently throughout the day in diverse settings. Cigarette-rolling machines were developed during the 1880s that could mass-produce and package cigarettes that were considerably less expensive than hand-rolled products. New strains of tobacco and new curing processes were developed that produced less irritating smoke that could be inhaled. These innovations were coupled with novel and aggressive advertising campaigns that glamorized smoking of particular brands of cigarettes, beginning with Camel cigarettes in 1913 (Slade, 1993). Free cigarettes were distributed in military rations to allied soldiers during World Wars I and II. Consequently, manufactured cigarettes became the predominant form of tobacco used in the United States (see Fig. 13–1) and most other Western countries by the mid-twentieth century. In addition to increasing substantially the number of people who used tobacco and the number of cigarettes consumed per person, early changes in cigarette design increased the surface area of respiratory epithelium exposed to the carcinogens in cigarettes. As mentioned, the popularization of manufactured cigarettes required adoption of new strains of tobacco leaf and new curing processes that released a milder, less irritating smoke (Doll, 1998a). The smoke from traditional tobacco products was highly alkaline and released nicotine in an unionized form that could be absorbed through the oropharyngeal mucosa (Henningfield et al., 1993; Slade, 1993). In contrast, ionized nicotine from manufactured cigarettes had to be inhaled into the trachea and large bronchi to facilitate rapid absorption. Whereas traditional tobacco products caused intense local contact of the lip and oropharyngeal tissues with tobacco leaf and tobacco carcinogens dissolved in saliva, inhaled smoke caused more extensive exposure of the larynx, trachea, and large bronchi. This compounded the preexisting risk of cancers of the oral cavity and pharynx with a massive increase in the risk of cancers of the trachea, bronchus, and lung (Thun et al., 2002).

PRESENT BURDEN OF TOBACCO-ATTRIBUTABLE MORTALITY The disease burden attributable to smoking is not static but varies with the number of people smoking and the duration and intensity of regular smoking in the population. In the United States, where manufactured cigarettes were introduced early during the twentieth century, cigarette smoking alone accounts for approximately 440,000 deaths each year

217

218

PART III: THE CAUSES OF CANCER

Figure 13–1. Adult (age ≥18 years) per capita consumption of various forms of tobacco in the United States, 1880–2000. (Source: Adapted from

NCI Smoking and Tobacco Control Monograph 8, 1997, p. 13. Data are from the U.S. Department of Agriculture.)

(Table 13–1) (Centers for Disease Control and Prevention, 2002a; U.S. Department of Health and Human Services, 2004), 20% of all deaths (McGinnis and Foege, 1993), and about 30% of cancer deaths (Doll and Peto, 1981). Smoking also accounts for an estimated $75.5 billion in health care costs in the United States (Centers for Disease Control and Prevention, 2002a). Lopez and colleagues at the World Health Organization (WHO) diagrammed the evolution of the epidemic of smoking and resultant disease in many countries (Fig. 13–2) (Lopez et al., 1994). The use of manufactured cigarettes typically increases first among young male adults. Stage 1 of the WHO paradigm is characterized by a low (<20%) prevalence of male cigarette smoking with as yet no apparent increase in smoking by women or rise in smoking-attributable diseases. Stage 2 of the epidemic is characterized by increases in smoking prevalence to more than 50% among men, early increases in smoking prevalence among women, a shift toward initiation at younger ages, and an increasing burden of lung cancer and other tobacco-attributable disease in men. Tobacco control activities are usually not well developed, and the health risks of tobacco are not well understood. Stage 3 of the epidemic is characterized by a marked downturn in smoking prevalence among men, a more gradual decline in women, and convergence of male and female smoking prevalence. Paradoxically, the burden of smoking-attributable disease and death continue to increase. Smoking-attributable deaths comprise 10%–30% of all deaths, about three-fourths of them in men. Health education about the diseases caused by smoking begins to decrease public acceptance of smoking among more educated subgroups of the population (Lantz et al., 1998). In stage 4 of the epidemic, smoking prevalence continues to decrease in both men and women. Deaths attributable to smoking in men peak at 30%–35% of all deaths (40%–45% of deaths in middle-aged men) and subsequently decline. Among women, smoking-attributable deaths rise to about 20%–25% of all deaths. Many industrialized countries are in or approaching this stage. However, even these countries vary considerably in their progress against tobacco and in their ability to sustain a national commitment to reduce tobacco use. Worldwide, the burden of smoking-attributable disease is shifting from the developed to the developing world. The number of deaths caused by smoking globally in the year 2000 was approximately 4.9 million, with roughly equal numbers in economically developed and developing countries (Mackay and Eriksen, 2002; Ezzati and Lopez,

2003a). By the years 2025–2030, if current smoking patterns persist, the burden is expected to be seven million and three million deaths, respectively, in developing and developed countries (Mackay and Eriksen, 2002). This rapid increase reflects the much larger number of current smokers who live in low- and middle-income countries (933 million) than in high-income countries (209 million) (Jha and Chaloupka, 2000). The only other exposures whose impact is known to be increasing with such rapidity are human immunodeficiency virus (HIV) infection and, in Western countries, obesity (Gonzalez et al., 2003). Not all countries follow the exact course of the Lopez et al. model (see Fig. 13–2). In China, for example, the prevalence of smoking among women has remained below 5% despite a high prevalence of cigarette smoking among men for several decades (Corrao et al., 2000). Countries such as Thailand have had markedly reduced per capita cigarette consumption early in the epidemic because of national policies that ban cigarette marketing and discourage smoking. The mix of diseases caused by smoking also varies depending on background risk in various countries. In the United States cardiovascular diseases account for an estimated 32% of deaths caused by active smoking (see Table 13–1) (U.S. Department of Health and Human Services, 2004). In China, where cardiovascular disease risk is generally low, smoking causes more premature deaths due to liver cancer than to heart disease (Liaw and Chen, 1998). In India, smoking appears to cause more deaths from tuberculosis than from any other condition (TATA Institute of Fundamental Reasearch, WHO, and CDC, 2000). Nevertheless, the WHO paradigm illustrates the natural history of the epidemic and its protracted course in the absence of effective national and international tobacco regulation.

EPIDEMIOLOGY OF TOBACCO USAGE The global consumption of manufactured cigarettes increased more than 100-fold during the twentieth century, reaching about 5500 billion cigarettes per year in the year 2000 (Fig. 13–3) (Mackay and Eriksen, 2002). The number of people who smoked tobacco worldwide in 2000 was approximately 1.3 billion (Shafey, 2003). Because no national surveys of smoking prevalence were conducted in the United States prior to 1955, historical information on the rise in cigarette smoking

219

Tobacco Table 13–1. Smoking-Attributable Mortality (percent and number of deaths, US 1995–1999) by Disease Relative Risk Estimates

Disease (ICD-9 Code)

Men

Attributable Deaths

Year First Considered

Year Formally Classified

Women

%

No.

Current

Former

Current

Former

Men

Women

Men

Women

1964 1964 1964 1967 1964 1964 1982 1964 1968 1990

1964/71* 1982 2004 1982 1964 1964/68* 2004 1979 1982 2004

10.9 6.8 2.0 2.3 14.6 23.3 NA 3.3 2.7 1.9

3.4 4.5 1.5 1.2 6.3 8.7 NA 2.1 1.7 1.3

5.1 7.8 1.4 2.3 13.0 12.7 1.6 2.2 1.3 1.1

2.3 2.8 1.3 1.6 5.2 4.5 1.1 1.9 1.1 1.4

75 73 29 23 83 88 NA 47 39 25

50 57 11 24 75 72 12 29 5 11

3,900 6,300 2,200 3,100 2,500 80,600 NA 3,700 2,800 800

1,300 1,600 600 3,400 600 44,200 500 1,100 200 300

1964

1968 2.8 1.5 1.8

1.6 1.2 1.2

3.1 1.6 1.6

1.3 1.2 1.2

42 15 19

37 11 9

22,100 29,300 18,800

7,100 23,500 10,500

1.0 1.0 1.3 3.1 1.0

4.0 1.5 1.8 7.1 2.2

1.3 1.0 1.0 2.1 1.1

40 9 27 65 15

44 6 9 50 15

3,900 4,700 1,600 6,500 700

3,600 5,300 900 3,100 900

1.4 15.6 6.8

2.2 12.0 13.1

1.1 11.8 6.8

23 91 82

14 81 75

8,800 9,900 34,900

6,800 7,800 29,800 410

cancers Lip, oral cavity, pharynx (140–150) Esophagus (150) Stomach (151) Pancreas (157) Larynx (161) Trachea, lung, bronchus (162) Cervix uteri (180) Urinary bladder (188) Kidney, other urinary tract (189) Acute myeloid leukemia (205)

cardiovascular disease Coronary heart disease (410–414) Age 35–64 Age 65+ Other heart disease (390–8, 415–7, 420–9) Cerebrovascular disease (430–438) Age 35–64 Age 65+ Atherosclerosis (440) Aortic aneurysm (441) Other arterial disease (442–448)

1964 1964

1973 1989

1964 1964 1964

1973 1979 1979

3.3 1.6 2.4 6.2 2.1

1964 1964 1964

2004 1964/1967* 1967

1.8 17.1 10.6

1964

1969

10

10

560

1964

1964

24

25

590

380

1972

1986

—

—

15,500

22,500

20

20

263,600

176,500

respiratory disease Pneumonia and influenza (480–487) Bronchitis, emphysema (491–492) Chronic airways obstruction (496)

pediatric diseases (765, 769, 770, 798.0) burn deaths (890–899) environmental tobacco smoke Total

Source: U.S. Department of Health and Human Services (2002). Reports of the Surgeon General on the Health Consequences of Smoking, 1964–2004. Morb Mortal Wkly Rep MMWR 51:300–303. *Lip cancer was classified as causal in 1964 other oropharngeal cancers in 1971. Lung cancer was classified as causal in men in 1964 and in women in 1968. Bronchitis was classified as causal in 1964; other chronic obstructive pulmonary diseases in 1967.

is limited to estimates of per capita consumption based on cigarette sales and census population data. Figure 13–4 illustrates that per capita consumption had already begun to increase in the United Kingdom by 1905 and in the United States by 1910. In contrast, postwar economic conditions delayed the major increase in consumption in Japan until the 1960s and in China until the 1970s (Forey et al., 2002). Per capita cigarette consumption is now decreasing in most Western countries but continuing to increase in many economically developing countries (Corrao et al., 2000). The first national survey of smoking prevalence among adults in the United States was conducted in 1955, when 57% of men and 28% of women age 18 years and older reported current cigarette smoking (Haenszel et al., 1956). Smoking prevalence was even higher in Britain between 1948 and 1952, where nearly 70% of men and more than 40% of women between the ages of 25 and 59 smoked cigarettes (Peto et al., 2000). Adult smoking prevalence has decreased in the United States since the 1964 U.S. Surgeon General Report on Smoking and Health (U.S. Public Health Service, 1964). The crude prevalence of current cigarette smoking in men, age ≥18 years, decreased from 51.9% in 1965 to 25.7% in 2000 (National Center for Health Statistics, 2002). The corresponding decrease in smoking prevalence among women was from 33.9% in 1965 to 21.0% in 2000. Although the overall prevalence of smoking among adults age ≥18 years decreased in many affluent countries over the last half century, the age at which smokers initiated the habit became progressively younger. Table 13–2 shows the average age of initiation among male and female smokers in birth cohorts from 1870 to 1970

in two American Cancer Society cohorts enrolled in 1959 and 1982 (Thun et al., 1997a, 2002) and in two National Health Interview surveys conducted during 1987–1988 (Burns et al., 1997a) and 1998 (http://www.cdc.gov/nchs/nhis.htm). A similar pattern in which successive generations begin smoking at progressively early ages has occurred in many other countries (U.S. Department of Health and Human Services, 1994). Most smokers in developed countries become addicted to tobacco use during adolescence. Hence the prevention of smoking initiation by adolescents is one of the critical goals of tobacco control (U.S. Department of Health and Human Services, 1994).

Birth Cohort Patterns in Smoking Prevalence The increase in smoking prevalence is not uniform across all age groups of the population but follows clear birth cohort patterns that reflect smoking initiation during the critical periods of adolescence and young adulthood (Giovino et al., 1995; Thun et al., 2002). Within each birth cohort (5- or 10-year interval of birth year), the uptake of smoking reflects social norms, peer behavior, tobacco marketing, and economic conditions that prevail during this period of vulnerability. In the United States, widespread cigarette smoking was adopted first by white men and then black men (Burns et al., 1997b). Male smoking prevalence increased across successive birth cohorts after 1885–1889, peaking among men born during 1925–1929 and then decreasing in later cohorts because of the growing publicity during the 1950s about the adverse health effects of smoking. Widespread cigarette smoking

220

PART III: THE CAUSES OF CANCER

Figure 13–2. Four stages of the tobacco epidemic. A descriptive model of the cigarette epidemic in developed and developing countries. (Source: Lopez et al., 1994.)

among women lagged behind that in men, with peak prevalence occurring in birth cohorts during 1930–1934 and 1935–1939. The increase in smoking initiation among adolescent girls in the United States around 1967 coincides with the introduction of several women’s brands and correlates strongly with increasing expenditures for tobacco advertising and promotion (Pierce et al., 1994; Giovino et al., 1995). A limitation of birth cohort analyses of smoking prevalence in the United States is that historical data must be reconstructed from surveys beginning in 1965 (Harris, 1983; Burns et al., 1997a) because national surveys were not conducted during the first half of the twentieth century. These reconstructions illustrate the progressive wave-like uptake of cigarette use by successive birth cohorts of Americans, but their quantitative accuracy can be questioned. Contemporary documentation of birth cohort increases in smoking is unfortunately still possible in countries where cigarette smoking is increasing, where

birth cohort trends can be monitored as they occur. These trends have important implications for understanding and communicating the epidemiology of tobacco-attributable diseases. Even large increases in age-specific smoking prevalence among young adults initially have little impact on the crude or age-adjusted prevalence for all adults. The full consequences of current smoking practices are not reflected in national lung cancer rates or in analytic studies of smoking until lifetime smoking patterns have been entrenched in a population for approximately 50 years (IARC, 2004). Studies that measure the risks from continued smoking or the benefits from smoking cessation underestimate these parameters in countries where the uptake of smoking has occurred more recently. Birth cohort patterns of cigarette smoking also explain why the downturn in lung cancer death rates seen in older age groups has occurred progressively later over time. In the United States, the age group with the highest lung cancer death rate in white men was age 65–69 in 1950–1959, age 70–74 in 1962–1969, age

Figure 13–3. Global cigarette consumption: billions of sticks, 1880–2000. (Source: The Tobacco Atlas. Geneva: World Health Organization, 2002. Adapted and reproduced with permission from WHO.)

221

Tobacco

Figure 13–4. Daily adult per capita cigarette consumption in the United States, United Kingdom, China, and Japan. (U.S. data: MMWR Morb Mortal Wkly Rep 48:986–993, 1999 and Tobacco Situation and Outlook, September 2001. U.K. and Japanese data: International Smoking Statistics. Oxford: Oxford University Press, 2002. Chinese data: Food and Agricultural Organization of the United Nations and the U.S. Department of Agriculture.)

75–79 during 1970–1985, and age 80—85 during 1986–1995 (Wingo et al., 2003). This shift coincides with the aging of birth cohorts with heaviest tobacco use.

TOBACCO PRODUCTS OTHER THAN CIGARETTES Although manufactured cigarettes are the predominant form of tobacco used, 15%–35% of global consumption involves other tobacco products (World Health Organization, 1997). The most common form of smoked tobacco in India involves bidis, traditionally hand-rolled in dried temburni leaf and tied with a string (Mackay and Eriksen, 2002; IARC, 2004). Cigars are defined as shredded tobacco wrapped in tobacco leaf or paper (U.S. Department of the Treasury, 1996). They vary in size from cigarette-sized cigarillos to cheroots and double coronas (Mackay and Eriksen, 2002). Kreteks are clove- and cocoa-flavored small cigars that originated in Indonesia but are available in the United States (IARC, 2004). Chuttas are coarsely prepared small cigars smoked exclusively in India, sometimes with the burning end held inside the mouth, a practice called “reverse smoking” (IARC, 2004). The previously widespread use of pipes is decreasing throughout Europe and the Americas (World Health Organization, 1997). In Moslem countries, tobacco is frequently smoked in water pipes (IARC, 2004).

The use of spit tobacco products predominates in certain populations. Betel chewing is common throughout much of Southeast Asia and the western Pacific. Betel leaves (Piper betle) are mixed with tobacco, areca nut (Areca catechu), lime, wood ash, or other substances to form a quid, pan, or nass (Gupta, 1992; IARC, 2004). The mixture is then chewed and/or retained in the mouth. The use of spit tobacco is estimated to cause nearly 100,000 deaths annually from oral cancer in southern Asia (Ezzati and Lopez, 2003a). In the United States alone, several million people use spit tobacco (Glover and Glover, 1992), mostly as moist snuff (Henningfield, et al., 2002). Moist snuff consists of finely ground tobacco with 20%–55% moisture content, often flavored with mint, wintergreen, or raspberry (Brunnemann and Hoffmann, 1992). A pinch (called a dip or rub) is placed between the gum and the cheek or under the tongue (Glover and Glover, 1992). The use of chewing tobacco remains common among baseball players and some rural populations in the United States. Chewing tobacco products may also be flavored with sugar, molasses, or licorice. In Sweden, a major form of tobacco use involves moist snuff or “snus,” reported to have lower nitrosamine content (Nilsson, 1998). Products other than manufactured cigarettes are often viewed as anachronistic, yet they cause substantial disease and can undergo rapid increases in usage with aggressive marketing. This was evidenced by the resurgence in the use of moist snuff and chewing tobacco by

Table 13–2. Average Age of Initiation Among Cigarette Smokers by Sex and Birth Cohort: CPS-I, CPS-II, NHIS 1987–1988, NHIS 1998 Study

1870

1880

1890

1900

1910

1920

1930

1940

1950

1960

1970

23.8 — — —

21.8 — — —

19.8 19.2 — —

18.9 18.8 — —

18.4 18.2 17.5 —

17.7 17.8 17.2 —

— 17.6 17.1 —

— 17.2 17.0 —

— 17.2 17.0 —

— — — 16.9

— — — 16.5a

42.4 — — —

39.3 — — —

34.2 33.7 — —

27.1 27.1 — —

21.8 22.4 22.9 —

19.8 20.9 21.0 —

— 19.6 19.4 —

— 18.4 18.7 —

— 18.1 17.5 —

— — — 17.2

— — — 16.2a

men CPS-I CPS-II NHIS 87-88 NHIS 98

women CPS-I CPS-II NHIS 87-88 NHIS 98

Sources: NCI Smoking and Tobacco Control Monograph 8, 1997, p. 313 and NHIS 1998 public use tapes (http://www.cdc.gov/nchs/nhis.htm). Thun et al. 2002. Tobacco use and cancer: an epidemiologic perspective for geneticists. Oncogene 21:7307–7325, with permission from Nature Publishing Group. Based on the age of initiation among current cigarette smokers in the analytic cohorts of Cancer Prevention Study I (CPS-I) and CPS-II and published data from the National Health Interview Surveys of 1987 and 1988, whites (Burns, 1994), and unpublished data from the National Health Interview Surveys of 1998, whites. a Age at initiation is biased toward a younger age because it is based on people aged 19–28 years.

222

PART III: THE CAUSES OF CANCER

Table 13–3. Classes of Carcinogens in Tobacco Smoke Class

No.

Aromatic hydrocarbons Monocyclic Polycyclic Aza-arenes N-Nitrosamines

1 10 3 7

Aromatic amines

3

Heterocyclic aromatic amines Aldehydes

8

Organic compounds Inorganic compounds

2 14 7

Example(s) Benzene Benzo[a]pyrene Dibenz[a,h]acridine N-Nitrosodiethylamine 4-Methylnitrosamino-1-(3-pyridyl)-1butanone (NNK) 2-Naphthylamine 4-Aminobiphenyl Formaldehyde Acetaldehyde 1-3-Butadiene Ethyl carbamate Arsenic, cadmium, chromium, hydrazine, nickel, polonium-210

Sources: Adapted from Hoffmann and Hoffmann (1997); IARC (1986), Hecht (1999b).

adolescents in the United States during the 1980s (Connolly et al., 1986) and the increase in premium cigar smoking in cigar bars during the 1990s (Gerlach et al., 1998). There is concern that the use of traditional tobacco products can be a pathway to the initiation or resumption of smoking manufactured cigarettes.

CHEMICAL COMPOSITION OF TOBACCO AND TOBACCO SMOKE Tobacco smoking generates both mainstream smoke (MS), drawn directly from the burning tobacco into the mouth, and sidestream smoke, released from the smoldering tobacco into the ambient air. The latter mixes with exhaled MS to make up environmental tobacco smoke (ETS). Synonyms for ETS exposure include passive, involuntary, or second-hand smoke exposure (IARC, 2004). Tobacco smoke is a complex, heterogeneous mixture that contains approximately 4000 identified chemicals (Hoffmann et al., 2001). Of these, at least 3000 are present in the tobacco leaf (Roberts, 1988); other compounds are generated during curing and/or combustion. At least 55 chemicals present in tobacco smoke are considered established carcinogens based on studies of laboratory animals or humans evaluated by the International Agency for Research on Cancer (IARC) (IARC, 1986; Hoffmann et al., 1997; Hecht, 1999b). Major classes of carcinogens in tobacco leaf and/or smoke are listed in Table 13–3. The composition of tobacco smoke is affected by many factors, including the tobacco leaf, smoking patterns, chemical additives, pH, type of paper and filter, and ventilation (U.S. Department of Health and Human Services, 2002). Numerous carcinogens are generated by combustion of tobacco, including many polycyclic aromatic hydrocarbons, N-nitrosamines, and aromatic amines as well as formaldehyde, phenolic compounds, and a variety of free radicals (IARC, 1986). Other carcinogens, such as arsenic, cadmium, chromium, nickel, and polonium 210 are incorporated into the tobacco plant from soil, pesticides, and phosphate fertilizers. Other compounds accumulate in tobacco during certain curing processes. For example, fermentation increases the concentration of N-nitrosamines in moist snuff and cigars (Hoffmann and Hoffmann, 1997); exposure of flue-cured tobacco to combustion products from gas heaters is also reported to increase tobacco-specific nitrosamines (TSNAs) (Peele et al., 2001). Two TSNAs of particular interest have been NNK (4-methylnitrosamino-1–3-pyridyl-1-butanone) and NNN (N-nitrosonornicotine). Their concentrations are greatly increased by the fermentation of tobacco for use in cigars and moist snuff and by the inclusion of ribs and stems in reconstituted tobacco (Hoffmann and Hoffmann, 1997). TSNAs have been proposed as major contributors to the increase in

adenocarcinoma of the lung in many countries, as these compounds induce adenocarcinoma of the lung in rodents, independent of the route of administration (Hecht, 1999b). Nicotine is the principal alkaloid present in tobacco and accounts for 0.05%–4.00% (by weight) of the tobacco leaf (U.S. Department of Health and Human Services, 1988). Absorption of nicotine from tobacco leaf or smoke is the major factor that induces physical addiction. Although nicotine itself is not carcinogenic, addiction to nicotine sustains tobacco use and prolongs exposure to other carcinogens. Furthermore, nicotine is transformed during curing and combustion to TSNAs, which are carcinogenic (Hecht, 2002a). In cultured lung epithelial cells, nicotine inhibits apoptosis, stimulates cell growth, and may function as a tumor promoter (Minna, 2003). Combustion of tobacco produces an aerosol with a vapor phase (about 90% of the total) and a particulate phase (Hecht, 1999b). The particulate phase is especially rich in carcinogens (Hecht, 1999b). Sidestream smoke contains higher concentrations of nicotine, carbon monoxide, benzene, and several polycyclic aromatic hydrocarbons than does mainstream smoke because combustion is less complete in smoldering than burning tobacco (IARC, 2004). ETS is comprised of a mixture of exhaled mainstream smoke, sidestream smoke, and a small amount of noninhaled smoke released during puffing, diluted in ambient air. ETS exposure is not accurately characterized in terms of “cigarette equivalents” because the ratio of specific components of ETS to nicotine is different in mainstream smoke and ETS. Cotinine concentration in saliva, urine, or serum provides a qualitative indication of recent exposure to ETS but not a quantitative measure of exposure to constituents of smoke other than nicotine (Hecht, 2002a).

ROLE OF NICOTINE ADDICTION IN SUSTAINING TOBACCO USE Physical dependence on nicotine is the critical factor that sustains tobacco use among tobacco users.Nicotine from tobacco binds with the nicotinic receptors for acetylcholine in the central and peripheral nervous systems. In the central nervous system (CNS), the receptors regulate the release of neurotransmitters such as dopamine, serotonin, and g-aminobutyric acid. Exposure to exogenous nicotine stimulates the production of additional nicotine receptors (Benowitz, 1996a). Abstinence from smoking triggers withdrawal symptoms of anxiety, irritability, weariness, constipation/diarrhea, insomnia, intense craving, and difficulty concentrating (Balfour and Fagerstrom, 1996; Arinami et al., 2000). Tobacco products vary in their delivery of nicotine in a form that is rapidly absorbed and pharmacologically active. Cigarettes and moist snuff increase plasma nicotine concentration almost immediately (Fig. 13–5), whereas the nicotine replacement products currently available provide much slower nicotine uptake (Benowitz, 1996a). Inhalation of cigarette smoke increases plasma nicotine and produces discernible CNS effects in as little as 7 seconds owing to the large surface area of the lungs. The uptake of nicotine from moist snuff depends on the pH of the product. Commercial brands produce a pH range in saliva from 8.0 (at which 50% of nicotine is free or un-ionized) to 7.0 (where only 10% of nicotine is un-ionized and can be rapidly absorbed) (Federal Register, 1995). Tobacco products that deliver nicotine rapidly reinforce positive associations with the behavioral aspects of smoking and increase the difficulty of cessation (O’Brien, 2001). For most users, tobacco use results in true drug dependence. Withdrawal symptoms among cigarette smokers who attempt to quit may equal the severity of withdrawal from opiates, amphetamines, and cocaine (Consensus Statement, 2000). The strength of the addiction is illustrated by the high failure rate among smokers who attempt to quit (O’Brien, 2001). Approximately 70% of current smokers express a desire to quit, yet fewer than 50% try to stop each year (Centers for Disease Control and Prevention, 2002b). Unassisted, about 2.5% of smokers succeed in quitting permanently on a single quit attempt (Centers for Disease Control and Prevention, 1994). The success rate approximately doubles with appropriate pharmacologicl and/or behavioral treatment.

Tobacco

223

Figure 13–5. Plasma nicotine concentration by various tobacco products. Venous blood concentrations (nanograms of nicotine per milliliter of blood) as a function of time for various nicotine delivery systems. Data on cigarettes, oral snuff, and nicotine gum are from Benowitz et al.: Clin

Pharmacol Ther 44:23, 1988; data on the nicotine nasal spray are from Schneider et al.: Clin Pharmacokinet 31:65, 1996; data on transdermal nicotine are from Benowitz: Drugs 45:161, 1993. (Source: Adapted from Henningfield et al.: Econ Neurosci 2:42–46, 2000.)

CIGARETTE YIELD AS MEASURED BY MACHINE SMOKING

range in salivary cotinine in people who smoke cigarettes with the same nicotine rating (Fig. 13–7) (Benowitz, 2001; Jarvis et al., 2001). This is because smokers, unlike smoking machines, seek to maintain their accustomed level of nicotine and can compensate for the change in cigarette design by taking larger and more frequent puffs, obstructing the ventilation holes that dilute the mainstream smoke, and inhaling the smoke more deeply into the lungs to increase the surface area for absorption (Djordjevic et al., 2000; Burns and Benowitz, 2001). There has been considerable debate about what impact, if any, design changes that have reduced machine-measured tar ratings have had on the carcinogenicity of cigarettes. This is discussed below in the section on variations in the carcinogenicity of cigarettes.

A complication when measuring exposure in epidemiologic studies, in addition to the role of nicotine addiction in regulating smoking behavior, is the incompletely documented impact of design changes in cigarettes on both the composition of smoke and smoking behavior. The most noticeable has been a series of changes that reduced the “yield” of tar and nicotine, as measured by machine smoking. A standardized method of testing was developed by the tobacco industry during the 1930s (Bradford et al., 1936) and adopted officially by the Federal Trade Commission (FTC) in 1969 to measure the average nicotine and “tar” yield from the various brands of cigarettes (Institute of Medicine, 2001). The FTC protocol specifies that a smoking machine take one 2-second (35 ml) puff per minute until the cigarette is consumed. Tar and nicotine are extracted from a special filter through which the machine has “smoked” the cigarette. Tar represents the total particulate matter after removing the nicotine and water (Institute of Medicine, 2001). The average sales-weighted FTC tar rating of U.S. cigarettes has decreased from 38 mg in 1954 to 12 mg in 1997 (Fig. 13–6) (U.S. Department of Health and Human Services, 1981, 1989; Kozlowski et al., 2001). During the same interval, the average nicotine rating decreased from 2.3 mg in 1954 to 0.9 mg in 2001. Most of the reduction in machine-measured tar and nicotine yield that occurred before 1970 resulted from the introduction of cellulose acetate filters. Further reductions in the FTC-rated yield, after 1970, were achieved by adding ventilation holes and porous paper to dilute the mainstream smoke, technology to puff the tobacco that decreased the amount of tobacco per cigarette, and modifications that caused cigarettes to burn faster so the testing machine had fewer puffs (Kozlowski et al., 2001). A major limitation of the FTC protocol is that the cigarette yield ratings, as measured by machine smoking, do not reliably predict the tar and nicotine exposure of individual smokers (Djordjevic et al., 2000; Burns and Benowitz, 2001). Smokers can compensate for design changes in cigarettes to extract greater amounts of nicotine and tar than the FTC rating would indicate. Studies that have measured salivary cotinine as an indicator of nicotine absorption demonstrate a wide

Figure 13–6. Sales-weighted tar and nicotine values for U.S. cigarettes as measured by a smoking machine using the Federal Trade Commission (FTC) method 1954–1998. (Values before 1968 are estimated from available data.) (Courtesy of D. Hoffmann, personal communication. Source: National Cancer Institute Smoking and Tobacco Control Monograph 13, 2002, p. 2.)

224

PART III: THE CAUSES OF CANCER

Figure 13–7. Wide range of saliva cotinine among smokers smoking cigarettes with a single FTC rating for nicotine. (Source: Jarvis et al., 2001.)

OTHER ASPECTS OF CIGARETTE DESIGN AND COMPOSITION THOUGHT TO INFLUENCE EXPOSURE Other aspects of cigarette design have also changed substantially over time and vary among countries, although few of these changes have been considered in epidemiologic studies. As mentioned above, the selection of special tobacco strains during the early development of manufactured cigarettes is thought to have altered the pH of smoke and the inhalation patterns of smokers (Doll, 1998a). However, the impact of these changes was either not measured systematically or not reported in the scientific literature. Two aspects of tobacco processing that increase the concentrations of TSNAs were the introduction of reconstituted tobacco during the 1950s (which includes tobacco ribs, stems, and leaves and releases higher concentrations of TSNAs) (IARC, 1986; Hoffmann and Hoffmann, 1997), and the fermentation of tobacco for use in cigars and moist snuff. The concentration of TSNAs in moist snuff far exceeds the limit allowed in other consumer products (Hoffmann et al., 1995). Of further concern is that hemoglobin adducts containing TSNAs have been observed in people who “dip” snuff (Hecht et al., 1994). Differences in the methods of curing tobacco underlie the distinction between blond tobacco, used by leading American and transnational brands, and black tobacco, which predominated in France, Spain, and several Latin American countries until the 1980s (Maxwell, 2002). Blond tobacco is produced by flue curing during which the tobacco is heated, whereas black tobacco is produced by air-curing with little or no use of artificial heat. Cigarettes made from black tobacco are reportedly more strongly associated with cancer of the bladder (De Stefani et al., 1991; Vineis, 1991; Bartsch et al., 1993) oropharynx (Boffetta, 1993; De Stefani et al., 1998), larynx (SanchoGarnier and Theobald, 1993), esophagus (De Stefani et al., 1993; Castellsague et al., 1999), and lung (Benhamou and Benhamou, 1993; Armadans-Gil et al., 1999) than are cigarettes made from blond tobacco. Substantial differences have also been observed in the concentration of nitrosamines, nitrates, and nicotine in cigarettes from various countries (Fischer et al., 1991; Gray et al., 1998). Relatively few studies have attempted to integrate information on the differences in chemical composition of cigarettes in various countries with detailed longitudinal information on smoking practices when examining international variations in smoking-attributable risk.

Various novel tobacco products have been developed that reportedly deliver less exposure to certain carcinogens or nicotine than conventional cigarettes. They have been designated “potential reduced-exposure products” (PREPs) by the Institute of Medicine (2001). PREPs include cigarettes and spit tobacco products made from modified tobacco with reduced nitrosamine content, cigarette-like products that deliver nicotine with less combustion of the tobacco, and pharmaceutical products that deliver nicotine, antidepressants, or other medications. Unlike drugs that have been evaluated by the U.S. Food and Drug Administration (FDA) for the treatment of tobacco dependence, most PREPs have not been assessed comprehensively for a sufficient time to determine their hazard compared to conventional tobacco use or their impact on the initiation or cessation of tobacco use.

EXPOSURE MEASUREMENT IN EPIDEMIOLOGIC STUDIES OF TOBACCO The best and most thoroughly validated external measures of tobacco exposure derive from self-reports (Shields, 2002). Most adults can report whether they have smoked 100 or more cigarettes in their lifetime and whether they now smoke every day or on some days, the definition of current smoking (Kovar and Poe, 1985). Parameters that can be determined from self-reports include smoking status (never, current, former), the number of cigarettes, cigars or pipes smoked daily, the use of spit tobacco, age of initiating tobacco use, and age at cessation. A meta-analysis of 26 studies that evaluated the validity of selfreported data on tobacco use found that self-reported smoking status predicted biochemical evidence of active smoking with 87% sensitivity and 89% specificity (Patrick et al., 1994; U.S. Department of Health and Human Services, 2001). The sensitivity and specificity of self-reported smoking were higher in studies of adults than in those of children. Smoking history is considered a more sensitive measure of intermittent smoking than are biochemical indices, as the half-life of cotinine is only about 17 hours (Benowitz, 1996b). However, selfreported data on the number of cigarettes consumed per day are thought to underestimate actual consumption by at least 20%. Estimates of per capita consumption based on questionnaire surveys consistently underestimate consumption based on cigarette sales data by 20%–30% (Todd, 1978).

Tobacco Despite their quantitative limitations, self-reported data on number of cigarettes smoked per day and/or years of smoking are consistently associated with a gradient of risk of developing many cancers. Strong evidence of a dose-response relation exists for all cancers designated causally related to smoking (IARC, 2004). Self-reported information may be less reliable, however, for reflecting more subtle differences or fluctuations in the intensity of smoking at various ages. Few studies have measured the intensity of adolescent smoking, fluctuations in the number of cigarettes smoked per day at different points in life, the number and duration of unsuccessful cessation attempts, the number of puffs taken per cigarette, average puff volume, depth of inhalation, or retention time in the lung. Most of these factors may not be measurable by questionnaire. However, they may introduce sufficient misclassification of lifetime tobacco exposure to make it difficult to document relatively small differences in the pathogenicity of cigarettes, given the potentially larger variations in smoking behavior. Another challenge concerns the difficulty of summarizing cumulative exposure to tobacco over a lifetime. The common practice of combining information on intensity and duration of smoking into a single variable of cumulative exposure (pack-years or cigarette-years) is contraindicated. In the British Doctors’ Study, Doll and Peto showed that the duration of smoking is a much stronger predictor of lung cancer risk than is the number of cigarettes smoked per day (Doll and Peto, 1978). Lung cancer risk increases with the fourth or fifth power of years of smoking but only the second power of cigarettes per day. Researchers increasingly recommend that pack-years no longer be used as an exposure variable (Leffondre et al., 2002), just as “ever” smoking is no longer considered an informative summary of the experience of current and former smokers.

Biomarkers Various biomarkers have been used in epidemiologic studies of tobacco and cancers to assess aspects of absorption, metabolism, excretion, and biologic activity of tobacco smoke (Institute of Medicine, 2001; Shields, 2002; IARC, 2004). The most thoroughly studied are measures that reflect internal exposure to nicotine and other chemicals in the smoke. Cotinine is the main proximate metabolite of nicotine and is considered the biomarker of choice for indicating exposure to tobacco during the last 2–3 days (Benowitz, 1996a). The concentration of cotinine in plasma, saliva, or urine can reliably differentiate active smoking from ETS exposure (IARC, 2004). Other biomarkers, such as thiocyanate in plasma or saliva, carbon monoxide in exhaled alveolar air, and blood carboxyhemoglobin concentrations, are less sensitive and/or less specific as markers of tobacco exposure than cotinine (Institute of Medicine, 2001). Studies of cotinine have been informative in demonstrating that persons exposed to ETS incur exposures that are less than in active smokers but greater than in nonsmokers. Other biomarkers reflect the systemic distribution or biologically effective dose of various components of tobacco smoke (Institute of Medicine, 2001). For example, metabolites of tobacco-specific nitrosamines can be measured in urine (Carmella et al., 1995; Atawodi et al., 1998; Shields, 2002) and other bodily fluids (Hecht, 2002b). Cigarette smoking increases the mutagenicity of urine (Atawodi et al., 1998; Vermeulen et al., 2000), the activation of certain enzymes in body tissue (Whyatt et al., 1995), and the presence of adducts from tobacco-specific nitrosamines or 4-aminobiphenyl attached to DNA, hemoglobin, or lymphocytes (Hecht et al., 1994; Dallinga et al., 1998; Hecht, 1999a). Adducts bound to cellular macromolecules persist longer than nicotine metabolites after abstinence from tobacco use. The number of chromosomal aberrations in cultured lymphocytes and extent of lipid peroxidation have been shown to correlate with the number of cigarettes smoked per day (Shields, 2002). A limitation of many of these biomarkers is that they reflect recent exposure to tobacco smoke rather than exposures in the remote past. Cotinine and carbon monoxide are affected by smoke exposure within the last few days, thiocyanate (from hydrogen cyanide) within the past few weeks (Jarvis, 1987). Although biomarkers are sensitive measures of current exposure to tobacco smoke, most do not reflect long-term usage.

225

Measures of Addiction One parameter that has not generally been measured in epidemiologic studies but that may prove informative in future studies is the level of nicotine addiction of individual smokers. The Fagerstrom index is a widely used index of addiction in behavioral studies (Fagerstrom, 1978; Fagerstrom et al., 1996) that measures six correlates of physical dependence. It includes questions such as: How soon after you wake up do you smoke your first cigarette? Do you find it difficult to refrain from smoking in places where it is forbidden? Do you smoke if you are so ill that you are in bed most of the day? Questions in the Fagerstrom index may correlate with parameters of tobacco exposure that are difficult to measure by questionnaire, such as greater puff volume, depth of inhalation, and retention time in the lung. If this were validated, the Fagerstrom index might provide a useful adjunct to the questions currently used to assess tobacco exposure.

EPIDEMIOLOGIC FINDINGS REGARDING TOBACCO AND CANCER Based predominantly on epidemiologic evidence, active cigarette smoking use is considered causally related to approximately 15 cancer sites. Twelve of these sites are included in the U.S. Surgeon General’s calculation of deaths attributable to cigarette smoking in the United States (see Table 13–1) (U.S. Department of Health and Human Services, 2004). For three other cancer sites (liver, nasal cavity/paranasal sinuses, nasopharynx), the IARC has designated the evidence for a causal relation with smoking as sufficient (IARC, 2004), but the Surgeon General does not currently include them when estimating deaths from smoking (U.S. Department of Health and Human Services, 2004). Table 13–1 indicates the year in which each cancer site or other condition was formally designated as being smoking-related by the Surgeon General. Only four associations were judged to be causal at the time of the first report on smoking and health in 1964: cancers of the lung and larynx and chronic bronchitis in men who smoked cigarettes and lip cancer in men who smoked pipes (U.S. Public Health Service, 1964). The relative risk estimates associated with each of these conditions exceeded 5.0 among current smokers compared to lifelong nonsmokers in early cohort studies in the United Kingdom (Doll and Hill, 1956, 1964, 1966), the United States (Hammond and Horn, 1958; Dorn, 1959; U.S. Public Health Service, 1964), and Canada (Best et al., 1961). Since then, the associations between smoking and many conditions have become considerably stronger as a consequence of earlier age of initiation among smokers. The relations are seen in women as well as men and have been observed in many studies of varying design in different populations. Many of the cancers associated with smoking are located in the respiratory, gastrointestinal, or genitourinary tracts. We discuss respiratory tract cancers first because the associations between smoking and these sites (especially cancers of the lung and larynx) are stronger than the associations with other cancers. The discussion of specific cancer sites is ordered by the International Classification of Diseases (ICD) code.

RESPIRATORY TRACT CANCERS Nasal Cavity and Paranasal Sinuses Cancers of the nasal cavity and paranasal sinuses are rare and were not separately designated as causally related to smoking until the IARC review in 2002 (IARC, 2004). Case-control studies in the United States (Brinton et al., 1984; Zheng et al., 1993; Caplan et al., 2000), the Netherlands (Hayes et al., 1999), other European countries (Mannetje et al., 1999), Hong Kong (Ng, 1986), and Japan (Fukuda and Shibata, 1990) reported relative risk estimates of about 2.0 in current smokers compared to lifelong nonsmokers for squamous cell carcinoma (IARC, 2004). This association is much weaker than the relation of cigarette smoking with cancers of the lung and larynx, but

226

PART III: THE CAUSES OF CANCER

the nasal cavity and paranasal sinuses are exposed to mainstream tobacco smoke only during exhalation.

Larynx Cancer of the larynx is second only to lung cancer in the strength of its association with cigarette smoking (see Table 13–1). The risk of death from laryngeal cancer among current smokers compared to lifelong nonsmokers in Cancer Prevention Study II (CPS-II) was similar for men [relative risk (RR) 14.6] and women (RR 13.0) (see Table 13–1). Several population-based case control studies report relative risks of ≥15 in men who smoke more than one pack of cigarettes per day (Tuyns et al., 1988; Falk et al., 1989; Zatonski et al., 1991; Zheng et al., 1992a; Hedberg et al., 1994). Because cancer of the larynx is rare among lifelong nonsmokers, some case-control studies combine light or former smokers with never-smokers in the reference category (Choi and Kahyo, 1991; Zatonski et al., 1991; Lopez-Abente et al., 1992; Hedberg et al., 1994), thereby attenuating the association with smoking. Many hospital-based case-control studies (Burch et al., 1981; Graham et al., 1981; Herity et al., 1982; Brownson and Chang, 1987; De Stefani et al., 1987; Franceschi et al., 1989; Sankaranarayanan et al., 1990; Ahrens et al., 1991; Choi and Kahyo, 1991; Freudenheim et al., 1992; Lopez-Abente et al., 1992) also weaken the association with smoking by including persons with other smokingrelated diseases in the control group. Risk increases with the duration and intensity of smoking and decreases rapidly after the cessation of smoking (Franceschi et al., 1989). The combination of tobacco smoking with heavy alcohol consumption greatly increases the relative risk for laryngeal cancer (Tuyns et al., 1988; Falk et al., 1989; Franceschi et al., 1990; Choi and Kahyo, 1991; Freudenheim et al., 1992; Zheng et al., 1992b; Baron et al., 1993; Dosemeci et al., 1997; Schlecht et al., 1999), although most studies have not been formally evaluated for statistical interaction.

Trachea, Bronchus, Lung Cigarette smoking is more strongly associated with lung cancer than with any other cancer site (see Table 13–1) (U.S. Department of Health and Human Services, 2004). The relative risk of death from lung cancer among current smokers compared to lifelong nonsmokers was approximately 23 in men and 13 in women in the American Cancer Society (ACS) CPS-II cohort; it increased to about 50 among male long-term (≥40 years) smokers of 40 cigarettes per day (Thun et al., 1997b). Cigarette smoking is strongly associated with all histologic types of lung cancer. In studies conducted during the 1950s, the association between smoking and lung cancer was largely with Kreyberg type I lung cancers (squamous and small-cell carcinomas) rather than with adenocarcinomas or large-cell carcinomas (Doll et al., 1957; Kreyberg, 1962; Wynder and Hoffmann, 1994). Smoking is still strongly associated with squamous and small-cell carcinomas but has become increasingly associated with adenocarcinoma and large-cell carcinomas located in the periphery of the lung (Thun et al., 1997b). The incidence of adenocarcinoma has also increased in many industrialized countries since the 1970s. This increase more closely follows birth cohort patterns relating to the introduction of filter-tip cigarettes and reconstituted tobacco beginning in the 1950s (Thun et al., 1997b) than period changes that would be expected from improvements in the technology of diagnosing peripheral lung cancer (Thun et al., 1997b). Two changes in cigarettes that may have contributed to the increase in adenocarcinomas were the introduction of filter-tip cigarettes and reconstituted tobacco beginning in the 1950s. In the United States, cigarette smoking causes an estimated 88% of lung cancer deaths in men and 72% in women (see Table 13–1) (U.S. Department of Health and Human Services, 2004). Lung cancer deaths from smoking comprise about 80% of all cancer deaths attributable to smoking but only 31% of all deaths from smoking (Centers for Disease Control and Prevention, 2002a). Hence, screening efforts to detect and treat lung cancer early cannot effectively prevent the more than twothirds of smoking-attributable deaths that involve diseases other than lung cancer.

The median delay between the initiation of smoking and death from lung cancer among smokers is approximately 50 years (Thun et al., 2002). The duration of regular smoking has been shown to be a substantially stronger determinant of lung cancer risk than is the number of cigarettes smoked per day (Doll and Peto, 1978; Flanders et al., 2003). Because of the protracted multistage development of solid tumors, the full impact of smoking on national lung cancer rates and on epidemiologic studies manifests only when regular smoking has been entrenched for many decades (IARC, 2004). This phenomenon is frequently misinterpreted, however. During the 1950s and 1960s, prominent scientists (Fisher, 1957, 1958a, 1958b, 1959) and politicians (Macdonald, 1957) interpreted the absence of a large increase in lung cancer mortality among women as evidence that smoking either did not cause lung cancer or that women were resistant to lung cancer, as many women started smoking after World War II (Doll, 1998a). Over time, the relative risk estimates associated with smoking continue to increase in both women and men with the aging of smokers who began at a young age and smoked intensively for many years. Other issues and current controversies regarding lung cancer from cigarette smoking are reviewed elsewhere. They involve the evidence that women are not more susceptible than men to developing lung cancer from an equivalent amount of smoking (Thun et al., 2002; Bain et al., 2004); IARC, 2004), the complex differences in smoking behaviors and lung cancer risk between African Americans and Caucasians, factors that affect the probability that a smoker will develop lung cancer, and the difficulty of distinguishing historical differences in smoking behavior from other factors influencing lung cancer rates in countries such as Japan.

GASTROINTESTINAL CANCERS Tobacco smoking is associated with cancer at all sites in the upper aerodigestive tract except the salivary glands (IARC, 2004). The association generally becomes weaker with progression from mouth to rectum. Early cohort analyses grouped many cancer sites into the single category of upper aerodigestive tract cancers. With longer followup of large cohort studies and the addition of case-control studies, many individual cancer sites and subsites have been examined separately.

Lip, Oral Cavity, Pharynx All forms of tobacco use (cigarettes, pipes, cigars, snuff, chewing tobacco, betel, other smoked and smokeless products) cause dysplasia and cancer (predominantly squamous cell carcinoma) of the oral cavity and pharynx. The magnitude of risk from cigar and pipe smoking is similar to the risk from cigarettes. On average, the RR of oropharyngeal cancer among persons who currently and exclusively smoke cigarettes is about 10.0 in men and 5.0 in women compared to that of lifelong nonsmokers (see Table 13–1) (Rogot and Murray, 1980; Franceschi et al., 1992; Muscat et al., 1996; U.S. Department of Health and Human Services, 2004). Tobacco use combined with heavy alcohol consumption magnifies the risk of either exposure alone. In one large population-based case-control study, the RR associated with smoking ≥40 cigarettes per day for ≥20 years was 7.4 in men and undefined among women who drank less than one alcoholic drink per week, but it was 37.7 in men and 107.9 in women who drank more than 30 alcoholic drinks per week (Blot et al., 1988). After cessation of smoking, the relative risk of oropharyngeal cancer decreases substantially within the first 10 years after quitting. Premalignant oral lesions such as leukoplakia and erythroplasia have been shown to regress after cessation of tobacco use (van der Waal et al., 1997; Martin et al., 1999).

Nasopharynx Only recently have large cohort studies accrued enough cases over prolonged follow-up to examine cancers of the nasopharynx separately. A 26-year follow-up of men in the U.S. Veterans Study documented an association with current cigarette smoking [odds ratio (OR)

Tobacco 3.9, 95% confidence interval (CI) 1.5–10.3] (Chow et al., 1993) that was stronger in men smoking more than two packs daily (OR 6.4, 95% CI 1.2–33.2). Cigarette smoking is also associated with cancer of the nasopharynx in six population-based case-control studies (Nam et al., 1992; Zhu et al., 1995; Vaughan et al., 1996; Armstrong et al., 2000; Cao et al., 2000; Yuan et al., 2000). None of these studies controlled for Epstein-Barr virus, the principal cause of nasopharyngeal carcinoma worldwide.

Esophagus Tobacco smoking was designated a cause of esophageal cancer by the U.S. Surgeon General in 1982 (U.S. Department of Health and Human Services, 1982) and by the IARC in 1986 (IARC, 1986). The risk of developing any esophageal cancer, irrespective of histologic subtype, is four to seven times higher among current smokers than lifelong nonsmokers in most studies (IARC, 2004). The risk increases with the amount and duration of smoking and decreases with earlier age at cessation (U.S. Department of Health and Human Services, 2004). Smoking combined with heavy alcohol consumption greatly increases risk (Inoue et al., 1994; Castellsague et al., 1999), with RR estimates sometimes exceeding 100 (Talamini et al., 2000). Even 1–24 ml of ethanol consumed per day increases the risk associated with each level of smoking (Castellsague et al., 1999). Cigar and/or pipe smokers incur risks of esophageal cancer similar to those of cigarette smokers (Kahn, 1966; Carstensen et al., 1987; Shanks and Burns, 1998). Smoking was associated with both squamous cell carcinoma and adenocarcinoma of the esophagus in case-control analyses that stratified by histologic subtype (Kabat et al., 1993; Inoue et al., 1994; Vaughan et al., 1995; Ahsan et al., 1997; Castellsague et al., 1999; Lagergren et al., 2000; Talamini et al., 2000). Several studies have suggested that black tobacco may confer a greater risk of esophageal cancer than blond tobacco (De Stefani et al., 1993; Castellsague et al., 1999).

Stomach More than 20 cohort studies and nearly 40 case-control studies have reported an association between tobacco smoking and stomach cancer, with the RR averaging approximately 1.6 in current cigarette smokers compared to that for never-smokers (IARC, 2004). Until recently, stomach cancer was not classified as smoking-related because of uncertainty about potential confounding by Helicobacter pylori infection and diet (IARC, 1986; U.S. Department of Health and Human Services, 1989). However, several case-control studies have stratified the analysis based on H. pylori seropositivity and reported substantially stronger associations between smoking and stomach cancer in persons who are seropositive for H. pylori than in uninfected individuals (Jedrychowski et al., 1993, 1999; Zaridze et al., 2000; Siman et al., 2001; Brenner et al., 2002; Wu et al., 2003). There is some evidence that tobacco smoking adversely affects the progression of intestinal metaplasia to dysplasia in H. pylori-affected people (Kneller et al., 1992; You et al., 1999). The most recent IARC review includes stomach cancer among the sites for which the evidence is considered sufficient in humans (IARC, 2004). Smoking appears to be associated with cancers of both the gastric cardia and noncardia; some casecontrol studies (Palli et al., 1992; Gammon et al., 1997) but not others (Ye et al., 1999) have reported a stronger association with cancer of the gastric cardia than with other subsites. It has been estimated (Tredaniel et al., 1997) that worldwide the proportion of stomach cancer attributable to smoking is 11% in men and 4% in women in economically developing countries and 17% among men and 11% among women in developed countries.

Colorectum The Surgeon General’s series of reports on the health consequences of smoking first considered smoking in relation to colorectal cancer in 2001 (U.S. Department of Health and Human Services, 2001) and more recently concluded that the evidence is suggestive but not suffi-

227

cient to infer a causal relation (U.S. Department of Health and Human Services, 2004). Increased risk of colorectal adenomatous polyps is associated with current cigarette smoking in 3 prospective studies (Giovannucci et al., 1994a, 1994b; Nagata et al., 1999) and 13 casecontrol studies (Kikendall et al., 1989; Cope et al., 1991; Monnet et al., 1991; Zahm et al., 1991; Olsen and Kronborg, 1993; Boutron et al., 1995; Martinez et al., 1995; Longnecker et al., 1996; Nagata et al., 1999; Potter et al., 1999; Almendingen et al., 2000; BreuerKatschinski et al., 2000; Inoue et al., 2000). With one exception (Kato et al., 1990), the relative risk estimates associated with current smoking range between 1.5 and 3.8 after adjusting for age and other covariates. Prospective cohort studies of colon cancer (Chute et al., 1991; Tverdal et al., 1993; Doll et al., 1994; Heineman et al., 1994; Chyou et al., 1996; Engeland et al., 1996; Nyren et al., 1996; Hsing et al., 1998; Sturmer et al., 2000), and rectal cancer (Chute et al., 1991; Tverdal et al., 1993; Doll et al., 1994; Heineman et al., 1994; Chyou et al., 1996; Engeland et al., 1996; Nyren et al., 1996) have generally reported RR estimates associated with current cigarette smoking of 1.2–1.4 for colon cancer and 1.4–2.0 for rectal cancer. However, there has not yet been a systematic meta-analysis that has evaluated this association across all studies, controlling for other factors known to increase or decrease the risk of colorectal cancer. Cancer of the anus, a malignancy with squamous or transitional cell histology, has repeatedly been found to be positively associated with cigarette smoking (Daling et al., 1992), although confounding by human papillomavirus has not been excluded.

Liver At least 22 cohort and 27 case-control studies have examined the relation between tobacco smoking and hepatocellular carcinoma. Most of the studies reported an association between smoking and liver cancer, with RR estimates of about 1.5–2.5 (Gonzalez et al., 2003). Liver cancer was not classified as smoking-related by the IARC in 1986, however, because of uncertainty about potential confounding by hepatitis virus infection and heavy consumption of alcohol (IARC, 1986; U.S. Department of Health and Human Services, 1989). Hepatitis B virus (HBV) infection causes most liver cancer worldwide, and hepatitis C virus (HCV) infection accounts for a large fraction of the disease in Japan, North Africa, and southern Europe (IARC, 1994). Heavy, but not moderate, consumption of alcohol also contributes to the risk (IARC, 1988). Recent studies have resolved these concerns about confounding. Several studies have reported a higher risk of liver cancer among nondrinking smokers compared to nondrinking nonsmokers (Chen et al., 1991; Goodman et al., 1995; Kuper et al., 2000). Smoking is also associated with liver cancer among Chinese (Liu et al., 1998) and Japanese (Tanaka et al., 1995) women in whom heavy alcohol consumption is rare (Gonzalez et al., 2003). Several studies have stratified on or adjusted for hepatitis B surface antigen (HbsAg) and antiHCV and found little attenuation of the association between smoking and liver cancer (Yu et al., 1991; Liaw and Chen, 1998; Kuper et al., 2000). The risk of chronic infection with hepatitis viruses was not higher in smokers than nonsmokers in one study (Evans et al., 2002); however, compared to never-smokers, smokers did experience greater risk of progression from chronic HBV and HCV infection to liver cirrhosis (Yu et al., 1991) and/or liver cancer (Tsukuma et al., 1993).

Gallbladder, Biliary Tract Cigarette smoking was associated with increased mortality from gallbladder and biliary tract cancer among current (RR 1.5, 95% CI 1.1–2.0) smokers in a 26-year follow-up of U.S. veterans (Chow et al., 1995). The association was stronger in persons who began smoking at a younger age or smoked more cigarettes per day. Findings have been inconsistent in case-control studies. Some studies have observed increased risk of biliary tract cancer in smokers (Chow et al., 1994; Moerman et al., 1994; Scott et al., 1999), whereas others have not (Yen et al., 1987; Zatonski et al., 1992; Moerman et al., 1994; Chalasani et al., 2000).

228

PART III: THE CAUSES OF CANCER

Pancreas Numerous cohort (n = 17) and case-control (n = 30) studies have reported an association between cigarette smoking and cancer of the exocrine pancreas (IARC, 2004). The risk of pancreatic cancer is also increased among persons who smoked exclusively cigars and/or pipes in most large studies (Shanks and Burns, 1998; Shapiro et al., 2000). The evidence that tobacco smoke is carcinogenic to the human pancreas was classified as sufficient by the IARC in 1986 (IARC, 1986) and by the U.S. Surgeon General in 1982 (U.S. Department of Health and Human Services, 1982). The RR of death from pancreatic cancer among male and female current cigarette smokers compared to neversmokers, is 2.3 for both men and women in the ACS CPS-II (U.S. Department of Health and Human Services, 2004). Risk decreases among persons who stop smoking compared to those who continue in both cohort and case-control studies (IARC, 2004; U.S. Department of Health and Human Services, 2004).

URINARY TRACT CANCERS Tobacco smoking is an established risk factor for cancer throughout the urinary tract (IARC, 1986, 2004; U.S. Department of Health and Human Services, 2004). Smoking was identified as an important cause of transitional cell carcinomas of the lower urinary tract (renal pelvis, ureter, urinary bladder, ureter) by the IARC in 1986 (IARC, 1986). More recently, the IARC designated the association between smoking and adenocarcinoma of the renal parenchyma as causal, although it was not as strong as that with transitional cell carcinoma (IARC, 2004). Renal adenocarcinoma accounts for most kidney cancers (Doll, 1996). The U.S. Surgeon General did not distinguish between adenocarcinoma and transitional cell carcinoma when estimating that cigarette smoking accounts for 26% of deaths from cancers of the kidney, ureter, and urethra in the United States (40% in men and 5% in women) (U.S. Department of Health and Human Services, 2004).

Renal Pelvis, Ureter Relatively few studies have specifically examined the association between smoking and cancers of the renal pelvis and ureter (McCredie et al., 1983; Jensen et al., 1988; Ross et al., 1989; McLaughlin et al., 1992b). These studies suggest that smoking is the strongest known risk factor for cancers of the renal pelvis and ureter and is the primary cause of these cancers worldwide (McLaughlin et al., 1992b). The relative risk estimates are at least as high and dose-response relations are steeper for cancers of the renal pelvis and ureter than for bladder cancer in studies in which both types of cancer have been investigated in the same geographic area (Jensen et al., 1987, 1988; McCredie and Stewart, 1992).

Renal Adenocarcinoma The 1986 IARC monograph on tobacco did not consider the evidence linking smoking to adenocarcinoma of the renal parenchyma with tobacco smoking to be conclusive because most of the available studies were either small or based on death certificates, which the committee did not believe could differentiate reliably between cancers of the renal parenchyma and those arising in the renal pelvis (IARC, 1986). A reevaluation by Doll et al. in 1996 (Doll, 1996) noted that all of the additional studies in Australia (McCredie et al., 1983; McCredie and Stewart, 1992), Canada (Kreiger et al., 1993), China (McLaughlin et al., 1992a), Denmark (Mellemgaard et al., 1994), Italy (Talamini et al., 1990), and the United States (Asal et al., 1988; Maclure and Willett, 1990; Hiatt et al., 1994) with more than 100 affected patients have found an increased risk in cigarette smokers compared to nonsmokers. Furthermore, the RR estimates in these studies were compatible with the estimates from a large collaborative study by McLaughlin et al. (Chow et al., 1995). The latter estimated the RR among current cigarette smokers compared to lifelong nonsmokers to be 1.1 in people smoking 1–10 cigarettes per day, 1.3 in

those smoking 11–20 cigarettes per day, and 2.1 in those smoking ≥21 daily (Doll, 1996). The average RR among former smokers was estimated to be 1.2. Studies that adjusted the RR estimates for other potential risk factors such as hypertension, use of diuretics and other medications for hypertension, and obesity did not find a substantial attenuation in the association with smoking (IARC, 2004).

Urinary Bladder Smoking is consistently associated with increased risk of transitional cell carcinoma of the urinary bladder in many epidemiologic studies (Brennan et al., 2000, 2001; IARC, 2004). The association between smoking and bladder cancer was classified as causal by the IARC in 1986 (IARC, 1986) and by the U.S Surgeon General in 1979 (U.S. Department of Health, Education, and Welfare, 1979). An estimated 41% of deaths from cancer of the urinary bladder in the United States are attributable to cigarette smoking (47% in men and 29% in women) (U.S. Department of Health and Human Services, 2004). The risk of bladder cancer among male current cigarettes smokers is approximately two to three times higher than that of never-smokers in prospective studies from the United States, Japan, and Europe (Brennan et al., 2000, 2001). Metabolites of heterocyclic aromatic amines, polycyclic aromatic hydrocarbons, and other carcinogens in tobacco can be detected in urine (IARC, 2004). Smokers have a higher prevalence of preneoplastic changes in the bladder (Auerbach and Garfinkel, 1986). DNA adducts have been detected in exfoliated urothelial cells from cigarette smokers (Talaska et al., 1991). There is limited evidence that the association between smoking and bladder cancer is stronger for smoking black (air-cured) than blond (fluecured) tobacco (Clavel et al., 1989; D’Avanzo et al., 1990; De Stefani et al., 1991; Lopez-Abente et al., 1991; Momas et al., 1994; Vineis et al., 1984). Meta-analyses of case-control (Marcus et al., 2000a) and case series (Marcus et al., 2000c) studies of smoking in relation to bladder cancer have reported a stronger association among persons with the slow acetylator N-acetyltransferase (NAT2) phenotype.

CANCERS OF THE REPRODUCTIVE TRACT AND BREAST Uterine Cervix Cancer of the uterine cervix, predominantly involving squamous cell carcinoma, is consistently associated with cigarette smoking in many studies. However, the association was not designated as causal by the IARC until 2002 (IARC, 2004) because of uncertainties about confounding by sexually transmitted diseases (Gonzalez et al., 2003). Since the identification of human papillomavirus (HPV) infection as the main cause of cervical cancer (IARC, 1995) studies have examined whether tobacco smoking acts as a cofactor with HPV infection in causing progression from preneoplastic lesions to cancer. Analyses either have been restricted to study participants who are positive for HPV DNA or have tried to adjust for HPV infection in the analysis. In the IARC multicenter, pooled analysis of invasive cervical cancer, restriction did not materially alter the association between smoking and risk (Plummer et al., 2001). Furthermore, in one cross-sectional study, smoking was more strongly associated with high-grade cervical intraepithelial neoplasia than with HPV infection (Deacon et al., 2000). Smoking is associated with a spectrum of cervical abnormalities, from dysplasia to cervical intraepithelial neoplasia, cervical cancer in situ, and invasive squamous cell carcinoma. Risk increases with increasing intensity and duration of smoking. Whereas smoking is associated with an increased risk of squamous cell carcinoma of the uterine cervix, it was associated with lower risk of adenocarcinoma of the cervix in one multicenter case-control study (Lacey et al., 2001).

Endometrium Endometrial cancer is the only human cancer reported to be inversely associated with cigarette smoking. More than 25 case-control studies

Tobacco and 5 cohort studies have examined the association between smoking and endometrial cancer (U.S. Department of Health and Human Services, 2001; IARC, 2004). Most report lower risk among current smokers than nonsmokers, with RR estimates of 0.2–0.9. In only a few studies has the association been statistically significant (Levi et al., 1987; Stockwell and Lyman, 1987; Elliott et al., 1990; Brinton et al., 1993; Weiderpass and Baron, 2001). Adjustment for other factors associated with decreased risk (oral contraceptives) or increased risk (obesity, late onset of menopause, menstrual disorders, infertility, hormone replacement therapy) does not materially alter the inverse association with smoking in these studies. The lower risk of endometrial cancer associated with current smoking was hypothesized by MacMahon et al. (1982) to be caused by a reduction in estrogen production (see discussion of smoking and estrogen, under Breast Cancer).

Vulva and Vagina Three case-control studies have examined the risk of cancer of the vulva in relation to cigarette smoking (Newcomb et al., 1984; Mabuchi et al., 1985; Brinton et al., 1990). All have found a positive association. None of the studies evaluated the possibility of confounding or biologic interaction with HPV infection (IARC, 1995). A report concluded that, “smoking may be associated with an increased risk for vulvar cancer, but the extent to which this association is independent of human papillomavirus infection is uncertain” (U.S. Department of Health and Human Services, 2001).

Ovary Eight cohort studies and nine case-control studies have assessed the relation between ovarian cancer and smoking (U.S. Department of Health and Human Services, 2001; IARC, 2004). Most studies found no relation, although few have considered the histologic subtypes of ovarian cancer or examined risk in relation to duration and intensity.

Breast Numerous case-control and cohort studies have examined the relation between cigarette smoking and breast cancer incidence or death rates (Palmer and Rosenberg, 1993; U.S. Department of Health and Human Services, 2001; IARC, 2004). No consistent association with either increased or decreased risk is seen with overall breast cancer incidence or with subgroups defined by menopausal status, estrogen receptor positivity (Cooper et al., 1989), or initiation of smoking during puberty or before first full-term pregnancy (Terry and Rohan, 2002). Cigarette smoking is also either unrelated or inversely related to benign breast disease. Thus, an extensive literature indicates no substantial overall association between cigarette smoking and breast cancer incidence in women or men (IARC, 2004). However, the possibility of a relation between smoking and breast cancer continues to be studied, in part because carcinogens in tobacco smoke cause mammary cancer in rodents (Ambrosone and Shields, 2001; Hecht, 2002c) and because DNA adducts containing polycyclic aromatic hydrocarbons have been found in exfoliated ductal epithelial cells in human breast milk (Gorlewska-Roberts et al., 2002; Thompson et al., 2002). One hypothesis that was proposed to explain the absence of an association in epidemiologic studies is that the antiestrogenic effects of smoking may reduce breast cancer risk and thus obscure the deleterious effect of carcinogens from tobacco on breast cancer risk. Several lines of indirect evidence suggest an antiestrogenic effect from active smoking, although the underlying mechanism for this effect is not known (Baron et al., 1990). Current smokers experience menopause 1–2 years earlier than never-smokers (Lesko et al., 1985; Brinton et al., 1986; Chu et al., 1990). Smokers also have a lower risk of endometrial cancer and endometriosis (both estrogen-responsive conditions), lower mammographic density (Sala et al., 2000; Vachon et al., 2000), and higher risk of osteoporotic fractures (Jensen et al.,

229

1985; Paganini-Hill and Hsu, 1994) than nonsmokers. However, the concentrations of estrogen (estrone and estradiol) measured in plasma or urine are not demonstrably lower in smokers (Freidman et al., 1987; Khaw et al., 1988; Longcope and Johnston 1988; Cauley et al., 1989; Schlemmer et al., 1990; Key et al., 1991; Berta et al., 1992; Cassidenti et al., 1992). Some researchers have proposed that the antiestrogenic effect from active smoking may reflect differential metabolism of estrogens through the 2-hydroxylation pathway, producing a metabolite with low estrogenic activity (Michnovicz et al., 1986; Key et al., 1996; Cook et al., 2003). A related issue is whether the inclusion of women exposed to second-hand smoke in the referent group in epidemiologic studies of breast cancer may obscure an increase in risk associated with active or passive smoking. The reasoning here is that second-hand smoke may contain enough carcinogens to increase a woman’s risk of breast cancer but not enough to trigger the antiestrogenic effect of active smoking. According to this theory, the inclusion of women with ETS exposure among the never-smoking comparison group in studies of active smoking may obscure the increase in risk associated with smoking. The practical dilemma is that most women in wealthy countries have been exposed to ETS, given the ubiquity of involuntary exposure over the last half century. Studies that attempt to define a small subset of women who report neither active nor passive smoking and use this as the referent group may introduce bias because such women are likely to be atypical in other respects that relate to breast cancer risk (IARC, 2004).

MALE GENITAL TRACT Prostate Studies have consistently shown no association between cigarette smoking and prostate cancer incidence (Hickey et al., 2001; Levi and La Vecchia, 2001; IARC, 2004). However, several cohort studies that have examined death rates from prostate cancer have reported higher mortality in smokers than in lifelong nonsmokers with RR estimates of 1.2–2.0 (Hsing et al., 1991; Tverdal et al., 1993; Adami et al., 1996; Coughlin et al., 1996; Rodriguez et al., 1997; Eichholzer et al., 1999; Giovannucci et al., 1999). Prostate cancer is a common disease among elderly men, however. The review by the IARC questions whether the association between smoking and increased death rates from prostate cancer represents a true causal effect from smoking, perhaps due to shortened survival, or a bias introduced by accelerated mortality from other smoking-attributable diseases.

Testes Cancer of the testes has not been extensively investigated but seems unrelated to smoking (Henderson et al., 1979; Brown et al., 1987; Gallagher et al., 1995).

Penis Studies of smoking in relation to squamous cell carcinoma of the penis have been reviewed elsewhere (Dillner et al., 2000, Moore et al., 2001). Several case-control studies (Hellberg et al., 1987; Maden et al., 1993; Harish and Ravi, 1995; Tsen et al., 2001) but not all (Brinton et al., 1991) that examined this endpoint reported increased risk among smokers. It is still uncertain whether the association is confounded by HPV exposure, however. The IARC has not classified the association as causally related to smoking.

NONMELANOMA SKIN CANCER Studies of the association of cigarette smoking with nonmelanoma skin cancer have yielded conflicting results. Several studies have reported increased risk of squamous cell but not basal cell carcinoma (Aubry and MacGibbon, 1985; Hunter et al., 1990; Karagas et al., 1992; Grodstein et al., 1995; Sahl et al., 1995; De Hertog et al., 2001).

230

PART III: THE CAUSES OF CANCER

Despite suggestions that cigarette smoking increases the metastatic potential of melanoma, no consistent smoking-related effect on its incidence or mortality has been observed (Osterlind et al., 1988; Westerdahl et al., 1996; Lear et al., 1998).

LEUKEMIA Of the hematopoietic malignancies, only acute myeloid leukemia is consistently associated with smoking (Doll 1996; U.S. Department of Health and Human Services, 2001). This association was designated causal by the IARC in 2002 (IARC, 2004) and by the U.S Surgeon General in 2004 (U.S. Department of Health and Human Services, 2004). The risk of leukemia was first observed to be higher in smokers than nonsmokers by Austin and Cole (1986). The relation was subsequently shown to be stronger for myeloid and monocytic leukemia than for lymphocytic subtypes in the U.S. Veterans cohort (Kinlen and Rogot, 1988). The association has been confirmed in other cohort (Garfinkel and Boffetta, 1990; Doll et al., 1994; Doll, 1996) and casecontrol studies, as reviewed elsewhere (IARC, 2004). Tobacco smoke contains several leukemogens, such as benzene and radioactive isotopes of polonium and lead. Smokers have much higher levels of benzene in the blood than do nonsmokers. Based on linear extrapolation from the known effects at high doses, Korte et al. (2000) estimated that benzene in cigarettes accounts for 12%–58% of smoking-induced myeloid leukemia.

OTHER CANCERS Thyroid Two studies have reported increased risk of thyroid cancer associated with smoking (Sokic et al., 1994; Memon et al., 2002), whereas others have reported lower risk among smokers (Galanti et al., 1996; Kreiger and Parkes, 2000; Rossing et al., 2000); still others reported no association (Williams and Horm, 1977; Rogot and Murray, 1980; McTiernan et al., 1984; Anonymous, 1990; Kolonel et al., 1990; Iribarren et al., 2001).

Other Nonepithelial Cancers Tobacco use is not consistently related to hematologic malignancies, other than leukemia, including lymphoma in aggregate, nonHodgkin’s lymphoma (Williams and Horm, 1977; Franceschi et al., 1989; Brown et al., 1992; Linet et al., 1992; Zahm et al., 1997; U.S. Department of Health and Human Services, 2001), or multiple myeloma (Boffetta et al., 1989; Mills et al., 1990; Heineman et al., 1992; Linet et al., 1992; Friedman, 1993; Adami et al., 1998). The

Figure 13–8. Risk of death from smoking-related cancers: CPS-II men and women, 1984–1991.

association with Hodgkin’s disease has not been adequately evaluated (U.S. Department of Health and Human Services, 2001). Primary neoplasms of the CNS are unrelated to active smoking in most studies (Hochberg et al., 1990; Ryan et al., 1992; Hurley et al., 1996; Zheng et al., 2001) but not all of them (Lee et al., 1997). The data regarding soft-tissue sarcoma are limited and mixed (Serraino et al., 1991; Franceschi and Serraino, 1992; Zahm et al., 1992).

BENEFITS OF SMOKING CESSATION Many of the detrimental effects of smoking can be prevented or reversed by smoking cessation. Strong evidence of the health benefits of smoking cessation first became available for lung cancer but now exists for nearly all the cancers related to smoking and for the major nonneoplastic smoking-attributable diseases (U.S. Department of Health and Human Services, 1990; IARC, 2004). The decrease in relative risk among persons who stop smoking compared to those who continue is an important aspect of the evidence for causation and indicates that continuing exposure influences even the late stages of carcinogenesis, presumably by affecting the transformation of premalignant clones into invasive cancer. The absolute risk of developing cancer or other smoking-related diseases does not decrease after smoking cessation but increases with age at a slower rate in persons who stop smoking than in those who continue. The risk of developing smoking-attributable diseases continues to diverge between the two groups over time. This is reflected by a progressive decrease (below unity) in the RR estimates comparing former to current smokers as time elapses since cessation. The relation between smoking cessation and the cumulative risk of death from any of the smoking-attributable cancers (as listed in Table 13–1 minus stomach cancer and myeloid leukemia) is depicted in Figure 13–8. This analysis is based on a 9-year follow-up of the CPSII cohort from 1984–1991, excluding the first 2 years of follow-up (1982–1984) to avoid bias from persons who have stopped smoking because of diseases caused by tobacco. Men and women who continue to smoke cigarettes have the highest cumulative risk of death from these cancers at all ages shown. Cumulative risk increases more slowly with age in persons who have stopped smoking than in those who continue. The cumulative risk is also lower, in absolute terms, with earlier age of quitting. Persons who have never smoked have the lowest cumulative risk. Figure 13–8 illustrates at least two important points. The first is that smoking cessation at any age avoids much of the future increase in risk seen with continued smoking. The health benefits are greatest when cessation occurs at an early age but are substantial even when cessation occurs by age 50 or 60 (Peto et al., 2000). The absolute risk among persons who quit smoking and their RR compared with those

Tobacco who continue are progressively smaller the earlier the age of cessation and the longer the time that has elapsed since cessation. Only in smokers who quit at younger ages does the relative risk of death from these cancers approach unity when compared with that of persons who have never smoked. However, the relevant comparison for a smoker is between the large risk from continued smoking and the much smaller risk from cessation. Most smokers have the option to quit smoking but not the possibility of returning to the status of a never-smoker. Second, analyses of cessation are more informative if they consider age at cessation rather than time since quitting, as the benefits of cessation are not constant across all ages. Furthermore, measuring the cumulative risk of developing a specified endpoint is more stable than measuring the annual incidence. It is also more meaningful from the perspective of the individual smoker because he or she passes through all of the previous time intervals.

FACTORS THAT MAY MODIFY CANCER RISK Genetic Influences on Smoking Behavior Genetic factors can, in principle, modify susceptibility across the entire spectrum of smoking initiation, addiction, carcinogen metabolism, DNA repair, and tumor suppression (Thun et al., 2002). Inherited genetic traits can influence an individual’s tendency to experiment with tobacco, become dependent on its use, stop smoking successfully, and/or relapse. The genetic contribution to smoking behavior is thought to be as least as great as for alcoholism (Arinami et al., 2000). Interestingly, genetic variation becomes an important determinant of smoking behavior only in cultures where tobacco use is prevalent and environmental factors impose fewer constraints on exposure to tobacco and tobacco marketing. The candidate genes that have received the most attention with respect to smoking behavior include the dopamine D1, D2, and D4 receptors; dopamine transporter and serotonin transporter genes; and the cytochrome P450 subfamily polypeptide 6 (CYP2A6) (Arinami et al., 2000). These genetic and metabolic factors are thought to affect reward pathways of the CNS by influencing the binding and metabolism of nicotine and other neurotransmitters. Research to identify the genetic and biologic determinants of addiction may contribute to developing more effective drugs for the treatment of tobacco dependence and to identifying appropriate pharmacologic and behavioral treatments (personalized genetic counseling) for individual smokers.

Genetic Susceptibility to Tobacco Carcinogenesis Increasingly, epidemiologists study tobacco-exposed populations to identify genetic traits that confer susceptibility or resistance to problems with carcinogen metabolism, DNA repair, and/or tumor suppression. Valid markers of genetic susceptibility might further our understanding of the mechanisms of carcinogenesis and/or help clarify etiologic relations that are presently confusing or inconsistent (Thun et al., 2002). For instance, available data are conflicting as to whether postmenopausal women with the slow phenotype of N-acetyltransferase are genetically more susceptible to developing breast cancer from tobacco use than other women (Ambrosone et al., 1996) or whether nonsmokers who are homozygous for the GTSM1 null allele are at particularly high risk of developing lung cancer from ETS exposure (Bennett et al., 1999; Weinberg and Sandler, 1999). It may be clinically valuable to identify subcategories of addicted smokers and individualize smoking cessation treatment or to identify high risk persons for enrollment in chemoprevention trials or special cancer screening programs that are not appropriate for the general population (Bartsch et al., 2000). So far, many of the published results concerning gene–environment or gene–gene–environment interactions involving smoking have not been reproducible, probably because of the inadequate sample size of these studies, overemphasis on subgroup analyses, and comparatively crude assessment of lifetime smoking behavior. Most published studies have evaluated a small number of single nucleotide polymor-

231

phisms (SNPs) rather than all the genetic variants relevant to specific metabolic pathways. Perhaps the strongest evidence of an interaction between smoking and genetic susceptibility derives from meta-analyses of studies of smoking and a variant of the N-acetyltransferase gene with respect to bladder cancer (Marcus et al., 2000a, 2000c). Persons with the slow acetylator (NAT-2) phenotype are known to be less efficient in detoxifying monoarylamines such as 4-aminobiphenyl. Compared to rapid acetylators, persons with the slow acetylator phenotype have been observed to have higher risk of developing bladder cancer associated with smoking. Although other genetic polymorphisms are likely to have modest but real effects on the risk of common diseases such as the cancers caused by smoking, much larger studies are needed to define these interactions than have been conducted in the past (Lohmueller et al., 2003). Studies of carcinogen metabolism and/or DNA repair are also needed to quantify lifetime tobacco exposure with more precision than have past studies to distinguish gradations in risk due to exposure from those caused by genetic susceptibility. Several reviews have proposed strategies to improve the design and interpretation of future molecular epidemiologic studies of gene–environment interactions, specifically as they concern tobacco exposure (Houlston, 1999; Bartsch et al., 2000; Brockton et al., 2000; Green et al., 2000; Geisler and Olshan, 2001). From a public health perspective, identifying genetic susceptibility factors that modify the risks of developing one or another tobaccorelated disease is less critical than is the application of measures known to reduce tobacco use. Genetic screening and counseling are unlikely to be effective in deterring smoking initiation among adolescents because teenagers who are most likely to smoke may be the least likely to participate in or accept genetic counseling. Genetic screening is also unlikely to reduce the cost of smoking cessation programs, as approximately 50% of long-term smokers die prematurely because of their tobacco use (Peto et al., 1994), and the expense of genetic screening would outweigh any savings from narrowing the population to be treated (Thun et al., 2002). Nevertheless, personalized genetic counseling may help guide the selection of pharmacologic and behavioral treatments for individual smokers and may motivate some individuals to quit.

Age at Initiation Beginning smoking at a young age is associated with higher risk of many smoking-related diseases. It is not clear, however, whether early age of initiation is detrimental simply because it leads to longer duration of smoking or whether there is additional risk because the immature lung is more vulnerable to early-stage carcinogenic events. One study has reported that among former smokers, the concentration of DNA adducts in nontumorous lung tissue increased with earlier age of smoking initiation (Wiencke et al., 1999). It is difficult to evaluate whether early age of initiation has an independent effect on lung cancer risk in Western countries because of the relatively narrow range of age of initiation and the strong inverse correlation between age of initiation and duration of smoking among current smokers. Most smokers in the United States become addicted between the ages of 14 and 21 years. The duration of smoking among current smokers is nearly collinear with attained age and moderately correlated with the age of initiation. It is difficult to distinguish the relation between duration of smoking and cancer risk from any additional contribution from early age at initiation or attained age (Leffondre et al., 2002). Distinguishing these factors is more relevant to mechanistic considerations than to public health, as the initiation of smoking and other forms of tobacco use by children is inherently undesirable.

VARIATIONS IN THE CARCINOGENICITY OF CIGARETTES Approximately 50 epidemiologic studies have examined the relation between cancer risk and changes that have occurred in cigarette design

232

PART III: THE CAUSES OF CANCER

since the 1950s (Burns et al., 2001a; Thun and Burns, 2001; West et al., 2003; IARC, 2004). Most of these studies compared the risks of smoking high-tar (>30 mg tar) unfiltered brands that predominated before the 1950s to those associated with smoking filter-tip cigarettes with an FTC tar rating of approximately 20–22 mg (West et al., 2003). These studies generally reported a higher risk of lung cancer in persons who smoke unfiltered cigarettes than in those who smoke filter tip brands. A similar pattern, with more limited data has been reported for cancers of the oropharynx, larynx, esophagus, and pancreas. In contrast, only five studies have compared the risks of lung cancer or other diseases among smokers across the range of tar yields that prevail today (Vutuc and Kunze, 1982; Vutuc and Kunze, 1983; Wilcox et al., 1988; Woodward et al., 1999; Harris et al., 2004). Presently, the available data do not support the premise that brands classified as low tar (FTC rating 8–14 mg) or very low tar (£7 mg) confer a lower risk of lung cancer than conventional filter-tip cigarettes (15–21 mg) (Vutuc and Kunze, 1982; Vutuc and Kunze, 1983; Wilcox et al., 1988; Woodward et al., 1999; Harris et al., 2004). Limitations of the epidemiologic studies of low tar cigarettes have been reviewed elsewhere (Burns et al., 2001a; Thun and Burns, 2001). These studies have had limited ability to control for factors associated with self-selection of lower-yield brands. Smokers who are less addicted may be more likely to switch from unfiltered to filter-tip products. Furthermore, studies of “reduced yield” cigarettes have not considered the indirect adverse effects that the marketing of these products may have had on the uptake and/or prolongation of smoking. Brands that deliver lower levels of machine-measured tar are promoted to smokers concerned about the adverse health effects of smoking. Some products are marketed explicitly or implicitly as an alternative to cessation. Addicted smokers may delay genuine efforts to stop smoking because of direct or implied health claims about reduced-yield products.

INTERACTIONS WITH OTHER EXPOSURES People who use tobacco and who also have substantial exposure to certain other occupational and/or environmental factors are known to incur a greater risk of some types of cancer than would be expected from either exposure alone (IARC, 2004). Only for a few exposures has the interaction with tobacco been evaluated systematically to assess whether statistical interaction exists on either an absolute or a relative scale. Three comprehensive reviews evaluated the epidemiologic studies that assessed the potential interaction between asbestos exposure and cigarette smoking with respect to lung cancer (Erren et al., 1999; Lee, 2001; Liddell, 2001). All of them found a departure from additivity but could not quantify with certainty the degree of statistical interaction on a multiplicative scale. The effect of cigarette smoking when combined with exposure to radon gas or other forms of ionizing radiation in relation to lung cancer has been reviewed comprehensively by the U.S. National Research Council’s Biological Effects of Ionizing Radiation VI (National Research Council Committee on Health Risks of Exposure to Radon, 1998). In general, the statistical interaction between smoking and radon appears submultiplicative but without strong evidence against multiplicative interaction (IARC, 2004). The combined effect of tobacco use and alcohol consumption has been examined extensively for cancers of the oral cavity, pharynx, larynx, and esophagus and to a lesser extent for cancers of the liver and pancreas (IARC, 2004). In the larger studies cancer risk consistently increased more rapidly with the combination of smoking and heavy drinking than with either exposure alone. Case-control studies that evaluated statistical interaction formally demonstrated a greater than multiplicative relation associated with joint exposure. The evidence is intriguing, albeit limited, that tobacco use combined with certain infectious agents may foster transformation of premalignant abnormalities into invasive cancer of the liver, stomach, uterine cervix, or lung. Several studies have suggested that tobacco smoking, in combination with infection with the hepatitis B virus (Chen et al., 1991; Yang et al., 2002; Sun et al., 2003), human papillomavirus (Ylitalo et

al., 1999; Kjellberg et al., 2000), H. pylori (Zaridze et al., 2000; Siman et al., 2001; Brenner et al., 2002), or tuberculosis (TATA Institute of Fundamental Research, WHO, and CDC, 2000; Brenner et al., 2001) may promote malignant progression. The evidence currently available regarding possible interactions between tobacco use and diet is also limited.

CIGARS, PIPES, AND SMOKED PRODUCTS BESIDES CIGARETTES Cigar and/or pipe smoking is strongly related to cancers of the oropharynx in many studies (Hammond and Horn, 1958; Kahn, 1966; Doll and Peto, 1976; Carstensen et al., 1987; Shanks and Burns, 1998; Iribarren et al., 1999; Shapiro et al., 2000). Persons who exclusively smoke pipes or cigars also have increased risk of cancer of the larynx (Kahn, 1966; Franceschi et al., 1990; Shanks and Burns, 1998; Shapiro et al., 2000), esophagus (Kahn, 1966; Doll and Peto, 1976; Carstensen et al., 1987; Shanks and Burns, 1998; Shapiro et al., 2000), lung (Hammond and Horn, 1958; Kahn, 1966; Doll and Peto, 1976; Lubin and Blot, 1984; Benhamou et al., 1986; Damber and Larsson, 1986; Carstensen et al., 1987; Steineck et al., 1988; Qiao et al., 1989; Chow, et al., 1992; Lange et al., 1992; Lubin et al., 1992; Tverdal et al., 1993; Ben Shlomo et al., 1994; Wald and Watt, 1997; Shanks and Burns, 1998; Boffetta et al., 1999; Iribarren et al., 1999, 2001), stomach (Kahn, 1966; Kneller et al., 1991; Tverdal et al., 1993), colon and/or rectum (Doll and Peto, 1976; Tverdal et al., 1993; Heineman et al., 1994; Hsing et al., 1998; Knekt et al., 1999; Chao et al., 2002), liver (Carstensen et al., 1987), pancreas (Kahn, 1966; Carstensen et al., 1987; Tverdal et al., 1993; Muscat et al., 1997; Partanen et al., 1997; Shanks and Burns, 1998; Shapiro et al., 2000; Iribarren et al., 2001), biliary tract (Chow et al., 1994), urinary bladder (Carstensen et al., 1987; Jensen et al., 1987; Shanks and Burns, 1998; Shapiro et al., 2000; Pitard et al., 2001), and renal pelvis (Jensen et al., 1988). The magnitude of risk for upper aerodigestive cancers is similar to that from cigarette smoking and is amplified by the combination of cigar and/or pipe smoking and alcohol consumption. Smoking of other tobacco products such as bidis is associated with cancers of the upper aerodigestive tract and to some extent lung cancer (IARC, 2004).

SECONDHAND SMOKE Many scientific consensus committees have concluded that exposure to environmental tobacco smoke (ETS) causes lung cancer in humans (National Research Council, 1986; U.S. Department of Health and Human Services, 1986; Australian National Health and Medical Research Council, 1987; U.S. Environmental Protection Agency, 1992; California Environmental Protection Agency, 1997; U.S. Department of Health and Human Services, 2002; IARC, 2004). ETS exposure (also called secondhand smoke, passive smoking, and involuntary exposure to tobacco smoke) was associated with increased lung cancer risk among nonsmokers married to smokers in more than 50 studies (IARC, 2004). A meta-analysis of these studies (Hackshaw et al., 1997) reported a pooled relative risk estimate of 1.24 (95% CI 1.13–1.36). Further evidence for causation is that persons with involuntary exposure to ETS breathe the same multitude of carcinogenic and toxic substances to which active smokers are exposed and that they absorb, metabolize, and excrete higher concentrations of tobaccospecific carcinogens than persons unexposed to ETS (Scherer and Richter, 1997; IARC, 2004). In contrast, studies have shown no consistent association between ETS exposure and breast cancer. The two largest prospective studies that control adequately for other risk factors for breast cancer found no association with ETS exposure (Wartenberg et al., 2000; Egan et al., 2002). Several other cohort studies (Hirayama, 1984; Jee et al., 1999) and case-control studies (Sandler et al., 1985a, 1985b; Smith et al., 1994; Morabia et al., 1996; Lash and Aschengrau, 1999; Johnson et al., 2000; Chang-Claude et al., 2002; Kropp and Chang-Claude, 2002) have reported an association between breast cancer risk and

Tobacco exposure to spousal smoking, whereas other studies have not (Marcus et al., 2000b; Nishino et al., 2001). The positive studies are difficult to interpret, however, either because they do not control for reproductive risk factors (Hirayama, 1984; Jee et al., 1999) or used as their referent group a small subset of women who reportedly had no exposure to either active smoking or ETS from any source (Morabia et al., 1996; Lash and Aschengrau, 1999; Johnson et al., 2000; Kropp and Chang-Claude, 2002). Given the ubiquitous presence of ETS exposure in many Western countries during the last half-century, there is concern that women who report no exposure to ETS are atypical in other respects that introduce bias (IARC, 2004).

SPIT TOBACCO The use of spit tobacco products (moist snuff, chewing tobacco, tobacco combined with betel leaf) is associated with increased risk of cancers of the oral cavity, pharynx, larynx, and esophagus (IARC, 2004). Relative risk estimates for cancer of the oral cavity range from 3 to 20 or more, with the risk increasing with the duration of use. Premalignant changes (leukoplakia) and cancers occur frequently on the anterior tongue, buccal mucosa, or gingiva, areas in direct contact with tobacco. Risk is higher for persons who use oral snuff than chewing tobacco, possibly because of the close mucosal contact generated by snuff dipping. In many cases, the premalignant changes may regress after cessation of exposure to tobacco products.

BIOLOGY OF TOBACCO USE AND CANCER Much of the evidence of the carcinogenicity of tobacco and tobacco smoke to humans derives from epidemiologic studies and is supported by extensive mechanistic evidence (Hecht, 2002a). As noted, many chemicals in the tobacco leaf and/or smoke cause cancer in humans or in experimental studies of animals. Certain carcinogens in tobacco smoke are known to cause specific types of cancer in occupationally exposed populations. Benzene, for instance, is an established cause of myeloid leukemia (Jarvis, 1987). It has been estimated that the concentration of benzene in tobacco smoke accounts for 12%–58% of the increased risk of acute myeloid leukemia attributable to smoking (Korte et al., 2000). The aromatic amines 4-aminobiphenyl and 2-naphthylamine are known to cause bladder cancer in occupationally exposed populations (IARC, 1987) and are thought to contribute to the increased risk of bladder cancer in smokers (Doll, 1998b). N-Nitrosodimethylamine, known to cause renal and other cancers in rats and rat offspring (U.S. Department of Health and Human Services, 2002), is also found in tobacco smoke (Doll, 1996). Carcinogens and procarcinogens from tobacco smoke reach most tissues throughout the body through either direct exposure to tobacco leaf and/or smoke or indirect exposure to substances dissolved in saliva and swallowed, absorbed, circulated systemically in the bloodstream, or accumulated and excreted in urine or stool. The fact that smoking was associated with so many types of cancer was initially counted as evidence against these relations being causal (Berkson, 1955, 1958). However, tobacco is now recognized irrefutably as the cause of multiple types of human cancer (IARC, 2004). Mutagens and carcinogens are detectable as adducts bound to DNA and proteins throughout the body. A tobacco-specific n-nitrosamine (Prokopczyk et al., 1997) and metabolites of benzo(a)pyrene (Melikian et al., 1999) have been found in the cervical mucus of smokers, and DNA adducts have been identified in cervical tissue (Melikian et al., 1999). Metabolites of the N-nitroso compound 4-(methylnitrosamino)-1-(3-pyridyl)1-butanone (NNK) have also been identified in the urine of smokers (Hecht, 2002b). Adducts of 4-aminobiphenyl have been demonstrated on the DNA of exfoliated bladder cells (Talaska et al., 1991), bladder biopsy tissues, and hemoglobin from smokers (Vineis et al., 1996; Airoldi et al., 2002; Gonzalez et al., 2003). Advances in biochemistry and molecular biology also provide considerable evidence about the biologic mechanisms by which tobacco

233

use may cause cancer. In the case of lung and other aerodigestive tract cancers, polycyclic aromatic hydrocarbons (PAHs) present in tobacco smoke have been shown to induce more frequent guanine to thymine transversions in circumscribed parts of the p53 and ras genes (Hainaut and Pfeifer, 2001; Gonzalez et al., 2003). These transversions have been proposed to be a molecular signature of the mutational effects of PAHs in tobacco smoke (Hainaut and Pfeifer, 2001). Tobacco use causes progressive genotypic and phenotypic changes in many organs that correspond to the stages of neoplastic development (Thun et al., 2002). Systematic autopsy studies of tissues affected by tobacco demonstrate an increase in the prevalence and/or severity of precancerous lesions and “field changes” from tobacco use in the larynx (Auerbach et al., 1970), bronchi (Auerbach et al., 1961, 1962), esophagus (Auerbach et al., 1965), and pancreas (Auerbach and Garfinkel, 1986). Both smoked and spit tobacco products cause leukoplakia in the oropharynx, a premalignant lesion that generally regresses after discontinuation of tobacco use. Biopsy studies of head and neck cancers demonstrate a “field effect,” in which clones of genetically damaged cells extend beyond the microscopically visible abnormalities (Westra and Sidransky, 1998). An autopsy study of pancreatic tissue noted an increased prevalence of atypical nuclei in ductal and parenchymal cells among smokers (Auerbach and Garfinkel, 1986).

ANIMAL STUDIES OF CARCINOGENICITY Experimental studies have established that tobacco smoke and its condensate are carcinogenic in various animal species (IARC, 2004). Topical application of tobacco smoke condensate to the skin of mice or rabbits induces benign and malignant epithelial tumors. Prolonged inhalation of tobacco smoke induces carcinoma of the larynx in Syrian hamsters. Topical application of benzo(a)pyrene to the cheek pouch of hamsters induces cancers of the oral cavity (Cheng et al., 1994). Injection of tobacco smoke condensates into gingival tissues of rabbits induces leukoplakia (Roffo, 1930). Tobacco smoke condensate and specific chemicals in tobacco smoke cause cancers of the rodent esophagus and forestomach when administered orally by gavage (U.S. Department of Health and Human Services, 2002). N-Nitrosodiethylamine induces esophageal cancer in the offspring of pregnant mice after intrauterine exposure by diet or gavage. It is difficult to identify animal models that exactly replicate the pulmonary exposure from smoking in humans because tobacco smoke is irritating and highly toxic to other species. Only humans inhale tobacco smoke voluntarily. Involuntary exposure causes other species to modify their breathing patterns toward shallow inhalation. Consequently, although inhalation exposure does cause lung tumors occasionally in some species, the principal evidence that tobacco causes lung cancer in humans derives from epidemiologic studies of humans.

OPPORTUNITIES FOR PREVENTION Two complementary approaches are needed to reduce the devastating effects of tobacco on health (U.S. Department of Health and Human Services, 2000). In the long term, progress depends on the systematic application of primary prevention measures that can reduce the initiation of tobacco use by young people and end the pandemic during the second half of the twenty-first century. In the near term, substantial reductions in smoking-attributable cancers and other diseases can be achieved by providing counseling and treatment to facilitate cessation among the 46 million Americans who currently smoke. A variety of community-based interventions have proven effective in reducing tobacco consumption and decreasing smoking initiation when these measures are applied as part of comprehensive tobacco control (Centers for Disease Control and Prevention, 1999). They include regulatory approaches (restrictions on tobacco marketing, laws ensuring clean indoor air, enforcement of laws restricting minors’ access to tobacco), economic approaches (increasing the price of cigarettes through excise taxes, thereby decreasing demand), and

234

PART III: THE CAUSES OF CANCER

countermarketing campaigns to redefine social norms regarding tobacco use. States such as California, Massachusetts, and Florida have implemented comprehensive tobacco control programs that have been particularly effective in reducing youths’ tobacco use (Bauer and Johnson, 2001). Nationally, the prevalence of cigarette smoking among U.S. high school students decreased sharply between 1997 and 2001 for males and females of the three largest racial and ethnic subgroups despite large increases in marketing expenditures by the tobacco industry (Burns et al., 2001b; Everett and Warren, 2001; Johnston, 2001; Kopstein, 2001; Centers for Disease Control and Prevention, 2002c). These temporal trends demonstrate that it is possible to prevent the uptake of tobacco use by young people if measures that have been proven to be effective are applied. About 70% of current smokers in the United States report that they want to quit smoking (Centers for Disease Control and Prevention, 2002b); 41% succeed in quitting for at least 1 day, but only about 5% remain abstinent at 3–12 months (Centers for Disease Control and Prevention, 2002b). Several pharmacologic and behavioral treatments have been shown to increase success rates (Centers for Disease Control and Prevention, 2002b), although they are widely underused (Fiore et al., 2000). Approved therapies include nicotine replacement (gum, patch, inhaler, nasal spray), sustained-release bupropion hydrochloride (Centers for Disease Control and Prevention, 2000; Anderson et al., 2001), and behavioral treatment. The latter methods are most effective when they include practical counseling (problem solving/skills training), social support as part of treatment, and help with securing social support outside of treatment (Anderson et al., 2001). The combination of pharmacologic and behavioral treatment enables 20%–25% of persons attempting to quit to remain abstinent 1 year after treatment (U.S. Department of Health and Human Services, 2000). Tobacco dependence frequently involves relapse, however, and is best characterized as a chronic disease that may require periodic treatment (U.S. Department of Health and Human Services, 2000). References Adami HO, Bergstrom R, Engholm G, et al. 1996. A prospective study of smoking and risk of prostate cancer. Int J Cancer 67:764–768. Adami J, Nyrem O, Bergstrom R, et al. 1998. Smoking and risk of leukemia, lymphoma, and multiple myeloma (Sweden). Cancer Causes Control 9:49–56. Ahrens W, Jockel KH, Patzak W, Elsner G. 1991. Alcohol, smoking, and occupational factors in cancer of the larynx: a case-control study. Am J Ind Med 20:477–493. Ahsan H, Neugut A, Gammon M. 1997. Association of adenocarcinoma and squamous cell carcinoma of the esophagus with tobacco-related and other malignancies. Cancer Epidemiol Biomarkers Prev 10:779–782. Airoldi L, Orsi F, Magagnotti C, et al. 2002. Determinants of 4-aminobiphenylDNA adducts in bladder cancer biopsies. Carcinogenesis 23:861–866. Almendingen K, Hofstad B, Trygg K, Hoff G, Hussain A, Vatn MH. 2000. Smoking and colorectal adenomas: a case-control study. Eur J Cancer Prev 9:193–203. Ambrosone C, Shields P. 2001. Smoking as a risk factor for breast cancer. In: Bowcock A, editor. Breast Cancer: Molecular Genetics, Pathogenesis, and Therapeutics, Vol. 146. Totowa, NJ: Humana Press, pp. 519–536. Ambrosone C, Freudenheim J, Graham S, et al. 1996. Cigarette smoking, Nacetyltransferase-2 genetic polymorphisms, and breast cancer risk. JAMA 276:1494–1501. Anderson CM, Burns DM, Major JM, Vaughn JW, Shanks TG. 2001. Changes in adolescent smoking behaviors in sequential birth cohorts. In: Burns D, Amacher R, Ruppert W, editors. National Cancer Institute. Changing Adolescent Smoking Prevalence. Smoking and Tobacco Control Monograph 14. NIH Publ. No. 02–5086. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, pp. 141–155. Anonymous. 1990. Deaths from chronic obstructive pulmonary disease in the United States, 1987. Stat Bull Metropolitan Insurance 71:20–26. Arinami T, Ishiguro H, Onaivi ES. Polymorphisms in genes involved in neurotransmission in relation to smoking. Eur J Pharmacol 2000; 410:215–226. Armadans-Gil L, Vaque-Rafart J, Rossello J, Olona M, Alseda M. 1999. Cigarette smoking and male lung cancer risk with special regard to type of tobacco. Int J Epidemiol 28:614–619. Armstrong RW, Imrey PB, Lye MS, Armstrong MJ, Yu MC, Sani S. 2000. Nasopharyngeal carcinoma in Malaysian Chinese: occupational

exposures to particles, formaldehyde and heat. Int J Epidemiol 29: 991–998. Asal NR, Risser DR, Kadamani S, Geyer JR, Lee ET, Cherng N. 1988. Risk factors in renal cell carcinoma. I. Methodology, demographics, tobacco, beverage use, and obesity. Cancer Detect Prev 11:359–377. Atawodi SE, Lea S, Nyberg F, et al.1998. 4-Hydroxy-1–3-(pyridyl)-1butanone-hemoglobin adducts as biomarkers of exposure to tobacco smoke: validation of a method to be used in multicenter studies. Cancer Epidemiol Biomarkers Prev 7:817–821. Aubry F, MacGibbon B. 1985. Risk factors of squamous cell carcinoma of the skin: a case-control study in the Montreal region. Cancer 55:907–911. Auerbach O, Garfinkel L. 1986. Histologic changes in pancreas in relation to smoking and coffee-drinking habits. Dig Dis Sci 31:1014–1020. Auerbach O, Hammond E, Garfinkel L. 1070. Histologic changes in the larynx in relation to smoking habits. Cancer 25:92–104. Auerbach O, Stout A, Hammond E, Garfinkel L. 1961. Changes in bronchial epithelium in relation to cigarette smoking and in relation to lung cancer. N Engl J Med 265:253–267. Auerbach O, Stout A, Hammond E, Garfinkel L. 1962. Bronchial epithelium in former smokers. N Engl J Med 267:119–125. Auerbach O, Stout A, Hammond E, Garfinkel L. 1965. Histologic changes in esophagus in relation to smoking habits. Arch Environ Health 11:4–15. Austin H, Cole P. 1986. Cigarette smoking and leukemia. J Chronic Dis 39:417–421. Australian National Health and Medical Research Council. 1987. Effects of Passive Smoking on Health. Report of the NHMRC Working Party on the Effects of Passive Smoking on Health. Canberra: Government Publishing Service. Bain C, Feskanich D, Speizer FE, Thun M, Hertzmark E, Rosner BA, Colditz GA. 2004. Lung cancer rates in men and women with comparable histories of smoking. J Natl Cancer Inst 96:826–834. Balfour D, Fagerstrom K. 1996. Pharmacology of nicotine and its therapeutic use in smoking cessation and neurodegenerative disorders. Pharmacol Ther 72:51–81. Baron A, Franceschi S, Barra S, Talamini R, La Vecchia C. 1993. A comparison of the joint effects of alcohol and smoking on the risk of cancer across sites in the upper aerodigestive tract. Cancer Epidemiol Biomarkers Prev 2:519–523. Baron JA, La Vecchia C, Levi F. 1990. The antiestrogenic effect of cigarette smoking in women. Am J Obstet Gynecol 162:502–514. Bartsch H, Malaveille C, Friesen M, Kadlubar FF, Vineis P. 1993. Black (aircured) and blond (flue-cured) tobacco cancer risk. IV. Molecular dosimetry studies implicate aromatic amines as bladder carcinogens. Eur J Cancer 8:1199–1207. Bartsch H, Nair U, Risch A, Rojas M, Wikman H, Alexandrov K. 2000. Genetic polymorphism of CYP genes, alone or in combination, as a risk modifier of tobacco-related cancers. Cancer Epidemiol Biomarkers Prev 9:3–28. Bauer U, Johnson T. 2001. Predictors of tobacco use among adolescents in Florida, 1998–1999. In: Burns D, Amacher R, Ruppert W, editors. Changing Adolescent Smoking Prevalence. Smoking and Tobacco Control Monograph No. 14. NIH Publ. No. 02–5086. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, pp. 121–140. Ben Shlomo Y, Smith G, Shipley M, Marmot M. 1994. What determines mortality risk in male former cigarette smokers? Am J Public Health 84:1235–1242. Benhamou E, Benhamou S. 1993. Black (air-cured) and blond (flue-cured) tobacco and cancer risk. VI. Lung cancer. Eur J Cancer 12:1778–1780. Benhamou S, Benhamou E, Flamant R. 1986. Lung cancer risk associated with cigar and pipe smoking. Int J Cancer 37:825–829. Bennett W, Alavanja M, Blomeke B, et al. 1999. Environmental tobacco smoke, genetic susceptibility, and risk of lung cancer in never-smoking women. J Natl Cancer Inst 91:2009–2014. Benowitz N. 1996a. Pharmacology of nicotine: addiction and therapeutics. Annu Rev Pharmacol Toxicol 36:597–613. Benowitz NL. 1996b. Cotinine as a biomarker of environmental tobacco smoke exposure. Epidemiol Rev 18:188–204. Benowitz N. 2001. Compensatory smoking of low-yield cigarettes. In: Shopland D, Burns D, Benowitz N, Amacher R, editors. Risks Associated with Smoking Cigarettes with Low Machine-Measured Yields of Tar and Nicotine. Smoking and Tobacco Control Monograph No. 13. NIH Publ. No. 02-5074. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, pp. 39–63. Berkson J. 1955. The statistical study of association between smoking and lung cancer. Proc Staff Meet Mayo Clin 30:319–348. Berkson J. 1958. Smoking and lung cancer: some observations on two recent reports. J Am Stat Assoc 53:28–38.

Tobacco Berta L, Frairia R, Fortunati N, Fazzari A, Gaidano G. 1992. Smoking effects on the hormonal balance of fertile women. Horm Res 37:45–48. Best E, Josie G, Walker C. 1961. A Canadian study of mortality in relation to smoking habits: a preliminary report. Can J Public Health 52:99– 106. Blot WJ, McLaughlin JK, Winn DM, et al. 1988. Smoking and drinking in relation to oral and pharyngeal cancer. Cancer Res 48:3282–3287. Boffetta P. 1993. Black (air-cured) and blond (flue-cured) tobacco and cancer risk. V. Oral cavity cancer. Eur J Cancer 9:1331–1335. Boffetta P, Pershagen G, Jockel K, et al. 1999. Cigar and pipe smoking and lung cancer risk: a multicenter study from Europe. J Natl Cancer Inst 91:697–701. Boffetta P, Stellman SD, Garfinkel L. 1989. A case-control study of multiple myeloma nested in the American Cancer Society prospective study. Int J Cancer 43:554–559. Boutron MC, Faivre J, Dop MC, Quipourt V, Senesse P. 1995. Tobacco, alcohol, and colorectal tumors: a multistep process. Am J Epidemiol 141:1038–1046. Bradford JA, Harlan WR, Hamner HR. 1936. Nature of cigarette smoke: technic of experimental smoking. Ind Eng Chem 28:836–839. Brennan P, Bogillot O, Cordier S, et al. 2000. Cigarette smoking and bladder cancer in men: a pooled analysis of 11 case control studies. Int J Cancer 86:289–294. Brennan P, Bogillot O, Greiser E, et al. 2001. The contribution of cigarette smoking to bladder cancer in women: pooled European data. Cancer Causes Control 12:411–417. Brenner AV, Wang Z, Kleinerman RA, et al. 2001. Previous pulmonary diseases and risk of lung cancer in Gansu Province, China. Int J Epidemiol 30:118–124. Brenner H, Arndt V, Bode G, Stegmaier C, Ziegler H, Stumer T. 2002. Risk of gastric cancer among smokers infected with Helicobacter pylori. Int J Cancer 98:446–449. Breuer-Katschinski B, Nemes K, Marr A, et al. 2000. Alcohol and cigarette smoking and the risk of colorectal adenomas. Dig Dis Sci 45:487–493. Brinton LA, Barrett RJ, Berman ML, Mortel R, Twiggs LB, Wilbanks GD. 1993. Cigarette smoking and the risk of endometrial cancer. Am J Epidemiol 137:281–291. Brinton LA, Blot WJ, Becker JA, et al. 1984. A case-control study of cancers of the nasal cavity and paranasal sinuses. Am J Epidemiol 119:896–906. Brinton LA, Li JY, Rong SD, et al. 1991. Risk factors for penile cancer: results from a case-control study in China. Int J Cancer 47:504–509. Brinton LA, Nasca PC, Mallin K, Baptiste K, Wilbanks GD, Richart RM. 1990. Case control study of smoking and cancer of the vulva. Obstet Gynecol 75:859–866. Brinton LA, Schairer C, Stanford JL, Hoover RN. 1986. Cigarette smoking and breast cancer. Am J Epidemiol 123:614–622. Brockton N, Little J, Sharp L, Cotton S. 2000. N-Acetyltransferase polymorphisms and colorectal cancer: a HuGE review. Am J Epidemiol 151:846–861. Brown LM, Everett G, Gibson R, Burmeister L, Schuman L, Blair A. 1992. Smoking and risk of non-Hodgkin’s lymphoma and multiple myeloma. Cancer Causes Control 3:49–55. Brown LM, Pottern LM, Hoover RN. 1987. Testicular cancer in young men: the search for causes of the epidemic increase in the United States. J Epidemiol Community Health 41:349–354. Brownson RC, Chang JC. 1987. Exposure to alcohol and tobacco and the risk of laryngeal cancer. Arch Environ Health 42:192–196. Brunnemann K, Hoffmann D. 1992. Chemical composition of smokeless tobacco products. In: Stotts RC, Schroeder KL, Burns DM, Shopland DR, editors. Smokeless Tobacco or Health—An International Perspective. Monograph 2. Bethesda, MD: National Institutes of Health, National Cancer Institute, pp. 96–108. Burch JD, Howe GR, Miller AB, Semenciw R. 1981. Tobacco, alcohol, asbestos, and nickel in the etiology of cancer of the larynx: a case-control study. J Natl Cancer Inst 67:1219–1224. Burns DM. 1994. Tobacco smoking. In: Samet JM, editor. Epidemiology of lung cancer. New York: Marcel Dekker, pp. 15–49. Burns D, Benowitz NL. 2001. Public health implications of changes in cigarette design and marketing. In: Shopland D, Burns D, Benowitz N, Amacher R, editors. Risks Associated with Smoking Cigarettes with Low Machine-Measured Yields of Tar and Nicotine. Smoking and Tobacco Control Monograph 13. NIH Publ. No. 02-5074. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, pp. 1–12. Burns D, Lee L, Shen L, et al. 1997a. Cigarette smoking behavior in the United States. In: Shopland D, Burns D, Garfinkel L, Samet J, editors. Changes in Cigarette-Related Disease Risks and Their Implication for Prevention and Control. Smoking and Tobacco Control Monograph 8. NIH Publ. No.

235

97-4213. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, pp. 13–112. Burns D, Major J, Shanks T, Thun M, Samet J. 2001a. Smoking lower yield cigarettes and disease risks. In: Shopland D, Burns D, Benowitz N, Amacher R, editors. Risks Associated with Smoking Cigarettes with Low Machine-Measured Yields of Tar and Nicotine. Smoking and Tobacco Control Monograph 13. NIH Publ. No. 02–5074. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, pp. 65–158. Burns D, Majors J, Vaughn J, Anderson C, Shanks T. 2001b. Cigarette smoking among adolescents in California, 1990–1996. Changing Adolescent Smoking Prevalence. Smoking and Tobacco Control Monograph No. 14. NIH Publ. No. 02–5086. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, pp. 69–100. Burns D, Shanks T, Choi W, Thun M, Heath CJ, Garfinkel L. 1997b. The American Cancer Society Cancer Prevention Study I. 12-Year Follow-up of One Million Men and Women. In: Shopland D, Burns D, Garfinkel L, Samet J, editors. Changes in Cigarette-Related Disease Risks and Their Implication for Prevention and Control. Smoking and Tobacco Control Monograph 8. NIH Publ. No. 97-4213. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, pp. 113–304. California Environmental Protection Agency. 1997. Health Effects of Exposure to Environmental Tobacco Smoke: Final Report. Sacramento: California Environmental Protection Agency, Office of Environmental Health Hazard Assessment. Cao SM, Liu Q, Huang QH, Yang CW, Huang TB. 2000. Analysis for risk factors of nasopharyngeal carcinoma in Sihui City. Cancer 19:987–989. Caplan LS, Hall HI, Levine RS, Zhu K. 2000. Preventable risk factors for nasal cancer. Ann Epidemiol 10:186–191. Carmella SG, Akerkar S, Hecht SS. 1995. Intraindividual and interindividual differences in metabolites of the tobacco-specific lung carcinogen 4(methylnitrosoamino)-1-(3-pyridyl)-1-butanone (NNK) in smokers’ urine. Cancer Epidemiol Biomarkers Prevent 4:635–644. Carstensen J, Pershagen G, Eklund G. 1987. Mortality in relation to cigarette and pipe smoking: 16 years’ observation of 25,000 Swedish men. J Epidemiol Community Health 41:166–172. Cassidenti DL, Pike MC, Vijod AG, Stanczyk FZ, Lobo RA. 1992. A reevaluation of estrogen status in postmenopausal women who smoke. Am J Obstet Gynecol 166:1444–1448. Castellsague X, Munoz N, De Stefani E, et al. 1999. Independent and joint effects of tobacco smoking and alcohol drinking on the risk of esophageal cancer in men and women. Int J Cancer 82:657–664. Cauley JA, Gutai JP, Kuller LH, LeDonne D, Powell JG. 1989. The epidemiology of serum sex hormones in postmenopausal women. Am J Epidemiol 129:1120–1131. Centers for Disease Control and Prevention. 1994. Health objectives for the nation; cigarette smoking among adults—United States, 1993. MMWR Morb Mortal Wkly Rep 43:625–630. Centers for Disease Control and Prevention. 1999. Best Practices for Comprehensive Tobacco Control Programs. Atlanta: CDC, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. Centers for Disease Control and Prevention. 2000. Use of FDA-approved pharmacologic treatments for tobacco dependence—United States, 1984–1998. MMWR Morb Mortal Wkly Rep 49:665–668. Centers for Disease Control and Prevention. 2002a. Annual smokingattributable mortality, years of potential life lost, and economic costs— United States, 1995–1999. MMWR Morb Mortal Wkly Rep 51:300–303. Centers for Disease Control and Prevention. 2002b. Cigarette smoking among adults—United States, 2002. MMWR Morb Mortal Wkly Rep 51:642–645. Centers for Disease Control and Prevention. 2002c. Youth risk behavior surveillance—United States, 2001. MMWR Surveill Summ 51(SS-4). Chalasani N, Baluyat A, Ismail A, et al. 2000. Cholangiocarcinoma in patients with primary sclerosing cholangitis: a multicenter case-control study. Hepatology 1:247–248. Chang-Claude J, Kropp S, Jager B, Bartsch H, Risch A. 2002.Differential effect of NAT2 on the association between active and passive smoke exposure and breast cancer risk. Cancer Epidemiol Biomarkers Prev 11:698–704. Chao A, Thun M, Henley S, Jacobs E, McCullough M, Calle E. 2002. Cigarette smoking, use of other tobacco products and stomach cancer mortality in US adults: The Cancer Prevention Study II. Int J Cancer 101:380–389. Chen CJ, Liang KY, Chang AS, et al. 1991. Effects of hepatitis B virus, alcohol drinking, cigarette smoking and familial tendency on hepatocellular carcinoma. Hepatology 13:398–406.

236

PART III: THE CAUSES OF CANCER

Cheng K, Duffy S, Day S, Lam T. 1994. Oesophageal cancer in never-smokers and never-drinkers. Int J Epidemiol 23:444–450. Choi SY, Kahyo H. 1991. Effect of cigarette smoking and alcohol consumption in the etiology of cancers of the digestive tract. Int J Cancer 49:381–386. Chow W, McLaughlin J, Hrubec Z, Fraumeni J. 1995. Smoking and biliary tract cancer in a cohort of US Veterans. Br J Cancer 72:1556–1558. Chow W, McLaughlin J, Menck H, Mack T. 1994. Risk factors for extrahepatic bile duct cancers: Los Angeles County, California (USA). Cancer Causes Control 5:267–272. Chow W, Schuman L, McLaughlin J, et al. 1992. A cohort study of tobacco use, diet, occupation, and lung cancer mortality. Cancer Causes Control 3:247–254. Chow W-H, McLaughlin J, Hrubec Z, Nam J-M, Blot W. 1993. Tobacco use and nasopharyngeal carcinoma in a cohort of US veterans. Int J Cancer 55:528–540. Chu SY, Stroup NE, Wingo PA, Lee NC, Peterson HB, Gwinn ML. 1990. Cigarette smoking and the risk of breast cancer. Am J Epidemiol 131:244–253. Chute C, Willet W, Colditz G, et al. 1991. A prospective study of body mass, height, and smoking and risk of colorectal cancer in women. Cancer Causes Control 2:117–124. Chyou P-H, Nomura A, Stemmermann G. 1996. A prospective study of colon and rectal cancer among Hawaii Japanese men. Ann Epidemiol 6:276–282. Clavel J, Cordier S, Boccon-Gibod L, Hemon D. 1989. Tobacco and bladder cancer in males: increased risk for inhalers and smokers of black tobacco. Int J Cancer 44:605–610. Connolly GN, Winn DM, Hecht SS, Henningfield JE, Walker B Jr, Hoffmann D. 1986. The reemergence of smokeless tobacco. N Engl J Med 314:1020–1027. Consensus Statement. 2000. A clinical practice guideline for treating tobacco use and dependence. JAMA 283:3244–3254. Cook L, Doherty J, Weiss N, Chen C. 2003. Endometrial cancer: epidemiology and molecular endocrinology. In: Henderson B, Ponder B, Ross R, editors. Hormones, Genes and Cancer. London: Oxford University Press. Cooper JA, Rohan TE, Cant EL, Horsfall DJ, Tilley WD. 1989. Risk factors for breast cancer by estrogen receptor status: a population-based casecontrol study. Br J Cancer 59:119–125. Cope GF, Wyatt JI, Pinder IF, Lee PN, Heatley RV, Kelleher J. 1991. Alcohol consumption in patients with colorectal adenomatous polyps. Gut 32:70–72. Corrao M, Guindon G, Sharma N, Shokoohi D, editors. 2000. Tobacco Control: Country Profiles. Atlanta: American Cancer Society. Coughlin SS, Neaton JD, Sengupta A. 1996. Cigarette smoking as a predictor of death from prostate cancer in 348,874 men screened for the Multiple Risk Factor Intervention Trial. Am J Epidemiol 143:1002–1006. Daling JR, Sherman KJ, Hislop TG, et al. 1992. Cigarette smoking and the risk of anogenital cancer. Am J Epidemiol 135:180–189. Dallinga JW, Pachen DM, Winjnhoven SW, et al. 1998. The use of 4aminobiphenyl hemoglobin adducts and aromatic DNA adducts in lymphocytes of smokers as biomarkers of exposure. Cancer Epidemiol Biomarkers Prev 7:571–577. Damber L, Larsson L. 1986. Smoking and lung cancer with special regard to type of smoking and type of cancer: a case-control study in north Sweden. Br J Cancer 53:673–681. D’Avanzo B, Negri E, La Vecchia C, et al. 1990. Cigarette smoking and bladder cancer. Eur J Cancer 26:714–718. De Hertog SA, Wensveen CA, Bastiaens MT, et al. 2001. Relation between smoking and skin cancer. J Clin Oncol 19:231–238. De Stefani E, Barrios E, Fierro L. 1993. Black (air-cured) and blond (fluecured) tobacco and cancer risk. III. Oesophageal cancer. Eur J Cancer 29A:763–766. De Stefani E, Boffetta P, Oreggia F, Mendilaharsu M, Deneo-Pellegrini H. 1998. Smoking patterns and cancer of the oral cavity and pharynx: a casecontrol study in Uruguay. Oral Oncol 34:340–346. De Stefani E, Correa P, Fierro L, Fontham E, Chen V, Zavala D. 1991. Black tobacco, mate, and bladder cancer: a case-control study from Uruguay. Cancer 67:536–540. De Stefani E, Correa P, Oreggia F, et al. 1987. Risk factors for laryngeal cancer. Cancer 60:3087–3091. Deacon JM, Evans CD, Yule R, et al. 2000. Sexual behaviour and smoking as determinants of cervical HPV infection and of CIN3 among those infected: a case-control study nested within the Manchester cohort. Br J Cancer 83:1565–1572. Dillner J, von Krogh G, Horenblas S, Meijer CJ. 2000. Etiology of squamous cell carcinoma of the penis. Scand J Urol Nephrol Suppl 205:189–193.

Djordjevic M, Stellman S, Zang E. 2000. Doses of nicotine and lung carcinogens delivered to cigarette smokers. J Natl Cancer Inst 92:106–111. Doll R. 1996. Cancers weakly related to smoking. Br Med Bull 52:35–49. Doll R. 1998a. The first reports on smoking and health. In: Lock S, Reynolds L, Tansey E, editors. Ashes to Ashes: the History of Smoking and Health. Amsterdam: Rodopi, pp. 130–142. Doll R. 1998b. Uncovering the effects of smoking: historical perspective. Stat Methods Med Res 7:87–117. Doll R, Hill A. 1956. Lung cancer and other causes of death in relation to smoking: a second report on the mortality of British doctors. BMJ 2:1071–1081. Doll R, Hill A. 1964. Mortality in relation to smoking: ten years’ observations of British doctors. BMJ 1:1399–1410, 1460–1467. Doll R, Hill A. 1966. Mortality of British doctors in relation to smoking: observations on coronary thrombosis. Natl Cancer Inst Monogr 19: 205–268. Doll R, Peto R. 1976. Mortality in relation to smoking: 20 years’ observations on male British doctors. BMJ 2:1525–1536. Doll R, Peto R. 1978. Cigarette smoking and bronchial carcinoma: dose and time relationships among regular smokers and lifelong non-smokers. J Epidemiol Community Health 32:303–313. Doll R, Peto R. 1981. The causes of cancer: quantitative estimates of avoidable risks of cancer in the United States today. J Natl Cancer Inst 66:1191–1308. Doll R, Hill A, Kreyberg L. 1957. The significance of cell type in relation to the aetiology of lung cancer. Br J Cancer 2:43–48. Doll R, Peto R, Wheatley K, Gray R, Sutherland I. 1994. Mortality in relation to smoking: 40 years’ observations on male British doctors. BMJ 309:901–911. Dorn H. 1959. Tobacco consumption and mortality from cancer and other diseases. Public Health Rep 74:581–593. Dosemeci M, Gokmen I, Unsal M, Hayes RB, Blair A. 1997. Tobacco, alcohol use, and risks of laryngeal and lung cancer by subsite and histologic type in Turkey. Cancer Causes Control 8:729–737. Egan K, Stampfer MJ, Hunter D, et al. 2002. Active and passive smoking in breast cancer: prospective results from the Nurses’ Health Study. Epidemiology 13:138–145. Eichholzer M, Stahelin HB, Ludin E, Bernasconi F. 1999. Smoking, plasma vitamins C, E, retinol, and carotene, and fatal prostate cancer: seventeenyear follow-up of the prospective Basel study. Prostate 38:189–198. Elliott EA, Matanoski GM, Rosenshein NB, Grumbine FC, Diamond EL. 1990. Body fat patterning in women with endometrial cancer. Gynecol Oncol 39:253–258. Engeland A, Andersen A, Haldorsen T, Tretli S. 1996. Smoking habits and risk of cancers other than lung cancer: 28 years’ follow-up of 26,000 Norwegian men and women. Cancer Causes Control 7:497–506. Erren TC, Jacobsen M, Piekarski C. 1999. Synergy between asbestos and smoking on lung cancer risks. Epidemiology 10:405–411. Evans A, Chen G, Ross E, Shen F, Lin W, London W. 2002. Eight-year followup of the 90,000-person Haimen City cohort. I. Hepatocellular carcinoma mortality, risk factors, and gender differences. Cancer Epidemiol Biomarkers Prev 11:369–376. Everett S, Warren C. 2001. Trends and Subgroup Differences in Tobacco Use Among High School Students in the United States, 1991–1997. Changing Adolescent Smoking Prevalence. Smoking and Tobacco Control Monograph No. 14. NIH Publ. No. 02–5086. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, pp. 35–50. Ezzati M, Lopez AD. 2003a. Estimates of global mortality attributable to smoking in 2000. Lancet 362:847–852. Ezzati M, Lopez AD. 2003b. Measuring the accumulated hazards of smoking: global and regional estimates for 2000. Tob Control 12:79–85. Fagerstrom KO. 1978. Measuring degree of physical dependence to tobacco smoking with reference to individualization of treatment. Addict Behav 3:235–241. Fagerstrom KO, Kunze M, Schoberger R, et al. 1996. Nicotine dependence versus smoking prevalence: comparisons among countries and categories of smokers. Tob Control 5:52–56. Falk RT, Pickle LW, Brown LM, Mason TJ, Buffler PA, Fraumeni JF Jr. 1989. Effect of smoking and alcohol consumption on laryngeal cancer risk in coastal Texas. Cancer Res 49:4024–4029. Federal Register. 1995. Notices. Washington, DC: Department of Health and Human Services, Food and Drug Administration, pp. 274–288. Fiore M, Bailey W, Cohen S, et al. 2000. Treating Tobacco Use and Dependence. Rockville, MD: U.S. Department of Health and Human Services, Public Health Service. Fischer S, Speigelhalder B, Preussman R. 1991. Tobacco specific nitrosamines in commercial cigarettes; possibilities for reducing exposure. In: Rele-

Tobacco vance to Human Cancer on N-Nitroso Compounds, Tobacco Smoke, and Mycotoxins, Vol. 105. Lyon: IARC, pp. 489–493. Fisher R. 1957. Dangers of cigarette smoking. BMJ 2:43. Fisher R. 1958a. Cancer and smoking. Nature 182:596. Fisher R. 1958b. Lung cancer and cigarettes? Nature 182:168. Fisher R. 1959. Smoking: the Cancer Controversy. Some Attempts to Assess the Evidence. Edinburgh: Oliver & Boyd. Flanders W, Lally C, Zhu B-P, Henley S, Thun M. 2003. Lung cancer mortality in relation to age, duration of smoking, and daily cigarette consumption: results from Cancer Prevention Study II. Cancer Res 63:6556–6562. Forey B, Hamling J, Lee P, Wald N. 2002. International Smoking Statistics: A Collection of Historical Data from 30 Economically Developed Countries. London: Oxford University Press. Franceschi S, Serraino D. 1992. Risk factors for adult soft tissue sarcoma in northern Italy. Ann Oncol 3(Suppl 2):S85–S88. Franceschi S, Barra S, La Vecchia C, Bidoli E, Negri E, Talamini R. 1992. Risk factors for cancer of the tongue and mouth, a case-control study in northern Italy. Cancer 70:2227–2233. Franceschi S, Serraino D, Bidoli E, et al. 1989. The epidemiology of nonHodgkin’s lymphoma in the north-east of Italy; a hospital-based casecontrol study. Leuk Res 13:465–472. Franceschi S, Talamini R, Barra S, et al. 1990. Smoking and drinking in relation to cancers of the oral cavity, pharynx, larynx, and esophagus in northern Italy. Cancer Res 50:6502–6507. Freidman A, Ravnikar V, Barbieri R. 1987. Serum steroid hormone profiles in postmenopausal smokers and nonsmokers. Fertil Steril 47:398–401. Freudenheim JL, Graham S, Byers TE, et al. 1992. Diet, smoking, and alcohol in cancer of the larynx: a case-control study. Nutr Cancer 17:33–45. Friedman G. 1993. Cigarette smoking, leukemia, and multiple myeloma. Ann Epidemiol 3:425–428. Fukuda K, Shibata A. 1990. Exposure-response relationships between woodworking, smoking or passive smoking, and squamous cell neoplasms of the maxillary sinus. Cancer Causes Control 1:165–168. Galanti MR, Hansson L, Lund E, et al. 1996. Reproductive history and cigarette smoking as risk factors for thyroid cancer in women: a population-based case-control study. Cancer Epidemiol Biomarkers Prev 5:425–431. Gallagher RP, Huchcroft S, Phillips N, et al. 1995. Physical activity, medical history, and risk of testicular cancer. Cancer Causes Control 5:398– 406. Gammon M, Schoenberg J, Ahsan H, et al. 1997. Tobacco, alcohol, and socioeconomic status and adenocarcinomas of the esophagus and gastric cardia. J Natl Cancer Inst 89:1277–1284. Garfinkel L, Boffetta P. 1990. Association between smoking and leukemia in two American Cancer Society prospective studies. Cancer 65:2356–2360. Geisler S, Olshan A. 2001. GSTM1, GSTT1, and the risk of squamous cell carcinoma of the head and neck: a mini-HuGE review. Am J Epidemiol 154:95–105. Gerlach K, Cummings K, Hyland A, Gilpin E, Johnson M, Pierce J. 1998. Trends in cigar consumption and smoking prevalence. In: Shopland D, Burns D, Hoffmann D, Cummings K, Amacher R, editors. Cigars: Health Effects and Trends. Smoking and Tobacco Control Monograph 9. NIH Publ. No. 98–4302. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, pp. 21–53. Giovannucci E, Colditz G, Stampfer M, et al. 1994a. A prospective study of cigarette smoking and risk of colorectal adenoma and colorectal cancer in U.S. women. J Natl Cancer Inst 86:192–199. Giovannucci E, Rimm E, Ascherio A, et al. 1999. Smoking and risk of total and fatal prostate cancer in United States health professionals. Cancer Epidemiol Biomarkers Prev 8:277–282. Giovannucci E, Rimm E, Stampfer M, Colditz G, Ascherio A, Willet W. 1994b. A prospective study of cigarette smoking and risk of colorectal adenoma and colorectal cancer in U.S. men. J Natl Cancer Inst 86:183–191. Giovino G, Henningfield J, Tomar S, Escobedo L, Slade J. 1995. Epidemiology of tobacco use and dependence. Epidemiol Rev 17:48–65. Glover ED, Glover PN. 1992. The smokeless tobacco problem: risk groups in North America. In: Shopland DR, Stotts RC, Schroeder KL, Burns DM, editors. Smokeless Tobacco or Health: An International Perspective. Smoking and Tobacco Control Monograph 2. NIH Publ. No. 93–3461. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, pp. 3–18. Gonzalez CA, Pera G, Agudo A, et al. 2003. Smoking and the risk of gastric cancer in the European Prospective Investigation Into Cancer and Nutrition (EPIC). Int J Cancer 107:629–634. Goodman MT, Moriwaki H, Vaeth M, Akiba S, Hayabuchi H, Mabuchi K. 1995. Prospective cohort study of risk factors for primary liver cancer in Hiroshima and Nagasaki, Japan. Epidemiology 6:36–41.

237

Gorlewska-Roberts K, Green B, Fares M, Ambrosone CB, Kadlubar FF. 2002. Carcinogen-DNA adducts in human breast epithelial cells. Environ Mol Mutagen 39:184–192. Graham S, Mettlin C, Marshall J, Priore R, Rzepka T, Shedd D. 1981. Dietary factors in the epidemiology of cancer of the larynx. Am J Epidemiol 113:675–680. Gray N, Boyle P, Zatonski W. 1998. Tar concentrations in cigarettes and carcinogen content. Lancet 352:787–788. Green J, Banks E, Berrington A, Darby S, Deo H, Newton R. 2000. NAcetyltransferase 2 and bladder cancer: an overview and consideration of the evidence for gene-environment interaction. Br J Cancer 83: 412–417. Grodstein F, Speizer FE, Hunter DJ. 1995. A prospective study of incident squamous cell carcinoma of the skin in the nurses’ health study. J Natl Cancer Inst 87:1061–1066. Gupta P. 1992. Smokeless tobacco use in India. In: Shopland D, Stotts R, Schroeder K, Burns D, editors. Smokeless Tobacco or Health: An International Perspective. Smoking and Tobacco Control Monograph 2. NIH Publ. No. 93–3461. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, pp. 19–25. Hackshaw AK, Law MR, Wald NJ. 1997. The accumulated evidence on lung cancer and environmental tobacco smoke. BMJ 315:980–988. Haenszel W, Shimkin M, Miller H. 1956. Tobacco smoking patterns in the United States. Public Health Monograph No. 45. PHS Publ. No. 463. Washington, DC: U.S. Government Printing Office. Hainaut P, Pfeifer GP. 2001. Patterns of p53 G to T transversions in lung cancers reflect the primary mutagenic signal of DNA damage by tobacco smoke. Carcinogenesis 22:367–374. Hammond E, Horn D. 1958. Smoking and death rates: report on forty-four months of follow-up of 187,783 men. II. Death rates by cause. JAMA 166:1294–1308. Harish K, Ravi R. 1995. The role of tobacco in penile carcinoma. Br J Urol 75:375–377. Harris J. 1983. Cigarette smoking among successive birth cohorts of men and women in the United States during 1900–80. J Natl Cancer Inst 71:473–479. Harris JE, Thun MJ, Mondul AM, Calle EE. 2004. Cigarette tar yields in relation to mortality from lung cancer in the Cancer Prevention Study II prospective cohort, 1982–8. BMJ 328:72–79. Hayes RB, Bravo-Otero E, Kleinman DV, et al. 1999. Tobacco and alcohol use and oral cancer in Puerto Rico. Cancer Causes Control 10:27–33. Hecht S. 1999a. DNA adduct formation tobacco-specific N-nitrosamines. Mutat Res 424:127–142. Hecht S. 1999b. Tobacco smoke carcinogens and lung cancer. J Natl Cancer Inst 91:1194–1210. Hecht S. 2002a. Cigarette smoking and lung cancer: chemical mechanisms and approaches to prevention. Lancet Oncol 3:461–469. Hecht S. 2002b. Human urinary carcinogen metabolites: biomarkers for investigating tobacco and cancer. Carcinogenesis 23:907–922. Hecht SS. 2002c. Tobacco smoke carcinogens and breast cancer. Environ Mol Mutagen 39:119–126. Hecht SS, Carmella SG, Foiles PG, Murphy SE. 1994. Biomarkers for human uptake and metabolic activation of tobacco-specific nitrosamines. Cancer Res 54(Suppl 7):1912s–1917s. Hedberg K, Vaughan TL, White E, Davis S, Thomas DB. 1994. Alcoholism and cancer of the larynx: a case-control study in western Washington (United States). Cancer Causes Control 5:3–8. Heineman E, Zahm S, McLaughlin J, Vaught J. 1994. Increased risk of colorectal cancer among smokers: results of a 26-year follow-up of US veterans and a review. Int J Cancer 59:728–738. Heineman E, Zahm S, McLaughlin J, Vaught J, Hrubec Z. 1992. A prospective study of tobacco use and multiple myeloma: evidence against an association. Cancer Causes Control 3:31–36. Hellberg D, Valentin J, Eklund T, Nilsson S. 1987. Penile cancer: is there an epidemiological role for smoking and sexual behaviour? BMJ 295: 1306–1308. Henderson BE, Benton B, Jing J, Yu MC, Pike MC. 1979. Risk factors for cancer of the testis in young men. Int J Cancer 23:598–602. Henningfield J, Cohen C, Pickworth W. 1993. Psychopharmacology of nicotine. In: Orleans C, Slade J, editors. Nicotine Addiction, Principles and Management. New York: Oxford University Press, pp. 24–45. Henningfield J, Rose C, Giovino G. 2002. Brave new world of tobacco disease prevention: promoting dual tobacco-product use. Am J Prev Med 23:226–230. Herity B, Moriarty M, Daly L, Dunn J, Bourke GJ. 1982. The role of tobacco and alcohol in the aetiology of lung and larynx cancer. Br J Cancer 46:961–964.

238

PART III: THE CAUSES OF CANCER

Hiatt RA, Tolan K, Quesenberry CP Jr. 1994. Renal cell carcinoma and thiazide use: a historical, case-control study (California, USA). Cancer Causes Control 5:319–325. Hickey K, Do K-A, Green A. 2001. Smoking and prostate cancer. Epidemiol Rev 23:115–125. Hirayama T. 1984. Cancer mortality in nonsmoking women with smoking husbands based on a large-scale cohort study in Japan. Prev Med 13:680–690. Hochberg F, Toniolo P, Cole P, Salcman M. 1990. Nonoccupational risk indicators of glioblastoma in adults. J Neurooncol 8:55–60. Hoffmann D, Hoffmann I. 1997. The changing cigarette, 1950–1995. J Toxicol Environ Health 50:307–364. Hoffmann D, Djordjevic MV, Fan J, Zang E, Glynn T, Connolly GN. 1995. Five leading U.S. commercial brands of moist snuff in 1994: assessment of carcinogenic N-nitrosamines. J Natl Cancer Inst 87:1862–1869. Hoffmann D, Djordjevic M, Hoffmann I. 1997. The changing cigarette. Prev Med 26:427–434. Hoffmann D, Hoffmann I, El Bayoumy K. 2001. The less harmful cigarette: a controversial issue: a tribute to Ernst L. Wynder. Chem Res Toxicol 14:767–790. Houlston R. 1999. Glutathione S-transferase M1 status and lung cancer risk: a meta-analysis. Cancer Epidemiol Biomarkers Prev 8:675–682. Hsing A, McLaughlin J, Chow W-H, et al. 1998. Risk factors for colorectal cancer in a prospective study among U.S. white men. Int J Cancer 77:549–553. Hsing A, McLaughlin J, Hrubec Z, Blot W, Fraumeni J. 1991. Tobacco use and prostate cancer: 26-year follow-up of US veterans. Am J Epidemiol 133:437–441. Hunter DJ, Colditz GA, Stampfer MJ, Rosner B, Willett WC, Speizer FE. 1990. Risk factors for basal cell carcinoma in a prospective cohort of women. Ann Epidemiol 1:13–23. Hurley SF, McNeil JJ, Donnan GA, Forbes A, Salzberg M, Giles GG. 1996. Tobacco smoking and alcohol consumption as risk factors for glioma: a case-control study in Melbourne, Australia. J Epidemiol Community Health 50:442–446. IARC. 1986. Tobacco Smoking. IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Vol. 38. Lyon: IARC. IARC.1987. Overall Evaluations of Carcinogenicity: An Updating of IARC Monographs. IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Vol. Suppl. 7. Lyon: IARC. IARC. 1988. Alcohol drinking. IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Vol. 44. Lyon: IARC. IARC. 1994. Hepatitis Viruses. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Vol. 59. Lyon: IARC. IARC. 1995. Human Papillomaviruses. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Vol. 64. Lyon: IARC. IARC. 2004. Tobacco Smoke and Involuntary Smoking. IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Vol. 83. Lyon: IARC. Inoue H, Kiyohara C, Marugame T, et al. 2000. Cigarette smoking, CYP1A1 MspI and GSTM1 genotypes, and colorectal adenomas. Cancer Res 60:3749–3752. Inoue M, Tajima K, Hirose K, Kuroishi T, Gao CM, Kitoh T. 1994. Life-style and subsite of gastric cancer—joint effect of smoking and drinking habits. Int J Cancer 56:494–499. Institute of Medicine. 2001. Clearing the smoke: assessing the science base for tobacco harm reduction. In: Stratton K, et al., editors. Washington, DC: National Academy Press. Iribarren C, Haselkorn T, Tekawa IS, Friedman GD. 2001. Cohort study of thyroid cancer in a San Francisco Bay area population. Int J Cancer 93:745–750. Iribarren C, Tekawa I, Sidney S, Friedman G. 1999. Effect of cigar smoking on the risk of cardiovascular disease, chronic obstructive pulmonary disease, and cancer in men. N Engl J Med 340:1773–1780. Jarvis M, Boreham R, Primatesta P, Feyerabend C, Bryant A. 2001. Nicotine yield from machine-smoked cigarettes and nicotine intakes in smokers: evidence from a representative population survey. J Natl Cancer Inst 93:134–138. Jarvis MA. 1987. Uptake of environmental tobacco smoke. In: O’Neill IK, Brunnemann KD, Dodet B, Hoffman D, editors. Passive Smoking, Vol. 9. Lyon: IARC, pp. 43–58. Jedrychowski W, Boeing H, Wahrendorf J, Popiela T, Tobiasz-Adamczyk B, Kulig J. 1993. Vodka consumption, tobacco smoking and risk of gastric cancer in Poland. Int J Epidemiol 22:606–613. Jedrychowski W, Popiela T, Drews M, et al. 1999. Effect of Helicobacter pylori infection, smoking and dietary habits on the occurrence of antrum intestinal metaplasia: clinico-epidemiological study in Poland. Pol J Pathol 50:289–295.

Jee SH, Ohrr H, Kim IS. 1999. Effects of husbands’ smoking on the incidence of lung cancer in Korean women. Int J Epidemiol 28:824–828. Jensen J, Christianson C, Rodbro P. 1985. Cigarette smoking, serum estrogens, and bone loss during hormone-replacement therapy early after menopause. N Engl J Med 313:973–975. Jensen O, Knudsen J, McLaughlin J, Sorensen B. 1988. The Copenhagen casecontrol study of renal pelvis and ureter cancer: role of smoking and occupational exposures. Int J Cancer 41:557–561. Jensen O, Wahrendorf J, Blettner M, Knudsen J, Sorensen B. 2000. The Copenhagen case-control study of bladder cancer: role of smoking in invasive and non-invasive bladder tumours. J Epidemiol Community Health 41:30–6. Jha P, Chaloupka F. 2000. The economics of global tobacco control. BMJ 321:358–361. Johnson KC, Hu J, Mao Y. 2000. Passive and active smoking and breast cancer risk in Canada, 1994–97; the Canadian Cancer Registries Epidemiology Research Group. Cancer Causes Control 11:211–221. Johnston L. 2001. Changing Demographic Patterns of Adolescent Smoking Over the Past 23 years: National Trends from the Monitoring the Future Study. Changing Adolescent Smoking Prevalence. Smoking and Tobacco Control Monograph No. 14. NIH Publ. No. 02–5086. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, pp. 9–34. Kabat GC, Ng SK, Wynder EL. 1993. Tobacco, alcohol intake, and diet in relation to adenocarcinoma of the esophagus and gastric cardia. Cancer Causes Control 4:123–132. Kahn H. 1966. The Dorn study of smoking and mortality among U.S. veterans: report on eight and one-half years of observation. In: Haenszel W, editor. Epidemiological Study of Cancer and Other Chronic Diseases. Bethesda, MD: U.S. Department of Health, Education, and Welfare, Public Health Service, National Cancer Institute, pp. 1–126. Karagas MR, Stukel TA, Greenberg ER, Baron JA, Mott LA, Stern RS. 1992. Risk of subsequent basal cell carcinoma and squamous cell carcinoma of the skin among patients with prior skin cancer; Skin Cancer Prevention Study Group. JAMA 267:3305–3310. Kato I, Tominaga S, Matsuura A, Yoshii Y, Shirai M, Kobayashi S. 1990. A comparative case-control study of colorectal cancer and adenoma. Jpn J Cancer Res 81:1101–1108. Key TJ, Pike MC, Baron JA, et al. 1991. Cigarette smoking and steroid hormones in women. J Steroid Biochem Mol Biol 39:529–534. Key TJ, Pike MC, Brown JB, Hermon C, Allen DS, Wang DY. 1996. Cigarette smoking and urinary oestrogen excretion in premenopausal and postmenopausal women. Br J Cancer 74:1313–1316. Khaw KT, Tazuke S, Barrett-Connor E. 1988. Cigarette smoking and levels of adrenal androgens in postmenopausal women. N Engl J Med 318:1705–1709. Kikendall J, Bowen P, Burgess M, et al. 1989. Cigarettes and alcohol as independent risk factors for colonic adenomas. Gastroenterology 97: 660–664. Kinlen L, Rogot E. 1988. Leukemia and smoking habits among United States veterans. BMJ 297:657–659. Kjellberg L, Hallmans G, Ahren AM, et al. 2000. Smoking, diet, pregnancy and oral contraceptive use as risk factors for cervical intra-epithelial neoplasia in relation to human papillomavirus infection. Br J Cancer 82:1332–1338. Knekt P, Hakama M, Jarvinen R, Pukkala E, Dich J, Hakulinen T. 1999. Risk of colorectal and other gastro-intestinal cancers after exposure to nitrate, nitrite and N-nitroso compounds: a follow-up study. Int J Cancer 80:852–856. Kneller RW, McLaughlin JK, Bjelke E, et al. 1991. A cohort study of stomach cancer in a high-risk American population. Cancer 68:672–678. Kneller RW, You WC, Chang YS, et al. 1992. Cigarette smoking and other risk factors for progression of precancerous stomach lesions. J Natl Cancer Inst 84:1261–1266. Kolonel LN, Hankin LH, Wilkens LR, Fukunaga FH, Hinds MW. 1990. An epidemiologic study of thyroid cancer in Hawaii. Cancer Causes Control 3:223–234. Kopstein A. 2001. Trends in adolescent smoking in the United States: data from the national household survey on drug abuse 1994 through 1998. Changing Adolescent Smoking Prevalence. Smoking and Tobacco Control Monograph No. 14. NIH Publ. No. 02–5086. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, pp. 51–68. Korte JE, Hertz-Picciotto I, Schulz MR, Ball LM, Duell E. 2000. The contribution of benzene to smoking-induced leukemia. Environ Health Perspectives 108:333–339. Kovar M, Poe G. 1985.The National Health Interview Survey Design, 1973–84, and Procedures, 1975–83. Vital and Health Statistics, Series 1,

Tobacco No. 18. Hyattsville, MD: US Department of Health and Human Services, Public Health Service, National Center for Health Statistics. Kozlowski L, O’Connor RJ, Sweeney CT. 2001. Cigarette design. In: Shopland D, Burns D, Benowitz N, Amacher R, editors. Risks Associated With Smoking Cigarettes With Low Machine-Measured Yields of Tar and Nicotine. Smoking and Tobacco Control Monograph 13. NIH Publ. No. 02–5074. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, pp. 13–38. Kreiger N, Parkes R. 2000. Cigarette smoking and the risk of thyroid cancer. Eur J Cancer 15:1969–1973. Kreiger N, Marrett LD, Dodds L, Hilditch S, Darlington GA. 1993. Risk factors for renal cell carcinoma: results of a population-based case-control study. Cancer Causes Control 4:101–110. Kreyberg L. 1962. Histological lung cancer types: a morphological and biological correlation. Acta Pathol Microbiol Immunol Scand Suppl 157:11–92. Kropp S, Chang-Claude J. 2002. Active and passive smoking and risk of breast cancer by age 50 years among German women. Am J Epidemiol 156:616–626. Kuper H, Tzonou A, Lagiou P, et al. Diet and hepatocellular carcinoma: a casecontrol study in Greece. Nutr Cancer 38:6–12. Lacey JV Jr, Frisch M, Brinton LA, et al.2001. Associations between smoking and adenocarcinomas and squamous cell carcinomas of the uterine cervix (United States). Cancer Causes Control 12:153–161. Lagergren J, Bergstrom R, Lindgren A, Nyren O. 2000. The role of tobacco, snuff and alcohol use in the aetiology of cancer of the oesophagus and gastric cardia. Int J Cancer 85:340–346. Lange P, Nyboe J, Appleyard M, Jensen G, Schnohr P. 1992. Relationship of the type of tobacco and inhalation pattern to pulmonary and total mortality. Eur Respir J 5:1111–1117. Lantz PM, House JS, Lepkowski JM, Williams DR, Mero RP, Chen J. 1998. Socioeconomic factors, health behaviors, and mortality. JAMA 279:1703–1708. Lash TL, Aschengrau A. 1999. Active and passive cigarette smoking and the occurrence of breast cancer. Am J Epidemiol 149:5–12. Lear JT, Tan BB, Smith AG, et al. 1998. A comparison of risk factors for malignant melanoma, squamous cell carcinoma and basal cell carcinoma in the UK. Int J Clin Pract 52:145–149. Lee M, Wrensch M, Miike R. 1997. Dietary and tobacco risk factors for adult onset glioma in the San Francisco Bay Area (California, USA). Cancer Causes Control 8:13–24. Lee PN. 2001. Relation between exposure to asbestos and smoking jointly and the risk of lung cancer. Occup Environ Med 58:145–153. Leffondre K, Abrahamowicz M, Siemiatycki J, Rachet B. 2002. Modeling smoking history: a comparison of different approaches. Am J Epidemiol 156:813–823. Lesko SM, Rosenberg L, Kaufman DW, et al. 1985. Cigarette smoking and the risk of endometrial cancer. N Engl J Med 313:593–596. Levi F, La Vecchia C. 2001. Tobacco smoking and prostate cancer: time for an appraisal. Ann Oncol 12:733–738. Levi F, La Vecchia C, Decarli A. 1987. Cigarette smoking and the risk of endometrial cancer. Eur J Cancer Clin Oncol 23:1025–1029. Liaw KM, Chen CJ. 1998. Mortality attributable to cigarette smoking in Taiwan: a 12-year follow-up study. Tob Control 7:141–148. Liddell FD. 2001. The interaction of asbestos and smoking in lung cancer. Ann Occup Hyg 45:341–356. Linet M, McLaughlin JK, Hsing AW, et al. 1992. Is cigarette smoking a risk factor for non-Hodgkin’s lymphoma or multiple myeloma? Results from the Lutheran Brotherhood Cohort Study. Leuk Res 16:621–624. Liu BQ, Peto R, Chen ZM, et al. 1998. Emerging tobacco hazards in China. 1. Retrospective proportional mortality study of one million deaths. BMJ 317:1411–1422. Lohmueller KE, Pearce CL, Pike M, Lander ES, Hirschorn JN. 2003. Meta-analysis of genetic association studies supports a contribution of common variants to susceptibility to common disease. Nat Genet 2:177–182. Longcope C, Johnston CCJ. 1988. Androgen and estrogen dynamics in preand postmenopausal women: a comparison between smokers and nonsmokers. J Clin Endocrinol Metab 67:379–383. Longnecker M, Chen M, Probst-Hensch N, et al. 1996. Alcohol and smoking in relation to the prevalence of adenomatous colorectal polyps at sigmoidoscopy. Epidemiology 7:275–280. Lopez A, Collishaw N, Piha T. 1994. A descriptive model of the cigarette epidemic in developed countries. Tob Control 3:242–247. Lopez-Abente G, Gonzalez C, Errezola M, et al. 1991. Tobacco smoke inhalation pattern, tobacco type, and bladder cancer in Spain. Am J Epidemiol 134:830–839.

239

Lopez-Abente G, Pollan M, Monge V, Martinez-Vidal A. 1992. Tobacco smoking, alcohol consumption, and laryngeal cancer in Madrid. Cancer Detect Prev 16:265–271. Lubin J, Blot W. 1984. Assessment of lung cancer risk factors by histologic category. J Natl Cancer Inst 73:383–389. Lubin J, Li J, Xuan X, et al. 1992. Risk of lung cancer among cigarette and pipe smokers in southern China. Int J Cancer 51:390–395. Mabuchi K, Bross DS, Kessler II. 1985. Epidemiology of cancer of the vulva: a case-control study. Cancer 55:1843–1848. Macdonald IG. 1957. Hearings before a subcommittee of the Committee on Government Operations, House of Representatives, 85th Congress, First Session, 85th Congress. Mackay J, Eriksen M. 2002. The Tobacco Atlas. Geneva: World Health Organization, p. 128. Maclure M, Willett W. 1990. A case-control study of diet and risk of renal adenocarcinoma. Epidemiology 1:430–440. MacMahon B, Cole P, Trichopoulos D, Brown J. 1982. Cigarette smoking and urinary estrogens. N Engl J Med 307:1062–1065. Maden C, Sherman KJ, Beckmann AM, et al. 1993. History of circumcision, medical conditions, and sexual activity and risk of penile cancer. J Natl Cancer Inst 85:19–24. Mannetje A, Kogevinas M, Luce D, et al. 1999. Sinonasal cancer, occupation, and tobacco smoking in European women and men. Am J Ind Med 36:101–107. Marcus PM, Hayes RB, Vineis P, et al. 2000a. Cigarette smoking, N-acetyltransferase 2 acetylation status, and bladder cancer risk: a case-series meta-analysis of a gene-environment interaction. Cancer Epidemiol Biomarkers Prev 9:461–467. Marcus PM, Newman B, Millikan RC, Moorman PG, Baird DD, Qaqish B. 2000b. The associations of adolescent cigarette smoking, alcoholic beverage consumption, environmental tobacco smoke, and ionizing radiation with subsequent breast cancer risk (United States). Cancer Causes Control 11:271–278. Marcus PM, Vineis P, Rothman N. 2000c. NAT2 slow acetylation and bladder cancer risk: a meta-analysis of 22 case-control studies conducted in the general population. Pharmacogenetics 10:115–122. Martin G, Brown J, Eifler C, Houston G. 1999. Oral leukoplakia status six weeks after cessation of smokeless tobacco use. J Am Dent Assoc 130:945–954. Martinez ME, McPherson RS, Annegers JF, Levin B. 1995. Cigarette smoking and alcohol consumption as risk factors for colorectal adenomatous polyps. J Natl Cancer Inst 87:274–279. Maxwell JC. 2002. The Maxwell Consumer Report: 2001 International Tobacco Report. Parts One and Two. Richmond, VA: Davenport. McCredie M, Stewart JH. 1992. Risk factors for kidney cancer in New South Wales. I. Cigarette smoking. Eur J Cancer 12:2050–2054. McCredie M, Stewart JH, Ford JM. 1983. Analgesics and tobacco as risk factors for cancer of the ureter and renal pelvis. J Urol 130:28–30. McGinnis J, Foege W. 1993. Actual causes of death in the United States. JAMA 1993; 270:2207–2212. McLaughlin JK, Gao YT, Gao RN, et al. 1992a. Risk factors for renal-cell cancer in Shanghai, China. Int J Cancer 52:562–565. McLaughlin JK, Silverman DT, Hsing AW, et al. 1992b. Cigarette smoking and cancers of the renal pelvis and ureter. Cancer Res 52:254–257. McTiernan AM, Weiss NS, Daling JR. 1984. Incidence of thyroid cancer in women in relation to reproductive and hormonal factors. Am J Epidemiol 120:423–435. Melikian AA, Sun P, Prokopczyk B, et al. 1999. Identification of benzo[a]pyrene metabolites in cervical mucus and DNA adducts in cervical tissues in humans by gas chromatography-mass spectrometry. Cancer Lett 146:127–134. Mellemgaard A, Engholm G, McLaughlin JK, Olsen JH. 1994, Risk factors for renal cell carcinoma in Denmark. I. Role of socioeconomic status, tobacco use, beverages, and family history. Cancer Causes Control 5:105–113. Memon A, Darif M, Al-Saleh K, Suresh A. 2002. Epidemiology of reproductive and hormonal factors in thyroid cancer: evidence from a case-control study in the Middle East. Int J Cancer 97:82–89. Michnovicz JJ, Hershcopf RJ, Naganuma H, Bradlow HL, Fishman J. 1986. Increased 2-hydroxylation of estradiol as a possible mechanism for the anti-estrogenic effect of cigarette smoking. N Engl J Med 315:1305– 1309. Mills PK, Newell GR, Beeson WL, Fraser GE, Philips RL. 1990. History of cigarette smoking and risk of leukemia and myeloma: results from the Adventist Health Study. J Natl Cancer Inst 82:1832–1836. Minna JD. 2003. Nicotine exposure and bronchial epithelial cell nicotinic acetylcholine receptor expression in the pathogenesis of lung cancer. J Clin Invest 111:31–33.

240

PART III: THE CAUSES OF CANCER

Moerman CJ, Bueno de Mesquita HB, Runia S. 1994. Smoking, alcohol consumption and the risk of cancer of the biliary tract; a population-based case-control study in The Netherlands. Eur J Cancer Prev 5:427–436. Momas I, Daures J, Festy B, Bontoux J, Gremy F. 1994. Bladder cancer and black tobacco cigarette smoking: some results from a French case-control study. Eur J Epidemiol 10:599–604. Monnet E, Allemand H, Farina H, Carayon P. 1991. Cigarette smoking and the risk of colorectal adenoma in men. Scand J Gastroenterol 26:758–762. Moore TO, Moore AY, Carrasco D, et al. 2001. Human papillomavirus, smoking, and cancer. J Cutan Med Surg 5:323–328. Morabia A, Bernstein M, Heritier S, et al. 1996. Relation of breast cancer with passive and active exposure to tobacco smoke. Am J Epidemiol 143:918–928. Muscat JE, Richie JP Jr, Thompson S, Wynder EL. 1996. Gender differences in smoking and risk for oral cancer. Cancer Res 56:5192–5197. Muscat JE, Stellman SD, Hoffmann D, Wynder EL. 1997. Smoking and pancreatic cancer in men and women. Cancer Epidemiol Biomarkers Prev 6:15–19. Nagata C, Shimizu H, Kametani M, Takeyema N, Ohnuma T, Matsushita S. 1999. Cigarette smoking, alcohol use, and colorectal adenoma in Japanese men and women. Dis Colon Rectum 42:337–342. Nam JM, McLaughlin JK, Blot WJ. 1992. Cigarette smoking, alcohol, and nasopharyngeal carcinoma: a case-control study among U.S. whites. J Natl Cancer Inst 84:619–622. National Center for Health Statistics. 2002. Health, United States, 2002: With Chartbook on Trends in the Health of Americans. DHHS Publ. No. 1232. Hyattsville, MD: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics. National Research Council. 1986. Environmental Tobacco Smoke: Measuring Exposures and Assessing Health Effects. Washington, DC: National Academy Press. National Research Council Committee on Health Risks of Exposure to Radon (BIER VI). 1998. Health Effects of Exposure to Radon. Biological Effects of Ionizing Radiation. Washington, DC: National Academy Press. Newcomb PA, Weiss NS, Daling JR. 1984. Incidence of vulvar carcinoma in relation to menstrual, reproductive, and medical factors. J Natl Cancer Inst 73:391–396. Ng TP. 1986. A case-referent study of cancer of the nasal cavity and sinuses in Hong Kong. Int J Epidemiol 15:171–175. Nilsson R. 1998. A qualitative and quantitative risk assessment of snuff dipping. Regul Toxicol Pharmacol 28:1–16. Nishino Y, Tsubono Y, Tsuji I, et al. 2001. Passive smoking at home and cancer risk: a population-based prospective study in Japanese nonsmoking women. Cancer Causes Control 12:797–802. Nyren O, Bergstrom R, Nystrom L, et al. 1996. Smoking and colorectal cancer: a 20-year follow-up study of Swedish construction workers. J Natl Cancer Inst 88:1302–1307. O’Brien CP. 2001. Drug addiction and drug abuse. In: Hardman JG, Limbird LE, Goodman Gilman A, editors. Goodman & Gilman’s The Pharmacological Basics of Therapeutics. New York: McGraw-Hill, pp. 621–642. Olsen J, Kronborg O. 1993. Coffee, tobacco and alcohol as risk factors for cancer and adenoma of the large intestine. Int J Epidemiol 22:398–402. Osterlind A, Tucker MA, Stone BJ, Jensen OM. 1988. The Danish case-control study of cutaneous malignant melanoma. IV. No association with nutritional factors, alcohol, smoking or hair dyes. Int J Cancer 42:825–828. Paganini-Hill A, Hsu G. 1994. Smoking and mortality among residents of a California retirement community. Am J Public Health 84:992–995. Palli D, Bianchi S, Decarli A, et al. 1992. A case-control study of cancers of the gastric cardia in Italy. Br J Cancer 65:263–266. Palmer JR, Rosenberg L. 1993. Cigarette smoking and the risk of breast cancer. Epidemiol Rev 15:145–156. Parkin D, Pisani P, Ferlay J. 1999. Estimates of the worldwide incidence of 25 major cancers in 1990. Int J Cancer 80:827–841. Partanen TJ, Vainio HU, Ojajarvi IA, Kauppinen TP. 1997. Pancreas cancer, tobacco smoking and consumption of alcoholic beverages: a case-control study. Cancer Lett 116:27–32. Patrick DL, Cheadle A, Thompson DC, Diehr P, Koepsell T, Kinne S. 1994. The validity of self-reported smoking: a review and meta-analysis. Am J Public Health 84:1086–1093. Peele DM, Riddick MG, Edwards ME. 2001. Formation of tobacco-specific nitrosamines in flue-cured tobacco. Recent Adv Tob Sci 27:3–12. Peto R, Lopez A. 2001. Future worldwide health effects of current smoking patterns. In: Koop CE, Pearson CE, Schwartz MR, editors. Critical Issues in Global Health. San Francisco: Jossey-Bass. Peto R, Darby S, Deo H, Silcocks P, Whitley E, Doll R. 2000. Smoking, smoking cessation and lung cancer in the UK since 1950: combination of national statistics with two case-control studies. BMJ 321:323–329.

Peto R, Lopez A, Boreham J, Thun M, Heath CJ. 1994. Mortality from smoking in developed countries 1950–2000. New York: Oxford University Press. Pierce JP, Lee L, Gilpin EA. 1994. Smoking initiation by adolescent girls, 1944 through 1988: an association with targeted advertising. JAMA 271: 608–611. Pitard A, Brennan P, Clavel J, et al. 2001. Cigar, pipe, and cigarette smoking and bladder cancer risk in European men. Cancer Causes Control 12:551–556. Plummer M, Herrero R, Bosch X, et al. 2001. Smoking and cervical cancer: pooled analysis of a multicentric case-control study. Presented at the 19th International Papillomavirus Conference, Florianopolis, Brazil, Abstract 0–193. Potter JD, Bigler J, Fosdick L, et al. 1999. Colorectal adenomatous and hyperplastic polyps: smoking and N-acetyltransferase 2 polymorphisms. Cancer Epidemiol Biomarkers Prev 8:69–75. Prokopczyk B, Cox JE, Hoffmann D, Waggoner SE. 1997. Identification of tobacco-specific carcinogen in the cervical mucous of smokers and nonsmokers. J Natl Cancer Inst 89:868–873. Qiao Y, Taylor P, Yao S, et al. 1989. Relation of radon exposure and tobacco use to lung cancer among tin miners in Yunnan Province, China. Am J Ind Med 16:511–521. Roberts D. 1988. Natural tobacco flavor. Rec Adv Tob Sci 14:49–81. Rodriguez C, Tatham L, Thun M, Calle E, Heath CJ. 1997. Smoking and fatal prostate cancer in a large cohort of adult men. Am J Epidemiol 145:466–475. Roffo A. 1930. Leucoplasie, expermentale produite par le tabac. Rev Sud Am Chir 1:321–330. Rogot E, Murray JL. 1980. Smoking and causes of death among U.S. veterans: 16 years of observation. Public Health Rep 95:213–222. Ross RK, Paganini-Hill A, Landolph J, Gerkins V, Henderson BE. 1989. Analgesics, cigarette smoking, and other risk factors for cancer of the renal pelvis and ureter. Cancer Res 49:1045–1048. Rossing MA, Cushing KL, Voigt LF, Wicklund KG, Daling JR. 2000. Risk of papillary thyroid cancer in women in relation to smoking and alcohol consumption. Epidemiology 11:49–54. Ryan P, Lee MW, North B, McMichael AJ. 1992. Risk factors for tumors of the brain and meninges: results from the Adelaide Adult Brain Tumor Study. Int J Cancer 51:20–27. Sahl WJ, Glore S, Garrison P, Oakleaf K, Johnson SD. 1995. Basal cell carcinoma and lifestyle characteristics. Int J Dermatol 34:398–402. Sala E, Warren R, McCann J, Duffy S, Luben R, Day N. 2000. Smoking and high-risk mammographic parenchymal patterns: a case-control study. Breast Cancer Res 2:59–63. Sancho-Garnier H, Theobald S. 1993. Black (air-cured) and blond (flue-cured) tobacco and cancer risk. II. Pharynx and larynx cancer. Eur J Cancer 2:273–276. Sandler DP, Everson RB, Wilcox AJ. 1985a. Passive smoking in adulthood and cancer risk. Am J Epidemiol 121:37–47. Sandler DP, Wilcox AJ, Everson RB. 1985b. Cumulative effects of lifetime passive smoking on cancer risk. Lancet 1:312–315. Sankaranarayanan R, Duffy SW, Nair MK, Padmakumary G, Day NE. 1990. Tobacco and alcohol as risk factors in cancer of the larynx in Kerala, India. Int J Cancer 45:879–882. Scherer G, Richter E. 1997. Biomonitoring exposure to environmental tobacco smoke (ETS): a critical reappraisal. Hum Exp Toxicol 16:449– 459. Schlecht NF, Franco EL, Pintos J, et al. 1999. Interaction between tobacco and alcohol consumption and the risk of cancers of the upper aero-digestive tract in Brazil. Am J Epidemiol 150:1129–1137. Schlemmer A, Jensen J, Riis BJ, Christiansen C. 1990. Smoking induces increased androgen levels in early post-menopausal women. Maturitas 12:99–104. Scott TE, Carroll M, Cogliano FD, Smith BF, Lamorte WW. 1999. A casecontrol assessment of risk factors for gallbladder carcinoma. Dig Dis Sci 44:1619–1625. Serraino D, Franceschi S, Talamini R, Frustaci S, La Vecchia C. 1991. Nonoccupational risk factors for adult soft-tissue sarcoma in northern Italy. Cancer Causes Control 2:157–164. Shafey O, Dolwick S, Guindon GE. 2003. Tobacco Control: Country Profiles. Atlanta: American Cancer Society. Shanks T, Burns D. 1998. Disease consequences of cigar smoking. In: Shopland D, Burns D, Hoffmann D, Cummings K, Amacher R, editors. Cigars: Health Effects and Trends. Smoking and Tobacco Control Monograph 9. NIH Publ. No. 98–4302. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, pp. 105–160. Shapiro J, Jacobs E, Thun M. 2000. Cigar smoking in men and risk of death from tobacco-related cancers. J Natl Cancer Inst 92:333–337.

Tobacco Shields PG. 2002. Tobacco smoking, harm reduction, and biomarkers. J Natl Cancer Inst 94:1435–1444. Siman JH, Forsgren A, Berglund G, Floren CH. 2001. Tobacco smoking increases the risk for gastric adenocarcinoma among Helicobacter pyloriinfected individuals. Scand J Gastroenterol 36:208–213. Slade J. 1989. The tobacco epidemic: lessons from history. J Psychoactive Drugs 21:281–291. Slade J. 1993. Nicotine delivery devices. In: Orleans C, Slade J, editors. Nicotine Addiction, Principles and Management. New York: Oxford University Press, pp. 3–23. Smith SJ, Deacon JM, Chilvers CED. 1994. Alcohol, smoking, passive smoking and caffeine in relation to breast cancer risk in young women; UK National Case-Control Study Group. Br J Cancer 70:112–119. Sokic SI, Adanja BJ, Marinkovic JP, Vlajinac HD. 1994. Case-control study of risk factors in laryngeal cancer. Neoplasma 41:43–47. Steineck G, Norell S, Feychting M. 1988. Diet, tobacco and urothelial cancer: a 14-year follow-up of 16,477 subjects. Acta Oncol 27:323–327. Stockwell HG, Lyman GH. 1987. Cigarette smoking and the risk of female reproductive cancer. Am J Obstet Gynecol 157:35–40. Sturmer T, Glynn R, Lee I, Christen W, Hennekens C. 2000. Lifetime cigarette smoking and colorectal cancer incidence in the Physicians’ Health Study. J Natl Cancer Inst 92:1178–1181. Sun CA, Wu DM, Lin CC, et al. 2003. Incidence and cofactors of hepatitis C virus-related hepatocellular carcinoma: a prospective study of 12,008 men in Taiwan. Am J Epidemiol 157:674–682. Talamini G, Capelli P, Zamboni G, et al. 2000. Alcohol, smoking and papillomavirus infection as risk factors for esophageal squamous-cell papilloma and esophageal squamous-cell carcinoma in Italy. Int J Cancer 86:874–878. Talamini R, Baron AE, Barra S, et al. 1990. A case-control study of risk factor for renal cell cancer in northern Italy. Cancer Causes Control 1:125–131. Talaska G, Schamer M, Skipper P, et al. 1991. Detection of carcinogen-DNA adducts in exfoliated urothelial cells of cigarette smokers: association with smoking, hemoglobin adducts, and urinary mutagenicity. Cancer Epidemiol Biomarkers Prev 1:61–66. Tanaka K, Hirohata T, Fukuda K, Shibata A, Tsukuma H, Hiyama T. 1995. Risk factors for hepatocellular carcinoma among Japanese women. Cancer Causes Control 6:91–98. TATA Institute of Fundamental Research, WHO, and CDC. 2000. Proceedings of the International Scientific Expert Meeting on the Possible Causality Between Smoking and Tuberculosis: Evidence for a Causal Link Between Smoking and Tuberculosis, Thiruvanathapuram, Kerala, India. Terry P, Rohan T. 2002. Cigarette smoking and breast cancer risk: a review of the literature. Cancer Epidemiol Biomarkers Prev 11:953–971. Thompson PA, DeMarini DM, Kadlubar FF, et al. 2002. Evidence for the presence of mutagenic arylamines in human breast milk and DNA adducts in exfoliated breast ductal epithelial cells. Environ Mol Mutagen 39:134–142. Thun M, Burns D. 2001. Health impact of “reduced yield” cigarettes: a critical assessment of the epidemiological evidence. Tob Control 10(Suppl I):i4–i11. Thun M, Day-Lally C, Meyers D, et al. 1997a. Trends in tobacco smoking and mortality from cigarette use in Cancer Prevention Studies I (1959 through 1965) and II (1982 through 1988). In: Shopland D, Burns D, Garfinkel L, Samet J, editors. Changes in Cigarette-Related Disease Risks and Their Implication for Prevention and Control. Smoking and Tobacco Control Monograph 8. NIH Publ. No. 97–4213. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute, pp. 305–382. Thun M, Henley S, Calle E. 2002. Tobacco use and cancer: an epidemiologic perspective for geneticists. Oncogene 21:7307–7325. Thun M, Lally C, Flannery J, Calle E, Flanders W, Heath C. 1997b. Cigarette smoking and changes in the histopathology of lung cancer. J Natl Cancer Inst 89:1580–1586. Todd GF. 1978. Cigarette consumption per adult of each sex in various countries. J Epidemiol Community Health 32:289–293. Tredaniel J, Boffetta P, Buiatti E, Saracci R, Hirsch A. 1997. Tobacco smoking and gastric cancer: review and meta-analysis. Int J Cancer 72:565–573. Tsen HF, Morgenstern H, Mack T, Peters RK. 2001. Risk factors for penile cancer: results of a population-based case-control study in Los Angeles County (United States). Cancer Causes Control 12:267–277. Tsukuma H, Hiyama T, Tanaka S, et al. 1993. Risk factors for hepatocellular carcinoma among patients with chronic liver disease. N Engl J Med 328:1797–1801. Tuyns AJ, Esteve J, Raymond L, et al. 1988. Cancer of the larynx/hypopharynx, tobacco and alcohol: IARC international case-control study in Turin and Varese (Italy), Zaragoza and Navarra (Spain), Geneva (Switzerland) and Calvados (France). Int J Cancer 41:483–491.

241

Tverdal A, Thelle D, Stensvold I, Leren P, Bjartveit K. 1993. Mortality in relation to smoking history: 13 years’ follow-up of 68,000 Norwegian men and women 35–49 years. J Clin Epidemiol 46:475–487. U.S. Department of Agriculture. 2002. Tobacco Situation and Outlook Report. Washington, DC: US Department of Agriculture, Market and Trade Economics Division, Economics Research Service. U.S. Department of Health, Education, and Welfare. 1979. Smoking and Health. A Report of the Surgeon General. DHEW Publ. No. (PHS) 79–50066. U.S. Department of Health, Education, and Welfare, Public Health Service, Office of the Assistant Secretary for Health, Office on Smoking and Health. U.S. Department of Health and Human Services. 1981. The Health Consequences of Smoking: The Changing Cigarette. A Report of the Surgeon General. Rockville, MD: U.S. Department of Health and Human Services, Public Health Service, Office on Smoking and Health. U.S. Department of Health and Human Services. 1982. The Health Consequences of Smoking: Cancer. A Report of the Surgeon General. Rockville, MD: U.S. Department of Health and Human Services, Public Health Service, Office on Smoking and Health. U.S. Department of Health and Human Services. 1986. The Health Consequences of Involuntary Smoking: A Report of the Surgeon General. Washington, DC: Department of Health and Human Services. U.S. Department of Health and Human Services. 1988. Nicotine Addiction: a Report of the Surgeon General. DHHS Publication No. (CDC) 88–8406, Rockville, MD: Department of Health and Human Services. U.S. Department of Health and Human Services. 1989. Reducing the Health Consequences of Smoking: 25 Years of Progress. A Report of the Surgeon General. Rockville, MD: U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control, Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. U.S. Department of Health and Human Services. 1990. The Health Benefits of Smoking Cessation: A Report of the Surgeon General. Rockville, MD: U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control, Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. U.S. Department of Health and Human Services. 1994. Preventing Tobacco Use Among Young People: A Report of the Surgeon General. Atlanta: U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. U.S. Department of Health and Human Services. 2000. Reducing Tobacco Use: A Report of the Surgeon General. Atlanta: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. U.S. Department of Health and Human Services. 2001. Women and Smoking: A Report of the Surgeon General. Atlanta: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. U.S. Department of Health and Human Services. 2002. 10th Report on Carcinogens. Washington, DC: U.S. Department of Health and Human Services, Public Health Service, National Toxicology Program. U.S. Department of Health and Human Services. 2004. The Health Consequences of Tobacco Use: A Report of the Surgeon General. Atlanta: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health. U.S. Department of the Treasury, Bureau of Alcohol, Tobacco and Firearms. 1996. Tobacco: Subpart B. Definitions of cigars: Federal Register, pp. 141–143. U.S. Environmental Protection Agency. 1992. Respiratory Health Effects of Passive Smoking: Lung Cancer and Other Disorders. Washington, DC: U.S. Environmental Protection Agency. U.S. Public Health Service. Smoking and Health. 1964. Report of the Advisory Committee to the Surgeon General of the Public Health Service. Washington, DC: U.S. Department of Health, Education, and Welfare, Public Health Service, Center for Disease Control. Vachon CM, Kuni CC, Anderson K, Anderson VE, Sellers TA. 2000. Association of mammographically defined percent breast density with epidemiologic risk factors for breast cancer (United States). Cancer Causes Control 11:653–662. Van der Waal I, Schepman KP, van der Meij EH, Smeele LE. 1997. Oral leukoplakia: a clinicopathological review. Oral Oncol 33:291–301. Vaughan TL, Davis S, Kristal A, Thomas DB. 1995. Obesity, alcohol, and tobacco as risk factors for cancers of the esophagus and gastric cardia:

242

PART III: THE CAUSES OF CANCER

adenocarcinoma versus squamous cell carcinoma. Cancer Epidemiol Biomarkers Prev 4:85–92. Vaughan TL, Shapiro JA, Burt RD, et al. 1996. Nasopharyngeal cancer in a low-risk population: defining risk factors by histological type. Cancer Epidemiol Biomarkers Prev 5:587–593. Vermeulen R, Wegh H, Bos RP, Kromhout H. 2000. Weekly patterns in smoking habits and influence on urinary cotinine and mutagenicity levels: confounding effects of nonsmoking policies in the workplace. Cancer Epidemiol Biomarkers Prev 9:1205–1209. Vineis P. 1991. Black (air-cured) and blond (flue-cured) tobacco and cancer risk. I. Bladder cancer. Eur J Cancer 27:1491–1493. Vineis P, Esteve J, Terracini B. 1984. Bladder cancer and smoking in males: types of cigarettes, age at start, effect of stopping and interaction with occupation. Int J Cancer 34:165–170. Vineis P, Talaska G, Malaveille C, et al. 1996. DNA adducts in urothelial cells: relationship with biomarkers of exposure to arylamines and polycyclic aromatic hydrocarbons from tobacco smoke. Int J Cancer 65: 314–316. Vutuc C, Kunze M. 1982. Lung cancer risk in women in relation to tar yields of cigarettes. Prev Med 11:713–716. Vutuc C, Kunze M. 1983. Tar yields of cigarettes and male lung cancer risk. J Natl Cancer Inst 71:435–437. Wald N, Watt H. 1997. Prospective study of effect of switching from cigarettes to pipes or cigars on mortality from three smoking related diseases. BMJ 314:1860–1863. Wartenberg D, Calle EE, Thun MJ, Heath CW Jr, Lally C, Woodruff T. 2000. Passive smoking exposure and female breast cancer mortality. J Natl Cancer Inst 92:1666–1673. Weiderpass E, Baron JA. 2001. Cigarette smoking, alcohol consumption, and endometrial cancer risk: a population-based study in Sweden. Cancer Causes Control 12:239–247. Weinberg C, Sandler D. 1999. Gene-by-environment interaction for passive smoking and glutathione S-transferase M1? J Natl Cancer Inst 91:1985–1986. West KA, Brognard J, Clark AS, et al. 2003. Rapid Akt activation by nicotine and a tobacco carcinogen modulates the phenotype of normal human airway epithelial cells. J Clin Invest 111:81–90. Westerdahl J, Olsson H, Masback A, Ingvar C, Jonsson N. 1996. Risk of malignant melanoma in relation to drug intake, alcohol, smoking and hormonal factors. Br J Cancer 73:1126–1131. Westra W, Sidransky D. 1998. Phenotypic and genotypic disparity in premalignant lesions: of calm waters and crocodiles. J Natl Cancer Inst 90:1500–1501. Whyatt RM, Garte SJ, Cosma G, et al. 1995. CYP1A1 messenger RNA levels in placental tissue as a biomarker of environmental exposure. Cancer Epidemiol Biomarkers Prev 4:147–153. Wiencke JK, Thurston SW, Kelsey KT, et al. 1999. Early age at smoking initiation and tobacco carcinogen DNA damage to the lung. J Natl Cancer Inst 91:614–619. Wilcox HB, Schoenberg JB, Mason TJ, Bill JS, Stemhagen A. 1988. Smoking and lung cancer: risk as a function of cigarette tar content. Prev Med 17:263–272. Williams RR, Horm JW. 1977. Association of cancer sites with tobacco and alcohol consumption and socioeconomic status of patients: interview study from the Third National Cancer Survey. J Natl Cancer Inst 58:525–547. Wingo P, Cardinez C, Landis S, et al. 2003. Long-term trends in cancer mortality in the United States, 1930–1998. Cancer 97(Suppl 12):3133– 3275. Woodward M, Moohan M, Tunstall-Pedoe H. 1999. Self-reported smoking, cigarette yields and inhalation biochemistry related to the incidence of

coronary heart disease: results from the Scottish Heart Health Study. J Epidemiol Biostat 4:285–295. World Health Organization. Tobacco or Health: A Global Status Report. Geneva: World Health Organization. Wu AH, Crabtree JE, Bernstein L, et al. 2003. Role of Helicobacter pylori CagA+ strains and risk of adenocarcinoma of the stomach and esophagus. Int J Cancer 103:815–821. Wynder E, Hoffmann D. 1994. Smoking and lung cancer: scientific challenges and opportunities. Cancer 54:5284–5295. Yang HI, Lu SN, Liaw YF, et al. 2002. Hepatitis B e antigen and the risk of hepatocellular carcinoma. N Engl J Med 347:168–174. Ye W, Ekstrom AM, Hansson LE, Bergstrom R, Nyren O. 1999. Tobacco, alcohol and the risk of gastric cancer by sub-site and histologic type. Int J Cancer 83:223–229. Yen S, Hsieh CC, MacMahon B. 1987. Extrahepatic bile duct cancer and smoking, beverage consumption, past medical history, and oral-contraceptive use. Cancer 59:2112–2116. Ylitalo N, Sorensen P, Josefsson A, et al. 1999. Smoking and oral contraceptives as risk factors for cervical carcinoma in situ. Int J Cancer 81:357–365. You WC, Li JY, Blot WJ, et al. 1999. Evolution of precancerous lesions in a rural Chinese population at high risk of gastric cancer. Int J Cancer 83:615–619. Yu M, Tong M, Govindarajan S, Henderson B. 1991. Nonviral risk factors for hepatocellular carcinoma in a low-risk population, the non-Asians of Los Angeles County, California. J Natl Cancer Inst 83:1820–1826. Yuan JM, Wang XL, Xiang YB, Gao YT, Ross RK, Yu MC. 2000. Non-dietary risk factors for nasopharyngeal carcinoma in Shanghai, China. Int J Cancer 85:364–369. Zahm S, Cocco P, Blair A. 1991. Tobacco smoking as a risk for colon polyps. Am J Public Health 81:846–849. Zahm S, Weisenburger D, Holmes F, Cantor K, Blair A. 1997. Tobacco and non-Hodgkin’s lymphoma: combined analysis of three case-control studies (United States). Cancer Causes Control 8:159–166. Zahm SH, Heineman EF, Vaught JB. 1992. Soft tissue sarcoma and tobacco use: data from a prospective cohort study of United States veterans. Cancer Causes Control 3:371–376. Zaridze D, Borisova E, Maximovitch D, Chkhikvadze V. 2000. Alcohol consumption, smoking and risk of gastric cancer: case-control study from Moscow, Russia. Cancer Causes Control 11:363–371. Zatonski W, Becher H, Lissowska J, Wahrendorf J. 1991. Tobacco, alcohol, and diet in the etiology of laryngeal cancer: a population-based case-control study. Cancer Causes Control 2:3–10. Zatonski WA, La Vecchia C, Przewozniak K, Maisonneuve P, Lowenfels AB, Boyle P. 1992. Risk factors for gallbladder cancer: a Polish case-control study. Int J Cancer 51:707–711. Zheng T, Cantor KP, Zhang Y, Chiu BC, Lynch CF. 2001. Risk of brain glioma not associated with cigarette smoking or use of other tobacco products in Iowa. Cancer Epidemiol Biomarkers Prev 10:413–414. Zheng W, Blot WJ, Shu XO, et al. 1992a. A population-based case-control study of cancers of the nasal cavity and paranasal sinuses in Shanghai. Int J Cancer 52:557–561. Zheng W, Blot WJ, Shu XO, et al. 1992b. Diet and other risk factors for laryngeal cancer in Shanghai, China. Am J Epidemiol 136:178–191. Zheng W, McLaughlin JK, Chow WH, Chien HT, Blot WJ. 1993. Risk factors for cancers of the nasal cavity and paranasal sinuses among white men in the United States. Am J Epidemiol 138:965–972. Zhu K, Levine RS, Brann EA, Gnepp DR, Baum MK. 1995. A populationbased case-control study of the relationship between cigarette smoking and nasopharyngeal cancer (United States). Cancer Causes Control 6:507–512.

14

Alcohol JAMES R. MARSHALL AND JO FREUDENHEIM

A

lcohol has been part of the warp and woof of most human societies for thousands of years. People enjoy consuming it: they enjoy its narcotic effects and use it as a social lubricant. Nonetheless, its proscription in many cultures is testimony to the significant costs of alcohol use, the most glaring of which are the consequences of alcohol abuse, including social disruption, addiction, and physical debilitation. There is little doubt that alcohol is a scourge for those who become addicted. In addition, excessive intake over the long term causes severe damage to the liver, kidney, and circulatory system. It even causes cancer at several sites. Research is making new advances in identifying the causal pathways by which alcohol may have an impact. In addition, more is being learned about the molecular factors that facilitate or hinder the impact of alcohol on cancer risk. Identifying the linkages of alcohol to cancer risk poses severe epidemiologic challenges. For alcohol intake to affect cancer risk appears to require decades of elevated exposure. Alcohol exposure is generally difficult to measure, and it is correlated with a number of other practices, especially cigarette smoking, that increase cancer risk. Thus, identifying the specific etiologic importance of alcohol is difficult. Because of the social isolation and disruption associated with alcohol abuse, it is doubly difficult to obtain data from those who most abuse alcohol. This chapter briefly summarizes the present understanding of the mechanisms by which alcohol might affect cancer risk and then proceeds to evaluate the molecular genetic factors that appear relevant to alcohol metabolism and hence the impact of alcohol on cancer risk. Brief reviews of the means by which alcohol’s effects can be studied and their limitations follow. The role of alcohol in cancer at major cancer sites is then used to gauge the likely importance of alcohol to cancer risk and prevention. These sites are either ones for which there is a substantial literature linking alcohol to risk, or they are associated with significant morbidity and mortality.

ALCOHOL IN CANCER ETIOLOGY: POSSIBLE EXPOSURE MECHANISMS Epidemiologic associations have been found between alcohol consumption and risk of cancer at various sites. A number of possible mechanisms may explain this putative relation, including the effects of alcohol on carcinogen metabolism, effects of acetaldehyde, interactions of alcohol with nutritional factors, effects of alcohol on hormone levels, and physical effects of alcohol on tissues. There is evidence that links all kinds of alcoholic beverages to such risks. We focus here on the effects of ethanol consumption on cancer etiology. There may be additional effects related to the compounds found in alcoholic beverages. Ethanol may affect the activation of carcinogens because of its effects on the induction of several of the p450 cytochromes. There is evidence from an animal model that chronic alcohol exposure can lead to induction of the cytochrome p450s 2E1, 1A1, 2B1, and 3A1 (Roberts et al., 1995). It may also inhibit phase II enzymes, affecting the clearance of carcinogens (Singletary and Gapstur, 2001). Acetaldehyde has been identified as a carcinogen by the International Agency for Research on Cancer (IARC). There is evidence from animal and cell models that it has mutagenic effects and can affect the

cell cycle, apoptosis, and DNA repair (Seitz et al., 1998). Exposure to this alcohol metabolite may explain part of the observed associations between alcohol consumption and cancer risk. Although most alcohol metabolism occurs in the liver, there is alcohol dehydrogenase activity in a number of other tissues (Saleem et al., 1984) with the possibility of exposure to acetaldehyde. Microbial metabolism of alcohol to acetaldehyde may play a role in the carcinogenesis of the upper gastrointestinal tract (Seitz et al., 1998). Acetaldehyde–DNA adducts may form as a result of alcohol exposure. Furthermore, there may be increased levels of malondialdehyde and 4-hydroxynonenal, lipid peroxidation products that may be produced as a result of acetaldehyde metabolism. These substances can bind DNA and affect gene transcription; and they may, with acetaldehyde, bind to proteins to affect cell functioning (Eriksson, 2001). Some of the observed associations of alcohol and cancer may be caused by alcohol-related effects on the status of nutrients of significance to cancer etiology. For example, alcohol consumption has a negative impact on the absorption, utilization, and excretion of folate (Herbert and Kshitish, 1994). Folate may be related to risk of cancer of the large bowel (Freudenheim et al., 1991; Giovannucci et al., 1993); and in relation to breast cancer risk, women with low folate levels may be particularly vulnerable to the effects of alcohol (Zhang et al., 1999; Sellers et al., 2002). Alcohol consumption may also affect nutritional status for other nutrients that have been suggested to be related to cancer risk. Compared with nondrinkers, drinkers may have lower concentrations of some carotenoids and vitamin C (Lieber, 2000; Singletary and Gapstur, 2001). Alcohol may interfere with vitamin A absorption, increase its degradation (Seitz et al., 1998), and interfere with the synthesis of retinoic acid from retinal (Lieber, 2000; Agarwal, 2001). There may be effects on vitamin D (Lieber, 2000). The metabolism of alcohol by CYP2E1 leads to increased production of reactive oxygen species, which in turn may exacerbate the effects of lower levels of antioxidant vitamins. Furthermore, acetaldehyde metabolism enzymes, molybdenum hydroxylase (XOR) and aldehyde oxidase (AOR) result in the production of reactive oxygen species (Wright et al., 1999). Alcohol can have both acute and chronic effects on blood hormone concentrations. Increased concentrations of estrogens and urinary metabolites of estrogen and decreased sex hormone-binding globulin have been found to be associated with alcohol consumption (Dorgan et al., 1994; Muti et al., 1998; Ginsburg, 1999; Onland-Moret et al., 2005). Chronic heavy alcohol consumption also affects aromatization of androgens to estrogens in animal models (Purohit, 2000). These mechanisms may be important for hormone-sensitive tumors such as breast and prostate cancers. Physical properties of alcohol may also be of significance, particularly for tissues that come in direct contact with alcoholic beverages, such as the upper gastrointestinal tract. It has been hypothesized that the solvent properties of alcohol may enhance the effects of exposure to carcinogens such as those in tobacco. As detailed below, there is evidence of interactive effects of alcohol consumption and smoking on the risk of several cancers. Moreover, local exposure to alcohol may have proliferative effects on tissues (Seitz et al., 1998). Alcohol may affect tissues because of its effects on membrane fluidity (Simonotti et al., 1995). Alcohol has been shown to have myriad effects on human physiology. Depending on the dose and frequency of use, this exposure

243

244

PART III: THE CAUSES OF CANCER

appears to have an effect on cancer incidence. Important differences among individuals in the way alcohol is metabolized may stem from differences in genetics, nutrition, and factors such as exposure to other carcinogens. Understanding this variation allows us to better understand the role of alcohol in cancer etiology and progression.

METABOLISM OF ALCOHOL AND ALCOHOLIC BEVERAGES: ROLE OF GENETIC VARIATION AND OTHER FACTORS Alcohol is absorbed by the small intestine and delivered to the blood and then the liver. Absorption by the gut is virtually complete. Although approximately 90% of alcohol metabolism occurs in the liver (Agarwal, 2001), there is evidence of metabolism of alcohol by the intestinal tract and of activity of alcohol-metabolizing enzymes in other tissues (Saleem et al., 1984). The increased concentration of alcohol in the blood after ingestion depends on the rate of gastric emptying and of first-pass metabolism of alcohol by the intestinal tract and the liver (Li et al., 2001). Unmetabolized ethanol is distributed to other body organs, with the relative concentrations in each organ depending on the water content of the organ (Ramchandani et al., 2001). There are three possible pathways for oxidation of alcohol to acetaldehyde: by alcohol dehydrogenases (ADHs) in the cytosol of the cell; by the microsomal ethanol-oxidizing system (MEOS) in the endoplasmic reticulum; and by the catalase in the peroxisomes. The primary pathway is by the ADHs. The MEOS system is induced, leading to alcohol metabolism by cytochrome p450 2E1, at extremely high levels of intake and with chronic alcohol consumption (Lieber, 2000). Catalase, dependent on H2O2, likely does not play a significant role in alcohol metabolism in vivo. Acetaldehyde is further metabolized by acetaldehyde dehydrogenases (ALDHs) (Lieber, 2000). Nonoxidative condensation of alcohol catalyzed by fatty acid ethyl ester synthases may contribute to alcoholrelated tissue damage, particularly in the heart (Agarwal, 2001). Furthermore, acetaldehyde may be metabolized to acetate by enzymes other than ALDH (e.g., XOR and AOR). Several factors affect the rate of alcohol metabolism and subsequent breakdown of acetaldehyde: genetic variation, nutritional factors, other exogenous factors. Research regarding genetic variation in alcohol metabolism indicates that enzyme variants explain part of the variation in drinking behavior and in susceptibility to diseases associated with alcohol consumption, including cancer. There are genetic variants of both the ADHs and the ALDHs. In addition to affecting the rate of alcohol metabolism, these genes may affect the amount of alcohol consumed because of their effects on the rate of clearance of acetaldehyde. Acetaldehyde accumulation produces unpleasant sensations such as flushing that can affect drinking behaviors. Genetic variation in other genes, such as those in the dopamine pathway, may affect the experience of reward from alcohol consumption and may have an effect on alcohol consumption itself (Kitson, 1999). The alcohol dehydrogenases are a family of enzymes; the genes for these proteins map closely to each other on chromosome 4 (Edenberg, 2000). The nomenclature for the genes has recently been changed (Duester et al., 1999). The genes have been grouped into classes; those in class one, thought to be key to alcohol metabolism, are expressed in large amounts in the liver (Edenberg, 2000). They include ADH1A, ADH1B, and ADH1C (ADH1, AHD2, and AHD3, respectively, in the old nomenclature; the new nomenclature is used here). The enzymes are dimeric; the two subunits can be the gene product from the same or a different allele at the same locus or from different ADHs. There are at least seven genetic loci for humans, resulting in at least 20 isozymes that vary in terms of the required alcohol concentration for metabolism, the maximal rate of metabolism, the preferred substrate, and activity by tissue type (Agarwal, 2001). Several ADH genes with common variants that affect function have been identified. In the ADH1B gene, there are three known alleles, designated *1, *2, and *3 (Edenberg, 2000).The *2 and *3 differ from the *1 in a single amino

acid. The maximum rate of reaction (Vmax) for the enzyme coded by the *2 is approximately 40-fold greater than for that coded by the *1. It appears that there are differences in alcohol consumption between those with the *2 variant compared to those with the *1 (Eriksson, 2001; Hasin et al., 2002), and there is evidence that those carrying the *2 are at lower risk for alcoholism (Chen et al., 1996; Eriksson, 2001). The *2 is found in 70% or more of Asian populations (Eriksson, 2001); the prevalence for Caucasians is generally much lower, less than 5% (Borras et al., 2000; Eriksson, 2001). However, prevalence may be as high as 20% among Ashkenazi Jews (Eriksson, 2001; Carr et al., 2002). One study showed an interaction of this variant with alcohol consumption and breast cancer risk (Stumer et al., 2002). The enzyme coded by the *3 variant can catalyze alcohol metabolism much more efficiently than the other two, leading to a more rapid drop in blood alcohol concentrations. This allele has been identified only among those of African ancestry. There is some evidence that this variant too can affect alcoholism (Ehlers et al., 2001). There are two alleles in the ADH1C gene. The difference in maximum rate of reaction is much smaller than for the ADH1B alleles, differing by a factor of about two. The *1 codes for the more rapid variant enzyme (Bosron and Li, 1986). A number of studies have found an increased risk of cancer at a number of sites associated with the *1 variant (Coutelle et al., 1997; Harty et al., 1997; Freudenheim et al., 1999). In another study, no association with breast cancer was observed, but there was evidence of variation in blood estrogen concentrations associated with the variant (Hines et al., 2000). There may be linkage disequilibrium between this gene and the ADH1B*2 allele in European populations (Borras et al., 2000); observed associations for ADH1C may be confounded by the ADH1B variant. There are two variants in the ADH2 gene; one in the promoter region has been shown to affect function. This ADH is active at higher alcohol concentrations (Edenberg et al., 1999; Stromberg et al., 2002). ALDH also exists in multiple forms. It is believed that ALDH2, a mitochondrial enzyme, is primarily responsible for acetaldehyde oxidation. One variant of ALDH2 is found in approximately half of the Japanese and Chinese populations. In vitro, this variant, ALDH2*2, has virtually no activity; but individuals who are homozygous for this variant are highly sensitive to alcohol consumption, reacting with flushing of the skin, increased heart rate, and increased skin temperature (Ramchandani et al., 2001). Heterozygotes tend to be sensitive as well, though less so than homozygotes. Because of this reaction, those with the ALDH2*2 gene tend to drink less and are less likely to have alcohol dependence (Chen et al., 1996, 1998). Among those who metabolize acetaldehyde slowly but drink nonetheless, there may be increased organ damage. For example, Matsuo et al. (2001) and Yokoyama (2001) observed that ALDH2*2 was associated with a substantial increase in the risk of esophageal cancer. Several polymorphisms have been identified in CYP2E1, which as noted, is induced by activation of the MEOS system; it is not clear that the polymorphisms affect function. In a study of Japanese men, one variant in the 5¢-flanking region did not affect the amount of alcohol consumed. However, among heavy drinkers, it was associated with alcoholic liver disease (Tanaka et al., 1997). Among African American drinkers, one variant, a 96-bp insertion in the regulatory region that affects gene activity, increases activity in drinkers (McCarver et al., 1998; McCarver, 2001). Other factors may also affect the rate of alcohol metabolism. Both endogenous and exogenous factors affect the rate of gastric emptying and the subsequent rate of absorption of ethanol from the gastrointestinal tract. The rate at which alcohol is consumed and the foods that are consumed with it affect this process. Independent of the effect of concurrent food consumption on alcohol absorption, consumption of other foodstuffs with alcohol increases the elimination of alcohol. This may result from blood flow in the liver, enzyme activity, or other factors. This effect does not depend on the relative food composition of fat, protein, or carbohydrate. Body height and body weight affect the concentration of alcohol in the blood after ingestion of a fixed amount, clearly the result of both endogenous and exogenous factors. The genders differ in terms of the elimination of alcohol. Some but not all of this difference can be attributed to differences in lean body

Alcohol mass (Li et al., 2001). There are also differences in first-pass metabolism between men and women under age 50 (Lieber, 2000). It has been estimated that 50% of the variance in the rate of alcohol elimination is heritable (Li et al., 2001).

EPIDEMIOLOGIC RESEARCH ON ALCOHOL AND CANCER: STRATEGIC OPTIONS Among the first data used to describe associations between alcohol and cancer risk are those derived from ecologic study designs. These designs evaluate the correlation of population measures of alcohol consumption with the population risk of a given cancer or set of cancers. The unit of analysis is a group or population sector, rather than an individual: it may be a state or province within a nation, or it may be the nation itself. Thus, for example, Macfarlane et al. (1995) studied the correlation of alcohol consumption and mortality due to upper aerodigestive tract cancer among men in 25 industrialized nations and found that alcohol intake and lung cancer were strong copredictors of subsequent oral cancer. Limitations of this method are, first, that aggregates, rather than individuals, are studied, so the causal dynamics of ecologic correlations are not well identified. In addition, the roster of ecologic unit characteristics that can be studied is large, so identifying which of these characteristics are key and which are not is often not straightforward. The case-control study design has been an important source of information regarding the importance of alcohol in cancer epidemiology. This study design compares alcohol exposure among cases—individuals who have a given cancer—to exposure among controls—individuals who are comparable but do not have cancer. The reference period for this analysis is a fixed time span prior to the interview; the researcher seeks to ensure that recent illness, which could influence an individual’s alcohol intake, does not distort the true association of alcohol and cancer risk. It has been well documented that this method has great potential to provide unbiased estimates of the association of alcohol and cancer (Cornfield, 1951, 1956; Miettinen, 1976; Miettinen and Cook, 1981). However, a critical limitation of this method in practice is that samples of cases might not represent all cases, and that samples of controls may not represent all those who conceivably could have the cancer under study but do not. Because some patterns of alcohol consumption might alter the probability that individuals are contacted and agree to be interviewed, identifying and obtaining appropriate samples of those with and without cancer may be difficult. In addition, the study participants usually must attempt to recall their exposure at some point in the past, a period often specified as 1 or 2 years before the interview. People can indeed recall some facets of their prior alcohol exposure (Marshall et al., 1992; Russell et al., 1997; McCann et al., 1999). On the other hand, it is likely that some degree of inaccuracy in subject recalls of prior exposure would prove a likely source of bias (Marshall et al., 1980, 1981). Whether the inaccuracy in reports is not markedly different for cases and controls is not clear. The prospective design has several important advantages in that there is generally no requirement for recall of intake in the past. Furthermore, if a sample is followed well, concerns regarding sampling bias can be dismissed (Willett, 1998). The prospective design requires that alcohol exposure data be collected from a sample of individuals, and that these individuals be followed prospectively, with their cancer incidence noted over an extended time span. The pattern of cancer risks of individuals in the various alcohol exposure categories is then used to address the association of exposure and risk. The first limitation of this method is that either an extremely large sample must be collected or the sample must be followed for an exceedingly long time. The logistics of such an effort are considerable, and loss to follow-up can have a substantial impact on study results. Again, the fact that some patterns of alcohol use might decrease the probability that subjects are accessible for follow-up and may affect who participates in a particular cohort presents a challenge to research on alcohol intake and cancer.

245

BIAS IN STUDIES OF THE EPIDEMIOLOGY OF ALCOHOL AND CANCER Studies of the impact of alcohol on cancer risk in human populations are susceptible to several sources of bias. Largely, these result especially from the impact of some patterns of alcohol consumption on study participation. First, the samples of subjects enrolled in studies might be different from the populations of individuals who are at noteworthy risk; associations observed in these samples might be different from the associations that would be observed if the populations of at-risk individuals were studied. For example, in a pooled analysis of alcohol intake and breast cancer risk, nearly 30% of the participants described themselves as nondrinkers, and the average alcohol intake ranged from that contained in one-third to one drink per day (SmithWarner et al., 1998). However, many in some Western industrialized societies consume a good deal more than this. Clearly, the problem of enrolling, obtaining data from, and maintaining contact with individuals whose alcohol intake might place them at altered cancer risk presents a challenge to epidemiologic research on alcohol and cancer. Second, the information collected on alcohol consumption might be prone to general inaccuracies in reporting. A larger concern is that individuals with greater alcohol consumption might be more likely to underreport their consumption than individuals with average or belowaverage consumption. Both of these factors could decidedly decrease the congruence of observed and true associations of alcohol use and cancer risk (Marshall et al., 1981; Gregorio et al., 1985; Freudenheim and Marshall, 1988). An important dimension of alcohol exposure is the timing and duration of that exposure: it is possible that the impact of alcohol intake varies by cancer. The lag between intake and risk may be a few years for some cancers and several decades for others. The appropriate time span must be identified and reported accurately. The farther in the past, the more likely it is that its role cannot be accurately assessed. Furthermore, average intake is generally queried, whereas the patterns of intake may be important. Aspects of alcohol consumption such as the number of drinks per drinking occasion and whether alcohol is consumed with meals may affect the biologic impact of consumption. Third, alcohol intake is part of a network of behavioral patterns that might have additional effects on cancer risk. More specifically, alcohol intake is highly associated with several facets of tobacco use, dietary practice, physical activity, occupation, and environmental exposure. It may, in addition, be associated with several physical conditions, such as oral and gastrointestinal health. An accurate description of the importance of alcohol intake to cancer risk requires accurate information on exposure to these other factors (Marshall and Hastrup, 1996) as a first step to the appropriate evaluation of their relevance. These challenges to alcohol and cancer research are exacerbated by the possibility that much of the deleterious impact of alcohol on risk for some cancers is concentrated among individuals who have exceedingly high intake. Corrao et al. (1999) concluded that the number of high quality studies addressing the shape of the dose-response between alcohol intake and cancer risk was not sufficient to permit conclusions to be drawn. As noted herein, heavy alcohol intake, binge drinking, and alcohol abuse have decided long-term effects on digestion and metabolism. Long-term alcohol abuse negatively affects liver and kidney function, whereas moderate alcohol intake does not; it is likely that these extraordinary abuse effects are transmitted to several organ sites. Alcohol abusers suffer from a number of competing risks of premature debilitation and death. Studying the effects of alcohol in light of the difficulties of identifying, contacting, obtaining data from, and maintaining contact with individuals whose alcohol intake is heavy or excessive remains a serious challenge.

SITE-SPECIFIC FINDINGS Oral Cavity/Pharynx One of the more consistent findings in the epidemiology of oral cancer is that alcohol intake is associated with increased risk. This association has been noted in both prospective and case-control studies.

246

PART III: THE CAUSES OF CANCER

Boffetta et al. (2001) evaluated the standardized incidence ratio as a measure of oral cancer risk among men and women with a history of hospitalization for alcoholic cirrhosis or pancreatitis: among each group of patients, the risk of oral cancer was increased about 12 times that of the general population. Kjaerheim et al. (1998) evaluated a cohort of nearly 11,000 Norwegian men and found the top category of alcohol consumption to be associated with a nearly fourfold elevation of oral cancer risk. Because oral and pharyngeal cancers are relatively uncommon, case-control studies have played an important role. Wynder et al. (1957), reporting one of the first widely cited case-control studies, found that alcohol intake was associated with a more than fivefold and nearly eightfold elevations, respectively, of oral and pharyngeal cancer risk. Martinez (1969) observed that elevated alcohol intake in excess of four drinks per day was more common in oral and pharyngeal cancer cases than among controls. Bross and Coombs (1976) reported that alcohol intake was associated with an oral cancer relative risk of approximately 3.4. Graham et al. (1977) used the case-control study design, with noncancer hospital controls, to consider alcohol consumption in the context of tobacco consumption and compromised dentition, finding that consumption of two or more drinks per day was associated with a relative risk of almost 3; they observed, however, that the impact of alcohol consumption was diminished among heavy smokers. Tuyns et al. (1988) found elevated alcohol intake to multiply the relative risk of pharyngeal cancer by approximately 10- to 12fold. Considering the effects of alcohol consumption on the risks of oral and pharyngeal cancer combined, Blot et al. (1988) found that elevated alcohol intake multiplied the relative risk of oral cancer by approximately ninefold. Constructing an index of lifetime consumption of alcohol and representing this consumption by the total estimated kilograms of ethanol intake, Schlecht et al. (2001) found the top category of alcohol intake among nonsmokers to be associated with an approximate doubling of oral cancer risk. Franceschi et al. (1990, 1992) and Zheng et al. (1990) found the oral cancer risk to be about tripled by elevated alcohol intake. Identifying the specific importance of alcohol versus a key behavioral correlate of alcohol intake—tobacco—is not straightforward. Both alcohol intake and tobacco ingestion are associated with substantially increased risk of oral cancer, and the ingestion of each is highly correlated with that of the other. Both are measured with some error, so it is difficult to identify their separate contributions to risk. Alcohol and tobacco may interact to influence risk, so the impact of alcohol could be altered by the amount of smoking. The nature of the dose-response between alcohol intake and oral cancer risk may well be related to the impact of alcohol in the various populations sampled. For example, Gronbaek et al. (1998) found in a prospective analysis that intake of two or three alcoholic drinks per day was associated with an approximate doubling of oral cancer risk, consumption of six to nine drinks per day was associated with a relative risk of 5, and consumption of 10 or more drinks per day was associated with a relative risk of nearly 12. On the other hand, Mashberg et al. (1993), observing a flattening of the relative risk with consumption of 10 or more drinks per day, noted that this flattening had been observed in other studies. The variability of alcohol intake observed in various studies is considerable and noteworthy. The quantities of alcohol consumed by Franceschi’s participants were much larger than in many other studies in the literature. In the study by Franceschi et al. (1999) the lightest drinking category was for intake of three or fewer drinks per day. The lightest drinking category in the study of Mashberg et al. (1993) was up to one drink per day, and the maximum category in the study by Graham et al. (1977) was two or more drinks per day. An approximate doubling of the relative risk of oral cancer was observed in Southern India; however, more than 50% of subjects never consumed alcohol, and only 10% consumed as many as two drinks per day (Balaram et al., 2002). Dal Maso et al. (2002), observed that alcohol consumption apart from meals is associated with a greater elevation of oral and pharyngeal cancer risk than is consumption only with meals. Similarly, alcohol intake is often accompanied by diminished oral health, and poor oral health is associated with increased risk of oral

cancer. Wynder et al. (1957), observed an excess of edentia among oral cancer patients compared to controls, and Graham et al. (1977) observed a relative risk associated with poor dentition in excess of 4.5. The finding that mouthwash use is associated with a slight increase of oral cancer risk (Weaver et al., 1979; Blot et al., 1983; Wynder et al., 1983; Marshall et al., 1992) could be interpreted as evidence that poor oral hygiene, with mouthwash serving as a surrogate indicator for poor oral hygiene or integrity, increases risk. Diminished oral integrity could well stem from several dimensions of alcohol intake: direct toxicity due to alcohol and alcohol by-products, poor diet, and the impact of tobacco use. Identifying the specific importance of alcohol and distinguishing it from the impact of poor oral hygiene, poor diet, and smoking presents serious analytic challenges. Several investigations have addressed the impact of the type of alcohol consumed. This might be important because one of the potential means by which alcohol could affect oral cancer risk would be through its contact with oral and pharyngeal tissues. Thus, Wynder et al. (1957) found whiskey to be associated with the greatest increase in the relative risk of oral cancer, although other American studies did not find this to be true (Keller and Terris 1965; Williams and Horm, 1977). Martinez (1969) observed little evidence that any one type of alcohol was more harmful than any other: however, wine consumption was extremely low in the population Martinez studied in Puerto Rico. Gronbaek et al. (1998) prospectively compared wine, beer, and spirit consumption, finding wine to be associated with a much lower, even diminished relative risk. Indeed, the only increased risk associated with alcohol consumption was seen among those who consume no wine. The greatest elevation of relative risk was among those for whom 30% or more of consumption was spirits. Similarly, studies in Cuba (Garrote et al., 2001) and Brazil (Schlecht et al., 2001) showed that the greatest elevation of risk was associated with spirits consumption. Consumption of two liquors commonly consumed in Greece—ouzo and tsipouro—was associated with greater elevations of relative risk than was consumption of wine (Zavras et al., 2001).

Larynx Because laryngeal cancer is also relatively uncommon among cancers, studying the role of alcohol has depended substantially on the casecontrol method (Wynder et al., 1956, 1976). According to Wynder et al. (1956), the quantity of alcohol consumed increased the risk, which persisted within categories of tobacco consumption. Although this report distinguished the extrinsic from the intrinsic larynx, it did not reveal large differences in the effects of alcohol on risk. A later report (Wynder et al., 1976) indicated that a substantial change had taken place in the male/female ratio, suggesting that this change corresponded to a rapid increase in smoking among women. The report also confirmed that increased relative risk of laryngeal cancer was associated with alcohol consumption (Wynder et al., 1976). The first of the studies reported by Wynder et al. (1957) showed increased risk for those who consumed greater proportions of beer and whiskey. Zatonski et al. (1991) found increased risk in Poland only for consumption of vodka: the risk associated with vodka consumption rose in a dose-response pattern according to years of consumption, so that with more than 30 years of regular vodka consumption the relative risk of laryngeal cancer, adjusted for cigarette consumption, was multiplied by approximately 10 times. Wynder et al. (1976) reported that a more important factor than total alcohol intake was occasional binge drinking. The importance of the drinking pattern is also seen in the finding of Dal Maso et al. (2002) that consumption of alcohol apart from meals, versus consumption only with meals, was associated with near doubling of the risk. As the effects of both alcohol and tobacco appear relevant to laryngeal cancer risk, distinguishing their separate roles and their interaction poses a challenge. Wynder et al. (1956, 1976), utilizing early statistical techniques to control for limited numbers of other variables, concluded that the approximately twofold elevation of risk they observed with alcohol intake was independent of cigarette and other smoking. Later studies, including those of Williams and Horm (1977), Graham et al. (1981), Tuyns et al. (1988), Falk et al. (1989),

Alcohol Sankrayrayanan et al. (1990), Franceschi et al. (1994), and Dosemeci et al. (1997), using newer techniques to control for tobacco, observed similar patterns. Other studies controlling for tobacco intake, including those of Burch et al. (1981), Elwood et al. (1984), Olsen et al. (1985), De Stefani et al. (1987), Zatonski et al. (1991), Choi et al. (1991), Maier et al. (1992), Freudenheim et al. (1992), and Muscat and Wynder (1992), have observed larger (generally at least fourfold) relative risks associated with alcohol intake. Only one large study that adjusted for tobacco intake—that reported by Zheng et al. (1992a, 1992b)—showed no elevation of risk with alcohol intake. In the population Zheng et al. studied, alcohol intake was relatively low: half of the subjects never consumed alcohol; and among those who drank alcohol, the top quartile began at four drinks per day. Bosetti et al. (2002) considered the impact of alcohol among nonsmokers and of smoking among nondrinkers. Their findings indicated that among the few alcohol consumers who used no tobacco the relative risk of laryngeal cancer was increased about 2.5 times. For comparison, the relative risk associated with cigarette smoking among nondrinkers is between 13 and 14. Altieri et al. (2002) compared the impact of stopping alcohol consumption to that of stopping smoking. They found that cessation of tobacco consumption had a substantial effect on relative risk, leading to a nearly fourfold decrease in risk within 10 years, but that cessation of alcohol consumption had little impact on risk for nearly 20 years. These findings can either be interpreted as evidence that alcohol alone does not affect risk or that alcohol-induced damage is irreversible. To date, many studies of laryngeal cancer have treated the larynx as a single organ. In terms of exposure, though, two regions can be identified: the laryngeal-pharyngeal junction, which is exposed directly to alcoholic beverages, and the main body of the larynx—the endolarynx—which has no exposure to liquids. Wynder’s 1956 report distinguished the two (Wynder et al., 1956). It is certainly possible that the part of the larynx exposed to liquids could be more influenced by alcoholic beverage consumption than the sector exposed only to air. Tuyns et al. (1988) distinguished the endolarynx from the hypopharynx, or junctional area, and found substantial effects of alcohol on cancer risk in both regions. Controlling for cigarette consumption, they reported that at each level alcohol intake was associated with increased risk. The importance of extremely high intake was noteworthy: among light smokers, the relative risks associated with 0–40, 41–80, 81–120, and >120 g of alcohol per day, respectively, were 1.0, 1.6, 2.3, and 3.8. Among those whose smoking was maximal—those who consumed more than 25 cigarettes per day—the corresponding relative risks were 11.5, 18.5, 23.6, and 43.2. For the hypopharynx and epilarynx, however, relative risks among light smokers ranged, respectively, from 1.0 to 3.0 to 5.5 to 14.7 among those consuming 0–40, 41–80, 81–120, and 121 or more drinks per day; among heavy smokers, the corresponding relative risks ranged from 4.9 to 18.4 to 37.6 to 135.5. Clearly, then, although intake of alcohol has a greater impact on those regions of the larynx directly exposed to alcohol, relative risk also is modified in regions that are not directly exposed (Tuyns et al., 1988). As with oral cancer, given the massive impact of tobacco intake on laryngeal cancer and the close connection of tobacco and alcohol intake, it is possible that a substantial proportion of the association of alcohol intake and laryngeal cancer stems from incomplete control for tobacco intake. The association of alcohol and cancer of the endolarynx, not directly exposed to alcohol, tends to be modest: around 2. The latter association can possibly be attributed in large part to resonant confounding by cigarette smoking.

Esophagus Few organs except the oral cavity, hypopharynx, and esophagus are directly exposed to the predigested agents in alcoholic beverages. Thus, observations that esophageal cancer patients were often heavy alcoholic beverage consumers (Wynder and Bross, 1961) led to the development of a substantial epidemiologic literature linking alcoholic beverage intake and esophageal cancer risk. The importance of alcohol vis-à-vis tobacco has been of concern. Although the esophagus is not

247

directly exposed to tobacco smoke, it is bathed in residual condensates of tobacco smoke that accumulate in the mouth and pharynx. Addressing the possible confounding effects of alcohol and tobacco, Wynder and Bross (1961) observed that the then-recent rise in lung cancer associated with cigarette smoking had not been accompanied by a parallel rise in esophageal cancer. The association of heavy alcohol use with substantially increased risk of esophageal cancer is quite consistent, with a large number of studies indicating relative risks of 8–10 for heavy consumption (Wynder and Bross, 1961; Martinez, 1969; Tuyns et al., 1977, 1979; Pottern et al., 1981; Vassallo et al., 1985; Victora et al., 1987; Yu et al., 1988; Cheng et al., 1992; Gao et al., 1994; Hu et al., 1994; Castellsague et al., 1999; Gallus et al., 2001; Wu et al., 2001; Boonyaphiphat et al., 2002). Dal Maso et al. (2002) observed a relative risk of more than 13 for consumption of eight or more drinks per day. Among these studies, there appears to be a sharp rise in risk associated with the upper ranges of consumption. A number of small studies, including those of Bradshaw and Schonland (1974), De Jong et al. (1974), Notani (1988), De Stefani et al. (1990), Sankaranarayanan et al. (1991), Sharp et al. (2001), and Yang et al. (2004) have observed smaller or negligible relative risk elevations. Increases in risk were also observed in prospective studies by Schmidt and Popham (1981) and Hirayama (1990). With the exception of Wynder and Bross (1961), all of these relative risks are derived with some adjustment, either by matching or in analysis, for smoking practices. Although the evidence tends to indict alcohol as more strongly related than cigarette smoking to esophageal cancer risk, the difference between these associations is not great. In addition, tobacco and alcohol may interact to increase risk. For example, Castellsague et al. (1999) found that among nonsmokers the top category of alcohol consumption was associated with a 14-fold elevation of risk; among heavy smokers, this top category was associated with a 51-fold elevation of risk. Gao et al. (1994) found that alcohol consumption among nonsmokers was associated with a relative risk of 4, whereas consumption among the heaviest smokers was associated with a relative risk of 12. Ke et al. (2002) found a pattern that suggests a three-way interaction involving consumption of alcohol, tobacco, and black Chinese tea. Ke et al. observed that the consumption of this tea at a high temperature was associated with a risk elevation and that it interacted with cigarette smoking to increase risk, much as alcohol does. There does not appear to be a strong pattern identifying any specific form of alcohol as more likely than others to increase esophageal cancer risk. Several studies have shown that the consumption of distilled spirits is associated with the greatest risk modification (Wynder and Bross, 1961; Tuyns, 1977, 1979; Pottern et al., 1981; Victora, 1988; Hu et al., 1994; Gronbaek et al., 1998); whereas others (Vassallo et al., 1985; Sankanarayanan et al., 1991; Cheng et al., 1992) observed little evidence of such a pattern. Segal et al. (1988) observed the greatest risk association among native South Africans with the consumption of traditional beer; and Yu et al. (1988) in China and Zambon et al. (2000) in Italy and Switzerland observed wine consumption to be a greater risk factor than either beer or spirit consumption. De Jong et al. (1974) found increased relative risk for those whose major form of alcohol was a distilled product, samsu. A number of facets of alcohol consumption could be related to its impact. Dal Maso et al. (2002) observed that alcohol consumption outside of meals was associated with little alteration of risk independent of quanitity. Early studies of esophageal cancer focused on squamous cell cancer (Wynder and Bross, 1961; Bradshaw and Schonland, 1974; Vassallo et al., 1985; Victora et al., 1987; Cheng et al., 1995; Talamini et al., 2000; Bollschweiler et al., 2002; Ke et al., 2002; Wu et al., 2002). More recently, however, there has been a steep increase in the incidence of adenocarcinoma of the esophagus, so that for whites in the United States the rate of adenocarcinoma of the esophagus is now as high as that of squamous cell cancer. Several recent studies have focused on the difference between the risk patterns for adenocarcinoma versus squamous cell cancer of the esophagus (Pera and Pera, 2001). Among whites, the rates for squamous cell cancer and

248

PART III: THE CAUSES OF CANCER

adenocarcinoma of the esophagus are approximately equal; among African Americans, the incidence of squamous cell carcinoma is 20 times that of adenocarcinoma. The epidemiologic divergence of these cancers with respect to alcohol is becoming clearer (Pera and Pera, 2001; Mayne and Navvaro, 2002): Wu et al. (2001), Lagergren et al. (2000), Dhillon et al. (2001), Bollschweiler et al. (2002), and Gao et al. (1994) found alcohol to be weakly or not associated with adenocarcinoma of the esophagus. In certain regions such as Assam, India, where cigarette smoking is less common than betel nut chewing (Phukan et al., 2001), alcohol appears to impart a risk elevation similar to that seen among smokers in the West; it may interact with betel nut chewing to increase risk, much as it does with cigarette smoking.

Stomach Stomach cancer in the United States and other Western countries has declined substantially over the past five decades. The reasons for this decline are not entirely clear, but it has made stomach cancer rare enough to be difficult to study in these populations. In addition, changes in exposure to the critical factor or factors responsible for this decline may confound assessments of the importance of alcohol. One of the key characteristics of this literature is the unanimity with which relatively strong prospective studies have failed to document evidence of etiologic importance for alcohol (Hakulinen et al., 1974; Jensen et al., 1978; Jensen, 1979, 1983; Kono and Ikeda, 1979; Klatsky et al., 1981, 1988; Schmidt and Popham, 1981; Kono et al., 1983, 1985, 1986; Adami et al., 1988, 1992a, 1992b; Hirayama, 1989; Carstensen et al., 1990; Hirayama, 1990; Tonnesen et al., 1994; Galanis et al., 1998). These studies approach stomach cancer in a broad range of populations. The studies reported by Hirayama, Klatsky et al., Adami et al., and Galanis et al. are based on general population samples, whereas those of Kono et al. were reported for physicians. The studies reported by Schmidt and Popham, Hakulinen et al., and Tonnesen et al. are based on samples of alcoholics or problem drinkers, whereas those reported by Carstensen et al., Jensen, and Dean et al. (1979) are based on data from brewery workers. In a cohort of hepatitis B surface antigen (HBsAg)-positive blood donors (Oshima et al., 1984), no association of risk and alcohol consumption was observed. These studies together appear to consider a broad range of use patterns and risk groups, including members of the general population, those who have used alcohol to excess, and those who, by virtue of ready access to alcoholic beverages, are likely to consume large amounts of alcohol. It is possible, of course, that within each of these populations there is too little variance in exposure to allow adequate examination of risk variance. There is considerable variance among case-control studies regarding the associations observed between alcohol intake and gastric cancer risk. Studies reported by D’Avanzo et al. (1994), Ferraroni et al. (1989), Chen et al. (2000), Chow et al. (1999), Ye et al. (1999), Setiawan et al. (2000), Rao et al. (2002), and Hansson et al. (1994) revealed no association. Chow et al. (1999) observed a curvilinear pattern, with the risk lower among light drinkers than among nondrinkers, the risk then increasing among the heaviest drinkers to a level slightly but insignificantly higher than among nondrinkers. Haenszel et al. (1972) observed a consistent elevation of risk among first generation Japanese migrants living in Hawaii but no elevation among the second generation. Jedrychowsky et al. (1993) observed a strong association of increased risk and vodka drinking. This risk was stronger among those who consumed vodka on an empty stomach, and it was stronger yet among those who reported consuming vodka before breakfast. On the other hand, several studies revealed increased risk among drinkers: Munoz et al. (2001) observed elevated relative risk among alcohol users and a higher relative risk among ex-users. Ji et al. (1996), Lee et al. (1990), Wu-Williams et al. (1990), Zaridze et al. (2000), De Stefani et al. (1998), Hu et al. (1988), and Boeing et al. (1991) observed increased risk with increased alcoholic beverage consumption. None of these risks was large, most being in the vicinity of 1.5–3.0. It is difficult to discern a pattern in these data: wine and beer

were equally likely to increase risk in the study of Haenszel et al. (1972), whereas beer and liquor were linked to increased risk and wine to decreased risk in the study of Boeing et al. (1991). Ye et al. (1999) observed decreased risk among wine drinkers and increased risk among whisky drinkers. Interactions among alcohol and smoking have been observed. Chen et al. (2000) observed that alcohol was associated with a relative risk of about 1.5. Among smokers, this relative risk was 3.0, whereas among ex-smokers it was 1.7. Hansson et al. (1994) observed a strong, statistically significant interaction of tobacco and alcohol, so that among nonsmokers alcohol use was associated with decreased risk; among smokers, it was associated with a significantly increased risk. As attention has focused on the distinction between adenocarcinoma and squamous cell cancer of the esophagus, a distinction is being drawn between cancer of the gastric cardia and of the distal stomach. It has been suspected that cancer of the gastric cardia and adenocarcinoma of the esophagus might share risk factors. Several epidemiologic inquiries have considered cancer of the gastric cardia apart from that of the distal stomach. Ji et al. (1996) found no risk associated with alcohol consumption. Wu-Williams et al. (1990), Zaridze et al. (2000), and Jedrychowski et al. (1993), however, did find increased risk associated with alcohol consumption. Jedrychowski et al. found, as for cancer of the distal stomach, an association of vodka consumption with increased risk.

Large Bowel Little of the alcohol consumed by an individual directly contacts the large bowel; as already noted, it is absorbed into the bloodstream through the stomach and the small bowel. Some in the blood comes in contact with the large bowel. More likely to come in contact with the lumen of the large bowel are residual nonalcoholic constituents of alcoholic beverages; these, which clearly differ for beer, wine, and distilled liquors, could have a bearing on the colonic and rectal content and environment. The recent case-control literature tends to show that alcohol increases risk, although a large study by Tavani et al. (1998) observed no increased risk with alcohol intake. A report by Slattery et al. (1999) based on factor analysis indicated that a factor dominated by alcohol intake, but also characterized by low intake of whole grain and fruit intake and elevated intake of fish, was associated with marginally increased colorectal cancer risk. Sharpe et al. (2002) found increased risk of cancer of the distal colon and rectum with alcohol intake; no such pattern was apparent for cancer of the proximal colon. On the other hand, the small study of Jedrychowski et al. (2001) found sizable increases in risk with the consumption of alcohol. A large study conducted in Hawaii among Pacific Rim populations, with careful control for ethnicity, age, and gender, reported by Le Marchand et al. (1997), revealed increased risk with increased alcohol intake. In a later analysis of the same data, Le Marchand et al. (1999) found a strong interaction between alcohol intake and family history; the association of intake and risk was much stronger for those with a family history of colon cancer. A small study reported by Matsuo et al. (2002) indicated an interaction of the ALDH genotype and alcohol as a factor in colon cancer risk. The prospective literature is not entirely consistent, with mortalitybased studies reported by Doll et al. (1994), Kono et al. (1986), Klatsky et al. (1981), and Camargo et al. (1997) revealing no increase in risk. Gapstur et al. (1994) observed no increased risk with alcohol intake among the Iowa Women’s Study. Hirayama’s report (1990) of the experience of 265,000 Japanese indicated no increase in colon cancer risk but a statistically significant increase in rectal cancer risk with alcohol intake. Jensen, comparing cancer morbidity in a cohort of Danish brewery workers—individuals who might be expected to have elevated beer consumption—to that of the Danish population, observed no elevation of risk. Flood et al. (2002), who followed a large cohort of women who were part of a breast cancer screening program, observed only a minimal association of risk and alcohol intake with alcohol consumption, and Camargo et al. (1997) observed no elevation of risk among members of the Physician’s Health Study. On the

Alcohol other hand, Chyou et al. (1996) observed, in a cohort of Japanese, a substantial risk elevation with increased alcohol intake. Ma et al. (1997) studying members of the Physician’s Health Study, observed that the association of alcohol intake and risk was null or negative among those homozygous for the wild-type allele of the methylene tetrahydrofolate reductase gene. For those homozygous for the mutant allele, alcohol was associated with increased risk. Glynn et al. (1996) following members of the a-tocopherol–b-carotene chemoprevention trial, Giovannucci et al. (1995), following members of the Health Professional follow-up study, and Hsing et al. (1998), following a cohort of insurance policy holders have also observed a positive association of alcohol and risk. Evidence that adenoma is a premalignant lesion that may lead to colon cancer has led researchers to believe that alcohol is associated with increased adenoma prevalence. Ferraroni et al. (1989) and Boutron et al. 1995 were among the few who found alcohol to be associated with increased risk. On the other hand, Giovannucci et al. (1993), Manus et al. (1997), Martinez et al. (1995), Yamada et al. (1997), and Todoroki et al. (1995) observed positive associations between alcohol intake and adenoma prevalence. The analysis by Baron et al. (1998) revealed a positive association between alcohol intake and polyp recurrence. In light of the suspicion that the predecessor of adenoma is the hyperplastic polyp, Kearney et al. (1995) found that alcohol was positively associated with the prevalence of hyperplastic polyps, which can be taken as additional evidence of the etiologic significance of alcohol intake. Thus, evidence that alcohol intake increases the risk of colorectal cancer appears to be accumulating. This evidence is modestly more consistent for cancer of the distal cancer and rectum than it is for the proximal colon.

Liver As already noted, the liver is the primary organ where alcohol is metabolized, and metabolism of alcohol takes precedence over a number of normal liver functions. The liver can apparently handle alcohol in modest quantities, but heavy alcohol intake exacts a severe burden. The first priority for the liver of a heavy alcohol consumer is metabolism of alcohol. Lieber (1990) has described this process. Heavy alcohol consumption may cause the liver to store fat and protein, rather than metabolize it. The presence of stored fat and protein over time swells the liver and impairs its ability to function. Inflammation and necrosis, alcoholic hepatitis, may follow. Continued excessive alcohol intake may eventually cause irreversible replacement of functional liver cells by fibrous scar tissue. Consequences include abnormalities in vitamin and mineral metabolism, reduced ability to counter toxic substance accumulation (Lieber, 1990), and increased risk of liver cancer. Although our understanding of the specific mechanisms by which alcohol might increase the risk of liver cancer is limited, prospective studies quite consistently point to heavy alcohol consumption as increasing the risk of liver cancer. Prospective studies of alcoholic beverage manufacturing employees (Dean et al., 1979; Jensen, 1979; Carstenson et al., 1990) and of alcohol abusers (Hakulinen et al., 1974; Schmidt and Popham 1981; Tonnesen et al., 1994) revealed an increased risk of liver cancer. Prospective studies of more typical populations (Kono et al., 1986; Shibata et al., 1986; Hirayama, 1990) documented increased risk among those with increased alcohol intake. Klatsky et al. (1981) followed a cohort in the San Francisco Bay area and showed no elevation of risk, but the risk in that population was extremely low, and the sample was small. Oshima et al. (1984) showed, in a cohort of high risk individuals infected with hepatitis B, that risk of liver cancer increased in a doseresponse pattern with increasing alcohol intake; the relative risk at the highest intake level indicated an eightfold elevation of risk. Mori et al. (2000) showed, in a small cohort of individuals characterized by hepatitis B and C infections, that alcohol intake increases risk and interacts synergistically with hepatitis infection to increase risk. Case-control studies have also consistently documented a positive association of risk and alcoholic beverage consumption (Bulatao-

249

Jayme et al., 1982; Stemhagen et al., 1983; Hardell et al., 1984; Inaba et al., 1984; Trichopoulos et al., 1987; LaVecchia et al., 1988; Tsukuma et al., 1990; Braga et al., 1997; Donato et al., 2002; Yuan et al., 2004). Stemhagen et al. (1983) also observed increased risk among those involved in wine manufacturing. These studies tend to show that risk increases in at least a dose-response pattern, so those with the heaviest intake are clearly at the greatest risk. The study by Lu et al. (1988) is one of few efforts not to observe an association of risk and intake. There is little consistency, however, in the type of alcohol that most increases risk. For example, in Japan, where Shibata et al. (1986) collected data, the most commonly consumed alcoholic beverage, and hence the form of alcohol leading to increased risk, is a distilled rice liquor. In Italy, where LaVecchia et al. (1998) conducted their study, wine is the beverage most commonly consumed and hence is the risk factor. Recent efforts have also focused on the possibility that the associations observed could be confounded. Bulatao-Jayme (1982) controlled statistically for aflatoxin as derived from an index of reported intake of foods contaminated by aflatoxin. The alcohol association persisted. Inaba et al. (1984), Trichopoulos et al. (1987), Tsukuma et al. (1990), and Donato et al. (2002) controlled for hepatitis B infection. Donato et al. (2002) also evaluated possible confounding by hepatitis C. These findings show that alcohol tends to interact with hepatitis, so its impact is greater in the presence of hepatitis B or C than in its absence.

Pancreas The pathway by which alcohol reaches the pancreas is not generally direct, as most alcohol is absorbed into the blood through the stomach or the intestine and then metabolized in the liver. Nonetheless, there is abundant evidence that long-term, excessive alcohol intake exacts severe costs on the pancreas. Among men in their thirties to early forties in western industrialized countries, most male chronic pancreatitis patients have a history of extended alcohol abuse (Steer et al., 1995); alcohol and tobacco have been identified as distinct risk factors for pancreatitis (Yen et al., 1982; Talamini et al., 1999). Pancreatitis is important to the effects of alcohol inasmuch as 4% of chronic pancreatitis patients develop pancreatic cancer. The evidence that alcohol contributes substantially to the risk of pancreatic cancer, though, is mixed. Summarizing the evidence up to 1986, Velema et al. (1986) indicated that any increased risk is probably quite modest. Indeed, several strong prospective studies (Hakulinen et al., 1974; Dean et al., 1979; Jensen, 1979; Klatsky et al., 1981; Kono et al., 1986; Hirayama, 1990; Friedman and Van den Eeden, 1993; Michaud et al., 2001; Stolzenberg-Solomon et al., 2001; Isaksson et al., 2002; Lin et al., 2002) observed no association of risk and alcohol consumption practices. Among the prospective studies detecting an association (Schmidt and Popham, 1981; Heuch et al., 1983; Carstensen et al., 1990; Zheng et al., 1993; Tonnesen et al., 1994; Harnack et al., 1997; Ye et al., 2002), only those led by Heuch et al. and Zheng et al. showed more than a weak association of risk with intake. Lagergren et al. (2000) interpreted the association they observed as within a range that could be readily attributed to the resonance of confounding by cigarette smoking. Case-control studies also provide little evidence of an association of an alcohol–pancreatic cancer association (Mack et al., 1986; Clavel et al., 1989; Bouchardy et al., 1990; Farrow and Davis, 1990; Baghurst et al., 1991; Ghadirian et al., 1991; Bueno de Mesquita et al., 1992; Lyon et al., 1992; Mizuno et al., 1992; Zatonsky et al., 1993; Tavani et al., 1997; Villeneuve et al., 2000). The case-control studies revealing an association (Cuzick and Babiker 1989; Partanen et al., 1997; Soler et al., 1998; Silverman, 2001) are in a distinct minority. Among even these studies, only that by Cuzick and Babiker suggested a more than modest increase in risk. The studies showed no tendency for specific types of alcohol to have more impact than others. Several of the case-control studies went to extensive efforts to address the problem of the rapid and severe course of pancreatic cancer, either by organizing rapid case-ascertainment systems or utilizing carefully designed procedures for obtaining interviews from significant others of cases

250

PART III: THE CAUSES OF CANCER

and from controls. Even with these efforts, the association of alcohol and risk appeared extremely weak or null. The congruence of these findings with those of the substantial number of prospective studies suggests that any effect of alcohol intake on pancreatic cancer risk is slight to negligible.

Female Breast The literature on alcohol consumption and female breast cancer is marked by relatively consistent evidence of a modest impact on risk (World Cancer Research Fund, 1997; Lenz et al., 2000; Singletary and Gapstur, 2001; Collaborative Group on Hormonal Factors in Breast Cancer, 2002). That is, the relative risk associated with alcohol consumption is not great; it has been estimated that there is approximately 10% increase in risk for an increase in average consumption of one drink per day (Longnecker et al., 1988, 1994, 1995a, 1995b; SmithWarner et al., 1998). Because of the high prevalence of alcohol consumption, though, the attributable risk associated with alcohol consumption may exceed 10% (Mezzetti et al., 1998). Although there is some evidence that there is a linear increase in risk with alcohol consumption (Smith-Warner et al., 1998), there is also some evidence that risk does not start to increase until intake is above a certain threshold (Howe et al., 1990, 1991). If the latter were true, studies in populations with a low prevalence of alcohol intake might be limited in their power to detect an effect. Given the importance of menopause to an altered risk of breast cancer, the grouping of pre- and postmenopausal breast cancers in several case-control studies may be a serious limitation. On the other hand, most of these studies controlled statistically for menopausal status, so there is little evidence of strong effect modification by menopausal status and hence little reason to suspect that the analysis in these studies is significantly flawed. Many of the studies, including those reported by Byers and Funch (1982), Le et al. (1984, 1986), Miller et al. (1978, 1992), Chu et al. (1989), Mannisto et al. (1996), Freudenheim et al. (1999), Sneyd et al. (1991), Kato et al. (1992), Rosenberg et al. (1982, 1990), Iscovich et al. (1989), Morabia et al. (1990), Marcus et al. (2000), Baumgartner et al. (2002), and Kinney et al. (2000), found little evidence of an association. Toniolo et al. (1989), Katsouyanni et al. (1994), and Kinney et al. (2000) observed slight evidence of a positive trend with increased intake, but the association was not statistically significant. Baumgartner et al. (2000) found evidence of increased risk for the highest levels of intake but a significantly decreased risk among those with moderate intake. Among studies that focused on premenopausal breast cancer, many appear to show little in the way of increased risk. Strong studies, including those by Harris et al. (1988), Chu et al. (1989), Ewertz et al. (1991), Sneyd et al. (1991), Meara et al. (1989), Kato et al. (1992), Enger et al. (1999), Smith et al. (1994), Britton et al. (2002), Adami et al. (1988), and Kropp et al. (2001), provided only weak evidence of an association of alcohol intake and premenopausal breast cancer risk. On the other hand, studies reported by Bowlin et al. (1997), Longnecker et al. (1988, 1994, 1995a, 1995b), Rohan and McMichael (1988), Ranstam and Olsson (1995), Tavani et al. (1997, 1998), Swanson et al. (1997), Viel et al. (1997), Britton et al. (2002), Kropp et al. (2001) and Petri et al. (2004) tended to show increased risk with increased intake. Trentham-Dietz et al. (2000), in an analysis of carcinoma in situ, observed significantly increased risk paralleled by increased risk among women diagnosed with invasive breast cancer. Studies focused on postmenopausal breast cancer appear more likely to reveal increased risk with increased alcohol intake. Null or nearly null findings, including those by Chu et al. (1989), Ewertz et al. (1991), Sneyd et al. (1991), Meara et al. (1989), Ferraroni et al. (1989), Hirose et al. (1995), Katsouyanni et al. (1994), Enger et al. (1999), and Cade et al. (1998) have been reported. The point estimates from these studies tend to be positive, but they are within the range that might be expected by chance fluctuations. Rohan and McMichael (1988) observed a significant trend for increased intake to accompany increased risk. Ranstam and Olsson (1995) and Longnecker et al. (1988, 1995a, 1995b) observed statistically significant associations.

What is noteworthy about these case-control reports is the rarity with which they reveal other than positive associations. Although the associations tend to be statistically nonsignificant, they are almost never negative, and it is even less common that they are negative and statistically significant. In the face of this uncertainty, the evidence from cohort studies is particularly pertinent. The evidence from cohort studies is striking in its consistency. Studies reported by Garfinkel et al. (1988), Harvey et al. (1987), Friedenreich et al. (1993), Rohan and McMichael (1988), Jain et al. (1991, 1999), Schatzkin et al. (1987, 1989), Gapstur et al. (1992, 1994), Sellers et al. (2001, 2002), Hiatt et al. (1984, 1988), Boice et al. (1995), Thun et al. (1997), Fuchs et al. (1994), Willett et al. (1987), Holmburg et al. (1994, 1996), Van den Brandt et al. (1990, 1995, 2000), Horn-Ross et al. (2002), Feigelson et al. (2001), and Chen et al. (2002) revealed positive, and in most cases statistically significant, associations of intake and risk. Only those studies reported by Zhang et al. (1999), Adami et al. (1988, 1992a, 1992b), Barrett-Connor et al. (1990), Simon et al. (1991), and Graham et al. (1977, 1978, 1981) found no association of intake with risk.

Lung Consumption of alcoholic beverages does result in exposure of the lung to alcohol. As already noted, most alcohol is absorbed into the blood from the stomach and small intestine. Transported via the circulatory system to the liver, it is metabolized to acetaldehyde prior to additional breakdown. It is clearly possible for alcohol or acetaldehyde to come into contact with pulmonary tissue. Either of these could function as carcinogens or as solvents to potentiate other carcinogens, such as those in tobacco smoke. It is important, however, that alcohol comes into contact with pulmonary tissue in much smaller amounts and concentrations than with oral, pharyngeal, or esophageal tissue. Nonetheless, there is evidence linking alcohol consumption and lung cancer risk. Positive case-control study results, including those reported by Bandera et al. (1997), De Stefani et al. (1993, 1996), Rachtan and Sokoloski (1997), Carpenter et al. (1998), and Stockwell and Matanoski (1984, 1985) are balanced by negative studies, including those by Koo (1988), Mayne et al. (1994), Holst et al. (1988), and Kabat and Wynder (1984). The results reported by Koo are especially important in that the study was conducted among nonsmoking women; it is possible that among these subjects confounding by cigarette smoking is minimized. Among prospective studies, positive associations have been reported by Potter et al. (1992), Hirayama et al. (1990), Klatsky et al. (1981), Pollack et al. (1984), Carstensen et al. (1990), Dean et al. (1979), Doll et al. (1994), Prescott et al. (1999), and Chyou et al. (1995). The reports by Klatsky et al., Dean et al., and Doll et al. focused on mortality rather than on morbidity; given the lethality of lung cancer, the distortions induced by this focus are probably slight. More importantly, several prospective studies, including those by Gordon and Kannel (1984), Kono et al. (1986), Stemmerman et al. (1990), and Yuan et al. (1997) are negative. Especially important null results are those based on the Finnish a-tocopherol and b-carotene intervention trial among heavy smokers (Woodson et al., 1999) and those based on the b-carotene and retinal intervention among those with heavy exposure to tobacco and asbestos (Omenn et al., 1996). That cigarette smoking is an overwhelming risk factor for lung cancer and is strongly correlated with alcohol consumption cannot be overlooked. The likelihood of confounding of the alcohol–lung cancer association by the combination of the cigarette smoking–lung cancer association and the cigarette smoking–alcohol consumption association is extremely high; modest imprecision in the measurement of smoke exposure can readily induce resonant confounding into measures of the impact of alcohol consumption. Even miniscule (5%–10% misclassification of cigarette smoke exposure can resonate, so a null but correlated variable can appear to alter risk by 20%–50% (Marshall and Hastrup, 1996; Marshall et al., 1999). It is not necessary that the variables used to operationalize exposure be misclassified. For example, cigarettes per day and the duration of smoking can be assessed perfectly; however, to the degree that other factors (e.g., brand, filter, depth of inhalation) add further variance to smoke

251

Alcohol exposure, the investigator’s measures are less than perfect, and the opportunity for resonant confounding is excellent. Conventional methods of covariate adjustment would not eliminate this possibility of resonant confounding. Thus, in light of the overwhelming importance of cigarette smoking and the likelihood that confounding by cigarette smoking has not been completely removed, reviews by Korte et al. (2002) and White et al. (2002) that alcohol is not likely to be a significant risk factor for lung cancer appear justified.

Prostate Some ingested alcohol reaches the prostate, and the by-products of alcohol metabolism can be expected to reach and affect this organ as well. The totality of the epidemiologic evidence, however, is that alcohol is not a significant risk factor for prostate cancer. It is pertinent that the advent of the prostate-specific antigen (PSA) screening test has radically altered the identification and treatment of prostate cancer, sizably increasing its incidence (Gann, 1997). Prostate cancer is diagnosed frequently but is infrequently a cause of death (Marshall and Wood, 2002). As a result of widespread application of the PSA test, the identification of asymptomatic and possibly non-lifethreatening cases has increased substantially. Clearly, these changes have had an impact on the epidemiology of alcohol in regard to prostate cancer. The extant epidemiologic literature on alcohol and prostate cancer, comprehensively reviewed by Dennis (2000) and Dennis and Hayes (2001), is dominated by results suggesting that alcohol is not associated with prostate cancer risk. Exceptions include two case-control studies (Hayes et al., 1996; Sharpe and Siemiatycki, 2001). The report by Hayes et al. is noteworthy for its inclusion of large numbers of both European Americans and African Americans. The elevation of risk Hayes et al. observed resulted from extremely heavy alcohol consumption: the intermediate intake category was 22–56 drinks per week; and the heavy drinking category was 57 or more drinks per week. On the other hand, three other case-control studies reported no association (Jain et al., 1998; Lumey et al., 1998; Crispo et al., 2004). The report of Jain et al., as large as that of Hayes et al., observed no alteration of risk. Two prospective studies—those of Adami et al. (1992a, 1992b) and Tonneson et al. (1994)—observed an increased risk among those with elevated alcohol intake. On the other hand, prospective studies reported by Hakulinen et al. (1974), Jensen (1979), Carstensen et al. (1990), Hiatt et al. (1994), Sorensen et al. (1998), Breslow et al. (1999), and Schuurman et al. (1999) reported no evidence of an association of alcohol intake and prostate cancer risk. There is limited evidence that prostate cancer and alcohol are linked and considerable evidence that they are not.

CONCLUSIONS Clearly, alcohol consumption, a behavior, can be modified. Given evidence that alcohol consumption increases the risks of some cancers, modification of alcohol consumption represents a ready means of decreasing cancer risk. The degree to which this cancer risk might be altered depends on the significance of the cancers for which alcohol has a role and the importance of alcohol for each of those cancers. Table 14–1 presents estimates of the total numbers of cancer cases in 2002 and the total numbers of deaths for the cancer sites reviewed in this chapter. These estimates amount to some 800,000 cases and 360,000 deaths. The right-hand column presents our judgment as to whether the totality of the evidence supports a causal role for alcohol consumption. This chapter has noted that our understanding of the importance of alcohol in cancer is limited. Although large quantities of data have been collected, the complexities by which alcohol exacts its effects, as well as the difficulty of studying those effects, are significant barriers to understanding. It is not particularly satisfying to weigh the epidemiologic evidence and then merely posit that the evidence does or does not support a causative role for alcohol. Given the limitations of our methods and understanding, one could perform a meta-analysis or

Table 14–1. Estimated Cancer Cases and Deaths and Probable Role of Alcohol: United States, 2002 Cancer Site Oral cavity/pharynx Esophagus Stomach Liver Pancreas Large bowel Larynx Lung Female breast Prostate

No. of Cases

No. of Deaths

Alcohol as a Significant Factor

28,900 13,100 21,600 16,600 30,300 148,300 8,900 169,400 205,000 189,000

7,400 12,600 12,400 14,100 29,700 56,600 3,700 154,900 40,000 30,200

+ + + + + + -

Source: Cancer Facts & Figures (2002). Chicago: American Cancer Society.

otherwise squeeze the data to extrude some single number that summarizes the etiologic importance of alcohol for each cancer site. Whether this number would have significant meaning, though, could be heartily debated. Thus, our judgment is that the evidence implicates alcohol in oropharyngeal, esophageal, liver, large bowel, laryngeal, and breast cancer. Alcohol appears to be much more important to some sites than to others. It has a large impact on the risks of oral and esophageal cancer and has a much more modest impact on the risks of large bowel and breast cancer. For some sites, the evidence is difficult to dispute, whereas for others disputes will continue. In the United States in 2002, the incidence of the cancer that appeared to be alcohol-related added up to slightly fewer than 300,000 cases and 135,000 deaths. The evidence suggests that alcohol has some component causal role and that control of alcohol intake would result in reduced incidence of and mortality attributable to these cancer sites. References Adami HO, Hsing AW, McLaughlin JK, et al. 1992. Alcoholism and liver cirrhosis in the etiology of primary liver cancer. Int J Cancer 51:898– 902. Adami HO, Lund E, Bergstrom R, Meirik O. 1988. Cigarette smoking, alcohol consumption and risk of breast cancer in young women. Br J Cancer 58:832–837. Adami, HO, McLaughlin JK, Hsing AW, et al. 1992. Alcoholism and cancer risk: a population-based cohort study. Cancer Causes Control 3:419–425. Agarwal DP. 2001. Genetic polymorphisms of alcohol metabolizing enzymes. Pathol Biol 49:703–709. Altieri A, Bosetti C, Talamini R, Gallus S, Franceschi S, Levi F, Dal Maso L, Negri ELV. 2002. Cessation of smoking and drinking and the risk of laryngeal cancer. Br J Cancer 87:1227–1229. Baghurst PA, McMichael AJ, Slavotinek AH, et al. 1991. A case-control study of diet and cancer of the pancreas. Am J Epidemiol 134:167–179. Balaram P, Sridhar H, Rajkumar T, Vaccarella S, Herrero R, Nandakumar A, Ravichandran K, Ramdas K, Sankaranarayanan R, Gajalakshmi V, Munoz N, Franceschi S. 2002. Oral cancer in southern India: the influence of smoking, drinking, paan-chewing and oral hygiene. Int J Cancer 98:440–445. Bandera EV, Freudenheim J, Marshall J, Zielezny M, Priore RL, Brasure J, Baptiste MS, Graham S. 1997. Diet and alcohol consumption and lung cancer risk in the New York State cohort (United States). Cancer Causes Control 8:828–840. Baron JA, Sandler RS, Haile RW, Mandel JS, Mott LA, Greenberg ER. 1998. Folate intake, alcohol consumption, cigarette smoking, and risk of colorectal adenomas. J Natl Cancer Inst 90:57–62. Barrett-Connor E, Garland C, McPhillips JB, Khaw KT, Wingard DL. 1990. A prospective, population-based study of androstenedione, estrogens, and prostatic cancer. Cancer Res 50:169–173. Baumgartner KB, Annegers JF, McPherson RS, Frankowski RF, Gilliland FD, Samet JM. 2002. Is alcohol intake associated with breast cancer in hispanic women? The New Mexico Women’s Health Study. Ethn Dis 12:460–469. Blot WJ, McLaughlin JK, Winn DM, et al. 1988. Smoking and drinking in relation to oral and pharyngeal cancer. Cancer Res 48:3282–3287. Blot WJ, Winn DM, Fraumeni JF. 1983. Oral cancer and mouthwash. J Natl Cancer Inst 70:251–253.

252

PART III: THE CAUSES OF CANCER

Boeing H, Frentzel-Beyme R, Berger M, et al. 1991. Case-control study on stomach cancer in Germany. Int J Cancer 47:858–864. Boffetta P, Ye W, Adami HO, Mucci LA, Nyren O. 2001. Risk of cancers of the lung, head and neck in patients hospitalized for alcoholism in Sweden. Br J Cancer 85:678–682. Boice JD, Mandel JS, Doody MM. 1995. Breast cancer among radiologic technologists. JAMA 274:394–401. Bollschweiler E, Wolfgarten E, Nowroth T, Rosendahl U, Monig SP, Holscher AH. 2002. Vitamin intake and risk of subtypes of esophageal cancer in Germany. J Cancer Res Clin Oncol 128:575–580. Boonyaphiphat P, Thongsuksai P, Sriplung H, Puttawibul P. 2002. Lifestyle habits and genetic susceptibility and the risk of esophageal cancer in the Thai population. Cancer Lett 186:193–199. Borras E, Coutelle C, Rosell A, Fernandez-Muizi F, Broch M, Crosas B, Hjelmqvist L, Lorenzo A, Gutierrez C, Santos M, Szczepanik M, Heilig M, Quattrocchi P, Farres J, Vidal F, Richart C, Mach T, Bogdal J, Jorncall H, Seitz H, Couzigou P, Pares X. 2000. Genetic polymorphism of alcohol dehydrogenase in Europeans: the ADH2*2 allele decreases the risk for alcoholism and is associated with ADH3*1. Hepatology 31:984–989. Bosetti C, Gallus S, Franceschi S, Levi F, Bertuzzi M, Negri E, Talamini, La Vecchia C. 2002. Cancer of the larynx in non-smoking alcohol drinkers and in non-drinking tobacco smokers. Br J Cancer 87:516– 518. Bosron WF, Li TK. 1986. Genetic polymorphisms of human liver alcohol and aldehyde dehydrogenases, and their relationship to alcohol metabolism and alcoholism. Hepatology 6:502–510. Bouchardy C, Clavel F, La Vecchia C, Raymond L, Boyle P. 1990. Alcohol, beer and cancer of the pancreas. Int J Cancer 45:842–846. Boutron MC, Faivre J, Dop MC, Quipourt V, Senesse P. 1995. Tobacco, alcohol, and colorectal tumors: a multistep process. Am J Epidemiol 141:1038–1046. Bowlin SJ, Leske MC, Varma A, Nasca P, Weinstein A, Caplan L. 1997. Breast cancer risk and alcohol consumption: results from a large case-control study. Int J Epidemiol 26:915–923. Bradshaw E, Schonland M. 1974. Smoking, drinking, and oesophageal cancer in African males of Johannesburg, South Africa. Br J Cancer 30:157–163. Braga C, La Vecchia C, Negri E, Franceschi S. 1997. Attributable risks for hepatocellular carcinoma in northern Italy. Eur J Cancer 33:629–634. Breslow RA, Wideroff L, Graubard BI, Erwin D, Reichman ME, Ziegler RG, Ballard-Barbash R. 1999. Alcohol and prostate cancer in the NHANES I epidemiological follow-up study. Ann Epidemiol 9:254–261. Britton JA, Gammon MD, Schoenberg JB, Stanford JL, Coates RJ, Swanson CA, Potischman N, Malone KE, Brogan DJ, Daling JR, Brinton LA. 2002. Risk of breast cancer classifed by joint estrogen receptor and progesterone receptor status among women 20–44 years of age. Am J Epidemiol 156:507–516. Bross IDJ, Coombs J. 1976. Early onset of oral cancer among women who drink and smoke. Oncology 33:136–139. Brown LM, Hoover R, Silverman D, Baris D, Hayes R, Swanson GM, Schoenberg J, Greenberg R, Liff J, Schwartz A, Dosemeci M, Pottern L, Fraumeni JF. 2001. Excess incidence of squamous cell esophageal cancer among US black men: role of social class and other risk factors. Am J Epidemiol 153:114–122. Bueno De Mesquita HB, Maisonneuve P, Moerman CJ, Runia S, Boyle P. 1992. Lifetime consumption of alcoholic beverages, tea and coffee and exocrine carcinoma of the pancreas: a population-based case-control study in The Neatherlands. Int J Cancer 50:514–522. Bulatao-Jayme J, Almero E, Castro MA, et al. 1982. A case-control dietary study of primary liver cancer risk from aflatoxin exposure. Int J Epidemiol 11:112–119. Burch JD, Howe GR, Miller AB, et al. 1981. Tobacco, alcohol, asbestos, and nickel in the etiology of cancer of the larynx: a case-control study. J Natl Cancer Inst 76:1219–1224. Byers T, Funch DF. 1982. Alcohol and breast cancer. Lancet 1:799–800. Cade J, Thomas E, Vail A. 1998. Case-control study of breast cancer in south east England: nutritional factors. J Epidemiol Community Health 52:105–110. Camargo CA Jr, Hennekens CH, Gaziano JM, Glynn RJ, Manson JE, Stampfer MJ. 1997. Prospective study of moderate alcohol consumption and mortality in US male physicians (original investigation). Arch Intern Med 157:79–85. Carpenter CL, Morgenstern H, London SJ. 1998. Alcoholic beverage consumption and lung cancer risk among residents of Los Angeles County. J Nutr 128:694–700. Carr LG, Foroud T, Stewart T, Castelluccio P, Edenberg HJ. 2002. Influence of ADH1B polymorphism on alcohol use and it subjective effects in a Jewish population. Am J Med Genet 112:138–143.

Carstensen JM, Bygren LO, Hatschek T. 1990. Cancer incidence among Swedish brewery workers. Int J Cancer 45:393–396. Castellsague X, Munoz N, De Stefani E, Victora CG, Castelletto R, Rolon PA, Quintana MJ. 1999. Independent and joint effects of tobacco smoking and alcohol drinking on the risk of esophageal cancer in men and women. Int J Cancer 82:657–664. Chen MJ, Chiou YY, Wu DC, Wu SL. 2000. Lifestyle habits and gastric cancer in a hospital-based case-control study in Taiwan. Am J Gastroenterol 95:3242–3249. Chen WJ, Chen C-C, Yu J-M, Cheng ATA. 1998. Self-reported flushing and genotypes of ALDH2, ADH2, and ADH3 among Taiwanese Han. Alcohol Clin Exp Res 22:1048–1052. Chen WJ, Loh EW, Hsu Y-P, Cheng ATA, Chen C-C, Yu J-M. 1996. Alcohol metabolizing genes and alcoholism among Taiwanese Han men; independent effect of ADH2, ADH3, and ALDH2. Br J Psychiatry 168:762–767. Chen WY, Colditz GA, Rosner B, Hankinson SE, Hunter DJ, Manson JE, Stampfer MJ, Willett WC, Speizer FE. 2002. Use of postemenopausal hormones, alcohol, and risk for invasive breast cancer. Ann Intern Med 137:798–804. Cheng KK, Day NE, Duffy SW, Lam TH, Fox M, Wong J. 1992. Pickled vegetables in the aetiology of oesophagel cancer in Hong Kong Chinese. Lancet 339:1314–1318. Cheng KK, Duffy SW, Day NE, Lam TH, Chung SF, Badrinath P. 1995. Stopping drinking and risk of oesophageal cancer. BMJ 310:1094– 1097. Choi SY, Kahyo H, et al. 1991. Effect of cigarette smoking and alcohol consumption in the aetiology of cancer of the oral cavity, pharynx and larynx. Int J Epidemiol 20:878–885. Chow WH, Swanson CA, Lissowska J, Groves FD, Sobin LH, NasierowskaGuttmejer A, Radziszewski J, Regula J, Hsing AW, Jagannatha S, Zatonski W, Blot WJ. 1999. Risk of stomach cancer in relation to consumption of cigarettes, alcohol, tea and coffee in Warsaw, Poland. Int J Cancer 81:871–876. Chu SY, Lee NC, Wingo PA, Webster LA. 1989. Alcohol consumption and the risk of breast cancer. Am J Epidemiol 130:867–877. Chyou PH, Nomura A, Stemmermann GN. 1995. Diet, alcohol, smoking and cancer of the upper aerodigestive tract: a prospective study of among Hawaii Japanese men. Int J Cancer 60:616–621. Chyou PH, P Nomura AMY, Stemmermann GN. 1996. A prospective study of colon and rectal cancer among Hawaii Japanese men. Ann Epidemiol 6:276–282. Clavel F, Benhamou E, Auguier A, Tararyre M, Flamant R. 1989. Coffee, alcohol, smoking, and cancer of the pancreas: a case-control study. Int J Cancer 43:17–21. Collaborative Group on Hormonal Factors in Breast Cancer. 2002. Alcohol, tobacco and breast cancer: collaborative reanalysis of individual data from 53 epidemiological studies, including 58,515 women with breast cancer and 95,067 women without the disease. Br J Cancer 87: 1234–1245. Cornfield J. 1951. A method of estimating comparative rates from clinical data: applications to cancer of the lung, breast and cervix. J Natl Cancer Inst 11:1269–1275. Cornfield J. 1956. A statistical problem arising from retrospective studies. In: Neyman, J. editor. Proceedings of the Third Berkeley Symposium. IV. Berkeley: University of California Press, pp. 135–148. Corrao G, Bagnardi V, Zambon A, Arico S. 1999. Exploring the dose-response relationship between alcohol consumption and the risk of several alcoholrelated conditions: a meta-analysis. Addiction 94:1551–1573. Coutelle C, Ward PJ, Fleury B, Quattrocchi P, Chambrin H, Iron A, Couzigou P, Cassaigne A. 1997. Laryngeal and oropharyngeal cancer, and alcohol dehydrogenase 3 and glutathione S-transferase M1 polymorphisms. Hum Genet 99:319–325. Crispo A, Talamini R, Gallus S, Negri E, Gallow A, Bossett C, LaVecchia C, Dal Haso L, Montella M. 2004. Alcohol and the risk of prostate cancer and benign prostatic hypertrophy. Urology 64:717–722. Cuzick J, Babiker AG. 1989. Pancreatic cancer, alcohol, diabetes mellitus and gallbladder disease. Int J Cancer 43:415–421. Dal Maso L, La Vecchia C, Polesel J, Talamini R, Levi F, Conti E, Zambon P, Negri E, Franceschi S. 2002. Alcohol drinking outside meals and cancers of the upper aero-digestive tract. Int J Cancer 102:435–437. D’Avanzo B, La Vecchia C, Franceschi S. 1994. Alcohol consumption and the risk of gastric cancer. Nutr Cancer 22:57–64. De Jong UW, Breslow N, Goh Ewe Hong J, Sridharan M, Shanmugaratnam K. 1974. Aetiological factors in oesophageal cancer in Singapore Chinese. Int J Cancer 13:291–303. De Stefani E, Boffetta P, Carzoglio J, Mendilaharsu S, Deneo-Pellegini H. 1998. Tobacco smoking and alcohol drinking as risk factors for stomach

Alcohol cancer: a case-control study in Uruguay. Cancer Causes Control 9:321–329. De Stefani E, Correa P, et al. 1987. Risk factors for laryngeal cancer. Cancer 60:3087–3091. De Stefani E, Correa P, Fierro L, Fontham ET, Chen V, Zavala D. 1993. The effect of alcohol on the risk of lung cancer in Uruguay. Cancer Epidemiol Biomarkers Prev 2:21–26. De Stefani E, Fierro L, Correa P, Fontham ET,, Ronco A, Larringaga M, Balbi J, Medilaharsu M. 1996. Mate drinking and risk of lung cancer in males: a case-control study from Uruguay. Cancer Epidemiol Biomarkers Prev 5:515–519. De Stefani E, Munoz N, et al. 1990. Mate drinking, alcohol, tobacco, diet and esophageal cancer in Uruguay. Cancer Res 50:426–431. Dean G, MacLennan R, et al. 1979. Causes of death of blue-collar workers at a Dublin brewery, 1954–1973. Br J Cancer 40:581–589. Dennis LK. 2000. Meta-analysis for combining relative risks of alcohol consumption and prostate cancer. Prostate 42:56–66. Dennis LK, Hayes RB. 2001. Alcohol and prostate cancer. Epidemiol Rev 23:110–114. Dhillon PK, Farrow DC, Vaughan TL, Chow WH, Risch HA, Gammon MD, Mayne ST, Stanford JL, Schoenberg JB, Ahsan H, Dubrow R, West AB, Rotterdam H, Blot WJ, Fraumeni JF. 2001. Family history of cancer and risk of esophageal and gastric cancers in the United States. Int J Cancer 93:148–152. Doll R, Peto R, Hall E, Wheatley K, Gray R. 1994. Mortality in relation to consumption of alcohol: 13 years’ observations on male British doctors. BMJ 309:911–918. Donato F, Tagger A, Gelatti U, Parrinello G, Boffetta P, Albertini A, Decarli A, Trevisi P, Ribero ML, Martelli C, Porru S, Nardi G. 2002. Alcohol and hepatocellular carcinoma: the effect of lifetime intake and hepatitis virus infections in men and women. Am J Epidemiol 155:323–331. Dorgan JF, Reichman ME, Judd JT, Brown CC, Longscope C, Schatzkin A, Campbell WS, Franz C, Kahle L, Taylor PR. 1994. The relation of reported alcohol ingestion to plasma levels of estrogens and androgens in premenopausal women (Maryland, U.S.A.). Cancer Causes Control 5:53–60. Dosemeci M, Gokmen I, Unsal M, Hayes RB, Blair A. 1997. Tobacco, alcohol use, and risks of laryngeal and lung cancer by subsite and histologic type in Turkey. Cancer Causes Control 8:729–737. Duester G, Farres J, Felder MR, Holmes RS, Hoog JO, Pares X, Plapp BV, Yin SJ, Jornvall H. 1999. Recommended nomenclature for the verte-brate alcohol dehydrogenase gene family. Biochem Pharmacol 58:389–395. Edenberg HJ. 2000. Regulation of the mammalian alcohol dehydrogenase genes. Prog Nucl Acids Res Mol Biol 64:295–341. Edenberg HJ, Jerome RE, Li M. 1999. Polymorphism of the human alcohol dehydrogenase 4 (ADH4) promoter affects gene expression. Pharmacogenetics 9:25–30. Ehlers CL, Gilder DA, Harris L, Carr LG. 2001. Association of the ADH2*3 allele with a negative family history of alcoholism in African American young adults. Alcohol Clin Exp Res 25:1773–1777. Elwood JM, Person JCG, Skippen DH, et al. 1984. Alcohol, smoking, and socioeconomic factors in the etiology of cancer of the oral cavity, pharynx, and larynx. Int J Cancer 34:603–612. Enger SM, Ross RK, Paganani-Hill A, Longnecker MP, Bernstein L. 1999. Alcohol consumption and breast cancer oestrogen and progesterone receptor status. Br J Cancer 79:1308–1314. Eriksson CJP. 2001. The role of acetaldehyde in the actions of alcohol (update 2000). Alcohol Clin Exp Res 25 (5 Suppl ISBRA):15s–32s. Ewertz M. 1991. Alcohol consumption and breast cancer risk in Denmark. Cancer Causes Control 2:247–252. Falk RT, Pickle LW, Brown LM, et al. 1989. Effect of smoking and alcohol consumption on laryngeal cancer risk in coastal Texas. Cancer Res 49:4024–4029. Farrow DC, Davis S. 1990. Risk of pancreatic cancer in relation to medical history and the use of tobacco, alcohol and coffee. Int J Cancer 45:816–820. Feigelson HS, Calle EE, Robertson AS, Wingo PA, Thun MJ. 2001. Alcohol consumption increases the risk of fatal breast cancer (United States). Cancer Causes Control 12:895–902. Ferraroni M, Negri D, La Vecchia C, et al. 1989. Socioeconomic indicators, tobacco and alcohol in the aetiology of digestive tract neoplasms. Int J Epidemiol 18:556–562. Flood A, Caprario L, Chaterjee N, Lacey JV, Schairer C, Schatzkin A. 2002. Folate, methionine, alcohol, and colorectal cancer in a prospective study of women in the United States. Cancer Causes Control 13:551–561. Franceschi S, Barra S, La Vecchia C, et al. 1992. Risk factors for cancer of the tongue and the mouth: a case-control study from northern Italy. Cancer 70:2227–2233.

253

Franceschi S, Biddi E, Negri E, et al. 1994. Alcohol and cancers of the upper aerodigestive tract in men and women. Cancer Epidemiol Biomarkers Prev 3:299–304. Franceschi S, Levi F, La Vecchia C, Conti E, Dal Maos L, Barzan L, Talamini R. 1999. Comparison of the effect of smoking and alcohol drinking between oral and pharyngeal cancer. Int J Cancer 83:1–4. Franceschi S, Talamini R, Barra S, et al. 1990. Smoking and drinking in relation to cancers of the oral cavity, pharynx, larynx and esophagus in northern Italy. Cancer Res 50:6502–6507. Freudenheim J, Marshall J. 1988. The problem of profound mismeasurement and the power of epidemiologic studies of diet and cancer. Nutr Cancer 11:243–250. Freudenheim J, Graham S, Marshall JR, Haughey BP, Cholewinski S, Wilkinson G. 1991. Folate intake and carcinogenesis of the colon and rectum. Int J Epidemiol 20:368–374. Freudenheim JL, Ambrosone CB, Moysich KB, Vena JE, Graham S, Marshall JR, Muti P, Laughlin R, Nemoto T, Harty LC, Crits GA, Shields P. 1999. Alcohol dehydrogenase 3 genotype modification of the association of alcohol consumption with breast cancer. Cancer Causes Control 10:369–377. Freudenheim JL, Graham S, Byers TE, et al. 1992. Diet, smoking, and alcohol in cancer of the larynx: a case-control study. Nutr Cancer 17:33–45. Friedenreich CM, How GR, Miller AB, Jain MG. 1993. A cohort study of alcohol consumption and risk of breast cancer. Am J Epidemiol 137:512–520. Friedman GD, Van Den Eden SK. 1993. Risk factors for pancreatic cancer: an exploratory study. Int J Epidemiol 22:30–37. Fuchs CS, Giovannucci EL, Colditz GA, Hunter DJ, Speizer FE, Willett WC. 1994. A prospective study of family history and the risk of colorectal cancer. N Engl J Med 331:1669–1674. Galanis DJ, Kolonel LN, Lee J, Nomura A. 1998. Intakes of selected foods and beverages and the incidence of gastric cancer among the Japanese residents of Hawaii: a prospective study. Int J Epidemiol 27:173– 180. Gallus S, Bosetti C, Franceschi S, Levi F, Simonato L, Negri E, La Vecchia C. 2001. Oesophageal cancer in women: tobacco, alcohol, nutritional and hormonal factors. Bri J Cancer 85:341–345. Gann PH. 1997. Interpreting recent trends in prostate cancer incidence and mortality. Epidemiology 8:117–120. Gao YT, McLaughlin JK, Blot WJ, et al. 1994. Risk factors for esophageal cancer in Shanghai, China. I. Role of cigarette smoking and alcohol drinking. Int J Cancer 58:192–196. Gapstur SM, Potter JD, Folsom AR. 1994. Alcohol consumption and colon and rectal cancer in postmenopausal women. Int J Epidemiol 23:50–57. Gapstur SM, Potter JD, Sellers TA, Folsom AR. 1992. Increased risk of breast cancer with alcohol consumption in postmenopausal women. Am J Epidemiol 136:1221–1231. Garfinkel L, Boffetta P, Stellman SD. 1988. Alcohol and breast cancer: a cohort study. Prev Med 17:686–693. Garrote LM, Herrero R, Reyes RM, Vaccarella S, Anta JF, Ferbeye L, Munoz N, Franceschi S. 2001. Risk factors for cancer of the oral cavity and oropharynx in Cuba. Br J Cancer 85:46–54. Ghadirian P, Simard A, Baillargeon J. 1991. Tobacco, alcohol and coffee and cancer of the pancreas. Cancer 67:2664–2670. Ginsburg ES. 1999/Estrogen, alcohol and breast cancer risk. J Steroid Biochem Mol Biol 69:299–306. Giovannucci E, Stampfer MJ, Colditz GA, et al. 1993. Folate, methionine, and alcohol intake and risk of colorectal adenoma. J NatL Cancer Inst 85:875–884. Giovannucci E, Ascherio A, Rimm EB, Colditz GA, Stampfer MJ, Willett WC. 1995. Physical activity, obesity, and risk for colon cancer and adenoma in men. Ann Intern Med 122:327–334. Giovannucci E, Rimm EB, Ascherio A, Stampfer MJ, Colditz GA, Willett WC. 1995. Alcohol, low-methionine–low-folate diets, and risk of colon cancer in men. J Natl Cancer Inst 87:265–273. Glynn SA, Albanes D, Pietinen P, Brown CC, Rautalahti M, Tangrea JA, Taylor PR, Virtamo J. 1996. Alcohol consumption and risk of colorectal cancer in a cohort of Finnish men. Cancer Causes Control 7:214–223. Gordon T, Kannel WB. 1984. Drinking and mortality: the Framingham Study. Am J Epidemiol 120:97–107. Graham S, Dagal H, Swanson M, Mittelman A, Wilkinson G. 1978. Diet in the epidemiology of cancer of the colon and rectum. J Natl Cancer Inst 61:709–714. Graham S, Dayal H, Rohrer T, Swanson M, Sultz H, Shedd D, Fischman S. 1977. Dentition, diet, tobacco, and alcohol in the epidemiology of oral cancer. J Natl Cancer Inst 59:1611–1618. Graham S, Mettlin C, Marshall J, et al. 1981. Dietary factors in the epidemiology of cancer of the larynx. Am J Epidemiol 113:675–680.

254

PART III: THE CAUSES OF CANCER

Gregorio DI, Marshall JR, Zielezny M. 1985. Fluctuations in odds ratios due to variance differences in case-control studies. Am J Epidemiol 121:767–774. Gronbaek M, Becker U, Johansen D, Tonnesen H, Jensen G, Sorensen TIA. 1998. Population based cohort study of the association between alcohol intake and cancer of the upper digestive tract. BMJ 317:844–847. Haenszel W, Kurihara M, Segi M, et al. 1972. Stomach cancer among Japanese in Hawaii. J Natl Cancer Inst 49:969–988. Hakulinen T, Lehtimaki L, Lehtonen M, Teppo L. 1974. Cancer morbidity among two male cohorts with increased alcohol consumption in Finland. J Natl Cancer Inst 52:1711–1714. Hansson LE, Baron J, Nyren O, Bergstrom R, Wolk A, Adami HO. 1994. Tobacco, alcohol and the risk of gastric cancer: a population-based casecontrol study in Sweden. Int J Cancer 57:26–31. Hardell L, Bengtsson NO, Jonsson U, et al. 1984. Aetiological aspects of primary liver cancer, with special regard to alcohol, organic solvents and acute intermittent prophyria: an epidemiological investigation. Br J Cancer 50:389–397. Harnack LJ, Anderson KE, Zheng W, Folsom AR, Sellers TA, Kushi LH. 1997. Smoking, alcohol, coffee, and tea intake and incidence of cancer of the exocrine pancreas: the Iowa Women’s Health Study. Cancer Epidemiol Biomarkers Prev 6:1081–1086. Harris RE, Wynder EL. 1988. Breast cancer and alcohol consumption: a study in weak associations. JAMA 259:2867–2871. Harty LC, Caporaso NE, Hayes RB, Winn DM, Bravo-Otero E, Blot WJ, Kleinman DV, Brown LM, Armenian HK, Fraumeni JF, Shields PG. 1997. Alcohol dehydrogenase 3 genotype and risk of oral cavity and pharyngeal cancers. J Natl Cancer Inst 89:1698–1705. Harvey EB, Schairer C, Brinton LA. 1987. Alcohol consumption and breast cancer. JAMA 78:657–661. Hasin D, Aharonovich E, Liu X, Mamman Z, Matseoane K, Carr LG, Li TK. 2002. Alcohol dependence symptoms and alcohol dehydrogenase 2 polymorphism: Israeli Ashkenazis, Sephardics, and recent Russian immigrants. Alcohol Clin Exp Res 26:1315–1321. Hayes RB, Morris Brown L, Schoenberg J, Greenberg RS, Silverman DT, Schwartz AG, Swanson GM, Benichou J, Liff J, Hoover RN, Pottern LM. 1996. Alcohol use and prostate cancer risk in US blacks and whites. Am J Epidemiol 143:692–697. Herbert V, Kshitish CD. 1994. Folic Acid and Vitamin B12. In: Shils ME, Olson JA, Shike M, editors. Modern Nutrition in Health and Disease. Phildelphia: Lea & Febiger, pp. 402–425. Heuch I, Kvole G, Jacobsen BK, et al. 1983. Use of alcohol, tobacco and coffee, and risk of pancreatic cancer. Br J Cancer 48:637–643. Hiatt RA, Bawol RD. 1984. Alcoholic beverage consumption and breast cancer incidence. Am J Epidemiol 120:676–683. Hiatt RA, Armstrong MA, Klatsky AL, Sidney S. 1994. Alcohol consumption, smoking, and other risk factors, and prostate cancer in a large health plan cohort in California (United States). Cancer Causes Control 5:66–72. Hiatt RA, Klatsky AL, Armstrong MA. 1988. Alcohol consumption and the risk of breast cancer in a prepaid health plan. Cancer Res 48:2284–2287. Hines LM, Hankinson SE, Smith-Warner SA, Spiegelman D, Kelsey JL, Colditz GA, Willet WC, Hunter DJ. 2000. A prospective study of the effect of alcohol consumption and ADH3 genotype on plasma steroid hormone levels and breast cancer risk. Cancer Epidemiol Biomarkers Prev 9:1099–1105. Hirayama T. 1989. Association between alcohol consumption and cancer of the sigmoid colon: observations from a Japanese cohort study. Lancet 2:725–727. Hirayama T. 1990. Contribution of a long-term prospective cohort study to the issue of nutrition and cancer, with special reference to the role of alcohol drinking. In: Crawford B, editor. Recent Progress in Research on Nutrition and Cancer. New York: Wiley-Liss, pp. 179–187. Hirose K, Tajima K, Hamajima N, Inoue M, Takezaki T, Kuroishi T, Yoshida M, Tokudome S. 1995. A large-scale, hospital-based case-control study of risk factors of breast cancer according to menopausal status. Jpn J Cancer Res 86:146–154. Holmberg L, Ohlander EM, Byers T, et al. 1994a. Diet and breast cancer risk: results from a population-based, case-control study in Sweden. Arch Intern Med 154:1805–1811. Holmberg L, Ohlander EM, Byers T, Zack M, Wolk A, Bruce A, Bergstrom R, Bergkvist L, Adami HO. 1996. A search for recall bias in a case-control study of diet and breast cancer. Int J Epidemiol 25:235–244. Holst PA, Kromhut D, Brand R. 1988. For debate: pet birds as an independent risk factor for lung cancer. BMJ 297:1319–1321. Horn-Ross PL, Hoggatt KJ, West DW, Krone MR, Stewart SL, Anton H, Bernstei CL, Deapen D, Peel D, Pinder R, Reynolds P, Ross RK, Wright W, Ziogas A. 2002. Recent diet and breast cancer risk: the California teachers study (USA). Cancer Causes Control 13:407–415.

Howe G, Hirohata T, Hislop GT, et al. 1990. Dietary factors and risk of breast cancer: combined analysis of 12 case-control studies. J Natl Cancer Inst 82:561–569. Howe G, Rohan T, De Carli A, et al. 1991. The association between alcohol and breast cancer risk: evidence from the combined analysis of six dietary case-control studies. Int J Cancer 47:707–710. Hsing AW, McLaughlin JK, Chow WH, Schuman LM, Co Chien HT, Gridley G, Bjelke E, Wacholder S, Blot WJ. 1998. Risk factors for colorectal cancer in a prospective study among U.S. white men. Int J Cancer 77:549–553. Hu J, Nyren O, Wolk A, et al. 1994. Risk factors for oesophageal cancer in northeast China. Int J Cancer 57:38–46. Hu JF, Zhang SF, Jia EM, et al. 1988. Diet and cancer of the stomach: a casecontrol study in China. Int J Cancer 41:331–335. Inaba Y, Maruchi N, Matsuda M, et al. 1984. A case-control study on liver cancer with special emphasis on the possible aetiological role of schistosomiasis. Int J Epidemiol 13:408–412. Isaksson B, Jonsson F, Pedersen NL, Larsson J, Feychting M, Permert J. 2002. Lifestyle factors and pancratic cancer risk: a cohort study from the Swedish Twin Registry. Int J Cancer 98:480–482. Iscovich JM, Iscovich RB, Howe G, Shiboski S, Kaldor JM. 1989. A casecontrol study of diet and breast cancer in Argentina. Int J Cancer 44:770–776. Jain M, Hislop GT, Howe G, Burch JD, Ghadirian P. 1998. Alcohol and other beverage use and prostate cancer risk among Canadian men. Int J Cancer 78:707–711. Jain MG, Hislop GT, Howe GR, Ghadirian P. 1999. Plant foods, antioxidants, and prostate cacer risk: findings from case-control studies in Canada. Nutr Cancer 34:173–184. Jain M, Howe GR, St Louis P, et al. 1991. Coffee and alcohol as determinants of risk of pancreas cancer: a case-control study from Toronto. Int J Cancer 47:384–389. Jedrychowski W, Boeing H, Wahrendorf J, Popiela T, Tobiasz Adamczyk TB, Kulig J. 1993. Vodka consumption, tobacco smoking and risk of gastric cancer in Poland. Int J Epidemiol 22:606–613. Jedrychowski W, Steindorf K, Popiela T, Wahrendorf J, Tobiasz Adamczyk TB, Kulig J, Penar Q. 2001. Risk of colorectal cancer from alcohol consumption at lower vitamin intakes. a hospital-based case-control study in Poland. Rev Environ Health 16:213–222. Jensen OM. 1979. Cancer morbidity and causes of death among Danish brewery workers. Int J Cancer 23:454–463. Jensen OM. 1983. Cancer risk among Danish male Seventh-Day Adventists and other Temperance Society members. J Natl Cancer Inst 70:1011–1014. Jensen OM, Tuyns AJ, Pequignot G. 1978. Usefulness of population controls in retrospective studies of alcohol consumption: experience from a casecontrol study of esophageal cancer in Ille-et-Vilaine, France. Q J Stud Alcohol 39:175–182. Ji BT, Chow WH, Yang G, McLaughlin JK, Gao R, Zheng W, Shu XO, Jin F, Fraumeni JF, Gao YT. 1996. The influence of cigarette smoking, alcohol, and green tea consumption on the risk of carcinoma of the cardia and distal stomach in Shanghai, China. Cancer 77:2449–2457. Kabat G, Wynder EL. 1984. Lung cancer in nonsmokers. Cancer 53:1214–1221. Kato I, Miura S, Kasumi F, Iwase T, Tashiro H, Fujita Y, Koyama K, Ikeda T, Fujiwara K, Saotome K, Asaishi K, Abe R, Nihei M, Ishida T, Yokoe T, Yamamoto H, Murata M. 1992. A case-control study of breast cancer among Japanese women: with special reference to family history and reproductive and dietary factors. Breast Cancer Res Treat 24:51–59. Katsouyanni K, Trichopoulou A, Stuver S, et al. 1994. Ethanol and breast cancer: an association that may be both confounded and casual. Int J Cancer 58:356–361. Ke L, Yu P, Zhang ZX, Huang SS, Huang G, Ma XH. 2002. Congou tea drinking and oesophageal cancer in south China. Br J Cancer 86:346–347. Kearney J, Giovannucci E, Rimm EB, et al. 1995. Diet, alcohol, and smoking and the occurrence of hyperplastic polyps of the colon and rectum. Cancer Causes Control 6:45–56. Keller AZ, Terris M. 1965. The association of alcohol and tobacco with cancer of the mouth and pharynx. Am J Public Health 55:1578–1585. Kinney AY, Millikan RC, Lin YH, Moorman PG, Newman B. 2000. Alcohol consumption and breast cancer among black and white women in North Carolina (United States). Cancer Causes Control 11:345–357. Kitson KE. 1999. Regulation of alcohol and aldehyde dehydrogenase activity: a metabolic balancing act with important social consequences. Alcohol Clin Exp Res 23:955–957. Kjaerheim K, Gaard M, Andersen A. 1998. The role of alcohol, tobacco, and dietary factors in upper aerogastric tract cancers: a prospective study of 10,900 Norwegian men. Cancer Causes Control 1:99–108.

Alcohol Klatsky AL, Amstrong MA, Friedman GD, Hiatt RA. 1988. The relations of alcohol beverage use to colon and rectal cancer. Am J Epidemiol 128:1007–1015. Klatsky AL, Friedman GD, Sievelaub AB. 1981. Alcohol and mortality: a tenyear Kaiser-Permanente experience. Ann Intern Med 95:139–145. Kono S, Ikeda M. 1979. Correlation between cancer mortality and alcoholic beverage in Japan. Br J Cancer 40:449–455. Kono S, Ikeda M, Ogata M, et al. 1983. The relationship between alcohol and mortality among Japanese physicians. Int J Epidemiol 12:437–441. Kono S, Ikeda M, Tokudome S, et al. 1985. Alcohol and cancer in male Japanese physicians. J Cancer Res Oncol 109:82–85. Kono S, Ikeda M, Tokudome S, Nishizum M, Kuratsune M. 1986. Alcohol and mortality: a cohort study of male Japanese physicians. Int J Epidemiol 15:527–532. Koo LC. 1988. Dietary habits and lung cancer risk among Chinese females in Hong Kong who never smoked. Nutr Cancer 11:155–172. Korte JE, Brennan P, Henley SJ, Boffetta P. 2002. Dose-specific meta-analysis and sensitivity analysis of the relation between alcohol consumption and lung cancer risk. Am J Epidemiol 155:496–506. Kropp S, Becher H, Nieters A, Chang-Claude J. 20001. Low-to-moderate alcohol consumption and breast cancer risk by age 50 years among women in Germany. Am J Epidemiol 154:624–634. Lagergren J, Bergstrom R, Lindgen A, Nyren O. 2000. The role of tobacco, snuff and alcohol use in the aetiology of cancer of the oesophagus and gastric cardia. Int J Cancer 85:340–346. La Vecchia C, Negri E, De Carli A, et al. 1988. Risk factors for hepatocellular carcinoma in northern Italy. Int J Cancer 42:872–876. Le MG, Hill C, Kramar A, et al. 1984. Alcoholic beverage consumption and breast cancer in a French case-control study. Am J Epidemiol 120:350–357. Le MG, Moulton LH, Hill C, et al. 1986. Consumption of dairy produce and alcohol in a case-control study of breast cancer. J Natl Cancer Inst 77:633–636. Lee HH, Wu HY, Chuang YC, et al. 1990. Epidemiologic characteristics and multiple risk factors of stomach cancer in Taiwan. Anticancer Res 10:875–881. Le Marchand L, Wilkens LR, Hankin JH, Kolonel LN, Lyu LC. 1999. Independent and joint effects of family history and lifestyle on colorectal cancer risk: implications for prevention. Cancer Epidemiol Biomarkers Prev 8:45–51. Le Marchand L, Wilkens LR, Kolonel LN, Hankin JH, Lyu LC. 1997. Associations of sedentary lifestyle, obesity, smoking, alcohol use, and diabetes with the risk of colorectal cancer. Cancer Res 57:4787–4794. Lenz S, Goldberg M, Labreche F, Parent ME, Valois, MF. 2002. Association between alcohol consumption and postmenopausal breast cancer: results of a case-control study in Montreal, Quebec, Canada. Cancer Causes Control 13:701–710. Li TK, Yin SJ, Crabb DW, O’Connor S, Ramchandani VA. 2001. Genetic and environomental influences on alcohol metabolism in humans. Alcohol Clin Exp Res 25:136–144. Lieber CS. 1990. The nutritional effects of alcohol. In: Herbert V, SubakSharpe GJ, Hammock DA, editors. The Mount Sinai School of Medicine Complete Book of Nutrition. New York: St. Martin’s Press, pp. 341–357. Lieber CS. 2000. Alcohol: its metabolism and interaction with nutrients. Annu Rev Nutr 20:395–430. Lin Y, Tamakoshi A, Kawamura T, Inaba Y, Kikuchi S, Motohashi Y, Kurosawa M, Ohno Y. 2002. Risk of pancreatic cancer in relation to alcohol drinking, coffee consumption and medical history: findings from the Japan Collaborative Cohort Study for Evaluation of Cancer Risk. Int J Cancer 99:742–746. Longnecker MP. 1994. Alcoholic beverage consumption in relation to risk of breast cancer: meta-analysis and review. Cancer Causes Control 5:73–82. Longnecker MP, Berlin JA, Orza MJ, et al. 1988. A meta-analysis of alcohol consumption in relation to risk of breast cancer. JAMA 260:652–656. Longnecker MP, Newcomb PA, Mittendorf R, Greenberg ER, Clapp RW, Bogdan GF, Baron J, MacMahon B, Willet WC. 1995. Risk of breast cancer in relation to lifetime alcohol consumption. J Natl Cancer Inst 87:923–929. Longnecker MP, Paganani-Hill A, Ross RK. 1995. Lifetime alcohol consumption and breast cancer risk among postmenopausal women in Los Angeles. Cancer Epidemiol Biomarkers Prev 4:721–725. Lu SN, Lin TN, Chen CJ, et al. 1988. A case-control study of primary hepatocellular carcinoma in Taiwan. Cancer 62:2051–2055. Lumey LH, Pittman B, Wynder EL. 1998. Alcohol use and prostate cancer in US whites: no assocation in a confirmatory study. Prostate 36:250–255. Lyon JL, Mahoney AW, French TK, Moser R Jr. 1992. Coffee consumption and the risk of cancer of the exocrine pancreas: a case-control study in a low-risk population. Epidemiology 3:164–170.

255

Ma J, Stampfer MJ, Giovannucci E, Artigas C, Hunter DJ, Fuchs C, Willet WC, Selhub J, Hennekens CH, Rozen R. 1997. Methylenetetrahydrofolate reductase polymorphism, dietary interactions, and risk of colorectal cancer. Cancer Res 57:1098–1102. MacFarlane GJ, Zhen T, Marshall JR, Boffetta P, Niu S, Brasure J, Merletti F, Boyle P. 1995. Alcohol, tobacco, diet and the risk of orak cancer: a pooled analysis of three case-control studies. Eur J Cancer Part B Oral Oncol 31B:181–187. Mack TM, Yu MC, Hanisch R, Henderson BE. 1986. Pancreas, smoking, beverage consumption, and past medical history. J Natl Cancer Inst 76:49–60. Maier H, Gewelke U, Dietz A, Heller WD. 1992. Risk factors of cancer of the larynx: results of the Heidelberg case-control study. Otolaryngol Head Neck Surg 107:577–582. Mannisto S, Pietinen P, Pyy M, Palmgren J, Eskelinen M, Uusitupa M. 1996. Body-size indicators and risk of breast cancer according to menopause and estrogen-receptor status. Int J Cancer 68:8–13. Manus B, Adang RP, Ambergen AW, Bragelmann R, Armbrecht U, Stockbrugger RW. 1997. The risk factor profile of recto-sigmoid adenomas: a prospective screening study of 665 patients in a clinical rehabilitation centre. Eur J Cancer Prev 6:38–43. Marcus PM, Newman B, Millikan RC, Moorman PG, Baird DD, Qaqish B. 2000. The associations of adolescent cigarette smoking, alcoholic beverage consumption, environmental tobacco smoke, and ionizing radiation with subsequent breast cancer risk (United States). Cancer Causes Control 11:271–278. Marshall J, Wood DP Jr. 2002. Prostate. In: Franco EL, Rohan T, editors. Cancer Precursors. New York: Springer-Verlag, pp. 333–343. Marshall J, Hastrup JL. 1996. Mismeasurement and the resonance of strong confounders: uncorrelated error. Am J Epidemiol 143:1069–1078. Marshall J, Hastrup JL, Ross JS. 1999. Mismeasurement and the resonance of strong confounders: correlated errors. Am J Epidemiol 150:88–96. Marshall J, Priore RL, Graham S, Brasure J. 1981. On the distortion of risk estimates in multiple exposure level case control studies. Am J Epidemiol 113:464–473. Marshall J, Priore RL, Haughey BP, Rzepka T, Graham S. 1980. Spouse-subject interviews and the reliability of diet studies. Am J Epidemiol 112:675–683. Marshall JR, Graham S, Haughey BP, et al. 1992. Smoking, alcohol, dentition and diet in the epidemiology of oral cancer. Eur J Cancer Part B Oral Oncol 28:9–15. Martinez I. 1969. Factors associated with cancer of the oesophagus, mouth and pharynx in Puerto Rico. J Natl Cancer Inst 42:1069–1094. Martinez ME, McPherson RS, Annegers JF, Levin B. 1995. Cigarette smoking and alcohol consumption as risk factors for colorectal adenomatous polyps. J Natl Cancer Inst 87:274–279. Mashberg A, Boffetta P, Winkelman R, Garfinkel L. 1993. Tobacco smoking, alcohol drinking, and cancer of the oral cavity and oropharynx among U.S. veterans. Cancer 72:1369–1375. Matsuo K, Hamajima N, Hirai T, Kato T, Koike K, Inoue M, Takezaki T, Tajima K. 2002. Aldehyde dehydrogenase 2 (ALDH2) genotype affects rectal cancer susceptibility due to alcohol consumption. J Epidemiol 12:70–76. Matsuo K, Hamajima N, Shinoda M, Hatooka S, Inoue, M Takezaki T, Tajima K. 2001. Gene–environment interaction between an aldehyde dehydrogenase-2 (aldh2) polymorphism and alcohol consumption for the risk of esophageal cancer. Carcinogenesis 22:913–916. Mayne ST, Navarro SA. 2002. Diet, obesity and reflux in the etiology of adenocarcinomas of the esophagus and gastric cardia in humans. J Nutr 132:3467S–3470S. Mayne ST, Janerich DT, Greenwald P, et al. 1994. Dietary beta-carotene and lung cancer risk in U.S. never smokers. J Natl Cancer Inst 86:33–38. McCann S, Freudenheim JL, Trevisan M, Russell M, Markovic N, Marshall JR. 1999. Recent alcohol intake as estimated by the Health Habits and History Food Frequency Questionnaire, the Harvard Semi-Quantitative Food Frequency Questionnaire and a more detailed Alcohol Intake Questionnaire. Am J Epidemiol 150:334–340. McCarver DG. 2001. ADH2 and CYP2E1 genetic polymorphisms: risk factors for alcohol-related birth defects. Drug Metab Dispos 29:562–565. McCarver DG, Byun R, Hines RN, Hichme M, Wegenek W. 1998. A genetic polymorphism in the regulatory sequences of human CYP2E1: assocaiton with increased chlorzoxazone hydroxylation in the prsence of obesity and ethanol intake. Toxicol Appl Pharmacol 152:276–281. Meara J, McPherson K, Roberts M, et al. 1989. Alcohol, cigarette smoking and breast cancer. Br J Cancer 60:70–73. Mezzetti M, La Vecchia C, Decarli A, Boyle P, Talamini R, Franceschi S. 1998. Population attributable risk for breast cancer: diet, nutrition, and physical exercise. J Natl Cancer Inst 90:389–394. Michaud DS, Giovannucci E, Willet WC, Colditz GA, Fuchs CS. 2001. Coffee and alcohol consumption and the risk of pancreatic cancer in two

256

PART III: THE CAUSES OF CANCER

prospective United States cohorts. Cancer Epidemiol Biomarkers Prev 10:429–437. Miettinen OS. 1976. Estimability and estimation in case-referent studies. Am J Epidemiol 103:226–235. Miettinen OS, Cook EF. 1981. Confounding: essence and detection. Am J Epidemiol 114:593–603. Miller AB, Baines CJ, To T, Wall C. 1992. Canadian National Breast Screening Study. 1. Breast cancer detection and death rates among women aged 40 to 49 years. 2. Breast cancer detection and death rates among women aged 50 to 59 years. Can Med Assoc J 147:1459–1488. Miller AB, Kelly A, Choi NW, Matthews V, Morgan RW, Munan L, Burch JD, Feather J, Howe GR, Jain M. 1978. A study of diet and breast cancer. Am J Epidemiol 107:499–509. Mizuno S, Watanabe S, Nakamura K, et al. 1992. A multi-institute case-control study on the risk factors of developing pancreatic cancer. Jpn J Clin Oncol 22:286–291. Morabia A, Wynder EL. 1990. Epidemiology and natural history of breast cancer: implications for the body weight-breast cancer controversy. Surg Clin North Am 70:739–752. Mori M, Hara M, Wada I, Hara T, Yamamoto K, Honda M, Naramoto J. 2000. Prospective study of hepatitis B and C viral infections, cigarette smoking, alcohol consumption, and other factors associated with hepatocellular carcinoma risk in Japan. Am J Epidemiol 151:131–139. Munoz N, Plummer M, Vivas J, Moreno V, De Sanjose S, Lopez G, Oliver W. 2001. A case-control study of gastric cancer in Venezuela. Int J Cancer 93:417–423. Muscat JE, Wynder EL. 1992. Tobacco, alcohol, asbestos, and occupational risk factors for laryngeal cancer. Cancer 69:2244–2251. Muti P, Trevisan M, Micheli A, Krogh V, Bolelli G, Sciajno R, Schunemann HJ, Berrino F. 1998. Alcohol consumption and total estradiol in premenopausal women. Cancer Epidemiol Biomarkers Prev 7:189–193. Notani PN. 1988. Role of alcohol in cancers of the upper alimentary tract: use of models in risk assessment. J Epidemiol Community Health 42:187–192. Olsen J, Sabroe S, Fasting U. 1985. Interaction of alcohol and tobacco as risk factors in cancer of the laryngeal region. J Epidemiol Community Health 39:165–168. Olsen J, Sabroe S, Ipsen J. 1985. Effect of combined alcohol and tobacco exposure on risk of cancer of the hypopharynx. J Epidemiol Community Health 39:304–307. Omenn GS, Goodman GE, Thornquist MD, Balmes J, Cullen MR, Glass A, Keogh JP, Meyskens FL Jr, Valanis B, Williams JH Jr, Barnhart S, Cherniack MG, Brodkin CA, Hammar S. 1996. Risk factors for lung cancer and for intervention effects in CARET, the Beta-Carotene and Retinol Efficacy Trial. J Natl Cancer Inst 88:1550–1559. Onland-Moret NC, Peeters PH, van der Schouw YT, Grobbee DE, van Gils CH. 2005. Alcohol and endogenous sex steroid levels in postmenopausal women: a cross-sectional study. J Clin Endocrinol Metab 90:1414–1419. Oshima A, Tsukuma H, Hiyama T, et al. 1984. Follow-up study of HBsAgpositive blood donors with special reference to effect of drinking and smoking on development of liver cancer. Int J Cancer 34:775–779. Partanen TJ, Vainio HU, Ojajarvi IA, Kauppinen TP. 1997. Pancreas cancer, tobacco smoking and consumption of alcoholic beverages: a case-control study. Cancer Lett 116:27–32. Pera M, Pera M. 2001. Recent changes in the epidemiology of esophageal cancer. Surgi Oncol 10:81–90. Petri AL, Tjonneland A, Gamborg M, Johansen D, Hoidrup S, Sorensen TJ, Gronbaek M. 2004. Alcoholic intake, type of beverage; and risk of breast cancer in pre- and post-menopausal women. Alcohol Clin Exp Res 28:1084–1090. Phukan RK, Ali MS, Chetia CK, Mahanta J. 2001. Betel nut and tobacco chewing; potential risk factors of cancer of oesophagus in Assam, India. Br J Cancer 85:661–667. Pollack ES, Nomura A, Heibrun LK, et al. 1984. Prospective study of alcohol consumption and cancer. N Engl J Med 310:617–621. Potter JD, Sellers TA, Folsom AR, McGovern PG. 1992. Alcohol, beer, and lung cancer in postmenopausal women. Ann Epidemiol 2:587– 596. Pottern LM, Morris LE, Bolt WJ, et al. 1981. Esophageal cancer among black men in Washington, D.C. I. Alcohol, tobacco, and other risk factors. J Natl Cancer Inst 67:777–783. Prescott E, Gronbaek M, Becker U, Sorensen TIA. 1999. Alcohol intake and the risk of lung cancer: influence of type of alcoholic beverage. Am J Epidemiol 149:463–470. Purohit V. 2000. Can alcohol promote aromatization of androgens to estrogens? A review. Alcohol 22:123–127. Rachtan J, Sokolowski A. 1997. Risk factors for lung cancer among women in Poland. Lung Cancer 18:137–145.

Ramchandani VA, Bosron WF, Li TK. 2001. Research advances in ethanol metabolism. Pathol Biol 49:676–682. Ranstam J, Olsson H. 1995. Alcohol, cigarette smoking, and the risk of breast cancer. Cancer Detect Prev 19:487–493. Rao DN, Ganesh B, Dinshaw KA, Mohandas KM. 2002. A case-control study of stomach cancer in Mumbai, India. Int J Cancer 99:727–731. Roberts BJ, Shoaf SE, Song BJ. 1995. Rapid change in cytochrome p4502E1 (cyp2E1) activity and other p450 isozymes following ethanol withdrawl in rats. Biochem Pharmacol 49:16. Rohan TE, McMichael AJ. 1988. Alcohol consumption and risk of breast cancer. Int J Cancer 41:695–699. Rosenberg L, Palmer JR, Miller DR, Clarke EA, Shapiro S. 1990. A casecontrol study of alcoholic beverage consumption and breast cancer. Am J Epidemiol 131:6–14. Rosenberg L, Slone D, Shapiro S, et al. 1982. Breast cancer and alcoholicbeverage consumption. Lancet 1:267–270. Russell M, Marshall JR, Trevisan M, Freudenheim JL, Chan AWK, Markovic N, Vana JE, Priore RL. 1997. Test-retest reliability of the cognitive lifetime drinking history. Am J Epidemiol 146:975–981. Saleem MMNM, Al-Tamer YY, Skursky L, Al-Habbal Z. 1984. Alcohol dehyrdrogenase activity in the human tissues. Biochem Med 31:1–9. Sankaranarayanan R, Duffy SW, Nair MK, Padmakumary G, Day NE. 1990. Tobacco and alcohol as risk factors in cancer of the larynx in Kerala. Int J Cancer 45:879–882. Sankaranarayanan R, Duffy SW, Padmakumary G, et al. 1991. Risk factors for cancer of the oesophagus in Kerala, India. Int J Cancer 49:485– 489. Schatzkin A, Carter CL, Green SB, Kreger BE, Splansky GL, Anderson KM, Helsel WE, Kannel WB. 1989. Is alcohol consumption related to breast cancer? Results from the Framingham heart study. J Natl Cancer Inst 81:31–35. Schatzkin A, Jones DY, Hoover RN, et al. 1987. Alcohol consumption and breast cancer in the epidemiologic follow-up study of the first national health and nutrition examination survey. N Engl J Med 316:1169–1173. Schatzkin A, Piantadosi S, Miccozzi M, Bartee D. 1989. Alcohol consumption and breast cancer: a cross-national correlation study. Int J Epidemiol 18:28–31. Schlecht NF, Pintos J, Kowalski LP, Franco EL. 2001. Effect of type of alcoholic beverage on the risks of upper aerodigestive tract cancers in Brazil. Cancer Causes Control 7:579–587. Schmidt W, Popham RE. 1981. The role of drinking and smoking in mortality from cancer and other causes in male alcoholics. Cancer 47:1031– 1041. Schuurman AG, Zeegers MPA, Goldbohm RA, Van Den Brandt PA. 1999. A case-cohort study on prostate cancer risk in relation to family history of prostate cancer. Epidemiology 10:192–195. Segal I, Reinach SG, De Beer M. 1988. Factors associated with oesophageal cancer in Soweto, South Africa. Br J Cancer 58:681–686. Seitz HK, Poschl G, Simanowski UA. 1998. Alcohol and cancer. Recent Dev Alcohol 14:67–95. Sellers TA, Kushi LH, Cerhan JR, Vierkant RA, Gapstur SM, Vachon CM, Olson JE, Therneau TM, Folsom AR. 2001. Dietary folate intake, alcohol, and risk of breast cancer in a prospective study of postmenopausal Women. Epidemiology 12:420–428. Sellers TA, Vierkant RA, Cerhan JR, Gapstur SM, Vachon CM, Olson JE, Pankratz S, Kushi LH, Folsom AR. 2002. Interaction of dietary folate intake, alcohol and risks of hormone-receptor defined breast cancer in a prospective study of postmenopausal women. Cancer Epidemiol Biomarkers Prev 11:1104–1107. Setiawan VW, Zhang ZF, Yu GP, Li YL, Lu ML, Tsai CJ, Cordova D, Wang MR, Guo CH, Yu SZ, Kurtz RC. 2000. GSTT1 and GSTM1 null genotypes and the risk of gastric cancer: a case-control study in a Chinese population. Cancer Epidemiol Biomarkers Prev 9:73–80. Sharp L, Chilvers CED, Cheng KK, McKinney PA, Logan RFA, CookMozaffari P, Ahmed A, Day NE. 2001. Risk factors for squamous cell carcinoma of the oesophagus in women: a case-control study. Br J Cancer 85:1667–1670. Sharpe CR, Siemiatycki J. 2001. Case-control study of alcohol consumption and prostate cancer risk in Montreal, Canada. Cancer Causes Control 12:589–598. Sharpe CR, Siemiatycki J, Rachet B. 2002. Effects of alcohol consumption on the risk of colorectal cancer among men by anatomical subsite (Canada). Cancer Causes Control 13:483–491. Shibata A, Hirohata T, Toshima H, et al. 1986. The role of drinking and cigarette smoking in the excess deaths from liver cancer. Jpn J Cancer 77:287–295. Silverman DT. 2001. Risk factors for pancreatic cancer: a case-control study based on direct interviews. Teratog Carcinog Mutagen 21:7–25.

Alcohol Simon MS, Carman W, Wolfe R, et al. 1991. Alcohol consumption and the risk of breast cancer: a report from the Tecumseh Community Health Study. J Clin Epidemiol 44:755–761. Simonotti P, Brusamolino N, Pelegrini P, Viani P, Clemonte G, Roggi C, Cestaro B. 1995. Evaluation of the effect of alcohol consumption on erythrocyte lipids and vitamins in a health population. Alcohol Clin Exp Re 19:517–522. Singletary KW, Gapstur SM. 2001. Alcohol and breast cancer: review of epidemiologic and experimental evidence and potential mechanisms. JAMA 286;2143–2151. Slattery ML, Edwards SL, Boucher KM, Anderson K, Caan BJ. 1999. Lifestyle and colon cancer: an assessment of factors associated with risk. Am J Epidemiol 150:869–877. Smith SJ, Deacon JM, Chilvers CED, et al. 1994. Alcohol, smoking, passive smoking and caffeine in relation to breast cancer risk in young women. Br J Cancer 70:112–119. Smith-Warner SA, Spiegelman D, Yaun SS, Van Den Brandt PA, Folsom AR, Goldbohm A, Graham S, Holmberg L, Howe GR, Marshall JR, Miller AB, Potter JD, Speizer FE, Willet WC, Wolk A, Hunter DJ. 1998. Alcohol and breast cancer in women. a pooled analysis of cohort studies. JAMA 279:535–540. Sneyd MJ, Paul C, Spears GFS, et al. 1991. Alcohol consumption and risk of breast cancer. Int J Cancer 48:812–815. Soler M, Chatenoud L, La Vecchia C, Franceschi S, Negri E. 1998. Diet, alcohol, coffee and pancreatic cancer: final results from an Italian study. Eur J Cancer Prev 7:455–460. Sorensen HT, Friis S, Olsen JH, Thulstrup AM, Mellemkjaer L, Linet M, et al. 1998. Risk of liver and other types of cancer in patients with cirrhosis: a nationwide cohort study in Denmark. Hepatology 28:921–925. Steer ML, Waxman I, Freedman S. 1998. Medical progress: chronic pancreatitis. N Engl Med J 332:1482–1490. Stemhagen A, Slade J, Altman R, et al. 1983. Occupational risk factors and liver cancer. a retrospective case-control study of primary liver cancer in New Jersey. Am J Epidemiol 117:443–454. Stemmermann GN, Nomura A, Chyon P, et al. 1990. Prospective study of alcohol intake and large bowel cancer. Dig Dis Sci35:1414–1420. Stockwell HG, Matanoski GM. 1984. Association of alcohol consumption and cancer. N Engl J Med 311:256. Stockwell HG, Matanoski GM. 1985. A case-control study of lung cancer in painters. J Occup Med 27:125–126. Stolzenberg-Solomon RZ, Pietinen P, Barrett MJ, Taylor PR, Virtamo J, Albanes D. 2001. Dietary and other methyl-group availability factors and pancreatic cancer risk in a cohort of male smokers. Am J Epidemiol 153:680–687. Stromberg P, Svensson S, Hedberg JJ, Nordling E, Hoog JO. 2002. Identifications and characterization of two allelic forms of human alcohol dehydrogenase 2. Cell Mol Life Sci 59:552–559. Stumer T, Wang-Gohrke S, Arndt V, Boeing H, Kong X, Kreienberg R, Brenner H. 2002. Interaction between alcohol dehydrogenase II gene, alcohol consumption and risk for breast cancer. Br J Cancer 87:519–523. Swanson CA, Coates RJ, Malone KE, Gammon MD, Schoenberg JB, Brogan DJ, McAdams M, Potischman N, Hoover RN, Brinton LA. 1997. Alcohol consumption and breast cancer risk among women under age 45 years. Epidemiology 8:231–237. Talamini G, Bassi C, FalconI M, Sartori N, Salvia R, Rigo L, Castagnini A, Cavallini G, Pequignot G. 1999. Alcohol and smoking as risk factors in chronic pancreatitis and pancreatic cancer. Dig Dis Sci 44:1303–1311. Talamini G, Capelli P, Zamboni G, Mastromauro M, Pasetto M, Castagnini A, Angelini G, Bassi C, Scarpa A. 2000. Alcohol, smoking and papillomavirus infection as risk factors for esophageal squamous-cell papilloma and esophageal squamous-cell carcinoma in Italy. Int J Cancer 86:874–878. Tanaka F, Shiratori Y, Yokosuka O, Imazeki F, Tsukada Y, Omata M. 1997. Polymorphism of alcohol-metabolizing genes affects drinking behavior and alcoholic liver disease in Japanese men. Alcohol Clin Exp Res 21:596–601. Tavani A, Ferraroni M, Mezzetti M, Franceschi S, Lo Re A, La Vecchia C. 1998. Alcohol intake and risk of cancers of the colon and rectum. Nutr Cancer 30:213–219. Tavani A, Pregnolato A, Negri E, La Vecchia C. 1997. Alcohol consumption and risk of pancreatic cancer. Nutr Cancer 27:157–161. Thun MJ, Peto R, Lopez AD, Monaco JH, Henley SJ, Heath CW, Doll R. 1997. Alcohol consumption and mortality among middle-aged and elderly U.S. adults. N Engl J Med 337:1705–1714. Todoroki I, Kono S, Shinchi K, Honjo S, Sakurai Y, Wakabayashi K, Imanishi K, Nishikawa H, Ogawa S, Katsurada M. 1995. Relationship of cigarette smoking, alcohol use, and dietary habits with sigmoid colon adenomas. Ann Epidemiol 5:478–483.

257

Toniolo P, Riboli E, Protta F, et al. 1989. Breast cancer and alcohol consumption: a case-control study in northern Italy. Cancer Res 49:5203– 5206. Tonnesen H, Moller H, Andersen JR, Jensen E, Juel K. 1994. Cancer morbidity in alcohol abusers. Br J Cancer 69:327–332. Trentham-Dietz A, Newcomb PA, Storer BE, Remington PL. 2000. Risk factors for carcinoma in situ of the breast. Cancer Epidemiol Biomarkers Prev 9:697–703. Trichopoulos D, Day NE, Kaklamani E, et al. 1987. Hepatitis B virus, tobacco smoking and ethanol consumption in the etiology of hepatocellular carcinomas. Int J Cancer 39:45–49. Tsukuma H, Hiyama T, Oshima A, et al. 1990. A case-control study of hepatocellular carcinoma in Osaka, Japan. Int J Cancer 45:231–236. Tuyns AJ, Esteve J, Raymond L, Berrino F, Benhamou E, Blanchet F, Boffetta P, Crosignani P, Del Moral A, Lehmann W, Merletti F, Pequignot G, Riboli E, Sancho-Garnier H, Terracini B, Zubiri A, Zubiri L. 1988. Cancer of the larynx/hypopharynx, tobacco and alcohol: IARC international case-control study in Turin and Varese (Italy), Zaragoza and Navarra (Spain), Geneva (Switzerland) and Calvados (France). Int J Cancer 41:483–491. Tuyns AJ, Pequignot G, Abbatucci JS. 1979. Oesophageal cancer and alcohol consumption: importance of type of beverage. Int J Cancer 23:443– 447. Tuyns AJ, Pequignot G, Jensen OM. 1977. Le cancer de l’oesophage en IlleetVillaine en fonction des niveaux de consommation d’alcool et de tabac: des risques qui se multiplient. Bull Cancer 64:45–60. Van Den Brandt PA, Goldbohm RA, Van’t Veer P. 1995. Alcohol and breast cancer: results from The Netherlands cohort study. Am J Epidemiol 141:907–915. Van Den Brandt PA, Goldbohm RA, Van’t Veer P, Volovics A, Hermus RJJ, Sturmans F. 1990. A large-scale prospective cohort study on diet and cancer in The Netherlands. J Clin Epidemiol 43:285–295. Van Den Brandt PA, Spiegelman D, Yaun SS, Adami HO, Beeson L, Folsom AR, Fraser G, Goldbohm RA, Graham S, Kushi L, Marshall JR, Miller AB, Rohan T, Smith-Warner SA, Speizer FE, Willett WC, Wolk A, Hunter DJ. 2000. Pooled analysis of prospective cohort studies on height, weight, and breast cancer risk. Am J Epidemiol 152:514–527. Vassallo A, Correa P, De Stefani E, et al. 1985. Esophageal cancer in Uruguay: a case-control study. J Natl Cancer Inst 75:1005–1009. Velema JP, Walker AM, Gold EB. 1986. Alcohol and pancreatic cancer: insufficient epidemiologic evidence for a casual relationship. Epidemiol Rev 8:28–41. Victora CG, Munoz N, Day NE, Barcelos LB, Peccin DA, Braga NM. 1987. Hot beverages and oesophageal cancer in southern Brazil: a case-control study. Int J Cancer 39:710–716. Viel JF, Perarnau JM, Challier B, Faivre-Nappez I. 1997. Alcoholic calories, red wine consumption and breast cancer among premenopausal women. Eur J Epidemiol 13:639–643. Villeneuve PJ, Johnson KC, Hanley AJ, Mao Y. 2000. Alcohol, tobacco and coffee consumption and the risk of pancreatic cancer: results from the Canadian enhanced surveillance system case-control project; Canadian Cancer Registries Epidemiology Research Group. Eur J Cancer Prev 9:49–58. Weaver A, Fleming SM, Smith DB. 1979. Mouthwash and oral cancer: carcinogen or coincidence? J Oral Surg 37:250–253. White IR, Altmann DR, Nanchahal K. 2002. Alcohol consumption and mortality: modelling risks for men and women at different ages. BMJ 325:191–197. Willett W. 1998. Nutritional Epidemiology, 2nd ed. Monographs in Epidemiology and Biostatistics 30. New York: Oxford University Press. Willet WC, Stampfer MJ, Colditz GA, et al. 1987. Moderate alcohol consumption and the risk of breast cancer. N Engl J Med 316:1174–1180. Williams RR, Horm JW. 1977. Association of cancer sites with tobacco and alcohol consumption and socioeconomic status of patients: interview study from the Third National Cancer Survey. J Natl Cancer Inst 58:525–547. Woodson K, Albanes D, Tangrea JA, Rautalahti M, Virtamo J, Taylor PR. 1999. Association between alcohol and lung cancer in the alpha-tocopherol, beta-carotene cancer prevention study in Finland. Cancer Causes Control 10:219–226. World Cancer Research Fund. 1997. Food, Nutrition and the Prevention of Cancer: A Global Perspective. Washington, DC: American Institute for Cancer Research. Wright RM, McManaman JL, Repine JE. 1999. Alcohol-induced breast cancer: a proposed mechanism. Free Radical Biol Med 26:348–354. Wu MT, Lee YC, Chen CJ, Yang PW, Lee CJ, Wu DC, Hsu HK, Ho CK, Kao EL, Lee JM. 2001. Risk of betel chewing for oesophageal cancer in Taiwan. Br J Cancer 85:658–660.

258

PART III: THE CAUSES OF CANCER

Wu MT, Lee JM, Wu DC, Ho CK, Wang YT, Lee YC, Hsu HK, Kao EL. 2002. Genetic polymorphisms of cytochrome p4501a1 and oesophageal squamous-cell carcinoma in Taiwan. Br J Cancer 87:529–532. Wu-Williams AH, Yu MC, Mack TM. 1990. Life-style, work-place, and stomach cancer by subsite in young men of Los Angeles County. Cancer Res 50:2569–2576. Wynder EL, Bross IJ. 1961. A study of etiological factors in cancer of the esophagus. Cancer 14:389–413. Wynder EL, Bross IJ, Day E. 1956. A study of environmental factors in cancer of the larynx. Cancer 9:86–110. Wynder EL, Bross IJ, Feldman RM. 1957. A study of etiological factors in cancer of the mouth. Cancer 10:1300–1323. Wynder EL, Covey LS, Mabuchi K, et al. 1976. Environmental factors in cancer of the larynx: a second look. Cancer 38:1591–1601. Wynder EL, Kabat G, Rosenber S, Levenstein M. 1983. Oral cancer and mouthwash use. J Natl Cancer Inst 70:255–260. Yamada K, Araki S, Tamura M, Sakai I, Takahashi Y, Kashihara H, Kono S. 1997. Case-control study of colorectal carcinoma in situ and cancer in relation to cigarette smoking and alcohol use (Japan). Cancer Causes Control 8:780–785. Yang CX, Wang HY, Wang ZM, Du HZ, Tao DM, Mu XY, Chen HG, Lei Y, Matsuo K, Tajima K. 2005. Risk factors for esophageal cancer: a case control study in South-western China. Asian Pac J Cancer Prev 6:48–53. Ye W, Ekstrom AM, Hansson LE, Bergstrom R, Nyren O. 1999. Tobacco, alcohol and the risk of gastric cancer by sub-site and histologic type. Int J Cancer 83:223–229. Ye W, Lagergren J, Weiderpass E, Nyren O, Adami HO, Ekbom A. 2002. Alcohol abuse and the risk of pancreatic cancer. Gut 51:236–239. Yen S, Hsieh CC, MacMahon B. 1982. Consumption of alcohol and tobacco and other risk factors for pancreatitis. Am J Epidemiol 116:407–414. Yokoyama A, Muramatsu T, Omori T, Yokoyama T, Matsushita S, Higuchi S, Maruyama K, Ishii H. 2001. Alcohol and aldehyde dehydrogenase gene polymorphisms and oropharyngolaryngeal, esophageal and stomach cancers in Japanese alcoholics. Carcinogenesis 22:433–439. Yu MC, Garabrant DH, Peters JM, et al. 1988. Tobacco, alcohol, diet, occupation, and carcinoma of the esophagus. Cancer Res 48:3843–3848.

Yuan JM, Ross RK, Gao YT, Henderson BE, Yu MC. 1997. Follow up study of moderate alcohol intake and mortality among middle aged men in Shanghai, China. BMJ 314:18–23. Yuan JM, Govindarajan S, Anakawa K, Yu MC. 2004. Synergism of alcohol, diabetes, and viral hepatitis on the risk of hepatocellular carcinoma in blacks and whites in the U.S. Cancer 101:1009–1017. Zambon P, Talamini R, La Vecchia C, Dal Maos L, Negri E, Tognazzo S, Simonato L, Franceschi S. 2000. Smoking, type of alcoholic beverage and squamous-cell oesophageal cancer in northern Italy. Int J Cancer 86:144–149. Zaridze D, Borisova E, Maximovitch D, Chkhikvadze V. 2000. Alcohol consumption, smoking and risk of gastric cancer: case-control study from Moscow, Russia. Cancer Causes Control 11:363–371. Zatonski W, Becher H, Lissowska J, Wahrendorf J. 1991. Tobacco, alcohol, and diet in the etiology of laryngeal cancer: a population-based case-control study. Cancer Causes Control 2:3–10. Zatonski W, Boyle P, Przewozruak K, et al. 1993. Cigarette smoking, alcohol, tea and coffee consumption and pancreas cancer risk: a case-control study from Opole, Poland. Int J Cancer 53:601–607. Zavras AI, Douglass CW, Joshipura K, Wu T, Laskaris G, Petridou E, Dokianakis G, Segas J, Lefantzis D, Nomikos P, Wang YF, Diehl SR. 2001. Smoking and alcohol in the etiology of oral cancer: gender-specific risk profiles in the south of Greece. Oral Oncol 37:28–35. Zhang SF, Hunter DJ, Hankinson SE, Giovannucci E, Rosner BA, Colditz GA, Speizer FE, Willet WC. 1999. A prospective study of folate intake and the risk of breast cancer. JAMA 281:1632–1637. Zheng TZ, Boyle P, Hu JF, et al. 1990. Tobacco smoking, alcohol consumption, and risk of oral cancer: a case-control study in Beijing, People’s Republic of China. Cancer Causes Control 1:173–179. Zheng W, Blot WJ, Shu X, et al. 1992. Risk factors for laryngeal cancer in Shanghai, China. Am J Epidemiol 136:178–191. Zheng W, Blot WJ, Shu X, et al. 1992. Risk factors for oral and pharyngeal cancer in Shanghai, with emphasis on diet. Cancer Epidemiol Biomarkers Prev 1:441–448. Zheng W, McLaughlin JK, Gridley G, et al. 1993. A cohort study of smoking, alcohol consumption, and dietary factors for pancreatic cancer (United States). Cancer Causes Control 4:477–482.

15

Ionizing Radiation JOHN D. BOICE, JR.

C

onclusive evidence that ionizing radiation causes cancer comes from the studies of radium dial painters, underground miners, pioneering radiologists, patient populations, and Japanese atomic bomb survivors (IARC, 2000, 2001). While the single most important study is of the survivors of the atomic bombs, there are well over 100 epidemiologic studies of patient populations linking radiation to cancer. A wealth of knowledge on carcinogenic effects also has been derived from experimental studies in animals and in cell culture (UNSCEAR, 1986, 1988, 1993, 1994, 2000, 2001; IARC 2000, 2001; NCRP, 2001). No other environmental carcinogen, with the possible exception of tobacco, has been as extensively studied. Both the United Nations (UNSCEAR, 2000) and the National Academy of Sciences (2005) periodically publish authoritative volumes on the effects of radiation, and most recently informative monographs have been published by the International Agency for Research on Cancer (IARC, 2000, 2001). Many human cancers have been convincingly linked to radiation, with a few notable exceptions such as chronic lymphocytic leukemia (CLL), Hodgkin disease, malignant melanoma, cervical cancer, testes cancer, and prostate cancer. The important questions, however, are not whether radiation causes cancer but how does it cause cancer? How much cancer is caused by radiation? Is risk substantially diminished when exposure is spread over time? How long does the risk last after exposure? Which organs are particularly sensitive and why do they vary in sensitivity? Although radiation can be readily detected and quantified, and precise radiation protection guidelines exist, there remains uncertainty about the shape of the dose-effect curve at low doses for sparsely ionizing radiation such as X-rays or gamma rays; the influence of physical exposure conditions such as dose rate (fractionation or protraction of dose) and type of radiation; and the influence of various biological modifiers of risk such as age and sex and possibly genetic predisposition. Many of these issues are of current scientific, public health, and radiation protection interest. Since our last review in 1996 (Boice et al., 1996a), the attention given to the health effects of ionizing radiation has intensified and the number of published epidemiologic studies has mushroomed (UNSCEAR, 2000; IARC, 2000, 2001). Plutonium at high levels was convincingly linked to increased cancer risks among earlier nuclear workers at the Mayak facility in Russia (Koshurnikova et al., 2002a). The carcinogenicity of uranium was extensively studied and reviewed because of concerns over exposure to depleted uranium munitions (CRS, 2001; IOM, 2001; IARC, 2001). The Chernobyl reactor accident in the Soviet Union in 1986 distributed radionuclides around the world, but the ensuing health effects appear limited to thyroid cancer among children who lived in nearby communities (UNSCEAR, 2000). The evidence that indoor radon gas is an important risk factor for lung cancer has accumulated (NAS, 1999). The dosimetry for the atomic bomb survivors in Japan underwent major revisions, but is expected to have only a modest effect on the risk estimates (Straume et al., 2003; Preston et al., 2004). The National Institutes of Health updated their radio-epidemiological tables on the probability that a given dose of radiation may have caused a specific cancer in an individual (NIH, 2003). Early studies linking residence near nuclear installations (Forman et al., 1987) and preconception radiation (Gardner et al., 1990) to childhood leukemia were not borne out in subsequent international investigations (UNSCEAR, 1994; Doll et al., 1994; Wakeford, 2003). Participants at some nuclear weapons tests appeared

to be at slight increased risk of leukemia but earlier excesses of multiple myeloma disappeared (Muirhead et al., 2003). New studies were published of patients (Travis et al., 2002, 2003b; IARC, 2000) and workers (Muirhead et al., 1999; Sont et al., 2001; Cardis et al., 2005a). Studies of patients given Thorotrast are coming to an end (IARC, 2001; Travis et al., 2003a) as are those of radium dial painters (Fry, 1998). The increased survival of cancer patients, particularly children, has raised concern about the late effects of curative treatments (Neglia et al., 2001; Meadows, 2003). Offspring of survivors of childhood cancer have been studied for possible trans-generational effects of radiotherapy (Boice et al., 2003d). There have been remarkable advances in cellular biology that may provide insights into the mechanism of carcinogenesis (Brooks, 2005); intriguing cellular phenomena include the bystander effect, genomic instability (Morgan, 2003), radiation hormesis (Kaiser, 2003), cellular repair, and apoptosis (Rothkamm and Lobrich, 2003). Throughout this chapter we will describe specific studies of exposed populations, touching on strengths and limitations, the need for caution in interpretation, and implications for public health, radiation protection, and carcinogenesis in general. It is noted that the UNSCEAR (1994; 2000) reports provide comprehensive summaries of the strengths and weaknesses of most studies.

SOURCES OF EXPOSURE Radiation is a natural part of our environment and is continually emitted from rocks, soil, plants, and water. Over 25 radioactive elements occur naturally in the environment, including uranium, thorium, radon, radium, and potassium. Other sources of radiation originate from the sun and from outer space. These terrestrial and cosmic sources of radiation continue to bathe us in a sea of low-level radiation throughout our lives (UNSCEAR, 2000). Epidemiologic studies have not found such background radiation exposure to cause cancer (Boice, 2002). The greatest population exposure to ionizing radiation comes from these natural background sources, about 2.4 millisievert (mSv) (0.24 rem) per year (UNSCEAR, 2000). Cosmic rays account for 0.40 mSv/year, which vary by altitude; terrestrial radiations contribute 0.50 mSv/year, which vary according to the distribution in soil of radioactive elements such as uranium; internally deposited radionuclides provide 0.30 mSv/year such as potassium-40; and radon weighs in with 1.2 mSv/year, confined mainly to lung. The greatest source of manmade radiation is from medical uses (0.40 mSv/year), with exposures increasing directly with age. Occupation, nuclear power, fallout from testing nuclear weapons, and consumer products make only a minor contribution (0.01 mSv/year). The average per capita dose from all sources of radiation, excluding radon, is thus about 1.6 mSv (0.160 rem) per year. In comparison, radiation workers are permitted up to 50 mSv in a single year of work (ICRP, 1991). Based on linear extrapolation from high-dose data (>200 mSv) it can be inferred that only a small fraction, perhaps 2% of all cancers, might be attributable to all sources of radiation (1.6 mSv/yr) (Jablon and Bailar, 1980; Harvard, 1996; Berrington de González and Darby, 2004), excluding indoor radon, which has been suggested as an important cause of lung cancer (NAS, 1999). Such estimates should be interpreted with caution given the substantial uncertainties in applying risks from high-dose studies to low-dose and low dose-rate situations.

259

260

PART III: THE CAUSES OF CANCER

It appears, though, that a reduction in medical X-ray exposures and possibly indoor radon are the only ways to reduce population exposure and presumed radiogenic cancer risks.

Table 15–1. Ranking of Various Tissues Regarding Carcinogenic Influence of Low-Linear Energy Transfer Radiation*

POPULATIONS STUDIED

Type of Cancer

Knowledge of radiation effects has come from medically exposed patients, occupational groups, atomic bomb survivors, and persons exposed to radioactive fallout or naturally occurring radiation such as radon (Fig. 15–1) (Boice and Fraumeni, 1984; Upton et al., 1986, 2003; Boice, 1997a; UNSCEAR, 2000; NAS, 1990, 1999, 2005; NCRP, 2001; NRPB, 2003a; IARC 2000, 2001). It is interesting that the variation in radiation risk reported in exposed populations, whether measured on a relative or absolute scale, is not that great and most differences can be explained by differences in dose, age, and follow-up distributions (Table 15–1) (UNSCEAR, 2000).

cancers frequently associated with radiation with authoritative risk estimates

Medical Exposures Chest Exposures Tuberculosis. Frequent X-ray fluoroscopic examinations to monitor lung-collapse treatments for tuberculosis (TB) during 1935 to 1954 increased the risk of breast cancer among 2573 women in Massachusetts who received an average of 88 chest fluoroscopies, compared with 2367 non-exposed women with TB (Boice and Monson, 1977; Boice et al., 1991c). The follow-up was 97% complete. Excess breast cancers (147 observed vs. 113.6 expected) were related to dose in a linear manner. The excess did not appear until 10 to 15 years after exposure and remained high throughout 50 years of observation. Exposures during the adolescent and teenage years carried the greatest risk, and exposures after age 40 carried the least risk (suggesting an important promotional role for hormonal factors). Radiationabsorbed dose to the breast per fluoroscopy was estimated as about 0.9 centigray (cGy) (rad), and the average cumulative dose as 79 cGy. The excess absolute risk was estimated as 10.7 cases per 104 personyears per gray (PY-Gy), and the relative risk (RR) as 1.61 at 1 Gy. A large Canadian series was generally consistent with these findings, although based on mortality rather than incidence data (Howe and McLaughlin, 1996; Miller et al., 1989). The breast appears to be one of the most sensitive tissues to the carcinogenic action of radiation (Table 15–1). Compared with other studies, fractionated high doserate exposures seem similar to single exposures of the same total dose in their ability to induce breast cancer (Boice and Land, 1979; Little and Boice, 1999; Preston et al., 2002a). No excesses of leukemia, lymphoma, or lung cancer have been reported among TB patients repeatedly exposed to fluoroscopic X-rays (Davis et al., 1989). The mean dose to lung tissue was 84 cGy (84 rad). Similarly, the large Canadian TB fluoroscopy study included 1178 lung cancer deaths and no increase in lung cancer occurred despite an average lung dose of 1.02 Gy (102 rad) (Howe, 1995). In animal experiments, splitting dose over time also appears to reduce the risk of radiogenic lung cancer to a much greater extent than the reduction seen for radiogenic breast cancer (Ullrich et al., 1987; UNSCEAR, 1986). Uncertainties in dosimetry limit precise quantification of radiation risks from these TB series. On the other hand, practically all environmental, occupational, and nontherapeutic medical exposures involve radiation doses no greater than those received by TB patients from a single fluoroscopy, yet cumulative exposures were high enough to engender measurable excess risks. Risk estimates based on these studies of frequent low-dose exposure are thus relatively free of the problems of extrapolation to low doses that characterize studies of brief, high-dose exposures, and may be more directly relevant to public health concerns.

Mastitis. A survey of 601 women treated with radiation for acute postpartum mastitis in New York revealed 56 (or 9.3%) breast cancers compared with 59 (or 4.8%) in 1239 women not irradiated (Shore et al., 1986). Treatment involved 1 to 11 exposures (mean, 3.4), 3.8 Gy to the irradiated breast. Doses were delivered by carefully calibrated

Range of Risk Estimates RR at 1 Gy

Excess Risk ¥104 PY-Gy

Leukemia

1.2–5.4

0.5–2.7

Thyroid

1.3–3.5

0.4–9.1

Female breast

1.1–3.4

0.1–9.1

Comment

Especially myeloid leukemia, short latency Low mortality; little risk if exposed >20 yr Little risk if exposed >40 yr

cancers occasionally associated with robust risk estimates Lung

1.0–2.0

0.0–4.6

Stomach Colon

1.0–1.5 1.0–1.7

0.0–5.7 0.0–3.2

Esophagus Bladder

1.2–1.8 1.1–1.8

0.2–0.6 0.1–1.0

Ovary Brain and nervous system

1.0–2.3 1.0–5.1

0.1–0.7 0.0–2.1

Liver

1.0–1.5

0.0–1.6

Interaction with smoking complex Major A-bomb effect Not seen after cervical cancer Both low- and highdose effect Mainly after highdose childhood exposures Major Thorotrast effect; A-bomb virus interaction

cancers rarely associated †with radiation with uncertain risk estimates Kidney Salivary glands Non-Hodgkin lymphoma

1.0–1.7 1.1–1.7

0.0–1.1 0.1–0.2

‡

‡

Myeloma

1.0–5.2

0.0–0.9

Skin

1.0–2.0

0.1–2.5

Rectum

1.0–1.2

0.0–0.1

Uterus

1.0–1.01

0.0–0.5

Bone

1.0–1.1

0.0–0.2

1.2

‡

Connective tissues

Limited evidence Some evidence Little evidence, possible high-dose effect Uncertainty whether association causal Effect may be limited to high doses (or UV necessary) Effect may be limited to high doses Effect may be limited to high doses Effect may be limited to high doses Effect may be limited to high doses

cancers never or sporadically associated with radiation with no risk estimates†† Chronic lymphocytic leukemia Pancreas Hodgkin disease Prostate Testis Cervix Certain childhood cancers** Supporting tissues of skeleton‡‡

‡

‡

Absent?

‡

‡

‡

‡

‡

‡

‡

‡

‡

‡

‡

‡

Little evidence Little evidence Little evidence Little evidence Little evidence Absent?

‡

‡

Little evidence

*Relative rankings are based on the studies summarized in this chapter, and consider the frequency that the observed cancer is reported in irradiated populations, the strength of the associations found, the significance of the association, and the availability of reliable estimates of radiation risk per unit organ dose (cf NAS, 1980). Risk coefficents mainly from UNSCEAR, 2000 with supplemental information from Thompson et al., 1994; Pierce et al., 1996; Preston et al., 2003; NAS, 1990; UNSCEAR, 1994; IARC 2000, 2001. Range of risk estimates are associated with difference in age distributions, follow-up times, and other factors among exposed populations; †Association is inconsistently found and/or available estimates of risk are highly uncertain; ‡No reliable estimates available; **Sites for which radiation-induced cancers have not yet been reported or confirmed; Retinoblastoma, Wilms’ tumor, neuroblastoma, and others of embryonic origin; ‡‡Muscles, tendons, and synovial membranes of joints.

–

–

–

–

+++ +++ + + – –

– –

– ++/-

– +/–

+

+

+

– – +++

–

+++ +++ – +/-

– –

+ +

–

+++

+

+ –

– – – –

+ +++

+++ – – – –

–

+

– – – –

+ –

–

–

– – – –

m ctu

sta

rea

ho mp

Lu ng

Liv er

nc

+

–

–

+/–

+++

–

– +++ +++ +++

– –

– + +

– –

– –

– –

+ –

+

–

+ –

– –

+ +++ +++

+

–

–/+

– +

–

+

–

+++ –

–

– +

–

+ +

+ + + – + +

+++

–

–

Ute rus

–

+++

Th yro id

–

s

–

ste

+

ch

–

Te

–

ma

–

Sto

–

Sa liva ry Gla nd Sk in

+

te

Re

–

+++ + +

s

+++

Ov ary

Pro

ia

+

em

Pa

–

My elo ma

+

Ly

–

ma

s gu ha

t, F as

– –

–

+++

uk

+

–

–

Le

+

–

Kid ne y

–

op

–

Es

–

+++

Co lon

–

Bre

in, Bra

–

Ce rvi x

S CN

TS ,S ne Bo

A-Bomb Radiation Japanese A-Bomb Marshall Islanders Weapons Test Part Medical Benign GYN Spondylitis (x-ray) Tinea Capitis Thymus Tonsil Hemangioma Breast Mastitis/Benign Peptic Ulcer Cervical Cancer Childhood Cancer Hodgkin's Disease Breast Cancer TB - Fluoroscopy Diagnostic X-ray In Utero X-ray Scoliosis Radionuclides Thorotrast (Th-232) Spondylitis (Ra-224) Diagnostic (I -131) Hyperthyroidism (I-131) Thyroid Ca (I-131) P-Vera (P-32) Occupation Ra Dial Painters (Ra) Radiologists/Rad Tech Underground Miners Nuclear Workers Uranium Processors Chernobyl Cleanup Mayak Workers (Pu) Environmental Natural Background Indoor Radon Hanford Thyroid Chernobyl Fallout Weapons Test Fallout Techa River

Bla dd er

Epidemiologic Study

em

ale

TYPE OF CANCER

–

– +

–

+

– – –

– – – –

+++ –/+ – – +

+++ + –

+

–

–

– – – – –

– +++ – –

– – – –

–

+++

+

–

–

–

– + – –

–

– + +

Figure 15–1. Distribution of various types of cancer associated with radiation in different populations. 䊏 = strong association; 䊏 = meaningful association; + = suggested but unconfirmed or questionable association; - = no significant association found, although study was reasonably powerful; +/- = there was more than one study with conflicting findings; Blank = no or minimal data.

+

+ –

– –

+ + – +

– –

–

–

– – –

– – – –

– – –

+

– – – –

– – –

– – –

– – +++ +

Note that the “strength” of the association is not related to level of risk (e.g., the strong association between radiotherapy for cervical cancer and cancer of the rectum is based on very high-dose exposure and the risk coefficient is quite small) (see Table 15–1).

262

PART III: THE CAUSES OF CANCER

X-ray therapy units and were accurately known. Excess risk was not apparent until 10 years after treatment, and was approximately linear in dose up to 3.5 Gy, when it declined. The absolute risk was estimated as 3.5 cases/104 PY-Gy, and the RR at 1 Gy as 1.4. Interestingly, the age-specific absolute risk estimates obtained for mastitis patients (who were especially healthy, having just given birth) were very similar to those computed for TB fluoroscopy patients and A-bomb survivors (Boice et al., 1979; Preston et al., 2002a).

Other. More is known about radiation-induced breast cancer than perhaps any other malignancy (Boice, 2001a). Risk of breast cancer among women given radiation therapy (mean, 5.5 Gy) for benign breast disorders in Sweden was inversely related to age at exposure (Mattson et al., 1993). Use of a nonirradiated comparison group minimized the possibility that the underlying breast disease was the sole reason for the breast cancer excess, although the authors recognized the possibility that women with more serious breast disease (conceivably at higher risk for breast cancer) were selected for radiotherapy. Women with scoliosis exposed to multiple diagnostic X-rays (mean, 12 cGy) during adolescence were at increased risk of breast cancer (Hoffman et al., 1989; Doody et al., 2000). The numbers were small, however, and factors associated with scoliosis, such as nulliparity, could have influenced risk. Breast cancer excesses have been reported following radiotherapy for Hodgkin disease (Travis et al., 2003b; van Leeuwen et al., 2003; Bhatia et al., 1996; Hancock et al., 1993) and breast cancer (Boice et al., 1992a). Interestingly, a breast cancer risk has been demonstrated at very high therapeutic doses (i.e., >10 Gy), although at a much lower level than predicted from low-dose studies (Travis et al., 2003b). Infants treated for thymic enlargement are at increased breast cancer risk later in life (Hildreth et al., 1989). Risk estimates derived from protracted exposures to treat hemangiomas in newborns were seven times lower than the acute exposures for thymic irradiation, suggesting an ameliorating effect when dose is spread over time (Lundell et al., 1996, 1999; Preston et al., 2002a). Radiation (mean, 2.8 Gy) to manage primary breast cancer appeared to result in the same age-specific RRs for secondary breast cancer as seen in other studies (Boice et al., 1992a), implying that radiation might interact with underlying host factors (e.g., late parity), in a multiplicative manner. Excess risk was apparent only among women treated before age 45. A concern over possible interaction between host factors and X-ray mammographic exposures does not appear warranted, however, because this procedure is usually recommended for women in midlife who are well past the ages of greatest breast tissue sensitivity to radiation carcinogenesis (Storm et al., 1992). Studies of radiotherapy during childhood do not indicate a multiplicative interaction with breast cancer risk factors such as age at first birth (Holmberg et al., 2001). It has been suggested that the ataxia-telangiectasia gene predisposes heterozygotes to radiogenic breast cancer (Swift et al., 1991), but the epidemiological and radiologic evidence to date provides little support for this possibility (Boice and Miller, 1992b; ICRP, 1978; Broeks et al., 2000; Shafman et al., 2000; Offit et al., 2002). Spine Irradiation Ankylosing Spondylitis. The mortality experience of 14,556 persons treated between 1935 and 1954 in 87 British radiotherapy clinics for ankylosing spondylitis, a rheumatoid condition of the spine, has been carefully evaluated through 1991 with 98% follow-up (Darby et al., 1987; Weiss et al., 1994, 1995). Radiation doses, estimated for each leukemia fatality and for a 7% sample of the population, averaged about 4.4 Gy for the active bone marrow, several gray for the esophagus, stomach, upper colon, pancreas, lung, and main bronchial tree, about 2.64 Gy mean total body dose, and substantially lower doses for organs not in the treatment field (Lewis et al., 1988). Some epidemiological evaluations focused on those patients, nearly half the total, who received a single course of treatment, typically 10 exposures over 1 month. Leukemia risk (60 observed vs. 21.5 expected) peaked 1 to 5 years after radiotherapy and gradually declined, but not to baseline levels; CLL was not significantly increased (7 observed vs. 4.8 expected). The

dose response for leukemia was irregular and essentially flat, possibly reflecting reduced leukemogenesis in the most heavily irradiated portions of the marrow due to cell killing or related to the fractionated nature of the exposures (Smith and Doll, 1982; Mole and Major, 1983). Compared to general population rates, the RR of leukemia can be estimated as 1.08 at 1 Gy and the absolute excess risk as 0.15/104 PY-Gy, although much higher estimates were suggested in the latest follow-up (Weiss et al., 1995). The dose-response relationship for radiation-induced leukemia was further examined in a pooled analysis with two other populations: Japanese atomic bomb survivors and women treated for cervical cancer (Little et al., 1999). A total of 383 leukemias were observed among 283,139 study subjects. Excluding chronic lymphocytic leukemia, the optimal relative risk model had a dose response with a purely quadratic term representing induction and an exponential term consistent with cell killing at high doses; the addition of a linear induction term did not improve the fit of the model. Experimental data in mice find a similar relationship between myeloid leukemia and radiation dose (Major and Mole, 1978). The relative risk in the combined analyses decreased with increasing time since exposure and with increasing attained age. Non-leukemia cancer deaths among the spondylitics were increased by 26%, with the excess occurring in the more heavily irradiated tissue such as the lung, esophagus, central nervous system, bone, nonHodgkin lymphoma, and multiple myeloma. An earlier reported excess of stomach cancer was no longer apparent. For non-leukemia cancers, the estimated RR was 1.14 at 1 Gy and the absolute excess risk was 4.67 per 104 PY-Gy. Unlike the experience of other irradiated populations, the RR for non-leukemia cancer was as great during the first 5 years after exposure as it was later, and then declined to nearnormal levels after 25 years. The temporal pattern was dominated by lung cancer, which may have been affected by variations in smoking habits, as well as the influence of cigarette smoke on relatively immobile lungs. Noncancer mortality, including benign lung conditions, was increased by 51%, and was attributed to conditions associated with spondylitis and not the radiation exposure. Some of the early cancer excesses may also reflect correlates of the underlying disease (e.g., ulcerative colitis is associated with both spondylitis and colon cancer), other therapies such as cytotoxic medications (Spiess et al., 1989), or possibly preexisting metastatic lesions causing pain that was misdiagnosed as ankylosing spondylitis. A study of 1201 nonirradiated patients with less severe disease, however, revealed no increased risk of leukemia or other cancers (Smith et al., 1977; Weiss et al., 1994). Historically, the study of patients treated for ankylosing spondylitis was a landmark investigation, which quantified leukemia risk in terms of dose to bone marrow. The last follow-up was through 1991 (Weiss et al., 1994). The mean total body dose was estimated as 2.64 Gy with the vertebrae receiving 18.6 Gy. For all cancers except leukemia the RR at 1 Gy was estimated to be between 1.11 and 1.18. Significant increases were seen for leukemia and cancers of the esophagus, pancreas, lung, bones, connective tissue, prostate, bladder, kidney, nonHodgkin lymphoma, and multiple myeloma. Radiation risk estimation is imprecise because of the absence of individual dosimetry on all but a sample. Organ dose could vary by several magnitudes depending on type of treatment.

Head and Neck Thymus. In the 1930s and 1940s, newborn children often received radiation therapy to shrink enlarged thymus glands. The fifth mail survey of 2856 irradiated persons identified 30 thyroid cancers versus 1 in 5055 untreated siblings (Shore et al., 1985). The followup was 88% complete. Females were at two to three times greater absolute risk than males, and the risk among Jews seemed especially high. The data were consistent with a linear dose response (mean, 1.2 Gy), risk remained high even after 40 years, and fractionation did not appear to reduce risk. Risk estimates were 2.9 cases/104 PY-Gy and RR = 9.90 at 1 Gy (Shore, 1992). Benign thyroid adenomas also occurred more frequently among exposed persons than among their siblings, 86 vs. 11, respectively (7.0/104PY-Gy). The RR at 1 Gy for nodules was 6.0 (Shore et al., 1993).

Ionizing Radiation The incidence of thyroid neoplasms rose abruptly during adolescence, suggesting the influence of thyroid stimulating hormone as a promoting or secondary factor. Childhood irradiation may also be particularly damaging if rapidly proliferating cells injured by radiation are more likely to develop abnormally than cells irradiated in later life with limited growth potential. Indeed, the rapid growth of the thyroid gland, from 1 to 2 g at birth to 18 g at maturity, may have influenced risk. Excess breast cancers have occurred, suggesting that the immature breast is also susceptible to the carcinogenic effects of radiation (Hildreth et al., 1989). Significant excesses of leukemia and cancer of the skin have been reported (Hempelmann et al., 1975; Hildreth et al., 1985).

Ringworm of the Scalp. Among 10,834 children in Israel who received X-ray therapy to the scalp for tinea capitis, 43 thyroid cancers were observed versus 11 expected based on two comparison groups (Ron et al., 1989). Cases were ascertained from tumor registry records and from searching pathology records of all major hospitals in Israel. The dose to the thyroid was particularly low, 9 cGy (rad) on average, although alignment errors might have greatly increased the exposure for some children (Schafer et al., 2001). The dose response was consistent with linearity up to 50 cGy. The absolute excess risk was 13 cancers/104PY-Gy, and the RR at 1 Gy was 31. Comparable numbers for benign tumors were 15 tumors/104PY-Gy and RR = 11 at 1 Gy. Risk was greatest among persons under age 5 at irradiation, and was most prominent 10 or more years later. The pattern of radiation risk over time was best described on the basis of a constant multiplication of the background rate, although few persons were observed for more than 30 years. Evidence for a low-dose effect is tempered by several possibilities: (1) careless irradiation techniques or restless children resulted in direct thyroid exposure, (2) radiotherapy given in countries of birth was missed, and (3) a remarkable difference in risk by ethnic origin suggested a genetic susceptibility or a bias in dose assessment. Recent analyses have attempted to account for some of the dosimetry uncertainties (Schafer et al., 2001; Lubin et al., 2004). A dose-response relationship for brain cancer and other neural tumors was reported (Ron et al., 1988b). The brain dose was estimated as 1.5 Gy. For all neural tumors (especially meningiomas), the absolute excess risk was 1.14 tumors/104PY-Gy and the RR at 1 Gy was 5.9. Significant excesses of leukemia and cancers of bone and connective tissue also occurred (Ron et al., 1988a). A preliminary report suggested that breast cancer might be elevated among children exposed at ages 5 to 9 years; however, the data are not easily interpreted since the excess resulted from a peculiar deficit among the comparison group rather than an elevation among the exposed, and no increase was seen among children exposed at younger or older ages (Modan et al., 1989; UNSCEAR 1994). Basal cell carcinomas of the skin, but not malignant melanoma, were significantly increased (mean, 7 Gy); the absolute excess risk was 0.31/104 PY-Gy and the RR at 1 Gy was 1.7 (Ron et al., 1991). Among 2224 children with tinea capitis treated in New York, no significant excess of thyroid cancer was found (2 observed), but thyroid adenomas, leukemia, and intracranial tumors were elevated (Shore et al., 1976, 2003). Cancer ascertainment was by questionnaire and a comparison group of 1380 patients were given topical medication to treat ringworm. Radiotherapy likely contributed to excess skin cancers, especially for anatomical areas exposed to ultraviolet radiation from the sun (Shore et al., 1984, 2002). Basal cell carcinomas of the face were significantly increased among white but not black patients, indicating the importance of susceptibility to UV as a cofactor. Patients with psoriasis treated with 8-methoxy-psoralen and longwave ultraviolet radiation (PUVA) also have been found to develop cancer in skin previously treated with low-energy X-rays (Stern et al., 1984, 1998). Tonsils. Excess thyroid cancer has occurred among 5379 predominantly Jewish persons irradiated in Chicago during childhood for mostly tonsil and nasopharyngeal conditions (Favus et al., 1976; Schneider et al., 1985). The tracing was about 68% complete. Intensive clinical screening of 1922 persons included thyroid scans. About 37.5% (1108) of the 3610 persons for whom vital status was known

263

had nodular thyroid disease, and in 297 of these (or 8% overall) it was malignant. A large number of small tumors were apparently detected only at screening; about 80% were less than 1.5 cm. The clinical significance of these small tumors is uncertain and they may be relatively harmless (Crile et al., 1979). An evaluation of individual dose estimates for 3843 subjects revealed a linear relationship between thyroid dose and cancer, radiation risk was inversely related to age at exposure, excess RR for men and women were similar, and RR decreased after 30 years of follow-up (Schneider et al., 1993). While overall rates of thyroid cancer dramatically increased after 1974 when screening programs began, the estimates of radiation risk did not vary significantly compared with those obtained before 1974 (Ron et al., 1992). Interpretation of radiation risks is hindered for several reasons: (1) a nonirradiated comparison group, comparably screened, was not available, (2) the significance of small, clinically silent cancers is uncertain, and (3) the follow-up was incomplete. The dose-response evaluations and consistency with other studies, however, support the reported results. Based on clinical examinations, an excess of thyroid nodularity was reported among 1590 individuals treated with radiation for lymphoid hyperplasia compared with 1499 persons treated with surgery only (Pottern et al., 1990). There was a strong dose-response gradient (mean, 24 cGy), similar RR at 1 Gy for males and females (8.0 and 7.0, respectively), and an inverse relationship with age at exposure. Much higher risks were suggested, however, by self-reported conditions from a mailed questionnaire, apparently due to underreporting of thyroid nodularity among the surgical comparison groups.

Hemangiomas. In Sweden in the 1920s through the 1950s, tens of thousands of infants with skin hemangiomas were treated with radium-226 or X-rays (Lundell et al., 1994, 1999; Furst et al., 1988). For those children with hemangiomas near the neck, a significant increase in thyroid cancer was found. The average dose was 1 Gy. Interestingly, compared with the studies of acute exposure for thymic irradiation during infancy, the protracted exposures from radium-226 plaques were about three times lower (UNSCEAR, 2000) and suggested that low doses at a low rate of delivery are less carcinogenic, perhaps related to repair. For those children with hemangiomas on the chest region, an elevated risk of breast cancer was found. The doses to breast tissue could reach several Gy among children less than 1 year of age when treated. Again, the risks were about seven times lower than those from acute exposures to treat thymic enlargement (Preston et al., 2002a). Other treatment sequela included breast hypoplasia. This study confirms that very young children are susceptible to radiationinduced breast cancer with the malignancy appearing many decades after exposure (i.e., even 50 years later). Other Head and Neck. The pattern of thyroid cancer incidence in birth cohorts in Connecticut appeared to coincide with the widespread use of radiation to treat benign head and neck conditions between 1920 and 1959; rates were also lower for persons born in the 1960s when such irradiation was discouraged (Pottern et al., 1980). Prior radiotherapy in childhood may account for 9% of all thyroid cancers (Ron et al., 1987). Excess salivary gland and neural tumors can occur after childhood irradiation (Schneider et al., 1998; Land, 1986). A series of 18,030 children treated with radiation for skin hemangioma identified increased rates of thyroid cancer and softtissue sarcoma (Furst et al., 1988). Persons treated with nasopharyngeal radium applicators to shrink lymphoid tissue around the eustachian tube were not found to be at significant increased risk of thyroid or other cancers (Hazen et al., 1966; Ronckers et al., 2001, 2002a, 2002b), although a small non-significant excess of brain tumors was reported following treatments in childhood (Yeh et al., 2001). A comprehensive pooled analysis of studies of radiogenic thyroid cancer found that linearity best described the dose response; a downturn or leveling of risk appears at very high doses (>10 Gy); risk was greatest for childhood exposures (RR = 8.7 at 1 Gy); the attributable risk at 1 Gy was 88%; and spreading dose over time appears to lower risk, possibly due to cellular repair processes (Ron et al., 1995). Practically all studies of thyroid irradiation, whether from external exposures or

264

PART III: THE CAUSES OF CANCER

Figure 15–2. Characteristic wavelike pattern of leukemia risk over time since exposure, seen among women with cervical cancer treated with radiation. (Source: Boice et al., 1985, J Natl Cancer Inst 74:955–975, p. 966.)

internal radio iodines, fail to find a significant increased risk when exposures occur after age 20 (Inskip, 1999). The one exception is a study of 8144 patients treated for painful arthritic conditions of the cervical spine with radiotherapy (Damber et al., 2002). Thyroid dose was of the order of 1 Gy and a low, but marginally significant excess risk was observed (SIR 1.60, n = 22).

Pelvic/Abdominal Irradiation Cervical Cancer. In an international study involving 31 radio-

phocytes also revealed a dose-effect pattern similar to that seen for leukemia (Kleinerman et al., 1989). A subsequent international cancer registry study indicated that the risk of second primary cancers remains high for the duration of life (Kleinerman et al., 1995). After a minimum latent interval of about 10 years, risk of second tumors following radiotherapy increased with time and reached nearly twofold among long-term survivors (Fig. 15–3). Overall, the excess RR for all second cancers was only about 10% (Boice et al., 1985a). A comprehensive dosimetry program provided organ dose estimates for individual patients based on their actual radiotherapy records. Dose-response information was provided for 18 solid tumors, including cancers of the stomach, uterus, rectum, bladder, vagina, kidney, ovary, thyroid, and breast (Boice et al., 1988b). Risk estimates for this study of incident cancers and leukemia were generally lower than those from the mortality studies of atomic bomb survivors and spondylitics (UNSCEAR, 1988). Possibly the protracted nature of the treatment for cervical cancer allowed more time for the repair of radiation damage than is possible from acute exposures; the high therapeutic doses might have resulted in substantial cell killing; or factors associated with cervical cancer might have confounded the observations. Interestingly, the excess number of cancers in the international cervical cancer study is larger than observed in the study of atomic bomb survivors (Boice et al., 1985; Preston et al., 2003). Among women with intact ovaries, radiotherapy was linked to a significant 35% reduction in breast cancer risk, attributable in all likelihood to the cessation of ovarian function (Boice et al., 1989). Radiotherapy lowers estrogen and androgen levels, even among postmenopausal women, which may contribute to the low breast cancer risk seen among women exposed past the age of menopause (Inskip et al., 1994a). Recent studies of young women treated for Hodgkin disease indicate a substantial reduction in breast cancer risk following ablation of the ovaries by chemotherapy or radiotherapy (Travis et al., 2003b; van Leeuwen et al., 2003).

therapy centers, over 30,000 women with cervical cancer were observed clinically and with blood studies for up to 10 years (mean, 5 years). Despite large radiation doses (5–15 Gy) to the pelvic bone marrow, no excess leukemia or lymphoma was observed (Boice and Hutchison, 1980). This study served as the basis for an expanded survey of more than 200,000 women from 15 countries (Boice et al., 1985a, 1987). The international cervical cancer study is second only to the study of atomic bomb survivors in terms of size, excess cancers, dosimetry program, statistical methodology, biochemical components, number of personnel, and length of follow-up. A small but significant risk of leukemia then became apparent, together with the characteristic wavelike pattern of risk over time (Fig. 15–2). Risk was modeled to account for the inhomogeneous distribution of dose to active bone marrow throughout the body, and for the possibility of cell killing at high doses. Risk increased with increasing dose up to about 4 Gy, and then decreased at the highest doses. The RR at 1 Gy was estimated to be 1.7. Evaluations of chromosome aberrations in circulating lym-

Uterine Bleeding. Among 2067 women who received X-ray treatment of 5–10 Gy to their ovaries for metropathia hemorrhagica in Scotland, 12 leukemia deaths occurred (5.9 expected) following an estimated bone marrow dose of 1.3 Gy (Darby et al., 1994). Cancers of heavily irradiated pelvic sites were also significantly increased, including the colon and bladder, but not cancers of the ovary and rectum. Breast cancer occurred significantly below expectation, even among postmenopausal women, possibly related to the cessation of ovarian function after radiation castration. Similar findings were reported from a Swedish investigation of 788 women irradiated as early as 1912 (Ryberg et al., 1990a). An increase in heart disease in both studies may be related to the cessation of ovarian function

Figure 15–3. Characteristic pattern of radiation-induced solid tumors over time since exposure, seen for heavily irradiated sites among cervical

cancer patients treated with radiation. (Source: Boice et al., 1985, J Natl Cancer Inst 74:955–975, p. 964.)

Ionizing Radiation (Ryberg et al., 1990b). An incidence study of 1893 women given radiotherapy for benign gynecologic diseases in Connecticut also found significant excesses of leukemia (12 vs. 5.3), uterine sarcomas, cancers of urinary organs, and lymphomas, including myeloma (Wagoner, 1984). In a subsequent study of 9770 women treated mainly with radium implants for bleeding disorders in New England, 64 leukemia deaths occurred and 39.6 were expected (Inskip et al., 1993). A comparison group of 3185 women was also studied. The RR at 1 Gy was estimated to be 2.9; the mean dose to bone marrow was 1.11 Gy. The lower doses used to treat benign menstrual disorders appear to have been more leukemogenic than the higher doses used to treat cervical cancer, presumably because of less cell killing (Kleinerman et al., 1994). Cancers of heavily irradiated sites were also significantly elevated, including the colon, uterus, and bladder, but not the rectum or cervix, lymphomas, or myeloma (Inskip et al., 1990; 1993). Only the excesses of leukemia and cancers of the bladder and uterus were consistent across these three studies. Inconsistencies might be related to different dose distributions within organs from the different radiation modalities (radium implants vs. external X-rays), differences in the extent of surgical procedures or presenting conditions, differences in the comparison populations, or simply chance.

Peptic Ulcer. In a survey of 1831 patients with peptic ulcer treated with radiation (15 Gy) and 1778 non-exposed ulcer patients, significant increases were reported for cancers of the stomach, pancreas, lung, and leukemia (Griem et al., 1994; Carr et al., 2002). The excess of pancreatic cancer was attributed to possible miscoding of stomach cancers on death certificates. An increase in heart disease was attributed to the selection of less healthy patients for radiotherapy over surgery. Radiation combined with surgery appeared to induce carcinogenic processes that greatly enhanced the development of stomach cancer, possibly mediated by hypoacidity and bile reflux. Estimated RR at 1 Gy for stomach and lung cancers were 1.15 and 1.66, respectively. Corresponding excess risks were 0.25 and 2.33/104 PY-Gy. A substantial difference in absolute (but not relative) risk estimates for stomach cancer between ulcer patients and A-bomb survivors points to the difficulty in generalizing from one population to another when baseline disease rates differ appreciably. Infertility. Another new series evaluated mortality among 968 Israeli women given radiotherapy for infertility (Ron et al., 1999). Doses were, on average, 1.0, 0.8, 0.6, and 0.4 Gy to the ovary, brain, colon, and bone marrow, respectively. No unusual cancer patterns were observed, in all likelihood due to the small study size. These findings were consistent with an earlier investigation among 816 women similarly treated in New York City (Ron et al., 1994a). Prenatal Irradiation Studies of fetal exposures have been interpreted as evidence that low doses of ionizing radiation (<10 cGy) can cause childhood cancers (Brenner et al., 2003; Doll and Wakeford, 1997). The evidence, however, is not entirely convincing (Boice and Miller, 1999; UNSCEAR, 1994; ICRP, 2003b). While there are many epidemiologic studies reporting a small increased relative risk, they are all casecontrol studies and no cohort study finds a significant risk. Further, the largest case-control study, the Oxford Survey, reports the same relative risk for practically every childhood cancer, whether acute lymphocytic leukemia or Wilms tumor, which suggests strongly some underlying bias. No study has been able to reconstruct individual fetal doses, except the study of the atomic bomb survivors pregnant at the time of exposure, which failed to find an association with childhood leukemia or cancer; and a small cohort study of 1000 women examined during early pregnancy on a routine basis (i.e., not selected), which reported fetal doses of between 1.5 and 3 cGy (rad) based on measurements and calculations and also found no association with leukemia or childhood cancers (Griem et al., 1967; Oppenheim et al., 1974). A recent nation-wide nested case-control study in Sweden, which relied upon medical records to determine prenatal X-ray exposure, rather than interview, also failed to find a significant elevation in childhood leukemia, although sampling variability was such that a

265

small risk could not be excluded (Naumburg et al., 2001). It is likely that the question of whether low-dose prenatal exposure is causally related to excess cancers will remain unanswered, but a causal assumption seems prudent for the purposes of radiation protection (MacMahon, 1985, 1989; ICRP, 2003b). Most, but not all, case-control studies are consistent with a 40% increased risk of childhood leukemia associated with low-dose intrauterine exposure to diagnostic radiation of between 1 and 10 cGy just before birth (Stewart et al., 1958; MacMahon, 1962; Bithell and Stewart, 1975; Monson and MacMahon, 1984). These studies have been extensively reviewed (UNSCEAR, 1972, 1986, 1994; NAS, 1972, 1980; Miller and Boice, 1986; Doll and Wakeford, 1997; Boice and Miller, 1999). It has been postulated, however, that selection factors, related to the medical reasons why women receive prenatal Xrays, and/or response bias were responsible for the increased leukemia risk and not the X-ray exposures themselves. The absence of any childhood leukemia (and only one childhood cancer) in atomic bomb survivors exposed in utero (Jablon and Kato, 1970) supported the possibility of bias, as did Miller’s observation (1969) that it was peculiar that diagnostic X-rays would increase all childhood malignancies by about the same percentage (50%) when there is such a remarkable degree of variability between tissues in their response to radiation at other ages. A response bias seems likely since cases apparently overreported the number of prenatal X-rays received, controls apparently under-reported the number of prenatal X-rays received, and associations with X-rays received before pregnancy were apparent (ICRP, 2003b). An increase in childhood lymphomas is also peculiar in that lymphoma is not convincingly linked to radiation in any human study (Boice and Miller, 1999; Miller and Boice, 1986). Further, many childhood cancers are primarily embryonal and developmental in origin, which are not known to be induced by radiation; and most pelvimetry X-rays occur just prior to birth and after the time when embryonic neoplasms have been initiated (Miller, 1995; Boice and Miller, 1999). Animal experiments do not suggest an enhanced sensitivity to leukemia induction following irradiation during fetal stages (UNSCEAR, 1986; ICRP, 2003b). Although the arguments fall short of being definitive because of the combination of biological and statistical uncertainties involved, they raise a serious question of whether the great consistency in elevated RRs, including embryonal tumors and lymphomas, may be due to biases in the Oxford Survey study rather than a causal association (ICRP, 2003b). The indication of a leukemia risk for preconception irradiation in one study (Graham et al., 1966), when no genetic effects were noted in the much larger A-bomb survivor study (Neel and Schull, 1991), and the finding of an excess risk in white, but not black, children prenatally exposed (Diamond et al., 1973), further suggested that fetal Xrays might just be an indicator of a poor future health experience. A small prospective study in Chicago evaluated about 1000 unselected children whose mothers received X-ray pelvimetry as a matter of hospital policy, and not medical indications, and no excess cancers were found (Oppenheim et al., 1974). Court Brown and coworkers (1960) studied nearly 40,000 children irradiated in utero and observed nine cases of leukemia versus an expected number of 10.5. The sample sizes of the prospective studies, however, were such that an increased risk of 40%–50% could not be excluded. Several large cohort studies of twins, however, also fail to find childhood leukemia to be increased (Boice and Miller, 1999). A New England study was extended to include 1342 childhood cancers among 1,429,499 births (Monson and MacMahon, 1984). The RR associated with prenatal X-ray was 1.52 for leukemia and 1.27 for other cancers, and there was no evidence that the associations were due to confounding. A reanalysis of the large Oxford Survey of childhood cancer in England concluded that X-raying 1000 fetuses with 1 cGy (1 rad) would yield about two or three extra cases of childhood cancer in the first 15 years of life (Bithell and Stiller, 1988). An early report of increased adult cancer following fetal irradiation of A-bomb survivors was not substantiated with further follow-up (Yoshimoto et al., 1994). An even fuller account of the findings through age 46 was recently reported (Delongchamp et al., 1997). Interpretation remains equivocal, in large part because the sample size is small with only 10

266

PART III: THE CAUSES OF CANCER

cancer deaths occurring among the in utero exposed. While the radiation risks appear compatible between the prenatal exposed group and the children aged 0 to 5 years at exposure, there are several biological inconsistencies. The risk of leukemia was inversely related to dose, two of the eight solid cancers are of types not known to be inducible by radiation, and two others followed very low dose, 0.10 Sv. Evidence against bias to explain completely the prenatal X-ray associations comes from the demonstration of a dose-response relationship for childhood leukemia based on number of X-ray films taken (Stewart and Kneale, 1970); and from the observation that the excess risk was as great among twins for whom X-ray pelvimetry was far more frequent (55%) than among singletons (15%) simply because of a greater likelihood to determine fetal positioning before delivery (Mole, 1974). This observation was confirmed in a case-control study of twins born in Connecticut (Harvey et al., 1985). Nonetheless, it is argued that number of X-rays is not necessarily equivalent to fetal dose, and that twin studies are somewhat difficult to interpret. For example, despite substantial exposure to prenatal X-rays, cohort studies consistently find twins to be at significantly low risk of childhood cancer compared to single births (UNSCEAR, 1986, 1994; Inskip et al., 1991; Boice and Miller, 1999). Further, while there is no reason to believe that the fetus should be immune to the leukemogenic effects of ionizing radiation, there is also little reason to believe that the risk should be substantially greater for exposures just before birth than for exposures in early childhood. Finally, it is paradoxical that the three largest data sets on radiation-induced leukemia find no evidence for an excess risk of any type of leukemia below 20 cGy with a quadratic dose-response most consistent with the data for each histologic type (Little et al., 1999; UNSCEAR, 2000), and yet it is fetal doses of the order of 1–2 cGy for which a prenatal association is found. Thus, while it is established that prenatal X-ray is associated with an increased risk of childhood leukemia, the magnitude of the hazard, and even the causal nature of the cancer association, remain uncertain (UNSCEAR, 1986, 1994; MacMahon, 1989; ICRP, 2003b).

among 484 persons with thyroid cancer and matched controls revealed no association between X-rays to the head and neck and thyroid cancer (Inskip et al., 1995). Childhood leukemia and postnatal X-rays were investigated in two case-control interview studies: a Canadian study of 491 cases of acute lymphocytic leukemia reported an association, with a suggestion that polymorphisms in repair genes might modify risk (Infante-Rivard et al., 2000), whereas a larger interview study in the United States of 1841 cases found no link between leukemia and diagnostic X-rays (Shu et al., 2002). Past exposures to dental or medical radiography of the head and neck have been correlated with meningiomas, gliomas, and salivary gland tumors in some case-control interview studies (Preston-Martin et al., 1980, 1982, 1988) but not in others (Kuijten et al., 1990). Radiation doses were not known, but might have been substantial. In one investigation, an association between skull X-rays and brain cancer was thought to be due to early symptoms of brain cancer prompting the radiographic examinations (Howe et al., 1989). Multiple fluoroscopic chest X-rays appear to increase the risk of breast cancer, but not lung cancer or leukemia among tuberculosis patients (Davis et al., 1989; Howe, 1995), and not cancer or leukemia among children undergoing heart catheterization (McLaughlin et al., 1993a; Modan et al., 2000). High numbers of spinal X-rays to monitor scoliosis in childhood were linked to increases in breast cancer, but not leukemia (Doody et al., 2000). Limitations of many of these studies include the potential for response bias in interview surveys; incomplete verification of the actual numbers of X-rays; limited dosimetry; and study sizes too small to detect risks on the order of currently accepted estimates (cf Boice and Land, 1979; Ron, 2003). The possible contribution of diagnostic radiology to the cancer burden appears small in comparison with other causes (Evans et al., 1986; Harvard, 1996; Berrington de González and Darby, 2004).

General Diagnostic Radiation

Radiotherapy for Cancer Adult Treatments. The most serious consequence of curative

Studies linking diagnostic radiation with adult leukemia are inconsistent and are complicated by the inherent difficulty in estimating radiation doses when procedures and exposures change frequently over time (Preston-Martin and Pagoda, 2003). Difficulties are further compounded when exposure determination is based solely on interviews and not actual medical records (Berrington de Gonzalez et al., 2003; Ron, 2003). Further, it is often not clear whether the X-ray was performed because of early symptoms for conditions not yet diagnosed. For example, a report from England was later retracted when the author attributed the concentration of X-rays within 5 years of the leukemia diagnosis to symptoms related to preclinical disease, including an increased susceptibility to infection (Stewart, 1973). Excesses of chronic myelogenous leukemia (CML) in some studies appeared restricted to those who received extremely large numbers of X-rays (Gibson et al., 1972). A study at the Mayo Clinic, which included accurate estimates of bone marrow doses, found no link between leukemia and diagnostic X-rays but the numbers were small (Linos et al., 1980). Another hospital-based case-control study in Japan reported no association between nearly 300 adult leukemias and diagnostic X-rays, but exposure assessment was based on questionnaire responses (Yuasa et al., 1997). A report from California found an association between diagnostic radiography, particularly low-back X-rays, and CML based on personal interviews of 136 cases and 136 neighborhood controls (Preston-Martin et al., 1989). Based on X-ray records with two prepaid health plans, however, a subsequent study concluded that persons with leukemia and lymphoma are X-rayed frequently just before diagnosis for conditions related to the development or natural history of their disease (Boice et al., 1991b). The possibility of small increases in myeloma could not be discounted among persons who received rather extensive X-ray exposures. Diagnostic X-rays, however, were not linked to multiple myeloma in an interview study of 399 cases and 399 controls in the United Kingdom (Cuzick and DeStavola, 1988) or in a study of 540 cases and a similar number of controls in the United States (Hatcher et al., 2001). An evaluation of medical X-ray records

therapies for cancer is the heightened risk of developing a new cancer (Boice et al., 1985b; Boice, 1993a; Neugut et al., 1999; Little, 2001a; van Leeuwen and Travis, 2001; IARC, 2000; Meadows, 2003). Leukemia has been linked to high-dose radiotherapy, but to a lesser extent than seen in patients treated with lower doses for nonmalignant diseases (Curtis et al., 1985; Boivin et al., 1986). Large international studies have revealed twofold leukemia risks following radiotherapy for cervical and uterine cancer, whereas much higher risks were predicted based on simple linear risk extrapolation (Boice et al., 1987; Curtis et al., 1994; Kleinerman et al., 1995). When such high doses are delivered to small volumes of tissue, cell killing likely predominates over cell transformation and overall leukemia risk is reduced. A variety of risk coefficients per Gy has been observed among medical studies of partial-body exposures (Fig. 15–4). Past treatments with systemic chemotherapy and radiotherapy together appeared to enhance the risk of leukemia over 17-fold among patients with breast cancer (Curtis et al., 1992). Increased leukemia risks have also been reported after radiotherapy for Hodgkin disease (Tucker et al., 1988; Kaldor et al., 1990) and non-Hodgkin lymphoma (Travis et al., 1991, 1994). Total or hemibody irradiation for non-Hodgkin lymphoma (NHL), a unique treatment that exposes large volumes of bone marrow to relatively low therapeutic doses, was seen to increase leukemia risk, although chemotherapy may also have played a role (Greene et al., 1983; Travis et al., 1996). Total or hemibody irradiation prior to bone marrow transplant has been linked to increases in solid cancers (Curtis et al., 1997; Socie et al., 2000). Cancers of the lung and breast are elevated after radiotherapy for Hodgkin disease (Tucker et al., 1988; van Leeuwen et al., 1989, 1995, 2003; Swerdlow et al., 2000; Hancock et al., 1993; Travis et al., 1995a, 2002, 2003, 2005; Gilbert et al., 2003). Cigarette smoking and high lung doses enhanced the risk of lung cancer in a multiplicative fashion among women treated for Hodgkin disease (Gilbert et al., 2003), whereas an interaction with alkylating agents appeared additive (Travis et al., 2002). Non-Hodgkin lymphoma risk is increased after

Ionizing Radiation

Figure 15–4. Relative risk of leukemia by bone marrow dose for atomic bomb survivors (NAS, 1990), and women treated for benign gynecologic disease (BGD) menstrual bleeding (Inskip et al., 1993), and cancers of the cervix (Boice et al., 1987), endometrium (Curtis et al., 1994), and breast (Curtis et al., 1992). (IARC 2000, p. 224.)

Hodgkin disease (Travis et al., 1991) and cervical cancer (Boice et al., 1988b), but it is possibly related to immune deficiencies. Only very high doses of radiation seem to elevate the risk of cancers of the rectum and uterus and sarcomas of the bone and soft tissues (Boice et al., 1988b; UNSCEAR, 2000; IARC, 2000). Excess bladder cancer has been reported after 10 Gy, but not colon or liver cancer (Boice et al., 1988b). Second breast cancer has been linked to radiotherapy for primary breast cancer, but only among women less than 45 of age at exposure (Boice et al., 1992a). Lung cancer has been recognized as a late effect following radiotherapy for breast cancer (Inskip et al., 1994b; Inskip and Boice, 1994; Travis et al., 1995b; Neugut et al., 1994; Rubino et al., 2003). Significantly low rates of breast cancer can follow ovarian doses greater than about 6 Gy (Boice et al., 1989; Travis et al., 2003b). Treatment for brain cancer is linked to significant increases of second tumors, with children showing higher risks than adults (Inskip, 2003). Many studies, however, used general population rates for comparison, which may not be appropriate if the underlying disease predisposes to second cancers. Other treatments, such as alkylating agents or other drugs of high toxicity, might also influence subsequent cancer risk (Boice, 2001c). Only a few studies attempted to quantify risk in terms of radiation-absorbed dose to organs (Boice et al., 1988b; Travis et al., 2002, 2003). It appears that radiotherapy may account for only 5%-10% of second cancers among cancer patients, with cigarette smoking, alcohol, hormonal factors, chemotherapy, and other cofactors playing more important roles (Boice et al., 1985a, 1985b). Because most of the classic studies in radiation epidemiology are coming to an end due to the death of most members in the exposed cohorts, studies of cancer patients offer continued opportunities to enhance our knowledge of low-dose radiation effects. Scatter doses to organs outside the treatment beams are low, the numbers of exposed patients are large, and the dosimetry can be accurately performed. Risk estimates from radiotherapy studies are generally much lower than those from the atomic bomb investigation (Little, 2001a, 2001b).

Childhood Treatments. Children treated for cancer are at high risk for developing new cancers (Tucker et al., 1984; Hawkins et al., 1987; Breslow et al., 1988; de Vathaire et al., 1989; Little et al., 1998b; Garwicz et al., 2000; Neglia et al., 2001; Mertens et al., 2001). Only a few series, however, have estimated radiation doses to specific organs, or evaluated dose-response relationships. Radiotherapy was not found to increase the risk of leukemia in one study (Tucker et al., 1987b), possibly because of the predominance of cell-killing effects over oncogenic transformation at such high levels. A more recent study reported a leukemia risk following radiotherapy (Hawkins et al.,

267

1992), possibly because of interactive effects with chemotherapeutic agents. Radiogenic thyroid cancer has been reported at 30 Gy (de Vathaire et al., 1988, 1999; Tucker et al., 1991) with a downturn in risk at higher doses (Sigurdson et al., 2005). A dose response over the range of 10–60 Gy has been seen for bone cancer (Tucker et al., 1987a). Interestingly, the RR for radiogenic bone cancer in children with retinoblastoma (who possess an underlying predisposition to develop osteosarcoma) was similar to that among children irradiated for other cancers; cumulative and absolute risks, however, were much greater among children with retinoblastoma. Children with hereditary retinoblastoma have a deletion in chromosome 13 that predisposes to osteosarcoma, and radiation appears to cause a second mutation in an osteoblast that leads to a high rate of osteosarcoma development (NAS, 1990). A survey of 1602 children with retinoblastoma revealed significant increases in cancers of the bone, connective tissue, brain, and skin melanoma for which radiotherapy appeared to further enhance the inborn susceptibility to cancer development (Eng et al., 1993; Kleinerman et al., 2004). A dose response for soft-tissue sarcomas was reported following high-dose radiotherapy for retinoblastoma (Wong et al., 1997). Radiogenic bone cancers have been reported after childhood cancer in other series (Draper et al., 1986; Hawkins et al., 1996; Le Vu et al., 1998), and among children treated for Ewing sarcoma (Strong et al., 1979). Children treated for medulloblastoma who have basal cell nevus syndrome develop multiple basal cell carcinomas in irradiated skin (Strong, 1977). Children with leukemia treated with cranial irradiation also are at high risk for developing brain malignancies (Neglia et al., 1991; Walter et al., 1998; Little et al., 1998a), with genetic susceptibility playing a possible interactive role (Relling et al., 1999). Children with Li Fraumeni syndrome, involving an inherited defect in the p53 gene, also appear to be at high risk of secondary cancers following radiotherapy (Hisada et al., 1998). Because children treated for cancer are now surviving for many years, late effects are of special concern (Meadows, 2003; Schwartz, 2003). Accordingly, radiation treatments have been modified to lower dose regimens and even excluded altogether (Donaldson et al., 1999).

Military Exposures from Atomic and Thermonuclear Weapons Japanese Atomic Bomb Survivors The Life Span Study (LSS) of the Radiation Effects Research Foundation (RERF) includes about 93,000 atomic bomb survivors and 27,000 non-exposed comparison subjects. Recent analyses of cancer mortality cover 1950 to 2000 (Preston et al., 2003, 2004). The first comprehensive report on cancer incidence was completed based on data from the Hiroshima and Nagasaki Tumor Registries during the period 1958 to 1987 (Thompson et al., 1994). These and other recent studies of cancer risk among A-bomb survivors, reflect increases in numbers of cancer cases associated with the natural aging of the population, notably among the youngest survivors, who appear to be at greatest relative risk of radiation-associated cancer. The findings also reflect modifications in dosimetry, and advances in statistical methods for the analysis of cohort survival data that facilitate modeling of age, sex, time since exposure, and other cofactors as modifiers of radiation dose-response relationships (Preston et al., 2003, 2004; UNSCEAR, 2000; Pierce and Preston, 2000; Pierce et al., 1996).

The Latest Atomic Bomb Dosimetry. In the early 1980s, the accuracy of the Tentative 1965 Dosimetry (T65D), used to estimate doses for individuals, was questioned. A new dosimetry, called Dosimetry System 1986 or DS86 was developed, and revised estimates were computed for 86,000 of the 93,000 exposed survivors in the LSS, called the DS86 sub-cohort. Under reasonable assumptions about the relative biological effectiveness (RBE) of neutrons, risk estimates based upon the new DS86 dosimetry were 1.5 to 2 times greater than those under the old T65 dosimetry (Pierce et al., 1996). Subsequent neutron activation measurements of metal and other minerals in Hiroshima suggested that the DS86 dosimetry may have underestimated the neutron exposure, and a new dosimetry system was

268

PART III: THE CAUSES OF CANCER

developed; however, the impact is likely to be small (Straume et al., 2003; Preston et al., 2003, 2004). The new dosimetry, called Dosimetry System 2002 or DS02 has just recently been incorporated into the new follow-up data (Preston et al., 2004). Some uncertainties in applying the risk estimates to other populations are related to extrapolating high-dose risks to low-dose exposure circumstances, applying a Dose and Dose Rate Effectiveness Factor (DDREF) to account for the differences between brief versus prolonged exposures, and assumptions pertaining the Relative Biological Effectiveness (RBE) of neutrons for specific organs, rather than using the colon dose as surrogate, and whether the RBE varies by dose (Mossman, 2003; Baker and Hoel, 2003; Kellerer et al., 2002; Kellerer, 2002b; Kellerer and Nekolla, 1997).

Leukemia, Multiple Myeloma, and Malignant Lymphoma. Leukemia was the first radiation-induced cancer reported among A-bomb survivors, with the risk peaking within 10 years of exposure (Moloney and Kastenbaum, 1955; Brill et al., 1962; Ichimaru et al., 1986; Preston et al., 1994). The dose response for leukemia appeared to follow a linear-quadratic relationship with some flattening at doses over 3–4 Gy (Pierce et al., 1996). A quadratic (Fig. 15–5) and even a threshold response appears consistent with the data depending upon assumptions made (Kellerer, 2002b; Baker and Hoel, 2003; Little et al., 1999; Chomentowski et al., 2000; Kellerer and Nekolla, 1998). Differences between Hiroshima and Nagasaki were no longer significant, although there were intercity differences in the frequency of chromosomal translocations in the level of response to any given dose (i.e., always higher for Hiroshima) (Kodama et al., 2001). Based on mortality data, the overall RR at 1 Gy was 5.62, and the absolute excess per 104 PY-Gy was 2.61 (Pierce et al., 1996). Multiple myeloma was significantly increased based on mortality (Pierce et al., 1996) but not incidence data (Preston et al., 1994). Malignant lymphoma was not related to radiation exposure. Radiation-related risks were seen for acute lymphocytic and myeloid leukemias, and chronic myeloid leukemia (CML), but not chronic lymphocytic leukemia (CLL) or adult T-cell leukemia (Ichimaru et al., 1986; Preston et al., 1994; Little et al., 1999; UNSCEAR, 2000). A sharp peak in CML occurred within 5 years of exposure, notably among children under 15 at time of bombings (ATB), but also at older ages (Ishimaru, 1979). Excess risks have declined since this early peak. In absolute terms, CML risks in Hiroshima are much larger than those in Nagasaki. Based on these differences, it was suggested that CML may be a neutron-dependent cancer (Ishimaru, 1979; Mole, 1975). However, CML was much less common among unexposed survivors in Nagasaki than in Hiroshima, and, in relative terms, there was no difference in CML risk. For acute leukemias, primarily myeloid, the temporal pattern of the excess risk depended on age at exposure. Among survivors under 20 when exposed, absolute excess risks peaked within 10 years and then

Figure 15–5. Observed and modelled relative risk of acute myeloid, acute lymphocytic, and chronic myeloid leukemia in a combined analysis of data for the Japanese atomic bomb survivors, women treated for cervical cancer, and patients treated for ankylosing spondylitis. The values are spe-

fell rapidly. For ages 20 to 35, the peak was less pronounced and the fall less rapid. For older survivors, the excess varied little over time. For all leukemias combined, males had almost twice the absolute excess risk as females, but similar RR.

Breast. Female breast cancer has been comprehensively studied (Land et al., 2003; Preston et al., 2003). The dose response was linear in both cities, and age at exposure and attained age strongly influenced risk. Risk was highest for exposures under age 20, although there was little variation under this age associated with exposures during infancy or menarche. Risk declined with increasing age at exposure and was low among women exposed after age 40. Excess risk did not appear until 10 years after exposure and not before about age 30. Measures of RR remained roughly constant over time after exposure, within age cohorts, with the notable exception of early-onset breast cancer before age 35 among women exposed before age 20 whose risk (RR = 14.5 at 1 Sv) was sufficiently high to suggest a possible genetically susceptible subgroup (Land et al., 1993a, 2003). Estimates of risk at 1 Gy are roughly 2.6 for the RR, and 6.7 excess cancers/104 PY-Gy (Thompson et al., 1994). Interestingly, an early age at first birth, parity, and lactation appeared to protect against both baseline and radiationinduced breast cancer (Land et al., 1994). Lung. Excess lung cancers of all major types have been reported, including adenocarcinoma, squamous cell carcinoma, and small-cell carcinoma (Yamamoto et al., 1987; Thompson et al., 1994). Among A-bomb survivors and uranium miners, it appears that the typical radiogenic lung cancer is small cell and the atypical is adenocarcinoma (Land et al., 1993b). Except for the absence of effect among those exposed under 10, age at exposure had little influence on the level of risk once smoking history was take into account. An apparent anomalous increase in risk with increasing age at exposure was attributed to birth cohort variations in lung cancer rates (i.e., related to variations in the smoking effect seen predominately in males) (Pierce et al., 2003). Females had higher RR than males; however, this difference was substantially reduced when adjusted for smoking history. The interaction between smoking history and radiation has recently been determined to be consistent with additivity (Pierce et al., 2003). The RR at 1 Gy was 1.54, and the absolute excess was 1.93/104 PY-Gy (Pierce et al., 1996). Thyroid. Thyroid cancer was the first solid tumor reported to be increased among A-bomb survivors (Wood et al., 1969). Subsequent surveys found higher background rates among those who receive biennial clinical examinations, high prevalences of occult cancer less than 1.5 cm in size at autopsy, significant excesses of papillary and follicular carcinoma, and no significant elevations for medullary or anaplastic cancer (Prentice et al., 1982; Sampson et al., 1969; Ezaki et al.,

cific to an attained age of 50 years, after exposure at 25 years, and depict the dose-response at doses less than 1 Sv. (Source: Little et al., 1999; Reproduced from UNSCEAR, 2000, p. 347.)

Ionizing Radiation 1986). The dose response was consistent with linearity; RRs were similar for males and females and highest among survivors less than age 20 at exposure (Thompson et al., 1994). Since women are about three times more likely to develop thyroid cancer than men, equality of the RRs implies higher absolute risks among women. For persons younger than 10, between 10 and 20, and 20 and older at the time of bombings (ATB), the RRs at 1 Gy were 10.5, 4.02, and 1.10, respectively (2.25 overall). Corresponding absolute excess risks were 4.4, 2.7, and 0.21/104 PY-Gy (1.61 overall), indicating only a small risk following adult exposures.

Other Cancers. Significant excess risks were found for both mortality and incidence of cancers of the stomach, colon, lung, breast, ovary, esophagus, and urinary bladder (Preston et al., 2003; Pierce et al., 1996; Thompson et al., 1994; Ron et al., 1994b). Mortality due to primary liver cancer was not significantly increased (Pierce et al., 1996), although incidence data revealed a significant excess RR, which appeared potentiated by the presence of hepatitis infection (Cologne et al., 1999). There was a significant increase in schwannomas and certain other tumors of the central nervous system, but not gliomas, meningiomas, and pituitary tumors (Preston et al., 2002b). Nonmelanoma skin cancers, but not squamous cell carcinomas, were linked to ionizing radiation dose (Ron et al., 1998a) as were salivary gland tumors (Saku et al., 1997). For all cancers, excluding leukemia, the dose-response relationship was clearly consistent with linearity (Fig. 15–6). Because the relative standard deviation increases greatly at low doses, however, there could be substantial deviations from linearity in the range below 200 mSv (Kellerer, 2000). No significant risks were seen for cancers of the rectum, pancreas, uterus, prostate, testes, bone, brain, oral cavity and pharynx, or nasal passages and larynx. For radiogenic cancers other than leukemia, the minimal latent period was at least 10 years, and the RR for all cancers as a group, except leukemia, declined with increasing attained age (Preston et al., 2003). Relative and absolute risks differed significantly by age at exposure, with younger survivors having higher risks. For nonleukemia deaths, the dose-response gradient was consistent with linearity until recently when a linear-quadratic model provided a better fit (Preston et al., 2004). After 50 years of follow-up through 2000, 55% of the leukemias (112 of 202) and 5% of the non-leukemia cancer deaths (477 of 10,085) have been attributed to the atomic radiation in 1945 (Preston et al., 2004; Pierce et al., 1996). About 1% of the 47,529 deaths from all causes could be attributable to the atomic bomb radiations. The average loss of life among all exposed atomic bomb survivors is somewhat less than 4 months (Cologne and Preston, 2000).

269

The RR for mortality at 1 Gy was 1.40 and the absolute excess was 10.6/104 PY-Gy for the period from 1950 to 1990 (Pierce et al., 1996). The estimated average RR at 1 Gy based on the solid tumor incidence data was 1.63 while the average excess absolute risk was 29.7/104 PYGy (Thompson et a1., 1994).

General Comments. The atomic bomb survivor studies provide an important human experience from which estimates of radiation risk can be derived. The population is large and not selected because of disease or occupation. Estimates of doses are based upon a comprehensive program that included interviews with virtually all proximal survivors. Despite its strengths, the study is not large enough to provide direct evidence of the effects of low doses (<~0.2 Gy); inferences about low-dose risk therefore depend upon dose-response models fitted to data obtained at a wide range of doses and upon principles of radiation biology and biophysics (Kellerer, 2002b; Heidenreich et al., 1997; Pierce and Preston, 1997; Hoel and Li, 1998; Baker and Hoel, 2003). The statistically important high-dose data are surprisingly few; there were only about 3000 survivors who received over 1 Gy. Because the exposures were acute and the dose rate was high, the data provide no direct information on the effects of protracted, low dose-rate exposures. Additional factors that might affect interpretations include: (1) restricting the study population to 5-year survivors for mortality and 13-year survivors for incidence, (2) possible effects of thermal or mechanical injury on those close to the hypocenter, (3) possible effects of poor nutrition or health problems on subsequent cancer risks, (4) possible effects of the substantial changes in cancer risk factors (e.g., diet and smoking) occurring over time in the population since 1945, (5) inaccuracies in dose assessment, including the neutron contribution, (6) inaccuracies and possible bias in death certificate diagnoses of cancer, and (7) the relative absence of healthy men of military age in the cities in 1945 (UNSCEAR, 2000). The shape of the dose-response curve in the low-dose domain may be affected by biased recording of cancer deaths on certificates when it was known that the subject was a survivor (Pierce et al., 1996; UNSCEAR, 2000). Increasingly sophisticated techniques are being incorporated to adjust for biases in dose estimates. With regard to possible biases introduced by selection factors, various studies suggest that some such effects, if present initially, tend to disappear over time (Howe et al., 1988). There remain lingering uncertainties on how to best apply the risk estimates from the atomic bomb survivor study where exposures occurred briefly in 1945. There have been changes in birth cohort exposures, such as tobacco consumption and diet, which have affected background cancer rates and radiation inferences (Pierce et al., 2003). Although linearity has described the relationship between solid cancers and radiation dose there are several proposed models (linear, linear quadratic, quadratic, threshold) that can be used to extrapolate data from high doses to low doses (Upton, 2003; NCRP, 2001; Tubiana, 1998, 2005), and results may also be influenced by assumptions relating to the small neutron component (i.e., whether the RBE varies by dose and whether organ-specific neutron doses are used) (Preston et al., 2003; Baker and Hoel, 2003; Kellerer et al., 2002; Kellerer, 2002b; Kellerer and Nekolla, 1997). Exposure to Radionuclides Radium (224Ra)

Figure 15–6. Excess relative risks for solid tumors, adjusted to men aged 30 at the time of exposure, in the Life Span Study of survivors of the atomic bombings. (Source: Pierce et al., 1996, Radiat Res 146:1–27, p. 9.)

In a study of 899 German patients repeatedly injected with 224Ra to manage bone tuberculosis and ankylosing spondylitis, 56 malignant bone tumors developed versus less than one expected (Mays and Speiss, 1984; Nekolla et al., 2000). Similar to the pattern seen for radiogenic leukemia in other studies, risk peaked at 6 to 8 years, and decreased to normal level after about 33 years. New dosimetry of bone surface doses (mean 30.6 Gy) revealed that children were at higher risk than adults. At high doses, increased duration of dose administration was related to increased risk. The dose response was best described by a linear-quadratic-exponential equation. Average dose to the bone volume was estimated as 4.2 Gy, the excess absolute risk as 0.8 cancers/104PY-Gy, and the lifetime risk as 2.0% per person per gray (NAS, 1972, 1988). Using the more appropriate but higher estimate

270

PART III: THE CAUSES OF CANCER

of dose to bone surface, the ERR/Sv ranged between 0.04 and 0.45 for exposures at age 60 years and 5 years, respectively, and the lifetime attributable risk was 0.18 per gray or 0.009 per sievert using a radiation weighting factor of 20 (Nekolla et al., 2000). Risk per unit dose to bone volume was about 10 times greater than that seen in radium dial painters who ingested 226Ra and 228Ra (Rowland et al., 1978), reflecting the different dose distributions in bone. Radium 224 has a short half-life (3.62 days) and releases its energy on bone surfaces where the “critical” cells for osteosarcoma induction, the endosteal cells, are located. In contrast, 226Ra is a bone volume seeker with a very long half-life (1600 years), and distributes its energy more uniformly throughout the bone, the dose to the bone matrix being essentially irrelevant to risk. For the same average bone volume dose, the risk from 224Ra will always be greater than that from 226 Ra because more radiation will reach the endosteal cells. Precise quantification of risk in the study of German patients is limited, however, because of the nonuniform distribution of 224Ra in bone and the possible effects of the underlying disease or other medications on cancer risk. Protracted exposure to radium alpha particles appeared more carcinogenic than acute exposures (Mays and Spiess, 1984), but is apparent only at the highest doses (Nekolla et al., 2000). This “protraction-enhancement” from alpha-emitting radiation might be related to several factors: (1) less killing of premalignant cells, (2) exposing more cells, (3) increasing the stimulus for cell division, and/or (4) preventing repair of local damage. In contrast, for X-rays and gamma rays, a decrease in the carcinogenic effectiveness of a given dose generally occurs when dose rates are decreased or protraction times are increased (NCRP, 1980; UNSCEAR, 1993). Small excesses of leukemia and cancers of the liver, kidney, bladder, breast, and soft tissue were reported (Spiess et al., 1989; Nekolla et al., 1999). The authors point out that phenylbutazone taken to relieve pain associated with spondylitis may be related to acute forms of leukemia. The increase in breast cancer was among women (28 observed vs. 8 expected) and also among men (2 observed vs. 0.2 expected). The alpha-particle dose to breast tissue was estimated to be about 0.1 Gy (and the equivalent dose of about 2–2.5 Sv). A radiation risk is thus plausible although other possible explanations include previous treatments by lung collapse with repeated chest fluoroscopies, a higher proportion of women who were nulliparous and thus at higher risk of breast cancer than the general population, an underlying genetic condition since there was a disproportionate number of early-onset breast cancers, or perhaps increased surveillance and detection of breast cancers in this cohort (Boice, 2001a). There was no reported dose response. Further, breast cancer among radium dial painters, described below, was not linked to larger amounts of ingested radium. Among 1577 German patients with spondylitis given 224Ra, but in smaller doses (0.56 Gy), four skeletal tumors (but no osteosarcomas) and 13 leukemias occurred (Wick et al., 1999). In contrast, one skeletal tumor and seven leukemias developed in 1462 non-exposed spondylitics. Breast cancer was not increased.

Iodine 131 Thyrotoxicosis. In a cooperative study of over 30,000 patients with thyrotoxicosis, leukemia was not linked to radioiodine treatments (Saenger et al., 1968). Among 18,400 patients treated with 131I, 17 leukemias developed, contrasted with 16 in 10,700 patients who received surgery only. The value of a non-exposed comparison group was underscored in that an excess of leukemia was suggested when general population rates were used for comparison. The dose to the whole body was low (7–13 cGy). An extended follow-up revealed no increases of cancer following I-131 therapy (Ron et al., 1998b). A suggested increase in mortality due to thyroid cancer might have been related to preexisting disease at the time of treatment. Other studies of patients treated for hyperthyroidism have also failed to link 131I with leukemia (Holm et al., 1991; Hall et al., 1992a, 1992b; Franklyn et al., 1999). Thyroid cancer has not been consistently correlated with 131I therapy (Holm et al., 1991; Ron et al., 1998b; Franklyn et al., 1999), possibly

because of the cellular destruction and loss of thyroid function that follows a dose of 10–100 Gy to the thyroid. No cancer has been convincingly linked to 131I treatments for hyperthyroidism. Excess cancers of organs such as the bladder that concentrate iodine (Hoffman, 1984) and of the breast (Goldman et al., 1988), small bowel, and lung (Franklyn et al., 1999) have been suggested in some studies, but were not confirmed in a large incidence series of 10,552 patients observed for up to 30 years (Holm et al., 1991) or in a large mortality series (Ron et al., 1998b).

Thyroid Cancer. In a study of 258 persons treated with highdose 131I for inoperable thyroid cancer, four leukemias were observed versus 0.08 expected based on general population rates (Edmonds and Smith, 1986). Small excesses of bladder cancer and breast cancer were also noted. A slight excess of leukemia (4 vs. 1.6) was reported among 834 patients treated with 131I for thyroid cancer in Sweden, but cancers of the bladder and breast were not excessive (Hall et al., 1991). The doses to the bone marrow and other organs in these series were large and likely between 0.5 and 1.0 Gy. The Swedish study was combined with two from France and Italy, confirming the increase in leukemia and revealing a salivary gland cancer risk (Rubino et al., 2003). Details of follow-up for the other two countries were not provided so interpretations are somewhat uncertain. Diagnostic 131I. A study of more than 35,000 Swedish patients failed to link the incidence of any cancer with diagnostic doses of 131I (Holm et al., 1988, 1989; Hall et al., 1996; Dickman et al., 2003). The dose to the thyroid was 1.1 Gy (110 rad) and a substantial excess of thyroid cancer was anticipated. The absence of an effect was originally thought to be due to a lower carcinogenic effect from internal 131 I beta particles compared to external X-rays or gamma rays, perhaps related to the protracted nature of the exposure (half-life = 8 days) or to the distribution of dose within the gland from 131I. However, because age at exposure significantly modifies the effectiveness of radiation to cause thyroid cancer (Thompson et al., 1994; Ron et al., 1995), the absence of an increased risk in the Swedish series might merely reflect the small number of exposed children and adolescents. [Interestingly, a recent study of external radiotherapy to treat arthritis of the cervical spine in Sweden reported a low but significantly increased risk of thyroid cancer among adults, RR of 1.6 at 1 Gy (Damber et al., 2002).] The latest follow-up included more patients under age 20 and still no increase was observed (Dickman et al., 2003), consistent with other series of children administered radioactive iodine for diagnostic purposes (Hahn et al., 2001; Hamilton et al., 1989), and providing additional evidence for a lower risk per unit of radiation following I-131 exposures (Boice, 2005). Patients examined because of a suspicion of a thyroid tumor were found to be at risk of thyroid cancer independent of 131I exposure. It is also possible that thyroid cancer risks were affected by subsequent surgery or hormonal medications. A series of nearly 14,000 patients in Germany given 131I also failed to identify a thyroid cancer risk (Globel et al., 1984). It is noted that no study of subjects exposed only to I-131 provides clear evidence for an increase in thyroid cancer, including the recent Hanford thyroid study, which involved examining more than 3000 persons exposed to I-131 as children during reactor releases (Davis et al., 2004; NAS, 2000). Although Chernobyl studies clearly indicate excess thyroid cancer risks among those exposed as children, risk may have been influenced by other shorter-lived and more penetrating radioactive iodines as well as surveillance and iodine dietary deficiencies (UNSCEAR, 2000; Balanov et al., 2003; Shakhtarin et al., 2003; Cardis et al., 2005b). Phosphorus 32 Among 1222 patients treated for polycythemia vera (PV), a blood disease characterized by overproduction of red cells, leukemia developed in 11% of 228 patients treated with 32P, 9% of 79 treated with Xrays, and 16% of 72 treated with both X-rays and 32P, in contrast to 1% of 133 nonirradiated patients (Modan and Lilienfeld, 1965). It is possible that the bone marrow of patients with PV may be unusually sensitive to radiation. However, the causal nature of the association

Ionizing Radiation 32

with P was not entirely clear for several reasons: (1) the incidence of leukemia was also associated with spleen size at the time of treatment, suggesting that biological factors determining treatment, rather than the treatment itself, could be associated with leukemia (UNSCEAR, 1972), (2) the underlying myeloproliferative disease may predispose to leukemia, (3) biases in the selection of patients being treated could not be discounted (that is, patients treated with “additional radiation” had to be removed from either the non-exposed, 32P only, or X-ray only groups), and (4) PV patients are exposed to other medications, including powerful cytotoxic drugs, that could increase leukemia risk. A randomized clinical trial, however, found that 9 of 156 (6%) patients treated with 32P developed leukemia in contrast to 1 of 134 (1%) treated by phlebotomy (Berk et al., 1981). Patients treated with chlorambucil were at highest risk (16 of 141, 11%). Subsequent studies confirm the association between 32P and leukemia, but also fail to find a dose-response relationship; an enhancement of risk with chemotherapy maintenance, in particular hydroxyurea, was apparent (Najean et al., 1996; Najean and Rain, 1997).

Thorotrast A colloidal solution of thorium dioxide (Thorotrast) was used between 1928 and 1955 as a contrast agent during radiographic procedures (IARC, 2001). The thorium, however, remained in body tissue for life and resulted in continuous alpha particle exposure at a low dose rate. The annual dose from a typical injection of 25 ml of Thorotrast was about 25 cGy to liver and 16 cGy to bone marrow. Surveys in Denmark, Germany, Japan, Portugal, Sweden, and the United States show substantial excesses of liver cancer, including angiosarcoma and cholangiocarcinoma, and acute myeloid leukemia (NAS, 1988; Taylor et al., 1989; Andersson et al., 1993, 1997; dos Santos Silva et al., 2003; Mori et al., 1999; Nyberg et al., 2002; van Kaick et al., 1995, 1999; Travis et al., 2001, 2003a). Hemangioendothelioma of the liver appears uniquely related to Thorotrast. Among 2326 exposed persons in the German Thorotrast study, 396 (17%) have died from liver cancer in contrast to only 2 (0.1%) among 1890 controls (van Kaick et al., 1989; NAS, 1988). Despite continuous exhalation of thoron (220Rn), excess lung cancer has not been consistently seen (Hofmann and Hornik, 1999; Travis et al., 2003a), suggesting that the risk from low doses of radon may be overestimated or the distribution of dose within the lung may be important. Small increases in bone cancer have also been noted, possibly due to translocating 224Ra from Thorotrast deposits (NAS, 1988). Cancer risk remained high for up to 50 years after initial injection (Travis et al., 2003a). The relative effectiveness of alpha particles to cause leukemia appears very similar to that expected from external irradiation at doses to bone marrow on the order of 1.3 Gy (Boice, 1993b). The nonuniform deposition of thorium in the liver and bone marrow likely resulted in very high local doses, which may be the important determinant of cancer risk (Guilmette and Mays, 1992). If so, the convention of averaging dose over the entire organ would be misleading. Risk estimation is also hindered for the following reasons: (1) the chemical nature of thorium, a heavy metal, may be related to risk, (2) the average dose to the liver was about 5 Gy, and a portion of this radiant energy, expended in necrotic tissue, was probably not essential for carcinogenesis, (3) except for the recent studies in Scandinavia the completeness of patient follow-up was generally poor, and (4) the combination of necrosis and liver regeneration might influence risk. Further, Thorotrast was often administered to diagnose and evaluate liver diseases that may intrinsically have contributed to the development of subsequent cancer, although such patients were excluded from the German study. Nonetheless, Thorotrast appears to be one of the most carcinogenic exposures known to man, with cumulative lifetime incidences of cancer estimated to be as high as 86% (Andersson and Storm, 1992).

Occupational Exposures Radium Dial and Clock Painters Among 1474 women employed in the US radium dial industry before 1930, 61 bone sarcomas and 21 head carcinomas have occurred

271

(Rowland et al., 1978; Stebbings et al., 1984; Fry, 1998; IARC, 2001). The habit of licking paint brushes to make fine tips resulted in the ingestion of large quantities of bone-seeking 226Ra (mean to bone, 17 Gy) and some quantities of 228Ra. Cancers in mastoid air cells or paranasal sinuses (head carcinomas) likely were caused by radon gas emitted as a decay product of radium. The “latency period” for osteosarcoma was not related to dose (Polednak, 1978). Age at first exposure did not influence risk (Carnes et al., 1997). Risk was estimated as 0.1 bone cancers/104PY-Gy (NAS, 1972). No excess of leukemia was observed (10 vs. 9.24 expected) among US dial painters (Spiers et al., 1983) or among 1100 English radium luminizers (Baverstock and Papworth, 1989), suggesting that the stem cells in the bone marrow necessary for leukemogenesis must not be sufficiently damaged from the alpha particles emanating from the radium deposited in the bone (Priest, 1989, 2001). Early reports linking breast cancer with radium or external gamma ray exposures were not confirmed (Stebbings et al., 1984; Baverstock and Papworth, 1989). Multiple myeloma was increased in the US study, but correlated with duration of employment (a surrogate for gamma ray exposure), rather than radium intake. Liver cancer was not increased. The British study reported only one osteosarcoma, but the systemic intake of radium was much lower than for the United States. Other than for radiogenic cancer, there was no general life-shortening effect (Stehney et al., 1978). A quadratic equation of the form I = (8 + bD2)e-aD fits the osteosarcoma data, whereas a linear form, I = c + bD, fits the head carcinoma data. Marshall and coworkers (1977, 1978) developed an elaborate two-target model proposing that two successive initiating events and a later promoting event are required for osteosarcoma induction. The initiation events remove the ability of a cell to stop dividing; the promotion event is a signal to divide associated with natural remodeling of bone. The model also allowed for the competitive effects of cell killing. No bone cancers occurred below 10 Gy suggesting a possible threshold or “practical” threshold for radiation-induced bone cancer (Priest, 2001). The actual dose-incidence curve determined for radium dial painters must be considered tentative for the following reasons: (1) the estimation of dose was made many years after the ingestion of radium, (2) the nonuniform distribution of radium in bone likely resulted in “hot spots” that caused extensive cell killing, (3) the dose responsible for tumor induction cannot be distinguished from the “irrelevant” or “wasted” dose received after initiation, (4) the relative effectiveness and contribution of the alpha particle emissions cannot easily be separated from the other radiations accompanying radium decay, and (5) the fraction of the total dose to the endosteal cells cannot be specified precisely.

Radiologists The first cancer attributed to ionizing radiation occurred on the hand of a radiologist in 1902 (NAS, 1990), and leukemia was first associated with chronic exposure in studies of radiologists (March, 1944). Leukemia, aplastic anemia, and skin cancer were excessive among radiologists who practiced during the early part of this century before radiation protection guidelines were commonplace, but these risks appear to have disappeared among more recent radiologists (Matanoski et al., 1975; Berrington et al., 2001; Wang et al., 1990a, 2002a). Multiple myeloma was increased among US radiologists practicing in later years (Lewis, 1963; Matanoski et al., 1975), but not among English or Chinese radiologists. Cancers of the pancreas and lung were increased among the pioneering radiologists in the United Kingdom, but not in the United States or China. Suggested increases of breast, thyroid, and bone cancers were correlated with radiation work in China only (Wang et al., 1990a). Neither leukemia nor cancer was reported to be in excess among US Army X-ray technologists, who likely received much lower total doses (Jablon and Miller, 1978). A survey of 145,000 radiologic technologists in the United States (Boice et al., 1992c) also found little evidence for increased mortality (Doody et al., 1998; Mohan et al., 2003). An evaluation of 600 prevalent breast cancers reported on mail questionnaires did not reveal an association between measures of radiation exposure and breast cancer risk (Boice et al., 1995). A smaller mortality survey suggested an increase in breast

272

PART III: THE CAUSES OF CANCER

cancer among women who began practice before 1950 (Mohan et al., 2002), but the results might have been confounded by pregnancy histories since Catholic nuns made up a high percentage of these early workers or they may be partially due to a survival bias and a methodological artefact related perhaps to a mismatch in the age distributions between exposed and comparison cohorts defined by calendar years of employment (UNSCEAR, 2000). To be included, women certified before 1950 had to have survived to the late 1980s and complete a mailed questionnaire. Those who died before questionnaire was mailed were excluded and this group did not show an increased risk overall (Mohan et al., 2003). The comparison group was technologists who were certified after 1950 and who survived to complete the mailed questionnaire in the 1980s, and thus were much younger with few elderly women. There appeared little overlap in the age distributions of elderly women who are at highest natural risk of breast cancer. A generally higher mortality rate among US radiologists from all causes was originally interpreted as evidence of an acceleration of the aging process by radiation. Other than the loss of life due to cancer deaths, however, nonspecific life-shortening has not been demonstrated in animal experiments or seen in British radiologists, or radium dial painters (UNSCEAR, 1982). Mortality analyses among A-bomb survivors, however, leave open the possibility that excess non-cancer deaths may have occurred following high doses over about 1 Gy (Preston et al., 2003). The absence of accurate estimates of radiation dose is a serious limitation of these studies. Cumulative doses were likely between 1 and 8 Gy during the early part of this century, and it is possible that radiologists who developed cancer were those who scorned safety measures and received even greater doses. It was not uncommon for X-ray workers to be given time off from work because of severe depression of white blood cell counts. Radiologists also receive more personal (non-occupational) exposures to diagnostic and therapeutic radiation than other specialists (Jessup and Silverman, 1981). Nonetheless, these studies indicate that leukemia and skin cancer can result from repeated, presumably small, radiation exposures received over a period of many years if the cumulative dose is sufficiently high.

Workers at Nuclear Shipyards A proportional mortality study of naval shipyard workers suggested an increased risk of cancer and leukemia among nuclear workers in Portsmouth, New Hampshire (Najarian and Colton, 1978), which was not borne out in a subsequent cohort study (Rinsky et al., 1981). Radiation exposure histories, ascertained from next-of-kin by newspaper reporters, did not correlate with employment records. Further, relatives of workers who died from cancer were more likely to be located and interviewed, which, in combination with a lower all-cause mortality among nuclear workers, contributed to the spurious result (Greenberg et al., 1985). Case-control studies of leukemia and lung cancer also found no association with radiation work (Rinsky et al., 1988; Stern et al., 1986). A recent follow-up attributed an excess of lung cancer to factors associated with SES and asbestos exposure (Yiin et al., 2005). A comprehensive evaluation of workers at eight nuclear shipyards found no increase in any cancer except mesothelioma, attributable in all likelihood to asbestos exposures (Matanoski, 1993; Boice, 2001b).

Workers at Nuclear Installations Studies of workers at individual nuclear facilities are in large part inconsistent because of the relatively low doses workers received and the limited study sizes (UNSCEAR, 2000). The one exception, noted below, is of the Mayak workers in Russia who received rather massive exposures to plutonium and external gamma rays. Large-scale studies of workers involved with the milling and processing of uranium have provided no evidence for a carcinogenic effect, related in all likelihood to the low radioactivity of uranium and to the limited distribution of dose within the body due to the chemical nature of the heavy metal (CRS, 2001; IARC, 2001). Insights into the effects of chronic long-term exposure to ionizing radiation, however, may come from the combination of large international studies, notably of 407,000 workers in 15 countries (Cardis et al., 2005). However, the recent find-

ings are equivocal (Wakeford, 2005). The significant risk for solid cancers was entirely due to an exceptionally high risk for lung cancer, suggesting the confounding influence of smoking, and significance depended on only one of the 15 countries (Canada). Further the individual country results where disimilar to those previously published, and leukemia was no longer significantly elevated. Previously, only leukemia has been convincingly linked to occupational exposure; a significant excess was based on a few workers who received relatively high doses (>0.4 Gy) in an international study of workers in three countries (Cardis et al., 1995) and was reported among Russian workers exposed to greater than 1 Gy (Koshurnikova et al., 1996, 1997). Except at Mayak, female workers in the nuclear industry have not been found to be at increased risk of cancer (McGeoghegan et al., 2003; Frome et al., 1990). The mortality experience of nearly 31,500 male and 12,600 female workers employed between 1944 and 1978 at the Hanford nuclear installation in Richland, Washington, has been reported by several investigators. An early proportional mortality analysis on 3520 certified deaths (Mancuso et al., 1977) was widely criticized and discounted (Hutchison et al., 1979; NCRP, 1980; NAS, 1980). Conclusions were inconsistent with subsequent follow-up studies (Gilbert et al., 1993a,b). The most recent analyses revealed a strong “healthy worker” effect, a significant deficit of cancer mortality including leukemia, and no evidence for increasing risk with increasing film badge exposure for any cancer. A previously reported excess of multiple myeloma was no longer significant. Multiple myeloma was also not associated with cumulative dose in a case-control study at four US nuclear facilities, but the small numbers and questionable control selection and case inclusion criteria limit interpretation (Wing et al., 2000). Leukemia was elevated at the Oak Ridge National Laboratory (ORNL), but risk was inversely related to dose (Checkoway et al., 1985; Gilbert et al., 1993b). Early reports found multiple myeloma to be excessive at Sellafield, although based on only two cases receiving over 50 cGy (Smith and Douglas, 1986), and at Hanford (Gilbert et al., 1989), but subsequent follow-ups failed to confirm these associations (Douglas et al., 1994; Omar et al., 1999; Gilbert et al., 1993b). Prostate cancer was increased at the UK Atomic Energy Authority and the UK Atomic Weapons Establishment (Beral et al., 1988; Rooney et al., 1993), but negatively linked to radiation among US workers (Gilbert et al., 1993b). Lung cancer is often found to be significantly low in nuclear workers (Gilbert et al., 1993b). Studies of nuclear workers to date provide no consistent patterns of increased cancer risk. An analysis of data on workers at the ORNL has received considerable criticism (Wing et al., 1991). An excess of leukemia, including CLL, was emphasized, although risk decreased with increasing levels of exposure. Confounding by smoking likely contributed to the correlations reported for solid cancers, making the data difficult to interpret since lung cancer dominated the analysis (Gilbert, 1992, 1993b). Compared to the general population, the ORNL workers were at a 28% significantly reduced risk of dying from lung cancer. Interestingly, Oak Ridge workers hired during WWII were previously reported to be at high risk of lung cancer unrelated to radiation exposure (Frome et al., 1990). This “unhealthy worker effect” appeared due to the selection out of the workforce of physically fit individuals to serve in the armed forces. A study of 3145 Department of Energy and US Navy personnel who received greater than 50 mSv (5 rem) in any calendar year found no significant deviations in cancer mortality and no excess deaths due to leukemia (2 observed vs. 4.3 expected) (Fry et al., 1996). The median cumulative dose was 153 mSv (15.3 rem) and 287 workers had cumulative doses over 400 mSv (40 rem). The absence of an increased risk of leukemia is of interest but the small numbers and possible under-ascertainment of deaths are limitations. Another recent follow-up of a large national registry of 125,000 radiation workers in the United Kingdom revealed marginally significant increased risks due to leukemia, excluding CLL, and the risk of other cancers was not significantly elevated (Muirhead et al., 1999). Nonetheless, these findings were consistent with A-bomb survivor data predictions indicating the inherent methodologic difficulties in excluding the possibility of a very small risk in any epidemiologic investigation. Overall, the workers were at low cancer risk compared

273

Ionizing Radiation to the general population (SMR = 0.82). This report, similar to the three country international series (Cardis et al., 1995), should be interpreted with caution because the leukemia risk seemed apparent only at one facility, Sellafield, where cumulative exposures more than 40 cGy have occurred and where potential exposure to leukemogenic chemicals during fuel reprocessing activities was possible. A report from Canada exemplifies the difficulties in detecting risks following low-dose radiation exposures (Gribbin et al., 1993). A careful study of nearly 9000 workers revealed a significant deficit of cancer (SMR = 0.87), and there were no significant correlations with radiation for any site or combination of sites. Less than 1% of workers received greater than 5 cGy. Risk estimates were computed and stated to be consistent with extrapolations from A-bomb survivor data; that is, RR = 1.0036 at 1 cGy. Although true, the data were equally consistent with no effect at all, and reflect the extremely small excess risk expected at such low exposure levels and the associated low power to detect such risk. National studies have been published in the United Kingdom (Muirhead et al., 1999), Canada (Ashmore et al., 1998; Sont et al., 2001), and Japan (Iwasaki et al., 2003) but, although large, the results are not consistent. The United Kingdom and Japanese findings were not statistically significant, despite large numbers, and the Canadian findings provided estimates that were much larger than those seen in other studies appeared biased (Gilbert et al., 2001). The likelihood of bias was suggested for the following reasons: (1) the SMR was unusually low, 0.59 for all causes in males, and indicates an under-ascertainment of deaths perhaps related to the probabilistic matching procedures employed, (2) significantly high risks were seen for circulatory disease and accidents following low-dose exposures, and while the recent atomic bomb data suggest an association with circulatory disease, it is only following much higher doses and the risk coefficient is much lower than that for cancer (Preston et al., 2003), and (3) the cancer sites with the highest risk coefficients included cancers of the rectum, pancreas, and testes, which are sites rarely, if ever, found increased in any radiation study (Gilbert, 2001). The first comprehensive effort to combine series of nuclear workers involved three US studies (Gilbert et al., 1993b). Excess relative risk estimates per Sv were 0.0 for all cancer and -1.0 for leukemia. The authors concluded that data extrapolations from higher dose studies are unlikely to underestimate risks at lower doses. Another analysis combined data from 75,006 employees in three nuclear establishments in the United Kingdom (Carpenter et al., 1994). Excess RR estimates per Sv were -0.02 for all cancers and 4.2 for leukemia. Leukemia was elevated only in one of the three nuclear establishments. Significant increases in cancer of the pleura suggest that employees were exposed to other hazardous agents such as asbestos. In 1995, 95,673 nuclear industry workers in three countries were analyzed (Cardis et al., 1995). Leukemia was increased, but not other cancers. Interpretations are limited because, overall, only about 9 leukemia deaths of the 3976 total cancer deaths could be attributable to radiation, and there was no increase of cancer deaths. This international study has been expanded to include over 400,000 workers in 15 countries, including utility workers in the United States (Cardis et al., 2005a; Telle-Lamberton et al., 2004). Studies of nuclear utility workers may eventually provide useful information on radiation risks because of relatively higher exposures and larger numbers (Jablon and Boice, 1993), although this was not found to be the case in a recent US study (Howe et al., 2004). Even combinations of larger studies may have difficulty in providing risk estimates of useful precision, however, because the sample sizes and ranges of exposures appear small for acceptable power at the most likely effect level (Land, 1980; Cook-Mozaffari et al., 1987) and these large statistical uncertainties are compounded by uncertainties resulting from potential confounding and bias (Gilbert, 2001). The average cumulative doses for workers employed in research and development or weapons production, for example, are about 3 cGy (Muirhead et al., 1999), with only about 5%–15% over 5 cGy (Gilbert et al., 1993a,b; Beral et al., 1988; Cardis et al., 1995) and much smaller percentages over 40 cGy (Table 15–2). While such worker studies are nonetheless important and allow a direct assessment of risks resulting from exposure to radiation at low doses and dose rates, it is unlikely

Table 15–2. Comparison of Radiation Studies No. Subjects with Cumulative Dose Exceeding Subjects IARC Three County Study UK National Registry of Radiation Workers National Dose Registry of Canada USA Shipyard Workers Nuclear Industry Workers of Japan Atomic Bomb Survivors

No. Subjects

No. Cancer Deaths

100 mSv

400 mSv

95,673

3,976

10,007

1,752

124,743

3,598

9,580

NA

206,620

1,632

2,926

70,730

1,724

4,238

NA

175,939

2,138

4,161

NA

86,572

7,578

17,264

5,489

234

Source: Cardis et al., 1995; Muirhead et al., 1999; Sont et al., 2001; Pierce et al., 1996; Boice et al., 2001b; Gilbert et al., 2001. NA, not available but small.

that they can replace atomic bomb survivors as the primary source of data for risk estimation (Gilbert, 2001; Doll, 1999). Occupational studies of radiation workers must be interpreted carefully for the following reasons: (1) film badge or thermoluminescent dosimeter (TLD) exposures are imperfect measures of organ doses, (2) the dose from natural background radiation (about 7 cGy in 70 years) is often greater than the occupational dose, (3) other occupational and non-occupational carcinogens are usually not considered, (4) ascertainment bias is possible if the working population receives better medical care and more accurate cancer diagnoses recorded on death certificates than the general population. This surveillance bias was suggested as a possible explanation for the initial report of excess multiple myeloma (in the absence of a leukemia excess) seen among Hanford workers (NIH, 1985). However, chance might have been responsible for the early excess, which was no longer significant in the latest follow-up (Gilbert et al., 1993a). Further, there is the possibility of under-ascertainment of dose from neutrons, bias resulting from excluding dose from internally deposited radionuclides, bias from occupational dose received at other nuclear facilities after terminating employment or from medical exposures (Gilbert and Fix, 1995; UNSCEAR, 2000), and the potential selection biases associated with entry into the work force and continued employment.

Workers Exposed to Plutonium At one time, plutonium was thought to be one of the most toxic elements known to humans. Animal experiments clearly indicate that excessive exposure to plutonium can cause cancers of the lung, bone, and liver; however, the evidence in humans was sketchy (NAS, 1988) until the publication of Mayak worker studies in the late 1990s which provide convincing evidence of the carcinogenicity of plutonium in humans (IARC, 2001). Excesses of cancer of the lung (Koshurnikova et al., 1997, 1998; Tokarskaya et al., 1997; Kreisheimer et al., 2003; Khokhryakov et al., 1998; Gilbert et al., 2004), bone (Koshurnikova et al., 2000), and liver (Gilbert et al., 2000) have been linked to plutonium inhalation among workers first employed in 1948 to 1958 at the Mayak nuclear complex in the southern Urals. The risk, however, was apparent only for enormous quantities of plutonium intake (i.e., for body burdens >3 kBq), and there was little evidence for effects below this level. The inhalation of plutonium at Mayak was so enormous that deterministic effects occurred (e.g., plutonium pneumosclerosis—an interstitial lung disease causing functional lung deficiencies and fibrosclerotic change) (Claycamp et al., 2000). Thus while the demonstration that plutonium is carcinogenic in humans is very important, the Russian studies do not demonstrate a risk at low doses, which may be equally as important. The absence of a cancer risk at low doses of plutonium intake is consistent with studies of

274

PART III: THE CAUSES OF CANCER

plutonium workers in the United States (Voelz et al., 1997; Gilbert et al., 1993a; Wilkinson et al., 1987) and the United Kingdom (Beral et al., 1988; Omar et al., 1999). Leukemia also was not linked to plutonium exposure among Mayak workers, but it was strongly related to gamma ray exposure (Koshurnikova et al., 1996, 1997). This makes biological sense since the chemistry of plutonium apparently results in little to no radiation to the bone marrow compartments where stem cells reside. Alpha particles emitted from bone depositions apparently do not penetrate to any meaningful degree to bone marrow (Priest, 1989). These observations are consistent with studies of radium dial painters where increases in bone cancer but not leukemia are seen (IARC, 2001). The picture of cancer risk among Mayak workers following lowlinear energy transfer (LET) exposures is not as clear as for high-LET plutonium inhalation, except perhaps for the increase in leukemia noted above. The early studies (Koshurnikova et al., 1997, 1998) and the most recent analysis (Kreisheimer et al., 2003) reveal no risk of lung cancer among reactor workers who had minimum potential exposure to plutonium, consistent with the studies of fractionated exposures among tuberculosis patients treated with lung collapse where high doses received over a period of 3 to 5 years were not linked to a significant increase of lung cancer (Howe, 1995; Davis et al., 1989). The most recent analysis (Kreisheimer et al., 2003) estimated the excess relative risk for the plutonium alpha particles to be 0.23/Sv (95% CI: 0.16–0.31), and the excess relative risk for external gammaray exposure to be negligible at 0.058/Sv (95% CI: -0.072–0.20), in contrast to the inferred relative risk for smoking of 16.5 (95% CI: 12.6–20.5). It appears that highly fractionated exposures to low-LET radiation might carry little if any risk of radiation-induced lung cancer. A recent paper of Mayak workers addressed external and internal risks and reinforced previous reports that very high exposures to plutonium can result in an increased cancer risk to the lung, bone, and liver and that high gamma ray exposures can increase the risk of leukemia (Shilnikova et al., 2003). Risks were somewhat lower than those seen among atomic bomb survivors. A reported association between gamma rays and solid cancers is problematic because of the difficulties in adequately controlling for plutonium effects. The use of a surrogate measure of plutonium exposure is questioned because of the biological uncertainty in lumping together organs with widely different doses due to the heterogeneous deposition of plutonium deposition throughout body tissue. The precision gained by including the plutonium workers may be at the cost of creating uncontrolled confounding and distortion. An association between plutonium with cancers other than of the bone, liver, and lung, also suggested bias in that these other organs received little to no plutonium exposure. The studies of plutonium workers in Russia must be interpreted with caution for the following reasons: (1) substantial limitations in dosimetry, (2) incomplete ascertainment of vital status, and (3) incomplete information on potential confounders such as smoking, alcohol, and chemical exposures before and after Mayak employment (IARC, 2001). Further, cause of death information comes from four different sources (autopsies, death certificates, relatives, and other medical records) and is thus not consistent and bias is indicated since the highdose workers were targeted for autopsies, which accounted for 43% of the cause of death information. When compared with other studies that relied on one consistent source (i.e., death certification), any differences in risk may reflect this ascertainment bias as well as differences due to the nonuniform ascertainment of cause of death information and not radiation quality (low LET, high LET) or radiation delivery (acute, fractionated, prolonged). The high autopsy rate might also produce spurious associations for organs that are minimally exposed to plutonium. Also, it is questioned whether a low-LET effect can be convincingly detected in organs that are also heavily exposed to plutonium. The huge exposures to plutonium caused deterministic effects in some workers. There are inconsistent reports of findings in different studies that are yet to be resolved.

Workers Exposed to Uranium Many uranium workers were employed during the early years of uranium processing and manufacturing, including during the Manhat-

tan Project, and inhaled or ingested relatively large amounts of uranium. These workers have been observed for many years and their cancer risks have been evaluated. Fourteen epidemiologic studies have been conducted of more than 120,000 workers at uranium processing, enriching, metal fabrication, and milling facilities (CRS, 2001; IOM, 2001; Harley et al., 1999). These studies overall found no cancer to be significantly increased. The summary risk for all cancers taken together was close to that expected; that is 7442 cancers were observed compared with 8178 expected (SMR = 0.91). The consistency of the finding from the 14 epidemiologic studies of workers employed in uranium processing throughout the world is noteworthy. A large-scale case-control study of lung cancer among four uranium processing operations found no association with estimated lung dose (Dupree et al., 1995). In contrast, studies of underground uranium and other hard rock miners, discussed below, revealed a substantial increase in lung cancer attributable to radon and its decay products (Lubin et al., 1995a). Descriptive correlation studies also find no excess cancers among populations residing near uranium milling, mining, or processing facilities (Mason et al., 1972; Boice et al., 2003a, 2003b, 2003c). There are several possible reasons why uranium is not found to cause cancer in humans and why it is not considered a human carcinogen (IARC, 2001): uranium is not very radioactive (it decays very slowly) and its chemical properties are often such that any inhaled or ingested uranium is excreted rather quickly from the body (Harley et al., 1999).

Underground Miners Radon and radon decay products have caused lung cancers among underground miners for at least 400 years. In 1556, Agricola described a mysterious lung disease afflicting miners of the Black Forest regions of Eastern Europe (Agricola, 1950). Inhaled radon and its decay products were subsequently indicted as the culprit (Tomásˇek et al., 1994), although in some mines arsenic and other factors also contributed to the untimely deaths from lung cancer. Alpha particles emitted during the radon decay cascade resulted in large depositions of energy in bronchial cells, apparently jumbling DNA and eventually leading to lung cancer. Naturally occurring radon is arguably the most consequential radiation exposure to the world’s populations (UNSCEAR, 2000; IARC, 2001) and may account for up to 10% of the lung cancers occurring in the United States and other countries (Lubin et al., 1995a; NAS, 1999; Darby et al., 2003). Studies of 3366 white and 780 nonwhite underground uranium miners in the United States provided the first quantitative evidence that breathing radon and its decay products for long periods of time increased the risk of death from lung cancer (Lundin et al., 1971; NAS, 1972). Among white Colorado Plateau miners who worked at least one month underground, a substantial excess of respiratory cancer was observed: 185 deaths versus 38.4 expected based on general population rates (Waxweiller et al., 1981). The mean dose to lung was very large and estimated to be about 600 Working Level Months (WLM) (or about 30 Sv) (Lubin et al., 1995a). Most lung cancers developed in cigarette smokers, but an excess risk was also seen among 516 nonsmoking miners exposed to very high radon levels (Roscoe et al., 1989). Radiation and smoking appeared to interact in a way that enhances risk, though somewhat less than multiplicative (NAS, 1988; Hornung et al., 1998). Dose-response data for lung cancer in the US miners were difficult to interpret because of the following uncertainties: (1) lung doses for individual miners, which had to be estimated based on infrequent measurements in more than 2500 mines, (2) the relationship between exposure (the concentration of radioactive materials in mine atmospheres) and actual dose to respiratory tissue, (3) the dose to individual cells, which could vary depending on cell type, the thickness of the epithelial and overlapping mucous layers, and the clearance rate of absorbed radioactive particles, (4) the doses received in nonuranium mines, and (5) the contribution to risk of cigarette smoking and of pollutants, like diesel exhaust, in mine atmospheres (NAS, 1991,1999; Abelson, 1991). Many of the above concerns were addressed in subsequent studies in other countries where radon exposures were lower and better char-

Ionizing Radiation

275

Indoor Radon

Figure 15–7. Relative risk of lung cancer by cumulative radon concentrations for 11 cohort studies, combined, of underground miners. (Source: Lubin et al., 1994a, NIH Publ No. 94–3644, p. 31.)

acterized than in Colorado mines (NAS, 1988), and most recently within the framework of an international cooperation forged between investigators in Australia, Canada, China, the Czech Republic, England, France, Sweden, and the United States (Lubin et al., 1994a, 1995a; NAS, 1999). Eleven major studies of uranium, fluorspar, tin, and iron miners were combined; analysis included 65,000 men and 2700 lung cancers. It was estimated that 40% (or 1080) of all lung cancers were attributable to underground radon exposure. In comparison, the study of 87,000 atomic bomb survivors recorded 939 lung cancer deaths, of which 7% (or 67) were attributable to the atomic radiation (Pierce et al., 1996). Perhaps the most striking epidemiological feature of the 11 underground miner studies is that a straight line adequately describes the relationship between cumulative radon exposure and lung cancer risk, although there were substantial differences in the individual risks estimates. The overall excess relative risk (ERR) per WLM was 0.49% (Fig. 15–7). The exposure-response trend for never smokers was three-fold greater than the trend for smokers. The ERR diminished with time after exposure and also with attained age. Age at first exposure did not affect risk, even among Chinese children exposed under age 15. For equal total exposures, exposures of long duration (and low rate) were more harmful than exposures of short duration (and high rate). This so-called “inverse dose rate” (or protraction enhancement effect) occurred only at the highest cumulative exposures, and was not apparent at cumulative exposure below about 100 WLM (about 5 Sv) (Lubin et al., 1995b; Hornung et al., 1998). Based on biophysical principles, very low doses, such as experienced in domestic situations, would not be expected to show a dose-rate effect (or protraction enhancement effect) because multiple alpha particle traversals of single cell nuclei from radon decay would be rare (Brenner, 1994). All the estimates of risk are interpreted carefully, however, because concomitant exposures of miners to agents such as arsenic or diesel exhaust may modify the radon effect and adjustment for known lung carcinogens did result in lower radon risks. No excess leukemia or lymphoma has been reported, although miners were heavily exposed to uranium, radon, and their decay products, including gamma ray emissions (Tomásˇek et al., 1993). No solid cancer, other than lung cancer, has been convincingly linked to radon exposures in underground mines (Darby et al., 1995; NAS, 1999). The evidence that radon causes lung cancer among underground miners comes mainly from exposures over 100 WLM (about 5 Sv), and significant risks below about 50 WLM (about 2.5 Sv) have not been observed (NAS, 1988), although data are consistent with a linear exposure-response relationship (Lubin et al., 1997b; Boice, 1997b). In contrast, average levels of indoor radon are estimated as 0.2 WLM (about 0.01 Sv to lung) per year (NAS, 1991).

Although it has long been recognized that radon and its decay products causes lung cancer among underground miners (NAS, 1999), the possible hazard to homeowners exposed to much lower levels was not appreciated until the mid 1980s when the home of a Limerick Nuclear Power Plant worker in Pennsylvania was found to have levels of 2800 pCi per L (i.e., more than 50 times higher than the annual occupational limit of exposure for uranium miners) (MMWR, 1985). Indoor radon accounts for over half of all radiation exposures received by the general population, and, based on extrapolations from underground miner studies, may cause between 6600 and 24,000 lung cancer deaths per year in the United States (Lubin et al., 1994a, 1995a; NAS, 1999). Because the entire population, over 280 million citizens, breathes in radon with every breath, even a small risk can translate into large numbers of estimated deaths. A flurry of research activities all over the world has attempted to quantify the lung cancer risk from residential radon (Lubin and Boice, 1997; NAS, 1999; Lubin et al., 2004; Krewski et al., 2005; Darby et al., 2005). Early studies were equivocal because of small numbers and problems with exposure assessment. In 1979, lung cancer was linked to living in stone houses compared to wood houses (Axelson et al., 1979). A Swedish national study revealed a positive association between radon and lung cancer based on comprehensive measurements and large numbers (Pershagen et al., 1994). A comprehensive case-control study of 308 women diagnosed with lung cancer in China found no association between lung cancer and increasing radon exposure (Blot et al., 1990). Year-long measurements of radon were made in current residences, and 20% of the readings exceeded 4 pCi/liter, which is the action level for remediation in the United States. Similar negative findings were reported in a large study in Canada of more than 750 lung cancer cases (Letourneau et al., 1994). A study of 600 incident lung cancer cases in Missouri among nonsmoking women also found no overall association with radon, and the population attributable risk to radon was estimated to be at most 2% (Alavanja et al., 1994, 1995). This study had several methodological strengths: it was an incidence survey with exposure measurements made close in time to the lung cancer diagnosis, it focused on nonsmokers to enhance the probability of detecting an effect and eliminating direct smoking as a confounder, there was relatively little migration of residents, and a comprehensive dosimetry program was in place. A subsequent study of predominantly smoking women in Missouri also found no association with radon when exposureresponse trends were based on standard radon dosimeter measurements, but a radon risk was reported based on a novel surface monitor approach to estimate cumulative radon exposure in household objects made of glass (Alavanja et al., 1999). This method, however, has not been validated, and uncertainties, especially in houses where persons smoked, are not inconsequential (IARC, 2001). Somewhat surprisingly, a large ecological survey reporting a significant inverse relationship between county measures of radon and lung cancer mortality continues to generate controversy despite the presence of numerous analytical studies (Cohen, 1995). Such ecological studies, however, have intrinsic methodological difficulties, such as the inability to adjust for mobility and individual smoking habits (Stidley and Samet, 1993; Gilbert, 1994; NAS, 1999; Puskin, 2003). Other descriptive correlation analyses subsequently found an inverse relationship between lung cancer mortality and radon, but case-control studies conducted within the same geographical areas revealed positive associations (Lagarde and Pershagen, 1999; Smith et al., 1998; Darby et al., 1998, 2001). In these instances, aberrant dose-response findings in the ecological analyses disappeared when methodologically sounder investigations were conducted using individual radon exposure estimates (rather than group estimates) and individual smoking histories in the analysis (IARC, 2001). Recent large-scale studies of lung cancer risk and residential radon in the United Kingdom, Iowa, Western Germany and Eastern Germany, and China confirm that radon levels at sufficiently high levels is an important cause of lung cancer (IARC, 2001; Lubin et al., 2004a). A recent study in China of underground dwellers is of special

276

PART III: THE CAUSES OF CANCER 1994; Neuberger and Gesell, 2002), although one investigation in Sweden was positive (Lagarde et al., 2001). Pooling data sets should help to define the possible level of risk associated with indoor radon (Darby et al., 2005; Krewski et al., 2005; Lubin et al., 2004a), but today the public health assessments rely upon the much stronger data from underground miners.

Natural Background Radiation Terrestrial Radiation

Figure 15–8. Relative risk of lung cancer by residential radon concentrations for seven case-control studies (Lubin, 1994; Blot et al., 1990; Schoenberg et al., 1990; Pershagen et al., 1992, 1994; Letorneau et al., 1994; Alavanja et al., 1994). The EPA action level for indoor radon is 150 Bq/m3 (4 pCi/l), which would result in approximately 0.8 Working Level Months in a year. (Source: Lubin and Boice, 1997, J Natl Cancer Inst 87:49–57, p. 52.)

note since radon levels were quite high, mobility was low, and smoking was controlled for in the analyses (Wang et al., 2002b). A significant trend was observed and findings were consistent with extrapolations from underground miners, although a risk under about 150 Bq/m3 (4 pCi/L) was not apparent. Indoor radon has not been linked to increases in childhood leukemia (Lubin et al., 1998; UK 2002a, UK 2002b). Recent meta-analyses of the residential case-control studies indicate that the lung cancer risk from indoor radon is not likely to be markedly greater than that predicted from miners, and confirm that the negative exposure response reported in some ecological studies is likely due to model mis-specification or uncontrolled confounding (Lubin and Boice, 1997; NAS, 1999; IARC, 2001). Increasing the number of studies in their initial meta-analysis from three to eight, Lubin and Boice (1997) analyzed a total of 4263 lung cancer case subjects and 6612 control subjects. The combined trend in the RR was significantly different from zero, and an estimated RR of 1.14 (95% CI: 1.0–1.3) at 150 Bq/m3 (4 pCi/L) was found, although there were significant differences in the study-specific estimates of the exposure response (Fig. 15–8). While meta-analyses are valuable for identifying differences among studies and for summarizing results, they should be interpreted cautiously when expected RRs are low as with indoor radon exposure, when there is study heterogeneity, when there is the potential for confounding and exposure misclassification, and when different types of measuring devices are used (Kreienbrock et al., 1999). Studies of indoor radon also must be interpreted with caution because of inherent difficulties in accurately estimating exposures that occurred many years ago based on current measurements (Lubin et al., 1990b, 1995c). The expected risk associated with average residential exposures is accordingly low (RR < 1.2), which necessitates accurate adjustment for the effect of smoking, including involuntary exposures. Exposure reconstruction is complicated further because of mobility (i.e., persons reside in many homes throughout life), home modifications, and uncertain estimates of time actually spent in the home. Studies of nonsmokers, which minimize possible confounding by tobacco and should have higher radon-related relative risks, have been generally negative or inconsistent (Blot et al., 1990; Alavanja et al.,

Descriptive epidemiologic studies (geographical correlation studies or ecological surveys) describe data that already exist on a population, and patterns of disease occurrence in time and in place are evaluated. Exposure is measured or estimated for groups of people and not individuals. Correlation studies attempting to link cancer mortality with natural background radiation have generally been negative (NAS, 1990), and are fraught with uncertainties in doses actually received by individuals, low and narrow ranges of cumulative doses, the potential that important demographic and lifestyle factors distort the correlations, migration patterns, selection factors for place of residence, and geographic variations in the accuracy of cancer diagnoses (Pochin, 1976; Boice, 2001b). The most extensive investigation on the possible health effects of naturally occurring radiation was conducted in China on a stable population of 80,640 persons who received three times the amount of background radiation as 32,651 inhabitants of a comparison region (Wei, 1980; Wei and Sugahara, 2000; Boice, 2001b). Cancer was not increased among residents of the high background area. Thyroid nodularity, a sensitive indicator of low-dose radiation effects, was also found to be similar among female residents of the high (14 cGy) and low (5 cGy) radiation areas based on clinical screenings of 2000 elderly women (Wang et al., 1990b). Differences in chromosome aberrations in circulating lymphocytes indicated that the background radiation levels were meaningfully different. A dose of 9 cGy accumulated gradually over a lifetime apparently produced many fewer thyroid tumors than seen following a similar dose of Xrays received briefly in childhood (Ron et al., 1989). Other large descriptive surveys have been reported from Ireland (Allwright et al., 1983), the United Kingdom (Muirhead et al., 1991; Richardson et al., 1995), Japan (Noguchi et al., 1986), India (Nambi and Soman, 1987), Sweden (Edling et al., 1982), and the United States (Mason and Miller, 1974; Amsel et al., 1982). Large-scale analytic studies of childhood leukemia and cancer have also revealed no correlation between natural levels of gamma rays (UKCCSI, 2002a) or radon (UKCCSI, 2002b; Lubin et al., 1998). The comprehensive UK Childhood Cancer Study, for example, measured gamma ray exposures in the homes of 3838 children with cancer and 7629 control children, and found no measurable risk from natural gamma ray exposures. These overwhelmingly negative results suggest that the carcinogenic risk of low natural levels of radiation is unlikely to be substantial (IARC, 2000).

Cosmic Radiation Air crew, pilots, flight attendants, and frequent flyers are exposed to neutrons and recoil protons from cosmic rays during high-altitude flights (UNSCEAR, 2000). Annual exposure may be about 1–2 mSv, which, even after flying for some 30 years, is likely too low a dose to detect a radiation effect (IARC, 2001). Nonetheless, there have been numerous studies conducted in North America, Europe, and Scandinavia (Boice et al., 2000; Pukkala et al., 2003; Blettner et al., 2003; Langner et al., 2003; Zeeb et al., 2003). Results are generally consistent in not finding any convincing evidence for a radiation effect. Increased rates of melanoma and breast cancer appear related to lifestyle factors such as increased sun exposure during leisure time activities and delayed childbearing. The statistical power to detect a radiation effect is likely even lower than previously assumed because estimated equivalent doses in mSv had been based on radiation weighting factors for neutrons and protons that were too high (Cox and Kellerer, 2003). Current understanding is that the per unit absorbed dose to cosmic radiation is less hazardous (i.e., the equivalent dose in mSv is lower than previously computed).

Ionizing Radiation

Fallout Marshall Islands Before the Chernobyl accident in 1986, the most convincing evidence that radioiodines could cause thyroid cancer came from the study of the Marshall Islanders (Robbins and Schneider, 2000). Over 200 native residents of four inhabited atolls east of Bikini Island were accidentally exposed to nuclear fallout from the BRAVO weapons test in 1954 (Conard, 1984; Robbins and Adams, 1989). Whole-body gamma ray doses were estimated as 0.11 Gy, 1.1 Gy, and 1.9 Gy to the Marshallese on three atolls. Mean thyroid dose, from gamma radiation plus radio iodines, was estimated as 3–52 Gy to children, depending upon age, and 1.6–12 Gy to adults. During 32 years of observation, 60 (or 24%) of 253 subjects developed thyroid nodules and cancer, excluding seven occult papillary carcinomas found incidentally during surgery. Thyroid cancers appeared in 7 of 130 women and 2 of 113 men; and in 6 of 127 children under age 19 at exposure and in 3 of 126 older natives. The offspring of 2 of 12 women pregnant at the time of the test developed thyroid nodules. The earliest thyroid lesion appeared 9 years after exposure. Impairment of thyroid function and some clinically evident hypothyroidism occurred at doses between 3.9 and 21 Gy. Growth retardation was apparent in some children. One leukemia occurred in each group of exposed and non-exposed (135) islanders. Two pituitary tumors developed in women exposed as young children, suggesting a possible link with thyroid injury. Calculated risk coefficients were 8.3/104 PY-Gy and 1.5/104PY-Gy for nodules and cancer, respectively. Risk estimates are uncertain, however, because the large thyroid doses may have caused lethal cellular damage that decreased the number of cells at risk for malignant transformation, frequent surgery for benign tumors and nodules might have removed tissue destined to develop into cancer, increased levels of TSH, secondary to thyroid hypofunction, and prophylactic thyroid hormone treatments might have influenced risk, and the effects of gamma ray exposures could not be distinguished from those of internal radioiodines. Further, dosimetry for the radioiodines is complex. Most beta particle energy from 131I is supposedly deposited in the colloid of the large follicles without reaching the critical follicular cells (NRC, 1996), and the low dose rate would tend to minimize risk. In contrast, the shorter-lived and more energetic beta-emitting isotopes (132I, 133I, and 135I) contributed two to three times the dose of 131I and exposed the thyroid more uniformly and at a higher dose rate. Thus, while the study clearly shows that short-lived iodines are carcinogenic, the contribution of 131I could not be distinguished. A study of 7000 Marshall Islanders from 14 atolls, including several not previously studied, revealed a linear relationship between thyroid nodules and proximity to Bikini (Hamilton et al., 1987). An estimate of 11 nodules/104PY-Gy was based on clinical examinations.

Utah Thyroid. Clinical examinations of 2945 children in the sixth to twelfth grades from 1965 to 1971 were conducted in two counties in Utah and Nevada that received fallout from nuclear weapons tests in the 1950s, and also on 2271 children in Arizona exposed to negligible fallout (Rallison et al., 1975). No cancers and 18 nodules (1.3%) were detected in 1378 children living in the exposed counties at the time of the weapons tests. In contrast, two thyroid cancers and 19 thyroid nodules (1.4%) occurred in 1313 non-exposed children who moved into the area after the time of fallout; and no cancers and 20 nodules (0.9%) were found in 2140 non-exposed Arizona children. Nearly 75% of the original cohort of children was reexamined in 1985 to 1986 and the authors concluded that radio iodines from fallout were responsible for a small excess of thyroid neoplasms (Kerber et al., 1993). The study is remarkable for the comprehensive dose reconstruction based on milk and green vegetable consumption, radionuclide deposition, and milk production and deposition. Estimates of individual thyroid doses ranged from 0 to 4.6 Gy and averaged 0.17 Gy in Utah. A significant association with dose was found for period prevalence of all neoplasms taken together (n = 19), but not for benign (n = 11) or malignant (n = 8) neoplasms separately. No excess of nonneoplastic thyroid nodules (n = 34) was found. Although the estimated

277

radiation risks were roughly consistent with those reported in studies of children exposed to external X-rays or gamma rays, the small number of cases limit the conclusions that can be drawn. In addition to the small sample size, other concerns include: (1) the apparent lack of a correlation between thyroid neoplasms and proximity to the Nevada test site (NTS) (Rallison et al., 1990), (2) the dosimetry relied on recall of dietary habits of 30 years ago, which must be uncertain, (3) recall bias was possible because the dietary questionnaire was administered after the thyroid examination, (4) observation bias was possible in that nurse practitioners referred nearly twice as many subjects for clinical thyroid examination from the high fallout areas— 32% from Utah and 17% from Arizona, and (5) the absence of a clear association between dose and the larger number of non-neoplastic thyroid nodules is unusual because nodules have been related to fallout radiation among residents of the Marshall Islands (Conard, 1984) and among children exposed to cranial irradiation (Ron et al., 1989). Gilbert et al. (1998) correlated thyroid cancer mortality and incidence rates within US counties to 131-I dose estimates, taking geographic location, age at exposure, and birth cohort into account. Estimates of cumulative dose were not associated with thyroid cancer incidence or mortality. A subgroup analysis of persons potentially exposed to fallout before 1 year of age indicated an increased risk, but there was no a priori reason to signal out this age group for emphasis and the absence of an increased risk from dose received at ages 1 to 4 years is not consistent with studies of children exposed to external radiation sources (Ron et al., 1995). As in all ecological studies, dose to individuals was unknown. Estimates of milk consumption were made for populations living in large geographical areas (i.e., the county), and substantial migration of young people from their counties of birth has occurred for educational and occupational reasons.

Leukemia. A geographical study of childhood cancer mortality within Utah suggested an increase of leukemia in “high-fallout” counties near the NTS (Lyon et al., 1979). However, there was a significantly low risk of other childhood cancers in the high-exposure counties, and no overall association with total childhood cancers (Land, 1979; NCRP, 1980). Attempts to duplicate the leukemia findings using county mortality statistics from the National Center for Health Statistics (NCHS) also failed (Land et al., 1984). It appeared that the earlier finding reflected an anomalously low rate of childhood leukemia rates in southern Utah during 1944 to 1949, and not an association with fallout. An increased ascertainment of childhood leukemias in the high-exposure counties also may have been related to the arrival of a hematologist during this time. A telephone survey of Mormon Church members of certain communities in the southwest claimed that fallout had caused exceptionally high risks of leukemia and cancer (Johnson, 1984). Self-reported cancers, obtained by volunteer interviewers, were not verified. Some purported risks were far greater than seen among A-bomb survivors exposed to near-lethal doses of more than 4 Gy, and suggested biases in the study methodology (Cook-Mozaffari et al., 1987; MacMahon, 1989). A geographical analysis of NCHS mortality data for counties in Utah covered by the telephone survey found a significant deficit of cancer, although leukemia was increased (Machado et al., 1987). A case-control study of over 1000 individuals who died of leukemia in southwestern Utah identified a weak association between estimated bone marrow dose and all leukemia, though the trend was not significant (Stevens et al., 1990). Significant risks, however, were observed for acute leukemia among those under age 20 when exposed to fallout, consistent with that expected based on other studies of exposed populations. The increasing trends seen for CLL, a tumor not known to be elevated after irradiation, and the difficulty in estimating doses retrospectively add caution to causal interpretations. Semipalatinsk The former Soviet Union tested nuclear weapons in Kazakhstan from 1949 to 1962 at the Semipalatinsk Test Site (Grosche et al., 2002). A study of thyroid disease prevalence has been conducted among heavily exposed residents near the test site, but no results have been published to date (Gilbert et al., 2002). Difficulties in reconstructing doses appear

278

PART III: THE CAUSES OF CANCER

substantial and early estimates of high exposures have not been borne out (Salomaa et al., 2002; Simon et al., 2003). A recent survey of Kazakhstan settlements exposed to fallout reported increased risks for practically all cancers, including those not convincingly linked to radiation (Bauer et al., 2005). Limitations include the lack of individual doses, incomplete follow-up, inconsistency and quality of case ascertainment, and uncertain comparability with the selected comparison group.

Nordic Countries Trends in childhood leukemia within Nordic countries were evaluated for possible changes that might be related to fallout from atmospheric nuclear weapons testing in the 1950s and 1960s (Darby et al., 1992). Estimates of fetal bone marrow exposure, primarily from cesium 137 (137Cs), were low and about 0.14 mSv, and no increase in leukemia incidence could be tied to such levels. A 7-year cumulative exposure was estimated to be 1.5 mSv, and although no correlation with exposure was found, small increases were not inconsistent with a possible radiation effect. There was no evidence for a preconception effect based on estimated parental testicular dose. Another correlation study of birth cohorts in the Nordic countries was conducted that suggested an association between fallout and thyroid cancer, although alternate explanations seemed plausible (Lund and Galanti, 1999). These analyses suffer from the same uncertainties of all ecological surveys in that the doses to individuals are unknown. Further, there have been a great many new exposures in childhood occurring after World War II, other than low-level radioactive fallout, which might influence the incidence, diagnosis, and reporting of leukemia and thyroid cancer over time. For example, time trends and geographical distributions of thyroid cancer in England and Wales appear more likely the result of pre-existing benign thyroid diseases, or factors that cause them, than the results of radioactive fallout from weapons testing (dos Santos Silva et al., 1993).

Nuclear Weapons Test Participants Nevada. No excess in total cancer mortality (112 vs. 117.5) was found among 3017 of 3217 participants in military maneuvers during the 1957 nuclear test “SMOKY” (Caldwell et al., 1983). Leukemia, however, was significantly elevated; 10 cases were observed, including the index case that prompted the investigation and one case that developed after radiation therapy for lymphoma, versus 4.0 expected. Lower cancer frequencies were generally noted among the military units with the highest exposures based on film badge doses (mean, 0.46 cGy). A survey of 46,186 military participants in two test series conducted at the NTS and three in the Pacific also found no excess of non-leukemia deaths (990 vs. 1187) (Robinette et al., 1985). Excluding SMOKY, 46 leukemia deaths occurred versus 52.4 expected, suggesting that the leukemia excess among SMOKY participants was either due to chance or to circumstances peculiar to that shot (or its participants). Another mortality study involved nearly 70,000 soldiers, sailors, and airmen who participated in one of five US nuclear weapons tests in the 1950s and nearly 65,000 comparable nonparticipants, or referents (IOM, 2000). Participants and referents had similar risks of death from cancer (SMR = 0.74 for both groups). Leukemia was also below expectation among participants and referents with SMRs of 0.74 and 0.64, respectively. A non-significant increase in leukemia (RR = 1.15) was highlighted contrasting participants with referents directly, but seems likely attributed to a methodologic bias (i.e., a differential ascertainment of cause of death between exposed and nonexposed groups), with 4.6% and 7.6% of the deaths being of unknown cause, respectively. Proportionally distributing the unknown causes of death into the observed categories for both groups results in a recomputed relative risk consistent with 1.0. Multiple myeloma also was not linked to participation during the weapons tests.

United Kingdom. Cancer mortality and incidence among 21,357 participants in the United Kingdoms’ atmospheric nuclear weapons tests in Australia and the Pacific Ocean between 1952 and 1967 and in 22,333 matched controls were evaluated in a third follow-

up (Muirhead et al., 2003). Mortality from all causes (SMR 0.89 and 0.88), and from all cancers (SMR 0.93 and 0.92) were similar. Multiple myeloma was no longer different between participants and controls (RR = 1.11 based on 35 incident cases for both groups) and increases seen in earlier reports were attributed to chance (Darby et al., 1993). Mortality from leukemia (excluding CLL) among participants was equal to that predicted from national rates (SMR = 0.98 based on 40 cases), but was significantly low among controls (SMR = 0.58 based on 23 cases). A significant RR in leukemia mortality was related, then, to the deficit among the controls rather than an excess among the exposed, and may be a chance finding in view of the low rates among the controls and the generally small radiation doses recorded. Liver and prostate cancer risks were significantly higher among participants than controls, whereas controls had significantly higher levels of lung (marginally) and kidney cancer. Studies of presumably very low doses usually show similar random patterns of cancer occurrence.

Cancer around Nuclear Installations Systematic Surveys. Reports of small clusters of childhood leukemia around nuclear installations in the United Kingdom in the 1980s prompted several large-scale systematic surveys around the world (UNSCEAR, 1994). Lymphoid leukemia among persons under age 25 was found to be generally increased in populations living near nuclear fuel reprocessing or weapons production facilities in the United Kingdom, but not plants that generated electricity (CookMozaffari et al., 1987; Forman et al., 1987). Mortality from Hodgkin disease at ages 0 to 24 was also increased, whereas mortality from lymphoid leukemia at ages 25 to 64 was significantly reduced. Overall, there was no general increase in cancer deaths in the vicinity of nuclear installations. Interestingly, a study from Britain evaluated residents of areas where construction of nuclear power stations had only been considered, or just recently completed. Excesses of childhood leukemia and Hodgkin disease, and deficits of adult leukemia, were reported that were similar to those previously identified in areas with operating nuclear facilities (Cook-Mozaffari et al., 1989). The authors concluded that the unexpected increases in some childhood cancers around nuclear installations are unlikely to be due to environmental radiation pollution, but rather to other risk factors yet to be identified. An infective agent associated with large migrations of people into these areas, for example, has been proposed as a likely explanation (Kinlen et al., 1991, 1993a, 1995, 1997; 2002; Doll et al., 1994; Doll, 1999), although not universally accepted (Inskip, 1997; Cartwright et al., 2001; Law et al., 2003). In the largest ecological survey to date, more than 900,000 cancer deaths in 113 counties in the United States containing or adjacent to 62 nuclear facilities were compared to 1,800,000 cancer deaths in control counties with similar population and socioeconomic characteristics (Jablon et al., 1991). Overall, and for specific groups of nuclear installations, there was no evidence that mortality for any cancer, including childhood leukemia, was higher in counties with nuclear reactors than in the control counties. For childhood leukemia, the RR in the study counties versus their controls after plant start-up was 1.03; before start-up it was 1.08. For all leukemia, the RRs were 0.98 after start-up and 1.02 before start-up. Systematic studies in France (Hill and Laplanche, 1990; Hattchouel et al., 1995, 1996), Germany (Michaelis et al., 1992; Michaelis, 1998; Martignoni, 2003), Canada (McLaughlin et al., 1993b), and other US areas (Grosche et al., 1999; Boice et al., 2003a, 2003b) also failed to identify clusters of childhood cancer around nuclear facilities. One analysis reported an association with childhood leukemia around a reprocessing plant in France (Viel et al., 1995; Pobel and Viel, 1997; Guizard et al., 2001), but an infectious agent associated with population mixing might have been the cause (Boutou et al., 2002). Small-area data were analyzed in England and Wales and, except for Sellafield, there was little evidence that childhood leukemia was related to proximity to nuclear installations (Bithell et al., 1994). The ecological correlation analyses are not without problems: (1) radiation dose to the population is unknown, although likely much

Ionizing Radiation below natural background (Darby and Doll, 1987; UNSCEAR, 1988), (2) mortality is not the best indicator of cancer hazard due to inaccuracies of death certificates, and cancer registration may also be incomplete and variable between areas, (3) other important risk factors often cannot be identified, (4) the county or district may be too large an administrative unit to detect localized increases in cancer rates, (5) the many comparisons made with regard to individual cancers, ages, and time frames increase the likelihood that chance plays a role in highlighting areas with seemingly high, or low, cancer rates, and (6) unknown factors associated with migration and selection of residence and occupation could contribute to cancer occurrence in these areas (IARC, 2000).

Clusters. In 1983, a team of investigative television reporters from Yorkshire set out to evaluate the risk of cancer in workers at the Sellafield (Windscale) nuclear fuel reprocessing complex in West Cumbria, UK. Learning that neither cancer nor leukemia was excessive in these workers (Smith and Douglas, 1986), the reporters focused on an apparent cluster of seven young people who developed leukemia between 1950 and 1983 in Seascale, a village about 3 kilometers south of Sellafield. A government report confirmed that childhood leukemia (4 vs. 0.25) was elevated in the region near Sellafield (Black, 1984). An assessment of total radiation exposure of the population revealed that natural background contributed the greatest amount (66%), with Sellafield discharges contributing only 16%. Thus, environmental pollution from radioactive releases seemed an unlikely culprit. Additional studies found that the excess of leukemia occurred entirely among individuals born in Seascale (5 vs. 0.53) and not among children born elsewhere (0 vs. 0.54), suggesting that factors present in early life or before birth might be important (Gardner et al., 1987). A subsequent case-control study, discussed below, raised the possibility that parental exposure among Sellafield workers might explain the cluster (Gardner et al., 1990). Other studies around nuclear facilities have failed to provide clear insights into the reasons, other than chance or selection, for apparent clusterings of childhood cancer (MacMahon, 1992; Draper et al., 1993; Inskip, 1997). In some investigations, findings were entirely dependent upon the selection of particular geographic and calendar time groupings. Even the Seascale cluster might be considered suspect, because it was the occurrence of the cases that determined both the geographic boundary and the age definition of the cluster. Recall that the TV reporters first went to Sellafield, not Seascale, and were seeking excesses of cancer among adult workers, not leukemia among young people in the general population. Preconception. The most provocative (and controversial) finding from the Seascale studies was the association between leukemia and preconception irradiation of the fathers working at Sellafield (Gardner et al., 1990). If true, the apparent cluster might be explained in terms of occupational rather than environmental radiation exposure. The numbers were small, however, and the association relied on only four high-exposed fathers. Other correlates of occupation, such as chemical exposures, were not evaluated. Most of the nine fathers who worked at Sellafield were chemists or were involved with chemical processing; parental exposure to chemicals has been suggested as a possible risk factor for leukemia in offspring (Buckley et al., 1989). The leukemia and lymphoma diagnoses evaluated occurred over a 35-year period and included “children” up to age 25. Conceivably, medical care, cancer diagnoses, and cancer registration might be slightly better among skilled workers at a nuclear facility, resulting in spurious associations between parental exposure and childhood cancer. The study was also at odds with the prospective investigation of children of the atomic bomb survivors where no excess of cancer, chromosome aberrations, or genetic mutations in blood proteins were observed (UNSCEAR, 1988; Neel and Schull, 1991; Izumi et al., 2003; Schull, 2003). Childhood cancer also has not been increased among offspring of long-term survivors of cancer treated with radiation (Mulvihill et al., 1987; Boice et al., 2003d). A preconception effect also has not been seen among Ontario radiation workers (McLaughlin et al., 1993c), Scottish radiation workers (Kinlen et al., 1993b),

279

Danish Thorotrast patients, British radiation workers (Draper et al., 1997; Sorahan et al., 2003), and US radiation workers at Hanford, Idaho and Oak Ridge (see Wakeford, 2000). Experimental studies indicate that chronic exposures, such as experienced among Sellafield workers, produce fewer mutations than equivalent acute doses (UNSCEAR, 1988). One experimental study, however, has suggested that preconception X-rays might induce heritable tumors (mainly of the lung) in mice (Nomura,1982). On the other hand, if radiation were acting to cause heritable mutations, an epidemic of known heritable diseases and congenital malformations, and not leukemia, might be expected in the Sellafield area whereas none was noted (Evans, 1990). A similar but small study at the Dounreay nuclear facility failed to replicate the Sellafield preconception findings (Urquhart et al., 1991). Another recent report purports to corroborate the preconception effect found at Sellafield (McKinney et al., 1991). Unfortunately, radiation exposures were not verified or quantified, and, perhaps more importantly, the study included many of the same fathers previously evaluated in the investigation by Gardner and coworkers (1990) that raised the hypothesis (Smith, 1991). Excluding these overlapping individuals, as would be appropriate for an independent assessment, meaningfully reduced the evidence for an association. A similar criticism can be raised for yet another evaluation of the same population, which included the Gardner cases (Dickinson and Parker, 2002). A study of 10,363 children who were born to fathers who worked at Sellafield evaluated the geographical distribution in Cumbria of the paternal dose received prior to conception (Parker et al., 1993). Paternal doses were consistently higher among fathers of children born outside Seascale. Since childhood leukemia was not increased in these areas of West Cumbria despite the higher preconception exposures, the authors concluded that paternal exposure to radiation before conception is unlikely to be a causal factor for childhood leukemia. An explanation of the Seascale cluster in terms of preconception radiation of the fathers appears now to have been a provocative hypothesis that was unsubstantiated by further studies (Doll et al., 1994; UNSCEAR, 1994; Little et al., 1996; Wakeford, 2002, 2003). Other than chance, one hypothesis being pursued is the possibility that childhood leukemia may occur as a rare response to an unidentified infection whose transmission is facilitated when large numbers of people come together, such as might occur when large industrial complexes are built in rural areas (Kinlen, 1997; Doll, 1999).

Nuclear Reactor Accidents Three Mile Island. The nuclear reactor accident at Three Mile Island released little radioactivity into the environment and resulted in population exposure that was much less than what was received from natural background. Any presumed increase in cancer at these levels would be negligible and undetectable (Upton, 1981). An ecological survey did not link increased cancer rates with estimated patterns of radiation releases (Hatch et al., 1990), nor were any peculiar mortality patterns noted (Jablon et al., 1991). Other studies (e.g., Wing et al., 1997; Hatch et al., 1997; Talbot et al., 2003) have given inconsistent results and provide little evidence for an effect of radiation, related in part because individual doses are unknown but very low (<1 mSv), smoking information was not available, and chance comes into play when many different cancer types are studied (UNSCEAR, 2000; IARC, 2000).

Chernobyl. In contrast, the 1986 accident at Chernobyl resulted in a massive release of radioactivity. Thirty firemen who quelled the smoldering fires of the burning reactor died shortly after being exposed to lethal amounts of radiation, over 100,000 citizens living in the surrounding area were evacuated, vast areas in the Ukraine, Russia, and Belarus were contaminated with radioactive debris spewing from the blazing nuclear core, and hundreds of thousands of workers from all over the former Soviet Union were sent to clean up the contaminated environment and to entomb the extinguished nuclear reactor in a “sarcophagus” (UNSCEAR, 2000). Studies have not linked increases in childhood leukemia to estimated exposure to radiation from the

280

PART III: THE CAUSES OF CANCER

Chernobyl fallout (Parkin et al., 1996; Ivanov et al., 1993, 1996, 1998). The only late health affect ascribed to the reactor accident is a remarkable increase in the incidence, but not mortality, of thyroid cancer following childhood exposure to radioactive iodines (Moysich et al., 2002; Williams, 2002). Although initial controlled studies found no association between radiation and thyroid neoplasia, the screening survey may have been conducted too early to detect an effect (Mettler et al., 1992). Nonetheless, the abrupt increase in childhood thyroid cancer in Belarus and the Ukraine (Kazakov et al., 1992; Likhtarev et al., 1995; Jacob et al., 1998, 1999) may have been partially due to increased medical surveillance and reporting (Ron et al., 1992). The first carefully conducted case-control study provided evidence for a dose response over categories or estimated radiation dose to thyroid (Astakhova et al., 1998). Radiation risk estimates are uncertain because of the very short latency, which suggests surveillance bias, the limited ability to estimate thyroid doses for individuals and assigning ecological estimates of dose from groups to individuals, the contribution of short-lived radioactive iodines other than I-131 to dose, and the possible influence of iodine deficiency and genetic susceptibility (UNSCEAR, 2000; Moysich et al., 2002; Shakhtarin et al., 2003). A recent case-control study reported a dose-response between radioiodines, primarily I-131, and thyroid cancer (Cardis et al., 2005b). The increased risk, however, appeared concentrated among children with diets deficient in stable iodine. A significant reduction in radiation risk was also seen among children taking potassium iodine tablets years after exposure, suggesting that a dysfunctional thyroid gland may have played an important role (Boice, 2005). Over 600,000 workers were sent to Chernobyl after the accident to clean up the environment and entomb the reactor (UNSCEAR, 2000). Allowable occupational exposures were 0.35 Gy, suggesting that future health studies might be informative. Initial cohort studies of cancer and thyroid nodularity among cleanup workers from the Baltic countries, however, failed to find associations with estimated radiation exposure from recorded doses and biological measures of dose (Inskip et al., 1997; Rahu et al., 1997). The follow-up may have been too short, sample size too small, but perhaps more importantly, the level of radiation exposure appeared too low to expect to detect a radiation effect in the liquidator population (Littlefield et al., 1998; Cardis et al., 1996). In a much larger study of 168,000 Russian recovery operation workers, however, Ivanov et al. (1997a) reported an increased risk of leukemia and thyroid cancer compared with national rates. Comparing cancer incidence in these workers to that in a general population is problematic, because the cleanup workers had a higher level of medical surveillance, especially of their thyroid glands (Boice and Holm, 1997; UNSCEAR 2000). A large excess would not be expected for thyroid cancer since adult exposure to either external radiation or to I-131 appears to carry little if any future risk (UNSCEAR, 2000; Dickman et al., 2003). Subsequent study also failed to find a correlation with estimated radiation dose to thyroid among Russian recovery workers (Ivanov et al., 2002). Bias in the leukemia findings was evident when the same data were re-analyzed in a case-control manner taking into account individual radiation doses to workers: a significant radiationrelated risk for leukemia was no longer apparent (Ivanov et al., 1997b). The discrepancy between the cohort study and the case-control study was ascribed to an over-ascertainment of leukemia among recovery workers compared with under-reporting of leukemia in the general population used for comparison in the cohort, misdiagnosis of leukemia, and the inclusion of CLL, which is not considered a radiation disease (Boice, 1997c; Boice and Holm, 1997; UNSCEAR, 2000). Combined studies of leukemia among recovery workers are ongoing (Kesniiniene et al., 2002).

Releases into the Environment Techa River. The occupational study of the worker population at the Mayak facility has provided information on health effects of plutonium (Koshurnikova et al., 2002a). There are three other populations in the Southern Urals in Russia that may provide information on the effects of chronic low-dose exposure from radioactive releases into the environment: the Techa River population (Kosenko, 1996), the

Ozyorsk population (Koshurnikova et al., 2002b), and the population in the East Ural Radioactive Trace (EURT), which resulted from an explosion in 1957 (the Kyshtym accident) in a storage tank at the Mayak facility (Kellerer, 2002a). During 1949 to 1956, high-level radioactive waste was dumped into the Techa River, and nearly 124,000 people living downstream received internal exposure from the ingestion of radioactive strontium and cesium and external exposure to gamma rays from the decay of Cs-137 and other radionuclides (Kossenko and Degteva, 1994; Kossenko, 1996; Kossenko et al., 1997; 2002). Approximately 30,000 are now being observed for health effects and increased risks for leukemia and solid cancers have been reported (Krestinina et al., 2005). The uncertainties in dosimetry are substantial and the latest revisions may have underestimated dose (and thus overestimated risks) (Kellerer, 2002a). Other serious limitations include: (1) the poor follow-up with nearly 50% of the cohort being lost, (2) unknown cause of death for 30% of the deaths, (3) unaccounted airborne doses from radionuclide releases from the Mayak stacks and from the EURT (Kyshtym accident), (4) differential exposures to medical X-rays and medical care among the villages downstream, (5) toxic chemical exposures in the liquid wastes released from the Mayak facility or from agricultural chemicals, (6) problems of separating the health effects from internal and external exposures, and (7) whether it is valid to base risk estimates on organ-specific doses when the population experienced substantially higher bone-marrow exposure and likely “abscopal” radiation effects (Kossenko et al., 2002; Kellerer, 2002a; UNSCEAR, 2000).

Hanford Thyroid Disease Study. Large amounts of radioactive iodine, specifically I-131, were released into the atmosphere between 1944 and 1957 from the Hanford Nuclear Site. A study of thyroid disease among 5199 persons born between 1940 and 1946 in seven counties in Washington State was conducted (Davis et al., 2004). A complex dose reconstruction program was performed and the follow-up involved thyroid screenings of 3440 of the 5199 former children. The most important contributor to dose was milk consumption. The mean and median doses were 18.6 cGy and 10.0 cGy, respectively (range, 0–284.0 cGy). Examination included ultrasound, thyroid palpation, and blood tests. Eleven categories of thyroid disease, ultrasound-detected abnormalities, and hyperparathyroidism were evaluated. There were 19 thyroid cancer and 249 benign thyroid nodules among the participants. There was no evidence for an association between estimated I-131 dose to thyroid and any thyroid disease or condition. The study, similar to medical studies (Dickman et al., 2003), involved exposure to I-131 only and the contribution of other radionuclides of iodine or of external exposure appeared negligible. In contrast to studies of external radiation (Ron et al., 1995), there was no evidence of an increased risk of thyroid neoplasia for I-131 doses of the order of 10 cGy (UNSCEAR, 2000). The limitations of the study include: (1) difficulties of reconstructing doses after several decades based on dietary recall and past information on radionuclide releases some 40 years ago and (2) the relatively low participation rate of about 66%. Nonetheless, contrasts between high and low exposure countries also failed to reveal an association. It is noteworthy that the estimated ERR at 1 Gy of 0.7 for I-131 is a factor of 11 lower than seen among children treated with radiotherapy or exposed to the atomic bomb (Ron et al., 2005). BASIC CONCEPTS Radiation Radiation generally refers to energy emitted from a source, such as heat or light from the sun, radio waves from a broadcast antenna, microwaves from a radar unit or cellular telephone, X-rays from an X-ray tube, or gamma rays from radioactive elements. Radiation that can remove electrons from atoms is called ionizing and includes electromagnetic rays such as X-rays and gamma rays and energetic particles such as protons, fission nuclei, and alpha and beta particles. Neutrons, unlike these other particles, have no charge and cannot ionize directly. Instead they impart energy to protons through elastic

Ionizing Radiation collisions, and the protons then cause the subsequent ionizations. The amount of energy absorbed in matter as a result of radiation interactions is called the dose, which is measured in gray (Gy): 1 Gy = 1 joule per kilogram. In the past, the standard unit for dose was the rad (1 rad = 100 ergs per gram), but the conversion is simple: 1 Gy = 100 rad. An acute whole-body dose of about 4 Gy (400 rad) is lethal about half the time in humans. Nonionizing radiations, such as radiowaves, do not possess enough energy to strip electrons from atoms. Visible light forms the boundary between ionizing and non-ionizing radiation with higher frequency (shorter wavelength) radiation possessing enough energy to cause ionization whereas lower frequency (longer wavelength) radiation is not able to remove electrons from atoms. Microwaves, such as used in ovens (2450 MHz) or cellular telephones (850–1900 MHz), and extremely low-frequency electromagnetic fields (60 Hz) from household appliances or electrical power transmission lines are all nonionizing. Although nonionizing electromagnetic fields have generated great public concern, no consistent or credible evidence has linked exposures to nonionizing electromagnetic fields to cancer in humans or animals (Ahlbom et al., 2001; NRPB, 2001; Boice and McLaughlin, 2002; Breckenkamp et al., 2003; NRPB, 2003b). Ionizing radiation is absorbed randomly by atoms and molecules in cells and can alter molecular structure. These alterations can be amplified by biological processes to result in observable effects. The biological effects, however, depend not only on the total absorbed dose but also on the linear energy transfer (LET), or ionization density, of the type of radiation. LET is a measure of the energy loss per unit distance traveled and depends on the velocity, charge, and mass of a particle or on X-ray or gamma-ray energy. High-LET radiations such as alpha particles (helium nuclei) release energy in short tracks of dense ionizations. Low-LET, or sparsely ionizing, radiations such as X-rays produce ionizing events that are not close together. Depending on the biological end point, the effect per Gy may differ widely as a function of LET but is usually greater for high-LET radiation. Not all types of radiation are similar in their ability to produce a specific effect (e.g., cell death, chromosome aberration, or cancer) and the magnitude of the effect can be influenced by the rate at which the radiation is received over time (Fig. 15–9). The relative biological effectiveness (RBE) of radiation characterizes its ability to produce a specific disorder compared to a standard, usually X-rays or gamma rays. The unit of biological equivalent dose is the sievert (Sv). The sievert represents the absorbed dose in gray multiplied by an appropriate radiation weighting factor (specific to the type of radiation) and other possible modifying factors (ICRP, 1991, 2003a). For X-rays, gamma rays, and electrons this weighting factor is 1, whereas for alpha particles the weighting factor is 20. The previous unit of equivalent

Figure 15–9. Influence of dose, dose rate, and type of radiation on the cumulative incidence of myeloid leukemia in male RF mice. (Source: Upton AC et al., 1970, Radiat Res 41:467–491, p. 476.)

281

dose was the rem with 1 Sv = 100 rem. An RBE of 20 for neutrons at 0.1 Gy (10 rad), for example, would imply that the biological effect from 0.1 Gy of neutrons is the same as that from 2.0 Gy (200 rad) of gamma rays. The sievert has also been applied to assess the effects of mixed-field exposures. For example, the equivalent dose of an exposure to 0.1 Gy of gamma ray and 0.1 Gy of neutrons, with gamma rays as the standard and an RBE of 20 for neutrons, would be 2.1 Sv (210 rem). Another unit that was designed only for use in radiation protection is the effective dose, also expressed in sievert, which essentially expresses the equivalent dose to a particular tissue in terms of wholebody risk. Effective dose allows one to compare risks of partial body exposures either from external or internal radiations with those from whole body exposure by applying both radiation and tissue weighting factors (ICRP, 1991). For example, indoor radon results in an annual equivalent dose of about 24 mSv to lung tissue (1.2 mGy absorbed dose to lung times a radiation weighting factor of 20 for alpha particles), which converts to an effective dose of 2.4 mSv (24 mSv times a tissue weighting factor of 0.1) (NAS, 1990). This quantity is not appropriate for individual risk assessment because it does not relate directly to tissue dose and it cannot be measured, only calculated. Effective dose is helpful when estimating the total consequences or detriment from different types of radiation exposures to different tissues in an individual. However, effective dose is a quantity intended for use in radiological protection and was not developed for use in epidemiological studies or other specific investigations of human exposure. For these other studies, absorbed dose in the organs of interest and specific data relating to the RBE of the radiation type in question are the most relevant quantities to use (Cox and Kellerer, 2003).

Estimation of Cancer Risks: Epidemiological Studies Epidemiological studies have conclusively linked high-dose radiation to increased cancer risks. Estimates of risk for low-level exposures, however, are based on extrapolations and require assumptions about the shape of the dose-effect relationship and about the mechanisms of radiation carcinogenesis. These assumptions are often guided by experimental investigations and radiobiology theory. Similar to other epidemiological investigations, however, studies of radiogenic cancers are also susceptible to the problems of inappropriate comparison groups; biases of selection, observation, and response; multiple comparisons; cluster analysis pitfalls; inadequate control of confounding factors; inadequate measures of dose and cancer outcome; and low statistical power due to small numbers of excess cancers at low doses (Land, 1980; Beebe, 1984; UNSCEAR, 1988, 1994, 2000; MacMahon, 1989; NAS, 1990). These problems are especially acute when risks at very low dose levels (and thus very low effect levels) are being studied. The United Nations Scientific Committee on the Effects of Atomic Radiation (1994, 2000) provides a very informative review of the pitfalls in conducting epidemiologic studies of irradiated populations overall as well as summarizes the strengths and weaknesses of specific studies. It is emphasized that not all epidemiological studies are equally informative or of equal quality. Some have low statistical power and provide little information on risks; others are so susceptible to potential or actual biases that findings have little or no validity. It is important to consider methodological issues when interpreting the evidence from different studies, and it is the consistency of findings in different studies conducted by different investigators in different parts of the world that is most informative. An appropriate comparison group is important. Comparisons with general population rates could be misleading when the irradiated population is ill, because the disease itself could be related to the occurrence of cancer (confounding by indication), and other competing risks or treatments could influence subsequent cancer incidence (such as chemotherapy among cancer patients). Many employed populations are also self-selected, with different expectations of disease than the general population (selection bias for employment and survival). Interpretation should be made cautiously, however, when an elevated “risk” is related more to a decreased incidence in the comparison group than an increased incidence among the exposed.

282

PART III: THE CAUSES OF CANCER

More thorough case-finding (differential disease ascertainment) among an irradiated population than among the general population used to compute expected numbers could bias study results in a positive direction (ascertainment bias). Another important source of bias is failure to obtain follow-up information on large numbers of the population, especially if the follow-up is differential by exposure. Followup could be biased in some hospital-based studies if patients who developed a second cancer are more likely to return to the hospital than patients who are free of disease. For ascertainment of disease by questionnaire, response bias could further exacerbate the follow-up bias, especially if healthy subjects were more (or less) likely to respond than those who are ill. Special screening of exposed but not non-exposed persons is an extreme but not uncommon example of observational bias. It could, for example, account for many of the excess benign tumors seen in persons with histories of head and neck irradiation during childhood. In case-control studies of cancer and diagnostic X-rays, respondents might be more reliable witnesses if they recently developed a cancer (response bias). Radiation dose should be accurately estimated, and dose-response gradients add meaningfully to causal interpretations. If doses are not known for individuals, an excess in a presumably low-dose group could be concentrated in those few individuals who received high doses or a negative study might miss an excess in these few individuals. There are a number of notable studies for which no or extremely little information on radiation dose is known for individuals (e.g., studies of prenatal X-ray and childhood leukemia, thyroid cancer and Chernobyl radiation, leukemia among radiologists, cancer among X-ray technologists, indoor radon and lung cancer, cancer among weapons test participants, and of course all ecological and descriptive correlation studies). Using effective dose instead of organ-specific absorbed dose in epidemiologic studies is also incorrect (Cox and Kellerer, 2003). When the radiogenic response is low, chance events are also more likely to be misinterpreted, and the influence of confounding factors more worrisome (Beebe, 1984; UNSCEAR, 2000). Care should be exercised in performing statistical tests on investigations generated because of an apparent cluster of disease in the study population that prompted the investigation. The statistical problems of cluster analysis are complex, especially in light of possible a posteriori selection of time and geographic groupings. When multiple comparisons are made, investigators should be careful not to over-interpret unexpected findings, whether positive or negative. The confounding influence of other carcinogenic exposures or lifestyle factors should be evaluated. Estimates of lung cancer risks obtained without adjusting for smoking or estimates of occupational risks obtained without controlling for toxic chemical exposures may be misleading. Other radiation exposures, such as excessive medical X-rays or environmental background radiation, are not often considered in low-dose studies, despite the possibility that their contribution to the total radiation burden may be greater than the exposure under study (Spiers, 1979). It is important that exposed populations be observed for substantial periods of time to detect excesses of solid tumors. If the A-bomb survivor studies had been terminated after about 15 years, the leukemia effect would have been observed but not the solid tumor effect. Evidence that the exposure precedes the disease is obviously important. In the case of diagnostic radiation and adult leukemia, however, it was likely that many X-rays occurred during the early stages of cancer when symptoms were being evaluated but before any diagnosis was actually made or onset recognized. Findings should be consistent with other studies. Excesses in sites not frequently seen after radiation, such as prostate cancer or Hodgkin disease, in the absence of elevations in leukemia or other sensitive sites, should be interpreted cautiously. There are substantial difficulties in studying directly the carcinogenic effects of very small radiation doses, such as impracticably large study size requirements and the inability to control for other factors that influence cancer incidence (Beebe, 1984; UNSCEAR, 2000). An unfortunate consequence of small study sizes (and in studies of cancer and low radiation dose, one million persons may be a small study size)

is that when a positive result is obtained (and reported), it is likely to be mainly the result of random variation and, if so, must yield an overestimate of risk (Land, 1980). Further, because positive studies appear more likely to be written up and published than negative studies, such reporting bias might operate in studies of low-dose radiation, as seems to be the case for nonionizing radiation (NRPB, 2001) and studies of maternal medical X-ray and Down syndrome (Carter et al., 1961). Another potential problem is the use of complex statistical models when the underlying statistical assumptions do not hold, or when the number of adjustments and assumptions are so enormous that the estimates of risk depend more on the assumed model than on the data themselves. Epidemiological methods may be capable of directly detecting RRs perhaps as low as 1.3–1.4 (i.e., 30%–40% relative excesses). However, the RRs of interest following low doses of radiation (1–10 cGy) are on the order of 1.01 to 1.05. Thus not much should be anticipated from direct observations at 1 cGy (1 rad), and indirect approaches must be used to estimate low-dose effects. Such approaches include studies of populations exposed to a wide range of doses, both low and high, and the development of models to interpolate the effects at high and moderate doses to those at low doses.

FUTURE RESEARCH Epidemiological studies of irradiated human populations should focus on producing refined estimates of cancer risk due to low-level exposure and on understanding the mechanisms of carcinogenesis. Studies of populations receiving a distribution of doses, both low level and high level, should receive highest priority for the following reasons: (1) such populations are rare and it is important that they be thoroughly studied while there is an opportunity to do so, (2) estimates of risks are obtainable with reasonable expenditures of resources, and (3) the possibility for confounding by factors other than radiation is minimized. Studies of populations exposed to high doses are also valuable because many questions about age effects, tissue sensitivity, time

Table 15–3. Factors that Influence Estimates of Risk Associated with Exposure to Radiation Factor Dose dependence Dose rate Radiation quality Sex Age Latency Co-factors Genetic susceptibility Outcome Background rates Tumor type Cellular factors

Comment Cell killing at high doses, repair at low doses. Higher risk for brief exposure, than prolonged exposures. Higher risk for high-LET than low-LET radiations. Somewhat higher risk for women. Somewhat higher risk for people exposed at a young age. Risk varies by time after exposure. Smoking enhances the risk associated with radon and may potentiate the effect of radiotherapy; chemotherapy may interact with radiotherapy. High-dose radiotherapy of susceptible patients may enhance their risk for malignancies, such as bone cancer after retinoblastoma. Cancer incidence may differ appreciably from cancer mortality (e.g., for the thyroid). Radiation risk varies for different cancers in relation to the background rate (on a relative or absolute scale). Cancer sites differ in inducibility, and some cancers have not been convincingly linked to radiation. Radiation damage can be repaired, but some errors occur. The extent of cellular repair at low doses is not known. The relevance of genomic instability and the ‘bystander effect’ is yet to be determined. Apoptosis or selection of damaged cells not to divide may occur following low doses.

Source: IARC, 2000. LET, linear energy transfer.

Ionizing Radiation response, and the interaction of radiation with other factors can be addressed. Populations exposed to fractionated or protracted radiation over long periods should be pursued, especially if cumulative levels of more than 25 cGy (rad) are reached. Separation of ionizing events in time likely simulates spatial separation of events at low doses, and any resulting dose response should be linear and directly applicable to estimation of low-level effects. The opportunities for studies of irradiated populations can be seen in the list of factors that influence radiation risks but for which definitive understanding is still lacking (Table 15–3, modified from IARC, 2000; Upton, 1987; Little, 1981). Current issues or opportunities include clarification of the level of risk from residential radon and new technologies such as CT imaging; pooling data of worker and patient studies; creation of national registries of radiation workers, especially in the utility industry (Zielinski et al., 1997; Goldsmith and Brooks, 1997); clarification of the role genetic susceptibility might play; and identification of any radiation fingerprints or signatures in tumor tissue. The classic studies in radiation epidemiology have or soon will be coming to an end, such as radium dial painters and atomic bomb survivors, but the issues of radiation risks will remain. One overlooked opportunity is the study of patients receiving medical radiation: the numbers exposed are enormous and increase each year, radiation doses can be quantified with great precision, follow-up can be complete, and scatter radiation to tissues that are free of disease can provide information on low-dose risk. Because of the multifactorial nature of carcinogenesis, studies should be encouraged that integrate biochemical and molecular measures of response as well as measures of underlying genetic susceptibility (Kleinerman et al., 1994; Bigbee et al., 1998; Littlefield et al., 1998; Kodama et al., 1991; ICRP, 1998; UNSCEAR, 2000; Bernstein et al., 2004). Such studies are needed to provide insights into the mechanisms of carcinogenesis, and to help in setting guidelines for occupational, medical, and environmental exposures to radiation. Over 100 years of research on the effects of ionizing radiation on human populations has provided a wealth of knowledge that is continually being refined and synthesized into exposure guidelines, compensation schemes, and mechanistic models of cancer development (ICRP, 1991; NIH, 2003; UNSCEAR, 1993, 1994, 2000; NAS, 2005; NRC, 2003, 2005). References Abelson PH. 1991. Mineral dusts and radon in uranium mines. Science 254:777. Agricola C, De Re Metallica. 1950. Basel, 1556. New York: Dover Publications. English reprint (Hoover translation), pp. 214. Ahlbom A, Cardis E, Green A, Linet M, Savitz D, Swerdlow A. ICNIRP (International Commission for Non-Ionizing Radiation Protection) Standing Committee on Epidemiology. 2001. Review of the epidemiologic literature on EMF and Health. Environ Health Perspect 109 Suppl 6:911–933. Alavanja MCR, Brownson RC, Benichou J, et al. 1995. Attributable risk of lung cancer in lifetime nonsmokers and long-term ex-smokers (Missouri, United States). Cancer Causes Control 6:209–216. Alavanja MCR, Brownson RC, Lubin JH, et al. 1994. Residential radon exposure and lung cancer among nonsmoking women. J Natl Cancer Inst 86:1829–1837. Alavanja MC, Lubin JH, Mahaffey JA, Brownson RC. 1999. Residential radon exposure and risk of lung cancer in Missouri. Am J Public Health 89:1042–1048. Allwright SP, Colgan PA, McAulay IR, Mullins E. 1983. Natural background radiation and cancer mortality in the Republic of Ireland. Int J Epidemiol 12:414–418. Amsel J, Waterbor JW, Oler J, et al. 1982. Relationship of site-specific cancer mortality rates to altitude. Carcinogenesis 3:461–465. Andersson M, Cartensen B, Visfeldt J. 1993. Leukemia and other related hematological disorders among Danish patients exposed to Thorotrast. Radiat Res 134:224–233. Andersson M, Storm HH. 1992. Cancer incidence among Danish Thorotrastexposed patients. J Natl Cancer Inst 84:1318–1325. Andersson M. 1997. Long-term effects of internally deposited alpha-particle emitting radionuclides. Epidemiological, pathological and molecular-biological studies of Danish Thorotrast-administered patients and their offspring. Dan Med Bull 44:169–190. Ashmore JP, Krewski D, Zielinski JM, et al. 1998. First analysis of mortality and occupational radiation exposure based on the National Dose Registry of Canada. Am J Epidemiol 148:564–574.

283

Astakhova LN, Anspaugh LR, Beebe GW, et al. 1998. Chernobyl-related thyroid cancer in children of Belarus: A case-control study. Radiat Res 150:349–356. Axelson O, Edling C, Kling H. 1979. Lung cancer and residency—a case-referent study on the possible impact of exposure to radon and its daughters in dwellings. Scand J Work Environ Health 5:10–15. Baker GS, Hoel DG. 2003. Corrections in the atomic bomb data to examine low dose risk. Health Phys 85:709–720. Balonov M, Kaidanovsky G, Zvonova I, et al. 2003. Contributions of shortlived radioiodines to thyroid doses received by evacuees from the Chernobyl area estimated using early in vivo activity measurements. Radiat Prot Dosimetry 105:593–599. Bauer S, Gusev BI, Pivina LM, et al. 2005. Radiation exposure due to local fallout from Soviet atmospheric nuclear weapons testing in Kazakhstan: solid cancer mortality in the Semipalatinsk historical cohort, 1960–1999. Radiat Res 164:409–419. Baverstock KF, Papworth DG. 1989. The UK radium luminiser survey. In: Taylor DM, Mays CW, Gerber GB, Thomas RG, eds. Risks from Radium and Thorotrast. BIR Report 21. London: Br Inst Radiol, pp. 72–76. Beebe GW. 1984. A methodologic assessment of radiation epidemiology studies. Health Phys 46:745–762. Beral V, Fraser P, Carpenter L, et al. 1988. Mortality of employees of the Atomic Weapons Establishment. Br Med J 297:757–770. Berk PD, Goldberg JD, Silverstein MN, et al. 1981. Increased incidence of acute leukemia in polycythemia vera associated with chlorambucil therapy. N Engl J Med 304:441–447. Bernstein JL, Langholz B, Haile RW, et al. 2004. Study design: evaluating gene-environment interactions in the etiology of breast cancer: The WECARE Study. Breast Cancer Res 6:R191–R206. Berrington A, Darby SC, Weiss HA, Doll R. 2001. One hundred years of observation on British radiologists: mortality from cancer and other causes 1897–1997. Br J Radiol 74:507–519. Berrington de Gonzalez A, Ekbom A, Glass AG, et al. 2003. Comparison of documented and recalled histories of exposure to diagnostic X-rays in case-control studies of thyroid cancer. Am J Epidemiol 157:652–663. Berrington de Gonzalez A, Darby S. 2004. Risk of cancer from diagnostic Xrays: estimates for the UK and 14 other countries. Lancet 363:345–351. Bhatia S, Robison LL, Oberlin O, et al. 1996. Breast cancer and other second neoplasms after childhood Hodgkin’s disease. N Engl J Med 334: 745–751. Bigbee WL, Jensen RH, Rahu M, et al. 1996. Glycophorin A biodosimetry of Chernobyl cleanup workers from the Baltic countries. Br Med J 312:1078–1079. Bithell JF, Dutton SJ, Neary NM. 1994. Distribution of childhood leukaemias and non-Hodgkin’s lymphomas near nuclear installations in England and Wales. Br Med J 309:501–505. Bithell JF, Stewart AM. 1975. Pre-natal irradiation and childhood malignancy: A review of the British data from the Oxford Survey. Br J Cancer 31:271–287. Bithell JF, Stiller CA. 1988. A new calculation of the carcinogenic risk of obstetric X-raying. Stat Med 7:857–864. Black D. 1984. Investigation of the possible increased incidence of cancer in West Cumbria. Report of the Independent Advisory Group. London, Her Majesty’s Stationery Office. Blettner M, Zeeb H, Auvinen A, et al. 2003. Mortality from cancer and other causes among male airline cockpit crew in Europe. Int J Cancer 106:946–952. Blot WJ, Xu Z-Y, Boice JD Jr, et al. 1990. Indoor radon and lung cancer in China. J Natl Cancer Inst 82:1025–1030. Boice JD Jr, Bigbee WL, Mumma MT, Blot WJ. 2003a. Cancer incidence in municipalities near two former nuclear materials processing facilities in Pennsylvania. Health Phys 85:678–690. Boice JD Jr, Bigbee WL, Mumma MT, Blot WJ. 2003b. Cancer mortality in counties near two former nuclear materials processing facilities in Pennsylvania. Health Phys 85:691–700. Boice JD Jr, Blettner M, Auvinen A. 2000. Epidemiologic studies of pilots and aircrew. Health Phys 79:576–584. Boice JD Jr, Blettner M, Kleinerman RA, et al. 1987. Radiation dose and leukemia risk in patients treated for cancer of the cervix. J Natl Cancer Inst 79:1295–1311. Boice JD Jr, Blettner M, Kleinerman RA, et al. 1989. Radiation dose and breast cancer risk in patients treated for cancer of the cervix. Int J Cancer 44:7–16. Boice JD Jr, Day NE, Andersen A, et al. 1985a. Second cancers following radiation treatment for cervical cancer: An international collaboration among cancer registries. J Natl Cancer Inst 74:955–975. Boice JD Jr, ed. 1997a. Implications of New Data on Radiation Cancer Risk. Bethesda, MD: National Council on Radiation Protection and Measurements.

284

PART III: THE CAUSES OF CANCER

Boice JD Jr, Engholm G, Kleinerman RA, et al. 1988b. Radiation dose and second cancer risk in patients treated for cancer of the cervix. Radiat Res 116:3–55. Boice JD Jr, Fraumeni JF Jr, eds. 1984. Radiation Carcinogenesis: Epidemiology and Biological Significance. New York: Raven Press. Boice JD Jr, Harvey E, Blettner M, et al. 1992a. Cancer in the contralateral breast after radiotherapy for breast cancer. N Engl J Med 326:781–785. Boice JD Jr, Land CE, Preston D. 1996a. Ionizing radiation. In: Schottenfeld D, Fraumeni JF Jr, eds. Cancer Epidemiology and Prevention. New York: Oxford University Press, pp. 319–354. Boice JD Jr, Land CE, Shore RE, et a1.1979. Risk of breast cancer following low-dose radiation exposure. Radiology 131:589–597. Boice JD Jr, Linet M. 1994. Chernobyl, childhood cancer and chromosome 21 [editorial]. Br Med J 309:139–140. Boice JD Jr, Lubin JH. 1996b. Lung cancer risks: Comparing radiation with tobacco. Radiat Res 146:356–357. Boice JD Jr, Mandel JS, Doody MM, et al. 1992c. A health survey of radiologic technologists. Cancer 69:586–598. Boice JD Jr, Mandel JS, Doody MM. 1995. Breast cancer among radiological technologists. JAMA 274:394–401. Boice JD Jr, McLaughlin JK. 2002. Epidemiologic studies of cellular telephones and cancer risk—a review. SSI rapport: 2002:16. Stockholm: Swedish Radiation Protection Authority, 2002:1–38. Available at: http://www.ssi.se/ssi_rapporter/pdf/ssi_rapp_2002_16.pdf. Accessed January 8, 2004. Boice JD Jr, Miller RW. 1992b. Risk of breast cancer in ataxiatelangiectasia [letter]. N Engl J Med 326:1357–1358. Boice JD Jr, Miller RW. 1999. Childhood and adult cancer after intrauterine exposure to ionizing radiation. Teratology 59:227–233. Boice JD Jr, Monson RR. 1977. Breast cancer in women after repeated fluoroscopic examinations of the chest. J Natl Cancer Inst 59:823–832. Boice JD Jr, Morin MM, Glass AG, et al. 1991b. Diagnostic X-ray procedures and risk of leukemia, lymphoma and multiple myeloma. JAMA 265:1290–1294. Boice JD Jr, Mumma M, Schweitzer S, Blot WJ. 2003c. Cancer mortality in populations living near uranium mining and milling activities in Texas, 1950–2001. J Radiol Prot 23:247–262. Boice JD Jr, Preston D, Davis FG, et al. 1991c. Frequent chest X-ray fluoroscopy and breast cancer incidence among tuberculosis patients in Massachusetts. Radiat Res 125:214–222. Boice JD Jr, Storm HH, Curtis RE, et al, eds. 1985b. Multiple primary cancers in Connecticut and Denmark. Natl Cancer Inst Monogr 68. Washington, DC: U.S. Govt Print Off, pp. 1–437. Boice JD Jr, Tawn EJ, Winther JF, et al. 2003d. Genetic effects of radiotherapy for childhood cancer. Health Phys 85:65–80. Boice JD Jr. 2001c. Second malignancies after chemotherapy. In: Perry MC, ed. The Chemotherapy Source Book. 3rd ed. Lippincott, Williams & Wilkins. Boice JD Jr. 1993a. Second cancer after Hodgkin’s disease—the price of success? J Natl Cancer Inst 85:4–5. Boice JD Jr. 1993b. Leukemia risk in Thorotrast patients [letter]. Radiat Res 136:301–302. Boice JD Jr. 1997b. Radon, your home or mine? [commentary]. Radiat Res 147:135–137. Boice JD Jr. 1997c. Leukaemia, Chernobyl and epidemiology [editorial]. J Radiol Prot 17:129–133. Boice JD Jr. 2001a. Radiation and breast carcinogenesis. Med Pediatr Oncol 36:508–513. Boice JD Jr. 2001b. Study of health effects of low-level radiation in USA nuclear shipyard workers. J Radiol Prot 21:400–403. Boice JD Jr. 2002. Study of health effects in areas of high background radiation in China. J Radiol Prot 22:102–104. Boice JD Jr. 2005. Radiation-induced thyroid cancer—what’s new? J Natl Cancer Inst 97:703–705. Boice JD Jr., Holm LE. 1997. Radiation estimates for leukaemia and thyroid cancer among Russian emergency workers at Chernobyl [letter]. Radiat Environ Biophys 36:213–214. Boice JD, Hutchison GB. 1980. Leukemia in women following radiotherapy for cervical cancer: Ten-year follow-up of an international study. J Natl Cancer Inst 65:115–129. Boice JD, Land CE. 1979. Adult leukemia following diagnostic X-rays? Am J Public Health 69:137–145. Boivin J-F, Hutchison GB, Evans FB, et al. 1986. Leukemia after radiotherapy for first primary cancers of various anatomic sites. Am J Epidemiol 123:993–1003. Boutou O, Guizard AV, Slama R, et al. 2002. Population mixing and leukaemia in young people around the La Hague nuclear waste reprocessing plant. Br J Cancer 87:740–745.

Breckenkamp J, Berg G, Blettner M. 2003. Biological effects on human health due to radiofrequency/microwave exposure: A synopsis of cohort studies. Radiat Environ Biophys 42:141–154. Brenner DJ, Doll R, Goodhead DT, et al. 2003. Cancer risks attributable to low doses of ionizing radiation: assessing what we really know. Proc Natl Acad Sci USA 100:13761–13766. Brenner DJ. 1994. The significance of dose rate in assessing the hazards of domestic radon exposure. Radiat Res 67:76–79. Breslow NE, Norkool PA, Olshan A, et al. 1988. Second malignant neoplasms in survivors of Wilms’ tumor: A report from the National Wilms’ Tumor Study. J Natl Cancer Inst 80:592–595. Brill AB, Tomonaga M, Heyssel RM. 1962. Leukemia in man following exposure to ionizing radiation: A summary of the findings in Hiroshima and Nagasaki, and a comparison with other human experience. Ann Intern Med 56:590–609. Broeks A, Russell NS, Floore AN, et al. 2000. Increased risk of breast cancer following irradiation for Hodgkin’s disease is not a result of ATM germline mutations. Int J Radiat Biol 76:693–698. Brooks AL. 2005. Paradigm shifts in radiation biology: their impact on intervention for radiation-induced disease. Radiat Res 164:454–461. Buckley JD, Robison LL, Swotinsky R, et al. 1989. Occupational exposures of parents of children with acute nonlymphocytic leukemia: A report from the Children Cancer Study Group. Cancer Res 49:4030–4037. Caldwell GG, Kelley D, Zack M, et al. 1983. Mortality and cancer frequency among military nuclear test (SMOKY) participants, 1957 through 1979. JAMA 250:620–624. Cardis E, Anspaugh L, Ivanov V K, et al. 1996. Estimated long term health effects of the Chemobyl accident. In: Proc Int Conf on One Decade After Chernobyl: Summing Up the Consequences of the Accident STVPUB/ 1001. Vienna: IAEA, pp. 241–279. Cardis E, Esteve J. 1992. International collaborative study of cancer risk among nuclear industry workers. I. Report of the feasibility study. Lyon, France: International Agency for Research on Cancer. (Internal report no. 92/001). Cardis E, Gilbert ES, Carpenter L, et al. 1995. Effects of low doses and low dose rates of external ionizing radiation: Cancer mortality among nuclear industry workers in three countries. Radiat Res 142:117–132. Cardis E, Vrijheid M, Blettner M, et al. 2005a. Risk of cancer after low doses of ionising radiation: retrospective cohort study in 15 countries. BMJ 331:77. Cardis E, Kesiniene A, Ivanov V, et al. 2005b. Risk of thyroid cancer after exposure to 131I in childhood. J Natl Cancer Inst 97:727–732. Carnes BA, Groer PG, Kotek TJ. 1997. Radium dial workers: Issues concerning dose response and modeling. Radiat Res 147:707–714. Carpenter L, Higgins C, Douglas A, et al. 1994. Combined analysis of mortality in three United Kingdom nuclear industry workforces, 1946–1988. Radiat Res 138:224–238. Carr ZA, Kleinerman RA, Stovall M, et al. 2002. Malignant neoplasms after radiation therapy for peptic ulcer. Radiat Res 157:668–677. Carter CO, Evans KA, Stewart AM. 1961. Maternal radiation and Down’s syndrome (Mongolism). Lancet ii:1042. Cartwright RA, Dovey GJ, Kane EV, Gilman EA. 2001. The onset of the excess of childhood cancer in Seascale, Cumbria. J Public Health Med 23:314–322. Checkoway H, Mathew RM, Shy CM, et al. 1985. Radiation, work experience, and cause specific mortality among workers at an energy research laboratory. Br J Ind Med 42:525–533. Chomentowski M, Kellerer AM, Pierce DA. 2000. Radiation dose dependences in the atomic bomb survivor cancer mortality data: A model-free visualization. Radiat Res 153:289–294. Claycamp HG, Okladnikova ND, Azizova TV, et al. 2000. Deterministic effects from occupational radiation exposures in a cohort of Mayak PA workers: Data base description. Health Phys 79:48–54. Cohen BL. 1995. Test of the linear-no threshold theory of radiation carcinogenesis for inhaled radon decay products. Health Phys 68:157–174. Cologne JB, Preston DL. 2000. Longevity of atomic-bomb survivors. Lancet 356:303–307. Cologne JB, Tokuoka S, Beebe GW, Fukuhara T, Mabuchi K. 1999. Effects of radiation on incidence of primary liver cancer among atomic bomb survivors. Radiat Res 152:364–373. Conard RA. 1984. Late radiation effects in Marshall Islanders exposed to fallout twenty-eight years ago. In: Boice JD Jr, Fraumeni JF Jr, eds. Radiation Carcinogenesis: Epidemiology and Biological Significance. New York: Raven Press, pp. 57–71. Cook-Mozaffari PJ, Ashwood FL, Vincent T, et al. 1987. Cancer incidence and mortality in the vicinity of nuclear installations, England and Wales, 1959–1980. OPCS Studies on Medical and Population Subjects 51. London: Her Majesty’s Stationery Office.

Ionizing Radiation Cook-Mozaffari P, Darby S, Doll R. 1989. Cancer near potential sites of nuclear installations. Lancet 2:1145–1147. Court Brown WM, Doll R, Hill AB. 1960. Incidence of leukemia after exposure to diagnostic radiation in utero. Br Med J 2:1539–1545. Cox R, Kellerer AM. 2003. A current view on radiation weighting factors and effective dose. (ICRP Publ 92). Ann ICRP 33:1–4. Crile G Jr, Esselstyn CB Jr, Hawk WA. 1979. Needle biopsy in the diagnosis of thyroid nodules appearing after radiation. N Engl J Med 301:997–999. CRS (Council of the Royal Society). 2001. The Health Hazards of Depleted Uranium Munitions. Part 1. Science Advice Section. London: The Royal Society. Curtis RE, Boice JD Jr, Kleinerman RA, et al. 1985. Summary: Multiple primary cancers in Connecticut, 1935–82. Natl Cancer Inst Monogr 68:219–242. Curtis RE, Boice JD Jr, Stovall M, et al. 1992. Risk of leukemia after chemotherapy and radiation treatment for breast cancer. N Engl J Med 326:1745–1751. Curtis RE, Boice JD Jr, Stovall M, et al. 1994. Relation of leukemia risk to radiation dose after cancer of the uterine corpus. J Natl Cancer Inst 86:1315–1324. Curtis RE, Rowlings PA, Deeg HJ, et al. 1997. Solid cancers after bone marrow transplantation. N Engl J Med 336:897–904. Cuzick J, DeStavola B. 1988. Multiple myeloma—a case-control study. Br J Cancer 57:516–520. Damber L, Johansson L, Johansson R, Larsson LG. 2002. Thyroid cancer after X-ray treatment of benign disorders of the cervical spine in adults. Acta Oncol 41:25–28. Darby SC, Doll R, Gill SK, et al. 1987. Long term mortality after a single treatment course with X-rays in patients treated for ankylosing spondylitis. Br J Cancer 55:179–190. Darby SC, Doll R. 1987. Fallout, radiation doses near Dounreay, and childhood leukaemia. Br Med J 294:603–607. Darby SC, Kendall GM, Fell TP, et al. 1993. Further follow-up of mortality and incidence of cancer in man from the United Kingdom who participated in the United Kingdom’s atmospheric nuclear weapon tests and experimental programmes. Br Med J 307:1530–1535. Darby SC, Olsen JH, Doll R, et al. 1992. Trends in childhood leukaemia in the Nordic countries in relation to fallout from atmospheric nuclear weapons testing. Br Med J 304:1005–1009. Darby SC, Reeves G, Key T, et al. 1994. Mortality in a cohort of women given X-ray therapy for metropathia haemorrhagica. Int J Cancer 56:793–801. Darby S, Whitley E, Silcocks P, et al. 1998. Risk of lung cancer associated with residential radon exposure in south-west England: A case-control study. Br J Cancer 78:394–408. Darby SC, Whitley E, Deo H, Doll R. 2001. A parallel analysis of individual and ecological data on residential radon and lung cancer in south-west England. J R Stat Soc Series A 164:193–203. Darby SC, Whitley E, Howe GR, et al. 1995. Radon and cancers other than lung cancer in underground miners: A collaborative analysis of 11 studies. J Natl Cancer Inst 87:378–384. Darby S, Hill D, Auvinen A, et al. 2005. Radon in homes and risk of lung cancer: collaborative analysis of individual data from 13 European casecontrol studies. BMJ 330:223. Davis FG, Boice JD Jr, Hrubec Z, et al. 1989. Cancer mortality in a radiationexposed cohort of Massachusetts tuberculosis patients. Cancer Res 49:6130–6136. Davis S, Kopecky KJ, Hamilton TE, et al. 2004. Hanford Thyroid Disease Study Team. Thyroid neoplasia, autoimmune thyroiditis, and hypothyroidism in persons exposed to iodine 131 from the Hanford nuclear site. JAMA 292:2600–2613. Davis S, Kopecky KJ, Hamilton TE. Hanford Thyroid Disease Study. Final Report. 2002. Fred Hutchinson Cancer Research Center, Seattle, WA. (CDC Contract No. 200–89–0716), June 21, 2002. Available at: http://www.cdc.gov/nceh/radiation/hanford/htdsweb/pdf/htdsreport.pdf. Accessed January 8, 2004. De Vathaire F, François P, Hill C, et al. 1989. Role of radiotherapy and chemotherapy in the risk of second malignant neoplasms after cancer in childhood. Br J Cancer 59:792–796. De Vathaire F, François P, Schweisguth O, et al. 1988. Irradiated neuroblastoma in childhood as potential risk factor for subsequent thyroid tumor. Lancet 2:455. De Vathaire F, Hardiman C, Shamsaldin A, et al. 1999. Thyroid carcinomas after irradiation for a first cancer during childhood. Arch Intern Med 159:2713–2719. Delongchamp RR, Mabuchi K, Yoshimoto Y, Preston DL. 1997. Cancer mortality among atomic bomb survivors exposed in utero or as young children, October 1950–May 1992. Radiat Res 147:385–395. Diamond EL, Schmerler H, Lilienfeld AM. 1973. The relationship of intrauterine radiation to subsequent mortality and development of leukemia in children. Am J Epidemiol 97:283–313.

285

Dickinson HO, Parker L. 2002. Leukaemia and non-Hodgkin’s lymphoma in children of male Sellafield radiation workers. Int J Cancer 99:437–444. Dickman PW, Holm L-E, Lundell G, et al. 2003. Thyroid cancer risk after thyroid examination with 131I: A population-based cohort study in Sweden. Int J Cancer 106:580–587. Doll R, Evans HJ, Darby SC. 1994. Paternal exposure not to blame. Nature 367:678–680. Doll R, Wakeford R. 1997. Risk of childhood cancer from fetal irradiation. Br J Radiol 70:130–139. Doll R. 1999. The Seascale cluster: A probable explanation. Br J Cancer 81:3–5. Donaldson SS, Hancock SL, Hoppe RT. 1999. The Janeway lecture. Hodgkin’s disease-finding the balance between cure and late effects. Cancer J Sci Am 5:325–333. Doody MM, Lonstein JE, Stovall M, et al. 2000. Breast cancer mortality after diagnostic radiography: Findings from the U.S. Scoliosis Cohort Study. Spine 25:2052–2063. Doody MM, Mandel JS, Lubin JH, Boice JD Jr. 1998. Mortality among United States radiologic technologists, 1926–1990. Cancer Causes Control 9:67–75. dos Santos Silva I, Malveiro F, Jones ME, Swerdlow AJ. 2003. Mortality after radiological investigation with radioactive Thorotrast: A follow-up study of up to fifty years in Portugal. Radiat Res 159:521–534. dos Santos Silva I, Swerdlow AJ. 1993. Thyroid cancer epidemiology in England and Wales: Time trends and geographical distribution. Br J Cancer 67:330–340. Douglas AJ, Omar RZ, Smith PG. 1994. Cancer mortality and morbidity among workers at the Sellafield plant of British Nuclear Fuels. Br J Cancer 70:1232–1243. Draper GJ, Little MP, Sorahan T, et al. 1997. Cancer in the offspring of radiation workers: A record linkage study. Br Med J 315:1181–1188. Draper GJ, Sanders BM, Kingston JE. 1986. Second primary neoplasms in patients with retinoblastoma. Br J Cancer 53:661–671. Draper GJ, Stiller CA, Cartwright RA, et al. 1993. Cancer in Cumbria and in the vicinity of the Sellafield nuclear installation, 1963–90. Br Med J 306:89–94. Dupree EA, Watkins JP, Ingle JN, et al. 1995. Uranium dust exposure and lung cancer risk in four uranium processing operations. Epidemiology 6:370–375. Edling C, Comba P, Axelson O, Flodin U. 1982. Effects of low-dose radiation—a correlation study. Scand J Work Environ Health 8 Suppl 1:59–64. Edmonds CJ, Smith T. 1986. The long-term hazards of the treatment of thyroid cancer with radioiodine. Br J Radiol 59:45–51. Elkind MM, Benjamin SA, Sinclair WK, et al. 1991. Oncogenic mechanisms in radiation-induced cancer. Cancer Res 51:2740–2747. Eng C, Li FP, Abramson DH, et al. 1993. Mortality from second tumors among long-term survivors of retinoblastoma. J Natl Cancer Inst 85:1121–1128. Evans HJ. 1990. Gardner report. Leukaemia and radiation. Nature 345:16–17. Evans JS, Wennberg JE, McNeil BJ. 1986. The influence of diagnostic radiography on the incidence of breast cancer and leukemia. N Engl J Med 315:810–815. Ezaki H, Ishimaru T, Hayashi Y, et al. 1986. Cancer of the thyroid and salivary glands. In: Shigematsu I, Kagan A, eds. Cancer in Atomic Bomb Survivors. GANN Monograph on Cancer Research No. 32. Tokyo: Plemun Press, pp. 129–142. Favus MJ, Schneider AB, Stachura ME, et al. 1976. Thyroid cancer occurring as a late consequence of head-and-neck irradiation: Evaluation of 1056 patients. N Engl J Med 294:1019–1025. Forman D, Cook-Mozaffari P, Darby S, et al. 1987. Cancer near nuclear installations. Nature 329:499–505. Franklyn JA, Maisonneuve P, Sheppard M, et al. 1999. Cancer incidence and mortality after radioiodine treatment for hyperthyroidism: A populationbased cohort study. Lancet 353:2111–2115. Frome EL, Cragle DL, McLain RW. 1990. Poisson regression analysis of the mortality among a cohort of World War II nuclear industry workers. Radiat Res 123:138–152. Fry SA. 1998. Studies of U.S. radium dial workers: An epidemiological classic. Radiat Res 150(5 Suppl):S21–S29. Fry SA, Dupree EA, Sipe AH, et al. 1996. A study of mortality and morbidity among persons occupationally exposed to > 50 mSv in a year: Phase I, mortality through 1984. Appl Occup Environ Hyg 11:334–343. Fürst CJ, Lundell M, Holm L-E, et al. 1988. Cancer incidence after radiotherapy for skin hemangioma: A retrospective cohort study in Sweden. J Natl Cancer Inst 80:1387–1392. Gardner MJ, Hall AJ, Downes S, et al. 1987. Follow-up study of children born to mothers resident in Seascale, West Cumbria (birth cohort). Br Med J 295:822–827. Gardner MJ, Snee MP, Hall AJ, et al. 1990. Results of case-control study of leukaemia and lymphoma among young people near Sellafield nuclear plant in West Cumbria. Br Med J 300:423–429.

286

PART III: THE CAUSES OF CANCER

Garwicz S, Anderson H, Olsen JH, et al. 2000. Second malignant neoplasms after cancer in childhood and adolescence: A population-based casecontrol study in the 5 Nordic countries. The Nordic Society for Pediatric Hematology and Oncology. The Association of the Nordic Cancer Registries. Int J Cancer 88:672–678. Gibson R, Graham S, Lilienfeld A, et al. 1972. Irradiation in the epidemiology of leukemia among adults. J Natl Cancer Inst 48:301–311. Gilbert ES, Cragle DL, Wiggs LD, et al. 1993a. Updated analyses of combined mortality data for workers at the Hanford Site, Oak Ridge National Laboratory, and Rocky Flats Weapons Plant. Radiat Res 136:408–421. Gilbert ES, Fix JJ. 1995. Accounting for bias in dose estimates in analyses of data from nuclear worker mortality studies. Health Phys 68:650–660. Gilbert ES, Koshurnikova NA, Sokolnikov M, et al. 2000. Liver cancers in Mayak workers. Radiat Res 154:246–252. Gilbert ES, Koshurnikova NA, Sokolnikov ME, et al. 2004. Lung cancer in Mayak workers. Radiat Res 162:505–516. Gilbert ES, Land CE, Simon SL. 2002. Health effects from fallout. Health Phys 82:726–735. Gilbert ES, Omohundro JA, Buchanan JA, et al. 1993b. Mortality of workers at the Hanford site: 1945–1986. Health Phys 64:577–590. Gilbert ES, Petersen GR, Buchanan JA. 1989. Mortality of workers at the Hanford site: 1945–1981. Health Phys 56:11–25. Gilbert ES, Stovall M, Gospodarowicz M, et al. 2003. Lung cancer after treatment for Hodgkin’s disease: Focus on radiation effects. Radiat Res 159:161–173. Gilbert ES, Tarone R, Bouville A, Ron E. 1998. Thyroid cancer rates and 131I doses from Nevada atmospheric nuclear bomb tests. J Natl Cancer Inst 90:1654–1660. Gilbert ES. 1992. Mortality of workers at the Oak Ridge National Laboratory. Health Phys 62:260–261. Gilbert ES. 1994. Smoking as an explanation for the negative relationship between exposure to radon and certain types of cancer. Health Phys 67:197–198. Gilbert ES. 2001. Invited commentary: Studies of workers exposed to low doses of radiation. Am J Epidemiol 153:319–322; discussion 323–324. Glöbel B, Glöbel H, Oberhausen E. 1984. Epidemiologic studies on patients with iodine-131 diagnostic and therapy. In: Kaul A, Neider R, Pensko J, et al., eds. Radiation Risk-Protection. vol II. IRPA. Köln: Fachverband fur Strahlenschutz eV., pp. 565–568. Goldman MB, Maloof F, Monson RR, et al. 1988. Radioactive iodine therapy and breast cancer. A follow-up study of hyperthyroid women. Am J Epidemiol 127:969–980. Goldsmith R, Brooks BG. 1997. Toward a U.S. national radiation dose registry . . . and more. Eur J Cancer 33(Suppl 3):S22–S24. Graham S, Levin ML, Lilienfeld AM, et al. 1966. Preconception, intrauterine and postnatal irradiation as related to leukemia. Natl Cancer Inst Monogr 19:347–371. Greenberg ER, Rosner B, Hennekens C, et al. 1985. An investigation of bias in a study of nuclear shipyard workers. Am J Epidemiol 121:301–308. Greene MH, Young RC, Merrill JM, et al. 1983. Evidence of a treatment dose response in acute nonlymphocytic leukemias which occur after therapy of non-Hodgkin’s lymphoma. Cancer Res 43:1891–1898. Gribbin MA, Weeks JL, Howe GR. 1993. Cancer mortality (1956–1985) among male employees of Atomic Energy Canada Limited with respect to occupational exposure to external low-linear-energy-transfer ionizing radiation. Radiat Res 133:375–380. Griem ML, Kleinerman RA, Boice JD Jr, et al. 1994. Cancer following radiotherapy for peptic ulcer. J Natl Cancer Inst 86:842–849. Griem ML, Meier P, Dobben GD. 1967. Analysis of the morbidity and mortality of children irradiated in fetal life. Radiology 88:347–349. Grosche B, Lackland D, Mohr L, et al. 1999. Leukaemia in the vicinity of two tritium-releasing nuclear facilities: a comparison of the Kruemmel Site, Germany, and the Savannah River Site, South Carolina, USA. J Radiol Prot 19:243–252. Grosche B, Land C, Bauer S, et al. 2002. Fallout from nuclear tests: Health effects in Kazakhstan. Radiat Environ Biophys 41:75–80. Guilmette RA, Mays DM, eds. 1992. Total body evaluation of a Thorotrast patient. Health Phys 63:1–100. Guizard AV, Boutou O, Pottier D, et al. 2001. The incidence of childhood leukaemia around the La Hague nuclear waste reprocessing plant (France): A survey for the years 1978–1998. J Epidemiol Community Health 55:469–474. Hahn K, Schnell-Inderst P, Grosche B, Holm LE. 2001. Thyroid cancer after diagnostic administration of iodine-131 in childhood. Radiat Res 156:61–70. Hall P, Berg G, Bjelkengren G, et al. 1992a. Cancer mortality after iodine-131 therapy for hyperthyroidism. Int J Cancer 50:886–890. Hall P, Boice JD Jr, Berg G, et al. 1992b. Leukaemia incidence after iodine131 exposure. Lancet 340:1–4.

Hall P, Holm L-E, Lundell G, et al. 1991. Cancer risks in thyroid cancer patients. Br J Cancer 64:159–163. Hall P, Mattson A, Boice JD Jr, et al. 1996. Thyroid cancer following diagnostic iodine-131 exposure. Radiat Res 145:86–92. Hamilton TE, van Belle G, LoGerfo JP. 1987. Thyroid neoplasia in Marshall Islanders exposed to nuclear fallout. JAMA 258:629–635. Hamilton PM, Chiacchierini RP, Kaczmarek RG. 1989. A Follow-Up Study of Persons who had Iodine-131 and Other Diagnostic Procedures during Childhood. Report No. FDA 89-8276. Rockville, MD: U.S. Food and Drug Administration. Hancock SL, Tucker MA, Hoppe RT. 1993. Breast cancer after treatment of Hodgkin’s disease. J Natl Cancer Inst 85:25–31. Harley NH, Foulkes EC, Hiborne LH, et al. 1999. Depleted Uranium: A Review of the Scientific Literature as It Pertains to Gulf War Illness. Santa Monica: RAND. Harvard Report on Cancer Prevention. 1996. Volume I: Causes of Human Cancer. Cancer Causes Control 7 (Suppl). Harvey EB, Boice JD Jr, Honeyman M, et al. 1985. Prenatal X-ray exposure and childhood cancer in twins. N Engl J Med 312:541–545. Hatch MC, Beyea J, Nieves JW, et al. 1990. Cancer near the Three Mile Island nuclear plant: radiation emissions. Am J Epidemiol 132:397–412. Hatch M, Susser M, Beyea J. 1997. Comments on “A reevaluation of cancer incidence near the Three Mile Island nuclear plant”. Environ Health Perspect 105:12. Hatcher JL, Baris D, Olshan AF, et al. 2001. Diagnostic radiation and the risk of multiple myeloma (United States). Cancer Causes Control 12:755–761. Hattchouel JM, Laplanche A, Hill C. 1995. Leukaemia mortality around French nuclear sites. Br J Cancer 71:651–653. Hattchouel JM, Laplanche A, Hill C. 1996. Cancer mortality around French nuclear sites. Ann Epidemiol 6:126–129. Hawkins MM, Draper GJ, Kingston JE: 1987. Incidence of second primary tumours among childhood cancer survivors. Br J Cancer 56:339–347. Hawkins MM, Kinnier-Wilson LM, Burton HS, et al. 1996. Radiotherapy, alkylating agents, and risk of bone cancer after childhood cancer. J Natl Cancer Inst 88:270–278. Hawkins MM, Kinnier-Wilson LM, Stovall MA, et al. 1992. Epipodophyllotoxins, alkylating agents, and radiation and risk of secondary leukaemia after childhood cancer. Br Med J 304:951–958. Hazen RW, Pifer JW, Tayooka ET, et al. 1966. Neoplasms following irradiation of the head. Cancer Res 26:305–311. Heidenreich WF, Paretzke HG, Jacob P. 1997. No evidence for increased tumor rates below 200 mSv in the atomic bomb survivors data. Radiat Environ Biophys 36:205–207. Hempelmann LH, Hall WJ, Phillips M, et al. 1975. Neoplasms in persons treated with X-rays in infancy: Fourth survey in 20 years. J Natl Cancer Inst 55:519–530. Hildreth NG, Shore RE, Dvoretsky PM. 1989. The risk of breast cancer after irradiation of the thymus in infancy. N Engl J Med 321:1281–1284. Hildreth NG, Shore RE, Hempelmann LH, et al. 1985. Risk of extrathyroid tumors following radiation treatment in infancy for thymic enlargement. Radiat Res 102:378–391. Hill C, LaPlanche A. 1990. Overall mortality and cancer mortality around French nuclear sites. Nature 347:755–757. Hisada M, Garber JE, Fung CY, et al. 1998. Multiple primary cancers in families with Li-Fraumeni syndrome. J Natl Cancer Inst 90:606–611. Hoel DG, Li P. 1998. Threshold models in radiation carcinogenesis. Health Phys 75:241–250. Hoffman D. 1984. Effects of I-131 therapy in the United States. In: Boice JD Jr, Fraumeni JF Jr, eds. Radiation Carcinogenesis: Epidemiology and Biological Significance. New York: Raven Press, pp. 273–280. Hoffman DA, Lonstein JE, Morin MM, et al. 1989. Breast cancer in women with scoliosis exposed to multiple diagnostic X-rays. J Natl Cancer Inst 81:1307–1312. Hofmann W, Hornik S. 1999. Lung dosimetry for thorotrast patients: implications for inhalation of radon progeny. Radiat Res 152(6 Suppl):S93–S96. Holm L-E, Hall P, Wiklund KE, et al. 1991. Cancer risk after iodine-131 therapy for hyperthyroidism. J Natl Cancer Inst 83:1072–1077. Holm L-E, Wiklund KE, Lundell GE, et al. 1988. Thyroid cancer after diagnostic doses of iodine-131: A retrospective cohort study. J Natl Cancer Inst 80:1132–1138. Holm L-E, Wiklund KE, Lundell GE, et al. 1989. Cancer risk in population examined with diagnostic doses of 131-I. J Natl Cancer Inst 81:302–306. Holmberg E, Holm LE, Lundell M, et al. 2001. Excess breast cancer risk and the role of parity, age at first childbirth and exposure to radiation in infancy. Br J Cancer 85:362–366. Hornung RW, Deddens JA, Roscoe RJ. 1998. Modifiers of lung cancer risk in uranium miners from the Colorado Plateau. Health Phys 74:12–21. Howe GR. 1995. Lung cancer mortality between 1950 and 1987 after exposure to fractionated moderate-dose-rate ionizing radiation in the Canadian

Ionizing Radiation fluoroscopy chart study and a comparison with lung cancer mortality in the atomic bomb survivor study. Radiat Res 142:295–304. Howe GR, Burch JD, Chiarelli AM, et al. 1989. An exploratory case-control study of brain tumors in children. Cancer Res 49:4349–4352. Howe GR, Chiarelli AM, Lindsay JP. 1988. Components and modifiers of the healthy worker effect: evidence from three occupational cohorts and implications for industrial compensation. Am J Epidemiol 128:1364–1375. Howe GR, McLaughlin J. 1996. Breast cancer mortality between 1950 and 1987 after exposure to fractionated moderate-dose-rate ionizing radiation in the Canadian fluoroscopy cohort study and a comparison with breast cancer mortality in the atomic bomb survivors study. Radiat Res 145:694–707. Howe GR, Zablotska LB, Fix JJ, et al. 2004. Analysis of the mortality experience amongst U.S. nuclear power industry workers after chronic low-dose exposure to ionizing radiation. Radiat Res 162:517–526. Hutchison GB, MacMahon B, Jablon S, et al. 1979. Review of report by Mancuso, Stewart and Kneale of radiation exposure of Hanford workers. Health Phys 37:207–220. IARC. International Agency for Research on Cancer. 2000. IARC monographs on the evaluation of carcinogenic risks to humans. Vol 75. Ionizing Radiation, Part 1: X- and Gamma (g)-Radiation, and Neutrons. Lyon, France: IARC. IARC. International Agency for Research on Cancer. 2001. IARC monographs on the evaluation of carcinogenic risks to humans. Vol 78. Ionizing Radiation, Part 2: Some Internally Deposited Radionuclides. Lyon, France: IARC. Ichimaru M, Ohkita T, Ishimaru T. 1986. Leukemia, multiple myeloma, and malignant lymphoma. In: Shigematsu I, Kagan A, eds. Cancer in Atomic Bomb Survivors. GANN Monograph on Cancer Research No. 32. Tokyo: Plemun Press, pp. 113–128. ICRP (International Commission on Radiological Protection). 1991. Publication 60. 1990 Recommendations of the International Commission on Radiological Protection. Ann ICRP 21:1–201. ICRP (International Commission on Radiological Protection). 1991. 1990 Recommendation of the International Commission on Radiological Protection. Oxford: Pergamon Press. ICRP. 1998. Publication 79. Genetic susceptibility to cancer. Ann ICRP 28:1–157. ICRP. 2003a. Publication 92. Relative biological effectiveness (RBE), quality factor (Q), and radiation weighting factor (w(R)). Ann ICRP 33:1–117. ICRP. 2003b. Publication 90. Biological effects after prenatal irradiation (embryo and fetus). Ann ICRP 33:1–206. Infante-Rivard C, Mathonnet G, Sinnett D. 2000. Risk of childhood leukemia associated with diagnostic irradiation and polymorphisms in DNA repair genes. Environ Health Perspect 108:495–498. Inskip H. 1997. Childhood leukaemia near nuclear sites re-examined. Lancet 349:969–970. Inskip PD, Boice JD Jr. 1994. Radiotherapy-induced lung cancer among women who smoke [editorial]. Cancer 73:1541–1543. Inskip PD, Eby NL, Cookfair D, et al. 1994a. Serum estrogen and androgen levels following treatment for cervical cancer. Cancer Epidemiol Biomarkers Prev 3:37–45. Inskip PD, Ekbom A, Galanti MR, et al. 1995. Medical diagnostic X-rays and thyroid cancer. J Natl Cancer Inst 87:1613–1621. Inskip PD, Hartshorne MF, Tekkel M, et al. 1997. Thyroid nodularity and cancer among Chernobyl cleanup workers from Estonia. Radiat Res 147:225–235. Inskip PD, Harvey EB, Boice JD Jr, et al. 1991. Incidence of childhood cancer in twins. Cancer Causes Control 2:315–324. Inskip PD, Kleinerman RA, Stovall, et al. 1993. Leukemia, lymphoma, and multiple myeloma following pelvic radiotherapy for benign disease. Radiat Res 135:108–124. Inskip PD, Monson RR, Wagoner JK, et al. 1990. Cancer mortality following radium treatment for uterine bleeding. Radiat Res 123:331–344. Inskip PD, Stovall M, Flannery JT. 1994b. Lung cancer risk and radiation dose among women treated for breast cancer. J Natl Cancer Inst 86:983–988. Inskip PD. 2001. Thyroid cancer after radiotherapy for childhood cancer. Med Pediatr Oncol 36:568–573. Inskip PD. 2003. Multiple primary tumors involving cancer of the brain and central nervous system as the first or subsequent cancer. Cancer 98:562–570. IOM (Institute of Medicine). 2001. Committee on the Health Effects Associated with Exposures During the Gulf War. Gulf War and Health. Volume 1. Depleted Uranium, Pyridostigmine Bromide, Sarin, Vaccines. Washington, DC: National Academy Press. IOM (Institute of Medicine). 2000. The Five Series Study: Mortality of Military Participants in US Nuclear Weapons Tests. Washington, DC: National Academy Press. Ishimaru T, Otake M, Ichimaru M. 1979. Dose-response relationship of neutrons and gamma rays to leukemia incidence among atomic bomb survivors in Hiroshima and Nagasaki by type of leukemia, 1950–1971. Radiat Res 77:377–394.

287

Ivanov EP, Tolochko G, Lazarev VS, et al. 1993. Childhood leukaemia after Chernobyl. Nature 365:702. Ivanov EP, Tolochko GV, Shuvaeva LP, et al. 1996. Childhood leukemia in Belarus before and after the Chernobyl accident. Radiat Environ Biophys 35:75–80. Ivanov EP, Tolochko GV, Shuvaeva LP, et al. 1998. Infant leukemia in Belarus after the Chernobyl accident. Radiat Environ Biophys 37:53–55. Ivanov VK, Tsyb AF, Gorsky AI, et al. 1997a. Leukaemia and thyroid cancer in emergency workers of the Chernobyl accident: Estimation of radiation risks (1986–1995). Radiat Environ Biophys 36:9–16. Ivanov VK, Tsyb AF, Konogorov AP, et al. 1997b. Case-control analysis of leukaemia among Chernobyl accident emergency workers residing in the Russian Federation, 1986–1993. J Radiol Prot 17:137–157. Ivanov VK, Tsyb AF, Petrov AV, et al. 2002. Thyroid cancer incidence among liquidators of the Chernobyl accident. Absence of dependence of radiation risks on external radiation dose. Radiat Environ Biophys 41:195– 198. Iwasaki T, Murata M, Ohshima S, et al. 2003. Second analysis of mortality of nuclear industry workers in Japan, 1986–1997. Radiat Res 159:228–238. Izumi S, Koyama K, Soda M, Suyama A. 2003. Cancer incidence in children and young adults did not increase relative to parental exposure to atomic bombs. Br J Cancer 89:1709–1713. Jablon S, Bailar JC III. 1980. The contribution of ionizing radiation to cancer mortality in the United States. Prev Med 9:219–226. Jablon S, Boice JD Jr. 1993. Mortality among workers at a nuclear power plant in the United States. Cancer Causes Control 4:427–430. Jablon S, Hrubec Z, Boice JD Jr. 1991. Cancer in populations living near nuclear facilities: A survey of mortality nationwide and incidence in two states. JAMA 265:1403–1408. Jablon S, Kato H. 1970. Childhood cancer in relation to prenatal exposure to atomic-bomb radiation. Lancet 2:1000–1003. Jablon S, Miller RW. 1978. Army technologists: 29-year followup for cause of death. Radiology 126:677–679. Jacob P, Goulko G, Heidenreich WF, et al. 1998. Thyroid cancer risk to children calculated. Nature 392:31–32. Jacob P, Kenigsberg Y, Zvonova I, et al. 1999. Childhood exposure due to the Chernobyl accident and thyroid cancer risk in contaminated areas of Belarus and Russia. Br J Cancer 80:1461–1469. Jessup GL, Silverman C. 1981. Personal usage of medical radiological procedures by radiologists, pathologists, and their families. Am J Epidemiol 114:53–62. Johnson CJ. 1984. Cancer incidence in an area of radioactive fallout downwind from the Nevada Test Site. JAMA 251:230–236. Kaiser J. 2003. Hormesis. A healthful dab of radiation? Science 302:378. Kaldor JM, Day NE, Clarke EA, et al. 1990. Leukemia following Hodgkin’s disease. N Engl J Med 322:7–13. Kazakov VS, Demidchik EP, Astakhova LN. 1992. Thyroid cancer after Chernobyl. Nature 359:21. Kellerer AM. 2002a. The Southern Urals radiation studies. A reappraisal of the current status. Radiat Environ Biophys 41:307–316. Kellerer AM. 2002b. Radiation risk-historical perspective and current issues. J Radiol Prot 22:A1–A10. Kellerer AM, Nekolla E. 1997. Neutron versus gamma-ray risk estimates. Inferences from the cancer incidence and mortality data in Hiroshima. Radiat Environ Biophys 36:73–83. Kellerer AM, Walsh L, Nekolla EA. 2002c. Risk coefficient for gamma-rays with regard to solid cancer. Radiat Environ Biophys 41:113–123. Kerber RA, Till JE, Simon SL, et al. 1993. A cohort study of thyroid disease in relation to fallout from nuclear weapons testing. JAMA 270:2076–2082. Kesniiniene A, Cardis E, Tenet V, et al. 2002. Studies of cancer risk among Chernobyl liquidators: materials and methods. J Radiol Prot 22:A137–141. Khokhryakov VF, Kellerer AM, Kreisheimer M, Romanov SA. 1998. Lung cancer in nuclear workers of Mayak. A comparison of numerical procedures. Radiat Environ Biophys 37:11–17. Kinlen LJ. 1997. Infection and childhood leukaemia near nuclear sites. Lancet 7:1702. Kinlen LJ, Clarke K, Balkwill A. 1993b. Paternal preconceptional radiation exposure in the nuclear industry and leukaemia and non-Hodgkin’s lymphoma in young people in Scotland. Br Med J 6:1153–1158. Kinlen LJ, Dickson M, Stiller CA. 1995. Childhood leukaemia and nonHodgkin’s lymphoma near large rural construction sites, with a comparison with Sellafield nuclear site [published correction appears in Br Med J 1995;310:911]. Br Med J 310:763–768. Kinlen LJ, Hudson CM, Stiller CA. 1991. Contacts between adults as evidence for an infective origin of childhood leukaemia: An explanation for the excess near nuclear establishments in West Berkshire. Br J Cancer 64:549–554.

288

PART III: THE CAUSES OF CANCER

Kinlen LJ, O’Brien F, Clarke K, et al. 1993a. Rural population mixing and childhood leukaemia: effects of the North Sea oil industry in Scotland, including the area near Dounreay nuclear site. Br Med J 306:743–748. Kinlen LJ, Tiplady P. 2002. The “Seascale cluster”. J Public Health Med 24:342–343. Kleinerman RA, Tucker MA, Tarone RE, et al. 2005. Risk of new cancers after radiotherapy in long-term survivors of retinoblastoma: an extended follow-up. J Clin Oncol 23:2272–2279. Kleinerman RA, Boice JD Jr, Storm HH, et al. 1995. Second primary cancer after treatment for cervical cancer. An international cancer registry study. Cancer 76:442–452. Kleinerman RA, Littlefield LG, Tarone RE, et al. 1989. Chromosome aberrations in peripheral lymphocytes and radiation dose to active bone marrow in patients treated for cancer of the cervix. Radiat Res 119:176–190. Kleinerman RA, Littlefield LG, Tarone RE, et al. 1994. Chromosome aberrations in lymphocytes from women irradiated for benign and malignant gynecological disease. Radiat Res 139:40–46. Kodama Y, Pawel D, Nakamura N, et al. 2001. Stable chromosome aberrations in atomic bomb survivors: Results from 25 years of investigation. Radiat Res156:337–346. Koshurnikova NA, Bolotnikova MG, Ilyin LA, et al. 1998. Lung cancer risk due to exposure to incorporated plutonium. Radiat Res 149:366–371. Koshurnikova NA, Bysogolov GD, Bolotnikova MG, et al. 1996. Mortality among personnel who worked at the Mayak complex in the first years of its operation. Health Phys 71:90–93. Koshurnikova NA, Gilbert ES, Shilnikova NS, et al. 2002a. Studies on the Mayak nuclear workers: health effects. Radiat Environ Biophys 41:29–31. Koshurnikova NA, Gilbert ES, Sokolnikov M, et al. 2000. Bone cancers in Mayak workers. Radiat Res 154:237–245. Koshurnikova NA, Mushkacheva GS, Shilnikova NS, et al. 2002b. Studies on the Ozyorsk population: Health effects. Radiat Environ Biophys 41:37–39. Koshurnikova NA, Shilnikova NS, Okatenko PV, et al. 1997. The risk of cancer among nuclear workers at the “Mayak” production association: preliminary results of an epidemiological study. In: Boice JD Jr, ed. Implications of New Data on Radiation Cancer Risk. NCRP Proc 18:113–122. Kossenko MM. 1996. Cancer mortality among Techa River residents and their offspring. Health Phys 71:77–82. Kossenko MM, Degteva MO. 1994. Cancer mortality and radiation risk for the Techa river population. Sci Total Environ 142:73–89. Kossenko MM, Degteva MO, Vyushkova OV. 1997. Issues in the comparison of risk estimates for the population in the Techa River region and atomic bomb survivors. Radiat Res 148:54–63. Kossenko MM, Preston DL, Krestinina LY. 2002. Studies on the extended Techa river cohort: cancer risk estimation. Radiat Environ Biophys 41:45–48. Kreienbrock L, Poffijn A, Tirmarche M, et al. 1999. Intercomparison of passive radon-detectors under field conditions in epidemiological studies. Health Phys 76:558–563. Kreisheimer M, Sokolnikov ME, Koshurnikova NA, et al. 2003. Lung cancer mortality among nuclear workers of the Mayak facilities in the former Soviet Union. An updated analysis considering smoking as the main confounding factor. Radiat Environ Biophys 42:129–135. Krestinina LY, Preston DL, Ostroumova EV, et al. 2005. Protracted radiation exposure and cancer mortality in the Techa River Cohort. Radiat Res 164:602–611. Krewski D, Lubin JH, Zielinski JM, et al. 2005. Residential radon and risk of lung cancer: a combined analysis of 7 North American case-control studies. Epidemiology 16:137–145. Kuijten RR, Bunin GR, Nass CC, et al. 1990. Gestational and familial risk factors for childhood astrocytoma: results of a case-control study. Cancer Res 50:2608–2612. Lagarde F, Axelsson G, Damber L, et al. 2001. Residential radon and lung cancer among never-smokers in Sweden. Epidemiology 12:396–404. Lagarde F, Pershagen G. 1999. Parallel analyses of individual and ecologic data on residential radon, cofactors, and lung cancer in Sweden. Am J Epidemiol 149: 268–274. Land CE. 1979. The hazards of fallout or of epidemiologic research? [editorial] N Engl J Med 300:431–432. Land CE. 1980. Estimating cancer risks from low doses of ionizing radiation. Science 209:1197–1203. Land CE. 1986. Carcinogenic effects of radiation on the human digestive tract and other organs. In: Upton AC, et al., eds. Radiation Carcinogenesis. New York: Elsevier, pp. 347–378. Land CE, Boice JD Jr, Shore RE, et al. 1980. Breast cancer risk from low-dose exposures to ionizing radiation: results of parallel analysis of three exposed populations of women. J Natl Cancer Inst 65:353–376. Land CE, Hayakawa N, Machado SG, et al. 1994. A case-control interview study of breast cancer among Japanese A-bomb survivors. II. Interactions with radiation dose. Cancer Causes Control 5:167–176.

Land CE, McKay FW, Machado SG. 1984. Childhood leukemia and fallout from the Nevada nuclear tests. Science 223:139–144. Land CE, Shimosato Y, Saccomanno G, et al. 1993b. Radiation-associated lung cancer: A comparison of the pathology of lung cancers in uranium miners and survivors of the atomic bombings of Hiroshima and Nagasaki. Radiat Res 134:234–243. Land CE, Tokunaga M, Koyama K, et al. 2003. Incidence of female breast cancer among atomic bomb survivors, Hiroshima and Nagasaki, 1950–1990. Radiat Res 160:707–717. Land CE, Tokunaga M, Nakamura N. 1993a. Early-onset breast cancer among Japanese A-bomb survivors: Evidence of a radiation susceptible subgroup? Lancet 342:237. Langlois RG, Bigbee WL, Kyoizumi S, et al. 1987. Evidence for increased somatic cell mutations at the glycophorin A locus in atomic bomb survivors. Science 236:445–448. Langner I, Blettner M, Gundestrup M, et al. 2003. Cosmic radiation and cancer mortality among airline pilots: results from a European cohort study (ESCAPE). Radiat Environ Biophys Nov 28 [Epub ahead of print]. Law GR, Parslow RC, Roman E. 2003. United Kingdom Childhood Cancer Study Investigators. Childhood cancer and population mixing. Am J Epidemiol 158:328–336. Le Vu B, de Vathaire F, Shamsaldin A, et al. 1998. Radiation dose, chemotherapy and risk of osteosarcoma after solid tumours during childhood. Int J Cancer 77:370–377. Létourneau EG, Krewski D, Choi NW, et al. 1994. Case-control study of residential radon and lung cancer in Winnipeg, Manitoba, Canada. Am J Epidemiol 140:310–322. Lewis CA, Smith PG, Stratton IM, et al. 1988. Estimated radiation doses to different organs among patients treated for ankylosing spondylitis with a single course of X-rays. Br J Radiol 61:212–220. Lewis EB. 1963. Leukemia, multiple myeloma, and aplastic anemia in American radiologists. Science 142:1492–1494. Likhtarev IA, Sobolev BG, Kairo IA, et al. 1995. Thyroid cancer in the Ukraine. Nature 375:365. Linos A, Gray JE, Orvis AL, et al. 1980. Low-dose radiation and leukemia. N Engl J Med 302:1101–1105. Little JB. 1981. Influence of noncarcinogenic secondary factors on radiation carcinogenesis. Radiat Res 87:240–250. Little MP, Boice JD Jr. 1999. Comparison of breast cancer incidence in the Massachusetts tuberculosis fluoroscopy cohort and in the Japanese atomic bomb survivors. Radiat Res 151:218–224. Little MP, de Vathaire F, Charles MW, et al. 1998b. Variations with time and age in the risks of solid cancer incidence after radiation exposure in childhood. Stat Med 17:1341–1355. Little MP, de Vathaire F, Shamsaldin A, et al. 1998a. Risks of brain tumour following treatment for cancer in childhood: modification by genetic factors, radiotherapy and chemotherapy. Int J Cancer 78:269–275. Little MP, Wakeford R, Charles MW, Andersson A. 1996. A comparison of the risks of leukaemia and non-Hodgkin’s lymphoma in the first generation offspring of the Danish Thorotrast patients with those observed in other studies of parental pre-conception irradiation. J Radiol Prot 16:25–36. Little MP, Weiss HA, Boice JD Jr, et al. 1999. Risks of leukemia in Japanese atomic bomb survivors, in women treated for cervical cancer, and in patients treated for ankylosing spondylitis. Radiat Res 152:280–292. Little MP. 2001a. Cancer after exposure to radiation in the course of treatment for benign and malignant disease. Lancet Oncol 2:212–220. Little MP. 2001b. Comparison of the risks of cancer incidence and mortality following radiation therapy for benign and malignant disease with the cancer risks observed in the Japanese A-bomb survivors. Int J Radiat Biol 77:431–464. Littlefield LG, McFee AF, Salomaa SI, et al. 1998. Do recorded doses overestimate true doses received by Chernobyl cleanup workers? Results of cytogenetic analyses of Estonian workers by fluorescence in situ hybridization. Radiat Res 150:237–249. Lubin JH, Boice JD Jr, Edling C et al. 1994a. Lung cancer following radon exposure among underground miners: A joint analysis of 11 studies. NIH Publ No. 94–3644. Washington, DC: US Govt Print Office. Lubin JH, Boice JD Jr, Edling C, et al. 1995a. Lung cancer in radon-exposed miners and estimation of risk from indoor exposure. J Natl Cancer Inst 87:817–827. Lubin JH, Boice JD Jr, Edling C, et al. 1995b. Radon-exposed underground miners and the inverse dose-rate (protraction enhancement) effects. Health Phys 69:494–500. Lubin JH, Boice JD Jr, Samet JM. 1995c. Errors in radon dosimetry might explain mixed results for indoor studies. Radiat Res 144:329–341. Lubin JH, Boice JD Jr. 1997a. Lung cancer risk from residential radon: Meta-analysis of eight epidemiologic studies. J Natl Cancer Inst 87: 49–57.

Ionizing Radiation Lubin JH, Liang Z, Hrubec Z, et al. 1994b. Radon exposure in residences and lung cancer among women: combined analysis of three studies. Cancer Causes Control 5:114–128. Lubin JH, Linet MS, Boice JD Jr, et al. 1998. Case-control study of childhood acute lymphoblastic leukemia and residential radon exposure. J Natl Cancer Inst 90:294–300. Lubin JH, Samet JM, Weinberg C. 1990b. Design issues in epidemiologic studies of indoor exposure to Rn and risk of lung cancer. Health Phys 59:807–817. Lubin JH, Tomásˇek L, Edling C, et al. 1997b. Estimating lung cancer mortality from residential radon using data from low exposure of miners. Radiat Res 147:126–134. Lubin JH, Wang ZY, Boice JD Jr, et al. 2004a. Risk of lung cancer and residential radon in China: pooled results of two studies. Int J Cancer 109:132–137. Lubin JH, Schafer DW, Ron E, et al. 2004b. A reanalysis of thyroid neoplasms in the Israeli tinea capitis study accounting for dose uncertainties. Radiat Res 161:359–368. Lund E, Galanti MR. 1999. Incidence of thyroid cancer in Scandinavia following fallout from atomic bomb testing: An analysis of birth cohorts. Cancer Causes Control 10:181–187. Lundell M, Hakulinen T, Holm LE. 1994. Thyroid cancer after radiotherapy for skin hemangioma in infancy. Radiat Res 140:334–339. Lundell M, Mattsson A, Hakulinen T, Holm LE. 1996. Breast cancer after radiotherapy for skin hemangioma in infancy. Radiat Res 145:225–230. Lundell M, Mattsson A, Karlsson P, et al. 1999. Breast cancer risk after radiotherapy in infancy: A pooled analysis of two Swedish cohorts of 17,202 infants. Radiat Res 151:626–632. Lundin FD Jr, Wagoner JK, Archer VE. 1971 Radon Daughter Exposure and Respiratory Cancer. Quantitative and Temporal Aspects. National Institute for Occupational Safety and Health and National Institute of Environmental Health Sciences Joint Monograph No. 1. Washington, DC, U.S. Department of Health, Education, and Welfare, Public Health Service. Lyon JL, Klauber MR, Gardner JW, et al. 1979. Childhood leukemias associated with fallout from nuclear testing. N Engl J Med 300:397–402. Machado SG, Land CE, McKay FW. 1987. Cancer mortality and radioactive fallout in southwestern Utah. Am J Epidemio1125:44–61. MacMahon B. 1962. Prenatal X-ray exposure and childhood cancer. J Natl Cancer Inst 28:1173–1191. MacMahon B. 1985. Prenatal X-ray exposure and twins. N Engl J Med 312:576–577. MacMahon B. 1989. Some recent issues in low-exposure radiation epidemiology. Environ Health Perspect 81:131–135. MacMahon B. 1992. Leukemia clusters around nuclear facilities in Britain. Cancer Causes Control 3:283–288. Major IR, Mole RH. 1978. Myeloid leukemia in X-ray irradiated CBA mice. Nature 272:455–456. Mancuso TF, Stewart A, Kneale G. 1977. Radiation exposures of Hanford workers dying from cancer and other causes. Health Phys 33:369–385. March HC. 1944. Leukemia in radiologists. Radiology 43:275–278. Marshall JH, Groer PG, Shlenker RA. 1978. Dose to endosteal cells and relative distribution factors for radium-224 and plutonium-239 compared to radium-226. Health Phys 35:91–101. Marshall JH, Groer PG. 1977. A theory of the induction of bone cancer by alpha radiation. Radiat Res 71:149–192. Martignoni KD. 2003. New study on leukaemia and malignant lymphoma in the vicinity of the Kruemmel power plant in northern Germany. J Radiol Prot 23:349–351. Mason TJ, Fraumeni JF Jr, McKay FW Jr. 1972. Uranium mill tailings and cancer mortality in Colorado. J Natl Cancer Inst 49:661–664. Mason TJ, Miller RW. 1974. Cosmic radiation at high altitudes and U.S. cancer mortality, 1950–1969. Radiat Res 60:302–306. Matanoski G. 1993. Nuclear shipyard workers study. Radiat Res 133:126–127. Matanoski GM, Seltser R, Sartwell PE, et al. 1975. The current mortality rates of radiologists and other physician specialists: Specific causes of death. Am J Epidemiol 101:199–210. Mattson A, Ruden B, Hall R, et al. 1993. Radiation-induced breast cancer and long term follow-up of radiotherapy for benign breast disease. J Natl Cancer Inst 85:1679–1685. Mays CW, Spiess H. 1984. Bone sarcomas in patients given radium-224. In: Boice JD Jr, Fraumeni JF Jr, eds. Radiation Carcinogenesis: Epidemiology and Biological Significance. New York: Raven Press, pp. 241–252. McGeoghegan D, Gillies M, Riddell AE, Binks K. 2003. Mortality and cancer morbidity experience of female workers at the British Nuclear Fuels Sellafield plant, 1946–1998. Am J Ind Med 44:653–663. McKinney PA, Alexander FE, Cartwright RA, et al. 1991. Parental occupations of children with leukaemia in West Cumbria, North Humerside, and Gateshend. Br Med J 302:681–687.

289

McLaughlin JR, Clarke EA, Nishri D, et al. 1993b. Childhood leukaemia in the vicinity of Canadian nuclear facilities. Cancer Causes Control 4:51–58. McLaughlin JR, King WD, Anderson TW, et al. 1993c. Paternal radiation exposure and leukaemia in offspring: The Ontario case-control study. Br Med J 307:959–966. McLaughlin JR, Kreiger N, Sloan MP, et al. 1993a. An historical cohort study of cardiac catheterization during childhood and the risk of cancer. Int J Epidemiol 22:584–591. Meadows AT. 2003. Pediatric cancer survivors: Past history and future challenges. Curr Probl Cancer 27:112–126. Mertens AC, Yasui Y, Neglia JP, et al. 2001. Late mortality experience in fiveyear survivors of childhood and adolescent cancer: The Childhood Cancer Survivor Study. J Clin Oncol 19:3163–3172. Mettler FA, Williamson MR, Royal HD, et al. 1992. Thyroid nodules in the population living around Chernobyl. JAMA 268:616–619. Michaelis J, Keller B, Haaf G, et al. 1992. Incidence of childhood malignancies in the vicinity of West German nuclear power plants. Cancer Causes Control 3:255–263. Michaelis J. 1998. Recent epidemiological studies on ionizing radiation and childhood cancer in Germany. Int J Radiat Biol 73:377–381. Miller AB, Howe GR, Sherman GJ, et al. 1989. Mortality from breast cancer after irradiation during fluoroscopic examinations in patients being treated for tuberculosis. N Engl J Med 321:1285–1289. Miller RW, Boice JD Jr. 1986. Radiogenic cancer after prenatal or childhood exposure. In: Upton AC, et al., eds. Radiation Carcinogenesis. New York: Elsevier, pp. 379–386. Miller RW. 1969. Delayed radiation effects in atomic-bomb survivors. Science 166:569–574. Miller RW. 1995. Delayed effects of external radiation exposure: a brief history. Radiat Res 144:160–169. MMWR. Morbidity and Mortality Weekly Reports. 1985. Health hazards associated with elevated levels of radon—Pennsylvania. JAMA 254:3168. Modan B, Chetrit A, Alfandary E, et al. 1989. Increased risk of breast cancer after low-dose irradiation. Lancet 1:629–631. Modan B, Keinan L, Blumstein T, Sadetzki S. 2000. Cancer following cardiac catheterization in childhood. Int J Epidemiol 29:424–428. Modan B, Lilienfeld AM. 1965. Polycythemia vera and leukemia-the role of radiation treatment: A study of 1222 patients. Medicine 44:305–344. Mohan AK, Hauptmann M, Freedman DM, et al. 2003. Cancer and other causes of mortality among radiologic technologists in the United States. Int J Cancer 103:259–267. Mohan AK, Hauptmann M, Linet MS, et al. 2002. Breast cancer mortality among female radiologic technologists in the United States. J Natl Cancer Inst 94:943–948. Mole RH. 1974. Antenatal irradiation and childhood cancer: Causation or coincidence? Br J Cancer 30:199–208. Mole RH. 1975. Ionizing radiation as a carcinogen: practical questions and academic pursuits. Br J Radiol 48:157–169. Mole RH, Major IR. 1983. Myeloid leukaemia frequency after protracted exposure to ionizing radiation: experimental confirmation of the flat doseresponse found in ankylosing spondylitis after a single treatment course with X-rays. Leuk Res 7:295–300. Moloney WC, Kastenbaum MA. 1955. Leukemogenic effects of ionizing radiation on atomic bomb survivors in Hiroshima City. Science 121:308–309. Monson RR, MacMahon B. 1984. Prenatal X-ray exposure and cancer in children. In: Boice JD Jr, Fraumeni JF Jr, eds. Radiation Carcinogenesis: Epidemiology and Biological Significance. New York: Raven Press, pp. 97–105. Morgan WF. 2003. Non-targeted and delayed effects of exposure to ionizing radiation: I. Radiation-induced genomic instability and bystander effects in vitro. Radiat Res 159:567–580. Mori T, Kido C, Fukutomi K, et al. 1999. Summary of entire Japanese thorotrast follow-up study: Updated 1998. Radiat Res 152(6 Suppl):S84–S87. Mossman KL. 2003. Bomb dosimetry unlikely to change risk estimates. Nature 425:237. Moysich KB, Menezes RJ, Michalek AM. 2002. Chernobyl-related ionising radiation exposure and cancer risk: An epidemiological review. Lancet Oncol 3:269–279. Muirhead CR, Bingham D, Haylock RG, et al. 2003. Follow up of mortality and incidence of cancer 1952–98 in men from the UK who participated in the UK’s atmospheric nuclear weapon tests and experimental programmes. Occup Environ Med 60:165–172. Muirhead CR, Butland BK, Green BM, et al. 1991. Childhood leukaemia and natural radiation. Lancet 337:503–504.Muirhead CR, Goodill AA, Haylock RG, et al. 1999. Occupational radiation exposure and mortality: second analysis of the National Registry for Radiation Workers. J Radiol Prot 19:3–26. Mulvihill JJ, Myers MH, Connelly RR, et al. 1987. Cancer in offspring of longterm survivors of childhood and adolescent cancer. Lancet 2:813–817.

290

PART III: THE CAUSES OF CANCER

Najarian T, Colton T. 1978. Mortality from leukemia and cancer in shipyard nuclear workers. Lancet 1:1018–1020. Najean Y, Rain JD. 1997. Treatment of polycythemia vera: Use of 32P alone or in combination with maintenance therapy using hydroxyurea in 461 patients greater than 65 years of age. The French Polycythemia Study Group. Blood 89:2319–2327. Najean Y, Rain JD, Dresch C, et al. 1996. Risk of leukaemia, carcinoma, and myelofibrosis in 32P- or chemotherapy-treated patients with polycythaemia vera: a prospective analysis of 682 cases. The “French Cooperative Group for the Study of Polycythaemias”. Leuk Lymphoma 22 (suppl 1):111–119. Nambi KS, Soman SD. 1987. Environmental radiation and cancer in India. Health Phys 52:653–657. NAS (National Academy of Sciences). 1972. Advisory Committee on the Biological Effects of Ionizing Radiations (The BEIR Report): The Effects on Populations of Exposure to Low Levels of Ionizing Radiation. Washington DC: U.S. Government Printing Office. NAS (National Academy of Sciences). 1980. Committee on the Biological Effects of Ionizing Radiations (BEIR III). The Effects on Populations of Exposure to Low Levels of Ionizing Radiation: 1980. Washington, DC: National Academy Press. NAS (National Academy of Sciences). 1988. Health Risks of Radon and Other Internally Deposited Alpha-Emitters (BEIR IV Report). Washington, DC: National Academy Press. NAS (National Academy of Sciences). 1990. Health Effects of Exposure to Low Levels of Ionizing Radiation (BEIR V). Washington, DC: National Academy Press. NAS (National Academy of Sciences). 1991. Comparative Dosimetry of Radon in Mines and Homes. Washington, DC: National Academy Press. NAS (National Academy of Sciences). 1999. Committee on Health Risks of Exposure to Radon. Health Effects of Exposure to Radon. BEIR VI. Washington, DC: National Academy Press. NAS (National Academy of Sciences). 2000. Review of the Hanford Thyroid Disease Study Draft Final Report. Washington, DC: National Academy Press. NAS (National Academy of Sciences). 2005. Committee to Assess Health Risks from Exposure to Low Levels of Ionizing Radiation. Health Risks from Exposure to Low Levels of Ionizing Radiation: BEIR VII – Phase 2. Washington, DC: The National Academies Press. Naumburg E, Bellocco R, Cnattingius S, et al. 2001. Intrauterine exposure to diagnostic x rays and risk of childhood leukemia subtypes. Radiat Res 156:718–723. NCRP (National Council on Radiation Protection and Measurements). 1980. Report No. 64. Influence of Dose and Its Distribution in Time on DoseResponse Relationships for Low-LET Radiations. Washington, DC: NCRP. NCRP. 2001. Report No. 136. Evaluation of the Linear-Nonthreshold DoseResponse Model for Ionizing Radiation. Bethesda, MD: NCRP. Neel JV, Schull WJ, eds. 1991. The Children of Atomic Bomb Survivors. A Genetic Study. Washington, DC: Natl Acad Press, pp 1518. Neglia JP, Friedman DL, Yasui Y, et al. 2001. Second malignant neoplasms in five-year survivors of childhood cancer: Childhood cancer survivor study. J Natl Cancer Inst 93:618–629. Neglia JP, Meadows AT, Robison LL, et al. 1991. Second neoplasms after acute lymphoblastic leukemia in childhood. N Engl J Med 325:1330–1336. Nekolla EA, Kellerer AM, Kuse-Isingschulte M, et al. 1999. Malignancies in patients treated with high doses of radium-224. Radiat Res 152(6 suppl):S3–7. Nekolla EA, Kreisheimer M, Kellerer AM, et al. 2000. Induction of malignant bone tumors in radium-224 patients: Risk estimates based on the improved dosimetry. Radiat Res 153:93–103. Neuberger JS, Gesell TF. 2002. Residential radon exposure and lung cancer: Risk in nonsmokers. Health Phys 83:1–18. Neugut AI, Meadows AT, Robinson E. 1999. Multiple Primary Cancer. Philadelphia: Lippincott. Neugut AI, Robinson E, Lee WC, et al. 1994. Lung cancer after radiotherapy for breast cancer. Cancer 71:3054–3057. NIH (National Institutes of Health). 1985. Report of the National Institutes of Health Ad Hoc Working Group to Develop Radioepidemiological Tables, NIH Publ No 85–2748, U.S. DHHS, Public Health Service, NIH: Washington, DC. NIH (National Institutes of Health). 2003. Report of the NCI-CDC Working Group to Revise the 1985 NIH Radioepidemiological Tables. NIH Publ No. 03–5387. U.S. DHHS, NIH: Washington, DC. Noguchi K, Shimizu M, Anzai I. 1986. Correlation between natural radiation exposure and cancer mortality in Japan (I). J Radiat Res (Tokyo) 27:191–212. Nomura T. 1982. Prenatal exposure to X-rays and chemicals induces heritable tumours and anomalies in mice. Nature 296:575–577. NRPB (National Radiological Protection Board). 2001 ELF. Electromagnetic Fields and the Risk of Cancer. vol 12, no 1. Chilton: NRPB.

NRPB (National Radiological Protection Board). 2003a. Risk of Leukaemia and Related Malignancies following Radiation Exposure: Estimates for the UK Population: Report of an Advisory Group on Ionising Radiation. vol 14, no 1. Chilton: NRPB. Available at: http://www.nrpb.org/publications/ documents_of_nrpb/pdfs/doc_14_1.pdf. Accessed January 16, 2004. NRPB (National Radiological Protection Board). 2003b. Health Effects from Radiofrequency Electromagnetic Fields. Report of an Independent Advisory Group on Non-ionising Radiation. vol 14, no 2. Chilton: NRPB. Available at: http://www.nrpb.org/publications/documents_of_nrpb/pdfs/ doc_14_2.pdf. Accessed January 16, 2004. NRC (National Research Council). 1996. Committee on Evaluation of 1950s Air Force Human Health Testing in Alaska Using Iodine-131. The Arctic Aeromedical Laboratory’s Thyroid Function Study: A Radiological Risk and Ethical Analysis. Washington, DC: The National Academy Press. NRC (National Research Council). 2003. A Review of the Dose Reconstruction Program of the Defense Threat Reduction Agency. Committee to Review the Dose Reconstruction Program of the Defense Threat Reduction Agency. Board on Radiation Effects Research, Division on Earth and Life Studies. Washington, DC, National Academies Press. NRC (National Research Council). 2005. Assessment of the scientific information for the Radiation Exposure Screening and Education Program. Washington, DC: The National Academies Press. Nyberg U, Nilsson B, Travis LB, et al. 2002. Cancer incidence among Swedish patients exposed to radioactive thorotrast: A forty-year follow-up survey. Radiat Res 157:419–425. Offit K, Gilad S, Paglin S, et al. 2002. Rare variants of ATM and risk for Hodgkin’s disease and radiation-associated breast cancers. Clin Cancer Res 8:3813–3819. Omar RZ, Barber JA, Smith PG. 1999. Cancer mortality and morbidity among plutonium workers at the Sellafield plant of British Nuclear Fuels. Br J Cancer 79:1288–1301. Oppenheim BE, Griem ML, Meier P. 1974. Effects of low-dose prenatal irradiation in humans: Analysis of Chicago lying-in data and comparison with other studies. Radiat Res 57:508–544. Parker L, Craft AW, Smith J, et al. 1993. Geographical distribution of preconceptional radiation doses to fathers employed at the Sellafield nuclear installation, West Cumbria. Br Med J 307:966–971. Parkin DM, Clayton D, Black RJ, et al. 1996. Childhood leukaemia in Europe after Chernobyl: 5 year follow-up. Br J Cancer 73:1006–1012. Pershagen G, Akerblom G, Axelson O, et al. 1994. Residential radon exposure and lung cancer in Sweden. N Engl J Med 330:159–164. Pierce DA, Preston DL. 1997. No evidence for increased tumor rates below 200 mSv in the atomic bomb survivors’ data. Radiat Environ Biophys 36:209–210; discussion 211–212. Pierce DA, Preston DL. 2000. Radiation-related cancer risks at low doses among atomic bomb survivors. Radiat Res 154:178–186. Pierce DA, Sharp GB, Mabuchi K. 2003. Joint effects of lung cancer and smoking on lung cancer risk among atomic bomb survivors. Radiat Res 159:511–520. Pierce DA, Shimizu Y, Preston DL, et al. 1996. Studies of the mortality of atomic bomb survivors. Report 12, Part I. Cancer: 1950–1990. Radiat Res 146:1–27. Pobel D, Viel JF. 1997. Case-control study of leukaemia among young people near La Hague nuclear reprocessing plant: the environmental hypothesis revisited. BMJ 314:101–106. Pochin EE. 1976. Problems involved in detecting increased malignancy rates in areas of high natural radiation background. Health Phys 31:148–151. Polednak AP. 1978. Bone cancer among female radium dial workers: Latency periods and incidence rates by time after exposure: brief communication. J Natl Cancer Inst 60:77–82. Pottern LM, Kaplan MM, Larsen PR, et al. 1990. Thyroid nodularity after childhood irradiation for lymphoid hyperplasia: a comparison of questionnaire and clinical findings. J Clin Epidemiol 43:449–460. Pottern LM, Stone BJ, Day NE, et al. 1980. Thyroid cancer in Connecticut, 1935–1975: An analysis by cell type. Am J Epidemiol 112:764–774. Prentice RL, Kato H, Yoshimoto K, et al. 1982. Radiation exposure and thyroid cancer incidence among Hiroshima and Nagasaki residents. Natl Cancer Inst Monogr 62:207–212. Preston D, Kusumi S, Tomonaga M, et al. 1994. Cancer incidence in atomic bomb survivors, Part III: Leukemia, lymphoma, and multiple myeloma, 1950–87. Radiat Res 137:S68–S97. Preston DL, Mattsson A, Holmberg E, et al. 2002a. Radiation effects on breast cancer risk: A pooled analysis of eight cohorts. Radiat Res 158:220–235. Preston DL, Ron E, Yonehara S, et al. 2002b. Tumors of the nervous system and pituitary gland associated with atomic bomb radiation exposure. J Natl Cancer Inst 94:1555–1563. Preston DL, Shimizu Y, Pierce DA, et al. 2003. Studies of mortality of atomic bomb survivors. Report 13: Solid cancer and noncancer disease mortality: 1950–1997. Radiat Res 160:381–407.

Ionizing Radiation Preston DL, Pierce DA, Shimizu Y, et al. 2004. Effect of recent changes in atomic bomb survivor dosimetry on cancer mortality risk estimates. Radiat Res 162:377–389. Preston-Martin S, Paganini-Hill A, Henderson BE, et al. 1980. Case-control study of intracranial meningiomas in women in Los Angeles County, California. J Natl Cancer Inst 65:67–73. Preston-Martin S, Pogoda JM. 2003. Estimation of radiographic doses in a casecontrol study of acute myelogenous leukemia. Health Phys 84:245–259. Preston-Martin S, Thomas DC, White SC, et al. 1988. Prior exposure to medical and dental X-rays related to tumors of the parotid gland. J Natl Cancer Inst 80:943–949. Preston-Martin S, Thomas DC, Yu MC, et al. 1989. Diagnostic radiography as a risk factor for chronic myeloid and monocytic leukaemia (CML). Br J Cancer 59:639–644. Preston-Martin S, Yu MC, Benton B, et al. 1982. N-nitroso compounds and childhood brain tumors: A case-control study. Cancer Res 42:5240–5245. Priest ND. 1989. Alpha-emitters in the skeleton: An evaluation of the risk of leukaemia following intakes of plutonium 239. In: Taylor DM, Mays CW, Gerber GB, Thomas RG, eds. Risks from Radium and Thorotrast. London: British Institute of Radiology; Report 21, pp. 159–165. Priest ND. 2001. Toxicity of depleted uranium. Lancet 357:244–246. Pukkala E, Aspholm R, Auvinen A, et al. 2003. Cancer incidence among 10,211 airline pilots: A Nordic study. Aviat Space Environ Med 74:699–706. Puskin JS. 2003. Smoking as a confounder in ecologic correlations of cancer mortality rates with average county radon levels. Health Phys 84: 526–532. Rahu M, Tekkel M, Veidebaum T, et al. 1997. The Estonian study of Chernobyl cleanup workers: II. Incidence of cancer and mortality. Radiat Res 147:653–657. Rallison ML, Dobyns BM, Keating FR, et al. 1975. Thyroid nodularity in children. JAMA 233:1069–1072. Rallison ML, Lotz TM, Bishop M, et al. 1990. Cohort study of thyroid disease near the Nevada Test Site: A preliminary report. Health Phys 59:739–746. Relling MV, Rubnitz JE, Rivera GK, 1999. High incidence of secondary brain tumours after radiotherapy and antimetabolites. Lancet 354:34–39. Richardson S, Monfort C, Green M, et al. 1995. Spatial variation of natural radiation and childhood leukaemia incidence in Great Britain. Stat Med 14:2487–2501. Rinsky RA, Melius JM, Hornung RW, et al. 1988. Case-control study of lung cancer in civilian employees at the Portsmouth Naval Shipyard, Kittery, Maine. Am J Epidemiol 127:55–64. Rinsky RA, Zumwalde RD, Waxweiler RJ, et al. 1981. Cancer mortality at a naval nuclear shipyard. Lancet 1:231–235. Robbins J, Adams WH. 1989. Radiation effects in the Marshall Islands. In: Nagataki S, ed. Radiation and the Thyroid. Tokyo: Excerpta Medica, pp 11–24. Robbins J, Schneider AB. 2000. Thyroid cancer following exposure to radioactive iodine. Rev Endocr Metab Disord 1:197–203. Robinette CD, Jablon S, Preston TL. 1985. Mortality of Nuclear Weapons Test Participants. Washington, DC: Natl Acad Press. Ron E, Auvinen A, Alfandary E, et al. 1999. Cancer risk following radiotherapy for infertility or menstrual disorders. Int J Cancer 82:795–798. Ron E, Boice JD Jr, Hamburger S, et al. 1994a. Mortality following radiation treatment for infertility of hormonal origin or amenorrhoea. Int J Epidemiol 23:1165–1173. Ron E, Doody MM, Becker DV, et al. 1998b. Cancer mortality following treatment for hyperthyroidism. Cooperative Thyrotoxicosis Therapy Followup Study Group. JAMA 280:347–355. Ron E, Kleinerman RA, Boice JD Jr, et al. 1987. A population-based casecontrol study of thyroid cancer. J Natl Cancer Inst 79:1–12. Ron E, Lubin J, Schneider AB. 1992. Thyroid cancer incidence. Nature 360:113. Ron E, Lubin JH, Shore RE, et al. 1995. Thyroid cancer after exposure to external radiation: A pooled analysis of seven studies. Radiat Res 141:259–277. Ron E, Modan B, Boice JD Jr. 1988a. Mortality after radiotherapy for ringworm of the scalp, Am J Epidemiol 127:713–725. Ron E, Modan B, Boice JD Jr, et al. 1988b. Tumors of the brain and nervous system after radiotherapy in childhood. N Engl J Med 319:1033–1039. Ron E, Modan B, Preston D, et al. 1989. Thyroid neoplasia following low-dose radiation in childhood. Radiat Res 120:516–531. Ron E, Modan B, Preston D, et al. 1991. Radiation-induced skin carcinoma of the head and neck. Radiat Res 125:318–325. Ron E, Preston DL, Kishikawa M, et al. 1998a. Skin tumor risk among atomicbomb survivors in Japan. Cancer Causes Control 9:393–401. Ron E, Preston DL, Mabuchi K, et al. 1994b. Cancer incidence in atomic bomb survivors. Part IV: Comparison of cancer incidence and mortality. Radiat Res 137:S98–S112. Ron E. 2003. Cancer risks from medical radiation. Health Phys 85:47–59.

291

Ronckers CM, Land CE, Hayes RB, et al. 2002a. Late health effects of childhood nasopharyngeal radium irradiation: Nonmelanoma skin cancers, benign tumors, and hormonal disorders. Pediatr Res 52:850–858. Ronckers CM, Land CE, Verduijn PG, et al. 2001. Cancer mortality after nasopharyngeal radium irradiation in the Netherlands: A cohort study. J Natl Cancer Inst 93:1021–1027. Ronckers CM, Van Leeuwen FE, Hayes RB, et al. 2002b. Cancer incidence after nasopharyngeal radium irradiation. Epidemiology 13:552–560. Rooney C, Beral V, Maconochie N, et al. 1993. Case-control study of prostate cancer in the United Kingdom Atomic Energy Authority employees. Br Med J 307:1391–1397. Roscoe RJ, Steenland K, Halperin WE, et al. 1989. Lung cancer mortality among nonsmoking uranium miners exposed to radon daughters. JAMA 262:629–633. Rothkamm K, Lobrich M. 2003. Evidence for a lack of DNA double-strand break repair in human cells exposed to very low X-ray doses. Proc Natl Acad Sci U S A 100:5057–5062. Rowland RE, Stehney AF, Lucas HF Jr. 1978. Dose-response relationships for female radium dial workers. Radiat Res 76:368–383. Rubino C, de Vathaire F, Diallo I, et al. 2002. Radiation dose, chemotherapy and risk of lung cancer after breast cancer treatment. Breast Cancer Res Treat 75:15–24. Rubino C, de Vathaire F, Dottorini ME, et al. 2003. Second primary malignancies in thyroid cancer patients. Br J Cancer 89:1638–1644. Ryberg M, Lundell M, Nilsson B, Pettersson F. 1990a. Malignant disease after radiation treatment of benign gynaecological disorders. A study of a cohort of metropathia patients. Acta Oncol 29:563–567. Ryberg M, Nilsson B, Pettersson F. 1990b. Cardiovascular death after radiotherapy for benign bleeding disorders. The Radiumhemmet metropathia cohort 1912–1977. J Intern Med 227:95–99. Saenger EL, Thoma GE, Tompkins EA. 1968. Incidence of leukemia following treatment of hyperthyroidism: preliminary report of the Cooperative Thyrotoxicosis Therapy Follow-up Study. JAMA 205:855–862. Saku T, Hayashi Y, Takahara O, et al. 1997. Salivary gland tumors among atomic bomb survivors, 1950–1987. Cancer 79:1465–1475. Salomaa S, Lindholm C, Tankimanova MK, et al. 2002. Stable chromosome aberrations in the lymphocytes of a population living in the vicinity of the Semipalatinsk nuclear test site. Radiat Res 158:591–596. Sampson RJ, Key CR, Buncher CR, et al. 1969. Thyroid carcinoma in Hiroshima and Nagasaki. I. Prevalence of thyroid carcinoma at autopsy. JAMA 209:65–70. Schafer DW, Lubin JH, Ron E, et al. 2001. Thyroid cancer following scalp irradiation: A reanalysis accounting for uncertainty in dosimetry. Biometrics 57:689–697. Schull WJ. 2003. The children of atomic bomb survivors: a synopsis. J Radiol Prot 23:369–384. Schneider AB, Lubin J, Ron E, et al. 1998. Salivary gland tumors after childhood radiation treatment for benign conditions of the head and neck: Dose-response relationships. Radiat Res 149:625–630. Schneider AB, Ron E, Lubin J, et al. 1993. Dose-response relationships for radiation-induced thyroid cancer and thyroid nodules: Evidence for the prolonged effects of radiation on the thyroid. J Clin Endocrinol Metab 77:362–364. Schneider AB, Shore-Freedman E, Ryo UY, et al 1985. Radiation-induced tumors of the head and neck following childhood irradiation. Prospective studies, Medicine 64:1–15. Schwartz CL. 2003. Health status of childhood cancer survivors [editorial]. JAMA 290:1641–1643. Shafman TD, Levitz S, Nixon AJ, et al. 2000. Prevalence of germline truncating mutations in ATM in women with a second breast cancer after radiation therapy for a contralateral tumor. Genes Chromosomes Cancer 27:124–129. Shakhtarin VV, Tsyb AF, Stepanenko VF, et al. 2003. Iodine deficiency, radiation dose, and the risk of thyroid cancer among children and adolescents in the Bryansk region of Russia following the Chernobyl power station accident. Int J Epidemiol 32:584–591. Shilnikova NS, Preston DL, Ron E, et al. 2003. Cancer mortality risk among workers at the Mayak nuclear complex. Radiat Res 159:787–798. Shore RE. 1992. Issues and epidemiological evidence regarding radiationinduced thyroid cancer. Radiat Res 131:98–111. Shore RE, Albert RE, Pasternack BS. 1976. Follow-up study of patients treated by X-ray epilation for tinea capitis: Resurvey of post-treatment illness and mortality experience. Arch Environ Health 31:21–28. Shore RE, Albert RE, Reed M, et al. 1984. Skin cancer incidence among children irradiated for ringworm of the scalp. Radiat Res 100:192–204. Shore RE, Hildreth N, Dvoretsky P, et al. 1993. Benign thyroid adenomas among persons x-irradiated in infancy for enlarged thymus gland. Radiat Res 134:217–223. Shore RE, Hildreth N, Woodard E, et al. 1986. Breast cancer among women given X-ray therapy for acute postpartum mastitis. J Natl Cancer Inst 77:689–696.

292

PART III: THE CAUSES OF CANCER

Shore RE, Moseson M, Harley N, Pasternack BS. 2003. Tumors and other diseases following childhood X-ray treatment for ringworm of the scalp (Tinea capitis). Health Phys 85:404–440. Shore RE, Moseson M, Xue X, et al. 2002. Skin cancer after X-ray treatment for scalp ringworm. Radiat Res 157:410–418. Shore RE, Woodward E, Hildreth N, et al. 1985. Thyroid tumors following thymus irradiation. J Natl Cancer Inst 74:1177–1184. Shu XO, Potter JD, Linet MS, et al. 2002. Diagnostic X-rays and ultrasound exposure and risk of childhood acute lymphoblastic leukemia by immunophenotype. Cancer Epidemiol Biomarkers Prev 11:177–185. Sigurdson AJ, Ronckers CM, Mertens AC, et al. 2005. Primary thyroid cancer after a first tumour in childhood (the Childhood Cancer Survivor Study): a nested case-control study. Lancet 365:2014–2023. Simon SL, Baverstock KF, Lindholm C; World Health Organization; Radiation and Nuclear Safety Authority in Finland; National Cancer Institute. 2003. A summary of evidence on radiation exposures received near to the Semipalatinsk nuclear weapons test site in Kazakhstan. Health Phys 84: 718–725. Smith BJ, Field RW, Lynch CF. 1998. Residential 222Rn exposure and lung cancer: Testing the linear no-threshold theory with ecologic data. Health Phys 75:11–17. Smith PG, Doll R, Radford EP. 1977. Cancer mortality among patients with ankylosing spondylitis not given X-ray therapy. Br J Radiol 50:728–734. Smith PG, Doll R. 1982. Mortality among patients with ankylosing spondylitis after a single treatment course with X-rays. Br Med J 284:449–460. Smith PG, Douglas AJ. 1986. Mortality of workers at the Sellafield plant of British Nuclear Fuels. Br Med J 293:845–854. Smith PG. 1991. Case-control studies of leukaemia clusters. Br Med J 302:672–673. Socie G, Curtis RE, Deeg HJ, et al. 2000. New malignant diseases after allogeneic marrow transplantation for childhood acute leukemia. J Clin Oncol 18:348–357. Sont WN, Zielinski JM, Ashmore JP, et al. 2001. First analysis of cancer incidence and occupational radiation exposure based on the National Dose Registry of Canada. Am J Epidemiol 153:309–318. Sorahan T, Haylock RG, Muirhead CR, et al. 2003. Cancer in the offspring of radiation workers: An investigation of employment timing and a reanalysis using updated dose information. Br J Cancer 89:1215–1220. Spiers FW. 1979. Background radiation and estimated risks from low-dose irradiation. Health Phys 37:784–789. Spiers FW, Lucas HF, Rundo J, et al. 1983. Leukaemia incidence in the U.S. dial workers. Health Phys 44(suppl 1):65–72. Spiess H, Mays CW, Chmelevsky D. 1989. Malignancies in patients injected with radium 224. In: Taylor DM, Mays CW, Gerber GB, eds. Risks from Radium and Thorotrast. BIR Report 21. London: Brit Inst Radiol, pp. 7–12. Stebbings JH, Lucas HF, Stehney AF. 1984. Mortality from cancers of major sites in female radium dial workers. Am J Ind Med 5:435–459. Stehney AF, Lucas HF Jr, Rowland RE. 1978. Survival times of women radium dial workers first exposed before 1930. In: Late Biological Effects of Ionizing Radiation. vol 1. Vienna: International Atomic Energy Agency, pp. 333–351. Stern FB, Waxweiler RA, Beaumont JJ, et al. 1986. A case control study of leukemia at a naval nuclear shipyard. Am J Epidemiol 123:980–992. Stern RS, Laird N, Melski J, et al. 1984. Cutaneous squamous cell carcinoma in patients treated with PUVA. N Engl J Med 310:1156–1161. Stern RS, Liebman EJ, Vakeva L. 1998. Oral psoralen and ultraviolet-A light (PUVA) treatment of psoriasis and persistent risk of nonmelanoma skin cancer. PUVA Follow-up Study. J Natl Cancer Inst 90:1278–1284. Stevens W, Thomas DC, Lyon JL, et al. 1990. Leukemia in Utah and radioactive fallout from the Nevada test site: A case-control study. JAMA 264:585–591. Stewart A. 1973. The carcinogenic effects of low level radiation: A re-appraisal of epidemiologists methods and observations. Health Phys 24:223–240. Stewart A, Kneale GW. 1970. Radiation dose effects in relation to obstetric Xrays and childhood cancers. Lancet 1:1185–1188. Stewart A, Webb J, Hewitt D. 1958. A survey of childhood malignancies. Br Med J 1:1495–1508. Stidley CA, Samet JM. 1993. A review of ecologic studies of lung cancer and indoor radon. Health Phys 65:234–251. Storm HH, Andersson M, Boice JD Jr, et al. 1992. Adjuvant radiotherapy and risk of contralateral breast cancer. J Natl Cancer Inst 84:1245–1250. Straume T, Rugel G, Marchetti AA, et al. 2003. Measuring fast neutrons in Hiroshima at distances relevant to atomic-bomb survivors. Nature 424:539–542. Strong LC. 1977. Genetic and environmental interactions. Cancer 40:1861–1866. Strong LC, Herson J, Osborne BM, et al. 1979. Risk of radiation-related subsequent malignant tumors in survivors of Ewing’s sarcoma. J Natl Cancer Inst 62:1401–1406.

Swerdlow AJ, Barber JA, Hudson GV, et al. 2000. Risk of second malignancy after Hodgkin’s disease in a collaborative British cohort: The relation to age at treatment. J Clin Oncol 18:498–509. Swift M, Morrell D, Massey RB, et al. 1991. Incidence of cancer in 161 families affected by ataxia-telangiectasia. N Engl J Med 325:1831–1836. Talbott EO, Youk AO, McHugh-Pemu KP, Zborowski JV. 2003. Long-term follow-up of the residents of the Three Mile Island accident area: 1979–1998 [published correction appears in Environ Health Perspect 111:A453, 2003]. Environ Health Perspect 111:341–348. Taylor DM, Mays CW, Gerber GB, et al., eds. 1989. Risks from Radium and Thorotrast, BIR Report 21. London: Br Inst Radiol. Telle-Lamberton M, Bergot D, Gagneau M, et al. 2004. Cancer mortality among French Atomic Energy Commission workers. Am J Ind Med 45:34–44. Thompson DE, Mabuchi K, Ron E, et al. 1994. Cancer incidence in atomic bomb survivors. Part II: solid tumors, 1958–1987. Radiat Res 137:S17–S67. Tokarskaya ZB, Okladnikova ND, Belyaeva ZD, Drozhko EG. 1997. Multifactorial analysis of lung cancer dose-response relationships for workers at the Mayak nuclear enterprise. Health Phys 73:899–905. Tomásˇek L, Darby S, Fearn T, et al. 1994. Patterns of lung cancer mortality among uranium miners in West Bohemia with varying rates of exposure to radon and its progeny. Radiat Res 137:251–261. Tomásˇek L, Darby SC, Swerdlow AJ, et al. 1993. Radon exposure and cancer other than lung cancer among uranium miners in West Bohemia. Lancet 341:919–923. Travis LB, Curtis RE, Bennett WP, et al. 1995a. Lung cancer after Hodgkin’s disease. J Natl Cancer Inst 87:1324–1327. Travis LB, Curtis RE, Boice JD Jr, et al. 1991. Second cancers following nonHodgkin’s lymphoma. Cancer 67:2002–2009. Travis LB, Curtis RE, Inskip PD, et al. 1995b. Re: Lung cancer risk and radiation dose among women treated for breast cancer. J Natl Cancer Inst 87:60–61. Travis LB, Curtis RE, Stovall M, et al. 1994. Risk of leukemia following treatment for non-Hodgkin’s lymphoma. J Natl Cancer Inst 86:1450–1457. Travis LB, Gospodarowicz M, Curtis RE, et al. 2002. Lung cancer following chemotherapy and radiotherapy for Hodgkin’s disease. J Natl Cancer Inst 94:182–192. Travis LB, Hauptmann M, Gaul LK, et al. 2003. Site-specific cancer incidence and mortality after cerebral angiography with radioactive Thorotrast. Radiat Res 160:691–706. Travis LB, Hauptmann M, Gaul LK, et al. 2003a. Site-specific cancer incidence and mortality after cerebral angiography with radioactive thorotrast. Radiat Res 160:691–706. Travis LB, Hill DA, Dores GM, et al. 2003b. Breast cancer following radiotherapy and chemotherapy among young women treated for Hodgkin’s disease. JAMA 290:465–475. Travis LB, Hill D, Dores GM, et al. 2005. Cumulative absolute breast cancer risk for young women treated for Hodgkin lymphoma. J Natl Cancer Inst 97:1428–1437. Travis LB, Land CE, Andersson M, et al. 2001. Mortality after cerebral angiography with and without radioactive Thorotrast: An international cohort of 3,143 two-year survivors. Radiat Res 156:136–150. Travis LB, Weeks J, Curtis RE, et al. 1996. Leukemia following low-dose total body irradiation and chemotherapy for non-Hodgkin’s lymphoma. J Clin Oncol 14:565–571. Tubiana M. 1998. The report of the French Academy of Science: “Problems associated with the effects of low doses of ionising radiation”. J Radiol Prot 18:243–248. Tubiana M. 2005. Dose-effect relationship and estimation of the carcinogenic effects of low doses of ionizing radiation: the joint report of the Academie des Sciences (Paris) and of the Academie Nationale de Medecine. Int J Radiat Oncol Biol Phys 63:317–319. Tucker MA, Coleman CN, Cox RS, et al. 1988. Risk of second cancers after treatment for Hodgkin’s disease. N Engl J Med 318:76–81. Tucker MA, D’Angio GJ, Boice JD Jr, et al. 1987a. Bone sarcomas linked to radiotherapy and chemotherapy in children. N Engl J Med 317:588–593. Tucker MA, Meadows AT, Boice JD Jr, et al. 1984. Cancer risk following treatment of childhood cancer. In: Boice JD Jr, Fraumeni JF Jr eds. Radiation Carcinogenesis: Epidemiology and Biological Significance. New York: Raven Press, pp. 211–224. Tucker MA, Meadows AT, Boice JD Jr, et al. 1987b. Leukemia after therapy with alkylating agents for childhood cancer. J Natl Cancer Inst 78: 459–464. Tucker MA, Morris Jones PH, Boice JD Jr, et al. 1991. Therapeutic radiation at a young age is linked to secondary thyroid cancer. Cancer Res 51:2885–2888. UK Childhood Cancer Study Investigators. 2002a. The United Kingdom Childhood Cancer Study of exposure to domestic sources of ionising radiation: 1: radon gas. Br J Cancer 86:1721–1726.

Ionizing Radiation UK Childhood Cancer Study Investigators. 2002b. The United Kingdom Childhood Cancer Study of exposure to domestic sources of ionising radiation: 2: gamma radiation. Br J Cancer 86:1727–1731. Ullrich RL, Jernigan MC, Satterfield LC, et al. 1987. Radiation carcinogenesis: time-dose relationships. Radiat Res 111:179–184. UNSCEAR (United Nations Scientific Committee on the Effects of Atomic Radiation). 1972. Ionizing Radiation Levels and Effects. vol 2: Effects. Publ E.72.IX.18. New York; United Nations. UNSCEAR (United Nations Scientific Committee on the Effects of Atomic Radiation). 1982. Ionizing Radiation: Sources and Biological Effects. Publ E.82.IX.8. New York: United Nations. UNSCEAR (United Nations Scientific Committee on the Effects of Atomic Radiation). 1986. Genetic and Somatic Effects of Ionizing Radiation. Publ E.86.IX.9. New York: United Nations. UNSCEAR (United Nations Scientific Committee on the Effects of Atomic Radiation). 1988. Sources, Effects, and Risks of Ionizing Radiation. Publ E.88.IX.7. New York: United Nations. UNSCEAR (United Nations Scientific Committee on the Effects of Atomic Radiation). 1993. Sources and Effects of Ionizing Radiation. Publ E.94.IX.2. New York: United Nations. UNSCEAR (United Nations Scientific Committee on the Effects of Atomic Radiation). 1994. Sources and Effects of Ionizing Radiation. E.94.IX.11. New York: United Nations. UNSCEAR (United Nations Scientific Committee on the Effects of Atomic Radiation). 2000. UNSCEAR 2000 Report to the General Assembly, with Scientific Annexes. Sources and Effects of Ionizing Radiation. E.00.IX.4. New York, United Nations. UNSCEAR (United Nations Scientific Committee on the Effects of Atomic Radiation). 2001. United Nations Scientific Committee on the Effects of Atomic Radiation. Report to the General Assembly, with Scientific Annexes. Hereditary Effects of Radiation. New York: United Nations. Upton AC, Randolph ML, Conklin JW, et al. 1970. Late effects of fast neutrons and gamma-rays in mice as influenced by the dose rate of irradiation: induction of neoplasia. Radiat Res 41:467–491. Upton AC, Albert RE, Burns FJ, et al., eds. 1986. Radiation Carcinogenesis. New York: Elsevier, pp. 1–459. Upton AC. 1981. Health impact of the Three Mile Island accident. Ann NY Acad Sci 365:63–75. Upton AC. 1987. Biological basis of radiological protection and its application to risk assessment. Br J Radiol 60:1–16. Upton AC. 2003. The state of the art in the 1990’s: NCRP Report No. 136 on the scientific bases for linearity in the dose-response relationship for ionizing radiation. Health Phys 85:15–22. Urquhart JD, Black RJ, Muirhead MJ, et al. 1991. Case-control study of leukaemia and non-Hodgkin’s lymphoma in children in Caithness near the Dounreay nuclear installation. Br Med J 302:687–692. van Kaick G, Dalheimer A, Hornik S, et al. 1999. The german thorotrast study: Recent results and assessment of risks. Radiat Res 152(6 suppl):S64–S71. van Kaick G, Karaoglou A, Kellerer AM, eds. 1995. Health effects of Internally Deposited Radionuclides: Emphasis on Radium and Thorium. Singapore: World Scientific. Van Kaick G, Wesch H, Lührs H, et al. 1989. The German Thorotrast Study— report on 20 years follow-up. In: Taylor DM, Mays CW, Gerber GB, Thomas RG, eds. Risks from Radium and Thorotrast. BIR Report 21. London: Br Inst Radiol, pp. 98–104. van Leeuwen FE, Klokman WJ, Stovall M, et al. 1995. Roles of radiotherapy and smoking in lung cancer following Hodgkin’s disease. J Natl Cancer Inst 87:1530–1537. van Leeuwen FE, Klokman WJ, Stovall M, et al. 2003. Roles of radiation dose, chemotherapy, and hormonal factors in breast cancer following Hodgkin’s disease. J Natl Cancer Inst 95:971–980. van Leeuwen FE, Somers R, Taal BG, et al. 1989. Increased risk of lung cancer, non-Hodgkin’s lymphoma, and leukemia following Hodgkin’s disease. J Clin Oncol 7:1046–1058. van Leeuwen FE, Travis LB. 2001. Adverse effects of treatment: second cancers. In: DeVita VT, Hellman S, Rosenberg SA, eds. Cancer—Principles & Practices of Oncology. 6th ed. Philadelphia: Lippincott, Williams & Wilkins, pp. 2939–2964. Viel JF, Pobel D, Carre A. 1995. Incidence of leukaemia in young people around the La Hague nuclear waste reprocessing plant: A sensitivity analysis. Stat Med 14:2459–2472. Voelz GL, Lawrence JN, Johnson ER. 1997. Fifty years of plutonium exposure to the Manhattan Project plutonium workers: An update. Health Phys 73:611–619. Wagoner JK. 1984. Leukemia and other malignancies following radiation therapy for gynecological disorders. In: Boice JD Jr, Fraumeni JF Jr, eds. Radiation Carcinogenesis: Epidemiology and Biological Significance. New York: Raven Press, pp. 153–159.

293

Wakeford R. 2000. Study of childhood cancer and paternal preconceptional irradiation at USA nuclear facilities. J Radiol Prot 20:331–332. Wakeford R. 2002. Cancer in offspring after paternal preconceptional irradiation. J Radiol Prot 22:191–194. Wakeford R. 2003. Childhood leukaemia and radiation exposure of fathers— the end of the road, perhaps? J Radiol Prot 23:359–362. Wakeford R. 2005. Cancer risk among nuclear workers. J Radiol Prot 25:225–228. Walter AW, Hancock ML, Pui CH, et al. 1998. Secondary brain tumors in children treated for acute lymphoblastic leukemia at St. Jude Children’s Research Hospital. J Clin Oncol 16:3761–3767. Wang JX, Inskip PD, Boice JD Jr, et al. 1990a. Cancer incidence among medical diagnostic X-ray workers in China, 1950 to 1985. Int J Cancer 45:889–895. Wang JX, Zhang LA, Li BX, et al. 2002a. Cancer incidence and risk estimation among medical X-ray workers in China, 1950–1995. Health Phys 82:455–466. Wang Z, Lubin JH, Wang L, et al. 2002b. Residential radon and lung cancer risk in a high-exposure area of Gansu Province, China. Am J Epidemiol 155:554–564. Wang ZY, Boice JD Jr, Wei LX, et al. 1990b. Thyroid nodularity and chromosome aberrations among women in areas of high background radiation in China. J Natl Cancer Inst 82:478–485. Waxweiler RJ, Roscoe RJ, Archer VE, et al. 1981. Mortality follow-up through 1977 of the white underground uranium miners cohort examined by the United States Public Health Service. In: Gomez M, ed. Radiation Hazards in Mining. New York: American Institute of Mining metallurgical and Petroleum Engineering, Inc, pp. 823–830. Wei L, Sugahara T, eds. 2000. High Background Radiation Area in China. J Radiat Res (Tokyo) 41(suppl):1–76. Wei L. 1980. Health survey in high background radiation areas in China. Science 209:877–880. Weiss HA, Darby SC, Doll R. 1994. Cancer mortality following X-ray treatment for ankylosing spondylitis. Int J Cancer 59:327–338. Weiss HA, Darby SC, Fearn T, et al. 1995. Leukemia mortality after X-ray treatment for ankylosing spondylitis. Radiat Res 142:1–11. Wick RR, Nekolla EA, Gossner W, Kellerer AM. 1999. Late effects in ankylosing spondylitis patients treated with 224Ra. Radiat Res 152(6 suppl):S8–S11. Wilkinson GS, Tietjen GL, Wiggs LD, et al. 1987. Mortality among plutonium and other radiation workers at a plutonium weapons facility. Am J Epidemiol 125:231–250. Williams D. 1994. Chernobyl, eight years on. Nature 371:556. Williams D. 2002. Cancer after nuclear fallout: Lessons from the Chernobyl accident. Nat Rev Cancer 2:543–549. Wing S, Richardson D, Armstrong D, Crawford-Brown D. 1997. A reevaluation of cancer incidence near the Three Mile Island nuclear plant: The collision of evidence and assumptions. Environ Health Perspect 105: 52–57. Wing S, Richardson D, Wolf S, et al. 2000. A case control study of multiple myeloma at four nuclear facilities. Ann Epidemiol 10: 144–153. Wing S, Shy CM, Wood JL, et al. 1991. Mortality among workers at Oak Ridge National Laboratory. Evidence of radiation effects in follow-up through 1984. JAMA 265:1397–1402. Wong FL, Boice JD Jr, Abramson DH, et al. 1997. Cancer incidence after retinoblastoma. Radiation dose and sarcoma risk. JAMA 278:1262– 1267. Wood JW, Tamagaki H, Neriishi S, et al. 1969. Thyroid carcinoma in atomic bomb survivors Hiroshima and Nagasaki. Am J Epidemiol 89: 4–14. Yamamoto T, Kopecky KJ, Fujikura T, et al. 1987. Lung cancer incidence among Japanese A-bomb survivors, 1950–80. J Radiat Res 28:156–171. Yeh H, Matanoski GM, Wang Ny, et al. 2001. Cancer incidence after childhood nasopharyngeal radium irradiation: a follow-up study in Washington County, Maryland. Am J Epidemiol 153:749–756. Yiin JH, Schubauer-Berigan MK, Silver SR, et al. 2005. Risk of lung cancer and leukemia from exposure to ionizing radiation and potential confounders among workers at the Portsmouth Naval Shipyard. Radiat Res 163:603–613. Yoshimoto Y, Delongchamp R, Mabuchi K. 1994. In-utero exposed atomic bomb survivors: Cancer risk update. Lancet 344:345–346. Yuasa H, Hamajima N, Ueda R, et al. 1997. Case-control study of leukemia and diagnostic radiation exposure. Int J Hematol 65: 251–261. Zeeb H, Blettner M, Langner I, et al. 2003. Mortality from cancer and other causes among airline cabin attendants in Europe: a collaborative cohort study in eight countries. Am J Epidemiol 158:35–46. Zielinski JM, Krewski D, Ashmore JP, Cardis E. 1997. The use of national registers of radiation exposure in occupational radiation risk assessment. Eur J Cancer 33 (suppl 3):S3–S6.

16

Solar Radiation ADELE C. GREEN AND DAVID C. WHITEMAN

S

olar ultraviolet radiation (UVR), a ubiquitous environmental carcinogen, is part of the spectrum of electromagnetic radiation emanating from the sun. UVR is also generated by artificial sources encountered in a wide range of settings. The ultraviolet spectral region spans wavelengths 10 nm–400 nm, and is divided into several bands (Table 16–1). Although UVR is classified as non-ionizing, UV photons are sufficiently energetic to destabilize electron configurations within molecules such as DNA and have a biological effect. In terms of solar UVR, only UVB and UVA have biological significance, because shorter wavelengths are absorbed by the atmosphere. Ultraviolet radiation is typically absorbed over a surface and is measured as a radiant exposure. The term “irradiation” represents the dose of radiant energy delivered to an area within a given time, and is measured in joules (J), J/m2 (Josefsson, 1993). The rate at which UV energy reaches a surface is termed “irradiance” measured in watts (W), W/m2. The total irradiance from any given source of UVR is derived by summing the wavelength-specific irradiances across the spectrum of wavelengths emitted. A “biologically effective UVR irradiance” (UVReff) for a given source is determined by weighting the irradiance at each emitted wavelength by its ability to cause the biological effect of interest (e.g., DNA mutation, erythema), and then summing these weighted values across all wavelengths.

METHODS OF MEASUREMENT Ultraviolet radiation can be measured at the earth’s surface by a variety of detector instruments, depending upon the research questions being asked. Instruments can be categorized as broadband or narrowband radiometers, or spectroradiometers, each having different requirements for calibration and maintenance. Broadband radiometers measure UVR irradiance over a broad spectral band, integrated over the wavelengths of radiation known to exert a particular biological effect (e.g., erythema). The output from broadband radiometers is a single number, calculated from the voltage generated by filtered UVR striking a photodiode. These simple instruments are portable and robust although relatively insensitive to changes in irradiance at specific wavelengths. Networks of radiometers were first established during the 1970s and 1980s, often in remote locations, and subsequently have provided long-term information about levels of biologically effective UVR at the earth’s surface. Popular early models (e.g., Robertson-Berger meters) have been largely replaced by newer models that have less technical error and are more stable under extremes of temperature, although some limitations remain (Huber et al., 2002). Narrowband radiometers also integrate UVR flux across wavebands, but with finer resolution and are able to measure UVR across bandwidths as narrow as 2 nm. Spectroradiometers are the most sophisticated UVR detection instruments, able to detect small changes in UV flux at specific wavelengths, such as occurs following depletions in atmospheric ozone from time to time (Roy et al., 1997). Notwithstanding their cost and the skill required to operate them, networks of spectroradiometers have been established worldwide to monitor changes in surface irradiance. For example, the United States Environmental Protection Agency operates a network of 14 spectroradiometers located in US National Parks in a broad range of latitudes and altitudes (Table 16–2) as part of the Park Research

294

and Intensive Monitoring of Ecosystems Network (PRIMENet) (Fig. 16–1). In addition, spectroradiometers have been placed in seven urban areas across the United States (Boston MA; Boulder, CO; Gaithersburg, MD; Research Triangle Park, NC; Atlanta, GA; Riverside CA; Albuquerque, NM) to measure ambient UVR to which humans are exposed. Similar networks of UV spectroradiometers have been established in other countries, contributing to a global network coordinated by The World Meteorological Organization (WMO) to monitor UV irradiance.

DETERMINANTS OF SOLAR IRRADIANCE AND SOLAR DOSE The irradiance at the surface of the earth is much less than in the upper atmosphere due to scattering and absorption by atmospheric constituents. Solar UVR is scattered by dust, aerosols, and molecules of air. Ozone (O3) in the stratosphere is continuously generated and dissociated and absorbs all UVC and a proportion of UVB from solar UVR before reaching the earth. The thickness of the ozone mantle varies according to season, latitude, and meteorological conditions. Since 1974 it has been recognized (Rowland and Molina, 1974) that chlorofluorocarbons and other gases generated by human activities disturb the natural balance of generation and destruction of ozone leading to overall depletion of the ozone layer and an increase of effective UVB in the biosphere. Naturally occurring ozone in the outer atmosphere is distinct from ozone occurring at ground level as a product of industry, with adverse effects on the human respiratory system. Stratospheric ozone levels have been depleted by industrial pollutants at the same time as ground level ozone levels have risen, though the effect of the latter on attenuating UVC is negligible. Natural phenomena also affect atmospheric ozone levels, as seen after the Mt. Pinatubo volcano in the Philippines erupted in 1991, releasing vast quantities of dust particles that added to the breakdown of ozone (Ryan et al., 1996). Despite international agreements to reduce the emission of chlorofluorocarbons and halt the depletion of ozone, mathematical models suggest that UVR levels are likely to continue rising well into the twenty-first century (Slaper et al., 1996). Other factors that directly determine UV irradiance (WMO, 1998), especially UVB, at the surface of the earth include: (1) latitude: UVR decreases with increasing distance from the equator, (2) altitude: UVR increases with increasing height above sea level, and (3) season: UVR levels are higher in summer than winter and this is more pronounced with increasing latitude (Fig. 16–1). Time of day and cloud cover are also determinants; 20%–30% of daily UVR is received between 11:00 and 13:00 hours when the solar angle is near the zenith (IARC, 1992), and heavy clouds attenuate UV ground irradiance by up to 90% (Roy et al., 1997), though light clouds over a blue sky hardly alter sunburn effectiveness (IARC, 1992). Air pollution particles also reduce UVR. UVA levels are less affected than UVB by atmospheric factors, and thus UVA irradiance is more stable throughout the day. At ground level, surface reflectance can augment a person’s total UVR exposure. The albedo is the fraction of solar radiation reflected by the ground, ground cover, and bodies of water. Surfaces with a high albedo include snow, dry sand, foaming surf, and pale concrete.

295

Solar Radiation Table 16–1. The Ultraviolet Radiation Spectrum

8000

Wavelength 10 nm– 120 nm– 200 nm– 280 nm– 320 nm–400 nm

UV, ultraviolet.

The Ultraviolet Index The UV Index was developed in 1995 by the WMO, World Health Organization (WHO), the United Nations Environment Program (UNEP), and the International Commission on Non-Ionizing Radiation Protection (ICNIRP) to standardize the way in which the public was advised about levels of biologically effective UVR in a given locality. The UV Index can be either a forecast value (based on computer modeling and expected weather conditions) or an actual measurement. In the United States, the predicted daily UV Index is calculated from a computer model that incorporates information about ozone concentration (based on current satellite measurements), the day of the year, latitude, altitude, and predicted cloud cover. These calculations permit an estimate of the UVR flux at solar noon for each wavelength in the UV spectrum (280 nm–400 nm). The estimates are then weighted according to their effectiveness at inducing erythema (McKinlay-Diffey erythema action spectrum) and summed across the spectrum to produce a single value. The UV Index is this figure divided by a conversion factor and rounded to the nearest whole number. One UV Index unit has the same value around the world (25 mW/m2 of UVR), allowing straightforward comparison of international values. Mid-summer values for the UV Index for North American cities are typically greater than 10.

Daily erythemal UV (J/m^2/day)

Extreme UV Far UV UV-C UV-B UV-A

5000

4000

3000

2000

0 0

25

50

75 100 125 150 175 200 225 250 275 300 325 350

Day of year 8000

Denali

7000

6000

5000

4000

3000

2000

1000

0

Humans are exposed to sunlight on a daily basis, yet accurate and reproducible methods for measuring the dose of solar radiation received by an individual have only been developed in the past few decades. In epidemiologic studies, human sun exposure has been measured in a variety of ways, ranging from dosimetry to personal recall of past exposure.

Table 16–2. Location of PRIMENet Radiometers in US National Parks

Virgin Islands, VI Hawaii Volcanoes, HI Everglades, FL Big Bend, TX Great Smoky Mountains, TN Sequoia/Kings Canyon, CA Shenandoah, VA Canyonlands, UT Rocky Mountain, CO Acadia, ME Theodore Roosevelt, ND Olympic, WA Glacier, MT Denali, AK

6000

1000

MEASURING PERSONAL EXPOSURE TO SOLAR RADIATION

Name of National Park (State)

Chicago

7000

Daily erythemal UV (J/m^2/day)

Ultraviolet Band

Latitude

Longitude

Elevation (m)

Date Commenced

18.3 19.4 25.4 29.3 35.6

64.8 155.3 80.7 103.2 83.8

250 1243 2 1052 564

May 1998 Feb 1999 Mar 1997 Feb 1997 Jan 1997

36.5

118.8

610

Aug 1998

38.5 38.5 40.0 44.4 46.9

78.4 109.8 105.5 68.3 103.4

1073 1814 2891 122 870

Mar 1997 Sept 1997 Apr 1998 Mar 1998 Sept 1998

48.1 48.5 63.7

123.4 113.1 149.0

2 968 640

Dec 1997 Sept 1997 Oct 1997

0

25

50

75 100 125 150 175 200 225 250 275 300 325 350

Day of year

A

Figure 16–1. The daily UV irradiation in year 2000 measured at six US sites monitored by the EPA PRIMENet program. These measures represent the daily dose of biologically effective UVR at ground level, integrated over the McKinlay-Diffey erythema action spectrum. The variation in values between sites largely reflects differences in latitude and altitude, whereas the day-to-day variation within sites is due to cloud cover, ozone, and season. These UV data were supplied by the National UV Monitoring Center at the University of Georgia. (Source: Downloaded from http://www.oz.physast.uga.edu on March 25, 2003.)

Dosimetry The two most widely used dosimeters in human research are the Bacillus biological dosimeter and the polysulfone badge. The biological dosimeter comprises Bacillus subtilis spores immobilized on a polyester sheet, which are inactivated in a dose-dependent manner upon exposure to UVR. The unit of measurement is the optical density of re-suspended spores after irradiation, and biological activity is measured by comparing the ratio of the optical density of the exposed spores to that of a control biofilm stored in the dark. Spores are quite stable under room temperature storage conditions (Quintern et al., 1997; Moehrle and Garbe, 2000). Dosimeters of this type have been used in diverse settings, including schools (Munakata et al., 1998), mountains (Moehrle et al., 2003), and the Arctic (Cockell et al., 2001).

296

PART III: THE CAUSES OF CANCER 8000

8000

Hawaii

Riverside

7000

Daily erythemal UV (J/m^2/day)

Daily erythemal UV (J/m^2/day)

7000

6000

5000

4000

3000

2000

6000

5000

4000

3000

2000

1000

1000

0

0 0

25

50

75 100 125 150 175 200 225 250 275 300 325 350

0

25

50

Day of year 8000

Day of year 8000

Albuquerque

7000

Boulder

7000

Daily erythemal UV (J/m^2/day)

Daily erythemal UV (J/m^2/day)

75 100 125 150 175 200 225 250 275 300 325 350

6000

5000

4000

3000

2000

6000

5000

4000

3000

2000

1000

1000

0

0 0

25

50

75 100 125 150 175 200 225 250 275 300 325 350

0

Day of year

B

25

50

75 100 125 150 175 200 225 250 275 300 325 350

Day of year

C

Figure 16–1. (cont.)

Polysulfone is UVR-sensitive polymer that has an action spectrum closely approximating the erythemal response of human skin when formulated as a film 30–45 um thick (Davis et al., 1976). Of particular importance for dosimetry, polysulfone’s optical properties deteriorate in a monotonic fashion with increasing doses of UVR. These dosimeters are especially useful in field studies, being inexpensive, stable, and easy to apply, though they become saturated at relatively modest doses of UVR requiring participants in exposure studies to periodically change badges (Diffey, 1987). Portable data-logging dosimeters have been developed (Diffey and Saunders, 1995) comprising a photodetector sensitive to UVR in a selected range linked to a hand-held computer to record readings. Numerous field studies have deployed dosimeters and dataloggers on humans outdoors (Table 16–3). Within any given population, highest doses of UVR are experienced by outdoor workers who may receive up to 80% of ambient radiation on horizontal skin surfaces (such as the vertex of the head and the top of the shoulders), and 40% on vertical skin surfaces (such as the chest and face) (Diffey et al., 1977; Holman et al., 1983). Very high UV doses have been recorded among outdoor workers in diverse settings including alpine environments (Moehrle et al., 2003) and the Canadian Arctic (Cockell et al., 2001). Outdoor recreational pursuits such as tennis, golf,

cycling, and sailing also result in high levels of exposure, particularly on the shoulders, back, and hands (Herlihy et al., 1994; Moehrle et al., 2000). Several studies have shown that children typically receive doses of 3%–6% of ambient sunlight each day, regardless of geographic location (Diffey and Gies, 1998; Gies et al., 1998). There is some evidence that children in elementary school spend more time outdoors and, on average, receive about twice the daily dose of UVR received by high school children (Diffey et al., 1996). The accumulated data suggest that there should be large differences in average UVR doses received by populations residing in areas with different ambient UVR, though comparative data are scarce. One study comparing relative UVReff doses received by schoolchildren in northern Australia and northern England (Diffey and Gies, 1998) showed that the median daily solar dose received by Australian children was about twice that received by English children. Very few English children received more than three Standard Erythemal Doses (SED) (one SED = 100 J m-2; three SED is enough to induce sunburn in most Caucasian skin types), yet one-third of the Australian children received at least this dose. These data provide evidence that fair-skinned people living in warm, low-latitude environments commonly experience biologically harmful doses of solar radiation.

Table 16–3. Examples of Dosimetry Studies on Human Subjects Dosimeter type

Reference

Country

Subjects

Methods

Children (n = 126) Boy’s baseball team (BB) 14– 18 yrs Boy’s baseball camp 10–16 yrs Girl scout (GS) camp 8–13 yrs

PSB at cheek, forehead, and arm

Children (n = 14) attending summer camp

PSB at cheek and wrist

USA

Welders in a machine shop

PSB at chest, helmet, interior helmet, and protective spectacles

UK

Children aged 9–15 yrs (n = 180)

PSB at shoulder Ambient UV also monitored. Diaries recorded time outdoors

(Diffey and Gies, 1998)

UK/Australia

PSB at shoulder UK: 2 PSB/week QLD: 1 PSB/day Ambient UV also monitored

Polysulphone badges

(Moise et al., 1999)

Australia

UK Children aged 9–10 yrs (n = 90) QLD children 11–12 years (n = 112) Neonates and toddlers (n = 115)

Polysulphone badges

(Gies, 1995)

Australia

PE teachers (n = 16) Groundsmen (n = 11) Nurseryman (n = 1) Lifeguards (n = 8)

PSB at chest and shoulder + diary

Polysulphone badges

(Melville et al., 1991)

USA

Polysulphone badges

(Rosenthal et al., 1990)

USA

Polysulphone badges

(Tenkate and Collins, 1997) (Diffey et al., 1996)

Polysulphone badges

Polysulphone badges

Polysulphone badges

(Gies et al., 1998)

Australia

Polysulphone badges

(Herlihy et al., 1994)

Australia

Polysulphone badges

(Kimlin et al., 1998)

Australia

Polysulphone badges

(Holman et al., 1983)

Australia

Bacillus subtilis spores + data logging dosimeter

(Cockell et al., 2001)

Canadian arctic

Polysulphone badges

(O’Riordan et al., 2000)

Australia

School children aged 12 yrs (n = 112) Adult males (n = 69) Adult females (n = 25)

PSB at chest and shoulder and diary measures

PSB at shoulder on 4 school days and 4 holidays PSB at cheek, back of hand, chest, shoulder, back, front thigh, back calf) for 6 activities

Outdoor workers Indoor workers Schoolchildren (n = not stated) Teacher PE teacher Gardener Roofer Bricklayer (n = 5) Field scientists at 75 N

PSB at shoulder

PSB at head, chest, back, shoulder, forearm, dorsum of hand 0930– 1530

Infants < 12 months (n = 21)

PSB on right wrist of child

Biofilm badges (left-side collar) and datalogger used by scientists while undertaking activities in field camp

Findings Mean dose (% of PAE) BB team arm BB team cheek BB team forehead BB camp arm BB camp cheek BB camp forehead GS camp arm GS camp cheek GS camp forehead Mean dose (% of PAE) Wrist Cheek Wrist-cheek ratio Interior helmet dose Headband

49.2 21.2 3.8 33.2 19.8 2.3 17.0 9.5 9.5 13.1 8.5 1.5 9 mJ/cm2 / 8 hr = 3 MPE 13 mJ/cm2 / 8 hr = 4.3 MPE

Median (% of TAE) Weekday Elementary boys 6.9 Elementary girls 6.4 High school boys 3.7 High school girls 3.7 Median doses English children 0.9 SED Queensland children 1.7 SED Both groups ~5% of TAE Dose range Infants Toddlers Ambient UVReff dose Mean dose PE teachers Mean shoulder Mean chest Groundsmen Mean shoulder Mean chest Lifeguards Mean shoulder Mean chest Median dose Boys Girls

Weekend 6.9 5.5 5.4 4.6

(% of TAE) 0.5–1.2 2.0–3.1 3454–6507 J/m2 (% of TAE) 19.2% (SD 9) 8.5% (SD 4) 18% (SD 11) 8.2% (SD 6) 12% (SD 10) 10% (SD 10) (% of TAE) 7.7%–8.6% of TAE 4.3%–6.3% of TAE

Median shoulder measures (% of PAE) Tennis 58% (SD 29) Swimming 74% (SD 49) Sailing 59% (SD 20) Golf 65% (SD 18) Walking 17% (SD 8) Gardening 23% (SD 11) Two-day total UVeff (summer) Outdoor 4.5 (SD 1.0) MED Indoor 2.0 (SD 1.0) MED Schoolchildren 2.0 (SD 1.0) MED Mean dose (% of TAE) Teacher 11% (SD 7) PE teacher 53% (SD 13) Gardener 70% (SD 7) Roofer 67% (SD 10) Bricklayer 66% (SD 10) Mean dose (% of TAE) Clear days 27% (SD 9) Cloudy days 12% (SD 24) Mean ambient dose (SED) Clear days 4.54 (SD 1.27) SED Cloudy days 1.9 (SD 0.48) SED Median dose (% of TAE) Infants 1.1% (SD 3.8)

(continued)

297

298

PART III: THE CAUSES OF CANCER

Table 16–3. (cont.) Dosimeter type Bacillus subtilis spores

Bacillus subtilis spores

Reference (Thieden et al., 2001)

(Moehrle et al., 2003)

Country Denmark

Germany Alpine environments

Subjects Indoor workers (n = 44)

Mountain guides (n = 9)

Methods

Findings

Biofilm badges + diary taken during a mean working period 17 days (4–23 days) and a mean holiday period 17 days (8–26 days)

Biofilm badges (attached in vertical plane to cap) worn every day for 1 year for all activities

Work Doses Wrist

0.7 (SE 0.1) SED/day 3.6% (SE 0.4) (Copenhagen)

Holiday Doses (Denmark) Wrist 2.1 (SE 0.3) SED/day 11.1% (SE 1.4) (Copenhagen) Holiday Doses (Southern Europe) Wrist 4.9 (SE 0.6) SED/day 44.6% (SE 8.3) (Copenhagen) Median annual exposure 1273 (312–1770) SED Median daily exposure 5.7 (0.6–24.2) SED

PSB, polysulfone badge; PAE, “personal ambient exposure”, defined as the ambient ultraviolet exposure received by a PSB on a horizontal surface during the time a subject was out of doors and exposed to sunlight; MPE, American Conference of Governmental Industrial Hygienists (ACGIH) Maximum Permissible Exposure limit for UV radiation = 30 J/m2 per 8-hr shift using the ACGIH action spectrum; TAE, “total ambient exposure”, defined as the total ambient ultraviolet exposure received by a PSB on a horizontal surface during a full day; SED, Standard erythemal dose using the CIE action spectrum = 100 J/m2 (CIE, 1997); SD, standard deviation; MED, Minimum erythemal dose using the CIE action spectrum = 200 J/m2 (Diffey, 1996); SE, standard error of the mean.

Personal Recollection Because “dosimetry” methods for measuring solar exposure are prospective they cannot be used in case-control studies, which rely on measures of past exposure. Accurate recall of past sun exposure is difficult; nevertheless epidemiologists have attempted to estimate past sun exposure in two main ways (Whiteman et al., 2001). The first approach has been to estimate an individual’s average exposures to solar UVR during defined time periods according to their places of residence, based on the fact that ambient solar UVR increases with proximity to the equator (Elwood and Diffey, 1993). Epidemiologists have commonly used these ambient exposure measures when conducting ecological studies, such as comparing the incidence of melanoma among native residents and migrants with contrasting expe-

riences of sun exposure (Iscovich and Howe, 1998). Although place of residence is a relatively crude measure of personal sun exposure, it is a reasonable proxy for solar dose when comparing populations (Diffey et al., 1996; Diffey and Gies, 1998) and has the practical advantage of being easy to recall. The other approach has been to ask study participants to recall salient exposures such as numbers of sunburns, time spent outdoors in summer, or recreational activities associated with high levels of sun exposure (Berwick and Chen, 1995; English et al., 1998). These measures have been used widely in epidemiologic studies of skin cancer, despite the potential for recall bias (Weinstock et al., 1991), their modest reliability (Berwick and Chen, 1995; Whiteman and Green, 1997; English et al., 1998), and indeterminable external validity (Table 16–4).

Table 16–4. Studies Testing Reliability of Epidemiologic Measures of Historical Sun Exposure Reference

Country

Participants

Methods

Statistic

(Weinstock et al., 1991)

USA

355 female participants (121 melanoma cases + 234 controls)

Nested case-control study First questionnaire (prediagnosis) 1982 Second questionnaire (postdiagnosis) 1984–1986

Spearman’s r

(Berwick and Chen, 1995)

USA

100 participants (50 melanoma cases + 50 controls)

Case-control study First questionnaire 1987– 1989 Re-test 1990

Kappa

(Whiteman and Green, 1997)

Australia

Australia

Case-control study Parents and children separately questioned about phenotype and sun exposure 1994 Case-control study First questionnaire 1988 Re-test 1993

Weighted kappa

(English et al., 1998)

204 children + 204 parents (51 melanoma cases + 153 controls) 190 participants (75 BCC cases + 115 controls)

Intraclass correlation coefficient

Weighted kappa

Results Correlation of tanning ability Q1–Q2 Prevalent cases 0.78 (0.68–0.85) Incident cases 0.59 (0.31–0.78) Controls 0.76 (0.70–0.81) Mean change in tanning ability score Q1–Q2 Prevalent cases 0.03 (0.06) Incident cases -0.24 (0.12) Controls 0.06 (0.04) Ever burned 0.37 Ever freckled 0.57 No. of burns 0.69 First burn 0.34 Last burn 0.42 Tanning ability 0.41 Eye color 0.88 Peeling sunburn 0.24 Blistering sunburn 0.35 Time spent outdoors overall Aged 8–14 yr Aged 15–19 yr Aged 20–24 yr Aged 25–34 yr Aged 35–39 yr Painful sunburn Blistering sunburn Vacation sun exposure

0.77 0.55 (0.43–0.65) 0.77 (0.70–0.83) 0.73 (0.65–0.79) 0.74 (0.66–0.80) 0.73 (0.65–0.79) 0.53 (0.41–0.66) 0.54 (0.37–0.70) 0.30 (0.19–0.40)

Solar Radiation

Biological Markers Because of the difficulties in measuring past sun exposure by questionnaire-based methods, and because dosimetry is limited to prospective or cross-sectional studies, researchers have sought “objective” measures of historical sun exposure based upon biological changes that correlate with exposure (Green and Battistutta, 1990; English et al., 1998). An episode of sunburn can be seen as an integrated measure of acute solar skin damage occurring in inadequately pigmented or protected skin (Green et al., 1985) and thus recall of the number of sunburns experienced serves as an indicator of accumulated acute exposures to intense solar UVR. Photoageing encompasses a range of visible and microscopic changes to the skin associated with chronic exposure to UVR. Dilated small blood vessels or telangiectasia and solar comedones are seen as well as loss of visible surface architecture, which can be quantified by silicone-rubber skin casts (Fritschi and Green, 1995). This rapid, painless, non-invasive method for measuring skin changes has been used in several field studies (Green, 1991; Fritschi et al., 1995; English et al., 1998) though it has only modest correlation with self-reported time outdoors (English et al., 1998) or histological signs of solar elastosis (Fritschi et al., 1995), namely large amounts of thickened elastic fibers (Kligman, 1986). Mutations in genomic DNA have been proposed as biomarkers for UV exposure (Nakazawa et al., 1994) but highly specific repair mechanisms (Ouhtit et al., 1998) confound this method. Mitochondrial DNA is not so efficiently repaired and is thus more likely to be a valid biomarker for UV exposure (Birch-Machin et al., 1998). The latter technique has yielded promising results in a small controlled study (Kawasaki et al., 2000), but must be validated in larger studies before it can be accepted as a reliable measure of chronic UVR exposure.

ARTIFICIAL SOURCES OF ULTRAVIOLET RADIATION EXPOSURE While most people’s only ultraviolet exposure comes from sunlight, artificial sources of UVR may contribute a material proportion of total UV exposure for some. Lamps emitting UVR are broadly classifiable as incandescent, typically emitting little UVR, or electrical discharge (NRPB, 1995). The latter include fluorescent lamps, the most common type, able to produce a range of UVR emissions, and high-intensity discharge lamps including high-pressure mercury, mercury, metal halide, and xenon lamps, which are potent sources but emit little UVR when an outer envelope is used. People are exposed (or potentially exposed) to artificial UVR sources in three main settings: occupational, medical, and cosmetic. (Although a widespread source of artificial UVR, fluorescent lighting does not present a hazard on acute or chronic exposure at illumination levels commonly used (Whillock, 1988).) In the occupational setting, uses of UVR lamps are diverse, ranging from curing ink and paint to use for germicidal purposes. Normally shields protect workers from material UVR. Prolonged and close exposure to gas welding and cutting processes can result in high doses of UVR while arc welding emits intense UVR that may be hazardous both to the eyes and skin (NRPB, 1995). Artificial sources of UVA radiation have several common medical applications including the diagnosis of certain skin and dental conditions, which fluoresce under a Wood’s light (a specially adapted fluorescent lamp that emits UVA), and phototherapy and photochemotherapy. PUVA is an established photochemotherapeutic regimen that combines UVA with photosensitizing psoralens and is used to manage skin diseases such as psoriasis, cutaneous T-cell lymphoma (mycosis fungoides), and vitiligo. Sunbeds for dermatologic patients can be fitted with lamps emitting varying proportions of UVB ranging from about 0.7% in conventional UVA lamps to almost 5% in newer lamps, which thus require shorter exposure times to achieve equivalent erythema (Das et al., 2002). Different UV dose regimens will influence the occurrence of long-term carcinogenic side effects. The use of photochemotherapy showed that tanning without sunburn could occur as a side effect of controlled exposure to UVA;

299

so arose the tanning industry (Gange et al., 1985). Although UV exposure from tanning lamps produces the types of DNA damage associated with photocarcinogenesis (Woollons et al., 1997), many surveys across Europe (McGinley et al., 1998; Moseley et al., 1998; Amir et al., 2000; Monfrecola et al., 2000; Kuluncsics et al., 2002; Szepietowski et al., 2002) and in North (Rhainds et al., 1999; Geller et al., 2002) and South America (Chouela et al., 1999), have confirmed the wide use of commercial tanning units especially by women. Moreover, these surveys indicate that exposure to tanning units can result in UVR doses similar to those received from sunbathing at a Mediterranean resort and that sunburn is a commonly experienced ill effect. The risks of using tanning lamps are often not explained to clients but even when known they are often ignored and many tanning establishments use no form of control of cumulative UV exposure to the skin or eyes.

BIOLOGIC MECHANISMS FOR ULTRAVIOLET RADIATION CARCINOGENESIS DNA Damage and DNA Repair The principal mechanism of UVR carcinogenesis at the molecular level is DNA damage. The chemical structure of DNA, particularly the density of pyrimidine bases (thymine (T) and cytosine (C)), renders the molecule a potent absorber of UVB photons (NRPB, 1995). Incident UVB radiation directly damages the integrity of DNA by inducing bulky lesions, cyclobutane pyrimidine dimers (CPDs) and (6–4) pyrimidine dimer photoproducts, between adjacent pyrimidine nucleotide bases that interfere with essential genomic functions such as transcription and replication. In contrast, UVA damages DNA indirectly by producing reactive oxygen species, causing oxidative damage to bases (especially guanine to 8-oxo-7,8 dihydroxyguanine) and single-strand breaks. UVA sensitizers, both endogenous (e.g., bilirubin, porphyrins) and exogenous (e.g., medications, cosmetics, sunscreens) may modulate the action of UVA. Skin cells have evolved highly efficient mechanisms to repair DNA damaged by UVR (NRPB, 1995; de Gruijl et al., 2001; Ravanat et al., 2001; Sarasin and Giglia-Mari, 2002). As a temporary stress response, RNA and DNA synthesis in the cell are halted and the p53 protein stabilized, arresting the cell cycle and enabling repair to take place. Enzymatic nucleotide excision is the main repair pathway through which UVB-induced photoproducts are removed to restore the original DNA sequence. If photoproducts are not repaired before cell division, the usual process of DNA replication by DNA polymerases cannot occur. Other enzymes (translesional polymerases) that bypass the UVinduced lesions can replicate the DNA sequence with lower fidelity, however. This can lead to fixed mutations when the wrong base is incorrectly substituted into the DNA sequence. The importance of DNA repair pathways has been shown in people with xeroderma pigmentosum (XP) who lack enzymes that specifically repair UV-induced lesions (Kraemer et al., 1994). Even within the population at large, there is considerable variation in DNA repair efficiency, prompting investigations of whether people with less efficient DNA repair mechanisms are at higher risk of UVR-associated cancers (Wei et al., 1993; Hall et al., 1994; Dybdahl et al., 1999) with conflicting findings to date. Actively transcribed genes are repaired more quickly than infrequently expressed genes and since the pattern of transcriptional activity differs by cell type, so too will the distribution of mutations.

Mutagenesis and Gene Expression Point mutations at dipyrimidine sites, mainly C to T transitions, and CC to TT tandem transitions, are recognized specifically as “signature mutations” of shortwave UVR. Corruption of signaling pathways needed for normal cell growth and homeostasis is a general feature of human and mouse skin cancers. So far UV signature mutations have been found in only a few genes controlling cell proliferation, the TP53 gene in squamous cell carcinoma (SCC) and basal cell carcinoma (BCC), the Patched (PTCH) gene in BCCs from XP patients (de Gruijl et al., 2001), and CDKN2A in melanoma (Pollock et al., 1995).

300

PART III: THE CAUSES OF CANCER

The p53 protein plays a central role in cell cycle regulation, apoptosis, and DNA repair and when this function is lost through mutation or deletion, cells have a reduced capacity to repair DNA damage, facilitating cell transformation. UVB-specific mutations in the TP53 gene have been detected in about 50% of BCCs and more than 90% of SCCs in humans (Brash et al., 1991; Rady et al., 1992; Campbell et al., 1993). Such mutations are also seen in over 60% of actinic keratoses (Taguchi et al., 1994) and in “normal” skin from sun-exposed sites (Nakazawa et al., 1994), suggesting that TP53 mutation is an early event in the evolution of skin cancers. In SCC, activation of the Ras oncogenic pathway appears to be involved as well. In BCC, disruption of the Hedgehog (Hh) signaling pathway is implicated along with dysfunctional TP53 tumor suppression (de Gruijl et al., 2001). In cutaneous melanoma, activation of the Ras oncogenes as well as inactivation of the INK4a (also known as CDKN2A) locus (Pollock et al., 1995; van-Elsas et al., 1996) involved in transcription and intracellular signaling are implicated. Some corroboration of the role of TP53 mutations comes from mouse skin tumors where mutations have been detected in SCCs, but not papillomas, and all located at dipyrimidine sites (Kress et al., 1992; Kanjilal et al., 1995). Good naturally occurring animal models for human melanoma are lacking although Xiphorus hybrids, opossums, and Angora goats among others have been studied (Setlow et al., 1989; Setlow et al., 1993; Green et al., 1996). They generally confirm a causal role of UVR with varying levels of support for action spectra lying in the UVB range, and involvement of CPDs as the main premutagenic lesion. Besides the effects of UVR on expression of regulatory genes, the cell’s initial response to UVR exposure leads to activation of genes with protective functions such as the glutathione S-transferase (GST) family of genes that code for a group of enzymes that can detoxify the byproducts of oxidative stress. Variation in enzyme activity may partly underlie differential susceptibility to UVR (Kerb et al., 1997; Carless et al., 2002). The production of heme oxygenase and metallothioneins may also protect against UVR-induced oxidative stress (NRPB, 1995; Ablett, 2003). Sunlight is essential for the biosynthesis of vitamin D, which becomes an active steroid hormone with a wide range of downstream effects after coupling with vitamin D receptors (VDR). These effects include down-regulation of proto-oncogenes such as c-myc, c-fos, and c-jun (Studzinski and Moore, 1995). As a corollary, it is postulated that exposure to low levels of UVR may actually increase a person’s risk of some types of cancer through increased expression of proliferative genes. The UVR response is communicated to neighboring epidermal cells by increased secretion of signaling molecules including various interleukins and keratinocyte-mediated growth factors (NRPB, 1995; Imokawa et al., 1996) and to melanocytes by factors like endothelin 1 (Imokawa et al., 1992). The long-term effect of many of these signals is to reduce UVR penetration of the skin by inducing keratinocytes to proliferate, and increasing the number and activity of melanocytes (Staricco and Pinkus, 1957; Mitchell, 1963; Pathak et al., 1965; Quevedo et al., 1965; Stierner et al., 1989). In particular, the human melanocortin 1 receptor gene (MC1R) encodes a receptor for the melanocyte stimulating hormone (MSH) (Chhajlani and Wikberg, 1992), which thereby regulates the quality and quantity of melanin pigments in skin cells. Numerous variants of the MC1R gene have been characterized, for example those strongly associated with red hair color (Valverde et al., 1995; Box et al., 1997; Smith et al., 1998), through the production of red-yellow pheomelanin and as such, are markers for high susceptibility to UVR. Some of these MC1R variants confer an increased risk of melanoma, presumably through reduced photoprotection, although other mechanisms may also be involved (Palmer et al., 2000). MC1R variants have also recently been implicated in the development of prostate cancer, possibly reflecting an interaction with UVR (Luscombe et al., 2001).

Immunosuppression Ultraviolet radiation (UV) interacts with the immune system to suppress immune responses locally in exposed skin as well as systemi-

cally as shown by extensive experimental evidence in rodents. It was shown in the early 1980s that UV inhibits the function of Langerhans cells, the antigen-presenting cells in the skin by inducing T cells with suppressor activity (Elmets et al., 1983). Most of these UV-induced suppressor/regulatory T cells are of the CD4 type and on antigenic stimulation release the immunosuppressive cytokine interleukin-10 (IL-10) (Schwarz, 1999). UV-induced NKT cells also appear to be involved (Moodycliffe et al., 2000). Induction of an immune response at a site remote to the site of UV irradiation attests to the existence of UV-induced systemic immunosuppression mediated by cytokines produced by keratinocytes and entering the circulation (Schwarz and Schwarz, 2002). UV-induced DNA damage is seen as the most important molecular mediator of UV-induced immunosuppression and it has been shown in mice that IL-12, which antagonizes UV-induced immunosuppression, can accelerate the removal of UV-induced DNA lesions, probably by inducing DNA repair (Schwarz et al., 2002). Other experimental studies have shown that UV radiation suppresses not only immune induction in naïve animals but also that UVR, specifically UVA, can also suppress the recall of established immune reactions (Nghiem et al., 2001). The evidence that immune surveillance plays a role in protecting against skin cancer development is indirect, derived mainly from the high incidence of skin cancers in organ transplant patients treated with immunosuppressive drugs, an effect that is amplified in immunosuppressed patients who experience high sun exposure (Bouwes Bavinck et al., 1996).

Interaction with Viruses Exposure to UVR may result in activation of quiescent viral genomes incorporated into the host cell genome following viral infection, especially in immunosuppressed patients. The interaction of UVR with HPV is of interest since cutaneous HPV may be tumorigenic in immunocompetent people (Boxman et al., 2000, 2001) as well as in the immunosuppressed (Harwood et al., 2000). A consensus p53binding motif has been detected (Purdie et al., 1999) in the upstream regulatory region of HPV77, a cutaneous HPV type found in SCCs in immunosuppressed patients (Shamanin et al., 1994). This sequence responds to normal p53 activation by UVR with the stimulation of HPV77 promoter activity. Study of the interaction of other HPV types with UVR shows that UVR-induced release of pro-inflammatory cytokines by keratinocytes can be enhanced by the presence of cutaneous types HPV20 and 27 and these cytokines can increase or decrease the promoter activities of these types (Ruhland and de Villiers, 2001). An interaction between UVR, E6 proteins from various HPV types and a cellular apoptotic protein, Bak, has also been shown (Jackson et al., 2000). Such anti-apoptotic activity may be a common mechanism whereby survival of HPV-associated lesions exposed to UVR is enhanced and UVR damage can persist (Harwood and Proby, 2002).

CANCERS ASSOCIATED WITH ULTRAVIOLET RADIATION EXPOSURE Keratinocyte Cancers There is a large body of observational and experimental evidence that UVR is the principal environmental cause of keratinocyte cancers (basal cell carcinomas and squamous cell carcinomas of the skin) (see Chapter 64). Salient points are that keratinocyte cancers are more common among those living in areas of high solar irradiance, and residents of regions with low solar irradiance develop more lesions when they migrate to areas with higher ambient UVR. Both BCC and SCC of the skin occur almost exclusively on sun-exposed body sites among fair-skinned populations, although there are notable sites of predilection for each. Recently, field trials have demonstrated that regular use of sunscreen reduces the incidence of actinic keratoses (AK) (Thompson et al., 1993) and SCC tumors in humans (Green et al., 1999), while XP patients who are unable to repair UVB-specific DNA mutations have rates of BCC and SCC thousands of times higher than normal

301

Solar Radiation (Kraemer et al., 1994). Animal studies have confirmed the causal role of sunlight in keratinocyte cancer, with early experiments in mice demonstrating that wavelengths in the UVB spectrum are potent skin carcinogens (de Gruijl et al., 1993).

Melanoma Cutaneous melanoma is an aggressive cancer of the pigment-producing cells of the skin. The causal role of UVR exposure is complex and is partly determined by host propensity to develop numerous nevi, and other as yet unidentified factors (Whiteman and Green, 1998) (see Chapter 63). The evidence linking UVR exposure with cutaneous melanoma is similar to that for keratinocyte cancer, namely that melanomas develop more commonly among fair-skinned migrants to higher UVR environments than among the population of origin (Whiteman et al., 2001); and sun-exposed sites such as the face and ears are most commonly affected per unit of skin area, but intermittently exposed sites such as the trunk and proximal limbs are also often affected (Osterlind et al., 1988; Green et al., 1993). UVR mutations are occasionally detected in cutaneous melanomas (van-’t-Veer et al., 1989; Maestro and Boiocchi, 1994; van-Elsas et al., 1996). Mouse models support observational human studies showing that sun exposure in early life has particularly adverse effects (Noonan et al., 2001). Numerous epidemiologic studies have examined the possible association of UVR with ocular melanoma (Vajdic et al., 2002) and it appears that occupational sun exposure, especially farming, is associated with higher risks of melanoma of the choroid and ciliary body (Vajdic et al., 2002).

Other Cancers There has been speculation that exposure to UVR may indirectly affect a person’s risk of developing cancers at sites other than skin (Table 16–5), but few studies have tested the hypothesis directly.

Non-Hodgkin Lymphoma Some studies suggest that the incidence of non-Hodgkin lymphoma (NHL) increases with proximity to the equator (Adami et al., 1995; Freedman et al., 1997; Adami et al., 1999), while others have found no such effect (Newton et al., 1996; Freedman et al., 1997; Newton, 1997). Several large population-based record-linkage studies have inferred sun exposure from occupational records, but have failed to identify any increased risk of NHL among outdoor workers (Freedman et al., 1997; Adami et al., 1999). Cancer registry data consistently show that NHL is more common among people with a prior history of BCC or SCC, suggesting shared causal mechanisms (Hemminki et al., 2003).

Prostate Cancer The observations that prostate cancer is twice as prevalent among US blacks as US whites and that the disease is more common in northern latitudes prompted speculation that exposure to UVR may reduce the risk of developing cancer of the prostate (Schwartz and Hulka, 1990; Hanchette and Schwartz, 1992). An English study found epidemiologic measures of sun exposure were associated with lower risks of developing prostate cancer (Luscombe et al., 2001; Luscombe et al., 2001) and a US study found lower risks of dying from prostate cancer among people living in areas of high ambient UVR (Freedman et al., 2002), but failed to find any consistent associations with measures of occupational exposure.

Colon Cancer Routinely collected data suggest that colon cancer mortality is lowest in areas of the United States receiving the highest levels of sunlight (Garland and Garland, 1980; Freedman et al., 2002), and that mortality is lower among people with a history of outdoor occupations (Freedman et al., 2002). Large, prospective cohort studies have shown that people with colon cancer or colorectal cancer have lower vitamin D levels at baseline compared with cancer-free controls as determined by dietary records (Garland et al., 1985; Martinez et al., 1996) and serum measurement (Garland et al., 1989) but an association between sun exposure and colon cancer is yet to be confirmed.

Breast Cancer Breast cancer mortality among US women is inversely correlated with ambient levels of solar radiation (Garland et al., 1990). At least one cohort study has showed lower risks of breast cancer among women with higher levels of sun exposure across a range of measures (John et al., 1999) and a record-based case-control study has reported similar findings (Freedman et al., 2002).

Ovarian Cancer While ovarian cancer mortality is lower among US women residing in environments of high UV insolation (Lefkowitz and Garland, 1994; Freedman et al., 2002), and there is speculation that UVR may reduce the risk of ovarian cancer through VDR pathways, there is as yet insufficient evidence to draw firm conclusions.

OPPORTUNITIES FOR PREVENTION Many health promotion and health education strategies have been developed to change knowledge, attitudes, or behaviors associated with reducing exposure to the sun.

Table 16–5. Solar Radiation and Internal Cancers

Cancer Type

Postulated Direction of Risk with Increasing UVR Exposure

NHL

Increase

Prostate

Decrease

Colon

Decrease

Breast

Decrease

Ovary

Decrease

NHL, non-Hodgkin lymphoma.

Epidemiologic Findings

Postulated Mechanism

Inverse latitude gradient Associated with SCC and melanoma Latitude gradient More common in blacks Higher risk with low UV exposure Latitude gradient Lower risk with outdoor work Higher risk with low vitamin D

Immunosuppression ?DNA damage

Latitude gradient Lower risk with outdoor work Higher risk with low vitamin D Latitude gradient

Vitamin D

Vitamin D

Vitamin D

Vitamin D

References (Adami et al., 1995; Newton et al., 1996; Newton, 1997; Adami et al., 1999) (Hanchette and Schwartz, 1992; Luscombe et al., 2001; Luscombe et al., 2001; Luscombe et al., 2001) (Garland and Garland, 1980; Garland et al., 1985; Garland et al., 1989; Martinez et al., 1996; Freedman et al., 2002) (Garland et al., 1990; John et al., 1999; Freedman et al., 2002) (Lefkowitz and Garland, 1994; Freedman et al., 2002)

302

PART III: THE CAUSES OF CANCER

Strategies Aimed at General Population Public health campaigns to reduce sun exposure of white-skinned populations have become widespread. Several have been pioneered in Australia, where a generally susceptible population experiences high levels of sun exposure and has extremely high rates of skin cancer. Probably the earliest systematic attempt to raise public awareness about the role of sunlight in the causation of skin cancer was “The Queensland Melanoma Project” (Smith, 1979), although its major aim was to encourage early presentation of suspicious pigmented skin lesions. The “Slip! Slop! Slap!” campaign commenced in Australia in the early 1980s, exhorting people to wear hats, protective clothing, and sunscreen when outdoors in the sun (Montague et al., 2001). In 1988 and 1989 it was broadened into the comprehensive “SunSmart” campaign involving extensive media advertising, educational resources for schools, and structural changes such as providing shade in public places and reducing the costs of sunscreens. Evaluation of the “SunSmart” project by annual telephone surveys during the summers of 1988, 1989, and 1990 showed substantial changes in knowledge and attitude about the health effects of sunlight in the target population (Hill et al., 1993), the largest change being a 10%–20% decrease in prevalence of positive beliefs about tanning. Behavior changed as well: the adjusted odds ratio for sunburn on the weekend before interview was 0.59 (95% CI: 0.43–0.81) in 1990 compared with 1988. Community campaigns designed to reduce sun exposure have also been evaluated in Hawaii (home-delivered educational comic book) (King et al., 1983; Putnam et al., 1985), Texas (daily reporting of UVB radiation levels in television, radio, and print media) (Boutwell, 1995), Scotland (national sun awareness week) (Fleming et al., 1997), England (magazine advertising and in-flight information to reduce sun exposure on vacation) (Cameron and McGuire, 1990), and rural Australia (targeting cancer risk behaviors through community information outlets and the media) (Hancock et al., 1996). Post-exposure evaluations of each of these media campaigns also indicated increases in knowledge and favorable changes in attitude among the target population, but less noticeable behavior changes. In general women are more likely to recall health messages and change their attitudes than men.

Strategies Aimed at Small Groups Evidence for the effectiveness of interventions designed to change individual sun protection knowledge, attitudes, or behaviors comes from controlled trials where an “intervention” group has been compared with a “control” group for the outcomes of interest, and from pre-test/post-test intervention studies. Typically, such studies have taken place in defined settings, such as the workplace (Borland et al., 1991; Girgis et al., 1994), school (Hughes et al., 1993; Buller et al., 1994; Loescher et al., 1995; Reding et al., 1995; Buller et al., 1996), or recreational settings (Olson et al., 1997; Dietrich et al., 1998) with the intention of improving sun protection behaviors while undertaking outdoor activities. Avoidance of exposure in the workplace is often not feasible, so the focus has been to encourage workers to adopt practical protection measures, wearing hats and appropriate clothing and applying sunscreen. For example, an educational package for outdoor telephone linesman in Australia (Borland et al., 1991) consisting of advice from an occupational nurse and reading materials for each worker led to a modest (6%–11%) but significant increase in workers wearing protective clothing. Another randomized controlled trial among outdoor electrical workers in Australia showed that a skin examination by a dermatologist and advice from a health education officer (Girgis et al., 1994) resulted in a significant increase from 50% to 66% of outdoor workers using a high-level of solar protection. Although such results are encouraging, the longer-term sustainability of the behavior changes must be monitored. Because sun exposure during childhood is a potent risk factor for skin cancer, and because behaviors adopted in early life are likely to persist, many studies have examined the efficacy of interventions aimed at children or their parents to reduce children’s sun exposure. For example, a simple educational program aimed at mothers of newborns and consisting of reading materials and

a reminder postcard in summer, significantly reduced the amount of time that infants were exposed to the sun (Bolognia et al., 1992). In a review of programs aimed at children and parents or caregivers and whole communities (Buller and Borland, 1999), it was again found that changes in knowledge and attitude were consistently greater than changes in behavior. Multiple presentations (typically through the school curriculum) have a consistent effect on children’s behavior in the sun while single, short presentations only improve knowledge. Similarly community-wide interventions with repeated presentations about protection have been the most effective and though expensive, should ultimately be the most cost effective. Within the community, teenagers remain a subgroup whose behavior it is difficult to influence despite possible high levels of knowledge about harmful sun exposure. Interventions aimed at adolescents must address social influences that drive the perception of desirability of suntanning (Grant-Petersson et al., 1999). Likewise, messages delivered to younger adults (college students, etc.) may be more effective at changing intended behavior if they emphasize the adverse effects of solar exposure on appearance rather than stressing health risks (Jones and Leary, 1994).

Governmental Regulatory Measures and Public Policy Many countries have enacted national environmental policies to halt emission of noxious substances affecting the ozone layer and reverse the increase in UVR occurring as a result. Setting quantitative limits of sun exposure for the population is clearly not possible, but limits do apply to artificial sources where level of irradiation and duration of involuntary exposure of workers and the public can be controlled (ICNIRP, 1999). Many countries provide information and guidelines through national, state, and local governments. For example, in the Netherlands, United Kingdom, and Australia (Netherlands Health Council, 1994; TCCA, 2001; NRPB, 2002) explicit national objectives exist, which include an increase in knowledge of the Global Solar UV Index generally and of UV protection among health care workers and the safe operation of solaria. Additional goals in Australia include an increase in the amount of natural or constructed shade in public places and improved protection of outdoor workers. The latter are to be achieved by the inclusion of sun protection policies in industrial awards; by support for tax deduction claims by employers/employees for provision/acquisition of sun protection items; and by incorporation of provisions for sun protection practices into occupational health and safety legislation (TCCA, 2001).

FUTURE RESEARCH In the area of UVR measurement, further attention is needed in the development and maintenance of databases linking ground level UVR measurement programs and reproducible biological dosimetry on sentinel populations and available skin cancer and other relevant health data in corresponding populations. This will enable inter alia monitoring trends of human UV exposure and success of environmental strategies and prevention programs. Regarding epidemiologic studies, specific investigations of the role of UVR in causation of internal (non-skin) cancers, and the role of DNA repair and of viral activation in risk of skin cancer should be undertaken. New technologies might be researched to reduce the deleterious effects of sun exposure such as treatments delivered directly onto the skin to specifically repair damaged genes. This is an example of “postprimary” prevention. For example, experimental systems have been developed that carry a UV-related repair enzyme (T4 endonuclease V) into the skin, which double the rate of removal of CPDs from DNA (Yarosh et al., 1992; Yarosh et al., 1999, 2001). Finally, implementation and evaluation programs of behavioral interventions, especially among children and during their careers, are high research priorities.

Solar Radiation References Ablett E, Whiteman DC, Boyle GM, et al. 2003. Induction of metallothionein in human skin by routine exposure to sunlight: evidence for a systemic response and enhanced induction at certain body sites. J Invest Dermatol 120:318–324. Adami J, Frisch M, Yuen J, et al. 1995. Evidence of an association between non-Hodgkin’s lymphoma and skin cancer. Br Med J 310:1491–1495. Adami J, Gridley G, Nyren O, et al. 1999. Sunlight and non-Hodgkin’s lymphoma: A population-based cohort study in Sweden. Int J Cancer 80:641–645. Amir Z, Wright A, Kernohan EE, et al. 2000. Attitudes, beliefs and behaviour regarding the use of sunbeds amongst healthcare workers in Bradford. Eur J Cancer Care (Engl) 9:76–79. Berwick M, Chen Y. 1995. Reliability of reported sunburn history in a casecontrol study of cutaneous malignant melanoma. Am J Epidemiol 141:1033–1037. Birch-Machin MA, Tindall M, Turner R, et al. 1998. Mitochondrial DNA deletions in human skin reflect photo- rather than chronologic aging. J Invest Dermatol 110:149–152. Bolognia JL, Berwick M, Fine JA, et al. 1992. Sun protection in newborns. A comparison of educational methods. Am J Dis Child 145:1125–1129. Borland RM, Hocking B, Godkin GA, et al. 1991. The impact of a skin cancer control education package for outdoor workers. Med J Aust 154:686–688. Boutwell WB. 1995. The Under Cover Skin Cancer Prevention Project. A community-based program in four Texas cities. Cancer 75:657–660. Bouwes Bavinck JN, Hardie DR, Green A, et al. 1996. The risk of skin cancer in renal transplant recipients in Queensland, Australia. A follow-up study. Transplantation 61:715–721. Box NF, Wyeth JR, O’Gorman LE, et al. 1997. Characterization of melanocyte stimulating hormone receptor variant alleles in twins with red hair. Hum Mol Genet 6:1891–1897. Boxman IL, Russell A, Mulder LH, et al. 2000. Case-control study in a subtropical Australian population to assess the relation between nonmelanoma skin cancer and epidermodysplasia verruciformis human papillomavirus DNA in plucked eyebrow hairs. Int J Cancer 86:118–121. Boxman IL, Russell A, Mulder LH, et al. 2001. Association between epidermodysplasia verruciformis-associated human papillomavirus DNA in plucked eyebrow hair and solar keratoses. J Invest Dermatol 117: 1108–1112. Brash D, Rudolph JA, Simon JA, et al. 1991. A role for sunlight in skin cancer: UV-induced p53 mutations in squamous cell carcinoma. Proc Natl Acad Sci U S A 88:10124–10128. Buller DB, Borland R. 1999. Skin cancer prevention for children: A critical review. Health Educ Behav 26:317–343. Buller DB, Buller MK, Beach B, et al. 1996. Sunny days, healthy ways: Evaluation of a skin cancer prevention curriculum for elementary school-aged children. J Am Acad Dermatol 35:911–922. Buller MK, Loescher LJ, Buller DB. 1994. “Sunshine and skin health”: A curriculum for skin cancer prevention education. J Cancer Educ 9: 155–162. Cameron IH, McGuire C. 1990. “Are you dying to get a suntan?”—the preand post-campaign survey results. Health Education Journal 49:166–170. Campbell C, Quinn A, Ro Y, et al. 1993. p53 mutations are common and early evens that precede tumor invasion in squamous cell neoplasia of the skin. J Invest Dermatol 100:746–748. Carless MA, Lea RA, Curran JE, et al. 2002. The GSTM1 null genotype confers an increased risk for solar keratosis development in an Australian Caucasian population. J Invest Dermatol 119:1373–1378. Chhajlani V, Wikberg JE. 1992. Molecular cloning and expression of the human melanocyte stimulating hormone receptor cDNA. FEBS Lett 309: 417–420. Chouela E, Pellerano G, Bessone A, et al. 1999. Sunbed use in Buenos Aires, Argentina. Photodermatol Photoimmunol Photomed 15:100–103. Cockell CS, Scherer K, Horneck G, et al. 2001. Exposure of arctic field scientists to ultraviolet radiation evaluated using personal dosimeters. Photochem Photobiol 74:570–578. Das S, Lloyd JJ, Walshaw D, et al. 2002. Response of psoriasis to sunbed treatment: Comparison of conventional ultraviolet A lamps with new higher ultraviolet B-emitting lamps. Br J Dermatol 147:966–972. Davis A, Deane GH, Diffey BL. 1976. Possible dosimeter for ultraviolet radiation. Nature 261:169–170. de Gruijl FR, Sterenborg JCM, Forbes PD, et al. 1993. Wavelength dependence of skin cancer induction by UV irradiation of albino hairless mice. Cancer Res 53:53–60. de Gruijl FR, van Kranen HJ, Mullenders LH. 2001. UV-induced DNA damage, repair, mutations and oncogenic pathways in skin cancer. J Photochem Photobiol B 63:19–27.

303

Dietrich AJ, Olson AL, Sox CH, et al. 1998. A community-based randomized trial encouraging sun protection for children. Pediatrics 102:E64. Diffey BL. 1987. A comparison of dosimeters used for solar ultraviolet radiometry. Photochem Photobiol 46:55–60. Diffey BL, Gibson CJ, Haylock R, et al. 1996. Outdoor ultraviolet exposure of children and adolescents. Br J Dermatol 134:1030–1034. Diffey BL, Gies HP. 1998. The confounding influence of sun exposure in melanoma. Lancet 351:1101–1102. Diffey BL, Kerwin M, Davis A. 1977. The anatomical distribution of sunlight. Br J Dermatol 97:407–410. Diffey BL, Saunders PJ. 1995. Behavior outdoors and its effects on personal ultraviolet exposure rate measured using an ambulatory datalogging dosimeter. Photochem Photobiol 61:615–618. Dybdahl M, Frentz G, Vogel U, et al. 1999. Low DNA repair is a risk factor in skin carcinogenesis: A study of basal cell carcinoma in psoriasis patients. Mutat Res 433:15–22. Elmets CA, Bergstresser PR, Tigelaar RE, et al. 1983. Analysis of the mechanism of unresponsiveness produced by haptens painted on skin exposed to low dose ultraviolet radiation. J Exp Med 158:781–794. Elwood JM, Diffey BL. 1993. A consideration of ambient solar ultraviolet radiation in the interpretation of studies of the aetiology of melanoma. Melanoma Res 3:113–122. English DR, Armstrong BK, Kricker A. 1998. Reproducibility of reported measurements of sun exposure in a case-control study. Cancer Epidemiol Biomarkers Prev 7:857–863. Fleming C, Newell J, Turner S, et al. 1997. A study of the impact of Sun Awareness Week 1995. Br J Dermatol 136:719–724. Freedman DM, Dosemeci M, McGlynn K. 2002. Sunlight and mortality from breast, ovarian, colon, prostate, and non-melanoma skin cancer: A composite death certificate based case-control study. Occup Environ Med 59:257–262. Freedman DM, Zahm SH, Dosemeci M. 1997. Residential and occupational exposure to sunlight and mortality from non-Hodgkin’s lymphoma: Composite (threefold) case-control study. Br Med J 314:1451–1455. Fritschi L, Battistutta D, Strutton GM, et al. 1995. A non-invasive measure of photoageing. Int J Epidemiol 24:150–154. Fritschi L, Green A. 1995. Sun damage in teenagers’ skin. Aust J Public Health 19:383–386. Gange RW, Blackett AD, Matzinger EA, et al. 1985. Comparative protection efficiency of UVA- and UVB-induced tans against erythema and formation of endonuclease-sensitive sites in DNA by UVB in human skin. J Invest Dermatol 85:362–364. Garland C, Shekelle RB, Barrett-Connor E, et al. 1985. Dietary vitamin D and calcium and risk of colorectal cancer: A 19-year prospective study in men. Lancet 1:307–309. Garland CF, Comstock GW, Garland FC, et al. 1989. Serum 25-hydroxyvitamin D and colon cancer: Eight-year prospective study. Lancet 2:1176–1178. Garland CF, Garland FC. 1980. Do sunlight and vitamin D reduce the likelihood of colon cancer? Int J Epidemiol 9:227–231. Garland FC, Garland CF, Gorham ED, et al. 1990. Geographic variation in breast cancer mortality in the United States: A hypothesis involving exposure to solar radiation. Prev Med 19:614–622. Geller AC, Colditz G, Oliveria S, et al. 2002. Use of sunscreen, sunburning rates, and tanning bed use among more than 10,000 US children and adolescents. Pediatrics 109:1009–1014. Gies P, Roy C, Toomey S, et al. 1998. Solar UVR exposures of primary school children at three locations in Queensland. Photochem Photobiol 68: 78–83. Girgis A, Sanson-Fisher R, Watson A. 1994. A workplace intervention for increasing outdoor workers’ use of solar protection. Am J Public Health 84:77–81. Grant-Petersson J, Dietrich AJ, Sox CH, et al. 1999. Promoting sun protection in elementary schools and child care settings: The SunSafe Project. J Sch Health 69:100–106. Green A, Battistutta D. 1990. Incidence and determinants of skin cancer in a high-risk Australian population. Int J Cancer 46:356–361. Green A, MacLennan R, Youl P, et al. 1993. Site distribution of cutaneous melanoma in Queensland. Int J Cancer 53:232–236. Green A, Neale R, Kelly R, et al. 1996. An animal model for human melanoma. Photochem Photobiol 64:577–580. Green A, Siskind V, Bain C, et al. 1985. Sunburn and malignant melanoma. Br J Cancer 51:393–397. Green A, Williams G, Neale R, et al. 1999. Daily sunscreen application and betacarotene supplementation in prevention of basal-cell and squamouscell carcinomas of the skin: A randomised controlled trial. Lancet 354:723–729. Green AC. 1991. Premature ageing of the skin in a Queensland population. Med J Aust 155:473–474, 477–478.

304

PART III: THE CAUSES OF CANCER

Hall J, English D, Artuso M, et al. 1994. DNA repair capacity as a risk factor for non-melanocytic skin cancer—A molecular epidemiologic study. Int J Cancer 58:179–184. Hanchette CL, Schwartz GG. 1992. Geographic patterns of prostate cancer mortality. Evidence for a protective effect of ultraviolet radiation. Cancer 70:2861–2869. Hancock L, Sanson-Fisher R, Redman S, et al. 1996. Community action for cancer prevention: overview of the Cancer Action in Rural Towns (CART) project, Australia. Health Prom Int 11:277–290. Harwood CA, Proby CM. 2002. Human papillomaviruses and non-melanoma skin cancer. Curr Opin Infect Dis 15:101–114. Harwood CA, Surentheran T, McGregor JM, et al. 2000. Human papillomavirus infection and non-melanoma skin cancer in immunosuppressed and immunocompetent individuals. J Med Virol 61:289–297. Hemminki K, Jiang Y, Steineck G. 2003. Skin cancer and non-Hodgkin’s lymphoma as second malignancies. Markers of impaired immune function? Eur J Cancer 39:223–229. Herlihy E, Gies PH, Roy CR, et al. 1994. Personal dosimetry of solar UV radiation for different outdoor activities. Photochem Photobiol 60:288– 294. Hill D, White V, Marks R, et al. 1993. Changes in sun-related attitudes and behaviours, and reduced sunburn prevalence in a population at high risk of melanoma. Eur J Cancer Prev 2:447–456. Holman CD, Gibson IM, Stephenson M, et al. 1983. Ultraviolet irradiation of human body sites in relation to occupation and outdoor activity: Field studies using personal UVR dosimeters. Clin Exp Dermatol 8:269–277. Huber M, Blumthaler M, Schreder J, et al. 2002. Effect of ambient temperature on Robertson-berger-type erythemal dosimeters. Applied Optics 41:4273–4277. Hughes BR, Altman DG, Newton JA. 1993. Melanoma and skin cancer: Evaluation of a health education programme for secondary schools. Br J Dermatol 128:412–417. IARC. 1992. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. Solar and Ultraviolet Radiation. 55: Lyon, International Agency for Research on Cancer. Lyon: IARC. ICNIRP (International Commission on Non-Ionizing Radiation Protection). 1999. Guidelines on limiting exposure to non-ionizing radiation. Munich: ICNIRP. Imokawa G, Yada Y, Miyagishi M. 1992. Endothelins secreted from human keratinocytes are intrinsic mitogens for human melanocytes. J Biol Chem 267:24675–24680. Imokawa G, Yukihiro Y, Kimura M, et al. 1996. Granulocyte/macrophage colony-stimulating factor is an intrinsic keratinocyte-derived growth factor for human melanocytes in UVA-induced melanosis. Biochem J 313:625–631. Iscovich J, Howe GR. 1998. Cancer incidence patterns (1972–91) among migrants from the Soviet Union to Israel [see comments]. Cancer Causes Control 9:29–36. Jackson S, Harwood C, Thomas M, et al. 2000. Role of Bak in UV-induced apoptosis in skin cancer and abrogation by HPV E6 proteins. Genes Dev 14:3065–3073. John EM, Schwartz GG, Dreon DM, et al. 1999. Vitamin D and breast cancer risk: the NHANES I Epidemiologic follow-up study, 1971–1975 to 1992. National Health and Nutrition Examination Survey. Cancer Epidemiol Biomarkers Prev 8:399–406. Jones JL, Leary MR. 1994. Effects of appearance-based admonitions against sun exposure on tanning intentions in young adults. Health Psychol 13:86–90. Josefsson WAP. 1993. Monitoring Ultraviolet radiation. Environmental UV Photobiology 73–85. Kanjilal S, Strom SS, Clayman GL, et al. 1995. P53 mutations in nonmelanoma skin cancer of the head and neck: Molecular evidence for field cancerization. Cancer Res 55:3604–3609. Kawasaki K, Suzuki T, Ueda M, et al. 2000. CC to TT mutation in the mitochondrial DNA of normal skin: Relationship to ultraviolet light exposure. Mutat Res 468:35–43. Kerb R, Brockmoller J, Reum T, et al. 1997. Deficiency of glutathione S-transferases T1 and M1 as heritable factors of increased cutaneous UV sensitivity. J Invest Dermatol 108:229–232. Kimlin MG, Parisi AV, Wong JC. 1998. Quantification of personal solar UV exposure of outdoor workers, indoor workers and adolescents at two locations in Southeast Queensland. Photodermatol Photoimmunol Photomed 14:7–11. King PH, Murfin GD, Yanagisako KL, et al. 1983. Skin cancer/melanoma knowledge and behavior in Hawaii: Changes during a community-based cancer control program. Prog Clin Biol Res 130:135–144. Kligman LH. 1986. Photoaging: Manifestations, prevention, treatment. Dermatol Clin 4:517–528.

Kraemer KH, Lee MM, Andrews AD, et al. 1994. The role of sunlight and DNA repair in melanoma and nonmelanoma skin cancer. The xeroderma pigmentosum paradigm. Arch Dermatol 130:1018–1021. Kress S, Sutter C, Strickland PT, et al. 1992. Carcinogen-specific mutational pattern in the p53 gene in ultraviolet B radiation induced squamous cell carcinomas of mouse skin. Cancer Res 52:6400–6403. Kuluncsics Z, Kerekgyarto T, Grof P, et al. 2002. Biological UV dosimeters in quality control of tanning tubes. Photochem Photobiol 76:391– 396. Lefkowitz ES, Garland CF. 1994. Sunlight, vitamin D, and ovarian cancer mortality rates in US women. Int J Epidemiol 23:1133–1136. Loescher LJ, Emerson J, Taylor A, et al. 1995. Educating preschoolers about sun safety. Am J Public Health 85:939–943. Luscombe CJ, French ME, Liu S, et al. 2001. Outcome in prostate cancer associations with skin type and polymorphism in pigmentation-related genes. Carcinogenesis 22:1343–1347. Luscombe CJ, French ME, Liu S, et al. 2001. Prostate cancer risk: Associations with ultraviolet radiation, tyrosinase and melanocortin-1 receptor genotypes. Br J Cancer 85:1504–1509. Luscombe CJ, Fryer AA, French ME, et al. 2001. Exposure to ultraviolet radiation: Association with susceptibility and age at presentation with prostate cancer. Lancet 358:641–642. Maestro R, Boiocchi M. 1994. Sunlight and melanoma: An answer from MTS1 p16. Science 267:15–16. Martinez ME, Giovannucci EL, Colditz GA, et al. 1996. Calcium, vitamin D, and the occurrence of colorectal cancer among women. J Natl Cancer Inst 88:1375–1382. McGinley J, Martin CJ, Mackie RM. 1998. Sunbeds in current use in Scotland: A survey of their output and patterns of use. Br J Dermatol 139:428– 438. Melville SK, Rosenthal FS, Luckmann R, et al. 1991. Quantitative ultraviolet skin exposure in children during selected outdoor activities. Photodermatol Photoimmunol Photomed 8:99–104. Mitchell RE. 1963. The effect of prolonged solar radiation on melanocytes of the human epidermis. J Invest Dermatol 41:199–212. Moehrle M, Dennenmoser B, Garbe C. 2003. Continuous long-term monitoring of UV radiation in professional mountain guides reveals extremely high exposure. Int J Cancer 103:775–778. Moehrle M, Garbe C. 2000. Personal UV dosimetry by Bacillus subtilis spore films. Dermatology 200:1–5. Moehrle M, Heinrich L, Schmid A, et al. 2000. Extreme UV exposure of professional cyclists. Dermatology 201:44–45. Moise AF, Gies HP, Harrison SL. 1999. Estimation of the annual solar UVR exposure dose of infants and small children in tropical Queensland, Australia. Photochem Photobiol 69:457–463. Monfrecola G, Fabbrocini G, Posteraro G, et al. 2000. What do young people think about the dangers of sunbathing, skin cancer and sunbeds? A questionnaire survey among Italians. Photodermatol Photoimmunol Photomed 16:15–18. Montague M, Borland R, Sinclair C. 2001. Slip! Slop! Slap! and SunSmart, 1980–2000: skin cancer control and 20 years of population-based campaigning. Health Educ Behav 28:290–305. Moodycliffe AM, Nghiem D, Clydesdale G, et al. 2000. Immune suppression and skin cancer development: Regulation by NKT cells. Nat Immunol 1:521–525. Moseley H, Davidson M, Ferguson J. 1998. A hazard assessment of artificial tanning units. Photodermatol Photoimmunol Photomed 14:79–87. Munakata N, Ono M, Watanabe S. 1998. Monitoring of solar-UV exposure among schoolchildren in five Japanese cities using spore dosimeter and UV-coloring labels. Jpn J Cancer Res 89:235–245. Nakazawa H, English D, Randell P, et al. 1994. UV and skin cancer: Specific p53 gene mutation in normal skin as a biologically relevant exposure measurement. Proc Natl Acad Sci U S A 91:360–364. Netherlands Health Council. 1994. Risks of UV radiation Committee: UV radiation from sunlight. The Hague: Health Council of the Netherlands. Newton R. 1997. Solar ultraviolet radiation is not a major cause of primary cutaneous non-Hodgkin’s lymphoma. Br Med J 314:1483–1484. Newton R, Roman E, Fear N, et al. 1996. Non-Hodgkin’s lymphoma and solar ultraviolet radiation. Data are inconsistent. Br Med J 313:298. Nghiem DX, Kazimi N, Clydesdale G, et al. 2001. Ultraviolet a radiation suppresses an established immune response: Implications for sunscreen design. J Invest Dermatol 117:1193–1199. Noonan FP, Recio JA, Takayama H, et al. 2001. Neonatal sunburn and melanoma in mice. Nature 413:271–272. NRPB (National Radiological Protection Board). 1995. Board statement on effects of ultraviolet radiation on human health and health effects from ultraviolet radiation. Chilton, Didcot, UK: National Radiological Protection Board.

Solar Radiation NRPB (National Radiological Protection Board). 2002. Advice on protection against ultraviolet radiation. Didcot, Oxfordshire: National Radiological Protection Board. Olson AL, Dietrich AJ, Sox CH, et al. 1997. Solar protection of children at the beach. Pediatrics 99:E1. O’Riordan DL, Stanton WR, Eyeson Annan M, et al. 2000. Correlations between reported and measured ultraviolet radiation exposure of mothers and young children. Photochem Photobiol 71:60–64. Osterlind A, Hou-Jensen K, Moller-Jensen O. 1988. Incidence of cutaneous malignant melanoma in Denmark 1978–1982. Anatomic site distribution, histologic types and comparison with non-melanoma skin cancer. Br J Cancer 58:385–391. Ouhtit A, Nakazawa H, Armstrong BK, et al. 1998. UV-radiation-specific p53 mutation frequency in normal skin as a predictor of risk of basal cell carcinoma. J Natl Cancer Inst 90:523–531. Palmer JS, Duffy DL, Box NF, et al. 2000. Melanocortin-1 receptor polymorphisms and risk of melanoma: Is the association explained solely by pigmentation phenotype? Am J Hum Genet 66:176–186. Pathak MA, Sinesi SJ, Szabo G. 1965. The effect of a single dose of ultraviolet radiation on epidermal melanocytes. J Invest Dermatol 45:520–528. Pollock PM, Yu F, Qiu L, et al. 1995. Evidence for u.v. induction of CDKN2 mutations in melanoma cell lines. Oncogene 11:663–668. Purdie KJ, Pennington J, Proby CM, et al. 1999. The promoter of a novel human papillomavirus (HPV77) associated with skin cancer displays UV responsiveness, which is mediated through a consensus p53 binding sequence. EMBO J 18:5359–5369. Putnam GL, Yanagisako KL. 1985. Skin cancer comic book: Evaluation of a public educational vehicle. J Audiov Media Med 8:22–25. Quevedo WC, Szabo G, Virks J, et al. 1965. Melanocyte populations in UVirradiated human skin. J Invest Dermatol 45:295–298. Quintern LE, Furusawa Y, Fukutsu K, et al. 1997. Characterization and application of UV detector spore films: The sensitivity curve of a new detector system provides good similarity to the action spectrum for UV-induced erythema in human skin. J Photochem Photobiol B 37:158–166. Rady P, Scinicariello F, Wagner RF, Jr., et al. 1992. p53 mutations in basal cell carcinomas. Cancer Res 52:3804–3806. Ravanat JL, Douki T, Cadet J. 2001. Direct and indirect effects of UV radiation on DNA and its components. J Photochem Photobiol B 63:88–102. Reding DJ, Fischer V, Gunderson P, et al. 1995. Skin cancer prevention: A peer education model. Wis Med J 94:77–81. Rhainds M, De Guire L, Claveau J. 1999. A population-based survey on the use of artificial tanning devices in the Province of Quebec, Canada. J Am Acad Dermatol 40:572–576. Rowland FS, Molina MJ. 1974. Stratospheric sink for chlorofluoromethanes chlorine atom-catalysed destruction of ozone. Nature 249:810–812. Roy CR, Gies HP, Toomey S. 1997. Monitoring UV-B at the earth’s surface. Australian Meteorological Magazine 46:203–210. Ruhland A, de Villiers EM. 2001. Opposite regulation of the HPV 20-URR and HPV 27-URR promoters by ultraviolet irradiation and cytokines. Int J Cancer 91:828–834. Ryan KG, Smith GJ, Rhoades DA, et al. 1996. Erythemal ultraviolet insolation in New Zealand at solar zenith angles of 30 and 45. Photochem Photobiol 63:628–632. Sarasin A, Giglia-Mari G. 2002. p53 gene mutations in human skin cancers. Exp Dermatol 11(Suppl 1):44–47. Schwartz GG, Hulka BS. 1990. Is vitamin D deficiency a risk factor for prostate cancer? (Hypothesis). Anticancer Res 10:1307–1311. Schwarz A, Schwarz T. 2002. Molecular determinants of UV-induced immunosuppression. Exp Dermatol 11(Suppl 1):9–12. Schwarz A, Stander S, Berneburg M, et al. 2002. Interleukin-12 suppresses ultraviolet radiation-induced apoptosis by inducing DNA repair. Nat Cell Biol 4:26–31. Schwarz T. 1999. Ultraviolet radiation-induced tolerance. Allergy 54:1252–1261. Setlow R, Grist E, Thompson K, et al. 1993. Wavelengths effective in induction of malignant melanoma. Proc Natl Acad Sci U S A 90:6666–6670. Setlow RB, Woodhead AD, Grist E. 1989. Animal model for ultraviolet-radiation induced melanoma: Platyfish-swordtail hybrid. Proc Natl Acad Sci U S A 86:8922–8926. Shamanin V, Glover M, Rausch C, et al. 1994. Specific types of human papillomavirus found in benign proliferations and carcinomas of the skin in immunosuppressed patients. Cancer Res 54:4610–4613.

305

Slaper H, Velders GJ, Daniel JS, et al. 1996. Estimates of ozone depletion and skin cancer incidence to examine the Vienna Convention achievements. Nature 384:256–258. Smith R, Healy E, Siddiqui S, et al. 1998. Melanocortin 1 receptor variants in an Irish population. J Invest Dermatol 111:119–122. Smith T. 1979. The Queensland Melanoma Project—An exercise in health education. Br Med J 1:253–254. Staricco RJ, Pinkus H. 1957. Quantitative and qualitative data on the pigment cells of adult human epidermis. J Invest Dermatol 28:33–44. Stierner U, Rosdahl I, Augustsson A, et al. 1989. UVB irradiation induces melanocyte increase in both exposed and shielded human skin. J Invest Dermatol 92:561–564. Studzinski GP, Moore DC. 1995. Sunlight—can it prevent as well as cause cancer? Cancer Res 55:4014–4022. Szepietowski JC, Nowicka D, Soter K, et al. 2002. Tanning salons in southwest Poland: A survey of safety standards and professional knowledge of the staff. Photodermatol Photoimmunol Photomed 18:179–182. Taguchi M, Watanabe S, Yashima K, et al. 1994. Aberrations of the tumor suppressor p53 gene and p53 protein in solar keratosis in human skin. J Invest Dermatol 103:500–503. The Cancer Council Australia. 2001. National Cancer Prevention Policy 2001–2003. Canberra: The Cancer Council Australia. Tenkate TD, Collins MJ. 1997. Personal ultraviolet radiation exposure of workers in a welding environment. Am Ind Hyg Assoc J 58:33–38. Thieden E, Agren MS, Wulf HC. 2001. Solar UVR exposures of indoor workers in a Working and a Holiday Period assessed by personal dosimeters and sun exposure diaries. Photodermatol Photoimmunol Photomed 17: 249–255. Thompson SC, Jolley D, Marks R. 1993. Reduction of solar keratoses by regular sunscreen use. N Engl J Med 329:1147–1151. Vajdic CM, Kricker A, Giblin M, et al. 2002. Sun exposure predicts risk of ocular melanoma in Australia. Int J Cancer 101:175–182. Valverde P, Healy E, Jackson I, et al. 1995. Variants of the melanocyte-stimulating hormone receptor gene are associated with red hair and fair skin in humans. Nature Genetics 11:328–330. van-Elsas A, Zerp SF, van-der-Flier S, et al. 1996. Relevance of ultravioletinduced N-ras oncogene point mutations in development of primary human cutaneous melanoma [see comments]. Am J Pathol 149:883–893. van-’t-Veer LJ, Burgering BM, Versteeg R, et al. 1989. N-ras mutations in human cutaneous melanoma from sun-exposed body sites. Mol Cell Biol 9:3114–3116. Wei Q, Matanoski G, Farmer E, et al. 1993. DNA repair and aging in basal cell carcinoma: a molecular epidemiologic study. Proc Natl Acad Sci U S A 90:1614–1618. Weinstock MA, Colditz GA, Willett WC, et al. 1991. Recall report bias and reliability in the retrospective assessment of melanoma risk. Am J Epidemiol 133:240–245. Whillock MJ, Clark I, McKinlay AF, Todd C, Mundy S. 1988. Ultraviolet radiation levels associated with the use of fluorescent general lighting, UVA and UVB lamps in the workplace and home. Chilton: National Radiological Protection Board. Whiteman D, Green A. 1997. Wherein lies the truth? Assessment of agreement between parent proxy and child respondents. Int J Epidemiol 26:855–859. Whiteman D, Green A. 1998. Melanoma and sun exposure: where are we now? Int J Dermatol. 38:481–489. Whiteman DC, Whiteman CA, Green AC. 2001. Childhood sun exposure as a risk factor for melanoma: A systematic review of epidemiologic studies. Cancer Causes Control 12:69–82. WMO (World Meteorological Organization). 1998. Scientific assessment of ozone depletion: 1998. World Meteorological Organization Global Ozone Research and Monitoring Project. Woollons A, Clingen PH, Price ML, et al. 1997. Induction of mutagenic DNA damage in human fibroblasts after exposure to artificial tanning lamps. Br J Dermatol 137:687–692. Yarosh D, Alas LG, Yee V, et al. 1992. Pyrimidine dimer removal enhanced by DNA repair liposomes reduces the incidence of UV skin cancer in mice. Cancer Res 52:4227–4231. Yarosh D, Klein J, O’Connor A, et al. 2001. Effect of topically applied T4 endonuclease V in liposomes on skin cancer in xeroderma pigmentosum: a randomised study. Lancet 357:926–929. Yarosh DB, O’Connor A, Alas L, et al. Photoprotection by topical DNA repair enzymes: molecular correlates of clinical studies. Photochem Photobiol 69:136–140.

17

Electromagnetic Fields and Radiofrequency Radiation DAVID A. SAVITZ AND ANDERS AHLBOM

E

lectromagnetic fields are characterized by their frequency or wavelength, with the wavelength inversely proportional to the frequency. At the lower end of the electromagnetic spectrum it is customary to refer to the frequency, while at the upper end one usually uses wavelengths. The energy is directly proportional to the frequency; the higher frequency generates more energy, as described by Planck’s law. The electromagnetic spectrum can be divided into an ionizing and a nonionizing segment (Fig. 17–1). In the nonionizing portion, the energy is too weak to break chemical bonds and thus to form ions. This chapter is concerned exclusively with the nonionizing part of the electromagnetic spectrum, and within that range, only with frequencies up to 300 GHz. This range of frequencies includes fields that are generated in connection with production, transmission, distribution, and use of electric power. Such fields usually have a frequency of 50 or 60 Hz. This band also includes the fields that are used for mobile telephone communication, both in the phones themselves and at the base stations. This technology typically uses frequencies from 900 MHz up 2500 MHz, although new technology is likely to extend this band. We will refer to these frequencies as ELF (extremely low frequency) and RF (radiofrequency). Other frequencies in the nonionizing range of the electromagnetic spectrum are used in commercial and research applications, and a few, such as radar, have corresponding epidemiologic research on potential health effects. However, the interest of researchers and the public has focused on these two frequency bands and these are the principal focus of our review. The ELF magnetic fields in the environment are usually characterized by their magnetic flux density, which is measured in units of Tesla (T) or rather microTesla (mT) in the range typically encountered. Electric fields, which are distinct from magnetic fields in the ELF range, are typically characterized in terms of Volts per meter (V/m). Higher frequency fields, including the RF fields, are characterized by their current density, measured as amperes per square meter (A/m2). Radiofrequency fields have wavelengths of a few centimeters or less, depending on the actual frequency. Therefore, energy is deposited in the body, mainly within a couple of centimeters from the body surface. The only well-established consequence of this energy deposition is heating. The rationale for existing guidelines is to prevent excessive heating, locally or in the whole body. The unit of measurement that addresses the potential for tissue heating is the specific absorption rate (SAR) measured in Watts per kilogram (W/kg). Some mobile phones are labeled with their SAR value. However, for practical reasons, this cannot be measured directly inside the body, and is instead established based on models and theoretical calculations. The actual field levels that people commonly encounter (e.g., in connection with mobile phone use), are below the exposure guideline levels, but of the same order of magnitude. However, fields from base stations are orders of magnitude below the guideline levels.

and the human body is through induction of electric currents. The current density induced in the body is a function of the external magnetic field flux density, which is why such fields are typically characterized by their level in units of mT. Both theoretical calculations and laboratory research indicate that high internal current densities cause acute biological effects, in contrast to the uncertainty in the very low range of magnetic flux density commonly encountered. Present exposure guidelines are based on these acute effects resulting from electric currents, and are intended to prevent neurological effects by restricting the internal current density. The environmental flux densities required to produce such internal current densities are orders of magnitude above what one normally encounters in the general environment, with the exception of certain work environments where high electric fields can momentarily result in current densities near the limits.

Interaction of Nonionizing Radiation with Biological Systems

Public Concern with Nonionizing Radiation

Extremely low-frequency fields have a long wavelength, with 50 Hz corresponding to a wavelength of 3500 km, approximately the earth’s radius. As a consequence of this extremely long wavelength, such fields essentially pass through the body without depositing any energy directly. The established mechanism of interaction between such fields

306

Potential Mechanisms of Carcinogenicity For the low frequency fields of concern, there is no known mechanism of biological interaction besides current induction and heating. Indeed, given the small amount of energy that is deposited in connection with exposure to these fields, other mechanisms or biological effects such as mutation are unlikely. Therefore, researchers who have sought to understand the possible implications of exposure to these fields on cancer risk have sought previously unknown mechanisms. Interest in laboratory studies of possible carcinogenicity has been driven largely by epidemiological data suggesting an association with cancer risk and ELF fields as discussed in detail below. However, despite much effort and some enticing leads from isolated studies, thus far the mechanistic experimental studies and the animal toxicology literature fail to provide a clear, replicable indication of effects that would support the hypothesis that these fields are carcinogenic (National Research Council, 1997; Portier and Wolfe, 1998). No indication of increased leukemia risk in experimental animals has been observed. Over the years, several putative carcinogenic mechanisms have been considered and extensively examined. The cellular research in particular has been extensive and scientists have addressed possibilities including genetic mutations, gene transcriptions, heat-shock proteins, and signal transduction. For RF fields, an early study on genetically modified mice with an effect on lymphoma rates triggered interest (Repacholi et al., 1997). However, a later study failed to replicate these findings (Utteridge et al., 2002), consistent with the evolution of research on ELF fields over many years. It seems likely that small thermal effects may indeed explain some of the effects that have been seen in laboratory experiments. Some recent cancer experimental studies see positive results and are of potential interest, but it is too early to assess their validity and relevance (Tice et al., 2002).

Despite important distinctions between ELF electric and magnetic fields and RF radiation with regard to the physics, biophysics, and epidemiologic evidence, important similarities regarding the motivation for evaluation of carcinogenicity exist. As noted above, laboratory research provides little or no basis for a concern with adverse health outcomes. The impetus for epidemiologic study has instead come

307

Electromagnetic Fields and Radiofrequency Radiation Visible light Power lines, electric appliances AM TV radio FM radioRadar

Infrared

Gamma rays

UV-light X-rays

103

10

105

107

109

1011

1013

1015

1017

1019

1021 Hz

0

100

104

Low frequency

106

108

Radio waves

1010

1012

Micro waves

1014

1016

1018

1020

1022

Ionizing radiation

Figure 17–1. Electromagnetic spectrum.

largely from public concern and the profound policy implications associated with the technologies that generate exposure. Humans have developed (and continue to develop) technologies that result in unprecedented exposures that can spread with remarkable speed (e.g., mobile telephones). Not surprisingly, the invisible byproducts of these widely used technologies raise questions about potential health consequences. The public often challenges the dissemination of such technology (e.g., siting of high-tension power lines or mobile telephone base stations), claims that illness has resulted from such exposures, and argues for regulation to limit the health and environmental concerns. Researchers interested in the most biologically plausible approaches to understanding the etiology of cancer or seeking the greatest reduction in cancer incidence and mortality often view the study of such environmental agents with justified skepticism. However, the societal need for informed regulatory policy and the political demand to respond to the public’s concern are likely to continue to encourage examination of the potential carcinogenicity of nonionizing radiation.

EXTREMELY LOW-FREQUENCY ELECTRIC AND MAGNETIC FIELDS Exposure Assessment Moving from the principles governing the generation of electric and magnetic fields to the practical aspects of exposure assessment in epidemiologic studies is not straightforward due to the complexity of the human environment. The spatial and temporal variability in these fields, combined with the movement of people in the environment, requires simplifying assumptions to achieve a feasible approach to exposure assessment. In the frequency range of interest, 50 to 60 Hz, electric fields and magnetic fields must be considered separately. Electric fields are produced wherever there is electric potential, whether or not current is flowing, and are readily shielded or distorted by a wide variety of materials. Thus, the electric fields from power lines do not predict exposure levels within homes, given trees and building materials between the lines and the occupied environment. Appliance location and grounding practices make it quite difficult to accurately estimate electric fields within homes or extrapolate measured fields across locations or time. Magnetic fields are not perturbed by commonly encountered materials, so that the fields just outside the home (from power lines, for example), are quite similar to the levels just inside the home. Magnetic fields are produced by the flow of electric current, so that the use of electricity in homes, appliances, and workplaces determine environmental magnetic field levels. Research on residential exposures has focused almost exclusively on magnetic fields, since those fields are related most directly to power lines and were implicated in initial epidemiologic studies (Wertheimer and Leeper, 1979, 1982). Magnetic fields are far more feasible to study because of established markers like power lines. In the occupational environment, exposure sources are more diverse and it is less clear whether a given electri-

cal occupation is associated with elevated electric fields, magnetic fields, or both. Much of the focus in occupational epidemiology is on magnetic fields as well, but electric fields have also been considered directly (Miller et al., 1996; Guenel et al., 1996; Green et al., 1999). Prediction of exact levels of magnetic (or electric) fields at a specific time is difficult, but the interest in potential carcinogenicity is focused on average levels over prolonged periods of time (i.e., months to years). Lacking a clear biologic mechanism, the focus on such averages is largely a default assumption, and arguments can also be made for focusing on peak exposures, variability in exposure, or time above some low threshold. Often, such indices of exposure are highly correlated with one another (Armstrong et al., 1990; Savitz et al., 1994), so that estimation of the time-weighted average is a reasonable goal for epidemiologic research. With the more attainable goal of finding markers of long-term timeweighted average magnetic field exposure, epidemiologists have found a number of useful indicators. Residential exposures vary, to some extent, in predictable ways based on the configuration and proximity of outside power lines (Kaune et al., 1987, 1994). Persons who live in homes with greater proximity to lines carrying more current, on average, tend to have increased magnetic field exposure, which was the basis for the original Wertheimer-Leeper wire coding system (Wertheimer and Leeper, 1979). Certain appliances that are used for prolonged periods of time in close proximity to the user, such as electric blankets, are capable of influencing magnetic field exposure, though for such sources the levels vary across different location on the body. Occupations that involve regular and prolonged proximity to energized electrical equipment have been found to produce elevated levels of magnetic and often electric field exposure (Armstrong et al., 1990). In each case, there is a well-defined and relatively stable source of fields, and people are located in sufficiently close proximity for extended periods to increase their time-weighted average exposure. Redefining the goal to the ascertainment of peak exposures or indices of variability, for example, may alter the value of such markers and make other determinants more important. Even for the assessment of time-weighted average exposures, the limitations in such measures should be appreciated. There is often substantial variability over short distances, more so for appliances and analogous sources in the workplace (e.g., electrical machinery), so that movement of individuals affects the exposure. The need to generalize across categories or homes or workplaces to conduct meaningful epidemiologic studies of large numbers of individuals requires compromises in accuracy of assignment. The goal is not to accurately identify a few homes or workplaces in detail; the goal is to summarize exposure for hundreds or thousands of such locations. In contrast to many environmental agents, electric and magnetic fields exposure is not limited by a single dominant source. With rare exceptions, ambient levels in homes, use of electrical appliances, exposure levels in the workplace and other frequently occupied environments will all contribute materially to the total magnetic field exposure, and there are

308

PART III: THE CAUSES OF CANCER

often diverse contributors within each of those environments. Lacking a biological marker, exposure assessment has been the preeminent concern from the outset and remains the chief limitation in epidemiologic studies of carcinogenicity.

Exposure Sources and Populations Studied Residential Exposure Several approaches have been developed to estimate long-term residential exposure to magnetic fields. All approaches are indirect relative to historical monitoring data, including present-day measurements as an indicator of past exposure, and all seek to approximate average levels. Since the rarity of most of the cancers of interest makes true prospective studies extremely inefficient, all studies to date have required retrospective exposure assessment. The need for indicators of historical exposure has resulted in a focus on power lines near the home as a stable marker, and to the use of present-day measurements as a potential surrogate for past exposure levels. Wire configuration codes were developed by Wertheimer and Leeper (1979) as a means of estimating the contribution of nearby power lines to residential magnetic fields, with homes that are nearer to lines carrying larger amounts of current resulting in assignment to a higher exposure stratum. Although the correspondence between estimated and measured magnetic field exposure is far from perfect (Wertheimer and Leeper, 1979; Savitz et al., 1988; Severson et al., 1988; Kleinerman et al., 1997), the predicted gradient has been observed to at least some extent in a number of settings in which it has been applied. In particular, homes assigned the highest level, Very High Current Configuration, and to a lesser extent, Ordinary High Current Configuration (Wertheimer and Leeper, 1982) have been found to have higher average magnetic fields than homes in the other categories (Kaune et al., 1987; Tarone et al., 1998). This system is applicable to homes in North America, primarily, where distribution lines are generally above ground and thus amenable to visual inspection. The wire code method infers current flow based on structural characteristics of the lines. In some settings, notably northern Europe, more refined approaches to estimation can be applied where exposures are determined more by transmission lines. Historical records of current flow along transmission lines combined with information on the geometry of the power lines relative to the homes allows investigators to make more precise estimates not just of present-day exposures, but the historical exposures of etiologic interest (Feychting and Ahlbom, 1993). Measurement of magnetic fields in homes is quite feasible with current instrumentation, subject to respondent cooperation, unlike wire codes, which can be done by visual inspection only. Systematic protocols for assessing magnetic field levels in homes that summarize over spatial variability within the home and across short-term temporal fluctuations have been developed (Feychting and Ahlbom, 1993; Kleinerman et al., 1997). Although these measured values are often mistakenly viewed as a “gold standard” even though they were not collected in the historical period of interest, they have come to be viewed as the best available marker in settings in which the determinants of exposure are too complex to estimate readily based on engineering considerations alone. The main argument against their value as a marker is their susceptibility to changes over time as a result of modifications to the environment since the historical period of interest due to such factors as wiring changes and movement of appliances. However, they also are believed to capture subtle determinants that remain stable over the time between disease development and conduct of the study. As research interest has focused on the rarer, more highly elevated magnetic fields of 0.3 mT or above (Ahlbom et al., 2000; Greenland et al., 2000), the relative interest in measured magnetic fields as the primary exposure indicator has been enhanced.

Residential Exposure and Childhood Cancers Wertheimer and Leeper published the first study on childhood cancer and EMF in 1979. That study initiated the intense scientific and public

interest in biological effects of EMF during the following two decades. This original study was a case-control study on childhood cancer mortality. The study used wire codes as markers for magnetic field exposure. Wire codes classify buildings according to proximity to and type of nearby power lines. The study found a substantial excess risk for children living in homes with wire codes indicative of higher magnetic field exposure. For childhood cancers in the aggregate, the relative risk was more than doubled for the high current configuration wire code compared to the low current configuration. For childhood leukemia, the corresponding relative risk was 3.0 (1.8–5.0) and for brain tumors it was 2.4 (1.0–5.4). The study was severely criticized for the procedures used to select subjects, for unblinded exposure assignment, and other methodological issues, including the assumption that wire codes could predict magnetic field exposure (Ahlbom et al., 2001). While a number of the general methodological issues were never fully resolved, subsequent measurement studies did show that the wire codes do predict magnetic field exposure, with buildings with higher wire codes on the average tending to have higher measured magnetic fields than buildings with lower wire codes. To either refute or confirm the results of this original study, some 20 studies have been conducted around the world, nearly all casecontrol studies (Table 17–1) (Wertheimer and Leeper 1979; Fulton et al., 1980; Tomenius, 1986; Savitz et al., 1988; Myers et al., 1990; London et al., 1991; Feychting and Ahlbom, 1993; Olsen et al., 1993; Verkasalo et al., 1993; Preston-Martin et al., 1996; Gurney et al., 1996; Tynes and Haldorsen, 1997; Linet et al., 1997; Michaelis et al., 1997; Dockerty et al., 1998; McBride et al., 1999; Green et al., 1999; United Kingdom Childhood Cancer Study Investigators, 1999). The more recent studies have built on the experiences of previous studies and have refined exposure assessment over time. For leukemia, a number of the later studies have tended to support the original results. This came as a surprise to many researchers and reviewers who had assumed the original results would not be repeated and that the interest in the issue would quickly fade. The first major study to follow was an attempt to repeat the first study as closely as possible, including a similar methodology for exposure assessment, also conducted in the Denver area. The results were similar to those of the original study (Savitz et al., 1988), though associations were somewhat lower in magnitude. This study by no means resolved the issue, but it helped encourage other researchers to take the issue seriously, and fostered a series of methodological studies, particularly focused on exposure assessment. A group of Nordic studies used a different method for exposure assessment based on calculated historical fields from transmission lines and took advantage of existing population registries (Feychting et al., 1993; Olsen et al., 1993; Verkasalo et al., 1993; Tynes and Haldorsen, 1997). Thus, these studies had more accurate exposure data and minimized the risk of selection bias, the two greatest deficiencies of the US studies. On the other hand, the number of cases was limited because of the small populations in the Nordic countries and the rarity of childhood cancer. Each of the first three Nordic studies also found evidence of a positive association between magnetic field exposure and leukemia, both individually and when pooled (Feychting and Ahlbom, 1993; Olsen et al., 1993; Verkasalo et al., 1993; Ahlbom et al., 2000). The last Nordic study to get published, the Norwegian, found no excess risk, but had only two exposed cases (Tynes and Haldorsen, 1997). The US National Cancer Institute together with the Children’s Cancer Group performed the biggest study at the time with magnetic field measurements performed during a 24-hour period in the children’s homes (Linet et al., 1997). Because of the prospective identification of case and controls, the measurements could be conducted shortly after diagnosis, usually within 1 year. Thus, this study addressed several of the concerns raised in relation to previous studies. The results, however, were somewhat compromised by a low response proportion, with some indications that it might have introduced selection bias (Kleinerman et al., 2000). The results were initially reported as completely negative, which they were with respect to wire codes; this interpretation was challenged by many reviewers who focused on the evidence of an association with elevated measured magnetic fields,

Table 17–1. Characteristics of Studies and Results on Relation between Magnetic Field Exposure and Childhood Cancer

Reference Wertheimer and Leeper, 1979

Fulton et al., 1980

Tomenius, 1986

Savitz et al., 1988

Myers et al., 1990

London et al., 1991

Feychting and Ahlbom, 1993

Olsen et al., 1993

Verkasalo et al., 1993

PrestonMartin et al., 1996

Study Population Denver resident born in Colorado Cases: <19 yrs, deaths (1950– 1973) Controls: birth certificates Rhode Island resident Cases: <20 yrs Controls: birth certificates Stockholm County Sweden Cases: <19 yrs (1958–1973) Controls: birth certificates Denver resident Cases: <15 yrs (1976–1983) Controls: random digit dialing Yorkshire, England Cases: <15 yrs (1970–1979) Controls: birth register Los Angeles County resident Case: <10 yrs (1980–1987) Controls: friends and random digit dialing Sweden resident within 300 m of 220 or 400 kV power line Cases: <15 yrs (1960–1985) Controls: selected at random from cohort to match cases Denmark resident Cases: <15 yrs (1960–1986) Controls: Central Population Registry Finland resident within 500 m of 110–400 kV power line Cases: <17 yrs (1974–1990) Los Angeles County resident Cases: <20 yrs (1984–1991) Controls: random digit dialing

Primary Exposure Measure(s) Wire code of diagnosis/ death home

Study Design Casecontrol

Wire code CaseCases: all control lifetime homes Controls: birth homes Front door Casemeasurement control at birth and diagnosis residences Wire code spot magnetic field measurements child’s bedroom low power

Casecontrol

Distance of home to nearest overhead line; estimated MF strength Wire code and 24-hr child’s bedroom magnetic field measurement in home lived in longest low power Historical calculated fields

Casecontrol

Magnetic Field Measurements RR (95% CI) (high category)

Wire Codes RR (95% CI) (high category)

Cancers (Numbers Cases, Controls) All cancers (328, 328) Leukemia (155, 155) Brain tumors (66, 66)

2.3—(HCC) 3.0 (1.8–5.00) (HCC) 2.4 (1.0–5.4) (HCC)

— — —

Leukemia (119, 240)

1.0—(HCC)

—

All cancers (1033, 890) Leukemia (243, 212) Brain tumors (294, 253)

— — —

1.8—(≥0.3 mT) 0.3—(≥0.3 mT) 3.7—(≥0.3 mT)

2.2 (1.0–5.2) (VHCC) 2.8 (0.9–8.0) (VHCC) 1.9 (0.5–8.0) (VHCC) 1.1 (0.5–2.6) (<25 m distance) 0.4 (0.0–4.3) (≥0.1 mT)

1.4 (0.6–2.9) (≥0.25 mT) 1.9 (0.7–5.6) (≥0.25 mT) 1.0 (0.2–4.8) (≥0.25 mT)

2.2 (1.1–4.3) (VHCC)

1.2 (0.5–2.8) (≥0.125 mT)

wire

All code cancers 320 Leukemia 97 Brain 59 tumors Controls 259 All cancers (374, 588)

Casecontrol Leukemia Controls

magnetic field 128 36 25 207

wire code

magnetic field

211 205

162 143

—

Nested Casecontrol

All cancers (141) Leukemia (38) Brain tumors (33) Controls (554)

1.3 (0.6–2.7) (≥0.3 mT) 3.8 (1.4–9.3) (≥0.3 mT) 1.0 (0.2–3.9) (≥0.3 mT)

Historical calculated fields

Casecontrol

All cancers (1707, 4788) Leukemia (833, 1666) Brain tumors (624, 1872)

5.6 (1.6–19) (≥0.4 mT) 6.0 (0.8–44) (≥0.4 mT) 6.0 (0.8–44) (≥0.4 mT)

Historical calculated fields

Cohort

All cancers (140) Leukemia (35) Brain tumors (39)

1.5 (0.74–2.7) (≥0.2 mT) 1.6 (0.32–4.5) (≥0.2 mT) 2.3 (0.75–5.4) (≥0.2 mT)

Wire code at diagnosis, first, and longest residence

Casecontrol

Brain tumors Controls

wire code

magnetic fields

281 250

106 99

1.2 (0.6–2.2) (VHCC)

1.7 (0.6–5.0) (≥0.3 mT)

(continued)

309

Table 17–1. (cont.)

Reference Gurney et al., 1996

Study Population

Primary Exposure Measure(s)

Seattle and Wire code of surrounding diagnosis western home Washington State Cases: <20 yrs (1984–1990) Controls: random digit dialing Tynes and Norway resident Historical Haldorsen, in census ward calculated 1997 with high-voltage fields power line Cases: <15 yrs (1965– 1989) Controls: selected at random from cohort to match cases Linet et al., 9 Mid-Atlantic Wire code 1997 and midwestern residences states in US, >70% of 5 yrs Cases: <15 yrs before (1989–1993) diagnosis; Controls: random time-weighted digit dialing average magnetic field measurements all residences combined >70% of 5 yrs before diagnosis Michaelis Northwest 24-hr child’s et al., Germany (Lower bedroom MF 1997 Saxony) and measurement Berlin resident Cases: <15 yrs (1991–1995) Controls: government office residents’ registry Dockerty New Zealand 24-hr child’s et al., resident bedroom 1998 Cases: <15 yrs magnetic field (1990–1993) measurement Controls: birth certificate McBride Canada 5 Wire code of et al., provinces home at 2 yrs 1999 before Cases: <15 yrs diagnosis 48(1990–1994) hr personal Controls: measurement provincial 24-hr child’s health insurance bedroom 2 yrs rolls before diagnosis Green et al., Southern Ontario Wire code; Spot 1999 Canada resident magnetic field Cases: <15 yrs measurements; (1985–1993) 48-hr personal Controls: monitoring telephone marketing lists

310

Study Design

Cancers (Numbers Cases, Controls)

Casecontrol

Brain tumors (120, 240)

Nested Casecontrol

All cancers (532, 2112) Leukemia (139, 546) Brain tumors (144, 599)

Casecontrol

Acute lymphoblastic leukemia Controls

Magnetic Field Measurements RR (95% CI) (high category)

Wire Codes RR (95% CI) (high category) 0.5 (0.2–1.6) (VHCC)

—

0.9 (0.5–1.8) (≥0.14 mT) 0.3 (0.0–2.1) (≥0.14 mT) 0.7 (0.2–2.1) (≥0.14 mT)

wire code

magnetic field

402 402

624 615

0.9 (0.5–1.6) (VHCC)

1.2 (0.9–1.8) (≥0.3 mT) 1.72 (1.0–2.9) (≥0.3 mT)

Casecontrol

Leukemia (176, 414)

—

2.3 (0.8–6.7) (≥0.2 mT)

Casecontrol

Leukemia (115, 117)

—

15.5 (0.3–7.6) (≥0.2 mT)

Casecontrol

Leukemia wire code (303, 309) 48-hr personal monitoring (293, 339) 24-hr child’s bedroom (272, 304)

Casecontrol

Leukemia wire code (79, 125) Spot meas. (21, 46) 48-hr personal monitoring (88, 133)

0.8 (0.4–1.6) (VHCC)

1.0 (0.7–1.6) (≥0.2 mT) 1.27 (0.7–2.3) (≥0.2 mT)

1.5 (0.3–8.7) (OHCC–VHCC) 1.1 (0.3–4.1) (≥0.4 mT) 4.5 (1.3–15.9) (≥0.14 mT)

311

Electromagnetic Fields and Radiofrequency Radiation Table 17–1. (cont.)

Reference UKCCS, 1999

Study Population England, Wales, Scotland resident Cases: <15 yrs (1992–1995) Controls: Family Health Services Authorities register

Primary Exposure Measure(s)

Study Design

In-home Casemagnetic field control measurements Phase I—90 min. measurement in family room and spot measurements in child’s bedroom. Phase II (highest 10%) 48-hr measurement in child’s bedroom School spot measurements.

Cancers (Numbers Cases, Controls)

Wire Codes RR (95% CI) (high category)

All cancers (2265, 2270) Leukemia (1094, 1096) Brain tumors (390, 393)

— — —

Magnetic Field Measurements RR (95% CI) (high category) 0.9 (0.3–2.3) (≥0.4 mT) 1.7 (0.4–7.1) (≥0.4 mT) 0 cases/2 controls (≥0.4 mT)

HCC, high current configuration; OHCC, ordinary high current configuration; VHCC, very high current configuration.

particularly in the uppermost categories considered. The publication was accompanied by an editorial proclaiming that there was now sufficient evidence to conclude that electric and magnetic field exposure was not associated with childhood leukemia or other health risks (Campion, 1997). The largest study so far is a nationwide populationbased childhood cancer case-control study in the United Kingdom that also used an extensive measurement program (United Kingdom Childhood Cancer Study Investigators, 1999). This study found no support for an association with leukemia. Despite the size of study, however, the small numbers in the very highest exposure categories limited precision. In an attempt to combine the available studies in a systematic way, a pooled analysis of all childhood leukemia studies with exposure assessment fulfilling certain quality criteria was conducted using the primary data in each of the included studies (Ahlbom et al., 2000). All nine studies with either calculated fields or 24-hour or longer exposure measurements were included. Use of original data had two important advantages. First, it was possible to apply exposure definitions consistently across the studies, which allowed the same cut points, the same reference categories, and the same models to calculate mean magnetic field values. Second, because of the large number of subjects in the combined data set, it was possible to assess associations for higher exposure levels than could be done in the individual studies. The main result of this meta-analysis was that no risk elevation was seen below 0.4 mT but the relative risk for ≥0.4 mT was 2.0 (95% CI = 1.3–3.1). Interestingly enough, the individual study with the largest impact on this pooled result was the NCI study that initially was considered by some to prove that magnetic fields were not associated with any health risks. However, a sensitivity analysis showed that the results would hold even without including that study. A stratified analysis showed similar results in the group of Nordic studies as in the rest of the studies. The merit of this comparison is that the Nordic studies are less likely to be affected by selection bias, but have more imprecise estimates than the rest of the studies because of smaller numbers. The authors concluded that chance is an unlikely explanation to the excess risk, but that selection bias may have accounted for some of it. Another group published a similar pooled analysis, but this analysis was based on wider inclusion criteria and, thus, a larger number of studies (Greenland et al., 2000). Despite a number of different decisions regarding inclusion and analytic methods, the results of these two analyses are quite similar. Exposure to 0.4 mT is very rare, with less than 1% of the controls in the aggregation of studies exposed

at that level (Ahlbom et al., 2000). However, if a causal effect of such exposures were proven, one would have to consider carefully whether this is likely to be a true biological threshold or merely a level at which the effect becomes discernible. The literature on other types of childhood cancer differs from that of leukemia in both volume of studies and findings. Initially, the brain tumor results were as strongly in support of a magnetic field effect as the leukemia results. The two Denver studies (Wertheimer and Leeper, 1979; Savitz et al., 1988) reported odds ratios of 2.4 (1.0–5.4) and 1.9 (0.5–8.0), respectively. The next round of studies of childhood brain tumors was consistently negative (Feychting and Ahlbom 1993; Preston-Martin et al., 1996; Gurney et al., 1995; United Kingdom Childhood Cancer Study Investigators, 1999). The only exceptions are two Nordic studies, but they have only two exposed cases in each study and, indeed, in one of the studies it was two tumors in one subject (Olsen et al., 1993; Verkasalo et al., 1993). The challenges of studying childhood leukemia are exacerbated for childhood brain tumors, which are even rarer. The largest study available has 390 cases of childhood brain tumor of which none was classified as having a magnetic field above 0.4 mT (United Kingdom Childhood Cancer Study Investigators, 1999). Compared to childhood leukemia, there is little support for an association between magnetic fields and childhood brain tumors but there is also far less extensive evidence. For other pediatric cancers, such as lymphoma, the information is even sparser and only scattered data exist.

Residential Exposure and Adult Cancers Initial concerns with residential magnetic field exposure and childhood cancer (Wertheimer and Leeper, 1979) were followed shortly after by a report on adult cancers (Wertheimer and Leeper, 1982). A similar association across multiple cancer types was suggested, but research on residential exposures and adult cancers has not generated the sustained interest and thus the volume of research and attention that childhood leukemia has generated. Leukemia in adults is generally too rare to examine effectively in cohort studies (McDowall, 1986; Schreiber et al., 1993), so that most of the pertinent evidence comes from case-control studies (Li et al., 1997). The initial report by Wertheimer and Leeper (1987) was actually negative with regard to leukemia mortality. A case-control study in Western Washington State (Severson et al., 1988) addressed acute non-lymphocytic leukemia and found no associations for either wire

312

PART III: THE CAUSES OF CANCER

codes or measured magnetic fields. A large study in Yorkshire, England with limited exposure data provided a modest suggestion that living within 50 m of an overhead transmission power line (relative risk = 1.3, 95% CI = 1.0–1.7) or having a calculated magnetic field above 0.3 mT (relative risk = 1.9, 0.8–4.4) might be associated with increased risk (Youngson et al., 1991). The study in Sweden (Feychting and Ahlbom, 1994) considered all forms of leukemia (325 cases, twice that number of matched controls), and focused on estimated historical exposure, a function of proximity to transmission lines and line loadings. Using the estimated exposure nearest the time of diagnosis, leukemia in the aggregate showed no association whereas both acute and chronic myeloid leukemia were increased with exposures above 0.2 mT, with odds ratios of 1.7 (0.7–3.8) and 1.7 (0.8–3.5), respectively. When duration of residence was incorporated to estimate cumulative 15-year exposure, the measures of association for >2 mT –years increased to 1.5 (1.0–2.4) for total leukemia, 2.3 (1.0–4.6) for acute myeloid leukemia, and 2.1 (0.9–4.7) for chronic myeloid leukemia. An assessment of transmission line exposures in Finland (Verkasalo et al., 1996) included 203 leukemia cases and found no evidence of increased risk with proximity to power lines or estimated magnetic field exposure. A study in Taiwan (Li et al., 1997) included 870 leukemia cases and used other cancers as controls. Leukemia was positively associated with living within 50 m of a transmission line (odds ratio = 2.0, 95% CI = 1.5–4.9) and with estimated magnetic fields above 0.2 mT (odds ratio = 1.4, 95% CI = 1.0–1.9). Acute myeloid leukemia showed little association whereas acute lymphocytic and chronic myeloid leukemia did show an association. Thus, there are sporadic suggestions of an association between indices of magnetic field exposure and one or more forms of adult leukemia. However, the quality of the exposure information varies across studies, the specific types of leukemia associated with exposure are not consistent, and the overall volume of research is limited. Studies of residential magnetic field exposure and breast cancer (Wertheimer and Leeper, 1987; Coogan and Aschengrau, 1998; Feychting et al., 1998; Davis et al., 2002) have provided evidence in the aggregate against a positive association. Modest associations were found in initial study in Denver (Wertheimer and Leeper, 1987), the study on Cape Cod based on living within 500 feet of a transmission line or substation (a questionable marker of magnetic field exposure) (Coogan and Aschengrau, 1998) and for younger women and those with estrogen-receptor positive tumors in the Swedish study (Feychting et al., 1998). However the more precise estimates from the Swedish study (Feychting et al., 1998), a case-control study in Taiwan (Li et al., 1997), and the report from a large case-control study in Seattle, Washington (Davis et al., 2002) indicate no association between magnetic field exposure based on calculated historical exposure or wire codes. All estimates of relative risk were very close to the null with reasonable precision. The few studies that have examined a wide array of cancers (McDowall, 1986; Wertheimer and Leeper, 1987; Schreiber et al., 1993; Verkasalo et al., 1996) have provided some suggestions of associations with one cancer type or another, but without replication, these results are of limited importance. Focused studies have included nervous system cancers (Wertheimer and Leeper, 1987; Feychting and Ahlbom, 1994; Li et al., 1997). After the initial report of an association by Wertheimer and Leeper (1987), the subsequent two studies of nervous system cancer provided essentially no evidence of an association between exposure and disease.

Appliance Exposure Appliances that use electricity produce electric and magnetic fields to varying degrees. Some devices with small motors (e.g., hair dryers, electric pencil sharpeners) produce extremely high magnetic fields on the order of a milliTesla, whereas low-resistance devices such as computers produce more modest levels of magnetic fields. The key issue for addressing the chronic exposures of potential health concern is to focus on those appliances that are used for extended periods in close proximity to the user. With that focus, much of the research has been

directed towards electric blanket and other bed heating devices (e.g., heated water beds, electric mattress pads). It is important to distinguish between electric blanket users who turn the device on to warm the bed and then shut it off, eliminating exposure, and those who use the electric blanket throughout the night. There have been some efforts to refine exposure further by examining the heat setting or background room temperature. There is no doubt that other appliances contribute to the time-weighted average and virtually certain that appliances are an important determinant of peak exposure. In addition, because of the ways in which appliances are used and the fields they produce, exposures at specific locations on the body are likely to be strongly influenced by electric appliance exposures (Mader and Peralta, 1992). However, epidemiologic research on those other appliances is quite limited and will not be addressed in detail here.

Childhood Cancer: Exposures to Electric Appliances In Utero and Postnatally Regarding childhood cancer, exposures both in utero (by the mother during pregnancy) and postnatally have been considered. Furthermore, a number of studies examined a broad array of electric appliances but only results for electric blankets and heated waterbeds will be considered in detail (Table 17–2). While the evaluation of those other appliances generated sporadic indications of associations, their uncertain meaning regarding electric and magnetic field exposure and lack of replication tempers the evidence of their importance. Four studies provide relevant data on leukemia, with only one (Hatch et al., 1998) suggesting an increased risk associated with in utero electric blanket exposure. Use of heated waterbeds by mothers tended to be inversely associated with leukemia in the offspring (Table 17–2). Childhood use of electric blankets and heated waterbeds is far less common than adult use, and thus risk estimates are imprecise, in some cases based on cell sizes of one or two. Subject to that caveat, there was some consistent suggestion that childhood use of electric blankets is associated with an increased risk of leukemia. Four studies also provided information on childhood central nervous system cancers. Two showed weak positive associations with electric blanket use prenatally (Savitz et al., 1990; Dockerty et al., 1998), and once again, heated waterbed use prenatally was inversely related to cancer risk. Childhood use of these appliances showed little indication of an association with nervous system cancers. Across these studies, a few other positive associations were found with electric appliance use, in some cases rather sizable in magnitude. London et al. (1991) noted associations of 1.5 or greater with five or more exposed cases between leukemia and the child’s use of black and white televisions, dial clocks, curling irons, electric hair dryers, and video games. Using the same threshold (relative risk of 1.5 or more, five or more exposed cases), Dockerty et al. (1998) found associations between childhood leukemia and mother’s use of a computer monitor or vacuum cleaner, and child’s use of electric heat in the room. In the largest study of leukemia (Hatch et al., 1998), mother’s use of heating pads and child’s use of hair dryers, curling irons, video arcade machine, and video games connected to televisions were associated to varying degrees with increased risk. In a study of childhood brain tumors, Gurney et al. (1996) reported relative risks of 1.5 or more based on five or more exposed cases for child’s exposure to portable black and white televisions, bedside digital clocks, incubators, and baby monitors. Dockerty et al. (1998) found such an association only for electric heating in the child’s bedroom in the 2 years before the reference date. Interpretation of the associations between electric appliances and childhood cancer is limited by the uncertainty regarding the relationship between self-reported use and actual exposure, susceptibility to reporting biases, and confounding by socioeconomic status and other factors. Electric blankets, the most thoroughly studied appliances with respect to exposure, underwent modifications approximately 10 to 15 years ago that markedly reduced magnetic field exposures, but the age of the blanket and thus the actual exposure in these studies is uncertain. Certainly parents of children who have developed cancer would

313

Electromagnetic Fields and Radiofrequency Radiation Table 17–2. Electric Blanket and Heated Water Bed Use and Childhood Cancer Reference

Study Location

Electric Blankets

Heated Water Bed

leukemia IN UTERO EXPOSURE Savitz et al., 1990 London et al., 1991 Dockerty et al., 1998 Hatch et al., 1998

Denver, Colorado area Los Angeles County New Zealand Midwest and Eastern US

1.3 (0.7–2.6) 1.2 (0.7–2.3) 0.8 (0.4–1.6) 1.6 (1.1–2.3)

0.3 (0.1–1.2) 0.7 (0.3–1.3) 0.6 (0.3–1.5) 0.9 (0.7–1.2)

Denver, Colorado area Los Angeles County New Zealand Midwest and Eastern US

1.5 (0.5–5.1) 7.0 (0.9–121.8) 2.2 (0.7–6.4) 2.8 (1.5–5.0)

0.7 (0.2–2.5) 1.0 (0.5–2.3) 0.8 (0.3–2.7) 1.2 (0.9–1.6)

Denver, Colorado area West Coast of US Seattle New Zealand

1.8 (0.9–4.0) 0.9 (0.5–1.2) 0.9 (0.5–1.6) 1.6 (0.6–4.3)

0.5 (0.2–2.0) 0.9 (0.6–1.3) 0.7 (0.4–1.3) 0.4 (0.1–1.9)

Denver, Colorado area West Coast of US Seattle New Zealand

1.2 (0.3–5.7) 1.0 (0.6–1.7) 0.5 (0.2–1.4) 1.6 (0.4–1.4)

0.3 (0.1–2.7) 1.2 (0.7–2.0) 0.8 (0.3–1.9) 5.5 (0.4–85.4)

postnatal exposure Savitz et al., 1990 London et al., 1991 Dockerty et al., 1998 Hatch et al., 1998

cns tumors IN UTERO EXPOSURE Savitz et al., 1990 Preston-Martin et al., 1996 Gurney et al., 1996 Dockerty et al., 1998 POSTNATAL EXPOSURE Savitz et al., 1990 Preston-Martin et al., 1996 Gurney et al., 1996 Dockerty et al., 1998

be susceptible to more complete reporting of electric appliance use, possibly even through some familiarity with the hypothesis regarding magnetic fields. For a number of associations (e.g., black and white televisions), the role of socioeconomic status may be critical, and the well-known tendency for control selection by random-digit dialing to under-represent the poorest controls (Olson et al., 1992, 2000) may not have been fully addressed in the analysis. Finally, to the extent that children who go on to develop leukemia have an increased tendency towards symptoms or illnesses that might result in electric blanket use, there could be bias in the reported associations. Although not easily addressed because of the rarity of the exposure, the association between childhood use of electric blankets and leukemia would warrant further examination.

Adult Cancers: Use of Electric Blanket or Other Appliances Because breast cancer is a very active topic of epidemiologic research, driven by its profound public health impact, and electric blanket use is easy to incorporate into questionnaires, there have been a sizable number of large studies with pertinent results (Table 17–3). The initial evaluation by Vena et al. (1991, 1994) suggested a small increased relative risk (around 1.5) for those women who used electric blankets throughout the night. The subsequent six studies generated effect esti-

mates very close to the null, but it should be noted that a number of these did not distinguish between women who used the blanket throughout the night and those who used the electric blanket only to warm the bed initially. Nevertheless, this replicated epidemiologic evidence has largely put the concern with electric blanket use and breast cancer to rest. Appliance use has been examined sporadically in relation to other cancers, including leukemia (Preston-Martin et al., 1988; Lovely et al., 1994), testicular cancer (Verreault et al., 1990), and prostate cancer (Zhu et al., 1999). The most suggestively positive study (Lovely et al., 1994) linking hand-held electric motors (e.g., shavers, hair dryers) to acute non-lymphocytic leukemia, was found to be solely due to the responses of proxy respondents, suggesting exposure misclassification (Sussman and Kheifets, 1996). A small association (odds ratio = 1.4, 95% CI: 0.9–2.2) was found for prostate cancer (Zhu et al., 1994). The research on these other cancers is insufficient for drawing conclusions or even for providing much guidance regarding the need for additional research.

Occupational Exposure The study of occupational exposure to electric and magnetic fields and cancer began with a series of reports linking job titles to mortality data bases (Milham, 1982) or cancer registries (Wright et al., 1982), and

Table 17–3. Electric Blanket Use and Breast Cancer Reference

Study Location

Case Definition

Exposure Measure

Main Results

Ever/never use Continuous through night Ever/never use Continuous through night Regular use of bed heating Ever/never use Continuous through night Ever/never use Ever/never use Continuous through night Ever/never use Ever/never use of bed warming

1.2 (0.8–1.7) 1.4 (0.9–2.2) 1.1 (0.9–1.4) 1.6 (0.8–3.1) 1.0 (0.7–1.4) 1.0 (0.9–1.2) 1.0 (0.9–1.2) 1.1 (1.0–1.2) 0.9 (0.7–1.1) 0.9 (0.7–1.2) 0.9 (0.8–1.1) 1.1 (0.8–1.3)

Vena et al., 1994

Western New York

Premenopausal

Vena et al., 1995

Western New York

All cases

Coogan and Aschengrau, 1998 Gammon et al., 1998

Cape Cod, Massachusetts Atlanta, New Jersey, Seattle

All cases Premenopausal

Laden et al., 2000 Zheng et al., 2000

United State (Nurses Health Study) Connecticut

All cases All cases

McElroy et al., 2001 Davis et al., 2002

Massachusetts, New Hampshire, Wisconsin Seattle, Washington

All cases All cases

314

PART III: THE CAUSES OF CANCER

consolidation of such sources to seek patterns of association (Savitz and Calle, 1987). From the outset, the cancer of greatest interest was leukemia, with a sustained but secondary interest in brain cancer. The subsequent 20 years produced a large volume of research, with metaanalyses in the early 1990s including 38 studies of EMF and leukemia (Kheifets et al., 1997) and 29 studies of EMF and central nervous system cancers (Kheifets et al., 1995). Despite the advantage of a large numbers of cases, these studies based solely on job title are seriously limited regarding the strength of inferences that can be drawn. The primary concern is the validity of job title as a marker of EMF exposure. Although intuition and a general familiarity with job tasks would suggest that certain jobs must have elevated EMF exposure on average, this assumption may be in error in some cases, there is no quantification possible for examining cumulative exposure, and it is not possible to incorporate more detailed characteristics of the workplace or job tasks to refine exposure assessment. Furthermore, the potential for confounding by other workplace exposures associated with those same jobs used to indicate EMF exposure cannot be addressed through job title information alone. Starting in the early 1990s, a series of studies with more sophisticated exposure assessment began to appear in the epidemiologic literature. In general, these studies focused on large occupational sectors, such as the electric utility industry (Sahl et al., 1993; Thériault et al., 1994), or electric railway workers (Tynes et al., 1994), and developed job-exposure matrices for workers in that setting. Expert judgment combined with measurements was used to link job titles, tasks, and locations, with exposure to electric and magnetic fields as well as to other potentially carcinogenic agents as potential confounding factors. This approach, applied to a single industrial sector, offers a substantial advantage in developing the needed expertise and measurements for making more accurate exposure assignments. The major disadvantage can be limited precision for studying such rare cancers as leukemia and brain cancer. In addition, in many but not all geographic settings, the evaluation was restricted to mortality rather than cancer incidence.

Occupational Exposure and Leukemia The 38 pertinent studies included in the meta-analysis by Kheifets et al. (1997) provided indications of a very small increased risk of leukemia in potentially exposed occupations, with relative risks on the order of 1.2 to 1.4. There was little discernible pattern of results across measures of study quality, level of EMF exposure, or leukemia subtype. Unfortunately, despite an increase in the volume and quality of the research during the 1990s (Table 17–4), this ambiguous pattern persisted. Across 10 studies that included reasonably large numbers of workers and intensive exposure assessment efforts, the pattern of a marginally increased risk associated with exposure persisted. Selected studies appeared to provide a much clearer indication of increased risk with increasing magnetic field exposure, most notably a study of the general population of Sweden in which a job-exposure matrix was applied (Floderus et al., 1993; Feychting et al., 1997) and the CanadaFrance study of electric utility workers (Thériault et al., 1993). However, studies of similar size and quality (Sahl et al., 1993; Savitz and Loomis, 1995; Johansen and Olsen, 1999) did not find any increased risk of leukemia with increased exposure. Although the data do not lend themselves to a meta-analysis, given the modest number of studies and diversity of methods, one effort was made to pool the results from three large studies of electric utility workers (Sahl et al., 1993; Thériault et al., 1994; Savitz and Loomis, 1995) that had obtained ostensibly divergent results (Kheifets et al., 1999). When integrated using a common protocol, the pattern was suggestive of a very weak association, with a relative risk for leukemia of 1.09 per 10 mT-years (95% CI: 0.98–1.21). After a decade of intensive study, the pattern of association had changed little from the original job title studies—sporadic indications of stronger associations, but overall support for a very weak relationship, so small that it is difficult to draw inferences. What is much different now, however, is the conclusions that can be drawn from the failure of enhanced methods to yield notably stronger associations. This suggests either that the improve-

ments in exposure assignment and control for confounding are of little value or there is simply not an association of discernible magnitude present to be detected, both of which remain plausible candidate explanations.

Occupational Exposure and Brain Cancer The evolution of research and patterns of results are strikingly similar for brain cancer as for leukemia. A large number of studies addressed EMF and brain cancer by examining job titles in relation to cancer mortality or diagnosis. Although there is some variability across populations, the overall tendency is towards a small association, a relative risk on the order of 1.2 based on a meta-analysis (Kheifets et al., 1995). Once again, although there is some variability across studies, there is not a tendency for studies with more favorable methodologic features or for jobs more certain to have elevated exposures to show stronger associations. The nine studies with more sophisticated exposure assessment published during the 1990s recapitulate that pattern (Table 17–4). Some studies provide rather clear support for an association, at least with specific types of brain cancer (Floderus et al., 1993; Savitz and Loomis, 1995) whereas a preponderance of the studies suggests either no association or a negligibly weak one. The pooled analysis of the utility worker cohort in the United States, Canada, and France yielded a relative risk of 1.12 per 10 mT-years (95% CI: 0.98–1.28) (Kheifets et al., 1999). The conclusions from the literature on occupational EMF and brain cancer can only be the same as those for leukemia. There are isolated suggestions of strong associations, but taken in the aggregate, the evidence suggests either no effect or one that is below the level at which epidemiologic research is likely to be able to discern it with confidence.

Occupational Exposure and Breast Cancer Interest in EMF and breast cancer generated by the hypotheses of Stevens et al. (1992) has encouraged research into the role of EMF in breast cancer specifically, but also into the hypothesized mediator of such an effect, pineal melatonin production. The time course of the anticipated relationship between EMF and melatonin production lends itself to both experimental and observational studies, and there is a small literature on each. The experimental studies, in which humans are exposed to known levels of EMF but masked to their exposure condition, have generally not detected alterations in melatonin synthesis (Graham et al., 1996). The few studies of occupational EMF and melatonin production have yielded some indications of an association but only for peculiar indices of exposure (Burch et al., 1998) or combined with low sunlight exposure (Burch et al., 1999). The epidemiologic literature addressing breast cancer directly is limited in scope with some reports on male breast cancer raising concern (Tynes and Andersen, 1990; Demers et al., 1991) but with the evolution of the literature, the association between EMF and this outcome has not been corroborated (Rosenbaum et al., 1994; Stenlund and Floderus, 1997). Several studies have addressed occupational EMF and female breast cancer. A large database on occupation and mortality in the United States was used to examine risk of breast cancer associated with electrical occupations. An overall odds ratio of 1.4 (1.0–1.8) was found for electrical workers (Loomis et al., 1994). However, application of a slightly different analytic approach to the same data did not corroborate those findings (Cantor et al., 1995). A more extensive analysis of job titles associated with elevated EMF exposure and breast cancer was conducted in a large case-control study (Coogan et al., 1996), which permitted adjustment for potential confounding factors. The potential for reproductive risk factors for breast cancer to vary by occupation raises a plausible scenario of confounding. A high potential for occupational exposure, but not medium or low potential, was associated with a slightly increased risk (OR = 1.4, 95% CI: 1.0–2.1), more pronounced among premenopausal women (OR = 2.0, 95% CI: 1.0–3.8). Female utility workers in Denmark showed little evidence of increased breast cancer risk (SIR = 1.1, 95% CI: 0.9–1.3), with almost all such workers in the lowest magnetic field exposure category. A large cohort study of Swedish workers similarly found no

315

Electromagnetic Fields and Radiofrequency Radiation

Table 17–4. Summary of the Principal Studies of Occupational Electromagnetic Fields and Leukemia in Relation to Brain Cancer Using MeasurementBased Job Exposure Matrices Reference Setting and Industry Matanoski et al., 1993 US, telephone workers Floderus et al., 1993 Sweden, general population

Sahl and Kelsh, 1993 California, electric utility Theriault et al., 1994 Canada/France, electric utility Tynes et al., 1994 Norway, railway Savitz and Loomis, 1995 US, electric utility Guenel et al., 1996 France, electric utility Miller et al., 1996 Ontario, Canada, electric utility Feychting et al., 1997 Sweden, general population

Harrington et al., 1997 England, electric utility Rodvall et al., 1998 Sweden, general population Johansen and Olsen, 1998 Denmark, electric utility

Leukemia (RR, 95% CI) Exposure Level

Brain Cancer (RR, 95% CI) RR (95% CI)

Exposure Level

RR (95% CI)

>Median (mean)

2.5 (0.7–8.6)

Not available

2nd Quartile 3rd Quartile 4th Quartile CLL/2nd Quartile CLL/3rd Quartile CLL/4th Quartile >Median >99th %ile >Median >90th %ile CLL/>Median AML/>Median

0.9 (0.6–1.4) 1.2 (0.8–1.9) 1.6 (1.1–2.4) 1.1 (0.5–2.3) 2.2 (1.1–4.3) 3.0 (1.6–5.8) 1.0 (0.8–1.4) 1.1 (0.8–1.4) 1.5 (0.9–2.6) 1.8 (0.8–4.0) 1.5 (0.5–4.0) 3.2 (1.2–8.3)

2nd Quartile 3rd Quartile 4th Quartile

1.0 (0.7–1.6) 1.5 (1.0–2.2) 0.4 (0.9–2.1)

Low High Electric field—Low Electric field—High 30–<50th %ile 50–<70th %ile 70–<90th %ile >=90th %ile Electric fields >50–75th %ile >75–90th %ile >90th %ile Electric—>33–67th %ile Electric—>67th %ile >33–67th %ile >67th %ile 0.13–0.19 uT 0.20+ uT AML/0.13–0.19 uT AML/0.20+ uT CLL/0.13–0.19 uT CLL/0.20+ uT Not available

1.0 (0.4–2.2) 0.6 (0.2–1.3) 0.4 (0.2–1.1) 1.0 (0.5–2.2) 1.0 (0.7–1.6) 1.1 (0.7–1.8) 1.0 (0.6–1.6) 1.1 (0.6–2.1)

>Median >99th %ile >Median >90th %ile Astrocytoma/>Median Glioblastoma/>Median Benign tumors/>Median Low High Electric field—Low Electric field—High 30–<50th %ile 50–<70th %ile 70–<90th %ile >=90th %ile Electric fields >50–75th %ile >75–90th %ile >90th %ile Not available

1.0 (0.6–1.5) 0.8 (0.5–1.3) 1.5 (0.9–2.8) 2.0 (0.8–5.0) 1.5 (0.9–2.8) 1.3 (0.5–3.8) 2.3 (0.8–6.7) 0.8 (0.3–2.0) 0.9 (0.4–2.3) 0.7 (0.3–1.7) 1.2 (0.5–2.8) 1.6 (1.0–2.6) 1.5 (0.8–2.6) 1.7 (0.9–3.0) 2.3 (1.2–4.6)

0.13–0.19 uT 0.20+ uT

1.0 (0.7–1.6) 1.0 (0.6–1.7)

>33–67th %ile >67th %ile Glioma/0.2–0.4 uT Glioma/>0.4 uT

1.1 (0.6–2.0) 1.0 (0.5–1.9) 1.1 (0.4–2.7) 1.9 (0.8–5.0)

Background, Low, Medium, High

0.5, 0.9, 0.7, 0.7

1.0 (0.5–2.0) 0.7 (0.3–1.9) 0.4 (0.1–1.3) 2.1 (0.6–7.2) 4.5 (1.0–19.7) 1.7 (0.6–4.8) 1.6 (0.5–5.1) 1.4 (1.0–2.2) 1.7 (1.1–2.7) 2.1 (0.9–5.0) 2.7 (0.9–7.9) 1.4 (0.7–2.5) 1.9 (1.0–3.8)

Not available Background, Low, Medium, High

1.0, 1.0, 0.9, 1.1

2.5 (1.0–6.2) 1.4 (0.5–4.5) 3.1 (1.1–8.7)

AML, acute myeloid leukemia; CLL, chronic lymphocytic leukemia.

discernible increased risk of breast cancer in jobs associated with higher magnetic field exposure (Floderus et al., 1999). A comparable study in Norway reported slightly increased risk of incident breast cancer among women who worked in jobs in the highest versus lowest exposure category (RR = 1.14, 95% CI: 1.10–1.19) (Kliukiene et al., 1999). The one study to integrate consideration of residential and occupational magnetic field exposure (Forssén et al., 2000) found no association overall with breast cancer, but imprecise evidence for increased risks among women under age 50, particularly those with estrogenreceptor positive tumors. Magnetic field exposures were estimated based on a measurement survey in conjunction with a case-control study in North Carolina to allow for individual exposure estimation (van Wijngaarden et al., 2001). Estimated exposures among breast cancer cases and controls were similar, overall, with some suggestion of an association for exposures 10 to 20 years before diagnosis. Within that window, premenopausal women had odds ratios of 1.4 to 1.7 for above-background exposure, but there was no dose-response gradient. Cumulatively, the literature on occupational EMF exposure and breast cancer is quite limited, with studies thus far either relying on large

databases or inferring exposure as an add-on to case-control studies. Nonetheless, there is some convergent support for a weak positive association, with modestly elevated relative risks for premenopausal women.

Summary and Interpretation The enhanced sophistication over the past 25 years in epidemiologic methods applied to the study of magnetic fields and cancer is notable, particularly residential exposure and childhood leukemia, and occupational exposure and adult leukemias and brain cancer. The principal challenge was and remains exposure assessment, and as a result the study protocols for assessing residential exposure have expanded to include sampling over prolonged periods of time and evaluation in relation to personal monitoring data, combined with detailed evaluation of power lines and related electrical constructions. Analogously, occupational exposure assessment includes extensive measurement protocols that can be used by experts to develop job-exposure matrices. In both settings, residential and occupational, extraordinarily large studies have been conducted, sufficient to generate rather precise risk

316

PART III: THE CAUSES OF CANCER

estimates even for these rare cancers of interest. In many ways, the results of these efforts are analogous as well. The evidence regarding residential exposure and childhood leukemia suggests that any discernible effect of magnetic fields is to be found in the upper reaches of the observed exposure distribution, above 0.2 mT, quite possibly above 0.3 or 0.4 mT. Unfortunately, in that range, even the largest studies generate rather imprecise findings, so that further progress would require new study settings that are more favorable regarding the proportion of homes with elevated exposure. Those studies would have to retain the strengths or improve on the structural features of previous studies regarding control selection and non-response. Despite all the research conducted on this topic, the question of whether residential magnetic fields cause childhood leukemia is unresolved and may well remain so. Observed associations appear not to be large and are not discernible in the range of exposure that is common in the population, yet the potential for an effect across a wider exposure range or one that is stronger than what has been observed to date cannot be discounted with confidence. Unfortunately, because the methods applied to this topic have become so sophisticated, prospects for notably better research in the foreseeable future are limited. The strongest links between occupational electric and magnetic field exposures and cancer have been found for leukemia and brain cancer. In both instances, both the preliminary studies based on job title and more sophisticated assessments have yielded mixed findings. The range, however, is generally between a modest effect, relative risks of 1.5 or less, and no effect. Once again, given the body of large, detailed studies, and the goal of distinguishing between modest increments in these cancers and no increment, it is difficult to have much optimism that new studies in the near future will markedly change the overall state of evidence. There are avenues of inquiry in which the research is less advanced, and thus more in a state of flux. Breast cancer, notably more common than childhood or adult leukemias or brain cancers, has come into focus. The extensive literature on electric blanket use indicates no association with breast cancer. A much more limited literature on residential exposure also is reassuring of an absence of association thus far, and the similarly limited research on occupational exposure and breast cancer contains a few preliminary suggestions of potential for increased risk among women in jobs with elevated exposure. While limited efforts have been made to address a wide range of other cancers through occupational cohort studies, case-control studies of residential exposure, and the incorporation of information on electric blankets in case-control studies, none of these lines of research has much momentum at this time.

RADIOFREQUENCY RADIATION Exposure Assessment Assessment of radiofrequency radiation exposure is perhaps even more complex than assessment of exposure to power-frequency electric and magnetic fields. A range of frequencies can be encountered from a diverse array of exposure sources, including radar and specialized industrial equipment (e.g., heat sealing), television, and radio transmission towers, and use of mobile telephones and related modes of wireless communication. For sources that are close to those exposed, such as mobile telephones, the exposure varies considerably across different parts of the body. Furthermore, the exact physical nature of the exposure changes as technology changes. For example, mobile telephone communication had been primarily analog until recently when it changed to digital, potentially obviating the relevance of all prior epidemiologic research that pertained to persons exposed to cellular telephones using analog mode. It is common in epidemiology to have exposures that are extremely complex if the goal is a precise moment-to-moment quantification, but for many concerns, we are able to identify markers or indices that adequately capture the chronic exposure of interest. In regard to radiofrequency radiation, it is not yet clear how effective we have been in finding such “handles” on exposure.

Exposure Sources and Populations Studied Residential Exposure Residential proximity to broadcast towers, including those used for transmitting radio, television, microwave, and mobile telephone communications, have raised concern regarding the potential for increased risk of cancer. The public has a long history of concern and resistance to the siting of such towers, for reasons involving aesthetics and property value, as well as health concerns. Much of the epidemiologic research has been conducted in response to such concerns, driven either solely by the exposure source or by a perceived cancer cluster among persons living in the vicinity. While living closer to such towers may, on average, be associated with greater exposure to the agent of interest, the spatial distribution of such exposures is much more complex, depending on topography, physical barriers between the tower and the home, and shielding by the home itself. A great temptation in studies addressing towers and cancer is to be inclusive of populations that are sufficiently large to study such rare outcomes as childhood leukemia, but to achieve this needed population size, a large number of persons with little or no exposure are included. One of the first studies of this issue applied sophisticated statistical techniques to address childhood leukemia in relation to a microwave tower southwest of San Francisco (Selvin et al., 1992). Childhood leukemia was also the focus in a perceived cluster in Hawaii, with an increased risk found for children living within 2.6 miles of radio towers (odds ratio = 2.0, 95% CI: 0.1–8.3) (Maskarinec et al., 1994). This study illustrates both the challenge of severely inadequate study size and the selectivity in defining the geography of reported clusters. A more systematic and sophisticated assessment of cancer incidence in relation to television towers was conducted in Sydney, Australia (Hocking et al., 1996). The investigators attempted to calculate the power density at varying distances from the source, with an indication that at the center the power density is around 1 mW/cm2 whereas it increased to a maximum of 8 mW/cm2 at 2 km from the center; clearly, distance from the source is a poor proxy for such exposures. Leukemia incidence and mortality in both children and adults was modestly elevated comparing the populations in the vicinity of the towers to more distant populations, with relative risk estimates in the range of 1.2 to 1.7 and good precision for these rather sizable populations. Brain tumors did not show such a pattern. Leukemia and lymphoma incidence in relation to radio and television transmitters in Great Britain was evaluated by Dolk et al. (1997a,b). A reported cluster in proximity to the Sutton Coldfield transmitter prompted a detailed small area analysis that did in fact suggest increased risk of adult leukemia proximal to the tower with decreasing incidence with increasing distance. Childhood leukemia was too rare for meaningful evaluation. However, an evaluation of 20 such television and radio transmitters with similar analytic methods did not show such a pattern for adult or childhood leukemia. An updated evaluation that included more recent cancer incidence data in relation to the Sutton Coldfield transmitter showed a markedly weaker association (Cooper et al., 2001). Communication towers from the Vatican Radio station, which transmits worldwide, raised public concern in nearby neighborhoods (Michelozzi et al., 2002). Very small numbers of leukemia cases were observed in proximity to the towers, with one childhood leukemia and two adult leukemias within 2 km of the tower. Only among adult males was there any indication of diminishing cancer occurrence with increasing distance, providing little support (or evidence) pertaining to the concern with radio transmission towers and cancer incidence. Cumulatively, the research on community exposures to radiofrequency radiation and cancer is at a very early stage of development. Diverse exposure sources, poorly estimated population exposures, small numbers of cases, and selective investigation (and quite possibly selective publication) result in a body of literature that is of limited value. Within the bounds of that uncertainty, there are some suggestions of a possible link to leukemia (Hocking et al., 1996; Dolk et al., 1997a), but more systematic investigations of populations chosen based solely on their potential exposures and not on prior indication of a cancer cluster (Selvin et al., 1992; Dolk et al., 1997b) provide

317

Electromagnetic Fields and Radiofrequency Radiation evidence against the hypothesis of increased risk associated with proximity to towers. Subsequent studies must combine more systematic exposure assessment with adequate population sizes, and ideally consider multiple geographic areas to avoid bias from reported clusters.

Mobile Telephones Epidemiologic studies in four different countries have evaluated cancer risk among mobile phone users (Table 17–5). As noted previously, exposure assessment is profoundly difficult for a number of reasons. The studies available to date have used two different approaches: three research groups have exploited the network operators’ subscriber lists to identify users and to classify the users according to duration of use (Rothman et al., 1996a; Johansen et al., 2000, 2001; Auvinen et al., 2002), while three other groups have conducted case-control studies with mobile telephone use information collected through interviews or questionnaires (Hardell et al., 1999; Muscat et al., 2000; Inskip et al., 2001). The studies that use subscriber lists run small risks of selection bias and also of differential information bias. On the other hand, the operator data omit information on the use of hands-free devices or whether the listed person is indeed the principal user of the telephone. Because of this exposure misclassification, the possibility of diluted effects, if there are any, is substantial. The United States subscriber list study is restricted to a short followup period; an extended follow-up was made impossible due to legal restrictions driven by concerns with confidentiality. Only a 1-year follow-up was possible, making the substantive findings of little value (Rothman et al., 1996; Dreyer et al., 1999), but the work has resulted in two important methodological papers (Rothman et al., 1996b; Funch et al., 1996). The Danish study identified over 400,000 subscribers that were matched to the national cancer registry for the period of 1982 to 1996 (Johansen et al., 2001). The total number of cancers in the cohort was over 3000 and the number of brain tumors was 135

in males and 19 in females. The brain tumor standardized incidence ratio values were 0.95 (0.79–1.12) and 1.03 (0.62–1.61) for men and women, respectively. These results appear reassuring but the interpretation is hampered by the fact that there are few long-term users, it is unknown whether the subscriber was in fact also the user, and the amount of use is unknown. The Finnish subscriber list study has similar strengths and limitations (Auvinen et al., 2002) and is of similar size. Although the results were largely negative, there was an increased risk for glioma among analogue telephone users (RR = 2.0; 1.0–4.1) and for analogue and digital users combined (RR = 1.7; 0.9–3.5). When assessing the interview and questionnaire case-control studies, both selection bias and differential information bias must be considered as a serious threat to validity in all studies. The Swedish group has done essentially two different studies, resulting in four publications (Hardell et al., 1999, 2001, 2002a,b). The methodology in the first study (Hardell et al., 1999; Hardell et al., 2001) has been questioned on several grounds. Their stated response proportion, in excess of 90%, was challenged since the number of reported cases was only one-third of the numbers in the official cancer registry (Ahlbom and Feychting, 1999); this issue was never satisfactorily resolved (Hardell et al., 1999). Second, since they reported no overall excess risk of brain tumors in total but only for tumors on the same side as the telephone was used, it was questioned whether recall bias might be responsible for all or part of this excess risk (Rothman, 2000). The latest reports of this group do not have obvious methodological problems (Hardell et al. 2002 a,b). Again they report an increased risk in the temporal area on the side of the head where the phone was used (OR = 2.5, 95% CI: 1.3–4.9); the tumor type with highest relative risk was acoustic neuroma (OR = 3.5, 95% CI: 1.8–6.8). The other two groups conducted studies in the United States and both groups relied on cases identified in hospitals and on hospital controls (Inskip et al et al., 2001;

Table 17–5. Mobile Telephone Use and Cancer Reference Rothman et al., 1996 Dreyer et al., 1999

Hardell et al., 1999/ Hardell et al., 2001

Muscat et al., 2000

Inskip et al., 2001

Johansen et al., 2001 Auvinen et al., 2002

Muscat et al., 2002 Hardell et al., 2002a,b

Study Population and Design Mobile phone use subscribers in 1994 followed 1 year in mortality registry. Size of cohort = 256,284. See Rothman et al., 1996. One network company added. Size of cohort = 285,561. SMR analysis based on those with portable antennas. The Uppsala/Örebro region of Sweden 1994–1996 and Stockholm region 1995–1996. Case-control study with population controls and mailed questionnaire. Hospital-based case-control study between 1994–1998 with personal interviews. Cases = 469, controls = 422. Hospital-based case-control study on intracranial tumors between 1994– 1998 with personal interviews. Cases = 782, controls = 799. All users in Denmark during 1982– 1995; linkage with cancer registry. Size of cohort = 420,095. Case control study: cases from cancer registry in 1996 and controls from population registry. Information on cellular phone use obtained from network companies. Cases = 398, controls = 398. See Muscat et al., 2000 Sweden case-control study, 1997– 2000, controls from population registers; 1303 total tumors, 159 acoustic neuromas.

Exposure

Outcome

Results; RR (95% CI)

Portable antenna vs. mobile antenna

Total mortality

0.86 (0.47–1.53) [90% CI]

Frequency of use (min/d) and length of service (y)

Cause-specific mortality

Use of mobile phone in total and ipsilateral use

Malignant brain tumor

No association for total mortality or total cancer mortality; too few cases for specific diagnoses 2.42 (0.97–6.05) (anatomic areas w/highest exposure)

Ever use, frequency and length of use

Brain cancer

0.7 (0.3–1.4) for frequent use (>10.1 h/mo)

Length of use

Glioma, meningioma, acoustic neuroma

Ever use, length of use

Brain and nervous tumors; leukemia

0.9 (0.5–1.6) (glioma) 0.7 (0.3–1.7) (meningioma) 1.4 (0.6–3.5) (ac neuroma) (>100 h) 1.2 (0.6–2.3) (brain and nervous) (≥3 y)

Length of use

All brain tumors, glioma, meningioma, salivary gland

1.5 (0.9–2.5) (all brain) 1.7 (0.9–3.5) (glioma)

See above Use of mobile phone

Acoustic neuromas All tumors

0.9 (regular use) 1.3 (1.0–1.6) (analogue) 1.0 (0.8–1.2) (digital) 3.5 (1.8–6.8) (analogue) 1.2 (0.7–2.2) (digital)

Acoustic neuroma

318

PART III: THE CAUSES OF CANCER

Muscat et al., 2000), leaving them vulnerable to selection bias. Both these studies were entirely negative with regard to any indications of an association between mobile telephone use and brain tumors. Muscat and colleagues (2002) also published a paper on neuromas that did not identify an association with mobile telephone use. The results in these nine reports based on six distinct studies provide an overall indication that use of mobile telephones is not associated with an increased risk of cancer (Table 17–5). The only exceptions are the glioma excess risk in the Finnish study and the acoustic neuroma increase in the Swedish study. Individually and collectively, the studies have methodologic limitations precluding firm conclusions. Beyond the technical aspects of the studies, it must be remembered that the exposures of interest are with a form of technology that may differ from the one in place at present. So long as the technology evolves, epidemiologic evidence will be a step behind in evaluating the exposures of current concern. In addition, there is an inherent inability to address health effects that may become manifest only with a long latency. Widespread use of mobile telephones is still a relatively recent phenomenon and only with the passage of time can we begin to ask about effects that occur after a 20- or 30-year delay.

Occupational Exposure The research addressing potential cancer risks in relation to radiofrequency radiation is also at a very early stage of development, with studies thus far limited in their sophistication in exposure assessment and many of them severely limited regarding precision. Attention has focused on leukemia and brain cancer, analogous to the literature on ELF electric and magnetic field exposure. However, in contrast to the relatively large populations with documented occupational exposure to ELF electric and magnetic fields, there are few studies of large populations with clear or even likely exposure to radiofrequency radiation. The job groups included in the tabulation pertaining to radiofrequency radiation exposure and leukemia and brain cancer (Table 17–6) are subject to uncertainty, with some more likely to have consistently elevated exposure to radiofrequency radiation than others (Swerdlow, 1999). Workers who routinely use radar (Groves et al., 2002) and dielectric radiofrequency heat sealers (Lagorio et al., 1997) are more certain to have some exposure above background than are general electronics technicians (Garland et al., 1990), for example, though without a systematic measurement protocol, it is difficult to prove that this is so. Beyond the reports included in the table, there are many more that have generated estimates of associations between jobs with possible radiofrequency exposure and cancer from routinely collected data. In fact, there is substantial overlap with the tabulations used to make inferences regarding ELF electric and magnetic field exposure as synthesized by Kheifets et al. (1995; 1997). Such groups as radio and television repairmen or radio and telegraph operators are included (Wright, 1983; Calle and Savitz, 1985; Milham, 1985, 1988a,b; Pearce

et al., 1989) and could be argued to incur radiofrequency radiation exposure as well as power-frequency electric and magnetic field exposure. The discussion here is focused on those papers that at least included expert evaluation and inference regarding radiofrequency radiation exposure potential, if not actual workplace measurements. Collectively, the evidence regarding both leukemia and brain cancer is severely limited by small numbers of cases as well as problems in assessing exposure. The largest and most sophisticated studies (Morgan et al., 2000; Groves et al., 2002) provide very little support for an association, with the most suggestive finding the relative risk of 1.5 (95% CI: 1.0–2.2) comparing US Navy personnel who were exposed to radar to those not exposed. Inverse associations of comparable magnitude were found for brain cancer in that study. Isolated evidence for an association with brain cancer has been reported as well (Thomas et al., 1987), but the large association reported by Szmigielski (1996) is based on an incompletely documented study that is therefore difficult to evaluate. The research thus far provides very limited evidence regarding an association between radiofrequency radiation and leukemia or brain cancer. Other cancer sites have been explored sporadically. Both breast cancer (RR = 1.5, 95% CI: 1.1–2.0) and endometrial cancer (RR = 1.9, 95% CI: 1.0–3.2) were found to be increased among women who worked as radio and telegraph operators on Norwegian merchant ships (Tynes et al., 1996), but the potential for confounding by reproductive factors could not be addressed fully. Other studies have been limited in size but have not tended to corroborate this finding for breast cancer (Lagorio et al., 1997; Morgan et al., 2000; Groves et al., 2002). Testicular cancer was increased among men with self-reported occupational exposure to microwaves and other radio waves (OR = 3.1) but not for such exposure based on the inferences of an industrial hygienist (Hayes et al., 1990). Testicular cancer was a concern based on a cluster among police officers who used hand-held traffic radar guns (Davis and Mostofi, 1993), and often rest such devices in their lap when not in use. Ocular melanoma was linked to self-reported microwave exposure or radar, with an odds ratio of 2.1 (95% CI: 1.1–4.0) (Holly et al., 1996) and to occupational exposure to radio sets, with an odds ratio of 3.3 (95% CI: 1.2–9.2) (Stang et al., 2001). In isolation, these reports raise some interest that might be incorporated in future studies of those cancers, but the findings are often isolated and largely not addressed directly in other studies. Self-reported exposure to radiofrequency radiation in case-control studies is of unproven validity, and likely to be rather poor.

SUMMARY AND INTERPRETATION Each of the three avenues of inquiry pertaining to carcinogenicity of radiofrequency radiation exposure—residential exposures from towers,

Table 17–6. Studies of Occupational Exposure to Radiofrequency Radiation in Relation to Leukemia and Brain Cancer Reference

Location

Thomas et al., 1987 Milham, 1988 Garland et al., 1990

US US US

Muhm et al., 1992 Tynes et al., 1996 Szmigielski, 1996 Grayson, 1996 Lagorio et al., 1997 Morgan et al., 2000

US Norway Poland US Italy US

Groves et al., 2002

US

Exposure

Outcome

Leukemia RR (95% CI)

Brain Cancer RR (95% CI)

Job title and interpretation Amateur radio operators Electronics technician Aviation technician Fire control technician Electromagnetic pulse Radio and telegraph operators Military personnel Air Force personnel Dielectric radiofrequency heat sealers Motorola workers (vs. US) <Median vs. unexposed >=Median vs. unexposed Navy personnel (vs. US) Exposed vs. unexposed

Mortality Mortality Incidence Incidence Incidence Incidence Incidence Incidence Incidence Mortality Mortality Mortality Mortality Mortality Mortality

— 1.2 (0.9–1.7) 1.1 (0.4–2.5) 0.3 (0.0–1.9) 0.3 (0.0–2.5) 5.4 (0.7–19.7) 1.1 (0.1–4.1) 7.7 — — * 0.8 (0.4–1.4) 0.6 (0.3–1.3) 0.6 (0.3–1.0) 1.0 (0.8–1.2) 1.5 (1.0–2.2)

1.6 (1.0–2.4) 1.4 (0.9–2.0) — — — — 1.0 (0.3–2.3) 1.9 (1.1–3.5) 1.4 (1.0–1.9) * 0.5 (0.2–1.1) 1.0 (0.4–2.2) 0.9 (0.4–1.9) 0.9 (0.7–1.1) 0.6 (0.4–1.0)

*Insufficent number of cases for analysis (1 leukemia, 1 brain cancer case).

Electromagnetic Fields and Radiofrequency Radiation mobile telephone use, and occupational exposure—is at a very early stage of development and far from the point of being capable of yielding firm evidence. Only for mobile telephone use is there sufficient information to offer even a guarded inference that strong associations based on past technology and a short latency are unlikely. The other avenues of research are so weak regarding exposure assessment and study size that the hypothesis of an association with cancer remains largely unaddressed. Residential exposure studies suggest that proximity to broadcast towers is not associated with markedly increased risks of certain cancers, but given how poorly distance corresponds to exposure level, more accurate exposure assignment could well yield a different pattern of results. Occupational studies are likewise so fraught with uncertainty regarding exposure classification that it is difficult to take much comfort in the lack of clear indications of an effect thus far. The degree of urgency for continued epidemiologic evaluation of the quality required to more effectively address the potential associations of interest is a question to be considered. On the one hand, the biologic basis for potential carcinogenic effects is limited, though the small margin between current exposures to mobile telephones and physiologic effects makes a slightly stronger case on the basis of biophysics than the concern with ELF electric and magnetic fields. The greatest justification for continued scrutiny, however, is not biophysics but the widespread dissemination of these new technologies. This rapid change produces increased exposure of the general population to cellular communication towers and to hand-held electronic communication devices. There is no end to such expansion in sight and the evolving technology will keep changing the characteristics of the agent of concern. The most compelling basis for epidemiologic scrutiny of these novel physical exposures is in public health surveillance—to monitor the health experience of populations to detect unanticipated, in fact unlikely, adverse effects of a new technology. Providing scientific answers to public and policy-driven concerns is a legitimate role of epidemiology, even if it is unlikely to greatly advance knowledge of cancer etiology and prevention. References Ahlbom A, Feychting M. 1999. Re: Use of cellular phones and the risk of brain tumours: A case-control study [letter]. Int J Oncol 15:1045. Ahlbom A, Cardis E, Green A, Linet M, Savitz D, Swerdlow A. 2001. Review of the epidemiologic literature on EMF and health. Environ Health Perspect 109:911–933. Ahlbom A, Day N, Feychting M, Roman E, Skinner J, Dockerty J, Linet M, McBride M, Michaelis J, Olsen JH, Tynes T, Verkasalo P. 2000. A pooled analysis of magnetic fields and childhood leukemia. Br J Cancer 83:692–698. Ahlbom A, Feychting M. 1999. Re: Use of cellular phones and the risk of brain tumours: A case-control study. Int J Oncol 15:1045. Armstrong BG, Deadman JE, Thériault G. 1990. Comparison of indices of ambient exposure to 60-Hertz electric and magnetic fields. Bioelectromagnetics 11:337–347. Auvinen A, Hietanen M, Luukkonen R, Koskela R-S. 2002. Brain tumors and salivary gland cancers among cellular telephone users. Epidemiology 13:356–569. Burch JB, Reif JS, Yost MG, Keefe TJ, Pitrat CA. 1998. Nocturnal excretion of a urinary melatonin metabolite among electric utility workers. Scand J Work Environ Health 24:183–189. Burch JB, Reif JS, Yost MG, Keefe TJ, Pitrat CA. 1999. Reduced excretion of a melatonin metabolite in workers exposed to 60 Hz magnetic fields. Am J Epidemiol 150:27–36. Calle EE, Savitz DA. 1985. Leukemia in occupational groups with presumed exposure to electrical and magnetic fields. N Engl J Med 313:1476–1477. Campion EW. 1997. Power lines, cancer, and fear. N Engl J Med 337:44– 46. Cantor KP, Dosemeci M, Brinton LA, Stewart PA. 1995. Re: Breast cancer mortality among female electrical workers in the United States [letter]. J Natl Cancer Inst 87:3. Coogan PF, Aschengrau A. 1998. Exposure to power frequency magnetic fields and risk of breast cancer in the Upper Cape Cod Cancer Incidence Study. Arch Environ Health 53:359–367. Coogan PF, Clapp RW, Newcomb PA, Wenzl TB, Bogdan G, Mittendorf R, Baron JA, Longnecker MP. 1996. Occupational exposure to 60-hertz magnetic fields and risk of breast cancer in women. Epidemiology 7:459–464.

319

Cooper D, Hemmings K, Saunders P. 2001. Re: Cancer incidence near radio and television transmitters in Great Britain. I. Sutton Coldfield transmitter; II. All high power transmitters. Am J Epidemiol 153:202–204. Davis S, Mirick DK, Stevens RG. 2002. Residential magnetic fields and the risk of breast cancer. Am J Epidemiol 155:446–454. Davis RL, Mostofi FK. 1993. Cluster of testicular cancer in police officers exposed to hand-held radar. Am J Indust Med 24:231–233. Demers PA, Thomas DB, Rosenblatt KA, Jimenez LM, McTiernan A, Stalsberg H, Stemhagen A, Thompson WD, Curnen MG, Satariano W, Austin DF, Isacson P, Greenberg RS, Key C, Kolonel LN, West DW. 1991. Occupational exposure to electromagnetic fields and breast cancer in men. Am J Epidemiol 134:340–347. Dockerty JD, Elwood JM, Skegg DCG, Herbison GP. 1998. Electromagnetic field exposures and childhood cancers in New Zealand. Cancer Causes Control 9:299–309. Dolk H, Elliott P, Shaddick G, Walls P, Thakrar B. 1997. Cancer incidence near radio and television transmitters in Great Britain .2. All high power transmitters. Am J Epidemiol 145:10–17. Dolk H, Shaddick G, Walls P, Grundy C, Thakrar B, Kleinschmidt I, Elliott P. 1997. Cancer incidence near radio and television transmitters in Great Britain .1. Sutton Coldfield transmitter. Am J Epidemiol 145:1–9. Dreyer NA, Loughlin JE, Rothman KJ. 1999. Cause-specific mortality in cellular telephone users. J Am Med Assoc 282:1814–1816. Feychting M, Ahlbom A. 1993. Magnetic fields and cancer in children residing near Swedish high-voltage power lines. Am J Epidemiol 138:467–481. Feychting M, Ahlbom A. 1994. Magnetic fields, leukaemia, and central nervous system tumors in Swedish adults residing near high-voltage power lines. Epidemiology 5:501–509. Feychting M, Forssen U, Floderus B. 1997. Occupational and residential magnetic field exposure and leukemia and central nervous system tumors. Epidemiology 8:384–389. Feychting M, Forssén U, Rutqvist LE, Ahlbom A. 1998. Magnetic fields and breast cancer in Swedish adults residing near high-voltage power lines. Epidemiology 9:392–397. Floderus B, Persson T, Stenlund C, Wennberg A, Ost A, Knave B. 1993. Occupational exposure to electromagnetic fields in relation to leukemia and brain tumors: A case-control study in Sweden. Cancer Causes Control 4:465–476. Floderus B, Stenlund C, Persson T. 1999. Occupational magnetic field exposure and site-specific cancer incidence: A Swedish cohort study. Cancer Causes Control 10:323–332. Forssén UM, Feychting M, Rutqvist LE, Floderus B, Ahlbom A. 2000. Occupational and residential magnetic field exposure and breast cancer in females. Epidemiology 11:24–29. Fulton JP, Cobb S, Preble L, Leone L, Forman E. 1980. Electrical wiring configurations and childhood leukemia in Rhode Island. Am J Epidemiol 111:292–296. Funch DP, Rothman KJ, Loughlin JE, Dreyer NA. 1996. Utility of telephone company records for epidemiologic studies of cellular telephones. Epidemiology 7:299–302. Gammon MD, Schoenberg JB, Britton JA, Kelsey JL, Stanford JL, Malone KE, Coates RJ, Brogan DJ, Potischman N, Swanson CA, Brinton LA. 1998. Electric blanket use and breast cancer risk among younger women. Am J Epidemiol 148:556–563. Garland FC, Shaw E, Gorham ED, Garland CF, White MR, Sinsheimer PJ. 1990. Incidence of leukemia in occupations with potential electromagnetic field exposure in United States Navy personnel. Am J Epidemiol 132:293–303. Graham C, Cook MR, Riffle DW, Gerkovich MM, Cohen HD. 1996. Nocturnal melatonin levels in human volunteers exposed to intermittent 60 Hz magnetic fields. Bioelectromagnetics 17:263–273. Grayson JK, 1996. Radiation exposure, socioeconomic status, and brain tumor risk in the US Air Force: A nested case-control study. Am J Epidemiol 143:480–486. Green LM, Miller AB, Agnew DA, Greenberg ML, Li J, Villeneuve PJ, Tibshirani R. 1999. Childhood leukemia and personal monitoring of residential exposures to electric and magnetic fields in Ontario, Canada. Cancer Causes Control 10:233–243. Greenland S, Sheppard AR, Kaune WT, Poole C, Kelsh MA. 2000. A pooled analysis of magnetic fields, wire codes, and childhood leukemia. Epidemiology 11:624–634. Groves FD, Page WF, Gridley G, Lisimaque L, Stewart PA, Tarone RE, Gail MH, Boice JD Jr, Beebe GW. 2002. Cancer in Korean War navy technicians: Mortality survey after 40 years. Am J Epidemiol 155:810–818. Guenel P, Nicolau J, Imbernon E, Chevalier A, Goldberg M. 1996. Exposure to 50-Hz electric field and incidence of leukemia, brain tumors, and other cancers among French electric utility workers. Am J Epidemiol 144:1107–1121.

320

PART III: THE CAUSES OF CANCER

Gurney JG, Mueller BA, Davis S, Schwartz SM, Stevens RG, Kopecky KJ. 1996. Childhood brain tumor occurrence in relation to residential power line configurations, electric heating sources, and electric appliance use. Am J Epidemiol 143:120–128. Hardell L, Hallquist A, Mild KH, Carlberg M, Påhlson A, Lilja A. 2002b. Cellular and cordless telephones and the risk for brain tumours. Eur J Cancer Prev 11:377–386. Hardell L, Hansson Mild K, Carlberg M. 2002a. Case-control study on the use of cellular and cordless phones and the risk for malignant brain tumours. Int J Radiat Biol 78:931–936. Hardell L, Mild KH, Påhlson A, Hallquist A. 2001. Ionizing radiation, cellular telephones and the risk for brain tumours. Eur J Cancer Prev 10:523–529. Hardell L, Näsman Å, Påhlson A, Hallquist A, Mild KH. 1999. Use of cellular telephones and the risk for brain tumours: A case-control study. Int J Oncol 15:113–116. Harrington JM, McBride DI, Sorahan T, Paddle GM, van Tongeren M. 1997. Occupational exposure to magnetic fields in relation to mortality from brain cancer among electricity generation and transmission workers. Occup Environ Med 54:7–13. Hatch EE, Linet MS, Kleinerman RA, Tarone RE, Severson RK, Hartsock CT, Haines C, Kaune WT, Friedman D, Robison LL, Wacholder S. 1998. Association between childhood acute lymphoblastic leukemia and use of electrical appliances during pregnancy and childhood. Epidemiology 9:234–245. Hayes RB, Brown LM, Pottern LM, Gomez M, Kardaun JW, Hoover RN, O’Connell KJ, Sutzman RE, Javadpour N. 1990. Occupation and risk for testicular cancer: A case-control study. Int J Epidemiol 19:825–831. Hocking B, Gordon IR, Grain HL, Hatfield GE. 1996. Cancer incidence and mortality and proximity to TV towers. Med J Aust 165:601–605. Holly EA, Aston DA, Ahn DK, Smith AH. 1996. Intraocular melanoma linked to occupations and chemical exposures. Epidemiology 7:55–61. Inskip PD, Tarone RE, Hatch EE, Wilcosky TC, Shapiro WR, Selker RG, Fine HA, Black PM, Loeffler JS, Linet MS. 2001. Cellular-telephone use and brain tumors. New Engl J Med 344:79–86. Johansen C, Boice JD Jr, Mc Laughlin JK, Christensen HC, Olsen JH. 2000. Mobile phones and malignant melanoma of the eye. Br J Cancer 86:348–349. Johansen C, Boice JD, McLaughlin JK, Olsen JH. 2001. Cellular telephones and cancer—A nationwide cohort study in Denmark. J Natl Cancer Inst 93:203–207. Johansen C, Olsen JH. 1998. Risk of cancer among Danish utility workers— A nationwide cohort study. Am J Epidemiol 147:548–555. Kaune WT, Darby SD, Gardner SN, Hrubec Z, Iriye RN, Linet MS. 1994. Development of a protocol for assessing time-weighted-average exposures of young children to power-frequency magnetic fields. Bioelectromagnetics 15:33–51. Kaune WT, Friedman DR, Haines CM, Muirhead CR, Boice JD Jr, Robison LL. 1997. Magnetic field exposure assessment in a case-control study of childhood leukemia. Epidemiology 8:575–583. Kaune WT, Stevens RG, Callahan NJ, Severson RK, Thomas DB. 1987. Residential magnetic and electric fields. Bioelectromagnetics 8:315–335. Kheifets LI, Afifi AA, Buffler PA, Zhang ZW, Matkin CC. 1997. Occupational electric and magnetic field exposure and leukemia. A meta analysis. J Occup Environ Med 39:1074–1091. Kheifets LI, Afifi AA, Buffler PA, Zhang ZW. 1995. Occupational electric and magnetic field exposure and brain cancer: A meta-analysis. J Occup Environ Med 37:1327–1341. Kheifets LI, Gilbert ES, Sussman SS, Guenel P, Sahl JD, Savitz DA, Theriault G. 1999. Comparative analyses of the studies of magnetic fields and cancer in electric utility workers: Studies from France, Canada, and the United States. Occup Environ Med 56:567–574. Kleinerman RA, Kaune WT, Hatch EE, Wacholder S, Linet MS, Robison LL, Niwa S, Tarone RE. 2000. Are children living near high-voltage power lines at increased risk of acute lymphoblastic leukemia? Am J Epidemiol 151:512–515. Kleinerman RA, Linet MS, Hatch EE, Wacholder S, Tarone RE, Severson RK, Kaune WT, Friedman DR, Haines CM, Muirhead CR, Boice JD Jr, Robison LL. 1997. Magnetic field exposure assessment in a case-control study of childhood leukemia. Epidemiology 8:575–583. Kliukiene J, Tynes T, Martinsen JI, Blaasaas KG, Andersen A. 1999. Incidence of breast cancer in a Norwegian cohort of women with potential workplace exposure to 50 Hz magnetic fields. Am J Ind Med 36:147–154. Laden F, Neas LM, Tolbert PE, Holmes MD, Hankinson SE, Spiegelman D, Speizer FE, Hunter DJ. 2000. Electric blanket use and breast cancer in the Nurses’ Health Study. Am J Epidemiol 152:41–49. Lagorio S, Rossi S, Vecchia P, De Santis M, Bastianini L, Fusilli M, Ferrucci A, Desideri E, Comba P. 1997. Mortality of plastic-ware workers exposed to radiofrequencies. Bioelectromagnetics 18:418–421.

Li CY, Theriault G, Lin RS. 1997. Residential exposure to 60-Hz magnetic fields and adult cancers in Taiwan. Epidemiology 8:25–30. Linet MS, Hatch EE, Kleinerman RA, Robison LL, Kaune WT, Friedman DR, Severson RK, Haines CM, Hartsock CT, Niwa S, Wacholder S, Tarone RE. 1997. Residential exposure to magnetic fields and acute lymphoblastic leukemia in children. N Engl J Med 337:1–7. London SJ, Thomas DC, Bowman JD, Sobel E, Cheng T-C, Peters JM. 1991. Exposure to residential electric and magnetic fields and risk of childhood leukemia. Am J Epidemiol 134:923–937. Loomis DP, Savitz DA, Ananth CV. 1994. Breast cancer mortality among female electrical workers in the United States. J Natl Cancer Inst 86:921–925. Lovely RH, Buschborn RL, Slavich AL, Anderson LE, Hansen NH, Wilson BW. 1994. Adult leukemia and personal appliance use: A preliminary study. Am J Epidemiol 140:510–517. Mader DL, Peralta SB. 1992. Residential exposure to 60-Hz magnetic fields from appliances. Bioelectromagnetics 13:287–301. Maskarinec G, Cooper J, Swygert L. 1994. Investigation of increased incidence in childhood leukemia near radio towers in Hawaii: Preliminary observations. J Environ Pathol Toxicol Oncol 13:33–37. Matanoski GM, Elliott EA, Breysse PN, Lynberg MC. 1993. Leukemia in telephone linemen. Am J Epidemiol 137:609–619. McBride ML, Gallagher RP, Theriault G, Armstrong BG, Tamaro S, Spinelli JJ, Deadman JE, Fincham S, Robson D, Choi W. 1999. Power-frequency electric and magnetic fields and risk of childhood leukemia in Canada. Am J Epidemiol 149:831–842. McDowall ME. 1986. Mortality of persons resident in the vicinity of electric transmission facilities. Br J Cancer 53:271–279. McElroy JA, Newcomb PA, Remington PL, Egan KM, Titus-Ernstoff L, Trentham-Dietz A, Hampton JM, Baron JA, Stampfer MJ, Willett WC. 2001. Electric blanket or mattress cover use and breast cancer incidence in women 50–79 years of age. Epidemiology 12:613–617. Michaelis J, Schuz J, Meinert R, Menger M, Grigat JP, Kaatsch P, Kaletsch U, Miesner A, Stamm A, Brinkmann K, Karner H. 1997. Childhood leukemia and electromagnetic fields: Results of a population-based casecontrol study in Germany. Cancer Causes Control 8:167–174. Michelozzi P, Capon A, Kirchmayer U, Forastiere F, Biggeri A, Barca A, Perucci CA. 2002. Adult and childhood leukemia near a high-power radio station in Rome, Italy. J Am J Epidemiol 155:1096–1103. Milham S. 1982. Mortality from leukemia in workers exposed to electrical and magnetic fields [letter]. N Engl J Med 307:249. Milham S Jr. 1985. Silent keys: Leukaemia mortality in amateur radio operators. Lancet 1:812. Milham S Jr. 1988a. Increased mortality in amateur radio operators due to lymphatic and hematopoietic malignancies. Am J Epidemiol 127:50– 54. Milham S Jr. 1988b. Mortality by license class in amateur radio operators. Am J Epidemiol 128:1175–1176. Miller AB, To T, Agnew DA, Wall C, Green LM. 1996. Leukemia following occupational exposure to 60-Hz electric and magnetic fields among Ontario electric utility workers. Am J Epidemiol 144:150–160. Morgan RW, Kelsh MA, Zhao K, Exuzides KA, Heringer S, Negrete W. 2000. Radiofrequency exposure and mortality from cancer of the brain and lymphatic/hematopoietic systems. Epidemiology 11:118–127. Muhm JM. 1992. Mortality investigation of workers in an electromagnetic pulse test program. J Occp Med 34:289–292. Muscat JE, Malkin MG, Shore RE, Thompson S, Neugut AI, Stellman SD, Bruce J. 2002. Handheld cellular telephones and risk of acoustic neuroma. Neurology 58:1304–1306. Muscat JE, Malkin MG, Thompson S, Shore RE, Stellman SD, McRee D, Neugut AI, Wynder E. 2000. Handheld cellular telephone use and risk of brain cancer. J Am Med Ass 284:3001–3007. Myers A, Clayden AD, Cartwright RA, Cartwright SC. 1990. Childhood cancer and overhead power lines. A case-control study. Br J Cancer 62: 1008–1014. National Research Council. 1997. Possible Health Effects of Exposure to Residential Electric and Magnetic Fields. Washington DC: National Research Council, National Academy Press. Olsen JH, Nielsen A, Schulgen G. 1993. Residence near high voltage facilities and risk of cancer in children. BMJ 307:891–895. Olson SH, Kelsey JL, Pearson TA, Levin B. 1992. Evaluation of random digit dialing as a method of control selection in case-control studies. Am J Epidemiol 135:210–222. Olson SH, Mignone L, Harlap S. 2000. Selection of control groups by using a commercial database and random digit dialing. Am J Epidemiol 152:585–592. Pearce N, Reif J, Fraser J. 1989. Case-control studies of cancer in New Zealand electrical workers. Int J Epidemiol 18:55–59.

Electromagnetic Fields and Radiofrequency Radiation Portier CJ, Wolfe MS, eds. 1998. Assessment of health effects from exposure to power-line frequency electric and magnetic fields. NIH Publication No. 98–3981. Research Triangle Park, North Carolina: Natl Inst Environ Health Sciences. Preston-Martin S, Navidi W, Thomas D, Lee PJ, Bowman J, Pogoda J. 1996. Los Angeles study of residential magnetic fields and childhood brain tumors. Am J Epidemiol 143:105–119. Preston-Martin S, Peters JM, Yu MC, Garabrant DH, Bowman JD. 1988. Myelogenous leukemia and electric blanket use. Bioelectromagnetics 9:207–213. Repacholi MH, Basten A, Gebski V, Noonan D, Finnie J, Harris AW. 1997. Lymphomas in E mu-Pim1 transgenic mice exposed to pulsed 900 MHZ electromagnetic fields. Radiat Res 147:631–640. Rodvall Y, Ahlbom A, Stenlund C, Preston-Martin S, Lindh T, Spannare B. 1998. Occupational exposure to magnetic fields and brain tumours in central Sweden. Eur J Epidemiol 14:563–569. Rosenbaum PF, Vena JE, Zielezny MA, Michalek AM. 1994. Occupational exposures associated with male breast cancer. Am J Epidemiol 139:30–36. Rothman KJ, Chou CK, Morgan R, Balzano Q, Guy AW, Funch DP, PrestonMartin S, Mandel J, Steffens R, Carlo G. 1996. Assessment of cellular telephone and other radio frequency exposure for epidemiologic research. Epidemiology 7:291–298. Rothman KJ, Loughlin JE, Funch DP, Dreyer NA. 1996. Overall mortality of cellular telephone customers. Epidemiology 7:303–305. Rothman KJ. 2000. Epidemiologic evidence on health risks of cellular telephones. Lancet 356:1837–1840. Sahl JD, Kelsh MA, Greenland S. 1993. Cohort and nested case-control studies of hematopoietic cancers and brain cancer among electric utility workers. Epidemiology 4:104–114. Savitz DA, Calle EE. 1987. Leukemia and occupational exposure to electromagnetic fields. Review of epidemiologic surveys. J Occup Med 29: 47–51. Savitz DA, John EM, Kleckner RC. 1990. Magnetic field exposure from electric appliances and childhood cancer. Am J Epidemiol 131:763–773. Savitz DA, Loomis DP. 1995. Magnetic field exposure in relation to leukemia and brain cancer mortality among electric utility workers. Am J Epidemiol 141:123–134. Savitz DA, Ohya T, Loomis DP, Senior RS, Bracken TD, Howard RL. 1994. Correlations among indices of electric and magnetic field exposure in electric utility workers. Bioelectromagnetics 15:193–204. Savitz DA, Wachtel H, Barnes FA, John EM, Tvrdik JG. 1988. Case-control study of childhood cancer and exposure to 60-hertz magnetic fields. Am J Epidemiol 128:21–38. Schreiber GH, Swaen GM, Meijers JM, Slangen JJ, Sturmans F. 1993. Cancer mortality and residence near electricity transmission equipment: A retrospective cohort study. Int J Epidemiol 22:9–15. Selvin S, Schulman J, Merrill DW. 1992. Distance and risk measures for the analysis of spatial data: a study of childhood cancers. Soc Sci Med 34:769–777. Severson RK, Stevens RG, Kaune WT, Thomas DB, Heuser L, Davis S, Sever LE. 1988. Acute nonlymphocytic leukemia and residential exposure to power frequency magnetic fields. Am J Epidemiol 128:10–20. Stang A, Anastassiou G, Ahrens W, Bromen K, Bornfeld N, Jöckel KH. 2001. The possible role of radiofrequency radiation in the development of uveal melanoma. Epidemiology 12:7–12. Stenlund C, Floderus B. 1997. Occupational exposure to magnetic fields in relation to male breast cancer and testicular cancer: A Swedish case-control study. Cancer Causes Control 8:184–191. Stevens RG, Davis S, Thomas DB, Anderson LE, Wilson BW. 1992. Electric power, pineal function, and the risk of breast cancer. FASEB J 6:853–860. Sussman SS, Kheifets LI. 1996. Re: “Adult leukemia risk and personal appliance use: A preliminary study”. Am J Epidemiol 143:743–744. Swerdlow AJ. 1999. Measurement of radiofrequency radiation exposure in epidemiological studies. Radiat Prot Dosimetry 83:149–153. Szmigielski S. 1996. Cancer morbidity in subjects occupationally exposed to high frequency (radiofrequency and microwave) electromagnetic radiation. Sci Total Environ 180:9–17. Tarone RE, Kaune WT, Linet MS, Hatch EE, Kleinerman RA, Robison LL, Boice Jr, Wacholder S. 1998. Residential wire codes: Reproducibility and relation with measured magnetic fields. Occup Environ Med 55:333–339. Theriault G, Goldberg M, Miller AB, Armstrong B, Guenel P, Deadman J, Imbernon E, To T, Chevalier A, Cyr D, Wall C. 1994. Cancer risks asso-

321

ciated with occupational exposure to magnetic fields among electric utility workers in Ontario and Quebec, Canada, and France: 1970–1989. Am J Epidemiol 139:550–572. Thomas TL, Stolley PD, Stemhagen A, Fontham ET, Bleecker ML, Stewart PA, Hoover RN. 1987. Brain tumor mortality risk among men with electrical and electronics jobs: A case-control study. J Natl Cancer Inst 79:233–238. Tice RR, Hook GG, Donner M, McRee DI, Guy AW. 2002. Genotoxicity of radiofrequency signals. I. Investigation of DNA damage and micronuclei induction in cultured human blood cells. Bioelectromagnetics 23: 113–126. Tomenius L. 1986. 50-Hz electromagnetic environment and the incidence of childhood tumors in Stockholm County. Bioelectromagnetics 7:191–207. Tynes T, Haldorsen T. 1997. Electromagnetic fields and cancer in children residing near Norwegian high-voltage power lines. Am J Epidemiol 145:219–226. Tynes T, Jynge H, Vistnes AI. 1994. Leukemia and brain tumors in Norwegian railway workers, a nested case-control study. Am J Epidemiol 139:643–653. Tynes T, Andersen A, Langmark F. 1992. Incidence of cancer in Norwegian workers potentially exposed to electromagnetic fields. Am J Epidemiol 136:81–88. Tynes T, Hannevik M, Andersen A, Vistnes AI, Haldorsen T. 1996. Incidence of breast cancer in Norwegian female radio and telegraph operators. Cancer Causes Control 7:197–204. Tynes T, Andersen A. 1990. Electromagnetic fields and male breast cancer. Lancet 336:1596. United Kingdom Childhood Cancer Study Investigators. Exposure to powerfrequency magnetic fields and the risk of childhood cancer. 1999. Lancet 354:1925–1931. Utteridge TD, Gebski V, Finnie JW, Vernon-Roberts B, Kuchel TR. 2002. Long term exposure of E-mu-Pim 1 transgenic mice to 898.4 MHz microwaves does not increase lymphoma incidence. Radiat Res 158:357–364. Van Wijngaarden E, Nylander-French L, Millikan RC, Savitz DA, Loomis D. 2001. Population-based case-control study of occupational exposure to electromagnetic fields and breast cancer. Ann Epidemiol 11:297–303. Vena JE, Freudenheim JL, Marshall JR, Laughlin R, Swanson M, Graham S. 1995. Re: Risk of pre-menopausal breast cancer and use of electric blankets and Use of electric blankets and risk of postmenopausal breast cancer. The authors reply [letter]. Am J Epidemiol 142:1345. Vena JE, Freudenheim JL, Marshall JR, Laughlin R, Swanson M, Graham S. 1994. Risk of premenopausal breast cancer and use of electric blankets. Am J Epidemiol 140:974–979. Vena JE, Graham S, Hellmann R, Swanson M, Brasure J. 1991. Use of electric blankets and risk of postmenopausal breast cancer. Am J Epidemiol 134:180–185. Verkasalo PK, Pukkala E, Hongistro MY, Valjus JE, Jarvinen PJ, Heikkila KV, Koskenvuo M. 1993. Risk of cancer in Finnish children living close to power lines. Br Med J 307:895–899. Verkasalo PK, Pukkala E, Kaprio J, Heikkila KV, Koskenvuo M. 1996. Magnetic fields of high voltage power lines and risk of cancer in Finnish adults: Nationwide cohort study. Br Med J 313:1047–1051. Verreault R, Weiss NS, Hollenbach KA, Strader CH, Daling JR. 1990. Use of electric blankets and risk of testicular cancer [comment]. Am J Epidemiol 131:759–762. Wertheimer N, Leeper E. 1982. Adult cancer related to electrical wires near the home. Int J Epidemiol 11:345–355. Wertheimer N, Leeper E. 1987. Magnetic field exposure related to cancer subtypes. Ann N Y Acad Sci 502:43–54. Wertheimer N, Leeper E. 1979. Electrical wiring configurations and childhood cancer. Am J Epidemiol 109:273–284. Wright WE, Peters JM, Mack TM. 1982. Leukaemia in workers exposed to electrical and magnetic fields [letter]. Lancet 2:1160–1161. Youngson JH, Clayden AD, Myers A, Cartwright RA. 1991. A case/control study of adult haematological malignancies in relation to overhead powerlines. Br J Cancer 63:977–985. Zheng T, Holford TR, Mayne ST, Owens PH, Zhang B, Boyle P, Carter D, Ward B, Zhang Y, Zahm SH. 2000. Exposure to electromagnetic fields from use of electric blankets and other in-home electrical appliances and breast cancer risk. Am J Epidemiol 151:1103–1111. Zhu K, Weiss NS, Stanford JL, Daling JR, Stergachis A, McKnight B, Brawer MK, Levine RS. 1999. Prostate cancer in relation to the use of electric blanket or heated water bed. Epidemiology 10:83–85.

18

Occupation JACK SIEMIATYCKI, LESLEY RICHARDSON, AND PAOLO BOFFETTA

O

ccupational carcinogens occupy a special place among the different classes of environmental carcinogens. The occupational environment has been a most fruitful one for investigating the pathogenesis of human cancer. Indeed, nearly half of all recognized human carcinogens are occupational carcinogens. Although it is important to discover occupational carcinogens for the sake of preventing occupational cancer, the potential benefit of such discoveries goes beyond the factory walls since most occupational exposures find their way into the general environment, sometimes at higher concentrations than in the workplace.

HISTORY AND BACKGROUND In 1775, Sir Percivall Pott, one of the leading British surgeons of the day, described some cases of cancer of the scrotum among English chimney sweeps. He ascribed this condition, which was known in the trade as “soot wart,” to the chimney sweeps’ pitifully dirty working conditions and to the “lodgment of soot in the rugae of scrotum” (Pott, 1775). In the ensuing century, the syndrome became widely known, but it remained the only recognized occupationally caused cancer until the latter part of the 19th century. In 1875, Volkmann described a syndrome identical to “chimney sweeps cancer” of the scrotum among a group of coal tar and paraffin workers (Volkmann, 1875). Apparent clusters of scrotal cancer were thereafter reported among shale oil workers (Bell, 1876) and mule spinners in the cotton textile industry (Morley, 1911; Southam and Wilson, 1922). By 1907 the belief in the carcinogenicity of “pitch, tar, and tarry substances” was widespread enough that skin cancers among exposed workers were officially recognized as compensable in the United Kingdom. Other types of cancer were also implicated as occupationally induced. In the late 19th century, Härting and Hesse (1879) reported that pulmonary cancer was common among metal miners and Rehn (1895) reported a striking cluster of bladder cancer cases among workers from a local plant that produced dyestuffs from coal tar. Following the accumulation of several of these clinical case reports of high-risk occupations, the scientific investigation of cancer etiology began in earnest at the beginning of the 20th century with experimental animal research. A major breakthrough came with the experiments of Yamagiwa and Ichikawa (1918) in which they succeeded in inducing skin tumors in rabbit ears by applying coal tar. Several important experimental discoveries were made in the next 20 years, particularly by an English group led by Kennaway. In a series of experiments they managed to isolate dibenz(a,h)anthracene and benzo(a)pyrene, both polycyclic aromatic hydrocarbons (PAHs) and active ingredients in coal tar (Kennaway and Hieger, 1930; Cook et al., 1932; Hieger, 1933). These compounds may have been responsible for many of the excess risks of scrotal cancer in various groups exposed to soot and oils (Waldron, 1983). Several other PAHs were subsequently shown to be carcinogenic to laboratory animals, but so were substances of many other chemical families. For instance, 2-naphthylamine was shown to cause bladder tumors in dogs and this was thought to explain the bladder cancers seen earlier among dyestuffs workers. During the first half of the 20th century, there were additional reports of high-risk occupation groups. Respiratory cancer risks were reported in such diverse occupational settings as nickel refineries (Bridge, 1933), coal carbonization processes (Kuroda and Kawahata,

322

1936), chromate manufacture (Machle and Gregorius, 1948), manufacture of sheep-dip containing inorganic arsenicals (Hill and Faning, 1948), and asbestos products manufacture (Merewether, 1949). This occurred before the smoking-induced epidemic of lung cancer was at its peak, when the background risks of lung cancer were low. The era of modern cancer epidemiology began around 1950 with several studies of smoking and lung cancer. In the field of occupational cancer epidemiology, this era saw the conduct of some important studies of gas workers (Doll, 1952), asbestos workers (Doll, 1955), and workers producing dyestuffs in the chemical industry (Case et al., 1954). The findings of these early studies were important in highlighting significant workplace hazards and the methods that these pioneering investigators developed for studying occupational cohorts have strongly influenced the conduct of occupational cancer research. Subsequently, and especially with the flowering of “environmentalism” in the 1960s as a component of social consciousness, there was a sharp increase in the amount of research aimed at investigating links between the environment and cancer. Particular attention was paid to the occupational environment for several reasons. Most of the historic observations of environmental cancer risks were discovered in occupationally exposed populations. As difficult as it is to characterize and study groups of workers, it is much harder to study groups of people who share other characteristics, such as diet or general environmental pollution. Not only are working populations easier to delineate but, often, company personnel and industrial hygiene records permit some, albeit crude, form of quantification of individual workers’ exposure to workplace substances. Also, the pressure of organized labor was an important force in attracting attention to the workplace. Finally, the workplace is a setting where people have been exposed to high levels of many substances that could potentially be harmful. Nonetheless, since many occupational exposures can also occur in the general environment, the cancer risks borne by workers have implications well beyond the workplace. The burst of epidemiologic research on cancer and environment was accompanied by extensive experimental work aimed at testing the carcinogenic potential of different substances. Whereas this was carried out in an uncoordinated fashion in the early years, national bodies, most notably the National Toxicology Program in the United States, have implemented systematic strategies to test large numbers of substances with standardized state-of-the-art long-term animal studies (Bucher, 2002).

SOURCES OF EVIDENCE ON RISK TO HUMANS DUE TO CHEMICALS Direct evidence concerning carcinogenicity of a substance can come from epidemiologic studies among humans or from classic experimental studies of animals (usually rodents). Additional evidence comes from the results of studies of chemical structure-activity analysis, pharmacokinetics, mutagenicity, cytotoxicology, and other aspects of toxicology.

Epidemiology Epidemiologic research provides the most relevant data for identifying occupational carcinogens and characterizing their effects in humans. It can also contribute to the understanding of the mechanism

Occupation of action of occupational carcinogens. Such research requires the juxtaposition of information on illness or death due to cancer among workers and information on their past occupations, industries, and/or occupational conditions. A third, optional data set that would improve the validity of inferences drawn from that juxtaposition is the set of concomitant risk factors that may confound the association between occupation and disease. Because of long induction periods for most cancers, current epidemiologic studies would not provide direct evidence on carcinogenic risk due to recently introduced industrial agents. Even for substances that have been with us for a long time, there are obstacles. All humans experience, over their lifetime, an idiosyncratic and bewildering pattern of exposures. Not only is it impossible to completely and accurately characterize the lifetime exposure profile of an individual, but even if we could it is a daunting statistical task to tease out the effects of a myriad of specific substances. The ascertainment of valid cancer diagnoses is also problematic since subjects are often traced via routine record sources (notably, death certificates), which may be error-prone or in which cancers with long survival are poorly represented. Confounding by factors other than the one under investigation is of course an issue in occupational cancer epidemiology, as it is in other areas of epidemiology. But the problem is sometimes particularly acute in occupational epidemiology because of some highly correlated co-exposures in the occupational environment. The number of subjects available for epidemiologic study is often limited and this compromises the statistical power to detect hazards. Despite these challenges, epidemiology has made significant contributions to our knowledge of occupational carcinogens.

Animal Experimentation Partly in consequence of the difficulty of generating adequate data among humans and partly because of the benefits of the experimental approach, great efforts have been devoted to studying the effects of substances in controlled animal experiments. Results generated by animal studies do bear on carcinogenicity among humans (Shubik, 1979; Berenblum, 1979; Wilbourn et al., 1986; Montesano et al., 1986; Rall et al., 1987). Certain fundamental genetic and cellular characteristics are similar among all mammalian species. Most recognized human carcinogens have been reported to be carcinogenic in one or more animal species; and there is some correlation between species in the target organs affected and in the carcinogenic potency (Wilbourn et al., 1986; Allen et al., 1988; Gold et al., 1997). Still, there are several reasons for caution in extrapolating from animal evidence to humans. The animal experiment is designed not to emulate the human experience but to maximize the sensitivity of the test to detect animal carcinogens. Doses administered are usually orders of magnitude higher than levels to which humans are exposed. The route of exposure is sometimes unrealistic (e.g., injection or implantation) and the controlled and limited pattern of co-exposures is unlike the human situation. The “lifestyle” of the experimental animal is not only different from that of humans, but it is unlike that of its species in the wild. Animals used are typically from pure genetic strains and susceptibility to carcinogens may be higher in such populations than in genetically heterogeneous human populations. Metabolic, immunologic, DNA repair systems, life spans, and other physiologic characteristics differ between species. Tumors seen in animals often occur at sites that do not have a counterpart among humans (e.g., forestomach or Zymbal’s glands) or sites that are much more rarely affected among humans (e.g., pituitary gland). The behaviour of many tumors generated in experimental animals does not mimic that of malignant neoplasms in humans, and the malignant phenotype is sometimes unclear. Quantitative extrapolation of effects from rodents to humans depends on unverifiable mathematical assumptions concerning dose equivalents, dose-response curves, safety factors, etc. Different reasonable assumptions can lead to wildly divergent estimates. Recent evidence indicates that several experimental carcinogens operate via mechanisms that may not be relevant to humans. A case in point is that of kidney tumors in male rats following exposure to various organic chemicals and mixtures including gaso-

323

line, which is apparently caused by precipitation of a2-microglobulin, a gender- and species-specific protein (Swenberg and LehmanMcKeeman, 1999). Gold et al. (1998) have shown that even between two species as close on the phylogenetic scale as mice and rats, the predictive value of carcinogenicity is only in the range of 75%. Despite efforts to investigate the scientific basis for inter-species extrapolation and despite resources that have been devoted to testing chemicals in animal systems, there remain serious disagreements about the predictive value of animal experimentation (Purchase, 1986; Gold et al., 1989; Tomatis and Bartsch, 1990; Cohen, 1995; Ashby, 1996; Freedman et al., 1996; Tomatis et al., 1996; Gottmann et al., 2001; Bucher, 2002).

Short-Term Tests and Structure-Activity Relationships To mitigate the lengthy and costly process of animal carcinogenesis testing, a number of rapid, inexpensive, and ingenious tests have been developed, to detect presumed correlates of or predictors of carcinogenicity (Montesano et al., 1986; Ashby and Tennant, 1988; Zeiger, 1998; Waters et al., 1999; Weisburger, 1999). However, neither alone nor in combination have these approaches proven to be consistently predictive of animal carcinogenicity, much less human carcinogenicity (Tennant et al., 1990; Huff et al., 1996; Zeiger, 1998; Kim and Margolin, 1999). Their role is in screening chemicals for animal testing and in complementing the results of animal experiments. Further advances in molecular biology and pathology may soon provide new tools to detect intermediate events relevant to the process of carcinogenesis. Examples of such events include: (1) alterations in signaling pathways following the interaction of the carcinogens with relevant receptors, (2) interaction of carcinogens or active metabolites with DNA and other macromolecules, (3) alterations in the regulation of genes involved in carcinogenesis, (4) stimulation of cell duplication and proliferation. For the most part, the evidence on mechanisms of carcinogenesis is now available only in experimental systems, but there is hope that this will lead to the development of tools that can be used in humans exposed to environmental carcinogens (Buffler et al., 2004).

LISTING OCCUPATIONAL CARCINOGENS This section, and the accompanying tables, is based on an article by Siemiatycki et al. (submitted).

Difficulties in Listing Occupational Carcinogens Although it seems like a simple enough task, it is very difficult to draw up an unambiguous list of occupational carcinogens. The first source of ambiguity concerns the definition of an occupational carcinogen. Most occupational exposures are also found in the general environment, and/or in consumer products; most general environmental exposures and consumer products, including medications, foods, and others, are found in some occupational environments. The distinctions can be quite arbitrary. For instance, while tobacco smoke, sunlight, and immunosuppressive medications are not primarily considered to be occupational exposures, there certainly are workers whose occupations bring them into contact with these agents. Also, while asbestos, benzene, and radon gas are considered to be occupational carcinogens, they are also found widely among the general population, and indeed it is likely that many more people are exposed to these substances outside than inside the occupational environment. There is no simple rule to earmark “occupational” carcinogens as opposed to “nonoccupational” ones. Further, some carcinogens are chemicals that are used for research purposes and to which few people would ever be exposed, whether occupationally or non-occupationally. Our operational criterion for designating occupational carcinogens is outlined below. A second source of ambiguity derives from the rather idiosyncratic nature of the evidence. In some instances, we know that an occupational or industrial group is at excess risk of cancer and we have a good idea of the causative agent (e.g., scrotal cancer among chimney

324

PART III: THE CAUSES OF CANCER

sweeps and PAHs in soot (Waldron, 1983); lung cancer among asbestos miners and asbestos fibers (IARC, 1977)). In some instances, we know that a group experienced excess risk but the causative agent is unknown or at least unproven (e.g., lung cancer among painters (IARC, 1989c); bladder cancer among workers in the aluminium industry (IARC, 1987a)). The strength of the evidence for an association can vary. For some associations the evidence of excess risk seems incontrovertible (e.g., liver angiosarcoma and vinyl chloride monomer (IARC, 1979b); bladder cancer and benzidine (IARC, 1982)). For some associations the evidence is suggestive (e.g., lung cancer and diesel engine exhaust (IARC, 1989a); bladder cancer and employment as a painter (IARC, 1989c)). Among the many substances in the industrial environment for which there are no human data concerning carcinogenicity, there are hundreds that have been shown to be carcinogenic in some animal species and thousands that have been shown to have some effect in assays of mutagenicity or genotoxicity. These considerations complicate the attempt to devise a list of occupational carcinogens.

International Agency for Research on Cancer Monographs For this task we drew on the authoritative Monograph Programme of the International Agency for Research on Cancer (IARC)—Evaluation of the Carcinogenic Risk of Chemicals to Humans (IARC, 1987a). The objective of the IARC Programme, which has been operating since 1971, is to publish critical reviews of epidemiological and experimental data on carcinogenicity for chemicals, groups of chemicals, industrial processes, other complex mixtures, physical agents, and biological agents to which humans are known to be exposed; to evaluate the data in terms of human risk; and to indicate where additional research efforts are needed. Substances are selected for evaluation on the basis of two main criteria: (1) humans are exposed, and (2) there is reason to suspect that

the substance may be carcinogenic. Direct evidence concerning carcinogenicity of a substance can come from epidemiological studies among humans or from experimental studies of animals (usually rodents). Additional evidence comes from the results of studies of chemical structure-activity analysis, absorption and metabolism, physiology, mutagenicity, cytotoxicology, and other aspects of toxicity. In the IARC Monographs, all types of data contribute to the evaluation. We will outline the IARC process, because it is important to understand how decisions are made to interpret properly its output. IARC evaluations are carried out during specially convened meetings that typically last a week. The meetings may evaluate only one agent, such as silica, they may address a set of related agents, or they may even address exposure circumstances such as an occupation or an industry. For each such meeting, and there have typically been three per year, IARC convenes an international Working Group, usually involving from 15 to 30 experts on the topic(s) being evaluated, from four perspectives: (1) exposure and occurrence of the substances being evaluated, (2) human evidence of cancer risk (i.e., epidemiology), (3) animal carcinogenesis, and (4) other data relevant to the evaluation of carcinogenicity and its mechanisms. The Working Group is asked to review all of the literature relevant to an assessment of carcinogenicity. In the first part of the meeting four subgroups (based on the four perspectives mentioned above) review and revise drafts prepared by members of the subgroup, and each subgroup develops a joint review and evaluation of the evidence on which they have focused. Subsequently, the entire Working Group convenes in plenary and proceeds to derive a joint text. They determine whether the epidemiological evidence supports the hypothesis that the substance causes cancer, and, separately, whether the animal evidence supports the hypothesis that the substance causes cancer. The judgments are not simply dichotomous (yes/no); rather, they allow the Working Group to express a range of opinions on each of the dimensions evaluated. Table 18–1 shows the categories into which the Working Groups are asked

Table 18–1. Classifications Used in the IARC Monographs to Characterize Evidence of Carcinogenicity Category of Evidence

In Humans

In Animals

Sufficient evidence of carcinogenicity

A causal relationship has been established between exposure to the agent, mixture, or exposure circumstance, and human cancer. That is, a positive relationship has been observed between the exposure and cancer in studies in which chance, bias, and confounding could be ruled out with reasonable confidence.

A causal relationship has been established between the agent or mixture and an increased incidence of malignant neoplasms or of an appropriate combination of benign and malignant neoplasms in (a) two or more species of animals or in (b) two or more independent studies in one species carried out at different times or in different laboratories or under different protocols.

Limited evidence of carcinogenicity

A positive association has been observed between exposure to the agent, mixture, or exposure circumstance, and cancer for which a causal interpretation is considered to be credible, but chance, bias, or confounding could not be ruled out with reasonable confidence.

The data suggest a carcinogenic effect but are limited for making a definitive evaluation because, e.g. (a) the evidence of carcinogenicity is restricted to a single experiment; or (b) there are unresolved questions regarding the adequacy of the design, conduct, or interpretation of the study; or (c) the agent or mixture increases the incidence only of benign neoplasms or lesions of uncertain neoplastic potential, or of certain neoplasms that may occur spontaneously in high incidences in certain strains.

Inadequate evidence of carcinogenicity

The available studies are of insufficient quality, consistency, or statistical power to permit a conclusion regarding the presence or absence of a causal association between exposure and cancer, or no data on cancer in humans are available.

The studies cannot be interpreted as showing either the presence or absence of a carcinogenic effect because of major qualitative or quantitative limitations, or no data on cancer in experimental animals are available.

Evidence suggesting lack of carcinogenicity

There are several adequate studies covering the full range of levels of exposure that human beings are known to encounter, which are mutually consistent in not showing a positive association between exposure to the agent, mixture, or exposure circumstance and any studied cancer at any observed level of exposure.

Adequate studies involving at least two species are available, which show that, within the limits of the tests used, the agent or mixture is not carcinogenic.

325

Occupation to classify each substance, when examining only the epidemiological evidence and when examining only the animal experimental evidence. The operational criteria for making these decisions leave room for interpretation, and the scientific evidence itself is open to interpretation. It is not surprising then that the evaluations are sometimes difficult and contentious. The overall evaluation of human carcinogenicity is based on the epidemiological and animal evidence of carcinogenicity, plus any other relevant evidence on genotoxicity, mutagenicity, metabolism, mechanisms, or other. Epidemiological evidence, where it exists, is given greatest weight. Direct animal evidence of carcinogenicity is next in importance, with increasing attention paid to mechanistic evidence that can inform the relevance of the animal evidence for human risk assessment. Table 18–2 shows the categories for the overall evaluation, and how they are derived from human, animal, and other evidence. Each substance is classified into one of the following classes (which IARC refers to as “groups”): carcinogenic (group 1), probably carcinogenic (group 2A), possibly carcinogenic (group 2B), not classifiable (group 3), probably not carcinogenic (group 4). However, the algorithm implied by Table 18–2 is only indicative and the Working Group may derive an overall evaluation that departs from the strict interpretation of the algorithm. For example, neutrons have been classified as human carcinogens (group 1) despite the absence of epidemiological data, because of overwhelming experimental evidence and mechanistic considerations (IARC, 2000a). The IARC process relies on consensus, and this is usually achieved, but sometimes, differing opinions among experts leads to split decisions. In the end, the published evaluations reflect the views of at least a majority of participating experts. The results of IARC evaluations are published in readily available and user-friendly volumes and summaries are published on a web site (IARC, 2003). For our purpose, there are several limitations to bear in mind. First, IARC does not provide any explicit indication as to whether the substance evaluated should be considered as an occupational exposure. Second, while the Working Groups certainly study the evidence in relation to cancer sites, until recently the formal evaluations did not identify which sites of cancer may be at risk. Site-specific information must be gleaned from the Working Group’s report and other literature. Third, the evaluations are anchored in the time that the Working Group met and reviewed the evidence; it is possible that evidence that appeared after the IARC review could change the evaluation.

Current Knowledge on Occupational Carcinogens From 1972 to 2003, the IARC Monograph Programme published 83 volumes, representing evaluations of more than 880 substances, complex mixtures, and industrial processes. Of these, 89 have been classed as human carcinogens, 64 as probable and 264 as possible human carcinogens (IARC, 2003). We reviewed each one and earmarked those we consider to be “occupational exposures”. A substance was considered an occupational exposure if there are, or have been, significant numbers of workers exposed to the substance at significant levels. The operational threshold for significant numbers that we tried to implement was: >10,000 workers exposed worldwide or >1000 in any country, presently or at any time in the past. The operational threshold for significant levels that we tried to implement was not explicit; it depended, inter alia, on the range of exposure levels to the agent. In fact, the knowledge base for determining how many workers are or have been exposed, and at what levels, is very fragmentary. We relied on available documentation such as the IARC Monographs, NIOSH surveys (US Department of Health and Human Services, 2003), and informed guesses. We tended to apply looser criteria for labeling definite and probable carcinogens as “occupational” than we did for possible carcinogens. Among the agents listed are several that are more prevalent and more important in non-occupational environments, such as aflatoxins, sunlight, involuntary tobacco smoking, and radon. Among the various agents that we did not include as occupational exposures were: hormones, pharmaceuticals, microbiological agents, and dietary constituents. However, there are workers exposed to many of these. Pharmaceuticals represent a special case. Many have been evaluated and many are considered to be carcinogenic. While the main population exposed consists of patients undergoing therapy, there can also be exposure to workers who produce the drugs, and to health care workers who administer them. But because the exposure doses are orders of magnitude higher among patients than among workers, we have not listed these as occupational carcinogens. Analogously, we have not listed carcinogenic viruses, notably HIV, HBV, and HCV, though health care workers may be at risk. For the purpose of this exposition, and with the criteria given above, we present the following lists derived from the IARC Monographs:

• 28 human occupational carcinogens (IARC group 1)—Table 18–3; • 27 probable human occupational carcinogens (IARC group 2A)— Table 18–4;

Table 18–2. Classifications and Guidelines Used by IARC Working Groups in Evaluating Human Carcinogenicity Based on the Synthesis of Epidemiological, Animal, and Other Evidence* Combinations That Fit in This Class Group 1 2A 2B 3 4

Description of Group The agent, mixture, or exposure circumstance is carcinogenic to humans. The agent, mixture, or exposure circumstance is probably carcinogenic to humans. The agent, mixture, or exposure circumstance is possibly carcinogenic to humans. The agent, mixture, or exposure circumstance is not classifiable as to its carcinogenicity to humans. The agent, mixture, or exposure circumstance is probably not carcinogenic to humans.

Epidemiological Evidence Sufficient Less than sufficient Limited Inadequate or not available Limited Inadequate or not available Inadequate or not available Inadequate or not available Suggesting lack of carcinogenicity Inadequate or not available

Animal Evidence

Other Evidence

Any Sufficient Sufficient Sufficient Less than sufficient Sufficient Limited Limited Not elsewhere classified Suggesting lack of carcinogenicity Suggesting lack of carcinogenicity

Any Strongly positive Less than strongly positive Strongly positive Any Less than strongly positive Strongly positive Less than strongly positive Any Strongly negative

*This table shows our interpretation of the IARC guidelines used by the Working Groups to derive the overall evaluation from the combined epidemiological, animal, and other evidence. However, the Working Group can, under exceptional circumstances, depart from these guidelines in deriving the overall evaluation. For example, the overall evaluation can be downgraded if there is less than sufficient evidence in humans and strong evidence that the mechanism operating in animals is not relevant to humans. For details of the guidelines refer to the Preamble of the IARC Monographs (IARC, 2003).

Table 18–3. Substances and Mixtures That Have Been Evaluated by IARC as Definite (Group 1) Human Carcinogens and Are Occupational Exposures Occupation or Industry in Which Substance Found*

IARC Volume and Year†

Human Evidence‡

Ionizing radiation and sources thereof, including notably, X-rays, gamma rays, neutrons, and radon gas

Radiologists, technologists, nuclear workers, radium-dial painters, underground miners, plutonium workers, clean-up workers following nuclear accidents, aircraft crew

Vol. 75 (2000a) Vol. 78 (2001a)

Sufficient

Sufficient

Solar radiation

Outdoor workers

Vol. 55 (1992b)

Sufficient

Sufficient

Suppl. 7 (1987)

Sufficient

Sufficient

Suppl. 7 (1987)

Sufficient

Sufficient

Lung Mesothelioma Larynx Gastrointestinal tract Mesothelioma

Vol. 68 (1997b)

Sufficient

Sufficient

Lung

Suppl. 7 (1987)

Sufficient

Inadequate

Vol. 62 (1995b)

Sufficient

Inadequate

Lung Mesothelioma Nasal cavities and paranasal sinuses

Suppl. 7 (1987)

Sufficient

Limited

Skin Lung Liver (angiosarcoma)

Vol. 58 (1993a)

Sufficient

Sufficient

Lung

Vol. 58 (1993a)

Sufficient

Sufficient

Lung

Vol. 49 (1990a)

Sufficient

Sufficient

Lung Nasal sinuses

Vol. 49 (1990a)

Sufficient

Sufficient

Lung Nasal cavity and sinuses

Suppl. 7 (1987)

Sufficient

Limited

Leukemia

Suppl. 7 (1987)

Sufficient

Sufficient

Skin Lung Bladder

Suppl. 7 (1987)

Sufficient

Inadequate

Skin Bladder Lung Nasal sinuses

Substance or Mixture

Animal Evidence‡

Site(s)**

physical agents Bone Leukemia Lung Liver Thyroid Others Melanoma Skin

respirable dusts and fibers Asbestos

Erionite Silica, crystalline

Talc containing asbestiform fibers Wood dust

Mining and milling; byproduct manufacture; insulating; shipyard workers; sheet-metal workers; asbestos cement industry; plumbers and pipe-fitters Waste treatment; sewage; agricultural waste; in air pollution control systems; cement aggregates; building materials Granite and stone industries; ceramics, glass, and related industries; foundries and metallurgical industries; abrasives; construction; farming Manufacture of pottery, paper, paint, and cosmetics Logging and sawmill workers; pulp and paper and paperboard industry; woodworking trades (e.g., furniture industries, cabinetmaking, carpentry and construction); used as filler in plastic and linoleum production

metals and metal compounds Arsenic and arsenic compounds

Beryllium

Cadmium and cadmium compounds Chromium compounds, hexavalent

Selected nickel compounds, including combinations of nickel oxides and sulfides in the nickel refining industry

Non-ferrous metal smelting; production, packaging, and use of arsenic-containing pesticides; sheep dip manufacture; wool fibre production; mining of ores containing arsenic Beryllium extraction and processing; aircraft and aerospace industries, electronics and nuclear industries; jewellers Cadmium-smelter workers; battery production workers; cadmium-copper alloy workers; dyes and pigments production; electroplating process Chromate production plants; dyes and pigments; plating and engraving; chromium ferro-alloy production; stainless-steel welding; in wood preservatives; leather tanning; water treatment; inks; photography; lithography; drilling muds; synthetic perfumes; pyrotechnics; corrosion– resistance Nickel refining and smelting; welding

wood and fossil fuels and their byproducts Benzene

Coal tars and pitches

Mineral oils, untreated and mildly treated

326

Production; solvents in the shoe production industry; chemical, pharmaceutical, and rubber industries; printing industry (rotogravure plants, bindery departments); gasoline additive Production of refined chemicals and coal tar products (patent-fuel); coke production; coal gasification; aluminum production; foundries; road paving and construction (roofers and slaters) Production; used as lubricant by metal workers, machinists, engineers; printing industry (ink formulation); used in cosmetics, medicinal, and pharmaceutical preparations

327

Occupation Table 18–3. (cont.) Substance or Mixture Shale oils or shalederived lubricants Soots

Occupation or Industry in Which Substance Found*

IARC Volume and Year†

Human Evidence‡

Suppl. 7 (1987)

Sufficient

Sufficient

Skin

Vol. 35 (1985)

Sufficient

Inadequate

Skin Lung Esophagus

Suppl. 7 (1987)

Sufficient

Sufficient

Liver (angiosarcoma) Liver (hepatocellular)

Production; chemical intermediate; alkylating agent; laboratory reagent; plastic manufacturing; ion-exchange resins and polymers

Suppl. 7 (1987)

Sufficient

Sufficient

Lung (oat cell)

Production; dyestuffs and pigment manufacture Production; dyestuffs and pigment manufacture Production; dyestuffs and pigment manufacture

Suppl. 7 (1987)

Sufficient

Sufficient

Bladder

Suppl. 7 (1987)

Sufficient

Sufficient

Bladder

Suppl. 7 (1987)

Sufficient

Sufficient

Bladder

Vol. 60 (1994b)

Limited

Sufficient

Leukemia

Vol. 69 (1997a)

Limited

Sufficient

All sites combined Lung Non-Hodgkin lymphoma Sarcoma

Feed production industry, workers loading and unloading cargo; rice and maize processing Workers in bars and restaurants, office workers Production; used in research laboratories, military personnel

Vol. 82 (2002c)

Sufficient

Sufficient

Liver

Vol. 83 (2004)

Sufficient

Sufficient

Lung

Suppl. 7 (1987)

Sufficient

Limited

Pickling operations; steel industry; petrochemical industry; phosphate acid fertilizer manufacturing

Vol. 54 (1992a)

Sufficient

Not available

Larynx Lung Pharynx Larynx Lung

Mining and processing; used as fuels or chemical-plant feedstocks; lubricant in cotton textile industry Chimney sweeps; heating-unit service personnel; brick masons and helpers; building demolition workers; insulators; firefighters; metallurgical workers; work involving burning of organic materials

Animal Evidence‡

Site(s)**

monomers Vinyl chloride

Production; production of polyvinyl chloride and co-polymers; refrigerant before 1974; extraction solvent; in aerosol propellants

intermediates in plastics and rubber manufacturing Bis(chloromethyl)ether and chloromethyl methyl ether (technical grade)

aromatic amine dyes 4-Aminobiphenyl Benzidine 2-Naphthylamine

pesticides Ethylene oxide 2,3,7,8-Tetrachlorodibenzopara-dioxin (TCDD)

Production; chemical industry; sterilizing agent (hospitals, spice fumigation) Production; use of chlorophenols and chlorophenoxy herbicides; waste incineration; PCB production; pulp and paper bleaching

others Aflatoxin Involuntary (passive) smoking Mustard gas Strong inorganic-acid mists containing sulfuric acid

*Not necessarily an exhaustive list of occupations/industries in which this agent is found. Not all workers in these occupations/industries are exposed. The term “production” is used to indicate that this substance is man-made and that workers may be exposed in the production process; †This is the most recent IARC evaluation. For those referenced as Supplement 7, it is possible that the 1987 review was quite perfunctory and that the essential evidence was cumulated at an earlier date; ‡As judged by the IARC Working Group. The notation “Not available” was added by the authors to signify those substances for which there was no evidence at all; **As judged by us. Regular script indicates that the evidence for an association with this site was strong. Italics indicate that the evidence was suggestive.

• 110 possible human occupational carcinogens (IARC group 2B)— Table 18–5; • 18 occupations and industries that definitely, probably, or possibly entail excess risk of cancer (IARC groups 1, 2A, and 2B)—Table 18–6. These tables only include agents and circumstances that were reviewed and published by the IARC Monograph Programme as of 2003. There are probably very few agents that have never been evaluated by IARC, but for which there is now significant evidence of carcinogenicity. As discussed above, the evaluations are rooted in the information base that was available at the time of the IARC evaluation. As evidence accumulates, the evaluation of an agent can change, as has already occurred in some cases (e.g., cadmium, acrylonitrile). This is why we have included in the tables a reference

to the IARC volume in which the substance was evaluated and its date. Evaluations with early dates are more vulnerable to being out of date. In a special review published in 1987, all substances and occupations covered in the first 15 years of the Programme were reevaluated (IARC, 1987a). Thus, every substance for which the Supplement 7 reference is cited had an earlier Monograph. For many of the substances, there was little, if any, new information, and consequently, we have quoted the original Monograph for those without any new data in 1987. For those substances referenced as Supplement 7, new data were available for the re-evaluation. For the agents in Tables 18–3 to 18–5, we devised a set of subheadings to help the reader digest the long lists of often obscure chemical names. The subheadings are: physical agents; respirable dusts and fibers; metals and metal compounds; polyaromatic hydrocarbons;

Table 18–4. Substances and Mixtures That Have Been Evaluated by IARC as Probable (Group 2A) Human Carcinogens and Are Occupational Exposures Occupation or Industry in Which Substance Found*

Substance or Mixture

IARC Volume and Year†

Human Evidence‡

Animal Evidence‡

Site(s)**

physical agents Ultraviolet radiation (A, B, and C) from artificial sources

Arc welding; industrial photoprocesses; sterilization and disinfection; phototherapy; operating theatres; research laboratories; UV fluorescence in food industry; insect traps

Vol. 55 (1992b)

Inadequate

Sufficient

Melanoma

Vol. 32 (1983b)

Not available

Sufficient

Vol. 32 (1983b)

Not available

Sufficient

Vol. 32 (1983b)

Not available

Sufficient

Lung Bladder Skin Lung Bladder Skin Lung Bladder Skin

Vol. 35 (1985) Vol. 46 (1989a)

Limited Limited

Sufficient Sufficient

Skin Lung Bladder

Vol. 57 (1993b)

Inadequate

Sufficient

Bladder

Vol. 60 (1994b)

Inadequate

Sufficient

polyaromatic hydrocarbons Benz[a]anthracene Benzo[a]pyrene Dibenz[a,h]anthracene

Work involving combustion of organic matter; foundries; steel mills; firefighters; vehicle mechanics Work involving combustion of organic matter; foundries; steel mills; firefighters; vehicle mechanics Work involving combustion of organic matter; foundries; steel mills; firefighters; vehicle mechanics

wood and fossil fuels and their byproducts Creosotes Diesel engine exhaust

Brickmaking; wood preserving Railroad workers; professional drivers; dock workers; mechanics

intermediates in plastics and rubber manufacturing 4,4¢-Methylene bis(2chloroaniline) (MOCA) Styrene-7,8-oxide

Production; curing agent for roofing and wood sealing Production; styrene glycol production; perfume preparation; reactive diluent in epoxy resin formulations; as chemical intermediate for cosmetics, surface coating, and agricultural and biological chemicals; used for treatment of fibers and textiles; in fabricated rubber products

chlorinated hydrocarbons

a-Chlorinated toluenes Polychlorinated biphenyls Tetrachloroethylene

Production; dye and pesticide manufacture Production; electrical capacitor manufacturing

Vol. 71 (1999a) Suppl. 7 (1987)

Limited Limited

Sufficient Sufficient

Lung Liver and biliary tract

Production; dry cleaning; metal degreasing

Vol. 63 (1995a)

Limited

Sufficient

Cervix Esophagus Non-Hodgkin lymphoma

Trichloroethylene

Production; dry cleaning; metal degreasing

Vol. 63 (1995a)

Limited

Sufficient

Liver and biliary tract Non-Hodgkin lymphoma Renal cell

Chemical industry; water and wastewater treatment; textile, steel, and lumber, industries; petroleum refining; mineral processing; sugar production; hospitals Chemical and rubber industries Production and use of resins, glycerine, and propylene-based rubbers; used as a solvent Production; production of vinyl bromide polymers and monoacrylic fibers for carpet backing material; rubber and plastic production Production; polyvinylfluoride and fluoropolymer production

Vol. 60 (1994b)

Inadequate

Sufficient

Pancreas

Vol. 71 (1999a) Vol. 71 (1999a)

Limited Inadequate

Sufficient Sufficient

Lympho-hematopoietic Lung CNS

Vol. 71 (1999a)

Not available

Sufficient

Vol. 63 (1995a)

Not available

Sufficient

Suppl. 7 (1987)

Inadequate

Sufficient

Bladder

Vol. 77 (2000b) Vol. 77 (2000b)

Limited Limited

Sufficient Sufficient

Bladder Bladder

Production; manufacture of pharmaceuticals, pesticides and dyes

Vol. 71 (1999)

Inadequate

Sufficient

Production; fungicide Production; pest control; petroleum refining and waterproofing; in leaded gasoline; chemical intermediate and solvent in gums, waxes, resins, dyes, and pharmaceutical preparations

Vol. 53 (1991c) Vol. 71 (1999a)

Not available Inadequate

Sufficient Sufficient

monomers Acrylamide

1,3-Butadiene Epichlorohydrin Vinyl bromide Vinyl fluoride

aromatic amine dyes Benzidine-based dyes 4-Chloro-ortho-toluidine ortho-Toluidine

Production; used in textile, paper, leather, rubber, plastics, printing, paint and lacquer industries Dye and pigment manufacture; textile industry Production; manufacture of dyestuffs, pigments, optical brightener, pharmaceuticals and pesticides; rubber vulcanizing; clinical laboratory reagent; cleaners and janitors.

intermediates in the production of dyes Dimethylcarbamoyl chloride

pesticides Captafol Ethylene dibromide

328

329

Occupation Table 18–4. (cont.) Substance or Mixture Non-arsenical insecticides

Occupation or Industry in Which Substance Found*

IARC Volume and Year†

Human Evidence‡

Animal Evidence‡

Production; pest control and agriculture workers; flour and grain mill workers

Vol. 53 (1991c)

Limited

Not available

Diethyl sulfate Formaldehyde

Ethanol production Production; pathologists; medical laboratory technicians; plastics; textile industry

Vol. 71 (1999a) Vol. 62 (1995b)

Not available Limited

Sufficient Sufficient

Tris(2,3-dibromopropyl) phosphate

Production; used in the textile industry; in phenolic resins (for electronics industry), paints, paper coatings and rubber.

Vol. 71 (1999a)

Inadequate

Sufficient

Site(s)** Brain Leukemia Lung Multiple myeloma Non-Hodgkin lymphoma

others Leukemia Nasal sinuses Nasopharynx

*Not necessarily an exhaustive list of occupations/industries in which this agent is found. Not all workers in these occupations/industries are exposed. The term “production” is used to indicate that this substance is man-made and that workers may be exposed in the production process; †This is the most recent IARC evaluation. For those referenced as Supplement 7, it is possible that the 1987 review was quite perfunctory and that the essential evidence was cumulated at an earlier date; ‡As judged by the IARC Working Group. The notation “Not available” was added by the authors to signify those substances for which there was no evidence at all; **We judged CNS, central nervous system that the evidence was suggestive.

wood and fossil fuels and their byproducts; monomers; intermediates in plastics and rubber manufacturing; chlorinated hydrocarbons; aromatic amine dyes; azo dyes; intermediates in the production of dyes; pesticides; nitro compounds; others. Tables 18–3 to 18–5 indicate some of the main occupations or industries in which each listed substance is found, and the strength of evidence from human and animal studies. In Tables 18–3 and 18–4, we show the type(s) of cancer affected, with an indication of the strength of evidence for each type listed. Information on target organ is not shown in Table 18–5 because, for agents listed as possible carcinogens, evidence concerning humans is either conflicting or not available at all. The Monograph Programme has occasionally addressed cancer risk in various occupations and industries, as well as agents. However, whereas the Monograph programme aims at a systematic evaluation of agents and complex mixtures, it is not intended to provide a systematic review of cancer risk by industries and occupations. That is, those reviews were conducted where there were particular concerns or anticipated insights regarding specific potential carcinogens. Sometimes this was done when there appeared to be strong evidence of risk in an occupation, but little indication of what the responsible agent might be (e.g., rubber industry; painter). Sometimes the impetus for an occupation or industry review came from the attempt to evaluate some agent, but the realization occurred that the evidence regarding that agent was rooted in epidemiological evidence regarding some occupation or industry (e.g., glass industry; hairdresser). Table 18–6 shows those that IARC has evaluated as definitely, probably, or possibly entailing a carcinogenic risk. Because there has been no pretense of exhaustiveness in evaluating occupations and industries, the absence of an occupation or industry in Table 18–6 does not carry the same significance as the absence of an agent in Tables 18–3 to 18–5. That is, it does not signify that there is no known risk for that occupation or industry. Since our inclusion criteria admitted substances to which workers were exposed in the past, we included some substances that have been banned or virtually eliminated in some countries, such as mustard gas, bis(chloromethyl)ether, tris(2,3-dibromopropyl) phosphate, and 4,4¢-methylene bis(2-chloroaniline) (MOCA), as well as some industries that no longer exist (e.g., production of auramine and magenta). These are mentioned partly for historic interest, and partly because it is possible that these might yet be used in some places at some time. It is important to note that the substances, occupations, and industries listed in Tables 18–3 to 18–6 are not mutually exclusive. Certainly, some of the occupations and industries listed in Table 18–6 may be there because of some of the substances that are listed in Tables

18–3 to 18–5. But further, the substances relate to each other in complicated ways. There are some families of substances that include some specific substances that are also listed (e.g., non-arsenical insecticides, which includes DDT; benzidine-based dyes, which includes benzidine). Also there are some complex mixtures (e.g., diesel exhaust) that contain some substance on the list (e.g., nitro-PAHs) and that may be responsible for the carcinogenicity of the mixture. The listing of affected cancer sites in Tables 18–3 and 18–4 does not come explicitly out of the IARC Monographs. Sometimes the affected target organ(s) was rather evident, but sometimes it required that we evaluate the evidence, including evidence published more recently than the IARC evaluation in question. Table 18–7 shows the same agents listed in Tables 18–3 and 18–4, but organized by site of cancer. As we did in the previous tables, we indicate clearly which associations are strong and which are only suggestive. Lung cancer is the target organ that has most often been identified.

The Evolution of Knowledge To appreciate how knowledge has evolved, we searched for information on the current occupational carcinogens at two earlier time periods. As mentioned above, IARC carried out a comprehensive cumulative synthesis in 1987 (IARC, 1987a). In that report, the results were presented with the same rating system (1, 2A, 2B, 3) as is used today, rendering the lists comparable. In 1964, even before the establishment of IARC, the World Health Organization commissioned an expert panel to survey available knowledge on human carcinogens (WHO, 1964). In the WHO report, there was no explicit rating system. It was a discursive presentation of knowledge and opinions that we attempted, with some license, to translate into a simple system corresponding to definite, probable/possible, or not mentioned. From these two reports, we searched for references to the 168 substances presented in Tables 18–3 to 18–5, and that are currently considered to be definite, probable, or possible occupational carcinogens. Table 18–8 shows how the current occupational carcinogens were considered in two earlier times. Half of today’s recognized definite occupational carcinogens were already recognized as such by 1964, in the early period of cancer epidemiology. Nearly 90% were considered to be definite or probable as of 15 years ago. In contrast, over 95% of today’s probable and possible occupational carcinogens had not even been mentioned as of 1964, and about one-third were not mentioned as of 1987. While it is possible for the classification of agents to change over time in either direction, in practice there have been rather few instances of agents being “downgraded” between successive periods. Notable counter-examples are:

Table 18–5. Substances and Mixtures That Have Been Evaluated by IARC as Possible (Group 2B) Human Carcinogens, and Are Occupational Exposures Substance or Mixture

Occupation or Industry in Which Substance Found*

IARC Volume and Year†

Human Evidence‡

Animal Evidence‡

respirable dusts and fibers Palygorskite (long fibers >5 mm) Refractory ceramic fibers Special-purpose glass fibers such as E-glass and “475” glass fibers

Miners and millers; production of waste absorbents, fertilizers, and pesticides Production; furnace insulators; ship builders; heat resistant fabric manufacture High-efficiency air filtration media; battery separator media

Vol. 68 (1997b) Inadequate

Sufficient

Vol. 81 (2002a)

Inadequate

Sufficient

Vol. 81 (2002a)

Not available

Sufficient

Vol. 47 (1989c) Vol. 52 (1991a) Suppl. 7 (1987)

Inadequate Inadequate Inadequate

Sufficient Sufficient Sufficient

Vol. 58 (1993a) Vol. 49 (1990a)

Inadequate Inadequate

Sufficient Sufficient

Vol. 63 (1995a)

Not available

Sufficient

Suppl. 7 (1987)

Inadequate

Sufficient

Vol. 65 (1996) Inadequate Vol. 45 (1989b) Inadequate Vol. 45 (1989b) Inadequate

Sufficient Limited Sufficient

Vol. 45 (1989b) Inadequate Vol. 46 (1989a) Inadequate

Limited Limited

Vol. 82 (2002c)

Inadequate

Sufficient

Vol. 32 (1983b) Vol. 32 (1983b) Vol. 32 (1983b) Vol. 32 (1983b)

Not available Not available Not available Not available

Sufficient Sufficient Sufficient Sufficient

Vol. 32 (1983b) Not available

Sufficient

Vol. 32 (1983b) Not available

Sufficient

Vol. 32 (1983b) Not available

Sufficient

Vol. 32 (1983b) Not available Vol. 32 (1983b) Not available

Sufficient Sufficient

Vol. 71 (1999a) Vol. 71 (1999a) Vol. 39 (1986a) Vol. 71 (1999a) Vol. 82 (2002c)

Inadequate Inadequate Not available Not available Limited

Sufficient Sufficient Sufficient Sufficient Limited

Vol. 71 (1999a)

Inadequate

Sufficient

Vol. 7 (1974a) Vol. 63 (1995a)

Not available Not available

Sufficient Limited

Vol. 71 (1999a)

Inadequate

Sufficient

Vol. 71 (1999a) Vol. 16 (1978)

Not available Not available

Sufficient Sufficient

Vol. 71 (1999a) Not available Vol. 77 (2000b) Inadequate Vol. 79 (2001b) Inadequate

Limited Sufficient Sufficient

Vol. 71 (1999a) Not available Vol. 60 (1994b) Inadequate

Sufficient Sufficient

Vol. 71 (1999a)

Sufficient

metals and metal compounds Antimony trioxide Cobalt and cobalt compounds Lead and inorganic lead compounds Methylmercury compounds Nickel—metallic and alloys

Ore processing; glass and ceramic production Miners; processing of copper and nickel ore; glass and ceramic production Lead smelters; plumbers; solderers; occupations in battery recycling smelters Pesticide and fungicide production; paint industry Nickel miners; metal fabrication, grinding, electroplating, and welding

wood and fossil fuels and their byproducts Benzofuran Bitumens, extracts of steamrefined and air-refined Carbon black Diesel fuel, marine Fuel oils residual (heavy) Gasoline Gasoline engine exhaust Naphthalene

Production; intermediate in coumarone-indene resin polymerization; coke production; coal gasification and combustion Production/refining; road construction; roofing and flooring Production; paint, ink, plastic, and rubber industries Petroleum refineries; marine fuel; distribution Petroleum refineries; distribution; marine fleet; majority of large diesel engines operated on land; industrial heating systems Petroleum refineries; transportation; mechanics and service station attendants Transportation and vehicle maintenance workers; drivers; toll attendants; traffic controllers Production; insecticide, resin, and pharmaceutical production

polyaromatic hydrocarbons Benzo(b)fluoranthene Benzo( j)fluoranthene Benzo(k)fluoranthene Dibenz[a,h]acridine Dibenz(a,j)acridine Dibenzo(a,e)pyrene Dibenzo(a,h)pyrene Dibenzo(a,i)pyrene Dibenzo(a,l)pyrene

Work involving combustion of organic matter Work involving combustion of organic matter Work involving combustion of organic matter Production; used in dye synthesis; biochemical laboratory workers; work involving combustion of organic matter Production; used in dye synthesis; work involving combustion of organic matter Production; biochemical laboratory workers; work involving combustion of organic matter Production; biochemical laboratory workers; work involving combustion of organic matter Work involving combustion of organic matter Production; biochemical laboratory workers; work involving combustion of organic matter

monomers Acrylonitrile Chloroprene Ethyl acrylate Isoprene Styrene Toluene diisocyanates Urethane Vinyl acetate

Production; acrylic textile fiber and plastic production Production; manufacture of polychloroprene (synthetic rubber) Production; plastic molding occupations using acrylate resins Production; synthetic rubber and plastics industries Polyester resin manufacture, production of packaging materials and fiberglass-reinforced polyester Production; production of polyurethane foams and wire coating; insulation workers; ship builders Production; amino-resin production Production; plastics, paint, and adhesive industries

intermediates in plastics and rubber manufacturing Acetaldehyde Acetamide 2,4-Diaminotoluene 1,2-Epoxybutane Ethylbenzene Ethylene thiourea Phenyl glycidyl ether Propylene oxide

Acetic acid production workers; dyestuff, plastic, and synthetic rubber industries Production; plastics and chemical industries Production; chemical intermediate in TDI production; dyes for textiles; leather; furs; wood; biological stain; photo developer Production; metal degreasing; plastics industry Production; ink, paint and plastic production Production; vulcanization in the rubber industry; manufacture of ethylenebisdithiocarbamate pesticides; electroplating baths; dyes, pharmaceuticals; synthetic resins Production; epoxy resins; casting and molding Production; polyurethane foam and glycol production; fumigant

chlorinated hydrocarbons Carbon tetrachloride

330

Production; industrial degreasing occupations; dry cleaners; refrigerant production

Inadequate

Table 18–5. (cont.) Substance or Mixture Chlorinated paraffin of average carbon-chain length C12 Chloroform 1,2-Dichloroethane Dichloromethane Hexachloroethane

Occupation or Industry in Which Substance Found*

IARC Volume and Year†

Human Evidence‡

Animal Evidence‡

Production; PVC processing industry

Vol. 48 (1990b) Not available

Sufficient

Refrigerant production; dyes, solvents, and pesticides Vinyl chloride production workers Production; painters and furniture restorers; pharmaceutical and electronic production Production; aluminum refinery; industrial firefighters

Vol. 73 (1999b) Inadequate Vol. 71 (1999a) Inadequate Vol. 71 (1999a) Inadequate

Sufficient Sufficient Sufficient

Vol. 73 (1999b) Inadequate

Sufficient

Suppl. 7 (1987) Inadequate Vol. 16 (1978) Not available Vol. 57 (1993b) Inadequate

Sufficient Sufficient Sufficient

Vol. 79 (2001b) Not available Vol. 1 (1972) Not available

Sufficient Sufficient

Vol. 57 (1993b) Vol. 29 (1982c) Vol. 29 (1982c) Vol. 48 (1990b) Vol. 57 (1993b) Vol. 39 (1986a) Vol. 57 (1993b)

Not available Inadequate Not available Not available Not available Not available Not available

Sufficient Sufficient Sufficient Sufficient Sufficient Sufficient Sufficient

Vol. 8 (1975) Suppl. 7 (1987) Vol. 57 (1993b) Vol. 57 (1993b) Vol. 8 (1975) Vol. 8 (1975) Vol. 8 (1975) Vol. 8 (1975) Vol. 8 (1975)

Not available Not available Not available Not available Not available Not available Not available Not available Not available

Sufficient Sufficient Sufficient Sufficient Sufficient Sufficient Sufficient Sufficient Sufficient

Vol. 8 (1975)

Not available

Sufficient

Production; manufacture of dyes, pigments, and perfumes Production; manufacture of dyes and pigments; dye for leather, paper, plastics, rubber, textiles, and laboratories Production; synthesis of anthraquinone dyes

Vol. 27 (1982b) Not available Suppl. 7 (1987) Inadequate

Sufficient Sufficient

Vol. 27 (1982b) Not available

Sufficient

Production; manufacture of dyes and pigments

Suppl. 7 (1987)

Inadequate

Sufficient

Production; manufacture of the dye intermediates ortho-anisidine and ortho-dianisidine Production; manufacture of dyes

Vol. 65 (1996)

Not available

Sufficient

Vol. 27 (1982b) Not available

Sufficient

Production; manufacture of diisocyanates and munitions Production; manufacture of diisocyanates and munitions Production; manufacture of dyestuffs, detergents, and cosmetics Underground miners using diesel-powered machinery Production; ink, paint, and explosives industries Production; manufacture of azidopyrene; particulate emissions Production; used only as a laboratory chemical; probably present before 1980 in carbon black used in photocopy machines Production; diesel fuel additive; TNT manufacturing

Vol. 65 (1996) Vol. 65 (1996) Vol. 65 (1996) Vol. 46 (1989a) Vol. 71 (1999a) Vol. 46 (1989a) Vol. 46 (1989a)

Inadequate Inadequate Not available Not available Not available Not available Not available

Sufficient Sufficient Sufficient Sufficient Sufficient Sufficient Sufficient

Vol. 65 (1996)

Not available

Sufficient

Production; in miticides in greenhouses, nurseries, and orchards Production; termite control Production; insecticide Production; defoliant Production; fungicide, bactericide, and nematocide Production; nonsystemic insecticide Production; pesticide, nematocide, and soil fumigant Production; pesticide Production; insecticide and miticide Production; termite control

Vol. 5 (1974b) Vol. 79 (2001b) Vol. 20 (1979a) Suppl. 7 (1987) Vol. 73 (1999b) Vol. 53 (1991c) Vol. 71 (1999a) Vol. 73 (1999b) Vol. 53 (1991c) Vol. 79 (2001b)

Not available Inadequate Not available Limited Not available Inadequate Inadequate Inadequate Inadequate Inadequate

Sufficient Sufficient Sufficient Inadequate Sufficient Sufficient Sufficient Sufficient Sufficient Sufficient

aromatic amine dyes Auramine (technical grade) Benzyl violet 4B CI Basic Red 9

Production; textiles; plastic; printing Production; food, drugs; cosmetics; textiles Production; textiles and printing; biological stains (basic fushin dye in laboratories) 2,4-Diaminoanisole Dyestuff industry; barbers and cosmetologists; furriers 3,3¢-Dimethylbenzidine Production; dye or intermediate in dye and pigment production, (o-tolidine) polyurethane elastomers, coating, plastics, clinical laboratories 2,6-Dimethylaniline (2,6-xylidine) Production; dyestuffs and pharmaceutical manufacturing 3,3¢-Dichlorobenzidine Production; dyestuff manufacturing 4,4¢-Diaminodiphenyl ether Production; polyamide-type resin manufacturing Disperse Blue 1 Production; hair coloring; textiles and plastics HC Blue No. 1 Production; hair dye 4,4¢-Methylenedianiline Production; production of diisocyanates, polyisocyanates, and epoxy resins Magenta containing CI Production; textiles and printing; biological stains Basic Red 9 in laboratories; photography

azo dyes ortho-Aminoazotoluene para-Aminoazobenzene CI Acid Red 114 CI Direct Blue 15 Citrus red No. 2 para-Dimethylaminoazobenzene Oil orange SS Ponceau 3R Ponceau MX Trypan blue

Production; textiles and leather Production; textiles and leather Production; textiles and leather Production; textiles and paper Production; used for food coloring Production; textiles; laboratories Production; dyes/pigments for varnishes, oils, fats, and waxes Production; textiles Production; textiles; leather; inks; paper; wood stains; food; biology laboratories Production; textiles and printing; biological stains in life science laboratories; used by ophthalmologists

intermediates for manufacture of dyes para-Cresidine 3,3¢-Dimethoxybenzidine (ortho-dianisidine) 2-Methyl-1-nitroanthraquinone (of uncertain purity/impurity) 4,4¢-Methylene bis(2methylaniline) 2-Nitroanisole 4,4¢-Thiodianiline

nitro compounds 2,4-Dinitrotoluene 2,6-Dinitrotoluene Nitrobenzene 2-Nitrofluorene 2-Nitropropane 1-Nitropyrene 4-Nitropyrene Tetranitromethane

pesticides Aramite Chlordane Chlordecone Chlorophenoxy herbicides Chlorothalonil DDT (p,p¢-DDT) 1,2-Dibromo-3-chloropropane para-Dichlorobenzene Dichlorvos Heptachlor

(continued)

331

332

PART III: THE CAUSES OF CANCER

Table 18–5. (cont.) Substance or Mixture Hexachlorobenzene

Hexachlorocyclohexanes (most common form is Lindane) Mirex Nitrofen Sodium ortho-phenylphenate Toxaphene (polychloronated camphenes)

Occupation or Industry in Which Substance Found* Production; in chlorinated pesticides and fungicides; dye manufacture and synthesis of organic chemicals and rubber; plasticizer for PVC; wood preservative; byproduct of the production of a number of chlorinated solvents Production; woodworkers; farm workers Production; fire-retardant additive; insecticide; workers at hazardous waste sites Production; herbicide Production; fungicide; chemical intermediate Production; insecticide

IARC Volume and Year†

Human Evidence‡

Animal Evidence‡

Vol. 79 (2001b) Inadequate

Sufficient

Suppl. 7 (1987)

Inadequate

Sufficient

Vol. 20 (1979a)

Not available

Sufficient

Vol. 30 (1983a) Not available Vol. 73 (1999b) Not available Vol. 79 (2001b) Inadequate

Sufficient Sufficient Sufficient

Vol. 40 (1986c) Vol. 71 (1999a) Vol. 71 (1999a)

Not available Not available Not available

Sufficient Sufficient Sufficient

Vol. 71 (1999a) Vol. 71 (1999a)

Inadequate Inadequate

Sufficient Sufficient

others Butylated hydroxyanisole (BHA) Catechol Diglycidyl resorcinol ether 1,4-Dioxane Hydrazine

Nitrilotriacetic acid and its salts Polychlorophenols and their sodium salts (mixed exposure) Potassium bromate Thiourea Welding fumes

Production; food and pharmaceutical industries Production; insecticide and pharmaceutical production; tanneries Production; liquid spray epoxy resin in electrical, tooling, adhesive, laminating applications; production of epoxy resins, rubber; aerospace industry Production; chlorinated solvents; textile processing; mixed with pesticides Production; manufacture of agricultural chemicals and chemical blowing agents; water treatment; spandex fibers; rocket fuel; oxygen scavenger in water boilers and heating systems; scavenger for gases; plating metals on glass and plastics; solder fluxes; photographic developers; reactant in fuel cells in the military; reducing agent in electrodeless nickel plating; chain extender in urethane; textile dyes; explosives Production; textiles; electroplaters; tanners Herbicide production, wood, textile, and leather manufacturing Production; bakeries Production; photoprocessing; dyes; rubber industry Metal fabricating industry

Vol. 73 (1999b) Not available Vol. 71 (1999a) Limited

Sufficient Inadequate

Vol. 73 (1999b) Not available Vol. 79 (2001b) Not available Vol. 49 (1990a) Limited

Sufficient Sufficient Inadequate

*Not necessarily an exhaustive list of occupations/industries in which this agent is found. Not all workers in these occupations/industries are exposed. The term “production” is used to indicate that this substance is man-made and that workers may be exposed in the production process; †This is the most recent IARC evaluation. For those referenced as Supplement 7, it is possible that the 1987 review was quite perfunctory and that the essential evidence was cumulated at an earlier date; ‡As judged by the IARC Working Group. The notation “Not available” was added by the authors to signify those substances for which there was no epidemiological evidence at all.

• 3,3 dichlorobenzene, which was considered a definite carcinogen in 1964 and was only considered as possible as of 1987 and as of 2002; • acrylonitrile and propylene oxide, which were considered probable carcinogens in 1987, and only as possible in 2002; • ionizing radiation, which is a special case, was considered a definite carcinogen in 1964 and is so considered today; but it had not been reviewed by IARC before the 1990s, so we had to classify it as “unrated” in 1987. The number of occupational agents rated by IARC as group 1 carcinogens has tapered off since 1987, while the proportion of group 2B evaluations increased. This reflects the fact that, when the Monograph Programme began, there was a “backlog” of agents for which strong evidence of carcinogenicity had accumulated, and, naturally, these were the agents that IARC initially selected for review. Once the agents with strong evidence had been dealt with, IARC started dealing with others.

Interpreting the Lists There is sometimes a tendency to interpret tables of carcinogens in too categorical a fashion. Although it may be convenient for lobbyists and regulators to divide the world of chemicals and occupational circumstances into “good guys” and “bad guys,” such a dichotomy is simplistic. The determination that a substance or circumstance is carcinogenic depends on the strength of evidence at a given point in time. The evidence is sometimes clear-cut (which would correspond

to evaluations in group 1) but more often it is not. The balance of evidence can change in either direction as new data emerge. The characterization of an occupation or industry group as a “high-risk group” is strongly rooted in time and place. For instance, the fact that some groups of nickel refinery workers experienced excess risks of nasal cancer does not imply that all workers in all nickel refineries will be subject to such risks. The particular circumstances of the industrial process, raw materials, impurities, and control measures may produce risk in one nickel refinery but not in another or in one historic era but not in another. The same can be said of rubber production facilities, aluminum refineries, and other industries and occupations. Labeling a chemical substance as a carcinogen in humans is a more timeless statement than labeling an occupation or industry as a high-risk group. However, even such a statement requires qualification. Different carcinogens produce different levels of risk and for a given carcinogen there may be vast differences in the risks incurred by different people exposed under different circumstances. Indeed there may also be interactions with other factors, environmental or genetic, that produce no risk for some exposed workers and high risk for others. This raises the issue of quantitative risk assessment, which is an important tool in preventing occupational cancer. Unfortunately, our tables provide no basis for gauging the strength of the effect of each carcinogen, either in relative risk or in absolute risk terms or in terms of dose-response relationships. The IARC evaluations provide no such indications, and while it would be most precious to have such information, for many agents, the information base on dose-response to support such quantification is fragmentary.

333

Occupation

Table 18–6. Occupations or Industries Evaluated by IARC as Definitely (Group 1) or Probably (Group 2A) or Possibly (Group 2B) Entailing Excess Risk of Cancer among Workers Occupation or Industry

Suspected Substance

IARC Volume and Year*

Group

Site(s)† Lung Bladder Bladder Leukemia Nose and paranasal sinuses Bladder

Aluminium production

Pitch volatiles; aromatic amines

Suppl. 7 (1987)

1

Auramine manufacture Boot and shoe manufacture and repair

2-Naphthylamine; auramine; other chemicals; pigments Leather dust; benzene and other solvents

Suppl. 7 (1987) Suppl. 7 (1987)

1 1

Carpentry and joinery Coal gasification

Wood dust Coal tar; coal-tar fumes; PAHs

Suppl. 7 (1987) Vol. 34 (1984b)

2B 1

Coke production

Coal-tar fumes

Suppl. 7 (1987)

1

Dry cleaning Furniture and cabinet making Hairdressers and barbers

Solvents and chemicals used in “spotting” Wood dust Dyes (aromatic amines, amino-phenols with hydrogen peroxide); solvents; propellants; aerosols

Vol. 63 (1995a) Suppl. 7 (1987) Vol. 57 (1993b)

2B 1 2A

Hematite mining, underground, with radon exposure Iron and steel founding Isopropanol manufacture, strong-acid process

Radon daughters; silica

Suppl. 7 (1987)

1

PAHs, silica, metal fumes; formaldehyde Diisopropyl sulfate; isopropyl oils; sulfuric acid

Suppl. 7 (1987) Suppl. 7 (1987)

1 1

Magenta; ortho-toluidine; 4,4¢-methylene bis(2-methylaniline); ortho-nitrotoluene

Vol. 57 (1993b)

1

Vol. 47 (1989c)

1

Magenta manufacture Painters Petroleum refining

PAHs

Vol. 45 (1989b)

2A

Printing processes Production of art glass, glass containers and pressed ware Rubber industry

Solvents, inks Lead; arsenic; antimony oxides; silica; asbestos; other metal oxides; PAHs Aromatic amines; solvents

Vol. 65 (1996) Vol. 58 (1993a)

2B 2A

Suppl. 7 (1987)

1

Textile manufacturing industry

Textile dust in manufacturing process; dyes and solvents in dyeing and printing operations

Vol. 48 (1990b)

2B

Skin (including scrotum) Bladder Lung Skin (scrotum) Lung Bladder Kidney Nose and sinonasal cavities Bladder Lung Non-Hodgkin lymphoma Ovary Lung Lung Paranasal sinuses Larynx Lung Bladder Lung Bladder Stomach Bladder Brain Leukemia skin Lung Bladder Stomach Larynx Leukemia Lung

*This is the most recent IARC evaluation. For those referenced as Supplement 7, it is possible that the 1987 review was quite perfunctory and that the essential evidence was cumulated at an earlier date; †As judged by us. Regular script indicates that the evidence for an association with this site was strong. Italics indicate that the evidence was suggestive. PAH, polycyclic aromatic hydrocarbons.

Incompleteness of Current Lists Many of the recognized definite occupational carcinogens were first suspected before the era of modern epidemiology (i.e., before 1950). The significance of this observation is unclear. It may be that there were only a limited number of strong occupation-cancer associations, and these were sufficiently obvious that they could produce observable clusters of cases for astute clinicians to notice. It may be that levels of exposure to occupational chemicals were so high before the 1950s as to produce high cancer risks and cancer clusters, but that improvements in industrial hygiene in industrialized countries have indeed decreased risks to levels that are difficult to detect. While the evaluation of the hypothesis of an agent causing human cancer depends critically on epidemiological and experimental evidence, the initial suspicion can be provoked by epidemiological surveillance, by experimental evidence, or by clinical cluster observations. Indeed, most definite occupational carcinogens were first suspected on the basis of case reports by clinicians or pathol-

ogists (Doll, 1975). These discoveries were usually coincidental (Siemiatycki et al., 1981). It is thus reasonable to suspect that there may be some, perhaps many, as yet undiscovered occupational carcinogens.

REVIEW ARTICLES One can conceive of a hypothetical matrix with hundreds of occupational circumstances (occupations or occupational chemicals) on one axis, and scores of different types of cancer on the other axis. There will be thousands of cells in such a matrix, and this is the scope of the literature that pertains to this chapter. The published literature in the field of occupational cancer includes hundreds of publications per year, potentially addressing hundreds of possible associations between exposure (occupations or substances) and disease (different types of cancer). The more recent IARC Monographs include good reviews, and these are referenced in Tables 18–3 to 18–6. To aid readers in finding material relevant to their interests, we have compiled a

334

PART III: THE CAUSES OF CANCER

Table 18–7. Definite or Probable Occupational Carcinogens and Carcinogenic Circumstances, Categorized by Site Site

Strength of Evidence*

Pharynx and nasopharynx Nasal cavities and paranasal sinuses

Suggestive Strong

Liver (hepatocellular) Pancreas Larynx

Suggestive Suggestive Suggestive Suggestive Strong Suggestive Strong Suggestive Suggestive Suggestive Strong

Lung

Suggestive Strong

Esophagus Stomach Gastrointestinal tract Liver and biliary tract Liver (angiosarcoma)

Suggestive

Lung (oat cell) Bone Melanoma Skin

Strong Strong Strong Suggestive Strong

Mesothelioma CNS Sarcoma Cervix Ovary Kidney Kidney (renal cell) Bladder

Suggestive Strong Suggestive Suggestive Suggestive Suggestive Suggestive Suggestive Strong Suggestive

Brain Thyroid Non-Hodgkin lymphoma

Suggestive Strong Suggestive

Lympho-hematopoietic system Multiple myeloma Leukemia

Suggestive Suggestive Strong Suggestive Suggestive Strong

Other sites All sites combined

High-Risk Substance or Circumstance Mustard gas; formaldehyde Boot and shoe manufacture and repair; furniture and cabinet making; isopropanol manufacture, strong acid process; selected nickel compounds, including combinations of nickel oxides and sulfides in the nickel refining industry; wood dust Chromium compounds, hexavalent; formaldehyde; mineral oils, untreated and mildly treated Soots; tetrachloroethylene Painters; rubber industry Asbestos Aflatoxin; ionizing radiation Polychlorinated biphenyls; trichloroethylene Vinyl chloride Arsenic and arsenic compounds Vinyl chloride Acrylamide Isopropanol manufacture, strong acid process; inorganic acid mists containing sulfuric acid; mustard gas Asbestos; rubber industry Aluminum production; arsenic and arsenic compounds; asbestos; beryllium; cadmium and cadmium compounds; chromium compounds, hexavalent; coal gasification; coke production; hematite mining, underground, with radon exposure; involuntary (passive) smoking; ionizing radiation; iron and steel founding; selected nickel compounds, including combinations of nickel oxides and sulfides in the nickel refining industry; painters; silica, crystalline; soots; talc containing asbestiform fibers Benz[a]anthracene; benzo[a]pyrene; a-chlorinated toluenes; coal tars and pitches; dibenz[a,h]anthracene; diesel engine exhaust; epichlorohydrin; hairdressers and barbers; inorganic acid mists containing sulfuric acid; isopropanol manufacture, strong acid process; mineral oils, untreated and mildly treated; non-arsenical insecticides;mustard gas; production of art glass, glass containers and pressed ware; rubber industry; 2,3,7,8-tetrachlorodibenzopara-dioxin (TCDD) Bis(chloromethyl)ether and chloromethyl methyl ether (technical grade) Ionizing radiation Solar radiation Ultraviolet radiation (A, B, and C) from artificial sources Arsenic and arsenic compounds; coal tars and pitches; coal gasification; coke production; dibenz[a,h]anthracene; mineral oils, untreated and mildly treated; shale oils or shale-derived lubricants; solar radiation; soots Benz[a]anthracene; benzo[a]pyrene; creosotes Asbestos; erionite; talc containing asbestiform fibers Epichlorohydrin 2,3,7,8-Tetrachlorodibenzo-para-dioxin (TCDD) Tetrachloroethylene Hairdressers and barbers Coke production Trichlorethylene Aluminum production; 4-aminobiphenyl; auramine manufacture; benzidine; coal gasification; magenta manufacture; 2-naphthylamine; rubber industry Benz[a]anthracene; benzidine-based dyes; benzo[a]pyrene; boot and shoe manufacture and repair; 4-chloro-ortho-toluidine; coal tars and pitches; coke production; di-benz[a,h]anthracene; diesel engine exhaust; hairdressers and barbers; 4,4¢-methylene bis(2-chloroaniline) (MOCA); mineral oils, untreated and mildly treated; ortho-toluidine; painters; petroleum refining Non-arsenical insecticides; petroleum refining Ionizing radiation Hairdressers and barbers; non-arsenical insecticides; 2,3,7,8-tetrachlorodibenzo-paradioxin (TCDD), tetrachloroethylene; trichloroethylene 1,3-Butadiene Non-arsenical insecticides Benzene; boot and shoe manufacture and repair; ethylene oxide; ionizing radiation Formaldehyde; non-arsenical insecticides; petroleum refining; rubber industry Ionizing radiation† 2,3,7,8-Tetrachlorodibenzo-para-dioxin (TCDD)‡

*This designation refers to the strength of evidence regarding each site, as judged by us; †There is suggestive evidence of an effect of ionizing radiation on several sites in addition to those with which it is associated in this table; ‡The evidence for an association with TCDD only becomes strong when data are combined for all cancer sites. CNS, central nervous system.

bibliography of relatively recent review articles that complement the IARC references in those tables. Table 18–9 contains references to articles on occupational circumstances that have not recently been evaluated by IARC, that complement the IARC reviews, or that concern occupational risk factors for specific sites of cancer. Some of these are review articles or chapters, and some are original contributions that happen to include a good review section.

COMMENTS AND CONTROVERSIES REGARDING SPECIFIC TOPICS As it is not feasible to discuss in detail all the possible associations between cancer and occupational exposure, we have selected a small number of topics for special attention. These tend to be occupational circumstances where the evidence is particularly controversial, or

335

Occupation Table 18–8. Evolution in Knowledge Regarding Current (2003) IARC Occupational Carcinogens Earlier Evaluation Current Rating

Past Rating

1 (n = 28)

Table 18–9. Recent Reviews on Selected Topics Related to Occupation and Cancer* Topic

IARC 1987

WHO 1964

1 2A 2B 3 Unrated Total

19 4 1 0 4 28

册

13

1 2A 2B 3 Unrated Total

0 16 6 2 3 27

册

1 2A 2B 3 Unrated Total

0 2 63 9 39 113

册

cancer sites Nasal cavity and pharynx

2A (n = 27)

2B (n = 110)

Esophagus 4 n.a. 11 28 0

Stomach Colorectum

0 n.a. 27 27 1 5 n.a. 107 113

Liver Pancreas

Larynx Lung Bone Melanoma Mesothelioma

where there are general lessons that can be learned from their consideration.

Soft-tissue sarcoma Breast

Polycyclic Aromatic Hydrocarbons (PAHs) Polycyclic aromatic hydrocarbons comprise a large family of chemical compounds that are produced during incomplete combustion of organic material, and in particular fossil fuels. PAHs are found in many occupations and industries, and they are found in such nonoccupational settings as vehicle roadways, homes heated by burning fuel, barbecued foods, cigarette smoke, and many more. As described above, the earliest known occupational carcinogens were coal-derived soots, oils, and fumes that caused skin cancers. Animal experiments showed that several of the chemicals found in these complex mixtures were carcinogenic. These chemicals were in the family of polycyclic aromatic hydrocarbons. When epidemiologic evidence accumulated on lung cancer risks among workers exposed to complex mixtures derived from coal, petroleum, and wood, it was widely felt that the responsible agents were likely to be PAHs. Several of the complex mixtures (coal tars and pitch, mineral oils, shale oils, soots) that are classified as IARC group 1 carcinogens include PAHs and several of the industries in which cancer risks have been identified (e.g., coal gasification, coke production, aluminum production, iron and steel founding) are industries in which PAHs are prevalent. Paradoxically, however, there are no specific PAHs on the group 1 list. The highest classed PAHs are in group 2A. This is because it is virtually impossible to epidemiologically isolate the effect of one versus another of the components of these carcinogenic mixtures. Because of the non-feasibility of measuring all PAHs when they are measured for industrial hygiene purposes, there has typically been a single marker of PAHs, often benzo(a)pyrene. While this marker may be available for epidemiologic purposes, it cannot be assumed that this is the only PAH present or how its presence is correlated with those of other PAHs. Similar considerations apply to urinary 1-OH-pyrene, the most widely used biomarker of internal PAH dose, whose excretion depends on the composition of the mixture of PAH and on metabolic pathways under the control of polymorphic genes. It is possible that biomarker and genetic studies will provide the additional information that would permit the determination that specific PAHs are definite human carcinogens.

Review

Uterine cervix and endometrium Ovary Prostate Testis Kidney Bladder Brain

Multiple myeloma Hematopoietic and lymphatic tumors

Maier and Tisch, 2000 Yu and Yuan, 2002 Sullivan et al., 1998 Parent et al., 2000 Engel et al., 2002 Cocco et al., 1996 Parent et al., 1998 Raj et al., 2003 Gamble, 1994 Dumas et al., 2000 Goldberg et al., 2001 Wogan, 2000 Fryzek et al., 1997 Weiderpass et al., 1998 Ojajarvi et al., 2000 Ojajarvi et al., 2001 Lowenfels and Maisonneuve, 2002 Koufman and Burke, 1997 Maier and Tisch, 2000 Steenland et al., 1996 Alberg and Samet, 2003 Fuchs and Pritchard, 2002 Rockley et al., 1994 Lutz et al., 1999 Desmond and Soong, 2003 Hodgson and Darnton, 2000 McDonald and McDonald, 1996 Wong, 2001 Britton, 2002 Zahm and Fraumeni, 1997 DeBruin and Josephy, 2002 Kheifets and Matkin, 1999 Wolff and Weston, 1997 Labreche and Goldberg, 1997 Wolff et al., 1996 Brody and Rudel, 2003 Weiderpass et al., 2001 Shen et al., 1998. Parent and Siemiatycki, 2001 Pollan et al., 2001 McLaughlin, 1993 McLaughlin and Blot, 1997 McLaughlin and Lipworth, 2000 Siemiatycki et al., 1994 Negri and La Vecchia, 2001 Kogevinas, et al., 2003 Carozza et al., 2000 Wesseling et al., 2002 Navas-Acien et al., 2002 De Roos et al., 2003 Bezabeh et al., 1996 Feychting, 1996 Hotz and Lauwerys, 1997 Kheifets et al., 1997 Miller et al., 1997 Savitz and Andrews, 1997 Hardell and Axelson, 1998 Zeeb and Blettner, 1998 McCunney, 1999 Baker et al., 1999 Baris and Zahm, 2000 Hayes et al., 2001 Becker et al., 2001 Weisenburger and Chiu, 2002

substances Acrylonitrile Aromatic amines Arsenic

Collins and Acquavella, 1998 Leonard et al., 1999 Vineis and Pirastu, 1997 Snyderwine et al., 2002 Hayes, 1997 Schwartz, 1997

(continued)

336

PART III: THE CAUSES OF CANCER

Table 18–9. (cont.) Topic Asbestos

Benzene Coal-petroleum products

Diesel emissions Dyes

Electromagnetic fields

Ethylene oxide Formaldehyde Ionizing radiation

Metals and metal compounds

Man-made mineral fibers Mustard gas Organochlorines

Polycyclic aromatic hydrocarbons (PAHs) Pesticides

Table 18–9. (cont.) Review

Topic

Jakobsson et al., 1994 Gamble, 1994 Dement and Brown, 1994 Landrigan et al., 1999 Browne and Gee, 2000 Osinubi et al., 2000 Hodgson and Darnton, 2000 Wong, 2001 Becker et al., 2001 Weisenburger and Chiu, 2002 Bergsagel et al., 1999 Krewski et al., 2000 O’Connor et al., 1999 Raabe, 1993 Enterline, 1993 Infante, 1993 Partanen and Boffetta, 1994 Wong and Raabe, 1995 Mastrangelo et al., 1996 Tolbert, 1997 Boffetta et al., 1997 Ahmed, 2001 Mehlman, 1996 Morgan et al., 1997 Stober et al., 1998 Boffetta and Silverman, 2001 Skov and Lynge, 1994 La Vecchia and Tavani, 1995 Correa et al., 2000a Correa et al., 2000b La Vecchia and Tavani, 2001 Yu et al., 2002 La Vecchia and Tavani, 2002 Feychting, 1996 Miller et al., 1997 Caplan et al., 2000 Kheifets, 2001 Ahlbom et al., 2001 Thier and Bolt, 2000 McLaughlin, 1994 Liteplo and Meek, 2002 Committee on the Biological Effects of Ionizing Radiation. 1990 Boice and Lubin, 1997 Miller, 1997 Ron, 1998 Samet and Eradze, 2000 Schubauer-Berigan and Wenzl, 2001 MacMahon, 1994 Costa, 1997 Eisenbud et al., 1997 Hayes, 1997 Benke et al., 1998 Oller, 2002 Seilkop and Oller, 2003 Verougstraete et al., 2003 Steenland and Stayner, 1997 Osinubi et al., 2000 Berrigan, 2002 Blair and Kazerouni, 1997 Karalliedde et al., 2000 Dacre and Goldman, 1996 Longnecker et al., 1997 Hoffmann, 1996 Snedeker, 2001 Calle et al., 2002 Mastrangelo et al., 1996 Boffetta et al., 1997 Bohnen and Kurland, 1995 Doe and Paddle, 1994 Zahm et al., 1997 Dich et al., 1997 Sathiakumar and Delzell, 1997 Georgellis et al., 1999

Silica

Solar radiation Solvents

Vinyl chloride Wood

Review Steenland and Stayner, 1997 Weill and McDonald, 1996 Hadfield, 1998 Soutar et al., 2000 Finkelstein, 2000 Tsuda et al., 2001 English et al., 1997 Wang et al., 2001 Labreche and Goldberg, 1997 Lynge et al., 1997 McLaughlin and Blot, 1997 Wartenberg et al., 2000 Bruning and Bolt, 2000 Cohen et al., 2002 McLaughlin and Lipworth, 1999 Boffetta et al., 2003 Nylander and Dement, 1993 Demers et al., 1995 Demers et al., 1995 Blot et al., 1997 Toren et al., 1996

occupations and industries Aluminum production Farming

Firefighters Rubber industry Welders

general reviews

Benke et al., 1998 Zahm and Blair, 1993 Van der Gulden and Vogelzang, 1996 Blair and Zahm, 1995 Acquavella et al., 2003 Guidotti, 1995 Golden et al., 1995 Roth, 1999 Acquavella and Leonard, 2001 Moulin, 1997 Ward et al., 1997 Blair and Kazerouni, 1997 Frumkin and Thun, 2000 Sullivan and Krieger, 2001

*The articles and book chapters listed here contain good recent reviews of topics that are not explicitly referenced in this chapter, but that pertain to some aspects of occupational cancer. For some topics there may be an IARC Monograph on the topic as well, and these would have been cited in Tables 18–3 to 18–6. In such cases, the articles listed here may complement or update the corresponding IARC publication. The topics covered here include not only occupations and occupational substances, but also sites of cancer from an occupational perspective. Most of the articles are review articles, but some are metaanalyses or single study reports with good literature reviews. The material is organized first by site of cancer, then by occupational substances, and then by occupation/industry.

Diesel and Gasoline Engine Emissions Engine emissions are common in many workplaces and are ubiquitous environmental pollutants. Some experimental evidence and some epidemiologic evidence suggests that emissions from diesel-powered engines may be lung carcinogens, but the epidemiologic evidence is inconclusive (Nauss et al., 1995; Katsouyanni and Pershagen, 1997; Boffetta et al., 1997). The difficulty of drawing inferences regarding the effect of diesel exhaust is in part due to some methodological limitations and in part due to the indirect nature of the evidence. Namely, most of the studies have used certain job titles (most often, truck driver) as proxies for occupational exposure to diesel exhaust. Few studies were able to control for the potential confounding effect of cigarette smoking, and of other occupational exposures. Many of the studies had low statistical power and/or insufficient follow-up time. Finally, the relative risk estimates in most studies have ranged from 1.0 to 1.5, making it difficult to exclude the possibility of chance or bias. The number of diesel-powered vehicles is increasing in many countries. Because of the significant scientific and public policy implications (Weeks, 1998; Silverman, 1998), it is important to derive more definitive inferences regarding the potential human carcinogenicity of diesel emissions. Although it is not definitive, there is less evidence, both experimental and epidemiologic, for a carcino-

Occupation genic effect of exposure to gasoline engine emission than to diesel emission. Engine emission provides an example of a common dilemma in occupational and environmental cancer risk assessment. A chemical analysis of both gasoline and diesel exhaust shows the presence of many substances that are considered carcinogenic, notably some nitroPAHs that are classed by IARC as 2A and 2B. Should the presence of a carcinogen within a complex mixture automatically trigger a labeling of the mixture as carcinogenic, irrespective of the epidemiologic evidence on the mixture? There is no wide consensus on this issue, but it has important consequences. For instance, it would mean not only that diesel exhaust should be considered a probable human carcinogen, even apart from any epidemiologic evidence, but also that gasoline engine emission, which is only considered to be a possible human carcinogen, would automatically be labeled as a probable human carcinogen.

Asbestos Few health issues have sparked as much public concern, controversy, and expense as has asbestos-related cancer risk. Asbestos is a term describing a family of naturally occurring fibrous silicates that have varied chemical and physical compositions, and that have been widely used in industrial and consumer products for more than a century. The main fibre types are called chrysotile and amphibole. Exposure to asbestos fibers occurs in many occupations, including mining and milling, manufacture of asbestos-containing products, and use of these products. Currently, construction and maintenance workers constitute the largest group of asbestos-exposed workers, resulting from application and removal of asbestos products, and building demolition. Asbestos has been one of the most ubiquitous workplace exposures in the twentieth century. Case reports linking asbestos with lung cancer started to appear in the 1930s and 1940s (Lynch and Smith, 1935), but the first formal investigations were published in the 1950s and 1960s (Doll, 1955; Selikoff, 1990). In the early 1960s, reports appeared linking asbestos exposure to a hitherto unrecognized tumor of the pleura and peritoneum called mesothelioma (Wagner et al., 1960). By the mid-1960s, it was clear that the very high and virtually uncontrolled exposure conditions prevalent up to then could induce lung cancer and mesothelioma. While asbestos production and use has declined dramatically in most industrialized countries since 1975, public concern and controversy have not (Doll and Peto, 1985; Nicholson, 1986; Stone, 1991; Upton et al., 1991; IPCS (International Programme on Chemical Safety), 1998; Collegium Ramazzini, 1999; Siemiatycki, 2001). Asbestos fibers are highly persistent and widespread in the environment. Measurements carried out in all kinds of nonoccupational settings have detected asbestos fibers, and it has become clear that asbestos is a widespread environmental pollutant, albeit at much lower levels than in the workplace. Also, because of long latency periods, we are still seeing the cancer impact of high occupational exposure levels experienced 20 to 40 years ago, and we will for some time to come. Finally, asbestos use has been increasing in developing countries. Several countries have banned use of asbestos, while others have instituted regulatory limits orders of magnitude below levels that had been known to produce harmful effects. The availability of alternative non-asbestos substitution products makes such strategies feasible. Despite the vast scientific literature on the health effects of asbestos, there are unresolved scientific questions that have important implications in setting public policy regarding continued use of asbestos, remediation, and compensation for asbestos-related disease. Apart from mesothelioma and lung cancer, it is uncertain whether there are other target organs for asbestos carcinogenicity. For both mesothelioma and lung cancer, there has been considerable variation in estimates of cancer risk from different studies, and the search for explanations of this great variability has centered on such issues as the industry in which asbestos is used, the dimensions of the fibers, and especially on the type of fibre. Since exposure levels are much lower

337

than they used to be, another unresolved issue is the nature of the risk due to low levels of asbestos exposure. Risk assessment models have been developed to extrapolate from high to low exposure levels, but these models have not been validated (Camus et al., 1998).

Cadmium and Cadmium Compounds Cadmium has been produced and used in alloys and various compounds for several end products including batteries, pigments, electroplating, and some plastics (IARC, 1993a). Exposure varies widely between industries in both types of cadmium compounds and level of exposure. Following reports in a few small cohorts of excess cases of prostate cancer among workers in battery plants, an early IARC Working Group concluded that there was moderately persuasive evidence of an excess risk of prostate cancer as a result of cadmium exposure (IARC, 1973b; IARC, 1976). They noted in passing that one of the cohorts also reported an excess of lung cancer. In the following decade a number of additional cohort studies were undertaken in cadmium-exposed workers (IARC, 1993a). There was no additional evidence of an increase in prostate cancer risk. However, the evidence on lung cancer, which was unremarkable in the first few studies, became much more pronounced as additional data were accumulated. By 1993, another IARC Working Group pronounced cadmium a group 1 carcinogen, but solely on the basis of its association with lung cancer. Still, the assessment of carcinogenicity of cadmium highlighted several methodological problems. The number of long-term, highly exposed workers was small, the historical data on exposure to cadmium was limited, and the ability to define and examine a gradient of exposure was limited to one study. Confounding by cigarette smoking in relation to lung cancer was difficult to address. Control of the confounding effect of co-exposure to other metals, particularly arsenic and nickel, was limited.

Styrene Styrene is one of the most important industrial chemicals. The major uses are in plastics, latex paints and coatings, synthetic rubbers, polyesters, and styrene-alkyd coatings (Collins and Richey, 1992). These products are used in construction, packaging, boats, automotive (tires and body parts) and household goods (e.g., carpet backing). Nearly 18 million tons were used worldwide in 1998. It has been estimated that as many as one million workers in the United States may be exposed to styrene, and the numbers worldwide would be much greater. In addition, there is widespread low-level environmental exposure. The first evidence of a possible cancer risk came from case reports of leukemia and lymphoma among workers in various styrene-related industries (Block, 1976; Lemen and Young, 1976; Nicholson et al., 1978). A number of cohort studies have been carried out since then in Europe and the United States in various industries (Bond et al., 1992; Wong et al., 1994; Kogevinas et al., 1994; Kolstad et al., 1995; Delzell et al., 2001). The interpretation of these studies has been bedevilled by four main problems: (1) the different types of industries in which these studies were carried out make it difficult to compare results across studies, (2) within most industries, styrene is only one of several chemical exposures and these tend to be highly correlated with styrene exposure, (3) the pattern of results has been unpersuasive, though there are a couple of hints of excess risk of leukemia in some subgroups of some cohorts, and finally (4) the classification of hematopoietic malignancies is complicated (IARC, 2002c). The substantial body of epidemiologic evidence can reasonably be interpreted as showing no cancer risk, or it can be interpreted as showing suggestions of risk of leukemia in some subgroups of some cohorts. The IARC Working Group leaned in the latter direction as they categorized the human evidence as “limited” rather than “inadequate.” The studies already conducted have been large and there have been several of them. It is not clear that another study would resolve the issue. Nor does the experimental evidence provide clear guidance. The animal experimental evidence is equivocal and human biomarker studies show some signs of DNA adduct formation.

338

PART III: THE CAUSES OF CANCER

1,3-Butadiene Concern about the possible carcinogenicity of 1,3-butadiene in humans derives from the results of animal experiments, which showed an increased incidence of leukemia in mice and, to a lesser extent, rats (IARC, 1999a). Data on the carcinogenicity of butadiene in humans derive essentially from studies conducted among workers employed in the production of the monomer and in the production of styrenebutadiene rubber (SBR), where high exposure levels occurred in the past. In the most informative study, Delzell et al. (1996), Macaluso et al. (1996), and Sathiakumar et al. (1998) evaluated the mortality experience of a cohort including 15,649 workers at eight SBR plants comprising most SBR-exposed workers in North America. Among the different causes of death, there was suggestive evidence of an excess risk for leukemia, particularly when including deaths in which leukemia was listed as a contributory cause. Depending on the subgroup analyzed and the exposure period included, the SMRs varied from near 1.3 to 2.2, with lower 95% confidence limits ranging from near 1.0 to 1.5. Retrospective, quantitative estimates of exposure to 1,3-butadiene, styrene, and benzene were developed. The relative risk of leukemia increased with cumulative butadiene exposure. In a further study of a cohort of 2795 American workers employed in the manufacture of 1,3-butadiene, the SMR for lymphohaematopoietic neoplasms was 1.5 (95% CI: 1.1–2.0), but that for leukaemia was 1.1 (95% CI: 0.6–1.9) (Divine and Hartman, 1996). There was no exposure-response trend. The results of additional studies in either the butadiene production industry (Cowles et al., 1994; Ward et al., 1996) or polymer production, including SBR (McMichael et al., 1976; Bond et al., 1992), have been limited by very small numbers. In conclusion, the most informative cohort study of workers exposed to butadiene showed an increased risk of leukemia, but these results have not been clearly replicated. It will be difficult to find exposed populations in which to try to replicate these findings.

Vinyl Chloride Vinyl chloride (VC) is a large volume industrial chemical with many practical applications. In the early 1970s, clinicians observed a cluster of cases of angiosarcoma of the liver among a group of workers in a plant using VC (Creech and Johnson, 1974). The tumor is so rare that they were struck by the cluster. Within a very short time, other similar clusters were reported and the association was quickly accepted as causal (Tabershaw and Gaffey, 1974; IARC, 1974a). The discovery was facilitated by the rarity of the tumor, the strength of the association, and the fact that there are no other known risk factors for this tumor, and thus little danger of confounding. Early cohort studies confirmed the strong effect of vinyl chloride on risk of angiosarcoma of the liver, and also raised questions about a possible association with lung cancer. In fact, the data were suggestive enough in the 1980s that an effect on lung cancer was considered likely (IARC, 1987a; Doll, 1988). However, subsequent studies have failed to demonstrate such an effect, and it is likely that the early reports were distorted by confounding or chance (Boffetta et al., 2003). While there is growing evidence that the lung is not a target organ, it is becoming more plausible, as a result of recent meta-analyses (Boffetta et al., 2003), that exposure to VC may cause hepatocellular carcinoma as well as liver angiosarcoma. Detecting an association of moderate strength with a fairly rare tumor that has a long latency is difficult, and it will take more data to confirm it. A further complication is whether some of the hepatocellular carcinomas are in fact misdiagnosed angiosarcomas. An additional source of potential bias and confusion derives from the observation, in the two multi-center cohort studies (Mundt et al., 2000; Ward et al., 2001) that diagnostic misclassification may occur between liver angiosarcoma and soft-tissue sarcomas and, given the rarity of soft-tissue sarcomas, this could artificially create the appearance of an association with soft-tissue sarcomas. Because of the drastic decrease in exposure levels that took place in the vinyl chloride industry after the discovery of its carcinogenic

activity, it is unlikely that there will be new cohorts of highly exposed workers to investigate. It is conceivable that new data can be generated from further follow-up of existing cohorts; however, the maximum latent period for most cancers is likely to be approaching, and additional cancers are increasingly likely to reflect background and other risk factors other than vinyl chloride. Molecular epidemiology provides another avenue for exploring the carcinogenic effects of VC. A mutation in the p53 gene, which is relatively specific for vinyl chloride exposure, has been identified (Marion and Boivin-Angele, 1999; Barbin, 2000). There are sparse data on analysis of p53 mutation in other tumors of exposed workers; results on hepatocellular carcinoma, however, are consistent with a carcinogenic effect (Weihrauch et al., 2000). There is a need for larger studies of p53 mutation in tumors other than liver angiosarcoma that arose in vinyl chlorideexposed workers.

Radium and Radon Radium and radon provide an interesting contrast from the point of view of prevention strategies. Both radium and radon gas induce tumors in exposed workers through ionizing radiation. Radium was used by dial painters and caused osteosarcomas. Radon gas caused lung cancer in miners. The risk due to radium was easily eliminated by, in effect, eliminating the occupation of radium dial painting. Mining cannot be eliminated, and radon gas is an inevitable exposure in mines. The best strategy here is to find a cost-effective way to reduce exposures by engineering methods, while also improving the epidemiologic database on dose-response relationships. Radon also provides one of the most successful examples of the use of high-dose occupational data for the purpose of extrapolation to lower-dose environmental exposure levels (NAS (National Academy of Sciences), 1999).

How Evidence Has Been Accumulated on Selected Associations Table 18–10, based on a similar table in Monson (1996), shows the evolution of evidence regarding ten recognized occupational risk factors. For each association, the table indicates when the first suspicions were published and some of the significant pieces of evidence that came into play subsequently. The table also gives some synthetic information about the nature of the epidemiologic findings. Typically, the association was first suspected on the basis of a clinical observation, which was followed up by suggestive, but inconclusive cohort studies and then by more rigorous and more persuasive cohort studies. For most recognized carcinogens, the interval between the first clinical report and the general acceptance of the association was measured in decades. The length of the interval was great in the early period, in part because of the lack of expertise in epidemiologic research and resources to conduct such studies. For three more recent “discoveries,” those relating asbestos to mesothelioma, vinyl chloride to angiosarcoma of the liver, and chloroethers to lung cancer, the interval between the first publication of a suspicious cluster, and the general acceptance of a causal association was only a matter of a few years. As a rule, early reports tended to manifest higher relative risk estimates than more recent reports. This is likely due to several reasons, including: the greater likelihood that outlier results will get noticed and reported, and real improvements in the industrial hygiene conditions that have indeed had the effect of decreasing risks of cancer. While it is instructive to study the history of the evolution of knowledge for recognized carcinogens, it is just as useful to understand that the trajectories of suspicion and recognition are not necessarily monotonic. That is, there are also examples of associations that have been considered possible or likely in the past, that are now considered as unlikely. We have already alluded to the example of cadmium exposure and prostate cancer risk. Early studies hinted at an association (Potts, 1965; Kipling and Waterhouse, 1967; Lemen et al., 1976; Sorahan and Waterhouse, 1983), but more recent and stronger studies have tended to refute the hypothesis (Thun et al., 1985; Kazantzis

339

Occupation

Table 18–10. Selected Milestone Publications Illustrating Development of Information in Humans on Selected Well-Established Occupational Cancers Material/ Cancer Radon/lung

Benzidine/ bladder

Nickel and nickel compounds/nasal Arsenic/respiratory

Asbestos/lung

Benzene/leukemia

Chloroethers/lung Vinyl chloride/liverangiosarcoma

Reference

Location

Study Population

Study Type

Evidence of Effect

Härting and Hesse, 1879 Peller, 1939 Archer et al., 1962 Archer et al., 1976 Howe et al., 1987 Rehn, 1895 Scott, 1952 Case et al., 1954 Meigs et al., 1986 Annual Report, 1933 Doll, 1958 Kaldor et al., 1986 Henry, 1934 Hill and Faning, 1948 Lee and Fraumeni, 1969 Lee-Feldstein, 1986 Pinto et al., 1978 Enterline et al., 1987b Lynch and Smith, 1935 Doll, 1955 Selikoff et al., 1964 McDonald et al., 1980 Dement et al., 1983 Seidman et al., 1986 Mallory et al., 1939 Vigliana and Saita, 1964 Ishimaru et al., 1971 Aksoy et al., 1974 Infante et al., 1977 Rinsky et al., 1987 Yin et al., 1987 Figueroa et al., 1973 DeFonso and Kelton, 1976 McCallum et al., 1983 Creech and Johnson, 1974 Monson et al., 1974 Waxweiler et al., 1976 Fox and Collier, 1977

Germany Czechoslov. USA USA Canada Germany England Great Britain Connecticut Wales Wales Wales England England Montana Montana Washington Washington South Carolina England USA Canada USA USA UK Italy Japan Turkey Ohio Ohio China Philadelphia Philadelphia United Kingdom Kentucky Kentucky USA Great Britain

Miners Miners Uranium miners Uranium miners Uranium miners Dye workers Dye workers Dye workers Benzidine makers Nickel refineries Nickel refineries Nickel refineries Sheep-dip makers Arsenical packers Smelter workers Smelter workers Smelter workers (urine index) Smelter workers (air index) Asbestos textile workers Asbestos workers Insulation workers Chrysotile miners Asbestos textile workers Amosite workers Various occupations Various occupations Various occupations Shoemakers Pliofilm makers Pliofilm makers Benzene producers Chemical workers Chemical workers Chloroether makers PVC makers PVC makers PVC makers PVC makers

Case series Cohort Cohort Cohort Cohort Case series Case series PMR Cohort Case series PMR Cohort Case series PMR Cohort Cohort Cohort Cohort Single case Cohort Cohort Cohort Cohort Cohort Case series Case series Case series Case series Cohort Cohort Cohort Case series Cohort Cohort Case series PMR Cohort Cohort

Moderate Moderate Strong Strong Strong Weak Moderate Strong Strong Moderate Strong Strong Weak Moderate Strong Strong Strong Strong Weak Weak Moderate Strong Strong Strong Weak Weak Moderate Moderate Moderate Strong Strong Moderate Moderate Strong Weak Strong Strong Moderate

et al., 1992). For the possible association between man-made mineral fibers (MMMF) and lung cancer, the impetus and suspicion came from the similarity in physical characteristics between MMMF and asbestos. However, large American and European cohort studies have failed to demonstrate an excess risk (Boffetta et al., 1997; Marsh et al., 2001; Kjaerheim et al., 2002). Still, the absolute exposure levels to MMMF have been so much lower than they have been to asbestos, that it may justly be asked whether the differential evidence of lung carcinogenicity between asbestos and MMMF is likely due to exposure levels rather than to inherent carcinogenic properties of the two classes of fibers. A third example is that of ethylene oxide and leukemia. There were reports from Sweden among producers and some users of ethylene oxide that hinted at excess risks of leukemia (Hogstedt et al., 1979; Hogstedt et al., 1986). However, larger American studies have subsequently shown no such risk (Stayner et al., 1993; Teta et al., 1993). A fourth example is that concerning acrylonitrile and lung cancer. Some American and British studies published in the early 1980s indicated possible excess risks (O’Berg, 1980; Werner and Carter, 1981; Delzell and Monson, 1982). But a series of large studies from Europe and the United States subsequently failed to demonstrate any risk of lung cancer. Finally, suspicions have been voiced for a long time about the possible association between formaldehyde and lung cancer. But a series of large studies have failed to demonstrate such an effect (Acheson et al., 1984; Blair et al., 1986; Bertazzi et al., 1989; Andjelkovich et al., 1995). It is certainly clear that reports of case clusters, or suspicions based on experimental findings or individual epidemiologic studies are not sufficient to predict the ultimate judgment regarding an association. Since random chance and error, supplemented by publication bias, will

inevitably lead to the publication of some false positive results, it is important to seek replication of findings.

PERCENTAGE OF CANCER ATTRIBUTABLE TO OCCUPATION In the late 1970s and early 1980s, there was a heated debate, fueled as much by sociopolitical considerations as by scientific ones, as to the relative importance of occupational and environmental exposures in producing cancer. Several investigators tried to address the following question: what proportion of cancer in the population is attributable to occupational exposures (referred to as the population attributable risk percent or etiologic fraction)? This question had great social importance; the answer would affect both the disbursement of cancer research funds and the priority to be given to cleaning up the workplace. Very divergent estimates were made by various authors, ranging from less than 1% to about 40% (Higginson and Muir, 1976; Cole, 1977; Wynder and Gori, 1977; Bridbord et al., 1978; Higginson, 1980; Peto and Schneiderman, 1981; Hirayama, 1981; Milham, 1981; Doll and Peto, 1981; Vineis and Simonato, 1991). In part, the variability in estimates was a result of the fact that the population attributable risk percent varies according to the prevalence of and degree of exposure to the risk factor in the population and the relative risk induced by exposure to the factor. The greater the prevalence and the greater the risk due to exposure, the greater will be the resulting number of cases. Since prevalence of exposure to occupational carcinogens differs from place to place and from time to time, it is inevitable that the

340

PART III: THE CAUSES OF CANCER

proportions of cancer attributable to occupational carcinogens will also vary. The second major reason for the wide variability of published estimates was the variety of the methods and assumptions used to derive the estimates. The methods used to estimate the proportions of cancer attributable to occupational carcinogens have for the most part been very crude and even subjective. The best ones have been based on lists of recognized carcinogens such as those shown in Tables 18–3 to 18–6; but as stated above, such a list may represent the tip of the iceberg of occupational carcinogens. Not only is our knowledge deficient in listing carcinogens, but for those already discovered, there is very little reliable quantitative information available on the prevalence of exposure to these substances, and indeed on the “exposureresponse” relationships that they induce. Perhaps the most widely cited estimates of cancer risk attributable to occupation were those produced by Doll and Peto (1981). In a sweeping survey of etiologic factors in cancer, they estimated attributable risks for a host of types of factors, including occupation, diet, air pollution, geophysical factors, and others. They estimated that in the late 1970s, in the United States, 4% of all cancer deaths were attributable to occupational exposures. Among males the proportion was about 7% and among females about 1%. A majority of cancers attributable to occupation (about two-thirds) were lung cancers. Other types with sizable attributable numbers were leukemia and bladder cancer. Doll and Peto admitted that their estimates were imprecise, but guessed that they were within a twofold range of the true values. This seems quite narrow in light of the limitations of the data they used. Despite the passage of time and the publication of thousands of research results that can bear on this, there have been few comprehensive attempts to update or improve on these estimates. Perhaps the most valid attempt to date to estimate the proportion of cancer attributable to occupation was that carried out by Nurminen and Karjalainen (2001). The target population for the estimate was Finland. They assembled three types of data: a list of occupationcancer associations, an estimate of the numbers of Finns exposed to each of the occupations or occupational exposures, and an estimate of the degree of risk (e.g., relative risk) associated with the level of exposure. As far as possible, they used local (i.e., Finnish) data, but where local data were unavailable, they extended the information base to northern Europe and then to North America and the rest of the world. Their judgement on which associations to include were liberal; that is, they did not limit themselves to associations that were group 1 or even group 2A under the IARC classification. Apart from the relative exhaustiveness of their procedures, the main virtue of these estimates was the fact that Finland has one of the best systems in the world for estimating the numbers of workers exposed to different substances (Kauppinen et al., 1998). For all sites of cancer and all carcinogens combined, Nurminen and Karjulainen estimated that 14% of cancers among Finnish males and 2% among Finnish females (8% combining the sexes) were attributable to occupational exposures. This concords with an estimate for Finland derived by other means (Dreyer et al., 1997). The largest numbers of attributable cases were generated by the following associations: asbestos-lung cancer; radon gas-lung cancer; and environmental tobacco smoke-lung cancer. For the United States the most convincing recent estimate is that produced by Steenland et al. (2003). They estimated that 2% to 5% of cancer deaths in the United States were attributable to occupational exposures. Estimates in these ranges, while smaller than the risk attributable to cigarette smoking, nevertheless translate to a significant burden of preventable disease and death. This burden is not randomly distributed in society, but rather, it preferentially affects blue collar workers (Nicholson, 1984).

METHODOLOGICAL CONSIDERATIONS—STUDY DESIGN, EXPOSURE ASSESSMENT, CONFOUNDING, AND INTERACTIONS The main stages in occupational cancer epidemiology are detection/discovery of hazards, which can be broken down into hypothesis-

generation and hypothesis-testing, and characterization of risks. This categorization is simplistic. In reality, a given piece of research may serve two or three of these stages, and the operational distinctions among them are ambiguous. However, it is a useful conceptual framework. Before the 1950s, the generation of hypotheses relied primarily on astute clinicians to notice clusters of cancer among groups of workers, and the investigation of hypotheses was carried out by means of industry-based historical cohort studies. Thereafter, new approaches were introduced, including attempts to generate hypotheses from analyses of routine record sources (such as death certificates) and from casecontrol studies. For testing hypotheses and characterization of hazards, there was increasing use of case-control methods. The various approaches that are used in occupational cancer epidemiology can be divided in two major families: community-based studies and industrybased studies. The following sections describe some of the salient features of these designs and their advantages and disadvantages in this area.

Industry-Based Studies In an industry-based study, the population under investigation is defined on the basis of belonging to a union or working for a company, or some other work-related institution. Because of the long latency of cancer, the study design typically used is a historical cohort design (Checkoway et al., 2004). A given workforce is generally exposed to a relatively narrow range of occupational substances, and for this reason the prime role of cohort studies has been and remains to investigate specific associations (or to “test hypotheses” or characterize relationships), rather than to generate hypotheses. But this is an oversimplification; a typical cohort study produces results on possible associations between one or more exposures and many types of cancer. Since it is often difficult or costly in practice to constitute an appropriate group of unexposed subjects with whom to compare the exposed and since the cohort usually constitutes a very small fraction of the entire population, it is expedient and often acceptable to take the disease or death rates in the entire population (national or regional) as a close approximation of those in the unexposed. The latter are easily available from published statistics or databases. When the disease experience of the exposed cohort is compared with that of the entire population, it is possible to take into account such basic demographic variables as age, sex, and race. The most common statistical approach is indirect standardization and the resulting parameter is called a standardized mortality ratio (SMR) or standardized incidence ratio (SIR). There are two significant advantages of the cohort approach, both relating to exposures of workers. First is the opportunity it affords to focus on a group of workers with relatively high exposure levels, thereby improving the chances of detecting a risk. Secondly, by focusing on a single industry or company, it is sometimes possible to derive detailed and valid data on the exposure histories of study subjects.

Exposure Assessment in Cohort Studies It is common for companies to maintain job history records for each worker, and these are often maintained for decades. Depending on the nature of the industry, the company, and the relationship established between the investigator and the company, it may be possible to obtain detailed historic exposure measurements and these might be linkable to the job histories of individual workers. It may also be possible to consult company hygienists or engineers or other workers who can inform the investigator about past conditions and exposure circumstances. The cooperation of employers is usually a sine qua non to conduct such studies. It is sometimes possible to obtain quite high-quality historic exposure information and to use this in assessing and characterizing hazards (Rappaport and Smith, 1991; Armstrong et al., 1994b; Checkoway et al., 2004). Notable examples include studies on formaldehyde (Blair et al., 1986; Blair and Stewart, 1990), asphalt workers (Burstyn et al., 2003), acrylonitrile (Swaen et al., 1998; Stewart et al., 1998), and nickel compounds (Grimsrud et al., 2002). In some historic

Occupation examples, such as in certain cohorts of asbestos workers, there were no available quantitative data on exposure levels, but the industrial process was thought to be so “simple” that only one substance was thought to be worth considering as an explanation for the excess risk of the entire cohort (Selikoff et al., 1973). Such reasoning may be acceptable in a few industries, such as the extractive industries, but most industrial processes entail diverse mixtures of exposures. The success at characterizing past exposures will depend on the skill and resources of the investigating team and the availability of adequate industrial hygiene data. Ingenious methods have been brought to bear by industrial hygienists working with epidemiologists to evaluate historic exposures to specific substances in various cohorts (Smith et al., 1993). The mechanism for abstracting exposure data from company records and assigning exposure retrospectively to study subjects may vary, but a common tactic is to develop an industry-specific job exposure matrix (JEM). A JEM is simply an automatic set of indicators showing which exposures may occur in which occupations (Hoar et al., 1980; Siemiatycki, 1996). A JEM is defined by a job axis, a substance axis, and the entries indicating whether a worker in a given job is considered exposed to a given substance. The entries in the JEM can be as simple as a binary yes/no indicator, or of various quantitative or semi-quantitative indices showing, for instance, the probability that workers are exposed, the concentration of exposure, the percentage of time that workers are exposed, and other parameters. In an industry-based cohort study, a JEM would ideally involve quite a detailed job axis, with assessments by company, by work site, by department, by task, and by era. The quality of exposure data in an industry-based cohort study is partly determined by whether or not the company has had a good tradition of industrial hygiene, including the nature and quality of past measurements. Even if long-standing industrial hygiene measurements are available, there is particular concern about the representativity of those measurements. Since measurements were often taken for compliance purposes, the times and locations of environmental sampling may have been unrepresentative. A novel approach to exposure assessment in cohort studies is occasioned by the increasing practice of monitoring exposure levels among workers by means of biological markers of exposure. In industries where this is done and where sampling results are maintained for long periods, it can be used to construct a historical cohort. This approach has been used in assessing the risk of cancer in workers exposed to solvents such as tricholorethylene (Anttila et al., 1995).

Other Considerations in Cohort Studies Sample size is often a problem in historical cohort studies because it may take many thousands of person-years to generate a number of cancers that provide stable estimates of risk. However, the sample size is usually beyond the investigator’s control; that is, the number of workers who worked in the plant is fixed. Sometimes, multiple similar plants can be identified and enrolled and pooled to enhance the statistical power. One of the main drawbacks of using a cohort with a national reference population is the inability to ensure that the exposed and unexposed groups are comparable, apart from age, sex, and race. In fact, it is likely that a cohort consisting of current or ex-employees of an industry would differ in some respects from the entire national population, which includes both active and inactive persons. Some issues relating to confounding will be discussed below. If individuals within the cohort can be categorized as to exposure to certain substances, or by occupation or by work site, it is possible to carry out “internal analyses.” Such analyses have the triple virtues of minimizing any threat of confounding (Siemiatycki et al., 1988), providing presumptive evidence on dose-response patterns, and focusing attention on subsets where risk may be localized. Further modifications to the basic cohort design are the nested casecontrol design and the case-cohort design (Rothman and Greenland, 1998; Checkoway et al., 2004). For both of these, subjects are all selected within the cohort, and additional detailed information can be collected on occupational exposures and/or potential confounders. These designs represent an efficient approach when it is not feasible

341

to collect data on exposure variables or key confounders for the entire study sample, but it would be for a subset. It is very difficult to envisage carrying out cohort studies of workers at small companies, or in industries that are not unionized. This is particularly unfortunate because these may be the workplaces that entail the greatest hazards. Similarly, industry-based studies have rarely been conducted in developing countries, because of difficulties in accessing company records, lack of well-organized unions, and problems in conducting mortality or cancer incidence follow-up in most developing countries.

Community-Based Studies In a community-based study, the population is typically defined on the basis of living in a given geographic area or falling in the catchment area of a set of health-care providers. The two principal study designs in community-based studies are routine record studies and questionnaire-based case-control studies. In some countries, occupations are recorded on death certificates, on cancer registers, or on databases (e.g., census data) that can be linked to cancer data. Such routine record data sets can be configured to derive cancer risks in different occupations. Routine record studies can be relatively inexpensive and often entail very large numbers of subjects. Questionnaire-based casecontrol studies provide the opportunity to collect information on lifetime occupation histories and on other relevant cofactors directly from cancer patients or close relatives, and appropriate controls. From this, it is possible to estimate cancer risks in relation to various occupational circumstances.

Job Titles Both routine-record studies and case-control studies provide the opportunity to conduct analyses based on job titles. A limiting factor of such analyses is the validity of the job title information collected. With an interviewer-administered questionnaire, quite valid job histories can be obtained (Baumgarten et al., 1983; Bond et al., 1985; Bourbonnais et al., 1988). The picture is less optimistic with routine records. The validity of occupations recorded on routine records such as death certificates or tumor registers is typically mediocre (Wigle et al., 1982; Steenland and Beaumont, 1984; Armstrong et al., 1994b). In addition, routine records typically contain only one of the subject’s jobs, usually the most recent, which may include “retired”. In a large sample of Montreal workers, it was estimated that, on average, about 62% of working years were spent in the job of longest duration, and about 50% in the last job (Siemiatycki, 1996). Some of the defects of approaches based on routine records can be attenuated by linking records of more valid sources of occupational data with mortality or cancer registry records. Examples are in Canada (Howe and Lindsay, 1983), where a government-run labor force survey was linked to mortality records, and especially in the Nordic countries where census data have been linked to the cancer registry (Lynge and Thygesen, 1990; Andersen et al., 1999). Analyses using job titles are useful. Several associations with cancer have been discovered by means of analyses on job titles. Such analyses are most valid and valuable when the workers have a relatively homogeneous exposure profile. Examples might include miners, motor vehicle drivers, butchers, and cabinetmakers. Whatever attempts are made to derive specific exposures in community-based studies, it is nevertheless worthwhile to also conduct the statistical analyses to evaluate risks by job titles. However, job titles are limited as descriptors of occupational exposures (Siemiatycki et al., 1981). On the one hand, many job titles cover workers with very diverse exposure profiles. On the other hand, many exposures are found to occur across many occupation categories. In such circumstances, epidemiologic analyses by job title may entail too much noise to allow for a signal to be detected. Several approaches have been used to ascertain exposures in community-based studies.

Self-Reported Checklist of Exposures In a questionnaire-based study, subjects can be asked what they were exposed to. Attempts to assess the validity of self-reports have tended

342

PART III: THE CAUSES OF CANCER

to show poor validity (Joffe, 1992; Kromhout et al., 1992; Fritschi et al., 1996). Under most circumstances, self-reported exposure is probably an inadequate and ill-advised tool for collecting exposure information in community-based studies.

Job Exposure Matrix Whether the job titles of study subjects are obtained via routine records or in an interview, it is possible to derive a list of exposures by means of a job exposure matrix (JEM). Creating a JEM for use in community-based studies entails much greater challenges than in an industrybased study. The job axis can consist of a crude or a fine occupation classification. Within that can be embedded one or more of the following classification variables: industry, generic task indicators, calendar time. From the crudest to the most detailed job axis, the number of categories can increase by orders of magnitude, perhaps from a few hundred to hundreds of thousands. The substance axis can be short (only a few substances) or very long. It can be composed of very general classes (e.g., dusts, solvents) or of highly specific ones (e.g., nickel subsulfide, trichloroethylene). The entries in the JEM can be as simple as a binary yes/no indicator, or of various quantitative or semiquantitative indices showing, for instance, the probability that workers are exposed, the concentration of exposure, the percentage of time that workers are exposed, and other parameters. The creation of a JEM requires either expert judgment or an adequate database or both. While the basic idea is seductively simple, the creation of a general purpose JEM with large numbers of occupations and large numbers of substances is so difficult that there have not been more than a handful of attempts to create community-based JEMs (Bouyer and Hemon, 1993). The JEMs that have been created have been based on crude rather than fine job axes, or they have involved a narrow range of substances. The resources devoted to JEM development have varied, but few had a team of experts working over a protracted period to make the thousands of decisions necessary for a JEM. Each was created in a particular time and place, and, even if it was valid for its own time and place, it is not clear that a JEM can be used elsewhere. Notable examples of JEMs that have been the result of considerable time investments on the parts of their authors are that of NIOSH (Sieber et al., 1991) in the United States, Berrino et al. (2003) for southern Europe, and that of Kauppinen et al. (1998) for Finland. There is little empirical evidence on the validity of JEMs, but the evidence that has been produced indicates rather mediocre validity (Linet et al., 1987; Siemiatycki et al., 1989; Dewar and Siemiatycki, 1991; Kromhout et al., 1992; Stucker et al., 1993; Luce et al., 1993). In any case, none has been validated as an instrument that can be used internationally. This somewhat pessimistic view of JEMs pertains to community-based JEMs, which must cover a very broad industrial/occupational spectrum. Whereas the development of a JEM is fraught with difficulties in the context of a communitybased study that spans all occupations and industries, it is much more feasible to envisage such a project within a single industry or company.

Expert Assessment Retrospective exposure assessment can be carried out by appropriately trained experts, based on information that can be obtained from study subjects. In interview studies, it is possible to ask subjects to provide detailed descriptions of their jobs. Experts in industrial hygiene can use these descriptions as well as documentary and other information to infer a list of potential exposures to which the subject was exposed in this job. There are many aspects that will influence the quality of the information elicited (Gérin and Siemiatycki, 1991), such as: number of experts, their experience and their training in exposure assessment, the quality and vividness of job descriptions obtained from subjects, amount of time allowed for coding exposures, the number of substances on the experts’ checklist, and availability of local resources such as industrial hygiene information (Siemiatycki et al., 1991). There is very little information available on the validity of expert assessment of exposures (Siemiatycki, 1996). The main drawback of this approach is its great expense, and the paucity of exposure assessment experts to do this work.

Biomonitoring If biologic specimens can be obtained from study subjects (cases and controls), these could conceivably be analyzed for markers of past exposure to different occupational substances (Schulte and Perera, 1993; Vineis et al., 1999). While the promise of using biomarkers of past exposure is a tantalizing one, at present its realization is limited by several obstacles. First, it may not be feasible to obtain biologic specimens from study subjects, particularly population controls. Noninvasive techniques, such as buccal swabs for extraction of DNA, are being developed to overcome this limitation, but are still not available for the full range of biomarkers of exposure. Second, there are not many known, validated biomarkers of exposure. Third, there are even fewer for which the half life is in the order of years or decades, which would be necessary for a study of occupational carcinogens. Finally, in case-control studies, the assessment of biomarkers may be distorted by the sequella of or treatment for cancer. The increasing availability of large-scale biobanks, in which biological samples, usually serum, urine, or DNA are stored prospectively for long periods, offers a potentially powerful tool to overcome some of the limitations of retrospective use of biomarkers. This approach has been used, for example, in studying DNA adducts formed by PAHs and subsequent risk of lung cancer (Perera et al., 1989).

Confounding Our discussion of various designs has necessarily involved allusion to different potential sources of error and bias, including confounding. In this section, we briefly review some forms of confounding that are of particular concern in occupational cancer studies.

Non-Occupational Confounders The threat of confounding must be taken seriously in any epidemiologic research, but the plausibility is certainly high in a cohort study with an external referent. Factors like social class, ethnicity, place of residence, physical fitness, physical activity, diet, smoking habits, and others may well differ between a given cohort or a given occupation group and the entire population of the country or state. The choice of an appropriate referent population (national, state, local) would have to be made in recognition of the likelihood of bias in relation to such factors, as well as the availability of statistically stable rates. One such bias is known as the “healthy worker effect”, and alludes to the fact that an industrial workforce may well be selected on characteristics that favor good health (Arrighi and Hertz-Picciotto, 1994). This probably pertains more to chronic diseases other than cancer such as cardiovascular or neurological disease. While it is difficult to ensure strict comparability, it should not be considered axiomatic that the possibility of confounding is synonymous with the presence of significant confounding. For instance, one of the most common concerns in this regard is that, given the very strong association between tobacco smoking and lung cancer risk, lung cancer risks may be distorted in studies that do not have access to data on smoking among study subjects. However, a few studies have shown that the presence or absence of smoking data does not, in most circumstances, make a huge difference to the relative risk estimates derived for associations between lung cancer and various occupation groups (Axelson, 1980; Blair et al., 1988; Siemiatycki et al., 1994a). In these empirical studies, the confounding bias rarely exceeded 20%. These considerations apply to any studies (routine-record, cohort, or case-control) in which one occupational group is compared with others. In case-control studies, there is greater opportunity to control for these various types of confounding bias. It depends on the cleverness of the design and on local circumstances whether the possibility is realized. As a general rule in occupational cancer studies, the greatest efforts on the investigator’s part should be on ascertaining disease status and exposure status. These are the “first-order” variables that must be measured with as little error as possible. Non-occupational confounding variables deserve attention, but their possible impact on relative risk estimates is indirect and usually of less consequence than error in measuring the two primary variables.

Occupation

Occupational Confounders The most difficult form of bias to contend with is the form that occurs among occupational substances, if there is an attempt to analyze risks in relation to substances. This difficulty arises because: a most occupational environments where there is chemical exposure involve many substances; b there may be quite high correlations in exposure patterns among different substances; c there is considerable error in exposure assessment; and; d this error may be correlated across different substances. These conditions create problems in interpreting both univariate and multivariate analyses. It has two kinds of consequences. First, when an analysis seems to find an association between a particular substance and cancer risk, it is usually appropriate to be circumspect about concluding that this substance, and not another one that has not been measured, or that has been measured with too much error, is the responsible agent. The second consequence is that when multiple substances are analyzed in models that mutually adjust the effects of different substances, it is possible that a true association will be missed because the effects of the true carcinogen will be partially “siphoned off ” by the inclusion of other variables with which it is correlated. Namely, the fact that the true carcinogen is measured with error will lead to residual confounding in estimates of other exposures, and the fact that measurement error may be correlated as well, will exacerbate the problem. This difficulty has no obvious solution, except to maximize the quality of the exposure assessment. Analyses of one occupation exposure variable at a time should be part of the strategy of analysis, and where there are different results between these one-at-a-time analyses and the multiple substance model analyses, there may be evidence that there is a carcinogen in the picture, but it is hard to pinpoint. An analogous confounding effect may occur, although to a much lesser degree, in studies based on job title analyses, because “dirty jobs” tend to cluster in occupational histories.

Joint Effects Between Smoking and Lung Carcinogens As discussed above, workers exposed to a given carcinogen are inevitably also exposed to many other substances and circumstances, at work or outside work; and some of these may modify the effect of the carcinogen in question. Such phenomena are sometimes referred to as interactions or as effect modification. The nature of joint effects can be addressed between any two risk factors, such as age and benzene exposure for leukemia, or benzene and ionizing radiation for leukemia. But the issue has been addressed most notably regarding the joint effects of smoking and asbestos on lung cancer risk (Saracci, 1987; Liddell and Armstrong, 2002). In an early cohort of asbestos workers, Selikoff et al. (1968) found that the workers exposed to both smoking and asbetos incurred a relative risk (53.6) that was approximately equal to the multiple of the relative risks due to smoking alone (10.9) and asbestos alone (5.2.). Subsequent empirical evidence on smoking and asbestos in lung cancer etiology has not corroborated the strong multiplicative relationship. There have now been several reviews and meta-analyses of the joint effects of smoking and asbestos (Erren et al., 1999; Lee, 2001; Liddell, 2001; IARC, in preparation). While individual studies provide estimates that range from additive effects (Liddell and Armstrong, 2002) to multiplicative effects, the consensus seems to lie in between. In most studies, the validity of inferences has been compromised by low statistical power to detect interactions, and by questionable data on smoking habits of cohort members. In part because few studies had smoking data, and in part because the statistical power to detect interactions is much less than the power to detect main effects, there have been few attempts to document the joint effects between smoking and other occupational lung carcinogens. Armstrong et al. (1994a) found that the joint effect of smoking and coal tar pitch volatiles in an aluminum refinery on lung cancer risk was compatible with both additive and multiplicative models. Results on the interaction between tobacco smoking and radon exposure are not consistent, though the joint effect seems to be greater than additive (IARC, in preparation). In the case of arsenic and silica, the avail-

343

able evidence also suggests an effect that is more than additive but less than multiplicative (Saracci and Boffetta, 1994). In any case, the presence or absence of interaction on one scale or another is not necessarily an indication of biological interaction between the two factors (Siemiatycki and Thomas, 1981; Greenland, 1993). Statistical interaction and biological interaction are two separate phenomena. Statistical interaction is an arbitrary construct that is rooted in the statistical models used to analyse data. However, it has been argued that from a public health point of view, it is appropriate to use the additive model of joint effects as a reference point, and that departures from additivity are noteworthy (Rothman, 1974).

MOLECULAR EPIDEMIOLOGY AND GENEENVIRONMENT INTERACTIONS As in other areas of epidemiology, the revolution in molecular biology and genetics is creating opportunities and challenges in occupational cancer research. In general terms, molecular methods provide opportunities in three areas: exposure assessment, markers of early carcinogenic effects (along the continuum from pre-cancerous to cancerous), and studies of individual susceptibility and geneenvironment interaction.

Exposure Assessment Biomarkers of exposure are used to improve precision and validity of exposure assessment in occupational epidemiology. Two classes of exposure biomarkers have been developed: markers of the agent itself and markers of interaction between the agent and macromolecules. These are respectively referred to as markers of internal exposure (or internal dose) and markers of biologically effective dose. The use of internal exposure markers to measure the agent of interest, or its metabolites, in the target organ or in blood or urine, offers several advantages: (1) it integrates different sources of exposure, (2) it takes into account individual characteristics related to uptake, partition, metabolism, and excretion, and (3) it provides a quantification of exposure. Markers of biologically effective dose target the products of the interaction between the agent and macromolecules, mainly adducts formed with DNA or proteins. The detection of adducts can provide evidence of a role of the agent of interest in the carcinogenic process, as it has in the case of ethylene oxide (IARC, 1994b). Among the obstacles to their use are possible measurement errors due to DNA repair activity and to technical difficulties in the assays. However, such exposure markers have important limitations: (1) they often reflect recent exposure, depending on the half-life of the chemicals under study, (2) non-occupational sources of exposure might confound the measured level of the marker, (3) interpretation of results might be difficult when the metabolism of the agent of interest is not fully characterized. Nonetheless, markers of internal dose have been successfully used for some agents, including trichloroethylene and tetrachloroethylene (Anttila et al., 1995).

Markers of Early Carcinogenic Effect Early markers of neoplasia are potentially useful endpoints for epidemiological studies since they can reduce the time interval between exposure and detectable effect and they can also increase the prevalence of occurrence of the outcome. But for the marker to be of use in this way, it would have to have high predictive value for cancer occurrence. There is a continuum of possible effects from mutations and cytogenetic abnormalities to pre-clinical neoplastic manifestations. Generally speaking, the demonstrated predictive validity of putative markers of early effects is greater for effects that are later in the natural progression. Unfortunately, at present, none of the target organs that are prime sites for occupational carcinogens (e.g., lung, bladder) have tests with high predictive value (such as cytological abnormalities in cervical smears used to predict cervical cancer in the Papanicolau test).

344

PART III: THE CAUSES OF CANCER

Among the earlier stage effects that might be useful in studies of occupational cancer, two stand out: p53 gene mutation and chromosomal alterations. Among several genes mutated in human cancer, p53 is of special interest because it is mutated in many cancers and it presents different patterns of mutations that may reflect the action of specific carcinogens. While the greatest amount of data available on agent-specific mutations in p53 concerns exposure to tobacco smoke, distinct patterns of mutations have been reported in lung cancer among workers exposed to radon (Taylor et al., 1994; Hollstein et al., 1997), as well as in liver angiosarcoma of workers exposed to vinyl chloride (Marion and Boivin-Angele, 1999). Chromosomal alterations, such as translocation, deletion, and amplification, measured in cultured peripheral lymphocytes, have long been suspected of being early markers of cancer, and have even been used as diagnostic tools to monitor workers exposed to known mutagens or carcinogens. Follow-up studies of workers have confirmed that such chromosomal aberrations indeed represent independent predictors of cancer risk, after the effects of smoking and occupational exposure have been taken into account (Hagmar et al., 1998; Bonassi et al., 2000).

Markers of Susceptiblity and Gene-Environment Interaction Susceptibility to the action of carcinogens varies among individuals, based on genetic background. With respect to occupational cancer, the most relevant aspect of individual susceptibility is represented by polymorphisms in low-penetrance genes involved in different aspects of carcinogenesis, including activation or detoxification of occupational agents, DNA repair, cell cycle control, and apoptosis. Geneenvironment interaction describes the situation where the effect of these genes depends on the presence or absence of an environmental exposure, and vice versa. Gene-environment interaction studies in occupational cancer involve the collection of data on both genetic factors and occupational factors, with a view to assessing risks among subjects who share a common genetic trait and occupational exposure (Vineis et al., 1999; Brennan, 2002). In theory, such studies can provide new information on occupational risk factors, genetic risk factors, and their interactions. It is possible that some occupational carcinogens are poorly detectable using conventional methods because they only affect a subset of the population, and it is plausible that the power to detect such an effect would be enhanced by focusing attention on the susceptible subgroup. A gene-occupation interaction study could provide the opportunity for focusing attention within the stratum of susceptible individuals, and thereby enhance the likelihood of finding carcinogens. In practice, this has not yet turned out to be a common or productive paradigm. Most research on gene-environment interactions has been anchored in known carcinogens, with the aim of determining whether certain genetic polymorphisms confer an increased susceptibility to the carcinogen in question. A prime example is the work on risk of bladder cancer among subjects with the N-acetyltransferase slow polymorphism (Vineis et al., 2001). This relationship has been studied among subjects with and without occupational exposure to selected occupational carcinogens, mainly aromatic amines. For the most part, it appears that subjects with the slow acetylation genotype are at excess risk of disease, particularly among subjects exposed to occupational carcinogens. The value of this approach is, however, questionable from a public health point of view. That is, even if we could earmark part of the population as being susceptible to the action of an occupational carcinogen, how would that information be used? It is much more attractive and effective to conceive of public health action geared to the elimination of occupational hazards than to the selective targeting of genetically susceptible individuals. It seems that we will still have to identify carcinogens using traditional methods.

CANCER PREVENTION Over the past 50 years, it is likely that the number of occupationally induced cancers has decreased in western countries. This is the result

of different trends. The decline in blue collar heavy industry and the corresponding growth of the white collar industry has decreased numbers of workers in particularly “dirty occupations”. At the same time, many industries have instituted procedures and processes that provide much cleaner work sites than in the past (Symanski et al., 1998). The motivations are complex and multi-dimensional. In part, this is a byproduct of epidemiologic research carried out in the past (Zenz, 1988; Levy and Wegman, 1988; Monson, 1990; Rom, 1992). The identification and characterization of occupational carcinogens triggers regulatory actions in many countries, and these are intended to reduce the permissible exposure levels. Such actions may range from substitution of one substance in an industrial process for another, modification of the industrial procedures or ventilation/emission control procedures, or the use of protective equipment by workers. But the real benefits of such regulations may be quite nonspecific. That is, while regulations concerning a particular carcinogen may reduce the risk of cancer in relation to that carcinogen, cleaning up an industrial process induces reduction in exposure to many substances, some of which may be in the hidden part of the iceberg of occupational carcinogens. It is impossible to estimate how many cancer cases have been prevented as a result of the cumulative efforts of occupational epidemiologists, but it is certainly a partial success story. In addition, there has been a growing realization on the part of many industries, that good industrial hygiene makes good business sense. The cautionary tales of companies that have suffered from regulatory or legal opprobrium, as well as compensation costs, as a result of being identified as a “cancer-causing company” has served as an incentive to others to clean up. Setting standards for regulatory purposes is a difficult task that relies on epidemiologic, toxicologic, and other data (Holmberg and Winell, 1977; Corn, 1983; ACGIH, 2001). Historically, these standards were usually based on considerations of acute toxicity; but increasingly, cancer has become a key endpoint. The main problem is the lack of reliable epidemiologic data on exposure-response relationships. For the most part, the regulators must rely on animal data, with complex mathematical models used to translate the animal experience into terms that are relevant for human risk assessment.

Some Examples of Prevention Strategies One approach, which can only be implemented if a known carcinogen has not yet been introduced into industrial practice, is to ban its introduction. This approach is usually only possible if the agent has been used and shown to be carcinogenic in one country and that information is then used in another country. For example, following reports from the United States on the increased bladder cancer risk among workers exposed to 4-aminobiphenyl, its introduction was banned in the United Kingdom (Swerdlow, 1990). Substitution of products known to be carcinogenic has been used successfully, as in the example of asbestos and man-made mineral fibers. Attempts to reduce exposure levels to known or suspect carcinogens have been more common. Successful examples include: (1) the virtual avoidance of radiation-related cancer risk among nuclear industry workers in the United States and western Europe, based on knowledge of the effects of radioactivity (Boice et al., 1996), (2) the significant decrease in angiosarcoma risk among workers in the vinyl chloride industry following recognition of its carcinogenicity in the 1970s (Boffetta et al., 2003), (3) the significant decrease in lung cancer risk among American workers exposed to chloromethyl ethers following recognition of its carcinogenicity in the 1950s (Maher and DeFonso, 1987), and (4) the significant decrease in nasal cancer risk among Norwegian nickel refinery workers following recognition of cancer risks in the 1930s (Magnus et al., 1982). The decrease in exposure to occupational carcinogens might be due to reduced emission, improved ventilation, or use of personal protection of the workers. As a general rule, the first two approaches are more efficient than the use of protection equipment in achieving a durable reduction in exposure. Reduction of emission can be easily achieved for chemicals produced under controlled conditions, such as

Occupation intermediates formed during chemical manufacturing processes. However, reduction of exposure at the source might be difficult to achieve for substances that are used under less controlled conditions, such as motor exhausts. Screening occupationally exposed workers has been proposed as an additional measure to prevent cancer death. However, there is no evidence of efficacy for any of the cancers for which it has been proposed. This is the case inter alia for lung cancer and mesothelioma among asbestos-exposed workers (using chest X-rays or cytological examination of sputum) and bladder cancer among workers exposed to aromatic amines (using cytological or mutagenicity analysis of urothelial cells in the urine). There has been significant improvement in occupational hygiene conditions in large industries in developed countries (Symanski et al., 1998). The challenge is to extend this improvement to smaller enterprises and to developing countries, where there remain significant problems of exposure to such agents as asbestos, crystalline silica, and pesticides (Pearce et al., 1994).

USE OF EPIDEMIOLOGICAL DATA IN WORKERS COMPENSATION Many countries have schemes in place to compensate workers who have developed cancer as a result of workplace exposures. Such schemes were set up primarily for the compensation of workers experiencing occupational injuries or diseases specific to the occupation, such as pneumoconiosis. For these cases, the cause of the injury or illness is rarely a matter of dispute. Increasingly, cancers, which have no objective features specific to the occupational cause, have been the object of compensation claims. For these diseases, evidence for an occupational cause rests on epidemiologic studies. It is not possible to identify with certainty which specific workers have had cancer caused by work exposures. For such diseases, different jurisdictions adopt different approaches. Compensation is not usually considered unless a causal association between work and a cancer has been well established (e.g., the association being recognized by IARC as group 1 or 2A). Unfortunately, the fact that the IARC monographs do not formally mention target organs tends to complicate the acceptance of potential causality. Other sources are used, which can be controversial. One concept that has recently attracted interest is the probability of causation (PC) (Armstrong and Theriault, 1996). The PC can be thought of as the chance that a specific case of cancer was caused by the exposure in question. Under certain assumptions it can be defined from the risks of cancer in people exposed (Re) and unexposed (Ru) as the excess risk divided by the total risk in exposed people: PC = (Re - Ru)/Re or PC = (RR - 1)/RR, where RR = Re/Ru. Note that at a relative risk of two, the PC is 50%. If a worker’s exposure status is known and if the RR of cancer at that exposure level is known from independent epidemiologic studies, it is possible to compute the PC. Further, it is possible to compute the PCs due to different factors, such as an occupational exposure and smoking, under different scenarios of interaction between the two factors. How the PC is used for the purpose of determining compensation depends on local social and legal considerations. The above formulation of the issue has been challenged by Greenland and Robins (1988) and by Greenland (1999). They argue that this formulation only considers cases of disease that would not have occurred at all in the absence of exposure, and ignores cases whose onset was accelerated by exposure. They argue that if causation is understood to include both the production of disease where it would not have otherwise occurred and the acceleration of the etiologic process, this approach tends to underestimate the true probability of causation.

345

STRUCTURAL CHALLENGES AND OBSTACLES TO CONDUCT EPIDEMIOLOGICAL RESEARCH ON OCCUPATIONAL CANCER Perceived Need for Research In the 1960s and 1970s, the field of occupational cancer research was one of the most thriving areas of epidemiological research. This was fed by the social trends that raised the profile of environmentalism and workers’ health and by important discoveries of occupational carcinogens such as asbestos. There was a perception that research on environmental causes of cancer was important and that it would be feasible to make breakthroughs. Workers’ organizations were active and vocal in calling for improved working conditions, and for the research that would support such action. Many young investigators, influenced by the zeitgeist of the 1960s, were ideologically drawn to a research area that would dovetail with their political and social interests. In contrast, at the beginning of the new millenium, we perceive a waning of interest and enthusiasm. What has happened? The reasons are complex, but may well include the following. The political/social climate that supported work on occupational health has greatly changed. In western countries, the economies and workforces have shifted and there are fewer blue collar industrial workers than there were 30 years ago. Union membership, especially in blue collar unions, has declined, and the unions have become less militant. These trends have been fostered by technology (e.g., computerization and robotization) and by globalization. To a certain extent, “dirty jobs” have been eliminated or exported from western to developing countries. The bottom line is that a smaller fraction of the western workforce is involved in traditional “dirty jobs.” Another factor is that, as mentioned above, most large workplaces have become much cleaner, at least in some industrialized countries. Another reason for the deflation of interest in this area, is that some people’s expectations for quick and dramatic discoveries of “smoking guns” like asbestos did not pan out. The expectations were unrealistic, but that was not clear at the time. There was a widespread belief that there were many cancer-causing hazards in the workplace and it would only be a matter of shining some light in the right places to find them. There was much more epidemiological research in the 1970s, 1980s, and 1990s than there had been in the preceding decades. While this research produced a large number of important findings, these were incremental in the overall scheme of things, and, for some, did not seem proportional to the effort. In the face of these social and economic changes, and the ostensible diminishing returns from research in occupational cancer, is this an area of investigation that should be fostered? Our answer is an unambiguous “Yes!” for the following reasons and with the following caveats. 1. In industrialized countries a large fraction of the workforce still works in circumstances that bring it into contact with chemical agents. Even if the fraction is less than it was a century ago, it is still sizeable, and will remain so for the foreseeable future. While industrial design and hygiene have succeeded in lowering exposures in many industries, there remain pockets where exposure levels remain high. 2. The story of occupational hygiene conditions in developing countries is less rosy. Enormous numbers of people are now working in insalubrious conditions. As life expectancy in these populations rises with increasing affluence and improved living conditions and medical care, the numbers of cancer cases and most likely the numbers of occupationally related cancers are steadily increasing. There is a tremendous opportunity for epidemiologists to investigate occupation-cancer relationships in developing countries. 3. There are many thousands of chemicals in workplaces. Many of them are obscure and involve relatively few workers; but many involve exposure for thousands of workers. Of these, only a small fraction have been adequately investigated with epidemiological data. 4. The industrial environment is constantly evolving with the introduction of new and untested chemicals. We must maintain a

346

5. 6.

7.

8.

PART III: THE CAUSES OF CANCER

monitoring capacity to detect “new” occupational carcinogens. A recent example of a suspected carcinogen is indium phosphide in the semiconductor industry (IARC, in preparation). The occupational environment is one that lends itself to preventive intervention. Many chemicals in the workplace find their way into the general environment, either via industrial effluent or via their use in consumer products. Hazards identified in the workplace often have an importance that goes beyond the factory walls. The discovery of occupational carcinogens is important to understanding the principles of carcinogenesis: workers represent a ‘natural experiment’ of high exposure to a potentially carcinogenic agent. The ability to detect hazards is increasing with improvement of methods for exposure assessment and outcome assessment, as well as the tendency to use larger study sizes.

research is in part the result of collaboration by industry. While this is not a simple issue, consideration should be given to requiring industry to maintain data and to make it accessible to bona fide researchers. This would of course have to be done in respect of the legitimate economic and proprietary concerns of industry.

PRIORITIES This chapter covers an enormous matrix of occupational exposures by types of cancer. It is very difficult to enumerate priorities for substantive research. But there are some general principles that can guide priority setting. The principles that we offer would have different significance in different areas of the world and thus the choice of priorities would naturally differ. The following factors must be considered:

• Suspicion: If there is some evidence of carcinogenicity, but it is Manpower to Conduct Occupational Cancer Research Let us imagine again the hypothetical matrix mentioned above with hundreds of occupational exposures on one axis and scores of different types of cancer on the other, resulting in thousands of cells of potential associations. While it is possible and important to establish priorities among these cells for research, it remains that there are indeed many of these cells that deserve attention on the part of epidemiology, and indeed, for evaluating consistency, it is important to have multiple pieces of evidence regarding those cells. The relatively small cadre of occupational cancer epidemiologists in the world cannot conduct all the studies needed. One part of the problem is the bald fact that there are not enough researchers in this area. The solution is not to attract epidemiologists who work in other areas. All areas of epidemiological research are under-staffed. There has been increasing recognition of the contribution that epidemiology can make to different areas of health and medicine. The total manpower of epidemiologists has not kept pace with the needs and the opportunities. Leaders in the field and users of epidemiological results must exert influence to drastically increase the numbers of epidemiologists so that there will be enough to satisfy the needs in specific areas such as occupational cancer.

Human Subjects Research—Data Access, Privacy, Ethics Another development that affects all areas of epidemiology, but occupational cancer research in particular, is the trend towards increasing obstacles to data access as a result of concerns about privacy and confidentiality. These developments have taken different forms and have occurred at different times in different countries. But the trends in most countries of North America and Europe are unmistakable. Without entering into a detailed discussion of these issues here, we only wish to draw attention to the looming crisis that would be upon us if the most draconian laws envisaged by privacy zealots are enacted. It may become so onerous to conduct historical cohort studies, or record linkage studies or case-control studies, that for all intents and purposes, such research would be prohibited. While epidemiologists are not necessarily expert in influencing public policy, we cannot afford the luxury of sitting by and letting the legislative political process or the “privacy and ethics lobby” deal with the issue without taking the social needs of epidemiological research into account. Epidemiologists must get involved, make alliances, and make sure that everybody knows that the price of such laws and regulations, if applied to research, will be the inability to identify causes of disease and therefore the inability to prevent future disease.

Access to Industry Data We have mentioned above that collaboration from the industry is an essential component of success in this field, and the extraordinary contribution of studies from Nordic countries to occupational cancer

• • • •

• •

• •

• •

•

inconclusive (e.g., IARC group 2A or group 2B), this should be prioritized. Prevalence of the exposure: More prevalent exposures (i.e., more workers exposed) should have higher priority. Relevance to non-occupational exposure: Exposures that are also found in non-occupational settings or consumer products should have higher priority. Degree of exposure: Exposures that are poorly controlled should have higher priority. Agents not previously studied: While it is often useful to conduct research locally on agents that have already been shown to be carcinogenic elsewhere, there is some danger of neglecting to investigate topics that are less strongly suggested by the literature. This may be particularly relevant for newly industrialized countries. Need for risk characterization: If there is a particular need to characterize dose-response relationships or effect modification, this should be prioritized. Feasibility: Feasibility of conducting research depends on several factors, such as: availability of cohorts, cooperation of companies and/or unions, availability of funding. Absence of these would be contra-indicative. Exposure assessment: Research should only be supported if it entails adequate exposure assessment. This often requires participation of trained experts (e.g., industrial hygienists, chemists). Sample size and statistical power: The publication of studies entailing little statistical power to detect hazards may do more harm than good. Research should only be supported if it entails sufficient sample sizes. Community-based case-control studies on occupational cancer must be much larger than they have traditionally been, with 500 as a minimum number of cases and controls in most instances. Historical cohort studies must be large enough to generate scores of cancer deaths or cases. Multicentric and multi-national trials should be encouraged: This leads to expanding expertise. Opportunity for training: Given the lack of expertise in occupational cancer epidemiology in many countries, priority should be given to those projects entailing a strong training component for collaborators from such countries. Opportunity to elucidate mechanisms of carcinogenesis: There may, in specific instances, be particular opportunities to elucidate mechanisms that can lead to a broader understanding of cancer etiology.

References ACGIH. 2001. 2001 TLVs and BEIs. Cincinnati: American Conference of Governmental Industrial Hygienists. Acheson E, Barnes H, Gardner M, Osmond C, Pannett B, Taylor C. 1984. Formaldehyde in the British chemical industry: An occupational cohort study. Lancet 1:611–616. Acquavella J, Doe J, Tomenson J, Chester G, Cowell J, Bloemen L. 2003. Epidemiologic studies of occupational pesticide exposure and cancer: Regulatory risk assessments and biologic plausibility. Ann Epidemiol 13:1–7. Acquavella JF, Leonard RC. 2001. A review of the epidemiology of 1,3butadiene and chloroprene. Chem Biol Interact 135:43–52.

Occupation Ahlbom IC, Cardis E, Green A, Linet M, Savitz D, Swerdlow A, ICNIRP (International Commission for Non-Ionizing Radiation Protection) Standing Committee on Epidemiology. 2001. Review of the epidemiologic literature on EMF and Health [review] [219 refs]. Environ Health Perspect 109:911–933. Ahmed F. 2001. Toxicology and human health effects following exposure to oxygenated or reformulated gasoline [review] [130 refs]. Toxicol Lett 123:89–113. Aksoy M, Erdem S, DinCol G. 1974. Leukemia in shoe-workers exposed chronically to benzene. Blood 44:837–841. Alberg AJ, Samet JM. 2003. Epidemiology of lung cancer. Chest 123:21S–49S. Allen BC, Crump KS, Shipp AM. 1988. Correlation between carcinogenic potency of chemicals in animals and humans. Risk Anal 8:531–550. Andersen A, Barlow L, Engeland A, Kjaerheim K, Lynge E, Pukkala E. 1999. Work-related cancer in the Nordic countries. Scand J Work Environ Health 25:1–114. Andjelkovich D, Janszen D, Brown M, Richardson R, Miller F. 1995. Mortality of iron foundry workers: 4. Analysis of a subcohort exposed to formaldehyde. J Occup Environ Med 37:826–837. Anttila A, Pukkala E, Sallmen M, Hernberg S, Hemminki K. 1995. Cancer incidence among Finnish workers exposed to halogenated hydrocarbons. J Occup Environ Med 37:797–806. Archer V, Gillam J, Wagoner J. 1976. Respiratory disease mortality among uranium miners. Ann N Y Acad Sci 271:280–293. Archer V, Magnuson J, Holaday Deal. 1962. Hazards to health in uranium mining and milling. J Occup Med 4:55–60. Armstrong B, Theriault G. 1996. Compensating lung cancer patients occupationally exposed to coal tar pitch volatiles. Occup Environ Med 53:160–167. Armstrong B, Tremblay C, Baris D, Theriault G. 1994a. Lung cancer mortality and polynuclear aromatic hydrocarbons—A case-cohort study of aluminum production workers in Arvida, Quebec, Canada. Am J Epidemiol 139:250–262. Armstrong BK, White E, Saracci R. 1994b. Principles of Exposure Measurement in Epidemiology. Oxford: Oxford University Press. Arrighi HM, Hertz-Picciotto I. 1994. The evolving concept of the healthy worker survivor effect. Epidemiology 5:189–196. Ashby J. 1996. Alternatives to the 2-species bioassay for the identification of potential human carcinogens. Hum Exp Toxicol 15:183–202. Ashby J, Tennant R. 1988. Chemical structure, Salmonella mutagenicity and extent of carcinogenicity as indicators of genotoxic carcinogenesis among 222 chemicals tested in rodents by US NCI/NTP (MYR 01277). Mutat Res 204:17–115. Axelson O. 1980. Aspects of confounding and effect modification in the assessment of occupational cancer risk. J Toxicol Environ Health 6:1127–1131. Baker P, Inskip H, Coggon D. 1999. Lymphatic and hematopoietic cancer in teachers. Scand J Work Environ Health 25:5–17. Barbin A. 2000. Etheno-adduct-forming chemicals: from mutagenicity testing to tumor mutation spectra. Mutat Res 462:55–69. Baris D, Zahm SH. 2000. Epidemiology of lymphomas. Curr Opin Oncol 12:383–394. Baumgarten M, Siemiatycki J, Gibbs GW. 1983. Validity of work histories obtained by interview for epidemiologic purposes. Am J Epidemiol 118:583–591. Becker N, Berger J, Bolm-Audorff U. 2001. Asbestos exposure and malignant lymphomas—A review of the epidemiological literature. Int Arch Occup Environ Health 74:459–469. Bell J. 1876. Paraffin epithelioma of the scrotum. Edinburgh Med J 22:135–137. Benke G, Abramson M, Sim M. 1998. Exposures in the alumina and primary aluminium industry—An historical review. Ann Occup Hyg 42:173–189. Berenblum I. 1979. Carcinogenicity testing for control of environmental tumor development in man. Isr J Med Sci 15:473–479. Bergsagel DE, Wong O, Bergsagel PL, Alexanian R, Anderson K, Kyle RA, Raabe GK. 1999. Benzene and multiple myeloma: Appraisal of the scientific evidence. Blood 94:1174–1182. Berrigan D. 2002. Respiratory cancer and exposure to man-made vitreous fibers: A systematic review. Am J Ind Med 42:354–362. Berrino F, Richiardi L, Boffetta P, Esteve J, Belletti I, Raymond L, Troschel L, Pisani P, Zubiri L, Ascunce N, Guberan E, Tuyns A, Terracini B, Merletti F. 2003. Occupation and larynx and hypopharynx cancer: A job-exposure matrix approach in an international case-control study in France, Italy, Spain and Switzerland. Cancer Causes Control 14:213–223. Bertazzi PA, Pesatori A, Guercilena S, Consonni D, Zocchetti C. 1989. [Carcinogenic risk for resin producers exposed to formaldehyde: Extension of follow-up]. Med Lav 80:111–122. Bezabeh S, Engel A, Morris C, Lamm S. 1996. Does benzene cause multiple myeloma—An analysis of the published case-control literature. Environ Health Perspect 104:1393–1398.

347

Blair A, Kazerouni N. 1997. Reactive chemicals and cancer [review]. Cancer Causes Control 8:473–490. Blair A, Steenland K, Shy C, O’Berg M, Halperin W, Thomas T. 1988. Control of smoking in occupational epidemiologic studies: Methods and needs. Am J Ind Med 13:3–4. Blair A, Stewart PA. 1990. Correlation between different measures of occupational exposure to formaldehyde. Am J Epidemiol 131:510–516. Blair A, Zahm S. 1995. Agricultural exposures and cancer. Environ Health Perspect 103:205–208. Blair A, Stewart P, O’Berg M, Gaffey W, Walrath J, Ward J, Bales R, Kaplan SCD. 1986. Mortality among industrial workers exposed to formaldehyde. J Natl Cancer Inst 76:1071–1084. Block J. 1976. A Kentucky study: 1950–1975. In: Ede L, ed. Proceeding of NIOSH Styrene-Butadiene Rubber Briefing, Covington, Kentucky, April 30, 1976. Cincinnati: National Institute for Occupational Safety and Health, pp. 28–32. Blot W, Chow W, Mclaughlin J. 1997. Wood dust and nasal cancer risk—A review of the evidence from North America [review]. J Occup Environ Med 39:148–156. Boffetta P, Jourenkova N, Gustavsson P. 1997. Cancer risk from occupational and environmental exposure to polycyclic aromatic hydrocarbons [review]. Cancer Causes Control 8:444–472. Boffetta P, Matisane L, Mundt K, Dell L. 2003. Meta-analysis of studies of occupational exposure to vinyl chloride in relation to cancer mortality. Scand J Work Environ Health 29:220–229. Boffetta P, Saracci R, Andersen A, Bertazzi PA, Changclaude J, Cherrie J, Ferro G, Frentzelbeyme R, Hansen J, Olsen J, Plato N, Teppo L, Westerholm P, Winter PD, Zocchetti C. 1997. Cancer mortality among man-made vitreous fiber production workers. Epidemiology 8:259–268. Boffetta P, Silverman DT. 2001. A meta-analysis of bladder cancer and diesel exhaust exposure. Epidemiology 12:125–130. Bohnen N, Kurland L. 1995. Brain tumor and exposure to pesticides in humans: A review of the epidemiologic data. J Neurol Sci 132:110–121. Boice J, Land C, Preston D. 1996. Ionizing radiation. In: Schottenfeld D, Fraumeni JF, eds. Cancer Epidemiology and Prevention. New York: Oxford University Press, 319–354. Boice J, Lubin J. 1997. Occupational and environmental radiation and cancer. Cancer Causes Control 8:309–322. Bonassi S, Hagmar L, Stromberg U, Montagud AH, Tinnerberg H, Forni A, Heikkila P, Wanders S, Wilhardt P, Hansteen IL, Knudsen LE, Norppa H. 2000. Chromosomal aberrations in lymphocytes predict human cancer independently of exposure to carcinogens. European Study Group on Cytogenetic Biomarkers and Health. Cancer Res 60:1619–1625. Bond G, Sobel W, Shellenberger R, Flores G. 1985. Validation of work histories obtained from interviews. Am J Epidemiol 122:536–537. Bond G, Bodner K, Olsen G, Cook R. 1992. Mortality among workers engaged in the development or manufacture of styrene-based products—An update. Scand J Work Environ Health 18:145–154. Bourbonnais R, Meyer F, Thériault G. 1988. Validity of self reported work history. Br J Ind Med 45:29–32. Bouyer J, Hemon D. 1993. Studying the performance of a job exposure matrix. Int J Epidemiol 22:S65–S71. Brennan P. 2002. Gene-environment interaction and aetiology of cancer: What does it mean and how can we measure it? Carcinogenesis 23:381– 387. Bridbord K, Decoufle P, Fraumeni J Jr, Hoel D, Hoover R, Rall D, Saffiotti U, Schneiderman M, Upton A. 1978. Estimates of the fraction of cancer in the United States related to occupational factors. Bethesda: National Cancer Institute, National Institute of Environmental Health Sciences, and National Institute for Occupational Safety and Health (unpublished). Bridge J. 1933. Annual Report of the Chief Inspector for the Year 1932. London: HMSO. Britton M. 2002. The epidemiology of mesothelioma. Semin Oncol 29:18–25. Brody JG, Rudel RA. 2003. Environmental pollutants and breast cancer. Environ Health Perspect 111:1007–1019. Browne K, Gee JBL. 2000. Asbestos exposure and laryngeal cancer. Ann Occup Hyg 44:239–250. Bruning T, Bolt HM. 2000. Renal toxicity and carcinogenicity of trichloroethylene: Key results, mechanisms, and controversies. Crit Rev Toxicol 30:253–285. Bucher J. 2002. The National Toxicology Program rodent bioassay: Designs, interpretations, and scientific contributions. Ann N Y Acad Sci 982:198–207. Buffler P, Baan R, Bird M, Boffetta P, eds. 2004. Mechanisms of Carginogenesis: Contributions of Molecular Epidemiology 157. Lyon: IARC. Burstyn I, Boffetta P, Kauppinen T, Heikkila P, Svane O, Partanen T, Stucker I, Frentzel-Beyme R, Ahrens W, Merzenich H, Heederik D, Hooiveld M, Langard S, Randem BG, Jarvholm B, Bergdahl I, Shaham J, Ribak J,

348

PART III: THE CAUSES OF CANCER

Kromhout H. 2003. Estimating exposures in the asphalt industry for an international epidemiological cohort study of cancer risk. Am J Ind Med 43:3–17. Calle EE, Frumkin H, Henley SJ, Savitz DA, Thun MJ. 2002. Organochlorines and breast cancer risk. CA Cancer J Clin 52:301–309. Camus M, Siemiatycki J, Meek B. 1998. Nonoccupational exposure to chrysotile asbestos and the risk of lung cancer. N Engl J Med 338:1565–1571. Caplan LS, Schoenfeld ER, O’Leary ES, Leske MC. 2000. Breast cancer and electromagnetic fields—A review. Ann Epidemiol 10:31–44. Carozza SE, Wrensch M, Miike R, Newman B, Olshan AF, Savitz DA, Yost M, Lee M. 2000. Occupation and adult gliomas. Am J Epidemiol 152:838–846. Case R, Hosker M, McDonald D, Pearson J. 1954. Tumors of the urinary bladder in workmen engaged in the manufacture and use of certain dyestuff intermediates in the British chemical industry. Part I. The role of aniline, benzidine, alpha-naphthylamine and beta-naphthylamine. Br J Ind Med 11:75. Checkoway H, Pearce N, Crawford-Brown DJ, Kriebel D. 2004. Research Methods in Occupational Epidemiology. New York: Oxford University Press. Cocco P, Ward M, Buiatti E. 1996. Occupational risk factors for gastric cancer—An overview [review]. Epidemiol Rev 18:218–234. Cohen JT, Carlson G, Charnley G, Coggon D, Delzell E, Graham JD, Greim H, Krewski D, Medinsky M. 2002. A comprehensive evaluation of the potential health risks associated with occupational and environmental exposure to styrene. J Toxicol Environ Health B Crit Rev 5:1–263. Cohen S. 1995. Human relevance of animal carcinogenicity studies. Regul Toxicol Pharmacol 21:75–80; discussion 81–86. Cole P. 1977. Cancer and occupation: Status and needs of epidemiologic research. Cancer 39:1788–1791. Collegium Ramazzini. 1999. Call for an international ban on asbestos. J Occup Environ Med 41:830–832. Collins D, Richey F Jr. 1992. Synthetic organic chemicals. In: Kent J, ed. Riegel’s Handbook of Industrial Chemistry. New York: Van Nostrand Reinhold, pp. 800–862. Collins J, Acquavella J. 1998. Review and meta-analysis of studies of acrylonitrile workers. Scand J Work Environ Health 24:71–80. Committee on the Biological Effects of Ionizing Radiation. 1990. Health Effects of Exposure to Low Levels of Ionizing Radiation. BEIR V. Washington, DC: National Academy Press. Cook J, Hieger I, Kennaway E, Mayneord W. 1932. The production of cancer by pure hydrocarbons. Proc R Soc Lond B Biol Sci 111:455–484. Corn M. 1983. Regulations, standards and occupational hygiene within the U.S.A. in the 1980s. Ann Occup Hyg 27:91–105. Correa A, Jackson L, Mohan A, Perry H, Helzlsouer K. 2000a. Use of hair dyes, hematopoietic neoplasms, and lymphomas: a literature review. II. Lymphomas and multiple myeloma. Cancer Invest 18:467–479. Correa A, Mohan A, Jackson L, Perry H, Helzlsouer K. 2000b. Use of hair dyes, hematopoietic neoplasms, and lymphomas: A literature review. I. Leukemias and myelodysplastic syndromes. Cancer Invest 18:366–380. Costa M. 1997. Toxicity and carcinogenicity of Cr(VI) in animal models and humans. Crit Rev Toxicol 27:431–442. Cowles S, Tsai S, Snyder P, Ross C. 1994. Mortality, morbidity, and haematological results from a cohort of long-term workers involved in 1,3butadiene monomer production. Occup Environ Med 51:323–329. Creech J, Jr, Johnson M. 1974. Angiosarcoma of liver in the manufacture of polyvinyl chloride. J Occup Med 16:150–151. Dacre JC, Goldman M. 1996. Toxicology and pharmacology of the chemical warfare agent sulfur mustard [review] [347 refs]. Pharmacol Rev 48:289–326. De Roos AJ, Stewart PA, Linet MS, Heineman EF, Dosemeci M, Wilcosky T, Shapiro WR, Selker RG, Fine HA, Black PM, Inskip PD. 2003. Occupation and the risk of adult glioma in the United States. Cancer Causes Control 14:139–150. DeBruin L, Josephy P. 2002. Perspectives on the chemical etiology of breast cancer [review] [171 refs]. Environ Health Perspect 110:119–128. DeFonso L, Kelton S Jr. 1976. Lung cancer following exposure to chloromethyl methyl ether. An epidemiological study. Arch Environ Health 31:125–130. Delzell E, Macaluso M, Sathiakumar N, Matthews R. 2001. Leukemia and exposure to 1,3-butadiene, styrene and dimethyldithiocarbamate among workers in the synthetic rubber industry. Chem Biol Interact 135–136: 515–534. Delzell E, Sathiakumar N, Hovinga M, Macaluso M, Julian J, Larson R, Cole P, Muir D. 1996. A follow-up study of synthetic rubber workers. Toxicology 113:182–189. Delzell E, Monson RR. 1982. Mortality among rubber workers: VI. Men with potential exposure to acrylonitrile. J Occup Med 24:767–769.

Dement JM, Harris R, Jr, Symons M, Shy C. 1983. Exposures and mortality among chrysotile asbestos workers. Part II: Mortality. Am J Ind Med 4:421–433. Dement J, Brown D. 1994. Lung cancer mortality among asbestos textile workers: A review and update. Ann Occup Hyg 38:525–532. Demers PA, Kogevinas M, Boffetta P, Leclerc A, Luce D, Gerin M, Battista G, Belli S, Bolmaudorf U, Brinton LA, Colin D, Comba P, Hardell L, Hayes RB, Magnani C, Merler E, Morcet JF, Prestonmartin S, Matos E, Rodella S, Vaughan TL, Zheng W, Vainio H. 1995a. Wood dust and sino-nasal cancer: pooled reanalysis of twelve case-control studies. Am J Ind Med 28:151–166. Demers P, Boffetta P, Kogevinas M, Blair A, Miller B, Robinson C, Roscoe R, Winter P, Colin D, Matos E, Vainio H. 1995b. Pooled reanalysis of cancer mortality among five cohorts of workers in wood-related industries. Scand J Work Environ Health 21:179–190. Desmond RA, Soong SJ. 2003. Epidemiology of malignant melanoma. Surg Clin North Am 83:1–29. Dewar R, Siemiatycki JGM. 1991. Loss of statistical power associated with the use of a job-exposure matrix in occupational case-control studies. Appl Occup Environ Hyg 6:508–515. Dich J, Zahm S, Hanberg A, Adami H. 1997. Pesticides and cancer. Cancer Causes Control 8:420–443. Divine B, Hartman C. 1996. Mortality update of butadiene production workers. Toxicology 113:169–181. Doe J, Paddle G. 1994. The evaluation of carcinogenic risk to humans—Occupational exposures in the spraying and application of insecticides. Regul Toxicol Pharmacol 19:297–308. Doll R. 1952. The causes of death among gas-workers with special reference to cancer of the lung. Br J Ind Med 9:180. Doll R. 1955. Mortality from lung cancer in asbestos workers. Br J Ind Med 12:81. Doll R. 1958. Cancer of the lung and nose in nickel workers. Br J Ind Med 15:217–223. Doll R. 1975. 7th Walter Hubert Lecture: Pott and the prospects for prevention. Br J Cancer 32:263–272. Doll R. 1988. Effects of exposure to vinyl chloride. An assessment of the evidence. Scand J Work Environ Health 14:61–78. Doll R, Peto J. 1985. Asbestos: Effects on Health of Exposure to Asbestos. London: Her Majesty’s Stationery Office. Doll R, Peto R. 1981. The Causes of Cancer. Oxford: Oxford University Press. Dreyer L, Andersen A, Pukkala E. 1997. Avoidable cancers in the Nordic countries. Occupation. APMIS Suppl 105:68–79. Dumas S, Parent ME, Siemiatycki J, Brisson J. 2000. Rectal cancer and occupational risk factors: a hypothesis-generating, exposure-based case-control study. Int J Cancer 87:874–879. Eisenbud M, Kotin P, Miller F, Rogers A, Trichopoulos D, Deubner D, Powers M, Emly M, Hanes H, Stonehouse J, Wilson P. 1997. Is beryllium carcinogenic in humans? J Occup Environ Med 39:205–208. Engel LS, Vaughan TL, Gammon MD, Chow WH, Risch HA, Dubrow R, Mayne ST, Rotterdam H, Schoenberg JB, Stanford JL, West AB, Blot WJ, Fraumeni JF. 2002. Occupation and risk of esophageal and gastric cardia adenocarcinoma. Am J Ind Med 42:11–22. English D, Armstrong B, Kricker A, Fleming C. 1997. Sunlight and cancer [review]. Cancer Causes Control 8:271–283. Enterline P. 1993. Review of new evidence regarding the relationship of gasoline exposure to kidney cancer and leukemia. Environ Health Perspect 101:101–103. Enterline P, Henderson V, Marsh G. 1987. Exposure to arsenic and respiratory cancer. A reanalysis. Am J Epidemiol 125:929–938. Erren T, Jacobsen M, Piekarski C. 1999. Synergy between asbestos and smoking on lung cancer risks. Epidemiology 10:405–411. Feychting M. 1996. Occupational exposure to electromagnetic fields and adult leukaemia—A review of the epidemiological evidence [review]. Radiat Environ Biophys 35:237–242. Figueroa W, Raszkowski R, Weiss W. 1973. Lung cancer in chloromethyl methyl ether workers. N Engl J Med 288:1096–1097. Finkelstein MM. 2000. Silica, silicosis, and lung cancer: A risk assessment. Am J Ind Med 38:8–18. Fox A, Collier P. 1977. Mortality experience of workers exposed to vinyl chloride monomer in the manufacture of polyvinyl chloride in Great Britain. Br J Ind Med 34:1–10. Freedman D, Gold L, Lin T. 1996. Concordance between rats and mice in bioassays for carcinogenesis. Regul Toxicol Pharmacol 23:225– 232. Fritschi L, Siemiatycki J, Richardson L. 1996. Self-assessed versus expertassessed occupational exposures. Am J Epidemiol 144:521–527. Frumkin H, Thun M. 2000. Chapter 16. Carcinogens. In: Levy B, Wegman D, eds. Occupational Health: Recognizing and Preventing Work-Related

Occupation Disease and Injury. Philadelphia: Lippincott Williams & Wilkins, pp. 335–353. Fryzek J, Garabrant D, Greenson J, Schottenfeld D. 1997. A review of the epidemiology and pathology of pancreas cancer [review]. Gi Cancer 2:99–110. Fuchs B, Pritchard D. 2002. Etiology of osteosarcoma [review] [107 refs]. Clin Orthop 40–52. Gamble J. 1994. Asbestos and colon cancer: a weight-of-the-evidence review. Environ Health Perspect 102:1038–1050. Georgellis A, Kolmodin-Hedman B, Kouretas D. 1999. Can traditional epidemiology detect cancer risks caused by occupational exposure to pesticides? J Exp Clin Cancer Res 18:159–166. Gérin M, Siemiatycki J. 1991. The occupational questionnaire in retrospective epidemiologic studies: Recent approaches in community-based studies. Appl Occup Environ Hyg 6:495–501. Gold LS, Bernstein L, Magaw R, Slone TH. 1989. Interspecies extrapolation in carcinogenesis: Prediction between rats and mice. Environ Health Perspect 81:211–219. Gold L, Slone T, Ames B. 1997. Chapter 4. Overview and update of analyses of the Carcinogenic Potency Database. In: Gold LS, Zeiger E, eds. Handbook of Carcinogenic Potency and Genotoxicity Databases. Boca Raton: CRC Press. pp. 661–685. Gold LS, Slone TH, Ames BN. 1998. What do animal cancer tests tell us about human cancer risk?: Overview of analyses of the carcinogenic potency database. Drug Metab Rev 30:359–404. Goldberg MS, Parent ME, Siemiatycki J, Desy M, Nadon L, Richardson L, Lakhani R, Latreille B, Valois MF. 2001. A case-control study of the relationship between the risk of colon cancer in men and exposures to occupational agents. Am J Ind Med 39:531–546. Golden A, Markowitz S, Landrigan P. 1995. The risk of cancer in firefighters. Occup Med 10:803–820. Gottmann E, Kramer S, Pfahringer B, Helma C. 2001. Data quality in predictive toxicology: Reproducibility of rodent carcinogenicity experiments. Environ Health Perspect 109:509–514. Greenland S. 1993. Basic problems in interaction assessment. Environ Health Perspect 101 Suppl 4:59–66. Greenland S. 1999. Relation of probability of causation to relative risk and doubling dose: a methodologic error that has become a social problem. Am J Public Health 89:1166–1169. Greenland S, Robins JM. 1988. Conceptual problems in the definition and interpretation of attributable fractions. Am J Epidemiol 128:1185–1197. Grimsrud TK, Berge SR, Haldorsen T, Andersen A. 2002. Exposure to different forms of nickel and risk of lung cancer. Am J Epidemiol 156:1123–1132. Guidotti T. 1995. Occupational mortality among firefighters: Assessing the association. J Occup Environ Med 37:1348–1356. Hadfield M. 1998. Is silicon a human carcinogen? [review]. J Environ Sci Health Part C Environ Carcinog Ecotoxicol Rev 16:123–134. Hagmar L, Bonassi S, Stromberg U, Brogger A, Knudsen LE, Norppa H, Reuterwall C. 1998. Chromosomal aberrations in lymphocytes predict human cancer: A report from the European Study Group on Cytogenetic Biomarkers and Health (ESCH). Cancer Res 58:4117–4121. Hardell L, Axelson O. 1998. Environmental and occupational aspects on the etiology of non-Hodgkin’s lymphoma. Oncol Res 10:1–5. Härting F, Hesse W. 1879. Der Lungenkrebs, die Bergkrankheit in den Schneeberger Gruben. Vrtljhrssch Gerichtl Med 30:296–309. Hayes R. 1997. The carcinogenicity of metals in humans [review] [164 refs]. Cancer Causes Control 8:371–385. Hayes RB, Yin SN, Dosemeci M, Linet M. 2001. Benzene and lymphohematopoietic malignancies in humans. Am J Ind Med 40:117–126. Henry S. 1934. Industrial Maladies. London: Legge. Hieger I. 1933. The isolation of a cancer-producing hydrocarbon from coal tar. J Chem Soc [Perkin 1] 395. Higginson J. 1980. Proportion of cancers due to occupation. Prev Med 9:180–188. Higginson J, Muir CS. 1976. The role of epidemiology in elucidating the importance of environmental factors in human cancer. Cancer Detect Prev 1:79–105. Hill A, Faning E. 1948. Studies in the incidence of cancer in a factory handling inorganic compounds of arsenic. I. Mortality experience in the factory. Br J Ind Med 5:1–6. Hirayama T. 1981. Proportion of cancer attributable to occupation obtained from a census, population-based, large cohort study in Japan. In: Peto R, Schneiderman M, eds. Banbury Report 9. Quantification of Occupational Cancer. Cold Spring Harbor: Cold Spring Harbor Laboratory, pp. 631–649. Hoar SK, Morrison AS, Cole P, Silverman DT. 1980. An occupation and exposure linkage system for the study of occupational carcinogenesis. J Occup Med 22:722–726.

349

Hodgson JT, Darnton A. 2000. The quantitative risks of mesothelioma and lung cancer in relation to asbestos exposure. Ann Occup Hyg 44:565–601. Hoffmann W. 1996. Organochlorine compounds—risk of non-Hodgkins lymphoma and breast cancer. Arch Environ Health 51:189–192. Hogstedt C, Aringer L, Gustavsson A. 1986. Epidemiologic support for ethylene oxide as a cancer-causing agent. JAMA 255:1575–1578. Hogstedt C, Malmqvist N, Wadman B. 1979. Leukemia in workers exposed to ethylene oxide. JAMA 241:1132–1133. Hollstein M, Bartsch H , Wesch H, Kure EH, Mustonen R, Muhlbauer KR, Spiethoff A, Wegener K, Wiethege T, Muller KM. 1997. p53 gene mutation analysis in tumors of patients exposed to alpha-particles. Carcinogenesis 18:511–516. Holmberg B, Winell M. 1977. Occupational health standards. An international comparison. Scand J Work Environ Health 3:1–15. Hotz P, Lauwerys R. 1997. Hematopoietic and lymphatic malignancies in vehicle mechanics [review] [283 refs]. Crit Rev Toxicol 27:443–494. Howe G, Lindsay J. 1983. A follow-up study of a ten-percent sample of the Canadian labor force. I. Cancer mortality in males, 1965–73. J Natl Cancer Inst 70:37–44. Howe G, Nair R, Newcombe H, Miller A, Burch J, Abbatt J. 1987. Lung cancer mortality (1950–80) in relation to radon daughter exposure in a cohort of workers at the Eldorado Port Radium uranium mine: possible modification of risk by exposure rate. J Natl Cancer Inst 79:1255–1260. Huff J, Weisburger E, Fung V. 1996. Multicomponent criteria for predicting carcinogenicity: dataset of 30 NTP chemicals. Environ Health Perspect 104(suppl 5):1105–1112. IARC. 1972. Some inorganic substances, chlorinated hydrocarbons, aromatic amines, N-nitroso compounds, and natural products. IARC Monogr Eval Carcinog Risk Chem Man 1. Lyon: IARC. IARC. 1973a. Certain polycyclic aromatic hydrocarbons and heterocyclic compounds. IARC Monogr Eval Carcinog Risk Chem Man 3. Lyon: IARC. IARC. 1973b. Some inorganic and organometallic compounds. IARC Monogr Eval Carcinog Risk Chem Man 2. Lyon: IARC. IARC. 1974a. Some anti-thyroid and related substances, nitrofurans and industrial chemicals. IARC Monogr Eval Carcinog Risk Chem Man 7. Lyon: IARC. IARC. 1974b. Some organochlorine pesticides. IARC Monogr Eval Carcinog Risk Chem Man 5. Lyon: IARC. IARC. 1975. Some aromatic azo compounds. IARC Monogr Eval Carcinog Risk Chem Man 8. Lyon: IARC. IARC. 1976. Cadmium, nickel, some epoxides, miscellaneous industrial chemicals and general considerations on volatile anaesthetics. IARC Monogr Eval Carcinog Risk Chem Man 11. Lyon: IARC. IARC. 1977. Asbestos. IARC Monogr Eval Carcinog Risk Chem Man 14. Lyon: IARC. IARC. 1978. Some aromatic amines and related nitro compounds—Hair dyes, colouring agents and miscellaneous industrial chemicals. IARC Monogr Eval Carcinog Risk Chem Man 16. Lyon: IARC. IARC. 1979a. Some halogenated hydrocarbons. IARC Monogr Eval Carcinog Risk Chem Hum 20. Lyon: IARC. IARC. 1979b. Some monomers, plastics and synthetic elastomers, and acrolein. IARC Monogr Eval Carcinog Risk Chem Hum 19. Lyon: IARC. IARC. 1981. Wood, leather and some associated industries. IARC Monogr Eval Carcinog Risk Chem Hum 25. Lyon: IARC. IARC. 1982a. The rubber industry. IARC Monogr Eval Carcinog Risk Chem Hum 28. Lyon: IARC. IARC. 1982b. Some aromatic amines, anthraquinones and nitroso compounds, and inorganic fluorides used in drinking-water and dental preparations. IARC Monogr Eval Carcinog Risk Chem Hum 27. Lyon: IARC. IARC. 1982c. Some industrial chemicals and dyestuffs. IARC Monogr Eval Carcinog Risk Chem Hum 29. Lyon: IARC. IARC. 1983a. Miscellaneous pesticides. IARC Monogr Eval Carcinog Risk Chem Hum 30. Lyon: IARC. IARC. 1983b. Polynuclear aromatic compounds, part 1, chemical, environmental and experimental data. IARC Monogr Eval Carcinog Risk Chem Hum 32. Lyon: IARC. IARC. 1984a. Polynuclear aromatic compounds, part 2, carbon blacks, mineral oils and some nitroarenes. IARC Monogr Eval Carcinog Risk Chem Hum 33. Lyon: IARC. IARC. 1984b. Polynuclear aromatic compounds, part 3, industrial exposures in aluminium production, coal gasification, coke production, and iron and steel founding. IARC Monogr Eval Carcinog Risk Chem Hum 34. Lyon: IARC. IARC. 1985. Polynuclear aromatic compounds, part 4, bitumens, coal-tars and derived products, shale-oils and soots. IARC Monogr Eval Carcinog Risk Chem Hum 35. Lyon: IARC. IARC. 1986a. Some chemicals used in plastics and elastomers. IARC Monogr Eval Carcinog Risk Chem Hum 39. Lyon: IARC.

350

PART III: THE CAUSES OF CANCER

IARC. 1986b. Some halogenated hydrocarbons and pesticide exposures. IARC Monogr Eval Carcinog Risk Chem Hum 41. Lyon: IARC. IARC. 1986c. Some naturally occurring and synthetic food components, furocoumarins and ultraviolet radiation. IARC Monogr Eval Carcinog Risk Chem Hum 40. Lyon: IARC. IARC. 1986d. Tobacco Smoking. IARC Monogr Eval Carcinog Risk Chem Hum 38. Lyon: IARC. IARC. 1987a. Overall evaluations of carcinogenicity: an updating of IARC Monographs Volumes 1 to 42. IARC Monogr Eval Carcinog Risk Chem Hum Suppl 7. Lyon: IARC. IARC. 1987b. Silica and some silicates. IARC Monogr Eval Carcinog Risk Chem Hum 42. Lyon: IARC. IARC. 1988. Man-made mineral fibers and radon. IARC Monogr Eval Carcinog Risks Hum 43. Lyon: IARC. IARC. 1989a. Diesel and gasoline engine exhausts and some nitroarenes. IARC Monogr Eval Carcinog Risks Hum 46. Lyon: IARC. IARC. 1989b. Occupational exposures in petroleum refining; crude oil and major petroleum fuels. IARC Monogr Eval Carcinog Risks Hum 45. Lyon: IARC. IARC. 1989c. Some organic solvents, resin monomers and related compounds, pigments and occupational exposures in paint manufacture and painting. IARC Monogr Eval Carcinog Risks Hum 47. Lyon: IARC. IARC. 1990a. Chromium, nickel and welding. IARC Monogr Eval Carcinog Risks Hum 49. Lyon: IARC. IARC. 1990b. Some flame retardants and textile chemicals, and exposures in the textile manufacturing industry. IARC Monogr Eval Carcinog Risks Hum 48. Lyon: IARC. IARC. 1991a. Chlorinated drinking-water; chlorination by-products; some other halogenated compounds; cobalt and cobalt compounds. IARC Monogr Eval Carcinog Risks Hum 52. Lyon: IARC. IARC. 1991b. Coffee, tea, mate, methylxanthines and methylglyoxal. IARC Monogr Eval Carcinog Risks Hum 51. Lyon: IARC. IARC. 1991c. Occupational exposures in insecticide application, and some pesticides. IARC Monogr Eval Carcinog Risks Hum 53. Lyon: IARC. IARC. 1992a. Occupational exposures to mists and vapours from strong inorganic acids; and other industrial chemicals. IARC Monogr Eval Carcinog Risks Hum 54. Lyon: IARC. IARC. 1992b. Solar and ultraviolet radiation. IARC Monogr Eval Carcinog Risks Hum 55. Lyon: IARC. IARC. 1993a. Beryllium, cadmium, mercury, and exposures in the glass manufacturing industry. IARC Monogr Eval Carcinog Risks Hum 58. Lyon: IARC. IARC. 1993b. Occupational exposures of hairdressers and barbers and personal use of hair colourants; some hair dyes, cosmetic colourants, industrial dyestuffs and aromatic amines. IARC Monogr Eval Carcinog Risks Hum 57. Lyon: IARC. IARC. 1994a. Preamble. IARC Monogr Eval Carcinog Risks Hum. Lyon: IARC. IARC. 1994b. Some industrial chemicals. IARC Monogr Eval Carcinog Risks Hum 60. Lyon: IARC. IARC. 1995a. Dry cleaning, some chlorinated solvents and other industrial chemicals. IARC Monogr Eval Carcinog Risks Hum 63. Lyon: IARC. IARC. 1995b. Wood dust and formaldehyde. IARC Monogr Eval Carcinog Risks Hum 62. Lyon: IARC. IARC. 1996. Printing processes and printing inks, carbon black and some nitro compounds. IARC Monogr Eval Carcinog Risks Hum 65. Lyon: IARC. IARC. 1997a. Polychlorinated dibenzo-para-dioxins and polychlorinated dibenzofurans. IARC Monogr Eval Carcinog Risks Hum 69. Lyon: IARC. IARC. 1997b. Silica, some silicates, coal dust and para-aramid fibrils. IARC Monogr Eval Carcinog Risks Hum 68. Lyon: IARC. IARC. 1999a. Re-evaluation of some organic chemicals, hydrazine and hydrogen peroxide. IARC Monogr Eval Carcinog Risks Hum 71. Lyon: IARC. IARC. 1999b. Some chemicals that cause tumors of the kidney or urinary bladder in rodents and some other substances. IARC Monogr Eval Carcinog Risks Hum 73. Lyon: IARC. IARC. 2000a. Ionizing radiation, part 1. X-radiation and g-radiation, and neutrons. IARC Monogr Eval Carcinog Risks Hum 75. Lyon: IARC. IARC. 2000b. Some industrial chemicals. IARC Monogr Eval Carcinog Risks Hum 77. Lyon: IARC. IARC. 2001a. Ionizing radiation, part 2. Some internally deposited radionuclides. IARC Monogr Eval Carcinog Risks Hum 78. Lyon: IARC. IARC. 2001b. Some thyrotropic agents. IARC Monogr Eval Carcinog Risks Hum 79. Lyon: IARC. IARC. 2002a. Man-made vitreous fibers. IARC Monogr Eval Carcinog Risks Hum 81. Lyon: IARC. IARC. 2002b. Non-ionizing radiation, part 1. Static and extremely lowfrequency (ELF) electric and magnetic fields. IARC Monogr Eval Carcinog Risks Hum 80. Lyon: IARC.

IARC. 2002c. Some traditional herbal medicines, some mycotoxins, naphthalene and styrene. IARC Monogr Eval Carcinog Risks Hum 82. Lyon: IARC. IARC. 2003. IARC Monographs Programme on the Evaluation of Carcinogenic Risks to Humans. Available at: http://193.51.164.11/default.html. Accessed August 14, 2003. IARC. Cobalt in hard-metals and cobalt sulfate, gallium arsenide, indium phosphide and vanadium pentoxide. IARC Monogr Eval Carcinog Risks Hum 86. Lyon: IARC. In preparation. IARC. 2004. Tobacco smoke and involuntary smoking. IARC Monogr Eval Carcinog Risks Hum 83. Lyon: IARC. Infante P. 1993. State of the science on the carcinogenicity of gasoline with particular reference to cohort mortality study results. Environ Health Perspect 101:105–109. Infante P, Rinsky R, Wagoner J, Young R. 1977. Leukaemia in benzene workers. Lancet 2:76–78. IPCS (International Programme on Chemical Safety). 1998. Chrysotile Asbestos. In: Environmental Health Criteria series. Geneva: World Health Organization. Ishimaru T, Okada H, Tomiyasu T, Tsuchimoto T, Hoshino T, Ichimaru M. 1971. Occupational factors in the epidemiology of leukemia in Hiroshima and Nagasaki. Am J Epidemiol 93:157–165. Jakobsson K, Albin M, Hagmar L. 1994. Asbestos, cement, and cancer in the right part of the colon. Occup Environ Med 51:95–101. Joffe M. 1992. Validity of exposure data derived from a structured questionnaire. Am J Epidemiol 135:564–570. Kaldor J, Peto J, Easton D, Doll R, Hermon C, Morgan L. 1986. Models for respiratory cancer in nickel refinery workers. J Natl Cancer Inst 77:841–848. Karalliedde L, Wheeler H, Maclehose R, Murray V. 2000. Possible immediate and long-term health effects following exposure to chemical warfare agents [review] [95 refs]. Public Health 114:238–248. Katsouyanni K, Pershagen G. 1997. Ambient air pollution exposure and cancer [review]. Cancer Causes Control 8:284–291. Kauppinen T, Toikkanen J, Pukkala E. 1998. From cross-tabulations to multipurpose exposure information systems—A new job-exposure matrix. Am J Ind Med 33:409–417. Kazantzis G, Blanks R, Sullivan K. 1992. Is cadmium a human carcinogen? In: Nordberg GF, Herber RFM, Alessio L, editors. Cadmium in the Human Environment: Toxicity and Carcinogenicity. Lyon: IARC, 435–446. Kennaway E, Hieger I. 1930. Carcinogenic substances and their fluorescence spectra. Br Med J 1:1044–1046. Kheifets LI. 2001. Electric and magnetic field exposure and brain cancer: A review. Bioelectromagnetics S120–S131. Kheifets L, Afifi A, Buffler P, Zhang Z, Matkin C. 1997. Occupational electric and magnetic field exposure and leukemia. A meta-analysis. J Occup Environ Med 39:1074–1091. Kheifets L, Matkin C. 1999. Industrialization, electromagnetic fields, and breast cancer risk [review] [71 refs]. Environ Health Perspect 107:145–154. Kim B, Margolin B. 1999. Prediction of rodent carcinogenicity utilizing a battery of in vitro and in vivo genotoxicity tests. Environ Mol Mutagen 34:297–304. Kipling M, Waterhouse J. 1967. Cadmium and prostatic carcinoma. Lancet 1:730–731. Kjaerheim K, Boffetta P, Hansen J, Cherrie J, Chang-Claude J, Eilber U, Ferro G, Guldner K, Olsen J, Plato N, Proud L, Saracci R. 2002. Lung cancer among rock and slag wool production workers. Epidemiology 13:445–453. Kogevinas M, Ferro G, Andersen A, Bellander T, Biocca M, Coggon D, Gennaro V, Hutchings S, Kolstad H, Lundberg I, Lynge E, Partanen T, Saracci R. 1994. Cancer mortality in a historical cohort study of workers exposed to styrene. Scand J Work Environ Health 20:251–261. Kogevinas M, ‘t Mannetje A, Cordier S, Ranft U, Gonzalez CA, Vineis P, Chang-Claude J, Lynge E, Wahrendorf J, Tzonou A, Jockel KH, Serra C, Porru S, Hours M, Greiser E, Boffetta P. 2003. Occupation and bladder cancer among men in Western Europe. Cancer Causes Control 14:907–914. Kolstad H, Juel K, Olsen J, Lynge E. 1995. Exposure to styrene and chronic health effects: mortality and incidence of solid cancers in the danish reinforced plastics industry. Occup Environ Med 52:320–327. Koufman J, Burke A. 1997. The etiology and pathogenesis of laryngeal caracinoma [review]. Otolaryngol Clin North Am 30:1–19. Krewski D, Snyder R, Beatty P, Granville G, Meek B, Sonawane B. 2000. Assessing the health risks of benzene: A report on the benzene state-ofthe-science workshop. J Toxicol Environ Health A 61:307–338. Kromhout H, Heederik D, Dalderup L, Kromhout D. 1992. Performance of two general job-exposure matrices in a study of lung cancer morbidity in the Zutphen cohort. Am J Epidemiol 136:698–711.

Occupation Kuroda S, Kawahata K. 1936. Uber die gewerbliche Entstehung des Lungenkrebses bei Generatorgasarbeitern. Z Krebsforsch 45:36–39. La Vecchia C, Tavani A. 1995. Epidemiological evidence on hair dyes and the risk of cancer in humans. Eur J Cancer Prev 4:31–43. La Vecchia C, Tavani A. 2001. Hair dyes and bladder cancer: An update. Eur J Cancer Prev 10:205–208. La Vecchia C, Tavani A. 2002. Hair dyes and lymphoid neoplasms: An update. Eur J Cancer Prev 11:409–412. Labreche F, Goldberg M. 1997. Exposure to organic solvents and breast cancer in women: a hypothesis [review] [125 refs]. Am J Ind Med 32:1–14. Landrigan P, Nicholson W, Suzuki Y, Ladou J. 1999. The hazards of chrysotile asbestos: A critical review [review] [21 refs]. Ind Health 37:271–280. Lee A, Fraumeni J Jr. 1969. Arsenic and respiratory cancer in man: An occupational study. J Natl Cancer Inst 42:1045–1052. Lee-Feldstein A. 1986. Cumulative exposure to arsenic and its relationship to respiratory cancer among copper smelter employees. J Occup Med 28:296–302. Lee P. 2001. Relation between exposure to asbestos and smoking jointly and the risk of lung cancer. Occup Environ Med 58:145–153. Lemen R, Young R. 1976. Investigation of health hazards in styrene-butadiene rubber facilities. In: Ede L, ed. Proceeding of NIOSH Styrene-Butadiene Rubber Briefing, Covington, Kentucky, April 30, 1976. Cincinnati: National Institute for Occupational Safety and Health, pp. 3–8. Lemen RA, Lee JS, Wagoner JK, Blejer HP. 1976. Cancer mortality among cadmium production workers. Ann N Y Acad Sci 271:273–279. Leonard A, Gerber G, Stecca C, Rueff J, Borba H, Farmer P, Sram R, Czeizel A, Kalina I. 1999. Mutagenicity, carcinogenicity, and teratogenicity of acrylonitrile [review] [188 refs]. Mutat Res 436:263–283. Levy B, Wegman D. 1988. Occupational Health. Boston: Little, Brown and Company. Liddell F. 2001. The interaction of asbestos and smoking in lung cancer. Ann Occup Hyg 45:341–356. Liddell F, Armstrong B. 2002. The combination of effects on lung cancer of cigarette smoking and exposure in Quebec chrysotile miners and millers. Ann Occup Hyg 46:5–13. Linet MS, Stewart WF, Van Natta ML, McCaffrey LD, Szklo M. 1987. Comparison of methods for determining occupational exposure in a casecontrol interview study of chronic lymphocytic leukemia. J Occup Med 29:136–141. Liteplo RG, Meek ME. 2002. Inhaled formaldehyde: Exposure estimation, hazard characterization, and exposure-response analysis. J Toxicol Environ Health B Crit Rev 6:85–114. Longnecker M, Rogan W, Lucier G. 1997. The human health effects of DDT (dichlorodiphenyl-trichloroethane) and PCBs (polychlorinated biphenyls) and an overview of organochlorines in public health [review]. Annu Rev Public Health 18:211–244. Lowenfels A, Maisonneuve P. 2002. Epidemiologic and etiologic factors of pancreatic cancer. Hematol Oncol Clin North Am 16:1–16. Luce D, Gerin M, Berrino F, Pisani P, Leclerc A. 1993. Sources of discrepancies between a job exposure matrix and a case by case expert assessment for occupational exposure to formaldehyde and wood dust. Int J Epidemiol 22:S113–S120. Lutz JM, Cree IA, Foss AJ. 1999. Risk factors for intraocular melanoma and occupational exposure. Br J Ophthalmol 83:1190–1193. Lynch K, Smith W. 1935. Pulmonary asbestosis III: Carcinoma of lung in asbesto-silicosis. Am J Cancer 24:56–64. Lynge E, Anttila A, Hemminki K. 1997. Organic solvents and cancer [review]. Cancer Causes Control 8:406–419. Lynge E, Thygesen L. 1990. Occupational cancer in Denmark. Cancer incidence in the 1970-census population. Scand J Work Environ Health 16:1–35. Macaluso M, Larson R, Delzell E, Sathiakumar N, Hovinga M, Julian J, Muir D, Cole P. 1996. Leukemia and cumulative exposure to butadiene, styrene and benzene among workers in the synthetic rubber industry. Toxicology 113:190–202. Machle W, Gregorius F. 1948. Cancer of the respiratory system in the United States chromate-producing industry. Public Health Rep 63:1114–1127. MacMahon B. 1994. The epidemiological evidence on the carcinogenicity of beryllium in humans. J Occup Med 36:15–24. Magnus K, Andersen A, Hogetveit AC. 1982. Cancer of respiratory organs among workers at a nickel refinery in Norway. Int J Cancer 30:681–685. Maher KV, DeFonso LR. 1987. Respiratory cancer among chloromethyl ether workers. J Natl Cancer Inst 78:839–843. Maier H, Tisch M. 2000. Occupation and head and neck cancer. Onkologie 23:34–40. Mallory T, Gall E, Brickley W. 1939. Chronic exposure to benzene (benzol). III The pathologic results. Journal of Industrial Hygiene and Toxicology 21:355–377.

351

Marion MJ, Boivin-Angele S. 1999. Vinyl chloride-specific mutations in humans and animals. IARC Sci Publ 315–324. Marsh G, Buchanich J, Youk A. 2001. Historical cohort study of US man-made vitreous fiber production workers: VI. Respiratory system cancer standardized mortality ratios adjusted for the confounding effect of cigarette smoking. J Occup Environ Med 43:803–808. Mastrangelo G, Fadda E, Marzia V. 1996. Polycyclic aromatic hydrocarbons and cancer in man [review]. Environ Health Perspect 104:1166–1170. McCallum R, Woolley V, Petrie A. 1983. Lung cancer associated with chloromethyl methyl ether manufacture: An investigation at two factories in the United Kingdom. Br J Ind Med 40:384–389. McCunney RJ. 1999. Hodgkin’s disease, work, and the environment—A review. J Occup Environ Med 41:36–46. McDonald J, Liddell F, Gibbs G, Eyssen G, McDonald A. 1980. Dust exposure and mortality in chrysotile mining, 1910–75. Br J Ind Med 37:11–24. McDonald J, McDonald A. 1996. The epidemiology of mesothelioma in historical context. Eur Respir J 9:1932–1942. McLaughlin J. 1993. Renal cell cancer and exposure to gasoline—A review. Environ Health Perspect 101:111–114. McLaughlin J. 1994. Formaldehyde and cancer: A critical review. Int Arch Occup Environ Health 66:295–301. McLaughlin J, Blot W. 1997. A critical review of epidemiology studies of trichloroethylene and perchloroethylene and risk of renal-cell cancer [review]. Int Arch Occup Environ Health 70:222–231. McLaughlin J, Lipworth L. 1999. A critical review of the epidemiologic literature on health effects of occupational exposure to vinyl chloride [review] [59 refs]. J Epidemiol Biostat 4:253–275. McLaughlin J, Lipworth L. 2000. Epidemiologic aspects of renal cell cancer [review] [112 refs]. Semin Oncol 27:115–123. McMichael A, Spirtas R, Gamble J, Toussey P. 1976. Mortality among rubber workers: Relationship to specific jobs. J Occup Med 18:178–185. Mehlman M. 1996. Dangerous and cancer-causing properties of products and chemicals in the oil-refining and petrochemical industry—Part XXII: health hazards from exposure to gasoline containing methyl tertiary butyl ether: study of New Jersey residents [review] [56 refs]. Toxicol Ind Health 12:613–627. Meigs JW, Marrett LD, Ulrich FU, Flannery JT. 1986. Bladder tumor incidence among workers exposed to benzidine: a thirty-year follow-up. J Natl Cancer Inst 76:1–8. Merewether E. 1949. Asbestosis and carcinoma of the lung. In: Annual Report of the Chief Inspector of Factories for the year 1947. London: HMSO, pp. 79–81. Milham S Jr. 1981. Proportion of cancer due to occupation in Washington State. In: Peto R and Schneiderman M, eds. Banbury Report 9. Quantification of Occupational Cancer. Cold Spring Harbor: Cold Spring Harbor Laboratory, pp. 513–522. Miller A. 1997. Occupational radiation carcinogenesis (overview). Eur J Cancer 33:S7–S9. Miller R, Neuberger J, Gerald K. 1997. Brain cancer and leukemia and exposure to power-frequency (50- to 60-Hz) electric and magnetic fields. Epidemiol Rev 19:273–293. Monson RR. 1990. Occupational Epidemiology. Boca Raton: CRC Press. Monson RR. 1996. Occupation. In: Cancer Epidemiology and Prevention. Schottenfeld D, Fraumeni JF Jr, eds. New York: Oxford University Press, pp. 373–405. Monson R, Peters J, Johnson M. 1974. Proportional mortality among vinylchloride workers. Lancet 2:397–398. Montesano R, Bartsch H, Vainio H, Wilbourn J, Yamasaki H. 1986. Long-Term and Short-Term Assays for Carcinogens: A Critical Appraisal. Lyon: International Agency for Research on Cancer. Morgan W, Reger R, Tucker D. 1997. Health effects of diesel emissions. Ann Occup Hyg 41:643–658. Morley J. 1911. The lymphatics of the scrotum in relation to the radical operation for scrotal epithelioma. Lancet 2:1545–1547. Moulin J. 1997. A meta-analysis of epidemiologic studies of lung cancer in welders. Scand J Work Environ Health 23:104–113. Mundt K, Dell L, Austin R, Luippold R, Noess R, Bigelow C. 2000. Historical cohort study of 10,109 men in the North American vinyl chloride industry, 1942–72: update of cancer mortality to 31 December 1995. Occup Environ Med 57:774–781. NAS (National Academy of Sciences). 1999. The Health Effects of Exposure to Indoor Radon (BEIR VI). Washington, DC: National Academy Press. Nauss K, Busby W, Cohen A, Green G, Higgins M, McClellan R, Rosenkranz H, Sawyer R, Upton A, Watson A, Watts W, Winer A. 1995. Critical issues in assessing the carcinogenicity of diesel exhaust: a synthesis of current knowledge. In: Diesel Exhaust: A Critical Analysis of Emissions, Exposure, and Health Effects. Health Effects Institute, Cambride, MA: Health Effects Institute, pp. 11–61.

352

PART III: THE CAUSES OF CANCER

Navas-Acien A, Pollan M, Gustavsson P, Plato N. 2002. Occupation, exposure to chemicals and risk of gliomas and meningiomas in Sweden. Am J Ind Med 42:214–227. Negri E, La Vecchia C. 2001. Epidemiology and prevention of bladder cancer. Eur J Cancer Prev 10:7–14. Nicholson WJ. 1984. Quantitative estimates of cancer in the workplace. Am J Ind Med 5:341–342. Nicholson W. 1986. Airborne Asbestos Health Assessment Update. Washington, DC: Office of Health and Environmental Assessment, U.S. Environmental Protection Agency. Nicholson W, Selikoff I, Seidman H. 1978. Mortality experience of styrenepolystyrene polymerization workers. Initial findings. Scand J Work Environ Health 4 Suppl 2:247–252. Nurminen M, Karjalainen A. 2001. Epidemiologic estimate of the proportion of fatalities related to occupational factors in Finland. Scand J Work Environ Health 27:161–213. Nylander L, Dement J. 1993. Carcinogenic effects of wood dust—Review and discussion. Am J Ind Med 24:619–647. O’Berg M. 1980. Epidemiologic study of workers exposed to acrylonitrile. J Occup Med 22:245–252. O’Connor S, Farmer P, Lauder I. 1999. Benzene and non-Hodgkin’s lymphoma [review] [112 refs]. J Pathol 189:448–453. Ojajarvi A, Partanen T, Ahlbom A, Boffetta P, Hakulinen T, Jourenkova N, Kauppinen T, Kogevinas M, Vainio H, Weiderpass E, Wesseling C. 2001. Risk of pancreatic cancer in workers exposed to chlorinated hydrocarbon solvents and related compounds: A meta-analysis. Am J Epidemiol 153:841–850. Ojajarvi IA, Partanen TJ, Ahlbom A, Boffetta P, Hakulinen T, Jourenkova N, Kauppinen TP, Kogevinas M, Porta M, Vainio HU, Weiderpass E, Wesseling CH. 2000. Occupational exposures and pancreatic cancer: A meta-analysis. Occup Environ Med 57:316–324. Oller A. 2002. Respiratory carcinogenicity assessment of soluble nickel compounds [review] [50 refs]. Environ Health Perspect 110:841– 844. Osinubi OYO, Gochfeld M, Kipen HM. 2000. Health effects of asbestos and nonasbestos fibers. Environ Health Perspect 108:665–674. Parent ME, Siemiatycki J. 2001. Occupation and prostate cancer. Epidemiol Rev 23:138–143. Parent M, Siemiatycki J, Fritschi L. 1998. Occupational exposures and gastric cancer. Epidemiology 9:48–55. Parent ME, Siemiatycki J, Fritschi L. 2000. Workplace exposures and oesophageal cancer. Occup Environ Med 57:325–334. Partanen T, Boffetta P. 1994. Cancer risk in asphalt workers and roofers: Review and meta-analysis of epidemiologic studies. Am J Ind Med 26:721–740. Pearce N, Matos E, Vainio H, Boffetta P, Kogevinas M. 1994. Occupational Cancer in Developing Countries. Lyon: IARC (International Agency for Research on Cancer). Peller S. 1939. Lung cancer among mine workers in Joachimsthal. Hum Biol 11:130–143. Perera F, Mayer J, Jaretzki A, Hearne S, Brenner D, Young TL, Fischman HK, Grimes M, Grantham S, Tang MX, et al. 1989. Comparison of DNA adducts and sister chromatid exchange in lung cancer cases and controls. Cancer Res 49:4446–4451. Peto R, Schneiderman M. 1981. Banbury Report 9. Quantification of Occupational Cancer. Cold Spring Harbor: Cold Spring Harbor Laboratory. Pinto SS, Henderson V, Enterline PE. 1978. Mortality experience of arsenicexposed workers. Arch Environ Health 325–330. Pollan M, Gustavsson P, Cano MI. 2001. Incidence of testicular cancer and occupation among Swedish men gainfully employed in 1970. Ann Epidemiol 11:554–562. Pott P. 1775. Chirurgical observations relative to the cataract, the polypus of the nose, the cancer of the scotum, the different kinds of ruptures and the mortification of the toes and feet. London: T. J. Carnegy. Potts C. 1965. Cadmium proteinuria: The health of battery workers exposed to cadmium oxide dust. Ann Occup Hyg 8:55–61. Purchase I. 1986. Carcinogenic risk assessment: Are animals good surrogates for man? In: Bannasch P, ed. Cancer Risks: Strategies for Elimination. Berlin: Springer Verlag, pp. 65–79. Raabe G. 1993. Review of the carcinogenic potential of gasoline. Environ Health Perspect 101:35–38. Raj A, Mayberry JF, Podas T. 2003. Occupation and gastric cancer. Postgrad Med J 79:252–258. Rall D, Hogan M, Huff J, Schwetz B, Tennant R. 1987. Alternatives to using human experience in assessing health risks. Annu Rev Public Health 8:355–385. Rappaport S, Smith T. 1991. Exposure Assessment for Epidemiology and Hazard Control. Chelsea, Michigan: Lewis Publishers.

Rehn L. 1895. Blasengeschwulste bei Fuchsin-Arbeitern. Arch Klin Chir 50:588–600. Rinsky R, Smith A, Hornung R, Filloon T, Young R, Okun A, Landrigan P. 1987. Benzene and leukemia. An epidemiologic risk assessment. N Engl J Med 316:1044–1050. Rockley P, Trieff N, Wagner R, Tyring S. 1994. Nonsunlight risk factors for malignant melanoma .1. Chemical agents, physical conditions, and occupation. Int J Dermatol 33:398–406. Rom W. 1992. Environmental and Occupational Medicine. Boston: Little, Brown and Company. Ron E. 1998. Ionizing radiation and cancer risk—Evidence from epidemiology [review]. Radiat Res 150:S 30–S 41. Roth VS. 1999. Rubber industry epidemiology. Occup Med 14:849–856. Rothman KJ. 1974. Synergy and antagonism in cause-effect relationships. Am J Epidemiol 99:385–388. Rothman KJ, Greenland S. 1998. Modern Epidemiology. Philadelphia: Lippincott-Raven Publishers. Samet JM, Eradze GR. 2000. Radon and lung cancer risk: Taking stock at the millenium. Environ Health Perspect 108:635–641. Saracci R. 1987. The interactions of tobacco smoking and other agents in cancer etiology. Epidemiol Rev 9:175–193. Saracci R, Boffetta P. 1994. Interactions of tobacco smoking with other causes of lung cancer. In: Samet JM, ed. Epidemiology of Lung Cancer. New York: Marcel Dekker, pp. 465–493. Sathiakumar N, Delzell E. 1997. A review of epidemiologic studies of triazine herbicides and cancer [review]. Crit Rev Toxicol 27:599–612. Sathiakumar N, Delzell E, Hovinga M, Macaluso M, Julian J, Larson R, Cole P, Muir D. 1998. Mortality from cancer and other causes of death among synthetic rubber workers. Occup Environ Med 55:230–235. Savitz D, Andrews K. 1997. Review of epidemiologic evidence on benzene and lymphatic and hematopoietic cancers [review] [74 refs]. Am J Ind Med 31:287–295. Schubauer-Berigan MK, Wenzl TB. 2001. Leukemia mortality among radiation-exposed workers. Occup Med 16:271–287. Schulte PA, Perera FP. 1993. Molecular Epidemiology: Principles and Practices. Troy, Missouri: Harcourt Brace and Company. Schwartz R. 1997. Arsenic and the skin [review] [144 refs]. Int J Dermatol 36:241–250. Scott T. 1952. The incidence of tumours in a dyestuffs factory. Br J Ind Med 9:127–132. Seidman H, Selikoff I, Gelb S. 1986. Mortality experience of amosite asbestos factory workers: Dose-response relationships 5 to 40 years after onset of short-term work exposure. Am J Ind Med 10:479–514. Seilkop SK, Oller AR. 2003. Respiratory cancer risks associated with low-l evel nickel exposure: An integrated assessment based on animal, epidemiological, and mechanistic data. Regul Toxicol Pharmacol 37: 173–190. Selikoff I, Churg J, Hammond E. 1964. Asbestos exposure and neoplasia. JAMA 118:22–26. Selikoff I. 1990. Historical developments and perspectives in inorganic fiber toxicity in man. Environ Health Perspect 88:269–276. Selikoff I, Hammond E, Churg J. 1968. Asbestos exposure, smoking, and neoplasia. JAMA 204:106–112. Selikoff I, Hammond E, Seidman H. 1973. Cancer risk of insulation workers in the United States. In: Bogovski P, Timbrell V, Gilson JC, Wagner JC, eds. Biological Effects of Asbestos. Lyon: International Agency for Research on Cancer, pp. 209–216. Shen N, Weiderpass E, Anttila A, Goldberg M, Vasamaneuvonen K, Boffetta P, Vainio H, Partanen T. 1998. Epidemiology of occupational and environmental risk factors related to ovarian cancer [review]. Scand J Work Environ Health 24:175–182. Shubik P. 1979. Identification of environmental carcinogens: Animal test models. In: Clark Griffin A, Shaw CR, eds. Carcinogens: Identifications and Mechanisms of Action. New York: Raven Press, pp. 37–47. Sieber WK Jr., Sundin DS, Frazier TM, Robinson CF. 1991. Development, use, and availability of a job exposure matrix based on National Occupational Hazard Survey data. Am J Ind Med 20:163–174. Siemiatycki J. 1996. Exposure assessment in community-based studies of occupational cancer. Occup Hyg 3:41–58. Siemiatycki J. 2001. Should Canadian health care professionals support the call for a worldwide ban on asbestos? CMAJ 164:495–497. Siemiatycki J, Day NE, Fabry J, Cooper JA. 1981. Discovering carcinogens in the occupational environment: A novel epidemiologic approach. J Natl Cancer Inst 66:217–225. Siemiatycki J, Dewar R, Krewski D, Desy M, Richardson L, Franco E. 1994a. Are the apparent effects of cigarette smoking on lung and bladder cancers due to uncontrolled confounding by occupational exposures? Epidemiology 5:57–65.

Occupation Siemiatycki J, Dewar R, Nadon L, Gérin M. 1994b. Occupational risk factors for bladder cancer: results from a case-control study in Montreal, Quebec, Canada. Am J Epidemiol 140:1061–1080. Siemiatycki J, Dewar R, Richardson L. 1989. Costs and statistical power associated with five methods of collecting occupation exposure information for population-based case-control studies. Am J Epidemiol 130:1236– 1246. Siemiatycki J, Nadon L, Lakhani R, Bégin D, Gérin M. 1991. Chapter 4. Exposure assessment. In: Siemiatycki J, ed. Risk Factors for Cancer in the Workplace. Boca Raton: CRC Press, pp. 45–114. Siemiatycki J, Richardson L, Straif K, Latreille B, Lakhani R, Campbell S, Rousseau MC, Boffetta P. Listing occupational carcinogens. Environ Health Perspect. 112:1447–1459. Siemiatycki J, Thomas DC. 1981. Biological models and statistical interactions: An example from multistage carcinogenesis. Int J Epidemiol 10:383–387. Siemiatycki J, Wacholder S, Dewar R, Wald L, Bégin D, Richardson L, Rosenman K, Gérin M. 1988. Smoking and degree of occupational exposure: Are internal analyses in cohort studies likely to be confounded by smoking status? Am J Ind Med 13:59–69. Silverman D. 1998. Is diesel exhaust a human lung carcinogen? Epidemiology 9:4–6. Skov T, Lynge E. 1994. Cancer risk and exposures to carcinogens in hairdressers. Skin Pharmacol 7:94–100. Smith T, Hammond S, Wong O. 1993. Health effects of gasoline exposure .1. Exposure assessment for US distribution workers. Environ Health Perspect 101:13–21. Snedeker SM. 2001. Pesticides and breast cancer risk: A review of DDT, DDE, and dieldrin. Environ Health Perspect 109:35–47. Snyderwine E, Sinha R, Felton J, Ferguson L. 2002. Highlights of the eighth international conference on carcinogenic/mutagenic N-substituted aryl compounds. Mutat Res 506–507:1–8. Sorahan T, Waterhouse J. 1983. Mortality study of nickel-cadmium battery workers by the method of regression models in life tables. Br J Ind Med 40:293–300. Soutar CA, Robertson A, Miller BG, Searl A, Bignon J. 2000. Epidemiological evidence on the carcinogenicity of silica: Factors in scientific judgement. Ann Occup Hyg 44:3–14. Southam A, Wilson S. 1922. Cancer of the scrotum: The aetiology, clinical features, and treatment of the disease. Br Med J 2:971–973. Stayner L, Steenland K, Greife A, Hornung R, Hayes R, Nowlin S, Morawetz J, Ringenburg V, Elliot L, Halperin W. 1993. Exposure-response analysis of cancer mortality in a cohort of workers exposed to ethylene oxide. Am J Epidemiol 138:787–798. Steenland K, Beaumont J. 1984. The accuracy of occupation and industry data on death certificates. J Occup Med 26:288–296. Steenland K, Loomis D, Shy C, Simonsen N. 1996. Review of occupational lung carcinogens [review]. Am J Ind Med 29:474–490. Steenland K, Stayner L. 1997. Silica, asbestos, man-made mineral fibers, and cancer. Cancer Causes Control 8:491–503. Steenland K, Burnett C, Lalich N, Ward E, Hurrell J. 2003. Dying for work: The magnitude of US mortality from selected causes of death associated with occupation. Am J Ind Med 43:461–482. Stewart P, Zaebst D, Zey J, Herrick R, Dosemeci M, Hornung R, Bloom T, Pottern L, Miller B, Blair A. 1998. Exposure assessment for a study of workers exposed to acrylonitrile. Scand J Work Environ Health 24:42–53. Stober W, Abel U, McClellan R. 1998. Revisiting epidemiological key studies on occupational diesel exhaust exposure and lung cancer in truck drivers. Inhal Toxicol 10:1045–1078. Stone R. 1991. No meeting of the minds on asbestos. Science 254:928–931. Stucker I, Bouyer J, Mandereau L, Hemon D. 1993. Retrospective evaluation of the exposure to polycyclic aromatic hydrocarbons—Comparative assessments with a job exposure matrix and by experts in industrial hygiene. Int J Epidemiol 22:S106–S112. Sullivan J, Krieger G. 2001. Clinical Environmental Health and Toxic Exposures. Philadelphia: Lippincott Williams & Wilkins. Sullivan P, Eisen E, Woskie S, Kriebel D, Wegman D, Hallock M, Hammond S, Tolbert P, Smith T, Monson R. 1998. Mortality studies of metalworking fluid exposure in the automobile industry—vi—A case-control study of esophageal cancer. Am J Ind Med 34:36–48. Swaen G, Bloemen L, Twisk J, Scheffers T, Slangen J, Collins J, Tenberge W, Sturmans F. 1998. Mortality update of workers exposed to acrylonitrile in the netherlands. Scand J Work Environ Health 24:10–16. Swenberg J, Lehman-McKeeman L. 1999. Alpha 2-urinary globulin associated nephropathy as a mechanism of renal tubule cell carcinogenesis in male rats. In: Capen C, Dybing E, Rice J, Wilbourn J, eds. Species Differences in Thyroid, Kidney and Urinary Bladder Carcinogenesis. Lyon: IARC, 95–118.

353

Swerdlow AJ. 1990. Effectiveness of primary prevention of occupational exposures on cancer risk. In: Hakama M, Beral V, Cullen JW, Parkin DM, eds. Evaluating Effectiveness of Primary Prevention of Cancer. Lyon: International Agency for Research on Cancer, pp. 23–56. Symanski E, Kupper LL, Hertzpicciotto I, Rappaport SM. 1998. Comprehensive evaluation of long term trends in occupational exposure—part 2— Predictive models for declining exposures. Occup Environ Med 55:310–316. Tabershaw IR, Gaffey WR. 1974. Mortality study of workers in the manufacture of vinyl chloride and its polymers. J Occup Med 16:509–518. Taylor JA, Watson MA, Devereux TR, Michels RY, Saccomanno G, Anderson M. 1994. p53 mutation hotspot in radon-associated lung cancer. Lancet 343:86–87. Tennant R, Spalding J, Stasiewicz S, Ashby J. 1990. Prediction of the outcome of rodent carcinogenicity bioassays currently being conducted on 44 chemicals by the National Toxicology Program. Mutagenesis 5:3–14. Teta M, Benson L, Vitale J. 1993. Mortality study of ethylene oxide workers in chemical manufacturing—A 10 year update. Br J Ind Med 50:704–709. Thier R, Bolt H. 2000. Carcinogenicity and genotoxicity of ethylene oxide: New aspects and recent advances [review] [91 refs]. Crit Rev Toxicol 30:595–608. Thun M, Schnorr T, Smith A, Halperin W, Lemen R. 1985. Mortality among a cohort of U.S. cadmium production workers—An update. J Natl Cancer Inst 74:325–333. Tolbert P. 1997. Oils and cancer [review]. Cancer Causes Control 8:386–405. Tomatis L, Bartsch H. 1990. The contribution of experimental studies to risk assessment of carcinogenic agents in humans. Exp Pathol 40:251– 266. Tomatis L, Kaldor J, Bartsch H. 1996. Experimental studies in the assessment of human risk. In: Schottenfeld D, Fraumeni JF Jr, eds. Cancer Epidemiology and Prevention. New York: Oxford University Press, pp. 11–27. Toren K, Persson B, Wingren G. 1996. Health effects of working in pulp and paper mills: Malignant diseases [comment] [review] [48 refs]. Am J Ind Med 29:123–130. Tsuda T, Mino Y, Babazono A, Shigemi J, Otsu T, Yamamoto E. 2001. A casecontrol study of the relationships among silica exposure, gastric cancer, and esophageal cancer. Am J Ind Med 39:52–57. Upton A, Barrett J, Becklake Meal. 1991. Asbestos in Public and Commercial Buildings: A Literature Review and Synthesis of Current Knowledge. Report to: Health Effects Institute-Asbestos Research (HEI-AR). Cambridge, Ma: Health Effects Institute. US Department of Health and Human Services. 2003. EHP Online. Available at: http://ehp.niehs.nih.gov/roc/toc10.html. Accessed August 20, 2003. Van der Gulden J, Vogelzang P. 1996. Farmers at risk for prostate cancer. Br J Urol 77:6–14. Verougstraete V, Lison D, Hotz P. 2003. Cadmium, lung and prostate cancer: A systematic review of recent epidemiological data. J Toxicol Environ Health B Crit Rev 6:227–255. Vigliani E, Saita G. 1964. Benzene and leukemia. N Engl J Med 271:872–876. Vineis P, Malats N, Lang M, d’Errico A, Caporaso N, Cuzick J, Boffetta P. 1999. Metabolic Polymorphisms and Susceptibility to Cancer. Lyon: International Agency for Research on Cancer. Vineis P, Marinelli D, Autrup H, Brockmoller J, Cascorbi I, Daly A, Golka K, Okkels H, Risch A, Rothman N, Sim E, Taioli E. 2001. Current smoking, occupation, N-acetyltransferase-2 and bladder cancer: A pooled analysis of genotype-based studies. Cancer Epidemiol Biomarkers Prev 10:1249– 1252. Vineis P, Pirastu R. 1997. Aromatic amines and cancer [review]. Cancer Causes Control 8:346–355. Vineis P, Simonato L. 1991. Proportion of lung and bladder cancers in males resulting from occupation: A systematic approach. Arch Environ Health 46:6–15. Volkmann R. 1875. Ueber Theer-, Paraffin und Russkrebs (Schornsteinfegerkrebs). In: Beitrage zur Chirurgie. Leipzig: Druck und Verlag von Breitkopf und Hartel, pp. 370–381. Wagner J, Sleggs C, Marchand P. 1960. Diffuse pleural mesothelioma and asbestos exposure in North Western Cape Province. Br J Ind Med 17:260–271. Waldron A. 1983. A brief history of scrotal cancer. Br J Ind Med 40:390–401. Wang SQ, Setlow R, Berwick M, Polsky D, Marghoob AA, Kopf AW, Bart RS. 2001. Ultraviolet A and melanoma: a review. J Am Acad Dermatol 44:837–846. Ward E, Boffetta P, Andersen A, Colin D, Comba P, Deddens J, De Santis M, Engholm G, Hagmar L, Langard S, Lundberg I, McElvenny D, Pirastu R, Sali D, Simonato L. 2001. Update of the follow-up of mortality and cancer incidence among European workers employed in the vinyl chloride industry. Epidemiology 12:710–718.

354

PART III: THE CAUSES OF CANCER

Ward E, Burnett CA, Ruder A, Davis-King K. 1997. Industries and cancer. Cancer Causes Control 8:356–370. Ward E, Fajen J, Ruder A, Rinsky R, Halperin W, Fessler-Flesch C. 1996. Mortality study of workers employed in 1,3-butadiene production units identified from a large chemical workers cohort. Toxicology 113:157–168. Wartenberg D, Reyner D, Scott CS. 2000. Trichloroethylene and cancer: Epidemiologic evidence. Environ Health Perspect 108:161–176. Waters M, Stack H, Jackson M. 1999. Short-term tests for defining mutagenic carcinogens. IARC Sci Publ 499–536. Waxweiler R, Stringer W, Wagoner J, Jones J, Falk H, Carter C. 1976. Neoplastic risk among workers exposed to vinyl chloride. Ann N Y Acad Sci 271:40–48. Weeks J. 1998. Reducing risk of lung cancer from diesel exhaust in underground mines. Am J Ind Med 34:203–206. Weiderpass E, Partanen T, Kaaks R, Vainio H, Porta M, Kauppinen T, Ojajarvi A, Boffetta P, Malats N. 1998. Occurrence, trends and environmental etiology of pancreatic cancer [review]. Scand J Work Environ Health 24:165–174. Weiderpass E, Pukkala E, Vasama-Neuvonen K, Kauppinen T, Vainio H, Paakkulainen H, Boffetta P, Partanen T. 2001. Occupational exposures and cancers of the endometrium and cervix uteri in Finland. Am J Ind Med 39:572–580. Weihrauch M, Lehnert G, Kockerling F, Wittekind C, Tannapfel A. 2000. p53 mutation pattern in hepatocellular carcinoma in workers exposed to vinyl chloride. Cancer 88:1030–1036. Weill H, Mcdonald J. 1996. Occupational lung disease: 1. Exposure to crystalline silica and risk of lung cancer: the epidemiological evidence. Thorax 51:97–102. Weisburger J. 1999. Carcinogenicity and mutagenicity testing, then and now. Mutat Res 437:105–112. Weisenburger DD, Chiu BCH. 2002. Does asbestos exposure cause nonHodgkin’s lymphoma or related hematolymphoid cancers? A review of the epidemiologic literature. Clin Lymphoma 3:36–40. Werner J, Carter J. 1981. Mortality of United Kingdom acrylonitrile polymerisation workers. Br J Ind Med 38:247–253. Wesseling C, Pukkala E, Neuvonen K, Kauppinen T, Boffetta P, Partanen T. 2002. Cancer of the brain and nervous system and occupational exposures in Finnish women. J Occup Environ Med 44:663–668. WHO. 1964. Prevention of Cancer. Report of a WHO Committee, Technical Report Series 276. Geneva: World Health Organization. Wigle D, Mao Y, Howe G. 1982. Comparison of occupation on survey and death records. Can J Public Health 73:242–247. Wilbourn J, Haroun L, Heseltine E, Kaldor J, Partensky C, Vainio H. 1986. Response of experimental animals to human carcinogens: An analysis

based upon the IARC Monographs programme. Carcinogenesis 7:1853–1863. Wogan GN. 2000. Impacts of chemicals on liver cancer risk. Semin Cancer Biol 10:201–210. Wolff M, Collman G, Barrett J, Huff J. 1996. Breast cancer and environmental risk factors: epidemiological and experimental findings [review] [116 refs]. Annu Rev Pharmacol Toxicol 36:573–596. Wolff MS, Weston A. 1997. Breast cancer risk and environmental exposures [review] [76 refs]. Environ Health Perspect 105:891–896. Wong O. 2001. Malignant mesothelioma and asbestos exposure among auto mechanics: Appraisal of scientific evidence. Regul Toxicol Pharmacol 34:170–177. Wong O, Raabe G. 1995. Cell-type-specific leukemia analyses in a combined cohort of more than 208,000 petroleum workers in the United States and the United Kingdom, 1937–1989. Regul Toxicol Pharmacol 21:307–321. Wong O, Trent L, Whorton M. 1994. An updated cohort mortality study of workers exposed to styrene in the reinforced plastics and composites industry. Occup Environ Med 51:386–396. Wynder E, Gori G. 1977. Guest editorial: Contribution of the environment to cancer incidence: An epidemiologic exercise. J Natl Cancer Inst 58:825–832. Yamagiwa K, Ichikawa K. 1918. Experimental study of the pathogenesis of carcinoma. J Cancer Res 3:1–29. Yin S, Li G, Tain F, Fu Z, Jin C, Chen Y, Luo S, Ye P, Zhang J, Wang G, et al. 1987. Leukaemia in benzene workers: a retrospective cohort study. Br J Ind Med 44:124–128. Yu MC, Skipper PL, Tannenbaum SR, Chan KK, Ross RK. 2002. Arylamine exposures and bladder cancer risk. Mutat Res 506:21–28. Yu MC, Yuan JM. 2002. Epidemiology of nasopharyngeal carcinoma. Semin Cancer Biol 12:421–429. Zahm SH, Fraumeni JF. 1997. The epidemiology of soft tissue sarcoma [review]. Semin Oncol 24:504–514. Zahm SH, Ward MH, Blair A. 1997. Pesticides and cancer [review]. Occup Med 12:269–289. Zahm S, Blair A. 1993. Cancer among migrant and seasonal farmworkers— An epidemiologic review and research agenda. Am J Ind Med 24:753–766. Zeeb H, Blettner M. 1998. Adult leukaemia—What is the role of currently known risk factors [review]. Radiat Environ Biophys 36:217–228. Zeiger E. 1998. Identification of rodent carcinogens and noncarcinogens using genetic toxicity tests: Premises, promises, and performance. Regul Toxicol Pharmacol 28:85–95. Zenz C. 1988. Occupational Medicine: Principles and Practical Applications. Chicago: Year Book Medical Publishers.

19

Air Pollution JONATHAN M. SAMET AND AARON J. COHEN

I

n the early decades of the 20th century, lung cancer was a rare disease. By mid-century, however, it was evident that an epidemic of lung cancer was occurring among males in the United States and a number of European countries and a parallel epidemic soon followed among women. As this epidemic was first identified, clinicians caring for the patients offered two primary hypotheses to explain the rising frequency of lung cancer: the serious air pollution in many cities and the increasingly common use of cigarettes (Hoffman, 1929; Hoffman, 1931; Macklin and Macklin, 1940; Overholt and Rumel, 1940; Stocks and Campbell, 1955). The air pollution hypothesis was also supported by the generally higher rates of lung cancer in urban compared with rural areas. These two non-exclusive hypotheses were given equal weight by Doll and Hill (Doll and Hill, 1950a) as a rationale for their first case-control study of lung cancer in London. However, the initial formal epidemiological studies quickly indicted cigarette smoking as the predominant cause of the disease (Doll and Hill, 1950b; Wynder and Graham, 1950; Levin et al., 1950). Nonetheless, there have been persistent concerns that air pollution may cause lung cancer and other cancers. These concerns have been prompted by the release of carcinogens into outdoor air from industrial sources, power plants, and motor vehicles, and by the recognition that indoor air is also contaminated by respiratory carcinogens. Epidemiological studies have also continued to provide an indication that air pollution increases lung cancer risk. During recent decades, a number of airborne carcinogens, viewed as potential threats to public health, have been particularly controversial. The energy crisis of the early 1970s led to increased manufacture of diesel-powered vehicles; recognition that diesel-soot particles were mutagenic raised concern that the increasing numbers of dieselpowered vehicles would increase lung cancer risk for the population. Three indoor carcinogens received widespread attention from the scientific community and the public during the 1980s and 1990s: tobacco smoke inhaled by nonsmokers, radon, and asbestos fibers. Policies and programs for reducing the risks to the population of each of these carcinogens became extremely controversial with questioning as to whether the risks were exaggerated and if the high costs of control, in the case of radon and asbestos, were justified. The scientific evidence on secondhand smoke was repeatedly challenged by the tobacco industry as it sought to maintain controversy, even as review panels concluded that inhaling secondhand smoke does cause lung cancer (National Research Council (NRC) and Committee on Passive Smoking, 1986; International Agency for Research on Cancer (IARC), 1986). In the United States, there has also been extensive litigation around indoor asbestos. For these and other inhaled carcinogens, substantial research has been motivated by concern for the public’s health and by the need to develop the scientific foundation for control strategies. This chapter provides an overview of the evidence on outdoor and indoor air pollution and lung cancer, as well as other types of malignancy. The topic of secondhand smoke and lung cancer is also addressed in another chapter in this volume. The evidence on air pollution and lung cancer is now extensive and this review is selective, emphasizing the most recent findings, primarily from the epidemio-

The views expressed in this chapter are those of the authors and do not necessarily reflect the views of the Health Effects Institute (HEI) or its sponsors.

logic literature. A 1990 monograph provides a more complete review of the earlier literature on outdoor air pollution (Tomatis, 1990) and a 1994 book comprehensively covers the epidemiologic literature on lung cancer, including indoor and outdoor air pollution (Samet, 1994). This chapter focuses on outdoor, primarily urban, air pollution and indoor air pollution and lung cancer in developed countries. Mounting evidence indicates that people in developing countries may have exposures to indoor and outdoor environments that rival or even dramatically exceed those found in developed western countries (Smith, 2000). Indoor air pollution from coal combustion and cooking fumes has been linked to increased lung cancer risk in homes in China and Hong Kong (Smith and Liu, 1994). Rising pollution of outdoor air in the mega-cities of the developing world may also pose a risk for lung cancer, although the carcinogens in the air of these cities have not yet been well characterized. These topics are addressed elsewhere (Smith, 1988; Cohen et al., 2004).

PERSPECTIVE ON EXPOSURES TO INHALED CARCINOGENS Exposures to inhaled carcinogens that may cause lung cancer or other cancers take place in a variety of settings, including the home, the workplace, and other public and commercial locations, and outdoors. The concept of total personal exposure provides a useful framework for conceptualizing exposures to inhaled carcinogens and evaluating the contributions of outdoor and indoor air pollution (National Research Council (NRC) and Committee on Advances in Assessing Human Exposure to Airborne Pollutants, 1991). Total personal exposure represents the integrated exposure to an agent, as that exposure is received in multiple microenvironments (i.e., environments having a relatively homogeneous concentration of the agent of interest during a specified time period). For an inhaled carcinogen such as benzo-(a)pyrene, relevant microenvironments over the day might include outdoor air contaminated by vehicle exhaust and industrial emissions, workplace exposures from an industrial process, and air contaminated by tobacco smoke in a bar. For each microenvironment, the exposure received depends on the concentration of the carcinogen and the time spent in the microenvironment. Total personal exposure is the sum of the product of concentration with time spent in each microenvironment. The actual lung dose of the carcinogen will further depend on the exposed person’s ventilation rate and pattern, physical characteristics of the agent, and other factors determining the extent and site of lung deposition (National Research Council (NRC) and Panel on Dosimetric Assumptions Affecting the Application of Radon Risk Estimates, 1991). Doses to other organs (e.g., the urinary bladder) depend on subsequent uptake, metabolism, distribution, and elimination. The relevant microenvironments will vary with the carcinogen of concern. Because most time is typically spent at home, exposures to contaminants in indoor air of residences may be dominant for many pollutants, such as radon. Based on time-activity patterns, the workplace is also a significant locus of exposure, not only to specific occupational carcinogens but to other carcinogens, including secondhand smoke. Generally, adults spend little time outdoors but exposures to outdoor carcinogens could also take place indoors from the entry of outdoor pollutants into the air of a home.

355

356

PART III: THE CAUSES OF CANCER

The related concepts of total personal exposure and of microenvironments are also relevant to the development of control strategies. Some environments are shared and public and any initiative to control exposure requires action at a societal level; other environments are private and control lies with individuals. Thus, regulations may be needed to control air toxics released by industrial processes while education may be the cornerstone of initiatives to reduce exposures to carcinogens in the home.

Epidemiologic Approaches Epidemiologists have used three approaches to examine the role of air pollution, whether indoor or outdoor, in the etiology of lung cancer: cohort, case-control, and ecologic designs. Historically, some of the earliest studies of outdoor air pollution were based on ecological approaches, comparing lung cancer mortality rates in urban and rural areas (Stocks and Campbell, 1955). This approach has also been used subsequently in exploratory assessments of the association of indoor radon with lung cancer. However, the recent evidence comes largely from cohort and case-control studies, which have used more refined approaches for exposure assessment than earlier studies. The cohort studies of outdoor air pollution have involved the follow-up of populations, usually defined with respect to residence in an area or areas, over some time interval with assignment of exposure based on residence location. Cohort members are classified at the level of a geographic unit with respect to air pollution exposure and at the individual level with respect to other factors related to lung cancer occurrence, such as age and cigarette smoking. The case-control approach provides investigators with an efficient way to estimate the relative risk of lung cancer in relation to air pollution exposure without having to collect information on an entire population. In the casecontrol design, cases of lung cancer that have occurred in the population are ascertained and classified according to their exposure to outdoor or indoor air pollution; a sample of the study population— the controls—is selected and similarly classified according to their exposures. In contrast to the cohort and case-control designs, ecologic, or aggregate-level, studies do not collect information on individual subjects, but instead compare lung (or other cancer) incidence or mortality rates for geographic regions distinguished by different overall levels of air pollution, making use of routinely collected data on both lung cancer rates and region-wide measurements of air pollution. This design has also been used for testing the hypothesis that indoor radon exposure causes lung cancer (Stidley and Samet, 1993). In these studies, surrogate measures of exposure have been used based on geologic features, housing characteristics, or measurements of indoor concentrations. Relative risks can be estimated from ecologic data, but the interpretation of these estimates is more complicated than for estimates derived from cohort and case-control studies. In most cases, data

are not available for adequately taking into account inter-individual and between-region differences in other lung cancer risk factors. The findings of ecological studies regarding indoor radon have been particularly contentious, in part because of the difficulty of taking other factors into account. All three designs potentially suffer from a common problem when applied to the study of air pollution and lung cancer: the difficulty of characterizing accurately the subjects’ exposure to air pollution. As described above, an individual’s exposure to carcinogens in outdoor and indoor environments may be complex and occur in multiple microenvironments, and therefore it may be difficult to estimate exposure for the purpose of epidemiologic analysis. Some of the exposure assessment approaches used are listed in Table 19–1 (Samet and Jaakkola, 1999); these range from categorical indicators, such as selfreport and residence location, to the contemporary use of geographic information systems. The early lung cancer studies often classified exposure to outdoor air pollution based on having an urban or rural residence location, a crude indicator of current or lifetime exposure. In some subsequent studies, approaches based on residence location were refined by more fully capturing the residential history and incorporating duration of residence into exposure indexes. Some of the most recent investigations of outdoor air pollution have incorporated data from stationary monitoring sites for one or more air pollutants that are indicative of combustion emissions; generally, a measure of airborne particles has been used. Estimation of exposures has been further refined in several recent studies by using the monitoring data in combination with an air pollution model to assign concentrations to the residence locations of study participants. For example, in a cohort study in the Netherlands, Hoek et al. (2002) used a geographic information system to assign exposures to residence locations, based on monitoring data, and also on proximity to roadways. The prediction model used traffic-related variables, such as traffic density and distance from a roadway, to estimate population exposure (Brauer et al., 2003). There has also been extensive attention given to the development of approaches for estimating exposures to several indoor carcinogens, particularly secondhand smoke and radon, and to a lesser extent asbestos. While the association of secondhand smoke with lung cancer is not specifically covered in this chapter, the conceptual approaches to assessing exposure to secondhand smoke are covered because of their relevance to the topic of air pollution and cancer. For each of these agents, the validity of exposure assessment approaches and the degree of misclassification and resulting bias in risk estimates have figured prominently in the interpretation of the epidemiological findings. The microenvironmental model has been the foundation of exposure assessment for these agents. For radon, individual- and population-level exposures are largely driven by exposures at home, reflecting both the time spent at home and the higher concentrations in houses, compared with other types of buildings. The earliest studies

Table 19–1. Methods of Assessing Exposure Source of Information Source strength Geographical information Dispersion models Outdoor-indoor penetration Stationary monitoring Questionnaires and interviews Personal monitoring Human samples Toxicological models Source: Samet and Jaakkola (1999).

Type of Information Emission rate (mass per time), traffic density Distance of place of residence from the source Spatio-temporal concentration distributions from modeling of emission rates, meteorology, air chemistry, geography Modeling from outdoor concentration, building, and ventilation characteristics Concentration over time modeling from concentration of pollutants in microenvironments Source strength, distance from the source, time activity Continuous or cumulated concentrations over time Concentration of biomarkers or exposure in human tissue or hair Concentration and dose of pollutants in target organs modeling from concentration, breathing rate, metabolism

357

Air Pollution used characteristics of residences as surrogates for potential exposure to radon. The concentration of radon can be feasibly measured in homes, using passive and relatively inexpensive measuring devices to measure the average concentration across intervals ranging from days to a year. More recently, a technique for longer-term exposure estimation has been developed that uses levels of surface radioactivity in glass, for example, that covered a photo owned for many years and moved from home to home (Steck et al., 2002). Issues related to misclassification based on these approaches are considered in the section of this chapter on indoor radon. For secondhand smoke, the array of approaches has included qualitative assessments based on participants’ perceptions of being exposed, description of the numbers of smokers, and their patterns of smoking (i.e., a characterization of the strength of the source, measurements of atmospheric markers of secondhand smoke, and measurements of biomarkers) (Jaakkola and Samet, 1999). Multiple microenvironments are relevant, including the home, the workplace, public places and entertainment locations, and transportation environments. In investigating secondhand smoke and lung cancer, exposures across the full lifespan are potentially relevant and researchers face the challenge of reconstructing exposures based primarily on study participant recollection of smoking patterns in various microenvironments, extending across childhood and adulthood in some studies. As discussed elsewhere, the extent of misclassification with this approach and its consequences for characterizing the risks of lung cancer associated with secondhand smoke have been a focus for dispute between researchers and the tobacco industry and its consultants (Wu, 1999; Samet and Wang, 2000). Even the more refined approaches of recent studies produce exposure estimates that are unavoidably subject to misclassification. The resulting misclassification of exposure may produce biased results in studies of air pollution and lung cancer and create the misimpression that results from different studies are not in agreement (Lubin et al., 1995b; Rothman and Greenland, 1998; Lubin et al., 1995b). Such misclassification can spuriously elevate or diminish estimates of effects depending on how misclassification differs between subjects with and without lung cancer. When the extent of misclassification of exposure is the same for those with and without lung cancer (i.e., nondifferential misclassification), estimates of effect are, in most cases, attenuated. When the risk of lung cancer increases directly and monotonically with exposure, non-differential misclassification of exposure can obscure this pattern (Birkett, 1992; Dosemeci et al., 1990; Birkett, 1992). Most studies attempt to collect data on other lung cancer risk factors, such as cigarette smoking, which could confound the air pollution relative risks. Errors in the measurement of potential confounders can introduce bias, even if air pollution exposures are estimated with relatively little error, and the result may be either over- or underestimation of the air pollution effect (Greenland, 1980).

RISK ASSESSMENT The concerns about outdoor and indoor air pollution and lung cancer largely relate to the risk posed to populations, not individuals. Risk assessment, a framework for organizing information about risks to populations, has been applied to the problem of outdoor air pollution and lung cancer for policy-making purposes. In the United States, the Clean Air Act Amendments of 1990 require the agency to address a list of 189 hazardous air pollutants using a risk-based approach; many of these agents are carcinogens.

OUTDOOR AIR POLLUTION Exposures to Carcinogens in Outdoor Air Ambient air, particularly in densely populated urban environments, contains a variety of known human carcinogens, including organic compounds such as benzo [a] pyrene and benzene, inorganic compounds such as arsenic and chromium, and radionuclides (Table 19–2) (International Agency for Research on Cancer (IARC), 1987). These substances are present as components of complex mixtures, which may include carbon-based particles to which the organic compounds are adsorbed, oxidants such as ozone, and sulfuric acid in aerosol form. The combustion of fossil fuels for power generation or transportation is the source of most of the organic and inorganic compounds, oxidants, and acids, and contributes heavily to particulate air pollution in most urban settings. The radionuclides result from fuel combustion as well as from mining operations. The plausibility of outdoor air pollution as a cause of cancer is supported by studies that have addressed its genotoxicity using in vitro assay systems and biomarkers. Urban air samples have mutagenic activity in the Salmonella assay, with much of the activity attributable to polycyclic aromatic hydrocarbons (PAHs) (DeMarini et al., 1998). Incinerator emissions are also mutagenic; their activity reflects PAH content and nitroarenes (DeMarini et al., 1998). Recently, investigators have used biomarkers of exposure to ambient air pollutants to better characterize patterns of exposure. These studies have been primarily focused on polycyclic aromatic hydrocarbons and also markers of oxidative stress. In a 1992 study in Poland (Perera et al., 1992), Perrera and colleagues compared levels of markers of molecular and genetic damage in a highly industrialized and a comparison area. In residents of the more polluted area, levels of the PAH-DNA adducts were higher, as was the frequency of indicators of DNA damage. Autrup and colleagues (Autrup et al., 1999) investigated bulky carcinogen-DNA adducts and also 2-amino-apidic semialdehyde, a marker of oxidative stress, in nonsmoking Danish bus drivers and postal workers. Levels of adducts tended to be higher in the bus drivers in central Copenhagen but the markers of oxidative stress did not follow this pattern. In a study in Germany, markers of exposure to

Table 19–2. Selected Known Carcinogens in Urban and Rural Ambient Air Substance

Urban Air

Rural Air

3

Inorganic particulates (ng/m ) Arsenic Asbestos Chromium Nickel Radionuclides (Ci/m3) 210 Pb 212 Pb 222 Rn Gaseous and particulate organic species (ng/m3) Benzene Benzo[a]pyrene Benzene-soluble organics

2–130 10–100 5–120 10–1000

<0.5–5 — <1–10 <10

1 * 10-15 –30 * 10-15 0.1 * 10-15 –4 * 10-15 20 * 10-125 –1000 * 10-12

5.5 * 10-15–10 * 10-15 0.03 * 10-15 –0.06 * 10-15 0.1 * 10-12 –20 * 10-12

5–90 1–50 1000–2000

— — 200–300

Source: International Agency for Research on Cancer (IARC), 1987.

358

PART III: THE CAUSES OF CANCER

benzo-(a)-pyrene tended to be higher in city dwellers, compared with suburban dwellers. In the AULIS project in Greece, biomarkers of genotoxicity were assessed in residents of the city of Athens and of the rural town of Halkida (Kyrtopoulos et al., 2001; Georgiadis et al., 2001). The mean levels of bulky DNA adducts in lymphocytes were higher in Halkida residents than in residents of Athens, and in the Halkida residents increased levels were found for those exposed to secondhand smoke. The findings in relation to residence location were unanticipated, given prior studies tended to show higher levels of markers of genotoxicity in urban workers and residents (Kyrtopoulos et al., 2001). By contrast, in a study of students in Copenhagen, measured personal exposure to small particles was associated with 7-hydro8-oxo-2¢-deoxyguanosine, a marker of DNA damage in lymphocytes; other associations between exposure and biomarker levels were not found (Sorensen et al., 2003). The differences in findings may reflect the differing levels of exposure and the markers considered. Unfortunately, there are few long-term trend data for ambient levels of known carcinogenic products of fossil fuel combustion that could be used to estimate long-term exposures for epidemiologic purposes. Available data indicate that over recent decades there have been improvements in some indices of air quality in the United States and some other developed countries. The U.S. Environmental Protection Agency collects data on six pollutants for which the U.S. government has promulgated national air quality standards, including airborne particles and SO2. Particulate matter with an aerodynamic diameter <10 m, PM10, has been the criteria pollutant of greatest current interest with respect to lung cancer because particles of size 10 m or less can be inhaled into the lung and generally originate from combustion processes and may carry carcinogenic substances, such as polycylic aromatic hydrocarbons, on their surfaces. More recently, the U.S. Environmental Protection Agency has begun to monitor that component of PM10 having an aerodynamic diameter less than 2.5 m, PM2.5, which includes those particles penetrating mostly deeply into the lung’s small airways and alveoli. Comprehensive monitoring of PM10 has only been in place since 1988; before 1988, the U.S. Environmental Protection Agency monitored only total suspended particulates (TSP), which include particles too large to be inhaled into the lung. Decreases in TSP were observed over the 1970s and early 1980s, but little change was noted through 1990. From 1988 to 1992 the average annual mean concentration of PM10 fell by 17% (U.S. Environmental Protection Agency (EPA), 1993). The U.S. EPA reports a 13% decrease in PM10 for the period 1993 to 2002, and, since 1970, aggregate emissions of the six principal pollutants have been cut 48% (U.S. Environmental Protection Agency (EPA), 2003). In the eastern United States, SO2 emissions decreased approximately 33 percent from 1983 to 2002. Nationally, average SO2 ambient concentrations have been cut approximately 54% over the same period. Reductions in SO2 concentrations and emissions since 1990 are primarily due to controls implemented under EPA’s Acid Rain Program. Sulfate reductions since 1999 are partly responsible for some improvement in ambient fine particle concentrations, particularly in the southeastern United States (U.S. Environmental Protection Agency (EPA), 2003). The data discussed above refer for the most part to ambient air pollution over relatively large geographic areas. However, the exposure of human populations to carcinogens in ambient air may be the result of proximity to more localized sources (Levy et al., 2000) such as small businesses (e.g., automotive body or chrome plating shops), municipal facilities (e.g., waste incinerators), or areas with high vehicular traffic. Studies in Harlem show that diesel particle exposures to pedestrians can be substantial (Kinney et al., 2000) and that high school students are exposed to a variety of toxic air pollutants throughout the day (Kinney et al., 2002). Levy and colleagues collected spot air samples along sidewalks in a section of Boston with heavy diesel bus traffic, showing the existence of “hot spots” for exposure to benzo(a)-pyrene, a marker for combustion emissions (Levy et al., 2000).

Combustion Products As noted above, the combustion of fossil fuels for transportation and power generation contributes to the presence of many known or sus-

pected carcinogens in ambient air. A discussion of some of the potentially more significant pollutants in terms of exposure prevalence and/or lung carcinogenicity follows.

Polycyclic Organic Matter Polycyclic organic matter, or POM, as defined by the U.S. EPA in the Federal Clean Air Act comprises a large and varied class of chemical compounds, including polycyclic aromatic hydrocarbons (PAH), and nitro-PAHs, which are known carcinogens and mutagens (International Agency for Research on Cancer (IARC), 1987). The compounds that comprise POM have common chemical features (one or more benzene rings and a boiling point >100°C), and are found in both the particulate and gas phases of ambient air, depending on their exact chemical structure (e.g., those with >5 benzene rings tend to be associated with the particle phase). In addition to those compounds released directly into the environment by combustion processes, others are created from primary combustion products, such as those emitted by diesel engines, via chemical and photochemical reactions in the ambient environment (Greenberg, 1988; Natusch, 1978; Winer and Busby Jr., 1995; Natusch, 1978; Greenberg, 1988). Although the combustion of fossil fuels is a ubiquitous source of POM in the urban ambient environment, it is not the only source of human exposure to POM, and for some individuals it may not be the predominant source. Other human exposure to POM comes from inhaling wood and tobacco smoke, and from diet (e.g., from the consumption of grilled meat). Urban air contains a mixture of polycyclic organic compounds, but certain specific constituents, such as benzo-[a]-pyrene, have been extensively studied and are known to be carcinogenic. Benzo-[a]pyrene has been frequently used as a surrogate or marker for combustion source air pollution in epidemiologic studies and for risk assessment (see below). The literature on cancer risk in relation to occupational and environmental exposure to PAHs has been reviewed by Boffetta et al. (1997), who concluded that PAHs are associated with increased lung cancer risk in a variety of occupational settings, and with increased lung cancer risk in urban populations. Mixtures of polycyclic compounds encountered in occupational settings, such as cokeoven workers in the steel industry, and coal gasification workers (Doll et al., 1972; Redmond, 1983; Doll et al., 1972), also are known to cause increased occurrence of lung cancer in exposed workers (International Agency for Research on Cancer (IARC), 1984). The levels of POM encountered in the ambient urban environment, however, are substantially less than those encountered in heavily exposed occupational settings.

Particles Like POM, particulate air pollution is not a single entity, but rather a chemically and physically diverse group of pollutants, derived from sources as diverse as crystal dust and sea spray, and the combustion of diesel fuel (Koutrakis and Sioutas, 1996). Largely because carbonaceous particles produced by the combustion of fossil fuels are in the respirable range (generally <1.0 m in diameter) and known human carcinogens, such as PAHs, are adsorbed to their surfaces, attention has focused on combustion-source particles in urban air as potential lung carcinogens. However, evidence from animal experiments showed that rats exposed to high levels of relatively pure carbon particles developed lung tumors at the same rate as rats exposed to diesel exhaust particles, suggesting that sufficient concentrations of particles per se might, under some conditions, be carcinogenic (Mauderly, 1997). The relevance of these findings for humans is controversial as carcinogenicity was observed at doses in which lung clearance mechanisms were overwhelmed (Winer and Busby Jr., 1995). The combustion of fossil fuels for power generation and transportation produces gaseous pollutants, such as sulphur dioxide (SO2) and oxides of nitrogen (NOx), which are converted into fine particles in the atmosphere. Epidemiologic studies provide no consistent evidence of increased lung cancer risk from occupational exposure to SO2; however, IARC has classified strong sulphuric acid aerosol as a known human carcinogen based on epidemiologic findings of increased lung and laryngeal cancer in heavily exposed occupa-

Air Pollution tional groups (International Agency for Research on Cancer (IARC), 1992).

Diesel Diesel exhaust is a ubiquitous component of urban ambient air pollution throughout the world, although few studies have estimated its proportional contribution. In one of the few studies on the proportional contribution of diesel exhaust to ambient air pollution, Cass and Gray estimated that diesel exhaust contributed 7% of the fine particulate matter (<2 mm) in the Los Angeles air basin in 1982 (Cass and Gray, 1995). Based on evidence from animal experiments and epidemiologic studies of occupationally exposed groups, diesel exhaust is considered by IARC to be a probable human carcinogen (2A) (International Agency for Research on Cancer (IARC), 1989), although, as noted above, the mechanism by which exposure to diesel exhaust might produce lung cancer in humans remains to be determined. The evidence for the carcinogenicity of diesel exhaust has been extensively and repeatedly reviewed (International Agency for Research on Cancer (IARC), 1989), as discussed below.

Other The combustion of fossil fuels contributes known or suspected individual carcinogenic chemicals to urban ambient air. In addition to such known human lung carcinogens as arsenic, chromium, and nickel, several others are worthy of note, and are discussed briefly below.

Radionuclides. Alpha-emitting radionuclides can also be measured in outdoor air. These include isotopes of lead, radium, thorium, and uranium (Natusch, 1978). They are naturally present in fossil fuels and are emitted as byproducts of combustion. The contribution of this route of exposure to the radiation dose received by the general population is negligible (National Research Council (NRC) and Committee on the Biological Effects of Ionizing Radiation, 1990). 1,3-butadiene. 1,3-butadiene is a volatile organic compound that has been employed since the 1930s in the production of synthetic rubber, and industrial emissions contribute to its presence in some urban areas. However, 1,3-butadiene is also emitted in automotive exhaust, accounting for the preponderance of emissions in the United States, and for much of the human exposure as well. Emissions tend to be substantially greater for vehicles with more than two axles (Sapkota and Buckley, 2003). Levels of 1,3-butadiene in ambient air are generally in the range of 1–10 ppb, about 1000-fold less than occupational exposure levels (Health Effects Institute, 1993; International Agency for Research on Cancer (IARC), 1992; International Agency for Research on Cancer (IARC), 1992). 1,3-butadiene has been classified by IARC as a probable human carcinogen (2A) based largely on the results of animal experiments, which indicated increases in tumors at multiple sites, including the lung. Epidemiologic studies of occupationally exposed populations (rubber workers and butadiene monomer production workers) have consistently observed increases in hematopoietic cancers, but not cancers of the respiratory system (International Agency for Research on Cancer (IARC), 1992). Aldehydes. Various aldehydes classified as hazardous air pollutants by the U.S. EPA (e.g., formaldehyde and acetaldehyde) are present in urban ambient air due largely to the combustion of gasoline and diesel fuel (Health Effects Institute, 1993). Exposures to formaldehyde and acetaldehyde in outdoor air tend to be highly correlated. In the case of formaldehyde, outdoor concentrations have generally been observed to be in the range of 1–20 mg/m3, although under certain conditions (in heavy traffic or air pollution episodes) levels of 100 mg/m3 have been observed (International Agency for Research on Cancer (IARC), 1995). The combustion of alternative fuels such as methanol, and oxygenated fuels containing the additive MBTE, yields aldehydes (Health Effects Institute, 1993). Formaldehyde is also present in indoor air. Formaldehyde has been classified by IARC as a probable human carcinogen (2A) (International Agency for Research on Cancer

359

(IARC), 1995), based on evidence from animal experiments and epidemiologic studies in occupational groups of exposure-related excess nasal and naso-pharyngeal cancer. There is no consistent evidence that occupational exposure to formaldehyde is associated with increased lung cancer risk (International Agency for Research on Cancer (IARC), 1995).

Point Sources Residential proximity to industrial point sources of air pollution is a potential source of exposure to known or suspected carcinogens. Fossil fuel-fired (i.e., coal, oil, and natural gas) electrical power plants emit known or suspected carcinogens (Natusch, 1978), including metals such as chromium and nickel, radionuclides such as radon and uranium, and POM such as benzo [a] pyrene. Non-ferrous metal smelters emit inorganic arsenic and other metals, and sulphur dioxide (Pershagen et al., 1977). Municipal solid waste incinerators emit heavy metals (e.g., lead and cadmium), PAHs, organic compounds (such as dioxins), and acidic gases (World Health Organization, 1988). Unfortunately, these sources of air pollution are more frequently located in or near poor, working-class communities, whose residents may, for a variety of reasons, be more susceptible to the effects of these pollutants. A number of reports describing cancer rates around particular sources have been published, some addressing specific chemicals or disasters (e.g., dioxin in Seveso, Italy) (Bertazzi et al., 1993), and radiation from the U.S. nuclear reactor at Three Mile Island, Pennsylvania (Hatch et al., 1990; Hatch et al., 1991). Such reports are not considered in this chapter. In 1990, Pershagen (1990) reviewed a number of epidemiologic studies of lung cancer occurrence and residential proximity to industrial point sources of air pollution. Eleven studies estimated lung cancer risk associated with proximity to non-ferrous metal smelters. Of these, five ecologic studies observed relative risks in males between 1.2 and 2.0, but only one accounted for employment at the smelter itself, and data on smoking were not available. These studies did not consistently observe elevations in risk among women. Six case-control studies presented conflicting results; several showed no association associated with residential proximity, but did not account for either employment at the facility or smoking habits. Two studies that did account for these factors observed relative risks in males of 1.6 and 2.0. Ecologic studies of residential proximity to diverse industrial sources (e.g., petro-chemical plants and steel mills) have generally observed increased relative risks of lung cancer, but have been unable to control for level confounders at the individual level, such as cigarette smoking and employment at the industrial facility itself. In a model approach for addressing point sources, Elliott and colleagues (1996) carried out an ecologic study of cancer incidence among 14 million people living near 72 municipal solid waste incinerators in Great Britain. Cancer rates and residential proximity to the incinerators were measured at the level of postal code (which is roughly analogous to neighborhood); the relative risks (compared with national incidence rates) were adjusted for age, sex, geographic region, and an index of socioeconomic status. For several cancers (e.g., stomach, colo-rectal, liver, and lung) excess relative risk was inversely related to distance of the residence from the incinerator. Lung cancer relative risks (95% confidence intervals) were 1.08 (1.07, 1.09) and 1.06 (1.05, 1.07) for residence distance 0–3 km and 0–7.5 km, respectively. However, Elliott and colleagues also observed equal elevations in lung cancer risk in the areas proximal to the incinerators before the construction of the facilities, leading them to conclude that residual confounding by unmeasured characteristics of the postal codes accounted for the apparent associations with proximity to the incinerators.

Fibers Asbestos fibers are present in ambient air in contemporary rural and urban environments, and apparently have been present in the ambient environment for at least 10,000 years. The literature on levels of asbestos in ambient air was reviewed as part of a comprehensive report

360

PART III: THE CAUSES OF CANCER

on the health effects of asbestos in public buildings (Health Effects Institute et al., 1991). The median levels of asbestos fibers in rural environments in which there were no known natural sources of asbestos were on the order of 0.01–0.001 ng/m3, and few individual measurements exceeded 1 ng/m3. In urban environments, where there is both a greater prevalence of asbestos-containing materials and a higher frequency of release of fibers from sources such as building materials and vehicular brake linings, a considerably greater proportion of individual measurements exceeded 1 ng/m3 and the median levels ranged from 0.02–10 ng/m3 (Health Effects Institute et al., 1991). Epidemiologic studies of occupational groups, such as asbestos miners and asbestos textile production workers who were exposed to asbestos at concentrations several orders of magnitude greater than those cited above, have consistently observed increased rates of lung cancer. Based on a large body of experimental and epidemiologic evidence, asbestos is considered to be a known human carcinogen (Health Effects Institute et al., 1991). Lung cancer occurrence was not increased in relation to exposure at levels observed in the ambient urban environment in a study of chrysotile-asbestos mining regions (Camus et al., 1998).

EPIDEMIOLOGIC EVIDENCE ON OUTDOOR AIR POLLUTION AND CANCER Epidemiologic Studies of Ambient Air Pollution and Lung Cancer The first studies on lung cancer and air pollution provided comparisons of rates of lung cancer mortality in urban and rural areas. These studies were carried out at the time that the strong role of smoking as a cause of lung cancer was first appreciated and consequently, some of the studies attempted to take smoking into account (Table 19–3). Most studies found overall excesses of lung cancer mortality of about 30%–40% in the urban areas. The attribution of these excesses to differences in air quality was strengthened by evidence of urban/rural differences in ambient levels of carcinogens, such as benzo-[a]-pyrene, in urban areas under investigation and by the persistence of the urban excess after adjustment for cigarette smoking in some of the studies. In addition, the studies of Buell et al. (Buell et al., 1967), Haenszel et al. (Haenszel and Loveland, 1962; Haenszel and Taeuber, 1964; Haenszel and Taeuber, 1964), and Samet et al. (Samet et al., 1987) found that the effect of living in an urban area increased with longer duration of urban residence. Doll and Peto in their widely cited 1981 monograph (Doll and Peto, 1981), The Causes of Cancer, cast doubt on the causal role of air pollution because early research had not accounted for the pattern of an earlier age of starting to smoke among urban dwellers as cigarette smoking became increasingly prevalent in the early 20th century. However, Dean et al. (1978) controlled for age at beginning smoking and found that the urban-rural gradient persisted. Moreover, cancer incidence data assembled by the International Agency for Research on Cancer over the past decade continue to show some evidence of urban-rural differences, as patterns of smoking have become more uniform (International Agency for Research on Cancer (IARC), 1997). However, the so-called “urban factor” may reflect factors in addition to prolonged exposure to urban air pollution, including indoor air pollution, aspects of diet, and occupation. Ecologic studies have been carried out that compared lung cancer rates across urban areas with differing levels of air pollution (Table 19–4). These studies showed relative excesses of lung cancer in the more polluted areas of similar or slightly higher magnitude than the urban/rural studies. Taking advantage of a “natural experiment,” Archer (1990) analyzed respiratory cancer mortality in two Utah counties with very low smoking rates. The counties were similar in many respects, with low and nearly equal respiratory cancer mortality rates, until a steel mill constructed during World War II caused substantial increases in air pollution in one of the counties. Within about 15 years after the opening of the steel mill, there were higher rates in the nowpolluted county that have persisted. A third neighboring county, unaf-

fected by pollution from the steel mill, but with higher smoking rates had higher lung cancer rates than either of the other two counties, underscoring the comparatively strong effect of cigarette smoking on lung cancer risk. However, because incidence, exposure, and covariate data were all on the aggregate, or ecologic, level, it was not possible to account adequately for intra-individual and between-area differences in other risk factors. Several case-control and cohort studies used air pollution monitoring data to quantify spatial variation in long-term average exposure to air pollution and estimate the exposures of study subjects (Table 19–4). Most studies observed increases in lung cancer risk associated with air pollution exposure, characterized variously, after adjustment for age, smoking, and occupational exposure. The prospective cohort studies provide some of the strongest evidence, as they include information on potential confounding factors and include some direct measurements of air pollution. Most of the studies have been carried out in the United States, although a recent report found increased lung cancer risk associated with urban air pollution in a Dutch cohort (Hoek et al., 2002). Dockery et al. (1993) reported the results of a cohort study of 8111 adults living in six US cities and observed for 14–16 years through 1989. Lung cancer relative risks were estimated with respect to average levels of PM2.5 (particulate matter <2.5 mm in aerodynamic diameter) and several other pollutants in each city. The researchers observed a 26% increase in all-cause mortality risk and a 37% excess lung cancer risk for a difference in fine particulate matter, PM2.5, equal to the difference between the most polluted and the least polluted city, after adjustment for differences in age, sex, cigarette smoking, obesity, and education. The American Cancer Society (ACS) Study (Pope III et al., 1995; Pope III et al., 2002) is the largest cohort study of air pollution and lung cancer. The air pollution component has been drawn from the ACS Cancer Prevention Study (CPS) II, involving an ongoing prospective cohort of approximately 1.2 million US adults most recently observed through 1998, with individual data on risk factors for cancer and other chronic diseases. In the air pollution component, residence locations of approximately 500,000 cohort members have been linked with air pollution data for metropolitan areas throughout the United States. Exposure was assigned as the annual average pollution concentration in the city-of-residence. For a 10-mg/m3 change in the ambient PM2.5 concentration, the investigators reported relative risks from 1.04–1.06 for all-cause mortality, and 1.08–1.14 for lung cancer, depending on the exposure data that were included. The excess relative risks associated with air pollution were comparable for neversmokers and smokers and for women and men. In most of the cohort studies, the exposures have been assigned based on air pollution data at one point in time or averages over extended periods. Consequently, most analyses have not addressed the exposure-time-response relationship of lung cancer risk with air pollution exposure. Recent re-analysis of both Six-Cities and ACS cohorts suggests that more recent exposures may have the strongest effects on all-cause mortality (Krewski et al., 2000). The reanalysis also found that in both cohorts the level of attained education was inversely related to the air pollution relative risk. The AHSMOG study (Mills et al., 1991) followed a cohort of 6000 Seventh Day Adventists in Southern California, initially from 1977–1982, and most recently through 1992 (Abbey et al., 1995). The prevalence of cigarette smoking was very low and dietary patterns were uniform and relatively healthy. Exposure to PM10 (particulate matter <10 mm in aerodynamic diameter) was associated with lung cancer mortality in males exposed to ≥43 days per year with PM10 concentration >100 mg/m3 (RR = 2.38). Ozone was associated with lung cancer at exposures >100 parts per billion for ≥551 hours /year (RR = 4.19). More recent analyses of AHSMOG sub-cohorts observed from 1977 by Beeson et al. (Beeson et al., 1998) and McDonnell et al. (McDonnell et al., 2000) show elevations in lung cancer risk in men associated with exposures to ozone and PM10, and estimated PM2.5 exposures, respectively. The current evidence suggests that lung cancer attributable to air pollution may occur among both smokers and nonsmokers, and, there-

361

Air Pollution Table 19–3. Urban/Rural Differences in Lung Cancer Risk Study

Locale/Time

Exposure Classification

England and Wales 1952–1954 710 male cases 12,000 hospital controls United States 1958–1959 2191 white male cases 31,516 white male controls United States 1958–1959 683 white female cases 34,339 white female controls Northern Ireland 1960–1962 1040 cases 12,932 controls New Mexico 1980–1982 504 cases 774 controls

Nonsmokers: Mortality ratio of urban residence compared with rural residence Mortality ratio of urban residence compared with rural residence Mortality ratio of urban residence compared with rural residence For nonsmokers: living in Belfast compared with living in “truly rural districts For cigarette smokers: ≥25 years residence (Hispanic) or ≥50 years residence (nonHispanic) in counties with ≥50,000 residents with at least 75% urban residents compared with 0 years residence Adult residence in town/city compared with country/small town Residence in industrial area compared with residential area Residence in center of city compared with residential area Urban residence compared with rural

Risk Estimate

Comments

case-control Stocks and Campbell, 1955 Haenszel et al., 1962 Haenszel and Taeuber, 1964 Dean, 1966 Samet et al., 1987

Holowaty et al., 1991 Barbone et al., 1995

Sasco et al., 2002 Zatloukal et al., 2003

Ontario, Canada 1983–1985 51 female cases 45 female controls Italy 1979–1981; 1985–1986 755 cases 755 controls

Morocco 1996–1998 118 cases 235 controls Czech Republic 1998–2002 145 adeno-carcinoma cases

Residence in a city of >100,000 compared with residence in a city of <100,000

221 other lung cancer cases 1624 controls

9.35 (NS)

Age-standardized by the age distribution of Liverpool

1.43 (NS)

Adjusted for age and smoking

1.27 (NS)

Adjusted for age and smoking

Men 3.80 (NS)

Age-standardized by the age distribution of Northern Ireland

Women 1.55 (NS) Hispanic 0.9 (0.4 –2.0) Non-Hispanic 1.0 (0.6–2.0)

Adjusted for smoking habit

2.3 (0.49–13.7)

Adjusted for smoking

1.4 (1.0–2.1)

Adjusted for age, cigarette smoking, occupational exposure, and SES

1.5 (1.0–2.2) 0.92 (NS)

No adjustments

Adenocarcinoma 0.44(0.31–0.62)

Adjusted for age

Other lung cancers 0.65 (0.48–0.88)

cohort Hammond and Horn, 1958

United States 1952–1955 187,783 white men

Buell et al., 1967

California 1959–1962 69,868 men

Hammond, 1972

United States 1959–1965 >1,000,000 adults

Cederlof et al., 1975

Sweden 1960–1970 55,000 adults

Residence in city 10,000– 50,000 compared with suburb or town residence Residence in city >50,000 compared with suburb or town residence Residence in Los Angeles County compared with residence in all other CA counties, except the San Francisco Bay area and San Diego County For men, not occupationally exposed to dust, fumes, etc.: Observed/Expected deaths For nonsmokers: Town residence compared with rural residence

3.12 (NS)

Adjusted for age No lung cancer deaths in rural never-smokers

1.97 (NS) 1.26 (NS)

Adjusted for age and cigarette smoking

Metropolitan: 0.98 Rural: 0.89

Adjusted for age and smoking

Men: 0.6 Women: 0.6

Adjusted for age

NS, not significant; PAH, polycyclic aromatic hydrocarbons; SES, socioeconomic status; TSP, total suspended particulates.

fore, both residual confounding and effect modification of the air pollution relative risk due to cigarette smoking must be considered in interpreting this evidence. Most studies report air pollution relative risks adjusted for cigarette smoking, but the adjustment may not have controlled completely for potential confounding. Most cohort studies have information on cigarette smoking only at the beginning of followup. The possibility that changes in tobacco use were correlated with air pollution exposure cannot be excluded as patterns of smoking cessation have varied geographically in the United States. On the other hand, the association of lung cancer with air pollution was largely

unaffected in the Six-Cities study (Dockery et al., 1993), when longitudinal information on cigarette smoking was used in a recent reanalysis (Krewski et al., 2000), and several case-control studies have adjusted for smoking using time-varying information and found increased risk for air pollution exposure (e.g., Nyberg et al. 2000). As discussed above, several studies show increased lung cancer risks among self-reported never-smokers, but the numbers in any single study are very small and the effect estimates imprecise. The joint effect of smoking and air pollution cannot be estimated with high precision for the same reason.

Table 19–4. Epidemiologic Studies of Outdoor Air Pollution and Lung Cancer Study

Locale/Time

Exposure Classification

Risk Estimate

Comments

ecologic Henderson et al., 1975

Buffler et al., 1988 Archer, 1990

Los Angeles, CA 1968– 1970, 1972 County death certificates (7722 lung cancer deaths) Texas 1979–1981 1300 white males Utah 1980–1987 County mortality statistics

Address in a higher PAH area compared with address in a lower PAH area

1.06 (NS)

Adjusted for 1970 age distribution

TSP, 1969–1971 (per mg/m3)

1.04 (NS)

Adjusted for age, smoking, and SES

County with mean TSP of 63 mg/m3 from 1980–1985 compared with county with mean TSP of 40 mg/m3 from 1980–1986

1.60* (NS)

Adjusted for 1980 age distribution

Residence in an area with “higher” air pollution compared with a “lower” area, as determined by Be, Ti, Fe, Cu, Dr, Ni, SO2, dust fall, benzo[a]pyrene, and suspended matter levels High air pollution at residence compared with intermediate air pollution, as determined by smoke, SO2, insoluble deposits, and ferric oxide levels >50 years residence in areas of medium or high air pollution compared with 0–29 years of residence TSP > 150 mg/m3 and SO2 > 104 mg/m3 compared with lower levels (for men) TSP > 150 mg/m3 and/or SO2 > 104 mg/m3 compared with lower levels (for women) Level of particulate deposition: >0.298 g/m2/day compared with <0.175 g/m2/day

Men 1.55 (NS) Women 1.04 (NS)

No adjustments

1.69 (NS)

Adjusted for age and smoking

1.58 (1.09–2.29)

Adjusted for age and occupation

Men 1.06 (1.06–1.99) Women 1.17 (0.70–1.96)

Adjusted for age

1.4 (1.1–1.8)

Adjusted for age, cigarette smoking, occupational exposure, and SES

SO2 (per 10 mg/m3) NO2 (per 10 mg/m3)

1.01 (0.98–1.03) 1.10 (0.97–1.23)

Adjusted for age, selection year, smoking, radon, SES, occupational exposures, and employment in high-risk occupations

Per 1000 hours/year in which the annual concentration of TSP is greater than 200 mg/m3 Residence in city with high live level of fine particles compared with residence in city with low level TSP of postal code above 200 mg/m3 for at least 42 days of the year compared with postal codes that did not meet those TSP conditions Residence in the most polluted (fine particles) areas compared with the least polluted

1.72* (0.81–3.65)

Adjusted for gender, education, and years of past smoking

1.37 (0.81–2.31)

Adjusted for age, gender, smoking, education, and body mass index

1.72* (0.81–3.65)

Adjusted for age, education, and years of past smoking

1.03 (0.80–1.33)

Adjusted for age, gender, race, cigarette smoking, exposure to passive cigarette smoke, body mass index, alcohol consumption, education, and occupational exposure Adjusted for smoking, alcohol consumption, and education

case-control Hitosugi, 1968

Japan 1958–1965 216 cases 4500 controls

Dean et al., 1978

England 1963–1972 616 cases 10,025 controls

Vena, 1982

New York State 1957–1965 417 white male cases 752 controls

Jedrychowski et al., 1990

Cracow, Poland 1980–1985 1099 adult cases 1673 adult controls

Barbone et al., 1995

Italy 1979–1981; 1985– 1986 755 male cases 755 controls Sweden 1985–1990 1042 cases 2364 controls

Nyberg et al., 2000

cohort Mills et al., 1991

California 1977–1982 6000 nonsmoking Seventh-Day Adventists

Dockery et al., 1993

6 US Cities 1979–1989 8111 adults

Abbey et al., 1995

California 1977–1987 6340 nonsmoking Seventh-Day Adventists

Pope III et al., 1995

151 US Cities (CPS II) 1982–1989 552,138 adults

Beeson et al., 1998

California 1977–1992 6338 nonsmoking Seventh-Day Adventists

McDonnell et al., 2000

362

California 1977–1992 6338 nonsmoking Seventh-Day Adventists

For men: Risk associated with an interquartile range increase in the following: 100 ppb ozone: Mean concentrations of PM10: Mean concentrations of SO2: For men: Per 24.3 mg/m3 increase in PM2.5:

3.56 (1.35–9.42) 5.21 (1.94–13.99) 2.14 (1.36–3.37) 2.23 (0.56–8.94)

Adjusted for smoking, alcohol consumption, and education

363

Air Pollution Table 19–4. (cont.) Study

Locale/Time

Exposure Classification

Risk Estimate

Hoek et al., 2002

The Netherlands 1986– 1994 5000 adults

Per 30 mg/m3 increase in NO2

1.81† (0.98–3.34)

Pope III et al., 2002

151 US Cities (CPS II) 1982–1998 ª500,000 adults

Per 10 mg/m3 change in fine particles measuring <2.5 mm in diameter

1.14 (1.04–1.23)

Nafstad et al., 2003

Oslo, Norway 1974–1998 16,209 adult men

Per 10 mg/m3 increase in nitrogen oxide (NOx) Per 10 mg/m3 increase in SO2

1.10 (1.043–1.17)

Comments Adjusted for age, gender, education, Quetelet index, occupation, active and passive cigarette smoking, and neighborhood socioeconomic status score Adjusted for age, gender, race, smoking, education, marital status, body mass, alcohol consumption, occupational exposure, and diet Adjusted for tobacco smoke exposure, education, age, and NOx or SO2

0.96 (0.88–1.04)

*Risk estimate for “respiratory cancers”; †Risk estimate for “cardiopulmonary mortality”. NS, not significant; PAH, polycyclic aromatic hydrocarbons; TSP, total suspended particulates.

Limitations of approaches to exposure estimation also contribute to uncertainty in risk estimates. The ACS and Six-Cities studies estimated the exposure of each participant based solely on long-term average concentrations in their metropolitan area of residence. This approach may accurately reflect exposure to pollutants that are distributed homogenously over large areas for several decades, but if exposure at finer spatial and temporal scales is relevant, exposure misclassification is unavoidable. Newer European and North American studies are now incorporating spatial statistical methods to estimate individual long-term exposure histories, linking residential histories, measurements of traffic density on nearby streets, and long-term records of specific air pollutants. With the resulting estimates, the effect of exposure can be estimated with consideration of exposure variation in time and space and perhaps less misclassification (Nyberg et al., 2000; Reynolds et al., 2001; Hoek et al., 2002). Hoek et al. observed larger relative effects on mortality from cardiopulmonary diseases from air pollution in proximity to major roads than from larger-scale urban and regional air pollution (Hoek et al., 2002), and Nyberg et al. estimated the highest lung cancer relative risks for exposure 20 years or more before diagnosis (Nyberg et al., 2000). By providing exposure estimates at the individual level these studies also reduce the possibility of aggregate-level (ecologic) bias.

Risk Assessment project (World Health Organization, 2002; Cohen et al., 2004). Outdoor air pollution, characterized as fine particulate matter, PM2.5, was estimated to be responsible for about 0.80 million (1.2% of world total) premature deaths and 6.4 million (0.5% of world total) Years of Life Lost (YLL) in the populations of the world’s large cities (>100,000). This burden is borne predominantly by developing countries, with 30% of attributable YLLs occurring in the Western Pacific Region, including China, and 19% in Southeast Asian Region, including India. Exposure to air pollution was estimated to contribute to 62,000 (10,000, 114,000 95% uncertainty interval) lung cancer deaths and 576,000 (92,000, 1 million 95% uncertainty interval) YLL per year worldwide. Sixty-two percent (62%) of this burden was estimated to occur in China and India. In Chinese cities, where air pollution levels are many-fold greater than those in the cities of the developed West, outdoor air pollution may contribute to as much as 9.7% of lung cancer overall, and perhaps a larger proportion in non-smoking women (Cohen et al., 2004). Unfortunately, lacking suitable studies in developing countries, these estimates were based on extrapolating the relative risk estimates from the US American Cancer Society study (Pope et al., 2002) to China, India, and other settings where differences in health status and the air pollution mixture introduce large uncertainties.

Risk Attribution

Diesel Exhaust and Cancer of the Lung and Bladder

Estimates of the population—attributable risk of lung cancer due to outdoor air pollution have been based on markedly different methods and their range spans an order of magnitude. Basing their estimate on past and then current estimates of benzo-[a]-pyrene in urban air and extrapolation from occupational studies of PAH-exposed workers, Doll and Peto (Doll and Peto, 1981) estimated that less than 1% of future lung cancer would be due to air pollution from the burning of fossil fuels. They did note, however, that perhaps 10% of then current lung cancer in large cities might have been due to air pollution. In 1990, the US Environmental Protection Agency (US Environmental Protection Agency (EPA), 1990) estimated that 0.2% of all cancer, and probably less than 1% of lung cancer, could be attributed to air pollution. This estimate was obtained by applying the unit risks for over 20 known or suspected human carcinogens found in outdoor air to estimates of the ambient concentrations and numbers of persons potentially exposed. Karch and Schneiderman (1981), using data from the American Cancer Society (CPS-I) study and US Census data, estimated that the “urban factor” accounted for 12% of lung cancer in 1980. They predicted that 1980 levels of total suspended particulates would be associated with a lung cancer rate ratio of 1.3, slightly less than the 47% increase observed for total suspended particles in the recent report of findings in the Six-Cities Study. The worldwide health impact of urban outdoor air pollution, including its impact on lung cancer, was recently estimated as part the World Health Organization’s (WHO) Global Burden of Disease Comparative

The evidence for the carcinogenicity of diesel exhaust has been extensively and repeatedly reviewed, in part because of the implications for regulation and transportation management. The evidence comes from studies of worker groups and not the general population for which exposure cannot yet be assessed (International Agency for Research on Cancer (IARC), 1989). Current epidemiologic evidence indicates that workers exposed to diesel exhaust for prolonged periods in a variety of occupational settings are at increased risk of lung cancer. Recent quantitative reviews of this literature estimate summary excess relative risks of 20%–50%, which cannot be explained by differences between exposed and nonexposed workers in cigarette smoking or other known causes of lung cancer. Occupational exposure to diesel exhaust has also been linked to increased risk of bladder cancer, but the evidence is less consistent. Interpretation of the degree of excess risk, not to mention quantitative risk assessment, is difficult because few of the studies provide quantitative estimates of the levels of diesel exhaust particulate or other constituents to which workers were historically exposed. More than 40 studies currently provide assessments of the risk of lung cancer associated with occupational exposure to diesel exhaust (Cohen and Higgins, 1995; Boffetta et al., 1997; Bhatia et al., 1998; Lipsett and Campleman, 1999). The review by Lipsett and Campleman (1999) provides meta-analytic summary estimates of results of 31 studies grouped according to occupational group and key features of study design (e.g., control for smoking and study population).

364

PART III: THE CAUSES OF CANCER Menck Hall Leupker Ahlberg

Siemiatycki Boffetta Hayes Boffetta Steenland Burns Williams 0.1

1

10

Relative Risk (95% CI) Figure 19–1. Relative risk of lung cancer in railroad workers. (Source: Cohen and Nikula, 1999.)

The most frequently studied occupational groups have been railroad workers and truck drivers. The studies of truck drivers show a 20%–50% excess incidence and/or mortality from lung cancer, which persists when cigarette smoking is accounted for in data analysis (Fig. 19–1). Some studies included only small numbers of subjects and, therefore offer highly imprecise estimates of the relative risk. However, the upper bounds of the 95% confidence intervals indicate that few of the results are consistent with more than a tripling of risk and the larger studies (e.g., (Steenland et al., 1990)) are consistent with less than a doubling. The summary estimate of the relative risk for occupation as a truck driver derived by Lipsett and Campleman from nine studies was 1.47 (95% CI: 1.33,1.63) (Lipsett and Campleman, 1999). Studies of railroad workers have also consistently observed excess relative risks on the order of 30%–50%, comparable in magnitude to the estimates from the studies of truck drivers after analytic control for cigarette smoking (Fig. 19–2). For railroad workers, the estimate from the largest study (Garshick et al., 1988), is consistent with a doubling of risk at most. Lipsett and Campleman estimated a summary relative risk of 1.45 (1.08, 1.93) from an analysis of six studies (Lipsett and Campleman, 1999). The possibility of confounding by cigarette smoking has been a major concern with respect to the interpretation of these studies.

Howe Hall Garshick Boffeta Schenker Garshick Burns Siemiatycki Williams Boffetta 0.1

1

10

Relative Risk (95% CI) Figure 19–2. Relative risk of lung cancer in truck drivers. (Source: Cohen and Nikula, 1999.)

Although confounding by factors such as asbestos exposure, social class, diet, and exposure to particulate matter from sources other than diesel exhaust may have affected the results of individual studies (Nauss and The Diesel Working Group, 1995; Cohen and Higgins, 1995), confounding by cigarette smoking is the most plausible source of bias that might explain the consistent observation of small relative increases in lung cancer risk in several occupations and in studies of widely varying design. Cohen and Higgins (1995) and Larkin et al. (Larkin et al., 2000) used data on the relative risk of smoking and lung cancer and the plausible distributions of smoking prevalence in occupational groups to calculate the likely impact of confounding by cigarette smoking in the large railroad workers’ cohort study (Garshick et al., 1988), and concluded that uncontrolled confounding by cigarette smoking could not explain the elevated relative risks. Lipsett and Campleman estimated a summary smoking-adjusted relative risk of 1.47 (95% CI: 1.29, 1.67) among occupational groups considered to have substantial exposure to diesel exhaust (Lipsett and Campleman, 1999). Cohen and Higgins (1995) reviewed the evidence for increased risk of cancers other than lung. For the most part, the epidemiologic studies that provided consistent evidence of increased risk of lung cancer did not show increased risk for cancer at any other site, although associations have been reported for various other cancers. The only possible exception is for bladder cancer, for which numerous case-control studies found increased risk associated with exposure to diesel exhaust, particularly among truck drivers. Cohort studies, however, have found less consistent results in other diesel-exposed occupations such as railroad workers (Table 19–5). Cohen and Higgins considered several possible explanations for this apparent discrepancy, including statistical imprecision, exposure misclassification, and confounding of the bladder cancer relative risk in truck drivers by factors related to urinary stasis due to delayed urination. Based on the epidemiologic and toxicologic evidence diesel exhaust is considered by IARC to be a “probable human carcinogen” (2A)(International Agency for Research on Cancer (IARC), 1989). The IARC reviewers judged the evidence from animal data “sufficient” to conclude that diesel exhaust was a human carcinogen, but considered that the epidemiologic data were “limited” in this regard. More recently, the World Health Organization (WHO) evaluated the currently available data and reached the same general conclusions (World Health Organization, 1996). However, they added that no human data were “suitable for estimating unit risk.” Several US regulatory agencies have recently estimated the risk of lung cancer attributable to exposure to diesel exhaust. Both the EPA (US Environmental Protection Agency (EPA) and National Center for Environmental Assessment, 2002) and the California EPA (California Environmental Protection Agency (Cal EPA), 1998) rejected the use of the available animal bioassays for quantitative risk assessment because of concerns that the mechanisms that produced lung cancer in rats at high doses were not relevant in humans and based their risk estimates on occupational epidemiology studies, after concluding that the available evidence is consistent with a causal relationship between occupational exposure and lung cancer risk. California EPA calculated a unit risk value (California Environmental Protection Agency (Cal EPA), 1998), which ranged from 1 to 24 excess deaths in 10,000 people per 1 mg/m3 diesel PM lifetime exposure, and the USEPA estimated a possible range of lung cancer risk from environmental exposure to diesel exhaust (0.1 to 10 excess deaths in 10,000 people per 1 mg/m3 diesel PM lifetime exposure). The USEPA, however, acknowledges recent improvements in diesel technology, and uncertainty as to the relevance of these risk estimates to present engines (US Environmental Protection Agency (EPA) and National Center for Environmental Assessment, 2002). The Health Effects Institute (HEI), a non-profit research institute funded equally by government and industry, conducted an independent assessment of the uses of current epidemiologic studies for quantitative risk assessment of diesel exhaust. HEI critically evaluated several studies that provided some quantitative estimates of exposure to diesel exhaust and had been, or were being, considered as the basis for quantitative risk assessment: studies of railroad workers (Garshick

Table 19–5. Bladder Cancer Reference

Exposure

Howe et al., 1980

Occupational exposure to diesel and traffic fumes

Gottlieb and Pickle, 1981

Occupation as transport operative

Howe et al., 1983

Job title within railway work

Rushton et al., 1983

Occupational exposures

Silverman et al., 1983

Occupation: Industry— trucking Occupation title: Truck driver

Schenker et al., 1984

Occupational diesel exposure

Schoenberg et al., 1984

Occupation: Garage and/or gas station worker Motor vehicle mechanic Driver and/or deliveryman

Population/Location CASE-CONTROL STUDY British Columbia, Nova Scotia, and Newfoundland, Canada 1974–1976 CASES: 480 men with incident bladder cancer CONTROLS: 480 male population controls, matched on neighborhood and age CASE-CONTROL STUDY Louisiana 1968–1975 CASES: 237 men who died from bladder cancer CONTROLS: 237 men [dead] who did not die from bladder cancer, matched on race, parish, and age COHORT STUDY Canada 1965–1977 43,826 male Canadian National Railway pensioners who retired before 1965 Outcome: 175 deaths due to bladder cancer COHORT England 1967–1975 Maintenance men employed for at least 1 year at London Transport garages and Chiswick Works Outcome: 12 deaths due to bladder cancer CASE-CONTROL STUDY Conducted as part of the National Bladder Cancer Study Detroit metropolitan area 1977–1978 CASES: 303 white males with transitional or squamous cell carcinoma of the lower urinary tract CONTROLS: 296 white male population controls COHORT United States 1967–1979 2519 white men with at least 10 years of railway service by 1967 Outcome: 3 bladder cancer deaths CASE-CONTROL STUDY New Jersey 1978–1979 CASES: 658 men aged 21– 84 with incident carcinoma of the urinary bladder CONTROLS: 1258 male population

Findings

Comments

Never exposed: OR: 1.0 [reference] Ever exposed: OR: 2.8 [0.8–11.8]

Adjusted for smoking ORs were calculated using logistic regression 152 female case-control pairs were also in the study, but their exposure to diesel and traffic fumes was not reported

Never transport operative: OR: 1.00 [reference] Ever transport operative: ORwhite men: 2.06 ORblack men: 3.10

Crude ORs reported 110 female case-control pairs were also in the study, but their occupational exposure information was not reported

SMRall employees: 1.03 SMRfreight loader: 3.40

It was never explicitly stated what population was used to establish the expected mortality rates; however, the rates were standardized by calendar year of observation and age at observation The mortality observed in this cohort was compared with the expected mortality for the male population of England and Wales

Observed/expected deaths: 1.39 (P = 0.16)

Industry—Trucking: Never employed: ORcrude: 1.0 [reference] Ever employed: ORcrude: 2.2 [1.1–4.4] Occupation title of “truck driver”: Never: ORcrude: 1.0 [reference] Ever: ORcrude: [1.4–4.4]

Observed/expected deaths: 0.76 [0.15–2.21]

Never employed in a “highrisk” job: OR: 1.00 [reference] Ever employed as a garage and/or gas station worker: OR: 2.35 [1.47–3.78] Ever employed as a motor vehicle mechanic: OR: 1.26 [0.87–1.84] Ever employed as a driver and/or deliveryman: OR: 1.16 [0.93–1.46]

ORs were adjusted for age and smoking (except when noted): “The effect of potential confounding by age, smoking, or employment in other high-risk industries or occupations was controlled by stratification in the analysis.” Confidence intervals were not provided for all risk estimates Expected deaths were calculated by multiplying the person-years of survival in 5-year age-time groups by the age-time cause-specific mortality rates for US white men Adjusted for other employment categories, age, and duration of cigarette smoking ORs were calculated using logistic regression

(continued)

365

Table 19–5. (cont.) Reference

Exposure

Population/Location

Findings

Comments

Hoar and Hoover, 1985

Truck driving: Duration of truck driving in years

Duration of truck driving: Never: OR: 1.0 [reference] 1–4 years: OR: 1.4 [0.6–3.3] 5–9 years: OR: 2.9 [1.2–6.7] >10 years: OR: 1.8 [0.8–4.1] P value for trend: 0.006 (one-tailed)

OR calculated using logistic regression “When necessary, the effects of county of residence and other confounding variables, such as cigarette smoking, were taken into account by stratification.” It is unclear when and what adjustments were done

Smith et al., 1985

Occupation: Auto or truck mechanics, Chemically related occupation Stratified by smoking

CASE-CONTROL STUDY New Hampshire and Vermont 1975–1979 CASES: 325 white NH or VT residents with deaths due to bladder cancer CONTROLS: 673 people who died during the study period (two separate control series were combined [number of controls from each series not provided]: one series matched on state, sex, race, age, and year of death, and the other series additionally matched on county of residence) NESTED CASE-CONTROL STUDY New Jersey 1977–1978 Participants in the NCI National Bladder Cancer Study CASES: 2008 men with incident, histologically confirmed carcinoma of the urinary bladder CONTROLS: 4046 men, frequency-matched on age and sex within each geographic area

RRs were calculated by modeling (the exact method is not stated) RR adjusted for age, education, and coffee drinking

Wong et al., 1985

Occupations with potential exposure to diesel exhaust

NONSMOKERS: Never exposed: RR: 1.0 [reference] Ever a mechanic for at least 6 months: RR: 1.33 [0.77– 2.31] Ever had a chemically related occupation for at least 6 months: RR: 1.53 [1.13–2.07] SMOKERS: Never exposed: RR: 1.0 [reference] Ever a mechanic for at least 6 months: RR: 1.21 [0.90– 1.63] Ever had a chemically related occupation for at least 6 months: RR: 0.99 [0.81–1.20] SMR: 118.1 [77.7–171.9]

Wynder et al., 1985

Occupational exposure to diesel exhaust: Occupation type Probable exposure (minimal, moderate, high)

Silverman et al., 1986

Motor exhaust-related occupation

366

RETROSPECTIVE COHORT STUDY California 1964–1978 34,156 male members of the Operating Engineers Local Union No. 3, San Francisco Outcome: 27 deaths due to bladder cancer CASE-CONTROL STUDY 6 US cities 1981–1983 CASES: 194 men age 20–80 with histologically confirmed primary bladder cancer CONTROLS: 582 male hospital controls, matched on age, race, year of interview, and hospital of admission

CASE-CONTROL STUDY United States (National Bladder Cancer Study) 1977–1978 CASES: 1909 white male bladder cancer patients CONTROLS: 3569 male population controls, frequency matched on age and geographic area

Occupation type: Never exposed to diesel exhaust: OR: 1.00 [reference] Warehousemen, materials handlers: OR: 0.85 [0.18–4.14] Bus and truck drivers: OR: 0.90 [0.44–1.97] Railroad workers: OR: 2.00 [0.34–11.61] Heavy equipment operators, mechanics: OR: 0.75 [0.16–3.53] Never employed in a motor exhaust-related occupation: OR: 1.0 [reference] Ever truck driver or deliveryman: OR: 1.3 [1.1–1.4]

The US national age-sex-racecause-specific mortality rates for 5-year periods from 1964 –1978 were used to calculate the expected deaths

Adjusted ORs were calculated using the maximum likelihood method

Table 19–5. (cont.) Reference

Exposure

Population/Location

Findings

CASE-CONTROL STUDY Greater La Plata area, Argentina 1983–1985 CASES: 117 men and women with incident bladder cancer CONTROLS: 117 hospital controls, and 117 community controls, matched on gender and age CASE-CONTROL STUDY Copenhagen, Denmark 1979–1981 CASES: 371 men and women with bladder cancer, as identified though a cancer registry CONTROLS: 790 population controls, frequency matched on age and gender SURVEY Sweden 1961–1979 11,702 men with bladder cancer, as identified through the Swedish CancerEnvironment Registry

Never truck or railway driver: OR: 1.00 [reference] Ever truck or railway driver: OR: 4.31

Adjusted for age and smoking ORs estimated using logistic regression Confidence interval not provided

Land Transport Industry: Never worked: OR: 1.00 [reference] Ever worked: OR: 1.55 [1.06–2.28]

All risk estimates were adjusted for age and gender; the risk estimates for the duration of years worked were also adjusted for smoking ORs were calculated using logistic regression

SIR: 1.04

CASE-CONTROL STUDY Hamilton County, Ohio 1960–1982 CASES: 731 male bladder cancer deaths CONTROLS: 95,057 all other male deaths PROSPECTIVE COHORT STUDY United States CPS II 1982–1988 1.2 million men and women

Never employed as a truck driver: OR: 1.00 [reference] Ever employed as a truck driver <20 years: OR: 1.06 ≥20 years: OR: 12.00

Adjusted for age and region The number of expected cases was obtained by applying the 5-year birth cohort- and sex-specific rate for bladder cancer in the general Swedish population Crude ORs Confidence intervals were not provided

Iscovich et al., 1987

Occupation: Truck or railway driver

Jensen et al., 1987

Occupation

Malker et al., 1987

Major industry of employment: transportation and communication

Steenland et al., 1987

Occupation: driver

Boffetta et al., 1988

Diesel exhaust exposure

Claude et al., 1988

Occupation: Gas station attendant Locomotive driver Truck driver

CASE-CONTROL STUDY Northern Germany 1977–1984 CASES: 531 men with histologically confirmed bladder cancer CONTROLS: 531 male hospital controls, matched on age

Risch et al., 1988

Diesel or traffic fumes

CASE-CONTROL STUDY Canada (Edmonton, Calgary, Toronto, Kingston) 1979– 1982 CASES: 826 men and women with newly diagnosed bladder cancer CONTROLS: 792 population controls matched on age, sex, and area of residence

Siemiatycki et al., 1988

Occupational exposure to diesel exhaust

CASE-CONTROL STUDY Montreal, Canada Dates unspecified CASES: 452 men age 35–70 with bladder cancer CONTROLS: 2196 men with other cancers (excluding lung or kidney cancer patients)

Never diesel exhaust exposure: RR: 1.00 [reference] Ever diesel exhaust exposure: RR: 1.04 [not significant] Never gas station attendant: OR: 1.00 [reference] Ever gas station attendant: OR: 0.33 [0.04–2.87] Never locomotive driver: OR: 1.00 [reference] Ever locomotive driver: OR: 3.00 [1.02–8.80] Never truck driver: OR: 1.00 [reference] Ever truck driver: OR: 1.78 [1.12–2.83] MEN Occupation: Never job with contact with diesel or traffic fumes: OR: 1.0 [reference] Ever job with contact with diesel or traffic fumes: OR: 1.53 [1.17–2.00] WOMEN Occupation: Never job with contact with diesel or traffic fumes: OR: 1.0 [reference] Ever job with contact with diesel or traffic fumes: OR: 0.62 [0.23–1.57] Never exposed: OR: 1.0 [reference] Ever exposed: OR: 1.0 [0.8–1.2]

Comments

RRs were calculated using the Mantel and Haenszel method and were standardized by age, smoking, and occupational exposures 144 females with bladder cancer were in the study but were not used for this analysis

ORs calculated using continuous conditional logistic regression ORs adjusted for lifetime cigarette consumption

OR calculated using MantelHaenszel analyses OR adjusted for age, SES, ethnic group, cigarette smoking, and blue-/whitecollar job history

(continued)

367

Table 19–5. (cont.) Reference

Exposure

Bonassi et al., 1989

Occupational exposure to polycyclic aromatic hydrocarbons

Paradis et al., 1989

Occupation: Bus Driver Employed <30 years Employed ≥30 years

Schumacher et al., 1989

Occupation: Driver

Silverman et al., 1989b

Occupation: Motor vehicle driver

Silverman et al., 1989a

Occupation: Motor vehicle driver

Iyer et al., 1990

Diesel exhaust exposure from main occupation or from self-reported occupational exposure

Steineck et al., 1990

Exposure to industrial diesel exhaust

Burns and Swanson, 1991

Occupation: Bus and truck transport

368

Population/Location

Findings

Comments

CASE-CONTROL STUDY Bormida Valley, Italy 1972–1982 CASES: 121 men and women with incident bladder cancer CONTROLS: 342 population controls, matched on age and gender RETROSPECTIVE COHORT STUDY Montreal, Canada 1962–1985 COHORT: 2134 male bus drivers hired before January 1, 1957 4 deaths from bladder cancer CASE-CONTROL STUDY Utah 1977–1978 CASES: 417 men and women with bladder cancer CONTROLS: 877 controls frequency matched on age and gender

Never exposed: OR: 1.00 [reference] Ever worked as truck driver: OR: 1.88 [0.44–8.00] Ever worked as auto mechanic: OR: 1.84 [0.43–7.84]

Adjusted for cigarette smoking ORs calculated using logistic regression

Employed <30 years: SMR: 97 [12–350] Employed ≥30 years: SMR: 37 [4–134]

Expected number of deaths calculated from age, sex, and cause-specific death rates of Montreal city men for the years of 1971 and 1981

MEN Never employed as driver: OR: 1.00 [reference] Employed as driver at least 10 years: OR: 1.24 [0.77– 2.01

For men, OR was adjusted for smoking (by stratification) For women, crude OR reported

CASE-CONTROL STUDY United States (National Bladder Cancer Study) 2100 white males with bladder cancer 3874 white male population controls, matched on age and geographic area CASE-CONTROL STUDY United States (National Bladder Cancer Study) CASES: 126 nonwhite men with bladder cancer CONTROLS: 126 nonwhite male population controls, matched on age and geographic area CASE-CONTROL STUDY United States (6 cities, unspecified) Dates unspecified CASES: 136 men with urinary bladder cancer CONTROLS: 272 hospital controls diagnosed with nontobacco-related diseases, individually matched on sex, age, race, hospital, and year of interview CASE-CONTROL STUDY Stockholm, Sweden 1985–1987 CASES: 320 men with incident bladder cancer CONTROLS: 363 male population controls CASE-CONTROL STUDY Detroit metropolitan area Dates unspecified (1973–?) CASES: 2160 men and women with bladder cancer CONTROLS: 3979 colon and rectum cancer patients

WOMEN Never employed as driver: OR: 1.00 [reference] Ever employed as driver: OR: 1.13 [0.02–22.0] Never drove motor vehicles: OR: 1.0 [reference] Ever drove motor vehicles: OR: 1.2 [1.1–1.4]

Adjusted for smoking Adjusted ORs were calculated using the maximum likelihood method

Never driver of motor vehicles: OR: 1.0 [reference] Ever driver of motor vehicles: OR: 1.0 [0.6– 1.5]

Adjusted for smoking Adjusted ORs were calculated using the maximum likelihood method

Never exposed: OR: 1.0 [reference] Ever exposed: MantelHaenszel OR: 1.06 [0.64– 1.76] Logistic regression OR: 1.24 [0.77–2.00]

OR adjusted for smoking status and education status

Never exposed: OR: 1.0 [reference] Ever exposed: 1.7 [0.9–3.3]

OR calculated using logistic regression

Never worked in an “exposed” occupation: OR: 1.0 [reference] Ever worked in bus or truck transport: OR: 0.6 [0.4–1.0]

OR calculated using unconditional logistic regression OR adjusted for cigarette smoking, race, gender, and age at diagnosis

Table 19–5. (cont.) Reference

Exposure

Population/Location

Findings Never locomotive driver: OR: 1.0 [reference] Ever locomotive driver: OR: 3.0 [1.0–8.8] Never truck driver: OR: 1.0 [reference] Ever truck driver: OR: 1.8 [1.1–2.8] Never worked in railway transport: OR: 1.00 [reference] Ever worked in railway transport: OR: 0.80 [0.49–1.30] Never worked in road transportation: OR: 1.00 [reference] Ever worked in road transportation: OR: 1.02 [0.62–1.69] Occupation: Never motor transport worker: OR: 1.0 [reference] Ever motor transport worker <10 years: OR: 1.4 [0.9–2.0] Ever motor transport worker ≥10 years: OR: 1.3 [0.9–1.9] Never motor vehicle driver: OR: 1.0 [reference] Ever motor vehicle driver: OR: 0.9 [0.5–1.7]

Kunze et al., 1992

Occupation: Locomotive driver Truck driver

CASE-CONTROL STUDY Germany 1977–1985 CASES: 675 men and women with bladder cancer CONTROLS: 675 male and female hospital controls, matched on gender and age

Cordier et al., 1993

Occupation: Railway transport Road transportation

CASE-CONTROL STUDY France 1984 –1987 CASES: 658 men with bladder cancer CONTROLS: 658 male hospital controls, matched on age, residence, and hospital

Siemiatycki et al., 1994

Occupation and occupational exposures

CASE-CONTROL STUDY Montreal, Quebec, Canada 1979–1986 CASES: 484 persons with bladder cancer CONTROLS: 1,879 controls with cancers at other sites and 533 population controls (combined)

Porru et al., 1996

Job title

Soll-Johanning et al., 1998

Occupation: Urban bus driver or tramway employee: Ever employed >3 months Never employed or employed £3 months

Boffetta et al., 2001

Occupational exposure to diesel engine emissions (according to occupations listed in censuses): Low/medium/high probability of exposure

CASE-CONTROL STUDY Northern Italy 1992–1993 CASES: 275 men with histologically confirmed bladder cancer CONTROLS: 397 male hospital controls, matched on age RETROSPECTIVE COHORT STUDY Copenhagen, Denmark 18,174 bus drivers or tramway employees during 1900–1994 177 incident cases of bladder cancer diagnosed between 1943–1992 PROSPECTIVE COHORT STUDY Sweden Employed Swedish population (according to 1960 and 1970 censuses) Followed for cancer incidence 1971–1989 Exposed men: 7.4 million person-years Exposed women: 240,000 person years

Zeegers et al., 2001

Occupational exposure to CASE-COHORT STUDY diesel exhaust: By exposure The Netherlands tertile COHORT: 58,279 men aged 55–69 at baseline Followed 1986–1992 CASES: 532 incident cases of cancer of the urinary bladder, ureters, renal pelvis, or urethra CONTROLS: 1688 randomly chosen cohort members

Comments Crude ORs provided

Adjusted for smoking ORs calculated using logistic regression 107 female case-control pairs were in the study, but their occupational exposures did not involve exposure to air pollution

80 female cases and 182 female controls were also in the study, but their occupational exposures did not involve exposure to air pollution

Using the entire Danish population as the reference group: SIR: 1.4 [1.2–1.6] Using the population of Copenhagen as the reference group: SIR: 1.1 [0.9–1.3]

Statistical tests based on the assumption that the observed number of cancer cases followed the Poisson distribution

MEN Probability of Exposure: Unexposed: RR: 1.0 [reference] Low: OR: 0.99 [0.94–1.09] Medium: OR: 0.84 [0.79–0.89] High: OR: 0.98 [0.92–1.04] WOMEN Probability of Exposure: Unexposed: RR: 1.0 [reference] Low: OR: 0.94 [0.63–1.39] Medium: OR: 0.50 [0.16–1.55] High: OR: 0.68 [0.28–1.63] No exposure: RR: 1.00 [reference] Low exposure tertile: RR: 0.98 [0.64–1.51] Medium exposure tertile: RR: 0.98 [0.62–1.55] High exposure tertile: RR: 1.21 [0.78–1.88] P value for trend = 0.50

RR estimated using Poisson regression RR adjusted for age, calendar period, geographical region, and urban/rural residence

RR calculated with exponentially distributed failure time regression models RR adjusted for age, cigarettes smoked per day, and years of cigarette smoking

(continued)

369

370

PART III: THE CAUSES OF CANCER

Table 19–5. (cont.) Reference Soll-Johanning et al., 2003

Exposure Cumulated time of employment

Population/Location

Findings

NESTED CASE-CONTROL STUDY

>10 YEARS LAG TIME:

Copenhagen, Denmark COHORT: 18,174 bus drivers or tramway employees during 1900–1994 CASES: 84 incident bladder cancer cases CONTROLS: 606 randomly chosen cohort members

Cumulated employment: <3 months: OR: 1.21 [0.12–12.34] 3 months–<2 years: OR: 1.00 [reference] 2–<10 years: OR:1.23 [0.46–3.30] 10–<20 years: OR: 1.61 [0.57–4.55]

Comments OR calculated using unconditional logistic regression OR adjusted for smoking

OR, odds ratio; SES, socioeconomic status.

et al., 1987; Garshick et al., 1988; Woskie et al., 1988b; Woskie et al., 1988a) and studies of teamsters (Steenland et al., 1990; Zaebst et al., 1991). Reanalysis of the railroad workers studies showed an inverse relationship between lung cancer risk and duration of employment in exposed job categories. HEI recommended against using the railroad workers’ data as the basis for a quantitative risk assessment in ambient settings, and recommended further analysis of the Teamster data with regard to reducing uncertainties in exposure estimates (Health Effects Institute and Diesel Epidemiology Expert Panel, 1999). HEI subsequently funded several pilot studies to assess the feasibility of providing better quantitative estimates of exposure to diesel exhaust that could be used in subsequent epidemiologic studies. Their recent review of the results of these feasibility studies resulted in recommendations for research to obtain better information for diesel quantitative risk assessment. HEI noted that “the emphasis in [diesel health] risk assessment has shifted to quantifying the burden of lung cancer and other diseases” and that “this shift requires quantitative estimates of the relation between disease risk and the level and conditions of exposure.” (Health Effects Institute and Diesel Epidemiology Working Group, 2002).

Childhood Cancer and Exposure to Traffic-Related Air Pollution Motor vehicle exhaust contains a number of known carcinogens, including the leukemogen, benzene. Some studies report increased risk of leukemia in occupational groups exposed to diesel and other motor vehicle exhausts (e.g., Flodin et al., 1988) but others do not (Cohen and Higgins, 1995). A recent meta-analysis of 48 studies reported that parental occupational exposure to vehicular exhaust is associated with an increased risk of childhood cancer (Colt and Blair, 1998), and parental motor vehicle-related occupation was associated with an increased risk of childhood brain tumors in a recent, large European case-control study (Cordier et al., 2001). Until recently, however, few studies had estimated the risk of childhood cancer from exposure of the child to traffic-related air pollution. In 1989 Savitz and Feingold, using data from an earlier case-control study of residential exposure to electric and magnetic fields (Savitz and Feingold, 1989), reported that vehicular traffic density at the residence of cancer cases was associated with an increased risk of leukemia and brain cancer among children in Denver. In the intervening decade, as interest in the health effects of air pollution grew and spatial statistical methods, such as GIS, became more widely available, this relationship has been studied by others with inconsistent results (Table 19–6). Pearson et al. (2000), in a further reanalysis of the Denver data, corroborated Savitz and Feingold’s results using a more detailed characterization of traffic density, but Langholz and colleagues (Langholz et al., 2002) failed to find an association of childhood leukemia with traffic density in Los Angeles after controlling for residential exposure to EMF.

Several Scandinavian studies estimated the outdoor residential concentrations of motor vehicle-related pollutants, in addition to traffic density. Two studies found little evidence of an association with cancer risk with either index (Nordlinder and Jarvholm, 1997; RaaschouNielsen et al., 2001), but one small study (Feychting et al., 1998) reported an increased risk of all childhood cancers, including elevations in the relative risk of leukemia and brain cancer, associated with residential ambient levels of nitrogen dioxide.

INDOOR AIR POLLUTION Carcinogens in indoor air have large potential implications for risk because people may spend substantial amounts of time indoors. Indoor air pollution may stem from incoming outdoor air or originate indoors from tobacco smoking, building materials, soil gases, household products, and combustion from heating and cooking (Spengler, 1991). In more developed countries, two of the most important indoor pollutants that influence lung cancer risk in never-smokers are passive smoking (US Department of Health and Human Services (USDHHS), 1986; Samet and Wang, 2000) and radon (National Research Council (NRC) and Committee on the Biological Effects of Ionizing Radiations, 1988; National Research Council (NRC) et al., 1998). Asbestos exposure may pose a risk to building occupants, but the resulting risk is estimated to be minimal (Health Effects Institute et al., 1991). Other carcinogens are present in indoor air, but most epidemiologic research has focused on involuntary smoking, radon, and asbestos. Of major concern in the developing world is the indoor air contamination resulting from the use of unprocessed solid fuels, notably coal, for cooking and space heating (Chen et al., 1990).

Asbestos Asbestos, a well-established occupational carcinogen, refers to several forms of fibrous, naturally occurring silicate minerals that have been widely used in products found in homes and public and commercial buildings (Health Effects Institute et al., 1991). The epidemiologic evidence on asbestos and lung cancer dates to the 1950s, although clinical case series had previously led to the hypothesis that asbestos causes lung cancer (Wedler, 1944; Lynch and Smith, 1939). In a retrospective cohort study published in 1955, Doll observed that asbestos textile workers at a factory in the United Kingdom had a 10-fold elevation in lung cancer risk and that the risk was most heavily concentrated during the time frame before regulations were implemented to limit asbestos dust in factories (Doll, 1955). A sevenfold excess of lung cancer was subsequently observed among insulation workers in the United States (Selikoff et al., 1964; Selikoff et al., 1979). The peak incidence occurred 30–35 years after the initial exposure to asbestos (Selikoff, 1980). Subsequently, numerous studies of specific worker groups and of the general population have repeatedly documented the carcinogenicity of asbestos (Health Effects Institute et al., 1991). The

Table 19–6. Air Pollution and Childhood Cancers Reference Savitz and Feingold, 1989

Exposure Traffic density (vehicles/ day)

Population/Location CASE-CONTROL STUDY Denver, Colorado 1976–1982 CASES: 328 children age 0–14 years with any incident cancer (98 cases of leukemia) CONTROLS: 262 population controls matched on age and telephone exchange area

Nordlinder and Jarvholm, 1997

Benzene from gasoline and car exhaust, estimated by the number of cars per m2

ECOLOGIC STUDY Sweden (277 municipalities) Leukemia incidence data from 1975–1985 for persons aged 0–24 years at diagnosis were collected from the National Swedish Cancer Registry

Feychting et al., 1998

NO2 concentration (mg/m3) of outdoor air, using the 99th percentile of NO2 content in the outdoor air for 1-hour averages over 1 year

CASE-CONTROL STUDY Sweden 1960–1985 CASES: 142 children age 0–15 diagnosed with cancer (39 leukemia cases and 33 cases with CNS tumors) who lived within 300 m of any of the 220 and 400 kV power lines in Sweden CONTROLS: 550 population controls, matched on age, residence location, and living near the same power line

Harrison et al., 1999

Distance to a major road; distance to a petrol station

CASE-CONTROL STUDY England 1990–1994 CASES: 130 children age 0–15 diagnosed with leukemia CONTROLS: 130 children age 0–15 diagnosed with a solid cancer

Pearson et al., 2000

Distance-weighted traffic density (in vehicles per day [VPD])

CASE-CONTROL STUDY Used case-control data from Savitz et al. 1988 Denver, Colorado 1976–1982 CASES: 320 children aged 0–14 years with any incident cancer (97 cases of leukemia) CONTROLS: 259 population controls matched on age, gender, and telephone exchange area

Findings ALL CANCER SITES: <500 vehicles/day: OR: 1.0 [reference] ≥500 vehicles/day: OR: 1.7 [1.0–2.8] ALL LEUKEMIAS: <500 vehicles/day: OR: 1.0 [reference] ≥500 vehicles/day: OR: 2.1 [1.1–4.0] Cancer incidence rate (per 106 person-years): ACUTE LYMPHOCYTIC LEUKEMIA: <5 cars/km2: 21.1 [17.0–25.8] 5–9 cars/km2: 19.9 [16.4–24.0] 10–19 cars/km2: 22.5 [19.1–26.4] ≥20 cars/km2: 18.6 [11.3–21.4] ALL CANCER SITES: £39 mg/m3 NO2: OR: 1.0 [reference] 40–49 mg/m3 NO2: OR: 1.3 [0.4–4.3] ≥50 mg/m3 NO2: OR: 2.7 [0.9–8.5] ALL LEUKEMIAS: £39 mg/m3 NO2: OR: 1.0 [reference] 40–49 mg/m3 NO2: OR: 1.7 [0.2–14.6] ≥50 mg/m3 NO2: OR: 2.7 [0.3–20.6] ALL LEUKEMIAS: >100 m from major road: OR: 1.00 [reference] <100 m from major road: OR: 1.61 [0.90–2.87] >100 m from petrol station: OR: 1.00 [reference] <100 m from petrol station: OR: 1.99 [0.73–5.43] ALL LEUKEMIAS: 750-ft distance-weighted traffic density: <500 VPD: OR: 1.0 [reference] 500–4999 VPD: OR: 0.94 [0.53–1.66] 5000–9999 VPD: OR: 2.04 [1.05–3.95] 10000–14999 VPD: OR: 0.48 [0.10–2.21] 15000–19999 VPD: OR: 1.04 [0.20–5.34] ≥20000 VPD: OR: 8.28 [2.09–32.80]

Comments ORs were calculated using the Mantel-Haenszel method

There was no association between the number of cars per day and the incidence of chronic myeloid leukemia

ORs adjusted for magnetic fields and socioeconomic status ORs were calculated using conditional logistic regression

No adjustments

ORs were calculated using “standard statistical programs for stratified analysis”

(continued)

371

Table 19–6. (cont.) Reference

Exposure

Population/Location

Raaschou-Nielsen et al., 2001

Traffic density: vehicles per day (VPD) during childhood period NO2 and benzene levels at the front door of each dwelling (in 1000 ppb-days) during childhood period

CASE-CONTROL STUDY Denmark 1968–1991 CASES: 1989 children age 15 or younger with cancer CONTROLS: 5506 child population controls, matched on gender, age, and calendar time

Reynolds et al., 2001

Traffic density: average number of cars per day (ADT)

Langholz et al., 2002

Traffic density distribution in vehicles per day (VPD)

CASE-CONTROL STUDY San Diego County, California 1988–1994 CASES: 90 children under age 5 diagnosed with childhood leukemia CONTROLS: 349 children not known to have developed any cancer, matched by gender and birth date CASE-CONTROL STUDY California 1978–1984 CASES: 212 children age 0–10 diagnosed with leukemia CONTROLS: 202 children (friends and random community controls), matched on age and gender

Reynolds et al., 2002

Vehicle density (vehicles per square mile) Road density (miles of road per square mile) Traffic density (vehicle miles traveled per day per square mile)

372

LONGITUDINAL STUDY California 1988–1994 6988 children aged 0–15 with a newly diagnosed cancer (2443 with leukemias, 1351 with gliomas) 46 million child-years of observation using population data from 1990 census

Findings ALL CANCER SITES: Traffic density: <500 VPD: OR: 1.0 [reference] 500–<5000 VPD: OR: 0.9 [0.8–1.0] 5000–<10,000 VPD: OR: 0.8 [0.6–1.1] >10,000 VPD: OR: 1.0 [0.7–1.6] Benzene concentration: <3.8 in 1000 ppb-days: OR: 1.0 [reference] 3.8–<10.4 in 1,000 ppbdays: OR: 0.9 [0.8–1.1] 10.4–<27.0 in 1000 ppbdays: OR: 1.0 [0.7–1.3] ≥27.0 in 1000 ppb-days: OR: 0.5 [0.2–1.0] NO2 concentration: <11.5 in 1000 ppb-days: OR: 1.0 [reference] 11.5–<29.4 in 1000 ppbdays: OR: 1.1 [0.9–1.3] 29.40–<57.8 in 1000 ppbdays: OR: 1.1 [0.8–1.5] ≥57.8 in 1000 ppb-days: OR: 1.2 [0.6–2.3] Total ADT within 500 feet: No attributed segments: OR: 2.00 [0.97–4.10] <10,000: OR: 1.0 [reference] 10,000–19,999: OR: 1.38 [0.61–3.15] ≥20,000: OR: 1.59 [0.76– 3.34] ALL LEUKEMIAS: Quintiles of traffic density distribution: 1st (£2300 VPD): OR: 1.0 [reference] 2nd (2301–5996 VPD): OR: 1.6 [0.8–3.6] 3rd(5997–13,263 VPD): OR: 1.1 [0.5–2.4] 4th(13,264 –28,496 VPD): OR: 1.1 [0.5–2.2] 5th(≥28,497 VPD): OR: 1.4 [0.7–3.0] ALL CANCER SITES: Vehicle density percentile: 1st–24th: RR: 1.00 [reference] 25th–49th: RR: 1.04 [0.97–1.12] 50th–74th: RR: 0.98 [0.91–1.06] 75th–89th: RR: 1.10 [1.01–1.19] 90th+: RR: 0.98 [0.89–1.08] ALL LEUKEMIAS: Vehicle density percentile: 1st–24th: RR: 1.00 [reference] 25th–49th: RR: 1.05 [0.93–1.19] 50th–74th: RR: 1.04 [0.92–1.19] 75th–89th: RR: 1.18 [1.03–1.35] 90th+: RR: 1.02 [0.86–1.20]

Comments Traffic density ORs are crude Benzene and NO2 exposure ORs are adjusted for urban development, geographic region, type of residence, electromagnetic fields, mother’s age, and birth order ORs calculated using logistic regression Risk estimated also reported for exposure during pregnancy period

ORs adjusted for birth year, gender, race/ethnicity, and median family income of block group ORs calculated using conditional logistic regression

ORs were adjusted for wire-code ORs were calculated using conditional logistic regression

RRs were adjusted for age, race/ethnicity, and gender RRs were calculated using Poisson regression Area-specific child-years of observation over the study period were generated by multiplying 1990 population of each block group by seven and applying the appropriate growth factor for each age/race/gender group

373

Air Pollution Table 19–6. (cont.) Reference Reynolds et al., 2003

Exposure Hazardous Air Pollution (HAP) exposure score, calculated by combining cancer potency factors with outdoor HAP concentrations modeled by the EPA

Population/Location

Findings

Comments

LONGITUDINAL STUDY California 1988–1994 6989 children under 15 years with an invasive cancer listed in California Cancer Registry (2443 leukemia cases and 1351 glioma cases) 46 million child-years of observation using population data from 1990 census

ALL CANCER SITES: Percentile of exposure score: 1–24: RR: 1.0 [reference] 25–74: RR: 1.04 [0.98– 1.09] 75–89: RR: 1.06 [0.98– 1.14] ≥90: RR: 1.06 [0.97–1.16] ALL LEUKEMIAS: Percentile of exposure score: 1–24: RR: 1.0 [reference] 25–74: RR: 1.10 [0.99–1.23] 75–89: RR: 1.11 [0.96– 1.28] ≥90: RR: 1.21 [1.03–1.42] P value for trend <0.05

RRs adjusted for age, race/ ethnicity, and gender RRs calculated using Poisson regression Area-specific child-years of observation over the study period were generated by multiplying 1990 population of block group by seven and applying the appropriate growth factor for each age/ race/gender group Stratified results available by type of emission source (mobile, area, point)

OR, odds ratio.

risk of lung cancer has been noted to increase with increased exposure to asbestos and to be associated with the principal commercial forms of asbestos. Asbestos and cigarette smoking are both independent causes of lung cancer, but in combination they act synergistically to increase the risk of lung cancer in a manner that is compatible with a multiplicative effect (Hammond et al., 1979; International Agency for Research on Cancer (IARC), 2002). With the recognition in the 1970s and 1980s that asbestoscontaining materials were prevalent in indoor environments and a potential source of airborne fibers, concern was raised as to risks to the general population for both mesothelioma and lung cancer. In the United States, programs were initiated to either remove or manage asbestos-containing materials to limit releases of fibers. Because the concentrations were generally very low, the risks could not be directly addressed using epidemiological approaches. A risk assessment was carried out by the Health Effects Institute that used all available measurements of indoor asbestos fiber concentrations and a concentrationrisk relationship from studies of workers (Health Effects Institute et al., 1991). The concentration information considered by the Health Effects Institute showed that exposures in indoor environments should generally be quite low and not elevated in comparison with levels in urban outdoor air, which is contaminated by fibers from brake linings and other sources. The US Environmental Protection Agency estimates that, if an individual were to continuously breathe air containing asbestos at an average of 0.000004 fibers/cm3 over his or her entire lifetime, that person would theoretically have no more than a 1 in 1 million increased chance of developing cancer as a direct result of breathing air containing this chemical. Similarly, the EPA estimates that breathing air containing 0.00004 fibers/cm3 would result in not greater than a 1 in 100,000 increased chance of developing cancer, and air containing 0.0004 fibers/cm3 would result in not greater than a 1 in 10,000 increased chance of developing cancer (US Environmental Protection Agency (EPA), 1999). This risk was considered sufficiently low as to not warrant removal of asbestos from buildings and the Environmental Protection Agency now promotes in-place management of asbestos-containing materials.

Radon Radon is an inert gas that is produced naturally from radium in the decay series of uranium. It decays into a series of relatively short-lived particle progeny; two of the short-lived decay products emit alpha particles that can, by virtue of their high energy and mass, cause damage

to the DNA of cells of the respiratory epithelium. Laboratory experiments involving exposure of cells to single alpha particles show that even one “hit” from an alpha particle causes permanent change in a cell; this finding implies that radon exposure may result in increased risk at any level (National Research Council (NRC) et al., 1998). Epidemiologic studies of underground miners of uranium and other ores have long established exposure to radon daughters as a cause of lung cancer (National Research Council (NRC) and Committee on the Biological Effects of Ionizing Radiation, 1988; National Research Council (NRC) et al., 1998; Lubin et al., 1995a). In fact, radon exposure was probably the first occupational respiratory carcinogen to be identified, based on the extremely high rates of lung cancer documented in the underground metal miners of Schneeberg and Joachimsthal and the high levels of radon measured in the mines (National Research Council (NRC) et al., 1998; Proctor, 1999). While lesser risks have been observed for more recent worker cohorts, the newer epidemiological studies still show clear evidence of existing cancer risk (National Research Council (NRC) et al., 1998). Cigarette smoking and radon decay products synergistically influence lung cancer risk in a manner that is supra-additive but sub-multiplicative (National Research Council (NRC) et al., 1998; Lubin et al., 1995a). Radon is of broader societal interest because it is a ubiquitous indoor air pollutant, entering buildings in soil gas. On average, indoor exposures to radon for the general population are much less than received by occupational groups such as uranium miners. For example, even the lowest historical radon concentration in a uranium mine is still roughly an order of magnitude higher than in the average home (Lubin et al., 1995a). However, measurements of indoor radon concentrations in homes show a log-normal distribution with all homes having some radon detectable at an average around 1 picocurie per liter (37 becquerels per m3), and there is a highly skewed tail extending to extremely high levels, as high as those in uranium mines where workers had increased lung cancer risk. As a basis for developing control strategies, risk assessments have been carried out to estimate the magnitude of the problem and to ascertain the extent to which indoor radon concentrations must be reduced to protect public health. Because of the shape of the concentration distribution, the risks conveyed by the lower and typical exposures are of particular concern. Since the problem of indoor radon was first widely recognized in the early 1980s, case-control studies have been carried out to estimate the risks directly, adding to the information from cohort studies of underground miners, generally exposed at much higher levels. In the first of these studies, radon exposure was generally inferred based on a surrogate measure, but the more recent studies have incorporated

374

PART III: THE CAUSES OF CANCER

measurements of radon concentrations in the homes of lung cancer cases and controls. The results of most of the individual studies have now been reported; interpretation of their findings is limited by exposure misclassification, a particularly difficult problem in estimating lifetime exposure (Lubin et al., 1995a). Measurement of implanted polonium-210 in the surface of glass (e.g., window glass or glass covering a picture) has been proposed as an indicator of long-term concentration and applied in several epidemiological studies (Field et al., 2002; Steck et al., 2002; Brownson and Alavanja, 2000). Nonetheless, the risk estimates from individual case-control studies have been relatively imprecise. A meta-analysis of the findings through the mid-1990s did show evidence for a significant and positive exposureresponse relationship that was consistent with the extrapolated risks in the miner studies (National Research Council (NRC) et al., 1998). Subsequently, the number of studies completed has doubled and, as this chapter was prepared, pooled analyses had been completed or were in progress for the North American and European studies, respectively. For the purpose of risk assessment and policy formulation, risk models have been developed based on the findings in the underground miners (National Research Council (NRC) et al., 1998). There are inherent uncertainties in extending such models to the general population, including particularly the extrapolation of findings from uranium miners to the generally lower exposures indoors, differences between mines and homes that may influence the dosimetry of radon decay products in the lungs, and a lack of information in the miner data on women and children. Effect modification by smoking is a further uncertainty with the evidence from the miners indicating synergism, although the analyses indicate a combined effect of smoking and radon that is sub-multiplicative (National Research Council (NRC) et al., 1998). From the policy perspective, the most critical uncertainty is whether radon causes lung cancer at all exposures—that is, is there a threshold exposure below which radon does not cause cancer? Biological evidence supports the assumption that a single hit to a cell by an alpha particle causes permanent cellular change, an assumption leading to a non-threshold dose-response relationship (National Research Council (NRC) et al., 1998; Hei et al., 1994; Hei et al., 1997). Additionally, the case-control study findings to date provide evidence for increased risk down to levels of exposure not substantially greater than that associated with typical indoor concentrations. When combined in a metaanalysis, there is a significant association between indoor radon and lung cancer in the general population that is quantitatively comparable with risk models derived from the underground miners. This coherence lends support to using extrapolation of the miner data with a linear, non-threshold model to estimate the risk of indoor radon. Consequently, the risk model developed by the National Research Council’s Biological Effects of Ionizing Radiation (BEIR) VI Committee (National Research Council (NRC) et al., 1998), and subsequently applied by the US Environmental Protection Agency incorporates a model for risk that is linear without a threshold. The assumptions made by the Environmental Protection Agency (US Environmental Protection Agency (EPA), 1992b) and the Biological Effects of Ionizing Radiation (BEIR) IV and VI Committees of the National Research Council (National Research Council (NRC) and Panel on Dosimetric Assumptions Affecting the Application of Radon Risk Estimates, 1991; National Research Council (NRC) et al., 1998) lead to estimates that approximately 21,000 lung cancer deaths per year in the United States are attributable to radon (with an uncertainty range of 8000 to 45,000), an estimate that makes indoor radon the second-leading cause of lung cancer (US Environmental Protection Agency (EPA) and Office of Radiation and Indoor Air, 2003).

Secondhand Smoke In 1981 reports were published from Japan (Hirayama, 1981) and from Greece (Trichopoulos et al., 1981) that indicated increased lung cancer risk in nonsmoking women married to cigarette smokers. Subsequently this controversial association has been examined in investigations conducted in the United States and other countries. The

association of secondhand smoke (SHS) with lung cancer derives biological plausibility from the presence of carcinogens in sidestream smoke and the lack of a documented threshold dose for respiratory carcinogens in active smokers (US Department of Health and Human Services (USDHHS), 1982; International Agency for Research on Cancer (IARC), 1986; International Agency for Research on Cancer (IARC), 1986; International Agency for Research on Cancer (IARC), 2002). Moreover, genotoxic activity had been demonstrated for many components of SHS (Lofroth, 1989; Claxton et al., 1989; Weiss, 1989; Claxton et al., 1989; Hillerdal, 1996). Experimental exposure of nonsmokers to ETS leads to their excreting NNAL, a tobacco-specific carcinogen, in their urine (Hecht et al., 1993), as does living with a smoker (Anderson et al., 2001). Nonsmokers exposed to SHS also have increased concentrations of adducts of tobacco-related carcinogens (Maclure et al., 1989; Crawford et al., 1994). Time trends of lung cancer mortality in nonsmokers have been examined with the rationale that temporally increasing exposure to SHS should be paralleled by increasing mortality rates (Enstrom, 1979; Garfinkel, 1981). These data provide only indirect evidence on the lung cancer risk associated with involuntary exposure to tobacco smoke. Epidemiologists have directly tested the association between lung cancer and involuntary smoking utilizing conventional designs: the case-control and cohort studies. The results of both study designs may be affected by inaccurate assessment of exposure to SHS, by inaccurate information on personal smoking habits that leads to classification of smokers as nonsmokers, by failure to assess and control for potential confounding factors, and by the misdiagnosis of a cancer at another site as a primary cancer of the lung. Methodological investigations suggest that accurate information can be obtained by interview in an epidemiological study on the smoking habits of a spouse (i.e., never or ever smoker) (Pron et al., 1988; Coultas et al., 1989; Cummings et al., 1989; Lubin, 1999). However, information concerning quantitative aspects of the spouse’s smoking is reported with less accuracy. Misclassification of current or former smokers as never- smokers may introduce a positive bias because of the concordance of spouse smoking habits (Lee, 1988; Lee, 1995; Wu, 1999). The extent to which this bias explains the numerous reports of association between spouse smoking and lung cancer has been controversial (Wald et al., 1986; Lee, 1988; US Environmental Protection Agency (EPA), 1992a; Lee, 1988, 1992). Use of spouse smoking alone to represent exposure to SHS does not cover exposures outside of the home (Friedman et al., 1983) or necessarily all exposure inside the home. The International Agency for Research on Cancer has conducted a 13-center study to assess the contribution of the home and work environments to exposures of nonsmoking women to SHS (Saracci and Riboli, 1989). Overall, the data show that some women married to smokers receive little exposure at home and that the number of cigarettes smoked per day by the husband is only moderately correlated with “actual” exposure, as opposed to average exposure. The study shows a widely varying proportion of women exposed to SHS among the centers. In some countries, including the United States, smoking prevalence varies markedly with indicators of income and education, more recently tending to rise sharply with decreasing educational level and income (US Department of Health and Human Services (USDHHS), 1989). In general, exposure to SHS follows a similar trend, and critics of the findings on ETS and lung cancer have argued that uncontrolled confounding by lifestyle, occupation, or other factors may explain the association. In fact, current data for the United States do indicate a generally less healthy lifestyle in those with greater SHS exposure (Matanoski et al., 1995). However, other than a few occupational exposures at high levels, as well as indoor radon, risk factors for lung cancer in never-smokers that might confound the SHS association cannot be proffered and the relevance to past studies of these current associations of potential confounders with SHS exposure is uncertain. Hirayama’s (1981) early report was based on a prospective cohort study of 91,540 nonsmoking women in Japan. Standardized mortality ratios (SMRs) for lung cancer increased significantly with the amount smoked by the husbands. The findings could not be explained by con-

Air Pollution founding factors and were unchanged when follow-up of the study group was extended (Hirayama, 1984). Based on the same cohort, Hirayama also reported significantly increased risk for nonsmoking men married to wives smoking 1 to 19 cigarettes and 20 or more cigarettes daily (Hirayama, 1984). In 1981, Trichopoulos et al. (Trichopoulos et al., 1981) also reported increased lung cancer risk in nonsmoking women married to cigarette smokers. These investigators conducted a case-control study in Athens, Greece, which included cases with a diagnosis other than for adenocarcinoma or terminal bronchial (alveolar) carcinoma. The positive findings reported in 1981 were unchanged with subsequent expansion of the study population (Trichopoulos et al., 1983). By 1986, the evidence had mounted and three consensus reports published in that year concluded that SHS was a cause of lung cancer. The International Agency for Research on Cancer of the World Health Organization (International Agency for Research on Cancer (IARC), 1986) concluded that “passive smoking gives rise to some risk of cancer.” In its monograph on tobacco smoking, the agency supported this conclusion on the basis of the characteristics of sidestream and mainstream smoke, the absorption of tobacco smoke materials during involuntary smoking, and the nature of dose-response relationships for carcinogenesis. In the same year, The National Research Council (National Research Council (NRC) and Committee on Passive Smoking, 1986) and the US Surgeon General (US Department of Health and Human Services (USDHHS), 1986) also concluded that involuntary smoking increases the incidence of lung cancer in nonsmokers. In reaching this conclusion, the National Research Council (National Research Council (NRC) and Committee on Passive Smoking, 1986) cited the biological plausibility of the association between exposure to ETS and lung cancer and the supporting epidemiological evidence. Based on a pooled analysis of the epidemiological data adjusted for bias, the report concluded that the best estimate for the excess risk of lung cancer in nonsmokers married to smokers was 25%. The 1986 report of the Surgeon General (US Department of Health and Human Services (USDHHS), 1986) characterized involuntary smoking as a cause of lung cancer in nonsmokers. This conclusion was based on the extensive information already available on the carcinogenicity of active smoking, on the qualitative similarities between SHS and mainstream smoke, and on the epidemiological data on involuntary smoking. In 1992 the US Environmental Protection Agency (US Environmental Protection Agency (EPA), 1992a) published its risk assessment of SHS as a carcinogen. The agency’s evaluation drew on the toxicologic evidence on SHS and the extensive literature on active smoking. A meta-analysis of the 31 studies published to that time was central in the decision to classify SHS as a class A carcinogen—namely a known human carcinogen. The meta-analysis considered the data from the epidemiologic studies by tiers of study quality and location and used an adjustment method for misclassification of smokers as neversmokers. Overall, the analysis found a significantly increased risk of lung cancer in never-smoking women married to smoking men; for the studies conducted in the United States, the estimated relative risk was 1.19 (90% CI: 1.04, 1.35). Critics of the report have raised a number of concerns including the use of meta-analysis, reliance of 90% rather than 95% confidence intervals, uncontrolled confounding, and information bias. The report, however, was endorsed by the Agency’s Science Advisory Board, and its conclusion is fully consistent with the 1986 reports. Subsequent to the 1992 risk assessment, findings of many additional studies in the United States and other countries have been reported (International Agency for Research on Cancer (IARC), 2002). The multicenter study of Fontham and colleagues (Fontham et al., 1994) is the largest report to date, having 651 cases and 1253 controls. It shows a significant increase in overall relative risk (OR = 1.26, 95% CI: 1.04, 1.54). There was also a significant risk associated with occupational exposure to SHS. Findings of an autopsy study conducted in Greece also strengthened the plausibility of the lung cancer/SHS association. Trichopoulos and colleagues (Trichopoulos et al., 1992) examined autopsy lung specimens from 400 persons 35 years of age and older, to assess airways

375

changes. Epithelial lesions were more common in nonsmokers married to smokers than in nonsmokers married to nonsmokers. A 1997 meta-analysis (Hackshaw et al., 1997) included 37 published studies. The excess risk of lung cancer for smokers married to nonsmokers was estimated as 24% (95% CI: 13%, 36%). Adjustment for potential bias and confounding by diet did not alter the estimate. This meta-analysis supported the 1998 conclusion of the U.K. Scientific Committee on Tobacco and Health (Scientific Committee on Tobacco and Health and HSMO, 1998) that SHS is a cause of lung cancer. A 2002 report by the International Agency for Research on Cancer (International Agency for Research on Cancer (IARC), 2002) reiterated the conclusion with the following statement: “. . . evidence is sufficient to conclude that involuntary smoking is a cause of lung cancer in never-smokers. The magnitudes of the observed risks are reasonably consistent with predictions based on studies of active smoking in many populations.” More recently, this conclusion was supported in a letter to the British Medical Journal by Hackshaw (Hackshaw, 2003). The extent of the lung cancer hazard associated with involuntary smoking in the United States and in other countries remains subject to some uncertainty, however (US Department of Health and Human Services (USDHHS), 1986; Weiss, 1986). The epidemiological studies provide varying and imprecise measures of risk, and exposures have not been characterized for large and representative population samples. Nevertheless, risk estimation procedures have been used to describe the lung cancer risk associated with involuntary smoking, but assumptions and simplifications must be made to use this method. The estimates of lung cancer deaths attributable to passive smoking have received widespread media attention and have figured prominently in the evolution of public policy on passive smoking. In 1990 Repace and Lowrey (1990) reviewed the risk assessments of lung cancer and passive smoking and estimated the numbers of lung cancer cases in United States nonsmokers attributable to passive smoking. The range of the nine estimates, covering both neversmokers and former smokers, provided by Repace and Lowery was from 58 to 8124 lung cancer deaths for the year 1988, with an overall mean of 4500 or 5000 excluding the lowest estimate of 58. The bases for the individual estimates included the comparative dosimetry of tobacco smoke in smokers and nonsmokers using presumed inhaled dose or levels of nicotine or cotinine, the epidemiological evidence, and modeling approaches. The 1992 estimate of the Environmental Protection Agency, based on the epidemiologic data was about 3000, including 1500 and 500 deaths in never-smoking women and men, respectively, and about 100 in long-term former smokers of both sexes (US Environmental Protection Agency (EPA), 1992a). These calculations illustrate that passive smoking must be considered an important cause of lung cancer death from a public health perspective; exposure is involuntary and not subject to control. The specific risk assessments require assumptions concerning the extent and degree of exposure to SHS, exposure-response relationships, and the lifetime expression of the excess risk associated with passive smoking at different ages. Moreover the calculations do not consider the potential contributions of other exposures, such as occupational agents and indoor radon. The current decline in the prevalence of active smoking and the implementation of strong clean indoor air policies will reduce the relevance of estimates based on past patterns of smoking behavior.

DEVELOPING COUNTRIES Current knowledge about ambient air pollution and lung cancer is based largely on the experience of populations of Western industrialized nations. The populations of the developing countries, however, are exposed to levels of air pollution from combustion sources in both ambient and indoor environments that rival or exceed those commonly observed in the industrialized West. Within the developing countries, the highest exposures, particularly among women, have been to indoor air pollution from the combustion of coal and biomass fuels for cooking and heating (Smith and Liu, 1994). For example, typical

376

PART III: THE CAUSES OF CANCER

concentrations of coal smoke in rural Chinese homes exceeded 500 mg/m3 and frequently exceeded 1 mg/m3 (Zhang and Smith, 1996). Smith and Liu (Smith and Liu, 1994) reviewed the epidemiologic literature on indoor air pollution and lung cancer in developing countries and found consistent evidence of increased rates of lung cancer associated with indoor cooking and heating with coal in studies done largely in China. A much smaller group of studies revealed no consistent association of lung cancer with indoor use of biomass fuels. Mumford and colleagues inferred that smoky coal was likely to be a major determinant of the geographic distribution of lung cancer in Xuan Wei, China (Mumford et al., 1987), a finding corroborated by an animal model (Mumford et al., 1990). Case-control studies conducted elsewhere in mainland China (Xu et al., 1989; Wu-Williams et al., 1990) and Taiwan (Ger et al., 1993) further implicated coal use as a risk factor for lung cancer. Another case-control study in Shanghai, where most of the homes are unheated, reported no association between use of coal and lung cancer risk (Gao et al., 1987). Exposure to coal burning in the pre-adult years was associated with lung cancer risk in a case-control study of women in Los Angeles County (Wu et al., 1985). Strong evidence for a role of “smoky” coal burning comes from a study of lung cancer in Xuanwei, China (Lan et al., 2002). Incidence dropped for those persons changed permanently from using an unvented firepit to a stove with a chimney. There has been little research on ambient air pollution and lung cancer among urban residents of the burgeoning cities of the developing countries, although mounting levels of urban air pollution, from local stationary and, increasingly, mobile sources are recognized as an important environmental problem by international public health and economic agencies (WHO World Health Report, 2002). In the cities of the poorest developing countries, WHO’s Global Environmental Monitoring System observed average ambient concentrations of total suspended particles of 300 mg/m3 although levels in locales where coal is used for fuel, such as poor communities in South Africa, may exceed 1g/m3 (World Health Organization, 1990). Estimated levels of PM10 in cities in developing countries also exceed those commonly encountered in Europe and North America (Pandey et al., 2004). One might predict that the high levels of ambient air pollution found in cities in the developing world would be associated with greater excess lung cancer occurrence than has been observed in Western industrialized settings. Although there are currently few relevant investigations, a case-control study in Shenyang, China, observed a 200% increase in lung cancer risk after adjustment for age, education, and smoking, among residents in “smoky” areas of the city and a 50% increase among those in “somewhat or slightly smoky” areas (Xu et al., 1989). As a greater proportion of the world’s population moves from rural communities to the rapidly expanding and highly polluted cities of Asia and the Southern Hemisphere, there is a clear need to address the large gap in epidemiologic research on outdoor air pollution and lung cancer in the developing world. These studies will present even greater challenges than those in the industrialized West. In addition to the generic problem of estimating long-term exposure to air pollution discussed above, the ambient air pollution mixture in urban centers in the developing countries is changing, due in part to the increase in vehicular traffic. Careful planning will be required to characterize these changes as they will have occurred over time, including choosing and measuring indicator pollutants for different pollution sources. In addition, the current dramatic increases in cigarette smoking in the developing world (Peto and Lopez, 2003), and the thoroughly predictable consequences, will complicate the interpretation of studies of air pollution and lung cancer.

FUTURE DIRECTIONS There remains a persistent basis for concern that indoor air pollution may cause lung cancer in smokers and nonsmokers. Carcinogens can be measured in indoor and outdoor environments and toxicologic data indicate the potential for human carcinogenicity. Epidemiologic research shows evidence of effects of indoor and outdoor air pollution on lung cancer risk, albeit weak for some agents. On the other hand,

levels of exposure to many agents would not be expected to greatly increase risk and the multifactorial etiology of lung cancer lowers the signal-to-noise ratio. Research is still needed on air pollution and lung cancer to guide policies for protection of public health. Direct epidemiologic observation of exposed populations can provide the best information for evaluating the magnitude of outdoor air pollution-related excess lung cancer. In general, large-scale epidemiologic studies of air pollution and lung cancer are needed if we are to obtain sufficiently informative data. Large numbers of cases will be necessary to measure the effects of air pollution among women and ethnic minorities and to measure the joint effects of air pollution and other factors, such as occupation and smoking. Without improved epidemiologic methods, however, even large studies may fail to inform. Current development of biologic markers of exposure to, and molecular effects of, PAHs represents one approach to improving epidemiologic methods. Markers of genetic susceptibility are also needed. In addition, and of equal importance, methods for the retrospective estimation of lifetime exposure to air pollutants should be developed and tested, so that large case-control and retrospective cohort studies can be feasibly conducted. These methods could combine time-activity information with data from national aerometric databases, such as those maintained by the US Environmental Protection Agency. This effort should include development of methods to characterize, quantify, and adjust for exposure measurement error. For lung cancer, urban and relatively unpolluted areas with established population-based tumor registries might be targeted. References Abbey DE, Lebowitz MD, Mills PK, Petersen FF, Beeson WL, Burchette RJ. 1995. Long-term concentrations of particulates and oxidants and development of chronic disease in a cohort of nonsmoking California residents. Inhal Toxicol 7:19–34. Anderson KE, Carmella SG, Ye M, Bliss RL, Le C, Murphy L, Hecht SS. 2001. Metabolites of a tobacco-specific lung carcinogen in nonsmoking women exposed to environmental tobacco smoke. J Natl Cancer Inst 93:378–381. Archer VE. 1990. Air pollution and fatal lung disease in three Utah counties. Arch Environ Health 45:325–334. Autrup H, Daneshvar B, Dragsted LO, Gamborg M, Hansen M, Loft S, Okkels H, Nielsen F, Nielsen PS, Raffn E, Wallin H, Knudsen LE. 1999. Biomarkers for exposure to ambient air pollution—comparison of carcinogen-DNA adduct levels with other exposure markers and markers for oxidative stress. Environ Health Perspect 107:233–238. Barbone F, Bovenzi M, Cavalleri F, Stanta G. 1995. Air pollution and lung cancer in Trieste, Italy. Am J Epidemiol 141:1161–1169. Beeson WL, Abbey DE, Knutsen SF. 1998. Long-term concentrations of ambient air pollutants and incident lung cancer in California adults: Results from the AHSMOG study. Adventist Health Study on Smog. Environ Health Perspect 106:813–823. Bertazzi A, Pesatori AC, Consonni D, Tironi A, Landi MT, Zocchetti C. 1993. Cancer incidence in a population accidentally exposed to 2,3,7,8tetrachlorodibenzo-para-dioxin. Epidemiology 4:398–406. Bhatia R, Lopipero P, Smith AH. 1998. Diesel exhaust exposure and lung cancer. Epidemiology 9:84–91. Birkett NJ. 1992. Effect of nondifferential misclassification on estimates of odds ratios with multiple levels of exposure. Am J Epidemiol 136:356–362. Boffetta P, Dosemeci M, Gridley G, Bath H, Moradi T, Silverman D. 2001. Occupational exposure to diesel engine emissions and risk of cancer in Swedish men and women. Cancer Causes Control 12:365–374. Boffetta P, Jourenkova N, Gustavsson P. 1997. Cancer risk from occupational and environmental exposure to polycyclic aromatic hydrocarbons. Cancer Causes Control 8:444–472. Boffetta P, Stellman SD, Garfinkel L. 1988. Diesel exhaust exposure and mortality among males in the American Cancer Society prospective study. Am J Ind Med 14:403–415. Bonassi S, Merlo F, Pearce N, Puntoni R. 1989. Bladder cancer and occupational exposure to polycyclic aromatic hydrocarbons. Int J Cancer 44:648–651. Brauer M, Hoek G, van Vliet P, Meliefste K, Fischer P, Gehring U, Heinrich J, Cyrys J, Bellander T, Lewne M, Brunekreef B. 2003. Estimating long-term average particulate air pollution concentrations: Application of traffic indicators and geographic information systems. Epidemiology 14:228–239.

Air Pollution Brownson RC, Alavanja MC. 2000. Previous lung disease and lung cancer risk among women (United States). Cancer Causes Control 11:853–858. Buell P, Dunn JE, Breslow L. 1967. Cancer of the lung and Los Angeles type air pollution. Cancer 20:2139–2147. Buffler PA, Cooper SP, Stinnett S, Contant C, Shirts S, Hardy RJ, Agu V, Gehan B, Burau K. 1988. Air pollution and lung cancer mortality in Harris County, Texas, 1979–1981. Am J Epidemiol 128:683–699. Burns PB, Swanson GM. 1991. Risk of urinary bladder cancer among blacks and whites: the role of cigarette use and occupation. Cancer Causes Control 2:371–379. California Environmental Protection Agency (Cal EPA). Identification of Particulate Emissions from Diesel-Fueled Engines as a Toxic Air Contaminant. Res. 98–35. 1998. Berkeley, CA, Office of Environmental Health Hazard Assessment, California Environmental Protection Agency. Camus M, Siemiatycki J, Meek B. 1998. Nonoccupational exposure to chrysotile asbestos and the risk of lung cancer. N Engl J Med 338: 1565–1571. Cass GR, Gray HA. 1995. Regional emissions and atmospheric concentrations of diesel engine particulate matter: Los Angeles as a case study. In: Health Effects of Diesel Engine Emissions: Characterization and Critical Analysis. Cambridge, MA: The Health Effects Institute, pp. 125–137. Cederlof R, Friberg L, Hrubec Z, Lorich U. 1975. The Relationship of Smoking and Some Social Covariables to Mortality and Cancer Morbidity. Stockholm: Karolinska Institute, Department of Environmental Hygiene. Chen BH, Hong CJ, Pandey MR, Smith KR. 1990. Indoor air pollution in developing countries. World Health Stat Q 43:127–138. Claude JC, Frentzel-Beyme RR, Kunze E. 1988. Occupation and risk of cancer of the lower urinary tract among men. A case-control study. Int J Cancer 41:371–379. Claxton LD, Morin RS, Hughes TJ, Lewtas J. 1989. A genotoxic assessment of environmental tobacco smoke using bacterial bioassays. Mutat Res 222:81–99. Cohen AJ, Anderson HR, Ostro B, Pandey KD, Krzyzanowski M, Kuenzli N, Gutschmidt K, Pope CA, Romieu I, Samet JM, Smith K. 2004. Mortality impacts of urban air pollution. In: Ezzati M, Lopez AD, Rodgers A, Murray CJL, eds. Comparative Quantification of Health Risks: Global and Regional Burden of Disease Attributable to Selected Major Risk Factors. Geneva: World Health Organization. Cohen AJ, Higgins MWP. 1995. Health Effects of Diesel Exhaust: Epidemiology. Diesel Exhaust: A Critical Analysis of Emissions, Exposure, and Health Effects. A Special Report of the Institute’s Working Group. Cambridge, MA: Health Effects Institute, pp. 251–292. Cohen AJ, Nikula K. 1999. The health effects of diesel exhaust: laboratory and epidemiologic studies. In: Holgate ST, Samet JM, Koren HS, Maynard RL, eds. Air Pollution and Health. San Diego: Academic Press, pp. 707–745. Colt JS, Blair A. 1998. Parental occupational exposures and risk of childhood cancer. Environ Health Perspect 106(Suppl 3):909–925. Cordier S, Clavel J, Limasset JC, Boccon-Gibod L, Le Moual N, Mandereau L, Hemon D. 1993. Occupational risks of bladder cancer in France: A multicentre case-control study. Int J Epidemiol 22:403–411. Cordier S, Mandereau L, Preston-Martin S, Little J, Lubin F, Mueller B, Holly E, Filippini G, Peris-Bonet R, McCredie M, Choi NW, Arsla A. 2001. Parental occupations and childhood brain tumors: results of an international case-control study. Cancer Causes Control 12:865–874. Coultas DB, Peake GT, Samet JM. 1989. Questionnaire assessment of lifetime and recent exposure to environmental tobacco smoke. Am J Epidemiol 130:338–347. Crawford FG, Mayer J, Santella RM, Cooper TB, Ottman R, Tsai WY, SimonCereijido G, Wang M, Tang D, Perera FP. 1994. Biomarkers of environmental tobacco smoke in preschool children and their mothers. J Natl Cancer Inst 86:1398–1402. Cummings KM, Markello SJ, Mahoney MC, Marshall JR. 1989. Measurement of lifetime exposure to passive smoke. Am J Epidemiol 130:122–132. Dean G. 1966. Lung cancer and bronchitis in Northern Ireland, 1960–2. Br Med J 1:1506–1514. Dean G, Lee PN, Todd GF, Wicken AJ. 1978. Report on a Second Retrospective Mortality Study in Northeast England. Parts 1 and 2. London: Tobacco Research Council. DeMarini DM, Shelton ML, Abu-Shakra A, Szakmary A, Levine JG. 1998. Spectra of spontaneous frameshift mutations at the hisD3052 allele of Salmonella typhimurium in four DNA repair backgrounds. Genetics 149:17–36. Dockery DW, Pope CA, III, Xu X, Spengler JD, Ware JH, Fay ME, Ferris BG, Jr., Speizer FE. 1993. An association between air pollution and mortality in six U.S. cities. N Engl J Med 329:1753–1759. Doll R. 1955. Mortality from lung cancer in asbestos workers. Br J Ind Med 12:81–86.

377

Doll R, Hill AB. 1950a. A study of the aetiology of carcinoma of the lung. Br Med J 2:740–748. Doll R, Hill AB. 1950b. Smoking and carcinoma of the lung. Br Med J 2:739–748. Doll R, Peto R. 1981. The Causes of Cancer. Oxford: Oxford University Press. Doll R, Vessey MP, Beasley RWR, Buckley AR, Fear EC, Fisher REW, Gammon EJ, Gunn W, Hughes GO, Lee K, Normon-Smith B. 1972. Mortality of gasworkers. Final report of a prospective study. Br J Ind Med 29:394–406. Dosemeci M, Washolder S, Lubin JH. 1990. Does nondifferential misclassification of exposure always bias a true effect toward the null value? Am J Epidemiol 132:746–748. Elliott P, Shaddick G, Kleinschmidt I, Jolley D, Walls P, Beresford J, Grundy C. 1996. Cancer incidence near municipal solid waste incinerators in Great Britain. Br J Cancer 73:702–710. Enstrom JE. 1979. Rising lung cancer mortality among nonsmokers. J Natl Cancer Inst 62:755–760. Feychting M, Svensson D, Ahlbom A. 1998. Exposure to motor vehicle exhaust and childhood cancer. Scand J Work Environ Health 24:8–11. Field RW, Smith BJ, Steck DJ, Lynch CF. 2002. Residential radon exposure and lung cancer: Variation in risk estimates using alternative exposure scenarios. J Expo Anal Environ Epidemiol 12:197–203. Flodin U, Fredriksson M, Persson B, Axelson O. 1988. Chronic lymphatic leukaemia and engine exhausts, fresh wood, and DDT: A case-referent study. Br J Ind Med 45:33–38. Fontham ETH, Correa P, Reynolds P, Wu-Williams A, Buffler PA, Greenberg RS, Chen VW, Alterman T, Boyd P, Austin DF, Liff J. 1994. Environmental tobacco smoke and lung cancer in nonsmoking women: A multicenter study. JAMA 271:1752–1759. Friedman GD, Petitti DB, Bawol RD. 1983. Prevalence and correlates of passive smoking. Am J Public Health 73:401–405. Gao YT, Blot WJ, Zheng W, Ershow AG, Hsu CW, Levin LI, Zhang R, Fraumeni JF, Jr. 1987. Lung cancer among Chinese women. Int J Cancer 40:604–609. Garfinkel L. 1981. Time trends in lung cancer mortality among nonsmokers and a note on passive smokers. J Natl Cancer Inst 66:1061–1066. Garshick E, Schenker MB, Muñoz A, Segal M, Smith TJ, Woskie SR, Hammond SK, Speizer FE. 1987. A case-control study of lung cancer and diesel exhaust exposure in railroad workers. Am Rev Respir Dis 135:1242–1248. Garshick E, Schenker MB, Muñoz A, Segal M, Smith TJ, Woskie SR, Hammond SK, Speizer FE. 1988. A retrospective cohort study of lung cancer and diesel exhaust exposure in railroad workers. Am Rev Respir Dis 137:820–825. Georgiadis P, Topinka J, Stoikidou M, Kaila S, Gioka M, Katsouyanni K, Sram R, Autrup H, Kyrtopoulos SA. 2001. Biomarkers of genotoxicity of air pollution (the AULIS project): Bulky DNA adducts in subjects with moderate to low exposures to airborne polycyclic aromatic hydrocarbons and their relationship to environmental tobacco smoke and other parameters. Carcinogenesis 22:1447–1457. Ger L-P, Hsu W-L, Chen K-T, Chen C-J. 1993. Risk factors of lung cancer by histological category in Taiwan. Anticancer Res 13:1491–1500. Gottlieb MS, Pickle LW. 1981. Bladder cancer mortality in Louisiana. J La State Med Soc 133:6–9. Greenberg A. 1987. Analyses of polycyclic aromatic hydrocarbons. In: Daisey JM, Lioy MJY, eds. Toxic Air Pollution: A Comprehensive Study of Non-Criteria Pollutants. Boca Raton, FL: Lewis Publishers, pp. 93–118. Greenland S. 1980. The effect of misclassification in the presence of covariates. Am J Epidemiol 112: 564–569. Hackshaw A. 2003. Passive smoking: paper does not diminish conclusion of previous reports. Br Med J 327:501–502. Hackshaw AK, Law MR, Wald NJ. 1997. The accumulated evidence on lung cancer and environmental tobacco smoke. Br Med J 315:980–988. Haenszel W, Loveland DB, Sirken MG. 1962. Lung-cancer mortality as related to residence and smoking histories. I. White Males. J Natl Cancer Inst 28:947–1001. Haenszel W, Taeuber KE. 1964. Lung-cancer mortality as related to residence and smoking history: II. White females. J Natl Cancer Inst 32:803–838. Hammond EC. 1972. Smoking habits and air pollution in relation to lung cancer. In: Lee DHK, ed. Environmental Factors in Respiratory Disease. New York: Academic Press, pp. 177–198. Hammond EC, Horn D. 1958. Smoking and death rates report on forty-four months of follow-up of 187,783 men. Part I. Total mortality. Part II. Death rates by cause. JAMA 166:1294–1308. Hammond EC, Selikoff IJ, Seidman H. 1979. Asbestos exposure, cigarette smoking and death rates. Ann NY Acad Sci 330:473–490. Harrison RM, Leung PL, Somervaille L, Smith R, Gilman E. 1999. Analysis of incidence of childhood cancer in the West Midlands of the United

378

PART III: THE CAUSES OF CANCER

Kingdom in relation to proximity to main roads and petrol stations. Occup Environ Med 56:774–780. Hatch MC, Beyea J, Nieves JW, Susser M. 1990. Cancer near the Three Mile Island nuclear plant: radiation emissions. Am J Epidemiol 132:397–412; discussion 413–417. Hatch MC, Wallenstein S, Beyea J, Nieves JW, Susser M. 1991. Cancer rates after the Three Mile Island nuclear accident and proximity of residence to the plant. Am J Public Health 81:719–724. Health Effects Institute. 1993. Research priorities for mobile air toxics. [HEI Communication Number 2]. Cambridge, Ma: Health Effects Institute. Health Effects Institute, Asbestos Research Committee, Literature Review Panel. 1991. Asbestos in public and commercial buildings: A literature review and a synthesis of current knowledge. Cambridge, MA: Health Effects Institute. Health Effects Institute, Diesel Epidemiology Expert Panel. 1999. Diesel emissions and lung cancer: Epidemiology and quantitative risk assessment. Special Report. Cambridge, MA: Health Effects Institute. Health Effects Institute, Diesel Epidemiology Working Group. 2002. Research directions to improve estimates of human exposure and risk from diesel exhaust. Special Report. Boston: Health Effects Institute. Hecht SS, Carmella SG, Murphy SE, Akerkar S, Brunneman KD, Hoffmann D. 1993. A tobacco-specific lung carcinogen in the urine of men exposed to cigarette smoke. N Engl J Med 329:1543–1546. Hei TK, Piao CQ, Willey JC, Thomas S, Hall EJ. 1994. Malignant transformation of human bronchial epithelial cells by radon-simulated alphaparticles. Carcinogenesis 15:431–437. Hei TK, Wu LJ, Liu SX, Vannais D, Waldren CA, Randers-Pehrson G. 1997. Mutagenic effects of a single and an exact number of alpha particles in mammalian cells. Proc Natl Acad Sci U S A 94:3765–3770. Henderson BE, Gordon RJ, Menck H, Soohoo J, Martin SP, Pike MC. 1975. Lung cancer and air pollution in Southcentral Los Angeles County. Am J Epidemiol 101:477–488. Hillerdal G. 1996. Long-term survival of patients with lung cancer from a defined geographical area before and after radiological screening. Lung Cancer 15:21–30. Hirayama T. 1981. Non-smoking wives of heavy smokers have a higher risk of lung cancer: A study from Japan. Br Med J (Clin Res Ed) 282:183–185. Hirayama T. 1984. Cancer mortality in nonsmoking women with smoking husbands based on a large-scale cohort study in Japan. Prev Med 13: 680–690. Hitosugi M. 1968. Epidemiological study of lung cancer with special reference to the effect of air pollution and smoking habit. Bull Inst Public Health 17:236–255. Hoar SK, Hoover R. 1985. Truck driving and bladder cancer mortality in rural New England. J Natl Cancer Inst 74:771–774. Hoek G, Brunekreef B, Goldbohm S, Fischer P, van den Brandt PA. 2002. Association between mortality and indicators of traffic-related air pollution in the Netherlands: A cohort study. Lancet 360:1203–1209. Hoffman FL. 1929. Cancer of the lungs. American Review of Tuberculosis 19:392. Hoffman FL. 1931. Cancer and smoking habits. Ann Surg 93:50–67. Holowaty EJ, Risch HA, Miller AB, Burch JD. 1991. Lung cancer in women in the Niagara Region, Ontario: A case-control study. Can J Public Health 82:304–309. Howe GR, Burch JD, Miller AB, Cook GM, Esteve J, Morrison B, Gordon P, Chambers LW, Fodor G, Winsor GM. 1980. Tobacco use, occupation, coffee, various nutrients, and bladder cancer. J Natl Cancer Inst 64:701–713. Howe GR, Fraser D, Lindsay J, Presnal B, Yu SZ. 1983. Cancer mortality (1965–77) in relation to diesel fume and coal exposure in a cohort of retired railway workers. J Natl Cancer Inst 70:1015–1019. International Agency for Research on Cancer (IARC). 1984. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans: Polynuclear aromatic compounds, Part 3. Industrial exposures in aluminum production, coal gasification, coke production, and iron and steel founding. [34]. Lyons, France: International Agency for Research on Cancer. International Agency for Research on Cancer (IARC). 1986. IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans: Tobacco Smoking. [Monograph 38]. Lyon, France: World Health Organization, IARC. International Agency for Research on Cancer (IARC). 1987. IARC Monographs on the evaluation of carcinogenic risks to humans: Supplement 7, overall evaluations of carcinogenicity, an updating of IARC monographs. [1–42]. Lyons, France: International Agency for Research on Cancer. International Agency for Research on Cancer (IARC). 1989. IARC Monographs on the evaluation of carcinogenic risks to humans: Diesel and gasoline engine exhausts and some nitroarenes. [Volume 46]. Lyons, France: International Agency for Research on Cancer.

International Agency for Research on Cancer (IARC). 1992. IARC Monographs on the evaluation of carcinogenic risks to humans: occupational exposures to mists and vapours from strong inorganic acids; and other industrial chemicals. [54]. Lyons, France: International Agency for Research on Cancer. International Agency for Research on Cancer (IARC). 1995. IARC Monographs on the evaluation of carcinogenic risks to humans: wood dust and formaldehyde. [62]. Lyons, France: International Agency for Research on Cancer. International Agency for Research on Cancer (IARC). 1997. IARC Monographs on the evaluation of carcinogenic risks to humans. Volume 68. Silica, some silicates, coal dust and para-aramid fibrils. Lyon, France: World Health Organization, IARC. International Agency for Research on Cancer (IARC). 2002. Tobacco smoke and involuntary smoking. IARC monograph 83. [83]. Lyon, France: International Agency for Research on Cancer. Iscovich J, Castelletto R, Esteve J, Munoz N, Colanzi R, Coronel A, Deamezola I, Tassi V, Arslan A. 1987. Tobacco smoking, occupational exposure and bladder cancer in Argentina. Int J Cancer 40:734–740. Iyer V, Harris RE, Wynder EL. 1990. Diesel exhaust exposure and bladder cancer risk. Eur J Epidemiol 6:49–54. Jaakkola MS, Samet JM. 1999. Summary: Workshop on health risks attributable to ETS exposure in the workplace. Environ Health Perspect 107(Suppl 6):823–828. Jedrychowski W, Becher H, Wahrendorf J, Basa-Cierpialek Z. 1990. A casecontrol study of lung cancer with special reference to the effect of air pollution in Poland. J Epidemiol Community Health 44:114–120. Jensen OM, Wahrendorf J, Knudsen JB, Sorenson BL. 1987. The Copenhagen case-referent study on bladder cancer. Risks among drivers, painters and certain other occupations. Scand J Work Environ Health 13:129– 134. Karch NJ, Schneiderman MA. 1981. Explaining the Urban Factor in Lung Cancer Mortality. A Report of the Natural Resources Defense Council. Washington, D.C.: Clement Associates, Inc. Kinney PL, Aggarwal M, Northridge ME, Janssen NA, Shepard P. 2000. Airborne concentrations of PM (2.5) and diesel exhaust particles on Harlem sidewalks: A community-based pilot study. Environ Health Perspect 108:213–218. Kinney PL, Chillrud SN, Ramstrom S, Ross J, Spengler JD. 2002. Exposures to multiple air toxics in New York City. Environ Health Perspect 110(Suppl 4):539–546. Koutrakis P, Sioutas C. 1996. Physico-chemical properties and measurement of ambient particles. In: Spengler J, Wilson R, eds. Particles in our air, concentrations in health effects. Cambridge, MA: Harvard University Press. Krewski D, Burnett RT, Goldberg MS, Hoover K, Siemiatycki J, Abrahamowicz M, White WH, et al. 2000. Reanalysis of the Harvard Six Cities Study and the American Cancer Society Study of particulate air pollution and mortality. Investigators’ reports parts I and II. Cambridge, MA: Health Effects Institute. Kunze E, Chang-Claude J, Frentzel-Beyme R. 1992. Life style and occupational risk factors for bladder cancer in Germany. A case-control study. Cancer 69:1776–1790. Kyrtopoulos SA, Georgiadis P, Autrup H, Demopoulos NA, Farmer P, Haugen A, Katsouyanni K, Lambert B, Ovrebo S, Sram R, Stephanou G, Topinka J, Stefanou G, Demopoulos N. 2001. Biomarkers of genotoxicity of urban air pollution. Overview and descriptive data from a molecular epidemiology study on populations exposed to moderate-to-low levels of polycyclic aromatic hydrocarbons: the AULIS project. Mutat Res 496: 207–228. Lan Q, Chapman RS, Schreinemachers DM, Tian L, He X. 2002. Household stove improvement and risk of lung cancer in Xuanwei, China. J Natl Cancer Inst 94:826–835. Langholz B, Ebi KL, Thomas DC, Peters JM, London SJ. 2002. Traffic density and the risk of childhood leukemia in a Los Angeles case-control study. Ann Epidemiol 12:482–487. Larkin EK, Smith TJ, Stayner L, Rosner B, Speizer FE, Garshick E. 2000. Diesel exhaust exposure and lung cancer: adjustment for the effect of smoking in a retrospective cohort study. Am J Ind Med 38: 399–409. Lee PN. 1988. Misclassification of Smoking Habits and Passive Smoking. Berlin: Springer Verlag. Lee PN. 1995. “Marriage to a smoker” may not be a valid marker of exposure in studies relating environmental tobacco smoke to risk of lung cancer in Japanese non-smoking women. Int Arch Occup Environ Health 67:287–294. Levin ML, Goldstein H, Gerhardt PR. 1950. Cancer and tobacco smoking. A preliminary report. JAMA 143:336–338.

Air Pollution Levy JI, Houseman EA, Ryan L, Richardson D, Spengler JD. 2000. Particle concentrations in urban microenvironments. Environ Health Perspect 108:1051–1057. Lipsett M, Campleman S. 1999. Occupational exposure to diesel exhaust and lung cancer: a meta-analysis. Am J Public Health 89:1009–1017. Lofroth G. 1989. Environmental tobacco smoke: overview of chemical composition and genotoxic components. Mutat Res 222:73–80. Lubin JH. 1999. Estimating lung cancer risk with exposure to environmental tobacco smoke. Environ Health Perspect 107:879–883. Lubin JH, Boice JD, Jr., Edling C, Hornung RW, Howe GR, Kunz E, Kusiak RA, Morrison HI, Radford EP, Samet JM, Tirmarche M, Woodward A, Yao SX, Pierce DA. 1995a. Lung cancer in radon-exposed miners and estimation of risk from indoor exposure. J Natl Cancer Inst 87:817–827. Lubin JH, Boice JD, Jr., Samet JM. 1995b. Errors in exposure assessment, statistical power, and the interpretation of residential radon studies. Radiat Res 144:329–341. Lynch KM, Smith WA. 1939. Pulmonary asbestosis. V. A report of bronchial carcinoma and epithelial metaplasia. Am J Cancer 36:567–573. Macklin MT, Macklin CC. 1940. Does chronic irritation cause primary carcinoma of the human lung? Arch Pathol 30:924–955. Maclure M, Katz RB, Bryant MS, Skipper PL, Tannenbaum SR. 1989. Elevated blood levels of carcinogens in passive smokers. Am J Public Health 89:1381–1384. Malker HS, McLaughlin JK, Silverman DT, Ericsson JL, Stone BJ, Weiner JA, Malker BK, Blot WJ. 1987. Occupational risks for bladder cancer among men in Sweden. Cancer Res 47:6763–6766. Matanoski G, Kanchanaraksa S, Lantry D, Chang Y. 1995. Characteristics of nonsmoking women in NHANES I and NHANES I epidemiologic follow-up study with exposure to spouses who smoke. Am J Epidemiol 142:149–157. Mauderly JL. 1997. Relevance of particle-induced rat lung tumors for assessing lung carcinogenic hazard and human lung cancer risk. Environ Health Perspect 105:1337–1346. McDonnell WF, Nishino-Ishikawa N, Petersen FF, Chen LH, Abbey DE. 2000. Relationships of mortality with the fine and coarse fractions of long-term ambient PM10 concentrations in nonsmokers. J Expo Anal Environ Epidemiol 10:427–436. Mills PK, Abbey D, Beeson WL, Petersen F. 1991. Ambient air pollution and cancer in California Seventh-day Adventists. Arch Environ Health 46:271–280. Mumford JL, He XZ, Chapman RS, Cao SR, Harris DB, Li XM, Xian YL, Jiang WZ, Xu CW, Chuang JC, et al. 1987. Lung cancer and indoor air pollution in Xuan Wei, China. Science 235:217–220. Mumford JL, Helmes CT, Lee XM, Seidenberg J, Nesnow S. 1990. Mouse skin tumorigenicity studies of indoor coal and wood combustion emissions from homes of residents in Xuan Wei, China with high lung cancer mortality. Carcinogenesis 11:397–403. Nafstad P, Haheim LL, Oftedal B, Gram F, Holme I, Hjermann I, Leren P. 2003. Lung cancer and air pollution: a 27 year follow up of 16 209 Norwegian men. Thorax 58:1071–1076. National Research Council (NRC), Committee on Advances in Assessing Human Exposure to Airborne Pollutants. 1991. Human Exposure Assessment for Airborne Pollutants: Advances and Opportunities. Washington, D.C.: National Academy Press. National Research Council (NRC), Committee on Health Risks of Exposure to Radon, Board on Radiation Effects Research, Commission on Life Sciences. 1998. Health Effects of Exposure to Radon (BEIR VI). Washington, D.C.: National Academy Press. National Research Council (NRC), Committee on Passive Smoking. 1986. Environmental Tobacco Smoke: Measuring Exposures and Assessing Health Effects. Washington, D.C.: National Academy Press. National Research Council (NRC), Committee on the Biological Effects of Ionizing Radiation. 1988. Health Risks of Radon and Other Internally Deposited Alpha-Emitters: BEIR IV. Washington, D.C.: National Academy Press. National Research Council (NRC), Committee on the Biological Effects of Ionizing Radiation. 1990. Health Effects of Exposure to Low Levels of Ionizing Radiation: BEIR V. Washington, D.C.: National Academy Press. National Research Council (NRC), Panel on Dosimetric Assumptions Affecting the Application of Radon Risk Estimates. 1991. Comparative Dosimetry of Radon in Mines and Homes. Companion to BEIR IV Report. Washington, D.C.: National Academy Press. Natusch DF. 1978. Potentially carcinogenic species emitted to the atmosphere by fossil-fueled power plants. Environ Health Perspect 22:79–90. Nauss K, The Diesel Working Group. 1995. Critical issues in assessing the carcinogenicity of diesel exhaust: A synthesis of current knowledge. In: Diesel Exhaust: A Critical Analysis of Emissions, Exposure, and Health Effects. Cambridge, MA: The Health Effects Institute, pp. 11–61.

379

Nordlinder R, Jarvholm B. 1997. Environmental exposure to gasoline and leukemia in children and young adults—an ecology study. Int Arch Occup Environ Health 70:57–60. Nyberg F, Gustavsson P, Jarup L, Bellander T, Berglind N, Jakobsson R, Pershagen G. 2000. Urban air pollution and lung cancer in Stockholm. Epidemiology 11:487–495. Overholt RH, Rumel WR. 1940. Clinical studies of primary carcinoma of the lung. An analysis of seventy-five cases, twenty-one of which were treated by pneumonectomy or lobectomy. JAMA 114:735–742. Pandey KD, Wheeler D, Ostro B, Deichmann U, Hamilton K, Bolt K. 2004. Ambient particulate matter concentrations in residential areas of world cities: new estimates based on global model of ambient particulates (GMAPS). The Development Research Group and the Environment Department, World Bank, Washington, DC. Paradis G, Theriault G, Tremblay C. 1989. Mortality in a historical cohort of bus drivers. Int J Epidemiol 18:397–402. Pearson RL, Wachtel H, Ebi KL. 2000. Distance-weighted traffic density in proximity to a home is a risk factor for leukemia and other childhood cancers. J Air Waste Manag Assoc 50:175–180. Perera FP, Hemminki K, Gryzbowska E, Motykiewicz G, Michalska J, Santella RM, Young TL, Dickey C, Brandt-Rauf P, DeVivo I, et al. 1992. Molecular and genetic damage in humans from environmental pollution in Poland [published erratum appears in Nature 1993 Feb 11; 361(6412):564]. Nature 360:256–258. Pershagen G. 1990. Air pollution and cancer. In: Vainio H, Sorsa M, McMichael AJ, eds. Complex Mixtures and Cancer Risk. Lyon, France: International Agency for Research on Cancer, pp. 240–251. Pershagen G, Elinder CG, Bolander AM. 1977. Mortality in a region surrounding an arsenic smelting plant. Environ Health Perspect 19:133–137. Peto R, Lopez AD. 2003. Smoking Data: Clinical Trial Service Unit. Clinical Trial Service Unit & Epidemiological Studies Unit: Oxford University. Pope CA, III, Burnett RT, Thun MJ, Calle EE, Krewski D, Ito K, Thurston GD. 2002. Lung cancer, cardiopulmonary mortality, and long-term exposure to fine particulate air pollution. JAMA 287:1132–1141. Pope CA, III, Thun MJ, Namboodiri MM, Dockery DW, Evans JS, Speizer FE, Heath CW, Jr. 1995. Particulate air pollution as a predictor of mortality in a prospective study of U.S. adults. Am J Resp Crit Care Med 151:669–674. Porru S, Aulenti V, Donato F, Boffetta P, Fazioli R, Cosciani-Cunico S, Alessio L. 1996. Bladder cancer and occupation: A case-control study in northern Italy. Occup Environ Med 53:6–10. Proctor RN. 1999. The Nazi War on Cancer. Princeton, NJ: Princeton University Press. Pron GE, Burch JD, Howe GR, Miller AB. 1988. The reliability of passive smoking histories reported in a case-control study of lung cancer. Am J Epidemiol 127:267–273. Raaschou-Nielsen O, Hertel O, Thomsen BL, Olsen JH. 2001. Air pollution from traffic at the residence of children with cancer. Am J Epidemiol 153:433–443. Redmond CK. 1983. Cancer mortality among coke oven workers. Environ Health Perspect 52:67–73. Repace JL, Lowrey AH. 1990. Risk assessment methodologies for passive smoking-induced lung cancer. Risk Anal 10:27–37. Reynolds P, Elkin E, Scalf R, Von Behren J, Neutra RR. 2001. A case-control pilot study of traffic exposures and early childhood leukemia using a geographic information system. Bioelectromagnetics(Suppl 5):S58–S68. Reynolds P, Von Behren J, Gunier RB, Goldberg DE, Hertz A, Smith D. 2002. Traffic patterns and childhood cancer incidence rates in California, United States. Cancer Causes Control 13:665–673. Reynolds P, Von Behren J, Gunier RB, Goldberg DE, Hertz A, Smith DF. 2003. Childhood cancer incidence rates and hazardous air pollutants in California: An exploratory analysis. Environ Health Perspect 111:663–668. Risch HA, Burch JD, Miller AB, Hill GB, Steele R, Howe GR. 1988. Occupational factors and the incidence of cancer of the bladder in Canada. Br J Ind Med 45:361–367. Rothman KJ, Greenland S. 1998. Modern Epidemiology. Philadelphia: Lippincott-Raven. Rushton L, Alderson MR, Nagarajah CR. 1983. Epidemiological survey of maintenance workers in London Transport Executive bus garages and Chiswick Works. Br J Ind Med 40:340–345. Samet JM. 1994. Epidemiology of Lung Cancer. New York: Marcel Dekker, Inc. Samet JM, Humble CG, Skipper BE, Pathak DR. 1987. History of residence and lung cancer risk in New Mexico. Am J Epidemiol 125:800–811. Samet JM, Jaakkola JJK. 1999. The epidemiologic approach to investigating outdoor air pollution. In: Holgate ST, Samet JM, Koren HS, Maynard RL, eds. Air Pollution and Health. San Diego: Academic Press, pp. 431–460.

380

PART III: THE CAUSES OF CANCER

Samet JM, Wang SS. 2000. Environmental Tobacco Smoke. In: Lippmann M, ed. Environmental Toxicants: Human Exposures and Their Health Effects. New York: Van Nostrand Reinhold Company, Inc., pp. 319–375. Sapkota A, Buckley TJ. 2003. The mobile source effect on curbside 1,3butadiene, benzene, and particle-bound polycyclic aromatic hydrocarbons assessed at a tollbooth. J Air Waste Manag Assoc 53:740–748. Saracci R, Riboli E. 1989. Passive smoking and lung cancer: Current evidence and ongoing studies at the International Agency for Research on Cancer. Mutat Res 22:117–127. Sasco AJ, Merrill RM, Dari I, Benhaim-Luzon V, Carriot F, Cann CI, Bartal M. 2002. A case-control study of lung cancer in Casablanca, Morocco. Cancer Causes Control 13:609–616. Savitz DA, Feingold L. 1989. Association of childhood cancer with residential traffic density. Scand J Work Environ Health 15:360–363. Schenker MB, Smith TF, Muñoz A, Woskie SR, Speizer FE. 1984. Diesel exposure and mortality among railway workers: Result of a pilot study. Br J Ind Med 41:320–327. Schoenberg JB, Stemhagen A, Mogielnicki AP, Altman R, Abe T, Mason TJ. 1984. Case-control study of bladder cancer in New Jersey. I. Occupational exposures in white males. J Natl Cancer Inst 72:973–981. Schumacher MC, Slattery ML, West DW. 1989. Occupation and bladder cancer in Utah. Am J Ind Med 16:89–102. Scientific Committee on Tobacco and Health, HSMO. Report of the Scientific Committee on Tobacco and Health. 011322124x. 2-5-1998. The Stationary Office. Selikoff IJ. 1980. Latency of asbestos disease among insulation workers in the United States and Canada. Cancer 46:2736–2740. Selikoff IJ, Churg J, Hammond EC. 1964. Asbestos exposure and neoplasia. JAMA 188:22–26. Selikoff IJ, Hammond EC, Seidman H. 1979. Mortality experience of insulation workers in the United States and Canada. Ann NY Acad Sci 330:91–116. Siemiatycki J, Dewar R, Nadon L, Gerin M. 1994. Occupational risk factors for bladder cancer: results from a case-control study in Montreal, Quebec, Canada. Am J Epidemiol 140:1061–1080. Siemiatycki J, Gerin M, Stewart P, Nadon L, Dewar R, Richardson L. 1988. Associations between several sites of cancer and ten types of exhaust and combustion products. Scand J Work Environ Health 14:79–90. Silverman DT, Hoover RN, Albert S, Graff KM. 1983. Occupation and cancer of the lower urinary tract in Detroit. J Natl Cancer Inst 70:237–245. Silverman DT, Hoover RN, Mason TJ, Swanson GM. 1986. Motor exhaustrelated occupations and bladder cancer. Cancer Res 46:2113–2116. Silverman DT, Levin LI, Hoover RN. 1989a. Occupational risks of bladder cancer in the United States: II Nonwhite men. J Natl Cancer Inst 81:1480–1483. Silverman DT, Levin LI, Hoover RN, Hartge P. 1989b. Occupational risks of bladder cancer in the United States: I. White men. J Natl Cancer Inst 81:1472–1480. Smith EL, Miller ER, Woolson RF, Brown CK. 1985. Bladder cancer risk among auto and truck mechanics and chemically related occupations. Am J Public Health 75:881–883. Smith KR. 1988. Air pollution, assessing total exposure in developing countries. Environment 30:16. Smith KR. 2000. Inaugural article: national burden of disease in India from indoor air pollution. Proc Natl Acad Sci U S A 97:13286–13293. Smith KR, Liu Y. 1994. Indoor air pollution in developing countries. In: Samet JM, ed. Epidemiology of Lung Cancer. New York: Marcel Dekker, Inc., pp. 151–184. Soll-Johanning H, Bach E, Jensen SS. 2003. Lung and bladder cancer among Danish urban bus drivers and tramway employees: A nested case-control study. Occup Med (Lond) 53:25–33. Soll-Johanning H, Bach E, Olsen JH, Tuchsen F. 1998. Cancer incidence in urban bus drivers and tramway employees: A retrospective cohort study. Occup Environ Med 55: 594–598. Sorensen M, Autrup H, Hertel O, Wallin H, Knudsen LE, Loft S. 2003. Personal exposure to PM2.5 and biomarkers of DNA damage. Cancer Epidemiol Biomarkers Prev 12:191–196. Spengler JD. 1991. Sources and concentrations of indoor air pollution. In: Samet JM, Spengler JD, eds. Indoor Air Pollution. A Health Perspective. Baltimore: Johns Hopkins University Press, pp. 33–67. Steck DJ, Alavanja MC, Field RW, Parkhurst MA, Bates DJ, Mahaffey JA. 2002. 210Po implanted in glass surfaces by long term exposure to indoor radon. Health Phys 83:261–271. Steenland K, Burnett C, Osorio AM. 1987. A case-control study of bladder cancer using city directories as a source of occupational data. Am J Epidemiol 126:247–257. Steenland NK, Silverman DT, Hornung RW. 1990. Case-control study of lung cancer and truck driving in the Teamsters Union. Am J Public Health 80:670–674.

Steineck G, Plato N, Gerhardsson M, Norell SE, Hogstedt C. 1990. Increased risk of urothelial cancer in Stockholm during 1985–87 after exposure to benzene and exhausts. Int J Cancer 45:1012–1017. Stidley CA, Samet JM. 1993. A review of ecologic studies of lung cancer and indoor radon. Health Phys 93:234–251. Stocks P, Campbell JM. 1955. Lung cancer death rates among nonsmokers and pipe and cigarette smokers. An evaluation in relation to air pollution by benzo[a]pyrene and other substances. Br Med J 2:923–929. Tomatis L. Air Pollution and Human Cancer. 1990. Berlin: Springer-Verlag. European School of Oncology Monographs. Trichopoulos D, Kalandidi A, Sparros L. 1983. Lung cancer and passive smoking: conclusion of Greek study. Lancet 2:677–678. Trichopoulos D, Kalandidi A, Sparros L, MacMahon B. 1981. Lung cancer and passive smoking. Int J Cancer 27:1–4. Trichopoulos D, Mollo F, Tomatis L, Agapitos E, Delsedime L, Zavitsanos X, Kalandidi A, Katsouyanni K, Riboli E, Saracci R. 1992. Active and passive smoking and pathological indicators of lung cancer risk in an autopsy study. JAMA 268:1697–1701. US Department of Health and Human Services (USDHHS). 1982. The health consequences of smoking: Cancer. A report of the Surgeon General. DHHS Publication No. (PHS) 82-50179. Washington, D.C.: U.S. Department of Health and Human Services, Public Health Service, Office on Smoking and Health. US Department of Health and Human Services (USDHHS). 1986. The health consequences of involuntary smoking: A report of the Surgeon General. DHHS Publication No. (CDC) 87-8398. Washington, D.C.: U.S. Government Printing Office. US Department of Health and Human Services (USDHHS). 1989. Reducing the health consequences of smoking. 25 years of progress. A report of the Surgeon General. Washington, D.C.: U.S. Government Printing Office. US Environmental Protection Agency (EPA). 1990. Cancer Risk from Outdoor Exposure to Air Toxics. Volume 1, Final Report. EPA-450/1-90-004a. Research Triangle Park, North Carolina, U.S. Government Printing Office. US Environmental Protection Agency (EPA). 1992a. Respiratory health effects of passive smoking: Lung cancer and other disorders. EPA/600/006F. Washington, D.C.: U.S. Government Printing Office. US Environmental Protection Agency (EPA). 1992b. Technical Support Document for the 1992 Citizen’s Guide to Radon. Washington, D.C.: U.S. Government Printing Office. US Environmental Protection Agency (EPA). 1993. National Air Quality and Emissions Trends Report, 1992. EPA-454/R-93-031. Research Triangle Park, North Carolina, Office of Air and Radiation/Office of Air Quality Planning and Standards. US Environmental Protection Agency (EPA). 1999. Integrated risk information system (IRIS) on asbestos. Washington, D.C.: National Center for Environmental Assessment, Office of Research and Development. US Environmental Protection Agency (EPA). 2003. Latest findings on national air quality. 2002 status and trends. EPA 454/K-03-001. Research Triangle Park, NC, US Environmental Protection Agency, Office of Air Quality Planning and Standards. US Environmental Protection Agency (EPA), National Center for Environmental Assessment. 2002. Health Assessment Document for Diesel Engine Exhaust. EPA/600/8-90/057F. Washington, D.C.: USEPA. US Environmental Protection Agency (EPA), Office of Radiation and Indoor Air. 2003. EPA assessment of risks from radon in homes. Washington, D.C.: USEPA. Vena JE. 1982. Air pollution as a risk factor in lung cancer. Am J Epidemiol 116:42–56. Wald NJ, Nanchahal K, Thompson SG, Cuckle HS. 1986. Does breathing other people’s tobacco smoke cause lung cancer? Br Med J (Clin Res Ed) 86:1217–1222. Wedler HW. 1944. Asbestosis and pulmonary carcinoma (Asbestose und lungenkrebs) [abstract]. Bulletin of Hygiene 19:362. Weiss SJ. 1989. Tissue destruction by neutrophils. N Engl J Med 320:365–376. Weiss ST. 1986. Passive smoking and lung cancer. What is the risk? Am Rev Respir Dis 133:1–3. Winer AM, Busby Jr. WF. 1995. Atmospheric transport and transformation of diesel emissions. Diesel exhaust: A critical analysis of emissions, exposure, and health effects. 1995. Cambridge, MA: The Health Effects Institute, pp. 83–105. Wong O, Morgan RW, Kheifets L, Larson SR, Whorton MD. 1985. Mortality among members of a heavy construction equipment operators union with potential exposure to diesel exhaust emissions. Br J Ind Med 42: 435–448. World Health Organization. 1988. Emissions of heavy metal and PAH compounds from municipal solid waste incinerators: control technology and health effects. Report on a WHO meeting, Florence. Copenhagen: WHO.

Air Pollution World Health Organization. 1990. Global estimates for health situation assessment and projections. HST/90.2. Geneva: WHO. World Health Organization. 1996. Diesel fuel and exhaust emissions. Environmental Health Criteria 171. International Programme on Chemical Safety. Geneva: WHO. World Health Organization. 2002. Global estimates of burden of disease caused by the environment. Geneva: WHO. Woskie SR, Smith TJ, Hammond SK, Schenker MB, Garshick E, Speizer FE. 1988a. Estimation of the diesel exhaust exposure of railroad workers: II. National and historical exposures. Am J Ind Med 13:395–404. Woskie SR, Smith TJ, Hammond SK, Schenker MB, Garshick E, Speizer FE. 1988b. Estimation of the diesel exhaust exposures of railroad workers: I. Current exposures. Am J Ind Med 13:381–394. Wu AH. 1999. Exposure misclassification bias in studies of environmental tobacco smoke and lung cancer. Environ Health Perspect 107:873–877. Wu AH, Henderson BE, Pike MC, Yu MC. 1985. Smoking and other risk factors for lung cancer in women. J Natl Cancer Inst 85:747–751. Wu-Williams AH, Dai XD, Blot W, Xu ZY, Sun XW, Xiao HP, Stone BJ, Yu SF, Feng YP, Ershow AG, et al. 1990. Lung cancer among women in northeast China. Br J Cancer 62:982–987. Wynder EL, Dieck GS, Hall NE, Lahti H. 1985. A case-control study of diesel exhaust exposure and bladder cancer. Environ Res 37:475–489.

381

Wynder EL, Graham EA. 1950. Tobacco smoking as a possible etiologic factor in bronchiogenic carcinoma. A study of six hundred and eighty-four proved cases. JAMA 143:329–336. Xu ZY, Blot WJ, Xiao HP, Wu A, Feng YP, Stone BJ, Sun J, Ershow AG, Henderson BE, Fraumeni JF, Jr. 1989. Smoking, air pollution, and the high rates of lung cancer in Shenyang, China. J Natl Cancer Inst 81: 1800–1806. Zaebst DD, Clapp DE, Blade LM, Marlow DA, Steenland K, Hornung RW, Scheutzle D, Butler J. 1991. Quantitative determination of trucking industry workers’ exposures to diesel exhaust particles. Am Ind Hyg Assoc J 52:529–541. Zatloukal P, Kubik A, Pauk N, Tomasek L, Petruzelka L. 2003. Adenocarcinoma of the lung among women: risk associated with smoking, prior lung disease, diet and menstrual and pregnancy history. Lung Cancer 41:283–293. Zeegers MP, Swaen GM, Kant I, Goldbohm RA, van den Brandt PA. 2001. Occupational risk factors for male bladder cancer: results from a population based case cohort study in the Netherlands. Occup Environ Med 58:590–596. Zhang J, Smith KR. 1996. Hydrocarbon emissions and health risks from cookstoves in developing countries. J Expo Anal Environ Epidemiol 6:147–161.

20

Water Contaminants KENNETH P. CANTOR, MARY H. WARD, LEE E. MOORE, AND JAY H. LUBIN

A

lthough drinking water quality in industrialized nations is considered the finest in the world, there is concern that various waterborne contaminants may contribute to the overall burden of environmental carcinogenesis. This awareness is not new. In a classic report published over four decades ago, W.C. Hueper (1960) warned of the potential carcinogenic dangers presented by the increased use of drinking water contaminated with natural and man-made pollutants. This concern has been fueled by more recent reports showing the widespread distribution of many proved or suspected carcinogens in finished drinking water, predominantly in trace amounts. Further evidence comes from epidemiologic studies linking some contaminants with an elevated risk of cancer. In order to assess water contaminants that may contribute to the human cancer burden, and if so, the potential for prevention, we will address the epidemiologic evidence for several contaminants and include information on their levels and environmental distribution, as well as individual susceptibility, where data exist. We will also discuss future research priorities and mention strategies for prevention. Other approaches to assessing the impact of drinking water quality in environmental carcinogenesis, such as extrapolation of dose-response data from animal studies, will not be covered. The three categories of drinking water contaminants that may be carcinogenic and that have been studied most systematically are arsenic, disinfection by-products, and nitrate. In addition, radionuclides, microbiological agents, organic compounds from human commerce, and asbestiform particles have been reported to cause cancer, either as they occur in drinking water or in other media, giving rise to suspicion about their carcinogenicity when ingested. This chapter is organized according to exposure. Although these groups are discussed separately, drinking water is a complex mixture of agents that may interact in various ways.

INORGANIC ARSENIC Exposure Arsenic is a widely distributed metalloid with a mean concentration of about 2 mg/kg in the earth’s crust (World Health Organization, 2001b). In natural waters, inorganic arsenic occurs primarily as pentavalent arsenate, As(V), or trivalent arsenite, As(III) (World Health Organization, 2001a). The greatest public health concern derives from the ingestion of drinking water contaminated with inorganic arsenic from geologic formations (National Research Council: Subcommittee on Arsenic in Drinking Water, 1999). Decreasing use of arsenical pesticides has resulted in lower direct and waterborne exposures (Bates et al., 1992). Until the 1940s, arsenical medications were a staple of the pharmacopeia, and patients were treated for a wide range of ailments, especially dermatological conditions (Cuzick et al., 1992). Currently, arsenic trioxide is used to treat patients with acute promyelocytic leukemia (Soignet et al., 2001; Tallman et al., 2002). Other exposed populations include workers who inhale airborne dusts during mining and processing of arseniccontaining ores (primarily lead, copper, and tin) or who manufacture and apply arsenical-based pesticides (Lubin et al., 2000). Food may account for ingestion of 8–10 mg/day (Tao and Bolger, 1998; Dabeka et al., 2004). Cigarette smokers incur an additional burden of about

382

1.5 mg arsenic per cigarette (Agency for Toxic Substances and Disease Registry, 2000). Tens of millions of people worldwide are exposed to arsenic in drinking water at levels that have been associated with increased risk of internal cancers. The most extensive exposure to levels above 50 mg/L occurs in Bangladesh (>25 million persons exposed) and West Bengal (>6 million persons exposed) (Rahman et al., 2001). This exposure is recent, the result primarily of widespread introduction of tube wells in the 1980s to reduce exposure to microbial agents in surface water sources. By 1993, arsenic contamination of tube wells was confirmed, leading to enormous public health concerns (Smith et al., 2000). High levels have been found in many other regions worldwide, including areas of Argentina (Hopenhayn-Rich et al., 1996a), northern Chile (Smith et al., 1998), China (areas of Inner Mongolia and Xinjiang Autonomous Regions) (Wang et al., 1997; Tian et al., 2001; Yang et al., 2002c), Finland (Kurttio et al., 1999), northern Mexico (Cebrian et al., 1983; Rosas et al., 1999) and the United States (California, Nevada, Alaska, Michigan, New England, New Mexico, and Utah; Focazio et al., 1999). Elevated levels in drinking water were also identified in Nepal (Shrestha et al., 2003), Thailand (Pavittranon et al., 2003), Vietnam (Berg et al., 2001), Hungary (Borzsonyi et al., 1992), Ghana (Boadu et al., 2001), and elsewhere (Smedley and Kinniburgh, 2002). Although arsenic contamination of water supplies arises mainly from natural sources, in some regions human activities such as mining and processing of sulfide-related ores directly contaminates water or modifies the chemical characteristics of water that results in the mobilization of soil-bound arsenic. In the United States, the new maximum contaminant level (MCL) for arsenic in drinking water in public water systems is 10 mg/L (National Research Council: Subcommittee on Arsenic in Drinking Water, 1999; Environmental Protection Agency, 2001; National Research Council: Subcommittee to Update the 1999 Arsenic in Drinking Water Report, 2001). The reduction from the former MCL of 50 mg/L affects U.S. water systems serving approximately 13 million people (Environmental Protection Agency, 2001). The World Health Organization similarly recommends 10 mg/L (Focazio et al., 1999).

Non-neoplastic Outcomes Perhaps the earliest recognized consequences of arsenic ingestion are cutaneous effects, such as hyperpigmentation, hypopigmentation, palmar-plantar hyperkeratoses, and leukomelanosis (National Research Council: Subcommittee on Arsenic in Drinking Water, 1999). The link between arsenic ingestion and cutaneous effects has been confirmed by numerous ecologic and, now, analytic epidemiologic studies (Allen-Price and Birm, 1960; Cebrian et al., 1983; Mazumder et al., 1998; Rahman et al., 1999; Haque et al., 2003). A disease latency for dermal effects is as brief as 2 to 10 years. Arsenic in drinking water has been linked to a wide variety of other noncancer outcomes, including gastrointestinal effects, vascular effects (peripheral vascular diseases, such as blackfoot disease [BFD], a dry gangrene of the extremities), cardiovascular disease, cerebrovascular disease, and hypertension, diabetes mellitus, and peripheral neuropathy (National Research Council: Subcommittee on Arsenic in Drinking Water, 1999). Ingested arsenic also has been associated with chronic cough, shortness of breath, and other respiratory effects (Mazumder et al., 2000; Milton et al., 2003). However, relatively few

383

Water Contaminants analytic studies of these conditions have been conducted, and further confirmation is needed, particularly for arsenic concentrations in drinking water at low to moderate levels, below 50–100 mg/L. In Argentina, arsenic levels in cord blood were nearly as high as in maternal blood, demonstrating that arsenic easily crosses the placental barrier (Concha et al., 1998b; Glickman et al., 1989). Further studies have linked ingestion of high levels of arsenic and adverse reproductive and developmental effects (Marienfeld et al., 1986; Hopenhayn-Rich et al., 2000; Ahmad et al., 2001).

Cancer Epidemiologic Studies In the absence of an acceptable animal model (Vahter, 1999; World Health Organization, 2001b; Vahter, 2002), inorganic arsenic has been designated a human carcinogen based on data from studies in human populations (International Agency for Research on Cancer, 1980). In 1980, the International Agency for Research on Cancer (IARC) found that there was sufficient evidence to establish arsenic as a carcinogen for skin and lung cancers, the latter site via inhalation of airborne arsenic (International Agency for Research on Cancer, 1980). More recent epidemiologic data for exposure to arsenic in water was reviewed by IARC in 2004 (International Agency for Research on Cancer, 2004). Adequate information had become available to classify arsenic in drinking water as a human carcinogen for skin, lung, and bladder cancers. Information for other cancer sites was deemed inadequate. Several publications provide extensive summaries of epidemiologic studies in humans and experimental animal studies (National Research Council: Subcommittee on Arsenic in Drinking Water, 1999; National Research Council: Subcommittee to Update the 1999 Arsenic in Drinking Water Report, 2001; World Health Organization, 2001b). Here we provide a brief review of these data. In 1887, Hutchinson noted an unusual number of skin cancers among patients treated with arsenicals (Hutchinson, 1887). In 1947, Neubauer cited reports of 143 skin cancer cases related to arsenic that had appeared in the literature, as well as reports of internal cancers (Neubauer, 1947). Many other case reports and preliminary reports of suspected elevated rates of skin and internal cancers related to ingestion of medicinal arsenic, drinking water, or wine contaminated with arsenical pesticides appeared before quantitative estimates were developed from epidemiologic studies (Sommers and McManus, 1953; Roth, 1957; Rosset, 1958; Robson and Jelliffe, 1963; Bergoglio, 1964; Jackson and Grainge, 1975; Popper et al., 1978; Prystowsky et al., 1978; Reymann et al., 1978; Nagy et al., 1980; Falk et al., 1981; Roat et al., 1982; Robertson and Low-Beer, 1983). The earliest quantitative epidemiologic studies of cancer were ecological in design and carried out primarily in areas of chronic arsenicism and highly contaminated water supplies, including Taiwan, Argentina, and Chile. Most studies relied on mortality statistics that permitted the assessment of arsenic effects for a variety of outcomes. These ecologic studies had no information on individual exposures to arsenic and little information on potential confounding factors. More recent efforts have included cohort and case-control studies with individual estimates of exposure and outcome and, in many instances, information on multiple potential confounding factors. In southwest coastal Taiwan, villages switched from surface to ground water (artesian wells) for drinking in the 1920s to improve the bacteriologic quality of drinking water. Unexpectedly, the newly tapped aquifers were contaminated with arsenic. The first quantitative studies to assess excess cancer and other adverse effects were conducted in this region. The first was a cross-sectional study by Tseng et al. (1968) of skin lesion prevalence, including skin cancer, in 37 villages. A total of 40,421 inhabitants were examined, and increasing prevalence was found with increasing exposure to arsenic. For example, among males 60 years of age and over, the prevalence of skin cancer per 1000 persons, was 46.1, 163.4, and 255.3 in villages grouped by concentration with <0.30, 0.30–0.59, and ≥0.60 mg/L (ppm) average arsenic concentration in drinking water. Among females 60+ years of age, skin cancer prevalence was 9.1, 62.0, and 110.1 per 1000 persons. Subsequent studies in this part of Taiwan

Table 20–1. Age-Standardized Mortality Rates (per 100,000) for Selected Cancers in Endemic Areas of Taiwan Blackfoot Disease–Endemic Area Cancer Site (ICD8 Code) All sites (140–209) Liver (155) Lung (162) Skin (173) Prostate (185) Bladder (188) Kidney (189)

Sex

≥0.60†

0.30–0.59†

<0.30†

General Population in Taiwan

M F M F M F M F M

434.7 369.4 68.8 31.8 87.9 83.8 28.0 15.1 8.4

258.9 182.6 42.7 18.8 64.7 40.9 10.7 10.0 5.8

154.0 113.3 32.6 14.2 35.1 26.5 1.6 1.6 0.5

128.1 85.5 28.0 8.9 19.4 9.5 0.8 0.8 1.5

M F M F

89.1 91.5 21.6 33.3

37.8 35.1 13.1 12.5

15.7 16.7 5.4 3.6

3.1 1.4 1.1 0.9

Source: Adapted from Chen et al. (1988a). † Well-water concentration (ppm).

revealed high risk for several internal cancers (Chen et al., 1985; Chen et al., 1988; Wu et al., 1989; Tsai et al., 1999). Dose-response gradients in age-standardized sex-specific mortality were observed for cancers of the liver, lung, skin, bladder, and kidney in both sexes, and prostate among men, as shown in Table 20–1. These findings were confirmed in mortality analyses that included the entire country (Chen and Wang, 1990). Studies of cancer mortality were also conducted in Cordoba Province, Argentina (Hopenhayn-Rich et al., 1996a; Hopenhayn-Rich et al., 1998) and Region II, Chile (Smith et al., 1998) where large populations have been exposed to high levels of drinking water arsenic. Consistent with the Taiwan studies, excess mortality in these South American studies was found for skin, lung, bladder, and kidney cancers. In Region II of Chile, the population-weighted average concentration of arsenic in drinking water was 568 mg/L from 1955 to 1969 and was lowered thereafter (Smith et al., 1998). Among males in Region II, Standardized Mortality Ratios (SMRs) for the years 1989–1993 among males for skin, bladder, lung, kidney, and liver cancers were 7.7, 6.0, 3.8, 1.6, and 1.1. Among females, the respective SMRs were 3.2, 8.2, 3.1, 2.7, and 1.1. With the exception of liver cancer, the 95% confidence interval for these point estimates excluded 1.0. It is noteworthy that cancer mortality in Region II of Chile was still elevated 20 years after arsenic levels in drinking water started to decrease (Smith et al., 1998). Levels of arsenic in the drinking water of southwest Taiwan and the affected areas of Chile and Argentina were above 150–200 mg/L, levels considered to be quite high. To date, assessment of risk at lower levels has relied on extrapolation from the high risks observed in the mortality data from southwest Taiwan (Smith et al., 1992; Environmental Protection Agency, 1988; Morales et al., 2000). Based on a linear extrapolation from the study of Chen et al. (Chen et al., 1988a), Smith et al. calculated that the lifetime risk of dying from cancer of the liver, lung, kidney, or bladder from drinking 1 L/day of water with 50 mg/L arsenic could be as high as 13 per 1000 persons (Smith et al., 1992). A later evaluation by Morales et al. applied several mathematical approaches to modeling, using mortality data for bladder, lung, and liver cancer from southwest Taiwan (Morales et al., 2000). Regardless of the model, calculated risks were of a similar magnitude as those estimated by Smith et al. (1992): “the results are sobering and indicate that current standards (50 mg/L at the time) are not adequately protective against cancer” (Morales et al., 2000). Additional support for the carcinogenicity of inorganic arsenic comes from cohort and casecontrol studies, conducted in populations exposed to generally high levels of ingested arsenic from medicinal (Cuzick et al., 1992), industrial (Tsuda et al., 1995), or natural sources (Chiou et al., 1995; Chiou et al., 2001). Cuzick et al. followed a cohort of 478 patients treated

384

PART III: THE CAUSES OF CANCER

with Fowler’s solution (potassium arsenite) and observed five bladder cancer deaths, with 1.6 expected (SMR = 3.07) (Cuzick et al., 1992). The average total dose per case was 1383 mg. Chiou et al. conducted two cohort studies in Taiwan, the first in the southwest and the second in the Liang Basin, an arsenic-contaminated area in the northeast (Chiou et al., 1995; Chiou et al., 2001). Incident bladder cancer in both regions was in excess, with risk increasing with increasing levels of waterborne levels, as estimated by cumulative and average levels of exposure. Associations with lung cancer were observed in southwest Taiwan (results for lung cancer not yet reported from the northeast). Arsenic-associated lung cancers in Taiwan appear to have a higher proportion of squamous and small cell carcinomas and a lower proportion of adenocarcinomas than lung cancers in the general population (Guo et al., 2004). In northeast Taiwan, among a cohort of 8102 persons followed for an average of 4 years, relative risks for transitional cell carcinomas (N = 10) were 1.9, 8.1, and 15.1 for arsenic levels of 10.1–50.0, 50.1–100.0, and >100.0 mg/L, respectively, relative to £10.0 mg/L. In a case-control study of incident lung cancer in Regions I–III of Chile, Ferreccio et al. found monotonically increasing risks with increasing average water concentration of inorganic arsenic, reaching an odds ratios of 8.5 (95% CI = 3.6 to 20.2) among persons with average arsenic in the range 200–400 mg/L, relative to <10 mg/L (Ferreccio et al., 2000). Northern Chile is unique in that historical levels of arsenic in drinking water can be estimated with high confidence, because most people in this desert region have lived in towns or cities where water is supplied by municipal water companies, and arsenic levels have been measured for many years. Direct observation of cancer risk at lower levels of exposure has been more challenging. Given the long latencies expected for internal cancers, it is necessary to estimate levels that occurred several decades prior to diagnosis, and this has been a vexing issue in many studies. In a Utah cohort exposed to lower levels of arsenic in water, there was no increase in mortality for a priori at-risk cancer sites, although the authors reported an increasing trend for prostate cancer mortality (Lewis et al., 1999). Estimates of arsenic levels in this study were imprecise, possibly explaining the absence of definitive results (National Research Council: Subcommittee to Update the 1999 Arsenic in Drinking Water Report, 2001). Three case-control studies of incident bladder cancer (Bates et al., 1995; Steinmaus et al., 2003; Bates et al., 2004), with modest sample sizes (117, 181, and 114 cases, respectively), found arsenic-related elevated risk among smokers with estimated exposures that took place at least three decades before diagnosis. Another small study (61 cases) found unusually elevated bladder cancer risk with waterborne arsenic levels, also restricted to smokers (Kurttio et al., 1999), in this case with shorter latency. A study of basal and squamous cell carcinomas that estimated internal dose by measuring toenail arsenic found slightly elevated risk for both diagnoses among persons above the 97th percentile of the distribution of arsenic in toenails (0.345–0.81 mg/g) (Karagas et al., 2001). However, toenail arsenic concentration as an indicator of exposure (Karagas et al., 2000) reflects internal arsenic levels about 9–12 months prior to sampling. To enhance its usefulness in estimates of longer duration exposures, this metric should be combined with information of residential (water source) stability. Risk estimates from studies of bladder cancer are highly variable for exposure to levels in water below 100 mg/L, and direct measurement of the risk level at low arsenic exposures based on available epidemiologic results are problematic. More precise estimates of risk in this exposure range await the completion of larger studies that include special attention to estimating historical exposures and collection of information to permit control for other risk factors on an individual level.

Summary of Cancer Outcomes Figure 20–1 shows summary statistics from selected studies. Results show generally consistent associations between high levels of arsenic in drinking water and lung, bladder, kidney, and nonmelanoma skin cancers. The SMR data also show evidence for an association between ingested arsenic and liver cancer but are less definitive, although 10 of 11 SMRs for liver cancer exceed 1. Results for cancer of the prostate are suggestive, but the relationship was examined in only three studies

and thus data are too limited to draw conclusions. Dose-response analysis in epidemiologic studies provide little support for a threshold concentration below which no risk is incurred. With large numbers of people exposed to arsenic in drinking water in excess of the current MCL of 10 mg/L, the full scope of the public health consequences of arsenic in drinking water remains to be clarified.

Biomarkers of Exposure and Susceptibility Urinary Speciation and Metabolism. Arsenic is metabolized primarily through methylation and then excreted in the urine. As illustrated in Figure 20–2 (Chen et al., 2003b), ingested inorganic As(V) is rapidly reduced to As(III), which then accepts a methyl group from S-adenosyl methionine (SAM) to produce the methylated pentavalent species monomethylarsonic acid, MMA(V). MMA(V) is then reduced to trivalent MMA(III). The cycle is repeated to yield the pentavalent species of dimethylarsinic acid, DMA(V), which can be reduced. Methylated (V) forms are excreted faster, are less reactive with tissues, and are less cytotoxic than inorganic arsenic or trivalent forms (Buchet et al., 1981; Marafante et al., 1985; National Research Council: Subcommittee on Arsenic in Drinking Water, 1999; Vahter and Concha, 2001). Intermediate metabolites may play an important role in carcinogenesis (Aposhian et al., 2000; Le et al., 2000). Urinary excretion of arsenic metabolites differs markedly among species (Vahter, 1994; Vahter, 1999). Inorganic and methylated amounts of urinary arsenic provide a quantitative measure of recent exposure and serve as a marker of biotransformation to methylated metabolites. Most studies of arsenicexposed populations show average values of 10%–30% inorganic arsenic, 10%–20% MMA, and 60%–70% DMA. However, there is substantial interindividual variation in methylation capacity (HopenhaynRich et al., 1993; Hopenhayn-Rich et al., 1996b; Hopenhayn-Rich et al., 1996c; Chiou et al., 1997; Concha et al., 1998a; National Research Council: Subcommittee on Arsenic in Drinking Water, 1999; Vahter, 2000; National Research Council: Subcommittee to Update the 1999 Arsenic in Drinking Water Report, 2001). Methylation efficiency may directly influence cancer risks, thereby providing important information for risk assessment of potentially susceptible subpopulations. The few studies that have addressed the issue have shown a relationship between arsenic methylation capacity and skin and bladder cancer risks (Hsueh et al., 1997; Yu et al., 2000; Chen et al., 2003a; Chen et al., 2003b). The second methylation step, from monomethyl to dimethylarsinic acid, produces less toxic metabolites. However, methylation may increase rather than decrease arsenic toxicity through the production of short-lived trivalent forms of arsenic that appear to inhibit enzyme activity, increase cell toxicity, and have genotoxic properties that may influence cancer development (Kitchin, 2001).

Nutritional Susceptibility. Inadequate nutrition or low levels of micronutrients such as vitamins B2, B6, B12, and folic acid may increase individual susceptibility to arsenic-induced cancers by influencing methylation capacity (Marafante and Vahter, 1984; Buchet and Lauwerys, 1985; Marafante et al., 1985; Buchet and Lauwerys, 1987; Styblo et al., 1995; Zakharyan et al., 1995; Li et al., 1996; Zakharyan and Aposhian, 1999). Low intakes of methionine, protein, and choline impair arsenic methylation in experimental animals (Vahter and Marafante, 1987). It has been proposed that certain populations, such as the rural Taiwanese, may be more susceptible to the carcinogenic effects of arsenic due to deficiencies in dietary factors, including protein (Buchet and Lauwerys, 1987; Lin and Yang, 1988; Bates et al., 1992; Engel and Receveur, 1993; Carlson-Lynch et al., 1994; Beck et al., 1995). Low intake of vitamins and other nutrients was associated with an odds ratio of 11.5 for skin cancer in arsenic-exposed Taiwanese villagers (Hsueh et al., 1995). In a prospective cohort study in Taiwan, skin cancer cases had significantly lower serum beta-carotene levels than matched controls (Hsueh et al., 1997). In West Bengal, people below 80% of the standard body weight for their age and sex had a 1.6-fold increase in the prevalence of arsenic-associated skin keratoses (Mazumder et al., 1998). Evaluating the effects of diet on arsenic health risks remains important, as the major arsenic risk

Figure 20–1. Standardized mortality ratios (SMR) for males (M) and females (F) in representative studies of arsenic-exposed populations. For studies with a trend analysis, SMRs shown for highest exposure category. Studies identified by location and reference.

386

PART III: THE CAUSES OF CANCER

Figure 20–2. Proposed arsenic methylation pathway in the human body. (Source: Chen et al., 2003b.)

assessments to date have been based on data from Taiwanese populations where nutritional sufficiency is uncertain (Environmental Protection Agency, 1988; Chen et al., 1992; Smith et al., 1992; Engel and Receveur, 1993; Mushak and Crocetti, 1995; National Research Council: Subcommittee on Arsenic in Drinking Water, 1999; Environmental Protection Agency, 2000; Morales et al., 2000; Steinmaus et al., 2000). Selenium increases the nonenzymatic methylation of inorganic arsenic (Zakharyan and Aposhian, 1999) and has been associated with reduced arsenic toxicity and increased arsenic excretion in animal and laboratory tests (Taketani et al., 1991; Schrauzer, 1992; Berry and Galle, 1994; Glattre et al., 1995; Biswas et al., 1999; Flora et al., 1999). Both selenium and arsenic affect DNA methylation in vivo and in vitro, suggesting that cytosine DNA methyltransferase, selenium, and arsenic compete for methyl donation from SAM.

Genetic Susceptibility. Genetic susceptibility to the toxic effects of arsenic is suggested by variations in arsenic metabolism within and among populations, as described above. Our understanding of the enzymes in the metabolic pathway of arsenic methylation to MMA and DMA is growing rapidly, with concomitant knowledge about the genes that code for these enzymes. Monomethylarsonic acid (MMA-V) reductase is suggested to be in the omega class of glutathione-S-transferases, and genetic polymorphisms of this gene have been linked with variations in relative concentrations in urine of inorganic arsenic, DMA, and MMA (Zakharyan et al., 2001; Marnell et al., 2003). Experimental studies suggest a role for cysteine and glutathione in the conversion of As(V) to As(III) (Buchet and Lauwerys, 1985; Thompson, 1993; Zakharyan et al., 1995; Styblo et al., 1996), and deficiencies in glutathione have been linked to increased arsenic toxicity (Hirata et al., 1990; Huang et al., 1993; Oya-Ohta et al., 1996; Shimizu et al., 1998). Methylenetetrahydrofolate reductase (MTHFR) may be involved in arsenic detoxification. Support for this pathway comes from experimental studies and an anecdotal report of a patient, chronically exposed to arsenic, with MTHFR deficiency and hyperhomocysteinemia who manifested arsenic poisoning symptoms. Similarly exposed family members who were not MTHFR-deficient remained asymptomatic (Brouwer et al., 1992; Vahter, 2000). Cytogenetic Studies. That arsenic is genotoxic comes from several lines of evidence. Possible mechanisms of DNA damage include inhibition of DNA repair enzymes (Li and Rossman, 1989; Lee-Chen et al., 1994; Wiencke et al., 1997; Hu et al., 1998; Hamadeh et al., 2002; Andrew et al., 2003), induction of DNA hypo- and hypermethylation (Schroeder and Mass, 1997; Zhao et al., 1997; Moore et al., 2003a), increased gene amplification (Lee et al., 1988), and oxidative damage (Kitchin and Ahmad, 2003). Chromosomal damage is evidenced by induction of micronuclei in human lymphocytes and in exfoliated bladder cells (Jha et al., 1992; Warner et al., 1994; Vega et al., 1995; Dulout et al., 1996; Moore et al., 1996; Gonsebatt et al., 1997; Moore et al., 1997a; Moore et al., 1997b). The molecular mechanisms by which arsenic promotes cancer also may be studied by comparing tumor DNA from arsenic-exposed and unexposed individuals. Such studies of skin tumors suggest the pos-

sibility of arsenic-specific mutational patterns in the p53 gene (Hsieh et al., 1994; Szekeres and De Giacomoni, 1994; Boonchai et al., 2000). However, p53 protein expression and mutation prevalence did not differ in bladder tumors from arsenic exposed and unexposed patients (Moore et al., 2003b). When the same tumors were screened for genetic aberrations using comparative genomic hybridization, bladder tumors associated with higher levels of arsenic in drinking water showed an increased number of chromosomal gains and losses. Most arsenic-related chromosome changes were also associated with tumor stage and grade, suggesting that arsenic-induced tumors may be more aggressive and result in increased mortality (Moore et al., 2002).

DISINFECTION BY-PRODUCTS Exposure Disinfection of drinking water with chlorine has brought large-scale health benefits to humankind. However, the well-documented reduction of morbidity and mortality due to control of waterborne microbes may be partially offset by small increases in cancer risk. Disinfection by-products (DBP), a complex mixture of chlorinated and brominated compounds, result from the unintentional halogenation of organic material common in untreated waters, mostly naturally occurring humic and fulvic acids. Hundreds of compounds have been identified, with chloroform, other trihalomethanes (THM), and haloacetic acids (HAA) the most commonly found, jointly accounting for 30% to 70% of covalently bound halogen (mostly chlorine and bromine) (Krasner et al., 1989; Stevens et al., 1990). Organically bound bromine usually originates as bromide in source waters, oxidized by chlorine to highly reactive Br0. A large number of higher molecular weight, nonvolatile compounds, such as halogenated mono- and dicarboxylic acids, acetonitriles, aldehydes, ketones, ethers, and others, have been identified (Singer and Chang, 1989; Stevens et al., 1989). Identification of additional compounds continues, especially those related to alternative disinfection practices (Richardson et al., 2002; Weinberg et al., 2002; Plewa et al., 2004). It is uncertain which compounds or compound families pose a risk of cancer. Among the relatively few compounds tested, several are carcinogenic in rodent bioassays (Dunnick and Melnick, 1993; Boorman, 1999; International Agency for Research on Cancer, 2004). Concentrates of the overall mixture and some mixture components are mutagenic in Ames Salmonella and other in vitro testing systems (Loper et al., 1978). A highly mutagenic chlorinated hydroxyfuranone (“MX”), also a rodent carcinogen, has been identified and is widely found in treated waters (Holmbom et al., 1981; Holmbom et al., 1984; Kronberg and Vartiainen, 1988; Kronberg and Christman, 1989; Daniel et al., 1993; Komulainen et al., 1997; Wright et al., 2002). The primary route of exposure for the low-molecular-weight, volatile THM is inhalation and dermal absorption during bathing and showering, as demonstrated by measurements in whole blood (Jo et al., 1990; Backer et al., 2000; Nieuwenhuijsen et al., 2000; Arbuckle et al., 2002; Miles et al., 2002; Whitaker et al., 2003). Elevated THM levels are found in the blood and exhaled breath of swimmers in chlo-

Water Contaminants rinated swimming pools, especially indoor pools (Aggazzotti et al., 1990; Aggazzotti et al., 1993; Lindstrom et al., 1997; Kim et al., 2002). Exposure to the higher molecular weight, usually more polar compounds, is primarily through consumption. Di- and trichloroacetic acid in urine have been measured after oral exposure (Froese et al., 2002; Calafat et al., 2003), with trichloroacetic acid more closely related to levels in consumed drinking water (Kim et al., 1999). Transdermal absorption of these polar compounds is negligible (Xu et al., 2002). Levels of chlorination by-products are generally higher in treated surface water than well water. THM concentrations in treated surface water occasionally range up to several hundred parts per billion (ppb), whereas concentrations in treated groundwater sources are typically below 15 or 20 ppb. Many epidemiologic studies have taken advantage of this characteristic, using “water source” (surface or ground) as a surrogate for DBP exposure. In other studies, measured or estimated THM levels served as indicator compounds for the full mixture. The current maximum contaminant level (MCL) permitted in the United States is 80 ppb total THM, taken as an annual average, with welldefined requirements for the timing of measurements and location of sampling within water distribution systems (Environmental Protection Agency, 1998). Epidemiologic data, while not used directly for quantitative risk assessment, have been important in motivating and informing these regulations.

Epidemiologic Studies Epidemiologic evaluation of cancer risk and chlorination byproducts started with ecologic studies shortly after THM were first detected in drinking water in 1974 (Bellar and Lichtenberg, 1974; Rook, 1974). Age-adjusted, site-, sex-, and race-specific incidence (Carlo and Mettlin, 1980; Bean et al., 1982b; Isacson et al., 1983; Flaten, 1992; Koivusalo et al., 1994; Koivusalo et al., 1995) or mortality (Page et al., 1976; Kuzma et al., 1977; Salg, 1977; Cantor et al., 1978; Hogan et al., 1979) rates for regions such as U.S. counties were used as outcomes and characteristics (chlorinated vs nonchlorinated, surface vs ground, THM level, or estimated mutagenicity) of the predominant regional drinking water source as the exposure variable. These ecologic studies, conducted in the United States, Norway, Finland, and elsewhere, provided the basis for further work by identifying the cancer sites and defining exposures of concern. They have received considerable review (International Agency for Research on Cancer, 1991; Cantor, 1997). Cancers of the bladder, colon, and rectum were the most commonly observed cancer sites linked with DBP. The next group of studies were of case-control design, using mortality records as the source of study subjects (Alavanja et al., 1977; Brenniman et al., 1980; Young et al., 1981; Gottlieb and Carr, 1982; Crump and Guess, 1982; Lawrence et al., 1984). Most focused on cancers of the bladder, colon, or rectum and linked the death certificate address with water supply information to estimate exposure. These studies typically had limited information on other risk factors and lacked detail on historical levels of DBP. Their results generally supported the findings of ecologic studies in showing excess risk for bladder, colon, and rectal cancers among highly exposed subjects. Cohort and case-control studies, using individual-level measures of effect and exposure, as well as control for other risk factors, have also been conducted. These studies vary in the number of years that water levels were estimated prior to diagnosis and the way in which DBP exposures were estimated. Here, we review these studies by cancer site. Bladder cancer has received the greatest attention in analytic studies, with data from three cohorts (Wilkins and Comstock, 1981; Doyle et al., 1997; Koivusalo et al., 1997), six case-control studies (McGeehin et al., 1993; Cantor et al., 1987; King and Marrett, 1996; Freedman et al., 1997; Cantor et al., 1998; Koivusalo et al., 1998), and a pooled analysis (Villanueva et al., 2004). Elevated risk was found with long-term exposure to DBP in all but one study. Among cohort studies, in a population of nearly 31,000 persons in Washington County, Maryland, Wilkins and Comstock (1981) found elevated (but not statistically significant) bladder cancer incidence among men

387

(RR = 1.80), with risk increasing with duration of residence of at least 12 years with a chlorinated surface water source, as compared with persons who used unchlorinated ground water (Wilkins and Comstock, 1981). In Iowa, among 28,237 women age 55–69 at enrollment, there was no association of bladder cancer with consumption of chlorinated surface water for at least 10 years (Doyle et al., 1997). Estimates of DBP levels in these two cohorts were limited to information pertaining to the residence at the time of cohort enrollment. In a historical cohort study from Finland, Koivusalo found elevated risk for bladder cancer associated with model-based estimates of historical water mutagenicity (Koivusalo et al., 1997). Case-control studies of bladder cancer were conducted in Maryland, Iowa, Colorado, Ontario (Canada), Finland, and independently, in 10 areas of the United States. In each instance, a long-term profile (from 35 years to lifetime) of water type and/or estimates of THM concentration or mutagenicity served as the exposure index, and risk estimates were adjusted for age, sex (if applicable), smoking, and other potential confounders. In 10 areas of the United States, bladder cancer risk increased with the amount of tap water consumed, with the increase related to the duration at residences served by chlorinated surface water, measured over the lifetime of respondents (Cantor et al., 1987). Among respondents with 60+ years of exposure at places served by surface water, a risk of 2.0 was found for persons in the highest tap water consumption quintile relative to those in the lowest. In Colorado, increasing risk with duration of chlorinated surface water use was found in a smaller case-control study of bladder cancer and disinfection methods (327 cases, 261 controls) (McGeehin et al., 1993). A nested case-control study from Washington County, Maryland, found increasing risk with duration at enrollment residences served by chlorinated surface water (OR = 1.4 for the highest duration category, 40+ years) (Freedman et al., 1997). In Finland, an OR of 1.22 (0.92–1.62) was observed per increment of estimated water mutagenicity (3000 net revertants/liter) among respondents with 30+ years of exposure information (Koivusalo et al., 1998). In Iowa, the OR for long-term use of chlorinated surface water (60+ years) was 1.5 (CI = 0.9 to 2.6), as contrasted with long-term users of nonchlorinated groundwater, with excess risk confined to male subjects (OR = 1.9 [1.1–3.6]); among women, the OR was 0.7 (0.2–2.4) (Cantor et al., 1998). A study from Ontario, Canada, found an OR of 1.63 (1.08–2.46) among respondents with 35+ years exposure to a THM level of ≥50 mg/L as compared with those exposed for less than 10 years (King and Marrett, 1996). Overall findings were positive from each casecontrol study, with some inconsistencies regarding sex-specific risks and whether risks occurred primarily among ever- or never-smokers. A pooled analysis of data from six published and unpublished casecontrol studies showed an association with long term (a 40-year window) THM exposure among men but not women (Villanueva et al., 2004). Among men, the OR increased to 1.43 (1.2–1.7) for a longterm average level of >50 mg/L (p [trend] <0.001) and an OR of 1.51 (1.2–1.8) for total ingestion of >1000 mg THM (p [trend] <0.001) (Figure 20–3). Incident colon cancer was evaluated in case-control studies in North Carolina (Cragle et al., 1985), Wisconsin (Young et al., 1987), Iowa (Hildesheim et al., 1998), and Ontario (King et al., 2000), and in a cohort study in Iowa (Doyle et al., 1997). Rectal cancer was also evaluated in the Iowa and Ontario case-control studies and in the Iowa cohort study. In the Iowa cohort, a dose-response relation with increasing levels of chloroform was observed for colon but not rectal cancer incidence (Doyle et al., 1997). In North Carolina, excess colon cancer was found among those more than 60 years of age but not among younger subjects (Cragle et al., 1985). Few study details are available. In Wisconsin, no association of colon cancer with estimates of past THM was noted (Young et al., 1987). However, response rates were below 50% and thus limit interpretation. In Ontario, colon cancer was associated with long-term (40 year) estimates of THM levels, primarily among men, with odds ratios increasing to 2.10 (1.2–2.8) among men with ≥35 years at a water source with ≥75 mg/L (King et al., 2000). There was no association with rectal cancer risk. Colon cancer risk in Iowa was not associated with lifetime estimates of exposure to chlorinated surface water or to average THM level (Hildesheim et al.,

388

PART III: THE CAUSES OF CANCER assay as a biomarker of genotoxicity in exfoliated bladder cells, a cross-sectional study of 3 Australian communities (228 participants) found no difference in the prevalence of micronuclei with varying levels of DBP in drinking water (38 to 157 mg/L), intake dose, or tobacco use (Ranmuthugala et al., 2003).

Summary

Figure 20–3. Exposure-response plot using natural splines for bladder cancer and average THM level over 40 years in a pooled analysis of six case-control studies. (Source: Villaneuva et al., 2004.)

1998). However, in Iowa, an association with rectal cancer was found for both men and women, with OR increasing monotonically to 2.61 (1.4–5.0) for subjects with at least 60 years chlorinated surface water exposure (Hildesheim et al., 1998). Findings for colon and rectal cancer from case-control studies differ, possibly due to differences in methodology or regional water quality. Other cancer sites with long-term exposure data evaluated in analytic studies include kidney, brain, pancreas, and esophagus as well as childhood leukemia. In Finland, a study of 703 incident kidney cancer cases and 914 controls found an OR of 1.31 (CI = 1.00 to 1.74) for an increase of 3000 net revertants/liter, an estimate of drinking water mutagenicity among respondents with ≥30 years of estimable exposure (Koivusalo et al., 1998). Evidence for an association was weaker in an Iowa case-control study with lifetime estimates of water source and long-term THM estimates (Cantor et al., 2005a). A companion study in Iowa that evaluated brain cancer (glioma) found elevated risk among men, but not women, with the OR among men increasing to 2.5 (1.2–5.0) for those exposed 40+ years to chlorinated surface water (Cantor et al., 1999). Two case-control studies of pancreas cancer that evaluated exposure to DBP level at the time of diagnosis were inconsistent (IJsselmuiden et al., 1992; Kukkula and Lofroth, 1997). In a third case-control study, pancreas cancer was not associated with lifetime estimates of duration of exposure to chlorinated surface water or to average THM (Cantor et al., 2005b). In a study of childhood acute lymphocytic leukemia (ALL) from Quebec Province, no association was found with indicators of DBP, both during pregnancy and postnatally (Infante-Rivard et al., 2001). In a nested case-control study in Shanghai, China, esophageal cancer deaths in men were linked to drinking-water mutagenicity, an indicator of DBP and/or other pollutants (Tao et al., 1999). The overall OR was 1.80 (1.12–2.88) for longterm residence in places with elevated mutagenicity in Ames testing systems.

Biomarkers of Susceptibility There is a large and growing body of in vitro and in vivo animal data on the metabolism and mechanism of action of various DBP (Daniel et al., 1993; Boorman, 1999; International Agency for Research on Cancer, 2004). However, relatively few human studies of metabolism or susceptibility have been conducted. A case-case analysis in a subset of subjects from the study of childhood ALL in Quebec cited above analyzed genetic polymorphisms in the THM metabolizing enzymes CYP2E1 and GSTT1 (Infante-Rivard et al., 2002). Elevated but not statistically significant odds ratios were observed in cases carrying the GSTT1 null and CYP2E1 (variant G-1259C, allele CYP2E1*5) compared to cases without the variant genotype. Using the micronucleus

Mounting evidence suggests an association of bladder cancer with long-term exposure to DBP in drinking water, a complex and variable mixture that contains many compounds that are carcinogenic and mutagenic in laboratory tests. Data for colon and rectal cancer from two well-conducted case-control studies with long-term estimates of exposure are conflicting. Information from studies of other cancers, in which exposures were estimated for at least four decades before diagnosis, are more limited. Findings from at least one such study of renal cell carcinoma and another of glioma provide a limited suggestion of elevated risk. Suggestive evidence also exists for esophageal cancer, which has been less studied. For all cancer sites other than perhaps bladder cancer, there is need for additional work. The growing database indicates that where elevated risk exists, it is likely related to long-term exposures, requiring that exposure assessment accounts for at least 35 years of experience prior to cancer diagnosis. Where associations have been observed, relative risks are modest (in the range 1.3–2.6). However, substantial numbers of people are exposed, and modest relative risks can translate to large attributable risks, implying an important public health impact.

NITRATE Exposure Over the past 50 years, humans have doubled the natural rate of nitrogen deposition onto the land, largely through application of manufactured nitrogen fertilizers but also through combustion of fossil fuels and replacement of natural vegetation with nitrogen-fixing crops such as soybeans (Vitousek et al., 1997). Nitrate levels in groundwater under agricultural land can be several- to a hundred-fold higher than levels under natural vegetation (Nolan and Stoner, 2000). Groundwater is used for drinking water by 90% of the rural population in the United States (Solley et al., 1993). Many rural residents have private wells, which are not regulated by the U.S. Environmental Protection Agency (EPA) and for which there are few comprehensive assessments of water quality. The U.S. Geological Survey assessed available private well measurements in U.S. aquifers sampled in 1970–1992 and found that 9% of samples exceeded the Maximum Contaminant Level (MCL) of 10 mg/L as nitrate-nitrogen (nitrate-N) (Mueller et al., 1995). Other private well surveys show that a significant proportion of wells, particularly those less than 30 m in depth in agricultural areas, exceed the MCL for nitrate (Johnson and Kross, 1990; National Center for Environmental Health, 1998). Almost all public water supplies have nitrate levels below the MCL; however, in the past few decades, nitrate levels in some public supplies in agricultural areas have risen to levels approaching the MCL (Ward et al., 2003). The EPA established an MCL for nitrate in drinking water at 10 mg/L nitrate-N to protect against methemoglobinemia, to which infants are especially susceptible. The MCL was not based on risk of cancer or other health outcomes. Nitrate is a precursor compound in the formation of N-nitroso compounds (NOC), which are among the most potent known animal carcinogens. NOC cause cancer in every animal species tested and at multiple organ sites (Lijinsky, 1986). Approximately 5% of nitrate ingested in drinking water or in the diet is reduced to nitrite by bacteria. Nitrate reduction occurs primarily in the mouth but can also occur in the stomach, small intestine, colon, and bladder when nitrate-reducing bacteria are present (National Research Council, 1995). Primarily in the stomach, nitrite reacts with amines and amides to form nitrosamines and nitrosamides (a reaction called nitrosation), the two main types of NOC. Nitrate is found in many foods, with the highest levels occurring in green leafy and root vegetables (National Academy of Sciences,

Water Contaminants 1981), and average daily intakes have been estimated to be in the range 30–130 mg/day (National Research Council, 1995). NOC formation is blocked by ascorbic acid, alpha-tocopherol, and other polyphenols. Dietary nitrate intake thus may not result in substantial NOC formation due to the presence of these nitrosation inhibitors in most vegetables (Mirvish, 1983). Drinking water can contribute the majority of nitrate intake when levels are near the MCL (Chilvers et al., 1984). Sources of exposure to preformed NOC include preserved meats and fish, beer, certain occupations, cosmetics, and some drugs; however, it is estimated that 45%–75% of human exposure to NOC comes from in vivo formation (Tricker, 1997). Endogenous NOC formation after ingesting water with elevated nitrate has been demonstrated in human volunteers (Moller et al., 1989; Mirvish et al., 1992). NOC have been also measured in human feces after dosing with nitrate in drinking water, indicating that nitrosation can occur in the intestines, most likely due to the action of colonic bacteria (Rowland et al., 1991). In each of these three studies, increased nitrosation was observed at nitrate levels exceeding the MCL. Animal studies indicate that long-term exposure to lower NOC concentrations has a stronger carcinogenic effect compared with short-term higher exposures (Lijinsky, 1986). Epidemiologic evidence for other water contaminants also indicates that exposure over many decades may be more important than high episodic exposures (King and Marrett, 1996; Cantor et al., 1998). However, few epidemiologic studies of drinking water nitrate and cancer have assessed long-term historical exposures.

Epidemiologic Studies Most early epidemiologic studies were ecologic mortality studies of stomach cancer that used exposure estimates based on drinking water nitrate measurements concurrent with the time period of cancer mortality. Results were mixed, with some studies showing positive associations, many showing no association, and a few showing inverse associations (studies through 1995 reviewed in Cantor [1997]). Most ecologic studies of stomach cancer incidence also relied on recent exposure data for public water supplies (Gilli et al., 1984; Barrett et al., 1998; Van Leeuwen et al., 1999), with the exception of a study in Slovakia (Gulis et al., 2002). In Slovakia, the average nitrate levels in public water supplies over a 20-year period (highest quartile: 6– 10 mg/L nitrate-N) showed a positive correlation with stomach cancer incidence among women but not men (Gulis et al., 2002). A study in Italy (Gilli et al., 1984) also found elevated stomach cancer incidence in communities with drinking water nitrate levels above 4.5 mg/L nitrate-N; however, a study in England (Barrett et al., 1998) did not. A Canadian study reporting low levels (75th percentile ≥2.63 mg/L nitrate-N) (Van Leeuwen et al., 1999) found a negative correlation. Two recent ecologic studies of stomach cancer mortality with levels above the MCL found positive correlations. A study in Hungary (Sandor et al., 2001) with approximately 10 years of historic measurements found elevated SMRs at nitrate levels above 18 mg/L nitrateN. A study in 258 municipalities in Valencia, Spain found elevated stomach cancer mortality at nitrate levels over about 11 mg/L nitrateN (Morales-Suarez-Varela et al., 1995). Many of the recent ecologic studies of cancer incidence and publicsupply drinking water nitrate levels focused on other cancer sites including the brain (Barrett et al., 1998; Van Leeuwen et al., 1999), bladder (Van Leeuwen et al., 1999; Gulis et al., 2002), colon (Van Leeuwen et al., 1999; Gulis et al., 2002), esophagus (Barrett et al., 1998), kidney (Gulis et al., 2002), and non-Hodgkin lymphoma (NHL) (Law et al., 1999; Van Leeuwen et al., 1999; Cocco et al., 2003). With the exception of the study in Slovakia (Gulis et al., 2002), these studies also relied on nitrate measurement data close in time to the period of cancer incidence. Incidence of NHL was elevated among men and women exposed to public supply nitrate levels of 6–10 mg/L nitrateN in Slovakia (Gulis et al., 2002). Three other ecologic studies with somewhat lower distributions of nitrate found no association with NHL (Law et al., 1999; Van Leeuwen et al., 1999; Cocco et al., 2003). In Nebraska, NHL incidence was elevated in counties where 20% or

389

more of private wells were above the MCL compared to counties with <10% of wells above that level (Weisenburger, 1991). Two studies, from Yorkshire, England, and Ontario, Canada, found no association with public-supply nitrate levels and brain (Barrett et al., 1998) or central nervous system (Van Leeuwen et al., 1999) cancers. Esophageal cancer incidence was also not associated with nitrate levels in the English study (Barrett et al., 1998). Elevated nitrate levels were associated with an increased incidence of colon but not rectal cancer in Slovakia (Gulis et al., 2002) and were positively correlated with colon cancer in a Canadian study (Van Leeuwen et al., 1999). There was no association between nitrate levels and bladder and kidney cancer incidence in Slovakia (Gulis et al., 2002) and with bladder cancer in Canada (Van Leeuwen et al., 1999). The Canadian study (Van Leeuwen et al., 1999) also found no correlation with ovarian cancer incidence. Prostate cancer mortality in Spain (MoralesSuarez-Varela et al., 1995) was positively correlated with nitrate levels. No correlation was observed for bladder or colon cancer mortality in the Spanish study. In the past decade, several case control and cohort studies have evaluated historical nitrate levels in public water supplies (largely below 10 mg/L nitrate-N) and risk of several cancers. Most were able to evaluate potential confounders and factors affecting nitrosation. Characteristics of these analytic studies and summary findings are summarized in Table 20–2. A cohort study of older women in Iowa (Weyer et al., 2001) found a positive association with the longterm average nitrate levels at the current residence and bladder and ovarian cancers. They observed significant inverse associations for uterine and rectal cancer and no association for NHL, leukemia, rectum, pancreas, kidney, lung, and melanoma. Relative risks for colon cancer were elevated in the second and third quartiles of water nitrate level. Some analyses were limited by small numbers. Casecontrol studies of bladder (Ward et al., 2003), colon and rectum (DeRoos et al., 2003), and pancreas (Coss et al., 2004) cancers in Iowa found no association with risk and average nitrate levels over an almost 30-year period or with other exposure metrics. Each of the studies evaluated the interaction between nitrosation inhibitors or NOC precursors and nitrate intake from drinking water. For colon cancer, there was a significant positive interaction between exposures of 10 or more years above 5 mg/L nitrate-N and low vitamin C and high meat intake. A case-control study of NHL in Nebraska (Ward et al., 1996) found a significant positive association between the average nitrate level in public water supplies over 40 years and risk among men and women. In the highest quartile of nitrate level (≥4.0 mg/L nitrateN), risk was elevated twofold. A case-control study of NHL in Minnesota (Freedman et al., 2000) with lower levels of nitrate found an inverse association among those with the highest level (>1.5 mg/L nitrate-N). Case-control studies in Nebraska (Ward et al., 2005) and Germany (Steindorf et al., 1994) found no association with long-term average nitrate levels in public water supplies and adult brain cancer. A cohort study in the Netherlands (van Loon et al., 1998) found no association between stomach cancer risk and quintiles of water nitrate intake determined from public-supply levels and tap water intake. Public-supply nitrate levels were not associated with stomach cancer in a death certificate–based case-control study in Wisconsin (Rademacher et al., 1992). Lower nitrate levels (highest tertile >0.45 mg/L nitrate-N) showed a weak but significant positive association in a death certificate–based case-control study in Taiwan (Yang et al., 1998). A few case-control studies evaluated well water use and cancer risk. In Germany (Boeing et al., 1991), stomach cancer risk was positively associated with private well water use compared with public supply use. In Colombia, well water use in the first 10 years of life was associated with an increased risk of stomach cancer precursor lesions (Cuello et al., 1976). In the states of California and Washington (Mueller et al., 2001), there was no overall association with well water use during pregnancy and subsequent risk of brain cancer in the offspring. Dipstick measurements of the nitrate levels in water supplies, which often occurred many years after the pregnancy, were not associated with risk.

Table 20–2. Analytic Epidemiologic Studies of Cancer and Exposure to Nitrate in Drinking Water First Author (Year) Country

Study Design, Years Regional Description

Boeing (1991) Germany

Hospital-based case-control Incidence, 1985–1988 Hospitals from two regions Population based case-control Incidence, 1987–1988 Rhein–Neckar–Odenwald area Population-based case-control Incidence, 1983–1986 66 counties in Eastern Nebraska

Steindorf (1994) Germany Ward (1996) U.S.A.

Exposure Description* Public water supply or other (mostly private well)

Stomach

Levels in municipal supplies after 1970 (highest quartile: >5.7 mg/L) Average level in public water supplies 1947 to early 1980s grouped into quartiles (lowest: <1.6; highest: ≥4.0 mg/L); Ever exposure ≥10 mg/L Intake from water determined from public supply levels (1986) and volume of tap water and beverages made with tap water (quintiles; mean level in highest quintile 3.7 mg/day)

Brain

Van Loon (1998) The Netherlands

Prospective cohort Incidence, 1986–1992

Freedman (2000) U.S.A.

Population-based case-control Incidence, 1980–1982 Minnesota excluding four largest cities

Mueller (2001) U.S.A

Population-based case-control 19 counties in San Francisco, California area, and western Washington State, 1984–1990

Weyer (2001) U.S.A.

Cohort of women ages 55–69 Incidence, 1986–1998 State of Iowa

Avg level (1955–1988) in public water supplies for residence at enrollment (highest quartile: >2.46 mg/L)

Ward (2003) U.S.A.

Population-based case-control Incidence, 1986–1989 Iowa

Average level in public water supplies 1960–1987 (highest quartile men: 3.1 mg/L; women: 2.4 mg/L); years of exposure ≥10 mg/L Average level in public supplies 1960–1987 grouped in 4 levels (lowest: £1.0; highest: >5 mg/L); years of exposure >5 and >10 mg/L

DeRoos (2003) Population-based case-control U.S.A. Incidence, 1986–1989 Iowa

Coss (2004) U.S.A.

Population-based case control Incidence, 1985–1987 Iowa

Ward (2005) U.S.A.

Population-based case-control Incidence, 1989–1993 66 counties in Eastern Nebraska

Cancer Sites Included

Avg level in public supplies 1947–1980 (157 towns) categorized into 3 levels: £0.5, >0.5 to £1.5, >1.5 mg/L) Water source (private well, public supply) during pregnancy; nitrate and nitrite measured for those still at residence during pregnancy

Average level in public supplies 1960–1987 (highest quartile: >2.8 mg/L); years of exposure ≥7.5,10 mg/L Average level in public water supplies 1960–1986

Comments

Positive association with well water use, nitrate levels not measured No association with average nitrate level

Non-Hodgkin lymphoma

Significant positive trend with quartile level among men and women: OR highest vs. lowest quartile, OR = 2.0 (95% CI 1.1–3.6)

Excluded those with >10% of years after 1944 with unknown nitrate level

Stomach

No association with quintiles of water nitrate intake (highest quintile: RR = 0.88)

RR adjusted for other stomach cancer risk factors, (e.g., education, vitamin C, beta-carotene, family history of stomach cancer, stomach disorders; freezer/ refr. use) Excluded those with >10% of years after 1946 with unknown nitrate level

Non-Hodgkin lymphoma

No increased risk with increasing exposure level. OR for >1.5 mg/L (3 cases, 4 controls) was 0.3 (95% CI 0.1–0.9). Childhood brain No overall association with water source. Well use in western Washington State increased risk (OR = 2.6, 95% CI 1.3–5.2); well use in Los Angeles inversely associated with risk (OR = 0.2, 0.1–0.8). Non-Hodgkin Positive associations with lymphoma, average nitrate level for leukemia, colon, bladder (highest quartile rectum, OR = 2.83) and ovary pancreas, (OR = 1.84) and inverse kidney, bladder, associations for uterus breast, ovary, (highest quartile OR = uterine corpus, 0.55) and rectal cancer lung/bronchus, (OR = 0.47) melanoma Bladder Inverse association with quartiles of average level among men; no association among women. Similar results for years ≥10 mg/L. Colon No association with average Rectum level or number of years >5 and 10 mg/L; significantly elevated risk among subgroups < median vitamin C intake or > median meat intake and 10 or more years >5 mg/L Pancreas No significant association with average nitrate or number of years ≥7.5 or 10 mg/L Brain (gliomas)

*Nitrate levels presented in the publications as mg/L nitrate were converted to mg/L as nitrate-nitrogen.

390

Summary of Findings

No association with quartiles of average nitrate level

Excluded those who used private well or bottled water and who resided <10 years at enrollment residence. Adjusted for smoking, vitamins C and E, dietary nitrate, water source. Excluded those with >30% years after 1959 with unknown nitrate level Excluded those with >30% years after 1959 with unknown nitrate level

Excluded those with >30% years after 1959 with unknown nitrate level Excluded those with >30% years after 1959 with unknown nitrate level

Water Contaminants

391

Biomarkers of Genetic Damage

Epidemiologic Studies

In response to the reported increase in internal cancers due to consumption of well water with high levels of nitrate, two cross-sectional studies of the genotoxic effects of nitrate in drinking water were conducted. Both included individuals drinking well water with nitrate concentrations ranging from 11 to 65 mg/L as nitrate-N. The first study found no increase in the frequency of peripheral lymphocyte sister chromatid exchanges with water nitrate level (Kleinjans et al., 1991). A subsequent study employed the hypoxanthine-guanine phosphoribosyltransferase (HPRT) variant frequency test in peripheral lymphocytes (van Maanen et al., 1996). An increased prevalence of HPRT was observed among subjects who drank water with medium and high levels of nitrate. An inverse correlation between the labeling index in lymphocytes and nitrate exposure was suggestive of an exposurerelated immunosuppressive effect. Such studies examining the genotoxic potential of nitrate in drinking water can make valuable contributions to our understanding of the adverse effects of nitrate exposure and warrant further exploration.

County and local cancer mortality and incidence rates, calculated from routinely collected information, have been examined in places with hazardous waste sites or with municipal water supplies with documented contamination. An early study from Iowa described elevated county lymphoma rates in places served by rivers with measurable levels of dieldrin (DeKraay, 1978). A bladder cancer cluster was noted in a Northwestern Illinois community with exposure to several contaminants in the community drinking water supply, including trichloroethylene, tetrachlorethylene, and 1,1-dichloroethane, that leached from a regional landfill (Mallin, 1990). In a nationwide study, age-adjusted cancer mortality rates from 339 U.S. counties with 593 hazardous waste sites listed by the EPA were compared with rates from 2726 other counties (Griffith et al., 1989). Significant associations were found for lung, bladder, stomach, colon, and rectal cancers in white males and females, esophageal cancer in white males, and breast cancer in white females. In New Jersey, significant positive associations between chemical toxic waste disposal sites and eight cancer sites, especially stomach and lung, were found in one or more subpopulations living in 194 municipalities with more than 10,000 population, after adjustment for sociodemographic characteristics (Najem et al., 1985). In Great Britain, there was no excess risk of incident bladder cancer, brain cancer, hepatobiliary cancer, or adult or childhood leukemias among populations living within 2 km of 9565 landfill sites that operated at some time in the period 1982–1997 (Jarup et al., 2002). A household survey and review of medical and vital statistics records in communities near the Stringfellow Hazardous Waste Site in Riverside County, California (Baker et al., 1988), did not reveal excesses of cancer or adverse birth outcomes, but increases in other conditions were reported. The authors noted that these findings may be explained by differential reporting from neighboring residents. A Finnish community with drinking water contaminated with chlorophenols, probably from sawmills, experienced an elevated incidence of soft-tissue sarcoma and non-Hodgkin lymphoma (Lampi et al., 1992). These findings are of concern because these tumors have been linked with exposure to the closely related chlorinated phenoxy-acetic acids and/or their dioxin contaminants (Lilienfeld and Gallo, 1989). Primary liver cancer in China was strongly linked to consuming drinking water from ditches highly polluted with agricultural runoff that presumably contained a variety of organic and other chemicals (Delong, 1979). In Clinton County, Pennsylvania, the location of the Drake Superfund Site, bladder cancer mortality among white males was significantly elevated (Budnick et al., 1984). A health survey in Hardeman County, Tennessee, where leachate from a pesticide waste dump contaminated the drinking water, found significant differences in hepatic profiles of exposed and unexposed members of the population that included alkaline phosphatase, albumin, total bilirubin, and SGOT (Clark et al., 1982). This may be of relevance to cancer induction, because detoxification of potentially carcinogenic compounds or conversion of procarcinogens to direct-acting carcinogens may be linked to enzyme profiles. The most frequently detected contaminants in the air of exposed homes were carbon tetrachloride and tetrachloroethylene. Cancer incidence of several sites was elevated in the vicinity of a solidwaste landfill in Montreal, Quebec (Goldberg et al., 1995; Goldberg et al., 1999). Exposures to chemicals from the Montreal landfill were assumed to be airborne. Statewide surveys of volatile organic chemicals (VOCs) in New Jersey drinking water have served as estimates of waterborne levels in studies of cancer incidence (Tucker, 1981; Fagliano et al., 1990; Cohn et al., 1993; Cohn et al., 1994). Leukemia incidence among females (but not males) in 27 New Jersey towns was associated with an index of volatile organic chemicals in municipal drinking water (Fagliano et al., 1990). A subsequent study in a 75town study area of New Jersey showed associations of trichloroethylene and/or perchloroethylene in drinking water with chronic myologenous leukemia and NHL among females and chronic lymphocytic leukemia among males and females (Cohn et al., 1993; Cohn et al., 1994).

Summary Although the number of epidemiologic studies with individual exposure data has increased in the past decade, there are still few studies for any single cancer site, making interpretation difficult. The recent analytic studies have generally included historical data of nitrate levels from public water supplies and have evaluated potential confounders and factors affecting nitrosation. However, the population using private wells that often have higher drinking water nitrate levels has usually been excluded due to a lack of long-term nitrate measurement data. Formation of NOC from drinking water nitrate intakes has been demonstrated in humans at nitrate levels above the MCL of 10 mg/L nitrate-N (Moller et al., 1989; Mirvish et al., 1992); however, further studies are needed to determine to what extent nitrosation occurs in vivo at the intermediate levels (5–10 mg/L) observed in many public water supplies. Nitrate levels in water supplies have been increasing in many areas worldwide; therefore, additional studies, preferably of populations with well-characterized and higher exposures, are warranted.

ORGANIC CHEMICALS FROM HUMAN COMMERCE Exposure Contamination of underground and surface water with organic chemicals from industrial, agricultural, commercial, and domestic sources, as well as from hazardous waste disposal sites, is increasingly found. Usually, the contamination is geographically restricted; however, chemicals such as agricultural pesticides and organic solvents used for cleaning cesspools may affect extensive aquifers (U.S. Environmental Protection Agency, Office of Pesticide Programs, 1988). Public water supplies are not immune. Leachate from lined water mains has resulted in system-wide exposure to perchloroethylene (Webler and Brown, 1993). Epidemiologic assessment of the health impact of such drinking water contaminants is challenging. This is due to the difficulty in estimating the levels, timing, and specific chemicals involved in past exposures, the relatively small populations typically exposed to high contaminant levels, and the problem of deciding which health end-points or intermediate biological markers to examine (Phillips and Silbergeld, 1985; Williams and Jalaludin, 1998; Vrijheid, 2000). When health effects are observed, it is often not possible to sort out multiple exposures. Of special concern are studies of hazardous waste sites, because the route of exposure often involves contamination of groundwater used for drinking. Aggregation of data from populations with similar exposures, a common practice used to enhance statistical power of epidemiologic observations from independent studies, is of limited utility when the composition and levels of specific exposures defy clear definition.

392

PART III: THE CAUSES OF CANCER

A cluster of childhood leukemia cases associated with contaminated community drinking water in Woburn, Massachusetts, has been the subject of scientific study, legal and political controversy, as well as a popular book and commercial film (Cutler et al., 1986; Lagakos et al., 1986; MacMahon et al., 1986; Byers et al., 1988; Harr, 1995; Costas et al., 2002; Grossman et al., 2002). Woburn (population 37,000) had been an industrial site for more than 130 years. Two of eight drinking water wells were discovered in 1979 to be contaminated with trichloroethylene (267 ppb), tetrachloroethylene (21 ppb), trichlorotrifluoroethane (23 ppb), and dichloroethylene (28 ppb) and subsequently closed. Forty-eight EPA priority pollutants and elevated levels of 22 metals were found in 61 additional test wells drilled to sample the groundwater (Lagakos et al., 1986). An elevated rate of childhood leukemia was found and statistically linked in space and time to the contamination (Lagakos et al., 1986; Costas et al., 2002). Associations were also observed for 2 of 5 categories of congenital anomalies and 2 of 9 categories of childhood disorders. In addition, lymphocyte abnormalities were noted among family members of cases (Byers et al., 1988). A population in upper Cape Cod, Massachusetts, was exposed to tetrachloroethylene from improperly cured vinyl-lined water distribution pipes (Webler and Brown, 1993). Population-based case-control studies in this region that gathered information on multiple risk factors found elevated risk for lung and breast cancers associated with estimated levels of tetrachloroethylene exposure, especially among subjects with latencies of long duration (Paulu et al., 1999; Aschengrau et al., 2003). In addition, suggestive associations were observed for leukemia as well as bladder, kidney, and colorectal cancers (Aschengrau et al., 1993; Paulu et al., 1999). No association was noted with brain or pancreas cancer. The elevated risks were observed primarily among subjects with water levels over the 90th percentile of the distribution. The 90th percentile of tetrachloroethylene level varied by study. It fell in the range 29.2–53.4 mg/L for all latencies. Documented groundwater contamination from a U.S. Department of Defense Superfund site in this region of Massachusetts resulted in other chemical exposures to some cases and controls. However, the number of exposed study subjects was too small for adequate evaluation of risk (Ozonoff et al., 1994). Love Canal in Niagara Falls, New York, served as a toxic waste disposal site from 1947 to 1952, and residential housing was later built adjacent to the area. Most exposure to residents likely was via inhalation, as volatile chemicals outgassed from the groundwater and contaminated the air in neighborhood houses. An assessment of cancer incidence rates for the period 1955–1977 revealed elevated lung cancer that was not consistent across age groups (Janerich et al., 1981). Other cancers were not elevated, but the statistical power to detect elevated rates of less common sites was limited. Rates of chromosomal aberrations and sister chromatid exchange frequencies were as expected (Heath et al., 1984). An elevated incidence of low birth weight was found for the period when chemicals were being dumped (Vianna and Polan, 1984).

Summary A straightforward summary is not possible of this heterogeneous group of studies of cancer after exposure to water contaminated with a variety of chlorinated organics and other industrial chemicals. Such studies are intrinsically opportunistic. They frequently involve exposure to poorly defined mixtures of chemicals, and where associations are observed, it may not be possible to link effects with exposure to specific chemicals. Descriptions of other unique exposure scenarios and their health impacts are summarized elsewhere (National Research Council: Committee on Environmental Epidemiology, 1991; National Research Council: Committee on Environmental Epidemiology, 1997). Health surveillance of potentially exposed populations in the vicinity of toxic waste disposal sites should continue to be of high priority.

RADIONUCLIDES Traces of natural and anthropogenic radioactivity from radionuclides are found in drinking water supplies throughout the United States. Levels vary geographically with local soil and rock conditions and may be increased by industrial and other point discharges, such as radiopharmaceutical production and use, radionuclide and weapons production, and nuclear power generation. The principal naturally occurring species include Ra-226, Ra-228, U, Rn-222, Pb-210, Po210, Th-230, and Th-232 (Cothern et al., 1986). In 1976, the EPA first promulgated regulations setting standards for maximum allowable levels of radionuclides in drinking water. Among other requirements, the regulations mandated monitoring of the almost 60,000 U.S. public water supplies (Federal Register, Vol. 41, No. 133, pp. 28404–9, 9 July 1976). Data from this monitoring program and other surveys have been used to estimate the occurrence of uranium (Cothern and Lappenbusch, 1983), radon-222 (Cross et al., 1985), radium (Mays et al., 1985), and total radioactivity (Hess et al., 1985) and the contribution of drinking water to total natural background radiation (Cross et al., 1985; Cothern et al., 1986; Nazaroff et al., 1987; National Research Council: Committee on Risk Assessment of Exposure to Radon in Drinking Water, 1999). Revised regulations have been issued (U.S. Environmental Protection Agency, 2000). Most of the population dose of alpha radiation derived from drinking water is from airborne radon released from water in dish and clothes washers, showers, baths, toilets, and from cooking, drinking, and cleaning (Nazaroff et al., 1987). The primary exposure is to the lungs from inhalation. Radon in water adds only a small increment to radon in air, with an estimated transfer coefficient of 1.0 ¥ 10-4. For example, 10,000 becquerel per cubic meter (Bq/m3) in water roughly contributes 1 Bq/m3 in air (National Research Council: Committee on Risk Assessment of Exposure to Radon in Drinking Water, 1999). Ingested radon is not thought to be important as an environmental determinant of cancer (National Research Council, 1977; National Research Council: Committee on Risk Assessment of Exposure to Radon in Drinking Water, 1999). On the average, water contributes less than 2% of airborne household radon. Although the predominant source of indoor airborne Rn-222 in most U.S. houses is the soil underlying and adjacent to the foundation, in some circumstances groundwater constitutes the predominant source (Hess et al., 1983). Radon-222 concentrations in groundwater vary over an extremely large range, from nearly zero to more than 106 Bq/m3, with a populationweighted average of 6900 Bq/m3 (Hess et al., 1985; National Research Council: Committee on Risk Assessment of Exposure to Radon in Drinking Water, 1999). At the higher end of the range, radon originating from water can make a substantial contribution to total exposure. Most epidemiologic studies of cancer and radon in dwellings have evaluated lung cancer as related to airborne radon without regard to its primary sources. Public water supplies that use surface water tend to have radon concentrations less than 4000 Bq/m3, whereas private water supplies are higher by factors of 3 to 20 (Hess et al., 1985). A committee of the National Academy of Sciences estimated that radon released from water may contribute 160 additional lung cancer deaths per year in the United States, about 0.8% of the total radon-associated lung cancer deaths. The committee also estimated ingestion of radon in water causes about 20 deaths from stomach cancer (National Research Council: Committee on Risk Assessment of Exposure to Radon in Drinking Water, 1999). Several investigations of cancer and radioactivity in drinking water are ecologic in design. A study of county leukemia incidence rates in Florida found a relative risk of 1.5 for total leukemia and 2.0 for acute myeloid leukemia in high groundwater radium vs. low radium counties (Lyman et al., 1985). In Maine, a study of county cancer mortality found an association between female lung cancer for 1950–1969 and average county radon concentrations in water (Hess et al., 1983). In Iowa, the incidence rates for cancers of the lung and bladder among males and cancers of the lung and breast among females were elevated in towns with a radium-226 level in the water supply exceeding 5.0 picocurie per liter (pCi/L) (Bean et al., 1982a) (1 pCi/L is equivalent to 37 Bq/m3 [National Research Council, 1999]). These associations

Water Contaminants could not be explained by smoking patterns, water treatment factors, or sociodemographic factors. A subsequent study in 59 Iowa towns revealed a small, increasing trend for total leukemia incidence with radium content in drinking water, consistent with either no or a small effect (Fuortes et al., 1990). Using county-level estimates of radon in groundwater, a North Carolina study noted no association with total adult leukemia deaths or mortality from esophagous, stomach, colon, or female breast cancer (Collman et al., 1988). Small, non-significant excesses were found for bone cancer among men and chronic leukemias (lymphocytic and nonlymphocytic) among women. A later study using similar methods found associations with all childhood cancers and childhood leukemia (Collman et al., 1991). A case-cohort study in Finland with 35 leukemia cases, which considered exposures on an individual-level, found no association with levels of uranium, radon, or radium-226 in drinking water (Auvinen et al., 2002). Several case-control studies have examined the association of airborne radon in homes with lung cancer risk; however, the contribution of drinking water to total household radon was not considered in these investigations. In the southern Urals of Russia, industrial releases to the Techa River from a plutonium production facility starting in 1950 resulted in substantial population exposure to long-lived radionuclides of cesium, strontium, and other elements (Balonov, 1997). Villages along the River were especially affected. Exposures occurred via air, water, and contaminated soil. The exposed population, which was partially relocated from the most contaminated areas, is being followed for subsequent health effects, including cancer incidence and mortality, with mounting evidence of excess leukemia and other malignancies (Akleyev et al., 1995; Kossenko et al., 2000; Kossenko et al., 2002; Reid et al., 2002). Refining of the dosimetry for individuals in the Techa River cohorts continues (Degteva et al., 2002; Shutov et al., 2002). Increased risk of osteosarcoma among radium dial painters after radium (-226 and -228) ingestion, and among radium-treated patients, suggests the possibility of elevated risk among consumers of drinking water with high radium levels (National Research Council, 1990). An early study in Iowa and Illinois observed associations of bone cancer mortality with level of radium-226 in drinking water, using town-level data (Petersen et al., 1966). Compared with towns having <1 pCi/L in water, relative risks in higher level Iowa and Illinois towns were 1.35 (avg. radium-226 = 3.4 pCi/L) and 1.19 (5.3 pCi/L), with males having higher risk than females in both states. In two recordbased case-control studies from Wisconsin, osteosarcoma was not associated with gross alpha level or with levels of radium-226 in community water systems serving study subjects (Moss et al., 1995; Guse et al., 2002). Elevated risk of bone cancer in Ontario youths (<25 years) was found in two case-control studies, the first based on information from death and birth registries and the second on interviews of incident cases and other-cancer controls that included lifetime residential information (Finkelstein, 1994; Finkelstein and Kreiger, 1996). Radium-226 was measured in household samples, and those with >0.2 pCi/L were considered exposed. The relative risk among all exposed in the combined studies was 1.48 (1.01–2.51) for all bone sarcoma, primarily osteosarcoma, with evidence of dose-response relationships (Finkelstein and Kreiger, 1996). Levels at the birth address conferred higher risk than average lifetime levels. In an ecologic study in 117 water service areas in New Jersey, an association of osteosarcoma with radium levels was found among males, but not females, with male rates about threefold in places with ≥4 pCi/L, compared with <0.5 pCi/L (Cohn et al., 2003). Dose-response gradients were observed, and risk was greater among males 25 and over. Further evaluation of the association between osteosarcoma risk and radium in drinking water is warranted, especially in large studies with individual information.

MICROBIOLOGICAL AGENTS Viruses, bacteria, and protozoa are the principal microbes in drinking water that can be transmitted to man by ingestion. Their concentra-

393

tions and sources vary greatly in different raw water sources, but the dramatic decreases and continued low incidence of waterborne infectious disease over the past 100 years underscore the effectiveness and importance of disinfection in maintaining acceptable low levels of these agents in finished water. An important characteristic is that these organisms exist in water primarily as aggregates, either with other microorganisms or with solid particulates; this has a bearing on their effectiveness as pathogens and on methods for their removal. These pathogens are generally not thought to be important as waterborne carcinogens, and they are discussed as noncarcinogens in a National Research Council review of this subject (National Research Council, 1977). There has been no evidence of waterborne transmission for several viruses with known or suggested links with human cancer (principally hepatitis B, human papilloma virus, Epstein-Barr virus, and human immunodeficiency virus-1 [HIV-1]).

Schistosoma Haematobium At least one waterborne microorganism, Schistosoma haematobium, has been associated with elevated bladder cancer rates in many tropical and subtropical countries where schistosomiasis is endemic (International Agency for Research on Cancer, 1994b; Johansson and Cohen, 1997), such as Egypt (Mustacchi and Shimkin, 1958; Bedwani et al., 1998), Sudan, Kenya, Senegal (el Mawla et al., 2001), Zambia (Elem and Purohit, 1983), and Zimbabwe (Thomas et al., 1990; Parkin et al., 1994; Vizcaino et al., 1994). Schistomiasis-related bladder cancer is usually of squamous-cell origin, whereas most other bladder carcinomas originate in transitional cells. In endemic areas, infections with S. haematobium start at an early age, persist for many years, and are related to bladder cancers of early onset. In a series of Egyptian bladder cancer patients, the average age of those found with schistosome eggs at cystoscopy was 46.7 years (El-Bolkainy et al., 1981). In places where schistosomal infection is not endemic, the median age of bladder cancer patients is 69 years (National Cancer Institute, 1992). It is not yet known whether the carcinogenic effects of S. haematobium infections are due to chronic inflammation of the bladder epithelium or the release of chemical carcinogens such as Nnitroso compounds from the eggs (Cheever, 1978; Tricker et al., 1991; Mostafa et al., 1994; Badawi et al., 1995). Infections with S. mansoni and S. japonicum, which primarily affect the liver and intestines, may increase risk of hepatocellular carcinoma and colon cancer (Ishii et al., 1994).

Helicobacter Pylori Helicobacter pylori infection is a risk factor for stomach cancer, gastric lymphoma, and possibly other cancers (International Agency for Research on Cancer, 1994a; Ahmad et al., 2003; Bjorkholm et al., 2003). The mechanism of action is not well understood but may involve production of carcinogenic N-nitroso compounds. In addition to low socioeconomic status, poor hygiene, and crowded living conditions (Brown, 2000), contaminated drinking water has been identified as a risk factor for infection, as measured serologically or by other methods, in many settings. Investigations with positive drinking water associations for infection have been conducted in areas such as rural Japan (Karita et al., 2003), Germany (Mazari-Hiriart et al., 2001), Kazakhstan (Nurgalieva et al., 2002), China (Shi et al., 1997; Brown et al., 2002; Guo et al., 2002), Ethiopia (Lindkvist et al., 1999), Gambia (Bunn et al., 2002), Bolivia (Glynn et al., 2002), Colombia (Goodman et al., 1996), and Peru (Klein et al., 1991). In a few other settings, water source was not linked with infection (Teh et al., 1994; Guo et al., 2002). Contamination of drinking water by H. pylori has also been detected directly in drinking water samples (Baker and Hegarty, 2001; Horiuchi et al., 2001; Mazari-Hiriart et al., 2001; Lu et al., 2002). Water chlorination at concentrations typically used in drinking water treatment is effective in inactivating H. pylori (Johnson et al., 1997).

394

PART III: THE CAUSES OF CANCER

ASBESTIFORM FIBERS Asbestos is a vernacular term referring to a large group of “naturally occurring hydrated silicate minerals possessing fibrous morphology and commercial utility” (National Research Council, 1977). Six minerals are called “asbestos”: actinolite, amosite, anthophyllite, crocidolite, chrysotile, and tremolite. Asbestos fibers are widely distributed in the aqueous environment, with higher concentrations usually found near cities and industrial centers. Asbestos fibers reach drinking water primarily through weathering from natural deposits such as serpentine, release from asbestos-cement pipes, and from processes associated with mining and production of various ores, including taconite, a type of iron ore (Langer et al., 1979; Millette et al., 1983a). Concentrations vary enormously, ranging from a barely detectable background level of 104 fibers/liter, to more than 1011 fibers/liter (Millette et al., 1983a). For human carcinogenesis, the size, shape, and crystalline structure of these fibers is as important as concentration, and these characteristics are modified by physicochemical processes resulting from exposure to water or gastric fluid (Seshan, 1983). Airborne asbestos fibers are human carcinogens of the respiratory tract and, possibly, because of lung clearance and subsequent swallowing, the gastrointestinal tract. Epidemiologic studies of populations served by water containing high concentrations of asbestos have failed to yield conclusive results. All but two studies (Polissar et al., 1984; Andersen et al., 1993) involving asbestos in drinking water are ecologic. Related studies in the San Francisco Bay area (Kanarek et al., 1980; Conforti et al., 1981) found associations between measured (naturally occurring) asbestos in drinking water and census tract incidence rates for cancers of the esophagus, stomach, and pancreas in both sexes (Conforti et al., 1981), of the lung in males, and of the gall bladder and peritoneum in females (Kanarek et al., 1980). However, potentially confounding factors such as diet, smoking, and occupation could not be adequately controlled. Mortality in Quebec communities was associated with asbestos in drinking water for cancers of the stomach (males), pancreas (females), and lung (males) (Wigle, 1977). Neither an ecologic study (Polissar et al., 1982) nor a case-control study (Polissar et al., 1984) in the Puget Sound region, both based on incident cancer, found overall patterns consistent with an asbestos effect. However, positive associations for male stomach cancer, based on small numbers, were observed in the case-control study. A cohort study of incident cancer among Norwegian lighthouse keepers exposed to asbestos in their drinking water also reported excess stomach cancer (Andersen et al., 1993). The excess was 2.4-fold and statistically significant among persons first exposed at least 20 years prior to diagnosis. Duluth, Minnesota, had elevated asbestos in its drinking water from 1955 to 1973, due to contamination of Lake Superior with tailings from an iron ore (taconite) processing facility 60 miles away. Cancer mortality (Mason et al., 1974) and incidence rates (Levy et al., 1976; Sigurdson et al., 1981) in Duluth or its county were compared to other Minnesota cities (or counties). Some excesses of gastrointestinal mortality and morbidity were observed, with inconsistent patterns. Cancer incidence as related to drinking water distributed by asbestos-cement water mains has been evaluated in Connecticut (Harrington et al., 1978; Meigs et al., 1980), Utah (Sadler et al., 1984), and Woodstock, New York (Howe et al., 1989) and mortality studied in Escambria County, Florida (Millette et al., 1983b), with inconsistent findings. In Utah, an association was found for kidney cancer (males) and leukemia (females), and in Woodstock, New York, for cancer of the oral cavity. Laboratory studies cannot provide conclusive evidence, principally because of methodological problems and the absence of suitable animal model systems (National Research Council, 1977). Although little attention has been focused in recent years on carcinogenic effects of ingested asbestiform fibers, epidemiologic and experimental research should nevertheless continue, as the available data cannot rule out adverse health effects of these carcinogens in water. In addition, asbestos in drinking water should be considered as a potential risk factor whenever large case-control studies of stomach, kidney, pan-

creas, or other cancers are conducted in regions with potential sources of asbestos-contaminated water.

INORGANIC SOLUTES OTHER THAN ARSENIC Trace Metals Other than arsenic, there has been relatively little study of trace metals in water and cancer risk. Trace metals are present in a wide range of concentrations in U.S. drinking waters (usually under 100 mg/L) and can increase or decrease during water treatment. The sources include leaching from soil or distribution systems, industrial or mining activities, or water treatment itself. An early correlational study (Berg and Burbank, 1972) found several significant links between site-specific cancer mortality in U.S. sites and drinking water concentrations of 8 trace metals in 16 major water basins. The most frequent associations were with beryllium, cadmium, and lead. An Iowa study observed geographic correlations between town-level cancer incidence of lung and bladder cancers and levels of nickel in drinking water (Isacson et al., 1985). The authors suggested that nickel was not directly implicated but an indicator of other contamination. A Norwegian study in 97 municipalities evaluated the correlation of 17 inorganic ions in drinking water with 16 groups of cancer morbidity (Flaten and Bolviken, 1991). Several significant associations were found, many of which could have been due to chance.

Hardness, Magnesium, and Calcium Water hardness is primarily due to magnesium and calcium in water and is defined as the concentrations of these cations expressed in terms of calcium carbonate. Definitions of “soft” and “hard” water vary somewhat. Water with <60 mg/L is generally considered soft and water with >180 mg/L very hard. Water hardness, and magnesium and calcium levels specifically, have received limited attention in epidemiologic studies of cancer. In an early geographic correlation study from England and Wales that used relative proportionate mortality as the outcome measure, no association was found between water hardness and cancers of the stomach, bladder, and female breast (Stocks, 1973). There was a suggestion of a positive link with prostate cancer. An ecologic study that compared rates among three communities in Los Angeles observed no association of total cancer with water hardness among males or females (Allwright et al., 1974). In Taiwan, casecontrol studies based on death certificates reported observations of cancer mortality of several anatomic sites with total water hardness (Yang et al., 1997a; Yang and Hung, 1998; Yang et al., 1999a; Yang et al., 1999b; Yang et al., 1999c) and with calcium and magnesium specifically (Yang et al., 1997b; Yang and Chiu, 1998; Yang et al., 1998; Yang et al., 2000; Yang et al., 2002a; Yang et al., 2002b). Protective associations with total water hardness were found for cancers of the colon (Yang and Hung, 1998), rectum (Yang et al., 1999c), esophagus (Yang et al., 1999a), pancreas (Yang et al., 1999b), and stomach (Yang et al., 1997a). Calcium levels in drinking water were negatively associated with cancers of the colon (Yang et al., 1997b), rectum (Yang and Chiu, 1998), and stomach (Yang et al., 1998), whereas magnesium levels were negatively associated with esophageal (Yang et al., 2002b) and prostate cancers (Yang et al., 2000). There was no association with liver cancer (Yang et al., 2002a). These studies of total water hardness, calcium, and magnesium in water are limited in that they examined mortality, not incidence; information on risk factors other than age, sex, and urbanization were not available; and levels were estimated at the time of death, not during a presumably lengthy latent period. Findings for waterborne calcium and lower risk of colorectal cancer are consistent with protective associations observed in studies of dietary calcium (Giovannucci, 2003). Further study of these initial observations of protection associated with water hardness and magnesium and calcium levels is warranted.

Fluoride Fluoride is found in most natural waters at concentrations that rarely exceed 5.0 mg/L and are usually less than 1.0 mg/L. Its natural source

Water Contaminants is mainly the dissolution from minerals, such as apatite, amphibole, and fluorite. The principal source in many communities is the successful prophylactic addition of fluoride to prevent dental caries, with a usual level of 0.7–1.0 mg/L. Fluoride in drinking water came under suspicion as a carcinogen when an analysis of cancer mortality showed that the overall rates in the 10 largest U.S. cities that practiced water fluoridation were significantly higher than those of the 10 largest that did not, with sitespecific differences for cancers of the urinary and gastrointestinal tracts, female breast, and ovary (Yiamouyiannis and Burk, 1977). These differences were not found when relevant sociodemographic variables were taken into account (Hoover et al., 1976; Chilvers, 1983). Additional studies, almost all of ecologic design, provided no supporting evidence. Given the important public health implications of this question, the epidemiologic findings on fluoridation and cancer risk were independently reviewed by an international panel (International Agency for Research on Cancer, 1982) and by three separate expert committees convened in the United States (National Research Council, 1977; U.S. Public Health Service.Committee to Coordinate Environmental Health and Related Programs, 1991) and Great Britain (Knox, 1985). These reviews concluded that the available evidence does not support the hypothesis that fluoride in drinking water influences cancer risk. However, concern was again fueled when “equivocal evidence” of excess osteosarcomas was reported in a lifetime sodium fluoride feeding study in rodents was found (Bucher et al., 1991). In the period since, epidemiologic assessments have found no time trend nor geographic pattern of bone cancer or osteosarcoma consistent with a causal role for fluoride in drinking water (Hrudey et al., 1990; Hoover et al., 1991; Mahoney et al., 1991; McGuire et al., 1991). The available evidence, quite extensive at this time, does not indicate that fluoride in drinking water is associated with elevated risk of cancer. However, most studies to date are ecologic in design, and further confidence in the safety of water fluoridation would be enhanced by showing no association in analytic studies with individual information on exposure and outcome.

RESEARCH PRIORITIES AND PREVENTION STRATEGIES Based on existing evidence, the water we drink contains a complex mixture of known or suspected carcinogenic substances, typically found at trace level concentrations (<100 ppb). The following relationships have been demonstrated or suggested: (1) inorganic arsenic (and possibly other trace metals) with cancer of several sites, including nonmelanoma skin cancer, bladder, lung, and kidney, (2) synthetic organic chemicals (especially disinfection byproducts) with cancers of the urinary bladder and possibly the large bowel, (3) radium with osteosarcoma, (4) radon in water with lung cancer, through its contribution to airborne levels in homes, (5) nitrate with cancers of the gastrointestinal tract and other sites, and (6) water hardness and magnesium and calcium content with protection at several cancer sites. For these waterborne substances, hypothesis-generating work, as represented by aggregate (ecological) studies, has been adequately conducted. To improve our understanding of the important relationships, further case-control or cohort studies based on individual information are indispensable, especially using refined chemical measures of exposure and appropriate biomarkers whenever possible. However, in the absence of precise data on levels that occurred 20 years or more in the past, three paths of epidemiologic investigation are possible: (1) estimating past exposures to contaminants through the use of models that apply our current understanding of water chemistry to knowledge of historical sources and treatment modalities, (2) refinement and continued use of surrogate exposure data, and (3) studies of persons occupationally exposed to elevated concentrations of specific waterborne substances of concern, such as chloroform or arsenic. Although the last option is necessary to determine the carcinogenic potential of specific substances and to quantify risk for standard setting, it cannot be the only approach taken. In many cases, water is a unique source of exposure to the chemicals or mixtures of concern.

395

Future epidemiologic investigations should profit from the recognized limitations of previous work. Specific areas include the following: (1) Studies relating individual exposure to individual risk of cancer are generally preferred to ecologic studies. Except in selected circumstances, aggregate studies have already met their goals. (2) The estimation of exposures for many decades in the past is of central importance to account for latency of adult cancers. (3) In most circumstances, small relative risks are expected, and large sample sizes adequate to detect the small differences in risk are necessary. (4) Because the carcinogenic risk conferred by drinking water contaminants is likely to be relatively small, the most valuable studies are those where information is available, at an individual level, on potentially confounding factors, such as occupation, diet, smoking, physical activity, and other factors that may contribute to risk of the cancer site under consideration. (5) Better estimates of past exposure are needed, including more information on groundwater, types of treatment practices, presence and type of upstream discharges for surface sources, and surveillance of major chemical contaminants in water supplies. This often requires close collaboration with environmental chemists, hydrogeologists, specialists in geographic information systems (GIS), and others. (6) Investigations should collect and validate information on individual consumption of tap water and other fluids, including water used in cooking and in beverages such as coffee and tea. (7) Understanding of carcinogenic risks and mechanisms of action will be enhanced by analysis of tissue (blood, urine, toenails, etc.) for markers of exposure and biological effect. Strategies to minimize exposure to water contaminants include various watershed protection programs and water treatment options, some of which are expensive and technologically complex. Pollution control and watershed management can limit sources of industrial, domestic, and agricultural pollution and should be a part of any strategy to minimize contamination of water sources. The use of alternate, unpolluted source waters can reduce exposures. However, this may not be an option because these sources are often unavailable or not economically feasible. Therefore, options to reduce chemical exposures are generally limited to the following: (1) reduction of disinfectant byproducts through application of alternate disinfection procedures or more selective use of existing disinfection methods, and (2) removal of organic, inorganic, and particulate contaminants through use of more advanced water treatment technologies. Disinfection by-products are a special class of contaminants, as they arise from interaction with chemicals intentionally added to water during treatment. Chlorine has generally served well as a costeffective disinfectant to prevent transmission of waterborne disease. The use of alternate disinfectants such as ozone, chlorine dioxide, and chloramine will reduce chlorinated by-products. However, they may result in other potentially toxic by-products (Richardson et al., 2002) and may not be feasible or effective for all communities. Before alternative disinfectants are considered, their effectiveness and potential by-products must be fully evaluated. In water treatment plants, moving the point of chlorination from early to late in the process can result in lowering the concentration of chlorinated by-products. However, caution is required to ensure this process does not increase microbial risk. Concerns about cancer risks that may be associated with longterm exposure to disinfected water must be tempered with consideration of the benefits provided by water disinfection, especially to vulnerable populations such as young children. Disinfection is the final barrier against transmission of waterborne pathogens, but it must not be the sole barrier. Source protection is required and water filtration necessary for many surface waters. Properly designed and operated filtration plants make disinfection more effective by reducing microbiological contamination, turbidity, and other substances that exert chlorine demand and may interfere with the efficacy of the process. The amount of disinfectant required is also reduced, thereby decreasing disinfection by-products. Advanced water treatment processes, such as activated carbon adsorption, ion exchange, desalinization, and membrane microfiltration, can be used to remove contaminants not usually removed during conventional water filtration/disinfection. These processes are usually designed for specific contaminants or classes of contaminants;

396

PART III: THE CAUSES OF CANCER

however, some processes also reduce or remove a wide range of organic and microbial contaminants. They are often expensive and complex to operate. Advanced treatment may therefore be an option only for affluent, more technologically advanced communities. Where advanced treatment is feasible, chemical exposures from drinking water can be reduced. However, it is not possible at present to quantify the reduction in carcinogenic risks. Further epidemiologic studies can help quantify the relative benefits of different levels of water treatment, but some communities may prefer not to wait for this information before reducing exposures. References Agency for Toxic Substances and Disease Registry. 2000. Toxicological Profile for Arsenic. Atlanta, Ga.: Public Health Service, U.S. Department of Health and Human Services. Aggazzotti G, Fantuzzi G, Righi E, Tartoni P, Cassinadri T, Predieri G. 1993. Chloroform in alveolar air of individuals attending indoor swimming pools. Arch Environ Health 48:250–254. Aggazzotti G, Fantuzzi G, Tartoni PL, Predieri G. 1990. Plasma chloroform concentration in swimmers using indoor swimming pools. Arch Environ Health 45:175–179. Ahmad A, Govil Y, Frank BB. 2003. Gastric mucosa–associated lymphoid tissue lymphoma. Am J Gastroenterol 98:975–986. Ahmad SA, Sayed SU, Barua S, Khan MH, Jalil A, Hadi SA, Talukder HK. 2001. Arsenic in drinking water and pregnancy outcomes. Environ Health Perspect 109:629–631. Akleyev AV, Kossenko MM, Silkina LA, Degteva MO, Yachmenyov VA, Awa A, Akiyama M, Veremeyeva GA, Vozilova AV, Kyozumi S. 1995. Health effects of radiation incidents in the southern Urals Stem Cells 13(Suppl 1):58–68. Alavanja M, Goldstein I, Susser M. 1977. A case-control study of gastrointestinal and urinary tract cancer mortality and drinking water chlorination. In: Jolley RL, Gorchev H, Hamilton DH, Jr., eds. Water Chlorination, Environmental Impact and Health Effects, Vol. 2. Ann Arbor, Mich.: Ann Arbor Science Publishers, pp. 395–409. Allen-Price ED, Birm MD. 1960. Uneven distribution of cancer in West Devon with particular reference to the divers water-supplies. Lancet i:1235–1238. Allwright SP, Coulson A, Detels R, Porter CE. 1974. Mortality and waterhardness in three matched communities in Los Angeles. Lancet ii: 860–864. Andersen A, Glattre E, Johansen BV. 1993. Incidence of cancer among lighthouse keepers exposed to asbestos in drinking water. Am J Epidemiol 138:682–687. Andrew AS, Karagas MR, Hamilton JW. 2003. Decreased DNA repair gene expression among individuals exposed to arsenic in United States drinking water. Int J Cancer 104:263–268. Aposhian HV, Gurzau ES, Le XC, Gurzau A, Healy SM, Lu X, Ma M, Yip L, Zakharyan RA, Maiorino RM, Dart RC, Tircus MG, Gonzalez-Ramirez D, Morgan DL, Avram D, Aposhian MM. 2000. Occurrence of monomethylarsonous acid in urine of humans exposed to inorganic arsenic. Chem Res Toxicol 13:693–697. Arbuckle TE, Hrudey SE, Krasner SW, Nuckols JR, Richardson SD, Singer P, Mendola P, Dodds L, Weisel C, Ashley DL, Froese KL, Pegram RA, Schultz IR, Reif J, Bachand AM, Benoit FM, Lynberg M, Poole C, Waller K. 2002. Assessing exposure in epidemiologic studies to disinfection by–products in drinking water:report from an international workshop. Environ Health Perspect 110 (Suppl 1):53–60. Aschengrau A, Ozonoff D, Paulu C, Coogan P, Vezina R, Heeren T, Zhang Y. 1993. Cancer risk and tetrachloroethylene (PCE) contaminated drinking water in Massachusetts. Arch Environ Health 48:284–292. Aschengrau A, Rogers S, Ozonoff D. 2003. Perchloroethylene-contaminated drinking water and the risk of breast cancer:additional results from Cape Cod, Massachusetts, USA. Environ Health Perspect 111:167– 174. Auvinen A, Kurttio P, Pekkanen J, Pukkala E, Ilus T, Salonen L. 2002. Uranium and other natural radionuclides in drinking water and risk of leukemia:a case-cohort study in Finland. Cancer Causes Control 13:825–829. Backer LC, Ashley DL, Bonin MA, Cardinali FL, Kieszak SM, Wooten JV. 2000. Household exposures to drinking water disinfection by—products: Whole blood trihalomethane levels. J Expo Anal Environ Epidemiol 10:321–326. Badawi AF, Mostafa MH, Probert A, O’Connor PJ. 1995. Role of schistosomiasis in human bladder cancer:evidence of association, aetiological factors, and basic mechanisms of carcinogenesis. Eur J Cancer Prev 4:45–59.

Baker DB, Greenland S, Mendlein J, Harmon P. 1988. A health study of two communities near the Stringfellow waste disposal site. Arch Environ Health 43:325–334. Baker KH, Hegarty JP. 2001. Presence of Helicobacter pylori in drinking water is associated with clinical infection. Scand J Infect Dis 33:744–746. Balonov MI. 1997. Internal exposure of populations to long-lived radionuclides released into the environment. In: Cardew G, Bock G, Lake JV. eds. Health Impacts of Large Releases of Radionuclides: Ciba Foundation Symposium No. 203. Chichester, U.K.: Wiley, pp. 120–138. Barrett JH, Parslow RC, McKinney PA, Law GR, Forman D. 1998. Nitrate in drinking water and the incidence of gastric, esophageal, and brain cancer in Yorkshire, England. Cancer Causes Control 9:153–159. Bates MN, Rey OA, Biggs ML, Hopenhayn C, Moore LE, Kalman D, Steinmaus C, Smith AH. 2004. Case-control study of bladder cancer and exposure to arsenic in Argentina. Am J Epidemiol 159:381–389. Bates MN, Smith AH, Cantor KP. 1995. Case-control study of bladder cancer and arsenic in drinking water. Am J Epidemiol 141:523–530. Bates MN, Smith AH, Hopenhayn-Rich C. 1992. Arsenic ingestion and internal cancers: A review. Am J Epidemiol 135:462–476. Bean JA, Isacson P, Hahne RMA, Kohler J. 1982a. Drinking water and cancer incidence in Iowa:II. Radioactivity in drinking water. Am J Epidemiol 116:924–932. Bean JA, Isacson P, Hausler WJ, Kohler J. 1982b. Drinking water and cancer incidence in Iowa:I. Trends and incidence by source of drinking water and size of municipality. Am J Epidemiol 116:912–923. Beck BD, Boardman PD, Hook GC, Rudel RA, Slayton TM, Carlson-Lynch H. 1995. Response to Smith et al. [Letter]. Environ Health Perspect 103:15–17. Bedwani R, Renganathan E, El Kwhsky F, Braga C, Abu Seif HH, Abul AT, Zaki A, Franceschi S, Boffetta P, La Vecchia C. 1998. Schistosomiasis and the risk of bladder cancer in Alexandria, Egypt. Br J Cancer 77:1186–1189. Bellar TA, Lichtenberg JJ. 1974. Determining volatile organics at microgramper-litre levels by gas chromatograpy. J Am Water Works Assoc 66:739–744. Berg JW, Burbank F. 1972. Correlations between carcinogenic trace metals in water supplies and cancer mortality. Ann N Y Acad Sci 199:249–264. Berg M, Tran HC, Nguyen TC, Pham HV, Schertenleib R, Giger W. 2001. Arsenic contamination of groundwater and drinking water in Vietnam:a human health threat. Environ Sci Technol 35:2621–2626. Bergoglio RM. 1964. Mortalidad por cancer en zonas de aguas arsenicales de la Provincia de Cordoba, Republica Argentina. Pren Med Argent 51:994–998. Berry JP, Galle P. 1994. Selenium-arsenic interaction in renal cells: Role of lysosomes. Electron microprobe study. J Submicrosc Cytol Pathol 26:203–210. Biswas S, Talukder G, Sharma A. 1999. Prevention of cytotoxic effects of arsenic by short–term dietary supplementation with selenium in mice in vivo. Mutat Res 441:155–160. Bjorkholm B, Falk P, Engstrand L, Nyren O. 2003. Helicobacter pylori: Resurrection of the cancer link. J Intern Med 253:102–119. Boadu M, Osae EK, Golow AA, Serfor-Armah Y, Nyarko BJB. 2001. Determination of arsenic in some water bodies, untreated ore and tailing samples at Konongo in the Ashanti region of Ghana and its surrounding towns and villages by instrumental neutron activation analysis. J Radioanal Nucl Chem 249:581–585. Boeing H, Frentzel-Beyme R, Berger M, Berndt V, Gores W, Korner M, Lohmeier R, Manarcher A, Mannl HFK, Meinhardt M, Muller R, Ostermeier H, Paul F, Schwemmle K, Wagner KH, Wahrendorf J. 1991. Case-control study on stomach cancer in Germany. Int J Cancer 47: 858–864. Boonchai W, Walsh M, Cummings M, Chenevix–Trench G. 2000. Expression of p53 in arsenic-related and sporadic basal cell carcinoma. Arch Dermatol 136:195–198. Boorman GA. 1999. Drinking water disinfection byproducts:review and approach to toxicity evaluation. Environ Health Perspect 107 (Suppl 1):207–217. Borzsonyi M, Bereczky A, Rudnai P, Csanady M, Horvath A. 1992. Epidemiologic studies on humans–subjects exposed to arsenic in drinking water in southeast Hungary. Arch Toxicol 66:77–78. Brenniman GR, Vasilomanolakis-Lagos J, Amsel J, Namekata T, Wolff AH. 1980. Case-control study of cancer deaths in Illinois communities served by chlorinated or nonchlorinated water. In: Jolley RL, Brungs WA, Cumming RB. eds. Water Chlorination:Environmental Impact and Health Effects, Vol. 3. Ann Arbor, Mich.: Ann Arbor Science: pp. 1043–1057. Brouwer OF, Onkenhout W, Edelbroek PM, de Kom JF, de Wolff FA, Peters AC. 1992. Increased neurotoxicity of arsenic in methylenetetrahydrofolate reductase deficiency. Clin Neurol Neurosurg 94:307–310.

Water Contaminants Brown LM. 2000. Helicobacter pylori:epidemiology and routes of transmission. Epidemiol Rev 22:283–297. Brown LM, Thomas TL, Ma JL, Chang YS, You WC, Liu WD, Zhang L, Pee D, Gail MH. 2002. Helicobacter pylori infection in rural China:demographic, lifestyle and environmental factors. Int J Epidemiol 31:638– 645. Bucher JR, Hejtmancik MR, Toft JD, II, Persing RL, Eustis SL, Haseman JK. 1991. Results and conclusions of the National Toxicology Program’s rodent carcinogenicity studies with sodium fluoride. Int J Cancer 48:733–737. Buchet JP, Lauwerys R. 1985. Study of inorganic arsenic methylation by rat liver in vitro:relevance for the interpretation of observations in man. Arch Toxicol 57:125–129. Buchet JP, Lauwerys R. 1987. Study of factors influencing the in vivo methylation of inorganic arsenic in rats. Toxicol Appl Pharmacol 91:65–74. Buchet JP, Lauwerys R, Roels H. 1981. Comparison of the urinary excretion of arsenic metabolites after a single oral dose of sodium arsenite, monomethylarsonate, or dimethylarsinate in man. Int Arch Occup Environ Health 48:71–79. Budnick LD, Sokal DC, Falk H, Logue JN, Fox JM. 1984. Cancer and birth defects near the Drake superfund site, Pennsylvania. Arch Environ Health 39:409–413. Bunn JE, MacKay WG, Thomas JE, Reid DC, Weaver LT. 2002. Detection of Helicobacter pylori DNA in drinking water biofilms:implications for transmission in early life. Lett Appl Microbiol 34:450–454. Byers VS, Levin AS, Ozonoff DM, Baldwin RW. 1988. Association between clinical symptoms and lymphocyte abnormalities in a population with chronic domestic exposure to industrial solvent–contaminated domestic water supply and a high incidence of leukaemia. Cancer Immunol Immunother 27:77–81. Calafat AM, Kuklenyik Z, Caudill SP, Ashley DL. 2003. Urinary levels of trichloroacetic acid, a disinfection by-product in chlorinated drinking water, in a human reference population. Environ Health Perspect 111:151–154. Cantor KP. 1997. Drinking water and cancer. Cancer Causes Control 8:292–308. Cantor KP, Hildesheim ME, Alavanja MCR, Craun GF, Dosemeci M, Lubin J, Lynch CF. 2005a. Drinking water source and chlorination byproducts in Iowa. IV. Risk of renal cell carcinoma (submitted). Cantor KP, Hildesheim ME, Alavanja MCR, Craun GF, Dosemeci M, Lubin J, Lynch CF. 2005b. Drinking water source and chlorination byproducts in Iowa. V. Risk of pancreas cancer (submitted). Cantor KP, Hoover R, Hartge P, Mason TJ, Silverman DT, Altman R, Austin DF, Child MA, Key CR, Marrett LD, Myers MH, Narayana AS, Levin LI, Sullivan JW, Swanson GM, Thomas DB, West DW. 1987. Bladder cancer, drinking water source, and tap water consumption: A case–control study. J Natl Cancer Inst 79:1269–1279. Cantor KP, Hoover R, Mason TJ, McCabe LJ. 1978. Associations of cancer mortality with halomethanes in drinking water. J Natl Cancer Inst 61:979–985. Cantor KP, Lynch CF, Hildesheim ME, Dosemeci M, Lubin J, Alavanja M, Craun G. 1998. Drinking water source and chlorination byproducts. I. Risk of bladder cancer. Epidemiology 9:21–28. Cantor KP, Lynch CF, Hildesheim ME, Dosemeci M, Lubin J, Alavanja M, Craun GF. 1999. Drinking water source and chlorination byproducts in Iowa. III. Risk of brain cancer. Am J Epidemiol 150:552–560. Carlo GL, Mettlin CJ. 1980. Cancer incidence and trihalomethane concentrations in a public drinking water system. Am J Public Health 70:523–525. Carlson–Lynch H, Beck BD, Boardman PD. 1994. Arsenic risk assessment. Environ Health Perspect 102:354–356. Cebrian ME, Albores A, Aguilar M, Blakely E. 1983. Chronic arsenic poisoning in the north of Mexico. Human Toxicol 2:121–133. Cheever AW. 1978. Schistosomiasis and neoplasia. J Natl Cancer Inst 61: 13–18. Chen C-J, Chen CW, Wu M-M, Kuo T-L. 1992. Cancer potential in liver, lung, bladder and kidney due to ingested inorganic arsenic in drinking water. Br J Cancer 66:888–892. Chen C-J, Chuang Y-C, Lin T-M, Wu H-Y. 1985. Malignant neoplasms among residents of a blackfoot disease-endemic area in Taiwan: High-arsenic artesian well water and cancers. Cancer Res 45:5895–5899. Chen C-J, Kuo T-L, Wu M-M. 1988a. Arsenic and cancers. [Letter]. Lancet i:414–415. Chen CJ, Wu MM, Lee SS, Wang JD, Cheng SH, Wu HY. 1988b. Atherogenicity and carcinogenicity of high-arsenic artesian well water. Multiple risk factors and related malignant neoplasms of blackfoot disease. Arteriosclerosis 8:452–460. Chen C-J, Wang C-J. 1990. Ecological correlation between arsenic level in well water and age-adjusted mortality from malignant neoplasms. Cancer Res 50:5470–5474.

397

Chen YC, Guo YLL, Su HJJ, Hsueh YM, Smith TJ, Lee MS, Chao SC, Lee JYY, Christiani DC. 2003a. Arsenic methylation and skin cancer risk in southwestern Taiwan. J Occup Environ Med 45:241–248. Chen YC, Su HJ, Guo YL, Hsueh YM, Smith TJ, Ryan LM, Lee MS, Christiani DC. 2003b. Arsenic methylation and bladder cancer risk in Taiwan. Cancer Causes Control 14:303–310. Chilvers C. 1983. Cancer mortality and fluoridation of water supplies in 35 US cities. Int J Epidemiol 12:397–404. Chilvers C, Inskip H, Caygill C, Bartholomew B, Fraser P, Hill M. 1984. A survey of dietary nitrate in well–water users. Int J Epidemiol 13:324–331. Chiou H-Y, Hsueh Y-M, Liaw K-F, Horng S-F, Chiang M-H, Pu Y-S, Lin JSN, Huang C-H, Chen C-J. 1995. Incidence of internal cancers and ingested inorganic arsenic: A seven-year follow-up study in Taiwan. Cancer Res 55:1296–1300. Chiou H-Y, Wei M-L, Tseng C-H, Hsu Y-H, Chien Y-H, Chen C-J. 2001. Incidence of transitional cell carcinoma and arsenic in drinking water:A follow-up study of 8102 residents in an arseniasis-endemic area in northeastern Taiwan. Am J Epidemiol 153:411–418. Chiou HY, Hsueh YM, Hsieh LL, Hsu LI, Hsu YH, Hsieh FI, Wei ML, Chen HC, Yang HT, Leu LC, Chu TH, Chen-Wu C, Yang MH, Chen CJ. 1997. Arsenic methylation capacity, body retention, and null genotypes of glutathione S-transferase M1 and T1 among current arsenic-exposed residents in Taiwan. Mutat Res 386:197–207. Clark CS, Meyer CR, Gartside PS, Majeti VA, Specker B, Balistreri WF, Elia VJ. 1982. An environmental health survey of drinking water contamination by leachate from a pesticide waste dump in Hardeman County, Tennessee. Arch Environ Health 37:9–18. Cocco P, Broccia G, Aru G, Casula P, Muntoni S, Cantor KP, Ward MH. 2003. Nitrate in community water supplies and incidence of non-Hodgkin’s lymphoma in Sardinia, Italy. J Epidemiol Community Health 57:510–511. Cohn P, Bove F, Klotz J, Berkowitz M, and Fagliano J. 1993. Drinking Water Contamination and the Incidence of Leukemia and Non-Hodgkin’s Lymphoma. Trenton, N.J.: Department of Health. Cohn P, Klotz J, Bove F, Berkowitz M, Fagliano J. 1994. Drinking water contamination and the incidence of leukemia and non-Hodgkin’s lymphoma. Environ Health Perspect 102:556–561. Cohn P, Skinner R, Burger S, Fagliano J, and Klotz J. 2003. Radium in Drinking Water and the Incidence of Osteosarcoma. Trenton, N.J.: New Jersey Department of Environmental Protection. Collman GW, Loomis DP, Sandler DP. 1988. Radon-222 concentration in groundwater and cancer mortality in North Carolina. Int Arch Occup Environ Health 61:13–18. Collman GW, Loomis DP, Sandler DP. 1991. Childhood cancer mortality and radon concentration in drinking water in North Carolina. Br J Cancer 63:626–629. Concha G, Nermell B, Vahter MV. 1998a. Metabolism of inorganic arsenic in children with chronic high arsenic exposure in northern Argentina. Environ Health Perspect 106:355–359. Concha G, Vogler G, Lezcano D, Nermell B, Vahter M. 1998b. Exposure to inorganic arsenic metabolites during early human development. Toxicol Sci 44:185–190. Conforti PM, Kanarek MS, Jackson LA, Cooper RC, Murchio JC. 1981. Asbestos in drinking water and cancer in the San Francisco Bay area: 1969–1974 incidence. J Chron Dis 34:211–224. Coss A, Cantor KP, Reif JS, Lynch CF, Ward MH. 2004. A case-control study of pancreatic cancer and drinking water nitrate. Am J Epidemiol 159:693–701. Costas K, Knorr RS, Condon SK. 2002. A case-control study of childhood leukemia in Woburn, Massachusetts: The relationship between leukemia incidence and exposure to public drinking water. Sci Total Environ 300:23–35. Cothern CR, Lappenbusch WL, Michel J. 1986. Drinking-water contribution to natural background radiation. Health Phys 50:33–47. Cothern DR, Lappenbusch WL. 1983. Occurrence of uranium in drinking water in the U.S. Health Phys 45:89–99. Cragle DL, Shy CM, Struba RJ, Siff EJ. 1985. A case-control study of colon cancer and water chlorination in North Carolina. In: Jolley RL, Bull RJ, Davis WP, Katz S, Roberts MHJr, Jacobs VA. eds. Water Chlorination:Chemistry, Environmental Impact and Health Effects, Vol. 5. Chelsea, Mich.: Lewis Publishers, Inc., pp. 153–160. Cross FT, Harley NH, Hofmann W. 1985. Health effects and risks from radon222 in drinking water. Health Phys 48:649–670. Crump KS, Guess HA. 1982. Drinking water and cancer: Review of recent epidemiological findings and assessment of risks. Annu Rev Public Health 3:339–357. Cuello C, Correa P, Haenszel W, Gordillo G, Brown C, Archer M, Tannenbaum S. 1976. Gastric cancer in Colombia. I. Cancer risk and suspect environmental agents. J Natl Cancer Inst 57:1015–1020.

398

PART III: THE CAUSES OF CANCER

Cutler JJ, Parker GS, Rosen S, Prenney B, Healey R, Caldwell GG. 1986. Childhood leukemia in Woburn, Massachusetts. Publ Health Rep 101:201–205. Cuzick J, Sasieni P, Evans S. 1992. Ingested arsenic, keratoses, and bladder cancer. Am J Epidemiol 136:417–421. Dabeka RW, Mckenzie Ad, Lacroix GMA, Cleroux C, Bowe S, Graham RA, Conacher HBS, Verdier P. 2004. Survey of arsenic in total diet food composites and estimation of the dietary-intake of arsenic by Canadian adults and children. J AOAC Int 76:14–25. Daniel FB, Meier JR, DeAngelo AB. 1993. Advances in research on carcinogenic and genotoxic by–products of chlorine disinfection:chlorinated hydroxyfuranones and chlorinated acetic acids. Ann Ist Super Sanita 29:279–291. Degteva MO, Shagina NB, Tolstykh EI, Vorobiova MI, Napier BA, Anspaugh LR. 2002. Studies on the Techa river populations:dosimetry. Radiat Environ Biophys 41:41–44. DeKraay WH. 1978. Pesticides and lymphoma in Iowa. J Iowa Med Soc 1978:50–53. Delong S. 1979. Drinking water and liver cell cancer. Chin Med J 92:748– 756. DeRoos AJ, Ward MH, Lynch CF, Cantor KP. 2003. Nitrate in public water systems and the risk of colon and rectum cancers. Epidemiology 14:640–649. Doyle TJ, Zheng W, Cerhan JR, Hong C-P, Sellers TA, Kushi LH, Folsom AR. 1997. The association of drinking water source and chlorination byproducts with cancer incidence among postmenopausal women in Iowa: A prospective cohort study. Am J Public Health 87:1168–1176. Dulout FN, Grillo CA, Seoane AI, Maderna CR, Nilsson R, Vahter M, Darroudi F, Natarajan AT. 1996. Chromosomal aberrations in peripheral blood lymphocytes from native Andean women and children from Northwestern Argentina exposed to arsenic in drinking water. Mutat Res 370:151–158. Dunnick JK, Melnick RL. 1993. Assessment of the carcinogenic potential of chlorinated water:Experimental studies of chlorine, chloramine, and trihalomethanes. J Natl Cancer Inst 85:817–822. El–Bolkainy MN, Mokhtar NM, Ghoneim MA, Hussein MH. 1981. The impact of Schistosomiasis on the pathology of bladder carcinoma. Cancer 48:2643–2648. el Mawla NG, el Bolkainy MN, Khaled HM. 2001. Bladder cancer in Africa: Update. Semin Oncol 28:174–178. Elem B, Purohit R. 1983. Carcinoma of the urinary bladder in Zambia. A quantitative estimation of Schistosoma haematobium infection. Br J Urology 55:275–278. Engel RR, Receveur O. 1993. Re: “Arsenic ingestion and internal cancers: A review” [Letter]. Am J Epidemiol 138:896–897. Environmental Protection Agency. 1988. Special Report on Ingested Inorganic Arsenic: Skin Cancer, Nutritional Essentiality. Risk Assessment Forum. Washington, D.C.: Environmental Protection Agency. Environmental Protection Agency. 1998. National Primary Drinking Water Regulations; Disinfectants and Disinfection Byproducts. Federal Register 63, No. 241:69389–69476. Environmental Protection Agency. 2000. National Primary Drinking Water Regulations; Arsenic and clarifications to compliance and new source contaminants monitoring. Federal Register 65, No.121:38888–38983. Environmental Protection Agency. 2001. National Primary Drinking Water Regulations; Arsenic and clarifications to compliance and new source contaminants monitoring: 40 CFR Parts 9, 141, and 142. Federal Register 66, No. 14:6975–7066. Fagliano J, Berry M, Bove F, Burke T. 1990. Drinking water contamination and the incidence of leukemia: An ecologic study. Am J Public Health 80:1209–1212. Falk H, Herbert J, Crowley S, Ishak KG, Thomas LB, Popper H, Caldwell GG. 1981. Epidemiology of hepatic angiosarcoma in the United States: 1964–1974. Environ Health Perspect 41:107–113. Ferreccio C, Gonzalez C, Milosavjlevic V, Marshall G, Sancha AM, Smith AH. 2000. Lung cancer and arsenic concentrations in drinking water in Chile. Epidemiology 11:673–679. Finkelstein MM. 1994. Radium in drinking water and the risk of death from bone cancer among Ontario youths. Can Med Assoc J 151:565– 571. Finkelstein MM, Kreiger N. 1996. Radium in drinking water and risk of bone cancer in Ontario youths: A second study and combined analysis. Occup Environ Med 53:305–311. Flaten TP. 1992. Chlorination of drinking water and cancer incidence in Norway. Int J Epidemiol 21:6–15. Flaten TP, Bolviken B. 1991. Geographical associations between drinking water chemistry and the mortality and morbidity of cancer and some other diseases in Norway. Sci Total Environ 102:75–100.

Flora SJ, Kannan GM, Kumar P. 1999. Selenium effects on gallium arsenide induced biochemical and immunotoxicological changes in rats. Chem Biol Interact 122:1–13. Focazio MJ, Welch AH, Watkins SA, Helsel DR, Horn MA. 1999. A Retrospective Analysis on the Occurrence of Arsenic in Ground-Water Resources of the United States and Limitations in Drinking-Water-Supply Characteristics. Reston, Va.: U.S. Geological Survey Water-Resources Investigations Report 99–4279. Freedman DM, Cantor KP, Lee NL, Chen LS, Lei HH, Ruhl CE, Wang SS. 1997. Bladder cancer and drinking water:a population-based case-control study in Washington County, Maryland. United States. Cancer Causes Control 8:738–744. Freedman DM, Cantor KP, Ward MH, Helzlsouer KJ. 2000. A case-control study of nitrate in drinking water and non-Hodgkin’s lymphoma in Minnesota. Arch Environ Health 55:326–329. Froese KL, Sinclair MI, Hrudey SE. 2002. Trichloroacetic acid as a biomarker of exposure to disinfection by-products in drinking water: A human exposure trial in Adelaide, Australia. Environ Health Perspect 110: 679–687. Fuortes L, McNutt LA, Lynch C. 1990. Leukemia incidence and radioactivity in drinking water in 59 Iowa towns. Am J Public Health 80:1261–1262. Gilli G, Corrao G, Favilli S. 1984. Concentrations of nitrates in drinking water and incidence of gastric carcinomas: First descriptive study of the Piemonte region, Italy. Sci Total Environ 34:35–48. Giovannucci E. 2003. Diet, body weight, and colorectal cancer: A summary of the epidemiologic evidence. J Womens Health 12:173–182. Glattre E, Mravcova A, Lener J, Vobecky M, Egertova E, Mysliveckova M. 1995. Study of distribution and interaction of arsenic and selenium in rat thyroid. Biol Trace Elem Res 49:177–186. Glickman LT, Schofer FS, McKee LJ, Reif JS, Goldschmidt MH. 1989. Epidemiologic study of insecticide exposures, obesity, and risk of bladder cancer in household dogs. J Toxicol Environ Health 28:407–414. Glynn MK, Friedman CR, Gold BD, Khanna B, Hutwagner L, Iihoshi N, Revollo C, Quick R. 2002. Seroincidence of Helicobacter pylori infection in a cohort of rural Bolivian children: Acquisition and analysis of possible risk factors. Clin Infect Dis 35:1059–1065. Goldberg MS, Al-Homsi N, Goulet L, Riberdy H. 1995. Incidence of cancer among persons living near a municipal solid waste landfill site in Montreal, Québec. Arch Environ Health 50:416–424. Goldberg MS, Siemiatyck J, Dewar R, Desy M, Riberdy H. 1999. Risks of developing cancer relative to living near a municipal solid waste landfill site in Montreal, Quebec, Canada. Arch Environ Health 54:291–296. Gonsebatt ME, Vega L, Salazar AM, Montero R, Guzman P, Blas J, del Razo LM, Garcia-Vargas G, Albores A, Cebrian ME, Kelsh M, OstroskyWegman P. 1997. Cytogenetic effects in human exposure to arsenic. Mutat Res 386:219–228. Goodman KJ, Correa P, Tengana Aux HJ, Ramirez H, DeLany JP, Guerrero PO, Lopez QM, Collazos PT. 1996. Helicobacter pylori infection in the Colombian Andes:a population-based study of transmission pathways. Am J Epidemiol 144:290–299. Gottlieb MS, Carr JK. 1982. Case-control cancer mortality study and chlorination of drinking water in Louisiana. Environ Health Perspect 46:169– 177. Griffith J, Duncan RC, Riggan WB, Pellom AC. 1989. Cancer mortality in U.S. counties with hazardous waste sites and ground water pollution. Arch Environ Health 44:69–74. Grossman LA, Vaughn RG, Zeveloff SI. 2002. A Documentary Companion to A Civil Action. New York: Foundation Press. Gulis G, Czompolyova M, Cerhan JR. 2002. An ecologic study of nitrate in municipal drinking water and cancer incidence in Trnava District, Slovakia. Environ Res 88:182–187. Guo HR, Wang NS, Hu H, Monson RR. 2004. Cell type specificity of lung cancer associated with arsenic ingestion. Cancer Epidemiol Biomarkers Prev 13:638–643. Guo Z, Li Y, Xu Z, Ji F, Wang L, Chen K. 2002. A case-control study on risk factors of Helicobacter pylori infection in out-patients with stomach diseases. Zhonghua Yu Fang Yi Xue Za Zhi 36:187–190. Guse CE, Marbella AM, George V, Layde PM. 2002. Radium in Wisconsin drinking water: An analysis of osteosarcoma risk. Arch Environ Health 57:294–303. Hamadeh HK, Trouba KJ, Amin RP, Afshari CA, Germolec D. 2002. Coordination of altered DNA repair and damage pathways in arsenite-exposed keratinocytes. Toxicol Sci 69:306–316. Haque R, Guha Mazumder DN, Samanta S, Ghosh N, Kalman D, Smith MM, Mitra S, Samtra A, Lahiri S, Das S, De BK, Smith A. 2003. Arsenic in drinking water and skin lesions: Dose response data from West Bengal, India. Epidemiology 14:174–182. Harr J. 1995. A Civil Action. New York: Random House.

Water Contaminants Harrington JM, Craun GF, Meigs JW, Landrigan PJ, Flannery JT, Woodhull RS. 1978. An investigation of the use of asbestos cement pipe for public water supply and the incidence of gastrointestinal cancer in Connecticut, 1935–1973. Am J Epidemiol 107:96–103. Heath CW, Nadel MR, Zack MM, Chen ATL, Bender MA, Preston RJ. 1984. Cytogenetic findings in persons living near the Love Canal. JAMA 251:1437–1440. Hess CT, Michel J, Horton TR, Prichard HM, Coniglio WA. 1985. The occurrence of radioactivity in public water supplies in the United States. Health Phys 48:553–586. Hess CT, Weiffenbach CV, Norton SA. 1983. Environmental radon and cancer correlations in Maine. Health Phys 45:339–348. Hildesheim ME, Cantor KP, Lynch CF, Dosemeci M, Lubin J, Alavanja M, Craun G. 1998. Drinking water source and chlorination byproducts. II. Risk of colon and rectal cancers. Epidemiology 9:29–35. Hinwood AL, Jolley DJ, Sim MR. 1999. Cancer incidence and high environmental arsenic concentrations in rural populations: results of an ecological study. Int J Environ Health Res 9:131–141. Hirata M, Tanaka A, Hisanaga A, Ishinishi N. 1990. Effects of glutathione depletion on the acute nephrotoxic potential of arsenite and on arsenic metabolism in hamsters. Toxicol Appl Pharmacol 106:469–481. Hogan MD, Chi P-Y, Hoel DG. 1979. Association between chloroform levels in finished drinking water supplies and various site-specific cancer mortality rates. J Environ Pathol Toxicol 2:873–887. Holmbom B, Voss RH, Mortimer RD, Wong A. 1981. Isolation and identification of an Ames mutagenic compound present in draft chlorination effluents. Tappi 64:172–174. Holmbom B, Voss RH, Mortimer RD, Wong A. 1984. Fractionation, isolation, and characterization of Ames mutagenic compounds in kraft chlorination effluents. Environ Sci Technol 18:333–337. Hoover RN, Devesa S, Cantor K, Fraumeni JF Jr. 1991. Appendix F. Time trends for bone and joint cancers and osteosarcomas in the Surveillance, Epidemiology and End Results (SEER) Program, National Cancer Institute. In: Review of Fluoride: Benefits and Risks. Report of the Ad Hoc Subcommittee on Fluoride of the Committee to Coordinate Environmental Health and Related Programs. Washington, D.C.: Public Health Service, DHHS. Hoover RN, McKay FW, Fraumeni JF. 1976. Fluoridated drinking water and the occurrence of cancer. J Natl Cancer Inst 57:757–768. Hopenhayn-Rich C, Biggs ML, Fuchs A, Bergoglio R, Tello E, Nicolli H, Smith AH. 1996a. Bladder cancer mortality associated with arsenic in drinking water in Argentina. Epidemiology 7:117–124. Hopenhayn-Rich C, Biggs ML, Kalman DA, Moore LE, Smith AH. 1996b. Arsenic methylation patterns before and after changing from high to lower concentrations of arsenic in drinking water. Environ Health Perspect 104:1200–1207. Hopenhayn-Rich C, Biggs ML, Smith AH. 1998. Lung and kidney cancer mortality associated with arsenic in drinking water in Córdoba, Argentina. Int J Epidemiol 27:561–569. Hopenhayn-Rich C, Biggs ML, Smith AH, Kalman DA, Moore LE. 1996c. Methylation study of a population environmentally exposed to arsenic in drinking water. Environ Health Perspect 104:620–628. Hopenhayn-Rich C, Browning SR, Hertz-Picciotto I, Ferreccio C, Peralta C, Gibb H. 2000. Chronic arsenic exposure and risk of infant mortality in two areas of Chile. Environ Health Perspect 108:667–673. Hopenhayn-Rich C, Smith AH, Goeden HM. 1993. Human studies do not support the methylation threshold hypothesis for the toxicity of inorganic arsenic. Environ Res 60:161–177. Horiuchi T, Ohkusa T, Watanabe M, Kobayashi D, Miwa H, Eishi Y. 2001. Helicobacter pylori DNA in drinking water in Japan. Microbiol Immunol 45:515–519. Howe HL, Wolfgang PE, Burnett WS, Nasca PC, Youngblood L. 1989. Cancer incidence following exposure to drinking water with asbestos leachate. Publ Health Rep 104:251–256. Hrudey SE, Soskolne CL, Berkel J, Fincham S. 1990. Drinking water fluoridation and osteosarcoma. Can J Public Health 81:415–416. Hsieh LL, Chen HJ, Hsieh JT, Jee SH, Chen GS, Chen CJ. 1994. Arsenicrelated Bowen’s disease and paraquat-related skin cancerous lesions show no detectable ras and p53 gene alterations. Cancer Lett 86: 59–65. Hsueh YM, Cheng GS, Wu MM, Yu HS, Kuo TL, Chen CJ. 1995. Multiple risk factors associated with arsenic-induced skin cancer—Effects of chronic liver disease and malnutritional status. Br J Cancer 71:109– 114. Hsueh YM, Chiou HY, Huang YL, Wu WL, Huang CC, Yang MH, Lue LC, Chen GS, Chen CJ. 1997. Serum beta-carotene level, arsenic methylation capability, and incidence of skin cancer. Cancer Epidemiol Biomarkers Prev 6:589–596.

399

Hu Y, Su L, Snow ET. 1998. Arsenic toxicity is enzyme specific and its affects on ligation are not caused by the direct inhibition of DNA repair enzymes. Mutat Res 408:203–218. Huang H, Huang CF, Wu DR, Jinn CM, Jan KY. 1993. Glutathione as a cellular defense against arsenite toxicity in cultured Chinese hamster ovary cells. Toxicology 79:195–204. Hueper WC. 1960. Cancer hazards from natural and artificial water pollutants. In: Faber HA, Bryson LJ, eds. Proceedings, Physiological Aspects of Water Quality. Washington, D.C.: U.S. Public Health Service, pp 181–193. Hutchinson J. 1887. Arsenic cancer. Br Med J 2:1280–1281. IJsselmuiden CB, Gaydos C, Feighner B, Novakoski WL, Serwadda D, Caris LH, Vlahov D, Comstock GW. 1992. Cancer of the pancreas and drinking water: A population-based case-control study in Washington County, Maryland. Am J Epidemiol 136:836–842. Infante-Rivard C, Amre D, Sinnett D. 2002. GSTT1 and CYP2E1 polymorphisms and trihalomethanes in drinking water:effect on childhood leukemia. Environ Health Perspect 110:591–593. Infante-Rivard C, Olson E, Jacques L, Ayotte P. 2001. Drinking water contaminants and childhood leukemia. Epidemiology 12:13–19. International Agency for Research on Cancer. 1980. IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Volume 23: Some Metals and Metallic Compounds. Lyon: IARC. International Agency for Research on Cancer. 1982. Inorganic fluorides. In IARC Monographs on the Evaluation of Carcinogenic Risk of Chemicals to Humans, Volume 27: Some Aromatic Amines, Anthraquinones and Nitroso Compounds, and Inorganic Fluorides. Lyon: IARC, pp. 235–303. International Agency for Research on Cancer. 1991. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Volume 52: Chlorinated Drinking-Water; Chlorination By-Products; Some Other Halogenated Compounds; Cobalt and Cobalt Compounds. Lyon: IARC. International Agency for Research on Cancer. 1994a. Infection with Helicobacter pylori. In: IACR Monographs on the Evaluation of Carcinogenic Risks to Humans, Volume 61: Schistosomes, Liver Flukes, and Helicobacter pylori. Lyon: International Agency for Research on Cancer, pp. 177–240. International Agency for Research on Cancer. 1994b. Infection with Schistosomes. In: IACR Monographs on the Evaluation of Carcinogenic Risks to Humans, Volume 61: Schistosomes, Liver Flukes, and Helicobacter pylori. Lyon: IARC, pp. 45–119. International Agency for Research on Cancer. 2004. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Volume 84: Some Drinking-Water Disinfectants and Contaminants, including Arsenic. Lyon: IARC. Isacson P, Bean JA, Lynch C. 1983. Relationship of cancer incidence rates in Iowa municipalities to chlorination status of drinking water. In: Jolley RL, Brungs WA, Cotruvo JA, Cumming RB, Mattice JS, Jacobs VA, eds. Water Chlorination:Environmental Impact and Health Effects, Vol. 4. Ann Arbor, Mich.: Ann Arbor Science, pp. 1353–1363. Isacson P, Bean JA, Splinter R, Olson DB, Kohler J. 1985. Drinking water and cancer incidence in Iowa: III. Association of cancer with indices of contamination. Am J Epidemiol 121:856–869. Ishii A, Matsuoka H, Aji T, Ohta N, Arimoto S, Wataya Y, Hayatsu H. 1994. Parasite infection and cancer with special emphasis on Schistosoma japonicum infections (Trematoda). A review. Mutat Res 305:273–281. Jackson R, Grainge JW. 1975. Arsenic and cancer. Can Med Assoc J 113:396–401. Janerich DT, Burnett WS, Feck G, Hoff M, Nasca P, Polednak AP, Greenwald P, Vianna N. 1981. Cancer incidence in the Love Canal area. Science 212:1404–1407. Jarup L, Briggs D, de Hoogh C, Morris S, Hurt C, Lewin A, Maitland I, Richardson S, Wakefield J, Elliott P. 2002. Cancer risks in populations living near landfill sites in Great Britain. Br J Cancer 86:1732–1736. Jha AN, Noditi M, Nilsson R, Natarajan AT. 1992. Genotoxic effects of sodium arsenite on human cells. Mutat Res 284:215–221. Jo WK, Weisel CP, Lioy PJ. 1990. Routes of chloroform exposure and body burden from showering with chlorinated tap water. Risk Anal 10: 575–580. Johansson SL, Cohen SM. 1997. Epidemiology and etiology of bladder cancer. Semin Surg Oncol 13:291–298. Johnson CJ, Kross BC. 1990. Continuing importance of nitrate contamination of groundwater and wells in rural areas. Am J Indust Med 18:456. Johnson CH, Rice EW, Reasoner DJ. 1997. Inactivation of Helicobacter pylori by chlorination. Appl Environ Microbiol 63:4969–4970. Kanarek MS, Conforti PM, Jackson LA, Cooper RC, Murchio JC. 1980. Asbestos in drinking water and cancer incidence in the San Francisco Bay area. Am J Epidemiol 112:54–72. Karagas MR, Stukel TA, Morris JS, Tosteson TD, Weiss JE, Spencer SK, Greenberg ER. 2001. Skin cancer risk in relation to toenail arsenic

400

PART III: THE CAUSES OF CANCER

concentrations in a US population-based case-control study. Am J Epidemiol 153:559–565. Karagas MR, Tosteson TD, Blum J, Klaue B, Weiss JE, Stannard V, Spate V, Morris JS. 2000. Measurement of low levels of arsenic exposure: A comparison of water and toenail concentrations. Am J Epidemiol 152:84–90. Karita M, Teramukai S, Matsumoto S. 2003. Risk of Helicobacter pylori transmission from drinking well water is higher than that from infected intrafamilial members in Japan. Dig Dis Sci 48:1062–1067. Kim H, Haltmeier P, Klotz JB, Weisel CP. 1999. Evaluation of biomarkers of environmental exposures: Urinary haloacetic acids associated with ingestion of chlorinated drinking water. Environ Res 80:187–195. Kim H, Shim J, Lee S. 2002. Formation of disinfection by–products in chlorinated swimming pool water. Chemosphere 46:123–130. King WD, Marrett LD. 1996. Case-control study of bladder cancer and chlorination by-products in treated water. Cancer Causes Control 7: 596–604. King WD, Marrett LD, Woolcott CG. 2000. Case-control study of colon and rectal cancers and chlorination by-products in treated water. Cancer Epidemiol Biomark Prev 9:813–818. Kitchin KT. 2001. Recent advances in arsenic carcinogenesis:modes of action, animal model systems, and methylated arsenic metabolites. Toxicol Appl Pharmacol 172:249–261. Kitchin KT, Ahmad S. 2003. Oxidative stress as a possible mode of action for arsenic carcinogenesis. Toxicol Lett 137:3–13. Klein P, Graham DY, Gaillour A, Opekun A, Smith EO. 1991. Water source as risk factor for Helicobacter pylori infection in Peruvian children. Lancet 337:1503–1506. Kleinjans JC, Albering HJ, Marx A, van Maanen JM, van Agen B, ten Hoor F, Swaen GM, Mertens PL. 1991. Nitrate contamination of drinking water: evaluation of genotoxic risk in human populations. Environ Health Perspect 94:189–193. Knox EG. 1985. Fluoridation of Water and Cancer: A Review of the Epidemiological Evidence. Report of the Working Party on Fluoridation of Water and Cancer. London: Her Majesty’s Stationary Office. Koivusalo M, Hakulinen T, Vartiainen T, Pukkala E, Jaakkola JJK, Tuomisto J. 1998. Drinking water mutagenicity and urinary tract cancers: A population-based case-control study in Finland. Am J Epidemiol 148: 704–712. Koivusalo M, Jaakkola JJK, Vartiainen T, Hakulinen T, Karjalainen S, Pukkala E, Tuomisto J. 1994. Drinking water mutagenicity and gastrointestinal and urinary tract cancers: An ecological study in Finland. Am J Public Health 84:1223–1228. Koivusalo M, Pukkala E, Vartiainen T, Jaakkola JJK, Hakulinen T. 1997. Drinking water chlorination and cancer—a historical cohort study in Finland. Cancer Causes Control 8:192–200. Koivusalo M, Vartiainen T, Hakulinen T, Pukkala E, Jaakkola JJK. 1995. Drinking water mutagenicity and leukemia, lymphomas, and cancer of the liver, pancreas, and soft tissue. Arch Environ Health 50:269–276. Komulainen H, Kosma V-M, Vaittinen S-L, Vartianen T, Kaliste-Korhonen E, Lotjonen S, Tuominen RK, Tuomisto J. 1997. Carcinogenicity of the drinking water mutagen 3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)furanone in the rat. J Natl Cancer Inst 89:848–856. Kossenko MM, Hoffman DA, Thomas TL. 2000. Stochastic effects of environmental radiation exposure in populations living near the Mayak Industrial Association: Preliminary report on study of cancer morbidity. Health Phys 79:55–62. Kossenko MM, Preston DL, Krestinina LY, Degteva MO, Startsev NV, Thomas T, Vyushkova VP, Anspaugh LR, Napier BA, Kozheurov VP, Ron E, Akleyev AV. 2002. Studies on the extended Techa river cohort: Cancer risk estimation. Radiat Environ Biophys 41:45–48. Krasner SW, McGuire MJ, Jacangelo JG, Patania NL, Reagan KM, Aieta EM. 1989. The occurrence of disinfection by-products in U.S. drinking water. J Am Water Works Assoc 81 (Aug):41–53. Kronberg L, Christman RF. 1989. Chemistry of mutagenic by-products of water chlorination. Sci Total Environ 81–82:219–230. Kronberg L, Vartiainen T. 1988. Ames mutagenicity and concentration of the strong mutagen 3-chloro-4-(dichloromethyl)-5-hydroxy-2(5H)furanone and of its geometric isomer E-2-chloro-3-(dichloromethyl)4-oxo-butenoic acid in chlorine-treated tap waters. Mutat Res 206: 177–182. Kukkula M, Lofroth G. 1997. Chlorinated drinking water and pancreatic cancer: A population-based case-control study. Eur J Public Health 7:297–301. Kurttio P, Pukkala E, Kahelin H, Auvinen A, Pekkanen J. 1999. Arsenic concentrations in well water and risk of bladder and kidney cancer in Finland. Environ Health Perspect 107:705–710. Kuzma RJ, Kuzma CM, Buncher CR. 1977. Ohio drinking water source and cancer rates. Am J Public Health 67:725–729.

Lagakos SW, Wessen BJ, Zelen M. 1986. An analysis of contaminated well water and health effects in Woburn, Massachusetts. J Am Stat Assoc 81:583–596. Lampi P, Hakulinen T, Luostarinen T, Pukkala E, Teppo L. 1992. Cancer incidence following chlorophenol exposure in a community in southern Finland. Arch Environ Health 47:167–175. Langer AM, Maggiore CM, Nicholson WJ, Rohl AN, Rubin IB, Selikoff IJ. 1979. The contamination of Lake Superior with amphibole gangue minerals. Ann N Y Acad Sci 330:549–572. Law G, Parslow R, McKinney P, Cartwright R. 1999. Non-Hodgkin’s lymphoma and nitrate in drinking water:a study in Yorkshire, United Kingdom. J Epidemiol Community Health 53:383–384. Lawrence CE, Taylor PR, Trock BJ, Reilly AA. 1984. Trihalomethanes in drinking water and human colorectal cancer. J Natl Cancer Inst 72: 563–568. Le XC, Lu X, Ma M, Cullen WR, Aposhian HV, Zheng B. 2000. Speciation of key arsenic metabolic intermediates in human urine. Anal Chem 72:5172–5177. Lee–Chen SF, Yu CT, Wu DR, Jan KY. 1994. Differential effects of luminol, nickel, and arsenite on the rejoining of ultraviolet light and alkylation–induced DNA breaks. Environ Mol Mutagen 23:116–120. Lee TC, Tanaka N, Lamb PW, Gilmer TM, Barrett JC. 1988. Induction of gene amplification by arsenic. Science 241:79–81. Levy BS, Sigurdson E, Mandel J, Laudon E, Pearson J. 1976. Investigating possible effects of asbestos in city water: Surveillance of gastrointestinal cancer incidence in Duluth, Minnesota. Am J Epidemiol 103:362–368. Lewis DR, Southwick JW, Ouellet–Hellstrom R, Rench J, Calderon RL. 1999. Drinking water arsenic in Utah: A cohort mortality study. Environ Health Perspect 107:359–365. Li J–H, Rossman TG. 1989. Inhibition of DNA ligase activity by arsenite: A possible mechanism of its comutagenesis. Mol Toxicol 2:1–9. Li YN, Gulati S, Baker PJ, Brody LC, Banerjee R, Kruger WD. 1996. Cloning, mapping and RNA analysis of the human methionine synthase gene. Hum Mol Genet 5:1851–1858. Lijinsky W. 1986. The significance of N–nitroso compounds as environmental carcinogens. J Environ Sci Health C4:1–45. Lilienfeld DE, Gallo MA. 1989. 2,4-D, 2,4,5-T, and 2,3,7,8-TCDD: An overview. Epidemiol Rev 28–58. Lin SM, Yang MH. 1988. Arsenic, selenium, and zinc in patients with Blackfoot disease. Biol Trace Elem Res 15:213–221. Lindkvist P, Enquselassie F, Asrat D, Nilsson I, Muhe L, Giesecke J. 1999. Helicobacter pylori infection in Ethiopian children: A cohort study. Scand J Infect Dis 31:475–480. Lindstrom AB, Pleil JD, Berkoff DC. 1997. Alveolar breath sampling and analysis to assess trihalomethane exposures during competitive swimming training. Environ Health Perspect 105:636–642. Loper JC, Lang DR, Schoeny RS, Richmond BB, Gallagher PM, Smith CC. 1978. Residue organic mixtures from drinking water show in vitro mutagenic and transforming activity. J Toxicol Environ Health 4:919–938. Lu Y, Redlinger TE, Avitia R, Galindo A, Goodman K. 2002. Isolation and genotyping of Helicobacter pylori from untreated municipal wastewater. Appl Environ Microbiol 68:1436–1439. Lubin JH, Pottern LM, Stone BJ, Fraumeni JFJ. 2000. Respiratory cancer in a cohort of copper smelter workers: results from more than 50 years of follow-up. Am J Epidemiol 151:554–565. Lyman GH, Lyman CG, Johnson W. 1985. Association of leukemia with radium groundwater contamination. JAMA 254:621–626. MacMahon B, Prentice RL, Rogan WJ, Swan SH, Robins JM, Whittemore AS. 1986. Comments and rejoinder on Lagakos, Wessen, and Zelen article on contaminated well water and health effects in Woburn, Massachusetts. J Am Stat Assoc 81:597–614. Mahoney MC, Nasca PC, Burnett WS, Melius JM. 1991. Bone cancer incidence rates in New York State: Time trends and fluoridated drinking water. Am J Public Health 81:475–479. Mallin K. 1990. Investigation of a bladder cancer cluster in Northwestern Illinois. Am J Epidemiol 132 (Suppl 1):S96–S106. Marafante E, Vahter M. 1984. The effect of methyltransferase inhibition on the metabolism of [74As] arsenite in mice and rabbits. Chem Biol Interact 50:49–57. Marafante E, Vahter M, Envall J. 1985. The role of the methylation in the detoxification of arsenate in the rabbit. Chem Biol Interact 56:225–238. Marienfeld CJ, Collins M, Wright H, Reddy R, Shoop G, Rust P. 1986. Cancer mortality and the method of chlorination of public drinking water: St. Louis City and St. Louis County, Missouri. J Environ Pathol Toxicol 7:151–157. Marnell LL, Garcia-Vargas GG, Chowdhury UK, Zakharyan RA, Walsh B, Avram MD, Kopplin MJ, Cebrian ME, Silbergeld EK, Aposhian HV. 2003. Polymorphisms in the human monomethylarsonic acid (MMA V)

Water Contaminants reductase/hGSTO1 gene and changes in urinary arsenic profiles. Chem Res Toxicol 16:1507–1513. Mason TJ, McKay FW, Miller RW. 1974. Asbestos-like fibers in Duluth water supply. JAMA 228:1019–1020. Mays CW, Rowland RE, Stehney AF. 1985. Cancer risk from the lifetime intake of Ra and U istopes. Health Phys 48:635–648. Mazari-Hiriart M, Lopez-Vidal Y, Castillo-Rojas G, Ponce de Leon S, Cravioto A. 2001. Helicobacter pylori and other enteric bacteria in freshwater environments in Mexico City. Arch Med Res 32:458–467. Mazumder DNG, Haque R, Ghosh N, De BK, Santra A, Chakraborti D, Smith AH. 2000. Arsenic in drinking water and the prevalence of respiratory effects in West Bengal, India. Int J Epidemiol 29:1047–1052. Mazumder DNG, Haque R, Gosh N, De BK, Santra, Chakraborty D, Smith AH. 1998. Arsenic levels in drinking water and the prevalence of skin lesions in West Bengal, India. Int J Epidemiol 27:871–877. McGeehin MA, Reif JS, Becher J, Mangione EJ. 1993. A case-control study of bladder cancer and water disinfection methods in Colorado. Am J Epidemiol 138:492–501. McGuire SM, Vanable ED, McGuire MH, Buckwalter JA, Douglass CW. 1991. Is there a link between fluoridated water and osteosarcoma? J Am Dent Assoc 122:38–45. Meigs JW, Walter SD, Heston JF, Millette JR, Craun GF, Woodhull RS, Flannery JT. 1980. Asbestos cement pipe and cancer in Connecticut 1955–1974. J Environ Health 42:187–191. Miles AM, Singer PC, Ashley DL, Lynberg MC, Mendola P, Langlois PH, Nuckols JR. 2002. Comparison of trihalomethanes in tap water and blood. Environ Sci Technol 36:1692–1698. Millette JR, Clark PJ, Stober J, Rosenthal M. 1983a. Asbestos in water supplies of the United States. Environ Health Perspect 53:45–48. Millette JR, Craun GF, Stober JA, Kraemer DF, Tousignant HG, Hildago E, Duboise L, Benedict J. 1983b. Epidemiology study of the use of asbestoscement pipe for the distribution of drinking water in Escambia County, Florida. Environ Health Perspect 53:91–98. Milton AH, Hasan Z, Rahman A, Rahman M. 2003. Non-cancer effects of chronic arsenicosis in Bangladesh: Preliminary results. J Environ Sci Health A 38:301–305. Mirvish SS. 1983. The etiology of gastric cancer: Intragastric nitrosamide formation and other theories. J Natl Cancer Inst 71:629–647. Mirvish SS, Grandjean AC, Moller H, Fike S, Maynard T, Jones L, Rosinsky S, Nie G. 1992. N-Nitrosoproline excretion by rural Nebraskans drinking water of varied nitrate content. Cancer Epidemiol Biomark Prev 1:455–461. Moller H, Landt J, Pedersen E, Jensen P, Autrup H, Jensen OM. 1989. Endogenous nitrosation in relation to nitrate exposure from drinking water and diet in a Danish rural population. Cancer Res 49:3117–3121. Moore LE, Huang WY, Chung J, Hayes RB. 2003a. Epidemiologic considerations to assess altered DNA methylation from environmental exposures in cancer. Ann N Y Acad Sci 983:181–196. Moore LE, Smith AH, Eng C, DeVries S, Kalman D, Bhargava V, Chew K, Ferreccio C, Rey OA, Hopenhayn C, Biggs ML, Bates MN, Waldman FM. 2003b. P53 alterations in bladder tumors from arsenic and tobacco exposed patients. Carcinogenesis 24:1785–1791. Moore LE, Smith AH, Eng C, Kalman D, DeVries S, Bhargava V, Chew K, Moore ID, Ferreccio C, Rey OA, Waldman FM. 2002. Arsenic-related chromosomal alterations in bladder cancer. J Natl Cancer Inst 94:1688–1696. Moore LE, Smith AH, Hopenhayn-Rich C, Biggs ML, Kalman DA, Smith MT. 1997a. Decrease in bladder cell micronucleus prevalence after intervention to lower the concentration of arsenic in drinking water. Cancer Epidemiol Biomarkers Prev 6:1051–1056. Moore LE, Smith AH, Hopenhayn-Rich C, Biggs ML, Kalman DA, Smith MT. 1997b. Micronuclei in exfoliated bladder cells among individuals chronically exposed to arsenic in drinking water. Cancer Epidemiol Biomarkers Prev 6:31–36. Moore LE, Warner ML, Smith AH, Kalman D, Smith MT. 1996. Use of the fluorescent micronucleus assay to detect the genotoxic effects of radiation and arsenic exposure in exfoliated human epithelial cells. Environ Mol Mutagen 27:176–184. Morales-Suarez-Varela MM, Llopis-Gonzalez A, Tejerizo-Perez L. 1995. Impact of nitrates in drinking water on cancer mortality in Valencia, Spain. Eur J Epidemiol 11:15–21. Morales KH, Ryan L, Kuo TL, Wu MM, Chen CJ. 2000. Risk of internal cancers from arsenic in drinking water. Environ Health Perspect 108:655–661. Moss ME, Kanarek MS, Anderson HA, Hanrahan LP, Remington PL. 1995. Osteosarcoma, seasonality, and environmental factors in Wisconsin, 1979–1989. Arch Environ Health 50:235–241. Mostafa MH, Helmi S, Badawi AF, Tricker AR, Spiegelhalder B, Preussmann R. 1994. Nitrate, nitrite and volatile N-nitroso compounds in the urine of

401

Schistosoma haematobium and Schistosoma mansoni infected patients. Carcinogenesis 15:619–625. Mueller BA, Newton K, Holly EA, Preston-Martin S. 2001. Residential water source and the risk of childhood brain tumors. Environ Health Perspect 109:551–556. Mueller DK, Hamilton PA, Helsel DR, Hitt KJ, Ruddy BC. 1995. Nutrients in Ground Water and Surface Water of the United States—An Analysis of Data through 1992. Water-Resources Investigations Report 95–4031. Denver: U.S.Geological Survey. Mushak P, Crocetti AF. 1995. Risk and revisionism in arsenic cancer risk assessment. Environ Health Perspect 103:684–689. Mustacchi P, Shimkin MB. 1958. Cancer of the bladder and infestation with Schistosoma hematobium. J Natl Cancer Inst 20:825–842. Nagy G, Nemeth A, Bodor F, Fiesor E. 1980. Cases of bladder cancer caused by chronic arsenic poisoning. Orv Hetil 121:1009–1011. Najem GR, Louria DB, Lavenhar MA, Feuerman M. 1985. Clusters of cancer mortality in New Jersey municipalities; with special reference to chemical toxic waste disposal sites and per capita income. Int J Epidemiol 14:528–537. National Academy of Sciences Committee on Nitrite and Alternative Curing Agents in Foods. 1981. The Health Effects of Nitrate, Nitrite, and N-Nitroso Compounds. Washington, D.C.: National Academy of Sciences. National Cancer Institute. 1992. Cancer Statistics Review: 1973–1989, NIH Pub. No. 92–2789. Bethesda, Md.: National Cancer Institute. National Center for Environmental Health, Centers for Disease Control. A Survey of the Quality of Water Drawn from Domestic Wells in Nine Midwest States. 1998. Atlanta, Ga.: CDC. National Research Council. 1977. Drinking Water and Health. Volume 1. Washington, D.C.: National Academy of Sciences. National Research Council. 1990. Health Effects of Exposure to Low levels of Ionizing Radiation. BEIR V. Washington, D.C.: National Academy Press. National Research Council. 1995. Nitrate and Nitrite in Drinking Water. Washington, D.C.: National Academy Press. National Research Council. 1999. Health Effects of Exposure to Radon. BEIR VI. Washington, D.C.: National Academy Press. National Research Council: Committee on Environmental Epidemiology. 1991. Environmental Epidemiology. Volume 1: Public Health and Hazardous Wastes. Washington, D.C.: National Academy Press. National Research Council: Committee on Environmental Epidemiology. 1997. Environmental Epidemiology, Volume 2: Use of the Gray Literature and Other Data in Environmental Epidemiology. Washington, D.C.: National Academy Press. National Research Council: Committee on Risk Assessment of Exposure to Radon in Drinking Water. 1999. Risk Assessment of Radon in Drinking Water. Washington, D.C.: National Academy Press. National Research Council: Subcommittee on Arsenic in Drinking Water. 1999. Arsenic in Drinking Water. Washington, D.C.: National Academy Press. National Research Council: Subcommittee to Update the 1999 Arsenic in Drinking Water Report. 2001. Arsenic in Drinking Water: 2001 Update. Washington, D.C.: National Academy Press. Nazaroff WW, Doyle SM, Nero AV, Sextro RG. 1987. Potable water as a source of airborne 222Rn in U.S. dwellings: A review and assessment. Health Phys 52:281–295. Neubauer O. 1947. Arsenical cancer: A review. Br J Cancer 1:192–251. Nieuwenhuijsen MJ, Toledano MB, Elliott P. 2000. Uptake of chlorination disinfection by-products; a review and a discussion of its implications for exposure assessment in epidemiological studies. J Expo Anal Environ Epidemiol 10:586–599. Nolan BT, Stoner JD. 2000. Nutrients in groundwaters of the conterminous United States, 1992–1995. Environ Sci Technol 34:1156–1165. Nurgalieva ZZ, Malaty HM, Graham DY, Almuchambetova R, Machmudova A, Kapsultanova D, Osato MS, Hollinger FB, Zhangabylov A. 2002. Helicobacter pylori infection in Kazakhstan: effect of water source and household hygiene. Am J Trop Med Hyg 67:201–206. Oya-Ohta Y, Kaise T, Ochi T. 1996. Induction of chromosomal aberrations in cultured human fibroblasts by inorganic and organic arsenic compounds and the different roles of glutathione in such induction. Mutat Res 357:123–129. Ozonoff D, Aschengrau A, Coogan P. 1994. Cancer in the vicinity of a Department of Defense superfund site in Massachusetts. Toxicol Ind Health 10:119–141. Page T, Harris RH, Epstein SS. 1976. Drinking water and cancer mortality in Louisiana. Science 193:55–57. Parkin DM, Vizcaino AP, Skinner ME, Ndhlovu A. 1994. Cancer patterns and risk factors in the African population of southwestern Zimbabwe, 1963–1977. Cancer Epidemiol Biomarkers Prev 3:537–547.

402

PART III: THE CAUSES OF CANCER

Paulu C, Aschengrau A, Ozonoff D. 1999. Tetrachloroethylene-contaminated drinking water in Massachusetts and the risk of colon–rectum, lung, and other cancers. Environ Health Perspect 107:265–271. Pavittranon S, Sripaoraya K, Ramchuen S, Kachamatch S, Puttaprug W, Pamornpusirikul N, Thaicharuen S, Rujiwanitchkul S, Walueng W. 2003. Laboratory case identification of arsenic in Ronpibul village, Thailand. 2000–2002. J Environ Sci Health A 38:213–221. Petersen NJ, Samuels LD, Lucas HF, Abrahams SP. 1966. An epidemiologic approach to low-level radium 226 exposure. Publ Health Rep 81: 805–814. Phillips AM, Silbergeld EK. 1985. Health effects studies of exposure from hazardous waste sites—where are we today? Am J Ind Med 8:1–7. Plewa MJ, Wagner ED, Jazwierska P, Richardson SD, Chen PH, McKague AB. 2004. Halonitromethane drinking water disinfection byproducts:chemical characterization and mammalian cell cytotoxicity and genotoxicity. Environ Sci Technol 38:62–68. Polissar L, Severson RK, Boatman ES. 1984. A case-control study of asbestos in drinking water and cancer risk. Am J Epidemiol 119:456–471. Polissar L, Severson RK, Boatman ES, Thomas DB. 1982. Cancer incidence in relation to asbestos in drinking water in the Puget Sound region. Am J Epidemiol 116:314–328. Popper H, Thomas LB, Telles NC, Falk H, Selikoff IJ. 1978. Development of hepatic angiosarcoma in man induced by vinyl chloride, thorotrast, and arsenic. Am J Pathol 92:349–369. Prystowsky SD, Elfenbein GJ, Lamberg SI. 1978. Nasopharyngeal carcinoma associated with long–term arsenic ingestion. Arch Dermatol 114:602–603. Rademacher JJ, Young TB, Kanarek MS. 1992. Gastric cancer mortality and nitrate levels in Wisconsin drinking water. Arch Environ Health 47:292–297. Rahman M, Tondel M, Chowdhury IA, Axelson O. 1999. Relations between exposure to arsenic, skin lesions, and glucosuria. Occup Environ Med 56:277–281. Rahman MM, Chowdhury UK, Mukherjee SC, Mondal BK, Paul K, Lodh D, Biswas BK, Chanda CR, Basu GK, Saha KC, Roy S, Das R, Palit SK, Quamruzzaman Q, Chakraborti D. 2001. Chronic arsenic toxicity in Bangladesh and West Bengal, India—A review and commentary. J Toxicol Clin Toxicol 39:683–700. Ranmuthugala G, Pilotto L, Smith W, Vimalasiri T, Dear K, Douglas R. 2003. Chlorinated drinking water and micronuclei in urinary bladder epithelial cells. Epidemiology 14:617–622. Reid ME, Duffield–Lillico AJ, Garland L, Turnbull BW, Clark LC, Marshall JR. 2002. Selenium supplementation and lung cancer incidence:an update of the nutritional prevention of cancer trial. Cancer Epidemiol Biomarkers Prev 11:1285–1291. Reymann F, Moller R, Nielsen A. 1978. Relationship between arsenic intake and internal malignant neoplasms. Arch Dermatol 114:378–381. Richardson SD, Simmons JE, Rice G. 2002. Disinfection byproducts: The next generation. Environ Sci Technol 36:198A–205A. Roat JW, Wald A, Mendelow H, Pataki KI. 1982. Hepatic angiosarcoma associated with short-term arsenic ingestion. Am J Med 73:933–936. Robertson DAF, Low-Beer TS. 1983. Long term consequences of arsenical treatment for multiple sclerosis. Br Med J 286:605–606. Robson AO, Jelliffe AM. 1963. Medicinal arsenic poisoning and lung cancer. Br Med J 5351:207–209. Rook JJ. 1974. Formation of haloforms during chlorination of natural waters. J Soc Water Treat Exam 23:234–243. Rosas I, Belmont R, Armienta A, Baez A. 1999. Arsenic concentrations in water, soil, milk and forage in Comarca Lagunera, Mexico. Water Air Soil Pollut 112:133–149. Rosset M. 1958. Arsenical dermatoses associated with carcinomas of the internal organs. Can Med Assoc J 78:416–419. Roth F. 1957. The sequelae of chronic arsenic poisoning in Moselle vintners. Ger Med Mon 2:172–175. Rowland IR, Granli T, Bockman OC, Key PE, Massey RC. 1991. Endogenous N-nitrosation in man assessed by measurement of apparent total N-nitroso compounds in faeces. Carcinogenesis 12:1395–1401. Sadler TD, Rom WN, Lyon JL, Mason JO. 1984. The use of asbestos-cement pipe for public water supply and the incidence of cancer in selected communities in Utah. J Community Health 9:285–293. Salg J. 1977. Cancer Mortality Rates and Drinking Water in 346 Counties of the Ohio River Valley Basin, Ph.D. Dissertation. Chapel Hill: University of North Carolina. Sandor J, Kiss I, Farkas O, Ember I. 2001. Association between gastric cancer mortality and nitrate content of drinking water:ecological study on small area inequalities. Eur J Epidemiol 17:443–447. Schrauzer GN. 1992. Selenium. Mechanistic aspects of anticarcinogenic action. Biol Trace Elem Res 33:51–62.

Schroeder M, Mass MJ. 1997. CpG methylation inactivates the transcriptional activity of the promoter of the human p53 tumor suppressor gene. Biochem Biophys Res Commun 235:403–406. Seshan K. 1983. How are the physical and chemical properties of chrysotile asbestos altered by a 10-year residence in water and up to 5 days in simulated stomach acid? Environ Health Perspect 53:143–148. Shi Y, Li J, Li M, Wang L, Guo W, Qin D, Geng C, Forman D, Newell DG, Peto R, Blot WJ. 1997. Association of Helicobacter pylori infection with precancerous lesions and stomach cancer:a case–control study in Yangzhong County. Chin Med Sci J 12:175–180. Shimizu M, Hochadel JF, Fulmer BA, Waalkes MP. 1998. Effect of glutathione depletion and metallothionein gene expression on arsenic-induced cytotoxicity and c-myc expression in vitro. Toxicol Sci 45:204–211. Shrestha RR, Shrestha MP, Upadhyay NP, Pradhan R, Khadka R, Maskey A, Maharjan M, Tuladhar S, Dahal BM, Shrestha K. 2003. Groundwater arsenic contamination, its health impact and mitigation program in Nepal. J Environ Sci Health A 38:185–200. Shutov VN, Travnikov IG, Bruk GY, Golikov VY, Balonov MI, Howard BJ, Brown J, Strand P, Kravtsova EM, Gavrilov AP, Kravtsova OS, Mubasarov AA. 2002. Current contamination by 137Cs and 90Sr of the inhabited part of the Techa River basin in the Urals. J Environ Radioact 61:91–109. Sigurdson EE, Levy BS, Mandel J, McHugh R, Michienzi LJ, Jagger H, Pearson J. 1981. Cancer morbidity investigations: Lessons from the Duluth study of possible effects of asbestos in drinking water. Environ Res 25:50–61. Singer PC, Chang SD. 1989. Correlations between trihalomethanes and total organic halides formed during water treatment. J Am Water Works Assoc 81;8:61–65. Smedley PL, Kinniburgh DG. 2002. A review of the source, behaviour and distribution of arsenic in natural waters. Appl Geochem 17:517–568. Smith AH, Goycolea M, Haque R, Biggs ML. 1998. Marked increase in bladder and lung cancer mortality in a region of Northern Chile due to arsenic in drinking water. Am J Epidemiol 147:660–669. Smith AH, Hopenhayn-Rich C, Bates MN, Goeden HM, Hertz-Picciotto I, Duggan HM, Wood R, Kosnett MJ, Smith MT. 1992. Cancer risks from arsenic in drinking water. Environ Health Perspect 97:259–267. Smith AH, Lingas EO, Rahman M. 2000. Contamination of drinking-water by arsenic in Bangladesh: A public health emergency. Bull WHO 78:1093–1103. Soignet SL, Frankel SR, Douer D, Tallman MS, Kantarjian H, Calleja E, Stone RM, Kalaycio M, Scheinberg DA, Steinherz P, Sievers EL, Coutre S, Dahlberg S, Ellison R, Warrell RP. 2001. United States multicenter study of arsenic trioxide in relapsed acute promyelocytic leukemia. J Clin Oncol 19:3852–3860. Solley WB, Pierce RR, Perlman HA. 1993. Estimated use of water in the United States in 1990. USGS Circular 1081. Reston, Va.: U.S. Geological Survey. Sommers SC, McManus RG. 1953. Multiple arsenical cancers of the skin and internal organs. Cancer 6:347–359. Steindorf K, Schlehofer B, Becher H, Hornig G, Wahrendorf J. 1994. Nitrate in drinking water. A case-control study on primary brain tumours with an embedded drinking water survey in Germany. Int J Epidemiol 23: 451–457. Steinmaus C, Moore L, Hopenhayn-Rich C, Biggs ML, Smith AH. 2000. Arsenic in drinking water and bladder cancer. Cancer Invest 18:174–182. Steinmaus C, Yuan Y, Bates MN, Smith AH. 2003. Case-control study of bladder cancer and drinking water arsenic in the western United States. Am J Epidemiol 158:1193–1201. Stevens AA, Moore LA, Miltner RJ. 1989. Formation and control of nontrihalomethane disinfection by-products. J Am Water Works Assoc 81;8:54–60. Stevens AA, Moore LA, Slocum CJ, Smith BL, Seeger DR, Ireland JC. 1990. By-products of chlorination at ten operating utilities. In: Jolley RL, Condie LW, Johnson JD, Katz S, Minear RA, Mattice JS, Jacobs VA. eds. Water Chlorination: Chemistry, Environmental Impact, and Health Effects, Vol. 6. Chelsea, Mich.: Lewis Publishers, pp. 579–604. Stocks P. 1973. Mortality from cancer and cardiovascular diseases in the county boroughs of England and Wales classified according to the sources and hardness of their water supplies, 1958–1967. J Hyg Lond 71:237–252. Styblo M, Delnomdedieu M, Thomas DJ. 1996. Mono- and dimethylation of arsenic in rat liver cytosol in vitro. Chem Biol Interact 99:147–164. Styblo M, Yamauchi H, Thomas DJ. 1995. Comparative in vitro methylation of trivalent and pentavalent arsenicals. Toxicol Appl Pharmacol 135:172–178. Szekeres G, De Giacomoni P. 1994. Ki-67 and p53 expression in cutaneous Bowen’s disease: An immunohistochemical study of fixed-embedded tissue sections. Acta Derm Venereol 74:95–97.

Water Contaminants Taketani S, Kohno H, Tokunaga R, Ishii T, Bannai S. 1991. Selenium antagonizes the induction of human heme oxygenase by arsenite and cadmium ions. Biochem Int 23:625–632. Tallman MS, Nabhan C, Feusner JH, Rowe JM. 2002. Acute promyelocytic leukemia: Evolving therapeutic strategies. Blood 99:759–767. Tao SH, Bolger PM. 1998. Dietary arsenic intakes in the United States:FDA total diet study, September 1991–December 1996. Food Addit Contam 16:465–472. Tao X, Zhu H, Matanoski GM. 1999. Mutagenic drinking water and risk of male esophageal cancer: A population-based case-control study. Am J Epidemiol 150:443–452. Teh BH, Lin JT, Pan WH, Lin SH, Wang LY, Lee TK, Chen CJ. 1994. Seroprevalence and associated risk factors of Helicobacter pylori infection in Taiwan. Anticancer Res 14:1389–1392. Thomas JE, Bassett MT, Sigola LB, Taylor P. 1990. Relationship between bladder cancer incidence, Schistosoma haematobium infection, and geographical region in Zimbabwe. Trans R Soc Trop Med Hyg 84: 551–553. Thompson DJ. 1993. A chemical hypothesis for arsenic methylation in mammals. Chem Biol Interact 88:89–14. Tian D, Ma H, Feng Z, Xia Y, Le XC, Ni Z, Allen J, Collins B, Schreinemachers D, Mumford JL. 2001. Analyses of micronuclei in exfoliated epithelial cells from individuals chronically exposed to arsenic via drinking water in inner Mongolia, China. J Toxicol Environ Health A 64:473–484. Tricker AR. 1997. N-nitroso compounds and man: Sources of exposure, endogenous formation and occurrence in body fluids. Eur J Cancer Prev 6:226–268. Tricker AR, Mostafa MH, Spiegelhalder B, Preussmann R. 1991. Urinary nitrate, nitrite and N-nitroso compounds in bladder cancer patients with schistosomiasis (bilharzia). IARC Sci Publ 105:178–181. Tsai SM, Wang TN, Ko YC. 1999. Mortality for certain diseases in areas with high levels of arsenic in drinking water. Arch Environ Health 54:186–193. Tseng WP, Chu HM, How SW. 1968. Prevalence of skin cancer in an endemic area of chronic arsenicism in Taiwan. J Natl Cancer Inst 40:453–463. Tsuda T, Babazono A, Yamamoto E, Kurumatani N, Mino Y, Ogawa t, Kishi Y, Aoyama H. 1995. Ingested arsenic and internal cancer:A historical cohort study followed for 33 years. Am J Epidemiol 141:198–209. Tucker RK. 1981. Groundwater Quality in New Jersey: An Investigation of Toxic Contaminants. Trenton, N.J.: New Jersey Department of Environmental Protection. U.S. Environmental Protection Agency. 2000. National Primary Drinking Water Regulations; Radionuclides; Final Rule. Federal Register 65, No. 236:76707–76753. U.S. Environmental Protection Agency. 1988. Office of Pesticide Programs. Pesticides in ground water data base: 1988 interim report. Washington, D.C.: U.S. Environmental Protection Agency. U.S. Public Health Service. Committee to Coordinate Environmental Health and Related Programs. 1991. Review of Fluoride: Benefits and Risks. Washington, D.C.: Department of Health and Human Services. Vahter M. 1994. Species differences in the metabolism of arsenic compounds. Appl Organomet Chem 8:175–182. Vahter M. 1999. Methylation of inorganic arsenic in different mammalian species and population groups. Sci Prog 82 (Pt 1):69–88. Vahter M. 2000. Genetic polymorphism in the biotransformation of inorganic arsenic and its role in toxicity. Toxicol Lett 112–113:209–217. Vahter M. 2002. Mechanisms of arsenic biotransformation. Toxicology 181:211–217. Vahter M, Concha G. 2001. Role of metabolism in arsenic toxicity. Pharmacol Toxicol 89:1–5. Vahter M, Marafante E. 1987. Effects of low dietary intake of methionine, choline or proteins on the biotransformation of arsenite in the rabbit. Toxicol Lett 37:41–46. Van Leeuwen JA, Waltner-Toews D, Abernathy T, Smit B, Shoukri M. 1999. Associations between stomach cancer incidence and drinking water contamination with atrazine and nitrate in Ontario (Canada) agroecosystems, 1987–1991. Int J Epidemiol 28:836–840. van Loon AJ, Botterweck AA, Goldbohm RA, Brants HA, van Klaveren JD, van den Brandt PA. 1998. Intake of nitrate and nitrite and the risk of gastric cancer:a prospective cohort study. Br J Cancer 78:129–135. van Maanen JMS, Welle IJ, Hageman G, Dallinga JW, Mertens PLJM, Kleinjans JCS. 1996. Nitrate contamination of drinking water: relationship with HPRT variant frequency in lymphocyte DNA and urinary excretion of N-nitrosamines. Environ Health Perspect 104:522–528. Vega L, Gonsebatt ME, Ostrosky-Wegman P. 1995. Aneugenic effect of sodium arsenite on human lymphocytes in vitro: An individual susceptibility effect detected. Mutat Res 334:365–373. Vianna NJ, Polan AK. 1984. Incidence of low birth weight among Love Canal residents. Science 226:1217–1219.

403

Villanueva CM, Cantor KP, Cordier S, Jaakkola JJK, King WD, Lynch CF, Porru S, Kogevinas M. 2004. Disinfection byproducts and bladder cancer: A pooled analysis. Epidemiology 15:357–367. Vitousek PM, Aber JD, Howarth RW, Likens GE, Matson PA, Schindler DW, Schlesinger WH, Tilman DG. 1997. Human alteration of the global nirogen cycle: Sources and consequences. Ecol Appl 7:737–750. Vizcaino AP, Parkin DM, Boffetta P, Skinner MEG. 1994. Bladder cancer: epidemiology and risk factors in Bulawayo, Zimbabwe. Cancer Causes Control 5:517–522. Vrijheid M. 2000. Health effects of residence near hazardous waste landfill sites: A review of epidemiologic literature. Environ Health Perspect 108 (Suppl 1):101–112. Wang GQ, Huang YZ, Xiao BY, Qian XC, Yao H, Hu Y, Gu YL, Zhang C, Liu KT. 1997. Toxicity from water containing arsenic and fluoride in Xinjiang. Fluoride 30:81–84. Ward MH, Cantor KP, Riley D, Merkle S, Lynch CF. 2003. Nitrate in public water supplies and risk of bladder cancer. Epidemiology 14:183– 190. Ward MH, Heineman EF, McComb RD, Merkle S, Weisenburger DD. 2005. Drinking water nitrate and risk of glioma. J Occup Environ Med 47:1260–1267. Ward MH, Mark SD, Cantor KP, Weisenburger DD, Correa-Villasenor A, Zahm SH. 1996. Drinking water nitrate and the risk of non-Hodgkin’s lymphoma. Epidemiology 7:465–471. Warner ML, Moore LE, Smith MT, Kalman DA, Fanning E, Smith AH. 1994. Increased micronuclei in exfoliated bladder cells of individuals who chronically ingest arsenic-contaminated water in Nevada. Cancer Epidemiol Biomarkers Prev 3:583–590. Webler T, Brown HS. 1993. Exposure to tetrachloroethylene via contaminated drinking water pipes in Massachusetts: A predictive model. Arch Environ Health 48:293–297. Weinberg HS, Krasner SW, Richardson SD, Thruston AD Jr. 2002. The Occurrence of Disinfection Byproducts (DBPs) of Health Concern in Drinking Water: Results of a Nationwide DBP Occurrence Study. EPA 600/R–02/068. Washington, DC: EPA. Weisenburger DD. 1991. Potential health consequences of ground-water contamination by nitrates in Nebraska. In: Bogardi I, Kuzelka RD. eds. Nitrate Contamination. Berlin: Springer-Verlag, pp. 309–315. Weyer PJ, Cerhan JR, Kross BC, Hallberg GR, Kantamneni J, Breuer G, Jones MP, Zheng W, Lynch CF. 2001. Municipal drinking water nitrate level and cancer risk in older women: The Iowa Women’s Health Study. Epidemiology 12:327–338. Whitaker HJ, Nieuwenhuijsen MJ, Best NG. 2003. The relationship between water concentrations and individual uptake of chloroform: A simulation study. Environ Health Perspect 111:688–694. Wiencke JK, Yager JW, Varkonyi A, Hultner M, Lutze LH. 1997. Study of arsenic mutagenesis using the plasmid shuttle vector pZ189 propagated in DNA repair proficient human cells. Mutat Res 386:335–344. Wigle DT. 1977. Cancer mortality in relation to asbestos in municipal water supplies. Arch Environ Health 32:185–189. Wilkins JRI, Comstock GW. 1981. Source of drinking water at home and sitespecific cancer incidence in Washington County, Maryland. Am J Epidemiol 114:178–190. Williams A, Jalaludin B. 1998. Cancer incidence and mortality around a hazardous waste depot. Aust N Z J Public Health 22:342–346. World Health Organization. 2001a. Arsenic in drinking water: Fact sheet No. 210. Geneva: World Health Organization. World Health Organization. 2001b. Environmental Health Criteria 224: Arsenic and Arsenic Compounds. Second Edition. The Environmental Health Criteria Series. Geneva: World Health Organization, pp. 1–521. Wright JM, Schwartz J, Vartiainen T, Maki-Paakkanen J, Altshul L, Harrington JJ, Dockery DW. 2002. 3-Chloro-4-(dichloromethyl)-5-hydroxy2(5H)-furanone (MX) and mutagenic activity in Massachusetts drinking water. Environ Health Perspect 110:157–164. Wu M-M, Kuo T-L, Hwang Y-H, Chen C-J. 1989. Dose-response relation between arsenic concentration in well water and mortality from cancers and vascular diseases. Am J Epidemiol 130:1123–1132. Xu X, Mariano TM, Laskin JD, Weisel CP. 2002. Percutaneous absorption of trihalomethanes, haloacetic acids, and haloketones. Toxicol Appl Pharmacol 184:19–26. Yang CY, Cheng MF, Tsai SS, Hsieh YL. 1998. Calcium, magnesium, and nitrate in drinking water and gastric cancer mortality. Jpn J Cancer Res 89:124–130. Yang CY, Chiu HF. 1998. Calcium and magnesium in drinking water and risk of death from rectal cancer. Int J Cancer 77:528–532. Yang CY, Chiu HF, Cheng MF, Tsai SS, Hung CF, Lin MC. 1999a. Esophageal cancer mortality and total hardness levels in Taiwan’s drinking water. Environ Res 81:302–308.

404

PART III: THE CAUSES OF CANCER

Yang CY, Chiu HF, Cheng MF, Tsai SS, Hung CF, Tseng YT. 1999b. Pancreatic cancer mortality and total hardness levels in Taiwan’s drinking water. J Toxicol Environ Health A 56:361–369. Yang CY, Chiu HF, Chiu JF, Cheng MF, Kao WY. 1997a. Gastric cancer mortality and drinking water qualities in Taiwan. Arch Environ Contam Toxicol 33:336–340. Yang CY, Chiu HF, Chiu JF, Tsai SS, Cheng MF. 1997b. Calcium and magnesium in drinking water and risk of death from colon cancer. Jpn J Cancer Res 88:928–933. Yang CY, Chiu HF, Tsai SS, Chang CC, Chuang HY. 2002a. Magnesium in drinking water and the risk of death from liver cancer. Magnes Res 15:223–228. Yang CY, Chiu HF, Tsai SS, Cheng MF, Lin MC, Sung FC. 2000. Calcium and magnesium in drinking water and risk of death from prostate cancer. J Toxicol Environ Health A 60:17–26. Yang CY, Chiu HF, Tsai SS, Wu TN, Chang CC. 2002b. Magnesium and calcium in drinking water and the risk of death from esophageal cancer. Magnes Res 15:215–222. Yang CY, Hung CF. 1998. Colon cancer mortality and total hardness levels in Taiwan’s drinking water. Arch Environ Contam Toxicol 35:148–151. Yang CY, Tsai SS, Lai TC, Hung CF, Chiu HF. 1999c. Rectal cancer mortality and total hardness levels in Taiwan’s drinking water. Environ Res 80:311–316. Yang LS, Peterson PJ, Williams WP, Wang WY, Hou SF, Tan JA. 2002c. The relationship between exposure to arsenic concentrations in drinking water and the development of skin lesions in farmers from Inner Mongolia, China. Environ Geochem Health 24:293–303.

Yiamouyiannis J, Burk D. 1977. Fluoridation and cancer. Age-dependence of cancer mortality related to artificial fluoridation. Fluoride 10:102– 125. Young TB, Kanarek MS, Tsiatis AA. 1981. Epidemiologic study of drinking water chlorination and Wisconsin female cancer mortality. J Natl Cancer Inst 67:1191–1198. Young TB, Wolf DA, Kanarek MS. 1987. Case-control study of colon cancer and drinking water trihalomethanes in Wisconsin. Int J Epidemiol 16:190–197. Yu RC, Hsu KH, Chen CJ, Froines JR. 2000. Arsenic methylation capacity and skin cancer. Cancer Epidemiol Biomarkers Prev 9:1259– 1262. Zakharyan R, Wu Y, Bogdan GM, Aposhian HV. 1995. Enzymatic methylation of arsenic compounds:assay, partial purification, and properties of arsenite methyltransferase and monomethylarsonic acid methyltransferase of rabbit liver. Chem Res Toxicol 8:1029–1038. Zakharyan RA, Aposhian HV. 1999. Arsenite methylation by methylvitamin B12 and glutathione does not require an enzyme. Toxicol Appl Pharmacol 154:287–291. Zakharyan RA, Sampayo-Reyes A, Healy SM, Tsaprailis G, Board PG, Liebler DC, Aposhian HV. 2001. Human monomethylarsonic acid (MMA(V) ) reductase is a member of the glutathione-S-transferase superfamily. Chem Res Toxicol 14:1051–1057. Zhao CQ, Young MR, Diwan BA, Coogan TP, Waalkes MP. 1997. Association of arsenic-induced malignant transformation with DNA hypomethylation and aberrant gene expression. Proc Natl Acad Sci U S A 94: 10907–10912.

21

Diet and Nutrition WALTER C. WILLETT

T

he possibility that diet may be important in the cause and prevention of cancer in humans has received major attention only in the past 25 years, despite long-standing knowledge that tumor incidence in animals can be influenced by nutritional manipulation (Tannenbaum, 1942). In an extensive review of causes of cancer, Doll and Peto (Doll and Peto, 1981) suggested that 35% of cancer deaths in the United States might be due to dietary factors. However, this estimate was highly uncertain, as they believed that the effect of diet could actually be as low as 10% or as high as 70%. The specific dietary factors that may cause or prevent cancer were also uncertain and have been the focus of much subsequent investigation. Possible relationships of diet with specific cancer sites are discussed within the appropriate chapters of this book. For this reason and because information on diet and cancer is still rapidly evolving, this chapter will be devoted primarily to a review of approaches and methods that are used to study relationships of diet and cancer, including considerations of their strengths and limitations. Issues involved in epidemiological studies of diets are discussed in more detail elsewhere (Willett, 1998). This chapter will conclude by briefly noting some dietary factors that have received particular attention as being possibly related to risk of certain cancers.

GENERAL APPROACHES TO THE STUDY OF DIET AND CANCER Hypotheses and supporting evidence relating dietary factors to cancer can be obtained from a variety of sources, including in vitro studies, animal experiments, small intervention studies using biochemical or molecular outcomes, epidemiological observations, and randomized trials using cancer or premalignant changes as the end points.

In Vitro Studies and Animal Experiments Many substances that cause mutations among microorganisms also cause cancer in animals and humans. This observation underlies the usefulness of microbial mutagenicity tests (such as the Ames test), which have been widely used to study components of human diets. These tests are attractive because results are available in only days and at a relatively low cost. Although these tests are helpful in directing human research and elucidating mechanisms of action, they cannot by themselves provide information that is directly relevant to humans (Ames et al., 1987). For example, many substances that influence the risk of cancer are not mutagenic. They may act, for example, by affecting the permeability of host tissues to carcinogens, by altering hormonal balances that inhibit or promote tumor growth, by changing the immune response of the host, or by affecting the rate of cell division, which in turn influences the likelihood that a mutation is reproduced. Because these functions are not all replicated in bacterial testing systems, false-negative and false-positive results will appear. Experimental exposure of laboratory animals to substances that may influence cancer incidence is more likely to simulate the effect of a chemical or food on the incidence of cancer in humans. High doses of potential carcinogens that do not reflect human experience have typically been used to induce tumors, and the response to different diets was then assessed. Animal models are now available with genetic alterations, such as a mutated tumor suppressor gene, that are more

likely to simulate human carcinogenesis. However, species differ in many ways, such as the function of their enzymatic systems to activate or deactivate potentially carcinogenic substances. Such factors preclude direct extrapolation of findings from animal experiments to humans, even though they may provide critical direction for research and aid in the interpretation of epidemiological studies.

Metabolic and Biochemical Studies Another approach to the study of diet and cancer involves metabolic or biochemical studies in humans. For example, Goldin and coworkers (Goldin et al., 1981) have studied the effect of diet on estrogen profiles, which in turn are thought to be related to the risk of breast cancer. Markers of DNA damage provide another salient outcome. These studies do not address the relations between dietary intake and the occurrence of cancer directly, but they can also be invaluable in the interpretation of other forms of evidence.

Epidemiological Studies Epidemiological studies of diet and cancer constitute a relatively new area of research. Until the 1980s, many nutritionists and epidemiologists thought that the difficulties of assessing the diets of free-living human beings over extended periods of time made large-scale studies impossible. However, approaches for assessing dietary intake have been developed and have been shown to be informative. These include standardized questionnaires to assess intakes of foods from which nutrient intakes can be calculated, biochemical determinations of body tissues, and anthropometric measurements. Because the measurement of dietary intake is a central issue, these methods will be discussed in more detail later.

Correlation Studies Early epidemiological investigations of diet and cancer consisted largely of “ecological” or “correlational” studies—comparisons of disease rates in populations with the population per capita consumption of specific dietary factors. The dietary information in such studies is usually based on “disappearance” data, meaning the national figures for food produced and imported minus the food that is exported, fed to animals, or otherwise not available for humans. Many of the correlations based on such information are remarkably strong; for example, the correlation between meat intake and incidence of colon cancer is 0.85 for men and 0.89 for women (Armstrong and Doll, 1975). The use of international correlational studies to evaluate the relationships between diet and cancer has several strengths. Most importantly, the contrasts in dietary intake are typically large. For example, within the United States, most individuals consume between 25% and 45% of their calories from fat (Willett et al., 1987), whereas the mean fat intake for populations in different countries varies from approximately 15% to 42% of calories (Goodwin and Boyd, 1987). Second, the average of diets for persons residing in a country are likely to be more stable over time than are the diets of individual persons within the country; for most countries, the changes in per capita dietary intakes over a decade or two are relatively small. Finally, the cancer rates on which international studies are based are usually derived from relatively large populations and are therefore subject to only small random errors.

405

406

PART III: THE CAUSES OF CANCER

The primary problem of such correlational studies is that many potential determinants of cancer other than the dietary factor under consideration may vary between areas with a high and low incidence of disease. Such confounding factors can include genetic predisposition; other dietary factors, including the availability of total energy intake; and other environmental or lifestyle practices. For example, with few exceptions, countries with a low incidence of colon cancer tend to be economically undeveloped. Therefore, any variable related to industrialization will be similarly correlated with incidence of colon cancer. Indeed, the correlation between gross national product and colon cancer mortality rate is 0.77 for men and 0.69 for women (Armstrong and Doll, 1975). More complex analyses can be conducted of such ecological data that control for some of the potentially confounding factors. For example, McKeown-Eyssen and Bright-See (McKeown-Eyssen and Bright-See, 1985) found that an inverse association of per capita dietary fiber intake and national colon cancer mortality rates persisted after adjustment for fat intake. Most correlational studies are also limited by the use of disappearance data that are only indirectly related to intake and are likely to be of variable quality. For example, the higher “disappearance” of calories per capita for the United States compared with most countries is probably related in part to wasted food in addition to higher actual intake. In addition, aggregate data for a geographical unit as a whole may be only weakly related to the diets of those individuals at risk of disease. As an extreme example, the interpretation of correlational data regarding alcohol intake and breast cancer is complicated because, in some cultures, most of the alcohol is consumed by men, but it is the women who develop breast cancer. These issues of data quality can potentially be addressed by collecting information on actual dietary intake in a uniform manner from the population subgroups of interest. This has been done in a study conducted in 65 geographical areas within China that are characterized by an unusually large variation in rates of many cancers (Chen et al., 1990). Another serious limitation of the international correlational studies is that they cannot be independently reproduced, which is an important part of the scientific process. Although the dietary information can be improved and the analyses can be refined, the data will not really be independent even as more information becomes available over time; the populations, their diets, and the confounding variables will be the same. Thus, it is not likely that many new insights will be obtained from further ecological studies among countries. The role of correlational studies in nutritional epidemiology is controversial. Clearly, these analyses have stimulated much of the current research on diet and cancer, and in particular they have emphasized the major differences in cancer rates among countries. Traditionally, such studies have been considered the weakest form of evidence, primarily due to the potential for confounding by factors that are difficult to measure and control (Kinlen, 1983). Some have argued that such studies provide strong evidence for evaluating hypotheses relating diet to cancer (Hebert and Miller, 1988; Prentice et al., 1988). On balance, ecological studies have unquestionably been useful but are far from conclusive regarding the relationships between dietary factors and disease and may sometimes be highly misleading.

Special Exposure Groups Subgroups within a population that consume unusual diets provide an additional opportunity to learn about the relation of dietary factors and disease. These groups are often defined by religious or ethnic characteristics and provide many of the same strengths as ecological studies. In addition, the special populations often live in the same general environment as the comparison group, which may somewhat reduce the number of alternative explanations for any differences that might be observed. For example, the observation that colon cancer mortality in the largely vegetarian Seventh-Day Adventists is only about half that expected (Phillips et al., 1980) has been used to support the hypothesis that meat consumption is a cause of colon cancer. Findings based on special exposure groups are subject to many of the same limitations as ecological studies. Many factors, both dietary and nondietary, are likely to distinguish these special groups from the comparison population. Thus, another possible

explanation for the lower colon cancer incidence and mortality among the Seventh-Day Adventist population is that differences in rates are attributable to a lower use of alcohol and tobacco or higher vegetable consumption. Given the many alternative explanations, such studies may be particularly useful when a hypothesized association is not observed. For example, the finding that the breast cancer mortality rate among the Seventh-Day Adventists is not appreciably different from the rate among the general United States population provides fairly strong evidence that eating meat is not a major cause of breast cancer.

Migrant Studies and Secular Trends Migrant studies have been particularly useful in addressing the possibility that the correlations observed in the ecological studies are due to genetic factors. For most cancers, populations migrating from an area with its own pattern of cancer incidence rates acquire rates characteristic of their new location (Adelstein et al., 1979; McMichael and Giles, 1988; Staszewski and Haenszel, 1965; Ziegler et al., 1993), although, for a few tumor sites, this change occurs only in later generations (Buell, 1973; Haenszel et al., 1972). Therefore, genetic factors cannot be primarily responsible for the large differences in cancer rates among these countries. Migrant studies may also be useful for examining the latency or relevant time of exposure. Major changes in the rates of a disease within a population over time provide evidence that nongenetic factors play an important role in the etiology of that disease. In Iceland, for example, rates of breast cancer rose dramatically over the first half of this century (Bjarnason et al., 1974). These secular changes clearly demonstrate that environmental factors, possibly including diet, are primary causes of this disease, even though genetic factors may still influence who becomes affected given an adverse environment.

Case-Control and Cohort Studies Many of the weaknesses of correlational studies are potentially avoidable in case-control studies (in which information about previous diet is obtained from diseased patients and compared to that of subjects without the disease) or cohort investigations (in which information on diet is obtained from disease-free subjects who are then followed to determine disease rates according to levels of dietary factors). In such studies, the confounding effects of other factors can be controlled either in the design (by matching subjects to be compared on the basis of known risk factors, or by restriction) or in the analysis (by any of a variety of multivariate methods) if information has been collected on the confounding variables. Furthermore, dietary information can be obtained for the individuals actually affected by disease rather than using the average intake of the population as a whole. Case-control studies generally provide information more efficiently and rapidly than cohort studies because the number of subjects is typically far smaller and no follow-up is necessary. However, concerns have existed whether consistently valid results can be obtained from case-control studies of dietary factors and disease because of the inherent potential for methodological bias. This potential for bias is not unique for diet but is likely to be unusually serious for several reasons. Due to the limited range of variation in diet within most populations and some inevitable error in measuring intake, realistic relative risks in most studies of diet and disease are likely to be modest, say of the order 0.5 to 2.0. These relative risks may seem small but would be quite important because the prevalence of exposure is high. Given typical distributions of dietary intake, these relative risks are usually based on differences in means for cases and controls (or those who become cases and those who remain noncases in prospective studies) of only about 5%. Thus, a systematic error of even 3% or 4% can seriously distort such a relationship. In case-control studies, it is plausible that biases (due to selection or recall) of this magnitude could often occur, and it is extremely difficult to exclude the possibility that this degree of bias has occurred in any specific study. Hence, it would not be surprising if case-control studies of dietary factors provide inconsistent findings. The selection of an appropriate control group for a study of diet and cancer is also usually problematic. One common practice is to use

407

Diet and Nutrition patients with another disease for comparison, with the assumption that the exposure under study is unrelated to the condition of this control group. However, because diet may influence the incidence of many diseases, it is often difficult to identify disease groups that are, with confidence, unrelated to the aspect of diet under investigation. A common alternative is to use a sample of persons from the general population as the control group. In many areas, particularly large cities, participation rates are low; it is now common for only 60% or 70% of eligible population controls to complete an interview (Hartge et al., 1984). Because diet is particularly associated with the level of general health consciousness, the diets of those who participate are likely to differ from those who do not. For example, controls who participate are likely to be more health-conscious and thus to consume more fruits and vegetables and less animal fat, which would tend to create artifactual inverse and positive associations, respectively, with cancer risk. Direct evidence regarding the magnitude of biases in casecontrol studies of diet is limited. In two large prospective studies of diet and cancer, diets of patients with breast cancer and a sample of control participants were also assessed retrospectively. In one study, no evidence of recall bias was observed (Friedenreich et al., 1991), but in the other the combination of recall and selection bias did seriously distort associations with fat intake (Giovannucci et al., 1993b). Even if many studies arrive at correct conclusions, distortion of true associations in a substantial percentage would produce an inconsistent body of published data, making a coherent synthesis difficult or impossible for a specific diet and cancer relationship. Methodological sources of inconsistency may be particularly troublesome in nutritional epidemiology because of the inherent biological complexity resulting from nutrient–nutrient interactions. Because the effect of one nutrient may depend on the level of another (which can differ between studies and may not have been measured), such interactions may result in apparently inconsistent findings in epidemiological studies. Thus, compounding biological complexity with methodological inconsistency may result in an uninterpretable literature. Prospective cohort studies avoid most of the potential sources of methodological bias associated with case-control investigations. Because the dietary information is collected before the diagnosis of disease, illness cannot affect the recall of diet. Although losses to follow-up that vary by level of dietary factors can result in distorted associations in a cohort study, follow-up rates tend to be rather high because participants have already provided evidence of willingness to participate, and they may also be followed passively by means of disease registries and vital record listings (Stampfer et al., 1984). In addition to being less susceptible to bias, prospective cohort studies provide the opportunity to obtain repeated assessments of diet over time and to examine the effects of diet on a wide variety of diseases, including total mortality, simultaneously. The primary constraints on prospective studies of diet are practical. Even for common cancers such as those of the lung, breast, or

2

Case-Control

colon, it is necessary to enroll tens of thousands of subjects. The use of structured, self-administered questionnaires has made studies of this size possible, although still expensive. Enough cohort studies have now become available to evaluate whether findings from casecontrol studies are replicable prospectively. Case-control studies of fat and breast cancer have been summarized in a pooled analysis by Howe and colleagues (Howe et al., 1990); a small but statistically significant positive overall association was seen, although highly significant heterogeneity was also observed among studies. In contrast, when large cohort studies were pooled, no association was seen, and there was no heterogeneity among studies (Hunter et al., 1996) (see Figure 21–1). Even more strikingly, positive associations have been seen consistently in case-control studies of dietary fat and lung cancer, whereas these associations are consistently absent in prospective studies (Smith-Warner et al., 2002). Similarly, suggestions of inverse associations between intake of foods high in betacarotene and risk of lung cancer were not supported in a pooled analysis of prospective studies (Smith-Warner et al., 2002). These inconsistencies between case-control and cohort studies strongly suggest that concerns regarding the potential for biases in the former are justified, and that conclusions about diet and cancer should not rely on such studies. However, for diseases of somewhat low frequency, even very large cohorts will not accumulate a sufficient number of cases within a reasonable amount of time, and casecontrol studies will continue to play a role in nutritional epidemiology. Care in design and caution in the interpretation will be important. Due to current uncertainty about measuring diets in early life, whether either study design will be able to address the influence of childhood diet on disease occurring decades later is currently unclear.

Controlled Trials The most rigorous evaluation of a dietary hypothesis is the randomized trial, optimally conducted as a double-blind experiment. The principal strength of a randomized trial is that potentially distorting variables should be distributed at random between the treatment and control groups, thus minimizing the possibility of confounding by these extraneous factors. In addition, it is sometimes possible to create a larger contrast between the groups being compared by use of an active intervention. Such experiments among humans, however, are best justified after considerable nonexperimental data have been collected to ensure that benefit is reasonably probable and that an adverse outcome is unlikely. Experimental studies are particularly practical for evaluating hypotheses that minor components of the diet, such as trace elements or vitamins, can prevent cancer, as these nutrients can be formulated into pills or capsules. Even if feasible, randomized trials of dietary factors and disease are likely to encounter several limitations. The time between change in

Prospective

1

0.5

*p, Heterogeneity < 0.001

**p, Heterogeneity = 0.24

Study

S HS b N HS a SM C N YS C C CS S IW H To S ta l* *

LC

N

N

L ta l* To

K

I

J

H

G

F

E

D

C

B

0 A

Relative Risk

1.5

Figure 21–1. Case-control and cohort studies of dietary fat and breast cancer. Case-control data abstracted from Howe and colleagues (Howe et al., 1990) and prospective data from Hunter and colleagues. (Source: Hunter et al., 1996.)

408

PART III: THE CAUSES OF CANCER

the level of a dietary factor and any expected change in the incidence of disease is typically uncertain. Therefore, trials must be of long duration, and the possibility that any lack of difference between treatment groups may be due to insufficient duration is difficult to exclude. Compliance with the treatment diet is likely to decrease during an extended trial, particularly if treatment involves a real change in food intake, and the control group may well adopt the dietary behavior of the treatment group if the treatment diet is thought to be beneficial. Such trends, which were found in the Multiple Risk Factor Intervention Trial of coronary disease prevention (Multiple Risk Factor Intervention Trial Research Group, 1982), may obscure a real benefit of the treatment. A related potential limitation of trials is that participants who enroll in such studies tend to be highly selected on the basis of health consciousness and motivation. Therefore, it is possible that the subjects at highest potential risk on the basis of their dietary intake, and thus susceptible to intervention, are seriously underrepresented. For example, if low beta-carotene intake is thought to be a risk factor for lung cancer, and a trial of beta-carotene supplementation is conducted among a health-conscious population that includes few individuals with low beta-carotene intake, no effect might be observed simply because most members of the study population were already receiving the maximal benefit of this nutrient through their usual diet. In such an instance, it would be useful to measure dietary intake of betacarotene before starting the trial. Because the effect of supplementation is likely to be greatest among those with low dietary intakes, it would be possible to exclude those with high intakes (the potentially nonsusceptibles) either before randomization or in subanalyses at the conclusion of the study. This requires, of course, a reasonable measurement of dietary intake. Trials are sometimes said to provide a better quantitative measurement of the effect of an exposure or treatment because the difference in exposure between groups is better measured than in an observational study. Although this contrast may at times be better defined in a trial (it is usually clouded by some degree of noncompliance), trials still usually produce an imprecise measure of the effect of exposure because of marginally adequate sample sizes and ethical considerations that require stopping soon after a statistically significant effect is seen. For example, with a P-value close to 0.05, the 95% confidence interval will extend from no effect to a strong effect that is usually implausible. In an observational study, an ethical imperative to stop does not exist when statistical significance occurs; continued accumulation of data can provide increasing precision regarding the relation between exposure and disease. A trial can provide unique information on the latent period between change in an exposure and change in diet; since spontaneous changes in diet are typically not clearly demarcated in time, the estimation of latent periods for dietary effects will usually be difficult in observational studies. Although all hypotheses would ideally be evaluated in randomized trials, this will sometimes be impossible for practical or ethical reasons. For example, our knowledge of the effects of cigarette smoking on risk of lung cancer is based on observational studies, and it is similarly unlikely that randomized trials could be conducted to examine the effect of alcohol use on human breast cancer risk. It remains unclear whether trials of sufficient size, duration, and degree of compliance can be conducted to evaluate many hypotheses that involve major behavioral changes in eating patterns, such as a reduction in fat intake (Michels and Willett, 1992).

MEASUREMENT OF DIET IN EPIDEMIOLOGICAL STUDIES The complexity of the human diet represents a daunting challenge to anyone contemplating a study of its relation to cancer. The foods we consume each day contain literally thousands of specific chemicals, some known and well quantified, some characterized only poorly, and others completely undescribed and currently unmeasurable. In human diets, intakes of various components tend to be intercorrelated. With

few exceptions, all individuals are exposed; for example, everyone eats fat, fiber, and vitamin A. Thus, dietary exposures can rarely be characterized as present or absent; rather, they are continuous variables, sometimes with rather limited range of variation between persons. Furthermore, individuals are generally not aware of the content of the foods that they eat; hence, the consumption of nutrients is usually determined indirectly. The chemicals that constitute our food can be described by the non–mutually exclusive categories given in Table 21–1.

Nutrients versus Foods Throughout nutrition in general and in much of the existing cancer literature, diet has usually been described in terms of its nutrient content. Alternatively, diet can be described in terms of foods or food groups. The primary advantage of representing diets as specific compounds, such as nutrients, is that such information can be directly related to our fundamental knowledge of biology. From a practical perspective, the exact structure of a compound must usually be known if it is to be synthesized and used for supplementation. In epidemiological studies, measurement of total intake of a nutrient (as opposed to using the contribution of only one food at a time) provides the most powerful test of a hypothesis, particularly if many foods contribute to intake of that nutrient. For example, in a particular study it is quite possible that total fat intake could be clearly associated with risk of disease, whereas none of the contributions to fat intake by individual foods would be significantly related to disease on its own. The use of foods to represent diet has several practical advantages when examining relationships with disease. Particularly when suspicion exists that some aspect of diet is associated with risk but a specific hypothesis has not been formulated, an examination of the relations of foods and food groups with risk of disease will provide a means to explore the data. Associations observed with specific foods may lead to a hypothesis relating to a defined chemical substance. For example, observations that higher intakes of green and yellow vegetables were associated with reduced rates of lung cancer led to the hypothesis that beta-carotene might protect DNA from damage caused by free radicals and singlet oxygen (Peto et al., 1981). The finding by Graham and co-workers (Graham et al., 1978) that intake of cruciferous vegetables was inversely related to risk of colon cancer supported the suggestion that indole compounds contained in these vegetables may be protective (Wattenberg and Loub, 1978). Even more seriously than the lack of a well-formulated hypothesis, the premature focus on a specific nutrient that turns out to have no relation with disease may lead to the erroneous conclusion that diet has no effect. Mertz (Mertz, 1984) has pointed out that foods are not fully represented by their nutrient composition, noting as an example that milk and yogurt produce different physiologic effects despite a similar nutrient content. Furthermore, the valid calculation of a nutri-

Table 21–1. Aspects of Diet Related to Cancer Category

Examples

Essential nutrients

Vitamins, specific fatty acids, amino acids, and minerals Proteins, carbohydrates, fats, and alcohol Preservatives (nitrates, BHT, salt) and coloring and flavoring agents Pesticides, herbicides, fungicides, growth hormones Aflatoxin Cadmium, lead, polychlorinated biphenyls Heterocyclic amines from cooked meat

Major energy sources Additives Agricultural chemical contaminants Microbial toxin contaminants Inorganic contaminants Chemicals formed in cooking or processing of food Natural toxins Other natural compounds

Safrole (in natural root beer), pyrrolizidine alkaloid (in comfrey tea), hydrazines (in mushrooms) Protease inhibitors, indoles, cholesterol

Diet and Nutrition ent intake from data on food consumption requires that reasonably accurate food composition information is available, which markedly constrains the scope of dietary chemicals that may be investigated because such information exists for only several dozen commonly studied nutrients. Epidemiological analyses based on foods, as opposed to nutrients, are generally most directly related to dietary recommendations because individuals and institutions ultimately manipulate nutrient intake largely by their choice of foods. Even if the intake of a specific nutrient is convincingly shown to be related to risk of cancer, this is not sufficient information on which to make dietary recommendations. Because foods are an extremely complex mixture of different chemicals that may compete with, antagonize, or alter the bioavailability of any single nutrient contained in that food, it is not possible to predict with certainty the health effects of any food solely on the basis of its content of one specific factor. For example, there is concern that high intake of nitrates may be deleterious, particularly with respect to gastrointestinal cancer. However, the primary sources of nitrates in our diets are green, leafy vegetables, which, if anything, appear to be associated with reduced risk of cancer at several sites. Similarly, because of the high cholesterol content of eggs, their avoidance has received particular attention in diets aimed at reducing the risk of coronary heart disease; per capita consumption of eggs declined by 25% in the United States between 1948 and 1980 (Welsh and Marston, 1982). However, eggs are more than cholesterol capsules; they provide a rich source of essential amino acids and micronutrients and are relatively low in saturated fat. It is thus difficult to predict the net effect of egg consumption on risk of coronary heart disease, much less the effect on overall health, without empirical evidence. Given the strengths and weaknesses of using nutrients or foods to represent diet, it appears that an optimal approach to epidemiological analyses will employ both. In this way, a potentially important finding is least likely to be missed. Moreover, the case for causality is strengthened when an association is observed with overall intake of a nutrient and also with more than one food source of that nutrient, particularly when the food sources are otherwise different. This provides, in some sense, multiple assessments of the potential for confounding by other nutrients; if an association was observed for only one food source of the nutrient, other factors contained in that food would tend to be similarly associated with disease. As an example, the hypothesis that alcohol intake causes breast cancer was strengthened by observing not only an overall association between alcohol intake and breast cancer risk, but also by independent associations with both beer and liquor intake, thus making it less likely that some factor other than alcohol in these beverages was responsible for the increased risk. One practical drawback of using foods to represent diet is their large number and complex, often reciprocal, interrelationships that are largely due to individual behavioral patterns. Many reciprocal relationships emerge upon perusal of typical data sets; for example, eaters of dark bread tend not to eat white bread, margarine users tend not to eat butter, and skim milk users tend not to use whole milk. This complexity is, of course, one of the reasons to compute nutrient intakes that summarize the contributions of all foods. An intermediate solution to the problem posed by the complex interrelationships among foods is to use food groups or to compute the contribution of nutrient intake from various food groups. For example, Manousos and co-workers (Manousos et al., 1983) combined the intakes of foods from several predefined groups to study the relation of diet to risk of colon cancer; they observed increased risk among subjects with high meat intake and with low consumption of vegetables. The computation of nutrient intakes from different food groups is illustrated by the partitioning of dietary fiber, an extremely heterogeneous collection of substances, into fiber from grains, fiber from fruits, and fiber from vegetables. This circumvents the inadequacy of food composition databases and also provides information in a form that is directly useful to individuals faced with decisions regarding choices of foods. In general, maximal information will be obtained when analyses are conducted at the levels of nutrients, foods, and food groups.

409

The Dimension of Time The assessment of diet in studies of cancer incidence is further complicated by the dimension of time. Because our understanding of the pathogenesis of most cancers is limited, considerable uncertainty exists about the period in time before diagnosis for which diet might be relevant. For some cancers, aspects of diet may be important during childhood even though the disease occurs decades later. For other cancers, it has been suggested that diet may act as a promoting or inhibiting factor; thus, intake over a continuous period up to just prior to diagnosis may be important. Ideally, data on dietary intake at different points prior to diagnosis could help to resolve these issues. However, individuals rarely make clear changes in their diet at identifiable points in time; more typically, eating patterns evolve over periods of years. Thus, epidemiologists are usually forced to direct questions about diet to a period several years before diagnosis of cancer with the hope that diet at this point in time will represent, or at least be correlated with, diet during the critical period in cancer development. Fortunately, diets of individuals do tend to be correlated from year to year, so that some imprecision in identification of critical periods of exposure may not be serious. For most nutrients, correlations for repeated assessments of diet at intervals from 1 to about 10 years tend to be of the order 0.6 to 0.7 (Byers et al., 1987; Rohan and Potter, 1984; Willett et al., 1985b), with decreasing correlations over longer intervals (Byers et al., 1983). For scientists accustomed to measurements made under highly controlled conditions in a laboratory, this may seem like a low degree of reproducibility. However, these correlations are similar to other biological measurements made in free-living populations, such as serum cholesterol (Shekelle et al., 1991) and blood pressure (Rosner et al., 1977). Even though diets of individuals tend to have a strong element of consistency over intervals of years, they are characterized by marked variation from day to day (Beaton et al., 1979). This variation differs from nutrient to nutrient, being moderate for total energy intake, but extreme for cholesterol and vitamin A. For this reason, even perfect information about diet on any single day or the average of a small number of days will poorly represent long-term average intake, which is likely to be more relevant to cancer etiology.

General Methods of Dietary Assessment Three general approaches have been used to assess dietary intake: information about intake of foods that can be used directly or to calculate intake of nutrients, biochemical measurements of blood or other body tissues that provide indicators of diet, and measures of body dimensions or composition that reflect the long-term effects of diet. Because the interpretation of data on diet and cancer is heavily influenced by the methods used to assess diet, features of these methods and their limitations will be considered.

Methods Based on Food Intake Short-Term Recall and Diet Records. The 24-hour recall, in which subjects are asked to report their food intake during the previous day, has been a widely used dietary assessment method. It has been the basis of most national surveys, including NHANES, and numerous prospective studies of coronary heart disease. Interviews are conducted by nutritionists or trained interviewers, usually using visual aids such as food models or shapes to obtain data on quantities of foods. The 24-hour recall requires about 10 to 20 minutes for an experienced interviewer; although usually conducted in person, it has also been done by telephone using a two-dimensional chart that is mailed beforehand to assist in the estimation of portion sizes (Posner et al., 1982). This method has the advantages of requiring no training or literacy and minimal effort on the part of the participant. Dietary records or food diaries are detailed meal-by-meal recordings of types and quantities of foods and beverages consumed during a specified period, typically 3 to 7 days. Ideally, subjects weigh each portion of food before eating, although this is frequently impossible for all meals. Alternatively, household measures can be used to estimate portion sizes. The method places a considerable burden upon the

410

PART III: THE CAUSES OF CANCER

subject, thus limiting its application to those who are literate and highly motivated. In addition, the effort involved in keeping diet records may increase awareness of food intake and induce an alteration in diet. However, diet recording has the distinct advantages of not depending on memory and allowing direct measurements of portion sizes. The validity of 24-hour recalls has been assessed by observing the actual intake of subjects in a controlled environment and interviewing them the next day. In such a study, Karvetti and Knuts (Karvetti and Knuts, 1985) observed that subjects both erroneously recalled foods that were not actually eaten and omitted foods that were eaten; correlations between nutrients calculated from observed intakes with calculations from the recalled information ranged from 0.58 to 0.74. In a similar study among elderly persons, Madden and co-workers (Madden et al., 1976) found correlations ranging from 0.28 to 0.87. Relatively few validation studies have been conducted of diet recordings. In a comparison of nitrogen intake calculated from diet records with intake based on analyses of replicate meals, Bingham and Cummings (Bingham and Cummings, 1985) found a correlation of 0.97. The most serious limitation of the 24-hour recall method is that dietary intake is highly variable from day to day. Diet records reduce the problem of day-to-day variation because the average of a number of days is used. For nutrients that vary substantially, however, even a week of recording will still not provide an accurate estimate of an individual’s intake (Beaton et al., 1979). The variability in intake of specific foods is even greater than for nutrients (Salvini et al., 1989), so that only very commonly eaten foods can be studied by this method. The problem of day-to-day variation is not an issue if the objective of a study is to estimate a mean intake for a population, as might be the goal in an ecological study. However, in case-control or cohort investigations, accurate estimation of individual intakes is necessary. Practical considerations and issues of study design further limit the application of short-term recall and diet record methods in epidemiological studies. Because they provide information on current diet, their use will typically be inappropriate in case-control studies because the relevant exposure will have occurred earlier and diet may have changed as a result of the cancer or its treatment. A few exceptions may occur, such as in the case of very early tumors or premalignant lesions. Although the average of multiple days of 24-hour recalls or diet recording could theoretically be used in prospective studies of diet and cancer, the costs may be prohibitive because of the large numbers of subjects required and substantial expense involved in collecting this information and processing it. These methods, however, can play an important role in the validation or calibration of other methods of dietary assessment that are more practical for epidemiological studies.

Food Frequency Questionnaire. Because short-term recall and diet record methods are generally expensive, inappropriate for assessment of past diet, and may be unrepresentative of usual intake, investigators have sought alternative methods for measuring long-term dietary intake. Burke (Burke, 1947) developed a detailed dietary history interview that attempted to assess an individual’s usual diet; this included a 24-hour recall, a menu recorded for 3 days, and a checklist of foods consumed over the preceding month. This method was time-consuming and expensive because a highly skilled professional was needed for both the interview and processing of information. The checklist, however, was the forerunner of the more structured dietary questionnaires in use today. During the 1950s, Stephanik and Trulson (Stephanik and Trulson, 1962), Heady (Heady, 1961), Wiehl and Reed (Wiehl and Reed, 1960), and Marr (Marr, 1971) developed food frequency questionnaires and evaluated their role in dietary assessment. Heady, using diet records collected by British bank clerks, demonstrated that the frequencies with which foods were used correlated highly with the total weights of the same foods consumed over a several-day period, thus providing the theoretical basis for the food frequency method. Multiple investigators have converged toward the use of food frequency questionnaires as the method of dietary assessment best suited for most epidemiological studies of diet and cancer. During recent years, substantial refinement, modification, and

evaluation of food frequency questionnaires have occurred, so that data derived from their use has become considerably more interpretable. The basic food frequency questionnaire consists of two components: A food list and a frequency response section for subjects to report how often each food was eaten. Questions related to further details of quantity and composition may be appended. A basic decision in designing a questionnaire is whether the objective is to measure intake of a few specific foods or nutrients or whether a comprehensive assessment of dietary intake is desired. A comprehensive assessment is generally desirable whenever possible. It is often impossible to anticipate at the beginning all the questions regarding diet that will appear important at the end of a study; a highly restricted food list may not have included an item that is, in retrospect, important. Furthermore, total food intake, represented by energy consumption, may be related to disease outcome and thus confound the effects of specific nutrients or foods. Even if total energy intake is not related to a disease outcome, adjustment for total intake may increase the accuracy of specific nutrient measurements (Willett, 2001, Willett and Stampfer, 1986). Nevertheless, epidemiological practice is usually a compromise between the ideal and reality, and it may simply not be possible to include a comprehensive diet assessment in a particular interview or questionnaire, especially if diet is not the primary focus of the study. Because diets tend to be reasonably correlated from year to year, most investigators have asked subjects to describe their frequency of using foods in reference to the preceding year. This provides a full cycle of seasons so that, in theory, the responses should be independent of the time of year. In case-control studies, the time frame could be in reference to a period or a specified number of years previously. Typically, investigators have provided a multiple-choice response format, with the number of options usually ranging from five to ten. Another approach is to use an open-ended format and provide subjects the option of answering in terms of frequency per day, week, or month (Block et al., 1986). In theory, an open-ended frequency response format might provide for some enhanced precision in reporting because the frequency of use is truly a continuous rather than a categorical variable. However, the overall increment in precision is unlikely to be large, because the estimation of the frequency that a food is used is inherently an approximation. Several options exist for collecting additional data on serving sizes. The first is to collect no additional information on portion sizes at all; that is, to use a simple frequency questionnaire. A second possibility is to specify a portion size as part of the question on frequency; for example, to ask how often a “glass” of milk is consumed rather than only how often milk is consumed, which has been termed a semiquantitative food frequency questionnaire. A third alternative is to include an additional question for each food to describe the usual portion size, in words (Block et al., 1986), using food models (Morgan et al., 1978), or pictures of different portion sizes (Hankin et al., 1983). Because most of the variation in intake of a food is explained by frequency of use rather than differences in serving sizes, several investigators have found that portion size data are relatively unimportant (Block et al., 1990, Pickle and Hartman, 1985; Samet et al., 1984). Cummings and co-workers (Cummings et al., 1987) found that adding questions on portion sizes to a simple frequency questionnaire only slightly improved estimation of calcium intake; others have found that the use of food models in an in-person interview did not increase the validity of a self-administered, semiquantitative food frequency questionnaire (Hernandez-Avila et al., 1988). These findings have practical implications because the cost of data collection by mail or telephone is far less than the cost of personal interviews, which are necessary if food models are to be used for assessing portion sizes. Cohen and co-workers (Cohen et al., 1990) also found that the portion size information included in the Block questionnaire added only slightly to validity assessed by comparison with diet records (average correlation 0.41 without portion sizes and 0.43 with portion sizes). Food frequency questionnaires are extremely practical in epidemiological applications because they are easy for subjects to complete,

411

Diet and Nutrition often as a self-administered form. Processing is readily computerized and inexpensive, so that even prospective studies involving tens of thousands of subjects are feasible.

Validity of Dietary Assessment Methods The interpretation of epidemiological data on diet and cancer depends directly on the validity of the methods used to measure dietary intake. This is particularly true when no association is found because one possible explanation could be that the method used to measure diet was not able to discriminate among persons. A substantial body of evidence has accumulated regarding the validity of food frequency questionnaires, which has been the method most commonly used in epidemiological studies. In evaluating the validity of a dietary assessment method, the choice of a standard for comparison is a critical issue. As for all variables, no perfect standard exists. Thus, a desirable feature for the comparison method is that its errors be independent from the method being evaluated so that an artificial correlation will not be observed (Willett, 1998). For this reason, biochemical indicators of diet are probably the optimal standard. Their greatest limitation is that markers of diet do not exist for most of the nutrients of current interest, such as total fat, fiber, and sucrose intake. Moreover, the available biochemical indicators of diet are likely to be quite imprecise measures of diet because they are influenced by many factors such as differences in absorption and metabolism, short-term biological variation, and laboratory measurement error. However, the capacity to demonstrate a correlation between a questionnaire estimate of nutrient intake and a biochemical indicator provides useful qualitative evidence of validity. Such correlations have been reported for questionnaire estimates of dietary carotenoids, folic acid, vitamin E, vitamin B6, and specific fatty acids (Willett, 1998). If a biomarker is sensitive to change in a dietary factor, even if not specific to that dietary factor, it can still provide a quantitative estimate of the measurement error in a dietary method. For example, on the basis of many metabolic studies, increasing intake of fat reduces fasting plasma triglyceride levels. Demonstration that fat intake measured by a dietary questionnaire is inversely related to fasting triglyceride levels provides qualitative information on the validity, and comparing the slope from a regression of plasma triglycerides on fat intake with the slope anticipated from controlled feeding studies provides a quantitative assessment of measurement error (Willett et al., 2001). In an application of this approach, the slope using food frequency questionnaires was not substantially lower than that from controlled feeding studies—it was actually somewhat larger. Thus, associations between fat intake estimated by the questionnaire and

cancer risks are not likely to be seriously underestimated due to measurement error. Kipnis and colleagues (Kipnis et al., 2003) have used doubly labeled water and urinary nitrogen to assess protein intake that was also measured by 24-hour recalls and a food frequency questionnaire. This confirmed the reduction in error achieved by adjustment for total energy intake. Although these authors found evidence of correlation in errors between the 24-hour recalls and the food frequency questionnaire, the estimates of validity for the questionnaire after accounting for correlated errors were similar to those reported in previous validation studies (Willett, 2003). Validation studies of dietary questionnaires have also been conducted by comparing computed intakes with those based on other dietary assessment methods. Among the possible comparison methods, diet records are particularly attractive because their errors are likely to be minimally correlated with errors in the food frequency questionnaires, as they do not depend on memory, and scales can be used to assess portion sizes. Several of the most comprehensive validation studies involving comparisons of questionnaires completed at about a 1-year interval, before and after multiple diet records collected during the intervening months, are summarized in Table 21–2 (Block et al., 1990; Goldbohm et al., 1994; Pietinen et al., 1988a; Pietinen et al., 1988b; Rimm et al., 1992; Willett et al., 1985b). Similar degrees of misclassification were seen in these studies; for questionnaires completed at the end of the 1-year recording of diet (which corresponds to the time frame of the questionnaires), correlations adjusted for total energy intake tended to be mainly between 0.5 and 0.7. Although the degree of measurement error associated with nutrient estimates calculated from food frequency questionnaires appears to be similar to that for many epidemiological measures, these errors will lead to important underestimates of relative risks. Less commonly appreciated, the errors will also result in observed confidence intervals that are inappropriately narrow; this is of particular concern when no association is seen because the entire interest is then in the range of possible relative risks that are reasonably excluded by the data. In part generated by the interest in diet and cancer and the recognized issue of measurement error in assessing dietary intake, considerable effort has been directed to the development of methods that provide corrected estimates of relative risks and confidence intervals based on quantitative assessments of measurement error (Byar and Gail, 1989; Rosner et al., 1992; Willett, 1998). Thus, validation studies of dietary questionnaires can provide important estimates of error that can be used to quantitatively interpret the influence of error on observed associations. Based on such analyses, it can be shown that important associations will generally not be missed by typical dietary questionnaires

Table 21–2. Comparison of Food Frequency Questionnaires with Other Dietary Assessment Methods Source

Population

Comparison Methods

Interval Between Methods

Reference Period

Range of Correlations

Willett et al. (1985b)

Registered nurses (N = 194)

Diet record

1 month to 1 year

Previous year

0.36 Vitamin A without supplements to 0.75 vitamin C 0.51 vitamin A to 0.73 polyunsaturated fat

Pietinen et al. (1988a, 1988b)

Finnish men (N = 189)

1–6 months

1 year

Block et al. (1990)

260

Twelve 2-day diet records (vs. 273item questionnaire) Three 4-day diet records

1–12 months

6 months

0.37 vitamin A to 0.74 vitamin C, with supplements, average = 0.55

Rimm et al. (1992)

127 U.S. health professionals

Two 1-week diet records

1–12 months

1 year

Goldbohm et al. (1994)

59 men 48 women

Three 3-day diet records

3–15 months

1 year

0.28 for iron to 0.86 for vitamin C with supplements, average = 0.59 0.33 for B1 to 0.75 for polyunsaturated fat, average = 0.64

Comments

Adjustment for energy had little effect on correlations Correlations were similar in low-fat and usual diet group. Variable portion sizes added little to correlations. Mean correlation increased to 0.65 with adjustment for variation in diet records Adjustment for sex and energy intake had little effect except for fat intake, changing from 0.72 to 0.52

412

PART III: THE CAUSES OF CANCER

(Rosner et al., 1989), although sample sizes for studies will need to be several times larger than those estimated, assuming that measurement error did not exist (Walker and Blettner, 1985). Caution is necessary when interpreting validation studies when the standard method is based on calculated nutrient intakes, such as diet records. For nutrients that vary substantially from one specimen of a food to another, the calculated values may be highly correlated because intakes of each of the foods is measured well by both methods, but neither method is actually measuring the nutrient well. This can occur when the nutrient contents of the food vary greatly with the soil on which it was grown or raised, as is true for selenium, or when the nutrient is sensitive to degradation during processing or storage.

Biochemical Indicators of Diet The use of biochemical measurements made on blood or other tissues to indicate intake of a nutrient is attractive because this does not depend on the memory or knowledge of the subject. Furthermore, such measurements can be made in retrospect; for example, using blood specimens that have been collected and stored for other purposes.

Choice of Tissues for Analysis. Most commonly, serum or plasma has been used in epidemiological studies to measure biochemical indicators of diet. However, consideration should also be given to red blood cells, subcutaneous fat, hair, and nails. The choices should be governed by the ability of the tissue to reflect dietary intake of the factor of interest; the time-integrating characteristics of the tissue; practical considerations in collecting, transporting, and storing the specimen; and cost. These considerations are examined in detail elsewhere for a number of dietary factors (Hunter, 1998); some general comments are provided here. Red Blood Cells. For a number of dietary factors, red cells are less sensitive to short-term fluctuations in diet than plasma or serum and may thus provide a better index of long-term exposure. Nutrients that can be usefully measured in red cells include some fatty acids, folic acid, and selenium. Subcutaneous Fat. Composed primarily of fatty acids, the adipose tissue turns over slowly among individuals with relatively stable weight. For at least some fatty acids, the half-life is of the order 600 days, which makes this an ideal indicator of long-term diet in epidemiological studies. Fat-soluble vitamins such as retinol, vitamin E, and carotenoids are also measurable in subcutaneous fat, although their relations to diet are yet to be clearly established. Hair and Nails. Hair and nails incorporate many elements into their matrix during formation, and for many heavy metals these may be the tissues of choice because these elements tend to be rapidly cleared from the blood. Nails appear to be an excellent tissue for the assessment of long-term selenium intake because of their capacity to integrate exposure over time (Morris et al., 1983). Because the hair and nails can be cut at various times after formation (a few weeks for hair close to the scalp and approximately 1 year for the great toe), an index of exposure can be obtained that may be little affected by recent experiences. This can be a particular advantage in the context of a casecontrol study of diet and cancer. Contamination poses the greatest problem for measurements in hair due to its extensive exposure to the environment and great surface area; these problems are generally much less for nails but still need to be considered.

Limitations of Biochemical Indicators. Although the use of biochemical indicators for assessing diet is attractive, no practical indicators exist for many of the dietary factors implicated in the etiology of cancer. Even when tissue levels of a nutrient can be measured, these levels are often highly regulated and thus reflect dietary intake poorly; blood retinol, cholesterol, and sodium are good examples. Just as with dietary intake, the blood levels of some nutrients fluctuate substantially over time, so that one measurement may not provide a good reflection of long-term intake. Furthermore, experience has provided sobering evidence that the tissue levels of many nutri-

ents can be affected by the presence of cancer, even several years prior to diagnosis (Wald et al., 1986), rendering the use of many biochemical indicators treacherous in most case-control studies. Despite these limitations, careful application of biochemical indicators can provide unique information about dietary intake, particularly for nutrients that cannot be accurately calculated from data on food intake.

Anthropometry and Measures of Body Composition The influence of energy balance at various times in life is likely to have important effects on the incidence of some cancers. Energy balance is better reflected by measurements of body dimensions and composition than by assessments based on the difference between energy intake and expenditure (largely physical activity), because both of these variables are measured with considerable error (Willett, 1998). The most common use of anthropometric measurements is in the calculation of obesity using either indices such as Quetelet index or body mass index (weight divided by the second power of height), or relative weight (weight standardized for height). Remarkably valid estimates of weight and height can be obtained even by questioning individuals (Stunkard and Albaum, 1981), including their recall for several decades earlier (Rhoads and Kagan, 1983). Thus, estimates of obesity can be obtained easily for large prospective investigations or retrospectively in the context of case-control studies. The major limitation of obesity estimates based on height and weight is that an assessment of weight cannot differentiate between fat and lean body mass. For this reason, these are imperfect measures of obesity. Until recently, studies of the validity of BMI as a measure of obesity have used as a “gold standard” body fat expressed as a percent of total weight, usually determined by underwater weighing. However, BMI is actually a measure of fat mass adjusted for height rather than a measure of percent body fat. When fat mass, determined from densitometry, is adjusted for height and used as the standard, the correlation with BMI is approximately 0.90, indicating a substantially higher degree of validity than has generally been appreciated (Spiegelman et al., 1992). Moreover, in the same study, fat mass adjusted for height was correlated more strongly with biologically relevant variables such as blood pressure and fasting blood glucose than was percent body fat. Using the simple measure of BMI, much has been learned about the relation of positive energy balance with cancer, as well as with coronary heart disease. The electrical impedance method (Hodgdon and Fitzgerald, 1987) is a simple method of measuring adiposity that could potentially be used in epidemiological studies, but this needs further evaluation. The use of one or a small number of skin-fold thicknesses does not appear to be appreciably more accurate than weight and height in the estimation of overall adiposity but can provide additional information on the distribution of body fat. The ratio of waist-to-hip circumferences has received considerable attention in relation to cardiovascular disease, diabetes, and blood pressure (Bjorntorp, 1987; Hartz et al., 1984). This ratio has also been of interest with respect to hormonally sensitive cancers, as it has been suggested that central fat functions differently than peripheral fat with respect to estrogen metabolism (Vague, 1956). Height has often been ignored as a variable of potential interest in epidemiological studies. However, it can provide unique information on energy balance during the years before adulthood, a time period that may be important in the development of some tumors that occur many years later. For example, in many studies height has been positively associated with risk of breast cancer (Swanson et al., 1988). Furthermore, this information can be valid even in the context of case-control studies because height will usually be unaffected even if illness has caused recent weight loss. Caution is indicated, however, if associations are not found with height because it is possible that, in some populations, few individuals will have been sufficiently deprived of energy intake during development to reduce their longitudinal growth. In such populations, height will primarily reflect genetic factors. Further information on the measurement and interpretation of body dimensions and composition is provided elsewhere (Lohman et al., 1988; Willett, 1998).

Diet and Nutrition

METHODOLOGICAL ISSUES IN NUTRITIONAL EPIDEMIOLOGY Between-Person Variation in Dietary Intake In addition to the availability of a sufficiently precise method for measuring dietary intake, an adequate degree of variation in diet is necessary to conduct observational studies within populations; if no variation in diet exists among persons, no associations can be observed. Some have argued that the diets within populations such as the United States are too homogeneous to study relationships with diet (Goodwin and Boyd, 1987; Hebert and Miller, 1988; Prentice et al., 1988). The true between-person variation in diet is difficult to measure directly, but it cannot be measured by the questionnaires used by epidemiologists because the observed variation will combine true differences with those due to measurement error; more quantitative methods must be used for this purpose. The fat content of the diet varies less among persons than does any other specific nutrient (Beaton et al., 1979); for men in our prospective study (Rimm et al., 1992), the mean fat intake assessed by the mean of two 1-week diet records for those in the top quintile was 40% of calories, whereas for those in the bottom quintile 24% of calories were derived from fat. Although this is not a large range of fat intake and is certainly smaller than the variation among countries, it is of considerable interest because it corresponds closely to the changes recommended by many organizations. Other nutrients vary much more among persons than does total fat intake (Beaton et al., 1979; Willett, 1998). Evidence that measurable and informative variation in diet exists within the United States population is provided by several sources. First, the correlations between food frequency questionnaires and independent assessments of diet found in the validation studies noted above could not have been observed if variation in diet did not exist. For the same reason, the correlations between questionnaire estimates of nutrient intakes and biochemical indicators of intake provide solid evidence of variation. In addition, the ability to find associations between dietary factors and incidence of disease (particularly when based on prospective data) indicates that measurable and biologically relevant variation exists. For example, reproducible relationships have been demonstrated between specific types of dietary fat and cardiovascular disease (Hu et al., 1997; Kushi et al., 1985; Shekelle et al., 1981) and between dietary fiber and risk of coronary heart disease (Prasad et al., 1992) and diabetes (Hu et al., 2001). Although accumulated evidence has indicated that informative variation in diets exists within the United States population and that these differences can be measured, it is important that findings be interpreted in the context of that variation. For example, a lack of association with fat intake within the range of 25% to 40% of energy should not be interpreted to mean that fat intake has no relation with risk of disease under any circumstances. It is possible that the relation is nonlinear and that risk changes at lower levels of fat intake; for example, at 20% of total energy.

Implications of Total Energy Intake Energy balance is likely to have important associations with some cancers; however, this cannot be studied directly by examining the relation of energy intake with risk of cancer because energy intake largely reflects factors other than over- or undereating in relation to requirements (Willett and Stampfer, 1986). The implications of total energy intake can be appreciated by realizing that variation among persons is, to a large degree, secondary to differences in body size and in physical activity. Persons also appear to differ somewhat in metabolic efficiency; inefficient persons require higher energy intake for the same level of function. However, these differences in metabolic efficiency are not practically measurable in epidemiological studies. Because virtually all nutrient intakes tend to be correlated with total energy intake, much of the variation in intake of specific nutrients is secondary to factors that may be unrelated to risk of disease. The effect of extraneous variation is, of course, to increase misclassification of diet and attenuate associations. Although the interrelations of diet and factors that determine variation in total energy intake are complex and

413

beyond the scope of this discussion, failure to adjust for total energy intake may result in lack of significant associations. When total energy intake is related to risk of disease—for example, when physical activity is protective—failure to consider total energy intake in the analysis can be particularly serious because it will confound associations with specific nutrients. The example of coronary heart disease is instructive: Because of the inverse relation with total energy intake, specific nutrients such as saturated fat will also tend to be inversely related to risk. Adjustment for total energy intake is necessary to avoid misleading conclusions; several statistical methods can be used (Willett, 1989a; Willett, 1998). However, the most commonly employed method, division by total energy, which is also called a nutrient density, is not an adequate solution because the inverse of energy intake can then be confounding; this may or may not be a serious problem in any particular study. Appropriate adjustment for energy intake can be a nontrivial issue in some studies; the direction of association can be reversed, such as in the relation between saturated fat intake and myocardial infarction (Gordon et al., 1981) and between fiber intake and risk of colon cancer (Lyon et al., 1987). Unfortunately, total energy intake has been either not measured or, when associated with disease, not appropriately accounted for in many studies on diet and cancer, thus rendering the interpretation unclear. Nutrient intakes adjusted for total energy can be viewed conceptually as measures of nutrient composition rather than as measures of absolute intake. Measures of nutrient composition are most relevant to personal decisions and public health policy because individuals must alter nutrient intakes primarily by manipulating the composition of their diets rather than their total energy intake. This reasoning underlies, for example, the use of fat as a percent of calories in expressing a dietary objective (Committee on Diet Nutrition and Cancer et al., 1982; U. S. Department of Agriculture & U.S. Department of Health and Human Services, 2000).

ASSOCIATIONS OF DIET WITH SPECIFIC CANCERS In this section, several aspects of diet that have been considered in relation to cancer etiology will be noted. This is intended to provide a sense of the scope of the issues and direction to other sources of information, including the chapters in this book relating to specific cancers. A comprehensive, critical review would require a volume in itself.

Energy Balance In animal studies, restriction of energy intake sufficient to reduce growth consistently and substantially reduces the occurrence of many tumors (Ross and Bras, 1971; Weindruch and Walford, 1982). In humans, excessive energy intake, as assessed by indices of obesity, is strongly related to risks of cancers of the endometrium, colon, breast (after menopause), kidney, esophagus (adenocarcinoma), and biliary system (International Agency for Research on Cancer, 2002), and recent evidence suggests that greater adiposity is also associated with death due to leukemia, lymphoma, myeloma, and cancer of the pancreas (Calle et al., 2003). Among premenopausal women, obesity appears to be protective for breast cancer (Huang et al., 1997; Le Marchand et al., 1988), perhaps because heavier women have more anovulatory menstrual cycles, and a high BMI at age 18–20 continues to predict lower risk of breast cancer through the postmenopausal years. Because weight in young women is inversely related to risk of breast cancer, and weight after menopause is positively related to risk, weight gain during midlife has been consistently and strongly associated with postmenopausal breast cancer (International Agency for Research on Cancer, 2002). Height, which in part reflects energy balance during development, is positively related to risk of breast cancer (Swanson et al., 1988; Tretli, 1989; Valaoras et al., 1969; van den Brandt et al., 2000). Height has also been associated with colon cancer (Chute et al., 1991), but the data are limited.

414

PART III: THE CAUSES OF CANCER

Dietary Fat Fat intake has been singled out (Committee on Diet Nutrition and Cancer et al., 1982; National Research Council (U.S.) & Committee on Diet and Health, 1989) as the aspect of diet most importantly related to risk of cancer. The most important sites for which associations have been suggested are cancers of breast, colon, and prostate. Much of the support for these relationships is based on the striking international correlations between per capita fat intake and rates of these malignancies (Armstrong and Doll, 1975; Prentice et al., 1988). Positive associations with breast cancer have been seen in some case-control studies (World Cancer Research Fund & American Institute for Cancer Research, 1997), and in a pooled analysis a highly significant positive association was seen (Howe et al., 1990). However, prospective studies have consistently shown no positive association (Hunter et al., 1996). In an updated pooled analysis of large prospective studies, the RR for a 25 g/day increment in energy-adjusted total fat intake was 1.03 (Smith-Warner et al., 2001a). This discordance between the casecontrol and cohort studies suggests that concerns about recall and selection biases in case-control studies of diet are justified. In a recent analysis limited to young women, intake of animal fat, but not vegetable fat, while premenopausal was associated with greater risk of breast cancer (Cho et al., 2003), suggesting that meat and dairy products may contain constituents other than fat that influence risk of breast cancer. For colon cancer, associations with red meat intake, and processed meat in particular, have been seen frequently (Norat et al., 2002), but fat intake itself does not appear to be related to risk (Howe et al., 1997; World Cancer Research Fund & American Institute for Cancer Research, 1997). Associations have also been reported between animal fat intake and risk of prostate cancer (Giovannucci et al., 1993a; Graham et al., 1983; Heshmat et al., 1985; Kolonel et al., 1988; Le Marchand et al., 1994; World Cancer Research Fund & American Institute for Cancer Research, 1997), but there is little evidence that vegetable fat increases risk. This suggests that fat per se is not an etiologic factor for prostate cancer, although the prospective studies are few. Overall, the more powerful and less biased studies have not supported an association between fat intake during midlife and cancer incidence. The effects of diet during earlier periods of life remain to be explored.

Fiber Interest in the relation between fiber intake and colon cancer is largely the result of Burkitt’s observation of low rates of colon cancer in areas of Africa where fiber consumption and stool bulk were high (Burkitt, 1971). Fiber has been hypothesized to dilute potential carcinogens and speed their transit through the colon. Inverse associations with total fiber intake have been seen in most case-control studies (Howe et al., 1992; Trock et al., 1990). When specific sources of fiber have been examined, fiber intake from fruits or vegetables have been most consistently associated with lower incidence, whereas fiber from cereals has either been related to an increased risk or not associated with colon cancer (Willett, 1989b). However, dietary fiber has not been associated with risk of colon cancer in most prospective studies (Fuchs et al., 1999; Terry et al., 2001). A modest inverse association was observed in a multicenter cohort study in Europe (Bingham et al., 2003), but the confounding by intake of folic acid (see below) was not considered, which is plausible because whole grains, fruits, and vegetables are the primary dietary sources of both nutrients.

Folic Acid Folic acid deficiency has long been known to cause tumors in animals, possibly by influencing gene expression through DNA methylation or by increasing the incorporation of uracil in DNA (Blount and Ames, 1994). Considerable evidence from both case-control and cohort studies supports an inverse association between folate intake and risk of colon cancer (Giovannucci, 2002a), and this association appears stronger among regular alcohol consumers (Giovannucci et al., 1995).

An association between a functional polymorphism in the folic acid metabolizing gene, methylene tetrahydrofolate reductase, and incidence of colon cancer adds support for a causal relationship. An inverse association between folate intake or blood level and risk of breast cancer among regular alcohol consumers has also been reported in multiple studies (Rohan et al., 2000; Zhang et al., 1999; Zhang et al., 2003).

Preformed Vitamin A and Carotenoids Because vitamin A plays a central role in regulating cell differentiation, reason exists to suspect a relation to cancer incidence. In some animal models, preformed vitamin A or chemical analogues can inhibit tumor development, even when administered well after the carcinogen (Lippman and Meyskens, 1989). Vegetable precursors of vitamin A, the carotenoids, have been less studied in animals but have reduced tumor incidence in some models, particularly skin cancers (MathewsRoth, 1989). Whether carotenoids act by virtue of conversion to retinol, the main circulating form of vitamin A with physiologic activity, or by other mechanisms, such as being an antioxidant or quencher of singlet oxygen (Peto et al., 1981), remains uncertain. Epidemiological studies of preformed vitamin A or carotenoids in relation to cancer incidence have used either questionnaire assessments of diet or blood measurements. The distinction between these two forms of vitamin A has practical implications because preformed vitamin A is obtained only from foods derived from animal sources or vitamin supplements, whereas carotenoid precursors of vitamin A are obtained almost entirely from plant sources. The initial reports of an inverse relation between total vitamin A intake and risk of lung cancer (Bjelke, 1975; Mettlin et al., 1979) did not clearly distinguish between these sources. In a subsequent prospective study by Shekelle and coworkers (Shekelle et al., 1981), an apparent protective effect of total vitamin A for lung cancer was found to be entirely attributable to carotenoid sources; preformed vitamin A was unrelated to risk of this disease. Subsequent case-control studies based on dietary intake (Gregor et al., 1980; Hinds et al., 1984; Samet et al., 1985; Wu et al., 1985; Ziegler et al., 1984) and beta-carotene measurements in prospectively collected blood (Menkes et al., 1986; Nomura et al., 1985; Stahelin et al., 1984) provided evidence for a protective relationship between carotenoid intake and risk of lung cancer after controlling for cigarette smoking but little support for any relationship with preformed vitamin A (Hunter and Willett, 1994; Willett, 1990). Large cohort studies have been less supportive of a protective role for beta-carotene and risk of lung cancer; little overall relationship has been seen, although a small increase in risk at very low intakes could not be excluded (Michaud et al., 2000). Neither preformed vitamin A nor carotenoid intake has been consistently associated with risk of colon cancer, although these relationships have been examined in several studies. Randomized trials using supplements of beta-carotene have not supported a protective effect. In a randomized study among patients with previous skin cancer, beta-carotene did not prevent the development of new skin cancer (Greenberg et al., 1990). Similarly, beta-carotene— as well as vitamins C and E—did not influence the recurrences of adenomatous colon polyps (Greenberg et al., 1994). In the large Finnish trial among men at high risk of lung cancer (Group, 1994), a statistically significant 18% increase in incidence of this malignancy was seen in those randomized to beta-carotene, and in a trial among persons at high risk for lung cancer, a combined beta-carotene/preformed vitamin A supplement also increased risk by 28% (Omenn et al., 1996). In the Physicians’ Health Study, no effect of beta-carotene supplements on lung or overall cancer incidence was seen (Hennekens et al., 1996). Many explanations have been offered for the lack of benefit (or even harm) in the randomized trials, including interference with the metabolism of other beneficial carotenoids, the relatively high dose of beta-carotene in the supplements, or induction of carcinogenactivating enzymes. Also, the lack of benefit from beta-carotene supplements does not exclude the possibility that other protective nutrients or biologically active constituents in fruits and vegetables could be protective. However, the premise that greater consumption

Diet and Nutrition of fruits and vegetables is protective has not been supported by the results of large cohort studies of diet and lung cancer; apart from a possible small increase in risk with very low intake, the dose-response relationship is flat (Smith-Warner et al., 2003). Thus, the basis for the enthusiasm for finding a powerful chemopreventive factor in fruits and vegetables appears to have been flawed. The findings from casecontrol studies could have been due to biases discussed earlier, and the studies based on beta-carotene levels in prospectively collected blood could have resulted from incomplete control of confounding by smoking because smoking reduces beta-carotene levels (Willett, 1998). Although the best epidemiologic evidence and randomized trials both indicate no benefit of high beta-carotene intake for lung cancer, it is sobering to consider the resources that have been invested in this hypothesis.

Vitamin C Vitamin C has been hypothesized to reduce cancer risk by its antioxidant properties, by blocking the conversion of nitrates and nitrogencontaining compounds to carcinogens under conditions found in the stomach (Mirvish et al., 1972) and in food (Raineri and Weisburger, 1975), and by other mechanisms (Cameron et al., 1979). In casecontrol studies, evidence for a protective effect of vitamin C has been seen for laryngeal cancer (Graham et al., 1981), oral cancer (Winn et al., 1984), esophageal cancer (Mettlin et al., 1981; Ziegler et al., 1981), stomach cancer (Correa et al., 1985), and cervical dysplasia (Romney et al., 1985; Wassertheil-Smoller et al., 1981). As discussed elsewhere (Block and Menkes, 1989), the interpretation of these data, while supporting a preventive effect of vitamin C, is also compatible with the possibility that other factors in fruits and vegetables are the primary cause of reduced cancer risk. However, given the experience with betacarotene and lung cancer, supportive data from large cohort studies or randomized trials would be needed to be confident of these relationships.

Vitamin E The concept that vitamin E might reduce risk of human cancer derives from its role as a potent intracellular antioxidant and from suggestive animal studies (Bieri et al., 1983; Mersens et al., 1989). Wald reported an extremely strong inverse relation between prediagnostic blood levels of vitamin E and risk of breast cancer (Wald et al., 1984); however, this was later found to be an artifact of differential handling of case and control specimens (Wald et al., 1988). In a separate study population, Wald reported a protective association between prediagnostic blood vitamin E levels and cancer at all sites combined but provided evidence that this was likely due to an effect of preclinical disease on vitamin E levels rather than a preventive effect of this micronutrient (Wald et al., 1987). Also using blood collected from a large cohort, Menkes and co-workers (Menkes et al., 1986) reported an inverse association with lung cancer; this was not confirmed in a similar study by Nomura and co-workers (Nomura et al., 1985). Although associations have generally not been seen for other cancer sites (Nomura et al., 1985), an inverse association was reported in a Finnish cohort between prediagnostic serum vitamin E levels and cancers at all sites combined, which was particularly strong for nonsmoking-related cancers (Knekt et al., 1988). In the recent Finnish trial (Group, 1994), vitamin E supplementation did not influence risk of lung cancer and was associated with a lower risk of prostate cancer, but no relation was seen in the United States cohort study (Chan et al., 1999). The overall data thus remain unclear at this time; this may be the result of only modest numbers of specific cancers in the studies that have been reported, but a large beneficial effect of vitamin E against cancer seems unlikely.

Selenium That higher intake of selenium might reduce the risk of some human cancers has been suggested by numerous animal experiments (Combs and Combs, 1986) and ecological studies in which indices of selenium

415

intake have been inversely related to risk of cancer, both internationally (Schrauzer et al., 1977) and within the United States (Shamberger et al., 1976). The correlational data have been particularly strong for cancers of the colon and breast. Inverse associations have been seen between serum selenium levels and subsequent risk of all cancers combined in several large prospective studies (Kok et al., 1987; Salonen et al., 1984; Salonen et al., 1985; Willett et al., 1983), but no evidence of a protective effect has been found in other similar investigations (Coates et al., 1988; Menkes et al., 1986; Nomura et al., 1987; Peleg et al., 1985; Virtamo et al., 1987). In two prospective studies that together included a large number of breast cancer cases, no relation was found between selenium in nails and subsequent risk of this malignancy (Hunter et al., 1990; van Noord et al., 1987). A strong inverse association between nail selenium levels and risk of lung cancer was seen in a prospective study from Holland (van den Brandt et al., 1993), and an inverse association between nail selenium levels and risk of prostate cancer was seen in U.S. men (Yoshizawa et al., 1998). In a randomized trial of selenium supplementation for prevention of skin cancer, no benefit was seen for this end point (Clark et al., 1996). Although based on a small number of cases, large reductions in the incidence of prostate and colon cancer were seen. A randomized trial is now ongoing, using a 2 ¥ 2 factorial design that will evaluate the chemopreventive effects of selenium and vitamin E (Klein, 2003).

Calcium Higher calcium intake has reduced bowel tumors in animals (Lamprecht and Lipkin, 2001), possibly by precipitation of bile acids that are thought to promote tumorigenesis. In a randomized trial, calcium supplementation reduced recurrence by about 20% (Baron et al., 1999). In large prospective studies, high calcium intake has consistently been associated with a modest and nonsignificantly lower risk of colon cancer. In a pooled analysis of cohort studies, those with the highest calcium intake had a 20% lower risk of colon cancer that was statistically significant (Cho et al., 2004). Thus, the convergence of many lines of evidence support a modest benefit of higher calcium intake in colon cancer prevention. However, the practical implications are currently unclear because high calcium intake or dairy product consumption has been associated with higher risk of prostate cancer in many studies (Chan and Giovannucci, 2001; Giovannucci, 2002b).

Salt Ecological associations have led to the suggestion that high salt intake—salt being traditionally used in many societies for the preservation of food—might increase the risk of gastric cancer (Howson et al., 1986; Joossens and Geboers, 1987). Salt is hypothesized to act as a local irritant that may compromise the gastric mucosal barrier and thus facilitate the action of local carcinogens. Positive associations of salt intake with risk of gastric cancer have been seen in several casecontrol studies (Fontham et al., 1986; Haenszel et al., 1972; You et al., 1988). This hypothesis is compatible with the striking decline in gastric cancer in most industrialized countries over this century as refrigeration reduced the need of salt for preservation. However, economic improvements also enhanced the supply of fresh fruits and vegetables on a year-round basis and improved hygiene, which may have reduced the transmission of Helicobacter infection. A strong association between salted fish intake during childhood and risk of nasopharyngeal cancer has been noted among several Chinese populations (You et al., 1988).

Fruits and Vegetables Higher intake of fruits and vegetables has been widely believed to reduce risk of many specific cancers (Block et al., 1992; World Cancer Research Fund & American Institute for Cancer Research, 1997), and an increase in consumption has been a key element of most dietary recommendations and strategies to prevent cancer. However, several factors may have led to an overstatement of the potential benefits. First, most of the studies supporting such conclusions have been case-

416

PART III: THE CAUSES OF CANCER

control investigations and are thus susceptible to bias. Reporting of evidence may have been selective, in part because several dozen fruit and vegetable items are typically included in dietary questionnaires, and publication of all associations is not feasible. Thus, “positive” findings supporting prior beliefs may be more likely to be published than null or contrary findings. Also, summaries of evidence have sometimes counted an inverse association with one or a few foods as support for a benefit of fruits and vegetables when this may have simply been the result of chance due to multiple comparisons. These potential biases are greatly reduced if large prospective studies are examined systematically; recent findings from such analyses provide a much more tempered view of the benefits to be derived from an overall increase in fruit and vegetable consumption. As described above, a pooled analysis of large prospective studies of lung cancer, which included more than 3000 incident cases in men and women, did not find an overall relationship with consumption of fruits and vegetables, although a small increase in risk in the lowest category of intake was observed (Smith-Warner et al., 2003). A similar lack of overall association with fruit and vegetable consumption was seen in the pooled analysis of breast cancer (7377 cases among 351,825 women) (Smith-Warner et al., 2001b). In the pooled analysis from the Nurses’ Health Study and Health Professionals’ Follow-up Study, no association was seen with total fruit and vegetable consumption and risk of colon cancer (Michels et al., 2000). Also, in a combined analysis of these two cohorts, no association was seen between overall consumption of fruits and vegetables and total cancer incidence (Hung et al., 2004). In this analysis, a statistically significant inverse association was seen between overall fruit and vegetable intake and risk of cardiovascular disease, which indicates that biologically important variation in consumption did exist in these cohorts. Thus, more recent evidence that is less susceptible to methodological bias does not support a major benefit from increasing fruit and vegetable consumption for reducing cancer risk. Despite the largely null findings for overall fruit and vegetable intake, the possibility remains that higher intakes of specific fruits and vegetables, or specific substances in fruits and vegetables, may reduce the risk of specific cancers. For example, evidence noted above suggests that higher intake of folic acid can reduce risk of colon cancer. However, if folic acid is the primary active factor in fruits and vegetables, fruit and vegetable consumption would represent a seriously misclassified measure of folic acid intake because many of these foods contain only small amounts of this vitamin. Also, fortification and supplementation provide a large proportion of the folic acid in the United States food supply, and this would tend to mask the contribution from fruits and vegetables. As another example, higher intake of lycopene, the carotenoid responsible for the red color of tomato products, has been associated with lower risk of prostate cancer in several studies (Giovannucci, 2002c), even though overall fruit and vegetable consumption was not associated with risk of this cancer. It is also possible that biologically active substances in fruits and vegetables exist but in doses that have effects too small to be detected in epidemiological studies, and that purified amounts administered pharmacologically will reduce cancer. Despite these caveats, the best available evidence suggests that the recommendation to increase consumption of fruits and vegetables is good general advice but that this cannot be relied upon to reduce risk of cancer.

Natural Carcinogens and Products of Food Preparation Ames (Ames et al., 1987) has reviewed the large number of naturally occurring compounds in foods that are known to be mutagenic or carcinogenic in laboratory settings. Many of these toxic compounds have probably been developed by plants during evolution as a form of protection. In addition, a wide variety of mutagens and carcinogens are formed during the process of cooking meat and other foods, even under conditions that do not involve charring (Sugimura, 1986). With few exceptions—such as the bracken fern, which may cause bladder cancer when used regularly as a tea (Pamukcu et al., 1970)—little evidence exists that these compounds contribute substantially to human

cancer. However, the role of carcinogens that occur naturally or that are formed during cooking has not been studied adequately with respect to human cancer and deserves further investigation.

PREVENTION OF CANCER BY DIETARY MEANS Intense interest in the possible prevention of cancer by dietary modification has generated a series of recommendations from numerous private and governmental bodies worldwide and within the United States (Committee on Diet Nutrition and Cancer et al., 1982; Greenwald and Sondick, 1986; World Cancer Research Fund & American Institute for Cancer Research, 1997). Those of the National Academy of Sciences (Committee on Diet Nutrition and Cancer et al., 1982) are typical: 1. The consumption of both saturated and unsaturated fats should be reduced, from an average of approximately 40% of total calories to 30%. 2. The intake of fruits, vegetables, and whole grain cereal products in the diet should be emphasized. 3. The consumption of food preserved by salt-curing (including saltpickling) or smoking should be minimized. 4. Efforts should continue to be made to minimize contamination of foods with carcinogens from any source, including those that are natural or occurring inadvertently during production, processing, and storage. 5. Further efforts should be made to identify mutagens in food and to expedite testing for their carcinogenicity. Where feasible and prudent, mutagens should be removed or minimized. 6. If alcoholic beverages are consumed, this should be done in moderation. The potential impact of these changes in diet is difficult to quantify. Based on similar recommendations, the U.S. National Cancer Institute suggested that within 10 years, cancer of the colon and rectum would be reduced by 50%; cancer of the breast by 25%; cancers of the prostate, endometrium, and gallbladder by 15%; and the cancers of the stomach, esophagus, pancreas, ovaries, liver, lung, and bladder by a possible but not precisely quantifiable amount (Greenwald and Sondick, 1986). In this same publication, the large uncertainty of these estimates was noted, including the possibility that no reductions for some of these cancers might occur. Dietary recommendations for the United States and other Western countries have primarily focused on ways to reduce cancers of the breast, large bowel, and prostate, and to a lesser extent pancreas and ovary because these are the most important cancers not caused by smoking or alcohol that are plausibly related to diet. In other parts of the world different cancer sites, such as stomach and esophagus, are dominant. These upper gastrointestinal cancers in particular are likely influenced by dietary factors; although uncertainty exists about the specific alterations of diet that will be effective, a change toward the contemporary United States diet is likely to be beneficial. However, a more precise understanding of etiologic factors is needed to avoid exchanging one pattern of cancer for another that includes high rates of breast and colon cancer. The actual benefit in cancer reduction that might be realized by dietary change is, of course, a function of both the biological relationships and the success of an intervention. The most certain effects of diet on cancer incidence are those related to total energy balance, especially with cancers of the (postmenopausal) breast, colon, kidney, and endometrium. In the United States, 20% of cancer mortality in women and 15% of cancer mortality in men are estimated to be attributable to overweight and obesity (Calle et al., 2003). In addition to reducing incidence, avoidance of midlife weight gain is likely to contribute to lower breast cancer mortality, in part due to a lower casefatality rate among lean women. Although reduction in obesity among premenopausal women could lead to an increase in breast cancer incidence, breast cancer mortality in this group is not related to body weight due to delayed diagnosis among overweight women (Tretli, 1989; Willett et al., 1985a). Thus, weight control deserves to be high

Diet and Nutrition on the agenda for cancer prevention, not far behind efforts to reduce cigarette smoking. Weight control programs need to be inextricably intertwined with promotion of physical activity, which also can reduce risk of several cancers independent of its effect on body weight (International Agency for Research on Cancer, 2002). Because rates of overweight and obesity are rapidly increasing in the United States and many countries (Flegal et al., 2002; Seidell, 2002), intensified weight control efforts are needed to avoid further increase in cancer incidence. Although the overall trends are discouraging, most motivated individuals can control their weight; this is manifested by the 50% lower prevalence of obesity among those with higher education compared to groups with low educational levels (Flegal et al., 2002). Epidemiologic data also provide strong evidence that reduction in alcohol consumption will reduce cancers of the upper gastrointestinal tract, liver, colon, and breast (International Agency for Research on Cancer, 1988; Smith-Warner et al., 1998). Although heavy consumption (more than two alcoholic drinks per day) has long been known to be associated with cancer incidence, modest increases in risk of breast cancer are seen with 1–2 drinks per day. Considerable effort has been given to increasing fruit and vegetable consumption with the expectation that this could have a major impact on cancer incidence. This has been the rationale for the “Five-a-Day” program of the U.S. National Cancer Institute. However, the likely benefits appear to have been seriously overstated, largely due to biases in case-control studies. Whether any relationship exists with overall cancer incidence in the United States is unclear. Association may exist between intakes of specific foods and specific cancers that are obscured by combining all fruits and vegetables, and some evidence suggests that benefits exist for cancers of the esophagus and stomach, which are uncommon in the United States but extremely important elsewhere. Also, some studies suggest that only very low intakes of fruits and vegetables are associated with excess risk, and that increasing intake beyond about three servings/day may have little further effect. Specific components of fruits and vegetables may be protective, as folic acid appears to be for colon cancer, but overall fruit and vegetable intake represents a poor indicator of intake. Despite the diminished prospects for reducing cancer incidence by increasing fruit and vegetable consumption, evidence is strong that this will reduce risk of coronary heart disease, the most important cause of death in the United States. The primary focus of most dietary recommendations thus far has been on intake of dietary fat; it has been suggested that intake should be reduced to an average of about 30% of energy (Committee on Diet Nutrition and Cancer et al., 1982; National Research Council (U.S.) & Committee on Diet and Health, 1989; World Cancer Research Fund & American Institute for Cancer Research, 1997). Anticipated benefits have related primarily to cancers of breast and colon and possibly prostate, endometrium, and ovary. The recommendation to reduce fat intake has often been in part justified on the basis of being harmless and also being likely to result in a reduced risk of coronary heart disease. However, this recommendation is overly simplistic and not without potential harm. Total fat intake is not related to risk of coronary heart disease (Hu and Willett, 2002; National Research Council (U.S.) & Committee on Diet and Health, 1989); and unsaturated fats, contained primarily in vegetable oils, are likely to reduce risk of coronary heart disease, whereas saturated and trans fats are likely to increase the risk (Hu and Willett, 2002). The relation of type of fat to risk of human cancer remains unclear; what evidence there is suggests that any positive association pertains mainly to fat from animal sources. Without a careful, selective reduction in the type of fat, it is quite conceivable that some individuals might reduce their intake of vegetable fat and thereby increase their risk of cardiovascular disease. Rather than focusing on fat intake, a reduction in red meat and dairy fat consumption is better supported; associations with red meat have been seen with cancers of the colon and prostate in multiple studies. Whether these relationships are due to the fat content of meat is uncertain, and some data for colon cancer suggest other components contribute to risk. The level of scientific certainty that is appropriate for launching public health interventions aimed at changing diets to prevent cancer

417

is an issue of legitimate debate. As has been pointed out elsewhere (Greenwald and Sondick, 1986), waiting to change dietary behavior until the scientific evidence is almost certain could mean the loss of thousands of lives. On the other hand, promulgating policies that turn out to be wrong has costs in terms of diverting attention and resources from interventions that are effective, such as smoking cessation and mammography, and in the loss of credibility, which is essential for any public health program. Fortunately, the dietary recommendations noted above, with the qualification regarding type of fat, are generally consistent with those designed to reduce the incidence of coronary heart disease (Consensus Conference, 1985; Hu and Willett, 2002), and for which the evidence is more compelling. Specific actions for implementing dietary recommendations have been developed by the U.S. National Cancer Institute (Greenwald and Sondick, 1986). These include encouraging federal agencies and industries to include cancer prevention dietary recommendations in federal food production, marketing, and distribution policies; encouraging the production of leaner meat products and lower fat content of dairy products; informing the public about the relationship between diet and cancer; expanding nutrition labeling to cover the full range of mass-marketed foods so consumers can be better informed and make wiser shopping decisions; and developing diet and cancer programs that make use of the mass media and other high-technology communication approaches. State and local agencies are also encouraged to review school curricula to reflect newer knowledge of diet and cancer risks and strategies for risk reduction; to review school menus in relation to the cancer control objectives; to promote diet and cancer information programs; to encourage restaurants to provide sufficient information to allow patrons to choose nutritious foods; to coordinate governmental planning activities to ensure that attention is given to reducing dietary risk factors for cancer; to promote dissemination of information about proper food selection to protect against cancer risk; and to include information on diet and cancer in existing food, nutrition, and health programs with the use of innovative approaches to reach high-risk groups. These approaches remain valid, but the emphasis will need to be shifted from reducing total fat intake if there is to be an appreciable impact on cancer incidence. Because of the importance of weight control, excessive energy intake from all sources and increasing physical activity will need to be emphasized. To accomplish this, health care providers will need to be more engaged in weight counseling, physical education programs will need to be increased and improved, and the physical infrastructure of America will need to be modified to encourage walking and bicycle riding for transportation and a wide range of enjoyable forms of recreational activity.

SUMMARY A wide variety of evidence based on comparisons of cancer rates in different geographic areas, migrating populations, religious orders, and rapid changes in rates over time strongly suggests that the high rates of breast, colon, and other important cancers in the affluent countries are due to environmental rather than genetic factors. Furthermore, aspects of diet are likely to be important etiological factors for some of these cancers. Similarly, the incidence rates of many cancers that are still of great importance in other parts of the world, such as those of the oral cavity, esophagus, and stomach, are also likely to be influenced by dietary factors. Despite the substantial evidence that dietary factors are likely to be important, the specific aspects of food and nutrient intake that are either causative or preventive remain inconclusively defined for most of these cancers. The most consistent evidence is that excessive energy intake in relation to level of physical activity increases the risk of many of these cancers and that this is the second most important avoidable cause of cancer mortality in many countries after cigarette smoking. Also, convincing evidence indicates that excessive alcohol consumption increases the risk of several important cancers, including those of the upper gastrointestinal tract, liver, colon, and breast. Substantial evidence, although not entirely consistent, indicates that red meat or animal fat intake is associated with risks of colon

418

PART III: THE CAUSES OF CANCER

and prostate cancers. Evidence that overall increases in fruit and vegetable consumption will reduce cancer incidence appreciably has become much weaker than believed earlier, although modest benefits for some specific cancers cannot be excluded. Better evidence supports a protective effect of folic acid against risk of colon cancer, but this is more reliably achieved by taking a multiple vitamin than by increasing fruit and vegetable intake. Individual and policy decisions regarding dietary changes should consider not only the possible benefits for cancer reduction but also the effects on coronary heart disease, because this remains the dominant cause of mortality in the United States and its relation with diet is better established. References Adelstein AM, Staszewski J, Muir CS. 1979. Cancer mortality in 1970–1972 among Polish–born migrants to England and Wales. Br J Cancer 40:464–475. Ames BN, Magaw R, Gold LS. 1987. Ranking possible carcinogenic hazards. Science 236:271–280. Armstrong B, Doll R. 1975. Environmental factors and cancer incidence and mortality in different countries, with special reference to dietary practices. Int J Cancer 15:617–631. Baron JA, Beach M, Mandel JS, van Stolk RU, Haile RW, et al. 1999. Calcium supplements for the prevention of colorectal adenomas. The Calcium Polyp Prevention Study Group. N Engl J Med 340:101–107. Beaton GH, Milner J, Corey P, McGuire V, Cousins M, et al. 1979. Sources of variance in 24-hour dietary recall data: implications for nutrition study design and interpretation. Am J Clin Nutr 32:2546–2549. Bieri JG, Corash L, Hubbard VS. 1983. Medical uses of vitamin E. N Engl J Med 308:1063–1071. Bingham SA, Cummings JH. 1985. Urine nitrogen as an independent validatory measure of dietary intake: A study of nitrogen balance in individuals consuming their normal diet. Am J Clin Nutr 42:1276–1289. Bingham SA, Day NE, Luben R, Ferrari P, Slimani N, et al. 2003. Dietary fibre in food and protection against colorectal cancer in the European Prospective Investigation into Cancer and Nutrition (EPIC): An observational study. Lancet 361:1496–1501. Bjarnason O, Day N, Snaedal G, Tulinius H. 1974. The effect of year of birth on the breast cancer age–incidence curve in Iceland. Int J Cancer 13:689–696. Bjelke E. 1975. Dietary vitamin A and human lung cancer. Int J Cancer 15:561–565. Bjorntorp P. 1987. Classification of obese patients and complications related to the distribution of surplus fat. Am J Clin Nutr 45(s):1120–1125. Block G, Hartman AM, Dresser CM, Carroll MD, Gannon J, Gardner L. 1986. A data–based approach to diet questionnaire design and testing. Am J Epidemiol 124:453–469. Block G, Menkes M. 1989. Ascorbic acid in cancer prevention. In: Moon TE, Micozzi MS, eds. Nutrition and Cancer Prevention. New York: Marcel Dekker, Inc., pp. 341–388. Block G, Patterson B, Subar A. 1992. Fruit, vegetables, and cancer prevention: A review of the epidemiological evidence. Nutr Cancer 18:1–29. Block G, Woods M, Potosky A, Clifford C. 1990. Validation of a self– administered diet history questionnaire using multiple diet records. J Clin Epidemiol 43:1327–1335. Blount BC, Ames BN. 1994. Analysis of uracil in DNA by gas chromatography–mass spectrometry. Anal Biochem 219:195–200. Buell P. 1973. Changing incidence of breast cancer in Japanese–American women. J Natl Cancer Inst 51:1479–1483. Burke BS. 1947. The dietary history as a tool in research. J Am Diet Assoc 23:1041–1046. Burkitt DP. 1971. Epidemiology of cancer of the colon and rectum. Cancer 28:3–13. Byar DP, Gail MH. 1989. Errors-in-variables workshop. Statist Med 8:1027– 1029. Byers T, Marshall J, Anthony E, Fiedler R, Zielezny M. 1987. The reliability of dietary history from the distant past. Am J Epidemiol 125:999–1011. Byers TE, Rosenthal RI, Marshall JR, Rzepka TF, Cummings KM, Graham S. 1983. Dietary history from the distant past: A methodological study. Nutr Cancer 5:69–77. Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ. 2003. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med 348:1625–1638. Cameron E, Pauling L, Leibovitz B. 1979. Ascorbic acid and cancer: A review. Cancer Res 39:663–681. Chan JM, Giovannucci E. 2001. Dairy products, calcium, and vitamin D and risk of prostate cancer. Epidemiol Rev 23:87–92.

Chan JM, Stampfer MJ, Ma J, Rimm EB, Willett WC, Giovannucci EL. 1999. Supplemental vitamin E intake and prostate cancer risk in a large cohort of men in the United States. Cancer Epidemiol Biomarkers Prev 8:893– 899. Chen J, Campbell TC, Junyao L, Peto R. 1990. Diet, Life-Style, and Mortality in China: A Study of the Characteristics of 65 Chinese Counties. Oxford, England: Oxford University Press. Cho E, Smith–Warner S, Spiegelman D, Beeson W, van den Brandt P, et al. 2004. Dairy foods and calcium and colorectal cancer: A pooled analysis of 10 cohort studies (in press). Cho E, Spiegelman D, Hunter DJ, Chen WY, Stampfer MJ, et al. 2003. Premenopausal fat intake and risk of breast cancer. J Natl Cancer Inst 95:1079–1085. Chute CG, Willett WC, Colditz GA, Stampfer MJ, Baron JA, et al. 1991. A prospective study of body mass, height, and smoking on the risk of colorectal cancer in women. Cancer Causes Control 2:117–124. Clark LC, Combs GF, Jr., Turnbull BW, Slate EH, Chalker DK, et al. 1996. Effects of selenium supplementation for cancer prevention in patients with carcinoma of the skin. A randomized controlled trial. Nutritional Prevention of Cancer Study Group. JAMA 276:1957–1963. Coates RJ, Weiss NS, Daling JR, Morris JS, Labbe RF. 1988. Serum levels of selenium and retinol and the subsequent risk of cancer. Am J Epidemiol 128:515–523. Cohen NL, Lans MJ, Ferris AM, et al. 1990. The contributions of portion size data to estimating nutrient intakes of food frequency questionnaires. Rep. Research Bulletin 730, Massachusetts Agriculture Experiment Station, University of Massachusetts, Amherst, MA. Combs GF, Jr, Combs SB. 1986. Selenium and cancer. In: The Role of Selenium in Nutrition. New York: Academic Press, pp. 413. Committee on Diet Nutrition and Cancer, Assembly of Life Sciences, National Research Council. 1982. Diet, Nutrition, and Cancer. Washington, DC: National Academy Press. Consensus Conference. 1985. Lowering blood cholesterol to prevent heart disease. J Am Med Assoc 253:2080–2086. Correa P, Fontham E, Pickle LW, Chen V, Lin Y, Haenszel W. 1985. Dietary determinants of gastric cancer in south Louisiana inhabitants. J Natl Cancer Inst 75:645–654. Cummings SR, Block G, McHenry K, Baron RB. 1987. Evaluation of two food frequency methods of measuring dietary calcium intake. Am J Epidemiol 126:796–802. Doll R, Peto R. 1981. The causes of cancer: Quantitative estimates of avoidable risks of cancer in the United States today. J Natl Cancer Inst 66:1191–1308. Flegal KM, Carroll MD, Ogden CL, Johnson CL. 2002. Prevalence and trends in obesity among US adults, 1999–2000. JAMA 288:1723–1727. Fontham E, Zavala D, Correa P, Rodriguez E, Hunter F, et al. 1986. Diet and chronic atrophic gastritis: A case–control study. J Natl Cancer Inst 76:621–627. Friedenreich CM, Howe GR, Miller AB. 1991. An investigation of recall bias in the reporting of past food intake among breast cancer cases and controls. Ann Epidemiol 1:439–453. Fuchs CS, Colditz GA, Stampfer MJ, Speizer FE, Giovannucci E, et al. 1999. Dietary fiber and the risk of colorectal cancer and adenoma in women. N Engl J Med 340:169–176. Giovannucci E. 2002a. Epidemiologic studies of folate and colorectal neoplasia: A review. J Nutr 132:2350S–2355S. Giovannucci E. 2002b. Nutritional and environmental epidemiology of prostate cancer. In: Kantoff PW, Carroll PR, D’Amico AV, eds. Prostate Cancer: Principles and Practice. Philadelphia: Lippincott Williams & Wilkins, pp. 117–139. Giovannucci E. 2002c. A review of epidemiologic studies of tomatoes, lycopene, and prostate cancer. Exp Biol Med (Maywood) 227:852–859. Giovannucci E, Rimm EB, Ascherio A, Stampfer MJ, Colditz GA, Willett WC. 1995. Alcohol, low-methionine–low-folate diets, and risk of colon cancer in men. J Natl Cancer Inst 87:265–273. Giovannucci E, Rimm EB, Colditz GA, Stampfer MJ, Ascherio A, et al. 1993a. A prospective study of dietary fat and risk of prostate cancer. J Natl Cancer Inst 85:1571–1579. Giovannucci E, Stampfer MJ, Colditz GA, Manson JE, Rosner BA, et al. 1993b. A comparison of prospective and retrospective assessments of diet in the study of breast cancer. Am J Epidemiol 137:502–511. Goldbohm RA, van den Brandt PA, Brants HAM, van’t Veer P, Al M, et al. 1994. Validation of a dietary questionnaire used in a large–scale prospective cohort study on diet and cancer. Eur J Clin Nutr 48:253– 265. Goldin BR, Adlercreutz H, Dwyer JT, Swenson L, Warram JH, Gorbach SL. 1981. Effect of diet on excretion of estrogens in pre– and postmenopausal women. Cancer Res 41:3771–3773.

Diet and Nutrition Goodwin PJ, Boyd NF. 1987. Critical appraisal of the evidence that dietary fat intake is related to breast cancer risk in humans. J Natl Cancer Inst 79:473–485. Gordon T, Kagan A, Garcia–Palmieri M, Kannel WB, Zukel WJ, et al. 1981. Diet and its relation to coronary heart disease and death in three populations. Circulation 63:500–515. Graham S, Dayal H, Swanson M, Mittelman A, Wilkinson G. 1978. Diet in the epidemiology of cancer of the colon and rectum. J Natl Cancer Inst 61:709–714. Graham S, Haughey B, Marshall J, Priore R, Byers T, et al. 1983. Diet in the epidemiology of carcinoma of the prostate gland. J Natl Cancer Inst 70:687–692. Graham S, Mettlin C, Marshall J, Priore R, Rzepka T, Shedd D. 1981. Dietary factors in the epidemiology of cancer of the larynx. Am J Epidemiol 113:675–680. Greenberg E, Baron J, Stukel T, Stevens M, Mandel J, et al. 1990. A clinical trial of beta carotene to prevent basal-cell and squamous-cell cancers of the skin. The Skin Cancer Prevention Study Group. N Engl J Med 323:789–795. Greenberg ER, Baron JA, Tosteson TD, Freeman DH, Beck GJ, et al. 1994. A clinical trial of antioxidant vitamins to prevent colorectal adenoma. N Engl J Med 331:141–147. Greenwald P, Sondick EJ, eds. 1986. Cancer Control Objectives for the Nation: 1985–2000, Vols. 2. NCI Monograph, U.S. Dept HHS, NIH Pub. No. 86–2880. Gregor A, Lee PN, Roe FJC, et al. 1980. Comparison of dietary histories in lung cancer cases and controls with special reference to vitamin A. Nutr Cancer 2:93–97. Group TA–TBCCPS. 1994. The effect of Vitamin E and beta carotene on the incidence of lung cancer and other cancers in male smokers. N Engl J Med 330:1029–1035. Haenszel W, Kurihara M, Segi M, Lee RKC. 1972. Stomach cancer among Japanese in Hawaii. J Natl Cancer Inst 49:969–988. Hankin JH, Nomura AM, Lee J, Hirohata T, Kolonel LN. 1983. Reproducibility of a dietary history questionnaire in a case–control study of breast cancer. Am J Clin Nutr 37:981–985. Hartge P, Brinton LA, Rosenthal JFA, Cahill JI, Hoover RN, Waksberg J. 1984. Random digit dialing in selecting a population-based control group. Am J Epidemiol 120:825–833. Hartz AJ, Rupley DC, Rimm AA. 1984. The association of girth measurements with disease in 32,856 women. Am J Epidemiol 119:71–80. Heady JA. 1961. Diets of bank clerks. Development of a method of classifying the diets of individuals for use in epidemiologic studies. J R Stat Soc 124:336–361. Hebert JR, Miller DR. 1988. Methodologic considerations for investigating the diet-cancer link. Am J Clin Nutr 47:1068–1077. Hennekens CH, Buring JE, Manson JE, Stampfer M, Rosner B, et al. 1996. Lack of effect of long-term supplementation with beta carotene on the incidence of malignant neoplasms and cardiovascular disease. N Engl J Med 334:1145–1149. Hernandez–Avila M, Master C, Hunter DJ, Buring J, Philips J, et al. 1988. Influence of additional portion size data on the validity of a semi-quantitative food frequency questionnaire (abstract). Am J Epidemiol 128:891. Heshmat MY, Kaul L, Kovi J, Jackson MA, Jackson AG, et al. 1985. Nutrition and prostate cancer: A case-control study. Prostate 6:7–17. Hinds MW, Kolonel LN, Hankin JH, Lee J. 1984. Dietary vitamin A, carotene, vitamin C, and risk of lung cancer in Hawaii. Am J Epidemiol 119:227–237. Hodgdon JA, Fitzgerald PI. 1987. Validity of impedance predictions at various levels of fatness. Human Biol 59:281–298. Howe GR, Aronson KJ, Benito E, Castelleto R, Cornee J, et al. 1997. The relationship between dietary fat intake and risk of colorectal cancer—evidence from the combined analysis of 13 case-control studies. Cancer Causes Control 8:215–228. Howe GR, Benito E, Castelleto R, Cornee J, Esteve J, et al. 1992. Dietary intake of fiber and decreased risk of cancers of the colon and rectum:evidence from the combined analysis of 13 case-control studies. J Natl Cancer Inst 84:1887–1896. Howe GR, Hirohata T, Hislop TG, Iscovich JM, Yuan JM, et al. 1990. Dietary factors and risk of breast cancer: Combined analysis of 12 case-control studies. J Natl Cancer Inst 82:561–569. Howson CP, Hiyama T, Wynder EL. 1986. The decline in gastric cancer: Epidemiology of an unplanned triumph. Epidemiol Rev 8:1–27. Hu F, Stampfer MJ, Manson JE, Rimm E, Colditz GA, et al. 1997. Dietary fat intake and the risk of coronary heart disease in women. N Engl J Med 337:1491–1499. Hu FB, van Dam RM, Liu S. 2001. Diet and risk of Type II diabetes: The role of types of fat and carbohydrate. Diabetologia 44:805–817.

419

Hu FB, Willett WC. 2002. Optimal diets for prevention of coronary heart disease. JAMA 288:2569–2578. Huang Z, Hankinson SE, Colditz GA, Stampfer MJ, Hunter DJ, et al. 1997. Dual effects of weight and weight gain on breast cancer risk. J Am Med Assoc 278:1407–1411. Hung HC, Joshipura K, Jiang R, Hu F, Hunter D, Smith-Warner SA, Colditz GA, Rosner B, Spiegelman D, Willett WC. 2004. Fruit and vegetable intake and the risk of major chronic disease. J Natl Cancer Inst 90:1577– 1584. Hunter D. 1998. Biochemical indicators of dietary intake. In: Nutritional Epidemiology. New York: Oxford University Press. Hunter DA, Willett WC. 1994. Vitamin A and cancer: Epidemiological evidence in humans. In: Blomhoff R, ed. Vitamin A in Health and Disease. New York: Marcel Dekker. Hunter DJ, Morris JS, Stampfer MJ, Colditz GA, Speizer FE, Willett WC. 1990. A prospective study of selenium status and breast cancer risk. JAMA 264:1128–1131. Hunter DJ, Spiegelman D, Adami HO, Beeson L, van den Brandt PA, et al. 1996. Cohort studies of fat intake and the risk of breast cancer:a pooled analysis. N Eng J Med 334:356–361. International Agency for Research on Cancer. 1988. Alcohol drinking. Lyon, France: IARC, 1988; IARC Mongr Eval Carcinog Risks Humans 44:194–207. International Agency for Research on Cancer. 2002. Weight control and Physical activity. Lyon: IARC Press. Joossens JV, Geboers J. 1987. Dietary salt and risks to health. Am J Clin Nutr 45(s):1277–1288. Karvetti R–L, Knuts LR. 1985. Validity of the 24–hour dietary recall. J Am Diet Assoc 85:1437–1442. Kinlen LJ. 1983. Fat and cancer. Br Med J (Clin Res Ed) 286:1081–1082. Kipnis V, Subar AF, Midthune D, Freedman LS, Ballard–Barbash R, et al. 2003. Structure of dietary measurement error: Results of the OPEN biomarker study. Am J Epidemiol 158:14–21; discussion 2–6. Klein EA. 2003. Clinical models for testing chemopreventative agents in prostate cancer and overview of SELECT: The Selenium and Vitamin E Cancer Prevention Trial. Recent Results Cancer Res 163:212–225; discussion 64–66. Knekt P, Aromaa A, Maatela J, Aaran RK, Nikkari T, et al. 1988. Serum vitamin E and risk of cancer among Finnish men during a 10-year follow-up. Am J Epidemiol 127:28–41. Kok FJ, de Bruijn AM, Hofman A, Vermeeren R, Valkenburg HA. 1987. Is serum selenium a risk factor for cancer in men only? Am J Epidemiol 125:12–16. Kolonel LN, Yoshizawa CN, Hankin JH. 1988. Diet and prostatic cancer: A case-control study in Hawaii. Am J Epidemiol 127:999–1012. Kushi LH, Lew RA, Stare FJ, Ellison CR, Lozy M, et al. 1985. Diet and 20year mortality from coronary heart disease: The Ireland–Boston Diet–Heart study. N Engl J Med 312:811–818. Lamprecht SA, Lipkin M. 2001. Cellular mechanisms of calcium and vitamin D in the inhibition of colorectal carcinogenesis. Ann N Y Acad Sci 952:73–87. Le Marchand L, Kolonel LN, Earle ME, Mi MP. 1988. Body size at different periods of life and breast cancer risk. Am J Epidemiol 128:137– 152. Le Marchand L, Kolonel LN, Wilkens LR, Myers BC, Hirohata T. 1994. Animal fat consumption and prostate cancer:a prospective study in Hawaii. Epidemiology 5:276–282. Lippman SM, Meyskens FL. 1989. Retinoids for the prevention of cancer. In: Moon TE, Micozzi MS, eds. Micronutrients and Cancer Prevention. New York: Marcel Dekker, Inc. pp. 243–272. Lohman TG, Roche AF, Martorell R. 1988. Anthropometric Standardization Reference Manual. Champaign, IL: Human Kinetics Books. Lyon JL, Mahoney AW, West DW, Gardner JW, Smith KR, et al. 1987. Energy intake: Its relationship to colon cancer risk. J Natl Cancer Inst 78:853– 861. Madden JP, Goodman SJ, Guthrie HA. 1976. Validity of the 24-hour recall. Analysis of data obtained from elderly subjects. J Am Diet Assoc 68:143–147. Manousos O, Day NE, Trichopoulos D, Gerovassilis F, Tzanou A, Polychronopoulou A. 1983. Diet and colorectal cancer: A case-control study in Greece. Int J Cancer 32:1–5. Marr JW. 1971. Individual dietary surveys: Purposes and methods. World Rev Nutr Diet 13:105–164. Mathews-Roth MM. 1989. Beta-carotene, canthaxanthin, and phytoene. In: Moon TE, Micozzi MS, eds. Micronutrients and Cancer Prevention. New York: Marcel Dekker, Inc. pp. 273–290. McKeown-Eyssen GE, Bright-See E. 1985. Dietary factors in colon cancer: International relationships. An update. Nutr Cancer 7:251–253.

420

PART III: THE CAUSES OF CANCER

McMichael AJ, Giles GG. 1988. Cancer in migrants to Australia: Extending the descriptive epidemiological data. Cancer Res 48:751–756. Menkes MS, Comstock GW, Vuilleumier JP, Helsing KJ, Rider AA, Brookmeyer R. 1986. Serum beta-carotene, vitamins A and E, selenium, and the risk of lung cancer. N Engl J Med 315:1250–1254. Mersens WJ, Bhagavan HN, Mertz W. 1989. Alpha-tocopherols (vitamin E). In: Moon TE, Micozzi MS, eds. Micronutrients and Cancer Prevention. New York: Marcel Dekker, Inc., pp. 341–388. Mertz W. 1984. Food and nutrients. J Am Dietet Assoc 84:769–770. Mettlin C, Graham S, Priore R. 1981. Diet and cancer of the esophagus. Nutr Cancer 2:143–147. Mettlin C, Graham S, Swanson M. 1979. Vitamin A and lung cancer. J Natl Cancer Inst 62:1435–1438. Michaud DS, Feskanich D, Rimm EB, Colditz GA, Speizer FE, et al. 2000. Intake of specific carotenoids and risk of lung cancer in 2 prospective US cohorts. Am J Clin Nutr 72:990–997. Michels KB, Giovannucci E, Joshipura KJ, Rosner BA, Stampfer MJ, Fuchs CS, Colditz GA, Speizer FE, Willett WC. 2000. Prospective study of fruit and vegetable consumption and incidence of colon and rectal cancers. J Natl Cancer Inst 92:1740–52. Erratum in: J Natl Cancer Inst 2001. 6:879. Michels KB, Willett WC. 1992. The Women’s Health Initiative: Daughter of politics or science? Principles and Practices of Oncology 6:1–11. Mirvish SS, Wallcave L, Eagen M, Shubik P. 1972. Ascorbate-nitrite reaction: Possible means of blocking the formation of carcinogenic N-nitroso compounds. Science 177:65–68. Morgan RW, Jain M, Miller AB, Choi NW, Matthews V, et al. 1978. A comparison of dietary methods in epidemiologic studies. Am J Epidemiol 107:488–498. Morris JS, Stampfer MJ, Willett W. 1983. Dietary selenium in humans: Toenails as an indicator. Biol Trace Elem Res 5:529–537. Multiple Risk Factor Intervention Trial Research Group. 1982. Multiple Risk Factor Intervention Trial: Risk factor changes and mortality results. J Am Med Assoc 248:1465–1477. National Research Council (U.S.), Committee on Diet and Health. 1989. Diet and health: Implications for reducing chronic disease risk. Washington, DC: National Academy Press. Nomura A, Heilbrun LK, Morris JS, Stemmermann N. 1987. Serum selenium and the risk of cancer, by specific sites: Case-control analysis of prospective data. J Natl Cancer Inst 79:103–108. Nomura AM, Stemmermann GN, Heilbrun LK, Salkeld RM, Vuilleumier JP. 1985. Serum vitamin levels and the risk of cancer of specific sites in men of Japanese ancestry in Hawaii. Cancer Res 45:2369–2372. Norat T, Lukanova A, Ferrari P, Riboli E. 2002. Meat consumption and colorectal cancer risk: Dose-response meta-analysis of epidemiological studies. Int J Cancer 98:241–256. Omenn GS, Goodman GE, Thornquist MD, Balmes J, Cullen MR, et al. 1996. Effects of a combination of beta carotene and vitamin A on lung cancer and cardiovascular disease. N Engl J Med 334:1150–1155. Pamukcu AM, Price JM, Bryan GT. 1970. Assay of fractions of bracken fern (Pteris aquilina) for carcinogenic activity. Cancer Res 30:902–905. Peleg I, Morris S, Hames CG. 1985. Is serum selenium a risk factor for cancer? Med Oncol Tumor Pharmacother 2:157–163. Peto R, Doll R, Buckley JD, Sporn MB. 1981. Can dietary beta-carotene materially reduce human cancer rates? Nature 290:201–208. Phillips RL, Garfinkel L, Kuzma JW, Beeson WL, Lotz T, Brin B. 1980. Mortality among California Seventh–day Adventists for selected cancer sites. J Natl Cancer Inst 65:1097–1107. Pickle LW, Hartman AM. 1985. Indicator foods for vitamin A assessment. Nutr Cancer 7:3–23. Pietinen P, Hartman AM, Haapa E, Rasanen L, Haapakoski J, et al. 1988a. Reproducibility and validity of dietary assessment instruments II. A qualitative food-frequency questionnaire. Am J Epidemiol 128:667–676. Pietinen P, Hartman AM, Haapa E, Rasanen L, Haapakoski J, et al. 1988b. Reproducibility and validity of dietary assessment instruments. I. A selfadministered food use questionnaire with a portion size picture booklet. Am J Epidemiol 128:655–666. Posner BM, Borman CL, Morgan JL, Borden WS, Ohls JC. 1982. The validity of a telephone-administered 24-hour dietary recall methodology. Am J Clin Nutr 36:546–553. Prasad MP, Krishna TP, Pasricha S, Krishnaswamy K, Quereshi MA. 1992. Esophageal cancer and diet—A case-control study. Nutr Cancer 18: 85–93. Prentice RL, Kakar F, Hursting S, Sheppard L, Klein R, Kushi LH. 1988. Aspects of the rationale for the Women’s Health Trial. J Natl Cancer Inst 80:802–814. Raineri R, Weisburger JH. 1975. Reduction of gastric carcinogens with ascorbic acid. Ann N Y Acad Sci 258:181–189.

Rhoads GG, Kagan A. 1983. The relation of coronary disease, stroke, and mortality to weight in youth and middle age. Lancet 1:492–495. Rimm EB, Giovannucci EL, Stampfer MJ, Colditz GA, Litin LB, Willett WC. 1992. Reproducibility and validity of a expanded self-administered semiquantitative food frequency questionnaire among male health professionals. Am J Epidemiol 135:1114–1126. Rohan TE, Jain MG, Howe GR, Miller AB. 2000. Dietary folate consumption and breast cancer risk. J Natl Cancer Inst 92:266–269. Rohan TE, Potter JD. 1984. Retrospective assessment of dietary intake. Am J Epidemiol 120:876–887. Romney SL, Duttagupta C, Basu J, Palan PR, Karp S, et al. 1985. Plasma vitamin C and uterine cervical dysplasia. Am J Obstet Gynecol 151:976–980. Rosner B, Hennekens CH, Kass EH, Miall WE. 1977. Age-specific correlation analysis of longitudinal blood pressure data. Am J Epidemiol 106:306– 313. Rosner B, Spiegelman D, Willett WC. 1992. Correction of logistic regression relative risk estimates and confidence intervals for random within–person measurement error. Am J Epidemiol 136:1400–1413. Rosner B, Willett WC, Spiegelman D. 1989. Correction of logistic regression relative risk estimates and confidence intervals for systematic within–person measurement error. Statistics Med 8:1051–1069. Ross MH, Bras G. 1971. Lasting influence of early caloric restriction on prevalence of neoplasms in the rat. J Natl Cancer Inst 47:1095–1113. Salonen JT, Alfthan G, Huttunen JK, Puska P. 1984. Association between serum selenium and the risk of cancer. Am J Epidemiol 120:342–349. Salonen JT, Salonen R, Lappeteläinen R, Meänpää PH, Alfthan G, Puska P. 1985. Risk of cancer in relation to serum concentrations of selenium and vitamins A and E: Matched case-control analysis of prospective data. Br Med J Clin Res Ed 290:417–420. Salvini S, Hunter DJ, Sampson L, Stampfer MJ, Colditz GA, et al. 1989. Foodbased validation of a dietary questionnaire: The effects of week-to-week variation in food consumption. Int J Epidemiol 18:858–867. Samet JM, Humble CG, Skipper BE. 1984. Alternatives in the collection and analysis of food frequency interview data. Am J Epidemiol 120:572–581. Samet JM, Skipper BJ, Humble CG, Pathak DR. 1985. Lung cancer risk and vitamin A consumption in New Mexico. Am Rev Respir Dis 131:198– 202. Schrauzer GN, White DA, Schneider CJ. 1977. Cancer mortality correlation studies—III: Statistical associations with dietary selenium intakes. Bioinorg Chem 7:23–31. Seidell JC. 2002. Prevalence and time trends of obesity in Europe. J Endocrinol Invest 25:816–822. Shamberger RJ, Tytko SA, Willis CE. 1976. Antioxidants and Cancer Part VI. Selenium and age-adjusted human cancer mortality. Arch Environ Health 31:231–235. Shekelle RB, Lepper M, Liu S, Maliza C, Raynor WJJ, et al. 1981. Dietary vitamin A and risk of cancer in the Western Electric Study. Lancet 2:1186–1190. Shekelle RB, Rossof AH, Stamler J. 1991. Dietary cholesterol and incidence of lung cancer: The Western Electric Study. Am J Epidemiol 134:480–484. Smith–Warner S, Spiegelman D, Yaun S–S, Albanes D, Beeson W, et al. 2003. Fruits, vegetables and lung cancer: A pooled analysis of cohort studies. Int J Cancer 107:1001–1011. Smith–Warner SA, Ritz J, Hunter DJ, Albanes D, Beeson WL, et al. 2002. Dietary fat and risk of lung cancer in a pooled analysis of prospective studies. Cancer Epidemiol Biomarkers Prev 11:987–992. Smith–Warner SA, Spiegelman D, Adami HO, Beeson WL, van den Brandt PA, et al. 2001a. Types of dietary fat and breast cancer: A pooled analysis of cohort studies. Int J Cancer 92:767–774. Smith–Warner SA, Spiegelman D, Yaun SS, Adami HO, Beeson WL, et al. 2001b. Intake of fruits and vegetables and risk of breast cancer: A pooled analysis of cohort studies. JAMA 285:769–776. Smith–Warner SA, Spiegelman D, Yaun SS, Adami HO, van den Brandt PA, et al. 1998. Alcohol and breast cancer in women: A pooled analysis of cohort studies. J Am Med Assoc 279:535–540. Spiegelman D, Israel RG, Bouchard C, Willett WC. 1992. Absolute fat mass, percent body fat, and body-fat distribution: Which is the real determinant of blood pressure and serum glucose? Am J Clin Nutr 55:1033–1044. Stahelin HB, Rosel F, Buess E, Brubacher G. 1984. Cancer, vitamins, and plasma lipids: Prospective Basel study. J Natl Cancer Inst 73:1463–1468. Stampfer MJ, Willett WC, Speizer FE, Dysert DC, Lipnick R, et al. 1984. Test of the National Death Index. Am J Epidemiol 119:837–839. Staszewski J, Haenszel W. 1965. Cancer mortality among the Polish-born in the United States. J Natl Cancer Inst 35:291–297. Stephanik PA, Trulson MF. 1962. Determining the frequency of foods in large group studies. Am J Clin Nutr 2:335–343.

Diet and Nutrition Stunkard AJ, Albaum JM. 1981. The accuracy of self-reported weights. Am J Clin Nutr 34:1593–1599. Sugimura T. 1986. Studies on environmental chemical carcinogenesis in Japan. Science 233:312–318. Swanson CA, Jones DY, Schatzkin A, Brinton LA, Ziegler RG. 1988. Breast cancer risk assessed by anthropometry in the NHANES I epidemiological follow-up study. Cancer Res 48:5363–5367. Tannenbaum A. 1942. The genesis and growth of tumors. III: Effects of a high fat diet. Cancer Res 2:468–475. Terry P, Giovannucci E, Michels KB, Bergkvist L, Hansen H, et al. 2001. Fruit, vegetables, dietary fiber, and risk of colorectal cancer. J Natl Cancer Inst 93:525–533. Tretli S. 1989. Height and weight in relation to breast cancer morbidity and mortality. A prospective study of 570,000 women in Norway. Int J Cancer 44:23–30. Trock B, Lanza E, Greenwald P. 1990. Dietary fiber, vegetables, and colon cancer:critical review and meta-analyses of the epidemiologic evidence. J Natl Cancer Inst 82:650–661. U.S. Department of Agriculture, U.S. Department of Health and Human Services. 2000. Nutrition and Your Health: Dietary Guidelines for Americans, Fifth Edition. Rep. Home and Garden Bulletin No. 232. Washington, DC: U.S. Government Printing Office. Vague J. 1956. The degree of masculine differentiation of obesity: A factor determining predisposition to diabetes, atherosclerosis, gout and uriccalculous disease. Am J Clin Nutr 4:20–34. Valaoras VG, MacMahon B, Trichopoulos D, Polychronopoulou A. 1969. Lactation and reproductive histories of breast cancer patients in greater Athens, 1965–67. Int J Cancer 4:350–363. van den Brandt PA, Goldbohm RA, van’t veer P, Bode P, Dorant E, et al. 1993. A prospective cohort study on selenium status and the risk of lung cancer. Cancer Res 53:4860–4865. van den Brandt PA, Spiegelman D, Yaun SS, Adami HO, Beeson L, et al. 2000. Pooled analysis of prospective cohort studies on height, weight, and breast cancer risk. Am J Epidemiol 152:514–527. van Noord PAH, Collette HJA, Maas MJ, de Waard F. 1987. Selenium levels in nails of premenopausal breast cancer patients assessed prediagnostically in a cohort-nested case–referent study among women screened in the DOM project. Int J Epidemiol 16:318–322. Virtamo J, Valkeila E, Alfthan G, Punsar S, Huttunen JK, Karvonen MJ. 1987. Serum selenium and risk of cancer: A prospective follow-up of nine years. Cancer 60:145–148. Wald N, Boreham J, Bailey A. 1986. Serum retinol and subsequent risk of cancer. Br J Cancer 54:957–961. Wald NJ, Boreham J, Hayward JL, Bulbrook RD. 1984. Plasma retinol, betacarotene, and vitamin E levels in relation to the future risk of breast cancer. Br J Cancer 49:321–324. Wald NJ, Nicolaides–Bouman A, Hudson GA. 1988. Plasma retinol, betacarotene and vitamin E levels in relation to the future risk of breast cancer (letter). Br J Cancer 57:235. Wald NJ, Thompson SG, Densem JW, Boreham J, Bailey A. 1987. Serum vitamin E and subsequent risk of cancer. Br J Cancer 56:69–72. Walker AM, Blettner M. 1985. Comparing imperfect measures of exposure. Am J Epidemiol 121:783–790. Wassertheil–Smoller S, Romney SL, Wylie–Rosett J, Slagle S, Miller G, et al. 1981. Dietary vitamin C and uterine cervical dysplasia. Am J Epidemiol 114:714–724. Wattenberg LW, Loub WD. 1978. Inhibition of polycyclic aromatic hydrocarbon–induced neoplasia by naturally occurring indoles. Cancer Res 38:1410–1413. Weindruch R, Walford RL. 1982. Dietary restriction in mice beginning at 1 year of age: Effect on life-span and spontaneous cancer incidence. Science 215:1415–1418.

421

Welsh SO, Marston RM. 1982. Review of trends in food use in the United States, 1909 to 1980. J Am Diet Assoc 81:120–128. Wiehl DG, Reed R. 1960. Development of new or improved dietary methods for epidemiological investigations. Am J Pub Health 50:824–828. Willett W. 1989a. An overview of issues related to the correction of nondifferential exposure measurement error in epidemiologic studies. Statistics Med 8:1031–1040. Willett W. 1989b. The search for the causes of breast and colon cancer. Nature 338:389–394. Willett W. 2001. Commentary: Dietary diaries versus food frequency questionnaires—A case of undigestible data. Int J Epidemiol 30:317–319. Willett W. 2003. Invited Commentary: OPEN Questions. Am J Epidemiol 158:22–24. Willett WC. 1990. Vitamin A and lung cancer. Nutr Rev 48:201–211. Willett WC. 1998. Nutritional Epidemiology, Second Edition. New York: Oxford University Press. Willett WC, Browne ML, Bain C, Lipnick RJ, Stampfer MJ, et al. 1985a. Relative weight and risk of breast cancer among premenopausal women. Am J Epidemiol 122:731–740. Willett WC, Sampson L, Stampfer MJ, Rosner B, Bain C, et al. 1985b. Reproducibility and validity of a semiquantitative food frequency questionnaire. Am J Epidemiol 122:51–65. Willett WC, Stampfer M, Chu N, Spiegelman D, Holmes M, Rimm E. 2001. Assessment of questionnaire validity for measuring total fat intake using plasma lipid levels as criteria. Am J Epidemiol 154:1107–1112. Willett WC, Stampfer MJ. 1986. Total energy intake: Implications for epidemiologic analyses. Am J Epidemiol 124:17–27. Willett WC, Stampfer MJ, Colditz GA, Rosner BA, Hennekens CH, Speizer FE. 1987. Dietary fat and the risk of breast cancer. N Engl J Med 316:22–28. Willett WC, Stampfer MJ, Underwood BA, Speizer FE, Rosner B, Hennekens CH. 1983. Validation of a dietary questionnaire with plasma carotenoid and alpha–tocopherol levels. Am J Clin Nutr 38:631–639. Winn DM, Ziegler RG, Pickle LW, Gridley G, Blot WJ, Hoover RN. 1984. Diet in the etiology of oral and pharyngeal cancer among women from the southern United States. Cancer Res 44:1216–1222. World Cancer Research Fund, American Institute for Cancer Research. 1997. Food, Nutrition and the Prevention of Cancer: A Global Perspective. Washington, DC: American Institutue for Cancer Research. Wu AH, Henderson BE, Pike MC, Yu MC. 1985. Smoking and other risk factors for lung cancer in women. J Natl Cancer Inst 74:747–751. Yoshizawa K, Willett WC, Morris SJ, Stampfer MJ, Spiegelman D, et al. 1998. Study of prediagnostic selenium level in toenails and the risk of advanced prostate cancer. J Natl Cancer Inst 90:1219–1224. You WC, Blot WJ, Chang YS, Ershow AG, Yang ZT, et al. 1988. Diet and high risk of stomach cancer in Shandong, China. Cancer Res 48:3518– 3523. Zhang S, Hunter DJ, Hankinson SE, Giovannucci EL, Rosner BA, et al. 1999. A prospective study of folate intake and the risk of breast cancer. JAMA 281:1632–1637. Zhang SM, Willett WC, Selhub J, Hunter DJ, Giovannucci EL, et al. 2003. Plasma folate, vitamin B6, vitamin B12, and homocysteine and risk of breast cancer. J Natl Cancer Inst 95:373–380. Ziegler RG, Hoover RN, Pike MC, Hildesheim A, Nomura AMY, et al. 1993. Migration patterns and breast cancer risk in Asian–American women. J Natl Cancer Inst 85:1819–1827. Ziegler RG, Mason TJ, Stemhagen A, Hoover R, Schoenberg JB, et al. 1984. Dietary carotene and vitamin A and risk of lung cancer among white men in New Jersey. J Natl Cancer Inst 73:1429–1435. Ziegler RG, Morris LE, Blot WJ, Pottern LM, Hoover R, Fraumeni JF, Jr. 1981. Esophageal cancer among black men in Washington, D.C. II. Role of nutrition. J Natl Cancer Inst 67:1199–1206.

22

Obesity and Body Composition RACHEL BALLARD-BARBASH, CHRISTINE FRIEDENREICH, MARTHA SLATTERY, AND INGER THUNE

O

ne explanation for the increasing incidence of cancers in developing countries has been the hypothesis that a chronic state of positive energy balance promotes tumor growth. (Tannenbaum 1940a,b) first explored this hypothesis in animal models and demonstrated that both increased calories and fat increased the occurrence and number of breast cancer tumors in mice. Similarly, epidemiologic studies in humans in the 1970s demonstrated that heavier women were at increased risk of breast and endometrial cancer (De Waard et al., 1974; Blitzer et al., 1976). Other population-based studies in the early 1980s suggested that risk increased with increasing weight for several other cancer sites, specifically colon, prostate, gallbladder (among women), and kidney (Lew and Garfinkel, 1979; Hartz et al., 1984; Garfinkel, 1986). Since then, numerous epidemiologic studies have examined the association of weight and other anthropometric measures with cancer incidence and have explored potential biological mechanisms to explain observed associations. The majority of these studies have focused on breast, colon, endometrial, renal cell, esophageal/gastric, and prostate cancer; limited data are available on body size associations with other cancer sites. The aim of this review is to provide a comprehensive overview of the state of scientific evidence for the association between obesityrelated risk factors and cancer, with a focus on those cancers with sufficient evidence for review: those of the colon, rectum, esophagus/ gastric cardia, renal cell, endometrium, ovary, breast, prostate, thyroid, lung, and head and neck. The specific objectives of this review are to:

• Summarize the epidemiologic literature for the association between

• • • •

cancer risk and obesity-related measures of weight or body mass index (BMI; defined most commonly as weight in kilograms [kg] divided by height in meters squared), fat deposition patterns (commonly approximated by waist circumference or by the ratio of waistto-hip circumferences [WHR] or by ratios of truncal to extremity skinfolds), and weight change; Highlight gaps in scientific knowledge for the association of obesity-related factors and cancer risk; Identify the possible underlying biological mechanisms for these associations; Comment on areas for future research; and Estimate the public health impact of obesity on cancer risk at the international level.

Because of space limitations, the epidemiologic evidence on height and cancer risk is not included in this review but has been well summarized in several previous reviews (Ballard-Barbash, 1999; Gunnell et al., 2001).

METHODS A search was conducted on MEDLINE and PUBMED for all publications on weight, body mass index, anthropometric factors, and specific cancers in human populations and was supplemented by a manual search of all major relevant journals. The literature search included all publications up to February 2003. Studies included in this review focused on some aspect of anthropometric risk factors in relation to cancer risk for the cancers noted above. Major review

422

papers or books were also identified and reviewed. Summary figures of risk estimates for the highest compared to lowest quantile of BMI and the major cancer sites were generated. Because of the very large number of studies in breast cancer, studies with at least 100 cases each of pre- and postmenopausal breast cancer cases are included in the breast cancer figures. Because of differences in risk for colon and rectal cancer and the limited number of studies on rectal cancer alone, only data on colon cancer were summarized in the colon cancer figure. Biological mechanisms potentially involved in the association between obesity and specific cancers were summarized in a table. In addition, population attributable risks estimated for the European Union by Bergström et al. (2001) were compared to current estimates for the United States and summarized in a table. A summary statement for each cancer site was developed following the criteria for convincing, probable, possible, and no association as summarized in the 1997 World Cancer Research Fund and American Institue for Cancer Research (World Cancer Research Fund, 1997) report on Food, Nutrition and the Prevention of Cancer: A Global Perspective.

METHODOLOGIC AND MEASUREMENT ISSUES Existing methods for assessing the etiologic and prognostic influences of energy balance and body size on cancer have limitations that should be considered when evaluating the epidemiologic evidence and planning future research. Briefly summarized, these limitations fall into three areas: validity of exposure assessment, difficulty of comparisons across studies, and insufficient sample sizes to explore fully the many factors that relate to body size and fat mass and that could potentially modify observed associations. The measurement of anthropometric indices is generally well standardized and documented (Lohman et al., 1988; WHO, 1995). In fact, the standardization, accuracy, and reliability of these measures are better than for many other cancer-related exposures. Weight and height are the most standardized measurements and least subject to variability. More recent investigations have used interviewers to measure various anthropometric indices from study participants, hence reducing the possibility of random and systematic error. However, many studies of cancer etiology have relied on self-reported weight and height. Studies on self-reported height suggest a reasonable degree of accuracy with a bias in overreporting height that is somewhat greater in men compared to women and that increases with age (Rowland, 1990) or is limited to people over age 60 (Kuczmarski et al., 2001), presumably due to age-related loss in height. Studies on self-reported weight indicate the presence of significant misreporting at the extremes of weight, with heavier people underreporting and lighter people overreporting their weight (Rowland, 1990; Stevens et al., 1990; Must et al., 1993). Therefore, studies of chronic disease that rely on self-reported height and weight will underestimate the risk associated with these measures (Gunnell et al., 2000). Nonetheless, compared to estimates of correlation between reported versus measured assessment of other exposures, such as dietary intake, correlation coefficients between recalled and measured weight suggest that weight is recalled with a reasonable degree of accuracy. For example, in one U.S. study, correlations between reported and measured weights for

Obesity and Body Composition elderly subjects asked to recall weight currently and from 4 and 28 years previously were reported to be 0.98 for current, 0.94 for 4-year, and 0.82 for 28-year recall (Stevens et al., 1990). Measurement of skinfolds and circumferences is less reliable than the measurement of weight and height (Lohman et al., 1988; WHO, 1995). However, reliability is improved with standardization in measurement technique. It is not feasible to obtain these anthropometric measures from selfreport. Waist circumference is currently considered to be the most convenient and simple measure of abdominal or central adiposity for epidemiologic research (WHO, 1995, 2000). Percent body fat can be estimated from bioelectric impedance but only has been reported in one study of cancer etiology (Lahmann, 2003). Bioelectric impedance is currently the most promising field method for estimating body composition in large epidemiologic studies because it is portable, inexpensive, easy to use, and highly precise. Issues related to use of different methods for estimating body composition for chronic disease epidemiology are well summarized by Baumgartner and colleagues (1995). Comparison across studies is difficult because most studies have examined risk by quantile distributions (most commonly tertile or quartile) for the BMI or other anthropometric measures used. As the distribution of BMI varies across populations, these quantile groups are not comparable across studies, making comparison difficult. With increasing interest in understanding the risk for many chronic diseases by standard WHO BMI categories, investigators have begun to examine risk by these categories: underweight as BMI of less than 18.50, normal weight as BMI of 18.50–24.99, overweight as BMI of 25.00 or higher. This latter overweight category is further subdivided into four categories: preobese, 25.00–29.99; obese class I, 30.00–34.99; obese class II, 35.00–39.99; and obese class III, ≥40.00 (WHO, 2000). More recent meta-analyses have the ability to examine risk by these broad weight categories; however, analyses should not be limited to these categories as risk may vary within them, depending on the chronic diseases and the populations examined. Research to date suggests often complex and varying associations of body size with different cancer sites. Although cancer as a whole now exceeds coronary artery disease as a cause of death in the United States population under age 85, the incidence of site-specific cancer is much lower than diseases such as diabetes mellitus and coronary artery disease, limiting the sample sizes of studies for less common cancers. Therefore, delineation of the association of body size and cancer requires site-specific studies, which can often be done only through multicollaborative efforts for less common cancers. Furthermore, as more factors are identified as potentially influencing body size and cancer associations, delineation of the underlying mechanisms requires further subgroup analysis, thereby increasing sample size requirements. The “gold standard” technologies required, such as doubly-labeled water for measuring energy expenditure or computerized tomography for measuring body composition, are expensive, not easily portable, and, thus, not feasible for large samples. Therefore, it is not possible to directly measure positive energy balance in large epidemiologic studies of cancer risk or prognosis. In addition, although it is possible to estimate energy balance from its components—energy intake and expenditure—current methods to assess diet and physical activity rely on self-report of food intake and physical activity, a methodology with substantial reporting error. Consequently, because large epidemiologic studies cannot measure energy balance through either doubly-labeled water techniques or through accurate estimation of energy intake and expenditure from self-reports, the most valid measure of either persistent or recurrent states of positive energy balance is weight gain during adult life. Given findings from clinical metabolic research of complex interactions among different steroid hormones, such as sex steroids and insulin, ranging from receptor cross-reactivity to postreceptor potentiation, statistical modeling methods that allow for this complex interaction are needed and have not been used. In the future, statistical approaches for complex systems, including factor analysis and assessment of effect modification and interaction, may advance understanding of the metabolic factors underlying body size and cancer associations.

423

COLORECTAL CANCER Summary of Findings Data from case-control and cohort studies provide convincing evidence of an approximately twofold increase in risk of colon cancer among men with a BMI of 30 or more. Risk estimates for men are most commonly reported to be about 2. Among women, less extensive data provide evidence of a probably increased risk of colon cancer from a high BMI, with risk estimates usually between 1.0 and 1.5. Among obese women who are estrogen positive (defined as premenopausal or postmenopausal and taking hormone replacement therapy [HRT]), a twofold increase in risk has been reported. Most of the literature has found no association between BMI and rectal cancer. Meta-analyses that provide a more quantitative statement of the association across several studies have not been published.

Overview Many aspects of the associations between body weight, obesity, and colorectal cancer have been examined. Colorectal cancer has been studied by examining risk associated with colon and rectal cancers combined, as well as for each site separately. Although many studies have examined associations with colon cancer, few have considered rectal cancer specifically. Additionally, unlike many cancers, associations between body size and the precursor lesion, adenomatous polyps, have been reported. Several consistent differences in associations by gender, site, and location within the colon have been observed. Stronger associations have been observed for men than for women; for colon than for rectal cancer; and, within the colon, for distal than for proximal tumors. Studies examining the associations between body size, obesity, and colorectal cancer have used many different indicators of body size. Most studies have relied on BMI, usually from several years before diagnosis. Few studies have used the current WHO criteria for overweight (BMI ≥ 25) and obesity (BMI ≥ 30), making comparisons and interpretations across studies difficult. Most studies have categorized BMI based on distribution in the population, and although the lower limit for the upper category of BMI is often near 30, at other times it is around 26. Limited data exist on waist and hip circumferences as indicators of fat pattern and its impact on colorectal cancer risk. Most studies focus on adult BMI, with little information available on weight change or early life body size as possible predictors of risk. Because of differences in reported associations for colon and rectal cancer, evidence for these sites is summarized separately.

Colon Cancer Epidemiology Many studies have evaluated associations between body size and colon cancer, and reported associations have been fairly consistent across case-control and cohort studies. Studies that have examined associations with obesity report slightly stronger associations than those for overweight. For men, most studies report significant increased risk, with risk estimates of 2.0 or greater, although risk estimates range from 1.2 to 3.0 (Graham et al., 1978; West et al., 1989; Gerhardsson de Verdier et al., 1990; Le Marchand et al., 1992; Lee and Paffenberger, 1992a; Giovannucci et al., 1995; Caan et al., 1998; Singh et al., 1998; Schoen et al., 1999) (Figures 22–1a and 22–1b). Most studies have adjusted for important confounding factors, including physical activity, dietary intake, and smoking, although some report age-adjusted estimates of association only (Lund Nilsen and Vatten, 2001). Associations observed for women are generally weaker than those observed for men and are often not statistically significant. For women, risk estimates have generally ranged from 0.7 to 2.7 (Potter and McMichael, 1983; Graham et al., 1988; Kune et al., 1990; Bostick et al., 1994; Dietz et al., 1995; Le Marchand et al., 1997; Martinez et al., 1997; Caan et al., 1998; Singh et al., 1998; Ford, 1999; Terry et al., 2001, 2002). Some studies report a similar twofold increase in risk of colon cancer for both men and women for the highest BMI quantile examined (approximately 30) (West et al., 1989;

424

PART III: THE CAUSES OF CANCER

A

Figure 22–1. A, Summary of risk estimates and confidence intervals from epidemiologic studies of BMI and colon cancer in men. B, Summary of risk estimates and confidence intervals from epidemiologic studies of BMI and colon cancer in women.

B

Ford, 1999). Few studies have included African Americans. The one study reporting associations for African Americans was limited to 99 cases and reported no significant association between BMI and colorectal cancer (Dales et al., 1979). Some studies have had sufficiently large sample sizes such that associations for proximal and distal colon tumors could be considered separately. Stronger associations for distal tumors than proximal tumors are often seen (Dietz et al., 1995; Martinez et al., 1997; Caan et al., 1998), although the study by Le Marchand et al. (1997) showed stronger associations with proximal rather than distal tumors, and the study of women by Terry et al. (2001) did not observe a significant association for either proximal or distal tumors. In a study by Le Marchand et al. (1992), the greatest risk associated with a high BMI was observed for sigmoid tumors, which most studies include as part of the distal colon. Some studies have examined waist-to-hip ratio and colon cancer risk (Martinez et al., 1997; Caan et al., 1998; Russo et al., 1998;

Schoen et al., 1999). Similar results have been observed from both case-control and cohort studies, and suggest that a higher WHR increases risk of colon cancer.

Biological Mechanisms and Tumor Mutations Associations between body size and colon cancer risk may be modified by several factors. Interactions between BMI, physical activity, and hormones appear to provide the most insight into possible biological mechanisms. High levels of vigorous physical activity have been shown to modify the risk associated with obesity; for example, risk is not as markedly increased among obese men and women who are physically active (Lee and Paffenberger, 1992a; Le Marchand et al., 1997; Slattery et al., 1997). Interactions between obesity and estrogen or HRT may explain gender differences in risk associated with BMI (Terry et al., 2002). Studies have shown that being overweight or obese increases risk of colon cancer only among women who are premenopausal (Terry et al., 2002). In another study that examined

Obesity and Body Composition estrogen-positive women (defined as premenopausal or postmenopausal and taking HRT), risks associated with a high BMI were similar to associations observed in men, with risk estimates for a BMI of 30 or more being twofold greater than that of women who are lean (Slattery et al., 2003). However, women who were estrogen negative (defined as postmenopausal and not taking HRT) did not experience increased risk from being overweight or obese. In men it has been observed that, with advancing age, the risk associated with being overweight or obese declines (Slattery et al., 2003). This change in risk over time could be the result of declining androgen levels that operate in a similar fashion as estrogen in regulating risk associated with obesity. These interactions between physical activity and BMI, and estrogen and BMI, suggest that at least one way BMI may influence colorectal risk is through its influence on estrogen and on the interaction between estrogen and insulin. Evaluations of the association between BMI and colon cancer that have examined specific tumor mutations have provided insight into possible mechanisms of action and the role of body size in possible specific disease pathways. The few published studies suggest that BMI may be involved in several pathways. One study reported that an elevated BMI was associated with Ki-ras mutations in codons 12 and 13 (Slattery et al., 2001a). The BMI associations appeared to be more specific for Ki-ras mutations in women than in men. BMI appeared to be equally associated with tumors with and without p53 mutations (Slattery et al., 2002). Obesity was reported as being associated only with tumors that were stable (i.e., negative for microsatellite instability) in women, although in men, obesity was associated with both stable and unstable tumors (Slattery et al., 2001b).

Rectal Cancer The small number of studies that have examined associations between body size and rectal cancer generally report results for few cases. From existing data it appears that body size is not associated with rectal cancer among men or women (Graham et al., 1988; Gerhardsson de Verdier et al., 1990; Dietz et al., 1995; Le Marchand et al., 1997; Russo et al., 1998; Terry et al., 2001). In contrast, one study of women reported reduced risk of rectal cancer among those with a BMI of more than 25 relative to those with a BMI of less than 22 (Potter and McMichael, 1986). Unlike colon cancer, studies of rectal cancer have not evaluated interactions between BMI and other indicators of body size with diet and lifestyle factors. This lack of examination of interaction is most probably because of the relatively limited number of rectal cancer cases available for analysis, making studies of interaction imprecise and problematic. Evaluation of BMI and other indicators of body size with specific mutations in tumors have not been reported.

Colorectal Adenomas Studies evaluating the association between body size and the occurrence of adenomas have examined associations by polyp size and type. Most studies of colorectal adenomas have examined BMI alone and have observed similar associations as for colon cancer (Neugut et al., 1991; Shinchi et al., 1994; Davidow et al., 1996; Giovannucci et al., 1996). Some studies do not detect associations with adenomas and body size (Terry et al., 2001, 2002). In a study that included both cancer and adenomas, an increased risk with a high BMI was observed for large adenomas only, but not for colorectal tumors (Boutron-Ruault et al., 2001). Most studies have observed a 1.5- to 2.5-fold increase in risk of adenomas among the group with the highest BMI. A large WHR also has been associated with adenomas, with stronger associations being observed for larger adenomas (Kono et al., 1991; Shinchi et al., 1994; Kono et al., 1999). Similar to studies of rectal cancer, studies of adenomas have been limited in terms of examining interactions with other diet and lifestyle factors, primarily because of small sample sizes available to examine interactions. One study by Martinez et al. (1999) observed no differences in risk for adenomas among those with and without Ki-ras mutations.

425

ADENOCARCINOMAS OF THE ESOPHAGUS (AE) AND GASTRIC CARDIA (AC) Summary of Findings Data from case-control and cohort studies provide evidence of a probable and fairly consistent association of a twofold or greater increased risk of adenocarcinoma of the esophagus and gastric cardia from overweight and obesity. Most studies have evaluated risk using a BMI of around 30 or more as the upper level of BMI, although some studies use a much lower BMI as the upper level. The increased risk associated with obesity has been observed in both men and women. Metaanalyses that provide a more quantitative statement of the association across several studies have not been published.

Epidemiology The incidence of adenocarcinoma of the esophagus and gastric cardia has risen dramatically in the past two decades in developed nations (Li and Morbahan, 2000; Walther et al., 2001). For reasons not entirely clear, body size and obesity are associated with AE and AC, with the increases in the incidence of these diseases paralleling the increases in obesity observed in Western cultures. An important feature of the observed associations between AE and AC and body size is that they are limited to adenocarcinomas and generally are not observed for squamous cell cancer or other histological types (Vaughan et al., 1995; Chow et al., 1998; Lagergren et al., 1999). However, one study reported a sixfold increased risk of squamous cell esophageal tumors and no increased risk of AE or AC adenocarcinomas among individuals with a high BMI (Kabat et al., 1993). In some studies, tumor location seems an important feature for AC, with associations stronger for tumors located more distally (Ji et al., 1997). Associations have remained after adjustment for important confounding factors, such as cigarette smoking and diet (Brown et al., 1995; Chow et al., 1998; Wu et al., 2001). Most studies show stronger associations for AE than for AC (Figure 22–2); risk estimates are inconsistent and range from the null to sixfold or greater with BMIs of 27 or higher. The study by Lagergren et al. (1999) revealed that associations were stronger for AE (OR = 7.6; 95% C.I., 3.8–15.2) than for AC (OR = 2.3; 95% C.I., 1.5–3.6); among people with a BMI of more than 30, the odds ratio was 16.2 for AE (95% C.I., 6.3–41.4). In some instances, the reported BMI level at which a risk is observed is modest. For example, in one study of AE in British women, a sixfold increased risk was observed for a BMI higher than 22.7 (Cheng et al., 2000). In the study by Chow et al. (1998), a three- to fourfold increase in risk of AE and a twofold increase in risk of AC was observed with a BMI of about 27.5 or greater. Limited data support a dose-response effect, with higher BMIs resulting in higher risk. Wu et al. (2001) observed a significant doseresponse, with risk increasing as BMI increased during all periods of adult life (i.e., for BMI at ages 20, 40, and recent adult BMI). Weight gain of 46 pounds or more was shown to be as important a predictor as weight itself (Ji et al., 1997; Chow et al., 1998). Body size at young age and 20 years before diagnosis also appear to be important contributors to risk (Ji et al., 1997; Lagergren et al., 1999). In some studies, the risk estimates observed for body size at younger ages are larger than those for recent body size. In the study of British women reported by Cheng et al. (2000) BMI at age 20 years appeared to be an important contributor to risk. In addition, associations have been reported as stronger for nonsmokers than smokers, and for younger people (<59) than older people, and as similar for those with and without gastroesophageal reflux disease, for men and women, and for various levels of education (Lagergren et al., 1999; Tretli et al., 1999).

Biological Mechanisms The underlying biological mechanisms for the association between BMI and AE/AC are not clear. Although it has been proposed that the mechanism involves esophageal reflux, limited data show that the risk

426

PART III: THE CAUSES OF CANCER

Figure 22–2. Summary of risk estimates and confidence intervals from epidemiologic studies of BMI and esophageal or gastric cardia adenocarcinoma in men and women.

associated with BMI is the same for those with and without symptomatic reflux disease (Lagergren et al., 1999). Other hypotheses for obesity potentially increasing risk include effects of diabetes or insulin-related mechanisms (Cheng et al., 2000), and the delay in esophageal motility associated with obesity (Li and Mobarhan, 2000).

RENAL CELL CANCER Summary of Findings Data from case-control and cohort studies provide convincing evidence of a consistent association of increased risk of renal cell cancer from overweight and obesity that is greater in women than in men. Estimates from a meta-analysis that calculated risks for men and women combined suggest this increase in risk corresponds to a 36% increase for an overweight person and an 84% increase for an obese person (Bergström et al., 2001).

Epidemiology Risk for renal cell cancer is increased in heavier men and women in virtually all studies (Wynder et al., 1974; McLaughlin et al., 1984, 1992; Goodman et al., 1986; Yu et al., 1986; Asal et al., 1988; Kadamini et al., 1989; Maclure and Willett, 1990; Partanen et al., 1991; McCredie and Stewart, 1992; Benhamou et al., 1993; Finkle et al., 1993; Kreiger et al., 1993; Hiatt et al., 1994; Lindblad et al., 1994; Mellemgaard et al., 1994, 1995; Muscat et al., 1995; Chow et al., 1996, 2000; Boeing et al., 1997; Yuan et al., 1998). The one exception is a hospital-based case-control study (Talamini et al., 1990) (Figures 22–3a and 22–3b). In contrast, no association between body size and tumors of the renal pelvis has been demonstrated (McCredie and Stewart, 1992; Chow et al., 2000). Most studies demonstrate a linear dose-response relationship between body weight or BMI and renal cell cancer. However, the increase in relative risk appears to be higher for women than for men, perhaps best demonstrated in the largest multicenter study that included men and women (Mellemgaard

et al., 1995). In that study, the relative risk for women in the top quartile of BMI was 2.2 compared to a risk of 1.2 for men in the top quartile of BMI. A similar pattern was observed in the very obese group, with higher increases in risk observed for women compared with men. For very obese women with a BMI in the top 5% (>38.1), risk was markedly increased to 3.6. In contrast, for a similar group of very obese men, risk was 1.6. In a meta-analysis including 11 studies, risk of incident renal cell cancer was increased by 6% and 7% for each unit increase in BMI for men and women, respectively (Bergström et al., 2001). This increase in risk corresponds to a 36% increase in risk for an overweight person and an 84% increase for an obese person. Data are limited on adult weight gain and renal cell cancer, and they have not been reported except for three studies. In a 1995 multicenter study, Mellemgaard et al. (1995) examined the slope and variability (coefficient of variation) of BMI, two measures that investigators have conceptualized as representing the rate of weight gain and frequency of weight fluctuations, respectively. In this study, the rate of weight change appeared to be an independent predictor of risk for women but not for men, though the coefficient of variation of BMI was not associated with renal cell cancer after adjustment for BMI in women. One other study, by Chow et al. (2000), examined the association of change in BMI over a period of 6 years and renal cell cancer risk in men and found a non-statistically significant increased risk of 1.6 with an increase in BMI of 2.5 units. Additionally, Bergström et al. (2001) observed that weight gain, especially among subjects with high BMI at age 20 (OR = 1.9, C.I., 1.2–3.0), was associated with an increased renal cancer risk. A number of other factors, including hypertension, diabetes mellitus, and tobacco use, have been consistently associated with risk of renal cell cancer. Each of these factors is related to obesity: Obesity is positively and causally related to hypertension and diabetes mellitus, and tobacco use is associated with leanness. A limited number of studies of obesity and renal cell cancer have adjusted for hypertension and/or smoking (Kreiger et al., 1993; Behnamou et al., 1993; Mellemgaard et al., 1994; Muscat et al., 1995; Chow et al., 1996,

427

Obesity and Body Composition

A

B

2000). However, none has fully explored potential interactions among these factors. Limited data suggest that risk from obesity is higher among heavier men and women with hypertension (Muscat et al., 1995) and who are nonsmokers (Chow et al., 2000). A limited number of studies have found amphetamines, used in the treatment of obesity and other conditions, to be positively associated with renal cell cancer (Wolk et al., 1996).

Figure 22–3. A, Summary of risk estimates and confidence intervals from epidemiologic studies of BMI and renal cell cancer in women. B, Summary of risk estimates and confidence intervals from epidemiologic studies of BMI and renal cell cancer in men.

et al., 1987), or through paracrine growth factors (Concolino et al., 1989). In a 1996 review on nutrition and renal cell cancer, Wolk et al. (1996) also proposed the hypothesis that the metabolic syndrome associated with upper body obesity, including hypertension, glucose intolerance, increased levels of insulin-like growth factor (IGF), and in women, anovulation and increased androgen production, might result in renal damage that would increase the susceptibility of the kidney to other carcinogens.

Biological Mechanisms The precise mechanisms by which higher body mass might influence renal cell carcinogenesis are not clearly defined. Hormonal factors, particularly sex steroids, have been found to promote renal cell proliferation and growth by direct endocrine receptor–mediated effects (Ronchi et al., 1984), by regulation of receptor concentrations (Traish

ENDOMETRIAL CANCER Summary of Findings Data from case-control and cohort studies provide convincing evidence of a consistent association of increased risk of endometrial

428

PART III: THE CAUSES OF CANCER

cancer from overweight and obesity with a doubling to tripling of risk observed among obese women. The lifetime risk of being diagnosed with endometrial cancer in the United States is about 3%, whereas the cumulative risk increases to 9% to 10% in obese women (Schottenfeld, 1995). Based on the prevalence of obesity (Flegal et al., 2002) and estimates of relative risk in U.S. women, excessive weight and a central pattern of fat distribution may account for 17% to 46% of endometrial cancer incidence in postmenopausal women. Data suggest that obesity may account for about 40% of the worldwide variation in cumulative rates of endometrial cancer (IARC, 2002). Meta-analyses that provide a more quantitative statement of the association across several studies have not been published.

Epidemiology Adenocarcinoma of the endometrium is the most common cancer of the female reproductive organs. International variation in endometrial cancer rates may represent differences in the distribution of known risk factors, which include obesity, hormone replacement therapy (HRT), ovarian dysfunction, diabetes mellitus, infertility, nulliparity, and tamoxifen use (Purdie et al., 2001). Comparisons of cumulative rates (ages 0–84) of endometrial cancer with reported data on the prevalence of obesity among women in developed and developing countries suggest that obesity may account for about 40% of the worldwide variation in cumulative rates of endometrial cancer (Akhmedkhanov et al., 2001). Thus, a substantial portion of the international variation in the incidence of endometrial cancer could be explained by differences in the prevalence of obesity. A consistent positive association between obesity and endometrial cancer risk has been observed in most cohort studies focusing on this association (Ewertz et al., 1984; Folsom et al., 1989, Tretli and Magnus, 1990; Le Marchand et al., 1991; Møller et al., 1994; Törnberg and Carstensen, 1994; Terry et al., 1999; Jain et al., 2000; Furberg and Thune, 2003). Excess risk persists even after adjustment for several factors, such as parity, HRT, and smoking (Le Marchand et al., 1991; Jain et al., 2000). Both Le Marchand et al. (1991) and

Figure 22–4. Summary of risk estimates and confidence intervals from epidemiologic studies of BMI and endometrial cancer.

Törnberg and Carstensen (1994) observed a stronger association between body mass index and endometrial cancer risk with increasing age than have others (Furberg and Thune, 2003). In a study of 47,003 women who were older than 55 years at entry and followed for 25 years, Törnberg and Carstensen (1994) observed a threefold increased risk (RR = 3.16; P < 0.0001) when comparing heavy to lean women (BMI ≥ 28 vs. BMI > 22). Consistent with the observation from cohort studies, most casecontrol studies (Blitzer et al., 1976; La Vecchia et al., 1984, 1991; Lapidus et al., 1988; Austin et al., 1991; Shu et al., 1991; Brinton et al., 1992; Levi et al., 1992; Shu et al., 1992; Swanson et al., 1993; Inoue et al., 1994; Olson et al., 1995; Baanders-van Halewyn et al., 1996; Kalandidi et al., 1996; Shoff and Newcomb et al., 1998) except three (Koumantaki et al., 1989; Parslov et al., 2000; Beard et al., 2000) observed that women with a BMI of 25 or higher have a two- to threefold increase in risk (Figure 22–4). High BMI was associated with an even greater excess risk in a study in Hawaii that included Japanese, Caucasian, native Hawaiian, Filipino, and Chinese populations (Goodman et al., 1997). Compared to lean women (BMI < 21.1), overweight women (BMI > 27.3) had a fourfold increased risk (RR = 4.3, P < 0.0001). Another large, informative Swedish case-control study with 709 endometrial cancer cases found an even larger increased risk of endometrial cancer with increasing weight (Weiderpass et al., 2000). Compared to lean women (BMI < 22.5), obese women (BMI of 30–33.99) had a threefold increased risk (OR = 2.9, C.I., 2.0–4.0), and markedly obese women (BMI > 34) a sixfold increased risk (OR = 6.3, C.I., 4.2–9.5). The effect of BMI did not vary by age, menopause, or use of contraceptives. Thus, a linear increase in risk with increasing BMI has been observed. Weight gain during adulthood is especially interesting, and studies show a linear dose-response relationship between weight gain and endometrial cancer (Le Marchand et al., 1991; Shu et al., 1992; Swanson et al., 1993; Olson et al., 1995; Terry et al., 1999). Two of these studies did not adjust for young adult weight or BMI (Swanson et al., 1993; Olson et al., 1995). The studies that adjusted for young adult or baseline weight or BMI found differing results. In two studies

Obesity and Body Composition that adjusted for either adolescent or early adult weight or BMI, the association between adult weight gain and endometrial cancer remained after adjustment (Le Marchand et al., 1991; Xu et al., 2002). In one other study, adjustment for baseline weight eliminated the association between adult weight gain and endometrial cancer (Terry et al., 1999). In one study in Shanghai, weight loss from ages 20 to 30 was inversely associated with endometrial cancer risk (Xu et al., 2002). The distribution of body fat, including such measures as skinfold (subscapular) and WHR, also has been examined in several studies (Folsom et al., 1989; Elliott et al., 1990; Austin, 1991; Schapira et al., 1991; Shu et al., 1992; Swanson et al., 1993). Evidence suggests that fat distribution may be important in endometrial cancer, with upperbody obesity particularly increasing risk. A case-control study from Shanghai found that the distribution of subscapular skinfolds was a better predictor for endometrial cancer risk than WHR, even after adjustment for BMI (Shu et al., 1992). Interaction between obesity and physical activity has been proposed as one explanation for the observation that some studies have observed a stronger positive association between obesity and endometrial cancer risk in older than younger women (Le Marchand et al., 1991; Törnberg and Carstensen, 1994). This finding is consistent with some population evidence that older women are less active than younger women (IARC, 2002). Recently, the established link between diabetes and endometrial cancer risk suggests that hyperglycemia (Furberg and Thune, 2003) and hyperinsulinemia (Weiderpass et al., 2000; Anderson et al., 2001) may influence risk.

Biological Mechanisms Changes in metabolism and hormonal activity that occur in obesity may be important mechanisms that explain the biologically plausible link between increased endometrial cancer risk and obesity. The normal menstrual cycle reflects the complex balance between the proliferative actions of estrogen and the antiestrogenic and secretory transforming actions of progesterone on the endometrium (Hale et al., 2002). A shift to a positive energy balance might contribute to an unfavorable sex hormone profile in women (Key et al., 2001; Jasienska et al., 2001). Among premenopausal women, obesity may induce a progesterone deficiency during the luteal phase as a result of anovulatory cycles, amenorrhea, and irregular menstrual periods. This situation may lead to an increased proliferation and decreased desquamation of endometrial cells. Among postmenopausal women, several lines of evidence suggest that obesity is associated with increased lifetime exposure to estrogen that may also increase risk: increased aromatization of androgens (androstenedione) to estrone in adipose tissue (Key and Pike, 1988a; Zeleniuch-Jacquotte, 2001), increased secretion of androstenedione in adrenal glands, and decreased sex hormone binding globulin (SHBG), resulting in an increase in bioavailable estradiol (free and albumin bound). Therefore, through different mechanisms for pre- and postmenopausal women, obesity alters sex steroid concentration and metabolism in ways that increase endometrial cancer risk. Experimental and epidemiological evidence suggest that insulin-like growth factor-I (IGF-I), a hormone with strong mitogenic and antiapoptotic actions, may be important in endometrial carcinogenesis. A high BMI may alter IGF-I blood levels, although this has not been examined in studies of endometrial cancer risk. In addition, energy restriction is known to enhance DNA repair, moderate oxidative damage to DNA, and reduce oncogene expression (Lipman et al., 1989). Another possible biological mechanism is related to the established link between diabetes and endometrial cancer (Furberg and Thune, 2003) and the fact that obesity is associated with insulin resistance, hyperinsulinemia, and diabetes mellitus. Evidence is increasing that insulin is a growth factor for tumor formation (Yu and Rohan, 2000). The mechanisms underlying insulin-mediated neoplasia may include enhanced DNA synthesis with resultant tumor cell growth, inhibition of apoptosis, and an altered sex hormone milieu. Hence, in overweight and obese women, coexisting metabolic disturbances may act synergistically to facilitate malignant transformation of glandular endometrial cells.

429

OVARIAN CANCER Summary of Findings The results from cohort and case-control studies do not indicate any clear association between overweight or obesity and epithelial ovarian cancer risk. Several recent studies, however, suggest a possible positive association.

Epidemiology Epithelial ovarian cancer incidence varies between countries and within countries and ethnic groups, but only a small percentage has been attributed to family history. The majority of ovarian cancer is sporadic, suggesting that part of the disease may be explained by modifiable lifestyle-related factors. In contrast to some cohort studies that found no association between older adult obesity and ovarian cancer (Møller et al., 1994; Törnberg et al., 1994; Mink et al., 1996; Fairfield et al., 2002), two studies (Lew and Garfinkel, 1979; Fairfield et al., 2002) have observed a direct association between obesity at some ages and ovarian cancer incidence. In a recent report from the Nurses Health Study with 402 cases of epithelial ovarian cancer (Fairfield et al., 2002), a twofold increase in premenopausal ovarian cancer risk was found among women with a BMI at age 18 of 25 or higher compared to a BMI less than 20. However, no association between recent BMI or adult weight change and ovarian cancer risk was observed in this study. An inconsistent association also has been observed in case-control studies (Figure 22–5). Only six of these studies (Casagrande et al., 1979; CASH, 1987; Farrow et al., 1989; Purdie et al., 1995; Mori et al., 1998; Lubin et al., 2003) found a direct association, whereas other studies found no association (Shu et al., 1989; Hartge et al., 1989; Mink et al., 1996; Ness et al., 2000; Greggi et al., 2000; Lukanova et al., 2002) and some a negative association (Byers et al., 1983; Parrazzini et al., 1997; Riman et al., 2001). One case-control study that found no association between BMI and ovarian cancer did observe a statistically significant increased risk with elevated WHR (Mink et al., 1996).

Biological Mechanisms Several reproductive factors, including the number of pregnancies, early menarche, late menopause, and late first pregnancy, influence ovarian cancer risk and could potentially play a role in the obesity–ovarian cancer risk association. All of these factors except for the number of pregnancies may increase risk. Thus, it is plausible that obesity may influence ovarian cancer risk through changes in sex steroids levels. However, at present the epidemiologic data have not established consistent direct association between overweight or obesity and ovarian cancer.

BREAST CANCER Summary of Findings Extensive data from case-control and cohort studies provide convincing evidence of a modest increased risk of postmenopausal breast cell cancer from overweight and obesity and a larger twofold increase in risk for adult weight gain. Estimates from a meta-analysis of cohort studies found gradual increases in risk of postmenopausal breast cancer to a BMI of 28, after which risk did not increase further. The relative risk for a BMI of 28 compared with a BMI of less than 21 was 1.26 (van den Brandt et al., 2000). However, risk estimates vary by age at diagnosis, history of HRT, and estrogen receptor status of the tumor. Another meta-analysis found that this increased risk for postmenopausal breast cancer corresponded to a 12% increase for an overweight woman and a 25% increase for an obese woman (Bergström et al., 2001). Data show a stronger association for adult weight gain, with a doubling of risk among women who have never used HRT and who gained over 20 kg from age 18 (Huang et al., 1997). A meta-analysis of cohort studies found an inverse association

430

PART III: THE CAUSES OF CANCER

Figure 22–5. Summary of risk estimates and confidence intervals from epidemiologic studies of BMI and ovarian cancer.

between BMI and premenopausal breast cancer, with a relative risk of 0.54 for women with a BMI higher than 31 compared with women with a BMI of less than 21 (van den Brandt et al., 2000). This estimate is consistent with the reduction in risk of 0.6 to 0.7 observed in many studies and does not appear to be present for BMIs less than 28. An extensive literature on breast cancer prognosis and survival suggests that women who are heavier at the time of diagnosis and gain weight following diagnosis experience a worse prognosis irrespective of menopausal status at diagnosis.

Epidemiology Animal model research in the 1930s first examined the hypothesis that a positive energy balance increased risk for breast cancer (Tannenbaum, 1940a, 1940b). The epidemiological evidence on weight or BMI and breast cancer risk varies by menopausal status (Figures 22–6a and 22–6b), age at diagnosis, hormone receptor status of the breast cancer, and exposure to exogenous estrogens. Because studies of breast cancer have used so many different BMI cutpoints that are not consistent with current WHO criteria of overweight and obese, the following review uses the term heavier women to describe the upper BMI groups rather than the terms overweight and obese. The most informative studies have distinguished between pre- and postmenopausal breast cancer; examined the effect of weight, weight gain, and central body fat at various ages; and have examined the differential effects of endogenous and exogenous estrogens. With the emergence of a possible IGF-mediated pathway for several cancers, recent studies have also begun to explore the potential interactions of IGF with body size. Most studies find that heavier women have a decreased risk of premenopausal breast cancer (Paffenbarger et al., 1980; Willett et al., 1985; Hislop et al., 1986; Le Marchand et al., 1988a; London et al., 1989; Swanson et al., 1989; Tretli, 1989; Bouchardy et al., 1990; Brinton et al, 1992; Harris et al., 1992; Pathak et al., 1992; Vatten and Kvinnsland, 1992; Törnberg and Carstensen, 1994; Francheschi et al., 1996; Swanson et al., 1996; Huang et al., 1997; Chie et al., 1998; Coates et al., 1999; Peacock et al., 1999; Enger et al., 2000; Hall et al., 2000; de Vasconcelos et al., 2001; Friedenreich et al., 2002). Other studies find no association between BMI and premenopausal breast cancer (Hirose et al., 2001; Shu et al., 2001; Yoo et al., 2001). Relative risks of approximately 0.6 to 0.7 have been reported whether

weight or BMI is assessed at the time of diagnosis or at earlier times during childhood, adolescence, or adulthood (Hislop et al., 1986; Kolonel et al., 1986; London et al., 1989; Brinton et al., 1992). Early studies suggested that the protective effect among heavier women was limited to early-stage disease due to poorer detection of small tumors (Willett et al., 1985; Swanson et al., 1989). However, subsequent studies in these same groups suggest that detection bias does not explain the increased risk for breast cancer observed among lean premenopausal women (London et al., 1989; Brinton et al., 1992; Swanson et al., 1996). A large case-control study of 1588 cases found that risk was increased about twofold among women who were tall and thin compared with women who were heavy and short (Swanson et al., 1996). A meta-analysis of seven cohorts comprising 723 incident cases of invasive breast cancer in premenopausal women found an inverse association between BMI and premenopausal breast cancer; the relative risk was 0.54 for women with a BMI higher than 31 compared to women with a BMI less than 21 (van den Brandt et al., 2000). This estimate is consistent with the reduction in risk of 0.6 to 0.7 observed in many studies and does not appear to be present for BMIs less than 28. Conversely, most studies have found that heavier women are at increased risk of postmenopausal breast cancer (Valaoras et al., 1969; de Waard et al., 1974; Choi et al., 1978; Paffenberger et al., 1980; Kalish, 1984; Lubin et al., 1985; Hislop et al., 1986; Kolonel et al., 1986; Le Marchand et al., 1988a; Negri et al., 1988; Tao et al., 1988; Ingram et al., 1989; Tretli et al., 1989; Folsom et al., 1990; Hseih et al., 1990; Parazzini et al., 1990; Chu et al., 1991; Harris et al., 1992; Pathak et al., 1992; Sellers et al., 1992, 2002; Radimer et al., 1993; Francheschi et al., 1996; Yong et al., 1996; Ziegler et al., 1996; Trentham-Dietz et al., 1997; Chie et al., 1998; Galanis et al., 1998; Magnusson et al., 1998; Enger et al., 2000; Hall et al., 2000; Lam et al., 2000; Li et al., 2000; Trentham-Dietz et al., 2000; Shu et al., 2001; Yoo et al., 2001; Friedenreich et al., 2002; Wenten et al., 2002; Lahmann et al., 2003). A meta-analysis of seven cohorts comprising 3208 cases of invasive postmenopausal breast cancer found gradual increases in risk to a BMI of 28 after which risk did not increase further; the relative risk for a BMI of 28 compared with a BMI of less than 21 was 1.26 (van den Brandt et al., 2000). The majority of studies on BMI and breast cancer risk have adjusted for major breast cancer risk factors, including reproductive factors. Few studies have examined in detail the effect of confounding or interactions with diet and

A

B

Figure 22–6. A, Summary of risk estimates and confidence intervals from epidemiologic studies of BMI and breast cancer in premenopausal women (limited to studies with at least

100 cases). B, Summary of risk estimates and confidence intervals from epidemiologic studies of BMI and breast cancer in postmenopausal women (limited to studies with at least 100 cases).

432

PART III: THE CAUSES OF CANCER

physical activity. Only one study in Vermont that had data on breast density from screening mammograms controlled for the effect of breast density (Lam et al., 2000). Because BMI is inversely related to breast density, adjustment for breast density resulted in an increase in the risk estimations at all levels of BMI; the OR increased from 1.9 to 2.5 after adjustment for breast density among obese women. When examined, risk estimates for the association between obesity and breast cancer vary by age at diagnosis, history of HRT, estrogen receptor status of the tumor, and possibly family history of breast cancer. Risk has been found to increase with age at diagnosis in some studies that include a substantial number of postmenopausal women older than 65 years (Yong et al., 1996; van den Brandt et al., 2000; Friedenreich et al., 2002). In one study, risk estimates increased from 1.1 among women younger than 60 years to 1.8 among women older than 65 years (Yong et al., 1996). The effect of exogenous estrogen or estrogen receptor status of tumors has been examined with stratified analyses in more recent studies. In these studies, obesity-related risk has been higher among women who have never used HRT (Huang et al., 1997; Lam et al., 2000; Friedenreich et al., 2002; Lahmann et al., 2003). One of the largest cohort studies in the United States found a statistically significant BMI and estrogen replacement interaction, with no increase in risk (RR of 1.1) among all women but an increase in risk (RR of 1.6) among heavier women who had not used HRT (Huang et al., 1997). At least three studies have examined risk by BMI and estrogen receptor status of the breast tumor (Enger et al., 2000; Yoo et al., 2001; Sellers et al., 2002). In one U.S. study, risk for a BMI of 27 compared with a BMI of 22 was 2.4 for tumors that were both estrogen and progesterone receptor positive (Enger et al., 2000). In another study, risk estimates were 2.0 and 2.2 for a BMI of 30.7 compared with a BMI of 23 for estrogen receptor positive and progesterone positive tumors, respectively (Sellers et al., 2002). In both of these studies, obesityrelated risk was not increased for estrogen and progesterone receptor negative tumors. In a Japanese study, risk did not vary by estrogen or progesterone status of the tumor (Yoo et al., 2001). However, the women in this study were lean; the upper quartile of BMI of 22 in this study is lower than the lowest quartile of BMI in most U.S. studies. Data are very limited on variation in BMI-related risk for postmenopausal breast cancer by family history. In several studies from the Iowa Women’s Health Study, heavier postmenopausal women with a family history of breast cancer have a greater risk of developing breast cancer than do heavier women without a family history (Sellers et al., 1992; Sellers et al., 2002). In one study in Japan, no differences were observed in associations of weight, BMI, or change in BMI by family history for premenopausal breast cancer, although somewhat stronger associations were seen for weight and change in BMI and postmenopausal breast cancer among postmenopausal women with a family history of breast cancer, confirming earlier results by Sellers and colleagues (Hirose et al., 2001). Only one study has examined variation in BMI-related risk for premenopausal breast cancer by family history of breast cancer (Swerdlow et al., 2002). In that study, the inverse association commonly observed between BMI and breast cancer risk was only observed in women without a family history of breast cancer. Weight or BMI at birth, during childhood, and early in adulthood have been examined relative to breast cancer. The data on birth weight and breast cancer are limited by a very small number of cases, with most studies having fewer than 100 cases. Some studies find no association (Le Marchand et al., 1988b; Ekbom et al., 1997), or a nonsignificant increased risk (Hilakivi-Clarke et al., 2001); others find an increase in risk with increasing birth weight for premenopausal but not postmenopausal breast cancer (Berstein, 1988; Ekbom et al., 1992; Michels et al., 1996; Sanderson et al., 1996; Innes et al., 2000) or a stronger increase in risk for premenopausal compared to postmenopausal breast cancer (De Stavola et al., 2000; Kaijser et al., 2001). In most studies, heavier weight or BMI during teenage and young adulthood (18–20 years) is associated with a 10% to 30% decrease in breast cancer risk for both pre- and postmenopausal breast cancer. This decreased risk is most often not statistically significant

(Paffenbarger et al., 1980; Willett et al., 1985; London et al., 1989; Folsom et al., 1990; Lund et al., 1990; Chu et al., 1991; Brinton and Swanson, 1992; Sellers et al., 1992; Ursin et al., 1995; Huang et al., 1997; Magnusson et al., 1998; Coates et al., 1999; Peacock et al., 1999; Enger et al., 2000; Hirose et al., 2001; de Vasconcelos et al., 2001; Wenten et al., 2002). During the middle decades of life, the risk associated with BMI remains inverse for premenopausal breast cancer and increases with age for postmenopausal breast cancer. Increases in central adiposity have been associated with a 1.4- to 2.0-fold increase in breast cancer risk among postmenopausal women in most studies (Ballard-Barbash et al., 1990a; Folsom et al., 1990; Schapira et al., 1990; Bruning et al., 1992a, 1992b; Sellers et al., 1992; Ng, 1997; Shu et al., 2001; Friedenreich et al., 2002). However, not all studies show this association (den Tonkelaar et al., 1992; Petrek, 1993; Lahmann et al., 2003). Data on central adiposity and premenopausal breast cancer do not suggest a consistent association between measures, such as waist circumference or WHR, and breast cancer risk (Francheschi et al., 1996; Mannisto et al., 1996; Swanson et al., 1996; Ng et al., 1997; Kaaks et al., 1998; Huang et al., 1999; Sonnenschein et al., 1999; Hall et al., 2000; Shu et al., 2001; Friedenreich et al., 2002). Only five of these studies observed statistically significant increases in risk (Mannisto et al., 1996; Ng et al., 1997; Sonnenschein et al., 1999; Hall et al., 2000; Shu et al., 2001). The association in postmenopausal women may be modified by a family history of breast cancer and ovarian cancer. In the Iowa Women’s Health Study, among women with elevated WHR, only women with a positive family history of breast cancer were at increased risk. The combination of a high WHR with a family history of breast and ovarian cancer was associated with a more than fourfold increase in risk of breast cancer (Folsom et al., 1990; Sellers et al., 1993). The most consistent body size predictor of postmenopausal breast cancer risk is adult weight gain (Lubin, 1985; Le Marchand et al., 1988b; Ballard-Barbash, 1990b; Folsom et al., 1990; Brinton et al., 1992; Radimer et al., 1993; Barnes-Josiah et al., 1995; Ziegler et al., 1996; Huang et al., 1997; Magnusson et al., 1998; Jernström and BarrettConnor, 1999; Enger et al., 2000; Li et al., 2000; Trentham-Dietz et al., 2000; Shoff et al., 2000; Hirose et al., 2001; Shu et al., 2001; Friedenreich et al., 2002; Wenten et al., 2002; Lahmann et al., 2003). This association has been seen in cohort studies that found no association between BMI at baseline and subsequent development of breast cancer and that also adjusted for baseline BMI (Ballard-Barbash et al., 1990a; Folsom et al., 1990; Huang et al., 1997). Findings from one of the largest cohort studies suggest that a doubling of risk was associated with a weight gain of more than 20 kg after the age of 18 years, but only among women who had never used postmenopausal HRT (Huang et al., 1997). Other studies have observed similar results, with increases in risk either limited to or much larger among women who have gained more weight and who have never used HRT compared to current users (Harris et al., 1992; Magnusson et al., 1998; Trentham-Dietz et al., 2000; Friedenreich et al., 2002; Lahman et al., 2003). Consistent with findings for BMI and premenopausal breast cancer, weight gain appears to be associated with a reduced or no significant increased risk of premenopausal breast cancer in most studies (Le Marchand et al., 1988a; London et al., 1989; Brinton et al., 1992; Huang et al., 1997; Coates et al., 1999; Peacock et al., 1999; Hirose et al., 2001; Shu et al., 2001; de Vasconcelos et al., 2001; Friedenreich et al., 2002; Wenten et al., 2002). However, a study by Wenten et al. (2002) in New Mexico found differences in risk for Hispanic compared to non-Hispanic white women. In that study, no association was found between weight gain and risk of premenopausal breast cancer among non-Hispanic white women; in contrast, a non-statistically significant but nearly twofold increased risk was observed with more than 14 kg of weight gain among Hispanic white women. The only study that has examined the effect of percent body fat measured by bioelectric impedance as well as several other measures of body size, fat mass, and distribution found the strongest association for percent body fat, with a doubling of risk for women with a percent body fat of over 36% compared to women with a percent body fat of less than 27%. Similar to results reported for BMI and weight gain, risk for percent body fat was stronger among women who had

Obesity and Body Composition never used HRT. Among women with a percent body fat of more than 36%, the risk estimate among women who did not use HRT was 3.4 compared to 1.0 among women using HRT (Lahmann et al., 2003).

Breast Cancer Prognosis Extensive data indicate that heavier women experience poorer survival and increased likelihood of recurrence in most studies, irrespective of menopausal status and after adjustment for stage and treatment (Greenberg et al., 1985; McNee et al., 1987; Hebert et al., 1988; Mohle-Boetani et al., 1988; Lees et al., 1989; Verrault et al., 1989; Coates et al., 1990; Tretli et al., 1990; Kyogoku et al., 1990; Vatten et al., 1991; Senie et al., 1992; Giuffrida et al., 1992; Bastarrachea et al., 1994; Zhang et al., 1995; den Tonkelaar et al., 1995; Maehle and Tretli, 1996). The effect of weight or BMI on prognosis appears to be limited to or more pronounced among women with stage I and II disease (Verreault et al., 1989; Tretli and Magnus, 1990), estrogen receptor and progesterone receptor positive status (Coates et al., 1990; Giuffrida et al., 1992; Maehle and Tretli, 1996), and negative nodes (Mohle-Boetani et al., 1988; Newman et al., 1997). The most precise risk estimates for BMI and breast cancer prognosis are derived from large population-based cohorts of breast cancer cases. In the largest cohort of more than 8000 women with breast cancer, risk varied by stage at diagnosis (Tretli and Magnus, 1990). Among women with stage I disease, women in the upper quintile of BMI had a 70% increased risk of dying from breast cancer. Among women with stage II disease, women in the upper quintile had a 40% increased risk. BMI was not associated with risk among women with late stage III and IV disease (Tretli and Magnus, 1990). In a subset of 1238 women from this cohort who had unilateral breast cancer treated with modified radical mastectomy and were followed for 15 years, the risk of dying from breast cancer relative to BMI varied markedly by hormone receptor status (Maehle and Tretli, 1996). Although women with estrogen receptor and progesterone receptor positive tumors had a 46% reduced risk of dying from breast cancer, the risk within hormone receptor positive and negative groups varied by BMI. Among women with hormone receptor positive tumors, obese women had a threefold higher risk of death than did thin women. Conversely, among women with hormone receptor negative tumors, thin women had a sixfold higher risk of death than did obese women, even after adjustment for lymph node status, tumor diameter, and mean nuclear area. One study that examined the association of BMI with distant recurrence and death found a 70% to 80% increase in risk of both distant recurrence and death among women with a BMI of 27.8 or greater (Goodwin et al., 2002a). This study also examined the association of fasting insulin, IGF, and estradiol to these outcomes and found increased risk of distant recurrence and death among women with elevated insulin, but not with elevated IGF-1, IGF-II, and estradiol. A subsequent study in this same sample found that the binding proteins of IGFBP-1 and IGFBP-3 were inversely associated with risk of distant recurrence, but only IGFBP-1 was also inversely associated with risk of death (Goodwin et al., 2002b). Weight gain is reported in the majority of women undergoing adjuvant therapy for breast cancer (Heasman et al., 1985; Goodwin et al., 1988, 1999; Camoriano et al., 1990; Demark-Wahnefried, 1993). Weight gain associated with treatment is lowest among women not receiving systemic therapy, intermediate among women receiving hormonal therapy, and more pronounced among women receiving adjuvant chemotherapy and those who undergo menopause after diagnosis and treatment. To identify optimal interventions to prevent weight gain during treatment, research has begun to examine whether changes in energy intake and expenditure during and after treatment are associated with weight gain (Demark-Wahnefried et al., 1993, 2001; De Waard et al., 1993). Although data on the association of postdiagnosis weight gain and prognosis are limited, the largest study of 646 premenopausal women found that women who gained more than 5.9 kg were 1.5 times as likely to relapse and 1.6 times more likely to die than women who gained less weight (Camoriano et al., 1990).

433

Biological Mechanisms Ovarian Hormone Metabolism The major focus of research on mechanisms underlying body size and breast and endometrial cancer risk has been the effects of adiposity on endogenous hormonal, predominantly estrogen, metabolism; this research has been extensively reviewed (Kirschner et al., 1981; Key and Pike, 1988b; Bernstein and Ross, 1993; Key et al., 2001). Hypotheses have evolved from a focus on estrogen excess, to the combined effects of estrogen and progesterone, and most recently, to attempts to delineate factors defining bioavailability and effects of estrogens and androgens and their metabolites on specific end organs. Increases in overall and central adiposity have been associated with increases in insulin, androgens, and triglycerides; decreases in SHBG; and increases in total and free estradiol (Evans et al., 1984; Haffner et al., 1989; Kirschner et al., 1990; Kaye et al., 1991; Bruning et al., 1992a, 1992b; Potischman et al., 1996). A number of these hormonal changes increase the bioavailability of estradiol and its metabolites and may also directly promote tumor growth. The bioavailability of estradiol is dependent on the degree of binding and the strength of binding to several protein carriers. SHBG is the predominant protein carrier of estradiol and the percentage of free estradiol is generally inversely related to the level of SHBG. However, estradiol is also transported and bound, though less tightly, to albumin. Increases in free fatty acids, such as triglycerides, has been reported to increase the level of free estradiol by displacing estradiol from SHBG where it is tightly bound, to albumin, where it is less tightly bound. Therefore, both decreases in SHBG and increases in triglycerides may result in increases in free estradiol. Key and Pike (1988a, 1988b) first proposed that the effect of adiposity on the bioavailability of estrogen was modulated by menopausal changes in estrogen and progesterone production, and so explained the apparent contradictory findings for premenopausal and postmenopausal breast cancer. Before menopause, ovarian estrogen production overwhelms changes in estrogen metabolism related to the overall level of adiposity. Consequently, estradiol in ovulatory cycles does not differ measurably in obese compared to lean women. In contrast, estradiol levels are reduced in anovulatory cycles that are more frequent in obese than lean premenopausal women. Conversely, obese premenopausal women have been found to have markedly reduced progesterone levels, both due to increased frequency of anovulation and decreased progesterone production in the luteal phase. With the onset of menopause, the decreased risk associated with premenopausal obesity declines over time. In postmenopausal women, the overall level of adiposity results in increases in estrogenic activity by increases in estrogen production from androgens (Kirschner et al., 1981), decreased estrogen-protein binding (Bruning, 1987) due to decreases in SHBG (Moore et al., 1987; Kaye et al., 1991; Bruning et al., 1992b) and increases in triglycerides (Bruning et al., 1992b). Furthermore, the C-2 hydroxy metabolite of estrogen has been proposed to be a less tumorigenic metabolite of estrogen, and C-2 versus 16-hydroxylation of estradiol is reported to be decreased in obese women (Schneider et al., 1983). A well-designed study using contemporary, high-quality assays for sex steroids found that the increases in estrone, estradiol, free estradiol, and albumin-bound estradiol associated with increases in BMI were not present in premenopausal women but were statistically significant among postmenopausal women (Potischman et al., 1996).

Insulin, Insulin-like Growth Factors, and Other Hormones Insulin, insulin-like growth factors, and binding proteins (BPs) may promote hormone-dependent tumors through direct effects on tumor cells (Foekens et al., 1989; Yee et al., 1989; Turner et al., 1997), and through indirect effects on estrogens and possible interactions with estrogen at the estrogen receptor on breast cancer cells (Clayton et al., 1997; El-Tanani et al., 1997). A comprehensive review of epidemiological research by Yu and Rohan (2000) summarized the mitogenic and antiapoptotic actions of IGFs on various cancer cells, their syngergistic effects with other growth factors and steroids, and the associations of IGFs and binding proteins with cancer. The research on

434

PART III: THE CAUSES OF CANCER

breast cancer is briefly reviewed here. Five of seven studies that have examined the effect of IGF-I have found an increased risk of premenopausal breast cancer associated with increased IGF-I (Peyrat et al., 1993; Bruning et al., 1995; Del Giudice et al., 1998; Hankinson et al., 1998; Toniolo et al., 2000). However, only one of these studies (Bruning et al., 1995) found a statistically significant increased risk. Three studies found no association between IGF-1 and postmenopausal breast cancer (Bruning et al., 1995; Hankinson et al., 1998; Jernström and Barrett-Connor, 1999), and one study found a difference in mean IGF-I in cases and controls (Agurs-Collins et al., 2000). C-peptide, a marker of hyperinsulinemia, has been associated with breast cancer risk in two studies (Bruning et al., 1992a; Yang et al., 2001). One study that observed an increase risk of postmenopausal breast cancer with weight gain did not find that this increased risk was explained by levels of IGF-I, fasting insulin, proinsulin, or C-peptide (Jernström and Barrett-Conner, 1999). Few other studies have examined whether insulin, IGF, or related growth factors might explain part of the increased risk associated with BMI, weight gain, or central body fat. Data on the association between IGF-I and BMI or fat mass are mixed (Yu and Rohan, 2000), with some studies finding no association (Kelly et al., 1990; Landin-Wilhelmsen et al., 1994; Goodman-Gruen et al., 1997; Janssen et al., 1998; Kaklamani et al., 1999), some an inverse association (Colletti et al., 1991; Veldhuis et al., 1995; Marin et al., 1993; Maccario et al., 1999), and some, particularly in lower BMI ranges, a positive association (IARC, 2002). However, IGF-1 is hypothesized to be associated with the degree of muscle mass, and is related to the preservation of muscle mass in animal studies. Studies have not examined whether increases in muscle mass also may explain the increased premenopausal breast cancer risk observed among tall and lean women.

Other Hormonal Hypotheses Animal studies have examined the effect of various other hormonal measures, such as growth hormone, or cortisol, on breast cancer development. These factors may vary by body weight, fat mass, or body fat distribution, but studies in humans have not examined these as potential mechanisms for weight-related breast cancer risk. One study has examined the association of leptin, a hormone that reflects total fat mass, with premenopausal breast cancer and found a non-significant lower level of leptin in cases compared to controls (Mantzoros et al., 1999). This finding is consistent with the inverse association between BMI and premenopausal breast cancer. No published studies have examined associations of leptin with postmenopausal breast cancer. Circulating leptin is closely related to insulin levels and adiposity. It increases the amount of adipose tissue in the body (Bennett et al., 1997; Niskanen et al., 1997) by regulating food intake and energy balance (Larsson et al., 1998; van Aggel-Leijssen et al., 1999) and interacting with other endocrine systems (Licino et al., 1997; Haffner et al., 1997). Insulin increases leptin gene expression, stimulates leptin protein production in rodents, and regulates leptin and SHBG protein levels in vivo and in vitro (Segal et al., 1996; Johannsson et al., 1998; Russell et al., 1998; Ho et al., 1999).

Other Mechanisms With the recent emergence of studies suggesting that physical activity may reduce risk of breast cancer, a limited number of analyses have examined the possibility of effect modification or interaction between physical activity and various weight-related measures (Thune et al., 1997). Although some studies have included physical activity in a large list of covariates within multivariate models to adjust for confounding, no analyses have reported whether effect modification by physical activity is observed in risk estimates for weight-related measures and breast cancer. Similarly, although diet has been proposed as a major mediator of the association of weight or BMI with postmenopausal breast cancer risk, no analyses have been designed to examine this issue. The large error in self-reporting of energy, greater in heavier compared to lighter individuals (Heitman, 1993, Subar et al., 2003), limits the ability to examine this issue with currently available data.

PROSTATE CANCER Summary of Findings Extensive data from cohort and case-control studies provide convincing evidence of no association between body mass index and prostate cancer. Data on other anthropometric measures, such as abdominal obesity, muscle mass, or weight gain, are too limited to allow for any definitive conclusion.

Epidemiology The association of BMI and incidence of prostate cancer has been examined in 33 studies conducted worldwide. Of these, 13 are cohort studies (Figure 22–7) (Nomura et al., 1985; Severson et al., 1988; Mills et al., 1989; Chyou et al., 1994; Le Marchand et al., 1994; Thune and Lund, 1994; Andersson et al., 1997; Cerhan et al, 1997; Giovannucci et al., 1997; Lund Nilsen and Vatten, 1999; Clarke and Whittemore, 2000; Habel et al., 2000; Putnam et al., 2000; Schuurman et al., 2000; Lee et al., 2001). The remaining 20 are casecontrol studies (Wynder et al., 1971; Graham et al., 1983; Talamini et al., 1986; Ross et al., 1987; Kolonel et al., 1988; Yu et al., 1988; Mettlin et al., 1989; West et al., 1991; Andersson et al., 1995; Whittemore et al., 1995; Andersson et al., 1996; Grönberg et al., 1996; Ilic et al., 1996; Key et al., 1997; Demark-Wahnefried et al., 1997a; Hayes et al., 1999; Hsieh et al., 1999; Villeneuve et al., 1999; Bairati et al., 2000; Hsing et al., 2000; Spitz et al., 2000; Sharpe and Siemiatycki, 2001). Of these 33 incidence studies, seven observed an increased prostate cancer risk among men who were in the highest category of BMI (Thune and Lund, 1994; Grönberg et al., 1996; Andersson et al., 1997; Key et al., 1997; Putnam et al., 2000; Spitz et al., 2000) or weight (Chyou et al., 1994) as compared to those in the lowest category. In these seven studies, the increased risk was small, 10% on average, across these categories of BMI, and the range in risk estimates was from 0.5 to 4.4. The remaining studies found no association between obesity and increased prostate cancer risk. Most epidemiological studies of obesity and prostate cancer have used BMI as a measure of obesity. However, the ranges for BMI used in these studies have not been standardized. Residual confounding that might have concealed weak associations with anthropometry is a possibility in these studies, particularly because several did not have a full examination of all possible confounding factors. Nonetheless, given the large number of studies conducted, the consistency of the results, the low magnitude of the association, and the limited evidence of a dose-response effect, the lack of association between body mass index and prostate cancer is convincing. In addition to examining BMI, several of these investigations evaluated the possible associations of weight with prostate cancer risk. Consistently, no associations were found for increased risk with weight. It is possible, however, that other anthropometric measures may be more strongly associated with prostate cancer risk. Body mass index measures both lean body mass and adiposity. Other anthropometric measures of body fat distribution may be more strongly related to prostate cancer risk because this cancer is androgen-dependent, and lean body mass is related to androgen levels. Fewer studies have examined lean body mass or abdominal adiposity, thereby limiting any conclusive assessments of the associations. Severson et al. (1988) found an increased prostate cancer risk in a Japanese population associated with the area of muscle in the arm and not with the fat in the arm. A Norwegian cohort study found no increased risk with lean body mass (Lund Nilsen and Vatten, 1999). Similarly, an American study found no association with lean body mass or fat mass area as measured in the arm (Clarke and Whittemore, 2000). Weight gain throughout the lifetime and increased abdominal adiposity also are possible prostate cancer risk factors, although few studies have evaluated these variables. Schuurman et al. (2000) found a significant positive trend for increasing quantiles of BMI at age 20, but no increased risk for high weight gain from age 20 to time of interview. Hsing et al. (2000) found that abdominal adiposity, as measured by WHR, increased prostate cancer risk in Chinese men. The observed increased risk with WHR was independent of BMI, socioeconomic

435

Obesity and Body Composition

Figure 22–7. Summary of risk estimates and confidence intervals from epidemiologic studies of BMI and prostate cancer.

status, physical activity, and dietary intake, including total calories. Though the associations with WHR were strong and consistent in this study, additional empirical evidence is needed to confirm this association. Some effect modification by stage of disease may exist for obesity and prostate cancer. In the Health Professionals Follow-up Study, Giovannucci et al. (1997) observed a 58% increased risk among men with metastatic prostate cancer for those in the highest quintile of WHR, though no such increased risk was found for all stages of prostate cancer combined. In this same study, preadult obesity appeared to protect against the occurrence of prostate cancer, and these associations were particularly evident for metastatic disease (Giovannucci et al., 1997). Further effect modification may occur with race. In a study in the United States with African American and white men, Hayes et al. (1999) found an increased risk of advanced prostate cancer among black men in the highest quartile of BMI at age 25. No similarly increased risks were found for white men or for men with less advanced prostate cancer for either race. Additional factors that may modify the effect of this putative association have not been fully explored. One such factor is smoking. A case-control study in Canada found an interaction between smoking and BMI. Men with high BMI who were smokers were at increased risk as compared to nonsmokers within the same BMI category (Sharpe and Siemiatycki, 2001).

Biological Mechanisms Prostate cancer is a hormone-mediated cancer in that endogenous hormones regulate the growth and function of the prostate gland (Henderson et al., 1982), and administration of large quantities of testosterone can induce prostate cancer in rodents (Noble, 1977). Men who have high estrogen levels rarely develop prostate cancer (Glantz, 1964). A Western lifestyle, characterized by a positive energy balance as a result of high-energy intake and low activity levels, has consistently been implicated as a risk factor for prostate cancer in international comparisons and experimental studies (Hebert et al., 1998; Bosland et al., 1999; Mukherjee et al., 1999). The evidence, however, from individual-level epidemiologic studies on the associations between dietary intake, obesity, and physical activity and prostate cancer risk is weaker (Friedenreich and Thune, 2001; Schulman et al., 2001). Despite the lack of an association between BMI and prostate cancer from epidemiologic studies, several studies of biological markers that are related to obesity suggest an association of these markers with prostate cancer. Leptin, an adipocyte-derived hormone that regulates satiety and energy expenditure, has been shown to increase prostate cancer risk. It is suggested as a key link between a state of continual positive energy balance and the transition from premalignant lesions to overt prostate cancer (Stattin et al., 2001). In addition, higher serum insulin

436

PART III: THE CAUSES OF CANCER

levels increase prostate cancer risk independent of abdominal adiposity (Hsing et al., 2001). Despite these associations, it is unclear whether the concomitant hormonal and metabolic changes that occur because of abdominal adiposity and insulin resistance are the biological mechanisms for the link between insulin and prostate cancer risk. Two other biomarkers for prostate cancer risk that may be related to BMI and either fat or muscle mass have been identified from studies in diverse populations: IGF-I (Mantzoros et al., 1997; Chan et al., 1998; Wolk et al., 1998) and serum testosterone (Shaneyfelt et al., 2000). A meta-analysis of published studies on hormonal predictors of prostate cancer risk found that men with serum testosterone or IGF-1 levels in the upper quartile of the population distribution have an approximately twofold higher prostate cancer risk (Shaneyfelt et al., 2000). Levels of dihydrotestosterone and estradiol were not found to have as strong an association with prostate cancer risk (Shaneyfelt et al., 2000). Overall, exposure to endogenous androgenic hormones appears to be positively associated with prostate cancer risk (Bosland, 2000). A number of studies have explored the associations between various sex steroids and BMI in men (Glass, 1989; Kato et al., 1992; Andersson et al., 1997; Demark-Wahnefried et al., 1997b; Tymchuk et al., 1998; Tamani et al., 2001; Sarma et al., 2002). However, how BMI may play a role in the association of these measures with prostate cancer is complex and not well understood.

mones. Proposed mechanisms include increases in obesity-related postmenopausal estrogen based on evidence that exogenous estrogens are weakly associated with thyroid cancer risk (La Vecchia et al., 1999; Negri et al., 1999). In addition, detection bias due to more frequent examination of the thyroid gland because of screening for hypothyroidism among heavier women may be a factor in countries, such as the United States, where this practice may be more prevalent.

CANCERS OF THE LUNG AND THE HEAD AND NECK Overview An extensive literature exists on the commonly observed inverse association between obesity and cancers of the lung and the head and neck. This literature was summarized in the 2002 International Agency for Research on Cancer (IARC) report on weight control and physical activity (IARC, 2002). In the past 10 years, studies examining the inverse association of obesity and these cancers have been designed to examine the extent of confounding by the effect of tobacco use or preexisting disease that may result in weight loss.

Summary of Findings THYROID CANCER Summary of Findings Data from case-control and cohort studies are suggestive of an increased risk of thyroid cancer among heavier individuals (BMI cutpoints not defined) that is present only for women. Estimates from a meta-analysis suggest a 20% increase in risk in women (Dal Maso et al., 2000).

Epidemiology A meta-analysis (Dal Maso et al., 2000) summarized and reanalzyed evidence from the existing case-control studies on obesity and thyroid cancer (McTiernan et al., 1984; Preston-Martin et al., 1987; Ron et al., 1987; Linos et al., 1989; Kolonel et al., 1990; Francheschi et al., 1991; Levi et al., 1991; Goodman et al., 1992; Glattre et al., 1993; PrestonMartin et al., 1993; Wingren et al., 1993; Hallquist et al., 1994; D’Avanzo et al., 1995; Galanti et al., 1997; Negri et al., 1999). The authors reported that relative risks for the upper tertile of BMI at the time of diagnosis was above unity for 9 of the 12 studies in women, with a relative risk overall for the highest compared to the lowest tertile of 1.2 (95% CI, 1.0–1.4), P for trend of 0.04. The overall relative risk for the highest compared to the lowest tertile in men was 1.0, with a nonsignificant P for trend. Similarly, in five case-control studies that have examined this factor, no association has been observed for thyroid cancer risk and BMI between the ages of 17 to 20 (McTiernan et al., 1984; Preston-Martin et al., 1987; Linos et al., 1989; Preston-Martin et al., 1993; Wingren et al., 1993). One case-control study of papillary thyroid cancer in women (Rossing et al., 2000), published after the Dal Maso (2000) meta-analysis, found that women who weighed 185 pounds or more had a 1.5-fold increased risk. Another case-control study in women found no association between BMI and thyroid cancer (Mack et al., 2002) but did observe a non-statistically significant increase in risk with weight gain. The only cohort study to examine the association between obesity and thyroid cancer risk found no statistically significant association with BMI at the baseline interview or with weight gain since age 20 (Iribarren et al., 2001).

Biological Mechanisms Several mechanisms have been proposed for the modest association observed in many case-control studies in women. Most relate to potential interactions between thyroid hormones and other steroid hor-

Data from case-control and cohort studies suggest an inverse association between BMI and cancers of the lung and head and neck. This inverse association moves to the null after adjustment for confounding by tobacco use and, therefore, is not thought to be causal in nature.

Lung Cancer The majority of cohort (Knekt et al., 1991; Lee and Paffenbarger, 1992b; Chyou et al., 1994; Drinkard et al., 1995; Kark et al., 1995; Henley et al., 2002) and case-control studies (Kabat and Wynder, 1992; Goodman and Wilkins, 1993; Kubik et al., 2002; Olson et al., 2002) have found an inverse association between body size and lung cancer, although it is generally less striking after adjustment for potential confounding by smoking status. In addition, several recent studies have examined the association of BMI and lung cancer in subgroups such as current and nonsmokers (Rauscher et al., 2000; Henley et al., 2002) and found little association among never or former smokers. A study limited to never and former smokers in New York (Rauscher et al., 2000) found a doubling of risk for lung cancer among obese men and women that was not statistically significant. A cohort study from the American Cancer Society’s Cancer Prevention Study II, which examined the effect of the potential for bias from tobacco use and preexisting disease, also evaluated the role of weight and BMI among never smokers and found no association (Henley et al., 2002). This study showed a lessening in the observed association between BMI and lung cancer among the entire cohort after adjusting for preexisting disease and removing early follow-up as an additional method of adjusting for preexisting disease. These findings are interpreted to indicate that leanness is unlikely to be a causal factor in the development of lung cancer. However, it is likely that the causal factors for lung cancer among never smokers are very different than causal factors among current smokers. Therefore, it is not clear that analyses of associations between BMI and lung cancer among never smokers can contribute much to understanding whether confounding by tobacco use explains the majority of the inverse association observed between BMI and lung cancer among smokers. The only study that has examined the association of body fat distribution by waist circumference and lung cancer risk found evidence of a differential association between waist circumference and histologic subtypes of lung cancer (Olson et al., 2002). In that study among middle-aged Iowa women, risk for small cell and squamous cell lung cancer was increased by threefold among women in the highest quintile of waist circumference; in contrast, no association was observed between waist circumference and adenocarcinoma of the lung.

Obesity and Body Composition

Head and Neck Cancers Although tobacco and alcohol use account for more than 90% of cancers of the head and neck in developing countries (IARC, 1986, 1988), leanness also has been associated with increased risk of head and neck cancers in some case-control studies. No cohort studies examining this association have been published. Six case-control studies found an inverse association between weight and/or BMI and cancer of the oral cavity and pharynx (McLaughlin et al., 1988; Marshall et al., 1992; Day et al., 1993; Kabat et al., 1994; D’Avanzo et al., 1996; Francheschi et al., 2001) and one found a non-statistically significant risk estimate of 1.5 among heavier women (Negri et al., 2000). Two other studies found weaker inverse associations between BMI and laryngeal cancer (Muscat and Wynder, 1992; D’Avanzo et al., 1996). Similar to findings for lung cancer, several studies found either no association (Kabat et al., 1994; Francheschi et al., 2001) or weaker associations (D’Avanzo et al., 1996) between BMI and oral cancer among never smokers compared to current smokers. These results suggest that BMI is not causally related to risk of head and neck cancers. However, similar to lung cancer, the causal factors contributing to head and neck cancer among never smokers may be different than those among current or former smokers.

CANCER SITES WITH INSUFFICIENT EVIDENCE FOR CONCLUSIONS Limited epidemiologic evidence exists on weight or BMI for a number of other cancers, including leukemia, non-Hodgkin lymphoma, malignant melanoma, and testicular, pancreatic, bladder, and cervical cancers. However, data for these and other cancer sites are too limited to allow specific conclusions to be made and are not summarized here due to space limitations. A summary of the evidence for several of these other cancer sites is provided in the February 2002 IARC publication Weight Control and Physical Activity (IARC, 2002).

DATA ON CANCER MORTALITY Because of space limitations and the difficulty of disentangling the effect of screening and treatment from individual-based risk factors, such as obesity, when examining cancer mortality as an outcome, this review does not provide a detailed overview of the evidence on obesity and cancer mortality. However, a recent study from a large U.S.-based cohort provides several new insights (Calle et al., 2003). In this study, the effect of obesity on cancer mortality was examined in a cohort of more than 900,000 adults who developed more than 57,000 cancers in 16 years of follow-up. The study is unique in that it provided estimates for multiple cancers within one cohort and one report and had a much larger sample size than any other single report to date. Therefore, the study was able to provide estimates for multiple cancers, including more precise estimates for less common cancers, and examined the effect of overweight and obesity among nonsmokers. The study found that BMI was associated with higher rates of death in both men and women for cancers of the esophagus, colon and rectum, liver, gallbladder, pancreas, kidney, non-Hodgkin lymphoma, and multiple myeloma. BMI was also associated with higher rates of death from cancers of the stomach and prostate in men and cancers of the breast, uterus, cervix, and ovary in women. Based on rates of overweight and obesity present in the United States in 1999 to 2000, the study estimated that 14% of cancer deaths in men and 20% of cancer deaths in women could be attributed to overweight and obesity.

FUTURE DIRECTIONS Studies are needed that have sufficient power and complete data on a number of factors to adequately control for all confounders of the association between BMI and cancer, and to study subgroups postulated

437

to be at very low or high risk. In addition, more studies on weight and height at an early age in life with a long follow-up are needed to clarify the role of weight and weight gain throughout the life span. For some cancer sites, a number of studies corrected only for age, thus allowing uncontrolled confounding to influence the results. Important confounders to consider include physical activity, other anthropometric factors such as height, smoking status, dietary intake, and other lifestyle factors. A full examination of all anthropometric risk factors and cancer is also needed. For some cancer sites, such as breast cancer, additional measures such as body fat distribution, weight gain, and body composition have been examined. However, for most cancer sites, very few studies have considered any factor other than BMI. Moreover, studies that examine a full range of possible effect modifiers need to be conducted because relatively few investigations have examined risks within subgroups of the population defined on characteristics such as ethnic origin/race, other lifestyle risk factors (e.g., physical activity, dietary intake, smoking status), HRT, genetic factors (e.g., family history of cancer or other comorbid diseases, such as diabetes that may influence the health consequences of obesity), or other possible risk factors, such as breast density in the case of breast cancer. Studies that provide data specific to racial and ethnic groups are increasingly important. At present, most of the data on obesity and cancer are based on population samples drawn from Europe, the United States, Canada, Australia, Japan, and China, and therefore have largely included whites of European descent, Chinese, Japanese, and Hawaiians. Data are very limited for African-American men and women or those of North, Central, or South American Latina descent or for many populations in Asia and Africa. In addition, data on these groups are seldom reported separately even within large case-control or cohort studies that may include these populations. To clarify the extent to which anthropometric factors are risk factors for specific cancers, further research is needed to delineate the role of site-specific biological mechanisms, particularly for endogenous hormones, insulin, insulin-related growth factors, leptin, cortisol, and related hormones that may influence the development of body fat, skeletal structure, and musculature and that may underlie observed associations with cancer outcomes. Evidence is evolving rapidly about genetic factors that may have particular importance in the metabolism of sex steroids and insulin or in the development of obesity. Future studies are needed to examine how specific genes predict underlying metabolic profiles and subsequent cancer risk. An overview of the potential biological mechanisms that have been either hypothesized or examined as possible explanations for the association of obesity with specific cancers is presented in Table 22–1. This overview demonstrates that many of these cancers are hypothesized to be hormone based, either in terms of sex steroids or, more recently, in terms of insulin, leptin, or insulin-related growth factors. Although not explored to any extent in humans, research in animal models suggest that factors such as cortisol or vitamin D may also explain some of the risk associated with obesity. Using animal models, these mechanisms have largely been explored by studies of caloric restriction rather than by studies of increases in energy expenditure (Hursting et al., 2003). Two other areas of emerging research should be briefly noted. With improved detection and treatment, many people are living longer and healthier lives with cancer, and therefore, studies on the effect of obesity on cancer prognosis are needed. At present, research in prognosis has been largely limited to breast cancer and should be extended to other cancers. In addition, to achieve a better understanding of what interventions may reduce risk, studies of the effect of specific dietary and exercise programs to control weight are needed, particularly studies that can elucidate the effect of such interventions on underlying mechanisms that may influence carcinogenesis.

EXTENT OF THE OBESITY EPIDEMIC AND MANAGEMENT STRATEGIES The WHO 2000 report Obesity: Preventing and Managing the Global Epidemic is the most recent comprehensive source summarizing the

438

PART III: THE CAUSES OF CANCER

Table 22–1. Biological Mechanisms Potentially Involved in the Association Between Obesity and Specific Cancers Mechanism Altered endogenous estrogen levels Altered endogenous androgen levels Altered metabolic hormone levels Interaction between sex steroid and metabolic hormones Altered nervous and immune system functioning Increased cell number Positive energy balance Concentration of adverse factors in adipose tissue Increased gastric reflux Delayed esophageal transit time Increased incidence of metabolic syndrome (hypertension, glucose intolerance, increased IGF, and in women anovulation and increased androgens)

Explanation

Cancer Sites

Higher estrogen, higher testosterone, and lower SHBG Higher estrogen, lower testosterone, and lower SHBG Elevated leptin and insulin, increased IGF, higher cholesterol

Breast, endometrial, renal Prostate Prostate, breast, endometrium, colon, esophageal, gastric Prostate, breast, endometrium, colon

Greater sympathetic activity, immune system dysfunction Larger pool of cells to undergo malignant transformation High dietary fat intake, lower physical activity Greater concentration of growth factors or carcinogens Increased intra-abdominal fat mass increases intraabdominal pressure and hence increases gastric reflux Obesity leads to delayed esophageal transit time and increased exposure time Contributes to renal damage and susceptibility to other carcinogens

Prostate, breast Breast, colon, theoretically relevant to all All All Esophageal Esophageal Renal

IGF, insulin-like growth factor; SHBG, sex hormone binding globulin.

global epidemic of obesity. Within that report, the WHO MONICA study provides a full spectrum of data comparing rates of obesity worldwide for the period 1983–1986. At that time, the international prevalence of obesity, defined as a BMI of 30 and higher, ranged from 5% to 20% for men and 10% to 40% in women (WHO, 2000). The WHO 2000 report highlighted several major features of the international patterns of body weight and associated health outcomes. For the first time, more people were classified as experiencing obesity than were suffering from starvation in developing as well as developed countries. The report also noted the rapidly increasing rates of obesity in some Asian countries, such as Japan and China, where the prevalence of obesity traditionally had been very low. Similar to patterns observed in the United States, rates of obesity in other countries are generally higher among women than men and in urban as compared to rural communities. Although data on children are more limited, evidence suggests that obesity is also increasing in children worldwide in developed and developing countries. In 2000, the WHO initiated a global strategy for preventing and controlling noncommunicable diseases that includes a focus on combatting the worldwide rise in obesity through reducing unhealthy diets and physical inactivity. In the United States, the prevalence of obesity has changed dramatically over the past 40 years. It was relatively stable at approximately 10% for men and 15% for women during the early 1960s to the late 1970s. During the late 1980s and early 1990s rates of obesity increased, and the most current estimates from the 1999–2000 National Health and Nutrition Examination Survey (NHANES) indicate that the prevalence of obesity has increased to 27.5% for men and 33.4% for women (Flegal et al., 2002). Rates tend to increase with age to about age 70; at later ages, they decline slightly. Rates are highest among non-Hispanic black women who experience a 50% prevalence of obesity. When the prevalence of overweight, or a BMI of 25 and higher, is considered, about 65% of the U.S. adult population is affected. Of particular concern are increases in rates of overweight among children and adolescents. Prevalence rates, which were approximately 5% during the 1960s, have tripled to over 15% in 1999–2000 among school-aged children and adolescents. Rates rose by 10 percentage points or more between 1988 and 1994 and between 1999 and 2000 for both Mexican American and non-Hispanic black adolescents (Ogden et al., 2002). Estimating the consequences and cost of obesity internationally is complex because obesity and overweight are risk factors for multiple chronic diseases other than cancer, including non–insulin-dependent diabetes mellitus, insulin resistance, gallbladder disease, hyperlipi-

demia, coronary heart disease, hypertension, osteoarthritis, sleep apnea, low-back pain, polycystic ovarian disease, impaired fertility, reproductive hormone abnormalities, surgical complications, and anesthesia complications. International estimates of the economic cost have ranged from 2% to about 7% of national health care costs, with the United States reporting the highest cost of 7% (WHO, 2000). In 1990, this cost was estimated to be $45.8 billion dollars from the direct cost of obesity-associated diseases and $23.3 billion from lost productivity (Wolf and Colditz, 1994). In an updated analysis, Wolf and Colditz (1998) estimated the total U.S. direct cost attributable to obesity to be $99.2 billion, which represented 5.7% of the total U.S. national health expenditure that year. A comprehensive overview of the treatment and prevention of obesity is beyond the scope of this review. These issues are addressed in detail in the WHO 2000 report Obesity: Preventing and Managing the Global Epidemic and other reviews (Kopelman, 2000). Advances in the field of obesity, particularly in terms of the genetics of obesity and its implication for developing more targeted and effective treatment approaches, have been highlighted in issues of Science (February 2003) and Nature (April 2000). Preventing and controlling obesity requires comprehensive societal efforts and must target lifelong health habits at the individual level. In addition, continued research is needed to find better means of identifying individuals at risk of developing obesity and its adverse consequences because of their genetic, familial, or other environmental risk factors. At present, with the rapid increase in obesity in developed and developing countries among all age groups, including children, prevention has become an urgent worldwide priority. Issues of particular concern in cancer control include the need to develop better strategies to facilitate the avoidance of weight gain during adult life and to identify whether specific interventions during treatment can alleviate treatment-related weight gain that has been identified to worsen breast cancer prognosis, but may also be an issue for other cancers. Key public health recommendations and strategies relevant to cancer are reviewed in the 2002 IARC report on Weight Control and Physical Activity (IARC, 2002).

CONCLUSIONS AND POPULATION ATTRIBUTABLE RISK At present, data provide convincing evidence of a positive association of overweight and obesity with cancers of the colon (among men), renal cell, postmenopausal breast, endometrium, and probable evidence of a positive association with colon cancer (among women),

Obesity and Body Composition

439

Table 22–2. Population Attributable Rates (PAR in %) for Weight-Related Cancers for the European Union (EU) from 1982 to 1996 and the United States from 1999 to 2000 Men

Women

Overweight

Obese

Overweight

Obese

6.9 15.2

4.2 10.3

5.0 11.1 17.2 4.1

5.7 13.4 22.0 4.5

9.1 19.4

8.5 19.0

8.5 18.3 26.8 6.9

9.8 21.7 33.4 7.6

eu* Colon Renal Endometrium Breast (postmenopausal)

usa Colon Kidney Endometrium Breast (postmenopausal)

Overweight/obesity prevalence rates used: EU overweight and obesity rates were, respectively, 35% and 19% for men, and 50% and 13% for women. For the United States from 1999 to 2000, the rates were increased to 67% and 28% for men and 62% and 33% for women. *Source of RR and EU PARs is Bergström et al. (2000).

adenocarcinoma of the esophagus and gastric cardia, and thyroid cancer (among women). On the other hand, epidemiologic studies do not demonstrate an association of weight or BMI with prostate cancer incidence. In addition, limited evidence does not suggest any consistent direct association between overweight or obesity and ovarian cancer. The consistent inverse association between overweight or obesity and lung, and head and neck, cancers is reduced with adjustment for tobacco use; therefore, overweight and obesity are not considered to be protective for these cancers. The inverse association between overweight and obesity and premenopausal breast cancer persists after adjustment for multiple potential confounding factors. Extensive data provide convincing evidence of a positive association between overweight and obesity and adverse changes in breast cancer prognosis and mortality. Estimates of the population attributable risk (PAR) of cancer due to overweight and obesity have been summarized in the 2002 IARC report and are modest for some cancers, such as colon and postmenopausal breast cancer. Approximately 9% to 11% of these cancers are attributable to overweight or obesity. PAR estimates are more substantial for renal cell, esophageal, and endometrial cancer: 25%, 37%, and 39%, respectively. However, these estimates were based on international rates of overweight and obesity from the 1990s (IARC, 2002) and are higher in the United States today given the continued increase in the prevalence of overweight and obesity. Table 22–2 provides a comparison of PARs for five major weight-related cancers for the European Union as estimated by Bergström and colleagues (2000) and for the United States based on rates of obesity and overweight from the 1999–2000 NHANES survey. The PARs estimated by Bergström and colleagues (2001) were based on relative risk estimates from metaanalyses and from rates of obesity in European Union countries from 1982 to 1996. Because of increases in rates of overweight and obesity in the United States, PARs are higher for the United States than those estimated for the European Union. Figure 22–8 shows the potential for substantial increases in PARs with increases in rates of overweight/obesity and relative risk estimates of overweight/obesity with cancer. For example, in a country such as Japan, which has obesity rates of less than 3%, the PAR for a cancer with a twofold increase due to obesity, such as endometrial cancer, is less than 3%. The PAR estimate for obesity and endometrial cancer within the European Union published by Bergström and colleagues (2000) was 22% when the prevalence of obesity was 13% for women within the European Union. In contrast, in the United States, whose rate of obesity in women is 33%, the PAR for endometrial cancer is 33%. The potential for marked increases in the PAR for obesity and weight-

Figure 22–8. Population attributable risk (PAR) by overweight/obesity prevalence rates and levels of relative risk. Source: Population Attributable Risk (PAR) from Bergstrom et al., 2001 for renal cell cancer. Updated PAR estimates are based on NHANES 1999–2000 obesity prevalence. PARs are 䊏䊐 US, EU males and 䊉䊊 US, EU females, respectively. 䊉 RR_1.5; 䉱 RR_2.0; 䊏 RR_3.0.

related cancers is further demonstrated in Figure 22–9, which demonstrates the relatively strong association between rates of obesity and endometrial cancer prevalence internationally. Similar associations can be seen for other cancers as well. With the expectation that the epidemic of obesity is likely to continue, if not accelerate, in the near term, overweight and obesity will become increasingly important contributors to cancer risk internationally. Acknowledgments The authors would like to acknowledge the contributions of the following people to this review: Anita Ambs, National Cancer Institute, for managing the distribution of materials for the review and the development of figures; David Tran, Scientific Consulting Group, for literature reviews; Darrell Anderson, Scientific Consulting Group, for development of summary tables, figures, and final preparation of the manuscript; and Anne Rodgers, for editing.

Figure 22–9. Association between obesity prevalence (BMI ≥ 30 kg/m2) and endometrial cancer cumulative rates in selected countries.

440

PART III: THE CAUSES OF CANCER

References Agurs-Collins T, Adams–Campbell LL, Kim KS, Cullen KJ. 2000. Insulin-like growth factor-1 and breast cancer risk in postmenopausal AfricanAmerican women. Cancer Detect Prev 24:199–206. Akhmedkhanov A, Zeleniuch–Jacquotte A, Toniolo P. 2001. Role of exogenous and endogenous hormones in endometrial cancer: Review of the evidence and research perspectives. Ann N Y Acad Sci 943:296–315. Anderson KE, Anderson E, Mink PJ, Hong CP, Kushi LH, Sellers TA, Lazovich D, Folsom AR. 2001. Diabetes and endometrial cancer in the Iowa women’s health study. Cancer Epidemiol Biomarkers Prev 10:611–616. Andersson SO, Baron J, Wolk A, Lindgren C, Bergström R, Adami HO. 1995. Early life risk factors for prostate cancer: A population-based case-control study in Sweden. Cancer Epidemiol Biomarkers Prev 4:187–192. Andersson SO, Baron J, Bergström R, Lindgren C, Wolk A, Adami HO. 1996. Lifestyle factors and prostate cancer risk: A case-control study in Sweden. Cancer Epidemiol Biomarkers Prev 5:509–513. Andersson SO, Wolk A, Bergström R, Adami HO, Engholm G, Englund A, Nyren O. 1997. Body size and prostate cancer: A 20-year follow-up study among 135006 Swedish construction workers. J Natl Cancer Inst 89:385–389. Asal NR, Risser DR, Kadamani S, Geyer JR, Lee ET, Cherng N. 1988. Risk factors in renal cell carcinoma: I. Methodology, demographics, tobacco, beverage use, and obesity. Cancer Detect Prev 11:359–377. Austin H, Austin JM, Jr, Partridge EE, Hatch KD, Shingleton HM. 1991. Endometrial cancer, obesity, and body fat distribution. Cancer Res 51:568–572. Bairati I, Larouche R, Meyer F, Moore L, Fradet Y. 2000. Lifetime occupational physical activity and incidental prostate cancer. Canada. Cancer Causes Control 11:759–764. Ballard–Barbash R, Schatzkin A, Carter CL, Kannel WB, Kreger BE, D’Agostino RB, Splansky GL, Anderson KM, Helsel WE. 1990a. Body fat distribution and breast cancer in the Framingham Study. J Natl Cancer Inst 82:286–290. Ballard-Barbash R, Schatzkin A, Taylor PR, Kahle LL. 1990b. Association of change in body mass with breast cancer. Cancer Res 50:2152–2155. Ballard-Barbash R. 1999. Energy balance, anthropometrics and cancer. In: Heber D, Blackburn GL, Go VLW, Holland JL, eds. Nutritional Oncology. San Diego: Academic Press, pp. 137–153. Barnes-Josiah D, Potter JD, Sellers TA, Himes JH. 1995. Early body size and subsequent weight gain as predictors of breast cancer incidence. Iowa, United States. Cancer Causes Control 6:112–118. Bastarrachea J, Hortobagyi GN, Smith TL, Kau SW, Buzdar AU. 1994. Obesity as an adverse prognostic factor for patients receiving adjuvant chemotherapy for breast cancer. Ann Intern Med 120:18–25. Baumgartner RN, Heymsfield SB, Roche AF. 1995. Human body composition and the epidemiology of chronic disease. Obes Res 3:73–95. Beard CM, Hartmann LC, Keeney GL, Crowson CS, Malkasian GD, O’Brien PC, Melton LJ, III. 2000. Endometrial cancer in Olmsted County, MN: Trends in incidence, risk factors and survival. Ann Epidemiol 10:97–105. Benhamou S, Lenfant MH, Ory-Paoletti C, Flamant R. 1993. Risk factors for renal-cell carcinoma in a French case-control study. Int J Cancer 55:32–36. Bennett FI, McFarlane-Anderson N, Wilks R, Luke A, Cooper RS, Forrester TE. 1997. Leptin concentration in women is influenced by regional distribution of adipose tissue. Am J Clin Nutr 66:1340–1344. Bergström A, Pisani P, Tenet V, Wolk A, Adami HO. 2001. Overweight as an avoidable cause of cancer in Europe. Int J Cancer 91:421–430. Bernstein L, Ross RK. 1993. Endogenous hormones and breast cancer risk. Epidemiol Rev 15:48–65. Berstein LM. 1988. Newborn macrosomy and cancer. Adv Cancer Res 50: 231–278. Blitzer PH, Blitzer EC, Rimm AA. 1976. Association between teen-age obesity and cancer in 56,111 women: all cancers and endometrial carcinoma. Prev Med 5:20–31. Boeing H, Schlehofer B, Wahrendorf J. 1997. Diet, obesity and risk for renal cell carcinoma: results from a case control-study in Germany. Z Ernahrungswiss 36:3–11. Bosland MC, Oakley-Girvan I, Whittemore AS. 1999. Dietary fat, calories, and prostate cancer risk. J Natl Cancer Inst 91:489–491. Bosland MC. 2000. The role of steroid hormones in prostate carcinogenesis. J Natl Cancer Inst Monogr 27:39–66. Bostick RM, Potter JD, Kushi LH, Sellers TA, Steinmetz KA, McKenzie DR, Gapstur SM, Folsom AR. 1994. Sugar, meat, and fat intake, and nondietary risk factors for colon cancer incidence in Iowa women. United States. Cancer Causes Control 5:38–52. Bouchardy C, Le MG, Hill C. 1990. Risk factors for breast cancer according to age at diagnosis in a French case-control study. J Clin Epidemiol 43:267–275.

Boutron-Ruault MC, Senesse P, Meance S, Belghiti C, Faivre J. 2001. Energy intake, body mass index, physical activity, and the colorectal adenomacarcinoma sequence. Nutr Cancer 39:50–57. Brinton LA, Berman ML, Mortel R, Twiggs LB, Barrett RJ, Wilbanks GD, Lannom L, Hoover RN. 1992. Reproductive, menstrual, and medical risk factors for endometrial cancer: Results from a case-control study. Am J Obstet Gynecol 167:1317–1325. Brinton LA, Swanson CA. 1992. Height and weight at various ages and risk of breast cancer. Ann Epidemiol 2:597–609. Brown LM, Swanson CA, Gridley G, Swanson GM, Schoenberg JB, Greenberg RS, Silverman DT, Pottern LM, Hayes RB, Schwartz AG. 1995. Adenocarcinoma of the esophagus: Role of obesity and diet. J Natl Cancer Inst 87:104–109. Bruning PF. 1987. Endogenous estrogens and breast cancer a possible relationship between body fat distribution and estrogen availability. J Steroid Biochem 27:487–492. Bruning PF, Bonfrer JM, van Noord PA, Hart AA, Jong-Bakker M, Nooijen WJ. 1992a. Insulin resistance and breast-cancer risk. Int J Cancer 52:511–516. Bruning PF, Bonfrer JM, Hart AA, van Noord PA, van der Hoeven H, Collette HJ, Battermann JJ, Jong-Bakker M, Nooijen WJ, de Waard F. 1992b. Body measurements, estrogen availability and the risk of human breast cancer: A case-control study. Int J Cancer 51:14–19. Bruning PF, Van Doorn J, Bonfrer JM, van Noord PA, Korse CM, Linders TC, Hart AA. 1995. Insulin-like growth-factor-binding protein 3 is decreased in early-stage operable pre-menopausal breast cancer. Int J Cancer 62:266–270. Byers T, Marshall J, Graham S, Mettlin C, Swanson M. 1983. A case-control study of dietary and nondietary factors in ovarian cancer. J Natl Cancer Inst 71:681–686. Caan BJ, Coates AO, Slattery ML, Potter JD, Quesenberry CP Jr, Edwards SM. 1998. Body size and the risk of colon cancer in a large case-control study. Int J Obes Relat Metab Disord 22:178–184. Calle EE, Rodriguez C, Walkr-Thurmond K, Thune MJ. 2003. Overweight, obesity and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med 348:1625–1638. Camoriano JK, Loprinzi CL, Ingle JN, Therneau TM, Krook JE, Veeder MH. 1990. Weight change in women treated with adjuvant therapy or observed following mastectomy for node-positive breast cancer. J Clin Oncol 8:1327–1334. Casagrande JT, Louie EW, Pike MC, Roy S, Ross RK, Henderson BE. 1979. “Incessant ovulation” and ovarian cancer. Lancet 2:170–173. CASH STUDY. 1987. The reduction in risk of ovarian cancer associated with oral–contraceptive use. The Cancer and Steroid Hormone Study of the Centers for Disease Control and the National Institute of Child Health and Human Development. N Engl J Med 316:650–655. Cerhan JR, Torner JC, Lynch CF, Rubenstein LM, Lemke JH, Cohen MB, Lubaroff DM, Wallace RB. 1997. Association of smoking, body mass, and physical activity with risk of prostate cancer in the Iowa 65+ Rural Health Study. United States. Cancer Causes Control 8:229–238. Chan JM, Stampfer MJ, Giovannucci E, Gann PH, Ma J, Wilkinson P, Hennekens CH, Pollak M. 1998. Plasma insulin-like growth factor-I and prostate cancer risk: A prospective study. Science 279:563–566. Cheng KK, Sharp L, McKinney PA, Logan RF, Chilvers CE, Cook-Mozaffari P, Ahmed A, Day NE. 2000. A case-control study of oesophageal adenocarcinoma in women: A preventable disease. Br J Cancer 83:127– 132. Chie WC, Li CY, Huang CS, Chang KJ, Lin RS. 1998. Body size as a factor in different ages and breast cancer risk in Taiwan. Anticancer Res 18:565–570. Choi NW, Howe GR, Miller AB, Matthews V, Morgan RW, Munan L, Burch JD, Feather J, Jain M, Kelly A. 1978. An epidemiologic study of breast cancer. Am J Epidemiol 107:510–521. Chow WH, McLaughlin JK, Mandel JS, Wacholder S, Niwa S, Fraumeni JF Jr. 1996. Obesity and risk of renal cell cancer. Cancer Epidemiol Biomarkers Prev 5:17–21. Chow WH, Blot WJ, Vaughan TL, Risch HA, Gammon MD, Stanford JL, Dubrow R, Schoenberg JB, Mayne ST, Farrow DC, Ahsan H, West AB, Rotterdam H, Niwa S, Fraumeni JF Jr. 1998. Body mass index and risk of adenocarcinomas of the esophagus and gastric cardia. J Natl Cancer Inst 90:150–155. Chow WH, Gridley G, Fraumeni JF Jr, Jarvholm B. 2000. Obesity, hypertension, and the risk of kidney cancer in men. N Engl J Med 343: 1305–1311. Chu SY, Lee NC, Wingo PA, Senie RT, Greenberg RS, Peterson HB. 1991. The relationship between body mass and breast cancer among women enrolled in the Cancer and Steroid Hormone Study. J Clin Epidemiol 44: 1197–1206.

Obesity and Body Composition Chyou PH, Nomura AM, Stemmermann GN. 1994. A prospective study of weight, body mass index and other anthropometric measurements in relation to site-specific cancers. Int J Cancer 57:313–317. Clarke G, Whittemore AS. 2000. Prostate cancer risk in relation to anthropometry and physical activity: the National Health and Nutrition Examination Survey I Epidemiological Follow-Up Study. Cancer Epidemiol Biomarkers Prev 9:875–881. Clayton SJ, May FE, Westley BR. 1997. Insulin-like growth factors control the regulation of oestrogen and progesterone receptor expression by oestrogens. Mol Cell Endocrinol 128:57–68. Coates RJ, Clark WS, Eley JW, Greenberg RS, Huguley CM Jr, Brown RL. 1990. Race, nutritional status, and survival from breast cancer. J Natl Cancer Inst 82:1684–1692. Coates RJ, Uhler RJ, Hall HI, Potischman N, Brinton LA, Ballard-Barbash R, Gammon MD, Brogan DR, Daling JR, Malone KE, Schoenberg JB, Swanson CA. 1999. Risk of breast cancer in young women in relation to body size and weight gain in adolescence and early adulthood. Br J Cancer 81:167–174. Colletti RB, Copeland KC, Devlin JT, Roberts JD, McAuliffe TL. 1991. Effect of obesity on plasma insulin-like growth factor-I in cancer patients. Int J Obes 15:523–527. Concolino G, Lubrano C, Santonati A, Di Silverio F, Catizone A, Conti C. 1989. Epidermal growth factor (EGF) binding by normal and neoplastic human renal tissue. J Tumor Marker Oncol 4:89–97. D’Avanzo B, La Vecchia C, Franceschi S, Negri E, Talamini R. 1995. History of thyroid diseases and subsequent thyroid cancer risk. Cancer Epidemiol Biomarkers Prev 4:193–199. D’Avanzo B, La Vecchia C, Talamini R, Franceschi S. 1996. Anthropometric measures and risk of cancers of the upper digestive and respiratory tract. Nutr Cancer 26:219–227. Dal Maso L, La Vecchia C, Franceschi S, Preston-Martin S, Ron E, Levi F, Mack W, Mark SD, McTiernan A, Kolonel L, Mabuchi K, Jin F, Wingren G, Galanti MR, Hallquist A, Glattre E, Lund E, Linos D, Negri E. 2000. A pooled analysis of thyroid cancer studies. V. Anthropometric factors. Cancer Causes Control 11:137–144. Dales LG, Friedman GD, Ury HK, Grossman S, Williams SR. 1979. A casecontrol study of relationships of diet and other traits to colorectal cancer in American blacks. Am J Epidemiol 109:132–144. Davidow AL, Neugut AI, Jacobson JS, Ahsan H, Garbowski GC, Forde KA, Treat MR, Waye JD. 1996. Recurrent adenomatous polyps and body mass index. Cancer Epidemiol Biomarkers Prev 5:313–315. Day GL, Blot WJ, Austin DF, Bernstein L, Greenberg RS, Preston-Martin S, Schoenberg JB, Winn DM, McLaughlin JK, Fraumeni JF Jr. 1993. Racial differences in risk of oral and pharyngeal cancer: Alcohol, tobacco, and other determinants. J Natl Cancer Inst 85:465–473. De Stavola BL, Hardy R, Kuh D, Silva IS, Wadsworth M, Swerdlow AJ. 2000. Birthweight, childhood growth and risk of breast cancer in a British cohort. Br J Cancer 83:964–968. de Vasconcelos AB, Azevedo e Silva Mendonca G, Sichieri R. 2001. Height, weight, weight change and risk of breast cancer in Rio de Janeiro, Brazil. Sao Paulo Med J 119:62–66. de Waard F, Baanders-van Halewijn EA. 1974. A prospective study in general practice on breast-cancer risk in postmenopausal women. Int J Cancer 14:153–160. de Waard F, Ramlau R, Mulders Y, de Vries T, van Waveren S. 1993. A feasibility study on weight reduction in obese postmenopausal breast cancer patients. Eur J Cancer Prev 2:233–238. Del Giudice ME, Fantus IG, Ezzat S, McKeown-Eyssen G, Page D, Goodwin PJ. 1998. Insulin and related factors in premenopausal breast cancer risk. Breast Cancer Res Treat 47:111–120. Demark-Wahnefried W, Winer EP, Rimer BK. 1993. Why women gain weight with adjuvant chemotherapy for breast cancer. J Clin Oncol 11:1418– 1429. Demark-Wahnefried W, Conaway MR, Robertson CN, Mathias BJ, Anderson EE, Paulson DF. 1997a. Anthropometric risk factors for prostate cancer. Nutr Cancer 28:302–307. Demark-Wahnefried W, Rimer BK, Winer EP. 1997b. Weight gain in women diagnosed with breast cancer. J Am Diet Assoc 97:519–526, 529. Demark-Wahnefried W, Peterson BL, Winer EP, Marks L, Aziz N, Marcom PK, Blackwell K, Rimer BK. 2001. Changes in weight, body composition, and factors influencing energy balance among peremenopausla breast cancer patients receiving adjuvant chemotherapy. J Clin Oncol 19:2367– 2369. den Tonkelaar I, de Waard F, Seidell JC, Fracheboud J. 1995. Obesity and subcutaneous fat patterning in relation to survival of postmenopausal breast cancer patients participating in the DOM-project. Breast Cancer Res Treat 34:129–137.

441

Dietz AT, Newcomb PA, Marcus PM, Storer BE. 1995. The association of body size and large bowel cancer risk in Wisconsin (United States) women. Cancer Causes Control 6:30–36. Drinkard CR, Sellers TA, Potter JD, Zheng W, Bostick RM, Nelson CL, Folsom AR. 1995. Association of body mass index and body fat distribution with risk of lung cancer in older women. Am J Epidemiol 142:600–607. Ekbom A, Trichopoulos D, Adami HO, Hsieh CC, Lan SJ. 1992. Evidence of prenatal influences on breast cancer risk. Lancet 340:1015–1018. Ekbom A, Hsieh CC, Lipworth L, Adami HQ, Trichopoulos D. 1997. Intrauterine environment and breast cancer risk in women: a population–based study. J Natl Cancer Inst 89:71–76. El Tanani MK, Green CD. 1997. Interaction between estradiol and growth factors in the regulation of specific gene expression in MCF-7 human breast cancer cells. J Steroid Biochem Mol Biol 60:269–276. Elliott EA, Matanoski GM, Rosenshein NB, Grumbine FC, Diamond EL. 1990. Body fat patterning in women with endometrial cancer. Gynecol Oncol 39:253–258. Enger SM, Ross RK, Paganini-Hill A, Carpenter CL, Bernstein L. 2000. Body size, physical activity, and breast cancer hormone receptor status: results from two case-control studies. Cancer Epidemiol Biomarkers Prev 9:681–687. Evans DJ, Hoffmann RG, Kalkhoff RK, Kissebah AH. 1984. Relationship of body fat topography to insulin sensitivity and metabolic profiles in premenopausal women. Metabolism 33:68–75. Ewertz M, Machado SG, Boice JD Jr, Jensen OM. 1984. Endometrial cancer following treatment for breast cancer: A case-control study in Denmark. Br J Cancer 50:687–692. Fairfield KM, Willett WC, Rosner BA, Manson JE, Speizer FE, Hankinson SE. 2002. Obesity, weight gain, and ovarian cancer. Obstet Gynecol 100:288–296. Farrow DC, Weiss NS, Lyon JL, Daling JR. 1989. Association of obesity and ovarian cancer in a case-control study. Am J Epidemiol 129:1300–1304. Finkle WD, McLaughlin JK, Rasgon SA, Yeoh HH, Low JE. 1993. Increased risk of renal cell cancer among women using diuretics in the United States. Cancer Causes Control 4:555–558. Flegal KM, Carroll MD, Ogden CL, Johnson CL. 2002. Prevalence and trends in obesity among US adults, 1999–2000. JAMA 288:1723–1727. Foekens JA, Portengen H, Janssen M, Klijn JG. 1989. Insulin-like growth factor-1 receptors and insulin-like growth factor-1-like activity in human primary breast cancer. Cancer 63:2139–2147. Folsom AR, Kaye SA, Potter JD, Prineas RJ. 1989. Association of incident carcinoma of the endometrium with body weight and fat distribution in older women: Early findings of the Iowa Women’s Health Study. Cancer Res 49:6828–6831. Folsom AR, Kaye SA, Prineas RJ, Potter JD, Gapstur SM, Wallace RB. 1990. Increased incidence of carcinoma of the breast associated with abdominal adiposity in postmenopausal women. Am J Epidemiol 131:794–803. Ford ES. 1999. Body mass index and colon cancer in a national sample of adult US men and women. Am J Epidemiol 150:390–398. Franceschi S, Levi F, Negri E, Fassina A, La Vecchia C. 1991. Diet and thyroid cancer: A pooled analysis of four European case-control studies. Int J Cancer 48:395–398. Franceschi S, Favero A, La Vecchia C, Baron AE, Negri E, Dal Maso L, Giacosa A, Montella M, Conti E, Amadori D. 1996. Body size indices and breast cancer risk before and after menopause. Int J Cancer 67:181–186. Franceschi S, Dal Maso L, Levi F, Conti E, Talamini R, La Vecchia C. 2001. Leanness as early marker of cancer of the oral cavity and pharynx. Ann Oncol 12:331–336. Friedenreich CM, Thune I. 2001. A review of physical activity and prostate cancer risk. Cancer Causes Control 12:461–475. Friedenreich CM, Courneya KS, Bryant HE. 2002. Case-control study of anthropometric measures and breast cancer risk. Int J Cancer 99:445–452. Furberg AS, Thune I. 2003. Metabolic abnormalities (hypertension, hyperglycemia and overweight), lifestyle (high energy intake and physical inactivity) and endometrial cancer risk in a Norwegian cohort. Int J Cancer 104:669–676. Galanis DJ, Kolonel LN, Lee J, Le Marchand L. 1998. Anthropometric predictors of breast cancer incidence and survival in a multi-ethnic cohort of female residents of Hawaii, United States. Cancer Causes Control 9:217–224. Galanti MR, Hansson L, Bergström R, Wolk A, Hjartaker A, Lund E, Grimelius L, Ekbom A. 1997. Diet and the risk of papillary and follicular thyroid carcinoma: a population-based case-control study in Sweden and Norway. Cancer Causes Control 8:205–214. Gallus S, Bosetti C, Franceschi S, Levi F, Simonato L, Negri E, La Vecchia C. 2001. Oesophageal cancer in women: Tobacco, alcohol, nutritional and hormonal factors. Br J Cancer 85:341–345.

442

PART III: THE CAUSES OF CANCER

Garfinkel L. 1986. Overweight and mortality. Cancer 58:1826–1829. Gerhardsson de Verdier M, Hagman U, Steineck G, Rieger A, Norell SE. 1990. Diet, body mass and colorectal cancer: a case–referent study in Stockholm. Int J Cancer 46:832–838. Giovannucci E, Ascherio A, Rimm EB, Colditz GA, Stampfer MJ, Willett WC. 1995. Physical activity, obesity, and risk for colon cancer and adenoma in men. Ann Intern Med 122:327–334. Giovannucci E, Colditz GA, Stampfer MJ, Willett WC. 1996. Physical activity, obesity, and risk of colorectal adenoma in women. United States. Cancer Causes Control 7:253–263. Giovannucci E, Rimm EB, Stampfer MJ, Colditz GA, Willett WC. 1997. Height, body weight, and risk of prostate cancer. Cancer Epidemiol Biomarkers Prev 6:557–563. Giuffrida D, Lupo L, La Porta GA, La Rosa GL, Padova G, Foti E, Marchese V, Belfiore A. 1992. Relation between steroid receptor status and body weight in breast cancer patients. Eur J Cancer 28:112–115. Glantz GM. 1964. Cirrhosis and carcinoma of the prostate gland. J Urol 91:291–293. Glass AR. 1989. Endocrine aspects of obesity. Med Clin North Am 73:139–160. Glattre E, Haldorsen T, Berg JP, Stensvold I, Solvoll K. 1993. Norwegian casecontrol study testing the hypothesis that seafood increases the risk of thyroid cancer. Cancer Causes Control 4:11–16. Goodman MT, Morgenstern H, Wynder EL. 1986. A case-control study of factors affecting the development of renal cell cancer. Am J Epidemiol 124:926–941. Goodman MT, Kolonel LN, Wilkens LR. 1992. The association of body size, reproductive factors and thyroid cancer. Br J Cancer 66:1180–1184. Goodman MT, Wilkens LR. 1993. Relation of body size and the risk of lung cancer. Nutr Cancer 20:179–186. Goodman MT, Hankin JH, Wilkens LR, Lyu LC, McDuffie K, Liu LQ, Kolonel LN. 1997. Diet, body size, physical activity, and the risk of endometrial cancer. Cancer Res 57:5077–5085. Goodman-Gruen D, Barrett-Connor E. 1997. Epidemiology of insulin-like growth factor-I in elderly men and women. The Rancho Bernardo Study. Am J Epidemiol 145:970–976. Goodwin PJ, Panzarella T, Boyd NF. 1988. Weight gain in women with localized breast cancer—A descriptive study. Breast Cancer Res Treat 11:59–66. Goodwin PJ, Boyd NF. 1990. Body size and breast cancer prognosis: A critical review of the evidence. Breast Cancer Res Treat 16:205–214. Goodwin PJ, Ennis M, Pritchard KI, McCready D, Koo J, Sidlofsky S, Trudeau M, Hood N, Redwood S. 1999. Adjuvant treatment and onset of menopause predict weight gain after breast cancer diagnosis. J Clin Oncol 17:120–129. Goodwin PJ, Ennis M, Pritchard KI, Trudeau ME, Koo J, Hartwick W, Hoffma B, Hood N. 2002. Insulin-like growth factor binding proteins 1 and 3 and breast cancer outcomes. Breast Cancer Res Treat 74:65–76. Goodwin PJ, Ennis M, Pritchard KI, Trudeau ME, Koo J, Madarnas Y, Hartwick W, Hoffman B, Hood N. 2002. Fasting insulin and outcome in early-stage breast cancer: results of a prospective cohort study. J Clin Oncol 20:42–51. Graham S, Dayal H, Swanson M, Mittelman A, Wilkinson G. 1978. Diet in the epidemiology of cancer of the colon and rectum. J Natl Cancer Inst 61:709–714. Graham S, Haughey B, Marshall J, Priore R, Byers T, Rzepka T, Mettlin C, Pontes JE. 1983. Diet in the epidemiology of carcinoma of the prostate gland. J Natl Cancer Inst 70:687–692. Graham S, Marshall J, Haughey B, Mittelman A, Swanson M, Zielezny M, Byers T, Wilkinson G, West D. 1988. Dietary epidemiology of cancer of the colon in western New York. Am J Epidemiol 128:490–503. Greenberg ER, Vessey MP, McPherson K, Doll R, Yeates D. 1985. Body size and survival in premenopausal breast cancer. Br J Cancer 51:691–697. Greggi S, Parazzini F, Paratore MP, Chatenoud L, Legge F, Mancuso S, La Vecchia C. 2000. Risk factors for ovarian cancer in central Italy. Gynecol Oncol 79:50–54. Gronberg H, Damber L, Damber JE. 1996. Total food consumption and body mass index in relation to prostate cancer risk: A case-control study in Sweden with prospectively collected exposure data. J Urol 155:969–974. Gunnell D, Berney L, Holland P, Maynard M, Blane D, Frankel S, Smith GD. 2000. How accurately are height, weight and leg length reported by the elderly, and how closely are they related to measurements recorded in childhood? Int J Epidemiol 29:456–464. Gunnell D, Okasha M, Smith GD, Oliver SE, Sandhu J, Holly JM. 2001. Height, leg length, and cancer risk: A systematic review. Epidemiol Rev 23:313–342. Habel LA, Van Den Eeden SK, Friedman GD. 2000. Body size, age at shaving initiation, and prostate cancer in a large, multiracial cohort. Prostate 43:136–143.

Haffner SM, Katz MS, Stern MP, Dunn JF. 1989. Relationship of sex hormone binding globulin to overall adiposity and body fat distribution in a biethnic population. Int J Obes 13:1–9. Haffner SM, Miettinen H, Karhapaa P, Mykkanen L, Laakso M. 1997. Leptin concentrations, sex hormones, and cortisol in nondiabetic men. J Clin Endocrinol Metab 82:1807–1809. Hale GE, Hughes CL, Cline JM. 2002. Endometrial cancer: Hormonal factors, the perimenopausal “window of risk,” and isoflavones. J Clin Endocrinol Metab 87:3–15. Hall IJ, Newman B, Millikan RC, Moorman PG. 2000. Body size and breast cancer risk in black women and white women: The Carolina Breast Cancer Study. Am J Epidemiol 151:754–764. Hallquist A, Hardell L, Degerman A, Boquist L. 1994. Thyroid cancer: reproductive factors, previous diseases, drug intake, family history and diet. A case-control study. Eur J Cancer Prev 3:481–488. Hankinson SE, Willett WC, Colditz GA, Hunter DJ, Michaud DS, Deroo B, Rosner B, Speizer FE, Pollak M. 1998. Circulating concentrations of insulin-like growth factor-I and risk of breast cancer. Lancet 351:1393– 1396. Harris RE, Namboodiri KK, Wynder EL. 1992. Breast cancer risk: Effects of estrogen replacement therapy and body mass. J Natl Cancer Inst 84:1575– 1582. Hartge P, Schiffman MH, Hoover R, McGowan L, Lesher L, Norris HJ. 1989. A case-control study of epithelial ovarian cancer. Am J Obstet Gynecol 161:10–16. Hartz AJ, Rupley DC, Rimm AA. 1984. The association of girth measurements with disease in 32,856 women. Am J Epidemiol 119:71–80. Hayes RB, Ziegler RG, Gridley G, Swanson C, Greenberg RS, Swanson GM, Schoenberg JB, Silverman DT, Brown LM, Pottern LM, Liff J, Schwartz AG, Fraumeni JF Jr, Hoover RN. 1999. Dietary factors and risks for prostate cancer among blacks and whites in the United States. Cancer Epidemiol Biomarkers Prev 8:25–34. Heasman KZ, Sutherland HJ, Campbell JA, Elhakim T, Boyd NF. 1985. Weight gain during adjuvant chemotherapy for breast cancer. Breast Cancer Res Treat 5:195–200. Hebert JR, Augustine A, Barone J, Kabat GC, Kinne DW, Wynder EL. 1988. Weight, height and body mass index in the prognosis of breast cancer: Early results of a prospective study. Int J Cancer 42:315–318. Hebert JR, Hurley TG, Olendzki BC, Teas J, Ma Y, Hampl JS. 1998. Nutritional and socioeconomic factors in relation to prostate cancer mortality: a cross-national study. J Natl Cancer Inst 90:1637–1647. Heitmann BL. 1993. The influence of fatness, weight change, slimming history and other lifestyle variables on diet reporting in Danish men and women aged 35–65 years. Int J Obes Relat Metab Disord 17:329–336. Henderson BE, Ross RK, Pike MC, Casagrande JT. 1982. Endogenous hormones as a major factor in human cancer. Cancer Res 42:3232–3239. Henley SJ, Flanders WD, Manatunga A, Thun MJ. 2002. Leanness and lung cancer risk: Fact or artifact? Epidemiology 13:268–276. Hiatt RA, Tolan K, Quesenberry CP Jr. 1994. Renal cell carcinoma and thiazide use: A historical, case-control study. California, USA. Cancer Causes Control 5:319–325. Hilakivi-Clarke L, Forsen T, Eriksson JG, Luoto R, Tuomilehto J, Osmond C, Barker DJ. 2001. Tallness and overweight during childhood have opposing effects on breast cancer risk. Br J Cancer 85:1680–1684. Hirose K, Tajima K, Hamajima N, Takezaki T, Inoue M, Kuroishi T, Miura S, Tokudome S. 2001. Association of family history and other risk factors with breast cancer risk among Japanese premenopausal and postmenopausal women. Cancer Causes Control 12:349–358. Hislop TG, Coldman AJ, Elwood JM, Brauer G, Kan L. 1986. Childhood and recent eating patterns and risk of breast cancer. Cancer Detect Prev 9:47–58. Ho SC, Tai ES, Eng PH, Ramli A, Tan CE, Fok AC. 1999. A study in the relationships between leptin, insulin, and body fat in Asian subjects. Int J Obes Relat Metab Disord 23:246–252. Hsieh CC, Trichopoulos D, Katsouyanni K, Yuasa S. 1990. Age at menarche, age at menopause, height and obesity as risk factors for breast cancer: associations and interactions in an international case-control study. Int J Cancer 46:796–800. Hsieh CC, Thanos A, Mitropoulos D, Deliveliotis C, Mantzoros CS, Trichopoulos D. 1999. Risk factors for prostate cancer: A case-control study in Greece. Int J Cancer 80:699–703. Hsing AW, Deng J, Sesterhenn IA, Mostofi FK, Stanczyk FZ, Benichou J, Xie T, Gao YT. 2000. Body size and prostate cancer: a population-based case-control study in China. Cancer Epidemiol Biomarkers Prev 9:1335– 1341. Hsing AW, Chua S Jr, Gao YT, Gentzschein E, Chang L, Deng J, Stanczyk FZ. 2001. Prostate cancer risk and serum levels of insulin and leptin: A population-based study. J Natl Cancer Inst 93:783–789.

Obesity and Body Composition Huang Z, Hankinson SE, Colditz GA, Stampfer MJ, Hunter DJ, Manson JE, Hennekens CH, Rosner B, Speizer FE, Willett WC. 1997. Dual effects of weight and weight gain on breast cancer risk. JAMA 278:1407–1411. Huang Z, Willett WC, Colditz GA, Hunter DJ, Manson JE, Rosner B, Speizer FE, Hankinson SE. 1999. Waist circumference, waist : hip ratio, and risk of breast cancer in the Nurses’ Health Study. Am J Epidemiol 150: 1316–1324. IARC Working Group. 1986. International Agency for Research on Cancer: Tobacco smoking. IARC monographs on the evaluation of the carcinogenic risks to humans, vol 38. Lyon, France: IARC Press. IARC Working Group. 1988. International Agency for Research on Cancer: Alcohol drinking. IARC monographs on the evaluation of the carcinogenic risks to humans, vol 44. Lyon, France: IARC Press. IARC Working Group. 2002. IARC Handbooks of Cancer Prevention, vol. 6. Weight control and physical activity. Lyon, France: IARC Press, pp. 1–315. Ilic M, Vlajinac H, Marinkovic J. 1996. Case-control study of risk factors for prostate cancer. Br J Cancer 74:1682–1686. Ingram D, Nottage E, Ng S, Sparrow L, Roberts A, Willcox D. 1989. Obesity and breast disease. The role of the female sex hormones. Cancer 64:1049–1053. Innes K, Byers T, Schymura M. 2000. Birth characteristics and subsequent risk for breast cancer in very young women. Am J Epidemiol 152:1121–1128. Inoue M, Okayama A, Fujita M, Enomoto T, Tanizawa O, Ueshima H. 1994. A case-control study on risk factors for uterine endometrial cancer in Japan. Jpn J Cancer Res 85:346–350. Inoue M, Ito LS, Tajima K, Yamamura Y, Kodera Y, Takezaki T, Hamajima N, Hirose K, Kuroishi T, Tominaga S. 2002. Height, weight, menstrual and reproductive factors and risk of gastric cancer among Japanese postmenopausal women: Analysis by subsite and histologic subtype. Int J Cancer 97:833–838. Iribarren C, Haselkorn T, Tekawa IS, Friedman GD. 2001. Cohort study of thyroid cancer in a San Francisco Bay area population. Int J Cancer 93:745–750. Jain MG, Howe GR, Rohan TE. 2000. Nutritional factors and endometrial cancer in Ontario, Canada. Cancer Control 7:288–296. Janssen JA, Stolk RP, Pols HA, Grobbee DE, de Jong FH, Lamberts SW. 1998. Serum free IGF-I, total IGF-I, IGFBP-1 and IGFBP-3 levels in an elderly population: relation to age and sex steroid levels. Clin Endocrinol (Oxf) 48:471–478. Jasienska G, Thune I. 2001. Lifestyle, hormones, and risk of breast cancer. BMJ 322:586–587. Jernstrom H, Barrett-Connor E. 1999. Obesity, weight change, fasting insulin, proinsulin, C-peptide, and insulin-like growth factor-1 levels in women with and without breast cancer: The Rancho Bernardo Study. J Womens Health Gend Based Med 8:1265–1272. Ji BT, Chow WH, Yang G, McLaughlin JK, Gao RN, Zheng W, Shu XO, Jin F, Fraumeni JF Jr, Gao YT. 1997. Body mass index and the risk of cancers of the gastric cardia and distal stomach in Shanghai, China. Cancer Epidemiol Biomarkers Prev 6:481–485. Johannsson G, Karlsson C, Lonn L, Marin P, Bjorntorp P, Sjostrom L, Carlsson B, Carlsson LM, Bengtsson BA. 1998. Serum leptin concentration and insulin sensitivity in men with abdominal obesity. Obes Res 6:416–421. Kaaks R, van Noord PA, den Tonkelaar I, Peeters PJ, Riboli E, Grobbee DE. 1998. Breast-cancer incidence in relation to height, weight and body-fat distribution in the Dutch “DOM” cohort. Int J Cancer 76:647–651. Kabat GC, Wynder EL. 1992. Body mass index and lung cancer risk. Am J Epidemiol 135:769–774. Kabat GC, Ng SK, Wynder EL. 1993. Tobacco, alcohol intake, and diet in relation to adenocarcinoma of the esophagus and gastric cardia. Cancer Causes Control 4:123–132. Kabat GC, Chang CJ, Wynder EL. 1994. The role of tobacco, alcohol use, and body mass index in oral and pharyngeal cancer. Int J Epidemiol 23:1137–1144. Kadamani S, Asal NR, Nelson RY. 1989. Occupational hydrocarbon exposure and risk of renal cell carcinoma. Am J Ind Med 15:131–141. Kaijser M, Lichtenstein P, Granath F, Erlandsson G, Cnattingius S, Ekbom A. 2001. In utero exposures and breast cancer: a study of opposite-sexed twins. J Natl Cancer Inst 93:60–62. Kaklamani VG, Linos A, Kaklamani E, Markaki I, Mantzoros C. 1999. Age, sex, and smoking are predictors of circulating insulin-like growth factor 1 and insulin-like growth factor-binding protein 3. J Clin Oncol 17:813–817. Kalandidi A, Tzonou A, Lipworth L, Gamatsi I, Filippa D, Trichopoulos D. 1996. A case-control study of endometrial cancer in relation to reproductive, somatometric, and life-style variables. Oncology 53:354– 359.

443

Kalish LA. 1984. Relationships of body size with breast cancer. J Clin Oncol 2:287–293. Kark JD, Yaari S, Rasooly I, Goldbourt U. 1995. Are lean smokers at increased risk of lung cancer? The Israel Civil Servant Cancer Study. Arch Intern Med 155:2409–2416. Kato I, Nomura A, Stemmermann GN, Chyou PH. 1992. Determinants of sex hormone levels in men as useful indices in hormone-related disorders. J Clin Epidemiol 45:1417–1421. Kaye SA, Folsom AR, Soler JT, Prineas RJ, Potter JD. 1991. Associations of body mass and fat distribution with sex hormone concentrations in postmenopausal women. Int J Epidemiol 20:151–156. Kelly PJ, Eisman JA, Stuart MC, Pocock NA, Sambrook PN, Gwinn TH. 1990. Somatomedin-C, physical fitness, and bone density. J Clin Endocrinol Metab 70:718–723. Key TJ, Pike MC. 1988a. The dose-effect relationship between “unopposed” oestrogens and endometrial mitotic rate: its central role in explaining and predicting endometrial cancer risk. Br J Cancer 57:205–212. Key TJ, Pike MC. 1988b. The role of oestrogens and progestagens in the epidemiology and prevention of breast cancer. Eur J Cancer Clin Oncol 24:29–43. Key TJ, Silcocks PB, Davey GK, Appleby PN, Bishop DT. 1997. A case-control study of diet and prostate cancer. Br J Cancer 76:678–687. Key TJ, Allen NE, Verkasalo PK, Banks E. 2001. Energy balance and cancer: The role of sex hormones. Proc Nutr Soc 60:81–89. Kirschner MA, Ertel N, Schneider G. 1981. Obesity, hormones, and cancer. Cancer Res 41:3711–3717. Kirschner MA, Samojlik E, Drejka M, Szmal E, Schneider G, Ertel N. 1990. Androgen-estrogen metabolism in women with upper body versus lower body obesity. J Clin Endocrinol Metab 70:473–479. Knekt P, Heliovaara M, Rissanen A, Aromaa A, Seppanen R, Teppo L, Pukkala E. 1991. Leanness and lung-cancer risk. Int J Cancer 49:208–213. Kolonel LN, Nomura AM, Lee J, Hirohata T. 1986. Anthropometric indicators of breast cancer risk in postmenopausal women in Hawaii. Nutr Cancer 8:247–256. Kolonel LN, Yoshizawa CN, Hankin JH. 1988. Diet and prostatic cancer: A case-control study in Hawaii. Am J Epidemiol 127:999–1012. Kolonel LN, Hankin JH, Wilkens LR, Fukunaga FH, Hinds MW. 1990. An epidemiologic study of thyroid cancer in Hawaii. Cancer Causes Control 1:223–234. Kono S, Shinchi K, Ikeda N, Yanai F, Imanishi K. 1991. Physical activity, dietary habits and adenomatous polyps of the sigmoid colon: A study of self-defense officials in Japan. J Clin Epidemiol 44:1255–1261. Kono S, Handa K, Hayabuchi H, Kiyohara C, Inoue H, Marugame T, Shinomiya S, Hamada H, Onuma K, Koga H. 1999. Obesity, weight gain and risk of colon adenomas in Japanese men. Jpn J Cancer Res 90:805–811. Kopelman PG. 2000. Obesity as a medical problem. Nature 404:635–643. Koumantaki Y, Tzonou A, Koumantakis E, Kaklamani E, Aravantinos D, Trichopoulos D. 1989. A case-control study of cancer of endometrium in Athens. Int J Cancer 43:795–799. Kreiger N, Marrett LD, Dodds L, Hilditch S, Darlington GA. 1993. Risk factors for renal cell carcinoma: results of a population-based case-control study. Cancer Causes Control 4:101–110. Kubik AK, Zatloukal P, Tomasek L, Petruzelka L. 2002. Lung cancer risk among Czech women: a case-control study. Prev Med 34:436–444. Kuczmarski MF, Kuczmarski RJ, Najjar M. 2001. Effects of age on validity of self–reported height, weight, and body mass index: findings from the Third National Health and Nutrition Examination Survey, 1988–1994. J Am Diet Assoc 101:28–34. Kune GA, Kune S, Watson LF. 1990. Body weight and physical activity as predictors of colorectal cancer risk. Nutr Cancer 13:9–17. Kyogoku S, Hirohata T, Takeshita S, Nomura Y, Shigematsu T, Horie A. 1990. Survival of breast-cancer patients and body size indicators. Int J Cancer 46:824–831. La Vecchia C, Franceschi S, Decarli A, Gallus G, Tognoni G. 1984. Risk factors for endometrial cancer at different ages. J Natl Cancer Inst 73:667–671. La Vecchia C, Parazzini F, Negri E, Fasoli M, Gentile A, Franceschi S. 1991. Anthropometric indicators of endometrial cancer risk. Eur J Cancer 27:487–490. La Vecchia C, Ron E, Franceschi S, Dal Maso L, Mark SD, Chatenoud L, Braga C, Preston-Martin S, McTiernan A, Kolonel L, Mabuchi K, Jin F, Wingren G, Galanti MR, Hallquist A, Lund E, Levi F, Linos D, Negri E. 1999. A pooled analysis of case-control studies of thyroid cancer. III. Oral contraceptives, menopausal replacement therapy and other female hormones. Cancer Causes Control 10:157–166. Lagergren J, Bergström R, Nyren O. 1999. Association between body mass and adenocarcinoma of the esophagus and gastric cardia. Ann Intern Med 130:883–890.

444

PART III: THE CAUSES OF CANCER

Lahmann PH, Lissner L, Gullberg B, Olsson H, Berglund G. 2003. A prospective study of adiposity and postmenopausal breast cancer risk: The Malmo Diet and Cancer Study. Int J Cancer 103:246–252. Lam PB, Vacek PM, Geller BM, Muss HB. 2000. The association of increased weight, body mass index, and tissue density with the risk of breast carcinoma in Vermont. Cancer 89:369–375. Landin-Wilhelmsen K, Wilhelmsen L, Lappas G, Rosen T, Lindstedt G, Lundberg PA, Bengtsson BA. 1994. Serum insulin-like growth factor I in a random population sample of men and women: relation to age, sex, smoking habits, coffee consumption and physical activity, blood pressure and concentrations of plasma lipids, fibrinogen, parathyroid hormone and osteocalcin. Clin Endocrinol (Oxf) 41:351–357. Lapidus L, Helgesson O, Merck C, Bjorntorp P. 1988. Adipose tissue distribution and female carcinomas. A 12-year follow-up of participants in the population study of women in Gothenburg, Sweden. Int J Obes 12:361–368. Larsson H, Elmstahl S, Berglund G, Ahren B. 1998. Evidence for leptin regulation of food intake in humans. J Clin Endocrinol Metab 83:4382– 4385. Le Marchand L, Kolonel LN, Earle ME, Mi MP. 1988a. Body size at different periods of life and breast cancer risk. Am J Epidemiol 128:137–152. Le Marchand L, Kolonel LN, Myers BC, Mi MP. 1988b. Birth characteristics of premenopausal women with breast cancer. Br J Cancer 57:437–439. Le Marchand L, Wilkens LR, Mi MP. 1991. Early-age body size, adult weight gain and endometrial cancer risk. Int J Cancer 48:807–811. Le Marchand L, Wilkens LR, Mi MP. 1992. Obesity in youth and middle age and risk of colorectal cancer in men. Cancer Causes Control 3:349–354. Le Marchand L, Kolonel LN, Wilkens LR, Myers BC, Hirohata T. 1994. Animal fat consumption and prostate cancer: A prospective study in Hawaii. Epidemiology 5:276–282. Le Marchand L, Wilkens LR, Kolonel LN, Hankin JH, Lyu LC. 1997. Associations of sedentary lifestyle, obesity, smoking, alcohol use, and diabetes with the risk of colorectal cancer. Cancer Res 57:4787–4794. Lee IM, Paffenbarger RS Jr. 1992a. Quetelet’s index and risk of colon cancer in college alumni. J Natl Cancer Inst 84:1326–1331. Lee IM, Paffenbarger RS Jr. 1992b. Re: “Body mass index and lung cancer risk”. Am J Epidemiol 136:1417–1419. Lee IM, Sesso HD, Paffenbarger RS Jr. 2001. A prospective cohort study of physical activity and body size in relation to prostate cancer risk (United States). Cancer Causes Control 12:187–193. Lees AW, Jenkins HJ, May CL, Cherian G, Lam EW, Hanson J. 1989. Risk factors and 10-year breast cancer survival in northern Alberta. Breast Cancer Res Treat 13:143–151. Levi F, Franceschi S, La Vecchia C, Negri E, Gulie C, Duruz G, Scazziga B. 1991. Previous thyroid disease and risk of thyroid cancer in Switzerland. Eur J Cancer 27:85–88. Levi F, La Vecchia C, Negri E, Parazzini F, Franceschi S. 1992. Body mass at different ages and subsequent endometrial cancer risk. Int J Cancer 50:567–571. Lew EA, Garfinkel L. 1979. Variations in mortality by weight among 750,000 men and women. J Chronic Dis 32:563–576. Li CI, Stanford JL, Daling JR. 2000. Anthropometric variables in relation to risk of breast cancer in middle-aged women. Int J Epidemiol 29: 208–213. Li SD, Mobarhan S. 2000. Association between body mass index and adenocarcinoma of the esophagus and gastric cardia. Nutr Rev 58:54–56. Licinio J, Mantzoros C, Negrao AB, Cizza G, Wong ML, Bongiorno PB, Chrousos GP, Karp B, Allen C, Flier JS, Gold PW. 1997. Human leptin levels are pulsatile and inversely related to pituitary–adrenal function. Nat Med 3:575–579. Lindblad P, Wolk A, Bergström R, Persson I, Adami HO. 1994. The role of obesity and weight fluctuations in the etiology of renal cell cancer: A population-based case-control study. Cancer Epidemiol Biomarkers Prev 3:631–639. Linos A, Linos DA, Vgotza N, Souvatzoglou A, Koutras DA. 1989. Does coffee consumption protect against thyroid disease? Acta Chir Scand 155: 317–320. Lipman JM, Turturro A, Hart RW. 1989. The influence of dietary restriction on DNA repair in rodents: A preliminary study. Mech Ageing Dev 48:135–143. Lohman TG, Roche AF, Martorell R. 1988. Anthropometric Standardization Reference Manual. Champaign, IL: Human Kinetics Books. London SJ, Colditz GA, Stampfer MJ, Willett WC, Rosner B, Speizer FE. 1989. Prospective study of relative weight, height, and risk of breast cancer. JAMA 262:2853–2858. Lubin F, Ruder AM, Wax Y, Modan B. 1985. Overweight and changes in weight throughout adult life in breast cancer etiology. A case-control study. Am J Epidemiol 122:579–588.

Lubin F, Chetrit A, Freedman LS, Alfandary E, Fishler Y, Nitzan H, Zultan A, Modan B. 2003. Body mass index at age 18 years and during adult life and ovarian cancer risk. Am J Epidemiol 157:113–120. Lukanova A, Toniolo P, Lundin E, Micheli A, Akhmedkhanov A, Muti P, Zeleniuch–Jacquotte A, Biessy C, Lenner P, Krogh V, Berrino F, Hallmans G, Riboli E, Kaaks R. 2002. Body mass index in relation to ovarian cancer: A multi-centre nested case-control study. Int J Cancer 99:603–608. Lund E, Adami HO, Bergström R, Meirik O. 1990. Anthropometric measures and breast cancer in young women. Cancer Causes Control 1:169–172. Lund Nilsen TI, Vatten LJ. 1999. Anthropometry and prostate cancer risk: a prospective study of 22,248 Norwegian men. Cancer Causes Control 10:269–275. Lund Nilsen TI, Vatten LJ. 2001. Prospective study of colorectal cancer risk and physical activity, diabetes, blood glucose and BMI: Exploring the hyperinsulinaemia hypothesis. Br J Cancer 84:417–422. Maccario M, Ramunni J, Oleandri SE, Procopio M, Grottoli S, Rossetto R, Savio P, Aimaretti G, Camanni F, Ghigo E. 1999. Relationships between IGF–I and age, gender, body mass, fat distribution, metabolic and hormonal variables in obese patients. Int J Obes Relat Metab Disord 23:612–618. Mack WJ, Preston–Martin S, Bernstein L, Qian D. 2002. Lifestyle and other risk factors for thyroid cancer in Los Angeles County females. Ann Epidemiol 12:395–401. Maclure M, Willett W. 1990. A case-control study of diet and risk of renal adenocarcinoma. Epidemiology 1:430–440. Maehle BO, Tretli S. 1996. Pre-morbid body-mass-index in breast cancer: reversed effect on survival in hormone receptor negative patients. Breast Cancer Res Treat 41:123–130. Magnusson C, Baron J, Persson I, Wolk A, Bergström R, Trichopoulos D, Adami HO. 1998. Body size in different periods of life and breast cancer risk in post-menopausal women. Int J Cancer 76:29–34. Manjer J, Kaaks R, Riboli E, Berglund G. 2001. Risk of breast cancer in relation to anthropometry, blood pressure, blood lipids and glucose metabolism: a prospective study within the Malmo Preventive Project. Eur J Cancer Prev 10:33–42. Mannisto S, Pietinen P, Pyy M, Palmgren J, Eskelinen M, Uusitupa M. 1996. Body–size indicators and risk of breast cancer according to menopause and estrogen–receptor status. Int J Cancer 68:8–13. Mantzoros CS, Tzonou A, Signorello LB, Stampfer M, Trichopoulos D, Adami HO. 1997. Insulin-like growth factor 1 in relation to prostate cancer and benign prostatic hyperplasia. Br J Cancer 76:1115–1118. Mantzoros CS, Bolhke K, Moschos S, Cramer DW. 1999. Leptin in relation to carcinoma in situ of the breast: A study of pre-menopausal cases and controls. Int J Cancer 80:523–526. Marin P, Kvist H, Lindstedt G, Sjostrom L, Bjorntorp P. 1993. Low concentrations of insulin-like growth factor-I in abdominal obesity. Int J Obes Relat Metab Disord 17:83–89. Marshall JR, Graham S, Haughey BP, Shedd D, O’Shea R, Brasure J, Wilkinson GS, West D. 1992. Smoking, alcohol, dentition and diet in the epidemiology of oral cancer. Eur J Cancer B Oral Oncol 28B:9–15. Martinez ME, Giovannucci E, Spiegelman D, Hunter DJ, Willett WC, Colditz GA. 1997. Leisure-time physical activity, body size, and colon cancer in women. Nurses’ Health Study Research Group. J Natl Cancer Inst 89:948–955. Martinez ME, Maltzman T, Marshall JR, Einspahr J, Reid ME, Sampliner R, Ahnen DJ, Hamilton SR, Alberts DS. 1999. Risk factors for Ki–ras protooncogene mutation in sporadic colorectal adenomas. Cancer Res 59:5181–5185. McCredie M, Stewart JH. 1992. Risk factors for kidney cancer in New South Wales, Australia. II. Urologic disease, hypertension, obesity, and hormonal factors. Cancer Causes Control 3:323–331. McLaughlin JK, Mandel JS, Blot WJ, Schuman LM, Mehl ES, Fraumeni JF Jr. 1984. A population-based case-control study of renal cell carcinoma. J Natl Cancer Inst 72:275–284. McLaughlin JK, Gridley G, Block G, Winn DM, Preston-Martin S, Schoenberg JB, Greenberg RS, Stemhagen A, Austin DF, Ershow AG. 1988. Dietary factors in oral and pharyngeal cancer. J Natl Cancer Inst 80:1237–1243. McLaughlin JK, Gao YT, Gao RN, Zheng W, Ji BT, Blot WJ, Fraumeni JF Jr. 1992. Risk factors for renal-cell cancer in Shanghai, China. Int J Cancer 52:562–565. McNee RK, Mason BH, Neave LM, Kay RG. 1987. Influence of height, weight, and obesity on breast cancer incidence and recurrence in Auckland, New Zealand. Breast Cancer Res Treat 9:145–150. McTiernan AM, Weiss NS, Daling JR. 1984. Incidence of thyroid cancer in women in relation to reproductive and hormonal factors. Am J Epidemiol 120:423–435.

Obesity and Body Composition Mellemgaard A, Engholm G, McLaughlin JK, Olsen JH. 1994. Risk factors for renal–cell carcinoma in Denmark. III. Role of weight, physical activity and reproductive factors. Int J Cancer 56:66–71. Mellemgaard A, Lindblad P, Schlehofer B, Bergström R, Mandel JS, McCredie M, McLaughlin JK, Niwa S, Odaka N, Pommer W. 1995. International renal-cell cancer study. III. Role of weight, height, physical activity, and use of amphetamines. Int J Cancer 60:350–354. Mettlin C, Selenskas S, Natarajan N, Huben R. 1989. Beta-carotene and animal fats and their relationship to prostate cancer risk. A case-control study. Cancer 64:605–612. Michels KB, Trichopoulos D, Robins JM, Rosner BA, Manson JE, Hunter DJ, Colditz GA, Hankinson SE, Speizer FE, Willett WC. 1996. Birthweight as a risk factor for breast cancer. Lancet 348:1542–1546. Mills PK, Beeson WL, Phillips RL, Fraser GE. 1989. Cohort study of diet, lifestyle, and prostate cancer in Adventist men. Cancer 64:598–604. Mink PJ, Folsom AR, Sellers TA, Kushi LH. 1996. Physical activity, waist-tohip ratio, and other risk factors for ovarian cancer: a follow-up study of older women. Epidemiology 7:38–45. Mohle–Boetani JC, Grosser S, Whittemore AS, Malec M, Kampert JB, Paffenbarger RS Jr. 1988. Body size, reproductive factors, and breast cancer survival. Prev Med 17:634–642. Møller H, Mellemgaard A, Lindvig K, Olsen JH. 1994. Obesity and cancer risk: A Danish record-linkage study. Eur J Cancer 30A:344–350. Moore JW, Key TJ, Bulbrook RD, Clark GM, Allen DS, Wang DY, Pike MC. 1987. Sex hormone binding globulin and risk factors for breast cancer in a population of normal women who had never used exogenous sex hormones. Br J Cancer 56:661–666. Mori M, Harabuchi I, Miyake H, Casagrande JT, Henderson BE, Ross RK. 1988. Reproductive, genetic, and dietary risk factors for ovarian cancer. Am J Epidemiol 128:771–777. Mori M, Nishida T, Sugiyama T, Komai K, Yakushiji M, Fukuda K, Tanaka T, Yokoyama M, Sugimori H. 1998. Anthropometric and other risk factors for ovarian cancer in a case-control study. Jpn J Cancer Res 89:246–253. Mukherjee P, Sotnikov AV, Mangian HJ, Zhou JR, Visek WJ, Clinton SK. 1999. Energy intake and prostate tumor growth, angiogenesis, and vascular endothelial growth factor expression. J Natl Cancer Inst 91:512–523. Muscat JE, Wynder EL. 1992. Tobacco, alcohol, asbestos, and occupational risk factors for laryngeal cancer. Cancer 69:2244–2251. Muscat JE, Hoffmann D, Wynder EL. 1995. The epidemiology of renal cell carcinoma. A second look. Cancer 75:2552–2557. Must A, Willett WC, Dietz WH. 1993. Remote recall of childhood height, weight, and body build by elderly subjects. Am J Epidemiol 138:56– 64. Negri E, La Vecchia C, Bruzzi P, Dardanoni G, Decarli A, Palli D, Parazzini F, Rosselli del Turco M. 1988. Risk factors for breast cancer: Pooled results from three Italian case-control studies. Am J Epidemiol 128:1207– 1215. Negri E, Ron E, Franceschi S, Dal Maso L, Mark SD, Preston-Martin S, McTiernan A, Kolonel L, Kleinerman R, Land C, Jin F, Wingren G, Galanti MR, Hallquist A, Glattre E, Lund E, Levi F, Linos D, Braga C, La Vecchia C. 1999. A pooled analysis of case-control studies of thyroid cancer. I. Methods. Cancer Causes Control 10:131–142. Negri E, Franceschi S, Bosetti C, Levi F, Conti E, Parpinel M, La Vecchia C. 2000. Selected micronutrients and oral and pharyngeal cancer. Int J Cancer 86:122–127. Ness RB, Grisso JA, Klapper J, Schlesselman JJ, Silberzweig S, Vergona R, Morgan M, Wheeler JE. 2000. Risk of ovarian cancer in relation to estrogen and progestin dose and use characteristics of oral contraceptives. SHARE Study Group. Steroid Hormones and Reproductions. Am J Epidemiol 152:233–241. Neugut AI, Lee WC, Garbowski GC, Waye JD, Forde KA, Treat MR, FenoglioPreiser C. 1991. Obesity and colorectal adenomatous polyps. J Natl Cancer Inst 83:359–361. Newman SC, Lees AW, Jenkins HJ. 1997. The effect of body mass index and oestrogen receptor level on survival of breast cancer patients. Int J Epidemiol 26:484–490. Ng EH, Gao F, Ji CY, Ho GH, Soo KC. 1997. Risk factors for breast carcinoma in Singaporean Chinese women: The role of central obesity. Cancer 80:725–731. Niskanen LK, Haffner S, Karhunen LJ, Turpeinen AK, Miettinen H, Uusitupa MI. 1997. Serum leptin in obesity is related to gender and body fat topography but does not predict successful weight loss. Eur J Endocrinol 137:61–67. Noble RL. 1977. The development of prostatic adenocarcinoma in Nb rats following prolonged sex hormone administration. Cancer Res 37: 1929–1933. Nomura A, Heilbrun LK, Stemmermann GN. 1985. Body mass index as a predictor of cancer in men. J Natl Cancer Inst 74:319–323.

445

Ogden CL, Flegal KM, Carroll MD, Johnson CL. 2002. Prevalence and trends in overweight among US children and adolescents, 1999–2000. JAMA 288:1728–1732. Olson JE, Yang P, Schmitz K, Vierkant RA, Cerhan JR, Sellers TA. 2002. Differential association of body mass index and fat distribution with three major histologic types of lung cancer: Evidence from a cohort of older women. Am J Epidemiol 156:606–615. Olson SH, Trevisan M, Marshall JR, Graham S, Zielezny M, Vena JE, Hellmann R, Freudenheim JL. 1995. Body mass index, weight gain, and risk of endometrial cancer. Nutr Cancer 23:141–149. Paffenbarger RS Jr, Kampert JB, Chang HG. 1980. Characteristics that predict risk of breast cancer before and after the menopause. Am J Epidemiol 112:258–268. Parazzini F, La Vecchia C, Negri E, Bruzzi P, Palli D, Boyle P. 1990. Anthropometric variables and risk of breast cancer. Int J Cancer 45:397– 402. Parazzini F, Moroni S, La Vecchia C, Negri E, dal Pino D, Bolis G. 1997. Ovarian cancer risk and history of selected medical conditions linked with female hormones. Eur J Cancer 33:1634–1637. Parslov M, Lidegaard O, Klintorp S, Pedersen B, Jonsson L, Eriksen PS, Ottesen B. 2000. Risk factors among young women with endometrial cancer: A Danish case-control study. Am J Obstet Gynecol 182:23– 29. Partanen T, Heikkila P, Hernberg S, Kauppinen T, Moneta G, Ojajarvi A. 1991. Renal cell cancer and occupational exposure to chemical agents. Scand J Work Environ Health 17:231–239. Pathak DR, Whittemore AS. 1992. Combined effects of body size, parity, and menstrual events on breast cancer incidence in seven countries. Am J Epidemiol 135:153–168. Peacock SL, White E, Daling JR, Voigt LF, Malone KE. 1999. Relation between obesity and breast cancer in young women. Am J Epidemiol 149:339–346. Petrek JA, Peters M, Cirrincione C, Rhodes D, Bajorunas D. 1993. Is body fat topography a risk factor for breast cancer? Ann Intern Med 118:356–362. Peyrat JP, Bonneterre J, Hecquet B, Vennin P, Louchez MM, Fournier C, Lefebvre J, Demaille A. 1993. Plasma insulin-like growth factor-1 (IGF1) concentrations in human breast cancer. Eur J Cancer 29A:492–497. Potischman N, Swanson CA, Siiteri P, Hoover RN. 1996. Reversal of relation between body mass and endogenous estrogen concentrations with menopausal status. J Natl Cancer Inst 88:756–758. Potter JD, McMichael AJ. 1983. Large bowel cancer in women in relation to reproductive and hormonal factors: A case-control study. J Natl Cancer Inst 71:703–709. Potter JD, McMichael AJ. 1986. Diet and cancer of the colon and rectum: A case-control study. J Natl Cancer Inst 76:557–569. Preston-Martin S, Bernstein L, Pike MC, Maldonado AA, Henderson BE. 1987. Thyroid cancer among young women related to prior thyroid disease and pregnancy history. Br J Cancer 55:191–195. Preston-Martin S, Jin F, Duda MJ, Mack WJ. 1993. A case-control study of thyroid cancer in women under age 55 in Shanghai (People’s Republic of China). Cancer Causes Control 4:431–440. Purdie D, Green A, Bain C, Siskind V, Ward B, Hacker N, Quinn M, Wright G, Russell P, Susil B. 1995. Reproductive and other factors and risk of epithelial ovarian cancer: An Australian case-control study. Survey of Women’s Health Study Group. Int J Cancer 62:678–684. Purdie DM, Green AC. 2001. Epidemiology of endometrial cancer. Best Pract Res Clin Obstet Gynaecol 15:341–354. Putnam SD, Cerhan JR, Parker AS, Bianchi GD, Wallace RB, Cantor KP, Lynch CF. 2000. Lifestyle and anthropometric risk factors for prostate cancer in a cohort of Iowa men. Ann Epidemiol 10:361–369. Radimer K, Siskind V, Bain C, Schofield F. 1993. Relation between anthropometric indicators and risk of breast cancer among Australian women. Am J Epidemiol 138:77–89. Rauscher GH, Mayne ST, Janerich DT. 2000. Relation between body mass index and lung cancer risk in men and women never and former smokers. Am J Epidemiol 152:506–513. Riman T, Dickman PW, Nilsson S, Correia N, Nordlinder H, Magnusson CM, Persson IR. 2001. Risk factors for epithelial borderline ovarian tumors: results of a Swedish case-control study. Gynecol Oncol 83:575–585. Ron E, Kleinerman RA, Boice JD Jr, LiVolsi VA, Flannery JT, Fraumeni JF Jr. 1987. A population-based case-control study of thyroid cancer. J Natl Cancer Inst 79:1–12. Ronchi E, Pizzocaro G, Miodini P, Piva L, Salvioni R, Di Fronzo G. 1984. Steroid hormone receptors in normal and malignant human renal tissue: Relationship with progestin therapy. J Steroid Biochem 21:329–335. Ross RK, Shimizu H, Paganini-Hill A, Honda G, Henderson BE. 1987. Casecontrol studies of prostate cancer in blacks and whites in southern California. J Natl Cancer Inst 78:869–874.

446

PART III: THE CAUSES OF CANCER

Rossing MA, Voigt LF, Wicklund KG, Daling JR. 2000. Reproductive factors and risk of papillary thyroid cancer in women. Am J Epidemiol 151:765–772. Rowland ML. 1990. Self–reported weight and height. Am J Clin Nutr 52: 1125–1133. Russell CD, Petersen RN, Rao SP, Ricci MR, Prasad A, Zhang Y, Brolin RE, Fried SK. 1998. Leptin expression in adipose tissue from obese humans: Depot-specific regulation by insulin and dexamethasone. Am J Physiol 275:E507–E515. Russo A, Franceschi S, La Vecchia C, Dal Maso L, Montella M, Conti E, Giacosa A, Falcini F, Negri E. 1998. Body size and colorectal-cancer risk. Int J Cancer 78:161–165. Sanderson M, Williams MA, Malone KE, Stanford JL, Emanuel I, White E, Daling JR. 1996. Perinatal factors and risk of breast cancer. Epidemiology 7:34–37. Sarma AV, Jaffe CA, Schottenfeld D, Dunn R, Montie JE, Cooney KA, Wei JT. 2002. Insulin-like growth factor-1, insulin-like growth factor binding protein-3, and body mass index: Clinical correlates of prostate volume among black men. Urology 59:362–367. Schapira DV, Kumar NB, Lyman GH, Cox CE. 1990. Abdominal obesity and breast cancer risk. Ann Intern Med 112:182–186. Schapira DV, Kumar NB, Lyman GH, Cavanagh D, Roberts WS, LaPolla J. 1991. Upper-body fat distribution and endometrial cancer risk. JAMA 266:1808–1811. Schneider J, Bradlow HL, Strain G, Levin J, Anderson K, Fishman J. 1983. Effects of obesity on estradiol metabolism: Decreased formation of nonuterotropic metabolites. J Clin Endocrinol Metab 56:973–978. Schoen RE, Tangen CM, Kuller LH, Burke GL, Cushman M, Tracy RP, Dobs A, Savage PJ. 1999. Increased blood glucose and insulin, body size, and incident colorectal cancer. J Natl Cancer Inst 91:1147–1154. Schottenfeld D. 1995. Epidemiology of endometrial neoplasia. J Cell Biochem Suppl 23:151–159. Schulman CC, Ekane S, Zlotta AR. 2001. Nutrition and prostate cancer: Evidence or suspicion? Urology 58:318–334. Schuurman AG, Goldbohm RA, Dorant E, van den Brandt PA. 2000. Anthropometry in relation to prostate cancer risk in the Netherlands Cohort Study. Am J Epidemiol 151:541–549. Segal KR, Landt M, Klein S. 1996. Relationship between insulin sensitivity and plasma leptin concentration in lean and obese men. Diabetes 45:988–991. Sellers TA, Kushi LH, Potter JD, Kaye SA, Nelson CL, McGovern PG, Folsom AR. 1992. Effect of family history, body-fat distribution, and reproductive factors on the risk of postmenopausal breast cancer. N Engl J Med 326:1323–1329. Sellers TA, Gapstur SM, Potter JD, Kushi LH, Bostick RM, Folsom AR. 1993. Association of body fat distribution and family histories of breast and ovarian cancer with risk of postmenopausal breast cancer. Am J Epidemiol 138:799–803. Sellers TA, Davis J, Cerhan JR, Vierkant RA, Olson JE, Pankratz VS, Potter JD, Folsom AR. 2002. Interaction of waist/hip ratio and family history on the risk of hormone receptor-defined breast cancer in a prospective study of postmenopausal women. Am J Epidemiol 155:225–233. Senie RT, Rosen PP, Rhodes P, Lesser ML, Kinne DW. 1992. Obesity at diagnosis of breast carcinoma influences duration of disease-free survival. Ann Intern Med 116:26–32. Severson RK, Grove JS, Nomura AM, Stemmermann GN. 1988. Body mass and prostatic cancer: a prospective study. BMJ 297:713–715. Shaneyfelt T, Husein R, Bubley G, Mantzoros CS. 2000. Hormonal predictors of prostate cancer: A meta-analysis. J Clin Oncol 18:847–853. Sharpe CR, Siemiatycki J. 2001. Joint effects of smoking and body mass index on prostate cancer risk. Epidemiology 12:546–551. Shinchi K, Kono S, Honjo S, Todoroki I, Sakurai Y, Imanishi K, Nishikawa H, Ogawa S, Katsurada M, Hirohata T. 1994. Obesity and adenomatous polyps of the sigmoid colon. Jpn J Cancer Res 85:479–484. Shoff SM, Newcomb PA. 1998. Diabetes, body size, and risk of endometrial cancer. Am J Epidemiol 148:234–240. Shoff SM, Newcomb PA, Trentham-Dietz A, Remington PL, Mittendorf R, Greenberg ER, Willett WC. 2000. Early-life physical activity and postmenopausal breast cancer: effect of body size and weight change. Cancer Epidemiol Biomarkers Prev 9:591–595. Shu XO, Gao YT, Yuan JM, Ziegler RG, Brinton LA. 1989. Dietary factors and epithelial ovarian cancer. Br J Cancer 59:92–96. Shu XO, Brinton LA, Zheng W, Gao YT, Fan J, Fraumeni JF Jr. 1991. A population-based case-control study of endometrial cancer in Shanghai, China. Int J Cancer 49:38–43. Shu XO, Brinton LA, Zheng W, Swanson CA, Hatch MC, Gao YT, Fraumeni JF Jr. 1992. Relation of obesity and body fat distribution to endometrial cancer in Shanghai, China. Cancer Res 52:3865–3870.

Shu XO, Jin F, Dai Q, Shi JR, Potter JD, Brinton LA, Hebert JR, Ruan Z, Gao YT, Zheng W. 2001. Association of body size and fat distribution with risk of breast cancer among Chinese women. Int J Cancer 94:449–455. Singh PN, Fraser GE. 1998. Dietary risk factors for colon cancer in a low-risk population. Am J Epidemiol 148:761–774. Slattery ML, Potter J, Caan B, Edwards S, Coates A, Ma KN, Berry TD. 1997. Energy balance and colon cancer—Beyond physical activity. Cancer Res 57:75–80. Slattery ML, Anderson K, Curtin K, Ma K, Schaffer D, Edwards S, Samowitz W. 2001a. Lifestyle factors and Ki–ras mutations in colon cancer tumors. Mutat Res 483:73–81. Slattery ML, Potter JD, Curtin K, Edwards S, Ma KN, Anderson K, Schaffer D, Samowitz WS. 2001b. Estrogens reduce and withdrawal of estrogens increase risk of microsatellite instability-positive colon cancer. Cancer Res 61:126–130. Slattery ML, Curtin K, Ma K, Edwards S, Schaffer D, Anderson K, Samowitz W. 2002. Diet activity, and lifestyle associations with p53 mutations in colon tumors. Cancer Epidemiol Biomarkers Prev 11:541–548. Slattery ML, Ballard–Barbash R, Edwards S, Caan BJ, Potter JD. 2003. Body mass index and colon cancer: An evaluation of the modifying effects of estrogen (United States). Cancer Causes Control 14:75–84. Sonnenschein E, Toniolo P, Terry MB, Bruning PF, Kato I, Koenig KL, Shore RE. 1999. Body fat distribution and obesity in pre- and postmenopausal breast cancer. Int J Epidemiol 28:1026–1031. Spitz MR, Strom SS, Yamamura Y, Troncoso P, Babaian RJ, Scardino PT, Wheeler T, Amos CI, von Eschenbach A, Kagan J. 2000. Epidemiologic determinants of clinically relevant prostate cancer. Int J Cancer 89: 259–264. Stattin P, Soderberg S, Hallmans G, Bylund A, Kaaks R, Stenman UH, Bergh A, Olsson T. 2001. Leptin is associated with increased prostate cancer risk: A nested case-referent study. J Clin Endocrinol Metab 86:1341– 1345. Stevens J, Keil JE, Waid LR, Gazes PC. 1990. Accuracy of current, 4-year, and 28-year self-reported body weight in an elderly population. Am J Epidemiol 132:1156–1163. Subar AF, Kipnis V, Troiano R, Terry MB, Midthune D, Schoeller D, Bingham S, Runswick S, Sharbaugh C, Ballard-Barbash R, Sunshine J, Schatzkin A. 2003. The Observing Protein and Energy Nutrition (OPEN) Study 2002. Am J Epidemiol 158:1–12. Swanson CA, Brinton LA, Taylor PR, Licitra LM, Ziegler RG, Schairer C. 1989. Body size and breast cancer risk assessed in women participating in the Breast Cancer Detection Demonstration Project. Am J Epidemiol 130:1133–1141. Swanson CA, Potischman N, Wilbanks GD, Twiggs LB, Mortel R, Berman ML, Barrett RJ, Baumgartner RN, Brinton LA. 1993. Relation of endometrial cancer risk to past and contemporary body size and body fat distribution. Cancer Epidemiol Biomarkers Prev 2:321–327. Swanson CA, Coates RJ, Schoenberg JB, Malone KE, Gammon MD, Stanford JL, Shorr IJ, Potischman NA, Brinton LA. 1996. Body size and breast cancer risk among women under age 45 years. Am J Epidemiol 143:698–706. Swerdlow AJ, De Stavola BL, Floderus B, Holm NV, Kaprio J, Verkasalo PK, Mack T. 2002. Risk factors for breast cancer at young ages in twins: An international population-based study. J Natl Cancer Inst 94:1238–1246. Talamini R, La Vecchia C, Decarli A, Negri E, Franceschi S. 1986. Nutrition, social factors and prostatic cancer in a Northern Italian population. Br J Cancer 53:817–821. Talamini R, Baron AE, Barra S, Bidoli E, La Vecchia C, Negri E, Serraino D, Franceschi S. 1990. A case-control study of risk factor for renal cell cancer in northern Italy. Cancer Causes Control 1:125–131. Tamimi R, Mucci LA, Spanos E, Lagiou A, Benetou V, Trichopoulos D. 2001. Testosterone and oestradiol in relation to tobacco smoking, body mass index, energy consumption and nutrient intake among adult men. Eur J Cancer Prev 10:275–280. Tannenbaum A. 1940a. The initiation and growth of tumors: Introduction. I. Effects of underfeeding. Am J Cancer 38:335–350. Tannenbaum A. 1940b. Relationship of body weight to cancer incidence. Arch Pathol 30:509–517. Tao SC, Yu MC, Ross RK, Xiu KW. 1988. Risk factors for breast cancer in Chinese women of Beijing. Int J Cancer 42:495–498. Terry P, Baron JA, Weiderpass E, Yuen J, Lichtenstein P, Nyren O. 1999. Lifestyle and endometrial cancer risk: A cohort study from the Swedish Twin Registry. Int J Cancer 82:38–42. Terry P, Giovannucci E, Bergkvist L, Holmberg L, Wolk A. 2001. Body weight and colorectal cancer risk in a cohort of Swedish women: Relation varies by age and cancer site. Br J Cancer 85:346–349. Terry PD, Miller AB, Rohan TE. 2002. Obesity and colorectal cancer risk in women. Gut 51:191–194.

Obesity and Body Composition Thune I, Lund E. 1994. Physical activity and the risk of prostate and testicular cancer: A cohort study of 53,000 Norwegian men. Cancer Causes Control 5:549–556. Thune I, Brenn T, Lund E, Gaard M. 1997. Physical activity and the risk of breast cancer. N Engl J Med 336:1269–1275. Toniolo P, Bruning PF, Akhmedkhanov A, Bonfrer JM, Koenig KL, Lukanova A, Shore RE, Zeleniuch-Jacquotte A. 2000. Serum insulin-like growth factor-I and breast cancer. Int J Cancer 88:828–832. Tornberg SA, Carstensen JM. 1994. Relationship between Quetelet’s index and cancer of breast and female genital tract in 47,000 women followed for 25 years. Br J Cancer 69:358–361. Traish AM, Wotiz HH. 1987. Prostatic epidermal growth factor receptors and their regulation by androgens. Endocrinology 121:1461–1467. Trentham-Dietz A, Newcomb PA, Storer BE, Longnecker MP, Baron J, Greenberg ER, Willett WC. 1997. Body size and risk of breast cancer. Am J Epidemiol 145:1011–1019. Trentham-Dietz A, Newcomb PA, Egan KM, Titus-Ernstoff L, Baron JA, Storer BE, Stampfer M, Willett WC. 2000. Weight change and risk of postmenopausal breast cancer (United States). Cancer Causes Control 11:533–542. Tretli S. 1989. Height and weight in relation to breast cancer morbidity and mortality. A prospective study of 570,000 women in Norway. Int J Cancer 44:23–30. Tretli S, Magnus K. 1990. Height and weight in relation to uterine corpus cancer morbidity and mortality. A follow–up study of 570,000 women in Norway. Int J Cancer 46:165–172. Tretli S, Haldorsen T, Ottestad L. 1990. The effect of pre-morbid height and weight on the survival of breast cancer patients. Br J Cancer 62:299–303. Tretli S, Robsahm TE. 1999. Height, weight and cancer of the oesophagus and stomach: A follow-up study in Norway. Eur J Cancer Prev 8:115–122. Turner BC, Haffty BG, Narayanan L, Yuan J, Havre PA, Gumbs AA, Kaplan L, Burgaud JL, Carter D, Baserga R, Glazer PM. 1997. Insulin-like growth factor-I receptor overexpression mediates cellular radioresistance and local breast cancer recurrence after lumpectomy and radiation. Cancer Res 57:3079–3083. Tymchuk CN, Tessler SB, Aronson WJ, Barnard RJ. 1998. Effects of diet and exercise on insulin, sex hormone-binding globulin, and prostate-specific antigen. Nutr Cancer 31:127–131. Ursin G, Paganini-Hill A, Siemiatycki J, Thompson WD, Haile RW. 1995. Early adult body weight, body mass index, and premenopausal bilateral breast cancer: Data from a case-control study. Breast Cancer Res Treat 33:75–82. Valaoras VG, MacMahon B, Trichopoulos D, Polychronopoulou A. 1969. Lactation and reproductive histories of breast cancer patients in greater Athens, 1965–67. Int J Cancer 4:350–363. van Aggel–Leijssen DP, van Baak MA, Tenenbaum R, Campfield LA, Saris WH. 1999. Regulation of average 24 h human plasma leptin level; the influence of exercise and physiological changes in energy balance. Int J Obes Relat Metab Disord 23:151–158. van den Brandt PA, Spiegelman D, Yaun SS, Adami HO, Beeson L, Folsom AR, Fraser G, Goldbohm RA, Graham S, Kushi L, Marshall JR, Miller AB, Rohan T, Smith–Warner SA, Speizer FE, Willett WC, Wolk A, Hunter DJ. 2000. Pooled analysis of prospective cohort studies on height, weight, and breast cancer risk. Am J Epidemiol 152:514–527. Vatten LJ, Foss OP, Kvinnsland S. 1991. Overall survival of breast cancer patients in relation to preclinically determined total serum cholesterol, body mass index, height and cigarette smoking: a population–based study. Eur J Cancer 27:641–646. Vatten LJ, Kvinnsland S. 1992. Prospective study of height, body mass index and risk of breast cancer. Acta Oncol 31:195–200. Vaughan TL, Davis S, Kristal A, Thomas DB. 1995. Obesity, alcohol, and tobacco as risk factors for cancers of the esophagus and gastric cardia: adenocarcinoma versus squamous cell carcinoma. Cancer Epidemiol Biomarkers Prev 4:85–92. Veldhuis JD, Liem AY, South S, Weltman A, Weltman J, Clemmons DA, Abbott R, Mulligan T, Johnson ML, Pincus S. 1995. Differential impact of age, sex steroid hormones, and obesity on basal versus pulsatile growth hormone secretion in men as assessed in an ultrasensitive chemiluminescence assay. J Clin Endocrinol Metab 80:3209–3222. Verreault R, Brisson J, Deschenes L, Naud F. 1989. Body weight and prognostic indicators in breast cancer. Modifying effect of estrogen receptors. Am J Epidemiol 129:260–268. Villeneuve PJ, Johnson KC, Kreiger N, Mao Y. 1999. Risk factors for prostate cancer: results from the Canadian National Enhanced Cancer Surveillance System. The Canadian Cancer Registries Epidemiology Research Group. Cancer Causes Control 10:355–367. Walther C, Zilling T, Perfekt R, Møller T. 2001. Increasing prevalence of adenocarcinoma of the oesophagus and gastro-oesophageal junction: A study

447

of the Swedish population between 1970 and 1997. Eur J Surg 167: 748–757. Weiderpass E, Persson I, Adami HO, Magnusson C, Lindgren A, Baron JA. 2000. Body size in different periods of life, diabetes mellitus, hypertension, and risk of postmenopausal endometrial cancer. Sweden. Cancer Causes Control 11:185–192. Wenten M, Gilliland FD, Baumgartner K, Samet JM. 2002. Associations of weight, weight change, and body mass with breast cancer risk in Hispanic and non-Hispanic white women. Ann Epidemiol 12:435–4. West DW, Slattery ML, Robison LM, Schuman KL, Ford MH, Mahoney AW, Lyon JL, Sorensen AW. 1989. Dietary intake and colon cancer: Sex- and anatomic site-specific associations. Am J Epidemiol 130:883–894. West DW, Slattery ML, Robison LM, French TK, Mahoney AW. 1991. Adult dietary intake and prostate cancer risk in Utah: A case-control study with special emphasis on aggressive tumors. Cancer Causes Control 2:85– 94. Whittemore AS, Kolonel LN, Wu AH, John EM, Gallagher RP, Howe GR, Burch JD, Hankin J, Dreon DM, West DW. 1995. Prostate cancer in relation to diet, physical activity, and body size in blacks, whites, and Asians in the United States and Canada. J Natl Cancer Inst 87:652–661. WHO Consultation on Obesity. 2000. Obesity: preventing and managing the global epidemic. Report of a WHO Consultation WHO Technical Reports Series, No. 894. Geneva, Switzerland: World Health Organization. WHO Expert Committee. 1995. Physical status: the use and interpretation of anthropometry. Report of a WHO Expert Committee. WHO Technical Reports Series, No. 854. Geneva, Switzerland: World Health Organization. Willett WC, Browne ML, Bain C, Lipnick RJ, Stampfer MJ, Rosner B, Colditz GA, Hennekens CH, Speizer FE. 1985. Relative weight and risk of breast cancer among premenopausal women. Am J Epidemiol 122:731–740. Wingren G, Hatschek T, Axelson O. 1993. Determinants of papillary cancer of the thyroid. Am J Epidemiol 138:482–491. Wolf AM, Colditz GA. 1994. The cost of obesity: The US perspective. Pharmacoeconomics 5:34–37. Wolf AM, Colditz GA. 1998. Current estimates of the economic cost of obesity in the United States. Obes Res 6:97–106. Wolk A, Lindblad P, Adami HO. 1996. Nutrition and renal cell cancer. Cancer Causes Control 7:5–18. Wolk A, Mantzoros CS, Andersson SO, Bergström R, Signorello LB, Lagiou P, Adami HO, Trichopoulos D. 1998. Insulin-like growth factor 1 and prostate cancer risk: A population-based, case-control study. J Natl Cancer Inst 90:911–915. WORLD Cancer Research Fund and American Institute for Cancer Research. 1997. Food, Nutrition and the Prevention of Cancer: A Global Perspective. Washington, DC: American Institute for Cancer Research. Wu AH, Wan P, Bernstein L. 2001. A multiethnic population-based study of smoking, alcohol and body size and risk of adenocarcinomas of the stomach and esophagus (United States). Cancer Causes Control 12:721–732. Wynder EL, Mabuchi K, Whitmore WF Jr. 1971. Epidemiology of cancer of the prostate. Cancer 28:344–360. Wynder EL, Mabuchi K, Whitmore WF Jr. 1974. Epidemiology of adenocarcinoma of the kidney. J Natl Cancer Inst 53:1619–1634. Xu W, Dai Q, Ruan Z, Cheng J, Jin F, Shu X. 2002. Obesity at different ages and endometrial cancer risk factors in urban Shanghai, China. Zhonghua Liu Xing Bing Xue Za Zhi 23:347–351. Yang G, Lu G, Jin F, Dai Q, Best R, Shu XO, Chen JR, Pan XY, Shrubsole M, Zheng W. 2001. Population-based, case-control study of blood C-peptide level and breast cancer risk. Cancer Epidemiol Biomarkers Prev 10:1207–1211. Yee D, Paik S, Lebovic GS, Marcus RR, Favoni RE, Cullen KJ, Lippman ME, Rosen N. 1989. Analysis of insulin-like growth factor I gene expression in malignancy: Evidence for a paracrine role in human breast cancer. Mol Endocrinol 3:509–517. Yong LC, Brown CC, Schatzkin A, Schairer C. 1996. Prospective study of relative weight and risk of breast cancer: the Breast Cancer Detection Demonstration Project follow-up study, 1979 to 1987–1989. Am J Epidemiol 143:985–995. Yoo K, Tajima K, Park S, Kang D, Kim S, Hirose K, Takeuchi T, Miura S. 2001. Postmenopausal obesity as a breast cancer risk factor according to estrogen and progesterone receptor status. Japan. Cancer Lett 167:57–63. Yu H, Harris RE, Wynder EL. 1988. Case-control study of prostate cancer and socioeconomic factors. Prostate 13:317–325. Yu H, Rohan T. 2000. Role of the insulin-like growth factor family in cancer development and progression. J Natl Cancer Inst 92:1472–1489. Yu MC, Mack TM, Hanisch R, Cicioni C, Henderson BE. 1986. Cigarette smoking, obesity, diuretic use, and coffee consumption as risk factors for renal cell carcinoma. J Natl Cancer Inst 77:351–356.

448

PART III: THE CAUSES OF CANCER

Yuan JM, Castelao JE, Gago-Dominguez M, Ross RK, Yu MC. 1998. Hypertension, obesity and their medications in relation to renal cell carcinoma. Br J Cancer 77:1508–1513. Zeleniuch–Jacquotte A, Akhmedkhanov A, Kato I, Koenig KL, Shore RE, Kim MY, Levitz M, Mittal KR, Raju U, Banerjee S, Toniolo P. 2001. Postmenopausal endogenous oestrogens and risk of endometrial cancer: Results of a prospective study. Br J Cancer 84:975–981. Zhang S, Folsom AR, Sellers TA, Kushi LH, Potter JD. 1995. Better breast cancer survival for postmenopausal women who are less overweight and eat less fat. The Iowa Women’s Health Study. Cancer 76:275–283.

Zhang ZF, Kurtz RC, Sun M, Karpeh M Jr, Yu GP, Gargon N, Fein JS, Georgopoulos SK, Harlap S. 1996. Adenocarcinomas of the esophagus and gastric cardia: Medical conditions, tobacco, alcohol, and socioeconomic factors. Cancer Epidemiol Biomarkers Prev 5:761–768. Ziegler RG, Hoover RN, Nomura AM, West DW, Wu AH, Pike MC, Lake AJ, Horn-Ross PL, Kolonel LN, Siiteri PK, Fraumeni JF Jr. 1996. Relative weight, weight change, height, and breast cancer risk in Asian-American women. J Natl Cancer Inst 88:650–660.

23

Physical Activity I-MIN LEE AND YUKO OGUMA

T

he notion that physical activity is essential for well-being is not a new idea. As far back as 2500 b.c., we find records of organized exercise for health promotion by the ancient Chinese (Lyons and Petrucelli, 1978). Another example from history is the Greek physician Hippocrates, who lived around 400 b.c. and who also believed in the value of physical activity for well-being and for treating disease and disability (Lyons and Petrucelli, 1978). In fact, Hippocrates favored moderate levels of physical activity but disapproved of what he believed to be excess training by the professional athlete of his time, a debate that continues to the present day. With regard to epidemiologic studies of physical activity and cancer risk, perhaps the earliest such studies were published in 1922. Using mortality data among men belonging to different occupations in England and, to a lesser extent Australia and the United States, an Australian physician observed that cancer mortality rates declined with increasing “amount of hard muscular work” (Cherry, 1922). That same year, investigators from the United States wrote that “patients, clinically diagnosed as precancerous, improved in a most satisfactory manner on a prescription for increased daily exercise, as rope skipping for women, golf and brisk daily walks for men” (Sivertsen and Dahlstrom, 1922). They then embarked on a formal study of cancer mortality rates among men in Minnesota. These investigators arrived at the same conclusion as their Australian colleague, noting that cancer mortality rates declined with greater physical exertion on the job. The theory of an inverse relation between physical activity and cancer risk did not attract much attention after these pioneer studies. No studies of this topic were published subsequently until the 1950s, and only a few studies were published between the 1950s and the mid1980s. Since the mid-1980s, however, there has been great interest in this topic, with a large body of epidemiologic data now available on the association between physical activity and the risk of developing cancer (Tables 23–1 to 23–10). This large body of literature has resulted, for the first time, in specific recommendations targeting physical activity as a cancer preventive measure. In 2002, the American Cancer Society published their guidelines on nutrition, physical activity, and cancer (Byers et al., 2002). With regard to physical activity, these guidelines recommended regular physical activity for the individual to decrease the risk of developing colon and breast cancers. The guidelines further targeted community action, recommending that public, private, and community organizations work to create social and physical environments that support the adoption and maintenance of physically active behavior. In addition to the American Cancer Society, the International Agency for Research on Cancer of the World Health Organization in 2002 also recognized that physical activity reduces the risk of colon and breast cancers and, possibly, other cancers (IARC, 2002). As with the American Cancer Society, the International Agency for Research on Cancer acknowledged that the environment is an important factor to be modified in striving to promote physical activity and the control of weight gain in the individual. In this chapter, we will discuss the many epidemiologic studies of physical activity and cancer prevention whose findings have led to the recommendations above, targeting physical activity as a cancer preventive measure.

PREVALENCE OF PHYSICAL ACTIVITY IN THE UNITED STATES Two major cross-sectional surveys in the United States provide information on the prevalence of physical activity, with both using self-reports. The National Health Interview Survey (NHIS) is a population-based survey of the civilian noninstitutionalized population. This survey has periodically collected information on physical activity since 1975. The most recent data, from the 2000 survey, indicate that 31.8% of U.S. adults aged ≥18 years engaged in at least some leisure-time physical activity, 30.6% engaged in regular leisure-time physical activity (defined as light or moderate intensity activity ≥5 times/week for ≥30 minutes each time, or vigorous activity ≥3 times/week for ≥20 minutes each time) (Barnes and Schoenborn, 2002). The prevalence of leisure-time physical activity was higher in men than women and declined with older age. Higher education and income levels were associated with higher prevalence, while widowed persons had a lower prevalence compared with other marital groups. With regard to geographic distribution, adults living in the Northeast and West were more likely to have engaged in leisure-time physical activity, compared with those living in the Midwest or South. White adults reported more physical activity than did adults from other race groups. A comparison of these data with data from previous surveys is difficult because the physical activity questions used in this survey were very different from those used in previous surveys. Possibly adding to this difficulty is the fact that physical activity recommendations in the United States have changed over time. Prior to 1995, recommendations called for vigorous intensity exercise for at least 3 times/week for at least 20 minutes (ACSM, 1985). In 1995, a new recommendation was issued by the Centers for Disease Control and Prevention (CDC)and the American College of Sports Medicine (ACSM) asking instead for 30 minutes of moderate intensity physical activity, which could be accumulated over the course of the day, almost daily (Pate et al., 1995). This recommendation was not meant to replace the previous recommendations; rather, it was meant to provide a choice for less intense kinds of activities. It heightened awareness of physical activities that would not have been “counted” under the old recommendation, such as gardening and yardwork. Subsequently, a National Institutes of Health Consensus Development Panel (NIH, 1996) and the Surgeon General’s Report on Physical Activity and Health (U.S. Department of Health and Human Services, 1996) issued similar recommendations. Therefore, perception of what “counts” as physical activity may have influenced self-reports of physical activity in surveys after 1995, making comparisons over time difficult. The other major national survey collecting information on physical activity is the Behavioral Risk Factor Surveillance System (BRFSS) (Anonymous, 2004). This is a population-based survey of the civilian noninstitutionalized U.S. population aged ≥18 years, conducted using random-digit telephone dialing. Data on physical activity have been collected annually since 1990. The most recent data, from 2002, indicate that the prevalence of no leisure-time physical activity in the preceding month was 25.1%. Prior to this, the prevalence had been stable at between 30–31% from 1988 to 1996, later declining to 28.4%,

449

450

PART III: THE CAUSES OF CANCER

27.8%, and 26.2% for 1998, 2000, and 2001, respectively. With regard to the proportion achieving recommended levels of physical activity, the data are difficult to compare because surveys prior to 2001 asked about leisure-time physical activity only, but the 2001 survey included household and transportation activities as well (Anonymous, 2003).

PREVALENCE OF PHYSICAL ACTIVITY WORLDWIDE There is no internationally agreed definition or measure of physical activity. Additionally, sparse data exist for many countries throughout the world. The World Health Organization’s World Health Report 2002 attempted to obtain comparable measures of physical activity worldwide (WHO, 2002). The report decided that estimates of physical inactivity were most widely available and most comparable. It defined physical inactivity as doing very little or no physical activity in four domains: at work, for transport, in domestic duties, or during leisure time. Most data were available for leisure-time activity, fewer data existed for occupational activity, while hardly any data were available for transport- and domestic-related activity. The global estimate for the prevalence of physical inactivity among adults aged ≥15 years was 17%, ranging from 11% for parts of Africa to 24% for parts of Europe. The United States, grouped together with Canada and Cuba, had a prevalence of physical inactivity of 20%. The different estimates of the prevalence of physical activity or inactivity in the United States obtained from different surveys using different methodology underscores the difficulty in making comparisons internationally when there is no uniformly accepted definition or method for measuring this risk factor. Even slight differences with regard to the protocol used in physical activity surveys can change estimates by a large amount (Macera et al., 2001).

BIOLOGICAL BASES FOR AN INVERSE RELATION BETWEEN PHYSICAL ACTIVITY AND RISK OF CANCER The precise mechanism or mechanisms responsible for lower rates of cancer among persons who are physically active remain unknown. Several categories of plausible mechanisms have been suggested, which relate to (1) alteration of sex hormone levels, (2) alteration of body fat, (3) change in intestinal transit time, and (4) change in immune function. Sex hormones have powerful mitogenic and proliferative effects and are thus important in the etiology of reproductive cancers. With regard to breast cancer, investigators have proposed that physical activity reduces the risk of developing this cancer via its effects on menstrual function and female sex hormone levels. Girls who participate in physical activity and sports tend to have a later age at menarche and are more likely to have cycles that are anovulatory (Apter, 1996; Bernstein et al., 1987; Merzenich et al., 1993; Moisan et al., 1991; Warren, 1980). Even moderate amounts of physical activity— the equivalent of perhaps 2 hours/week—appear sufficient to have these effects (Bernstein et al., 1987). Because early age of menarche is associated with increased breast cancer risk, this could plausibly explain, in part, lower rates of breast cancer with higher levels of physical activity. Additionally, early age at menarche is associated with more rapid onset of regular, ovulatory cycles during adolescence and higher levels of estrogen during reproductive life (Apter et al., 1989; Vihko and Apter, 1984). Thus, physical activity also may reduce the cumulative exposure to sex hormones over a woman’s lifetime (Apter, 1996). Physical activity is associated with changes in female sex steroid hormones in adult women as well. Among adult premenopausal women, higher levels of physical activity are associated with lower levels of circulating estrogen and progesterone (Bullen et al., 1985; Ellison and Lager, 1986), a shortened luteal phase (Beitins et al., 1991; Loucks et al., 1989), increased frequency of cycles that are anovulatory (Broocks et al., 1990; Ellison and Lager, 1986; Russell et al., 1984), and increased incidence of oligomenorrhea and amenorrhea (Loucks and Horvath, 1985). As for postmenopausal women,

increased physical activity is associated with decreased levels of serum estradiol and estrone, even after taking into account differences in body mass index (Cauley et al., 1989). Additionally, increased circulating sex hormone binding globulin (SHBG) concentrations have been observed in women who are physically active (Haffner et al., 1995), which would result in lower amounts of free, active estrogens in circulation. In general, more subtle differences in female sex hormone levels are observed with moderate amounts of physical activity (e.g., among women where the mean level of activity was the equivalent of some 4 hours/week of moderate intensity activity [Cauley et al., 1989]), whereas more extreme changes are seen with very high levels of physical activity (e.g., among women who increased their running mileage by 30 or 50 miles/week [Boyden et al., 1983]). These changes in female sex hormone levels with physical activity, in particular, changes in estrogen levels, also may be expected to decrease the risk of developing endometrial cancer among physically active women. In men, modulation of androgen levels by physical activity has been postulated to decrease the risk of prostate cancer. Although androgen levels may be acutely elevated after a session of aerobic exercise (Fahrner and Hackney, 1998; MacConnie et al., 1986), basal levels appear lower, within physiological range, among highly trained men compared with sedentary men (Hackney et al., 1990; Struder et al., 1999; Struder et al., 1998; Wheeler et al., 1984), even after taking body mass index into account (Struder et al., 1998, 1999; Wheeler et al., 1984). Some studies have observed that both total and free androgen levels are decreased in male athletes (Hackney et al., 1990; Wheeler et al., 1984); others, only free androgen levels (Cooper et al., 1998; Struder et al., 1998, 1999), which may be a consequence of higher levels of sex-hormone binding globulin (Cooper et al., 1998). Not all studies, however, have reported lower basal androgen levels in male athletes (Cooper et al., 1998; MacConnie et al., 1986). The amount of physical activity associated with lower androgen levels is high, such as that in elite marathon runners (Hackney et al., 1990) and men who ran >65 km/week for >20 years (Struder et al., 1998, 1999). A moderate exercise program of walking 3 times/week, 30–60 minutes/session for 20 weeks did not lower testosterone levels among previously sedentary men (Struder et al., 1999). Favorable changes in adiposity represent a second major pathway through which physical activity may influence the risk of several obesity-related cancers, such as postmenopausal breast cancer, endometrial cancer, and colorectal cancer. However, studies that have controlled for body weight still have observed lower risks of these cancers among individuals who are physically active (discussed below). Greater body weight is associated with altered estrogen metabolism (Schneider et al., 1983), and obese postmenopausal women have higher levels of estrogen than their lean counterparts because of the conversion of androstenedione to estrogen in adipose tissue (Siiteri, 1987). In addition to alterations in sex hormone levels, obesity also is associated with insulin resistance, hyperinsulinemia, and hypertriglyceridemia, and may increase the level of insulin-like growth factors (IGF) (Ballard-Barbash et al., 1997). Hyperinsulinemia by itself also can produce an increase in circulating IGF-1 and a decrease in IGF binding proteins, which increases the availability of IGFs (Cotterill et al., 1992). Because insulin and IGFs have been implicated in the etiology of several cancers, such as breast, prostate, and colon cancers (Chan et al., 1998; Giovannucci, 2001; Hankinson et al., 1998; Li et al., 2001; Mantzoros et al., 1997; Toniolo et al., 2000), this may represent an additional pathway, separate from the sex hormone pathway, through which physical activity has the potential to influence cancer development. A third commonly cited explanation for lower rates of, specifically, colon cancer among physically active persons relates to change in intestinal transit time. This hypothesis states that physical activity speeds up transit time within the colon, decreasing exposure to carcinogens, cocarcinogens or promoters in the fecal stream. However, while some studies have shown faster transit time among physically active persons, not all studies have supported this (Bingham and Cummings, 1989; Coenen et al., 1992; Holdstock et al., 1970; Lampe

Physical Activity et al., 1991; Oettle, 1991). Some of the inconsistency may be partly due to different methods of measuring transit time that have varying degrees of precision. A fourth major pathway through which physical activity may prevent the development of cancer is via its effects on the innate immune system (Nieman, 1994; Pedersen and Ullum, 1994; Woods and Davis, 1994). In examining the influence of physical activity on the immune system, investigators have studied the change in numbers of various immune system cells that play a surveillance role, such as natural killer (NK) cells, lymphokine-activated killer (LAK) cells, and cytotoxic T-lymphocytes, and cells of the monocyte-macrophage system. They also have assessed several measures of immune system function in response to exercise. These measures include immune cell proliferation in response to mitogens, natural killer cell cytolytic activity, immunoglobulin synthesis, and cytokine production. Because the immune system is responsible for regulating susceptibility to infections, investigators have studied susceptibility to upper respiratory infection, as a marker of immune function, among persons of different physical activity levels. The available evidence suggests that moderate levels of physical activity can enhance the immune system. However, prolonged and intense exercise (e.g., running a marathon) may have immunosuppressive effects instead, leading to the hypothesis of a J- or U-shaped relation between immune function and level of physical activity.

EPIDEMIOLOGIC STUDIES OF PHYSICAL ACTIVITY AND CANCER RISK Measurement of Physical Activity The many different methods used in research studies to measure physical activity are beyond the scope of this chapter; detailed descriptions can be obtained from other sources (e.g., Montoye et al., 1996; Res Q Exerc Sport, 2000). The ensuing discussion will relate to methods used in epidemiologic studies of physical activity and cancer. There have been very few studies initiated specifically to examine the associations of physical activity or physical fitness with health. Examples of such studies include those conducted among British civil servants, which were designed to look at coronary heart disease (Morris et al., 1973, 1990), and the Harvard Alumni Health Study (Paffenbarger et al., 1978, 1993) and Aerobics Center Longitudinal Study (Blair et al., 1989, 1995), both designed to examine chronic diseases. However, many additional studies, designed to investigate other exposures, have collected information on physical activity in order to be able to control for confounding. The investigators of these other studies have appropriately taken advantage of their available data to examine associations between physical activity and cancer as well. Because these studies were not designed specifically to address questions regarding physical activity, details related to physical activity— the kinds of activities carried out; total volume of energy expended; the intensity, duration, and frequency of activities; and so forth—often were not available. Rather, global measures providing an indicator of the total volume of physical activity tended to be available instead. In addition, heightened awareness of and interest in the specific details of physical activity (outside of the exercise science and sports medicine community) only occurred after the 1995 CDC/ACSM recommendation that required moderate instead of vigorous intensity activity and accumulated bouts rather than a single session of activity (Pate et al., 1995). Because cancer is a relatively rare disease, cohort studies of physical activity and cancer risk involve large numbers of subjects followed for long duration. Thus, the most feasible method of assessing physical activity, and the method used to date by the majority of such studies, is using physical activity questionnaires. These questionnaires have assessed physical activity in one or more of the following domains: leisure-time, occupation, transport, and household activities. They have ranged from very simple formats, such as a single question asking about recreational exercise with the response options of “much exercise,” “moderate exercise,” or “little or no exercise” (e.g., Albanes et al., 1989), to detailed questions asking subjects to list all recreational

451

activities with their associated duration and frequency, in addition to inquiring about walking and climbing stairs to gauge nonrecreational activity (e.g., Lee et al., 1991). Many studies, especially those with more detailed assessments of physical activity, have examined the reliability and validity of their questionnaires in smaller groups of subjects. Commonly used standards against which validity has been assessed include other physical activity questionnaires, physiologic parameters known to be influenced by physical activity, diet, physical activity diaries, and mechanical or electronic devices used to measure bodily movement (Kriska and Caspersen, 1997). Apart from physical activity questionnaires, two other methods of estimating physical activity have been used in cohort studies of physical activity and cancer risk. They are, first, imputing occupational levels of physical activity based on job title (e.g., Garabrant et al., 1984) and, second, using physical fitness (either measured directly using treadmill tests [e.g., Oliveria et al., 1996] or indirectly assessed using heart rate [e.g., Gann et al., 1995]) as an indicator of usual levels of physical activity. Case-control studies of physical activity and cancer prevention also have primarily relied on physical activity questionnaires, either administered by an interviewer or completed by subjects themselves. Because interest is in physical activity prior to the development of cancer (and, perhaps, many years prior if the relevant period of exposure is, say, during adolescence), this is the most feasible method. Preexisting records of past physical activity (apart from some school, military, and job records, which may not be easily accessible) are almost nonexistent for most populations. There have been no clinical trials of physical activity and cancer prevention because of the expense and lack of feasibility in maintaining compliance among large numbers of individuals over long periods. However, randomized clinical trials of physical activity and markers of cancer risk—such as body weight; levels of sex hormones and their binding proteins; levels of insulin, IGF and binding proteins, and immune function—have been conducted (McTiernan, 2003). Such studies also contribute information to potential biological underpinnings for an inverse relation between physical activity and cancer incidence. In several of these trials, the assigned level of physical activity has corresponded to some 30–60 minutes of moderate-intensity physical activity most days of the week. This represents an attempt to define an efficacious level of physical activity that is likely to be accepted by a very sedentary U.S. population.

Physical Activity and Risk of Developing Cancer Data from existing epidemiologic studies suggest that physical activity has different associations with different site-specific cancers (Tables 23–1 to 23–10). Thus, we will discuss the different cancers separately. Below, the cancers are ordered according to the data available.

Colorectal Cancer Colorectal cancer has been the most commonly studied cancer with regard to physical activity. Many investigators have examined two or more of these outcomes—colon, rectal, or colorectal cancer—within the same study. Tables 23–1 and 23–2 summarize the data from cohort and case-control studies of physical activity and the risk of developing colorectal cancer. Because it appears that the associations are different for colon and rectal cancers, these two cancers will be examined separately. At least 50 studies, 21 cohort and 29 case-control, have examined the association between physical activity and the risk of developing colon cancer. (In addition, 17 studies, 11 cohort and 6 case-control, have investigated the relation between physical activity and the risk of developing colorectal cancer. These 17 studies will not be discussed further because the available data indicate that physical activity is inversely related to the risk of colon cancer but appears to have little association with the risk of rectal cancer. Thus, when combining the two site-specific cancers, the individual associations with physical activity may be obscured.) Of the 50 studies, 46 included male subjects, while 28 enrolled female subjects. These studies have been

Table 23–1. Summary of Results from Cohort Studies of Physical Activity and Colorectal Cancer Relative Risk, Most vs. Least Active (95% Confidence Interval) Study Site, Reference

Subjects

Colon

Rectum

USA Polednak, 1976 Chicago Persky et al., 1981

8393 college men

Sweden Gerhardsson et al., 1986 USA Paffenbarger et al., 1987

1,100,000 men, 20–64 yr

0.77 (0.67–0.83)

0.91 (0.83–1.00)

2 cohorts of 6351 longshoremen, and 16,936 college alumni 35–74 yr 430,000 men 25,000 women 11,888 retired men and women 868,620 men and women

1.4

Infinity (1.1 per 10,000 person-years divided by 0)

Men: 0.74 Women: 0.71

Men: 0.78 Women: 1.00

16,477 twin men and women, born 1886–1925 2,000,000 men and women, 20–64 yr

0.3 (0.1–0.8)

Washington state Vena et al., 1987 California Wu et al., 1987 USA Garfinkel & Stellman, 1988 Sweden Gerhardsson et al., 1988 Denmark Lynge & Thygesen, 1988

1.06 Heart rates significantly different between survivors and decedents only for first cohort

USA Waterbor et al., 1988 USA Albanes et al., 1989 Switzerland Marti & Minder, 1989 Hawaii Severson et al., 1989 Framingham Ballard-Barbash et al., 1990 USA Lee et al., 1991 USA Paffenbarger et al., 1992 China Chow et al., 1993

985 major league baseball players 5138 men and 7407 women, 25–74 yr 1,862,015 men, 15–79 yr 8006 men, born 1900–1919 1906 men and 2308 women, 30–62 yr

Finland Pukkala et al., 1993 Iowa Bostick et al., 1994 Sweden Chow et al., 1994

10,118 teachers, 20–75+ yr 35,215 women, 55–69 yr All employed persons in Sweden

USA Lee & Paffenbarger, 1994 USA Giovannucci et al., 1995 USA Steenland et al., 1995

17,607 male Harvard alumni, 34–79 yr

1.08 (0.81–1.46)

47,723 male health professionals, 40–75 yr 14,407 men and women, 25–74 yr

0.53 (0.32–0.88)

Norway Thune & Lund, 1996

53,242 men and 28,274 women, 20–69 yr 21,807 male physicians, 40–84 yr 89,488 female nurses, 40–65 yr

USA Lee et al., 1997 USA Martinez et al., 1997

452

1.57 million men and 1.55 million women, 30+ years

1.18

2.5 (0.9–10) for occupation; 0.8 (0.5– 1.4) for leisure-time Men: 1.04 (0.70–1.61) Women: 1.64 (0.64–5.88)

Adjusted for age Adjusted for age Adjusted for age, sex Adjusted for age

Adjusted for age Men: 1.0 (0.5–2.0) Women: 0.8 (0.4–1.7)

1.05

0.71 (0.51–0.99)

Adjusted for age Adjusted for age, BMI, smoking Adjusted for age, calendar year

1.76 (0.94–3.01)

0.89 (0.82–0.96)

Adjusted for age, weight, smoking, blood pressure, cholesterol Ajusted for age, social class, population density

Men: 0.40 (0.2–0.8) Women: 0.89 (0.5–1.6) Men: 1.22 Women: 1.37

Men: 0.72 (0.53–0.99) Women: 0.58 (0.37–0.94)

Comments Crude

3 cohorts of 1223, 1899, and 5784 men, 40–64 yr

17,148 male Harvard alumni, 30–79 yr 56,683 college men and women

Colorectum

Adjusted for age Adjusted for age

1.41 (0.84–2.36)

Adjusted for age, BMI

Men: 0.6 (0.3–1.0) Women: 0.9 (0.6–1.7) 0.50 (0.27–0.93)

1.72 (0.38–7.71)

Adjusted for age

1.10

2.17

Adjusted for age, sex

Men: 0.85 (0.74–0.97) Women: 0.83 (0.67–1.01) 1.61 (0.74–3.05)

Adjusted for age

Adjusted for age

0.27 (0.01–1.52)

Adjusted for age

0.95 (0.68–1.39)

Adjusted for age, height, parity, diet, vitamin A, vitamin E RR for farming, forestry, fishing, and hunting. Adjusted for age Adjusted for age, BMI, smoking, family history

Men: 0.8 Women: 0.9 1.71 (0.88–3.31)

Men: 0.98 (0.47–2.08) Women: 1.06 (0.45–2.50) Men: 0.97 (0.63–1.50) Women: 0.63 (0.39–1.04) 1.1 (0.7–1.6) 0.54 (0.33–0.90)

Adjusted for age, BMI, smoking, family history, screening, diet, aspirin Adjusted for age, income, BMI, smoking, alcohol, recreational activity Adjusted for age, BMI, geographic region Adjusted for age, BMI, alcohol, randomized treatment assignment Adjusted for age, BMI, smoking, hormone therapy, diet, aspirin

Table 23–1. (cont.) Relative Risk, Most vs. Least Active (95% Confidence Interval) Study Site, Reference USA Hsing et al., 1998

Subjects

Colon

Rectum

17,633 US men from Lutheran Brotherhood, 35+ yr 352,849 men and 510,850 women, mean age 52–53 yr

0.4 (0.2–0.9)

Finland Colbert et al., 2001

29,133 men, 50–69 yr

0.45 (0.26–0.78)

UK Wannamethee et al., 2001

7735 men, 40–59 yr

USA Will et al., 1998

Colorectum

Comments

0.4 (0.2–0.9) Men: 0.74 (0.64–0.86) Women: 0.90 (0.77–1.06) 0.50 (0.26–0.97)

0.95 (0.48–1.88)

RR for farmers vs. professionals Adjusted for age, smoking, alcohol, leisure activity Adjusted for age, race, education, BMI, smoking, diet, aspirin, pregnancy (women) Adjusted for age, randomized treatment assignment (colon, rectum), BMI, smoking (colon) Adjusted for age, social class, BMI, smoking, alcohol

BMI, body mass index; RR, relative risk.

Table 23–2. Summary of Results from Case-Control Studies of Physical Activity and Colorectal Cancer Colon

Study Site, Reference Germany Husemann et al., 1980 Los Angeles Garabrant et al., 1984 Buffalo Vena et al., 1985 USA Whittemore et al., 1985 Yugoslavia Vlajinac et al., 1987 Utah Slattery et al., 1988 Missouri Brownson et al., 1989 Sweden Fredriksson et al., 1989 Yugoslavia Jarebinski et al., 1989

Cases; Controls

Rectum Relative Risk (95% Confidence Interval)

Cases; Controls

Relative Risk (95% Confidence Interval)

Colorectum

Cases; Controls 59 men and 46 women; 99

Relative Risk (95% Confidence Interval) 0.45

Comments Crude

Men, 20–64 yr 2950; 31,724

0.6 (0.5–0.6)

Men 1213; 31,724

1.0 (0.8–1.3)

Adjusted for age

Men, 30–79 yr 210; 1431

0.48

Men, 30–79 yr 276; 1431 College men and women 53; 212

0.85

Adjusted for age

2.0 (1.0–5.0)

Adjusted for age, sex

Men: 43; 43 Women: 45; 45 Men: 110; 180 Women: 119; 204 Men 1993; 9965 Men: 165; 306 Women: 164, 317 30–75 yr

0.82

Adjusted for age, sex, neighborhood Adjusted for age, BMI, diet

Men: 0.70 (0.38–1.29) Women: 0.48 (0.27–0.87) 0.7 (0.5–1.0

Adjusted for age

Men: 0.82 Women: 0.68

Adjusted for age, county Men, 31–86 yr: 56; 56 Women, 31–86 yr: 42; 42 Men, 25–44 yr 41; 41

Los Angeles Peters et al., 1989

1.0

0.7 (0.2–2.5)

Spain Benito et al., 1990 Stockholm, Sweden Gerhardsson de Verdier et al., 1990 Japan Kato et al., 1990

Men: 163; 512 (both sexes) Women: 189; 512 (both sexes) Men, 20+ yr 1716; 16,600

Men: 0.6 (0.2–1.7) Women: 0.4 (0.2–1.1) 0.57 (0.46– 0.71)

Men: 107; 512 (both sexes) Women: 110; 512 (both sexes) Men, 20+ yr 1611; 16,600

Melbourne, Australia Kune et al., 1990

Men: 388; 398 Women: 327; 329 (total for colon and rectum)

Men: 1.44 (0.70–3.00) Women: 0.15 (0.00–1.40)

Men: 388; 398 Women: 327; 329 (total for colon and rectum)

Adjusted for age, sex, residence Men, 25–44 yr 147; 147

1.1 (0.5– 2.3)

Men: 151; 158 Women: 135, 137 <80 yr

0.7

Adjusted for age, race, neighborhood, education Adjusted for age, sex

Men: 1.1 (0.4–3.3) Women: 0.7 (0.2–2.0)

Adjusted for age, sex, BMI, diet, browned meat surface

0.80 (0.65–0.98)

Adjusted for age, residence, marital status, smoking, alcohol Adjusted for age, BMI, diet

Men: 1.58 (0.80–3.20) Women: 1.89 (0.50–6.60)

(continued)

453

Table 23–2. (cont.) Colon

Study Site, Reference Utah Slattery et al., 1990 North America & China Whittemore et al., 1990

Missouri Brownson et al., 1991 USA Markowitz et al., 1992 USA Thun et al., 1992

Cases; Controls

Relative Risk (95% Confidence Interval)

Men: 0.7 (0.4–1.4) Women: 0.5 (0.3–0.9) Men (NA): 179; 698 0.6 (0.5–0.8) Women (NA): 114; 494 Not significant Men (China): 95; 678 Women (China): 78; 618 Men, 20+ yr 1838; 17,147

0.8 (0.7–1.0)

Men 307; 1164

0.5 (0.3–0.8)

Men: 611; 3051 Women: 539; 2695

Men: 0.60 (0.28–1.27) Women: 0.90 (0.41–1.96) Men: 0.6 (0.3–1.3) Women: not significant 1.1 (0.6–2.0)

Men: 83; 314 Women: 13; 57 14–97 yr

Sweden Arbman et al., 1993

Men: 51; 203 Women: 47; 227 <76 yr Men 93; 2127

0.6 (0.3–1.1)

Men, 15–64 yr 1651; 1976

0.85 (0.74– 0.98)

Men: 74; 254 Women: 57; 209

Wisconsin Marcus et al., 1994

Women <75 yr 536; 2315

Men: 0.7 (0.4–1.3) Women: 0.9 (0.4–2.0) 0.46 (0.19– 1.10)

Japan Kotake et al., 1995 New England Longecker et al., 1995 Seattle White et al., 1996

Men: 111; 111 Women: 76; 76 Men 163; 703

0.91 (0.35– 3.33) 0.57 (0.33– 0.97)

Men: 251; 233 Women: 193; 194 30–62 yr

Men: 0.69 (0.42–1.13) Women: 0.74 (0.41–1.34)

Hawaii Marchand et al., 1997 California, Minnesota, Utah Slattery et al., 1997a California, Minnesota, Utah Slattery et al., 1997b Italy La Vecchia et al., 1999

454

Cases; Controls

Colorectum

Relative Risk (95% Confidence Interval)

Cases; Controls

Relative Risk (95% Confidence Interval)

Men: 109; 177 Women: 79; 133 40–79 yr

Turkey Vetter et al., 1992

Turkey Dosemici et al., 1993 New Zealand Fraser and Pearce, 1993 Italy Vineis et al., 1993

Rectum

Adjusted for age, BMI, diet Men (NA): 105; 698 Women (NA): 75; 494 Men (China): 131; 678 Women (China): 128; 618 Men, 20+ yr 812; 17,147 Men: 133; 1164

0.6 (0.4–0.9)

Adjusted for age, sex, BMI, diet, time in North America

Not significant

0.8 (0.6–1.3)

Adjusted for age, smoking

0.6 (0.3–1.1)

0.5 (0.3– 0.8)

Men and women, 30–79 yr 2073; 2466 Men: 688; 2073 Women: 537; 2081 23–74 yr

Men: 0.74 (0.58–0.95) Women: 0.91 (0.66–1.24) Men: 0.61 (0.47–0.79) Women: 0.63 (0.48–0.83) No family history: 0.68 (0.56–0.83) Family history; 0.95 (0.48– 1.89)

Adjusted for age, race, geography, recreational activity Adjusted for age, BMI, diet, aspirin family history Adjusted for age, smoking

Men: 48; 203 Women: 31; 227 <76 yr Men 120; 2127 Men, 15–64 yr 1046; 1976

2.1 (1.2–3.9)

Adjusted for age, sex

0.7 (0.3–1.4)

Adjusted for age, socioeconomic status, smoking Adjusted for age

0.79 (0.66–0.96)

Adjusted for age

Men: 103; 103 Women: 73; 73 Men 242; 703

Adjusted for age, BMI, family history, screening Adjusted for age, sex

0.53 (0.18–1.52) 1.19 (0.70–2.04)

Adjusted for age, race, BMI, smoking, diet, family history Adjusted for age, county Men: 698; 698 Women: 494; 494

Men and women, 30–79 yr 2053; 2410

Comments

Men: 0.6 (0.4–0.8) Women: 0.7 (0.5–1.1)

Adjusted for age, BMI, smoking, alcohol, diet, family history Adjusted for age, BMI, family history, diet, aspirin/ NSAIDs Adjusted for age, BMI, family history, diet, aspirin/ NSAIDs Adjusted for age, sex, center, education, diet

455

Physical Activity Table 23–2. (cont.) Colon

Study Site, Reference

Cases; Controls

Taiwan Tang et al., 1999

Men: 43; 92 Women: 27; 71 33–80 yr

Italy Tavani et al., 1999

Men: 688; 2073 Women: 537; 2081 <75 yr

Poland Steindorf et al., 2000 California, Minnesota, Utah Slattery et al., 2002

Rectum Relative Risk (95% Confidence Interval) Men: 0.19 (0.05–0.77) Women: 0.63 (0.18–2.18) Men: 0.64 (0.44–0.93) Women: 0.49 (0.33–0.72)

Cases; Controls Men: 49; 92 Women: 44; 71 33–80 yr Men: 435; 2073 Women: 286; 2081 <75 yr

Relative Risk (95% Confidence Interval)

Colorectum

Cases; Controls

Men: 0.44 (0.13–1.49) Women: 0.84 (0.28–2.46) Men: 1.32 (0.86–2.03) Women: 0.88 (0.48–1.60)

Comments Adjusted for age, smoking, diet, alcohol Adjusted for age, education, center, alcohol, diet

Men: 95; 95 Women: 85; 85 Men: 1099; 1290 Women: 894; 1120

Relative Risk (95% Confidence Interval)

0.6 (0.5–0.7)

0.45 (0.24– 0.84)

Adjusted for age, sex, education, diet Adjusted for age

BMI, body mass index; NSAIDs, nonsteroidal anti-inflammatory drugs.

conducted in many countries in North America, Europe, Asia, Australia, and New Zealand. Most of the studies assessed physical activity during middle age and later years. The totality of evidence indicates that physically active individuals have a lower risk of developing colon cancer, compared with inactive individuals. When comparing most active subjects with least active subjects, the median relative risk across all studies is 0.7. In general, case-control studies have observed larger effects than cohort studies, with the median relative risk being 0.6 for case-control studies and 0.8 for cohort studies. When we examined men and women separately, the corresponding median relative risk across all studies was 0.7 for men and 0.6 for women. The relative risk estimates from studies examining the association of physical activity with colon cancer risk do not appear to be confounded by other factors that predict the development of this cancer. Age-adjusted relative risks are similar to multivariate relative risks that have been adjusted for many different variables including body mass index, smoking, diet (such as energy intake, intake of fiber, micronutrients, vegetables, and meat), use of nonsteroidal anti-inflammatory drugs (NSAIDs), and screening (Giovannucci et al., 1995; Martinez et al., 1997; Slattery and Potter, 2002). Whether the association of physical activity with colon cancer risk is modified by other factors has not been well investigated. There is a suggestion that the inverse relation is stronger among individuals with heavier body mass index (Lee and Paffenbarger, 1994; Slattery and Potter, 2002), higher energy intake, and high-risk (Western style) diet (Slattery and Potter, 2002). These observations are congruent with one of the hypothesized mechanisms for physical activity and decreased risk of colon cancer—that physical activity reduces colon cancer incidence because of its beneficial effects on insulin resistance and hyperinsulinemia (Giovannucci, 2001). Another study also reported that family history modified the relation between physical activity and colon cancer risk (La Vecchia et al., 1999). How much physical activity is needed in order to lower the risk of colon cancer? Few studies have collected sufficiently detailed information on physical activity to answer this question. A study of male college alumni reported that at least 1000 kcal/week was associated with significantly lower risk (Lee et al., 1991). This level of energy expenditure can be achieved by approximately 2.5 hours/week of moderate-intensity physical activity. However, another study of male health professionals found that risk was significantly reduced only among men in the top quintile of physical activity, who expended a median of 46.8 MET-hours/week (MET-metabolic equivalent; 1.0 MET is the resting metabolic rate) (Giovannucci et al., 1995) or approximately 12 hours/week of moderate-intensity activity. Among

women, a study of female nurses observed that at least 21 METhours/week was related to significantly lower risk (Martinez et al., 1997), or approximately 5 hours/week of moderate-intensity physical activity. Yet another study of men and women found that physical activity, amounting to the equivalent of at least 60 minutes/day of vigorous activity, was necessary to significantly lower risk (Slattery et al., 1997a). Based on these data, it appears that at least 30–60 minutes/day of moderate to vigorous intensity physical activity is required to significantly lower the risk of colon cancer. Approximately two-thirds of the studies on physical activity and colon cancer risk have classified subjects according to at least three levels of physical activity, allowing for assessment of dose-response. However, many of these studies did not formally test for a significant trend across the different levels of physical activity. There is some indication of a dose-response relation: of the studies with at least three levels of physical activity, more than half reported a significant trend of declining risk with higher levels of physical activity, or observed relative risks consistent with an inverse dose-response that was not tested for statistical significance. With regard to rectal cancer, many of the studies above that investigated colon cancer also have separately examined cancers occurring in the rectum. At least 30 studies, 12 cohort and 18 case-control, have reported findings on the relation between physical activity and the risk of developing rectal cancer. Twenty-nine studies included male subjects; 14 studies, female subjects. As with the colon cancer studies, these studies of rectal cancer have been carried out in North America, Europe, Asia, and Australia. Although individual studies have reported significant inverse associations, the data on the whole do not support an association between physical activity and the risk of developing rectal cancer. Across all studies, the median relative risk comparing most with least active subjects is 1.0. Among men and women, the corresponding relative risks are 1.0 and 0.9, respectively.

Female Breast Cancer At least 57 studies, 28 cohort and 29 case-control, have investigated whether physical activity is associated with the risk of developing breast cancer in women. Tables 23–3 and 23–4 summarize the data from these studies, which have been conducted in North America, Europe, Asia, and Australia. Several of these studies investigated premenopausal and postmenopausal women separately or included only women belonging to one of these menopausal categories. Overall, the data support an inverse relation between physical activity and breast cancer incidence rates, with a smaller magnitude of association than that observed for colon cancer. Across all studies, the median relative risk for developing breast cancer, comparing most

456

PART III: THE CAUSES OF CANCER

Table 23–3. Summary of Results from Cohort Studies of Physical Activity and Female Breast Cancer Study Site, Reference USA Frisch et al., 1985 USA Paffenbarger et al., 1987 Washington state Vena et al., 1987 USA Garfinkel & Stellman, 1988 USA Albanes et al., 1989 USA Paffenbarger et al., 1992 Finland Vihko et al., 1992 Finland Pukkala et al., 1993 China Zheng et al., 1993 Framingham Dorgan et al., 1994 USA Steenland et al., 1995

Subjects

Relative Risk, Most vs. Least Active (95% Confidence Interval)

Comments

5398 college alumnae, 21–80 yr

0.54 (0.29–1.00)

4706 college alumnae

0.96

Adjusted for age, BMI, smoking, oral contraceptives, hormone therapy, reproductive variables, family history Adjusted for age

25,000 women

0.74

Adjusted for age, calendar year

868,620 men and women

1.23

Adjusted for age

7407 women, 25–74 yr

1.00 (0.63–3.33)

Adjusted for age

4706 college alumnae 2370 college alumnae, 40–50 yr 4213 teachers

0.96 0.88 (0.54–1.43) 1.28

Adjusted for age, sex Adjusted for age, BMI, family history Adjusted for age

10,118 teachers, 20–75+ yr

1.35 (0.95–1.87)

Adjusted for age

Population of urban Shanghai 30+ yr, not given (2736 cases) 2321 women, 35–68 yr

0.61

Adjusted for age

1.6 (0.9–2.9)

14,407 men and women, 25–74 yr (no separate numbers provided) 428,653 women, 30+ yr

1.16 (0.65–2.08)

Adjusted for age, education, occupation, alcohol, reproductive variables Adjusted for age, income, BMI, smoking, alcohol, menopause

0.84 (0.71–1.01)

Adjusted for age

23,341 women, >24 yr

0.91 (0.67–1.23)

Adjusted for age

25,624 women, 20–54 yr

Adjusted for age, county, BMI, height, parity

563,395 women, 30+ yr

Leisure: 0.63 (0.42–0.95) Occupation: 0.48 (0.25–0.92) 0.95 (0.75–1.20)

USA Rockhill et al., 1998 USA Sesso et al., 1998 Sweden Moradi et al., 1999 USA Rockhill et al., 1999 Finland Luoto et al., 2000

116,671 female nurses, 25–42 yr

1.1 (0.8–1.5)

1566 college alumnae

0.73 (0.46–1.14)

2 cohorts of 704,904 and 983,370 women 85,364 female nurses, 34–59 yr

0.91 (0.91–1.00) 0.91 (0.83–1.00) 0.82 (0.70–0.97)

30,548 women, 15–64 yr

Leisure: 1.01 (0.72–1.42) Occupation: 0.87 (0.62–1.24)

Iowa Moore et al., 2000 USA Wyrwich et al., 2000 USA Wyshak and Frisch, 2000 USA Breslow et al., 2001 Netherlands Dirx et al., 2001 USA Drake, 2001 USA Lee et al., 2001b Sweden Moradi et al., 2002

37,105 postmenopausal women

0.95 (0.83–1.10)

3131 women, 70–98 yr

0.42 (0.19–0.95)

3940 college alumnae

0.61 (0.44–0.84)

6160 women

0.58 (0.31–1.07)

62,537 women, 55–69 yr

0.76 (0.58–0.99)

4520 women

1.32

39,322 female health professionals, 45+ yr 9539 twin women, 42–70 yr

All: 0.80 (0.58–1.12) Postmenopausal: 0.67 (0.44–1.02) 42–50 yr: 1.3 (0.7–2.5) 51–70 yr: 0.6 (0.4–1.0);

USA Michels-Blanck et al., 1996 USA Fraser & Shavlik, 1997 Norway Thune et al., 1997 USA Calle et al., 1998

Adjusted for age, race, education, BMI, smoking, alcohol, oral contraceptives, hormone therapy, reproductive variables, breast cysts, family history Adjusted for age, height, alcohol, oral contraceptives, reproductive variables, benign breast disease, family history Adjusted for age, BMI Adjusted for age, calendar year, residence, socioeconomic status Adjusted for age, BMI at 18, height, hormone therapy, reproductive variables, benign breast disease, family history Adjusted for age, education, BMI, reproductive variables, occupational activity Adjusted for age, education, BMI, reproductive variables, leisure activity Adjusted for age, education, BMI at 18, BMI, waist–hip ratio, hormone therapy, reproductive variables, family history Adjusted for age, education, BMI, prior cancer, prior doctor visits Adjusted for age, oral contraceptives, hormone therapy, reproductive variables, family history, current exercise Adjusted for age, BMI at 25, height, adult weight change Adjusted for age, education, height, alcohol, energy intake, reproductive variables, benign breast disease, family history Crude Adjusted for age, BMI, alcohol, oral contraceptives, hormone therapy, reproductive variables, family history Adjusted for age

BMI, body mass index.

with least active women, is 0.8. As with the colon cancer studies, effect sizes were larger in case-control studies (corresponding median relative risk, 0.7) than cohort studies (0.9). In studies where data could be obtained separately for premenopausal and postmenopausal women, the median relative risk was 0.8 for women who were premenopausal and 0.7 for those who were postmenopausal. The

estimates of relative risk in most studies have been controlled for potential confounders such as body mass index, alcohol intake, use of oral contraceptives and hormone therapy, reproductive variables (ages at menarche and menopause, menopausal status, parity, age at first birth, breast feeding), benign breast disease, and family history.

Table 23–4. Summary of Results from Case-Control Studies of Physical Activity and Female Breast Cancer Study Site, Reference Turkey Dosemici et al., 1993 USA Bernstein et al., 1994 Australia Friedenreich & Rohan, 1994 Japan Hirose et al., 1995 USA Mittendorf et al., 1995 USA Taioli et al., 1995 USA Coogan et al., 1996 Italy D’Avanzo et al., 1996 Washington state McTiernan et al., 1996 Washington state Chen et al., 1997 USA Coogan et al., 1997

Cases; Controls

Relative Risk, Most vs. Least Active (95% Confidence Interval)

241; 244

1.43 (0.29–5.00)

Adjusted for age, socioeconomic status, smoking

Women, <40 yr 545; 545 Women, 20–74 yr 451; 451

0.42 (0.27–0.64)

Adjusted for age, BMI, oral contraceptives, reproductive variables, family history Adjusted for age, BMI, energy intake

1186; 23,163

Premenopausal: 0.74 (0.55–0.99) Postmenopausal: 0.72 (0.53–0.97) 0.5 (0.4–0.7)

Women, 17–74 yr 6888; 9539 617; 531 Women, <75 yr 340; 492 Women, 20–74 yr 2569; 2588 Women, 50–64 yr 537; 492 Women, 21–45 yr 747; 961 Women, <74 yr 4863; 6783

0.73 (0.50–1.05)

Premenopausal: 0.6 (0.2–1.6) Postmenopausal: 1.1 (0.5–2.6) 0.96 (0.82–1.11) Leisure: 0.76 (0.55–1.05) Occupation: 0.54 (0.33–0.89) Premenopausal: 0.9 (0.6–1.4) Postmenopausal: 0.6 (0.4–1.0) 0.95 (0.73–1.23) 0.82 (0.63–1.08)

Japan Hu et al., 1997 Atlanta, Seattle, New Jersey Gammon et al., 1998

Women, 26–75 yr 157; 369 Women, <45 yr 1668; 1505

Premenopausal: 0.72 (0.38–1.38) Postmenopausal: 1.39 (0.61–3.13) 1.01 (0.81–1.25)

Italy Mezzetti et al., 1998 Japan Ueji et al., 1998 USA Carpenter et al., 1999 Massachusetts Coogan & Aschengrau, 1999 Switzerland Levi et al., 1999 North Carolina Marcus et al., 1999 Los Angeles Enger et al., 2000 Sweden Moradi et al., 2000

Women, 20–74 yr 2569; 2588 Women, 26–69 yr 148; 296 Women, 55–64 yr 1123; 904 233; 670

0.67 (0.52–0.86)

Maine, Massachusetts, New Hampshire, Wisconsin Shoff et al., 2000 The Netherlands Verloop et al., 2000 USA Adams-Campbell et al., 2001 Canada Friedenreich et al., 2001a Friedenreich et al., 2001b New Mexico Gilliland et al., 2001 USA Lee et al., 2001a China Matthews et al., 2001 Germany Steindorf et al., 2003

Comments

Adjusted for age, year of first visit Adjusted for age, state, BMI, alcohol, reproductive variables, benign breast disease, family history Adjusted for age, education, BMI, reproductive variables Adjusted for age, state, education, BMI, alcohol, reproductive variables, benign breast disease, family history Adjusted for age, center, energy intake, reproductive variables, benign breast disease, family history Adjusted for age, education Adjusted for age Adjusted for age, education, BMI, physical activity 14–22 yr, alcohol, reproductive variables, benign breast disease, family history Adjusted for age, BMI, reproductive variables

0.9 (0.4–1.9)

Adjusted for age, center, education, income, race, marital status, BMI, BMI at 20, smoking, alcohol, energy intake, oral contraceptive, hormone therapy, reproductive variables, breast biopsy, family history Adjusted for age, center, education, BMI, alcohol, energy intake, b-carotene, vitamin E, menopause Adjusted for age, education, BMI, height, reproductive variables, family history Adjusted for age, BMI, reproductive variables, family history, interviewer Adjusted for age, education, vital status, total duration of work

Women, <75 yr 246; 374 527 White and 337 African American; 790 Premenopausal: 424; 714 Postmenopausal: 760; 2092 Postmenopausal women, 50–74 yr 3347; 3455 Postmenopausal women 4614; 5817

Leisure: 0.50 (0.30–0.81) Occupation: 0.51 (0.26–0.98) 0.6 (0.3–1.1)

Adjusted for age, education, reproductive variables, benign breast disease, family history Adjusted for age, race, sampling design

Premenopausal: 0.46 to 0.94 Postmenopausal: 0.43 to 0.69 0.77 (0.67–0.91)

Adjusted for age, socioeconomic status, BMI, alcohol, hormone therapy, reproductive variables, family history Adjusted for age, BMI, height, oral contraceptives, hormone therapy, reproductive variables

0.55 (0.39–0.78)

Adjusted for age, education, BMI at 18, reproductive variables, family history

Women, 20–54 yr 918; 918 Black women: 704; 1408 21–69 yr

0.60 (0.38–0.93)

Adjusted for age, education, smoking, family history

0.6 (0.4–0.8)

Adjusted for age, education, BMI at 18, reproductive variables, history of breast cancer

1233; 1237

Premenopausal: 1.07 (0.72–1.61) Postmenopausal: 0.70 (0.52–0.94)

Adjusted for age, education, waist–hip ratio, smoking, alcohol, hormone therapy, benign breast disease, family history

Hispanic: 332; 388 Non-Hispanic: 380; 456 35–74 yr 394; 788

Hispanic: 0.30 (0.18–0.49) Non-Hispanic: 0.67 (0.43–1.06)

Adjusted for age, oral contraceptives, reproductive variables

1.10 (0.73–1.67)

1459; 1556

0.40 (0.27–0.60)

Premenopausal women 360; 886

0.83 (0.60–1.14)

Adjusted for age, randomized treatment, BMI, alcohol, oral contraceptives, hormone therapy, reproductive variables, family history Adjusted for age, education, income, reproductive variables, breast fibroadenoma, family history Adjusted for age, height, BMI change, alcohol, reproductive variables, family history

Leisure: 0.35 (0.17–0.73) Occupation: 0.55 (0.27–1.12) 0.55 (0.37–0.83)

BMI, body mass index.

457

458

PART III: THE CAUSES OF CANCER

Most of the studies assessed physical activity during adulthood, from early adulthood through middle age and older. Several studies, in particular the case-control studies, also examined physical activity during preadolescence and the teenage years. There is no clear pattern of a stronger association between physical activity at a particular age. Apart from age, there has been little investigation of whether the relation between physical activity and breast cancer incidence is modified by other variables, with inconsistent findings. One report found a stronger association among parous than nulliparous women (Bernstein et al., 1994), but the opposite was observed in another study (Friedenreich et al., 2001a). When investigators examined body weight as an effect modifier, a stronger inverse relation between physical activity and breast cancer risk was seen among lean compared with heavier women in one study (Thune et al., 1997), but the opposite was reported in another (Shoff et al., 2000). Yet another study noted a stronger relation in women with little, compared with much, weight gain during adulthood (Carpenter et al., 1999). Two studies that investigated this relation separately for tumors that were positive for estrogen- and/or progesterone-receptors did not find any difference by hormone-receptor status (Enger et al., 2000; Lee et al., 2001b). Several studies have tried to quantify the level of physical activity required for a decreased risk of breast cancer. Investigators have reported significantly lower rates of breast cancer among women exercising at least 1 hour/week (Rockhill et al., 1999); exercising at least 3.8 hours/week (primarily vigorous exercise) (Bernstein et al., 1994); exercising to keep fit at least 4 hours/week (Thune et al., 1997); and exercising vigorously at least 7 hours/week (Adams-Campbell et al., 2001). Other investigators have observed significantly lower rates among women expending at least 1500 kcal/week (Lee et al., 2001b) (~4 hours/week of moderate-intensity activity); at least 15.3 METhours/week (~4 hours/week of moderate-intensity activity) (Ueji et al., 1998); and at least 17.6 MET-hours/week (4–5 hours/week of moderate-intensity activity) (Carpenter et al., 1999). Two reports from the same study found significantly lower rates of breast cancer only among women who exercised vigorously on a daily basis during ages 14–22 years (Mittendorf et al., 1995; Shoff et al., 2000). Finally, in a study where total physical activity over the lifetime was assessed, significantly lower breast cancer rates were seen in women who expended ≥42.7 hours/week per year in total activity (Friedenreich et al., 2001a, 2001b). Thus, it appears that at least 4–7 hours/week of moderate to vigorous intensity physical activity is required, which is similar to that observed for colon cancer. Approximately three-quarters of the studies examining physical activity classified women according to three or more levels of physical activity, allowing assessment of a dose-response relation. Of these, there appears to be an inverse dose-response, with about three-fifths reporting either a significant inverse trend across levels of physical activity, or relative risks consistent with an inverse dose-response, which was not tested for statistical significance.

Prostate Cancer There have been at least 36 studies, 22 cohort and 14 case-control, examining the association between physical activity and the risk of developing prostate cancer. These studies have been conducted in North America, Europe, and Asia. Tables 23–5 and 23–6 summarize the main results from these investigations. The epidemiologic data on the whole do not provide support for an inverse relation between physical activity and the risk of this cancer. In fact, several studies have reported strong direct associations with relative risks of 2 or higher among most active men compared with least active men (Ilic et al., 1996; Le Marchand et al., 1991; Sung et al., 1999; West et al., 1991). The median relative risk across all studies, comparing most with least active men, is 0.9. For cohort studies, this median relative risk is 0.9; for case-control studies, it is 0.8. Apart from age and race, few risk factors for prostate cancer have been established. Therefore, it is unclear what other factors may confound the relation between physical activity and prostate cancer risk. Several of the studies in Tables 23–5 and 23–6 have also adjusted the relative risk estimates for body mass index, smoking, alcohol consumption, diet (energy and fat intake), medical history (in particular, diabetes,

benign prostatic hyperplasia, vasectomy), and/or family history of prostate cancer. Although prostate cancer incidence is higher in African Americans than in other races, few data are available specifically for this group of men (Whittemore et al., 1995; Yu et al., 1988). It has been proposed that the association between physical activity and prostate cancer risk may be modified by age, as prostate cancer in younger men may be more genetically determined and less susceptible to environmental influences. However, the data have not been clear in this regard (Cerhan et al., 1997; Le Marchand et al., 1991; Lee et al., 1992). One consideration for prostate cancer is the effect of screening. Prostate-specific antigen (PSA) screening for early detection became widespread in the United States in the 1990s. If physically active men also are more health conscious, this may result in higher observed rates of prostate cancer among these men because of increased detection. An investigation of physical activity and prostate cancer, diagnosed in 1988 or earlier, among Harvard University alumni found an almost halving of prostate cancer incidence rates among men aged 70 years or older who expended ≥4000 kcal/week, compared with those expending <1000 kcal/week (Lee et al., 1992). However, an updated analyses of these men, examining prostate cancer diagnosed after 1988, did not support the earlier observations (Lee et al., 2001c). These inconsistent findings may have been due to bias arising from increased screening for prostate cancer among the most active men. Lending support to this argument is another study of health professionals, where investigators observed an approximate halving of risk of advanced prostate cancer among very active men compared with sedentary men (Giovannucci et al., 1998). The diagnosis of advanced prostate cancers is unlikely to be affected by screening. No association was noted for all prostate cancer, which included less advanced stage prostate cancers that are more susceptible to screening bias.

Lung Cancer At least 21 studies, 15 cohort and 6 case-control, have been published on the relation between physical activity and the risk of developing lung cancer. Fifteen studies included men and eight included women. These studies have been carried out in the United States and in Europe (Tables 23–7 and 23–8). The data are suggestive of an inverse relation—the median relative risk for lung cancer, comparing most with least active subjects, is 0.8. As with the previous site-specific cancers, the association was stronger for case-control studies (corresponding median relative risk, 0.5) than for cohort studies (0.8). While there have been fewer studies of women, the inverse relation between physical activity and lung cancer risk appears stronger in women (median relative risk comparing most with least active women, 0.5) than in men (0.8). A major concern with lung cancer is confounding by cigarette smoking. Almost all studies controlled for cigarette smoking; typically, the number of cigarettes smoked and the duration of smoking (or combined into a pack-years variable). However, it is difficult to be certain that the effect of cigarette smoking was completely controlled for. Residual confounding, as well as confounding by other smoking related factors (e.g., whether low-tar cigarettes or filter tips were used, the depth of inhalation when smoking, passive smoking, etc.) might still be present. Although these issues can be eliminated by conducting studies among individuals who have never smoked, lung cancer occurs at a very low frequency in these individuals and so such studies would have very limited statistical power to detect an effect of physical activity. None of the studies were able to investigate never-smokers, although nonsmokers (never and past smokers) have been separately investigated, with similar associations observed for nonsmokers and smokers (Lee and Paffenbarger, 1994; Lee et al., 1999). An interesting observation that lends some support to a true inverse association between physical activity and lung cancer risk was made in a Norwegian study (Thune and Lund, 1997). In this study, investigators examined the association separately by histologic type of lung cancer. They found significant inverse associations for small cell cancer and adenocarcinoma, but not for squamous cell cancer, in men. Although cigarette smoking increases the risk of all lung cancers, the

459

Physical Activity Table 23–5. Summary of Results from Cohort Studies of Physical Activity and Prostate Cancer Study Site, Reference USA Polednak, 1976 USA Paffenbarger et al., 1987 Washington state Vena et al., 1987 USA Albanes et al., 1989 Oahu Severson et al., 1989 USA Lee et al., 1991 China Hsing et al., 1994 USA Lee et al., 1994 Norway Thune & Lund, 1994 Chicago USA Steenland et al., 1995 Gann et al., 1995 USA Oliveria et al., 1996 Iowa Cerhan et al., 1997 USA Giovannucci et al., 1998 Finland Hartman et al., 1998 USA Clarke & Whittemore, 2000 USA Liu et al., 2000 Iowa Putnam et al., 2000 Norway Lund Nilsen et al., 2000 USA Lee et al., 2001c UK Wannamethee et al., 2001 Sweden Norman et al., 2002

Subjects

Relative Risk, Most vs. Least Active (95% Confidence Interval)

Comments

8393 college men

1.63

Crude

6531 longshoremen, 35–74 yr 16,936 college alumni, 35–74 yr 430,000 men

0.65 0.56 0.85

Adjusted for age Adjusted for age, BMI, smoking Adjusted for age, calendar year

5138 men, 25–74 yr

0.6 (0.3–1.0)

Adjusted for age

8006 men, born 1900–1919

1.05 (0.75–1.47)

Adjusted for age, BMI

17,719 college alumni, 30–79 yr

0.88 (0.64–1.23)

Adjusted for age

1.75 million men, 30+ yr

0.92 (0.7–1.1)

Adjusted for age

17,607 college alumni, 30–79 yr

0.97 (0.81–1.32)

Adjusted for age, BMI, smoking, family history

53,242 men, 19–50 yr

Leisure-time: 0.87 (0.57–1.34) Occupation: 0.81 (0.50–1.30) 1.26 (1.02–1.51) 0.76 (0.44–1.32)

Adjusted for age, geographic region, BMI

22,380 men 14,407 men and women, 25–74 yr (no separate numbers provided) 12,975 men, 20–80 yr 1050 men, 65–101 yr

Adjusted for age, education, BMI, smoking, total Adjusted for age, income, BMI, smoking, alcohol cholesterol, systolic blood pressure, glucose Adjusted for age, BMI, smoking

Physical fitness: 0.26 (0.10–0.63) Physical activity: 0.37 (0.14–0.98) 1.9 (1.0–3.5)

Adjusted for age, BMI, smoking

47,542 health professionals, 40–75 yr 29,133 men, 50–69 yr

0.90 (0.76–1.07)

Adjusted for age, height, smoking, diet, diabetes, vasectomy

0.9 (0.73–1.14)

5377 men

0.53 (0.31–0.91)

Adjusted for age, randomized treatment, residence, smoking, benign prostatic disease Adjusted for age, race, education, family history

22,071 physicians, 40–84 yr

1.12 (0.91–1.37)

1572 men, 40–86 yr

0.9 (0.5–1.5)

Adjusted for age, randomized treatment, BMI, height, smoking, hypercholesterolemia, hypertension, diabetes mellitus, multivitamins Adjusted for age

22,895 men, 40+ yr

0.80 (0.62–1.03)

Adjusted for age

8922 college alumni

1.04 (0.79–1.38)

Adjusted for age, BMI, smoking, alcohol, family history

7735 men, 40–59 yr

0.25 (0.06–0.99)

Adjusted for age, social class, BMI, smoking, alcohol

2 cohorts of 1,348,971 and 1,377,629 men

0.90 (0.87–0.93) 0.90 (0.88–0.95)

Adjusted for age, calendar year, residence

BMI, body mass index.

association is particularly strong for squamous cell and small cell lung cancers but less marked for adenocarcinoma. This may provide some indirect evidence the observed association in this study was not reflecting confounding by smoking. However, no association was observed between physical activity and lung cancer risk among women in the study.

Endometrial Cancer To date, at least 15 studies, 4 cohort and 11 case-control, have examined whether physical activity is related to the risk of endometrial cancer (Tables 23–9 and 23–10). Although this body of evidence is small, the data are suggestive of an inverse relation. The median relative risk for developing this cancer, comparing most with least active women, is 0.7. Although cohort studies of this topic have tended not to adjust for body mass index, most of the case-control studies have. Increasing body weight is strongly related to risk of endometrial cancer; however, obesity also may be viewed as an intermediate event on the causal pathway, in which case investigators would not want to control for

body weight. Regardless, case-control studies that have controlled for body mass index still have observed significant inverse associations between physical activity and endometrial cancer risk. Another variable that must be considered is the use of postmenopausal hormone therapy, in particular, use of unopposed estrogen, which increases the risk of endometrial cancer. Fewer than half of the studies investigating physical activity and endometrial cancer risk have accounted for the use of postmenopausal hormone therapy. Women using postmenopausal hormone therapy have a better health profile than those who do not (Grodstein et al., 1996); such women are more physically active compared with those who have never used such therapy (Lee et al., 2001b). Therefore, studies that did not control for the use of such hormones may be underestimating the magnitude of the inverse association.

Other Cancers Other cancers that have been examined with regard to an association with physical activity include ovarian cancer (Bertone et al., 2001, 2002; Cottreau et al., 2000; Dosemeci et al., 1993; Mink et al., 1996;

Table 23–6. Summary of Results from Case-Control Studies of Physical Activity and Prostate Cancer Study Site, Reference

Case; Controls

USA Whittemore et al., 1984 USA Whittemore et al., 1985a USA Yu et al., 1988 Missouri Brownson et al., 1991 Hawaii Le Marchand et al., 1991 Utah West et al., 1991 Turkey Dosemici et al., 1993 Sweden Andersson et al., 1995 USA and Canada Whittemore et al., 1995

College men 243; 1972 College men 77; 8084 Men 1162; 3124 Men, 20+ yr 2878; 17,147 Men 452; 899 Men 358; 679 Men 27; 2127 Men, <80 yr 256; 252 Men 1655; 1645

Yugoslavia Ilic´ et al., 1996 Taiwan Sung et al., 1999 Canada Villeneuve et al., 1999 Canada Bairati et al., 2000 China Lacey Jr et al., 2001

Men 101; 202 Men, 50–80+ yr 90; 180 Men, 50–74 yr 1623; 1623 Men 64; 546 Men, 50–94 yr 238; 471

Relative Risk, Most vs. Least Active (95% Confidence Interval)

Comments

1.7 (1.1–2.6)

Adjusted for age, college, class year

2.0 (1.0–5.0)

Adjusted for age

Whites: 0.77 (0.63–1.00) African Americans: 0.71 (0.38–1.25) 0.7 (0.6–0.8)

Adjusted for age Adjusted for age, smoking

1.67 (1–2.5)

Adjusted for age, ethnicity

0.50 (0.19–1.25)

Crude

0.3 (0.03–2.0)

Adjusted for age, socioeconomic status, smoking

0.7 (0.4–1.1)

Adjusted for age, urbanization, farming

Whites: 0.92 African Americans: 1.20 Chinese Americans: 0.70 Japanese Americans: 0.84 3.87 (2.09–7.16)

Adjusted for energy and fat intake

2.16 (1.18–3.96) 0.6 (0.4–0.9)

Adjusted for medical conditions, occupational exposures, number of brothers, number of sexual partners Adjusted for education, BMI, diet Adjusted for age, residence, race, income, BMI, smoking, alcohol, diet family history Adjusted for age, education, smoking, vitamin supplements, energy intake Adjusted for age, marital status, education, BMI, waist–hip ratio, energy intake

0.2 (0.1–0.7) 1.4 (0.9–2.1)

BMI, body mass index.

Table 23–7. Summary of Results from Cohort Studies of Physical Activity and Lung Cancer Study Site, Reference England and Wales Stukonis & Doll, 1969 USA Polednak, 1976 Chicago Persky et al., 1981 USA Paffenbarger et al., 1987 USA Garfinkel & Stellman, 1988 USA Albanes et al., 1989 Oahu Severson et al., 1989 USA Paffenbarger et al., 1992 Finland Pukkala et al., 1993 USA Lee & Paffenbarger, 1994 USA Steenland et al., 1995 Norway Thune et al., 1997 USA Lee et al., 1999 UK Wannamethee et al., 2001 Finland Colbert et al., 2002 BMI, body mass index.

460

Subjects

Relative Risk, Most vs. Least Active (95% Confidence Interval)

Comments

men, 20–64 yr

1.09

Adjusted for age

8393 men

1.07

Crude

3 cohorts of 1223, 1899, and 5784 men, 40–64 yr

Heart rates not significantly different between survivors and decedents in all cohorts 0.61 0.4 0.58 Men: 1.19 Women: 1.36

Adjusted for age, weight, smoking, blood pressure, cholesterol

3 cohorts of 3686 and 6351 longshoremen, and 16,936 college alumni, 35–74 yr 868,620 men and women

Adjusted for age, smoking, blood pressure Adjusted for age Adjusted for age, BMI, smoking Adjusted for age

5138 men, 25–74 yr

1.11 (0.67–3.33)

Adjusted for age

8006 men, born 1900–1919

0.70 (0.48–1.01)

Adjusted for age, BMI, smoking

56,683 college men and women

1.22

Adjusted for age, sex

10,118 female teachers, 20–75+ yr 17,607 male Harvard alumni, 30–79 yr 14,407 men and women, 25–74 yr 53,242 men and 28,274 women, 20–49 yr 13,905 male Harvard alumni

1.38 (0.45–3.21)

Adjusted for age

0.62 (0.45–0.85)

Adjusted for age, BMI, smoking, family history

Men: 0.79 (0.45–1.41) Women: 0.71 (0.35–2.78) Men: 0.71 (0.52–0.97) Women: 0.99 (0.35–2.78) 0.61 (0.41–0.89)

Adjusted for age, income, BMI, smoking, alcohol

7735 men, 40–59 yr

0.76 (0.40–1.43)

Adjusted for age, social class, BMI, smoking, alcohol

27,087 male smokers, 50–69 yr

Leisure-time: 1.03 (0.93–1.14) Occupation: 0.81 (0.63–1.05)

Adjusted for age, randomized treatment, education, BMI, smoking, energy intake, vegetable intake

Adjusted for age, geographical region, BMI, smoking Adjusted for age, BMI, smoking

Table 23–8. Summary of Results from Case-Control Studies of Physical Activity and Lung Cancer Study Site, Reference Missouri Brownson et al., 1991 Iowa Sellers et al., 1991 lowa Potter et al., 1992 Turkey Dosemici et al., 1993 Czech Republic Kubík et al., 2001 Czech Republic Kubík et al., 2002

Relative Risk, Most vs. Least Active (95% Confidence Interval)

Case; Controls White men 4700; 17,147 Women, 55–69 yr 152; 1900 Women, 55–69 yr 109; 1900 Men 1148; 2127 Women, 25–84 yr 140; 280 Women, 25–89 yr 269; 1079

Comments

1.25 (1.11–1.67)

Adjusted for age, job, smoking

0.5 (0.3–0.9)

Adjusted for education, smoking

0.41 (0.20–0.81)

Adjusted for education, smoking, beer consumption

1.0 (0.77–1.25)

Adjusted for age, socioeconomic status, smoking

0.38 (0.2–0.7)

Adjusted for age, residence, education, smoking

0.42 (0.29–0.62)

Adjusted for age, residence, education, smoking

BMI, body mass index.

Table 23–9. Summary of Results from Cohort Studies of Physical Activity and Endometrial Cancer Study Site, Reference China Zheng et al., 1993 Finland Pukkala et al., 1993 Sweden Moradi et al., 1998 Sweden Terry et al., 1999

Relative Risk, Most vs. Least Active (95% Confidence Interval)

Subjects

Comments

Population of urban Shanghai 30+ yr, not given (452 cases) 10,118 teachers, 20–75+ yr

0.73

Adjusted for age

1.33 (0.58–2.62)

Adjusted for age

989,270 women, <40–85+ yr

0.76 (0.67–0.85)

11,659 twin women, born 1886–1925

0.10 (0.04–0.60)

Adjusted for age, residence, calendar year, socioeconomic status Adjusted for age, weight, parity

Table 23–10. Summary of Results from Case-Control Studies of Physical Activity and Endometrial Cancer Study Site, Reference

Case; Controls

Relative Risk, Most vs. Least Active (95% Confidence Interval)

Comments

Turkey Dosemici et al., 1993 Switzerland and Italy Levi et al., 1993

31; 244

2.00 (0.11–10.00)

Adjusted for age, socioeconomic status, smoking

274; 572

0.12 (0.04–0.33)

China Shu et al., 1993 Illinois, Pennsylvania, Minnesota, North Carolina Sturgeon et al., 1993 Japan Hirose et al., 1996 Greece Kalandidi et al., 1996

Women, 18–74 yr 268; 268 Women, 20–74 yr 405; 297

0.59 (0.23–1.67)

Adjusted for age, study center, education, BMI, energy intake, parity, menopause, oral contraceptives, hormone therapy, energy intake Adjusted for age, BMI, number of pregnancies, energy intake

0.4 (0.11–1.43)

Adjusted for age, study area, education, BMI, smoking, parity, oral contraceptives, hormone therapy

Women, 20–80+ yr 145; 26,751 145; 298

0.60 (0.38–0.93)

Adjusted for age, calendar year

0.41 (0.18–0.91)

Hawaii Goodman et al., 1997 New York state Olson et al., 1997 Sweden Moradi et al., 2000

Women, 18–84 yr 332; 511 232; 631

0.7

Postmenopausal women, 50–74 yr 709; 3368 85; 668

Leisure: 0.81 (0.56–1.00) Occupation: 0.71 (0.53–1.0)

Adjusted for age, education, BMI, height, smoking, alcohol, coffee, energy intake, age at menarche, parity, miscarriages, induced abortions, oral contraceptives, hormone therapy Adjusted for age, BMI, energy intake, pregnancy history, oral contraceptives, hormone therapy Adjusted for age, education, BMI, smoking, parity, age at menarche, menopause, hormone therapy, diabetes mellitus Adj for age, smoking, BMI, parity, age at last birth, oral contraceptives, hormone therapy, menopause

Mexico Salazar-Martínez et al., 2000 Washington State Women, 45–74 yr Littman et al., 2001 822; 1111

0.67 (0.42–1.09)

0.47 (0.26–0.86) 0.83 (0.59–1.15)

Adjusted for age, BMI, smoking, menopause, hypertension, diabetes mellitus, anovulatory index Adjusted for age, county, BMI, hormone therapy

BMI, body mass index.

461

462

PART III: THE CAUSES OF CANCER

Pukkala et al., 1993; Tavani et al., 2001; Zheng et al., 1993), testicular cancer (Brownson et al., 1991; Coldman et al., 1982; Dosemeci et al., 1993; Gallagher et al., 1995; Paffenbarger et al., 1987; Srivastava and Kreiger, 2000; Thune and Lund, 1994; United Kingdom Testicular Cancer Study Group, 1994; Whittemore et al., 1984), pancreatic cancer (Brownson et al., 1991; Hanley et al., 2001; Lee and Paffenbarger, 1994; Lee et al., 2003; Michaud et al., 2001; Paffenbarger et al., 1987; Pukkala et al., 1993; Waterbor et al., 1988), kidney cancer (Bergstrom et al., 1999; Bergstrom et al., 2001; Brownson et al., 1991; Goodman et al., 1986; Lindblad et al., 1994; Mellemgaard et al., 1994; Mellemgaard et al., 1995; Paffenbarger et al., 1987; Pukkala et al., 1993; Whittemore et al., 1984), bladder cancer (Brownson et al., 1991; Dosemeci et al., 1993; Paffenbarger et al., 1987; Severson et al., 1989; Wannamethee et al., 2001; Whittemore et al., 1984) and hematopoetic cancers (Brownson et al., 1991; Cerhan et al., 2002; Paffenbarger et al., 1987; Wannamethee et al., 2001). The data are too sparse to make any conclusions regarding whether an association exists for any of these cancers.

DISCUSSION At present, there is a large body of epidemiologic data on the relation between physical activity and the risk of developing cancer. Although the direct evidence on this relation comes only from observational studies, randomized clinical trials have provided indirect evidence by examining the association of physical activity with markers of cancer risk, such as body weight and hormone levels. Moreover, several plausible biological mechanisms support the hypothesis that higher levels of physical activity decrease the incidence of various cancers. The data are clearest for colon and breast cancer, with case-control and cohort studies supporting a moderate, inverse relation between physical activity and the development of these cancers. Although individual studies have reported as much as a halving or more of colon and breast cancer incidence rates among active compared with sedentary subjects, the overall data suggest a more modest reduction in risk. Active persons have perhaps a 30% reduction in colon cancer risk compared with those who are sedentary. For breast cancer, the risk reduction appears somewhat smaller, for an average of perhaps 20%. There also appears to be a dose-response relation, with risk decreasing at higher levels of physical activity. However, there is little information regarding what additional amounts and intensity of physical activity are associated with additional risk reductions; it also is unclear what the magnitude of the additional decrements in risk may be. This lack of information stems from the fact that most of the studies examining the relation between physical activity and the risk of cancer were not designed specifically to address this question and, thus, did not collect detailed information on physical activity. The available evidence suggests that at least 30–60 minutes/day of moderate to vigorous intensity physical activity is required to significantly lower the risk of colon and breast cancer. There are fewer data on lung and endometrial cancers. On average, the data from epidemiologic studies suggest that active persons have perhaps a 30% reduction in risk of lung cancer compared with those who are sedentary. A major concern, however, is how well cigarette smoking has been controlled for in these studies. For endometrial cancer, the data also indicate a 30% risk reduction; however, because many studies did not control for postmenopausal hormone therapy, this may be a biased underestimate. With regard to rectal cancer, the data overall do not support any association with physical activity, while conflicting data exist for prostate cancer. Too few data exist regarding the other site-specific cancers to make reasonable conclusions. In general, case-control studies of physical activity and cancer risk have reported inverse associations of larger magnitude than cohort studies. There may be several explanations for this. First, recall bias may be operating. However, this is not very likely because an inverse association between physical activity and cancer risk only has been widely publicized in the past couple of years, and most of the studies on this topic were conducted earlier. A second reason may be an underestimate of the association in cohort studies because of random mis-

classification resulting from a change in subjects’ physical activity. Most of the cohort studies assessed physical activity only once at baseline and followed subjects for many years for the development of cancer, without updating physical activity levels over time. However, case-control studies were able to ascertain physical activity that was more proximate to the development of the cancer in cases and the reference date in controls. A related concern is that cohort studies also may not be assessing physical activity level during the appropriate window of time, as these studies generally ascertained physical activity only at baseline. On the other hand, many case-control studies have assessed physical activity at different times during the lifetime. In spite of this, the data have not been clear about what the appropriate timing of physical activity should be in order to decrease the risk of colon or breast cancers. The discussion in this chapter has focused primarily on the role of physical activity in the primary prevention of cancer. Few data are available regarding the effects of physical activity on health outcomes among persons who already have cancer. These data come from studies published mainly in the psychology and nursing literature and can be divided into two groups: those where physical activity carried out concurrently with medical treatment was of interest, and those where physical activity carried out after the completion of medical treatment was investigated. The former studies are framed around a coping model, where exercise is viewed as a way of mitigating the common side effects or symptoms of cancer treatment. There have been few such studies, mostly involving fewer than 100 patients (Courneya, 2003). Based on these limited data, the findings suggest that physical activity during cancer treatment is associated with better outcomes with regard to symptoms (such as nausea, fatigue, sleep disturbance), anxiety and depression, weight gain, and functional capacity. The second group of studies, investigating physical activity in patients who have completed medical treatment, has as its focus health promotion (Courneya, 2003). There also have been few such studies, generally enrolling fewer than 100 patients. The health outcomes investigated have been short-term outcomes such as physical fitness, immune function, and quality of life (e.g., depression, anxiety, and satisfaction with life). These limited data suggest that physical activity can improve these short-term outcomes. At present, very few data are available on whether physical activity in cancer patients can favorably influence the progression of the disease or its prognosis. A recently published study indicated that among women with breast cancer, physical activity of 3 MET-hours/week (approximately 1 hour/week of walking at 3 mph) or more was associated with reduced mortality (Holmes et al., 2005). Although we have learned much about the relation between physical activity and cancer risk, particularly in the past two decades, much more work is needed on this topic. It is important to clarify the mechanisms that underpin this association. An understanding of the different mechanisms that operate can help explain why physical activity has an effect on certain site-specific cancers, but not others, and what cancers might be influenced. In addition to mechanistic studies, exploring subgroups in epidemiologic studies also may provide indirect information regarding biological pathways. For example, finding a stronger inverse association between physical activity and colon cancer incidence rates among overweight individuals helps to support the hypothesis that one path through which physical activity operates is via the insulin pathway. Although the overall epidemiologic data provide strong support for an inverse association between physical activity and certain cancers, a clear limitation of the available data is lack of information regarding physical activity details. Because the U.S. public is so sedentary, it is important to ascertain the minimum, effective level of physical activity to decrease cancer risk. Additionally, in order to provide focused public health messages on cancer prevention, details on amount, intensity, frequency, and duration of physical activity need to be clarified. The timing of physical activity (what age or ages?), as well as whether changing from a sedentary to an active lifestyle makes any difference, are further important details. Because there is already a large body of epidemiologic data on colon and breast cancers, further studies examining the association between physical activity, crudely

Physical Activity assessed, and the risk of these cancers are not helpful. Future studies of these cancers should seek to address the gaps in knowledge, rather than confirming what is known. In order to provide useful information on physical activity details, physical activity questionnaires— which clearly are most appropriate for large-scale studies—must be reliable and valid for the population studied. Further work needs to be carried out with respect to questionnaire development, particularly when used in women and minorities. Apart from the main cancers discussed above, the data are sparse on the association of physical activity with the risk of developing other site-specific cancers. Finally, little information is available on whether physical activity can improve survival among patients with cancer; physical activity appears to help such patients tolerate cancer treatment better and also may be associated with better short-term health outcomes. In conclusion, physical activity clearly reduces the rates of cardiovascular disease, the leading cause of death in U.S. men and women (U.S. Department of Health and Human Services, 1996). Recently, a large body of data also indicates that physical activity can reduce the risk of cancer, the second leading cause of death in the United States, in particular, colon cancer in men and women and breast cancer in women. References ACSM (American College of Sports Medicine). 1985. Guidelines for exercise testing and exercise prescription. Philadelphia: Lea & Febiger. Adams–Campbell LL, Rosenberg L, Rao RS, Palmer JR. 2001. Strenuous physical activity and breast cancer risk in African-American women. J Natl Med Assoc 93:267–75. Albanes D, Blair A, Taylor PR. 1989. Physical activity and risk of cancer in the NHANES I population. Am J Public Health 79:744–50. Andersson SO, Baron J, Wolk A, Lindgren C, Bergstrom R, Adami HO. 1995. Early life risk factors for prostate cancer: A population-based case-control study in Sweden. Cancer Epidemiol Biomarkers Prev 4:187–92. Anonymous. 2003. Prevalence of physical activity, including lifestyle activities among adults—United States, 2000–2001. MMWR Morb Mortal Wkly Rep 52:764–769. Anonymous. 2004. Prevalence of no leisure-time physical activity 35 states and the District of Columbia, 1988–2002. MMWR Morb Mortal Wkly Rep 53:82–86. Apter D. 1996. Hormonal events during female puberty in relation to breast cancer risk. Eur J Cancer Prev 5:476–82. Apter D, Reinila M, Vihko R. 1989. Some endocrine characteristics of early menarche, a risk factor for breast cancer, are preserved into adulthood. Int J Cancer 44:783–7. Arbman G, Axelson O, Fredriksson M, Nilsson E, Sjodahl R. 1993. Do occupational factors influence the risk of colon and rectal cancer in different ways? Cancer 72:2543–9. Bairati I, Larouche R, Meyer F, Moore L, Fradet Y. 2000. Lifetime occupational physical activity and incidental prostate cancer (Canada). Cancer Causes Control 11:759–64. Ballard-Barbash R, Birt DF, Kestin M, King IB. 1997. Perspectives on integrating experimental and epidemiologic research on diet, anthropometry and breast cancer. J Nutr 127:936S–9S. Ballard-Barbash R, Schatzkin A, Albanes D, Schiffman MH, Kreger BE, et al. 1990. Physical activity and risk of large bowel cancer in the Framingham Study. Cancer Res 50:3610–3. Barnes PM, Schoenborn CA. 2003. Physical activity among adults: United States, 2000. Advance data from vital and health statistics; no. 333. Hyattsville, MD: National Center for Health Statistics. Beitins IZ, McArthur JW, Turnbull BA, Skrinar GS, Bullen BA. 1991. Exercise induces two types of human luteal dysfunction:confirmation by urinary free progesterone. J Clin Endocrinol Metab 72:1350–8. Benito E, Obrador A, Stiggelbout A, Bosch FX, Mulet M, et al. 1990. A population-based case-control study of colorectal cancer in Majorca. I. Dietary factors. Int J Cancer 45:69–76. Bergstrom A, Moradi T, Lindblad P, Nyren O, Adami HO, Wolk A. 1999. Occupational physical activity and renal cell cancer: A nationwide cohort study in Sweden. Int J Cancer 83:186–91. Bergstrom A, Terry P, Lindblad P, Lichtenstein P, Ahlbom A, et al. 2001. Physical activity and risk of renal cell cancer. Int J Cancer 92:155–7. Bernstein L, Henderson BE, Hanisch R, Sullivan-Halley J, Ross RK. 1994. Physical exercise and reduced risk of breast cancer in young women. J Natl Cancer Inst 86:1403–8. Bernstein L, Ross RK, Lobo RA, Hanisch R, Krailo MD, Henderson BE. 1987. The effects of moderate physical activity on menstrual cycle patterns

463

in adolescence: implications for breast cancer prevention. Br J Cancer 55:681–5. Bertone ER, Newcomb PA, Willett WC, Stampfer MJ, Egan KM. 2002. Recreational physical activity and ovarian cancer in a population–based case–control study. Int J Cancer 99:431–6. Bertone ER, Willett WC, Rosner BA, Hunter DJ, Fuchs CS, et al. 2001. Prospective study of recreational physical activity and ovarian cancer. J Natl Cancer Inst 93:942–8. Bingham SA, Cummings JH. 1989. Effect of exercise and physical fitness on large intestinal function. Gastroenterology 97:1389–99. Blair SN, Kohl HW 3rd, Barlow CE, Paffenbarger RS Jr, Gibbons LW, Macera CA. 1995. Changes in physical fitness and all-cause mortality. A prospective study of healthy and unhealthy men. JAMA 273:1093–8. Blair SN, Kohl HW 3rd, Paffenbarger RS Jr, Clark DG, Cooper KH, Gibbons LW. 1989. Physical fitness and all-cause mortality. A prospective study of healthy men and women. JAMA 262:2395–401. Bostick RM, Potter JD, Kushi LH, Sellers TA, Steinmetz KA, et al. 1994. Sugar, meat, and fat intake, and non–dietary risk factors for colon cancer incidence in Iowa women (United States). Cancer Causes Control 5: 38–52. Boyden TW, Pamenter RW, Stanforth P, Rotkis T, Wilmore JH. 1983. Sex steroids and endurance running in women. Fertil Steril 39:629–32. Breslow RA, Ballard-Barbash R, Munoz K, Graubard BI. 2001. Long–term recreational physical activity and breast cancer in the National Health and Nutrition Examination Survey I epidemiologic follow-up study. Cancer Epidemiol Biomarkers Prev 10:805–8. Broocks A, Pirke KM, Schweiger U, Tuschl RJ, Laessle RG, et al. 1990. Cyclic ovarian function in recreational athletes. J Appl Physiol 68:2083–6. Brownson RC, Chang JC, Davis JR, Smith CA. 1991. Physical activity on the job and cancer in Missouri. Am J Public Health 81:639–42. Brownson RC, Zahm SH, Chang JC, Blair A. 1989. Occupational risk of colon cancer. An analysis by anatomic subsite. Am J Epidemiol 130:675–87. Bullen BA, Skrinar GS, Beitins IZ, von Mering G, Turnbull BA, McArthur JW. 1985. Induction of menstrual disorders by strenuous exercise in untrained women. N Engl J Med 312:1349–53. Byers T, Nestle M, McTiernan A, Doyle C, Currie-Williams A, et al. 2002. American Cancer Society guidelines on nutrition and physical activity for cancer prevention: Reducing the risk of cancer with healthy food choices and physical activity. CA Cancer J Clin 52:92–119. Calle EE, Murphy TK, Rodriguez C, Thun MJ, Heath CW Jr. 1998. Occupation and breast cancer mortality in a prospective cohort of US women. Am J Epidemiol 148:191–7. Carpenter CL, Ross RK, Paganini–Hill A, Bernstein L. 1999. Lifetime exercise activity and breast cancer risk among post-menopausal women. Br J Cancer 80:1852–8. Cauley JA, Gutai JP, Kuller LH, LeDonne D, Powell JG. 1989. The epidemiology of serum sex hormones in postmenopausal women. Am J Epidemiol 129:1120–31. Cerhan JR, Janney CA, Vachon CM, Habermann TM, Kay NE, et al. 2002. Anthropometric characteristics, physical activity, and risk of nonHodgkin’s lymphoma subtypes and B-cell chronic lymphocytic leukemia: a prospective study. Am J Epidemiol 156:527–35. Cerhan JR, Torner JC, Lynch CF, Rubenstein LM, Lemke JH, et al. 1997. Association of smoking, body mass, and physical activity with risk of prostate cancer in the Iowa 65+ Rural Health Study (United States). Cancer Causes Control 8:229–38. Chan JM, Stampfer MJ, Giovannucci E, Gann PH, Ma J, et al. 1998. Plasma insulin-like growth factor-I and prostate cancer risk: A prospective study. Science 279:563–6. Chen CL, White E, Malone KE, Daling JR. 1997. Leisure-time physical activity in relation to breast cancer among young women (Washington, United States). Cancer Causes Control 8:77–84. Cherry T. 1922. A theory of cancer. Med J Australia 1:425–38. Chow WH, Dosemeci M, Zheng W, Vetter R, McLaughlin JK, et al. 1993. Physical activity and occupational risk of colon cancer in Shanghai, China. Int J Epidemiol 22:23–9. Chow WH, Malker HS, Hsing AW, McLaughlin JK, Weiner JA, et al. 1994. Occupational risks for colon cancer in Sweden. J Occup Med 36:647–51. Clarke G, Whittemore AS. 2000. Prostate cancer risk in relation to anthropometry and physical activity: The National Health and Nutrition Examination Survey I Epidemiological Follow-Up Study. Cancer Epidemiol Biomarkers Prev 9:875–81. Coenen C, Wegener M, Wedmann B, Schmidt G, Hoffmann S. 1992. Does physical exercise influence bowel transit time in healthy young men? Am J Gastroenterol 87:292–5. Colbert LH, Hartman TJ, Malila N, Limburg PJ, Pietinen P, et al. 2001. Physical activity in relation to cancer of the colon and rectum in a cohort of male smokers. Cancer Epidemiol Biomarkers Prev 10:265–8.

464

PART III: THE CAUSES OF CANCER

Colbert LH, Hartman TJ, Tangrea JA, Pietinen P, Virtamo J, et al. 2002. Physical activity and lung cancer risk in male smokers. Int J Cancer 98:770–3. Coldman AJ, Elwood JM, Gallagher RP. 1982. Sports activities and risk of testicular cancer. Br J Cancer 46:749–56. Coogan PF, Aschengrau A. 1999. Occupational physical activity and breast cancer risk in the upper Cape Cod cancer incidence study. Am J Ind Med 36:279–85. Coogan PF, Clapp RW, Newcomb PA, Mittendorf R, Bogdan G, et al. 1996. Variation in female breast cancer risk by occupation. Am J Ind Med 30:430–7. Coogan PF, Newcomb PA, Clapp RW, Trentham-Dietz A, Baron JA, Longnecker MP. 1997. Physical activity in usual occupation and risk of breast cancer (United States). Cancer Causes Control 8:626–31. Cooper CS, Taaffe DR, Guido D, Packer E, Holloway L, Marcus R. 1998. Relationship of chronic endurance exercise to the somatotropic and sex hormone status of older men. Eur J Endocrinol 138:517–23. Cotterill AM, Holly JM, Taylor AM, Davies SC, Coulson VJ, et al. 1992. The insulin-like growth factor (IGF)-binding proteins and IGF bioactivity in Laron-type dwarfism. J Clin Endocrinol Metab 74:56–63. Cottreau CM, Ness RB, Kriska AM. 2000. Physical activity and reduced risk of ovarian cancer. Obstet Gynecol 96:609–14. Courneya KS. 2003. Exercise in cancer survivors:biopsychosocial outcomes. Med Sci Sports Exerc 35:1846–1852. D’Avanzo B, Nanni O, La Vecchia C, Franceschi S, Negri E, et al. 1996. Physical activity and breast cancer risk. Cancer Epidemiol Biomarkers Prev 5:155–60. Dirx MJ, Voorrips LE, Goldbohm RA, van den Brandt PA. 2001. Baseline recreational physical activity, history of sports participation, and postmenopausal breast carcinoma risk in the Netherlands Cohort Study. Cancer 92:1638–49. Dorgan JF, Brown C, Barrett M, Splansky GL, Kreger BE, et al. 1994. Physical activity and risk of breast cancer in the Framingham Heart Study. Am J Epidemiol 139:662–9. Dosemeci M, Hayes RB, Vetter R, Hoover RN, Tucker M, et al. 1993. Occupational physical activity, socioeconomic status, and risks of 15 cancer sites in Turkey. Cancer Causes Control 4:313–21. Drake DA. 2001. A longitudinal study of physical activity and breast cancer prediction. Cancer Nurs 24:371–7. Ellison PT, Lager C. 1986. Moderate recreational running is associated with lowered salivary progesterone profiles in women. Am J Obstet Gynecol 154:1000–3. Enger SM, Ross RK, Paganini-Hill A, Carpenter CL, Bernstein L. 2000. Body size, physical activity, and breast cancer hormone receptor status: results from two case-control studies. Cancer Epidemiol Biomarkers Prev 9: 681–7. Fahrner CL, Hackney AC. 1998. Effects of endurance exercise on free testosterone concentration and the binding affinity of sex hormone binding globulin (SHBG). Int J Sports Med 19:12–5. Fraser G, Pearce N. 1993. Occupational physical activity and risk of cancer of the colon and rectum in New Zealand males. Cancer Causes Control 4:45–50. Fraser GE, Shavlik D. 1997. Risk factors, lifetime risk, and age at onset of breast cancer. Ann Epidemiol 7:375–82. Fredriksson M, Bengtsson NO, Hardell L, Axelson O. 1989. Colon cancer, physical activity, and occupational exposures. A case-control study. Cancer 63:1838–42. Friedenreich CM, Bryant HE, Courneya KS. 2001a. Case-control study of lifetime physical activity and breast cancer risk. Am J Epidemiol 154: 336–47. Friedenreich CM, Courneya KS, Bryant HE. 2001b. Relation between intensity of physical activity and breast cancer risk reduction. Med Sci Sports Exerc 33:1538–45. Friedenreich CM, Rohan TE. 1995. Physical activity and risk of breast cancer. Eur J Cancer Prev 4:145–51. Frisch RE, Wyshak G, Albright NL, Albright TE, Schiff I, et al. 1985. Lower prevalence of breast cancer and cancers of the reproductive system among former college athletes compared to non-athletes. Br J Cancer 52:885–91. Gallagher RP, Huchcroft S, Phillips N, Hill GB, Coldman AJ, et al. 1995. Physical activity, medical history, and risk of testicular cancer (Alberta and British Columbia, Canada). Cancer Causes Control 6:398–406. Gammon MD, Schoenberg JB, Britton JA, Kelsey JL, Coates RJ, et al. 1998. Recreational physical activity and breast cancer risk among women under age 45 years. Am J Epidemiol 147:273–80. Gann PH, Daviglus ML, Dyer AR, Stamler J. 1995. Heart rate and prostate cancer mortality: Results of a prospective analysis. Cancer Epidemiol Biomarkers Prev 4:611–6. Garabrant DH, Peters JM, Mack TM, Bernstein L. 1984. Job activity and colon cancer risk. Am J Epidemiol 119:1005–14.

Garfinkel L, Stellman SD. 1988. Mortality by relative weight and exercise. Cancer 62:1844–50. Gerhardsson de Verdier M, Steineck G, Hagman U, Rieger A, Norell SE. 1990. Physical activity and colon cancer:a case–referent study in Stockholm. Int J Cancer 46:985–9. Gerhardsson M, Floderus B, Norell SE. 1988. Physical activity and colon cancer risk. Int J Epidemiol 17:743–6. Gerhardsson M, Norell SE, Kiviranta H, Pedersen NL, Ahlbom A. 1986. Sedentary jobs and colon cancer. Am J Epidemiol 123:775–80. Gilliland FD, Li YF, Baumgartner K, Crumley D, Samet JM. 2001. Physical activity and breast cancer risk in hispanic and non-hispanic white women. Am J Epidemiol 154:442–50. Giovannucci E. 2001. Insulin, insulin-like growth factors and colon cancer: A review of the evidence. J Nutr 131:3109S–20S. Giovannucci E, Ascherio A, Rimm EB, Colditz GA, Stampfer MJ, Willett WC. 1995. Physical activity, obesity, and risk for colon cancer and adenoma in men. Ann Intern Med 122:327–34. Giovannucci E, Leitzmann M, Spiegelman D, Rimm EB, Colditz GA, et al. 1998. A prospective study of physical activity and prostate cancer in male health professionals. Cancer Res 58:5117–22. Goodman MT, Hankin JH, Wilkens LR, Lyu LC, McDuffie K, et al. 1997. Diet, body size, physical activity, and the risk of endometrial cancer. Cancer Research. 57:5077–85. Goodman MT, Morgenstern H, Wynder EL. 1986. A case-control study of factors affecting the development of renal cell cancer. Am J Epidemiol 124:926–41. Grodstein F, Stampfer MJ, Manson JE, Colditz GA, Willett WC, et al. 1996. Postmenopausal estrogen and progestin use and the risk of cardiovascular disease. N Engl J Med 335:453–61. Hackney AC, Sinning WE, Bruot BC. 1990. Hypothalamic–pituitary– testicular axis function in endurance–trained males. Int J Sports Med 11:298–303. Haffner SM, Newcomb PA, Marcus PM, Klein BE, Klein R. 1995. Relation of sex hormones and dehydroepiandrosterone sulfate (DHEA-SO4) to cardiovascular risk factors in postmenopausal women. Am J Epidemiol 142:925–34. Hankinson SE, Willett WC, Colditz GA, Hunter DJ, Michaud DS, et al. 1998. Circulating concentrations of insulin–like growth factor-I and risk of breast cancer. Lancet 351:1393–6. Hanley AJ, Johnson KC, Villeneuve PJ, Mao Y. 2001. Physical activity, anthropometric factors and risk of pancreatic cancer: results from the Canadian enhanced cancer surveillance system. Int J Cancer 94:140–7. Hartman TJ, Albanes D, Rautalahti M, Tangrea JA, Virtamo J, et al. 1998. Physical activity and prostate cancer in the Alpha-Tocopherol, Beta-Carotene (ATBC) Cancer Prevention Study (Finland). Cancer Causes Control 9:11–8. Hirose K, Tajima K, Hamajima N, Inoue M, Takezaki T, et al. 1995. A large-scale, hospital-based case-control study of risk factors of breast cancer according to menopausal status. Jpn J Cancer Res 86: 146–54. Hirose K, Tajima K, Hamajima N, Takezaki T, Inoue M, et al. 1996. Subsite (cervix/endometrium)-specific risk and protective factors in uterus cancer. Jpn J Cancer Res 87:1001–9. Holdstock DJ, Misiewicz JJ, Smith T, Rowlands EN. 1970. Propulsion (mass movements) in the human colon and its relationship to meals and somatic activity. Gut 11:91–9. Holmes MD, Chen WJ, Feskanich D, Kroenke CH, Colditz GA. 2005. Physical activity and survival after breast cancer diagnosis. JAMA 293: 2479–2486. Hsing AW, McLaughlin JK, Chow WH, Schuman LM, Co Chien HT, et al. 1998. Risk factors for colorectal cancer in a prospective study among U.S. white men. Int J Cancer 77:549–53. Hsing AW, McLaughlin JK, Zheng W, Gao YT, Blot WJ. 1994. Occupation, physical activity, and risk of prostate cancer in Shanghai, People’s Republic of China. Cancer Causes Control 5:136–40. Hu YH, Nagata C, Shimizu H, Kaneda N, Kashiki Y. 1997. Association of body mass index, physical activity, and reproductive histories with breast cancer: a case-control study in Gifu, Japan. Breast Cancer Res Treatment 43:65–72. Husemann B, Neubauer MG, Duhme C. 1980. [Sedentary occupation and recto–sigmoidal neoplasms]. Onkologie 3:168–71. IARC. 2002. Weight Control and Physical Activity. IARC Working Group on the Evaluation of Cancer-Preventive Agents. Lyon: IARC. Ilic M, Vlajinac H, Marinkovic J. 1996. Case–control study of risk factors for prostate cancer. Br J Cancer 74:1682–6. Jarebinski M, Adanja B, Vlajinac H. 1989. Case-control study of relationship of some biosocial correlates to rectal cancer patients in Belgrade, Yugoslavia. Neoplasma 36:369–74.

Physical Activity Kalandidi A, Tzonou A, Lipworth L, Gamatsi I, Filippa D, Trichopoulos D. 1996. A case-control study of endometrial cancer in relation to reproductive, somatometric, and life-style variables. Oncology 53:354–9. Kato I, Tominaga S, Ikari A. 1990. A case-control study of male colorectal cancer in Aichi Prefecture, Japan: With special reference to occupational activity level, drinking habits and family history. Jpn J Cancer Res 81: 115–21. Kotake K, Koyama Y, Nasu J, Fukutomi T, Yamaguchi N. 1995. Relation of family history of cancer and environmental factors to the risk of colorectal cancer: A case-control study. Jpn J Clin Oncol 25:195–202. Kriska AM, Caspersen CJ, eds. 1997. A collection of physical activity questionnaires for health related research. Med Sci Sports Exerc 29 (suppl). Kubik A, Zatloukal P, Boyle P, Robertson C, Gandini S, et al. 2001. A casecontrol study of lung cancer among Czech women. Lung Cancer 31: 111–22. Kubik AK, Zatloukal P, Tomasek L, Petruzelka L. 2002. Lung cancer risk among Czech women: A case-control study. Prev Med 34:436–44. Kune GA, Kune S, Watson LF. 1990. Body weight and physical activity as predictors of colorectal cancer risk. Nutr Cancer 13:9–17. La Vecchia C, Gallus S, Talamini R, Decarli A, Negri E, Franceschi S. 1999. Interaction between selected environmental factors and familial propensity for colon cancer. Eur J Cancer Prev 8:147–50. Lacey JV, Jr., Deng J, Dosemeci M, Gao YT, Mostofi FK, et al. 2001. Prostate cancer, benign prostatic hyperplasia and physical activity in Shanghai, China. Int J Epidemiol 30:341–9. Lampe JW, Slavin JL, Apple FS. 1991. Iron status of active women and the effect of running a marathon on bowel function and gastrointestinal blood loss. Int J Sports Med 12:173–9. Le Marchand L, Kolonel LN, Yoshizawa CN. 1991. Lifetime occupational physical activity and prostate cancer risk. Am J Epidemiol 133:103–11. Le Marchand L, Wilkens LR, Kolonel LN, Hankin JH, Lyu LC. 1997. Associations of sedentary lifestyle, obesity, smoking, alcohol use, and diabetes with the risk of colorectal cancer. Cancer Res 57:4787–94. Lee IM, Cook NR, Rexrode KM, Buring JE. 2001a. Lifetime physical activity and risk of breast cancer. Br J Cancer 85:962–5. Lee IM, Manson JE, Ajani U, Paffenbarger RS Jr., Hennekens CH, Buring JE. 1997. Physical activity and risk of colon cancer:the Physicians’ Health Study (United States). Cancer Causes Control 8:568–74. Lee IM, Paffenbarger RS Jr. 1994. Physical activity and its relation to cancer risk:a prospective study of college alumni. Med Sci Sports Exerc 26: 831–7. Lee IM, Paffenbarger RS Jr, Hsieh C. 1991. Physical activity and risk of developing colorectal cancer among college alumni. J Natl Cancer Inst 83: 1324–9. Lee IM, Paffenbarger RS Jr, Hsieh CC. 1992. Physical activity and risk of prostatic cancer among college alumni. Am J Epidemiol 135:169–79. Lee IM, Rexrode KM, Cook NR, Hennekens CH, Burin JE. 2001b. Physical activity and breast cancer risk: The Women’s Health Study (United States). Cancer Causes Control 12:137–45. Lee IM, Sesso HD, Oguma Y, Paffenbarger RS Jr. 2003. Physical activity, body weight, and pancreatic cancer mortality. Br J Cancer 88:679–83. Lee IM, Sesso HD, Paffenbarger RS Jr. 1999. Physical activity and risk of lung cancer. Int J Epidemiol 28:620–5. Lee IM, Sesso HD, Paffenbarger RS Jr. 2001c. A prospective cohort study of physical activity and body size in relation to prostate cancer risk (United States). Cancer Causes Control 12:187–93. Levi F, La Vecchia C, Negri E, Franceschi S. 1993. Selected physical activities and the risk of endometrial cancer. Br J Cancer. 67:846–51. Levi F, Pasche C, Lucchini F, La Vecchia C. 1999. Occupational and leisure time physical activity and the risk of breast cancer. Eur J Cancer 35: 775–8. Li BD, Khosravi MJ, Berkel HJ, Diamandi A, Dayton MA, et al. 2001. Free insulin-like growth factor-I and breast cancer risk. Int J Cancer 91:736–9. Lindblad P, Wolk A, Bergstrom R, Persson I, Adami HO. 1994. The role of obesity and weight fluctuations in the etiology of renal cell cancer: A population-based case-control study. Cancer Epidemiol Biomarkers Prev 3:631–9. Littman AJ, Voigt LF, Beresford SA, Weiss NS. 2001. Recreational physical activity and endometrial cancer risk. Am J Epidemiol 154:924–33. Liu S, Lee IM, Linson P, Ajani U, Buring JE, Hennekens CH. 2000. A prospective study of physical activity and risk of prostate cancer in US physicians. Int J Epidemiol 29:29–35. Longnecker MP, Gerhardsson de Verdier M, Frumkin H, Carpenter C. 1995. A case-control study of physical activity in relation to risk of cancer of the right colon and rectum in men. Int J Epidemiol 24:42–50. Loucks AB, Horvath SM. 1985. Athletic amenorrhea: A review. Med Sci Sports Exerc 17:56–72.

465

Loucks AB, Mortola JF, Girton L, Yen SS. 1989. Alterations in the hypothalamic-pituitary-ovarian and the hypothalamic-pituitary-adrenal axes in athletic women. J Clin Endocrinol Metab 68:402–11. Lund Nilsen TI, Johnsen R, Vatten LJ. 2000. Socio-economic and lifestyle factors associated with the risk of prostate cancer. Br J Cancer 82:1358–63. Luoto R, Latikka P, Pukkala E, Hakulinen T, Vihko V. 2000. The effect of physical activity on breast cancer risk: A cohort study of 30,548 women. Eur J Epidemiol 16:973–80. Lynge E, Thygesen L. 1988. Use of surveillance systems for occupational cancer: data from the Danish National system. Int J Epidemiol 17: 493–500. Lyons AS, Petrucelli RJ. 1978. Medicine: An Illustrated History. New York: Harry N. Abrams. MacConnie SE, Barkan A, Lampman RM, Schork MA, Beitins IZ. 1986. Decreased hypothalamic gonadotropin-releasing hormone secretion in male marathon runners. N Engl J Med 315:411–7. Macera CA, Ham SA, Jones DA, Kimsey CD, Ainsworth BE, Neff LJ. 2001. Limitations on the use of a single screening question to measure sedentary behavior. Am J Public Health 91:2010–2. Mantzoros CS, Tzonou A, Signorello LB, Stampfer M, Trichopoulos D, Adami HO. 1997. Insulin-like growth factor 1 in relation to prostate cancer and benign prostatic hyperplasia. Br J Cancer 76:1115–8. Marcus PM, Newcomb PA, Storer BE. 1994. Early adulthood physical activity and colon cancer risk among Wisconsin women. Cancer Epidemiol Biomarkers Prev 3:641–4. Marcus PM, Newman B, Moorman PG, Millikan RC, Baird DD, et al. 1999. Physical activity at age 12 and adult breast cancer risk (United States). Cancer Causes Control 10:293–302. Markowitz S, Morabia A, Garibaldi K, Wynder E. 1992. Effect of occupational and recreational activity on the risk of colorectal cancer among males: A case-control study. Int J Epidemiol 21:1057–62. Marti B, Minder CE. 1989. [Physical occupational activity and colonic carcinoma mortality in Swiss men 1979–1982]. Soz Praventivmed 34:30–7. Martinez ME, Giovannucci E, Spiegelman D, Hunter DJ, Willett WC, Colditz GA. 1997. Leisure-time physical activity, body size, and colon cancer in women. Nurses’ Health Study Research Group. J Natl Cancer Inst 89: 948–55. Matthews CE, Shu XO, Jin F, Dai Q, Hebert JR, et al. 2001. Lifetime physical activity and breast cancer risk in the Shanghai Breast Cancer Study. Br J Cancer 84:994–1001. McTiernan A. 2003. Intervention studies in exercise and cancer prevention. Med Sci Sports Exerc 35:1841–1845. McTiernan A, Stanford JL, Weiss NS, Daling JR, Voigt LF. 1996. Occurrence of breast cancer in relation to recreational exercise in women age 50–64 years. Epidemiology 7:598–604. Mellemgaard A, Engholm G, McLaughlin JK, Olsen JH. 1994. Risk factors for renal-cell carcinoma in Denmark. III. Role of weight, physical activity and reproductive factors. Int J Cancer 56:66–71. Mellemgaard A, Lindblad P, Schlehofer B, Bergstrom R, Mandel JS, et al. 1995. International renal-cell cancer study. III. Role of weight, height, physical activity, and use of amphetamines. Int J Cancer 60:350–4. Merzenich H, Boeing H, Wahrendorf J. 1993. Dietary fat and sports activity as determinants for age at menarche. Am J Epidemiol 138:217–24. Mezzetti M, La Vecchia C, Decarli A, Boyle P, Talamini R, Franceschi S. 1998. Population attributable risk for breast cancer:diet, nutrition, and physical exercise. J Natl Cancer Inst 90:389–94. Michaud DS, Giovannucci E, Willett WC, Colditz GA, Stampfer MJ, Fuchs CS. 2001. Physical activity, obesity, height, and the risk of pancreatic cancer. JAMA 286:921–9. Michels–Blanck H, Byers T, Mokdad AH, Will JC, Calle EE. 1996. Menstrual patterns and breast cancer mortality in a large U.S. cohort. Epidemiology 7:543–6. Mink PJ, Folsom AR, Sellers TA, Kushi LH. 1996. Physical activity, waist-tohip ratio, and other risk factors for ovarian cancer: A follow-up study of older women. Epidemiology 7:38–45. Mittendorf R, Longnecker MP, Newcomb PA, Dietz AT, Greenberg ER, et al. 1995. Strenuous physical activity in young adulthood and risk of breast cancer (United States). Cancer Causes Control 6:347–53. Moisan J, Meyer F, Gingras S. 1991. Leisure physical activity and age at menarche. Med Sci Sports Exerc 23:1170–5. Montoye HJ, Kemper HCG, Saris WHM, Washburn RA. 1996. Measuring Physical Activity and Energy Expenditure. Champaigne, IL: Human Kinetics. Moore DB, Folsom AR, Mink PJ, Hong CP, Anderson KE, Kushi LH. 2000. Physical activity and incidence of postmenopausal breast cancer. Epidemiology 11:292–6.

466

PART III: THE CAUSES OF CANCER

Moradi T, Adami HO, Bergstrom R, Gridley G, Wolk A, et al. 1999. Occupational physical activity and risk for breast cancer in a nationwide cohort study in Sweden. Cancer Causes Control 10:423–30. Moradi T, Adami HO, Ekbom A, Wedren S, Terry P, et al. 2002. Physical activity and risk for breast cancer: A prospective cohort study among Swedish twins. Int J Cancer 100:76–81. Moradi T, Nyren O, Bergstrom R, Gridley G, Linet M, et al. 1998. Risk for endometrial cancer in relation to occupational physical activity: A nationwide cohort study in Sweden. Int J Cancer 76:665–70. Moradi T, Nyren O, Zack M, Magnusson C, Persson I, Adami HO. 2000a. Breast cancer risk and lifetime leisure-time and occupational physical activity (Sweden). Cancer Causes Control 11:523–31. Moradi T, Weiderpass E, Signorello LB, Persson I, Nyren O, Adami HO. 2000b. Physical activity and postmenopausal endometrial cancer risk (Sweden). Cancer Causes Control 11:829–37. Morris JN, Chave SP, Adam C, Sirey C, Epstein L, Sheehan DJ. 1973. Vigorous exercise in leisure–time and the incidence of coronary heart-disease. Lancet 1:333–9. Morris JN, Clayton DG, Everitt MG, Semmence AM, Burgess EH. 1990. Exercise in leisure time:coronary attack and death rates. Br Heart J 63:325–34. Nieman DC. 1994. Exercise, upper respiratory tract infection, and the immune system. Med Sci Sports Exerc 26:128–39. NIH. 1996. Physical activity and cardiovascular health. NIH Consensus Development Panel on Physical Activity and Cardiovascular Health. JAMA 276:241–6. Norman A, Moradi T, Gridley G, Dosemeci M, Rydh B, et al. 2002. Occupational physical activity and risk for prostate cancer in a nationwide cohort study in Sweden. Br J Cancer 86:70–5. Oettle GJ. 1991. Effect of moderate exercise on bowel habit. Gut 32:941–4. Oliveria SA, Kohl HW 3rd, Trichopoulos D, Blair SN. 1996. The association between cardiorespiratory fitness and prostate cancer. Med Sci Sports Exerc 28:97–104. Olson SH, Vena JE, Dorn JP, Marshall JR, Zielezny M, et al. 1997. Exercise, occupational activity, and risk of endometrial cancer. Ann Epidemiol 7:46–53. Paffenbarger RS, Jr, Hyde RT, Wing AL. 1987. Physical activity and incidence of cancer in diverse populations:a preliminary report. Am J Clin Nutr 45:312–7. Paffenbarger RS Jr, Hyde RT, Wing AL, Lee IM, Jung DL, Kampert JB. 1993. The association of changes in physical-activity level and other lifestyle characteristics with mortality among men. N Engl J Med 328: 538–45. Paffenbarger RS Jr, Lee IM, Wing AL. 1992. The influence of physical activity on the incidence of site-specific cancers in college alumni. Adv Exp Med Biol 322:7–15. Paffenbarger RS Jr, Wing AL, Hyde RT. 1978. Physical activity as an index of heart attack risk in college alumni. Am J Epidemiol 108:161–75. Pate RR, Pratt M, Blair SN, Haskell WL, Macera CA, et al. 1995. Physical activity and public health. A recommendation from the Centers for Disease Control and Prevention and the American College of Sports Medicine. JAMA 273:402–7. Pedersen BK, Ullum H. 1994. NK cell response to physical activity: Possible mechanisms of action. Med Sci Sports Exerc 26:140–6. Persky V, Dyer AR, Leonas J, Stamler J, Berkson DM, et al. 1981. Heart rate: A risk factor for cancer? Am J Epidemiol 114:477–87. Peters RK, Garabrant DH, Yu MC, Mack TM. 1989. A case-control study of occupational and dietary factors in colorectal cancer in young men by subsite. Cancer Res 49:5459–68. Polednak AP. 1976. College athletics, body size, and cancer mortality. Cancer 38:382–7. Potter JD, Sellers TA, Folsom AR, McGovern PG. 1992. Alcohol, beer, and lung cancer in postmenopausal women. The Iowa Women’s Health Study. Ann Epidemiol 2:587–95. Pukkala E, Poskiparta M, Apter D, Vihko V. 1993. Life-long physical activity and cancer risk among Finnish female teachers. Eur J Cancer Prev 2:369–76. Putnam SD, Cerhan JR, Parker AS, Bianchi GD, Wallace RB, et al. 2000. Lifestyle and anthropometric risk factors for prostate cancer in a cohort of Iowa men. Ann Epidemiol 10:361–9. Res Q Exerc Sport. 2000. Measurement of physical activity: The Cooper Institute Conference Series. Res Q Exerc Sport 71(suppl):1–158. Rockhill B, Willett WC, Hunter DJ, Manson JE, Hankinson SE, Colditz GA. 1999. A prospective study of recreational physical activity and breast cancer risk. Arch Intern Med 159:2290–6. Rockhill B, Willett WC, Hunter DJ, Manson JE, Hankinson SE, et al. 1998. Physical activity and breast cancer risk in a cohort of young women. J Natl Cancer Inst 90:1155–60.

Russell JB, Mitchell D, Musey PI, Collins DC. 1984. The relationship of exercise to anovulatory cycles in female athletes:hormonal and physical characteristics. Obstet Gynecol 63:452–6. Salazar-Martinez E, Lazcano-Ponce EC, Lira-Lira GG, Escudero-De los Rios P, Salmeron-Castro J, et al. 2000. Case-control study of diabetes, obesity, physical activity and risk of endometrial cancer among Mexican women. Cancer Causes Control 11:707–11. Schneider J, Bradlow HL, Strain G, Levin J, Anderson K, Fishman J. 1983. Effects of obesity on estradiol metabolism:decreased formation of nonuterotropic metabolites. J Clin Endocrinol Metab 56:973–8. Sellers TA, Potter JD, Folsom AR. 1991. Association of incident lung cancer with family history of female reproductive cancers: The Iowa Women’s Health Study. Genet Epidemiol 8:199–208. Sesso HD, Paffenbarger RS Jr., Lee IM. 1998. Physical activity and breast cancer risk in the College Alumni Health Study (United States). Cancer Causes Control 9:433–9. Severson RK, Nomura AM, Grove JS, Stemmermann GN. 1989. A prospective analysis of physical activity and cancer. Am J Epidemiol 130:522–9. Shoff SM, Newcomb PA, Trentham-Dietz A, Remington PL, Mittendorf R, et al. 2000. Early-life physical activity and postmenopausal breast cancer:effect of body size and weight change. Cancer Epidemiol Biomarkers Prev 9:591–5. Shu XO, Hatch MC, Zheng W, Gao YT, Brinton LA. 1993. Physical activity and risk of endometrial cancer. Epidemiology 4:342–9. Siiteri PK. 1987. Adipose tissue as a source of hormones. Am J Clin Nutr 45:277–82. Sivertsen I, Dahlstrom AW. 1922. The relation of muscular activity to carcinoma: A preliminary report. J Cancer Res 6:365–78. Slattery ML, Abd-Elghany N, Kerber R, Schumacher MC. 1990. Physical activity and colon cancer: A comparison of various indicators of physical activity to evaluate the association. Epidemiology 1:481–5. Slattery ML, Edwards SL, Ma KN, Friedman GD, Potter JD. 1997a. Physical activity and colon cancer: A public health perspective. Ann Epidemiol 7:137–45. Slattery ML, Potter J, Caan B, Edwards S, Coates A, et al. 1997b. Energy balance and colon cancer—beyond physical activity. Cancer Res 57: 75–80. Slattery ML, Potter JD. 2002. Physical activity and colon cancer: Confounding or interaction? Med Sci Sports Exerc 34:913–9. Slattery ML, Schumacher MC, Smith KR, West DW, Abd-Elghany N. 1988. Physical activity, diet, and risk of colon cancer in Utah. Am J Epidemiol 128:989–99. Srivastava A, Kreiger N. 2000. Relation of physical activity to risk of testicular cancer. Am J Epidemiol 151:78–87. Steenland K, Nowlin S, Palu S. 1995. Cancer incidence in the National Health and Nutrition Survey I. Follow-up data:diabetes, cholesterol, pulse and physical activity. Cancer Epidemiol Biomarkers Prev 4:807–11. Steindorf K, Schmidt M, Kropp S, Chang–Claude J. 2003. Case-control study of physical activity and breast cancer risk among premenopausal women in Germany. Am J Epidemiol 157:121–30. Steindorf K, Tobiasz-Adamczyk B, Popiela T, Jedrychowski W, Penar A, et al. 2000. Combined risk assessment of physical activity and dietary habits on the development of colorectal cancer. A hospital-based case-control study in Poland. Eur J Cancer Prev 9:309–16. Struder HK, Hollmann W, Platen P, Rost R, Weicker H, et al. 1999. Neuroendocrine system and mental function in sedentary and endurance-trained elderly males. Int J Sports Med 20:159–66. Struder HK, Hollmann W, Platen P, Rost R, Weicker H, Weber K. 1998. Hypothalamic-pituitary-adrenal and -gonadal axis function after exercise in sedentary and endurance trained elderly males. Eur J Appl Physiol Occup Physiol 77:285–8. Stukonis M, Doll R. 1969. Gastric cancer in man and physical activity at work. Int J Cancer 4:248–54. Sturgeon SR, Brinton LA, Berman ML, Mortel R, Twiggs LB, et al. 1993. Past and present physical activity and endometrial cancer risk. Br J Cancer 68:584–9. Sung JF, Lin RS, Pu YS, Chen YC, Chang HC, Lai MK. 1999. Risk factors for prostate carcinoma in Taiwan: A case-control study in a Chinese population. Cancer 86:484–91. Taioli E, Barone J, Wynder EL. 1995. A case-control study on breast cancer and body mass. The American Health Foundation–Division of Epidemiology. Eur J Cancer 31A:723–8. Tang R, Wang JY, Lo SK, Hsieh LL. 1999. Physical activity, water intake and risk of colorectal cancer in Taiwan: A hospital-based case-control study. Int J Cancer 82:484–9. Tavani A, Braga C, La Vecchia C, Conti E, Filiberti R, et al. 1999. Physical activity and risk of cancers of the colon and rectum: An Italian casecontrol study. Br J Cancer 79:1912–6.

Physical Activity Tavani A, Gallus S, La Vecchia C, Dal Maso L, Negri E, et al. 2001. Physical activity and risk of ovarian cancer: An Italian case-control study. Int J Cancer 91:407–11. Terry P, Baron JA, Weiderpass E, Yuen J, Lichtenstein P, Nyren O. 1999. Lifestyle and endometrial cancer risk: A cohort study from the Swedish Twin Registry. Int J Cancer 82:38–42. Thun MJ, Calle EE, Namboodiri MM, Flanders WD, Coates RJ, et al. 1992. Risk factors for fatal colon cancer in a large prospective study. J Natl Cancer Inst 84:1491–500. Thune I, Brenn T, Lund E, Gaard M. 1997. Physical activity and the risk of breast cancer. N Engl J Med 336:1269–75. Thune I, Lund E. 1994. Physical activity and the risk of prostate and testicular cancer: A cohort study of 53,000 Norwegian men. Cancer Causes Control 5:549–56. Thune I, Lund E. 1996. Physical activity and risk of colorectal cancer in men and women. Br J Cancer 73:1134–40. Thune I, Lund E. 1997. The influence of physical activity on lung-cancer risk: A prospective study of 81,516 men and women. Int J Cancer 70: 57–62. Toniolo P, Bruning PF, Akhmedkhanov A, Bonfrer JM, Koenig KL, et al. 2000. Serum insulin-like growth factor-I and breast cancer. Int J Cancer 88: 828–32. Ueji M, Ueno E, Osei-Hyiaman D, Takahashi H, Kano K. 1998. Physical activity and the risk of breast cancer: A case–control study of Japanese women. J Epidemiol 8:116–22. United Kingdom Testicular Cancer Study Group. 1994. Aetiology of testicular cancer: Association with congenital abnormalities, age at puberty, infertility, and exercise. United Kingdom Testicular Cancer Study Group. BMJ 308:1393–9. US Department of Health and Human Services 1996. U.S. Department of Health and Human Services. Physical activity and health. A report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Disease Control and Prevention and Health Promotion. Vena JE, Graham S, Zielezny M, Brasure J, Swanson MK. 1987. Occupational exercise and risk of cancer. Am J Clin Nutr 45:318–27. Vena JE, Graham S, Zielezny M, Swanson MK, Barnes RE, Nolan J. 1985. Lifetime occupational exercise and colon cancer. Am J Epidemiol 122:357–65. Verloop J, Rookus MA, van der Kooy K, van Leeuwen FE. 2000. Physical activity and breast cancer risk in women aged 20–54 years. J Natl Cancer Inst 92:128–35. Vetter R, Dosemeci M, Blair A, Wacholder S, Unsal M, et al. 1992. Occupational physical activity and colon cancer risk in Turkey. Eur J Epidemiol 8:845–50. Vihko R, Apter D. 1984. Endocrine characteristics of adolescent menstrual cycles: Impact of early menarche. J Steroid Biochem 20:231–6. Vihko VJ, Apter DL, Pukkala EI, Oinonen MT, Hakulinen TR, Vihko RK. 1992. Risk of breast cancer among female teachers of physical education and languages. Acta Oncol 31:201–4. Villeneuve PJ, Johnson KC, Kreiger N, Mao Y. 1999. Risk factors for prostate cancer: Results from the Canadian National Enhanced Cancer Surveil-

467

lance System. The Canadian Cancer Registries Epidemiology Research Group. Cancer Causes Control 10:355–67. Vineis P, Ciccone G, Magnino A. 1993. Asbestos exposure, physical activity and colon cancer: A case-control study. Tumori 79:301–3. Vlajinac H, Jarebinski M, Adanja B. 1987. Relationship of some biosocial factors to colon cancer in Belgrade (Yugoslavia). Neoplasma 34:503–7. Wannamethee SG, Shaper AG, Walker M. 2001. Physical activity and risk of cancer in middle-aged men. Br J Cancer 85:1311–6. Warren MP. 1980. The effects of exercise on pubertal progression and reproductive function in girls. J Clin Endocrinol Metab 51:1150–7. Waterbor J, Cole P, Delzell E, Andjelkovich D. 1988. The mortality experience of major-league baseball players. N Engl J Med 318:1278–80. West DW, Slattery ML, Robison LM, French TK, Mahoney AW. 1991. Adult dietary intake and prostate cancer risk in Utah: A case-control study with special emphasis on aggressive tumors. Cancer Causes Control 2:85–94. Wheeler GD, Wall SR, Belcastro AN, Cumming DC. 1984. Reduced serum testosterone and prolactin levels in male distance runners. JAMA 252: 514–6. White E, Jacobs EJ, Daling JR. 1996. Physical activity in relation to colon cancer in middle-aged men and women. Am J Epidemiol 144:42–50. Whittemore AS, Kolonel LN, Wu AH, John EM, Gallagher RP, et al. 1995. Prostate cancer in relation to diet, physical activity, and body size in blacks, whites, and Asians in the United States and Canada. J Natl Cancer Inst 87:652–61. Whittemore AS, Paffenbarger RS Jr, Anderson K, Lee JE. 1984. Early precursors of urogenital cancers in former college men. J Urol 132: 1256–61. Whittemore AS, Paffenbarger RS Jr, Anderson K, Lee JE. 1985. Early precursors of site-specific cancers in college men and women. J Natl Cancer Inst 74:43–51. Whittemore AS, Wu-Williams AH, Lee M, Zheng S, Gallagher RP, et al. 1990. Diet, physical activity, and colorectal cancer among Chinese in North America and China. J Natl Cancer Inst 82:915–26. WHO. 2002. The World Health Report 2002. Reducing risks, promoting healthy life. Geneva, Switzerland: WHO. Will JC, Galuska DA, Vinicor F, Calle EE. 1998. Colorectal cancer: Another complication of diabetes mellitus? Am J Epidemiol 147:816–25. Woods JA, Davis JM. 1994. Exercise, monocyte/macrophage function, and cancer. Med Sci Sports Exerc 26:147–56. Wu AH, Paganini-Hill A, Ross RK, Henderson BE. 1987. Alcohol, physical activity and other risk factors for colorectal cancer: A prospective study. Br J Cancer 55:687–94. Wyrwich KW, Wolinsky FD. 2000. Physical activity, disability, and the risk of hospitalization for breast cancer among older women. J Gerontol A Biol Sci Med Sci 55:M418–21. Wyshak G, Frisch RE. 2000. Breast cancer among former college athletes compared to non-athletes: A 15-year follow-up. Br J Cancer 82:726–30. Yu H, Harris RE, Wynder EL. 1988. Case-control study of prostate cancer and socioeconomic factors. Prostate 13:317–25. Zheng W, Shu XO, McLaughlin JK, Chow WH, Gao YT, Blot WJ. 1993. Occupational physical activity and the incidence of cancer of the breast, corpus uteri, and ovary in Shanghai. Cancer 71:3620–4.

24

Exogenous Hormones JAMES V. LACEY, JR., GRAHAM A. COLDITZ, AND DAVID SCHOTTENFELD

T

he sex steroid hormones cause or contribute to the pathogenesis of reproductive organ cancers putatively through mitogenic and mutagenic effects (U.S. Department of Health and Human Services, 2002). By augmenting cell proliferation rates, steroid hormones stimulate clonal expansion of mutated stem cells and enhance tumorigenesis. In general, steroid hormones enter cells by passive diffusion and exhibit activity when binding as a ligand with high-affinity receptor proteins. The ligand-activated receptors then interact with specific DNA-response elements, and the DNA-bound receptors then positively or negatively influence target gene transcription. Sex steroid hormone metabolites may also act as genotoxins (Liehr, 2000) by covalently binding to guanine or adenine, thus depurinating and structurally altering DNA (Jefcoate et al., 2000).

ORAL CONTRACEPTIVES Oral contraceptives (OCs) were introduced in the 1960s. Early OC preparations contained both an estrogen (usually 100 mg–150 mg ethinyl estradiol) and a progestogen (usually 1–5 mg of progestogen). The hormone levels decreased over succeeding decades. Current OC formulations contain the equivalent of 30 mg of ethinyl estradiol and 150 mg of levonorgestrel. Ethinyl estradiol and 19-nortestosterone derivatives, such as levonorgestrel or norethisterone, are the most commonly used estrogen and progestogens, respectively. By inhibiting pituitary follicle-stimulating hormone and luteinizing hormone, combined OCs block the preovulatory luteinizing hormone surge, which prevents ovulation and, therefore, pregnancy. The progestogen component of combined OCs may also alter physiologic responses at the cervix to hinder fertilization and in the endometrium to obstruct blastocyst implantation (International Agency for Research on Cancer, 1999). Current OC formulations combine estrogen and progestogen for 20–22 days each cycle, followed by 7 days without hormones. Some formulations incorporate progestins at the same dose each day, whereas others use progestin doses that vary during the cycle (Petitti, 2003). These formulations replaced earlier-generation sequential oral contraceptive pills, which administered estrogen for the first 16 days of the cycle, followed by 5–7 days of estrogen and progestogen. (These sequential pills increased endometrial cancer risk and were removed from most markets worldwide in the mid-1970s.) Although many early epidemiologic studies of oral contraceptives included exposure to both sequential and combined OCs, this chapter focuses on the combined OCs. Large numbers of U.S. women first used OCs in the 1960s. As with any new medical intervention, cohorts of early users have highly correlated exposures: long-term use cannot be distinguished from first use at younger ages or last use at older ages. The combined associations with duration, latency, and recency will influence risk-versus-benefit decisions (e.g., to determine the minimum duration of use necessary to generate a 50% reduced (or increased) relative risk). At present, only the most recent epidemiologic studies have begun to assemble the data necessary to evaluate these issues.

POSTMENOPAUSAL HORMONE THERAPY For most women in developed countries, menopause occurs around age 50. Changes in menstrual cycle length and fluctuating endogenous

468

estradiol and progesterone production often prompt vasomotor, urogenital, sleep, and mood disturbances (Kenemans, 1999). Menopausal hormone therapy refers to exogenous estrogens, progestins, or combinations of estrogen and progestins taken by perimenopausal or postmenopausal women to relieve these symptoms (Nelson et al., 2002).

TERMINOLOGY AND NOMENCLATURE Although historically referred to as hormone replacement therapy or estrogen replacement therapy, exogenous hormones do not replace the endogenous steroid hormones that circulate at lower levels after menopause. Terms such as postmenopausal hormone therapy, unopposed estrogens, and estrogen plus progestin are now favored. Unopposed estrogen regimens are fairly uniform. Most U.S. estrogen use for the past 20 years has been continuous estrogen, where estrogen is taken each day, rather than cyclic estrogen, in which estrogen is taken for all but the last 5–7 days of the cycle. In contrast, estrogen plus progestin regimens vary widely. Cyclic or sequential regimens refer to daily estrogen plus progestin pills taken for 7, 7–10, or 10–14 days per cycle. Continuous estrogen plus progestin regimens combine daily estrogen and daily progestin. In this chapter, we applied a uniform descriptive nomenclature. Formulations are distinguished as unopposed estrogen or estrogen plus progestin. When formulation was not stated or could not be determined, or when it was not necessary to distinguish formulation, we used the term hormone therapy. Estrogen plus progestin formulations are further differentiated as sequential estrogen plus progestin regimens (e.g., progestin taken for less than 15 days per cycle) or continuous estrogen plus progestin regimens (e.g., progestin taken each day of the cycle).

TEMPORAL PATTERNS OF HORMONE THERAPY USE Patterns of hormone therapy use have dramatically changed in the past 40 years (Fig. 24–1). Unopposed estrogen use increased in the 1960s, fell in the late 1970s, climbed again in the 1980s, stabilized in the late 1990s, and declined in the early 2000s (Beral et al., 1999; Hersh et al., 2004). New information on potential risks caused both major downturns: increased endometrial cancer risks in the 1970s, and increased breast cancer, coronary heart disease, and stroke risks in the early 2000s. By 1999, when an estimated 24% of all U.S. women over age 40 took hormone therapy, unopposed estrogen was the most popular formulation: 17% of women over age 40 were current unopposed estrogen users, and 10%—approximately 6.5 million women—were current users who had been taking estrogens for at least 5 years (Brett and Reuben, 2003). Women with hysterectomy accounted for 90% of unopposed estrogen use. An estimated 62% of U.S. women over age 40 with hysterectomy had used estrogen at some point in their lives, and 45% were current unopposed estrogen users. The advent of estrogen plus progestin therapy, largely facilitated by oral medroxyprogesterone acetate, rejuvenated hormone therapy use in the early 1980s, and the total number of U.S. hormone therapy prescriptions increased until the early 2000s. Increasing estrogen plus progestin use among women without hysterectomy explained most of

Exogenous Hormones

469

No. of Dispensed Prescriptions, in Millions

60

50

Non-contraceptive estrogen

40

Oral and transdermal menopausal estrogens Oral estrogen

30 Non-contraceptive progestin

Oral medroxyprogesterone

20

Oral estrogen plus progestin

10

the late-1990s rise (Hersh et al., 2004). However, the population’s cumulative estrogen plus progestin use still remains relatively low: by 1999, 7% of all women over age 40 were current estrogen plus progestin users, and only 3%—roughly 2 million U.S. women—were current users with at least 5 years of total use (Brett and Reuben, 2003). The combination estrogen plus progestin pill accounted for 2% of all hormone therapy prescriptions in 1995 but 26% of 2001 prescriptions (Hersh et al., 2004). In 2002–2003, the number of hormone therapy prescriptions sharply declined, presumably in response to professional and public attitudes about risks and benefits. Between 2001 and 2003, combination estrogen plus progestin use fell by 56% to an estimated 2.5 million women (Hersh et al., 2004). As a result of the smaller pool of estrogen plus progestin users, population-based studies must confront the challenges associated with studying rare and dynamic exposures. This chapter on exogenous hormones will concentrate on estrogen and progesterone effects on tissues of the breast, endometrium, uterine cervix, ovary, and colon. Unless noted, all measures of association [odds ratios (ORs), relative risks (RRs), and hazard ratios (HRs)] for OCs include women who never used OCs as the referent group. Similarly, associations for hormone therapy reflect no menopausal hormone use as the referent group.

BREAST CANCER Oral Contraceptives Epidemiologic studies of OCs and breast cancer have employed diverse study designs, which vary in analytic complexity, sensitivity, and specificity (Hankinson et al., 1997; Malone et al., 1993; Stanford et al., 1989), to assess potential risks related to age at initiation, duration of exposure, cumulative dose, type or formulation of estrogen and progestogen, and cessation of use. The recent International Agency for Research on Cancer (IARC) summary reported that the “weight of the evidence suggests a small increase in the relative risk for breast cancer among current and recent [OC] users,” where “recent” described use within the previous 10 years (International Agency for Research on Cancer, 1999). The review included more than 50,000 women with breast cancer from more than 60 published studies of combined oral contraceptives and breast cancer. Only current and recent use increased risk, for breast cancer was not associated with OC duration, type, or dose. The IARC report’s conclusion expanded on the report made a few years earlier, when a combined reanalysis of data from 54 studies conducted in 25 countries, including 53,297 women with breast cancer and 100,239 women without breast cancer, concluded that current and recent (within the past 10 years) combined OC use slightly increased the relative risk (RR = 1.24, 95% CI, 1.15–1.33) (Collaborative Group

04

02

00

98

96

94

92

90

88

Year

20

20

20

19

19

19

19

19

84

82

80

78

76

74

72

70

68

86

19

19

19

19

19

19

19

19

19

19

19

19

66

0

Figure 24–1. Trends in U.S. menopausal hormone therapy use, 1966–2003. (Source: Adapted from Beral et al., 1999, and Hersh et al., 2004.)

on Hormonal Factors in Breast Cancer, 1996). That increased risk disappeared 10 years after cessation of use (RR = 1.01, 95% CI, 0.96–1.05). Recent use appeared to be more important than other exposure measures, such as duration, age at first use, dose, or formulation. Neither design issues nor other risk factors introduced statistical heterogeneity for the association with recent use. Despite these cohesive summaries, questions remain about potential associations limited to subgroups or surveillance biases in the literature (International Agency for Research on Cancer, 1999). Some large studies, such as the Cancer and Steroid Hormone (CASH) casecontrol study, showed no association with OC use (OR = 1.0; 95% CI, 0.9–1.1) (Centers for Disease Control, 1986), but others, such as the Nurses’ Health Study (Romieu et al., 1989), the Four State Study (Newcomb et al., 1996), and the U.S. WISH study (Brinton et al., 1995) reported significantly increased risks. Studies since have explored these and other issues. The Women’s Contraceptive and Reproductive Experiences (Women’s CARE) Study, a case-control study with 4575 case patients diagnosed between ages 35 and 64, generally provided reassuring evidence of no increased breast cancer risk among OC users (Davidson and Helzlsouer, 2002). Marchbanks et al. reported ORs of 1.0 (95% CI, 0.8–1.3) for current OC use and 0.9 (95% CI, 0.8–1.0) for previous use (Marchbanks et al., 2002). As in the IARC review and combined reanalysis, risk was not associated with duration of use, interval since last use, younger age at first use, or OC dose, and the results were similar across other risk factor strata. However, data on women younger than age 35 suggested potential increased risks. The large hospital-based case-control study of Rosenberg et al. reported increased risks for women between ages 25 and 34 years (N = 323 of 3540 cases) (Rosenberg et al., 1996). In the population-based case-control study of Althuis and Brinton, the 545 (of 3307) case patients under age 35 who had recent (i.e., within the previous 5 years) OC use were at increased risk (OR = 2.3, 95% CI, 1.4–3.6) (Althuis and Brinton, 2002). At the other end of the age spectrum, a Swedish study with 3016 cases patients and 3263 controls, ages 50–74 years, showed no associations with OC use (Van Hoften et al., 2000). The consideration of risk in women with a family history or mutations in the BRCA1 or BRCA2 genes has generated interest. Despite some evidence of increased risk due to OC use among these high-risk women, the data to date are limited by small sample size (Grabrick et al., 2000; Ursin et al., 1997) or potential biases related to use of prevalent case patients from select high-risk clinics (Narod et al., 2002). The sample size and level of detail provided by the 1996 combined reanalysis established a high standard of evidence for evaluating potential breast cancer risks among OC users. The studies published since, although rigorously designed, are unlikely to provide new interpretations of or conclusions about breast cancer risks among OC users.

470

PART III: THE CAUSES OF CANCER

Postmenopausal Hormone Therapy By the mid-1990s, the numerous studies of breast cancer among hormone therapy users provided inconsistent evidence of moderately increased risks. To better understand these potential associations, the Collaborative Group on Hormonal Factors in Breast Cancer combined and reanalyzed 51 epidemiologic studies, which represented approximately 90% of the worldwide literature on the topic. The main analysis included 17,949 postmenopausal cases and 35,916 postmenopausal controls with known and consistently defined age at menopause. Among current or recent (within 5 years) past users, the RRs increased with increasing duration of use (RR = 1.4, 95% CI, 1.2–1.5 for ≥5 years of use, and RR = 1.02, 95% CI, 1.01–1.04 per year of use). Neither ever-use nor increasing duration of use increased risk among former (last use at least 5 years ago) users (RR = 0.9, 95% CI, 0.7–1.1 for ≥5 years of use). The Group estimated that the cumulative excess risk of breast cancer after 10 years of hormone therapy use was 6 per 1000 women (95% CI, 3–9); after 15 years of use, it was 12 per 1000 women (95% CI, 5–20). Most of the included studies did not distinguish unopposed estrogen from estrogen plus progestin, but the individual study periods and the available data indicated that the majority of formulations were unopposed estrogen (Collaborative Group on Hormonal Factors in Breast Cancer, 1997). Based largely on this comprehensive analysis, the increased breast cancer risks among hormone therapy users appeared to be limited to long-term recent users, disappeared 5 years after cessation of use, and produced tumors with a favorable prognosis (Willett et al., 2000). Since that report, studies of every major design have expanded the literature, and the risk profile, for hormone therapy and breast cancer. Most data indicate risks are higher for estrogen plus progestin than for unopposed estrogen formulations. Particular progestins may especially increase risk. (Table 24–1). In the United States, Schairer et al. analyzed data from a cohort of 46,355 postmenopausal women followed from 1980 through 1995, in whom 2082 incident breast cancers developed (Schairer et al., 2000). Increasing durations of estrogen plus progestin use (≥4 years, RR = 1.9, 95% CI, 1.1–3.3) generated stronger associations than increasing durations of unopposed estrogen use (≥16 years, RR = 1.5, 95% CI, 1.1–2.2). Assuming a linear relationship between hormone use and risk, each year of unopposed estrogen use increased the RR by 0.04 (95% CI, 0.01–0.06) and each year of estrogen plus progestin use increased the RR by 0.13 (95% CI, 0.02–0.29). As in the collaborative analysis, risks were limited to current and recent users and declined with increasing time since last use. Porch et al. (Porch et al., 2002) explored data from the Women’s Health Study, a clinical trial of aspirin and vitamin E for cancer and heart disease prevention. The trial included 17,835 postmenopausal women, and 411 breast cancers occurred between 1993 and 2000. Five or more years of estrogen plus progestin (RR = 1.8, 95% CI, 1.3–2.4) but not unopposed estrogen (RR = 1.0, 95% CI, 0.7–1.5) use significantly increased risk. The study only captured hormone use at baseline, but the estrogen plus progestin association was limited to the continuous estrogen plus regimen (RR = 1.8, 95% CI, 1.3–2.5); no association appeared for sequential estrogen plus progestin (RR = 1.0, 95% CI, 0.7–1.5). Case-control studies generally included larger numbers of estrogen plus progestin users. The population-based case-control study by Ross et al. in Los Angeles county, with 1897 cases and 1637 controls, observed stronger associations with estrogen plus progestin (Ross et al., 2000). Unopposed estrogen use significantly increased risk only in long-term users (15 or more years), but the OR per 5 years of use was 1.06 (95% CI, 0.97–1.15). In women receiving estrogen plus progestin therapy, the OR per 5 years of use was 1.24 (95% CI, 1.07–1.45). This increase reflected higher risks for sequential regimens (OR = 1.38, 95% CI, 1.13–1.68) than continuous regimens (OR = 1.09, 95% CI, 0.88–1.35). A larger study from Newcomb et al. reported similar increases for both regimens (Newcomb et al., 2002). In 5298 cases and 5571 controls, the ORs for sequential and continuous use were 1.6 (95% CI, 1.0–2.6) and 1.5 (95% CI, 1.2–2.1), respectively. The peryear increases for unopposed estrogen use (OR = 1.02) and estrogen plus progestin use (OR = 1.04) were similar to the earlier studies. Both

of these studies included too few former estrogen plus progestin users to reliably assess risk after cessation. European studies report similar increased risks for estrogen plus progestin, but with important differences. Testosterone-related progestins account for the majority of European formulations, whereas progesterone-related progestins (such as medroxyprogesterone acetate) are far more common in the United States. In a Swedish study from Magnusson et al., which included 3345 cases and 3454 controls, the OR per year of unopposed estrogen use was 1.03 (95% CI, 0.98–1.08) and the OR per year of estrogen plus progestin use (89% of which was testosterone-derived progestins) was 1.07 (95% CI, 1.02–1.11) (Magnusson et al., 1999). Here, the per-year increase was higher for continuous regimens (OR = 1.19, 95% CI, 1.09–1.31) than sequential regimens (OR = 1.03, 95% CI, 0.94–1.13). Two cohort studies from Sweden published similar patterns of risk, despite low statistical power and potential residual confounding, especially by age at menopause (Jernstrom et al., 2003; Olsson et al., 2003). A cohort of more than 10,000 Danish nurses showed higher risks among users of continuous regimens (RR = 4.2, 95% CI, 2.6–6.8) than sequential regimens (RR = 1.9, 95% CI, 1.3–3.0) (Stahlberg et al., 2004). Women who used sequential regimens with progesterone-derived progestins also were at increased risk (RR = 3.0, 95% CI, 1.8–5.1), although these associations were based on relatively few cancers. The U.K. Million Women Study is the largest hormone therapy study to date (Beral, 2003). Between 1996 and 2001, the study recruited 1,084,110 women ages 50 to 64 who were attending the National Health Service Breast Screening Program. As with the Collaborative Group, the large sample size allowed the main analyses to include only postmenopausal women with a defined time since menopause. For each hormone therapy formulation, longer duration of use was associated with higher risks among the current users. The RRs for 5 or more years of current unopposed estrogen use and current estrogen plus progestin use were 1.3 (95% CI, 1.2–1.4), and 2.2 (95% CI, 2.1–2.4), respectively. The increased risks were similar regardless of estrogen type (e.g., conjugate equine estrogens vs. estradiol) or dose (e.g., £0.625 mg/day vs. >0.625 mg/day), progestin type (e.g., progesterone-derived vs. testosterone-derived), or formulations (e.g., oral vs. transdermal). The RR for conjugated equine estrogen plus medroxyprogesterone acetate for 5 or more years was 2.4 (95% CI, 2.1–2.8). With those associations, the authors estimated that hormone therapy accounted for an extra 20,000 breast cancers in U.K. women 50 to 64 years of age over the previous 10 years, with most due to estrogen plus progestin. Although randomized trials may not be designed solely to verify risks reported in observational studies, the Women’s Health Initiative includes two randomized clinical trials that were initiated in 1992 to evaluate whether hormone therapy reduces the risk of coronary heart disease and cardiovascular disease (Women’s Health Initiative Study Group, 1998). These trials offer a unique picture of the potential breast cancer risks associated with hormone therapy. One trial compared estrogen plus progestin (daily 0.625 mg/day conjugated equine estrogen plus 2.5 mg/day medroxyprogesterone acetate) to placebo in women with an intact uterus. The other compared unopposed estrogen (daily 0.625 mg/day conjugated equine estrogen) to placebo in women who had a prior hysterectomy. Unfavorable risk-benefit estimates forced the early termination of both trials. The estrogen plus progestin trial included 16,608 women, ages 50–79. After 5.6 years of observation, more breast cancers occurred in women taking hormones (N = 245) than placebo (N = 185); the HR was 1.24 (unweighted 95% CI, 1.02–1.50) (Chlebowski et al., 2003). The increased risk appeared 3–4 years into the trial and was even higher in women who used hormone therapy before the trial and in analyses that accounted for the substantial dropouts and drop-ins. The magnitude of the increased risk was nearly identical to the results from earlier observational studies, but this trial reported significantly more advanced-stage tumors among women taking estrogen plus progestin. Tumor grade or histology did not differ in the two groups. The unopposed estrogen trial included 10,379 women who had a hysterectomy (Anderson et al., 2004). It continued for a slightly longer period (average = 6.8 years) than the estrogen plus progestin trial

Table 24–1. Breast Cancer and Menopausal Hormone Therapy: Selected Studies Author

Year

Colditz

1995

Design and Location Cohort, USA

Study Details 1976–1992

Breast Cancers 1935 cases (930 no HT, 270 ET, 110 EPT)

Stanford

1995

Case-control, Washington, USA

Ages 50–64 1988–1990

537 cases

Persson

1997

Cohort, Sweden

Ages 40–70 1990–1995

435 cases

Persson

1999

Cohort, Sweden

198 cases

Magnusson

1999

Case-control Sweden

60,298 women, 1987–1993 Ages 50–74 1993–1995

Schairer

2000

Cohort, USA

46,355 women, 1979–1995

2082 cases

Ross

2000

Case-control, Los Angeles county, USA

Ages 55–72, 1987–1989

1897 cases

Kirsh

2002

Case-control, Ontario, Canada

Ages 20–74, 1995–1996

404 cases

Newcomb

2002

Li

2002

Case-control, Ages 50–79, 5298 cases Massachusetts, 1992–1994 New Hampshire, Wisconsin, USA Case-control, Ages 30–74, 149 Hispanic New Mexico, 1992–1994 cases USA 217 non-Hispanic white cases

Chen

2002

Case-control, Washington, USA

1990–1995

705 cases

WHI

2003

16,608 women, ages 50– 79, 1993– 2002

430 cases (285 among EPT and 185 among placebo)

Olsson

2003

Double-blind, placebocontrolled, randomized clinical trial, USA, EPT vs. placebo Cohort, Sweden

29,508 women, 1990–2001

556 cases

2563 cases

Control for Age at Menopause

Result

2-year intervals, plus menopause type

Increased RR with duration for current users ET, RR = 1.32 (1.14–1.61); EPT, = 1.41 (1.15–1.74) All analyses Current EPT, except duration RR = 0.9 (0.6–1.2) of HT Current ET, RR = 0.9 (0.7–1.3) None HT ≥ 10 yr, RR = 2.0 (1.0–4.0) EPT ≥ 11 yr, RR = 2.4 (0.7–8.6) 3 strata (<50, ET ≥ 6 yr, 50–54, ≥55 yr) RR = 1.1 (0.7–1.7) EPT RR = 1.7 (1.1–2.6) 5 strata (<45, 45– OR per yr of ET, 49, 50–51, 52–54, 1.03 (0.98–1.08) ≥55 yr), plus OR per yr of EPT, menopause type 1.07 (1.02–1.11)

Narrow intervals, with unknown age at menopause excluded from subanalyses Continuous age at menopause, with exclusion of women with hysterectomy without bilateral oophorectomy Continuous age at menopause

8 strata

None

Evaluated, but not included because not a confounder None

Yes, but not described

Recent (within previous 4 yr) use: ET, RR = 1.2 (1.0–1.4) EPT, RR = 1.4 (1.1–1.8) OR per 5 yr of ET, 1.06 (0.97–1.15) OR per 5 yr of EPT, 1.24 (1.07–1.45)

OR per yr of ET, 1.03 (0.97–1.09) OR per yr of EPT, 1.15 (1.01–1.33) ET ≥ 10 yr, OR = 1.74 (0.93–3.24) EPT ≥ 10 yr, OR = 3.48 (1.00–12.11) OR per yr of ET, 1.02 (1.01–1.03) OR per yr of EPT, 1.04 (1.01–1.08) Av. dur. ET = 10.1 yr Av. dur. EPT = 4.7 yr OR increased with duration of ET among both ethnic groups after control for P. Progestin use not related to risk. OR for current ET, 1.17 (0.85–1.60) OR for current EPT, 1.49 (1.04–2.12) EPT HR = 1.24 (1.02–1.50), P < 0.001 Significant trend with duration of use

Comment Higher RR in older women Significant increase in death from breast cancer No association with duration of ET or duration of EPT EPT ≥ 11 yr based on 4 cases RRs higher for recent use than distant past use Higher OR for testosteronederived progestin. Higher OR for continuous than sequential EPT regimens. Highest RRs for EPT and ductal breast cancer. Too few lobular to evaluate alone. OR higher per 5 yr of sequential EPT, (1.38), than per 5 yr of continuous EPT (1.09)

Results similar when restricted to invasive ductal or lobular cancer (all but 22 cases included)

RR diminished with time since last use for ET, but limited power for EPT latency Limited power. Median dur. of EPT: 18 mo. Hispanic women, 48 mo. nonHispanic. ET used for substantially longer durations. Higher ORs among leaner women. Limited data on continuous EPT regimens. Substantial noncompliance: 42% of women in EPT arm stopped therapy and 38% of women in placebo arm stopped therapy

Sequential EPT ≥ 4 yr, HR = 3.11 (0.99–9.83) Continuous EPT ≥4 yr, HR = 6.28 (3.17–12.41)

(continued)

471

472

PART III: THE CAUSES OF CANCER

Table 24–1. (cont.) Author

Year

Million Women Study collaborators

2003

Anderson (WHI)

2004

Design and Location

Study Details

Breast Cancers

Cohort, United Kingdom 1,084,110 women, 50– 64 years of age

1,084,110 women, ages 50– 64, 1996– 2001

9364 incident breast cancers 637 breast cancer deaths

Double-blind, placebocontrolled, randomized clinical trial, USA, ET vs. placebo

10,739 women, ages 50– 79, 1993– 1998 through 2003

218 cases (94 among EPT and 124 among placebo)

Control for Age at Menopause Assumed menopause at age 53 for women with hysterectomy; Assumed age at menopause comparable to others of same age. None

Result

Comment

Current ET use: RR = 1.30 (1.21–1.40)

RRs increased regardless of pattern of progestin use; with duration of use of hormones, mortality increased significant for current users, but not broken out for type of hormone used. Substantial noncompliance: more than 50% of women stopped therapy by end of study

ET ≥ 10 yr, RR = 1.37 (1.22–1.54) Current EPT use: RR = 2.00 (1.88–2.12) EPT ≥ 10 yr, RR = 2.31 (2.08–2.56) ET HR = 0.77 (0.59–1.01)

EPT, estrogen plus progestin therapy; ET, unopposed estrogen therapy; HR, hazard ratio; HT, hormone therapy; OR, odds ratio; RR, relative risk; WHI, Women’s Health Initiative.

before being stopped. Women assigned to unopposed estrogen developed fewer breast cancers (N = 94) than women assigned to placebo (N = 124), for a reduced risk (HR = 0.77, unweighted 95% CI, 0.59–1.01). These lower risks among unopposed estrogen users first appeared 2 years into the trial, and they remained through the duration of follow-up. Neither screening—all participants received annual mammograms—nor prior hormone therapy use appeared to account for this difference. One other randomized, placebo-controlled, clinical trial, the Heart and Estrogen/Progestin Replacement Study Follow-up (HERS II) compared the same estrogen plus progestin regimen to placebo (Hulley et al., 2002). In 2321 older, postmenopausal women with an intact uterus who completed an average of 4.1 years on trial medication plus 2.7 years on open-label follow-up, more breast cancers developed in women taking estrogen plus progestin (N = 49) than placebo (N = 39). The increased risk, however, was not statistically significant (HR = 1.3, 95% CI, 0.8–1.9). The cohort study from Schairer et al. (Schairer et al., 2000) noted an increased risk for tumors with ductal or lobular histology. Between 1987 and 1999, the U.S. national SEER program showed a 4% increase in breast cancer incidence (Li et al., 2003), which represented 1.5 additional cases of ductal carcinoma and 6.2 additional cases of lobular carcinoma per 100,000 women per year. The concurrent increase in the use of estrogen plus progestin raises questions about whether hormone therapy might account for the increased incidence of these tumor types. Evaluating risk for breast cancer subtypes faces several methodologic challenges: low statistical power, especially for rare subtypes; changes in diagnostic criteria or in pathologists’ application of those criteria; and variable analytic approaches can complicate interpretation of results. The Iowa Women’s Health Study reported that hormone therapy use increased risk for nonductal and nonlobular carcinomas only (N = 82) (Gapstur et al., 1999), whereas the Malmo Preventive Cohort reported stronger increased risks with hormone therapy for lobular tumors (Manjer et al., 2001). Neither study differentiated unopposed estrogen from estrogen plus progestin. The case-control data also show stronger increased risks for particular tumor types. The Newcomb et al. data reported similar per-year increased risks for both formulations for both ductal and lobular cancers (Newcomb et al., 2002). However, even in that large study, only 32 breast cancer patients had used estrogen plus progestin. Another report from the same source population used 219 lobular breast cancer cases, 2172 ductal breast cancer cases, and 3179 controls. For both unopposed estrogen and estrogen plus progestin, recent use generated higher relative risks for lobular cancers than ductal cancers (Newcomer et al., 2003). Daling et al. conducted a study of

841 ductal cancer cases, 145 lobular cancer cases, and 1210 controls (Daling et al., 2002). More than 5 years of unopposed estrogen use and more than 5 years of continuous estrogen plus progestin regimen use were each more strongly associated with lobular cancer than ductal cancer. Finally, in a detailed analysis of 975 postmenopausal breast cancer cases diagnosed in western Washington State, Li and colleagues evaluated type of hormone therapy and risk of specific histologic subtypes of breast cancer (Li et al., 2003). In 656 ductal breast cancer cases, 196 lobular cases, and 1007 controls, estrogen plus progestin use was associated with higher increased risks for lobular cancers than ductal cancers. These patterns held for both sequential and continuous estrogen plus progestin use. The combined epidemiologic and clinical trial data support a causal role for unopposed estrogen and estrogen plus progestin formulations in the occurrence of postmenopausal breast cancer. The increased risk per year of use is higher for estrogen plus progestin than for unopposed estrogen. Five years and 10 years of estrogen plus progestin use increase the cumulative breast cancer incidence by 5–7 per 1000 and 18–20 per 1000, respectively (Chlebowski et al., 2003; Beral, 2003). Estrogen plus progestin therapy likely increases the risk of lobular cancers and, to a lesser extent, ductal cancers. Hormone therapy would be more likely to account for the increasing incidence of breast cancer since the late 1980s if estrogen plus progestin were associated with both the rare lobular cancers and the much more common ductal cancers (Newcomb et al., 2002). The pattern of risks that are limited to current and recent users and that decline to pre-use level by 5 years after cessation of use is based largely on data from studies of unopposed estrogen. To date, studies of estrogen plus progestin have included almost exclusively current or recent users. Additional data are clearly required to directly test the hypothesis that risk associated with estrogen plus progestin also disappears quickly after cessation of use. The increased breast cancer risks, plus elevated risks of stroke, coronary heart disease, and pulmonary embolism, observed with menopausal hormone therapy (Anderson et al., 2004; Rossouw et al., 2002) challenged the perception that use of menopausal hormones ensured postmenopausal health. Although no longer considered appropriate for disease prevention (Fletcher and Colditz, 2002), exogenous hormones effectively reduce the risk of osteoporotic fractures (Nelson et al., 2002) and treat menopausal symptoms (Nelson, 2004). Other hormone-related treatments, such as selective estrogen receptor modulators (SERMs), have been proposed as potential alternatives to hormone therapy (Draper, 2003). The SERMs are structurally diverse steroid-like hormone molecules that differentially bind to estrogen receptors and modulate receptor

Exogenous Hormones expression in a tissue-specific manner. Steroidal estrogens are generally agonists in breast and endometrial tissues, but SERMs exert selective agonist or antagonist effects on various target tissues. Two SERMs have been approved for use in the United States. Tamoxifen citrate, a triphenylethylene that was developed more than 30 years ago, is currently used to prevent and treat breast cancer, and raloxifene hydrochloride, a benzothiophene, is currently used to prevent and treat osteoporosis (Jordan et al., 2001). Tamoxifen clearly prevents estrogen-receptor positive (but not estrogen receptor-negative) breast cancer (Cuzick et al., 2003). Data from randomized, double-blind, placebo-controlled clinical trials in the United Kingdom, the United States, Italy, and an international collaboration showed an approximate 40% decreased risk among women at high risk of breast cancer who received 20 mg of tamoxifen daily for up to 5 years (Cuzick et al., 2003), although the Italian study showed no decreased breast cancer risk among women not considered at high risk of breast cancer (Veronesi et al., 2003). The benefits for high-risk women must, however, be balanced against significantly increased risks of endometrial cancer, pulmonary embolism, deep vein thrombosis, cerebrovascular thrombosis, and cataracts (Freedman et al., 2003). Tamoxifen increases endometrial cancer risks by acting as an agonist in the uterus. Raloxifene, in contrast, acts as a uterine antagonist. In 7705 postmenopausal women in the Multiple Outcomes of Raloxifene Evaluation (MORE) trial, 60 mg of raloxifene once or twice daily reduced vertebral fracture incidence, the main trial outcome (Ettinger et al., 1999), and significantly reduced breast cancer incidence, a secondary trial outcome (RR = 0.24, 95% CI, 0.13–0.44; 13 breast cancers in raloxifene group vs. 27 breast cancers in placebo group) (Cummings et al., 1999). Unlike tamoxifen, raloxifene did not increase endometrial cancer risk. Raloxifene’s potential breast cancer chemopreventive effect makes it closer to the “ideal” SERM (Jordan et al., 2001). Current clinical trials, such as the Study of Tamoxifen and Raloxifene (STAR) and the Rationale and Overview of the Raloxifene Use for the Heart (RUTH) trials, are directly comparing and evaluating tamoxifen and raloxifene for breast cancer, endometrial cancer, and other chronic disease end points. However, as the current hormone therapy literature demonstrates, clinical and epidemiologic studies will need to definitely show both long-term efficacy and safety before SERMs’ role expands from chemoprevention to complement or replace the current role of menopausal hormone therapy in sustaining women’s health.

UTERINE CANCER Exquisitely sensitive to endogenous and exogenous hormones, the uterine corpus provides some of the most definitive data on the cancer risks and benefits associated with hormone use: decreased risks among oral contraceptive users and increased risks among unopposed estrogen users. Other aspects of exogenous hormone associations, such as risks after oral contraceptive cessation or among estrogen plus progestin users, are not fully understood.

Oral Contraceptives Combined OC use decreases the relative risk of endometrial cancer by approximately one-half and longer durations of use further reduce risk (International Agency for Research on Cancer, 1999). Cohort studies from the northeastern United States (Trapido, 1983), Royal College of General Practitioners (Beral et al., 1988), and United Kingdom (Vessey and Painter, 1995) described lower risks among users but were limited by the small number of women who developed endometrial cancer. Case-control data are therefore considered more robust. The CASH study included 433 cases and 3191 population-based controls and reported ORs near 1.0 for OC use for less than 12 months. The ORs for 12–23 months (OR = 0.7) and more than 24 months (OR = 0.4) of use were below 1.0. All categories of time since first use or last use were significantly below 1.0 (Centers for Disease Control, 1987a). Another multicenter U.S. study of 405 cases and 297 controls

473

showed even more dramatic declines with longer duration of use (OR = 0.2, 95% CI, 0.1–0.5 for 10 or more years of use) (Stanford et al., 1993). Other early European (La Vecchia et al., 1986) and World Health Organization (Rosenblatt and Thomas, 1991) data show similar associations, as do recent large studies from China (Xu et al., 2004) and Sweden (Weiderpass et al., 1999b). The Swedish study provided some of the most informative details about the association with OCs. In 709 cases and 3368 controls, OCs were associated with a 30% reduced risk (95% CI, 10%–50%) (Weiderpass et al., 1999b). Use for less than 3 years was not significantly associated with any reduction, but use for 10 or more years substantially reduced risk (OR = 0.2, 95% CI, 0.1–0.4). Each year of use reduced risk by 10% (95% CI, 6%–14%). There was some suggestion that OC users were at reduced risk for up to 20 years after their last use but that protection disappeared 30 years after last use. Two U.S. case-control studies also suggested that protection declines as former users enter their second decade after stopping. Voigt and colleagues combined data from two periods 10 years apart (Voigt et al., 1994) in a study of 316 cases and 501 controls. Very short-term use (1 year or less) was not associated with endometrial cancer, but a significantly reduced risk of endometrial cancer was associated with use of at least 5 years in total duration. Increasing duration of use produced inverse ORs even as time since last use increased from 10 years or less to more than 10 years. The study from Stanford and colleagues (Stanford et al., 1993) indicated that the reduced OR among OC users receded toward 1.0 as time since last use increased from <15 years to 20 or more years. Recent use appeared to be more important than duration of use. These consistent associations may extend to newer pill types. The data from the CASH study (Centers for Disease Control, 1987a), Voigt et al. (Voigt et al., 1994), and Swedish (Weiderpass et al., 1999b) studies showed similar ORs for high-, medium-, and low-dose progestin pills, but studies from Rosenblatt and Thomas (Rosenblatt and Thomas, 1991) and Hulka et al. (Hulka et al., 1982) found more pronounced protection among users of higher-dose progestins. The study from Weiderpass et al. (Weiderpass et al., 1999b) found that OCs reduced risk regardless of subsequent menopausal hormone therapy use or body mass index, but other studies reported diminished protection in heavier women or hormone therapy users (Stanford et al., 1993). The inverse association between OC use and endometrial cancer has been reported in almost all studies. At present, epidemiologic studies cannot confidently project that the duration of the protective effect will be sustained beyond 10 to 15 years after cessation of OC use (Table 24–2).

Menopausal Hormone Therapy The increased endometrial cancer risks among unopposed estrogen users provided the impetus for the estrogen plus progestin formulations. However, nearly 30 years after the initial risks among unopposed estrogen users surfaced, the magnitude of risk or extent of protection associated with estrogen plus progestin is ambiguous.

Unopposed Estrogens The increased risks among unopposed estrogen users described in the landmark 1975 case-control studies (Smith et al., 1975; Ziel and Finkle, 1976) have been replicated by almost all subsequent investigations. Herrinton and Weiss (Herrinton and Weiss, 1993) reviewed data from 19 studies, and Grady and colleagues (Grady et al., 1995) assessed more than 30 studies. Any estrogen use significantly increased risk in almost a linear fashion: the pooled RRs for 1–5, 5–10, and 10 or more years of use were 2.8 (95% CI, 2.3–3.5), 5.9 (95% CI, 4.7–7.5), and 9.5 (95% CI, 7.4–12.3), respectively. The elevated risks declined with increasing time since last use, from 4.1 (95% CI, 2.9–5.7) in women with less than 1 year since last use, to 2.3 (95% CI, 1.8–3.1) in women with last use at least 5 years ago (Grady et al., 1995). Both reviews reported increased risks regardless of dose or regimen. Three of the 17 case-control studies included by Herrinton and Weiss

474

PART III: THE CAUSES OF CANCER

Table 24–2. Endometrial Cancer and Combined Oral Contraceptives: Selected Case-Control Studies Cases Reference

China

Weiderpass (1999)

Sweden

Voigt (1994)

USA

Stanford (1993) USA

Jick (1993)

USA

Rosenblatt and Australia, Chile, China, Thomas (1991) Israel, Mexico, Phillipines, Thailand Levi et al. Switzerland (1991)

Controls

No. (Years Diagnosed)

COC Use (%)

833 (1997–2001) 709 (1994–1995)

17

833

25

22

3368

33

Household 40–59 surveys and random-digitdialing Random-digit20–74 dialing

316 (1975–1977 & 1985–1987) 405 (1987–1990)

22

501

34

20

297

36

Health plan members

50–64

142 (1979–1989)

18

1042

26

Hospital

25–59

220 (1979–1988)

17

1537

30

Hospital

£75

122 (1988–1990)

14

309

27

Study Location

Xu (2004)

Age Range (Years)

Controls Population registers Population registers

30–69 50–74

No.

COC Years Use (%) of Use OR £2 >2 <1 1–4 5–9 ≥10 <5 <5 ≥5 ≥5 <1 1–2 3–4 5–9 ≥10 1 2–5 >6

£2 >2 <2 2–5 >5

0.8 0.4 1.2 0.7 0.6 0.2

0.7 0.3 0.3 0.7 0.2 0.4 0.8 0.3 Dose Hi Lo 0.1 1.0 0.2 0.1 1.0 0.5 0.3

Years Since Last Use

OR

£10 >10 £10 >10 <10 10–14 15–19 ≥20

1.0 0.9 0.3 0.4 0.1 0.3 0.4 0.7

1–10 11–15 16–20 ≥21

0.4 0.4 0.5 0.6 Dose Hi Lo 0.1 0.4 0.2 0.1

£10 >10

COC, combined oral contraceptives; OR, odds ratio.

(Herrinton and Weiss, 1993) evaluated low-dose estrogen (0.3 mg) use but found risks similar to the common higher dose of 0.625 mg. Doses that are today considered high—1.25 mg—were associated with the highest pooled RRs in the Grady et al. report, but there was no difference in pooled RRs between today’s standard (0.625 mg; RR = 3.4, 95% CI, 2.0–5.6) and lower (0.3 mg; RR = 3.9, 95% CI, 1.6–9.5) doses, or between cyclic (RR = 3.0, 95% CI, 2.4–3.8) and continuous (RR = 2.9, 95% CI, 2.2–3.8) unopposed estrogen regimens. The summary RRs were higher for conjugated equine estrogens (RR = 2.5, 95% CI, 2.1–2.9) than synthetic estrogens (RR = 1.3, 95% CI, 1.1–1.6), and for cancers that had not invaded the myometrium (RR = 6.2, 95% CI, 4.5–8.4) than for cancers that had (RR = 3.8, 95% CI, 2.9–5.1). Subsequent studies indicate that risk among long-term users remained elevated for more than 5 years after cessation. Green and colleagues reported significantly increased risks among longer duration estrogen users for at least 8 years after last use (Green et al., 1996), and Newcomb and Trentham-Dietz (Newcomb and Trentham-Dietz, 2003) reported a statistically significant twofold increased risk for unopposed estrogen users whose last use occurred 10 or more years ago (Table 24–3).

Estrogen plus Progestins In the review by Grady and colleagues, three (Brinton and Hoover, 1993; Jick et al., 1993; Voigt et al., 1991) of the four (Bergkvist et al., 1989; Brinton and Hoover, 1993; Jick et al., 1993; Voigt et al., 1991) studies that assessed estrogen plus progestin formulations reported non-significantly elevated ORs. The single study (Voigt et al., 1991) that differentiated progestin regimens reported a higher OR for sequential regimens with progestin used fewer than 10 days per month (OR = 2.0, 95% CI, 0.7–5.3) than sequential regimens with progestin used for at least 10 days per month (OR = 0.9, 95% CI, 0.3–2.4). Additional data emerged in the past decade, yet the association with estrogen plus progestin regimen remains uncertain. Population-based studies generally included large samples of cases and controls, but

low prevalence of estrogen plus progestin use has impeded most investigations. The prevailing view that sequential, but not continuous, estrogen plus progestin increases risk draws support from two large casecontrol studies. In 1997, the data of Pike et al., from 833 cases and 791 controls, revealed a pattern of risks related to the number of days progestin was taken (Pike et al., 1997). For sequential estrogen plus progestin with fewer than 10 days of progestin each cycle, the increased risk was only slightly lower than risk among unopposed estrogen users. For any use of sequential estrogen plus progestin with more than 10 days of progestin per cycle (OR = 1.1, 95% CI, 0.8–1.4) and for continuous estrogen plus progestin (OR = 1.1, 95% CI, 0.8–1.4), risk was indistinguishable from nonusers. At least 70 cases reported use of these estrogen plus progestin regimens. There were slight but non-significant positive associations with increasing duration of continuous estrogen plus progestin. Two years later, Weiderpass et al. provided data from 709 cases and 3368 controls (Weiderpass et al., 1999a). Any sequential estrogen plus progestin use and 5 or more years of use generated statistically significant twofold increased risks. Continuous estrogen plus progestin use was associated with a 30% decreased risk (OR = 0.7, 95% CI, 0.7–1.0), and 5 or more years of use was more strongly associated with a decreased endometrial cancer risk (OR = 0.2, 95% CI, 0.1–0.8). However, this apparent protection was based on two exposed cases and limited to continuous regimens with testosterone-derived progestins, such as norethisterone or levonorgestrel, which are more potent than the progesterone-derived progestins, such as medroxyprogesterone acetate (Weiderpass et al., 1999a). Three case-control studies added further complexity to the literature. One Seattle study included 788 cases and 1122 controls, 9 and 33 of whom, respectively, had used continuous estrogen plus progestin. The OR was 0.6 (95% CI, 0.3–1.3) for ever-use, and there was no association with duration of use (Hill et al., 2000). An earlier publication from this group reported significantly increased risks with increasing duration of sequential estrogen plus progestin (progestin included for 10–21 days per cycle; OR = 2.5, 95% CI, 1.1–5.5) for 5

475

Exogenous Hormones Table 24–3. Endometrial Cancer and Menopausal Estrogen Therapy: Selected Case-Control Studies Cases Reference

Study Location

Controls

Age Range (Years)

No. (Years Diagnosed)

Controls ET Use (%)

No.

ET Use (%)

Newcomb and Trentham-Dietz (2003)

USA

Licensed driver lists and Medicare lists

40–79

591 (1991–1994)

19

2045

8

Jain (2000)

Canada

Residential lists Population registers

30–79

512 (1994–1998) 709 (1994–1995)

15

513

11

Weiderpass (1999a, 1999b)

Sweden

50–74

14 (b)

3368

USA

Random-digitdialing Random-digitdialing

45–74

Cushing (1998)

USA

Pike (1997)

USA

Neighborhood controls and Medicare lists

50–74

Green (1996)

USA

Random-digitdialing

45–64

45–64

730 (1985–1991) 484 (1985–1991)

5

11d

20 (c) Shapiro (1998)

c

57

1002

18

30

790

12

833 (1987–1993)

51

791

33

661

49

865

21

Years of Use ORa

Years Since Last Use

<5 ≥5

2.1 5.5

Current <10 ≥10

6.2b 1.5 2.3

Per yr <3 ≥3 <2 2–4 5–9 10–14 ≥15 Per yr <5 <5 ≥5 ≥5 <5 ≥5 Per yr <3 ≥3

1.1 1.3 4.1 1.7 2.1 3.3 8.4 12.6 1.2

<3 ≥3

3.4 1.7

<5 ≥5 <5 ≥5 <1 ≥1

2.0 7.5 1.8 6.3 2.4 1.2

£8 £8 >8 >8 £2 3–5 6–10 11–15 >15 Per 5 yr <4 4–8 9–12 >12

1.7 3.0 1.1 1.7e 5.6e

1.3 2.2 4.5 5.3 24.2 2.2

OR

Dosef £2 >2 £2 >2 <2 2–9 ≥10

£2 3–8 >8 £2 3–8 >8 £2 3–8 >8 £2 3–8 >8

0.3 0.625 1.25 2.3 2.3 7.8 1.9 1.5 1.2 9.2 11.2 16.3 g 2.1 3.0 2.5 1.9 1.6

2.5 1.7 1.7 5.1 2.1 1.5 8.6 5.7 2.0 19.8 12.4 5.4

ET, onopposed estrogen therapy; OR, odds ratio. a Blank OR indicates ORs for “Years since last use” reflect combinations of duration and recency. b Not stratified by duration. c Medium potency estrogens. d Low potency estrogens. e Crude OR calculated from published report. f mg/day. g Not estimatable.

or more years of use) (Beresford et al., 1997). A Canadian study of 512 cases and 513 controls reported non-significant positive associations with both sequential and continuous estrogen plus progestin, but low participant response and statistical power compel a cautious interpretation (Jain et al., 2000) The most recent case-control study reported strong but imprecise associations with both sequential and continuous regimens. Newcomb and Trentham-Dietz (Newcomb and Trentham-Dietz, 2003), in data from 591 cases and 2045 controls, found significant associations with sequential estrogen plus progestin when progestin was added for fewer than 10 days per cycle (OR = 2.4, 95% CI, 1.0–5.9) and with continuous estrogen plus progestin (OR = 2.3, 95% CI, 1.3–4.0), but not with sequential estrogen plus progestin when progestin was added for 10–21 days per cycle (OR = 1.1, 95% CI, 0.6–2.1). Current cohorts contribute minimal data on estrogen plus progestin. A Swedish record linkage reported slightly elevated risks associated with estrogen plus progestin (Persson et al., 1999) but had low statistical power and lacked data on other risk factors.

Additional analyses should clarify the potential risks or benefits associated with continuous estrogen plus progestin use. Reducing or reversing the risks associated with unopposed estrogen motivated prescribing the combined formulations to women with an intact uterus (Whitehead and Fraser, 1987). The prevailing views on how many days per cycle progestin must be given to abolish the increased risk associated with unopposed estrogen have evolved since the early 1980s. Original calls for 7 days of progestin use were replaced by the need for at least 10 days (Pike and Ross, 2000). The 10–14 days of progestin use in sequential estrogen plus progestin regimens are theoretically insufficient to counter the increased cellular proliferation stimulated by the unopposed estrogen component (Key and Pike, 1988), and the initial epidemiologic data from studies of cancer endpoints confirm that view. The continuous regimen’s daily progestin eliminates all unopposed estrogen exposure and therefore is predicted to not increase risk. But recent studies report similar associations for sequential and continuous regimens; these await replication. At present, only limited data on long duration (i.e., 5 or more years) of

476

PART III: THE CAUSES OF CANCER ment Study Follow-up (HERS II) reported two cancers in women taking hormones and eight in women taking placebo (HR = 0.3, 95% CI, 0.1–1.2) (Hulley et al., 2002). The Postmenopausal Estrogen/ Progestin Interventions (PEPI) Trial, the smallest of the three, reported no endometrial cancers over a 3-year follow-up (The Writing Group for the PEPI Trial, 1996). However, the extensive surveillance and progesterone-based treatment of potential endometrial cancer precursors in clinical trial settings (Holinka, 2001) may alter the natural history of endometrial lesions such that the external validity of trial data on invasive cancer end points is reduced. Nonetheless, continued observation of these study populations will provide valuable data on potential risks associated with continuous estrogen plus progestin (Table 24–4).

estrogen plus progestin use have been published. Additional followup of new and existing studies will further influence the interpretation of the current data. Three randomized clinical trials demonstrated that the continuous estrogen plus progestin regimen most widely used in the United States (0.625 mg/day conjugated equine estrogen plus 2.5 mg/day medroxyprogesterone acetate) does not increase risk relative to placebo. Two of the trials reported lower risks, but neither was statistically significant. The WHI estrogen plus progestin trial, the largest of the three, reported a hazard ratio of 0.8 (95% CI, 0.5–1.4) based on 27 cancers among women assigned to estrogen plus progestin and 31 cancers among women assigned to placebo (Anderson et al., 2003). The Heart and Estrogen/Progestin Replace-

Table 24–4. Endometrial Cancer and Menopausal Estrogen plus Progestin Therapy: Selected Case-Control Studies Cases Reference

Study Location

Newcomb and USA TrenthamDietz (2003)

Jain (2000)

Hill (2000)

Weiderpass (1999)

Pike (1997)

Canada

USA

Sweden

USA

Controls

Age Range (Years)

Licensed driver 40–79 lists and Medicare lists

Residential lists

Random-digitdialing

Population registers

Neighborhood controls and Medicare lists

30–79

45–74

50–74

50–74

Controls

No. EPT Use (Years Diagnosed) (%) 591 (1991–1994)

512 (1994–1998)

969 (1985–1991 and 1991–1994)

709 (1994–1995)

833 (1987–1993)

8

No. 2045

EPT Use EPT (%) Regimen* ORa

1 3 3

4 4

5 3

7 1 19

513

7 1 22

1325

10 12 20 2 17 8 9 16 1 3 12

3368

2 1 1 15

1 <1 <1 17 6 2 6 1 13 5 6

7 2 4 <1 9 6 2

6 5 <1

7 5 2

30 9 4 1 3

791

ORa

Any EPT: <5 ≥5 Per yr

1.4 2.3 1.1

Years Since Last Use OR

7

1 2 3

6 13 17 2 13 4 9 12 1 3 8

Years of Use

<10 10–21 >21

Any

2.4 1.1 2.3

Continuous

Continuous

<16b

Continuousc

1.1 <3 ≥3

0.7 1.7

0.6 1.5

<10

<3 ≥3

1.2 2.2

<3 ≥3

1.1 1.0

<5 ≥5 <5 ≥5

1.3 0.8 1.6 1.7

1.5

£3 >3

0.6 0.6

<5 <5 ≥5 ≥5 <5 ≥5 Per yr

1.5 2.9 1.1

<5 ≥5 Per yr

0.8 0.2 0.9

Any £2 3–4 ≥5 Per 5 yr

1.9 1.4 1.5 3.5 1.9

27 6 3 2 2

1.8 1.5

1.3 <3 ≥3

Sequential

Current Former

477

Exogenous Hormones Table 24–4. (cont.) Cases Reference

Beresford (1997)

Study Location

USA

Shapiro (1998) USA Jick (1993)

USA

Brinton (1993) USA

Controls

Random-digitdialing

Age Range (Years)

45–74

Random-digitdialing

45–74

Health plan members

50–64

Random-digitdialing

20–74

Controls

No. EPT Use (Years Diagnosed) (%)

394 (1985–1991)

730 (1985–1991) 172 (1979–1989) 300 (f) (1987–1990)

No.

EPT Use EPT (%) Regimen*

9 4 2 3

11 4 3 4

≥10

12 5 3 3

10 5 2 3

Continuous

14 8 3 1 4 7 3 1 3 9 14

4

788

13

1002

4 2 1 2 9 4 3 2 14

1720

207e

<10

10–21

≥10

ORa

Years of Use

ORa

Any £2 3–4 ≥5 Per 5 yr Any £2 3–4 ≥5 Per 5 yr

1.1 1.0 0.7 1.1 1.1 1.1 1.1 1.4 1.3 1.1

<3 3–4 ≥5

2.1 1.4 3.7

<3 3–4 ≥5

0.8 0.6 2.5

<3 ≥3

0.8d 1.2d

<3 ≥3 >0

2.2 1.3 1.8

Years Since Last Use OR

9

4

EPT, estrugen plus progestin therapy; OR, odds ratio. a Blank OR for EPT or years of use indicates ORs or “Years since last use” reflect combinations of regimen and duration or duration and recency. b Includes 6% of P used for 10–14 days per cycle. c Includes <1% of P used for at least 19 days per cycle. d Crude OR calculated from published report; e postmenopausal women only. *No. of days progestin.

CERVICAL CANCER Recent progress in understanding cervical cancer etiology brings a new focus to epidemiologic investigations of exogenous hormones and this tumor. Oncogenic types of human papillomavirus (HPV) are necessary but not sufficient causes of cervical cancer, and therefore we begin from the premise that OCs and hormone therapy may act as cofactors for cervical cancer, but they cannot be independent causes per se. The relevant research and public health question thus evolves from “Do exogenous hormones increase the risk of cervical cancer?” to “Do exogenous hormones influence the risk of progression from HPV infection to cervical cancer?” (Castellsague and Munoz, 2003). For further discussion of cervical cancer epidemiology and carcinogenesis, we refer the reader to Chapter 54.

Oral Contraceptives Early studies of OCs and cervical cancer did not account for HPV status. Therefore, increased cancer risks are almost certainly confounded by increased opportunity for HPV infection by sexual behaviors that are more common among women taking OCs than women not taking OCs (Skegg, 2002). Because cervical cancer screening may also be correlated with OC use, differences in prior screening may generate a detection bias among OC users. For those reasons, studies that use type-specific DNA-based HPV detection to measure HPV in both cases and controls, plus adjustment for other risk factors, are considered more reliable. Other recent publications (International Agency for Research on Cancer, 1999) provide extensive details about numerous studies conducted before the adoption of DNA-based HPV testing in epidemiologic studies. To address potential confounding by HPV, Green and colleagues (Green et al., 2003) reviewed 19 studies to evaluate the association

between OC use and HPV-positive status. Regardless of the OC definition (i.e., any use vs. none, current vs. past use, or short vs. long duration) or HPV-positive status classification (i.e., high-risk vs. lowrisk types, or detection with more accurate PCR-based techniques vs. detection via other methods), OC use was not associated with HPV status. Such reviews are constrained by the limitations of the source data, and not all of the studies included sufficient detail to fully address exposure misclassification or other potential confounders. Nonetheless, the authors concluded that residual confounding by HPV would not account for increased risks associated with OCs. Smith and colleagues (Smith et al., 2003) systematically reviewed 28 case-control and cohort studies published through 2002. Based on 12,531 women with cervical cancer, increasing duration of OC use increased cervical cancer risk, with an OR of 2.2 (95% CI, 1.9–2.4) for 10 or more years of use. Analyses restricted to DNA-based HPVpositive women produced a similar association (OR = 2.5, 95% CI, 1.6–3.9). Regardless of duration of use, current and recent users had higher risks than past users. Another recent review summarized data from seven studies that assessed OCs in the context of measured HPV status (Castellsague and Munoz, 2003). Results were mixed, but the authors concluded OCs might act as cofactors in cervical carcinogenesis. Risk estimates from studies with DNA-based HPV testing were lower than risk estimates from studies that did not measure HPV status or that used tests with lower sensitivity, such as serology. The IARC conducted the largest study to date that showed increased risks of carcinoma in situ and invasive cervical carcinoma in HPVpositive women with long-term OC use. In data combined from eight case-control studies conducted worldwide between 1985 and 1997, 5 or more years of OC use significantly increased the risk of invasive carcinoma (OR = 4.0, 95% CI, 2.0–8.0) and carcinoma in situ (OR = 3.4, 95% CI, 2.1–5.5) (Moreno et al., 2002). Shorter durations of use

478

PART III: THE CAUSES OF CANCER

increased risk only among recent users, which may reflect heightened surveillance among new OC users. Another large study from South Africa focused on invasive cancers—stage 1b or higher—to reduce potential confounding by screening and detection among OC users (Shapiro et al., 2003). In 524 cases and 254 HPV-positive controls, neither duration of use nor recent use was positively associated with cervical cancer. The majority of cervical cancers are squamous cell carcinomas. Some data on cervical adenocarcinomas, which account for approximately 15% of all cervical carcinomas (Wang et al., 2004), suggest that OCs particularly increase risk for this tumor type. Only recently have studies included relatively large numbers of cervical adenocarcinomas and used type-specific DNA-based HPV testing for both cases and controls. Two U.S. case-control studies found positive associations between adenocarcinoma in situ and both current use and longer duration of use (Lacey Jr. et al., 1999; Madeleine et al., 2001). Whether those results reflect true risks or detection bias is unclear. A possible pathogenic mechanism for an association with OC use is the effect of estrogen metabolites on HPV E6 and E7 oncogene expression (de Villiers, 2003). Current cohort studies of HPV-infected women should assist in determining the absolute and relative risks of OCs in conjunction with persistent HPV infection.

Menopausal Hormone Therapy Few studies have evaluated the potential cervical carcinoma risk among hormone therapy users. Earlier detection of cervical cancers (i.e., due to widespread screening via Papanicolaou tests) has generally decreased the mean age at diagnosis in the United States and reduced the likelihood that cervical carcinoma patients have the opportunity to accumulate substantial hormonal therapy exposure before diagnosis. However, the hypothesized role for OCs raises concern about exogenous hormone exposure via menopausal hormone therapy.

Unopposed Estrogens An Italian hospital-based case-control study (Parazzini et al., 1997) and a U.K. cohort study of mortality (Hunt et al., 1990) reported inverse associations with ever-use of unopposed estrogens, while a Scandanavian record linkage study reported no association with cervical cancer mortality (Schairer et al., 1997). All three preceded the widespread availability of HPV testing. A U.S. case-control study found a positive association between unopposed estrogen use and adenocarcinomas, but no association with squamous cell carcinomas (Lacey Jr. et al., 2000). This study included adjustment for HPV status, but, like the Italian study, was limited by low statistical power.

Estrogen plus Progestins Data on cervical cancer among estrogen plus progestin users are extremely sparse. Only six cases (three adenocarcinomas and three squamous cell carcinomas) in the study from Lacey and colleagues (Lacey Jr. et al., 2000) had used estrogen plus progestin, and thus the null associations reported therein are minimally informative. In the WHI estrogen plus progestin trial, hormone therapy had no effect on the incidence of cervical cancer or its precursor lesions (Anderson et al., 2003). Five women in the trial (two taking estrogen plus progestin, three taking placebo) developed cervical cancer. The data to date have raised some questions about potential associations, but data from more comprehensive studies are needed before hormone therapy is considered a cofactor for cervical carcinoma.

OVARIAN CANCER Whereas the breast, endometrium, and uterine cervix are target tissues for endogenous steroid hormones, the ovary is the primary source of those hormones during reproductive years. Somewhat surprisingly, then, epidemiologic associations between ovarian cancer and OC use (and, increasingly, menopausal hormone therapy use) resemble those observed with endometrial cancer.

Oral Contraceptives Epidemiologic studies consistently find decreased ovarian cancer risks among women who used OCs (International Agency for Research on Cancer, 1999). As with endometrial cancer, essentially all studies showed protection of approximately 40% to 50% relative to nonusers. The current controversies and gaps in the literature relate to temporal aspects of the effects of exposure. Reduced risks occur with increasing duration of use: each year of use reduces risk by approximately 7% to 10% (Kumle et al., 2004; Siskind et al., 2000). Different aspects of study design (e.g., cohort vs. case-control, population-based vs. hospital-based controls, North America vs. Europe vs. Asia and Oceania) or analysis (e.g., extent of adjustment for other risk factors) do not appear to modify the risk reductions. Two pooled analyses of case-control studies highlight the strength and consistency of these associations. In 12 U.S. studies with 2197 cases and 8893 controls, the OR for 6 or more years of use was 0.3 (95% CI, 0.2–0.4) (Whittemore et al., 1992). In six European studies (both hospital-based and population-based) with 2768 cases and 6274 controls, the OR for 5 or more years of use was 0.50 (95% CI, 0.33–0.76) (Bosetti et al., 2002). Cohort studies in Scandinavia (Kumle et al., 2004), the United Kingdom (Vessey and Painter, 1995), and the United States (Hankinson et al., 1995) reported similar associations, although based on fewer ovarian cancers. The current literature includes minimal data on risk 3–4 decades after last use. The large CASH study showed that protection persisted for at least 10 years (Centers for Disease Control, 1987b), but another large U.S. study revealed slight attenuation of protection more than 20 years after last use (Rosenberg et al., 1982). In data from Australia, where use reached higher levels sooner than in the United States or Europe, risks remained reduced for 20 years after last use (Siskind et al., 2000). A recent analysis of U.S. incidence and mortality rates suggested that the relative decline in incidence rates among users waned over time (Gnagy et al., 2000). Continued observation of existing cohort studies and new studies could provide additional data on potential temporal changes in the protection associated with OCs. Indication for OC use has not received extensive attention in studies to date. The available data suggest a slightly attenuated protection associated with noncontraceptive use, such as for treatment of gynecologic diseases (Ness et al., 2001), but additional data are needed to confirm that association. Studies are now beginning to provide estimates of ovarian cancer risk associated with lower-dose pills, whose use has been increasing since approximately 1980. Initial results show similar inverse associations. Some etiologic hypotheses predict that increased progesterone reduces risk (Risch, 1998), and thus the lower doses of progestin would be predicted to introduce less protection than the higher doses that were common in earlier pills. A large U.S. case-control study from Ness et al. found equivalent protection for high-dose and low-dose pills, and for pill use by time periods that roughly corresponded to the changes in pill potency (Ness et al., 2000). A reanalysis of the CASH study data concluded that high-potency progestin formulations were associated with a greater risk reduction than low-potency progestin formulations, but those data, collected between 1980 and 1982, did not contain enough use of low-dose estrogen and low-dose progestin formulations to allow definitive conclusions (Schildkraut et al., 2002). Ovarian cancer exhibits a stronger familial component than endometrial cancer, and therefore OCs have been considered for ovarian chemoprevention. Present data are inconsistent regarding whether OCs decrease risk in women at high risk due to mutations in the BRCA1 or BRCA2 genes. Two high-profile studies reached opposite conclusions. Using patients from high-risk families in a clinic-based setting, one case-control study reported equal protection among BRCA1 or BRCA2 mutation carriers and noncarriers (Narod et al., 1998). Using population-based cases and controls from Israel (where mutation prevalence is relatively high), another study reported protection only among women who did not carry deleterious mutations (Modan et al., 2001). Additional studies will be needed to evaluate the efficacy of OCs for prevention in high-risk populations (Tables 24–5 and 24–6).

479

Exogenous Hormones Table 24–5. Ovarian Cancer and Combined Oral Contraceptives: Selected Case-Control Studies Cases Age Range (Years)

Reference

Study Location

Ness (2001)

USA

Royar (2001)

Germany Residential lists

£75

Siskind (2000)

Australia Electoral registers

18–79

Sanderson (2000)

USA

Wittenberg (1999) USA

Controls

Random-digit- 20–69 dialing

Controls

No. COC Use (Years Diagnosed) (%)

No.

COC Use (%)

Years of Use £4 5–9 ≥10 1–2 3–5 6–10 11–15 16–20 ≥21 £1 2–5 6–10 10–15 >15 <5 <5 ≥5 ≥5

0.6 0.5 0.3 0.9 0.5 0.4 0.4 0.3 0.1 0.6 0.7 0.5 0.4 0.3

<5 ≥5

1.2c 0.4c

767 (1994–1998)

35*

1367

48

282 1993–1996

38

533

53

794 (1990–1993)

47

853

62

Random-digit- 20–79 dialing

276 1986–1988

45

388

64

Random-digit- 20–79 dialing

322 (1986–1988)

44

426

60

ORa

Years Since Last Use

OR

Current 1–5 6–10 11–20 ≥21

0.2 0.3 0.5 0.5 0.5

<1 1–4 5–9 10–19 ≥20 <10 ≥10 <10 ≥10

0.8b 1.5b 1.0b 1.4b Ref 0.8 1.0 0.7 0.4

£5 6–15 ≥15

0.6c 0.6c 1.2c

1.0d 0.6d

0.6d 0.6d 1.1d

COC, combined oral contraceptives; OR, odds ratio. *COCs for contraception. a Blank OR indicates ORs for “Years since last use” reflect combinations of duration and recency. b Among COC users only. c Mucinous cases only, N = 43. d Non-mucinous cases only, N = 279.

Table 24–6. Ovarian Cancer and Combined Oral Contraceptives: Selected Cohort Studies Reference Kumle (2004)

Cohort

Location

No.

Age at Enrollment (Years)

Observation Period

Norwegian-Swedish Women’s Lifestyle & Health Cohort

Norway & Sweden

106,841

30–49

1991–1992 through 2000

Schouten (2004)

Netherlands Cohort Netherlands Study Hankinson (1995) Nurses’ Health Study USA

2,412

55–69

1986–1995

107,865

30–55

1976–1988

Exposure Any COC use Duration of use <1 yr 1–4 yr 5–9 yr 10–14 yr ≥15 yr Time since last use 0–9 yr 10–14 yr 15–19 yr ≥20 yr Any OC use Any OC use Duration of use Current use Past, <1 yr Past, 1–2 yr Past, 3–4 yr Past, ≥5 yr Time since last use <5 yr 5–9 yr 10–14 yr ≥15 yr

No. No. Cancers Person-Years

RR

58

42,616

0.5

21 35 33 13 1

8,493 23,688 20,719 9,795 4,348

0.9 0.5 0.6 0.5 0.1

27 25 26 22 30

26,799 13,928 15,205 11,085 4,640

0.5 0.7 0.6 0.5 0.6*

94

592,056

1.1

4 26 23 10 19

22,486 116,757 133,287 79,785 156,137

1.9 1.2 1.1 0.8 0.7

12 19 27 19

99,891 156,377 145,766 71,489

0.9 0.8 1.0 1.1

COC, combined oral contraceptives; OC, oral contraceptives; RR, relative risk. *Crude RR calculated from published report.

Menopausal Hormone Therapy Ovarian cancer has not traditionally been included in the list of cancer sites for which hormone therapy increases risk, but recent publications indicate increased risks among women with long-term unopposed estrogen use (Noller, 2002). Data on estrogen plus progestin are

sparse, and drawing firm conclusions for this hormone therapy formulation is probably premature.

Unopposed Estrogens By the late 1990s, more than 1 dozen studies, each with at least 100 case patients, had been published. From sites across Europe, North

480

PART III: THE CAUSES OF CANCER

America, and Australia, with both hospital- and population-based controls, null associations appeared (Kaufman et al., 1989; La Vecchia et al., 1982; Purdie et al., 1995). A large pooled analysis found no association with hormone therapy use (La Vecchia et al., 1982), and a few investigations published non-significantly increased risks (Booth et al., 1989; Cramer et al., 1983; Risch, 1996). One positive (Parazzini et al., 1994) and one inverse (Hartge et al., 1988) association reached statistical significance. Two meta-analyses attempted to address the divergent study results. A 1998 report on 10 studies generated a weak but statistically significant positive association with ever-use: OR = 1.15, 95% CI, 1.05–1.27) (Garg et al., 1998). Individually excluding the studies that introduced significant heterogeneity yielded stronger associations (e.g., ORs of 1.40 and 1.41) for ever-use. An analysis of 15 studies reported a similar, although non-significant, summary association for all studies (OR = 1.1, 95% CI, 0.9–1.3) (Coughlin et al., 2000). After including only four homogeneous U.S. studies with population-based controls, the summary OR was 1.3 (95% CI, 1.0–1.6). Neither metaanalysis found any duration-response association. In a pooled analysis of four similar European case-control studies, ever-use of hormone therapy was also significantly associated with ovarian cancer (OR = 1.71, 95% CI, 1.30–2.25) (Negri et al., 1999). In 1999, Beral and colleagues proposed that failure to fully address confounding by hysterectomy and OC use—which both decrease subsequent ovarian cancer risk but increase the probability of hormone therapy use—might explain some of the inconsistent literature. Stratifying studies according to whether they adjusted for these factors, they reported a significantly increased pooled OR for those that adjusted and a null association for those that did not (Beral et al., 1999). However, of three (Purdie et al., 1999; Riman et al., 2002; Sit et al., 2002) large case-control studies published since that accounted for OCs, hysterectomy, and other risk factors, only one (Riman et al., 2002) reported positive associations with unopposed estrogen use. Almost 30 years ago, Hoover et al. noted elevated ovarian cancer incidence (eight observed vs. three expected) in a cohort of estrogen users (Hoover et al., 1978). Three cohort studies have recently published data consistent with that early observation. The American Cancer Society’s Cancer Prevention Study II has followed 240,073 postmenopausal women for cancer mortality since 1982. After 14 years of follow-up and 944 ovarian cancer deaths in the cohort, unopposed estrogen use significantly increased risk. Ten or more years of use significantly increased risk in current (RR = 2.2, 95% CI, 1.5–3.2) and former (RR = 1.6, 95% CI, 1.1–2.3) users (Rodriguez et al., 2001). A similar picture emerged from the Iowa Women’s Health Study, a study of 31,396 postmenopausal women followed for 14 years. Highest risks (RR = 2.5, 95% CI, 1.4–4.5) appeared among current

users with 5 or more years of use (Folsom et al., 2004). In a third cohort, the NCI’s Breast Cancer Detection Demonstration Project Follow-up Study of 44,241 postmenopausal women, any use of unopposed estrogen increased risk by 60% (RR = 1.6, 95% CI, 1.4–2.0) (Lacey Jr. et al., 2002). Unopposed estrogen use for 10–19 years or 20 or more years generated increased risks, especially among women with prior hysterectomy (RR = 2.0, 95% CI, 0.96–4.3 and RR = 3.4, 95% CI, 1.6–7.5, respectively). In each cohort, earlier analyses after short follow-up periods showed no or equivocal increased risks ( Mink et al., 1996; Rodriguez et al., 1995), which yielded to statistically significant increases after longer observations. However, none of the cohorts showed clear patterns of risk as time since last use increased.

Estrogen plus Progestins The study from Lacey Jr. et al. reported null associations with everuse and increasing duration of estrogen plus progestin use. With only 18 cancers among women who used estrogen plus progestin, this study was too small for reliable estimates (Lacey Jr. et al., 2002). A Swedish case-control study reported a significantly increased risk with ever-use of sequential (OR = 1.54, 95% CI, 1.15–2.05) but not continuous (OR = 1.02, 95% CI, 0.73–1.43) estrogen plus progestin (Riman et al., 2002). There was no clear pattern of increased risk with increasing duration, but the OR for 10 or more years of sequential estrogen plus progestin was 2.10 (95% CI, 0.99–4.48). Only two cases had used continuous estrogen plus progestin for 10 or more years (OR = 1.80, 95% CI, 0.37–8.82). Sit and colleagues reported no increased risk in association with estrogen plus progestin (Sit et al., 2002). The WHI estrogen plus progestin arm reported an elevated but non-significant hazard ratio among estrogen plus progestin users (HR = 1.58, 95% CI, 0.77–3.24, based on 32 total incident cancers) (Anderson et al., 2003). The small numbers of women with ovarian cancer who used estrogen plus progestin limit each of these studies. None to date has provided conclusive data on ovarian cancer risk among estrogen plus progestin users (Tables 24–7, 24–8, and 24–9).

COLORECTAL CANCER Oral Contraceptives Epidemiologic studies of OCs have shown consistent protection for ovarian and endometrial cancers and potential protection for colorectal cancer. A recent meta-analysis of 15 studies showed some heterogeneity between case-control studies but produced a significantly reduced summary estimate for any OC use and colorectal cancer: OR = 0.82, 95% CI, 0.74–0.92) (Fernandez et al., 2001). Not all of the studies separated tumors according to site (colon vs. rectum) or

Table 24–7. Ovarian Cancer and Menopausal Estrogen Therapy: Selected Case-Control Studies Cases Reference

Study Location

Sit (2002)

USA

Riman (2002)

Sweden

Purdie (1999)

Australia

Controls Random-digitdialing and Medicare lists Population registers

Random-digitdialing

Controls

Age Range (Years)

No. (Years Diagnosed)

ET Use (%)

No.

ET Use (%)

Years of Use

ORa

≥45

484 (1994–1998)

12

926

11

Any

0.9

50–74

655 (1993–1995)

9

3899

7

<1 1 2–4 5–9 ≥10 <5 <5 ≥5 ≥5 <1 1–3 >3

1.4 1.1 1.0 1.8 2.1

18–79

793 (1990–1993)

ET, unopposed estrogen therapy; OR, odds ratio. a Blank OR indicates ORs for “Years since last use” reflect combinations of duration and recency.

9

855

9

0.9 1.1 0.9

Years Since Last Use

OR

<3 ≥3 <3 ≥3 Current 1–5 >5

0.8 1.3 1.0 1.2 0.9 1.2 0.9

Table 24–8. Ovarian Cancer and Menopausal Estrogen plus Progestin Therapy: Selected Case-Control Studies Cases Reference

Study Location

Sit (2002)

USA

Riman (2002) Sweden

Age Range (Years)

Controls Random-digitdialing and Medicare lists Population registers

Controls

No. EPT Use (Years Diagnosed) (%)

No.

EPT Use (%)

≥45

484 (1994–1998)

8

926

7

50–74

655 (1993–1995)

17

3368

15

14 4 3 3 1 2

9 2 2 3 2 1

9 3 1 2 1 <1

7 2 1 2 2 <1

EPT Regimen* Any

Years OR of Use

ORa

Years Since Last Use ORa

1.1

<16b

Continuousc

Any <1 1 2–4 5–9 ≥10 <5 <5 ≥5 ≥5

1.5 1.9 1.5 1.3 1.1 2.1

1.0 1.3 0.8 0.9 0.9 1.8

<1 1 2–4 5–9 ≥10 <5 <5 ≥5 ≥5

<3 ≥3 <3 ≥3

1.3 0.7 1.3 0.3

<3 ≥3 <3 ≥3

1.2 2.9 2.8 d

EPT, estrogen plus progestin therapy; OR, odds ratio. a Blank OR indicates ORs for “Years since last use” reflect combinations of duration and recency. b Primarily P used for 10–14 days per cycle; c Primarily P used for 28 days per cycle; d Not estimable. *No. of days progestin.

Table 24–9. Ovarian Cancer and Menopausal Hormone Therapy: Selected Cohort Studies Reference

Cohort

Location

No.

Age at Enrollment (years)

Folsom (2004)

Iowa Women’s Health Study

USA

31,381

55–69

Lacey (2002)

Breast Cancer Detection Demonstration Project Follow-up Study

Rodriguez American (2001) Cancer Society Cancer Prevention Study II

USA

USA

44,241

211,581

36–89

30–55

Population Postmenopausal women

Postmenopausal women

Postmenopausal, hysterectomyfree women with known age at menopause

Observation Period 1986–2000

Exposure

1979–1998

ET use at baseline Former Current Duration of use at baseline Former, £5 yr Former, >5 yr Current, £5 yr Current, >5 yr Any ET use

1982–1996

Duration of ET use <4 yr 4–9 yr 10–19 yr ≥20 yr Time since last ET use Current or <2 yr 2–4 yr 5–9 yr 10–14 yr ≥15 yr EPT only use Duration of EPT only use <2 yr ≥2 yr ET use at baseline Duration of use <10 yr Last use within 15 yr Last use >15 yr ago ≥10 yr Last use within 15 yr Last use >15 yr ago

No. No. Cancers Person-Years RR 49 23

96,544 31,099

1.1 1.7

45 4 7 16 116

84,944 10,144 15,176 14,997 179,065

1.1 0.7 1.1 2.5 1.6

51 25 21 16

93,804 40,451 30,058 11,567

1.3 1.6 1.8 3.2

40 4 16 10 24 18

58,571 16,590 26,857 21,308 33,779 42,400

2.0 0.6 1.5 1.1 1.3 1.1

8 6 255*

12,809 19,521 625,984

1.6 0.8 1.2

45*

160,278

1.2

113*

256,545

1.1

19*

30,887

2.1

16*

26,394

1.3

EPT, estrogen plus progestin therapy; ET, unopposed estrogen therapy; RR, relative risk. *End point was fatal ovarian cancer.

481

482

PART III: THE CAUSES OF CANCER

Table 24–10. Colorectal Cancer and Combined Oral Contraceptives: Selected Case-Control Studies Cases

Controls

Study Location

Controls

Age Range (Years)

No. (Years Diagnosed)

COC Use (%)

No.

COC Use (%)

Years of Use

OR

Years Since Last Use

OR

Levi (2003)

Switzerland

Hospital-based

27–74

11

373

17

0.7 0.9

£15 >15

1.1 0.7

Italy

Hospital-based

<70

£5 >5

Fernandez (1998)

131 (1992–2001) 1232 (1985–1996)

7

2793

13

£2 >2

0.6 0.6

<10 ≥10

0.4 0.7

Reference

COC, combined oral contraceptives; OR, odds ratio.

provided details on OC use, but the ORs were similar for colon vs. rectal cancers and for short-duration (<5 years) versus longer duration (≥5 years) of use. The most pronounced reduction in risk appeared for use within the previous 10 years (OR = 0.46, 95% CI, 0.67–0.89). However, in a cohort of 267,400 women in China participating in a breast cancer screening trial, Rosenblatt and colleagues reported a significantly increased risk among women who had used OCs for at least 3 years (RR = 1.56, 95% CI, 1.01–2.40) (Rosenblatt et al., 2004). The formulation of the OC preparations used in China or the characteristics of the participants in the randomized trial may have contributed to the observed increased risks. Other cohort studies show potentially decreased risks. Using data from the Nurses’ Health Study, in which 501 colorectal cancers occurred over 12 years, Martinez and colleagues reported a significantly reduced risk among women who had used OCs for at least 8 years (RR = 0.60, 95% CI, 0.40–0.89) (Martinez et al., 1997). Other measures of exposure were non-significantly associated with colorectal cancer. The Breast Cancer Detection Demonstration Project Follow-up Study’s 1997 report of 330 colorectal cancers in a 10-year follow-up produced a RR of 1.0 (95% CI, 0.75–1.4), with no change in RR for duration of use (Troisi et al., 1997) (Tables 24–10 and 24–11).

Postmenopausal Hormone Therapy Unopposed Estrogens A 1998 meta-analysis from Hebert-Croteau summarized data from 19 studies published through 1996 (Hebert-Croteau, 1998). The casecontrol and cohort studies demonstrated substantial heterogeneity, but

attempts to control for that heterogeneity suggested that any use of hormone therapy—unopposed estrogen was not distinguished from estrogen plus progestin—was inversely associated with colon cancer (summary RR = 0.85, 95% CI, 0.73–0.99). Studies published after 1990 or that, in Hebert-Croteau’s view, provided better adjustment for confounding, produced slightly stronger protection. The summary risk reductions were stronger for current and long-term (at least 5 years) users. Two 1999 meta-analyses, one of 25 studies (Nanda et al., 1999) and one of 19 studies (Grodstein et al., 1999), reached similar conclusions: ever-use slightly reduced risk, and recent or longer use reduced risk more strongly than former or short-term use, respectively. The temporal aspects of the association deserve some attention. The Hebert-Croteau meta-analysis noted stronger inverse associations with hormone therapy in studies conducted after 1990. Two large cohort studies—the Nurses’ Health Study (Grodstein et al., 1998) and the Iowa Women’s Health Study (Folsom et al., 1995)—both reported initial null associations in early reports and stronger inverse associations in subsequent reports based on longer follow-up. This pattern resembles the ovarian cancer data (Lacey Jr et al., 2002) and reiterates the critical need for continued follow-up of hormone therapy users to understand long-term risks and benefits. Two recent studies added conflicting evidence about the potential association. In a case-control study based on pharmacy records in Washington state, Jacobs and colleagues (Jacobs et al., 1999) reported no association between recent unopposed estrogen use (within the past 10 years) and colon cancer. Neither increasing number of estrogen pills nor increasing cumulative estrogen dose was associated with colon cancer. A Canadian study also used prescription dispensing data to differentiate oral unopposed estrogens and transdermal estrogens

Table 24–11. Colorectal Cancer and Combined Oral Contraceptives: Selected Cohort Studies Reference

Cohort

Location

No.

Age at Enrollment (Years)

Observation Period

Rosenblatt (2004)

Shanghai Textile Industry Bureau employees

China

267,400

30–64

1989–1991

Troisi (1997)

Martinez (1997)

Breast Cancer Detection Demonstration Project Follow-up Study Nurses’ Health Study

OC, oral contraceptive; RR, relative risk.

USA

USA

57,259

89,448

31–90

34–59

1979–1989

1980–1992

End Point Colon cancer

Colorectal cancer

Colorectal cancer

No. Cancers

No. Person-Years

RR

Any OC use

92

352,851

1.1

Duration of use £6 mo 7–24 mo 25–36 mo ≥37 mo Any OC use

25 32 13 22 57

134,048 133,308 426,625 42,378 113,019

1.0 1.0 1.1 1.6 1.0

Duration of use <60 mo ≥60 mo Any OC use

38 19 166

75,124 37,895 498,648

1.0 1.1 0.8

Duration of use £11 mo 12–35 mo 36–95 mo ≥96 mo

40 43 45 23

112,653 126,984 153,644 81,940

0.9 1.0 0.8 0.6

Exposure

483

Exogenous Hormones Table 24–12. Colorectal Cancer and Hormone Therapy: Selected Case-Control Studies

Reference

Study Location

Csizmadi (2004)

Canada

Jacobs (1999)

Fernandez (1998)

USA

Italy

Controls Population-based registers

Health Plan Members

Hospital

Age Range (Years)

Cases

Controls

No. (Years Diagnosed)

No.

End Point

Exposure

4669

Colorectal cancer

Oral ET <3 yr ≥3 yr Transdermal ET <3 yrs ≥3 yr Recent ET usea No. ET tablets <750 ≥750 Total ET dose <375 mg ≥375 mg Any HT

≥50 yr

55–79

<75 yr

1197 (1992–1998)

441 (1984–1993)

1536 (1992–1996)

2180

3110

Colon cancer

Colon cancer

Rectal cancer

Duration £2 yr >2 yr Time since last use <10 yr ≥10 yr Any HT Duration £2 yr >2 yr Time since last use <10 yr ≥10 yr

OR 0.9 0.8 0.7 0.3 1.1 1.1 1.0 1.2 0.6b 0.6b 0.5b 0.7b 0.5b 0.5b 0.6b 0.4b 0.4b 0.5b

ET, unopposed estrogen therapy; HT, hormone therapy; OR, odds ratio. a Use within the past 10 years. b Postmenopausal women only, colon cancer, N = 822; rectal cancer, N = 453; controls N = 2188.

(Csizmadi et al., 2004). In 1197 cases and 4669 matched controls, ever-use of oral (OR = 0.82, 95% CI, 0.70–0.97) and transdermal (OR = 0.32, 95% CI, 0.13–1.79) estrogens were both significantly inversely associated with colorectal cancer. For both preparations, 3 or more years of use decreased risk to a greater degree than use for less than 3 years. In the WHI’s unopposed estrogen arm, roughly equal numbers of the 119 total colorectal cancers occurred in the estrogen (N = 61) and placebo (N = 58) arms, for a hazard ratio of 1.08 (95% CI, 0.75–1.55) (Anderson et al., 2004). Stratification by age showed non-significantly reduced hazard ratios in women ages 50–59 (0.59, 95% CI, 0.25–1.41) and 60–69 (0.88, 95% CI, 0.52–1.48) but a significantly increased hazard ratio in women ages 70–79 (2.09, 95% CI, 1.08–4.04).

Estrogen plus Progestin The case-control study from Jacobs and colleagues reported null associations for estrogen plus progestin use, similar to those for unopposed estrogen (Jacobs et al., 1999). In the HERS II study, colon cancer incidence was similar in women taking hormones and women taking placebo (HR = 0.81, 95% CI, 0.46–1.45, based on 47 total colon cancers). The WHI trial estrogen plus progestin trial reported 43 incident colorectal cancers in women taking hormones and 72 in women taking placebo, for a 44% decreased risk (HR = 0.56, 95% CI, 0.38–0.81) (Chlebowski et al., 2004). Separate analysis of the 19 rectal cancers produced a reduced but non-significant hazard ratio (0.66; 95% CI, 0.26–1.64). However, women taking estrogen plus progestin were significantly more likely than women taking placebo to be diagnosed with regional or metastatic disease. The observational studies provide a fairly consistent picture of decreased colorectal cancer risks among women taking menopausal hormones, and the study periods indicate most of that use would have been unopposed estrogen. However, the unopposed estrogen versus

placebo WHI trial showed no decreased risk, although the number of cancers was small. In contrast, the epidemiologic data on estrogen plus progestin use are limited, but the WHI trial showed significantly reduced risks among women taking continuous estrogen plus progestin regimens (Tables 24–12 and 24–13).

SUMMARY Oral Contraceptives At least three generations of women worldwide have had access to OCs for most of their reproductive lives, but epidemiologic research has not yet definitively characterized all potential cancer risks and benefits. Early users were primarily married women looking to space the timing of pregnancies. Recent users increasingly rely on OCs to delay or prevent a first pregnancy. It is not known whether the risk-benefit profiles are equivalent for these changing patterns of use. Understanding long-term cancer risks and benefits of OCs will therefore remain an important public health responsibility. Women who use OCs are less likely to develop ovarian cancer and endometrial cancer, but current data preclude extending that protection past 15 or 20 years. Balancing potential chemopreventive benefits with potential risks, especially for women at high risk of breast or ovarian cancers, urgently requires additional data.

Menopausal Hormone Therapy Women should no longer consider menopausal hormone therapy for prevention of chronic disease, but many women will continue to benefit from its relief of menopausal symptoms, which can last for years in some women. Recent declines in hormone therapy use could ultimately reduce the cancer burden in postmenopausal women, if

484

PART III: THE CAUSES OF CANCER

Table 24–13. Colorectal Cancer and Menopausal Hormone Therapy: Selected Cohort Studies

Reference

Cohort

Grodstein Nurses’ Health (1998) Study

Troisi (1997)

Breast Cancer Detection Demonstration Project Follow-up Study

Location

No.

Age at Enrollment (Years)

USA

59,002

34–59

USA

40,464

41–80

Observation Period

Population Postmenopausal women

1980–1994

Postmenopausal women

1979–1989

Exposure

End Point

Current HT use <5 yr ≥5 yr Former HT use <5 yr ≥5 yr Time since last HT use <5 yr ≥5 yr Any HT use

Colorectal cancer

No. No. PersonCancers Years RR

Colorectal cancer

Duration of HT use Recent use (within 1 yr) <5 yr ≥5 yr Former use (>1 yr ago) <5 yr ≥5 yr

30 59

75,299 0.6 90,903 0.7

77 34

99,458 0.8 36,029 0.9

31 74 155

51,365 0.7 77,065 0.9 152,323 1.0

12 28

20,910 0.8 37,842 0.8

71 44

59,404 1.1 34,166 1.0

HT = hormone theropy; RR, relative risk.

fewer breast cancers relatively occur in women who otherwise would have been exposed to hormones (Fig. 24–2). These changing patterns of use complicate but must not inhibit further study of the risks and benefits of hormone therapy. To under-

stand why the increased and decreased risks for breast and colorectal cancers, respectively, return to pre-use levels within 5 years of stopping could improve projections of absolute risk in hormone therapy users. To determine whether estrogen plus progestin increases risk for

Breast cancer EPT

0.03

Breast cancer EPT placebo

0.025

Breast cancer - ET

Cumulative hazard

Breast cancer - ET placebo

0.02

Colorectal cancer EPT Colorectal cancer EPT placebo

0.015

Colorectal cancer ET Colorectal cancer ET placebo

0.01

Ovarian cancer EPT Ovarian cancer EPT placebo

0.005

Endometrial cancer - EPT Endometrial cancer - EPT placebo

0 1

2

3

4

5

6

7

8

Time (years)

0.03 Breast cancer EPT

Figure 24–2. Cumulative cancer risks associated with estrogen plus progestin (EPT), unopposed estrogen (ET), and placebo in the Women’s Health Initiative. (a) Cumulative breast cancer risks (Chlebowski et al., 2003) (Anderson et al., 2004). (b) Cumulative endometrial cancer and ovarian cancer risks (Chlebowski et al., 2004) (Anderson et al., 2004). (c) Cumulative colorectal cancer risks (Anderson et al., 2003). ET placebo refers to women assigned to receive placebo in the unopposed estrogen trial. EPT placebo refers to women assigned to receive placebo in the estrogen plus progestin trial.

Cumulative hazard

0.025 Breast cancer EPT placebo

0.02

0.015 Breast cancer - ET

0.01

0.005 Breast cancer - ET placebo

0 1

(a)

2

3

5 4 Time (years)

6

7

8

485

Exogenous Hormones 0.03 Endometrial cancer - EPT

Cumulative hazard

0.025

0.02

Endometrial cancer - EPT placebo

0.015 Ovarian cancer EPT placebo

0.01

0.005 Ovarian cancer EPT

0 1

2

3

4 5 Time (years)

6

7

8

(b) 0.03 Colorectal cancer EPT

Cumulative hazard

0.025

0.02

Colorectal cancer EPT placebo

0.015 Colorectal cancer ET

0.01

0.005 Colorectal cancer ET placebo

0 1

2

3

4

5

6

7

8

Time (years)

(c)

endometrial, ovarian, and cervical cancers would help target screening efforts. To solidify risk estimates for women who use oral contraceptives and then hormone therapy would address a growing concern. Increasingly innovative epidemiologic investigations will facilitate these efforts. References Althuis MD, Brinton LA. 2002. Oral contraceptives and the risk of breast cancer. N Engl J Med 347:1448–1449. Anderson GL, Judd HL, Kaunitz AM, et al. 2003. Effects of estrogen plus progestin on gynecologic cancers and associated diagnostic procedures: The Women’s Health Initiative randomized trial. JAMA 290:1739–1748. Anderson GL, Limacher M, Assaf AR, et al. 2004. Effects of conjugated equine estrogen in postmenopausal women with hysterectomy: The Women’s Health Initiative randomized controlled trial. JAMA 291:1701–1712. Beral V. 2003. Breast cancer and hormone–replacement therapy in the Million Women Study. Lancet 362:419–427. Beral V, Banks E, Reeves G, Appleby P. 1999. Use of HRT and the subsequent risk of cancer. J Epidemiol Biostat 4:191–210. Beral V, Hannaford P, Kay C. 1988. Oral contraceptive use and malignancies of the genital tract. Results from the Royal College of General Practitioners’ Oral Contraception Study. Lancet 2:1331–1335. Beresford SA, Weiss NS, Voigt LF, McKnight B. 1997. Risk of endometrial cancer in relation to use of oestrogen combined with cyclic progestagen therapy in postmenopausal women. Lancet 349:458–461. Bergkvist L, Adami HO, Persson I, Hoover R, Schairer C. 1989. The risk of breast cancer after estrogen and estrogen-progestin replacement. N Engl J Med 321:293–297.

Figure 24–2. (cont.)

Booth M, Beral V, Smith P. 1989. Risk factors for ovarian cancer: A casecontrol study. Br J Cancer 60:592–598. Bosetti C, Negri E, Trichopoulos D, et al. 2002. Long-term effects of oral contraceptives on ovarian cancer risk. Int J Cancer 102:262–265. Brett KM, Reuben CA. 2003. Prevalence of estrogen or estrogen-progestin hormone therapy use. Obstet Gynecol 102:1240–1249. Brinton LA, Daling JR, Liff JM, et al. 1995. Oral contraceptives and breast cancer risk among younger women. J Natl Cancer Inst 87:827–835. Brinton LA, Hoover RN. 1993. Estrogen replacement therapy and endometrial cancer risk: Unresolved issues. The Endometrial Cancer Collaborative Group. Obstet Gynecol 81:265–271. Castellsague X, Munoz, N. 2003. Chapter 3: Cofactors in human papillomavirus carcinogenesis—role of parity, oral contraceptives, and tobacco smoking. J Natl Cancer Inst Monogr 20–28. Centers for Disease Control. 1986. Oral-contraceptive use and the risk of breast cancer. The Cancer and Steroid Hormone Study of the Centers for Disease Control and the National Institute of Child Health and Human Development. N Engl J Med 315:405–411. Centers for Disease Control. 1987a. Combination oral contraceptive use and the risk of endometrial cancer. The Cancer and Steroid Hormone Study of the Centers for Disease Control and the National Institute of Child Health and Human Development. JAMA 257:796–800. Centers for Disease Control. 1987b. The reduction in risk of ovarian cancer associated with oral–contraceptive use. The Cancer and Steroid Hormone Study of the Centers for Disease Control and the National Institute of Child Health and Human Development. N Engl J Med 316:650–655. Chlebowski RT, Hendrix SL, Langer RD, et al. 2003. Influence of estrogen plus progestin on breast cancer and mammography in healthy postmenopausal

486

PART III: THE CAUSES OF CANCER

women: The Women’s Health Initiative Randomized Trial. JAMA 289: 3243–3253. Chlebowski RT, Wactawski–Wende J, Ritenbaugh C, et al. 2004. Estrogen plus progestin and colorectal cancer in postmenopausal women. N Engl J Med 350:991–1004. Collaborative Group on Hormonal Factors in Breast Cancer. 1996. Breast cancer and hormonal contraceptives: Collaborative reanalysis of individual data on 53,297 women with breast cancer and 100,239 women without breast cancer from 54 epidemiological studies. Lancet 347: 1713–1727. Collaborative Group on Hormonal Factors in Breast Cancer. 1997. Breast cancer and hormone replacement therapy: Collaborative reanalysis of data from 51 epidemiological studies of 52,705 women with breast cancer and 108,411 women without breast cancer. Collaborative Group on Hormonal Factors in Breast Cancer. Lancet 350:1047–1059. Coughlin SS, Giustozzi A, Smith SJ, Lee NC. 2000. A meta–analysis of estrogen replacement therapy and risk of epithelial ovarian cancer. J Clin Epidemiol 53:367–375. Cramer DW, Hutchison GB, Welch WR, Scully RE, Ryan KJ. 1983. Determinants of ovarian cancer risk. I. Reproductive experiences and family history. J Natl Cancer Inst 71:711–716. Csizmadi I, Collet JP, Benedetti A, Boivin JF, Hanley JA. 2004. The effects of transdermal and oral oestrogen replacement therapy on colorectal cancer risk in postmenopausal women. Br J Cancer 90:76–81. Cummings SR, Eckert S, Krueger KA, et al. 1999. The effect of raloxifene on risk of breast cancer in postmenopausal women: Results from the MORE randomized trial. Multiple Outcomes of Raloxifene Evaluation. JAMA 281:2189–2197. Cuzick J, Powles T, Veronesi U, et al. 2003. Overview of the main outcomes in breast-cancer prevention trials. Lancet 361:296–300. Daling JR, Malone KE, Doody DR, et al. 2002. Relation of regimens of combined hormone replacement therapy to lobular, ductal, and other histologic types of breast carcinoma. Cancer 95:2455–2464. Davidson NE, Helzlsouer KJ. 2002. Good news about oral contraceptives. N Engl J Med 346:2078–2079. de Villiers EM. 2003. Relationship between steroid hormone contraceptives and HPV, cervical intraepithelial neoplasia and cervical carcinoma. Int J Cancer 103:705–708. Draper MW. 2003. The role of selective estrogen receptor modulators (SERMs) in postmenopausal health. Ann N Y Acad Sci 997:373–377. Ettinger B, Black DM, Mitlak BH, et al. 1999. Reduction of vertebral fracture risk in postmenopausal women with osteoporosis treated with raloxifene: Results from a 3–year randomized clinical trial. Multiple Outcomes of Raloxifene Evaluation (MORE) Investigators. JAMA 282:637–645. Fernandez E, La Vecchia C, Balducci A, Chatenoud L, Franceschi S, Negri E. 2001. Oral contraceptives and colorectal cancer risk: A meta-analysis. Br J Cancer 84:722–727. Fletcher SW, Colditz, GA. 2002. Failure of estrogen plus progestin therapy for prevention. JAMA 288:366–368. Folsom AR, Anderson JP, Ross JA. 2004. Estrogen replacement therapy and ovarian cancer. Epidemiology 15:100–104. Folsom AR, Mink PJ, Sellers TA, Hong CP, Zheng W, Potter JD. 1995. Hormonal replacement therapy and morbidity and mortality in a prospective study of postmenopausal women. Am J Public Health 85:1128–1132. Freedman AN, Graubard BI, Rao SR, McCaskill–Stevens W, Ballard–Barbash R, Gail MH. 2003. Estimates of the number of US women who could benefit from tamoxifen for breast cancer chemoprevention. J Natl Cancer Inst 95:526–532. Gapstur SM, Morrow M, Sellers TA. 1999. Hormone replacement therapy and risk of breast cancer with a favorable histology: Results of the Iowa Women’s Health Study. JAMA 281:2091–2097. Garg PP, Kerlikowske K, Subak L, Grady D. 1998. Hormone replacement therapy and the risk of epithelial ovarian carcinoma: A meta-analysis [see comments]. Obstet Gynecol 92:472–479. Gnagy S, Ming EE, Devesa SS, Hartge P, Whittenmore AS. 2000. Declining ovarian cancer rates in U.S. women in relation to parity and oral contraceptive use. Epidemiology 11:102–105. Grabrick DM, Hartmann LC, Cerhan JR, et al. 2000. Risk of breast cancer with oral contraceptive use in women with a family history of breast cancer. JAMA 284:1791–1798. Grady D, Gebretsadik T, Kerlikowske K, Ernster V, Petitti D. 1995. Hormone replacement therapy and endometrial cancer risk: A meta–analysis. Obstet Gynecol 85:304–313. Green J, Berrington dG, Smith JS, et al. 2003. Human papillomavirus infection and use of oral contraceptives. Br J Cancer 88:1713–1720. Green PK, Weiss NS, McKnight B, Voigt LF, Beresford SA. 1996. Risk of endometrial cancer following cessation of menopausal hormone use. Washington, United States. Cancer Causes Control 7:575–580.

Grodstein F, Martinez ME, Platz EA, et al. 1998. Postmenopausal hormone use and risk for colorectal cancer and adenoma. Ann Intern Med 128:705–712. Grodstein F, Newcomb PA, Stampfer MJ. 1999. Postmenopausal hormone therapy and the risk of colorectal cancer: A review and meta-analysis. Am J Med 106:574–582. Hankinson SE, Colditz GA, Hunter DJ, et al. 1995. A prospective study of reproductive factors and risk of epithelial ovarian cancer. Cancer 76: 284–290. Hankinson SE, Colditz GA, Manson JE, et al. 1997. A prospective study of oral contraceptive use and risk of breast cancer. Nurses’ Health Study, United States. Cancer Causes Control 8:65–72. Hartge P, Hoover R, McGowan L, Lesher L, Norris HJ. 1988. Menopause and ovarian cancer. Am J Epidemiol 127:990–998. Hebert-Croteau N. 1998. A meta-analysis of hormone replacement therapy and colon cancer in women. Cancer Epidemiol Biomarkers Prev 7:653–659. Herrinton LJ, Weiss NS. 1993. Postmenopausal unopposed estrogens. Characteristics of use in relation to the risk of endometrial carcinoma. Ann Epidemiol 3:308–318. Hersh AL, Stefanick ML, Stafford RS. 2004. National use of postmenopausal hormone therapy: Annual trends and response to recent evidence. JAMA 291:47–53. Hill DA, Weiss NS, Beresford SA, et al. 2000. Continuous combined hormone replacement therapy and risk of endometrial cancer. Am J Obstet Gynecol 183:1456–1461. Holinka CF. 2001. Design and conduct of clinical trials in hormone replacement therapy. Ann N Y Acad Sci 943:89–108. Hoover R, Fraumeni JF Jr, Gray LA Sr. 1978. Exogenous oestrogens and ovarian cancer. Lancet 1:325–326. Hulka BS, Chambless LE, Kaufman DG, Fowler WC Jr., Greenberg BG. 1982. Protection against endometrial carcinoma by combination-product oral contraceptives. JAMA 247:475–477. Hulley S, Furberg C, Barrett–Connor E, et al. 2002. Noncardiovascular disease outcomes during 6.8 years of hormone therapy: Heart and Estrogen/ progestin Replacement Study follow–up. HERS II. JAMA 288:58–66. Hunt K, Vessey M, McPherson K. 1990. Mortality in a cohort of long–term users of hormone replacement therapy: An updated analysis. Br J Obstet Gynaecol 97:1080–1086. International Agency for Research on Cancer. 1999. International Agency for Research on Cancer. IARC) Monographs on the Evaluation of Carcinogenic Risk to Humans. Hormonal contraception and postmenopausal hormonal therapy. World Health Organization, vol 72. Lyon: IARC, pp. 1–338, 399–564. Jacobs EJ, White E, Weiss NS, Heckbert SR, LaCroix A, Barlow WE. 1999. Hormone replacement therapy and colon cancer among members of a health maintenance organization. Epidemiology 10:445–451. Jain MG, Rohan TE, Howe GR. 2000. Hormone replacement therapy and endometrial cancer in Ontario, Canada. J Clin Epidemiol 53:385–391. Jefcoate CR, Liehr JG, Santen RJ, et al. 2000. Tissue–specific synthesis and oxidative metabolism of estrogens. J Natl Cancer Inst Monogr 95–112. Jernstrom H, Bendahl PO, Lidfeldt J, Nerbrand C, Agardh CD, Samsioe G. 2003. A prospective study of different types of hormone replacement therapy use and the risk of subsequent breast cancer: The women’s health in the Lund area (WHILA) study. Sweden. Cancer Causes Control 14:673–680. Jick SS, Walker AM, Jick H. 1993. Estrogens, progesterone, and endometrial cancer. Epidemiology 4:20–24. Jordan VC, Gapstur S, Morrow M. 2001. Selective estrogen receptor modulation and reduction in risk of breast cancer, osteoporosis, and coronary heart disease. J Natl Cancer Inst 93:1449–1457. Kaufman DW, Kelly JP, Welch WR, et al. 1989. Noncontraceptive estrogen use and epithelial ovarian cancer. Am J Epidemiol 130:1142–1151. Kenemans P. 1999. Menopause, HRT and menopausal symptoms. J Epidemiol Biostat 4:141–146. Key TJ, Pike, MC. 1988. The dose–effect relationship between ‘unopposed’ oestrogens and endometrial mitotic rate: its central role in explaining and predicting endometrial cancer risk. Br J Cancer 57:205–212. Kumle M, Weiderpass E, Braaten T, Adami HO, Lund E. 2004. Risk for invasive and borderline epithelial ovarian neoplasias following use of hormonal contraceptives: The Norwegian–Swedish Women’s Lifestyle and Health Cohort Study. Br J Cancer 90:1386–1391. La Vecchia C, Decarli A, Fasoli M, et al. 1986. Oral contraceptives and cancers of the breast and of the female genital tract. Interim results from a casecontrol study. Br J Cancer 54:311–317. La Vecchia C, Liberati A, Franceschi S. 1982. Noncontraceptive estrogen use and the occurrence of ovarian cancer. J Natl Cancer Inst 69:1207. Lacey JV Jr., Brinton LA, Abbas FM, et al. 1999. Oral contraceptives as risk factors for cervical adenocarcinomas and squamous cell carcinomas. Cancer Epidemiol Biomarkers Prev 8:1079–1085.

Exogenous Hormones Lacey JV Jr., Brinton LA, Barnes WA, et al. 2000. Use of hormone replacement therapy and adenocarcinomas and squamous cell carcinomas of the uterine cervix. Gynecol Oncol 77:149–154. Lacey JV Jr, Mink PJ, Lubin JH, et al. 2002. Menopausal hormone replacement therapy and risk of ovarian cancer. JAMA 288:334–341. Li CI, Malone KE, Porter PL, et al. 2003. Relationship between long durations and different regimens of hormone therapy and risk of breast cancer. JAMA 289:3254–3263. Li R, Gilliland FD, Baumgartner K, Samet J. 2002. Hormone replacement therapy and breast carcinoma risk in Hispanic and non-Hispanic women. Cancer 95:960–968. Liehr JG. 2000. Is estradiol a genotoxic mutagenic carcinogen? Endocr Rev 21:40–54. Madeleine MM, Daling JR, Schwartz SM, et al. 2001. Human papillomavirus and long–term oral contraceptive use increase the risk of adenocarcinoma in situ of the cervix. Cancer Epidemiol Biomarkers Prev 10:171–177. Magnusson C, Baron JA, Correia N, Bergstrom R, Adami HO, Persson I. 1999. Breast-cancer risk following long-term oestrogen- and oestrogenprogestin-replacement therapy. Int J Cancer 81:339–344. Malone KE, Daling JR, Weiss NS. 1993. Oral contraceptives in relation to breast cancer. Epidemiol Rev 15:80–97. Manjer J, Malina J, Berglund G, Bondeson L, Garne JP, Janzon L. 2001. Increased incidence of small and well–differentiated breast tumours in post–menopausal women following hormone–replacement therapy. Int J Cancer 92:919–922. Marchbanks PA, McDonald JA, Wilson HG, et al. 2002. Oral contraceptives and the risk of breast cancer. N Engl J Med 346:2025–2032. Martinez ME, Grodstein F, Giovannucci E, et al. 1997. A prospective study of reproductive factors, oral contraceptive use, and risk of colorectal cancer. Cancer Epidemiol Biomarkers Prev 6:1–5. Mink PJ, Folsom AR, Sellers TA, Kushi LH. 1996. Physical activity, waist–to–hip ratio, and other risk factors for ovarian cancer: A follow–up study of older women. Epidemiology 7:38–45. Modan B, Hartge P, Hirsh–Yechezkel G, et al. 2001. Parity, oral contraceptives, and the risk of ovarian cancer among carriers and noncarriers of a BRCA1 or BRCA2 mutation. N Engl J Med 345:235–240. Moreno V, Bosch FX, Munoz N, et al. 2002. Effect of oral contraceptives on risk of cervical cancer in women with human papillomavirus infection: The IARC multicentric case-control study. Lancet 359:1085– 1092. Nanda K, Bastian LA, Hasselblad V, Simel DL. 1999. Hormone replacement therapy and the risk of colorectal cancer: A meta-analysis. Obstet Gynecol 93:880–888. Narod SA, Dube MP, Klijn J, et al. 2002. Oral contraceptives and the risk of breast cancer in BRCA1 and BRCA2 mutation carriers. J Natl Cancer Inst 94:1773–1779. Narod SA, Risch H, Moslehi R, et al. 1998. Oral contraceptives and the risk of hereditary ovarian cancer. Hereditary Ovarian Cancer Clinical Study Group. N Engl J Med 339:424–428. Negri E, Tzonou A, Beral V, et al. 1999. Hormonal therapy for menopause and ovarian cancer in a collaborative re-analysis of European studies. Int J Cancer 80:848–851. Nelson HD. 2004. Postmenopausal estrogen for treatment of hot flashes: clinical applications. JAMA 291:1621–1625. Nelson HD, Humphrey LL, Nygren P, Teutsch SM, Allan JD. 2002. Postmenopausal hormone replacement therapy: Scientific review. JAMA 288:872–881. Ness RB, Grisso JA, Klapper J, et al. 2000. Risk of ovarian cancer in relation to estrogen and progestin dose and use characteristics of oral contraceptives. SHARE Study Group. Steroid Hormones and Reproduction. Am J Epidemiol 152:233–241. Ness RB, Grisso JA, Vergona R, Klapper J, Morgan M, Wheeler JE. 2001. Oral contraceptives, other methods of contraception, and risk reduction for ovarian cancer. Epidemiology 12:307–312. Newcomb PA, Longnecker MP, Storer BE, et al. 1996. Recent oral contraceptive use and risk of breast cancer. United States. Cancer Causes Control 7:525–532. Newcomb PA, Titus–Ernstoff L, Egan KM, et al. 2002. Postmenopausal estrogen and progestin use in relation to breast cancer risk. Cancer Epidemiol Biomarkers Prev 11:593–600. Newcomb PA, Trentham–Dietz A. 2003. Patterns of postmenopausal progestin use with estrogen in relation to endometrial cancer. United States. Cancer Causes Control 14:195–201. Newcomer LM, Newcomb PA, Potter JD, et al. 2003. Postmenopausal hormone therapy and risk of breast cancer by histologic type. United States. Cancer Causes Control 14:225–233. Noller KL. 2002. Estrogen replacement therapy and risk of ovarian cancer. JAMA 288:368–369.

487

Olsson HL, Ingvar C, Bladstrom A. 2003. Hormone replacement therapy containing progestins and given continuously increases breast carcinoma risk in Sweden. Cancer 97:1387–1392. Parazzini F, La Vecchia C, Negri E, et al. 1997. Case-control study of oestrogen replacement therapy and risk of cervical cancer. BMJ 315: 85–88. Parazzini F, La Vecchia C, Negri E, Villa A. 1994. Estrogen replacement therapy and ovarian cancer risk. Int J Cancer 57:135–136. Persson I, Weiderpass E, Bergkvist L, Bergstrom R, Schairer C. 1999. Risks of breast and endometrial cancer after estrogen and estrogen–progestin replacement. Cancer Causes Control 10:253–260. Petitti DB. 2003. Clinical practice. Combination estrogen–progestin oral contraceptives. N Engl J Med 349:1443–1450. Pike MC, Peters RK, Cozen W, et al. 1997. Estrogen–progestin replacement therapy and endometrial cancer. J Natl Cancer Inst 89:1110–1116. Pike MC, Ross, RK. 2000. Progestins and menopause: epidemiological studies of risks of endometrial and breast cancer. Steroids 65:659–664. Porch JV, Lee IM, Cook NR, Rexrode KM, Burin JE. 2002. Estrogen-progestin replacement therapy and breast cancer risk: The Women’s Health Study. United States. Cancer Causes Control 13:847–854. Purdie D, Green A, Bain C, et al. 1995. Reproductive and other factors and risk of epithelial ovarian cancer: An Australian case-control study. Survey of Women’s Health Study Group. Int J Cancer 62:678–684. Purdie DM, Bain CJ, Siskind V, et al. 1999. Hormone replacement therapy and risk of epithelial ovarian cancer. Br J Cancer 81:559–563. Riman T, Dickman PW, Nilsson S, et al. 2002. Hormone replacement therapy and the risk of invasive epithelial ovarian cancer in Swedish women. J Natl Cancer Inst 94:497–504. Risch HA. 1998. Hormonal etiology of epithelial ovarian cancer, with a hypothesis concerning the role of androgens and progesterone. J Natl Cancer Inst 90:1774–1786. Risch HA. 1996. Estrogen replacement therapy and risk of epithelial ovarian cancer. Gynecol Oncol 63:254–257. Rodriguez C, Calle EE, Coates RJ, Miracle–McMahill HL, Thun MJ, Heath CW Jr. 1995. Estrogen replacement therapy and fatal ovarian cancer. Am J Epidemiol 141:828–835. Rodriguez C, Patel AV, Calle EE, Jacob EJ, Thun MJ. 2001. Estrogen replacement therapy and ovarian cancer mortality in a large prospective study of US women. JAMA 285:1460–1465. Romieu I, Willett WC, Colditz GA, et al. 1989. Prospective study of oral contraceptive use and risk of breast cancer in women. J Natl Cancer Inst 81:1313–1321. Rosenberg L, Palmer JR, Rao RS, et al. 1996. Case-control study of oral contraceptive use and risk of breast cancer. Am J Epidemiol 143:25– 37. Rosenberg L, Shapiro S, Slone D, et al. 1982. Epithelial ovarian cancer and combination oral contraceptives. JAMA 247:3210–3212. Rosenblatt KA, Gao DL, Ray RM, Nelson ZC, Thomas DB. 2004. Contraceptive methods and induced abortions and their association with the risk of colon cancer in Shanghai, China. Eur J Cancer 40:590–593. Rosenblatt KA, Thomas, DB. 1991. Hormonal content of combined oral contraceptives in relation to the reduced risk of endometrial carcinoma. The WHO Collaborative Study of Neoplasia and Steroid Contraceptives. Int J Cancer 49:870–874. Ross RK, Paganini–Hill A, Wan PC, Pike MC. 2000. Effect of hormone replacement therapy on breast cancer risk: Estrogen versus estrogen plus progestin. J Natl Cancer Inst 92:328–332. Rossouw JE, Anderson GL, Prentice RL, et al. 2002. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: Principal results From the Women’s Health Initiative randomized controlled trial. JAMA 288:321–333. Schairer C, Adami HO, Hoover R, Persson I. 1997. Cause–specific mortality in women receiving hormone replacement therapy. Epidemiology 8:59–65. Schairer C, Lubin J, Troisi R, Sturgeon S, Brinton L, Hoover R. 2000. Menopausal estrogen and estrogen–progestin replacement therapy and breast cancer risk. JAMA 283:485–491. Schildkraut JM, Calingaert B, Marchbanks PA, Moorman PG, Rodriguez GC. 2002. Impact of progestin and estrogen potency in oral contraceptives on ovarian cancer risk. J Natl Cancer Inst 94:32–38. Shapiro S, Rosenberg L, Hoffman M, et al. 2003. Risk of invasive cancer of the cervix in relation to the use of injectable progestogen contraceptives and combined estrogen/progestogen oral contraceptives. South Africa. Cancer Causes Control 14:485–495. Siskind V, Green A, Bain C, Purdie D. 2000. Beyond ovulation: Oral contraceptives and epithelial ovarian cancer. Epidemiology 11:106–110. Sit AS, Modugno F, Weissfeld JL, Berga SL, Ness RB. 2002. Hormone replacement therapy formulations and risk of epithelial ovarian carcinoma. Gynecol Oncol 86:118–123.

488

PART III: THE CAUSES OF CANCER

Skegg DC. 2002. Oral contraceptives, parity, and cervical cancer. Lancet 359:1080–1081. Smith DC, Prentice R, Thompson DJ, Herrmann WL. 1975. Association of exogenous estrogen and endometrial carcinoma. N Engl J Med 293: 1164–1167. Smith JS, Green J, Berrington dG, et al. 2003. Cervical cancer and use of hormonal contraceptives: A systematic review. Lancet 361:1159–1167. Stahlberg C, Pedersen AT, Lynge E, et al. 2004. Increased risk of breast cancer following different regimens of hormone replacement therapy frequently used in Europe. Int J Cancer 109:721–727. Stanford JL, Brinton LA, Berman ML, et al. 1993. Oral contraceptives and endometrial cancer: Do other risk factors modify the association? Int J Cancer 54:243–248. Stanford JL, Brinton LA, Hoover RN. 1989. Oral contraceptives and breast cancer: Results from an expanded case-control study. Br J Cancer 60:375–381. The Writing Group for the PEPI Trial. 1996. Effects of hormone replacement therapy on endometrial histology in postmenopausal women. The Postmenopausal Estrogen/Progestin Interventions. PEPI) Trial. JAMA 275: 370–375. Trapido EJ. 1983. A prospective cohort study of oral contraceptives and cancer of the endometrium. Int J Epidemiol 12:297–300. Troisi R, Schairer C, Chow WH, Schatzkin A, Brinton L A, Fraumeni JF Jr. 1997. Reproductive factors, oral contraceptive use, and risk of colorectal cancer. Epidemiology 8:75–79. U.S. Department of Health and Human Services, Public Health Service, National Toxicology Program. 2002. Report on Carcinogens, Tenth Edition. Government Printing Office, Washington D.C. Triangle Park, NC: National Toxicology Program. Ursin G, Henderson BE, Haile RW, et al. 1997. Does oral contraceptive use increase the risk of breast cancer in women with BRCA1/BRCA2 mutations more than in other women? Cancer Res 57:3678–3681. Van Hoften C, Burger H, Peeters PH, Grobbee DE, Van Noord PA, Leufkens HG. 2000. Long–term oral contraceptive use increases breast cancer risk in women over 55 years of age: The DOM cohort. Int J Cancer 87: 591–594. Veronesi U, Maisonneuve P, Rotmensz N, et al. 2003. Italian randomized trial among women with hysterectomy: Tamoxifen and hormone-dependent breast cancer in high-risk women. J Natl Cancer Inst 95:160–165.

Vessey MP, Painter R. 1995. Endometrial and ovarian cancer and oral contraceptives—findings in a large cohort study. Br J Cancer 71: 1340–1342. Voigt LF, Deng Q, Weiss NS. 1994. Recency, duration, and progestin content of oral contraceptives in relation to the incidence of endometrial cancer. Washington, USA. Cancer Causes Control 5:227–233. Voigt LF, Weiss NS, Chu J, Daling JR, McKnight B, van Belle G. 1991. Progestagen supplementation of exogenous oestrogens and risk of endometrial cancer. Lancet 338:274–277. Wang SS, Sherman ME, Hildesheim A, Lacey JV Jr., Devesa S. 2004. Cervical adenocarcinoma and squamous cell carcinoma incidence trends among white women and black women in the United States for 1976–2000. Cancer 100:1035–1044. Weiderpass E, Adami HO, Baron JA, et al. 1999a. Risk of endometrial cancer following estrogen replacement with and without progestins. J Natl Cancer Inst 91:1131–1137. Weiderpass E, Adami HO, Baron JA, Magnusson C, Lindgren A, Persson I. 1999b. Use of oral contraceptives and endometrial cancer risk. Sweden. Cancer Causes Control 10:277–284. Whitehead MI, Fraser D. 1987. The effects of estrogens and progestogens on the endometrium. Modern approach to treatment. Obstet Gynecol Clin North Am 14:299–320. Whittemore AS, Harris R, Itnyre J, the Collaborative Ovarian Cancer Group. 1992. Characteristics relating to ovarian cancer risk: Collaborative analysis of 12 US case-control studies. Am J Epidemiol 136:1184– 1203. Willett WC, Colditz G, Stampfer M. 2000. Postmenopausal estrogens— opposed, unopposed, or none of the above. JAMA 283:534–535. Wittenberg J, Cook LS, Rossing MA, Weiss NS. 1999. Reproductive risk factors for mucinous and non-mucinous epithelial ovarian cancer. Epidemiology 10:761–763. Women’s Health Initiative Study Group. 1998. Design of the Women’s Health Initiative clinical trial and observational study. The Women’s Health Initiative Study Group. Control Clin Trials 19:61–109. Xu WH, Xiang YB, Ruan ZX, et al. 2004. Menstrual and reproductive factors and endometrial cancer risk: Results from a population-based casecontrol study in urban Shanghai. Int J Cancer 108:613–619. Ziel HK, Finkle WD. 1976. Association of estrone with the development of endometrial carcinoma. Am J Obstet Gynecol 124:735–740.

25

Pharmaceuticals Other Than Hormones LAUREL A. HABEL AND GARY D. FRIEDMAN

A

mong the substances in the environment that are of concern as possible carcinogens are therapeutic drugs. Drugs are deliberately ingested by, or injected into, people in amounts often much greater than other chemicals in the environment that are known or suspected to be carcinogenic. A relatively small number of drugs have been demonstrated to cause or promote the development of cancer in humans. Some others have these effects in experimental animals or in vitro systems, but their status as carcinogens in humans remains unclear. There are also a few medications that are known or suspected to be chemopreventive.

OVERVIEW OF PHARMACEUTICALS A drug may be defined as any exogenously administered substance that exerts a physiologic effect (McGettigan et al., 2000). Traditionally, most drugs were small chemicals with molecular weights in the hundreds, as well as a few that were natural human or animal hormones (Benet, 1996). Within the past two decades, a number of protein and peptide drugs, designed to interact with a particular receptor or enzyme, have been approved for clinical use and new drugs, including those with new mechanisms of action (e.g., gene-based therapies), are being developed at an ever-increasing rate (Nies and Spielberg, 1996). The concentration of a drug at its sites of action depends on the amount of the drug administered, the extent and rate of absorption, distribution, binding or localization in tissues, biotransformation, and excretion (Nies and Spielberg, 1996; McGettigan et al., 2000). The effect of most drugs results from their interaction with specialized proteins such as enzymes and cell surface receptors. Drug molecules may be present within body fluids either in their free state or bound to proteins or other constituents. It is usually the free or unbound fraction that interacts with the target proteins. There is considerable inter-individual variation in the response to a drug (Nies and Spielberg, 1996). Factors influencing response include age, sex, some disease conditions (e.g., liver cirrhosis), drug–drug interactions, tolerance, and genetic factors. Genetic factors are the major determinants of the normal variability of drug effects and are responsible for many significant quantitative and qualitative differences in pharmacologic activity.

CHEMICAL CARCINOGENESIS Chemical carcinogens, such as medications, may be classified by their genotoxicity (Klaassen, 1996a; Weisburger and Williams, 2000). Genotoxic carcinogens interact with DNA and result in structural changes at the level of the gene (e.g., adduct formation, mutation, chromosomal aneuploidy). These carcinogens are usually unreactive with DNA in their original state but are converted to reactive intermediates (or ultimate carcinogens) in the body by P450-dependent monooxygenases. Epigenetic carcinogens are not mutagenic but act through other mechanisms (e.g., cell proliferation) that predispose to cancer. Carcinogens may act at any step during the multistage process of carcinogenesis (Rothman and Poole, 1996). Several somatic events are necessary before a replicating tissue accumulates enough genetic

damage to go from a normal phenotype to a malignant one (Hoover, 1998). Although a number of chemical carcinogens have induction periods of 20 or more years (e.g., asbestos and mesothelioma), some chemical agents, including some medications, may act late in the carcinogenic process, very close in time to the clinical manifestation of disease. For example, menopausal hormones (estrogen plus progestin) have been shown to increase the incidence of breast cancer and decrease the incidence of colon cancer within only 5 years of initiating therapy (Rossouw et al., 2002) and some immunosuppressive agents increase the incidence of lymphomas within months of treatment initiation (Kinlen, 1996). In addition to chemicals that are themselves carcinogens, candidate carcinogenic and chemopreventive agents include compounds that influence or alter the endogenous synthesis of carcinogens, the rate and amount of absorption of carcinogenic chemicals, the metabolism of carcinogens or the generation of free radicals and activated forms of oxygen, or the covalent binding of carcinogens to DNA (Schottenfeld, 1996).

DRUG SAFETY Most drugs produce several effects, only one of which is usually desired (Nies and Spielberg, 1996). The spectrum of undesired effects range from very minor, nondeleterious effects (e.g., dry mouth) to effects that are life threatening. Drugs must go through a series of premarketing studies (Phases I–III clinical trials for both safety and efficacy) before they can be approved by the Food and Drug Administration (FDA) for marketing in the United States. Although this preapproval process is quite rigorous, it cannot identify all adverse effects (Wood et al., 1998). When a drug is first marketed, it has usually been tested in only a limited number of well-characterized patients (Nies and Spielberg, 1996). The selection of patients for clinical trials often excludes individuals with coexisting diseases and/or the very old and very young. For these reasons, the results may be poorly generalizable to the full population that is exposed to the drug once it is marketed. In addition, trials are likely to miss rare adverse effects or those with induction periods longer than the trials’ duration of follow-up. In the United States, postmarketing surveillance has been largely voluntary (Nies and Spielberg, 1996). Reporting of adverse events by individual health care providers and consumers to the manufacturer or the FDA is not required by law. The manufacturer, however, is legally obligated to forward reports of adverse events to the FDA. Pharmaceutical companies may be required by the FDA to conduct additional studies of approved drugs (also called phase IV studies) because of particular safety questions that were not fully addressed in premarketing studies or that arise following reports of unanticipated adverse events or new information from in vivo or in vitro studies.

PHARMACOEPIDEMIOLOGIC STUDIES Human data on drug carcinogenicity other than case reports have come from ad hoc pre- and postmarketing studies of specific drugs and from programs that allow the postmarketing evaluation of several drugs. Programs that have provided or could provide such information

489

490

PART III: THE CAUSES OF CANCER

include the Kaiser Permanente drug surveillance program (Friedman et al., 2000), the hospital-based Boston Collaborative Drug Surveillance Program and its successor, the Slone Epidemiology Unit at Boston University (Shapiro, 2000), studies based on data from Group Health Cooperative (Saunders et al., 2000), the Harvard Community Health Plan (Chan and Platt, 2000), the United Health Group (Shatin et al., 2000), Medicaid databases (Carson et al., 2000), general practices in the United Kingdom (Garcìa Rodrìguez et al., 2000; Mann, 2000), and studies involving record linkage in geographic areas such as Saskatchewan (Downey et al., 2000) and some European countries (Evans and MacDonald, 2000; Leufkens and Urquhart, 2000). In addition to hypotheses generated by drug surveillance, hypotheses about specific drug-cancer associations may arise from theoretical considerations of the chemical and pharmacokinetic properties of the compound, from results of in vitro tests for mutagenicity, genotoxicity or other effects (e.g., cell proliferation, apoptosis), or from reports of experiments in animals. Drug-cancer associations may also be suggested by results from large epidemiologic studies of multiple exposures and from long-term follow-up of clinical trial populations for adverse events. The carcinogenic or chemopreventive effect of a drug at one cancer site may lead to suspicion of similar effects at other sites. The International Agency for Research on Cancer (IARC) conducts systematic evaluations of the evidence for carcinogenicity of hundreds of chemicals, including many drugs. In these evaluations, definitive evidence for carcinogenicity in humans is provided by data from epidemiologic studies or clinical trials, although animal studies and other biological data provide contributing evidence. The evaluations to date have been published in a series of monographs (IARC, 1980; IARC, 1987; IARC, 1990; IARC, 1996; IARC, 2000; IARC, 2001). More than 200 medications have been evaluated and approximately 70 nonhormonal drugs have been categorized as definitely, probably, or possibly carcinogenic to humans (Table 25–1). At the time of these IARC evaluations, human data were unavailable on many of these medications, and were limited or inadequate for a number of others.

METHODOLOGIC ISSUES IN STUDIES OF DRUGS AND CANCER The indications, or contraindications, for treatment with the medication of interest are often the most important potential confounders to consider in drug studies. A medication may be associated with cancer at a specific site, solely because it is used to treat or it is contraindicated for a condition that is a risk factor for that cancer. Such confounding by indication can sometimes be at least partially addressed by comparing cancer risk among users of the medication of interest to risk among users of other medications used to treat the same condition and/or to risk among untreated individuals with this condition. Use of a medication may be related to the severity of such a risk factor, and therefore if severity is not taken into account, residual confounding may remain. In addition, a drug may be spuriously associated with the development of cancer when it is used or is contraindicated for early symptoms of the disease. Lagging the analysis to only include drug use greater than 1 or 2 years prior to diagnosis may help address this bias. Because users of some medications may be more likely to be followed by a health care provider and/or undergo routine cancer screening, their cancers may be detected earlier than the cancers of nonusers of these medications. Bias in relative risk estimates may also result when use of a medication increases the likelihood of undergoing diagnostic procedures because of a known or suspected relation between the medication and the cancer of interest (e.g., menopausal hormones and endometrial cancer). Users of a medication may differ from nonusers with respect to risk factors for a specific cancer. For example, users of oral contraceptives are also more likely to be smokers (West et al., 2000). Risk factors also may differ between current and past users of a medication and new versus prevalent users (Ray, 2003). In addition, patients often take more than one medication, and the use of some medications is often correlated with the use or nonuse of others. For example, women who

use menopausal hormones are more likely to use several other drugs, as well (Small et al., 2001). Use of two drugs may be negatively correlated when the use of one medication is a contraindication for the use of another. As with other exposure–disease relationships (Armstrong et al., 1992), there is likely to be an etiologically relevant exposure period during which use of a specific medication may cause or prevent a specific cancer. Inclusion of medication use outside this window may lead to misclassification of the exposure and reduce the likelihood of observing a true drug–cancer association. Although the critical time period is often unknown, collecting information on medication use with respect to time (e.g., time of first use, time of last use, duration of use) will allow for the examination of different time windows. Obtaining comprehensive information on the dosage is also important. Characterizing study subjects simply as exposed versus unexposed increases the rate of misclassification and biases the results toward the null (Collet and Boivin, 2000). As with other exposures, some medications may be carcinogenic only after a certain cumulative dose or duration of use, and an observation that cancer risk increases with increasing dose or duration supports a causal relation (Rothman, 1986). Unfortunately, many studies of the potential carcinogenic effect of medications have had only minimal information on drug exposure and/or have followed patients for short periods after initiating therapy. Accurate information on medication use is often difficult to obtain (West et al., 2000). Studies using self-reported information on historical medication use benefit from incorporating calendars of important life events, pictures of the medication(s), and by including questions about both the indications for and the actual use of the medication(s) of interest. When using dispensing records (either hard copy or electronic), information on whether or not the medication was actually taken is usually unavailable. Researchers have sometimes dealt with this problem by requiring two or more dispensings before classifying a study subject as exposed (e.g., Rossing et al., 2000). Use of a drug is usually a relatively rare exposure in a population, and studying the drug’s effects on the incidence of a specific cancer (usually a relatively rare disease) creates design challenges. Although large health surveys or electronic pharmacy databases linked to cancer registries can facilitate the identification of sufficient numbers of drug users and cancer end points, additional data usually must be collected to address potential confounding. Case-control or case-cohort studies nested within large cohorts of drug users and nonusers may still require sizable data collection efforts. More efficient methods for addressing or assessing the presence of confounding have recently been proposed, such as two-stage sampling and case only strategies (Suissa, 2000).

ASSOCIATIONS OF NONHORMONAL DRUGS WITH CANCER Presented below are summaries of available information for selected nonhormonal medications that are known or suspected to be carcinogenic or chemopreventive in humans. We have focused on those medications, or groups of medications, that have received a fair amount of attention in human studies.

ANALGESIC–ANTIPYRETIC AND ANTIINFLAMMATORY MEDICATIONS Phenacetin Phenacetin is an aniline derivative closely related to acetaminophen. It was widely used during the 1940s and 1950s as an ingredient in many nonprescription, combination analgesic–antipyretic preparations. These preparations were often used to excess because of mild, subjective stimulation or euphoria induced by phenacetin. Phenacetin is principally metabolized by the liver to N-acetyl-p-aminophenol (acetaminophen), conjugated, and excreted in the urine (Insel, 1996). However, a small proportion is metabolized in the kidney to p-

491

Pharmaceuticals Other Than Hormones Table 25–1. Drugs Considered Definitely, Probably, or Possibly Carcinogenic to Humans by the International Agency for Research on Cancer (IARC)*

Analgesics and anti-inflammatory drugs Phenacetin Analgesic mixtures containing phenacetin Phenazopyridine hydrochloride Anesthetics Chloroform Antibacterial drugs Chloramphenicol Nifuradene Furathiazole Panfuran S Antineoplastics Adriamycin (doxorubicin) Amsacrine† Azacytidine (5-azacytidine) Chlornaphazine Bischloroethyl nitrosourea (BCNU) Bleomycin Busulfan (Myleran) Chlorambucil (also immunosuppressant) CCNU (Lomustine) Methyl-CCNU Chlorozotocin Cisplatin Cyclophosphamide (also immunosuppressant) Dacarbazine Daunorubicin Etoposide† Etoposide in combination with Cisplatin and bleomycin† Melphalan Merphalan Mitomycin C Mitoxantrone† MOPP and other combined therapy Nitrogen mustard Nitrogen mustard N-oxide Procarbazine hydrochloride Streptozotocin Teniposide† Treosulphan Trimustine hydrochloride Thiotepa Uracil mustard Antifungals, antiprotozoans, antiparasitics Clofenotane Griseofulvin‡ g-Hexachloroclohexane Metronidazole (also antibacterial) Niridazole Antiseptic agents b-Propiolactone Antivirals Zalcitabine† Zidovudine (AZT)† Dermatological agents Arsenic salts (also antineoplastic) Coal tars 5-Methoxypsoralen Methoxsalen + UV Safrole Drugs for treating anemia Iron-dextran complex Drugs for treating cardiovascular disorders Phenoxybenzamine hydrochloride

Data Source Drug

Data Source Drug

Table 25–1. (cont.)

Humana

Animala

Overall Evaluationb

L S

S L

2A 1

I

S

2B

I

S

2B

L ND ND ND

I S S S

2A 2B 2B 2B

I I (ND) ND S L I S S I S ND I S

S S S L S L L S S L S S S

2A 2B 2A 1 2A 2B 1 1 2A 1 2A 2A 1

I ND L S

S S I I (ND)

2B 2B 2A 1

S ND ND L S L ND I ND L S ND S I

S S S I (ND) I S S S S I (ND) ND S S S

1 2B 2B 2B 1 2A 2B 2A 2B 2A 1 2B 1 2B

I I I I ND

S S L S S

2B 2B 2B 2B 2B

ND

S

2B

I I

S S

2B 2B

S S I S ND

L S S S S

1 1 2A 1 2B

I

S

2B

ND

S

2B

Drugs for treating central nervous disorders Oxazepam§ (antianxiety) Phenobarbital Phenytoin§ Drugs for treating thyroid disorders Methylthiouracil Propylthiouracil Thiouracil‡ Thiourea Immunosuppressants Azathioprine Cyclosporin Laxatives Danthron (dantron) Phenolphthalein† Miscellaneous and experimental drugs 2-amino-5-(5-nitro-2-furyl)1,3,4-thiadiazole Lasiocarpine N-Methyl-N-nitrosourea Nafenopin

Humana

Animala

Overall Evaluationb

I I I

S S S

2B 2B 2B

ND I I ND

S S S S

2B 2B 2B 2B

S S

L L

1 1

ND I

S S

2B 2B

ND

S

2B

ND ND ND

S S S

2B 2A 2B

a L, limited evidence of carcinogenicity; I, inadequate evidence of carcinogenicity; ND, no data; S, sufficient evidence of carcinogenicity. b Group 1, agent is carcinogenic to humans; 2A, agent is probably carcinogenic to humans; 2B, agent is possibly carcinogenic to humans. *From Volume 50, IARC (1990) except where noted. † Volume 76, IARC (2000). ‡ Volume 79, IARC (2001). § Volume 66, IARC (1996).

aminophenol and other known urinary tract carcinogens (Carpenter and Mudge, 1981). Phenacetin has been shown to be a potent promoter of tumor formation in previously initiated rat bladder tissue (Ito et al., 1984). An association between chronic abuse of combination analgesic– antipyretic preparations and a distinctive form of interstitial nephropathy with renal papillary necrosis was first described in the 1950s (Hultengren, 1961). Because phenacetin was included in nearly all preparations associated with the nephropathy, this agent has been most frequently implicated. However, phenacetin was rarely used as a single agent and other analgesics may also produce papillary necrosis (Prescott, 1982). Increased occurrence of transitional cell cancers of the renal pelvis and urinary bladder in persons with analgesic nephropathy was subsequently reported (Bengtsson et al., 1968; Hultengren, 1968; Johansson et al., 1974). Case-control studies have confirmed associations of chronic or excessive use of phenacetincontaining analgesics with both of these urinary cancers (Fokkens, 1979; McCredie et al., 1982; McCredie et al., 1983; Piper et al., 1985; McCredie and Stewart, 1988; McLaughlin et al., 1992; Kreiger et al., 1993; McCredie et al., 1993). A weaker association with renal cell carcinoma has also been observed, but the evidence has been mixed (McLaughlin et al., 1984; McCredie and Stewart, 1988; McCredie et al., 1995; Gago-Dominguez et al., 1999). The level of exposure at which risk for urothelial cancer begins to increase cannot be specified from most studies, but one study (Fokkens, 1979) found no increase in risk for bladder cancer if lifetime consumption was below 2.0 kg. Some evidence suggests that tissue damage and scarring in the renal pelvis are likely necessary steps in phenacetin’s causation of urothelial cancer (Bringuier et al., 1998; Stewart et al., 1999). Evidence is mixed regarding the nonphenacetin analgesics, acetaminophen, aspirin, and other nonsteroidal anti-inflammatory drugs (NSAIDs) and risk of urothelial cancers (McCredie et al., 1983; Piper et al., 1985; McCredie et al., 1988; McCredie and Stewart, 1988; Chow et al., 1994; Rosenberg et al., 1998b; Pommer et al., 1999).

492

PART III: THE CAUSES OF CANCER

Acetaminophen (paracetamol) has been of special concern because it is the chief metabolite of phenacetin (McCredie et al., 1993) and further studies are needed to clearly implicate or exonerate this drug. Although abuse of phenacetin-containing analgesics apparently increases risk for urothelial cancers, occasional use of such products in appropriate doses may not increase risk. Likewise, it has not been established whether other analgesics, alone or in combination with phenacetin, can also increase cancer risk. Because of the problems it has caused, phenacetin is no longer used in standard analgesic mixtures. In New South Wales, Australia, use of phenacetin was highly prevalent until sales were prohibited from 1979. The incidence of renal pelvis cancer declined slightly and bladder cancer more so after 1985, but the contribution of the phenacetin ban is difficult to distinguish from the effects of the decline in prevalence of cigarette smoking (McCredie et al., 1999).

Phenazopyridine Hydrochloride Oral administration of phenazopyridine, an orange-red azo dye used as a urinary tract analgesic alone or in combination with sulfonamide preparations, has been shown to increase the incidence of hepatic carcinomas in mice and to induce colon and rectum tumors in rats (National Cancer Institute, 1978). The only epidemiologic data in humans comes from the Kaiser Permanente study of prescription drugs. Among 2214 recipients of the drug, no significant excess was noted for cancer at any site after 3 to 7 years of follow-up (Friedman and Ury, 1980). Twelve additional years of follow-up have now been reported (Van Den Eeden and Friedman, 1995), again with no excess of cancer for any site or for all sites combined. In later observations with up to 21 years of follow-up, associations were noted with prostate cancer, lymphosarcoma, and multiple myeloma, but these need further evaluation (Friedman and Habel, unpublished data). Thus, the epidemiologic data are very limited and inconclusive as to carcinogenicity of phenazopyridine in humans.

Phenylbutazone Phenylbutazone, a potent anti-inflammatory drug, was commonly used in the past to treat rheumatic conditions. It is now little used because of concern about side effects and because several other potent nonsteroidal anti-inflammatory agents have become available. One of its feared although infrequent side effects is bone marrow depression with resulting leukopenia, agranulocytosis, or aplastic anemia (American Medical Association, 1986). There have been a sizeable number of case reports of leukemia after phenylbutazone administration, but such occurrences are difficult to ascribe to anything but coincidence, particularly as phenylbutazone was so commonly prescribed when they were reported (IARC, 1987). A follow-up study of 489 patients with rheumatoid arthritis revealed an incidence of non-Hodgkin’s lymphoma far greater than expected. Sixty percent of the hospital charts of those who developed this and other lymphohematopoietic malignancies showed evidence of phenylbutazone use, but there were reasons to doubt that phenylbutazone was the cause (Symmons, 1985; IARC, 1987). Two Kaiser Permanente studies found no significant excess risk of leukemia attributable to phenylbutazone (Friedman and Ury, 1980; Friedman, 1982). An association of musculoskeletal disease with leukemia was noted; this could underlie an apparent phenylbutazone-leukemia association. In a case-control study of chronic lymphocytic leukemia in Yorkshire, U.K., a statistically significant 2.2-fold increased risk was observed with use of phenylbutazone within 10 years before diagnosis. In this study, two musculoskeletal diseases, rheumatoid arthritis and osteoarthritis, were not associated with leukemia risk (Cartwright et al., 1987). In one experiment, exposure to phenylbutazone for 2 years led to renal tubular cell neoplasms in rats and an increased incidence of liver tumors in male but not female mice (Kari et al., 1995). In another, carcinogenic and tumor promoting effects were absent to minimal with 2-year administration of phenylbutazone to rats (Meakawa et al., 1987).

Evidence to date is limited and conflicting, and whether phenylbutazone is carcinogenic, or specifically leukemogenic, is still uncertain. On the positive side, this medication is one of the nonsteroidal antiinflammatory drugs that have an antiproliferative effect on human colon cancer cells in vitro (Hixson et al., 1994) and could thus have cancer-preventive effects on the large bowel.

Aspirin and Other NSAIDs as Chemopreventive Agents In the late 1980s and early 1990s, several epidemiologic studies reported an inverse association between use of aspirin and other NSAIDs and risk of colon cancer and adenomatous polyps (Baron and Sandler, 2000). A number of subsequent investigations, using both case-control and cohort designs, have supported these findings and have largely addressed early concerns that the association was due primarily to the avoidance of aspirin among individuals with early symptoms of the disease. In these observational studies, sustained use of NSAIDs has been consistently associated with a 30–50% reduction in risk of adenomatous polyps and colorectal cancer (Thun et al., 2002). However, the only large prevention trial with data on colon cancer incidence found no reduction in risk with 12 years of follow-up. It is not clear whether randomization reduced uncontrolled confounding present in the observational studies or that the dose and/or duration of aspirin use in the intervention trial (325 mg every other day for 5 years) was insufficient to influence cancer risk. A number of mechanisms for the chemopreventive effect of aspirin and other NSAIDs have been proposed. NSAIDs have been shown to restore apoptosis and to inhibit angiogenesis (Thun et al., 2002). There continues to be uncertainty about the molecular pathways responsible for the anticarcinogenic effects of NSAIDs. NSAIDs inhibit the activity of cyclooxygenase (COX), the rate limiting enzyme in the prostaglandin pathway. There are two isoforms of cyclooxygenase (COX-1 and COX-2). NSAIDs vary in their abilities to inhibit COX1 and COX-2 at different concentrations and in different tissues (Thun et al., 2002). Elevated COX-2 expression has been found in a number of premalignant and malignant conditions (Michalowski, 2002). Prostaglandin exhibits a number of biological activities expected to promote tumor growth and progression: it increases cell proliferation, angiogenesis, and bcl-2 (an antiapoptosis protein) levels and decreases natural killer cell activity and immune surveillance (Badawi, 2000). There are also experimental models suggesting that NSAIDs may affect apoptosis and other mechanisms through prostaglandinindependent pathways (Thun et al., 2002). Given the consistent findings of an inverse association between use of NSAIDs and colon cancer risk, the possible generalizability of the proposed biologic mechanisms, and a growing body of supporting animal and in vitro data, there has been increasing interest in examining whether these medications may be chemopreventive at other sites as well. Several observational studies have provided relative risk estimates for cancer at multiple sites or for cancer at specific sites (Khuder and Mutgi, 2001; Bosetti et al., 2002; Johnson et al., 2002; Corley et al., 2003; Perron et al., 2003). The number of studies providing estimates for any given site is still relatively small. At several sites (e.g., esophagus, prostate, breast, ovary), most studies have observed inverse associations; however, the relative risk estimates have often been only slightly less than 1.0 and not statistically significant. In addition, most studies have not had sufficient information on NSAID use to examine dose, duration, or recency of use in any detail. A number of clinical trials have now shown that the NSAID sulindac and the COX-2 inhibitor celecoxib suppress adenomatous polyps and cause regression of existing polyps among patients with familial adenomatous polyposis (FAP) (Thun et al., 2002). Several studies are underway to examine whether NSAIDs reduce adenomatous polyps or colon cancer in other high-risk populations. In addition, there are a number of small phase II and phase III clinical trials underway to examine the effectiveness of the COX-2 inhibitors in preventing or treating cancer at other sites (Michalowski, 2002). In summary, there is a growing body of evidence from experimental, clinical, and epidemiologic data supporting the chemopreventive

Pharmaceuticals Other Than Hormones activity of NSAIDs. Additional epidemiologic studies are needed to determine the most appropriate dose, duration, and type of NSAID that should be tested in large intervention studies of colon cancer prevention and to identify the subpopulations most likely to benefit. Studies are also needed to further examine the potential chemopreventive benefits of NSAIDs at other cancer sites.

ANTIBACTERIAL DRUGS Chloramphenicol Because the antibiotic chloramphenicol can cause aplastic anemia, and because some patients with chloramphenicol-induced aplastic anemia go on to develop leukemia (IARC, 1987), this drug has for many years been suspected of being a cause of leukemia. Adding to the suspicion, acute nonlymphocytic leukemia and aplastic anemia have other causal factors in common such as ionizing radiation and benzene exposure. A large, population-based epidemiological study supported the chloramphenicol-leukemia link. This case-control study included 309 children with leukemia and 618 matched controls in Shanghai, China. Chloramphenicol users showed a statistically significant 2.3-fold increased risk of leukemia. Supporting a causal interpretation was a dose-response relationship—the relative risk reaching 9.7 for all leukemias, 11.0 for acute lymphocytic leukemia, and 12.0 for acute nonlymphocytic leukemia when the duration of chloramphenicol use was more than 10 days. Also, a closely related drug, syntomycin, was associated with subsequent leukemia, whereas other drugs, including many antibiotics, were not. The greater prior use of chloramphenicol in the leukemia patients persisted when attention was restricted to the period at least 2 years before diagnosis, suggesting that the association was not attributable to treatment of leukemia- or preleukemiainduced infections with chloramphenicol (Shu et al., 1987). A later case-control study of 257 patients with leukemia and 324 matched controls in a health maintenance organization in northern California and Oregon showed an inverse chloramphenicol-leukemia association with an odds ratio of 0.4 (95% CI 0.2–0.97) based on 7 exposed cases. The authors noted that the association was inconsistent by gender, suggesting that it might have been spurious (Doody et al., 1996). Because use of oral chloramphenicol tablets was greatly restricted in the United States during much of the study period, many of the chloramphenicol notations may have been for eye drops, with little systemic exposure. Based on current evidence, chloramphenicol is probably leukemogenic. It should only be administered when no appropriate alternatives are available, and its duration of use should be as short as possible. It is still not clear whether topical application of chloramphenicol, particularly to the eyes, results in sufficient exposure to be of concern (Besamusca and Bastiaensen, 1986; Stevens and Mission, 1987).

Isoniazid Isoniazid, a drug commonly used to treat and prevent tuberculosis, is a derivative of hydrazine, a potent mutagen and carcinogen in animals (Balo, 1979). Hydrazine is released during metabolism of isoniazid and can be detected in the plasma and urine of subjects taking the drug (Blair et al., 1985), particularly in patients who metabolize the drug slowly (slow acetylator phenotype). Isoniazid induces lung tumors in mice after both intraperitoneal and oral administration (Balo, 1979). Two large follow-up studies in humans treated for tuberculosis (Stott et al., 1976; Clemmesen and Hjalgrim-Jensen, 1979) found increased cancer mortality, particularly from lung cancer, in isoniazid users compared to the general population. In general, these studies did not control for cigarette smoking, which may have been more frequent among persons with tuberculosis than in the general population, or for the known association of lung cancer with tuberculosis that is independent of isoniazid use (Zheng et al., 1987). An association of isoniazid with bladder cancer has also been reported (Miller et al., 1978). A large randomized trial of isoniazid chemoprophylaxis for household contacts of tuberculosis patients conducted by the U.S. Public

493

Health Service in the late 1950s provides the best evaluation of possible isoniazid carcinogenicity. This study avoided confounding by the presence of tuberculosis among isoniazid users and, by randomization, minimized differences in smoking between groups. Follow-up of 14 years for cancer mortality (Glassroth et al., 1977b) and 19 years for cancer incidence (Costello and Snider, 1980) failed to show increased risks for lung, bladder, or all cancer. No association of isoniazid with bladder cancer was found in two case-control studies (Glassroth et al., 1977a; Kantor et al., 1985). Although these findings are reassuring, the animal studies and the biological considerations noted above dictate longer term follow-up before a delayed carcinogenic effect of isoniazid can be ruled out.

ANTIDEPRESSANT DRUGS Emotional depression has long been considered a possible risk factor for cancer development, with conflicting findings (Linkins and Comstock, 1990). Based on a finding that amitriptyline and fluoxetine stimulated the growth of experimentally induced tumors in mice (Brandes et al., 1992) it was suggested that attention should be focused on the drugs used to treat depression (Brandes, 1992). The earlier antidepressant drugs were primarily the tricyclics, tetracyclics, and monoamine oxidase (MAO) inhibitors, and in response to this concern, some evidence was presented suggesting that two of the most frequently used tricyclic drugs, amitriptyline and imipramine, did not increase cancer risk (Friedman, 1992). After the introduction in the late 1980s of a new class of drugs, the selective serotonin reuptake inhibitors (SSRIs), pharmacologic treatment of depression became much more prevalent, and concerns about possible carcinogenic effects of all antidepressants increased. Most attention has been directed toward cancers affecting women. The findings regarding breast cancer have been mixed. The earliest studies found to have relevant data were both published in 1982. One showed both a positive association with use of antidepressant drugs, complicated by interaction with socioeconomic status (Wallace et al., 1982) and the other, a negative (protective) association (Danielson et al., 1982a). A twofold increased risk of breast cancer with use of tricyclics was found in a large Canadian case-control study (Sharpe et al., 2002), and subgroup analysis in another Canadian case-control study weakly suggested the possibility of a similar risk elevation with tricyclic use for over 2 years and elevated risk with only paroxetine, one of three SSRIs examined individually (Beebe et al., 2000; Cotterchio et al., 2000; Cotterchio and Kreiger, 2000; Lawlor, 2000). Another large case-control study, conducted in several hospitals in the United States, was largely negative, but a suggestion of increased risk appeared in one subgroup, users of SSRIs in the previous year (Kelly et al., 1999). A large retrospective cohort study, based in New Jersey, showed no elevation of breast cancer risk for antidepressants or specific antidepressant drugs, during follow-up of up to 7.5 years (Wang, 2001). In a case-control study in Massachusetts (Harlow and Cramer, 1995), risk for ovarian cancer was increased approximately twofold among women who had taken antidepressants before the SSRI era, in the tricyclic and MAO inhibitor categories. Risk was increased more with long-term use and first use before age 50 years. In a later casecontrol study (Harlow et al., 1998) these investigators again found an association with ovarian cancer and obtained enough information about specific drugs to conclude that the risk increase was confined to use of medications that operate through dopaminergic mechanisms or GABAergic pathways as opposed to serotoninergic pathways. Two later case-control studies (Coogan et al., 2000a; Dublin et al., 2002) did not confirm that antidepressant drugs were associated with ovarian cancer. The incidence of all cancers combined was not increased among users of several classes of antidepressant drugs in a large cohort study in Denmark (Dalton et al., 2000). In site-specific analyses, an association was observed between use of either tricyclic or tetracyclic antidepressants and non-Hodgkin’s lymphoma. Further negative evidence came from a nested case-control study in a cohort of cancer patients.

494

PART III: THE CAUSES OF CANCER

Use of antidepressant drugs was not associated with either recurrences of the cancer or development of second primary cancers (Weiss et al., 1998). Risk of malignant melanoma was not associated with use of either tricyclic or tetracyclic antidepressants in a case-control study in Sweden (Westerdahl et al., 1996). To sum up, there has been limited and inconsistent evidence that antidepressant drugs might be carcinogenic. No conclusions can be drawn at present, and several distinct classes of drugs are involved. If one or more classes of these drugs is later judged to be a causal factor for cancer, this does not necessarily incriminate the other classes, which differ in both form and mechanism of action.

ANTILIPEMICS During the past few decades, there has been a dramatic increase in prescriptions for lipid-lowering drugs (Newman and Hulley, 1996) and they are now among the most commonly prescribed medications in the United States (NDC Health, 2003). Hyperlipidemia is often a chronic condition, and individuals may be treated with one or more of these medications for 20 to 30 years or longer. The potential association between lipid levels and cancer risk has received some attention. Reports of an association between low cholesterol levels and cancer incidence or mortality have largely been attributed to cholesterol levels dropping due to preexisting cancer, as associations have often disappeared or have been attenuated when analyses were lagged to exclude cancer events within 1–2 years after the cholesterol measurement (Law et al., 1994; Eichholzer et al., 2000). In a few studies, high-density lipoprotein (HDL) levels have been positively associated with cancer risk (Boyd and McGuire, 1990). Confounding by indication may therefore be a possibility when conducting observational studies of lipid-lowering drugs and cancer risk.

Fibric Acid Derivatives Clofibrate and gemfibrozil are members of a class of lipid-lowering drugs derived from fibric acids. Fibrates lower the levels of triglyceride-rich lipoproteins, modestly raise high-density lipoprotein (HDL) levels, and have various effects on levels of low-density lipoprotein (LDL), depending on the specific drug used and a number of host and environmental factors (Witztum, 1996). Clofibrate was approved for use in the United States in1967 (IARC, 1996). It was one of the most frequently prescribed lipid-lowering drugs for a number of years until results of a WHO-sponsored cooperative clinical trial published in 1978 suggested that its use was associated with an increase in noncardiovascular disease and overall mortality (IARC, 1996). There was also a non-significant increase in deaths due to cancer. Risk of cancer was subsequently examined among participants of several clinical trials (including extended follow-up of the WHO study), a small case-control study, and a meta-analysis of six clinical trials (IARC, 1996). An excess in cancer risk was not observed among participants of the clinical trials. A nonsignificant excess of soft-tissue sarcoma was observed in the casecontrol study. Gemfibrozil was first marketed in the United States in 1982 and it continues to be a commonly used lipid-lowering medication. The results of two clinical trials, each with approximately 5 years of follow-up, showed no increase in cancer events in the gemfibroziltreated arms (IARC, 1996; Rubins et al., 1999). Although available human data are reassuring, results of animal studies provide some reason for concern (Newman and Hulley, 1996). Clofibrate and gemfibrozil increase liver tumors in some rodent systems (IARC, 1996; Newman and Hulley, 1996), although it is not clear whether the mechanism of action would be operative in humans (IARC, 1996). The most recently conducted IARC evaluations concluded that there was insufficient or limited evidence to classify these medications as carcinogenic in animals or humans (IARC, 1996). Additional, especially longer-term, follow-up studies of fibrate-treated patients are needed.

Statins Earlier rodent carcinogenicity studies suggested that a relatively new class of lipid-lowering drugs, 3-hydroxy-3-methylglutaryl-coezyme A (HMG CoA) reductase inhibitors, or statins, might be carcinogenic in humans (Newman and Hulley, 1996). However, the results of more recent animal and in vitro studies indicating that various statins also have anticancer activity (e.g., antiproliferative, antiangiogenic, proapoptotic) have generated a growing interest in their potential use in cancer management and prevention (Brower, 2003; Chan et al., 2003). The limited data from clinical trials of statin therapy, primarily among patients with prior heart disease and/or hypercholesterolemia, are conflicting. Non-significant increases or decreases in overall cancer and cancer at various sites have been reported (Pedersen et al., 1996; Sacks et al., 1996; Hebert et al., 1997; Pedersen et al., 2000; Downs et al., 2001; Bjerre and LeLorier, 2001; Heart Protection Study Collaborative Group, 2002; Pfeffer et al., 2002); a significant increase in risk of breast cancer was reported in one study (Sacks et al., 1996). Small numbers of cancer events in these trials have limited their power to detect modest or even moderate effects on cancer incidence. Three cohort studies using pharmacy registries have reported on the association between statins and cancer risk (Olsen et al., 1999; Blais et al., 2000; Beck et al., 2003). As with the clinical trials, two of these cohort studies (Blais et al., 2000; Olsen et al., 1999) had small numbers of cancer at any site and limited power to detect increases or decreases in risk. Information on potentially confounding factors, other than age and sex, was also limited. The first study (Olsen et al., 1999) found no association, but had only 41 cancers in the statin group. A second study (Blais et al., 2000) was able to at least in part control for confounding by indication by comparing risk of cancer in individuals treated with statins to individuals treated with other lipidlowering drugs. They found a significantly reduced risk of any cancer among statin users compared to users of bile acid–binding resins. Although not statistically significant, they also found a reduced risk of cancer of the skin, prostate, lung, breast, colon, bladder and kidney, and uterus and an increased risk of lymphoma. A third cohort study (Beck et al., 2003) compared risk of breast cancer among 13,500 statin users to that among 53,880 nonusers of lipid-lowering drugs. While there was a slight and non-significant increase in risk shortly after initiating statin therapy, 4 years after initiation risk was significantly decreased. Risk of breast cancer has also been examined in two case-control studies (NDC Health, 2003). In a hospital-based study (Coogan et al., 2002), risk of breast cancer in statin users was compared to risk in nonusers of statins. Controls were hospitalized patients with admissions unrelated to use of lipid-lowering drugs. In this study, risk of breast cancer was not associated with recent use or use of less than 3 years duration. Women who had used statins for 3 or more years were at approximately twice the risk of breast cancer (RR = 2.1, 95% CI 1.1–4.0), although this was largely due to risk of in situ (RR = 3.4, 95% CI 1.5–8.0) and not invasive (RR = 1.5, 95% CI 0.7–3.1) cancer. In a second case-control study (Kaye et al., 2002), risk of breast cancer in current and past statin users was not statistically different from that among women with no diagnosis of hyperlipidemia and no prescriptions for a lipid-lowering medication. Because lipid-lowering medications are so commonly prescribed, the human data are still so limited and inconsistent, and the animal and in vitro data are beginning to suggest both chemopreventive and carcinogenic activities, additional studies are needed.

ANTINEOPLASTICS Tremendous progress has been made in cancer therapy, and the nature of and basic approaches to treating cancer are evolving at a rapid pace. Antineoplastic drugs comprise a variety of agents. Major categories include alkylating agents, antimetabolites, antitumor antibiotics, vinca alkaloids and epipodophyllotoxins, platinum coordination complexes, biological response modifiers, and hormones and antagonists.

Pharmaceuticals Other Than Hormones Many antineoplastic drugs are cytotoxic and, unfortunately, several are known or suspected to greatly increase the risk of second cancers. Evaluation of the carcinogenic effects of these agents is difficult because the cancer being treated may itself be associated with subsequent new malignancies and because of the administration to a patient of more than one such drug plus, in many cases, radiation therapy. The complexity of the problem is illustrated by so-called MOPP therapy for Hodgkin disease. MOPP is a combination of (1) nitrogen mustard, an alkylating agent that damages DNA and prevents cell replication, (2) vincristine, a plant alkaloid that inhibits mitosis, (3) prednisone, an adrenal corticosteroid that damages lymphoid cells and interferes with their proliferation, and (4) procarbazine, a synthetic drug that inhibits DNA, RNA, and subsequent protein synthesis. Both nitrogen mustard and procarbazine have been shown to be carcinogenic in experimental animals, but neither has been studied epidemiologically as a single agent (IARC, 1987). MOPP therapy was introduced in 1967 and was found, along with radiation therapy alone and MOPP plus radiation, to cure 70% of all cases of Hodgkin disease. This resulted in many long-term survivors of a disease that had usually been fatal within a few years (Blayney et al., 1987). Several studies have shown a markedly increased risk of acute nonlymphocytic leukemia (ANLL) in persons treated with MOPP with or without radiation therapy, and a lesser and possibly nonelevated risk after radiation therapy alone (IARC, 1987). In a study of long-term survivors, there appeared to be a window of increased ANLL risk with peak incidence between 3 and 9 years after MOPP; those who survived at least 11 years appeared to be at no increased risk (Blayney et al., 1987). At present, it is not known whether one or more components or MOPP as a whole is responsible for the heightened risk of ANLL or how much is attributable to radiation therapy and, perhaps, other factors associated with Hodgkin disease itself. Also observed has been an excess of solid tumors among Hodgkin disease survivors, but the causes of this are obscure. Complicating the interpretation of the many studies that have been done are differences in the patients investigated, the methods and timing of MOPP administration, and the study criteria and methods, all of which could have contributed to the wide variation in the reported risks of ANLL and other cancers (IARC, 1987). The alkylating agents make up the largest group of nonhormonal drugs for which IARC considers there is sufficient evidence for carcinogenic effects in humans (see Table 25–1). These agents include busulfan, chlorambucil, melphalan, thiotepa, treosulphan, methylCCNU, and cyclophosphamide, all of which can cause ANLL, and the last of which can also cause bladder cancer. Chlornaphazine, which is no longer used, also increases the risk of bladder cancer (IARC, 1987). Etoposide, and the less widely used teniposide, are DNA topoisomerase II inhibitors. There is sufficient human evidence for the carcinogenicity of etoposide when combined with cisplatin and bleomycin (IARC, 2000). Although there is growing evidence that these drugs, especially etoposide, are carcinogenic when used alone, they are usually given in combination with other cytoxic therapies, making this difficult to establish for the reasons noted above. In addition to the demonstrated risks of cytotoxic drugs to patients, there has also been concern about occupational exposures among oncologists, nurses, pharmacists, and other health care workers who handle them. The Occupational Health and Safety Administration (OSHA) has published guidelines for controlling occupational exposure to these medications (OSHA, 1999). The risks for these personnel are not well defined, but prudence dictates that exposure be minimized.

ANTIFUNGALS, ANTIPROTOZOANS, ANTIPARASITICS Metronidazole Metronidazole is an antiprotozoal and antibacterial drug that is used to treat a variety of infections, mostly of the vagina and gastrointestinal tract. Because of its apparent carcinogenicity in rats and mice (IARC, 1987) there has been concern that it might also have this effect in humans (Bendesky et al., 2002). Follow-up studies have been performed in three settings. In one study, 771 women treated at the Mayo

495

Clinic with metronidazole for vaginal trichomoniasis showed statistically significant elevations in the incidence of lung cancer and, subsequently, of all cancers combined after 15 to 25 years of follow-up (Beard et al., 1979, 1988). In another study of 2236 female and 224 male patients who received metronidazole from a Kaiser Permanente Medical Care Program pharmacy, there was no significant increase in either total cancer or lung cancer incidence after 11 to 15 years of follow-up. An increase in cervical cancer was believed to be due to sexual behavior that led also to vaginitis treated with metronidazole (Friedman, 1980; Friedman and Selby, 1989; Van Den Eeden and Friedman, 1995). A third study group of 12,280 patients who received the drug at the Group Health Cooperative of Puget Sound (GHC) were followed for up to 21/2 years and showed no increase over expected cancer incidence (Danielson et al., 1982a). Another study in the same setting, which focused on breast cancer, showed a nonsignificantly elevated risk ratio of 1.1 (Danielson et al., 1982b). Longer follow-up— median 12.6 years—was available in still another study based in GHC. Five thousand two hundred twenty-two metronidazole users and 5222 matched nonusers showed a similar incidence of cancer. However, the incidence of solid-organ cancers was relatively greater in the metronidazole-exposed group if attention was restricted to those followed up for at least 12 and at least 15 years, but the differences were not statistically significant (Falagas et al., 1998). A retrospective cohort study evaluated the risk of cancer in children under age 5 years in relation to their mothers’ prenatal use of metronidazole (Thapa et al., 1998). No statistically significant associations were found, but the risk ratio for neuroblastoma was 2.6 (95% CI 0.89–7.59). There is still insufficient evidence to demonstrate that metronidazole is carcinogenic in humans or to be sure that it is safe in this regard. If there is an added risk, it is probably small. Longer follow-up and more studies are both needed.

Dapsone Dapsone has been used for the treatment of leprosy, malaria, and some other diseases. Although this drug has shown some evidence of carcinogenicity in rats, the findings in humans, based on the follow-up of patients treated for leprosy and Australian military personnel who received dapsone in Vietnam to prevent malaria, have been largely negative (Christie, 1993; Stolley and Zahm, 1995).

ANTITHYROID DRUGS The thionamides (methimazole, methylthiouracil, propylthiouracil, and thiouracil) are antithyroid drugs used to treat hyperthyroidism caused by Graves disease, toxic thyroid nodules, toxic multinodular goiter, and several other rare causes of hyperthyroidism (IARC, 2001). They are potent inhibitors of thyroid hormone synthesis. Along with a number of other chemicals, the thionamides have been found to produce thyroid tumors in rodents. Although there is sufficient evidence for the carcinogenicity of several of these medications (methylthiouracil, propylthiouracil, and thiouracil) in experimental animals, there are very few data on their potential carcinogenicity in humans (IARC, 2001). An unspecified excess of thyroid cancer deaths among patients treated with antithyroid medications was reported in a follow-up study of approximately 35,000 patients treated in the United States and United Kingdom for adult hyperthyroidism (Dobyns et al., 1974). The antithyroid drugs used were not specified, but thionamides were the drugs primarily used during this period (IARC, 2001). In a subsequent analysis based on additional follow-up, standardized mortality ratios were elevated for all cancers (SMR = 1.3, 95% CI 1.1–1.6) among patients treated exclusively with antithyroid drugs, primarily due to excesses in deaths from oral cancer and brain tumors (Ron et al., 1998). Radioactive iodine, the most common therapy for hyperthyroidism, was not associated with an increase in total cancer deaths; however, there was an excess of deaths due to thyroid cancer (SMR = 3.9; 95% CI 2.52–5.86). In the Kaiser Permanente study

496

PART III: THE CAUSES OF CANCER

of prescription drugs, there were 107 recipients of propylthiouracil, with subjects averaging more than 4 prescriptions each. After 11 to 15 years follow-up, no significant excess of cancer has been noted for any site (Selby et al., 1989). One thyroid cancer was observed (compared to 0.07 expected), but this cancer was diagnosed less than 1 year after the first prescription of propylthiouracil and may have been present before treatment began. With up to 21 years of follow-up, 9 propylthiouracil recipients have developed cancer, as compared with 11.8 expected (Friedman and Habel, unpublished data). Further study of the possible carcinogenicity of anti-thyroid drugs in humans is needed.

DERMATOLOGIC AGENTS Arsenicals Arsenic and compounds containing it have been associated with increased cancer risk in humans exposed in three different ways: occupationally, in drinking water, and in drugs (IARC, 1987). Arsenic may also be ingested in foods, where it may occur as pesticide residues on fruits and vegetables, or be derived from feed additives for livestock or poultry. It is also concentrated from environmental sources in some species of fish and shellfish (Klaassen, 1996b). Medicinal arsenic was formerly given orally for psoriasis, asthma, and a variety of other conditions in the form of inorganic trivalent compounds such as Fowler’s solution (potassium arsenite). As such, it has been shown to cause skin cancer (Cuzick et al., 1982). Suspected associations of other types of cancer with these drugs, based on case reports, have not been confirmed epidemiologically. These compounds are now rarely used. At present, medicinal arsenic is administered chiefly in the form of the organic compound, melarsoprol, for therapy of advanced stages of African trypanosomiasis, a tropical parasitic disease (Pepin and Milord, 1994). There is now renewed interest in arsenic trioxide as a potentially effective treatment of hematologic and lymphatic malignancies and some solid tumors (Murgo, 2001; Chen et al., 2002; Miller et al., 2002). These arsenical drugs have not been studied in relation to risk of subsequent cancer in humans, but arsenic trioxide is carcinogenic in other settings (IARC, 1987).

Coal Tars Occupational exposure to coal tar has been shown to cause cancer of the skin, including the scrotum, and lungs, and possibly of the urinary and digestive tracts as well (IARC, 1987). Coal tar in ointments has been applied to the skin in the treatment of psoriasis and eczematous dermatitis. Thus, concern has been raised about its possible carcinogenicity when used medicinally in this way. Sometimes its use is combined with ultraviolet radiation in what is known as the Goeckerman regimen (Pittelkow et al., 1981). Epidemiological data regarding cancer risk after coal tar ointment application have been inconclusive. In one case-control study, the risk of skin cancer after the Goeckerman regimen was increased 4.7-fold over that in matched psoriasis patients (Stern et al., 1980). However, in two 25-year follow-up studies, the observed increase in risk was much smaller, and not apparently greater than that in age-sex matched residents of the Dallas-Fort Worth, Texas area (Maughan et al., 1980; Pittelkow et al., 1981). Another follow-up study found a similar attack rate of skin cancer before and after the Goeckerman regimen was used (Menter and Cram, 1983). No skin cancers were observed in 84 Japanese patients with psoriasis who received the Goeckerman regimen, either alone or with PUVA therapy (psoralens plus long-wave ultraviolet light—see below) at other times. However, only 17 were followed up for more than 6 years (Torinuki and Tagami, 1988). Coal tar compounds are absorbed through the skin and can produce DNA adducts in human white blood cells as well as skin (Godschalk et al., 1998). In weighing the advantages and disadvantages of topical coal tar versus other dermatologic preparations, the possible small added risk of skin cancer should be considered and appropriate follow-up instituted if coal tar is selected.

Psoralens Methoxsalen (8-methoxypsoralen, or 8-MOP) is a photosensitizing agent, which is sometimes used with long-wave ultraviolet radiation in the treatment of severe psoriasis, vitiligo, and mycosis fungoides. It may be applied topically or given orally if large areas of skin are to be treated. Although methoxsalen alone has not been shown to cause cancer in humans or animals, the combination of the drug plus long-wave ultraviolet light (often abbreviated as PUVA) has been found to be carcinogenic for human and mouse skin (IARC, 1987). A number of studies have reported an increased incidence of squamous cell carcinoma of the skin among psoriatic patients treated with PUVA, with persons exposed to high doses having a several-fold higher risk than the general population (Stern et al., 1984; Bruynzeel et al., 1991; Lindelof et al., 1991). A much smaller increase has been reported for basal cell carcinoma (Stern et al., 1984; Bruynzeel et al., 1991). An increase in melanoma risk was recently observed among one of these cohorts of patients (Lindelof et al., 1999), although the number of melanoma cases was small (N = 4). The evidence to date thus dictates both caution in the use of PUVA and careful follow-up after it is used. A related compound, 5-methoxypsoralen (5-MOP), was used in sunscreens in France, Belgium, and Greece, with the objective of increasing tanning by admitting longer ultraviolet waves while protecting against radiation-induced skin damage by blocking shorter ultraviolet waves (Cartwright and Walter, 1983). Animal data indicate that the combination of ultraviolet radiation and 5-MOP increases risk of skin cancer in mice (IARC, 1986, 1987). Although there are few data in humans, sunscreens with 5-MOP were banned by the European Commission in 1995, after an epidemiologic study (Autier et al., 1995) reported an increased risk of melanoma among subjects who used these sunscreens, especially if they were naturally sun sensitive (Autier et al., 2000).

DRUGS FOR TREATING ANEMIA Iron Iron is a dietary element essential for the synthesis and function of hemoglobin, myoglobin, and intracellular enzymes such as cytochromes and other oxidases. Iron deficiency, as a result of dietary deficiency, excessive menstrual blood loss, or acute or chronic bleeding, is a frequently seen medical condition. Medicinal iron, in the form of iron dextran injections, or more frequently as orally administered ferrous salts, is commonly used to treat this condition. Reports of soft tissue tumors developing at the site of iron dextran injections appeared as early as 1960 (Robinson et al., 1960). However, no epidemiologic studies of this association have appeared to date, and the observations may well represent coincidence given the widespread use of this drug. Repeated injections of doses of iron dextran much larger than those used clinically have also induced formation of fibrosarcomas in some animal species, but not in others (Richmond, 1959). Concern that excessive body iron stores could lead to increased risk of both infection and cancer has been raised on the basis of biological considerations (Weinberg, 1984). Both tumor cells and bacteria require iron for cell growth. Pathogenic bacteria and neoplastic cells are specifically adapted to enhance iron acquisition from the host (Nielands, 1980; Weinberg, 1984). The hypoferremia and iron sequestration by the reticuloendothelial system, which are components of the host response to both infection and neoplasia, may be a defense that withholds iron from invading cells (Roeser, 1980; Weinberg, 1984). Iron overload, even in small amounts, enhances chemically induced tumor promotion in experiments on the skin of mice (Bhasin et al., 2003). Epidemiologic studies (Stevens et al., 1986; Selby and Friedman, 1988; Stevens et al., 1994; Knekt et al., 1994; Herrinton et al., 1995) have examined the association of iron stores with risk of cancer. Although there have been inconsistent associations observed with

Pharmaceuticals Other Than Hormones cancer of several particular sites (lung, stomach, colon and rectum, and hematologic malignancies), it is not clear that there is a general relation with cancer per se. Persons with hereditary hemochromatosis experience a markedly increased incidence of hepatocellular carcinoma and possible lesser increases in other cancers, including colon cancer and childhood leukemia. This has been attributed to the effects of the excess stored iron in hemochromatosis rather than to other effects of the hemochromatosis gene mutation (Dorak et al., 2002; Shaheen et al., 2003).

DRUGS FOR TREATING CARDIOVASCULAR DISORDERS Diuretics In the late 1980s and early 1990s, cohort (Fraser et al., 1990; Lindblad et al., 1993) and case-control (Yu et al., 1986; McLaughlin et al., 1988; Finkle et al., 1993; Kreiger et al., 1993; Hiatt et al., 1994) studies found a link between diuretic or antihypertensive use and subsequent risk of renal cell carcinoma in women, with ever users being at two- to fourfold increased risk. The relation existed both in hypertensives and in women who used diuretics for weight control. One study (Hiatt et al., 1994) specifically examined thiazides, noting an odds ratio of 4.0 (95% CI 1.5–11). A second study (Finkle et al., 1993) excluded women who had been diagnosed with the cancer within the 10 years following first use of diuretics, observing an odds ratio of 3.5 (95% CI 1.7–7.4). There was also evidence that risk increases with increasing cumulative dose (Finkle et al., 1993; Hiatt et al., 1994). Another study (Weinmann et al., 1994) found a slightly stronger association in men than in women, but the evidence in men has been inconsistent. Studies suggesting this association continued to appear (summarized in Felmeden and Lip [2001]). Meta-analysis showed about a twofold increased risk in three cohort studies and about a 1.5-fold increased risk in nine case-control studies (Grossman et al., 1999). Some recent evidence has pointed to renal cell carcinoma being primarily associated with the underlying condition, hypertension, rather than with diuretic therapy used to treat it (Yuan et al., 1998; Shapiro et al., 1999). Another underlying risk factor for renal cell carcinoma is obesity (Yuan et al., 1998), which predisposes to hypertension. Further complicating the picture, associations of diuretic therapy have been reported with cancer of the renal pelvis and ureter (Liaw et al., 1997) and with colon cancer (Tenenbaum et al., 2001), the latter being restricted to nonusers of aspirin. The possible biological basis of carcinogenic effects of diuretics and/or hypertension, if such effects exist, has not been elucidated. The situation is still not clear with regard to diuretics but the question is important for public health, as hypertension is common and diuretics are an effective and commonly used treatment for this condition. The data from a large three-arm randomized trial of antihypertensive drugs in Sweden should be somewhat reassuring in that patients who received “conventional drugs” (diuretics or betablockers) and followed for a median 5.3 years developed cancer at a rate statistically indistinguishable from that observed in those who received calcium channel blockers or angiotensin converting enzyme (ACE) inhibitors, or from the cancer incidence in the general population (Lindholm et al., 2001).

Reserpine There has been considerable interest in the possibility that reserpine (or rauwolfia), an antihypertensive drug, may cause breast cancer in women. About 20 epidemiological studies have yielded conflicting results, with relative risk estimates ranging from 0.6 to over 3 (Danielson et al., 1982a; Danielson et al., 1982b; Shapiro et al., 1984; Stanford et al., 1986; IARC, 1987). Although the studies that are methodologically most sound tend to show little if any increased risk (IARC, 1987), there has been some evidence that a larger risk may be found among women using the drug for long durations. One study sug-

497

gested that previous contradictory findings could be explained by the hypothesis that reserpine use for at least five years causes breast cancer occurring after age 50 years (Williams et al., 1978), but another study aimed at this specific hypothesis (Friedman, 1983) did not confirm it. A large case-control study found evidence for a moderately strong association of reserpine with breast cancer only for women who had taken the drug for at least 10 years, or where there had been a latency interval from first use to diagnosis of cancer of 10 years or more (Stanford et al., 1986). Animal experiments have given mixed results with limited evidence for carcinogenicity in rats and mice (IARC, 1987). A proposed mechanism linking reserpine to breast cancer is the drug’s stimulation of prolactin secretion. The degree of elevation of prolactin levels in women taking reserpine was found to be consistent with only small increases in breast cancer risk as judged by a postulated statistical model of breast cancer incidence (Ross et al., 1984) Genotoxic and mutagenic effects appear to be absent (von Poser et al., 1990; Tsutsui et al., 1994), but morphological changes in Syrian hamster embryo cells in culture, induced by reserpine, resembled those produced by known carcinogens (Tsutsui et al., 1994). The overall body of evidence to date suggests that reserpine may be used to treat hypertension for several years with little if any added risk of breast cancer. Furthermore, lower doses than were used previously may be sufficient for effective therapy (Fraser, 1996). Nevertheless, caution is still appropriate with regard to long-term use, pending additional studies.

Calcium Channel Blockers Calcium channel blockers (CCBs) represent a chemically and pharmacologically diverse group of agents that have been commonly used since the 1980s for the treatment of hypertension and angina (Robertson and Robertson, 1996; Mason, 1999). A report in 1996 (Pahor et al., 1996) raised concern that use of CCBs may be associated with an increased risk of cancer. In this interview-based prospective cohort study of 5052 elderly individuals 71 years and older (medication use [current] obtained at baseline, only), cancer incidence among those taking CCBs was 1.72 times higher (95% CI 1.27–2.34) than the incidence among nonusers of CCBs, after adjusting for several confounding factors. Increasing daily dose was associated with increasing risk. However, the number of events was small; only 47 cancer cases occurred among the CCB users. Risk of cancer associated with use of CCBs was subsequently examined in a number of studies, including several clinical trials (Borhani et al., 1996; Staessen et al., 1997; Kanamasa et al., 1999; Sajadieh et al., 1999), and case-control (Jick et al., 1997; Rosenberg et al., 1998a; Vaughan et al., 1998; Vezina et al., 1998) and cohort (Zacharski et al., 1990; Olsen et al., 1997; Fitzpatrick et al., 1997; Hole et al., 1998; Jonas et al., 1998; Michels et al., 1998; Trenkwalder et al., 1998; Cohen et al., 2000; Sorensen et al., 2000) studies. Although the majority of studies examined overall cancer incidence (Pahor et al., 1996; Jick et al., 1997; Olsen et al., 1997; Braun et al., 1998; Hole et al., 1998; Jonas et al., 1998; Michels et al., 1998; Rosenberg et al., 1998a; Trenkwalder et al., 1998), some of these studies also presented results for several individual cancer sites (Pahor et al., 1996; Olsen et al., 1997; Jick et al., 1997; Hole et al., 1998; Michels et al., 1998; Rosenberg et al., 1998a). Other studies focused on a single site such as breast (Fitzpatrick et al., 1997) or on related sites such as esophageal and gastric cancer (Vaughan et al., 1998). Most of the clinical trials and cohort studies had relatively few cases of cancer at any given site. Few studies collected detailed information on dose or duration of CCB use, involved long-term follow-up, or could adjust for confounding factors other than age and sex. In a cohort of women aged 65–100 participating in the Cardiovascular Health Study (Fitzpatrick et al., 1997) with information on multiple possible confounders, risk of breast cancer among users of CCBs was higher than risk among nonusers of CCBs (adjusted RR = 2.57, 95% CI 1.47–4.49) and higher than risk among users of other antihypertensive medications (diuretics, ACE inhibitors, beta-blockers) (adjusted RR = 2.91, 95% CI 1.41–6.00). In addition, there was a

498

PART III: THE CAUSES OF CANCER

suggestion that the association was strongest among women above the modal dose and also among women taking hormone replacement therapy. No association with breast cancer risk was observed in one hospital-based case-control study with information on lifetime medication use and multiple potential confounders (Rosenberg et al., 1998a), in an analysis of women participating in the Nurses Health Study (Michels et al., 1998), or in a record linkage study conducted in Denmark (Olsen et al., 1997). Weak to modest, and non-significant, increases in breast cancer risk were observed in three cohort studies that examined cancer at multiple sites (Pahor et al., 1996; Jick et al., 1997; Hole et al., 1998), two of which had some information on potential confounders. Only results from one (Pahor et al., 1996) of the six studies (Pahor et al., 1996; Jick et al., 1997; Olsen et al., 1997; Rosenberg et al., 1998a; Vezina et al., 1998; Braun et al., 1998) reporting on risk of prostate cancer suggested a modest increase in risk. Similarly, the results of only one (Jick et al., 1997) of six studies (Pahor et al., 1996; Jick et al., 1997; Olsen et al., 1997; Braun et al., 1998; Rosenberg et al., 1998a; Vezina et al., 1998) reporting on lung cancer suggested an increase in risk associated with use of CCBs. Other results on cancer risk associated with CCB use mostly have been null. Inhibition of apoptosis was the first hypothesized mechanism by which CCBs might increase cancer risk (Pahor et al., 1996). Although high doses of CCBs have been shown to inhibit apoptosis in some in vitro systems, the effect of CCBs on apoptosis appears to be complex (Mason, 1999). CCBs have also been shown to promote apoptosis in transformed cell lines and results of animal studies generally do not suggest increased tumor growth (Ahr et al., 1998). Therefore, while some laboratory and epidemiologic studies suggest that use of CCBs may increase the risk of cancer, the majority of studies do not support the carcinogenicity of these medications.

ACE Inhibitors Some of the observational studies comparing risk of cancer in users of CCBs to risk in users of other antihypertensive medications found that risk was lower in users of ACE inhibitors (Pahor et al., 1996; Jick et al., 1997; Lever et al., 1998; Hole et al., 1998). This, along with some in vitro and animal data suggesting that ACE inhibitors may retard cancer cell growth and inhibit angiogenesis, led some investigators to hypothesize that ACE inhibitors may be chemopreventive (Lever et al., 1998; Friis et al., 2001). One study compared the observed cancer incidence among 17,897 Danish individuals prescribed ACE inhibitors to expected rates based on county-specific rates and to cancer incidence among 47,579 individuals prescribed either beta-blockers or CCBs (Friis et al., 2001). Information on medications and cancer incidence was obtained from a population-based prescription database and cancer registry, respectively. Overall, observed rates were similar to those expected. This is consistent with several other observational studies (Fitzpatrick et al., 1997; Rosenberg et al., 1998a; Meier et al., 2000) and one clinical trial (Lindholm et al., 2001) that found no difference in risk between users of CCBs and users of ACE inhibitors, and/or found no increase or decrease in cancer risk when comparing users of ACE inhibitors to non-users of ACE inhibitors. However, in a population-based casecontrol study, risk of prostate cancer was increased modestly (RR = 1.5, 95% CI 1.0–2.2) among users of ACE inhibitors compared to risk among nonusers of any antihypertensive medication (Vezina et al., 1998). In summary, available epidemiologic data generally do not support a chemopreventive effect of ACE inhibitors.

DRUGS FOR TREATING CENTRAL NERVOUS SYSTEM DISORDERS Phenobarbital Barbiturates such as phenobarbital, secobarbital, and pentobarbital have been widely used since 1920 as sedatives and hypnotics and for treatment of epilepsy. The use of barbiturates has decreased since the 1960s, but they are still used extensively (IARC, 2001).

Barbiturates are potent inducers of hepatic microsomal enzymes known as mixed function oxidases, which are responsible for the metabolic activation as well as detoxification of many chemical carcinogens (Meyer et al., 1980). CYP2B1 and CYP2B2 are the primary P450 enzymes induced by phenobarbital, although the activities of several other CYP enzymes are also induced, including benzo[a]pyrene hydroxylase, UDP-glucuronosyl transferase, and several glutathioneS-transferases (IARC, 2001). Phenobarbital stimulates cell proliferation of normal hepatocytes and inhibits their intercellular communication (IARC, 2001). In mice and rats, it is also a known promoter of liver tumors after exposure to initiating carcinogens (Peraino et al., 1971; IARC, 2001). Phenobarbital has been found to promote liver cancer in monkeys and thyroid tumors in mice and rats. Phenobarbital does not appear to be genotoxic (IARC, 2001). An increased occurrence of hepatic cancer (11 observed cases versus 2.8 expected) was noted in one large follow-up study of persons treated with phenobarbital for epilepsy (Olsen et al., 1989). However, eight of the 11 cases had also been exposed to Thorotrast (a diagnostic radiopharmaceutical known to be carcinogenic) in the workup of their epilepsy. Other follow-up studies of persons with epilepsy (White et al., 1979; Shirts et al., 1986) found no excess of hepatic cancer. A greater than expected occurrence of brain tumors in these followup studies (White et al., 1979; Shirts et al., 1986; Olsen et al., 1989) was most likely due to use of anticonvulsants for early symptoms of preexistent brain tumors. The incidence of the tumors was increased mainly during the early years of treatment, reverting towards normal after 10 years of use. This was the opposite of what would be expected if exposure to phenobarbital were causally related to brain tumor. No increase in risk of brain tumors was observed among 5834 recipients of phenobarbital prescriptions in the Kaiser Permanente study of prescription drugs (Selby et al., 1989), most of whom received phenobarbital as a sedative rather than an anticonvulsant. In a case-control study, children with prenatal exposure to barbiturates had an excess of brain tumors (OR = 1.5, P = 0.03) (Gold et al., 1978). However, the association was based on only six discordant case-control pairs and was noted only when cases were compared to individuals with other cancers, not when compared with noncancer controls. An excess in risk was not observed in two other studies of children exposed in utero (Heinonen et al., 1977; Goldhaber et al., 1990). In addition, no increased risk of brain tumors was observed among 3237 offspring of women with epilepsy (Olsen et al., 1990). A 1.7-fold increase in the incidence of lung cancer was observed among users of three barbiturate preparations in the Kaiser Permanente study of prescription drugs (Friedman and Ury, 1980, 1983). The association persisted after adjustment for cigarette smoking (Friedman, 1981) and was seen among nonsmokers as well as smokers, although the association in nonsmokers was based on only four observed cases (vs. 2.7 expected). A subsequent analysis with an additional 16 years of follow-up produced similar results (Friedman and Habel, 1999). In this study, with a larger number of cancer cases, there was a suggestion of a dose-relationship and the increase in risk diminished over time. Increases of similar size in the incidence of lung cancer were noted in the three follow-up studies of persons treated for epilepsy (White et al., 1979; Shirts et al., 1986; Olsen et al., 1989). One of these cohorts formed the basis for a nested case-control study in which more detailed information about drug use, smoking history, and exposure to Thorotrast was obtained (Olsen et al., 1993). The earlier-observed association with lung cancer was appreciably lower after adjustment for smoking (OR = 1.2, 95% CI 0.7–2.2). The association of a history of use of barbiturates with lung cancer risk merits further investigation. Increasing doses of phenobarbital use have also been associated with a decrease in the occurrence of bladder cancer, such that persons whose cumulative dose was 720 g or greater were at one-fifth the risk of nonusers (95% CI 0.0–0.9) (Olsen et al., 1989; Olsen et al., 1993). The authors hypothesized that phenobarbital induces liver enzymes that deactivate bladder carcinogens. This hypothesis was supported by a small study that observed an inverse association between barbiturate

Pharmaceuticals Other Than Hormones use and risk of bladder cancer among current and former smokers, but not among never-smokers (Habel et al., 1998). In contrast, a recent case-control study of bladder cancer found no decrease in risk associated with phenobarbital use in the full study population, or among ever smokers (Castelao et al., 2003). However, there was a suggestion that subjects in the highest category of phenobarbital exposure were at increased risk of bladder cancer (OR = 2.5; 95% CI 0.9–6.8). In this study, only 21 cases and 15 controls were regular users of phenobarbital. Thus, although the data are considered sufficient to classify phenobarbital as an animal carcinogen (IARC, 2001) and there are other biological links that could explain a carcinogenic effect at some sites and a chemopreventive effect at others, the data from humans remain inconclusive.

Phenytoin (Diphenylhydantoin) Diphenylhydantoin, an anticonvulsant drug, can cause immune dysfunction in animals and man (Sorrell and Forbes, 1975), and the development of pseudolymphoma (Saltzstein and Ackerman, 1959; Gams et al., 1968), a hypersensitivity reaction that includes generalized lymphadenopathy and usually regresses on discontinuation of the medication. This drug has been shown to have liver tumor promoting activity in mice (Dethloff et al., 1996). Although no association has been noted with liver cancer in humans, numerous cases of Hodgkin and non-Hodgkin’s lymphoma have been reported in persons taking diphenylhydantoin (Rubinstein et al., 1985; Hyman and Sommers, 1966), sometimes preceded by, and arising in, pseudolymphoma. Occasional cases of leukemia (Gyte et al., 1985) and multiple myeloma (Matzner and Polliack, 1978) have also been noted. Two case-control studies (Charlton and Lunsford, 1971; Li et al., 1975) found increased frequencies of prior exposure to diphenylhydantoin in cases of Hodgkin or non-Hodgkin’s lymphoma, but each study was based on a very small number of exposed cases. Longitudinal studies have been inconclusive. In three cohorts of persons with epilepsy (White et al., 1979; Shirts et al., 1986; Olsen et al., 1989), incidence or mortality from lymphoma after 13 to 30 years of follow-up was slightly increased compared to general population rates. These studies could not distinguish phenytoin users from the small proportion of subjects who did not take the drug. In the Kaiser Permanente study of carcinogenicity of prescription drugs (Van Den Eeden and Friedman, 1995), incidence of non-Hodgkin’s lymphoma after 11 to 15 years of follow-up was increased among the 954 users of diphenylhydantoin (2 observed cases, 0.64 expected). However, for all hematologic malignancies, a slight deficit was noted (4 observed vs. 4.56 expected). Neither difference was statistically significant. Sixteen additional years of follow-up revealed no statistically significant associations with lymphomas or leukemias and statistically significant (P < 0.05) deficits in skin melanoma (0 cases observed, 3.75 expected) and prostate cancer were observed (6 cases observed, 15.11 expected) (Friedman and Habel, unpublished data). Excesses of central nervous system neoplasms are frequently noted in follow-up studies of persons with epilepsy (Clemmesen et al., 1974; White et al., 1979; Shirts et al., 1986) and of diphenylhydantoin users (Friedman and Ury, 1980). These associations are most likely due to inclusion of persons whose seizures are an early manifestation of slow-growing tumors such as low-grade astrocytomas (Mathieson, 1975) rather than to carcinogenic effects of either seizures or anticonvulsants. Other confounding factors in these studies are greater surveillance (detection bias) and more frequent exposure to diagnostic radiation (or, in earlier studies, to Thorotrast) in persons with epilepsy. Of some concern are case reports of neuroblastoma and lymphoma in children exposed to diphenylhydantoin in utero who display the constellation of congenital anomalies known as fetal hydantoin syndrome (Cohen, 1981). No epidemiologic assessment of cancer risk in this group has yet been reported. All in all, there is suggestive evidence that use of diphenylhydantoin may increase risk for non-Hodgkin lymphoma slightly, but the link is by no means well established.

499

DRUGS FOR TREATING GASTRIC ACIDITY AND PEPTIC ULCERS The histamine-2 receptor antagonists (or H2RAs) were introduced beginning in the mid-1970s for treatment and secondary prevention of peptic ulcer disease. These agents suppress gastric production of hydrochloric acid and are now widely used for a variety of conditions related to gastric acidity. Some of the H2RAs are now available overthe-counter, and they are among the most commonly used medications in the United States (NDC Health, 2003). As early as 1979 (Elder et al., 1979) theoretical concern was raised that H2RAs could cause gastric cancer. Inhibition of gastric acid secretion raises intragastric pH, which may allow bacterial colonization of the normally sterile stomach with increased production of carcinogenic nitrates, nitrites, and N-nitroso compounds. It was also suggested that certain H2RAs may act directly as carcinogens (Wormsley, 1984). Because cimetidine, but not the other H2RAs, inhibits the binding of dihydrotestosterone to androgen receptors, interferes with the oxidative metabolism of estrogen, and increases serum prolactin levels, it has been hypothesized that this medication may influence the risk of hormonally mediated malignancies, for example, by decreasing the risk of prostate cancer (Stolinsky, 1991). There also has been interest in the possibility that cimetidine may influence tumor growth (Sasson et al., 1999). Early reports of gastric cancer developing shortly after use of cimetidine (Taylor et al., 1981) appeared to be explained by use of the medication for preexistent but undiagnosed cancer (Piper, 1981). This also has been observed in several more recent studies (reviewed by La Vecchia and Tavani [2002]). A few studies have examined the potential association between H2RAs and risk of other digestive tract cancers (e.g., esophagus, colon), and results have generally not supported an association (La Vecchia and Tavani, 2002). Cancer incidence or mortality at several different sites has been examined in four cohorts of cimetidine users, each with up to 8 to 10 years of follow-up (Moller et al., 1989; Moller and Mellemgaard, 1992; Colin-Jones et al., 1992; Habel et al., 2000; Rossing et al., 2000). None of these studies found a statistically significant association between cimetidine use and risk of any cancer, or cancers of the breast or prostate. Two studies (Habel et al., 2000; Rossing et al., 2000) at least partially addressed potential confounding by indication by including comparisons of risk among cimetidine users with risk among users of another H2RA. However, all studies were unable to adjust for additional potential confounders other than age and sex. Only one study had enough cancer events among H2RA users to examine dose, latency, or recency with any degree of precision (Rossing et al., 2000). Although there was generally no evidence of an increased or decreased risk of breast cancer in this study, there was a suggestion of a modest increased risk of prostate cancer among those who had over 20 prescriptions. Although not statistically significant, two of the cohort studies observed a higher rate of lymphoma among the cimetidine users. The results of two subsequent case-control studies of non-Hodgkin’s lymphoma were inconsistent (Holly et al., 1999; Beiderbeck et al., 2003). In summary, available human data provide little support for an association between cimetidine and other H2RAs and cancer risk. However, larger studies with longer follow-up are necessary before any conclusions are possible.

IMMUNOSUPPRESSIVE AGENTS Immunosuppressive agents are administered routinely to prevent graft rejection in bone marrow and organ transplantation, to treat several cancers, and they are also used frequently, although in smaller doses, to treat rheumatoid arthritis, systemic lupus erythematosus, and other autoimmune disorders. There are four classes of immunosupressant drugs: cyclosporine and tacrolimus, adrenocortical steriods, cytotoxic drugs (e.g., azathioprine, cyclophosphamide, methotrexate), and

500

PART III: THE CAUSES OF CANCER

antibody reagents (Diasio and LoBuglio, 1996). Most studies of immunosuppressants and cancer risk have focused on agents used for organ or bone marrow transplantation or for cancer therapy. Earlier studies of organ transplant recipients, most of whom received either azathioprine, cyclophosphamide, or both (Penn, 1977; Fraumeni and Hoover, 1977; Kinlen et al., 1979) consistently found large increases in risk for certain cancers, including lymphoma (particularly non-Hodgkin’s or B-cell lymphoma), Kaposi’s sarcoma, squamous cell and basal cell skin cancer, malignant melanoma, and hepatocellular carcinoma. Slightly to modestly elevated risks, of the order two- to fourfold, were also suggested for lung and bladder cancer (Fraumeni and Hoover, 1977). The incidence of lymphoma and Kaposi’s sarcoma was increased within months after inititation of treatment. An unusual predilection was noted for the central nervous system as the primary or only site of occurrence of the lymphomas. Incidence of lymphoma, Kaposi’s sarcoma, nonmelanoma skin cancer, and other malignancies is also increased after use of the more current immunosuppressive regimens, such as cyclosporine and methotrexate, for organ transplantation (Penn and Brunson, 1988; Birkeland et al., 2000; Lindelof et al., 2000; Adami et al., 2003; Eurvard et al., 2003). Treatment with cyclosporine and other immunosuppressive agents for bone marrow transplantation also has been associated with an increased risk of malignant neoplasms, especially solid tumors (Kolb et al., 1999; Curtis et al., 2003; Baker et al., 2003). There have been reports of spontaneous regression of both lymphomas and Kaposi’s sarcoma upon cessation of cyclosporine therapy (Starzl et al., 1984; Bencini et al., 1988). Increased risk for both nonHodgkin’s lymphoma and squamous cell skin cancer has also been reported in patients receiving immunosuppressive therapy for other conditions such as rheumatoid arthritis (Kinlen, 1985), although the excess risk was considerably lower than that after transplantation. A potential confounder is the condition being treated, as an increased occurrence of non-Hodgkin’s lymphoma has been reported in rheumatoid arthritis patients even in the absence of immunosuppressive therapy (Isomaki et al., 1978). Chronic lymphoid stimulation from an autoimmune process such as rheumatoid arthritis could cause the increased occurrence of lymphoma. However, the increase in risk for lymphoma due to rheumatoid arthritis itself appears to be much less than the 13-fold increase (Kinlen, 1985) seen after immunosuppressive therapy.

LAXATIVES The most widely used commercial laxatives, over-the-counter and prescribed, are the stimulant preparations, which include the anthraquinone derivatives or anthranoids and the diphenylmethane derivatives, phenolphthalein and bisacodyl (Brunton, 1996). In vitro and animal data suggest that several anthranoids may be carcinogenic (Hallmann, 2000). Danthron, one of the anthranoids used in laxatives since the beginning of the 1900s, was banned by the FDA in 1987 after there was sufficient evidence of carcinogenicity in animals (IARC, 1990). Other anthranoids, such as senna, are under investigation but continue to be ingredients in stimulant laxatives. In 1996, the FDA reported the results of studies in rodents indicating that phenolphthalein is carcinogenic and genotoxic in several test systems (IARC, 2000). By 1999, phenolphthalein-containing laxatives were largely withdrawn from the United States and many markets in Europe. Despite in vitro and animal studies, results of human studies of commercially available stimulant laxatives have been largely negative. A number of studies have examined the potential association between commercial laxatives and cancers of the colon and rectum (Sonnenberg and Muller, 1993; IARC, 2000). Although some studies have observed an association, those able to adjust for constipation, another suspected risk factor for colorectal cancer, have generally found no association (Jacobs and White, 1998; Dukas et al., 2000b; Roberts et al., 2003). In addition, a cohort study (Dukas et al., 2000a) and an analysis of three case-control studies (Longnecker et al., 1997) found no consistent association between use of laxatives and risk of adenomatous polyps.

Use of phenolphthalein laxatives was examined in a hospital-based study of 18,163 cancer cases (16 sites or types). Controls (N = 12,204) were selected from patients admitted for a nonmalignant condition that was also judged to be unrelated to laxative use (trauma, acute infections, orthopedic conditions) (Coogan et al., 2000b). When laxative use was defined as use for at least 3 months beginning at least 2 years before hospital admission, neither use of phenolphthalein laxatives or use of anthranoid laxatives was statistically significantly associated (either positively or negatively) with any cancer site, with the exception of an increased risk of stomach cancer associated with use of anthranoid laxatives (RR = 2.0, 95% CI 1.1–3.6). An excess of tumors of the ovary has been observed in female mice exposed to phenolphthalein. However, in 410 epithelial ovarian cancer cases and 713 controls diagnosed from 1994 to 1998 in contiguous counties of Pennsylvania, New Jersey, and Delaware (Cooper et al., 2000), no association was observed between cancer risk and use phenolphthalein-containing laxatives.

MAGNITUDE OF THE PROBLEM OF CARCINOGENESIS DUE TO DRUGS For the past several decades, new drugs have appeared on the market at an ever-increasing rate, and a significant proportion of adults and children now take one or more medications regularly. Available data suggest that, aside from the well-known exceptions described here and in the chapter on hormones, most drugs, as ordinarily used in clinical practice, do not increase or decrease the risk of developing cancer. However, the power to detect relatively weak effects has been low for many drugs, given the rarity of most cancers. In addition, more followup will be needed to rule out longer term effects of many medications. In order to minimize the unnecessary use of harmful drugs, identify potential chemopreventive agents, and gain insight into mechanisms of carcinogenesis, it is important that epidemiologic research continue to include routine surveillance of commonly used medications and well-controlled studies of specific drug–cancer associations. Acknowledgments We gratefully acknowledge the support for our studies of drug carcinogenesis by Public Health Service Grants No. R37-CA19939 and No. R35-CA49761 from the National Cancer Institute. Joe V. Selby and Lisa Herrinton made extensive contributions to this chapter in the previous edition of Cancer Epidemiology and Prevention. Much of their work is included in the current updated chapter.

References Adami J, Gabel H, Lindelof B, et al. 2003. Cancer risk following organ transplantation: A nationwide cohort study in Sweden. Br J Cancer 89: 1221–1227. Ahr HJ, Bomhard E, Enzmann H, et al. 1998. Calcium channel blockers and the risk of cancer: a preclinical assessment. Cardiovasc Drugs Ther 12(2):157–169. American Medical Association, Department of Drugs, Division of Drugs and Technology. 1986. Drug Evaluations. 6th ed. Chicago: American Medical Association, pp. 1579–1580. Armstrong BK, White E, Saracci R. 1992. Exposure measurement. In: Armstrong BK, White E, Saracci R, eds. Principles of Exposure Measurement in Epidemiology. Oxford: Oxford University Press, pp. 1–21. Autier P, Dore JF, Schifflers E, et al. 1995. Melanoma and use of sunscreens: An EORTC case–control study in Germany, Belgium and France. The EORTC Melanoma Cooperative Group. Int J Cancer 61(6):749–755. Autier P, Severi G, Boyle P, Dore JF. 2000. RESPONSE more about: sunscreen use and duration of sun exposure: A double–blind, randomized trial. J Natl Cancer Inst 92(18):1532–1533. Badawi AF. 2000. The role of prostaglandin synthesis in prostate cancer. BJU Int 85(4):451–462. Baker KS, DeFor TE, Burns LJ, Ramsay NKC, Neglia JP, Robison LL. 2003. New malignancies after blood or marrow stem-cell transplantation in children and adults: Incidence and risk factors. J Clin Oncol 21: 1352–1358. Balo J. 1979. Role of hydrazine in carcinogenesis. Adv Cancer Res 30: 151–164. Baron JA, Sandler RS. 2000. Nonsteroidal anti-inflammatory drugs and cancer prevention. Annu Rev Med 51:511–523.

Pharmaceuticals Other Than Hormones Beard CM, Noller KL, O’Fallon WM, Kurland LT, Dahlin DC. 1988. Cancer after exposure to metronidazole. Mayo Clin Proc 63(2):147–153. Beard CM, Noller KL, O’Fallon WM, Kurland LT, Dockerty MB. 1979. Lack of evidence for cancer due to use of metronidazole. N Engl J Med 301(10):519–522. Beck P, Wysowski DK, Downey W, Butler-Jones D. 2003. Statin use and the risk of breast cancer. J Clin Epidemiol 56(3):280–285. Beebe KL, Zaninelli R, Trock B. 2000. Re: “Antidepressant medication use and breast cancer risk”. Am J Epidemiol 152(11):1104–1105. Beiderbeck AB, Holly EA, Sturkenboom MC, Coebergh JW, Stricker BH, Leufkens HG. 2003. Prescription medications associated with a decreased risk of non-Hodgkin’s lymphoma. Am J Epidemiol 157(6):510–516. Bencini PL, Marchesi L, Cainelli T, Crosti C. 1988. Kaposi’s sarcoma in kidney transplant recipients treated with cyclosporin. Br J Dermatol 118(5): 709–714. Bendesky A, Menendez D, Ostrosky-Wegman P. 2002. Is metronidazole carcinogenic? Mutat Res 511(2):133–144. Benet LZ. 1996. Introduction. In: Hardman JG, Limbird LE, Molinoff PB, Ruddon RW, Gilman AG, eds. Goodman and Gilman’s The Pharmacological Basis of Therapeutics. New York: McGraw-Hill, pp. 1–2. Bengtsson U, Angervall L, Ekman H, Lehmann L. 1968. Transitional cell tumors of the renal pelvis in analgesic abusers. Scand J Urol Nephrol 2(3):145–50. Besamusca FW, Bastiaensen LA. 1986. Blood dyscrasias and topically applied chloramphenicol in ophthalmology. Doc Ophthalmol 64(1):87– 95. Bhasin G, Kausar H, Sarwar AM, Athar M. 2003. Progressive iron overload enhances chemically mediated tumor promotion in murine skin. Arch Biochem Biophys 409(2):262–273. Birkeland SA, Lokkegaard H, Storm HH. 2000. Cancer risk in patients on dialysis and after renal transplantation. Lancet 355:1886–1887. Bjerre LM, LeLorier J. 2001. Do statins cause cancer? A meta-analysis of large randomized clinical trials. Am J Med 110(9):716–723. Blair IA, Mansilla Tinoco R, Brodie MJ, et al. 1985. Plasma hydrazine concentrations in man after isoniazid and hydralazine administration. Hum Toxicol 4(2):195–202. Blais L, Desgagne A, LeLorier J. 2000. 3-Hydroxy-3-methylglutaryl coenzyme A reductase inhibitors and the risk of cancer: A nested case-control study. Arch Intern Med 160(15):2363–2368. Blayney DW, Longo DL, Young RC, et al. 1987. Decreasing risk of leukemia with prolonged follow-up after chemotherapy and radiotherapy for Hodgkin’s disease. N Engl J Med 316(12):710–714. Borhani NO, Mercuri M, Borhani PA, et al. 1996. Final outcome results of the Multicenter Isradipine Diuretic Atherosclerosis Study (MIDAS). A randomized controlled trial. JAMA 276(10):785–791. Bosetti C, Gallus S, La Vecchia C. 2002. Aspirin and cancer risk: An update to 2001. Eur J Cancer Prev 11(6):535–542. Boyd NF, McGuire V. 1990. Evidence of association between plasma highdensity lipoprotein cholesterol and risk factors for breast cancer. J Natl Cancer Inst 82(6):460–468. Brandes LJ. 1992. Re: Depression, antidepressant medication, and cancer. Am J Epidemiol 136(11):1414–1417. Brandes LJ, Arron RJ, Bogdanovic RP, et al. 1992. Stimulation of malignant growth in rodents by antidepressant drugs at clinically relevant doses. Cancer Res 52(13):3796–3800. Braun S, Boyko V, Behar S, et al. 1998. Calcium channel blocking agents and risk of cancer in patients with coronary heart disease. Benzafibrate Infarction Prevention (BIP) Study Research Group. J Am Coll Cardiol 31(4): 804–808. Bringuier PP, McCredie M, Sauter G, et al. 1998. Carcinomas of the renal pelvis associated with smoking and phenacetin abuse: p53 mutations and polymorphism of carcinogen–metabolising enzymes. Int J Cancer 79(5):531–536. Brower V. 2003. Of cancer and cholesterol: studies elucidate anticancer mechanisms of statins. J Natl Cancer Inst 95(12):844–846. Brunton LL. 1996. Agents affecting gastrointestinal water flux and motility; emesis and antiemetics; bile acids and pancreatic enzymes. In: Hardman JG, Limbird LE, Molinoff PB, Ruddon RW, Gilman AG, eds. Goodman and Gilman’s The Pharmacological Basis of Therapeutics. New York: McGraw-Hill, pp. 917–936. Bruynzeel I, Bergman W, Hartevelt HM, et al. 1991. “High single–dose” European PUVA regimen also causes an excess of non-melanoma skin cancer. Br J Dermatol 124(1):49–55. Carpenter HM, Mudge GH. 1981. Acetaminophen nephrotoxicity: studies on renal acetylation and deacetylation. J Pharmacol Exp Ther 218(1): 161–167. Carson J, Ray W, Strom B. 2000. Medicaid Databases. In: Strom BL, ed. Pharmacoepidemiology. New York: John Wiley & Sons, pp. 307–324.

501

Cartwright LE, Walter JF. 1983. Psoralen-containing sunscreen is tumorigenic in hairless mice. J Am Acad Dermatol 8(6):830–836. Cartwright RA, Bernard SM, Bird CC, et al. 1987. Chronic lymphocytic leukaemia: Case control epidemiological study in Yorkshire. Br J Cancer 56(1):79–82. Castelao JE, Gago-Dominguez M, Yuan JM, Ross RK, Yu MC. 2003. Phenobarbital use and bladder cancer risk. Eur J Epidemiol 18(7):659–664. Chan KK, Oza AM, Siu LL. 2003. The statins as anticancer agents. Clin Cancer Res 9(1):10–19. Chan K, Platt R. 2000. Harvard Pilgrim Health Care/Harvard Vanguard Medical Associates. In: Strom BL, ed. Pharmacoepidemiology. New York: John Wiley & Sons, pp. 285–294. Charlton M, Lunsford D. 1971. Le sostanze di idantoina come possibile cause del linfoma maligno. Njnerva Ned; 62:2185. Chen Z, Chen GQ, Shen ZX, et al. 2002. Expanding the use of arsenic trioxide: leukemias and beyond. Semin Hematol 39(2 Suppl 1):22– 26. Chow WH, McLaughlin JK, Linet MS, Niwa S, Mandel JS. 1994. Use of analgesics and risk of renal cell cancer. Int J Cancer 59(4):467–470. Christie D. 1993. Did Vietnam veterans get cancer from dapsone? Med J Aust 159(8):500. Clemmesen J, Fuglsang-Frederiksen V, Plum CM. 1974. Are anticonvulsants oncogenic? Lancet 1(7860):705–707. Clemmesen J, Hjalgrim-Jensen S. 1979. Is isonicotinic acid hydrazide (INH) carcinogenic to man? A 24-year follow-up of 3371 tuberculosis cases. Ecotoxicol Environ Saf 3(4):439–450. Cohen MM, Jr. 1981. Neoplasia and the fetal alcohol and hydantoin syndromes. Neurobehav Toxicol Teratol 3(2):161–162. Cohen SM, Shirai T, Steineck G. 2000. Epidemiology and etiology of premalignant and malignant urothelial changes. Scand J Urol Nephrol Suppl (205):105–115. Colin-Jones DG, Langman MJ, Lawson DH, Logan RF, Paterson KR, Vessey MP. 1992. Postmarketing surveillance of the safety of cimetidine: 10 year mortality report. Gut 33(9):1280–1284. Collet JP, Boivin JF. 2000. Bias and confounding in pharmacoepidemiology. In: Strom BL, ed. Pharmacoepidemiology. New York: John Wiley & Sons, pp. 765–784. Coogan PF, Rosenberg L, Palmer JR, et al. 2000a. Risk of ovarian cancer according to use of antidepressants, phenothiazines, and benzodiazepines (United States). Cancer Causes Control 11(9):839–845. Coogan PF, Rosenberg L, Palmer JR, Strom BL, Zauber AG, Shapiro S. 2002. Statin use and the risk of breast and prostate cancer. Epidemiology 13(3):262–267. Coogan PF, Rosenberg L, Palmer JR, et al. 2000b. Phenolphthalein laxatives and risk of cancer. J Natl Cancer Inst 92(23):1943–1944. Cooper GS, Longnecker MP, Sandler DP, Ness RB. 2000. Risk of ovarian cancer in relation to use of phenolphthalein-containing laxatives. Br J Cancer 83(3):404–406. Corley DA, Kerlikowske K, Verma R, Buffler P. 2003. Protective association of aspirin/NSAIDs and esophageal cancer: A systematic review and metaanalysis. Gastroenterology 124(1):47–56. Costello HD, Snider DE, Jr. 1980. The incidence of cancer among participants in controlled, randomized isoniazid preventive therapy trial. Am J Epidemiol 111(1):67–74. Cotterchio M, Kreiger N. 2000. Re: “Antidepressant medication use and breast cancer risk” two of the authors reply. Am J Epidemiol 152(11):1105. Cotterchio M, Kreiger N, Darlington G, Steingart A. 2000. Antidepressant medication use and breast cancer risk. Am J Epidemiol 151(10):951– 957. Curtis RE, Rowlings PA, Deeg HJ, et al. 1997. Solid cancers after bone marrow transplantation. N Engl J Med 336:897–904. Cuzick J, Evans S, Gillman M, Price Evans DA. 1982. Medicinal arsenic and internal malignancies. Br J Cancer 45(6):904–911. Dalton SO, Johansen C, Mellemkjaer L, et al. 2000. Antidepressant medications and risk for cancer. Epidemiology 11(2):171–176. Danielson DA, Hannan MT, Jick H. 1982a. Metronidazole and cancer. JAMA 247(18):2498–2499. Danielson DA, Jick H, Hunter JR, Stergachis A, Madsen S. 1982b. Nonestrogenic drugs and breast cancer. Am J Epidemiol 116(2):329–332. Dethloff LA, Graziano MJ, Goldenthal E, Gough A, de la Iglesia FA. 1996. Perspective on the carcinogenic potential of phenytoin based on rodent tumor bioassays and human epidemiological data. Hum Exp Toxicol 15(4):335–348. Diasio RB, LoBuglio AF. 1996. Immunomodulators: immunosuppresive agents and immunostimulants. In: Hardman JG, Limbird LE, Molinoff PB, Ruddon RW, Gilman AG, eds. Goodman and Gilman’s The Pharmacological Basis of Therapeutics. New York: McGraw-Hill, pp. 1291– 1308.

502

PART III: THE CAUSES OF CANCER

Dobyns BM, Sheline GE, Workman JB, Tompkins EA, McConahey WM, Becker DV. 1974. Malignant and benign neoplasms of the thyroid in patients treated for hyperthyroidism: A report of the cooperative thyrotoxicosis therapy follow-up study. J Clin Endocrinol Metab 38(6): 976–998. Doody MM, Linet MS, Glass AG, et al. 1996. Risks of non-Hodgkin’s lymphoma, multiple myeloma, and leukemia associated with common medications. Epidemiology 7(2):131–139. Dorak MT, Burnett AK, Worwood M. 2002. Hemochromatosis gene in leukemia and lymphoma. Leuk Lymphoma 43(3):467–477. Downey W, Beck P, McNutt M, Stang M, Osei W, Nichol J. 2000. Health databases in Saskatchewan. In: Strom BL, ed. Pharmacoepidemiology. New York: John Wiley & Sons, pp. 325–346. Downs JR, Clearfield M, Tyroler HA, et al. 2001. Air Force/Texas Coronary Atherosclerosis Prevention Study (AFCAPS/TEXCAPS): Additional perspectives on tolerability of long-term treatment with lovastatin. Am J Cardiol 87(9):1074–1079. Dublin S, Rossing MA, Heckbert SR, Goff BA, Weiss NS. 2002. Risk of epithelial ovarian cancer in relation to use of antidepressants, benzodiazepines, and other centrally acting medications. Cancer Causes Control 13(1): 35–45. Dukas L, Platz EA, Colditz GA, Willet WC, Giovannucci EL. 2000a. Bowel movement, use of laxatives and risk of colorectal adenomatous polyps among women (United States). Cancer Causes Control 11(10):907–914. Dukas L, Willett WC, Colditz GA, Fuchs CS, Rosner B, Giovannucci EL. 2000b. Prospective study of bowel movement, laxative use, and risk of colorectal cancer among women. Am J Epidemiol 151(10):958–964. Eichholzer M, Stahelin HB, Gutzwiller F, Ludin E, Bernasconi F. 2000. Association of low plasma cholesterol with mortality for cancer at various sites in men: 17-y follow-up of the prospective Basel study. Am J Clin Nutr 71(2):569–574. Elder JB, Ganguli PC, Gillespie IE. 1979. Cimetidine and gastric cancer. Lancet 1(8124):1005–1006. Euvrard S Kanitakis J, Claudy A. 2003. Skin cancers after organ transplantation. N Engl J Med 348:1681–1691. Evans J, MacDonald T. 2000. The Tayside Medicines Monitoring Unit (MEMO). In: Strom BL, ed. Pharmacoepidemiology. New York: John Wiley & Sons, pp. 361–374. Falagas ME, Walker AM, Jick H, Ruthazer R, Griffith J, Snydman DR. 1998. Late incidence of cancer after metronidazole use: A matched metronidazole user/nonuser study. Clin Infect Dis 26(2):384–388. Felmeden DC, Lip GY. 2001. Antihypertensive therapy and cancer risk. Drug Saf 24(10):727–739. Finkle WD, McLaughlin JK, Rasgon SA, Yeoh HH, Low JE. 1993. Increased risk of renal cell cancer among women using diuretics in the United States. Cancer Causes Control 4(6):555–558. Fitzpatrick AL, Daling JR, Furberg CD, Kronmal RA, Weissfeld JL. 1997. Use of calcium channel blockers and breast carcinoma risk in postmenopausal women. Cancer 80(8):1438–1447. Fokkens W. 1979. Phenacetin abuse related to bladder cancer. Environ Res 20(1):192–198. Fraser GE, Phillips RL, Beeson WL. 1990. Hypertension, antihypertensive medication and risk of renal carcinoma in California Seventh-Day Adventists. Int J Epidemiol 19(4):832–838. Fraser HS. 1996. Reserpine: A tragic victim of myths, marketing, and fashionable prescribing. Clin Pharmacol Ther 60(4):368–373. Fraumeni JF Jr., Hoover R. 1977. Immunosurveillance and cancer: Epidemiologic observations. Natl Cancer Inst Monogr 47:121–126. Friedman GD. 1992. Re: antidepressant drugs, depression, and cancer: An editor comments. Am J Epidemiol 136:1415–1416. Friedman GD. 1980. Cancer after metronidazole. N Engl J Med 302(9):519. Friedman GD. 1981. Barbiturates and lung cancer in humans. J Natl Cancer Inst 67(2):291–295. Friedman GD. 1982. Phenylbutazone, musculoskeletal disease, and leukemia. J Chronic Dis 35(4):233–243. Friedman GD. 1983. Rauwolfia and breast cancer: No relation found in long term users age fifty and over. J Chronic Dis 36(5):367–370. Friedman GD, Habel LA. 1999. Barbiturates and lung cancer: A re-evaluation. Int J Epidemiol 28(3):375–379. Friedman GD, Selby JV. 1989. Metronidazole and cancer. JAMA 261(6):866. Friedman GD, Ury HK. 1980. Initial screening for carcinogenicity of commonly used drugs. J Natl Cancer Inst 65(4):723–733. Friedman GD, Ury HK. 1983. Screening for possible drug carcinogenicity: second report of findings. J Natl Cancer Inst 71(6):1165–1175. Friedman G, Habel LA, Boles M, McFarland B. 2000. Kaiser Permanente Medical Care Program: Division of Research, Northern California, and Center for Health Research, Northwest Division. In: Strom BL, ed. Pharmacoepidemiology. New York: John Wiley & Sons, pp. 263–284.

Friis S, Sorensen HT, Mellemkjaer L, et al. 2001. Angiotensin-converting enzyme inhibitors and the risk of cancer: A population-based cohort study in Denmark. Cancer 92(9):2462–2470. Gago-Dominguez M, Yuan JM, Castelao JE, Ross RK, Yu MC. 1999. Regular use of analgesics is a risk factor for renal cell carcinoma. Br J Cancer 81(3):542–548. Gams RA, Neal JA, Conrad FG. 1968. Hydantoin-induced pseudopseudolymphoma. Ann Intern Med 69(3):557–568. García Rodríguez LA, Pérez-Gutthann S, Jick S. 2000. The UK General Practice Research Database. In: Strom BL, ed. Pharmacoepidemiology. New York: John Wiley & Sons, pp. 375–386. Glassroth JL, Snider DE, Comstock GW. 1977a. Urinary tract cancer and isoniazid. Am Rev Respir Dis 116(2):331–333. Glassroth JL, White MC, Snider DE Jr. 1977b. An assessment of the possible association of isoniazid with human cancer deaths. Am Rev Respir Dis 116(6):1065–1074. Godschalk RW, Ostertag JU, Moonen EJ, Neumann HA, Kleinjans JC, van Schooten FJ. 1998. Aromatic DNA adducts in human white blood cells and skin after dermal application of coal tar. Cancer Epidemiol Biomarkers Prev 7(9):767–773. Gold E, Gordis L, Tonascia J, Szklo M. 1978. Increased risk of brain tumors in children exposed to barbiturates. J Natl Cancer Inst 61(4):1031–1034. Goldhaber MK, Selby JV, Hiatt RA, Quesenberry CP. 1990. Exposure to barbiturates in utero and during childhood risk of intracranial and spinal cord tumors. Cancer Res 50(15):4600–4603. Grossman E, Messerli FH, Goldbourt U. 1999. Carcinogenicity of cardiovascular drugs. Curr Hypertens Rep 1(3):212–218. Gyte GM, Richmond JE, Williams JR, Atwood JL. 1985. Hairy cell leukaemia occurring during phenytoin (diphenylhydantoin) treatment. Scand J Haematol 35(3):358–362. Habel LA, Bull SA, Friedman GD. 1998. Barbiturates, smoking, and bladder cancer risk. Cancer Epidemiol Biomarkers Prev 7(11):1049–1050. Habel LA, Levin TR, Friedman GD. 2000. Cimetidine use and risk of breast, prostate and other cancers. Pharmacoedpidemiol Drug Safety 9:149–155. Hallmann F. 2000. Toxicity of commonly used laxatives. Med Sci Monit 6(3):618–628. Harlow BL, Cramer DW. 1995. Self-reported use of antidepressants or benzodiazepine tranquilizers and risk of epithelial ovarian cancer: Evidence from two combined case-control studies (Massachusetts, United States). Cancer Causes Control 6(2):130–134. Harlow BL, Cramer DW, Baron JA, Titus–Ernstoff L, Greenberg ER. 1998. Psychotropic medication use and risk of epithelial ovarian cancer. Cancer Epidemiol Biomarkers Prev 7(8):697–702. Heart Protection Study Collaborative Group. 2002. MRC/BHF Heart Protection Study of cholesterol lowering with simvastatin in 20,536 high-risk individuals: A randomised placebo-controlled trial. Lancet 360(9326): 7–22. Hebert PR, Gaziano JM, Chan KS, Hennekens CH. 1997. Cholesterol lowering with statin drugs, risk of stroke, and total mortality. An overview of randomized trials. JAMA 278(4):313–321. Heinonen O, Slone D, Shapiro S. 1977. Birth defects and drugs in pregnancy. Littleton, Mass: PSG Publishing Co. Herrinton LJ, Friedman GD, Baer D, Selby JV. 1995. Transferrin saturation and risk of cancer. Am J Epidemiol 142(7):692–698. Hiatt RA, Tolan K, Quesenberry CP Jr. 1994. Renal cell carcinoma and thiazide use: A historical, case-control study (California, USA). Cancer Causes Control 5(4):319–325. Hixson LJ, Alberts DS, Krutzsch M, et al. 1994. Antiproliferative effect of nonsteroidal antiinflammatory drugs against human colon cancer cells. Cancer Epidemiol Biomarkers Prev 3(5):433–438. Hole DJ, Gillis CR, McCallum IR, et al. 1998. Cancer risk of hypertensive patients taking calcium antagonists. J Hypertens 16(1):119–124. Holly EA, Lele C, Bracci PM, McGrath MS. 1999. Case-control study of nonHodgkin’s lymphoma among women and heterosexual men in the San Francisco Bay Area, California. Am J Epidemiol 150(4):375–389. Hoover RN. 1998. Cancer prevention: better late than never? Ann Intern Med 128(9):771–772. Hultengren N. 1961. Renal papillary necrosis, a clinical study of 103 cases. Acta Chir Scand Suppl 277:1–84. Hultengren N. 1968. Carcinoma of the renal pelvis in renal papillary necrosis. Acta Chir Scand 130:314–320. Hyman GA, Sommers SC. 1966. The development of Hodgkin’s disease and lymphoma during anticonvulsant therapy. Blood 28(3):416–427. Insel P. 1996. Analgesic-antipyretic and antiinflammatory agents and drugs employed in the treatment of gout. In: Hardman JG, Limbird LE, Molinoff PB, Ruddon RW, Gilman AG, eds. Goodman and Gilman’s The Pharmacological Basis of Therapeutics. New York: McGraw-Hill, pp. 617–657.

Pharmaceuticals Other Than Hormones International Agency for Research on Cancer. 1980. Some pharmaceutical drugs. In: IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, Vol. 24. Lyon, France: IARC. International Agency for Research on Cancer. 1986. Some naturally occuring and synthetic food components, furocoumarins and ultraviolet radiation. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Vol. 40. Lyon, France: IARC. International Agency for Research on Cancer. 1987. Overall evaluations of carcinogenicity: An updating of IARC Monographs Volumes 1 to 42. IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans (Supplement 7). Lyon, France: IARC. International Agency for Research on Cancer. 1990. Pharmaceutical drugs. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Vol. 50. Lyon, France: IARC. International Agency for Research on Cancer. 1996. Some pharmaceutical drugs. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Vol. 66. Lyon, France: IARC. International Agency for Research on Cancer. 2000. Some antiviral and antineoplastic drugs and other pharmaceutical agents. IARC Monographs on the Evaluation of Carcinogenic Risk to Humans, Vol. 76. Lyon, France: IARC. International Agency for Research on Cancer. 2001. Some thyrotropic agents. IARC Monographs on the Evaluation of Carcinogenic Risk to Humans, Vol. 79. Lyon, France: IARC. Isomaki HA, Hakulinen T, Joutsenlahti U. 1978. Excess risk of lymphomas, leukemia and myeloma in patients with rheumatoid arthritis. J Chronic Dis 31(11):691–696. Ito N, Fukushima S, Shirai T, Hagiwara A, Imaida K. 1984. Drugs, food additives and natural products as promoters in rat urinary bladder carcinogenesis. IARC Sci Publ (56):399–407. Jacobs EJ, White E. 1998. Constipation, laxative use, and colon cancer among middle-aged adults. Epidemiology 9(4):385–391. Jick H, Jick S, Derby LE, Vasilakis C, Myers MW, Meier CR. 1997. Calciumchannel blockers and risk of cancer. Lancet 349(9051):525–528. Johansson S, Angervall L, Bengtsson U, Wahlqvist L. 1974. Uroepithelial tumors of the renal pelvis associated with abuse of phenacetincontaining analgesics. Cancer 33(3):743–753. Johnson TW, Anderson KE, Lazovich D, Folsom AR. 2002. Association of aspirin and nonsteroidal anti-inflammatory drug use with breast cancer. Cancer Epidemiol Biomarkers Prev 11(12):1586–1591. Jonas M, Goldbourt U, Boyko V, Mandelzweig L, Behar S, Reicher-Reiss H. 1998. Nifedipine and cancer mortality: Ten-year follow-up of 2607 patients after acute myocardial infarction. Cardiovasc Drugs Ther 12(2):177–181. Kanamasa K, Kimura A, Miyataka M, Takenaka T, Ishikawa K. 1999. Incidence of cancer in postmyocardial infarction patients treated with shortacting nifedipine and diltiazem. Secondary Prevention Group. Cancer 85(6):1369–1374. Kantor AF, Hartge P, Hoover RN, Fraumeni JF Jr. 1985. Tuberculosis chemotherapy and risk of bladder cancer. Int J Epidemiol 14(1):182– 184. Kari F, Bucher J, Haseman J, Eustis S, Huff J. 1995. Long-term exposure to the anti–inflammatory agent phenylbutazone induces kidney tumors in rats and liver tumors in mice. Jpn J Cancer Res 86(3):252–263. Kaye JA, Meier CR, Walker AM, Jick H. 2002. Statin use, hyperlipidaemia, and the risk of breast cancer. Br J Cancer 86(9):1436–1439. Kelly JP, Rosenberg L, Palmer JR, et al. 1999. Risk of breast cancer according to use of antidepressants, phenothiazines, and antihistamines. Am J Epidemiol 150(8):861–868. Khuder SA, Mutgi AB. 2001. Breast cancer and NSAID use: a meta-analysis. Br J Cancer 84(9):1188–1192. Kinlen LJ. 1985. Incidence of cancer in rheumatoid arthritis and other disorders after immunosuppressive treatment. Am J Med 78(1A):44– 49. Kinlen LJ. 1996. Immunologic factors, including AIDS. In: Schottenfeld D, Fraumeni JF Jr., eds. Cancer Epidemiology and Prevention. New York: Oxford University Press, pp. 532–545. Kinlen LJ, Sheil AG, Peto J, Doll R. 1979. Collaborative United KingdomAustralasian study of cancer in patients treated with immunosuppressive drugs. Br Med J 2(6203):1461–1466. Klaassen C. 1996a. Principles of toxicology and treatment of poisoning. In: Hardman JG, Limbird LE, Molinoff PB, Ruddon RW, Gilman AG, eds. Goodman and Gilman’s The Pharmacological Basis of Therapeutics. New York: McGraw-Hill, pp. 63–76. Klaassen C. 1996b. Heavy metals and heavy-metal antagonists. In: Hardman JG, Limbird LE, Molinoff PB, Ruddon RW, Gilman AG, eds. Goodman and Gilman’s The Pharmacological Basis of Therapeutics. New York: McGraw-Hill, pp. 1649–1671.

503

Knekt P, Reunanen A, Takkunen H, Aromaa A, Heliovaara M, Hakulinen T. 1994. Body iron stores and risk of cancer. Int J Cancer 56(3):379–382. Kolb HJ, Socie G, Duell T, et al. 1999. Malignant neoplasms in long-term survivors of bone marrow transplantation. Ann Intern Med 131:738–744. Kreiger N, Marrett LD, Dodds L, Hilditch S, Darlington GA. 1993. Risk factors for renal cell carcinoma: Results of a population-based case-control study. Cancer Causes Control 4(2):101–110. La Vecchia C, Tavani A. 2002. A review of epidemiological studies on cancer in relation to the use of anti-ulcer drugs. Eur J Cancer Prev 11(2): 117–123. Law MR, Thompson SG, Wald NJ. 1994. Assessing possible hazards of reducing serum cholesterol. BMJ 308(6925):373–379. Lawlor DA. 2000. Re: “Antidepressant medication use and breast cancer risk”. Am J Epidemiol 152(11):1104. Leufkens H, Urquhart J. 2000. Automated pharmacy record linkage in The Netherlands. In: Strom BL, ed. Pharmacoepidemiology. New York: John Wiley & Sons, pp. 347–360. Lever AF, Hole DJ, Gillis CR, et al. 1998. Do inhibitors of angiotensin-Iconverting enzyme protect against risk of cancer? Lancet 352(9123): 179–184. Li FP, Willard DR, Goodman R, Vawter G. 1975. Malignant lymphoma after diphenylhydantoin (dilantin) therapy. Cancer 36(4):1359–1362. Liaw KL, Linet MS, McLaughlin JK, et al. 1997. Possible relation between hypertension and cancers of the renal pelvis and ureter. Int J Cancer 70(3):265–268. Lindblad P, McLaughlin JK, Mellemgaard A, Adami HO. 1993. Risk of kidney cancer among patients using analgesics and diuretics: A population-based cohort study. Int J Cancer 55(1):5–9. Lindelof B, Sigurgeirsson B, Tegner E, et al. 1999. PUVA and cancer risk: The Swedish follow–up study. Br J Dermatol 141(1):108–112. Lindholm LH, Anderson H, Ekbom T, et al. 2001. Relation between drug treatment and cancer in hypertensives in the Swedish Trial in Old Patients with Hypertension 2: A 5-year, prospective, randomised, controlled trial. Lancet 358(9281):539–544. Linkins RW, Comstock GW. 1990. Depressed mood and development of cancer. Am J Epidemiol 132(5):962–972. Longnecker MP, Sandler DP, Haile RW, Sandler RS. 1997. Phenolphthaleincontaining laxative use in relation to adenomatous colorectal polyps in three studies. Environ Health Perspect 105(11):1210–1212. Mann R. 2000. Prescription-event monitoring. In: Strom BL, ed. Pharmacoepidemiology. New York: John Wiley & Sons, pp. 231–246. Mason RP. 1999. Calcium channel blockers, apoptosis and cancer: Is there a biologic relationship? J Am Coll Cardiol 34(7):1857–1866. Mathieson G. 1975. Pathologic aspects of epilepsy with special reference to the surgical pathology of focal cerebral seizures. Adv Neurol 8:107–138. Matzner Y, Polliack A. 1978. Myelomatosis after phenytoin therapy: A chance association? Scand J Haematol 21(4):309–312. Maughan WZ, Muller SA, Perry HO, Pittelkow MR, O’Brien PC. 1980. Incidence of skin cancers in patients with atopic dermatitis treated with ocal tar. A 25-year follow-up study. J Am Acad Dermatol 3(6):612–615. McCredie M, Ford JM, Stewart JH. 1988. Risk factors for cancer of the renal parenchyma. Int J Cancer 42(1):13–16. McCredie M, Ford JM, Taylor JS, Stewart JH. 1982. Analgesics and cancer of the renal pelvis in New South Wales. Cancer 49(12):2617–2625. McCredie M, Pommer W, McLaughlin JK, et al. 1995. International renal-cell cancer study. II. Analgesics. Int J Cancer 60(3):345–349. McCredie M, Stewart J, Smith D, Supramaniam R, Williams S. 1999. Observations on the effect of abolishing analgesic abuse and reducing smoking on cancers of the kidney and bladder in New South Wales, Australia, 1972–1995. Cancer Causes Control 10(4):303–311. McCredie M, Stewart JH. 1988. Does paracetamol cause urothelial cancer or renal papillary necrosis? Nephron 49(4):296–300. McCredie M, Stewart JH, Day NE. 1993. Different roles for phenacetin and paracetamol in cancer of the kidney and renal pelvis. Int J Cancer 53(2):245–249. McCredie M, Stewart JH, Ford JM, MacLennan RA. 1983. Phenacetin-containing analgesics and cancer of the bladder or renal pelvis in women. Br J Urol 55(2):220–224. McGettigan P, Henry D, Hennessy S. 2000. Basic principles of clinical pharmacology relevant to pharmacoepidemiology studies. In: Strom BL, ed. Pharmacoepidemiology. New York: John Wiley & Sons, pp. 41–62. McLaughlin JK, Blot WJ, Fraumeni JF Jr. 1988. Diuretics and renal cell cancer. J Natl Cancer Inst 80(5):378. McLaughlin JK, Gao YT, Gao RN, et al. 1992. Risk factors for renal-cell cancer in Shanghai, China. Int J Cancer 52(4):562–565. McLaughlin JK, Mandel JS, Blot WJ, Schuman LM, Mehl ES, Fraumeni JF Jr. 1984. A population-based case-control study of renal cell carcinoma. J Natl Cancer Inst 72(2):275–284.

504

PART III: THE CAUSES OF CANCER

Meakawa A, Onodera H, Tanigawa H, et al. 1987. Long-term studies on carcinogenicity and promoting effect of phenylbutazone in DONRYU rats. J Natl Cancer Inst 79(3):577–584. Meier CR, Derby LE, Jick SS, Jick H. 2000. Angiotensin-converting enzyme inhibitors, calcium channel blockers, and breast cancer. Arch Intern Med 160(3):349–353. Menter A, Cram DL. 1983. The Goeckerman regimen in two psoriasis day care centers. J Am Acad Dermatol 9(1):59–65. Meyer UA, Meier PJ, Hirsiger H, Giger U, Althaus FR. 1980. Induction of drug-metabolizing enzymes by phenobarbitone: structural and biochemical aspects. Ciba Found Symp 76:101–118. Michalowski J. 2002. COX-2 inhibitors: cancer trials test new uses for pain drug. J Natl Cancer Inst 94(4):248–249. Michels KB, Rosner BA, Walker AM, et al. 1998. Calcium channel blockers, cancer incidence, and cancer mortality in a cohort of U.S. women: the nurses’ health study. Cancer 83(9):2003–2007. Miller CT, Neutel CI, Nair RC, Marrett LD, Last JM, Collins WE. 1978. Relative importance of risk factors in bladder carcinogenesis. J Chronic Dis 31(1):51–56. Miller WH Jr., Schipper HM, Lee JS, Singer J, Waxman S. 2002. Mechanisms of action of arsenic trioxide. Cancer Res 62(14):3893–3903. Moller H, Lindvig K, Klefter R, Mosbech J, Moller JO. 1989. Cancer occurrence in a cohort of patients treated with cimetidine. Gut 30(11): 1558–1562. Moller H, Mellemgaard A. 1992. Prostate cancer incidence in cimetidine users. J Natl Cancer Inst 84(5):359–360. Murgo AJ. 2001. Clinical trials of arsenic trioxide in hematologic and solid tumors: overview of the National Cancer Institute Cooperative Research and Development Studies. Oncologist 6 Suppl 2:22–28. National Cancer Institute. 1978. Bioassay of phenazopyridine hydrochloride for possible carcinogenicity. Tech. Rep. Ser. No. 99, 78-1349. Washington, DC, U.S. Government Printing Office. NDC Health. 2003. Rx list: The top 200 prescriptions for 2002 by number of US prescriptions dispensed. Available at http://rxlist.com/top200.htm. Newman TB, Hulley SB. 1996. Carcinogenicity of lipid-lowering drugs. JAMA 275(1):55–60. Nielands J. Microbial metabolism of iron. 1980. In: Jacobs A, Worwood M, eds. Iron in Biochemistry and Medicine, II. London: Academic Press, pp. 529–572. Nies A, Spielberg S. 1996. Principles of therapeutics. In: Hardman JG, Limbird LE, Molinoff PB, Ruddon RW, Gilman AG, eds. Goodman and Gilman’s The Pharmacological Basis of Therapeutics. New York: McGraw-Hill, pp. 43–62. Olsen JH, Boice JD Jr., Fraumeni JF Jr. 1990. Cancer in children of epileptic mothers and the possible relation to maternal anticonvulsant therapy. Br J Cancer 62(6):996–999. Olsen JH, Boice JD Jr., Jensen JP, Fraumeni JF Jr. 1989. Cancer among epileptic patients exposed to anticonvulsant drugs. J Natl Cancer Inst 81(10):803–808. Olsen JH, Johansen C, Sorensen HT, et al. 1999. Lipid–lowering medication and risk of cancer. J Clin Epidemiol 52(2):167–169. Olsen JH, Sorensen HT, Friis S, et al. 1997. Cancer risk in users of calcium channel blockers. Hypertension 29(5):1091–1094. Olsen JH, Wallin H, Boice JD Jr., Rask K, Schulgen G, Fraumeni JF Jr. 1993. Phenobarbital, drug metabolism, and human cancer. Cancer Epidemiol Biomarkers Prev 2(5):449–452. OSHA. 1999. OSHA Technical Manual (TED 1-0.15A). Section IV, Chapter 2. Controlling occupational exposure to hazardous drugs. Available at http://www.osha–slc.gov/dts/osta/otm/otm_vi/otm_vi_2.html#2. Pahor M, Guralnik JM, Ferrucci L, et al. 1996. Calcium-channel blockade and incidence of cancer in aged populations. Lancet 348(9026):493– 497. Pedersen TR, Berg K, Cook TJ, et al. 1996. Safety and tolerability of cholesterol lowering with simvastatin during 5 years in the Scandinavian Simvastatin Survival Study. Arch Intern Med 156(18): 2085–2092. Pedersen TR, Wilhelmsen L, Faergeman O, et al. 2000. Follow-up study of patients randomized in the Scandinavian simvastatin survival study (4S) of cholesterol lowering. Am J Cardiol 86(3):257–262. Penn I. 1977. Development of cancer as a complication of clinical transplantation. Transplant Proc 9(1):1121–1127. Penn I, Brunson ME. 1988. Cancers after cyclosporine therapy. Transplant Proc 20(3 Suppl 3):885–892. Pepin J, Milord F. 1994. The treatment of human African trypanosomiasis. Adv Parasitol 33:1–47. Peraino C, Fry RJ, Staffeldt E. 1971. Reduction and enhancement by phenobarbital of hepatocarcinogenesis induced in the rat by 2-acetylaminofluorene. Cancer Res 31(10):1506–1512.

Perron L, Bairati I, Moore L, Meyer F. 2003. Dosage, duration and timing of nonsteroidal antiinflammatory drug use and risk of prostate cancer. Int J Cancer 106(3):409–415. Pfeffer MA, Keech A, Sacks FM, et al. 2002. Safety and tolerability of pravastatin in long-term clinical trials: Prospective Pravastatin Pooling (PPP) Project. Circulation 105(20):2341–2346. Piper DW. 1981. Cimetidine and cancer. Med J Aust 1(7):327. Piper JM, Tonascia J, Matanoski GM. 1985. Heavy phenacetin use and bladder cancer in women aged 20 to 49 years. N Engl J Med 313(5):292–295. Pittelkow MR, Perry HO, Muller SA, Maughan WZ, O’Brien PC. 1981. Skin cancer in patients with psoriasis treated with coal tar. A 25-year followup study. Arch Dermatol 117(8):465–468. Pommer W, Bronder E, Klimpel A, Helmert U, Greiser E, Molzahn M. 1999. Urothelial cancer at different tumour sites: role of smoking and habitual intake of analgesics and laxatives. Results of the Berlin Urothelial Cancer Study. Nephrol Dial Transplant 14(12):2892–2897. Prescott LF. 1982. Analgesic nephropathy: A reassessment of the role of phenacetin and other analgesics. Drugs 23(1–2):75–149. Ray WA. 2003. Evaluating medication effects outside of clinical trials: Newuser designs. Am J Epidemiol 158(9):915–920. Richmond H. 1959. Induction of sarcoma in the rat by iron-dextran complex. Br Med J (5127):947–949. Roberts MC, Millikan RC, Galanko JA, Martin C, Sandler RS. 2003. Constipation, laxative use, and colon cancer in a North Carolina population. Am J Gastroenterol 98(4):857–864. Robertson RM, Robertson D. 1996. Drugs used for the treatment of myocardial ischemia. In: Hardman JG, Limbird LE, Molinoff PB, Ruddon RW, Gilman AG, eds. Goodman and Gilman’s The Pharmacological Basis of Therapeutics. New York: McGraw-Hill, pp. 759–779. Robinson C, Bell O, Sturdy J. 1960. Possible association of malignant neoplasm with non-dextran injection. Br Med J (2):648. Roeser H. 1980. Iron metabolism in inflammation and malignant disease. In: Jacobs A, Worwood M, eds. Iron in Biochemistry and Medicine, II. London: Academic Press, pp. 605–640. Ron E, Doody MM, Becker DV, et al. 1998. Cancer mortality following treatment for adult hyperthyroidism. Cooperative Thyrotoxicosis Therapy Follow-up Study Group. JAMA 280(4):347–355. Rosenberg L, Rao RS, Palmer JR, et al. 1998a. Calcium channel blockers and the risk of cancer. JAMA 279(13):1000–1004. Rosenberg L, Rao RS, Palmer JR, et al. 1998b. Transitional cell cancer of the urinary tract and renal cell cancer in relation to acetaminophen use (United States). Cancer Causes Control 9(1):83–88. Ross RK, Paganini-Hill A, Krailo MD, Gerkins VR, Henderson BE, Pike MC. 1984. Effects of reserpine on prolactin levels and incidence of breast cancer in postmenopausal women. Cancer Res 44(7):3106–3108. Rossing MA, Scholes D, Cushing-Haugen KL, Voigt LF. 2000. Cimetidine use and risk of prostate and breast cancer. Cancer Epidemiol Biomarkers Prev 9(3):319–323. Rossouw JE, Anderson GL, Prentice RL, et al. 2002. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: Principal results from the Women’s Health Initiative randomized controlled trial. JAMA 288(3):321–333. Rothman KJ. 1986. Causal inference in epidemiology. In: Rothman KJ, ed. Modern Epidemiology. New York: Little, Brown and Company, pp. 7–21. Rothman K, Poole C. 1996. Causation and causal inference. In: Schottenfeld D, Fraumeni JF Jr., eds. Cancer Epidemiology and Prevention. New York: Oxford University Press, pp. 3–10. Rubins HB, Robins SJ, Collins D, et al. 1999. Gemfibrozil for the secondary prevention of coronary heart disease in men with low levels of highdensity lipoprotein cholesterol. Veterans Affairs High-Density Lipoprotein Cholesterol Intervention Trial Study Group. N Engl J Med 341(6): 410–418. Rubinstein I, Langevitz P, Shibi G. 1985. Isolated malignant lymphoma of the jejunum and long-term diphenylhydantoin therapy. Oncology 42(2): 104–106. Sacks FM, Pfeffer MA, Moye LA, et al. 1996. The effect of pravastatin on coronary events after myocardial infarction in patients with average cholesterol levels. Cholesterol and Recurrent Events Trial investigators. N Engl J Med 335(14):1001–1009. Sajadieh A, Storm HH, Hansen JF. 1999. Verapamil and risk of cancer in patients with coronary artery disease. DAVIT Study Group. Danish Verapamil Infarction Trial. Am J Cardiol 83(9):1419–1422, A9. Saltzstein S, Ackerman L. 1959. Lymphadenopathy induced by anticonvulsant drugs and minicking clinically and pathologically malignant lymphomas. Cancer 12:164–182. Sasson AR, Gamagami R, An Z, Wang X, Moossa AR, Hoffman RM. 1999. Cimetidine: an inhibitor or promoter of tumor growth? Int J Cancer 81(5):835–838.

Pharmaceuticals Other Than Hormones Saunders K, Davis R, Stergachis A. 2000. Group Health Cooperative of Puget Sound. In: Strom BL, ed. Pharmacoepidemiology. New York: John Wiley & Sons, pp. 247–262. Schottenfeld D. 1996. Principles and applications of cancer prevention. In: Schottenfeld D, Fraumeni JF Jr., eds. Cancer Epidemiology and Prevention. New York: Oxford University Press, pp. 1391–1409. Selby JV, Friedman GD. 1988. Epidemiologic evidence of an association between body iron stores and risk of cancer. Int J Cancer 41(5):677–682. Selby JV, Friedman GD, Fireman BH. 1989. Screening prescription drugs for possible carcinogenicity: Eleven to fifteen years of follow-up. Cancer Res 49(20):5736–5747. Shaheen NJ, Silverman LM, Keku T, et al. 2003. Association between hemochromatosis (HFE) gene mutation carrier status and the risk of colon cancer. J Natl Cancer Inst 95(2):154–159. Shapiro JA, Williams MA, Weiss NS, Stergachis A, LaCroix AZ, Barlow WE. 1999. Hypertension, antihypertensive medication use, and risk of renal cell carcinoma. Am J Epidemiol 149(6):521–530. Shapiro S. Case-control surveillance. 2000. In: Strom BL, ed. Pharmacoepidemiology. New York: John Wiley & Sons, pp. 209–230. Shapiro S, Parsells JL, Rosenberg L, Kaufman DW, Stolley PD, Schottenfeld D. 1984. Risk of breast cancer in relation to the use of rauwolfia alkaloids. Eur J Clin Pharmacol 26(2):143–146. Sharpe CR, Collet JP, Belzile E, Hanley JA, Boivin JF. 2002. The effects of tricyclic antidepressants on breast cancer risk. Br J Cancer 86(1): 92–97. Shatin D, Drinkard C, Stergachis A. 2000. UnitedHealth Group. In: Strom BL, ed. Pharmacoepidemiology. New York: John Wiley & Sons, pp. 295–306. Shirts SB, Annegers JF, Hauser WA, Kurland LT. 1986. Cancer incidence in a cohort of patients with seizure disorders. J Natl Cancer Inst 77(1):83–87. Shu XO, Gao YT, Linet MS, et al. 1987. Chloramphenicol use and childhood leukaemia in Shanghai. Lancet 2(8565):934–937. Small R, Friedman GD, Ettinger B. 2001. Concomitant medication use in postmenopausal women using estrogen therapy. Menopause 8(2):120– 126. Sonnenberg A, Muller AD. 1993. Constipation and cathartics as risk factors of colorectal cancer: A meta-analysis. Pharmacology 47 Suppl 1:224–233. Sorensen HT, Olsen JH, Mellemkjaer L, et al. 2000. Cancer risk and mortality in users of calcium channel blockers. A cohort study. Cancer 89(1): 165–170. Sorrell TC, Forbes IJ. 1975. Depression of immune competence by phenytoin and carbamazepine. Studies in vivo and in vitro. Clin Exp Immunol 20(2):273–285. Staessen JA, Fagard R, Thijs L, et al. 1997. Randomised double-blind comparison of placebo and active treatment for older patients with isolated systolic hypertension. The Systolic Hypertension in Europe (Syst-Eur) Trial Investigators. Lancet 350(9080):757–764. Stanford JL, Martin EJ, Brinton LA, Hoover RN. 1986. Rauwolfia use and breast cancer: a case–control study. J Natl Cancer Inst 76(5):817–822. Starzl TE, Nalesnik MA, Porter KA, et al. 1984. Reversibility of lymphomas and lymphoproliferative lesions developing under cyclosporin–steroid therapy. Lancet 1(8377):583–587. Stern RS, Laird N, Melski J, Parrish JA, Fitzpatrick TB, Bleich HL. 1984. Cutaneous squamous-cell carcinoma in patients treated with PUVA. N Engl J Med 310(18):1156–1561. Stern RS, Nichols KT, Vakeva LH. 1997. Malignant melanoma in patients treated for psoriasis with methoxsalen (psoralen) and ultraviolet A radiation (PUVA). The PUVA Follow-Up Study. N Engl J Med. 10;336(15): 1041–1045. Stern RS, Zierler S, Parrish JA. 1980. Skin carcinoma in patients with psoriasis treated with topical tar and artificial ultraviolet radiation. Lancet 1(8171):732–735. Stevens JD, Mission GP. 1987. Ophthalmic use of chloramphenicol. Lancet 2(8573):1456. Stevens RG, Beasley RP, Blumberg BS. 1986. Iron-binding proteins and risk of cancer in Taiwan. J Natl Cancer Inst 76(4):605–610. Stevens RG, Graubard BI, Micozzi MS, Neriishi K, Blumberg BS. 1994. Moderate elevation of body iron level and increased risk of cancer occurrence and death. Int J Cancer 56(3):364–369. Stewart JH, Hobbs JB, McCredie MR. 1999. Morphologic evidence that analgesic-induced kidney pathology contributes to the progression of tumors of the renal pelvis. Cancer 86(8):1576–1582. Stolinsky DC. 1991. Prostate cancer and histamine2 antagonists. J Natl Cancer Inst 83(21):1589. Stolley PD, Zahm SH. 1995. Nonhormonal drugs and cancer. Environ Health Perspect 103 Suppl 8:191–196. Stott H, Peto J, Stephens R, Fox W. 1976. An assessment of the carcinogenicity of isoniazid in patients with pulmonary tuberculosis. Tubercle 57(1):1–15.

505

Suissa S. 2000. Novel approaches to pharmacoepidemiology study design and statistical analysis. In: Strom BL, ed. Pharmacoepidemiology. New York: John Wiley & Sons, pp. 785–806. Symmons DP. 1985. Neoplasms of the immune system in rheumatoid arthritis. Am J Med 78(1A):22–28. Taylor TV, Lee D, Howatson AG, Anderson J, Macleod IB. 1981. Gastric cancer and cimetidine. J R Coll Surg Edinb 26(1):34–35. Tenenbaum A, Grossman E, Fisman EZ, et al. 2001. Long-term diuretic therapy in patients with coronary disease: increased colon cancerrelated mortality over a 5-year follow-up. J Hum Hypertens 15(6): 373–379. Thapa PB, Whitlock JA, Brockman Worrell KG, et al. 1998. Prenatal exposure to metronidazole and risk of childhood cancer: A retrospective cohort study of children younger than 5 years. Cancer 83(7):1461–1468. Thun MJ, Henley SJ, Patrono C. 2002. Nonsteroidal anti-inflammatory drugs as anticancer agents: Mechanistic, pharmacologic, and clinical issues. J Natl Cancer Inst 94(4):252–266. Torinuki W, Tagami H. 1988. Incidence of skin cancer in Japanese psoriatic patients treated with either methoxsalen phototherapy, Goeckerman regimen, or both therapies. A 10-year follow-up study. J Am Acad Dermatol 18(6):1278–1281. Trenkwalder P, Hendricks P, Hense HW. 1998. Treatment with calcium antagonists does not increase the risk of fatal or non-fatal cancer in an elderly mid-European population: Results from STEPHY II. Starnberg Study on Epidemiology of Parkinsonism and Hypertension in the Elderly. J Hypertens 16(8):1113–1116. Tsutsui T, Taguchi S, Hasegawa K, et al. 1994. Reserpine-induced cell transformation without detectable genetic effects in Syrian hamster embryo cells in culture. Carcinogenesis 15(1):11–14. Van Den Eeden SK, Friedman GD. 1995. Prescription drug screening for subsequent carcinogenicity. Pharmacoepidemiol Drug Saf 4:275–287. Vaughan TL, Farrow DC, Hansten PD, et al. 1998. Risk of esophageal and gastric adenocarcinomas in relation to use of calcium channel blockers, asthma drugs, and other medications that promote gastroesophageal reflux. Cancer Epidemiol Biomarkers Prev 7(9):749–756. Vezina RM, Lesko SM, Rosenberg L, Shapiro S. 1998. Calcium channel blocker use and the risk of prostate cancer. Am J Hypertens 11(12): 1420–1425. von Poser G, Andrade HH, da Silva KV, Henriques AT, Henriques JA. 1990. Genotoxic, mutagenic and recombinogenic effects of rauwolfia alkaloids. Mutat Res 232(1):37–43. Wallace RB, Sherman BM, Bean JA. 1982. A case-control study of breast cancer and psychotropic drug use. Oncology 39(5):279–283. Wang ZY. 2001. Arsenic compounds as anticancer agents. Cancer Chemother Pharmacol 48 Suppl 1:S72–S76. Weinberg ED. 1984. Iron withholding: A defense against infection and neoplasia. Physiol Rev 64(1):65–102. Weinmann S, Glass AG, Weiss NS, Psaty BM, Siscovick DS, White E. 1994. Use of diuretics and other antihypertensive medications in relation to the risk of renal cell cancer. Am J Epidemiol 140(9):792–804. Weisburger JH, Williams GM. 2000. The distinction between genotoxic and epigenetic carcinogens and implication for cancer risk. Toxicol Sci 57:4–5. Weiss SR, McFarland BH, Burkhart GA, Ho PT. 1998. Cancer recurrences and secondary primary cancers after use of antihistamines or antidepressants. Clin Pharmacol Ther 63(5):594–599. West SL, Strom BL, Poole C. 2000. Validity of pharmacoepidemiology drug and diagnosis data. In: Strom BL, ed. Pharmacoepidemiology. New York: John Wiley & Sons, pp. 661–706. Westerdahl J, Olsson H, Masback A, Ingvar C, Jonsson N. 1996. Risk of malignant melanoma in relation to drug intake, alcohol, smoking and hormonal factors. Br J Cancer 73(9):1126–1131. White SJ, McLean AE, Howland C. 1979. Anticonvulsant drugs and cancer. A cohort study in patients with severe epilepsy. Lancet 2(8140):458– 461. Williams RR, Feinleib M, Connor RJ, Stegens NL. 1978. Case-control study of antihypertensive and diuretic use by women with malignant and benign breast lesions detected in a mammography screening program. J Natl Cancer Inst 61(2):327–335. Witztum J. 1996. Drugs used in the treatment of hyperlipoproteinemias. In: Hardman JG, Limbird LE, Molinoff PB, Ruddon RW, Gilman AG, eds. Goodman and Gilman’s The Pharmacological Basis of Therapeutics. New York: McGraw-Hill, pp. 875–897. Wood AJ, Stein CM, Woosley R. 1998. Making medicines safer—the need for an independent drug safety board. N Engl J Med 339(25):1851– 1854. Wormsley KG. 1984. Assessing the safety of drugs for the long-term treatment of peptic ulcers. Gut 25(12):1416–1423.

506

PART III: THE CAUSES OF CANCER

Yu MC, Mack TM, Hanisch R, Cicioni C, Henderson BE. 1986. Cigarette smoking, obesity, diuretic use, and coffee consumption as risk factors for renal cell carcinoma. J Natl Cancer Inst 77(2):351–356. Yuan JM, Castelao JE, Gago-Dominguez M, Ross RK, Yu MC. 1998. Hypertension, obesity and their medications in relation to renal cell carcinoma. Br J Cancer 77(9):1508–1513.

Zacharski LR, Moritz TE, Haakenson CM, et al. 1990. Chronic calcium antagonist use in carcinoma of the lung and colon: A retrospective cohort observational study. Cancer Invest 8(5):451–458. Zheng W, Blot WJ, Liao ML, et al. 1987. Lung cancer and prior tuberculosis infection in Shanghai. Br J Cancer 56(4):501–504.

26

Infectious Agents NANCY E. MUELLER, BRENDA M. BIRMANN, JULIE PARSONNET, MARK H. SCHIFFMAN, AND SHERRI O. STUVER

T

here is substantial evidence that infectious agents play a causal role in a variety of human malignancies. These cancers are surprisingly diverse and include the liver, cervix, stomach. nasopharynx, bladder, and bile duct as well as Kaposi sarcoma (KS) and several lymphomas. In terms of the magnitude of associated cancer cases, the most important oncogenic agents are viruses. These include two hepatotropic viruses, hepatitis B virus (HBV) and hepatitis C virus (HCV) (IARC, 1994a); two herpesviruses, Epstein-Barr virus (EBV) and human herpesvirus 8/KS virus (HHV8) (IARC, 1997); the human papillomaviruses (HPV) (IARC, 1995); and the retrovirus human T-cell lymphotropic virus type I (HTLV-I) (IARC, 1996). In addition, two chronic parasitic infections, the liver fluke Opisthorchis viverrini and the blood fluke Schistosoma haematobium, have long been thought to be human carcinogens, and more recently a bacterium, Helicobacter pylori, has been implicated (IARC, 1994b) (Table 26–1). The public health impact of the oncogenic effects of these infections is substantial. It has been estimated that about 15% of all incident cases of cancer worldwide are attributable to infections; this accounts for 23% of all malignancies in economically developing countries and 7% in developed countries (Parkin et al., 1999). Because these malignancies tend to occur relatively early in life (Doll, 1978) (Fig. 26–1), the impact on person-years of life lost due to cancers caused by oncogenic infections is somewhat greater than for other carcinogenic exposures with similar attributable risk percent. Table 26–1 summarizes the major oncogenic infections. Several are quite prevalent infections, such as EBV, HPV, and H. pylori. Others are relatively uncommon infections such as HTLV-I, which is microendemic in relatively isolated populations in southern Japan, the Caribbean, South America, and parts of Africa. Endemic infection with O. viverrini is found in parts of Asia and the former Soviet Union, while that of S. haematobium is scattered throughout Africa. The epidemiological patterns of HHV8, HBV, and HCV infections are more variable. These are widely spread throughout the world, generally at a low prevalence but with pockets of relatively high endemicity. The evaluation of causality for these infectious agents as human carcinogens is difficult given their ubiquitous nature, the substantial length of time between infection and the occurrence of cancer, the nature of cofactors, and the rarity of malignancy among those infected (Evans and Mueller, 1990). That being said, the International Agency for Research on Cancer (IARC) has completed a series of monographs that reviewed the evidence for carcinogenicity for each. The Working Groups concluded that for all of the agents there is “sufficient evidence” for classification as a human carcinogen, with the exception of HHV8, which was classified as “probably carcinogenic” to humans. These judgments have rested heavily on the epidemiological and biomarker evidence. The agents share several biological characteristics. Each has the capacity to become persistent, creating a carrier state in which periodic or continuing transmission of the agent to new hosts occurs (Ahmed et al., 1996). This can occur as part of the normal life cycle of the agent, such as for herpesviruses EBV and HHV8, which persist as episomes in infected cells. Alternatively, the agent has the capacity to become a conditionally persistent or chronic infection, such as HBV, which can become a chronic infection if not initially cleared by an effective immune response.

For each of these agents, the occurrence of malignancy is a relatively uncommon sequela. This observation underscores the importance of the dynamic host–agent interaction. When host cellular immunity is severely compromised, for example, by induced immunosuppression in organ transplantation or by infection with the human immunodeficiency virus (HIV), the risk of malignancy secondary to the loss of control of latent oncogenic viral infections is substantial (Mueller, 1999). For chronic HBV and HCV infections, interferon (IFN) therapy to enhance cellular immunity sometimes clears the persistent infection and substantially reduces the risk of liver cancer. In most cases, the host immune competency at the time of infection is important in either clearing the infection, such as for HCV, or generating an adequate cytotoxic T-lymphocyte (CTL) response for setting a favorable host response to control latent infections, such as HPV. Factors that influence the initial host–agent interaction include age and route of infection, gender, and the presence of coinfections and other comorbidity. The general effect of age of infection is reflected in the unusual age-incidence patterns characteristic of these tumors (Fig. 26–1). A paradigm of cytokine-mediated immune function is useful in this setting (Mosmann and Moore, 1991; Lucey et al., 1996; Romagnani, 2000). The paradigm describes two counter-regulating cascades of cytokine responses to specific antigens. Cytokines are secreted by helper T-cells upon presentation of antigens and modulate the resulting immune response to an agent (along with cytokines secreted by other cells). Together they form a complex system of selective stimulation and cross-regulation that results in an immune response to an antigen or family of antigens. “Type 1” (or Th1) responses involve what have long been thought of as cell-mediated immunity, with inflammation and CTL responses. “Type 2” (or Th2) responses induce immunoglobulin (Ig)- based immunity. These two types of T-cell reaction are mutually counter-regulatory, with specific immune responses determined by a relative predominance of either type 1 or type 2 cytokines and effectors (Romagnani, 2000). Viewing immunity as a spectrum ranging from a predominantly humoral response to a predominately cellular immune response, there is a growing body of evidence linking polarized (likely systemic) or imbalanced responses with a variety of malignancies, autoimmune diseases, and allergic states (Lucey et al., 1996; Kero et al., 2001; Bach, 2002; Simpson et al., 2002). These characterizations of diseases by “type” of immune response have involved in vitro measures of cytokines, such as IL-10, a major type 2 cytokine, or in some cases, serologically detectable products of T-cells, such as sCD30, also a type 2 biomarker (Lucey et al., 1996). A type 1 inflammatory response is typically directed against viral, microbial, and other intracellular infections and is also responsible for delayed-type hypersensitivity. However, excessive type 1 immunity can cause tissue damage and some autoimmune conditions, such as multiple sclerosis and rheumatoid arthritis. In contrast, a type 2 response promotes IgE production and eosinophil function, which are involved in the pathogenesis of allergic inflammation and a variety of atopic conditions. At birth, the immune system of infants is skewed toward type 2 immunity. The counterbalancing type 1 immunity matures with age, likely due to infection with Th1 stimulatory antigens, such as viruses (Barrios et al., 1996; Krampera et al., 1999). The immune system of children who are sheltered from early contact with viruses may continue to be biased toward type 2 immunity, which promotes the

507

508

PART III: THE CAUSES OF CANCER

Table 26–1. Major Human Infection–Associated Malignancies Malignancy

Agent (Group)

Carcinomas Bladder Cervical Hepatocellular Bile duct Nasopharynx Stomach Lymphomas Adult T-cell Burkitt Hodgkin Sarcoma Kaposi

Schistosoma haematobium (blood fluke) HPV (papillomavirus) HBV (hepadnavirus) HCV (flavivirus) Opisthorchis viverrini (liver fluke) EBV (herpesvirus) Helicobacter pylori (bacterium) HTLV-I (retrovirus) EBV (herpesvirus) EBV (herpesvirus) HHV8 (herpesvirus)

EBV, Epstein-Barr virus; HBV, hepatitis B virus; HCV, hepatitis C virus; HHV8, human herpesvirus 8; HPV, human papillomavirus; HTLV-I, human T-cell lymphotropic virus type I.

Burkitt Hodgkin lymphoma

Figure 26–1. Age-specific incidence rates for major virus-associated malignancies. Adult T-cell leukemia/lymphoma (ATL): males, 1986–1987, Kyushu (Japan); Burkitt lymphoma: males, 1961–1975, West Nile District (Uganda); Hodgkin lymphoma: males, 1986–1990, SEER Program; liver, males, 1983–1987, Shanghai (China); cervix: 1982–1986, Cali (Colombia); nasopharynx: males, 1983–1987, Chinese (Singapore). (Source: Parkin et al., 1992.)

development of asthma and allergic conditions (Cookson and Moffatt, 1997; Shirakawa et al., 1997; Martinez and Holt, 1999). Although the infection-associated malignancies have generally been related to diminished cellular immunity (type 1), this may actually be accompanied by immune activation—as seen in HIV infection. Optimally, humans should have a “. . . well balanced Th1 and Th2 response, suited to the immune challenge” (Berger, 2000). A common feature of the natural history of all these oncogenic infectious agents is that risk of malignancy is associated with a chronic or increased level of replication, raising the probability of secondary genetic damage to the affected tissues. The mechanisms of oncogenesis are fairly well understood for some infections, whereas for others they are not at all defined. The oncogenic viruses have evolved strategies to interfere with cell cycle control and, in some cases, to induce transformation of infected cells. For example, in HPV-16 (the prototypic high-risk strain) infection, the viral product E6 inactivates the p53 tumor suppressor protein by inducing its rapid degradation. Similarly, the E7 protein binds with the tumor suppressor retinoblastoma protein, resulting in inactivation and degradation. E7 also activates cyclin-dependent kinase 2, which is involved in the transition from the G1 to the S phase of the cell cycle. In HPV-associated cervical cancer, E6 and E7 are usually selectively expressed (Alani and Munger, 1998). Similar viral protein–host gene interactions have been described for the EBV (Spender et al., 1999). In HTLV-I infection, the product of the regulatory gene Tax transactivates a number of host oncogenes while down-regulating the expression of the DNA repair-enzyme gene for b-polymerase (IARC, 1996). HTLV-I, EBV, HBV, HCV, and HHV8 all have evolved mechanisms to activate the NFKB pathway that transcriptionally activates multiple genes that influence cell cycle and enhance the survival of infected cells (Hiscott et al., 2001). The mechanisms by which H. pylori and the oncogenic flukes induce malignancy are not well defined, although a role for chronic inflammation is often cited. Chronic inflammation may also contribute to the oncogenesis of HCV and HBV. Because the great majority of carriers of these oncogenic infections do not develop subsequent malignancy (or other serious disease), a central problem for the epidemiologist is to define the natural history of infection and identify those factors that are related to the development of cancer. This approach was introduced to cancer epidemiologists by Francis et al. (1979, 1981) who described the natural history of feline leukemia (FeLV), a common retroviral infection among outbred cats in the United States at that time (Fig. 26–2), and noted its similarities to that of HBV infection. With FeLV infection, it is the small proportion of infected cats that are chronically viremic that are

Figure 26–2. An epidemiologic model of the natural history of FelV infection in a steady-state population of cats. (Source: Francis et al., 1979.)

Infectious Agents at high risk of developing a FeLV-positive leukemia or lymphoma. The determinants of chronic viremia, as identified by a peripheral blood smear assay, include early age of infection and/or heavy exposure by infectious saliva from multiple viremic cats. The evolution of malignancy occurs over a relatively long period of time. The understanding of the natural history of FeLV infection contributed to the development and implementation of a vaccine program. This model illustrates many of the principles that characterize oncogenic infections of humans. Natural history research for human oncogenic infections requires informative biomarkers of the agent (such as viral load) and host (such as abnormal antibody pattern) interaction. Interaction with other oncogenic exposures, such as immunosuppressive ultraviolet (UV) light, tobacco use, and unhealthy diet, should be considered. Prospective cohorts of carriers with serial biospecimens provide a rich venue for natural history studies. However, the small number of expected cases, necessitating the identification of intermediary end points, hampers such studies. Such data can be maximized by collaboration among cohorts. Gaining understanding of the natural history of these infections should help identify those carriers who are at high risk of malignancy and guide the development of interventions. This chapter summarizes the biological and epidemiologic features of each of the major oncogenic infections, beginning with the viruses, followed by H. pylori, and with a brief summary of the relevant parasites.

EPSTEIN-BARR VIRUS The EBV is a ubiquitous herpesvirus that infects the great majority of the world’s population. EBV establishes latent infection in Blymphocytes and can induce proliferation of latently infected cells. Although most EBV infections are benign, its oncogenic potential is well documented (Anagnostopoulos and Hummel, 1996). An IARC Working Group on the evaluation of carcinogenic risks to humans concluded that there is sufficient evidence for the carcinogenicity of EBV in the causation of Burkitt lymphoma (BL), Hodgkin lymphoma (HL), nasopharyngeal carcinoma (NPC), immunosuppression-related lymphoma, and sinonasal angiocentric T-cell lymphoma (IARC, 1997). A member of the gamma-herpesvirus family (gamma-one subgroup), the EBV (also termed human herpesvirus 4) is a linear, doublestranded 172-kb DNA molecule that codes for at least 100 viral proteins (Miller, 1990; Kieff, 1996). Like all herpes viruses, EBV consists of an outer envelope with external, glycoprotein spikes, a viral capsid, and a DNA core. There are two major EBV strains (1 and 2), which differ somewhat in their geographic distribution. EBV was identified in 1964 in a culture established from a BL tumor (Epstein et al., 1964; Epstein, 1999). It was soon found that Blymphocytes infected with EBV in vitro become “immortalized,” being capable of continual growth. In this process, the cells enlarge into blastoid forms, lose their contact inhibition, and develop new antigens on their surface. It was also determined that lymphoma could be induced in vivo in cotton-top marmosets and owl monkeys by the injection of EBV or EBV-infected lymphocytes (Epstein et al., 1973; Miller, 1974; Miller et al., 1977).

Natural History of EBV Infection The great majority of EBV infections are transmitted by oral contact with virus shed in saliva from an EBV carrier. EBV enters and multiplies in B-lymphocytes via the C3d complement receptor (CD21) (Fingeroth et al., 1984; IARC, 1997). It can also multiply in epithelial cells of the oropharynx, parotid gland, and uterine cervix (Miller et al., 1987). The presence of the virus in cervical secretions (Sixbey et al., 1986) suggests the possibility of sexual transmission. Transfusiontransmitted EBV infections also occur but rarely result in any serious disease in immune competent recipients. Infection with EBV persists for life. In infected cells, EBV establishes a latent infection or can undergo a lytic cycle of virus replication that results in cell death. The latent state is the predominant form of viral infection in which the EBV DNA persists as a closed, circular, nuclear molecule (episome) which associate randomly with chromosomes in a subpopulation of circulating B-lymphocytes-primarily

509

resting memory B-cells, and in the bone marrow (Rickinson and Kieff, 2001; Thorley-Lawson and Gross, 2004). In latently infected cells, viral genes coding for the Epstein-Barr nuclear proteins EBNA-1, -2, -3A, -3B, -3C, -LP can be expressed. Other latent proteins include the latent membrane proteins LMP-1, LMP-2A/-2B; and two abundant, small, nonpolyadenylated RNAs, EBER-1/-2. The functions of these latent proteins are complex and multiple, serving to transform infected cells and insuring the persistence of the virus in the infected host (Kieff and Rickinson, 2001). In the lytic cycle, the structural proteins of the virus are expressed as progeny virus particles are produced. These proteins are recognized as antigens by antibodies, primarily against the viral capsid antigen (VCA) and a complex of early antigens (EA); antibodies against the EA include the so-called diffuse (EA-D) and restricted (EA-R) forms. After a primary EBV infection, the virus remains latent in a small percentage of B-lymphocytes. It is also excreted intermittently from lytic infections of oropharyngeal epithelial cells (Rickinson and Kieff, 2001). At any time, perhaps 20% or more of EBV antibody positive persons excrete EBV in the oropharynx, maintaining infectious spread within the population (Niederman and Evans, 1997; Ling et al., 2003). It appears that viral replication does not normally occur in infected peripheral lymphocytes in resolved infection (Prang et al., 1997). Changes in immune competence may result in reactivation of latent EBV infection in B-lymphocytes. This reactivation is usually asymptomatic but, if sufficiently vigorous, it may lead to lymphoproliferative disease and B-cell lymphomas, as has been observed in HIV-infected patients, renal-transplant patients, and other severely immune-depressed individuals. Host control of EBV infection is primarily maintained by EBVspecific CTLs. In acute infection, CTLs against lytic antigens predominant (Khanna and Burrows, 2000). However, as the primary infection resolves, CTLs against the latent proteins become predominant. These are directed primarily against the EBNA-2, -3a, -3b, and -3c (IARC, 1997; Khanna and Burrows, 2000). CTLs against LMP-2 and, to a lesser degree, LMP-1 are also found (Meu et al., 2002). Although carriers produce antibodies to a variety of lytic and latent antigens, antibodies appear to afford very little or no protection in control of established infection. Rather, the prevalence and spectrum of antibodies appear to be an indirect index of the prevalence of viral protein expressed in the carrier. This is illustrated in the sequential cascade of antibody response that has been documented in seroconverters, primarily patients with infectious mononucleosis (IM) (Fig. 26–3). Typically, these patients first develop antibodies against the VCA and the EA, reflecting viral replication, and against the latent protein EBNA-2 that is expressed on EBV-transformed lymphoblastoid B-cells. This is followed by the appearance of antibodies to EBNA-1, expressed in both lytic and latent infection and the sole latent protein expressed on infected resting B-cells. The appearance of antiEBNA-1 parallels the generation of specific CTLs against the EBV latent proteins and the resolution of early infection. Antibodies against EBNA-2 subsequently diminish, resulting in an EBNA-1/EBNA-2 ratio of >1.0 (Rickinson and Kieff, 2001). A continuing EBNA1/EBNA-2 ratio of £1.0 suggests defective control of latent EBV infection (Henle et al., 1987). The serology of EBV reactivation has been evaluated prospectively in therapeutically immunosuppressed populations. In general, reactivation is marked by an increase in IgG titers against VCA and EA-R. The presence of IgM against VCA has also been noted (Ho et al., 1985). In addition, there is a decrease in titer or loss of antibody against EBNA (List et al., 1987; Riddler et al., 1994). EBV load in blood specimens varies in association with disease. Viral load has been measured in peripheral blood mononuclear cells (PBMCs) and in serum or plasma. In PBMCs, increased viral load has been detected in patients with IM (Tierney et al., 1994) and posttransplant lymphoproliferative disease (PTLD) (Lucas et al., 1998). In PTLD, the increases in viral load in PBMCs are sometimes dramatic, as is the fall in viral load after an intervention that restores cellular immunity, such as adoptive T cell immunotherapy (Rooney et al., 1995). “Naked” EBV DNA is not normally detected in serum or plasma drawn from healthy seropositives. It has been detected in the serum of patients with IM (Gan et al., 1994), PTLD (Swinnen et al.,

510

PART III: THE CAUSES OF CANCER In general, most infections occur in childhood. Factors influencing the probability of childhood infection with EBV include the amount of exposure to other children within the household, neighborhood, or in group care settings; the level of hygiene; the sharing and prechewing of food; and perhaps oral contact with adults (Evans and Niederman, 1989). For those individuals who escape childhood infection, the onset of dating and intimate kissing in adolescence is associated with the incidence of EBV infection—with a high probability of IM (Evans and Niederman, 1989). EBV infection is prevalent throughout the world, even in the most remote tribes along the Amazon (Black et al., 1974). Early seroepidemiologic studies established that the prevalence of infection varied with general socioeconomic conditions (Fig. 26–4). In economically developing populations, infection generally occurs early in life. In Ghana, for example, about 80% of the children are infected by the time they reach 18 months of age (Biggar et al., 1978). Infection typically occurs later in life in economically developed countries where transmission to children is less frequent because of less dense housing, smaller families, and better hygiene. For example, about half of entering American college students in the early 1970s lacked EBV antibody (Evans and Niederman, 1989). With the more recent social changes in the family in the United States including the widespread employment of women, and the increasing use of day care and nursery school for young children (Jamieson et al., 2001), this has likely changed.

EBV Biomarkers Figure 26–3. Parameters of the EBV-host balance during the incubation, symptomatic, and convalescent phases of acute IM, during subsequent asymptomatic virus carriage, and after the imposition of immunosuppressive therapy on an existing virus carrier state. The band represents the broad range within which most infected individuals lie; where the band extends to the baseline indicates that the relevant responses are detectable in some, but not all, infected individuals. (Source: Adapted from Rickinson AB, Kieff E. 2001. In: Knipe DM, Howley PM, eds. Fields Virology. 4th ed. Philadelphia: Lippincott–Raven, p. 2584. Based on Henle and Henle, 1979.)

1998), NPC (Mutirangura et al., 1998), and HL (R. Ambinder et al., personal communication). The viral DNA, detected in the plasma or serum of patients with tumors, does not reflect the presence of virion DNA but rather cellular DNA released by apoptosis or necrosis of tumor cells (Lo et al., 2000). The host response to primary EBV infection varies with age at the time of infection. When infection occurs early in life, it is usually subclinical (Chan et al., 2001). This is common in most underdeveloped countries and in developed populations among people with poorer socioeconomic living conditions. When infection with the virus is delayed until older childhood or young adult life, IM occurs in perhaps half of the primary infections (Evans and Niederman, 1989). When the infection is accompanied clinically with IM, most of the symptoms reflect the vigorous CTL response to EBNA-positive, infected B-lymphocytes. Children differ from adults in their antibody response to primary infection, as they rarely develop clinical IM, antiEA-D, or heterophile (diagnostic of IM) antibodies. The serologic mark of an established, latent EBV infection for all ages is that of IgG antibodies against VCA and against EBNA. The level (or titer) of these antibodies tends to remain constant for a patient’s lifetime, once the primary infection is resolved. After IM, these levels are established at a relatively high level (Niederman et al., 1970; Evans and Niederman, 1989). This is also true among children who are infected as infants (Melbye et al., 1984). Failure to develop anti-EBNA, after seroconversion, is associated with continuing viremia and development of lymphoproliferative disease (Purtilo et al., 1982).

Biomarkers for EBV DNA/RNA and viral-encoded proteins and antibodies have been used in epidemiologic studies of EBV-associated malignancies. For detection of the viral genome fragments in tumor tissue, the techniques that were initially used include the slot and Southern blot (Weiss et al., 1987), polymerase chain reaction (PCR) (Uhara et al., 1990), and in situ hybridization (Anagnostopoulos et al., 1989) assays. Other than the latter, these methods do not distinguish whether any EBV detected is from the neoplastic cells or infiltrating lymphocytes. What has become commonly used to characterize the EBVpositivity status of a tumor is in situ hybridization to detect the abundant EBERs that are actively transcribed in latently infected cells (Wu et al., 1990). The EBER-1 probes are viewed as the most sensitive in detecting EBV genome in paraffin-embedded tissues and are often combined with in situ hybridization with monoclonal antibodies to test for the presence of the latent viral gene products, such as LMP-1 (Pallesen et al., 1991; Gulley et al., 2002). The clonality of the episomal EBV genome can be determined using Southern blot probes for the EBV terminal repeats (Raab-Traub and Flynn, 1986). The systemic EBV viral load in peripheral blood mononuclear cells can be quantified by a variety of PCR techniques (Lechowicz et al., 2002). A surprisingly robust new biomarker is cell-free EBV DNA,

Epidemiology of Infection The epidemiology of EBV infection within a population reflects the prevalence of factors that influence the age of exposure to the virus.

Figure 26–4. Age distribution of antibodies to EBV in East Africa and the United States. (Source: Henle et al., 1969.)

511

Infectious Agents measured by PCR in serum or plasma (Mutirangura, 2001; van Esser et al., 2001; Lechowicz et al., 2002). We have found detectable levels in serum specimens collected several years preceding the development of EBV-positive HL (R. Ambinder et al., personal communication). The quantification of antibody levels (primarily IgG and IgA) against the major EBV antigens, VCA, EA complex (less so EA-R and EA-D), and EBNA complex, has been a major empiric tool in establishing an association with EBV and cancer and providing insight into the virus–host interaction. A schema developed by Rickinson and Kieff (2001) of the relative levels of several antibodies, after primary infection in asymptomatic carriers, and after immune suppression is useful in interpreting antibody profiles (Fig. 26–3). Historically, measurement of EBV antibodies has relied on immunofluorescence (IF)based tests (Reedman and Klein, 1973; Henle et al., 1974). Because the EBV-infected cell lines used in these IF assays may differ in the amount of antigen produced, titer levels from different laboratories may not be comparable but should be internally consistent, as evidenced in de Thé et al. (1978). Assays have been developed to measure anti-EBNA-1 and EBNA-2 (Lennette et al., 1993). Antibodies against other informative antigens include anti-ZEBRA (as a mark of viral reactivation) (Drouet et al., 1999) and EBV-specific DNase, which is predictive of NPC (Chien et al., 2001). Although enzyme immunoassay (EIA) methods are more commonly used in clinical settings to measure antibodies against the major antigens, such as VCA and EA, they are not widely used in epidemiologic studies, and the comparability of results with those based on IF assays is not always clear. Unfortunately, few laboratories now exist that are able to carry out quantitative determinations of EBV antibodies over the range of antigens in the volume required for epidemiologic studies.

The Major EBV-Associated Malignancies Burkitt Lymphoma Burkitt lymphoma (BL) is a malignancy that occurs endemically, primarily among young children, especially boys, in central Africa and New Guinea and sporadically among all ages in other parts of the world. Endemic areas correlate strongly with the presence of holoendemic malaria (Fig. 26–1). The risk of BL is thought to result from the enhanced proliferation of B-lymphocytes by early infection with EBV, interacting with the ongoing mitogenic effect of malaria. In these areas, mainly children are affected, with a peak incidence at about 8 years of age. The jaw and abdominal organs are the most frequent sites of the malignancy (Evans and Mueller, 1997). Elsewhere, BL is a rare tumor. Approximately one in three of these nonendemic BL cases are EBV-positive, compared to nearly 100% of the endemic tumors (Evans and Mueller, 1997). Of note, BL does occur as an opportunistic malignancy in HIV infection, more commonly in children (Biggar and Rabkin, 1992).

Independent of geographic origin, essentially all BL show one of the three following reciprocal chromosomal translocations involving the long arm of chromosome 8q: these include either 14q (75% of the cases), or the short arm of chromosome 2p (9%), or the long arm of chromosome 22q (16%). These translocations have been found consistently in more than 100 established BL cell lines, irrespective of EBV positivity. The translated portion of chromosome 8 includes the c-myc oncogene, which, when translocated adjacent to the heavy chain genes on 14q, or the K light chain gene on 2p, or the D light chain gene on 22q, becomes activated (derepressed) during antibody generation. Alteration of c-myc activity appears to be an essential event in the development of the malignant cell clone whose replication leads to BL (Leder, 1985a,b). The association of EBV with BL is defined by the demonstration of clonal episomal EBV in tumors, characteristically expressing only EBNA-1 (Klein, 1994). In terms of antibody patterns, a significantly higher prevalence of elevated IgG antibody titers to the VCA, EBNA, and, to a lesser extent, to EA-R, occurs than in age and gendermatched controls from the same area (Henle et al., 1969) (Table 26–2). The most important epidemiologic study that linked EBV infection with the risk of BL was a prospective study based in a cohort of 42,000 children, aged less than 9 years, in the West Nile district of Uganda, under the auspices of IARC (de Thé et al., 1978). Baseline blood specimens were obtained and stored. Fourteen BL cases occurred in the first 5 years of follow-up, and two additional cases were later identified (Geser et al., 1982). All patients with BL had EBV antibody present in the initial serum sample, taken between 7 and 54 months prior to diagnosis, and 12 of 13 EBV-associated BL cases had pretumor VCA antibody titers as high as, or higher, than any control bled at the same time. The risk for development of BL in children with titers two dilutions or more above the geometric mean titer (standardized for age, sex, and area) of the controls was estimated to be 30 times higher than those with normal levels. In addition, 9 of the 10 confirmed cases, from whom biopsies had been obtained at the time of diagnosis, had detectable EBNA or EBV DNA in the tumor tissue, including one case in whom the pretumor antibody level was normal. Antibodies to other herpesviruses as well as to measles virus were not elevated in the baseline bleeding, indicating a specificity of the EBV association. This study serves as a prototype for establishing the causal association of EBV and other oncogenic infections with subsequent malignancy.

Nasopharyngeal Carcinoma Nasopharyngeal carcinoma (NPC) occurs worldwide. These tumors are generally histologically classified as poorly or undifferentiated carcinomas and exhibit substantial infiltration of lymphocytes (Agathanggelou et al., 1995). The highest incidence is in the Far East, primarily in persons of Chinese descent and in related populations, wherever they may migrate (de Thé et al., 1989). Rates differ among various Chinese populations in the same area and are highest among

Table 26–2. Pattern of EBV Biomarkers in Associated Malignancies Marker Antibody VCA IgG IgA IgM EBNA Early antigen Diffuse Restricted Molecular Clonality Latent phenotype

Burkitt Lymphoma

Nasopharyngeal Carcinoma

Hodgkin Lymphoma

≠≠ — — —

≠ ≠≠ — ≠≠

≠ ≠ Ø ≠≠

— ≠*

≠ (IgA) —

≠ ±

Yes EBNA1+

Yes EBNA1+/LMP1±/ LMP2A+

Yes EBNA1+/LMP1/2A+

EBNA, Epstein-Barr nuclear protein; EBV, Epstein-Barr virus; LMP, latent membrane protein; VCA, viral capsid antigen. *Predictive of recurrence after treatment.

512

PART III: THE CAUSES OF CANCER

those from the South. The age-specific incidence among Chinese in Hong Kong and Singapore shows a rising peak from age 20 to 24 to about age 50, then drops off thereafter (Fig. 26–1). This type of age curve suggests an infectious etiology, with exposure occurring early in life (Doll, 1978). In Sweden in the 1980s, by contrast, the rise occurred some two decades later and continued through the end of life (Parkin et al., 1992). In the United States, a trimodal distribution has been noted by Green et al. (1977), who suggest that this may reflect varying etiologies operating in different age groups in a racially mixed population. The epidemiology of NPC is described in detail elsewhere (see Chapter 31). As in the case of BL, high antibody titers to VCA-IgG mark an association with EBV in NPC (Henle et al., 1970, 1973). Such high titers are found in more than 80% of NPC patients. Antibody to EA is also elevated, but usually it is the diffuse form (EA-D) (Table 26–2). The level of these antibodies increases as the stages of disease progress (Henle et al., 1977). Mathew et al. (1994) found that of 100 NPC patients with anti-VCA-IgA, 75% had IgG antibodies against ZEBRA, indicative of viral replication. NPC patients also have high antibody levels against EBNA (de Thé et al., 1989). Chatani et al. (1991) have reported that NPC cases have high titers against both EBNA-1 and EBNA-2. Evidence of EBV in NPC tissue has also been found regularly in biopsies (including carcinoma in situ) by several molecular techniques and occurs predominantly in epithelial cells, not in the infiltrating lymphocytes (Wolf et al., 1973; Klein et al., 1974; Pagano et al., 1975; Yeung et al., 1993). The viral genome is clonal (Raab-Traub and Flynn, 1986) and expresses a restricted pattern of viral latent proteins, with only EBNA-1, LMP-2A, and inconsistently LMP-1 detected (Fåhraeus et al., 1988; Henssinger et al., 2004) (Table 26–2). Serologically detectable cell-free EBV DNA can be found at diagnosis, and its return after treatment is predictive of recurrence (Mutirangura, 2001). A key study that provided strong evidence that EBV plays a primary causal role in NPC was conducted by Pathmanathan et al. (1995). The investigators screened 5362 nasopharyngeal biopsy samples from Malaysia as a means of identifying preinvasive lesions, including dysplasia and carcinoma in situ; they identified 11 biopsies with such lesions that lacked adjacent invasive carcinoma. Of these, eight had tissue available for further analysis. Three additional specimens were obtained from Guangzhou, China, for a total of 11 specimens. All 11 specimens were positive for the presence of EBV. Of the seven with sufficient tissue to test for viral clonality, six were found to have clonal EBV DNA in the lesion. A prospective serologic study conducted in Taiwan examined the role of EBV antibodies as predictive of subsequent NPC (Chien et al., 2001). The subjects were men residing in areas with the highest rates of NPC. A total of 9688 men were enrolled who had no evidence of NPC and who provided blood specimens. All specimens were tested for anti-VCA-IgA and anti-EBV DNase antibodies. The subjects were followed for up to 16 years, and 22 incident cases of NPC were identified, who were diagnosed after 1 year from enrollment. The prevalence at baseline of anti-EBV DNase was 12% and that for anti-VCA-IgA was 1.2%. The relative risk (RR) for developing NPC for persons with both markers at baseline, compared to those with neither, was 20.7 (95% confidence interval, 2.6–162). Other major risk factors for NPC include genetic susceptibility and related environmental exposures, especially dietary factors (IARC, 1997).

Hodgkin Lymphoma Hodgkin lymphoma (HL) is a relatively common malignancy of young adults in Westernized populations, where the age-incidence curve is generally bimodal (Fig. 26–1) (see Chapter 45 for a detailed review). Since EBV was first identified, it was considered a candidate virus for this lymphoma. The hypothesis that age at infection with EBV may play an important role in the development of HL was suggested by the epidemiological similarities between HL among young adults and IM in economically advantaged populations. For both diseases, risk has been associated with higher social class and small family size

(Gutensohn and Cole, 1977, 1981). These similarities suggest that exposure to the causative agent is delayed until young adult life. It has been almost always found that people with a history of IM have about a threefold increase in their risk of developing HL. In addition, a larger proportion of HL cases have elevated antibody titers to several EBV antigens, primarily VCA-IgG and EA-D, as compared to controls. Generally, some 30% to 40% of HL cases are found to have elevated titers. This has been a consistent finding in many studies over different geographic areas (Mueller, 1987). A pilot study based on two cases of HL occurring prospectively among 25,802 persons with banked serum samples showed that in comparison to matched-controls, EBV antibodies were significantly elevated prior to diagnosis, in contrast to three other herpesviruses whose antibody titers were not elevated (Evans and Comstock, 1981). This study was then extended to include sera for more than 240,000 persons from five serum banks. In this extended study, 43 persons who developed HL between 1 and 13 years after blood collection were identified and matched with 96 controls (Mueller et al., 1989). EBV antibody analysis of these sera confirmed the findings of the pilot study, in that the proportion of patients with elevated antibody against VCA-IgG, IgA, and EBNA complex was significantly higher than that of the controls; this was associated with a three- to sevenfold increased risk of HL. Of interest, the cases had significantly decreased levels of IgM against the VCA, compared to controls (Mueller et al., 1989). More than half (56%) of the sera from pre-HL cases had elevated titers of one or more EBV antibodies, compared to 35% of controls. This pattern of EBV antibodies was quite different from that seen in prediagnosis serum samples from non-Hodgkin lymphoma patients, obtained from the same serum banks (Mueller et al., 1991); no significant differences were found for antibody titers against CMV. A major breakthrough linking the EBV to HL was achieved by the application of molecular hybridization assays, as first reported by Weiss et al. (1987); these studies are summarized in detail elsewhere (Mueller, 1996). The studies consistently find that between 25% and 50% of HL cases are EBV-positive, as defined by the presence of EBV genome (which has been shown to be monoclonal) or of viral gene products. EBV positivity is more strongly associated with the histologic subtypes that connote more advanced disease, is higher in HL in children, and is somewhat lower among young adults (Glaser et al., 1997). As first reported by Pallesen et al. (1991), it has been further demonstrated that the EBV-positive tumor cells express a restricted latent infection phenotype of only EBNA-1+/LMP-1/ +LMP2A+ (Niedobitek et al., 1997; Ambinder and Weiss, 1999) (Table 26–2). In parallel we have found that the typical HL EBV antibody pattern segregates with tumor EBV-positivity, while that for EBV-negative HL appears to be normal in both a case-control study (Chang et al., 2004) and in prospectively obtained blood specimens (L. Levin et al., personal communication). Finally, the link to IM history was validated by Hjalgrim et al. (2003). In the follow-up of 38,000 Scandinavians with IM, the authors were able to characterize EBV status on biopsies from 29 of 40 cases of HL diagnosed ≥2 years after the diagnosis of IM. The RR of EBVpositive HL in this group associated with prior IM was 2.8 (1.7–4.6) and 1.1 (0.7–2.0) for EBV-negative HL.

Biological Causal Mechanisms The EBV has evolved a number of inherently oncogenic mechanisms that ensure its survival (Young and Murray, 2003; Dolcetti and Masucci, 2003). These include its ability to transform resting B-cells into permanently growing lymphoblastoid cell lines in vitro. The LMP-1 protein—which is expressed in both a lymphoid tumor, HL, and an epithelial tumor, NPC—is in itself an efficient oncogene. LMP1 plays a central role in transformation and in the up-regulation of a range of cellular factors, including the NFKB transcriptional pathway. What is actually astounding is that the great majority of humankind carry this virus through much of their lifetimes with no ill effects, being protected by the persistent surveillance of EBV-specific CTLs.

Infectious Agents This protection is seen most clearly when normal carriers undergo therapeutic immune suppression for organ transplantation. In a small proportion of such cases, a spontaneous lymphoproliferation can occur, which can evolve to an EBV-positive monoclonal B-cell lymphoma. The infusion of EBV-specific CTL can reverse this malignant process (Rooney et al., 1995).

Cofactors A consistent theme in the EBV-associated malignancies is the apparent effect of social environment on age at infection. The mechanism underlying these associations may likely relate to an age-related cytokine milieu. Other effects of social environment—such as that seen in the association of poverty and EBV-positive HL—may contribute. Genetic and dietary factors have been identified as likely of importance in NPC. Coinfection with HIV is a risk factor for both BL and for the EBV-positive HL. The increased risk among males that is seen for BL, NPC, and HL may reflect the protective effect of the relatively stronger type 1 immunity of women. Of potential interest, we have found in a recent case-control study on HL that aspirin, which selectively down-regulates the NFKB pathway, is apparently protective (Chang et al., 2004).

Prevention and Future Research EBV is largely an unavoidable infection. The epidemiologic evidence suggests that EBV—like many viral infections—is encountered best in childhood. That would argue against overly protecting children from infection and in favor of preschool and day care experience for those children who have limited exposure to other children at home. Given the potential oncogenicity of EBV as well as the morbidity of IM, the development of an effective vaccine against EBV is warranted. Potential interventions in early disease should be considered. For example, the therapeutic use of EBV-specific CTLs might be tested for early stage EBV-positive HL or NPC. The finding of an apparent protective effect of aspirin against HL needs to be confirmed. Finally, EBV continues to be implicated in a surprising array of malignancies. Some of these connections—such as with breast cancer (Labrecque et al., 1995)—may not prove to be valid (Hermann and Niedobitek, 2003). Others, such as for gastric cancer (Niedobitek et al., 1992), appear to be valid (Takada, 1999), and the underlying biology and epidemiology should be pursued. In either case, the application of the principles and biomarkers that have guided the established models should aid in validating and defining new EBV-associated malignancy.

HEPATITIS B VIRUS Hepatitis B virus (HBV) is a DNA virus within the Hepadnaviridae family, which includes other mammalian and avian hepadnaviruses, such as the woodchuck hepatitis virus and the duck hepatitis B virus (Ganem, 1996). These viruses have a very narrow host range as well as a striking tropism for hepatocytes. Among the DNA viruses, hepadnaviruses are highly unusual in that they replicate via reverse transcription of a full-length RNA copy of the virus genome. The spherical HBV virion, which is called a “Dane” particle, is about 42 nm in diameter. The outer lipid, bilayer envelope contains the hepatitis B surface antigen (HBsAg). Within the inner nucleocapsid core (HBcAg) is a partially double-stranded, circular genome approximately 3.2 kb in size as well as a DNA polymerase. The HBV genome contains four open reading frames for the S, C, pol, and X genes (Ganem, 1996; Lee, 1997). S encodes the envelope surface protein, HBsAg, and includes the two pre-S regions and the S region. Transcription beginning at the pre-S1 region produces the “large” S protein; at the pre-S2 region, the “middle” S protein; and at the S region, the “major” S protein. The C gene codes for the core protein, HBcAg, with the amino terminal region used for the synthesis of the HBV e antigen (HBeAg). The polymerase protein encoded by pol contains the DNA polymerase and reverse transcriptase func-

513

tions. The role of X in the virus life cycle is less clear, although in vitro experiments have shown that it can transactivate transcription of viral as well as cellular genes.

Natural History of HBV Infection Normally in acute HBV infection, HBsAg is usually the first marker detected, appearing within the first few weeks of infection (Hoofnagle and Schafer, 1986; Hollinger, 1996). HBeAg is present at about the same time as HBsAg and then is lost after a few days to a few weeks with subsequent seroconversion to antibody against HBeAG. Anti-HBcAG also appears during the early phase of infection, with the detection of IgM-specific antibody being a characteristic of acute infection. The HBc IgM antibody declines within about 6 months, but HBc IgG antibody generally persists indefinitely as a marker of past infection. Hepatitis symptoms tend to occur about 2 months postinfection, although the majority of acute infections are, in fact, subclinical. As HBV is cleared, antibody against HBsAG becomes detectable about 8 months after infection and subsequent to the disappearance of HBsAg. In contrast to recovery from acute infection, with chronic HBV infection, HBsAg, along with antibodies to HBc, remains persistently detectable, and HBs antibody does not develop (Hoofnagle and Schafer, 1986; Hollinger, 1996). HBeAg can persist for years; normally, about 10–15% of HBV chronic carriers will lose their HBeAg per year and seroconvert to anti-HBe seropositivity (Alward et al., 1985; Evans et al., 1997). The loss of HBeAg appears to be associated with latency of infection, with no or little replication. HBsAg also can disappear over time and antibodies against HBs develop, but this occurs at a very low rate of about 1% annually. HBV can become integrated into the host genome over the course of chronic infection and likely represents an important step toward hepatocarcinogenesis (Thomas, 1990; DeFranchis et al., 1993; Ganem, 1996). The integrated virus is not usually replicatively competent, although HBsAg is synthesized and expressed on the surface of the infected hepatocyte. The probability of chronic infection is closely and inversely related to the age at which infection with HBV occurs (Thomas, 1990; IARC, 1994a; Hollinger, 1996). Around 80% to 90%, or more, of infants born to infected mothers become chronic carriers (Beasley et al., 1977), whereas less than 10% of adolescents and adults will develop a persistent infection. Of children between the ages of 1 and 10, 20–40% will become chronic carriers, with the probability decreasing with increasing age over those years (McMahon et al., 1985). The situation is reversed with respect to the occurrence of clinically apparent hepatitis: about 30–50% of acutely infected adolescents and adults will experience symptomatic disease, compared to fewer than 5% of babies and young children. This difference in the course of initial infection is related to the competency of the host immune response to HBV. A strong immune response is required to clear the infection (Lee, 1997; Koziel, 1999). CTLs, primarily directed against HBV core antigens, kill infected hepatocytes. Cytokines, such as tumor necrosis factor-a (TNF-a) and IFN-g, also appear to be important in inhibiting HBV replication in acute resolving infection. This inflammatory response results in immune-mediated damage of the liver, leading to acute hepatitis. Nonetheless, the damaged hepatocytes are eventually replaced, neutralizing antibody develops, and the liver returns to normal. In chronic infection, the immune response is insufficient to completely eliminate the infection—as in the case of neonates who have immature type 1 immunity or in persons who are immunodeficient (Thomas, 1990; Hollinger, 1996). Thus, some infected hepatocytes persist and, in turn, infect other hepatocytes. The infected cells continue to be targets of the immune response, setting up cycles of hepatocellular injury and regeneration. Of note, infected young boys are slightly more likely to become chronic HBV carriers, despite an equal likelihood of infection at birth for both genders (London, 1981). About 70–90% of chronic HBV carriers will remain asymptomatic (Hoofnagle and Schafer, 1986). However, 10–30% will go on to develop chronic persistent or chronic active hepatitis. Of those carriers with chronic active hepatitis, approximately 1–2% per year will progress to cirrhosis. Both the hepatitis and cirrhosis can be clinically

514

PART III: THE CAUSES OF CANCER

silent. Hepatocellular carcinoma (HCC) develops at an annual rate of 1–6% among those with cirrhosis (Hollinger, 1996; Tabor, 1998). The latent period from infection to the occurrence of HCC is estimated to range from 30 to 50 years. About 15% to 25% of chronic HBV carriers will eventually die of some form of liver disease. Extrahepatic conditions related to HBV infection appear to be relatively rare and primarily include diseases associated with damage resulting from the formation of immune complexes (IARC, 1994a; Lee, 1997).

Epidemiology of Infection HBV is primarily a blood-borne infection that is transmitted by percutaneous and permucosal routes. The major modes of infection are via sexual, parenteral, and perinatal exposure. High levels of virus are found in blood and serum; more moderate viremia is evident in semen, vaginal fluid, and saliva. With respect to transmission by sexual contact, multiple partners and a history of sexually transmitted diseases appear to increase the risk of infection (Alter et al., 1989; IARC, 1994a; Hollinger, 1996). Sexual transmission of HBV is higher from men to women than the reverse. Parenteral exposure can occur via intravenous drug use as well as tattooing and body piercing. Medical personnel also are at risk of HBV transmission by contaminated needlestick or other injuries with sharp instruments. Before blood screening for antibody to hepatitis B core (HBc) antigen began, transfusion with contaminated units was a source of infection; in addition, hemodialysis has been a route of HBV transmission. Perinatal transmission occurs from the infected mother to her infant, usually by contact with maternal blood during delivery (Beasley, 1988). Some infected babies may acquire their infection in utero (Stevens et al., 1985; Li et al., 1986). The probability of perinatal infection is related to whether the carrier mother is also HBeAgpositive; the chance of infection is 70–90% for HBeAg-positive mothers versus 5–10% for HBeAg-negative, HBV-infected mothers (Beasley et al., 1977; Hollinger, 1996). Horizontal transmission of HBV also occurs in childhood. The actual modes of infection during this age period are not clear and likely multiple. In developing countries, the use of nondisposable, contaminated needles and syringes may play a role in transmission, particularly in the past (Ko et al., 1991). Transmission via skin lesions, human bites, and arthropods has also been suggested (Vall Mayans et al., 1990; IARC, 1994a). A study in Ghana found sharing of bath towels, of chewing gum or partially eaten candies, and of dental cleaning materials with an HBV carrier to be significant, independent risk factors for HBV infection in children aged 1 to 16 years (Martinson et al., 1998). Of interest, biting one’s fingernails in conjunction with scratching a carrier’s back also was associated with HBV seropositivity. The authors concluded that contact with objects contaminated with infectious saliva and/or blood represent important routes of HBV transmission in this age group. Childhood infection does appear to be associated with multiple infected siblings in the household (Whittle et al., 1991), also supporting possible contact-related transmission. The WHO estimated that there were 400 million HBV carriers in the world in the year 2000 (Lee, 1997). The prevalence of chronic infection is high (≥8% HBsAg positivity) in China, Southeast Asia, and sub-Saharan Africa as well as in parts of South America and in Alaska (IARC, 1994a). Moderate levels of endemicity (2–7%) are found in eastern and southern Europe, the Middle East, Japan, and southern and northern Asia. The lowest prevalence of infection (<2%) occurs in the Americas, including the United States, western and northern Europe, and Australia. The global distribution of chronic HBV infection is a function of the relative contribution of the different routes of infection on the level of endemicity. Perinatal infection is an important transmission route in high-endemic areas, particularly in China and other parts of Southeast Asia where up to half of the chronic infections are due to neonatal transmission (Beasley, 1988; IARC, 1994a). The higher frequency of perinatal infection in these countries, than in other endemic areas, appears to be related to the proportion of carrier mothers who are positive for HBeAg (Beasley et al., 1977). For reasons that are not known, the prevalence of HBeAg positivity is higher among infected Asian

women than among infected women in other endemic areas, such as Africa (IARC, 1994a). Childhood infection also plays an important role in the transmission of HBV in highly endemic regions as well as in areas with more moderate levels of chronic infection. In lowendemic areas, transmission occurs mainly during adulthood and adolescence, with sexual contact and parenteral exposure as the predominant modes of HBV infection. Although HBV infection of adults and adolescents takes place in areas with moderate or high levels of endemicity, its contribution to the prevalence of chronic carriers is relatively minor, as most of such infections would be cleared. In the United States, 0.33% of the population has been estimated to be HBV chronic carriers, based on data from NHANES III, conducted from 1988 to 1994 (McQuillan et al., 1999). The sample prevalence of infection was similar to that found in NHANES II (1976 to 1980). The proportion of persons ever-infected with HBV differed little between the two surveys: NHANES III, 4.9%, age-adjusted; NHANES II, 5.5%. Men were more likely to have been infected than were women (5.7% vs. 4.1%, respectively, NHANES III). The agespecific prevalence of infection in the United States appears to be quite low, until about the teenage years when it begins to increase, emphasizing the importance of sexual transmission and high-risk parenteral behaviors as routes of infection (Goldstein et al., 2002). Non-Hispanic blacks (12.8%) had higher levels of ever infection than did nonHispanic whites (2.8%) and Mexican Americans (4.8%).

HBV Biomarkers Several HBV antigens and antibodies to viral antigens can be routinely detected in serum, using commercially available assays (IARC, 1994a; Hollinger, 1996). As shown in Table 26–3, the chronic carrier state of infection is marked by the presence of HBsAg for more than 6 months. Antibody to HBsAg represents the neutralizing antibody and is indicative of recovery from infection and immunity to reinfection—the HBV vaccine induces the production of anti-HBs. HBeAg serves as a marker of high viral replication; in contrast, detection of antibody to HBe points to a low level of replication. Although HBcAg itself is not usually detectable in serum, both total antibody as well as IgM-specific antibody directed against HBcAg is used to distinguish past or chronic HBV infection from acute infection, respectively. Of note, assays for antibodies against HBcAg have been found to be not very specific, particularly at low titers; however, high titers, without the concurrent detection of HBsAg, can signify a low-level carrier. In addition, PCR methods can be used to detect the presence of HBV DNA in the serum.

Hepatocellular Carcinoma Chronic HBV infection has been established as a major cause of HCC (IARC, 1994a). The global incidence of HCC is strongly correlated with the prevalence of chronic HBV infection (Maupas and Melnick, 1981), with high rates of liver cancer found in Southeast Asia and subSaharan Africa, and low rates in the United States and Europe (Muñoz and Bosch, 1987). Researchers at IARC estimated that 52% of all liver cancer cases worldwide in 1990, about 229,000, could be attributed to chronic infection with HBV (Parkin et al., 1999). Most of the burden of HBV-associated liver cancer can be found in developing countries (attributable fraction, 59%), primarily Melanesia/Micronesia (76%), China (70%), and Africa (66%). In contrast, less than a quarter of liver cancer cases in developed countries are attributable to HBV; in North America and Australia, the estimate is 9%. A substantial body of epidemiologic evidence supports the etiologic role for chronic HBV infection in hepatocarcinogenesis (Muñoz and Bosch, 1987). Overwhelmingly consistent findings have been accumulated over two decades of research studies conducted in more than 25 different countries (IARC, 1994a). In the nearly 100 or more casecontrol studies evaluating this association, RR estimates have generally ranged from about 3 to 30, with a few reporting stronger associations (IARC, 1994a; Donato et al., 1998; Kuper et al., 2000a). Even more compelling evidence has been provided by more than 15 cohort studies that have followed HBV carriers over time (IARC,

515

Infectious Agents Table 26–3. Typical Serological Patterns in HBV Infection Anti-HBc Infection Status Acute infection* Chronic infection with high levels of viral replication Chronic infection with low levels of viral replication† Recovery from acute infection before development of anti-HBs Low titer; possible false positive High titer; possible “low-level carrier” Recovery from acute infection, indicating immunity Vaccine response‡ Susceptible to HBV infection

HBsAg

IgM

Total

HBeAg

Anti-HBe

Anti-HBs

+ + + -

+ + -

+ + + + + + + -

+ + -

+ + + + -

+ + -

Source: IARC Monographs, Vol. 59 (1994). anti-HBs, antibody to hepatitis B surface antigen; HBV, hepatitis B virus; HBsAg, hepatitis B surface antigen; anti-HBc, antibody to hepatitis B core antigen; HBeAg, hepatitis B envelope antigen; anti-HBe, antibody to hepatitis B envelope antigen. *Reactivated chronic disease may have this pattern with sensitive anti-HBc IgM assays. † Some patients may be seronegative for HBeAg and anti-HBe. ‡ In unvaccinated individuals, a high titer may represent immunity or be nonspecific; low titers are often nonspecific.

1994a; Mori et al., 2000; Evans et al., 2002). In these studies, RRs from 5 to more than 100 were observed. Probably the most well-known cohort study is that conducted in Taiwan by Beasley et al. (1981), which involved more than 19,000 male government employees followed for about 10 years and reported an RR of 103. The RR estimates tend to be higher in the cohort studies than in the case-controls because serum HBsAg (as biomarker of chronic infection) may disappear in some cases during the development of HCC (Huo et al., 1998). HBV DNA has been detected in the tumor tissue of more than 90% of HCC cases with HBsAg in their serum as well as in 10–20% of cases without HBsAg (Bréchot et al., 1982; Popper, 1988; Tabor, 1998). These HBV sequences appear to be clonally integrated into the genome of the infected hepatocyte, with no common integration site for the virus identified. Multiple copies of the HBV genome are found that could lead to the generation of random chromosomal abnormalities, including allelic loss (Tabor, 1994; Bréchot et al., 2000). Once integrated into the host genome, HBV no longer appears able to replicate, most likely as a result of the frequent rearrangement and deletions occurring in the viral genes. Of note, altered sequences from the X and pre-S2/S genes are preferentially retained.

Mechanisms of Carcinogenesis The exact mechanisms by which HBV induces hepatocarcinogenesis are essentially unknown (London and Buetow, 1988). HBV does not appear to be directly cytopathogenic. As discussed above, HBVassociated liver injury arises from the cell-mediated immune response to the virus and from bystander damage related to the release of cytokines. As much as 80% of HCC is observed to occur in conjunction with cirrhosis (Okuda et al., 1982; Hollinger, 1996; Tabor, 1998). Thus, chronic HBV infection could lead to HCC through continuous cycles of hepatocyte necrosis and regeneration that would “promote” acquired mutations that exist in the liver cells. Hepatocytes, in which HBV has become integrated, no longer express HBcAg and would not be targets for immune clearance, leading to their selective regeneration. Because it has the capability both to initiate, by virtue of its ability to integrate into the host genome, and potentially cause mutation as well as to promote hepatocarcinogenesis via constant immunemediated inflammation and cell turnover, HBV has been viewed as a “complete carcinogen” (Trichopoulos et al., 1987; Tabor, 1994). The woodchuck hepatitis virus appears to induce liver cancer via insertional mutagenesis at the myc gene location (Pineau and Tiollais, 1997). Although the evidence does not suggest that HBV acts in a similar manner, in vitro experiments have shown that proteins, encoded by the X and pre-S2/S regions of the HBV genome, are able to transactivate several cellular oncogenes, including c-myc and c-fos (Kekulé et al., 1990; Balsano et al., 1991; Tabor, 1994) which are overexpressed in HCC. As discussed, the X and pre-S2/S genes are often altered in the HBV inserts isolated from HCC tumor cells (Koshy and

Wells, 1991). These alterations may be important for transactivation. The native pre-S2/S middle protein does not appear to have transactivating properties, and the truncated X protein is a stronger transactivator than is the wild-type version. In addition, the X protein has been shown to bind the p53 tumor suppressor protein and to inhibit its function in HCC tissue as well as in transgenic mice (Tabor, 1997; Ueda et al., 1997). HCC has been reported to develop at a very high frequency in mice transgenic for the X gene, particularly in male mice (Kim et al., 1991; Ueda et al., 1997). In addition, in vitro studies indicate a possible role of the HBV X protein in cell cycle control and apoptosis (Bréchot et al., 2000). With respect to growth factors, the detection of transforming growth factor-a (TGF-a) is closely linked to that of HBsAg in HCC tissue as well as in nontumorous tissue (Tabor, 1994). It is possible that HBVrelated liver regeneration leads to increased levels of TGF-a, and/or that it acts in conjunction with HBV in the process of hepatocarcinogenesis. Last, some data exist that suggest that insulin-like growth factor-II is overexpressed in HBV-associated HCC (Lee et al., 1998). As previously mentioned, clonally integrated HBV DNA is found in the liver tissue from HCC cases who are negative for HBsAg. Indeed, low levels of HBV DNA can be detected in HCC cases as well as in healthy blood donors who are HBsAg-negative (Huo et al., 1998; Bréchot et al., 2000; Yotsuyanagi et al., 2001; Tabor, 2002). Frequently in such individuals, the only other marker of HBV infection found is anti-HBc. Such “silent” or “occult” HBV infections are believed to be etiologically relevant in the development of HCC, perhaps with other cofactors acting as promoters of hepatocarcinogenesis. The role of occult HBV coinfection may be particularly relevant with respect to the pathogenesis of HCV (Cacciola et al., 1999; Marusawa et al., 1999b; Tabor et al., 2002).

Cofactors The establishment of chronic HBV infection at a young age in childhood appears to increase the risk of HCC occurrence, based on data from case-control studies (Muñoz et al., 1989; Hsieh et al., 1992; Kuper et al., 2000c). In addition, HCC is much more frequent in persons younger than 40 in populations with a high incidence of liver cancer than in populations in which the risk of liver cancer is relatively low (Bosch, 1997). Male carriers are at a higher risk of HCC as well as of cirrhosis than are female carriers (Beasley, 1988; IARC, 1994a; Evans et al., 2002); this male predominance is more apparent in developing countries where HCC incidence and HBV endemicity are high. Although some of the gender difference may be due to higher prevalence of other cofactors in men (e.g., high alcohol consumption), endogenous hormones also may play a role. A cohort study in Taiwan found a significant fourfold increase rate of HCC for men in the highest tertile of testosterone level (Yu and Chen, 1993). In contrast, a cohort study in Shanghai reported a modest nonsignificant 50%

516

PART III: THE CAUSES OF CANCER

increased risk among HBV carriers with the highest tertile of testosterone (Yuan et al., 1995). In an additional study by the researchers in Taiwan, the relationship between fewer CAG repeats in the androgen receptor (AR) gene, which has been associated with prostate cancer in men and the incidence of HCC in male HBV carriers, was investigated (Yu et al., 2000). A twofold increased risk for HBV carriers with 20 or less CAG repeats in AR was observed; a possible synergistic effect of high testosterone levels also was suggested. Aflatoxins, which are produced by molds that contaminate food staples stored in hot, humid climates, have been determined to be a human carcinogen (IARC, 1993). They are hypothesized to be an important risk factor for HCC in some areas of the developing world (Montesano et al., 1997). In nested case-control studies of HCC among men in Shanghai (Qian et al., 1994) and Taiwan (Wang et al., 1996), a strong positive interaction was found between detectable urinary aflatoxins and HBsAg seropositivity. The particular p53 mutation, believed to be associated with aflatoxin exposure (the 249ser mutation), occurs at a higher frequency in HBV-associated HCC. Experimental evidence from studies in woodchucks and HBV-transgenic mice further support a synergistic effect of HBV chronic infection and aflatoxin consumption on the incidence of HCC (Montesano et al., 1997). A meta-analysis of HCC cases estimated a significant interaction between HBV positivity and a high level of aflatoxin exposure (Stern et al., 2001). However, HBV infection did not appear to modify the effect of high aflatoxin exposure on the proportion of HCCs with the 249ser mutation. It is known that the consumption of alcohol is hepatotoxic, although not necessarily mutagenic. Moreover, the epidemiologic data indicate that excessive drinking and alcoholic cirrhosis increase the risk of HCC (Kuper et al., 2000b; Donato et al., 2002). Several studies suggest that a more than additive effect of heavy alcohol consumption and chronic HBV infection exists with respect to the development of HCC, although the findings have been somewhat inconsistent (Bréchot et al., 1996; Chen et al., 1997; Yu et al., 1997; Kuper et al., 2000b; Mori et al., 2000; Donato et al., 2002). The proposed mechanism of action is that alcohol drinking further exacerbates HBV-induced hepatocellular damage. A role for smoking in hepatocarcinogenesis is biologically plausible (Kuper et al., 2000b). Whereas some studies have found a positive, although not necessarily significant, interaction between increased smoking and HBV infection for HCC (Chen et al., 1997; Mori et al., 2000), others have not (Tanaka et al., 1992; Kuper et al., 2000b; Evans et al., 2002). The evidence regarding an effect of HIV coinfection on the progression and pathogenesis of HBV has not been clear. Some studies have reported lower HBV levels as well as lower liver enzyme values in persons coinfected with HIV, implying a weaker immune response to HBV (Gilson et al., 1997; Colin et al., 1999). However, in a few studies, more liver disease has been observed in HBV carriers with HIV than in those without HIV (Colin et al., 1999; Thio et al., 2002), although not in others (Lee, 1997). A strong immune response during acute infection clearly is important with respect to clearance of HBV (Lee, 1997; Koziel, 1999). At the same time, among persons with chronic, persistent infection, progression to liver disease most likely is a consequence of HBV-specific immune-mediated damage. Thus, it seems plausible that host immune function, and factors affecting the balance between immune control of HBV infection and immuneinduced liver injury, should be important in the natural history of the infection.

Prevention and Future Research The best prevention of HBV-associated liver cancer is through the elimination of chronic HBV infection. The development of a safe and effective vaccine against HBV infection makes this goal attainable. HBV mass vaccination programs were begun in the early to mid-1980s in HBV-endemic populations, including in Africa, China, Taiwan, and the United States (Anand and Hollinger, 1997; Blumberg, 1997). Because the development of chronic HBV infection is much more likely in younger children, who also have the highest risk for HCC, immunization during infancy or early childhood is the most effective

strategy for reducing the burden of HBV. Accordingly, the childhood vaccination programs initiated in HBV-endemic areas have led to marked reductions in the prevalence of carriers in those populations (Chen et al., 1996; Blumberg, 1997). A most striking finding has been the observation of a significant decline in the incidence of childhood HCC in Taiwan between 1981 and 1994, subsequent to the introduction of the nationwide immunization effort against HBV (Chang et al., 1997). In the United States, the HBV vaccine has routinely been administered to infants since the early 1990s (Goldstein et al., 2002). As of 2002, programs targeted at universal immunization of infants existed in more than 100 countries worldwide (WHO, 2002). Interferon-a therapy is used to treat chronic hepatitis due to HBV infection. Wong and colleagues (1993) performed a meta-analysis of the published results from 15 clinical trials that evaluated the effect of IFN-a treatment of HBeAg-positive patients with chronic hepatitis B. About one-third showed a loss of HBeAg and/or HBV DNA. More recent studies have confirmed the 30–50% clearance of HBeAg or HBV DNA after IFN-a therapy for chronic hepatitis B, with lower rates of disease progression suggested for those who responded to treatment than for those who did not (Niederau et al., 1996). The effect of IFN on the development of HCC has been examined in a few studies of patients with HBV-related cirrhosis (Tabor, 2003). The studies, which were conducted in Singapore (Oon, 1992), Italy (Mazzella et al., 1996), and Japan (Ikeda et al., 1998), each reported fairly substantial reductions in the occurance of HCC in the treated cirrhotics compared to the untreated cirrhotics, despite differences in study design with respect to the treatment regimen and the length of follow-up. In the study in Italy, the HCC cases that occurred in the treated patients were only among those who did not respond to the IFN therapy. Thus, decreased disease progression, including the development of HCC among persons with HBV-induced liver injury, appears possible with IFN treatment. However, the side effects related to such therapy can be problematic. The development of more effective, less toxic treatments would be important in minimizing morbidity and mortality for those with chronic HBV infection. In parallel, the identification of those subgroups of HBV carriers (as defined by age, gender, and biomarkers, for example) who can benefit from therapeutic interventions should be determined.

HEPATITIS C VIRUS Hepatitis C virus (HCV) is a small, enveloped RNA virus. HCV has been classified as a separate genus in the Flaviviridae family, which includes such human pathogens as the yellow fever and dengue viruses as well as the hepatitis G virus (Lauer and Walker, 2001). The virus was first identified in 1988 by Choo, Kuo, and colleagues as the biologic agent likely responsible for most non-A, non-B hepatitis in the United States (Choo et al., 1989; Kuo et al., 1989). Using state-of-theart molecular techniques available at the time, the investigators screened an enormous number of clones prepared from the plasma of an experimentally infected chimpanzee. They eventually obtained a single cDNA clone that allowed them to isolate and sequence the virus as well as to construct an assay to identify antibodies to HCV proteins. The positive-sense, single-stranded viral genome, which is approximately 9.6 kb in length, codes for the core nucleocapsid protein, two envelope (E) glycoproteins (E1, E2), and six nonstructural (NS) proteins (NS2, NS3, NS4A, NS4B, NS5A, NS5B) (De Francesco, 1999). Neutralizing antibodies appear to be directed against the E2 protein (Koziel, 1997; Ferrari et al., 1999). Within the E2 coding sequence, there are two hypervariable regions; these regions are important with respect to the virus’s ability to evade neutralization by the immune system, by leading to the generation of multiple quasi-species of genetically related HCV variants within the infected host, similar to HIV (Pawlotsky, 2002). With respect to the nonstructural proteins, NS2, NS3, and NS4A have protease functions, with NS3 also demonstrating RNA helicase activity; NS5B functions as an RNAdependent RNA polymerase and is important in genome replication, perhaps through the synthesis of a negative-strand intermediate (De Francesco, 1999). Lacking reverse transcriptase, HCV does not integrate into the host genome (Choo et al., 1989).

Infectious Agents HCV infects hepatocytes and also may be lymphotropic (Wang et al., 1992; Bartolome et al., 1993). In vivo, the virus appears to replicate at a very high rate, producing as many as 1012 virus particles per day (Neumann et al., 1998). Until recently, HCV had not been grown successfully in cell culture, limiting the study of its replication and the viral life cycle. However, the generation of functional subgenomic replicons in a human hepatoma cell line (Lohmann et al., 1999) represented an important step toward the eventual development of in vitro systems that can product infectious virus particles (Wakita et al., 2005; Zhong et al., 2005; Lindenbach et al., 2005). At present, the chimpanzee is the only experimental animal that has been infected with HCV. No small animal model currently exists for studying the infection, although chimeric mice with human hepatocyte engraftments have been created and can sustain HCV infection of the xenogeneic cells (Mercer et al., 2001; Meuleman et al., 2005).

Natural History of HCV Infection In acute infection with HCV, virus RNA is detectable within 2 weeks after exposure (Hoofnagle, 2002). A few weeks later, serum alanine aminotransferase (ALT) levels, indicating liver dysfunction, begin to rise, with symptoms occurring soon after. However, less than 30% of persons newly infected with HCV will present with clinically apparent hepatitis. After the acute infection, periodic elevations in ALT can be observed and often parallel changes in HCV RNA load (Alter et al., 1992; Hoofnagle, 2002; Okayama et al., 2002). Antibodies against HCV become detectable about 8–12 weeks after exposure. The presence of HCV antibodies appears to be life-long; however, in some persons who have cleared their HCV infection, these antibodies may eventually disappear (Alter et al., 1992; Wiese et al., 2000; Hoofnagle, 2002). A vigorous CTL response is thought to be important for clearance of infection (Lechmann et al., 1996; Spengler et al., 1996; Koziel, 1997; Takaki et al., 2000). A strong T-lymphocyte response and low or undetected antibody levels have been found in acute hepatitis patients who clear their HCV infection, whereas a weak T-cell proliferative response and high antibody titers appear to characterize those patients who develop chronic hepatitis (Spengler et al., 1996). Some individuals are able to clear HCV infection, as marked by the loss of detectable RNA. However, the majority of newly acquired infections, as much as 75–85%, appear to become chronic as demonstrated by the persistence of HCV viremia for more than 6 months. Population-based, cross-sectional studies, of the prevalence of HCV RNA positivity among anti-HCV seropositives, have found similar proportions of persistence (Alter et al., 1997, 1999; Bellentani et al., 1999; Arduino et al., 2001; Hyams et al., 2001). However, the development of chronic infection has been reported to be lower, around 55% in some populations (Rodger et al., 2000; Hoofnagle, 2002; Seeff, 2002), including among young mothers infected with HCV by contaminated anti-D globulin (Kenny-Walsh et al., 1999; Wiese et al., 2000). In general, persistence of infection may be less frequent in women, younger persons, and individuals with symptomatic acute infection (Alter et al., 1999; Bellentani et al., 1999; Kenny-Walsh et al., 1999; Inoue et al., 2000; Wiese et al., 2000; Hoofnagle, 2002) and more frequent in African Americans and persons with immunodeficiency (Thomas DL et al., 2000; Seeff et al., 2001; Hoofnagle, 2002). Efforts to characterize the natural history of HCV have been difficult. Much of the problem is inherent to the infection itself, where onset and persistence are largely asymptomatic; progression to disease occurs after latency measured in decades, and cofactors likely play a critical role in the natural history of the infection (Seeff, 2002). Early data originated primarily from studies of transfused populations and patients with medically identified acute hepatitis or chronic liver disease (Kiyosawa et al., 1990; Alter et al., 1992; Takahashi et al., 1993; Tong et al., 1995; Seeff, 2002). Many of those studies were retrospective in nature or short in duration; often, they were not designed to specifically assess the natural history of HCV. More recently, additional information has been provided from retrospective-prospective, long-term cohorts of intravenous drug users (Rodger et al., 2000; Thomas DL et al., 2000), women exposed to contaminated anti-D

517

immunoglobulin (Kenny-Walsh et al., 1999; Wiese et al., 2000), and other community-based populations (Seeff et al., 2000), including our cohort in Miyazaki, Japan (Boschi-Pinto et al., 2000). Data from these studies suggest a lower rate of liver disease progression than that from the earlier studies. Although most HCV carriers are asymptomatic, abnormalities in ALT level do occur, indicative of chronic hepatitis and associated underlying liver injury; such elevations can be continuous or intermittent and frequently are mild (Hoofnagle, 2002). Of those chronically infected, 50–80% appear to develop chronic hepatitis of varying severity, likely within the first decade of infection (Kiyosawa et al., 1990; Alter et al., 1992, 1997; Takahashi et al., 1993; Kenny-Walsh et al., 1999; Rodger et al., 2000; Wiese et al., 2000; Seeff et al., 2001; Okayama et al., 2002). Freeman and colleagues (2001) have modeled the incidence of cirrhosis after 20 years of chronic infection, by study type: liver clinic series, post-transfusion cohorts, blood donor series, and community-based cohorts. Not surprisingly, the average 20-year frequency of progression to cirrhosis was higher in the studies of liver disease (22%, 95% CI, 18–26%) and post-transfusion patients (24%, 95% CI, 11–37%) than in the studies of blood donors (4%, 95% CI, 1–7%) and community-based populations (7%, 95% CI, 4–10%). Although older age at infection appeared to explain some of the increased occurrence of cirrhosis in the transfusion-exposed HCV carriers, other cofactors as well as underlying disease may also play a role. The presence of existing liver disease likely contributed to an overestimate of the frequency of progression in the liver clinic studies (Freeman et al., 2001). Thus, the proportion of HCV carriers who will develop cirrhosis, and subsequently progress to end-stage liver disease, may prove to be lower than initially thought. In any case, progression to more advanced liver disease is highly variable and will be influenced by characteristics of the infected host as well as of the virus (Poynard et al., 1997; Seeff, 2002). Once cirrhosis occurs, the risk of developing HCC increases at a rate that may range from 1% to 4% per year or higher (Alter and Seeff, 2000). The interval from infection to HCC has been estimated to be around 30 years, although latent periods of 40 to 50 years have been observed (Kiyosawa et al., 1990; Takahashi et al., 1993; Tong et al., 1995). The impact of HCV infection on overall mortality is less clear. Although the likelihood of death from liver disease may be higher in persons with HCV, the overall mortality rate may not be increased, as observed by us and others (Boschi-Pinto et al., 2000; Seeff et al., 2000, 2001). In fact, death from causes related to the route of infection (e.g., transfusion, intravenous drug use) is frequently observed in HCV carriers (Thomas DL et al., 2000; Seeff et al., 2001). The effect of HCV on the health of carriers is broader than that resulting only from damage to the liver. A number of extrahepatic conditions have been associated with HCV infection (Hoofnagle, 2002), including mixed essential cryoglobulinemia (Agnello et al., 1992; Cacoub et al., 1994), porphyria cutanea tarda (Ferri et al., 1993), siccalike syndrome (Haddad et al., 1992), and lichen planus (Jubert et al., 1994). Of these extrahepatic manifestations, cyroglobulinemia appears to be the most commonly observed. In addition, HCV has been reported to have a role in renal disease development, mainly with respect to glomerulonephritis (Rollino et al., 1991; Johnson et al., 1993). Studies also have linked HCV to the occurrence of diabetes (Allison et al., 1994; Mason et al., 1999; Mehta et al., 2001). HCV infection has also been associated with NHL in HCV-endemic populations (Ferri et al., 1994; Luppi and Torelli, 1996; Ohsawa et al., 1999; Montella et al., 2001). If the association of HCV infection with NHL proves to be valid, the potential mechanism of pathogenesis is likely to be indirect (De Re et al., 2000).

Epidemiology of Infection In terms of transmission, parenteral exposure to contaminated blood appears to be a particularly efficient route of HCV infection. Prior to the screening of the blood supply for anti-HCV, which began in 1990 in the United States, transfusion of contaminated blood and blood products was an important source of transmission (Alter et al., 2000). As a result, the prevalence of anti-HCV is extremely high among persons with hemophilia, ranging from 50% to 95% (Rall and

518

PART III: THE CAUSES OF CANCER

Dienstag, 1995). However, injecting drug use represents the most frequently reported risk factor for HCV infection in the United States and other Western populations, even among blood donors whose admitted intravenous drug use was often in the distant past (Conry-Cantilena et al., 1996; Alter et al., 1999, 2000; Murphy et al., 2000). Among active injection drug users, the level of infection is as high as observed for those with hemophilia. HCV appears to be acquired relatively quickly after intravenous drug use is begun and to be transmitted more readily than HIV. In some studies, intranasal cocaine use also was associated with anti-HCV seropositivity, independent of history of intravenous drug use (Conry-Cantilena et al., 1996), although other studies have not observed this association (Murphy et al., 2000). Other percutaneous exposures associated with the transmission of HCV include hemodialysis, organ transplantation, and accidental needlesticks in the health care setting (Rall and Dienstag, 1995; Alter et al., 2000). Data from Japan have implicated the use of nondisposable needles, syringes, and other medical instruments, after the Second World War and prior to the 1970s and 1980s, as a possible important route of HCV transmission in that country (Hayashi et al., 1995). Several folk remedies practiced in Japan that involve possible percutaneous exposure, such as acupuncture, also have been identified as potential sources of HCV infection (Kiyosawa et al., 1994; Noguchi et al., 1997). Studies in Italy and Taiwan also have reported an association between exposure to nondisposable needles and syringes and infection with HCV (Chen et al., 1995; Chiaramonte et al., 1996). In Egypt, parenteral antischistosomal therapy campaigns likely are responsible for the very high prevalence of HCV in that country; transmission likely resulted from insufficiently sterilized syringes during the mass intervention efforts (Frank et al., 2000). In addition, tattooing, body piercing, and sharing of razors have been reported as possible routes of HCV transmission, although the data have been rather limited and largely inconclusive (ConryCantilena et al., 1996; Alter et al., 2000; Murphy et al., 2000). Nonpercutaneous transmission of HCV also occurs, albeit at a relatively low rate. This route of infection appears to be rather inefficient, particularly in comparison to other blood-borne infections such as HBV and HIV, or even HTLV-I. Although HCV likely is transmissible via sexual contact, the data concerning the relative importance of sexual transmission in HCV infection are somewhat conflicting. An increased prevalence of HCV has been reported for sexually transmitted disease clinic populations and for persons with multiple sexual partners, including prostitutes (Rall and Dienstag, 1995; Alter et al., 1999, 2000; Terrault, 2002). However, such groups may engage in other high-risk behaviors, which could have confounded some of the reported findings. In contrast, almost no transmission has been observed in the long-term, monogamous sexual partners or spouses of HCV-infected hemophiliacs (Terrault, 2002) and recipients of contaminated anti-D immunoglobulin (Meisel et al., 1995). Although some studies from Japan have reported evidence of sexual transmission of HCV within marriage (Akahane et al., 1994; Chayama et al., 1995), our cohort study as well as other studies have not supported this route of infection as having a major impact on the spread of HCV in Japan (Nakashima et al., 1995; Tanaka et al., 1997; Okayama et al., 2002). A transmission rate of <1% per year has been estimated for persons in monogamous relationships; the frequency is somewhat higher if an individual has multiple sex partners or is HIV-infected (Terrault, 2002). Transmission from HCV-infected mothers to their babies has been reported in studies conducted in the United States, Japan, and Europe, although the probability is rather low, at 4–7% (Ohto et al., 1994; Zanetti et al., 1995; Roberts and Yeung, 2002). Perinatal transmission is hypothesized to take place in utero and/or at the time of delivery and not via breastfeeding. High virus titer as well as HIV coinfection in the mother appears to be important for transmission to occur. The seroprevalence of HCV infection ranges from about 0.1% to 5% in most populations worldwide (Lavanchy and McMahon, 2000). The WHO estimates that approximately 3% of the world’s population is infected with HCV. The overwhelming majority of countries with available data have a seroprevalence of less than 2.5%. More inter-

mediate prevalences, up to 10%, are found in Brazil, parts of Africa, and China. The highest reported level of infection is in Egypt, with estimates of 20–30% or more. Pockets of high seroprevalence also occur within areas of relatively low endemicity, as has been observed in some areas of Japan, where the frequency of anti-HCV positivity can be greater than 10% (Kiyosawa et al., 1994; Nakashima et al., 1995; Tanaka et al., 1997). In the United States, data from NHANES III estimate the prevalence of anti-HCV seroprevalence to be 1.8% in the general population (Alter et al., 1999). The prevalence of infection is higher among non-Hispanic blacks and Mexican Americans than among whites. A slightly higher prevalence of anti-HCV among men also has been reported in the United States (Murphy et al., 1996; Alter et al., 1999). With respect to age, there appears to be a peak in prevalence among 30- and 40-year-olds. An explanation for the unimodal age distribution of HCV seroprevalence in the United States is that it reflects HCV infection via intravenous drug use during the late 1960s through the early 1980s (Murphy et al., 1996; Alter et al., 2000). In populations with more endemic levels of infection, such as in Japan, we and others have found a somewhat different age- and sex-specific pattern of HCV seroprevalence (Hayashi et al., 1995; Nakashima et al., 1995; Okayama et al., 2002). There, the seroprevalence of anti-HCV tends to increase through the fifth decade and then is fairly stable. Also, men and women have about the same frequency of infection. This distribution of infection likely reflects differences in the timing and mode of infection between Japan and the United States.

HCV Biomarkers The HCV RNA genome can be detected by reverse-transcriptase PCR (RT-PCR) methods, with detection limits of the commercial assays as low as 50–100 IU/mL (Pawlotsky, 2002). Quantification of HCV RNA level can be obtained by RT-PCR or branched DNA signal amplification, although these assays are less sensitive than the qualitative assays to detect HCV RNA. Real-time PCR, based on the TaqMan system, may offer a more sensitive alternative for the quantification of HCV viral load (Takeuchi et al., 1999). In addition, enzyme immunoassays (EIAs) exist that can measure the amount of HCV core antigen (Aoyagi et al., 1999; Komatsu and Takahashi, 1999; Widell et al., 2002). Detection of core antigen appears to correlate very well with HCV RNA (Komatsu and Takahashi, 1999; Pawlotsky, 2002). Although the first versions of these assays could not detect HCV core antigen below 20 KIU/mL of HCV RNA (Bouvier-Alias et al., 2002), newer assays are more sensitive (Aoyagi et al., 1999; Tanaka et al., 2000) and may prove to be a more cost-effective, simpler means of detecting and quantifying HCV viremia on a population level. Antibodies to HCV are most frequently measured using EIAs (Pawlotsky, 2002). Second-generation versions of these assays employ recombinant proteins derived from the core, NS3, and NS4 regions; the third-generation EIAs also include antigens from the NS5 region. Immunoblot testing can be used to confirm anti-HCV seropositivity, which was more important for earlier, less specific versions of the EIAs, particularly in populations with a low prevalence of infection, such as blood donors. Six major genotypes of HCV, each with a number of subtypes, have been identified throughout the world (Simmonds et al., 1993). The six primary genotypes differ genetically from one another by approximately 30%; within a given genotype, the genetic variation is about 20% (Simmonds, 2000). The relative frequency of the HCV genotypes and subtypes does vary from country to country and even between different populations of the same country.

Hepatocellular Carcinoma IARC classifed HCV as a human carcinogen in 1993, primarily on the strength of epidemiologic data (IARC, 1994a). Parkin et al. (1999) have estimated that 25% of liver cancer cases occurring worldwide in 1990 could be attributable to HCV infection. The greatest burden excluding liver cancer caused by HCV is in Africa, Japan, and Oceania

Infectious Agents (not Australia/New Zealand), where the attributable fractions are greater than 30%. In Japan, it is believed that most of the recent increase in HCC observed in that country is due to chronic HCV infection (Okuda, 1997). Although there is not a striking global concordance between the prevalence of chronic HCV infection and incidence of liver cancer, as is evident for HBV, within Japan, HCV seroprevalence has been correlated with local HCC rates (Tanaka et al., 1994). The risk of HCC in persons with HCV infection is substantial, with RRs greater than 10 found in several cohort studies (Boschi-Pinto et al., 2000; Mori et al., 2000; Chang C-C et al., 1994; IARC, 1994a). This strong association has been echoed in numerous case-control studies across different populations (IARC, 1994a; Donato et al., 1998; Tagger et al., 1999; Kuper et al., 2000a), with associations of the order 2- to 50-fold and higher based on the seroprevalence in the source population. Persistent, chronic infection, as indexed by the detection of HCV RNA, appears to be etiologically relevant with respect to the development of HCC (Donato et al., 1998; Tagger et al., 1999). Virus-specific factors may be important with regard to the risk of HCV-associated liver disease progression and HCC. Genotype 1b has been reported to be more pathogenic and may contribute to hepatocarcinogenicity (Tanaka K et al., 1996; Davis, 1999; Tagger et al., 1999), although this association has been questioned (Davis, 1999; Tagger et al., 1999; Seeff, 2002). The effect of HCV load on disease development is also unclear. Some studies have shown a correlation of the quantity of serum HCV RNA with ALT level and stage of liver disease (Hagiwara et al., 1993; Gretch et al., 1994). Others have not observed such a relationship (McGuinness et al., 1996; Bonis et al., 1999; Thomas DL et al., 2000; Seeff, 2002). The relevance of quasispecies in the oncogenic progression of HCV-associated liver disease is not established, although there are reports of increased numbers of HCV quasispecies in those with advanced chronic hepatitis and with a poorer response to treatment (Hayashi et al., 1997; Davis, 1999; Seeff, 2002).

Mechanisms of Carcinogenesis The mechanisms underlying the hepatocarcinogenic effect of HCV are not well understood. As a nonintegrating virus, HCV is not likely to act as a cancer initiator in the usual sense. Because up to 90% of HCVassociated HCC arises within a cirrhotic liver (Alter and Seeff, 2000), HCV infection may lead to cancer through a promoting process of recurring cycles of cell death and regeneration, as would occur with cirrhosis (Tabor, 1998). It has been suggested that the presence of HCV-specific CTLs in liver tissue indicates T-cell–mediated toxicity with concomitant hepatocyte lysis (Koziel, 1997; Ferrari et al., 1999). HCV-associated liver cancer occurs in somewhat older patients and in conjunction with more severe disease than does HBV-associated HCC (Shiratori et al., 1995). Thus, an extended duration of liver cell damage—progressing from chronic hepatitis to cirrhosis to HCC— may be critical in hepatocarcinogenesis, resulting in an accumulation of mutations over an extended period in a stochastic manner (Kiyosawa et al., 1990; Tabor, 1998) (Fig. 26–5). It should be noted, however, that HCV-associated HCC has been reported in the absence of cirrhosis (Alberti et al., 1992; De Mitri et al., 1995). The HCV core protein may be able to interfere with cell-cycle control and thus inhibit apoptosis (Ray et al., 1996b; Chung and Liang, 1999). In particular, the core protein may affect apoptosis induced by TNF-a, which is important in the immune response to viral infection (Marusawa et al., 1999a). In addition, there have been reports that the core protein can repress transcription from the promoters of the genes for the tumor suppressors p53 and retinoblastoma protein (Kim et al., 1994; Ray et al., 1997). Moreover, the N-terminal half of the NS3 protein of HCV as well as the viral core protein have been shown to transform cells in vitro (Sakamura et al., 1995; Ray et al., 1996a). Of further interest is the observation that mice transgenic for the HCV core gene develop HCC (Moriya et al., 1998). Experimental data also suggest that HCV replication may increase the expression of TGF-a and insulin-like growth factor-II, both of which are likely important in hepatocyte transformation (Tanaka S et al., 1996).

519

Figure 26–5. Postulated natural history of HCV infection.

Cofactors A number of cofactors have been identified for HCV–induced HCC. Older age at infection and male gender appear to be associated with liver disease progression among HCV carriers (Poynard et al., 1997; Freeman et al., 2001; Seeff, 2002). The interaction of alcohol drinking with chronic HCV infection in liver disease pathogenesis has been evaluated in a number of epidemiologic studies, with the evidence convincingly supporting that heavy alcohol consumption and HCV infection act synergistically to increase the progression of liver disease and the development of HCC (Shen et al., 1996; Poynard et al., 1997; Tagger et al., 1999; Mori et al., 2000; Donato et al., 2002; Peters and Terrault, 2002). In cohort studies of HCC in Japan (Mori et al., 2000) and Taiwan (Sun et al., 2003), a positive interaction has been observed between cigarette smoking and HCV infection. The hypothesis that HBV and HCV infections interact in the development of HCC has been supported by a number of reports (Donato et al., 1998; Tagger et al., 1999). In a meta-analysis of 21 case-control studies, Donato et al. (1998) estimated a greater than additive effect of the two viruses, with an OR of 165 for dual infection, 17.3 for infection with HCV alone, and 22.5 for infection with HBV alone. Of note, a negative correlation between the detection of anti-HCV and HBsAg within an individual has been observed, suggesting virus interference (Tanaka et al., 1991; Donato et al., 1998; Tagger et al., 1999; Thomas DL et al., 2000). Thus, when coinfection does occur, it may lead to enhanced hepatic pathogenesis and cancer development. Coinfection with HIV also appears to be associated with more rapidly progressing HCV-associated liver disease (Thomas, 2002). A meta-analysis of eight cohort studies found a significantly elevated rate of decompensated liver disease in persons coinfected with HIV and HCV, compared to those infected only with HCV (Graham et al., 2001). Of note, HCV RNA levels are reported to be higher in HIV-infected individuals (Eyster et al., 1994; Telfer et al., 1994; Thomas, 2002). In addition, we have observed a positive interaction between HCV and HTLV-I, with respect to liver cancer occurrence in the Miyazaki Cohort Study (MCS) (Boschi-Pinto et al., 2000). Thus, it seems plausible that immune suppression inhibits the effective control of HCV, with a resulting enhancement of liver damage, exacerbation of liver disease, and promotion of the hepatocarcinogenic process. The host immune response to HCV infection likely is very important in the persistence of this infection and its induction of HCC. As with other chronic oncogenic infections controlled by CTLs, the antibody response may reflect persistent viral antigen expression and not a protective effect. Of interest, a study in Japan found an increased risk of HCC associated with very high antibody titers to HCV (Mori et al., 2000). In addition, several HLA genotypes have been found to

520

PART III: THE CAUSES OF CANCER

be related to both clearance of HCV (Barrett, et al., 1999; Mangia et al., 1999) and liver disease progression (Aikawa et al., 1996; Kuzushita et al., 1998; Mangia et al., 1999). In particular, significant associations have been reported for alleles from the HLA class II DRB1 and DQB1 loci.

Prevention and Future Research HCV is a serious public health problem in populations with high levels of infection. With the recent increases in HCC observed in the United States (El-Serag and Mason, 1999), there is growing concern that in a pattern similar to Japan (Okuda, 1997), this trend will continue over the next 20 years, primarily as a result of HCV infection (Armstrong et al., 2000). In terms of prevention, no vaccine currently exists. The development of a vaccine has been impeded by substantial virus strain variation as well as the existence of multiple quasi-species within the infected host and the need for a strong cellular immune response to clear and control infection (Prince, 1994). Fortunately, interruption of parenteral transmission has led to marked decreases in the incidence of HCV infection. The screening of the blood supply for HCV antibodies has substantially reduced post-transfusion hepatitis C cases in Japan (Japanese Red Cross, 1991), Taiwan (Wang et al., 1995), and the United States (Alter, 1997; Tobler and Busch, 1997). In addition, the implementation of disposable needle and syringe usage and other changes in medical practices appears to have been effective in diminishing HCV transmission in Japan (Okuda, 1997) and Taiwan (Sung, 1997). Although the prevalence of HCV among intravenous drug users is shockingly high, increased efforts to prevent and treat drug use as well as to reduce sharing of injection paraphernalia (Thorpe et al., 2002) will be important in limiting new infections in this high-risk population. Data from clinical trials suggest that treatment with IFN-a in combination with ribavirin results in sustained loss of HCV RNA in 40–50% of patients with chronic HCV, which is higher than the approximately 20% observed with IFN-a monotherapy (McHutchinson et al., 1998; Mangia et al., 2001). A number of studies have evaluated the direct effect of IFN-a treatment on the incidence of HCC (Tabor, 2003). Although issues have been raised concerning variations and problems in study design, the results of these studies suggest that IFNa may reduce the occurrence of HCC in treated individuals, with a stronger effect being found for responders (virologic or biochemical). Thus, chemotherapy may serve as a means of preventing the development of HCC in chronic HCV carriers. Continued improvements in treatments available (DiBisceglie and Hoofnagle, 2002) also would be beneficial in decreasing the burden of HCV-associated disease. Future research should focus on clarification of the oncogenic mechanisms of HCV infection, with particular focus on the role of host immunity. The delineation of the natural history of infection can lead to the identification of high-risk individuals and, potentially, to the modification of risk. Such studies in community-based and special exposure cohorts should also provide more realistic estimates of disease outcomes for infected individuals. These research challenges will require multidisciplinary teams of epidemiologists, virologists, immunologists, and clinicians if progress is to be made quickly and cost-effectively.

HUMAN HERPESVIRUS 8 Human herpesvirus 8 (HHV8), or Kaposi sarcoma (KS)-associated herpesvirus, is a gamma human herpesvirus with a restricted geographic distribution (Sarid et al., 1999). Similar to the EBV, most infections with HHV8 remain asymptomatic. Nevertheless, the oncogenic potential for HHV8 infection is suggested by its association with all types of KS (Chang Y et al., 1994; Moore and Chang, 1995; Buonaguro et al., 1996), with primary effusion or body cavity–based lymphomas (PEL or BCBL) (Cesarman et al., 1995), and with multicentric Castleman disease (MCD) (Soulier et al., 1995). However, many aspects of the natural history, epidemiology, and pathogenesis of HHV8 remain uncharacterized.

Morphologically, HHV8 resembles other herpesviruses. A core of linear double-stranded DNA is surrounded by a capsid, a protein-rich tegument, and an outer lipid envelope (Renne et al., 1996a; Wu L et al., 2000). The HHV8 genome is approximately 165 kb in size and encodes more than 85 viral proteins (Renne et al., 1996a; Russo et al., 1996). Many HHV8-encoded genes are oncogenes that are either homologues of human oncogenes or unique to HHV8 or the rhadinovirus species (reviewed by Moore and Chang, 2003). Though the HHV8 genome is generally highly conserved, five major variants (genotypes A through E) have been characterized to date, which have restricted geographic distributions (Zong et al., 1999; Biggar et al., 2000). The five genotypes have no apparent specific disease associations (Hayward, 1999). HHV8 was discovered by the creative application of representational difference analysis in 1994 in lesions from KS tumors (Chang Y et al., 1994) and is the first known human gamma-two herpesvirus, or rhadinovirus (Moore et al., 1996; Russo et al., 1996). HHV8 is closely related to several rhadinovirus species that infect nonhuman primates. The EBV, a gamma-one herpesvirus, is the closest relative to HHV8 among human herpesviruses (Moore et al., 1996; Russo et al., 1996). Although HHV8 is detectable in the malignant spindle cells from KS lesions, cultured KS cells lose their HHV8 infection after only a few passages (Sarid et al., 1999). In addition, HHV8 is difficult to transmit to primary cells in culture (reviewed in Ablashi et al., 2002). Thus, little is known about the effect of HHV8 on untransformed cells. In one primary cell line, successful infection with HHV8 resulted in the development of a spindle-cell morphology, typical of the malignant cell type in KS lesions (Cannon et al., 2000; Wu et al., 2001). The cells also lost contact inhibition and expressed the virus-encoded latent nuclear antigen (LANA-1 or LNA-1). In addition, approximately 10% of the spindle-shaped cells were observed to spontaneously enter the lytic cycle and express viral structural proteins (Cannon et al., 2000; Wu et al., 2001). In contrast, in a previously transformed cell line, latent HHV8 infection was maintained indefinitely after transmission, although the cells could be induced to enter the lytic cycle (Moses et al., 1999). Transition to lytic replication resulted in the development of the spindle-cell morphology in these cells.

Natural History of HHV8 Infection According to serologic studies, sexual intercourse and nonsexual intrafamilial transmission appear to predominate in the spread of HHV8. Sexual transmission is particularly efficient among male homosexuals, although no clear pattern of specific high-risk homosexual behavior has emerged to date (Schulz et al., 2002). HHV8 DNA has been detected in semen samples, but only rarely or in small quantities, and thus it is unclear whether semen is a relevant mode of exposure during sexual activity (Schulz et al., 2002). Close personal contact and saliva are likely routes of non-sexual intrafamilial transmission (Ablashi et al., 2002). Of interest, seropositivity for HHV8 was associated with HBV infection among Ugandan children and adolescents, suggesting that factors that favor HBV transmission may also increase exposure to HHV8 (Mayama et al., 1998). Perinatal transmission does not appear to occur frequently (Goedert et al., 1997; Mayama et al., 1998), but HHV8 appears to spread non-perinatally from mother to child and between siblings in populations of moderate to high endemicity (Gessain et al., 1999; Plancoulaine et al., 2004). In one endemic population, the risk of intrafamilial HHV8 transmission was related to the relatives’ HHV8 serostatus, but not to anti-HHV8 titer or viral load (Plancoulaine et al., 2004). Parenteral exposure and solid organ transplantation confer a low risk of HHV8 transmission (Parravicini et al., 1997; Engels et al., 1999; Pellet et al., 2003; Renwick et al., 2002). HHV8 appears to establish a lifelong persistent infection in the host, consistent with other herpesviruses (Ariyoshi et al., 1998; Biggar et al., 2003). In vitro studies suggest that heparan sulfate is the cellular receptor for HHV8 (Birkmann et al., 2001). Subsequent entry into the cell appears to occur by clathrin-mediated endocytosis (Akula et al., 2003). In vivo, the circulating cell types most frequently infected with HHV8 are B-lymphocyes and monocytes (Monini et al., 1999).

Infectious Agents In latently infected cells, HHV8 DNA exists as a circular episome in the cell nucleus (Renne et al., 1996a). Only a small subpopulation of infected cells undergoes lytic replication at any given time; in lytically replicating cells, the viral genome is linear (Renne et al., 1996a; Orenstein et al., 1997). Very few HHV8 genes are expressed in the latent phase (Renne et al., 1996b; Zhong et al., 1996). These include LANA-1 or LNA-1 and K12 or Kaposin. The proteins encoded by these and other constitutive latent cycle genes have cell-transforming capabilities and other complex functions (Ablashi et al., 2002). A second latent nuclear antigen, LANA-2, is expressed in PEL and MCD but not in KS lesions and is a potent inhibitor of programmed cell death (Rivas et al., 2001). A broader array of HHV8 genes is expressed in the lytic cycle, including capsid proteins and envelope glycoproteins involved in viral replication (Sarid et al., 1999; Schulz et al., 2002). Host control of HHV8 replication is likely to depend heavily on cellular immunity. Consistent with this assumption, CTLs, specific for the latent K12 protein, as well as for several envelope glycoproteins and other structural antigens, have been documented in HHV8seropositive subjects (Osman et al., 1999; Wang et al., 2001). Natural killer (NK) cells may also contribute significantly to the control of latent HHV8 (Sirianni et al., 2002). In addition, lytic cycle proteins and LANA-1 are specific targets of circulating antibodies, although it is unknown whether the humoral response further enhances overall control of HHV8 replication. The LANA-2 antigen does not appear to be immunogenic (Rivas et al., 2001). Based on limited data from seroconverters to HHV8 antibody positivity, primary infection with HHV8 leads to the appearance of HHV8 DNA in PBMCs, and of extracellular serum viremia (Goudsmit et al., 2000; Wang et al., 2001). Ex vivo immunologic experiments suggested that this primary infection elicits a lytic antigen-specific CTL response and increased secretion of the cytokine IFN-g, and that such responses may occur several months prior to the appearance of detectable circulating HHV8 antibodies (Wang et al., 2001). The observed lag between the cellular response and detection of HHV8 antibody may indicate that initial titers are very low, close to detectable limits of available serologic assays (Biggar et al., 2003). It is not uncommon, especially among HIV-infected men, for subjects to mount antibodies to only one (i.e., lytic or latent) antigen at seroconversion and to develop a broader humoral response over subsequent months to years (Goudsmit et al., 2000; Biggar et al., 2003). Overall, antibody reactivity is more commonly observed to lytic than to latent HHV8 antigens. No severe clinical illness has been described in association with primary HHV8 infection, although limited data are available. Among five HIV-negative male seroconverters, four reported transient and nonspecific symptoms, such as localized rash, lymphadenopathy, and fatigue around the estimated time of seroconversion (Wang et al., 2001). Infection appeared to occur without clinical signs or symptoms in the fifth man. Similarly, low-grade fever was a common manifestation in immunocompetent infants and children with presumed primary HHV8 infection (Andreoni et al., 2002). A skin rash, respiratory tract infection, and/or cervical lymphadenopathy were also observed in some of the children. Though based on small numbers, these reports do not suggest a difference between children and adults in the experience of primary infection. The nature of host–virus interactions after primary infection with HHV8 is poorly understood, and serologic profiles associated with viral reactivation are not characterized. A pattern of increasing HHV8 antibody titers precedes the diagnosis of KS in a subset of subjects (Renwick et al., 1998; Jacobson et al., 2000; Biggar et al., 2003). However, duration of HHV8 infection also predicts increasing titers in asymptomatic carriers (Biggar et al., 2003). Coinfection with HIV is associated with higher antibody titers as well as with a higher likelihood of discordant antibody seropositivity and periodic loss of detectable antibody (Biggar et al., 2003). These observations collectively suggest that antibody titers may be indicative of the quality of cellular immune control of HHV8. A proposal that antibody titers also reflect HHV8 viral load (Sitas et al., 1999) has been disputed (Campbell et al., 1999) and remains unresolved. In asymptomatic seropositive subjects, HHV8 DNA is infrequently detectable by PCR in PBMCs (Whitby et al., 1995; Engels et al., 2000). In a recent cross-

521

sectional study of 158 asymptomatic HHV8-seropositive adults in Italy, the prevalence of HHV8 DNA detection was 16.5% (Brown et al., 2005). HHV8 DNA positivity was predicted by childhood crowding and by clinical markers of anemia and thrombocytopenia; in addition, level of HHV8 viral load was inversely associated with red blood cell and platelet counts. Those findings, if confirmed, may indicate that early age at infection and chronic inflammation influence adult HHV8 viral load (Brown et al., 2005).

Epidemiology of Infection It is well accepted that HHV8 is not a ubiquitous infection, unlike EBV and other herpesviruses. Furthermore, the geographic distribution of HHV8 closely parallels that of KS. The factors that influence the epidemiology of HHV8 in a given population therefore reflect the predominant modes of HHV8 transmission and the prevalence of risk factors for KS (Sarid et al., 1999). For example, a higher seroprevalence of HHV8 infection is predicted by older age, Mediterranean or Eastern European origin, HIV infection, and homosexual or other high-risk sexual behaviors and their correlates (i.e., certain drug use or a history of sexually transmitted disease). In endemic areas, having a seropositive mother or sibling also predicts a higher risk of HHV8 antibody seropositivity (Plancoulaine et al., 2000). Seroprevalence does not differ markedly by gender but has been observed to increase with increasing age in both endemic and nonendemic populations (Lennette et al., 1996; Calabrò et al., 1998). In endemic areas, the agespecific increase in seroprevalence begins in childhood (Mayama et al., 1998; Gessain et al., 1999). In contrast, age-specific seroprevalences do not rise before adolescence or later teenage years in nonendemic regions (Lennette et al., 1996). In two recent reports, markers of lower socioeconomic status also predicted HHV8 seropositivity in endemic areas (Wojcicki et al., 2004; Mbulaiteye et al., 2005). Regions of low endemicity include the United States, the United Kingdom, and much of Northern Europe and Asia (Schulz et al., 2002). Reported general population seroprevalences typically range from 2% to 9% in the United States and between 3% and 5% in Northern Europe and Southeast Asia, depending on the serologic assay used. In contrast, intermediate levels of endemicity characterize Mediterranean populations, such as Italy and Greece, where prevalences of 14% to 30% have been reported. In Italy, reported seroprevalences are usually high in southern areas but closer to those of nonendemic populations in the North, in parallel with the distribution of “classic” (HIV-negative) KS (Calabrò et al., 1998; Whitby et al., 1998). HHV8 infection is considered highly endemic in parts of Africa, where the greatest population seroprevalences have been documented. For example, some countries in sub-Saharan Africa have reported seroprevalences of 50% to 60% (Schulz et al., 2002).

HHV8 Biomarkers Molecular biomarkers for HHV8 infection include HHV8 DNA, which is consistently demonstrable in KS and PEL specimens by PCR, in situ hybridization, Southern blot hybridization, and immunohistochemistry (Ablashi et al., 2002). Two HHV8 RNA transcripts can also be detected using in situ hybridization: T0.7 which likely encodes a latent protein and is expressed in resting tumor cells; and T1.1, which appears to be an untranslated transcript expressed exclusively during lytic infection (Sarid et al., 1999). These molecular technologies have been invaluable to studies of the virology and biology of HHV8. For epidemiologic investigations, however, these techniques are not as informative, due in part to the low prevalence (often <10%) of HHV8 DNA detected in asymptomatic HHV8-seropositive subjects (Whitby et al., 1995; Engels et al., 2000). Alternatively, a variety of serologic methods have been used in epidemiologic studies to detect antibody to HHV8. The most common approaches include IF-based assays, (Kedes et al., 1996; Lennette et al., 1996), ELISA (Simpson et al., 1996), and EIA (Chatlynne et al., 1998). The IFAs typically test serum reactivity to PEL cell lines that harbor HHV8 but not EBV or HIV (Ablashi et al., 2002). ELISA and EIA methods incorporate single or multiple recombinant

522

PART III: THE CAUSES OF CANCER

HHV8-encoded antigens or whole virus lysate (Chandran et al., 1998; Chatlynne et al., 1998; Simpson et al., 1996). The earliest assays demonstrated poor reproducibility, especially among healthy blood donors, in an interassay comparison analysis (Rabkin et al., 1998). More recently, developed tests have shown improved interassay reproducibility (Engels et al., 2000). In the absence of a gold standard of true HHV8 status, these serologic assays cannot be directly evaluated for accuracy. Furthermore, the methods are not standardized, which can complicate the comparison of results across different study populations. Of interest, a study comparing six laboratories that used different multi-assay algorithms to determine HHV8 serostatus reported improved but nonetheless variable interlaboratory agreement (Pellett et al., 2003). The use of assays to detect both antilytic and antilatent antibodies is generally recommended for improved sensitivity and specificity of HHV8 antibody detection. Clearly, further development of serologic assays is necessary to facilitate the epidemiologic study of HHV8.

The Major HHV8-Associated Malignancies Kaposi Sarcoma Four variants of KS are currently recognized, which are histologically identical but differ somewhat in their typical epidemiologic and clinical characteristics (Tappero et al., 1993). HHV8 is now considered to be a necessary cause of all four forms of KS (Moore and Chang, 1998). However, KS is a rare outcome of HHV8 infection, and thus cofactors must cooperate with HHV8 in KS pathogenesis. “Classic” or “sporadic” KS is the condition first characterized by Moritz Kaposi in 1872. Patients present with a pigmented skin lesion of multiple nodules and plaques that appear first on the lower limbs and may spread to arms, face, and torso over a period of years to decades. In a minority of cases, the lesions progress to involve visceral mucosal surfaces (Antman and Chang, 2000). Classic KS is rare and is diagnosed primarily in elderly men of Mediterranean or Eastern European origin. Its occurrence has been attributed in part to immune suppression related possibly to age, host genetics, a history of other cancer, or to regional factors such as malaria (Kaloterakis et al., 1995; Calabrò et al., 1998; Iscovich et al., 1999). Recent evidence from both epidemiologic (Touloumi et al., 1999; Goedert et al., 2002) and laboratory in vitro studies (Fiorelli et al., 1998; Monini et al., 1999) appears to contradict this hypothesis by suggesting that immune activation and inflammation predict classic KS, although these immunologic states may not be mutually exclusive. Of interest, cigarette smoking, which has immune-suppressive effects, has been found to be inversely associated with risk of classic KS (Goedert et al., 2002; Hatzakis A, personal communication). Of note, chronic inflammation was also implicated as a determinant of HHV8 viral load among asymptomatic HHV8-seropositive subjects from this Italian study population (Brown et al., 2005). In addition, infrequent bathing, a history of asthma, and, in men only, a history of allergy were linked with classic KS among Italian subjects who were HHV8 antibody seropositive (Goedert et al., 2002). Of note, HHV8-seropositive subjects who use topical steroids may also have an elevated risk, although it could not be determined whether the association reflected an effect of the underlying dermatitis or of the steroid use itself. Regardless, this association suggests that localized immunologic events in the skin also contribute to KS pathogenesis (Goedert et al., 2002). In contrast to EBV models, age and route of exposure to HHV8 does not appear to modulate risk of progression to classic KS (Goedert et al., 2002). These findings warrant confirmation in other HHV8-seropositive populations. A second KS variant, called “endemic” or “African” KS, is so named because of its high prevalence in sub-Saharan Africa. For decades, endemic KS has comprised a substantial proportion of all tumors diagnosed in some Central African countries, such as Uganda and Zambia (de-Thé et al., 1999; Antman and Chang, 2000). Endemic KS affects HIV-negative adults and children, in contrast to other variants. In adults, the condition is more common in men than in women and follows a relatively indolent clinical course, similar to classic KS. In children, however, the disease often involves lymph nodes in addi-

tion to the typical cutaneous sarcoma lesions and is more aggressive (Ziegler and Katongole-Mbidde, 1996). Predictors of endemic KS risk among HHV8-seropositive persons are not well characterized. Factors indicative of relative affluence were associated with an elevated risk of endemic KS in Ugandan cancer patients, suggesting a role for delayed age at primary infection with HHV8 (Ziegler et al., 2003). It is noteworthy that rarely or never wearing shoes was one of the lifestyle factors associated with endemic KS in the latter population; the practice of going barefoot likely reflects active farming and is therefore not inconsistent with an affluence-related effect on KS risk (Ziegler et al., 2003). This effect may reflect localized immune suppression of the skin due to blockage of the lower lymphatic system by fine soil particles, which can pass through the barefoot skin. If confirmed, this finding would appear consistent with the reported association of topical steroid use with classic KS (Goedert et al., 2002) and would thus expand the evidence that localized immune dysregulation of the skin is important in KS pathogenesis. A third variant, “iatrogenic” KS, occurs in patients undergoing immunosuppressive therapy after organ transplantation or for a variety of other conditions. KS is a rare outcome of iatrogenic immune suppression (Antman and Chang, 2000) but is diagnosed more frequently in patients at risk for classic KS, such as those of Mediterranean origin (Franceschi and Geddes, 1995). It is not known whether this is due to an increased likelihood of HHV8 infection in these patients or to other etiologic cofactors related to birthplace or ethnicity. The clinical course of iatrogenic KS can be relatively aggressive; however, some patients have been reported to regress upon withdrawal of immunosuppressive therapy (Sarid et al., 1999). The latter observation is compelling evidence that immune suppression is a significant risk factor for KS development. Thus, immune activation and immune suppression may each influence the natural history of HHV8 infection in some undescribed manner (Touloumi et al., 1999). The fourth KS variant, known as “epidemic” or AIDS-associated KS, occurs in HIV-positive persons, primarily in homosexual men, and is the most prevalent type of KS in many countries. Epidemic KS can also be the most clinically aggressive and disseminated type, affecting lymph nodes and visceral mucosal tissues as well as the skin (Sarid et al., 1999). There is strong epidemiologic evidence for an association of HHV8 with AIDS KS. Seroconversion to HHV8 antibody positivity both precedes and predicts development of AIDS-KS, which often occurs within 5 to 10 years in HIV-positive homosexual men (Renwick et al., 1998; Jacobson et al., 2000). The rate of progression to KS was substantially shorter when HIV-1 infection preceded seroconversion to HHV8 (Renwick et al., 1998; Jacobson et al., 2000), which may reflect promotion of KS development by HIV-related immune suppression. The latter interpretation is consistent with the findings that markers of more advanced immune suppression, including decreased CD4+ cell count (Renwick et al., 1998; Jacobson et al., 2000) and increased HIV-1 viral load (Jacobson et al., 2000), were independent predictors of KS in these subjects. However, Nawar and colleagues (2005) recently reported an apparent protective effect of cigarette smoking on KS risk among HHV8-seropositive homosexual men with AIDS, consistent with prior observations in classic KS and with a role for immune activation in epidemic KS etiology. The activation of HHV8 by the HIV-1 Tat protein is also believed to directly affect the proliferation of HHV8-infected cells, further enhancing HIV-1–associated risk of KS (Ariyoshi et al., 1998; Fiorelli et al., 1998). In addition, as suggested for endemic KS, relative affluence may increase the risk of AIDS-KS in HHV8- and HIV-I-seropositive men (Nawar et al., 2005). Table 26–4 summarizes the major risk factors that have been reported for KS. However, some of these associations are based on sparse data and require confirmation. Collectively, the observed predictors of the four types of KS suggest that immune dysfunction is a central factor in the pathogenesis of KS. Consistent with this view, diminished CTL responsiveness to HHV8 antigens is associated with KS pathogenesis (Osman et al., 1999), whereas restored NK cell cytotoxicity appeared to explain KS regression in AIDS patients successfully treated with highly active antiretroviral therapy (Sirianni et al., 2002). However, some of the

523

Infectious Agents Table 26–4. Risk Factors for KS, and the Strength of the Associations, as Reported in Epidemiologic Studies Risk Factor Demographic Male gender* Mediterranean origin*

Association Moderate to strong positive Moderate to strong positive

(KS Subtype; Comment)

References

(Classic, endemic, and epidemic) (Classic and iatrogenic; also includes persons of Jewish or Arabic ancestry) (Classic KS, including rare familial cases)

Franceschi and Geddes (1995) Franceschi and Geddes (1995)

Jacobson et al. (2000) Renwick et al. (1998) Goedert et al. (2002) Touloumi et al. (1999) Touloumi et al. (1999)

HLA-DR5* Immune-related HIV-1 seropositivity

Moderate to strong positive

Topical steroid use Serum neopterin Serum b2-microglobulin Medical history Cancer* Asthma Allergy Other factors Cigarette smoking

Strong positive Strong positive Strong positive

(Risk greatest when HIV infection precedes seroconversion to HHV8 antibody seropositivity) (Classic) (Classic) (Classic)

Strong positive Strong positive Strong positive

(Classic; strongest for history of lymphoma, leukemia) (Classic) (Classic, in men only)

Iscovich et al. (1999) Goedert et al. (2002) Goedert et al. (2002)

Strong inverse

(Classic; evidence for dose-response relationship)

Strong positive Moderate to strong positive

(Endemic) (Endemic; includes higher household income, younger age of leaving home, ownership of pigs or goats) (Classic and endemic)

Goedert et al. (2002); Nowar et al. (2005); A Hatzakis, personal communication Ziegler et al. (2003) Ziegler et al. (2003); Nowar et al. (2005) Franceschi and Geddes (1995)

Frequently barefoot Relative affluence Malaria-endemic area*

Strong positive

Moderate positive

Kaloterakis et al. (1995)

KS, Kaposi sarcoma. *Association established prior to the discovery of HHV8 or without consideration of HHV8 serostatus. Because of lack of restriction to HHV8-seropostives, some estimates may in part capture the risk of exposure to HHV8 rather than that of malignancy, whereas other associations may be underestimated.

immunity-related evidence appears contradictory, implicating immune activation in KS pathogenesis. Emerging evidence from in vitro studies suggests an important role for inflammatory cytokines and in particular for IFN-g in the early development of KS (Fiorelli et al., 1998; Monini et al., 1999). It is not clear how inflammatory and immune suppressive forces interact to result in KS nor is it known how HHV8 gene expression contributes to pathogenesis. To date, the characterization of HHV8-encoded genes suggests a partial overlap in the viral strategies for immune escape and induction of cell proliferation and thus offers unique insights into the interrelationship between viral immune evasion and tumorigenesis (Moore and Chang, 2003). These issues warrant further investigation.

Primary Effusion Lymphoma Primary effusion lymphoma (PEL) is a rare subtype of NHL that is characterized by detection of HHV8 DNA as well as other highly stringent diagnostic criteria (Cesarman et al., 1995, 1996). In brief, the malignancy is a lymphomatous effusion of apparent B-cell origin occurring in the pleural cavity and pericardium, with no discernable tumor mass. The vast majority of cases occur in AIDS patients, sometimes as a second malignancy after KS. The presence of HHV8 distinguishes PELs from other body-cavity–based lymphoma types (Ablashi et al., 2002). Most PEL lesions also harbor EBV. The pathogenetic role of HHV8 and EBV in PEL is not known, although expression of viral oncogenes is likely involved (Ablashi et al., 2002). The rarity of PEL even in HHV8-seropositive persons, however, suggests that strong cofactors affect the risk of this malignancy.

Multicentric Castleman Disease Multicentric Castleman disease (MCD) is the plasma cell subtype of Castleman disease, an atypical lymphoproliferative disorder that is poorly understood (Ablashi et al., 2002). MCD is considered a polyclonal, non-neoplastic disorder but is associated with immune dysregulation and with a high risk of malignancy, and in particular of KS and NHL. MCD occurs more commonly in older persons and in men. In AIDS-associated MCD, nearly all cases harbor HHV8, as do approximately half of the HIV-negative cases of MCD (Soulier et al., 1995; Ablashi et al., 2002). The role of HHV8 in MCD pathogenesis is not understood but may include an effect of the HHV8-encoded

functional homologue to the cytokine IL-6, a potent B-cell growth factor. Both human and viral IL-6 (vIL-6) are expressed in the lymphoid tissues of MCD patients.

Other Associations Rettig and colleagues (1997) created a stir when they reported finding HHV8 DNA by PCR in bone marrow stromal cells from 15 of 15 patients with multiple myeloma (MM). Specimens from two of eight patients with monoclonal gammopathy of uncertain significance (MGUS), a nonmalignant condition with a high risk of progression to MM, were also PCR-positive for HHV8 genome (Rettig et al., 1997). Because the cytokine IL-6 is an important growth factor for MM, and HHV8 encodes vIL-6, Rettig et al. (1997) proposed that HHV8 in the bone marrow microenvironment contributes to plasma cell transformation by paracrine secretion of vIL-6. Despite the biological plausibility of an association of HHV8 with MM, however, the majority of epidemiologic evidence does not support a link between HHV8 and MM (Ablashi et al., 2002). In addition to MM, HHV8 DNA was reportedly detected in diagnostic tissue specimens from a variety of diseases, including pemphigus, bullous pemphigoid, a variety of skin tumors, sarcoidosis, and Kikuchi disease. However, these reports have been disputed in the literature, and the collective evidence remains either inconclusive or not supportive of a causal role for HHV8 in the pathogenesis of these diseases (Ablashi et al., 2002).

Prevention and Future Research With inadequate knowledge of the routes of transmission of infectious HHV8, it is difficult to propose a comprehensive strategy to prevent infection. Nevertheless, an important component of such a strategy is clearly the avoidance of behaviors that increase the risk of HIV-1 infection. The possibility of interventions to prevent nonsexual modes of transmission of HHV8 are less certain, although it has been suggested, based on what is understood of the natural history of KS, that a vaccine and immunotherapies are worthy of exploration (Sarid et al., 1999). Therefore, a more precise understanding of routes of sexual and nonsexual horizontal transmission of infectious HHV8 virions remains an important focus for future research. The related question of host factors that confer susceptibility to infection, once exposed, also warrants study.

524

PART III: THE CAUSES OF CANCER

In addition, little is known about host control of HHV8 replication. In particular, studies of the host response to primary infection, including the impact of age and route of initial infection, would contribute valuable information. The serologic profiles pertaining to HHV8 reactivation during latent infection also warrant characterization, as do temporal patterns and determinants of viral load. Given the implied roles for both immune suppression and immune activation in KS pathogenesis, prospective epidemiologic studies that incorporate biomarkers of each of these processes will be valuable for clarifying the immunologic parameters that predict risk of KS. The opportunity to conduct these proposed studies will be heavily dependent on applications of biotechnology that permit more effective quantification of serum HHV8 antibodies and viral load. Perhaps in part because the clonality of KS lesions has been long disputed (Ablashi et al., 2002), few studies have addressed whether HHV8 itself is clonal within a KS lesion as well as among KS lesions in a given patient. The limited data available suggest that many KS lesions contain either oligo-or monoclonal HHV8 (Judde et al., 2000; Gessain and Duprez, 2005). These data are consistent with an early etiologic role for HHV8 in KS pathogenesis and with the view that KS progresses from a polyclonal disease to a monoclonal malignancy (Gessain and Duprez, 2005). Further evaluation of HHV8 clonality may thus help to resolve the debate over the clonality of KS lesions and contribute important insights into the pathogenesis of HHV8related malignancies. The informativeness of the future epidemiologic studies of HHV8 will likely be amplified by continuing advances in the areas of molecular biology, virology, and immunology. Indeed, multidisciplinary approaches to resolving the mysterious aspects of HHV8 infection are especially warranted. As a more comprehensive understanding of HHV8 natural history and pathogenesis emerges, the opportunity to develop effective preventive measures, including appropriate public health messages and immune-based preventive therapies, to reduce the global burden of KS and associated diseases may be substantially enhanced.

HUMAN PAPILLOMAVIRUSES Papillomaviruses cause warts (papillomas) that are either flat or raised. These ubiquitous viruses have evolved slowly with their animal hosts, and the infections are species-specific. Papillomavirus infections are virtually always benign. However, persistent infections with oncogenic genotypes of human papillomaviruses (HPVs) cause almost all cases of cervical cancer and a smaller fraction of other cancers, totaling a half million cases per year worldwide (Pisani et al., 2002; Bosch and de Sanjose, 2003). This brief review will describe the epidemiology of oncogenic HPV infections, with occasional comparisons to the even more common non-oncogenic types. The interested reader is referred to the chapters on anogenital and oral cancers for fuller discussion of HPV related to specific neoplastic outcomes. HPVs are small, nonenveloped, double-stranded DNA viruses of approximately 8000 bases (Munger, 2002; zur Hausen, 2002). The assembled 55-nm virus has icosahedral symmetry with 72 capsomers. The genome is circular, and only one strand contains open reading frames that are transcribed. There are eight open reading frames and an upstream regulatory region (URR), also called the long control region (LCR), or noncoding region (NCR), which contains an origin of replication and cis-acting transcriptional regulatory elements. The early region of the HPV genome contains six open reading frames (E1, E2, E4, E5, E6, and E7) that encode proteins necessary for viral replication and cell transformation. E6 and E7 are currently considered the major transforming proteins of the oncogenic types of HPV (zur Hausen, 2000, 2001; Munger and Howley, 2002), but E5 may also play a role (Fehrmann and Laimins, 2003). The late region codes for the two structural proteins of the viral capsid: L1, the major structural protein, and L2, the protein link to encapsulated DNA. With regard to taxonomy, the family Papillomaviridae is subdivided into two major groups, cutaneous and mucosal/genital, based on data from epidemiology, laboratory assays, genetic evolutionary analyses,

and clinical manifestations. Each of these groups is composed of several species (a, b . . .). Within species different types (not “strains”) are defined based on proportion of DNA sequence homology (de Villiers, 2001; Schiffman et al., 2005). Different types share less than 90% homology in the L1 region, which tends to be genetically conserved. HPV types are numbered chronologically in order of characterization, not relatedness. There are considerably more than 100 HPV types, of which approximately 90 have already been characterized and assigned numbers (de Villiers, 2001). The process of finding and characterizing novel HPV types continues; however, the types now being identified are mainly commensal and do not contribute meaningfully to cancer burden, with the possible exception of skin cancer. Within HPV types, even finer evolutionary branchings, called variants, have been defined and associated with oncogenicity (Yamada et al., 1997; Hildesheim et al., 2001; Xi et al., 2002). Epidemiologists make use of the genetic species of HPV to permit analyses that lack statistical power if each type were considered separately. Alternatively, the associations are presented for groupings of HPV types and disease. Most mucosal types cause benign flat lesions or no discernible pathology and are termed “low risk.” A few mucosal types (HPV 6 and 11 primarily) cause benign, exophytic, anogenital warts called condyloma acuminatum (Wiley et al., 2002). The proven oncogenic HPV types are in the mucosal group but not all in one clade. The a9 and a7 species, containing HPV 16 and HPV 18, respectively, are the most important. The oncogenic types include HPV 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, 68, and possibly types 26, 73, and 82 (Bosch et al., 1995; Munoz et al., 2003; Cogliano et al., 2005). HPV 16 is uniquely prevalent and carcinogenic and causes half of the cancer burden (Bosch and de Sanjose, 2003). Animal and human papillomavirus-related lesions have been recognized for more than a millennium (Burns, 1992). In the twentieth century, cottontail rabbit and bovine papillomaviruses provided the most important animal models of papillomavirus infections and viral carcinogenesis (Lancaster and Olson, 1982). Transmission of human warts by cell-free filtrate was first demonstrated in 1907. The personto-person infectivity and latency of genital HPV infections was established by the incidence of genital warts in women, 4 to 6 weeks after the return of their husbands (who had new penile warts) from military service (Barrett et al., 2003). Studies of HPV and cancer have predominantly addressed cervical neoplasia because of its high prevalence. The relationship of cervical carcinoma to sexual behavior, particularly lifetime number of sexual partners, has been suspected for over a century and well-established by epidemiologic studies since the 1960s (Brinton and Fraumeni, 1986). Advances in HPV-cancer cytopathology occurred over the last half of the 20th century as various groups described the preinvasive abnormalities related, on one hand, to warts and HPV, and on the other hand, to risk of subsequent invasive cervical cancer (Meisels, 1969; Richart and Barron, 1969; Koss, 1989). The molecular biological association of HPV infection and invasive cervical carcinoma was first demonstrated by zur Hausen and colleagues (2002) who found HPV DNA of previously uncharacterized types (now called HPV 16 and HPV 18) in cervical carcinomas. The early molecular epidemiologic studies of HPV in the 1980s established the plurality of HPV types and the association of specific types with different tissues and degrees of neoplasia (Fuchs et al., 1988; Lorincz et al., 1992).

Natural History of HPV Infection All HPV infections are usually transmitted by person-to-person, skinto-skin contact, as supported by the characteristic age-specific prevalences of outcomes of different HPV infections; for example, common cutaneous warts in young children (Vittorio et al., 1995). Male and female genital HPV infections (as measured by DNA, cytologic diagnosis, or presentation of overt genital warts) peak in early adulthood concurrent with the age of usual onset of sexual intercourse (Schiffman, 1992a; Kjaer et al., 2001). The appreciable prevalence of anogenital HPV infection, as measured by DNA, in women reporting only one lifetime sexual partner, mainly reflects their partners’ sexual experiences with other partners (Castellsague et al., 2002). As a rule,

Infectious Agents completely virginal women are not infected at the cervix, although nonintromissive sexual behaviors can lead to anogenital transmission (Winer et al., 2003). Autoinoculation and spread appears to be possible. Nonsexual transmission of anogenital infection is probably uncommon, although the point remains controversial (Dillner et al., 1999; Rice et al., 1999). Fomite transmission is possible but probably uncommon (Strauss et al., 2002). Vertical transmission of HPV types 6 or 11 during vaginal birth rarely cause laryngeal papillomas, which can be serious if they threaten airway obstruction (Silverberg et al., 2003). HPV infection can lead to a wide range of epithelial response, from no apparent lesion to flat or raised warts to invasive carcinoma. There is no known viremia or infection of nonepithelial tissues. Tissue effects are discussed more fully in other chapters as they relate to particular malignancies. The spectrum of HPV-related epithelial abnormalities leading to cancer was formerly considered a stepwise progression of increasingly severe intraepithelial neoplasia: grade 1 to grade 2 to grade 3 (including carcinoma in situ) to cancer (Richart and Barron, 1969). More recently, the biological meaning of this microscopically evident pathologic continuum has been questioned (Schiffman and Kjaer, 2003) and, for epidemiologic studies, it is more reliable and analytically useful to broadly classify tissue effects as acute HPV infection, with or without, associated lesions (which can be pathognomonic but are more often variable), precancer, or cancer. Acute HPV infections, even with potentially oncogenic, mucosal types, are usually self-limited. Cervical HPV infections, as defined by DNA detection, usually disappear within 2 years of first incident appearance (Richardson et al., 2003). However, for already prevalent infections, defined at cross-sectional screening, the median time to disappearance is approximately 1 year (Ho et al., 1998). The key immunity involved in the clearance of HPV infections is cell-mediated (type 1) response (Kadish, 2001; Berry and Palefsky, 2003; Pinto et al., 2003). Multiple warts of a single kind tend to regress concurrently, associated with an infiltration by antigen-presenting cells and lymphocytes. HIV/AIDS and iatrogenic immunosuppression, as part of organ transplantation, provide important insights into the effects of immunosuppression on HPV infection (Palefsky and Holly, 2003). These examples have established an association between immunosuppression, HPV infection, and anogenital precancerous lesions. An effect of immunosuppression on the rate of progression to invasive carcinoma is less clear. Recent data from two HIV cohorts of women indicate that immune status, as measured by CD4 counts, is more weakly linked to the detection of prevalent and incident HPV 16 cervical infection than other HPV type cervical infections (Strickler et al., 2003). These data suggest that HPV 16 is unique in avoiding immune surveillance and may help explain why HPV 16 is the most persistent HPV type, and the one linked to about half of cervical cancers worldwide. Antibodies against the HPV are detected in some, but not all, infected women (Wideroff et al., 1999; Carter et al., 2000). HPV persistence, lasting for years, has been directly linked in cohort studies to the development of precancer (Ho et al., 1995; Nobbenhuis et al., 1999). Precancer develops in <10% of women with point prevalent oncogenic HPV infection; in comparison, the risk of development of precancer among women with oncogenic HPV infections that persist for 5 years or more is much greater (Schiffman et al., 2005). The sine qua non of oncogenic types is that they produce precancer when they persist, whereas low-risk types rarely do. As an important etiologic clue, HPV 16 persists longer than other HPV types and also is most likely to cause precancer when it does persist; the absolute risk at 5 years approaches 40%. Cancer can be considered the most serious known outcome of HPV infection. Cervical cancer is by far the most common HPV-related cancer but still is rare compared to infection. The proportion of other anogenital cancers and oropharyngeal cancer attributable to HPV infection is lower than for cervical cancer, that is, they may be more multifactorial than cervical cancer (Herrero, 2003; Schiffman and Kjaer, 2003). The possible role of poorly defined cutaneous types of HPV in skin cancer etiology (Pfister, 2003) is discussed in the chapter on skin cancer.

525

Epidemiology of Infection Accurate estimates of the prevalence and incidence of HPV infection must be referred to a particular epithelial tissue, a particular HPV type or group of types, a specific specimen collection and testing protocol, and a defined population. Nonetheless, HPV is the most common sexually transmitted infection, and tens of millions of sexually active individuals have been infected at some time with at least one type of oncogenic HPV (Franceschi et al., 2002). Although infection is relatively benign it is a persistent infection that raises concern. However defined, the prevalence of oncogenic cervical types in a population depends on the age and sexual practices of the population (Bauer et al., 1993). In general, young sexually active individuals appear to experience the highest prevalence of oncogenic anogenital HPV infections, consistent with an “epidemic curve” after first sexual exposure. Again, HPV infections of the cervix have been best studied (Fig. 26–6). The drop in cervical HPV prevalence in women past their 20s is likely due to clearance or suppression of existing infections, combined with less exposure to new HPV types because of fewer new sexual partners. However, persistant infections increase with age (Castle et al., 2005). Of note, the point prevalence of HPV DNA in surveys of immunocompetent (HIV-uninfected) prostitutes is not always elevated, suggesting immunity in some women following intense exposure (Kjaer et al., 2000). Although it is unwise to think in terms of a simple summary estimate of anogenital HPV prevalence without reference to age and sexual behavior, oncogenic HPV DNA prevalences of 15–30% are typical, using sensitive detection methods, with an annual incidence of 10% or more among sexually active young women (Franco et al., 1999; Sellors et al., 2003). HPV 16 is the most common type in health and disease, perhaps because it persists effectively (Herrero et al., 2000; Munoz et al., 2003). The prevalence of concurrent infections with multiple types of HPV typically approaches 20–30% of infected women, using the best PCR methods (Herrero et al., 2000). It is not clear whether HPV types influence each other’s presence either directly or via the host immune system, but the available evidence suggests that each HPV type can be viewed as a separate infection with its own natural history (Thomas KK et al., 2000; Liaw et al., 2001). HPV infects the anogenital epithelium broadly. Vaginal and vulvar (introital) HPV infections are just as common and varied in type

Figure 26–6. The age-specific prevalence of oncogenic types of HPV infection varies by region, for reasons that are not yet understood. For example, among 20,810 women in the Portland Kaiser cohort (Sherman et al., 2003), the prevalence of oncogenic HPV types as measured by Hybrid Capture 2 decreased steadily with age until the oldest age groups. In contrast, among 9165 women in the Guanacaste Project in Costa Rica, the prevalence of the same 13 types (measured at similar analytic sensitivity to Hybrid Capture 2 using consensus primer PCR) turned back up in middle age. (Herrero et al., 2005).

526

PART III: THE CAUSES OF CANCER

as cervical HPV (Winer et al., 2003). Common HPV-induced penile lesions can be found in the urethral meatus, on the glans and shaft of the penis, and on the scrotum (Barrasso, 1992; Bleeker et al., 2002; Franceschi et al., 2002; Baldwin et al., 2003). The available data suggest that penile HPV prevalence is comparable to cervical prevalence. Anal HPV infection, linked to an increased risk of anal neoplasia, is also extremely common in both sexes, particularly, but not at all exclusively, among individuals who practice intromissive anal intercourse (Frisch et al., 1997; Palefsky et al., 1998). Large-scale HPV DNA studies began in the 1980s and, thus, there are no long-term data addressing historical trends. Epidemiologists would like to relate the predictors of HPV prevalence to the known demographic risk factors for cervical cancer. Long-noted correlations within the United States of cervical cancer with religion (e.g., decreased rates in Jewish women), race (increased rates in African Americans), SES (increased rates in poorer women), and occupation (husbands away from home) are concordant with the known predictors of HPV as a sexually transmitted infection (Bauer et al., 1993). However, decades after the “sexual revolution,” HPV is now so prevalent that all women with even a few sexual partners have a substantial risk of exposure. The historical epidemiologic associations are blurring. HPV infection is common globally, and its role in anogenital carcinogenesis does not vary by region (Bosch and de Sanjose, 2003; Munoz et al., 2003). However, interesting subtle variations occur. The geographic distribution of HPV DNA detection has been studied mainly in correlation with cervical cancer incidence rates, and some expected ecologic associations have been seen (Bosch and de Sanjose, 2003). In many study populations, the prevalence of current HPV infection, as measured by DNA or cytology, decreases strongly with age as it does in the United States, from a peak at 15–25 years. But in other regions, there is no decrease or a secondary upturn at older ages is seen (Schiffman and Kjaer, 2003), a pattern that presents one of the remaining challenges to HPV epidemiologists (Fig. 26–6). Varying age trends in HPV prevalence might be rooted in differing societal sexual practices, immunologic senescence, and/or cohort effects. There is emerging epidemiologic evidence for each. (Castle et al., 2005).

HPV Biomarkers After molecular biologists demonstrated in small studies that HPV DNA was present in cervical cancer tissue, it took several years to adapt HPV DNA measurement techniques to permit testing of noninvasively obtained cervical specimens collected at the time of pelvic examination by swab, scrape, brush, or lavage (Schiffman, 1992b). During the development phase, the adverse impact of HPV exposure misclassification on early studies of HPV and cervical neoplasia was so profound that it merits consideration by all epidemiologists using newly developed assays that are incompletely validated (Franco, 1992). HPV infection is most often measured by HPV DNA detection, although the detectability of HPV DNA in a single specimen (point prevalence) clearly differs from lifetime exposure to HPV (cumulative incidence). The improved DNA testing techniques now being used by most epidemiologists (Iftner and Villa, 2003) to assay all oncogenic types at once tend to yield roughly comparable results when performed expertly on adequate cervical cytologic specimens. Keratinized epithelial surfaces, such as the vulva, penis, or nongenital skin, have proven more difficult to test reliably. Excluding in situ methods, which have rarely been used by epidemiologists, there are two categories of HPV DNA tests: those that identify nucleic acids directly, and those that amplify nucleic acids first and then detect the amplified product. In the first category currently is Hybrid Capture 2, or HC2 (Schiffman et al., 2000; Terry et al., 2001). The amplification methods currently used for HPV epidemiology are PCR-based (Castle et al., 2002a; van Dorn et al., 2002; Gravitt et al., 2003; Iftner and Villa, 2003; Schiffman et al., 2005). Any data based on older techniques should be interpreted with caution. Very sensitive PCR techniques are available for research on extremely low levels of particular HPV types but have not been validated to test for all oncogenic types concurrently at a clinically relevant level. HPV serologic assay development and validation have lagged behind advances in DNA testing (Wang and Hildesheim, 2003).

Lacking a source of abundant, native viral antigen, virus-like particles (VLPs) have been synthesized for many types and used particularly to test case-case series and archived, premorbid sera in case-control studies of invasive cancers, especially from the huge Scandinavian serology banks (Bjorge et al., 1997, 2002; Strickler et al., 1998; Mork et al., 2001; van Doornum et al., 2003). Seropositivity is a useful type-specific measurement of exposure, although 25–55% of known infected women do not become persistently seroreactive (Wideroff et al., 1999; Carter et al., 2000). Newer, apparently more sensitive assays might improve this situation (Pastrana et al., 2004). HPV infections are limited to the epithelia and apparently induce weak immune responses. Though likely to be important, the mucosal antibody response to epithelial HPV infection is poorly understood.

Cervical Cancer Case-control studies in dozens of countries have confirmed, with virtually no exceptions, the extremely strong association of oncogenic HPV DNA detection with risk of cervical cancer and intraepithelial precursor lesions (Bosch et al., 2002). More than 90% of cervical cancers and precancers contain HPV DNA, with the inclusion of “possible” infections raising the proportions even higher (Bohmer et al., 2003; Munoz et al., 2003). The cancer-associated types agree well with the relative transforming properties of the viral types as defined in vitro and with “phylogenetic” studies grouping HPV types by genetic relatedness. Case-control studies of HPV variants show, remarkably, that risk of precancer and cancer is associated with specific variants, reaching prevalence ORs in the thousands (Wang and Hildesheim, 2003). The prospective evidence that oncogenic HPV infection precedes and predicts risk of cervical precancer and cancer is also strong and consistent (Wallin et al., 1999; Sherman et al., 2003). A single HPV measurement confers an increase in risk of approximately 10- to 30fold over the following decade, whereas persistent oncogenic HPV infection is even more tightly linked to risk of precancer (Kjaer et al., 2002). HPV16 and HPV18 are particularly important carcinogens (Khan et al., 2005). In summary, persistent oncogenic HPV infection is virtually necessary for the development of precancer and cancer of the cervix (Bosch et al., 2002). As discussed in the relevant chapters, the etiologic fractions of other anogenital cancers and oropharyngeal cancer due to HPV infection are lower but still substantial, although the available epidemiologic data for causality are weaker, partly because these cancers tend to be rare.

Biological Causal Mechanisms The molecular biology of HPV and other papillomaviruses, as model DNA tumor viruses, is under intensive study. The reader is referred elsewhere for recent reviews, as only a few critical points can be mentioned here (Kadish, 2001). It is assumed that the initial site of HPV infection is the germinal cells in the basal layer of the epithelium, secondary to minor epithelial injuries during sexual contact. HPV-induced lesions appear to be monoclonal, suggesting that each lesion derives from a single infected germinal cell (Park et al., 2003). Early viral transcripts are detectable in the basal and parabasal layers of the epithelium, whereas capsid production and virion assembly occur in the more superficial layers of the differentiated epithelium (Stoler et al., 1992). In early HPVinduced lesions (typically destined to regress), squamous differentiation in the more superficial layers of the epithelium becomes abnormal, but the cells continue to differentiate, such that there is only a minimal effect on the expansion of the proliferative (nondifferentiated or “immortalized”) compartment to approximately one-third or less of the full thickness of the epithelium. The accompanying expansion of the middle portion (spiny layer) of the epithelium characterizing HPV lesions, particularly condylomata, results from a reduced rate of squamous cell sloughing, rather than increased rate of cell turnover. The E6 and E7 proteins are principally responsible for HPV neoplastic effects, via interactions with pRb and p53-related cell-cycle pathways, respectively (Munger and Howley, 2002; Fehrmann and Laimins, 2003). The specific details of these molecular interactions are

Infectious Agents increasingly understood, making HPV an important model of viral carcinogenesis (Broker and Chow, 2001). In most infections, the HPV genome is maintained in the cell nucleus in an episomal state. In the majority of invasive cervical carcinomas, however, integration of HPV DNA into the host genome is found (Ziegert et al., 2003). Integration tends to occur throughout the cell genome. However, with reference to the viral genome, integration is not random. In cancers, the E6 and E7 open reading frames are preserved, with frequent disruption during integration of the E1 and E2 genes that normally inhibit E6 and E7. Thus, continuous production of E6 and E7 proteins appears to have a role in HPV carcinogenicity and, in fact, the E6 and E7 regions of the HPV genome are transcriptionally active in HPV-associated cervical carcinomas and derived cell lines. RNA-based assays for possible clinical use are currently being validated.

Cofactors For unknown reasons, HPV infection tends to cause cancer at boundary areas (“transformation zones”) where one kind of epithelium contacts and sometimes gradually replaces another. The cervix, tonsils, and anus are examples of tissues with transformation zones prone to HPV carcinogenesis. Apart from target tissue, a crucial current question is “What else besides HPV type predicts the risk of viral persistence and progression to precancer?” We hope eventually to study viral persistence and progression to precancer separately and prospectively, but that level of distinction has not yet been achieved. The three categories of factors that seem most likely to be responsible for persistence and progression include (1) viral factors, (2) environmental cofactors, and (3) host factors (Castellsague and Munoz, 2003). Apart from the powerful importance of HPV type and variant, no other critical viral factors have been identified. It appears that very high levels of HPV DNA (often seen in acute benign warts) are not (except for HPV16) more closely linked to persistence and progression than are low levels, although extremely low levels probably reflect an increased tendency toward clearance (Lorincz et al., 2002). Smoking is the environmental cofactor most consistently linked with persistence and progression of HPV infection (Castle et al., 2002b; Castellsague and Munoz, 2003). Smoking has both mutagenic and immunologic effects, and it is unclear which is primarily involved. Other possible environmental cofactors that interact with HPV include chronic inflammation, other sexually transmitted infections like Chlamydia trachomatis, and levels of antioxidant nutrients or folate (Castle and Giuliano, 2003). It is possible that sunlight and HPV interact in the etiology of skin cancer (Pfister, 2003). With regard to host factors that interact with HPV, multiparity increases the risk of progression to cervical precancer and cancer although the mechanism is unclear (Munoz et al., 2002, 2003). Parity might act by influencing immunity or by nutritional, traumatic, or even biochemical mechanisms (Gravitt and Castle, 2001). Long-duration hormonal contraceptive use may increase the risk of precancer and cancer (Moreno et al., 2002; Green et al., 2003; Smith et al., 2003) among HPV infected women if screening and treatment of intraepithelial precursor lesions does not intervene. There is some laboratory support for hormonal influences on HPV (de Villiers, 2003).

Prevention Oncogenic, genital HPV transmission would be prevented by abstinence. However, even conscientious condom use would not entirely prevent the spread of genital HPV infections because genital HPV infections can involve the scrotum and perineum (Manhart and Koutsky, 2002). Ultimately, prophylactic HPV vaccines that induce neutralizing antibodies might permit primary prevention of HPV and its clinical sequelae, including cancers (Lowy and Frazer, 2003). At least two very large and highly promising Phase III efficacy trials of HPV 16 and 18 prophylactic vaccines are now underway, based on virus-like particle antigens produced in insect cells or yeast (Koutsky et al., 2002; Lowy and Frazer, 2003; Harper et al., 2004; Villa et al., 2005). Early therapeutic trials of HPV infection per se, to permit treat-

527

ment of HPV-induced intraepithelial lesions and secondary prevention or even treatment of cervical cancer, are also underway at earlier phases (Eiben et al., 2003). Until polyvalent vaccines are available and widely used for multiple cohorts of women, HPV infections and HPV-induced lesions will still be highly prevalent (Sherman et al., 1998; Hughes et al., 2002). HPV control efforts will be aimed at preventing carcinoma, particularly cervical carcinoma. In the United States, the standard response to the detection of HPV-associated mild lesions is tending toward “watchful waiting” to permit clearance, whereas the treatment of precancerous cervical lesions is destruction of the lesion and the remaining cervical transformation zone usually by loop electrocautery (Cox et al., 2003). Chemopreventive therapies have, so far, been largely unsuccessful (Stanley, 2003). HPV DNA testing is now approved by the U.S. Food and Drug Administration for the clarification of equivocal cytology (“ASC-US” Pap tests) at all ages, and as an adjunct to screening cervical cytology among women 30 and older (Wright and Schiffman, 2003). Because virtually all cases of cervical cancer are caused by oncogenic HPV, the absence of HPV as measured by sensitive reliable tests is very reassuring, especially at older ages when fewer new infections are expected (Franco, 2003).

Future Research Although most future research will likely address the role of HPV in cervical cancer, but it will be important to understand the role of HPV in other cancers. Comparisons should be made of HPV natural history in different tissues. In particular, why are transformation zones so prone to HPV carcinogenesis? It will be very important to compare oncogenic to low-risk types of HPV at the variant level in population-based studies, combining epidemiology and molecular biology. Given the small size of the HPV genome and our reasonable understanding of HPV carcinogenesis, it should be feasible to understand which parts of the viral genome influence risk of persistence and progression from infection to precancer/cancer given viral persistence. As a related issue, through continued prospective studies, the trend of decreasing anogenital HPV infection rates with increasing age should be better understood. Contributions of cohort effects, sexual practices, and immunologic suppression must be distinguished. Secondary biomarkers indicating the interaction of host and virus are needed, particularly as diagnostic assays. Current HPV DNA detection methods, although sensitive, are inadequately specific to be the optimal answer for our testing needs because poor positive predictive value leads to excessive cost and iatrogenic morbidity (von Knebel, 2002; Wang and Hildesheim, 2003). As the highest priority, HPV immunology is likely to occupy epidemiologists studying HPV infection over the next decade. In the immediate future, the interactions of multiple HPV types in mixed infections should be clarified, as one pathway to understanding HPV immunity. Assays of cell-mediated immunity must be developed and applied (Pinto et al., 2003). The association of specific HLA locus haplotypes with the risk of invasive cervical carcinoma might yield indirect evidence of the role of specific aspects of cell-mediated immunity in the natural history of HPV infections (Hildesheim and Wang, 2002). The ultimate goal will be to define the successful immune response to HPV infection, in the context of stimulating cancer-preventive immunity by vaccination.

RETROVIRUSES: HTLV-I The human T-cell leukemia virus type I (HTLV-I) was the first identified human retrovirus. In 1980, Poiesz et al. made the original isolation from a cutaneous lymphoma in an African American. The same virus was isolated independently by Yoshida, Miyoshi, and Hinuma from a cell line established from a Japanese patient with T-cell leukemia (Yoshida et al., 1982). The malignancy, now called adult Tcell leukemia/lymphoma (ATL), an aggressive malignancy of mature T-lymphocytes, was first recognized as a distinct entity by Japanese

528

PART III: THE CAUSES OF CANCER

clinicians in the 1970s (Uchiyama et al., 1977). The syndrome is characterized by hypercalcemia, cutaneous involvement, and depressed cellular immunity (The T- and B-cell Malignancy Study Group, 1981). The evidence that the virus is a causal factor in the etiology of ATL is so compelling that it has been unchallenged: ATL being highly restricted to HTLV-I carriers who have an estimated lifetime risk of ATL of £5% (IARC, 1996). Yet little is known about the natural history of the infection and the role of other factors that affect oncogenesis. HTLV-II, a closely related retrovirus, has also been isolated (Kalyanaraman et al., 1982). This rare infection is primarily endemic among Amerindian groups but has reached a relatively high prevalence among American injection drug users (Liu et al., 2001). Although several early case reports linked HTLV-II infection with atypical hairy cell leukemia (Rosenblatt et al., 1986), this appears to be a very rare occurrence (Orland et al., 2003). HTLV-I is a prototypic retrovirus most closely related to the bovine leukemia virus. Its genome of approximately 9 kb in length codes for a minimal set of structural genes plus several regulatory genes (Cann and Chen, 1996). The coding areas common to all retroviruses are gag (which encodes the core proteins p19, p24); env (envelope glycoproteins gp21, gp46); and pol, which encodes polymerase (reverse transcriptase), endonuclease (ribonuclease), and integrase. A protease is encoded by a reading frame spanning the gag and pol gene regions. The genome is bounded by long terminal repeats. It has an additional reading frame that codes for regulatory proteins, including tax (p40), rex (p27, p21), and several alternatively spliced mRNAs (p12, p13, p30). (The taxprotein was sometimes termed p42 in early publications). HTLV-I preferentially infects and transforms CD4+ T-cells, which express an activated phenotype, although CD8+ T-cells can also be infected (Cann and Chen, 1996). The oncogenic potential of HTLV-I in infected cells is largely orchestrated by the transactivation and other properties of the Tax protein (Johnson et al., 2001). The Tax protein promotes genetic mutation of infected cells, inhibits cell-cycle control, and drives viral gene expression. The Tax protein constitutively activates the NFKB and SRF transcription pathways (Jeang, 2001). It activates key cellular genes, including IL-2, IL-2Ra, GM-CSF, and PTHRP, while down-regulating the b-polymerase and the cdkinhibitory p16ink4a genes. The fact that the integration sites of HTLV-I provirus differ between patients, and that no host oncogene has been identified within the provirus, argues that the mechanism of oncogenesis likely involves transactivation by Tax (Seiki et al., 1983, 1984; Gatza et al., 2003).

Natural History of HTLV-I Infection Unlike other retroviruses, such as FeLV or HIV-1, HTLV-I is a highly cell-associated infection, with predominantly cell-to-cell transmission and virtually no cell-free infectious virus produced in vitro (Igakura et al., 2003). This property accounts for its extremely low level of infectivity. Transmission primarily occurs by perinatal, blood, and sexual exposure. Perinatal infection principally occurs by prolonged breast feeding, presumably via infected lymphocytes. The transmission rate associated with lactation of ≥6 months is correlated with markers of high maternal proviral load; as such, it varies between groups but is generally about 20–30% (Hino et al., 1987; Hisada et al., 2002; Ando et al., 2003b). Interventions on breastfeeding of carrier mothers reduces, but does not eliminate, perinatal transmission (Ando et al., 2003a). The most efficient route of transmission is direct exposure to infected blood containing cellular components. In a prospective study of transmission by transfusion in Jamaica, 24 of 54 recipients of HTLV-I–positive, cellular blood components seroconverted while none of 12 recipients of positive, noncellular donor units, nor of 52 HTLV-I–negative units became infected (Manns et al., 1992). The screening of blood donors in endemic populations essentially eliminates this risk to transfusion recipients. However, the reuse of needles for injection and other applications, such as acupuncture, remains a potential source of infection. Sexual transmission is the most inefficient route, with infectivity of the exposing partner correlated

with markers of high viral load. Transmission is more common from infected men to their female partners than the reverse (Shioiri et al., 1997). In endemic populations, infection rates are quite low and stable among children, reflecting perinatal infection (Kusuhara et al., 1987). There is a slow increase of seroprevalence with advancing age, which plateaus among men at about age 50. Among older women, the seroprevalence continues to increase with age (Tajima et al., 1987; Mueller et al., 1990) (Fig. 26–7). Infection in adults appears to be due primarily to sexual exposure (Murphy et al., 1989; Stuver et al., 1993) and, to a much lesser extent, through transfusion. The divergence seroprevalence curves between genders after age 50 likely reflect the greater probability of heterosexual transmission from men than vice versa in this age group, but whether it is due to greater infectiousness of men, discontinuation of barrier contraception practices, or increased susceptibility following menopause is not known (Stuver et al., 1993). Upon primary infection, HTLV-I provirus integrates into the genome of infected cells, establishing latency. As reviewed by Rosenblatt et al. (1988), the apparent reservoir of latent infection is in peripheral blood T-lymphocytes. These cells can be immortalized by HTLV-I in vitro, where they continue to grow in the absence of extraneous IL-2. The cells immortalized are generally of an activated T-helper cell phenotype, CD4+ and CD25+ (IL-2 receptor a), the phenotype of essentially all cases of ATL. It is thought that the oncogenic potential of HTLV-I infection is counterbalanced by host virus–specific CTLs, many of which are directed against the Tax protein (IARC, 1996; Bangham, 2000). (Of interest, HTLV-I–specific CTL response has been reported in seven of 19 seronegative/PCR negative persons who had known high exposure to the virus but in none of 16 matched controls without risk factors for exposure [Nishimura et al., 1994]). Although carriers make antibodies to various proteins of the virus, these antibodies do not appear to be protective once the infection is established. In fact, antibody titers against structural proteins and against the Tax regulatory protein correlate with proviral load in asymptomatic carriers (Shioiri et al., 1993; Shinzato et al., 1993; Ishihara et al., 1994; Morand-Joubert et al., 1995), suggesting they mirror viral protein expression. ZuckerFranklin et al. (1997, 1998) have reported asymptomatic antibodynegative individuals who are positive only for the tax gene sequences; however, this has not been confirmed (Cowan et al., 1999; Dezzutti et al., 2003). Carriers generally maintain their anti-HTLV-I antibody levels, suggesting that some level of viral replication occurs. However, the extremely low mutation rate of the virus at the population level—estimated at 0.1% per century (Gessain et al., 1992)—indicates that virus replication via error-prone reverse transcription is generally quite limited. Instead, a remarkable feature of HTLV-I infection is that viral

Figure 26–7. Age-specific prevalence incidence of HTLV-1 antibody by gender in Miyazaki Cohort (Japan) and Jamaican food handlers. (Source: Jamaican data adapted from Murphy et al., 1991.)

Infectious Agents expansion in asymptomatic carriers appears to occur predominantly by clonal expansion of infected cells (Furukawa et al., 1992; Wattel et al., 1995). We have shown that such clones can persist in asymptomatic Japanese carriers for years, apparently protected from host CTL response (Etoh et al., 1997). This remarkable biology is shared with the simian T-cell leukemia virus, which is phylogenetically intimately related to HTLV-1 (Vandamme et al., 1998), in its natural hosts (Gabet et al., 2003). Paradoxically, it was reported by Mortreux et al. (2001) that somatic mutations do occur, both within the integrated provirus genomes and in flanking sequences, and accumulate proportionately with the prevalence of clones. The variation among clones could lead to selective evolution toward increased proliferation and potential malignancy. The authors reconcile the stability of the virus on the population level with the instability within infected clones of T-cells in vivo by proposing that the majority of somatically mutated proviruses are without virus progeny. The clinical effects of primary HTLV-I infection are not characterized. The spectrum of HTLV-I antibodies evolves relatively quickly in seroconverters; however, the appearance of anti-Tax is delayed by a number of months (Manns et al., 1991). Antibody titer level and proviral load appear to stabilize within a few years after seroconversion (Okayama et al., 2001; Manns et al., 1999b). Of note, asymptomatic carriers show evidence of diminished cell-mediated immunity (Tachibana et al., 1988; Murai et al., 1990). The disease outcomes of HTLV-I infection appear to be influenced by both age at infection and gender. Tajima and Hinuma (1984) first proposed that early age at infection may be a modifier of risk for ATL, as the sex ratio among ATL cases in Japan is essentially equal (with some male predominance) (Fig. 26–8), similar to the ratio of seroprevalence among children after perinatal infection, and quite different from that among older persons. Conversely, transfusion-acquired infection in adults is associated with the occurrence of HTLV-I–associated myelopathy (HAM; also termed tropical spastic paraparesis, or TSP), which is much more common in women (Mueller and Blattner, 1997). The viral markers in HAM/TSP patients include high proviral load, elevated antibody titers against the virus and against Tax, and an elevated level of CTL against the Tax protein, consistent with its autoimmune nature (Nagai et al., 2001).

Epidemiology of Infection The prevalence of HTLV-I infection is quite geographically restricted and exhibits a high level of clustering within endemic areas, consistent with interfamilial transmission (Stuver et al., 1992). Overall, it is estimated that there are 15–20 million HTLV-I carriers in the world. Areas with reported pockets of higher seroprevalence (≥15 %) include the southwestern islands of Japan, the Caribbean, parts of South

Figure 26–8. Age-specific incidence of ATL in Kyushu Prefecture, Japan, 1986–1987, by sex. (Source: Adapted from Tajima et al., 1990.)

529

America, intertropical Africa, the Mashhad region in Iran, and Melanesia (IARC, 1996). The epidemiology of HTLV-I infection presumably reflects the probable historic geographic distribution of infected populations, the limited transmissibility of the virus, and the corresponding low reproductive rate within populations. Phylogenetic analyses of viral isolates from endemic populations suggest that the virus is an ancient infection that originated in the Pacific Rim and spread by migrating populations throughout much of the world (Mansky, 2001). In phylogenetic trees, human isolates of HTLV-I interdigitate with simian counterparts, arguing that extensive interspecies transmission has occurred (Vandamme et al., 1998). There is extremely little variation between HTLV-I strains isolated from geographically diverse populations; for example, isolates from Japan, the Caribbean, and Africa share as much as 96–99% homology (Cann and Chen, 1996). There has been no demonstration of pathogenic differences between isolates from diverse endemic populations. Most of the epidemiologic data that are available on HTLV-I come from Japan and the Caribbean where substantial research has been conducted and to a much lesser degree from Africa. Unfortunately, little is known about the epidemiology of HTLV-I–associated disease in carriers in the United States, given the rarity of the infection. American carriers tend to be descendants or immigrants from endemic populations (Ho et al., 1991; Parks et al., 1991; Kaplan et al., 1993; Dosik et al., 1994).

HTLV-I Biomarkers Highly reliable assays are available and widely used in HTLV-I epidemiologic research. In the United States and Europe, HTLV-I antibodies are most commonly measured by EIAs, using whole virus lysate or recombinant antigens or peptides and in Japan, (Manns et al., 1999a) by the particle agglutination assay. Confirmation is generally done by Western blot, IF, or PCR testing (Matsumoto et al., 1990; Kinoshita et al., 1993; Busch et al., 1994; IARC, 1996). Assays to detect anti-Tax antibodies have been developed (Rudolph et al., 1994). In terms of indeterminate findings, cross-reactivity with Plasmodium falciparum has been reported (Hayes et al., 1991; Mahieux et al., 2000). The direct assessment and quantification of proviral DNA is done using PCR techniques (IARC, 1996). Clonality of the virus can be assessed by Southern blot and by inverse PCR (Etoh et al., 1999). Reverse-transcription PCR can be used to detect mRNA to assess viral activity (Okayama et al., 1997).

Adult T-Cell Leukemia/Lymphoma The strong geographic concordance between HTLV-I endemicity and the incidence of ATL in Japan provided the first major piece of evidence of a causal association. This concordance was carefully mapped by Japanese investigators in a series of four reports from the T- and B-cell Malignancy Study Group, published in the 1980s (Mueller, 1991). The clarity of the association was due to two factors: (1) the high level of geographic restriction and the micro-epidemiologic nature of HTLV-I infection, and (2) the vast majority of ATL in endemic populations occurs in HTLV-I carriers. The molecular evidence of the association was also consistent and persuasive. Antibodies against HTLV-I are detected in sera from nearly all ATL patients in endemic areas (Hinuma et al., 1981; Robert-Guroff et al., 1982). All seropositive cases show monoclonal integration of HTLV-I provirus in their malignant cells (Yoshida et al., 1985). A spectrum of premalignant states in HTLV-I carriers has been described. All are distinguishable by the detection on a peripheral blood smear of abnormal (“flower-like”) lymphocytes, which are characteristic of ATL (Kim and Durack, 1988). These cells are also phenotypically CD4+ and CD25+ and are thought to contain the HTLV-I genome (Matutes et al., 1986). The evolution of these syndromes into ATL has been reported. In an early report, five persons with low levels of circulating abnormal lymphocytes (£2%) were identified, who had presented clinically with either skin or infectious problems. These patients were followed for extended periods of time by Yamaguchi and

530

PART III: THE CAUSES OF CANCER

colleagues (1983); two cases of ATL developed. The same group (Yamaguchi et al., 1988) then described 15 individuals with polyclonal integration of HTLV-I and a low level of circulating abnormal lymphocytes. Fourteen of these were patients who were seen for a variety of complaints suggestive of immune dysfunction. Of these, one case showed a subsequent loss of circulating, detectable, abnormal lymphocytes, while one progressed to ATL during up to 3 years of followup. Kinoshita et al. (1985) followed 18 patients with 10% to 40% circulating, abnormal lymphocytes. Ten of the 18 presented with rash or infectious complications. Of these, 14 had monoclonally integrated HTLV-I genome. Of 13 followed more than 1 year, three developed ATL (mean follow-up, 4.1 years), three had persistent lymphocytosis (4.4 years), and seven had a regression of their abnormalities (2.8 years). The prospective occurrence of ATL within asymptomatic carrier cohorts has provided strong additional evidence for the causal relationship. Iwata et al. (1994) studied a total of 1997 individuals from a community-based cohort who resided in an HTLV-I–endemic island population in Nagasaki Prefecture. The prevalence of HTLV-I positivity was 25%. After a follow-up averaging 5.3 years, two deaths from ATL were identified among the 503 HTLV-I carriers. Similarly, a cohort of 3090 atomic-bomb survivors in Nagasaki, of whom 9% were HTLV-I seropositive, was identified. After a follow-up for a median of 8.9 years, one ATL case had been diagnosed among the 270 carriers (Arisawa et al., 1998). In a population of 23,922 subjects who had either visited the outpatient clinic or had received an annual health check-up at a hospital in Nagasaki between 1985 and 1996, a nested case-control study of ATL was conducted, based on stored serum samples. Among the 4007 (17%) seropositives, a total of 34 incident ATL cases were identified (Arisawa et al., 2002). In each of these prospective cohort studies, no case of ATL occurred in an HTLV-I seronegative subject, emphasizing the specificity of the HTLV-I/ATL association. The question then arises as to the value of viral and other biomarkers in predicting the subsequent risk of ATL among the virus carriers. Hisada et al. (1998) conducted a nested case-control study based on prediagnosis serum specimens from five incident ATL cases and 38 age, gender, and screen-matched HTLV-I seropositive controls from our Miyasaki Cohort Study. We found the OR associated with a twofold dilution of HTLV-I antibody titer was 1.6 (0.94–3.5), while paradoxically the prevalence of anti-Tax was low or undetectable for all five cases for up to 10 years preceding their diagnosis, OR = 0.74 (0.21–1.8). Prediagnosis PBMCs were available for four of these cases. The proviral load was measured and compared to 37 age- and gender-matched carriers. The median number of HTLV-I proviral copies (per 100,000 PBMC) in the cases’ earliest sample was 4930, significantly higher than that of the matched controls, 820. In one case, a clone of HTLV-I–infected cells with an integration site identical to that of the leukemic cells was found to exist up to 8 years preceding diagnosis, using inverse long PCR (Okayama et al., 2004). In the study by Arisawa et al. (2002), there were 27 incident cases with prediagnosis specimens for whom matched controls could be identified to compare for biomarkers in prediagnosis serum samples. A total of 158 controls were included. Blood specimens were tested for HTLV-I antibody, anti-Tax, and soluble CD25 (sCD25). The authors found that elevated levels of sCD25 and high HTLV-I antibody titers were independently predictive of subsequent ATL; the ORs were 20.5 (4.5–194) and 2.9 (0.98–9.5), respectively. Similar to the Hisada study above, the OR for the presence of anti-Tax antibodies was 0.59 (0.15–2.0). The authors found that, in the controls, there was a strong correlation of the presence of anti-Tax with antibody level of anti-HTLV-I as expected, but there was no correlation within the cases.

Biological Causal Mechanisms A general model of pathogenesis of ATL evolves from the virology and epidemiology of HTLV-I (Matsuoka, 2003; Mortreux et al., 2003). The sequential elements in this model include an initial infection early in life, extensive clonal expansion of infected T-cells, the loss of Tax expression, and acquired genetic and epigenetic mutations. The appar-

ent role of early infection suggests that an immature type 1 immunity at the time of infection increases the likelihood of developing a relatively high proviral load, perhaps compromising the subsequent immune maturation in such carriers. In proportion to high proviral load, an extraordinary number of clonal proliferations of HTLVI–infected cells can occur over an extended period of time. The cellular proliferation and genetic instability that are orchestrated by Tax may become autonomous. By chance, the loss of Tax protein production in a proliferating clone then leads to a veil of protection from host CTL. One can envision the system spinning forward, stochastically inducing additional genetic and epigenetic changes that may result in malignancy. The epidemiologic data based on the prospective development of ATL among HTLV-I carriers noted above confirm the hypothesis that higher proviral load is associated with the development of ATL. Of interest, the higher HTLV-I titers among persons who go on to develop ATL are not paralleled with antibody titers against the Tax protein, as is normally seen. Tax is not expressed on ATL cells, and following diagnosis, ATL cases exhibit a dissociation of anti-Tax antibody with HTLV-I antibodies (Yokota et al., 1989), likely reflecting the lack of active Tax protein expression. These two prospective studies confirm this finding and suggest that early in the pathogenesis of ATL, the premalignant clones become independent of ongoing Tax expression in the integrated HTLV-I provirus, thus escaping from CTL surveillance.

Cofactors It has long been noted that HTLV-I–infected men are more likely to develop ATL than women, despite the higher infection rate among women (Kondo et al., 1987, 1989; Arisawa et al., 2000). The explanation for this difference may lie in gender differences in cell-mediated immune competency, which is central to the control of established HTLV-I infection. We have found that among carriers, men are about four times as likely to have diminished responsiveness to tubercilin skin testing—a functional measure of type 1 immunity (Hisada et al., 1999). This observation did not change with adjustment for estimated early age at infection (Hisada et al., 2001). Among asymptomatic carriers, older men are more likely to have high proviral loads. Shinzato et al. (1991) reported that among 134 carriers from Nagasaki, those carriers with a high proviral load were all more than 30 years old and predominantly male. In the Miyasaki Cohort, a predominantly elderly population, we have found that carrier men are more likely to have a higher proviral load and detectable mRNA of the tax/rex gene than women (Tachibana et al., 1992; Okayama et al., 1997). That the effect of older, rather than younger age interacts with gender is underlined by two studies conducted among younger carriers. Etoh et al. (1999) found no difference between the genders for proviral load in a somewhat younger population of 256 Japanese carriers from Kumamoto. In a study of 20 younger asymptomatic carriers from the French West Indies, Wattel et al. (1992) also found no evidence of gender difference.

Genetic Factors Genetic susceptibility to HTLV-I–induced malignancy has been considered. Several investigations have evaluated the HLA phenotypes of ATL patients; however, the findings have been inconsistent (Tajima et al., 1984; Tanaka et al., 1984; Usuku et al., 1988; The T- and B-Cell Malignancy Study Group, 1988; Uno et al., 1988). Usuku et al. (1988), Sonoda et al. (1993), and Yashiki et al. (2001) have defined two subgroups, as identified by HLA phenotype in Kagoshima who differ in their cellular immune response to HTLV-I in vitro. “Low responders” are associated with the occurrence of ATL, while “high responders” appear to be associated with risk of HAM/TSP. The distribution of HLA among normal carriers in the population tends to be intermediate.

Strongyloidiasis It was initially proposed that risk of ATL was associated with undernutrition and with repeated exposure to filariasis in childhood, as both

Infectious Agents of these conditions were common in the past in endemic areas of Japan (Tajima and Hinuma, 1984). Neva et al. (1989) have reported no difference in the prevalence of antibody against Strongyloides stercoralis between HTLV-I carriers and controls in Jamaica. However, there is growing evidence that coinfection is detrimental to the HTLV-I carrier and associated with high proviral load. It appears that coinfected individuals have dysregulated immunity that diminishes their ability to control either infection. Porto et al. (2001) have reported that coinfected individuals’ type 2 immune responses that contribute to the control of the parasites, as measured by serum IgE and IL-5, are diminished. Conversely, Satoh et al. (2003) have found that coinfected individuals have a diminution of type 1 response, as indicated by decrease in anti-EBNA antibody levels for the EBV, with a corresponding increase in HTLV-I provirus load. Gabet et al. reported in 2000 that individuals with coinfection have five times higher proviral load, as compared to HTLV-I carriers who were not infected with S. stercoralis. Further, upon successful treatment for strongyloidiasis in one case, the HTLV-I proviral load dropped substantially and appeared to result from the extensive proliferation of a restricted number of infected clones. In contrast, in another case who had unsuccessful treatment for S. stercoralis, there was no significant effect on proviral load. Thus it appears that strongyloidiasis could enhance the risk of developing ATL in HTLV-I coinfected carriers.

Social Environment Finally, it appears that the strongest factor that influences the risk of ATL within an endemic population is social and economic environment. This assertion is based on the evidence that the relative frequency and age at diagnosis of ATL and HAM/TSP differ markedly between Japanese carriers and those in the Caribbean. This finding does not appear to be due to differences in the virus per se in these two major population groups. Figure 26–9 displays the estimated ageand sex-specific incidence curves of ATL for Japanese and Jamaican carriers (Kondo et al., 1989; Murphy et al., 1989). In Japan, the incidence of ATL peaks at about age 60 for men; among women the incidence is about one-third that of men but also peaks among the elderly (Hisada et al., 2001). In sharp contrast, ATL occurs at a much lower rate in Jamaican carriers and peaks in the forties, with little evidence of a male predominance in incidence. For HAM/TSP, the preponderance of disease burden is reversed, with about a 10-fold higher incidence of HTLV-I–associated neurologic disorder seen in the Caribbean population. In both populations, women are at higher risk for HAM/TSP. In addition, a syndrome termed infectious dermatitis has been identified among perinatally infected Jamaican children

531

(La Grenade et al., 1990). However, this condition has not been reported among infected Japanese children. Thus it appears that in the face of relatively optimum living conditions as in Japan, the effects of HTLV-I infection are less severe than in the face of poor living conditions, which are more prevalent in the Caribbean. In a comparative study of 51 age- and gender-matched HTLV-I carriers from Japan and Jamaica, we compared the viral markers for the two groups. We found that the mean antibody titer and detection of anti-Tax was significantly higher among the Jamaican carriers, while there was no significant difference in viral load (Hisada et al., 2004). The immunologic correlates with these population differences are currently under investigation.

Prevention and Future Research The efficacy of prevention of transmission has been demonstrated for both breast milk and blood-borne exposure. Policies implementing these two interventions in endemic populations should bring the reproductive rate of HTLV-I infection below that needed to maintain endemic infection. Although the risk of sexually transmitted infection is low, it is nonetheless real, requiring sensible and sensitive counseling for discordantly infected couples. In terms of epidemiologic research, what is essentially uncharted is the characterization of the immune profile of carriers in relation to viral markers; between genders, age groups, and host genetics; and between infected populations with differing disease outcomes. Such data should help flesh out the natural history of this infection and its variations by host and environment. To identify potential interventions in high risk carriers collaboration among the established cohorts is essential, given the rarity of prospective samples and data and the relatively low risk of disease among carriers. The most difficult route of transmission to interrupt is that of perinatal transmission in children who are not breastfed. This is a priority area that should not be neglected.

HELICOBACTER PYLORI Natural History of H. pylori Infection Helicobacter pylori is a Gram-negative, spiral-shaped organism that lives within the mucosa of the human stomach. The organism had been seen in histologic sections of the stomach since the early 20th century (Kreinitz, 1906), but not until the early 1980s was it recognized to cause disease—peptic ulcer disease, specifically. At that time, gastroenterologist Barry Marshall and colleagues in Perth, Australia, cultured the organism and also conducted epidemiologic studies connecting the organism to duodenal ulcer disease (Warren and Marshall, 1983; Marshall and Warren, 1984; Marshall et al., 1988). Subsequently, in two separate experiments, Marshall and Arthur Morris ingested H. pylori proving it caused the inflammatory process—gastritis with acute and chronic inflammation—that antecedes both ulcer disease and gastric adenocarcinoma (Walker et al., 1971; Marshall et al., 1985; Morris and Nicholson, 1987). From that time until today, a series of epidemiological studies has ensued to investigate the role of H. pylori in malignancy.

Epidemiology of Infection

Figure 26–9. Incidence rate of ATL among HTLV-1 carriers in Japan and the Caribbean by gender (Kondo et al., 1989; Murphy et al., 1989).

Helicobacter pylori is an extremely common organism. It is estimated that 50% of the world is infected and, once infected, a person typically remains so for their lifetime (Brown, 2000). In the United States, Europe, and Oceania, the prevalence of H. pylori is probably closer to 25%, with infection quite unusual in children but continuing to plague the majority of adults over the age of 60 years. Over the last half of the 20th century in the industrialized West, H. pylori infection prevalence has decreased by a remarkable 25% per decade (Parsonnet et al., 1992; Banatvala et al., 1993; Roosendaal et al., 1997). This transformation parallels societal improvements 50 years earlier in socioeconomic status and in public health and hygiene. Infection is strongly linked with socioeconomic status in childhood (Malaty and Graham, 1994; Brown, 2000). In the United States, independent of

532

PART III: THE CAUSES OF CANCER

socioeconomic status, H. pylori is two- to threefold more common in blacks and in Hispanics than in whites (Everhart et al., 2000). One twin study suggests that there is a small element of genetic predisposition to the acquisition of infection (Malaty et al., 1994). Some studies also find H. pylori infection to be more common in men than in women, although this finding has not been consistently reported (Replogle et al., 1995; Everhart et al., 2000). In developing countries, H. pylori infection is almost universally found in adults, and 50% of children may be infected by the age of 5 years (Brown, 2000). The difference in H. pylori prevalence in developing and developed countries is related to the mode of transmission. H. pylori is predominantly transmitted from person-to-person by the fecal–oral route. As with other organisms transmitted in this manner (e.g., hepatitis A and Shigella), improved sanitation and hygiene and decreased household crowding are accompanied by decreases in infection incidence. Because H. pylori infection, once established, does not spontaneously remit; the recent improvements in sanitation and hygiene lower childhood prevalence but have little effect on the prevalence in adults, who had acquired infection decades earlier. All H. pylori–infected hosts develop chronic gastric inflammation or gastritis. This inflammatory process is associated with a specific humoral response as well as a locally vigorous Th1 cell-mediated immune response, with elevated IL-12, IFN-g, and TNF-a secretion (Crabtree, 1996; D’Elios et al., 1997). H. pylori also induces IL-8 secretion by epithelial cells, resulting in recruitment of inflammatory cells to the mucosa (Huang, 1995). In the chronic phase of infection, IL-10 has been reported to suppress the Th1 response (Holck et al., 2003). Some speculate that this IL-10 counter-response may foster bacterial persistence (Chen et al., 2001).

Helicobacter pylori Biomarkers Large epidemiologic studies typically rely on serology (usually anti–H. pylori IgG) for diagnosis of infection. In older children and young and middle-aged adults who have not received treatment, serology has high sensitivity and specificity for ongoing infection (Vaira et al., 2002). In children less than 5 years, in the elderly, and those previously treated, however, H. pylori serology is less accurate. Moreover, not all tests work equally well in all populations (Bodhidatta et al., 1993). To be reliable, the tests should have been developed for a similar population to that in which the study is being conducted. Serology is the only useful test for determining between types of H. pylori infection in large population-based studies. Specifically, H. pylori has two major subtypes: one that contains a pathogenicity island of genes called the Cag pathogenicity island (PAI) and one that does not. One of the Cag PAI proteins, CagA, is highly immunogenic, and antibodies are readily detected serologically. In CagA H. pylori–infected patients, antibodies to CagA are more sensitive for infection than the more general H. pylori serologic tests (Vaira et al., 2002). Other genetic variants of H. pylori—including genotypes of the vacuolating cytotoxin (VacA), of a surface protein (BabA), and a protein expressed with bacterial adhesion to the mucosa (IceA) (van Doorn et al., 1998)—have been related to disease outcome. However, characterization of these variants requires endoscopy with bacterial culture, limiting their use in epidemiologic studies. Characterization by culture of the stomach is also relatively insensitive and potentially biasing, as not all strain types are equally cultivable ex vivo. Other diagnostic tests for H. pylori include urea breath tests and stool antigen tests. These tests have sensitivity and specificity comparable to the best serologic assays (Vaira et al., 2002). The breath tests exploit H. pylori’s production of the enzyme urease—which splits urea into carbon dioxide and water—for diagnostic purposes. After giving an oral solution of carbon-isotope–labeled urea, investigators measure the amount of the isotope in exhaled carbon dioxide. Although C14 isotope tests are inexpensive, they are not commercially available in the United States. The less radioactive C13 tests are available, but their expense may limit their use in large epidemiological studies. Stool antigen testing for H. pylori has not yet been used for many large studies although it is practically feasible.

Associated Gastric Carcinoma Observational studies linking H. pylori with gastric adenocarcinoma include (1) ecologic studies correlating H. pylori prevalence and gastric cancer incidence and mortality in various geographic regions (Forman et al., 1990; Eurogast Study Group, 1993), (2) temporal studies demonstrating similar time trends in H. pylori prevalence and cancer incidence or mortality (Sonnenberg, 1993; Rupnow et al., 2000), (3) myriad case-control studies comparing H. pylori infection prevalence in gastric cancer patients and various controls (Huang et al., 1998; Eslick et al., 1999; Xue et al., 2001), (4) nested case-control studies (Helicobacter and Cancer Collaborative Group, 2001), and (5) one large cohort study demonstrating the temporal association between H. pylori infection and subsequent adenocarcinoma (Uemura et al., 2001). The resulting data paint a convincing picture that H. pylori is the preeminent risk factor for adenocarcinoma of the gastric antrum and the distal stomach. (Cancers of the proximal stomach, i.e., the gastroesophageal junction and cardia, are not thought to be H. pylori–related tumors.) In meta-analyses of the heterogeneous studies, case-control studies averaged a relatively low risk of cancer with infection (OR ª 1.8); nested case-control studies found a higher risk (OR ª 3.0); and nested case-control studies with long follow-up (>10 years) yielded an even higher estimate (OR ª 5.9) (Xue et al., 2001; Helicobacter and Cancer Collaborative Group, 2001). In 1500 Japanese subjects followed for 8 years, only those with infection developed cancer, leading to an RR of infinity (Uemura et al., 2001). Although the nested case-control studies indicate that greater than 50% of gastric tumors are caused by H. pylori, gastric adenocarcinomas are not monomorphic. Different types of adenocarcinoma in the stomach have different risk factors. The types of gastric adenocarcinoma include the intestinal type (occurs more in men than in women, more in the elderly, and typically localize to the antrum), the diffuse type (seen equally in men and women, occur at younger age, and seen in antrum and body), and a much smaller number of lymphoepitheliomalike tumors (LELCs) (Lauren, 1965; Wang et al., 1999). H. pylori is a strong risk factor for both intestinal and diffuse tumors but not for LELCs (Huang et al., 1998; Wu MS et al., 2000). Gastric tumors also differ by their location in the stomach. Tumors in the antrum predominate in developing countries, but this is not the case in industrialized countries. Currently, 50% of gastric adenocarcinomas in the United States are in the cardia or gastroesophageal junction (Devesa et al., 1998). H. pylori has been strongly linked to tumors in the antrum and body but less consistently linked to tumors of the cardia (Huang et al., 1998). Some data even support H. pylori being protective against cancers of the gastroesophageal junction (Chow et al., 1998). Given the variability of tumors and the imperfect nature of diagnostic tests, the true proportion of gastric tumors caused by H. pylori infection is difficult to state with certainty. However, some have speculated that H. pylori may be necessary for intestinal and diffuse tumors of the distal stomach to occur—the tumors most strongly linked to infection (Peterson, 2002). Pivotal randomized clinical trials of H. pylori eradication in the prevention of gastric cancer are ongoing, although the logistics are daunting (Xia et al., 2003; Saito, 2003). Early analyses do not show that eradication of H. pylori prevents cancer per se, but rather evaluate whether H. pylori eradication prevents evolution or induces regression of the preneoplastic conditions that precede cancer. These data are inconclusive but suggest that H. pylori eradication may limit progression of preneoplasia to malignancy in a subset of subjects (Correa et al., 2000; Sung et al., 2000; Ley et al., 2004). Helicobacter pylori has been even more strongly linked to primary gastric lymphoma than to adenocarcinoma (Boot and De Jong, 2002). Primary gastric lymphomas derive from marginal zone parafollicular B cells within mucosal-associated lymphoid tissue (MALT) (Ahmad et al., 2003). This tissue, which consists of submucosal lymphoid follicles that mimic the Peyer patches in the ileum, does not exist in the stomach of uninfected individuals but occurs in all with H. pylori infection. In a nested case-control study, infection was found in greater than 85% of primary gastric lymphoma, OR = 6.3 (Parsonnet et al., 1994). The true strength of the association between H. pylori and lym-

Infectious Agents phoma, however, stems from clinical trials of H. pylori eradication. A series of investigators, beginning with Wotherspoon and colleagues (1993), found that eliminating H. pylori from the stomach results in complete regression in 70% to 80% of low-grade, superficial tumors (Boot and De Jong, 2002). A small proportion will recur within 10 years of follow-up and become resistant to H. pylori eradication therapy. Higher grade tumors in the absence of MALT can also occasionally regress in response to H. pylori therapy (Alsolaiman et al., 2003).

Mechanisms of Carcinogenesis Chronic inflammation and augmented cell proliferation are believed to be critical factors in the development of gastric adenocarcioma. The inflammatory process enhances production of free radicals by epithelial and inflammatory cells, leading to oxidative DNA damage (Baik et al., 1996; Farinati et al., 1998; Bagchi et al., 2002). Decreased antioxidant concentration (ascorbic acid, a-tocopherol, and bcarotene) in the gastric lumen and epithelium consequent to infection—which occurs irrespective of dietary antioxidants—enhances the probability of damage (Sobala et al., 1991; Zhang et al., 2000). Both through its DNA damaging effects as well as more directly through the action of its cytotoxin, H. pylori infection augments gastric epithelial apoptosis (Cover et al., 2003; Nagasako et al., 2003). The mucosal epithelium manifests enhanced epithelial proliferation that may overcompensate for this apoptotic cell loss (Peek et al., 1997; Rokkas et al., 1999). In many, but not all infected hosts, this process leads to the gradual destruction of the mucosal glands (atrophic gastritis) and subsequent replacement of the normal tissue with metaplastic cells (intestinal metaplasia). Over time, glandular destruction results in decreased gastric acid secretory capacity, favoring extension of H. pylori’s domain throughout the stomach. Extensive, multifocal atrophic gastritis and intestinal metaplasia, although unnecessary for carcinogenesis, are considered preneoplastic conditions and markers for increased cancer risk. Risk for cancer appears to be most acute with the strains that induce the most inflammation; specifically, those that contain the Cag PAI (Blaser et al., 1995; Parsonnet et al., 1997). Cag PAI strains typically coincide with a more virulent form of the vacuolating cytotoxin and are consequently more damaging to epithelial cells (Atherton et al., 1995). Additionally, the Cag PAI encodes a type IV secretion system that injects the CagA protein into epithelial cells (Stein et al., 2000). Within the epithelial cell, the CagA protein becomes phosphorylated and induces structural changes of the mucosal cell that give it a motile cellular phenotype. Individuals infected with strains containing Cag PAI have increased production of IL-1, TNF-a, IL-2, IL-6, and IFNg when compared with infections of other strains (Helicobacter and Cancer Collaborative Group, 2001). Cag PAI strains also induce greater cellular IL-8 production and enhance inflammatory cell recruitment (Crabtree et al., 1995). These changes, in turn, result in greater free-radical production and augmented DNA damage. Cag PAI–containing strains increase the risk for intestinal type tumors—which often contain the classic CpG mutations seen with inflammationrelated malignancies—threefold compared to non–Cag PAI strains; however, Cag PAI does not appear to engender greater risk for diffuse tumors than other strains (Parsonnet et al., 1997; Shibata et al., 2002). Thus, the mechanisms of H. pylori–related disease may vary for different tumors and different infections. The pathogenesis of MALT, which comprises up to 50% of gastric lymphomas (Isaacson et al., 2001), is poorly understood. It appears H. pylori infection stimulates T cells to drive B-cell proliferation in MALT (Hussell et al., 1993). The early stage of MALT lymphoma with low-grade histology and involvement of only superficial areas of the mucosa is completely reversible with H. pylori eradication therapy. On careful examination of lymphocytes remaining in the tissue, however, clonal cells from the lymphoma may still be found despite regression, and recurrent infection can result in recurrent tumor. Eventually, if untreated, a translocation t(11;18)(q21;q21) causes the B-cell proliferation to no longer respond to H. pylori eradication (Boot and De Jong, 2002).

533

Cofactors Only a small proportion of H. pylori–infected people develop gastric adenocarcinoma. Factors that contribute to the outcome in this small subset include bacterial factors (described above), host factors, environmental cofactors, and age of acquisition of infection. To date, the strongest host factors for cancer related to infection appears to be the IL-1b genotype and the genotype of the IL-1b endogenous receptor antagonist (El Omar et al., 2000). The disadvantageous IL-1b genotype both enhances the inflammatory response and decreases acid secretion within the stomach. Similar links to polymorphisms of TNF-a and IL-10 have also been described (El Omar et al., 2003). Combinations of these adverse cytokines markedly enhance risk, whereas none of these cytokine variations is related to cancer in uninfected hosts. Environmental cofactors are poorly understood. Among the most intriguing of the putative cofactors are coinfections with other infectious agents. Some maintain that decreased gastric acidity enhances H. pylori–related carcinogenesis by permitting bacterial growth of N-nitrosating organisms within the stomach (Correa et al., 1975; Sanduleanu et al., 2001). Based on a mouse model, others have posited that intestinal helminths mitigate the probability of cancer in the setting of H. pylori infection by diminishing the mucosal inflammatory response (Fox et al., 2000). Noninfectious factors that may alter cancer risk in the setting of H. pylori include dietary antioxidants, aspirin, and smoking (Ekstrom et al., 2000; Siman et al., 2001; Akre et al., 2001). Each of these hypotheses requires further substantiation. As with hepatitis B infection, age at acquisition of infection may contribute to infection outcome. Because the age at which infection is acquired is typically unknown, this is a difficult hypothesis to test. Some circumstantial data, however, support this possibility. First, H. pylori–infected first-born children, who typically have less exposure to oral–fecal pathogens than children later in the birth order, are less likely to get gastric cancer than others (Blaser et al., 1994). This suggests that later acquisition of infection may be protective. Second, fecal–oral pathogens that are on the wane (hepatitis A and Shigella) exhibit a characteristic change in age of incidence. Specifically, as the incidence of infection decreases, the average age at infection acquisition increases. A pattern of decreasing overall incidence of infection, increasing age of acquisition, and decreased average duration of infection does much to explain patterns of cancer and ulcer disease observed worldwide. Cofactors for lymphomagenesis are unknown. The PAI status of H. pylori infection does not appear to influence the occurrence of MALT lymphoma.

Prevention Helicobacter pylori is spontaneously disappearing from Western populations (Parsonnet et al., 1992; Banatvala et al., 1993; Roosendaal et al., 1997). It is increasingly rare to find infection in children born in the United States, Western Europe, Japan, or Australia. Primary prevention of infection may therefore be moot in the industrialized West. Improvements in household sanitation and hygiene, decreased household crowding, provision of clean water, and improved nutrition coincide with the organism’s demise. Exactly which of these factors is most responsible for H. pylori’s disappearance is unknown. Based on its primary mode of transmission—person-to-person spread—it is most likely that the combination of household sanitation and hygiene and decreased crowding have mitigated the incidence of infection. With that, analyses indicate that an H. pylori vaccine could be a costeffective strategy for cancer prevention (Rupnow et al., 1999; 2001). The years of use of such a vaccine would be limited in developed countries, as the organism would rapidly spiral out of existence. In developing countries and countries undergoing rapid economic growth where the incidence of H. pylori remains high and gastric cancer is common, a vaccine could be useful for the foreseeable future. The challenges of creating such a vaccine, however, are daunting. H. pylori is a chronic infection that can evade the human immune response for

534

PART III: THE CAUSES OF CANCER

decades. Some even posit that the natural host response benefits the organism. Yet, experimental and accidental ingestions of the organism indicate that some people do not get infected despite exposure. Much more work needs to be done to understand the correlates of this protective immunity. Secondary prevention is widely available. H. pylori is curable with combinations of antibiotic therapy. Although increased resistance to treatment is being observed, the diligent clinician can eradicate virtually any infection. Controversy remains over who should receive such treatment. A majority of the world’s population is infected with H. pylori, yet only a small percent will develop malignancy. Although this small ratio of benefited-to-treated challenges the probability of success, several analyses indicate that eradication of H. pylori could be cost-effective in preventing cancer (Parsonnet et al., 1996; Fendrick et al., 1999; Mason et al., 2002). To be cost-effective, however, such treatment would need to be given in midlife (Parsonnet et al., 1996). Unfortunately, it is unknown whether treatment of infection so late in its course diminishes cancer incidence. Studies on H. pylori preneoplastic conditions have not promised a large benefit in adults. It is hoped that randomized trials addressing cancer prevention more directly will definitively provide answers to this question. Until such studies are complete, different consensus conferences have advocated different strategies. The NIH Consensus, the oldest of the available recommendations, has not advocated treatment of asymptomatic patients to prevent cancer (NIH Consensus Development Panel on Helicobacter pylori in Peptic Ulcer Disease, 1994). In contrast, the more recent European Consensus has recommended treatment for people with a family history of gastric cancer, people who have had partial gastrectomy, those with advanced preneoplasia on gastric biopsy, and for virtually any one who desires treatment (Malfertheiner et al., 2002). Tertiary prevention (i.e., treatment of H. pylori after diagnosis of adenocarcinoma) is widely undertaken in patients who have early, potentially curable malignancies. A small, nonrandomized trial in Japan found that patients with mucosally resected early gastric cancer did not have disease recurrence after H. pylori eradication therapy; without H. pylori eradication, recurrence occurred at the rate of 10% per year (Uemura et al., 1997). Treatment of H. pylori is also widely recommended for patients with MALT lymphoma, where in early stages, H. pylori eradication may be curative.

Future Research The most critical area in the H. pylori–cancer field is to see whether, and when, H. pylori eradication prevents cancer. Can treatment reverse disease progression in midlife? At what age, if any, are these treatment benefits most evident? Are there specific strains that should be treated and others ignored? If so, how can this best be done? A second area of intense debate is whether deficits of H. pylori eradication outweigh benefits. Several studies indicate that H. pylori eradication therapy increases the risk of esophageal reflux disease (Falk, 2001) and possibly that of esophageal cancer (Henrik et al., 2001). The latter is among the most rapidly increasing tumors in the world. In favor of H. pylori eradication, the organism also causes ulcer disease and, like other chronic inflammatory processes, is suspected of contributing to atherogenesis. What is the balance of all of these conditions? Randomized trials of therapy that simply address H. pylori’s role in gastric cancer may not be enough to settle the debate regarding secondary prevention. We need to evaluate a variety of outcomes, including total mortality. Moreover, because strains of varying pathogenicity exist and have differing distributions worldwide, decisions to screen and treat may need to be made not broadly but regionally, based on local data regarding H. pylori genotypes, cofactors for disease, and the patterns of cancer risk. H. pylori infection is one of the strongest risk factors for cancer identified. Identifying its importance has been a far simpler endeavor than deciding how to address the problem. This debate will undoubtedly continue for the next decade. Yet it is to be hoped that we will ultimately have solid solutions for treatment and prevention of this curable carcinogen.

PARASITES Schistosomes Schistosomes form a genus of parasitic blood flukes, three of which have been evaluated with respect to their carcinogenicity to humans (IARC, 1994b): Schistosoma mansoni, Schistosoma haematobium, and Schistosoma japonicum. Schistosoma mansoni and S. haematobium account for more than 95% of human infections and are endemic in Africa and the eastern Mediterranean; S. mansoni is also found in Central and South America. Schistosoma japonicum occurs primarily in Southeast Asia, including China, the Philippines, and Indonesia. Infection results from exposure to contaminated freshwater during which the schistosomal larvae penetrate the skin. These larvae mature and travel through the bloodstream to the veins that drain the urinary bladder (S. haematobium) or the gastrointestinal tract (S. japonicum and S. mansoni), where they produce a large number of eggs. The eggs are then excreted in the host’s urine or feces, respectively. Many eggs also remain within the tissue of the bladder and ureters (S. haematobium) or the intestines and liver (S. japonicum and S. mansoni). The larvae hatched from eggs that are shed into freshwater then infect an intermediate snail host within which they multiply by an asexual process and are released back into the water. Detection of infection is based on the observation of eggs in urine (S. haematobium) or feces (S. japonicum and S. mansoni) (IARC, 1994b). Assays to detect parasite-specific antibodies or antigens appear to be of variable validity and have not generally been used in epidemiologic studies. The prevalence and intensity (eggs per milliliter of urine or per gram of feces) of infection increases from childhood through about age 20 and then declines. Manifestation of clinically apparent schistosomiasis results from the host immune response to the parasite eggs and the subsequent development of granulomas in the affected organs. The extent of associated morbidity appears to be related to the intensity and duration of the schistosomal infection, although many infected persons do not experience any significant symptoms. The adult worm can live in a human host for up to 30 years. A fairly large number of case series studies conducted in Africa have linked the presence of S. haematobium infection with bladder cancer (IARC, 1994b). Such schistosomal-related tumors are generally squamous cell in origin and occur at a relatively young age, in contrast to other bladder cancers. Correlation studies, again performed primarily in Africa, also support an association between S. haematobium endemicity and bladder cancer occurrence (Thomas et al., 1990; IARC, 1994b). Statistically significant RRs ranging from 2 to 15 have been found in about seven hospital-based case-control studies (IARC, 1994b; Vizcaino et al., 1994; Bedwani et al., 1998). Based on these data, S. haematobium has been classified by IARC as being carcinogenic to humans. With respect to the malignant potential of infection with S. japonicum, rates of colorectal and of liver cancer have shown significant geographic correlations with the impact of this parasite within Japan and China (IARC, 1994b). In the IARC assessment of the association of S. japonicum with cancer development, only three case-control studies for each of liver cancer and colorectal cancer provided relevant data. Although significant RRs were reported by these studies, of the order 2 to 10, the limited available evidence led IARC to conclude that infection with S. japonicum is “possibly carcinogenic” (Class 2B). Inadequate information existed to evaluate S. mansoni (IARC, 1994b). The hypothesized mechanism by which chronic S. haematobium infection induces bladder cancer is via immune-mediated inflammation of the tissues in which the parasite eggs are embedded (IARC, 1994b). The resultant cell turnover and regeneration likely leads to the promotion of existing mutations, with concomitant proliferation of and dysplastic changes in the squamous epithelium of the bladder and lower urinary tract. Moreover, during the inflammatory process, mutagenesis may occur from the production of carcinogenic metabolites as well as nitrosamines. An estimated 9500 cases of bladder cancer occurring in 1990 were attributable to infection with S. haematobium (Parkin et al., 1999). All of these cases were located in Africa or West Asia and represented

Infectious Agents one-third of the bladder cancers in those areas. Schistosomal infection can be effectively treated. Although costly, population-based chemotherapeutic interventions, combined with public health efforts to improve sanitation, safeguard water supplies, and increase health education at the community level, have led to substantial decreases in, and even eradication of, the burden of schistosomiasis in Asia, Latin America, major parts of the Middle East and the Caribbean, and other endemic areas (IARC, 1994b; WHO, 2002). However, the prevalence of schistosomes and the morbidity associated with these infections remain elevated in sub-Saharan Africa.

Liver Flukes Infection with three species of liver flukes has been hypothesized to be associated with cholangiocarcinoma of the liver: Opisthorchis viverrini, Opisthorchis felineus, and Clonorchis sinensis. These parasites can be found at highly endemic levels in Thailand and Laos (O. viverrini), China, Taiwan, and the Republic of Korea (C. sinensis), and the Russian Federation (O. felineus) (IARC, 1994b). Transmission to humans occurs via the consumption of raw fish contaminated with the infectious miracid, a stage of the liver fluke life cycle. Upon hatching the flukes migrate to the bile ducts where they mature and lay eggs that are shed in the feces. Subsequent fecal contamination of freshwater leads to the ingestion of the eggs by susceptible snails, who serve as the intermediate host for the reproductive stage. Larvae are then released and infect fish, which are the second intermediate host in the liver fluke’s life cycle. Infection with liver flukes is most usually determined by the detection of eggs in fecal specimens (IARC, 1994b). EIAs also have been used to measure helminth-specific antibodies. In endemic areas, the prevalence of infection is relatively low in young children, generally increases during the early to late teen years, and then plateaus (IARC, 1994b). Liver flukes establish a chronic infection, unless treatment is received, and can live up to 25 years in the infected human host. The epidemiologic evidence linking liver fluke infection with the occurrence of cholangiocarcinoma has originated from correlation studies and a small number of case-control studies. Much of the research has been conducted in Thailand and, thus, pertains to infection with O. viverrini (Parkin et al., 1993; IARC, 1994b). In that country, a strong correlation has been observed between cholangiocarcinoma rates and the prevalence of O. viverrini (Srivatnakul et al., 1991). A hospital-based case-control study reported a fivefold association between incident cholangiocarcinoma and elevated antibodies to this parasite (Parkin et al., 1991). In addition, in a cross-sectional study conducted in northeastern Thailand, an increasing odds of prevalent cholangiocarcinoma, as detected by ultrasonography, was found for increasing intensity of O. viverrini infection (Haswell-Elkins, 1994a). The data related to a carcinogenic effect of the other two liver fluke infections are more limited, particularly for O. fileneus (IARC, 1994b). Case series studies of cholangiocarcinoma and bile duct cancers occurring in Hong Kong and Korea demonstrated a likely relationship with C. sinensis infection. One hospital-based case-control study in Hong Kong and two in Korea observed RRs of 3 and 6, respectively, for the association of infection with C. sinensis and cholangiocarcinoma. IARC has categorized infection with O. viverrini as being a Class 1 human carcinogen and infection with C. sinensis as being “probably carcinogenic” (Class 2A) (IARC, 1994b). Chronic infection with liver flukes is believed to induce carcinogenesis through the mechanism of chronic inflammation of the intrahepatic bile ducts, resulting from direct irritation of epithelial cells by the liver flukes as well as from the host immune response to the parasites (Parkin et al., 1993; IARC, 1994b). The carcinogenic process is likely similar to that for schistosomal infection, with inflammationassociated changes playing an important role. Additional mutations also may be produced by the generation of reactive oxygen species (Parkin et al., 1993). As well, the induction of nitric oxides and nitrosamines may be particularly relevant with respect to the development of parasite-related cholangiocarcinoma (Haswell-Elkins, 1994b). Despite the high prevalence of these infections in endemic areas of Southeast Asia and Russia, cholangiocarcinoma is relatively rare in

535

affected populations. The number of cases attributable to O. viverrini infection in Thailand in 1990 was estimated to be about 300, although perhaps as high as 3000 (Parkin et al., 1999). Effective treatment is available for liver fluke infection, which would prevent further damage to the biliary tract and progression to cancer. Improvements in sanitation and education with respect to the consumption of raw fish also are important for reducing the burden of infection in endemic areas.

FUTURE DIRECTIONS Despite their diversity, these persistent infections share some common features in their demonstrated oncogenicity. These include the central importance of the dynamic host–agent interaction. Age and other host conditions of initial infection as well as gender are often important in influencing the host–agent interaction. The direct modification of cell-cycle control is a characteristic of most of these infections. Malignancy occurs as an occasional accident. The empirical use of agent-related biomarkers has illuminated the trail for untangling the epidemiology of these systems. A key area for future research in all of the oncogenic infections is the characterization of immune status in biospecimens gathered during the formative period of malignancy. An early model of such research is that of Schroeder et al. (1999), who demonstrated the predictive value of serum sCD23 for subsequent non-Hodgkin lymphoma in AIDS patients. The identification of the determinants of cytokine profile, including such factors as gender, social factors, and genetic polymorphisms, should fill in the evolving “pictures” of the natural history of these infection-associated malignancies. In terms of secondary prevention, the tools are in hand to help identify those at high risk in most of these infection-associated cancers. In fact, we have an example of infection-associated malignancy that can be cured by treatment of the underlying infection in MALT and an example of reversal of a premalignant disease in post-transplant lymphoproliferation with the infusion of EBV-specific CTLs. It is likely that interventions on the other oncogenic infections to prevent or reverse oncogenesis will be seen. New associations between infections and cancer continue to be proposed and can generate contentious scientific interest and great public concern. The prime current example is the question of whether SV40 is a human carcinogen (Klein et al., 2002). We propose that the validity of such assertions be addressed quickly and rigorously and be guided by the body of work summarized here. However, we should be open to new twists in this engaging set of puzzles. References Ablashi DV, Chatlynne LG, Whitman JE JR, Cesarman E. 2002. Spectrum of Kaposi’s sarcoma-associated herpesvirus, or human herpesvirus 8, diseases. Clin Microbiol Rev 15:439–464. Agathanggelou A, Niedobitek G, Chen R, Nicholls J, Yin W, Young LS. 1995. Expression of immune regulatory molecules in Epstein-Barr virusassociated nasopharyngeal carcinomas with prominent lymphoid stroma. Evidence for a functional interaction between epithelial tumor cells and infiltrating lymphoid cells. Am J Pathol 147:1152–1160. Agnello V, Chung RT, Kaplan LM. 1992. A role of hepatitis virus infection in type II cryoglobulinemia. New Engl J Med 327:1490–1495. Ahmad A, Govil Y, Frank BB. 2003. Gastric mucosa-associated lymphoid tissue lymphoma. Am J Gastroenterol 98:975–986. Ahmed R, Morrison LA, Knipe DM. 1996. Persistence of viruses. In: Fields BN, Knipe DM, Howley PM, et al., eds. Fields Virology. Third edition. Philadelphia: Lippincott–Raven, pp. 219–249. Aikawa T, Kojima M, Onishi H, et al. 1996. HLA-DRB1 and DQB1 alleles and haplotypes influencing the progression of hepatitis C. J Med Virol 49:274– 278. Akahane Y, Kojima M, Sugai Y, et al. 1994. Hepatitis C virus infection in spouses of patients with type C chronic liver disease. Ann Int Med 120:748–752. Akre K, Ekstrom AM, Signorello LB, Hansson LE, Nyren O. 2001. Aspirin and risk for gastric cancer: A population-based case-control study in Sweden. Br J Cancer 84:965–968. Akula SM, Naranatt PP, Walia N-S, Wang F-Z, Fegley B, Chandran B. 2003. Kaposi’s sarcoma-associated herpesvirus (human herpesvirus 8) infection of human fibroblast cells occurs through endocytosis. J Virol 77:7978–7990. Alani RM, Munger K. 1998. Human papillomaviruses. Sci Med 5:28–35.

536

PART III: THE CAUSES OF CANCER

Alberti A, Morsica G, Chemello L, et al. 1992. Hepatitis C viraemia and liver disease in symptom-free individuals with anti-HCV. Lancet 340:697–698. Allison ME, Wreghitt T, Palmer CR, Alexander GJ. 1994. Evidence for a link between hepatitis C virus infection and diabetes mellitus in a cirrhotic population. J Hepatol 21:1135–1139. Alsolaiman MM, Bakis G, Nazeer T, Macdermott RP, Balint JA. 2003. Five years of complete remission of gastric diffuse large B cell lymphoma after eradication of Helicobacter pylori infection. Gut 52:507–509. Alter HJ, Conry-Cantilena C, Melpolder J, et al. 1997. Hepatitis C in asymptomatic blood donors. Hepatology 26(Suppl 1):29S–33S. Alter HJ, Seeff LB. 2000. Recovery, persistence, and sequelae in hepatitis C virus infection: A perspective on long-term outcome. Sem Liver Dis 20:17–35. Alter MJ. 1997. Epidemiology of hepatitis C. Epidemiology 26(Suppl 1): 62S–65S. Alter MJ, Coleman PJ, Alexander WJ, et al. 1989. Importance of heterosexual activity in the transmission of hepatitis B and non-A, non-B hepatitis. JAMA 262:1201–1205. Alter MJ, Hutin YJF, Armstrong GL. 2000. Epidemiology of hepatitis C. In: Liang TJ, Hoofnagle JH, eds. Biomedical Research Reports, Hepatitis C. San Diego: Academic Press, pp. 169–183. Alter MJ, Kruszon-Moran D, Nainan OV, et al. 1999. The prevalence of hepatitis C virus infection in the United States, 1988 through 1994. N Engl J Med 341:556–562. Alter MJ, Margolis HS, Krawczynski K, et al. 1992. The natural history of community-acquired hepatitis C in the United States. N Engl J Med 327:1899–1905. Alward WLM, Mcmahon BJ, Hall DB, et al. 1985. The long-term serologic course of asymptomatic hepatitis B virus carriers and the development of primary hepatocellular carcinoma. J Infect Dis 15:604–609. Ambinder RF, Weiss LM. 1999. Association of Epstein-Barr virus with Hodgkin’s disease. In: Mauch PM, Armitage JO, Diehl V, et al., eds. Hodgkin’s Disease. Philadelphia: Lippincott Williams and Wilkins, pp. 79–98. Anagnostopoulos I, Herbst H, Niedobitek G, Stein H. 1989. Demonstration of monoclonal EBV genomes in Hodgkin’s disease and Ki-1 positive anaplastic large cell lymphoma by combined Southern blot and in situ hybridization. Blood 74:810–816. Anagnostopoulos I, Hummel M. 1996. Epstein-Barr virus in tumours. Histopathology 29:297–315. Anand BS, Hollinger FB. 1997. Prevention of hepatocellular carcinoma with vaccines. In: Okuda K, Tabor E, eds. Liver Cancer. New York: Churchill Livingstone, pp. 415–424. Ando Y, Matsumoto Y, Nakano S, et al. 2003a. Long-term follow-up of HTLV-I infection in bottle-fed children born to seropositive mothers. J Infect 46:9–11. Ando Y, Matsumoto Y, Nakano S, et al. 2003b. Long-term study of vertical HTLVI infection in children breast-fed by seropositive mothers. J Infect 46:177–179. Andreoni M, Sarmati L, Nicastri E, et al. 2002. Primary human herpesvirus 8 infection in immunocompetent children. JAMA 287:1295–1300. Antman K, Chang Y. 2000. Kaposi’s sarcoma. N Engl J Med 342:1027– 1038. Aoyagi K, Ohue C, Iida K, et al. 1999. Development of a simple and highly sensitive enzyme immunoassay for hepatitis C virus core antigen. J Clin Microbiol 37:1802–1808. Arduino JM, Stuver SO, Spiegelman D, et al. 2001. Assessment of markers of hepatitis C virus infection in a Japanese adult population. J Infect Dis 184:1229–1235. Arisawa K, Katamine S, Kamihira S, et al. 2002. A nested case-control study of risk factors for adult T-cell leukemia/lymphoma among human T-cell lymphotropic virus type-I carriers in Japan. Cancer Causes and Control 12:1–7. Arisawa K, Soda M, Akahoshi M, et al. 1998. Human T-lymphotropic type-I infection, antibody titers and cause-specific mortality among atomic-bomb survivors. Jpn J Cancer Res 89:797–805. Arisawa K, Soda M, Endo S, et al. 2000. Evaluation of adult T-cell leukemia/lymphoma incidence and its impact on non-Hodgkin’s lymphoma incidence in southwestern Japan. Int J Cancer 85:319–324. Ariyoshi K, Schim Van DER Loeff M, Cook P, et al. 1998. Kaposi’s sarcoma in the Gambia, West Africa is less frequent in human immunodeficiency virus type 2 than in human immunodeficiency virus type 1 despite a high prevalence of human herpesvirus 8. J Hum Virol 1:193–199. Armstrong GL, Alter MJ, Mcquillan GM, Margolis HS. 2000. The past incidence of hepatitis C virus infection: Implications for the future burden of chronic liver disease in the United States. Hepatology 31:777–782. Atherton JC, Cao P, Peek RM Jr, Tummuru Mk, Blaser MJ, Cover TL. 1995. Mosaicism in vacuolating cytotoxin alleles of Helicobacter pylori. Association of specific vacA types with cytotoxin production and peptic ulceration. J Biol Chem 270:17771–17777. Bach J-F. 2002. The effect of infections on susceptibility to autoimmune and allergic diseases. N Engl J Med 347:911–920. Bagchi D, Mcginn TR, Ye X, et al. 2002. Helicobacter pylori-induced oxidative stress and DNA damage in a primary culture of human gastric mucosal cells. Digest Dis Sci 47:1405–1412. Baik SC, Youn HS, Chung MH, et al. 1996. Increased oxidative DNA damage in Helicobacter pylori-infected human gastric mucosa. Cancer Res 56:1279–1282.

Baldwin SB, Wallace Dr, Papenfuss MR, et al. 2003. Human papillomavirus infection in men attending a sexually transmitted disease clinic. J Infect Dis 187:1064–1070. Balsano C, Avantaggiati ML, Natoli G, et al. 1991. Transactivation of c-fos and c-myc protooncogenes by both full–length and truncated versions of the HBV–X protein. In: Hollinger FB, Lemon SM, Margolis H, eds. Viral Hepatitis and Liver Disease. Baltimore: Williams and Wilkins, pp. 572–576. Banatvala N, Mayo K, Megraud F, Jennings R, Deeks JJ, Feldman RA. 1993. The cohort effect and Helicobacter pylori. J Infect Dis 168:219–221. Bangham CR. 2000. HTLV-I infections. J Clin Pathol 53:581–586, Barrasso R. 1992. HPV-related genital lesions in men. In: Munoz N, Bosch FX, Shah KV, Meheus A, eds. The Epidemiology of Cervical Cancer and Human Papillomavirus. IARC Sci Publ No. 119. Lyon: International Agency for Research on Cancer, pp. 85–92. Barrett S, Ryan E, Crowe J. 1999. Association of the HLA-DRB1*01 allele with spontaneous clearance in an Irish cohort infected with hepatitis C virus via contaminated anti-D immunoglobulin. J Hepatol 30:979–983. Barrett TJ, Silbar JD, Mcginley JP. 2003. Genital warts—a venereal disease. JAMA 154:333–334. Barrios C, Brawand P, Berney M, Brandt C, Lambert PH, Siegrist CA. 1996. Neonatal and early life immune responses to various forms of vaccine antigens qualitatively differ from adult responses: Predominance of a Th2-biased pattern which persists after adult boosting. Eur J Immunol 26:1489–1496. Bartolome J, Castillo I, Quiroga JA, Mavas S, Carreno V. 1993. Detection of hepatitis C virus RNA in serum and peripheral blood mononuclear cells. J Hepatol 17(Suppl 3):S90–S93. Bauer HM, Hildesheim A, Schiffman MH, et al. 1993. Determinants of genital human papillomavirus infection in low-risk women in Portland, Oregon. Sex Transm Dis 20:274–278. Beasley RP. 1988. Hepatitis B virus: The major etiology of hepatocellular carcinoma. Cancer 61:1942–1956. Beasley RP, Hwang L-Y, Lin C-C, Chien C-S. 1981. Hepatocellular carcinoma and hepatitis B Virus. A prospective study of 22,707 men in Taiwan. Lancet 2:1129–1133. Beasley RP, Trepo C, Stevens CE, et al. 1977. The e antigen and vertical transmission of hepatitis B surface antigen. Am J Epidemiol 105:94–98. Bedwani R, Renganathan E, EL Kwhsky F, et al. 1998. Schistosomiasis and the risk of bladder cancer in Alexandria, Egypt. Br J Cancer 77:1186–1189. Bellentani S, Pozzato G, Saccoccio G, et al. 1999. Clinical course and risk factors of hepatitis C virus related liver disease in the general population: Report from the Dionysos study. Gut 44:874–880. Berger A. 2000. Th1 and Th2 responses: What are they? Science 321:424. Berry JM, Palefsky JM. 2003. A review of human papillomavirus vaccines: From basic science to clinical trials. Front Biosci 8:s333–s345. Biggar RJ, Engels EA, Whitby D, Kedes DH, Goedert JJ. 2003. Antibody reactivity to latent and lytic antigens to human herpesvirus-8 in longitudinally followed homosexual men. J Infect Dis 187:12–18. Biggar RJ, Henle W, Fleisher G, et al. 1978. Primary Epstein-Barr virus infections in African infants. I. Decline of maternal antibodies and time of infection. Int J Cancer 22:239–243. Biggar RJ, Rabkin CS. 1992. The epidemiology of acquired immunodeficiency syndrome-related lymphomas. Current Opin Oncol 4:883–893. Biggar RJ, Whitby D, Marshall V, Linhares AC, Black F. 2000. Human herpesvirus 8 in Brazilian Amerindians: A hyperendemic population with a new subtype. J Infect Dis 181:1562–1568. Birkmann A, Mahr K, Ensser A, et al. 2001. Cell surface heparan sulfate is a receptor for human herpesvirus 8 and interacts with envelope glycoprotein K8.1. J Virol 75:11583–11593. Bjorge T, Dillner J, Anttila T, et al. 1997. Prospective seroepidemiological study of role of human papillomavirus in non-cervical anogenital cancers. BMJ 315:646–649. Bjorge T, Engeland A, Luostarinen T, et al. 2002. Human papillomavirus infection as a risk factor for anal and perianal skin cancer in a prospective study. Br J Cancer 87:61–64. Black FL, Hierholzer WJ, Pinheiro DP, et al. 1974. Evidence for persistence of infectious agents in isolated human populations. Am J Epidemiol 100:230–250. Blaser MJ, Chyou PH, Nomura A. 1994. Age at establishment of Helicobacter pylori infection and risk of gastric carcinoma, gastric ulcer and duodenal ulcer. A birth order and sibship size study [Abstract]. Gastroenterology 106:A53. Blaser MJ, Perez–Perez GI, Kleanthous H, et al. 1995. Infection with Helicobacter pylori strains possessing cagA is associated with an increased risk of developing adenocarcinoma of the stomach. Cancer Res 55:2111–2115. Bleeker MC, Hogewoning CJ, Van Den Brule AJ, et al. 2002. Penile lesions and human papillomavirus in male sexual partners of women with cervical intraepithelial neoplasia. J Am Acad Dermatol 47:351–357. Blumberg BS. 1997. The current state of the prevention of HBV infection, the carrier state and hepatocellular carcinoma. Res Virol 148:91–94. Bodhidatta L, Hoge CW, Churnratanakul S, et al. 1993. Diagnosis of Helicobacter pylori infection in a developing country: Comparison of two ELISAs and a seroprevalence study. J Infect Dis 168:1549–1553. Bohmer G, Van Den Brule AJ, Brummer O, Meijer CL, Petry KU. 2003. No confirmed case of human papillomavirus DNA-negative cervical intraepithelial

Infectious Agents neoplasia grade 3 or invasive primary cancer of the uterine cervix among 511 patients. Am J Obstet Gynecol 189:118–120. Bonis PA, Tong MJ, Blatt LM, Conrad A, Griffith JL. 1999. A predictive model for the development of hepatocellular carcinoma, liver failure, or liver transplantation for patients presenting to clinic with chronic hepatitis C. Am J Gastroenterol 94:1605–1612. Boot H, De Jong D. 2002. Gastric lymphoma: The revolution of the past decade. Scand J Gastroenterol 37(Suppl 236):27–36. Bosch FX. 1997. Global epidemiology of hepatocellular carcinoma. In: Okuda K, Tabor E, eds. Liver Cancer. New York: Churchill Livingstone, pp. 13–28. Bosch FX, De Sanjose S. 2003. Chapter 1: Human papillomavirus and cervical cancer—burden and assessment of causality. J Natl Cancer Inst Monogr 31:3–13. Bosch FX, Lorincz A, Munoz N, Meijer CJ, Shah KV. 2002. The causal relation between human papillomavirus and cervical cancer. J Clin Pathol 55:244– 265. Bosch FX, Manos MM, Munoz N, et al. 1995. Prevalence of human papillomavirus in cervical cancer: A worldwide perspective. International Biological Study on Cervical Cancer (IBSCC) Study Group. J Natl Cancer Inst 87:796–802. Boschi–Pinto C, Stuver S, Okayama A, et al. 2000. A follow-up study of morbidity and mortality associated with hepatitis C virus infection and its interaction with human T lymphotropic virus type I in Miyazaki, Japan. J Infect Dis 181:35–41. Bouvier–Alias M, Patel K, Dahari H, et al. 2002. Clinical utility of total HCV core antigen quantification: A new indirect marker of HCV replication. Hepatology 36:211–218. Brechot C, Gozuacik D, Murakami Y, Paterlini-Brechot P. 2000. Molecular bases for the development of hepatitis B virus (HBV)-related hepatocellular carcinoma (HCC). Semin Cancer Biology 10:211–231. Brechot C, Nalpas B, Feitelson MA. 1996. Interactions between alcohol and hepatitis viruses in the liver. Clin Lab Med 16:273–287. Brechot C, Pourcel C, Hadchouel M, et al. 1982. State of hepatitis B virus DNA in liver diseases. Hepatology 2:27S–34S. Brinton LA, Fraumeni JF, Jr. 1986. Epidemiology of uterine cervical cancer. J Chronic Dis 39:1051–1065. Broker TR, Chow LT. 2001. Human papillomaviruses: A window into the mechanism and regulation of eukaryotic cellular DNA replication. Dev Biol (Basel) 106:367–373. Brown EE, Whitby D, Vitale F, et al. 2005. Correlates of human herpesvirus-8 DNA detection among adults in Italy without Kaposi sarcoma. Int J Epidemiol 34:1110–1117. Brown LM. 2000. Helicobacter pylori: Epidemiology and routes of transmission. Epidemiol Rev 22:283–297. Buonaguro FM, Tornesello ML, Beth-Giraldo E, et al. 1996. Herpesvirus-like DNA sequences detected in endemic, classic, iatrogenic and epidemic Kaposi’s sarcoma (KS) biopsies. Int J Cancer 65:25–28. Burns DA. 1992. “Warts and all”—the history and folklore of warts: A review. J R Soc Med 85:37–40. Busch MP, Laycock M. Kleinman SH, et al. 1994. Accuracy of supplementary serologic testing for human T-lymphotropic virus types I and II in United States blood donors. Blood 83:1143–1148. Cacciola I, Pollicino T, Squadrito G, Cerenzia G, Orlando ME, Raimondo G. 1999. Occult hepatitis B virus infection in patients with chronic hepatitis C liver disease. N Engl J Med 341:22–26. Cacoub P, Fabiani FL, Musset L, et al. 1994. Mixed cryoglobulinemia and hepatitis C virus. Am J Med 96:124–132. Calabrò ML, Sheldon J, Favero A, et al. 1998. Seroprevalence of Kaposi’s sarcomaassociated herpesvirus/human herpesvirus 8 in several regions of Italy. J Hum Virol 3:207–213. Campbell TB, Borok M, Gwanzura L. 1999. HHV-8 peripheral-blood viral load and the titer of antibodies against HHV-8. N Engl J Med 341:1241–1242. Cann AJ, Chen ISY. 1996. Human T-cell leukemia viruses types I and II. In: Fields BN, Knipe DM, Howley PM, et al., eds. Fields Virology. Third edition. Philadelphia: Lippincott–Raven, pp. 1849–1880. Cannon JS, Ciufo D, Hawkins AL, et al. 2000. A new primary effusion lymphomaderived cell line yields a highly infectious Kaposi’s sarcoma herpesviruscontaining supernatant. J Virol 74:10187–10193. Carter JJ, Koutsky LA, Hughes JP, et al. 2000. Comparison of human papillomavirus types 16, 18, and 6 capsid antibody responses following incident infection. J Infect Dis 181:1911–1919. Castellsague X, Bosch FX, Munoz N, et al. 2002. Male circumcision, penile human papillomavirus infection, and cervical cancer in female partners. N Engl J Med 346:1105–1112. Castellsague X, Munoz N. 2003. Chapter 3: Cofactors in human papillomavirus carcinogenesis—role of parity, oral contraceptives, and tobacco smoking. J Natl Cancer Inst Monogr 31:20–28. Castle PE, Giuliano AR. 2003. Chapter 4: Genital tract infections, cervical inflammation, and antioxidant nutrients—assessing their roles as human papillomavirus cofactors. J Natl Cancer Inst Monogr 31:29–34. Castle PE, Schiffman M, Herrero R, et al. 2005. A prospective study of age trends in cervical human papillomavirus acquisition and persistence in Guanacaste, Costa Rica. J Infect Dis 191:1808–1816.

537

Castle PE, Schiffman M, Gravitt PE, et al. 2002a. Comparisons of HPV DNA detection by MY09/11 PCR methods. J Med Virol 68:417–423. Castle PE, Wacholder S, Lorincz AT, et al. 2002b. A prospective study of high-grade cervical neoplasia risk among human papillomavirus-infected women. J Natl Cancer Inst 94:1406–1414. Cesarman E, Chang Y, Moore PS, Said JW, Knowles DM. 1995. Kaposi’s sarcomaassociated herpesvirus-like DNA sequences in AIDS-related body-cavitybased lymphomas. N Engl J Med 332:1186–1191. Cesarman E, Nador R, Knowles DM. 1996. Body-cavity-based lymphoma in an HIV–seronegative patient without Kaposi’s sarcoma-associated herpesviruslike DNA sequences [Letter]. N Engl J Med 334:272–273. Chan KH, Tam JS, Peiris JS, Seto WH, Ng MH. 2001. Epstein-Barr virus (EBV) infection in infancy. J Clin Virol 21:57–62. Chandran B, Smith MS, Koelle DM, Corey L, Horvat R, Goldstein E. 1998. Reactivities of human sera with human herpesvirus-8-infected BCBL-1 cells and identification of HHV-8-specific proteins and glycoproteins and the encoding cDNAs. Virology 243:208–217. Chang C-C, Yu M-W, Lu C-F, Yang C-S, Chen C-J. 1994. A nested case-control study on association between hepatitis C virus antibodies and primary liver cancer in a cohort of 9,775 men in Taiwan. J Med Virol 43:276–280. Chang ET, Zheng T, Lennette ET, et al. 2004. Heterogeneity of risk factors and antibody profiles in Epstein-Barr genome-positive and -negative Hodgkin lymphoma. J Infect Dis 189:2271–2281. Chang ET, Zheng T, Weir E. 2004. Aspirin and the risk of Hodgkin’s lymphoma in a population-based case-control study. J Natl Cancer Inst 96:305–315. Chang MH, Chen CJ, Lai MS, et al. 1997. Universal hepatitis B vaccination in Taiwan and the incidence of hepatocellular carcinoma in children. Taiwan Childhood Hepatoma Study Group. N Engl J Med 336:1855–1859. Chang Y, Cesarman E, Pessin MS, et al. 1994. Identification of herpesvirus-like DNA sequences in AIDS-associated Kaposi’s sarcoma. Science 266:1865–1869. Chatani M, Teshima T, Inoue T, et al. 1991. Antibody response against the EpsteinBarr virus-coded nuclear antigen 2 (EBNA2) in nasopharyngeal carcinoma. Laryngoscope 101:626–629. Chatlynne LG, Lapps W, Handy M, et al. 1998. Detection and titration of human herpesvirus-8-specific antibodies in sera from blood donors, acquired immunodeficiency syndrome patients, and Kaposi’s sarcoma patients using a whole virus enzyme-linked immunosorbent assay. Blood 92:53–58. Chayama K, Kobayashi M, Tsubota A, et al. 1995. Molecular analysis of intraspousal transmission of hepatitis C virus. J Hepatol 22:431–439. Chen C-J, Yu M-W, Liaw Y-F. 1997. Epidemiological characteristics and risk factors of hepatocellular carcinoma. J Gastroenterol Hepatol 12(Suppl):S294– S308. Chen HL, Chang MH, Ni YH, et al. 1996. Seroepidemiology of hepatitis B virus infection in children: Ten years of mass vaccination in Taiwan. JAMA 276:906–908. Chen T-Z, Wu J-C, Yen F-S, et al. 1995. Injection with nondisposable needles as an important route for transmission of acute community-acquired hepatitis C virus infection in Taiwan. J Med Virol 46:247–251. Chen W, Shu D, Chadwick VS. 2001. Helicobacter pylori infection: Mechanism of colonization and functional dyspepsia. Reduced colonization of gastric mucosa by Helicobacter pylori in mice deficient in interleukin-10. J Gastroenterol Hepatol 16:377–383. Chiaramonte M, Stroffolini T, Lorenzoni U, et al. 1996. Risk factors in communityacquired chronic hepatitis C virus infection: A case-control study in Italy. J Hepatol 24:129–134. Chien Y–C, Chen J–Y, Liu M–Y, et al. 2001. Serologic markers of Epstein–Barr virus infection and nasopharyngeal carcinoma in Taiwanese men. N Engl J Med 345:1877–1882. Choo QL, Kuo G, Weiner AJ, et al. 1989. Isolation of a cDNA clone derived from a blood-borne non-A, non-B viral hepatitis genome. Science 244:359–362. Chow WH, Blaser MJ, Blot WJ, et al. 1998. An inverse relation between cagA+ strains of Helicobacter pylori infection and risk of esophageal and gastric cardia adenocarcinoma. Cancer Res 58:588–590. Chung RT, Liang TJ. 1999. Hepatitis C virus and hepatocellular carcinoma. In: Parsonnet J, ed. Microbes and Malignancy: Infection as a Cause of Human Cancers. New York: Oxford University Press, pp. 267–288. Cogliano V, Baan R, Straif K, Grosse Y, Secretan B, El Ghissassi F. WHO International Agency for Research on Cancer. 2005. Carcinogenicity of human papillomaviruses. Lancet Oncol 6:204. Colin J-F, Cazals-Hatem D, Loriot MA, et al. 1999. Influence of human immunodeficiency virus infection on chronic hepatitis B in homosexual men. Hepatology 29:1306–1310. Conry-Cantilena C, Vanraden M, Gibble J, et al. 1996. Routes of infection, viremia, and liver disease in blood donors found to have hepatitis C virus infection. N Engl J Med 334:1691–1696. Cookson Wo, Moffatt Mf. 1997. Asthma: An epidemic in the absence of infection? Science 275:41–42. Correa P, Fontham ET, Bravo JC, et al. 2000. Chemoprevention of gastric dysplasia: Randomized trial of antioxidant supplements and anti-Helicobacter pylori therapy. [Comment]. J Natl Cancer Inst 92:1881–1888. Correa P, Haenszel W, Cuello C, Tannenbaum S, Archer M. 1975. A model for gastric cancer epidemiology. Lancet 2:58–60.

538

PART III: THE CAUSES OF CANCER

Cover TL, Krishna US, Israel DA, Peek RM, Jr. 2003. Induction of gastric epithelial cell apoptosis by Helicobacter pylori vacuolating cytotoxin. Cancer Res 63:951–957. Cowan EP, Nemo GJ, Williams AE, et al. 1999. Absence of human Tlymphotropic virus type I tax sequences in a population of normal blood donors in the Baltimore, MD/Washington, DC, area: Results from a multicenter study. Transfusion 39:904–909. Cox JT, Schiffman M, Solomon D. 2003. Prospective follow-up suggests similar risk of subsequent cervical intraepithelial neoplasia grade 2 or 3 among women with cervical intraepithelial neoplasia grade 1 or negative colposcopy and directed biopsy. Am J Obstet Gynecol 188:1406–1412. Crabtree JE. 1996. Gastric mucosal inflammatory responses to Helicobacter pylori. Aliment Pharmacol Ther 10(Suppl 1):29–37. Crabtree JE, Covacci A, Farmery SM, et al. 1995. Helicobacter pylori induced interleukin-8 expression in gastric epithelial cells is associated with CagA positive phenotype. J Clin Pathol 48:41–45. D’Elios MM, Manghetti M, Decarli M, et al. 1997. T helper 1 effector cells specific for Helicobacter pylori in the gastric antrum of patients with peptic ulcer disease. J Immunol 158:962–967. Davis GL. 1999. Hepatitis C virus genotypes and quasispecies. Am J Med 107(Suppl 6B):21S–26S. De Francesco R. 1999. Molecular virology of the hepatitis C virus. J Hepatol 31(Suppl 1):47–53. De Franchis, R, Meucci R, Vecchi M, et al. 1993. The natural history of asymptomatic hepatitis B surface antigen carriers. Ann Int Med 118:191–194. De Mitri MS, Poussin K, Baccarini P, et al. 1995. HCV-associated liver cancer without cirrhosis. Lancet 345:413–415. De Re V, De Vita S, Marzotto A, et al. 2000. Sequence analysis of the immunoglobulin antigen receptor of hepatitis C virus-associated non-Hodgkin lymphomas suggests that the malignant cells are derived from the rheumatoid factor–producing cells that occur mainly in type II cryoglobulinemia. Blood 96:3578– 3584. De Thé G, Bestetti G, Van Beveren M, Gessain A. 1999. Prevalence of human herpesvirus 8 infection before the acquired immunodeficiency disease syndromerelated epidemic of Kaposi’s sarcoma in East Africa. J Natl Cancer Inst 91:1888–1889. De Thé G, Geser A, Day NE, et al. 1978. Epidemiological evidence for causal relationship between Epstein-Barr virus and Burkitt’s lymphoma from Ugandan prospective study. Nature 274:756–761. De Thé G, Ho JHC, Muir CS. 1989. Nasopharyngeal cancer. In: Evans AS, ed. Viral Infections of Humans: Epidemiology and Control. Third edition. New York: Plenum Press, pp. 737–767. De Villiers EM. 2001. Taxonomic classification of papillomaviruses. Papillomavirus Report 12:57–63. De Villiers EM. 2003. Relationship between steroid hormone contraceptives and HPV, cervical intraepithelial neoplasia and cervical carcinoma. Int J Cancer 103:705–708. Devesa SS, Blot WJ, Fraumeni JF, Jr. 1998. Changing patterns in the incidence of esophageal and gastric carcinoma in the United States. Cancer 83:2049– 2053. Dezzutti CS, Guenthner PC, Daniel S, et al. 2003. Detection of human T-lymphotropic virus (HTLV) tax sequences in New York City blood donors seronegative HTLV types 1 and 2. Clin Diagn Lab 10:715–717. Dibisceglie AM, Hoofnagle JH. 2002. Optimal therapy of hepatitis C. Hepatology 36(Suppl 1):S121–S127. Dillner J, Andersson-Ellstrom A, Hagmar B, Schiller J. 1999. High risk genital papillomavirus infections are not spread vertically. Rev Med Virol 9:23–29. Dolcetti R, Masucci MG. 2003. Epstein-Barr virus: Induction and control of cell transformation. J Cell Physiol 196:207–218. Doll R. 1978. An epidemiologic perspective of the biology of cancer. Cancer Res 38:3573–3583. Donato F, Boffetta P, Puoti M. 1998. A meta-analysis of epidemiological studies on the combined effect of hepatitis B and C virus infections in causing hepatocellular carcinoma. Int J Cancer 75:347–354. Donato F, Tagger A, Gelatti U, et al. 2002. Alcohol and hepatocellular carcinoma: The effect of lifetime intake and hepatitis virus infections in men and women. Am J Epidemiol 155:323–331. Dosik H, Goldstein MF, Poiesz J, et al. 1994. Seroprevalence of human Tlymphotropic virus in blacks from a selected central Brooklyn population. Cancer Invest 12:289–295. Drouet E, Brousset P, Fares F, et al. 1999. High Epstein-Barr virus serum load and elevated titers of anti-ZEBRA antibodies in patients with EBVharboring tumor cells of Hodgkin’s disease. J Med Virol 57:383–389. Eiben GL, Da Silva DM, Fausch SC, Le Poole IC, Nishimura MI, Kast WM. 2003. Cervical cancer vaccines: Recent advances in HPV research. Viral Immunol 16:111–121. Ekstrom AM, Serafini M, Nyren O, Hansson LE, Ye W, Wolk A. 2000. Dietary antioxidant intake and the risk of cardia cancer and noncardia cancer of the intestinal and diffuse types: A population–based case-control study in Sweden. Int J Cancer 87:133–140. El Omar EM, Carrington M, Chow WH, et al. 2000. Interleukin-1 polymorphisms associated with increased risk of gastric cancer. Nature 404:398–402.

El Omar EM, Rabkin CS, Gammon MD, et al. 2003. Increased risk of noncardia gastric cancer associated with proinflammatory cytokine gene polymorphisms. Gastroenterology 124:1193–1201. El-Sarag HB, Mason AC. 1999. Rising incidence of hepatocellular carcinoma in the United States. N Engl J Med 340:745–750. Engels EA, Eastman H, Ablashi DV, Wilks RJ, Braham J, Manns A. 1999. Risk of transfusion-associated transmission of human herpesvirus 8. J Natl Cancer Inst 91:1773–1775. Engels EA, Whitby D, Goebel PB, et al. 2000. Identifying human herpesvirus 8 infection: Performance characteristics of serologic assays. JAIDS 23: 346–354. Epstein MA. 1999. On the discovery of Epstein-Barr virus: A memoir. Epstein-Barr Virus Report 7:58–63. Epstein MA, Achong BG, Barr YM. 1964. Virus particles in cultured lymphoblasts from Burkitt’s lymphoma. Lancet i:702–703. Epstein MA, Hunt RD, Rabin H. 1973. Pilot experiments with EB virus in owl monkeys (Aotus Tririgotus): Reticuloproliferative disease in an inoculated animal. Int J Cancer 12:309–318. Eslick GD, Lim LL, Byles JE, Xia HH, Talley NJ. 1999. Association of Helicobacter pylori infection with gastric carcinoma: A meta-analysis. Am J Gastroenterol 94:2373–2379. Etoh K, Tamiya S, Yamaguchi K, et al. 1997. Persistent clonal proliferation of human T-lymphotropic virus type I-infected cells in vivo. Cancer Res 57: 4862–4867. Etoh K, Yamaguchi K, Tokudome S, et al. 1999. Rapid quantification of HTLV-I provirus load: Detection of monoclonal proliferation of HTLV-I-infected cells among blood donors. Int J Cancer 81:859–864. Eurogast Study Group. 1993. An international association between Helicobacter pylori infection and gastric cancer. Lancet 341:1359–1362. Evans AA, Chen G, Ross EA, Shen F–M, Lin W–Y, London WT. 2002. Eight year follow–up of the 90,000-person Haimen City Cohort: I. Hepatocellular carcinoma mortality, risk factors, and gender differences. Cancer Epidemiol, Biomarkers Prev 11:369–376. Evans AA, Fine M, London WT. 1997. Spontaneous seroconversion in hepatitis B e antigen-positive chronic hepatitis B: Implications for interferon therapy. J Infect Dis 176:845–850. Evans AS, Comstock GW. 1981. Presence of elevated antibody titers to Epstein-Barr virus before Hodgkin’s disease. Lancet 1:1183–1186. Evans AS, Mueller NE. 1990. Viruses and cancer-causal associations. Ann Epidemiol 1:71–92. Evans AS, Mueller NE. 1997. Epstein-Barr virus and malignant lymphomas. In: Evans AS, Kaslow R, eds. Viral Infections of Humans: Epidemiology and Control. Fourth edition. New York: Plenum Press, pp. 895–967. Evans AS, Niederman JC. 1989. Epstein-Barr virus. In: Evans AS, ed. Viral Infections of Humans: Epidemiology and Control, Third edition. New York: Plenum Press, pp. 265–292. Everhart JE, Kruszon-Moran D, Perez-Perez GI, Tralka TS, Mcquillan G. 2000. Seroprevalence and ethnic differences in Helicobacter pylori infection among adults in the United States. J Infect Dis 181:1359–1363. Eyster ME, Fried MW, Dibisceglie AM, Goedert JJ. 1994. Increasing hepatitis C virus RNA levels in hemophiliacs: Relationship to human immunodeficiency virus infection and liver disease. Blood 84:1020–1023. Fahraeus R, Fu HL, Ernbergi, et al. 1988. Expression of Epstein-Barr virus-encoded proteins in nasopharyngeal carcinoma. Int J Cancer 42:329–338. Falk GW. 2001. The possible role of Helicobacter pylori in GERD. Sem Gastrointestinal Dis 12:186–195. Farinati F, Cardin R, Degan P, et al. 1998. Oxidative DNA damage accumulation in gastric carcinogenesis. Gut 42:351–356. Fehrmann F, Laimins LA. 2003. Human papillomaviruses: Targeting differentiating epithelial cells for malignant transformation. Oncogene 22:5201–5207. Fendrick AM, Chernew ME, Hirth RA, Bloom BS, Bandekar RR, Scheiman JM. 1999. Clinical and economic effects of population-based Helicobacter pylori screening to prevent gastric cancer. Arch Intern Med 159:142–148. Ferrari C, Urbani S, Penna A, et al. 1999. Immunopathogenesis of hepatitis C virus infection. J Hepatol 31(Suppl 1):31–38. Ferri C, Baicchi U, La Civita L, et al. 1993. Hepatitis C virus-related autoimmunity in patients with porphyria cutanea tarda. Eur J Clin Invest 23:851–855. Ferri C, Caracciolo F, Zignego AL, et al. 1994. Hepatitis C virus infection in patients with non-Hodgkin’s lymphoma. Br J Haematol 88:392–394. Fingeroth JD, Weis JJ, Tedder TF, Strominger JL, Biro PA, Fearson DT. 1984. Epstein-Barr virus receptor of human B lymphocytes in the C3d receptor CR2. Proc Natl Acad Sci USA 81:4510–4514. Fiorelli V, Gendelman R, Sirianni MC, et al. 1998. g-Interferon produced by CD8+ T cells infiltrating Kaposi’s sarcoma induces spindle cells with angiogenic phenotype and synergy with human immunodeficiency virus-1 Tat protein: An immune response to human herpesvirus 8 infection? Blood 91:956–967. Forman D, Sitas F, Newell DG, et al. 1990. Geographic association of Helicobacter pylori antibody prevalence and gastric cancer mortality in rural China. Int J Cancer 46:608–611. Fox JG, Beck P, Dangler CA, et al. 2000. Concurrent enteric helminth infection modulates inflammation and gastric immune responses and reduces helicobacter-induced gastric atrophy. [Comment]. Nature Med 6:536–542.

Infectious Agents Franceschi S, Castellsague X, Dal Maso L, et al. 2002. Prevalence and determinants of human papillomavirus genital infection in men. Br J Cancer 86:705–711. Franceschi S, Geddes M. 1995. Epidemiology of classic Kaposi’s sarcoma, with special reference to Mediterranean population. Tumori 81:308–314. Francis DP, Essex M, Cotter S, et al. 1979. Feline leukemia virus infections: The significance of chronic viremia. Leuk Res 3:435–441. Francis DP, Essex M, Maynard JE. 1981. Feline leukemia virus and hepatitis B virus: A comparison of late manifestations. Prog Med Virol 27:127–132. Franco EL. 1992. Measurement errors in epidemiological studies of human papillomavirus and cervical cancer. In: Munoz N, Bosch FX, Shah KV, Meheus A, eds. The Epidemiology of Cervical Cancer and Human Papillomavirus. IARC Sci Publ No. 119. Lyon: International Agency for Research on Cancer, pp. 181–197. Franco EL. 2003. Chapter 13: Primary screening of cervical cancer with human papillomavirus tests. J Natl Cancer Inst Monogr 31:89–96. Franco EL, Villa LL, Sobrinho JP, et al. 1999. Epidemiology of acquisition and clearance of cervical human papillomavirus infection in women from a high-risk area for cervical cancer. J Infect Dis 180:1415–1423. Frank C, Mohamed MK, Strickland GT, et al. 2000. The role of parenteral antischistosomal therapy in the spread of hepatitis C virus in Egypt. Lancet 355:887–891. Freeman AJ, Dore GJ, Law MG, et al. 2001. Estimating progression to cirrhosis in chronic hepatitis C virus infection. Hepatology 34:809–816. Frisch M, Glimelius B, Van Den Brule AJ, et al. 1997. Sexually transmitted infection as a cause of anal cancer. N Engl J Med 337:1350–1358. Fuchs PG, Girardi F, Pfister H. 1988. Human papillomavirus DNA in normal, metaplastic, preneoplastic and neoplastic epithelia of the cervix uteri. Int J Cancer 41:41–45. Furukawa Y, Fujisawa J, Osame M, et al. 1992. Frequent clonal proliferation of human T-cell leukemia virus type I (HTLV-I)-infected T cells in HTLV-Iassociated myelopathy (HAM-TSP). Blood 60:1012–1016. Gabet A–S, Gessain A, Wattel E. 2003. High simian T-cell leukemia virus type 1 provirus loads combined with genetic stability as a result of cellassociated provirus replication in naturally infected, asymptomatic monkeys. Int J Cancer 107:74–83. Gabet A-S, Mortreux F, Talarmin A, et al. 2000. High circulating proviral load with oligoclonal expansion of HTLV-I bearing T cells in HTLV-1 carriers with strongyloidiasis. Oncogene 19:4954–4960. Gan YJ, Sullivan JL, Sixbey JW. 1994. Detection of cell-free Epstein-Barr virus DNA in serum during acute infectious mononucleosis. J Infect Dis 170:436–439. Ganem D. 1996. Hepadnaviridae and their replication. In: Fields BN, Knipe DM, Howley PM, et al., eds. Fields Virology, Third edition. Philadelphia: Lippincott–Raven, pp. 2703–2737. Gatza ML, Watt JC, Marriott SJ. 2003. Cellular transformation by the HTLV-I Tax protein, a Jack-of-all trades. Oncogene 22:5141–5149. Geser A, De The G, Lenoir G, et al. 1982. Final case reporting from the Ugandan prospective study of the relationship between EBV and Burkitt’s lymphoma. Int J Cancer 29:397–400. Gessain A, Duprez R. 2005. Spindle cells and their role in Kaposi’s sarcoma. Int J Biochem Cell Biol 37:2457–2465. Gessain A, Gallo RC, Franchini G. 1992. Low degree of human T-cell/lymphoma virus type I genetic drift in vivo as a means of monitoring viral transmission and movement of ancient human populations. J Virol. 66:2288–2295. Gessain A, Mauclere P, Van Beveren M, et al. 1999. Human herpesvirus 8 primary infection occurs during childhood in Cameroon, Central Africa. Int J Cancer 81:189–192. Gilson RJ, Hawkins AE, Beecham MR, et al. 1997. Interactions between HIV and hepatitis B virus in homosexual men: Effects on the natural history of infection. AIDS 11:597–606. Glaser SL, Lin RJ, Stewart SL, et al. 1997. Epstein-Barr virus-associated Hodgkin’s disease: Epidemiologic characteristics in international data. Int J Cancer 70:375–382. Goedert JJ, Kedes DH, Ganem D. 1997. Antibodies to human herpesvirus 8 in women and infants born in Haiti and the USA. Lancet 349:1368. Goedert JJ, Vitale F, Lauria C, et al. 2002. Risk factors for classical Kaposi’s sarcoma. J Natl Cancer Inst 94:1712–1720. Goldstein ST, Alter MJ, Williams It, et al. 2002. Incidence and risk factors for acute hepatitis B in the United States, 1982–1998: Implications for vaccination programs. J Infect Dis 185:713–719. Goudsmit J, Renwick N, Dukers Nhtm, et al. 2000. Human herpesvirus 8 infections in the Amsterdam Cohort Studies (1984–1997): Analysis of seroconversions to ORF65 and ORF73. Proc Natl Acad Sci USA 97:4838– 4843. Graham CS, Baden LR, Yu E, et al. 2001. Influence of human immunodeficiency virus infection on the course of hepatitis C virus infection: A meta-analysis. CID 33:562–569. Gravitt PE, Burk RD, Lorincz A, et al. 2003. A comparison between real-time polymerase chain reaction and hybrid capture 2 for human papillomavirus DNA quantitation. Cancer Epidemiol Biomarkers Prev 12:477–484. Gravitt PE, Castle PE. 2001. Chlamydia trachomatis and cervical squamous cell carcinoma. JAMA 285:1703–1704.

539

Green J, Berrington DG, Smith JS, et al. 2003. Human papillomavirus infection and use of oral contraceptives. Br J Cancer 88:1713–1720. Green MH, Fraumeni JF, Jr, Hoover R. 1977. Nasopharyngeal cancer among adults in United States: Racial variations by cell type. J Natl Cancer Inst 58:1267–1270. Gretch D, Coety I, Wilson J, et al. 1994. Assessment of hepatitis C virus RNA levels by quantitative competitive RNA polymerase chain reaction: High-titer viremia correlates with advanced stage of disease. J Infect Dis 169:1219–1225. Gulley ML, Glaser SL, Craig FE, et al. 2002. Guidelines for interpreting EBER in situ hybridization and LMP1 immunohistochemical tests for detecting Epstein-Barr virus in Hodgkin’s lymphoma. Am J Clin Pathol 117:259–267. Gutensohn (Mueller) N, Cole P. 1977. Epidemiology of Hodgkin’s disease in the young. Int J Cancer 19:595–604. Gutensohn (Mueller) N, Cole P. 1981. Childhood social environment and Hodgkin’s disease. N Engl J Med 304:135–140. Haddad J, Deny P, Munz-Gotheil C, et al. 1992. Lymphocytic sialadenitis of Sjogren’s syndrome associated with chronic hepatitis C virus liver disease. Lancet 339:321–323. Hagiwara H, Hayashi N, Mita E, et al. 1993. Quantitation of hepatitis C virus RNA in serum of asymptomatic blood donors and patients with type C chronic liver disease. Hepatology 17:545–550. Harper DM, Franco EL, Wheeler C, et al. 2004. Efficacy of a bivalent L1 virus-like particle vaccine in prevention of infection with human papillomavirus types 16 and 18 in young women: a randomized controlled trial. Lancet 164:1757– 1765. Haswell-Elkins MR, Mairiang E, Mairiang P, et al. 1994a. Cross-sectional study of Opisthorchis viverrini infection and cholangiocarcinoma in communities within a high-risk area in Northeast Thailand. Int J Cancer 59:505–509. Haswell–Elkins MR, Satarug S, Tsuda M, et al. 1994b. Liver fluke infection and cholangiocarcinoma: Model of endogenous nitric oxide and extragastric nitrosation in human carcinogenesis. Mutat Res 305:241–252. Hayashi J, Kishihara Y, Yamaji K, et al. 1995. Transmission of hepatitis C virus by health care workers in a rural area of Japan. Am J Gastroenterol 90:794–799. Hayashi J, Kishihara Y, Yamaji K, et al. 1997. Hepatitis C viral quasispecies and liver damage in patients with chronic hepatitis C virus infection. Hepatology 25:697–701. Hayes CG, Burans JP, Oberst RB. 1991. Antibodies to human T lymphotropic virus type I in a population from the Philippines: Evidence for crossreactivity with Plasmodium falciparum. J Infect Dis 163:257–262. Hayward GS. 1999. KSHV strains: The origins and global spread of the virus. Semin Cancer Biol 9:187–199. Helicobacter and Cancer Collaborative Group. 2001. Gastric cancer and Helicobacter pylori: A combined analysis of 12 case control studies nested within prospective cohorts. Gut 49:347–353. Henle G, Henle W, Clifford P, et al. 1969. Antibodies to the Epstein-Barr virus in Burkitt’s lymphoma and control groups. J Natl Cancer Inst 43:1147–1159. Henle G, Henle W, Horwitz CA. 1974. Epstein-Barr specific diagnostic tests in infectious mononucleosis. Hum Pathol 5:551–565. Henle W, Henle G. 1979. Seroepidemiology of the virus. In: Epstein MA, Achong BG, eds. The Epstein-Barr Virus. Berlin: Springer-Verlag, pp. 61–78. Henle W, Henle G, Andersson J, et al. 1987. Antibody responses to Epstein-Barr virus-determined nuclear antigen (EBNA-1 and EBNA-2) in acute and chronic Epstein-Barr virus infection. Proc Natl Acad Sci USA 84:570–574. Henle W, Henle G, Ho H-C, et al. 1970. Antibodies to Epstein-Barr virus in nasopharyngeal carcinoma, other head and neck neoplasms, and control groups. J Natl Cancer Inst 44:225–231. Henle W, Ho JH, Henle G, Chau JC, Kwan HC. 1977. Nasopharyngeal carcinoma: Significance of changes in Epstein-Barr virus-related antibody patterns following therapy. Int J Cancer 20:663–672. Henle W, Ho H-C, Kwan HC. 1973. Antibodies to the Epstein-Barr related antigens in nasopharyngeal carcinoma: Comparison of active cases with long term survivors. J Natl Cancer Inst 51:361–369. Henrik SJ, Forsgren A, Berglund G, Floren CH. 2001. Helicobacter pylori infection is associated with a decreased risk of developing oesophageal neoplasms. Helicobacter 6:310–316. Hermann K, Niedobitek G. 2003. Epstein-Barr virus-associated carcinomas: Facts and fiction. J Pathol 199:140–145. Herrero R. 2003. Chapter 7: Human papillomavirus and cancer of the upper aerodigestive tract. J Natl Cancer Inst Monogr 31:47–51. Herrero R, Castle PE, Schiffman M, et al. 2005. Epidemiologic profile of typespecific human papillomavirus infection and cervical neoplasia in Guanacaste, Costa Rica. J Infec Dis 191:1796–1807. Herrero R, Hildesheim A, Bratti C, et al. 2000. Population-based study of human papillomavirus infection and cervical neoplasia in rural Costa Rica. J Natl Cancer Inst 92:464–474. Heussinger N, Buttner M, Ott G, et al. 2004. Expression of the Epstein-Barr virus (EBV)-encoded latent membrane protein 2A (LMP2A) in EBV-associated nasopharyngeal carcinoma. J Pathol 203:696–699. Hildesheim A, Schiffman M, Bromley C, et al. 2001. Human papillomavirus type 16 variants and risk of cervical cancer. J Natl Cancer Inst 93:315–318. Hildesheim A, Wang SS. 2002. Host and viral genetics and risk of cervical cancer: A review. Virus Res 89:229–240.

540

PART III: THE CAUSES OF CANCER

Hino S, Doi H, Yoshikuni H, et al. 1987. HTLV-I carrier mothers with high–titer antibody are at high risk as a source of infection. Jpn J Cancer Res. 78:1156–1158. Hinuma Y, Nagata K, Hanaoka M, et al. 1981. Adult T-cell leukemia: Antigen in an ATL cell line and detection of antibodies to the antigen in human sera. Proc Nat Acad Sci USA 78:6476–6480. Hisada M, Maloney EM, Sawada T, et al. 2002. Virus markers associated with vertical transmission of human T lymphotropic virus type 1 in Jamaica. Clin Inf Dis 34:1551–1557. Hisada M, Okayama A, Shioiri S, Spiegelman DL, Stuver SO, Mueller NE. 1998. Risk factors for adult T-cell leukemia among carriers of human T-lymphotropic virus type I. Blood 92:3557–3561. Hisada M, Okayama A, Spiegelman D, Mueller NE, Stuver SO. 2001. Sexspecific mortality from adult T-cell leukemia among carriers of human T-lymphotropic virus type I. Int J Cancer 91:497–499. Hisada M, Stuver SO, Okayama A, et al. 2004. Persistent paradox of natural history of human T lymphotropic virus type I: parallel analy-ses of Japanese and Jamaican carriers. J Infect Dis 190:1605–1609. Hisada M, Stuver S, Okayama A, Mueller NE. 1999. Gender difference in skin reactivity to purified protein derivative among carriers of HTLV-I in Japan. JAIDS 22:302–307. Hiscott J, Kwon H, Génin P. 2001. Hostile takeovers: Viral appropriation of the NFKB pathway. J Clin Invest 107:143–151. Hjalgrim H, Askling J, Rostgaard K, et al. 2003. Characteristics of Hodgkin’s lymphoma after infectious mononucleosis. N Engl J Med 349:1324–1332. Ho GY, Bierman R, Beardsley L, Chang CJ, Burk RD. 1998. Natural history of cervicovaginal papillomavirus infection in young women. N Engl J Med 338:423–428. Ho GY, Burk RD, Klein S, et al. 1995. Persistent genital human papillomavirus infection as a risk factor for persistent cervical dysplasia. J Natl Cancer Inst 87:1365–1371. Ho GY, Nomura AM, Nelson K, Lee H, Polk BF, Blattner WA. 1991. Declining seroprevalence and transmission of HTLV-I in Japanese families who immigrated to Hawaii. Amer J Epidemiol 134:981–987. Ho M, Miller G, Atchinson RW, et al. 1985. Epstein-Barr virus infections and DNA hybridization studies in posttransplantation lymphoma and lymphoproliferative lesions: The role of primary infection. J Infect Dis 152:876–886. Holck S, Norgaard A, Bennedsen M, Permin H, Norn S, Andersen LP. 2003. Gastric mucosal cytokine responses in Helicobacter pylori-infected patients with gastritis and peptic ulcers. Association with inflammatory parameters and bacteria load. FEMS Immunol Med Microbiol 36:175–180. Hollinger FB. 1996. Hepatitis B virus. In: Fields BN, Knipe DM, Howley PM, et al., eds. Fields Virology. Third edition. Philadelphia: Lippincott–Raven, pp. 2739–2807. Hoofnagle JH. 2002. Course and outcome of hepatitis C. Hepatology 36(Suppl 1): S21–S29. Hoofnagle JH, Schafer DF. 1986. Serologic markers of hepatitis B virus infection. Sem Liver Dis 6:1–10. Hsieh C-C, Tzonou A, Zavitsanos X, Kaklamani E, Lan S-J, Trichopoulos D. 1992. Age at first establishment of chronic hepatitis B infection and hepatocellular carcinoma risk: A birth order study. Am J Epidemiol 136:1115–1121. Huang J, O’Toole PW, Doig P, Trust TJ. 1995. Stimulation of interleukin-8 production in epithelial cell lines by Helicobacter pylori. Infect Immun 63: 1732–1738. Huang JQ, Sridhar S, Chen Y, Hunt RH. 1998. Meta-analysis of the relationship between Helicobacter pylori seropositivity and gastric cancer. Gastroenterology 114:1169–1179. Hughes JP, Garnett GP, Koutsky L. 2002. The theoretical population-level impact of a prophylactic human papilloma virus vaccine. Epidemiology 13:631–639. Huo T-I, Wu, J-C, Lee P-C, et al. 1998. Sero-clearance of hepatitis B surface antigen in chronic carriers does not necessarily imply a good prognosis. Hepatology 28:231–236. Hussell T, Isaacson PG, Crabtree JE, Spencer J. 1993. The response of cells from low-grade B-cell gastric lymphomas of mucosa-associated lymphoid tissue to Helicobacter pylori. Lancet 342:571–574. Hyams KC, Riddle J, Rubertone M, et al. 2001. Prevalence and incidence of hepatitis C virus infection in the US military: A seroepidemiologic survey of 21,000 troops. Am J Epidemiol 153:764–770. IARC. 1993. Monographs on the Evaluation of Carcinogenic Risks to Humans. Volume 56. Some Naturally Occurring Substances: Food Items and Constituents, Heterocyclic Aromatic Amines and Mycotoxins. Lyon: International Agency for Research on Cancer. IARC. 1994a. Monographs on the Evaluation of Carcinogenic Risks to Humans. Volume 59. Hepatitis Viruses. Lyon: International Agency for Research on Cancer. IARC. 1994b. Monographs on the Evaluation of Carcinogenic Risks to Humans. Volume 61. Schistosomes, Liver Flukes and Helicobacter Pylori. Lyon: International Agency for Research on Cancer. IARC. 1995. Monographs on the Evaluation of Carcinogenic Risks to Humans. Volume 64. Papillomaviruses. Lyon: International Agency for Research on Cancer.

IARC. 1996. Monographs on the Evaluation of Carcinogenic Risks to Humans. Volume 67. Human Imunodeficiency Viruses and Human T–Cell Lymphotropic Viruses. Lyon: International Agency for Research on Cancer. IARC. 1997. Monographs on the Evaluation of Carcinogenic Risks to Humans. Volume 70. Epstein-Barr Virus and Kaposi’s Sarcoma Herpesvirus/Human Herpesvirus 8. Lyon: International Agency for Research on Cancer. Iftner T, Villa LL. 2003. Chapter 12: Human papillomavirus technologies. J Natl Cancer Inst Monogr 31:80–88. Igakura T, Stinchcombe JC, Goon PK, et al. 2003. Spread of HTLV-I between lymphocytes by virus-induced polarization of the cytoskeleton. Science 299:1713– 1716. Ikeda K, Saitoh S, Suzuki Y, et al. 1998. Interferon decreases hepatocellular carcinogenesis in patients with cirrhosis caused by the hepatitis B virus: A pilot study. Cancer 82:827–835. Inoue G, Horiike N, Michitaka K, Onji M. 2000. Hepatitis C virus clearance is prominent in women in an endemic area. J Gastroenterol and Hepatol 15:1054–1058. Isaacson PG, Muller–Hermelink HK, Piris MA, et al. 2001. Extranodal marginal zone B-cell lymphoma of mucosa-associated lymphoid tissue (MALT lymphoma). In: Jaffe ES, Harris NL, Stein H, Vardiman JW, eds. Pathology and Genetics. Tumours of Haematopoietic and Lymphoid Tissues. Lyon: International Agency for Research on Cancer, pp. 157–160. Iscovich J, Boffetta P, Winkelmann R, Brennan P. 1999. Classic Kaposi’s sarcoma as a second primary neoplasm. Int J Cancer 80:178–182. Ishihara S, Okayama A, Stuver S, et al. 1994. Association of HTLV-I antibody profile of asymptomatic carriers with proviral DNA levels of peripheral blood mononuclear cells. JAIDS 7:199–203. Iwata (Arisawa) K, Ito S-I, Saito H, et al. 1994. Mortality among inhabitants of an HTLV-I endemic area in Japan. Jpn J Cancer Res 85:231–237. Jacobson LP, Jenkins FJ, Springer G, et al. 2000. Interaction of human immunodeficiency virus type 1 and human herpesvirus type 8 infections on the incidence of Kaposi’s sarcoma. J Infect Dis 181:1940–1949. Jamieson A, Curry A, Martinez G. 2001. School enrollment in the United States – social and economic characteristics of students: October 1999. Washington, D.C., United States Census Bureau. Japanese RED Cross. 1991. Non-A, Non-B-Hepatitis Research Group. Effect of screening for hepatitis C virus antibody and hepatitis B virus core antibody on incidence of post-transfusion hepatitis. Lancet 338:1040–1041. Jeang K-T. 2001. Functional activities of the human T-cell leukemia virus type I Tax oncoprotein: Cellular signaling through NF-kappa B. Cytokine Growth Factor Rev 12:207–217. Johnson JM, Harrod R, Franchini G. 2001 Molecular biology and pathogenesis of the human T-cell leukaemia/lymphotropic virus type-I (HTLV-I). Int J Exp Pathol 82:135–147. Johnson RJ, Gretch DR, Yamabe H, et al. 1993. Membranoproliferative glomerulonephritis associated with hepatitis C virus infection. N Engl J Med 328:465–470. Jubert C, Pawlotsky JM, Pouget F, et al. 1994. Lichen planus and hepatitis C virusrelated chronic active hepatitis. Arch Dermatol 130:73–76. Judde JG, Lacoste V, Briere J, et al. 2000. Monoclonality or oligoclonality of human herpesvirus 8 terminal repeat sequences in Kaposi’s sarcoma and other diseases. J Natl Cancer Inst 92:729–736. Kadish AS. 2001. Biology of anogenital neoplasia. Cancer Treat Res 104:267–286. Kaloterakis A, Papasteriades C, Filiotou A, et al. 1995. HLA in familial and nonfamilial Mediterranean Kaposi’s sarcoma in Greece. Tissue Antigens 45:117–119. Kalyanaraman VS, Sarngadharan MG, Robert-Guroff M, et al. 1982. A new subtype of human T-cell leukemia virus (HTLV-II) associated with a T-cell variant of hairy cell leukemia. Science 218:571–573. Kaplan JE, Lal RB, Davidson M, Lanier AP, Lairmore MD. 1993. HTLV-I in Alaska Natives. JAIDS 6:327–328. Kedes DH, Operskalski E, Busch M, Kohn R, Flood J, Ganem D. 1996. The seroepidemiology of human herpesvirus 8 (Kaposi’s sarcoma-associated herpesvirus): Distribution of infection in KS risk groups and evidence for sexual transmission. Nature Med 2:918–924. Kekule AS, Lauer U, Meyer M, Caselmann WH, Hofschneider PH, Koshy R. 1990. The preS2/S region of integrated hepatitis B virus DNA encodes a transcriptional transactivator. Nature 343:457–461. Kenny-Walsh E. 1999. For the Irish Hepatology Research Group. Clinical outcomes after hepatitis C infection from contaminated anti-D immune globulin. N Engl J Med 340:1228–1233. Kero J, Gissler M, Hemminki E, Isolauri E. 2001. Could TH1 and TH2 diseases coexist? Evaluation of asthma incidence in children with celiac disease, type 1 diabetes, or rheumatoid arthritis: A register study. J Allergy Clin Immunol 108:781–783. Khan MJ, Castle PE, Lorincz AT, et al. 2005. The elevated 10-year risk of cervical precancer and cancer in women with human papillomavirus (HPV) type 16 or 18 and the possible utility of type-specific HPV testing in clinical practice. J Natl Cancer Inst 97:1072–1079. Khanna R, Burrows SR. 2000. Role of cytotoxic T lymphocytes in Epstein-Barr virus-associated diseases. Ann Rev Microbiol 54:19–48. Kieff E. 1996. Epstein-Barr virus and its replication. In: Fields BN, Knipe DM, Howley PM, et al., eds. Fields Virology, Third edition. Philadelphia: Lippincott–Raven, pp. 2343–2396.

Infectious Agents Kieff E, Rickinson AB. 2001. Epstein-Barr virus and its replication. In: Knipe DM, Howley PM, eds. Fields Virology, Fourth edition. Philadelphia: Lippincott Williams and Wilkins, pp. 2511–2573. Kim C, Koike K, Saito I, Miyamura T, Jay G. 1991. HBx gene of hepatitis B virus induces liver cancer in transgenic mice. Nature 351:317–320. Kim DW, Suzuki R, Harada T, Saito I, Miyamura T. 1994. Trans-suppression of gene expression by hepatitis C viral core protein. Jpn J Med Sci Biol 47:211–220. Kim JH, Durack DT. 1988. Manifestations of human T-lymphotropic virus type I infection. Am J Med 84:919–928. Kinoshita K, Amagasaki T, Ikeda S, et al. 1985. Preleukemic state of adult T-cell leukemia: Abnormal T lymphocytosis induced by human adult T-cell leukemia-lymphoma virus. Blood 56:120–127. Kinoshita T, Imamura J, Nagai H, et al. 1993. Absence of HTLV-I infection among seronegative subjects in an endemic area of Japan. Int J Cancer 54:16–19. Kiyosawa K, Sodeyama T, Tanaka E, et al. 1990. Interrelationship of blood transfusion, non-A, non-B hepatitis and hepatocellular carcinoma: Analysis by detection of antibody to hepatitis C virus. Hepatology 12:671–675. Kiyosawa K, Tanaka E, Sodeyama T, et al. 1994. Transmission of hepatitis C in an isolated area in Japan: Community-acquired infection. Gastroenterol 106:1596–1602. Kjaer SK, Chackerian B, Van Den Brule AJ, et al. 2001. High-risk human papillomavirus is sexually transmitted: Evidence from a follow-up study of virgins starting sexual activity (intercourse). Cancer Epidemiol Biomarkers Prev 10:101–106. Kjaer SK, Svare EI, Worm AM, Walboomers JM, Meijer CJ, Van Den Brule AJ. 2000. Human papillomavirus infection in Danish female sex workers. Decreasing prevalence with age despite continuously high sexual activity. Sex Transm Dis 27:438–445. Kjaer SK, Van Den Brule AJ, Paull G, et al. 2002. Type specific persistence of high risk human papillomavirus (HPV) as indicator of high grade cervical squamous intraepithelial lesions in young women: Population based prospective follow up study. BMJ 325:572. Klein G. 1994. The paradoxical coexistence of EBV and the human species. EpsteinBarr Virus Rep 1:5–9. Klein G, Giovanella BC, Lindahl T, et al. 1974. Direct evidence for the presence of Epstein-Barr virus DNA and nuclear antigen in malignant epithelial cells from patients with poorly differentiated carcinoma of the nasopharynx. Proc Natl Acad Sci USA 71:4737–4741. Klein G, Powers A, Croce C. 2002. Meeting Report: Association of SV40 with human tumors. Oncogene 21:1141–1149. Ko Y-C, Li S-C, Yen Y-Y, Yeh S-M, Hsieh C-C. 1991. Horizontal transmission of hepatitis B virus from siblings and intramuscular injection among preschool children in a familial cohort. Am J Epidemiol 133:1015–1023. Komatsu F, Takahashi K. 1999. Determination of serum hepatitis C (HCV) core protein using a novel approach for quantitative evaluation of HCV viremia in anti-HCV-positive patients. Liver 19:375–380. Kondo T, Kono H, Miyamoto N, et al. 1989. Age- and sex-specific cumulative rate and risk of ATL for HTLV-I carriers. Int J Cancer 43:1061–1064. Kondo T, Kono H, Nonaka H, et al. 1987. Risk of adult T-cell leukemia/ lymphoma in HTLV-I carriers. Lancet ii:159. Koshy R, Wells J. 1991. Deregulation of cellular gene expression by HBV transactivators in hepatocarcinogenesis. In: Tabor E, Di Bisceglie AM, Purcell RH, eds. Etiology, Pathology, and Treatment of Hepatocellular Carcinoma in North America. Volume 13. Houston: Advances in Applied Biotechnology Series, pp. 159–170. Koss LG. 1989. The Papanicolaou test for cervical cancer detection. A triumph and a tragedy. JAMA 261:737–743. Koutsky LA, Ault KA, Wheeler CM, et al. 2002. A controlled trial of a human papillomavirus type 16 vaccine. N Engl J Med 347:1645–1651. Koziel MJ. 1997. The role of immune responses in the pathogenesis of hepatitis C virus infection. J Viral Hepatitis 4(Suppl 2):31–41. Koziel MJ. 1999. Cytokines in viral hepatitis. Sem Liver Dis 19:157–169. Krampera M, Vinante F, Tavecchia L, et al. 1999. Progressive polarization towards a T helper/cytotoxic type-I cytokine pattern during age-dependent maturation of the immune response inversely correlates with CD30 cell expression and serum concentration. Clin Exper Immunol 117:291–297. Kreinitz W. 1906. Ueber das Auftreten von Spirochaeten verschiedener Form im Mageninhalt bei Carcinoma ventriculi. Deutsche Medizinische Wochenschrift 32:872. Kuo G, Choo QL, Alter HJ, et al. 1989. An assay for circulating antibodies to a major etiologic virus of human non-A, non-B hepatitis. Science 244:362–364. Kuper H, Hsieh CC, Stuver SO, et al. 2000c. Birth order, as a proxy for age at infection, in the etiology of hepatocellular carcinoma. Epidemiology 11:680–683. Kuper HE, Tzonou A, Kaklamani E, et al. 2000a. Hepatitis B and C viruses in the etiology of hepatocellular carcinoma; a study in Greece using third-generation assays. Cancer Causes and Control 11:171–175. Kuper HE, Tzonou A, Kaklamani E, et al. 2000b. Tobacco smoking, alcohol consumption and their interaction in the causation of hepatocellulsr carcinoma. Int J Cancer 85:498–502. Kusuhara K, Sonoda S, Takahashi K, et al. 1987. Mother-to-child transmission of human T-cell leukemia virus type I (HTLV–I): A fifteen year follow–up study in Okinawa, Japan. Int J Cancer 40:755–757.

541

Kuzushita N, Hayashi N, Moribe T, et al. 1998. Influence of HLA haplotypes on the clinical course of individuals infected with hepatitis C virus. Hepatology 27:240–244. Labrecque LG, Barnes DM, Fentiman IS, Griffin BE. 1995. Epstein-Barr virus in epithelial cell tumors: A breast cancer study. Cancer Res 55:39–45. LaGrenade L, Hanchard B, Fletcher V, Cranston B, Blattner W. 1990. Infective dermatitis of Jamaican children: A marker for HTLV-1 infection. Lancet 336:1345–1347. Lancaster WD, Olson C. 1982. Animal papillomaviruses. Microbiol Rev 46:191–207. Lauer GM, Walker BD. 2001. Hepatitis C virus infection. N Engl J Med 345:41–52. Lauren P. 1965. The two histological main types of gastric cancer: Diffuse and so-called intestinal type carcinoma. Acta Pathol Microbiol Scand 64:31–49. Lavanchy D, Mcmahon B. 2000. Worldwide prevalence and prevention of hepatitis C. In: Liang TJ, Hoofnagle JH, eds. Biomedical Research Reports, Hepatitis C. San Diego: Academic Press, pp. 185–201. Lechmann M, Ihlenfeldt HG, Braunschweiger I, et al. 1996. T and B cell responses to different hepatitis C virus antigens in patients with chronic hepatitis and healthy seropositives. Hepatology 24:790–795. Lechowicz MJ, Lin L, Ambinder RF. 2002. Epstein-Barr virus DNA in body fluids. Current Opin Oncol 14:533–537. Leder P. 1985a. Chromosomal dislocations among antibody genes in human cancer. In: Lenoir GM, O’Conor GM, Olweny CLM, eds. Burkitt’s Lymphoma: A Human Cancer Model. IARC Sci Publ No. 60. Lyon: International Agency for Research on Cancer, pp. 341–358. Leder P. 1985b. The state and prospect for molecular genetics in Burkitt’s lymphoma. In: Lenoir GM, O’Conor GM, Olweny CLM, eds. Burkitt’s Lymphoma: A Human Cancer Model. IARC Sci Publ No. 60. Lyon: International Agency for Research on Cancer, pp. 465–468. Lee WM. 1997. Hepatitis B virus infection. New Engl J Med 337:1733–1745. Lee YI, Lee S, Lee Y, et al. 1998. The human hepatitis B virus trans-activator X gene product regulates Sp1 mediated transcription of an insulin-like growth factor II promoter 4. Oncogene 16:2367–2380. Lennette ET, Blackbourn DJ, Levy JA. 1996. Antibodies to human herpesvirus type 8 in the general population and in Kaposi’s sarcoma patients. Lancet 348:858–861. Lennette ET, Rymo L, Yadav M, et al. 1993. Disease-related differences in antibody patterns against EBV-encoded nuclear antigens EBNA 1, EBNA 2, and EBNA 6. Eur J Cancer 29A:1584–1589. Ley C, Mohar A, Guarner J, et al. 2004. Helicobacter pylori eradication and gastric preneoplastic conditions; a randomized, double-blind, placebo-controlled trial. Cancer Epidemiol Biomarkers and Prev 13:4–10. Li L, Shen MH, Li YQ, et al. 1986. A study of mother-child transmission of hepatitis B virus. In: Proc. Update of the 1986 Shanghai International Symposium on Liver Cancer and Hepatitis, January 15–17, 1986. Liaw KL, Hildesheim A, Burk RD, et al. 2001. A prospective study of human papillomavirus (HPV) type 16 DNA detection by polymerase chain reaction and its association with acquisition and persistence of other HPV types. J Infect Dis 183:8–15. Lindenbach BD, Evans MJ, Syder AJ, et al. 2005. Complete replication of hepatitis C virus in cell culture. Science 309:623–626. Ling PD, Lednicky JA, Keitel WA, et al. 2003. The dynamics of herpesvirus and polyomavirus reactivation and shedding in healthy adults: A 14–month longitudinal study. J Infect Dis 187:1571–1580. List AF, Greco FA, Vogler LB. 1987. Lymphoproliferative diseases in immunocompromised hosts: The role of Epstein-Barr virus. J Clin Oncol 5:1673–1689. Liu H, Leung P, Glynn S, Murphy EL. 2001. Human T-lymphotrophic virus type II RFLP subtypes aO and b4/b5 are associated with different demographic and geographic characteristics in the United States. Virology 279:90–96. Lo YM, Leung SF, Chan LY, et al. 2000. Kinetics of plasma Epstein-Barr virus DNA during radiation therapy for nasopharyngeal carcinoma. Cancer Res 60:2351–2355. Lohmann V, Korner F, Koch J-O, Herian U, Theilmann L, Bartenschlager R. 1999. Replication of subgenomic hepatitis C virus RNAs in a hepatoma cell line. Science 285:110–113. London WT. 1981. Primary hepatocellular carcinoma: Etiology, pathogenesis and prevention. Human Pathol 12:1085–1097. London WT, Buetow K. 1988. Controversies in basic science: Hepatitis B virus and primary hepatocellular carcinoma. Cancer Invest 6:317–326. Lorincz AT, Castle PE, Sherman ME, et al. 2002. Viral load of human papillomavirus and risk of CIN3 or cervical cancer. Lancet 360:228–229. Lorincz AT, Reid R, Jenson AB, Greenberg MD, Lancaster W, Kurman RJ. 1992. Human papillomavirus infection of the cervix: Relative risk associations of 15 common anogenital types. Obstet Gynecol 79:328–337. Lowy DR, Frazer IH. 2003. Chapter 16: Prophylactic human papillomavirus vaccines. J Natl Cancer Inst Monogr 31:111–116. Lucas KG, Filo F, Heilman DK, Lee CH, Emanuel DJ. 1998. Semiquantitative Epstein-Barr virus polymerase chain reaction analysis of peripheral blood from organ transplant patients and risk for the development of lymphoproliferative disease. Blood 92:3977–3978. Lucey DR, Clerici M, Shearer GH. 1996. Type 1 and type 2 cytokine dysregulation in human infectious, neoplastic, and inflammatory diseases. Clin Microbiology Reviews 9:532–562.

542

PART III: THE CAUSES OF CANCER

Luppi M, Torelli G. 1996. The new lymphotropic herpesviruses (HHV-6, HHV-7, HHV-8) and hepatitis C virus (HCV) in human lymphoproliferative diseases: An overview. Haematologica 81:265–281. Mahieux R, Horal P, Mauclère P, et al. 2000. Human T-cell lymphotropic virus type 1 Gag indeterminant western blot patterns in central Africa: Relationship to Plasmodium falciparum infection. J Clin Microbiol 38:4049–4057. Malaty HM, Graham DY. 1994. Importance of childhood socioeconomic status on the current prevalence of Helicobacter pylori infection. Gut 35:742–745. Malaty HM, Engstrand L, Pedersen NL, Graham DY. 1994. Helicobacter pylori infection: genetic and environmental influences. Ann Intern Med 120: 982–986. Malfertheiner P, Megraud F, O’Morain C, et al. 2002. Current concepts in the management of Helicobacter pylori infection–The Maastricht 2–2000 Consensus Report. Aliment Pharmacol Ther 16:167–180. Mangia A, Gentile R, Cascavilla I, et al. 1999. HLA class II favors clearance of HCV infection and progression of the chronic liver damage. J Hepatol 30:984–989. Mangia A, Villani MR, Minerva N, et al. 2001. Efficacy of 5 MU of interferon in combination with ribavirin for naïve patients with chronic hepatitis C virus: A randomized controlled trial. J Hepatol 34:441–446. Manhart LE, Koutsky LA. 2002. Do condoms prevent genital HPV infection, external genital warts, or cervical neoplasia? A meta-analysis. Sex Transm Dis 29:725–735. Manns A, Hisada M, Lagrenade L. 1999a. Human T-lymphotropic virus type I infection. Lancet 353:1951–1958. Manns A, Miley WJ, Wilks RJ, et al. 1999b. Quantitative proviral DNA and antibody levels in the natural history of HTLV-I infection. J Infect Dis 180:1487–1493. Manns A, Murphy EL, Wilks RJ, et al. 1991. Early antibody profile during HTLVI seroconversion. Lancet 337:181–182. Manns A, Wilks RJ, Murphy EL, et al. 1992. A prospective study of transmission by transfusion of HTLV-I and risk factors associated with seroconversion. Int J Cancer 51:886–891. Mansky LM. 2001. Sorting out mutations in human T-cell leukemia virus type I proviruses during in vivo clonal expansion. J Natl Cancer Inst 93:336–337. Marshall BJ, Warren JR. 1984. Unidentified curved bacilli in the stomach of patients with gastritis and peptic ulceration. Lancet 1:1311–1314. Marshall BJ, Armstrong JA, Mcgeche DB, Glancy RJ. 1985. Attempt to fulfill Koch’s postulates for pyloric Campylobacter. Med J Aust 142:436–439. Marshall BJ, Goodwin CS, Warren JR, et al. 1988. Prospective double-blind trial of duodenal ulcer relapse after eradication of Campylobacter pylori. Lancet 2:1439–1442. Martinez FD, Holt PG. 1999. Role of microbial burden in aetiology of allergy and asthma. Lancet 354(Suppl 2):SII 12–15. Martinson FEA, Weigle KA, Royce RA, Weber DJ, Suchindran CM, Lemon SM. 1998. Risk factors for horizontal transmission of hepatitis B virus in a rural district in Ghana. Am J Epidemiol 147:478–487. Marusawa H, Hijikata M, Chiba T, Shimotohno K. 1999a. Hepatitis C virus core protein inhibits Fas- and tumor necrosis factor alpha-mediated apoptosis via NF-kB activation. J Virol 73:4713–4720. Marusawa H, Osaki Y, Kimura T, et al. 1999b. High prevalence of antihepatitis B virus serological markers in patients with hepatitis C virus related chronic liver disease in Japan. Gut 45:284–288. Mason AL, Lau JY, Hoang N, et al. 1999. Association of diabetes mellitus and chronic hepatitis C virus infection. Hepatology 29:328–333. Mason J, Axon AT, Forman D, et al. 2002. The cost-effectiveness of population Helicobacter pylori screening and treatment: A Markov model using economic data from a randomized controlled trial. Aliment Pharmacol Ther 16:559–568. Mathew A, Cheng H-M, Sam C-K, et al. 1994. A high incidence of serum IgG antibodies to the Epstein-Barr virus replication activator protein in nasopharyngeal carcinoma. Cancer Immunol Immunother 38:68–70. Matsumoto C, Mitsunaga S, Oguchi T, et al. 1990. Detection of human T-cell leukemia virus type I (HTLV-I) provirus in an infected cell line and in peripheral mononuclear cells of blood donors by the nested double polymerase chain reaction method: Comparison with HTLV-I antibody tests. J Virol 64:5290–5294. Matsuoka M. 2003 Human T-cell leukemia virus type I and adult T-cell leukemia. Oncogene 5131–5140. Matutes E, Dalgleish AG, Weiss RA, et al. 1986. Studies in healthy human T-cell leukemia lymphoma virus (HTLV-I) carriers from the Caribbean. Int J Cancer 38:41–45. Maupas P, Melnick JL. 1981. Hepatitis B infection and primary liver cancer. Prog Med Virol 27:1–5. Mayama S, Cuevas LE, Sheldon J, et al. 1998. Prevalence and transmission of Kaposi’s sarcoma-associated herpesvirus (human herpesvirus 8) in Ugandan children and adolescents. Int J Cancer 77:817–820. Mazzella G, Accogli E, Sottili S, et al. 1996. Alpha interferon treatment may prevent hepatocellular carcinoma in HCV-related liver cirrhosis. J Hepatol 24:141–147. Mbulaiteye SM, Biggar RJ, Pfeiffer RM, et al. 2005. Water, socioeconomic factors, and human herpesvirus 8 infection in Ugandan children and their mothers. J Acquir Immune Defic Syndr 38:474–479.

McGuinness PH, Bishop GA, Painter DM, Chan R, McCaughan GW. 1996. Intrahepatic hepatitis C RNA levels do not correlate with degree of liver injury in patients with chronic hepatitis C. Hepatology 23:676–687. McHutchison JG, Gordon SC, Schiff ER, et al. 1998. Interferon alfa-2b alone or in combination with ribavirin as initial treatment for chronic hepatitis C. N Engl J Med 339:1485–1492. McMahon BJ, Alward WLM, Hall DB, et al. 1985. Acute hepatitis B virus infection: Relation of age to the clinical expression of the disease and subsequent development of the carrier state. J Infect Dis 151:599–603. McQuillan GM, Coleman PJ, Kruszon-Moran D, Moyer LA, Lambert SB, Margolis HS. 1999. Prevalence of hepatitis B virus infection in the United States: The National Health and Nutrition Examination Surveys, 1976 through 1994. Am J Pub Health 89:14–18. Mehta SH, Strathdee SA, Thomas DL. 2001. Association between hepatitis C virus infection and diabetes mellitus. Epidemiol Rev 23:302–312. Meisel H, Reip A, Faltus B, et al. 1995. Transmission of hepatitis C virus to children and husbands by women infected with contaminated anti-D immunoglobulin. Lancet 345:1209–1211. Meisels A. 1969. Dysplasia and carcinoma of the uterine cervix. IV. A correlated cytologic and histologic study with special emphasis on vaginal microbiology. Acta Cytol 13:224–231. Melbye M, Ebbeson P, Levine PH, Bennike T. 1984. Early primary infection and high Epstein-Barr virus antibody titers in Greenland Eskimos at high risk for nasopharyngeal carcinoma. Int J Cancer 34:619. Mercer DF, Schiller DE, Elliott JF, et al. 2001. Hepatitis C virus replication in mice with chimeric human livers. Nat Med 7:927–933. Meu P, Leen A, Rickinson AB, et al. 2002. Identification and prevalence of CB8+ T-cell responses directed against Epstein-Barr virus-encoded latent membrane 1 and latent membrane protein 2. Int J Cancer 99:93–99. Meuleman P, Libbrecht L, De Vos R, et al. 2005. Morphological and biochemical characterization of a human liver in a uPA-SCID mouse chimers. Hepatology 41:847–856. Miller G. 1974. Oncogenicity of Epstein-Barr virus. J Infect Dis 130:187–205. Miller G. 1990. Epstein-Barr virus: Biology, pathogenesis and medical aspects. In: Fields B, Knipe DM, eds. Virology. Second edition. New York: Lippincott– Raven, pp. 1921–1958. Miller G, Katz BZ, Niederman JC. 1987. Some recent developments in the molecular epidemiology of Epstein-Barr virus infections. Yale J Biol Med 60:307– 316. Miller G, Shope T, Cooper D, et al. 1977. Lymphoma in cotton top marmosets after inoculation with Epstein-Barr virus: Tumor incidence, histologic spectrum, antibody responses, demonstration of viral DNA, and characterization of virus. J Exp Med 145:948–967. Monini P, Colombini S, Stürzl M, et al. 1999. Reactivation and persistence of human herpesvirus 8 infection in B-cells and monocytes by th-1 cytokines increased in Kaposi’s sarcoma. Blood 93:4044–4058. Montella M, Crispo A, Frigeri F, et al. 2001. HCV and tumors correlated with immune system: A case-control study in an area of hyperendemicity. Leuk Res 25:775–781. Montesano R, Hainaut P, Wild CP. 1997. Hepatocellular carcinoma: From gene to public health. J Natl Cancer Inst 89:1844–1851. Moore PS, Chang Y. 1995. Detection of herpesvirus-like DNA sequences in Kaposi’s sarcoma in patients with and those without HIV infection. N Engl J Med 332:1181–1185. Moore PS, Chang Y. 1998. Kaposi’s sarcoma (KS), KS-associated herpesvirus, and the criteria for causality in the age of molecular biology. Am J Epidemiol 147:217–221. Moore PS, Chang Y. 2003. Kaposi’s sarcoma-associated herpesvirus immunoevasion and tumorigenesis: two sides of the same coin? Ann Rev Microbiol 57:609–639. Moore PS, Gao SJ, Dominguez G, et al. 1996. Primary characterization of a herpesvirus agent associated with Kaposi’s sarcoma. J Virol 70:549–558. Morand-Joubert L, Mariotti M, Reed D, Petit J-C, Lefrere J-J. 1995. Correlation between viral DNA load and serum anti p19 antibody concentration in symptomless human T-lymphotropic virus type-I (HTLV-I)-infected individuals. Int J Cancer 60:156–159. Moreno V, Bosch FX, Munoz N, et al. 2002. Effect of oral contraceptives on risk of cervical cancer in women with human papillomavirus infection: The IARC Multicentric Case-Control Study. Lancet 359:1085–1092. Mori M, Hara M, Wada I, et al. 2000. Prospective study of hepatitis B and C viral infections, cigarette smoking, alcohol consumption, and other factors associated with hepatocellular carcinoma risk in Japan. Am J Epidemiol 151:131– 139. Moriya K, Fujie H, Shintani Y, et al. 1998. The core protein of hepatitis C virus induces hepatocellular carcinoma in transgenic mice. Nature Med 4:1065–1067. Mork J, Lie AK, Glattre E, et al. 2001. Human papillomavirus infection as a risk factor for squamous-cell carcinoma of the head and neck. N Engl J Med 344:1125–1131. Morris A, Nicholson G. 1987. Ingestion of Campylobacter pyloridis causes gastritis and raised fasting gastric pH. Am J Gastroenterol 82:192–199. Mortreux F, Gabet A-S, Wattel E. 2003. Molecular and cellular aspects of HTLV-1 associated leukemogenesis in vivo. Leuk 17:26–38.

Infectious Agents Mortreux F, Leclercq I, Gabet AS, et al. 2001. Somatic mutation in human T-cell leukemia virus type I provirus and flanking cellular sequences during clonal expansion in vivo. J Natl Cancer Inst 93:367–377. Moses AV, Fish KN, Ruhl R, et al. 1999. Long–term infection and transformation of dermal microvascular endothelial cells by human herpesvirus 8. J Virol 73:6892–6902. Mosmann TR, Moore KW. 1991. The role of IL-10 in crossregulation of TH1 and TH1 responses. Immunol Today 12:149–153. Mueller N. 1987. Epidemiological studies assessing the role of Epstein-Barr virus in Hodgkin’s disease. Yale J Biol Med 60:321–327. Mueller N. 1991. The epidemiology of HTLV-I infection. Cancer Causes Control 2:37–52. Mueller N. 1996. Hodgkin’s disease. In: Schottenfeld D, Fraumeni JF, Jr., eds. Cancer Epidemiology and Prevention. Second edition. New York: Oxford University Press, pp. 893–919. Mueller N. 1999. Overview of the epidemiology of malignancy in immunedeficiency. JAIDS 21(Suppl 1):S5–S10. Mueller NE, Blattner WA. 1997. Retroviruses – human T-cell lymphotropic virus. In: Evans AS, Kaslow RA, eds. Viral Infections of Humans. Fourth edition. New York: Plenum Publishing, pp. 785–814. Mueller N, Evans A, Harris NL, et al. 1989. Hodgkin’s disease and Epstein-Barr virus: Altered antibody pattern before diagnosis. New Engl J Med 320:696–701. Mueller N, Mohar A, Evans A, et al. 1991. Epstein-Barr virus antibody patterns preceeding the diagnosis of non-Hodgkin’s lymphoma. Int J Cancer 49:387–393. Mueller N, Tachibana N, Stuver S, et al. 1990. Epidemiologic perspectives of HTLV-I. In: Blattner WA, ed. Human Retrovirology: HTLV. New York: Lippincott–Raven, pp. 281–293. Munger K. 2002. The role of human papillomaviruses in human cancers. Front Biosci 7:d641–d649. Munger K, Howley PM. 2002. Human papillomavirus immortalization and transformation functions. Virus Res 89:213–228. Munoz N, Bosch FX. 1987. Epidemiology of hepatocellular carcinoma. In: Okuda K, Ishak KG, eds. Neoplasms of the Liver. Tokyo: Springer, pp. 3–19. Munoz N, Bosch FX, De Sanjose S, et al. 2003. Epidemiologic classification of human papillomavirus types associated with cervical cancer. N Engl J Med 348:518–527. Munoz N, Franceschi S, Bosetti C, et al. 2002. Role of parity and human papillomavirus in cervical cancer: The IARC Multicentric Case-Control Study. Lancet 359:1093–1101. Munoz N, Lingao A, Lao J, et al. 1989. Patterns of familial transmission of HBV and the risk of developing liver cancer: A case-control study in the Philippines. Int J Cancer 44:981–984. Murai K, Tachibana N, Shioiri S, et al. 1990. Suppression of delayed-type hypersensitivity to PPD and PHA in elderly HTLV-I carriers. JAIDS 3:1006–1009. Murphy EL, Bryzman SM, Glynn SA, et al. 2000. Risk factors for hepatitis C virus infection in United States blood donors. Hepatology 31:756–762. Murphy EL, Bryzman S, Williams AE, et al. 1996. Demographic determinants of hepatitis C virus seroprevalence among blood donors. JAMA 275:995–1000. Murphy EL, Figueroa JP, Gibbs WN, et al. 1991. Human T-lymphotropic virus type I (HTLV-I) seroprevalence in Jamaica. I. Demographic determinants. Am J Epidemiol 133:1114–1124. Murphy EL, Hanchard B, Figueroa JP, et al. 1989. Modeling the risk of adult T-cell leukemia/lymphoma in persons infected with human T-lymphotropic virus type I. Int J Cancer 43:250–253. Mutirangura A. 2001. Serum/plasma viral DNA: Mechanisms and diagnostic applications to nasopharyngeal and cervical cancer. Ann NY Acad Sci 945: 59–67. Mutirangura A, Pomthanakasem W, Theamboonlers A, et al. 1998. Epstein-Barr viral DNA in serum of patients with nasopharyngeal carcinoma. Clin Cancer Res 4:665–669. Nagai M, Kubota R, Greten TF, Schneck JP, Leist TP, Jacobson S. 2001. Increased activated human T cell lymphotropic virus type I (HTLV-I) Tax 11–19specific memory and effector CD8+ cells in patients with HTLV-I-associated myelopathy/tropical spastic paraparesis: Correlation with HTLV-I provirus load. J Infect Dis 183:197–205. Nagasako T, Sugiyama T, Mizushima T, Miura Y, Kato M, Asaka M. 2003. Upregulated Smad5 mediates apoptosis of gastric epithelial cells induced by Helicobacter pylori infection. J Biol Chem 278:4821–4825. Nakashima K, Ikematsu H, Hayashi J, Kishihara Y, Mitsutake A, Kashiwagi S. 1995. Intrafamilial transmission of hepatitis C virus among the population of an endemic area of Japan. JAMA 274:1459–1461. Nawar E, Mbulaiteye SM, Gallant JE, et al. 2005. Risk factors for Kaposi’s sarcoma among HHV8-seropositive homosexual men with AIDS. INT J Cancer 115:296–300. Neumann AU, Lam NP, Dahari H, et al. 1998. Hepatitis C viral dynamics in vivo and the antiviral efficacy of interferon-alpha therapy. Science 282:103–107. Neva FA, Murphy EL, Gam A, et al. 1989. Antibodies to Strongyloides stercoralis in healthy Jamaican carriers of HTLV-I. N Engl J Med 320:252–253. Niederau C, Heintges T, Lange S, et al. 1996. Long-term follow–up of HBcAg positive patients treated with interferon alpha for chronic hepatitis B. New Engl J Med 334:1422–1427.

543

Niederman JC, Evans AS. 1997. Epstein-Barr virus. In: Evans AS, Kaslow RA, eds. Viral Infections of Humans: Epidemiology and Control. Fourth edition. New York: Plenum Press, pp. 253–284. Niederman JC, Evans AS, Subrahmanjan L, Mccollum RV. 1970. Prevalence, incidence, and persistence of EBV antibody in young adults. N Engl J Med 282:361–365. Niedobitek G, Herst H, Young LS, et al. 1992. Epstein-Barr virus and carcinomas: Expression of the viral genome in an undifferentiated gastric carcinoma. Diagn Mol Pathol 1:103–108. Niedobitek G, Kremmer E, Herbst H, et al. 1997. Immunohistochemical detection of the Epstein-Barr virus-encoded latent membrane protein 2A in Hodgkin’s disease and infectious mononucleosis. Blood 90:1664–1672. NIH Consensus Development Panel On Helicobacter Pylori in Peptic Ulcer Disease. 1994. Helicobacter pylori in peptic ulcer disease. JAMA 272:65–69. Nishimura M, Kermode AG, Clerici M, et al. 1994. Demonstration of human T lymphotropic virus type I (HTLV-I)-specific cell responses from seronegative and polymerase chain reaction-negative persons exposed to HTLV-I. J Infect Dis 170:334–338. Nobbenhuis MA, Walboomers JM, Helmerhorst TJ, et al. 1999. Relation of human papillomavirus status to cervical lesions and consequences for cervical-cancer screening: A prospective study. Lancet 354:20–25. Noguchi S, Sata M, Suzuki H, Mizokami M, Tanikawa K. 1997. Routes of transmission of hepatitis C virus in an endemic rural area of Japan. Molecular epidemiologic study of hepatitis C virus infection. Scand J Infect Dis 29:23–28. Ohsawa M, Shingu N, Miwa H, et al. 1999. Risk of non-Hodgkin’s lymphoma in patients with hepatitis C virus infection. Int J Cancer 80:237–239. Ohto H, Terazawa S, Sasaki N, et al. 1994. Transmission of hepatitis C virus from mothers to infants. N Engl J Med 330:744–750. Okayama A, Stuver S, Iga M, et al. 2001. Sequential change of virus markers in seroconverters with community-acquired infection of human T lymphotropic virus type 1. J Infect Dis 183:1031–1037. Okayama A, Stuver S, Matsuoka M, et al. 2004. The role of HTLV-I proviral DNA load and clonality in the development of adult T-cell leukemia/lymphoma in asymptomatic carriers. Intl J Cancer 110:621–625. Okayama A, Stuver SO, Tabor E, et al. 2002. Incident hepatitis C virus infection in a community–based population in Japan. J Viral Hepatitis 9:43–51. Okayama A, Tachibana N, Ishihara S, et al. 1997. Increased expression of interleukin-2 receptor a on peripheral blood mononuclear cells in HTLV-I tax/rex mRNA-positive asymptomatic carriers. JAIDS 15:70–75. Okuda K. 1997. Hepatitis C virus and hepatocellular carcinoma. In: Okuda K, Tabor E, eds. Liver Cancer. New York: Churchill Livingstone, pp. 39–50. Okuda K, Nakashima T, Sakamoto K, et al. 1982. Hepatocellular carcinoma arising in noncirrhotic and highly cirrhotic livers: A comparative study of histopathology and frequency of hepatitis B markers. Cancer 49:450–455. Oon CJ. 1992. Long-term survival following treatment of hepatocellular carcinoma in Singapore: Evaluation of Wellferon in the prophylaxis of highrisk pre-cancerous conditions. Cancer Chemother Pharmacol 31(Suppl): S137–S142. Orenstein JM, Alkan S, Blauvelt A, et al. 1997. Visualization of human herpesvirus type 8 in Kaposi’s sarcoma by light and transmission electron microscopy. AIDS 11:F35–F45. Orland JR, Wang B, Nass C, et al. 2003. Mortality in a cohort of HTLV-I and HTLVII infected blood donors. Transfusion 43(Suppl 9):41A. Osman M, Kubo T, Gill J, et al. 1999. Identification of human herpesvirus 8-specific cytotoxic T-cell responses. J Virol 73:6136–6140. Pagano JS, Huang CH, Klein G, et al. 1975. Homology of Epstein-Barr viral DNA in nasopharyngeal carcinoma from Kenya, Taiwan, Singapore and Tunis. In: de The G, Epstein MA, zur Hausen, eds. Oncogenesis and Herpesviruses II. Lyon: International Agency for Research on Cancer, pp. 191–193. Palefsky JM, Holly EA. 2003. Chapter 6: Immunosuppression and co-infection with HIV. J Natl Cancer Inst Monogr 31:41–46. Palefsky JM, Holly EA, Ralston ML, Jay N. 1998. Prevalence and risk factors for human papillomavirus infection of the anal canal in human immunodeficiency virus (HIV)-positive and (HIV)-negative homosexual men. J Infect Dis 177:361–367. Pallesen G, Hamilton–Dutoit SJ, Rowe M, et al. 1991. Expression of Epstein-Barr virus latent gene products in tumour cells of Hodgkin’s disease. Lancet 337:320–322. Park TW, Riethdorf S, Schulz G, Riethdorf L, Wright T, Loning T. 2003. Clonal expansion and HPV-induced immortalization are early molecular alterations in cervical carcinogenesis. Anticancer Res 23:155–160. Parkin DM, Muir CS, Whelan SL, Gao YT, Ferlay J, Powell J, eds. 1992. Cancer Incidence in Five Continents, Volume VI. IARC Sci Publ 120. Lyon: International Agency for Research on Cancer, pp. 205, 438, 532. Parkin DM, Ohshima H, Srivatanakul P, Vatanasapt V. 1993. Cholangiocarcinoma: epidemiology, mechanisms of carcinogenesis and prevention. Cancer Epidemiol Biomarkers Prev 2:537–544. Parkin DM, Pisani P, Munoz N, Ferlay J. 1999. The global health burden of infection associated cancers. In: Newton R, Beral V, Weiss RA, eds. Cancer Surveys. Volume 33: Infections and Human Cancer. Plainview, N.Y.: Cold Spring Harbor Laboratory Press, pp. 5–33.

544

PART III: THE CAUSES OF CANCER

Parkin DM, Srivatanakul P, Khlat M, et al. 1991. Liver cancer in Thailand. I. A case-control study of cholangiocarcinoma. Int J Cancer 48:323–328. Parks WP, Lenes BA, Tomasulo PA, et al. 1991. Human T-cell lymphotropic virus infection among blood donors in South Florida. JAIDS 4:89–96. Parravicini C, Olsen SJ, Capra M, et al. 1997. Risk of Kaposi’s sarcomaassociated herpesvirus transmission from donor allografts among Italian posttransplant Kaposi’s sarcoma patients. Blood 90:2826–2829. Parsonnet J, Blaser MJ, Perez-perez GI, Hargrett-bean N, Tauxe RV. 1992. Symptoms and risk factors of Helicobacter pylori infection in a cohort of epidemiologists. Gastroenterology 102:41–46. Parsonnet J, Friedman GD, Orentreich N, Vogelman H. 1997. Risk for gastric cancer in people with CagA positive or CagA negative Helicobacter pylori infection. Gut 40:297–301. Parsonnet J, Hansen S, Rodriguez L, et al. 1994. Helicobacter pylori infection and gastric lymphoma. N Engl J Med 330:1267–1271. Parsonnet J, Harris RA, Hack HM, Owens DK. 1996. Modeling cost-effectiveness of Helicobacter pylori screening to prevent gastric cancer: A mandate for clinical trials. Lancet 348:150–154. Pastrana DV, Buck CB, Pang Y-YS, et al. 2004. Reactivity of human sera in a sensitive, high-throughput pseudovirus-based papillomavirus neutralization assays for HPV16 and HPV18. Virology 321:205–216. Pathmanathan R, Prasad U, Sadler R, Flynn K, Raab-Traub N. 1995. Clonal proliferations of cells infected with Epstein-Barr virus in preinvasive lesions related to nasopharyngeal carcinoma. N Engl J Med 333:693–698. Pawlotsky JM. 2002. Use and interpretation of virological tests for hepatitis C. Hepatology 36(Suppl 1):S65–S73. Peek RM Jr, Moss SF, Tham KT, et al. 1997. Helicobacter pylori cagA+ strains and dissociation of gastric epithelial cell proliferation from apoptosis. J Natl Cancer Inst 89:863–868. Pellett PE, Wright DJ, Engels EA, et al. 2003. Multicenter comparison of serologic assays and estimation of human herpesvirus 8 seroprevalence among US blood donors. Transfusion 43:1260–1268. Peters MG, Terrault NA. 2002. Alcohol use and hepatitis C. Hepatology 36(Suppl 1):S220–S225. Peterson WL. 2002. Review article: Helicobacter pylori and gastric adenocarcinoma. Aliment Pharmacol Ther 16(Suppl 1):40–46. Pfister H. 2003. Chapter 8: Human papillomavirus and skin cancer. J Natl Cancer Inst Monogr 31:52–56. Pineau P, Tiollais P. 1997. Animal hepadnaviruses as a model of carcinogenesis. In: Okuda K, Tabor E, eds. Liver Cancer. New York: Churchill Livingstone, pp. 189–199. Pinto LA, Edwards J, Castle PE, et al. 2003. Cellular immune responses to human papillomavirus (HPV)-16 L1 in healthy volunteers immunized with recombinant HPV-16 L1 virus-like particles. J Infect Dis 188:327–338. Pisani P, Bray F, Parkin DM. 2002. Estimates of the world-wide prevalence of cancer for 25 sites in the adult population. Int J Cancer 97:72–81. Plancoulaine S, Abel L, Tregouet D, et al. 2004. Respective roles of serological status and blood specific antihuman herpesvirus 8 antibody levels in human herpesvirus 8 intrafamilial transmission in a highly endemic area. Cancer Res 64:8782–8787. Plancoulaine S, Abel L, Van Beveren M, et al. 2000. Human herpesvirus 8 transmission from mother to child and between siblings in an endemic population. Lancet 356:1062–1065. Poiesz BJ, Ruscetti FW, Gazdar AF, et al. 1980. Detection and isolation of type C retrovirus particles from fresh and cultured lymphocytes of a patient with cutaneous T-cell lymphoma. Proc Natl Acad Sci USA 77:7415– 7419. Popper H. 1988. Relation between hepatocellular carcinoma and persistent hepatitis B infection. Appl Pathol 6:64–72. Porto AF, Neva FA, Bittencourt H, et al. 2001. HTLV-1 decreases Th2 type of immune response in patients with strongyloidiasis. Parasite Immunol 23:503–507. Poynard T, Bedossa P, Opolon P. 1997. Natural history of liver fibrosis progression in patients with chronic hepatitis C. Lancet 349:825–832. Prang NS, Hornef MW, Jager M, Wagner HJ, Wolf H, Schwarzmann FM. 1997. Lytic replication of Epstein-Barr virus in the peripheral blood: Analysis of viral gene expression in B lymphocytes during infectious mononucleosis and in the normal carrier state. Blood 89:1665–1677. Prince AM. 1994. Challenges for development of hepatitis C virus vaccines. FEMS Microbiol Rev 14:273–277. Purtilo DT, Sakamotok K, Barnabei V, et al. 1982. Epstein-Barr virus-induced diseases in boys with the X-linked lymphoproliferation syndrome (XLP). Am J Med 73:49–56. Qian GS, Ross RK, Yu MC, et al. 1994. A follow-up study of urinary markers of aflatoxin exposure and liver cancer risk in Shanghai, People’s Republic of China. Cancer Epidemiol Biomarkers Prev 3:3–10. Raab–Traub N, Flynn K. 1986. The structure of the termini of the EpsteinBarr virus as a marker of clonal cellular proliferation. Cell 47:883–889. Rabkin CS, Schulz TF, Whitby D, et al. 1998. Interassay correlation of human herpesvirus 8 serologic tests. J Infect Dis 178:304–309. Rall CJN, Dienstag JL. 1995. Epidemiology of hepatitis C virus infection. Sem Gastrointestinal Dis 6:3–12.

Ray RB, Lagging LM, Meyer K, Ray R. 1996a. Hepatitis C virus core protein cooperates with ras and transforms primary rat embryo fibroblasts to tumorigenic phenotype. J Virol 70:4438–4443. Ray RB, Meyer K, Ray R. 1996b. Suppression of apoptotic cell death by hepatitis C virus core protein. Virology 226:176–182. Ray RB, Steele R, Meyer K, Ray R. 1997. Transcriptional repression of p53 promoter by hepatitis C virus core protein. J Biol Chem 272:10983–10986. Reedman BM, Klein G. 1973. Cellular localization of an Epstein-Barr virus (EBV) associated complement-fixing antigen in producer and non-producer lymphoblastoid cell lines. Int J Cancer 11:499–520. Renne R, Langunoff M, Zhong W, Ganem D. 1996a. The size and conformation of Kaposi’s sarcoma-associated herpesvirus (human herpesvirus 8) DNA in infected cells and virions. J Virol 70:8151–8154. Renne R, Zhong W, Herndier B, et al. 1996b. Lytic growth of Kaposi’s sarcomaassociated herpesvirus (human herpesvirus 8) in culture. Nature Med 2:342– 346. Renwick N, Dukers NH, Weverling GJ, et al. 2002. Risk factors for human herpesvirus 8 infection in a cohort of drug users in The Netherlands, 1985–1996. J Infect Dis 185:1808–1812. Renwick N, Halaby T, Weverling GJ, et al. 1998. Seroconversion for human herpesvirus 8 during HIV infection is highly predictive of Kaposi’s sarcoma. AIDS 12:2481–2488. Replogle ML, Glaser SL, Hiatt RA, Parsonnet J. 1995. Biological sex as a risk factor for Helicobacter pylori infection in healthy young adults. Am J Epidemiol 142:856–863. Rettig MB, Ma HJ, Vescio RA, et al. 1997. Kaposi’s sarcoma-associated herpesvirus infection of bone marrow dendritic cells from multiple myeloma patients. Science 276:1851–1854. Rice PS, Cason J, Best JM, Banatvala JE. 1999. High risk genital papillomavirus infections are spread vertically. Rev Med Virol 9:15–21. Richardson H, Kelsall G, Tellier P, et al. 2003. The natural history of typespecific human papillomavirus infections in female university students. Cancer Epidemiol Biomarkers Prev 12:485–490. Richart RM, Barron BA. 1969. A follow-up study of patients with cervical dysplasia. Am J Obstet Gynecol 105:386–393. Rickinson AB, Kieff E. 2001. Epstein-Barr virus. In: Knipe Dm, Howley PM, eds. Fields Virology, Fourth edition. Philadelphia: Lippincott Williams and Wilkins, pp. 2575–2627. Riddler SA, Breinig MC, Mcknight JLC. 1994. Increased levels of circulating Epstein-Barr virus (EBV)-infected lymphocytes and decreased EBV nuclear antigen antibody responses are associated with the development of posttransplant lymphoproliferative disease in solid-organ transplant recipients. Blood 84:972–984. Rivas C, Thlick A-E, Parravicini C, Moore PS, Chang Y. 2001. Kaposi’s sarcomaassociated herpesvirus LANA2 is a B-cell-specific latent viral protein that inhibits p53. J Virol 75:429–438. Robert–Guroff M, Nakao Y, Notake K, et al. 1982. National antibodies to human retrovirus HTLV in a cluster of Japanese patients with adult T-cell leukemia. Science 215:975–978. Roberts EA, Yeung L. 2002. Maternal-infant transmission of hepatitis C virus infection. Hepatology 36(Suppl 1):S106–S113. Rodger AJ, Roberts S, Lanigan A, Bowden S, Brown T, Crofts N. 2000. Assessment of long-term outcomes of community hepatitis C infection in a cohort with sera stored from 1971–1975. Hepatology 32:582–587. Rokkas T, Ladas S, Liatsos C, et al. 1999. Relationship of Helicobacter pylori CagA status to gastric cell proliferation and apoptosis. Digest Dis Sci 44:487–493. Rollino C, Roccatello D, Giachino O, Basolo B, Piccoli G. 1991. Hepatitis C virus infection and membranous glomerulonephritis. Nephron 59:319–320. Romagnani S. 2000. T-cell subsets (Th1 versus Th2). Ann Allergy Asthma Immunol 85:9–18. Rooney CM, Smith CA, Ng CY, et al. 1995. Use of gene-modified virusspecific T lymphocytes to control Epstein-Barr-virus-related lymphoproliferation. Lancet 345:9–13. Roosendaal R, Kuipers EJ, Buitenwerf J, et al. 1997. Helicobacter pylori and the birth cohort effect: Evidence of a continuous decrease of infection rates in childhood. Am J Gastroenterol 92:1480–1482. Rosenblatt JD, Chen ISY, Wachsman W. 1988. Infection with HTLV-I and HTLVII: evolving concepts. Semin Hematol 25:230–246. Rosenblatt JD, Golde DW, Wachsman W, et al. 1986. A second isolate of HTLV-II associated with atypical hairy-cell leukemia. N Engl J Med 315:372–377. Rudolph DA, Coligan JE, Lal RB. 1994. Detection of antibodies to transactivator protein (p40taxI) of human T-cell lymphotropic virus type I by a synthetic peptide-based assay. Clin Diagn Lab Immunol 1:176–181. Rupnow MF, Owens DK, Shachter R, Parsonnet J. 1999. Helicobacter pylori vaccine development and use: A cost-effectiveness analysis using the Institute of Medicine Methodology. Helicobacter 4:272–280. Rupnow MF, Shachter RD, Owens DK, Parsonnet J. 2000. A dynamic transmission model for predicting trends in Helicobacter pylori and associated diseases in the United States. Emerg Infect Dis 6:228–237. Rupnow MFT, Shachter RD, Owens DK, Parsonnet JA. 2001. Vaccine for ulcers and gastric cancer: A rationale for a H. pylori vaccine [Abstract]. The Fourth

Infectious Agents Annual Conference on Vaccine Research. Arlington, VA. Available at http://www.nfid.org/conferences/vaccine01/abstracts/abs9–12.pdf. Russo JJ, Bohenzky RA, Chien M-C, et al. 1996. Nucleotide sequence of the Kaposi’s sarcoma-associated herpesvirus (HHV-8). Proc Natl Acad Sci USA 93:14862–14867. Saito D. 2003. Present state of Japanese intervention trial of H. pylori. [Japanese]. Nippon Rinsho – Jpn J Clin Med 61:50–55. Sakamuro D, Furukawa T, Takegami T. 1995. Hepatitis C virus nonstructural protein NS3 transforms NIH 3T3 cells. J Virol 69:3893–3896. Sanduleanu S, Jonkers D, De Bruine A, Hameeteman W, Stockbrugger RW. 2001. Double gastric infection with Helicobacter pylori and non-Helicobacter pylori bacteria during acid-suppressive therapy: Increase of pro-inflammatory cytokines and development of atrophic gastritis. Aliment Pharmacol Ther 15: 1163–1175. Sarid R, Olsen SJ, Moore PS. 1999. Kaposi’s sarcoma-associated herpesvirus: Epdemiology, virology, and molecular biology. Adv Virus Res 52:139–231. Satoh M, Kiyuna S, Shiroma Y, Toma H, Kokaze A, Sato Y. 2003. Predictive markers for development of strongyloidiasis in patients infected with both Strongyloides stercoralis and HTLV-1. Clin Exp Immunol 133:391–396. Schiffman M, Herrero R, Desalle R, et al. 2005. The carcinogenicity of human papillomavirus types reflects viral evolution. Virology 337:76–84. Schiffman M, Herrero R, Hildesheim A, et al. 2000. HPV DNA testing in cervical cancer screening: Results from women in a high-risk province of Costa Rica. JAMA 283:87–93. Schiffman M, Kjaer SK. 2003. Chapter 2: Natural history of anogenital human papillomavirus infection and neoplasia. J Natl Cancer Inst Monogr 31:14– 19. Schiffman M, Wheeler CM, Dasgupta A, Solomon D, Castle PE, ALTS Group. 2005. A comparison of a prototype PCR assay and hybrid capture 2 for detection of carcinogenic human papillomavirus DNA in women with equivocal or mildly abnormal papanicolaou smears. Am J Clin Pathol 124:1–11 Schiffman MH. 1992a. Recent progress in defining the epidemiology of human papillomavirus infection and cervical neoplasia. J Natl Cancer Inst 84: 394–398. Schiffman MH. 1992b. Validation of hybridization assays: Correlation of filter in situ, dot blot and PCR with Southern blot. In: Munoz N, Bosch FX, Shah KV, Meheus A, eds. The Epidemiology of Cervical Cancer and Human Papillomavirus. IARC Sci Publ NO. 119. Lyon: International Agency for Research on Cancer, pp. 169–179. Schroeder J, Saah AJ, Hoover DR, et al. 1999. Serum soluble CD23 level correlates with subsequent development of AIDS-related non-Hodgkin’s lymphoma. Cancer Epi Biomarkers Prev 8:979–984. Schulz TR, Sheldon J, Greensill J. 2002. Kaposi’s sarcoma associated herpesvirus (KSHV) or human herpesvirus 8 (HHV8). Virus Research 82:115–126. Seeff LB. 2002. Natural history of chronic hepatitis C. Hepatology 36(Suppl 1):S35–S46. Seeff LB, Hollinger FB, Alter HJ, et al. 2001. Long-term mortality and morbidity of transfusion-associated non-A, non-B, and type C hepatitis: A national heart, lung, and blood institute collaborative study. Hepatology 33:455–463. Seeff LB, Miller RN, Rabkin CS, et al. 2000. 45-year follow-up of hepatitis C virus infection in healthy young adults. Ann Internal Med 132:105–111. Seiki M, Eddy R, Shows TB, et al. 1984. Non-specific integration of the HTLV provirus genome into adult T-cell leukemia cells. Nature 309:640– 642. Seiki M, Hattori S, Hirayama V, et al. 1983. Human adult T-cell leukemia virus: Complete nucleotide sequence of the provirus genome integrated in leukemia cell DNA. Proc Natl Acad Sci USA 80:3618–3622. Sellors JW, Karwalajtys TL, Kaczorowski J, et al. 2003. Incidence, clearance and predictors of human papillomavirus infection in women. CMAJ 168:421–425. Shen C-Y, Lee H-S, Huang L-C, Tsai K-S, Chen D-S, Cheng ATA. 1996. Alcoholism, hepatitis B and C viral infections, and impaired liver function among Taiwanese aboriginal groups. Am J Epidemiol 143:936–942. Sherman ME, Lorincz AT, Scott DR, et al. 2003. Baseline cytology, human papillomavirus testing, and risk for cervical neoplasia: A 10-year cohort analysis. J Natl Cancer Inst 95:46–52. Sherman ME, Schiffman MH, Strickler H, Hildesheim A. 1998. Prospects for a prophylactic HPV vaccine: Rationale and future implications for cervical cancer screening. Diagn Cytopathol 18:5–9. Shibata A, Parsonnet J, Longacre TA, et al. 2002. CagA status of Helicobacter pylori infection and p53 gene mutations in gastric adenocarcinoma. Carcinogenesis 23:419–424. Shinzato O, Ikeda S, Saburo M, et al. 1991. Semiquantitative analysis of integrated genomes of human T-lymphotropic virus type I in asymptomatic virus carriers. Blood 78:2082–2088. Shinzato O, Kamihira S, Ikeda S, et al. 1993. Relationship between the anti-HTLV1 antibody level, the number of abnormal lymphocytes and the viral-genome dose in HTLV-1-infected individuals. Int J Cancer 54:208–212. Shioiri S, Stuver SO, Okayama A, et al. 1997. Intrafamilial transmission of HTLVI and its association with anti-tax antibody in an endemic population in Japan. Int J Cancer 75:15–18. Shioiri S, Tachibana N, Okayama A, et al. 1993. Analysis of anti-tax antibody of HTLV-I carriers in an endemic area in Japan. Int J Cancer 53:1–4.

545

Shirakawa T, Enomoto T, Shimazu, Hopkin JM. 1997. The inverse association between tuberculin responses and atopic disorder. Science 275:77–79. Shiratori Y, Shiima S, Imamura M, et al. 1995. Characteristic difference of hepatocellular carcinoma between hepatitis B- and C-viral infection in Japan. Hepatology 22:1027–1033. Silverberg MJ, Thorsen P, Lindeberg H, Grant LA, Shah KV. 2003. Condyloma in pregnancy is strongly predictive of juvenile-onset recurrent respiratory papillomatosis. Obstet Gynecol 101:645–652. Siman JH, Forsgren A, Berglund G, Floren CH. 2001. Tobacco smoking increases the risk for gastric adenocarcinoma among Helicobacter pylori-infected individuals. Scand J Gastroenterol 36:208–213. Simmonds P. 2000. Hepatitis C virus genotypes. In: Liang TJ, Hoofnagle JH, eds. Biomedical Research Reports, Hepatitis C. San Diego: Academic Press, pp. 53–70. Simmonds P, Holmes EC, Cha T-A, et al. 1993. Classification of hepatitis C virus into six major genotypes and a series of subtypes by phylogenetic analysis of NS-5 region. J Gen Virol 74:2391–2399. Simpson CR, Anderson WJ, Helms PJ, et al. 2002. Coincidence of immunemediated diseases driven by Th1 and Th2 subsets suggests a common aetiology. A population-based study using computerized general practice data. Clin Exp Allergy 32:37–42. Simpson GR, Schulz TF, Whitby D, et al. 1996. Prevalence of Kaposi’s sarcoma associated herpesvirus infection measured by antibodies to recombinant capsid protein and latent immunofluorescence antigen. Lancet 349:1133–1138. Sirianni MC, Vincenzi L, Topino S, et al. 2002. NK cell activity controls human herpesvirus 8 latent infection and is restored upon highly active antiretroviral therapy in AIDS patients with regressing Kaposi’s sarcoma. Eur J Immunol 32:2711–2720. Sitas F, Newton R, Boshoff C. 1999. Increasing probability of mother-to-child transmission of HHV-8 with increasing maternal antibody titer for HHV-8. N Engl J Med 340:1923. Sixbey JW, Lemon SN, Pagano JS. 1986. A second site for Epstein-Barr shedding: The uterine cervix. Lancet 2:1122–1124. Smith JS, Green J, Berrington DG, et al. 2003. Cervical cancer and use of hormonal contraceptives: A systematic review. Lancet 361:1159–1167. Sobala GM, Crabtree JE, Dixon MF, et al. 1991. Acute Helicobacter pylori infection: Clinical features, local and systemic immune response, gastric mucosal histology and gastric juice ascorbic acid concentrations. Gut 32:1415–1418. Sonnenberg A. 1993. The US temporal and geographic variations of diseases related to Helicobacter pylori. Am J Public Health 83:1006–1010. Sonoda S, Yashiki S, Fujiyoshi T, et al. 1993. Ethnically defined immunogenetic factors involved in the pathogenesis of HTLV-I associated diseases. Bio Bulletin 5:79–88. Soulier J, Grollet L, Oksenhendler E, et al. 1995. Kaposi’s sarcoma-associated herpesvirus-like DNA sequences in multicentric Castleman’s disease. Blood 86:1276–1280. Spender LC, Cannell EJ, Hollyoake M, et al. 1999. Control of cell cycle entry and apoptosis in B lymphocytes by Epstein-Barr virus. J Virol 73:4678–4688. Spengler U, Lechmann M, Irrgang B, Dumoulin FL, Sauerbruch T. 1996. Immune responses in hepatitis C virus infection. J Hepatol 24(Suppl 2):20–25. Srivatanakul P, Parkin DM, Jiang YZ, et al. 1991. The role of infection by Opisthorchis viverrini, hepatitis B virus, and aflatoxin exposure in the etiology of liver cancer in Thailand. A correlation study. Cancer 68:2411– 2417. Stanley M. 2003. Chapter 17: Genital human papillomavirus infections—current and prospective therapies. J Natl Cancer Inst Monogr 31:117–124. Stein M, Rappuoli R, Covacci A. 2000. Tyrosine phosphorylation of the Helicobacter pylori CagA antigen after cag-driven host cell translocation. Proc Natl Acad Sci USA 97:1263–1268. Stern MC, Umbach DM, Yu MC, London SJ, Zhang Z–Q, Taylor JA. 2001. Hepatitis B, aflatoxin B1, and p53 Codon 249 mutation in hepatocellular carcinomas from Guangxi, People’s Republic of China, and a meta-analysis of existing studies. Cancer Epidemiol Biomarkers Prev 10:617–625. Stevens CT, Toy PT, Tong MJ, et al. 1985. Perinatal Hepatitis B virus transmission in the United States: Prevention by passive-active immunization. JAMA 253:1740–1745. Stoler MH, Rhodes CR, Whitbeck A, Wolinsky SM, Chow LT, Broker TR. 1992. Human papillomavirus type 16 and 18 gene expression in cervical neoplasias. Hum Pathol 23:117–128. Strauss S, Sastry P, Sonnex C, Edwards S, Gray J. 2002. Contamination of environmental surfaces by genital human papillomaviruses. Sex Transm Infect 78:135–138. Strickler HD, Palefsky JM, Shah KV, et al. 2003. Human papillomavirus type 16 and immune status in human immunodeficiency virus-seropositive women. J Natl Cancer Inst 95:1062–1071. Strickler HD, Schiffman MH, Shah KV, et al. 1998. A survey of human papillomavirus 16 antibodies in patients with epithelial cancers. Eur J Cancer Prev 7:305–313. Stuver SO, Tachibana N, Okayama A, et al. 1993. Heterosexual transmission of human T-cell leukemia/lymphoma virus type I among married couples in Southwestern Japan: An initial report from the Miyazaki Cohort Study. J Infect Dis 167:57–65.

546

PART III: THE CAUSES OF CANCER

Stuver SO, Tachibana N, Okayama A, Romano F, Yokota T, Mueller N. 1992. Determinants of HTLV-I seroprevalence in Miyazaki Prefecture, Japan: A crosssectional study. JAIDS 5:12–18. Sun C-A, Wu D-M, Lin C-C, et al. 2003. Incidence and cofactors of hepatitis C virus-related hepatocellular carcinoma: A prospective study of 12,008 men in Taiwan. Am J Epidemiol 157:674–682. Sung JJ, Lin SR, Ching JY, et al. 2000. Atrophy and intestinal metaplasia one year after cure of H. pylori infection: A prospective, randomized study. Gastroenterology 119:7–14. Sung J-L. 1997. Prevention of hepatitis B and C virus infection for prevention of cirrhosis and hepatocellular carcinoma. J Gastrocenterol Hepatol 12(Suppl): S370–S376. Swinnen LJ, Gulley ML, Hamilton E, Schichman S. 1998. EBV DNA quantitation in serum is highly correlated with the development and regression of posttransplant lymphoproliferative disorder (PTLD) in solid organ transplant recipients. Blood 92(Suppl):S1291. Tabor E. 1994. Tumor suppressor genes, growth factor genes, and oncogenes in hepatitis B virus-associated hepatocellular carcinoma. J Med Virol 42:357–365. Tabor E. 1997. The role of tumor suppressor genes in the development of hepatocellular carcinoma. In: Okuda K, Tabor E, eds. Liver Cancer. New York: Churchill Livingstone, pp. 89–95. Tabor E. 1998. Viral hepatitis and liver cancer. In: Goldin RD, Thomas HC, Gerber MA, eds. Pathology of Viral Hepatitis. London: Arnold, pp. 161–177. Tabor E. 2002. Global epidemiology of hepatocellular carcinoma associated with hepatitis B and C virus infections. In: Margolis HS, Alter MJ, Liang TJ, Dienstag JL, eds. Viral Hepatitis and Liver Disease. Atlanta: International Medical Press, pp. 400–403. Tabor E. 2003. Interferon for preventing and treating hepatocellular carcinoma associated with the hepatitis B and C viruses. Dig Liver Dis 35:297–305. Tachibana N, Okayama A, Ishihara S, et al. 1992. High HTLV-I proviral DNA level associated with abnormal lymphocytes in peripheral blood from asymptomatic carriers. Int J Cancer 51:593–595. Tachibana N, Okayama A, Ishizaki J, et al. 1988. Suppression of tuberculin skin reaction in healthy HTLV-I carriers from Japan. Int J Cancer 42:829– 831. Tagger A, Donato F, Ribero ML, et al. 1999. Case-control study on hepatitis C virus (HCV) as a risk factor for hepatocellular carcinoma: The role of HCV genotypes and the synergism with hepatitis B virus and alcohol. Int J Cancer 81:695–699. Tajima K (T and B–Cell Malignancy Study Group). 1990. The 4th nation-wide study of adult T-cell leukemia/lymphoma (ATL) in Japan: Estimates of risk of ATL and its geographical and clinical features. Int J Cancer 45:237–243. Tajima K, Hinuma Y. 1984. Epidemiological features of adult T-cell leukemia virus. In: Mathe G, Reizenstein P, eds. Pathophysiological Aspects of Cancer Epidemiology. Oxford: Pergamon Press, pp. 75–87. Tajima K, Akaza T, Koike K, et al. 1984. HLA antigens and adult T-cell leukemia virus infection: A community-based study in the Goto Islands, Japan. Jpn J Clin Oncol 14:347–352. Tajima K, Kamura S, Ito S, et al. 1987. Epidemiological features of HTLV-I carriers and incidence of ATL in an ATL-endemic island: A report of the community-based cooperative study in Tsuschima, Japan. Int J Cancer 40:741–746. Takada K. 1999. Epstein-Barr virus and gastric carcinoma. Epstein-Barr Virus Rep 7:95–100. Takahashi M, Yamada G, Miyamoto R, Doi T, Endo H, Tsuji T. 1993. Natural course of chronic hepatitis C. Am J Gastroenterol 88:240–243. Takaki A, Wiese M, Maertens G, et al. 2000. Cellular immune responses persist and humoral responses decrease two decades after recovery from a single-source outbreak of hepatitis C. Nature Med 6:578–582. Takeuchi T, Katsume A, Tanaka T, et al. 1999. Real-time detection system for quantification of hepatitis C virus genome. Gastroenterol 118:636–642. Tanaka E, Ohue C, Aoyagi K, et al. 2000. Evaluation of a new enzyme immunoassay for hepatitis C virus (HCV) core antigen with clinical sensitivity approximating that of genomic amplification of HCV RNA. Hepatology 32:388–393. Tanaka H, Hiyama T, Okubo Y, Kitada A, Fujimoto I. 1994. Primary liver cancer incidence-rates related to hepatitis C virus infection: A correlation study in Osaka, Japan. Cancer Causes Control 5:61–65. Tanaka K, Hirohata T, Koga S, et al. 1991. Hepatitis C and hepatitis B in the etiology of hepatocellular carcinoma in the Japanese population. Cancer Res 51:2842–2847. Tanaka K, Hirohata T, Takeshita S, et al. 1992. Hepatitis B virus, cigarette smoking and alcohol consumption in the development of hepatocellular carcinoma: A case-control study in Fukuoka, Japan. Int J Cancer 51:509–514. Tanaka K, Ikematsu H, Hirohata T, Kashiwagi S. 1996. Hepatitis C virus infection and risk of hepatocellular carcinoma among Japanese: Possible role of type Ib(II) infection. J Natl Cancer Inst 88:742–746. Tanaka K, Sato H, Okochi K. 1984. HLA antigens in patients with adult T-cell leukemia. Tissue Antigens 23:81–83. Tanaka K, Stuver SO, Ikematsu H, et al. 1997. Heterosexual transmission of hepatitis C virus among married couples in southwestern Japan. Int J Cancer 72:50–55. Tanaka S, Takenaka K, Matsumata T, Mori R, Sugimachi K. 1996. Hepatitis C virus replication is associated with expression of transforming growth factor-alpha

and insulin-like growth factor-II in cirrhotic livers. Dig Dis Science 41:208–215. T- and B-cell Malignancy Study Group. 1981. Statistical analysis of immunologic, clinical and histopathologic data in lymphoid malignancies in Japan. Jpn J Clin Oncol 11:15–38. T- and B-cell Malignancy Study Group. 1988. The third nation-wide study of adult T-cell leukemia/lymphoma (ATL) in Japan: Characteristic patterns of HLA antigen and HTLV-I infection in ATL patients and their relatives. Int J Cancer 41:505–512. Tappero JW, Conant MA, Wolfe SF, Berger TG. 1993. Kaposi’s sarcoma: Epidemiology, pathogenesis, histology, clinical spectrum, staging criteria and therapy. J Am Acad Dermatol 28:371–395. Telfer PT, Brown D, Devereux H, Lee CA, Desheiki GM. 1994. HCV RNA levels and HIV infection: Evidence for a viral interaction in haemophilic patients. Br J Haematol 88:397–399. Terrault NA. 2002. Sexual activity as a risk factor for hepatitis C. Hepatology 36(Suppl 1):S99–S105. Terry G, Ho L, Londesborough P, Cuzick J, Mielzynska-lohnas I, Lorincz A. 2001. Detection of high-risk HPV types by the hybrid capture 2 test. J Med Virol 65:155–162. Thio CL, Seaberg EC, Skolasky R, et al. 2002. HIV-1, hepatitis B virus, and risk of liver-related mortality in the Multicenter Cohort Study (MACS). Lancet 360:1921–1926. Thomas DL. 2002. Hepatitis C and human immunodeficiency virus infection. Hepatology 36(Suppl 1):S201–S209. Thomas DL, Astemborski J, Rai RM, et al. 2000. The natural history of hepatitis C virus infection. Host, viral and environmental factors. JAMA 284:450–456. Thomas HC. 1990. The hepatitis B virus and the host response. J Hepatol 11(Suppl):S83–S89. Thomas JE, Bassett MT, Sigola LB, Taylor P. 1990. Relationship between bladder cancer incidence, Schistosoma haematobium infection, and geographical region in Zimbabwe. Trans R Soc Trop Med Hyg 84:551–553. Thomas KK, Hughes JP, Kuypers JM, et al. 2000. Concurrent and sequential acquisition of different genital human papillomavirus types. J Infect Dis 182:1097– 1102. Thorley-Lawson DA, Gross A. 2004. Persistence of the Epstein-Barr virus and the origins of associated lymphomas. NEJM 350:1328–1337. Thorpe LE, Ouellet LJ, Hershow R, et al. 2002. Risk of hepatitis C virus infection among young adult injection drug users who share injection equipment. Am J Epidemiol 156:645–653. Tierney RJ, Steven N, Young LS, Rickinson AB. 1994. Epstein-Barr virus latency in blood mononuclear cells – analysis of viral gene transcription during primary infection and in the carrier state. J Virol 68:7374–7385. Tobler LH, Busch MP. 1997. History of post-transfusion hepatitis. Clin Chem 43:1487–1493. Tong MJ, EL-Farra NS, Reikes AR, Co RL. 1995. Clinical outcomes after transfusion-associated hepatitis C. N Engl J Med 332:1463–1466. Touloumi G, Hatzakis A, Potouridou I, et al. 1999. The role of immunosuppression and immune-activation in classic Kaposi’s sarcoma. Int J Cancer 82:817–821. Trichopoulos D, Day NE, Kaklamani E, et al. 1987. Hepatitis B virus, tobacco smoking and ethanol consumption in the etiology of hepatocellular carcinoma. Int J Cancer 39:45–49. Uchiyama T, Yodoi J, Sagawa K, et al. 1977. Adult T-cell leukemia: Clinical and hematological features of 16 cases. Blood 50:481–492. Ueda H, Ohkoshi S, Kim C-M, Jay G. 1997. A transgenic mouse model of HBVrelated hepatocellular carcinoma. In: Okuda K, Tabor E, eds. Liver Cancer. New York: Churchill Livingstone, pp. 201–211. Uemura N, Mukai T, Okamoto S, et al. 1997. Effect of Helicobacter pylori eradication on subsequent development of cancer after endoscopic resection of early gastric cancer. Cancer Epidemiol Biomarkers Prev 6:639–642. Uemura N, Okamoto S, Yamamoto S, et al. 2001. Helicobacter pylori infection and the development of gastric cancer [Comment]. N Engl J Med 345: 784–789. Uhara H, Sato Y, Mukai K, et al. 1990. Detection of Epstein-Barr virus DNA in Reed-Sternberg cells of Hodgkin’s disease using the polymerase chain reaction and in situ hybridization. Jpn J Cancer Res 81:272–278. Uno H, Kawano K, Matsuoka H, et al. 1988. HLA and adult T cell leukemia: HLA genes controlling susceptibility to human T cell leukemia virus type I. Clin Exp Immunol 71:211–216. Usuku W, Sonoda S, Osame M, et al. 1988. HLA haplotype-linked high immune responsiveness against HTLV-I in HTLV-I associated myelopathy: Comparison with adult T-cell leukemia/lymphoma. Ann Neurol 23(Suppl):S143– S150. Vaira D, Gatta L, Ricci C, Miglioli M. 2002. Review article: Diagnosis of Helicobacter pylori infection. Aliment Pharmacol Ther 1(Suppl 1):16–23. Vall Mayans M, Calvet X, Bruix J, et al. 1990. Risk factors for hepatocellular carcinoma in Catalonia, Spain. Int J Cancer 46:378–381. Van Doorn LJ, Figueiredo C, Sanna R, et al. 1998. Clinical relevance of the cagA, vacA, and iceA status of Helicobacter pylori. Gastroenterology 115:58–66. Van Doorn LJ, Quint W, Kleter B, et al. 2002. Genotyping of human papillomavirus in liquid cytology cervical specimens by the PGMY line blot assay and the SPF(10) line probe assay. J Clin Microbiol 40:979–983.

Infectious Agents Van Doornum GJ, Korse CM, Buning-Kager JC, et al. 2003. Reactivity to human papillomavirus type 16 L1 virus-like particles in sera from patients with genital cancer and patients with carcinomas at five different extragenital sites. Br J Cancer 88:1095–1100. Van Esser JW, Niesters HG, Thijsen SF, et al. 2001. Molecular quantification of viral load in plasma allows for fast and accurate prediction of response to therapy of Epstein-Barr virus-associated lymphoproliferative disease after allogeneic stem cell transplantation. Br J Haematol 113:814–821. Vandamme AM, Salemi M, Desmyter J. 1998. The simian origins of the pathogenic human T-cell lymphotropic virus type I. Trends Microbiol 6:477–483. Villa LL, Costa RL, Petta CA, et al. 2005. Prophylactic quadrivalent human papillomavirus (types 6, 11, 16, and 18) L1 virus-like particle vaccine in young women: a randomized double-blind placebo-controlled multicentre phase II efficacy trial. Lancet Oncol 6:271–278. Vittorio CC, Schiffman MH, Weinstock MA. 1995. Epidemiology of human papillomaviruses. Dermatol Clin 13:561–574. Vizcaino AP, Parkin DM, Boffetta P, Skinner ME. 1994. Bladder cancer: Epidemiology and risk factors in Bulawayo, Zimbabwe. Cancer Causes Control 5:517–522. Von Knebel DM. 2002. New markers for cervical dysplasia to visualize the genomic chaos created by aberrant oncogenic papillomavirus infections. Eur J Cancer 38:2229–2242. Wakita T, Pietschmann T, Kato T, et al. 2005. Production of infectious hepatitis C virus in tissue culture from a cloned viral genome. Nat Med 11:791–96. Walker IR, Strickland RG, Ungar B, Mackay IR. 1971. Simple atrophic gastritis and gastric carcinoma. Gut 12:906–911. Wallin KL, Wiklund F, Angstrom T, et al. 1999. Type-specific persistence of human papillomavirus DNA before the development of invasive cervical cancer. N Engl J Med 341:1633–1638. Wang HH, Wu MS, Shun CT, Wang HP, Lin CC, Lin JT. 1999. Lymphoepitheliomalike carcinoma of the stomach: A subset of gastric carcinoma with distinct clinicopathological features and high prevalence of Epstein-Barr virus infection. Hepatogastroenterology 46:1214–1219. Wang JT, Sheu JC, Lin JT, Wang TH, Chen DS. 1992. Detection of replicative form of hepatitis C virus RNA in peripheral blood mononuclear cells. J Infect Dis 166:1167–1169. Wang JT, Wang TH, Lin JT, Lee CZ, Sheu JC, Chen DS. 1995. Effect of hepatitis C antibody screening in blood donors on post-transfusion hepatitis in Taiwan. J Gastroenterol Hepatol 10:454–458. Wang L-Y, Hatch M, Chen C-J, et al. 1996. Aflatoxin exposure and risk of hepatocellular carcinoma in Taiwan. Int J Cancer 67:620–625. Wang QJ, Jenkins FJ, Jacobson LP, et al. 2001. Primary human herpesvirus 8 infection generates a broadly specific CD8+ T-cell response to viral lytic cycle proteins. Blood 97:2366–2373. Wang SS, Hildesheim A. 2003. Chapter 5: Viral and host factors in human papillomavirus persistence and progression. J Natl Cancer Inst Monogr 31: 35–40. Warren JR, Marshall BJ. 1983. Unidentified curved bacilli in gastric epithelium in active chronic gastritis. Lancet 1:1273–1275. Wattel E, Mariotti M, Agis F, et al. 1992. Quantification of HTLV-I proviral copy number in peripheral blood of symptomless carriers from the French West Indies. JAIDS 5:943–946. Wattel E, Vartanian JP, Pannetier C, Wain-Hobson S. 1995. Clonal expansion of human T-cell leukemia virus type I-infected cells in asymptomatic and symptomatic carriers without malignancy. J Virol 69:2863–2868. Weiss L, Strickler JG, Warnke RA, et al. 1987. Epstein-Barr viral DNA in tissue of Hodgkin’s disease. Am J Pathol 129:86–91. Whitby D, Howard MR, Tenant-Flowers M, et al. 1995. Detection of Kaposi sarcoma associated herpesvirus in peripheral blood of HIV-infected individuals and progression to Kaposi’s sarcoma. Lancet 346:799–802. Whitby D, Luppi M, Barozzi P, Boshoff C, Weiss RA, Torelli G. 1998. Human herpesvirus 8 seroprevalence in blood donors and lymphoma patients from different regions of Italy. J Natl Cancer Inst 90:395–397. Whittle HC, Inskip H, Hall AJ, Mendy M, Downes R, Hoare S. 1991. Vaccination against hepatitis B and protection against chronic viral carriage in the Gambia. Lancet 337:747–750. WHO Expert Committee on the Prevention and Control of Schistosomiasis and Soil-Transmitted Helminthiasis. 2002. Geneva: World Health Organization. WHO Vaccine-Preventable Diseases: Monitoring System. Global Summary. 2002. Geneva: World Health Organization, Department of Vaccines and Biologicals. Widell A, Molnegren V, Pieksma F, Calmann M, Peterson J, Lee SR. 2002. Detection of hepatitis C core antigen in serum or plasma as a marker of hepatitis C viremia in the serological window–phase. Transfusion 12:107–113. Wideroff L, Schiffman M, Haderer P, et al. 1999. Seroreactivity to human papillomavirus types 16, 18, 31, and 45 virus-like particles in a case-control study of cervical squamous intraepithelial lesions. J Infect Dis 180:1424–1428. Wiese M, Berr F, Lafrenz M, Porse H, Oesen U, for the East German Hepatitis C Study Group. 2000. Low frequency of cirrhosis in a hepatitis C (genotype 1b) single-source outbreak in Germany: A 20–year multicenter study. Hepatology 32:91–96. Wiley DJ, Douglas J, Beutner K, et al. 2002. External genital warts: Diagnosis, treatment, and prevention. Clin Infect Dis 35(Suppl):S210–S224.

547

Winer RL, Lee SK, Hughes JP, Adam DE, Kiviat NB, Koutsky LA. 2003. Genital human papillomavirus infection: Incidence and risk factors in a cohort of female university students. Am J Epidemiol 157:218–226. Wojcicki JM, Newton R, Urban MI, et al. 2004. Low socioeconomic status and risk for infection with human herpesvirus 8 among HIV-1 negative, South African black cancer patients. Epidemiol Infect 132:1191–1197. Wolf H, Zur Hausen H, Becker V. 1973. EB viral genomes in epithelial nasopharyngeal carcinoma cells. Nat New Biol 244:245–257. Wong DKH, Cheung AM, O’Rourke K, Naylor CD, Detsky AS, Heathcote J. 1993. Effect of alpha-interferon treatment in patients with hepatitis B e antigenpositive chronic hepatitis B. A meta-analysis. Ann Intern Med 119:312– 323. Wotherspoon AC, Doglioni C, Diss TC, et al. 1993. Regression of primary low-grade B-cell gastric lymphoma of mucosa-associated lymphoid tissue type after eradication of Helicobacter pylori. Lancet 342:575–577. Wright TC, JR., Schiffman M. 2003. Adding a test for human papillomavirus DNA to cervical-cancer screening. N Engl J Med 348:489–490. Wu FY, Ahn J-H, Alcendor DJ, et al. 2001. Origin-independent assembly of Kaposi’s sarcoma-associated herpesvirus DNA replication compartments in transient cotransfection assays and association with the ORF-K8 protein and cellular PML. J Virol 75:1487–1506. Wu L, Lo P, Yu X, Stoops JK, Forghani B, Zhou ZH. 2000. Three-dimensional structure of the human herpesvirus 8 capsid. J Virol 74:9646–9654. Wu MS, Shun CT, Wu CC, et al. 2000. Epstein-Barr virus-associated gastric carcinomas: Relation to H. pylori infection and genetic alterations. Gastroenterology 118:1031–1038. Wu TC, Mann RB, Charache P, et al. 1990. Detection of EBV gene expression in Reed-Sternberg cells of Hodgkin’s disease. Int J Cancer 46:801–804. Xi LF, Carter JJ, Galloway DA, et al. 2002. Acquisition and natural history of human papillomavirus type 16 variant infection among a cohort of female university students. Cancer Epidemiol Biomarkers Prev 11:343–351. Xia HH, Wong BC, Lam SK. 2003. Chemoprevention of gastric cancer: Current status. Chinese Med J 116:5–10. Xue FB, Xu YY, Wan Y, Pan BR, Ren J, Fan DM. 2001. Association of H. pylori infection with gastric carcinoma: A meta analysis. World J Gastroenterol 7:801–804. Yamada T, Manos MM, Peto J, et al. 1997. Human papillomavirus type 16 sequence variation in cervical cancers: A worldwide perspective. J Virol 71:2463–2472. Yamaguchi K, Kiyokawa T, Nakada K, et al. 1988. Polyclonal integration of HTLVI proviral DNA in lymphocytes from HTLV-I seropositive individuals: An intermediate state between the healthy carrier state and smoldering ATL. Br J Haematol 68:169–174. Yamaguchi K, Nishima H, Kohrogi H, et al. 1983. A proposal for smoldering adult T-cell leukemia: A clinicopathologic study of five cases. Blood 62:758– 766. Yashiki S, Fujiyoshi T, Arima N, et al. 2001. HLA-A*26, HLA-B*4002, HLAB*4006, and HLA-B*4801 alleles predispose to adult T cell leukemia: The limited recognition of HTLV type 1 Tax peptide anchor motifs and epitopes to generate anti-HTLV type 1 Tax CD8+ cytotoxic T lymphocytes. AIDS Res Human Retroviruses 17:1047–1061. Yeung WM, Zong YS, Chiu CT, et al. 1993. Epstein-Barr virus carriage by nasopharyngeal carcinoma in situ. Int J Cancer 53:746–750. Yokota T, Cho M-J, Tachibana N, et al. 1989. The prevalence of antibody to p42 among ATL patients in comparison with healthy carriers in Japan. Int J Cancer 43:970–974. Yoshida M, Hattori S, Seiki M. 1985. Molecular biology of human T-cell leukemia. In: Voigt P, ed. Current Topics in Microbiology and Immunology, Volume 5, No. 115. Berlin: Springer–Verlag, pp. 157–175. Yoshiida M, Miyoshi I, Hinuma Y. 1982. Isolation and characterization of retrovirus from cell lines of human T-cell leukemia and its implication in the disease. Proc Natl Acad Sci USA 79:2031–2035. Yotsuyanagi H, Yasuda K, Moriya K, et al. 2001. Frequent presence of HBV in the sera of HBsAg-negative, anti-HBc-positive blood donors. Transfusion 41:1093–1099. Young LS, Murray PG. 2003. Epstein-Barr virus and oncogenesis: From latent genes to tumours. Oncogene 22:5108–5121. Yu M-W, Chen C-J. 1993. Elevated serum testosterone levels and risk of hepatocellular carcinoma. Cancer Res 53:790–794. Yu M-W, Cheng S-W, Lin M-W, et al. 2000. Androgen-receptor gene CAG repeats, plasma testosterone levels, and risk of hepatitis B-related hepatocellular carcinoma. J Natl Cancer Inst 92:2023–2028. Yu M-W, Hsu FC, Sheen IS, et al. 1997. Prospective study of hepatocellular carcinoma and liver cirrhosis in asymptomatic chronic hepatitis B virus carriers. Am J Epidemiol 145:1039–1047. Yuan JM, Ross RK, Stanczyk FZ, et al. 1995. A cohort study of serum testosterone and hepatocellular carcinoma in Shanghai, China. Int J Cancer 63:491–493. Zanetti AR, Tanzi E, Paccagnini S, et al. 1995. The Lombardy Study Group on Vertical HCV Transmission. Mother-to-infant transmission of hepatitis C virus. Lancet 345:289–291. Zhang ZW, Patchett SE, Perrett D, Domizio P, Farthing MJ. 2000. Gastric alphatocopherol and beta-carotene concentrations in association with Helicobacter pylori infection. Eur J Gastroenterol Hepatol 12:497–503.

548

PART III: THE CAUSES OF CANCER

Zhong J, Gastaminza P, Cheng G, et al. 2005. Robust hepatitis C virus infection in vitro. Proc Natl Acad Sci USA 102:9294–9299. Zhong W, Wang H, Herndier B, Ganem D. 1996. Restricted expression of Kaposi sarcoma-associated herpesvirus (human herpesvirus 8) genes in Kaposi sarcoma. Proc Natl Acad Sci USA 93:11883–11888. Ziegert C, Wentzensen N, Vinokurova S, et al. 2003. A comprehensive analysis of HPV integration loci in anogenital lesions combining transcript and genomebased amplification techniques. Oncogene 22:3977–3984. Ziegler JL, Katongole-Mbidde E. 1996. Kaposi’s sarcoma in childhood: An analysis of 100 cases from Uganda and relationship to HIV infection. Int J Cancer 65:200–203. Ziegler J, Newton R, Bourboulia D, et al. 2003. Risk factors for Kaposi’s sarcoma: A case-control study of HIV-seronegative people in Uganda. Int J Cancer 103:233–240. Zong J-C, Ciufo DM, Alcendor DJ, et al. 1999. High-level variability in the ORFK1 membrane protein gene at the left end of the Kaposi’s sarcoma-associated

herpesvirus genome defines four major virus subtypes and multiple variants or clades in different human populations. J Virol 73:4156–4170. Zucker-Franklin D, Pancake BA. 1998. Human T-cell lymphotropic virus type 1 tax among American blood donors. Clin Diagn Lab Immunol 5:831–835. Zucker-Franklin D, Pancake BA, Marmor M, Legler PM. 1997. Reexamination of human T cell lymphotropic virus (HTLV-I/II) prevalence. Proc Natl Acad Sci USA 94:6403–6407. Zur Hausen H. 2000. Papillomaviruses causing cancer: Evasion from host-cell control in early events in carcinogenesis. J Natl Cancer Inst 92:690–698. Zur Hausen H. 2001. Proliferation-inducing viruses in non-permissive systems as possible causes of human cancers. Lancet 357:381–384. Zur Hausen H. 2002. Papillomaviruses and cancer: From basic studies to clinical application. Nat Rev Cancer 2:342–350.

27

Immunologic Factors GARETH J. MORGAN, MARTHA S. LINET, AND CHARLES S. RABKIN

T

he immune system evolved from the need of multicellular organisms to protect against invasion by pathogens. Infectious agents are highly diverse, and the immune system has functional components that both detect and react to a wide variety of threats with great specificity. Malfunction of the system can lead to immune deficiency with increased susceptibility to infection, or, alternatively, to autoimmune disease with aberrant response directed against self components.

STRUCTURE AND FUNCTION OF THE IMMUNE SYSTEM Function of the Immune System The immune response is composed of a primordial innate immunity, which recognizes broad features shared by many infectious agents, and a more complex adaptive immunity, which develops after exposure to specific antigen and provides memory for subsequent re-exposure. The pathogen-associated molecular patterns detected by innate immunity are generally polysaccharides and polynucleotides that do not differ between pathogens are not found in the host, for which receptors are encoded in the host germ-line. In contrast, the epitopes recognized in adaptive immunity are usually derived from polypeptides, and the host receptors have a narrow specificity. The diversity of the adaptive immune response is generated by a unique feature of its cells. Lymphocytes are the only cells to undergo genetic recombination and mutation of their germ-line DNA as part of their normal life cycle in a process designated controlled genetic instability (Kelsoe et al., 1998; Vanasse et al., 1999). These alterations also include immunoglobulin and antigen receptor gene rearrangement during cellular maturation in the bone marrow and somatic hypermutations and class switch recombination, which then take place during the passage through the germinal centers of the lymph nodes. Abnormalities in either the control or the mechanisms underlying the process can result in immunodeficiency and/or lymphoproliferative diseases. Although the immune system has evolved to recognize external antigens primarily, the immune system also recognizes internal antigens. To prevent autoimmune disease, cells directed against selfantigens are deleted from the immune repertoire during immune modulation. Tumor specific or abnormally expressed proteins are processed and expressed by antigen processing cells and are substrates for recognition by the immune system, a form of immune surveillance against malignant change. Lymphocytes also have a highly variable life history. Exposure to specific antigen may lead to clonal expansion and effector cell differentiation. Derangements of these interactions are potentially important pathogenic processes and are regulated by a finely balanced system of cytokine and cell-to-cell interactions. The immune system is therefore composed of a network of recognition and effector cells whose clonal expansion and/or differentiation are governed by the timing and nature of external exposures.

Structure of the Immune System Effector cells of the innate immune system include neutrophils, natural killer (NK) cells, dendritic cells, and cells of the monocyte/macrophage lineage. Neutrophils and macrophages are phagocytes that engulf and eliminate foreign material. NK cells are lymphocytes that are preprogrammed to kill host cells that have

become virus-infected or cancerous. Macrophages and dendritic cells also contribute to adaptive immunity by processing foreign molecules and presenting them as antigens on their cellular surfaces. In the adaptive immune system, lymphocytes recognize and react to unique antigens. The initial response requires T lymphocytes, which mature in the thymus (hence “T”). Helper T cells are crucial regulators of immune responses, and cytotoxic T cells destroy target cells bearing antigen from intracellular infection, neoplastic transformation, or transplantation from a different host. B lymphocytes, which develop in the bone marrow (hence “B”), are the source of antibodyproducing plasma cells. Antigen encountered at the skin and mucosal surfaces is presented to the lymphocytes by dendritic cells after they have moved to the lymph nodes. In the lymph nodes and spleen, helper T cells interact with these dendritic cells to enable B cells to produce highly specific antibody and to generate cytotoxic T cells. The repertoire of specificities is maintained in the bone marrow by the continual production of naive lymphocytes through recombination of the immunoglobulin and T cell receptor genes in precursor B and T cells, respectively (Fig. 27–1). The spleen and lymph nodes are the structural components wherein the immune system selects, refines and expands B lymphocyte clones that have encountered antigens. Two major subsets of T lymphocytes are distinguished by the cell surface glycoproteins CD8 or CD4. CD8+ T cells are responsible for cytotoxic T lymphocyte activity, whereas CD4+ T cells perform important helper functions for both cell-mediated and antibody-mediated immune responses. T-cell recognition of antigens depends on their presentation in association with human leukocyte antigen (HLA) histocompatibility molecules. CD8+ T cells bind epitopes presented on HLA Class I molecules, which are expressed by most normal cells in the body. In contrast, CD4+ T cells bind epitopes associated with HLA Class II molecules, which are found only on antibody-presenting cells. Different helper functions are performed by subsets of CD4+ T cells that can be distinguished according to their secretion and responsiveness to major classes of cytokines. Type 1 cytokines include interleukin-2 (IL-2), interferon-gamma, and lymphotoxin, and contribute to Type 1 helper (Th1) cell response to viruses and other intracellular infections. Type 2 cytokines include IL-4, Il-5, IL-6, and IL-10, and mediate Type 2 helper (Th2) cell response to extracellular pathogens such as parasites. Appropriate modulation of these two responses is required for normal immune functional responses and involves activation followed by deletion of highly reactive clones (Hall et al., 2000). When the balance of Th1 and Th2 responses is shifted, a Th1 predominance leads to an increase in autoimmune diseases, whereas a Th2 predominance leads to allergic/atopic disorders. The lymph node has a distinct anatomic structure and is composed of a mantle zone of cells that have not entered the germinal center, and a marginal zone composed of cells that have been exposed to the processes in the germinal center. The germinal center selects for cells that optimally recognize antigen by the process of affinity maturation, whereby mutations are introduced into the immunoglobulin heavychain gene by somatic hypermutation and subsequently selected for their capacity to efficiently bind antigen. The level and extent of mutation can be used as a marker of a lymphocyte that is either pre or postgerminal center phase of development. After leaving the germinal center, B cells undergo class switching to alter the functional activities of the immunoglobulin molecules. These B cells then migrate to the bone marrow as antibody-producing plasma cells. A second

549

550

PART III: THE CAUSES OF CANCER

Dendritic Cell Antigen

Adaptive Immune Response

Immune Effector

B Cells

Antibodies

Helper T Cells TH1

TH2

Innate Immune Response

Figure 27–1. Diagrammatic representation of the structure of the immune system.

Cytotoxic T Cells

NK cells Monocytes Neutrophils

pathway exists whereby B cells can differentiate directly to plasma cells without going through a germinal center leading to cells producing low-affinity antibodies. T cells undergo a range of similar processes in the thymus. Once formed, B and T cells circulate throughout the body, and a specialized lymph system collects these cells from the tissues and delivers them back to the circulation.

of immunodeficiency is less clear, in part because it is difficult to measure minimal alterations in immune function. However, the recognition of genetic variation in immune control offers prospects for evaluating the mechanisms underlying these concepts.

Abnormalities of the Immune System

Immune function varies with age and environmental exposures. At birth, the adaptive immune response is undeveloped as the immune system has not encountered foreign antigen, and specific immunoglobulins are not produced (Burns and Leventhal, 2000; Malaguarnera et al., 2001). However, at this time immunoglobulins are present that have been acquired from the maternal circulation and from the colostrum of breast milk. Protection via this route is lost within weeks to months after birth as the half-life of immunoglobulins is relatively short. One of the most obvious changes in the immune system with increasing age is the involution of the thymus starting in adolescence. Subsequently, there are increases in memory cells, decreases in naive lymphocytes, and a reduction in the capacity to generate new T cells. With rising age, there is a reduction in the proliferative response of T cells and in the capacity to generate cytotoxic effector cells. Suppressor T cell function also changes with increasing age and may be associated with changes in autoantibodies reactive with self antigens. These antibodies are of low affinity, are found in 80% or more of elderly persons, and appear to have little pathologic significance. There are also age-related changes in B cells that may reflect a reduction in T-cell dependent antibody production (Bonafe et al., 2001; Ferguson et al., 1995; Ginaldi et al., 1999; Solana and Mariani, 2000). The breadth of the B-cell repertoire decreases with increasing age, as does the vigor of primary and secondary B-cell responses. Levels of antibody produced are also reduced. In addition, NK and macrophage activity decline with increasing age; the numbers of NK cells increase, but their functional activity decreases. Studies of the effects of aging on immune function in humans must be interpreted with caution due to potential differences in study populations. There is often substantial variability in the exclusionary criteria, which may result in failure to exclude or total exclusion of persons at the oldest ages, those who are afflicted with chronic or acute neoplastic or inflammatory diseases, or individuals receiving medications that may alter immune status. Some, but not all, investigators have used a common set of criteria recommended by certain consortia of investigators. Nevertheless, age is an important potential confounder when considering the possible role of immune function in cancer risk. The occurrence of various diseases at different ages must be evaluated as a consequence of changes in immune function (Asimov, 1983; Peto et al., 1975). Fundamental cellular and molecular mechanisms for declines in immune function

Immune alterations can affect all of the functions of the immune system, although the mechanisms vary that lead to cancer predisposition. Abnormalities of immune function can take several forms that may vary greatly in severity, ranging from primary or acquired forms of immune deficiency, to altered reactivity to antigenic stimulation associated with allergic disorders, to autoreactivity directed against an organ-specific antigen versus multiple antigens as occurs in systemic autoimmune diseases. In immunosuppressed persons, control of cells with latent viral infection is seriously compromised. Viral carcinogenesis is thought to be a major pathogenic mechanism in the immunosuppressed, outweighing malignant transformation arising by other mechanisms. Inherited immunodeficiency disorders, AIDS-associated lymphomas, and local inflammatory states each appear to involve abnormal T cell–B cell interactions. It is not always possible to distinguish the role played by immunosuppression from the underlying cause of the immunosuppression in these conditions. Allergy in the most general sense refers to all forms of altered reactivity to antigenic stimulation, whether or not the response is associated with an adverse clinical reaction to an antigen (e.g., hypersensitivity) or a protective effect (e.g., immunity). Clinically, allergic disorders usually refer to potentially harmful immune responses, and allergic reactions generally designate responses to environmental antigens including components of certain foods, drugs, pollen, and other agents. Allergic reactions occur as the expression of acquired immunologic responsiveness, involving preexisting specific antibodies and T cells. Allergic disorders include immediate hypersensitivity reactions, cytotoxic reactions, Arthus and other immune complex reactions, and a form of allergic disorders in which sensitized T cells react with antigen in the absence of antibodies. Autoimmune disorders, in which there is autoreactivity against self antigens, include organ-specific disorders (including insulin-dependent diabetes, multiple sclerosis, autoimmune thyroid disorders, and inflammatory bowel disease) and systemic autoimmune disorders (including systemic lupus erythematosus, scleroderma, rheumatoid arthritis, chronic graftversus-host disease and others). The genetically determined severe immunodeficiency disorders are associated with dramatically elevated risks of infection and lymphoproliferative malignancies. The impact on cancer risk of lesser degrees

Variation of Immune Function with Age

551

Immunologic Factors with age are incompletely understood. The question of whether the immune deficits in the elderly contribute to increased incidence of cancer in the elderly continues to be controversial, and data are unclear.

IMMUNE VARIATION AND CANCER RISK The primary and acquired severe immunodeficiency disorders have profound effects on risk of specific cancers, although generally not the most common types of epithelial cancers that occur in adults. The level of immune dysfunction seen with autoimmune conditions, such as systemic lupus erythematosus and rheumatoid arthritis, has been consistently associated with moderate increases in cancer risks. However, the role of mild dysfunction, such as occurs in atopy or asthma, is less firmly established in relation to cancer risk.

Primary Severe Immune Deficiency The inherited immunodeficiency syndromes are rare disorders that predispose to recurrent and persistent infection. Although affected individuals may be so severely immunosuppressed that they die from uncontrolled infection in early childhood, those that survive experience substantially increased risks of lymphoproliferative malignancy (Ochs and Rosen, 1999). These single-gene defects may be useful models for elucidating the relationship between abnormal immune function and cancer, although the mechanism leading to increased risk of malignancy is frequently obscure. The underlying molecular defects compromise critical cellular processes, such as the repair of doublestranded DNA breaks, which may directly contribute to malignant transformation. Thus, it may be difficult to ascribe the excess cancer risk to immunodeficiency per se. The types of immune deficiency and their underlying mechanism are summarized in Table 27–1. Data from case series and international registries of immunodeficiency disorders suggest that affected children have excess risks of certain cancers, most notably lymphoproliferative malignancies. However, it is difficult to quantify the risks due to the rarity of each of the conditions and uncertainty about the numbers of affected individuals, particularly those with less severe clinical manifestations of disease. Cancer excesses have been observed in young persons with congenital X-linked immunodeficiency disorders (Purtilo et al., 1977), ataxia-telangiectasia and Wiskott-Aldrich syndrome (Filipovich et al., 1992). Among 491 cancers occurring in the setting of primary immunodeficiency as reported to one registry, more than half were nonHodgkin lymphoma (NHL) (Filipovich et al., 1987). Infection with Epstein-Barr virus (EBV) appears to be an important factor in the pathogenesis of lymphoma arising in immunodeficiency disorders (IARC, 1997). The nature and mechanism of the underlying immunodeficiency disorder are important in determining the cancer risk. The X-linked hyperIgM syndrome (XHIM) is the consequence of a T-cell defect resulting from mutations in the CD40 ligand. Affected patients have normal or elevated IgM levels and lymphoid hyperplasia, which distinguishes XHIM from other causes of primary hypo-gammaglobulinemia (Ramesh et al., 1999). In addition to recurrent infection, XHIM is asso-

ciated with increased risk of malignancy. Lymphomas predominate, constituting 56% of all tumors in one series, but an increased incidence of liver and biliary tumors is a further unique feature of XHIM (Filipovich et al., 1994). A more frequently diagnosed deficiency of immunoglobulin production is common variable immunodeficiency (CVID), which is due to an incompletely characterized B and T cell defect. CVID appears in previously immunologically normal individuals with two age peaks of occurrence, at ages 1–5 and 16–20. Presenting features include respiratory and gut infections, but cancer may also develop as a serious complication. A registry study demonstrated a 30-fold excess of NHL among 220 patients with CVID (Cunningham-Rundles et al., 1987; Kinlen, 1985; Sneller et al., 1993). In a separate study, 19 cases of NHL were observed among 248 patients with CVID followed for 1–25 years (Cunningham-Rundles and Bodian, 1999). In addition to lymphoproliferative disorders, a 50-fold increased risk of gastric cancer has also been noted (Kinlen, 1985), thought to be due to a failure to control gastric Helicobacter pylori infection and consequent progression of chronic atrophic gastritis. The Epstein-Barr virus may also play a role in gastric cancer (IARC, 1997). Wiskott-Aldrich syndrome (WAS) is due to a mutation in the WAS gene on chromosome Xp11.23 that impairs T and B cell function. Cases present with thrombocytopenia, bloody diarrhea, eczema, and recurrent ear infection. Affected adolescents have a high risk of cancer, and the peak age of cancer incidence is likely to increase as patients are kept alive longer (Amiet, 1963; ten Bensel et al., 1966; Brand and Marinkovich, 1969; Cotelingham et al., 1985; Ochs and Rosen, 1999; Sullivan et al., 1994). In a large study, 13% developed cancer equating to a 100-fold increase in risk (Perry et al., 1980). Eighteen of the 21 cancers recorded were lymphoproliferative disease. Another inherited immunodeficiency affecting both T and B lymphocytes is X-linked lymphoproliferative disease (XLP), which is due to mutation of the SH2D1A gene. XLP is characterized by extreme sensitivity to EBV infection, which initiates a vigorous, uncontrolled polyclonal lymphocyte expansion. XLP is associated with greatly increased risk of Burkitt lymphoma and other lymphoproliferative disease, which occurs in 30% of affected individuals at a median age of 5 years (Purtillo et al., 1977). The tumors are almost always EBV positive and frequently involve the distal ileum; most are of B cell lineage, although T cell tumors account for 6% of cases (Schuster and Kreth, 1999). Excesses of Hodgkin lymphoma (HL) have also been found in relatives of XLP cases (Grufferman et al., 1977; Purtillo et al., 1977). The spectrum of lymphoid disorders and their molecular characteristics are reminiscent of AIDS-associated lymphomas. An entirely different mechanism is responsible for the excess of lymphoid tumors seen in the autoimmune lymphoproliferative syndrome (ALPS), which is caused by mutations in the FAS gene pathway that interfere with lymphocyte apoptosis. The abnormalities result in lymphoproliferation, chronic lymphadenopathy, splenomegaly, and autoimmunity. There is frequent transformation to lymphoid malignancy, either NHL or HL, although there are insufficient data to quantitate this risk. Other important mechanisms in cancer predisposition are the DNA breakage syndromes, which include ataxia telangiectasia (AT), Nijmegen breakage syndrome and Bloom syndrome. Inherited

Table 27–1. Inherited Immunodeficiency Disorders Associated with Malignancy Immune Defect Abnormal T and B cell function Defective antibody production DNA breakage syndromes

Disease

Underlying Genetic Abnormality

Associated Malignancies

Wiskott-Aldrich syndrome X-linked lymphoproliferative disease X-linked hyper-IgM (XHIM) Common variable immunodeficiency Ataxia telangiectasia (AT) Nijmegan breakage syndrome (NBS) Bloom syndrome (BLM)

Actin polymerization defect Tyrosine kinase defect CD40 ligand abnormality B- and T-cell defect DNA damage response Double-strand break repair DNA helicase

Lymphomas Burkitt lymphoma Lymphoma, liver, and biliary cancer Lymphoma, gastric cancer T-cell lymphomas B-cell lymphomas Acute leukemias

552

PART III: THE CAUSES OF CANCER

mutations within ATM, a gene important in DNA damage detection and apoptosis induction, gives rise to the syndrome of ataxiatelangiectasia and an increased risk of lymphoproliferative disorders (Hecht and Hecht, 1990; Khanna et al., 2000; Xu, 1999). The tumors are frequently T cell in immunophenotype and often have abnormal T cell receptor rearrangement. Experimental evidence suggests that ATM heterozygotes have increased sensitivity to radiation. Although it is not feasible to distinguish mutation carriers in the general population, heterozygotes are estimated to comprise about 1% of the population based on Hardy-Weinberg equilibrium. Relatives of AT patients appear to have excess mortality from certain cancers, with three studies indicating their breast cancer risk is increased up to sevenfold whereas their colon and rectal cancer risks are unaltered (Swift et al., 1991; Taylor et al., 1996). The ATM protein interacts with a complex of MRE11, NBS1, and Rad 50, which are members of the same pathway. Not surprisingly, mutation in any of these components results in a similar clinical pattern. Of particular interest is Nijmegan breakage syndrome, a rare disease due to a 5-bp deletion in the NBS1 gene (Weemaes et al., 1994). Like AT, NBS is associated with a predisposition to lymphoid tumors, but of B cell rather than T cell origin. Bloom syndrome is caused by mutation in the gene for the DNA helicase protein BLM (Mohaghegh and Hickson, 2001). This syndrome is characterized clinically by susceptibility to infection, chronic lung disease, diabetes mellitus, and an increased incidence of acute myeloid leukemia, acute lymphocytic leukemia, and non-Hodgkin lymphoma at an early age. It is clear that the profound changes in immune response caused by inherited single-gene defects can increase the risk of developing cancer, in particular tumors of the immune system itself. While nonHodgkin lymphoma is the most common manifestation, Hodgkin lymphoma is also in excess with cumulative incidence up to 25% (Filipovich et al., 1994). The most notable association is with an excess of EBV-related lymphoproliferative disease, although other viral-associated cancers may also be seen. As in post-transplant and AIDS-related lymphoproliferations, the primary immunodeficiencies are characterized by an inability of immunosurveillance to control the growth of virus-infected cells. Other pathological mechanisms may also operate independent of poor immunosurveillance, as defects in DNA repair and abnormalities of lymphocyte apoptosis are likewise associated with lymphoproliferative disease.

Acquired Severe Immunodeficiency The major iatrogenic cause of significant immunosuppression is immunosuppressive drug therapy given after transplantation or for other reasons. Infection with the human immunodeficiency virus (HIV) has also become a common cause of severe immunosuppression. In both settings there is a cancer excess, the tumors being predominantly aggressive variants of diffuse large B-cell lymphoma occurring at extranodal sites (Ballerini et al., 1993; Obrams and Grufferman, 1991; Shapiro et al., 1988). The control nervous system (CNS) is a characteristic site of involvement, and CNS lymphomas arising in immunodeficiency almost invariably contain EBV. The spectrum of tumors seen after transplantation suggests that viral-induced cancers are common and that impaired immunosurveillance of virally transformed cells is an important mechanism. The relative contribution of other mechanisms is difficult to ascertain. In retroviral infection, insertional mutagenesis is a theoretical mechanism that has not been firmly established. In both HIV infection and PTLD, there are aberrant T/B cell interactions, which can also alter the risk of lymphoproliferative disease. Both conditions are also associated with an excess of Kaposi sarcoma (KS), which is caused by human herpesvirus 8 (HHV-8) also known as KS-associated herpesvirus (KSHV).

Drugs That Cause Severe Immunosuppression Severe immunosuppression can be caused by several categories of medications, including alkylating drugs (such as cyclophosphamide), antimetabolites (e.g., methotrexate), drugs prescribed for preventing graft-versus-host disease (e.g., cyclosporine), and conditioning treatments for preparing a patient for bone marrow transplantation (such

as high-dose total body irradiation or VP 16). The extent of immune alteration varies from lower levels among those who have undergone splenectomy or used corticosteroids to profound immunosuppression associated with the use of cyclosporine or azathioprine. It is important to interpret cancer risks from immunosuppressive therapies in the context of their mechanisms of action and the medical conditions necessitating their use. Some of these agents—notably, the alkylators—act by directly damaging DNA. As in the primary immunodeficiencies, it may be difficult to distinguish the significance of immunosuppression from the effects of DNA damage as influences on cancer risk. It is generally believed that immunosuppression is most closely associated with increased risk of NHL, whereas the DNA damaging effects are more closely linked with increased risks of therapyrelated acute myeloid leukemia (tAML).

Post-transplant Immunosuppression Patients treated with immunosuppressive drugs subsequent to organ transplantation frequently develop post-transplantation lymphoproliferative disorders (PTLD) (Calne et al., 1979; Forman et al., 1987; Kinlen et al., 1983; Penn et al., 1971; Polson et al., 1988; Swinnen, 1999). Renal transplant recipients experience markedly increased relative risks of NHL, ranging from 20- to 59-fold excesses (Birkeland et al., 1995; Hoover and Fraumeni, 1973; Hughson et al., 1986; Kinlen et al., 1979; Opelz et al., 1993; Scherr and Mueller, 1996;). Higher risks, ranging from 48- to 336-fold, are reported among patients undergoing heart or bone marrow transplants (Anderson et al., 1978; Opelz et al., 1993; Penn, 1993). The magnitude of risk appears to be related to the level of immune suppression and risks appear to be lower among those transplanted in more recent years (Curtis et al., 1997). Patient groups at highest risk due to more immunosuppression include those who develop graft-versus-host disease and recipients of HLAmismatched or T cell depleted bone marrow transplants (Curtis et al., 1999; Swinnen et al., 1990, 1999). The PTLD can be subdivided into polyclonal disease (Cleary and Sklar, 1984; Frizzera et al., 1981; Hanto et al., 1981), which is treatable by reducing the level of immunosuppression, and true monoclonal diffuse large B-cell lymphoma. EBV is causally related to these conditions as evidenced by their sensitivity to T-cell clones that recognize EBV: Infusion of such clones can lead to complete resolution of the tumors (Purtillo et al., 1981; Reece et al., 1981). Treatment with immunosuppressive therapy in the absence of organ transplantation has also been linked with 10-fold elevated risks of NHL (Kinlen et al., 1979; Kinlen, 1985). Interpretation of this association is complicated because most patients treated with immunosuppressive drugs have autoimmune disorders, such as rheumatoid arthritis, that are independently linked with elevated risk of NHL (Kinlen, 1992; 1992b). Other malignant diseases with increased incidence after transplantation include melanoma and non-melanoma skin cancer, anogenital carcinoma, KS, and possibly some solid tumors (Ateenyi-Agaba, 1995; Fairley et al., 1994; Walder et al., 1971). The most common site of cancer in transplant recipients is the skin, with squamous cell and basal cell carcinomas accounting for the vast majority of cases (Kinlen et al., 1979; Caussy et al., 1990). Malignant melanoma and Merkel cell carcinoma are also increased in frequency (Greene et al., 1981). The incidence of these complications is proportional to the degree of immunosuppression, and is increased by solar exposure and decreased by skin pigmentation (Barr et al., 1989; Lishner et al., 1990; Lutzner et al., 1983). The anogenital cancers include cervical cancer, anal cancer, and cancers of the genitalia. Infection with the human papillomavirus (HPV) is a common etiologic factor for these tumors (Alloub et al., 1989; Austin, 1982; Caterson et al., 1984; Halpert et al., 1986). Anogenital HPV infection in immunosuppressed individuals has a high propensity for progression to preneoplastic dysplasia and carcinoma in situ, suggesting that loss of immune control of this infection plays an important role in these cancers (Fairley et al., 1994). KS incidence is also increased up to 100-fold in transplant recipients compared to the general population (Penn, 1982; Centers for Disease Control, 1981a, 1981b). The risk varies between populations (Qunibi et al., 1993, 1998), which may reflect differences in the preva-

Immunologic Factors lence of the etiologic agent, human herpesvirus 8/KS associated herpesvirus (HHV-8/KSHV; Chang et al., 1994). Post-transplant KS remains responsive to immune control, and frequently will resolve if immunosuppressive therapy is stopped (Penn, 1991; Starzl et al., 1984; Wilson et al., 1968). Other sites suggested to have an excess of cancer post-transplantation include the liver, colon and lung (Kinlen, 1996).

Acquired Immunodeficiency Syndrome and HIV Infection A number of infections depress immune responsiveness, but HIV is the only one that causes severe immunodeficiency. HIV-related acquired immunodeficiency syndrome (AIDS) is associated with increased risks of NHL, KS, and several other tumors (Biggar et al., 1984; Cote et al., 1991; Friedman-Kien et al., 1990; Gange and Jones, 1978; Goedert et al., 1995; Grulich et al., 1999; Knowles et al., 1988; Rabkin et al., 1991, 1992; Reynolds et al., 1993). In developed countries, the widespread application of antiretroviral therapies has dramatically ameliorated the incidence of KS and, to a lesser extent, AIDS-associated NHL (Pluda et al., 1993; Rabkin et al., 2001). Patients with AIDS are at very high risk of NHL, which is increased as much as 60 to more than 100 times background rates (Beral et al., 1991; Freter, 1990; Goedert et al., 1998; Kaplan et al., 1989; Rabkin et al., 1991). AIDS-related lymphomas tend to be clinically aggressive and histologically high grade, either diffuse large cell (including immunoblastic) or Burkitt/Burkitt-like (small noncleaved cell) NHL, although other B-cell subtypes may also occur (Ballerini et al., 1993; Obrams and Grufferman, 1991; Ziegler et al., 1984). Extranodal primary sites are frequently involved, particularly the CNS, which is an otherwise rare site for NHL. Primary CNS tumors are uniformly EBV positive, and EBV is found in 38% to 79% of systemic NHL in the setting of AIDS (IARC, 1997; Rosenburg et al., 1986; Saemundsen et al., 1981; Uccini et al., 1989). Genetic lesions may include c-myc rearrangements, ras mutation, and/or p53 loss/mutation (Ballerini et al., 1993; Subar et al., 1998). The risk factors for AIDS-associated NHL are poorly understood, but the absolute risk increases with age and is higher in men and in whites. In people with AIDS, the relative risk for NHL increases with duration of HIV infection and with decline in immune competence. A polymorphism in the HIV co-receptor CCR5 that protects against HIV infection in the homozygous state also partially protects against AIDSrelated NHL in heterozygotes, as AIDS patients carrying the CCR5delta 32 allele experience threefold lower risks of developing NHL (Dean et al., 1999; Rabkin et al., 1999). In addition to the risk of NHL, HIV-related immunosuppression is associated with markedly increased risk of KS, a complication of infection with human herpesvirus 8 (HHV-8, also known as Kaposi sarcoma–associated herpesvirus (KSHV; Beral et al., 1990). HHV8/KSHV is also responsible for the primary effusion lymphoma (PEL) seen almost exclusively in this population. How HHV-8 causes malignancy is actively under investigation. The HHV-8 genome has a high degree of homology with some cellular genes that are active in cell cycle regulation. These “pirated” genes include viral cyclins, anti-apoptotic factors (e.g., vFLIP, vBCL-2), viral cytokines and chemokines (e.g., vIL-6, vMIPs), and several trans-activating proteins that are also present in other herpesviruses. These genes can disrupt mitosis, interrupt apoptosis, increase angiogenesis, and block presentation of antigenic epitopes, which may contribute variably to the pathogenesis of KS (Ensoli et al., 2001). Because only a fraction of HHV-8 infected people develop KS, other factors must also be important. In HIV-infected people, HIV-tat protein may act synergistically with declining immune competence to increase risk of KS in people with AIDS. Compared with other HIV exposure groups, homosexual men have a roughly 10-fold higher risk of AIDS-related KS, which is thought to reflect their relatively higher prevalence of infection with HHV-8. However, the sensitivity, specificity, and predictive value of HHV-8 antibody tests have not been fully established, and the absolute prevalence of HHV-8 infection is uncertain in most populations. Even prior to the widespread use of highly active anti-retroviral therapy, the inci-

553

dence of KS has declined over the course of the AIDS epidemic. The fall in rates for KS among patients with AIDS may be related to decreased incidence of new HIV infections among homosexual men. Alternatively, the sexual behavior of HIV-infected homosexual men may have changed because of the AIDS epidemic, leading to a reduced prevalence of HHV-8 infection, and corresponding declines in KS. Although not designated as an AIDS-defining cancer, Hodgkin lymphoma incidence is elevated in people with AIDS, albeit with lower relative risks than for NHL or KS; over 80% of AIDS-associated HL tumors have detectable EBV (Alexander et al., 2000; Mbulaiteye et al., 2003). Increases in squamous carcinoma of the conjunctiva are seen in AIDS patients from Africa, and leiomyosarcoma incidence is specifically increased in HIV-infected children. Cervical and anal cancer are also increased in people with HIV infection, but the specific role of immunodeficiency is uncertain: people with HIV have a very high prevalence of HPV infection, the dominant risk factor for anogenital cancers (Daling et al., 1987; Frisch et al., 1993; Li et al., 1982; Mbulaiteye et al., 2003; McWhorter, 1988; Melbye et al., 1994, 1991).

Acquired Low-Level Immunosuppression An association between mild immune suppression and cancer risk is much less clear (Bebb et al., 2001; Beral and Newton, 1998; Connell et al., 1994; Hoover, 1992; Scherr and Mueller, 1996). One complicating factor is the difficulty of disentangling the immunosuppressive effects of a medication from the effects of the disease for which the medication was prescribed, particularly if the disease itself affects immune function (Cimino et al., 1988; Georgescu et al., 1997, 1999; Grardel et al., 1997; Hazelman, 1985, 1982; Jellinger et al., 1979, Kaldor et al., 1990; Kinlen, 1985; Moser et al., 1972; Neuhas et al., 1976; Pinals, 1976; Plotz et al., 1979, Ulrich et al., 1974; Uhl et al., 1974). Corticosteroids, for example, are used to treat asthma, ulcerative colitis and other autoimmune disorders. While steroids induce apoptosis in lymphocytes and increase susceptibility to infection, the importance of these effects on cancer predisposition is uncertain (Howshaw and Schwartz, 1980; Ilie et al., 1981; Klepp et al., 1978; Love and Sowa, 1975; Starzl et al., 1984). The antiepileptic drug phenytoin occasionally induces a benign lymphoproliferative syndrome, which can be misdiagnosed as a malignant lymphoma. Early reports associating phenytoin use with frank NHL (Anthony, 1970; Li et al., 1975) were refuted by subsequent studies of large numbers of treated patients demonstrating no excess risk (Clemmensen et al., 1974; Olsen et al., 1995). Use of amphetamines and/or other illicit drugs has been associated with increased risks of NHL in some studies (Abbondazo et al., 1995; Doody et al., 1992; Nelson et al., 1997), and reduced risks in others (Holly and Lele, 1999). Blood transfusion, particularly of the white blood cell component, is suspected of modulating the immune system. Initial reports of increased risk of NHL associated with receipt of blood transfusion have not been supported in larger epidemiological investigations (Adami et al., 1997; Cerhan et al., 1997; Maguire-Boston et al., 1999; Nelson et al., 1998). Cancer risks associated with surgical manipulation of the immune system have been examined in a number of settings, notwithstanding uncertainty about any immunosuppressive effects of surgical excision of lymphoid tissue (Vineis et al., 2000). In a cohort study of 383 children post-thymectomy, there was no evidence of an increase in NHL or other cancers (Johnson, 1968; Vessey et al., 1979), nor was there a cancer excess following splenectomy (Robinette and Fraumeni, 1977) or appendectomy (McVay, 1964; Moertel et al., 1974; Mueller et al., 1987; Ruuskanen et al., 1971). Tonsillectomy was associated with a slight excess of NHL in a registry-linkage study utilizing hospital records in Sweden (Liaw et al., 1997) as well as in a case-control study in Italy (La Vecchia et al., 1992), but a third study found no such relationship (Holly and Lele, 1999). Tonsillectomy with or without adenoidectomy has been inconsistently linked with increased risk of HL (Bonelli et al., 1990; Gutensohn et al., 1975; Johnson and Johnson, 1972; Shimaoka et al., 1973; Vianna et al., 1971, 1974). A small excess

554

PART III: THE CAUSES OF CANCER

of HL was seen among 55,169 Swedish patients post-tonsillectomy, with the risk more pronounced for tonsillectomy before age 12 (Liaw et al., 1997). Chronic lymphocytic leukemia (CLL) is associated with immune impairment, which is manifest clinically by an increased incidence of infective disease. CLL is also associated with an increased incidence of second primary malignancies, including lung cancer, melanoma, KS, soft tissue sarcoma and perhaps nonmelanoma skin cancer (Cartwright et al., 1987, 1997; Cuttner, 1992; Davis et al., 1987; Hisada et al., 2001; Goldin et al., 1999; Greene et al., 1978; Sgambati et al., 2001). The specific role of CLL-induced immune dysfunction is obscure, as affected patients may be more intensively screened for second primary malignancies due to their close medical supervision, and they are treated with agents such as alkylating drugs which may themselves be carcinogenic, either directly or through immune suppression. Chronic renal failure has been suggested to impair cellular immunity, which may be further compromised by dialysis (Ishikawa, 1987; Kinlen et al., 1980; MacDougall et al., 1987). A number of studies have suggested that a variety of cancers are increased in this setting (Matas et al., 1975; Sheil, 1991). Three follow-up studies of dialysis patients found an excess of NHL, but not other malignancies (Herr et al., 1979; Kinlen et al., 1980; Slifkin et al., 1977), whereas the largest and most recent study carried out in the US found no significant increase in any type of malignancy (Kantor et al., 1987). The inconsistencies among these studies may be due to differences in study design and/or temporal changes in the management of renal failure. Epidemiologic evidence links malaria coinfection with the distribution of endemic Burkitt lymphoma (IARC, 1985; Morrow et al., 1976). The underlying mechanism is uncertain, but the immune modulatory effects of chronic parasitemia offer potential explanations (IARC, 1997). Infected children have high serum levels of IgG and IgM, and adults have increased turnover of IgG, which may indicate B-cell hyperactivity. In addition, chronic malaria shifts the helper Tcell response from Th1 to Th2, which may impair cytotoxic T-cell control of EBV-infected lymphocytes. Defining and measuring minimal immune dysfunction is difficult in the context of population-based research. Alterations of immune function may be subtle, transient, and/or confounded by coexisting conditions. In contrast, variations in the genes controlling immune responsiveness may prove more robust than measures of actual function for studying immune associations with cancer risk. As exemplified by the highly polymorphic cytokine genes, inherited variability may underlie much of the differences in individual immune responsiveness. Epidemiologic evidence associating these variations with specific diseases implies a role for immune function in their pathogenesis.

Allergy Some hospital-based (Tielsch et al., 1987) and population-based case control studies (Bernstein and Ross, 1992; Doody et al., 1992; Holly and Lele, 1999) have associated a variety of allergic disorders with protection against NHL and Hodgkin lymphoma (Alderson, 1974; Bernstein and Ross, 1992; Chilvers et al., 1986; Dworwin et al., 1955; Fisherman et al., 1960; Gabriel et al., 1972; Holly and Lele, 1999; Logan and Saker, 1953; Mackay, 1966; Meers, 1973; Mills et al., 1992; Robinette and Fraumeni, 1978; Shapiro et al., 1971; Ure, 1969; Vena et al., 1985), but no such relationship was found in a hospitalbased study from Italy (La Vecchia et al., 1992). Allergic disorders were not associated with multiple myeloma in a population-based study from the United States (Koepsell et al., 1987), whereas a threefold excess risk was described in a population-based study in Canada (Gallagher et al., 1983). Furthermore, aggregated categories of allergic diathesis were not associated with multiple myeloma risk in several case-control studies (Cohen et al., 1987; Cuzick and de Stavola, 1988; Gallagher et al., 1983; Linet et al., 1987). Overall, there is little evidence that individuals with allergies have either reduced or increased risks of NHL, multiple myeloma, or other lymphoproliferative malignancies (Cooper et al., 1996).

Autoimmunity Many epidemiological studies have identified increased risks of NHL and, to a lesser extent, other lymphoproliferative disorders among patients with autoimmune diseases (Kamel et al., 1995; Leandro et al., 2001). Specific disorders that have been associated with lymphoid malignancy include the systemic diseases rheumatoid arthritis (Castor et al., 1985; Fries et al., 1985; Gridley et al., 1993; Hakulinen et al., 1985; Isomaki et al., 1979; Mellemkjaer et al., 1996; Pearce and Porter, 1986; Silman et al., 1988; Symmons, 1985), Felty’s syndrome (Gridley et al., 1994), and psoriasis (Hannuksela-Svahn et al., 2000), and the organ-specific Sjogren disease (Kassan et al., 1978; Kinlen, 1985) and Hashimoto’s thyroiditis (Fukuda et al., 1987; Goldman et al., 1990). Increased NHL risk has also been suggested for systemic lupus erythematosus (Abu-Shakra et al., 1996; Lindeman et al., 1976; Lipsmeyer, 1972; Mellemkjaer et al., 1997; Petersson et al., 1992). Myasthenia gravis has been linked in case reports with lymphoblastic lymphoma (Uner et al., 2001). The reported relative risks associated with autoimmune disorders generally range from two to fourfold in magnitude.

Inflammation Inflammation is one of the major effector arms of the immune response and results in the generation of genotoxic free radicals and reactive oxygen species, which can increase the risk of developing malignancy. Diseases in which such local inflammation appears to contribute to lymphomagenesis include Sjogren syndrome, chronic thyroiditis, chronic gastritis, and celiac sprue (Harris et al., 1967; Knowles, 2001; Swinson et al., 1983). In Sjogren syndrome, chronic inflammation of the salivary glands leads to reactive lymphoid hyperplasia that may evolve to lymphoma, with as much as 40-fold increase in relative risk (Talal and Bunim, 1964). The specificity of the relationship to inflammation is suggested by the characteristic histology of the associated types of non-Hodgkin lymphoma, which are almost invariably marginal zone lymphomas, in contrast to the follicular lymphomas constituting the majority of salivary gland lymphomas in patients without evidence of reactive lymphoid infiltration (Hyjek et al., 1988). Similarly, patients with Hashimoto and other forms of chronic thyroiditis have a very high risk for marginal zone lymphoma of the thyroid, which is an otherwise uncommon histologic subtype for extranodal lymphoma at this site (Holm et al., 1985; Hyjek and Isaacson, 1988; Kato et al., 1985; Van Krieken and Hoeve, 2000; Wotherspoon et al., 1991; Wotherspoon, 1998). Gastric inflammation induced by Helicobacter pylori infection leads to chronic gastritis and mucosal atrophy. In developed countries, these are pre-cancerous conditions associated with increased risks of gastric marginal zone lymphoma and gastric carcinoma. The so-called African enigma refers to the conspicuous rarity of these cancers in some parts of the world despite high prevalence of H. pylori. Celiac disease and its cutaneous manifestation dermatitis herpetiformis (Collin et al., 1996; Holmes et al., 1976; Leonard et al., 1983; Siguirsson et al., 1994) are associated with increased risks of T-cell lymphoma of the small bowel and skin, preferentially but not exclusively occurring at local sites of disease involvement. Schistosoma hematobium infection of the bladder is associated with excess risk of bladder carcinoma (IARC, 1999). Cohort studies of rheumatoid arthritis patients have confirmed that they have a modestly elevated risk of lymphoma, with no specific histologic subtype accounting for the excess (Baecklund et al., 1998; Baker et al., 1987; Baltus et al., 1983; Pearce and Porta, 1986). This autoimmune condition results in chronic joint inflammation and systemic circulation of reactive cells may explain the cancer association. Alternative explanations may involve abnormal T cell–B cell interactions thought to be important in the underlying pathogenesis of this disease. An association with HL as well as NHL has been noted in some studies (Mellemkjaer et al., 1996) but not others (Cibere et al., 1997). An association with multiple myeloma has been inconsistent, with significant excesses seen in 3 cohort (Hayes et al., 1997; Isomaki et al., 1978; Katusic et al., 1985) and 1 case-control study (Eriksson, 1993), while no significant excess was observed in 1 cohort study

Immunologic Factors (Prior et al., 1985) and 6 case-control studies (Cuzick and De Stavola, 1988; Doody et al., 1992; Gramenzi et al., 1991; Lewis et al., 1994; Linet et al., 1987; Pearce et al., 1986). Some of the inconsistency may reflect small underpowered investigations, and/or subjects misclassifying osteoarthritis as rheumatoid arthritis (Herrinton et al., 1996).

Chronic Antigenic Stimulation Chronic stimulation of antigen-directed (i.e., mature) B cells was previously theorized to pose a risk of malignant transformation. Epidemiological studies have evaluated this hypothesis in relation to several lymphoid cancers, including NHL (Tielsch et al., 1987), chronic lymphocytic leukemia (Linet et al., 1987; Rosenblatt et al., 1991), and multiple myeloma (Boffetta et al., 1989; Cuzick and De Stavola, 1988; Doody et al., 1992; Eriksson, 1993; Gallagher et al., 1983; Gramenzi et al., 1991; Koepsell et al., 1987; Lewis et al., 1994; Linet, 1996; Linet et al., 1987, 1988; Pearce et al., 1986; Vesterinen et al., 1993). A wide variety of purported exogenous and endogenous stimuli have been considered, including vaccinations, surgical implants, asthma desensitizing treatments, horse serum injections, allergies, inflammatory disorders, autoimmune and connective tissue diseases (Anderson et al., 1978; Comstock et al., 1975, 1978; Dworwin et al., 1971; Holly et al., 1997, 1999; Hoover, 1976; Kendrick et al., 1981; Kinlen et al., 1971, 1981; Rosenthal et al., 1961; Skegg, 1978; Snider et al., 1978; Stewart and Draper, 1971; Sutherland, 1982; Waaler, 1971). Overall, associations have not been consistent, but interpretation has been complicated by variations and poor definitions of the medical conditions, inconsistent categorization of specific types of immune dysfunction, nonstandardized interview methods, substantial proportions of proxy respondents, and multiple comparisons. Furthermore, the often long prodromal period for lymphoproliferative malignancies may have led to classifying a subject as exposed before diagnosis, when onset of a condition or use of a treatment actually followed occurrence of subclinical manifestations of cancer.

AGE-RELATED CHANGES IN IMMUNE FUNCTION Age-related changes in immune function appear to be linked with some malignant lymphoproliferations and related disorders that occur at the extremes of age. Examples include acute lymphoblastic leukemia (ALL) in children as well as chronic lymphocytic leukemia and multiple myeloma (and their pre-malignant counterparts) in the elderly (Linet and Devesa, 1991). Despite increases in knowledge about the pathogenesis of these disorders, etiologic factors remain mostly unknown, including the uncertain role of immune function in their initiation and promotion.

Cancer Risk and Childhood Exposures Pediatric ALL peaks in incidence at ages 2–4, a period when the developing immune system undergoes profound changes following exposures to infectious agents. Greaves (1997) has postulated that the common pre–B-cell leukemia represents an abnormal response to infections typically occurring in infants or very young children due to delayed exposure in western society. In support of this hypothesis, certain infections during infancy (Infante-Rivard et al., 2000; Neglia et al., 2000; Perrillat et al., 2002; Petridou et al., 1993; Smith et al., 1998; van Steensell-Moll, 1986), and specific vaccinations at particular ages (Auvinen et al., 2000; Groves et al., 1999, 2000) appear to protect against ALL. Other epidemiologic evidence includes the protective effect of social contact, as measured by daycare attendance (Ma et al., 2002; Petridou et al., 1997) and household crowding (Rosenbaum et al., 2000). A number of studies have associated being breastfed with reduced risk (Bener et al., 2001; Infante-Rivard et al., 2000; Perillat et al., 2002, Petridou et al., 1993; Shu et al., 1999; UKCCS, 2001).

555

In contrast to the extensive investigations in pediatric ALL, immune exposures have received relatively little study in adult lymphoproliferative disease. Two U.S. studies found a reduced risk of NHL with various, albeit different, vaccinations (Bernstein and Ross, 1992; Holly and Lele, 1999), whereas an Italian study found increased NHL risk associated with various chronic infectious diseases (Francheschi et al., 1989).

Lymphoproliferative Disorders in the Elderly Concomitant with structural and functional changes in the immune system, the elderly have dramatically increased incidence of several lymphoid malignancies, including multiple myeloma, follicular lymphoma, and chronic lymphocytic leukemia (Linet et al., 2001). Related precancerous abnormalities also increase with age, suggesting that progressive dysfunction of the aging immune system underlies its increasing vulnerability to malignant transformation (Salana et al., 2000). The role of environmental exposures in the etiology or progression of these clonal proliferations is undefined. Monoclonal gammopathy of unknown significance (MGUS) is a chronic asymptomatic proliferation of plasma cells that generate a serum paraprotein. The prevalence of MGUS rises with age, reaching as high as several percent among oldest age groups (Herrinton et al., 1996). The risk of progression to overt myeloma or related conditions has been estimated as 1% per year, and risk appears to remain elevated for more than 25 years (Blade et al., 1992; Gregersen et al., 2000; Herrington et al., 1993; Kyle et al., 2002). Risk factors for progression have not been identified, nor has the relationship with immunogenetic factors been clarified (Grufferman et al., 1989). Circulating lymphocytes with the t(14;18) characteristic of follicular lymphoma may be detected in up to half of apparently normal individuals (Aster et al., 1992; Indraccolo et al., 1999; Limpens et al., 1991; Lipkowitz et al., 1992; Liu et al., 1994; Ohshima et al., 1993; Segal et al., 1994). It is not yet known whether such cells indicate increased risk for subsequent lymphoma (Janz et al., 2003). A clonal proliferation of the same immunophenotype as CLL was detectable in 3.5% of adult outpatients, with males and elderly individuals having the greatest prevalence (Rawstron et al., 2000). Further studies are needed to confirm these findings, to provide information about the patterns of occurrence, and to examine any association with frank CLL.

CONCLUSIONS Experience with a wide range of immune deficient conditions indicates that loss of immune competence does not globally increase cancer incidence. Limited sets of malignancies are increased that are specific to the underlying immune abnormalities. A common feature is the prominent excess of NHL in many of these disorders. Loss of Epstein-Barr virus control appears contributory to some of these tumors, but EBV-negative lymphomas are also increased, indicating other etiologies must also be important. Several other viral-related cancers may also be increased, notably KS and anogenital carcinoma in situ, but others (e.g., liver cancer) apparently are not. The prominence of lymphoid malignancies among immunodeficiency-related tumors suggests that abnormal immune stimulation rather than immune suppression may be playing an important role. Notably, reversal of immunosuppression, either globally or by specific therapy (e.g., EBV-directed T cell infusion), can prevent and sometimes even treat secondary KS and NHL. Immune altering therapies warrant further investigation for treating and perhaps preventing lymphoid malignancies in the general population, but may not be widely applicable for the common solid tumors. References Abbondazo SL, Irey NS, Frizzera G. 1995. Dilantin-associated lymphadenopathy. Spectrum of histologic patterns. Am J Surg Pathol 19:675–686. Abu-Shakra M, Gladman DD, Urowitz MB. 1996. Malignancy in systemic lupus erythematosus. Arthritis Rheum 39(6):1050–1054.

556

PART III: THE CAUSES OF CANCER

Adami J, Nyren O, Bergstrom R, Ekbom A, McLaughlin JK, Hogman C, et al. 1997. Blood transfusion and risk of non-Hodgkin lymphoma: lack of association. Ann Intern Med 127:365–371. Alderson M. 1974. Mortality from malignant disease in patients with asthma. Lancet 2(7895):1475–1477. Alexander FE, Jarrett RF, Lawrence D, et al. 2000. Risk factors for Hodgkin’s disease by Epstein-Barr virus (EBV) status: prior infection by EBV and other agents. Br J Cancer 82:1117–1121. Alloub MI, Barr BB, McLaren KM, Smith IW, Bunney MH, Smart GE. 1989. Human papillomavirus infection and cervical intraepithelial neoplasia in women with renal allografts. BMJ 298(6667):153–156. Amiet A. 1963. Aldrich–Syndrom: Beobachtung zweier Falle. Ann Paediatr 201:315–335. Andersen E, Isager H. Pre-morbid factors in Hodgkin’s disease. II. 1978. BCGvaccination status, tuberculosis, infectious diseases, tonsillectomy, and appendectomy. Scand J Haematol 21(4):273–277. Anderson JL, Fowles RE, Bieber CP, Stinson EB. 1978. Idiopathic cardiomyopathy, age, and suppressor-cell dysfunction as risk determinants of lymphoma after cardiac transplantation. Lancet 2(8101):1174–1177. Asimov VN. 1983. Carcinogenesis and aging. Adv Cancer Res 40:365–424. Ateeyni-Agaba C. 1995. Conjunctival squamous cell carcinoma associated with HIV infection in Kampala, Uganda. Lancet 345:695–696. Austin DF. 1982. Etiological clues from descriptive epidemiology: squamous carcinoma of the rectum or anus. Natl Cancer Inst Monogr 62:89–90. Auvinen A, Hakulinen T, Groves F. 2000. Haemophilus influenzae type B vaccination and risk of childhood leukaemia in a vaccine trial in Finland. Br J Cancer 83:956–958. Baecklund E, Ekbom A, Sparén P, Feltelius N, and Klareskog L. Disease activity and risk of lymphoma in patients with rheumatoid arthritis: nested case-control study. BMJ 317:180–181. Baker GL, Kahl LE, Zee BC, Stolzer BL, Agarwal AK, Medsger TA Jr. 1987. Malignancy following treatment of rheumatoid arthritis with cyclophosphamide. Long-term case-control follow-up study. Am J Med 83(1):1–9. Baltus JA, Boersma JW, Hartman AP, Vandenbroucke JP. 1983. The occurrence of malignancies in patients with rheumatoid arthritis treated with cyclophosphamide: a controlled retrospective follow-up. Ann Rheum Dis 42(4):368–373. Barr BB, Benton EC, McLaren K, Bunney MH, Smith IW, Blessing K, Hunter JA. 1989. Human papilloma virus infection and skin cancer in renal allograft recipients. Lancet 1(8630):124–129. Bebb JR, Logan RP. 2001. Does the use of immunosuppressive therapy in inflammatory bowel disease increase the risk of developing lymphoma? Aliment Pharmacol 15(12):1843–1849. Bener A, Denic S, Galadari S. 2001. Longer breast-feeding and protection against childhood leukaemia and lymphomas. Eur J Cancer 37:234–238. Beral V, Newton R. 1998. Overview of the epidemiology of immunodeficiency–associated cancers. J Natl Cancer Inst Monogr 23:1–6. Beral V, Peterman T, Berkelman R, Jaffe H. 1991. AIDS–associated nonHodgkin lymphoma. Lancet 337:805–809. Beral V, Peterman TA, Berkelman RL, Jaffe HW. 1990. Kaposi’s sarcoma among persons with AIDS: a sexually transmitted infection? Lancet 335 (8682):123–128. Bernstein L, Ross RK. 1992. Prior medication use and health history as risk factors for non-Hodgkin’s lymphoma: preliminary results from a case-control study in Los Angeles County. Cancer Res 52(19 Suppl): 5510s–5515s. Biggar RJ, Horm J, Goedert JJ, Melbye M. 1987. Cancer in a group at risk of acquired immunodeficiency syndrome (AIDS) through 1984. Am J Epidemiol 126(4):578–586. Birkeland SA, Storm HH, Lamm LU, et al. 1995. Cancer risk after renal transplantation in the Nordic countries, 1964–1986. Int J Cancer 60:183–189. Blade J, Lopez-Guillermo A, Rozman C, Cervantes F, Salgado C, Aguilar JL, Vives-Corrons JL, Montserrat E. 1992. Malignant transformation and life expectancy in monoclonal gammopathy of undertermined significance. Br J Haematol 81:391–394. Boffetta P, Stellman SD, Garfinkel L. 1989. A case-control study of multiple myeloma nested in the American Cancer Society prospective study. Int J Cancer 43:554–559. Bonafe M, Valensin S, Gianni W, Marigliano V, Franceschi C. 2001. The unexpected contribution of immunosenescence to the leveling off of cancer incidence and mortality in the oldest old. Crit Rev Oncol Hematol 39:227–233. Bonelli L, Vitale V, Bistolfi F, Landucci M, Bruzzi P. 1990. Hodgkin’s disease in adults: association with social factors and age at tonsillectomy. A casecontrol study. Int J Cancer 15;45(3):423–427. Brand MM, Marinkovich VA. 1969. Primary malignant reticulosis of the brain in Wiskott–Aldrich syndrome. Report of cases. Arch Dis Child 44:536–542.

Burns EA, Leventhal EA. 2000. Aging, immunity, and cancer. Cancer Control 7(6):513–522. Calne RF, Rolles K, White DJG, et al. 1979. Cyclosporin A initially as the only immunosuppressant in 34 recipients of cadaveric organs: 32 kidneys, 2 pancreases and 2 livers. Lancet 2:1033–1036. Cartwright RA, Bernard SM, Bird CC, et al. 1987. Chronic lymphocytic leukaemia: case control epidemiological study in Yorkshire. Br J Cancer 56:79–82. Cartwright RA, McNally RJQ, Rowland DJ, Thomas J. 1997. The Descriptive Epidemiology of Leukaemia and Related Conditions in Parts of the United Kingdom, 1984–1993. London: Leukaemia Research Fund. Castor CW, Bull FE. 1985. Review of United States data on neoplasms in rheumatoid arthritis. Am J Med 78(1A):33–38. Caterson RJ, Furber J, Murray J, McCarthy W, Mahony JF, Sheil AG. 1984. Carcinoma of the vulva in two young renal allograft recipients. Transplant Proc 16 (2):559–561. Caussy D, Goedert JJ, Palefsky J, Gonzales J, Rabkin CS, DiGioia RA, Sanchez WC, Grossman RJ, Colclough G, Wiktor SZ, et al. 1990. Interaction of human immunodeficiency and papilloma viruses: association with anal epithelial abnormality in homosexual men. Int J Cancer 46(2):214–219. Centers for Disease Control. 1981a. Kaposi’s sarcoma and pneumocystis pneumonia among homosexual men–New York City and California. MMWR 30(25):305–308. Centers for Disease Control. 1981b. Follow up on Kaposi’s sarcoma and pneumocystis pneumonia. MMWR 30(33):409–410. Chang Y, Cesarman E, Pessin MS, Lee F, Culpepper J, Knowles DM, Moore PS. 1994. Identification of herpesvirus-like DNA sequences in AIDS-associated Kaposi’s sarcoma. Science 266(5192):1865– 1869. Chilvers C, Johnson B, Leach S, Taylor C, Vigar E. 1986. The common cold, allergy, and cancer. Br J Cancer 54(1):123–126. Cibere J, Sibley J, Haga M. 1997. Rheumatoid arthritis and the risk of malignancy. Arthrit Rheum 40:1580–1586. Cimino G, Lo Coco F, Cartoni C, et al. 1988. Immune-deficiency in Hodgkin’s disease (HD): a study of patients and healthy relatives in families with multiple cases. Eur J Cancer Clin Oncol 24:1595–1601. Cleary ML, Sklar J. 1984. Lymphoproliferative disorders in cardiac transplant recipients are multiclonal lymphomas. Lancet 2(8401):489–493. Cohen HJ, Bernstein RJ, Grufferman S. 1987. Role of immune stimulation in the etiology of multiple myeloma: a case control study. Am J Hematol 24(2):119–126. Collin P, Pukkala E, Reunala T. 1996. Malignancy and survival in dermatitis herpetiformis: a comparison with celiac disease. Gut 38:528–530. Comstock GW, Martinez I, Livesay VT. 1975. Efficacy of BCG vaccination in prevention of cancer. J Natl Cancer Inst 54(4):835–839. Comstock GW, Martinez I, Livesay VT. 1978. Efficacy of BCG vaccination in prevention of cancer: an update. J Natl Cancer Inst 60(4):785–788. Connell WR, Kamm MA, Dickson M, Balkwill AM, Ritchie JK, Lennard-Jones JE. 1994. Long–term neoplasia risk after azathioprine treatment in inflammatory bowel disease. Lancet 343(8908):1249–1252. Cooper GS, Kamel F, Sandler DP, Davey FR, Bloomfield CD. 1996. Risk of adult acute leukemia in relation to prior immune-related conditions. Cancer Epidemiol Biomarkers Prev 5:867–872. Cote TR, Howe HL, Anderson SP, Martin RJ, Evans B, Francis BJ. 1991. A systematic consideration of the neoplastic spectrum of AIDS: registry linkage in Illinois. AIDS 5(1):49–53. Cotelingham JD, Witebsky FG, Hsu SM, Blaese RM, Jaffe ES. 1985. Malignant lymphoma in patients with the Wiskott-Aldrich syndrome. Cancer Invest 3:515–522. Cunningham-Rundles C, Bodian C. 1999. Common variable immunodeficiency: clinical and immunological features of 248 patients. Clin Immunol 92:34–48. Cunningham-Rundles C, Siegal FP, Cunningham-Rundles S, Lieberman P. 1987. Incidence of cancer in 98 patients with common varied immunodeficiency. J Clin Immunol 7(4):294–299. Curtis RE, Rowlings PA, Deeg HJ, et al. 1997. Solid cancers after bone marrow transplantation. N Engl J Med 336:897–904. Curtis RE, Travis LB, Rowlings PA, Socie G, Kingma DW, Banks PM, et al. 1999. Risk of lymphoproliferative disorders after bone marrow transplantation: a multi-institutional study. Blood 94:2208–2216. Cuttner J. 1992. Increased incidence of hematologic malignancies in first–degree relatives of patients with chronic lymphocytic leukemia. Cancer Invest 10:103–109. Cuzick J, De Stavola B. 1988. Multiple myeloma—a case-control study. Br J Cancer 57:516–520. Daling JR, Weiss NS, Hislop TG, Maden C, Coates RJ, Sherman KJ, Ashley RL, Beagrie M, Ryan JA, Corey L. 1987. Sexual practices, sexually trans-

Immunologic Factors mitted diseases, and the incidence of anal cancer. N Engl J Med 317(16):973–977. Davignon L, Lemonde P, St-Pierre J, Frappier A. 1971. B.C.G. vaccination and leukaemia mortality. Lancet 1(7689):80–81. Davis JW, Weiss NS, Armstrong BK. 1987. Second cancers in patients with chronic lymphocytic leukemia. J Natl Cancer Inst 78(1):91– 94. Dean M, Jacobson LP, McFarlane G, et al. 1999. Reduced risk of AIDS lymphoma in individuals heterozygous for the CCR5-delta32 mutation. Cancer Res 59:3561–3564. Doody MM, Linet MS, Glass AG, et al. 1992. Leukemia, lymphoma, and multiple myeloma following selected medical conditions. Cancer Causes Control 3:449–456. Dworwin M, Diamond HD, Craver LF. 1955. Hodgkin’s disease and allergy. Cancer 8:128–131. Ensoli B, Sgadari C, Barillari G, Sirianni MC, Sturzl M, Monini P. 2001. Biology of Kaposi’s sarcoma. Eur J Cancer. 37(10):1251–1269. Eriksson M. 1993. Rheumatoid arthritis as a risk factor for multiple myeloma: a case-control study. Eur J Cancer 29A:259–263. Fairley CK, Sheil AG, McNeil JJ, Ugoni AM, Disney AP, Giles GG, Amiss N. 1994. The risk of ano-genital malignancies in dialysis and transplant patients. Clin Nephrol 41(2):101–105. Ferguson FG, Wikby A, Maxson P, Olsson J, Johansson B. 1995. Immune parameters in a longitudinal study of a very old population of Swedish people: a comparison between survivors and nonsurvivors. J Gerontol A Biol Sci Med Sci 50:378–382. Filipovich AH, Heinitz KJ, Robison LL, et al. 1987. The immunodeficiency cancer registry: a research resource. Am J Pediatr Hematol Oncol 9:183–184. Filipovich AH, Mathur A, Kamat D, Kersey JH, Shapiro RS. 1994. Lymphoproliferative disorders and other tumors complicating immunodeficiencies. Immunodeficiency 5:91–112. Filipovich AH, Mathur A, Kamat D, Shapiro RS. 1992. Primary immunodeficiencies: genetic risk factors for lymphoma. Cancer Res 52(19 Suppl): 5465s–5467s. Fisherman EW. 1960. Does the allergic diathesis influence malignancy? J Allergy 31:74–78. Forman SJ, Sullivan JL, Wright C, Ratech H, Racklin B, Blume KG. 1987. Epstein-Barr-virus-related malignant B cell lymphoplasmacytic lymphoma following allogeneic bone marrow transplantation for aplastic anemia. Transplantation 44(2):244–249. Freter CE. 1990. Acquired immunodeficiency syndrome-associated lymphomas. J Natl Cancer Inst Monogr (10):45–54. Friedman-Kien AE, Saltzman BR, Cao YZ, Nestor MS, Mirabile M, Li JJ, Peterman TA. 1990. Kaposi’s sarcoma in HIV-negative homosexual men. Lancet 20;335(8682):168–169. Fries JF, Bloch D, Spitz P, Mitchell DM. 1985. Cancer in rheumatoid arthritis: a prospective long-term study of mortality. Am J Med 78(1A):56–59. Frisch M, Melbye M, Moller H. 1993. Trends in incidence of anal cancer in Denmark. BMJ 306(6875):419–422. Frizzera G, Hanto DW, Gajl–Peczalska KJ, Rosai J, McKenna RW, Sibley RK, Holahan KP, Lindquist LL. 1981. Polymorphic diffuse B cell hyperplasias and lymphomas in renal transplant recipients. Cancer Res 41(11 Pt 1):4262–4279. Fukuda A, Hirohata T, Noguchi S, Ikeda M, Matsuo K, Yoshida A. 1987. Risks for malignancies in patients with chronic thyroiditis: a long-term followup study. Jpn J Cancer Res 78(12):1329–1334. Gabriel R, Dudley BM, Alexander WD. 1972. Lung cancer and allergy. Br J Clin Pract 26(5):202–204. Gallagher RP, Spinelli JJ, Elwood JM, et al. 1983. Allergies and agricultural exposure and risk factors for multiple myeloma. Br J Cancer 48:853–857. Gange RW, Jones EW. 1978. Kaposi’s sarcoma and immunosuppressive therapy: an appraisal. Clin Exp Dermatol 3(2):135–146. Ginaldi L, De Martinis M, D’Ostilio A, Marini L, Loreto MF, Quagli D. 1999. Immunological changes in the elderly. Aging (Milano) 11:281–286. Georgescu L, Paget SA. 1999. Lymphoma in patients with rheumatoid arthritis: what is the evidence of a link with methotrexate? Drug Safety 20(6):475–487. Georgescu L, Quinn GC, Schwartzman S, Paget SA. 1997. Lymphoma in patients with rheumatoid arthritis: association with the disease state or methotrexate treatment. Semin Arthritis Rheum 26(6):794–804. Goedert JJ, Cote TR. 1995. Conjunctival malignant disease with AIDS in USA. Lancet 346(8969):257–258. Goedert JJ, Cote TR, Virgo P, Scoppa SM, Kingma DW, Gail MH, Jaffe ES, Biggar RJ. 1998. Spectrum of AIDS–associated malignant disorders. Lancet 351(9119):1833–1839. Goldin LR, Sgambati M, Marti GE, et al. 1999. Anticipation in familial CLL. Am J Hum Genet 65:265–269.

557

Goldman MB, Monson RR, Maloof F. 1990. Cancer mortality in women with thyroid disease. Cancer Res 50(8):2283–2289. Gramenzi A, Buttino I, D’Avanzo B, Negri E, Franceschi S, La Vecchia C. 1991. Medical history and the risk of multiple myeloma. Br J Cancer 63(5):769–772. Grardel B, Fauquert P, Hardouin P. 1997. Malignancy in patients with rheumatoid arthritis treated with methotrexate. J Rheumatol 24(4):805–806; discussion 806–807. Greaves MF. 1997. Aetiology of acute leukaemia. Lancet 349:344–349. Greene MH, Hoover RN, Fraumeni JF Jr. 1978. Subsequent cancer in patients with chronic lymphocytic leukemia—-a possible immunologic mechanism. J Natl Cancer Inst 61(2):337–340. Greene MH, Young TI, Clark WH Jr. 1981. Malignant melanoma in renaltransplant recipients. Lancet 1(8231):1196–1199. Gregersen H, Mellemkjaer L, Salling Ibsen J, et al. 2000. Cancer risk in patients with monoclonal gammopathy of undetermined significance. Am J Hematol 63:1–6. Gridley G, Klippel JH, Hoover RN, Fraumeni JF. 1994. Incidence of cancer among men with Felty syndrome. Ann Intern Med 120:35–39. Gridley G, McLaughlin JK, Ekbom A, et al. 1993. Incidence of cancer among patients with rheumatoid arthritis. J Natl Cancer Inst 85:307–311. Groves F, Auvinen A, Hakulinen T. 2000. Haemophilus influenzae type b vaccination and risk of childhood leukaemia in a vaccine trial in Finland. Ann Epidemiol 10:474. Groves FD, Gridley G, Wacholder S, Shu XO, Robison LL, Neglia JP, Linet MS. 1999. Infant vaccinations and risk of childhood acute lymphoblastic leukaemia in the USA. Br J Cancer 81:175–178. Grufferman S, Cohen HJ, Delzell ES, et al. 1989. Familial aggregation of multiple myeloma and central nervous system diseases. J Am Geriatr Soc 37:303–309. Grufferman S, Cole P, Smith PG, et al. 1977. Hodgkin’s disease and siblings. N Engl J Med 296:248–250. Grulich AE, Wan X, Law MG, et al. 1999. Risk of cancer in people with AIDS. AIDS 13:839–843. Gutensohn N, Li FP, Johnson RE, Cole P. 1975. Hodgkin’s disease, tonsillectomy and family size. N Engl J Med 292(1):22–25. Hakulinen T, Isomaki H, Knekt P. 1985. Rheumatoid arthritis and cancer studies based on linking nationwide registries in Finland. Am J Med 78(1A):29–32. Hall MA, McGlinn E, Coakley G, Fisher SA, Boki K, Middleton D, Kaklami E, Moutsopolos H, Loughran TP, Ollier WE, Panayi GS, Lanchbury JS. 2000. Genetic polymorphisms of the IL12 p40 gene in immune-mediated disease Genes Immun 3:219–224. Halpert R, Fruchter RG, Sedlis A, Butt K, Boyce JG, Sillman FH. 1986. Human papillomavirus and lower genital neoplasia in renal transplant patients. Obstet Gynecol 68(2):251–258. Hannuksela-Svahn A, Pukkala E, Laara E, et al. 2000. Psoriasis, its treatment, and cancer in a cohort of Finnish patients. J Invest Dermatol 114: 587–590. Hanto DW, Frizzera G, Purtilo DT, Sakamoto K, Sullivan JL, Saemundsen AK, Klein G, Simmons RL, Najarian JS. 1981. Clinical spectrum of lymphoproliferative disorders in renal transplant recipients and evidence for the role of Epstein-Barr virus. Cancer Res 41(11 Pt 1):4253–4261. Harris OD, Cooke WT, Thompson H, Waterhouse JA. 1967. Malignancy in adult coeliac disease and idiopathic steatorrhoea. Am J Med 42(6): 899–912. Hayes RB, Yin SN, Dosemeci M, Li GL, Wacholder S, Travis LB, Li CY, Rothman N, Hoover RN, Linet MS. 1997. Benzene and the dose-related incidence of hematologic neoplasms in China. Chinese Academy of Preventive Medicine—-National Cancer Institute Benzene Study Group. J Natl Cancer Inst 89:1065–1071. Hazleman B. 1985. Incidence of neoplasms in patients with rheumatoid arthritis exposed to different treatment regimens. Am J Med 78(1A): 39–43. Hazleman BL. 1982. The comparative incidence of malignant disease in rheumatoid arthritics exposed to different treatment regimens. Ann Rheum Dis 41(Suppl 1):12–17. Hecht F, Hecht BK. 1990. Cancer in ataxia telangectasia patients. Cancer Genet Cytogenet 46:9–19. Herr HW, Engen DE, Hostetler J. 1979. Malignancy in uremia: dialysis versus transplantation. J Urol 121(5):584–586. Herrinton LJ, Demers PA, Koepsell TD, Weiss NS, Daling JR, Taylor JW, Lyon JL, Swanson GM, Greenberg RS. 1993. Epidemiology of the M–component immunoglobulin types of multiple myeloma. Cancer Causes Control 4:83–92. Herrinton LJ, Weiss NS, Olshan AF. 1996. Multiple myeloma. In: Schottenfeld D, Fraumeni JF, Jr., eds. Cancer Epidemiology and Prevention. New York: Oxford University Press, pp. 946–970.

558

PART III: THE CAUSES OF CANCER

Herrinton LJ. 1996. The epidemiology of monoclonal gammopathy of unknown significance: a review. Curr Top Microbiol Immunol 210: 389–395. Holly EA, Lele C. 1997. Non-Hodgkin’s lymphoma in HIV-positive and HIVnegative homosexual men in the San Francisco Bay Area: allergies, prior medication use, and sexual practices. J Acquir Immune Defic Syndr Hum Retrovirol 15:211–222. Holly EA, Lele C, Bracci P, et al. 1999. Case-control study of non-Hodgkin’s lymphoma among women and heterosexual men in the San Francisco Bay Area, California. Am J Epidemiol 150:375–389. Holm LE, Blomgren H, Lowhagen T. 1985. Cancer risks in patients with chronic lymphocytic thyroiditis. N Engl J Med 312(10):601–604. Holmes GKT, Stokes PL, Sorahan TM, Prior P, Waterhouse JAH, Cooke WT. 1976. Coeliac disease, gluten-free diet, and malignancy. Gut 17:612–619. Hoover RN, Fraumeni JF. 1973. Risk of cancer in renal–transplant recipients. Lancet 2:55–57. Hoover RN. 1976. Bacillus Calmette-Guerin vaccination and cancer prevention: a critical review of the human experience. Cancer Res 36:652–654. Hoover RN. 1992. Lymphoma risks in populations with altered immunity: a search for mechanism. Cancer Res (Suppl.) 52:5477s–5478s. Howshaw RA, Schwartz RA. 1980. Kaposi’s sarcoma after immunosuppressive therapy with Prednisone. Arch Dermatopathol 16:1280–1282. Hughson MD, Buchwald D, Fox M. 1986. Renal neoplasia and acquired cystic kidney disease in patients receiving long-term dialysis. Arch Pathol Lab Med 110(7):592–601. Hyjek E, Smith WJ, Isaacson PG. 1988. Primary B-cell lymphoma of salivary glands and its relationship to myoepithelial sialadenitis. Hum Pathol 19(7):766–776. Hyjek E, Isaacson PG. 1988. Primary B cell lymphoma of the thyroid and its relationship to Hashimoto’s thyroiditis. Hum Pathol 19(11):1315–1326. Ilie B, Brenner S, Lipitz R, Krakowski A. 1981. Kaposi’s sarcoma after steroid therapy for pemphigus foliaceus. Case report and short review of literature. Dermatologica 163(6):455–459. International Agency for Research on Cancer. 1995. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. Vol. 64: Human Papillomaviruses. Lyon: France, IARC. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. 1997. Vol. 70: EBV and KSHV/HHV–8. Lyon, France: IARC. IARC Working Group on the Evaluation of Carcinogenic Risks to Humans. 1999. Some chemicals that cause tumors of the kidney or urinary bladder in rodents, and some other substances. Lyon, France: IARC. Infante-Rivard C, Fortier I, Olson E. 2000. Markers of infection, breast-feeding and childhood acute lymphoblastic leukaemia. Br J Cancer 83:1559– 1564. Ishikawa I. 1987. Development of adenocarcinoma and acquired cystic disease of the kidney in hemodialysis patients. Princess Takamatsu Symp 18:77–86. Isomaki H, Hakulinen T, Joutsenlahti U. 1979. Lymphoma and rheumatoid arthritis. Lancet 1(8112):392. Isomaki HA, Hakulinen T, Joutsenlahti U. 1978. Excess risk of lymphomas leukemia and myeloma in patients with rheumatoid arthritis. J Chron Dis 31:691–696. Janz S, Potter M, Rabkin CS. 2003. Lymphoma- and leukaemia-associated chromosomal translocations in healthy individuals. Genes Chromosomes Cancer 36:211–223. Jellinger K, Kothbauer P, Weiss R, Sunder-Plassmann E. 1979. Primary malignant lymphoma of the CNS and polyneuropathy in a patient with necrotizing vasculitis treated with immunosuppression. J Neurol 220(4):259–268. Johnson SK, Johnson RE. 1972. Tonsillectomy history in Hodgkin’s disease. N Engl J Med 287(22):1122–1125. Johnson S. 1968. Effect of thymectomy on the induction of skin tumors by dibenzanthracene, and of breast tumors by dimethylbenzanthracene, in mice of the IF strain. Br J Cancer 22(4):755–761. Kafuko GW, Burkitt DP. 1970. Burkitt’s lymphoma and malaria. Int J Cancer 6(1):1–9. Kaldor JM, Day NE, Pettersson F, et al. 1990b. Leukemia following chemotherapy for ovarian cancer. N Engl J Med 322:1–6. Kamel OW, van de Rijn M, Hanasono MM, Warnke RA. 1995. Immunosuppression-associated lymphoproliferative disorders in rheumatic patients. Leuk Lymphoma 16(5–6):363–368. Kantor AF, Hoover RN, Kinlen LJ, McMullan MR, Fraumenti JF Jr. 1987. Cancer in patients receiving long–term dialysis treatment. Am J Epidemiol 126(3):370–376. Kaplan LD, Abrams DI, Feigal E, McGrath M, Kahn J, Neville P, Ziegler J, Volberding PA. 1989. AIDS–associated non–Hodgkin’s lymphoma in San Francisco. JAMA 261(5):719–724. Kassan SS, Thomas TL, Moutsopoulos HM, et al. 1978. Increased risk of lymphoma in sicca syndrome. Ann Intern Med 89:888–892.

Kato I, Tajima K, Suchi T, Aozasa K, Matsuzuka F, Kuma K, Tominaga S. 1985. Chronic thyroiditis as a risk factor of B-cell lymphoma in the thyroid gland. Jpn J Cancer Res 76(11):1085–1090. Katusic S, Beard CM, Kurland LT, Weis JW, Bergstralh E. 1985. Occurrence of malignant neoplasms in the Rochester, Minnesota, rheumatoid arthritis cohort. Am J Med 78(1A):50–55. Kendrick MA, Comstock GW. 1981. BCG vaccination and the subsequent development of cancer in humans. J Natl Cancer Inst 66(3):431–437. Khanna KK. 2000. Cancer risk and the ATM gene: a continuing debate. J Natl Cancer Inst 92:795–802. Kinlen LJ. 1985. Incidence of cancer in rheumatoid arthritis and other disorders after immunosuppressive treatment. Am J Med (Suppl 1A): 44–49. Kinlen L. 1992a. Immunosuppressive therapy and acquired immunological disorders. Cancer Res 52(19 Suppl):5474s–5476s. Kinlen LJ. 1982. Immunosuppressive therapy and cancer. Cancer Surv 1:565–583. Kinlen LJ. 1992b. Malignancy in autoimmune diseases. J Autoimmun 5(Suppl A):363–371. Kinlen LJ, Doll R, Peto J. 1983. The incidence of tumors in human transplant recipients. Transplant Proc 15:1039–1042. Kinlen LJ, Sheil AG, Peto J, Doll R. 1979. Collaborative United Kingdom–Australasian study of cancer in patients treated with immunosuppressive drugs. Br Med J 2(6203):1461–1466. Kinlen LJ, Pike MC. 1971. BCG vaccination and leukaemia: evidence of vital statistics. Lancet 2:398–402. Kinlen LJ, Eastwood JB, Kerr DN, Moorhead JF, Oliver DO, Robinson BH, de Wardener HE, Wing AJ. 1980. Cancer in patients receiving dialysis. Br Med J 280(6229):1401–1403. Kinlen LJ, Peto J, Doll R, Sheil AG. 1981. Cancer in patients treated with immunosuppressive drugs. Br Med J (Clin Res Ed) 282(6262):474. Kinlen LJ, Webster AD, Bird AG, Haile R, Peto J, Soothill JF, Thompson RA. 1985. Prospective study of cancer in patients with hypogammaglobulinaemia. Lancet 78(1A)(8423):263–266. Kinlen LJ. 1996. Immunologic factors, including AIDS. In: Schottenfeld D, Fraumeni JF Jr., eds. Cancer Epidemiology and Prevention. 2nd ed. New York: Oxford University Press. Klepp O, Dahl O, Stenwig JT. 1978. Association of Kaposi’s sarcoma and prior immunosuppressive therapy: a 5-year material of Kaposi’s sarcoma in Norway. Cancer 42(6):2626–2630. Knowles DM, Chamulak GA, Subar M, Burke JS, Dugan M, Wernz J, Slywotzky C, Pelicci G, Dalla-Favera R, Raphael B. 1988. Lymphoid neoplasia associated with the acquired immunodeficiency syndrome (AIDS). The New York University Medical Center experience with 105 patients (1981–1986). Ann Intern Med 108(5):744–753. Knowles DM. 2001. Malignant lymphomas and lymphoid hyperplasias that occur in the ocular adnexa (orbit, conjunctiva and eyelids). In: Knowles DM, ed. Neoplastic Hematopathology. Second ed. Philadelphia: Lippincott Williams & Wilkins. Koepsell TD, Daling JR, Weiss NS, Taylor JW, Olshan AF, Lyon JL, Swanson GM, Child M. 1987. Antigenic stimulation and the occurrence of multiple myeloma. Am J Epidemiol 126(6):1051–1062. Kyle RA, Therneau TM, Rajkumar SV, Offord JR, Larson DR, Plevak MF, Melton LJ. 2002. A long–term study of prognosis in monoclonal gammopathy of undetermined significance. N Engl J Med 346:564– 569. La Vecchia C, Negri E, Francheschi S. 1992. Medical history and the risk of non-Hodgkin’s lymphoma. Cancer Epidemiol Biomarkers Prev 1: 533–536. Leandro MJ, Isenberg DA. 2001. Rheumatic diseases and malignancy—is there an association? Scand J Rheumatol 30(4):185–188. Leonard JN, Tucker WFG, Fry JS, Coulter CAE, Boylston AW, McMinn RMH, et al. 1983. Increased incidence of malignancy in dermatitis herpetiformis. Br Med J 1:16–18. Lewis DR, Pottern LM, Brown LM, Silverman DT, Hayes RB, Schoenberg JB, et al. 1994. Multiple myeloma among U.S. blacks and whites: the role of chronic antigenic stimulation. Cancer Causes Control 5:529– 539. Li FP, Osborn D, Cronin CM. 1982. Anorectal squamous carcinoma in two homosexual men. Lancet 2(8294):391. Li FP, Willard DR, Goodman R, Vawter G. 1975. Malignant lymphomas after diphenylhydantoin (dilantin) therapy. Cancer 36:1359–1362. Liaw KL, Adami J, Gridley G, et al. 1997. Risk of Hodgkin’s disease subsequent to tonsillectomy: a population-based cohort study in Sweden. Int J Cancer 72:711–713. Limpens J, de Jong D, van Krieken JH, et al. 1991. Bcl-2/JH rearrangements in benign lymphoid tissues with follicular hyperplasia. Oncogene 6:2271–2276.

Immunologic Factors Lindeman RD, Pederson JA, Matter BJ, Laughlin LO, Mandal AK. 1976. Long–term azathioprine-corticosteroid therapy in lupus nephritis and idiopathic nephrotic syndrome. J Chronic Dis 29(3):189–204. Linet MS. 1996. Is chronic antigenic stimulation etiologically related to multiple myeloma? In: Schottenfeld DM, Fraumeni JF, eds. Cancer Epidemiology and Prevention. 2nd ed. New York: Oxford University Press, pp. 841–892. Linet MS, Devesa SS. 1991. Descriptive epidemiology of childhood leukemia. Br J Cancer 63:424–429. Linet MS, Devesa SS. 2001. Epidemiology of leukemia: overview and patterns of occurrence. In: Henderson ES, Lister TA, Greaves MF, eds. Leukemia. 7th ed. Philadelphia: WB Saunders. Linet MS, Harlow SD, McLaughlin JK. 1987. A case-control study of multiple myeloma in whites: chronic antigenic stimulation, occupation, and drug use. Cancer Res 47:2978–2981. Linet MS, McCaffrey LD, Humphrey RL, et al. 1986. Chronic lymphocytic leukemia and acquired disorders affecting the immune system: a casecontrol study. J Natl Cancer Inst 77:371–378. Linet MS, McLaughlin JK, Harlow SD, et al. 1988. Family history of autoimmune disorders and cancer in multiple myeloma. Int J Epidemiol 17:512–513. Linet MS, Van Natta ML, Brookmeyer R, et al. 1989. Familial cancer history and chronic lymphocytic leukemia. A case-control study. Am J Epidemiol 130:655–664. Lipsmeyer EA. 1972. Development of malignant cerebral lymphoma in a patient with systemic lupus erythematosus treated with immunosuppression. Arthritis Rheum 15(2):183–186. Lishner M, Patterson B, Kandel R, Fyles G, Curtis JE, Meharchand J, Minden MD, Messner HA. 1990. Cutaneous and mucosal neoplasms in bone marrow transplant recipients. Cancer 65(3):473–476. Logan J, Saker D. 1953. The incidence of allergic disorders in cancer. NZ Med 52:210–212. Love RR, Sowa JM. 1975. Myelomonocytic leukaemia following cyclophosphamide therapy of rheumatoid disease. Ann Rheum Dis 34(6):534– 535. Lutzner MA, Orth G, Dutronquay V, Ducasse MF, Kreis H, Crosnier J. 1983. Detection of human papillomavirus type 5 DNA in skin cancers of an immunosuppressed renal allograft recipient. Lancet 2(8347):422–424. Ma X, Buffler PA, Selvin S, Matthay KK, Wiencke JK, Wiemels JL, Reynolds P. 2002. Daycare attendance and risk of childhood acute lymphoblastic leukaemia. Br J Cancer 86:1419–1424. MacDougall ML, Welling LW, Wiegmann TB. 1987. Renal adenocarcinoma and acquired cystic disease in chronic hemodialysis patients. Am J Kidney Dis 9(2):166–171. Mackay WD. 1966. The incidence of allergic disorders and cancer. Br J Cancer 20(3):434–437. Maguire-Boston EK, Suman V, Jacobsen SJ, et al. 1999. Blood transfusion and risk of non-Hodgkin’s lymphoma. Am J Epidemiol 149:1113–1118. Malaguarnera L, Ferlito L, Di Mauro S, Imbesi RM, Scalia G, Malaguarnera M. 2001. Immunosenescence and cancer: a review. Arch Gerontol Geriatr 32(2):77–93. Matas AJ, Simmons RL, Kjellstrand CM, Buselmeier TJ, Najarian JS. 1975. Increased incidence of malignancy during chronic renal failure. Lancet 1(7912):883–886. Mbulaiteye SM, Parkin DM, Rabkin CS. 2003. Epidemiology of AIDS-related malignancies an international perspective. Hematol Oncol Clin North Am 17:673–696. McVay JR, Jr. 1964. The appendix in relation to neoplastic disease. Cancer 17:929–937. McWhorter WP. 1988. Allergy and risk of cancer. A prospective study using NHANESI followup data. Cancer 62(2):451–455. Meers PD. 1973. Allergy and cancer. Lancet 1(7808):884–885. Melbye M, Cote TR, Kessler L, Gail M, Biggar RJ. 1994. High incidence of anal cancer among AIDS patients. The AIDS/Cancer Working Group. Lancet 343(8898):636–639. Melbye M, Sprogel P. 1991. Aetiological parallel between anal cancer and cervical cancer. Lancet 338(8768):657–659. Mellemkjaer L, Andersen V, Linet MS, Gridley G, Hoover RN, Olsen JH. 1997. Non-Hodgkin’s lymphoma and other cancers among a cohort of patients with systemic lupus erythematosus. Arthritis Rheumatism 40:761–768. Mellemkjaer L, Linet MS, Gridley G, Frisch M, Moller H, Olsen JH. 1996. Rheumatoid arthritis and cancer risk. Eur J Cancer 32A:1753–1757. Mills PK, Beeson WL, Fraser GE, Phillips RL. 1992. Allergy and cancer: organ site-specific results from the Adventist Health Study. Am J Epidemiol 1992 136(3):287–295. Moertel CG, Nobrega FT, Elveback LR, Wentz JR. 1974. A prospective study of appendectomy and predisposition to cancer. Surg Gynecol Obstet 138:549–553.

559

Mohaghegh P, Hickson ID. 2001. DNA helicase deficiencies associated with cancer predisposition and premature ageing disorders. Hum Mol Genet 10:741–746. Morrow RH, Kisuule A, Pike MC, Smith PG. 1976. Burkitt’s lymphoma in the Mengo Districts of Uganda: epidemiologic features and their relationship to malaria. J Natl Cancer Inst 56(3):479–483. Moser KD, Araoz CA, Smith PL. 1972. Immunosuppressive therapy and reticulum cell sarcomas of the brain. A report of 2 cases. Panminerva Med 14(12):436–438. Mueller N, Swanson GM, Hsieh CC, Cole P. 1987. Tonsillectomy and Hodgkin’s disease: results from companion population-based studies. J Natl Cancer Inst 78(1):1–5. Neglia JP, Linet MS, Shu XO, Severson RK, Potter JD, Mertens AC, Wen W, Kersey JH, Robison LL. 2000. Patterns of infection and day care utilization and risk of childhood acute lymphoblastic leukaemia. Br J Cancer 82:234–240. Nelson RA, Levine AM, Bernstein L. 1998. Blood transfusions and the risk of intermediate- or high-grade non-Hodgkin’s lymphoma. J Natl Cancer Inst 90:1742–1743. Nelson RA, Levine AM, Marks G, Bernstein L. 1997. Alcohol, tobacco and recreational drug use and the risk of non-Hodgkin’s lymphoma. Br J Cancer 76:1532–1537. Neuhas K, Thorhurst J, Bertol O, Speck B. 1976. Multiple neoplasms in a case of rheumatoid arthritis treated with azathioprine. Praxis 10:2131. Obrams GI, Grufferman S. 1991. Epidemiology of HIV associated nonHodgkin’s lymphoma. Cancer Surv 10:91–102. Ochs HD, Rosen FS. 1999. The Wiskott-Aldrich syndrome. In: Ochs HD, Smith CIE, Puck JM, eds. Primary Immunodeficiency Diseases. A molecular and genetic approach. New York: Oxford University Press. Olsen JH, Schulgen G, Boice JD Jr, Whysner J, Travis LB, Williams GM, Johnson FB, McGee JO. 1995. Antiepileptic treatment and risk for hepatobiliary cancer and malignant lymphoma. Cancer Res 55:294–297. Opelz G, Henderson R, for the Collaborative Transplant Study. 1993. Incidence of non-Hodgkin’s lymphoma in kidney and heart transplant recipients. Lancet 342:1514–1516. Pearce N, Porta M. 1986. Association of non-Hodgkin’s lymphoma with rheumatoid arthritis. Am J Med 81:747–748. Pearce NE, Smith AH, Howard JK, Sheppard RA, Giles HJ, Teague CA. 1986. Case-control study of multiple myeloma and farming. Br J Cancer 54:493–500. Penn I. 1982. The occurrence of cancer in immune deficiencies. Curr Probl Cancer 6:1–64. Penn I. 1991. Cancer in the immunosuppressed organ recipient. Transplant Proc 23:1771–1772. Penn I. 1993. Tumors after renal and caridoac transplantation. Hematol Oncol Clin North Am 7:431–445. Penn I, Halgrimson CG, Starzl TE. 1971. De novo malignant tumors in organ transplant recipients. Transplant Proc 3(1):773–778. Peto R, Row FJC, et al. 1975. Cancer and aging in mice and men. Br J Cancer 32:411–426. Petridou E, Kassimos D, Kalmanti M, Kosmidis H, Haidas S, Flytzani V, Tond D, Trichopoulos D. 1993. Age of exposure to infections and risk of childhood leukaemia. Br Med J 307:774. Perrillat F, Clavel J, Auclerc MF, Baruchel A, Leverger G, Nelken B, Philippe N, Schaison G, Sommelet D, Vilmer E, Hemon D. 2002. Day-care, early common infections and childhood acute leukemia: a multicentre French case-control study. Br J Cancer 86:1064–1069. Perry GS 3rd, Spector BD, Schuman LM, Mandel JS, Anderson VE, McHugh RB, Hanson MR, Fahlstrom SM, Krivit W, Kersey JH. 1980. The Wiskott-Aldrich syndrome in the United States and Canada (1892–1979). J Pediatr 97(1):72–78. Petersson T, Pukkula E, Teppo L, Friman C. 1992. Increased risk of cancer in patients with systemic lupus erythematosus. Ann Rheum Dis 51:437–439. Petridou E, Trichopoulis D, Kalapothaki V, Pourtsidis A, Kogevinas M, Kalmanti M, Koliouskas D, Kosmidis H, Panagiotou JP, Piperopoulou F, Tzortzatou F. 1997. The risk profile of childhood leukaemia in Greece: a nationwide case-control study. Br J Cancer 76:1241–1247. Pinals RS. 1976. Azathioprine in the treatment of chronic polyarthritis: longterm results and adverse effects in 25 patients. J Rheumatol 3(2):140–144. Plotz PH, Klippel JH, Decker JL, Grauman D, Wolff B, Brown BC, Rutt G. 1979. Bladder complications in patients receiving cyclophosphamide for systemic lupus erythematosus or rheumatoid arthritis. Ann Intern Med 91(2):221–223. Pluda JM, Venzon DJ, Tosato G, Lietzau J, Wyvill K, Nelson DL, et al. 1993. Parameters affecting the development of non–Hodgkin’s lymphoma in patients with severe human immunodeficiency virus infection receiving antiretroviral therapy. J Clin Oncol 11:1099–1107.

560

PART III: THE CAUSES OF CANCER

Polson RJ, Neuberger J, Forman D, Calne R, Williams R. 1988. De novo malignancies after liver transplantation. Transplant Proc 20(1):94–97. Prior P. 1985. Cancer and rheumatoid arthritis: epidemiologic considerations. Am J Med 78(1A):15–21. Purtilo DT, DeFlorio D Jr, Hutt LM, Bhawan J, Yang JP, Otto R, Edwards W. 1977. Variable phenotypic expression of an X–linked recessive lymphoproliferative syndrome. N Engl J Med 297(20):1077–1081. Purtilo DT, Sakamoto K, Saemundsen AK, Sullivan JL, Synnerholm AC, Anvret M, Pritchard J, Sloper C, Sieff C, Pincott J, Pachman L, Rich K, Cruzi F, Cornet JA, Collins R, Barnes N, Knight J, Sandstedt B, Klein G. 1981. Documentation of Epstein-Barr virus infection in immunodeficient patients with life-threatening lymphoproliferative diseases by clinical, virological, and immunopathological studies. Cancer Res 41(11 Pt 1): 4226–4236. Qunibi W, Al-Fyrayh O, Almeshari K, Lin SF, Sun R, Heston L, Ross D, Rigsby M, Miller G. 1998. Serologic association of human herpesvirus eight with posttransplant Kaposi’s sarcoma in Saudi Arabia. Transplantation 65:583–585. Qunibi WY, Barri Y, Alfurayh O, Almeshari K, Khan B, Taher S, Sheth K. 1993. Kaposi’s sarcoma in renal transplant recipients: a report on 26 cases from a single institution. Transplant Proc 251402–251405. Rabkin CS, Biggar RJ, Horm JW. 1991. Increasing incidence of cancers associated with the immunodeficiency virus epidemic. Int J Cancer 47(5):692–696. Rabkin CS, Biggar RJ, Melbye M, Curtis RE. 1992. Second primary cancers following anal and cervical carcinoma: evidence of shared etiologic factors. Am J Epidemiol 136(1):54–58. Rabkin CS, Blattner WA. 1991. HIV infection and cancers other than nonHodgkin lymphoma and Kaposi’s sarcoma. Cancer Surv 10:151–160. Rabkin CS. 2001. Aids and cancer in the era of highly active antiretroviral therapy (HAART). Eur J Cancer 37:1316–1319. Ramesh N, Geha RS, Notarangelo LD. 1999. CD40 Ligand and the hyper-IgM syndrome. In: Ochs HD, Smith CIE, Puch JM, eds. Primary Immunodeficiency Diseases. A molecular and genetic approach. New York: Oxford University Press. Reece ER, Gartner JG, Seemayer TA, Joncas JH, Pagano JS. 1981. Epstein-Barr virus in a malignant lymphoproliferative disorder of B-cells occurring after thymic epithelial transplantation for combined immunodeficiency. Cancer Res 41(11 Pt 1):4243–4247. Reynolds P, Saunders LD, Layefsky ME, Lemp GF. 1993. The spectrum of acquired immunodeficiency syndrome (AIDS)-associated malignancies in San Francisco, 1980–1987. Am J Epidemiol 137(1):19–30. Robinette CD, Fraumeni JF Jr. 1978. Asthma and subsequent mortality in World War II veterans. J Chronic Dis 31(9–10):619–624. Robinette CD, Fraumeni JF Jr. 1977. Splenectomy and subsequent mortality in veterans of the 1939–45 war. Lancet 2:127–129. Rosenbaum PF, Buck GM, Brecher ML. 2000. Early child–care and preschool experiences and the risk of childhood acute lymphoblastic leukaemia. Am J Epidemiol 152:1136–1144. Rosenblatt KA, Koepsell TD, Daling JR, et al. 1991. Antigenic stimulation and the occurrence of chronic lymphocytic leukemia. Am J Epidemiol 134:22–28. Rosenburg NL, Hochberg FH, Miller G, et al. 1986. Primary central nervous system lymphoma related to Epstein-Barr virus in a patientwith acquired immunodeficiency syndrome. Ann Neurol 20:98–102. Rosenthal SR, Loewinsohn E, Graham ML, et al. 1961. BCG vaccination against tuberculosis in Chicago: a 20 year study statistically analysed. Paediatrics 28:622–641. Ruuskanen O, Vanha-Perrula T, Kouvalainen K. 1971. Tonsillectomy, appendicectomy and Hodgkin’s disease. Lancet 1:1127–1128. Saemundsen AK, Purtilo DT, Sakamoto K, Sullivan JL, Synnerholm AC, Hanto D, Simmons R, Anvret M, Collins R, Klein G. 1981. Documentation of Epstein-Barr virus infection in immunodeficient patients with lifethreatening lymphoproliferative diseases by Epstein-Barr virus complementary RNA/DNA and viral DNA/DNA hybridization. Cancer Res 41(11 Pt 1):4237–4242. Salana R, Mariani E. 2000. NK and NK/T cells in human senescence. Vaccine 18:1613–1620. Scherr PA, Mueller NE. 1996. Non–Hodgkin’s lymphomas. In: Schottenfeld D, Fraumeni JF, eds. Cancer Epidemiology and Prevention. Second ed. New York: Oxford University Press. Schuster V, Kreth HW. 1999. X–Linked lymphoproliferative disease. In: Ochs HD, Smith CIE, Puck JM, eds. Primary Immunodeficiency Diseases. A molecular and genetic approach. New York: Oxford University Press. Sgambati M, Linet M, Devesa SS. 2001. Epidemiology of chronic lymphocytic leukemia. In: Cheson BD, ed. Chronic Lymphocytic Leukemia: Scientific Advances and Clinical Developments. New York: Dekker, pp. 33–62.

Shapiro S, Heinonen OP, Siskind V. 1971. Cancer and allergy. Cancer 28(2): 396–400. Shapiro RS, McClain K, Frizzera G, Gajl–Peczalska KJ, Kersey JH, Blazar BR, Arthur DC, Patton DF, Greenberg JS, Burke B, et al. 1988. EpsteinBarr virus associated B cell lymphoproliferative disorders following bone marrow transplantation. Blood 71(5):1234–1243. Sheil AGR. 1991. Cancer Survey. In: Disney APS, ed. 14th Report of the Australia and New Zealand Dialysis and Transplants Registry, 1991. Woodville, South Australia: The Queen Elizabeth Hospital. Shimaoka K, Bross ID, Tiding J. 1973. Tonsillectomy and Hodgkins disease. N Engl J Med 288:634–635. Sigurgeirsson B, Agnarsson BA, Lindelof B. 1994. Risk of lymphoma in patients with dermatitis herpetiformis. Br Med J 308:13–15. Silman AJ, Petrie J, Hazleman B, Evans SJ. 1988. Lymphoproliferative cancer and other malignancy in patients with rheumatoid arthritis treated with azathioprine: a 20 year follow up study. Ann Rheum Dis 47:988–992. Skegg DCG. 1978. BCG vaccination and the incidence of lymphoma and leukaemia. Int J Cancer 21:18–21. Slifkin RF, Goldberg J, Neff MS, Baez A, Mattoo N, Gupta S. 1977. Malignancy in end–stage renal disease. Trans Am Soc Artif Intern Organs 23:34–40. Smith MA, Simon R, Strickler HD, et al. 1998. Evidence that childhood acute lymphoblastic leukemia is associated with an infectious agent linked to hygiene conditions. Cancer Causes Control 9:285–298. Sneller MC, Strober W, Eisenstein E, et al. 1993. NIH conference. New insights into common variable immunodeficiency. Ann Intern Med 118: 720–730. Snider DE, Comstock GW, Martinez I, Caras GJ. 1978. Efficacy of BCG vaccination in prevention of cancer: an update. J Natl Cancer Inst 60: 785–788. Solana R, Mariani E. 2000. NK and NK/T cells in human senescence. Vaccine 18:1613–20. Starzl TE, Nalesnik MA, Porter KA, Ho M, Iwatsuki S, Griffith BP, Rosenthal JT, Hakala TR, Shaw BW Jr, Hardesty RL, et al. 1984. Reversibility of lymphomas and lymphoproliferative lesions developing under cyclosporin-steroid therapy. Lancet 1(8377):583–587. Stewart A, Draper G. 1971. B.C.G. vaccination and leukaemia mortality. Lancet 1(7703):799. Subar M, Neri A, Inghirami G, Knowles DM, Dalla–Favera R. 1988. Frequent c-myc oncogene activation and infrequent presence of Epstein-Barr virus genome in AIDS-associated lymphoma. Blood 72(2):667–671. Sullivan KE, Mullen CA, Blaese RM, Winkelstein JA. 1994. A multiinstitutional survey of the Wiskott-Aldrich syndrome. J Pediatr 125: 876–885. Sutherland I. 1982. BCG and vole bacillus vaccination in adolescence and mortality from leukaemia. Stat Med 1(4):329–335. Swift M, Morrell D, Massey RB, Chase CL. 1991. Incidence of cancer in 161 families affected by ataxia-telangiectasia. N Engl J Med 325:1831–1836. Swinnen LJ, Costanzo, Nordin MR, et al. 1990. Increased incidence of lymphoproliferative disorders after immunosuppression with the monoclonal antibody OKT3 in cardiac transplant recipients. N Engl J Med 323:1723–1728. Swinnen LJ. 1999. Overview of posttransplant B-cell lymphoproliferative disorders. Semin Oncol 26(5 Suppl):21–25. Swinson CM, Slavin G, Coles EC, Booth CC. 1983. Coeliac disease and malignancy. Lancet 1(8316):111–115. Symmons DP. 1985. Neoplasms of the immune system in rheumatoid arthritis. Am J Med 78(1A):22–28. Talal N, Bunim JJ. 1964. The development of malignant lymphoma in the course of sjogren’s syndrome. Am J Med 36:529–540. Taylor AM, Metcalfe JA, Thick J, Mak YF. 1996. Leukemia and lymphoma in ataxia telangiectasia. Blood 87:423–438. ten Bensel RW, Stadlan EM, Krivit W. 1966. The development of malignancies in the course of the Aldrich syndrome. J Paediatr 68:761–767. Tielsch JM, Linet MS, Szklo M. 1987. Acquired disorders affecting the immune system and non–Hodgkin’s lymphoma. Prev Med 16:96–106. Uccini S, Monardo F, Ruco LP, Baroni CD, Faggioni A, Agliano AM, Gradilone A, Manzari V, Vago L, Costanzi G, et al. 1989. High frequency of EpsteinBarr virus genome in HIV-positive patients with Hodgkin’s disease. Lancet 1(8652):1458. Uhl GS, Williams JE, Arnett FC. 1974. Intracerebral Lymphoma in a patient with central nervous system lupus on cyclophosphamise. J Rheumatol 1:282–286. UK Childhood Cancer Study Investigators. 2001. Breastfeeding and childhood cancer. Br J Cancer 85(11):1685–1694. Ulrich J, Wuthrich R. 1974. Multiple sclerosis: reticulum cell sarcoma of the nervous system in a patient treated with immunosuppressive drugs. Eur Neurol 12(2):65–78.

Immunologic Factors Uner AH, Abuli H, Engin H, Akyol A, Ruacan S, Tan E, Gullu Ik Altundag K, Guler N. 2001. Myasthenia gravis and lymphoblastic lymphoma involving thethymus: a rare association. Leuk Lymph 42: 527–531. Ure DM. 1969. Negative assoication between allergy and cancer. Scott Med J 14(2):51–54. Van Krieken JH, Hoeve MA. 2000. Epidemiological and prognostic aspects of gastric MALT–lymphoma. Recent Results Cancer Res 156:3–8. van Steensel-Moll HA, Valkenburg HA, van Zanen GE. 1986. Childhood leukemia and infectious diseases in the first year of life: a register-based case-control study. Am J Epidemiol 124:590–594. Vanasse GJ, Concannon P, Willerford DM. 1999. Regulated genomic instability and neoplasia in the lymphoid lineage. Blood 94:3997–4010. Vena JE, Bona JR, Byers TE, Middleton E Jr, Swanson MK, Graham S. 1985. Allergy-related diseases and cancer: an inverse association. Am J Epidemiol 122(1):66–74. Vessey MP, Doll R, Norman-Smith B, Hill ID. 1979. Thymectomy and cancer: a further report. Br J Cancer 39:193–195. Vesterinen E, Pukkala E, Timonen T, Aromaa A. 1993. Cancer incidence among 78,000 asthamatic patients. Int J Epidemiol 22:976–982. Vianna NJ, Greenwald P, Polan A, Keogh MD, Davies JN. 1974. Letter: Tonsillectomy and Hodgkin’s disease. Lancet 2(7873):168–169. Vianna NJ, Greenwald P, Davies JN. 1971. Tonsillectomy and Hodgkin’s disease: the lymphoid tissue barrier. Lancet 1:431–432.

561

Vineis P, Crosignani P, Sacerdote C, et al. 2000. Haematopoietic cancer and medical history: a multicentre case control study. J Epidemiol Comm Health 54:431–436. Waaler HT. 1971. BCG and mortality. Lancet 2:1314. Walder BK, Robertson MR, Jeremy D. 1971. Skin cancer and immunosuppression. Lancet 2(7737):1282–1283. Weemaes CM, Smeets DF, van der Burgt CJ. 1994. Nijmegen breakage syndrome: a progress report. Int J Radiat Biol 66:S185–S188. Wilson RE, Hager EB, Hampers CL, et al. 1968. Immunologic rejection of human cancer transplanted with a renal allograft. N Engl J Med 278:479–483. Wotherspoon AC, Ortiz–Hidalgo C, Falzon MR, et al. 1991. Helicobacter pylori-associated gastritis and primary B-cell gastric lymphoma. Lancet 338:1175–1176. Wotherspoon AC. 1998. Gastric lymphoma of mucosa-associated lymphoid tissue and Helicobacter pylori. Annu Rev Med 49:289–299. Xu Y. 1999. ATM in lymphoid development and tumorigenesis. Adv Immunol 72:179–189. Ziegler JL, Beckstead JA, Volberding PA, Abrams DI, Levine AM, Lukes RJ, Gill PS, Burkes RL, Meyer PR, Metroka CE, et al. 1984. Non-Hodgkin’s lymphoma in 90 homosexual men. Relation to generalized lymphadenopathy and the acquired immunodeficiency syndrome. N Engl J Med 311(9):565–570.

28

Hereditary Neoplastic Syndromes NORALANE M. LINDOR, CARL J. LINDOR, AND MARK H. GREENE

E

xperience over the past several decades has demonstrated, unequivocally, that the study of rare familial clusters of both common and uncommon cancers is a remarkably productive enterprise, both scientifically and clinically. Beginning with alert clinical and laboratory observations, multidisciplinary investigations of cancerprone families have led to the identification of an ever-expanding list of clinical syndromes due to highly penetrant cancer susceptibility genes (Table 28–1). These data have had a critical impact on our understanding of the pathogenesis of the hereditary cancers at the individual, population, and laboratory levels, influencing both the management of high-risk patients as well as the study of the sporadic counterparts of the hereditary cancers. The history of the Li-Fraumeni syndrome is but one of many examples that could be cited, in which the identification of familial clusters of childhood sarcomas and breast cancer ultimately led to the identification of germ-line mutations in the p53 tumor suppressor gene as the genetic basis for this disorder, having a profound impact on the entire research spectrum, from clinical cancer genetics to the molecular biology of both inherited and sporadic cancers (Li et al., 1969, 1982, 1988; Malkin et al., 1990, 1997; Frebourg et al., 1995; Garber et al., 1997; Lee et al., 2001; Nichols et al., 2001). It is now widely accepted that most of the known hereditary cancer susceptibility genes are of the rare, highly penetrant variety, which may result in dramatic familial aggregations of malignancy, but these only account for a small minority of any particular form of cancer. However, there is growing evidence that some cancers may be due to the effects of genes with low penetrance, which do not generally produce familial aggregations. However, because the relevant alleles of these genes are relatively common in the general population, it has been estimated that they could account for a larger fraction of specific cancers than highly-penetrant genes such as BRCA1 and BRCA2 (Peto et al., 2001; Antoniou et al., 2003). The recent clinical availability of germ-line mutation testing for susceptibility genes related to the hereditary forms of such common tumors as breast, ovary, colorectum, and melanoma has served as a powerful catalyst for diverse research activities. The consequences and implications of this rapid evolution from molecular genetics to clinical cancer genetics for etiologically oriented investigators and clinicians are enormous. Time-worn strategies, such as the taking of a family history, must be reexamined to ensure that they meet the higher scientific standards now required by this new paradigm. Previously unfamiliar concepts, particularly those related to clinical genetics, must be integrated into the information base used by workers in this field, many of whom have no formal training in genetics. The need has never been greater for clinicians and epidemiologists to be well grounded in the biological and molecular basis of the diseases that they study. Laboratory research, which has been propelled forward by access to carefully annotated biological samples obtained from cancerprone families, is now challenged by an increasingly complex regulatory environment related to the ethical use of such specimens. And practitioners are being confronted by a host of new clinical issues, including those related to predictive risk assessment, genetic counseling, germ-line mutation testing for clinical decision-making, the duty to warn at-risk relatives versus their high-risk patient’s right to privacy and confidentiality, and, most importantly, the need for evidencebased, safe and effective management recommendations for high-risk individuals.

562

In this chapter, we will touch briefly upon some of these issues; their full exposition is beyond the scope of this review. We will then provide a thumbnail description of selected hereditary cancer syndromes to introduce the reader to the rich set of data that has developed in this fast-moving field. We will consider only some disorders for which a Mendelian mode of inheritance has been established and for which at least one germ-line susceptibility gene has been identified. A more comprehensive list is provided in Table 28–1.

FAMILY HISTORY OF CANCER Taking a family history has long been a standard component of the complete history and physical examination. Unfortunately, as the pace and complexity of medical practice have accelerated, this decidedly low-tech but nonetheless valuable tool often receives short shrift, thereby depriving both the patient and the health care provider of information that might have a significant impact on clinical decisions and patient outcome. In one survey of 100 unselected colorectal cancer patients, the medical record contained a family history in only 46% of subjects and, of those, only 80% were accurate (Church et al., 2000). Intervention increased the rate of recorded family history to 64%. The first point, therefore, is that taking an appropriately focused family history should receive increased emphasis in medical education (Offit, 1998; Weber et al., 2001; Lynch et al., 2002; Marsh et al., 2002). Clinical features suggesting the possibility of an underlying genetic predisposition to cancer have been identified (Table 28–2). These are not infallible. Certain modes of inheritance (e.g., autosomal recessive, X-linked recessive, imprinting), or specific clinical situations (such as an hereditary female cancer susceptibility disorder segregating through the paternal bloodline) are less readily recognized using these criteria but, in general, they have proven to be clinically useful. If reported family history of cancer is to be increasingly used as both a basis for clinical decision-making and research study eligibility or outcome (a strategy that is already widely practiced), then we must evaluate the accuracy and reliability of such information. This issue has been addressed both for reported personal (Bergmann et al., 1998) and family history of cancer in general (Napier et al., 1972; Love et al., 1985; Douglas et al., 1999; Sijmons et al., 2000; Ziogas et al., 2003), as well as for probands with cancers of the breast (Theis et al., 1994; Parent et al., 1997), colon and rectum (Kerber et al., 1997; Church et al., 2000; Kaballe et al., 2001), prostate (Bratt et al., 1999; King et al., 2002), and endometrium (Ivanovich et al., 2002), as well as sarcomas, (Bondy et al., 1994; Novakovic et al., 1996), brain tumors (Airewele et al., 1998), and colorectal adenomas (Aitken et al., 1995). The following highlights can be gleaned from this literature:

• An estimate on the upper bounds of reporting accuracy may be represented by the accuracy with which patients report their own cancer, which one study documents as 79% for all sites combined, and 91%, 90%, 71%, and 53% for breast cancer, prostate cancer, uterine cancer, and melanoma, respectively (Bergmann et al., 1998). • Accuracy of reported family history among relatives, for all cancers combined, ranges from 52% to 85%, with most reports clustering in the upper end of this range. Reports by probands who are members of high-risk families are more accurate than those provided by sporadic cancer patients (Love et al., 1985; Kerber et al., 1997; Sijmons et al., 2000).

563

Hereditary Neoplastic Syndromes Table 28–1. Hereditary Cancer Predisposition Syndromes* Disorder

Inheritance

OMIM Number(s) 208900; 251260 109400 135150 210900 601607 113705 600185 160980 215400 120435; 120436; 600259; 600258; 158320; 600678 158350 148500 133700; 133701; 166000; 600209 227650; 607139; 227660; 227645; 605724; 227646; 600901; 603467; 602956; 608111 137215 606764 151623 155600; 155601; 155755; 123829; 606719 131100 171400

1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Ataxia-telangiectasia Basal cell nevus syndrome Birt-Hogg-Dube syndrome Bloom syndrome Brain tumors in infancy, familial posterior fossa Breast/ovarian cancer, hereditary (BRCA1) Breast/ovarian cancer, hereditary (BRCA2) Carney complex Chordoma, hereditary Colon cancer, hereditary non polyposis

AR AD AD AR AD AD AD AD AD AD

11. 12. 13. 14.

Cowden syndrome Esophageal cancer, with tylosis Exostosis, hereditary multiple Fanconi anemia

AD AD AD AR

15. 16. 17. 18.

Gastric cancer, hereditary diffuse Gastrointestinal stromal tumor Li-Fraumeni syndrome Melanoma, hereditary multiple

AD AD AD AD

19. Multiple endocrine neoplasia type 1 20. Multiple endocrine neoplasia type 2A, 2B, and familial medullary thyroid cancer 21. Neuroblastoma, hereditary 22. Neurofibromatosis type 1 23. Neurofibromatosis type 2 24. Paraganglioma, hereditary 25. Peutz-Jeghers syndrome 26. Polyposis, familial adenomatous 27. Polyposis, familial juvenile 28. Renal cell carcinoma, hereditary, with leiomyomas 29. Renal cell carcinoma, hereditary nonpapillary/ clear cell 30. Renal cell carcinoma, hereditary papillary 30. Retinoblastoma, hereditary 31. Rothmund-Thomson syndrome 32. Thyroid, familial non-medullary 33. Thyroid, papillary with papillary renal neoplasia 34. Tuberous sclerosis complex 35. Von-Hippel Lindau syndrome 36. Werner syndrome 37. Wilms tumor, hereditary 38. Xeroderma pigmentosum

AD AD AD AD AD AD AD AD; AR AD AD

256700 162200 101000 168000; 601650; 606373; 185470 175200 175100; 135290 174900 605839; 150800; 164860

AD

144700

AD AD AR AD AD AD AD AR AD AR

168000; 601650; 605373; 185470 180200 268400 188550; 606240 605642 191100, 191092 193300 277700 601363; 605982 278700; 133510; 278720; 278730; 278740; 278760; 278780; 278750

AD, autosomal dominant; AR, autosomal recessive. *OMIM number refers to On Line Mendelian Inheritance in Man. Further information on all syndromes can be found by accessing OMIM at http://www3.ncbi.nlm.nih.gov/Omim/.

• Accuracy is greater for: First-degree versus second-degree relatives; Breast or colon cancers versus cancers of the female pelvis, abdomen other than gastrointestinal; Siblings versus parents, children; Deceased versus living relatives; More recent versus more remote diagnoses; Female versus male probands (inconsistent).

• The accuracy of reported age at cancer diagnosis (± 5 years) ranges from 89% to 97% (Aitken et al., 1995; Parent et al., 1997; Sijmons et al., 2000). • Attempts to verify reported cancers in the family have also led to the discovery of false-positive reports, as well as previously unreported cancers. • There is a significant correlation between the number of contacts made in the course of attempting verification and the likelihood that confirmation will be obtained (Ivanovich et al., 2002; Kadan-Lottick et al., 2003).

• When verified, nearly all families in which no relatives are reported to have had cancer prove to be true negatives (Aitken et al., 1995). One epidemiologic study estimated that the sensitivity and specificity of a reported family history of cancer were 87% and 97%, respectively (Aitken et al., 1995). Patient management was reported to have been changed by the diagnosis verification process in 5% to 11% of subjects (Theis et al., 1994; Douglas et al., 1999). This proportion is thought to be relatively low as a consequence of familial breast and colon cancers (which have the highest reported accuracy rates) comprising the vast majority of high-risk families currently being evaluated. One series targeting hereditary nonpolyposis colorectal cancer (HNPCC) families found sufficient numbers of both false-positive and false-negative reports of HNPCC-related tumors among relatives that they strongly recommended against basing colonoscopic surveillance recommendations on an unverified family history (Katballe et al., 2001). Families with a factitious history of cancer (a variant of Munchausen syndrome) have been reported as well (Evans et al., 1996; Kerr et al., 1998). A recent review of the

564

PART III: THE CAUSES OF CANCER

Table 28–2. Features That Suggest the Presence of a Hereditary Cancer Predisposition In the Individual Patient • Multiple primary tumors in the same organ • Multiple primary tumors in different organs • Bilateral primary tumors in paired organs • Multifocality within a single organ • Younger than usual age at tumor diagnosis • Rare histology • In the sex not usually affected • Associated with other genetic traits • Associated with congenital defects • Associated with an inherited precursor lesion • Associated with another rare disease • Associated with cutaneous lesions known to be related to cancer susceptibility disorders (e.g., the genodermatoses)

In the Patient’s Family • One first-degree relative with the same or a related tumor and one of the individual features listed • ≥2 first-degree relatives with tumors of the same site • ≥2 first-degree relatives with tumor types belonging to a known familial cancer syndrome • ≥2 first-degree relatives with rare tumors • ≥3 relatives in two generations with tumors of the same site or etiologically related sites

Source: Modified from Weber et al. (2001).

hereditary breast/ovarian cancer literature emphasized the value of a meticulous family history in the evaluation of adult-onset cancer susceptibility syndromes that lack a distinctive, non-neoplastic phenotype (Lynch et al., 2003). These findings have prompted a number of recommendations, including (a) the suggestion that attempting to verify cancer diagnoses in third-degree relatives yields little useful information, and thus probably represents an inefficient use of resources (Douglas et al., 1999); (b) reported breast cancer is sufficiently reliable as to not require verification (Sijmons et al., 2000); (c) verification efforts should be focused on tumor sites that are vaguely defined, (e.g., the female pelvis), and sites commonly involved with metastatic disease (e.g., liver, bone, brain) (Sijmons et al., 2000). In general, the available data on first-degree relatives suggest that a negative family history, or a positive family history of breast or colon cancer, are both surprisingly accurate.

POPULATION-BASED DATA REGARDING FAMILIAL CANCERS Etiologic Research In the context of investigating familial aggregations of cancer, Goldgar (2002) provides the epidemiologist’s perspective: “. . . the natural questions are what fraction of cancer incidence can be explained by inherited susceptibility, and how does inherited susceptibility interact with the surrounding environment and individual lifestyle choices to modify cancer risk”. These questions, along with the issue regarding the magnitude of risk, both relative and absolute, can only be addressed with population-based research strategies. Time-honored studies of multiple-case families, and even well-designed epidemiologic studies that are not population-based, are constrained by a variety of practical limitations (e.g., small sample sizes for rare disorders) and serious methodologic biases related to case ascertainment, generalizability, and so forth. Population-based studies may produce results that are generalizable to the population that was studied, as well as having the potential for being large enough to produce statistically reliable risk estimates, even for relatively rare cancers. Thus, for example, a detailed analysis of family history, coupled with extensive mutation testing in a population-based cohort of women with breast cancer, permitted Peto and colleagues (1999) to estimate that only 17% of familial breast cancer risk can be explained on the basis of germ-line mutations in BRCA1 and BRCA2. The current

evidence suggests that the remaining 83% is the result of the combined action of 15 or more common, low-penetrance alleles, yet to be identified (Goldgar, 2002). Fortunately, the Utah Population Database (Goldgar et al., 1994) and the Swedish Family-Cancer Registry (Hemminki et al., 1999; Dong et al., 2001) have been explored systematically in an effort to improve the level of evidence related to these aspects of familial cancer risk. A sample of the data available from these two registries is shown in Table 28–3, which summarizes the familial relative risk (FRR) of the same cancer among first-degree relatives of probands with a specified malignancy. The pattern of risks by site is relatively similar between the Utah and the Swedish data, with FRRs in the range of 2.0 to 3.0. Compared with other cancer sites, thyroid and testicular cancers have notably higher risks in both populations, in the 8 to 12 range. The Utah data were analyzed by age at cancer diagnosis in the proband, and they show significantly increased FRRs among relatives of probands with early-onset cancer, as described above among the clinical clues to an inherited cancer susceptibility disorder. For example, the FRR for breast cancer is 1.8 overall, but 3.7 among the relatives of women whose breast cancer was diagnosed at age <50 (Table 28–3). The Swedish Family-Cancer Registry has proved to be an extraordinarily rich source of population-based data regarding a multitude of familial cancer risk issues. It is based upon two-generation families born in Sweden after 1934 and their biological parents, including more than 9.6 million individuals, 2.1 million nuclear families, and data on more than 700,000 invasive and 130,000 in situ cancers (Hemminki et al., 1999; Dong and Hemminki, 2001). The final column in Table 28–3 displays data regarding the population-attributable fraction (PAF) related to site-specific familial cancer susceptibility in Sweden (Hemminki and Czene, 2002). The PAF is the proportion of cases that

Table 28–3. Familial Relative Risks and Population Attributable Fraction of the Same Cancer Among First-Degree Relatives of Cancer Probands by Primary Cancer Site Sweden†,‡

Utah* Site Prostate Breast Colorectal Lung Uterine corpus Melanoma Urinary bladder Non-Hodgkin lymphoma Brain/CNS Ovary Stomach Pancreas Kidney Thyroid Multiple myeloma Hodgkin lymphoma Soft tissue sarcoma Testicular Total, mean Total, median

FRR (Total)

FRR§ (Early Onset)

FRR (Child)

FRR (Sibling)

PAR (%)

2.2 1.8 2.5 2.6 1.3 2.1 1.5 1.7

4.1 3.7 4.5 2.5 1.8 6.4 5.0 2.4

2.8 1.9 1.9 1.7 — 2.5 1.5 1.7

9.4 2.0 4.4 3.2 — 3.4 3.3 2.4

20.5 10.6 6.9 3.8 3.9 1.4 2.0 1.2

2.0 2.0 2.1 1.2 2.5 8.5 4.3 1.2 2.0 8.6 2.1 2.2

9.0 — — — — — — — — — 3.8 4.1

1.7 2.9 1.7 — 1.6 9.5 4.2 — — 4.3 2.1 1.9

2.4 2.5 8.8 — 5.3 12.4 5.6 — — 8.5 3.4 3.5

1.2 4.9 1.5 1.0 1.9 3.6 1.0 0.9 0.1 2.7 — —

Source: Modified from Risch (2001). CNS, central nervous system; FRR, familial relative risk; PAF = PAR, population attributable risk. Cancers listed in order of decreasing population prevalence in Utah. *Goldgar et al. (1994). † Dong and Hemminki (2001). ‡ Hemminki and Czene (2002). § Early onset: less than age 50 for melanoma, breast, and brain; less than age 60 for all others.

Hereditary Neoplastic Syndromes is exposed to the risk factor of interest (here, positive family history of a particular cancer) and represents that fraction of cases that could be prevented if the risk factor were completely eliminated. These data provide substantial support for the earlier-stated claim that, for the most part, familial and inherited factors are thought to account for a relatively small proportion of any specific malignancy. For most sites, the PAF is between 1% and 3%. Interestingly, prostate, breast, and colorectal cancer have substantially higher PAFs: 20.5%, 10.6%, and 6.9%, respectively. Hemminki and colleagues have published an extraordinary array of reports from this resource including summary studies of familial risk related to siblings and parents on offspring (Hemminki et al., 1997; Hemminki and Vaittinen, 1999; Dong et al., 2001; Hemminki and Czene, 2002), spouses (Hemminki et al., 2001), gender (Hemminki and Li, 2002), women who have borne children to multiple partners (Li et al., 2002), twins (Hemminki and Jiang, 2002), discordant cancers between parents and offspring (Vaittinen et al., 1999), environmental factors (Czene et al., 2002), cancer site and histopathology (Hemminki et al., 2003), and in situ cancers (Hemminki et al., 1998), as well as site-specific analyses targeting many specific types of cancer (Hemminki et al., 1999, 2000, 2002; Anderson et al., 2000; Hemminki and Jiang, 2001; Hemminki and Li, 2001; Plna et al., 2001; Hemminki and Czene, 2002; Czene and Hemminki, 2002, 2003). Detailed consideration of this mountain of data is beyond the scope of the current review; the interested reader is referred to the primary literature just cited. These analyses demonstrate the remarkable potential inherent in population-based studies of familial cancer risk.

Population Screening It is difficult to discuss general population issues related to hereditary cancer syndromes without at least touching upon the issue of population screening for germ-line mutations in cancer susceptibility genes. At the lay level, it is commonly (and mistakenly) assumed that the availability of clinical testing for deleterious mutations in rare, highpenetrance genes such as BRCA1, BRCA2, MLH1, MSH2, and CDKN2A means that such testing can and should be routinely applied to anyone who has a concern about familial cancer risk. However, the rarity of these mutations in the general population typically results in test performance characteristics (sensitivity, specificity, positive predictive value, etc.) that are unacceptable. General issues related to the principles underlying population screening in the postgenomic era have been reviewed (Khoury et al., 2003); not surprisingly, there are no hereditary cancer syndromes among the various illustrative examples of population-level genetic screening that they discuss. The technical issues related specifically to population screening for hereditary cancer syndromes have been reviewed in a particularly thoughtful fashion by Grann and Jacobson (2002). There is one situation in current clinical practice that may represent the first opportunity for serious consideration regarding implementing population screening that targets a specific subset of the general population, and that relates to the occurrence of ovarian cancer among women of Ashkenazi Jewish heritage. The occurrence of 2 specific mutations in BRCA1, and 1 specific mutation in BRCA2, has been shown to account for nearly 90% of all deleterious mutations in these two genes among Ashkenazi Jews (Kauf et al., 2002). These three Ashkenazi founder mutations have a combined population frequency of approximately 2.5% (Warner et al., 1999), compared with 0.08% in the general North American/European Caucasian population (Parmigiani et al., 1998). Consequently, it is much more likely that one would detect a deleterious mutation in BRCA1 or BRCA2 if one were to screen a random sample of Ashkenazim than if one were to screen the general U.S. population. However, even this relatively high prevalence rate does not yield satisfactory screening test performance characteristics. But recently, several publications have demonstrated that the prevalence of BRCA founder mutations among Jewish women with ovarian cancer, unselected for age at diagnosis or family history, ranges from 29% to 41% (Moslehi et al., 2000; Modan et al., 2001). These very high prevalence rates have led some to suggest that clini-

565

cal germ-line mutation testing be considered for all Jewish women with ovarian cancer. This view has not yet become standard of clinical care, nor is it true screening at the general population level, but this scenario provides insight into how “population screening” is likely to evolve in the world of clinical cancer genetics.

CRITICAL GENETIC TERMINOLOGY A major issue that is unavoidable when two major disciplines (in this case, epidemiology and genetics) form a critical nexus of scientific inquiry relates to the relative unfamiliarity with basic concepts that each party brings to the other’s area of expertise. The remainder of this chapter is devoted to a thumbnail summary of selected hereditary cancer susceptibility disorders. To facilitate the reader’s understanding of the many issues raised in these brief descriptions, the following introduction to a number of fundamental genetic concepts is offered. These definitions are modified from publicly available online glossaries maintained by the National Cancer Institute (http://cancer.gov/dictionary) and the National Human Genome Research Institute (http://www.genome.gov/glossary.cfm), and from a standard genetics text (Strachan and Read, 1999). Allele: One of the alternative forms of a gene (or DNA sequence) at a particular location on a chromosome. Constitutional: An abnormality (mutation) that was present in the fertilized egg. Consequently, it is present in every cell in the body. Dominant: A gene that almost always results in a specific physical characteristic, for example, a disease, even though the patient’s genome possesses only one copy. The trait is expressed in heterozygotes. If the gene is not on a sex chromosome, then the disorder is said to be autosomal. Founder effect: The presence of a specific allele in a population as a consequence of that population having its origin from a small number of ancestors (“founders”), at least one of whom carried that allele. Genetic heterogeneity: A single phenotype (e.g., hereditary nonpolyposis colorectal cancer) is caused by multiple different genes. Genome: All the DNA contained in an organism or a cell. Genotype: The genetic makeup of an individual, either overall or at a specific locus. Heritability: The proportion of the etiology of a particular trait that is due to genetic causes. Heterozygous: Possessing two different forms of a particular gene, one inherited from each parent. Homozygous: Possessing two identical forms of a particular gene, one inherited from each parent. Imprinting: The determination of whether a particular gene is expressed, by which parent transmitted the allele. This mechanism may result in a pattern of disease within a family that can be difficult to recognize as a genetic disorder. In hereditary paraganglioma, for instance, offspring of female gene carriers do not develop disease even if they inherit the causative gene mutation, whereas 50% of offspring of male gene carriers do. Thus, in this instance, the disease will be manifest only when the mutated allele is transmitted to an offspring by the male parent. Locus: The particular physical place on a chromosome where a specific gene or DNA sequence is located. The plural is loci. Microsatellite: A short series (generally less than 100 base pairs) of tandem repeats of a very simple DNA sequence (usually 1 to 4 base pairs), such as (CA)n. Mismatch repair: A normal DNA repair process, mediated by a series of enzymes that replaces a mispaired nucleotide in the DNA double helix. These abnormalities typically occur due to an error in DNA replication. Mutation: A heritable change in the DNA sequence. Types include base substitution, base deletion, base insertion, and chromosomal abnormalities. Some mutations may alter protein function; many do not. Mutations may be caused by mistakes during cell division, or they may be caused by exposure to DNA-damaging agents in the

566

PART III: THE CAUSES OF CANCER

environment. Mutations can be harmful, beneficial, or have no effect. If they occur in cells that make eggs or sperm, they can be inherited; if mutations occur in other types of cells, they are not inherited. New (de novo) germ-line mutation: The first appearance of a new, usually autosomal dominant, mutation in an individual with no affected ancestors or siblings. This individual’s offspring are at risk of inheriting the new mutation. This is one explanation for the appearance of a genetic disorder in a person with a negative family history for that disorder. Phenotype: The observable traits or characteristics of an organism: for example, hair color, weight, or the presence or absence of a disease. Penetrance: The percent frequency with which a dominant or homozygous recessive gene or gene combination manifests itself in the phenotype of the carriers. Polymorphism: A common variation or mutation in DNA. The occurrence in the same population of multiple discrete allelic states of which at least two have “high” frequency (conventionally of 1% or more). In common usage, often implies a nonpathologic sequence variant. Recessive: A genetic disorder that appears only in individuals who have received two copies of a mutant gene, one from each parent. Such individuals are said to be homozygous at that genetic locus. Somatic cell: All body cells, except the germ cells (i.e., the sperm and the egg). Tumor suppressor: A gene that normally inhibits or controls cell division. When a tumor suppressor is mutated (inactivated), it may fail to keep a cancer from growing.

SYNOPSIS OF SELECTED HEREDITARY CANCER SUSCEPTIBILITY DISORDERS Ataxia-Telangiectasia (AT) (Includes AT Complementation Groups A, C, D, E, V1/V2; Louis-Barr Syndrome) Genetics: Autosomal recessive disorder due to mutations in the ATM gene at 11q22.3. Incidence: One case in 30,000 to one in 100,000. Gene frequency in the general population is estimated at about 0.2–1.0% (Gatti, 2001). Diagnosis: The neurologic features of AT dominate the clinical picture. Telangiectasias generally begin in sun-exposed areas and occur later than the neurologic symptoms (typically after age 7). Other cutaneous features include vitiligo, cafe-au-lait macules, and premature graying of the hair. Frequent sinopulmonary infections are related to a variable degree of immunodeficiency and cellular immunodeficiency (Gatti, 2001). Associated malignant neoplasms: One-third of all AT patients will develop cancer during their shortened lives (the oldest reported patients with AT as of 1985 were 52 and 49 [Scriver et al., 1995]), and 15% die of their cancer. Eighty-five percent of the associated malignant neoplasms involve lymphoreticular tissue, especially B-cell non-Hodgkin lymphoma and lymphocytic leukemias. Adult male patients, particularly those who are IgA deficient, have a 70-fold increased risk of gastric cancer. Increased rates of medulloblastomas, basal cell carcinomas, gliomas, and uterine cancers have been reported. Associated benign neoplasms: None reported. Associated risks for heterozygous carriers of ATM: Swift et al. (1991) reported that heterozygous men and women have relative risks of developing cancer in general of 2.3 and 3.1, with excess risks of cancer mortality of 3.0 and 2.6, respectively. The controversy concerning risks for heterozygous carriers has focused on the risk of female breast cancer. The following studies have reported increased risks of breast cancer in heterozygous carriers of ATM mutations: Swift et al. (1991), Easton et al. (1994), Vorechovsky et al. (1996), Athma et al. (1996), and Stankovic et al. (1998), as well as epidemiological studies (frequently cited in these papers as showing 3.8 RR). The following studies show no statisti-

cally significant increased risk of breast cancer: Fitzgerald et al. (1997), Izatt et al. (1999), Wooster et al. (1993), and Cortessis et al. (1993). Stankovic et al. (1998) reported that specific mutations in the ATM gene could account for observed breast cancer susceptibility. In their study, a relative risk of 12.7 was found for 7271TÆG (V2424G, a missense mutation) in both homozygotes and heterozygotes. In families with this mutation, there was a less severe neurological phenotype.

Basal Cell Nevus Syndrome (BCNS), Nevoid Basal Cell Carcinoma Syndrome (NBCCS), or Gorlin Syndrome Genetics: Autosomal dominant disorder due to mutations in the PTCH gene at chromosome 9q22.3. Incidence: Estimated at ~1 in 56,000. Diagnosis: The diagnostic criteria for BCNS have been published elsewhere (Evans et al., 1993). Associated benign neoplasms: Odontogenic keratocysts of the jaw and epidermal cysts of the skin occur in the majority of cases. In a minority of cases, meningioma or cardiac fibromas (in 2%) and ovarian fibromas (in 20%) may occur. Ectopic calcification of the falx cerebri is seen in over 90% of patients over age 20 years (Gorlin, l987). Associated malignant neoplasms: Multiple basal cell carcinomas begin to appear in the third decade, although 10% of gene carriers may never develop basal cell cancers (particularly in non-white populations). Up to 5% of children develop medulloblastoma with a peak incidence around age 2 years, compared with 7 years in sporadic medulloblastomas (Cowan et al., 1997). Ovarian fibrosarcoma develops rarely.

Bloom Syndrome Genetics: Autosomal recessive disorder due to mutations in the BLM gene at 15q26.1, a RecQ-like DNA helicase. Incidence: Unknown in the general population. The condition is extremely rare. The Bloom’s Syndrome Registry is composed of 168 people who were identified as having Bloom syndrome between 1960 and 1991. It represented the majority of Bloom syndrome patients worldwide at that time. Oddoux et al. (1999) reported a frequency of the Ashkenazi founder mutation as 1 of 231 in the New York Ashkenazi. Diagnosis: Growth deficiency (pre- and postnatal) with normal body proportion, except for mild microcephaly, a sun-sensitive erythema/telangiectasia and characteristic facies; males are sterile; females have reduced fertility and a shortened reproductive span; increased risk of infection; frequent occurrence of diabetes; cafeau-lait or hypopigmented macules; characteristic high-pitched voice. Learning disabilities are frequent, but overall intellect is usually normal. Specific diagnostic testing involves demonstrating increased frequency of sister chromatid exchange (SCE) using special cytogenetic techniques. Limited genetic testing for founder mutations is now available. Associated malignant neoplasms: Increased frequency at all ages, with acute leukemia and lymphoid neoplasms predominating before the age of 25. After age 20, carcinomas of the tongue, larynx, lung, esophagus, colon, skin, breast, and cervix are most common, with the age at diagnosis often 20 or more years younger than that generally expected for each tumor type. Obligatory heterozygotes have, until recently, been said to not have any increased risk of cancers. Gruber et al. (2002) have now reported that Ashkenazi Jews with colorectal cancer were more than twice as likely to carry the BLM founder mutation than Ashkenazi Jewish controls without colorectal cancer. The Bloom’s Syndrome Registry, as of January 1, 1996, describes the cancer statistics of 168 individuals with this disorder (German et al., 2001). Associated benign neoplasms: None known.

Hereditary Neoplastic Syndromes

Breast/Ovarian Cancer, Hereditary (BRCA1) Genetics: Autosomal dominant disorder due to mutations in the BRCA1 gene on chromosome 17q21, a tumor suppressor gene that is involved with cell-cycle regulation and DNA repair. More than 1200 distinct mutations, polymorphisms, and variants have been identified. Incidence: Determining true frequency of this gene has been challenging. Overall, BRCA1 and BRCA2 account for 6–10% of all breast and ovarian cancers in patients unselected for family history, suggesting an overall carrier frequency of one of these two genes in 1 in 100 to 1 in 2500 across different populations. Szabo and King (1997) have reviewed studies that showed that, in the United States, 39% of families with three or more cases of female breast and/or ovarian cancer had identifiable BRCA1 mutations (the range was from 9% in Iceland to 79% in Russia). Sixty-four percent of families with three or more cases of female breast and/or ovarian cancer had either BRCA1 or BRCA2 mutations. The cause of cancer in the remaining third of families is unknown. Eight percent of families in the United States with male and female breast cancer had BRCA1 mutations. Mutations in BRCA1 were found in 3.3% of American women unselected for family history, diagnosed with breast cancer between the ages of 20 and 74 years (Newman et al., l998). In the Ashkenazi Jew population, 1% carry the 185delAG mutation and this single mutation accounts for 21% of breast cancers diagnosed at or before the age of 40 (Fitzgerald et al., 1996). A second Ashkenazi founder mutation, 5382insC, has also been recognized. Three Ashkenazi founder mutations combined (BRCA1 185delAG, BRCA1 5382insC, and BRCA2 6174del T) account for over 90% of the identifiable mutations in hereditary breast cancer families in that population; these mutations have a population frequency of about 1.1%, 0.1%, and 1.5%, respectively (Roa et al., 1996). Diagnosis: Suspected on the basis of premenopausal breast cancer diagnosis or a pedigree suggestive of dominant inheritance of a predisposition to breast and/or ovarian cancer. Associated malignant neoplasms: Studies based upon multiple case families (usually four or more cases of breast cancer) estimated the cumulative risk of breast cancer in BRCA1 mutation carriers as 3% by age 30 years, 19% by age 40, 51% by age 50, 54% by age 60, and 85% by age 70. More recent studies suggest the risks are not this great outside of these multiplex families with more recent estimates of lifetime risk of breast cancer of 45–68% (Hopper et al., l999; Warner et al., 1999; Anglian Breast Cancer Study Group, 2000; Risch et al., 2001; Stagopan et al., 2001). The risk of ovarian cancer appears to vary between families, with estimates ranging from 16% to 63% (Easton et al., l995). Carriers of BRCA1 may develop papillary serous carcinoma of the peritoneum, indistinguishable from serous epithelial ovarian carcinoma. It may develop years after oophorectomy and can be multifocal in origin. Fallopian tube carcinoma also occurs excessively among BRCA1 mutation carriers with relative risks as high as 120 reported (Brose et al., 2002). Ford et al. (1994) reported the relative risk of colon cancer in BRCA1 gene carriers as 4.1, or 6%, by age 70, compared with a risk of 1–2% in the general population. Subsequent studies have failed to confirm this association. The risk of prostate cancer was 3.3, or 8%, by age 70. This association is increasingly accepted as real, although BRCA-related prostate cancers do not typically demonstrate a younger-than-usual age at diagnosis (Giusti et al., 2003). Male breast cancer is occasionally reported in BRCA1 families, but the magnitude of risk is unclear. Various other cancers have been inconsistently associated with BRCA1 mutation status (Brose et al., 2002). Associated benign neoplasms: None known.

Breast/Ovarian Cancer, Hereditary (BRCA2) Genetics: Autosomal dominant disorder due to mutations in the BRCA2 gene on chromosome 13q12.3, a tumor suppressor gene involved with DNA synthesis and repair.

567

Mutations: Nearly 1400 distinct mutations, polymorphisms and variants have been reported. A founder effect involving the 999del5 mutation was found in Iceland; this mutation was present in 38% of Icelandic men with breast cancer (n = 34) and in 10.4% of women with breast cancer (n = 541). The penetrance of this mutation in women in Iceland was 17% by age 50 and 37% by age 70, which is lower than other BRCA2 reports (Thorlacius et al., 1998). In Ashkenazi Jewish women, the 6174delT mutation may be present in 8% of women diagnosed with breast cancer before the age of 42 years and in 1.52% of the Ashkenazim (Berman et al., 1996; Roa et al., 1996). Bi-allelic mutations in BRCA2 have recently been described in patients with Fanconi’s anemia (Howlett et al., 2002). It appears that BRCA2 is also the gene that had previously been designated as FANC-D1. This remarkable observation has identified an unexpected nexus between the FA and BRCA DNA repair pathways that is currently the subject of active investigation. Incidence: This varies widely between populations; BRCA1 and BRCA2 mutations may account for 6–10% of all breast and ovarian cancers unselected for family history, suggesting an overall carrier frequency of one of these two genes in 1 in 100 to 1 in 2500. In a review by Szabo and King (1997), BRCA2 mutations have been identified in 25% of American families with three or more cases of female breast and/or ovarian cancer (values range from a low of 8% in Finland to a high of 64% in Iceland). In families with male and female breast cancer, BRCA2 mutations were found in 19% of American families and in 90% of Icelandic families. Diagnosis: Suspected on the basis of premenopausal breast cancer or a family tree showing a constellation of BRCA2-associated cancers with possible dominant inheritance. The presence of male breast cancer may be a clue pointing toward the involvement of BRCA2. Associated malignant neoplasms: Adenocarcinoma of the female breast (generally estrogen receptor positive, moderately differentiated); the risk of a contralateral breast cancer by age 70 was 52.3% (95% CI, 41.7–61.0%). Male breast cancer is more common in BRCA2 families than in BRCA1 families. The cumulative probability to age 70 of male breast cancer in BRCA2 mutation carriers was approximately 6%. The risk of ovarian cancer is significantly lower than that observed in BRCA1 mutation carriers, but is still greatly increased compared with the rates in the general population; it is an important element of the BRCA2 syndrome. Risch et al. (2001) noted that ovarian cancer in BRCA2 carriers is more likely to occur after age 60 than those found in BRCA1 carriers. They also underscored their finding that the overall penetrance of BRCA2 mutation was greater for males than for females: 53% versus 38%. Fallopian tube carcinoma has also been associated with BRCA2 mutations (Aziz et al., 2001). The presence of pancreatic cancer in a breast cancer family may be a significant predictor of a BRCA2 mutation. The Breast Cancer Linkage Consortium (1999) reported on 173 families from 20 countries; they showed statistically increased risks for cancers of the prostate (RR = 4.6), pancreas (RR = 3.5), gallbladder and bile duct (RR = 5.0), stomach (RR = 2.6), and melanoma (RR = 2.6). The cumulative risk of developing ovarian cancer by age 70 was 15.9% (95% CI, 8.8–22.5%). Associated benign neoplasms: None known.

Colorectal Cancer, Hereditary Non Polyposis (HNPCC; sometimes called Lynch Syndrome and Includes Muir-Torre Syndrome) Genetics: Autosomal dominant disorder due to mutations in hMLH1 at 3p21.3; hMSH2 at 2p22-p21; hPMS1 at 2q31-q33; hPMS2 at 7p22; and hMSH6 (GTBP) at 2p16; and maybe other genes. The products of the first five genes participate in a multimeric DNA mismatch repair complex. Other genes also participate in this complex, but germ-line mutations in those genes have not yet been confirmed. Note that not all families diagnosed with “HNPCC” by pedigree criteria have heritable defects involving this pathway, thus it is

568

PART III: THE CAUSES OF CANCER

important to define how the term “HNPCC” is being used (pedigree versus molecular definition) in discussions of this disorder. Incidence: HNPCC due to defective DNA mismatch repair genes may account for 2–3% of all colorectal cancers (Wijnen et al., 1998). Diagnosis: Diagnosis has traditionally relied on pedigree assessment. The Amsterdam I criteria (ACI) (Vasen et al., 1991) were developed in 1991 to assist in defining HNPCC for research purposes. The criteria are acknowledged to be overly restrictive for most clinical purposes, as up to 20% of true HNPCC families (determined by germ-line mutation identification) will not meet these criteria. Conversely, up to 50% of families that do fulfill the ACI criteria do not have evidence of a DNA mismatch repair defect. The Bethesda Guidelines (Rodriguez-Bigas et al., 1997; Umar et al., 2004) were developed to cast a broader net for the identification of families in which testing for microsatellite instability (MSI) should be considered. MSI is the tumor phenotype that indicates defective DNA mismatch repair. Note: A microsatellite unstable tumor phenotype is not diagnostic of HNPCC because sporadic age-related promoter methylation will also cause an MSI-H phenotype. Associated malignant neoplasms: Colorectal cancer, most often in the right colon, with average age at diagnosis of 45 years, is the salient feature of HNPCC. The lifetime risk of colorectal cancer is approximately 80%. Endometrial adenocarcinoma is the second most frequent syndromic neoplasm, with an average age at diagnosis of 45; the lifetime risk of endometrial adenocarcinoma is 30–60% (Lynch et al., 1996). Aarnio et al. (1999) reported cumulative risks for colon, endometrial, gastric and ovarian cancers as 85%, 60%, 13%, and 12%, respectively. In their population, 47 out of 50 families included had mutations in MLH1 gene. Site-specific cancer risks likely will vary between different studies, depending on how HNPCC is defined and which of the MMR genes predominate in that cohort. Several studies have now noted that the risk of endometrial cancer may be higher than the risk of colon cancer in gene mutation carrying women. MSH6 appears to have later onset of cancers and lower penetrance for the HNPCC-related tumors, except for endometrial cancer, which may be the predominant feature in MSH6 families. Sebaceous carcinomas are also found. Benign or malignant sebaceous skin tumors in combination with internal cancer have been called Muir-Torre syndrome; linkage and mutational analysis of both hMSH2 and hMLH1 have demonstrated that Muir-Torre syndrome is a form of HNPCC. There is probable increased risk of basal cell cancers and squamous cell cancers of the skin. Glioblastoma multiforme is associated with HNPCC. Brain tumor in combination with colorectal tumors is also called Turcot syndrome. Patients with FAP have an increased risk of medulloblastoma, and this, too, has been called Turcot syndrome. Rare homozygotes for germ-line DNA MMR gene mutations are reported. They have café-au-lait macules like those seen in neurofibromatosis, early onset hematologic or brain malignancies, in addition to very early onset HNPCC-spectrum tumors. Associated benign neoplasms: Colonic adenomas, keratoacanthomas, sebaceous adenomas, and epitheliomas.

Cowden Syndrome (Gingival Multiple Hamartoma Syndrome) Genetics: Autosomal dominant disorder due to mutation in the PTEN gene at 10q23. Mutations in the BMPR1A gene on 10q may cause a subset of Cowden syndrome (CS). Incidence: Estimated at approximately 1 in 200,000 to 1 in 250,000 in the Dutch population (Nelen et al., 1999). Diagnosis: The International Cowden Syndrome Consortium Operational Criteria have been published (Eng, 2000). Associated benign neoplasms: Verrucous skin lesions of the face and limbs and cobblestone-like hyperkeratotic papules of the gingiva and buccal mucosal, facial trichilemmomas, oral mucosal fibromas, and hyperkeratotic lesions of the hands and feet. Sixty percent of affected individuals had hamartomatous polyps of the stomach,

small bowel, and colon. Lipomas, giant fibroadenomas of the breast, cerebellar gangliocytomatosis, and hemangiomas are also common. Schrager et al. (1998) examined breast tissue from 19 women with CS. The women showed a spectrum of benign histopathological findings including ductal hyperplasia, intraductal papillomatosis, adenosis, lobular atrophy, fibroadenomas, and fibrocystic change. Seventeen (89%) showed features suggestive of a breast hamartoma. Associated malignant neoplasms: Schrager et al. (1998) identified an increased risk of cancer of the following types/sites: follicular cell carcinoma of the thyroid, carcinomas of the breast, and cancers of the urogenital and digestive tract. The highest risk is of breast cancer in females, approximately 70%, with ductal carcinoma as the most frequent breast malignancy observed (86% in one series). Endometrial cancer was recently added to the list of syndromedefining malignancies (Eng et al., 2000). Other cancers (melanoma, Merkel cell cancer, lung, and retinal glioma) have been reported in the context of CS, but it is not clear if the overall risk of these is different from that in the general population.

Fanconi Anemia (FA) (Includes Pancytopenia, Fanconi type) Genetics: Genetically heterogeneous autosomal recessive disorder due to mutation in FANCA at 16q24.3; FANCB, unmapped; FANCC at 9q22.3; FANCD1at 13q12.3 (BRCA2); FANCD2 at 3p26-p22; FANCE at 6p22-p21; FANCF at 11p15; FANCG at 9p13; FANCI, unmapped; FANCJ, unmapped; and FANCL at 15q15-q21.1. Incidence: Heterozygote frequency estimated at 1 in 300–600 in the general population, and 1 in 100 among Ashkenazim, in whom a FANCC founder mutation occurs. FA is associated with approximately 20% of all cases of childhood aplastic anemia. Diagnosis: Multiple clinical findings have been described in Fanconi anemia (Scriver et al., l995). Induction of chromosome breakage with the mitomycin-C chromosome stress test or diepoxybutane in vitro reliably identifies homozygotes (Auerbach, 1993; Kuffel et al., 1997). These tests show 10- to 100-fold excesses of chromatid breaks, gaps, characteristic formation of radial chromosomes, endoreduplications, and other types of nonhomologous recombination. Standard karyotyping does not demonstrate these features, and heterozygotes cannot be detected with these studies. Associated malignant neoplasms: Acute nonlymphatic leukemia, hepatocellular carcinoma (often after anabolic steroid therapy for aplastic anemia), and squamous cell carcinomas. Leukemia is the terminal event in 5–10% of affected individuals. FA also predisposes to solid tumors, especially squamous cell carcinomas, treatment of which is complicated by hypersensitivity to radiation therapy (reviewed by Oksuzoglu and Yalcin, 2002). Quantitative assessments of the solid tumor risks demonstrate dramatic excesses of cancers involving the head/neck, esophagus, and female genital tract (Kutler et al., 2003). Howlett et al. (2002) showed bi-allelic inactivation of BRCA2 in FA; BRCA2 is now thought to be the FANCD1 gene, and it has been suggested that this gene includes a predisposition to both breast cancer and brain tumors in FA families (Offit et al., 2003). Increased cancer risks have been suggested in FA heterozygotes, an observation which has not been confirmed. Associated benign neoplasms: Hepatic adenomas, likely related to treatment with anabolic steroids for the anemia.

Gastric Cancer, Hereditary Diffuse (HDGC) Genetics: Autosomal dominant disorder due to mutations in the CDH1 gene, encoding E-cadherin on chromosome 16q22 (Guilford et al., 1998). Incidence: Unknown; rare. Diagnosis: The first workshop of the International Gastric Cancer Linkage Consortium has proposed a working definition for HDGC (Caldas et al., l999). Associated malignant neoplasms: Pharoah et al. (2001) analyzed 11 HDGC families and estimated cumulative risk of diffuse gastric cancer by age 80 years was 67% for men (95% CI, 39–99%) and 83% for

Hereditary Neoplastic Syndromes women (95% CI, 58–99%). The risk of breast cancer (often lobular) in women was 39 (95% CI, 12–84). In the Maori gastric cancer families, the majority of cases occurred under the age of 40 (youngest was age 14). For other reported families, the age at diagnosis for gastric cancer in gene mutation carriers clustered around age 40. Associated benign neoplasms: None known. In all reported cases in which gene carriers have undergone prophylactic gastrectomy, multifocal islands of intramucosal signet ring gastric adenocarcinoma were identified, but a precancerous lesion was not evident (Chun et al., 2001; Huntsman et al., 2001).

Li-Fraumeni Syndrome (LFS) Including Li-Fraumenilike Syndrome (LFL) Genetics: Autosomal dominant disorder for which about 70% of LiFraumeni families have identifiable mutations in the p53 gene (also known as TP53), which is located on chromosome 17p13.1 (Varley et al., 1999). In LFS families negative for p53 mutations, germ-line CHEK2 mutations have been reported (Bell et al., l999; Vahteristo et al., 2001). Incidence: Unknown. Unselected children with adrenocortical carcinoma have a very high frequency of germ-line p53 mutations (50–100%) (Varley et al., 1999; Chompret et al., 2001). Mutations have been detected in approximately one-third of children with bone or soft-tissue sarcoma and in 9% of patients with rhabdomyosarcoma. Patients with multiple primary syndrome-related tumors had an estimated mutation frequency of 7–20% (Chompret et al., 2001). Diagnosis: The classical definition of LFS requires (1) one patient with sarcoma under the age of 45, (2) a first-degree relative under the age of 45 with cancer (type not specified), and (3) a third affected family member (first- or second-degree relative) with either sarcoma at any age or cancer (type not specified) under the age of 45 years (Li et al., 1988; Malkin et al., 1990). Families which resemble LFS, but which do not meet the above definition have been designated “Li-Fraumeni-like” (LFL). Various criteria for LFL have been proposed, with 8% to 22% of such families found to carry a p53 mutation (Birch et al., l994; Eeles et al., l995). Associated malignant neoplasms: Risk of developing any invasive cancer (excluding skin cancer) was almost 50% by age 30 (compared with 1% in the general population), and almost 90% by age 70. The tumor spectrum includes rhabdomyosarcoma and other soft tissue sarcomas, osteogenic and chondrosarcomas, breast cancer, brain cancer (especially glioblastomas), leukemia, and adrenocortical carcinoma. In some families, other tumors appear to be excessive, including lymphoma, melanoma, Wilms tumor, and laryngeal, lung, gonadal germ cell, pancreatic, gastric, and prostate cancers (Frebourg et al., 1995; Kleihues et al., 1997; Hisada et al., 1998). Clinical observations suggest an increased susceptibility to radiation-induced cancers (mainly sarcomas) within the treatment field of a prior cancer. Hisada et al. (1998) estimated that the probability of developing a second primary cancer in patients with LFS reached 57% by 30 years follow-up, documenting that this syndrome is associated with an exceptionally high risk of multiple primary malignancies. Risk decreased with increasing age at diagnosis of the first cancer. This risk was higher among patients who initially had soft tissue sarcoma (100% by 30 years follow-up). Based on small numbers, the estimated probability of developing a third cancer was 38% at 10 years. Associated benign neoplasms: None known.

Melanoma, Familial with or without Dysplastic Nevi (Includes Dysplastic Nevus Syndrome, Familial Atypical Mole-Malignant Melanoma [FAMMM] Syndrome, Familial Atypical Multiple Mole Melanoma-Pancreatic Carcinoma Syndrome [FAMMMPC], Melanoma-Astrocytoma Syndrome) Genetics: Autosomal dominant disorder due to genes in at least three susceptibility loci. CMM1 maps to chromosome 1p36 but no can-

569

didate gene has yet been identified. CMM2 maps to 9p21, and the gene is CDKN2A (which produces the p16ink4a protein). Approximately 20–25% of melanoma-prone kindreds are linked to CDKN2A. CDKN2A also encodes for a second protein, p14ARF, through an alternative splicing mechanism. The role of this protein in familial melanoma is uncertain. Rare germ-line mutations in the CDK4 gene (chromosome 12q14) have been reported. Incidence: It is estimated that 5–7% of melanoma patients are from genetically high-risk families, although higher and lower frequencies have been reported. Bishop et al. (2002) estimated age-specific penetrance by geographic region with an estimated overall penetrance across different ethnicities of 30% by age 50 years, 67% by age 80 years. Diagnosis: Family history of invasive melanoma in at least two firstdegree relatives. At least three affected relatives are required in geographical regions with high solar exposure, such as Australia. Early age at melanoma diagnosis and a tendency to develop multiple primary melanomas characterize these families. The diagnosis of dysplastic nevus syndrome requires the presence of 10 to 100 irregularly shaped, variably pigmented nevi (Tucker et al., 2002). There are clearly familial melanoma kindreds that do not manifest the dysplastic nevus syndrome. Associated malignant neoplasms: Members of kindreds with deleterious CDKN2A mutations have a melanoma risk that is increased by a factor of 75 and a risk of pancreatic cancer that is increased by a factor of 13 (Goldstein et al., 1995; Whelan, l995). The average age at diagnosis of melanoma in hereditary melanoma/dysplastic nevus families is 34 years, compared with 54 years for melanoma diagnosed in the general population. Families with pancreatic cancer include both early-onset and late-onset pancreatic carcinoma, as well as individuals with double primary melanoma and pancreatic cancer (Lynch et al., 2002). Astrocytomas occur with melanomas in rare families (Kaufman, l993; Bauhau et al., 1998). Bataille (2000) commented that clustering of other cancers such as breast in addition to pancreas and skin cancers are not uncommon in melanoma families. Borg et al. (2000) reported a 3.8-fold excess (95% CI, 1.6–7.5) of breast cancer in a series of melanoma families with CDKN2A mutations. Associated benign neoplasms: Dysplastic nevi.

Multiple Endocrine Neoplasia Type 1 (MEN1) (Wermer Syndrome; Includes Zollinger-Ellison Syndrome) Genetics: Autosomal dominant disorder due to mutation in the MEN1gene on chromosome 11q13. Incidence: Unknown. Diagnosis: This disorder is characterized by a high frequency of peptic ulcer disease and endocrine abnormalities, especially of the pituitary (44%), parathyroid (95%), pancreas (73%), and adrenal (16%). Hyperparathyroidism is the presenting symptom or is diagnosed simultaneously with the presenting symptom in 94% of cases (Bassett et al., 1998). Associated malignant neoplasms: Duodenal, thymic, bronchial, and gastric carcinoids; bronchial carcinoma; malignant schwannoma; ovarian tumors; pancreatic islet cell carcinomas and other malignant pancreatic endocrine tumors; and adrenocortical carcinomas (Calender, l995; Skogseid, l995). Associated benign neoplasms: Pancreatic islet-cell adenomas; parathyroid hyperplasias or single or multiple adenomas; pituitary adenomas that are either nonsecreting or that secrete prolactin (most frequent), growth hormone, ACTH, and luteinizing hormones (LH); multiple adrenocortical adenomas; gastrinomas (usually of the duodenum, but also seen in the pancreas); Cushing syndrome; pheochromocytomas; prolactinomas; glucagonomas; insulinomas; vasointestinal peptide tumors; angiofibromas in a high proportion of affected individuals; leiomyomas; collagenomas; thymomas; and lipomas.

570

PART III: THE CAUSES OF CANCER

Multiple Endocrine Neoplasia Type 2A, 2B (MEN2A and MEN2B), and Familial Medullary Thyroid Cancer (FMTC) Genetics: Autosomal dominant disorder due to mutations in the RET proto-oncogene on chromosome 10q11.2. Incidence: Estimated at 1 in 30,000. The germ-line mutation carrier frequency is unknown. Five to 10% of all thyroid cancers in clinical practice are of the medullary type. Among these, 20% are due to germ-line RET mutations. Diagnosis: MEN2A is diagnosed clinically when two or more of the following are present: medullary thyroid carcinoma, pheochromocytomas, or parathyroid adenoma/hyperplasia. MEN2B is characterized by MTC and presence of mucosal neuromas of lips, tongue, eyelids, enlarged corneal nerve fibers, distinctive facies, marfanoid body habitus. FMTC is diagnosed in families with multiple cases of MTC without pheochromoocytoma or parathyroid hyperplasia (Eng, l996). Associated malignant neoplasms: Medullary thyroid cancer (MTC), with the earliest age at diagnosis reported for a 2.7-year-old child with MEN2A. MEN2B displays an even younger age at onset and a more aggressive clinical course, with invasive MTC occurring even in infancy. MTC is almost always diagnosed before the age of 40 in MEN2A and 2B. Pheochromocytomas are present in up to 50% of individuals with MEN2A and 2B; about 4% undergo malignant transformation (Modilgliani, l995). Associated benign neoplasms: C-cell hyperplasia is nearly universal. Hyperparathyroidism is found in 20–30% of individuals with MEN2A; this condition is rare in MEN2B and is never found in FMTC (by definition). Ganglioneuromas of the GI tract, mucosal neuromas and thickened corneal nerves are present in nearly all patients with MEN2B (Eng, l996). Some families with MEN2A develop cutaneous lichen amyloidosis.

Neurofibromatosis Type 1 (NF1) Genetics: Due to mutations in the NF1 gene at 17q11.2. Incidence: One in 2500–5000; one-third to one-half of cases represent new mutation dominant disease. Diagnosis: Clinical features and genetics reviewed by Gutmann and Collins (2001). The National Institutes of Health (NIH) consensus criteria require two or more of the following for diagnosis: 1. Cafe-au-lait macules (CALs): in children, five or more that are ≥0.5 cm in diameter and, in adults, six or more that are ≥1.5 cm in diameter (note that 25% of the general population have 1–3 CALs); 2. Two or more neurofibromas of any type or one plexiform neurofibroma; 3. Multiple axillary or inguinal freckles; 4. Sphenoid wing dysplasia or congenital bowing or thinning of the long bone cortex (+/- pseudoarthrosis); 5. Bilateral optic nerve gliomas; 6. Two or more iris Lisch nodules (iris hamartomas); 7. A first-degree relative with NF type 1 by these criteria. Associated malignant neoplasms: Neurofibrosarcomas or malignant peripheral nerve sheath tumors develop in 3–5% of affected individuals, most often associated with plexiform neurofibromas. Conversely, it has been estimated that 50% of individuals with neurofibrosarcomas have NF. Also, there is increased risk of astrocytomas, carcinoid tumors (usually duodenal), pheochromocytomas, neuroblastomas, ependymomas, primitive neuroectodermal tumors, rhabdomyosarcomas (especially of the genitourinary tract), undifferentiated sarcomas, Wilms tumors, and juvenile chronic myelogenous leukemia (Korf, 2000). Optic gliomas affect 15–20% of patients (Listernick, l999). Associated benign neoplasms: Neurofibromas; pheochromocytomas (risk: 0.1–1.0%); meningiomas; optic nerve schwannomas.

Neurofibromatosis Type 2 (NF2) Genetics: Autosomal dominant disorder due to mutation in the NF2 gene at 22q12.2. Incidence: One in 37,000, of which 50% are caused by new germline mutations. NF2 is 10 times less common than NF1. Diagnosis: The 1987 NIH consensus conference criteria for NF2 have been modified (Mulvihill, 1990), and they now require one of the following two major criteria to be met: 1. bilateral 8th nerve masses seen by appropriate MR imaging techniques or 2. a first-degree relative with NF2 plus one of the following: a. unilateral eighth nerve mass or b. two of the following: meningioma, glioma, schwannoma, or juvenile posterior subcapsular lenticular opacity. Associated malignant neoplasms: Gliomas and ependymomas (often intramedullary). Associated benign neoplasms: Vestibular schwannomas (formerly designated, inaccurately, as “acoustic neuromas”), meningiomas (in up to half of NF2 patients), and both peripheral and spinal cord schwannomas.

Paraganglioma, Familial Genetics: Hereditary paraganglioma type 1 (PGL1) is caused by mutations in the SDHD gene, located on chromosome 11q23. PGL2 maps to chromosome 11q13.1; the causative gene is not yet identified. PGL3 is caused by mutations in the SDHC gene, located on chromosome 1q21. PGL4 is caused by mutations in the SDHB gene, which is located on chromosome 1p36. The three genes identified to date each encode a mitochondrial respiratory chain protein. Hereditary paraganglioma is an autosomal dominant disorder; however, a unique parent-of-origin effect is consistently reported for the SDHD gene: offspring of female gene carriers do not develop disease, whereas 50% of offspring of male gene carriers do. This pattern suggests genetic imprinting (Van der Mey et al., l989; Baysal, 2001). Examination of phenotypic expression of both germline SDHB and SDHC mutations reveals no parent-of-origin effects (Baysal et al., 2002). Incidence: Very rare. Estimated at 1 case per million in the Netherlands (Oosterwijk et al., 1996). An estimated 10% of unselected paraganglioma cases in the United States were familial versus 50% in the Netherlands (Van der Mey et al., 1989). Diagnosis: No validated criteria. Should be considered in any individual or kindred with multiple paragangliomas or glomus tumors. Multifocality appears to be a feature of hereditary PGL (Baysal et al., 2002). Associated malignant neoplasms: Paragangliomas can be clinically malignant despite being histologically benign. Four to 16% are reported to undergo malignant degeneration with metastases in 5%. Baysal (2001) noted that familial paraganglioma tumors may be associated with other tumor types, including astrocytomas, thyroid carcinomas, and parathyroid adenomas (risks not defined). Associated benign neoplasms: Paragangliomas; pheochromocytomas.

Polyposis, Familial Adenomatous (FAP/Gardner Syndrome, Includes Familial Multicentric Fibromatosis/Hereditary Desmoid Disease) Genetics: Autosomal dominant disorder due to mutations in the APC gene on 5q21-q22; Also, can be autosomal recessive disorder due to mutation in the MYH gene, a base excision repair gene on 1p34.3-p32.1 (Al-Tassan et al., 2002). Most mutations of APC are truncating. One specific missense mutation (I1307K) has been reported in 6% of Ashkenazi Jews and about 28% of Ashkenazim with a family history of colorectal cancer (Laken et al., 1997). This is associated with only a twofold increased risk of colorectal cancer.

Hereditary Neoplastic Syndromes Incidence: One in 6000 to 1 in 13,000. The frequency of gene mutations in the general population is unknown. Diagnosis: Based on characteristic polyposis (usually the presence of greater than 100 colorectal polyps). Associated malignant neoplasms: In untreated individuals with germ-line APC mutations, colon adenocarcinoma occurs at a mean of 39 years (7% by age 21; 87% by age 45) (Bussey, 1975). Duodenal carcinomas, especially around the ampulla of Vater, occur on the average 20 years later than the colon cancers, with a lifetime risk of 4–12% (Kadmon et al., 2001). Follicular or papillary thyroid cancer occurs in about 2% of affected individuals at a mean age of 28 years (Cetta et al., 2000). The risk of childhood hepatoblastomas is estimated as 1 in 150 and rarely occurs after age 6 years (Hughes and Michels, 1992). Gastric carcinomas arise in only 0.5% of individuals with FAP in Western cultures but are more common in FAP kindreds in Japan and Korea (Offerhous et al., 1999; Park et al., 1992). Lifetime risk of pancreatic cancer is 2%, and this may include islet cell tumors. Central nervous system tumors are reported in some families, 79% of which are medulloblastomas (Hamilton et al., 1995). The combination of multiple adenomatous colon polyps with a brain tumor has been called “Turcot syndrome.” Turcot syndrome can be seen in both FAP and HNPCC. Homozygous mutations in MYH gene can cause a phenotype indistinguishable clinically from classical FAP or can result in a more attenuated phenotype. Specific cancer risks have not yet been determined. Associated benign neoplasms: In APC-associated FAP, adenomatous polyps of the colon appear at a mean age of 16 years, but may occur before age of 10 years in gene carriers (<10%); they are present in more than 90% of gene carriers by age 20 years and in more than 95% by age 35 years. Duodenal polyps (especially periampullary) occur in 50–90% of patients (Kadmon et al., 2001). Hamartomatous gastric polyps occur in up to half of FAP patients. Ten percent of individuals with FAP may have adenomatous gastric polyps. The risk of malignancy in these polyps is low, except in the Japanese and Korean FAP populations (Bulow et al., l995). Dental abnormalities occur in 17% of individuals with FAP and include supernumerary or congenitally absent teeth and dentigerous cysts and osteomas of the jaw (Brett, l994). Other benign findings include sebaceous or epidermoid cysts, lipomas, congenital hypertrophy of the retinal pigment epithelium (CHRPE) (Burn et al., 1991). Osteomas can arise in any bone. Desmoid tumors are histologically benign clonal neoplasms comprised of fibrous tissue that cause significant morbidity or mortality in about 5% of individuals with FAP. Desmoid tumors are reported in 30% of FAP kindreds, with an overall lifetime risk of 8% for males with FAP and 15% for females; the risk of desmoid tumors is 25% if a first-degree relative with FAP has a desmoid versus 8% if third-degree relative is affected (Gurbuz et al., l994). Smith et al. (2000) found 13% of patients with FAP have adrenal masses >1 cm.

Renal Cell Carcinoma, Hereditary Papillary (RCCP or HPRC) Genetics: Autosomal dominant disorder due to mutations in the MET gene at 7q31.1-34. Mutations in MET cause a gain of function, indicating that it functions as an oncogene. Incidence: Papillary renal cell cancer accounts for 15–20% of all RCCs. Of all RCCs, approximately 2% represent familial cases. Schmidt et al. (1998) estimated that the age-dependent penetrance of the H1112R MET mutation reaches 100% by age 80. Diagnosis: Patients with either bilateral and multifocal papillary renal cell tumors without a family history, or patients with a single tumor or multifocal tumors and a first- or second-degree relative with papillary renal cell cancer should be suspected for HPRC. Associated malignant neoplasms: Papillary RCC, type 1. Case reports indicate that other cancers do occur in these families, but there are no quantitative data to indicate whether these occurrences are anything other than coincidental. Associated benign neoplasms: None known.

571

Retinoblastoma, Hereditary Genetics: Autosomal dominant due to mutations in RB1 at 13q14. Incidence: One in 15,000 to 20,000. The frequency of gene mutation carriers in the general population is unknown. Approximately 60% are unilateral and nonhereditary, 15% are unilateral and hereditary, and 25% are bilateral and hereditary. Males and females are equally affected. Approximately 20–30% of individuals with germline mutations have new germinal mutations. Diagnosis: Based on strabismus and/or leukocoria in infants or young children. Approximately 90% of all retinoblastomas are diagnosed before the age of 3 years, with the average age at diagnosis being 12 months in the case of bilateral disease versus 18 months in unilateral disease. Associated malignant neoplasms: Approximately 49% of the offspring of bilaterally affected parents develop retinoblastoma, whereas 42% of the offspring of unilaterally affect parents with hereditary disease are affected, suggesting some variability in gene penetrance. Age at onset has some predictive value with respect to the development of bilateral disease, with bilateral disease developing in 85%, 82%, 44%, and 6% of cases that present under the age of 6 months, ages 6–11 months, 12–23 months, and ≥24 months, respectively. Second malignant tumors in RB patients occur both within and outside of radiation treatment fields. Approximately 1% of patients with hereditary retinoblastoma develop a second nonocular tumor each year and, of these, at least 50% die of these second malignant neoplasms. For those who survive their second cancer, there is a 2% per year risk of developing a third tumor (Abramson and Frank, 1998). The cumulative incidence of second malignancies in hereditary bilateral retinoblastoma has been reported as 4.4% at 10 years, 18.3% at 20 years, and 26.1% at 30 years after diagnosis (Roarty et al., 1988). Lower cumulative risks have also been described (Moll et al., 2001). This study confirmed a greater risk in children diagnosed and irradiated before the age of 1 year but suggested that radiation did not increase the risk of second cancers beyond that accountable for by RB itself. However, others have reported that, while genetic predisposition has a substantial impact on risk of subsequent cancers in retinoblastoma patients, this risk is further increased by radiation treatment, as evidenced by a strong radiation dose-response relationship for all sarcomas (Wong et al., 1997). The second nonocular tumor observed most often was osteosarcoma (risk increased by 500-fold). Fibrosarcomas, chondrosarcomas, epithelial malignant tumors, Ewing sarcomas, leukemias, lymphomas, melanomas, brain tumors, and pinealoblastomas were also reported. RB patients are also at increased risk of breast tumors and Hodgkin disease (Wong et al., 1997) and of early-onset lung cancer (Kleinerman et al., 2000). Associated benign neoplasms: Retinomas (benign retinal tumors), and lipomas.

Rothmund-Thomson Syndrome (RTS) Genetics: Autosomal recessive disorder caused by mutations in the gene RECQL4 on chromosome 8q24.3. This gene is a RECQ helicase gene like those that cause Bloom and Werner syndromes. Incidence: Very rare. Through 1990, only 200 cases had been reported worldwide (Vennos et al., l992). Diagnosis: Based on clinical assessment with confirmation by mutation testing and exclusion of similar phenotypes, such as Bloom syndrome. The characteristic feature, found in nearly all patients diagnosed with RTS, is a sun-sensitive rash, which usually presents between 3 and 6 months (range, birth to 2 years), and which becomes generalized, resulting in poikiloderma. Associated malignant neoplasms: Wang et al. (2001) reported that 13 of 41 (32%) of RTS patients developed osteosarcoma (OS), with a median age at diagnosis of 12 years (range, 4–41 years). In all but one patient, the RTS diagnosis preceded diagnosis of OS. This prevalence is clearly excessive. One additional patient from this cohort developed Hodgkin lymphoma, while another has

572

PART III: THE CAUSES OF CANCER

recurrent squamous cell carcinoma (SCC) of the lip. One case each of parathyroid adenoma, “Hodgkin’s sarcoma,” fibrosarcoma, and gastric carcinoma have been reported (Vennos et al., 1992). It is uncertain if these associations are meaningful. Associated benign neoplasms: Warty dyskeratosis.

Tuberous Sclerosis Complex (TSC) Genetics: Autosomal dominant disorder due to mutations in either the TSC1 gene on chromosome 9q34 or the TSC2 gene on chromosome 16p13.3. Incidence: Estimated at approximately 1 in 30,000 individuals under the age of 65 and 1 in 15,000 under the age of 5. Up to 60% of cases represent new germline mutations. Diagnosis: The diagnostic criteria currently used for the tuberous sclerosis complex have been published (Roach et al., l998). Associated malignant neoplasms: There is a 6–14% incidence of childhood brain tumors in patients with tuberous sclerosis, of which more than 90% are subependymal giant-cell astrocytomas (Gomez, l988). Tuberous sclerosis is associated with renal cell cancer, which is reported as the second leading cause of death (27.5%) (Shepherd et al., l991). The age at diagnosis of renal cell cancer has been reported to be younger than that of sporadic renal cell cancer. Malignant angiomyolipomas have been reported, as have Wilms tumors. Associated benign neoplasms: Cortical or subcortical tubers (70%), subependymal glial nodules (90%), retinal hamartomas, and facial angiofibromas (80% of postpubertal patients), also called “adenoma sebaceum.” Cook et al. (1996) reported renal lesions in 61% (85 of 139 patients), including angiomyolipomas, renal epithelial cysts, or both. The angiomyolipomas were multiple in 91% and bilateral in 84%. Angiomyolipomas may rupture, causing life-threatening hemorrhage. Cardiac rhabdomyomas may present prenatally or perinatally. It is estimated that 51–86% of cardiac rhabdomyomas may be associated with tuberous sclerosis. Shagreen patches (connective tissue hamartomas) are present in 54–55% of affected individuals. Ungual fibromas are usually not present in children under the age of 5; they are found in 23% of affected children between the ages of 5 and 14 years and 88% of patients older than 30 years of age. Multiple dental enamel pits in secondary dentition are seen in 71% of patients with tuberous sclerosis compared with 0.88% in control subjects. Adrenal angiomyolipomas, adrenal adenomas, paragangliomas, pancreatic adenomas, and parathyroid adenomas are also reported (Gomez, l988).

von Hippel-Lindau (VHL) Disease Genetics: Autosomal dominant disorder due to mutation in the VHL gene at 3p25-26. This protein product plays a role in the transduction of growth signals generated by changes in ambient oxygen tension. Incidence: Approximately 1 in 30,000–40,000, with nearly complete penetrance by the age of 65, and an average life expectancy of 49 years. Diagnosis: Based on clinical criteria as follows: Central nervous system (CNS) and retinal hemangioblastoma; or, CNS or retinal hemangioblastoma, plus one of the following: multiple renal, pancreatic, or hepatic cysts; pheochromocytoma; or renal cancer; or Definite family history plus one of the following: CNS or retinal hemangioblastoma; multiple renal, pancreatic, or hepatic cysts; pheochromocytoma; or renal cancer. Associated malignant neoplasms: As reviewed by Michels (1987), malignant RCC occurs in 35–75% of affected individuals in autopsy series and in 25–38% in clinical series; the renal cell cancers are often multiple and bilateral. The average age at diagnosis is 41–45 years (range, 20–69). The mean age at renal cell cancer death is 44.5 years. Pancreatic cystadenocarcinoma or islet cell tumors tend to

cluster in certain families where the incidence ranges from 7.5% to 25%. APUDomas that produce vasoactive intestinal peptide and cause hypercalcemia have been reported. Pancreatic cancer does occur but is not common in VHL. Carcinoid tumors have been reported. Retinal and cerebellar hemangioblastomas are common VHL tumors. Associated benign neoplasms: Retinal angiomas may result in visual loss. These angiomas have been detected in children as young as the age of 4, but typically they become evident between the ages of 21 and 28 years. Pancreatic cysts, which can be multiple and occasionally large, are detected in 9–29% of patients; pancreatic angiomas or cystadenomas are found in 7% of patients and hemangioblastomas are rare. CNS hemangioblastomas, which are histologically benign tumors, occur in 50–79% of autopsy-confirmed cases and in 18–44% of patients in clinical series. Varying proportions of patients have renal lesions, including simple cysts, typically multiple and bilateral (76%), hemangiomas (7%), and benign adenomas (14%). Pheochromocytomas occur in a minority (4–17%) of VHL patients, but they tend to cluster in selected kindreds. Benign adenomas and paragangliomas of the sympathetic chain are also found infrequently in VHL. Epididymal cysts occur in 7–27% of patients (size: 0.5–2.0 cm). Benign epididymal papillary cystadenomas are found in 3–26% of autopsied males, and women with VHL have been reported to have broad ligament papillary cystadenomas. Hepatic cysts (in 17% of patients) have been reported in autopsy series, as have angiomas and cysts of the spleen (3–7%). Endolymphatic sac tumors (ELSTs) cause hearing loss in VHL patients (Manski et al., 1997).

Werner Syndrome (Adult Progeria) Genetics: Autosomal recessive disorder due to mutation in the WRN gene on chromosome 8p12-11.2, which encodes a DNA helicase. Incidence: Estimated at 1 in 50,000 to 1 in 1,000,000. Higher reported incidence rates from Japanese populations are likely due to high rates of consanguinity (Goto et al., 1981). Diagnosis: No consensus criteria have been published. Growth arrest at puberty; cataracts occur in the second and third decades; premature graying and balding; scleroderma-like changes of the limbs; diminution of muscle mass and of subcutaneous tissue; chronic pressure ulcers on the feet and ankles; premature arteriosclerosis; adult onset diabetes; endocrine failure; and localized soft tissue calcification. Intellect is normal. Old age appearance is evident by age 30–40. Associated malignant neoplasms: Goto et al. (1996) reviewed 34 non-Japanese cases of WS with 30 cancers and 124 Japanese cases with 127 cancers. The cancers were diagnosed between the ages of 25 and 64, except for one 20-year-old with osteosarcoma and a 24year-old with acute myelogenous leukemia. Among Japanese patients, there were 23 soft-tissue sarcomas, 21 melanomas (which occurred in unusual locations, especially on the nasal mucosa and on the feet), 9 osteosarcomas, and 14 hematologic disorders. In addition, there were 63 epithelial malignancies, most notably cancers of the thyroid (21), hepatobiliary system (7), stomach (6), and breast (6). Among the 30 non-Japanese cases, there were 7 soft tissue sarcomas, 4 osteosarcomas, 3 melanomas, and 4 hematologic disorders. Eleven non-epithelial tumors were reported, most notably two thyroid cancers. Associated benign neoplasms: Sixteen of 124 Japanese cases and 7 of the 30 non-Japanese cases had benign meningiomas.

Xeroderma Pigmentosum (XP) Genetics: Autosomal recessive disorder with eight identified genes: XP-A (9q34.1); XP-B (2q21); XP-C (3p25.1); XP-D (19q13.2); XP-

Hereditary Neoplastic Syndromes E (11p12-p11); XP-F (16p13.2-p13.1); XP-G (13q32-33); and XPVariant (6p21.1-p12). With one exception, these genes are involved in nucleotide excision repair (NER) of ultraviolet radiation-induced DNA pyrimidine dimers. Incidence: One in 1,000,000 in the United States; 1 in 40,000 in Japan. Diagnosis: Based on childhood onset of photosensitivity (blistering in 50%) and/or freckling, with progressive cutaneous changes leading to poikiloderma, irregular pigmentation, telangiectasia, photophobia, and very early development of skin and eye cancers. Skin changes in sun-exposed areas are evident in 50% of affected individuals by age 18 months, in 75% by age 4 years, and in 95% by age 15 years (Kraemer et al., l987; Scriver et al., l995). Approximately 20% of reported patients with XP have associated neurologic abnormalities. Complementation studies assigned Cockayne syndrome (CS) patients to the rare XP groups B, D, and G. Individuals with CS do manifest photosensitivity but do not develop skin cancer. CS is characterized by cachectic dwarfism with microcephaly, premature aged appearance as well as progressive mental, neurologic, and retinal degeneration. Trichothiodystrophy (TTD) also is allelic with XP complementation groups B and D, and it also lacks a predisposition to skin cancers. TTD patients have brittle hair and nails, ichthyotic skin, and physical and mental retardation. Approximately half of the patients display photosensitivity. The site of the mutation in the NER gene determines the clinical susceptibility to malignancies. Associated malignant neoplasms: There is a 2000-fold increased frequency of basal cell and squamous cell carcinomas of the skin, often with multiple tumors, by age 20. The relative risk of malignant melanoma is increased 700-fold. The incidence of squamous cell carcinoma of the sun-exposed tip of the tongue is increased 10,000fold. Increases in the risk of internal neoplasms, including brain tumors, lung cancers, gastric cancers, sarcomas, Wilms tumors in adults, and leukemias have been reported, but the reliability of these estimates is limited by very small numbers (Kraemer et al., 1987, 1994). The possibility that XP heterozygotes may be at increased risk of malignancy has been proposed, but this question awaits a definitive answer (Heim et al., 1992). Associated benign neoplasms: Conjunctival papillomas, actinic keratoses, lid epitheliomas, keratoacanthomas, angiomas, and fibromas. References Aarnio M, Sankila R, Pukkala E, et al. 1999. Cancer risk in mutation carriers of DNA-mismatch-repair genes. Int J Cancer 81(2):214–218. Abramson DH, Frank CM. 1998. Second non-ocular tumors in survivors of bilateral retinoblastoma: A possible age effect on radiation-related risk. Ophthalmology 105(4):573–580. Airwele G, Adatto P, Cunningham J, et al. 1998. Family history of cancer inpatients with glioma: A validation study of accuracy. J Natl Cancer Inst 90:543–544. Aitken J, Bain C, Ward M, Siskind V, MacLennan R. 1995. How accurate is self-reported family history of colorectal cancer? Am J Epidemiol 141:863–871. Al-Tassan N, Chmiel NH, Maynard J, et al. 2002. Biallelic germline mutations in MYH predispose to multiple colorectal adenoma and somatic G:C-toT:A mutations. Hum Molec Genet 11:2961–2967. Anderson H, Bladstrom A, Olsson H, Miller TR. 2000. Familial breast and ovarian cancer: A Swedish population-based register study. Am J Epidemiol 152:1154–1163. Anglian Breast Cancer Study Group. 2000. Prevalence and penetrance of BRCA1 and BRCA2 mutations in a population-based series of breast cancer cases. Br J Cancer 83:1301–1308. Antoniou AC, Easton DF. 2003. Polygenic inheritance of breast cancer: Implications for the design of association studies. Gen Epidemiol 25:190–202. Athma P, Rappaport R, Swift M. 1996. Molecular genotyping shows that ataxia-telangiectasia heterozygotes are predisposed to breast cancer. Cancer Genet Cytogenet 92(2):130–134. Auerbach AD. 1993. Fanconi anemia diagnosis and the diepoxybutane (DEB) test. Exp Hematol 21(6):731–733. Aziz S, Kuperstein G, Rosen B, et al. 2001. A genetic epidemiology study of carcinoma of the fallopian tube. Gynecol Oncol 80:341–345. Bahuau M, Vidaud D, Jenkins RB, et al. 1998. Germ-line deletion involving the INK4 locus in familial proneness to melanoma and nervous system tumours. Cancer Res 58:2298–2303.

573

Bassett JHD, Forbes SA, Pannett AAJ, et al. 1998. Characterization of mutations in patients with multiple endocrine neoplasia type 1. Am J Hum Genet 62:232–244. Bataille V. 2000. Genetics of familial and sporadic melanoma. Clin Exp Dermatol 25(6):464–470. Baysal BE. 2001. Genetics of familial paragangliomas: past, present, and future. Otolaryngol Clin North Am 34(5):863–879. Baysal BE, Willett–Brozick JE, Lawrence EC, et al. 2002. Prevalence of SDHB, SDHC, and SDHD germline mutations in clinic patients with head and neck paragangliomas. J Med Genet 39:178–183. Bell DW, Varley JM, Szydlo TE, et al. 1999. Heterozygous germ line hCHK2 mutations in Li-Fraumeni syndrome. Science 286(5449):2528–2531. Bergmann MM, Calle EE, Mervis CA, Miracle-McMahill HL, Thun MJ, Heath CW. 1998. Validity of self-reported cancers in a prospective cohort study in comparison with data from state cancer registries. Am J Epidemiol 147:556–562. Berman DB, Costalas J, Schultz DC, Grana G, Daly M, Godwin AK. 1996. A common mutation in BRCA2 that predisposes to a variety of cancers is found in both Jewish Ashkenazi and non-Jewish individuals. Cancer Res 56:3409–3414. Birch JM, Hartley AL, Tricker KJ, et al. 1994. Prevalence and diversity of constitutional mutations in the p53 gene among 21 Li-Fraumeni families. Cancer Res 54:1298–1304. Bishop DT, Demenais F, Goldstein AM, et al. 2002. The Melanoma Genetics Consortium. Geographical variation in the penetrance of CDKN2A mutations for melanoma. J Natl Cancer Inst 94(12):894–903. Bondy ML, Strom SS, Colopy MW, Brown BW, Strong LC. 1994. Accuracy of family history of cancer obtained through interviews with relatives of patients with childhood sarcoma. J Clin Epidemiol 47:89–96. Borg A, Sandberg T, Nilsson T, et al. 2000. High frequency of multiple melanomas and breast and pancreatic carcinomas in CDKN2A mutationpositive melanoma families. J Natl Cancer Inst 92:1260–1266. Bratt O, Kristofferson U, Lundgren R, Olsson H. 1999. Familial and hereditary prostate cancer in Southern Sweden: A population-based case-control study. Eur J Cancer 35:272–277. Breast Cancer Linkage Consortium. 1999. Cancer Risks in BRCA2 mutation carriers. J Natl Cancer Inst 91:1310–1316. Brett MCA, Hershman MJ, Glazer G. 1994. Other manifestations of familial adenomatous polyposis. In: Phillips FSK, Spigelman A, Thomason JPS, eds. Familial Adenomatous Polyposis and Other Polyposis Syndromes. London: Edward Arnold, p. 146. Brose MS, Rebbeck TR, Calzone KA, Stopfer JE, Nathanson KL, Weber BL. 2002. Cancer risk estimates for BRCA1 mutation carriers identified in a risk evaluation program. J Natl Cancer Inst 94:1365–1372. Bulow S, Alm T, Fausa O, et al. 1995. Duodenal adenomatosis in familial adenomatous polyposis. DAF Project Group. Int J Colorectal Dis 10: 43–46. Burke W, Daly W, Garber J, et al. 1997. Recommendations for follow-up care of individuals with an inherited predisposition to cancer. BRCA1 and BRCA2. JAMA 277:997–1003. Burn J, Chapman P, Delhanty J, et al. 1991. The UK Northern region genetic register for familial adenomatous polyposis coli: Use of age of onset, congenital hypertrophy of the retinal pigment epithelium, and DNA markers in risk calculations. J Med Genet 28:289–296. Bussey HJR. 1975. Familial polyposis coli. Family studies, histopathology, differential diagnosis and results of treatment. Baltimore: The Johns Hopkins University Press. Caldas C, Carneiro F, Lynch HT, et al. 1999. Familial gastric cancer: Overview and guidelines for management. J Med Genet 36:873–880. Calender A, Giraud S, Cougard P, et al. 1995. Multiple endocrine neoplasia type 1 in France: clinical and genetic studies. J Intern Med 238(3):263–268. Cetta F, Montalto G, Gori M, Curia MC, Cama A, Olschwang S. 2000. Germline mutations of the APC gene in patients with familial adenomatous polyposis-associated thyroid carcinoma: Results from a European cooperative study. J Clin Endocrinol Metab 85:286–292. Chompret A, Abel A, Stoppa-Lyonnet D, et al. 2001. Sensitivity and predictive value of criteria for p53 germline mutation screening. J Med Genet 38(1):43–47. Chun YS, Lindor NM, Smyrk TC, et al. 2001. Germline E-cadherin gene mutations: is prophylactic total gastrectomy indicated? Cancer 92:181–187. Church J, McGannon E. 2000. Family history of colorectal cancer: How often and how accurately is it recorded? Dis Colon Rectum 43:1540–1544. Cook JA, Oliver K, Mueller RF, Sampson J. 1996. A cross sectional study of renal involvement in tuberous sclerosis. J Med Genet 33:480–484. Cortessis V, Ingles S, Millikan R, et al. 1993. Linkage analysis of DRD2, a marker linked to the ataxia–telangiectasia gene, in 64 families with premenopausal bilateral breast cancer. Cancer Res 53(21):5083– 5086.

574

PART III: THE CAUSES OF CANCER

Cowan R, Hoban P, Kelsey A, Birch JM, Gattamaneni R, Evans DG. 1997. The gene for the naevoid basal cell carcinoma syndrome acts as a tumoursuppressor gene in medulloblastoma. Br J Cancer 76:141–145. Czene K, Lichtenstein P, Kemminki K. 2002. Environmental and heritable causes of cancer among 9.6 million individuals in the Swedish FamilyCancer Database. Int J Cancer 99:260–266. Czene K, Hemminki K. 2002. Kidney cancer in the Swedish Family-Cancer Database: familial risks and second primary malignancies. Kidney Int 61:1806–1813. Czene K, Hemminki K. 2003. Familial papillary renal cell tumors and subsequent cancers: A nationwide epidemiological study from Sweden. J Urol 169:1271–1275. Dong C, Hemminki K. 2001. Modification of cancer risk in offspring by sibling and parental cancers from 2,112,616 nuclear families. Int J Cancer 92:144–150. Douglas FS, O’Dair LC, Robinson M, Evans DGR, Lynch SA. 1999. The accuracy of diagnoses as reported in families with cancer: A retrospective study. J Med Genet 36:309–312. Easton DF, Ford D, Bishop DT. 1995. Breast and ovarian cancer incidence in BRCA1 mutation carriers. Breast Cancer Linkage Consortium. Am J Hum Genet 56:265–271. Easton DF. 1994. Cancer risks in A-T heterozygotes. Int J Radiat Biol 66:S177–S182. Eeles RA. 1995. Germline mutations in the TP53 gene. Cancer Surv 25:101–124 Eng C. 2000. Will the real Cowden syndrome please stand up: Revised diagnostic criteria. J Med Genet 37:828–830. Eng C, Schneider K, Fraumeni JF Jr, Li FP. 1997. Third international workshop on collaborative interdisciplinary studies of p53 and other predisposing genes in Li-Fraumeni syndrome. Cancer Epidemiol Biomarkers Prev 6:379–383. Eng C, Clayton D, Schuffenecker I, et al. 1996. The relationship between specific RET proto-oncogene mutations and disease phenotype in multiple endocrine neoplasia type 2. International RET mutation consortium analysis. JAMA 276:1575–1579. Evans DG, Kerr B, Cade D, Hoare E, Hopwood P. 1996. Fictitious breast cancer family history. Lancet 348:1034. Evans DG, Ladusans EJ, Rimmer S, Burnell LD, Thakker N, Farndon PA. 1993. Complications of the naevoid basal cell carcinoma syndrome: results of a population based study. J Med Genet 30(6):460–464. Fitzgerald MG, Bean JM, Hegde SR, et al. 1997. Heterozygous ATM mutations do not contribute to early onset of breast cancer. Nat Genet 15:307–310. Fitzgerald MG, MacDonald DJ, Krainer M, et al. 1996. Germline BRCA1 mutations in Jewish and non-Jewish women with early-onset breast cancer. N Engl J Med 334:143–149. Ford D, Easton DF, Bishop DT, et al. 1994. Risks of cancer in BRCA1mutation carriers. Lancet 343:692–695. Frebourg T, Barbier N, Yan YX, et al. 1995. Germline p53 mutations in 15 families with Li-Fraumeni syndrome. Am J Hum Genet 56:608–615. Garber JE, Goldstein AM, Kantor AF, Dreyfus MG, Fraumeni JF Jr, Li FP. 1991. Follow-up study of twenty-four families with Li-Fraumeni syndrome. Cancer Res 51:6094–6097. Gatti R. Ataxia-telangiectasia. 2001. In: Scriver CR, Beaudet AL, Sly WS, Valle D, eds. The Metabolic and Molecular Bases of Inherited Disease. 7th ed. New York: McGraw-Hill, pp. 705–732. German J, Ellis NA. 2001. Bloom Syndrome. In: Scriver CR, Beaudet AL, Sly WS, Valle D, eds. Metabolic & Molecular Bases of Inherited Disease. New York: McGraw-Hill, pp. 733–752. Giusti RM, Rutter JL, Duray PH, et al. 2003. A twofold increase in BRCA mutation related prostate cancer among Ashkenazi Israelis is not associated with distinctive histopathology. J Med Genet 40:787–792. Goldgar DE, Easton DF, Cannon-Albright LA, Skolnick MH. 1994. Systematic population-based assessment of cancer risk in first-degree relatives of cancer probands. J Natl Cancer Inst 86:1600–1608. Goldgar DE. 2002. Population aspects of cancer genetics. Biochimie 84:19–25. Goldstein AM, Fraser MC, Struewing JP, et al. 1995. Increased risk of pancreatic cancer in melanoma–prone kindreds with p16INK4a mutations. N Engl J Med 333:970–974. Gomez MR. 1998. Tuberous Sclerosis. 2nd ed. New York: Raven Press. Gorlin RJ. 1987. Nevoid basal-cell carcinoma syndrome. Medicine 66(2): 98–113. Goto M, Tanimoto K, Horiuchi Y, Sasazuki T. 1981. Family analysis of Werner’s syndrome: A survey of 42 Japanese families with a review of the literature. Clin Genet 19(1):8–15. Goto M, Miller RW, Ishikawa Y, Sugano H. 1996. Excess of rare cancers in Werner syndrome (adult progeria). Cancer Epidemiol Biomarkers Prev 5:239–246.

Grann VR, Jacobson JS. 2002. Population screening for cancer-related germline gene mutations. Lancet Oncol 3:341–348. Gruber SB, Ellis NA, Rennert G, Offit K. 2002. BLM heterozygosity and the risk of colorectal cancer. Science 297:2013. Guilford P, Hopkins J, Harraway J, et al. 1998. E-cadherin germline mutations in familial gastric cancer. Nature 392:402–405. Gurbuz AK, Giardiello FM, Petersen GM, et al. 1994. Desmoid tumours in familial adenomatous polyposis. Gut 35:377–381. Gutmann DH, Collins FS. 2001. Neurofibromatosis 1. In: ScriverCR, Beaudet AL, Sly WS, Valle D, Childs B, Kinzler KW, Vogelstein B, eds. The Metabolic and Molecular Bases of Inherited Disease. 8th ed. New York: McGraw-Hill, pp. 877–896. Hamilton SR, Liu B, Parsons RE, et al. 1995. The molecular basis of Turcot’s syndrome. N Engl J Med 332:839–847. Heim RA, Lench NJ, Swift M. 1992. Heterozygous manifestations in four autosomal recessive human cancer–prone syndromes: ataxia telangiectasia, xeroderma pigmentosum, Fanconi anemia and Bloom syndrome. Mutat Res 284:25–36. Hemminki K, Vaittinen P. 1997. Effect of paternal and maternal cancer on cancer in the offspring: A population-based study. Cancer Epidemiol Biomarkers Prev 6:993–997. Hemminki K, Vaittinen P. 1998. Familial risks for in situ cancers from the Family-Cancer Database. Cancer Epidemiol Biomarkers Prev 7:865–868. Hemminki K, Vaittinen P. 1999. Familial cancers in a nationwide Family-Cancer Database: age distribution and prevalence. Eur J Cancer 35:1109–1117. Hemminki K, Vaittinen, Dong C. 1999. Endometrial cancer in the FamilyCancer Database. Cancer Epidemiol Biomarkers Prev 8:1005–1010. Hemminki K, Dong C. 2000. Familial prostate cancer from the Family-Cancer Database. Eur J Cancer 36:229–234. Hemminki K, Dong C, Vaittinen P. 2001. Cancer risks to spouses and offspring in the Family-Cancer Database. Genetic Epidemiol 20:247–257. Hemminki K, Li X. 2001. A population-based study of familial soft tissue tumors. J Clin Epidemiol 54:411–416. Hemminki K, Li X. 2001. Familial carcinoid tumors and subsequent cancers: A nationwide epidemiologic study from Sweden. Int J Cancer 94:444–448. Hemminki K, Li X. 2001. Increased cancer risk in the offspring of women with colorectal carcinoma. Cancer 91:1061–1063. Hemminki K, Li X. 2001. Familial colorectal adenocarcinoma from the Swedish Family-Cancer Database. Int J Cancer 94:743–748. Hemminki K, Jiang Y. 2002. Risks among siblings and twins for childhood acute lymphoblastic leukemia. Leukemia (Baltimore) 16:297–298. Hemminki K, Czene K. 2002. Attributable risks of familial cancer from the Family-Cancer Database. Cancer Epidemiol Biomarkers Prev 11:1638– 1644. Hemminki K, Czene K. 2002. Age-specific and attributable risks of familial prostate carcinoma from the Family-Cancer Database. Cancer 95:1346– 1353. Hemminki K, Granstrom C, Czene K. 2002. Attributable risks for familial breast cancer by proband status and morphology: A nationwide epidemiologic study from Sweden. Int J Cancer 100:214–219. Hemminki K, Jiang Y. 2002. Familial and second gastric carcinomas: A nationwide epidemiologic study from Sweden. Cancer 94:1157–1165. Hemminki K, Jiang Y. 2002. Familial and second esophageal cancers: A nationwide epidemiologic study from Sweden. Int J Cancer 98:106–109. Hemminki K, Li X. 2002. Gender effects in familial cancer. Int J Cancer 102:184–187. Hemminki K, Li X. 2002. Cancer risks in twins: Results from the Swedish Family-Cancer Database. Int J Cancer 99:873–878. Hemminki K, Li Z. 2003. Familial risk of cancer by site and histopathology. Int J Cancer 103:105–109. Hisada M, Garber JE, Fung CY, Fraumeni JF Jr, Li FP. 1998. Multiple primary cancers in families with Li-Fraumeni syndrome. J Nat Cancer Inst 90(8): 606–611. Hopper JL, Southey MC, Dite GS, et al. 1999. Population-based estimate of the average age-specific cumulative risk of breast cancer for a defined set of protein-truncating mutation in BRCA1 and BRCA2. Australian Breast Cancer Family Study. Cancer Epidemiol Biomarkers Prev 8:741–747. Howlett NG, Taniguchi T, Olson S, et al. 2002. Biallelic inactivation of BRCA2 in Fanconi anemia. Science 297:606–609. Hughes LJ, Michels VV. 1992. Risk of hepatoblastoma in FAP. Am J Med Genet 43:1023–1025. Huntsman DG, Carneiro F, Lewis FR, et al. 2001. Early gastric cancer in young, asymptomatic carriers of germ-line E-cadherin mutations. N Engl J Med 344:1904–1909. Ivanovich J, Babb S, Goodfellow P, et al. 2002. Evaluation of the family history data collection process and the accuracy of reporting among a series of women with endometrial cancer. Clin Cancer Res 8:1849–1856.

Hereditary Neoplastic Syndromes Izatt L, Greenman J, Hodgson S, et al. 1999. Identification of germline missense mutations and rare allelic variants in the ATM gene in early-onset breast cancer. Genes Chromosomes Cancer 26(4):286–294. Kadan-Lottick NS, Friedman DI, Mertens AC, et al. 2003. Self-reported family history of cancer: The utility of probing questions. Epidemiology 14: 737–740. Kadmon M, Tandara A, Herfarth C. 2001. Duodenal adenomatosis in familial adenomatous polyposis coli. A review of the literature and results from the Heidelberg Polyposis Register. Int J Colorectal Dis 16:63–75. Katballe N, Juul S, Christensen M, Orntoft TF, Wikman FP, Laurberg S. 2001. Patient accuracy of reporting on hereditary non-polyposis colorectal cancer-related malignancy in family members. Br J Surgery 88:1228– 1233. Kauf ND, Perez-Segura P, Robson ME, et al. 2002. Incidence of non-founder BRCA1 and BRCA2 mutations in high-risk Ashkenazi breast and ovarian cancer families. J Med Genet 39:611–614. Kaufman DK, Kimmel DW, Parisi JE, Michels VV. 1993. A familial syndrome with cutaneous malignant melanoma and cerebral astrocytoma. Neurology 43:1728–1731. Kerber RA, Slattery ML. 1997. Comparison of self-reported and databaselinked family history of cancer data in a case-control study. Am J Epidemiol 146:244–248. Kerr B, Foulkes WD, Cade D, et al. 1998. False family history of breast cancer in the family cancer clinic. Eur J Surg Oncol 24:275–279. Khoury MJ, McCabe LL, McCabe ERB. 2003. Population screening in the age of genomic medicine. N Engl J Med 348:50–58. King TM, Tong L, Pack RJ, Spencer C, Amos CI. 2002. Accuracy of family history of cancer as reported by men with prostate cancer. Urology ••:546–550. Kleihues P, Schauble B, zur Hausen A, Esteve J, Ohgaki H. 1997. Tumors associated with p53 germline mutations: A synopsis of 91 families. Am J Pathol 150:1–13. Kleinerman RA, Tarone RE, Abramson DH, Seddon JM, Li FP, Tucker MA. 2000. Hereditary retinoblastoma and risk of lung cancer. J Nat Cancer Inst 92(24):2037–2039. Korf BR. 2000. Malignancy in neurofibromatosis type 1. Oncologist 5:477–485. Kraemer KH, Lee MM, Scotto J. 1987. Xeroderma pigmentosum—cutaneous, ocular and neurologic abnormalities in 830 published cases. Arch Derm 123:241–250. Kraemer KH, Lee MM, Andrews AD, Lambert WC. 1994. The role of sunlight and DNA repair in melanoma and non-melanoma skin cancer. The xeroderma pigmentosum paradigm. Arch Dermatol 130:1018–1021. Kuffel DG, Lindor NM, Litzow MR, Dewald GW. 1997. Mitomycin C chromosome stress test to identify hypersensitivity to bifunctional alkylating agents in patients with Fanconi anemia or aplastic anemia. Mayo Clinic Proc 72:579–580. Kutler DI, Singh B, Satagopan J, et al. 2003. A 20-year perspective on the International Fanconi Anemia Registry (IFAR). Blood 101:1249–1256. Laken SJ, Petersen GM, Gruber SB, et al. 1997. Familial colorectal cancer in Ashkenazim due to a hypermutable tract in APC. Nat Genet 17:79–83. Lee SB, Kim SH, Bell DW, et al. 2001. Destabilization of CHK2 by a missense mutation associated with Li-Fraumeni Syndrome. Cancer Res 61:8062–8067. Li FP, Fraumeni JF Jr, Mulvihill JJ, et al. 1988. A cancer family syndrome in 24 kindreds. Cancer Res 48:5358–5362. Li FP, Fraumeni JF Jr. 1982. Prospective study of a family cancer syndrome. JAMA 24:2692–2694. Li FP, Fraumeni JF Jr. 1969. Soft-tissue sarcomas, breast cancer, and other neoplasms. A familial syndrome? Ann Intern Med 71:747–752. Li X, Hemminki K. 2002. Cancer risks in women who had children with different partners from the Swedish Family-Cancer Database. Eur J Cancer Prev 11:433–438. Listernick R, Gutmann DH. 1999. Tumors of the optic pathway. In: Friedman JM, Gutmann DH, MacCollin M, Riccardi VM, eds. Neurofibromatosis: Phenotype, Natural History, and Pathogenesis. Baltimore: Johns Hopkins University Press, pp. 203–230. Love R, Evans AM, Josten DM. 1985. The accuracy of patient reports of a family history of cancer. J Chron Dis 38:289–293. Lynch HT, Snyder CL, Lynch JF, Riley BD, Rubinstein W. 2003. Hereditary breast-ovarian cancer at the bedside: Role of the medical oncologist. J Clin Oncol 21:740–753. Lynch HT, Lynch JF. 2002. Hereditary cancer: Family history, diagnosis, molecular genetics, ecogenetics and management strategies. Biochimie 84:3–17. Lynch HT, Brand RE, Hogg D, et al. 2002. Phenotypic variation in eight extended CDKN2A germline mutation familial atypical multiple mole melanoma-pancreatic carcinoma-prone families: The familial atypical mole melanoma pancreatic carcinoma syndrome. Cancer 92:84–96.

575

Lynch HT, Smyrk T, Lynch J. 1996. Hereditary nonpolyposis colorectal cancer (Lynch Syndrome)—An updated review. Cancer 1149–1167. Malkin D, Li FP, Strong LC, et al. 1990. Germline p53 mutations in a familial syndrome of breast cancer, sarcomas, and other neoplasms. Science 250:1233–1238. Malkin D, Friend SH, Li FP, Strong LC. 1997. Germ-line mutations of the p53 tumor-suppressor gene in children and young adults with second malignant neoplasms. N Engl J Med 336:734. Manski TJ, Heffner DK, Glenn GM, et al. 1997. Endolymphatic sac tumors. A source of morbid hearing loss in von Hippel-Lindau disease. JAMA 277(18):1461–1466. Marsh DJ, Zori RT. 2002. Genetic insights into familial cancer: update and recent discoveries. Cancer Letters 181:125–164. Michels VV. 1987. Hippel-Lindau disease. In: Gomez M, ed. Neurocutaneous Diseases. Boston: Butterworth, pp. 53–66. Modan B, Hartge P, Hirsh-Yechezkel G, et al. 2001. Parity, oral contraceptives and the risk of ovarian cancer among carriers and non-carriers of a BRCA1 or BRCA2 mutation. N Engl J Med 345:235–240. Modigliani E, Vasen HM, Raue K, et al. 1995. Pheochromocytoma in multiple endocrine neoplasia type 2: European study. The Euromen Study Group. J Intern Med 238:363–367. Moll AC, Imhof SM, Schouten-Van Meeteren AY, Kuik DJ, Hofman P, Boers M. 2001. Second primary tumors in hereditary retinoblastoma: a register-based study, 1945–1997: Is there an age effect on radiation-related risk? Ophthalmology 108(6):1109–1114. Moslehi R, Chu W, Karlan B, et al. 2000. BRCA1 and BRCA2 mutation analysis of 208 Ashkenazi Jewish women with ovarian cancer. Am J Hum Genet 66:1259–1272. Mulvihill JJ, Parry DM, Sherman JL, Pikus A, Kaiser-Kupfer MI, Eldridge R. 1990. NIH conference. Neurofibromatosis 1 (Recklinghausen disease) and neurofibromatosis 2 (bilateral acoustic neurofibromatosis). An update. Ann Intern Med 113:39–52. Napier BA, Metzner H, Johnson BC. 1972. Limitation of morbidity and mortality data obtained from family histories—a report from the Tecumseh Community Health Study. Am J Public Health 62:30–35. Nelen MR, Padberg GW, Peeters EA, et al. 1996. Localization of the gene for Cowden disease to chromosome 10q22-q23. Nat Genet 13:114–116. Newman B, Mu H, Butler LM, Millikan RC, Moorman PG, King M-C. 1998. Frequency of breast cancer attributable to BRCA1 in a population-based series of American women. JAMA 279:915–921. Nichols KE, Malkin D, Garber JE, Fraumeni JF Jr, Li FP. 2001. Germ-line p53 mutations predispose to a wide spectrum of early-onset cancers. Cancer Epidemiol Biomarkers Prev 10:83–87. Novakovic B, Goldstein AM, Tucker MA. 1996. Validation of family history of cancer in deceased family members. J Natl Cancer Inst 88:1492–1493. Oddoux C, Clayton CM, Nelson HR, Ostrer H. 1999. Prevalence of Bloom syndrome heterozygotes among Ashkenazi Jews. Am J Hum Genet 64:1241–1243. Offerhaus GJ, Entius MM, Giardiello FM. 1999. Upper gastrointestinal polyps in familial adenomatous polyposis. Hepatogastroenterology 46:667–669. Offit K, Levran O, Mullaney B, et al. 2003. Shared genetic susceptibility to breast cancer, brain tumors and Fanconi anemia. J Natl Cancer Inst 95:1548–1551. Offit K. 1998. Clinical cancer genetics and risk counseling. In: Offit K, ed. Clinical Cancer Genetics—Risk Counseling and Management. New York: Wiley-Liss, pp. 1–20. Oksuzoglu B, Yalcin S. 2002. Squamous cell carcinoma of the tongue in a patient with Fanconi’s anemia: A case report and review of the literature [Review]. Ann Hematol 81(5):294–298. Oosterwijk JC, Jansen JC, van Schothorst EM, et al. 1996. First experiences with genetic counselling based on predictive DNA diagnosis in hereditary glomus tumours (paragangliomas). J Med Genet 33(5):379–383. Parent ME, Ghadirian P, Lacroix A, Perret C. 1997. The reliability of recollections of family history: implications for the medical provider. J Cancer Educ 12:114–120. Park JG, Park KJ, Ahn YO, et al. 1992. Risk of gastric cancer among Korean familial adenomatous polyposis patients. Report of three cases. Dis Colon Rectum 35:996–998. Parmigiani G, Berry DA, Aguilar O. 1998. Determining carrier probabilities for breast cancer susceptibility genes BRCA1 and BRCA2. Am J Hum Genet 62:145–158. Peto J, Collins N, Barfoot R, et al. 1999. Prevalence of BRCA1 and BRCA2 mutations in patients with early-onset breast cancer. J Natl Cancer Inst 91:943–949. Peto J, Houlston RS. 2001. Genetics and the common cancers. Eur J Cancer 37:S88–S96. Pharoah PD, Guilford P, Caldas C. 2001. International Gastric Cancer Linkage Consortium. Incidence of gastric cancer and breast cancer in CDH1

576

PART III: THE CAUSES OF CANCER

(E-cadherin) mutation carriers from hereditary diffuse gastric cancer families. Gastroenterology 121(6):1348–1353. Plna K, Hemminki K. 2001. Familial bladder cancer in the national Swedish Family-Cancer Database. J Urol 166:2129–2133. Risch HA, McLaughlin JR, Cole DE, et al. 2001. Prevalence and penetrance of germline BRCA1 and BRCA2 mutations in a population series of 649 women with ovarian cancer. Am J Hum Genet 68:700–710. Risch N. 2001. The genetic epidemiology of cancer: interpreting family and twin studies and their implications for molecular genetics approaches. Cancer Epidemiol Biomarkers Prev 10:733–741. Roa BB, Boyd AA, Volcik K, Richards CS. 1996. Ashkenazi Jewish population frequencies for common mutations in BRCA1 and BRCA2. Nat Genet 14:185–187. Roach ES, Gomez MR, Northrup H. 1998. Tuberous sclerosis complex consensus conference: revised clinical diagnostic criteria. J Child Neurol 13:624–628. Roarty JD, McLean IW, Zimmerman LE. 1998. Incidence of second neoplasms in patients with bilateral retinoblastoma. Ophthalmology 95:1583. Rodriguez–Bigas MA, Boland CR, et al. 1997. A National Cancer Institute Workshop on Hereditary Nonpolyposis Colorectal Cancer Syndrome: Meeting highlights and Bethesda Guidelines. J Natl Cancer Inst 89: 1758–1762. Schmidt L, Junker K, Weirich G, et al. 1998. Two North American families with hereditary papillary renal carcinoma and identical novel mutations in the MET proto-oncogene. Cancer Res 58(8):1719–1722. Schrager CA, Schneider D, Gruener AC, Tsou HC, Peacocke M. 1998. Clinical and pathological features of breast disease in Cowden’s syndrome: an under recognized syndrome with an increased risk of breast cancer. Hum Pathol 29(1):47–53. Scriver CR, Beaudet AL, Sly WS, Valle D, eds. 1995. Metabolic and Molecular Basis of Inherited Disease. 7th ed. New York: McGraw-Hill. Shepherd CW, Gomez MR, Lie JT, Crowson CS. 1991. Causes of death in patients with tuberous sclerosis. Mayo Clin Proc 66:792–796. Sijmons RH, Boonstra QE, Reefhuis J, et al. 2000. Accuracy of family history of cancer: Clinical genetic implications. Eur J Hum Genet 8: 181–186. Skogseid B, Oberg K. 1995. Experience with multiple endocrine neoplasia type 1 screening. J Intern Med 238:255–261. Smith TG, Clark SK, Katz DE, Reznek RH, Phillips RK. 2000. Adrenal masses are associated with familial adenomatous polyposis. Dis Colon Rectum 43:1739–1742. Stagopan JM, Offit K, Foulkes W, et al. 2001. The lifetime risks of breast cancer in Ashkenazi Jewish carriers of BRCA1 and BRCA2 mutations. Cancer Epidemiol Biomarkers Prev 10:467–473. Stankovic T, Kidd AM, Sutcliffe A, et al. 1998. ATM mutations and phenotypes in ataxia–telangiectasia families in the British Isles: Expression of mutant ATM and the risk of leukemia, lymphoma, and breast cancer. Am J Hum Genet 62(2):334–345. Strachan T, Read AP. 2004. Human Molecular Genetics. 3rd ed. New York: Garland Science. Swift M, Morrell D, Massey RB, Chase CL. 1991. Incidence of cancer in 161 families affected by ataxia-telangiectasia. N Engl J Med 325:1831–1836. Szabo CI, King M–C. 1997. Invited Editorial. Population genetics of BRCA1 and BRCA2. Am J Hum Genet 60:1013–1020.

Theis B, Boyd N, Lockwood G, Tritchler D. 1994. Accuracy of family history in breast cancer patients. Eur J Cancer Prev 3:321–327. Thorlacius S, Struewing JP, Hartge P, et al. 1998. Population-based study of risk of breast cancer in carriers of BRCA2 mutation. Lancet 352: 1337–1339. Tucker MA, Fraser MC, Goldstein AM, et al. 2002. A natural history of melanoma and dysplastic nevi: An atlas of lesions in melanoma-prone families. Cancer 94:3192–3209. Umar A, Boland CR, Terdiman JP, et al. 2004. Revised Bethesda Guidelines for hereditary nonpolyposis colorectal cancer (Lynch syndrome) and microsatellite instability. J Natl Cancer Inst 96(4):261–268. Vahteristo P, Tamminen A, Karvinen P, et al. 2001. p53, CHK2, and CHK1 genes in Finnish families with Li-Fraumeni syndrome: Further evidence of CHK2 in inherited cancer predisposition. Cancer Res 61(15):5718– 5722. Vaittinen P, Hemminki K. 1999. Familial cancer risks in offspring from discordant parental cancers. Int J Cancer 81:12–19. Van der Mey AG, Maaswinkel-Mooy PD, Cornelisse CJ, Schmidt PH, van de Kamp JJ. 1989. Genomic imprinting in hereditary glomus tumors: evidence for new genetic theory. Lancet 2(8675):1291–1294. Varley JM, McGown G, Thorncroft M, et al. 1999. Are there low-penetrance TP53 alleles? Evidence from childhood adrenocortical tumors. Am J Hum Genet 65(4):995–1006. Vasen HFA, Wijnen JT, Menko FH, et al. 1996. Cancer risk in families with hereditary nonpolyposis colorectal cancer diagnosed by mutation analysis. Gastroenterology ••:1020–1027. Vasen HFA, Mecklin JP, Meera Khan P, Lynch HT. 1991. The International Collaborative Group on Hereditary Non-Polyposis Colorectal Cancer. Dis Colon Rectum 34:424–425. Vennos EM, Collins M, James WD. 1992. Rothmund-Thomson Syndrome: review of the world literature. J Am Acad Dermatol 27:750–762. Vorechovsky I, Luo L, Lindblom A, et al. 1996. ATM mutations in cancer families. Cancer Res 56(18):4130–4133. Wang LL, Levy ML, Lewis RA, et al. 2001. Clinical manifestations in a cohort of 41 Rothmund-Thomson syndrome patients. Am J Med Genet 102(1): 11–17. Warner E, Foulkes W, Goodwin P, et al. 1999. Prevalence and penetrance of BRCA1 and BRCA2 gene mutations in unselected Ashkenazi Jewish women with breast cancer. J Natl Cancer Inst 91:1241–1247. Weber W, Estoppey J, Stoll H. 2001. Familial cancer diagnosis. Anticancer Res 21:3631–3636. Whelan AJ, Bartsch D, Goodfellow PJ, et al. 1995. A familial syndrome of pancreatic cancer and melanoma with a mutation in the CDKN2 tumorsuppressor gene. N Engl J Med 333:975–977. Wijnen JT, Vasen HFA, Kahn M, et al. 1998. Clinical findings with implications for genetic testing in families with clustering of colorectal cancer. N Engl J Med 339:511–518. Wong FL, Boice JD Jr, Abramson DH, et al. 1997. Cancer incidence after retinoblastoma. Radiation dose and sarcoma risk. JAMA 278(15):1262– 1267. Wooster R, Ford D, Mangion J, et al. 1993. Absence of linkage to the ataxia telangiectasia locus in familial breast cancer. Hum Genet 92(1):91–94. Ziogas A, Anton-Culver H. 2003. Validation of family history data in cancer family registries. Am J Prev Med 24:190–198.

29

Genetic Modifiers of Cancer Risk NEIL E. CAPORASO

D

espite decades of accumulated evidence suggesting important contributions of inherited susceptibility to most types of cancer, the precise genetic determinants and how they act in concert with environmental exposures and host factors remain to be described. A notable area of success is the identification of major genes that follow Mendelian patterns of inheritance for certain hereditary cancer syndromes as well as somatically altered genes universally present in tumors (Futreal, 2004). However, all the identified genes together only account for a few percent of all cancer (Peto and Houlston, 2001). Understanding the contribution of genetic susceptibility to common disease and cancer outcomes was a key motivation for sequencing of the human genome (Green, 1991; Khoury, 1999) and emerging genome science (Gibson and Muse, 2002); however, establishing and verifying the precise relationships of genes, their variants, and disease phenotypes has proved to be challenging (Lander and Schork, 1994; Risch and Merikangas, 1996; Houlston and Tomlins, 2000). There is only weak to mixed evidence for verifiable associations of specific susceptibility genes with common human diseases, and neither of two recent reviews identified even one association considered “proven” for a common human cancer (Hirschhorn, 2002, Lohmueller, 2003). Nevertheless, published metanalyses support at least a few specific associations, such as increased risk of bladder cancer in NAT2 “deficient” subjects (slow acetylators). Many other genes or pathways are mechanistically plausible modifiers of one or more cancers but to date have been insufficiently studied in human populations to establish causal associations. Understanding how gene–environment and gene–gene effects contribute to susceptibility also awaits larger studies. It is anticipated that a broader understanding of the contribution of genetic modifiers to human cancer will improve mechanistic understanding, shed light on new modes of chemoprevention (Lippman and Spitz, 2001) and therapy, identify subgroups at altered risk, and help identify heretofore unsuspected environmental agents that contribute to risk.

CONTEXT The more frequent but less characterized genes that modify “sporadic” cancer, rather than the rare high-penetrance alleles of major genes that underlie familial cancer, is the topic of this chapter. Some of the distinctions between the two classes of inherited genes are depicted in Table 29–1 (Caporaso, 1995; Vineis, 2002). The high-penetrance genes that contribute to hereditary cancer syndromes involve only a few percent of all cancer. Potter (2001) observed that certain high-risk genes and their associated cancers (e.g., Rb and retinoblastoma) are minimally influenced by the environment, whereas others exhibit more variable penetrance (e.g., BRCA1 and breast cancer) as a result of other genes or environmental influences. In general, the role of the environment for the high-penetrance genes is less prominent, whereas for the common adult-onset tumors, putative susceptibility genes act in concert with exposures and other modifier genes to alter risk. The relative and absolute risks conferred by these genes are small to modest, and familial aggregation is not a conspicuous feature. The lower absolute risks (and lack of major implications for family members) make informed consent requirements somewhat less stringent for the study of susceptibility genes in the population (Beskow, 2001): however complex; psychological and social issuer are emerging with

expanded genetic testing (Lerman and Shields, 2004). To detect these modestly augmented relative and absolute risks, substantial study sizes involving many hundreds to thousands of study subjects are required. The allelic variants are common; although rare genes might also have weak effects, they would be extremely difficult to detect reliably because of power considerations. To date, these polymorphic variants (which affect more than 1% of the population) have been associated with small changes in absolute and relative risk, so their effects have been too small to justify population screening even for the more common tumors. However, as the pharmacogenetic literature documents (Evans, 2003; Weinshilboum, 2003), common gene variants are responsible for substantial differences in drug metabolism or in receptor function, so that the eventual identification of strong effects on “phenotypes” with clinical impact cannot be ruled out. However, even modest effects cannot be dismissed. For example, the NAT2 deficient genotype showed an odds ratio of 1.4 in relation to bladder cancer (i.e., a 40% increased risk for the 50% of the population that are “slow acetylators”) (Marcus, 2000). Although this risk would not be considered clinically important, the slow acetylation mechanism might account for >15% of all bladder cancers.

RATIONALE FOR THE STUDY OF LOW-PENETRANCE GENES The study of modifier or susceptibility genes can provide major insights not only into genetic susceptibility but also into interactions with other causal factors in the carcinogenic process. This general aim subsumes more specific goals including (1) implicating specific environmental (sometimes unsuspected) or genetic factors amenable to interventions, (2) understanding determinants of risk in individuals and the population, (3) suggesting specific individual, clinical, or population-based strategies for prevention, screening, or treatment, (4) refining clinical and etiologic subclassifications of disease, (5) identifying critical intermediate phenotypes (endophenotypes) that contribute to disease and may be more closely related to particular genes, and (6) identifying subgroups for special interventions. As the genetic architecture of cancer is increasingly revealed using new molecular technologies in the context of epidemiologic study designs, it will emerge that understanding cancer etiology will be incomplete without taking genetics into account. Broadly, population studies will be needed for gene discovery, population risk characterization, and evaluation of genetic information for diagnosis and prevention (Khoury, 2003).

THE ROLE OF LOW-PENETRANCE GENES IN CANCER SUSCEPTIBILITY Although the precise contribution of genetic variation to cancer and the best models to measure it are matters of vigorous debate (Lichtenstein et al., 2000; Risch, 2001), a hereditary contribution is widely posited based on evidence that includes increased risk of specific cancers in relatives of cases compared to relatives of controls, twin studies, hereditary cancer syndromes, and animal models. Nonspecific types of evidence suggested to implicate genetic susceptibility include

577

578

PART III: THE CAUSES OF CANCER

Table 29–1. Features of Genetic Modifiers Contrasted with HighPenetrance Genes Cancer Modifier Terms Study design Goal of study Study population Absolute risk Relative risk Role of environment Gene frequency Number of genes Clinical role Attributable risk Familiality Type of gene alteration Potential testing application Ethical concerns

Minor gene Low-penetrance gene Population-based (association) Gene characterization General population Low Low Major Generally high Many Generally minor Potentially high Not prominent Polymorphism, typically present in >1% Susceptibility Less stringent

Cancer Gene Major gene High-penetrance gene Family-based (linkage) Gene discovery Families High High Minor Low to rare One (or few) Increasingly well established Low Apparent Mutation Diagnostic Highly stringent

early age of onset and multifocal development of cancer in a single organ. Although these clinical clues are often characteristic of family cancer syndromes (Lindor and Greene, 1998), it is not established whether these features are also associated with low-penetrance candidate genes. In contrast to highly familial conditions due to a gene or a small number of genes segregating at particular genetic loci, “complex” diseases such as sporadic forms of cancer result from the combined effects of multiple genetic and environmental factors. Virtually all cancers exhibit some degree of aggregation among related individuals (Risch, 2001). Even for the cancers where a well-recognized gene accounts for some familial aggregations, additional genes are likely to be present (Peto and Houlston, 2001). Dite et al. (2003) concluded that BRCA1 and BRCA2 account for less than one-third of the excess breast cancers diagnosed before age 40 in relatives of cases attributable to familial factors. Using a population-based series of breast cancer, Pharoah et al. (2002) concluded that a log-normal distribution of genetic risk exists in which the half of the population at highest risk based on susceptibility genes accounts for 88% of the risk. These and similar results suggest that additional modifier loci remain to be discovered in the setting of highly familial as well as sporadic cancers (Mack and Peto, 2000). For certain cancers, an environmental factor might account for familial aggregation, particularly when Mendelian expectations are violated. Despite examples where environmental carcinogens may cause familial cancer (e.g., spouses of asbestos-exposed workers) (Li et al., 1978), the origin of most familial aggregation is genetic. However, a more subtle sharing of behavioral risk factors (perhaps influenced by genetics) may contribute to familial clustering at tobacco- and alcohol-related cancer sites. A considerable component of familiality is yet to be explained, and there is substantial room for low-penetrance genes to play a role. The role of genes is further supported in that for many cancers, good evidence exists that polymorphic genes contribute to the metabolic activation or elimination of carcinogenic species implicated in their etiology. For example, known polymorphic genes contribute to the activation or elimination of aromatic amines (bladder cancer), aflatoxin (liver cancer), and polycyclic aromatic hydrocarbons, and nitrosamines (lung cancer). For cancers where specific environmental agents are poorly understood (e.g., leukemias and brain cancer), a potential gain from the identification of susceptibility genes could be the linked substrates or cofactors that may implicate as yet unrecognized etiologic factors. Epidemiologic studies that evaluate whether relatives of cases have excess cancer compared to relatives of controls generally support a familial component of risk that persists after adjustment for known

risk factors. Tokuhata and Lilienfeld (1963) first showed that lung cancer mortality was increased in both smoking and nonsmoking relatives of lung cancer cases. Since then, many other investigators have confirmed elevated risk of lung cancer in first-degree relatives of cases, taking into account potential confounders such as active and passive smoking, age, and gender (Samet et al., 1986; Shaw et al., 1991; Amos et al., 1992; Bromen et al., 2000; Etzel et al., 2003). Studies in large populations have been conducted to evaluate sexand age-specific risks of cancer in relatives of cancer probands such as the Family Cancer Database in Sweden, which includes 0.5 million primary adult cancers (Hemminki et al., 1999). Such large-scale databases can take gender and age into account, although this design cannot consider risk factors such as shared environmental exposures (e.g., smoking). In a recent study, age-specific familial risks in offspring were increased at 24 of 25 sites for concordant cancer in only the parent and in 20 of 21 sites for sibling probands (Hemminki, 2004). Similar findings were reported in the Utah Cancer Registry involving a Mormon population (Cannon-Albright et al., 1994; Goldgar et al., 1994). These studies establish that virtually all tumors exhibit some degree of excess risk in relatives. The increased risk of cancer in relatives tended to be restricted to cancers of the same site, suggesting that modifier genes may have site-specific effects. Twin studies also support a hereditary contribution to a variety of cancers. Comparisons of concordance between monozygotic (MZ) and dizygotic (DZ) twin pairs are used to derive a heritability estimate, suggesting whether an observed pattern is due to hereditary and/or environmental influences. This estimate is subject to a variety of assumptions, for example, that the shared environment is the same for MZ twins and DZ twins. Generally, no gene–environment interaction is modeled. For cancer, the precise proportion of variance explained by heredity is uncertain (Braun et al., 1995; Ahlbom et al., 1997; Lichtenstein et al., 2000; Risch, 2001). However, a recent populationbased twin study with complete incidence data on 90,000 twins in Scandinavia revealed higher concordance rates in MZ pairs than in DZ pairs, and estimates of the proportion of susceptibility to cancer due to heritable effects ranged from 26% to 42% for cancer at five common sites (Hoover, 2000; Lichtenstein et al., 2000). Although the proportion of cancer attributable to genetic influences is unclear, surveys of the literature (Vineis, 1999, 2002; Garte and Taioli, 2000) focusing on low-penetrance genotypes generally estimate odds ratios for single genes that are less than 2 and suggest attributable risks for specific genotypes that range from 0% to 30%. However, the available data are too limited for drawing stable conclusions, but suggest (in agreement with twin studies) that attributable risks due to genetic susceptibility may be substantial but lower than those due to lifestyle and other environmental factors. It is likely, however, that the contributions of as yet unrecognized genes, as well as gene–environment and gene–gene interactions, make this range conservative (Caporaso, 2002). An early rationale for gene-cancer associations was based on the appreciation that carcinogens require metabolic activation to form active DNA-binding species. This drove the first pharmacogenetic investigations into genes that modify cancer risk in the population. The four candidate genes studied were GSTM1 (Seidegard et al., 1986), CYP2D6 (Ayesh et al., 1984), NAT2 (Lower et al., 1979), and CYP1A1 (Kellermann et al., 1973). Metabolic phenotyping (administration of a probe drug and measurement of metabolites or related approach) was used to infer the genetic variant a subject possessed in cancer cases and controls. The studies were relatively small and subject to various forms of bias (i.e., cancer in the host might alter metabolism irrespective of the underlying gene) as well as deficiencies in epidemiologic design (i.e., inadequate power, poor control selection, failure to collect appropriate exposure data). However, the studies had the advantage that the “metabolic phenotype” reflected a physiological integration of gene action on a target substrate and thus represented a biologically realistic assessment (Caporaso et al., 1991). By the late 1980s this approach was supplanted by direct genotype determinations, and in the late 1990s high-throughput genotyping made possible simultaneous evaluation of a few and then many candidate genes. The technological capacity currently exists to study

Genetic Modifiers of Cancer Risk substantial numbers of variants, although this development has been accompanied by formidable methodologic challenges (Table 29–2).

METHODOLOGICAL ISSUES Table 29–2 highlights methodological issues of particular importance in gene-based population studies, including confounding, power, misclassification, gene–environment misspecification, multiple comparisons, and the appropriate inclusion of biomarkers into study design. The validity of genotyping and systematic approaches to data evaluation are also key concerns (Rothman et al., 1995; Little et al., 2002; Haddow and Palomaki, 2004). All epidemiologic study designs used to evaluate effects of genes must take into account various forms of potential bias (selection, recall, misclassification and confounding) (Khoury et al., 1993; Rothman and Greenland, 1998; Garcia-Closas et al., 2004). Also fundamental are provisions for an appropriate analytic approach and proper control selection. Population-based controls are the gold standard, but high response rates are extremely difficult to achieve in studies that include biospecimens. Hospital controls can be more convenient and feasible but entail more opportunities for bias (Wacholder, 1992). Population association studies are commonly used for low-penetrance genes because of the requirement for substantial numbers of subjects (statistical power) and the need to take environmental exposures into account. Family-based approaches using linkage analysis have been successful in identifying the genes that account for many Mendelian traits including familial cancer syndromes but require DNA from well-characterized members of high-risk kindreds, making this strategy unsuitable for the detection of low-penetrance genes (Risch, 1996). Characteristics of the various population design options, including case-control (Caporaso et al., 1999), cohort (Hunter, 1997; Langholz et al., 1999) and case-case (case-only) approaches (Khoury and Flanders, 1996; Yang et al., 1997) have been described. Specific issues pertinent to gene studies in the population are summarized in Table 29–3. In particular, virtually all genetic polymorphisms exhibit ethnic, racial, and geographic variation. If both the gene and the disease vary by ethnicity, the conditions are met for uncontrolled confounding (population stratification). Population stratification may be dealt with by usual epidemiologic approaches to control confounding (Wacholder et al., 2000, 2004). Genomic control or the use of panels of genetic markers that establish ethnicity may have a role in selected settings (i.e., studies involving mixed ethnic groups that cannot be reliably distinguished by self-report) (Hoggart et al., 2003; Freedman et al., 2004; Marchini et al., 2004). A methodologic defect that has received much emphasis in critical reviews is inadequate study size. Only a handful of cancer–gene associations have achieved some degree of verification (Hirschorn, 2002); it is clear in retrospect that most of the hundreds of studies conducted during the genotyping era have been underpowered, if the most commonly reported effects from metanalyses are realistic. The situation with regard to power is actually worse, considering the multiple comparison issues, the need to take gene–environment and gene–gene effects into account, and the expected low prior probabilities of an increasing number of gene variants that will be tested. In general, study sizes in the thousands are required to address these issues adequately.

GENE SELECTION IN POPULATION STUDIES Given the thousands of candidate genes and the appreciation that prioritizing candidates offers statistical advantages, there is clearly a need to identify efficient approaches to gene and mutation/polymorphism selection when conducting population studies. Considerations that frame studies of genetic markers include (1) the number of genes to be studied, (2) the method of gene selection, and (3) the number and type of polymorphic variants or mutations to be studied. Regarding the number of genes to be studied, at one extreme is testing a single sequence variant of one gene. This approach has the

579

advantage of providing a straightforward framework for hypothesis testing. The early pharmacogenetic studies of cancer employed this approach using a “metabolic phenotype” (i.e., metabolism of a drug probe dependent upon the gene of interest), but given the resources required to design, field, and assemble epidemiologic studies including DNA collection, the testing of single gene variants in population settings is wasteful. At the other extreme is a whole genome search, in which anonymous markers (e.g., microsatellite markers spaced at every 5–10 cMs or single nucleotide polymorphisms [SNPs] at much closer intervals) are selected to cover the entire genome (Uhl, 2001). This approach has become increasingly more feasible with the availability of the sequence of the human genome, improved SNP databases, and highthroughput genotyping platforms (Taylor, 2001). Currently available chips contain 500,000 SNPs and those expected within a few years will contain millions. There are two distinct modes of whole genome study. Linkage analysis (evaluating the cosegregation of genetic markers and the pattern of disease transmission in order to identify chromosome regions likely to harbor disease-related genes) is appropriate when multiple-case families suggesting a Mendelian pattern of inheritance are available or when the familial risk is substantial (Risch and Merikangas, 1996). This approach has led to the successful positional cloning of numerous genes involved in hereditary cancer syndromes. However, for evaluating low-penetrance genes for common malignancies, linkage disequilibrium or whole genome mapping approaches are looming as an option (Bonnen, 2002; Weiss, 2002). There is widespread expectation that improvements in the quality and density of the emerging genetic and physical map, relatively inexpensive and reliable high-throughput SNP genotyping, improved data storage, and computational approaches will enable linkage disequilibrium mapping to discover candidate genes in population (association) studies. The population study designs will involve the testing of thousands to millions of SNPs in whole genome scans to identify loci that are in strong linkage disequilibrium (LD) with putative susceptibility genes. Pursuing this approach in complex diseases poses difficulties including the assessment of environmental risk factors, as well as informatics and technological challenges. Technologies to pool DNA that will reduce the number of genotypes required by a factor of 100 or more may be helpful (Barcellos, 1997; Hellard, 2002), although haplotyping and exposure assessment will be sacrificed. There is also uncertainty concerning proper sample size, extent of linkage disequilibrium likely to be present, number of markers required (Hellard, 2002; Gibson and Muse, 2002; Pfeiffer, 2002), potential advantages of conducting such studies in population isolates, such as Iceland (Peltonen, 2000), and the benefits of using haplotypes (Weiss and Clark, 2002). The use of LD-based methods are likely to work best if there is a single susceptibility allele, while substantial allelic heterogeneity poses difficulties (Pritchard, 2002). The current international HapMap project is designed to characterize “haplotype blocks” permitting more efficient SNP selection capturing an estimated 75% of common SNP variants (Kruglyak, 2005), thereby improving the chances to detect associations (Cardon and Abecasis, 2003). The low prior probability of the thousands of SNPs tested means a stringent p-value will be required to reduce the chance of false positives and follow-up of promising candidates in other studies will be required (Wacholder, 2004). To date, only a few such studies have been conducted involving cancer but the results are promising. A study of 254 breast cancer cases and >25,000 SNPs identified a 20 kb region (19p13.2) including the intercellular adhesion molecules (ICAM 1,4,5) with replication in two independent collections (Kammerer, 2004), although a more recent report was null (Cox, 2006). The majority of current studies, however, fall between these extremes. Most investigate a few to thousands of SNPs or other markers from a selected group of candidate genes, gene families, or regions of interest. Special strategies used to select genes for study include prior functional data, targeted gene families or pathways, or specific disease associations (Table 29–3). How does an investigator decide among the available approaches for gene selection? This complex challenge grows as facile genotyping expands the available choices for genomic information. Among the key considerations are

Table 29–2. Overview of Methodological Issues Topic

Selected Points and Problems

Candidate gene selection (see Table 29–3)

Selection of specific candidates includes a spectrum of options from single genes (hypothesis driven) to whole-genome searches (anonymous). Many technologies can nominate candidate genes including expression profiling, somatic mutations, loss of heterozygosity studies, mouse or other models, etc.

Ethnicity, race and population stratification

Population stratification (PS) is addressed by taking ethnicity into account in study design and control of confounding by restriction, matching, or adjustment. Genomic control involves using genetic information to classify individuals (Hoggart, 2003; Pritchard, 2002). Ethnic, racial and geographic variation (see review by Risch et al., 2003). Population isolates offer certain advantages for genetic studies (Peletonen, 2000). Inadequate study size (power) Multiple comparisons False positives due to low prior probabilities (Wacholder, 2004) Selection of risk models (Khoury, 1988)

Statistical issues

General failure to replicate

Disease effect(s)

General challenges: Phenotype definition Involvement of diverse biological pathways Polygenic mechanisms (implying small contributions of individual genes) First study and publication bias Multiple comparison issues and high false positive rate due to low priors (related to inadequate study size) Failure to study enough genetic variants within candidates (Neale and Sham, 2004) Prospective studies; Mendelian “randomization” Prevalent/incident cases

“Phenotype definition”

Variation within disease subgroups by histology or other disease characteristics; genetic heterogeneity

Bioinformatics

Tracking and using the vast amount of data generated by new technologies in the context of population studies will be a key challenge. Gene–environment interaction in the context of cancer requires larger study sizes. See references for specific methods issues. How do genes act in concert to influence common cancers? Epistasis (Moore, 2003).

Gene–environment interaction Gene–gene interaction

Quality control

Quality issues related to general epidemiological practice are treated elsewhere. Issues related to biospecimen and genotype determination.

Comment The literature provides examples of mechanistically plausible genes based on metabolism of putative carcinogens or key roles in carcinogenic mechanisms such as DNA repair. With the increasing availability of high throughput genotyping, proposals to conduct “genome searches” on population samples are increasingly feasible, although cost, DNA pooling, use of haplotypes, variability of LD structure, informatics limitations, and other problems remain challenges. Prioritizing polymorphisms based on functional changes is advocated (Pfeiffer, 2002; Tabor, 2002; Rebbeck, 2004; Zhu, 2004). Qualitative and quantitative differences in genes across racial/geographic groups exist and must be taken into account in design. PS’ is not a major cause of bias and can be dealt with using standard epidemiological approaches (Wacholder, 2000, 2002) although some argue “Ethnic” labels are inadequate to represent geographic variation in population genetic structure (Wilson, 2001). Genomic control, i.e., using a panel of highly polymorphic markers to assign ethnicity, may play a role in control of population stratification, but selfreported ancestry is less intrusive and correlates better with unknown environmental risks; genetic ancestry may have advantages under certain circumstances, i.e. recently admixed populations (Rosenberg, 2002). Studies have often been too small to achieve sufficient statistical power to identify associations with genes that have modest effects. Subgroup analysis is severely limited in small studies. Interaction (gene–environment, gene–gene, or higher order interactions) requires large study size (Foppa, 1997; Garcia-Closas, 1999). Testing numerous genes with low a priori expectations of effects will inevitably result in false positive associations unless alpha values or priors are adjusted (Bonferroni type or other) for the multiple tests conducted. Calculation of the False Probability Report Probability allows different stringencies for more likely genes than for less likely ones; no penalty for multiple testing; possible trade-offs of power against false positives (Wacholder, 2004). Claiming “subgroup” effects in the absence of overall findings is a very common flaw; it can be correct but probably not often. Metanalyses generally have demonstrated low confirmation rates for candidate genes studies in human disease (not just cancer; see Ioannidis, 2001; Hirschhorn, 2002; Lohmueller, 2003). For overview of sources of bias and implications, see Wacholder (2002, 2004), Tabor (2002), Little (2002), Cardon and Bell (2001), Botstein (2003), Rebbeck (2004). See Dragani (1996) and Comings (1998) for views on polygenic mechanism. See Chanock (2002) for implications of pleiotropic effects of genes.

Disease does not influence germline polymorphisms and therefore the case-control design is not problematic to test polymorphic genes in relation to disease, however, disease in the host or its treatment may alter assays of gene function, exposures, and metabolism, investigating those assays may require a prospective design. Including prevalent cases will introduce bias if for example, genes under investigation are directly or indirectly related to survival; see Shu (2004) or Yokomizo (2002). Cancer has traditionally posed less of a problem with phenotype definition than “behavioral phenotypes” such as alcoholism. Nevertheless, molecular tools are increasingly demonstrating molecular heterogeneity within cancer that will likely have genetic correlates, e.g., by prognosis (CLL and heavy-chain mutations) and lung cancer (adenocarcinoma has distinct expression and possible gene associations) (Yanagitani, 2003). Subgrouping disease may increase power in presence of etiologic heterogeneity. Enhanced data processing, management and statistical approaches will be required (Elkin, 2003). The web provides vast new resources for investigating the genetics of complex disease (see Table 29–4) (Pevsner, 2003). Various models have been described (Ottman, 1990); study size (Foppa, 1997; Garcia-Closas, 1998, 1999), misclassification (Garcia-Closas, 1998), case-only designs (Khoury, 1996). Study size issues are nominally similar to those required for detecting gene– environment interaction but there are important biological, molecular, and mechanistic differences. Modeling new approaches to combining genes in and across pathways will be a future direction, i.e. hierarchical (Hung, 2004), oligogenic (Sellers, 2005) models. Include blinded duplicate and replicate samples in runs. Include QC (standard operating procedures) at every stage of DNA processing from: collection, labeling, shipping, processing, DNA extraction, aliquoting, storage (freezer monitoring), DNA integrity, amount and assay testing, and shipping. Bias results if reasons for suboptimum biospecimen collection or assay variability are differential. Increasing capacity to obtain DNA from noninvasive sources and whole genome amplification.

Table 29–3. Selection of Candidate Genes Strategy to Select Genes Based on functional variants from available SNPs

Based on areas identified in linkage studies

Based on cytogenetic evidence

Based on animal models

Based on somatic mutations observed in tumors, LOH, CGH, etc.

Based on role in a related phenotype, disease or exposure Based on expression studies

Based on protein studies

Based on known etiologic exposures

Advantages (+) and Challenges (-) of the Particular Approach + Certain codon sequences or SNP typologies more likely alter function (Risch, 2000; Tabor, 2002; Zhu, 2004), i.e., cSNPS that alter the amino acid sequence of the encoded protein (Cargill, 1999). High-throughput screening may be more efficient (cost-effective) than selecting best candidates from known variants. Information from nonfunctional variants may contribute to understanding haplotypes or be in LD with important variants. + Regions identified may include major genes. Finer mapping may proceed using conserved haplotypes, linkage disequilibrium, or based on chromosome translocations, deletions, etc. LOH, CGH or multiple strategies (Gao, 2000; Dahia, 2005). Variants of major genes likely to play a role in nonfamilial case (see examples Table 29–3). Linkage signals are too weak to detect odds ratios in the range most common for cancer modifier genes; the environment is not taken into account. Generally there is a solid mechanistic basis for studying genes already implicated in families. + Successful identification of genes important in leukemias and some solid tumors. Cytogenetic regions implicated, e.g., 3p21 in lung (Protopopov, 2003; Wei, 1996), NPC (Xiong, 2004), and renal cancers (Sukosd, 2003) and in the origin or development of multiple solid tumors likely harbor tumor suppressor genes (Protopopov, 2003). May be less relevant to modifier genes with low penetrance. Cytogenetic abnormalities are more numerous and variable in solid tumors. Chromosome alterations in lymphocytes may independently predict cancer susceptibility (Bonassi, 2000) but have little localizing information. + Increasing number of model systems and ability to understand homologies suggests new pathways and specific candidates. Easier to demonstrate organ specific effects and obtain tissue. Controlled matings can elucidate genetics. Animal models (knockout and transgenic) may implicate susceptibility and resistance genes (Gonzalez, 1999; Kimura, 2000; Wang, 2005) and describe gene–gene interactions (Tripodis, 2001; Samuelson, 2005). For example, 60 loci contribute to lung cancer susceptibility in the mouse genome, including many with interactive effects (Dragani, 2000; Manenti, 2000). Can study both temporal and spatial events. Improved databases exist, i.e., http://tumor.informatics.jax.org (Naf, 2002). Animals don’t duplicate the complex exposures that underlie human cancer and are distinct from humans in clinical, histologic, and other aspects. Animal findings must be verified in human population studies. + “Two hit” model established for at least a few high-penetrance genes. Patterns of mutations may suggest gene pathways (Kimura, 2003) or specific genes. Certain molecular finding in tumors suggest oncogenes or TSGs. Somatic mutations may be more related to prognosis, disease histology, or clinical characteristics then etiologic factors (Coe, 2006). See Web sites that catalog somatic mutations (Bamford, 2004). + Genes involved in, e.g., one tobacco- or alcohol-related cancer are candidates for investigation in another. Genes involved in other conditions related to a particular exposure (i.e., insulin resistance may be related to abdominal obesity, heart disease, and prostate cancer) may be worth investigating in all those conditions. Genes involved in emphysema may mediate lung cancer risk as well (Minematsu, 2005) + Expression data provide a rationale for gene’s role in tissue of interest and suggest tumor mechanisms, candidate genes, and families (Crawford, 2000; Williams, 2003; Amatschek, 2004). Pathways may be suggested for tumors where current understanding of etiology and therapy is inadequate (Godard, 2003). Tumor differences by anatomic location, i.e., left and right colon may be elucidated (Glebov, 2003). Expression profiles may suggest or even define tumor subtypes and suggest critical discriminators (Tay, 2003). Organ specific genes, mechanisms, and pathways important for specific tumors are indicated (Sanchez-Carbayo, 2002). Provide insight into molecular pharmacology of new and established therapies (Clarke, 2003). Numerous factors unrelated to disease may influence expression in tissue. Obtaining a well-characterized sample of tumors from a generalizable sample of subjects with a cancer is challenging. There is enormous variability in expression; studies to date have been modest in size. Statistical problems include “dimensionality,” overfitting, poor power, and informatics limitations. Patient specific transcription profiles differ due to genetic instability inherent to tumors as well as technical factors. + Protein is “effector arm” of genome and therefore must be involved in action of altered genes involved in carcinogenesis. Current techniques do not always allow facile identification of protein/peptides involved in cancer to their genetic components. Proteomics is less mature technology. Validation and reproducibility are challenges. + Early mechanistic “pharmacogenetics” approach has historical precedent; processing compounds known to influence cancer risk has biological plausibility (Zhang, 2003). Epidemiologic (and clinical) observations suggest important molecular mechanism for investigation. Many categories of genes known to be involved in carcinogenesis do not fall in these categories and new genes may implicate novel mechanisms.

(continued)

581

582

PART III: THE CAUSES OF CANCER

Table 29–3. (cont.) Strategy to Select Genes Based on previous population studies

Based on evolutionarily conserved regions

Based on regions determined from whole genome searches/haplotype linkage disequilibrium

Based on epigenetic changes

Advantages (+) and Challenges (-) of the Particular Approach + Candidates with known gene frequencies, established prior hypotheses. Necessary to assess candidates in the population to verify their role, gauge public health impact, further explore subgroup and covariate relations. Such an approach will miss new or understudied genes or variants. Previous studies may have been flawed. + Conserved regions likely to harbor areas that involve critical functions that when altered perturb important biological processes. May miss some genes, for example, those involved in gene–environment effects, those that uniquely involve humans, recent mutations, etc. Selection may not conserve regions important to common cancers that affect individuals well beyond reproductive age. + Identify previously unknown genes and families. Use haplotype blocks to guide selection of polymorphisms (Bonnen, 2002; Cardon, 2003), identify blocks of genes in linkage disequilibrium. Fine-gene localization (Risch and Merikangas, 1996; Jorde, 2000). Locus heterogeneity complicates complex disease analysis (Weiss, 2002). Power may be limiting. Markers studied may not correspond to functional variants that account for associations. Frequency mismatch of marker gene and functional genes will reduce power. Variation of LD across populations, genome regions and between markers in close proximity requires further study (Shifman, 2003). Role of haplotypes in SNP selection still controversial (Belmont and Gibbs, 2004). + May account for plausible candidate genes that fail to show mutations or functional polymorphisms. Epigenetic silencing of tumor suppressor genes by promoter hypermethylation operates in some cancers (Baylin, 2000; Chang, 2003). Methylation profiles vary according to previous exposure (Anttila, 2003), histology, geographic origin (Toyooka, 2003), as well as germ line mutations (Paz, 2002).

the size and design of the study (case-control, cohort, sib pair), whether the study is primarily designed to test a specific hypothesis or generate new hypotheses (i.e., gene discovery or characterization), the degree to which tissue studies will be integrated, genotyping approach, laboratory resources, and the available budget. Generally, an investigator will amass evidence favoring candidate genes or genes families from the categories indicated in Table 29–2 and then rank them. Selection should balance the desire to include diverse categories of modifier genes and the opportunity to comprehensively cover many or all the genes in key pathways. The investigator must also decide how many polymorphisms/SNPs per gene should be investigated, the type of polymorphism (i.e., SNPs, microsatellite, functional or not, etc.), and whether selection will be such that haplotypes can be reconstructed. How many variants to investigate within a candidate gene is a controversial point. Neale and Sham (2004) argue for extensive study of risk-conferring variations within candidate genes as a gold standard, although this would not be necessary if functional variants were well understood and would be impractical for large numbers of genes. Within a particular haplotype block, SNP selection can be limited, as flanking SNPs will be in linkage disequilibrium and additional SNPs will not provide much further information. Common SNPs (>5–20%) are often favored because of power considerations. Given a limit on how many genes to test, the genes with the highest prior probabilities of disease are favored, as candidates with low priors will have high false positive rates. For presumptively positive findings involving randomly selected SNPs (given that there are potentially millions of SNPs that can be tested but the number of true associations are likely to be far fewer for any given cancer), replication will be mandatory. Underpowered studies will have reduced probability of detecting a true association between a gene and disease and will report a higher proportion of false positive findings than larger studies with adequate study power (Garcia-Closas et al., 2004; Wacholder et al., 2004). Evidence that can contribute to candidate gene selection can be loosely divided into mechanistic considerations and findings from population studies. Table 29–4 provides a semiquantitative summary of the mechanistic and population evidence for several candidate gene

families. In the table, both categories are given an estimated score (i.e., 0 = no data or null findings, 1 = some suggestive evidence that is short of confirmatory, 2 = strong evidence). The summed scores for the two categories could serve as one guide to setting prior expectations. For example, a summed score of 0–1 indicates a low prior expectation (i.e., the probability of a true association is quite low, e.g., p £ 0.001). A score of 2–3 indicates a medium prior of 0.01. A high score of 4 indicates the highest prior (i.e., p = 0.10) (Wacholder, 2004). Assessing the mechanistic component involves considering the action of the gene in relation to physiological, cellular, or biochemical functions potentially relevant to cancer. This can include metabolism/activation/inactivation of exogenous or endogenous factors related to cancer, whether the gene action is relevant to cancer in animals or model systems, and whether the gene is active in the tissue of interest (consistent with direct carcinogenic action). Also related to mechanism is the specific effect of the genetic mutation or polymorphism. Coding for a nonsynonymous mutation, a mutation in an evolutionarily conserved region, a mutation in an exon or regulatory region, or a mutation known to influence function increases relevance to cancer and should be given priority (Tabor, 2002; Zhu, 2004). Actual SNP selection can proceed using increasingly comprehensive databases that are searchable by gene, chromosome, and pathway, e.g., http://snp500cancer.nci.nih.gov). See Table 29–4 for other relevant web based resources. The population component of the evidence derives from epidemiologic studies. Genes that have exhibited associations with cancer, associated conditions, or precursors in highquality studies will be more attractive for investigation. The case to investigate unselected “random” SNPs is based on the availability of improving technologies that lower the cost of SNP testing, the time and difficulty in establishing functional evidence, certain exceptions to the usual rules (e.g., synonymous mutations found to be important in selected genes) (Duan et al., 2003), the possibility that associations may reflect linkage disequilibrium with the real functional markers (Chapman et al., 2003), and the certainty that many important genes remain undiscovered. It is possible that as yet unrecognized genes acting via novel mechanisms will require a systematic

Genetic Modifiers of Cancer Risk

583

Table 29–4. Web Sites Useful in Susceptibility Gene Investigations Locus Link http://www.ncbi.nlm.nih.gov/LocusLink Descriptive information about genetic loci Nomenclature http://www.gene.ucl.ac.uk/nomenclature/ Human Gene Nomenclature Committee (HUGO) assigns official names to genes and proteins Human Genome Informatics

Informatics related to the genome http://www-l.lanl.gov/HGhotlist.htlm

National Center for Biotechnology Information (NCBI) Comprehensive genetics information http://www.ncbi.nlm.nih.gov Online Mendelian Inheritance in Man (OMIM) PubMed http://www.ncbi.nlm.nih.gov/PubMed/ National Library of Medicine including Medline GeneCards SNP-related dbSNP SNP Consortium HGBASE HapMap Genetic Association Database

http://bioinfo.weizmann.ac.il/cards

Database of human genes, their products, and involvement in human disease

www.ncbi.nlm.nih.gov/SNP http://snp.cshl.org http://hgbase.cgr.ki.se/ http://www.hapmap.org/ http://geneticassociationdb.

Archive of human genetic association studies

A comprehensive review of bioinformatics and material on web sites covering sequencing, DNA, RNA, and protein, model organisms, genome analysis and the Human Genome Project and web resources related to the study of human disease, see Bioinformatics and Functional Genomics, Jonathan Pevsner, Wiley-Liss, 2003.

search in currently unexplored areas of the genome, through the use of random markers in linkage disequilibrium studies.

OVERVIEW OF CANDIDATE GENES Table 29–5 lists the candidate cancer susceptibility genes, as well as representative examples of genes and key references. Historically important reports, meta-analyses (boldface type in the table), and reviews are selected where available. The first candidate pharmacogenetic studies investigated genes thought to activate (i.e., Phase 1, genes such as CYP1A1 and CYP2D6) (Kellerman, 1973; Ayesh, 1984) or eliminate carcinogens (i.e., Phase 2 genes, such as NAT2 and GSTM1) (Lower, 1979; Siedegard, 1986) in relation to tobacco-related cancers. These Phase 1 and Phase 2 genes have been the most extensively studied, based on mechanistically plausible pharmacogenetic relationships of key exposures with known phenotypes. Mechanistic studies have documented the relation of these genes to exposure, intermediate markers, subtypes of disease, and prognosis (Bartsch, 2000). The best established Phase 1 association is CYP1A1 and lung cancer. There is also good evidence that two Phase 2 genes are associated with bladder cancer (NAT2, the slow acetylator phenotype or deficient genotype, and the GSTM1 null genotype). The evidence for other Phase 1 and 2 genes is inconclusive, although based on the associations observed and the known mechanisms involved, it would not be surprising if weak to modest effects are eventually established for combinations of genes from both of these classes. The gene families categories in Tables 29–5 and 29–6 are somewhat arbitrary, as genes commonly influence multiple processes (pleiotropy). For example, MTHFR may be classified in the nutrient category based on its involvement in folate metabolism, but it also influences oxidative metabolism, DNA repair, and chromosome integrity. Certain genes may both activate and detoxify (pro)carcinogens (i.e., have both Phase 1 and Phase 2 activity, such as mEH and SULT1A1). Gene polymorphisms involving hormonal pathways are another well-studied group. Despite plausible mechanisms and some suggestive findings (Medeiros, 2003) for the role of androgen-pathway genes in prostate cancer and estrogen-related genes in breast and endometrial cancer (Zheng, 2004), consistent evidence for associations is not yet evident.

DNA repair genes are a major focus of recent study with suggestive findings across a broad spectrum of malignancies. The variety of mechanisms of DNA repair and many mutations known to influence these complex processes suggest the need for further comprehensive study and emphasize the need to prioritize their study based on biological function and relevance to cancer (Berwick and Vineis, 2000). Immunologic and inflammatory mechanisms appear to be affected by susceptibility as indicated by the association of HLA types with selected diseases. The oldest known association (ABO group and gastric cancer; Aird, 1953) falls in this category, and a variety of processes involving cytokine, chemokine, and immunoglobulin gene processes are under active investigation and will likely be established to have importance in cancer susceptibility. Recently reported are associations linking aggressive NHL to TNF-alpha, and gastric cancer to proinflammatory cytokine polymorphisms (El-Omar et al., 2003). Because various oncogenes and tumor suppressor genes determine susceptibility to specific cancers in the setting of high-risk kindreds, polymorphic variants of these genes that influence but do not ablate function deserve further study as candidates that modify susceptibility. For example, the specific founder mutation in the NBS1 gene responsible for Nijmegan breakage syndrome is present in 9% of familial prostate cancer in a Polish population but also in 2.2% of nonfamilial prostate cancer (compared to 0.6% of controls) (Cybulski, 2004). Various polymorphisms of tumor suppressor genes and mismatch repair genes implicated in familial cancer are under study in the sporadic forms of the disease. Several other molecular mechanisms known to influence later stages of cancer including apoptosis, angiogenesis, cell-cycle control, telomere integrity, and many others have been investigated but only in a fragmentary manner. Dietary nutrients and energy balance including obesity all alter the risk of cancer and are under various degrees of genetic control. Only a handful of genes have been well studied, such as MTHFR. Because these genes act in pathways and in concert with environmental, metabolic, and genetic factors (Ulrich, 1999), large-scale studies will be needed to sort out their effects. Behavioral genes are a new category of genes. Although their primary role may be in influencing exposures such as alcohol, tobacco, or food intake that are major determinants of cancer in the population,

Table 29–5. Candidate Cancer Susceptibility Genes Gene Family

Gene Examples

Relevant Cancers

Mech

Pop

G-E or G-G

2

1

GE = alcohol GG = ALDH2

2

1

GE = alcohol GG-ADH3 GE = alcohol, HCV GG-ADH2

Alcohol

ADH3

Alcohol

ALDH2

Oral, Head and Neck, Esophagus, Colorectal Oral, Esophagus

Alcohol

CYP2E1

Oral

1

1

Angiogenesis Angiogenesis Angiogeneiss Angiogenesis (?) Apoptosis

Endostatin ecNOS ecNOS ACE FAS; BCL6

1 1 1 1 1

1 1 1 1 0

Behavior

OPRM1, many others

1

0

GE = alcohol, tobacco, obesity

Behavior Blood type Cell cycle

COMT ABO CHEK2

1 1 1

0 2 1

GE = same

Cell cycle

CCND1

1

1

Cell cycle Chemokine receptors Circadian Cytokines DNA repair

TGFB CCR2 and CCR5 PER1 IL-10 XRCC1, XRCC3 XPC, XPD, RAD51, OGG1, MGMT, and others.

Prostate Prostate, Lung Lung Breast Bladder AML Lymphoma All tobacco, alcohol, drug, obesity related cancers Same Gastric Breast Colorectal Prostate Lung Colorectal Breast NHL Breast NHL Esophagus, Lung, Colorectal, Prostate, Melanoma, Glioma, Bladder, Breast, Gastric, Pancreas, NPC, HCC, Skin

1 1 1 1 2

1 1 1 1 1

GE = HIV infection GG = other DNA repair genes; With BRCA1/2 (breast) GE

Drug transport

MDR1

Leukemia, Lung

1

1

Environmental Contaminant

NQO1

Bladder Esophagus Breast

1

1

Environmental contaminant Growth Factor

PON1

?

1

0

IGFBP3, IGF1

Breast

2

1

Hormone

Estrogen-related: CYP11A, COMT, CYP17, CYP19, CYP1B1, EDH17B2 ESR1 PR

Breast, (Prostate, Ovary, Endometrial, Liver)

2

1

Hormone

Other steroid hormones: i.e., progesterone oxytocin, prolactin

Breast, (Prostate, Endometrial), Liver

1

1

584

GE-benzene

GE = HCV (liver)

Reference Monzoni, 2001 (mechanism); Tiemersma, 2003 (colorectal adenomas) Brennan, 2003 (oropharyngeal) Yokoyama, 2003a, 2003b (Oro Yang, 2005 (esophagus)) Kato, 2003 Inghetti, 2001 Medeiros, 2002, 2003b (prostate) Cheon, 2000 (lung) Koh, 2003 Hazra, 2003 (bladder) Sibley, 2003 (AML) Lossos, 2001, Zhang, 2005 (lymphoma) McGue and Bouchard, 1998; Han, 1999, Caspi, 2003; Uhl, 2002; Crowley, 2003 (mechanism) Hariri, 2003 Aird, 1953; Boren, 1993 Zhang, 2003 Kilpivaara, 2003 (colorectal) Cybulski, 2004a (prostate) Qiuling, 2003 (lung) Lewis, 2003 (colorectal adenomas) Shu, 2004 Smith, 1997 Zhu, 2005 Cunningham, 2003 Friedberg, 2000, Savas, 2004 Mohrenweiser, 2003 (mechanistic) Butkiewicz, 2001; David-Beabes, 2001 Wang, 2003 (lung), Kirk, 2005 (HCC) Han, 2004 (breast) Stern, 2002; Kelsey, 2004 (bladder) Heather, 2002 (skin, XRCC1) Duell, 2002 (pancreas) Lee, 2002 (gastric), Han, 2005 (skin) Elahi, 2002 (oropharyngeal) Cho, 2003 (NPC) Rybicki, 2004 (prostate) Tomescu, 2001; Baccarelli, 2004 (melanoma) Goode, 2002 (review) Sinues, 2003 (lung) Jamroziak, 2004 (childhood ALL) Bauer, 2003; Moran, 1999 (benzene toxicity) Park, 2003 (mechanism) Zhang, 2003b (esophagus) Menzel, 2004 (breast) Battuello, 2004 Kelada, 2003 (general review) Ren, 2004 (breast) Fletcher, 2005 Mitrunen, 2003; Jefcoate, 2000 (mechanism) Cai, 2003 (breast); Zheng, 2004 (breast); Mitrunen, 2003 (breast) Thompson, 1998 (COMT) Tworoger, 2004 (genes and hormones), Gold, 2004 (ESR1) Hachey, 2003 (CYP1B1) Madigan, 2003 (prostate, CYP17); Berstein, 2002 (endometrial, CYP17) Zheng, 2004 (breast, CYP11A) Rossi, 2003 (HCC); Goodman, 2001 (ovary) Ntiais, 2003c (CYP17 and prostate) Maloney, 2003 (general), Zheng, 2004 (CYP11A and breast) DeVivo, 2003 (progesterone) Vondehaar, 1999 Rossi, 2003 (liver)

Table 29–5. (cont.) Gene Family

Gene Examples

Relevant Cancers

Mech

Pop

G-E or G-G

Reference

GE = mammographic density, estrogen therapy

Dietzsch, 2003 (esophagus). Liede, 2003; Lillie, 2004; Suter, 2003 (breast) van Gils, 2003 (SRD5A2, breast) Li, 2003; Soderstrom, 2002 (SRD5A2, prostate) Mononen, 2000 (prostate) Chang, 2003c; Medeiros, 2003a (prostate, ER and others) Liu, 2003 de Jong, 2003 (breast), Butsch Kovacic, 2005 (NPC) Snoek, 2000 (lung) de Jong, 2003 (breast) Lu, 2003 (NPC); Hirunsatit, 2000 (NPC) Howell, 2002 (lymphoma) Thye 2003 DuBois, 2003; Thun, 2002 (mechanism) Landi, 2003 (colorectal) Campa, 2004 (lung) Xu, 2005 Martinez, 2003 Thye, 2003 (mechanism) El-Omar, 2002, 2003; Taguchi, 2005 (gastric) Zienolddiny, 2004 (lung) Tsukasaki, 2001 (leukemia) Kirkpatrick, 2004 (cervix) Watson, 2003 Egeblad, 2002; Ghilardi, 2002 (breast) Yu, 2002; Su, 2005; Fang, 2005 (lung) Miao, 2003 (gastric) Sommer, 2003; Bretsky, 2003; Angele, 2003; Thompson, 2005; Lee, 2005 Meitz, 2002 Dunning, 1997 Crabtree, 2004 Hahnloser, 2003 Miller, 2003 Robledo, 2003 Kibel, 2003 Zhao, 2004 (mechanism) Van Eerdewegh, 2002 (asthma) Zhu, 2001 (lung) Egeblad and Werb, 2002 (cancer progression) Kallianpur, 2004 Sinues, 2003 Ma, 1998 (prostate); Chokkalingam, 2001 Ntais, 2003b (prostate); John, 2005 (prostate) Curtin, 2004; Ulrich, 1999 Goode, 2004, Ma, 1997 (colorectal) Giovannucci, 2003 (colorectal adenomas, alcohol); Cicek, 2004 (prostate) Jacques, 1996; Spotilla, 2003 (homocysteine) Bailey, 2003; Heijmans 2003 Gao, 2002 (gastric) Hung, 2004 (bladder) Lin, 2004 (bladder) Ergul, 2003; Chen 2005; Shrubsole, 2004 (lowfolate) (breast) Krajinovic, 2004 (ALL) Infante-Rivard, 2003 (mechanism leukemia) Esteller, 1997 (endometrium) Spotila, 2003 Robien 2003 (leukemia) Heinonen, 1998 Calle, 2003; Thompson, 2004 (mechanism)

Hormones

Androgen receptor, SRD5A2

Prostate, Breast, Esophagus (squamous)

2

1

Hormones Immune

Pepsinogen C HLA region genes

Gastric Lung, Breast, NPC, Lymphoma

1 2

1 1

Immune Inflammation

IFNGR1 COX2 and others

Gastric Colorectal, Lung, Breast

2 2

1 1

Inflammation Inflammation Interleukins

Multiple ODC IL-1, IL-8

Prostate Colorectal Gastric, Lung

1 2 2

1 1 1

GG GE-aspirin GE-H. pylori

Interleukins

TNFalpha

1

1

Lipid metabolism Matrix metalloprotein ases

APOE MMP-1, MMP-3, MMP-2

Leukemia, Lymphoma, Cervix Colorectal Lung, Breast, Gastric

1 1

1 1

GE-infection, (CIN in cervix) GE-gender

Major gene modifier

ATM

Breast

1

1

Major gene modifier Major gene modifier Major gene modifier

HPC BRCA1 APC

Prostate Breast Colorectal

2 2 1

1 0 1

Major gene modifier Major gene modifier Major gene modifier Metalloprotease family (MMPs)

MSR1 RET CDKN1* ADAM33, MM1

Prostate Thyroid Prostate Tobacco-related cancers and related conditions

1 1 1 1

1 1 0 0

Miscellaneous Miscellaneous Nutrient

HFE MDR-1 VDR

Breast Lung Prostate, Colon

1 1 1

1 0 0

Nutrient

MTHFR (also related, MTR, MS)

Colorectal, NHL, Gastric, Breast Endometrum; Prostate Leukemia (ALL) Bladder

2

1

Nutrients Obesity (energy balance)

Vitamin E Leptin (LEP), LEPR, PPAG

Prostate Endometrial, Kidney, Prostate, Colon, Esophagus, Breast (postmenopausal)

1 1

0 1

GG = NAT1/2

GE-calcium, Vitamin D GG-IDGF GE-folate, homocyst., Cardiovascular, diet, alcohol GG = one-carbon pathway

GE-weight, nutrition GE-insulin resistance, energy balance

(continued)

585

Table 29–5. (cont.) Gene Family

Gene Examples

Relevant Cancers

Mech

Pop

G-E or G-G

Phase 1

CYP1A1 (Related CYP1A2)

Lung, Prostate, Head and Neck, Renal, Endometrium, Acute Leukemia Colorectal

2

2

GE-tobacco; GG-GSTM1

Phase 1

MPO

Lung

2

1

GE = H. pylori

Phase 1

CYP1B1

Breast, Prostate, Endometrium Colorectal

1

1

Phase 1

CYP2A6 (CYP2A13)

Lung, Esophagus

2

1

Phase 1

CYP2D6

1

1

Phase 1

CYP3A4 CYP3A5

Lung, Bladder, Breast, others Prostate, Lung, Various

1

1

Phase 1

CYP2E1

Gastric, Lung, NPC, Prostate

1

1

Phase 1 Phase 1/2

CYP2C9 MEH

Lung Lung

0 2

1 1

Phase 2

NAT2 (NAT1)

Bladder Lung, Breast

2

2

GE-tobacco HAA

Phase 2

NAT2 (NAT1)

Colorectal, Gastric, HCC

2

1

GE-meat consumption

Phase 2

NQO1

1

1

GE-benzene

Phase 2

GST family

Lung, Bladder, Colorectal Lung, Bladder, Head and Neck, Breast, Thyroid, NHL, Esophagus, Skin (Malignant Melanoma, Basal Cell Carcinoma), Larynx, Renal, Ovary, Testicular Brain, Leukemia, Cervix, HCC

2

1

GG-CYP1A1 GE-smoking, young age (lung); GSTT1 and HPV (cervix), GST1/ M1-cruciferous vegetables (lung)

Phase 2

UGT1A1 (UGT1A7)

Lung, Oropharyngeal, Head and Neck, Pancreas, HCC

2

1

Phase 2

SULT1A1

Prostate, Lung

2

1

586

GE-tobacco, nicotine

Reference Kellermann, 1973 (lung) Kihara, 1995 (lung) Sugawara, 2003 (female tumors) Chang, 2003a (prostate); Longuemaux, 1999 (renal) D’Alo, 2004 (AML) Esteller, 1997 (endometrium); Landi, 2005 (colorectal) Hung, 2003 (nonsmokers) LeMarchaud 2003 Vineis 2003 Roe, 2002; Caporaso, 2002; Chevier, 2003; Kiyohara, 2005 Kantarci, 2002; Schabath, 2000 (lung) Roe, 2002 (gastric) Rylander-Rudqvist, 2003 (breast); Landi, 2005 (colorectal) Sasaki, 2003 (endometrium) Tanaka, 2002 (prostate) Chang, 2003b (prostate) Hecht, 2000; Wang, 2003; Minematsu, 2003 (mechanism) Tan, 2001 (lung and esophagus) Loriot, 2001 (lung) Oscarson, 2001 (nicotine) Ayesh, 1984 (first study) Lamba, 2002 (mechanism) Plummer, 2003 (prostate); Tayeb, 2003; Dally, 2003 (lung) Ferreira, 2003 (prostate) Kongruttanachok, 2001 (NPC) Park, 2003 (gastric) Garcia-Martin, 2002 Cajas-Salazar, 2003; Gsur, 2003 Lee, 2002 (review) Lower, 1979 (bladder, first study) Muckel, 2002 (mechanism) Wikman, 2001 (lung) Weber, 1987 (review) Lan, 2003 (gastric) LeMarchaud, 2002; Landi, 2005 (colorectal) Huang, 2003 (HCC, red meat) Brockton, 2000 (colorectal) Moran, 1999; Chen, 1999 (lung) Park, 2003 (bladder); van der Logt, 2005 Seidegard, 1986 (first study) Egan, 2004 (breast), Coughlin, 2002 Miller, 2002 (diverse GSTs) Evans, 2004 (head and neck) Hung, 2004 (bladder) Lee, 2004 (cervix) Rollinson, 2000; d D’Alo, 2004 (acute leukemia); Kirk, 2005 (HCC) Sweeney, 2000 (renal) DeRoos, 2003; Ezer, 2002 (brain); Brennan, 2005 (lung) McWilliams, 1995 (lung) Benhamou, 2002; Cotton, 2000 (GST and colorectal) Houlston, 1999 (GST and lung) Taioli, 2003 (GSTM1 and lung, age <45) Engel, 2002 (GSTM1 and bladder) Geisler, 2001 (GSTs and sq head and neck) Wiener, 2004; Burchell, 2003 (mech) Fang, 2002; Tseng, 2005 (HCC) Elahi, 2003 (oropharynx) Ockenga, 2003; Dugay, 2004 (pancreas) Muckel, 2002 (mechanism) Wang, 2002 (lung) Nowell, 2004 (prostate)

Table 29–5. (cont.) Gene Family

Gene Examples

Relevant Cancers

Mech

Pop

G-E or G-G

Pigmentation

MC1R

Melanoma

2

1

GE-sun, freckling, mole count

Pigmentation Proto-oncogene

MC1R HRAS1, HER2

Skin Cancer Lung, Breast, others

2 1

1 1

Same

Senescence Signal transduction TSG TSG

MnSOD ICAM PTEN P53

General Breast, Prostate HCC Lung, HCC, Cervix, Endometrial, Colorectal Esophagus

1 2 1 2

0 1 0 0

TSG TSG

FHIT CDH1

Cervix Breast, Lung, Gastric

1 1

0 1

TSG

KLK10 NKX3.1

Prostate Prostate

1 1

1 1

GE-hepatitis

Reference Sturm, 2003 (mechanism) Hayward, 2003; Hearle, 2003; Vajdic, 2003 (ocular melanoma) Sturm, 2003 Rutter, 2004; Tamimi, 2003 (breast) Lindstedt, 1999 (lung) Krontiris, 1993 (review) Lundberg, 2000 Kammerer, 2004 (breast) Chung, 2003 Jackson, 2003 (HCC) Koushik, 2004 (cervix) Roh, 2004 (endometrial) Gemignani, 2004 (colorectal) Hong, 2005 (esophagus) Jee, 2003 Toyooka, 2001 Humar, 2002 (gastric) Bharaj, 2002 Gelmann, 2002

Mechanism: 0 = no plausible mechanism related to cancer is established; 1 = possible cancer mechanism is associated with the gene but evidence weak or mechanism only weakly linked to cancer; 2 = cancer mechanism is well-established for the gene. Population evidence: 0 = no population data or null; 1 = some suggestive population data exists but too sparse to draw firm conclusions; 2 = association with cancer in population studies is reasonably well-established. Boldface type indicates data derive from a review, or meta- or pooled analysis.

Table 29–6. Specific Cancers and Associated Gene Families Cancer

Key Families and Genes

Bladder

Phase 1 and Phase 2, DNA repair, Apoptosis, Nutrient (MTHFR/folate)

Breast

Steroid hormone metabolism genes (estrogen, progesterone and androgen), carcinogen metabolism genes, DNA repair, Nutrients (especially folate pathways), Major gene modifiers (BRCAx, ATM), Immunologic (HLA), Phase 1 and Phase 2, Proto-oncogene and TSG, Nutrient, Signal transduction, Growth Factor, Circadian Phase 1 and Phase 2

Brain

Cervix

TSG, Phase 2, Nutrients

Colorectal (including adenomas)

Inflammation pathways; DNA repair, Nutrients, Cytokines; Phase 1; Phase 2; Major gene

Comment Park, 2003 (NQO1) Hazra, 2003; Sanchez-Carbayo, 2002 Stern, 2002 (XPD) Hung, 2004 (MTHFR, Phase 2s); Lin, 2004 (MTHFR) Marcus, 2000a (NAT2) Marcus, 2000b (NAT2 and tobacco) Vineis, 2001 (NAT2) Engel, 2002 (GSTM1) Zheng, 2004 (sex hormone); Gold, 2004 (estrogen receptor) Angele, 2003; Lee, 2005 (ATM); Zheng, 2004 (CYP11A) Egan, 2004 (Phase 2 review) Zhang, 2003 Han, 2004 (DNA repair genes) Rutter, 2003 (HER) Setiawan, 2004 (HSD17B1) Suter, 2003 (androgen repeats) Tamimi, 2003 (H-ras) Dunning, 1999 (review), Zhu, 2005 (PER1) Menzel, 2004 (NQO1 and p53) Chen, 2005 (MTHFR) Kammerer, 2004 (ICAM) Ren, 2004 (IGFBP3) GSTs (De Roos, 2003; Ezer, 2002) CYP2E1 (DeRoos, 2003) require further study EGRF dysregulation, Weiss (2003) Koushik, 2004 (p53) Lee, 2004 (GSTM1) Sharma, 2004 (GSTs) Landi, 2003 (inflammation) Martinez, 2003 (ODC and adenoma) LeMarchaud, 2001 (NATs) Herman, 1998; Ulrich, 1999; Giovannucci, 2000 Bailey, 2003; Curtin, 2004 (MTHFR)

Cancer

Key Families and Genes modifiers, Alcohol, Cell cycle, TSG

Endometrial

Estrogen/hormone; Phase 1; Nutrient; TSG

Esophagus

Phase 1 and Phase 2, Alcohol Metabolizing Genes, Hormones

Gastric

Hormone polymorphisms (pepsinogen); Phase 1, DNA repair, Nutrient, Phase 2, Immune

Head and Neck

Alcohol metabolizing genes; Phase 1 and 2; DNA Repair

Comment Lewis, 2003 (CCND1) Goode, 2004 (environmental risks) Hahnloser, 2003 (APC) Cotton, 2000 (review GSTM1) Brockton, 2000 (review NAT1/NAT2) LeMarchand, 2001 (colorectal) Tiemersma, 2003 (alcohol) Gemignani, 2004 (p53) Giovannucci, 2003 (MTHFR/ADH3/alcohol) Dugay, 2004 (estrogen) Berstein, 2002 (CYP17) Sasaki, 2003 (CYP1B1) Esteller, 1997 (Phase 1 and MTHFR); Sugawara, 2003 (CYP1A1) Roh, 2004 (p53 and p21); Jee, 2003 (FHIT) Hamajima, 2001; Yang, 2005 (ALDH2) Zhang, 2003 (cyclin D) Tan, 2001 (CYP2A6) Liu, 2003 (pepsinogen C) Humar, 2002 (CDH1) Park, 2003; Wu, 2002 (CYP2E1) Roe, 2002 (MPO) Gao, 2002 (MTHFR) Ming, 2003 (MMP2) El-Omar, 2002, 2003; Taguchi, 2005 (interleukin genes) Lan, 2001; Lan, 2003 (Phase 2) Hardy, 1997 (ADH3) Geisler and Olshan, 2001 (GST family) Elahi, 2003 (UGT1A1) Lee, 2002 (XRCC) Elahi, 2002 (OGG1) Evans, 2004 (GSTT1) Hasibe, 2003 (Phase 1 and 2) Brennan, 2003 (ADH3) Geisler, 2003 (GSTs)

(continued)

587

588

PART III: THE CAUSES OF CANCER

Table 29–6. (cont.) Cancer

Key Families and Genes

HCC

Alcohol metabolizing genes; Phase 1 and 2 Steroid hormone metabolizing genes DNA Repair

Leukemia (acute)

Phase 1 and 2, Apoptosis; Nutrition (folate metabolism); DNA repair; Drug transport

Lung

Phase 1 and 2; DNA repair; Nutrition, Inflammation/ Immune

Lymphoma

Immune response (HLA, cytokines), Phase 2, Folate metabolism, DNA repair, Apoptosis

Melanoma

NPC

DNA repair, Pigmentation, Major gene variants Immune (HLA, PIGR), DNA repair,

Comment Yu, 2002; Sakamoto, 2005 (ALDH2) Kato, 2003 (CYP2E1); Rossi, 2003 (hormones); Tseng, 2005 (UGP1A7) Munaka, 2003 (alcohol-related and Phase 1) Kirk, 2005 (GSTM1, XRCC1) Huang, 2003 (NAT2 and red meat) Sbley, 2003 (FAS) Krajinovic, 2004 (ALL and MTHFR) Infante-Rivard, 2003 (DNA repair) Bowen, 2003 (Phase 1) D’Alo, 2004 (AML and CYP1A1 and GSTT1) Jamroziak, 2004 (ALL and MDR1) Rollinson, 2000 (Phase 2); van der Logt, 2005 (NQO1) Robien, 2003 (leukemia) Gsur, 2003 (mEH) Snoek, 2000 (HLA) Xing, 2003 (XPD) Cajas-Salazar, 2003 (EH) Loriot, 2001; Tan, 2001 (CYP2A6) Wang, 2003 (CYP2A13) Chevier, 2003 (MTHFR) Chen, 1999 (NADPH) Sinuues, 2003 (MDR1) Quiling, 2003 (CCND1) Zienolddiny, 2004 (IL1B) Wikman, 2001 (NATs) Miller, 2002; McWilliams, 1995 (GSTM1) Benhamou, 2002 (GSTM1); Houlston, 1999 (GSTM1) LeMarchaud, 2003 (CYP1A1, exon 7) Rostami-Hodjegan, 1998 (CYP2D6) Taioli, 2003 (CYP1A1 and GST, < age 45) Vineis, 2003 (CYP1A1) Howell, 2002 (mechanism) Hishida, 2003 (thymidylate synthase), Zhang, 2005 (apoptosis) Cerosaletti, 2002 (DNA repair) Kerridge, 2002; Battuello, 2004 Cunningham, 2003 (inflammatory cytokines) Robien and Ulrich, 2002 Garcia-Borron, 2003; Millikan, 2005 (DNA repair) Hayward, 2003 (MC1R) Tomescu, 2001 (XPD) Bertram, 2002 (major genes) Lu, 2003; Jalbout, 2003 (HSP) Hirunsatit, 2003; Butschkovacic, 2005 (HLA)

Cancer

Key Families and Genes Phase 1, TSG

Ovary Pancreas

Phase 2; Galactose; Estrogen, Phase 1 DNA repair; Phase 1 and 2

Prostate

Steroid hormone genes; Growth factors; Cell cycle; Obesity/energy regulation; Nutrient TSG Phase 1 and 2 TSG; Major gene variant; Nutrients; DNA repair, Angiogenesis

Skin

DNA repair; Pigmentation

Renal

Phase 1, Phase 2, Angiogenesis

Thyroid

Tyrosine kinase receptors Phase 2

Testicular

Comment Cho, 2003 (DNA repair) Kongruttanachok, 2001 (CYP2E1) Cho, 2003 (Phase 1) Tiwawch, 2003 (p53) Jalbout, 2003 (HSP71-2) Coughlin, 2002 (GST) Fung, 2003 (GALT) Goodman, 2001 (CYP1B1) Duell, 2002a and 2002b (DNA repair) Ockenga, 2003 (UGT1A7) Soderstrom, 2002; Chang, 2003 (SRD5A2) Margiotti, 2002 (HSD) Allen, 2003; Li, 2003 (testosterone) Mononen, 2000 (androgen) Madigan, 2003 (CYP17) John, 2005 (VDR) Gelmann, 2003 (TSG) Nowell, 2004 (SULT1) Ho (insulin) Medeiros, 2002, 2003 (NOS) Fang, 1992 (misc. genes) Xu, 2002 (MSR1) Chang, 2003; Tanaka, 2002 (CYP1B1) Miller, 2003 (macrophage scavenger) Kibel, 2003 (CDKN1*) Cybulski, 2004b (NBS1) Plummer (CYP3A4) Tayeb, 2003 (CYP3A4) Wu, 2002; Ferreira, 2003; Chang, 2003 (Phase 1 and 2) Miller, 2003; Adler, 2003; Stanford, 2003; Severi, 2003; Meitz, 2002 (familial prostate genes) Chokkalingam, 2001 (vitamin D) Rybicki, 2004 (DNA repair) Simard, 2003 (overview); Ntais, 2003a (SRD5A2); Ntais, 2003b (vitamin D); Ntais, 2003c (CYP17) Nelson, 2002; Han, 2004 and 2005 (DNA repair) Sturm, 2003 (MC1R) Longuemaux, 1999 (CYP1A1) Sweeney, 2000 (GSTT1) Sweeney, 2000 (GSTs) Abe, 2002 (VEGF) Kimura, 2003 (RET/PTC-RASBRAF pathway) Youngren, 2003 (mechanism) Harries, 1997 (GSTP1)

*Boldface type indicates data derive from a pooled or meta-analysis.

understanding their role and mechanisms of action will offer new perspectives for prevention.

OVERVIEW OF CANCER-SPECIFIC ASSOCIATIONS Details of studies organized by tumor type are summarized in Table 29–5. The tobacco-related cancers currently provide the best examples of reasonably well-established associations, notably the NAT2 “slow acetylator” polymorphism and the GSTM1 “null” genotype with

bladder cancer. Lung cancer exhibits associations with a Phase 1 gene (e.g., CYP1A1) and a Phase 2 gene (e.g., GSTM1), but the precise role that these genes play in concert and the degree of gene–environment and gene–gene interaction await larger studies. Other tobaccorelated neoplasms such as pancreas cancer have not been as extensively studied, and because the degree of association with tobacco is less than with lung cancer, associations with genes involved in processing the exposures may be correspondingly weaker. Next most studied are the hormone-related malignancies such as breast, endometrial, and prostate cancers. In general, consistent asso-

Genetic Modifiers of Cancer Risk ciations with specific mechanistically plausible genes have not been demonstrated. Much larger and more comprehensive studies are planned in a coalition of cohort studies involving replication and pooling strategies. Colorectal cancer and adenomas have been widely studied, and nutrient-related genes (e.g., MTHFR) and inflammation-related genes (e.g., COX2) are the most plausible candidates. Genes involved in immunerelated and inflammatory processes are the best current candidates in gastric cancer. There are some suggestive findings in melanoma involving genes that influence skin pigmentation (e.g., MC1R) and in NHL in genes involved in immune processes, consistent with the known factors that influence their etiology. For the cancers where etiologic factors are poorly understood, such as testicular and brain cancers, and leukemia, susceptibility genes are not yet established. As the relation of cancer to susceptibility genes is investigated in larger population samples with greater numbers of genes, it is likely that strong relationships of genes with intermediate phenotypes or “endophenotypes” will be found. Genes involved in myriad precursor states and mechanisms relevant to cancer such as Helicobacter pylori infection, mammographic density (Lillie, 2004), addiction to alcohol or nicotine, enhanced inflammatory response to hepatitis C virus, chronic obstructive pulmonary disease, or poor DNA repair after sun damage may reveal many strong associations with these intermediate phenotypes that are stronger than with cancer. Another emerging trend is the investigation of histological and molecular subclassifications of cancer. Associations of polymorphic gene variants with specific disease histologies are illustrated by lung cancer, e.g., small cell lung cancer (Daly, 2003), adencarcinoma (Wang, 2003), or tendecy to early metastasis (Rostami-Hodjegan, 2003). Selected histologic variants of esophageal (Dally, 2003; Wang, 2003) or gastric cancer (Meireles, 2004), and folliclular center lymphoma (Lossos, 2001) are further examples. Related to this will be investigations of the genetics of cancer precursors such as atypical adenomatous hyperplasia (lung) (Marcus and Travis, 2002) or B-cell monoclonal lymphocytosis (B-cell malignancies) (Marti et al., 2003). It seems likely that certain genes and their families will be shown to alter risk for a spectrum of cancer types. Tobacco carcinogenmetabolizing genes may affect a number of smoking-related cancers, although the precise tobacco constituents, mechanisms, and levels of risk may vary by tumor type. Similarly, genes that affect energy balance, inflammation, DNA repair, and other critical pathways may have an effect common to several types of tumors as well as other chronic conditions. Cohort studies (where multiple tumor end-points may be studied) will be useful to address these questions. A complementary consideration is that a complex disease such as cancer will likely be influenced by many genes and pathways (i.e., a polygenic mechanism). Theoretical constructs (multistage theories of carcinogenesis) or simple consideration of the independent steps necessary for a cancer to take hold suggest that genes can operate on multiple targets (i.e., different carcinogens), directions (to accelerate, block, or dysregulate a process), and biological levels (i.e., DNA, cell, tissue, organ, etc.) compatible with the action of many genes.

META-ANALYSES Table 29–7 summarizes published pooled and meta-analyses exhibiting suggestive findings for well-studied genes and specific cancers. Although subject to well-known limitations as well as the rapid acceleration of molecular technologies that render older approaches archaic, these studies provide the current best assessments of reported gene–cancer associations. The much larger and more comprehensive studies expected in the next few years should provide a clearer picture for the role of these genes in cancer. Overall, we are on the early part of an exponential growth curve with regard to genetic investigations of complex diseases, so that the status of individual gene–cancer relationships and the pattern of overall associations that will emerge are difficult to predict based on the available data. Nevertheless, given the mechanistic data and the body of work represented by the metaanalyses, evidence-to date supports a role for metabolic genes in alter-

589

ing risk of tobacco-related cancers. Most convincing is the increased risk associated with GSTM1 null (lung and bladder cancer), CYP1A1 variant (lung cancer), and NAT2 “slow acetylators” (bladder cancer). Even with their limitations, pooled analyses also provide the only substantial data on gene–environment effects and combinations of genes. For example, increased risk of lung cancer due to at-risk variants of CYP1A1 and GSTM1 was observed in nonsmokers (Hung, 2003), those with both at-risk gene variants and in those below age 45 (Taioli, 2003). The available evidence also suggests that NAT2 slow acetylators have a more pronounced risk of bladder cancer among tobacco smokers (Marcus, 2000a, 2000b).

GENE–ENVIRONMENT INTERACTION All genes act in the context of the environment, and early models for gene–environment interaction were described by Ottman (1990). Gene–environment interaction may be defined as a differential effect of the environment on disease risk based on genotype, or a differential effect of a genotype based on environmental exposures. Smith and Day (1984) showed that detecting interactions of the same magnitude as main effects in 1:1 unmatched case-control studies required at least a fourfold increase in study size. Depending on whether an additive or multiplicative scale as a statistical model is defined, the presence or absence of interaction may entail some ambiguity. When genes are rare or environmental factors uncommon, a countermatching design may provide a way to improve efficiency (Andrieu et al., 2001). A likely place to observe combined effects in cancer risk involve genes that metabolize extrinsic carcinogens such as bladder cancer associated with the interaction of NAT2 genotype and exposure to aromatic amines in tobacco smoke (Marcus, 2000b). In addition, the MTHFR C677T polymorphism increases the rate of colorectal adenomas when folate levels are low (Ulrich, 1999), and individuals homozygous for the ODC A-allele who use aspirin are less likely to experience adenoma recurrence than those homozygous for the major G-allele (Martinez, 2003). Other gene–environment relationships are plausible based on preliminary study, for example, increased risk of colorectal cancer in rapid NAT2/CYP1A2 phenotypes in those who consume well-done red meat (LeMarchand, 2001), a protective effect of cruciferous vegetables limited to GSTM1/T1 null subjects (Brennon, 2005) and an increased risk of squamous esophageal cancer associated with alcohol drinking in ALDH2 “inactive” subjects (Yokoyama, 2003). Individuals with inactive ALDH2 generally consume less alcohol because this genotype results in slow elimination of acetaldehyde and causes “flushing” reactions. However, when these individuals consume alcohol, risks of esophageal and possibly other cancers are increased (Yokoyama, 2003). Because of limited efforts to consider genes and the environment in concert, little is known about the contribution of specific agents in complex mixtures as well as causal pathways in cancer risk. For example, there are many reported studies of CYP1A1 polymorphisms based on their putative role in activation of PAH carcinogens in tobacco-related cancer. Another group of studies assessed GSTM1, involved in the elimination of PAHs. However, only a few studies take environmental cofactors and both genes into account (Smith, 2001; Alexandrie, 2004), with some studies finding a combined effect of the genes that is more substantial. However, no studies have evaluated a comprehensive group of the genes involved in PAH metabolism or all the genes involved in the activation or detoxification of carcinogens present in tobacco smoke (e.g., nitrosamines, aryl amines, nicotine). The available designs for study of gene–environment interaction (Goldstein, 1999) and the impact of disease and exposure misclassification and sample size in population-based designs have been reviewed (Garcia-Closas et al., 1998, 1999, 2003, 2004). Case-only designs can be used when certain conditions are met (i.e., exposure is independent of genotype) (Yang, 1997) and main effects or additive models cannot be directly assessed. It has been suggested that the population attributable fraction due to interaction be considered in power calculations in order to gauge feasibility and public health impact (Yang, 2003). Whatever the design, it is important to obtain detailed

590

PART III: THE CAUSES OF CANCER Table 29–7. Selected Meta-analyses and Pooled Analyses of Candidate Genes in Relation to Specific Cancers Cancer Lung Lung Lung Lung Lung Lung Lung Lung Lung

Gene

No. Studies

GSTM1 GSTM1 MPO (G > A) CYP2D6 CYP1A1 and GSTM1 (Caucasian nonsmokers) CYP1A1 (MspI-Asians)2 CYP1A1 (MspI-Caucasians) CYP1A1, “exon 7”

23 21 ?

2.12 (1.43–3.13) 1.34 (1.21–1.48)1 0.52 (0.39–0.90)

Houlston, 1999 D’Errico, 1999 Schabath, 2000

16 14

1.28 (1.01–1.58) CYP1A1 ile(462)Val 2.99 (1.51–5.91) GSTM1 1.20 (0.89–1.63) 1.73 (1.30–2.31)

D’errico, 1999 Hung, 2003

10

1.04 (0.85–1.27)

D’errico, 1999

11

1.15 (0.95–1.39) het 1.54 (0.97–1.46) homo p for gene-dosage 0.03 2.36 (1.16–4.81) homozygous variant 1.17 (1.07–1.27) 1.41 (1.23–1.61) 0.70 (0.51–0.96) 0.65 (0.44–0.96) 0.70 (0.56–0.88) 0.70 (0.47–1.04) 1.37 (1.20–1.57) 1.40 (1.2–1.6) 1.30 (1.0–1.6) 1.42 (1.14–1.77) with increase in risk limited to current smokers (OR = 1.74, 0.96–3.15) 1.57 (1.36–1.81) 1.44 (1.23–1.68) 1.11 (1.02–1.39) 1.23 (1.06–1.42) 1.17 (0.98–1.40) 1.35 (0.95–1.82) 1.93 (1.63–2.30)

Le Marchaud, 2003

Lung Lung Lung Lung Lung Lung Bladder Bladder Bladder Bladder

CYP1A1 T3801C GSTM1 GSTM1 EPHX1 Exon 3 his/his EPHX1 (his/his)4 NQ01 (pro187ser)5 MP0 (G-463A)6 NAT2 NAT2 NAT2* ever smoking3 NAT2

Bladder Bladder Colorectal Head and Neck Head and Neck Head and Neck “Cancer”

GSTM1 GSTM1 NAT2 GSTM1 GSTT1 CYP1A1 (Val462) HRAS1 “rare”

4

22 43 12 8 8 3 22 16 22 16 7

11 17 12 28 18 23

Summary OR and 95% CI

Author and Year

D’errico, 1999

Vineis, 2003 Benhamou, 2002 McWilliams, 1995 Lee, 2002 Kiyohara, 2005 Kiyohara, 2005 Kiyohara, 2005 D’errico, 1999 Marcus, 2000a Marcus, 2000b Vineis, 2002

D’errico, 1999 Engels, 2002 D’errico, 1999 Hashibe, 2003 Hashibe, 2003 Hashibe, 2003 Krontiris, 1993

Caucasians only, 13 studies, OR = 1.21 (1.06–1.39). CYP1A1, Asians exon 7, 3 studies, OR = 2.25 (1.37–3.69); Caucasians 4 studies, OR = 1.30 (0.89–1.90). An interaction is this context means that smokers who are slow acetylators have a higher risk of bladder cancer than smokers who are rapid acetylators, and that slow acetylators who do not smoke are at similar risk to nonsmoking rapid acetylators. 4 Whites only. 5 Japanese. 6 Caucasians only. 1 2 3

data on clinical status, exposure, and related biological information derived from specimen evaluation to provide the best context for interpretation.

GENE–GENE EFFECTS Genes also act in concert with other genes, and animal models suggest that epistatic relationships between genes may be common (Moore, 2003). Epistasis was originally defined as “masking” of one gene’s effect by another but more broadly refers to differential phenotypic expression of a genotype at one locus dependent on a genotype at another. An example of genes thought to exhibit this type of effect among modifier loci in humans is the combination of the Phase 1 CYP1A1 minor allele and the Phase 2 GSTM1 null genotype (Kihara et al., 1995; Taioli et al., 2003; Vineis et al., 2004). Both of these genes seem to exhibit an independent effect on some tobacco-related cancers (see Table 29–6), but there is suggestive evidence that their combined effects may be supra-additive. Sample size and design considerations for population-based studies involving gene–gene interactions are described by Wang and Zhao (2003). The ability to recognize gene– gene effects is complicated by the degree to which genes act in pathways or flexible networks. For example, if a mutation in one gene alters the genetic pathway, it becomes difficult to detect combined effects (Greenspan, 2001). A data mining approach, multifactor dimensional-

ity reduction, has been used to identify putative higher-order gene–gene interactions in the inflammation pathway in prostate cancer (Xu, 2005). It is important to appreciate that genes operate in pathways and that feedback, homeostatic, and compensatory mechanisms are involved, so that it would be unusual for polymorphism in one particular gene to have a critical effect on a common cancer. In order to dissect the intrinsically weak effects of modifier genes, it is more realistic to evaluate genes comprehensively, including multiple SNPs and the joint effects of related genes in a particular pathway or family.

PROSPECTS The investigation of modifying genes in cancer induction is a dominant theme in molecular and genomic epidemiology at the beginning of the new millennium (Ioannidis, 2006). The integration of genomics and epidemiology in large population studies should allow application of advanced technologies and participation of scientists from diverse disciplines to advance our understanding of the role of genes across the cancer spectrum. A primary goal of epidemiology is to understand the role of genes in cancer causation and progression in large-scale epidemiologic platforms. Future studies will clarify their interactions with environmental exposures, their role at various stages of carcinogenesis including precursor states, their effects on molecular effectors in tissue, and their influence on outcome and response to therapy.

Genetic Modifiers of Cancer Risk While progress in identifying the precise hereditary determinants that contribute to common cancers has not been as rapid as some may have anticipated, meta-analyses support a number of associations of common genes with cancer and other complex diseases, and it is likely that we are at a very early stage of discovery. As high-throughput genotyping is combined with well-designed large-scale population studies supported by efficient biospecimen collection and informatics capability, specific genes and pathways that are critical to cancer should emerge. These insights will fulfill the promise of the genomics revolution and galvanize new approaches to cancer prevention, diagnosis, and treatment. References Abe A, Sato K, Habuchi I, Wang L, Tsuchiya N, Ohyama C, et al. 2002. Single nucleotide polymorphisms in the 3¢ untranslated region of VEGF gene in Japanese population with or without renal carcinoma. Tohoku J Exp Med 198(3):181–190. Adler D, Kanji N, Trpkov K, Fick G, Hughes RM. 2003. HPLC2/ELAC2 gene variants associated with incident prostate cancer. J Hum Genetics 48(12):634–638. Ahlbom A, Lichtenstein P, Malmstrom H, Feychting M, Hemminki K, Pedersen NL. 1997. Cancer in twins: Genetic and nongenetic familial risk factors. J Natl Cancer Inst 89(4):287. Aird I, Benthall HH, Roberts JAF. 1953. A relationship between cancer of the stomach and the ABO blood groups. BMJ 1:179–195. Alexandrie AK, Nyberg F, Warholm M. 2004. Influence of CYP1A1, GSTM1, GSTT1, and NQO1 genotypes and cumulative smoking dose on lung cancer risk in a Swedish population. Cancer Epidemiol Biomarkers Prev 13(6):908–914. Allen NE, Reichardt JK, Nguyen H, Key TJ. 2003. Association between two polymorphisms in the SRD5A2 gene and serum androgen concentrations in British men. Cancer Epidemiol Biomarkers Prev 12(6):578–581. Amatschek S, Koenig U, Auer H, Steinlein P, Pacher M, et al. 2004. Tissuewide expression profiling using cDNA subtraction and microarrays to identify tumor-specific genes. Cancer Res 64:844–856. Andrieu N, Goldstein AM, Thomas DC, Langholz B. 2001. Counter-matching in gene-environment interaction studies:efficiency and feasibility. Am J Epidemiol 1;153(3):265–274. Angele S, Romestaing P, Moullan N, Vuillaume M, Chapot B, et al. 2003. ATM haplotypes and cellular response to DNA damage: Association with breast cancer risk and clinical radiosensitivity. Cancer Res 63:8717–8725. Antonious AC, Easton DF. 2003. Polygenic inheritance of breast cancer:implications of design for association studies. Genet Epidemiol 25(3): 190–202. Anttila S, Hakkola J, Tuominen P, Elovaara E, Husgafvel–Pursiainen K, Karjalainen A, Hirvonen A, Nurminen T. 2003. Methyation of cytochrome P4501A1 in the lung is associated with tobacco smoking. Cancer Res 63:8623–8628. Ayesh R, Idle JR, Ritchie JC, Crothers MJ, Hetzel MR. 1984. Metabolic oxidation phenotypes as markers of lung cancer susceptibility. Nature 312:169–170. Bailey LB. 2003. Folate, methyl-related nutrients, alcohol and the MTHFR 677C > T polymorphism affect cancer risk: Intake recommendations. J Nutr 133(11):3748S–53S. Bamford S, Dawson E, Fortes S, Clements J, Pettett R, Dogan A, Flanagan A, Teague J, Futreal PA, Stratton MR, Wooster R. 2004. The COSMIC. Catalogue of Somatic Mutations in Cancer. Dataset and website. Br J Cancer 91(2):355–358. Barcellos LF, Klitz W, Field LL, Tobias R, Bowcock AM, Wilson R, Nelson MP, Nagatomi J, Thomson G. 1997. Association mapping of disease loci, by use of a pooled DNA genomic screen. Am J Hum Genet 61:734–747. Bartsch H, Nair U, Risch A, Rojas M, Wikman H, Alexandrov K. 2000. Genetic polymorphism of CYP genes, alone or in combination as a risk modifier of tobacco-related cancer. Cancer Epidemiol Biomarkers Prev 9(1):3–28. Battuello K, Furlong C, Fenske R, Austin MA, Burke W. 2004. Paraoxonase polymorphisms and susceptibility to organophosphate toxicity. In: Khoury MJ, Little J, Burke W, eds. Human Genome Epidemiology. New York: Oxford University Press, pp. 305–321. Baylin SB, Herman JG. 2000. DNA hypermethylation in tumorigenesis:epigenetics joins genetics. Trends Genet 16:168–174. Bauer AK, Faiola B, Abernathy DJ, Marchan R, Pluta LJ, Wong VA, et al. 2003. Genetic susceptibility to benzene-induced toxicity: Role of NADPH: quinone oxidoreductase. Cancer Res 63:929–935. Belmont JW, Gibbs RA. 2004. Genome-wide linkage disequilibrium and haplotype maps. Am J Pharmacogen 4(4):253–262.

591

Benhamou S, Lee WJ, Alexandrie AK, Boffetta P, et al. 2002. Meta- and pooled analyses of the effects of glutathione S-transferase M1 polymorphisms and smoking in lung cancer. Carcinogenesis 23(8):1343–1350. Berstein LM, Imyanitov EN, Gamajunova VB, Kovalevskij AJ, Kuligina ESH, Belogubova EV, Buslov KG, Karpova MB, Togo AV, Volkov ON, Kovalenko IG. 2002. CYP17 genetic polymorphism in endometrial cancer:are only steroids involved? Cancer Lett 180(1):47–53. Bertram CG, Gant RM, Barrett JH, Pinney E, Whitaker L, Turner F, Bataille V, et al. 2002. An assessment of CDKN2A variant Ala148Thr as a nevus/melanoma susceptibility allele. J Invest Dermatol 119(4):961–965. Berwick M, Vineis P. 2000. Markers of DNA repair and susceptibility to cancer in humans:an epidemiologic review. J Natl Cancer Inst 92(11):874–897. Beskow LM, Burke W, Merz JF, Barr PA, Terry S, Penchaszadeh VB, Gostin LO, Gwinn M, Khoury MJ. 2001. Informed consent for population-based research involving genetics. JAMA 286(18):2315–2321. Bharaj BB, Luo LY, Jung K, Stephan C, Diamandis EP. 2002. Identification of SNP Polymorphisms in the human kallikrein 10 (KLK10) gene and their association with prostate, breast, testicular, and ovarian cancers. Prostate 51(1):35–41. Bioinformatics and Functional Genomics. 2003. Jonathan Pevsner, John Wiley and Sons, Inc., Hoboken, New Jersey, Wiley-Liss. Bonassi S, Hagmar L, Stromberg U, Montagud AH, Tinnerberg H, et al. 2000. Chromosomal aberrations in lymphocytes predict human cancer independently of exposure to carcinogens. Cancer Res 60:1619–1625. Bonnen PE, Wang PJ, Kimmel M, Chakraborty R, Nelson DL. 2002. Haplotype and linkage disequilibrium architecture for human cancer-associated genes. Genome Res 12:1846–1853. Boren T, Falk P, Roth KA, Larson G, Normark S. 1993. Attachment of Helicobacter pylori to human gastric epithelium mediated by blood group antigens. Science 262(5141):1892–1895. Botstein D, Risch N. 2003. Discovering genotypes underlying human phenotypes:past successes for Mendelian disease, future approaches for complex disease. Nat Genet 33(Suppl):228–237. Bowen DT, Frew ME, Rollinson S, Roddam PL, Dring A, Smith MT, Langabeer SE, Morgan GJ. 2003. CYP1A1*2B (Val) allele is overrepresented in a subgroup of acute myeloid leukemia patients with poor-risk karyotype associated with NRAS mutation, but not associated with FLT3 internal tandem duplication. Blood 101(7):2770–2774. Braun MM, et al. 1995. A cohort study of twins and cancer. Cancer Epidemiol Biomarkers Prev 4:469–473. Brennan P, Hsu CC, Moullan N, Szeszenia-Dabrowska N, Lissowska J, Zaridze D, Rudnai P, Fabianova E, Mates D, Bencko V, Foretova L, Janout V, Gemignani F, Chabrier A, Hall J, Hung RJ, Boffetta, P, Canzian F. 2005. Effect of cruciferous vegetables on lung cancer in patients stratified by genetic status: a mendelian randomisation approach. Lancet 366(9496): 1558–1560. Brennan P, Lewis S, Hashibe M, Bell DA, Boffetta P, Bouchardy C, Caporaso N, Chen C, Coutelle C, et al. 2004. Pooled analysis of alcohol dehydrogenase and head and neck cancer: A HuGe review. Am J Epidemiol 159(1):1–16. Bretsky P, Haiman CA, Gilad S, Yahalom J, Grossman A, et al. 2003. The relationship between twenty missense ATM variants and breast cancer risk:the multiethnic cohort. Cancer Epidemiol Biomarkers Prev 12:733–738. Brockton N, Little J, Sharp L, Cotton SC. 2000. N–acetyltransferase polymorphisms and colorectal cancer:a HuGE review. Am J Epidemiol 151:846–861. Bromen K, Pohlabein H, Jahn I, Ahrens W, Jockel K-H. 2000. Aggregation of lung cancer in families: Results from a population-based case-control study in Germany. Am J Epidemiol 152(6):497–505. Burchell B. 2003. Genetic variation of human UDP-glucuronosyltransferase: implications in disease and drug glucuronidation. Am J Pharmacogen 2003;3(1):37–52. Butkiewicz D, Rustin M, Enewold L, Shields PG, Chorazy M, Harris CC. 2001. Genetic polymorphisms in DNA repair genes and risk of lung cancer. Carcinogenesis 32(4):593–597. Butsch Kovacic M, Martin M, Gao X, Fuksenko T, Chen CJ, Cheng YJ, Chen JY, Apple R, Hildesheim A, Carrington M. 2005. Variation of the Killer Cell Immunoglobulin-Like Receptors and HLA-C Genes in Nasopharyngeal Carcinoma. Cancer Epidemiol Biomarkers Prev 14(11):2673– 2677. Cai Q, Shu XO, Jin F, Dai Q, Wen W, Cheng JR, Gao YT, Zheng W. 2003. Genetic polymorphisms in the estrogen receptor alpha gene and risk of breast cancer: Results from the Shanghai Breast Cancer Study. Cancer Epidemiol Biomarkers Prev 12(9):853–859. Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ. 2003. Overweight, obesity and mortality from cancer in a prospectively studied cohort of US adults. New Engl J Med 348(17):1625–1638.

592

PART III: THE CAUSES OF CANCER

Cajas-Salazar N, Au WW, Zwischenberger JB, Sierra-Torres CH, Salama SA, Alpard SK, Tyring SK. 2003. Effect of epoxides hydrolase polymorphisms on chromosome aberrations and risk of cancer. Cancer Genet Cytogenet 145(2):97–102. Campa D, Zienolddiny S, Maggini V, Skaug V, Haugen A, Canzian F. 2004. Association of a common polymorphism in the cyclooxygenease 2 gene with risk of non–small cell lung cancer. Carcinogenesis 5(2):229–235. Cannon-Albright LA, Thomas A, Godlgar DE, Gholami K, et al. 1994. Familiality of cancer in Utah. Cancer Res 54:2378–2385. Caporaso N, Goldstein A. 1995. Cancer genes: Single and susceptibility: Exposing the difference. Pharmacogenetics 5:59–63. Caporaso N, Landi MT, Vineis P. 1991. Relevance of metabolic polymorphisms to human carcinogenesis. Pharmacogenetics 1:4–29. Caporaso N. 2002. Why have we failed to find the low penetrance genetic constituents of common cancers? Cancer Epidemiol Biomarkers Prev. 11(12):1544–1549. Caporaso N, Rothman N, Wacholder S. 1999. Case-control studies of common alleles and environmental factors. J Natl Cancer Inst 26:25–30. Cardon LR, Abecasis. 2003. Using haplotype blocks to map human complex trait loci. Trends Genet 19(3):135–140. Cardon LR, Bell JI. 2001. Association study designs for complex disease. Nat Rev Genet 2:91–99. Cargill M, Altshuler D, Ireland J, Sklar P, Ardlie K, et al. 1999. Characterization of single-nucleotide polymorphisms in the coding regions of human genes. Nat Genet 22:231–238. Carlson CS, Eberle MA, Kruglyak L, Nickerson DA. 2004. Mapping complex disease loci in whole-genome association studies. Nature 429:446–452. Caspi A, Sugden K, Moffitt TE, Taylor A, Craig IW, Harrington H, McClay J, et al. 2003. Influence of life stress on depression:Moderation by the polymorphism in the 5-HTT gene. Science 301:386–389. Cerosaletti KM, Morrison VA, Sabath DE, Willerford DM, Concannon P. 2002. Mutations and molecular variants of the NBS1 gene in non-Hodgkin lymphoma. Gene Chromosomes Cancer 35(3):282–286. Chang HW, Chan A, Kwong DLW, et al. 2003. Detection of hypermethylated RIZ1 gene in primary tumor, mouth and throat rinsing fluid, Nasopharyngeal swab and peripheral blood of NPC patient. Clinical Cancer Res 9:1033–1038. Chang BL, Zheng SL, Isaacs SD, Turner A, Hawkins GA, Wiley KE, Bleeker ER, Walsh PC, et al. 2003. Polymorphisms in the CYP1A1 gene are associated with prostate cancer risk. Int J Cancer 106(3):376–378. Chang BL, Zheng SL, Issacs SD, Turner A, Hawkins GA, Wiley KE, Bleecker ER, Walsh PC, Meyers DA, Issacs WB, Xu J. 2003. Polymorphisms in the CYP1B1 gene are associated with increased risk of prostate cancer. Br J Cancer 89(8):1524–1529. Chang BL, Zhang SL, Isaacs SD, Turner AR, Bleeker ER, Walsh PC, Meyers DA, Isaacs WB, Xu J. 2003. Evaluation of SRD5A2 sequence variants in susceptibility to hereditary and sporadic prostate cancer. Prostate 56(1):37–44. Channock S, Wacholder S. 2002. One gene and one outcome? No way. Trends Mol Med 8(6):266–269. Chapman JM, Cooper JD, Todd JA, Clayton DG. 2003. Detecting disease associations due to linkage disequilibrium using haplotype tags: a class of tests and the determinants of statistical power. Hum Hered 56:18–31. CHEK2–Breast Cancer Consortium. 2002. Low-penetrance susceptibility to breast cancer due to CHEK2*1100delC in noncarriers of BRCA1 or BRCA2 mutations. Nature Genet 31:55–59. Chen H, Lum A, Seifried A, Wilkens LR, Le Marchaud LL. 1999. Association of the NAD(P)H: quinone oxidoreductase 609C > T polymorphism with a decreased lung cancer risk. Cancer Res 59:3045–3048. Chen J, Gammon MD, Chan W, Palomeque C, Wetmur JG, et al. 2005. Onecarbon metabolism, MTHFR polymorphisms, and risk of breast cancer. Cancer Res 65(4):1606–1617. Cicek MS, Nock NL, Li L, Conti DV, Casey G, Witte JS. 2004. Relationship between MTHFR C677T and A1298C Genotypes and haplotypes and prostate cancer risk and aggressiveness. Cancer Epidemiol Biomarkers Prev 2004; 13(8). Cheon KT, Choi KH, Lee HB, Park SK, Rhee YK, Lee YC. 2000. Gene polymorphisms of endothelial nitric oxide synthase and angiotensinconverting enzyme in patients with lung cancer. Lung 178:351–360. Chevrier I, Stucker I, Houllier AM, Cencee S, Beaune P, Laurent-Puig P, Loriot MA. 2003. Myeloperoxidase:new polymorphisms and relation with lung cancer risk. Pharmacogenetics 13(12):729–739. Cho EY, Hildesheim A, Chen CJ, Hsu MM, Chen IH, Mittl BF, Levine PH, Liu MY, Chen JY, Brinton LA, Cheng YJ, Yang CS. 2003. Nasopharyngeal carcinoma and genetic polymorphisms of DNA repair enzymes XRCC1 and hOGG1. Cancer Epidemiol Biomarkers Prev 12:1100–1104. Chokkalingam AP, McGlynn KA, Gao Y-T, Pollak M, Deng J, et al. 2001. Vitamin D receptor gene polymorphisms, insulin-like growth factors and

prostate cancer risk: A population-based case-control study in China. Cancer Res 61:4333–4336. Chung T-W, Lee Y-C, Ko J-H, Kim C-H. 2003. Hepatitis B virus protein modulates the expression of PTEN by inhibiting the function of p53 a transcriptional activator in liver cells. Cancer Res 63:3453–3458. Clarke PA, George ML, Easdale GS, Cunningham D, Swift RI, Hill ME, Tait DM, Workman P. 2003. Molecular Pharmacology of cancer therapy in human colorectal cancer by gene expression profiling. Cancer Res 63: 6855–6863. Coe BP, Lee EH, Chi B, Girard L, Minna JD, Gazdar AF, Lam S, Macaulay C, Lam WL. 2006. Gain of a region on 7p22.3 containing MAD1L1 is the most frequent event in small-cell lung cancer lines. Gene Chromosomes Cancer 45(1):11–19. Comings DE. 1998. Why different rules are required for polygenic inheritance: Lessons from studies of the DRD2 gene. Alcohol 16(1):61–70. Conney AH. 2003. Enzyme induction and dietary chemicals as approaches to cancer chemoprevention: The seventh DeWitt S Goodman Lecture. Cancer Res 63:7005–7031. Cotton SC, Sharp L, Little J, Brockton N. 2000. Glutathione S-transferase polymorphisms and colorectal cancer: A HUGE Review. Am J Epid 151(1):7–32. Coughlin SS, Hall IJ. 2002. Glutathione S-transferase polymorphisms and risk of ovarian cancer. Genet Med 4(4):250–257. Coughlin SS, Hall IJ. 2002. A review of genetic polymorphisms and prostate cancer risk. Ann Epidemiology 12(3):182–196. Cox DG, Hankinson SE, Hunter DJ. 2006. Polymorphisms in the ICAM gene locus and not associated with breast cancer risk. Cancer Epidemiol Biomakers Prev 15(1):178–179. Crabtree MD, Fletcher C, Churchman M, Hodgson SV, Neale K, Phillips RK, Tomlinson IP. 2004. Analysis of candidate modifier loci for the severity of colonic familial adenomatous polyposis, with evidence for the importance of the N-acetyl transferases. Gut 53(2):271–276. Crawford EL, Khuder SA, Durham SJ, Frampton M, Utell M, Thilly W, Weaver DA, Ferencak WJ, Jenning CA, Hammersley JR, Olson DA, Willey JC. 2000. Normal bronchial epithelial cell expression of glutathione transferase P1, glutathione transferase M3, and glutathione peroxidase is low in subjects with bronchogenic carcinoma. Cancer Res 60(6):1609– 1618. Crowley JJ, Oslin DW, Patkar AA, Gottheil E, et al. 2003. A genetic association study of the mu opiod receptor and severe opiod dependence. Psychiatr Genet 13(3):169–173. Cunningham LM, Chapman C, Dunstan R, Bell MC, Joske DJ. 2003. Polymorphisms in the interleukin 10 polymorphism are associated with susceptibility to NHL. Leuk Lymphoma 44(2):251–255. Curtin K, Bigler J, Slattery ML, Caan B, Potter JD, Ulrich CM. 2004. MTHFR C677T and A1298C polymorphisms:diet, estrogen and risk of colon cancer. Cancer Epidemiol Biomarkers Prev 13:285–292. Cybulski C, Huzarski T, Gorski B, Masojc B, Mierzejewski M, Debniak T, Gliniewicz B, Matyjasik J, Zlowocka E, et al. 2004a. A novel founder CHEK2 mutations is associated with increased prostate cancer risk. Cancer Res 64:2677–2679. Cybulski C, Gorski B, Debniak T, Gliniewicz B, Mierzejewski M, Masojc B, Jakubowska A, Matyjasik J, Zlowocka E, et al. 2004b. NBS1 is a prostate cancer susceptibility gene. Cancer Res 64:1215–1219. Dahia PLM, Hao K, Rogus J, Colin C, Pujana MAG, et al. 2005. Novel Pheochromocytoma Susceptibility Loci Identified by Integrative Genomics. Cancer Res 65(21). Dally H, Edler L, Jager B, Schmezer P, Spiegelhalder B, Dienemann H, Drings P, Schultz V, Kayser K, Bartsch H, Risch A. 2003. The CYP3A4*1B allele increases risk for small cell lung cancer:effect of gender and smoking dose. Pharmacogenetics 13(10):607–618. D’Alo F, Voso MT, Guidi F, Massini G, Scardocci A, Sica S, Pagano L, Hohaus S, Leone G. 2004. Polymorphisms of CYP1A1 and GST and susceptibility to adult acute myeloid leukemia. Haematologica 89(6): 664–670. Daly AK. 2003. Candidate gene case-control studies. Pharmacogenomics 4(2):127–139. David-Beabes GL, London SJ. 2001. Genetic polymorphism of XRCC1 and lung cancer risk among African-Americans and Caucasians. Lung Cancer 34(3):333–339. D’errico, Malats N, Boffetta P. 1999. Review of studies of selected metabolic polymorphisms and cancer. In: Ryder W, ed. Metabolic polymorphisms and Susceptibility to Cancer. IARC Scientific Publications No. 148. Lyon: IARC, pp. 323–394. DeJong MM, Nolte IM, de Vries EG, Schaapveld M, Kleibeuker JH, Oosterom E, Oosterwijk JC, et al. 2003. The HLA class III subregion is responsible for an increased breast cancer risk. Hum Mol Genetics 12(18): 2311–2319.

Genetic Modifiers of Cancer Risk DeRoos AJ, Rothman N, Inskip PD, Linet MS, Shapiro WR, Selker RG, Fine HA, Black PM, Pittman GS, Bell DA. 2003. Genetic polymorphisms in GSTM1, -P1, -T1, and CYP2E1 and the risk of adult brain tumors. Cancer Epidemiol Biomarkers Prev 12(1):14–22. DeVivo I, Hankinson SE, Colditz GA, Hunter DJ. 2003. A functional polymorphism in the progesterone receptor gene is associated with an increase in breast cancer risk. Cancer Res 63:5236–5238. Diettzsch E, Laubscher R, Parker MI. 2003. Esophageal cancer risk in relation to GGC and CAG trinucleotide repeat lengths in the androgen receptor gene. Int J Cancer 107(1):38–45. Doll R. 1996. Nature and nurture:possibilities for cancer control. Carcinogenesis 17(2):177–184. Dragani TA, Hirohashi S, Juji T, Kawajiri K, et al. 2000. Population-based mapping of pulmonary adenoma susceptibility 1 locus. Cancer Res 60:5017–5020. Dragani TA, Manenti G, Pierotti MA. 1996. Polygenic inheritance of predisposition to lung cancer. Ann 1st Super Sanita 32(1):145–150. Duan J, Wainwright MS, et al. 2003. Synonymous mutations in the human dopamine receptor D2. DRD2. affect mRNA stability and synthesis of the receptor. Hum Mol Genet 12(3):205–216. DuBois RN. 2003. Evaluation of the whole prostaglandin biosynthetic pathway in lung cancer. Clin Cancer Res 9:1577–1578. Duell EJ, Holly EA, Bracci PM, Wiencke JK, Kelsey KT. 2002. A populationbased study of the Arg399Gln polymorphism in X-ray repair crosscomplementing group 1 (XRCC1) and risk of pancreatic adenocarcinoma. Cancer Res 62(16):4630–4636. Duell EJ, Holly EA, Bracci PM, Liu M, Wiencke JK, Kelsey KT. 2002. A population-based case-control study of polymorphisms in carcinogenmetabolizing genes, smoking, and pancreatic cancer. J Natl Cancer Inst 94(4):297–306. Dugay Y, McGrath M, Lepine J, Gagne J-F, Hankinson SE, et al. 2004. The functional UGT1A1 promoter polymorphism decreases endometrial cancer risk. Cancer Res 1202–1207. Dunning AM, Chiano M, Smith NR, Dearden J, et al. 1997. Common BRCA1 variants and susceptibility to breast and ovarian cancer in the general population. Hum Mol Genet 6:285–289. Dunning AM, Healey CS, Pharoah PDP, Tearne MD, Ponder BA, Easton DF. 1999. Systematic review of genetic polymorphisms and breast cancer risk. 8:843–854. Egan KM, Cai Q, Shu X-O, Jin F, Zhu T-L, Dai Q, Gao Y-T, Zheng W. 2004. Genetic polymorphisms in GSTM1, GSTP1, and GSTT1 and the risk for breast cancer:results from the Shanghai Breast Cancer Study and MetaAnalysis. Cancer Epid Biomarkers Prev 13:197–204. Egeblad M, Werb Z. 2002. New functions for the matrix metalloproteinases in cancer progression. Nat Rev 2:161–173. Elahi A, Bendaly J, Zheng Z, Muscat JE, Richie JP Jr., Schantz SP, Lazarus P. 2003. Detection of UGT1A10 polymorphisms and their association with orolaryngeal cancer risk. Cancer 98(4):872–880. Elkin PL. 2003. Primer on medical genomics. Part V: Bioinformatics. Mayo Clin Proc 78:57–64. El-Omar EM, Carrington M, Chow W-H, McColl EEL, Bream JH, Young HA, et al. 2002. The role of interleukin-1 polymorphisms in the pathogenesis of gastric cancer. Nature 404:398–402. El-Omar EM, Rabkin CS, Gammon MD, Vaughan TL, Risch HA, Schroenberg JB, et al. 2003. Increased risk of noncardia gastric cancer associated with proinflammatory cytokine gene polymorphisms. Gastroenterology 124(5):1193–1201. Engel LS, Taioli E, Pfeiffer R, Garcia-Closas M, Marcus PM, Lan Q, Boffetta P, Vineis P, Autrup H, Bell DA, Branch RA, Brockmoller J, Daly AK, Heckbert SR, Kalina I, Kang D, Katoh T, Lafuente A, Lin HJ, Romkes M, Taylor JA, Rothman N. 2002. Pooled analysis and meta-analysis of glutathione S-transferase M1 and bladder cancer: A HuGE review. Am J Epidemiol 156(5):492. Ergul E, Sazci A, Ytkan Z, Canturk NZ. 2003. Polymorphisms in the MTHFR gene are associated with breast cancer. Tumour Biol 24(6): 286–290. Esteller M, Garcia A, Martinez JM, Xercavins J, Reventos J. 1997. Germ line polymorphisms in cytochrome-P450 1A1 (C4887CYP1A1) and metylenetetrahydrofolate reductase (MTHFR) genes and endometrial cancer susceptibility. Carcinogenesis 18(12):2307–2311. Etzel CJ, Amos CI, Spitz MR. 2003. Risk for smoking related cancer among relatives of lung cancer patients. Cancer Res 63:8531–8535. Evans AJ, Henner WD, Eilers KM, Montalto MA, Wersinger EM, Andersen PE, Cohen JI, Everts EC, McWilliams JE, Beer TM. 2004. Polymorphisms of GSTT1 and related genes in head and neck cancer risk. Head Neck 26(1):63–70. Evans WE, McLeod HL. 2003. Pharmacogenomics–drug disposition, drug targets and side effects. N Engl J Med 348(6):538–549.

593

Ezer R, Alonso M, Pereira E, Kim M, Allen JC, Miller DC, Newcomb EW. 2002. Identification of glutathione S-transferase (GST) polymorphisms in brain tumors and association with susceptibility to pediatric astrocytomas. J Neurooncol 59(2):123–124. Fang J-L, Beland FA, Doerge DR, Wiener D, et al. 2002. Characterization of benzo(a)pyrene-trans-7,8-dihydrodiol glucuronidation by human tissue microsomes and overexpressed UDP-glucuronyltransferase enzymes. Cancer Res 62:1978–1986. Fang S, Jin X, Wang R, Li Y, Guo W, et al. 2005. Polymorphisms in the MMP1 and MMP3 promoter and non-small cell lung cancer in North China. Carcinogenesis 26(2):481–486. Fang W, Hartmann N, Chow DT, Riegal AT, Wellstein A. 1992. Pleiotrophin stimulates fibroblasts and endothelial and epithelial cells and is expressed in human cancer. J Biol Chem 267:25889–25897. Ferreira PM, Medeiros R, Vasconcelos A, Costa S, Pinto D, Morais A, Oliveira J, Lopes C. 2003. Association between CYP2E1 polymorphisms and susceptibility to prostate cancer. Eur J Cancer Prev 12(3):205–211. Fletcher O, Gibson L, Johnson N, et al. 2005. Polymorphisms and Circulating levels in the insulin-like growth factor system and risk of breast cancer: a systematic review. Cancer Epidemiol Biomarkers Prev 4(1):2–19. Foppa I, Spiegelman D. 1997. Power and sample size calculations for case–control studies of GE interactions with polycotomous exposure variables. Am J Epidemiol 146(7):596–604. Fraumeni JF. 1997. Clinical patterns of familial cancer. In: Mulvihill JJ, Miller RW, Fraumeni JF, Jr., eds. Genetics of Human Cancer. New York: Raven Press, pp. 223–233. Freedman ML, Reich D, Penney KL, McDonald GJ, Mignault AA, Patterson N, Gabriel SB, Topol EJ, et al. 2004. Assessing the impact of population stratification on genetic association studies. Nat Genet 36(4):388–393. Friedberg EC, Bond JP, Burns DK, Cheo DL, Greenblatt MS, Meira LB, Nahari D, Reis AM. 2000. Defective nucleotide excision repair in xpc mice and its association with cancer predisposition. Mutat Res 459(2): 99–108. Fung WLA, Risch H, McLaughlin J, Rosen B, Cole D, Vesprini D, Narod SA. 2003. The N314D Polymorphism of galactose-1-phosphate uridyl transferase does not modify the risk of ovarian cancer. Cancer Epidemiol Biomarkers Prev 12:678–680. Futreal PA, Coin L, Marshall M, Down T, Hubbard T, Wooster R, Rahman N, Stratton MR. 2004. A census of human cancer genes. Nat Rev 4:177–184. Gao S-W, Lange K. 2000. Genetic mapping of complex traits: Promises, problems and prospects. Theoretical Population Biol 57:1–11. Gao C, Wu J, Ding J, Zang Y, Li S, Su P, Hu X, Xu T, Toshiro T, Kazuo T. 2002. Polymorphism of MTHFR C677T and the risk of stomach cancer. Zhonghua Liu Xing Bing Xue Za Zhi 23(4):289–292. Garcia-Borron JC. 2003. Biochemical aspects of MC1R signaling through the camp pathway. Pigment Cell Res 16(5):7501. Garcia-Closas M, Wacholder S, Caporaso N, Rothman N. 2004. Inference issues in cohort and case-control studies of genetic effects and geneenvironment interactions. In: Khoury M, Little J, Burke W, eds. Human Genome Epidemiology. pp 127–144. Oxford University Press. Garcia-Closas M, Thompson WD, Robins JM. 1998. Differential misclassification and the assessment of gene-environment interaction in case-control studies. Am J Epidemiol 147:426–433. Garcia-Closas M, Rothman N, Lubin J. 1999. Misclassification in case-control studies of gene–environment interactions:assessment of bias and sample size. Cancer Epidemiol Biomarkers Prev 8:1043–1050. Garcia-Martin E, Martinez C, Ladero JM, Rodriguez-Lescure A, Agundez JA. 2002. Influence of cytochrome P450 CYP2C9 in lung cancer risk. Cancer Lett 180(1):41–46. Garte and Taioli. 2000. Ethnic and geographic distribution of metabolic gene polymorphisms in a large sample of healthy controls:results from GSEC study. AACR Proceedings, vol 41. Geisler SA, Olshan AF. 2001. GSTM1, GSTT1 and the risk of squamous cell carcinoma of the head and neck: A mini-HUGE review. Am J Epidemiol 154(2):95–105. Gelmann EP, Steadman DJ, Ma J, Ahronovitz N, Voeller HJ, Swope S, et al. 2002. Currence of NKX3.1 C154T polymorphism in men with and without prostate cancer and studies of its effect on protein function. Cancer Research 62:2654–2659. Gemignani F, Moreno V, Landi S, Moullan N, Chabrier A, Gutierrez-Enrqiuez S, Hall J, Guiano E, Pinado MA, Capella G, Canzian F. 2004. A TP53 polymorphism is associated with increased risk of a colorectal cancer and reduced levels of TP53 mRNA. Oncogene 23(10):1954–1956. Ghilardi G, Biondi ML, Caputo M, Leviti S, DeMonti M, Guagnellini E, Scorza R. 2002. A single nucleotide polymorphism in the matrix metalloproteinase-3 promoter enhances breast cancer susceptibility. Clinical Cancer Res 8:3820–3823.

594

PART III: THE CAUSES OF CANCER

Giovannucci E, Chen J, Smith-Warner SA, Rimm EB, Fuchs CS, Palomeque C, Willett WC, Hunter DJ. 2003. Methylenetetrahydrofolate reductase, alcohol dehydrogenase, diet, and risk of colorectal adenomas. Cancer Epidemiol Biomarkers Prev 12(10):970–979. Gibson G, Muse SV. 2002. A primer of genome science. Sunderland, Mass.: Sinauer Associates. Giovannucci E, Pollak MN, Platz EA, Willett WC, et al. 2000. A prospective study of plasma insulin-like growth factor-1 and binding protein-3 and risk of colorectal neoplasia in women. Cancer Epidemiol Biomarkers Prev 9(4):345–349. Glebov OK, Rodriguez LM, Nakahara K, Jenkins J, et al. 2003. Distinguishing right from left colon by the pattern of gene expression. Cancer Epidemiol Biomarkers Prev 12:755–762. Godard S, Getz G, Delorenzi M, Farmer P, Kobayashi H, Desbaillets I, Nozaki M, et al. 2003. Classification of human astrocyic gliomas on the basis of gene expression:a correlated group of genes with angiogenic activity emerges as a stron predictor of genotype. Cancer Res 63: 6613–6625. Gold B, Kalush F, Bergeron J, Scott K, Mitra N, Wilson K, et al. 2004. Estrgen receptor genotypes and haplotypes associated with breast cancer risk. Cancer Res 64:8891–8900. Goldgar DE, Easton DF, Cannon-Albright LA, Skolnick MH. 1994. Systematic population-based assessment of cancer risk in first-degree relatives of cancer probands. J Natl Cancer Inst 86:1600–1607. Goldstein A, Andrieu N. 1999. Detection of interaction involving identified genes:available study designs. J Natl Cancer Inst Monogr 26:49–54. Gonzalez FJ, Kimura S. 1999. Role of knockout mice in understanding the mechanisms of chemical toxicity and carcinogenesis. Cancer Lett 143:199–204. Goode EL, Ulrich CM, Potter JD. 2002. Polymorphisms in DNA repair genes and associations with cancer risk. Cancer Epidemiol Biomarkers Prev 11(12):1513–1530. Goode EL, Potter JD, Bigler J, Ulrich CM. 2004. Methionine synthase D919G polymorphism, folate metabolism, and colorectal adenoma risk. Cancer Epidemiol Biomarkers Prev 13:157–163. Goodman MT, McDuffie K, Kolonel LN, Tereda K, Donlon TA, Wilkens LR, Guo C, LeMarchand L. 2001. Case-control study of ovarian cancer and polymorphisms in genes involved in catecholestrogen formation and metabolism. Cancer Epid Biomarkers Prev 10(3):209–216. Green ED, Waterston RH. 1991. The Human Genome Project. JAMA 266(14):1966–1975. Greenspan RJ. 2003. The flexible genome. Nat Rev 2:383–387. Gsur A, Zidek T, Schnattinger K, Feik E, Haidinger G, Hollaus P, MohnStaudner A, et al. 2003. Association of microsomal epoxides hydrolase polymorphisms and lung cancer risk. Br J Cancer 89(4):702–706. Hachey DL, Dawling S, Roodi N, Parl FF. 2003. Sequential action of Phase 1 and II enzymes cytochrome P450 1B1 and glutathione S-transferase P1 in mammary estrogen metabolism. Cancer Res 63:8492–8499. Haddow JE, Palomaki GE. 2004. ACCE: A model process for evaluating data on emerging genetic tests. In: Khoury M, Little J, Burke W, eds. Human Genome Epidemiology. New York: Oxford University Press, pp. 217–233. Han J, Colditz GA, Liu JS, Hunter DJ. 2005. Genetic variation in XPD, sun exposure, and risk of skin cancer. Cancer Epidemiol Biomarkers Prev 14(6):1539–1544. Han C, McGue M, Iacono WG. 1999. Lifetime tobacco, alcohol and other substance use in adolescent Minnesota twins: Univariate and multivariate behavioral genetic analyses. Addiction 94(7):981–993. Han J, Hankinson SE, Ranu H, De Vivo I, Hunter DJ. 2004. Polymorphisms in DNA double strand break repair genes and breast cancer risk in the nurses’ health study. Carcinogenesis 25(2):189–195. Han J, Colditz GA, Samson LD, Hunter DJ. 2004. Polymorphisms in DNA double strand break repair genes and skin cancer risk. Cancer Res 64:3009–3013. Hanloser D, Petersen GM, Rabe K, Snow K, Lindor NM, et al. 2003. The APC E131Q in adenomatous polyps and colorectal cancers. Cancer Epidemiol Biomarkers Prev 12(10):1023–1028. Hardy LC, Caporaso NE, Hayes RB, Winn DM, Bravo-Otero E, Blot WJ, Kleinman DV, Brown LM, Armenian HK, Fraumeni JF Jr, Shields PG. 1997. Alcohol dehydrogenase 3 genotype and risk of oral cavity and pharyngeal cancers. J Natl Cancer Inst. 89(22):1698–1705. Hariri AR, Weinberger DR. 2003. Imaging genomics. Br Med Bull 65:259–270. Harries LW, Stubbins MJ, Forman D, Howard GC, Wolf CR. 1997. Identification of genetic polymorphisms at the glutathione S-transferase Pi locus and association with susceptibility to bladder, testicular and prostate cancer. Carcinogenesis 18(4):641–644. Hazra A, Chamberlain RM, Grossman B, Zhu Y, Spitz MR, Wu X. 2003. Death receptor 4 and bladder cancer risk. Cancer Res 63:1157–1159.

Heather HN, Kelsey KT, Mott LA, Karagas MR. 2002. The XRCC1 Arg399Gln polymorphism, sunburn, and non-melanoma skin cancer evidence for gene-environment interaction. Cancer Res 62:152–155. Hashibe M, Brennan P, Strange RC, Bhisey R, et al. 2003. Meta- and pooled analyses of GSTM1, GSTT1, GSTP1, and CYP1A1 genotypes and risk of head and neck cancer. Cancer Epidemiol Biomarkers Prev 12:1509–1517. Hayward NK. 2003. Genetics of melanoma predisposition. Oncogene 22(20):3053–3062. Hecht SS, Hochalter JB, Villalta PW, Murphy SE. 2000. 2¢ hydroxylation of nicotine by cytochrome P4502A6 and human liver microsomes: Formation of a lung carcinogen precursor. Proc Natl Acad Sci USA 97(23): 12493–12497. Hecht SS. 1999. Tobacco smoke carcinogens and lung cancer. J Natl Cancer Inst 91(14):1194–1210. Heinonen OP, Albanes D, Virtamo J, Taylor PR, Huttman JK, et al. 1998. Prostate cancer and supplementation with alpha-tocopherol and beta– carotene:incidence and mortality in a controlled trial. J Natl Cancer Inst 90:440–446. Heijmans BT, Boer JMA, Suchiman HED, Cornelisse CJ, Westendorp RGJ, Kromhout D, Feskens EJM, Slagboom PE. 2003. A common variant of the MTHFR gene (1p36) is associated with increased risk of cancer. Cancer Res 63:1249–1253. Hellard SL, Ballereau SJ, Visscher PM, Torrance HS, Pinson J, Morris SW, Thomson ML, et al. 2002. SNP genotyping on pooled DNAs:comparison of genoyping technologies and semi-automated method for DNA storage. Nucleic Acids Res 30(15):1–10. Hemminki K, Vaittinen P. 1999. Familial cancers in a nationwide family cancer database: Age distribution and prevalence Eur J Cancer 35(7): 1109–1117. Hemminki K, Li X, Czene K. 2004. Familial risk of cancer:data for clinical counseling and cancer genetics. Int J Cancer 108:109–114. Herman JG, Umar A., Polyak K, Graff JR, Ahuja N, Issa JP, Markowitz S, et al. 1998. Incidence and functional consequences of hMLH1 promoter hypermethylation in colorectal carcinoma. Proc Natl Acad Sci USA 95:6870–6875. Hirschhorn JN, Lohmueller K, Byrne E, Hirschhorn K. 2002. A comprehensive review of genetic association studies. Genet Med 4(2):45–61. Hirunsatit R, Kongruttanachok N, Shotelersuk K, et al. 2003. Polymeric immunoglobulin receptor polymorphism and the risk of nasopharyngeal cancer. BMC Genet 4(1):3. Hishida A, Matsuo K, Hamajima N, Ito H, Ogura M, Kagami Y, Taji H, Morishima Y, Emi N, Tajima K. 2003. Associations between polymorphisms in the thymidylate synthase and serine hydroxymethyltransferase genes and susceptibility to malignant lymphoma. Haematologica 88(2): 159–166. Hoggart CJ, Parra EJ, Shriver MD, Bonilla C, Kittles RA, Clayton DG, McKeigue PM. 2003. Control of confounding of genetic associations in stratified populations. Am J Hum Genet 72(6). Hong Y, Miao X, Zhang X, Ding F, et al. 2005. The Role of p53 and MDM2 polymorphisms in the risk of esophageal squamous cell carcinoma. Cancer Res 65(20):9573–9582. Hoover RN. 2000. Cancer—Nature, nurture, or both. N Engl J Med 343(2): 135–136. Houlston RS. 1999. Glutathione S-transferase M1 status and lung cancer risk: A meta-analysis. Cancer Epidemiol Biomarkers Prev 8(8):675–682. Houlston RS, Tomlinson IPM. 2000. Detecting low penetrance genes in cancer:the way ahead. J Med Genetics 37:161–167. Howell WM, Calder PC, Grimble RF. 2002. Gene polymorphisms, inflammatory diseases, and cancer. Proc Nutr Soc 61(4):447–456. Huang YES, Chern HD, Wu JC, Chao Y, Huang YH, Chang FY, Lee SD. 2003. Polymorphism of the NAT2 gene, red meat intake and susceptibility to HCC. Am J Gastroenterol 98(6):1417–1422. Humar B, Graziano F, Cascinu S, Catalano V, Ruzzo AM, Magnani M, Toro T, Burchill T, Futschik ME, Merriman T, Guilford P. 2002. Association of CDH1 haplotypes with susceptibility to sporadic diffuse gastric cancer. Oncogene 21(53):8192–8195. Hung RJ, Brennan P, Malaveille C, Porru S, Donato F, Boffetta P, Witte J. 2004. Using hierarchical modeling in genetic association studies with multiple markers: application to a case-control study of bladder cancer. Cancer Epidemiol Biomarkers Prev 13(6):1013–1021. Hung RJ, Boffetta P, Butkiewicz D, Cascorbi I, Clapper ML, et al. 2003. CYP1A1 and GSTM1 genetic polymorphisms and lung cancer risk in Caucasian non-smokers: A pooled analysis. Carcinogenesis 24(5):875–882. Hung RJ, Boffetta P, Brennan P, Malaveille C, Gelatti U, Placidi D, Car A, Hautefeuille A, Porru S. 2004. Genetic polymorphisms of MPO, COMT, MnSOD, NQO1, interactions with environmental exposures and bladder cancer risk. Carcinogenesis 25(6):973–978.

Genetic Modifiers of Cancer Risk Hung RJ, Boffetta P, Brennan P, Malaveille C, Hautefeuille A, Donato F, Gelatti U, Spaliviero M, Placidi D, Carta A, Scotto di Carlo A, Porru S. 2004. GST, NAT, SULT1A1, CYP1B1 genetic polymorphisms, interactions with environmental exposures and bladder cancer risk in a high-risk population. Int J Cancer 110(4):598–604. Hunter DJ. 1997. Methodological issues in the use of biological markers in cancer epidemiology:cohort studies. In: Toniolo P, Boffetta P, Shuker DEG, Rothman N, Hulka B, Pearce N. eds. Application of Biomarkers in Cancer Epidemiology. IARC Scientific Publications No. 142 Lyon: IARC pp. 39–46. Infante-Rivard C. 2003. Diagnostic x-rays, DNA repair genes and childhood acute lymphoblastic leukemia. Health Physics 85(1):60–64. Ioannidis JA, Gwinn M, Little J, Higgans JPT, Bernstein JL, Boffetta P, et al. 2006. A road map for efficient and reliable human genome epidemiology. Nature Genetics 38(1):3–5. Ioannidis JPA, Ntzani EE, et al. 2001. Replication validity of genetic association studies. Nat Genet 29:244–245. Iughetti P, Suzuki O, et al. 2001. A polymorphism of endostatin an angiogenesis inhibitor, predisposes for the development of prostatic adenocarcinoma. Cancer Res 61:7375–7378. Jackson PE, Kuang SY, Wang JB, Strickland PT, Munoz A, Kensler TW, Qian GS, Groopman JD. 2003. Prospective detection of codon 249 mutations in p53 in plasma of hepatocellular carcinoma patients. Carcinogenesis 24(10):1657–1663. Jacques PF, Bostom AG, Willians RR, Ellison RC, Eckfeldt JH, Rosenberg IH, Selhub, Rozen R. 1996. Relation between folate status, a common mutation in methyltetrahyrofolate reductase and plasma homocysteine concentrations. Circulation 93:7–9. Jalbout M, Bouaouina N, Gargouri J, Corbex M, et al. 2003. Polymorphism of the stress protein HSP70–2 gene is associated with susceptibility to NPC. Cancer Lett 193(1):75–81. Jamroziak K, Mlynarski W, Balcerczak E, Mistygacz M, Trelinska J, Mirowski M, Bodalski J, Robak T. 2004. Functional C3435T polymorphism of MDR1 gene: An impact on genetic susceptibility and clinical outcome of childhood acute lymphoblastic leukemia. Eur J Haematol 72(5):314–321. Jee SH, Um SJ, Lee JE, Kim JH, Namkoong SE, Park JS. 2003. The effect of codon 98 of the FHIT gene on cervical cancer in Korean women. Int J Gynecological Cancer 13(6):843–848. Jefcoate CR, Liehr JG, Santen RJ, Sutter TR, Yager JD, et al. 2000. Tissue specific synthesis and oxidative metabolism of estrogens. JNCI Monograph No. 27. New York: Oxford University Press, pp. 95–112. John EM, Schwartz GG, Koo J, Van Den Berg D, Ingles SA. 2005. Sun Exposure, vitamin D receptor gene polymorphisms, and risk of advanced prostate cancer. Cancer Res 2005; 65(12). Jorde LB. 2001. Linkage disequilibrium and the search for complex disease genes. Genome Res 1435–1443. Kaklamani VG, Hou N, Bian Y, Reich J, Offit K, Michel LS, Rubenstein WS, Rademaker A, Pasche B. 2003. TGFBR1*6A and cancer risk: A Metaanalysis of seven case-control studies. J Clin Oncol 21(17):3236–3243. Kallianpur AR, Hall LD, Yadav M, et al. 2004. Increased prevalence of the HFE C282Y hemochromatosis allele in women with breast cancer. Cancer Epidemiol Biomarkers Prev 13:205–212. Kammerer S, Roth RB, Reneland R, Marnellos G, et al. 2004. Large-scale association study identifies ICAM gene region as breast and prostate cancer susceptibility locus. Cancer Research 64:8906–8910. Kantarci OH, Lesnick TG, Yang P, Meyer RL, et al. 2002. Myeloperoxidase–463 (G > A) polymorphism associated with lower risk of cancer. Mayo Clin Proc 77(1):17–22. Kato S, Tajiri T, Matsukura N, Matsuda N, Tamiai N, Mamada H, Yoshida H, Kiyam T, Naito Z. 2003. Genetic polymorphisms of ADH2, CYP2E1 for liver cancer risk in HCV antibody-positive Japanese patients and the variations of CYP2E1 mRNA expression levels in the liver due to polymorphism. Scand J Gastroenterol 38(8):886–893. Kelada SN, Eaton DL, Wang SS, Rothman NR, Khoury MJ. 2003. The role for genetic polymorphisms in environmental health. Environ Health Perspect 111(8):1055–1064. Kelsey KT, Park S, Nelson HH, Karagas MR. 2004. A population-based casecontrol study of the XRCC1 Arg399gln polymorphism and susceptibility to bladder cancer. Cancer Epidemiol Biomarkers Prev 2004; 13(8). Kellermann G, Shaw CR, Luyten-Lkellermann M. 1973. Aryl hydrocarbon hydroxylase inducibility and bronchogenic carcinoma. New Engl J Med 289:934–937. Kerridge I, Lincz L, Scorgie F, Hickey D, Granter N, Spencer A. 2002. Association between xenobiotic gene polymorphisms and non-Hodgkins lymphoma risk. Br J Hematol 118(2):477–481. Khoury M, Little J, Burke W. 2004. Human genome epidemiology:scope and strategies. In: Khoury MJ, Little J, Burke W, eds. Human Genome Epidemiolgy. New York: Oxford University Press, pp. 3–16.

595

Khoury MJ, Adams MJ Jr., Flanders WD. 1988. An epigenetic approach to ecogenetics. Am J Hum Genet 42:89–95. Khoury MJ, Beaty T, Cohen BH. 1993. Fundamentals of Genetic Epidemiology. New York: Oxford University Press. Khoury M, Flanders W. 1996. Nontraditional epidemiologic approaches in the analysis of gene–environment interaction: Case-control studies with no controls. Am J Epidemiol 144:207–213. Khoury MJ. 1999. Human genome epidemiology. (HuGE): Translating advances in human genetics into population-based data for medicine and public health. Gen Med 1:71–73. Khoury MJ, McCabe LL, McCabe ERB. 2003. Population screening in the age of genomic medicine. New Engl J Med 348(1):50–58. Kibel AS, Suarez BK, Belani J, Oh J, Webster R, et al. 2003. CDKN1A and CDKN1B polymorphisms and risk of advanced prostate cancer. Cancer Res 63(9):2033–2036. Kihara M, Kihara M, Noda K. 1995. Risk of smoking for squamous and small cell carcinomas of the lung modulated by the combinations of CYP1A1 and GSTM1 gene polymorphisms in a Japanese population. Carcinogenesis 16(10):2331–2336. Kilpivaara O, Laiho P, Aaltonen LA, Nevanlinna H. 2003. CHEK2 1100delC and colorectal cancer. J Med Genet 40:110. Kimura ET, Nikiforova MN, Zhu Z, Knauf JA, Nikiforov YE, Fagin JA. 2003. High prevalence of BRAF mutations in thyroid cancer:genetic evidence for constituative activation of RET/PTC-RAS-BRAF signaling pathway in papillary thyroid cancer. Cancer Res 63:1454–1457. Kimura S, Gonzalez FJ. 2000. Applications of genetically manipulated mice in pharmacogenetics and pharmacogenomics. Pharmacology 61:147– 153. Kirk GD, Turner PC, Gong Y, Lesi OA, Mendy M, Goedert JJ, Hall AJ, Whittle H, Hainaut P, Montesano R, Wild CP. 2005. Hepatocellular carcinoma and polymorphisms in carcinogen-metabolizing and DNA repair enzymes in a population with aflatoxin exposure and hepatitis B virus endemicity. Cancer Epidemiol Biomarkers Prev 14(2):373–379. Kirkpatrick A, Bidwell J, van den Brule AJ, Meijer CJ, Pawade J, Glew S. 2004. TNF alpha polymorphism frequencies in HPV-associated cervical dysplasia. Gynecol Oncol 92(2):675–679. Kiyohara C, Yoshimasu K, Takayama K, Nakanishi Y. 2005. NQO1, MPO, and the risk of lung cancer: a Huge review. Genet Med 7(7):463. Koh W-P, Yuan J-M, Sun CC-L, van def Berg D, Seow A, Lee, H-P, Yu MC. 2003. Angiotensin I-converting enzyme. ACE. gene polymorphism and breast cancer risk among Chinese women in Singapore. Cancer Res 63:573–578. Kongruttanachok N, Sukdikul S, Setavarin S, Kerekhjanarong V, Supiyaphun P, Voravud N, Poovorawan Y, Mutirangura A. 2001. Cytochrome P450 2E1 polymorphism and nasopharyngeal carcinoma development in Thailand: a correlative study. BMC Cancer 1(1):4. Koushik A, Platt RW, Franco EL. 2004. P53 codon 72 polymorphism and cervical neoplasia: A meta-analysis review. Cancer Epidemiol Biomarkers Prev 13:11–12. Krajinovic M, Lamothe S, Labuda D, Lemieux-Blanchard E, Theoret Y, Moghrabi A, Sinnett D. 2004. Role of MTHFR genetic polymorphisms in the susceptibility to childhood lymphoblastic leukemia. Blood 103(1):252–257. Krontitis TG, Devlin B, Karp DD, Robert NJ, Risch N. 1993. An association between risk of cancer and mutations in the HRAS1 minisatellite locus. New Engl J Med 329(8):517–523. Kruglyak L. 2005. Power tools for human genetics Nature Genet 37(12): 1299–1300. Lamba JK, Lin YS, Schuetz EG, Thummel KE. 2002. Genetic contribution to the variable human CYP3A-mediated metabolism. Adv Drug Delivery Rev 54:1271–1294. Lan Q, Rothman N, Chow WH, Lissowska J, Doll MA, et al. 2003. No apparent association between NAT1 and NAT2 genotypes and risk of stomach cancer. Cancer Epidemiol Biomarkers Prev 12(4):384–386. Lan Q, Chow WH, Lissowska J, Hein DW, Buetow K, Engel LS, Ji B, Zatonski W, Rothman N. 2001. Glutathione S-transferase genotypes and stomach cancer in a population-based case-control study in Warsaw, Poland. Pharmacogenetics 11(8):655–661. Lander ES, Schork NJ. 1994. Genetic dissection of complex traits. Science 265:2037–2048. Landi S, Gemignani F, Moreno V, Gioia-Patricola L, Chabrier A, Guino E, Navarro M, de Oca J, Capella G, Canzian F, for the Bellvitage Colorectal Cancer Study Group. 2005. A comprehensive analysis of phase I and phase II metabolism gene polymorphisms and risk of colorectal cancer. Pharmacogenet Genomics 15(8):535–546. Landi S, Moreno V, Gioia-Patricola L, Guino E, Navarro M, de Oca J, Capella G, et al. 2003. Association of common polymorphisms in inflammatory gene interleukin (IL6, IL98), TNF alpha, NFKB1, and peroxisome

596

PART III: THE CAUSES OF CANCER

proliferator-activated receptor gamma with colorectal cancer. Cancer Res 63:3560–3566. Langholz B, Rothman N, Wacholder S, Thomas DC. 1999. Cohort Studies for characterizing measured genes. J Natl Cancer Inst Monogr 26:39–42. Lee KM, Choi JY, Park SK, Chung HW, Ahn B, Yoo KY, Han W, Noh DY, Ahn SH, Kim H, Wei Q, Kang D. 2005. Genetic polymorphisms of ataxia telangiectasia mutated and breast cancer risk. Cancer Epidemiol Biomarkers Prev 14(4):821–825. Lee SA, Kim JW, Roh JW, Choi JY, Lee KM, Yoo KY, Song YES, Kang. 2004. Genetic polymorphisms of GSTM1, p21, p53, and HPV infection with cervical cancer in Korean women. Gynecol Oncol 93(1):14–18. Lee SG, Kim B, Choi J, Kim C, Lee I, Song K. 2002. Genetic polymorphism of XRCC1 and risk of gastric cancer. Cancer Lett 187(1–2):53–60. Lee WJ, Brennan P, Boffetta P, London SJ, Benhamou S, et al. 2002. Microsomal epoxides hydrolase polymorphisms and lung cancer risk: A quantitative review. Biomarkers 7(3):230–241. LeMarchaud L, Guo C, Benhamou S, Bouchardy C, Cascorbi I, Clapper ML, et al. 2003. Pooled analysis of the CYP1A1 exon 7 polymorphism and lung cancer. United States. Cancer Causes Control 14(4):339–346. LeMarchand L, Hankin JH, Wilkens LR, Pierce LM, Franke A, Kolonel LN, Seifried A, Custer LJ, Chang W, Lum-Jones A, Donlon T. 2001. Combined effects of well-done red meat, smoking, and rapid Nacetyltransferase 2 and CYPA2 phenotypes in increasing colorectal cancer risk. Cancer Epidemiol Biomarkers Prev 10(12):1259–1266. Lerman C, Shields AE. 2004. Genetic testing for cancer susceptibility: the promise and the pitfalls. Nature Reviews 4:235–241. Lewis RC, Bostick RM, Xie D, Deng Z, Wargovich MJ, Fina MF, Roufai WM, Geisinger KR. 2003. Polymorphism of the cyclin D1 gene, CCND1, and risk for incident sporadic colorectal adenomas. Cancer Res 63(23): 8549–8953. Li FP, Lokich J, Lapey J, Neptune WB, Wilkins EW. 1978. Familial mesothelioma after intense asbestos exposure at home. JAMA 240(5):467. Li Z, Habuchi T, Mitsumori K, Kamoto T, Kinoshitu H, Segawa T, Ogawa O, Kato T. 2003. Association of V89L SRD5A2 polymorphism with prostate cancer development in a Japanese population. J Urol 169(9):2378–2381. Lichtenstein P, Holm NV, Verkasalo PK, Iliadou A, Kaprio J, et al. 2000. Environmental and heritable factors in the causation of cancer. New Engl J Med 343:78–85. Liede A, Zhang W, et al. 2003. Androgen receptor gene polymorphism and breast cancer susceptibility in the Philippines. Cancer Epidemiol Biomarkers Prev 12:848–852. Lillie EO, Bernstein L, Ingles SA, Gauderman WJ, Rivas GE, Gagalang V, Krontiris T, Ursin G. 2004. Polymorphism in the androgen receptor and mammographic density in women taking and not taking estrogen and progestin therapy. Cancer Res 64:1237–1241. Liu HJ, Guo XL, Dong M, Wang L, Yuan Y. 2003. Association between pepsinogen C gene polymorphism and genetic predisposition to gastric cancer. World J Gastroenterol 9(1):50–53. Lin J, Spitz MR, Wang Y, Schabath MB, Gorlov IP, Hernandez LM, Pillow PC, Grossman HB, Wu X. 2004. Polymorphisms of folate metabolic genes and susceptibility to bladder cancer: A case-control study. Carcinogenesis 25(9):1639–1647. Linder NM, Green MH, and the Mayo Familial Cancer Program. 1998. The Concise Handbook of Family Cancer Syndromes. J Natl Cancer Inst 90:4. Lindstedt BA, Ryberg D, Zienolddiny S, Khan H, Haugen A. 1999. Hras1 VNTR alleles as susceptibility markers for lung cancer: Relationship to microsatellite instability in tumours. Anticancer Res 19(6C):5523–5527. Lippman SM, Spitz MR. 2001. Lung cancer chemoprevention: An integrated approach. J Clin Oncol 19(18):74–82. Little J, Bradley L, Bray MS, Clyne M, Doeman J, Ellsworth DL. 2002. Reporting, appraising and integrating data on genotype prevalence and genedisease associations. Am J Epidemiol 156(4):300–310. Lohmueller KE, Pearce CL, Pike M, Lander ES, Hirschhorn JN. 2003. Metaanalysis of genetic association studies supports a contribution of common variants to susceptibility to common disease. Nat Genet 33:177–182. Longuemaux S, Delomenie C, Gallou C, Mejean A, Vincent-Viry M, et al. 1999. Candidate genetic modifiers of individual susceptibility to renal cell carcinoma:a study of polymorphic human xenobiotic-metabolizing enzymes. Cancer Res 59(12):2903–2908. Loriot MA, Rebuissou S, et al. 2001. Genetic polymorphism of cytochrome P450 2A6 in a case-control study of lung cancer in a French population. Pharmacogenetics 11(1):39–44. Lossos IS, Jones CD, Zehnder JL, Levy R. 2001. A polymorphism in the BCL6 gene is associated with follicle center lymphoma. Leuk Lymphoma 42(6):1343–1350. Lower GM Jr., Nilsson T, Nelson CE, Wolf H, Gamsky TE, Bryan GT. 1979. N-Acetyltransferase phenotype and risk of bladder cancer: Approaches in

molecular epidemiology. Preliminary results from Sweden and Denmark. Environ Health Perspec 29:71–79. Lu CC, Chen JC, Yang HB, Chan SH, Tsai ST. 2003. Genetic susceptibility to nasopharyngeal carcinoma within the HLA-A locus in Taiwanese. Int J Cancer 103(6):745–751. Lundberg AS, Hahn WC, Gupta P, Weinberg RA. 2000. Genes involved in senescence and immortalization. Curr Opin Cell Biol 12:705–709. Ma J, Stampfer MJ, Giovannucci E, Artigas C, Hunter DJ, Fuchs C, et al. 1997. Methylenetetrahydrofolate reductase polymorphism, dietary interactions, and risk of colorectal cancer. Cancer Res 57:1009–1102. Ma J, Stampfer MJ, Gann PH, Hough HL, Giovannucci E, Kelsey KT, et al. 1998. Vitamin D receptor polymorphisms, circulating vitamin D metabolites, and risk of prostate cancer in the United States. Cancer Epidemiol Biomarkers Prev 7:385–390. Madigan MP, Gao YT, Deng J, Pfeiffer RM, Chang BL, et al. 2003. CYP17 polymorphism in relation to prostate cancer and benign prostatic hyperplasia: A population-based study in China. Int J Cancer 107(2): 271–275. Maloney EK, McLaughlin JL, Dagdigian NE, Garrett LM, Conners KM, et al. 2003. An anti-insulin-like growth factor I receptor antibody that is a potent inhibitor of cancer cell proliferation. Cancer Res 63:5073–5083. Manenti G, Nomoto T, et al. 2000. Predisposition to lung tumorigenesis. Toxicol Lett 112–113:257–263. Marchini J, Cardon LR, Phillips MS, Donnelly P. 2004. The effect of human population structure on large genetic association studies. Nat Genet 36(5): 512–517. Marcus PM, Travis WD. 2002. Lung. In: Cancer Precursors. Franco EL, Rohan TE, eds. New York: Springer-Verlag. 201–219. Marcus PM, Vineis P, Rothman N. 2000a. NAT2 slow metabolism and bladder cancer risk: A meta-analysis of 22 case-control studies conducted in the general population. Pharmacogenetics 10(2):115–122. Marcus P, Hayes R, et al. 2000b. Cigarette smoking, N-acetyltransferase 2 acetylation status, and bladder cancer risk: A case-series meta-analysis of a gene-environment interaction Cancer Epidemiol Biomarkers Prev 9(5):461–467. Margiotti K, Kim E, Pearce CL, Spera E, Novelli G, Reichardt JK. 2002. Association of the G289S single nucleotide polymorphism in the HSD17B3 gene with prostate cancer in Italian men. Prostate 53(1):65–68. Marti GE, Rawstrom AC, Ghia P, Hillman P, Houlston PS, Kay N, Schleinitz TA, Caporaso N. 2005. Diagnostic criteria for monoclonal B-cell lymphocytosis. Br J Haematol 130(3):325–332. Martinez ME, O’Brien TG, Fultz KE, Babbar N, Yerushalmi H, Qu N, Guo Y, Boorman D, et al. 2003. Pronounced reduction in adenoma recurrence associated with aspirin use and a polymorphism in the ornithine decarboxylase gene. Proc Natl Acad Sci USA 100(13):7859–7864. McGue M, Bouchard TJ Jr., 1998. Genetic and Environmental influences on human behavior differences. Annu Rev Neurosci 21:1–24. McWilliams JE, Sanderson BJS, Harris EL, Richert-Boe KE, Henner WD. 1995. Glutathione S–transferase M1 (GSTM1) deficiency and lung cancer risk. Cancer Epidemiol Biomarkers Prev 4:589–594. Medeiros R, Morais A, Vasconcelos A, Costa S, Pinto D, Oliveira J, Lopes C. 2002. Endothelial nitric oxide synthase polymorphisms and genetic susceptibility to prostate cancer. Eur J Cancer Prev 11(4):343–350. Medeiros R, Vasconcelos A, Costa S, Pinto D, Morais A, Oliveira J, Lopees C. 2003a. Steroid hormone genotypes ARStuI and ER325 are linked to the progression of human prostate cancer. Cancer Genet Cytogenet 141(2):91–96. Medeiros R, Morais A, Vasconcelos A, Costa S, Carrilho S, Oliveira J, Lopes C. 2003b. Endothelial nitric oxide synthase gene polymorphisms and the shedding of circulating tumour cells in the blood of prostate cancer patients. Cancer Lett 189(1):85–90. Meireles SI, Cristo EB, Carvalho AF, Hirata R Jr, Pelosof A, Gomes LI, et al. 2004. Molecular classifiers for gastric cancer and nonmalignant diseases of the gastric mucosa. Cancer Res 64:1255–1265. Menzel HJ, Sarmanova J, Soucek P, Berberich R, Grunewald K, Haun M, Kraft HG. 2004. Association of NQO1 polymorphism with spontaneous breast cancer in two independent populations. Br J Cancer 90(10):1989–1994. Merikangas KR, Risch N. 2003. Genomic priorities and public health. Science 302:599–601. Meitz JC, Edwards SM, Easton DF, Murkin A, Ardern-Jones A, Jackson RA, Williams S, et al. 2002. HPC2/ELAC2 polymorphisms and prostate cancer risk: Analysis by age of onset. Br J Cancer 87(8):905–908. Miao X, Yu C, Tan W, Xiong P, Liang G, Lu W, Lin D. 2003. A functional polymorphism in the matrix metalloproteinase-2 gene promoter (1306C/T) is associated with risk of development but not metastasis of gastric cardia adenocarcinoma. Cancer Res 63(14):3987–3990. Miller DC, Zheng SL, Dunn RL, Sarma AV, Montie JE, et al. 2003. Germline mutations of the macrophage scavenger receptor 1 gene:association

Genetic Modifiers of Cancer Risk with prostate cancer risk in African American men. Cancer Res 63: 3486–3489. Miller DP, Liu G, De Vivo I, Lynch TJ, Wain JC, Su L, Christiani DC. 2002. Combinations of the variant genotypes of GSTP1, GSTM1, and p53 are associated with an increased lung cancer risk. Cancer Res 62:2819– 2823. Millikan RC, Hummer A, Begg C, Player J, de Cotret AR, Winkel S, Mohreweiser H, Thomas N, Armstrong B, Kricker A, Marrett LD, Gruber SB, Anton Culver H, Zanetti R, Gallagher RP, Dwyer T, Rebbeck T, Busam K, From L, Mujumdar U, Berwick M. 2005. Polymorphisms in nucleotide excision repair genes and risk of multiple primary melanoma: the Genes Environment and Melanoma study. Carcinogenesis 29; [Epub ahead of print]. Minematsu N, Nakamura H, Iwata M, Tateno H, Nakajima T, Takahashi S, Fujishima S, Yamaguchi K. 2003. Association of CYP2A6 delection polymorphism with smoking habit and development of pulmonary emphysema. Thorax 58:623–628. Mitrunen K, Hirvonen A. 2003. Molecular epidemiology of sporadic breast cancer. The role of polymorphic genes involved in oestrogen biosynthesis and metabolism. Mutat Res 544(1):9–41. Mononen N, Syrjakoski K, et al. 2000. Two percent of Finnish Prostate Cancer Patients have a germ-line mutation in the hormone-binding domain of the androgen receptor gene. Cancer Res 60:6479–6481. Monzoni A, Masutti F, Saccoccio G, Bellentani S, Tiribelli C, Giacca M. 2001. Genetic determinants of ethanol–induced liver damage. Mol Med 7(4):255–262. Moore JH. 2003. The ubiquitous nature of epistasis in determining susceptibility to common human diseases. Hum Hered 56:73–82. Moran JL, Siegel D, Ross D. 1999. A potential mechanism underlying the increased susceptibility of individuals with a polymorphism in the NAD(P)H: quinone oxidoreductase 1 (NQO1) to benzene toxicity. Proc Natl Acad Sci USA 96:8150–8155. Mohrenweiser HW, Xi T, Jones IM. 2003. Many common amino acid substitution variants identified in DNA repair genes during population screening are predicted to impact protein function. Proc AACR 44: #5826. Muckel E, Frandsen H, Glatt HR. 2002. Heterologous expression of human N-acetyltransferases 1 and 2 and sulfotransferase 1A1 in Salmonella typhimurium for mutagenicity testing of heterocyclic amines. Food Chem Toxicol 40:1063–1068. Munaka M, Kohshi K, Kawamoto T, Takasawa S, Nagata N, Itoh H, Oda S, Katoh T. 2003. Genetic polymorphisms of tobacco and alcohol-related metabolizing enzymes and risk of hepatocellular carcinoma. J Cancer Res Clin Oncol 129(6):355–360. Naf D, Krupke DM, Sundberg JP, Eppig JT, Bult CJ. 2002. The Mouse Tumor Biology Database: A public resource for cancer genetics and pathology of the mouse. Cancer Res 62:1235–1240. Neale BM, Sham PC. 2004. The future of association studies: Gene-based analysis and replication. Am J Hum Genet 75:353–362. Nowell S, Ratnasinghe DL, Ambrosome CB, Williams S, Teague-Ross T. 2004. Association of SULT1A1 phenotype and genotype with prostate cancer risk in African-Americans and Caucasians. Cancer Epidemiol Biomarkers Prev 13:270–276. Ntais C, Polycarpou A, Ioannidis JPA. SRD5A2. 2003a. Gene polymorphism and risk of prostate cancer: A meta-analysis. Cancer Epidemiol Biomarkers Prev 12(7):618–624. Ntais C, Polycarpou A, Ioannidis JPA. 2003b. Vitamin D receptor gene polymorphism and risk of prostate cancer: A meta-analysis. Cancer Epidemiol Biomarkers Prev 12:1395–1402. Ntais C, Polycarpou A, Ioannidis JP. 2003c. Association of the CYP17 gene polymorphism with the risk of prostate cancer: A meta-analysis. Cancer Epidemiol Biomarkers Prev 12(2):120–126. Ockenga J, Vogel A, Teich N, Keim V, Manns MP, Strassburg CP. 2003. UDP glucuronyltransferase (UGT1A7) gene polymorphisms increase the risk of chronic pancreatitis and pancreatic cancer. Gastroenterology 124(7): 1802–1808. Oscarson M. 2001. Genetic polymorphisms in the cytochrome P450 (CYP2A6) gene: Implications for interindividual differences in nicotine metabolism. Drug Metabolism Disposition 29(2):91–95. Ottman R. 1990. An epidemiologic approach to gene-environment interaction. Genet Epidemiol 7:177–185. Park SJ, Zhao H, Spitz MR, Grossman HB, Wu X. 2003. An association between NQO1 genetic polymorphism and risk of bladder cancer. Mutat Res 536(1–2):131–137. Park GT, Lee OY, Kwon SJ, Lee CG, Yoon BC, Hahm JS, Lee MH, Hoo Lee D, Kee CS, Sun HS. 2003. Analysis of CYP2E1 polymorphism for the determination of genetic susceptibility to gastric cancer in Koreans. J Gastroenterol Hepatol 18(11):1257–1263.

597

Paz MF, Avila S, Fraga MF, Pollan M, et al. 2002. Germ-line variants in methylgroup metabolism genes and susceptibility to DNA methylation in normal tissue and human primary tumors. Cancer Res 62(15):4519–4524. Peltonen L, Palotie A, Lange K. 2000. Use of population isolates for mapping complex traits. Nature 1:182–190. Peto J, Mack TM. 2000. High constant incidence in twins and other relatives of women with breast cancer. Nature Genet 4:411–414. Peto J, Houlston RS. 2001. Genetics and common cancers. Eur J Cancer 37:S88–S96. Pfeiffer RM, Rutter JL, Gail MH, Struewing J, Gastwirth JL. 2002. Efficiency of DNA pooling to estimate joint allele frequencies and measure linkage disequilibrium. Genet Epidemiol 22:94–102. Pharoah PDP, Antoniou A, Bobrow M, Zimmern RL, Easton DF, Ponder BAJ. 2002. Polygenic susceptibility to breast cancer and implications for prevention. Nat Genet 31(1):33–36. Plummer SJ, Conti DV, Paris PL, Curran AP, Casey G, Witte JS. 2003. CYP3A4 and CYP3A5 genotypes, haplotypes and risk of prostate cancer. Cancer Epidemiol Biomarkers Prev 12:928–932. Potter JD. 2001. At the interface of epidemiology, genetics and genomics. Nat Genet 2:142–147. Pritchard JK, Cox NJ. 2002. The allelic architecture of human disease genes: Common disease–common variant . . . or not? Hum Mol Genet 11(20): 2417–2423. Protopopov A, Kashuba V, Zabarovska VI, Muravenko OV, Lerman MI, Klein G, Zabarovsky ER. 2003. An integrated physical and gene map of the 3.5-Mb chromosome 3p21 (AP20) region implicated in major human epithelial malignancies. Cancer Res 63:404–412. Qiuling S, Yuxin Z, Suhua S, Cheng X, et al. 2003. Cyclin D1 gene polymorphism and susceptibility to lung cancer in a Chinese population. Carcinogenesis 24(9):1499–1503. Rebbeck TR, Ambrosone CB, Bell DA, Chanock SJ, Hayes RB, Kadlubar FF, Thomas DC. 2004. SNPs, haplotypes, and cancer: Applications in molecular epidemiology. Cancer Epidemiol Biomarkers Prev 13(5): 681–687. Ren Z, Cai Q, Shu X-O, Cai H, et al. 2004. Genetic polymorphisms in the IGFBP3 gene: association with breast cancer risk and blood IGFBP-3 protein levels in Chinese women, Cancer Epidemiol Biomarkers Prev 2004; 13(8). Risch NJ. 2000. Searching for genetic determinants in the new millennium. Nature 405:847–856. Risch N, Merikangas K. 1996. The future of genetic studies of complex human diseases. Science 273(5281):1516–1517. Risch N. 2001. The genetic epidemiology of cancer: interpreting family and twin studies and their implications for molecular genetic approaches. Cancer Epidemiol Biomarkers Prev 10:733–741. Risch N, Burchard E, Ziv E, Tang H. 2002. Categorization of humans in biomedical research:genes, race and disease. Genome Biol 3(7):2007–2012. Robien K, Ulrich CM. 2003. 5,10–Methylenetetrahydrofolate reductase polymorphisms and leukemia risk: A Huge minireview. AJE 157(7):571–582. Robledo M, Gil L, Pollan M, Cebrian A, Ruiz S, et al. 2003. Polymorphisms G691S/S904S of RET as genetic modifiers of MEN 2A. Cancer Res 63:1814–1817. Roe I, Nam S, Kim J, Shin J, Bang W, Yang M. 2002. Association of the myeloperoxidase -463G->A polymorphism with development of atrophy in Helicobacter pylori-infected gastritis. Am J Gastroenterol 97(7):1629–1634. Roh JW, Kim JW, Park NH, Song YES, Park IA, Park SY, Kang SB, Lee HP. 2004. P53 and p21 genetic polymorphisms and susceptibility to endometrial cancer. Gynecol Oncol 93(2):499–505. Rollinson S, Roddam P, Kane E, Roman E, Cartwright R, Jack A, Morgan GJ. 2000. Polymorphic variation within the glutathione S-transferase genes and adult acute leukemia. Carcinogenesis 21(1):43–47. Rosenberg NA, Pritchard JK, Weber JL, Cann HM, Kidd KK, Zhivototovsky LA, Feldman MW. 2002. Genetic structure of human populations. Science 298:2381–2385. Rossi L, Leveri M, Gritti C, DeSilvestri A, Zavaglia C, Sonzogni L, Silvestri L, Civardi E, Modelli MU, Silini EM. 2003. Genetic polymorphisms of steroid hormone metabolizing enzymes and risk of liver cancer in hepatitis C-infected patients. J Hepatol 39(4):564–570. Rostami-Hodjegan A, Lenna Rostami-Hodjegan MS, Woods HF, Tucker GT. 1998. Meta-analysis of studies of the CYP2D6 polymorphism in relation to lung cancer and Parkinson’s disease. Pharmacogenetics 3:227–238. Rothman KJ, Greenland S. 1998. Modern Epidemiology. Philadelphia: Lippincott–Raven. Rothman N, Wacholder S, Caporaso NE, Garcia-Closas M, Buetow K, Fraumeni JF Jr. 2001. The use of common genetic polymorphisms to enhance the epidemiologic study of environmental carcinogens Biochim Biophys Acta 1471(2):C1–10.

598

PART III: THE CAUSES OF CANCER

Rothman N, Stewart WF, Schulte PA. 1995. Incorporating biomarkers into cancer epidemiology: A matrix of biomarker and study design categories. Cancer Epidemiol Biomarkers Prev 4:301–311. Rutter JL, Chatterjee N, Wacholder S, Struewing J. 2003. The HER2 I655V polymorphism and breast cancer risk in Ashkenazim. Epidemiology 14:694–700. Rybicki BA, Conti DV, Moreira A, Cicek M, Casey G, Witte JS. 2004. DNA repair gene XRCC1 and XPD polymorphisms and risk of prostate cancer. Cancer Epidemiol Biomarkers Prev 13:23–29. Rylander-Rudqvist T, Wedren S, Granath F, Humphreys K, Ahlberg S, Weiderpass E, Oscarson M, Ingelman-Sundberg M, Persson I. 2003. Cytochrome P450 1B1 gene polymorphisms and postmenopausal breast cancer risk Carcinogenesis 24(9):1533–1539. Sakamoto T, Hara M, Higaki Y, Ichiba M, Horita M, Mizuta T, Eguchi Y, Yasutake T, Ozaki I, Yamamoto K, Onohara S, Kawazoe S, Shigematsu H, Koizumi S, Tanaka K. 2005. Influence of alcohol consumption and gene polymorphisms of ADH2 and ALDH2 on hepatocellular carcinoma in a Japanese population. Int J Cancer 26; [Epub ahead of print]. Samet JM, Humble CG, Pathak DR. 1986. Personal and family history of respiratory disease and lung cancer risk. J Natl Cancer Inst 76:217–222. Samuelson DJ, Aperavich BA, Haag JD, Gould MN. 2005. Fine mapping reveals multiple loci and a possible epistatic interaction within the mammary carcinoma susceptibility quantitative trait locus, Mcs5. Cancer Res 65:(21). Sanchez-Carbayo M, Socci ND, Charytonowicz E, Lu M, Prystowsky M, Childs G, Cordon-Cardo C. 2002. Molecular Profiling of bladder cancer using cDNA microarrays: Defining histogenesis and biological phenotypes. Cancer Res 62:6973–6980. Sasaki M, Tanaka Y, Kaneuchi M, Sakuragi N, Dahiya R. 2003. CYP1B1 gene polymorphisms have higher risk for endometrial cancer, and positive correlations with estrogen receptor alpha and estrogen receptor beta expressions. Cancer Res 63(14):3913–3918. Savas S, Kim DY, Ahmad F, Shariff M, Ozcelik H. 2004. Identifying functional genetic variants in DNA repair pathway using protein conservation analysis. Cancer Epidemiol Biomarkers Prev 13(5):801–807. Schabath MB, Spitz MR, Zhang X, Delclos GL. 2000. Genetic variants of myeloperoxidase and lung cancer risk. Carcinogenesis 21(6):1163–1166. Seidegard J, Pero RW, Miller DG, Beattie EJ. 1986. A glutathione transferase in human leukocytes as a marker of susceptibility to lung cancer. Carcinogenesis 7:751–753. Sellers TA, Schildkraut JM, Pankratz VS, Vierkant RA, Fredericksen ZS, Olson JE, Cunningham J, Taylor W, Liebow M, McPherson C, Hartmann LC, Pal T, Adjei AA. 2005. Estrogen bioactivation, genetic polymorphisms, and ovarian cancer. Cancer Epidemiol Biomarkers Prev 14:2536–2543. Setiawan VW, Hankinson SE, Colditz GA, et al. 2004. HSD17B1 Gene polymorphisms and risk of endometrial and breast cancer. Cancer Epidemiol Biomarkers Prev 13:213–219. Severi G, Giles GG, Southey MC, Tesoriero A, Tilley W, Neufing P, Morris H, English DR, McCredie MR, Boyle P, Hopper JL. 2003. ELAC2/ HPC2 polymorphisms, prostate-specific antigen levels, and prostate cancer. J Natl Cancer Inst 95(11):818–824. Sharma A, Sharma JK, Murthy NS, Mitra AB. 2004. Polymorphisms at GSTM1 and GSTT1 gene loci and susceptibility to cervical cancer in Indian population. Neoplasma 51(1):12–16. Shaw GL, Falk RT, Pickle LW, Mason TJ, Buffler PA. 1991. Lung cancer risk associated with cancer in relatives. J Clin Epidemiol 44(4–5):429–437. Shifman S, Kuypers J, Kokoris M, Yakir B, Darvasi A. 2003. Linkage disequilibrium patterns of the human genome across populations. Hum Mol Genet 12(7):771–776. Shrubsole MJ, Gao Y-T, Cai Q, et al. 2004. MTHFR polymorphisms, dietary folate intake, and breast cancer risk results from the Shanghai Breast Cancer Study. Cancer Epidemiol Biomarkers Prev 13:190–196. Shu X-O, Gao Y-T, Cai Q, Pierce L, Cai H, et al. 2004. Genetic polymorphisms in the TGF-B1 gene and breast cancer survival: A report from the Shanghai Breast Cancer Study. Cancer Res 64:836–839. Sibley K, Rollinson S, Allan JM, Smith AG, Law GR, Roddam PL, Skibola CF, Smith MT, Morgan GJ. 2003. Functional FAS promoter polymorphisms are associated with increased risk of acute myeloid leukemia. Cancer Res 63:4327–4330. Simard J, Dumont M, Labuda D, Sinnett D, Meloche C, El-Alfy M, Berger L, Lees E, Labrie F, Tavtigian SV. 2003. Prostate cancer susceptibility genes:lessons learned and challenges posed. Endocr Relat Cancer 1092:225–259. Sinues B, Fanlo A, Bernal ML, Mayayo E, Bello S, Rubio E, Isla D. 2003. MDR-1 C3435T genetic polymorphism and tobacco–related lung cancer. Oncology 64:183–185. Smith PG, Day NE. 1984. The design of case-control studies:the influence of confounding and interaction effects. Int J Epidemiol 13:356–365.

Smith GB, Harper PA, Wong JM, Lam MS, Reid KR, Petsikas D, Massey TE. 2001. Human lung microsomal cytochrome P4501A1 (CYP1A1) activities: Impact of smoking status and CYP1A1, AHR and GSTM1 genetic polymorphisms. Cancer Epidemiol Biomarkers Prev 10(8):839– 853. Snoek M, Albertella MR, et al. 2000. G7c, a novel gene in the mouse and human major histocompatibility complex class III region, possibly controlling lung cancer susceptibility. Immunogenetics 51(4–5):383– 386. Soderstrom T, Wadelius M, Andersson SO, Johansson JE, Johansson S, Granath F, Rane A. 2002. 5alpha-reductase 2 polymorphisms as risk factors in prostate cancer. Pharmacogenetics 12(4):307–312. Sommer SS, Jiang Z, Feng J, Buzin CH, Zheng J, et al. 2003. ATM missense mutations are frequent in patients with breast cancer. Cancer Genet Cytogenet 145(2):115–120. Spotilla LD, Jacques PF, Berger PB, Ballman KV, Ellison RC, Rozen R. 2003. Age dependence of the influence of MTHFR genotype on plasma homocysteine levels. Am J Epidemiol 158(9):871–877. Stanford JL, Sabacan LP, Noonan EA, Iwasaki I, Shu J, Feng Z, Ostrander EA. 2003. Association of HPC2/ELAC2 polymorphisms with risk of prostate cancer in a population-based study. Cancer Epidemiol Biomarkers Prev.12(9):876–881. Stern MC, Johnson LR, Bell DA, Taylor JA. 2002. XPD codon 751 polymorphism, metabolism genes, smoking, and bladder cancer risk. Cancer Epidemiol Biomarkers Prev 11(10 Pt 1):1004–1011. Sturm RA, Duffy DL, Box NF, Newton RA, Shepard AG, Chen W, Marks LH, Leonard JH, Martin NG. 2003. Genetic association and cellular function of the MC1R variant alleles in human pigmentation. Ann NY Acad Sci 994:348–358. Su L, Zhou W, Park S, Wain JC, Lynch TJ, Liu G, Christiani DC. 2005. Matrix metalloproteinase-1 promoter polymorphism and lung cancer risk. Cancer Epidemiol Biomarkers Prev 14:567–570. Sugawara T, Nomura E, Sagawa T, Sakuragi N, Fujimoto S. 2003. CYP1A1 polymorphism and risk of gynecological malignancy in Japan. Int J Gynecol Cancer 13(6):785–790. Sukosd F, Kuroda N, Beothe T, Kaur AP, Kovacs G. 2003. Deletion of chromosome 3p14.2-p25 involving the VHL and FHIT genes in conventional renal cancer. Cancer Res 63:455–457. Suter NM, Malone KE, Daling JR, et al. 2003. Androgen receptor CAGn and GGCn polymorphisms and breast cancer risk in a population-based casecontrol study of young women. Cancer Epidemiol Biomarkers Prev 12(2):127–135. Sweeney C, Farrow DC, Schwrtz SM, Eaton DL, Checkoway H, Vaughan TL. 2000. Glutathione-S-transferase M1, T1, and P1 polymorphisms as risk factors for renal cell carcinoma: A case-control study. Cancer Epidemiol Biomarkers Prev 9(4):449–454. Tabor HK, Risch NJ, Myers RM. 2002. Candidate-gene approaches for studying complex genetic traits:practical considerations. Nat Rev 3:1–7. Taguchi A, Ohmiya N, Shirai K, Mabuchi N, Itoh A, Hirooka Y, Niwa Y, Goto H. 2005. Interleukin-8 promoter polymorphism increases the risk of atrophic gastritis and gastric cancer in Japan. Cancer Epidemiol Biomarkers Prev 14(11):2487–2493. Taioli E, Gaspari L, Benhamou S, Boffetta P, et al. 2003. Polymorphisms in CYP1A1, GSTM1, GSTT1 and lung cancer below the age of 45 years. Int J Epidemiol 32(1):60–63. Tamimi RM, Hankinson SE, Ding S, Gagalang V, Larson GP, et al. 2003. The HRAS1 variable number of tandem repeats and risk of breast cancer. Cancer Epidemiol Biomarkers Prev 12:1528–1530. Tan W, Chen GF, Xing DY, Song CY, Kadlubar FF, Lin DX. 2001. Frequency of CYP2A6 gene deletion and its relation to risk of lung and esophageal cancer in Chinese population. Int J Cancer 95(2):96–101. Tanaka Y, Sasaki M, Kaneuchi M, Shiina H, Igawa D, Dahiya R. 2002. Polymorphisms of the CYP1B1 gene have a higher risk for prostate cancer. Biochem Biophys Res Commun 296(4):820–826. Tay ST, Leong SH, Yu K, Aggarwal A, Tan SY, Lee CH, et al. 2003. A combined comparative genomic hybridization and expression microarray analysis of gastric cancer reveals novel molecular subtypes. Cancer Res 63:3309–3316. Tayeb MT, Clark C, Haites NE, Sharp L, Murray GL, McLeod. 2003. CYP3A4 and VDR gene polymorphisms and the risk of prostate cancer in men with benign prostate hyperplasia. Br J Cancer 88(6):928–932. Taylor JG, Choi E-H, Foster CB, Chanock SJ. 2001. Using genetic variation to study human disease. Trends Mol Med 7(11):507–511. Thompson D, Antoniou AC, Jenkins M, Marsh A, Chen X, Wayne T, Tesoriero A, Milne R, Spurdle A, Thorstenson Y, Southey M, Giles GG, Khanna KK, Sambrook J, Oefner P, Goldgar D, Hopper JL, Easton D, ChenevixTrench G. 2005. KConFab investigators. two ATM variants and breast cancer risk. Hum Mutat 25(6):594–595.

Genetic Modifiers of Cancer Risk Thompson HJ, Zhu Z, Jiang W. 2004. Identification of the apoptosis cascade induced in mammary carcinomas by energy restriction. Cancer Res 64:1541–1545. Thompson PA, Shields PG, Fruedenheim JL, et al. 1998. Genetic polymorphisms in catechol-O-methyltransferase, menopausal status and breast cancer risk. Cancer Res 58:2107–2110. Thun MJ, Henley J, Patrono C. 2002. Nonsteroidal anti-inflammatory drugs as anticancer agents:mechanistic, pharmacologic and clinical issues. J Natl Cancer Inst 94(4):252–263. Thye T, Burchard GD, Nilius M, Muller-Myhsok B, Hostmann RD. 2003. Genomewide linkage analysis identifies polymorphism in the human interferon-gamma receptor affecting Helicobacter pylori infection. Am J Hum Genet 72(7):448–453. Tiemersma EW, Wark PA, Ocke MC, Bunschoten A, et al. 2003. Alcohol consumption, alcohol dehydrogenasae 3 polymorphism and colorectal adenomas. Cancer Epidemiol Biomarkers Prev 12:419–425. Tiwawech D, Srivatanakul P, Karaluk A, Ishida T. 2003. The p53 codon72 polymorphism in Thai nasopharyngeal carcinoma. Cancer Lett 198(1): 69–75. Tokuhata GK, Lilienfeld AM. 1963. Familial aggregation of lung cancer in humans. JNCI 30:289–312. Tomescu D, Kavanaugh G, Ha T, Campbell H, Melton DW. 2001. Nucleotide excision repair gene XPD polymorphisms and genetic predisposition to melanoma. Carcinogenesis 22(3):403–408. Toyooka KO, Toyooka S, Virmani AK, Sathyanarayana UG, et al. 2001. Loss of expression and aberrant methylation of the CDH13 (H–Cadherin) gene in breast and lung carcinomas. Cancer Res 61:4556–4560. Toyooka S, Maruyama R, Toyooka KO, McLerran D, et al. 2003. Smoke exposure, histologic type and geography-related differences in the methylation profiles of non-small cell lung cancer. Int J Cancer 103(2): 153–160. Tripodis N, Hart AM, Fijineman RJA, Demant P. 2001. Complexity of lung cancer modifiers: Mapping of thirty genes and twenty–five interactions in half of the mouse genome. J Natl Cancer Inst 93(19):1484–1491. Tseng CS, Tang KS, Lo HW, Ker CG, Teng HC, Huang CS. 2005. UDPglucuronosyltransferase 1A7 genetic polymorphisms are associated with hepatocellular carcinoma risk and onset age. Am J Gastroenterol 100(8):1758–1763. Tsukasaki K, Miller CW, Kubota T, Takeuchi S, Fujimoto T, Ikeda S, Tomonaga M, Koeffler HP. 2001. Tumor necrosis factor alpha polymorphism associated with increased susceptibility to development of adult T-cell leukemia/lymphoma in human T-lymphotropic virus type 1 carriers. Cancer Res 61:3770–3774. Tworoger SS, Chubak J, Aiello EJ, Ulrich CM, Atkinson C, Potter JD, Yasui Y, et al. 2004. Association of CYP17, CYP19, CYP1B1, and COMT polymorphisms with serum and urinary sex hormone concentrations in postmenopausal women. Cancer Epidemiol Biomarkers Prev 13:94–101. Uhl GR, Liu Q, Walther D, Hess J, Naiman D. 2001. Polysubstance abuse vulnerability gene: Genome scans for association using 1004 subjects and 1494 SNPs. Am J Hum Genet 69:1290–1300. Ulrich C, et al. 1999. Colorectal adenomas and the C677T MTHFR polymorphism. Evidence for gene-environment interaction. Cancer Epidemiol Biomarkers Prev 8:659–668. Vajdic C, Kricker A, Duffy DL, Aitken JF, Stark M, ter Huurne JA, et al. 2003. Ocular melanoma is not associated with CDKN2A or MC1R variants— a population-based study. Melanoma Res 13(4):409–413. van der Logt EM, Bergevoet SM, Roelofs HM, Te Morsche RH, Dijk YV, Wobbes T, Nagengast FM, Peters WH. 2006. Role of epoxide hydrolase, NAD(P)H: quinone oxidoreductase, cytochrome P450 2E1 or alcohol dehydrogenase genotypes in susceptibility to colorectal cancer. Mutation Res 593(1–2):39–49. Van Eerdewegh P, Little RD, Dupuis J, Del Mastro RG, Falls K, Simon J, Torrey D, et al. 2002. Association of the ADAM33 gene with asthma and bronchial hyperresponsiveness. Nature 418:426–430. Vineis P, Veglia F, Benhamou S, Butkiewicz D, Casorbi I, et al. 2003. CYP1A1 T3801 C polymorphism and lung cancer: A pooled analysis of 2451 cases and 3358 controls. Int J Cancer 1104(5):650–657. Vineis P, Marinelli D, Autrup H, Brockmoller J, Cascorbi I, Daly AK, et al. 2001. Current smoking, occupation, N-acetyltransferase and bladder cancer: A pooled analysis of genotype-based studies. Cancer Epidemiol Biomarkers Prev 10(12):1249–1252. Vineis P, Malats N, Lang M, d’Errico A, Caporaso, N, Cuzick J. Boffetta P. 1999. Metabolic polymorphisms and susceptibility to cancer. IARC Scientific Publication No. 1487. Lyon: IARC. Vineis P, Schulte P, McMichael AJ. 2001. Misconceptions about the use of genetic tests in populations. Lancet 357:709–712. Vineis P, Veglia F, Anttila S, Benhamou S, Clapper ML, Dolzan V, Ryberg D, Hirvonen A, Kremers P, Le Marchand L, Pastorelli R, Rannug A, Romkes

599

M, Schoket B, Strange RC, Garte S, Taioli E. 2004. CYP1A1, GSTM1 and GSTT1 polymorphisms and lung cancer: a pooled analysis of genegene interactions. Biomarkers 9(3):298–305. Vineis P. 2002. The relationship between polymorphisms of xenobiotic metabolizing enzymes and susceptibility to cancer. Toxicology 181/182: 457–462. Van Gils CH, Oland-Moret NC, Roest M, et al. 2003. The V89L Polymorphism in the 5-alpha-reductase type 2 gene and risk of breast cancer. Cancer Epidemiol Biomarkers Prev 12:1194–1199. Wacholder S, Chanock S, Garcia-Closas M, El ghormli L, Rothman N. 2004. Assessing the probability that a positive report is false: An approach for molecular epidemiology studies. JNCI 96(6):434–442. Wacholder S, Garcia-Closas M, Rothman N. 2002. Study of genes and environmental factors in complex disease. Lancet 359:1155. Wacholder S, McLaughlin JK, Silverman DT, Mandel JS. 1992. Selection of controls in case-control studies I. Principles. Am J Epidemiol 135(9): 1019–1050. Wacholder S, Rothman N, Caporaso N. 2000. Population stratification in epidemiologic studies of common genetic variants and cancer:quantification of bias. J Natl Cancer Inst 92(14):1151–1158. Wacholder S, Rothman N, Caporaso N. 2002. Counterpoint: Bias from population stratification is not a major threat to the validity of conclusions from epidemiological studies of common polymorphisms and cancer. Cancer Epidemiol Biomarkers Prev 11:513–520. Wang H, Tan W, Hao B, Miao X, Zhou G, He F, Lin D. 2003. Substantial reduction in risk of adenocarcinoma associated with genetic polymorphism in CYP2A13, the most active cytochrome P450 for the metabolic activation of tobacco-specific carcinogen NNK. Cancer Res 63:8057– 8061. Wang J, Deng Y, Li L, Kuriki K, Ding J, Pan X, et al. 2003. Association of GSTM1, CYP1A1 and CYPZE1 genetic polymorphisms with susceptibility to lung adenocarcinoma: A case control study in Chinese population. Cancer Sci 94(5):448–452. Wang S, Zhao H. 2003. Sample size needed to detect gene-gene interactions using association designs. Amer J Epidemiology 158(9):899–914. Wang XL, Raveendran M, Wang J. 2003. Genetic influence on cigaretteinduced cardiovascular disease. Prog Cardiovascular Dis 45(5):361– 382. Wang Y, Spitz MR, Tsou AM, Zhang K, Makan N, Wu X. 2002. Sulfotransferase (SULT) 1a1 polymorphism as a predisposition for lung cancer: A case-control analysis. Lung Cancer 35(2):137–142. Wang Y, Liang D, Spitz MR, Zhang K, Dong Q, Amos CI, Wu X. 2003. XRCC3 genetic polymorphism, smoking, and lung carcinoma risk in minority populations. Cancer 98(8):1701–1706. Watson MA, Gay L, Stebbings WS, Speakman CT, Bingham SA, Lottionov A. 2003. Apolipoprotein e gene polymorphism and colorectal carcinoma: Gender specific modulation of risk and prognosis. Clin Sci 104(5): 537–545. Weiss KM, Clark AG. 2002. Linkage disequilibrium and the mapping of complex human traits. Trends Genet 18(1):19–24. Weber WW. 1987. The Acetylator Genes and Drug Response. New York: Oxford University Press. Wei M-H, Latif F, Bader S, Kashuba V, Chen J-Y, Duh F-M, et al. 1996. Construction of a 600-kilobase cosmid clone contig and generation of a transcriptional map surrounding the lung cancer tumor suppressor gene (TSG) on human chromosome 3p21.3 progress toward isolation of a lung cancer TSG. Cancer Res 56:1487–1492. Weiss KM, Clark AG. 2002. Linkage disequilibrium and the mapping of complex human traits. Trends Genetics 18(1):19–24. Weinshilboum R. 2003. Inheritance and drug response. N Engl J Med 348:529–537. Wikman H, Thiel S, Jager B, et al. 2001. Relevance of N-acetyltransferase 1 and 2 (NAT1, NAT2) genetic polymorphisms in non-small cell lung cancer susceptibility. Pharmacogenetics 11(2):157–168. Wilson JF, Weale ME, Smith AC, Gratrix F, Fletcher B, et al. 2001. Population genetic structure of variable drug response. Nat Genet 29:265–271. Williams NS, Gaynor RB, Scoggin S, Verma U, Gokaslan T, Simmang C, et al. 2003. Identification and validation of genes involved in the pathogenesis of colorectal cancer using cDNA microarrays and RNA interference. Clin Cancer Res 9:931–946. Wu MS, Chen CJ, Lin MT, Wang HP, Shun CT, Sheu JC, Lin JT. 2002. Genetic polymorphism of cytochrome p450 2E1, glutathione S-transferase M1 and T1, and susceptibility to gastric carcinoma in Taiwan. Int J Colorectal Dis 17(5):338–343. Xing D, Qi J, Tan W, Miao X, et al. 2003. Association of genetic polymorphisms in the DNA Repair gene XPD with risk of lung cancer and esophageal cancer in a Chinese population in Beijing. Zhonghua Yi Xue Yi Chuan Xue Zhi 20(1):35–38.

600

PART III: THE CAUSES OF CANCER

Xiong W, Zeng ZY, Xia JH, Shen SR, Li XL, Hu DX, Tan C, et al. 2004. A susceptibility locus at chromosome 3p21 linked to familial nasopharyngeal cancer. Cancer Res 64:1972–1974. Xu J, Lowey J, Wiklund F, Sun J, et al. 2005. The Interaction of GFour Genes in the inflammation pathway significantly predicts prostate cancer risk, Cancer Epidemiol Biomarkers Prev 14(11):2563–2568. Xu J, Zheng SL, Turner A, Isaacs SD, Wiley KE, Hawkis GA, Chang B-L, et al. 2002. Associations between hOGG1 sequence variants and prostate cancer susceptibility. Cancer Res 62:2253–2257. Xu JF, Zheng SL, Komiya A, Mychaleckyj JC, Issacs JC, et al. 2002. Germline mutations and sequence variants of the macrophage scavenger receptor 1 gene are associated with prostate cancer risk. Nat Genet 32:321–325. Yanagitani N, Kohno T, Kim JG, Tamura T, et al. 2003. Identification of D19S246 as a novel lung adenocarcinoma susceptibility locus by genomic survey with 10-cM resolution microsatellite markers. Cancer Epidemiol Biomarkers Prev 12(4):366–371. Youngren KY, Nadeau JH, Matin A. 2003. Testicular cancer susceptibility in the 129.MOLF-Chr19 mouse strain:additive effects, gene interactions and epigenetic modifications. Hum Mol Genet 12(4):389–398. Yang CX, Matsuo K, Ito H, Hirose K, Wakai K, Saito T, Shinoda M, Hatooka S, Mizutani K, Tajima K. 2005. Esophageal cancer risk by ALDH2 and ADH2 polymorphisms and alcohol consumption: exploration of geneenvironment and gene-gene interactions, Asian Pac J Cancer Prev 6(3):256–262. Yang P, Yokomizo A, Tazelaar HD, Marks RS, Lesnick TG, Miller DL, Sloan JA, Edell ES, et al. 2002. Genetic determinants of lung cancer short-term survival: The role of glutathione-related genes. Lung Cancer 35(3):221– 229. Yang Q, Khoury MJ, Friedman JM, Flanders WD. 2003. On the use of population attributable fraction to determine sample size for casecontrol studies of gene-environment interaction. Epidemiology 14(2): 137–138. Yang Q, Khoury MJ, Flanders WD. 1997. Sample size requirements in caseonly designs to detect gene-environment interaction. Am J Epidemiol 146(9):713–720. Yokoyama A, Omori T. 2003a. Genetic polymorphisms of alcohol and aldehyde dehydrogeneases and risk for esophageal and head and neck cancers. Jpn J Clin Oncol 33(3):111–121. Yokoyama T, Yokoyama A, Kato H, Muto M, Omori T, Haneda T, Kumagai Y, Igaki H, Yokoyama M, Watanabe H, Yoshimizu H. 2003b. Alcohol flushing, alcohol and aldehyde dehydrogenase genotypes and risk for esophageal squamous cell carcinoma in Japanese men. Cancer Epidemiol Biomarkers Prev 12:1227–1233. Yu C, Pan K, Xing D, Liang G, Tan W, Zhang L, Lin D. 2002. Correlation between a single nucleotide polymorphism in the matrix metallopro-

teinase-2 promoter and risk of lung cancer. Cancer Res 62(22):6430– 6433. Yu SZ, Huang XE, Koide T, Cheng G, Chen GC, Harada K, Ueno Y, Sueoka E, Oda H, Tashiro F, Mizokami M, et al. 2002. Hepatitis B and C viruses infection, lifestyle and genetic polymorphism as risk factors for hepatocellular carcinoma in Haimen, China. Jpn J Cancer Res 93(12):1287– 1292. Zhang J, Li Y, Wang R, Wed D, Sarbia M, Kuang G, Wu M, Wei L, He M, Zhang L, Wang S. 2003. Association of cyclin D. G870A. polymorphism with susceptibility to esophageal and gastric carcinoma in a northern Chinese population. Int J Cancer 105(2):281–284. Zhang J, Schulz WA, Li W, Wang R, Zotz R, Wen D, Siegel D, Ross D, Gabbert HE, Sarbia M. 2003. Association of NAD(P)H:quinone oxidoreductase 1 (NQO1) C609T polymorphism with esophageal squamous cell carcinoma in a German Caucasian and northern Chinese population. Carcinogenesis 24(5):905–909. Zhang SM, Willett WC, Selhub J, Hunter DJ, et al. 2003. Plasma folate, viamin B6, vitamin B12, homocysteine and risk of breast cancer. J Natl Cancer Inst 95:373–380. Zhang Y, Lan Q, Rothman N, Zhu Y, Zahm SH, Wang SS, Holford TR, et al. 2005. A putative exonic splicing polymorphism in the BCL6 gene and the risk of non-Hodgkin lymphoma. J Natl Cancer Inst 97(21):1616–1618. Zhao Y-G, Xiao A-Z, Park HI, Newcomer RG, Yan M, Man Y-G, Heffelfinger SC, Sang Q-XA. 2004. Endometase/matrilysin-2 in human breast ductal carcinoma in situ and its inhibition by tissue inhibitors of metalloproteinases-2 and -4: A putative role in the initiation of breast cancer invasion. Cancer Res 64:590–598. Zheng W, Gao Y-T, Shu X-O, Wen W, Cai Q, Dai Q, Smith JR. 2004. Population-based case-control study of CYP11A gene polymorphism and breast cancer risk. Cancer Epidemiol Biomarkers Prev 13(5):709–714. Zheng W, Gao Y-T, Shu X-O, Wen W, Cai Q, Smith JR. 2004. Population-based case-control study of CYP11A Gene Polymorphism and Breast Cancer Risk. Cancer Epidemiol Biomarkers Prev 13(5):709–714. Zhu Y, Brown HN, Zhang Y, Stevens RG, Zheng TPeriod3 structural variation: a circadian biomarker associated with breast cancer in young woman. 2005. Cancer Epidemiol Biomarkers Prev 14(1):268–270. Zhu Y, Spitz MR, Lei L, Mills GB, Wu X. 2001. A single nucleotide polymorphism in the matrix metalloproteinase-1 promoter enhances lung cancer susceptibility. Cancer Res 61:7825–7829. Zhu Y, Spitz MR, Amos CI, Lin J, Schabath MB, Wu X. 2004. An evolutionary perspective on single-nucleotide polymorphism screening in molecular epidemiology. Cancer Res 64:2251–2257. Zienolddiny S, Ryberg D, Maggini V, Skaug V, Canzian F, Haugen A. 2004. Polymorphisms of interleukin-1 beta gene are associated with increased risk of non-small cell lung cancer. Int J Cancer 109(3):353–356.

IV CANCER BY TISSUE OF ORIGIN

This page intentionally left blank

30

Cancers of the Nasal Cavity and Paranasal Sinuses ALYSON J. LITTMAN AND THOMAS L. VAUGHAN

C

ancers arising in the sinuses and nasal cavity occur rarely in the United States and in most areas of the world. During 2002 it is estimated that 2000 people in the United States developed the disease. The incidence of sinonasal cancer (SNC) has remained relatively steady during the past 25 years at 0.7 cases per 100,000 persons per year, with a median survival of 5.7 years (Ries et al., 2003). A distinctive characteristic of SNC is its strong association with a number of occupational exposures, particularly dusts. The identification of clusters of sinonasal adenocarcinomas occurring among workers in the furniture, leather, and shoe manufacturing industries, as well as in nickel refining, has led to the designation of SNC as a “sentinel health event,” indicating a possibly harmful occupational environment when diagnosed in an employee of an at-risk industry (Rutstein et al., 1983).

CLASSIFICATION Anatomic Distribution An overview of the anatomy of the region is given in Figure 30–1. Anteriorly, the nasal cavity borders on the skin of the nose; posteriorly, the nasal cavity joins the nasopharynx. The paranasal sinuses are air-filled cavities within the maxillary, frontal, sphenoid, and ethmoid bones. Misclassification between sinonasal tumors and cancers of surrounding anatomic sites may affect observed population-based incidence and mortality rates and may also influence reported etiologic correlates. In the United States, the most common subsites for SNC are the nasal cavity (ICD-O 300) and maxillary sinus (ICD-O 310) (Table 30–1). Cancers of the ethmoid sinus (ICD-O 311), middle ear (ICDO 301), frontal sinus (ICD-O 312), sphenoid sinus (ICD-O 313), overlapping lesions of the accessory sinuses (ICD-O 318), and accessory sinus, NOS (ICD-O 319) are typically more rare.

Histopathology The normal epithelium of most of the sinonasal region is characterized by ciliated, pseudostratified columnar epithelium, which includes a substantial number of goblet cells that permit mucus secretion. The sinonasal respiratory epithelium creates a mucociliary blanket that allows clearance of particulate matter. The majority of cancers arising in the sinonasal region are squamous cell carcinomas (ICD-O codes 8050–8082) (Table 30–2). Other histologic types include: adenocarcinoma (ICD-O codes 8140–8211, 8260–8310); esthesioneurocytoma and neuroblastoma (ICD-O codes 9500–9561); carcinoma, NOS (ICD-O codes 8000–8012); melanoma (ICD-O codes 8720–8772); sarcoma (ICD-O codes 8800–9041, 9120–9240, 9370); undifferentiated carcinoma (ICD-O codes 8020–8041); transitional cell carcinoma (ICD-O codes 8120–8130); and other (ICD-O codes 8090–8094, 8240, 8246, 8410–8693, 9071–9081, 9364, 9473).

Precursor Neoplastic Lesions Under conditions of chronic inflammation, normal epithelium may become metaplastic and progress to dysplasia and then invasive cancer

over a period of years or decades. Examples include the lung (Gazdar, 1994), esophagus (Haggitt, 1994), and stomach (Correa, 1992). There is limited evidence that a similar histologic progression may also occur for some cancers arising in the nasal cavity and sinuses. For example, among persons thought to be at high risk of SNC due to long-term exposure to wood dust or formaldehyde, increased prevalence of metaplasia and dysplasia has been observed (Boysen and Solberg, 1982; Bussi et al., 2002; Edling et al., 1987). However, longitudinal studies necessary to quantify the extent to which these lesions predict neoplastic progression to cancer have not been carried out, due primarily to the rarity of SNC. In contrast, there is solid evidence linking papillomas, particularly inverted and oncocytic schneiderian types, to subsequent development of SNC (Kaufman et al., 2002; Lawson et al., 1995; Yoon et al., 2002). These uncommon neoplasms can occur throughout the sinuses and nasal cavities, with the lateral nasal wall being the most common site (Kaufman et al., 2002). Approximately 3%–13% eventually progress to invasive cancer (Bielamowicz et al., 1993; Kaufman et al., 2002; Lawson et al., 1995).

Molecular Genetic Characteristics Relatively few studies have investigated the molecular genetic characteristics of SNCs, with most focusing on the possible role of the p53 tumor suppressor gene in tumorigenesis. Inactivation of the p53 gene through mutation and allelic loss is one of the most common events in the pathogenesis of human tumors. Inactivation of the p53 gene may also play a role in SNC. Much of the research in this area has concentrated on persons with papillomas, with or without SNC. Among five papillomas with concomitant squamous cell carcinomas, three (60%) were positive for p53 overexpression, an indicator of likely mutation in the p53 gene (Franzmann et al., 1998). In contrast, none of the 25 papillomas without squamous cell carcinomas demonstrated p53 overexpression. More rare histologic types have been studied even less frequently. Wu et al. reported p53 mutations in 18% (2 of 11) of adenocarcinomas. In a study of seven transitional cell carcinomas, Gotte et al. did not observe any p53 protein overexpression or functionally relevant mutations, although in two cases they did observe markedly reduced expression of the putative tumor suppressor gene, FHIT (Gotte et al., 2000). Prevalence of p53 mutations appears to increase with increasing histologic abnormality. In a biopsy study of inverted papillomas, prevalence of p53 mutations was 0% in non-dysplasias, 57% in those with dysplasia, and 75% in those with squamous cell carcinoma (Caruana et al., 1997). Thus, p53 mutations may play a key role in the neoplastic progression of papilloma to invasive cancer. Several studies also indicate that presence of oncogenic human papillomavirus (HPV) (types 16 or 18) DNA (a putative SNC risk factor) is inversely related to overexpression of p53. In one study, 91% (21 of 23) of sinonasal squamous cell carcinomas without evidence of oncogenic HPV types manifested p53 overexpression, in comparison with none of three with oncogenic HPV types (Buchwald et al., 2001). A second study observed similar results (Caruana et al., 1997). Together, these studies suggest involvement of at least two separate neoplastic pathways, one involving HPV and the other involving p53.

603

604

PART IV: CANCER BY TISSUE OF ORIGIN Table 30–2. Distribution of Sinonasal Cancers by Histologic Type, SEER 9 Registries, 1973–1999 Total Histology Squamous cell Adenocarcinoma Carcinoma, NOS Esthesioneurocytoma and neuroblastoma Melanoma Sarcoma Undifferentiated carcinoma Transitional cell Other

Figure 30–1. Illustration of nasal cavity, paranasal sinuses, and surrounding structures.

DEMOGRAPHIC PATTERNS Incidence and survival data are from the Surveillance, Epidemiology, and End Results (SEER) program of the National Cancer Institute. All cases of SNC identified by nine SEER registries (Connecticut, San Francisco, western Washington, Iowa, Detroit, Atlanta, Utah, New Mexico, and Hawaii) were included. The SEER registries include data on persons diagnosed with cancer between 1973 and 1999, except in Seattle, where the registry started in 1974, and in Atlanta, where the registry started in 1975. Patients with in situ disease were excluded (n = 59). Incidence rates were age-adjusted to the 2000 US standard population. Overall 4025 cases of SNC were identified. Age-adjusted incidence rates between 1973 and 1999 for SNC were 0.73 cases per 100,000 persons per year among men and women of all races (Table 30–3). The male : female rate ratio was 1.7. As is typical of epithelial cancers, risk of SNC increases with age (Fig. 30–2). Among children, SNC is quite rare; between 1973 and 1999, 20 cases of SNC were observed in the nine SEER registry areas in children under 5 years old. From 50 through 80 years of age, incidence rates increase rapidly. In individuals under 20 years of age, incidence rates in females were similar to, or slightly higher than rates in males. Above age 20, however, incidence of SNC in men was nearly double that in women. Incidence among black males was slightly (22%) higher than among white males, whereas there was little difference by race among females.

N

%

2179 474 272 245 205 193 172 55 230

54.1 11.8 6.8 6.1 5.1 4.8 4.3 1.4 5.7

Black males had 0.55 times the rate of nasal cavity cancer and 2.3 times the rate of maxillary sinus cancer compared to white males (Fig. 30–3). The distribution of subsites among black women, and men and women of other races (data not shown) was similar to that of white males. Substantial differences in histologic distribution of SNC by gender were observed in both blacks and whites (Fig. 30–4). Males had approximately twice the risk of squamous cell carcinoma of females. In addition, regardless of gender, rates of adenocarcinoma, sarcoma, and carcinoma, NOS were higher among blacks than whites. Table 30–4 describes incidence rates by SEER registry and race. Rates varied only slightly by registry and were generally higher among black and other races compared to whites.

Time Trends Incidence rates of SNC have held remarkably constant since 1973 for all races together, and for white males and females, although some variation was noted for non-whites (Table 30–5). Although overall rates have remained constant, there have been modest changes over time in the incidence in some of the subsites (Fig. 30–5). Incidence rates of cancer of the nasal cavity have increased slightly between 1973 and 1999, rising from 0.29 per 100,000 (in 1973–1979) to 0.33 per 100,000 (in 1995–1999), whereas rates for maxillary sinus have decreased (from 0.30 per 100,000 to 0.23 per 100,000). The stage at which SNC is diagnosed has also shifted moderately between 1973 and 1999. A greater proportion of sinonasal tumors were diagnosed at regional and distant stage in the time interval 1995 to 1999 (35.4% and 25.8%, respectively) than in 1973 to 1979 (31.2% and 18.8%). A concomitant reduction in the proportion of localized tumors has also been seen during this time period, while the proportion of unstaged tumors has remained stable (data not presented).

Survival The overall 5-year relative survival for all persons diagnosed in the SEER program between 1973 and 1994 was 52.1% (Table 30–6).

Table 30–1. Distribution of Sinonasal Cancer by Subsite, SEER 9 Registries, 1973–1999 Table 30–3. Incidence Rate* of Sinonasal Cancer by Sex and Race, SEER 9 Registries, 1973–1999

Total Subsite*

N

%

1645 146 1460 369 40 119 110 136

40.9 3.6 36.3 9.2 1.0 3.0 2.7 3.4

Total Nasal cavity (300) Middle ear (301) Maxillary sinus (310) Ethmoid sinus (311) Frontal sinus (312) Sphenoid sinus (313) Overlapping lesions of the accessory sinuses (318) Accessory sinuses, NOS (319) *ICD-O codes listed in parentheses.

All races White Black Other†

Male

Female

N

Rate

N

Rate

N

Rate

4025 3322 351 352

0.73 0.71 0.81 0.94

2329 1901 211 217

0.97 0.93 1.13 1.23

1696 1421 140 135

0.56 0.55 0.57 0.68

*Rates are per 100,000 and age adjusted to the 2000 US standard; †Includes 30 people with unknown race.

Figure 30–2. Incidence rate of sinonasal cancer by age at diagnosis, US Surveillance, Epidemiology, and End Results 9 registries, 1973–1999.

Figure 30–3. Incidence rate of sinonasal cancer by primary site, sex, and race, US Surveillance, Epidemiology, and End Results 9 registries, 1973–1999.

Figure 30–4. Incidence rate of sinonasal cancer by histologic site, sex, and race, US Surveillance, Epidemiology, and End Results 9 registries, 1973–1999.

Table 30–4. Incidence Rate* of Sinonasal Cancer by SEER registry and race, 1973–1999 White SEER Registry Nine SEER registries San Francisco/Oakland Connecticut Detroit (Metropolitan) Hawaii Iowa New Mexico Seattle (Puget Sound) Utah Atlanta (Metropolitan)

Other†

Black

N

Rate

N

Rate

N

Rate

3322 496 573 611 63 491 223 485 176 204

0.71 0.72 0.71 0.79 1.09 0.63 0.77 0.72 0.58 0.72

351 65 36 156 1 10 4 13 1 65

0.81 0.75 0.82 0.84 0.16 1.17 0.73 0.87 0.60 0.80

352 106 8 15 168 3 17 25 3 7

0.94 0.93 1.08 1.40 0.94 1.06 1.00 0.72 0.66 0.78

*Rates are per 100,000 person-years and age adjusted to the 2000 US standard; †Includes 30 people with unknown race.

Figure 30–5. Incidence rate of sinonasal cancer by subsite and year of diagnosis, US Surveillance, Epidemiology, and End Results 9 registries, 1973–1999. Note: Regression lines based on a linear model of the incidence of SNC on year of diagnosis.

605

606

PART IV: CANCER BY TISSUE OF ORIGIN

Table 30–5. Incidence Rate* of Sinonasal Cancer by Sex, Race, and Year of Diagnosis in SEER 9 registries, 1973–1999 White Male Year of diagnosis 1973–1979 1980–1984 1985–1989 1990–1994 1995–1999

Other†

Black Female

Male

Female

Male

Female

N

Rate

N

Rate

N

Rate

N

Rate

N

Rate

N

Rate

451 335 348 388 379

0.99 0.96 0.90 0.95 0.86

290 249 299 283 300

0.50 0.54 0.61 0.54 0.53

47 42 34 40 48

1.22 1.40 1.07 0.98 1.07

19 19 29 35 38

0.44 0.49 0.59 0.67 0.67

29 30 52 43 63

1.18 1.18 1.62 1.02 1.25

27 16 34 29 29

0.90 0.64 0.91 0.61 0.51

*Rates are per 100,000 person-years and age adjusted to the 2000 US standard; †Includes 30 people with unknown race.

Five-year relative survival was substantially greater for whites (54.2%) than for blacks (37.6%). This difference was also observed by stage, particularly for localized (75.7% for whites vs. 59.3% for blacks) and distant stage (30.1% for whites and 16.9% for blacks). There were modest improvements over time in 5-year survival for those diagnosed with local or regional stage (Fig. 30–6). Substantial variability in relative survival by histologic type was observed, with the longest survival among those with esthesioneurocytoma, neuroblastoma, and adenocarcinoma, and shorter survival among those with squamous cell carcinoma and undifferentiated carcinomas (Fig. 30–7).

International Patterns Sinonasal cancer exhibits only modest variation in incidence across different populations. Incidence rates are highest in several countries in Africa, the Philippines, Japan, and France (Table 30–7) (International Agency for Research on Cancer, 2002). Countries with incidence rates less than 0.5 per 100,000 person-years among men with more than 10 cases include Algeria, Algiers; Jews living in Israel (born in Israel or born in Europe or America); and Delhi, India. In many cases, countries with high incidence rates (>1/100,000) for men also have high incidence rates among women (>0.6/100,000), but exceptions include Ecuador; blacks living in New Orleans; Chinese in Singapore; Japan; Iceland; Isere and Bas-Rhin France; Slovenia; Flanders, Belgium (excluding Limberg); and Oceania.

ENVIRONMENTAL FACTORS Overview

Although little evidence exists, a role for infectious agents is plausible, given their important role in other head and neck cancers such as Burkitt lymphoma and nasopharyngeal cancer. There are many challenges to studying the etiology of SNC, primarily due to the rarity of both the disease and exposure to suspected risk factors in the population. In early studies, mortality from SNC— rather than incidence—was used as the outcome. However, in studies based on SNC deaths, many cases may be missed because 5-year survival is relatively good. In addition, factors associated with death from SNC may not be the same factors that cause SNC. Furthermore, means by which cases were ascertained—death certificates—may have also led to bias. A study conducted in the United States demonstrated that misclassification on death certificates was common. Among persons who died of cancer within 2 years of SNC diagnosis, SNC was listed as the underlying cause of death on only 57% of death certificates (Percy et al., 1981). However, for this to lead to bias, survival patterns and/or accuracy of underlying causes of death listed on death certificates among individuals in specific occupational cohorts would have to differ appreciably from the survival patterns or accuracy of reporting of death in the general population. Several studies have benefited from existing data on occupational groups to conduct retrospective cohort studies. Cohort studies can facilitate the study of exposures that are uncommon or difficult to document. Yet investigation of a rare disease such as SNC, together with exposures that have a long latent/induction period, necessitates following up a large number of persons for a long period of time to estimate with precision excess mortality or incidence. Additionally, retrospective studies often lack quantitative exposure data or accurate

Many of the major risk factors identified for SNC are occupational in origin, often involving exposure to dusts such as from working with wood, leather, nickel, and possibly textiles and baking ingredients. Cigarette smoking also appears to be an important risk factor, although less so than for other head and neck cancers. Other environmental agents including metals such as chromium, and chemicals such as mineral oils and formaldehyde may also increase the risk of SNC.

Table 30–6. 1-, 5-, and 10-Year Relative Survival of Sinonasal Cancer by Stage in the US SEER Program, 1973–1998 Relative Survival (%) Stage All Localized Regional Distant Unstaged

N

%

1-Year*

5-Year†

10-Year‡

3878 1182 1461 833 402

100.0 30.5 37.7 21.5 10.4

79.4 93.6 79.3 58.2 80.8

52.1 74.7 47.1 26.8 53.5

42.0 63.7 35.4 20.3 38.2

*Diagnosed 1973–1998: Localized, n = 1142; Regional, n = 1416; Distant, n = 787; Unstaged, n = 394; All, n = 3739. †Diagnosed 1973–1994: Localized, n = 951; Regional, n = 1173; Distant, n = 619; Unstaged, n = 324; All, n = 3067. ‡Diagnosed 1973–1989: Localized, n = 751; Regional, n = 873; Distant, n = 430; Unstaged, n = 232; All, n = 2286.

Figure 30–6. Five-year survival after diagnosis of sinonasal cancer, by stage and year of diagnosis, US Surveillance, Epidemiology, and End Results 9 registries, 1973–1999. Note: Regression lines based on a quadratic model of cumulative 5-year survival on year of diagnosis and (year of diagnosis).

Figure 30–7. Cumulative relative survival following sinonasal cancer diagnosis by subsite, US Surveillance, Epidemiology, and End Results 9 registries, 1973–1999.

Table 30–7. Worldwide Incidence Rate (Per 100,000 Person-Years)* of Sinonasal Cancer‡ Men Location Africa Algeria, Algiers Zimbabwe, Harare: African Uganda, Kyadondo County Central and South America Ecuador, Quinto North America US, SEER: White US, SEER: Black US, Michigan, Detroit: Black US, Louisiana, New Orleans: Black Asia Israel: Jews Israel: Jews born in Europe or America China, Hong Kong Singapore: Chinese India, Delhi India, Poona Japan, Hiroshima Japan, Miyagi Prefecture Philippines, Manila Philippines, Rizal Europe Belgium, Limberg Belgium, Flanders (excluding Limburg) Germany, Saarland Iceland Russia, St. Petersburg Sweden France, Bas-Rhin France, Isere Slovenia Switzerland, Geneva Spain, Asturias Oceania US, Hawaii: Japanese Australia, Tasmania

Women

All Sites

Nasal Cavity

Maxillary Sinus

All Sites

Nasal Cavity

Maxillary Sinus

0.4 1.2 1.3

— 0.4 0.5

0.8 0.6 0.5

0.2 0.8 1.0

— 0.0 0.4

— 0.6 0.3

0.3

0.3

0.0

0.7

0.1

0.5

0.6 0.9 1.1 1.2

0.3 0.3 0.4 0.4

0.2 0.4 0.4 0.6

0.4 0.4 0.6 0.1

0.1 0.1 0.2 0.0

0.2 0.2 0.2 0.1

0.3 0.3 0.8 1.0 0.4 1.2 1.2 1.2 1.6 1.2

0.2 0.2 — 0.4 0.2 0.1 0.2 0.1 0.7 0.4

0.1 0.0 — 0.5 0.2 1.0 0.7 0.9 0.8 0.6

0.3 0.3 0.4 0.3 0.3 0.6 0.4 0.5 0.8 0.6

0.2 0.2 — 0.2 0.1 0.0 0.1 0.2 0.4 0.2

0.0 0.0 — 0.1 0.1 0.5 0.2 0.3 0.4 0.3

0.4 1.2 0.4 0.4 0.4 0.4 1.4 1.0 1.0 1.1 1.3

— 0.2 0.1 0.2 0.1 0.2 0.5 0.3 0.2 0.8 0.5

— 0.3 0.1 — 0.2 0.1 0.2 0.2 0.5 0.3 0.0

0.1 0.3 0.4 1.0 0.2 0.3 0.2 0.1 0.4 0.7 0.4

— 0.2 0.2 0.8 0.1 0.2 0.1 — 0.1 0.3 0.1

— 0.1 0.1 0.1 0.0 0.0 — — 0.3 0.3 0.2

0.7 1.2

0.1 0.4

0.4 0.1

0.8 0.3

0.3 0.2

0.3 —

*Rates based on less than 10 cases are in italics; ‡Age-standardized to the year 2000 world population. Source: International Agency for Research on Cancer, 2002.

607

608

PART IV: CANCER BY TISSUE OF ORIGIN

information about potential confounding factors. Conversely, casecontrol studies are efficient and have the ability to collect detailed exposure data on both occupational and non-occupational characteristics such as cigarette smoking or diet, which cannot usually be assessed using existing records. Disadvantages of case-control studies can include low prevalence of exposures and reliance on proxy respondents. In addition, occupational exposures of interest regarding SNC are typically rare in community-based studies, which can preclude examining details of the exposure-disease relationship. Etiologic factors may differ by subsite or histology, but such information is usually only available from population-based registries. For these reasons, meta-analyses or pooled studies often present an efficient means to evaluate details of exposure, but the heterogeneity of study methods and populations can limit their interpretability. Regardless of the study design, misclassification of both disease and exposure are concerns. With the introduction of new International Classification of Disease editions, classification of SNC has changed, potentially affecting comparison of rates across time within a country or between countries. Under the ICD-7, “cancer of the ethmoid” and “cancer of the maxilla(ry)” were classified under bone cancer. Under ICD-8 rules, the former term is classified with SNC but the latter is called bone cancer. The same is true in the ICD-9, with the added complication that the term “cancer of the nose” with no other modifiers is classified under ill-defined sites (Decoufle and Walrath, 1987; Fritz, 2000; Percy et al., 1979). Even given the same coding scheme, inaccuracy of the underlying and contributing cause of death information appears to be relatively common. For example, Olsen noted that in Denmark, up to 10% of cases classified as cancer of the internal nose were found to be of the skin of the nostrils and external nose when reviewed (Olsen, 1987). Because information on exposures have commonly been assessed through secondary sources, such as death certificates, occupational records, and proxy respondents, misclassification of exposure is often a concern. For example, jobs held only early in life may not be listed on the death certificate and may be less likely to be reported by proxy respondents. If those are the jobs (due to a long latent/induction period) more likely to be etiologically related to the cancers, relative risk (RR) estimates would be biased toward the null (Lerchen and Samet, 1986). Ever employed (or employed for at least 3–6 months) has often been the exposure metric; if employment was sometimes brief, this might result in non-differential misclassification and dilute any possible associations. In addition, exposure status assessed in case-control studies may be inaccurate because agents such as nickel, chromium, and formaldehyde are constituents of composite materials and may not be known to employees who are exposed (or their nextof-kin). For efficiency, several jobs were often grouped together. This grouping of jobs may mean that people with different work duties are included in the same occupational group, leading to heterogeneity within categories. Lastly, women’s occupations were often not listed on death certificates, impeding the identification of occupational exposures among women in studies using existing records only.

Occupational Factors Wood Dust Long recognized as a respiratory irritant and capable of producing severe allergic responses, exposure to wood dust has been identified as a carcinogen in the nasal cavities and paranasal sinuses as well. Dust from the processing of wood is commonly encountered: over 600,000 persons in the United States are occupationally exposed, and many more come in contact with wood dust through hobbies (International Agency for Research on Cancer and World Health Organization, 1995). Operations performed on wood can create dust in two ways: by shattering wood cells and by chipping out whole cells (Hinds, 1988). Shattering, which occurs almost exclusively during sanding operations, for example, results in a finer particle size and higher dust concentration than does chipping, which is more likely to occur in sawing and milling operations (Scheeper et al., 1995; Teschke et al., 1999). In addition, wood that has densely packed cells (i.e., hardwood) and

wood that is less elastic (i.e., dry wood) are more likely to be shattered during processing. Thus the amount and type of dust produced depends largely on the type of wood involved and the specific processes it undergoes. Clinical observations in the 1960s (Macbeth, 1965) in the High Wycombe area near Oxford, England suggested that workers involved in the manufacture of furniture from hardwoods had substantially increased risks of sinonasal adenocarcinoma, later estimated to be approximately 500-fold (Acheson et al., 1968). A subsequent study of cases of SNC occurring in England and Wales outside of the Oxford area confirmed a high relative risk nationwide for adenocarcinoma among furniture workers (RR = 95) and also observed a fivefold relative risk among other woodworkers, principally carpenters and joiners (Acheson et al., 1972). A large number of studies have examined the association between exposure to wood dust and SNC since the first studies in England. These studies have addressed a range of research questions, including: (1) how does the risk of SNC differ by occupation and wood dust concentration, (2) are the associations observed in England also present in other countries, (3) have excess risks declined over time, and (4) is the association between wood dust and SNC limited to adenocarcinoma, or are other histologic types affected? In 1994, the International Agency for Research on Cancer (IARC) reviewed studies addressing whether woodworking processes and exposures other than those entailed in furniture manufacturing result in an increased risk of SNC (International Agency for Research on Cancer and World Health Organization, 1995). Consistent increases in risk of adenocarcinoma, often 10-fold or higher, were observed for a number of occupations, including carpenters, woodworking machine operators, and cabinet makers. In most instances, RRs tended to increase with increasing duration of work and with increasing likelihood of exposure to high concentrations of wood dust. In a pooled reanalysis of 12 case-control studies of SNC from Europe, the United States and China, any employment in a primary wood industry job (defined as logging, pulp preparation, and sawmills) that entailed moderate or high wood dust exposure was associated with a RR = 14.8 (95% CI: 8.7–25.4) for adenocarcinoma among men (Demers et al., 1995). When analyzed by duration of employment, risk was estimated to increase by 8% (95% CI: 6%–11%) for each year of employment. Thus there is compelling evidence that a high level of exposure to wood dust in a variety of occupational settings, not just furniture manufacturing, substantially increases risk of sinonasal adenocarcinoma. However, it should be noted that the increases observed among contemporary woodworkers, as well as among woodworkers outside of Europe, are not as strong as those observed among the British furniture makers of the early 20th century. This is likely explained by differences among countries in types of wood used and other aspects of the manufacturing processes, as well as changes in industrial practices that decreased occupational exposures to wood dust over time (Jones and Smith, 1986; Teschke et al., 1999). Another issue is whether increased risks associated with woodworking extend to sinonasal cancers other than adenocarcinoma. In a pooled reanalysis of 12 studies, risk of squamous cell carcinoma was found to be only slightly increased for those ever exposed at a moderate or high level in a primary wood industry (OR = 1.2; 95% CI = 0.8–1.8), with an estimated 1% (95% CI: 0%–3%) increase in risk per year of employment (Demers et al., 1995). For the 11 cases exposed for 30 or more years, the OR was estimated to be 2.4 (95% CI: 1.1–5.0). However, substantial heterogeneity in findings was observed among the case-control studies. Together with the modest and inconsistent increases in risk observed, the evidence is not yet sufficient to consider squamous cell carcinomas of the sinonasal cavity to be conclusively linked with exposure to wood dust. While the ability of occupational exposure to wood dust to cause sinonasal adenocarcinomas in humans may be clear, the underlying mechanisms are not well understood. Three main hypotheses have been posited: impairment of mucociliary clearance, a direct effect of wood dust particles, and chronic inflammation. A large fraction of inhaled dust, particularly dust greater than 5 mm in diameter, is deposited in the nasal epithelium, which can reduce

Cancers of the Nasal Cavity and Paranasal Sinuses mucociliary clearance or even cause mucostasis (Andersen et al., 1977; Black et al., 1974). This impairment of mucociliary clearance, an important defense mechanism by which environmental contaminants are removed, can result in longer contact time between proliferating cells and mutagens and mitogens adherent to inhaled particulates (such as carcinogens in tobacco smoke). The wood dust particles themselves contain a number of different chemicals, the distribution of which is highly dependent on the type of wood. Several components of beech and oak, for example, are mutagenic in bacterial assays and can produce single-stranded chromosomal breaks in animal studies (International Agency for Research on Cancer and World Health Organization, 1995). Finally, and perhaps most importantly, long-term occupational exposure to wood dust can cause chronic inflammation of the sinonasal epithelium, and may increase cancer risk through the production of reactive oxygen species and by increasing the cellular proliferation rate (Cohen and Ellwein, 1991).

Leather Dust and Shoe Industry Employment in boot and shoe production has been strongly associated with SNC, with RRs in excess of 10, particularly for adenocarcinomas. Exposure to leather dust is thought to be the main carcinogen, although chemicals involved in tanning and leather manufacturing as well as wood fibers in fiberboard used to build up soles and heals are also suspected causes of SNC. Some chemical exposures in this industry such as formaldehyde and tannins are thought to be carcinogenic and may overlap with those found in woodworking.

Shoe and Boot Manufacture. Relative risks of SNC associated with shoe and boot manufacturing industry have varied, with the strongest associations observed in England, Wales, and Italy. A small study conducted in Belgium (Debois, 1969), followed by a larger one in the Northamptonshire area of England (Acheson et al., 1970) were the first published reports of increased risks of adenocarcinoma of the nasal cavity and sinuses among workers in the shoe and boot industry. Of the men diagnosed with SNC in Northamptonshire between 1953 and 1967, 59% worked in the boot and shoe industry, compared with 17% of the adult male population. Less than one adenocarcinoma (0.2) was expected, and seven were observed, yielding a relative risk of 35. Relative risks associated with boot and shoe manufacturing for squamous cell carcinoma (O/E = 7/1.4 = 4) and transitional cell carcinoma (O/E = 3/0.4 = 7.5) were also elevated, though not as high. Those who worked in the dustiest areas of the factory (e.g., sorting and cutting leather, and in the finishing room) had higher relative risks than those who worked in less dusty jobs. Other case-control (Cecchi et al., 1980; Merler et al., 1986) and cohort studies (Fu et al., 1996; Olsen, 1988; Pippard and Acheson, 1985) have observed positive associations with leather exposure, with stronger associations for those with higher purported dust exposures. In a mortality study of more than 5000 men known to have been employed in the shoe and boot industry in Great Britain in 1939, 10 cases were observed (vs. 1.87 expected) (Pippard and Acheson, 1985). Relative risks of SNC were elevated in shoemakers from England and Italy followed between 1939 and 1991 (English cohort: n = 4215; 12 observed, SMR = 8.1, 95% CI: 4.2–14.1; Italian cohort: n = 2008; 1 observed; SMR = 12.5, 95% CI: 0.3–69.7) (Fu et al., 1996). One small case-control study conducted in Italy observed particularly high ORs (OR = 47, 95% CI: 8.7–255.1) for SNC among men exposed to leather dust compared to those not exposed (Merler et al., 1986). In comparison with the unexposed category, sex-adjusted ORs were 7.5 (95% CI: 1.8–31.7) and 121 (95% CI: 17.3–844.3) for intermediate and heavy exposure, respectively (Merler et al., 1986). Results from studies conducted in the United States (Decoufle and Walrath, 1983; Decoufle and Walrath, 1987; Garabrant and Wegman, 1984; Walker et al., 1993; Walrath et al., 1987), Europe (Baxter and McDowall, 1986; Hayes et al., 1986a; Hernberg et al., 1983; Luce et al., 1993) and Japan (Takasaka et al., 1987), however, failed to observe increased risks, or observed associations weaker in magnitude than those seen in earlier studies. Type of oil used in finishing vegetabletanned leather has been presented as a possible reason for differences in results for studies conducted in the United States and the United

609

Kingdom. In the United States, animal oils are primarily used, while in the United Kingdom, mineral oils are used, which are considered carcinogenic (International Agency for Research on Cancer and World Health Organization, 1987; Randell, 1990). Exposure to leather dust was associated with an excess risk (OR = 2.7, 95% CI: 0.8–9.4) in a pooled reanalysis of studies among European men and women including 555 cases and 1705 controls diagnosed between 1970 and 1989 (t Mannetje et al., 1999). The association appeared stronger for adenocarcinomas (OR = 3.0, 95% CI: 1.3–6.7) than for squamous cell carcinomas (OR = 1.5, 95% CI: 0.7–3.0) (t Mannetje et al., 1999). According to IARC, boot and shoe manufacture and repair is considered a group 1 carcinogen (International Agency for Research on Cancer, 1981; International Agency for Research on Cancer and World Health Organization, 1987).

Tanning. Tanning is generally a wet process with few dusty operations, but chemicals that are used to convert hides and skins to leather may increase the risk of SNC. Chromium salts (along with natural tannins), which are used to process leather for handbags and upper parts of shoes, have been implicated as possible etiologic agents for the associations observed between SNC and leather tanning (Battista et al., 1995). While studies on leather tanning suggest an increased risk of SNC, most effect estimates have been less than five with wide confidence intervals (International Agency for Research on Cancer, 1981). The International Agency for Research on Cancer concluded that there was inadequate evidence for leather tanning and processing to be considered carcinogens. IARC was therefore not able to classify this exposure regarding its carcinogenicity to humans (group 3) (International Agency for Research on Cancer, 1981; International Agency for Research on Cancer & World Health Organization, 1987). Nickel The principal use of nickel is in its metallic form combined with other metals (such as iron, copper, chromium, and zinc) and nonmetals (chlorine, sulfur, and oxygen) as alloys. Major current uses of nickel are in the production of stainless and heat-resistant steels, nonferrous alloys, and superalloys that are used in electroplating, as catalysts, in the manufacture of alkaline (nickel-cadmium) batteries, coins, welding products, and in certain pigment and electronic products (International Agency for Research on Cancer and World Health Organization, 1990). In occupational settings, exposure to nickel may occur by skin contact or by inhalation of dusts, fumes, or mists containing nickel. A study published in the 1930s first identified a cluster of SNC cases among nickel refinery workers in Wales (Sunderman, 2001). Subsequent epidemiologic studies have generally confirmed a carcinogenic role for nickel, which is an irritant to nasal passages and the respiratory tract (Agency for Toxic Substances and Disease Registry, 1997). Increased incidence of respiratory cancer in nickel workers has been observed throughout the world, including in Germany, Japan, Russia, South Wales, Canada, and Norway (Brinton et al., 1984; Doll et al., 1990; Doll et al., 1970; Hernberg et al., 1983; Jarup et al., 1998; Magnus et al., 1982; Roberts et al., 1989), both in plants where the nickel carbonyl Mond process or electrolytic extraction is used (Doll et al., 1990). Since the initial discovery of increased risks among nickel workers, studies have attempted to identify the particular forms of nickel, processes and exposure levels that constitute a hazard, and also whether the excess risks have decreased or have been eliminated over time through changes in production practices. These questions have been challenging to answer because it is difficult to differentiate precisely among nickel species, and few good animal models with relevant routes of exposure (inhalation) exist. In addition, most epidemiologic studies have had insufficiently detailed exposure data to provide an understanding of the amounts and forms of nickel that are associated with increased cancer risks. Within these limitations, the emerging epidemiologic picture has pointed away from nickel carbonyl process and toward exposure to dust from preliminary processes, particularly calcination of impure nickel copper sulfides to nickel copper oxide (Doll et al., 1977). Risks appear to have decreased

610

PART IV: CANCER BY TISSUE OF ORIGIN

drastically since 1925, but some elevation in risks remain (Doll et al., 1990). Doll and colleagues observed some of the strongest associations between occupational exposure to nickel and SNC in a South Wales nickel refinery. Among employees who worked at least 5 years between 1929 and 1949 (and who were observed until 1972), the risk of death from carcinoma of the nasal sinus was 300 to 700 times the national average for those who started working before 1920, and about 100 times the national average for those first employed between 1920 and 1925 (Doll et al., 1977). Overall, among 1348 men first employed before 1930, there were 74 nasal cancer deaths (SMR = 211.2, 95% CI: 165.8–265.1) (Doll et al., 1990). The highest risks were associated with calcining, furnaces, and copper sulfate production (Doll et al., 1990). The relative risks increased sharply with increasing age at first exposure, but remained roughly constant throughout the period of follow-up (Kaldor et al., 1986). In Wales, risks decreased following changes in the refinery process (Doll et al., 1977), including modifications to the calciners to reduce dust exposure and introduction of personal protective equipment such as cotton face pads (Doll et al., 1977). Studies in Canada, England, Finland, Norway, and Sweden also observed large increases in relative risks among nickel refinery workers, though somewhat lower than in the study noted above. A cohort study conducted among 54,509 men involved in the extraction and refining of copper and nickel in Ontario, Canada between 1950 and 1985 observed standardized mortality ratios for SNC deaths of 36.1 (95% CI: 13.3–78.9) and 77.8 (95% CI: 46.8–121.4) based on 6 and 19 observed deaths (vs. 0.17 and 0.24 expected deaths) due to SNC, respectively, in men employed in the sinter plants in Sudbury, Ontario, or the leaching, calcining, and sintering department at Porte Colborne, Ontario (Doll et al., 1990; Roberts et al., 1989). Several studies were conducted in Norway among a cohort of 4764 male nickel refinery workers (Andersen et al., 1996; Magnus et al., 1982; Pedersen et al., 1973). Thirty-two cases of SNC were observed where only 1.8 were expected (SIR 18.0, 95% CI: 12.3–25.4) (Andersen et al., 1996). Risk of SNC increased with increasing exposure to nickel. The association was particularly strong for nickel oxide, and somewhat weaker for soluble nickel. In addition, in an earlier analysis limited to 2247 men, risk decreased with increasing calendar year; the RR among those first employed between 1916 and 1929 (and 25 to 34 years after their first employment) was over 100 (O/E = 3/0.022 = 136) and more than six times greater than the RR of SNC among men first employed between 1950 and 1959 (O/E = 1/0.052 = 19) (Magnus et al., 1982). Conversely, in a cohort study of workers at a plant producing nickel from silicate oxide ore (Goldberg et al., 1994), no SNC deaths were observed. This may be because the processing method of the silicate oxide ore or the content of the ore itself may confer a lower carcinogenic risk. Some recent studies suggest that RRs remained elevated even for those employed after the 1930s. Among 418 workers employed for at least three months between 1945 and 1985 in a nickel refinery in Finland, those who were exposed primarily to soluble nickel sulfate at levels below 0.5 mg/m3 as well as to low concentrations of other nickel compounds had a substantially increased risk of SNC (SIR 41.1, 95% CI: 5.0–148), positively associated with increasing latency and duration of employment (Anttila et al., 1998). The risk was elevated despite the fact that measured concentrations were generally lower than levels considered to be safe. Nickel smelter workers exposed to low levels of non-soluble nickel and to other metals and chemicals did not have elevated RRs. The association between non-refinery exposures to nickel (e.g., nickel alloy and stainless steel production, nickel/chromium plating, die-casting, and electroplating) and SNC has been less consistent. In a cohort of 2689 nickel/chromium platers in England and Wales observed between 1946 and 1983, the association was weakly positive. One death from SNC was observed, where only 0.3 were expected (Sorahan et al., 1987). In a cohort study in Sweden of 869 male battery workers exposed to nickel hydroxide and cadmium oxide, three cases of SNC were observed where only 0.36 were expected (SIR 8.3, 95% CI: 1.7–24.3); the risk remained significantly elevated after

applying a 10-year latency period (Jarup et al., 1998). However, because of the concomitant exposure to cadmium and nickel, either or both exposures may be causal agents. While one case-control study observed a 2.4-fold increased risk (95% CI: 0.9–6.6) associated with “welding, flame-cutting, soldering” chromium and nickel (Hernberg et al., 1983), these results were not confirmed in case-control studies in France or the United States (Brinton et al., 1984; Luce et al., 1993; Roush et al., 1980). In the former study, all but one of the nickelexposed cases had also been exposed to chromium, so confounding is a possibility (Hernberg et al., 1983). Several cohort studies have also failed to observe associations (Cornell and Landis, 1984; Cragle et al., 1984; Egedahl et al., 2001; Enterline and Marsh, 1982; Redmond, 1984; Shannon et al., 1991). In one of the largest negative studies, only one nasal cancer was observed (vs. 0.6 expected, SMR = 1.7, 95% CI: 0.04–9.3) among 11,567 nickel workers in Ontario engaged in mining, milling, and smelting and followed from 1950 until 1984, providing little evidence of increased death rates due to SNC (Shannon et al., 1991). The majority of these cohort studies were among nickel workers who were involved in other, apparently lower-risk processes. The average concentrations of nickel dust in mining and smelting is usually considerably higher than those encountered in refining operations because of the higher nickel content of the materials being handled in mining and smelting processes (International Agency for Research on Cancer and World Health Organization, 1990). Insufficient sample size and/or follow-up may have limited the power to observe relatively small excesses in mortality or incidence present among those first employed after 1930. Other environmental contaminants associated with nickel ore and its contents (e.g., arsenic, sulfur dioxide, and asbestos) have been hypothesized as possible reasons for heterogeneity in study results (Langer et al., 1980; Meininger et al., 1982). Analysis at one major smelter in New Caledonia indicated that nickel ores from that area were heavily contaminated with asbestos fibers (Langer et al., 1980). Thus, while nickel alone may be sufficient to cause cancer, its effects may be modified by the presence of other contaminants. Suggested carcinogenic mechanisms include binding to DNA and to nuclear proteins, and generation of reactive oxygen species. Nickel compounds may generate nickel ions at critical sites in target cells. After reviewing the epidemiologic data, Doll and colleagues contended that the increased risks of SNC were due to inhaling soluble nickel compounds (primarily nickel sulfate and nickel chloride carbonate) and certain insoluble nickel compounds (e.g., metallic nickel, oxidic nickel, and sulphidic nickel) (Doll et al., 1990). Based in part on its association with SNC, nickel compounds are considered a group 1 carcinogen to humans, and metallic nickel is considered possibly carcinogenic to humans (International Agency for Research on Cancer and World Health Organization, 1990).

Chromium Chromium is currently used in stainless, tool, and alloy steels (including those based on nickel and iron-nickel), heat- and corrosionresistant materials, special-purpose alloys, alloy cast iron pigments, metal plating, leather tanning, various industrial chemicals, and refractory materials for metallurgical furnaces. Hexavalent chromium (or chromium [VI]) has been implicated in the etiology of lung cancer (see Chapter 33), and is a suspected cause of SNC, although the evidence is weaker. Materials that contain chromium [VI] include paint and primer pigments, graphic arts supplies, fungicides, wood preservatives, and corrosion inhibitors (International Agency for Research on Cancer and World Health Organization, 1990). Occupational exposures to airborne dusts containing chromium metal may occur during production, welding, cutting, and grinding of chromium alloys. Hexavalent chromium is corrosive and may cause irritation to nasal membranes (Agency for Toxic Substances and Disease Registry, 2000). Cases of SNC have been reported in studies of primary chromate production workers in Japan, the United Kingdom, and the United States; chromate pigment production workers in Norway; and chromium platers in the United Kingdom, indicating a pattern of excess risk (International Agency for Research on Cancer and World Health Organization, 1990).

Cancers of the Nasal Cavity and Paranasal Sinuses In a cohort study of 2715 chromate production workers employed for at least 1 year between 1948 and 1977, two deaths from SNC occurred where only 0.28 were expected (SMR = 7.1, 95% CI: 0.9–25.8) (Alderson et al., 1981). In three British chromate plants, among workers exposed for at least 1 year between 1948 and 1977, two deaths from SNC occurred at one of the plants (vs. 0.15 expected, O/E = 13.3) (Davies et al., 1991). Similar excess risks of SNC death were observed in other cohort studies of occupational exposure to chromium (Enterline, 1974; Satoh et al., 1981). A case-control study conducted in France found no significant associations between chromium compounds or hexavalent chromium and SNC (Luce et al., 1993). In contrast, case-control studies in Denmark (see Nickel section) and the United States observed elevated relative risks of SNC (Brinton et al., 1984; Hernberg et al., 1983). Men with chromium/chromate exposure had 5.1 times (P < 0.05) the risk of SNC compared to men without chromium exposure (based on five exposed cases). The excess was associated mainly with use of chromate products in the building industry and in painting, rather than in the manufacture of chromates (Brinton et al., 1984). The IARC Working group determined that there was sufficient evidence (group 1) in humans for the carcinogenicity of chromium [VI] compounds as encountered in the chromate production, chromate pigment production, and chromium plating industries, based in part on data on SNC (International Agency for Research on Cancer and World Health Organization, 1990).

Mineral Oils Oils derived from petroleum are used in a wide variety of occupational settings and applications, including metalworking, print press operating, and cotton and jute spinning. In metalworking operations, fluids (commonly described as cutting oils) are applied to the working surface in a spray or stream, to cool or lubricate the metals being worked. The oil mist that is generated can result in dermal or inhalational exposure. A few case-control studies have investigated the association between exposure to mineral oils and SNC, with inconsistent results. Those with a job title associated with exposure to cutting oils had a 2.8-fold (95% CI: 1.4–5.7) increased risk of SNC (Roush et al., 1980). A French study did not observe an association with adenocarcinoma or squamous cell carcinoma, but did observe a 2.9-fold increased risk (95% CI: 1.0–8.0) for other types of SNC (Luce et al., 1993). Finally, a recent US study among US men who served in Vietnam failed to observe an association (Zhu et al., 2002). Known or suspected carcinogens in mineral oil formulations include polycyclic aromatic hydrocarbons and nitrosamines. However, the chemical composition of the oils varies by the type of petroleum used. Mineral oils derived from refined petroleum contain reduced amounts of polycyclic aromatic hydrocarbons and other impurities. Untreated or mildly treated mineral oils are classified by IARC as a group 1 carcinogen based on a number of cancer sites, including SNC. Highly refined oils are not classified by IARC (group 3) due to inadequate evidence for carcinogenicity in humans (International Agency for Research on Cancer and World Health Organization, 1987).

Formaldehyde Formaldehyde is used mainly in the production of resins, which have widespread use in producing adhesives and binders for the wood, plastics, leather, and related industries (International Agency for Research on Cancer and World Health Organization, 1995). Thus, employment in particleboard mills, wood-working, foundries, textiles, and leather industries often involves significant exposure to formaldehyde. However, because of the strong association of SNC with wood dust, and the frequency with which exposure to formaldehyde and wood dust occur together, it has been difficult to assess an independent effect of formaldehyde on SNC risk (Luce et al., 1993). Given the long and widespread use of formaldehyde in industrial production, and its extensive use in residential products, the discovery of its ability to cause nasal cancer in rats exposed to high doses of formaldehyde (>10 ppm) prompted widespread concern (Swenberg et al., 1980). However, studies in mice and monkeys, who are not obli-

611

gate nasal breathers, failed to observe similar high rates of cancer, suggesting that formaldehyde’s effects may be species-specific (International Agency for Research on Cancer and World Health Organization, 1995; Kerns et al., 1983). In addition, given the non-linear nature of the biologic data and inter-species differences in respiration, extrapolating results seen among rats exposed to high doses to humans exposed to lower doses may not be appropriate (Howlett et al., 1989). Nevertheless, observational data in humans suggests that formaldehyde can have adverse effects on nasal mucosa, leading to inflammation, increased cell proliferation, and dysplasia (Boysen et al., 1990; International Agency for Research on Cancer and World Health Organization, 1995). Formaldehyde is thought to be weakly genotoxic, with cross-linking properties; it also induces micronuclei in the nasal mucosa of exposed workers, a possible indicator of elevated cancer risk (Ballarin et al., 1992; Burgaz et al., 2001; Ma and Harris, 1988). Several mortality studies have been conducted among individuals with formaldehyde exposure, including workers exposed to formaldehyde in a large chemical-producing plant in Massachusetts, embalmers in New York State and California, pathologists in England, workers in the garment industry exposed to formaldehyde, workers in plastics manufacturing and in a research and development facility in New Jersey, and anatomists. None of these studies observed excess mortality due to SNC (Blair et al., 1986; Dell and Teta, 1995; Gardner et al., 1993; Hall et al., 1991; Harrington and Oakes, 1984; Hayes et al., 1990; Levine et al., 1984; Marsh, 1982; Stayner et al., 1985; Stroup et al., 1986; Walrath and Fraumeni, 1983; Walrath and Fraumeni, 1984). Limitations of these studies include insufficient power, as well as relatively low exposure to formaldehyde in cohort members. In studies that have observed an increased risk, residual confounding by other factors, notably wood dust, is a concern. Several casecontrol studies have more frequently observed associations between occupational exposures to formaldehyde and SNC, typically in the range of 2- to 3-fold increased risks (Hayes et al., 1986b; Olsen and Asnaes, 1986), although other studies have not (Roush et al., 1987b; Vaughan et al., 1986). Associations were typically stronger for adenocarcinomas, when histologic subtypes were analyzed separately. Risk of SNC among Danish men in 265 companies exposed to formaldehyde was 2.3-times higher (95% CI: 1.3–4.0; 13 cases) compared to men in the general population. Blue collar workers with no probable exposure to wood dust had a similarly elevated risk (SPIR = 3.0, 95% CI: 1.4–5.7) (Hansen and Olsen, 1995). In a study of all cases of SNC (n = 525) diagnosed in Denmark between 1970 and 1979, men with a job title indicating certain exposure to formaldehyde had a 2.8fold increased relative risk (95% CI: 1.8–4.3), which was reduced after adjustment for wood dust exposure (RR = 1.6, 95% CI: 0.7–3.6) (Olsen et al., 1984). The authors suggested that the results were in accordance with an additive effect of formaldehyde and wood dust. A recently published pooled reanalysis of 12 case-control studies observed a statistically significant elevated risk (men: OR = 3.0, 95% CI: 1.5–5.7; women: OR = 6.2, 95% CI: 2.0–19.7) of adenocarcinoma associated with exposure to formaldehyde (Luce et al., 2002). In analyses limited to subjects never exposed to wood or leather dust, the risk of adenocarcinoma associated with high formaldehyde exposure remained statistically significantly elevated for women only (men: OR = 1.9, 95% CI: 0.5–6.7; women: OR = 11.1, 95% CI: 3.2–38.0) (Luce et al., 2002). Risk of squamous cell carcinoma was weakly associated with formaldehyde exposure (men: OR = 1.2, 95% CI: 0.8–1.8; women: OR = 1.5, 95% CI: 0.6–3.8) (Luce et al., 2002). These summary estimates were heavily influenced by the strong association observed in a single French study for adenocarcinomas (Luce et al., 1993). Among those exposed to wood dust and formaldehyde, risk of adenocarcinoma was high (OR = 692, 95% CI: 92.0–5210), compared to those who were exposed to wood dust but not formaldehyde (OR = 130, 95% CI: 14.0–1191), or formaldehyde alone (based on 4 cases: OR = 8.1, 95% CI: 0.9–73.0). Three groups of authors have published meta-analyses on the association between formaldehyde and SNC (Blair et al., 1990; Collins et al., 1997; Partanen, 1993), coming to somewhat different conclusions. Based on results from 12 studies, low- to medium-level or duration of exposure was associated with a non-statistically significant 10% to

612

PART IV: CANCER BY TISSUE OF ORIGIN

20% increased risk in the first two meta-analyses (Blair et al., 1990; Partanen, 1993). Substantial exposure was not associated with SNC in the meta-analysis by Blair and coworkers (Blair et al., 1990), but was associated with a 70% increased risk (based on log-Gaussian risk ratio model, 95% CI: 1.0–2.8) in the meta-analysis by Partanen (Partanen, 1993). Collins and coauthors conducted a third meta-analysis including both published and unpublished results. Based on data from cohort studies, exposure to formaldehyde was associated with a decreased risk of SNC (meta RR = 0.3, 95% CI: 0.1–0.9). In contrast, summary results from the case-control studies indicated that exposure to formaldehyde was associated with an increased risk of SNC (meta RR = 1.8, 95% CI: 1.4–2.3). The authors noted considerable heterogeneity in the case-control studies, with the United States case-control studies having a meta RR of 1.0 (95% CI: 0.7–1.5), while the European studies had a meta RR of 2.9 (95% CI: 2.2–4.0). Collins et al. surmised that the case-control studies represented relatively low levels of exposure compared to the levels in cohort studies, and that a substantial number of subjects who were considered exposed in the casecontrol studies were probably misclassified (Collins et al., 1997). Such misclassification may be non-differential, leading to an underestimate of the true relative risk estimates. Cohort studies may also be unlikely to detect associations because death from SNC (rather than incidence) is particularly rare. It is also possible that differences in results from case-control and cohort studies are due to confounding by other factors associated both with formaldehyde exposure and SNC, which may be more likely in case-control studies, resulting in a bias away from the null. Formaldehyde does not penetrate the nasal sinuses, where most human SNCs originate (Feron et al., 2001). Therefore, it is possible that the ability to detect an association is hindered by disease heterogeneity in studies that included cases arising in sites other than the nasal cavity. Given the possible non-causal explanations and null studies, some doubt remains about a causal association between formaldehyde and SNC. Accordingly, IARC has concluded that there is limited evidence in humans for the carcinogenicity of formaldehyde, and classified it as a group 2a (probable) carcinogen (International Agency for Research on Cancer and World Health Organization, 1995).

Textiles Textile workers are exposed to both natural and synthetic textile dusts in the manufacturing process. Certain fibers are made from plant materials (e.g., cotton, linen, and rayon) and may produce similar exposures to those encountered by furniture workers and cabinet makers (Teschke et al., 1997). Treatment for crease resistance is the most widely used process for cellulosic textiles (e.g., cotton and viscose), which involves application of formaldehyde-based resins to produce “permanent press” fabrics (International Agency for Research on Cancer, 1990). Increased risks of SNC associated with employment in the textile industry have been observed throughout the world, including the United Kingdom, France, Italy, Sweden, the United States, and Hong Kong (Acheson et al., 1972; Acheson et al., 1981; Bimbi et al., 1988; Brinton et al., 1985; Bross et al., 1978; Comba et al., 1992; Dubrow and Gute, 1988; Luce et al., 1997; Malker et al., 1986; Ng, 1986; Olsen, 1988; Teschke et al., 1997). Relative risks have typically been less than 10. In a study conducted in Sweden, employment in the textile industry was associated with elevated risks of SNC among men (SIR = 2.4, 8 cases, P < 0.05) and women (SIR = 3.0, 3 cases, P > 0.05) (Malker et al., 1986). In a case-control study conducted in North Carolina and Virginia, female cases were more likely to report employment in the textile industry (OR = 2.1, 95% CI: 1.1–4.3) than controls; employment in the textile industry was similar among male cases and controls (OR = 0.8, 95% CI: 0.4–1.7) (Brinton et al., 1985). However, women who had worked less than 10 years were at higher risk (OR = 3.9, 95% CI: 1.4–11.0) than women who had worked for 10 years or more (OR = 1.4, 95% CI: 0.4–3.8). Others (Acheson et al., 1981; Malker et al., 1986; Olsen, 1988) who analyzed the association between employment in the textile industry and SNC separately among men and women, also observed stronger associations among

women. In a case-control study in Hong Kong in which SNC cases were compared to persons with other types of cancers, those employed in the textile industry had a nearly threefold increased risk (95% CI: 1.1–7.9); the association was even stronger for those who had worked more than 15 years (OR = 7.4, 95% CI: 1.2–45.1) (Ng, 1986). An association was also observed in a case-control study conducted in British Columbia including 48 cases. Those who were ever employed in the textile industry had 7.6 times the risk of SNC (95% CI: 1.4–56.6) compared to those never employed. After excluding from analysis the most recent 20 years of employment, ORs remained elevated (OR = 5.0, 95% CI: 0.8–43.0) (Teschke et al., 1997). In a pooled reanalysis of 12 case-control studies in seven countries, elevated risks were observed for squamous cell carcinoma among male fiber preparers (OR for ever employed = 5.1, 95% CI: 1.3–19.2), but the association was limited to those who had worked 10 years or less (OR = 13.5, P < 0.05) (Leclerc et al., 1997). No female cases were exposed. Odds ratios of squamous cell carcinoma were not elevated among spinners, weavers, knitters, or bleachers, or work in the textile industry. Female weavers (OR = 4.0, 95% CI: 0.97–16.2) and textile workers (OR = 2.6, 95% CI: 1.03–6.6) had elevated risks of adenocarcinoma (Leclerc et al., 1997). There were too few exposed cases to examine this association among men. Although data from Brinton and colleagues suggested a stronger association with cotton dust (Brinton et al., 1985), no specific effect of any textile fiber was found in the pooled reanalysis (Luce et al., 2002). Employment in the textile industry appears to be associated with adenocarcinomas among women, though elevated risks of squamous cell carcinoma have been observed among men. In 1990, IARC concluded that there was limited evidence that working in the textile manufacturing industry entailed a carcinogenic risk and classified work in this industry as possibly carcinogenic to humans (group 2b) (International Agency for Research on Cancer, 1990).

Radium Dial Painting Beginning in 1917, a self-luminescing paint was used for clock faces, military equipment, and wrist watches. This paint was made by combining a small quantity of radium and zinc sulfide with a glue binder (Brues and Kirsh, 1977; IARC Working Group on the Evaluation of Carcinogenic Risks to Humans et al., 2001). Until 1925, those employed in this industry (predominantly young women) applied the paint by first pointing the paintbrush with their lips. Studies of cancer risk among radium watch-dial painters in the United States, some of whom ingested radium-226, often in combination with radium-228, showed increases in SNC risk (Brues and Kirsh, 1977; Hasterlik et al., 1964; Polednak et al., 1978). It is believed that exposure to the radioactive paint resulted in migration of radium-222 (a decay product of radium-226) to the paranasal sinuses and mastoid process, exposing the sinonasal region to alpha-particles (Hasterlik et al., 1964).

Other Occupational Factors Increased risk of SNC has been observed among persons with occupations not already mentioned, such as fishermen (Ng, 1986), bakers and pastry cooks (Acheson et al., 1981; Luce et al., 1992; Malker et al., 1986), food processors (Malker et al., 1986; Olsen, 1988), plumbers (Malker et al., 1986), those involved in the manufacture of isopropyl alcohol by the strong acid process (Eckardt, 1974; Hueper, 1966), and hairdressers (Teschke et al., 1997). Several studies have also suggested an increased risk of squamous cell carcinoma among farm workers, but the results have been inconsistent (Bimbi et al., 1988; Fukuda et al., 1987; Luce et al., 1992; Malker et al., 1986; Ng, 1986; Olsen, 1988; Takasaka et al., 1987). Vietnam veterans thought to be exposed to pesticides such as Agent Orange and 2 methyl-4chlorophenoxyacetic acid were not observed to have an increased risk of nasal cancer (1990). However, in a recent case-control study, men exposed to pesticides containing 2,4,5-T had a 5.9-fold increased risk (95% CI: 1.5–23.7) (Zhu et al., 2002). In a pooled reanalysis of 12 case-control studies of SNC, an increased risk of squamous cell carcinoma was observed for women accountants; male sculptors, painters, photographers, and related creative artists; male orchard, vineyard, and related tree and shrub crop

613

Cancers of the Nasal Cavity and Paranasal Sinuses farmers and workers; male stationary engineers and related equipment operators; and male cooks and food preservers (Leclerc et al., 1997). However, despite pooling studies, some results were based solely on exposed cases from a single study. In analyses limited to those with adenocarcinoma who did not have occupational exposure to wood or leather dust, a significantly elevated odds ratio was observed for male technical salesmen, commercial travelers, and manufacturing agents; general farmers and farm workers; food and beverage processors; and transport equipment operators. ORs for farm workers and transport equipment operators were elevated only among those employed for less than 10 years (Leclerc et al., 1997). Due to small numbers, risk patterns associated with various occupations were difficult to discern among women (Leclerc et al., 1997). Some details of occupational exposures associated with SNC risk are summarized in Table 30–8.

Non-Occupational Environmental Factors Tobacco Active Cigarette Smoking. Studies of cigarette smoking have generally suggested an elevated risk of SNC, with RRs in the range of 1.2–3.0 (Caplan et al., 2000; Elwood, 1981; Fukuda and Shibata, 1990; Hayes et al., 1987; Leclerc et al., 1994; Ng, 1986; Shimizu et al., 1989; Strader et al., 1988; Zheng et al., 1992; Zheng et al., 1993). A few studies have failed to observe an association (Hernberg et al., 1983; Lareo et al., 1992). In a pooled reanalysis of eight European studies, a small excess risk was observed among exsmokers (OR = 1.3, 95% CI: 1.0–1.8), and a slightly weaker association was seen among current smokers (OR = 1.2, 95% CI: 0.9–1.6). In the studies that quantified usual consumption, risk increased with increasing number of cigarettes smoked per day (Fukuda and Shibata, 1990; Hayes et al., 1987; Strader et al., 1988). Studies in Japan and the Netherlands observed strong associations (RR = 4.6 and 5.0, respectively) for SNC among those who smoked 35 or more cigarettes per day. One case-control study, which examined the effects of cessation, reported a 60% risk reduction (OR = 0.4, 95% CI: 0.2–0.7) for SNC among those who had quit smoking 10 years or longer (Zheng et al., 1993). Cigarette smoking appears to be somewhat more strongly related to squamous cell carcinomas, and cancers arising in the

maxillary sinus (Brinton et al., 1984; Elwood, 1981; Fukuda and Shibata, 1990; Shimizu et al., 1989; Mannetje et al., 1999; Zheng et al., 1993). A US study observed an OR = 6.6 (95% CI: 1.7–29.6) for 40 or more pack-years of smoking in analyses limited to squamous cell carcinoma (Strader et al., 1988). Based on the available evidence, the IARC has determined that there is sufficient evidence in humans that tobacco smoking causes cancer of the nasal cavity and paranasal sinuses (International Agency for Research on Cancer and World Health Organization, 2002).

Passive Smoking. Environmental tobacco smoke contains exhaled mainstream smoke, as well as many other toxic agents generated by tobacco combustion. In several studies conducted in Japan (Fukuda and Shibata, 1990; Hirayama, 1984) and the United States (Zheng et al., 1993), persons exposed to passive smoke had an increased relative risk of SNC, which was similar in magnitude to that of active smoking. Risk of SNC increased with increasing passive exposure to cigarettes smoked per day (Hirayama, 1984) or number of smokers in the home (Fukuda and Shibata, 1990). Non-smoking wives with husbands who smoked 20 or more cigarettes per day had a 2.6-fold increased risk (95% CI: 1.0–6.3) compared to non-smoking wives whose husbands did not smoke (Hirayama, 1984). Cases with two or more smokers in a household had a 5.7-fold (P value for trend <0.05) increased risk of SNC compared to those with none (Fukuda and Shibata, 1990). A study conducted in the United States found that cases more often than controls had a spouse who smoked cigarettes (OR = 3.0, 95% CI: 1.0–8.9), after adjustment for age and alcohol use; the association was particularly strong for maxillary sinus cancer (OR = 4.8, 95% CI: 0.9–24.7), but there was no evidence of a trend with amount of cigarettes smoked by the spouse (Zheng et al., 1993). The IARC has concluded that the data are conflicting and sparse for an association between passive smoking and SNC based in part on the unlikely situation where effects produced in passive smokers are not observed to a greater extent in active smokers (International Agency for Research on Cancer and World Health Organization, 2002). Snuff. Based on a small number of studies, there is some support for an association between snuff use and SNC, particularly for maxillary sinus cancer. Several researchers have suggested that the rela-

Table 30–8. Selected Occupations and Exposures Associated with an Increased Risk of Sinonasal Cancer Suspected Carcinogen

Occupation/Industry

Strength and/or Consistency of Association

Wood dust

Furniture manufacturing; other wood working occupations

+++

Leather dust

Boot and shoe manufacture and repair Nickel smelting and refining

+++

Nickel dust

Chromium Mineral oil Formaldehyde Textile dust Radium-226

+++

Comments Hardwoods are more strongly associated with SNC risk; association stronger for adenocarcinomas, and in studies conducted in Europe Association stronger for adenocarcinoma

Group 1

Association strongest among those first employed before 1930

Nickel sulfate and combinations of nickel sulfides and oxides: Group 1; Metallic nickel and nickel alloys: Group 2b Chromium[VI]: Group 1; Metallic chromium and chromium[III] compounds: Group 3 Untreated or mildly treated mineral oils: Group 1; Highly refined oils: Group 3 Group 2a

Primary chromate production, chromate pigment production, chromium plating Metal working, print press operating, and cotton and jute spinning Wood, leather, textiles, plastics, and related industries

+

Known lung carcinogen

+

Textile manufacturing Radium dial painting

++ ++

Untreated or mildly treated mineral oils have higher contents of polyaromatic hydrocarbons and nitrosamines Wood dust often present and may confound association; animal data (rat model) are compelling Self-luminescing paint applied by “lip-pointing” from 1917–1925

+

IARC Classification

Group 1

Group 2b Group 1

+ Inconsistent and/or RR < 2; ++ somewhat consistent and/or 2 < RR < 10; +++ consistent and/or RR > 10. Group 1: Sufficient evidence for carcinogenicity; Group 2a: Probably carcinogenic; Group 2b: Possibly carcinogenic; Group 3: Inadequate evidence of carcinogenicity (not classifiable). SNC, sinuses and nasal cancer.

614

PART IV: CANCER BY TISSUE OF ORIGIN

tively high rates of nasal cancer in southern Africa may be due to nasal snuff usage (Baumslag et al., 1971; Keen, 1974). In the Bantu in the 1960s, cancer of the nasal and accessory sinuses comprised 45.5% of all respiratory tract cancers compared to 4.4% in whites (Baumslag et al., 1971). Snuff was found to be contaminated with a variety of trace metals including nickel and chromium, which may explain its apparent carcinogenicity in this population (Baumslag et al., 1971). In several studies conducted in the United States and Europe, where snuff is usually taken under the lip or tongue, snuff usage was more common among SNC cases than controls (Acheson et al., 1970; Brinton et al., 1984; Hayes et al., 1987).

Pipes, Smokeless Tobacco, and Cigars. Evidence for an association between other forms of tobacco and SNC are limited and inconsistent. Both cigar and pipe use were associated with elevated risks for adenocarcinoma (though not squamous cell carcinoma) in a case-control study (cigars: OR = 2.6, 90% CI: 1.0–7.3; pipes: OR = 2.2, 90% CI: 0.9–5.4) (Hayes et al., 1987). Other studies found little evidence of an association with pipes (Caplan et al., 2000), chewing tobacco, or cigars (Brinton et al., 1984; Zheng et al., 1993).

include preparations such as decongestants and corticosteroid nasal sprays. Common adverse reactions to corticosteroid preparations include irritation and dryness of the nasal mucosa (Strader et al., 1988). Long-term use of nasal preparations is hypothesized to create an environment more favorable for carcinogenesis through a local reduction of mucociliary transport and the induction of necrotic changes in the sinonasal epithelium (Strader et al., 1988). Few studies have investigated the association between nasal preparations and SNC risk. One study in the United States reported an increased risk associated with the use of nasal preparations (OR = 3.5, 95% CI: 1.7–7.0), particularly for tumors of the nasal sinuses (OR = 5.2, 95% CI: 1.9–13.4) and for those with adenocarcinomas (OR = 8.0, 95% CI: 1.2–50.2) (Strader et al., 1988). In a study conducted in Holland, investigators observed an increased risk associated with nasal sprays among men (OR = 5), but no increase among women (Hayes et al., 1984). Brinton and colleagues, however, found no association with regular use of nose drops and SNC after controlling for a history of sinus trouble (Brinton et al., 1984). Results from studies investigating this association have been difficult to interpret because it is possible that conditions leading to the use of nasal preparations, not the preparations themselves, are associated with an increased risk of SNC.

Diet and Alcohol Although multiple studies have observed a strong association between ingestion of salted fish and nasopharyngeal cancer (Yu and Henderson, 1996), only two studies have investigated the association between this and other dietary factors among SNC cases and controls (Zheng et al., 1992; Zheng et al., 1993). In a US case-control study, proxy respondents were asked to report usual intake of alcohol and five food groups during the decedent’s adult life (Zheng et al., 1993). Those who consumed salted and pickled foods daily had an increased risk of SNC (OR = 1.7, 95% CI: 0.8–3.6), and those who frequently consumed vegetables had a decreased risk (OR = 0.7, 95% CI: 0.3–1.3), relative to those who ate salted or pickled foods, or vegetables less than three times per week, respectively (Zheng et al., 1993). Exposure misclassification may be a concern because proxy respondents reported diet for both cases and controls. A case-control study conducted in Shanghai also observed increased risks of SNC associated with consumption of salt preserved fish, meat, and vegetables (OR = 1.9, 95% CI: 1.2–2.8), and decreased risks for higher intakes of oranges and tangerines (OR = 0.6, 95% CI: 0.4–0.9), adjusted for other major risk factors (Zheng et al., 1992). Alcohol consumption has been inconsistently associated with SNC. Several studies have observed no association (Brinton et al., 1984; Hayes et al., 1986b), while two observed elevated risks (Strader et al., 1988; Zheng et al., 1993). Risk of death due to SNC among those who consumed alcohol daily was 1.8 times (95% CI: 1.0–3.3) higher than among those who drank less than once per week (Zheng et al., 1993). In another study, usual intake of 21 or more drinks per week was associated with a 3.4-fold (95% CI: 1.4–8.4) increase in risk of all SNC types, and a 6.8-fold (95% CI: 2.1–22.0) increase in risk of squamous cell carcinoma (Strader et al., 1988).

Thorotrast Instillation Stabilized thorium-232 containing thorium dioxide (Thorotrast) was used extensively in medical practice during the 1930s through the 1950s for various radiological purposes, including visualization of the sinus tracts and paranasal sinuses (IARC Working Group on the Evaluation of Carcinogenic Risks to Humans et al., 2001). Instillation of Thorotrast could have lasting effects due to its half-life (1.39 ¥ 1010 years), resulting in long-term alpha-irradiation of the tissues from the daughter compounds (Rankow et al., 1974). Several cases of SNC have been reported to be associated with its use (Goren et al., 1980; Rankow et al., 1974). However, a recent cohort study of 240 individuals administered Thorotrast for visualization of the paranasal sinuses and observed for at least 40 years observed no SNC deaths (dos Santos Silva et al., 2003).

Nasal Preparations Nasal sprays and drops are often used to alleviate symptoms of allergic rhinitis, upper respiratory tract infections, and nasal polyps, and

Infectious Agents Two viruses (Epstein-Barr virus (EBV) and HPV) have been investigated as potential risk factors for SNC. EBV has been implicated in the etiology of nasopharyngeal cancer, particularly undifferentiated nasopharyngeal carcinomas (Yu and Henderson, 1996). A number of studies have also explored a potential role for EBV infection in sinonasal undifferentiated carcinomas (Cerilli et al., 2001; Gallo et al., 1995; Hwang et al., 1998; Jeng et al., 2002; Leung et al., 1995; Lopategui et al., 1994; Paulino et al., 2000) and sinonasal lymphoepithelial carcinoma. Reported EBV detection prevalences in undifferentiated sinonasal carcinomas have ranged from zero (Cerilli et al., 2001; Jeng et al., 2002; Paulino et al., 2000) to more than 50% (Zong). Thus, while an association may be present, it appears to be weaker than for nasopharyngeal cancer. HPV DNA has been detected in 33.3% of sinonasal papillomas analyzed whereas detection in malignant sinonasal lesions was somewhat less common (21.7%) (Syrjanen, 2003). HPV-16 was detected most often, followed by HPV-18. Nevertheless, detection rates of HPV DNA in SNCs have varied substantially among studies, and results have been inconsistent. Some of the differences in detection may be due to different technologies, with different sensitivities and specificities (Syrjanen, 2000). Limitations from published studies limit the ability to infer a causal association. Many included cases only, without a comparison group; thus it is difficult to determine whether prevalence of EBV or HPV is truly more common among SNC cases than controls. Second, sampling of serum or tissue was typically done after diagnosis. It is unclear whether existence of SNC or treatment for it may impact likelihood of detection of EBV or HPV. Presence of EBV or HPV DNA in SNC may represent chronic infection that existed prior to cancer diagnosis, or it may instead reflect transient reactivation of the virus in the neoplastic nasal cells as a result of the cancer (Gallo et al., 1995). Also, misclassification of disease is a concern with undifferentiated carcinomas in particular, which can appear similar to nasopharyngeal cancers that may have spread to the nasal cavity or sinuses. Finally, due to environmental or genetic cofactors, the association between EBV and SNC may be limited to certain ethnic groups, such as Asians. Apparent inconsistency of findings may be due to such effect modification.

HOST FACTORS Sinusitis, Nose Bleeds, Rhinitis, and Nasal Trauma/Injury Sinusitis is a common condition caused by a variety of infectious, anatomic, and inflammatory factors. Although several inflammatory conditions occurring in other organ systems, such as atrophic gastri-

Cancers of the Nasal Cavity and Paranasal Sinuses tis, are associated with subsequent development of cancer, there has been only limited support for an association between chronic sinusitis and SNC. One major reason is the difficulty in separating prior disease from early symptoms of the cancer. Only some studies have attempted to address this issue by asking about history of nasal diseases several years before diagnosis. As all studies to investigate this question have used a case-control design, recall bias is a particular concern because there was no independent assessment of infection. An early study conducted in the United States found elevated risks for “sinus trouble” (OR = 2.2, 95% CI: 1.3–3.7) and nose bleeds (OR = 1.9, 95% CI: 1.1–3.2) that were diagnosed 10 or more years prior to SNC diagnosis or interview (Brinton et al., 1984). Two studies conducted in Japan observed 2- to 3-fold increased risks of squamous cell carcinoma of the maxillary sinus (Fukuda and Shibata, 1988; Shimizu et al., 1989). In a case-control study conducted in France, men with a history of sinusitis or nose bleeds 10 or more years before interview had a 2.8-fold (95% CI: 1.4–5.7) and 3-fold increased risk (95% CI: 1.2–7.3), respectively of squamous cell carcinoma (Lareo et al., 1992). History of rhinitis was also associated with squamous cell carcinoma among men (OR = 2.9, 95% CI: 1.2–6.9), but the association was not significant when limited to rhinitis diagnosed at least 10 years prior to interview. Relative risks for adenocarcinomas were also elevated for a history of either nose bleeds or rhinitis 10 years prior to interview (OR = 3.3, 95% CI: 1.5–7.1; OR = 2.4, 95% CI: 1.2–5.0, respectively). Associations were generally similar, but not statistically significant among women (Lareo et al., 1992). In a study conducted in Shanghai, a diagnosis of sinusitis, rhinitis, nasal polyps, chronic otitis media, nose bleeds, or other disease of the nose or middle ear was associated with a 5-fold elevated risk (95% CI 2.9–9.0), after adjustment for other major risk factors (Zheng et al., 1992). Risk of SNC remained elevated among those with onset 10 or more years prior to diagnosis or interview (OR = 4.0, 95% CI: 2.1–7.4) (Zheng et al., 1992). Risk was particularly high for chronic sinusitis (OR = 12.3, 95% CI: 4.1–36.8) and recurrent nose bleeds (OR = 12.9, 95% CI: 5.6–29.6) (Zheng et al., 1992). A recent case-control study among US men diagnosed between 1984 and 1988 found that ever having sinus problems other than infection in the previous 5 years with first occurrence 5 years or more prior to interview date was not associated with SNC (OR = 1.1, 95% CI: 0.1–8.8) (Zhu et al., 2002). Several studies have also observed an increased relative risk associated with a history of nasal trauma or injury (Lareo et al., 1992; Tola et al., 1980), although one did not (Shimizu et al., 1989). Lareo and colleagues observed a 3.3-fold elevation in risk (95% CI: 1.5–7.1) of squamous cell carcinoma associated with nasal trauma occurring 10 or more years before diagnosis/interview date (Lareo et al., 1992). In a case-control study conducted in Finland, five cases and no controls reported a history of either probable fracture or lesion of the nasal septum (P = 0.028) (Tola et al., 1980).

Polyps Nasal polyps are an inflammatory disorder of the nose and paranasal sinuses. Solitary nasal polyps may be caused by acute or chronic sinusitis, and diffuse nasal polyposis may cause secondary sinusitis (Noble and Greene, 2001). Relative risks for SNC (particularly squamous cell carcinoma) associated with a history of polyps have generally been high. In studies in the United States, Japan, and France, a history of polyps was associated with relative risks in excess of five for squamous cell carcinoma (Fukuda and Shibata, 1988; Lareo et al., 1992) or for all histologies of SNC (Brinton et al., 1984; Zheng et al., 1992). The association was weaker for adenocarcinomas (OR = 2.7, 95% CI: 0.8–8.6) (Lareo et al., 1992). Two other studies, however, observed 70% increases in relative risks that were compatible with no association (Shimizu et al., 1989; Strader et al., 1988).

Estrogenic Hormones Estrogenic hormones have been hypothesized to alter risk for SNC with possible reductions in risk after menopause (Roush et al., 1987a). This theory is supported by a relative deficit in SNC incidence among

615

women in their 50s, as well as other data. However, little research has been conducted to investigate this hypothesis.

PATHOGENESIS A large number of factors have been implicated in the etiology of SNC, including various exposures largely encountered in occupational settings, such as dusts, byproducts from snuff and cigarette use, and infectious agents. Some of these agents are likely to act by directly damaging DNA and chromosomal integrity (e.g., carcinogens in tobacco smoke). Others, particularly dusts, may increase risk of SNC by causing chronic inflammation, with the production of potentially genotoxic reactive oxygen species; and by increasing cell proliferation. Cell proliferation increases the likelihood that a mutational event related to the carcinogenic process will occur via errors in replication or the conversion of endogenous or exogenous DNA adducts to mutations before DNA repair can occur (Butterworth and Goldsworthy, 1991). Many questions remain, however, regarding the mechanisms by which individual exposures increase risk, the effects of exposure cessation on subsequent risk, and the specific molecular pathways involved.

PREVENTIVE MEASURES The most effective primary prevention measure is likely to be avoidance of tobacco use. Caplan and colleagues (Caplan et al., 2000) estimated the population attributable risk percent for ever having smoked cigarettes (based on an OR of 2.4 for ever having smoked) of 53%. If one assumes a more conservative RR of 2.0 and a prevalence of ever smoking of 60% (t Mannetje et al., 1999), eliminating tobacco smoking could result in a 30% reduction in the incidence of SNC. Another important preventive measure is reduction in the frequency and level of exposure to occupational agents that have been strongly linked to the development of SNC. The most commonly encountered of these in the United States is wood dust. Most SNCs are advanced at presentation because they originate in a space that must be almost filled before signs and symptoms develop (Stupp et al., 2000). These relate to the impingement of the tumor on neighboring structures such as the eye, nose, and mouth. Therefore, nearly half of all SNCs are diagnosed at the regional or distant stage. Unfortunately, the rarity of the disease renders population-based screening measures prohibitively expensive. However, it may be feasible to screen persons at high risk on the basis of their occupational exposure history. While such a program is unlikely to identify a significant number of cancers, it might identify persons with preneoplastic lesions, such as metaplasia, dysplasia or papilloma, who could then be placed in a surveillance program for early detection of subsequent malignancy.

FUTURE DIRECTIONS A better understanding of the genetic and epigenetic pathways by which occupational and environmental exposures increase SNC risk might identify prevention opportunities not only for SNC, but for other respiratory cancers as well. In particular, little is known about the somatic genetic abnormalities underlying the initiation and progression of this disease, the order in which they occur, and the risk factors with which they are most closely associated. While certain types of papilloma have been closely associated with SNC, the underlying mechanism is not understood, nor are the factors associated with risk of subsequent neoplastic progression. Since much of the mucosa of the nasal cavity is accessible with minimally invasive procedures, the opportunity exists to take multiple biopsies over time to study the natural history of progression. A better understanding of the etiology and prevalence of precursor conditions such as dysplasia and inverted papillomas might help to identify a high-risk subgroup of exposed persons for subsequent surveillance and prevention activities. These

616

PART IV: CANCER BY TISSUE OF ORIGIN

and other precursor conditions also might be useful as intermediate outcomes in prevention studies. Clarifying the possible role of infectious agents, including HPV and EBV, in the development of SNC is another priority. Finally, the observed differences between blacks and whites in the incidence of SNC at various subsites suggests an area of research that might be fruitful in terms of identifying new risk factors and understanding the pathogenesis of this disease. References Acheson ED, Cowdell RH, Hadfield E, Macbeth RG. 1968. Nasal cancer in woodworkers in the furniture industry. Br Med J 2:587–596. Acheson ED, Cowdell RH, Jolles B. 1970. Nasal cancer in the Northamptonshire boot and shoe industry. Br Med J 1:385–393. Acheson ED, Cowdell RH, Rang E. 1972. Adenocarcinoma of the nasal cavity and sinuses in England and Wales. Br J Ind Med 29:21–30. Acheson ED, Cowdell RH, Rang EH. 1981. Nasal cancer in England and Wales: An occupational survey. Br J Ind Med 38:218–224. Agency for Toxic Substances and Disease Registry. 1997. Toxicological Profile for Nickel. US Department of Health and Human Services, Public Health Service: Atlanta, GA. Agency for Toxic Substances and Disease Registry. 2000. Toxicological Profile for Chromium. US Department of Health and Human Services, Public Health Service: Atlanta, GA. Alderson MR, Rattan NS, Bidstrup L. 1981. Health of workmen in the chromate-producing industry in Britain. Br J Ind Med 38:117–124. Andersen A, Berge SR, Engeland A, Norseth T. 1996. Exposure to nickel compounds and smoking in relation to incidence of lung and nasal cancer among nickel refinery workers. Occup Environ Med 53:708–713. Andersen HC, Andersen I, Solgaard J. 1977. Nasal cancer, symptoms and upper airway function in woodworkers. Br J Ind Med 34:201–207. Anttila A, Pukkala E, Aitio A, Rantanen T, Karjalainen S. 1998. Update of cancer incidence among workers at a copper/nickel smelter and nickel refinery. Int Arch Occup Environ Health 71:245–250. Ballarin C, Sarto F, Giacomelli L, Bartolucci GB, Clonfero E. 1992. Micronucleated cells in nasal mucosa of formaldehyde-exposed workers. Mutat Res 280:1–7. Battista G, Comba P, Orsi D, Norpoth K, Maier A. 1995. Nasal cancer in leather workers: An occupational disease. J Cancer Res Clin Oncol 121:1–6. Baumslag N, Keen P, Petering HG. 1971. Carcinoma of the maxillary antrum and its relationship to trace metal content of snuff. Arch Environ Health 23:1–5. Baxter PJ, McDowall ME. 1986. Occupation and cancer in London: An investigation into nasal and bladder cancer using the Cancer Atlas. Br J Ind Med 43:44–49. Bielamowicz S, Calcaterra TC, Watson D. 1993. Inverting papilloma of the head and neck: The UCLA update. Otolaryngol Head Neck Surg 109:71–76. Bimbi G, Battista G, Belli S, Berrino F, Comba P. 1988. [A case-control study of nasal tumors and occupational exposure.] Med Lav 79:280–287. Black A, Evans JC, Hadfield EH, Macbeth RG, Morgan A, Walsh M. 1974. Impairment of nasal mucociliary clearance in woodworkers in the furniture industry. Br J Ind Med 31:10–17. Blair A, Saracci R, Stewart PA, Hayes RB, Shy C. 1990. Epidemiologic evidence on the relationship between formaldehyde exposure and cancer. Scand J Work Environ Health 16:381–393. Blair A, Stewart P, O’Berg M, et al. 1986. Mortality among industrial workers exposed to formaldehyde. J Natl Cancer Inst 76:1071–1084. Boysen M, Solberg LA. 1982. Changes in the nasal mucosa of furniture workers. A pilot study. Scand J Work Environ Health 8:273–282. Boysen M, Zadig E, Digernes V, Abeler V, Reith A. 1990. Nasal mucosa in workers exposed to formaldehyde: A pilot study. Br J Ind Med 47: 116–121. Brinton LA, Blot WJ, Becker JA, Winn DM, Browder JP, Farmer JC Jr, Fraumeni JF, Jr. 1984. A case-control study of cancers of the nasal cavity and paranasal sinuses. Am J Epidemiol 119:896–906. Brinton LA, Blot WJ, Fraumeni JF, Jr. 1985. Nasal cancer in the textile and clothing industries. Br J Ind Med 42:469–474. Bross ID, Viadana E, Hooten L. 1978. Occupational cancer in men exposed to dust and other environmental hazards. Arch Environ Health 33:300–307. Brues AM, Kirsh IE. 1977. The fate of individuals containing radium. Trans Am Clin Climatol Assoc 88:211–218. Buchwald C, Lindeberg H, Pedersen BL, Franzmann MB. 2001. Human papilloma virus and p53 expression in carcinomas associated with sinonasal papillomas: A Danish Epidemiological study 1980–1998. Laryngoscope 111:1104–1110. Burgaz S, Cakmak G, Erdem O, Yilmaz M, Karakaya AE. 2001. Micronuclei frequencies in exfoliated nasal mucosa cells from pathology and anatomy

laboratory workers exposed to formaldehyde. Neoplasma 48:144– 147. Bussi M, Gervasio CF, Riontino E, Valente G, Ferrari L, Pira E, Cortesina G. 2002. Study of ethmoidal mucosa in a population at occupational high risk of sinonasal adenocarcinoma. Acta Otolaryngol 122:197–201. Butterworth BE, Goldsworthy TL. 1991. The role of cell proliferation in multistage carcinogenesis. Proc Soc Exp Biol Med 198:683–687. Caplan LS, Hall HI, Levine RS, Zhu K. 2000. Preventable risk factors for nasal cancer. Ann Epidemiol 10:186–191. Caruana SM, Zwiebel N, Cocker R, McCormick SA, Eberle RC, Lazarus P. 1997. p53 alteration and human papilloma virus infection in paranasal sinus cancer. Cancer 79:1320–1328. Cecchi F, Buiatti E, Kriebel D, Nastasi L, Santucci M. 1980. Adenocarcinoma of the nose and paranasal sinuses in shoemakers and woodworkers in the province of Florence, Italy (1963–77). Br J Ind Med 37:222–225. Cerilli LA, Holst VA, Brandwein MS, Stoler MH, Mills SE. 2001. Sinonasal undifferentiated carcinoma: Immunohistochemical profile and lack of EBV association. Am J Surg Pathol 25:156–163. Cohen SM, Ellwein LB. 1991. Genetic errors, cell proliferation, and carcinogenesis. Cancer Res 51:6493–6505. Collins JJ, Acquavella JF, Esmen NA. 1997. An updated meta-analysis of formaldehyde exposure and upper respiratory tract cancers. J Occup Environ Med 39:639–651. Comba P, Battista G, Belli S, et al. 1992. A case-control study of cancer of the nose and paranasal sinuses and occupational exposures. Am J Ind Med 22:511–520. Cornell RG, Landis JR. 1984. Mortality patterns among nickel/chromium alloy foundry workers. IARC Sci Publ 87–93. Correa P. 1992. Human gastric carcinogenesis: A multistep and multifactorial process—First American Cancer Society Award Lecture on Cancer Epidemiology and Prevention. Cancer Res 52:6735–6740. Cragle DL, Hollis DR, Newport TH, Shy CM. 1984. A retrospective cohort mortality study among workers occupationally exposed to metallic nickel powder at the Oak Ridge Gaseous Diffusion Plant. IARC Sci Publ 57–63. Davies JM, Easton DF, Bidstrup PL. 1991. Mortality from respiratory cancer and other causes in United Kingdom chromate production workers. Br J Ind Med 48:299–313. Debois J. 1969. Tumoren van de neusholte by houtebewerkers. Tijdschr Geneeskol 25:92–93. Decoufle P, Walrath J. 1983. Proportionate mortality among US shoeworkers, 1966–1977. Am J Ind Med 4:523–532. Decoufle P, Walrath J. 1987. Nasal cancer in the U.S. shoe industry: Does it exist? Am J Ind Med 12:605–613. Dell L, Teta MJ. 1995. Mortality among workers at a plastics manufacturing and research and development facility:1946–1988. Am J Ind Med 28:373–384. Demers PA, Kogevinas M, Boffetta P, et al. 1995. Wood dust and sino-nasal cancer: Pooled reanalysis of twelve case-control studies. Am J Ind Med 28:151–166. Doll R, Andersen A, Cooper WC, et al. 1990. Report of the International Committee on Nickel Carcinogenesis in Man. Scand J Work Environ Health 16:1–82. Doll R, Mathews JD, Morgan LG. 1977. Cancers of the lung and nasal sinuses in nickel workers: A reassessment of the period of risk. Br J Ind Med 34:102–105. Doll R, Morgan LG, Speizer FE. 1970. Cancers of the lung and nasal sinuses in nickel workers. Br J Cancer 24:623–632. dos Santos Silva I, Malveiro F, Jones ME, Swerdlow AJ. 2003. Mortality after radiological investigation with radioactive Thorotrast: A follow-up study of up to fifty years in Portugal. Radiat Res 159:521–534. Dubrow R, Gute DM. 1988. Cause-specific mortality among male textile workers in Rhode Island. Am J Ind Med 13:439–454. Eckardt RE. 1974. Annals of industry—noncasualties of the work place. J Occup Med 16:472–477. Edling C, Hellquist H, Odkvist L. 1987. Occupational formaldehyde exposure and the nasal mucosa. Rhinology 25:181–187. Egedahl R, Carpenter M, Lundell D. 2001. Mortality experience among employees at a hydrometallurgical nickel refinery and fertiliser complex in Fort Saskatchewan, Alberta (1954–95). Occup Environ Med 58:711–715. Elwood JM. 1981. Wood exposure and smoking: Association with cancer of the nasal cavity and paranasal sinuses in British Columbia. Can Med Assoc J 124:1573–1757. Enterline PE. 1974. Respiratory cancer among chromate workers. J Occup Med 16:523–526. Enterline PE, Marsh GM. 1982. Mortality among workers in a nickel refinery and alloy manufacturing plant in West Virginia. J Natl Cancer Inst 68:925–933.

Cancers of the Nasal Cavity and Paranasal Sinuses Feron VJ, Arts JH, Kuper CF, Slootweg PJ, Woutersen RA. 2001. Health risks associated with inhaled nasal toxicants. Crit Rev Toxicol 31:313–347. Franzmann MB, Buchwald C, Jacobsen GK, Lindeberg H. 1998. Expression of p53 in normal nasal mucosa and in sinonasal papillomas with and without associated carcinoma and the relation to human papillomavirus (HPV). Cancer Lett 128:161–164. Fritz AG. 2000. International classification of diseases for oncology: ICD-O. World Health Organization: Geneva. Fu H, Demers PA, Costantini AS, et al. 1996. Cancer mortality among shoe manufacturing workers: an analysis of two cohorts. Occup Environ Med 53:394–398. Fukuda K, Shibata A. 1988. A case-control study of past history of nasal diseases and maxillary sinus cancer in Hokkaido, Japan. Cancer Res 48:1651–1652. Fukuda K, Shibata A. 1990. Exposure-response relationships between woodworking, smoking or passive smoking, and squamous cell neoplasms of the maxillary sinus. Cancer Causes Control 1:165–168. Fukuda K, Shibata A, Harada K. 1987. Squamous cell cancer of the maxillary sinus in Hokkaido, Japan: A case-control study. Br J Ind Med 44:263–266. Gallo O, Di Lollo S, Graziani P, Gallina E, Baroni G. 1995. Detection of Epstein-Barr virus genome in sinonasal undifferentiated carcinoma by use of in situ hybridization. Otolaryngol Head Neck Surg 112:659–664. Garabrant DH, Wegman DH. 1984. Cancer mortality among shoe and leather workers in Massachusetts. Am J Ind Med 5:303–314. Gardner MJ, Pannett B, Winter PD, Cruddas AM. 1993. A cohort study of workers exposed to formaldehyde in the British chemical industry: An update. Br J Ind Med 50:827–834. Gazdar AF. 1994. The molecular and cellular basis of human lung cancer. Anticancer Res 14:261–267. Goldberg M, Goldberg P, Leclerc A, Chastang JF, Marne MJ, Dubourdieu D. 1994. A 10-year incidence survey of respiratory cancer and a case-control study within a cohort of nickel mining and refining workers in New Caledonia. Cancer Causes Control 5:15–25. Goren AD, Harley N, Eisenbud L, Levin S, Cohen N. 1980. Clinical and radiobiologic features of Thorotrast-induced carcinoma of the maxillary sinus. A case report. Oral Surg Oral Med Oral Pathol 49:237–242. Gotte K, Riedel F, Schafer C, Coy J, Hormann K. 2000. Cylindrical cell carcinomas of the paranasal sinuses do not show p53 alterations but loss of heterozygosity at 3p and 17p. Int J Cancer 85:740–742. Haggitt RC. 1994. Barrett’s esophagus, dysplasia, and adenocarcinoma. Hum Pathol 25:982–993. Hall A, Harrington JM, Aw TC. 1991. Mortality study of British pathologists. Am J Ind Med 20:83–89. Hansen J, Olsen JH. 1995. Formaldehyde and cancer morbidity among male employees in Denmark. Cancer Causes Control 6:354–360. Harrington JM, Oakes D. 1984. Mortality study of British pathologists 1974–80. Br J Ind Med 41:188–191. Hasterlik RJ, Finkel AJ, Miller CE. 1964. The cancer hazards of industrial and accidental exposure to radioactive isotopes. Ann N Y Acad Sci 114:832–837. Hayes RB, Blair A, Stewart PA, Herrick RF, Mahar H. 1990. Mortality of U.S. embalmers and funeral directors. Am J Ind Med 18:641–652. Hayes RB, Gerin M, Raatgever JW, de Bruyn A. 1986a. Wood-related occupations, wood dust exposure, and sinonasal cancer. Am J Epidemiol 124:569–577. Hayes RB, Kardaun JW, de Bruyn A. 1987. Tobacco use and sinonasal cancer: A case-control study. Br J Cancer 56:843–846. Hayes RB, Raatgever JW, de Bruyn A. 1984. Tumours of the nose and nasal sinuses: A case-control study. Technical Report. Rotterdam: Department of Public Health and Social Medicine, Erasmus University, pp. 19–22. Hayes RB, Raatgever JW, de Bruyn A, Gerin M. 1986b. Cancer of the nasal cavity and paranasal sinuses, and formaldehyde exposure. Int J Cancer 37:487–492. Hernberg S, Westerholm P, Schultz-Larsen K, et al. 1983. Nasal and sinonasal cancer. Connection with occupational exposures in Denmark, Finland and Sweden. Scand J Work Environ Health 9:315–326. Hinds WC. 1988. Basis for particle size-selective sampling for wood dust. Appl Industr Hygiene 3:67–72. Hirayama T. 1984. Cancer mortality in nonsmoking women with smoking husbands based on a large-scale cohort study in Japan. Prev Med 13: 680–690. Howlett CT, Mathias R, Friess S. 1989. Quantitative risk approaches for formaldehyde. Exp Pathol 37:119–127. Hueper WC. 1966. Occupational and environmental cancers of the respiratory system. Recent Results Cancer Res 3:105–107. Hwang TZ, Jin YT, Tsai ST. 1998. EBER in situ hybridization differentiates carcinomas originating from the sinonasal region and the nasopharynx. Anticancer Res 18:4581–4584.

617

International Agency for Research on Cancer Working Group on the Evaluation of Carcinogenic Risks to Humans, International Agency for Research on Cancer and World Health Organization. 2001. Ionizing Radiation, Part 2: Some Internally Deposited Radionuclides. Vol. 78. IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans. IARC: Lyon, France. International Agency for Research on Cancer. 1981. Wood, Leather, and Some Associated Industries. Vol. 25. International Agency for Research on Cancer: Lyon, France. International Agency for Research on Cancer. 1990. Some Flame Retardants and Textile Chemicals, and Exposures in the Textile Manufacturing Industry. Vol. 48. IARC Monographs. International Agency for Research on Cancer: Lyon, France. International Agency for Research on Cancer. 2002. Cancer in Five Continents, Vol. VIII. Vol. 8. IARC Scientific Publication No. 155. International Agency for Research on Cancer: Lyon, France. International Agency for Research on Cancer and World Health Organization. 1987. Overall Evaluations of Carcinogenicity: An Updating of IARC Monographs Volumes 1 to 42. Vol. Supplement 7. IARC Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans. International Agency for Research on Cancer: Lyon, France. International Agency for Research on Cancer and World Health Organization. 1990. Chromium, Nickel and Welding. Vol. 49. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. International Agency for Research on Cancer: Lyon, France. International Agency for Research on Cancer and World Health Organization. 1995. Wood Dust and Formaldehyde. Vol. 62. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. International Agency for Research on Cancer: Lyon, France. International Agency for Research on Cancer and World Health Organization. 2002. Tobacco Smoke and Involuntary Smoking. Vol. 83. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. International Agency for Research on Cancer: Lyon, France. Jarup L, Bellander T, Hogstedt C, Spang G. 1998. Mortality and cancer incidence in Swedish battery workers exposed to cadmium and nickel. Occup Environ Med 55:755–759. Jeng YM, Sung MT, Fang CL, et al. 2002. Sinonasal undifferentiated carcinoma and nasopharyngeal-type undifferentiated carcinoma: Two clinically, biologically, and histopathologically distinct entities. Am J Surg Pathol 26:371–376. Jones PA, Smith LC. 1986. Personal exposures to wood dust of woodworkers in the furniture industry in the High Wycombe area: A statistical comparison of 1983 and 1976/77 survey results. Ann Occup Hyg 30: 171–184. Kaldor J, Peto J, Easton D, Doll R, Hermon C, Morgan L. 1986. Models for respiratory cancer in nickel refinery workers. J Natl Cancer Inst 77: 841–848. Kaufman MR, Brandwein MS, Lawson W. 2002. Sinonasal papillomas: Clinicopathologic review of 40 patients with inverted and oncocytic schneiderian papillomas. Laryngoscope 112:1372–1377. Keen P. 1974. Trace elements in plants and soil in relation to cancer. S Afr Med J 48:2363–2364. Kerns WD, Pavkov KL, Donofrio DJ, Gralla EJ, Swenberg JA. 1983. Carcinogenicity of formaldehyde in rats and mice after long-term inhalation exposure. Cancer Res 43:4382–4392. Langer AM, Rohl AN, Selikoff IJ, Harlow GE, Prinz M. 1980. Asbestos as a cofactor in carcinogenesis among nickel-processing workers. Science 209:420–422. Lareo AC, Luce D, Leclerc A, Rodriguez J, Brugere J. 1992. History of previous nasal diseases and sinonasal cancer: A case-control study. Laryngoscope 102:439–442. Lawson W, Ho BT, Shaari CM, Biller HF. 1995. Inverted papilloma: A report of 112 cases. Laryngoscope 105:282–288. Leclerc A, Luce D, Demers PA, et al. 1997. Sinonasal cancer and occupation. Results from the reanalysis of twelve case-control studies. Am J Ind Med 31:153–165. Leclerc A, Martinez Cortes M, Gerin M, Luce D, Brugere J. 1994. Sinonasal cancer and wood dust exposure: Results from a case-control study. Am J Epidemiol 140:340–349. Lerchen ML, Samet JM. 1986. An assessment of the validity of questionnaire responses provided by a surviving spouse. Am J Epidemiol 123:481– 489. Leung SY, Yuen ST, Chung LP, Kwong WK, Wong MP, Chan SY. 1995. Epstein-Barr virus is present in a wide histological spectrum of sinonasal carcinomas. Am J Surg Pathol 19:994–1001. Levine RJ, Andjelkovich DA, Shaw LK. 1984. The mortality of Ontario undertakers and a review of formaldehyde-related mortality studies. J Occup Med 26:740–746.

618

PART IV: CANCER BY TISSUE OF ORIGIN

Lopategui JR, Gaffey MJ, Frierson HF Jr., et al. 1994. Detection of EpsteinBarr viral RNA in sinonasal undifferentiated carcinoma from Western and Asian patients. Am J Surg Pathol 18:391–398. Luce D, Gerin M, Leclerc A, Morcet JF, Brugere J, Goldberg M. 1993. Sinonasal cancer and occupational exposure to formaldehyde and other substances. Int J Cancer 53:224–231. Luce D, Gerin M, Morcet JF, Leclerc A. 1997. Sinonasal cancer and occupational exposure to textile dust. Am J Ind Med 32:205–210. Luce D, Leclerc A, Begin D, et al. 2002. Sinonasal cancer and occupational exposures: A pooled analysis of 12 case-control studies. Cancer Causes Control 13:147–157. Luce D, Leclerc A, Morcet JF, et al. 1992. Occupational risk factors for sinonasal cancer: A case-control study in France. Am J Ind Med 21: 163–175. Ma TH, Harris MM. 1988. Review of the genotoxicity of formaldehyde. Mutat Res 196:37–59. Macbeth RG. 1965. Malignant disease of the paranasal sinuses. J Laryngol Otol 79:596–612. Magnus K, Andersen A, Hogetveit AC. 1982. Cancer of respiratory organs among workers at a nickel refinery in Norway. Int J Cancer 30:681–685. Malker HS, McLaughlin JK, Blott WJ, et al. 1986. Nasal cancer and occupation in Sweden, 1961–1979. Am J Ind Med 9:477–485. Marsh GM. 1982. Proportional mortality patterns among chemical plant workers exposed to formaldehyde. Br J Ind Med 39:313–322. Meininger J, Raffinot P, Troly G. 1982. Cancer in nickel-processing workers in New Caledonia. Science 215:424–426. Merler E, Baldasseroni A, Laria R, et al. 1986. On the causal association between exposure to leather dust and nasal cancer: Further evidence from a case-control study. Br J Ind Med 43:91–95. Ng TP. 1986. A case-referent study of cancer of the nasal cavity and sinuses in Hong Kong. Int J Epidemiol 15:171–175. Noble J, Greene HL. 2001. Textbook of Primary Care Medicine. St. Louis: Mosby. Olsen JH. 1987. Epidemiology of sinonasal cancer in Denmark, 1943–1982. Acta Pathol Microbiol Immunol Scand [A] 95:171–175. Olsen JH. 1988. Occupational risks of sinonasal cancer in Denmark. Br J Ind Med 45:329–335. Olsen JH, Asnaes S. 1986. Formaldehyde and the risk of squamous cell carcinoma of the sinonasal cavities. Br J Ind Med 43:769–774. Olsen JH, Jensen SP, Hink M, Faurbo K, Breum NO, Jensen OM. 1984. Occupational formaldehyde exposure and increased nasal cancer risk in man. Int J Cancer 34:639–644. Partanen T. 1993. Formaldehyde exposure and respiratory cancer—a metaanalysis of the epidemiologic evidence. Scand J Work Environ Health 19:8–15. Paulino AF, Singh B, Carew J, Shah JP, Huvos AG. 2000. Epstein-Barr virus in squamous carcinoma of the anterior nasal cavity. Ann Diagn Pathol 4:7–10. Pedersen E, Hogetveit AC, Andersen A. 1973. Cancer of respiratory organs among workers at a nickel refinery in Norway. Int J Cancer 12:32–41. Percy C, Stanek E 3rd, Gloeckler L. 1981. Accuracy of cancer death certificates and its effect on cancer mortality statistics. Am J Public Health 71:242–250. Percy C, Van Holten V, World Health Organization and National Cancer Institute (US). 1979. Conversion of neoplasms by topography and morphology from the International Classification of Diseases for Oncology (ICD-O) to Chapter II, Neoplasms, 9th revision of the International Classification of Diseases, (ICD-9) 1975. NIH publication; no. 79-2007. US Dept. of Health Education and Welfare Public Health Service National Institutes of Health National Cancer Institute: Bethesda. Pippard EC, Acheson ED. 1985. The mortality of boot and shoe makers, with special reference to cancer. Scand J Work Environ Health 11:249–255. Polednak AP, Stehney AF, Rowland RE. 1978. Mortality among women first employed before 1930 in the US radium dial-painting industry. A group ascertained from employment lists. Am J Epidemiol 107:179–195. Randell RA. 1990. Nasal cancer and mineral oil. Lancet 335:56. Rankow RM, Conley J, Fodor P. 1974. Carcinoma of the maxillary sinus following thorotrast instillation. J Maxillofac Surg 2:119–126. Redmond CK. 1984. Site-specific cancer mortality among workers involved in the production of high nickel alloys. IARC Sci Publ 73–86. Ries LAG, Eisner MP, Kosary CL, Hankey BF. 2003. SEER Cancer Statistics Review, 1975–2000. National Cancer Institute: Bethesda, MD. Roberts RS, Julian JA, Muir DC, Shannon HS. 1989. A study of mortality in workers engaged in the mining, smelting, and refining of nickel. II: Mortality from cancer of the respiratory tract and kidney. Toxicol Ind Health 5:975–993. Roush GC, Meigs JW, Kelly JA, Flannery JT, Burdo H. 1980. Sinonasal cancer and occupation: A case-control study. Am J Epidemiol 111:183–193.

Roush GC, Schymura MJ, Stevenson JM, Holford TR. 1987a. Time and age trends for sinonasal cancer in Connecticut incidence and US mortality rates. Cancer 60:422–428. Roush GC, Walrath J, Stayner LT, Kaplan SA, Flannery JT, Blair A. 1987b. Nasopharyngeal cancer, sinonasal cancer, and occupations related to formaldehyde: A case-control study. J Natl Cancer Inst 79:1221–1224. Rutstein DD, Mullan RJ, Frazier TM, Halperin WE, Melius JM, Sestito JP. 1983. Sentinel Health Events (occupational): A basis for physician recognition and public health surveillance. Am J Public Health 73:1054–1062. Satoh K, Fukuda Y, Torii K, Katsuno N. 1981. Epidemiological study of workers engaged in the manufacture of chromium compounds. J Occup Med 23:835–838. Scheeper B, Kromhout H, Boleij JS. 1995. Wood-dust exposure during woodworking processes. Ann Occup Hyg 39:141–154. The Selected Cancers Cooperative Study Group. 1990. The association of selected cancers with service in the US military in Vietnam. III. Hodgkin’s disease, nasal cancer, nasopharyngeal cancer, and primary liver cancer. Arch Intern Med 150:2495–2505. Shannon HS, Walsh C, Jadon N, et al. 1991. Mortality of 11,500 nickel workers—extended follow up and relationship to environmental conditions. Toxicol Ind Health 7:277–294. Shimizu H, Hozawa J, Saito H, et al. 1989. Chronic sinusitis and woodworking as risk factors for cancer of the maxillary sinus in northeast Japan. Laryngoscope 99:58–61. Sorahan T, Burges DC, Waterhouse JA. 1987. A mortality study of nickel/chromium platers. Br J Ind Med 44:250–258. Stayner L, Smith AB, Reeve G, et al. 1985. Proportionate mortality study of workers in the garment industry exposed to formaldehyde. Am J Ind Med 7:229–240. Strader CH, Vaughan TL, Stergachis A. 1988. Use of nasal preparations and the incidence of sinonasal cancer. J Epidemiol Community Health 42:243–248. Stroup NE, Blair A, Erikson GE. 1986. Brain cancer and other causes of death in anatomists. J Natl Cancer Inst 77:1217–1224. Stupp R, Chung TDK, Collins SL, Dougherty MJ, Weichselbaum RR, Vokes EE. 2000. Head and Neck. In: Abeloff MD, ed. Clinical Oncology. New York: Churchill Livingstone, pp. 1296–1299. Sunderman FW Jr. 2001. Nasal toxicity, carcinogenicity, and olfactory uptake of metals. Ann Clin Lab Sci 31:3–24. Swenberg JA, Kerns WD, Mitchell RI, Gralla EJ, Pavkov KL. 1980. Induction of squamous cell carcinomas of the rat nasal cavity by inhalation exposure to formaldehyde vapor. Cancer Res 40:3398–3402. Syrjanen K. 2000. Papillomavirus Infections in Human Pathology. New York: Wiley & Sons. Syrjanen KJ. 2003. HPV infections in benign and malignant sinonasal lesions. J Clin Pathol 56:174–181. t Mannetje A, Kogevinas M, Luce D, et al. 1999. Sinonasal cancer, occupation, and tobacco smoking in European women and men. Am J Ind Med 36:101–107. Takasaka T, Kawamoto K, Nakamura K. 1987. A case-control study of nasal cancers. An occupational survey. Acta Otolaryngol Suppl 435:136–142. Teschke K, Marion SA, Vaughan TL, Morgan MS, Camp J. 1999. Exposures to wood dust in US industries and occupations, 1979 to 1997. Am J Ind Med 35:581–589. Teschke K, Morgan MS, Checkoway H, et al. 1997. Surveillance of nasal and bladder cancer to locate sources of exposure to occupational carcinogens. Occup Environ Med 54:443–451. Tola S, Hernberg S, Collan Y, Linderborg H, Korkala ML. 1980. A case-control study of the etiology of nasal cancer in Finland. Int Arch Occup Environ Health 46:79–85. Vaughan TL, Strader C, Davis S, Daling JR. 1986. Formaldehyde and cancers of the pharynx, sinus and nasal cavity: I. Occupational exposures. Int J Cancer 38:677–683. Walker JT, Bloom TF, Stern FB, Okun AH, Fingerhut MA, Halperin WE. 1993. Mortality of workers employed in shoe manufacturing. Scand J Work Environ Health 19:89–95. Walrath J, Decoufle P, Thomas TL. 1987. Mortality among workers in a shoe manufacturing company. Am J Ind Med 12:615–623. Walrath J, Fraumeni JF Jr. 1983. Mortality patterns among embalmers. Int J Cancer 31:407–411. Walrath J, Fraumeni JF Jr. 1984. Cancer and other causes of death among embalmers. Cancer Res 44:4638–4641. Yoon JH, Kim CH, Choi EC. 2002. Treatment outcomes of primary and recurrent inverted papilloma: An analysis of 96 cases. J Laryngol Otol 116:699–702. Yu M, Henderson B. 1996. Nasopharyngeal cancer. In: Schottenfeld D, Fraumeni JF, Jr., eds. Cancer Epidemiology and Prevention. New York: Oxford University Press, pp. 603–618.

Cancers of the Nasal Cavity and Paranasal Sinuses Zheng W, Blot WJ, Shu XO, et al. 1992. A population-based case-control study of cancers of the nasal cavity and paranasal sinuses in Shanghai. Int J Cancer 52:557–661. Zheng W, McLaughlin JK, Chow WH, Chien HT, Blot WJ. 1993. Risk factors for cancers of the nasal cavity and paranasal sinuses among white men in the United States. Am J Epidemiol 138:965–972. Zhu K, Levine RS, Brann EA, Hall HI, Caplan LS, Gnepp DR. 2002. Casecontrol study evaluating the homogeneity and heterogeneity of risk

619

factors between sinonasal and nasopharyngeal cancers. Int J Cancer 99:119–123. Zong Y, Lui K, Zhong B, Chen G, Wu W. 2001. Epstein-Barr virus infection of sinonasal lymphoepithelial carcinoma in Guangzhou. Chin Med J (Engl) 114:132–136.

31

Nasopharyngeal Cancer MIMI C. YU AND JIAN-MIN YUAN

C

ancer of the nasopharynx is a disease with a remarkable geographic and racial distribution worldwide. Except for a handful of populations, this is a rare human malignancy with an incidence well under 1 per 100,000 population per year. Regardless of race and geography, the commonest form of nasopharyngeal cancers are those arising from the epithelial cells lining the nasopharynx. These carcinomas (commonly referred to as NPC) constitute 75% to 95% of nasopharyngeal cancers in low-risk populations and virtually all nasopharyngeal cancers in high-risk populations (Parkin et al, 2002b). In this chapter, we shall discuss primarily the epidemiology of NPC.

CLASSIFICATION Anatomic Distribution The majority of NPCs arise in the lateral walls of the nasopharynx, especially from the pharyngeal recesses (fossae of Rosenmuller) and eustachian cushions. NPC may also arise in the superoposterior wall, in particular the roof of the nasopharynx, but only rarely in the anterior and inferior walls of the nasopharynx (Simons and Shanmugaratnam, 1982).

Histopathology On the basis of light microscopic studies, the WHO classified NPC into three histological types: (1) keratinizing squamous cell carcinoma, (2) differentiated non-keratinizing carcinoma, and (3) undifferentiated carcinoma. Keratinizing squamous cell carcinomas are further divided into well, moderately, or poorly differentiated squamous cell carcinomas (Shanmugaratnam and Sobin, 1991; IARC, 1997).

DEMOGRAPHIC PATTERNS International Patterns Most cancer registries only present incidence data for cancer of the nasopharynx as a whole. Therefore, rates for nasopharyngeal cancer, which, for most populations, are indistinguishable from their respective NPC rates, are used to compare worldwide incidence of NPC. In most parts of the world, annual incidence of NPC is below 1 per 100,000 for both men and women. Table 31–1 lists the handful of populations that deviate from this low-risk pattern. Highest rates are noted among the Cantonese who inhabit the central region of Guangdong Province in southern China, of which Hong Kong is a part (Fig. 31–1). Although all Chinese possess increased risk of NPC, rates generally decline as one travels from south to north China (Fig. 31–1). High rates comparable to the Hokkien-speaking Chinese in Taiwan are seen in natives of the Arctic region, of which the Northwest Territories in Canada is a part. Intermediate to high rates are observed among many indigenous people of Southeast Asia, including Thais, Vietnamese, Malays, and Filipinos. Finally, rates comparable to natives of Southeast Asia are seen among Arabs of North Africa (Table 31–1). There is a common feature across the populations demonstrating an elevated risk of NPC; consumption of preserved foods beginning at an early age is frequent among these peoples. Epidemiologic studies have linked childhood intake of locally consumed preserved foods to NPC development in all four groups of populations exhibiting

620

increased risk of NPC—Chinese, natives of Southeast Asia, natives of Arctic region, and Arabs of North Africa (see below). It is quite likely that the diverse groups of preserved foods from the different cultures actually share common carcinogenic constituents, or their precursors. It is interesting to note that nitrosamines/precursors and Epstein Barr virus (EBV) activating substances have been detected in a number of these preserved foods (see below).

Sex and Age Independent of race-ethnicity, rates of NPC are higher in men than women. For most populations, the male:female ratio is roughly 2–3 : 1 (Parkin et al., 2002b). Distributions of age-specific rates of NPC show distinct features across different populations. For most low-risk populations, NPC incidence rises monotonically with age, similar to the age distributions of most epithelial cancers. In contrast, among high-risk southern Chinese of both sexes, incidence of NPC increases with age until it peaks between 50 and 59 years, and shows a definite decline at older ages (Yu et al., 1981; Parkin et al., 2002b). The middle-age incidence peak noted among Southern Chinese can be interpreted as suggesting that exposure to the putative carcinogens occur relatively early in life. All Southern Chinese, irrespective of dialect group, share a popular practice of weaning their babies on preserved foods. Epidemiologic studies conducted in these populations have implicated higher risk with earlier age at exposure (see below), thus offering an explanation for the observed decline in NPC incidence following the middle-age peak. Several populations at low to moderate risk for NPC exhibit a minor peak in incidence among adolescents and young adults (early teens to early twenties); these include United States blacks and whites (Burt et al., 1992; SEER Program, 2000), natives (i.e., Thais and Kadazans) of Southeast Asia (Rothwell, 1979; Parkin et al., 2002b), and Indians (Balakrishnan, 1975; Parkin et al., 2002b).

Race-Ethnicity The high risk for NPC among Chinese is mainly confined to those residing in the southern provinces of Guangdong, Guangxi, Hunan, and Fujian (National Cancer Control Office and Nanjing Institute of Geography, 1979) (Fig. 31–1). Several distinct dialect groups inhabit this high-risk region, and these groups have been shown to exhibit varying rates of NPC. Specifically, the Cantonese men and women from central Guangdong show rates that are twice those of their counterparts in other dialect groups, including the Hakka, Hokkien, and Chiu Chau peoples (Yu et al., 1981; Li et al., 1985). Even after migration to Southeast Asia, the Cantonese continue to exhibit a twofold higher risk of NPC than the other dialect groups of south China (Lee et al., 1988). Southern Chinese migrants, irrespective of their country of migration, continue to exhibit high rates of NPC (Yu et al., 1981; Lee et al., 1988; Parkin et al., 2002b). However, successive generations of Southern Chinese living in low-risk countries including the United States (Yu et al., 1981) and Australia (Worth and Valentine, 1967) show continually declining rates. There is no evidence that China-born Chinese living in Southeast Asia experience higher rates of NPC relative to their Southeast Asian-born counterparts (Lee et al., 1988).

Nasopharyngeal Cancer Table 31–1. Populations at Increased Risk for Nasopharyngeal Cancer Age-Standardized (World) Incidence* Population

Male

Female

Chinese, Hong Kong Chinese, Taiwan Chinese, Shanghai Chinese, Tianjin Eskimos, Northwest Territories, Canada Thais, Bangkok Vietnamese, Hanoi Malays, Singapore Filipinos, Manila Tunisians, Sousse Algerians, Setif

21.4 8.9 4.2 1.7

8.3 3.4 1.5 0.5

9.2 4.5 10.4 6.8 7.2 4.9 6.3

6.0 1.6 4.6 2.0 2.5 1.5 2.2

Source: Data from Cancer Incidence in Five Continents, Volume VIII (Parkin et al., 2002b) and Cancer in Africa: Epidemiology and Prevention (Parkin et al., 2002a). *Per 100,000 person-years, 1993–1997.

Consumption of a number of preserved food items is an integral part of a traditional Southern Chinese diet. As stated above, some of the preserved foods have been linked to NPC development (see below). Chinese living in Southeast Asia tend to retain their traditional ways of life while their counterparts in the United States and other western countries are inclined to gradually adopt the lifestyles of their host countries. In other words, intake of preserved foods is likely to diminish over time with overseas Chinese living in western communities, but this trend would be absent from their Asian counterparts. The dialect groups of south China are distinct not only in their languages, but also in their dietary patterns, including the consumption of various preserved foods. Thus, the differential rates of NPC across the dialect groups are likely to be diet-related.

Socioeconomic Status Among populations with elevated risk of NPC, including Southern Chinese, indigenous people of Southeast Asia, and Arab of North Africa, lower social class is associated with a higher risk for NPC (Yu et al., 1981; Jeannel et al., 1990; Sriamporn et al., 1992). Consumption of preserved foods was found to be a major risk factor for NPC in these people (see below). Given that the NPC-associated preserved foods are among the least expensive foods available in those local populations, the observed inverse social class gradients with NPC risk should not be surprising.

621

In Los Angeles, the population-based cancer registry classifies cancer patients according to social class characteristics of their places of residence. Specifically, census information on income and/or educational levels of residents in the neighborhoods where cancer patients reside is used to rank cancer cases into one of five social class groupings. Figure 31–2 shows that in Asians, lower social class is related to higher NPC incidence in both men and women, although the trend is more distinct in men. Among non-Hispanic whites, an association between social class and disease risk is apparent only in men. Occupational exposure to dust and smoke has been found to be a risk factor for NPC in the United States (see below), and is likely one of the reasons for the observed inverse social class-NPC association that is confined to non-Hispanic white men.

Urbanization No difference in risk of NPC was found between urban and rural Southern Chinese populations, including those residing in Southeast Asia (Armstrong et al., 1979; Yu et al., 1981). In contrast, in the United States, non-Hispanic white residents of urban counties experience higher mortality rates of NPC relative to their counterparts in rural counties. For men, the urban/rural ratio of age-adjusted NPC mortality rate was 2.2, while the corresponding figure for women was 1.7 (Hoover et al., 1975).

Time Trends Table 31–2 presents the time trends in average annual incidence rates of nasopharyngeal cancer in Hong Kong and Singapore Chinese between 1973 and 1997. In Hong Kong Chinese, there is a monotonic decrease in NPC incidence over this 25-year period in both sexes. Rates during 1993–1997 were 35%–40% lower than the corresponding figures during 1973–1977. In contrast, rates of NPC were relatively stable among Singapore Chinese of both sexes over the 20-year period between 1973–1992. However, there is a visible drop in incidence for 1993 to 1997 relative to all previous 5-year periods, suggesting that rates in Singapore Chinese may follow the time trend noted in Hong Kong Chinese, after a lag time of 20 years. Use of salted fish to feed young children, a major risk factor for NPC in Hong Kong Chinese (see below), has been declining since the end of the Pacific War, as Hong Kong underwent rapid economic development (Geser et al., 1978; Yu et al., 1986). The economic transformation of Singapore, on the other hand, began around the time of its independence in the 1960s, some 20 years behind Hong Kong. The time trends in average annual rates of nasopharyngeal cancer in United States blacks and whites between 1973 and 1997 are given in Table 31–3. In US whites, there is a moderate but consistent decrease in NPC incidence over this 25-year period in both sexes.

Figure 31–1. Map of China showing provinces with high incidence of nasopharyngeal carcinoma.

622

PART IV: CANCER BY TISSUE OF ORIGIN

Figure 31–2. Incidence rates of nasopharyngeal cancer by socioeconomic status in Asians and non-Hispanic whites in Los Angeles County, California, 1972–1999.

Rates during 1993–1997 were roughly 20% lower than the corresponding figures during 1973–1977. Among US blacks, rates were relatively stable in women and showed suggestion of an increase in men (Table 31–3).

Survival Table 31–4 shows the 5-year relative survival rates of nasopharyngeal cancer patients in the United States during 1992 to 1999. In both sexes, blacks exhibited poorer survival relative to whites. In turn, the survival rates for all races are higher than the corresponding figures in whites for both men and women, suggesting that US Asians possessed the most favorable survival outcomes. Table 31–2. Time Trends in Average Annual Age-Standardized (World Population) Incidence Rates of Nasopharyngeal Cancer (Per 100,000 Person-Years) in Hong Kong and Singapore Chinese Hong Kong Chinese Period 1973–1977 1978–1982 1983–1987 1988–1992 1993–1997

Survival rates comparable to those observed in the United States have been reported for mostly Chinese patients in Hong Kong and Singapore. Lee et al. (1992) noted a 5-year relative survival rate of 57% among 2523 unselected cases treated during 1981–1985 at the Queen Elizabeth Hospital in Hong Kong. In Singapore, the relative survival rate in 270 patients treated at the Singapore General Hospital between July 1987 and December 1988 was 53% (Fong et al., 1996).

Incidence and Mortality in United States Table 31–5 gives the age-standardized incidence rates of nasopharyngeal cancer among all major racial-ethnic groups in Los Angeles County, California, between 1993 and 1997. Rates in Los Angeles

Table 31–4. Five-Year Relative Survival Rates (%) of Nasopharyngeal Cancer in the United States, 1992–1999

Singapore Chinese

Males

Females

Males

Females

32.9 30.0 28.5 24.3 21.5

14.4 12.9 11.2 9.5 8.3

19.4 18.1 18.1 18.5 16.7

7.5 7.9 7.4 7.3 5.5

Source: Data from Cancer Incidence in Five Continents Series (Waterhouse et al., 1982; Muir et al., 1987; Parkin et al., 1992; 1997; 2002b).

Males Females Total

Period 1973–1977 1978–1982 1983–1987 1988–1992 1993–1997

Blacks

Males

Females

Males

Females

0.59 0.58 0.55 0.50 0.47

0.25 0.23 0.23 0.21 0.21

0.81 0.90 0.96 0.96 0.92

0.33 0.39 0.37 0.22 0.32

Source: Data from the Surveillance, Epidemiology and End Results Program (2000).

Whites

All races

42.5 46.4 43.8

52.4 50.9 51.9

58.1 55.4 57.2

Source: Data from the Surveillance, Epidemiology and End Results Program (Ries et al., 2003).

Table 31–5. Age-Standardized Incidence Rates of Nasopharyngeal Cancer by Major Racial-Ethnic Groups in Los Angeles County, California, 1993–1997

Table 31–3. Time Trends in Average Annual Age-Standardized (World Population) Incidence Rates of Nasopharyngeal Cancer (Per 100,000 Person-Years) in United States Whites and Blacks Whites

Blacks

Males

Non-Hispanic White Hispanic white Black Chinese Filipino Japanese Korean

Females

No. Cases

Rate

No. Cases

Rate

66 21 22 64 26 3 6

0.5 0.5 0.9 7.6 3.7 0.6 1.1

35 13 10 23 15 0 3

0.3 0.2 0.3 2.4 1.6 — 0.5

Source: Data from Cancer Incidence in Five Continents, Volume VIII (Parkin et al., 2002b). Rates are per 100,000 person years.

Nasopharyngeal Cancer blacks and whites are similar to the corresponding rates reported for all SEER registries. The nationwide SEER rates for white men and women were 0.5 and 0.2, respectively. For black men and women, the rates were 0.9 and 0.3, respectively (Parkin et al., 2002b). The high rates noted in Los Angeles Chinese and Filipino are consistent with data on these two population groups in other parts of North America (Parkin et al., 2002b). The mortality rates of nasopharyngeal cancer in US blacks and whites resemble the pattern seen for incidence rates, being twice as high in blacks than whites. The age-adjusted (2000 US standard population) mortality rates in black men and women, and white men and women during 1996 to 2000 were 0.5, 0.2, 0.3, and 0.1 per 100,000 person-years, respectively (Ries et al., 2003).

ENVIRONMENTAL FACTORS Epstein-Barr Virus Infection The association between Epstein-Barr Virus (EBV) infection and NPC was first revealed by Old et al. (1966) who noticed that a high proportion of the sera from NPC patients in Africa and the United States reacted with antigen prepared from cultured Burkitt’s lymphoma cells. Since then, numerous studies have shown that NPC patients of all races and geographic locales consistently show highly elevated antibody titers to various EBV-associated antigens relative to control subjects (IARC, 1997). In a cohort study of roughly 10,000 men recruited from six townships in Taiwan and observed for an average of 15 years, baseline serum positivity for IgA antibodies against EBV capsid antigen and neutralizing antibodies against EBV DNase was found to be associated with a 33-fold increased risk for NPC. Even after excluding cases diagnosed within 5 years of follow-up, this EBV-NPC association remained highly significant with a 21-fold increase in NPC risk (Chien et al., 2001). Virtually all cases of NPC from high-risk areas and non-keratinizing/undifferentiated NPC from low- to intermediaterisk areas have detectable EBV DNA in their biopsy samples (IARC, 1997). However, there is strong evidence that the virus is not capable of inducing NPC by itself; other critical cofactors must also be present. EBV infection is ubiquitous throughout the world whereas NPC is rare except for a handful of populations (see above section titled “International Patterns of Incidence”). Throughout China, there is little variation in the prevalence of infection and the age at primary infection with EBV (Zeng, 1985), yet there is a more than 10-fold difference in NPC incidence between regions in North vs. South China (Parkin et al., 2002b).

Cantonese-Style Salted Fish and Other Preserved Foods An astute radiation oncologist named John Ho in Hong Kong first proposed in 1971 that Cantonese-style salted fish, a common item in the local diet and a popular weaning food, may be an etiological factor for NPC (Ho, 1971). A large number of case-control studies conducted in diverse (Cantonese, other Southern Chinese, Northern Chinese, and Thais) populations residing in different parts of Asia and North America have confirmed Ho’s hypothesis (Yu et al., 1986; 1988; 1989a; Ning et al., 1990; Sriamporn et al., 1992; Lee et al., 1994; Armstrong et al., 1998; Yuan et al., 2000a). Age at first exposure was established as an important determinant of risk in exposed individuals; earlier age at exposure was associated with a higher risk of disease (Yu et al., 1986; 1988; Ning et al., 1990). Experimental data have further strengthened the evidence for Cantonese-style salted fish as a human nasopharyngeal carcinogen. Rats fed this human food developed nasal cavity carcinomas in a dose-dependent manner (Huang et al., 1978; Yu et al., 1989b; Zheng et al., 1994), and presence of carcinogenic nitrosamines/precursors (Huang et al., 1981; Zou et al., 1994) and EBV activating substances (Shao et al., 1988) were repeatedly found in samples of this food. Cantonese-style salted fish is not a frequent food in many communities of Southern China. Studies conducted in these non-Cantonese

623

populations revealed that intake of a variety of other preserved foods also are related to increased risk of NPC. Again, the epidemiologic data suggest that higher disease risk is associated with an earlier age at exposure. These NPC-associated preserved foods include different types of salted and pickled leafy/stem vegetables and roots, fermented beans and bean pastes, salted and fermented eggs, and various seafood pastes (Yu et al., 1988; Ning et al., 1990; Lee et al., 1994; Armstrong et al., 1998; Ward et al., 2000; Yuan et al., 2000a). Presence of carcinogenic nitrosamines/precursors, and genotoxic and EBV activating substances were detected in many of these preserved foods (Shao et al., 1988; Poirier et al., 1989). Epidemiologic data seem to suggest that the carcinogenic potential of these non-fish preserved foods are not as high as that of the Cantonese-style salted fish. There is limited evidence that childhood consumption of salted fish, which is common among the indigenous peoples of Southeast Asia and natives of the Arctic region, is related to increased NPC risk in these populations with intermediate to high rates of NPC (Lanier et al., 1980; Armstrong and Armstrong, 1983). Samples of salted fish from Greenland have been found to contain carcinogenic nitrosamines/precursors, and genotoxic and EBV activating substances (Shao et al., 1988; Poirier et al., 1989). Preserved foods also are relatively common among the Arabs of North Africa who display elevated incidence of NPC. A well-designed case-control study examining diet and NPC risk in Tunisia showed statistically significant associations between NPC risk and childhood exposure to three common preserved foods (harissa, qaddid, and touklia) in the local diet (Jeannel et al., 1990). All three foods demonstrated presence of carcinogenic nitrosamines/precursors, and genotoxic and EBV activating substances (Shao et al., 1988; Poirier et al., 1989; Bouvier et al., 1995). Farrow et al. (1998) recently reported on the first study examining the role of preserved foods on NPC in a low-risk population, the black and white residents in five areas of the United States. Increased risk was significantly associated with increasing intake for non-keratinizing or undifferentiated NPC, the primary histologic subtypes of NPC in high-risk Southern Chinese (Chia et al., 2000).

Fresh Fruits and Vegetables Most case-control studies conducted among Southern Chinese noted a statistically significant deficit in intake of fresh fruit and vegetables among cases relative to control subjects. NPC cases were repeatedly shown to consume less citrus fruit, which are rich in vitamin C, an inhibitor of in vivo formation of nitrosamines. Other NPC-protective foods include orange vegetables (carrots and sweet potatoes), tomatoes, and various dark green vegetables, all of which are rich in carotenoids (Yu et al., 1986; 1989a; Ning et al., 1990; Lee et al., 1994; Armstrong et al., 1998; Yuan et al., 2000a). Given the close correlation between intake of fruit/vegetables and preserved foods (high consumers of the former tend to be low consumers of the latter), it is difficult to disentangle the effects of the two sets of foods in epidemiological data. Indeed, only some of the above studies demonstrated a statistically significant residual effect of fruits and vegetables after adjustment for intake of NPC-associated preserved foods. However, given its biological plausibility, a direct role of fresh fruits and vegetables in NPC risk reduction is quite likely.

Tobacco and Alcohol A number of epidemiologic studies conducted in high- and low-risk populations during the past decade have strongly implicated the nasopharynx as a tobacco-susceptible cancer site. However, the magnitude of risk associated with a given level of smoking is much less pronounced than those for other upper respiratory sites. Ever smokers exhibit a roughly 30%–100% excess risk relative to lifelong nonsmokers. Risk of NPC is positively associated with number of cigarettes smoked on a regular basis and inversely associated with age at starting to smoke regularly. Risk is reduced among ex-smokers relative to those who continue to smoke. Generally, risk is 2- to 4-fold among the heaviest smokers relative to lifelong nonsmokers (Yu et al.,

624

PART IV: CANCER BY TISSUE OF ORIGIN

1990; Nam et al., 1992; West et al., 1993; Chow et al., 1993; Zhu et al., 1995; Vaughan et al., 1996; Cheng et al., 1999; Yuan et al., 2000b). In comparison, relative risks of 20 or higher are repeatedly observed for cancers of the lung and larynx among heavy smokers (Centers for Disease Control, 1989). It appears that either the nasopharyngeal epithelium is less sensitive to the carcinogenic effects of tobacco constituents, or the exposure level of the target cells to these compounds is lower for the nasopharynx than other upper respiratory sites. Several studies in Chinese have examined the possible role of passive smoking, both during childhood and during adult years, on risk of NPC. Results are inconsistent (Yu et al., 1988; 1990; Cheng et al., 1999; Yuan et al., 2000b). Given the moderate dose-risk relationship between active smoking and NPC, a passive-smoking/NPC association, even if it exists, may not be observable in epidemiologic studies using traditional assessment of self-reported passive smoking with inherently large measurement errors. The possible association between alcohol use and NPC has been investigated in a number of case-control studies among Chinese in and outside of China and among primarily whites in the United States. Virtually all studies conducted in Chinese yielded null results (Cheng et al., 1999; Yuan et al., 2000b). On the other hand, the two US studies both noted substantial use of alcohol to be significantly related to NPC risk, after adjustment for cigarette smoking (Nam et al., 1992; Vaughan et al., 1996). Cigarette smoking and alcohol drinking are highly correlated lifestyle habits in the United States and other western societies. It is possible that the alcohol effect on NPC risk observed in the latter studies is the result of residual confounding from cigarette smoking.

Occupational Exposures Formaldehyde is a recognized nasal cavity carcinogen in rodents (IARC, 1995). Thus, the report by Blair et al. (1986) that industrial workers exposed to formaldehyde experienced a statistically significant excess risk of NPC raised an alarm and led to a number of subsequent investigations. Meta analysis of over 30 epidemiological studies yielded a statistically significant, dose-dependent association between formaldehyde exposure and risk of NPC (Blair et al., 1990; Partanen, 1993). The International Agency for Research on Cancer evaluated the carcinogenic risk of formaldehyde to humans (IARC, 1995) and concluded that “the epidemiological studies suggest a causal relationship between exposure to formaldehyde and nasopharyngeal cancer.” The nasopharynx traps primarily medium-sized particles (5–10 mm) in inspired air, including dust particles from wood (Armstrong et al., 2000). In conjunction with earlier studies, a pooled analysis involving close to 29,000 wood workers in Britain and the United States (Demers et al., 1995), and two recent large-scale case-control studies in separate Chinese populations (Armstrong et al., 2000; Hildesheim et al., 2001) have yielded strong evidence that intense exposure to wood dust (as occurs under occupational settings) is associated with a durationdependent, increased risk of NPC. There is some suggestion that exposure to chlorophenols, which serve as wood preservatives, independently contributes towards wood workers’ high risk for NPC (Mirabelli et al., 2000). Smoke particles from incomplete combustion of coal, wood, and other materials also are of the size and weight to be deposited mostly in the nasopharynx (Armstrong et al., 2000). There is some evidence that intense exposure to smoke may be a risk factor for NPC; studies conducted in Chinese and in the United States have reported increasing risk of NPC with increasing duration of exposure to smoke on the job (Henderson et al., 1976; Armstrong et al., 1983; Yu et al., 1990; Armstrong et al., 2000).

Herbal Drugs A number of Chinese herbs have been shown to contain EBV inducing substances (Zeng et al., 1994), raising the possibility that this might represent a cause of NPC in high-risk Chinese. Yu et al. (1986; 1988; 1989a) conducted a series of case-control studies in Southern China to examine if frequency of use of one of the most popular herbal

formulations, either during childhood or in adult years, was related to NPC risk. All of their results were null. Hildesheim et al. (1992), in a case-control study conducted in the Philippines, asked subjects about ever use of any herbal medicines, and noted a statistically significant 2.5-fold risk among ever users. However, the latter study is difficult to interpret for the following reasons. Recall bias is a serious concern when a non-specific question (such as general use of herbal medicines without naming specific formulations) was asked under a case-control setting. In addition, use of herbal medicines is part of the “traditional” lifestyle, an established risk factor for NPC in Chinese as well as Southeast Asians. In other words, use of herbal medicine may simply be a marker of the NPC-related lifestyle.

HOST FACTORS Nitrosamine-Metabolizing Genes The cytochrome P450 2E1 (CYP2E1) enzyme catalyzes the metabolic activation of low-molecular weight nitrosamines such as those detected in NPC-associated foods. A variant form of the gene that is detectable by Rsa I digestion (the c2 allele) has been shown to exhibit higher enzymatic activity. If dietary nitrosamines from preserved foods are indeed playing a direct role in NPC development, exposed individuals possessing varying CYP2E1 genotypes may experience differential levels of NPC risk. Hildeheim et al. (1997) compared the CYP2E1 genotypes of 364 NPC cases and 320 population control subjects in Taiwan. Subjects possessing the c2/c2 genotype experienced a statistically significant 2.6-fold risk relative to those with one or two copies of the wild-type allele. This first study of metabolic genotype in relation to NPC risk further strengthens the notion that nitrosaminecontaining preserved foods are important human nasopharyngeal carcinogens.

Human Leukocyte Antigen Genes Human leukocyte antigen (HLA) molecules on cell surface are thought to be critical for the identification of foreign antigens, including viral peptides, by the host’s immune system. Since EBV infection is a risk factor for NPC, it is possible that individuals possessing HLA alleles with reduced efficiency in triggering an immune response to EBV may experience increased risk of NPC (Goldsmith et al., 2002; Hildesheim et al., 2002). A number of case-control studies conducted among highrisk Southern Chinese have reported statistically significant associations between specific HLA alleles and NPC risk. Goldsmith et al. (2002) performed a meta-analysis on 13 published studies in Chinese, which used low-resolution serotyping techniques to identify distinct HLA alleles. The authors found enhanced NPC risk to be associated with HLA alleles A2, B14, and B46, while reduced risk was associated with alleles A11, B13, and B22. Hildesheim et al. (2002) used high-resolution, PCR-based genotyping techniques to identify HLA alleles in relation to NPC risk in a case-control study conducted in Taiwan. Results are consistent with those reported by Goldmith et al. (2002). Case-control studies conducted in non-Chinese populations also have suggested an association between HLA profile and NPC risk. These populations include intermediate-risk Southeast Asians and North Africans (Chan et al., 1985; Mokni-Baizig et al., 2001; Pimtanothai et al., 2002) and low-risk whites (Kruger et al., 1981; Simons and Shanmugaratnam, 1982; Burt et al., 1994). The small sample size and other design limitations of these studies together with their diverse results preclude any firm conclusion from these findings.

PREVENTIVE MEASURES There is strong evidence implicating dietary factors (exposure to preserved foods, especially during childhood) as a major cause of NPC in the handful of populations with raised incidence of this disease (southern Chinese, natives of the Arctic region and Southeast Asia, Arabs of north Africa). These at-risk populations should be educated

Nasopharyngeal Cancer about the carcinogenic potential of those local foods, and mothers especially should be discouraged from feeding those foods to their young children. References Armstrong RW, Armstrong MJ. 1983. Environmental risk factors and nasopharyngeal carcinoma in Selangor, Malaysia: A cross-ethnic perspective. Ecol Dis 2:185–198. Armstrong RW, Armstrong MJ, Yu MC, Henderson BE. 1983. Salted fish and inhalants as risk factors for nasopharyngeal carcinoma in Malaysian Chinese. Cancer Res 43:2967–2970. Armstrong RW, Imrey PB, Lye MS, Armstrong MJ, Yu MC, Sani S. 1998. Nasopharyngeal carcinoma in Malaysian Chinese: Salted fish and other dietary exposures. Int J Cancer 77:228–235. Armstrong RW, Imrey PB, Lye MS, Armstrong MJ, Yu MC, Sani S. 2000. Nasopharyngeal carcinoma in Malaysian Chinese: Occupational exposures to particles, formaldehyde and heat. Int J Epidemiol 29:991–998. Armstrong RW, Kannan Kutty M, Dharmalingam SK, Ponnudurai JR. 1979. Incidence of nasopharyngeal carcinoma in Malaysia, 1968–1977. Br J Cancer 40:557–567. Balakrishnan V. 1975. An additional younger-age peak for cancer of the nasopharynx. Int J Cancer 15:651–657. Blair A, Saracci R, Stewart PA, Hayes RB, Shy C. 1990. Epidemiologic evidence on the relationship between formaldehyde exposure and cancer. Scand J Work Environ Health 16:381–393. Blair A, Stewart P, O’Berg M, et al. 1986. Mortality among industrial workers exposed to formaldehyde. J Natl Cancer Inst 76:1071–1084. Bouvier G, Hergenhahn M, Polack A, Bornkamm GW, de The G, Bartsch H. 1995. Characterization of macromolecular lignins as Epstein-Barr virus inducer in foodstuff associated with nasopharyngeal carcinoma risk. Carcinogenesis 16:1879–1885. Burt RD, Vaughan TL, McKnight B. 1992. Descriptive epidemiology and survival analysis of nasopharyngeal carcinoma in the United States. Int J Cancer 52:549–556. Burt RD, Vaughan TL, Nisperos B, Swanson M, Berwick M. 1994. A protective association between the HLA-A2 antigen and nasopharyngeal carcinoma in US Caucasians. Int J Cancer 56:465–467. Centers for Disease Control. 1989. Reducing the Health Consequences of Smoking: 25 Years of Progress. A report of the Surgeon General 1989. Rockville, Maryland: US Department of Health and Human Services, Public Health Service, Publications CDC 89–8411. Chan SH, Chew CT, Prasad U, Wee GB, Srinivasan N, Kunaratnam N. 1985. HLA and nasopharyngeal carcinoma in Malays. Br J Cancer 51:389–392. Cheng YJ, Hildesheim A, Hsu MM, et al. 1999. Cigarette smoking, alcohol consumption and risk of nasopharyngeal carcinoma in Taiwan. Cancer Causes Control 10:201–207. Chia KS, Seow A, Lee HP, Shanmugaratnam K. 2000. Cancer Incidence in Singapore 1993–1997. Singapore: Singapore Cancer Registry, Report No. 5. Chien YC, Chen JY, Liu MY, et al. 2001. Serologic markers of Epstein-Barr virus infection and nasopharyngeal carcinoma in Taiwanese men. N Engl J Med 345:1877–1882. Chow WH, McLaughlin JK, Hrubec Z, Nam JM, Blot WJ. 1993. Tobacco use and nasopharyngeal carcinoma in a cohort of US veterans. Int J Cancer 55:538–540. Demers PA, Boffetta P, Kogevinas M, et al. 1995. Pooled reanalysis of cancer mortality among five cohorts of workers in wood-related industries. Scand J Work Environ Health 21:179–190. Farrow DC, Vaughan TL, Berwick M, Lynch CF, Swanson GM, Lyon JL. 1998. Diet and nasopharyngeal cancer in a low-risk population. Int J Cancer 78:675–679. Fong KW, Chua EJ, Chua ET, et al. 1996. Patient profile and survival in 270 computer tomography-staged patients with nasopharyngeal cancer treated at the Singapore General Hospital. Ann Acad Med Singapore 25:341–346. Geser A, Charnay N, Day NE, de The G, Ho HC. 1978. Environmental factors in the etiology of nasopharyngeal carcinoma: Report on a case-control study in Hong Kong. In: de The G, Ito Y, eds. Nasopharyngeal Carcinoma: Etiology and Control. IARC Scientific Publications No. 20. Lyon: International Agency for Research on Cancer. Goldsmith DB, West TM, Morton R. 2002. HLA associations with nasopharyngeal carcinoma in Southern Chinese: A meta-analysis. Clin Otolaryngol 27:61–67. Henderson BE, Louie E, SooHoo Jing J, Buell P, Gardner MB. 1976. Risk factors associated with nasopharyngeal carcinoma. N Engl J Med 295:1101–1106. Hildesheim A, Anderson LM, Chen CJ, et al. 1997. CYP2E1 genetic polymorphisms and risk of nasopharyngeal carcinoma in Taiwan. J Natl Cancer Inst 89:1207–1212.

625

Hildesheim A, Apple RJ, Chen CJ, et al. 2002. Association of HLA class I and II alleles and extended haplotypes with nasopharyngeal carcinoma in Taiwan. J Natl Cancer Inst 94:1780–1789. Hildesheim A, Dosemeci M, Chan CC, et al. 2001. Occupational exposure to wood, formaldehyde, and solvents and risk of nasopharyngeal carcinoma. Cancer Epidemiol Biomarkers Prev 10:1145–1153. Hildesheim A, West S, DeVeyra E, et al. 1992. Herbal medicine use, EpsteinBarr virus, and risk of nasopharyngeal carcinoma. Cancer Res 52:3048– 3051. Ho JH. 1971. Genetic and environmental factors in nasopharyngeal carcinoma. In: Nakahara W, Nishioka K, Hirayama T, Ito Y, eds. Recent Advances in Human Tumor Virology and Immunology. Tokyo: University of Tokyo Press, pp. 275–295. Hoover R, Mason TJ, McKay FW, Fraumeni JF. 1975. Geographic patterns of cancer mortality in the United States. In: Fraumeni JF, ed. Persons at High Risk of Cancer: An Approach to Cancer Etiology and Control. New York: Academic Press, pp. 343–360. Huang DP, Ho JH, Saw D, Teoh TB. 1978. Carcinoma of the nasal and paranasal regions in rats fed Cantonese salted marine fish. In: de The G, Ito Y, eds. Nasopharyngeal Carcinoma: Etiology and Control. IARC Scientific Publications No. 20. Lyon: International Agency for Research on Cancer, pp. 315–328. Huang DP, Ho JH, Webb KS, Wood BJ, Gough TA. 1981. Volatile nitrosamines in salt-preserved fish before and after cooking. Food Cosmet Toxicol 19:167–171. IARC. 1995. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Volume 62, Wood Dust and Formaldehyde. Lyon: International Agency for Research on Cancer. IARC. 1997. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. Volume 70. Epstein-Barr Virus and Kaposi’s Sarcoma Herpesvirus/Human Herpesvirus 9. Lyon: International Agency for Research on Cancer. Jeannel D, Hubert A, de Vathaire F, et al. 1990. Diet, living conditions and nasopharyngeal carcinoma in Tunisia—a case-control study. Int J Cancer 46:421–425. Kruger J, Ieromnimon V, Dahr W. 1981. Frequencies of HLA antigens in patients with NPC. In: Grundmann E, Krueger GR, Ablashi DV, eds. Cancer Campaign, Volume 5, Nasopharyngeal Carcinoma. Stuttgart: Gustav Fischer Verlag, pp. 201–203. Lanier A, Bender T, Talbot M, et al. 1980. Nasopharyngeal carcinoma in Alaskan Eskimos Indians, and Aleuts: A review of cases and study of Epstein-Barr virus, HLA, and environmental risk factors. Cancer 46:2100–2106. Lee AW, Poon YF, Foo W, et al. 1992. Retrospective analysis of 5037 patients with nasopharyngeal carcinoma treated during 1976–1985: Overall survival and patterns of failure. Int J Radiat Oncol Biol Phys 23:261– 270. Lee HP, Duffy SW, Day NE, Shanmugaratnam K. 1988. Recent trends in cancer incidence among Singapore Chinese. Int J Cancer 42:159–166. Lee HP, Gourley L, Duffy SW, Esteve J, Lee J, Day NE. 1994. Preserved foods and nasopharyngeal carcinoma: A case-control study among Singapore Chinese. Int J Cancer 59:585–590. Li CC, Yu MC, Henderson BE. 1985. Some epidemiologic observations of nasopharyngeal carcinoma in Guangdong, People’s Republic of China. Natl Cancer Inst Monogr 69:49–52. Mirabelli MC, Hoppin JA, Tolbert PE, Herrick RF, Gnepp DR, Brann EA. 2000. Occupational exposure to chlorophenol and the risk of nasal and nasopharyngeal cancers among US men aged 30 to 60. Am J Ind Med 37:532–541. Mokni-Baizig N, Ayed K, Ayed FB, et al. 2001. Association between HLA-A/-B antigens and -DRB1 alleles and nasopharyngeal carcinoma in Tunisia. Oncology 61:55–58. Muir C, Waterhouse J, Mack T, Powell J, Whelan SL. 1987. Cancer Incidence in Five Continents, Volume V. IARC Scientific Publications, No. 88. Lyon: International Agency for Research on Cancer. Nam JM, McLaughlin JK, Blot WJ. 1992. Cigarette smoking, alcohol, and nasopharyngeal carcinoma: A case-control study among US whites. J Natl Cancer Inst 84:619–622. National Cancer Control Office, Nanjing Institute of Geography. 1979. Atlas of Cancer Mortality in the People’s Republic of China. Shanghai: China Map Press. Ning JP, Yu MC, Wang QS, Henderson BE. 1990. Consumption of salted fish and other risk factors for nasopharyngeal carcinoma (NPC) in Tianjin, a low-risk region for NPC in the People’s Republic of China. J Natl Cancer Inst 82:291–296. Old LJ, Boyse EA, Oettgen HF, et al. 1966. Precipitating antibody in human serum to an antigen present in cultured Burkitt’s lymphoma cells. Proc Natl Acad Sci 56:1699–1704.

626

PART IV: CANCER BY TISSUE OF ORIGIN

Parkin DM, Ferlay J, Hamdi-Cherif M, et al. 2002a. Cancer in Africa: Epidemiology and Prevention. IARC Scientific Publications No. 153. Lyon: International Agency for Research on Cancer. Parkin DM, Muir CS, Whelan SL, Gao Y-T, Ferlay J, Powell J. 1992. Cancer Incidence in Five Continents, Volume VI. IARC Scientific Publications, No. 120. Lyon: International Agency for Research on Cancer. Parkin DM, Whelan SL, Ferlay J, Raymond L, Young J. 1997. Cancer Incidence in Five Continents, Volume VII. IARC Scientific Publications No. 143. Lyon: International Agency for Research on Cancer. Parkin DM, Whelan SL, Ferlay J, Teppo L, Thomas D. 2002b. Cancer Incidence in Five Continents, Volume VIII. IARC Scientific Publications No. 155. Lyon: International Agency for Research on Cancer. Partanen T. 1993. Formaldehyde exposure and respiratory cancer—a metaanalysis of the epidemiologic evidence. Scand J Work Environ Health 19:8–15. Pimtanothai N, Charoenwongse P, Mutirangura A, Hurley CK. 2002. Distribution of HLA-B alleles in nasopharyngeal carcinoma patients and normal controls in Thailand. Tissue Antigens 59:223–225. Poirier S, Bouvier G, Malaveille C, et al. 1989. Volatile nitrosamine levels and genotoxicity of food samples from high-risk areas for nasopharyngeal carcinoma before and after nitrosation. Int J Cancer 44:1088–1094. Ries LA, Eisner MP, Kosary CL, et al. 2003. SEER Cancer Statistics Review, 1975–2000. Bethesda, Maryland: National Cancer Institute (http://seer.cancer.gov/csr/1975–2000). Rothwell RI. 1979. Juvenile nasopharyngeal carcinoma in Sabah (Malaysia). Clin Oncol 5:353–358. SEER Program. 2000. SEER*Stat Database: Incidence—SEER 9 Regs, Aug 1999 Sub (1973–1997), National Cancer Institute (www.seer.cancer.gov), DCCPS Surveillance Research Program, Cancer Statistics Branch, Released April 2000, based on the August 1999 submission. Shanmugaratnam K, Sobin LH. 1991. Histological Typing of Tumours of the Upper Respiratory Tract and Ear. 2nd ed. Berlin: Springer-Verlag. Shao YM, Poirier S, Ohshima H, et al. 1988. Epstein-Barr virus activation in Raji cells by extracts of preserved food from high risk areas for nasopharyngeal carcinoma. Carcinogenesis 9:1455–1457. Simons MJ, Shanmugaratnam K. 1982. The Biology of Nasopharyngeal Carcinoma. UICC Technical Report Series, Volume 71. Geneva: International Union Against Cancer, pp. 46–54. Sriamporn S, Vatanasapt V, Pisani P, Yongchaiyudha S, Rungpitarangsri V. 1992. Environmental risk factors for nasopharyngeal carcinoma: A casecontrol study in northeastern Thailand. Cancer Epidemiol Biomarkers Prev 1:345–348. Vaughan TL, Shapiro JA, Burt RD, et al. 1996. Nasopharyngeal cancer in a low-risk population: Defining risk factors by histological type. Cancer Epidemiol Biomarkers Prev 5:587–593. Ward MH, Pan WH, Cheng YJ, et al. 2000. Dietary exposure to nitrite and nitrosamines and risk of nasopharyngeal carcinoma in Taiwan. Int J Cancer 86:603–609.

Waterhouse J, Muir CS, Shanmugaratnam K, Powell J. 1982. Cancer Incidence in Five Continents, Volume IV. IARC Scientific Publications, No. 42. Lyon: International Agency for Research on Cancer. West S, Hildesheim A, Dosemeci M. 1993. Non-viral risk factors for nasopharyngeal carcinoma in the Philippines: Results from a case-control study. Int J Cancer 55:722–727. Worth RM, Valentine R. 1967. Nasopharyngeal carcinoma in New South Wales, Australia. In: Muir CS, Shanmugaratnam K, eds. Cancer of the Nasopharynx. UICC Monograph Series, Volume 1. Copenhagen: UICC, pp. 73–76. Yu MC, Garabrant DH, Huang TB, Henderson BE. 1990. Occupational and other non-dietary risk factors for nasopharyngeal carcinoma in Guangzhou, China. Int J Cancer 45:1033–1039. Yu MC, Ho JH, Lai SH, Henderson BE. 1986. Cantonese-style salted fish as a cause of nasopharyngeal carcinoma: Report of a case-control study in Hong Kong. Cancer Res 46:956–961. Yu MC, Ho JH, Ross RK, Henderson BE. 1981. Nasopharyngeal carcinoma in Chinese—salted fish or inhaled smoke? Prev Med 10:15–24. Yu MC, Huang TB, Henderson BE. 1989a. Diet and nasopharyngeal carcinoma: A case-control study in Guangzhou, China. Int J Cancer 43:1077–1082. Yu MC, Mo CC, Chong WX, Yeh FS, Henderson BE. 1988. Preserved foods and nasopharyngeal carcinoma: A case-control study in Guangxi, China. Cancer Res 48:1954–1959. Yu MC, Nichols PW, Zou XN, Estes J, Henderson BE. 1989b. Induction of malignant nasal cavity tumours in Wistar rats fed Chinese salted fish. Br J Cancer 60:198–201. Yuan JM, Wang XL, Xiang YB, Gao YT, Ross RK, Yu MC. 2000a. Preserved foods in relation to risk of nasopharyngeal carcinoma in Shanghai, China. Int J Cancer 85:358–363. Yuan JM, Wang XL, Xiang YB, Gao YT, Ross RK, Yu MC. 2000b. Non-dietary risk factors for nasopharyngeal carcinoma in Shanghai, China. Int J Cancer 85:364–369. Zeng Y. 1985. Seroepidemiological studies on nasopharyngeal carcinoma in China. Adv Cancer Res 44:121–138. Zeng Y, Zhong JM, Ye SQ, et al. 1994. Screening of Epstein-Barr virus early antigen expression inducers from Chinese medicinal herbs and plants. Biomed Environ Sci 7:50–55. Zheng X, Luo Y, Christensson B, Drettner B. 1994. Induction of nasal and nasopharyngeal tumours in Sprague-Dawley rats fed with Chinese salted fish. Acta Otolaryngol 114:98–104. Zhu K, Levine RS, Brann EA, Gnepp DR, Baum MK. 1995. A populationbased case-control study of the relationship between cigarette smoking and nasopharyngeal cancer (United States). Cancer Causes Control 6:507–512. Zou XN, Lu SH, Liu B. 1994. Volatile N-nitrosamines and their precursors in Chinese salted fish—a possible etiological factor for NPC in china. Int J Cancer 59:155–158.

32

Cancer of the Larynx ANDREW F. OLSHAN

I

n the year 2000, worldwide, there were approximately 142,168 new cases of laryngeal cancer and 78,573 deaths (Parkin et al., 2001; Ferlay et al., 2001). It has been estimated that there will be 9880 (7920 males, 1961 females) new cases diagnosed and 3770 deaths from larynx cancer in the United States in the year 2005 (American Cancer Society, 2005). This chapter will focus on the most common and beststudied histologic type of laryngeal cancer, squamous cell carcinoma of the larynx. Information will also be presented on etiologic heterogeneity among anatomic subsites of laryngeal cancer. The identification of the major exposures related to its occurrence, such as tobacco use, alcohol use, and dietary factors, has provided a strong rationale and scientific basis for primary prevention.

CLASSIFICATION Anatomic Distribution The primary laryngeal subsite classification currently used in the International Classification of Diseases for Oncology (ICD-O, Third Edition, Fritz et al., 2000) includes glottis, supraglottis, subglottis, and laryngeal cartilage. Glottic tumors arise from the true vocal cords including the anterior and posterior commissures. Supraglottic tumors are superior to the true vocal cords and include the false vocal cords, arytenoids, epiglottis, and aryepiglottic folds. Subglottic tumors form below the apex of the ventricle and extend to the lower border of the cricoid cartilage. According to recent (1995–2000) data from the Surveillance, Epidemiology, and End Results Program of the United States National Cancer Institute (SEER, 2003) the distribution of tumor sites of 6210 cases of squamous cell carcinoma of the larynx was: 56.1% glottis, 32.7% supraglottis, 1.4% subglottis, 0.4% larynx cartilage, 3.5% tumors that overlap the boundaries of two or more subsites and whose point of origin cannot be determined, and 5.9% not otherwise specified. Males had more glottic tumors than did females (61% vs. 38%) and fewer supraglottic tumors (28% vs. 50%). Whites had more glottic tumors (58%) than did blacks (45%). Of 6021 cases of squamous cell laryngeal cancer with staging data (SEER 1995–2000), 48% were staged as localized. Glottic cancer had the most localized cases (62.7% vs. 31% of supraglottic, 18.4% of subglottic, and 42% of laryngeal cartilage). Some of these differences in stage can be explained by an earlier diagnosis of glottic cancer due to hoarseness and because of the more extensive lymphatic supply to other sites. Of all tumors, 53% among males and 44% among females were localized. The percentage of localized glottic tumors was similar regardless of sex (63% among males; 61% among females).

Precursor Lesions The development of laryngeal cancer is associated with a series of epithelial changes of the laryngeal mucosa including early lesions such as laryngeal keratosis (Schwartz, 1999). Laryngeal keratosis is a white keratotic plaque similar to oral leukoplakia. They are usually bilateral and found on the vocal cords. Sometimes they may also appear as a red patch (erythroplasia or erythroplakia). Keratoses are found adjacent to larynx cancer in 18%–43% of cases and have a malignant transformation rate of 1%–40% (Bouquot and Gnepp, 1991a); however, many of these studies included smaller selected samples and a popu-

lation study found a transformation rate of 0.9% (Bouquot and Gnepp, 1991b). Keratoses with atypia have a higher transformation rate (6%–40%) than those without atypia (0%–16%). One autopsy study reported that 19.6% of men (80.7% of whom were smokers) had detectable laryngeal keratosis at death (Muller and Krohn, 1980). The estimated annual incidence (per 100,000 persons) of laryngeal keratosis has been reported as 10.2 for men and 2.1 for women (Bouquot et al., 1989). Age-specific incidence data indicate that the rate does not rise with age, but declines after about age 64 years (Bouquot et al., 1991b). More men had diagnoses of laryngeal keratosis than women (64%–94%) and 91% of persons who had laryngeal keratosis with atypia were men (Bouquot and Gnepp, 1991a). Tobacco use has been associated strongly with the occurrence of laryngeal keratosis (Bouquot and Gnepp, 1991a). Voice abuse, alcohol use, and nutritional deficiencies have also been proposed as risk factors (Bouquot and Gnepp, 1991a). Approximately 7% of all laryngeal cancer cases reported to SEER (1995–2000) were classified as carcinoma in situ. Laryngeal carcinoma in situ has been shown to have an average transformation rate of 29.0% (range 3.5%–90%; Bouquot and Gnepp, 1991a). Further research on precursor lesions can provide important data on the early molecular events and mechanisms of laryngeal carcinogenesis. Larger population studies are especially important as a source of systematically identified cases to obtain risk factor data and bio-specimens.

Molecular Pathogenesis Much has been learned about the molecular mechanisms of head and neck carcinogenesis in recent years. It has been assumed that laryngeal cancer generally follows the proposed models, but comprehensive data specifically on the molecular pathogenesis of laryngeal cancer has not been yet obtained. Consistent with the Vogelstein model, the malignant phenotype is presumed to involve the accumulation of a series of genetic events (Ha and Califano, 2002), primarily inactivation of tumor suppressor genes and activation of oncogenes. It has been suggested that between 6 and 10 independent genetic events are involved in the development of squamous cell carcinoma of the head and neck (Renan, 1993). Cytogenetic studies provide evidence on critical chromosomal regions that have undergone losses or gains. These studies have reported deletions on 3p, 5q, 8p, 9p, 18q, and 21q using tumor samples and cell lines (Carey et al., 1993; Van Dyke et al., 1994). Recent studies that examined loss of heterozygosity (LOH) led to the development of a genetic progression model for head and neck cancer (Califano et al., 1996). The proposed sequence, location of events, and corresponding changes include: 9p21 LOH (normal to hyperplasia); 3p21, 17p13 LOH (hyperplasia to dysplasia); 11q13, 13q21, 14q32 LOH (dysplasia to carcinoma in situ); and 6p, 8, 4q27, 10q23 LOH (CIS to carcinoma). Some of these loci may contain candidate tumor suppressor genes. The 9p21 locus contains the putative tumor suppressor gene, p16MTS1/CDK41. Homozygous deletion of the p16 locus occurs in 50% of head and neck tumors; promotor hypermethylation is an alternative p16-gene-silencing mechanism (Reed et al., 1996). The 17p locus contains an important tumor suppressor gene, p53. Almost 50% of head and neck tumors contain a p53 mutation (Greenblatt et al., 1994). Specific mutational patterns (spectra) of the p53

627

628

PART IV: CANCER BY TISSUE OF ORIGIN

gene have been related to tobacco and other exposures (Brennan et al., 1995; Olshan et al., 1997). 3p is a frequent site of loss in head and neck tumors. The fragile histidine triad (FHIT) gene has been suggested as an important 3p locus in head and neck cancer, but requires further investigation (Ha and Califano, 2002). The 13q region contains the retinoblastoma (Rb) and BRCA2 genes, but these do not appear to be altered in head and neck tumors (Ha and Califano, 2002). The 11q13 region is amplified in about one-third of head and neck tumors; it contains the cyclin D1 gene (PRAD1). Other regions that show allele loss or gain in head and neck tumors are being investigated to determine specific critical genes.

DEMOGRAPHIC PATTERNS Incidence and Mortality Laryngeal cancer incidence rates during the period 1995–2000 were estimated using the latest SEER public use file (SEER, 2003). We restricted the rate calculations to the 5780 cases of malignant squamous cell carcinoma (ICDO, second edition codes C32, 8050–8077, Percy et al., 1990). Overall, the incidence (per 100,000, age-adjusted to the 2000 US standard population) is higher among males than females (7.0 vs. 1.5) and higher among blacks than whites (6.4 vs. 3.9) (Table 32–1). The rate is notably higher among black men (11.8 vs. 6.9 among white men). The rate among all other ethnic groups is lower at 2.0, based on only 252 cases. The rates for Hispanic groups were low (white Hispanic = 2.8, Non-white Hispanic = 0.4; SEER 1995–2000, 11 registries). Only 85 cases were reported among persons aged younger than 40 years at diagnosis. The incidence rate was 1.5 for persons aged 40–44 years at diagnosis and 6.1 for persons aged 50–54 years. The incidence rates sharply increase after age 54 years at diagnosis (aged 55–59 years = 10.0 per 100,000; 60–64 = 15.3; 65–69 = 19.0; 70–74 = 20.0; 75–79 = 16.9). Among the nine SEER registries (1995–2000), the incidence rate ranged from a low of 2.0 in Utah to 5.2 in Connecticut. Of 23,141 cases, the overall age-adjusted mortality (1995–2000) was 1.4 per 100,000; 2.7 among males and 0.5 among females (Table 32–1). The mortality rate among blacks was approximately twice that among whites, overall (2.9 vs. 1.3) and by sex (black males = 5.7, white males = 2.9; black females = 0.9, white females = 0.6). Mortality among Hispanics was lower (white Hispanic = 1.2, non-white Hispanic = 0.5).

Time Trends There is evidence that the incidence of laryngeal cancer has declined in the United States. Based upon SEER data, the age-adjusted rate was 4.8 per 100,000 in the year 1973, 5.0 in 1980, 4.9 in 1990, and 3.8 in 2000 (APC, annual percent change = -1.0). Among males, the rate declined from 9.1 in 1973 to 6.9 in 2000 (APC = -1.2); among females the rate increased slightly (APC = 0.2). A larger decline in incidence from 1973–2000 was seen for whites (APC = -1.0) than for blacks (APC = -0.2). Mortality from larynx cancer has declined slightly. According to data from TNCS and SEER, the estimated age-adjusted mortality rate

Table 32–1. Incidence, Mortality, and Survival United States, SEER Program

Incidence† Mortality‡ Survival**

Total

Males

Females

Whites

Blacks

Other*

4.0 1.4 65%

7.0 2.7 67%

1.5 0.5 59%

3.9 1.3 67%

6.4 2.9 53%

2.0 0.6 69%

*Rates per 100,000, age adjusted to the 2000 US population, SEER 1995–2000, 9 SEER registries; †Rates per 100,000, cause-specific morality, age adjusted to the 2000 US population, SEER 1995–2000; ‡Five-year, age-adjusted relative survival %, SEER 1992–2000; **Other racial/ethnic groups include: American Indian or Alaskan Native, Asian, or Pacific Islander.

was 1.7 per 100,000 in 1969, 1.6 in 1980, 1.6 in 1990, and 1.4 in 1999 (APC = -0.6). Among men, the rate was 3.5 in 1969 and 2.6 in 1999 (APC = -0.9). For females, the rate increased from 0.4 in 1969 to 0.5 in 1999 (APC = 1.1). White males had a larger decrease in mortality (APC = -1.2) than black males (APC = 0.8).

Survival SEER data for the years 1992–2000 were used to estimate the ageadjusted 5-year relative survival rates (Table 32–1; SEER, 2003). For all stages, the survival rate was 65%. Survival was poorer for females (59%) than males (67%). As has been shown in past studies, the 5year relative survival was lower among blacks (53%) compared to whites (67%). The median survival time was 6.5 years among whites and 4.1 years among blacks. The 5-year relative survival rates by stage were: 82% localized, 48% regional, 21% distant, and 59% unstaged. Survival rates were highest for glottic tumors (81%), and lower for supraglottic (48%), subglottic (51%), and laryngeal cartilage tumors (42%).

International Patterns Incidence data available from population-based registries on five continents vary considerably (Parkin et al., 2002). Age-adjusted incidence rates were determined for the years 1993–1997. Men had higher incidence rates in Uruguay (12.1 per 100,000), parts of India (Delhi = 9.4), Eastern Europe (Croatia = 12.7; Yugoslavia = 10.6; Poland, lower Silesia = 13.3), Southern Europe (France, Calvados = 11.7; France, Somme = 13.0; Italy, North East = 13.1), and Spain (Asturias = 15.3; Granada = 12.8; Murcia = 14.9; Zaragoza = 18.0). In comparison, incidence rates for men were lower than 5 per 100,000 in most registries in Africa, China, Japan, Northern Europe, United Kingdom, and Canada. Among women, the age-adjusted incidence rates are usually less than 1 per 100,000 and vary little. Rates among black women in the United States are higher than those among whites in the United States, and compared with most other registries. Worldwide, the highest age-adjusted mortality rates estimated for the year 2000 among men include: Eastern Europe (7.9 per 100,000), Southern Europe (5.1), Western Asia (4.9), South Central Asia (4.5), and South America (4.0). Lower death rates for men were seen for Eastern and Middle Africa (2.1 and 1.2, respectively) and Eastern Asia (1.1) (Ferlay et al., 2001). An analysis of cancer mortality in Europe from 1955–1994 showed persistent declines in mortality from laryngeal cancer in France, Italy, and Finland (Levi et al., 1999). Significant increases were noted for Germany and other central and Eastern European countries. The highest mortality rates for 1990–1994 were found in Hungary (9.4 per 100,000 males) and the lowest rates in Iceland (0.5 per 100,000 males). The authors attribute the increases in mortality to an epidemic of tobacco and alcohol use in areas of Central and Eastern Europe (Levi et al., 1999).

ENVIRONMENTAL FACTORS Tobacco A consistent association between cigarette smoking and cancer of the larynx has been demonstrated in several studies, beginning in the 1950s (Wynder et al., 1956; Rothman et al., 1980; IARC, 1986; Boyle et al., 1990; Austin and Reynolds, 1992). Case-control and cohort studies consistently have found a dose-response gradient of increasing risk with increasing use. Despite the generally strong association with tobacco use, the magnitude of effect varies. Among the largest studies conducted (845 total laryngeal cancer cases from France, Italy, Spain, and Switzerland), Tuyns et al. (1988) found adjusted odds ratios of 24.0 (CI = 11.8–48.7) for supraglottic cancer and 10.2 (CI = 5.4–19.3) for glottic cancer when average consumption was 26 or more cigarettes per day. Cancer at other laryngeal subsites had an OR of 9.4 (CI = 3.2–28.0). A recent hospital-based case-control study that included 527 cases and 1297 controls from Northern Italy and Switzer-

Cancer of the Larynx land reported odds ratios of 19.8 (CI = 11.0–32.9) for current smoking (versus never smokers); 42.9 (CI = 22.8–80.9) for 25 or more cigarettes per day; and 37.2 (CI = 20.2–68.5) for 40 or more years of smoking (Talamini et al., 2002). A study from Western New York State (250 cases) reported an odds ratio of 12.6 (CI = 5.0–31.5) for 46 or more pack-years of smoking (Freudenheim et al., 1992). A Polish study (249 male cases) reported an odds ratio of 59.7 (CI = 13.0–274.0) for use of more than 30 cigarettes per day (Zatonski et al., 1991). A German study (164 male cases) found an odds ratio of 9.1 (CI = 4.5–18.7) for more than 50 “tobacco-years” (tobacco-year = daily consumption of 20 cigarettes, 4 cigars, or 5 pipes for 1 year) of consumption (Maier et al., 1992). An odds ratio of 19.2 (CI = 5.0–73.4) for average tobacco use of more than 40 cigarettes per day was found in a Texas case-control study (151 male cases; Falk et al., 1989). A large Turkish study (832 male cases) reported a lower risk (OR = 6.6; CI = 4.2–10.3) for the highest consumption level (more than 20 cigarettes per day; Dosemeci et al., 1997). An odds ratio of 25 (CI = 9.9–63.2) for consumption of 20 or more cigarettes per day was found in a Shanghai, China, case-control study (201 cases; Zheng et al., 1992). Few studies have examined the association between smoking and laryngeal cancer among women. The largest study of women, including 68 cases and 340 controls from northern Italy and Switzerland, reported a large, although imprecise, odds ratio for tobacco use (OR = 46.0; CI = 16.0–131.0 for current smokers compared to never smokers; Gallus et al., 2003). The variations in effect estimates across studies and study areas may reflect differences in sample size, control groups, tobacco variable definitions, influence of other risk factors, as well as tobacco and cigarette type. Talamini et al. (2002) reported odds ratios of 27.1 for high-tar (>19 mg) cigarettes and 19.3 for low- or medium-tar (<19 mg) cigarette use. Individual studies from Europe (Tuyns et al., 1988) and South America (De Stefani et al., 1987) have reported higher odds ratios for “black” (air-cured) tobacco than for “blond” (flue-cured) tobacco. The risk was approximately twice as high with black tobacco use (Sancho-Garnier and Theobald, 1993). Experimental studies have reported greater carcinogenicity with black tobacco owing to its higher content of aromatic amines and tobacco-specific nitrosamines (Hecht and Hoffmann, 1988). Several studies reported a higher effect estimate for non-filter cigarettes compared with filter cigarettes. For example, a large European study found that filter-cigarette smokers had one-half the risk of laryngeal cancer (Tuyns et al., 1988) compared with non-filter-cigarette smokers. A pattern of decreasing risk with increasing time since smoking cessation has been reported. An Italian and Swiss study noted that risk steadily decreased 3 years after smoking stopped (Altieri et al., 2002). After 10 or more years since cessation, the risk was reduced by 70%. Among moderate (<20 cigarettes per day) and heavy smokers (≥20 cigarettes per day), odds ratios of 0.2 and 0.3 for larynx cancer, respectively, were found for ≥20 years since quitting smoking (Talamini et al., 2002). This group also concluded that duration of smoking was more important than the age when smoking started. There have not been many well-designed studies that have examined other forms of tobacco use and their association with laryngeal cancer. Studies of the Indian bidi cigarette (Jusawalla and Deshpande, 1971) and a study of hand-rolled cigarette use from Uruguay (De Stefani et al., 1992) have reported higher risks than among smokers of commercial cigarettes. Some earlier, smaller studies have noted a possible association between pipe or cigar use, although a more recent case-control study from the United States reported no association (Freudenheim et al., 1992). Smokeless tobacco use in the form of betel quid chewing has been associated in one study (Jusawalla and Deshpande, 1971). A small number of studies of dry and moist snuff and chewing tobacco have not reported a consistently elevated risk (Rodu and Cole, 2002). Several studies have examined the effects of tobacco on cancer risk in laryngeal cancer subsites. Most studies have reported larger relative risks for supraglottic than glottic cancer (Talamini et al., 2002). The large (845 cases) study from Spain, France, Italy, and Switzerland reported odds ratios of 24 (CI = 11.8–48.7), 10.2 (CI = 5.4–19.3), and

629

9.4 (CI = 3.2–28.0) for supraglottis, glottis, and epilarynx, respectively, for 26 or more cigarettes per day (Tuyns et al., 1988). A Turkish case-control study found for the supraglottis an odds ratio of 7.9 (CI = 4.6–13.7) for 21 or more pack-years of smoking compared with 3.6 (CI = 1.8–7.3) for the glottis (Dosemeci et al., 1997). A case-control study from Northern Italy and Switzerland reported an odds ratio of 54.9 (CI = 16.9–178.4) for supraglottic cancer among current smokers compared with 7.4 (CI = 4.9–11.3) for glottic cancer (Talamini et al., 2002). Among heavy smokers (≥25 cigarettes per day), the odds ratios were 112.4 (CI = 30.4–416.0) for supraglottic and 16.7 (CI = 9.1–30.5) for glottic cancer. The location of the supraglottic tissues appears more readily exposed to tobacco than other subsites.

Alcohol Investigations of cohorts of alcoholics, occupational alcohol producers and distributors, and of incidence and mortality in religious groups who consume less alcohol have indicated an overall association between alcohol use and laryngeal cancer (IARC, 1988). Many analytic studies have attempted to estimate the independent, unconfounded effect of alcohol by adjustment for smoking. Because cigarette and alcohol use are correlated, the potential independent role of alcohol consumption in the etiology of laryngeal cancer has not been fully elucidated. In general, most studies have found an association with heavy use of alcohol and some found a dose-response gradient with the amount consumed (Olsen et al., 1985; Destefani et al., 1987; Tuyns et al., 1988; Falk et al., 1989; Franceshi et al., 1990; Frauenheim et al., 1992; Maier et al., 1992; Hedberg et al., 1994; Dosemici et al., 1997; Schlecht et al., 1999; Talamini, et al., 2002). The effect estimates for the largest alcohol-consumption categories relative to the lowest have generally been more modest than for heavy tobacco use—usually between about two and five. An analysis of female cases and controls from Italy and Switzerland found an odds ratio of 4.3 (CI = 0.8–24.1) for more than five drinks per day (Gallus et al., 2003). A recent meta-analysis of 20 case-control studies of laryngeal cancer (3759 total cases) published between 1996 and 2000 found an overall dose-response gradient with a pooled relative risk (RR) of 3.9 for 100 g day-1 of alcohol consumption (Bagnardi et al., 2001). When restricted to studies that reported adjustment for tobacco, a pooled RR of 2.8 (CI = 2.4–3.3) for 100 g day-1 was found. Despite the doseresponse pattern with amount of drinking, previous studies have not found a clear trend of increasing duration of drinking with increasing risk (Franceschi et al., 1990; Talamini et al., 2002). Analysis of non-smokers or light smokers has generally shown an elevated risk for alcohol consumption (Burch et al., 1981; Elwood et al., 1984; Tuyns, 1988). For example, Bosetti et al. (2002a) combined data from two European case-control studies and reported an odds ratio of 2.5 (CI = 0.9–6.2) for heavy drinking (≥8 drinks per day compared with <8 drinks per day) among non-smokers. Although the original studies were relatively large, the number of non-smokers was small (40 non-smoking cases), which limited the precision of the effect estimates. Few studies have evaluated the effect of cessation of alcoholic beverage consumption. The results from these studies suggest that unlike smoking, cessation of drinking has an effect only after a long period of cessation. A European study found no consistent pattern of risk reduction association with cessation spanning less than 20 years (Altieri et al., 2002). Even in this relatively large study (527 cases), the effect estimate for time since drinking cessation of 20 or more years was unstable (OR = 0.5; CI = 0.2–1.9). Several studies have examined whether type of alcoholic beverage affects risk. Some studies have reported stronger associations for whiskey (Wynder et al., 1956), beer (Olsen et al., 1985), and wine (Franschei et al., 1990), although others found no differences (Talamini et al., 2002). To evaluate whether the non-alcohol constituents of distilled alcoholic beverages conferred any particular risk, Rothman et al. (1989) classified liquor type by “dark” (whiskeys, rum, cognac) and “light” (gin, vodka, light rum). Heavy consumption of dark liquor had a strong association with hypopharyngeal cancer

630

PART IV: CANCER BY TISSUE OF ORIGIN

(OR = 4.4), compared with heavy intake of light liquor (OR = 1.3). No similar pattern of association with dark liquor was found for laryngeal cancer. A case-control study reported elevated odds ratios for heavy consumption of cachaca, a strong Brazilian alcoholic beverage that is produced from sugar cane (Schlecht et al., 2001). Several studies have reported higher risks for supraglottic cancer compared with glottic and other laryngeal tumor subsites for all alcoholic beverage types (Guenel et al., 1988; Schlecht et al., 2001; Talamini et al., 2002), although others have not reported any difference (Olsen et al., 1985; Tuyns et al., 1988; Lopez-Abente et al., 1992; Dosemici et al., 1997). Several investigators have performed analyses to quantify the joint effects or interaction between tobacco and alcohol and the risk of laryngeal cancer. An important issue in this assessment is the scale or benchmarks used to describe and evaluate the interaction. Although opinions still differ, some authors have argued for appropriateness of the additive benchmark in causal interpretation (Rothman and Greenland, 1998). The IARC multicenter European study, one of the largest (Tuyns et al., 1988) including 845 male cases, found an odds ratio of 43.2 for the combined effect of cigarette smoking (26 or more cigarettes a day) and 121 or more grams of alcohol per day (compared with individuals consuming 0–7 cigarettes and 0–40 grams of alcohol daily). This is the expected odds ratio using a multiplicative null benchmark (ORtobacco+ & alcohol+ = ORtobacco+ & alcohol- ¥ ORtobacco- & alcohol+). Most other studies have also found that the joint effect of tobacco and alcohol approximated the multiplicative null or was sub-multiplicative (Olsen et al., 1985; Falk et al., 1989; Baron et al., 1993; Dosemici et al., 1997; Schlecht et al., 1999; Talamini et al., 2002). These findings indicate a departure from the expected additive null. However, few studies have directly quantified departures from the additive null (ORtobacco+ & alcohol+ - 1) = [(ORtobacco+ & alcohol- - 1) + (ORtobacco- & alcohol+ 1)]. Flanders and Rothman (1982) analyzed data from two studies: the Third National Cancer Survey (TNCS) and the earlier study by Wynder et al. (1976). Both included a relatively small number of cases (87 in the TNCS and 258 in Wynder’s study). Using a measure of interaction on the additive scale, they estimated a moderate amount of synergy between alcohol and tobacco (1.5, or 50% more risk than from the simple additivity of tobacco and alcohol). A Danish study of 326 cases also reported departures from addivity using the same synergy index as Rothman (Olsen et al., 1985). In this study, the magnitude of the synergy ranged from 2.2–6.1, which was greater than in the Flanders and Rothman analysis. Other studies have attempted to describe the fit of their data relative to the additive and multiplicative scales (Guenel et al., 1988; Falk et al., 1989) with results suggesting compatibility with the multiplicative form or an intermediate result. Interpreting these results would be easier if departures from a chosen form of interaction were quantified consistently. In summary, tobacco is the best-described and most important risk factor for laryngeal cancer. The reported effect estimates have been generally strong, consistent, and show a pronounced dose-response relationship. The etiologic role of alcohol is less clear, but evidence suggests both a weaker independent effect and a joint effect with tobacco. Supraglottic tumors have shown the strongest association with tobacco and alcohol compared with other subsites. The quantification and specific form of interaction has not been examined comprehensively and many of the studies are too small to evaluate interactions precisely, especially among subsites.

Occupation Many studies have examined associations between occupational asbestos exposure and increased risk of laryngeal cancer. Occupations and industries with potential asbestos exposure include cement workers, asbestos miners, shipyard workers, building materials workers, friction materials workers, Chrysotile asbestos workers, railroad workers, and textile workers. Individual studies have found both positive and negative results (e.g., Parnes et al., 1990; Liddell, 1990; Zheng et al., 1992; Muscat et al., 1992; Marchand et al., 2000; Berrino et al., 2003). Cohort studies of more highly exposed workers have had mixed findings, which are difficult to interpret because of potentially

uncontrolled confounding by smoking and alcohol consumption. Some case-control studies of asbestos exposure that adjusted for these potential confounders have reported odds ratios around two, while others have reported no association. Some studies (Smith 1990, De Stefani et al., 1998a) reported an interaction between asbestos and smoking, although others did not support these results (Muscat et al., 1992; Marchand et al., 2000). Several reviews of the epidemiologic findings in relation to asbestos and laryngeal cancer have concluded that the increased risk, if one exists, is very weak (Chan and Gee, 1988; Edelman, 1989). Two reviews dissented, offering evidence for a positive relationship (Doll and Peto, 1985; Smith et al., 1990). Doll and Peto estimated a summary relative risk of 1.4 from eight cohorts and suggested that smoking could not account fully for the excess risk. A meta-analysis of 27 asbestos-exposed occupational cohorts found a meta-standardized mortality ratio (SMR) of 133 (CI = 114–155) for laryngeal cancer (Goodman et al., 1999). These results, plus the positive findings from five case-control studies, led them to conclude that asbestos is one cause of laryngeal cancer. It is difficult to determine specific critical exposure levels or thresholds because these studies vary widely in the occupational settings they examined and in the methods used to measure exposure. More recently introduced synthetic fibers, man-made mineral vitreous fibers (MMVF) including mineral wool, refractory ceramic fibers, glass filaments, and microfibers, have not been as well studied as asbestos, but there has been concern over their potential carcinogenicity. The few studies of MMVF and laryngeal cancer have mixed results. A French study of workers in a glass-wool plant reported an increased incidence of laryngeal cancer and an Italian study noted an elevation in mortality. Other cohort studies have reported no association (Maier et al., 1991). A recent French case-control study found a weak elevation in risk (OR = 1.3; CI = 0.9–1.9) for laryngeal cancer with mineral wool exposure, after adjusting for smoking, alcohol, and asbestos exposure (Marchand et al., 2000). Although generally not as well studied as asbestos, there are several specific occupational exposures, job titles, and industries reported to be associated with an increased risk of laryngeal cancer (Maier et al., 1991). Workers involved in the production of mustard gas in Japan and England had a much higher-than-expected mortality from laryngeal cancer (Wada et al., 1968; Manning et al., 1981). Large relative risks have been reported in cohort studies of workers exposed to sulfuric acid mist in refineries and chemical plants (Maier et al., 1991). Elevated laryngeal cancer mortality from employment at plants that manufacture alcohol has been reported (Ahlborg et al., 1981; Soskolone et al., 1984; Steenland et al., 1988). It has been suggested that the association is related to exposure to sulfuric acid esters such as diethyl sulfate (Lynch et al., 1979). Using a job-exposure matrix, a Southern European case-control study reported significant exposure-response trends with solvent exposure based upon probability (“probable” exposure OR = 2.5; CI = 1.5–4.2) and duration of exposure (20 or more years OR = 2.0; CI = 1.2–3.4) (Berrino et al., 2003). An association with formaldehyde was also reported (Berrino et al., 2003). Some analyses of employment as a woodworker and exposure-based analyses of wood dust have found elevated risks of laryngeal cancer (Maier et al., 1991). The associations were stronger among cases with cancer of the endolarynx (glottis and supraglottis) than cancer of the epilarynx and hypopharynx (Bofetta et al., 2003). This study also reported an association with wood dust exposure, but only among men born before 1925, which suggests a possible change in processes and exposures over time (Berrino et al., 2003). A recent IARC review of wood dust concluded that it was not a laryngeal carcinogen (IARC, 1995). Cohort studies have indicated that men who refine nickel have an elevated incidence and mortality from laryngeal cancer (Pederson et al., 1973; Mangus et al., 1982; Maier et al., 1991). The cohort studies indicated that specific work in the roasting and melting operation may have been responsible for the association. A cohort study from New Caledonia in the South Pacific found no association with nickel mining or refining, but noted increased risks for mine and other dusts (Goldberg et al., 1994). Other case-control studies have not reported a spe-

Cancer of the Larynx cific association with nickel exposure (Wortley et al., 1992; Zheng et al., 1992; Gustavsson et al., 1997). Occupations that involve other metal-related operations and exposures, including sheet-metal workers, metal platers, plumbers, welders, electricians, metal grinders, pipe-fitters, and metal fabrication, have been reported to be associated with laryngeal cancer in various case-control studies (Maier et al., 1991; Wortley et al., 1992; Goldberg et al., 1997; Gustavsson et al., 1998; Bofetta et al., 2003). Although many of these studies adjusted for tobacco and alcohol use, most of these occupations involve exposure to a variety of potentially carcinogenic compounds including metal fumes, asbestos, radiation, and polycyclic aromatic hydrocarbons. Thus, identifying the specific occupational exposure(s) responsible for the elevated risk associated with these job titles is a notable challenge. Laryngeal cancer is associated with an array of other specific or classes of occupations, industries, and exposures. Multiple studies have reported associations with truck drivers, railway workers, leatherworkers, potters, butchers, barbers, textile workers, paint and print workers, mustard gas, plastic and rubber product manufacturing, automobile mechanics, diesel fume exposure, agriculture, machining fluids, and construction work (Maier et al., 1991; Muscat et al., 1992; Wortley et al., 1992; Boffetta et al., 2003). Important methodologic issues must be considered when assessing epidemiologic evidence for associations between occupational exposures and laryngeal cancer. Most of the case-control studies have attempted to adjust carefully for the potential confounding effects of tobacco and alcohol use. Residual confounding is possible but is unlikely to explain the associations reported. A critical issue, especially in case-control studies, is the misclassification of exposure based upon the reporting source (e.g., self-reported asbestos exposure) and use of job titles and industry and job-exposure matrices to infer exposure.

Diet Several case-control studies from Europe, North and South America, Asia, and elsewhere have reported associations with dietary factors. Relatively consistent inverse associations (increased intake, decreasing risk) have been reported for higher levels of fruit and vegetable consumption, after adjustment for tobacco and alcohol consumption (Riboli et al., 1996). A case-control study conducted in Shanghai, China, reported decreased risks (OR = 0.3–0.4) for high levels of citrus fruit and dark green vegetables (Zheng et al., 1992). An Italian study (La Vecchia et al., 1990) found decreased odds ratios for high intake of green vegetables (OR = 0.4) and fresh fruits (OR = 0.3). A recent study from Uruguay also found decreased risks for high intake of fruits (OR = 0.4) and raw vegetables (OR = 0.3) (De Stefani et al., 2000a). A large European study (727 cases) conducted in Spain, Switzerland, France, and Italy found elevated odds ratios for cancer of the endolarynx (glottis, supraglottis, and subglottis) among low consumers of fruits (OR = 1.4), citrus fruit (OR = 1.6), and vegetables (OR = 1.7; Estve et al., 1996). Most studies have also found a linear decrease in risk from lowest to highest consumption levels (Riboli et al., 1996). Additional research has been conducted to isolate the specific micronutrients that might be responsible for the food group associations (Riboli et al., 1996). Associations have been reported for vitamins A and C. Relative risk estimates of about three were reported for the lowest intake level of vitamin A; later studies reported associations specifically with b-carotene (Riboli et al., 1996). Odds ratios of 1.6 and 2.8 for the two lowest levels of vitamin C and a dose-response gradient were reported in a study from southwestern Europe (Esteve et al., 1996). A case-control study from Uruguay reported decreased risks related to greater consumption of tomatoes, tomato foods, and lycopene (De Stefani et al., 2000b). Although less consistent, some studies found elevated odds ratios for consumption of meat, eggs, and butter. Higher levels of retinol, total calories, protein, and fat in the diet were reported to increase risk (Riboli et al., 1996; Oreggia et al., 2001). Some studies noted decreased risks for fish and vegetable oil consumption (Riboli et al.,

631

1996). Moderately high polyunsaturated/saturated fatty acids ratios were found to have a reduced risk (Esteve et al., 1996). One study reported associations with decreased risks for self-reported iron and zinc but analysis of toenail samples did not support the findings (Rogers et al., 1993). Recent analyses of dietary data from a large study conducted in northern Italy and Switzerland (527 cases) examined food groups, nutrients, and cooking practices in more detail (Bosetti et al., 2002b). An analysis of food groups showed elevated risks (OR = 1.6–3.1) for high consumption of eggs, red meat, processed meats, fish, and sugars. Decreased risks (OR = 0.2–0.7) were observed for intake of high levels of raw vegetables, cooked vegetables, citrus fruit, and other fruits. The report also noted an elevated risk for mixed seed oils (OR = 2.2), but a decreased risk (OR = 0.4) for use of olive oil. Higher dietary fiber consumption was associated with a decreased risk of laryngeal cancer (Pelucchi et al., 2003). The odds ratios were 0.2 for fiber from vegetables, 0.5 for fiber from fruits, and 1.1 for fiber from grains. The same study also reported elevated odds ratios for high consumption of fried meat (1.6), fish (3.1), eggs (1.9), and potatoes (1.9) (Bosetti et al., 2002). The authors suggested that the findings might be explained, in part, by the genotoxic effects of heterocyclic amines and fat used in frying. An earlier study from Uruguay reported an association between total heterocyclic amine intake and cancer of the upper aerodigestive tract (De Stefani et al., 1998b). Few studies have examined whether the risks related to dietary factors vary across anatomic subsites of the larynx. European studies have found no important differences (Esteve et al., 1996; Bosetti et al., 2002). The potential interaction between tobacco and alcohol use and dietary factors has been evaluated in a small number of studies. Decreased risks associated with fruit and vegetable intake were similar across levels of tobacco and alcohol consumption in the European studies (Esteve et al., 1996; Bosetti et al., 2002). The Chinese study, which includes a high proportion of non-alcohol drinkers, also found a strong inverse association with higher intake of citrus fruit and vegetables (Zheng et al., 1992). Diet appears to be an independent risk factor for laryngeal cancer. High fruit and vegetable consumption is consistently associated with a decreased risk. More research is needed to understand the specific nutritional factors underlying this association. Other food groups that may increase risk also deserve additional study.

Viruses There has been considerable attention paid to the role of human papillomavirus (HPV) in the etiology of laryngeal cancer. The malignant transformation of laryngeal papillomas, possibly acquired during parturition, has long been the basis of interest in the etiologic role of HPV (Lindeberg and Elbrond, 1990; Lie et al., 1994; Go et al., 2003). Biologic plausibility for the potentially causal role of HPV comes from molecular evidence that the HPV E6 and E7 oncoproteins disrupt cell cycle control by degrading p53 and Rb (Dyson et al., 1989; Scheffner et al., 1990). Prevalence studies suggest HPV may be important in head and neck cancer because approximately 35% of these tumors have PCR-detectable HPV, especially the “high-risk” HPV16 and HPV18 types (McKaig et al., 1998). Recent molecular and epidemiologic studies provided additional evidence that HPV may have an etiologic role in head and neck cancer, especially oropharyngeal and tonsillar tumors (Gillson and Shah, 2001). Molecular studies of frozen tumor samples that used multiple HPV-detection methods (PCR, Southern blot, and in situ hybridization) and quantified E6 expression concluded that HPV, especially HPV type 16, is associated with a subset of oropharyngeal tumors, especially tonsillar cancer (Gillson et al., 2000, van Houten et al., 2001). Research that measured viral copy number and E6 mRNA expression also suggested that standard PCR methods to detect HPV have largely overestimated the prevalence of HPV-positivity (Van Houten et al., 2001; Franchesi et al., 2000; Ha et al., 2002) with respect to etiologically relevant HPV-positive samples. Ha et al. (2002), using quantitative PCR, concluded that HPV was unlikely to play an important role in oral cancer. They did not have

632

PART IV: CANCER BY TISSUE OF ORIGIN

enough oropharyngeal tumors to draw any conclusions of special interest about these sites. Interestingly, a cohort study with stored prediagnostic serum samples found a twofold increased risk of head and neck cancer with HPV seropositivity, after adjustment for cotinine levels (Mork et al., 2001). The nature of the association between HPV and laryngeal cancer is unresolved. Using PCR-based methods to detect HPV, a review of previous studies reported that the average prevalence of HPV in laryngeal carcinomas was 35%, with a range of 3%–85% (Lindeberg and Krogdahl, 1999). There were 19 studies with 4–100 case tumors included. Another review reported that using PCR methods, HPV16 was detected in 46% and HPV18 in 16% of laryngeal tumors (McKaig et al., 1998). The variability in prevalence in detected HPV has been attributed, in part, to the small number of tumors tested, differences in PCR primers used, and sample contamination during collection, sectioning, processing, and other routes throughout PCR analysis (McKaig et al., 1998; Lindeberg and Krogdahl, 1999). One review concluded that HPV does not have a major etiologic role in laryngeal cancer in persons without a history of laryngeal papillomatosis (Lindeberg and Krogdahl, 1999). On the other hand, other recent studies suggested that HPV may play an etiologic role in a subset of laryngeal tumors. Venuti et al. examined the prevalence, physical state (episomal or integrated), and expression of HPV oncoproteins in 25 laryngeal tumors (Venutti et al., 2000). The study found 13 (52%) were HPV-positive tumors, seven of which were HPV-16 positive (5 had HPV-6 and 1 had HPV-45). Viral integration was reported for 43% of HPV-16–positive cases and viral transcripts were detected in three of four HPV-16–positive cases available for analysis. Another study found a correlation between HPV-positivity and epidermal growth factor receptor (EGFR) expression (Almadori et al., 2001). EGFR has been reported to contribute to development and progression of laryngeal cancer. Finally, the Nordic cohort study, which used stored pre-diagnostic serum samples, found an association between seropositivity and laryngeal cancer (RR = 2.4; CI = 1.0–5.6; Monk et al., 2001). The overall prevalence of HPV, especially the high-risk types, in laryngeal tumors, coupled with the recent molecular study of integration and expression, suggest that HPV plays a role in the initiation or progression in at least a subset of laryngeal cancers. Nonetheless, additional molecular and epidemiologic work clearly is required to establish causality (McKaig et al., 1998; Franceschi et al., 2000).

Other Factors Gastroesophageal reflux (GER), the upward movement of gastric contents into the esophagus, is a relatively common condition in the Western world. Gastroesophageal reflux disease (GERD) is a more severe and chronic reflux condition that has been associated with esophageal cancer (Shaheen and Ransohoff, 2002). It is thought that GER can lead to chronic irritation and inflammation resulting in malignant transformation. GERD has been suggested as a potential risk factor for laryngeal cancer since 1960 (Assimakopoulos and Patrikakos, 2002). Small clinical studies, some including nonsmoking persons, appeared to support an association (Assimakopoulos and Patrikakos, 2002). There have been few epidemiologic studies of GERD using a comparison group. A recent study of the hospital records of 8228 US military veterans with laryngeal cancer and 7648 hospital controls found that after adjustment for smoking, age, race, and alcohol use, GERD was associated with an odds ratio of 2.4 (CI = 2.2–2.7) (El-Serag et al., 2001). A case-control study of 173 cases of head and neck cancer (48 larynx cases) found an association with a history of marijuana use (ever versus never use, adjusted for tobacco, alcohol use, and age, OR = 2.6; CI = 1.1–6.6; <55 years of age OR = 3.1; CI = 1.0–9.7), including a dose-response gradient and possible interaction with tobacco use and mutagen sensitivity (Zhang et al., 1999). One small (68 cases) case-control analysis of female cases and controls did not find any associations with reproductive or hormonal factors (Gallus et al., 2003).

Genetic Factors Familial cases of laryngeal cancer have been noted in case series reports (Gencik et al., 1986). Characteristics of familial cases included early age at diagnosis, concordance of tumor site and histology within families, and age at diagnosis among siblings. Recent case-control studies evaluated the etiologic importance of a family history of cancer. A Brazilian study reported an association between larynx cancer and a family history of any cancer among first-degree relatives (OR = 2.3; CI = 1.5–3.5); the strongest association was seen among siblings (Foulkes et al., 1995). Among all cases of head and neck cancer, a first-degree relative with a head and neck cancer had the strongest association. Other studies of oral and pharyngeal cancer also found elevated risks for a family history of cancer among first-degree relatives (Goldstein et al., 1994; Cooper et al., 1995). A case-control study of head and neck cancer that used a mutagen-sensitivity assay reported a possible interaction between family history of cancer among first-degree relatives and mutagen sensitivity using an in vitro bleomycin assay (Yu et al., 1999). Much effort has been devoted recently to examining genetic susceptibility factors for cancer, especially common variants of less penetrant genes (so-called minor genes) (Caporaso and Goldstein, 1995). These variants or polymorphisms (also called single-nucleotide polymorphisms) typically have a population prevalence of greater than 1%, sometimes as much as 50%, and are thought to be involved in geneenvironment interactions. There are now at least 20 reports from studies that evaluated the association between polymorphisms of genes encoding carcinogen metabolizing and DNA-repair proteins and the risk of head and neck cancer (Bartsch et al., 2000; Lazarus and Park, 2000; Geisler and Olshan, 2001; Goode et al., 2002). Epidemiologic evidence has been inconsistent. Moreover, few studies have examined laryngeal cancer specifically. Two phase I enzymes, CYP1A1 and CYP2E1, have been studied. CYP1A1 helps activate important classes of tobacco carcinogens such as polycyclic aromatic hydrocarbons (PAHs) and aromatic amines. A polymorphism at codon 462 (A > G, Ile > Val) results in higher microsomal enzyme activity compared with the wild-type allele. Some studies have reported a positive association with the CYP1A1 rapid allele and head and neck cancer (Bartsch et al., 2000). Only one study contained a sufficient number of larynx cancer cases (n = 272) and no increased risk was found (Matthias et al., 1998). CYP2E1 is involved in the metabolism of tobacco-specific nitrosamines and ethanol. Several CYP2E1 polymorphisms have been investigated, but no association with laryngeal cancer was reported in five studies (Bartsch et al., 2000). The study with the most laryngeal cancer cases found an odds ratio of 1.3 (CI = 0.7–2.6) for the intron6 mutation. The glutathione S-transferase detoxification enzyme (GSTM1, GSTM3, GSTT1, GSTP1) polymorphisms were investigated in a few studies. The results were inconsistent; four studies reported a positive association (ORs = 1.6–3.9) for the GSTM1-null (deletion) genotype and two of three studies found a positive association with the GSTT1-null genotype (Geisler and Olshan, 2001). The glutathione S-transferase enzyme pi is involved in the detoxification of tobacco smoke PAHs. Two studies of the GSTP1 Ile104Val polymorphism reported no increase in risk (Matthias et al., 1998; Jourenkova-Mironova et al., 1999). The GSTM3 variant was associated with an odds ratio of 2.0 in one study (JourenkovaMironova et al., 1999). Results from single studies have suggested associations for polymorphisms of genes also involved in the metabolism of tobacco smoke and other carcinogens. Positive associations were found for the EPHX1 high-activity genotypes and NAT2 slow or intermediate genotypes (Morita et al., 1999; JourenkovaMironova et al., 2000) and an inverse association with MPO -463G/A transition polymorphism (Cascorbi et al., 2000). UDP-glucuronosyltransferases (UGTs) metabolize a variety of endogenous compounds including the conjugation and detoxification of tobacco carcinogens (Grove et al., 2000; Ren et al., 2000). Zheng et al. (2001) reported an association between the UGT1A7 enzyme polymorphism resulting in lower-activity (57 cases; OR = 3.7; CI = 0.9–14.0) laryngeal cancer.

Cancer of the Larynx Four studies have examined the relationship between DNA-repairgene polymorphisms and head and neck cancer. Two studies had inconsistent findings for XRCC1 (Goode et al., 2002), one study found no association with an XPD variant, and one study found an odds ratio of 1.9 for an XPC variant (Goode et al., 2002). A case-control study including 125 orolaryngeal cases (44 laryngeal cases) reported an elevated odds ratio (OR = 4,1; CI = 1.3–13.0) for individuals homozygous for the Ser326Cys variant of the hOGG1 base excision repair enzyme (Elahi et al., 2002). None of the studies contained a sufficient number of laryngeal cancer cases to precisely estimate an effect measure. Several studies have examined the role of polymorphisms of alcohol metabolism enzymes in the etiology of head and neck cancer (Brennan et al., 2004). Alcohol dehydrogenase (ADH) is a phase I enzyme that helps metabolize ethanol to acetaldehyde. Aldehyde dehydrogenase (ALDH) catalyzes the conversion of acetaldehyde to acetic acid. Ethanol is metabolized in the liver by ADH and ALDH, but also in the oral cavity and other sites of the upper aerodigestive tract. ADH and ALDH include gene loci with known variants (Smith et al., 1986). ADH1C*1 and ADH1B*2 alleles are in linkage disequilibrium in some European populations and represent “fast” alleles with ethanol oxidation increased by about 2.5-fold compared with ADH1C*2 (Edenberg and Bosron, 1997; Osier et al., 1999; Borras et al., 2000). The ALDH2 gene contains an inactivating ALDH2*2 allele (Glu487Lys polymorphism). ALDH*2 homozygotes cannot oxidize acetaldehyde (Yoshida et al., 1984). Persons with homozygous and heterozygous ALDH2*2 genotypes experience a toxic build-up of acetaldehyde with flushing, increased heart rate, and nausea. The published studies of the ADH1C allele, primarily including oropharyngeal cancer, are inconsistent with one study that reported a strong association among heavy drinkers and others that reported either a weaker association or no association (Brennan et al., 2004). These studies included too few laryngeal cancer cases to perform a separate analysis. The ADH1B and ALDH2 at-risk genotypes have been examined in relation to laryngeal cancer in four studies. ADH1B-1 allele was associated with a sixfold increased risk among Japanese alcoholics (Yokoyama et al., 2001). The same group also reported an odds ratio of 29 for alcoholics with the ALDH-1 allele. Nomura et al. (2000) also reported an elevated odds ratio (OR = 2.9) for the ALDH1 allele. Another study found no association with ALDH (Katoh et al., 1999). The current data have not provided consistent evidence for a meaningful association between polymorphisms of metabolizing enzymes and DNA-repair enzymes and the risk of laryngeal cancer. To date, there have been relatively few studies. Previous studies were limited by relatively small sample sizes, especially for specific analyses of laryngeal cancer, and by including only one or two enzyme polymorphisms. Few studies have examined interaction with smoking or alcohol directly.

PATHOGENESIS A general model for the pathogenesis of laryngeal cancer can be considered based upon studies from epidemiology, pathology, and molecular biology. The strongest etiologic factor, tobacco, defines a pathway involving chronic exposure over many years resulting in cumulative DNA damage that, if not repaired, increases the probability of alterations of key tumor-suppressor and oncogenes. Prior to molecular somatic events, a series of activation and detoxification processes occur whereby many tobacco (and alcohol) carcinogens undergo Phase I and Phase II metabolism. There is great interest in investigating polymorphisms in the genes that encode these metabolism enzymes as susceptibility factors in head and neck cancer. DNA damage and DNA adduct formation are the next recognized somatic molecular events. DNA damage repair is a complex process and not completely understood. As with the evaluation of metabolizing enzymes, polymorphisms of DNA-repair genes are an active research area. The failure to repair DNA damage adequately can lead to the

633

induction and stabilization of somatic mutations. These mutations affect several cellular systems that prevent uncontrolled growth including cell cycle control, programmed cell death, immune surveillance, and other mechanisms that, once dysregulated, transform cells into the malignant state. At the molecular level, the specific events along this pathway have become better understood and include alterations of tumor-suppressor genes such as p16 and p53, as well as amplification of oncogenes such as cyclin D1. Some have attempted to map these events to the sequence of pathologic change from normal tissue to carcinoma (see “Molecular Pathogenesis” section). At some loci the putative target tumor-suppressor or oncogenes have been identified. For some genes, such as p53 and p16, specific mechanisms of inactivation and potential relationship with carcinogens have been explored. Although more limited data currently is available specifically for laryngeal cancer it has been suggested that it has a similar general model of pathogenesis. Despite advances in understanding the pathogenesis and molecular basis of head and neck cancer, including the role of tobacco, much work remains to refine this model.

PREVENTIVE MEASURES Primary Prevention Worldwide, separate and joint effects of alcohol and tobacco consumption cause most laryngeal cancer cases. A European study estimated attributable risks of 90% and 58% for the separate effects of tobacco and alcohol, respectively (Talamini et al., 2002), while another study reported an attributable risk of greater than 75% for the joint effects (Franceshi et al., 1990). Thus, primary prevention efforts should be aimed at reducing these products in the population. Studies have shown that smoking cessation decreases risk; one study reported a decline starting at 3 years after cessation. Although the independent and joint effects of alcohol do not appear as strong as tobacco alone, reduction in heavy alcohol consumption would appear prudent. Although tobacco and alcohol use constitute the major risk factors that warrant preventive measures, other exposures may contribute to risk, either independently or in combination with tobacco and alcohol, and may facilitate risk reduction. Increased consumption of diets rich in fruits and vegetables has been reported consistently. Although the specific micronutrients involved in the possible risk reduction have not been isolated, increased fruit and vegetable consumption would appear sufficient.

Chemoprevention Chemoprevention studies in head and neck cancer have focused on preventing second primary tumors and reversing premalignant lesions (Kim et al., 2002; Papadimitrakopoulou et al., 2002). Much of this work has examined the efficacy of retinoids. These trials have suggested that b-carotene and vitamin A can reverse early oral premalignant lesions (Hong et al., 1986; Stich et al., 1988). Other studies have evaluated lower doses and other agents such as vitamin E. A small trial of combination isotretinoin, a-tocopherol, and interferon alpha administered for 6 months to 23 persons with mild, moderate, or severe laryngeal dysplasia found about a 50% complete response at 6 and 12 months follow-up. This response was much higher than the response for oral lesions (Papadimitrakopoulou et al., 1999). Chemoprevention trials have also examined prevention of primary recurrences or second primary tumors. A randomized, double-blinded, placebo-controlled trial of daily high-dose 13-cRA (50–100 mg/m2) for 12 months among 103 patients with head and neck cancer (36 patients with laryngeal cancer) showed a significantly lower occurrence of second primary tumors among the 13-cRA group with a median follow-up of 32 months (Hong et al., 1990). However, these high-dose patients developed toxic effects requiring dose reduction. The same research group has undertaken a new randomized trial of low-dose (30 mg/day) 13cRA with daily use (vs. placebo) for 3 years with 4 years of followup in a group of patients with head and neck cancer (including laryngeal). These results have not been published yet.

634

PART IV: CANCER BY TISSUE OF ORIGIN

Screening and Early Detection The typical symptoms of laryngeal cancer, especially glottic tumors, offer some potential for early detection. The most common presentation for glottic tumors is the onset of chronic hoarseness. Because glottic tumors present early with voice changes, lower-stage detection and improved survival are evident for these tumors in comparison with laryngeal cancer at other sites. A Japanese mass screening trial of hoarseness with 941 persons reported encouraging results (Ono, 1980). Continued improvements in laryngoscopy and imaging techniques likely would permit earlier detection of laryngeal cancer (Thekdi and Ferris, 2002). Researchers also are exploring the use of molecular methods to screen smokers for head and neck cancer (Hu et al., 2002). Interesting results have been obtained using saliva samples to detect p53 mutations, promotor hypermethylation, and mitochondrial DNA mutations in exfoliated cells (Boyle et al., 1994; Fliss et al., 2000; Rosas et al., 2001). Serum has also been used to detect molecular markers related to cancer development such as microsatellite alterations and promotor methylation (Nawroz et al., 1996; Sanchez-Cespedes et al., 2000). The use of saliva and serum samples to detect molecular alterations associated with cancer progression may help to detect cancer earlier and to monitor patients for the development of recurrence and second primary tumors.

FUTURE DIRECTIONS Despite the strong and consistent association between tobacco and alcohol use and the risk of laryngeal cancer, several areas of research are warranted. These include refining risk factor patterns and enhancing understanding of laryngeal carcinogenesis mechanisms. Finally, there is a critical need for more early detection and prevention research. The incidence and mortality patterns of laryngeal cancer generally have been well described worldwide and in the United States. Research into the causes and prevention of the notably poorer survival among certain groups, such as blacks, deserves serious attention. Questions remain regarding certain aspects of the association between tobacco and alcohol and laryngeal cancer. These include: (1) differential effects of types of tobacco, (2) the precise nature of the independent effects of alcohol and joint effects with tobacco, (3) differential effects of alcohol type, and (4) heterogeneity of effect at different laryngeal subsites. Many occupations, industries, and exposures have been related to increased risk of laryngeal cancer. Additional research, especially with improved exposure assessment, is needed to pinpoint more precisely and validly specific agents in the workplace that might increase risk. Epidemiologic studies of diet have suggested that several food groups and nutrients may reduce risk. Additional studies conducted in diverse populations with improved dietary data collection and assessment should be conducted to confirm previous associations and improve understanding of the critical components of a complex diet association. The etiologic role of HPV in laryngeal cancer is uncertain. Studies of laryngeal tumors using sensitive and specific techniques to detect HPV and to determine integration status and viral gene expression are required. An active area of research is the attempt to identify genetic susceptibility factors that may modify risk, especially in combination with tobacco and alcohol exposure. Studies have been conducted examining polymorphisms of genes related to tobacco and alcohol metabolism and DNA repair. The studies generally have been limited in the number of laryngeal cancer cases and polymorphisms. Future studies must be designed carefully to ensure adequate power to estimate gene-gene and gene-exposure effects precisely. The general molecular-genetic progression model for head and neck cancer has been proposed. Further work is needed with laryngeal tumors to identify all the key genes that are altered and their places in the sequence. Expression microarrays can efficiently provide important clues to the set of genes that may be altered in different stages of

laryngeal carcinogenesis. Not only will the identification of critical target genes be of interest for understanding mechanisms, but it will also facilitate the continuing development of new molecular markers and techniques for rapid evaluation and application of molecular screening methods. Many of the aforementioned research questions can be examined in the context of more holistic molecular epidemiologic studies. Large population-based studies collecting exposure histories, germline DNA, and tumor samples would facilitate the broader examination of exposure, genotype, and somatic alteration. These relationships potentially could be examined within important subgroups. In some populations, these studies could also examine other issues such as health disparities related to social class and ethnicity. Finally, a critical area that cannot be overlooked is prevention and control. Significant progress is needed to develop more effective methods of early detection, smoking and alcohol cessation, and chemoprevention. References Ahlborg G, Hogstedt C, Sundell L, Aman CG. 1981. Laryngeal cancer and pickling house vapors. Scand J Work Environ Health 7:239–240. Almadori G, Cadoni G, Cattani P, et al. 2001. Human papillomavirus infection and epidermal growth factor receptor expression in primary laryngeal squamous cell carcinoma. Clin Cancer Res 7:3988–3993. Altieri A, Bosetti C, Talamini R, et al. 2002. Cessation of smoking and drinking and the risk of laryngeal cancer. Br J Cancer 87:1227–1229. American Cancer Society. 2005. Cancer Facts & Figures 2005. American Cancer Society, Inc., Atlanta, GA. Assimakopoulos D, Patrikakos G. 2002. The role of gastroesophageal reflux in the pathogenesis of laryngeal carcinoma. Am J Otolaryngol 23:351– 357. Austin DF, Reynolds P. 1996. Laryngeal cancer. In: Schottenfeld D, Faumeni JF, eds. Cancer Epidemiology and Prevention. 2nd ed. Oxford University Press, pp. 619–636. Bagnardi V, Biangiardo M, La Vecchia C, Corrao G. 2001. A meta-analysis of alcohol drinking and cancer risk. Br J Cancer 85:1700–1705. Baron AE, Franceschi S, Barra S, Talamini R, La Vecchia C. 1993. A comparison of the joint effects of alcohol and smoking on the risk of cancer across sites in the upper aerodigestive tract. Cancer Epidemiol Biomarkers Prev 2:519–523. Bartsch H, Nair U, Risch A, Rojas M, Wikman H, Alexandrov K. 2000. Genetic polymorphism of CYP genes, alone or in combination, as a risk modifier of tobacco-related cancers. Cancer Epidemiol Biomarkers Prev 9:3–28. Berrino F, Richiardi L, Boffetta P, et al. 2003. Occupation and larynx and hypopharynx cancer: A job-exposure matrix approach in an international case-control study in France, Italy, Spain and Switzerland. Cancer Causes Control 14:213–223. Boffetta P, Richiardi L, Berrino F, et al. 2003. Occupation and larynx and hypopharynx cancer: An international case-control study in France, Italy, Spain, and Switzerland. Cancer Causes Control 14:203–212. Borras E, Coutelle C, Rosell A, et al. 2000. Genetic polymorphism of alcohol dehydrogenase in europeans: The ADH2*2 allele decreases the risk for alcoholism and is associated with ADH3*1. Hepatology 31:984–989. Bosetti C, Gallus S, Franceschi S, et al. 2002a. Cancer of the larynx in nonsmoking alcohol drinkers and in non-drinking tobacco smokers. Br J Cancer 87:516–518. Bosetti C, La Vecchia C, Talamini R, et al. 2002b. Food groups and laryngeal cancer Risk: A case-control study from Italy and Switzerland. Int J Cancer 100:355–360. Bouquot JE, Gnepp DR. 1991a. Laryngeal precancer: A review of the literature, commentary, and comparison with oral leukoplakia. Head Neck 13:488–497. Bouquot JE, Weiland LH, Kurland LT. 1991b. Laryngeal keratosis and carcinoma in the Rochester, Minnesota population, 1935–1984. Cancer Detect Prev 5:83–91. Boyle JO, Mao L, Brennan JA, et al. 1994. Gene mutations in saliva as molecular markers for head and neck squamous cell carcinomas. Am J Surg 168:429–432. Boyle P, Zheng T, Macfarlane GJ, et al. 1990. Recent advances in the etiology and epidemiology of head and neck cancer. Curr Opin Oncol 2:539–545. Brennan JA, Boyle JO, Koch WM, et al. 1995. Association between cigarette smoking and mutation of the p53 gene in squamous cell carcinoma of the head and neck. N Engl J Med 332:712–717. Brennan P, Lewis S, Hashibe M, et al. 2004. Pooled-analysis of alcohol dehydrogenase genotypes and head and neck cancer—a HuGE review. Am J Epidemiol 159:1–16.

Cancer of the Larynx Burch JD, Howe GR, Miller AB, Semenciw R. 1981. Tobacco, alcohol, asbestos, and nickel in the etiology of cancer of the larynx: A case-control study. J Natl Cancer Inst 67:1219–1224. Califano J, van der Riet P, Westra W, et al. 1996. Genetic progression model for head and neck cancer: Implications for field cancerization. Cancer Res 56:2488–2492. Caporaso N, Goldstein A. 1995. Cancer genes: single and susceptibility: Exposing the difference. Pharmacogenetics 5:59–63. Carey TE, Van Dyke DL, Worsham MJ. 1993. Nonrandom chromosome aberrations and clonal populations in head and neck cancer. Anticancer Res 13:2561–2567. Cascorbi I, Henning S, Brockmoller J, et al. 2000. Substantially reduced risk of cancer of the aerodigestive tract in subjects with variant—463A of the myeloperoxidase gene. Cancer Res 60:644–649. Chan CK, Gee BL. 1988. Asbestos exposure and laryngeal cancer: An analysis of the epidemiologic evidence. J Occup Med 30:23–27. Cooper MP, Jovanovic A, Nauta JJP, et al. 1995. Role of genetic factors in the etiology of squamous cell carcinoma of the head and neck. Arch Otolaryngol Head Neck Surg 121:157–160. De Stefani E, Boffetta P, Oreggia F, et al. 2000a. Plant foods and risk of laryngeal cancer: A case-control study in Uruguay. Int J Cancer 87: 129–132. De Stefani E, Boffetta P, Oreggia F, Ronco A, Kogevinas M, Mendilaharsu M. 1998a. Occupation and the risk of laryngeal cancer in Uruguay. Am J Ind Med 33:537–542. De Stefani E, Correa P, Oreggia F, et al. 1987. Risk factors for laryngeal cancer. Cancer 60:3087–3091. De Stefani E, Oreggia F, Rivero S, Fierro L. 1992. Hand-rolled cigarette smoking and risk of cancer of the mouth, pharynx, and larynx. Cancer 70:679–682. De Stefani E, Ronco A, Mendilaharsu M, Deneo-Pellegrini H. 1998b. A casecontrol study on the role of heterocyclic amines in the etiology of upper digestive cancers in Uruguay. Nutr Cancer 32:43–8. De Stefani E, Oreggia F, Boffetta P, Deneo-Pellegrini H, Ronco A, Mendilaharsu M. 2000b. Tomatoes, tomato-rich foods, lycopene and cancer of the upper aerodigestive tract: A case-control in Uruguay. Oral Oncol 36:47–53. Doll R, Peto J. 1985. Effects on Health of Exposure to Asbestos. London:Health and Safety Commission: Her Majesty’s Stationery Office. Dosemeci M, Gokmen I, Unsal M, Hayes RB, Blair A. 1997. Tobacco, alcohol use, and risks of laryngeal and lung cancer by subsite and histologic type in Turkey. Cancer Causes Control 8:729–737. Dyson N, Howley PM, Munger K, Harlow E. 1989. The human papilloma virus—16 E7 oncoprotein is able to bind to the retinoblastoma gene product. Science 243:934–936. Edenberg HJ and Bosron WF. 1997. Alcohol dehydrogenases. In: Guengerich FP, ed. Comprehensive Toxicology. Vol 3: Biotransformation. New York: Pergamon Press, pp. 119–131. Edelman DA. 1989. Laryngeal cancer and occupational exposure to asbestos. Int Arch Occup Environ Health 61:223–227. Elahi A, Zheng Z, Park J, Eyring K, McCaffrey T, Lazarus P. 2002. The human OGG1 DNA repair enzyme and its association with orolaryngeal cancer risk. Carcinogenesis 23:1229–1234. El-Serag HB, Hepworth EJ, Lee P, Sonnenberg A. 2001. Gastroesophageal reflux disease is a risk factor for laryngeal and pharyngeal cancer. Am J Gastroenterol 96:2013–2018. Elwood JM, Pearson JC, Skippen DH, Jackson SM. 1984. Alcohol, smoking, social and occupational factors in the aetiology of cancer of the oral cavity, pharynx and larynx. Int J Cancer 34:603–612. Esteve J, Riboli E, Pequignot G, et al. 1996. Diet and cancers of the larynx and hypopharynx: The IARC multi-center study in southwestern Europe. Cancer Causes Control 7:240–252. Falk RT, Pickle LW, Brown LM, Mason TJ, Buffler PA, Fraumeni JF. 1989. Effect of smoking and alcohol consumption on laryngeal cancer risk in coastal Texas. Cancer Res 49:4024–4029. Ferlay J, Bray F, Pisani P, Parkin DM. 2001. GLOBOCAN 2000: Cancer Incidence, Mortality and Prevalence Worldwide, Version 1.0. IARC CancerBase No. 5. Lyon, IARC Press. Flanders WD, Rothman KJ. 1982. Interaction of alcohol and tobacco in laryngeal cancer. Am J Epidemiol 115:371–379. Fliss MS, Usadel H, Caballero OL, et al. 2000. Facile detection of mitochondrial DNA mutations in tumors and bodily fluids. Science 287:2017– 2019. Foulkes WD, Brunet J-S, Kowalski LP, Narod SA, Franco EL. 1995. Family history of cancer is a risk factor for squamous cell carcinoma of the head and neck in Brazil: A case-control study. Int J Cancer 63:769–773. Franceschi S, Munoz N, Snijders PJ. 2000. How strong and how wide is the link between HPV and oropharyngeal cancer? Lancet 356:871–872.

635

Franceschi S, Talamini R, Barra S, et al. 1990. Smoking and drinking in relation to cancers of the oral cavity, pharynx, larynx, and esophagus in Northern Italy. Cancer Res 50:6502–6507. Freudenheim JL, Graham S, Byers TE, et al. 1992. Diet, smoking, and alcohol in cancer of the larynx: A case-control study. Nutr Cancer 17:33–45. Fritz A, Percy C, Jack A, et al., eds. 2000. International Classification of Diseases for Oncology. 3rd ed. Geneva: World Health Organization. Gallus S, Bosetti C, Franceschi S, Levi F, Negri E, La Vecchia C. 2003. Laryngeal cancer in women: tobacco, alcohol, nutritional, and hormonal factors. Cancer Epidemiol Biomarkers Prev 12:514–517. Geisler SA, Olshan AF. 2001. GSTM1, GSTT1, and the risk of squamous cell carcinoma of the head and neck: A mini-huge review. Am J Epidemiol 154:95–105. Gencik A, Wey W, Muller H. 1986. High incidence of laryngeal carcinoma in a Swiss family. J Otorhinolaryngol Relat Spec 48:162–166. Gillison ML, Koch WM, Capone RB, et al. 2000. Evidence for a casual association between human papillomavirus and a subset of head and neck cancers. J Natl Cancer Inst 92:709–720. Gillison ML, Shah KV. 2001. Human papillomavirus-associated head and neck squamous cell carcinoma: Mounting evidence for an etiologic role for human papillomavirus in a subset of head and neck cancers. Curr Opin Oncol 13:183–188. Go C, Schwartz MR, Donovan DT. 2003. Molecular transformation of recurrent respiratory papillomatosis: Viral typing and p53 overexpression. Ann Otol Rhinol Laryngol 112:298–302. Goldberg M, Goldberg P, Leclerc A, Chastang JF, Marne MJ, Dubourdieu D. 1994. A 10-year incidence survey of respiratory cancer and a case-control study within a cohort of nickel mining and refining workers in New Caledonia. Cancer Causes Control 5:15–25. Goldberg P, Leclerc A, Luce D, Morcet J-F, Brugère J. 1997. Laryngeal and hypopharyngeal cancer and occupation: Results of a case-controls study. Occup Environ Med 54:477–482. Goldstein AM, Blot WJ, Greenberg RS, et al. 1994. Familial risk in oral and pharyngeal cancer. Eur J Cancer B Oral Oncol 30B:319–322. Goode EL, Ulrich CM, Potter JD. 2002. Polymorphisms in DNA repair genes and associations with cancer risk. Cancer Epidemiol Biomarkers Prev 11:1513–1530. Goodman M, Morgan RW, Ray R, Malloy CD, Zhao K. 1999. Cancer in asbestos-exposed occupational cohorts: A meta-analysis. Cancer Causes Control 10:453–465. Greenblatt MS, Bennett WP, Hollstein M, Harris CC. 1994. Mutations in the p53 tumor suppressor gene: Clues to cancer etiology and molecular pathogenesis. Cancer Res 54:4855–4878. Grove AD, Llewellyn GC, Kessler FK, White KL Jr, Crespi CL, Ritter JK. 2000. Differential protection by rat UDP-glucuronosyltransferase 1A7 against Benzo [a] pyrene-3, 6-quinone- versus Benzo [a] pyrene-induced cytotoxic effects in human lymphoblastoid cells. Toxicol Appl Pharmacol 162:34–43. Guenel P, Chastang JF, Luce D, Leclerc A, Brugere J. 1988. A study of the interaction of alcohol drinking and tobacco smoking among French cases of laryngeal cancer. J Epidemiol Commun Health 42:350–354. Gustavsson P, Jakobsson R, Johansson H, Lewin F, Norell S, Rutkvist L-E. 1997. Occupational exposures and squamous cell carcinoma of the oral cavity, pharynx, larynx, and oesophagus: A case-control study in Sweden. Occup Environ Med 55:393–400. Ha PK, Califano JA. 2002. The molecular biology of laryngeal cancer. Otolaryngol Clin N Am 35:993–1012. Ha PK, Pai SI, Westra WH, et al. 2002. Real-time quantitative PCR demonstrates low prevalence of human papillomavirus type 16 in premalignant and malignant lesions of the oral cavity. Clin Cancer Res 8:1203– 1209. Hecht SS, Hoffmann D. 1988. Tobacco-specific nitrosamines, an important group of carcinogens in tobacco and tobacco smoke. Carcinogenesis 9:875–884. Hedberg K, Vaughan TL, White E, Davis S, Thomas DB. 1994. Alcoholism and cancer of the larynx: A case-control study in western Washington (United States). Cancer Causes Control 5:3–8. Hong WK, Endicott J, Itri LM, et al. 1986. 13-cis-retinoic acid in the treatment of oral leukoplakia. N Engl J Med 315:1501–1505. Hong WK, Lippman SM, Itri LM, et al. 1990. Prevention of second primary tumors with isotretinoin in squamous-cell carcinoma of the head and neck. N Engl J Med 323:795–801. Hu YC, Sidransky D, Ahrendt SA. 2002. Molecular detection approaches for smoking associated tumors. Oncogene 21:7289–7297. IARC. 1986. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, vol. 38: Tobacco Smoking. Lyon: International Agency for Research on Cancer.

636

PART IV: CANCER BY TISSUE OF ORIGIN

IARC. 1988. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, vol. 44: Alcohol Drinking. Lyon: International Agency for Research on Cancer. IARC. 1995. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Vol. 62: Wood dust and formaldehyde. Lyon: International Agency for Research on Cancer. Jourenkova-Mironova N, Mitrunen K, Bouchardy C, Dayer P, Benhamou S, Hirvonen A. 2000. High-activity microsomal epoxide hydrolase genotypes and the risk of oral, pharynx, and larynx cancers. Cancer Res 60:534–536. Jourenkova-Mironova N, Voho A, Bouchardy C, et al. 1999. Glutathione Stransferase GSTM3 and GSTP1 genotypes and larynx cancer risk. Cancer Epidemiol Biomarkers Prev 8:185–188. Jusawalla DJ, Deshpande VA. 1971. Evaluation of cancer risk in tobacco chewers and smokers: An epidemiologic assessment. Cancer 28:244–252. Katoh T, Kaneko S, Kohshi K, et al. 1999. Genetic polymorphisms of tobaccoand alcohol-related metabolizing enzymes and oral cavity cancer. Int J Cancer 83:606–609. Kim ES, Hong WK, Khuri FR. 2002. Chemoprevention of aerodigestive tract cancers. Annu Rev Med 53:223–243. La Vecchia C, Negri E, D’Avanzo B, Franceschi S, Decarli A, Boyle P. 1990. Dietary indicators of laryngeal cancer risk. Cancer Res 50:4497– 4500. Lazarus P, Park JY. 2000. Metabolizing enzyme genotype and risk for upper aerodigestive tract cancer. Oral Oncol 36:421–431. Levi F, Lucchini F, Negri E, Boyle P, La Vecchia C. 1999. Cancer mortality in Europe, 1990–1994, and an overview of trends from 1955 to 1994. Eur J Cancer 30:1477–1516. Liddell FDK. 1990. Laryngeal cancer and asbestos. Br J Ind Med 47:289–291. Lie ES, Engh V, Boysen M, et al. 1994. Squamous cell carcinoma of the respiratory tract following laryngeal papillomatosis. Acta Otolaryngol 114:209–212. Lindeberg H, Elbrond O. 1990. Laryngeal papillomas: The epidemiology in a Danish subpopulation 1965–1984. Clin Otolaryngol 15:125–131. Lindeberg H, Krogdahl A. 1999. Laryngeal cancer and human papillomavirus: HPV is absent in the majority of laryngeal carcinomas. Cancer Lett 146:9–13. Lopez-Abente G, Pollan M, Monge V, Martinezvidal A. 1992. Tobacco smoking, alcohol consumption, and laryngeal cancer in Madrid. Cancer Detect Prev 16:265–271. Lynch J, Hanis NM, Bird MG, Murray KJ, Walsh JP. 1979. An association of upper respiratory cancer with exposure to diethyl sulfate. J Occup Med 211:333–341. Magnus K, Andersen A, Hogetveit AC. 1982. Cancer of respiratory organs among workers at a nickel refinery in Norway. Int J Cancer 30:681–685. Maier H, De Vries N, Snow GB. 1991. Occupational factors in the aetiology of head and neck cancer. Clin Otolaryngol 16:406–412. Maier H, Dietz A, Gewelke U, Heller W-D, Weidauer H. 1992. Tobacco and alcohol and the risk of head and neck cancer. Clin Investig 70:320–327. Manning KP, Skegg DCG, Stell PM, Doll R. 1981. Cancer of the larynx and other occupational hazards of mustard gas workers. Clin Otolaryngol 6:165–170. Marchand JL, Luce D, Leclerc A, et al. 2000. Laryngeal and hypopharyngeal cancer and occupational exposure to asbestos and man-made vitreous fibers: Results of a case-control study. Am J Ind Med 37:581–589. Matthias C, Bockmuhl U, Jahnke V, et al. 1998. The glutathione S-transferase GSTP1 polymorphism: Effects on susceptibility to oral/pharyngeal and laryngeal carcinomas. Pharmacogenetics 8:1–6. McKaig RG, Baric RS, Olshan AF. 1998. Human papillomavirus and head and neck cancer: Epidemiology and molecular biology. Head Neck 20:250–265. Morita S, Yano M, Tsujinaka T, et al. 1999. Genetic polymorphisms of drugmetabolizing enzymes and susceptibility to head-and-neck squamous-cell carcinoma. Int J Cancer 80:685–688. Mork J, Lie K, Glattre E, et al. 2001. Human papillomavirus infection as a risk factor for squamous-cell carcinoma of the head and neck. N Engl J Med 344:1125–1131. Muller KM, Krohn BR. 1980. Smoking habits and their relationship to precancerous lesions of the larynx. J Cancer Res Clin Oncol 96:211–217. Muscat JE, Wynder EL. 1992. Tobacco, alcohol, asbestos, and occupational risk factors for laryngeal cancer. Cancer 69:2244–2251. Nawroz H, Koch W, Anker P, Stroun M, Sidransky D. 1996. Microsatellite alterations in serum DNA of head and neck cancer patients. Nat Med 2:1035–1037. Nomura T, Noma H, Shibahara T, Yokoyama A, Muramatusu T, Ohmori T. 2000. Aldehyde dehydrogenase 2 and glutathione S-transferase M 1 polymorphisms in relation to the risk for oral cancer in Japanese drinkers. Oral Oncol 36:42–46.

Olsen J, Sabreo S, Fasting U. 1985. Interaction of alcohol and tobacco as risk factors in cancer of the laryngeal region. J Epidemiol Community Health 39: 165–168. Olshan AF, Weissler MC, Pei H, Conway K. 1997. p53 mutations in head and neck cancer: New data and evaluation of mutational spectra. Cancer Epidemiol Biomarkers Prev 6:499–504. Ono I. 1980. Mass screening for the detection of early carcinoma of the larynx. Nippon Jibiinkoka Gakkai Kaiho 83:1498–1502. Oreggia F, De Stefani E, Boffetta P, Brennan P, Deneo-Pellegrini H, Ronco AL. 2001. Meat, fat and risk of laryngeal cancer: A case-control study in Uruguay. Oral Oncol 37:141–145. Osier MV, Pakstis AJ, Soodyall H, et al. 2002. A global perspective on genetic variation at the ADH genes reveals unusual patterns of linkage disequilibrium and diversity. Am J Hum Genet 71:84–99. Papadimitrakopoulou VA, Clayman GL, Shin DM, et al. 1999. Biochemoprevention for dysplastic lesions of the upper aerodigestive tract. Arch Otolaryngol Head Neck Surg 125:1083–1089. Papadimitrakopoulou VA. 2002. Chemoprevention of head and neck cancer: An update. Curr Opin Oncol 14:318–322. Parkin DM, Bray F, Ferlay J, Pisani P. 2001. Estimating the world cancer burden: Globocan 2000. Int J Cancer 94:153–156. Parkin DM, Whelan SL, Ferlay J, Teppo L, Thomas D, eds. 2002. Cancer Incidence in Five Continents, Vol. VIII. IARC Scientific Publications No. 155. Parnes SM. 1990. Asbestos and cancer of the larynx: Is there a relationship? Laryngoscope 100:254–261. Pedersen E, Hogetveit AC, Andersen A. 1973. Cancer of respiratory organs among workers at a nickel refinery in Norway. Int J Cancer 12:32–41. Pelucchi C, Talamini R, Levi F, et al. 2003. Fibre intake and laryngeal cancer risk. Ann Oncol 14:162–167. Percy C, Van Holten V, Muir C, Eds. 1990. International Classification of Diseases for Oncology. 2nd ed. Geneva: World Health Organization. Reed AL, Califano J, Cairns P, et al. 1996. High frequency of p16 (CDKN2/MTS-1/INK4A) inactivation in head and neck squamous cell carcinoma. Cancer Res 56:3630–3633. Ren Q, Murphy SE, Zheng Z, Lazarus P. 2000. O-Glucuronidation of the lung carcinogen 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol (NNAL) by human UDP-glucuronosyltransferase 2B7 and 1A9. Drug Metab Dispos 28:1352–1360. Renan MJ. 1993. How many mutations are required for tumorigenesis? Implications from human cancer data. Mol Carcinog 7:139–146. Riboli E, Kaaks R, Esteve J. 1996. Nutrition and laryngeal cancer. Cancer Causes Control 7:147–156. Rodu B, Cole P. 2002. Smokeless tobacco use and cancer of the upper respiratory tract. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 93:511–515. Rogers MA, Thomas DB, Davis S, Vaughan TL, Nevissi AE. 1993. A casecontrol study of element levels and cancer of the upper aerodigestive tract. Cancer Epidemiol Biomarkers Prev 2:305–312. Rosas SL, Koch W, da Costa Carvalho MG, et al. 2001. Promoter hypermethylation patterns of p16, O6-methylguanine-DNA-methyltransferase, and death-associated protein kinase in tumors and saliva of head and neck cancer patients. Cancer Res 61:939–942. Rothman KJ, Greenland S. 1998. Concepts of Interaction, Modern Epidemiology. 2nd ed. Philadelphia: Lippincott Raven, pp. 329–342. Rothman KJ, Cann CI, Flanders D, Fried MP. 1980. Epidemiology of laryngeal cancer. Epidemiol Rev 2:195–209. Rothman KJ, Cann CI, Fried MP. 1989. Carcinogenicity of dark liquor. Am J Public Health 79:1516–1520. Sanchez-Cespedes M, Esteller M, Wu L, et al. 2000. Gene promoter hypermethylation in tumors and serum of head and neck cancer patients. Cancer Res 60:892–895. Sancho-Garnier H, Theobald S. 1993. Black (air-cured) and blond (flue-cured) tobacco and cancer risk. II: Pharynx and larynx cancer. Eur J Cancer 29A:273–276. Shaheen N, Ransohoff DF. 2002. Gastroesophageal reflux, Barrett esophagus, and esophageal cancer: Scientific review. JAMA 287:1972–1981. Scheffner M, Werness BA, Hulbregste JM, Levine AJ, Howley PT. 1990. The E6 oncoprotein encoded by human papillomavirus types 16 and 18 promotes the degradation of p53. Cell 63:1129–1136. Schlecht NF, Franco EL, Pintos J, et al. 1999. Interaction between tobacco and alcohol consumption and the risk of cancers of the upper aero-digestive tract in Brazil. Am J Epidemiol 150:1129–1137. Schlecht NF, Pintos J, Kowalski LP, Franco EL. 2001. Effect of type of alcoholic beverages on the risks of upper aerodigestive tract cancers in Brazil. Cancer Causes Control 12:579–587. Schwartz MR. 1999. Pathology of laryngeal tumors. In: Thawley SE, Panje WR, Batsakis JG, Lindberg RD, eds. Comprehensive Management of

Cancer of the Larynx Head and Neck Tumors, Vol. 2. 2nd ed. Philadelphia: WB Saunders, pp. 950–978. SEER. 2003. Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database: Incidence—SEER 9 Regs, Nov 2002 Sub (1973–2000), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2003, based on the November 2002 submission. Smith AH, Handley MA, Wood R. 1990. Epidemiological evidence indicates asbestos causes laryngeal cancer. J Occup Med 32:499–507. Smith M. 1986. Genetics of human alcohol and aldehyde dehydrogenases. Adv Hum Genet 15:249–290. Soskolne CL, Zeighami EA, Hanis NM, et al. 1984. Laryngeal cancer and occupational exposure to sulfuric acid. Am J Epidemiol 120:358–369. Steenland K, Schnorr T, Beaumont J, Halperin W, Bloom T. 1988. Incidence of laryngeal cancer and exposure to acid mist. Br J Ind Med 45:766–776. Stich HF, Hornby AP, Mathew B, Sankaranarayanan R, Nair MK. 1988. Response of oral leukoplakias to the administration of vitamin A. Cancer Lett 40:93–101. Talamini R, Bosetti C, La Vecchia C, et al. 2002. Combined effect of tobacco and alcohol on laryngeal cancer risk: A case-control study. Cancer Causes Control 13:957–964. Thekdi AA, Ferris RL. 2002. Diagnostic assessment of laryngeal cancer. Otolaryngol Clin North Am 35:953–969. Tuyns AJ, Esteve J, Raymond L, et al. 1988. Cancer of the larynx/hypopharynx, tobacco and alcohol: IARC international case-control study in Turin and Varese (Italy), Zaragoza and Navarra (Spain), Geneva (Switzerland) and Calvados (France). Int J Cancer 41:483–491. Van Dyke ML, Worsham MJ, Benninger MS, et al. 1994. Recurrent cytogenetic abnormalities in squamous cell carcinomas of the head and neck region. Genes Chromosomes Cancer 9:192–206. Van Houten VMM, Snijders PJF, van den Brekel MWM, et al. 2001. Biological evidence that human papillomaviruses are etiologically involved in a subgroup of head and neck squamous cell carcinomas. Int J Cancer 93:232–235.

637

Venuti A, Manni V, Morello R, De Marco F, Marzetti F, Marcante ML. 2000. Physical state and expression of human papillomavirus in laryngeal carcinoma and surrounding normal mucosa. J Med Virol 60:396– 402. Wada S, Nishimoto Y, Miyanishi M, Kambe S. 1968. Mustard gas as a cause of respiratory neoplasia in man. Lancet I:1161–1163. Wortley P, Vaughan TL, Davis S, Morgan MS, Thomas DB. 1992. A casecontrol study of occupational risk factors for laryngeal cancer. Br J Ind Med 49:837–844. Wynder EL, Bross IJ, Day E. 1956. A study of environmental factors in cancer of the larynx. Cancer 9:86–110. Wynder EL, Covey LS, Mabuchi K, Mushinski M. 1976. Environmental factors in cancer of the larynx: a second look. Cancer 38:1591–1601. Yokoyama A, Muramatsu T, Omori T, et al. 2001. Alcohol and aldehyde dehydrogenase gene polymorphisms and oropharyngolaryngeal, esophageal and stomach cancers in Japanese alcoholics. Carcinogenesis 22:433–439. Yoshida A, Huang IY, Ikawa M. 1984. Molecular abnormality of an inactive aldehyde dehydrogenase variant commonly found in Orientals. Proc Natl Acad Sci 81:258–261. Yu GP, Zhang ZF, Hsu TC, Spitz MR, Schantz SP. 1999. Family history of cancer, mutagens sensitivity, and increased risk of head and neck cancer. Cancer Lett 146:93–101. Zatonski W, Becher H, Lissowska J, Wahrendorf J. 1991. Tobacco, alcohol, and diet in the etiology of laryngeal cancer: A population-based case-control study. Cancer Causes Control 2:3–10. Zhang Z-F, Morgenstern H, Spitz MR, et al. 1999. Marijuana use and increased risk of squamous cell carcinoma of the head and neck. Cancer Epidemiol Biomarkers Prev 8:1071–1078. Zheng W, Blot WJ, Shu X-O, et al. 1992. Diet and other risk factors for laryngeal cancer in Shangai, China. Am J Epidemiol 136:178–191. Zheng Z, Park JY, Guillemette C, Schantz SP, Lazarus P. 2001. Tobacco carcinogen-detoxifying enzyme UGT1A7 and its association with orolaryngeal cancer risk. J Natl Cancer Inst 93:1411–1418.

33

Cancer of the Lung MARGARET R. SPITZ, XIFENG WU, ANNA WILKINSON, AND QINGYI WEI

L

ung cancer continues to be the most frequently diagnosed cancer worldwide and the leading cause of cancer deaths in the United States as in many other countries. Annually, about 1.35 million new cases are diagnosed worldwide (Parkin et al., 2005). It is estimated that in the United States in 2005 there will be 172,570 new cases of lung cancer and 163,510 deaths (American Cancer Society (ACS), 2005). These estimates represent 13% of all incident cancers in US men and 12% in women. Further, the deaths represent 31% of all mortality from cancer in US men and 27% in women. Lung cancer exceeded breast cancer as the leading cause of cancer deaths in women in 1987 (ACS, 2005). This chapter provides an update on the epidemiology of lung cancer (Blot and Fraumeni, 1996) with an emphasis on the rapidly expanding literature exploring host susceptibility to tobacco carcinogenesis. Also briefly reviewed are chemoprevention and lung screening trials.

CLASSIFICATION Histopathology While the basic classification of lung cancer has not changed, histologic criteria have evolved in parallel with technologic advances in molecular biology and immunohistochemical techniques (Gazdar and Linnoila, 1988). Four major lung cancer histological subtypes, as classified by conventional light microscopy, comprise over 90% of all cases (Janssen-Heijnen and Coebergh, 2001). These subtypes include squamous cell carcinoma (SCC), large cell carcinoma, adenocarcinoma (AC), and small cell carcinoma (Janssen-Heijnen and Coebergh, 2001). Over the past several decades, trends in histological subtype have shifted in many parts of the world. In the United States, over the past decade or two, AC has become the most common cell type in both men and women. In 1973–1977 data from the Surveillance, Epidemiology, and End Results (SEER, 2003) program, there were more than twice as many SCC as AC among males while AC was slightly more common among females. By the mid-1980s, however, the excess of SCC was only about 40% among men while AC out-numbered SCC by about 50% among women (Devesa et al., 1991). AC is now the most frequent subtype in many parts of the world, including the United States and Japan (Janssen-Heijnen and Coebergh, 2001; Li X et al., 2001; Harkness et al., 2002; Morita, 2002); however, SCC is still the most frequently occurring subtype in men in Europe and Australia (Fig. 33–1). Changes in classification and pathology techniques can account for only a fraction of this trend (Charloux et al., 1997). Rather, it is speculated that this shift reflects changes in the type of cigarette smoked. There have been decreases in the average nicotine and tar delivery of cigarettes from about 2.7 and 39 mg, in the 1950s to 1.0 and 13.5 mg, respectively, in the 1990s. Since the 1950s the percentage of filtertipped cigarettes has mushroomed. The tobacco in filter cigarettes is richer in nitrate than that of non-filter cigarettes. This raises the yield of N-nitrosamines (Wynder and Muscat, 1995). Stellman et al. (1997) have suggested that smokers of low-yield cigarettes have higher risks for AC while smokers of high-yield cigarettes were more likely to develop SCC. The former are thought to smoke more intensely and inhale more deeply to satisfy the need for nicotine, resulting in greater exposure of bronchioalveolar regions and smaller bronchi to the organ-

638

specific lung carcinogen, NNK. However, a recent analysis of data from the Iowa Women’s Health Study suggested that the excess risk for heavy smokers compared with never smokers was higher for AC (excess risk = 206) than SCC (excess risk = 122) (Yang et al., 2002). Thus AC may be more strongly associated with tobacco exposure than previously recognized.

Preneoplastic Lesions There are three well-recognized preneoplastic conditions—squamous dysplasia and carcinoma in situ (SD/CIS), atypical adenomatous hyperplasia (AAH), and diffuse idiopathic pulmonary neuroendocrine cell hyperplasia (DIP-NECH) (Travis, 2002). The latter is rare and associated with development of multiple carcinoid tumors. SD/CIS is graded according to atypical cells and mitotic figures, and considered a precursor for central bronchial carcinoma, unlike AAH, which is considered an important precursor lesion for peripheral parenchymal AC, and could serve as a surrogate end-point biomarker in chemoprevention trials. No data are available on the risk of progression of AAH or its prevalence in a smoking population (Kerr, 2001). It is thought that AAH may also progress from low to high grade to bronchio-alveolar carcinoma. Large cell neuroendocrine carcinoma is now recognized as a histologically high-grade non-small cell carcinoma. The histopathologic classification of lung cancer will continue to change as clinical practice and biologic knowledge evolve (Franklin, 2000).

DEMOGRAPHIC PATTERNS Trends in Incidence and Mortality In the United States, lung cancer incidence among males peaked in 1984 and has decreased on average -1.4% per year thereafter. Lung cancer mortality showed an increasing trend for males between 1930 and the early 1990s, when levels began to fall. The mortality rate decline among men since 1990 has been statistically significant with average declines ranging from -2.6% and -1.6% per year (Wingo et al., 1999). In contrast, among US women, lung cancer incidence and mortality continued to rise through the mid-1990s with an average 0.1% increase per year for incidence and 1.4% for mortality (Wingo et al., 1999). This increase in incidence for women has now slowed and decreased for the first time from 52.8/100,000 in 1998 to 49.1/100,000 in 2001 (ACS, 2005). A similar trend is not evident for female death rates, although the rate of increase has leveled off. Figure 33–2 depicts lung cancer mortality from 1975–2000 (SEER, 2003). In the United States, lung cancer incidence and mortality rates begin to increase notably between the ages of 45 and 54 years and rise progressively with increasing age until 75 years. Although rates are high in the 75–84 years age category, they begin to decline and substantially decrease in the 85+ category, probably due to competing mortality from other causes (Fig. 33–3). Smoking behavior varies by birth cohort, and rates of lung cancer incidence and mortality reflect previous trends in cigarette use categorized by birth cohort (Jemal et al., 2001; Parkin et al., 2001b). Lung cancer mortality peaked in the 1925–1930 birth cohort for white men and in the 1935–1940 birth cohort for white women (Devesa et al., 1999a). For men and women born later, the birth-cohort risk of lung

639

Cancer of the Lung

Figure 33–1. Distribution of histologic subtypes of lung cancer (age-adjusted standardized incidence) by gender in selected countries. (Source: Parkin et al., 1997).

cancer declined continuously until 1950 (Devesa et al., 1999a). Jemal et al. (2001) have pointed out that there has been an unexpected, but statistically significant, moderation in the rate of decrease of the birth cohort trend for mortality for whites born after 1950. They concluded that this reflected the impact of teenage smoking on lung cancer risk in people under the age of 45. Incidence rates peaked in 1979–1981 for men 50–59 years; in the mid-1980s for men 60–79 years and in the early 1990s for men 80 years and older (Wingo et al., 1999). Incidence rates for women peaked in the mid-1970s for those aged 40–49 years, and in the late 1980s for those aged 50–59. Incidence rates for women aged 60–69 have remained level but have continued to increase for older women (Wingo et al., 1999).

Gender and Ethnicity The ratio of male to female lung cancer incidence and mortality rates is proportionately higher in developing countries (Alberg and Samet,

2003). This is most likely due to the lower or rare prevalence of female cigarette use in those countries. Males have a 7.81% lifetime risk of developing lung cancer and a 7.52% lifetime risk of dying from lung cancer, compared with 5.80% and 4.87% respectively, for women. There is also an ethnic distinction among males in lifetime risk. African American males have an 8.46% risk of a lung cancer diagnosis and a 7.60% risk of death due to lung cancer. A similar difference does not exist between white and African-American females. At least 10 case-control studies have reported higher lung cancer relative risks for women smokers than men (reviewed in Thun et al., 2002). However, this has never been confirmed in cohort studies. In the United States, African Americans have the highest incidence and mortality due to lung cancer followed by whites. Hispanics, on the other hand, have the lowest rates (SEER, 2003) (Fig. 33–4). While the overall rate for American Indians/Alaska Natives is low, there is a wide range in these rates by geographic area with very low rates for those living in New Mexico and very high rates in Alaska natives (Wingo et al., 1999). Other countries have also noted ethnic

Figure 33–2. Lung cancer mortality rates per 100,000 in the United States from 1975 to 2000 by gender, and age-adjusted to the 2000 US standard population (SEER, 2003).

640

PART IV: CANCER BY TISSUE OF ORIGIN Table 33–1. Gender/Racial Differences in Median Age of Diagnosis and Death 1995–1999 SEER Total, All races Whites, total Blacks, total White males Black males White females Black females

Median Age of Diagnosis 70.07 70.0 66.0 70.0 66.0 71.0 66.0

Median Age of Death 71.07 71.0 67.0 71.0 67.0 72.0 68.0

Source: Adapted from Ries et al., 2002.

Figure 33–3. Age distribution of lung cancer incidence and mortality (per 100,000) (SEER, 2003).

differences in lung cancer rates. For example, in New Zealand, Maoris are the highest risk subgroup (Parkin et al., 2001b). Disparities in time trends are also present between white and African American males. Statistically significant decreases in lung cancer death rates are evident for black and Hispanic men, although white males exhibited a steeper decline between 1985 and 1998 (Gadgeel et al., 2001). While incidence rates are similar for white and African American women, statistically significant increasing death rates are noted for black and American Indian/Alaska Native women. Table 33–1 shows differences in median age of diagnosis and death by gender and ethnicity. For reasons that are not yet explained, both white men and women exhibit older ages at diagnosis and death compared with blacks.

Geographic Variation Patterns for lung cancer mortality appear to cluster within areas with historically higher prevalences of cigarette smoking (Devesa et al., 1999a; Jemal et al., 2001; Parkin et al., 2001b; Ries et al., 2002; Giovino, 2002), and these patterns have shifted significantly since the 1950s, largely due to regional and temporal variations in smoking trends (Devesa et al., 1999a, 1999b). From 1950–1969, the highest incidence in white males was observed along the Atlantic and Gulf

Figure 33–4. Lung cancer incidence and mortality rates per 100,000 by race/ethnicity (SEER, 2003).

Coasts and in northeastern urban areas. Between 1970 and 1994, the highest mortality rates spanned across much of the southeast, the Atlantic seaboard, the Gulf coast, and the Mississippi valley. Lung cancer mortality rates for white females have not obviously shifted their geographic pattern. Over a 5-year period in the US (1997–2001), Kentucky had the highest lung cancer incidence (139.5/100,000) and mortality (114.5/100,000) rates among males and had the highest female incidence (70.9/100,000) but Nevada had the highest female mortality (54.1/100,000). The lowest incidence and mortality rates for both sexes were observed in Utah (ACS, 2005). California is expected to lead the US both in lung cancer incident cases (15,150) and deaths (14,350) while Alaska is projected to have the lowest incidence (220) and mortality numbers (210). African American males have the highest rates of lung cancer mortality scattered in urban areas of the midwest and northeast. Globally, lung cancer variation in incidence rates is greater than fourfold among men and fivefold among women (Alberg and Samet, 2003). Only some of this variation could be attributed to data quality and varying diagnostic practices. Most cases are seen in the developed countries of North America, Western Europe, and Australia/New Zealand. In the developed world, lung cancer incidence is decreasing while in developing countries the incidence is beginning to reach the same level as in developed countries, especially in males (Parkin et al., 2001a).

Survival The ratio of incidence to mortality is 0.9 and rates of survival are poor. Currently, the estimated 5-year survival rate for lung cancer worldwide is 11% (Parkin et al., 1999). Age, tumor stage, histological subtype, and treatment impact survival. Thus, it is not surprising that developed countries have higher survival rates than developing countries (13% vs. 9%). The highest 5-year survival rates have been observed in Japan (21%) and North America (20%), while the lowest were seen in northwestern Europe (7%), southern Asia, Middle East, and North Africa (8%) (Parkin et al., 2005). Improvements in diagnostic and therapeutic technologies have contributed to an increase in survival in the United States. In 2000, the 1year relative survival rate was 42% compared with 37% in 1975 (ACS, 2005). For all stages combined, the 5-year relative survival rate is 15%, a rate that has doubled from 6% for 1950–1954 (Ries et al., 2002; ACS, 2005). The death rates begin to stabilize after 3 years (Clegg et al., 2002). Though lung cancer survival rates are poor overall, whites appear to have slightly better rates than African Americans (Fig. 33–5). The survival rates for both races have increased since 1974; however, presently, white males have a 15% survival rate compared with 13% for African American males (Jemal et al., 2005). Similar racial differences are also apparent in survival rates according to tumor stage; 17% of white subjects present with localized disease at diagnosis compared with 14% of African Americans. The comparable percentages for those presenting with distant disease are 39% and 40%, respectively (ACS, 2005). If lung cancer is detected while still localized, the 5-year survival rate is 49%. Decreases in survival beyond the

Cancer of the Lung

(a)

641

(b)

Figure 33–5. Lung cancer survival curves for males and females by race. SEER program 1992–1999 (SEER, 2003).

localized stage are astounding. Survival decreases to 16% if diagnosed at the regional stage, and to 2% if distant stage at diagnosis (ACS, 2005). Survival also decreases with age. Patients older than age 50 have a 19% relative survival compared with those younger than age 75 (9% survival rate) (Fig. 33–6). Observed survival rates are lower than relative rates especially in the older population because of competing causes of death. Histological subtype is also a determinant of survival. Patients with non-small cell carcinoma have a better prognosis than those with small cell carcinoma.

Socioeconomic Status Higher incidence and mortality rates are reported among men from lower than higher socioeconomic (SES) groups in many industrialized nations, including Canada (Mao et al., 2001), France (Marshall et al., 1999), Germany (Heinrich et al., 2000), Scotland (Hart et al., 2001), Spain (Fernandez and Borrell, 1999), and the United States (Krieger et al., 1999; Singh et al., 2002). The pattern among women is less consistent. Fernandez and Borrell (1999) report a negative association between lung cancer mortality rates and SES; others report no association (Mao et al., 2001), and still others report a positive association (Hart et al., 2001; Heinrich et al., 2000; Krieger et al., 1999). These studies used various indicators of SES such as occupation (Marshall et al., 1999), educational level (Fernandez and Borrell, 1999), income

Figure 33–6. Five-year observed and relative overall lung cancer survival rates by age. (Source: Adapted from Edwards et al., 2002.)

level (Mao et al., 2001), social class (Hart et al., 2001; Heinrich et al., 2000), and an index variable composed of 11 different indicators of SES (Singh et al., 2002). The rates generally are higher among those with lower levels of educational attainment. Smoking accounts for most but not all the observed differences in lung cancer (Hart et al., 2001; Mao et al., 2001).

TOBACCO By the early 1950s case-control studies in the United States and Great Britain clearly documented an association between smoking and lung cancer incidence (Levin et al., 1950; Wynder and Graham, 1950; Doll and Hill, 1952). This evidence became conclusive through surveys of large cohorts of smokers. Almost four decades have now passed since the 1964 US Surgeon General’s report on the causal relationship between smoking and lung cancer. Cigarette smoking is the single most important risk factor for lung cancer. It is estimated that about 90% of male lung cancer deaths and 75%–80% of female lung cancer deaths in the United States each year are caused by smoking (Hecht, 1999). Since only a fraction of smokers will develop lung cancer, it is assumed that genetic predisposition also plays a major role in tobacco carcinogenesis (discussed in greater detail below). Peto et al. (2000) have estimated that the cumulative risk of death from lung cancer by age 75 in the United Kingdom (in the absence of other causes of death) rose from 6% at 1950 rates to 16% at 1990 rates in male cigarette smokers, and from 1% to 10% in female cigarette smokers. The relative risk of lung cancer after smoking cessation begins to decrease after 5 years of cessation, but never achieves the risk of a nonsmoker. For men who stopped smoking at ages 60, 50, 40, and 30, the cumulative risks of lung cancer by age 75 were 10%, 6%, 3%, and 2%, respectively (Peto et al., 2000). Thus, those who stopped smoking, even well into middle age, avoid most of their subsequent risk of lung cancer, and stopping before middle age avoids more than 90% of the risk attributable to tobacco. Lung cancer is increasingly a disease of former smokers (Tong et al., 1996), who will constitute an increasing fraction of patients. The extent to which young people become persistent smokers will affect mortality rates in the middle or second half of the 21st century (Peto et al., 2000). There are more than 80 carcinogens in cigarette smoke that have been evaluated by the International Agency for Research on Cancer (IARC) (Smith et al., 2003), and for which it has been deemed that “sufficient evidence for carcinogenicity” exists in humans or laboratory animals. Among the polycyclic aromatic hydrocarbons (PAHs), benzo[a]pyrene has been the most extensively studied, and is a welldocumented lung carcinogen. Among the N-nitrosamines, NNK induces lung AC independent of route of administration (Hecht, 1999).

642

PART IV: CANCER BY TISSUE OF ORIGIN

The NNK dose of lifetime smokers is close to the lowest dose shown to induce lung tumors in rats (Hecht, 1998).

Passive Smoking Most of the carcinogens found in mainstream smoke are also present in sidestream smoke (environmental tobacco smoke, ETS), the material released from the burning tip of a cigarette. In fact, the amount of nitrosamine is substantially higher in sidestream than in mainstream smoke, although there is dilution with ambient air, so that passive uptake is far less. Evidence from epidemiologic studies, as well as from biomarker and toxicology studies, confirms a causal association between lung cancer risk and ETS (Hackshaw et al., 1997; Hackshaw, 1998). The two earliest studies in 1981 (Hirayama, 1981; Trichopolous et al., 1981) showed higher lung cancer risks for nonsmoking women whose husbands smoked than for those whose spouses were nonsmokers. The National Research Council (NRC) reviewed the evidence in 1986 and concluded that nonsmoking spouses were at about 30% increased risk of lung cancer (NRC, 1986). The 1986 Surgeon General Report also deemed passive exposure to be a risk factor for lung cancer (US DHHS, 1986). In 1992 the EPA classified ETS as a human class A carcinogen (US EPA, 1992), and estimated that it accounted for 3000 US lung cancer deaths annually. Several meta-analyses have been conducted to summarize the evidence from epidemiologic studies. From 1981 through the end of 1999, there were about 76 primary studies on ETS and 20 metaanalyses. Pershagen (1994) pooled the data of eight US studies and reported a pooled relative risk estimate (with 95% confidence interval) of 1.23 (1.02, 1.49). He also combined studies from Europe, United States, China, and Japan and reported pooled estimates of 1.23 (1.11, 1.36) for nonsmoking women and 1.82 (0.98, 3.37) for nonsmoking men. Law (1997) reported an approximately 20% increased risk associated with marriage to a smoking spouse. Boffetta et al. (2000) performed a meta-analysis of 30 studies of maternal smoking during pregnancy and reported no increased risk for lung cancer from childhood passive exposure. A meta-analysis by Taylor et al. (2001) included 43 primary studies and the pooled RR for never-smoking women was 1.29 (1.17, 1.43). They also computed sequential meta-analyses results for each year from 1981 and showed that since 1992, the risk has always been above 1.25. In 2000, Copas and Shi (2000) conducted a reanalysis of 37 published studies and concluded that a modest degree of publication bias led to a substantial reduction in the relative risk and to a weaker level of significance. Reynolds (1999) has summarized the evidence related to workplace exposures to ETS. The evidence on ETS and lung cancer is conclusive, but quantitative assessment could still be refined.

Harm Reduction Filtered cigarettes and those with lower tar and nicotine yields have been marketed as safer than high tar products. However, the unexpected result of the introduction of these products was increasing intensity of smoking, especially deeper inhalation, and substantial increase in exposure to tobacco-specific nitrosamines (Shields, 2002). For smokers who cannot or refuse to quit, alternatives are being considered, such as help in reducing the number of cigarettes smoked per day or smoking intensity, as well as new products that use different methods of curing tobacco and that might lead to reduced delivery of carcinogens. The use of potential exposure reduction products (PERPs) to evaluate risk reduction was reviewed recently by Shields (2002). The Institute of Medicine (2001) concluded that harm reduction is feasible, but that the evaluation of actual reduction will require substantial research. Medicinal nicotine is a safer alternative than modified tobacco products (Hatsukami et al., 2004).

Santillan et al., 2003). Both cigarette smoking and chronic respiratory diseases result in a continual cycle of injury and repair and therefore could play a key role in lung carcinogenesis (Fitzpatrick, 2001). Additionally, respiratory diseases may result in chronic immune stimulation leading to random pro-oncogenic mutations in actively dividing stem cells and an increased risk of cancer. Nonetheless, the causal nature of the association between respiratory diseases and lung cancer is still speculative, since both emphysema and chronic bronchitis are strongly influenced by smoking (Markewitz et al., 1999). Prior histories of inflammatory- and/or allergy-related diseases (e.g. hay fever and eczema) are reported in several studies to be inversely associated with lung cancer risk (Schabath et al., 2005a; Turner et al., 2005; Talbot-Smith et al., 2003; Castaing et al., 2005). It is hypothesized that this association is due the “immunosurveillance hypothesis” by which a stimulated and/or enhanced immune system is more efficient in detecting and destroying malignant cells (Talbot-Smith et al., 2003). Alternatively, it could be argued that anti-inflammatory medications, may in part contribute to this protective effect. At present, there is no proven hypothesis for a role of respiratory diseases in lung carcinogenesis and the empirical evidence, which is not entirely consistent, has been largely derived from observational epidemiologic data.

DIETARY FACTORS Considerable evidence exists that diet may be a co-factor in lung cancer risk (Byers, 1994). Animal models have provided data that dietary fat can promote chemically induced pulmonary tumors (Kroes et al., 1986) and that low levels of retinol are associated with increased tumor yield (Sporn and Roberts, 1984). It is likely that lung cancer risk may be confounded by the association between smoking status and dietary intake (Subar et al., 1990), since smokers tend to have less “healthy” dietary habits (Hebert and Kabat, 1990). Data from the Second National Health and Nutrition Examination Survey showed that smokers have lower intakes of vitamin C, folate, fiber, and vitamin A than non-smokers, and intake tended to decrease as cigarette consumption increased, especially for vitamin C, fiber, and folate (Subaretal et al., 1990). A negative linear trend was found between smoking intensity and intake of several categories of fruits and vegetables (Subar et al., 1990). Furthermore, smokers influence the dietary habits of household nonsmokers (Trobs et al., 2002).

Fat Rates of lung cancer are highest in countries with greatest consumption of fat even after controlling for smoking (Carroll and Kohr, 1975; Wynder et al., 1987; Xie et al., 1991). In most case-control studies, elevated risks (up to threefold) have been reported (Goodman et al., 1988; Jain et al., 1990; Alavanja et al., 1993; de Stefani et al., 1997; Pillow et al., 1997; Mohr et al., 1999), although one large study was negative (Swanson et al., 1997). Cohort data have been conflicting. A recent pooled analysis of eight prospective studies (Smith-Warner et al., 2002) with over 3000 lung cancer cases, found that fat intake was not associated with risk. For 5% incremental energy from fat, the pooled risk was 1.01. Breslow et al. (2000) studied the association between lung cancer mortality and diet using 1987 National Health Survey Interview data, and reported that intake of red meats was positively associated with lung cancer mortality while intake of dairy products appeared to be protective. However, more recent results from cohort and serum micronutrient studies, that avoid the problems of inaccurate reporting of diet and recall bias, have been statistically insignificant (Koo, 1997).

Fruits and Vegetables Prior Respiratory Diseases A prior history of respiratory diseases such as asthma, bronchitis, emphysema, hay fever, and pneumonia may modify lung cancer risk (Talbot-Smith et al., 2003; Schabath et al., 2005a; Turner et al., 2005;

The most consistent association with diet has been the lowered risk attributed to consumption of fresh vegetables and fruits, noted in many case-control and cohort studies, and replicated in many different countries (Colditz, 1985). Risk in those with the highest intake tended to

Cancer of the Lung

643

be about one-half relative to the lowest intake group. A recent casecontrol analysis reported a significant protective effect for dietary phytoestrogens, especially in current and never smokers (Schabath, 2005b).

In a nested case-control study, Hartman et al. (2001) reported that serum vitamin B6 levels but not folate and vitamin B12 were significantly associated with decreased risk of lung cancer. Le Marchand et al. (1989) found no association with folate intake.

Carotenoids

Selenium

Several, but not all, biochemical studies have shown reduced levels of beta-carotene in stored sera from persons who subsequently developed lung cancer (Byers, 1994). The protective effects of foods high in carotenoids have been replicated across different populations in the United States, Europe, and Asia. A few studies also suggested that the protective effect was greater for current smokers and recent quitters than nonsmokers (Ziegler et al., 1984). Recently, Holick and coworkers (2002) investigated associations between dietary betacarotene, lutein/zeaxanthin, lycopene, beta-cryptoxanthin, vitamin A, serum beta-carotene, and serum retinol and lung cancer risk in the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study (ATBC) cohort of Finnish male smokers. They found consumption of fruits and vegetables was associated with a relative risk for lung cancer of 0.73 (0.62, 0.86) for the highest vs. the lowest quintile. Lower risks of lung cancer were similarly observed for the highest vs. lowest quintiles of lycopene (28%), lutein/zeaxanthin (17%), beta-cryptoxanthin (15%), total carotenoids (16%), serum beta-carotene (19%), and serum retinol (27%). Given these strong observational data, findings from several beta-carotene supplementation trials were unexpected, as discussed later in this chapter.

Selenium has a role in cellular defense against oxidative stress. Of over 100 studies investigating the effects of selenium in carcinogenexposed animals about two-thirds observed a reduction in tumor incidence and/or preneoplastic endpoints (Combs and Combs, 1986; Patterson and Levander, 1997). Early ecologic epidemiological data suggested that the levels of selenium in tobacco from countries with high lung cancer incidence appear to be significantly lower than those of tobacco from the low-incidence countries (Bogden et al., 1981). This could reflect geographic difference in selenium concentration in the soils, implying similar differences in intake from other crops consumed there. A Dutch cohort study found that the relative risk of lung cancer for subjects in the highest compared with the lowest quintile of toenail selenium, was 0.50, with a significant inverse trend across quintiles (P < 0.006) (van den Brandt et al., 1993). However, in the Nurses Health cohort, toenail selenium levels were not associated with subsequent lung cancer risk (Garland et al., 1995). In a Finnish study, the relative risk of lung cancer in the highest compared with the lowest tertiles of serum selenium was 0.41 (0.17–0.94) (Knekt et al., 1998; Hartman et al., 2002). Kabuto et al. (1994) reported that risks were elevated only among those in the lowest quartiles of serum selenium (OR = 1.8). Combs and colleagues (1997) generated considerable interest in selenium when they reported in a secondary analysis of an intervention trial that selenium supplementation reduced the incidence of lung cancer. In a recent reanalysis of these data with 3 additional years of follow-up, the relative risk was 0.70, but not statistically significant. In subgroup analysis, there was a significant decrease in relative risk among individuals receiving supplementation with lowest baseline serum selenium levels (Reid et al., 2002). Thus low selenium status may contribute to the risk of lung cancer.

Isothiocyanates Fruits and vegetables also contain other micronutrients, including vitamin C (that inhibits the formation of nitrosamines), phenols, flavones, vitamin E, selenium, and isothiocyanates (ITCs). ITCs are non-nutrient compounds in cruciferous vegetables, predominantly of the Brassica genus that are effective inhibitors of tumorigenesis in animal model systems (Wattenberg, 1987; Chung, 1992). One proposed mechanism is downregulation of cytochrome P450 biotransformation enzymes together with induction of phase II detoxifying enzymes (Yang et al., 1994; Zhang and Talalay, 1994). Well-designed cohort and case-control studies have demonstrated the protective effect of cruciferous vegetables, and against lung cancer specifically (Verhoeven et al., 1996). Hecht et al. (1995) showed that consumption of average portions of vegetables can result in the release of tens of milligrams of ITCs with inhibition of the oxidative metabolism of NNK. Epidemiologic studies examining the protective association within smoking strata have had mixed results. Gao et al. (1993) found a protective effect most apparent in current smokers (OR = 0.3), as did Spitz et al. (2000). Steinmetz et al. (1993) reported an effect only among ex-smokers (OR = 0.4), while Koo (1988) found no association in never-smoking women.

Folate Folate deficiency is associated with alterations in DNA methylation and synthesis and disruption of DNA repair activities (Mason et al., 1996; Choi et al., 2000; Wei et al., 2003). Epidemiologic studies have been inconsistent. To date, three cohort studies (Bandera et al., 1997; Speizer et al., 1999; Voorrips et al., 2000), and two case-control studies (Bandera et al., 1992; Le Marchand et al., 1989) have evaluated the association between dietary folate intake and lung cancer risk, and another two case-control studies have investigated the association between serum folate levels and lung cancer (Hartman et al., 2001; Jatoi et al., 2001). Bandera et al. (1997) reported that intake of folate was inversely associated with lung cancer risk especially in heavy smokers. More recently, a significant protective effect of dietary folate on lung cancer incidence was found in men in a Netherlands cohort (Voorrips et al., 2000), with a stronger effect for current than former smokers. Shen et al. (2003) reported significant protective effects for folate in former smokers. However, in the Nurses’ Health Study, no association of total folate intake was observed (Speizer et al., 1999).

Alcohol Studies of the association between alcohol consumption and lung cancer risk are complicated by smoking status. Woodson et al. (1999) evaluated this association in the ATBC study of male smokers, and concluded that alcohol consumption was not a risk factor for lung cancer nor were its effects modified by smoking history. On the other hand, data from three prospective studies in Denmark suggested that high consumption of beer and spirits in men was associated with increased risk, whereas wine intake appeared to be protective (Prescott et al., 1999). A recent meta-analysis of both cohort and case-control studies concluded that smoking explained the elevated relative risks in studies of alcoholics and that strong misclassification of smoking status resulted in elevated smoking adjusted risks in cohort and case-control studies (Bandera et al., 2001). Overall evidence of a smoking-adjusted association was limited to very high consumption groups in cohort and hospital-based case-control studies. At lower levels the associations are likely explained by confounding (Bandera et al., 2001). Korte et al. (2002) noted smoking-adjusted associations between alcohol and lung cancer in hospital-based, but not population-based case-control studies. They also suggested that smoking status explained elevated risks in alcoholics.

OCCUPATIONAL FACTORS The IARC (1988, 1990) has categorized several occupational agents as known carcinogens including arsenic, asbestos, bischloromethyl ether, chromium, nickel, polycyclic aromatic compounds, radon, and vinyl chloride. Probable carcinogens include acrylonitrile, beryllium,

644

PART IV: CANCER BY TISSUE OF ORIGIN

cadmium, formaldehyde, acetaldehyde, man-made fibers, silica, and welding fumes. Estimates from case-control studies of the proportion of lung cancers attributed to occupational exposures have wide ranges but most estimates include values from 9%–15%. The number of cases reflects past high exposure and is likely to drop in the future unless new workplace carcinogens are introduced. Detailed review of all work exposures carcinogenic to the lung is beyond the scope of this chapter.

Asbestos Asbestos exposure has posed the largest carcinogenic threat in the workplace. In the 1950s, Doll (1955) first reported that asbestos textile workers had a 10 times higher risk of lung cancer. The peak incidence occurred about 30–35 years after initial exposure (Selikoff, 1980). Morgan (1992) noted that mortality from lung cancer was increased in 40 of 47 cohort studies of male asbestos-exposed workers, with an overall weighted-average SMR of 1.65, and a dose-dependent risk (Hughes and Weill, 1991). Although risk increases as amount of exposure increases, it is not clear whether the dose-response relationship is linear or exponential (Blot and Fraumeni, 1996). A characteristic of asbestos-related lung cancer is its synergistic relation with cigarette smoking. In a combined analysis of six occupational groups, asbestos was shown to interact with smoking in a more than additive, and likely multiplicative manner (Saracci, 1987).

Radon Radon is a well-established lung carcinogen that arises from the decay chain of uranium-238, with a half-life of 3.8 days. Two of the shortlived progeny in the commonest decay chain are poloniun-2218 and polonium-2214 that decay by emitting genotoxic alpha particles. Radon is the principal source of exposure to ionizing radiation in most countries. IARC (1988) declared radon to be a group 1 carcinogen. It is estimated that residential radon is probably responsible for about 2000 lung cancer deaths per year in the United Kingdom, or 6.5% of the total lung cancer deaths in 1998 (Darby et al., 2001). Radon is a ubiquitous indoor air pollutant, although at substantially lower levels than for occupational exposures. Studies of radonexposed mine workers have shown increasing lung cancer risks with increasing cumulative exposure to radon (see Darby and Samet, 1994 for a review). Animal studies have also provided confirmatory evidence of the causal association between radon exposure and lung cancer risk. The Biological Effects of Ionizing Radiations (BEIR) committee provided evidence of a multiplicative interaction between smoking and radon exposure (NRC, 1999). However, subsequent data support the idea that radon and smoking act more than additively, but less than multiplicatively (Darby and Samet, 1994).

Chromium Chromium compounds are used in a variety of industrial processes. It was estimated that about 2 million US workers in 1977 were exposed to these compounds (IARC, 1990). Studies of workers exposed through chromate production, zinc chromate pigment, chromium electroplating, and to ferrochromium alloy products have all shown elevated risks (Coultas, 1994).

Nickel Highest risks have been noted for nickel exposure from mining, smelting, and refining. Reforms in industrial processing have led to significantly lower levels of exposures and thus decreased lung cancer risks. Results from 10 cohort analyses suggest that oxidic and sulfidic nickels are responsible for most of the respiratory cancers (Report of the International Committee on Nickel Carcinogenesis in Man, 1990).

Polycyclic Aromatic Hydrocarbons These compounds result from pyrolysis or incomplete combustion of organic matter. Although exposure to PAHs is prevalent, there are few data on exposure to specific compounds or on quantitative measures of exposure. PAHs are responsible for some of the excess risk of lung cancers reported in a range of occupational groups including coke, foundry, steel and iron plant workers, aluminum industry, roofers, asphalt workers, truck drivers and railroad workers (diesel exhaust), metal workers, printing press workers, and tool and die makers (reviewed in Coultas, 1994). Dioxin contamination of defoliants and herbicides is also implicated (Fingerhut et al., 1991).

Particulate Matter As outdoor air pollution worsened in the 20th century and the lung cancer rates began to rise steeply, a causal link was postulated. Of the roughly 3000 chemicals identified in ambient air, about 10% have been studied for carcinogenic potential and many have been found to be carcinogenic in animals (Lewtas, 1991). However, there are still many chemicals that have not been studied in humans and many studies of point source exposures are flawed (Speizer and Samet, 1994). The Cancer Prevention II study of the ACS recently reported that fine particulate and sulfur-oxide-related pollution, but not coarse and total suspended particles, were associated with lung cancer mortality (Pope et al., 2002). This effect was strongest for never smokers.

GENETIC FACTORS

Despite negative results from experimental studies attempting to induce tumors in laboratory animals with arsenic, epidemiologic evidence from diverse occupational settings indicates that inhaled arsenicals are carcinogenic to the lung. These occupations include nonferrous smelter workers (especially copper), miners, pesticide manufacturers, and a variety of other industrial uses of inorganic arsenic. Concern has also been expressed about arsenical air pollution near smelters and arsenic-contaminated drinking water. For a detailed review, see Blot and Fraumeni (1994).

The tobacco-cancer relationship is one of the best-explored paradigms of the concept that heritable traits modify the effects of carcinogenic exposures. As discussed above, only a fraction of long-term smokers will develop lung cancer. Bach et al. (2003a) used data on over 18,000 subjects enrolled in the Beta-Carotene and Retinol Efficacy Trial (CARET) to derive a cancer risk prediction model (akin to the Gail model for breast cancer) that included age, gender, asbestos exposure, and smoking history. They concluded that the one-quarter of smokers at highest risk will account for about half of the lung cancer cases (Bach et al., 2003a). Such interindividual variation in susceptibility could be attributed in part to common variants in genes in a variety of cellular pathways, including metabolic, DNA repair, cell cycle, inflammation, and the microenvironment (e.g., metallopeptidases).

Chloromethyl Ethers

Familial Aggregation

Workers exposed to bis chloromethyl ethers (BCME) used primarily as chemical intermediates in the production of organic chemicals, have high rates of lung cancer, with a reduction in risk after cessation of exposure (see Coultas, 1994 for a review). Commercially available chloromethyl ether is of lower carcinogenicity than BCME, but it is usually contaminated by BCME (Laskin et al., 1975).

Studies of familial aggregation provide indirect evidence for the role of genetic factors in lung cancer etiology. Tokuhata and Lilienfeld (1963) first reported an excess of lung cancer mortality in relatives of 270 lung cancer probands, with a more pronounced effect among nonsmokers. Subsequently, Ooi et al. (1986) and Shaw et al. (1991) reported similar findings. Schwartz et al. (1996) and Kreuzer et al.

Arsenic

Cancer of the Lung (1998) concluded that lung cancer in a first-degree relative was associated with an increased risk of lung cancer in young (<46 years of age) cases with no elevated risk in the older group, but not Etzelatel (2003). Similar patterns of aggregation among relatives of young onset probands were reported by Bromen et al. (2000). Bailey-Wilson et al. (2004) mapped a major susceptibility locus in multiplex lung cancer families to chromosome 6q23–25.

Metabolic Polymorphisms The internal dose of tobacco carcinogens to which lung tissue is exposed is modulated by polymorphisms in genes encoding for enzymes responsible for activation and detoxification of these carcinogens. These polymorphisms, although generally associated with low cancer risks, are frequent in the population, and therefore attributable risks are high.

Cytochrome P-450 Cytochrome P-450 (CYP) enzymes are known to metabolize many established carcinogens. Interindividual variation in P-450 gene expression attributed to polymorphisms in the genes encoding for the proteins, leads to variability in the extent to which carcinogens are metabolized, and to differences in cancer risk (Smith et al., 1995).

Cytochrome P4501A1 (CYP1A1). This gene, located on 15q22–q24, encodes the enzyme, aryl hydrocarbon hydroxylase (AHH), one of the monooxygenases that initiates a multienzyme pathway activating both PAHs and benzo[a]pyrene into highly electrophilic metabolites and also catabolizing arylamines. AHH activity is highly inducible, varies up to several thousand-fold among subjects, and is more inducible in lung cancer patients than in controls (Shields et al., 1993). Thirteen variants have been validated in the NCI SNP500 Database (http://snp500cancer.nci.nih.gov/), four of which have been investigated in case-control studies. The first is an A > G transition causing the substitution of isoleucine by valine at codon 462 in exon 7, thereby affecting the encoded protein’s function. In a Japanese study (247 cases and 185 controls) this homozygous polymorphism was associated with a threefold statistically significant risk of lung cancer, especially for SCC (OR = 4.85) and small cell carcinoma (OR = 9.35) (Sugimura et al., 1998). Similar findings were reported for a Chilean population (60 cases and 174 controls) (Quinones et al., 2001) and a Chinese population (217 cases and 404 controls) (Song N et al., 2001). In a prospective study in China, lung cancer risk was not associated with the Ile462Val genotype (London et al., 2000a). However, lighter smokers with at least one valine allele had marginally increased risk (OR = 1.72), which became statistically significant in association with homozygous deletion of GSTM1 (OR = 2.80). A similar finding was reported in the Chilean study (Quinones et al., 2001). The second polymorphism, Msp1-restriction fragment length polymorphism, is a T-to-C transition 264 bp downstream of the polyadenylation signal. The presence of at least one copy of the Msp1 variant allele was associated with a 2.4-fold increased risk of SCC in subjects of Caucasian, Japanese, or Hawaiian origin (Le Marchand et al., 1998) and Chinese (Song N et al., 2001), but not in a Finnish population (Hirvonen et al., 1992), nor in a smaller Chinese study (Yin et al., 2001). In African and Mexican Americans, these two polymorphisms were associated with increased risk among light smokers (Ishibe et al., 1997), and a similar non-significant trend was noted by Bennett et al. (1999a). A third polymorphism is an African-American specific allele (also a T > C) and increased risk was associated with this variant in two of three studies (Kelsey et al., 1994; London et al., 1995; Taioli et al., 1998). A recent meta-analysis of 15 case-control studies indicated a positive but not statistically significant association between two CYP1A1 polymorphisms (Ile462Val and Msp1) and lung cancer risk (Houlston, 2000). The ORs (non-significant) for one or two copies of the variant Msp1 allele were 1.09 and 1.27, respectively. For Ile462Val, the respective ORs were 1.16 and 1.62 (Houlston, 2000). Another review summarized 25 studies on lung cancer (Bartsch et al.,

645

2000) and a greater-than-additive risk was associated with the at-risk combined genotype of CYP1A1 and GSTM1. This gene-gene interaction between polymorphisms of CYP1A1 and GSTM1 appears to have a biological relevance, because benzo(a)pyrene diol epoxide (BPDE)DNA adduct levels in lung parenchyma were found to be more pronounced in persons with both variants of these two genes (Alexandrov et al., 2002).

Cytochrome P4501A2 (CYP1A2). This is a CYP1A1related family member on chromosome 15 (q22-qter), which encodes a protein involved in the metabolic activation of heterocyclic and aromatic amines, certain nitro-aromatic compounds, and some PAHs to carcinogenic intermediates. There is substantial interindividual variability in CYP1A2 activity (Eaton et al., 1995). Only four variants have been validated in the SNP500 Cancer Database. One polymorphism is a C > A point mutation in intron 1. The variant A allele is associated with high enzyme activity in individuals exposed to tobacco smoke (Nakajima et al., 1999). Sachse et al. (1999) further showed that smokers homozygous for the variant A allele had a 1.6-fold higher metabolic activity compared with other genotypes. Another polymorphism is in the 5¢-flanking region of the gene where G is replaced by A at position -2964 (Nakajima et al., 1999). This point mutation resulted in a significant decrease in enzymatic activity in Japanese smokers. No study has yet investigated the role of CYP1A2 as a risk factor in lung cancer. However, Landi et al. (1996) reported that subjects with rapid CYP1A2 activity showed highest 4-aminobiphenylhemoglobin adduct levels. Cytochrome P4502A6 (CYP2A6). The hepatic CYP2A6 enzyme is responsible for most of the conversion of nicotine to cotinine (Messina et al., 1997), through oxidation to nicotine iminium ion, the rate-limiting step. Both in vivo and in vitro studies have shown interindividual variation in CYP2A6 activity, attributed to variations in the gene, mapped to chromosome 19. Three alleles were originally identified—wild-type 6*1 and 2 defective alleles, *2 and *3. The former is a point mutation resulting in a leucine to histidine conversion in codon 160 that results in a null allele with no activity towards the probe substrate. There are multiple mutations in the *3 allele; however, it has proven to have a very low frequency. There are now over 16 alleles identified; however, only a small number of these are known to have functional consequences for nicotine metabolism (*2, *3, *4, *7, *10) (Benowitz et al., 2001; Nakajima et al., 2001; Xu et al., 2002a; Yoshida et al., 2002), and many have not been examined for their functional impact (*5, *6, *9, *11–16). Others are very rare (e.g., *5, *6, *11) (Oscarson et al., 1999; Nakajima et al., 2001; Yoshida et al., 2002) and it is unclear what their significance is. A recently identified allele, 6*12, has a reduced metabolic capacity (Oscarson et al., 2002). The significance of these alleles has been reviewed (Nakajima et al., 2002; Xu et al., 2002b). It is suggested that individuals with impaired nicotine metabolism due to a defective copy of CYP2A6 would be protected from becoming tobacco dependent, and would find nicotine exposure unpleasant, and would experience greater aversive effects than individuals with unimpaired nicotine metabolism. Such individuals therefore are less likely to become dependent smokers, and are more likely to smoke fewer cigarettes and quit smoking more easily. There are major ethnic differences in the frequencies of the variant alleles being very low in Caucasians, while some of the alleles are prevalent (up to 20%) in Asian populations. This enzyme also activates tobacco-specific nitrosamines. Cytochrome P4502D61 (CYP2D6). This gene is located on chromosome 22q13.1 and encodes a protein that localizes to the endoplasmic reticulum and is known to metabolize as many as 20% of commonly prescribed drugs. The gene is highly polymorphic and 38 variants have been validated in the SNP500 Cancer Database. There is a poor metabolizer phenotype, characterized by a decreased ability to metabolize the enzyme’s substrates (Gonzalez et al., 1988). Of nine independent studies on the polymorphisms (CYP2D6*3-10) reviewed by Bartsch et al. (2000), four studies reported significant results. However, two meta-analyses did not provide any evidence for an

646

PART IV: CANCER BY TISSUE OF ORIGIN

association with lung cancer risk (Christensen et al., 1997; RostamiHodjegan et al., 1998).

Cytochrome P4502C9 (CYP2C9). This gene is located on chromosome 10q24 and encodes a key enzyme with an oxidative role in the metabolism of benzo[a]pyrene (Shou et al., 1994). The enzyme function differs by allelic variant (Yasumori et al., 1991; Rees et al., 1993). CYP2C9*2 is a fairly common C/T polymorphism that produces an Arg/Cys amino acid substitution and appears to encode an enzyme with impaired metabolism (Rees et al., 1993; Rettie et al., 1994; Furuya et al., 1995). Few studies have examined the effect of this CYP2C9*2 polymorphism on lung cancer risk. One study reported no association in either Caucasians or African Americans (London et al., 1996). However, significantly decreased risk of lung cancer associated with the CYP2C9*2 allele has been reported in males, but not females (Wu et al., 1998a). This study also noted a trend for decreased BPDE-induced lymphocytic chromatid breaks with increasing numbers of CYP2C9*2 alleles. Cytochrome P4502E1 (CYP2E1). This gene, located on chromosome 10q24.3, encodes an enzyme involved in the metabolic activation of N-nitrosamines, and its expression is greatly enhanced by tobacco exposures (Guengerich et al., 1991; Koop, 1992). The inducibility of CYP2E1 by different genotypes remains unknown. CYP2E1 is somewhat induced in the liver by cigarette smoke, but its pulmonary expression is fourfold more inducible compared with CYP1A1, suggesting that CYP2E1 may actively participate in pulmonary carcinogenesis induced by cigarette smoke (Villard et al., 1994). The structure and function of CYP2E1 are highly evolutionarily conserved across species. One DraI CYP2E1 restriction fragment length polymorphism (RFLP) located in intron 6 was associated with lung cancer in a Japanese population (Uematsu et al., 1991), but a subsequent study in US Caucasians and African Americans was unable to corroborate this conclusion (Kato et al., 1994). Two other linked polymorphisms detectable with RsaI and PstI in the 5¢-flanking region of CYP2E1 have been identified and are associated with an effect on transcription levels (Hayashi et al., 1991). The association between CYP2E1 PstI or RsaI genotypes and lung cancer is controversial (Ingelman-Sundberg et al., 1994; Stephens et al., 1994). This discrepancy may be partly attributed to ethnic differences in the allelic distribution. In Finnish, Brazilian, and Japanese populations, the CYP2E1 polymorphism was not associated with lung cancer risk (Hirvonen et al., 1993a; Kato et al., 1994; Sugimura et al., 1995; Watanabe et al., 1995). A Swedish study reported that individuals with the CYP2E1 RsaI or PstI c2 allele were at lower risk for developing lung cancer (Persson et al., 1993). A US study demonstrated that individuals homozygous for the wild-type c1 CYP2E1 allele had an elevated lung cancer risk; however, this association was only present in men and in ever smokers (Wu et al., 1998b). Patients with the susceptible CYP2E1 genotype were diagnosed with lung cancer at earlier ages compared with those with the other genotypes. Wu et al. (1997) also reported that the c2 allele was a risk factor in Mexican Americans, but not African Americans. Myeloperoxidase Myeloperoxidase (MPO) is a phase I metabolic enzyme found in macrophages and neutrophil granules that is involved in the oxidation of procarcinogens (Mallett et al., 1991). As part of a mediated immune response, neutrophils are recruited to sites of pulmonary inflammation, and release MPO locally. There is a G > A polymorphism, located in a hormone response element region 463 bp upstream of the MPO gene transcription start site that is associated with a marked decrease in transcriptional activity, because of the disruption of an SP1-binding site in an Alu element (Piedrafita et al., 1996). Since overall transcriptional activity is decreased in individuals with the variant A allele, less enzyme would ultimately be available for conversion of the B[a]P intermediate to the highly carcinogenic BPDE. In fact, a number of case-control studies have demonstrated that the variant allele is associated with apparent 40%–70% reductions in lung cancer risk (London et al., 1997; Schabath et al., 2000; Le Marchand et al., 2000; Cascorbi

et al., 2000; Feyler et al., 2002). However, a nested case-control study in Finnish male smokers found no protective association (Misra et al., 2001), nor did a recent large case-control study (Xu LL et al., 2002) of 988 cases and 1100 controls. Thus no conclusions can be drawn regarding this polymorphism and lung cancer risk.

Microsomal Epoxide Hydrolase The Microsomal Epoxide Hydrolase (mEPHX) gene is located on chromosome 1q42.1 (Skoda et al., 1988). Four mEPHX alleles result from the presence or absence of two point mutations (Hassett et al., 1994). In the exon 3 polymorphism, tyrosine is replaced by histidine at residue 113; this variant is referred to as the “slow allele” because there is a marked decrease in enzyme activity. In the exon 4 polymorphism, arginine replaces histidine at residue 139 resulting in an increase in activity, the “fast allele”. The third allele has two mutations, one at residue 113 and the other at residue 139. The fourth allele is wild-type at both residues, with normal enzyme activity. mEPHX is involved in the first-pass metabolism of highly reactive epoxide intermediates. The enzyme is localized in the endoplasmic reticulum, and inactivates polycyclic hydrocarbon oxides and epoxides. However, if these inactivated compounds undergo further epoxidation, highly mutagenic and carcinogenic polycyclic hydrocarbon diol epoxides may result (Hirvonen, 1999). Levels of mEPHX activity in the lung and lymphocytes are normally 10- to 100-fold less than those found in the liver, kidney, and testis (Omiecinski et al., 1993; Oesch et al., 1977; Seidegard and Ekstrom, 1997). Existing evidence suggests that polymorphisms in exon 3 and exon 4 of mEPHX increase susceptibility to lung cancer. One study demonstrated a significantly higher frequency of homozygous exon 3 variant alleles in patients with lung cancer compared with controls (Smith and Harrison, 1997). Benhamou et al. (1998) did not find an association between lung cancer risk and the homozygous or heterozygous genotypes of the exon 3 and 4 variant alleles. However, they did find a statistically significant increased lung cancer risk when the data were analyzed according to genotypes that predict for intermediate or high mEPHX activity. Zhao et al. (2002) demonstrated an increased risk for exon 4 homozygotes.

NAD(P)H Quinone Oxidoreductase 1 (NQO1) NAD(P)H quinone oxidoreductase 1 (NQO1), formerly referred to as DT-diaphorase, is located on chromosome 16q22.1 and encodes for an enzyme that protects cells against oxidative stress, by catalyzing the 2-electron reduction of quinones to hydroquinones (Begleiter et al., 1997). NQO1 decreases benzo[a]pyrene 3,6-quinone-induced DNA adduct formation (Joseph and Jaiswal, 1994). A polymorphism at position 187 of exon 6 results in a proline to serine substitution, and reduced enzymatic activity. There are undetectable or minute levels of mutant NQO1 protein in lung tissue from individuals with the wild-type (T/T) (Traver et al., 1997). Thus individuals with the variant allele may be predisposed to an increased tobacco-related cancer risk due to altered activation of tobacco procarcinogens. However, the NQO1 pathway is also a major mechanism responsible for the toxicity of quinones, including those arising from benzo[a]pyrene. Under specific circumstances, metabolism by NQO1 yields more active products that produce reactive oxygen species or generate alkylating species. Therefore, the beneficial or harmful effects of NQO1 vary by different substrates. Rosvold et al. (1995) reported that the rare allele was twice as common in cases as controls. Other studies confirmed that patients carrying at least one NQO1 variant allele had increased risks of lung cancer (Lewis et al., 2001; Sunaga et al., 2002). On the other hand, there was a protective effect for the homozygous variant allele in Japanese in Hawaii (Chen et al., 1999) and African and Mexican Americans (Wiencke et al., 1997). Yin et al. (2001) reported no association in a Chinese population.

Glutathione S-transferase The Glutathione S-transferase (GST) supergene family of enzymes catalyzes the detoxification of a variety of exogenous carcinogens and

Cancer of the Lung

647

therapeutic compounds by glutathione conjugation. The detoxifying capability of GST isoforms may be influenced by gene polymorphisms (Seidegard et al., 1988, 1990; Hirvonen et al., 1993b; Nazar-Stewart et al., 1993; Pemble et al., 1994; Wiencke et al., 1995). There are five GST enzyme classes based on biochemical characteristics: alpha (GSTA), mu (GSTM), theta (GSTT), pi (GSTP), and omega (GSTO). Genetic polymorphisms have been identified for GSTM1, GSTP1, and GSTT1. The significance of these polymorphisms in lung cancer has been comprehensively reviewed (Reszka and Wasowicz, 2001) and are now being intensively evaluated in pharmacogenetic studies as well.

quency of the AA genotype for GSTP exon 5. An interaction has been shown between GSTM1/GSTM3 and GSTP exon 5 in lung cancer risk in smokers in some studies (Jourenkova-Mironova et al., 1998; Kihara and Noda, 1999), but not others (Ozawa et al., 1997; Harris et al., 1998; Saarikoski et al., 1998; To-Figueras et al., 1999; Yamamura et al., 2000). No relationship between GSTP1 exon 6 and lung cancer risk has been reported (Saarikoski et al., 1998; Harris et al., 1998). However, Wang et al. (1999) showed a fivefold increased risk in younger men with exon 6, but not exon 5 variant allele. A similar result was reported by Zhou et al. (2002).

GSTM1. This gene, on 1p13.3, encodes an enzyme that belongs to the mu class. There are three polymorphisms. One results in a deletion with lack of a functional gene product. The other two, GSTM1*A and GSTM1*B, differ by a lysine to arginine substitution and the two are typically categorized together as a single functional phenotype. The homozygous null genotype has been correlated with a higher risk for smoking-related cancers, and occurs in approximately 50% of Caucasians. A meta-analysis of 19 case-control genotype studies reported an overall lung cancer risk of 1.14 associated with the GSTM1 null genotype (Houlston, 1999). The combination of GSTM1 null and CYP1A1 variant alleles is associated with significantly increased risk (Kawajiri et al., 1995; Fryer and Jones, 1999; Nair and Bartsch, 2001).

Glutathione Peroxidase. Glutathione peroxidase (GPX) catalyzes the reduction of hydrogen peroxide, organic hydroperoxide, and lipid peroxides by reduced glutathione and protects cells against oxidative damage. There are at least four isoenzymes encoded by four different genes (GPX1, GPX2, GPX3, and GPX4) and their locations are 3p21.3, 14q24.1, 5q23, and 19p13.3, respectively. These genes (except for GPX4) are highly polymorphic. GPX2 is a seleniumdependent glutathione peroxidase. Normal lung tissue exhibits 7.0fold variation in GPX activity (Carmichael et al., 1988). No studies have reported on the roles of the phenotypes or genotypes in lung cancer risk. N-acetyltransferase Isozymes (NAT1 and NAT2)

GSTT1. This gene, located on chromosome 22q11.23, catalyzes the conjugation of reduced glutathione to a variety of electrophilic and hydrophobic compounds. The theta class includes GSTT1 and GSTT2, which share 55% amino acid sequence identity and both are thought to have a role in carcinogenesis. Since GSTT1 is involved in both detoxification and activation reactions, it is somewhat difficult to predict the biological consequences of the polymorphism. Significant differences in the prevalence of the null genotype exist among various ethnic groups, ranging from 10%–65% (Rebbeck, 1997). Early studies suggested that the GSTT1-null genotype conferred a small, non-statistically significant increased risk of cancer, whereas those with both GSTM1-null and GSTT1-null genotypes were at significantly higher risk of lung cancer (Hirvonen et al., 1994; el-Zein et al., 1997, Spitz et al., 2000). Subsequent studies showed that the GSTT1 null genotype was not associated with lung cancer risk in English (Deakin et al., 1996), Northwestern Mediterranean (To-Figueras et al., 1997), or French populations (Stucker et al., 2002) nor in a pooled analysis of 651 cases and 983 controls (Stucker et al., 2001). Since ITCs induce GSTs and serve as a substrate for GSTs, Spitz et al. (2000) studied diet-gene interactions and noted that low ITC intake and either the GSTM1 or GSTT1 null genotypes were associated with risks of 2.22 and 3.19, respectively, in current smokers. For both null genotypes, the OR was 5.45. These effects were not evident in former smokers. Similarly, in a cohort of Chinese men, individuals without detectable levels of a urinary biomarker of total isothiocyanates and who were GSTM1 or GSTT1 null, were shown to be at increased lung cancer risk (London et al., 2000b). Some of the inconsistencies reported in the role of GST genotypes in lung cancer risk could be due to confounding from dietary factors (Zhao et al., 2001).

GSTP. The Glutathione S-transferase P (GSTP) gene, located on 11q13 (Saarikoski et al., 1998), is the most abundant isoform in the lungs (Anttila et al., 1993). It metabolizes many carcinogenic and pharmacologic compounds (Board et al., 2000). Single nucleotide substitutions in GSTP exon 5 (Ile105Val) and/or exon 6 (Ala114Val) appear to change enzyme activity. Watson et al. (1998) reported that individuals with the 105 valine allele demonstrated consistently lower levels of GST activity; unlike those with the 114 valine substitution. Three variant cDNAs, GSTP1*A (105Ile/114Ala), GSTP1*B (105Val/114Ala), and GSTP1*C (105Val/114Val), were identified by Ali-Osman et al. (1997) and Mannervik et al. (1992). Ryberg et al. (1997) reported that the G-allele in exon 5 was associated with higher levels of smoking-related DNA adducts in lung tissue with a positive trend between the number of adducts and the number of G-alleles. They also showed that lung cancer patients had a significantly higher frequency of the GG genotype and a lower fre-

Both NAT1 and NAT2 and a pseudogene NATP are located on 8p22, a region that often displays LOH. NAT1 and NAT2 have about 87% nucleotide homology in the coding regions, and the enzymes metabolize a number of aromatic and heterocyclic amine carcinogens present in cigarette smoke with acetyl transfer. O-acetylation typically results in the activation of aromatic and heterocyclic amine carcinogens, while N-acetylation usually deactivates these carcinogens. Many alleles have been identified for both NAT1 and NAT2, and attempts have been made to achieve consensus of NAT nomenclature (Hein et al., 2000a). NAT1 derives its entire transcript from a single exon. Seven missense and four silent substitutions have been identified thus far in the NAT2 coding exon (Hein et al., 2000b). NAT2*4 is considered the wild-type allele. NAT2 activity is highest in the liver and gastrointestinal tract; NAT1 activity is expressed in many extrahepatic tissues. The importance of rapid versus slow acetylator genotype differs by organ sites. Results have been most consistent for bladder and colorectal cancers. Early lung cancer studies showed a slight overrepresentation of rapid acetylators (Roots et al., 1988; Phillip et al., 1988; Nyberg et al., 1998). Subsequently it was shown (Cascorbi et al., 1996) that highest risk was in smokers with the homozygous rapid acetylator NAT2*4/*4 genotype. Two studies (Martinez, 1995; Bouchardy et al., 1998) failed to document an association between the NAT2 genotype and lung cancer risk. Bouchardy et al. (1998) reported an association between the low activity NAT1 alleles (*14 and *15) with lung cancer, but this was not confirmed in another study (Blum et al., 1990).

Genes Involved in Methyl Metabolism DNA methylation regulates gene expression through differentially methylated CpG dinucleotides in promoter regions of genes. Aberrant DNA cytosine methylation facilitates gene mutation through deamination of 5-methyl cytosine to thymine (Esteller et al., 2001). De novo hypermethylation in promoter CpG islands has been identified as a possible mechanism for tumor suppressor gene inactivation in human cancer cells (Eng et al., 2000). Aberrant methylation is now being correlated with a variety of epidemiologic variables (e.g., diet or family).

Methylenetetrahydrofolate Reductase Methylenetetrahydrofolate reductase (MTHFR) is localized to 1p36.3 (Goyette et al., 1994) and encodes an enzyme participating in catalyzing the reduction of methylenetetrahydrofolate to methyltetrahydrofolate, a cofactor for homocysteine methylation to methionine. Reduced MTHFR activity induces DNA hypomethylation, and MTHFR may be a target of altered methylation status (Virmani et al., 2002). Three common MTHFR polymorphisms, C677T, A1298C, and

648

PART IV: CANCER BY TISSUE OF ORIGIN

G1793A have been identified. The nucleotide 677 polymorphism results in an alanine to valine (C > T) substitution at codon 222 (A222V) (Goyette et al., 1994), and is associated with about 30% of the in vitro enzyme activity of those with the 677CC wild-type genotype, whereas heterozygotes (677CT) have about 65% of normal enzyme activity (Frosst et al., 1995). About 15% of the population is homozygous 677TT for the variant, which is associated with higher plasma homocysteine and reduced plasma folate levels (Deloughery et al., 1996). The second common MTHFR polymorphism a glutamate to alanine, A1298C, resulting in at codon 429 (A429E), also influences homocysteine and folate levels but to a lesser extent (van der Put et al., 1998). The third polymorphism, G1793A, resulting in an amino acid substitution of arginine to glutamine at position 1793 at codon 594 (R594Q) is relatively rare with unknown functional relevance (Rady et al., 2002). The common 677T variant was associated with increased risk of esophageal (Song C et al., 2001) and stomach (Gao et al., 2002) cancers, but not lung cancer (Shen et al., 2001), nor was there evidence for an interaction between these two MTHFR polymorphisms and dietary folate or alcohol use. The role of the G1793A variant in lung cancer is unknown.

et al., 2003). Compared with the highest DRC quartile in the controls, suboptimal DRC was associated with adjusted risks for lung cancer of 1.5, 1.8, and 1.9 for the second, third, and fourth quartiles, respectively (Ptrend < 0.001). Younger cases (<60 years), females, lighter smokers, or those with a family history of cancer exhibit the lowest DRC and the highest lung cancer risk among their subgroups (Wei et al., 2000; Spitz et al., 2003).

DNA Methyltransferase 3B (DNMT3B)

DNA Repair Gene Transcript (mRNA) Levels

These genes are required for the establishment and maintenance of genomic methylation patterns (Bachman et al., 2001). DNMT3B is located on chromosome 20q11.2 (Xie et al., 1999). mRNA expression of the three DNMTs (DNMT1, DNMT3A, and DNMT3B) is upregulated in SCC, while only DNMT1 and DNMT3B are up-regulated in nonsquamous cell lung cancer. A C-to-T single-base transition in a novel promoter of DNMT3B increases promoter activity (Shen et al., 2002). CT heterozygotes had greater than twofold risk of lung cancer (OR = 2.13) and TT homozygotes an OR of 1.42 (Shen et al., 2002). No polymorphisms in DNMT1 and DNMT3A have been reported yet.

Cheng et al. (2000) measured the relative expression levels of five NER genes (ERCC1, XPB/ERCC3, XPG/ERCC5, CSB/ERCC6, and XPC) in stimulated peripheral lymphocytes. Lung cancer cases were significantly more likely than controls to have reduced expression levels of XPG/ERCC5 (OR = 2.32) and CSB/ERCC6 (OR = 2.49), with a significant trend between reduced expression levels and increasing risk.

DNA Repair Capacity Defective repair of genetic damage is an important determinant of susceptibility to lung cancer. There is a substantial interindividual variation in DNA repair capacity within a given population with xeroderma pigmentosum or Bloom syndrome patients representing the extreme end of the repair spectrum (Hoeijmakers, 1994). These patients have defects in DNA repair capacity, increased frequency of chromosomal instability, hypersensitivity to carcinogenic exposure, and substantially increased risks of cancer (Hsu et al., 1991). An extensive review of the published literature concluded that most studies demonstrate that cancer cases have a significant decrease in DNA repair capacity compared with controls (Berwick and Vineis, 2000).

Host-Cell Reactivation Assay While there are many assays that measure the efficiency of multiple steps of excision repair individually (Athas et al., 1991), the ability to test the whole pathway is needed for population studies. Therefore, measuring the expression level of damaged reporter genes (host cell reactivation assay) is the assay of choice. This assay uses undamaged cells, is relatively fast, and is an objective way of measuring repair (Athas et al., 1991). In the assay, a BPDE-damaged non-replicating recombinant plasmid (pCMVcat) harboring a chloramphenicol acetyltransferase (CAT) reporter gene is introduced by transfection into lymphocytes. The presence of only one such unrepaired DNA lesion can block the transcription of an essential gene (Koch et al., 1993); therefore reactivated CAT enzyme activity is measured as a function of excision repair of the damaged bacterial gene (Athas et al., 1991). Both lymphocytes and skin fibroblasts from patients with basal cell carcinoma but not XP have lower excision-repair rates than individuals without cancer (Wei et al., 1993). Consequently, the repair capacity of lymphocytes can be considered a reflection of an individual’s overall repair capacity. A pilot study and a subsequently larger study of 764 lung cancer patients and 677 controls, demonstrated statistically significantly lower DRC in cases compared with controls (Wei et al., 1996a; Spitz

32

P-Postlabeling Assay of DNA Adducts

Tobacco-induced DNA adducts can be detected by a very sensitive 32Ppostlabeling assay (Reddy et al., 1986). There is a linear relationship between the levels of aromatic DNA adducts in human lung and the number of cigarettes smoked (Phillips et al., 1986). However, a relatively large variation is observed in the level of persistent DNA adducts in vivo (Everson et al., 1986). An in vitro induction assay of carcinogen-DNA adducts has been developed by treating stimulated lymphocytes with a relatively large dose (4 mM) of BPDE (Li et al., 1996). The levels of induced DNA adducts were significantly higher in lung cancer cases than in controls, with never- smoking and younger cases having the highest levels (Li et al., 2001).

Mutagen Sensitivity The mutagen sensitivity assay measures constitutional capacity to accomplish DNA repair by quantifying the chromatid breaks induced by in vitro exposure to a mutagen in short-term cultures of peripheral lymphocytes, after time for recovery (Hsu et al., 1991; Wu et al., 1998c). Challenge mutagens include BPDE and bleomycin. Bleomycin is radiomimetic and generates free oxygen radicals that induce single- and double-stranded breaks repaired by the base excision and/or double-strand break repair pathway(s). Numerous compounds in cigarette smoke cause oxidative damage; therefore, bleomycin sensitivity is relevant to the study of tobacco sensitivity. Several case-control studies have shown that mutagen sensitivity is a significant independent predictor of lung cancer risk (Spitz et al., 1995; Strom et al., 1995; Wu et al., 1995; Wu et al., 1998c, Wei et al., 1996b), with up to a fourfold increased risk (Wu et al., 1995). Higher risk estimates were noted for lighter smokers and younger patients, supporting the notion that mutagen sensitivity may constitute a susceptible phenotype. These findings have recently been confirmed by other investigators (Zheng et al., 2003). Increasing risk is associated with increased numbers of induced breaks per cell. An interaction between mutagen sensitivity and the CYP2E1 c1/c1 genotype has been demonstrated (Wu et al., 1997). BPDE sensitivity is also associated with a significantly elevated risk for lung cancer (Wu et al., 1998c; Wei et al., 1996c) and recently updated in 406 cases and 411 controls, with twofold higher risk in the highest quartile compared with the lowest. Mutagen-induced chromosome damage is not randomly distributed but may reflect the susceptibility of specific loci to damage by carcinogens (Wu et al., 1995). The short arm of chromosome 3 is a frequently identified genetic aberration in lung cancer (Whang-Peng et al., 1982), as well as in bronchial dysplasia (Hung 1995). Wu et al. (1995) reported that lymphocyte BPDE-induced 3p21.3 aberrations were significantly higher in cases than in controls (P < 0.0001).

Polymorphisms in DNA Repair Genes Polymorphisms in DNA repair genes may contribute to interindividual variation in DNA repair capacity, although subtle differences in DNA repair capacity due to a single polymorphism of a single gene in a very complex pathway are difficult to detect (Spitz et al., 2001; Qiao et al., 2002). Recently, the entire coding regions of several DNA

Cancer of the Lung repair genes on chromosome 19 (i.e., three nucleotide excision repair (NER) genes (ERCC1, XPD/ERCC2, and XPF/ERCC4) ), one doublestrand break (DSB) repair gene (XRCC3), and one base excision repair (BER) gene (XRCC1), were re-sequenced in 12 normal individuals (Shen et al., 1998) and later expanded to 37 genes in up to 164 individuals (Mohrenweiser et al., 2002). Although this study discovered 127 amino acid substitution variants, the significance of these variants is largely unknown; the implication is that variants that cause amino acid substitutions may impact protein function. Variants that do not cause amino acid changes or are located in introns may impact repair by causing splicing aberration, mRNA instability, or in linkage disequilibrium with other disease-causing genes. The NER pathway is important in repairing bulky DNA adducts induced by polycyclic aromatic hydrocarbons in tobacco.

XPA. XPA, one of seven genetic complementation groups that encode for proteins involved in the NER pathway, is located at 9q22.3. Mutations in the functional domains of the XPA gene could partially or completely inactivate its function (Cleaver and States, 1997; Miyamoto et al., 1992; Li et al., 1995), and thus DNA repair capacity. One polymorphism in the 5¢ non-coding region results in a single nucleotide substitution (A > G) that does not affect the amino acid sequence of the gene product (Butkiewicz et al., 2000). The 5¢noncoding region may regulate gene expression through transcriptional and post-transcriptional control mechanisms (Akiri et al., 1998; Larsen et al., 2002). An association between the XPA A/G polymorphism and lung cancer risk has been shown in a Korean population (Park et al., 2002a), with the G variant allele having a protective effect, especially in smokers (Wu et al., 2003) and being predictive of the more efficient repair phenotype. XPD. XPD (previously known as excision repair cross complementing group 2) is considered an essential gene (Taylor et al., 1997). Aside from its role in NER, it also serves a function in basal transcription as an evolutionary conserved ATP-dependent helicase within the multi subunit transcription factor, TFIIH, which is necessary for transcription by RNA polymerase II. The result of conservative mutations in XPD is subtle because helicase activity is not affected. However, multiple alterations in this gene may lead to obvious changes in phenotype (Wang et al., 1996). The XPD gene is highly conserved in eukaryotes, with homology to Rad3 and Rad15. Its chromosomal location is 19q13.2–q13.3. Two XPD polymorphisms have been studied with regard to lung cancer risk with somewhat inconsistent observations. There is a single G/A nucleotide polymorphism that codes for an amino acid substitution (Asp/Asn) at codon 312 in exon 10. Another polymorphism is an amino acid change (Lys/Gln) due to a nucleotide substitution (A > C) in exon 23. Though this particular polymorphism does not reside in a known helicase/ATPase domain and its functional relevance remains to be determined (Shen et al., 1998), it is located at an amino acid residue that is highly conserved throughout human, mouse, hamster, and fish XPD, suggesting a functional relevance. Increased lung cancer risk and suboptimal DRC in those with the variant alleles have been reported in one study (Spitz et al., 2001), but not another, in which the wild-type genotype of Lys/Lys 751 was associated with increased risk of suboptimal DRC (Lunn et al., 2000). One study reported no relationship with DRC (Moller et al., 1998). Excision Repair Cross Complementation Group 1 (ERCC1). The ERCC1 gene, located on 19q 13.2–13.3 (Mohrenweiser et al., 1989), codes for a 5¢ incision subunit of the NER complex. The ERCC1 and XPF proteins form a stable complex in vivo and in vitro (de Laat et al., 1999). Although no defect in ERCC1 has been reported in humans, cells from ERCC1-deficient mice have increased genomic instability and a repair-deficient phenotype (Melton et al., 1998). There are five known polymorphisms of ERCC1 gene that are validated in the NCI SNP Database, but none cause an aminoacid change. However, an A > C polymorphism at nucleotide 8092, in the 3¢ untranslated region, is thought to affect mRNA stability (Shen et al., 1998). Zienolddiny et al. (2005) reported that two polymor-

649

phisms in the ERCC1 gene were associated with increased risk of nonsmall cell lung cancer, although data from Shen et al. (2005) showed no such pattern.

X-Ray Repair Cross Complementation Group 1 (XRCC1). This gene, on 19q13.2, encodes a critical enzyme participating in the basal excision repair pathway (Mohrenweiser et al., 1989). A total of eight known variants have been validated, four of which cause an amino acid change. Although the functional relevance of these variants is unknown, codon 399 is within the XRCC1 BRCT domain (codons 314–402) (Zhang et al., 1998), which is highly homologous to BRCA1 (Masson et al., 1998). The R194W polymorphism is less likely to cause a significant change in the repair function. Elevated lung cancer risk associated with the 399R homozygous genotype has been detected in non-Hispanic whites (Divine et al., 2001), Koreans (Park et al., 2002b), and Chinese (Chen et al., 2001). In a cohort of Chinese tin miners, the R280H polymorphism was associated with increased lung cancer risk (OR = 1.8), and the R194W polymorphism with lower risk (Ratnasinghe et al., 2001). This finding was confirmed in Caucasian (Butkiewicz et al., 2001) and AfricanAmerican populations (OR = 0.2), but no increased risk was noted for the codon 399 variant allele (David-Beabes et al., 2001a).

X-Ray Repair Cross Complementation Group 3 (XRCC3). This gene, located at 14q32.3, encodes a protein that repairs DNA double-strand breaks. A single C/T polymorphism at position 241 in exon 7 results in an amino acid substitution (Thr/Met); however, its association with lung cancer is uncertain. David-Beabes et al. (2001b) found no significant association between the XRCC3 polymorphism and lung cancer. However, Wang et al., 2003 reported an elevated, but not statistically significant risk of lung cancer associated with the XRCC3 T-allele in African Americans and Mexican Americans, evident largely in heavy smokers. The distribution of the T-allele in African and Mexican Americans is lower (22%) than in Caucasians (>30%) (David-Beabes et al., 2001b).

hOGG1/hMMH. The hOGG1 gene, located on 3p25 is homologous to the yeast repair gene OGG1/MMH. It encodes, proteins responsible for the excision of 8-oxoguanine, a mutagenic base byproduct. Three variants have been validated in the NCI SNP Database, but only one causes an amino acid change (S326C) due to a C > G transversion. There are both somatic and polymorphic mutations and loss of heterozygosity (LOH) of the hOGG1 gene in lung tumors (Chevillard et al., 1998; Lu et al., 1997; Wikman et al., 2000). To date, five studies have investigated the role of the hOGG S326C polymorphism in lung cancer. This common polymorphism leads to hOGG1-Ser326 and hOGG1-Cys326 proteins and activity in the repair of 8-hydroxyguanine appears to be greater with the Ser326 protein than with the Cys326 protein (Kohno et al., 1998). Sugimura et al. (1999) found that Japanese with the hOGG1 Cys326Cys genotype were at an increased risk of SCC lung cancer, compared with those with the Ser326Cys or Ser326Ser genotypes (ORs = 3.0 and 2.2, respectively) but these risk estimates were not substantiated in a recent Japanese study (Ito et al., 2002). The risk appeared to be more pronounced in Hawaiian (OR = 3.6) than in Japanese (OR = 2.0) and Caucasians (OR = 1.6) (Le Marchand et al., 2002). A polymorphic allele 3 in hMMH/OGG1 exon 1 has been significantly related to risk of AC (OR = 3.2) among Japanese (Ishida et al., 1999). Cell Cycle Control Cell-cycle checkpoints are biochemical signaling pathways that sense DNA damage and elicit a complex cellular response that leads to cellcycle delay and activation of DNA repair mechanisms. This slowing or arrest of cell-cycle progression is an important characteristic of the cellular response to DNA damage. A recent in vitro functional assay showed that shorter durations of the S and G2 cell cycle phases after exposure to g-radiation were associated with increased lung cancer risk (Wu et al., 2005).

650

PART IV: CANCER BY TISSUE OF ORIGIN

p53 The p53 gene is located on 17p13.1. p53 missense mutations are the most frequent aberration in lung cancer (Bennett et al., 1999b). Three polymorphisms have been shown to affect the risk of lung cancer: a 16 base pair insertion in intron 3, a G > C transversion in exon 4 that codes for proline instead of arginine at residue 72, and a G-to-A transition at intron 6. Thomas et al. (1999) reported that the exon 4 Pro variant was less efficient in suppressing cell transformation and slower in inducing apoptosis than the wild-type. Several studies (Weston et al., 1992; Murata et al., 1996; Biros et al., 2001) concluded that there was no association between codon 72 genotypes and lung cancer. One study found that the Pro/Pro and Pro/Arg genotypes were associated with an increased risk of AC in Caucasians (Fan et al., 2000). This finding was previously noted especially in lighter smokers and younger onset patients (Jin et al., 1995), and a gene-dose effect for exon 4 M-allele was reported (Wu et al., 2002). Studies of introns 3 and 6 polymorphisms have yielded conflicting results. Birgander et al. (1995) found no association with lung cancer risk. However, Biros et al. (2001) reported a higher percentage of the intron 6 variant allele in lung cancer patients, and Wu et al. (2002) showed increased risks associated with both intronic polymorphisms.

Screening Study (LSS), assessed the feasibility of enrolling more than 3000 participants in a randomized CT trial and showed that at-risk subjects are willing to accept randomization and that the background rate of spiral CT was less than 2%. The second phase involves a randomized controlled trial at 30 centers of 50,000 individuals at high risk of developing lung cancer to see whether screening with low-dose helical CT can reduce lung cancer-specific mortality relative to chest radiographs. High risk is defined by age 55–74 years with a current or previous smoking history of at least 30 pack years; former smokers must have quit within the preceding 15 years. The experimental group will undergo screening with low-dose helical CT. The control group will undergo screening with chest radiographs. The study is powered to detect a 20% difference in lung cancer mortality. Marshall et al. (2001) assessed the potential cost effectiveness of a one-time helical CT scan using the experience of the ELCAP trial and concluded it was cost effective in a very high-risk cohort. However, a recent decision and cost-effectiveness analysis of helical CT screening by Mahadevia et al. (2003) concluded that even if the screening would truly prove to be efficacious it appears to be too expensive in most situations. Many of the assumptions used for such analyses will have to be reassessed at the end of the trial.

Chemoprevention PREVENTIVE MEASURES Screening There have been five randomized controlled trials assessing sputum cytology and chest X-rays that were recently reviewed by Bach et al. (2003b). Several trials, although differing in experimental design, concluded that screening did not reduce lung cancer-specific mortality. The Johns Hopkins and Memorial Sloan-Kettering trials (Flehinger et al., 1984; Tockman, 1986) were designed to address the incremental benefit of sputum cytology analysis rather than chest radiographs per se. Both tests were able to detect presymptomatic early lesions, and rates of resectability and survival were higher in the intervention groups, but this did not translate into reduced lung cancer mortality, the designated endpoint of the trial. The Mayo Lung Project focussed on the combined impact of cytology and chest X-ray compared with recommendation for annual screening (Fontana et al., 1986). Both this study and a Czechoslovakian trial showed advantages to the screened groups with respect to earlier stage at diagnosis, resectability, and survival (Fontana et al., 1986, 1991; Kubik and Polak, 1986). However, because the experimental groups in both studies demonstrated increases in cumulative lung cancer incidence above the control groups (p = 0.019), improvements in case fatality did not translate into significant reductions in lung cancer mortality. Although these studies had shortcomings (Marcus, 2001), they have proved to be central in shaping national lung cancer screening policy since then. A subsequent analysis to extend follow-up time through the end of 1996, documented lung cancer mortality rates of 4.4 and 3.9 per 1000 person-years for the treatment and controls arm, respectively (Marcus et al., 2000). This was attributed to over-diagnosis bias. The Prostate, Lung, Colorectal and Ovarian (PLCO) Screening Trial is also evaluating annual chest X-ray in a low-risk population. Low-dose spiral CT scanning that allows a low-resolution image of the entire thorax with low radiation exposure is engendering considerable enthusiasm. A non-randomized historical comparison of two screening strategies in Japan suggested that CT scans increased the ability to diagnose lung cancer at earlier stages (Kaneko et al., 1996; Kakinuma et al., 1999). Data from the observational Early Lung Cancer Action Project (ELCAP) documented the ability of low-dose spiral CT to detect noncalcified nodules and malignancies in the lung, of which 85% were stage I (Henschke et al., 1999). This was not a randomized controlled trial and thus could not provide information on the effect of screening in reducing lung cancer mortality. All studies had high false-positive results. For this reason another national screening trial, the National Lung Cancer Screening Trial (NLST), has been launched. Phase 1, the Lung

Several phase IIb retinoid chemoprevention trials in the early 1990s using metaplasia (Lee et al., 1994) and sputum atypia as an endpoint (Arnold et al., 1992; McLarty et al., 1995) were all negative as was a trial of the synthetic retinoid, fenretinide (4 HPR) (Kurie et al., 2000). More recently two studies using the organosulfur compound anethole dithiolethione (ADT) and 9 cis retinoic acid (in former smokers) provided some encouraging results (Lam et al., 2002; Kurie et al., 2003). Several large-scale phase III trials including the Beta-Carotene and Retinol Efficacy Trial (CARET), the Alpha-Tocopherol, BetaCarotene (ATBC) Prevention Trial, and the Intergroup Lung Study have demonstrated that retinoid treatment may actually increase lung cancer incidence in individuals who smoke (The ATBC Study Group, 1994; Omenn et al., 1996; Lippman et al., 2001). In contrast, retinoid treatment reduced recurrence and mortality in nonsmokers, and there was evidence of benefit (not statistically significant) in mortality in former smokers (Lippman et al., 2001). Neither the Physician’s Health Study (Cook et al., 2000) nor the Women’s Health Study (Lee et al., 1999) showed any difference in lung cancer incidence by study arm, although this was a secondary endpoint. Debate has focused on dosage, duration of the trials, and the isomeric form used (Greenwald, 2003). Recent preclinical data provide biologic plausibility for the adverse interaction of cigarette smoke and b-carotene that occurred in the ATBC and CARET studies. Antioxidant activity of carotenoids can cause oxidative stress through pro-oxidant activity, depending on redox potential and the host environment (Palozza, 1998), and therefore oxidative metabolites of b-carotene may inhibit retinoid signaling and tumorigenesis. As Greenwald (2003) points out, the next generation of clinical trials will benefit from in vivo and in vitro mechanistic studies, development of animal models, and a better characterization of the risk profile of the populations to be enrolled. The Selenium and Vitamin E Cancer Prevention Trial (SELECT) will examine the effects of selenium and vitamin E on lung cancer as a secondary endpoint in men.

FUTURE DIRECTIONS The ability to identify smokers with the highest risks of developing cancer has substantial implications for early detection and chemoprevention. Increased emphasis must be placed on correlating biomarker data derived from surrogate tissues (peripheral lymphocytes, serum) with molecular changes in the target tissue (lung) as well as intermediate tissue (bronchial brushings or lavage). Surrogate markers most representative of molecular changes in the target tissue can be used to develop quantitative risk assessment models for lung cancer. These markers require readily accessible samples (blood), are

Cancer of the Lung amenable to high throughput analysis and provide an opportunity for non-invasive evaluation of risk, physiologic and pathophysiological states, and as an adjunct to new screening modalities. Identification of protein patterns in serum using high-throughput proteomics linked to novel bioinformatics approaches is providing exciting data suggesting that predisposition to disease, early diagnosis, and evaluation of response to therapy can be performed on sera. Tumor DNA can be isolated from serum or plasma, as a source for screening specific transcripts or mutations in mitochondria or nuclear DNA sequences with a potential role in early detection. Linking of genome-wide polymorphism analysis, DNA copy number, epigenetic changes, transcriptional profiling, and proteomics provide powerful new approaches that cannot be successfully accomplished by any discipline independently. Pharmacogenetic profiles can be built to individualize therapy and to understand the functional consequences of chemoprevention, chemotherapy, or radiotherapy response. For rapid translation of these emerging new technologies, there will be a growing emphasis on molecular epidemiology and the need for well-characterized epidemiologic data and specimen repositories. References Akiri G, Nahari D, Finkelstein Y, Le SY, Elroy-Stein O, Levi BZ. 1998. Regulation of vascular endothelial growth factor (VEGF) expression is mediated by internal initiation of translation and alternative initiation of transcription. Oncogene 17:226–236. Alavanja MCR, Brown CC, Swanson C, Brownson RC. 1993. Saturated fat intake and lung cancer risk among nonsmoking women in Missouri. J Natl Cancer Inst 85:1906–1916. Alberg AJ, Samet JM. 2003. Epidemiology of lung cancer. Chest 123:21S– 49S. Alexandrov K, Cascorbi I, Rojas M, Bouvier G, Kriek E, Bartsch H. 2002. CYP1A1 and GSTM1 genotypes affect benzo[a]pyrene DNA adducts in smokers’ lung: Comparison with aromatic/hydrophobic adduct formation. Carcinogenesis 23:1969–1977. Ali-Osman F, Akande O, Antoun G, Mao JX, Buolamwini J. 1997. Molecular cloning, characterization and expression in Escherichia coli of full-length cDNAs of three human glutathione S-transferase Pi gene variants. Evidence for differential catalytic activity of the encoded proteins. J Biol Chem 272:10004–10012. American Cancer Society. 2005. Cancer Facts and Figures. Atlanta, GA: American Cancer Society, Inc. Anttila S, Hirvonen A, Vainio H, Husgafvel-Pursiainen K, Hayes JD, Ketterer B. 1993. Immunohistochemical localization of glutathione S-transferases in human lung cancer. Cancer Res 53:5643–5648. Arnold AM, Browman GP, Levine MN, et al. 1992. The effect of the synthetic retinoid etretinate on sputum cytology: Results from a randomised trial. Br J Cancer 65:7373–7343. Athas WF, Hedayati M, Matanoski GM, Farmer ER, Grossman L. 1991. Development and field-test validation of an assay for DNA repair in circulating lymphocytes. Cancer Res 51:5786–5793. Bach PB, Kattan MW, Thornquist MD, et al. 2003a. Variation in lung cancer risk among smokers. J Natl Cancer Inst 95:470–478. Bach PB, Kelley MJ, Tate RC, McCrory DC. 2003b. Screening for lung cancer: A review of the current literature. Chest 123:72S–82S. Bachman KE, Rountree MR, Baylin SB. 2001. Dnmt3a and Dnmt3b are transcriptional repressors that exhibit unique localization properties to heterochromatin. J Biol Chem 276:32282–32287. Bailey-Wilson JE, Amos CI*, Pinney SM, Petersen GM, de Andrade M, Wiest JS, Fain P, Schwartz AG, You M, Franklin W, Klein C, Gazdar A, Rothschild H, Mandal D, Coons T, Slusser J, Lee J, Gaba C, Kupert E, Perez A, Zhou X, Zeng D, Liu Q, Zhang Q, Seminara D, Minna J, Anderson MW. 2004. A major lung cancer susceptibility locus maps to chromosome 6q23–25. Am J Hum Genet 75:460–474. *equal contribution as first author. Bandera EV, Freudenheim JL, Graham S, et al. 1992. Alcohol consumption and lung cancer in white males. Cancer Causes Control 3:361–369. Bandera EV, Freudenheim JL, Marshall JR, et al. 1997. Diet and alcohol consumption and lung cancer risk in the New York State Cohort. Cancer Causes Control 8:828–840. Bandera EV, Freudenheim J, Vena JE. 2001. Alcohol consumption and lung cancer. Cancer Epidemiol Biomarkers Prev 10:813–821. Bartsch H, Nair U, Risch A, Rojas M, Wikman H, Alexandrov K. 2000. Genetic polymorphism of CYP genes, alone or in combination, as a risk modifier of tobacco-related cancers. Cancer Epidemiol Biomarkers Prev 9:3–28.

651

Begleiter A, Leith MK, Curphey TJ, Doherty GP. 1997. Induction of DTdiaphorase in cancer chemoprevention and chemotherapy. Oncol Res:371–382. Benhamou S, Reinikainen M, Bouchardy C, Dayer P, Hirvonen A. 1998. Association between lung cancer and microsomal epoxide hydrolase genotypes. Cancer Res 58:5291–5293. Bennett WP, Alavanja MC, Blomeke B, et al. 1999a. Environmental tobacco smoke, genetic susceptibility, and risk of lung cancer in never-smoking women. J Natl Cancer Inst 91:2009–2014. Bennett WP, Hussain SP, Vahakangas KH, Khan MA, Shields PG, Harris CC. 1999b. Molecular epidemiology of human cancer risk: Gene-environment interactions and p53 mutation spectrum in human lung cancer. J Pathol 187:8–18. Benowitz NL, Griffin C, Tyndale R. 2001. Deficient C-oxidation of nicotine continued. Clin Pharmacol Ther 70:567. Berwick M, Vineis P. 2000. Markers of DNA repair and susceptibility to cancer in humans: an epidemiologic review. J Natl Cancer Inst 92:874–897. Birgander R, Sjalander A, Rannug A, et al. 1995. P53 polymorphisms and haplotypes in lung cancer. Carcinogenesis 16:2233–2236. Biros E, Kalina I, Kohut A, Stubna J, Salagovic J. 2001. Germ line polymorphisms of the tumor suppressor gene p53 and lung cancer. Lung Cancer 31:157–162. Blot WJ, Fraumeni JF. 1994. Arsenic and lung cancer. In: Samet JM, ed. Epidemiology of Lung Cancer. New York: Marcel Dekker Inc, pp. 207–218. Blot WJ, Fraumeni JF. 1996. Cancers of the lung and pleura. In: Schottenfeld D, Fraumeni JF, eds. Cancer Epidemiology and Prevention. New York: Oxford University Press. Blum M, Grant DM, McBride W, Heim M, Meyer UA. 1990. Human arylamine N-acetyltransferase genes: Isolation, chromosomal localization, and functional experession. DNA Cell Biol 9:193–203. Board PG, Coggan M, Chelvanayagam G, Eeasteal S, et al. 2000. Identification, characterization and crystal structure of the Omega class glutathione transferases. J Biol Chem 275:24798–24806. Boffetta P, Tredaniel J, Grecco A. 2000. Risk of childhood cancer and adult lung cancer after childhood exposure to passive smoke: A meta-analysis. Environ Health Perspect 108:73–82. Bogden JD, Kemp FW, Buse M, et al. 1981. Composition of tobaccos from countries with high and low incidences of lung cancer. I. Selenium, polonium-210, Alternaria, tar, and nicotine. J Natl Cancer Inst 66:27–31. Bouchardy C, Mitrunen K, Wikman H, et al. 1998. N-Acetyltransferase NAT1 and NAT2 genotypes and lung cancer risk. Pharmacogenetics 8:291–298. Breslow RA, Graubard BI, Sinha R, Subar AF. 2000. Diet and lung cancer mortality: A 1987 National Health Interview Survey cohort study. Cancer Causes Control 11:419–413. Bromen K, Pohlabeln H, Ingeborg J, Ahrens W, Jockel K-H. 2000. Aggregation of lung cancer in families: Results from a population-based casecontrol study in Germany. Am J Epidemiol 152:497–505. Butkiewicz D, Rusin M, Harris CC, Chorazy M. 2000. Identification of four single nucleotide polymorphisms in DNA repair genes: XPA and XPB (ERCC3) in Polish population. Hum Mutat 15:577–578. Butkiewicz D, Rusin M, Enewold L, Shields PG, Chorazy M, Harris CC. 2001. Genetic polymorphisms in DNA repair genes and risk of lung cancer. Carcinogenesis 22:593–597. Byers T. 1994. Diet as a factor in the etiology and prevention of lung cancer. In: Samet JM, ed. Epidemiology of Lung Cancer. New York: Marcel Dekker Inc, pp. 335–348. Carmichael J, Forrester LM, Lewis AD, Hayes JD, Hayes PC, Wolf CR. 1988. Glutathione S-transferase isoenzymes and glutathione peroxidase activity in normal and tumour samples from human lung. Carcinogenesis 9:1617–1621. Carroll KK, Khor HT. 1975. Dietary fat in relation to tumorigenesis. Prog Biochem Pharmacol 10:308–353. Cascorbi I, Brockmoller J, Mrozikiewicz PM, Bauer S, Loddenkemper R, Roots I. 1996. Homozygous rapid arylamine N-acetyltransferases (NAT2) genotype as a susceptibility factor for lung cancer. Cancer Res 56: 3961–3966. Cascorbi I, Henning S, Brockmöller J, et al. 2000. Substantially reduced risk of cancer of the aerodigestive tract in subjects with variant -463A of the myeloperoxidase gene. Cancer Res 60:644–649. Castaing M, Youngson J, Zaridze D, Szeszenia-Dabrowska N, Rudnai P, Lissowska J, Fabianova E, Mates D, Bencko V, Foretova L, Navratilova M, Janout V, Fletcher T, Brennan P, Boffetta P. 2005. Is the risk of lung cancer reduced among eczema patients? Am J Epidemiol 162:542–547. Charloux A, Quoix E, Wolkove N, Small D, Pauli G, Kreisman H. 1997. The increasing incidence of lung adenocarcinoma: Reality or artifact? A review of the epidemiology of lung adenocarcinoma. Int J Epidemiol 26:14–23.

652

PART IV: CANCER BY TISSUE OF ORIGIN

Chen H, Lum A, Seifried A, Wilkens LR, Le Marchand L. 1999. Association of the NAD(P)H:quinone oxidoreductase 609C–> T polymorphism with a decreased lung cancer risk. Cancer Res 59:3045–3048. Chen S, Tang D, Xue K, et al. 2001. DNA repair gene XRCC1 and XPD polymorphisms and risk of lung cancer in a Chinese population. Carcinogenesis 23:1321–1325. Cheng L, Spitz MR, Hong WK, Wei Q. 2000. Reduced expression levels of nucleotide excision repair genes in lung cancer: A case-control analysis. Carcinogenesis 21:1527–1530. Chevillard S, Radicella JP, Levalois C, et al. 1998. Mutations in OGG1, a gene involved in the repair of oxidative DNA damage, are found in human lung and kidney tumours. Oncogene 16:3083–3086. Choi SW, Mason JB. 2000. Folate and carcinogenesis: An integrated scheme. J Nutr 130:129–132. Christensen PM, Gotzsche PC, Brosen K. 1997. The sparteine/debrisoquine (CYP2D6) oxidation polymorphism and the risk of lung cancer: A metaanalysis. Eur J Clin Pharmacol 51:389–393. Chung F-L. 1992. Chemoprevention of lung carcinogenesis by aromatic isothiocyanates. In: Wattenberg L, Lipkin M, Boone CW, Kelloff GJ, eds., Cancer Chemoprevention. Boca Raton, FL: CRC Press, Inc, pp. 227–245. Cleaver JE, States JC. 1997. The DNA damage-recognition problem in human and other eukaryotic cells: The XPA damage binding protein. Biochem J 328:1–12. Clegg LX, Li FP, Hankey BF, Chu K, Edwards BK. 2002. Cancer survival among US whites and minorities: A SEER population-based study. Arch Intern Med 162:1985–1993. Colditz GA, Branch LG, Lipnick RJ, et al. 1985. Increased green and yellow vegetable intake and lowered cancers deaths in an elderly population. Am J Clin Nutr 41:32–36. Combs GF, Combs SB. 1986. Selenium and cancer. In: Combs GF, Combs SB, eds. The Role of Selenium in Nutrition. San Diego, CA: Academic Press, pp. 413–462. Combs GF, Clark LC, Turnbull BW. 1997. Reduction of cancer risk with an oral supplement of selenium. Biomed Environ Sci 10:227–234. Cook NR, Lee IM, Manson JE, Buring JE, Hennekens CH. 2000. Effects of bcarotene supplementation on cancer incidence by baseline characteristics in the Physician’s Health Study (United States). Cancer Causes Control 11:617–626. Copas JB, Shi JQ. 2000. Reanalysis of epidemiological evidence on lung cancer and passive smoking. Br Med J 320:417–418. Coultas DB. 1994. Other occupational carcinogens. In: Samet JM, ed. Epidemiology of Lung Cancer. New York: Marcel Dekker, Inc, pp. 299–334. Darby SC, Hill D, Doll R. 2001. Radon: A likely carcinogen at all exposures. Ann Oncol 12:1341–1351. Darby SC, Samet JM. 1994. Radon. In: Samet JM, ed. Epidemiology of Lung Cancer. New York: Marcel Dekker, Inc, pp. 219–244. David-Beabes GL, London SJ. 2001a. Genetic polymorphism of XRCC1 and lung cancer risk among African-Americans and Caucasians. Lung Cancer 34:333–339. David-Beabes GL, Lunn RM, London SJ. 2001b. No association between the XPD (Lys751G1n) polymorphism or the XRCC3 (Thr241Met) polymorphism and lung cancer risk. Cancer Epidemiol Biomarkers Prev 10: 911–912. Deakin M, Elder J, Hendrickse C, et al. 1996. Glutathione S-transferase GSTT1 genotypes and susceptibility to cancer: Studies of interactions with GSTM1 in lung, oral, gastric and colorectal cancers. Carcinogenesis 17:881–884. de Laat WL, Jaspers NG, Hoeijmakers JH. 1999. Molecular mechanism of nucleotide excision repair. Genes Dev 13:768–785. Deloughery TG, Evans A, Sadeghi A, et al. 1996. Common mutation in methylenetetrahydrofolate reductase. Correlation with homocysteine metabolism and late-onset vascular disease. Circulation 94:3074–3078. de Stefani E, Deneo-Pellegrini, Mendilaharsu M, Carzoglio JC, Ronco A. 1997. Dietary fat and lung cancer: A case-control study in Uruguay. Cancer Causes Control 8:913–921. Devesa SS, Grauman DJ, Blot WJ, Fraumeni JF Jr. 1999a. Cancer surveillance series: Changing geographic patterns of lung cancer mortality in the United States, 1950 through 1994. J Natl Cancer Inst 91:1040–1050. Devesa SS, Grauman DJ, Blot WJ, Pennello GA, Hoover RN, Fraumeni JF. 1999b. Atlas of cancer mortality in the United States, 1950–94. Washington, DC: US Govt Print Off; [NIH Publ No. (NIH) 99-4564]. Devesa SS, Shaw GL, Blot WJ. 1991. Changing patterns of lung cancer incidence by histologic type. Cancer Epidemiol Biomarkers Prev 1:29–34. Divine KK, Gilliland FD, Crowell RE, et al. 2001. The XRCC1 399 glutamine allele is a risk factor for adenocarcinoma of the lung. Mutat Res 461:273–278. Doll R. 1955. Mortality from lung cancer in asbestos workers. Br J Ind Med 12:81–86.

Doll R, Hill AB. 1952. A study of the aetiology of carcinoma of the lung. Br Med J 2:1271–1286. Eaton DL, Gallagher EP, Bammler TK, Kunze KL. 1995. Role of cytochrome P4501A2 in chemical carcinogenesis: Implications for human variability in expression and enzyme activity. Pharmacogenetics 5:259–274. Edwards BK, Howe HL, Ries LAG, et al. 2002. Annual report to the nation on the status of cancer, 1973–1999, featuring implications of age and aging on U.S. cancer burden. Cancer 94:2766–2792. el-Zein R, Zwischenberger JB, Wood TG, et al. 1997. Combined genetic polymorphism and risk for development of lung cancer. Mutat Res 381:189–200. Eng C, Herman JG, Baylin SB. 2000. A bird’s eye view of global methylation. Nat Genet 24:101–102. Esteller M, Corn PG, Baylin SB, Herman JG. 2001. A gene hypermethylation profile of human cancer. Cancer Res 61:3225–3229. Etzel CJ, Amos CI, Spitz MR. 2003. Risk for smoking-related cancer among relatives of lung cancer patients. Cancer Res 63:8531–8535. Everson RB, Randerath E, Santella RM, Cefalo RC, Avitts TA, Randerath K. 1986. Detection of smoking-related covalent DNA adducts in human placenta. Science 231:54–57. Fan R, Wu MT, Miller D, et al. 2000. The p53 codon 72 polymorphism and lung cancer risk. Cancer Epidemiol Biomarkers Prev 9:1037–1042. Fernandez E, Borrell C. 1999. Cancer mortality by educational level in the city of Barcelona. Br J Cancer 79:684–689. Feyler A, Voho A, Bouchardy C, et al. 2002. Point: Myeloperoxidase -463G -> A polymorphism and lung cancer risk. Cancer Epidemiol Biomarkers Prev 11:1550–1554. Fingerhut MA, Halperin WE, Marlow DA, et al. 1991. Cancer mortality in workers exposed to 2,3,7,8-tetrachlorodibenzo p-dioxin. N Engl J Med 324:212–218. Fitzpatrick FA. 2001. Inflammation, carcinogenesis and cancer. Int Immunopharmacol 1:1651–1667. Flehinger BJ, Melamed MR, Zaman MB, Heelan RT, Perchick WB, Martini N. 1984. Early lung cancer detection: Results of the initial (prevalence) radiologic and cytologic screening in the Memorial Sloan-Kettering study. Am Rev Respir Dis 130:555–560. Fontana RS, Sanderson DR, Woolner LB, Taylor WF, Miller WE, Muhm JR. 1986. Lung cancer screening: The Mayo Program. J Occup Med 28:746–750. Fontana RS, Sanderson DR, Woolner LB, et al. 1991. Screening for lung cancer: A critique of the Mayo lung project. Cancer 67:1155–1164. Franklin WA. 2000. Diagnosis of lung cancer: Pathology of invasive and preinvasive neoplasia. Chest 117(Suppl 1):80S–89S. Frosst P, Blom HJ, Milos R, et al. 1995. A candidate genetic risk factor for vascular disease: A common mutation in methylenetetrahydrofolate reductase. Nat Genet 10:111–113. Fryer AA, Jones PW. 1999. Interactions between detoxifying enzyme polymorphisms and susceptibility to cancer. IARC Sci Publ 148:303–322. Furuya H, Fernandez-Salguero P, Gregory W, et al. 1995. Genetic polymorphism of CYP2C9 and its effect on warfarin maintenance dose requirement in patients undergoing anticoagulation therapy. Pharmacogenetics 5:389–392. Gadgeel SM, Severson RK, Jau Y, Weiss LK, Kalemkerian GP. 2001. Impact of race in lung cancer: Analysis of temporal trends from a surveillance, epidemiology, and end results database. Chest 120:55–63. Gao C, Wu J, Ding J, et al. 2002. Polymorphisms of methylenetetrahydrofolate reductase C677T and the risk of stomach cancer. Zhonghua Liu Xing Bing Xue Za Zhi 23:289–292. Gao C-M, Tajima K, Kuroishi T, Hirose K, Inouw M. 1993. Protective effects of raw vegetables and fruit against lung cancer among smokers and ex-smokers: A case-control study in the Tokai area of Japan. Jpn J Cancer Res 84:594–600. Garland M, Morris JS, Stampfer MJ, et al. 1995. Prospective study of toenail selenium levels and cancer among women. J Natl Cancer Inst 87: 497–505. Gazdar AF, Linnoila RI. 1988. The pathology of lung cancer—changing concepts and newer diagnostic techniques. Semin Oncol 15:215–225. Giovino GA. 2002. Epidemiology of tobacco use in the United States. Oncogene 21:7326–7340. Epidemiol 128:1241–1255. Gonzalez FJ, Skoda RC, Kimura S, et al. 1988. Characterization of the common genetic defect in humans deficient in debrisoquine metabolism. Nature 331:442–446. Goodman MT, Kolonel LN, Yoshizawa CN, Hankin JH. 1988. The effect of dietary cholesterol and fat on the risk of lung cancer in Hawaii. Am J Epidemiol 128:1241–1255. Goyette P, Sumner JS, Milos R, et al. 1994. Human methylenetetrahydrofolate reductase: Isolation of cDNA, mapping and mutation identification. Nat Genet 7:195–200.

Cancer of the Lung Greenwald P. 2003. b-carotene and lung cancer: A lesson for future chemoprevention investigations? J Natl Cancer Inst 95:E1. Guengerich FP, Kim D-H, Iwasaki M. 1991. Role of human cytochrome P-450 IIE1 in the oxidation of many low molecular cancer suspects. Chem Res Toxicol 4:168–179. Hackshaw AK. 1998. Lung cancer and passive smoking. Stat Methods Med Res 7:119–136. Hackshaw AK, Law MR, Wald NJ. 1997. The accumulated evidence on lung cancer and environmental tobacco smoke. Br Med J 315:980–988. Harkness EF, Brewster DH, Kerr KM, Fergusson RJ, MacFarlane GJ. 2002. Changing trends in incidence of lung cancer by histologic type in Scotland. Int J Cancer 102:179–183. Harris M J, Coggan M, Langton L, Wilson SR, Board PG. 1998. Polymorphism of the Pi class glutathione S-transferase in normal populations and cancer patients. Pharmacogenetics 8:27–31. Hart CL, Hole DJ, Gillis CR, Davey Smith G, Watt GCM, Hawthorne VM. 2001. Social class differences in lung cancer mortality: risk factor explanations using two Scottish cohort studies. Int J Epidemiol 30: 268–274. Hartman TJ, Woodson K, Stolzenberg-Solomon R, et al. 2001. Association of the B-vitamins pyridoxal 5¢-phosphate (B(6)), B(12), and folate with lung cancer risk in older men. Am J Epidemiol 153:688–694. Hartman TJ, Taylor PR, Alfthan G, et al. 2002. Toenail selenium concentration and lung cancer in male smokers (Finland). Cancer Causes Control 13: 923–928. Hassett C, Aicher L, Sidhu JS, Omiecinski CJ. 1994. Human microsomal epoxide hydrolase: Genetic polymorphism and functional expression in vitro of amino acid variants. Hum Mol Genet 3:421–428. Hatsukami DK, Lemmonds C, Zhang Y, Murphy SE, Le C, Carmella SG, Hecht SS. 2004. Evaluation of carcinogen exposure in people who used “reduced exposure” tobacco products. J Natl Cancer Inst 96:844–852. Hayashi SI, Watanabe J, Kawajiri K. 1991. Genetic polymorphisms in the 5¢flanking region change transcriptional regulation of the human cytochrome P450IIEI gene. J Biochem 110:559–565. Hebert JR, Kabat GC. 1990. Differences in dietary intake associated with smoking status. Eur J Clin Nutr 44:185–193. Hecht SS. 1998. Biochemistry, biology, and carcinogenicity of tobacco-specific N-nitrosamiines. Chem Res Toxicol 11:559–603. Hecht SS. 1999. Tobacco smoke carcinogens and lung cancer. J Natl Inst Cancer 91:1194–1210. Hecht SS, Chung F-L, Richie JP, et al. 1995. Effects of watercress consumption on metabolism of a tobacco-specific lung carcinogen in smokers. Cancer Epidemiol Biomarkers Prev 4:877–884. Hein DW, Grant DM, Stanton VP. 2000a. Single nucleotide polymorphisms can cause different structural folds of mRNA. Proc Natl Acad Sci USA 96:7871–7876. Hein DW, Doll MA, Adrian JF, et al. 2000b. Molecular genetics and epidemiology of NAT1 and NAT2 acetylation polymorphisms. Cancer Epidemiol Biomarkers Prev 9:29–42. Heinrich J, Mielck A, Schafer I, May W. 2000. Social inequality and environmentally-related diseases in Germany: Review of empirical results. Soc Prev Med 45:106–118. Henschke CI, McCauley DI, Yankelevitz DF, et al. 1999. Early Lung Cancer Action Project: Overall design and findings from baseline screening. Lancet 354:99–105. Hirayama T. 1981. Non smoking wives of heavy smokers have a higher risk of lung cancer: A study from Japan. Br Med J 282:149–155. Hirvonen A, Hugsgatvel-Pursiainen K, Karjalainen A, Anttila S, Vainio H. 1992. Point-mutational Msp1 and Ile-Val polymorphisms closely linked in the CYP1A1 gene: Lack of association with susceptibility to lung cancer in a Finnish study population. Cancer Epidemiol Biomarkers Prev 1:485–489. Hirvonen A, Husgafvel-Pursiainen K, Anttila S, Karjalainen A, Vainio H. 1993a. The human CYP2E1 gene and lung cancer: DraI and RsaI restriction fragment length polymorphism in a Finnish Cohort. Carcinogenesis 14:85–88. Hirvonen A, Husgafvel-Pursiainen K, Anttila S, Vainio H. 1993b. The GSTM1 null genotype as a potential risk modifier for squamous cell carcinoma of the lung. Carcinogenesis 14:1479–1481. Hirvonen A, Nylund L, Kociba P, Husgafvel-Pursiainen K, Vainio H. 1994. Modulation of urinary mutagenicity by genetically determined carcinogen metabolism in smokers. Carcinogenesis 15:813–815. Hirvonen A. 1999. Polymorphisms of xenobiotic-metabolizing enzymes and susceptibility to cancer. Environ Health Perspect 107:37–47. Hoeijmakers JHJ. 1994. Human nucleotide excision repair syndromes: Molecular clues to unexpected intricacies. Eur J Cancer 30:1912–1921. Holick CN, Michaud DS, Stolzenberg-Solomon R, et al. 2002. Dietary carotenoids, serum beta-carotene, and retinol and risk of lung cancer in

653

the alpha-tocopherol, beta-carotene cohort study. Am J Epidemiol 156:536–547. Houlston RS. 1999. Glutathione S-transferase M1 status and lung cancer risk: A meta-analysis. Cancer Epidemiol Biomarkers Prev 8:675–682. Houlston RS. 2000. CYP1A1 polymorphisms and lung cancer risk: A metaanalysis. Pharmacogenetics 10:105–114. Hsu TC, Spitz MR, Schantz SP. 1991. Mutagen sensitivity: A biologic marker of cancer susceptibility. Cancer Epidemiol Biomarkers Prev 1: 83–89. Hughes J, Weill H. 1991. Asbestosis as a necessary precursor of asbestosrelated lung cancer: Results of a prospective mortality study in asbestos products manufacturing. Br J Ind Med 48:229–233. Hung J, Kishimoto Y, Sugio K, Virmani A, McIntire DD, Minna JD, Gazdar AF. 1995. Allele-specific chromosome 3p deletions occur at an early stage in the pathogenesis of lung carcinoma. JAMA 273:558–563. Ingelman-Sundberg M, Johansson I, Persson I, et al. 1994. Genetic polymorphism of cytochrome P450: Interethnic differences and relationship to incidence of lung cancer. Pharmacogenetics 2:264–271. Institute of Medicine, Committee to assess the science base for tobacco harm and reduction, board on health promotion and disease prevention. 2001. Clearing and smoke: Assessing the science base for tobacco harm reduction. Washington DC: National Academy Press. International Agency for Research on Cancer. 1988. Man-made mineral fibres and radon. IARC Monogr Eval Carcinog Risks Hum 43. Lyon, France: IARC. International Agency for Research on Cancer. 1990. Chromium, nickel and welding IARC Monogr Eval Carcinog Risks Hum 49. Lyon, France: IARC. Ishibe N, Wiencke JK, Zuo ZF, McMillan A, Spitz MR, Kelsey KT. 1997. Susceptibility to lung cancer in light smokers associated with CYP1A1 polymorphisms in Mexican- and African-Americans. Cancer Epidemiol Biomarkers Prev 12:1075–1080. Ishida T, Takashima R, Fukayama M, et al. 1999. New DNA polymorphisms of human MMH/OGG1 gene: Prevalence of one polymorphism among lung-adenocarcinoma patients in Japanese. Int J Cancer 80:18–21. Ito H, Hamajima N, Takezaki T, Matsuo K, et al. 2002. A limited association of OGG1 Ser326Cys polymorphism for adenocarcinoma of the lung. J Epidemiol 12:258–265. Jain M, Burch JD, Howe GR, Risch HA, Miller AB. 1990. Dietary factors and risk of lung cancer: Results from a case control study, Toronto 1981–1985. Int J Cancer 45:287–293. Janssen-Heijnen MLG, Coebergh JWW. 2001. Trends in incidence and prognosis of the histological subtypes of lung cancer in North America, Australia, New Zealand and Europe. Lung Cancer 31:123– 137. Jatoi A, Daly BD, Kramer G, Mason JB. 2001. Folate status among patients with non-small cell lung cancer: A case-control study. J Surg Oncol 77:247–252. Jemal A, Chu KC, Tarone RE. 2001. Recent trends in lung cancer mortality in the United States. J Natl Cancer Inst 93:277–283. Jemal A, Murray T, Ward E, Samuels A, Tiwari RC, Ghafoor A, Feuer EJ, Thun MJ. 2005. Cancer Statistics 2005. Ca Cancer J Clin 55:10–30. Jin X, Wu X, Roth JA, et al. 1995. Higher lung cancer risk for younger African-Americans with the Pro/Pro p53 genotype. Carcinogenesis 16: 2205–2208. Joseph P, Jaiswal AK. 1994. NAD(P)H:quinone oxidoreductases 1 (DTdiaphorase) specifically prevents the formation of benzo[a]pyrene quinone-DNA adducts generated by cytochrome P4501A1 and P450 reductase. Proc Natl Acad Sci USA 91:8413–8417. Jourenkova-Mironova N, Wikman H, Bouchardy C, et al. 1998. Role of glutathione S-transferase GSTM1, GSTM3, GSTP1 and GSTT1 genotypes in modulating susceptibility to smoking-related lung cancer. Pharmacogenetics 8:495–502. Kabuto M, Imai H, Yonezawa C, et al. 1994. Prediagnostic serum selenium and zinc levels and subsequent risk of lung and stomach cancer in Japan. Cancer Epidemiol Biomarkers Prev 3:465–469. Kakinuma R, Ohmatsu H, Kaneko M, et al. 1999. Detection failures in spiral CT screening for lung cancer: Analysis of CT findings. Radiology 212:61–66. Kaneko M, Eguchi K, Ohmatsu H, et al. 1996. Peripheral lung cancer: Screening and detection with low-dose CT vs radiography. Radiology 201: 798–802. Kato S, Shields PG, Caporaso NE. 1994. Analysis of cytochrome P450 2E1 genetic polymorphisms in relation to human lung cancer. Cancer Epidemiol Biomarkers Prev 3:515–518. Kawajiri K, Watanabe J, Eguchi H, Hayashi S. 1995. Genetic polymorphisms of drug-metabolizing enzymes and lung cancer susceptibility. Pharmacogenetics 5(S70):3.

654

PART IV: CANCER BY TISSUE OF ORIGIN

Kelsey KT, Wiencke JK, Spitz MR. 1994. A race-specific genetic polymorphism in the CYP1A1 gene is not associated with lung cancer in African Americans. Carcinogenesis 15:1121–1124. Kerr KM. 2001. Pulmonary preinvasive neoplasia. J Clin Pathol 54:257–271. Kihara M, Noda K. 1999. Lung cancer risk of the GSTM1 null genotype is enhanced in the presence of the GSTP1 mutated genotype in male Japanese smokers. Nippon Koshu Eisei Zasshi 46:241–249. Knekt P, Marniemi J, Teppo L, Heliovaara M, Aromaa A. 1998. Is low selenium status a risk factor for lung cancer? Am J Epidemiol 148:975–982. Koch KS, Fletcher RG, Grond MP, et al. 1993. Inactivation of plasmid reporter gene expression by one benzo(a)pyrene diol-epoxide DNA adduct in adult rat hepatocytes. Cancer Res 53:2279–2286. Kohno T, Shinmura K, Tosaka M, et al. 1998. Genetic polymorphisms and alternative splicing of the hOGG1 gene, that is involved in the repair of 8-hydroxyguanine in damaged DNA. Oncogene 16:3219–3225. Koo LC. 1988. Dietary habits and lung cancer risk among Chinese females in Hong Kong who never smoked. Nutr Cancer 11:115–172. Koo LC. 1997. Diet and lung cancer 20+ years later: More questions than answers? Int J Cancer Suppl 10:22–29. Koop DR. 1992. Oxidative and reductive metabolism by cytochrome P4502E1. FASEB J 6:724–730. Korte JE, Brennan P, Henley SJ, Boffetta P. 2002. Dose-specific meta-analysis and sensitivity of the relation between alcohol consumption and lung cancer risk. Am J Epidemiol 155:496–506. Kreuzer M, Kreienbrock L, Gerken M, et al. 1998. Risk factors for lung cancer in young adults. Am J Epidemiol 147:1028–1037. Krieger N, Quesenberry C, Peng T, et al. 1999. Social class, race/ethnicity, and incidence of breast, cervix, colon, lung and prostate cancer among Asian, black, Hispanic, and white residents in the San Francisco Bay Area, 1988–92 (United States). Cancer Causes Control 10:525–537. Kroes R, Beems M, Bosland GS, Bunnik SJ. 1986. Nutritional factors in lung, colon, and prostate carcinogenesis in animal models. Fed Proc 45: 136–141. Kubik A, Polak J. 1986. Lung cancer detection: Results of a randomized prospective study in Czechoslovakia. Cancer 57:2427–2437. Kurie JM, Lee JS, Khuri FR, et al. 2000. N-(4-hydroxyphenyl)retinamide in the chemoprevention of squamous metaplasia and dysplasia of the bronchial epithelium. Clin Cancer Res 6:2973–2979. Kurie JM, Lotan R, Lee JJ, et al. 2003. Treatment of former smokers with 9cis-Reinoic acid reverses loss of retinoic acid receptor-beta expression in the bronchial epithelium: Results from a randomized placebo-clinical trial. J Natl Cancer Inst 95:206–214. Lam S, MacAulay C, le Riche JC, et al. 2002. A randomized phase Iib trial of anethole dithiolethione in smokers with bronchial dysplasia. J Natl Cancer Inst 94:1001–1009. Landi MT, Zocchetti C, Bernucci I, et al. 1996. Cytochrome P4501A2: Enzyme induction and genetic control in determining 4-aminobiphenyl-hemoglobin adduct levels. Cancer Epidemiol Biomarkers Prev 5:693–698. Larsen LK, Amri EZ, Mandrup S, Pacot C, Kristiansen K. 2002. Genomic organization of the mouse peroxisome proliferator activated receptor beta/delta gene. Alternative promoter usage and splicing yield transcripts exhibiting differential translational efficiency. J Biochem 365:11522– 11560. Laskin S, Drew RT, Cappiello V, Kuschner M, Nelson N. 1975. Inhalation carcinogenicity of alpha halo ethers. Arch Environ Health 30:70–72. Law MR, Morris JK, Wald NJ. 1997. Environmental tobacco smoke exposure and ischaemic heart disease: An evaluation of the evidence. Br Med J 315:973–980. Lee IM, Cook NR, Manson JE, Buring JE, Hennekens CH. 1999. b-carotene supplementation and incidence of cancer and cardiovascular disease: The Women’s Health Study. J Natl Cancer Inst 91:2102–2106. Lee JS, Lippman SM, Benner SE, et al. 1994. Randomized placebo-controlled trial of isotretinoin in chemoprevention of bronchial squamous metaplasia. J Clin Oncol 12:937–945. Le Marchand L, Donlon T, Lum-Jones A, Seifried A, Wilkens LR. 2002. Association of the hOGG1 Ser326Cys polymorphism with lung cancer risk. Cancer Epidemiol Biomarkers Prev 11:409–412. Le Marchand L, Seifried A, Lum A, Wilkens LR. 2000. Association of the myeloperoxidase -463G- A polymorphism with lung cancer risk. Cancer Epidemiol Biomarkers Prev 9:181–184. Le Marchand L, Sivaraman L, Pierce L, et al. 1998. Associations of CYP1A1, GSTM1, and CYP2E1 polymorphisms with lung cancer suggest cell type specificities to tobacco carcinogens. Cancer Res 58:4858–4863. Le Marchand L, Yoshizawa CN, Kolonel LN, Hankin JH, Goodman MT. 1989. Vegetable consumption and lung cancer risk: A population-based casecontrol study in Hawaii. J Natl Cancer Inst 81:1158–1164. Levin ML, Goldstein H, Gerhardt PR. 1950. Cancer and tobacco smoking: A preliminary report. JAMA 143:336–338.

Lewis SJ, Cherry NM, Niven RM, Barber PV, Povey AC. 2001. Polymorphisms in the NAD(P)H: Quinone oxidoreductase gene and small cell lung cancer risk in a UK population. Lung Cancer 34:177–183. Lewtas J. 1991. Experimental evidence for the carcinogenicity of air pollutants. In: Tomatis L, ed. Air Pollution and Human Cancer, pp. 49–62. Li D, Wang M, Cheng L, Spitz MR, Hittelman WN, Wei Q. 1996. In vitro induction of benzo(a)pyrene diol epoxide—DNA adducts in peripheral lymphocytes as a susceptibility marker for human lung cancer. Cancer Res 56:3638–3641. Li D, Firozi PF, Wang L, et al. 2001. Sensitivity to DNA damage induced by benzo(a)pyrene diol epoxide and risk of lung cancer—a case-control analysis. Cancer Res 61:1445–1450. Li L, Peterson CA, Lu X, Legerski RJ. 1995. Mutations in XPA that prevent association with ERCC1 are defective in nucleotide excision repair. Mol Cell Biol 15:1993–1998. Li X, Mutanen P, Hemminki K. 2001. Gender-specific incidence trends in lung cancer by histological type in Sweden. Eur J Cancer Prev 10:227–235. Lippman SM, Lee JJ, Karp DD, et al. 2001. Randomized phase III intergroup trial of isotretinion to prevent second primary tumors in stage I non-smallcell lung cancer. J Natl Cancer Inst 93:605–618. London SJ, Daly AK, Fairbrother KS, et al. 1995. Lung cancer risk in AfricanAmericans in relation to a race-specific CYP1A1 polymorphism. Cancer Res 55:6035–6037. London SJ, Daly AK, Leathart JBS, Navidi WC, Idle JR. 1996. Lung cancer risk in relation to the CYP2C9*1/CYP2C9*2 genetic polymorphism among African-Americans and Caucasians in Los Angeles County, California. Pharmacogenetics 6:527–533. London SJ, Lehman TA, Taylor JA. 1997. Myeloperoxidase genetic polymorphism and lung cancer risk. Cancer Res 57:5001–5003. London SJ, Yuan JM, Coetzee GA, Gao YT, Ross RK, Yu MC. 2000a. CYP1A1 I462V genetic polymorphism and lung cancer risk in a cohort of men in Shanghai, China. Cancer Epidemiol Biomarkers Prev 9:987–991. London SJ, Yuan JM, Chung FL, et al. 2000b. Isothiocyanates, glutathione S-transferase M1 and T1 polymorphisms and lung cancer risk: A prospective study of men in Shanghai, China. Lancet 356:724–729. Lu R, Nash HM, Verdine GL. 1997. A mammalian DNA repair enzyme that excises oxidatively damaged guanines maps to a locus frequently lost in lung cancer. Curr Biol 7:397–407. Lunn RM, Helzlsouer KJ, Parshad R, et al. 2000. XPD polymorphisms: Effects on DNA repair proficiency. Carcinogenesis 21:551–555. Mahadevia PJ, Fleisher LA, Frick KD, Eng J, Goodman SN, Powe NR. 2003. Lung cancer screening with helical computed tomography in older adult smokers: A decision and cost-effectiveness analysis. JAMA 289: 313–322. Mallet WG, Mosebrook DR, Trush MA. 1991. Activation of (±)-trans-7,8dihydroxy-7,8-dihydrobenzo[a]pyrene to diolepoxides by human polymorphonuclear leukocytes or myeloperoxidase. Carcinogenesis 12: 521–524. Mannervik B, Awasthi YC, Board PG, et al. 1992. Nomenclature for human glutathione transferases. Biochem J 282:305–308. Mao Y, Hu J, Ugnat A, Semenciw R, Finchma S, Canadian Cancer Registries Epidemiology Research Group. 2001. Socioeconomic status and lung cancer risk in Canada. Int J Epidemiol 30:809–817. Marcus PM. 2001. Lung cancer screening: An update. J Clin Oncol 19(18 Suppl):83S–86S. Marcus PM, Bergstralh EJ, Fagerstrom RM, et al. 2000. Lung cancer mortality in the Mayo Lung Project: Impact of extended follow-up. J Natl Cancer Inst 92:1308–1316. Markewitz BA, Owens MW, Payne DK. 1999. The pathogenesis of chronic obstructive pulmonary disease. Am J Med Sci 318:74–78. Marshall B, Chevalier A, Garillon C, Goldberg M, Coing F. 1999. Socioeconomic status, social mobility and cancer occurrence during working life: A case-control study among French electricity and gas workers. Cancer Causes Control 10:495–502. Marshall D, Simpson KN, Earle CC, Chu C-W. 2001. Potential cost-effectiveness of one-time screening for lung cancer (LC) in a high-risk cohort. Lung Cancer 32:227–236. Martinez C, Agundez JAG, Olivera M, Martin R, Laredo JM, Beitez J. 1995. Lung cancer and mutations at the polymorphic NAT2 gene locus. Pharmacogenetics 5:207–214. Mason JB, Levesque T. 1996. Folate: Effects on carcinogenesis and the potential for cancer chemoprevention. Oncology (Huntingt) 10:1727–1736. Masson M, Niedergang C, Schreiber V, Muller S, Demarcia JM, Demurcia G. 1998. XRCC1 is specifically associated with PARP polymerase and negatively regulates its activity following DNA damage. Mol Cell Biol 18:3563–3571. McLarty JW, Holiday DB, Girard WM, Yanagihara RH, Kummet TD, Greenberg SD. 1995. Beta-Carotene, vitamin A, and lung cancer chemo-

Cancer of the Lung prevention: Results of an intermediate endpoint study. Am J Clin Nutr 62(6 Suppl):1431S–1438S. Melton DW, Ketchen AM, Nunez F, Bonatti-Abbondandolo S, Squires S, Johnson RT. 1998. Cells from ERCC1-deficient mice show increased instability and a reduced frequency of S-phase-dependent illegitimate chromosome exchange but a normal frequency of homologous recombination. J Cell Sci 111:395–404. Messina ES, Tyndale RF, Sellers EM. 1997. A major role for CYP2A6 in nicotine C-oxidation by human liver microsomes. J Pharmacol Exp Ther 282:1608–1614. Misra RR, Tangrea JA, Virtamo J, et al. 2001. Variation in the promoter region of the myeloperoxidase gene is not directly related to lung cancer risk among male smokers in Finland. Cancer Lett 164:161–167. Miyamoto I, Miura N, Niwa H, Miyazaki J, Tanaka K. 1992. Mutational analysis of the structure and function of the xeroderma pigmentosum group A complementing protein. Identification of essential domains for nuclear localization and DNA excision repair. J Biol Chem 267: 12182–12187. Mohr DL, Blot WJ, Tousey PM, van Doren ML, Wolfe KW. 1999. Southern cooking and lung cancer. Nutr Cancer 35:34–43. Mohrenweiser HW, Carrano AV, Fertitta A, et al. 1989. Refined mapping of the three DNA repair genes, ERCC1, ERCC2, and XRCC1, on human chromosome 19. Cytogenet Cell Genet 52:11–14. Mohrenweiser HW, Xi T, Vazquez-Matias J, Jones IM. 2002. Identification of 127 amino acid substitution variants in screening 37 DNA repair genes in humans. Cancer Epidemiol Biomarkers Prev 11:1054–1064. Moller P, Knudsen LE, Frentz G, Dybdahl M, Wallin H, Nexo BA. 1998. Seasonal variation of DNA damage and repair in patients with non-melanoma skin cancer and referents with and without psoriasis. Mutat Res 407: 25–34. Morgan RW. 1992. Attitudes about asbestos and lung cancer. Am J Ind Med 22:437–441. Morita T. 2002. A statistical study of lung cancer in the annual of pathological autopsy cases in Japan, from 1958 to 1997, with reference to time trends of lung cancer in the world. Jpn J Cancer Res 93:15–23. Murata M, Tagawa M, Kimura M, Kimura H, Watanabe S, Saisho H. 1996. Analysis of a germ line polymorphism of the p53 gene in lung cancer patients; Discrete results with smoking history. Carcinogenesis 17: 261–264. Nair U, Bartsch H. 2001. Metabolic polymorphisms as susceptibility markers for lung and oral cavity cancer. IARC Sci Publ 54:271–290. Nakajima M, Kwon JT, Tanaka N, et al. 2001. Relationship between interindividual differences in nicotine metabolism and CYP2A6 genetic polymorphism in humans. Clin Pharmacol Ther 69:72–78. Nakajima M, Kuroiwa Y, Yokoi T. 2002. Interindividual differences in nicotine metabolism and genetic polymorphisms of human CYP2A6. Drug Metab Rev 34:865–877. Nakajima M, Yokoi T, Mizutani M, Kinoshita M, Funayama M, Kamataki T. 1999. Genetic polymorphism in the 5¢-flanking region of human CYP1A2 gene: Effect on the CYP1A2 inducibility in humans. J Biochem (Tokyo) 125:803–808. National Research Council (NRC). 1986. Committee on passive smoking. Environmental tobacco smoke: Measuring exposures and assessing health effects. Washington DC: National Academy Press. National Research Council (NRC). 1999. Committee on health exposures of radon: BEIR VI. Health effects of exposure to radon. Washington DC: National Academy Press. Nazar-Stewart V, Motulsky AG, Eaton DL, et al. 1993. The glutathione Stransferase mu polymorphism as a marker for the susceptibility to lung carcinoma. Cancer Res 53:2313–2318. Nyberg F, Hou S-M, Hemminki K, Lambert B, Pershagen G. 1998. Glutathione S-transferase M1 and N-acetyltransferase 2 genetic polymorphisms and exposure to tobacco smoke in nonsmoking and smoking lung cancer patients and population controls. Cancer Epidemiol Biomarkers Prev 7:875–883. Oesch F, Glatt H, Schimassmann H. 1977. The apparent ubiquity of epoxide hydrolase in rat organs. Biochem Pharmacol 26:603–607. Omenn GS, Goodman GE, Thornquist MD, et al. 1996. Effects of a combination of beta carotene and vitamin A on long cancer and cardiovascular disease. N Engl J Med 334:1150–1155. Omiecinski CJ, Aicher L, Holubkov R, Checkoway H. 1993. Human peripheral lymphocytes as indicators of microsomal epoxide hydrolase activity in liver and lung. Pharmacogenetics 3:150–158. Ooi WL, Eltaon RC, Chen VW, Bailey-Wilson JE, Rothschild H. 1986. Increased familial risk for lung cancer. J Natl Cancer Inst 76:217–222. Oscarson M, McLellan RA, Gullsten H, et al. 1999. Identification and characterization of novel polymorphisms in the CYP2A locus: Implications for nicotine metabolism. FEBS Lett 460:321–327.

655

Oscarson M, McLellan RA, Asp V, et al. 2002. Characterization of a novel CYP2A7/CYP2A6 hybrid allele (CYP2A6*12) that causes reduced CYP2A6 activity. Hum Mutat 20:275–283. Ozawa S, McDaniel LP, Tang YM, Schoket B, Vincze I, Kadlubar FF. 1997. Cyp2C9 and GSTP1 genetic polymorphisms in patients with smokingrelated lung cancer. Proc Am Assoc Cancer Res 38:212P. Palozza P. 1998. Prooxicant actions of carotenoids in biologic systems. Nutr Rev 56:257–265. Park JY, Park S, Choi J, et al. 2002a. Polymorphism of the DNA repair gene Xeroderma Pigmentosum Group A and risk of primary lung cancer. Cancer Epidemiol Biomarkers Prev 11:993–997. Park JY, Lee SY, Jeon HS, et al. 2002b. Polymorphism of the DNA repair gene XRCC1 and risk of primary lung cancer. Cancer Epidemiol Biomarkers Prev 11:23–27. Parkin DM, Bray F, Ferlay J, Pisani P. 2001a. Estimating the world cancer burden: Globocan 2000. Int J Cancer 94:153–156. Parkin DM, Bray F, Ferlay J, Pisani P. 2005. Global cancer statistics, 2002. Ca Cancer J Clin 55:74–108. Parkin DM, Bray FI, Devesa SS. 2001b. Cancer burden in the year 2000. The global picture. Eur J Cancer 37:S4–S66. Parkin DM, Pisani P, Ferlay J. 1999. Global cancer statistics. Ca Cancer J Clin 49:33–64. Patterson BH, Levander OA. 1997. Naturally occurring selenium compounds in cancer chemoprevention trials: A workshop summary. Cancer Epidemiol Biomarkers Prev 6:63–69. Pemble S, Schroeder KR, Spencer SR, Meyer DJ, Hallier E, Taylor JB. 1994. Human glutathione S-transferase theta (GSTT1): cDNA cloning and the characterization of a genetic polymorphism. Biochem J 300: 271–276. Pershagen G. 1994. Passive smoking and lung cancer. In: Samet JM, ed. Epidemiology of Lung Cancer. New York: Marcel Dekker, Inc, pp. 109–130. Persson I, Johansson I, Bergling H, et al. 1993. Genetic polymorphism of cytochrome P4502E1 in a Swedish population relationship to incidence of lung cancer. FEBS Lett 319:207–211. Peto R, Darby S, Deo H, Silcocks P, Whitley E, Doll R. 2000. Smoking, smoking cessation, and lung cancer in the UK since 1950: Combination of national statistics with two case-control studies. Br Med J 321:323– 329. Phillips DH, Hewer A, Grover PL. 1986. Aromatic DNA adducts in human bone marrow and peripheral blood leukocytes. Carcinogenesis 7: 2071–2075. Phillip PA, Fitzgerald DL, Cartwright RA, Peake MD, Rogers HJ. 1988. Polymorphic N-acetylation capacity in lung cancer. Carcinogenesis 9: 491–493. Piedrafita FJ, Molander RB, Vansant G, Orlova EA, Pfahl M, Reynolds WF. 1996. An Alu element in the myeloperoxidase promoter contains a composite SP1-thyroid hormone-retinoic acid response element. J Biol Chem 271:14412–14420. Pillow PC, Hursting SD, Duphorne CM, et al. 1997. Case-control assessment of diet and lung cancer risk in African Americans and Mexican Americans. Nutr Cancer 29:129–173. Pope CA, Burnett RT, Thun MJ, et al. 2002. Lung cancer, cardiopulmonary mortality, and long-term exposure to the fine particulate air pollution. JAMA 288:830. Prescott E, Gronbaek M, Becker U, Sorensen TI. 1999. Alcohol intake and the risk of lung cancer: Influence of type of alcoholic beverage. Am J Epidemiol 149:463–470. Qiao Y, Spitz MR, Guo Z, et al. 2002. Rapid assessment of repair of ultraviolet DNA damage with a modified host-cell reactivation assay using a luciferase reporter gene and correlation with polymorphisms of DNA repair genes in normal human lymphocytes. Mutat Res 509:165–174. Quinones L, Lucas D, Godoy J, et al. 2001. CYP1A1, CYP2E1 and GSTM1 genetic polymorphisms. The effect of single and combined genotypes on lung cancer susceptibility in Chilean people. Cancer Lett 174:35–44. Rady PL, Szucs S, Grady J, et al. 2002. Genetic polymorphisms of methylenetetrahydrofolate reductase (MTHFR) and methionine synthase reductase (MTRR) in ethnic populations in Texas; A report of a novel MTHFR polymorphic site, G1793A. Am J Med Genet 107:162–168. Ratnasinghe D, Yao SX, Tangrea JA, et al. 2001. Polymorphisms of the DNA repair gene XRCC1 and lung cancer risk. Cancer Epidemiol Biomarkers Prev 10:119–123. Rebbeck TR. 1997. Molecular epidemiology of the human glutathione Stransferase genotypes GSTM1 and GSTT1 in cancer susceptibility. Cancer Epidemiol Biomarkers Prev 6:733–743. Reddy MV, Randerath K. 1986. Nuclease P1-mediated enhancement of sensitivity of 32P-postlabeling test for structurally diverse DNA adducts. Carcinogenesis 7:1543–1551.

656

PART IV: CANCER BY TISSUE OF ORIGIN

Rees DLP, Gasser R, Meyer UA, Birkett DJ. 1993. Site directed mutation studies of human liver cytochrome P-450 isozymes in the CYP2C subfamily. Biochem 289:533–538. Reid ME, Duffield-Lillico AJ, Garland L, Turnbull BW, Clark LC, Marshall JR. 2002. Selenium supplementation and lung cancer incidence: An update of the nutritional prevention of cancer trial. Cancer Epidemiol Biomarkers Prev 11:1285–1291. Report of the International Committee on Nickel Carcinogenesis in Man. 1990. Scand J Work Environ Health 16:1–82. Reszka E, Wasowicz W. 2001. Significance of genetic polymorphisms in glutathione S-transferase multigene family and lung cancer risk. Int J Occup Med Environ Health 14:99–113. Rettie AE, Wienkers LC, Gonzalez FJ, Trager WF, Korzekwa KR. 1994. Impaired (S)-warfarin metabolism catalysed by the R144C allelic variant of CYP2C9. Pharmacogenetics 4:39–42. Reynolds P. 1999. Epidemiologic evidence for workplace ETS as a risk factor for lung cancer among nonsmokers: Specific risk estimates. Environ Health Perspect 107(Suppl):865–877. Ries LAG, Eisner MP, Kosary CL, et al., eds. 2002. SEER Cancer Statistics Review, 1973–1999. Bethesda, MD: National Cancer Institute. Roots I, Drakoulis N, Ploch M, et al. 1988. Debrisoquine hydroxylation phenotype, acetylation phenotypes, and ABO blood groups as genetic host factors of lung cancer risk. Klin Wochenshr 66:87–97. Rostami-Hodjegan A, Lennard MS, Woods HF, Tucker GT. 1998. Metaanalysis of studies of the CYP2D6 polymorphism in relation to lung cancer and Parkinson’s disease. Pharmacogenetics 8:227–238. Rosvold EA, McGlynn KA, Lustbader ED, Buetow KH. 1995. Identification of an NAD(P)H:quinone oxidoreductase polymorphism and its association with lung cancer and smoking. Pharmacogenetics 5:199–206. Ryberg D, Skaug V, Hewer A, et al. 1997. Genotypes of glutathione transferase M1 and P1 and their significance for lung DNA adduct levels and cancer risk. Carcinogenesis 18:1285–1289. Saarikoski ST, Voho A, Reinikainen M, et al. 1998. Combined effect of polymorphic GST genes on individual susceptibility to lung cancer. Int J Cancer 77:516–521. Sachse C, Brockmoller J, Bauer S, Roots I. 1999. Functional significance of a CÆA polymorphism in intron 1 of the cytochrome P450 CYP1A2 gene tested with caffeine. Br J Clin Pharmacol 47:445–449. Santillan AA, Camargo CA Jr, Colditz GA. 2003. A meta-analysis of asthma and risk of lung cancer (United States). Cancer Causes Control 14: 327–334. Saracci R. 1987. The interactions of tobacco smoking and other agents in cancer etiology. Epidemiol Rev 9:175–193. Schabath MB, Delclos GL, Martynowicz MM, Greisinger AJ, Wu X, Spitz MR. 2005a. Opposing effects of emphysema, hay fever, and select genetic variants on lung cancer risk. Am J Epidemiol 161:412–422. Schabath MB, Hernandez LM, Wu X, Pillow PC, Spitz MR. 2005b. Dietary phytoestrogens and lung cancer risk. JAMA 294(12):1493–1504. Schabath MB, Spitz MR, Zhang X, Delclos GL, Wu X. 2000. Genetic variants of myeloperoxidase and lung cancer risk. Carcinogenesis 21:1163– 1166. Schwartz AG, Yang P, Swanson GM. 1996. Familial aggregation of lung among nonsmokers and their relatives. Am J Epidemiol 144:554–562. Seidegard J, Ekstrom G. 1997. The role of human glutathione transferases and epoxide hydrolases in the metabolism of xenobiotics. Environ Health Perspect 105:791–799. Seidegard J, Vorachek WR, Pero RW, Pearson WR. 1988. Hereditary differences in the expression of the human glutathione transferase active on trans-stilbene oxide are due to a gene deletion. Proc Natl Acad Sci USA 85:7293–7297. Seidegard J, Pero RW, Markowiz MM, Roush G, Miller DG, Beattie J. 1990. Isoenzyme(s) of glutathione transferase (class Mu) as a marker for the susceptibility to lung cancer. Carcinogenesis 11:33–36. Selikoff IJ. 1980. Latency of asbestos disease among insulation workers in the United States and Canada. Cancer 46:2736–2740. Shaw GL, Falk RT, Pickle LW, Mason TJ, Buffler PA. 1991. Lung cancer risk associated with cancer in relatives. J Clin Epidemiol 44:429– 437. Shen H, Spitz MR, Wang LE, Hong WK, Wei Q. 2001. Polymorphisms of methylene-tetrahydrofolate reductase and risk of lung cancer: A casecontrol study. Cancer Epidemiol Biomarkers Prev 10:397–401. Shen H, Wang L, Spitz MR, Hong WK, Mao L, Wei Q. 2002. A novel polymorphism in human cytosine DNA-methyltransferase-3B promoter is associated with an increased risk of lung cancer. Cancer Res 62: 4992–4995. Shen H, Wei Q, Pillow PC, Amos CI, Hong WK, Spitz MR. 2003. Dietary folate intake and lung cancer risk in former smokers: A case-control analysis. Cancer Epidemiol Bomarkers Prev 12:980–986.

Shen M, Berndt SI, Rothman N, DeMarin DM, Mumford JL, He X, et al. 2005. Polymorphisms in the DNA nucleotide excision repair genes and lung cancer risk in Xuan Wei, China. Int J Cancer 116:768–773. Shen MR, Jones IM, Mohrenweiser HW. 1998. Nonconservative amino acid substitution variants exist at polymorphic frequency in DNA repair genes in healthy humans. Cancer Res 58:604–608. Shields PG. 2002. Tobacco smoking, harm reduction, and biomarkers. J Natl Cancer Inst 94:1435–1444. Shields PG, Caporaso NE, Falk RT, et al. 1993. Lung cancer, race and a CYP1A1 genetic polymorphism. Cancer Epidemiol Biomarkers Prev 2:481–485. Shou M, Korzekwa KR, Crespt CL, Gonzalez FJ, Gelboin HV. 1994. The role of 12 cDNA-expressed human, rodent, and rabbit cytochromes P450 in the metabolism of benzo[a]pyrene and benzo[a]pyrene trans-7,8cihydrodiol. Mol Carcinog 10:159–168. Singh GK, Miller BA, Hankey BF. 2002. Changing area socioeconomic patterns in US cancer mortality, 1950–1998: Part II—lung and colorectal cancers. J Natl Cancer Inst 94:916–925. Skoda RC, Demierre A, McBride OW, Gonzalez FJ, Meyers UA. 1988. Human microsomal xenobiotic epoxide hydrolase. Complementary DNA sequence, complementary DNA-directed expression in COS-1 cells, and chromosomal localization. J Biol Chem 263:1549–1554. Smith CAD, Harrison DJ. 1997. Association between polymorphism in gene for microsomal epoxide hydrolase and susceptibility to emphysema. Lancet 350:630–633. Smith CJ, Perfetti TA, Garg R, Hansch C. 2003. IARC carcinogens reported in cigarette mainstream smoke and their calculated log P values. Food Chem Toxicol 41:807–817. Smith G, Stanley LA, Sim E, Strange RC, Wolf CR. 1995. Metabolic polymorphisms and cancer susceptibility. Cancer Surv 25:27–66. Smith-Warner SA, Ritz J, Hunter DJ, et al. 2002. Dietary fat and risk of lung cancer in a pooled analysis of prospective studies. Cancer Epidemiol Biomarkers Prev 11:987–992. Song C, Xing D, Tan W, Wei Q, Lin D. 2001. Methylenetetrahydrofolate reductase polymorphisms increase risk of esophageal squamous cell carcinoma in a Chinese population. Cancer Res 61:3272–3275. Song N, Tan W, Xing D, Lin D. 2001. CYP 1A1 polymorphism and risk of lung cancer in relation to tobacco smoking: A case-control study in China. Carcinogenesis 22:11–16. Speizer FE, Colditz GA, Hunter DJ, Rosner B, Hennekens C. 1999. Prospective study of smoking, antioxidant intake, and lung cancer in middle-aged women (USA). Cancer Causes Control 10:475–482. Speizer FE, Samet JM. 1994. Air pollution and lung cancer. In: Samet JM, ed. Epidemiology of Lung Cancer. New York: Marcel Dekker, Inc, pp. 131–150. Spitz MR, Duphorne CM, Detry MA, et al. 2000. Dietary intake of isothiocyanates: Evidence of a joint effect with glutathione S-transferase polymorphisms in lung cancer risk. Cancer Epidemiol Biomarkers Prev 9:1017–1020. Spitz MR, Hsu TC, Wu XF, Fueger JJ, Amos CI, Roth JA. 1995. Mutagen sensitivity as a biologic marker of lung cancer risk in African Americans. Cancer Epidemiol Biomarkers Prev 4:99–103. Spitz MR, Wei Q, Dong Q, Amos CI, Wu X. 2003. Genetic susceptibility to lung cancer: The role of DNA damage and repair. Cancer Epidemiol Biomarkers Prev 12:689–698. Spitz MR, Wu X, Wang Y. 2001. Modulation of nucleotide excision repair capacity by XPD polymorphisms in lung cancer patients. Cancer Res 61:1354–1357. Sporn MB, Roberts AB. 1984. Role of retinoids in differentiation and carcinogenesis. J Natl Cancer Inst 73:1381–1386. Steinmetz KA, Potter JD, Folsom AR. 1993. Vegetables, fruit and lung cancer in the Iowa Women’s Health Study. Cancer Res 53:536–543. Stellman SD, Muscat JE, Thompson S, Hoffmann D, Wynder EL. 1997. Risk of squamous cell carcinoma and adenocarcinoma of the lung in relation to lifetime filter cigarette smoking. Cancer 80:382–388. Stephens EA, Taylor JA, Kaplan N, et al. 1994. Ethnic variation in the CYP2E1 gene: Polymorphism analysis of 695 African-Americans, EuropeanAmericans and Taiwanese. Pharmacogenetics 4:185–192. Strom SS, Wu X, Sigurdson AJ, et al. 1995. Lung cancer, smoking patterns, and mutagen sensitivity in Mexican-Americans. J Natl Cancer Inst 18: 29–33. Stucker I, Boffetta P, Antilla S, et al. 2001. Lack of interaction between asbestos exposure and glutathione S-transferase M1 and T1 genotypes in lung carcinogenesis. Cancer Epidemiol Biomarkers Prev 10:1253– 1258. Stucker I, Hirvonen A, de Waziers I, et al. 2002. Genetic polymorphisms of glutathione S-transferases as modulators of lung cancer susceptibility. Carcinogenesis 9:1475–1481.

Cancer of the Lung Subar AF, Harlan LC, Mattson ME. 1990. Food and nutrient intake differences between smokers and non-smokers in the US. Am J Public Health 80:1323–1329. Sugimura H, Hamada GS, Suzuki I, et al. 1995. CYP1A1 and CYP2E1 polymorphism and lung cancer, case-control study in Rio de Janeiro, Brazil. Pharmacogenetics 5:S145–148. Sugimura H, Kohno T, Wakai K, et al. 1999. hOGG1 Ser326Cy....s polymorphism and lung cancer susceptibility. Cancer Epidemiol Biomarkers Prev 8:669–674. Sugimura H, Wakai K, Genka K, et al. 1998. Association of Ile462Val (Exon 7) polymorphism of cytochrome P450 IA1 with lung cancer in the Asian population: Further evidence from a case-control study in Okinawa. Cancer Epidemiol Biomarkers Prev 7:413–417. Sunaga N, Kohno T, Yanagitani N, Sugimura H, Kunitoh H, Tamura T, Takei Y, Tsuchiya S, Saito R, Yokota J. 2002. Contribution of the NQO1 and GSTT1 polymorphisms to lung adenocarcinoma susceptibility. Cancer Epidemiol Biomarkers Prev 11(8):730–788. Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database. National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2003. Swanson CA, Brown CC, Sinha R, Kulldorff M, Brownson RC, Alavanja MCR. 1997. Dietary fats and lung cancer risk among women: The Missouri Women’s Health Study (United States). Cancer Causes Control 8:883–893. Taioli E, Ford J, Trachman J, Li Y, Demopoulos R, Garte S. 1998. Lung cancer risk and CYP1A1 genotype in African Americans. Carcinogenesis 19: 813–817. Taylor EM, Broughton BC, Botta E, et al. 1997. Xeroderma pigmentosum and trichothiodystrophy are associated with different mutations in the XPD (ERCC2) repair/transcription gene. Proc Natl Acad Sci USA 94:8658–8663. Taylor R, Cumming R, Woodward A, Black M. 2001. Passive smoking and lung cancer: A cumulative meta-analysis. Aust NZ J Public Health 25: 203–211. The Alpha-Tocopherol, Beta Carotene Cancer Prevention Study Group. 1994. The effect of vitamin E and beta carotene on the incidence of lung cancer and other cancers in male smokers. N Engl J Med 330: 1029–1035. Thomas M, Kalita A, Labrecque S, Pim D, Banks L, Matlashewski G. 1999. Two polymorphic variants of wild-type p53 differ biochemically and biologically. Mol Cell Biol 19:1092–1100. Thun MJ, Henley SJ, Calle EE. 2002. Tobacco use and cancer: An epidemiologic perspective for geneticists. Oncogene 21:7307–7325. Tockman M. 1986. Survival and mortality from lung cancer in a screened population: The Johns Hopkins study. Chest 57:44–53. To-Figueras J, Gene M, Gomez-Catalan J, et al. 1997. Glutathione Stransferase M1 (GSTM1) and T1 (GSTT1) polymorphisms and lung cancer risk among Northwestern Mediterraneans. Carcinogenesis 18: 1529–1533. To-Figueras J, Gene M, Gomez-Catalan J, et al. 1999. Genetic polymorphism of glutathione S-transferase P1 gene and lung cancer risk. Cancer Causes Control 10:65–70. Tokuhata GK, Lilienfeld AM. 1963. Familial aggregation of lung cancer in humans. J Natl Cancer Inst 30:289–312. Tong L, Spitz MR, Fueger JJ, Amos CI. 1996. Lung carcinoma in former smokers. Cancer 78:1004–1010. Traver RD, Siegel D, Beall HD, et al. 1997. Characterization of a polymorphism in NAD(P)H:quinone oxidoreductases 1 (DT-diaphorase). Br J Cancer 75:69–75. Travis WD. 2002. Pathology of lung cancer. Clinics Chest Med 23:65–81. Trichopoulos D, Kalandidi A, Sparros L, MacMahon B. 1981. Lung cancer and passive smoking. Int J Cancer 27:1–4. Trobs M, Renner T, Scherer G, et al. 2002. Nutrition, antioxidants, and risk factor profile of nonsmokers, passive smokers and smokers of the Prevention Education Program (PEP) in Nuremberg, Germany. Prev Med 34:600–607. Uematsu F, Kikuchi H, Motomiya M, et al. 1991. Association between restriction fragment length polymorphism of the human cytochrome P450IIEI gene and susceptibility to lung cancer. Jpn J Cancer Res 82:254–256. US Department of Health and Human Services (USDHHS). 1986. The health consequences of involuntary smoking: A report of the Surgeon General. USDHHS Publication No (CDC) 87–8398. Washington DC: US Government Printing Office. US Environmental Protection Agency (USEPA). 1992. Respiratory health effects of passive smoking: lung cancer and other disorders. EPA Publication No.600/006F. Washington DC: US Government Printing Office.

657

van den Brandt PA, Goldbohm RA, van’t Veer P, et al. 1993. A prospective cohort study on selenium status and the risk of lung cancer. Cancer Res 53:4860–4865. van der Put NM, Gabreels F, Stevens EM, et al. 1998. A second common mutation in the methylenetetrahydrofolate reductase gene: An additional risk factor for neural-tube defects? Am J Hum Genet 62:1044–1051. Verhoeven DTH, Goldbohm RA, van Poppel G, Verhagen H, van den Brandt P. 1996. Epidemiological studies on Brassica vegetables and cancer risk. Cancer Epidemiol Biomarkers Prev 5:733–748. Villard P H, Seree E, Lacarelle B, et al. 1994. Effect of cigarette smoke on hepatic and pulmonary cytochrome P450 in mouse: Evidence for CYP2E1 induction in lung. Biochem Biophys Res Commun 202: 1731–1737. Virmani AK, Tsou JA, Siegmund KD, et al. 2002. Hierarchical clustering of lung cancer cell lines using DNA methylation markers. Cancer Epidemiol Biomarkers Prev 11:291–297. Voorrips LE, Goldbohm RA, Brants HA, van Poppel GA, Sturmans F, Hermus RJ, van den Brandt PA. 2000. A prospective cohort study on antioxidant and folate intake and male lung cancer risk. Cancer Epidemiol Biomarkers Prev 9:357–365. Wang XW, Vermeulen W, Coursen JD, et al. 1996. The XPB and XPD DNA helicases are components of the p53-mediated apoptosis pathway. Genes Dev 10:1219–1232. Wang YC, Chen CY, Chen SK, Chang YY, Lin P. 1999. p53 codon 72 polymorphism in Taiwanese lung cancer patients: Association with lung cancer susceptibility and prognosis. Clin Cancer Res 5:129–134. Wang Y, Liang D, Spitz MR, Zhang K, Dong Q, Amos CI, Wu X. 2003. XRCC3 genetic polymorphism, smoking, and lung carcinoma risk in minority populations. Cancer 98:1701–1706. Watanabe J, Yang JP, Eguchi H, et al. 1995. An Rsa I polymorphism in the CYP2E1 gene does not affect lung cancer risk in a Japanese population. Jpn J Cancer Res 86:245–248. Watson MA, Stewart RK, Smith GBJ, Massey TE, Bell DA. 1998. Human glutathione S-transferase P1 polymorphisms: Relationship to lung tissue enzyme activity and population frequency distribution. Carcinogenesis 19:275–280. Wattenberg LW. 1987. Inhibitory effects of benzyl isothiocyanate administered shortly before diethylnitrosamine or benzo(a)pyrene on pulmonary and forestomach neoplasia in A/J mice. Carcinogenesis 8:1971–1973. Wei Q, Cheng C, Amos CI, et al. 2000. Repair of tobacco carcinogen-induced DNA adducts and lung cancer risk: A molecular epidemiology study. J Natl Cancer Inst 92:1764–1772. Wei Q, Cheng L, Hong WK, Spitz MR. 1996a. Reduced DNA repair capacity in lung cancer patients. Cancer Res 56:4103–4107. Wei Q, Gu J, Cheng L, et al. 1996c. Benzo[a]pyrene diol epoxide-induced chromosomal aberrations and risk of lung cancer. Cancer Res 56: 3975–3979. Wei Q, Matanoski GM, Farmer ER, Hedayati MA, Grossman L. 1993. DNA repair and aging in basal cell carcinoma: A molecular epidemiology study. Proc Natl Acad Sci USA 90:1614–1618. Wei Q, Shen H, Wang LE, Duphorne CM, Pillow PC, Guo Z, Qiao Y, Spitz MR. 2003. Association between low dietary folate intake and suboptimal cellular DNA repair capacity. Cancer Epidemiol Biomarkers Prev 12:963–969. Wei Q, Spitz MR, Gu J, et al. 1996b. DNA repair capacity correlates with mutagen sensitivity in lymphoblastoid cell lines Cancer Epidemiol Biomarkers Prev 5:199–204. Weston A, Perrin LS, Forrester K, et al. 1992. Allelic frequency of a p53 polymorphism in human lung cancer. Cancer Epidemiol Biomarkers Prev 1:481–483. Whang-Peng J, Kao SC, Lee EC, et al. 1982. A specific chromosome defect associated with human small cell lung cancer. Science 215:181–185. Wiencke JK, Pemble S, Ketterer B, Kelsey KT. 1995. Gene deletion of glutathione transferase theta1: Correlation with induced genetic damage and potential role in endogenous mutagenesis. Cancer Epidemiol Biomarkers Prev 4:253–260. Wiencke JK, Spitz MR, McMillan A, Kelsey KT. 1997. Lung cancer in Mexican-Americans and African-Americans is associated with the wildtype genotype of the NAD(P)H:quinone oxireductase polymorphism. Cancer Epidemiol Biomarkers Prev 6:87–92. Wikman H, Risch A, Klimek F, et al. 2000. hOGG1 polymorphism and loss of heterozygosity (LOH): Significance for lung cancer susceptibility in a Caucasian population. Int J Cancer 88:932–937. Wingo PA, Ries LA, Giovino GA, et al. 1999. Annual report to the nation on the status of cancer, 1973–1996, with a special section on lung cancer and tobacco smoking. J Natl Cancer Inst 91:675–690. Woodson K, Albanes D, Tangrea JA, Rautalahti M, Virtamon J, Taylor PR. 1999. Association between alcohol and lung cancer in the alpha-

658

PART IV: CANCER BY TISSUE OF ORIGIN

tocopherol, beta-carotene cancer prevention study in Finland. Cancer Causes Control 10:219–226. Wu X, Amos CI, Kemp BL, et al. 1998b. Cytochrome P450 2E1 DraI polymorphisms in lung cancer in minority populations. Cancer Epidemiol Biomarkers Prev 7:13–18. Wu X, Delclos GL, Annegers FJ, et al. 1995. A case-control study of wooddust exposure, mutagen sensitivity, and lung-cancer risk. Cancer Epidemiol Biomarkers Prev 4:583–588. Wu X, Gu J, Hong WK, Lee JJ, et al. 1998c. Benzo[a]pyrene diol epoxide and bleomycin sensitivity and susceptibility to cancer of upper aerodigestive tract. J Natl Cancer Inst 90:1393–1399. Wu X, Roth JA, Zhao H, Luo S, Zheng YL, Chiang S, Spitz MR. 2005. Cell cycle checkpoints, DNA damage/repair, and lung cancer risk. Cancer Res 65:349–357. Wu X, Shi H, Jiang H, et al. 1997. Association between chromosome P4502E1 genotype, mutagen sensitivity, cigarette smoking and susceptibility to lung cancer. Carcinogenesis 18:967–973. Wu X, Wen Y, Hong WK, Spitz MR. 1998a. Cytochrome P450 (CYP) 2C9 genetic polymorphism and lung cancer risk among Caucasians, Mexican Americans, and African Americans. Proc Am Soc Prev Oncology 21st Annual Meeting, Bethesda, MD. Wu X, Zhao H, Amos CI, et al. 2002. p53 genotypes and haplotypes associated with lung cancer susceptibility and ethnicity. J Natl Cancer Inst 94:681–690. Wu X, Zhao H, Wei Q, et al. 2003. XPA polymorphism associated with reduced lung cancer risk and a modulating effect on nucleotide excision repair capacity. Carcinogenesis 24:505–509. Wynder EL, Graham EA. 1950. Tobacco smoking as a possible etiologic factor in bronchiogenic carcinoma. A study of six hundred and eighty-four proved cases. JAMA 143:329–346. Wynder EL, Herbert JR, Kabat GC. 1987. Association of dietary fat and lung cancer. J Natl Cancer Inst 79:631–637. Wynder EL, Muscat JE. 1995. The changing epidemiology of smoking and lung cancer histology. Environ Health Perspect 103(Suppl 8):143–148. Xie J, Lesaffre E, Kesteloot H. 1991. The relationship between animal fat intake, cigarette smoking, and lung cancer. Cancer Causes Control 2:79–83. Xie S, Wang Z, Okano M, et al. 1999. Cloning, expression and chromosome locations of the human DNMT3 gene family. Gene 236:87–95. Xu C, Goodz S, Sellers EM, Tyndale RF. 2002b. CYP2A6 genetic variation and potential consequences. Adv Drug Deliv Rev 54:1245–1256. Xu C, Rao YS, Xu B, et al. 2002a. An in vivo pilot study characterizing the new CYP2A6*7, *8, and *10 alleles. Biochem Biophys Res Commun 290:318–324. Xu LL, Liu G, Miller DP, et al. 2002. Counterpoint: The myeloperoxidase -463G-> A Polymorphism does not decrease lung cancer susceptibility in Caucasians. Cancer Epidemiol Biomarkers Prev 11:1555–1559.

Yamamura K, Kiyohara C, Nakanishi Y, Takayama K, Hara N. 2000. Lung cancer risk and genetic polymorphism at the glutathione S-transferase P1 locus in male Japanese. Fukuoka Igaku Zasshi 91:203–206. Yang CS, Smith TJ, Hong J-Y. 1994. Cytochrome P450 enzymes as targets for chemoprevention against chemical carcinogenesis and toxicity: Opportunities and limitations. Cancer Res 54(Suppl):1982s–1986s. Yang P, Cerhan JR, Vierkant RA, et al. 2002. Adenocarcinoma of the lung is strongly associated with cigarette smoking: Further evidence from a prospective study of women. Am J Epidemiol 156:1114–1122. Yasumori T, Yamazoe Y, Kato R. 1991. Cytochromes P450 human-s (P4502C9) in mephenytoin hydroxylation polymorphisms in human livers: Differences in substrate and stereoselectivities among microheterogeneous P-4502C species expressed in yeasts. J Biochem 109:711–711. Yin L, Pu Y, Liu TY, Tung YH, Chen KW, Lin P. 2001. Genetic polymorphisms of NAD(P)H quinone oxidoreductase, CYP1A1 and microsomal epoxide hydrolase and lung cancer risk in Nanjing, China. Lung Cancer 33: 133–141. Yoshida R, Nakajima M, Watanabe Y, Kwon JT, Yokoi T. 2002. Genetic polymorphisms in human CYP2A6 gene causing impaired nicotine metabolism. Br J Clin Pharmacol 54:511–517. Zhang X, Morera S, Bates PA, et al. 1998. Structure of an XRCC1 BCRT domain: A new protein-protein interaction module. EMBO J 17:6404– 6411. Zhang Y, Talalay P. 1994. Anticarcinogenic activities of organic isothiocyanates: chemistry and mechanisms. Cancer Res 54(Suppl):1976s– 1981s. Zhao B, Seow A, Lee EJ, et al. 2001. Dietary isothiocyanates, glutathione S-transferase -M1, -T1 polymorphisms and lung cancer risk among Chinese women in Singapore. Cancer Epidemiol Biomarkers Prev 10:1063–1067. Zhao H, Spitz MR, Gwyn KM, Wu X. 2002. Microsomal epoxide hydrolase polymorphisms and lung cancer risk in non-Hispanic whites. Mol Carcinog 33:99–104. Zheng YL, Loffredo CA, Yu Z, et al. 2003. Bleomycin-induced chromosome breaks as a risk marker for lung cancer: A case-control study with population and hospital controls. Carcinogenesis 24(2):269–274. Zhou W, Liu G, Miller DP, et al. 2002. Gene-environment interaction for the ERCC2 polymorphisms and cumulative cigarette smoking exposure in lung cancer. Cancer Res 62:1377–1381. Ziegler RG, Mason BJ, Stemhagen A, et al. 1984. Dietary carotene and vitamin A and risk of lung cancer among white men in New Jersey. J Natl Cancer Inst 73:1429–1434. Zienolddiny S, Campa D, Lind H, Ryberg D, Skaug V, Stangeland L, Phillips DH, Canzian F, Haugen A. 2005. Polymorphisms of DNA repair genes and risk of non-small cell lung cancer. Carcinogenesis (e-pub Sept 2005).

34

Pleural and Peritoneal Neoplasms PAOLO BOFFETTA AND LESLIE T. STAYNER

P

rimary malignant neoplasms of the pleura and peritoneum originate from the mesothelial cells that line the respective cavities. The majority of these tumors are mesotheliomas; rare forms include lymphomas, synovial sarcomas, solitary fibrous tumors, and calcifying tumors (Churg et al., 2004). Mesothelioma is a relatively rare but very severe neoplasm, the pleura being the most common organ of origin, followed by the peritoneum. Mesothelioma may also very rarely develop from the pericardium, the tunica vaginalis of the testis, and the ovary. Symptoms are nonspecific and appear late in the development of the disease. A biopsy is usually necessary to establish the diagnosis, which in many cases represents a pathological and clinical challenge. Most tumors are diagnosed at advanced stage. Various treatment modalities, including radical surgery, chemotherapy, and radiation therapy, are used, but survival remains poor. Since the late 1950s, cases of pleural mesothelioma have been reported in miners from South Africa and American workers exposed to asbestos (Wagner et al., 1960; Mancuso and Coulter, 1963; Selikoff et al., 1964). As early as 1964, the causal link between exposure to asbestos and development of mesothelioma in humans was recognized by international panels (UICC, 1965). Use of asbestos started to decline in the 1970s; however, the incidence of pleural and peritoneal mesothelioma is still increasing in most countries in which reliable data are available. The strong causal role of asbestos, the rarity of the disease in populations not exposed to asbestos, and the diagnostic difficulties of mesothelioma complicate the epidemiology of this neoplasm, since knowledge of asbestos exposure might influence diagnosis. Studies based on autopsy series revealed that a sizable proportion of mesotheliomas may remain undiagnosed (e.g., 45% in a series of male cases from Trieste, Italy (Delendi et al., 1991).

MOLECULAR PATHOLOGY AND GENETICS Multiple chromosomal deletions have been identified in mesotheliomas, which are suggestive of a multistep carcinogenic process characterized by several genetic alterations. The most common alterations are deletions of 1p21-22, 3p21, 4q, 6q, 9p21, 13q13-14, and 14q (Churg et al., 2004). Other important alterations include monosomy of chromosome 22, losses of 4p and proximal 15q, as well as losses of 17p12, where the p53 gene is located. Loss of heterozygosity has been reported in all these sites. Chromosomal gains are reported less frequently, in particular at 1q, 5p, 7p, 8q, and 15q (Churg et al., 2004). Inactivation of tumor suppressor genes is a frequent finding in mesothelioma, in particular of the CDKN2A/ARF locus, which encodes for the genes p16INK4a and p14ARF (Churg et al., 2004). Homozygous deletions and promoter methylation of this locus are found frequently, resulting in very frequent loss of p16INK4a expression. Mutations and loss of heterozygosity of the NF2 gene, located on chromosome 22, are also common in mesotheliomas. Aberrant promoter methylation occurs frequently at the GPC3 gene. Activation of proto-oncogenes by mutation or amplification is not frequent in mesothelioma, although activation of several genes, such as c-fos, c-jun, met, and notch-1, has been documented in mesothelial cellular systems exposed to asbestos and SV40 (Churg et al., 2004). The mechanism of asbestos-related carcinogenesis in the pleura and peritoneum is not well understood. There is evidence to suggest

the involvement of reactive oxygen species in the mutations induced by asbestos fibers. Fibers have been shown to catalyze the formation of hydroxyl radicals in cell-free systems, and the production of hydroxyl radicals (Maples and Johnson, 1992) may result in the formation of pre-mutagenic DNA bases such as 8-hyroxydeoxyguanosine (8-OHdG). DNA-8-OHdG adducts have been demonstrated to occur in lacI transgenic rats exposed to crocidolite (Unfried et al., 2002). Other possible mechanisms include physical interference with mitosis and stimulation of cell proliferation (Kane, 1996).

DESCRIPTIVE EPIDEMIOLOGY Geographic Patterns The incidence of mesothelioma largely reflects past exposure to asbestos. Despite reports of cases in children (Cooper et al., 1989), the incidence is very low until the age of 50, due to the long latency between exposure and development of the disease. In most populations, pleural mesothelioma incidence rates are substantially higher among males than among females, which is due to the higher frequency of occupational exposure to asbestos among men. For example, in 1999 among whites in the United States the rate of mesothelioma was 2.3 per 100,000 among men and 0.4 per 100,000 among women. In the United States, mesothelioma incidence rates also appear to be somewhat higher among whites than among nonwhites (1.2 vs. 0.6 per 100,000 in 1999) (Ries et al., 2003), most likely reflecting ethnic differences in employment patterns. There are large differences in the rates of pleural mesothelioma in different parts of the world, which likely reflect past asbestos exposure, but might be due in part also to differences in diagnostic accuracy. Rates are low (typically up to 0.5 per 100,000 in men and up to 0.2 per 100,00 in women) in developing countries from which reliable data are available. Figure 34–1 shows incidence rates in men from selected developed countries: rates can be as high as 5 per 100,000 in areas with high past asbestos exposure, such as Genoa in Italy and Western Australia. Rates in women in developed countries are mostly between 0.1 and 0.5 per 100,000 (Fig. 34–2); however, they can be higher in selected areas (e.g., 1.2 per 100,000 in Genoa, Italy). The geographic patterns of peritoneal mesothelioma parallel that of the pleural form of the disease, although rates are consistently lower. However, while in high-risk, developed countries the ratio between pleural and peritoneal mesotheliomas is in the order of 10– 30 : 1, in low-risk countries this ratio is 3–10 : 1, suggesting that heavy exposure to asbestos increases predominantly the risk of pleural mesothelioma. The higher mesothelioma rates in North America, Europe, and Australia, as compared with Africa, South America or Asia, presumably reflect the longer history of use of asbestos in the industrialized countries. National incidence or mortality rates, however, mask the very uneven distribution of the disease. Blue-collar workers experience by far the highest incidence in most countries, reflecting the overwhelming importance of occupational exposure to asbestos in causing this disease. In addition, very high incidence has been reported in regions with heavy environmental exposure to asbestos or similar fibers. For example, in several villages in Central Turkey heavy environmental exposure to the zeolite fiber, erionite, has caused an epidemic of pleural and peritoneal mesothelioma.

659

660

PART IV: CANCER BY TISSUE OF ORIGIN New Zealand England Sweden Lithuania Denmark Croatia Israel Hong Kong USA (SEER-W) Canada

Figure 34–1. Mesothelioma incidence rates (ca. 1995) in selected countries—men.

0

Temporal Trends Although the production and commercial use of asbestos has largely ceased in the United States and Western Europe, the incidence of mesothelioma appears to have either leveled off or continues to rise in most of these countries because of the long latency of the disease (20–45 years). A number of studies have developed models of mesothelioma mortality or incidence as a function of age and birth cohort, which have been used to predict future occurrence of the disease. Analyses of data from several Western European countries (Peto et al., 1995; La Vecchia et al., 2000; Peto et al., 1999; Banaei et al., 2000; Kjaergaard and Andersson, 2000) have predicted a peak of pleural mesothelioma deaths occurring between 2010 and 2020. Countries in which no increasing trend in pleural mesothelioma incidence or mortality has been observed in the past decades include Austria: this pattern may be explained by patterns of use (Neuberger and Vutuc, 2003). An analysis of mesothelioma incidence published in 1997 in the United States suggested that the peak number of incident mesothelioma cases would have occurred in the year 2000 (Price, 1997), and in fact mesothelioma incidence does appear to have leveled off in the late 1990s in the United States (Ries et al., 2003). Pleural mesothelioma incidence in Sweden appears to have peaked after 1995 (Hemminki and Li, 2003b). The fact that mesothelioma incidence has declined earlier in Sweden and the United States than in other countries in Western Europe would appear to reflect the fact that the major

0.5

1

1.5

2

2.5

3

uses of asbestos declined earlier in these countries due to regulatory actions. Caution is required in interpreting predictions of future trends in mesothelioma incidence; for example, a recent analysis from the Netherlands of mortality data from 1969–1998 resulted in predictions for the period 2000–2028 that were 44% lower than those from a previous analysis based on data from 1969–1993 (Segura et al., 2003). Relatively little data are available on trends in mesothelioma incidence or mortality in countries outside of Western Europe, Australia, and North America. The number of mesothelioma cases in Japan has risen dramatically in the 1990s, following an important increase in the imports of asbestos in the 1970s (Murai, 2001; Morinaga et al., 2001); since the asbestos consumption curve for Japan has lagged behind that in Europe and North America, the full impact on mesothelioma incidence has not yet been observed (Takahashi et al., 1999).

Survival Based on US statistics from 1992–1999, the 5-year survival for pleural mesothelioma is 7%. Survival appears to be worse for males than females (5.2% vs. 13.9% respectively in the United States between 1992 and 1999) (Ries et al., 2003). Survival data from Europe are comparable to those from the United States: 5-year relative survival for cases diagnosed during 1985–1989 in 17 European countries was 6.7% in men and 7.9% in women (Berrino, 1999). It has been suggested

New Zealand England Sweden Lithuania Denmark Croatia Israel Hong Kong USA (SEER-W) Canada

Figure 34–2. Mesothelioma incidence rates (ca. 1995) in selected countries—women.

0

0.1

0.2

0.3

0.4

0.5

Pleural and Peritoneal Neoplasms Table 34–1. Selected Early Series of Pleural Mesothelioma Cases with Information on Asbestos Exposure

Reference Sleggs et al. (1961) Anspach (1965) Elmes et al. (1965) Hagerstrand et al. (1968)

S. Africa Former DDR Northern Ireland Sweden

N Patients

N with Asbestos Exposure

N with Lung Fibrosis

34 22 42 35

33 19 32 35

11 10 NA 3

Source: Data from Fletcher, 1972.

that non-asbestos–related cases may have better survival than those caused by asbestos, but there is little evidence to support this hypothesis (Law et al., 1984). A model based on gene expression data from oligonucleotide microarrays has been shown to be highly predictive of mesothelioma survival in one study (Gordon et al., 2003). This appears to be a promising approach with clear implications for clinical treatment, which needs further research.

Risk Factors Asbestos The link between asbestos exposure and mesothelioma development was first demonstrated in the 1960s. As early as 1973, the International Agency for Research on Cancer (IARC) confirmed the causal nature of association between asbestos exposure and mesothelioma; at that time, a substantial number of reports of high prevalence of asbestos exposure among mesothelioma patients were available from various countries (Table 34–1). An increased risk of mesothelioma has been convincingly shown in many occupational groups exposed to asbestos, such as miners, insulation workers, manufacturers of cement, textiles, and other asbestos-based products and shipyard workers. However, the widespread use of asbestos has caused important exposure in many industries, and cases of asbestos-related pleural mesothelioma have been reported among workers in very diverse trades, such as thermoelectric power plants (Crosignani et al., 1995), oil refining (Tsai et al., 1996), textile production (Paci et al., 1987; Paci et al., 1991), pulp and paper production (Jarvholm et al., 1988; Langseth and Andersen, 2000), petroleum industry (Gennaro et al., 2000), cigarette filter manufacture (Talcott et al., 1989), and railroad industry (Ohlson et al., 1984). In many Western countries, the classic circumstances of exposure to asbestos are nowadays of less importance, because of the ban on most if not all uses of asbestos and precautions taken when exposure is known. The greatest exposure is likely to occur among maintenance and construction workers (Koskinen et al., 2003). In many less developed countries, on the other hand, high levels of exposure are still prevalent in many industries (Algranti, 1998; Harris and Kahwa, 2003).

Industry-Based Studies. Table 34–2 reports the results of selected studies of cohorts of workers exposed to asbestos. Given the large body of evidence available, only studies of occupational groups primarily exposed to asbestos have been included. When interpreting the results of this table, one should consider that the estimate of the magnitude of the risk of pleural mesothelioma following asbestos exposure based on standardized mortality ratios (SMR) of pleural neoplasms or similar measures can suffer from a number of biases, as discussed in Table 34–3. Although some of these biases are common to studies of other exposures and diseases, the likelihood of bias is particularly important in the case of asbestos and mesothelioma because of the strength of the association and the possibility that diagnostic accuracy depends on knowledge of exposure. Because of the possible biases, we did not report the SMR or similar measures of mesothelioma risk in Table 34–2; rather, we listed the number of pleural and peritoneal mesotheliomas, which allow us to calculate the proportion

661

of mesothelioma deaths over total deaths. Furthermore, only populations with at least 100 deaths observed are included in the table, to reduce the random variability of the results. In some of the studies listed, the study population was defined according to presence of asbestosis rather than employment in a given industry; by definition these individuals primarily developed their disease as a consequence of occupational exposure to asbestos. One or more cases of pleural mesothelioma have been reported in all but 2 of the 51 populations listed in Table 34–2. The proportion of pleural mesothelioma compared to total deaths was 1% or more in 28 of 46 populations in which this ratio could be measured. A strong correlation was present between the percentage of mesotheliomas over total deaths and the SMR of lung cancer (correlation coefficient 0.57, P value 0.0001, Fig. 34–3).

Effect of Different Asbestos Fibers. Workers exposed to amphibole asbestos, including in particular crocidolite and amosite, experienced a higher risk of mesothelioma than workers exposed predominantly to the most widely used type of asbestos, chrysotile. The proportion of mesothelioma deaths over the total was lower in the cohorts of workers classified as exposed to pure chrysotile than in the other cohorts (Table 34–4 and Fig. 34–3). The difference in the proportion of total deaths as mesotheliomas was not significant between studies of workers exposed only to chrysotile or predominantly to chrysotile (P = 0.15), nor between studies of workers exposed to amphiboles or to mixed and unknown fibers (P = 0.8). However, the difference between studies of chrysotile vs. amphiboles/mixed/ unknown exposure was significant (P = 0.004). It is a matter of debate, however, whether exposure to pure chrysotile entails a risk of mesothelioma, or whether the relatively small risk detected in workers classified as exposed to chrysotile can be attributed to low-level contamination by (or concomitant exposure to) amphiboles (McDonald et al., 1989; Stayner et al., 1996; McDonald et al., 1997; Egilman et al., 2003). Studies of lung fiber burden have shown that crocidolite and amosite persist for a longer period in the lung than chrysotile (Churg and Wright, 1994). This finding might help to explain the lower risk of mesothelioma after inhalation of chrysotile compared with amphiboles, and suggests that risk of pleural neoplasms decreases more rapidly after cessation of exposure to the former type of asbestos. However, studies of pleural fiber burden also show a greater concentration of chrysotile fibers than amphiboles in the pleura (Suzuki and Yuen, 2001), and in the peripheral areas of the lung (Sebastien et al., 1975), likely reflecting the type of fiber workers were predominantly exposed to. Given the contamination of most commercially available chrysotile by amphiboles, and notably that of Canadian chrysotile by tremolite fibers, data from good quality studies on cancer risk among asbestos workers for whom amphibole exposure can be excluded with certainty are not available. Shape of Dose-Response Relationship. A quantitative relationship between mesothelioma risk and asbestos exposure can be derived from the occupational cohorts with good exposure data and sufficient latency. A widely accepted model involves a power function of time since first exposure and time since cessation of exposure of the form: I(t) = k * E * [(t - t1)n - (t - t2)n] where I(t) is the incidence of mesothelioma at time t caused by exposure at constant level E (expressed in fb/ml) starting at time t1 and ending at time t2 (Peto et al., 1985; HEI, 1991), k is a constant expressing the carcinogenic potency on the pleura, which is specific to industry and type of asbestos fiber, and n is 3 or 4. The formula assumes that the excess is equal to the total incidence; that is, no mesothelioma cases or deaths are expected without exposure. In the case of multiple exposure periods at different levels, the overall incidence will be: l( t ) = k * Â Ei Î( t - t1i )n - ( t - t 2 i )n ˚ i

Table 34–2. Results of Selected Cohort Studies of Mesothelioma Following Asbestos Exposure

Reference Jones et al. (1980) Rossiter and Coles (1980) Acheson et al. (1982) McDonald et al. (1982) Thomas et al. (1982) Finkelstein (1984) McDonald et al. (1984) Acheson et al. (1984) Ohlson et al. (1984) Peto et al. (1985) Ohlson and Hogstedt (1985) Kolonel et al. (1985) Alies-Patin and Valleron (1985) Newhouse et al. (1985) SzeszeniaDabrowska et al. (1986) Woitowitz et al. (1986) Seidman et al. (1986) Gardner et al. (1986) Hodgson and Jones (1986) Hughes et al. (1987) Amandus and Wheeler (1987) Enterline et al. (1987) Armstrong et al. (1988) Newhouse and Sullivan (1989) Raffn et al. (1989) Albin et al. (1990)

662

Exposure Circumstance

Asbestos type

Country

Period of Diagnostic employment Evidence

Sex

Gas mask P Cr manufacture Shipyard NA

UK

1938–1945

MR

F

UK

1947

BE

M

Gas mask Ch manufacture Cr Textile P Ch product manufacture Cement P Ch workers Cement Mix workers Friction Ch product manufacture Insulation Am manufacture

UK

1939

DC

F

UK USA

1939 1938–1958

DC DC

UK

1936–1977

Canada

Railroad Mix repair work Textile P Ch product manufacture Cement P Ch workers

Size of Cohort

Number of Lung Total Cancer Deaths Deaths

SMR of Lung Cancer Deaths†

Number of Pleura Mesothelioma Deaths

Number of Peritoneal Mesothelioma Deaths

578

166

12

[1.94*]

13

4

6 292

1 043

84

0.70*

31

NA

570

177

6

[1.25]

1

0

F M

757 4 137

219 1 392

13 NA

[2.10*] NA

3 10

2 4

DC

M

1 592

351

30

0.93

2

0

1955–1959

DC

M

535

108

26

4.90*

11

8

USA

1939–1958

DC

M

3 641

1 267

73

1.49*

0

0

UK

1945–1978

DC

M

4 820

333

57

[1.96*]

4

1

Sweden

1939–1980

DC

M

3 442

925

37

1.16

5

0

UK

1933–1974

DC

M

3 211

1 113

132

1.31*

10

1

Sweden

1943–1976

DC

M

1 216

220

11

1.23

0

0

Shipyard

Mix

USA

1950–1969

CR

M

5 191

668

61

[1.09]

8

0

Cement workers

Mix

France

1940–1977

DC

M

1 506

206

12

2.17*

3

1

Mix

Mix

UK

1933–1964

DC DC

M F

4 695 932

818 274

158 37

[2.50*] [7.40*]

31 14

NA NA

Mix

P Ch

Poland

1945–1985

DC

M

2 403

527

35

1.41

1

0

Mix

Mix

Germany 1930–1974

DC

PM

3 070

185

22

1.44

6

0

USA

1941–1945

BE

M

820

593

102

4.97*

8

9

UK

1941–1983

DC

MF

2 167

486

41

0.97

1

0

UK

1969–1981

DC

M

31 565

1 128

186

1.26*

35

NA

NA

9

1

Various Am product manufacture Cement Ch workers Mix Mix Cement workers Vermiculite miners

P Ch

USA

1937–1970

BE

M

6 931

2 143

NA

Tre, Act

USA

1970–1981

DC

M

569

161

20

2.23*

2

0

Mix

Mix

USA

1941–1967

DC

M

1 074

617

77

2.71*

6

2

Crocidolite miners

Cr

Australia 1943–1966

DC

PM

6 916

843

93

2.64*

33

1

UK

1941–1979

DC

M

8 404

2 055

240

1.08

11

NA

Denmark 1928–1984

CR

M

7 996

1 305

162

1.80*

10

NA

Sweden

CR

M

1 465

592

27

NA

12

NA

Friction P Ch product manufacture Cement Mix workers Cement P Ch workers

1907–1977

Table 34–2. (cont.)

Reference Neuberger and Kundi (1990) Piolatto et al. (1990) Hilt et al. (1991) Selikoff and Seidman (1991) Sanden et al. (1992) Sluis-Cremer et al. (1992) Menegozzo et al. (1993) Dement et al. (1994) Giaroli et al. (1994) Meurman et al. (1994) Magnani et al. (1996) Zhu and Wang (1996) Liddell et al. (1997) Oksa et al. (1997) SzeszeniaDabrowska et al. (1997) Battista et al. (1999) Germani et al. (1999) Tulchinsky et al. (1999) Puntoni et al. (2001) Yano et al. (2001) SzeszeniaDabrowska et al. (2002) Ulvestad et al. (2002)

Number of Lung Total Cancer Deaths Deaths

SMR of Lung Cancer Deaths†

Number of Pleura Mesothelioma Deaths

Number of Peritoneal Mesothelioma Deaths

Exposure Circumstance

Asbestos type

Country

Period of Diagnostic employment Evidence

Sex

Size of Cohort

Cement workers

P Ch

Austria

1950–1981

DC

NA

2 816

540

50

1.7*

7

3

Miners

Ch

Italy

1946–1987

DC

M

1 058

427

22

1.1

2

NA

Electrochemical Insulation workers

NA

Norway

NA

CR

NA

287

186

18

3.16*

6

NA

Mix

USA

1967

BE

M

17 800

4 951

1 008

3.75*

173

285

Shipyard

P Ch

Sweden

1977–1979

DC

M

3 893

NA

22

0.85

11

0

Miners

Am, Cr

S Africa

1945–1981

BE

M

7 317

1 225

63

1.72*

22

6

Railroad Mix constr. Work

Italy

1970–1989

DC

M

1 543

194

28

1.45*

3

2

Textile Ch product manufacture Cement P Ch workers Miners Antho

USA

1940–1965

DC

MF

3 022

1 259

88

1.97*

2

0

Italy

1952–1987

DC

NA

3 341

274

33

1.24

5

NA

Finland

1953–1967

CR

M

736

NA

76

2.88*

3

1

Italy

1950–1980

DC

M

2 605

1 147

162

2.48*

53

23

China

1972–1981

DC

M

NA

260

34

NA

1

NA

China Canada

1972–1981 1902–1971

DC DC

M M

NA 10 918

240 8 009

24 646

NA 1.37*

3 38

NA NA

Sprayers, Mix asbestosis p. Cement Mix workers

Finland

1955–1976

CR

PM

247

NA

43

[11*]

8

NA

Poland

1945–1980

CR

M

3 405

473

41

0.97

5

NA

Railroad workers

Mix

Italy

1945–1969

DC

M

734

199

26

1.24

5

NA

Asbestosis patients

Mix

Italy

1979‡

DC

F

631

277

16

4.83*

14

12

Cement workers

Mix

Israel

1953–1992

BE

M

3 057

NA

28

1.35

20

1

Ship NA construction repair Various Ch product manufacture Asbestosis Mix patients

Italy

1960–1981

DC

M

3 984

2 376

298

1.77*

60

NA

China

1972

MR

M

515

132

26

NA

1

1

Poland

1970–1997‡

DC

M

907

300

39

1.68*

3

0

Cement workers

Norway

1942–1976

CR

M

541

NA

33

3.1*

18

0

Cement workers

P Ch

Various Mix product manufacture Miners Ch Miners Ch

P Ch

*P < 0.05; †SMR are in square brackets if calculated from raw data; ‡Period of diagnosis. Act, actymolite; Am, amosite; Antho, anthophyllite; BE, best evidence; Ch, chrysotile; Cr, crocidolite; CR, cancer registry; DC, death certificate; F, females; M, males; MF, males and females; Mix, mixed exposure; MR, medical records; NA, not available; PCh, predominantly chrysotile; PM, predominantly males; SMR, standardized mortality ratios; Tre, tremolite.

663

664

PART IV: CANCER BY TISSUE OF ORIGIN

Table 34–3. Possible Sources of Bias in Quantifying Asbestos Carcinogenicity Based on Standardized Mortality Ratios of Pleural Neoplasms Source of Bias

Consequence

Rarity of the disease in the absence of asbestos exposure Poor sensitivity of disease assessment Poor specificity of disease assessment

% mesothelioma deaths

Poor sensitivity of exposure assessment

Effect on Risk Estimate

Lack of truly unexposed (reference) groups, since most mesotheliomas in the reference population occur in individuals exposed to asbestos Mesotheliomas classified as lung cancer or other neoplasms Inclusion of neoplasms not related to asbestos (e.g., mediastinal tumors, lymphomas) Misclassification of exposure (also in internal analyses)

Underestimate of the effect Possible overestimate of the effect (knowledge of exposure may influence diagnosis) Underestimate of the effect Underestimate of the effect (most likely)

Table 34–4. Proportion (%) of Mesotheliomas over Total Deaths by Type of Asbestos Fibers

12 10 8

Type of Asbestos

6 4

Pure chrysotile Predominantly chrysotile Amphiboles Mixed, unknown, other

2 0

N Studies*

Mean

Standard Deviation

8 11 6 21

0.49 1.19 2.91 2.64

0.39 1.30 2.61 2.23

*Studies listed in Table 34–2.

0

1

2

3

4

5

SMR of lung cancer Chrysotile Amphiboles

Predominantly chrysotile

Figure 34–3. Standardized mortality ratios of lung cancer and percentage of mesothelioma deaths of total deaths in selected occupational cohorts of asbestos-exposed workers, by fiber type.

where each ith period of exposure starts at time t1i and ends at time t2i. The model can be refined by applying a lag of 10 years. An analysis of the incidence of mesothelioma based on four cohorts with detailed information on time since first exposure and level of exposure resulted in the estimates of k reported in Table 34–5 (Nicholson, 1986; HEI, 1991). The lowest value is found in a cohort of asbestos textile workers exposed mainly to chrysotile (Peto et al., 1982); the value is intermediate for a cohort of American insulators exposed to both chrysotile and amphiboles (Selikoff et al., 1979); and it is highest in workers exposed to insulation material consisting only of amosite (Seidman et al., 1979). In the case of the fourth cohort, from Canada, a very high Km was estimated, despite the predominant exposure to chrysotile (Finkelstein, 1983); the assessment of exposure in this cohort, however, might suffer from systematic error (HEI, 1991). The issues related to quantitative estimate of mesothelioma risk from asbestos exposure have been reviewed (HEI, 1991; Peto et al., 1985).

Community-Based Studies. The strongest evidence of the risk of pleural mesothelioma after occupational exposure to asbestos

comes from industry-based studies, as reviewed above. In addition, several studies, mainly of case-control design, have been conducted in populations not selected for specific occupational exposures; while these investigations can suffer from selection and information bias, they are useful to identify the main industries and occupation at risk of mesothelioma in different populations, and to estimate the proportion of cases without recognized asbestos exposure. Table 34–6 summarizes these studies. The proportion of mesothelioma cases exposed to asbestos in the workplace varies according to the study population and the sensitivity of the method used to estimate exposure; in most studies, however, this proportion is in the range 60%–75%. In two studies, a detailed assessment of employment circumstances has led to a quantitative estimate of the risk following asbestos exposure (Iwatsubo et al., 1998; Rodelsperger et al., 2001). In both studies, a linear dose-response relationship has been derived, with a small but detectable increase in mesothelioma risk below a cumulative exposure of 1 fiber/ml-year, which is compatible with exposure limits currently implemented in many countries. However, caution should be used in interpreting these results since the level of exposure was estimated retrospectively by industrial hygienists, possibly resulting in quantitative underestimate of past exposure, which in turn would lead to an overestimate of the dose-response relationship (Siemiatycki and Boffetta, 1998).

Risk in Carriers of Pleural Plaques. Pleural plaques are characteristic patches of the parietal pleura. They represent the most common lesion found in individuals exposed to asbestos; they are asymptomatic and are detected radiologically. Although pleural plaques have been for a long time considered only a marker of past

Table 34–5. Estimated Km Values in Selected Cohorts (Nicholson, 1986; HEI, 1991) Country Reference United Kingdom, Peto et al. (1982) Unites States, Selikoff et al. (1979) United States, Seidman et al. (1979) Canada, Finkelstein (1983)

Industry

Fiber Type

N Observed Mesotheliomas

Average Exposure (fb/ml)

Km * 10-8

Textile workers Insulators Product manufacture Asbestos cement

P Ch Ch, Am Am P Ch

18 170 14 11

20 15 35 9

1.0 1.5 3.2 12.0

Source: Nicholson (1986); HEI (1991). Am, amosite; Ch, chrysotile; P Ch, predominantly chrysotile.

665

Pleural and Peritoneal Neoplasms Table 34–6. Community-Based Studies of Pleural Mesothelioma and Occupational Exposure to Asbestos

Reference

Population

Design

Exposure Assessment

N Cases

% Cases with Occupational Exposure to Asbestos

Iwatsubo et al. (1998)

France, 1987–1993

HCC

EE

405

71

McDonald et al. (2001) Rees et al. (1999)

United Kingdom, 1990–1996 South Africa, 1988–1990 Sweden, 1961–1979

CS*

JH

128

NA

HCC

EE

123†

96‡

RL

Census

318

NA

Rodelsperger et al. (2001)

Germany, 1988–1991

PCC

EE

125

91

Agudo et al. (2000)

Spain, 1993–1996

PCC

EE

132

61

Chellini et al. (1992) Cicioni et al. (1991)+

Italy, 1970–1988 California, USA, 1972–1988 Finland, 1985–1988

CS PCC

JEM EE

100 101

72 36

CS

EE

23

65

Malker et al. (1985)+

Tuomi et al. (1991)

Comments Linear relationship between asbestos exposure and mesothelioma risk at all doses Increased risk <1 f/ml-yr Risk in construction industry No cases with pure chrysotile exposure Highest risk from crocidolite mining Risk in sugar refining, cellulose, wood and pulp, shipbuilding, railroad manufacture High risk in metal and vehicle production Linear relationship between asbestos exposure and mesothelioma risk at all doses Increased risk <1 f/ml-yr Risk in machinery fitters, electricians, asbestos workers — Low sensitivity of exposure assessment because of poor job classification —

In all studies cases are identified from medical records except for + (cancer registry). *Comparison with census data; study restricted to subjects born > 1943; †Including seven peritoneal mesothelioma; ‡Including environmental exposure. CS, case-series; EE, expert evaluation; HCC, hospital-based case-control study; JEM, job-exposure matrix; PCC, population-based case-control; study; RL, record linkage.

asbestos exposure (Weiss, 1993), an increased risk of mesothelioma has been shown in several series of carriers. In an early study of shipyard workers from the United Kingdom followed up between 1961 and 1970, the cumulative incidence of mesothelioma was 3 of 408 carriers of plaques and 0 of 404 non-carriers (P = 0.08) (Fletcher, 1972). In an autopsy-based study from Italy, Bianchi and colleagues (1997) calculated an odds ratio of mesothelioma for presence of plaques equal to 12.7 (95% CI: 1.71–7.94) in men and 7.59 (95% CI: 1.71–45.6) in women, and a relationship between mesothelioma risk and size of the lesion. In a prospective study, the incidence of mesothelioma was compared between 1569 Swedish pleural plaque carriers and the national population, resulting in a standardized incidence ratio of 11.3 (95% CI: 5.13–21.3). While pleural plaques should be considered markers of mesothelioma risk, it is unclear whether they simply reflect a particularly high exposure, or they are a marker of individual susceptibility to both pleural reaction and cancer development following exposure to asbestos. An important problem in the interpretation of results of studies of pleural plaques is the poor sensitivity and specificity of their diagnosis based on imaging (Svenes et al., 1986).

Risk from Non-Occupational Exposures. In contrast to the many epidemiological studies available on asbestos-exposed workers, there are relatively few studies of the health effects of nonoccupational (household and residential) exposure to asbestos. One type of household exposure concerns cohabitants of asbestos workers and arises from dust brought home on clothes. Other household sources of asbestos exposure are represented by the installation, degradation, removal, and repair of asbestos-containing products. Residential exposure mainly results from outdoor pollution related to asbestos mining or manufacturing, in addition to natural exposure from the erosion of asbestos or asbestiform rocks. The assessment of non-occupational exposure to asbestos presents difficulties since levels are generally low, and the duration and frequency of exposure and the type of fiber are seldom known with precision. Several authors have reported cases of mesothelioma among family members of asbestos workers (e.g., Newhouse and Thompson, 1965; Anderson et al., 1976; Vianna and Polan, 1978; Dodoli et al., 1992). In particular, household exposure might occur via handling and

washing asbestos-contaminated clothes. Although a formal estimate of risk has been seldom calculated (e.g., RR = 10, 95% CI: 1.2–37 for hand-washing clothes in a case-control study in New York State) (Vianna and Polan, 1978), there is no doubt that cases of mesothelioma have occurred among family members of asbestos workers that are attributable to household exposure. Table 34–7 summarizes the studies on risk of pleural mesothelioma from residential exposure to asbestos. Such studies were available from different countries of the world; in most cases, exposure was defined as residence near a mine or another major source of asbestos exposure. Although cases with known occupational or household exposure have been excluded from these studies, the possibility of unrecognized exposure remains. The risk of mesothelioma was greatly increased in all studies listed in Table 34–7. Although there is strong evidence that environmental exposure to asbestos may increase mesothelioma risk, it is unclear what burden of the disease is caused attributable to this exposure circumstance. According to a model used by World Health Organization (1987), 5% of the European population experience residential exposure to asbestos. A recent meta-analysis estimated the relative risk of mesothelioma from environmental exposure to asbestos equal to 3.5 (95% CI: 1.8–7.0) (Bourdes et al., 2000); these results lead to an estimated number of 425 mesotheliomas in men and 56 in women occurring yearly in the European Union. However, the figure of 5% might overestimate the prevalence of exposure to circumstances comparable with those investigated in the studies listed in Table 34–7. A conservative estimate of 1% of exposed population leads to estimates of 92 mesotheliomas in men and 12 in women. It should be stressed that in specific areas such as Cyprus [tremolite (McConnochie et al., 1987)], Northern Greece [tremolite (Langer et al., 1987)], and Eastern Sicily [tremolite-actinolite amphibole (Paoletti et al., 2000)], in addition to those listed in Table 34–7, the prevalence of heavy environmental exposure is relatively high leading to a substantial burden of cancer.

Risk of Peritoneal Mesothelioma. Results for peritoneal mesothelioma were reported for 30 of the occupationally exposed populations listed in Table 34–2. In 14 of them, no cases were reported, and in none did peritoneal mesotheliomas represent more than 1% of total deaths. A strong correlation is present between percentage of

666

PART IV: CANCER BY TISSUE OF ORIGIN

Table 34–7. Studies on Risk of Mesothelioma from Environmental Exposure to Asbestos Country

SD

TF

South Africa* South Africa* Canada

Ec Co Co

A A C

Canada United States Italy

Ec Ec Co

C UM UM

Italy Italy, Spain, Switzerland United Kingdom

CC CC

UM UM

CC

UM

United Kingdom

CC

UM

China New Caledonia

Co CC

A A

Australia

CC

A

Source of Exposure

Ca

RR

95% CI

Reference

R. in mining area R. in mining area R. in mining area

61 8 7

8.7 26.7 1.3

6.7–11.4 11.5–52.5 0.5–3.0

R. in mining area R. in polluted city R. <1 km from asbestos cement plant R. in polluted city R. <2 km from potential source R. <0.5 miles from asbestos factory R. <0.5 km from potential source R. in polluted area Use of contaminated building materials R. >5 years in mining area‡

7 24 36

7.6 [2.0 expected] 6.6

3.4–14.9 4.1–11

Botha et al. (1986)* Kielkowski et al. (2000)† Theriault and Grand-Bois (1978)* Camus et al. (1998)† Berry (1997) Magnani et al. (1995)*

32 17

14.6 11.5

4.9–43.1 3.5–38.2

Magnani et al. (2001)* Magnani et al. (2000)

11

5.4

1.8–17

5

6.6

0.9–50

Newhouse and Thompson (1965) Howel et al. (1997)

70 14

• 40.9

NA 5.1–325

Luo et al. (2003) Luce et al. (2000)*

14

6.7

2.0–22.2

Hansen et al. (1998)

CC, case-control study; Co, cohort study; Ec, ecological study; SD, study design A, amphiboles; C, chrysotile; Ca, number of exposed cases; CI, confidence interval; R., residence; exp, expected cases; Ref, reference; Numbers in square brackets are weighted averages of gender-specific results; RR, relative risk; TF, predominant type of fiber; UM, unspecified and mixed *Results derived from raw data reported in the publication; †Women only ‡ Compared with residence <1 year; +Partially overlapping studies.

deaths from pleural and peritoneal mesothelioma (correlation coefficient 0.8, P < 0.0001). Studies of workers exposed only or predominantly to chrysotile resulted in lower mean percentage of total deaths from peritoneal mesothelioma than other studies (0.29% ± 0.57 vs. 1.57% ± 2.28, P = 0.06). In all studies with adequate number of cases, a strong association has been found between occupational exposure to asbestos and risk of peritoneal mesothelioma (Jones et al., 1980; Finkelstein, 1984; Seidman et al., 1986; Enterline et al., 1987; Neuberger and Kundi, 1990; Selikoff and Seidman, 1991; Sluis-Cremer et al., 1992; Magnani et al., 1996; Germani et al., 1999). In a study based on death certificates from 24 of the United States during 1984–1992, 657 deaths from peritoneal neoplasms were identified (Cocco and Dosemeci, 1999). An increased risk was found among men employed in the same occupations and industries that entail a risk of pleural mesothelioma, such as insulators and construction workers; results among women were hampered by small numbers. A relationship was found between peritoneal neoplasm risk and probability and intensity of exposure to asbestos as estimated with a job-exposure matrix.

Trends and Patterns of Asbestos Use. Between the 1980s and the 1990s, asbestos use has either ceased or decreased to a very small amount in United States and in Western Europe. On a global scale asbestos production in 2000 was less than half of that of 1975,

when it peaked above 5,000,000 tons (Fig. 34–4). However, the decline in production has slowed down since the late 1990s. Figure 34–5 shows the production of asbestos in 2000 by country, and Figure 34–6 reports the figures of consumption in the top 10 countries. Canada is the only developed country still producing a large amount of asbestos, and Japan is the only developed country in which consumption is still high. In addition to Russia and other former Soviet Republics, asbestos use has become an issue in developing countries, in particular those with the strongest industrial infrastructure, such as China, Thailand, and Brazil. Little data are available on industrial hygiene conditions and occupational and environmental exposure levels in developing countries (Algranti, 1998).

Fibers Other than Asbestos Erionite. Cases of mesothelioma have been reported in individuals without occupational exposure to asbestos from several areas of Anatolia, Turkey (Baris et al., 1979). In some areas, environmental exposure to asbestos was identified (Yazicioglu et al., 1980; Emri et al., 2002; Metintas et al., 1999). In other regions, however, and in particular in a very high-risk area from Central Cappadocia, no sources of asbestos exposure were identified, while erionite fibers were detected both in lung biopsies and in environmental samples (Baris et

Kazakhstan

Brazil

China

Colombia Russia Zimbabwe Other developing Other developed

Figure 34–4. Worldwide asbestos production, 1900–2000.

Canada

Figure 34–5. Worldwide asbestos producers, 2000.

667

Pleural and Peritoneal Neoplasms

cators). No cases have been reported in a small cohort of workers exposed to refractory ceramic fibers; the strong excess of mesothelioma among hamsters exposed by inhalation to this type of fiber (Mast et al., 1994), however, suggests prudence before concluding that refractory ceramic fibers do not pose a risk to humans. An increased risk of mesothelioma has been reported in sugar refinery workers from Sweden and Italy, which was attributed to exposure to organic fibers (Malker et al., 1983; Maltoni et al., 1995). These findings, however, have not been confirmed in studies such as conducted in Hawaii (Sinks et al., 1992) and Florida (Brooks et al., 1992), and might be due to concomitant exposure to asbestos.

Russia China Brazil India Thailand Japan Indonesia Korea Mexico Belarus

Ionizing Radiation 0

100

200

300

400

500

Tons x 1,000

Figure 34–6. Top asbestos consumers, 2000.

al., 1978; Baris et al., 1981). Epidemiological studies subsequently confirmed a causal role of erionite, a zeolite fiber, in causing pleural and peritoneal mesothelioma as well as lung cancer (Saracci et al., 1982; Baris et al., 1987). In one of the most contaminated villages, Kirian, pleural mesothelioma represented the cause of 21 of 50 deaths (42%) in the period 1979–1983 (Baris et al., 1987). An increased risk of mesothelioma has also been reported in Turkish immigrants from eronite-contaminated areas to Sweden (Metintas et al., 1999). Erionite is classified as human carcinogen (Group 1) by the International Agency for Research on Cancer (1987).

Other Fibers No excess mortality from mesothelioma has been reported among workers employed in the production of man-made vitreous fibers; among almost 14,000 deaths occurring in workers included in the available cohorts, only six were from mesothelioma (Table 34–8). Two of these cases had possible or probable concomitant exposure to asbestos. In two community-based studies, an increased risk of mesothelioma has been reported following estimated exposure to manmade vitreous fibers; after adjustment for asbestos exposure, the ORs were 1.5 (95% CI: 0.6–3.7) in a study from United States (Muscat and Wynder, 1991), and 3.1 (95% CI: 1.2–8.1) in a study from Germany (Rodelsperger et al., 2001). The apparent discrepancy of results between cohort and case-control studies might be explained by residual confounding by asbestos exposure in the latter type of investigation. An alternative explanation might be the high exposure level of individuals included in the case-control studies (predominantly appli-

Two cohorts of patients from Germany and Denmark undergoing radiological examinations with Thorotrast, a colloidal contrast agent that was used during the 1930s–1950s, mainly for cerebral arteriography, and emits 220Rn and other a-particles, have been studied for cancer incidence and mortality (Andersson et al., 1995; van Kaick et al., 1999). In both studies, an increased risk of mesothelioma, notably of the peritoneum, was detected, with relative risk in the order of 10 and a positive relationship between estimated radiation dose and mesothelioma risk. The mean latency between injection and tumor development was 30 years. In a further small, autopsy-based study from Japan, one case of peritoneal mesothelioma and three cases of peritoneal sarcoma were detected, which represented an excess over controls (1.1% vs. 0.2%, P < 0.01) (Ishikawa et al., 1995). It is likely that Thorotrast induces mesothelioma via a-particles from the agent deposited in the thoracic and abdominal organs adjacent to the pleura and peritoneum, such as the liver, the spleen, and the lymph nodes. Cases of mesothelioma have also been reported in cancer patients treated with radiotherapy (Antman et al., 1983; Cavazza et al., 1996). In the only available epidemiological study, the RR of mesothelioma following breast cancer was 1.78 (95% CI: 0.20–6.42) among patients treated with radiotherapy and 0.91 (95% CI: 0.24–2.32) in patients without radiotherapy (Neugut et al., 1997). No clear excess of mesothelioma has been detected in cohorts of individuals exposed to X- and gamma-radiation (IARC, 2000).

Simian Virus 40 (SV40) SV40 is a polyomavirus whose natural hosts are the rhesus monkey (Macaca mulatta) and other Asian macaques. In the early 1960s, it was found to contaminate vaccines prepared on cultures of rhesus kidney cells, including poliovaccine, of both the inactivated (“Salk”) and the attenuated oral (“Sabin”) types, as well as vaccines against adenovirus3 and adenovirus-7 (Fraumeni et al., 1963). Many millions of individuals have received contaminated vaccine around the world (an estimated 98 million in the United States); however, the individual

Table 34–8. Mesothelioma Mortality in Cohorts of Man-Made Synthetic Fiber Production Workers Study

Country

Total Number of Deaths

N Mesothelioma Deaths

US Europe France

9060 1281 N/A

0 1 0

191 437 161

0 0 0

Comments

glass wool Marsh et al. (2001) Boffetta et al. (1997) Moulin et al. (1986)

continuous filament Boffetta et al. (1997) Chiazze, Jr. et al. (1997) Watkins et al. (1997)

Europe US US

rock/slag wool Marsh et al. (2001)

US

1011

1

Boffetta et al. (1997)

Europe

1679

4

87

0

refractory ceramic fibers LeMasters et al. (2003)

US

Case not confirmed during pathology review Two cases with heavy asbestos exposure

668

PART IV: CANCER BY TISSUE OF ORIGIN

assessment of exposure is difficult because of inter-batch variability in contamination (Gazdar et al., 2002). Although contamination has occurred in most developed countries, it did not take place in some countries, such as Finland. The finding that injection of SV40 in hamsters produces mesothelioma in a high proportion of animals (Cicala et al., 1993) has raised the suspicion of a similar effect in humans. The results of selected studies of detection of SV40 DNA in human mesothelioma samples are reported in Table 34–9; in most series SV40 DNA sequences were detected in a high proportion (mainly in the range 40%–60%) of tumors. In three studies, however, all cases were free from viral DNA; this might be explained by lack of vaccine contamination in Finland and Turkey (Hirvonen et al., 1999; Emri et al., 2000), as well as low sensitivity of the assay based on archival material in an American study (Strickler et al., 1996). In addition to real differences in prevalence of contamination, inter-laboratory variability in the methods used to detect SV40 DNA might explain some of the discordant results (The International SV40 Working Group, 2001). The study by Strickler and colleagues (1996) is the only available investigation that has compared the prevalence of serological markers of SV40 infection based on 34 mesothelioma cases and 35 controls; antibody-positive sera were detected in three cases and one control. In one case series, the prevalence of SV40-positive tumors did not differ according to exposure to asbestos (Shivapurkar et al., 1999). The available cohort studies of individuals likely to have been exposed to contaminated vaccines have reported no excess risk of mesothelioma, but none had an adequate statistical power because of small size, short follow-up, or young age of cohort members (Mortimer et al., 1981; Geissler, 1990; Heinonen et al., 1973; CarrollPankhurst et al., 2001). Ecological studies from the United States and Sweden did not clearly show a higher incidence of mesothelioma among individuals likely to have been exposed to contaminated vaccines (Olin and Giesecke, 1998; Strickler et al., 1998; Fisher et al., 1999). In conclusion, an etiologic role of SV40 in human mesothelioma is suspected, given the results of experimental studies and the high proportion of positive tumors; however, additional evidence is needed from analytical epidemiology before a final conclusion can be reached (Klein et al., 2002).

Other Suspected Risk Factors Mesothelioma was not associated with tobacco smoking in either cohort or case-control studies (Hammond et al., 1979; Selikoff, 1977; Tagnon et al., 1980; Berry et al., 1985; Muscat and Wynder, 1991). In

a study of asbestos workers from the United Kingdom, although tobacco smoking did not increase the risk of mesothelioma, there was a suggestion of an interaction with asbestos exposure (Berry et al., 1985). The crude relative risks for heavy vs. moderate or low asbestos exposure, calculated from the raw data reported in the publication, were 2.0 (95% CI: 0.2–17), 1.4 (95% CI: 0.4–4.7), and 3.7 (95% CI: 1.1–13) among non-smokers, ex-smokers, and current smokers, respectively (P value of heterogeneity = 0.5). The increased risk among current smokers, albeit not significantly different from that among non-smokers and ex-smokers, suggests that tobacco smoking might facilitate the persistence of asbestos fibers in the peripheral lung. Limited information is available on the role of nutritional factors in determining mesothelioma risk. In a small case-control study, a protective effect of high intake of vegetables, and in particular cruciferous vegetables, was detected (Schiffman et al., 1988). In a further hospital-based case-control study from the United States, a decreased risk was detected for high consumption of carrots and tomatoes, but not of other foods rich in carotenoids, nor of cruciferous vegetables (Muscat and Huncharek, 1996). In this latter study, no association was detected between mesothelioma risk and body mass index. Although limited available evidence suggests a protective effect of a vegetablerich diet, the lack of data from prospective studies and the possible residual confounding effect of asbestos exposure caution against premature conclusions on the role of diet in human mesothelioma.

Genetic Factors While numerous familial clusters of mesothelioma have been reported (e.g., Dawson et al., 1992; Musti et al., 2002; Nalepa and Zielinski, 2001), they are generally based on cases who shared exposure to asbestos; it is therefore difficult to establish a genetic linkage based on these reports. A genetic predisposition to mesothelioma has been described in a six-generation study of villagers from erionite-exposed areas in Turkey (Roushdy-Hammady et al., 2001). However, some aspects of this study have been criticized (Saracci and Simonato, 2001). A significant association between pleural mesothelioma risk and familial history of cancer has been reported in two case-control studies (Huncharek et al., 1996; Heineman et al., 1996), providing suggestive evidence of a genetic predisposition. Relatively few studies have been conducted to evaluate the role of genetic polymorphisms in the etiology of mesothelioma, and they have generally lacked the statistical power to allow evaluation of potentially complex interactions. A high risk of mesothelioma has been reported in one study of 44 cases among workers who were highly exposed to asbestos and had both the null genotype for the glutathione S-

Table 34–9. Results of Selected Studies on Prevalence of SV40 DNA Sequences in Human Mesothelioma (Only Studies Including 20 or More Tumor Samples are Listed) Reference

Country

PCR Primer

N Samples

% Positive

PYV.for/PYV.rev SV.for3/SV.rev PYV.for/PYV.rev SV.for3/SV.rev SV.for3/SV.rev PYV.for/PYV.rev SV.for3/SV.rev SV.for3/SV.rev PYV.for/PYV.rev SV.for3/SV.rev SV.for2/SV.rev SV.for3/SV.rev SV.for2/SV.rev PYV.for/PYV.rev SV.for3/SV.rev PYV.for/PYV.rev PYV.for/PYV.rev SV.for2/SV.rev SV.for3/SV.rev

48 48 50* 50* 35* 21 42 25 26 26 26 118 49 25 28 23 83 29 66

75 60 0 0 86 48 90 32 42 100 12 48 0 56 46 39 60 0 48

Carbone et al. (1994)

US

Strickler et al. (1996)

US

De Luca et al. (1997) Galateau-Salle et al. (1998) Pass et al. (1998) Mutti et al. (1998) Griffiths et al. (1998)

N/A France US Italy UK

Shivapurkar et al. (1999) Hirvonen et al. (1999) Ramael et al. (1999) Dhaene et al. (1999) Strizzi et al. (2000) Procopio et al. (2000) Emri et al. (2000) Toyooka et al. (2001)

US, Canada, Germany Finland† Belgium Belgium Italy Italy Turkey† US

*Archival samples; †Country in which contaminated vaccine was not used.

Pleural and Peritoneal Neoplasms transferase gene, and the N-acetyltransferase-2 slow acetylator genotype (Hirvonen et al., 1995, 1996). A high frequency of micronuclei in peripheral blood lymphocytes has been observed among patients with pleural mesothelioma, and it has been suggested that these alterations may be a marker of susceptibility (Bolognesi et al., 2002).

Prevention Among the known causes of mesothelioma, asbestos plays a special role because of the high carcinogenic potency and the high prevalence of exposure. Control of exposure to asbestos and asbestiform fibers (e.g., erionite) represents the main strategy for preventing pleural and peritoneal mesothelioma. Control of occupational exposure has been successful in many circumstances, most notably the traditional high-risk industries such as manufacturing or application of asbestos products. The sharp decline in asbestos use in many high-income countries has further decreased the opportunities for exposure. However, several considerations mitigate an overly optimistic view on mesothelioma prevention. Firstly, circumstances of occupational exposure persist in highincome countries, mainly in relation with demolition and recycling of asbestos containing buildings and other materials. Secondly, use is still prevalent in many middle- and low-income countries, often under poor hygienic conditions (Kogevinas et al., 1994). Thirdly, any preventive effect resulting from control of asbestos exposure will become apparent only after several decades. Although environmental exposure to asbestos is not likely to be responsible for a large number of cases of mesothelioma, levels of contamination should be monitored and the main sources of exposure should be controlled. At the population level, Sweden is the only country for which a decline in the incidence of pleural mesothelioma has been observed, following sharp reduction in asbestos use since the 1970s (Hemminki and Li, 2003a). No other preventive measures seem justified by current knowledge on etiology and pathogenesis of mesothelioma. Although clinical surveillance of individuals with past exposure to asbestos is warranted, in view of the carcinogenic and other health consequences of asbestos exposure, there is no evidence that this leads to improved survival from mesothelioma. No genetic test can be recommended for implementation at the population level. References Acheson ED, Gardner MJ, Pippard EC, Grime LP. 1982. Mortality of two groups of women who manufactured gas masks from chrysotile and crocidolite asbestos: A 40-year follow-up. Br J Ind Med 39(4):344– 348. Acheson ED, Gardner MJ, Winter PD, Bennett C. 1984. Cancer in a factory using amosite asbestos. Int J Epidemiol 13(1):3–10. Agudo A, Gonz CA, Bleda MJ, et al. 2000. Occupation and risk of malignant pleural mesothelioma: A case-control study in Spain. Am J Ind Med 37(2):159–168. Albin M, Jakobsson K, Attewell R, Johansson L, Welinder H. 1990. Mortality and cancer morbidity in cohorts of asbestos cement workers and referents. Br J Ind Med 47(9):602–610. Algranti E. 1998. Asbestos: Current issues related to cancer and to uses in developing countries. Cad Saude Publica 14 Suppl 3:173–176. Alies-Patin AM, Valleron AJ. 1985. Mortality of workers in a French asbestos cement factory 1940–82. Br J Ind Med 42(4):219–225. Amandus HE, Wheeler R. 1987. The morbidity and mortality of vermiculite miners and millers exposed to tremolite-actinolite: Part II. Mortality. Am J Ind Med 11(1):15–26. Anderson HA, Lilis R, Daum SM, Fischbein AS, Selikoff IJ. 1976. Householdcontact asbestos neoplastic risk. Ann N Y Acad Sci 271:311–323. Andersson M, Wallin H, Jonsson M, et al. 1995. Lung carcinoma and malignant mesothelioma in patients exposed to Thorotrast: Incidence, histology and p53 status. Int J Cancer 63(3):330–336. Anspach M. 1965. [On the etiology of pleural calcifications.] Radiol Diagn (Berl) 6(3):341–347. Antman KH, Corson JM, Li FP, et al. 1983. Malignant mesothelioma following radiation exposure. J Clin Oncol 1(11):695–700. Armstrong BK, de Klerk NH, Musk AW, Hobbs MS. 1988. Mortality in miners and millers of crocidolite in Western Australia. Br J Ind Med 45(1):5–13. Banaei A, Auvert B, Goldberg M, Gueguen A, Luce D, Goldberg S. 2000. Future trends in mortality of French men from mesothelioma. Occup Environ Med 57(7):488–494.

669

Baris I, Simonato L, Artvinli M, et al. 1987. Epidemiological and environmental evidence of the health effects of exposure to erionite fibres: A four-year study in the Cappadocian region of Turkey. Int J Cancer 39(1): 10–17. Baris YI, Artvinli M, Sahin AA. 1979. Environmental mesothelioma in Turkey. Ann N Y Acad Sci 330:423–432. Baris YI, Sahin AA, Ozesmi M, et al. 1978. An outbreak of pleural mesothelioma and chronic fibrosing pleurisy in the village of Karain/Urgup in Anatolia. Thorax 33(2):181–192. Baris YI, Saracci R, Simonato L, Skidmore JW, Artvinli M. 1981. Malignant mesothelioma and radiological chest abnormalities in two villages in Central Turkey. An epidemiological and environmental investigation. Lancet 1(8227):984–987. Battista G, Belli S, Comba P, et al. 1999. Mortality due to asbestos-related causes among railway carriage construction and repair workers. Occup Med (Lond) 49(8):536–539. Berrino Feal. 1999. Survival of Cancer Patients in Europe: The EUROCARE2 Study. Lyon: IARC Scientific Publications. Berry G, Newhouse ML, Antonis P. 1985. Combined effect of asbestos and smoking on mortality from lung cancer and mesothelioma in factory workers. Br J Ind Med 42(1):12–18. Berry M. 1997. Mesothelioma incidence and community asbestos exposure. Environ Res 75(1):34–40. Bianchi C, Brollo A, Ramani L, Zuch C. 1997. Pleural plaques as risk indicators for malignant pleural mesothelioma: A necropsy-based study. Am J Ind Med 32(5):445–449. Boffetta P, Saracci R, Andersen A, et al. 1997. Cancer mortality among man-made vitreous fiber production workers. Epidemiology 8(3):259– 268. Bolognesi C, Filiberti R, Neri M, et al. 2002. High frequency of micronuclei in peripheral blood lymphocytes as index of susceptibility to pleural malignant mesothelioma. Cancer Res 62(19):5418–5419. Botha JL, Irwig LM, Strebel PM. 1986. Excess mortality from stomach cancer, lung cancer, and asbestosis and/or mesothelioma in crocidolite mining districts in South Africa. Am J Epidemiol 123(1):30–40. Bourdes V, Boffetta P, Pisani P. 2000. Environmental exposure to asbestos and risk of pleural mesothelioma: Review and meta-analysis. Eur J Epidemiol 16(5):411–417. Brooks SM, Stockwell HG, Pinkham PA, Armstrong AW, Witter DA. 1992. Sugarcane exposure and the risk of lung cancer and mesothelioma. Environ Res 58(2):195–203. Camus M, Siemiatycki J, Meek B. 1998. Nonoccupational exposure to chrysotile asbestos and the risk of lung cancer. N Engl J Med 338(22):1565–1571. Carbone M, Pass HI, Rizzo P, et al. 1994. Simian virus 40-like DNA sequences in human pleural mesothelioma. Oncogene 9(6):1781–1790. Carroll-Pankhurst C, Engels EA, Strickler HD, Goedert JJ, Wagner J, Mortimer EA, Jr. 2001. Thirty-five year mortality following receipt of. Br J Cancer 85(9):1295–1297. Cavazza A, Travis LB, Travis WD, et al. 1996. Post-irradiation malignant mesothelioma. Cancer 77(7):1379–1385. Chellini E, Fornaciai G, Merler E, et al. 1992. Pleural malignant mesothelioma in Tuscany, Italy (1970–1988): II. Identification of occupational exposure to asbestos. Am J Ind Med 21(4):577–585. Chiazze L, Jr., Watkins DK, Fryar C. 1997. Historical cohort mortality study of a continuous filament fiberglass manufacturing plant. I. White men. J Occup Environ Med 39(5):432–441. Churg Roggli VL, Galateau-Salle F, Cagle P, et al. 2004. Tumours of the Pleura. Pathology and Genetics of Tumours of the Respiratory Tract, Mediastinum and Thymus. Lyon: International Agency for Research on Cancer. Churg A, Wright JL. 1994. Persistence of natural mineral fibers in human lungs: an overview. Environ Health Perspect 102 Suppl 5:229–233. Cicala C, Pompetti F, Carbone M. 1993. SV40 induces mesotheliomas in hamsters. Am J Pathol 142(5):1524–1533. Cicioni C, London SJ, Garabrant DH, Bernstein L, Phillips K, Peters JM. 1991. Occupational asbestos exposure and mesothelioma risk in Los Angeles County: Application of an occupational hazard survey job-exposure matrix. Am J Ind Med 20(3):371–379. Cocco P, Dosemeci M. 1999. Peritoneal cancer and occupational exposure to asbestos: results from the application of a job-exposure matrix. Am J Ind Med 35(1):9–14. Cooper SP, Fraire AE, Buffler PA, Greenberg SD, Langston C. 1989. Epidemiologic aspects of childhood mesothelioma. Pathol Immunopathol Res 8(5–6):276–286. Crosignani P, Forastiere F, Petrelli G, et al. 1995. Malignant mesothelioma in thermoelectric power plant workers in Italy. Am J Ind Med 27(4): 573–576.

670

PART IV: CANCER BY TISSUE OF ORIGIN

Dawson A, Gibbs A, Browne K, Pooley F, Griffiths M. 1992. Familial mesothelioma. Details of 17 cases with histopathologic findings and mineral analysis. Cancer 70(5):1183–1187. De Luca A, Baldi A, Esposito V, et al. 1997. The retinoblastoma gene family pRb/p105, p107, pRb2/p130 and simian virus-40 large T-antigen in human mesotheliomas. Nat Med 3(8):913–916. Dement JM, Brown DP, Okun A. 1994. Follow-up study of chrysotile asbestos textile workers: cohort mortality and case-control analyses. Am J Ind Med 26(4):431–447. Dhaene K, Verhulst A, Van Marck E. 1999. SV40 large T-antigen and human pleural mesothelioma. Screening by polymerase chain reaction and tyramine-amplified immunohistochemistry. Virchows Arch 435(1):1–7. Dodoli D, Del Nevo M, Fiumalbi C, et al. 1992. Environmental household exposures to asbestos and occurrence of pleural mesothelioma. Am J Ind Med 21(5):681–687. Egilman D, Fehnel C, Bohme SR. 2003. Exposing the myth of , Anything but Chrysotile. A critique of the Canadian mining industry and McGill University chrysotile studies. Am J Ind Med 44(5):540–57. ELMES PC, McCaughey WT, Wade OL. 2002. Diffuse mesothelioma of the pleura and asbestos. Br Med J 1965(5431):350–353. Emri S, Demir A, Dogan M, Akay H, Bozkurt B, Carbone M, et al. Lung diseases due to environmental exposures to erionite and asbestos in Turkey. Toxicol Lett 127(1–3):251–257. Emri S, Kocagoz T, Olut A, Gungen Y, Mutti L, Baris YI. 2000. Simian virus 40 is not a cofactor in the pathogenesis of environmentally induced malignant pleural mesothelioma in Turkey. Anticancer Res 20(2A):891–894. Enterline PE, Hartley J, Henderson V. 1987. Asbestos and cancer: A cohort followed up to death. Br J Ind Med 44(6):396–401. Finkelstein MM. 1983. Mortality among long term employees of an Ontario asbestos-cement factory. Br J Ind Med 40:138–144. Finkelstein MM. 1984. Mortality among employees of an Ontario asbestoscement factory. Am Rev Respir Dis 129(5):754–761. Fisher SG, Weber L, Carbone M. 1999. Cancer risk associated with simian virus 40 contaminated polio vaccine. Anticancer Res 19(3B):2173–2180. Fletcher DE. 1972. A mortality study of shipyard workers with pleural plaques. Br J Ind Med 29(2):142–145. Fraumeni JF, Jr., Ederer F, Miller RW. 1963. An evaluation of the carcinogenicity of Simian Virus 40 in Man. JAMA 185(9):85–90. Galateau-Salle F, Bidet P, Iwatsubo Y, et al. 1998. SV40-like DNA sequences in pleural mesothelioma, bronchopulmonary carcinoma, and nonmalignant pulmonary diseases. J Pathol 184(3):252–257. Gardner MJ, Winter PD, Pannett B, Powell CA. 1986. Follow up study of workers manufacturing chrysotile asbestos cement products. Br J Ind Med 43(11):726–732. Gazdar AF, Butel JS, Carbone M. 2002. SV40 and human tumours: Myth, association or causality? Nat Rev Cancer 2(12):957–964. Geissler E. 1990. SV40 and human brain tumors. Prog Med Virol 37:211–222. Gennaro V, Finkelstein MM, Ceppi M, et al. 2000. Mesothelioma and lung tumors attributable to asbestos among petroleum workers. Am J Ind Med 37(3):275–282. Germani D, Belli S, Bruno C, et al. 1999. Cohort mortality study of women compensated for asbestosis in Italy. Am J Ind Med 36(1):129–134. Giaroli C, Belli S, Bruno C, et al. 1994. Mortality study of asbestos cement workers. Int Arch Occup Environ Health 66(1):7–11. Gordon GJ, Jensen RV, Hsiao LL, et al. 2003. Using gene expression ratios to predict outcome among patients with mesothelioma. J Natl Cancer Inst 95(8):598–605. Griffiths DJ, Nicholson AG, Weiss RA. 1998. Detection of SV40 sequences in human mesothelioma. Dev Biol Stand 94:127–136. Hagerstrand I, Meurman L, Odlund B. 1968. Asbestos bodies in the lungs and mesothelioma. A retrospective examination of a ten-year-autopsy material. Acta Pathol Microbiol Scand 72(2):177–191. Hammond EC, Selikoff IJ, Seidman H. 1979. Asbestos exposure, cigarette smoking and death rates. Ann N Y Acad Sci 330:473–490. Hansen J, de Klerk NH, Musk AW, Hobbs MS. 1998. Environmental exposure to crocidolite and mesothelioma: Exposure-response relationships. Am J Respir Crit Care Med 157(1):69–75. Harris LV, Kahwa IA. 2003. Asbestos: Old foe in 21st century developing countries. Sci Total Environ 307(1–3):1–9. Health Effects Institute—Asbestos Research. 1991. Asbestos in Public and Commercial Buildings: A Literature Review and Synthesis of Current Knowledge. Cambridge, USA. Heineman EF, Bernstein L, Stark AD, Spirtas R. 1996. Mesothelioma, asbestos, and reported history of cancer in first-degree relatives. Cancer 77(3): 549–554. Heinonen OP, Shapiro S, Monson RR, Hartz SC, Rosenberg L, Slone D. 1973. Immunization during pregnancy against poliomyelitis and influenza in relation to childhood malignancy. Int J Epidemiol 2(3):229–235.

Hemminki K, Li X. 2003a. Mesothelioma incidence seems to have leveled off in Sweden. Int J Cancer 103(1):145–146. Hemminki K, Li X. 2003b. Time trends and occupational risk factors for pleural mesothelioma in Sweden. J Occup Environ Med 45(4):456–461. Hilt B, Andersen A, Rosenberg J, Langard S. 1991. Cancer incidence among asbestos-exposed chemical industry workers: an extended observation period. Am J Ind Med 20(2):261–264. Hirvonen A, Mattson K, Karjalainen A, et al. 1999. Simian virus 40 (SV40)like DNA sequences not detectable in finnish mesothelioma patients not exposed to SV40-contaminated polio vaccines. Mol Carcinog 26(2): 93–99. Hirvonen A, Pelin K, Tammilehto L, Karjalainen A, Mattson K, Linnainmaa K. 1995. Inherited GSTM1 and NAT2 defects as concurrent risk modifiers in asbestos-related human malignant mesothelioma. Cancer Res 55(14):2981–2983. Hirvonen A, Saarikoski ST, Linnainmaa K, et al. 1996. Glutathione Stransferase and N-acetyltransferase genotypes and asbestos-associated pulmonary disorders. J Natl Cancer Inst 88(24):1853–1856. Hodgson JT, Jones RD. 1986. Mortality of asbestos workers in England and Wales 1971–81. Br J Ind Med 43(3):158–164. Howel D, Arblaster L, Swinburne L, Schweiger M, Renvoize E, Hatton P. 1997. Routes of asbestos exposure and the development of mesothelioma in an English region. Occup Environ Med 54(6):403–409. Hughes JM, Weill H, Hammad YY. 1987. Mortality of workers employed in two asbestos cement manufacturing plants. Br J Ind Med 44(3):161– 174. Huncharek M, Kelsey K, Muscat J, Christiani D. 1996. Parental cancer and genetic predisposition in malignant pleural mesothelioma: A case-control study. Cancer Lett 102(1–2):205–208. International Agency for Research on Cancer. 1973. Asbestos. IARC Monographs on the Evaluation of Carcinogenic Risk of Chemicals to Man. Vol. 2. Some inorganic and organometallic compounds. International Agency for Research on Cancer, Lyon, 17–47. International Agency for Research on Cancer. 1987. Erionite. IARC Monographs on the Evaluation of the Carcinogenic Risks to Humans. Overall Evaluations of Carcinogenicity: An Updating of IARC Monographs Volumes 1 to 42. Supplement 7. International Agency for Research on Cancer, Lyon, 203. International Agency for Research on Cancer. 1991. Autopsy in Epidemiology and Medical Research E. Riboli E and Delendi M, eds. IARC Scientific Publications, No 112, Lyon. International Agency for Research on Cancer. 2000. X-Radiation and GammaRadiation. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. Vol. 75. Ionizing Radiation, Part 1: X- and GammaRadiation, and Neutrons. International Agency for Research on Cancer, Lyon, 121–359. Ishikawa Y, Mori T, Machinami R. 1995. Lack of apparent excess of malignant mesothelioma but increased overall malignancies of peritoneal cavity in Japanese autopsies with Thorotrast injection into blood vessels. J Cancer Res Clin Oncol 121(9–10):567–570. Iwatsubo Y, Pairon JC, Boutin C, et al. 1998. Pleural mesothelioma: Dose-response relation at low levels of asbestos exposure in a French population-based case-control study. Am J Epidemiol 148(2):133– 142. Jarvholm B, Malker H, Malker B, Ericsson J, Sallsten G. 1988. Pleural mesotheliomas and asbestos exposure in the pulp and paper industries: A new risk group identified by linkage of official registers. Am J Ind Med 13(5):561–567. Jones JS, Smith PG, Pooley FD, et al. 1980. The consequences of exposure to asbestos dust in a wartime gas-mask factory. IARC Sci Publ (30): 637–653. Kane AB. 1996. Mechanisms of mineral fibre carcinogenesis. IARC Sci Publ (140):11–34. Kielkowski D, Nelson G, Rees D. 2000. Risk of mesothelioma from exposure to crocidolite asbestos: A 1995 update of a South African mortality study. Occup Environ Med 57(8):563–567. Kjaergaard J, Andersson M. 2000. Incidence rates of malignant mesothelioma in Denmark and predicted future number of cases among men. Scand J Work Environ Health 26(2):112–117. Klein G, Powers A, Croce C. 2002. Association of SV40 with human tumors. Oncogene 21(8):1141–1149. Kogevinas M, Boffetta P, Pearce N. 1994. Occupational exposure to carcinogens in developing countries. In: Pearce N, Matos E, Vainio H, Boffetta P, Kogevinas M, eds. Occupational Cancer in Developing Countries. Lyon: IARC Scientific Publications, pp 63–95. Kolonel LN, Yoshizawa CN, Hirohata T, Myers BC. 1985. Cancer occurrence in shipyard workers exposed to asbestos in Hawaii. Cancer Res 45(8): 3924–3928.

Pleural and Peritoneal Neoplasms Koskinen K, Pukkala E, Reijula K, Karjalainen A. 2003. Incidence of cancer among the participants of the Finnish Asbestos Screening Campaign. Scand J Work Environ Health 29(1):64–70. La Vecchia C, Decarli A, Peto J, Levi F, Tomei F, Negri E. 2000. An age, period and cohort analysis of pleural cancer mortality in Europe. Eur J Cancer Prev 9(3):179–184. Langer AM, Nolan RP, Constantopoulos SH, Moutsopoulos HM. 1987. Association of Metsovo lung and pleural mesothelioma with exposure to tremolite-containing whitewash. Lancet 1(8539):965–967. Langseth H, Andersen A. 2000. Cancer incidence among male pulp and paper workers in Norway. Scand J Work Environ Health 26(2):99–105. Law MR, Gregor A, Hodson ME, Bloom HJ, Turner-Warwick M. 1984. Malignant mesothelioma of the pleura: A study of 52 treated and 64 untreated patients. Thorax 39(4):255–259. LeMasters GK, Lockey JE, Yiin JH, Hilbert TJ, Levin LS, Rice CH. 2003. Mortality of workers occupationally exposed to refractory ceramic fibers. J Occup Environ Med 45(4):440–450. Liddell FD, McDonald AD, McDonald JC. 1997. The 1891–1920 birth cohort of Quebec chrysotile miners and millers: Development from 1904 and mortality to 1992. Ann Occup Hyg 41(1):13–36. Luce D, Bugel I, Goldberg P, et al. 2000. Environmental exposure to tremolite and respiratory cancer in New Caledonia: A case-control study. Am J Epidemiol 151(3):259–265. Luo S, Liu X, Mu S, Tsai SP, Wen CP. 2003. Asbestos related diseases from environmental exposure to crocidolite in Da-yao, China. I. Review of exposure and epidemiological data. Occup Environ Med 60(1):35–41. Magnani C, Agudo A, Gonzalez CA, et al. 2000. Multicentric study on malignant pleural mesothelioma and non-occupational exposure to asbestos. Br J Cancer 83(1):104–111. Magnani C, Dalmasso P, Biggeri A, Ivaldi C, Mirabelli D, Terracini B. 2001. Increased risk of malignant mesothelioma of the pleura after residential or domestic exposure to asbestos: A case-control study in Casale Monferrato, Italy. Environ Health Perspect 109(9):915–919. Magnani C, Terracini B, Ivaldi C, Botta M, Mancini A, Andrion A. 1995. Pleural malignant mesothelioma and non-occupational exposure to asbestos in Casale Monferrato, Italy. Occup Environ Med 52(6):362– 367. Magnani C, Terracini B, Ivaldi C, Mancini A, Botta M. 1996. [Tumor mortality and from other causes in asbestos cement workers at the Casale Montferrato plant.] Med Lav 87(2):133–146. Malker HR, Malker BK, Blot WJ. 1983. Mesothelioma among sugar refinery workers. Lancet 2(8354):858. Malker HS, McLaughlin JK, Malker BK, et al. 1985. Occupational risks for pleural mesothelioma in Sweden, 1961–79. J Natl Cancer Inst 74(1):61–66. Maltoni C, Pinto C, Valenti D, Carnuccio R, Amaducci E, Minardi F. 1995. Mesotheliomas following exposure to asbestos used in sugar refineries: Report of 12 Italian cases. Med Lav 86(5):478–483. Mancuso TF, Coulter EJ. 1963. Methodology in Industrial Health Studies. Arch Environ Health 6:210–226. Maples KR, Johnson NF. 1992. Fiber-induced hydroxyl radical formation: Correlation with mesothelioma induction in rats and humans. Carcinogenesis 13(11):2035–2039. Marsh GM, Gula MJ, Youk AO, Buchanich JM, Churg A, Colby TV. 2001. Historical cohort study of US man-made vitreous fiber production workers: II. Mortality from mesothelioma. J Occup Environ Med 43(9):757–766. Mast RW, Hesterberg TW, Glass LR, McConnell EE, Anderson R, Bernstein DM. 1994. Chronic inhalation and biopersistence of refractory ceramic fiber in rats and hamsters. Environ Health Perspect 102 Suppl 5:207– 209. McConnochie K, Simonato L, Mavrides P, Christofides P, Pooley FD, Wagner JC. 1987. Mesothelioma in Cyprus: The role of tremolite. Thorax 42(5): 342–347. McDonald AD, Case BW, Churg A, et al. 1997. Mesothelioma in Quebec chrysotile miners and millers: Epidemiology and aetiology. Ann Occup Hyg 41(6):707–719. McDonald AD, Fry JS, Woolley AJ, McDonald JC. 1984. Dust exposure and mortality in an American chrysotile asbestos friction products plant. Br J Ind Med 41(2):151–157. McDonald AD, McDonald JC, Pooley FD. 1982. Mineral fibre content of lung in mesothelial tumours in North America. Ann Occup Hyg 26(1–4):417–422. McDonald JC, Armstrong B, Case B, et al. 1989. Mesothelioma and asbestos fiber type. Evidence from lung tissue analyses. Cancer 63(8):1544–1547. McDonald JC, Edwards CW, Gibbs AR, et al. 2001. Case-referent survey of young adults with mesothelioma: II. Occupational analyses. Ann Occup Hyg 45(7):519–523.

671

Menegozzo M, Belli S, Bruno C, et al. 1993. [Mortality due to causes correlatable to asbestos in a cohort of workers in railway car construction.] Med Lav 84(3):193–200. Metintas M, Hillerdal G, Metintas S. 1999. Malignant mesothelioma due to environmental exposure to erionite: Follow-up of a Turkish emigrant cohort. Eur Respir J 13(3):523–526. Meurman LO, Pukkala E, Hakama M. 1994. Incidence of cancer among anthophyllite asbestos miners in Finland. Occup Environ Med 51(6):421–425. Morinaga K, Kishimoto T, Sakatani M, Akira M, Yokoyama K, Sera Y. 2001. Asbestos-related lung cancer and mesothelioma in Japan. Ind Health 39(2):65–74. Mortimer EA, Jr., Lepow ML, Gold E, Robbins FC, Burton GJ, Fraumeni JF, Jr. 1981. Long-term follow-up of persons inadvertently inoculated with SV40 as neonates. N Engl J Med 305(25):1517–1518. Moulin JJ, Mur JM, Wild P, Perreaux JP, Pham QT. 1986. Oral cavity and laryngeal cancers among man-made mineral fiber production workers. Scand J Work Environ Health 12(1):27–31. Murai Y. 2001. Malignant mesothelioma in Japan: analysis of registered autopsy cases. Arch Environ Health 56(1):84–88. Muscat JE, Huncharek M. 1996. Dietary intake and the risk of malignant mesothelioma. Br J Cancer 73(9):1122–1125. Muscat JE, Wynder EL. 1991. Cigarette smoking, asbestos exposure, and malignant mesothelioma. Cancer Res 51(9):2263–2267. Musti M, Cavone D, Aalto Y, Scattone A, Serio G, Knuutila S. 2002. A cluster of familial malignant mesothelioma with del(9p) as the sole chromosomal anomaly. Cancer Genet Cytogenet 138(1):73–76. Mutti L, De Luca A, Claudio PP, Convertino G, Carbone M, Giordano A. 1998. Simian virus 40-like DNA sequences and large-T antigen-retinoblastoma family protein pRb2/p130 interaction in human mesothelioma. Dev Biol Stand 94:47–53. Nalepa P, Zielinski M. 2001. [Two cases of malignant mesothelioma of the pleura within one family.] Pol Merkuriusz Lek 11(62):165–168. Neuberger M, Kundi M. 1990. Individual asbestos exposure: smoking and mortality–a cohort study in the asbestos cement industry. Br J Ind Med 47(9):615–620. Neuberger M, Vutuc C. 2003. Three decades of pleural cancer and mesothelioma registration in Austria where asbestos cement was invented. Int Arch Occup Environ Health 76(2):161–166. Neugut AI, Ahsan H, Antman KH. 1997. Incidence of malignant pleural mesothelioma after thoracic radiotherapy. Cancer 80(5):948–950. Newhouse ML, Berry G, Wagner JC. 1985. Mortality of factory workers in east London 1933–80. Br J Ind Med 42(1):4–11. Newhouse ML, Sullivan KR. 1989. A mortality study of workers manufacturing friction materials:1941–86. Br J Ind Med 46(3):176–179. Newhouse ML, Thompson H. 1965. Mesothelioma of pleura and peritoneum following exposure to asbestos in the London area. Br J Ind Med 22(4): 261–269. Nicholson WJ. 1986. Airborne Asbestos Health Assessment Update. Washington, DC: Office of Health and Environmental Assessment, U.S. Environmental Protection Agency. Ohlson CG, Hogstedt C. 1985. Lung cancer among asbestos cement workers. A Swedish cohort study and a review. Br J Ind Med 42(6):397–402. Ohlson CG, Klaesson B, Hogstedt C. 1984. Mortality among asbestos-exposed workers in a railroad workshop. Scand J Work Environ Health 10(5): 283–291. Oksa P, Pukkala E, Karjalainen A, Ojajarvi A, Huuskonen MS. 1997. Cancer incidence and mortality among Finnish asbestos sprayers and in asbestosis and silicosis patients. Am J Ind Med 31(6):693–698. Olin P, Giesecke J. 1998. Potential exposure to SV40 in polio vaccines used in Sweden during 1957: no impact on cancer incidence rates 1960 to 1993. Dev Biol Stand 94:227–233. Paci E, Dini S, Buiatti E, Seniori CA, Lenzi S, Zappa M. 1987. Malignant mesothelioma in non-asbestos textile workers in Florence. Am J Ind Med 11(3):249–254. Paci E, Zappa M, Paoletti L, et al. 1991. Further evidence of an excess of risk of pleural malignant mesothelioma in textile workers in Prato (Italy). Br J Cancer 64(2):377–378. Paoletti L, Batisti D, Bruno C, et al. 2000. Unusually high incidence of malignant pleural mesothelioma in a town of eastern Sicily: An epidemiological and environmental study. Arch Environ Health 55(6):392–398. Pass HI, Donington JS, Wu P, et al. 1998. Human mesotheliomas contain the simian virus-40 regulatory region and large tumor antigen DNA sequences. J Thorac Cardiovasc Surg 116(5):854–859. Peto J, Decarli A, La Vecchia C, Levi F, Negri E. 1999. The European mesothelioma epidemic. Br J Cancer 79(3–4):666–672. Peto J, Doll R, Hermon C, Binns W, Clayton R, Goffe T. 1985. Relationship of mortality to measures of environmental asbestos pollution in an asbestos textile factory. Ann Occup Hyg 29(3):305–355.

672

PART IV: CANCER BY TISSUE OF ORIGIN

Peto J, Hodgson JT, Matthews FE, Jones JR. 1995. Continuing increase in mesothelioma mortality in Britain. Lancet 345(8949):535–539. Peto J, Seidman H, Selikoff IJ. 1982. Mesothelioma mortality in asbestos workers: implications for models of carcinogenesis and risk assessment. Br J Cancer 45(1):124–135. Piolatto G, Negri E, La Vecchia C, Pira E, Decarli A, Peto J. 1990. An update of cancer mortality among chrysotile asbestos miners in Balangero, northern Italy. Br J Ind Med 47(12):810–814. Price B. 1997. Analysis of current trends in United States mesothelioma incidence. Am J Epidemiol 145(3):211–218. Procopio A, Strizzi L, Vianale G, et al. 2000. Simian virus-40 sequences are a negative prognostic cofactor in patients with malignant pleural mesothelioma. Genes Chromosomes Cancer 29(2):173–179. Puntoni R, Merlo F, Borsa L, Reggiardo G, Garrone E, Ceppi M. 2001. A historical cohort mortality study among shipyard workers in Genoa, Italy. Am J Ind Med 40(4):363–370. Raffn E, Lynge E, Juel K, Korsgaard B. 1989. Incidence of cancer and mortality among employees in the asbestos cement industry in Denmark. Br J Ind Med 46(2):90–96. Ramael M, Nagels J, Heylen H, et al. 1999. Detection of SV40 like viral DNA and viral antigens in malignant pleural mesothelioma. Eur Respir J 14(6):1381–1386. Rees D, Myers JE, Goodman K, et al. 1999. Case-control study of mesothelioma in South Africa. Am J Ind Med 35(3):213–222. Ries LAG, Eisner MP, Kosary CL, et al. 2003. SEER Cancer Statistics Review, 1975–2000. National Cancer Institute, Bethesda, MD. Rodelsperger K, Jockel KH, Pohlabeln H, Romer W, Woitowitz HJ. 2001. Asbestos and man-made vitreous fibers as risk factors for diffuse malignant mesothelioma: Results from a German hospital-based case-control study. Am J Ind Med 39(3):262–275. Rossiter CE, Coles RM. 1980. HM Dockyard, Devonport: 1947 mortality study. IARC Sci Publ (30):713–721. Roushdy-Hammady I, Siegel J, Emri S, Testa JR, Carbone M. 2001. Geneticsusceptibility factor and malignant mesothelioma in the Cappadocian region of Turkey. Lancet 357(9254):444–445. Sanden A, Jarvholm B, Larsson S, Thiringer G. 1992. The risk of lung cancer and mesothelioma after cessation of asbestos exposure: A prospective cohort study of shipyard workers. Eur Respir J 5(3):281–285. Saracci R, Simonato L. 2001. Familial malignant mesothelioma. Lancet 358(9295):1813–1814. Saracci R, Simonato L, Baris Y, Artvinli M, Skidmore J. 1982. The agemortality curve of endemic pleural mesothelioma in Karain, Central Turkey. Br J Cancer 45(1):147–149. Schiffman MH, Pickle LW, Fontham E, et al. 1988. Case-control study of diet and mesothelioma in Louisiana. Cancer Res 48(10):2911–2915. Sebastien P, Fondimare A, Bignon J, Monchaux G, Desbordes J, Bonnaud G. 1975. Topographical distributions of asbestos fibers in human lung in relation to occupational and non-occupational exposure. Inhaled Particles 2:435–436. Segura O, Burdorf A, Looman C. 2003. Update of predictions of mortality from pleural mesothelioma in the Netherlands. Occup Environ Med 60(1):50–55. Seidman H, Selikoff IJ, Gelb SK. 1986. Mortality experience of amosite asbestos factory workers: Dose-response relationships 5 to 40 years after onset of short-term work exposure. Am J Ind Med 10(5–6):479–514. Seidman H, Selikoff IJ, Hammond EC. 1979. Short-term asbestos work exposure and long-term observation. Ann N Y Acad Sci 330:61–89. Selikoff IJ. 1977. Air pollution and asbestos carcinogenesis: Investigation of possible synergism. IARC Sci Publ (16):247–253. Selikoff IJ, Churg J, Hammond EC. 1964. Asbestos exposure and neoplasia. JAMA 188:22–26. Selikoff IJ, Hammond EC, Seidman H. 1979. Mortality experience of insulation workers in the United States and Canada, 1943–1976. Ann N Y Acad Sci 330:91–116. Selikoff IJ, Seidman H. 1991. Asbestos-associated deaths among insulation workers in the United States and Canada, 1967–1987. Ann N Y Acad Sci 643:1–14. Shivapurkar N, Wiethege T, Wistuba II, et al. 1999. Presence of simian virus 40 sequences in malignant mesotheliomas and mesothelial cell proliferations. J Cell Biochem 76(2):181–188. Siemiatycki J, Boffetta P. 1998. Invited commentary: Is it possible to investigate the quantitative relation between asbestos and mesothelioma in a community-based study? Am J Epidemiol 148(2):143–147. Sinks T, Goodman M, Kolonel L, Anderson B. 1992. A case-control study of mesothelioma and exposure to biogenic silica fibers. Am J Epidemiol 136:1015. Sleggs CA, Marchand P, Wagner JC. 1961. Diffuse pleural mesotheliomas in South Africa. S Afr Med J 35:28–34.

Sluis-Cremer GK, Liddell FD, Logan WP, Bezuidenhout BN. 1992. The mortality of amphibole miners in South Africa, 1946–80. Br J Ind Med 49(8):566–575. Stayner LT, Dankovic DA, Lemen RA. 1996. Occupational exposure to chrysotile asbestos and cancer risk: A review of the amphibole hypothesis. Am J Public Health 86(2):179–186. Strickler HD, Goedert JJ, Fleming M, CS et al. 1996. Simian virus 40 and pleural mesothelioma in humans. Cancer Epidemiol Biomarkers Prev 5(6):473–475. Strickler HD, Rosenberg PS, Devesa SS, Hertel J, Fraumeni JF, Jr, Goedert JJ. 1998. Contamination of poliovirus vaccines with simian virus 40 (1955–1963) and subsequent cancer rates. JAMA 279(4):292–295. Strizzi L, Vianale G, Giuliano M, et al. 2000. SV40, JC and BK expression in tissue, urine and blood samples from patients with malignant and nonmalignant pleural disease. Anticancer Res 20(2A):885–889. Suzuki Y, Yuen SR. 2001. Asbestos tissue burden study on human malignant mesothelioma. Ind Health 39(2):150–160. Svenes KB, Borgersen A, Haaversen O, Holten K. 1986. Parietal pleural plaques: A comparison between autopsy and X-ray findings. Eur J Respir Dis 69(1):10–15. Szeszenia-Dabrowska N, Urszula W, Szymczak W, Strzelecka A. 2002. Mortality study of workers compensated for asbestosis in Poland, 1970–1997. Int J Occup Med Environ Health 15(3):267–278. Szeszenia-Dabrowska N, Wilczynska U, Szymczak W. 1986. [Risk of cancer in women occupationally exposed to asbestos dust.] Med Pr 37(4): 243–249. Szeszenia-Dabrowska N, Wilczynska U, Szymczak W. 1997. [Cancer risk in asbestos-cement industry workers in Poland.] Med Pr 48(5):473–483. Tagnon I, Blot WJ, Stroube RB, et al. 1980. Mesothelioma associated with the shipbuilding industry in coastal Virginia. Cancer Res 40(11):3875– 3879. Takahashi K, Huuskonen MS, Tossavainen A, et al. 1999. Ecological relationship between mesothelioma incidence/mortality and asbestos consumption in ten Western countries and Japan. J Occup Health 41:8–11. Talcott JA, Thurber WA, Kantor AF, et al. 1989. Asbestos-associated diseases in a cohort of cigarette-filter workers. N Engl J Med 321(18):1220–1223. Theriault GP, Grand-Bois L. 1978. Mesothelioma and asbestos in the Province of Quebec, 1969–1972. Arch Environ Health 33(1):15–19. Thomas HF, Benjamin IT, Elwood PC, Sweetnam PM. 1982. Further followup study of workers from an asbestos cement factory. Br J Ind Med 39(3):273–276. Toyooka S, Pass HI, Shivapurkar N, et al. 2001. Aberrant methylation and simian virus 40 tag sequences in malignant mesothelioma. Cancer Res 61(15):5727–5730. Tsai SP, Waddell LC, Gilstrap EL, Ransdell JD, Ross CE. 1996. Mortality among maintenance employees potentially exposed to asbestos in a refinery and petrochemical plant. Am J Ind Med 29(1):89–98. Tulchinsky TH, Ginsberg GM, Iscovich J, Shihab S, Fischbein A, Richter ED. 1999. Cancer in ex-asbestos cement workers in Israel, 1953–1992. Am J Ind Med 35(1):1–8. Tuomi T, Huuskonen MS, Tammilehto L, Vanhala E, Virtamo M. 1991. Occupational exposure to asbestos as evaluated from work histories and analysis of lung tissues from patients with mesothelioma. Br J Ind Med 48(1):48–52. UICC. 1965. Report and recommendations of the Working Group on Asbestos and Cancer. 132 ed. USA. Ulvestad B, Kjaerheim K, Martinsen JI, et al. 2002. Cancer incidence among workers in the asbestos-cement producing industry in Norway. Scand J Work Environ Health 28(6):411–417. Unfried K, Schurkes C, Abel J. 2002. Distinct spectrum of mutations induced by crocidolite asbestos: clue for 8-hydroxydeoxyguanosine-dependent mutagenesis in vivo. Cancer Res 62(1):99–104. van Kaick G, Dalheimer A, Hornik S, et al. 1999. The german thorotrast study: recent results and assessment of risks. Radiat Res 152(6 Suppl): S64–S71. Vianna NJ, Polan AK. 1978. Non-occupational exposure to asbestos and malignant mesothelioma in females. Lancet 1(8073):1061–1063. Virta RL. 2003. Worldwide asbestos supply and consumption. Trends from 1900 to 2000. U.S. Geological Survey. U.S. Department of the Interior. Washington, DC. Wagner JC, Sleggs CA, Marchand P. 1960. Diffuse pleural mesothelioma and asbestos exposure in the North Western Cape Province. Br J Ind Med 17:260–271. Watkins DK, Chiazze L Jr, Fryar C. 1997. Historical cohort mortality study of a continuous filament fiberglass manufacturing plant. II. Women and minorities. J Occup Environ Med 39(6):548–555. Weiss W. 1993. Asbestos-related pleural plaques and lung cancer. Chest 103(6):1854–1859.

Pleural and Peritoneal Neoplasms Woitowitz HJ, Lange HJ, Beierl L, et al. 1986. Mortality rates in the Federal Republic of Germany following previous occupational exposure to asbestos dust. Int Arch Occup Environ Health 57(3):161–171. World Health Organization. 1987. Air quality guidelines for Europe. WHO Regional Publications, European Series No. 23. Copenhagen: World Health Organization Regional Office for Europe, 182–199. Yano E, Wang ZM, Wang XR, Wang MZ, Lan YJ. 2001. Cancer mortality among workers exposed to amphibole-free chrysotile asbestos. Am J Epidemiol 154(6):538–543.

673

Yazicioglu S, Ilcayto R, Balci K, Sayli BS, Yorulmaz B. 1980. Pleural calcification, pleural mesotheliomas, and bronchial cancers caused by tremolite dust. Thorax 35(8):564–569. Zhu H, Wang Z. 1996. Cancer epidemiology investigation for workers exposed to asbestos. In: The Ministry of Public Health of China, ed. National Epidemiological Study on 8 Occupational Cancers. 1–121.

35

Cancers of the Oral Cavity and Pharynx SUSAN T. MAYNE, DOUGLAS E. MORSE, AND DEBORAH M. WINN

T

his chapter reviews the epidemiology of cancers of the oral cavity, pharynx, lip, and salivary glands. The oral cavity and pharyngeal (hereafter referred to as OCP) cancers are composed of the following tumor sites/codes (International Classification of Diseases for Oncology, 3rd ed; Fritz et al., 2000): oral cancers [base of tongue (C01), other and unspecified parts of the tongue (C02), gum (C03), floor of mouth (C04), palate (C05), other and unspecified part of mouth (C06)], and pharyngeal cancers [tonsil (C09), oropharynx (C10), pyriform sinus (C12), hypopharynx (C13), and other and ill-defined sites in lip, oral cavity, and pharynx (C14)]. The OCP cancer sites are discussed together because they share risk factors. Cancers of the lip (ICD-O C00) and major salivary glands [parotid gland (C07), other and unspecified major salivary glands (C08)] are considered separately at the end of the chapter, given their different epidemiologic characteristics. Cancers of the nasopharynx (ICD-O code C11) are discussed elsewhere (see Chapter 31). In the United States (1996–2000), invasive cancers of the OCP/lip/salivary gland account for 2.7% of cancers among men and 1.5% of cancers among women (Ries, 2003). It is estimated that 27,700 cases will be diagnosed with these malignancies in the United States in 2003 (18,200 men and 9500 women), and about 7200 will die from these cancers (Ries, 2003). The lifetime risk of being diagnosed with OCP/lip/salivary gland cancers for a US male is 1.4% and 0.7% for a US female (Ries, 2003). Worldwide, cancers of the OCP, lip, and salivary gland were responsible for an estimated 390,000 new cases and 207,000 deaths in 2000 (Parkin et al., 2001). The etiology of these cancers is reasonably well understood, with lifestyle factors, particularly tobacco and excessive alcohol exposures, accounting for the vast majority. Thus, efforts to reduce the burden of these cancers should emphasize primary prevention as discussed below.

ORAL CAVITY AND PHARYNGEAL CANCERS Classification Anatomic Distribution The distribution of OCP cancers by anatomical site can vary by geographic region. In the United States (1996–2000), age-adjusted incidence rates for all races combined were highest for cancers of the tongue (Table 35–1).

Histopathology Most OCP cancers are squamous cell carcinomas arising from the surface mucosa. Less frequently, other carcinomas, including those of the minor salivary glands, as well as sarcomas and a variety of rare and metastatic cancers may develop in the OCP. Based upon 1992–1999 data from 11 US SEER registries, squamous cell carcinoma accounted for approximately 90% of all invasive OCP cancers while an additional 5% of registered malignancies were classified as either adenocarcinoma or mucoepidermoid carcinoma (SEER, 2002). In the United States (1996–2000), 96% of malignant OCP cancers are histologically confirmed (SEER, 2003c).

Premalignant Lesions Oral cavity and pharyngeal cancers are often preceded clinically by precursor lesions and conditions, the most established of which include oral leukoplakia, erythroplakia, and oral submucous fibrosis

674

(OSF). Oral leukoplakia is a clinical term for “a predominantly white lesion of the oral mucosa that cannot be categorized as any other definable lesion” (Pindborg et al., 1997). Analogously, the term oral erythroplakia is used to designate similarly defined red lesions of the oral mucosa (Axell et al., 1996; Pindborg et al., 1997). Although not a lesion per se, OSF presents with “epithelial atrophy and fibrosis of the subepithelial connective tissue, resulting in stiffness of the oral mucosa” (Pindborg et al., 1997). Upon microscopic examination, premalignant lesions and conditions can exhibit oral epithelial dysplasia (OED), a histopathologic designation characterized by “cellular atypia and loss of normal maturation and stratification short of carcinoma in situ” (Pindborg et al., 1997). Oral leukoplakia has a higher transformation rate to cancer than normal oral mucosa and is therefore considered precancerous. On the other hand, many leukoplakias, particularly those with a clinically homogeneous appearance, are benign hyperkeratotic lesions with a low malignant potential (Waldron and Shafer, 1975; Gupta et al., 1989). The probability of malignant transformation is much higher in those leukoplakic lesions characterized clinically with a red, speckled, verrucous, or nodular component and in those with a histopathologic diagnosis of epithelial dysplasia (Silverman et al., 1984; Gupta et al., 1989; Schepman et al., 1998; Cowan et al., 2001). In one US study, the malignant transformation rate after a mean follow-up of 7.2 years was 6.5% for homogeneous leukoplakia, 23.4% for leukoplakias with a red component, and 36.4% for leukoplakia with microscopically diagnosed dysplasia (Silverman et al., 1984); other studies, however, have reported lower rates. There is also evidence that the risk of malignant transformation among leukoplakia cases may be higher among non-smokers than among smokers (Einhorn and Wersall, 1967; Silverman et al., 1984; Schepman et al., 1998) and among betel quid chewers relative to nonusers (Shiu et al., 2000). Erythroplakia is far more likely than most oral leukoplakias to contain epithelial dysplasia, putting these lesions at greater risk of impending malignancy; they are also more likely to contain carcinoma in situ, or invasive carcinoma (Shafer and Waldron, 1975; Mashberg, 1978). Estimates of the malignant transformation rate for OSF are limited; however, one Indian study reported a rate of 7.6% over a median observation period of 10 years (Murti et al., 1985). While the histopathologic examination of biopsied precancerous lesions and conditions remains the gold standard for assessing malignant risk, the use of various biomarkers, including loss of heterozygosity at specific microsatellite loci, and loss of pRb and accumulation of p53, has shown promise in predicting subsequent malignant transformation (Mao et al., 1996; Rosin et al., 2000; Soni et al., 2005).

Molecular Genetic Characteristics of Tumor Carcinogenesis of the oral cavity and pharynx is a consequence of multiple molecular events. Both genes and the environment (chronic exposure to tobacco, alcohol, and possibly certain viruses such as HPV16) are responsible for producing and promoting these molecular alterations. This multitude of molecular events affects numerous chromosomes and genes, and it is believed that it is the accumulation of multiple genetic changes that pushes cells towards malignancy (see “Pathogenesis” below). Alterations in genes involved in cell signaling, cell cycles, tumor suppression, and angiogenesis are all found in OCP cancers. Cytogenetic and molecular studies have demonstrated that somatic mutations that activate oncogenes (e.g., Ras, Myc, ErbB2,

Cancers of the Oral Cavity and Pharynx Table 35–1. Distribution of Oral and Pharyngeal Cancer Incidence (AgeAdjusted to 2000 US Standard Population) by Anatomical Site, Race, Hispanic Origin, and Sex, 1996–2000 Incidence Rate per 100,000 Person-Years

Total oral and pharynx, lip, and salivary gland* Total oral + pharynx* Anatomic subsite* Tongue Floor of mouth Gum/other mouth Tonsil Oropharynx Hypopharynx Other oral cavity/pharynx Race* White Black American Indian/Alaskan Native Asian or Pacific Islander Hispanic Origin† Hispanic Non-Hispanic Lip Race* White Black American Indian/Alaskan Native Asian or Pacific Islander Hispanic Origin† Hispanic Non-Hispanic Salivary gland Race* White Black American Indian/Alaskan Native Asian or Pacific Islander Hispanic Origin† Hispanic Non-Hispanic

Total

Males

Females

M : F Ratio

10.2

15.1

6.1

2.5

7.9

11.6

4.7

2.5

2.5 0.9 1.7 1.3 0.3 0.9 0.3

3.7 1.2 2.1 2.1 0.5 1.5 0.5

1.6 0.5 1.4 0.5 0.2 0.4 0.2

2.3 2.4 1.5 4.2 2.5 3.8 2.5

7.9 10.7 3.9 4.9

11.5 18.2 6.7 7.1

4.8 5.1 3.1

2.4 3.6 — 2.3

4.8 8.2 1.0

7.3 12.1 1.9

2.8 4.9 0.4

2.6 2.5 4.8

1.2

2.2

0.4

5.5 — — —

‡

‡

‡

‡

‡

‡

‡

‡

‡

‡

0.6 1.0 1.2

1.1 1.9 1.6

0.2 0.4 1.0

5.5 4.8 1.6

1.3 0.9

1.7 1.2

1.0 0.8

0.8

1.0

0.7

1.7 1.5 — 1.4

0.7 1.3

0.9 1.7

0.7 1.0

1.3 1.7

‡

‡

‡

Source: *Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov). SEER*Stat Database: Incidence—SEER 12 Regs, Nov 2002 (1973–2000), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2003 (SEER, 2003c). †Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov). SEER*Stat Database: Incidence— SEER 11 Regs, Nov 2002 Sub for Hispanics (1992–2000), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2003 (SEER, 2003d). Surveillance Research Program, National Cancer Institute SEER*Stat software (seer.cancer.gov/seerstat) version 5.0.20. (Surveillance Research Program). ‡ Less than 25 cases.

EGFR, bcl2, int-2, hst-1, ems-1, cyclinD1) as well as point mutations, deletions, or hypermethylation that lead to tumor suppressor gene inactivation (e.g., p16, TP53, PTEN, Rb) are involved in and contribute to the development of these cancers (Tassi and Wellstein, 2003).

Demographic Patterns Incidence and Mortality in the United States The incidence rate (1996–2000) for OCP cancers is 7.9 in 100,000 persons per year in the United States, with higher incidence rates in men (11.6/100,000) than women (4.7/100,000)(SEER, 2003c). Corresponding mortality rates in the United States (SEER 2003a) are 2.5 in 100,000 for men and women combined, and 3.5 in 100,000 for men and 1.4 in 100,000 for women. OCP cancer mortality shows marked geographic variation in the United States, with high rates noted for white males and females along the eastern seaboard. Among black males and females, rates are high in the mid-Atlantic and Florida, with

675

high rates also noted in parts of the Northeast for black males (Devesa et al., 1999).

United States Time Trends In the United States, age-adjusted (US 2000 population) incidence rates for OCP cancers are available for 9 SEER sites over the years 1973–2000 (SEER, 2003b). During that period, the most notable trend was for black males, for whom annual rates increased from around 16 in 100,000 in the early 1970s to over 25 in 100,000 during much of the 1980s before declining to less than 20 in 100,000 by the year 2000 (Fig. 35–1). Through the 1970s and into the 1980s, rates also increased for black females and less notably for whites before showing signs of decline during much of the 1990s. There are also differences in trends for different histologic types and by race and sex for specific anatomic sites. Between the 1975–1982 time period and the 1992–1998 time period, squamous cell carcinomas declined in both race and sex groups, whereas adenocarcinoma rates increased, a pattern also observed for esophageal cancer (Devesa et al., 1998). Among black and white males and females, declines were evident for lip cancer (except white females), gum, floor of mouth, other mouth, pyriform sinus, hypopharynx (except black males), and other and non-specified OCP cancer (except black females). Salivary gland cancer decreased by 10% among black men, but increased by 14% among white men, 5.9% among white females, and 39.7% among black females. Another cancer site that has been increasing over time is the tongue, which increased by 10.7% among white males and 5.2% among white females, but declined among black males (-5.5%) and females (-14.7%). Palate cancer increased among white males (8.0%), did not change among black males, and decreased among females by 15.2% among whites and 4.8% among blacks (Canto and Devesa, 2002).

Survival Overall 5-year relative survival (Table 35–2) for OCP cancers is 57.2% (55.7% in men and 60.4% in women; Ries et al., 2003). Survival varies by site, with lip cancers having the highest survival rates (94.4%), and hypopharynx (30.9%) and oropharynx cancers (37.3%) having the lowest rates. Survival is also heavily dependent on stage at time of diagnosis, with 5-year overall survival of 82.1% for localized disease, 47.9% for regional disease, and 26.1% for distant disease. OCP cancers are most commonly diagnosed as regional disease (48%), with localized disease the next most common (34%). Survival is much poorer for blacks than whites at every stage at diagnosis.

Age, Sex, Race, and Ethnicity As with most epithelial cancers, incidence rates for OCP cancer typically increase with age. Oral cancer is rare in children and young adults. By 35–39 years of age rates are 3.1 in 100,000 per year, increasing up to 41.1 by ages 65–69 up to 46.4 for 80–84 year olds, and declining to 41.3 in the oldest age group, 85 and older. Rates plotted on age, however, differ for black and white males and females (Fig. 35–2). As shown in Table 35–1, rates for total OCP cancer in males exceed those for females by 2.5-fold, with lip and tonsillar cancers having the highest sex ratios (4.8 and 4.2, respectively). Overall OCP rates are highest among black men (18.2/100,000), followed by white men (11.5/100,000) (Table 35–1). Among men rates for American Indian/Alaskan natives (6.7/100,000) and Asian or Pacific Islanders (7.1/100,000) are less than half that for black males. Among females, rates for whites and blacks are similar, 4.8 in 100,000 and 5.1 in 100,000 respectively, and are considerably higher than for Asian and Pacific Islanders, 3.1 in 100,000. Among Hispanics, rates for males are 7.3 in 100,000 and for females 2.8 in 100,000; corresponding rates for non-Hispanics are 12.1 in 100,000 and 4.9 in 100,000.

Socioeconomic Status Associations between low socioeconomic status, education and income, and OCP cancer have been observed in many studies in the United States (Williams and Horm, 1977; Greenberg et al., 1991; Kabat et al., 1994; Hayes et al., 1999) and other countries (Franco et

676

PART IV: CANCER BY TISSUE OF ORIGIN

Figure 35–1. Oral and pharyngeal cancer age-adjusted (2000 USA) incidence rates by race and sex, nine SEER Sites, all ages, 1973–2000 (excluding lip, major salivary glands, and nasopharynx).

al., 1989; Franceschi et al., 1990; Dikshit and Kanhere, 2000). Low versus high educational attainment, social class, and income are associated with increased smoking and alcohol use (Shohaimi et al., 2003). However, the effect of low income and education appears to be independent of these risk behaviors in some (Zheng et al., 1990b; Dikshit and Kanhere, 2000), but not all (Greenberg et al., 1991) of the studies that have controlled for risk behaviors.

Table 35–2. Five-Year Relative Survival Rates (%) for US Men and Women with Oral and Pharyngeal Cancers by Anatomical Site, Race and Sex, and Stage at Diagnosis, 1992–1999

Anatomic subsite Tongue Floor of mouth Gum/other mouth Tonsil Oropharynx Hypopharynx Other oral cavity/pharynx Lip Salivary gland Total all sites above

Overall

Males

Females

53.1 51.7 54.8 54.0 37.3 30.9 32.3 94.4 74.7 57.2

51.1 50.8 48.3 54.0 37.4 30.4 34.7 95.4 70.0 55.7

56.9 53.9 64.1 54.1 37.4 32.5 26.0 89.7 80.3 60.4

Whites

Total all sites above By stage Localized Regional Distant Unstaged Stage distribution Localized Regional Distant Unstaged

Blacks

Males

Females

Males

Females

58.9

61.2

30.7

50.6

82.7 50.3 25.7 44.1

82.6 48.8 29.0 48.0

62.7 28.0 16.1 19.5

77.7 39.1 34.9 51.0

35 47 9 9

41 43 8 8

15 60 15 10

28 51 13 8

Stage distribution at diagnosis (%) is also shown. Source: Ries et al., 2003.

Figure 35–2. Age-specific incidence rates for cancer of the oral cavity and pharynx by race and sex, 1990–2000, nine SEER Registries (excludes cancers of the lip, major salivary glands, and nasopharynx). (Source: SEER Program, November, 2002, released April, 2003.)

Cancers of the Oral Cavity and Pharynx Low occupational status has also been linked to OCP cancer risk (Elwood et al., 1984). One US population-based case-control study compared the percent of potential working life spent in employment by male cases with OCP cancer with controls. A low percentage of years worked, possibly an indicator of discontinuity of work history and social and economic instability, was linked to risk of these cancers among men (Greenberg et al., 1991), controlling for tobacco, alcohol, and other risk factors. Deprivation is also associated with oral cancer risk. In the Northeast of England, oral cancer incidence and mortality from the mid1970s to the early 1990s were both positively correlated with geographic area indicators of deprivation (O’Hanlon et al., 1997). Dietary or immunologic factors could potentially explain some of these findings.

International Incidence Rates and Trends, By Sex Oral cavity and pharyngeal cancer incidence varies by geographic area, and rates in a given region are almost always higher among males than females (Fig. 35–3). Based upon data reported in Cancer Incidence in Five Continents, Volume VIII for the period approximating 1993–1997, age-adjusted (world) annual incidence rates for males were highest in Somme and Bas-Rhin, France where rates exceeded 40 in 100,000 (Parkin et al., 2002). For females, the highest incidence rates were reported for South Karachi, Pakistan and Bangalore, India, with rates in excess of 10 in 100,000 per year. US rates are intermediate in comparison to other countries. The incidence of oral cavity

Males

677

(tongue and other mouth) cancer exceeds that of pharyngeal cancer in most geographic areas (Fig. 35–3); however, in some regions the reverse can occur, particularly among males. For example, in some areas of France (Manche, Bas-Rhin, Doubs) and Switzerland (Valais), pharyngeal cancer accounted for over 60% of the OCP cancer incidence among males. Trends in OCP cancer incidence also vary by geographic area and gender (Figs. 35–4 and 35–5). Based upon those regions of the world included in Cancer Incidence in Five Continents, volumes III through VIII and approximating the period 1968–1972 to 1993–1997 (Waterhouse et al., 1976; Waterhouse et al., 1982; Muir et al., 1987; Parkin et al., 1992; Parkin et al., 1997; Parkin et al., 2002), age-adjusted (world) rates for males registered a net decline of over 30% in Puerto Rico, Mumbai (Bombay), India, and Cali, Columbia as well as among populations living in Israel (non-Jews, Jews born in Israel) and Singapore (Indians, Malays) (Fig. 35–4). Over the same time period, ageadjusted rates increased over 100% in portions of Germany (Saarland), Poland (Warsaw city, Cracow), Spain (Zaragoza), and Japan (Miyagi, Osaka) as well as in Denmark. For females (Fig. 35–5), age-adjusted incidence rates fell over 30% in Puerto Rico, among Singaporean Indians and among Jews born in Israel, but at least doubled in areas of Germany (Saarland), and Switzerland (Geneva) as well as in Denmark and Alberta, Canada. During the last quarter of the 20th century, reports from the United States, Europe, and India identified increasing incidence and mortality trends among young adults (primarily males) for tongue

Females AMERICA US, SEER, Black US, Puerto Rico Uruguay, Montevideo Brazil, Campinas US, SEER, White Canada

Costa Rica EUROPE France, Somme Slovakia

Switzerland, Vaud Germany, Saarland Italy, North East Spain, Asturias Estonia

Austria, Vorarlberg Lithuania Czech Republic Denmark Poland, Warsaw City The Netherlands UK, England Sweden Finland ASIA

India, Ahmedabad Pakistan, South Karachi Taiwan

India, Bangalore Thailand, Bangkok Oral Cavity

Japan, Osaka

Pharynx

Israel, Jews China, Shanghai OCEANIA Australia, New South Wales New Zealand

50

40

30

20

10

0

10

Rate per 100,000 person-years

20

Figure 35–3. Age-adjusted (world) incidence rates for cancer of the oral cavity and pharynx (excluding the lip, major salivary glands, and nasopharynx), selected geographic regions, circa 1993–1997, all ages.

678

PART IV: CANCER BY TISSUE OF ORIGIN 2000). In Taiwan, increasing rates of OCP cancer observed during the 1980s and 1990s have been attributed to a rise in the consumption of alcohol and use of betel quid (Ho et al., 2002).

Migration Migrants tend to retain the risks of oral cancer from their country of origin. Migrants to Australia from the British Isles, Southern Europe, and Eastern Europe all had lower death rates for OCP cancer initially and after 30 years compared with Australian natives (McCredie et al., 1994; Warnakulasuriya, 2002). In the Thames region in Southern England, oral cancer incidence was strongly correlated with the percentage of the population in the local area who were Asian (Warnakulasuriya et al., 1999). Also, migrants from the Indian subcontinent to England and Wales have higher death rates from OCP cancer than natives (Swerdlow et al., 1995). These relationships presumably are due to the immigrants retaining their tobacco behaviors in their new home country (Khan et al., 2000).

Demographic Patterns: Oral Premalignant Lesions The prevalence of oral premalignant lesions and conditions varies by geographic region, population exposure patterns, and the case definition employed. The reported prevalence of oral leukoplakia in adult populations is generally within the range of less than 1%–5%, although substantially higher estimates have been reported for

Figure 35–4. Age-adjusted (world) incidence rates for cancers of the oral cavity and pharynx, circa 1968–1972 to 1993–1997, selected geographic regions, males, all ages (excludes cancers of the lip, major salivary glands, nasopharynx, and pharynx, NOS).

(Depue, 1986; Davis and Severson, 1987; Swango, 1996), mouth (Gupta, 1999), and OCP cancer (Franceschi et al., 1994; Devesa et al., 1995; Levi et al., 1995). In the United States (9 SEER sites), for example, incidence rates for cancer of the tongue among persons aged 0–39 years increased with an estimated annual percentage change of 6.7% from 1973 until 1985 before plateauing for the remainder of the century while rates for persons aged 40+ increased only modestly (Schantz and Yu, 2002; SEER, 2003c). The explanation for the reported increase in rates among young persons remains unclear. In regions of Europe (Denmark, Slovakia, Scotland, England/ Wales) and the United States (Connecticut) as well as in New Zealand, trends in age-standardized incidence rates reported for various time periods during the second half of the 20th century were, in many instances, birth-cohort based, particularly among males. Rates began to increase for cohorts born in the early decades of the century, and continued to rise for subsequent cohorts in a number of geographic regions (Moller, 1989; Macfarlane et al., 1992; Plesko et al., 1994; Cox et al., 1995; Hindle et al., 1996; Morse et al., 1999; Robinson and Macfarlane, 2003). In Slovakia, rising incidence rates during 1968–1989 were in keeping with increases in the per capita consumption of both tobacco and alcohol (Plesko et al., 1994) while in Denmark, Scotland, England, and Wales, trends in incidence were more consistent with changes in the consumption of alcohol than with that of tobacco (Moller, 1989; Macfarlane et al., 1992; Hindle et al.,

Figure 35–5. Age-adjusted (world) incidence rates for cancers of the oral cavity and pharynx, circa 1968–1972 to 1993–1997, selected geographic regions, females, all ages (excludes cancers of the lip, major salivary glands, nasopharynx, and pharynx, NOS).

Cancers of the Oral Cavity and Pharynx populations engaging in high-risk behaviors (Sciubba, 1995; Banoczy et al., 2001; Yang et al., 2001). There are few reports of erythroplakia prevalence, but available estimates are near or below 0.1% (Lay et al., 1982; Zain et al., 1997). OSF is observed primarily in areas of the Indian subcontinent, Southeast Asia, Taiwan, and Melanesia as well as in their migrant populations (Gupta and Warnakulasuriya, 2002). The prevalence of OSF generally ranges from less than 1%–3% (Trivedy et al., 2002), but can exceed 10% in some sub-populations (Gupta et al., 1998b; Yang et al., 2001).

Environmental Factors Tobacco Both tobacco use and alcohol consumption are well-established, important risk factors for OCP cancer, regardless of the type of tobacco product or alcoholic beverage. Tobacco can be used in smoked or unsmoked forms and formulated using a wide range of tobaccos, with varied processing procedures, inclusion of other ingredients, usage patterns, and vehicles for delivery, such as cigarette, bidi, pipe, and cigar.

Cigarette Smoking. Human evidence supporting a causal role for smoking in the etiology of OCP cancer comes from numerous casecontrol and cohort studies. In large cohort studies, smokers had 1.5–4.9-fold greater mortality rates of OCP cancer than nonsmokers (National Cancer Institute, 1997). Based on data from case-control studies in which cigarette smoking was the smoking product used by nearly all of the study participants, relative risks for current smokers range from 3–12 controlling statistically for alcohol use and other potential confounding factors (Blot et al., 1988; Franceschi et al., 1990; Zheng et al., 1990b; Mashberg et al., 1993; Kabat et al., 1994; Lewin et al., 1998; Schlecht et al., 1999). Corresponding ranges for former smokers are 1.1–4.5 (Blot et al., 1988; Franceschi et al., 1990; Zheng et al., 1990b; Mashberg et al., 1993; Kabat et al., 1994; Lewin et al., 1998; Bosetti et al., 2000). Among non-users of alcohol, the group in which confounding by alcohol can be most easily ruled out, risks range from twofold to fivefold (Zheng et al., 1990b; Castellsague and Munoz, 1999; Hayes et al., 1999). Strong positive dose-response relationships between amounts used per day are nearly always evident (Blot et al., 1988; Franceschi et al., 1990; Mashberg et al., 1993; National Cancer Institute, 1997; De Stefani et al., 1998; Hayes et al., 1999). Heavy smoking was associated with excess risks of at least threefold for smokers of more than 25 cigarettes per day (Franceschi et al., 1990; De Stefani et al., 1998) or 40 per day (Blot et al., 1988; Mashberg et al., 1993; National Cancer Institute, 1997; Hayes et al., 1999) except for lower risks in a large study in China (Zheng et al., 1990b), and in many a greater than fivefold increased risk with heavy smoking was observed. In many studies, modest elevated risks are evident for smokers of less than one pack per day with most of the excess risk generally among those smoking 10–19 cigarettes per day (Franceschi et al., 1990; Zheng et al., 1990b; Mashberg et al., 1993), but only in females in one large US study (Blot et al., 1988). In addition to being influenced by numbers of cigarettes smoked, risks also increase with years of use (Zheng et al., 1990b; Mashberg et al., 1993; De Stefani et al., 1998) and overall cumulative amounts used (usually measured by pack years) (Zheng et al., 1990b; Marshall et al., 1992; Mashberg et al., 1993; De Stefani et al., 1998; Lewin et al., 1998) in a wide range of cultures. Persons who inhale the smoke are at greater risk than those who do not (Lewin et al., 1998). In some (Mashberg et al., 1993; De Stefani et al., 1998; Lissowska et al., 2003), but not all studies (Blot et al., 1988; Franco et al., 1989; Kabat et al., 1994; Hayes et al., 1999), persons smoking filter cigarettes experience higher risks of OCP cancer compared to those smoking non-filtered cigarettes. Hand-rolled cigarettes are more strongly associated with risk than commercially made cigarettes (De Stefani et al., 1998). OCP cancer risks for use of black tobacco (a high tar content noncommercial cigarette rolled in cornhusk leaves) were similar to those for commercial cigarettes in a study in Brazil (Schlecht et al., 1999), but were higher in comparison with blond tobacco ciga-

679

rettes in Uruguay (De Stefani et al., 1998) and in Italy (Merletti et al., 1989). Cigarette smoking cessation leads to lower risks of OCP cancer after a period of years (Lissowska et al., 2003). In one study risks reached the level of never smokers within 10 years of cessation (Blot et al., 1988). In other studies in the United States and Sweden, risk levels reached that of nonsmokers only after about 20 years (Lewin et al., 1998; Hayes et al., 1999). In another study, risks did not reach those of nonsmokers even 20 or more years after quitting smoking commercial cigarettes (Schlecht et al., 1999). The decline with years of cessation also varies by the type of tobacco smoked and possibly the anatomic subsite involved. A Brazilian study found that odds ratios for OCP cancer among former smokers of commercial cigarettes were close to one but still slightly elevated even after 20 or more years of not smoking black tobacco (Schlecht et al., 1999). In another Brazilian study, tongue cancer rates were higher after 10 years of smoking cessation than for other mouth sites (Franco et al., 1989).

Bidi Smoking. Bidi smoking is widespread in many parts of South Asia, particularly India (Gupta, 1996). Bidis are similar to cigarettes except that the tobacco is rolled in paper or leaves, usually by hand, and then smoked. Oral cancer risks are higher in bidi smokers than in nonsmokers even when other smoking behaviors, smokeless tobacco, and/or alcohol are taken into account (Sankaranarayanan et al., 1989; Sankaranarayanan et al., 1990; Rao and Desai, 1998; Dikshit and Kanhere, 2000; Balaram et al., 2002). Cigars and Pipes. Declines in the prevalence of cigar smoking in the United States occurred until 1993, but increased 50% in the subsequent 4 years (Gerlach et al., 1998). Based on a cohort of more than 500,000 persons observed for 12 years, current cigar smoking increased mortality from OCP cancers by fourfold. Mortality rates for former smokers were 2.5 and were higher for persons who smoked three or more cigars per day, inhaled, or smoked cigars for 25 years or more (Shapiro et al., 2000). In case-control studies, risks ranged from 3–9 and also were dependent on dose measured by frequency, duration, or cumulative dose (Boffetta et al., 1999; Schlecht et al., 1999; Garrote et al., 2001). The prevalence of pipe smoking has also been declining in the United States, to only a few percent by the early 1990s (Nelson et al., 1996). Controlling for other confounders such as cigarettes and alcohol consumption, pipe smoking increases risks for OCP cancer (Zheng et al., 1990b; Boffetta et al., 1999; Schlecht et al., 1999). Environmental Tobacco Smoke. In a New York casecontrol study of OCP cancers, environmental tobacco smoke was linked to an odds ratio of 2.8, after controlling for multiple risk factors including alcohol, cigarette smoking, and marijuana use (Zhang et al., 2000). Risks were evident even among nonsmokers and were greater among the more heavily exposed groups. There was greater than multiplicative interaction between environmental tobacco smoke exposure and mutagen sensitivity. Smoking and Anatomic Subsites. Some intra-oral sites appear to be more susceptible to the effects of cigarette smoking. The most conspicuous example is for reverse smoking, in which the lit end of the cigarette is placed in the mouth. This behavior is associated with high risks of hard palate cancer (Reddy et al., 1975). Greater effects of smoking on pharynx compared with oral cavity cancer risk have been noted by some investigators (De Stefani et al., 1998; Lewin et al., 1998; Franceschi et al., 1999b), but not consistently in all studies (Hayes et al., 1999; Schlecht et al., 1999). Investigators in Milan observed that odds ratios for oral cavity sites were about two times higher than for pharyngeal sites at every joint smoking and drinking consumption level (Franceschi et al., 1999b). In larger case-control studies, where there are more cases in each anatomic subsite, the floor of the mouth was the intra-oral site with the highest cigarette smoking risks in both men and women (Macfarlane et al., 1995), but the soft palate sites were at highest risk in another

680

PART IV: CANCER BY TISSUE OF ORIGIN

(Boffetta et al., 1992). Differences among studies in grouping anatomic subsites make risk comparisons difficult. Pipe and cigar smoking risks are higher for floor of mouth/buccal mucosa (Blot et al., 1988) and soft palate sites (Boffetta et al., 1992) compared with other subsites.

Smokeless Tobacco—Western Countries. In many cultures, use of smokeless tobacco products, in which the tobacco product is not smoked during use, is common and also is associated with an increased risk of OCP cancer. These products come in many forms (National Cancer Institute, 2002); this is the source for the descriptions of the products described in the “Smokeless Tobacco” sections. The United States and Sweden are countries where smokeless tobacco use is common. In the United States, moist snuff and chewing tobacco are most common, whereas moist snuff accounts for almost all of the smokeless tobacco used in Sweden. In South Asia, many forms are used and are combined with other ingredients. United States and Swedish snuffs, which are placed and retained between the cheeks or lips and the gums, are made from air- or firecured tobacco that is cut into a powder or small strips. Chewing tobacco consists of tobacco leaves and sweeteners. A recent systematic review noted the difficulties in evaluating the literature on smokeless tobacco and OCP cancer including small sample sizes of smokeless tobacco users and lack of control for confounding by smoking. Nevertheless, there are studies that restricted the analysis to only nonsmokers; significantly elevated odds ratios were observed in two studies in the United States (Winn et al., 1981; Blot et al., 1988) and in one (Lewin et al., 1998) of two in Sweden (Lewin et al., 1998; Schildt et al., 1998). N-nitrosamines are present in smokeless tobacco; other carcinogens in these products include polonium (Hoffmann et al., 1986). The amounts of nitrosamines vary considerably, but tend to be lower in Swedish snuff than US-manufactured snuff and can be affected by factors such as storage (Djordjevic et al., 1993). Smokeless Tobacco and Related Products—Asian, African, Pacific Island Countries. Areca nuts, the fruit of the Areca catechu palm, are chewed alone or in combination with other ingredients, including the Piper betle leaf, tobacco, slaked lime, and various spices, by an estimated 600 million persons (about 10%–20% of the world population), particularly cultures in Asia and the South Pacific (Gupta and Warnakulasuriya, 2002). Common terms for these combinations of ingredients are betel quid or pan. Tobacco is commonly added in some cultures, but not in others (Gupta and Warnakulasuriya, 2002). The use of areca nut and areca nut-based preparations has been linked to an increased risk of both oral leukoplakia and OSF. In one Taiwanese study, nonsmokers who chewed areca nut quid without tobacco had a 10-fold increase in the risk of oral leukoplakia and a 39-fold increase in the risk of OSF (Lee et al., 2003). Other studies have reported odds ratios of 100 or more for the frequent use of areca nuts or areca nut-containing products in relation to OSF (Maher et al., 1994; Hazare et al., 1998; Shah and Sharma, 1998). A large case-control study conducted in India observed elevated oral cancer odds ratios of 6 for men and 42 for women for use of tobaccocontaining betel quid, adjusting for smoking (Balaram et al., 2002). Increasing risks with increasing amount and duration of tobaccocontaining pan have been observed, with dose being defined according to frequency and/or duration (Sankaranarayanan et al., 1989; Sankaranarayanan et al., 1990; Rao et al., 1994; Wasnik et al., 1998; Balaram et al., 2002). The buccal mucosa is the cancer site most strongly associated with the practice (Rao et al., 1994; Merchant et al., 2000). A greater than additive interaction among smoking, drinking, and betel on the risk of oral cancer has been observed (Ko et al., 1995). Both pan with and without tobacco are classified as human carcinogens by the International Agency for Research on Cancer (IARC) (IARC Working Group on the Evaluation of the Carcinogenic Risk of Chemicals to Humans et al., 1985; IARC Working Group on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, 2004). Areca nuts contain 3-(methylnitrosamino)proprionaldehyde, classified by the IARC as having limited evidence of carcinogenicity in humans,

and 3-(methylnitrosamino)proprionitrile, which is possibly carcinogenic (IARC Working Group on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, 2004). Other Asian tobacco practices in which the product is kept in the mouth and not smoked linked to increased oral cancer risks include khaini, used in India and consisting of tobacco, slaked lime paste, and sometimes areca nut (Ghosh et al., 1996) and tobacco used as a dentifrice (Wasnik et al., 1998). As in US and Swedish smokeless tobaccos, nitrosamines are also in many smokeless tobaccos used in other cultures. The Sudanese product, called toombak, consists of sodium bicarbonate added to tobacco that has been fermented for several weeks, and has extremely high nitrosamine levels (Idris et al., 1998). A case-control study found excess risks reaching four- to seven-fold among users (Idris et al., 1995). Nass and naswar, made from tobacco, slaked lime, ash, and coloring and flavoring agents, are used in Central Asia, Pakistan, and India. Case-control studies have linked use of these high nitrosaminecontaining mixtures (Zaridze et al., 1991) with premalignant oral lesions in Uzbekistan (Zaridze et al., 1985). In one study, naswar was associated with a 10-fold increased risk of oral cancer controlling for smoking and alcohol (Merchant et al., 2000).

Marijuana Smoking. In one case-control study, marijuana smoking was associated with nearly a threefold increased risk of head and neck cancer. Risks increased with frequency and years of use. Some carcinogens found in tobacco smoke are present in marijuana and may occur in greater amounts in marijuana, and also there may be mutagens derived from the marijuana that are not present in tobacco smoke. In this study there were greater than additive interactions between mutagen sensitivity and marijuana use, tobacco smoking and marijuana use, and possibly marijuana and alcohol drinking (Zhang et al., 1999). Tobacco and Oral Premalignant Lesions. Most studies investigating tobacco as a risk factor for oral premalignancy have focused on oral leukoplakia. Both smoking and smokeless forms of tobacco have been linked to an elevated risk of leukoplakia, and smoking tobacco has been associated with an increased risk of epithelial dysplasia (Mehta et al., 1981; Baric et al., 1982; Grady et al., 1990; Evstifeeva and Zaridze, 1992; Morse et al., 1996; Hashibe et al., 2000b; Shiu et al., 2000). The reported risks are generally highest in current users and tend to increase with frequency of use (Gupta, 1984b; Evstifeeva and Zaridze, 1992; Morse et al., 1996; Hashibe et al., 2000b). For example, in a recent Indian study, odds ratios for oral leukoplakia in relation to current and past smoking were 3.4 and 1.7, respectively, while the corresponding odds ratios for tobacco used primarily as an ingredient in betel quid or pan were 9.4 and 3.9 (Hashibe et al., 2000b). In the United States where smokeless tobacco is generally used without other major ingredients, one study of professional baseball players found that the prevalence of oral leukoplakia was 15and 87-fold greater among users of either chewing tobacco or snuff, respectively, relative to nonusers (Grady et al., 1990). Tobacco cessation is associated with a reduction in the risk of both oral leukoplakia and OED (Gupta et al., 1995; Morse et al., 1996), and leukoplakic lesions often show partial or complete remission when tobacco use is discontinued (Roed-Petersen, 1982; Martin et al., 1999). The body of evidence relating tobacco use to erythroplakia and OSF is limited. Smoking tobacco has not been consistently linked to either erythroplakia or OSF (Shah and Sharma, 1998; Hashibe et al., 2000a; Hashibe et al., 2002; Lee et al., 2003); however, studies conducted in Kerala, India have revealed strong associations between these premalignancies and the use of chewing tobacco, which is most often consumed as part of an areca nut-containing quid (Hashibe et al., 2000a; Hashibe et al., 2002). Alcohol Epidemiologic evidence demonstrates that alcohol is an independent risk factor for OCP cancers. Overall risks associated with alcohol consumption vary among populations. In a large US study, drinking 15–29 drinks per week was associated with increased risks of OCP cancer of

Cancers of the Oral Cavity and Pharynx threefold for men and twofold for women, controlling for smoking (Blot et al., 1988), whereas in an Italian study persons drinking more than 60 drinks per week had lower smoking-adjusted risks of about 3.5 (Franceschi et al., 1990). Based on a meta-analysis (Bagnardi et al., 2001), smoking-adjusted alcohol-related relative risks were 1.8 for drinking the equivalent of 2 drinks per day, 2.9 for 4, and 6.0 for 8. No excess risk was observed for drinking between 1 and 4 drinks per week in a large US study (Blot et al., 1988) and 1–7 drinks per week in a study in Puerto Rico (Hayes et al., 1999). Alcohol risks are evident among nonsmokers (Blot et al., 1988; Franco et al., 1989; Hayes et al., 1999), providing evidence of alcohol as a cancerinitiating agent. Risks increase with increasing frequency of consumption and cumulative lifetime consumption (Blot et al., 1988; Franco et al., 1989; Franceschi et al., 1990; Zheng et al., 1990b; Mashberg et al., 1993; Lewin et al., 1998; Hayes et al., 1999; Bagnardi et al., 2001). Duration of use of alcohol in years is not strongly associated with OCP cancer risk (Blot et al., 1988; Merletti et al., 1989; Franceschi et al., 2000). Cessation of alcohol drinking appears to be associated with a decreasing risk. In one study, reductions in risks were observed only after 10 years of nondrinking (Franceschi et al., 2000). Risks were similar to that of never drinkers only after 20 or more years of cessation among men in Puerto Rico (Hayes et al., 1999). The alcohol type that predominates in a given culture tends to be the one that is associated with the highest risks at equivalent levels of consumption; for example, hard liquor and beer in the United States (Blot et al., 1988; Merletti et al., 1989; Lewin et al., 1998) and wine in Italy (Franceschi et al., 1990). This may be due to greater overall alcohol consumption in users of the most common type of alcohol. In a large US study, darker alcoholic beverages were associated with a two-times higher risk than lighter-colored ones for each anatomic subsite (Day et al., 1994a). Regionally popular alcoholic beverages linked to an increased oral cancer risk include among others cachaça, a distilled sugar-cane product (Franco et al., 1989) consumed in Brazil, and sake in Japan (Takezaki et al., 1996). Ways of consuming alcoholic beverages that would tend to increase the contact of ethyl alcohol with the oral mucosa may also play an important role in modifying alcohol-related risks. A case-control study in Puerto Rico (Huang et al., 2003) showed that risks were greater for beverages consumed in a more concentrated form. Drinking alcohol outside of meals is associated with a higher and similar risk of oral and of pharyngeal cancer, respectively, than alcohol taken with meals at each level of alcohol drinking (Dal Maso et al., 2002). The anatomic sites that are at highest risk of cancer from drinking alcoholic beverages have varied from study to study (Blot et al., 1988; Boffetta et al., 1992; Franceschi et al., 1999b; Hayes et al., 1999). Most studies that have examined interactions of alcohol and tobacco on risk find that those with the heaviest alcohol and tobacco behaviors have relative risks that are greater than additive and exceed 30-fold in many studies (Franco et al., 1989; Zheng et al., 1990b; Mashberg et al., 1993; Franceschi et al., 1999b; Hayes et al., 1999; Lissowska et al., 2003). In addition, synergism between smoking and drinking was observed for the oral cavity as well as the pharynx (Blot et al., 1988; Franceschi et al., 1999b).

Nonsmokers and Nondrinkers. Data are limited on characteristics of OCP cancers occurring in nonsmokers who also do not drink alcohol. These patients are more likely to be female (consistent with tobacco and alcohol drinking rates among women vs. men) (Blot et al., 1988; Boffetta et al., 1992; Hayes et al., 1999), are less likely to have a family history of OCP, larynx, and esophageal cancer combined (Brown et al., 2001), and have a similar median number of genetic aberrations as tobacco and alcohol users (Singh et al., 2002). Mechanisms of Alcohol Carcinogenesis. Proposed mechanisms of alcohol carcinogenesis with some experimental laboratory support involve alcohol as an initiator of carcinogenesis through the action of specific carcinogens contained in particular alcoholic beverages or via metabolism of ethyl alcohol to carcinogenic acetalde-

681

hyde in oral tissues. Alcohol may also act as a promoter of carcinogenicity through enhancement of the permeability of the oral mucosa to carcinogens such as those from cigarette smoke, mucosal atrophy, potentiation of genotoxicity from tobacco, and inhibition of DNA repair (Wight and Ogden, 1998). While some carcinogens are present in some alcoholic beverages (Wight and Ogden, 1998), the consistency of findings of higher risks of OCP cancer with greater frequency of use of different alcoholic beverages used in different cultures throughout the world, even among nonsmokers, suggests that metabolism of ethyl alcohol is the most important contributor to alcoholic beveragerelated carcinogenesis. Additionally, if alcohol is also a promoter, this might explain the synergism between alcohol and tobacco use in greatly increasing risks of OCP cancer in users of both products.

Alcohol and Oral Premalignant Lesions. The role of alcohol consumption as an independent risk factor for oral leukoplakia has not been clearly established, with some investigations reporting generally weak to moderate associations and others finding no relationship (Gupta, 1984a; Macigo et al., 1995; Hashibe et al., 2000b; Evstifeeva and Zaridze, 1992; Lee et al., 2003). Associations between alcohol consumption and OSF are also equivocal (Hashibe et al., 2002; Lee et al., 2003). There is evidence implicating alcohol as a risk factor for both oral erythroplakia (Hashibe et al., 2000a) and epithelial dysplasia (Morse et al., 1996), but the number of studies is limited. Use of alcohol-containing mouthwashes has not been associated with an increased risk of OED (Morse et al., 1997). Diet/Nutrition Foods/Food Groups. The literature on food, nutrients, and risk of OCP cancers was comprehensively reviewed by the World Cancer Research Fund in 1997 (World Cancer Research Fund/American Institute for Cancer Research, 1997). As for foods, 13 of 15 case-control studies of fruit and/or vegetable intake and OCP cancer risk reported a significant protective association for at least one vegetable and/or fruit category. One cohort study from Japan similarly found an inverse association between risk of OCP cancer and consumption of green and yellow vegetables (Hirayama, 1985). Five of seven studies that examined vegetables as a category reported inverse associations while eight of ten studies reported likewise for fruit. Several studies on diet and risk of OCP cancer have been published subsequent to this review. In agreement with prior studies, the most consistent finding to emerge is an inverse association with fruit and/or vegetable intake (Levi et al., 1998; De Stefani et al., 1999; Franceschi et al., 1999a; Negri et al., 2000; Takezaki et al., 2000; Garrote et al., 2001; Petridou et al., 2002; Uzcudun et al., 2002; Weinstein et al., 2002; Lissowska et al., 2003; Sanchez et al., 2003). The Joint WHO/UN Food and Agriculture Organization Expert Committee recently concluded that an inverse association between fruit and vegetable intake and oral cavity cancers is “probable,” the highest level of evidence given for any dietary association with any tumor site (Joint WHO/FAO Expert Consultation, 2003). Statistically significant odds ratios for total vegetable intake and OCP risk range from 0.14–0.6; corresponding odds ratios for total fruit intake range from 0.2–0.6 (Chainani-Wu, 2002). Statistically significant dose-response relationships are noted in approximately half of these studies (Chainani-Wu, 2002). Several studies also report a positive association between meat intake and risk of OCP cancers (Levi et al., 1998; De Stefani et al., 1999; Franceschi et al., 1999a; Garrote et al., 2001; Petridou et al., 2002; Uzcudun et al., 2002). Case-control studies of diet and cancer risk are susceptible to many biases including recall bias, so supporting data from cohort studies are helpful to the interpretation of associations. One recent cohort study of diet and risk of upper aerogastric tract cancers (oral, pharyngeal, laryngeal, esophageal) in Norwegian men reported that high consumption of oranges was associated with a significant risk reduction while frequent consumption of beef and bacon was associated with increased cancer risk (Kjaerheim et al., 1998). Another cohort study of US women reported that consumption of the highest versus lowest tertile of yellow/orange vegetables was

682

PART IV: CANCER BY TISSUE OF ORIGIN

associated with a 31% reduction in the risk of OCP cancer (Kasum et al., 2002). Dietary patterns are known to track with smoking behaviors (Dallongeville et al., 1998) so the interpretation of dietary associations with cancers that are strongly linked to tobacco is complex. Smokers consume fewer fruits and vegetables on average compared with nonsmokers, and there is the possibility that associations observed may at least in part reflect residual confounding by smoking. However, inverse associations with fruit and vegetable intake are noted throughout the globe and with use of different types of tobacco products and with control for tobacco use statistically or among non-tobacco users. For example, a case-control study of precancerous lesions from Kerala, India found that among women (none of whom smoked), an inverse association with beta-carotene (found in fruits and vegetables) was noted (55% reduction in risk per 1000 mg consumed) (Gupta et al., 1999). Also, results from both animal studies and intervention trials of micronutrients support the concept that diet relates to risk of OCP cancers as discussed in greater detail below.

Nutrients. Fruits and vegetables are the primary food sources for certain micronutrients such as carotenoids and vitamin C. Given the consistent inverse association noted for fruits and vegetables and risk of OCP cancers, it should not be surprising that these same nutrients are frequently inversely associated with risk. The World Cancer Research Fund review (World Cancer Research Fund/American Institute for Cancer Research, 1997) noted that five of five case-control studies that examined an association between vitamin C intake and risk of OCP cancer reported statistically significant inverse associations, a finding supported by more recent studies as well (Negri et al., 2000). Similarly, dietary intake of carotene/carotenoids has been inversely associated with risk in numerous studies (Franco et al., 1989; Gridley et al., 1990; Negri et al., 1993; Zheng et al., 1993a; Negri et al., 2000). Pre-diagnostic bloods from patients who subsequently developed OCP cancer had lower serum concentrations of all individual carotenoids (Zheng et al., 1993b). Because fruits and vegetables are the primary contributors to carotenoid and vitamin C levels, observational studies are inadequate to determine whether these nutrients per se are involved in the etiology of OCP cancers. Human intervention trials (see “Preventive Measures,” below) and animal studies of these nutrients, however, are informative. The animal model most closely related to human oral carcinomas is the hamster buccal pouch model. Numerous studies have evaluated the ability of micronutrients to regress or inhibit tumors in this model. Beta-carotene has been widely studied; evidence that beta-carotene as a single nutrient can inhibit tumor formation in this animal model has been deemed “convincing” by the International Agency for Research on Cancer (International Agency for Research on Cancer, 1998).

tion of sera positive for HPV type 16 capsid antibody was 35% in controls, 51.4% overall in cases, and 75.7% in the 37 case subjects whose tumors were found to contain HPV16 DNA (Schwartz et al., 1998). The odds ratio for oral cancer was 6.8 for those who were HPV16positive in both sera and tissue as compared to controls with HPV negative sera. The largest study to date was a multi-country study that enrolled 1670 case patients and 1732 controls (Herrero et al., 2003). HPV DNA was found in biopsy specimens (case only analysis) from 3.9% of oral cavity cancers and 18.3% of oropharyngeal tumors. Having antibodies against HPV16 E6 or E7 (case-control analysis) was associated with 2.9- and 9.2-fold increased risks of oral and oropharyngeal cancers, respectively. The presence of viral DNA in tumors does not necessarily imply a causal role, because the cancer may have activated the virus, but prospective data also support an association. Using a nested casecontrol design, serum samples from 292 OCP cancer cases and 1568 matched controls were evaluated for antibodies against HPV16, HPV18, HPV33, and HPV73 (Mork et al., 2001). Subjects who were seropositive for HPV16 had a 2.2-fold increased risk of subsequent OCP cancer with an average of 9.4 years of follow-up. No risk was observed for other HPV types. Paraffin-embedded tissues from cases were also evaluated; 50% of oropharyngeal tumors (9/18) contained HPV16 DNA, as did 14% of tongue tumors (4/29). Other evidence supporting the temporal relationship between HPV infection and subsequent malignancy comes from analyses of tumor registry data. In a US study using SEER data, standardized incidence rates of buccal cavity cancers were elevated for both white (standardized incidence ratio 2.0) and black (standardized incidence ratio 3.5) women with prior cervical malignancies (Spitz et al., 1992). In another analysis of SEER data, patients with HPV-associated anogenital cancers had a 4.3-fold increased risk of tonsillar squamous cell carcinoma (Frisch and Biggar, 1999). An excess of tonsillar cancer among husbands of women with HPV-associated neoplastic lesions of the cervix was also noted in a Swedish study (Hemminki et al., 2000). HPV, especially16, may be an etiologic factor for a subset of OCP cancers characterized by a better prognosis (Gillison and Shah, 2001). In one study, HPV was detected in 62 (25%) of 253 cases (Gillison et al., 2000). High-risk HPV16 was detected in 90% of the HPVpositive tumors. OCP patients with HPV-positive tumors had a significant 59% reduction in the risk of cancer death as compared to those with HPV-negative tumors. In another study, patients with HPV16 DNA in their tumors had significantly reduced all-cause mortality (HR = 0.34) and disease-specific mortality (HR = 0.17) (Schwartz et al., 2001b). Future HPV vaccine trials in patients with HPV16-positive oropharyngeal cancers (Erdmann, 2003) may help to further clarify this association.

Mouthwash Use Diet and Oral Premalignant Lesions. Relatively few observational epidemiologic studies have investigated the relationship between oral precancer and diet. Most consistently, these investigations have identified an apparent, although not always statistically significant, inverse association with the consumption of at least some fruits and/or vegetables and risk of oral premaligancy (Gupta et al., 1998a; Gupta et al., 1999; Hashibe et al., 2000a; Morse et al., 2000).

Human Papillomaviruses A possible viral etiology for OCP cancers was suggested following the recognition that viral DNA (herpes simplex virus and/or human papillomavirus) was present in some OCP tumor tissues (Loning et al., 1985; Cox et al., 1993). Subsequent studies have particularly focused on HPV, with numerous studies showing that HPV DNA is present in OCP cancers, although with widely discordant prevalence rates (Shah, 1998), likely due to various assay methodologies and geographic differences in infection rates. HPV DNA has also been detected in normal buccal mucosa (Scully, 2002); however, case-control studies suggest that HPV DNA is more commonly detected in case as compared to control tissues. In a US study, infection with HPV16 was associated with a 6.2-fold increased risk of oral cancer (Maden et al., 1992). In a large US case-control study (284 cases, 477 controls), the propor-

Interest in a possible link between mouthwash use and oral cancer arose when Weaver et al. reported that 10 of 11 oral cancer cases who were both nondrinkers and nonsmokers had a history of frequent daily mouthwash use for over 20 years, with the majority of cases having used undiluted mouthwash high in alcohol content (25%) (Weaver et al., 1979). Since that report, a number of case-control studies have evaluated the association between mouthwash use and risk (Blot et al., 1983; Wynder et al., 1983; Kabat et al., 1989; Winn et al., 1991; Talamini et al., 2000; Garrote et al., 2001; Winn et al., 2001). Although some investigations have reported findings suggestive of an association in one or another subgroup or when evaluating one or another frequency- or duration-response trend, there has been an overall lack of consistency in findings across studies, and to date, a relationship between oral cancer and mouthwash use remains equivocal.

Oral Hygiene, Missing Teeth, and Other Dental Factors A growing number of case-control studies from disparate geographic regions have evaluated various measures of oral hygiene and dentition as potential risk factors for oral cancer while controlling for known confounders, including tobacco and alcohol. Studies often report a weak-to-moderate positive association between oral cancer and infre-

Cancers of the Oral Cavity and Pharynx quent toothbrushing; however, odds ratios do not always achieve statistical significance (Franco et al., 1989; Kabat et al., 1989; Zheng et al., 1990a; Velly et al., 1998; Talamini et al., 2000; Garrote et al., 2001). Similarly, a number of investigations have identified a positive association between oral cancer and having more, relative to fewer, missing teeth (Zheng et al., 1990a; Marshall et al., 1992; Bundgaard et al., 1995; Garrote et al., 2001; Balaram et al., 2002). There is, however, little evidence that wearing dentures per se increases risk (Franco et al., 1989; Kabat et al., 1989; Zheng et al., 1990a; Winn et al., 1991; Velly et al., 1998; Talamini et al., 2000; Garrote et al., 2001; Balaram et al., 2002). Other potential oral cancer risk factors such as broken or jagged teeth (Franco et al., 1989; Zheng et al., 1990a; Marshall et al., 1992; Velly et al., 1998) and a history of dental X-rays (Zheng et al., 1990a; Winn et al., 1991; Marshall et al., 1992) have received relatively little attention; findings to date suggest little, if any, excess risk. It is possible that the bacteria associated with poor oral hygiene and dentition may increase the amounts of acetaldehyde produced from ingested alcohol (Homann et al., 1997; Tillonen et al., 1999; Muto et al., 2000), thus providing an explanation for the relationship between poor oral hygiene and dentition and oral cancer.

Occupation Several studies have specifically examined the role of occupation in the etiology of OCP cancers, and working in certain occupational groups has been associated with increased risks of OCP cancers. Occupational groups where excess risk has been noted include butchers (Boffetta et al., 2000), male carpet installers (Huebner et al., 1992), machinists (Merletti et al., 1991; Huebner et al., 1992), male leatherworkers (Decoufle, 1979), textile workers (Whitaker et al., 1979), women in the electronics industry (Winn et al., 1982), sugarcane farmers (Coble et al., 2003), and a variety of other occupations wherein blue collar workers are exposed to dusts, inhaled organic agents, or inhaled inorganic agents (Maier et al., 1990). There is little consistency to these findings, and interpretation is further complicated in that smoking, drinking, and socioeconomic status can vary across occupational groupings, making it difficult to isolate specific occupational effects. A large study linked data from the Finnish Cancer Registry to occupational codes from the population census, and concluded that the role of direct occupational factors in the etiology of any oral/pharyngeal sites seems to be minimal (Pukkala et al., 1994). In OCP cancer cases with pronounced tobacco and alcohol consumption, it thus seems that occupation plays a minor role; however, a role in less heavily exposed cases cannot be ruled out.

Host Factors Diseases Predisposing to Risk A number of medical conditions in addition to oral premalignant lesions have been associated in cohort studies with an increased risk of OCP cancer. Among them are alcohol-related conditions including alcoholism (Boffetta et al., 2001b) and cirrhosis of the liver (Keller, 1983). Two other conditions are Fanconi anemia (Kutler et al., 2003; Rosenberg et al., 2003), an inherited condition possibly involving DNA repair deficits (Tischkowitz and Hodgson, 2003), and psoriasis, possibly due to confounding by tobacco and alcohol or cell cycle and cell proliferation problems that might be associated with both psoriasis and cancer development (Boffetta et al., 2001a).

Familial Aggregation There is a modest to moderate degree of aggregation of OCP cancer in families. Patients with OCP cancer are 2–3 times more likely than other persons to have first-degree family members with OCP cancers (Goldgar et al., 1994; Goldstein et al., 1994; Brown et al., 2001), with upper aerodigestive tract cancers (includes OCP plus larynx and/or esophagus) (Jovanovic et al., 1992; Jovanovic et al., 1993a; Jovanovic et al., 1993b; Li and Hemminki, 2003; Goldstein et al., 1994; Jovanovic et al., 1994a; Jovanovic et al., 1994b; Copper et al., 1995; Foulkes et al., 1995; Foulkes et al., 1996; Mork et al., 1999), and with tobacco- and alcohol-related cancers more generally (Li and Hem-

683

minki, 2003). The risks of OCP cancer associated with having a relative with upper aerodigestive tract cancers may be higher for younger than older cases (Brown et al., 2001) and higher for cases who have multiple primary OCP cancers (Mork et al., 1999). These patterns may be due to the increased prevalence of smoking and heavy alcohol drinking in the relatives of smokers and alcohol drinkers. However, these findings for first-degree relatives also may be due to shared inherited genotypes such as those involved in the metabolism of tobacco carcinogens and alcohol, DNA repair, or predisposition to tobacco and alcohol behaviors and addictions (Cheng et al., 2000b; Swan et al., 1997) and alcoholism (Cheng et al., 2000a). By measuring DNA repair assessed by a mutagen sensitivity assay, odds ratios of three for having one or seven for having multiple family members with cancer were observed in one study (Bondy et al., 1993) and a linear relationship of risk with number of relatives affected with cancer in another (Bondy et al., 1993; Yu et al., 1999). In the latter study, an additive interaction between having a family history of cancer and mutagen sensitivity was observed (Yu et al., 1999). Very limited data are available on risk behaviors for relatives of the probands in many of these studies of familial aggregation, making it difficult to distinguish between environmental and genetic influences. In one study in which the tobacco and alcohol behaviors of index cases with OCP cancers, their spouses, and the first-degree relatives were assessed, a fourfold risk of the relatives developing OCP cancers was observed after controlling statistically for these risk behaviors (Foulkes et al., 1996). This suggests an inherited component above the effect of tobacco and alcohol behaviors. In addition, segregation analysis findings in one study rejected the model of a purely environmental cause for oral cancer, a Mendelian model had a better fit (De Andrade et al., 1998), and risks for OCP cancers among heterozygotes for a susceptibility gene were restricted to those who smoked and drank alcohol. Tobacco and alcohol use as well as infrequent consumption of fruits and vegetables interact in a multiplicative fashion with family history of upper aerodigestive tract cancers, which also suggests both environmental and hereditary influences on risk (Foulkes, 1995; De Andrade et al., 1998).

Multiple Primaries and Cancers Associated with Oral and Pharyngeal Cancers Patients with OCP cancer are sometimes diagnosed with several primary cancers in the oral cavity and pharynx at the same time; for example, 7% in one study (Barbone et al., 1996). Second primary cancers subsequent to the first are also relatively common among persons with OCP cancer. Based on follow-up of patients diagnosed with cancer between 1935 and 1982, patients with OCP cancer experienced the highest risk of subsequent new primaries compared with patients with any other form of cancer. After a diagnosis of cancer of the tongue, other mouth, and pharynx, risks ranged from 9–25-fold for the development of new primaries in the tongue and other mouth, 12–16-fold in the esophagus, 3–8-fold in the larynx, and threefold in the lung (Curtis et al., 1985). Patients with cancers of the esophagus, larynx, lung, and cervix— cancer sites associated with cigarette smoking—also were at greater risk of cancer of the tongue, other oral cavity sites, and pharynx than expected. Digestive tract cancers, especially liver/biliary tract cancer, are also somewhat higher among patients with tongue and other mouth cancer (Curtis et al., 1985). Alcohol drinking is a known risk factor for liver cancer (IARC Working Group on the Evaluation of the Carcinogenic Risk of Chemicals to Humans et al., 1988). Persons with squamous cell carcinomas of the skin may also have an elevated risk of several OCP cancer sites. In examining the determinants of second primary aerodigestive tract cancers, several studies have observed an association between smoking prior to initial OCP cancer diagnosis and risk of second primary OCP cancers (Barbone et al., 1996; Cianfriglia et al., 1999). In one study, quitting smoking at or after diagnosis did not confer protection, but quitting before initial diagnosis did (Day et al., 1994b). Risk of second cancers was 40%–60% lower among those in the highest versus lowest quartile of prediagnostic intake of total vegetables and most vegetable subgroups in one study (Day et al., 1994c)

684

PART IV: CANCER BY TISSUE OF ORIGIN

and 60% lower among those in the highest beta-carotene tertile in another (Barbone et al., 1996).

Genetic Susceptibility Genes that may influence the occurrence of OCP cancer include those that affect metabolism of carcinogens and nutrients, DNA repair, and cell cycle control. Many of the studies are of patients with “head and neck” cancer, a grouping that includes patients with laryngeal and occasionally esophageal in addition to OCP cancer. It was possible in some studies of metabolic genes to distinguish OCP cancer findings from esophageal and laryngeal findings, and these are the studies cited in the discussion of metabolic genes. Nearly all of the studies of DNA repair and cell cycle control genes present findings only for patients with head and neck cancer—oral, pharyngeal, and laryngeal cancer combined. All studies have a case-control design with hospital-based or population-based study participants. Studies described below include those with at least 150 study participants. Nearly all of the studies included fewer than 300 cases.

Glutathione S-transferases (GST). This family of genes includes GSTM1 and GSTT. Among other functions, this class of genes facilitates detoxification of carcinogens such as benzo(a)pyrene, a carcinogen found in tobacco smoke. Deletions in these genes lead to reduced enzyme activity (Geisler and Olshan, 2001). GSTM1 deletion genotypes are common in a wide range of populations (Geisler and Olshan, 2001). A review article reported on 11 studies worldwide that examined GSTM1 and risk of OCP cancers (Geisler and Olshan, 2001). Based on this review and subsequent studies in Asian populations, five of eight studies of Asian populations showed higher risks of OCP cancers associated with having GSTM1 deletion genotypes (Nomura et al., 2000; Sreelekha et al., 2001; Geisler and Olshan, 2001; Buch et al., 2002), while none of four studies of Europeans and one of two US studies observed similar associations (Geisler and Olshan, 2001). A dose-response relationship was observed of increasing risk of oral cancer associated with the deletion genotype with increasing lifetime dose of cigarette smoking in one study (Sato et al., 2000) and chewing Indian tobacco in another (Buch et al., 2002). In studies of the GSTT gene, risks in three (Geisler and Olshan, 2001; Sreelekha et al., 2001) of four (Geisler and Olshan, 2001; Sreelekha et al., 2001; Buch et al., 2002) studies of Asians were elevated for the GSTT deletion genotype, and the odds ratios for the four European studies ranged from 1–2 (Geisler and Olshan, 2001). Cytochrome P450 Pathways (CYP). CYP1A1 influences activation of polycyclic aromatic hydrocarbons as well as aromatic amines, which are carcinogens present in tobacco smoke and other substances. The m2 vs. wildtype or m1 genotype has been associated with an increased risk of OCP cancers with odds ratios between 1.5 and 5.7 (Katoh et al., 1999; Morita et al., 1999; Bartsch et al., 2000; Sato et al., 2000; Sreelekha et al., 2001), while m1 findings were inconsistent (Matthias et al., 1998; Tanimoto et al., 1999). CYP2E1 findings have been inconsistent (Katoh et al., 1999; Bartsch et al., 2000) in terms of whether the c1 or c2 genotypes confer risk and associations were observed only in subpopulations that differed across studies, for example, among non-users of betel quid (Hung et al., 1997), those with a lower number of cigarette packyears (Liu et al., 2001), or heavy drinkers (Bouchardy et al., 2000). Limited and conflicting data are available on CYP2D6 (Bartsch et al., 2000). N-Acetyltransferases (NAT). NAT1 and NAT2 are involved in the detoxification of aromatic amines via acetylation. Aromatic amines are present in tobacco smoke and some occupational settings (Lazarus and Park, 2000). Case-control studies have found no association of NAT2 and oral (Katoh et al., 1998; Chen et al., 2001), pharyngeal (Morita et al., 1999), or oral, pharynx, and larynx cancer risk combined (Olshan et al., 2000), although increased oral cancer risk associated with having the rapid/intermediate genotypes and increasing numbers of alcoholic drinks consumed per week was found in one study (Chen et al., 2001).

Microsomal Epoxide Hydrolase (mEH). The microsomal epoxide hydrolase (mEH) gene is involved in metabolism of carcinogens such as benzo(a)pyrene, and a variant has been linked to oral cancer in one study (Jourenkova-Mironova et al., 2000). Alcohol Dehydrogenase (ADH2, ADH3). The alcohol dehydrogenase 3 gene (ADH3) is involved in the metabolism of alcohol to acetaldehyde, and the alcohol dehydrogenase 2 gene (ADH2) metabolizes acetaldehyde to acetic acid (Olshan et al., 2001). Acetaldehyde, the intermediate compound, is carcinogenic in animals (IARC, 1999). It also has been suggested that the ADH3 2–2 genotype may reduce the conversion of retinol to retinoic acid (Schwartz et al., 2001a). In US and European studies, the evidence for a role of ADH3 in OCP cancers is contradictory, with studies showing that having fastmetabolizing alleles places persons at greater risk (Harty et al., 1997; Bouchardy et al., 1998), lesser risk (Zavras et al., 2002), and neither greater nor lesser risk of OCP cancer (Bouchardy et al., 2000; Schwartz et al., 2001a; Sturgis et al., 2001). Effect modification by alcohol consumption has been noted in a few studies, for example higher risks in alcohol drinkers homozygous for the fast-metabolizing genotype (Harty et al., 1997). In a Japanese population, a mutant gene that inactivates ADH2 is common; the ADH2 inactive genotype was associated with a threefold increased risk of oral cancer (Nomura et al., 2000). DNA Repair. Case-control studies have compared head and neck cancer cases with controls with respect to indicators of DNA damage and repair and genes involved in these processes (Berwick and Vineis, 2000a; Goode et al., 2002). Bleomycin-induced sensitivity in lymphocytes, also known as mutagen sensitivity, is one measure of DNA damage to cells and thus an indicator of failure of DNA repair mechanisms. In studies that have controlled for tobacco, or tobacco and alcohol (Cloos et al., 1996a; Cloos et al., 1996b; Spitz et al., 1989; Spitz et al., 1993), mutagen sensitivity was associated with risks in the range of 4–10 (Berwick and Vineis, 2000b). Benzo(a)pyrene diol epoxide (BPDE) sensitivity, another indicator of DNA damage and repair, also has been linked to risk of oral premalignant lesions (Wu et al., 2002) and head and neck cancers (Cheng et al., 1998; Wang et al., 1998; Wu et al., 1998; Berwick and Vineis, 2000a; Li et al., 2001). In one study patients with multiple primary head and neck cancers had greater mutagen sensitivity compared with head and neck cancer patients with only one tumor (Schantz and Ostroff, 1997). Evidence for a multiplicative interaction between mutagen sensitivity and tobacco use has been described (Spitz et al., 1993). Another study reported odd ratios of 6 for bleomycin sensitivity, 14 for BPDE sensitivity, and 36 for both sensitivities (Wu et al., 1998). DNA repair genes investigated are those involving base excision repair, nucleotide excision repair, and mismatch repair (Cheng et al., 2002; Goode et al., 2002). The base excision repair gene hOGG1 is involved in the excision repair of 8-hydroxy-2¢-deoxyguanine from oxidatively damaged DNA (Goode et al., 2002). In one study, cases with oral cavity, tonsil, and oropharyngeal cancer were more likely to have the low activity genotype compared with community-based controls. Another gene in this family is XRCC1, which facilitates endonuclease activity and the creation of a scaffold for reconstruction of the damaged site (Goode et al., 2002). Positive (Hsieh et al., 2003; Sturgis et al., 1999) as well as negative (Olshan et al., 2002) associations with the XRCC1 variants have been observed. Variants in the nucleotide repair genes XPC (Shen et al., 2001), which is involved in recognizing DNA damage, and XPD, which is involved in unwinding of DNA after damaged DNA is recognized, were not associated with head and neck cancer risk (Sturgis et al., 2000; Sturgis et al., 2002). However, expression of a number of nucleotide excision repair genes including these two was reduced in a study of head and neck cancer cases relative to controls (Cheng et al., 2002). A study that controlled for smoking and alcohol (Wei et al., 1998) found that oral cancer cases were more likely to have low expression

Cancers of the Oral Cavity and Pharynx of hMLH1, which is involved in mismatch repair, compared with controls, but there was no difference for hMLH2, a related gene. Interactions have been noted. For example, no association between the hOGG1 326cys/326cys vs. 326ser hetero- or homozygote genotypes and OCP cancer risk was observed among nonsmokers or non-alcohol drinkers. However, these genotypes were associated with significantly increased risk among smokers and among drinkers (Elahi et al., 2002).

Cell Cycle Control. No association with risk of head and neck cancer was observed for two p16 tumor suppressor haplotypes (Zheng et al., 2001b). However, CPG island hypermethylation within the gene promoter region of the p16 and other genes resulting in transcriptional inactivation was common in head and neck cancer cases compared with controls (Rosas et al., 2001). Associations were observed in a study of checkpoint kinase 2 (Zheng et al., 2001a), another tumor suppressor gene, and cyclin D1 (Zheng et al., 2001b), which controls how cells with damage pass through a checkpoint in the cell cycle. Pathogenesis Clinical Applications of Molecular Markers for Oral and Pharyngeal Cancers Molecular analyses of OCP cancers are not currently incorporated into routine clinical practice. However, research studies indicate that molecular analysis of OCP cancer tumor specimens and oral epithelial cells, saliva, and blood has considerable potential clinical utility; for example, in early detection of malignancies, differentiation of OCP recurrences from second primary tumors, prognostic evaluation, and treatment planning for invasive malignancies and premalignancies (Hu et al., 2002).

Progression Model Evidence suggests that a series of losses of genetic material at specific chromosomal sites leads to the development of the progressive histologic changes that culminate in invasive OCP cancers (Sidransky, 2002). This model has been developed by comparing the frequency of each type of loss in normal tissue and lesions from patients with upper aerodigestive tract lesions of varying degrees of severity. Califano and colleagues suggested that chromosomal loss at 9p is associated with the development of benign squamous hyperplasia and some other precursor lesions, 3p and 17p with dysplasia, 11q, 13q, 14q with carcinoma in situ, and 6p, 8p and 8q, and 4q with invasive cancer (Califano et al., 1996). Mitochondrial C-tract alterations also increased with histological severity of lesions (Ha et al., 2002). These chromosomal losses correspond to aberrations in specific genes. For example, the earliest and most common aberration may be inactivation of the tumor suppressor gene p16 on 9p21 (Reed et al., 1996). The 8p21–22 losses may be associated with the gene for a tumor necrosis-related apoptosis-inducing ligand receptor, DR4 (Fisher et al., 2001). These aberrations can result from several different processes including hypermethylation of the promoting regions of tumor suppressor genes, deletions, and point mutations (Reed et al., 1996; Rosas et al., 2001).

Field Cancerization Second or multiple primary cancers occur relatively frequently in patients with OCP cancers. This may be the result of either of two processes, the development of independent, transformed cells or the spread of cells derived from a single transformed cell to other areas of the mucosa. Based on molecular analysis of tumor tissue from multiple sites within the same individuals, it appears that both phenomenona occur (Bedi et al., 1996; Califano et al., 1999; Partridge et al., 2001).

Preventive Measures Primary Prevention Most OCP cancers develop in persons with chronic exposures to tobacco, alone or in concert with exposure to alcohol, so most of these cancers should be preventable. Avoidance of tobacco initiation, along with promotion of tobacco cessation, is likely to have the greatest

685

impact in preventing these cancers. Evidence suggests that even relatively brief periods of tobacco cessation (e.g., 6–10 years) may significantly lower risk, although in some studies the full benefit was only realized after more prolonged cessation. Current declines in rates of OCP cancers in white men in the United States are consistent with the decline in smoking prevalence noted in this group. While tobacco prevention/control should be the cornerstone of prevention efforts for these cancers, reducing excessive exposures to alcohol is also an important aspect of primary prevention. Research studies do not indicate that low levels of alcohol consumption increase risk of OCP cancers, but heavier drinking clearly does and should be avoided. As is the case with tobacco, cessation of alcohol lowers risk (Franceschi et al., 2000) but estimates of the number of years required vary. Studies suggest that dietary practices also contribute to the etiology of OCP cancers. Campaigns to increase fruit and vegetable consumption, such as the US Five-A-Day For Better Health Program (Stables et al., 2002), may also contribute to reducing the incidence of these cancers. While the evidence for an association between dietary practices and risk is not considered definitive, this dietary pattern is prudent and likely to reduce the risk of a variety of cancers and other chronic diseases. Specific nutrients also are of interest for prevention; selected nutrient supplements have been evaluated in randomized clinical trials for efficacy as summarized below.

Chemoprevention: Oral Premalignant Lesions As discussed previously, oral precancerous lesions are indicators of field cancerization and some proportion may progress into malignancies. Chemoprevention of these lesions has direct clinical relevance and can be used to screen agents that may have efficacy in the prevention of OCP cancers. As reviewed elsewhere (Mayne et al., 2003), at least 13 randomized chemoprevention trials in oral leukoplakia have been conducted using the following agents: retinoids, vitamin A, betacarotene, the algae Spirulina fusiformis, tea, and bleomycin. Other agents (e.g, vitamin E, selenium, protease inhibitors (Bowman Birk Inhibitor Concentrate)) have been evaluated in non-randomized trials. Retinoids and beta-carotene have established efficacy in the regression of oral precancerous lesions (Mayne et al., 2003). The most widely studied retinoid (13-cis-retinoic acid) and carotenoid (beta-carotene), however, have been shown to interact with tobacco smoke to increase the risk of lung cancer (recurrences for 13-cis-retinoic acid (Lippman et al., 2001) and primary lung cancers for beta-carotene (AlphaTocopherol Beta-Carotene Cancer Prevention Study Group, 1994; Omenn et al., 1996)). Given this, long-term preventive therapy of these agents for oral precancerous lesions, at least in smokers, is unlikely. Some suggestions of efficacy have also been obtained in randomized trials of Spirulina fusiformis (a nutrient-rich algae (Mathew et al., 1995)), green tea (Li et al., 1999), and bleomycin (Epstein et al., 1994), but none is considered to have established chemopreventive efficacy in oral premalignant lesions.

Secondary Prevention Many patients diagnosed with OCP cancers will survive their primary cancer, particularly those with early-stage cancers (Table 35–2). These patients, however, are at substantial risk of failure due to local recurrences and second primary cancers; recent data indicate that 5% of Stage I/II patients will develop second primary tumors yearly (Khuri et al., 2001). These second cancers are a consequence of field cancerization, and are the leading cause of death in patients diagnosed with early-stage OCP cancers (Lippman and Hong, 1989). Retinoids, vitamin A, and beta-carotene have all been evaluated for secondary prevention in this setting. The first trial tested high-dose 13-cisretinoic acid vs. placebo in 103 OCP cancer patients (Hong et al., 1990). The rate of second primary tumors was significantly lower in the retinoid arm than in the placebo group, developing in two (4%) of the 13 cRA-treated patients compared with 12 (24%) of the placebotreated patients (P = .005). Side effects, however, were substantial. A multicenter US trial of a lower dose of 13-cis-retinoic acid (Khuri et al., 2003) saw no overall benefit, although there was a suggestion that local recurrences were reduced while patients were receiving drugs.

686

PART IV: CANCER BY TISSUE OF ORIGIN

Another randomized trial evaluated the synthetic retinoid etretinate in 316 patients following definitive therapy of OCP squamous cell carcinoma (Bolla et al., 1994). In this French trial, the rate of second primary tumor development in the two arms was not significantly different. The Euroscan trial evaluated 2 years of retinyl palmitate and N-acetylcysteine (factorial design) in preventing second primary tumors in 2592 patients with prior cancers of the oral cavity, larynx, and lung (van Zandwijk et al., 2000). Retinyl palmitate, N-acetylcysteine, or both produced no improvement in event-free survival, survival, or incidence of second primary tumors. Another trial evaluated beta-carotene vs. placebo in 264 patients with prior oral/pharyngeal/ laryngeal cancers (Mayne et al., 2001). Patients randomized to supplemental beta-carotene had a nonsignificant reduction in second OCP/larynx cancer (RR 0.69) and a nonsignificant increase in lung cancer (RR 1.44). The results of these trials thus indicate that agents with efficacy in oral premalignancy (e.g., 13-cis-retinoic acid, betacarotene) appear to have some efficacy in the prevention of invasive disease, but emphasize the need to identify agents with minimal toxicity and without adverse effects on other sites (e.g., lung). Fruit/vegetable intervention trials have been considered as another approach for secondary prevention of OCP/larynx cancers. Completed trials to date show that this patient population can successfully increase fruit/vegetable intake (Le Marchand et al., 1994), although data are not available to evaluate the efficacy of such a change with respect to cancer incidence.

Screening/Early Detection Screening for OCP cancer in asymptomatic persons generally involves inspection and palpation of the oral cavity. However, the sensitivity and specificity of screening for oral cancer in the United States has not been characterized, and controlled trials of screening also have not been done. Despite this, the American Cancer Society recommends that primary care doctors examine the mouth and throat as part of a routine cancer-related checkup. The US Preventive Services Task Force concluded that there is “insufficient evidence to recommend for or against routine screening of asymptomatic persons for oral cancer by primary care clinicians. All patients should be counseled to discontinue use of all forms of tobacco, and limit consumption of alcohol. Clinicians should remain alert to signs and symptoms of oral cancer and premalignancy in persons who use tobacco or regularly use alcohol” (US Preventive Services Task Force, 1996).

LIP CANCERS The ICD-O-3 topography code for the lip, C00, includes malignant neoplasms arising on the vermilion border, commissure, and labial mucosa, but excludes cancers originating on the skin of the lip (Fritz et al., 2000). The lower lip is most frequently affected, and the majority of lip cancers are squamous cell carcinomas. During the period 1992–1999 and based upon data from 11 SEER sites in the United States, 79% of all reported invasive lip cancers arose on the vermilion border, 80% were located exclusively on the lower lip, and 95% were squamous cell carcinomas (SEER, 2002). In the United States during 1996–2000, the age-adjusted incidence rate was 1 in 100,000 per year (Table 35–1) (Ries et al., 2003). Five-year relative survival for cancer of the lip is high and the mortality rate low (Table 35–2). Based upon SEER data from nine areas of the United States for the period 1992–1999, the 5-year relative survival for lip cancer was 95% for males and 90% for females. During 1996–2000, the age-adjusted (2000 US) mortality rate for cancer of the lip in the United States was less than 0.1 in 100,000 (0.05/100,000 for males and 0.01/100,000 for females) (Ries et al., 2003; SEER, 2003a). There is geographic variation in lip cancer incidence, with rates generally higher among men than women and in light-skinned populations relative to those of color (Fig. 35–6) (Dardanoni et al., 1984; Pogoda and Preston-Martin, 1996). For the period approximating 1993–1997 and for most regions of the world, the annual age-adjusted (world) incidence of lip cancer was generally less than 4 in 100,000 for males

and below 1 in 100,000 for females (Parkin et al., 2002). During the same time period, however, annual age-adjusted rates (world) approached or exceeded 10 in 100,000 for males living in regions of Canada (Newfoundland), Australia (Queensland, Tasmania), and Spain (Cuenca, Granada, Albacete) (Parkin et al., 2002). For females, the highest age-adjusted rates (2 to 3/100,000) were reported in Canada (Yukon), Australia (Tasmania, Queensland, South, Western), and Thailand (Khon Kaen) (Parkin et al., 2002). Over recent decades, the incidence of lip cancer has declined in many geographic regions of the world, particularly among males. For example, based upon SEER data for nine geographic regions in the United States, age-adjusted incidences rates declined over 70% for males and 40% for females during the period 1973–2000 (SEER, 2003b). Lip cancer is associated with low socioeconomic status (Dardanoni et al., 1984; Pukkala et al., 1994), rural residence (Lindqvist and Teppo, 1978; Doll, 1991; Schouten et al., 1996), and outdoor occupation, particularly fishing and farming (Keller, 1970; Spitzer et al., 1975; Khuder, 1999; Hakansson et al., 2001). Evidence also suggests that both solar radiation and tobacco are important risk factors for lip cancer. Solar radiation has long been considered a probable risk factor for lip cancer (Ahlbom, 1937; Ebenius, 1943). The fact that both outdoor occupation and rural residence are associated with an increased risk of lip cancer is consistent with a solar component although it does not preclude the possibility that other outdoor elements or associated factors are involved. The observation that lip cancer most often occurs on the lower lip has led some to suggest that the disparity in anatomical risk is a consequence of the differential exposure of the lower lip to direct sunlight (Ju, 1973). The finding that lip cancer is more frequent in light-skinned individuals and populations has led to the proposition that higher levels of melanin within the lip protect against the effects of solar radiation (Ebenius, 1943; Bernier and Clark, 1951), and that the use of lipstick by women may partially explain the generally lower incidence in women (Pogoda and Preston-Martin, 1996). Studies regarding the geographic distribution of lip cancer have not all been consistent with a primary etiologic role for ultraviolet (UV) light (Keller, 1970; Szpak et al., 1977; Lindqvist and Teppo, 1978; de Visscher et al., 1998); however, a case-control study of lip cancer in Los Angeles females found that lip cancer risk tended to increase with both average annual residential UV flux and average hours of outdoor activities (Pogoda and Preston-Martin, 1996). Studies have also identified strong associations between lip cancer and a previous history of skin cancer (Keller, 1970; Pogoda and Preston-Martin, 1996; Wassberg et al., 1999), although the decline in lip cancer incidence stands in contrast to increasing incidence rates for cutaneous malignant melanoma (SEER, 2003b). Another link in the relationship between sunlight exposure and lip cancer risk is chronic actinic cheilitis, a precancerous condition of the lip that presents as a hyperkeratosis interspersed with areas of erythema on the vermilion border. Actinic cheilitis is generally attributed to solar damage and may harbor epithelial dysplasia, carcinoma in situ, or squamous cell carcinoma (Nicolau and Balus, 1964; Kaugars et al., 1999). As with lip cancer, actinic cheilitis predominates on the lower lip and is seen most often in males, light-skinned persons, and outdoor workers. Evidence for a relationship between smoking tobacco and lip cancer risk is based largely upon findings from case-control studies. Most investigations conducted in the early through mid 20th century implicated primarily pipe smoking (Broders, 1920; Ahlbom, 1937; Ebenius, 1943; Levin et al., 1950; Spitzer et al., 1975). However, pipe smoking has declined dramatically in the United States (Nelson et al., 1996; Psoter and Morse, 2001), and studies during the latter decades of the 20th century identified cigarette smoking and/or smoking in general as risk factors (Keller, 1970; Wigle et al., 1980; Pogoda and PrestonMartin, 1996). Further evidence of a link between smoking and lip cancer is provided by the observation that lip cancer cases are at an elevated risk of second primary cancers known to be associated with smoking (lung, larynx) and vice versa (Curtis et al., 1985; Winn and Blot, 1985; Soderholm et al., 1994).

Cancers of the Oral Cavity and Pharynx Males

687

Females AMERICA Canada, Newfoundland US, SEER, White Colombia, Cali US, SEER, Black EUROPE Spain, Cuenca Italy, Sassari Slovakia Lithuania Finland Denmark Ireland Slovenia The Netherlands Sweden Poland, Warsaw City Czech Republic France, Isere England ASIA Israel, Jews India, Mumbai (Bombay) Thailand, Kohn Kaen Singapore, Chinese OCEANIA Australia, Queensland Hawalii, White New Zealand

16

14

12

10

8

6

4

2

0

2

4

Rate per 100,000 person-years

There is little consistent evidence that alcohol consumption (Wynder et al., 1957; Keller, 1970; Spitzer et al., 1975; Pogoda and Preston-Martin, 1996), syphilis (Wynder et al., 1957; Keller, 1970), or herpetic lesions (Spitzer et al., 1975; Lindqvist, 1979; Dardanoni et al., 1984; Pogoda and Preston-Martin, 1996) are important etiologic factors for lip cancer.

SALIVARY GLAND CANCER Salivary gland cancer (SGC) can arise in either the major or minor salivary glands. Current topography codes (ICD-O-3, ICD-10) for malignant neoplasms of the major salivary glands (C07-08) include the parotid, submandibular, and sublingual glands as well as their associated ducts, while cancers of the minor salivary glands are classified separately according to their anatomical site. Based largely upon these coding practices, population-based reports on the incidence of salivary gland cancer are generally restricted to malignant neoplasms of the major salivary glands of which the parotid is the most frequently affected. In one report providing population-based incidence data for both major and minor SGCs, the major glands accounted for 77% of all reported incident cases in Sweden during 1960–1989 while the minor glands accounted for the remaining 23% (Ostman et al., 1997). The histopathologic classification of SGCs is complex and has undergone change over time. In the United States (11 SEER areas) during the period 1992–1999, the majority of SGCs were classified histopathologically as either adenocarcinomas (26%) or mucoepider-

Figure 35–6. Age-adjusted (world) incidence rates for cancer of the lip, selected geographic regions, circa 1993–1997, all ages.

moid carcinomas (25%) followed by squamous cell (18%) and acinar cell (12%) carcinomas (SEER, 2002). Salivary gland cancers are rare, with reported annual incidence rates of 1.2 in 100,000 in the United States (12 SEER sites, 1996–2000; Table 35–1) (SEER, 2003c) and rates generally well below 2 in 100,000 in most geographic regions during the period 1993–1997 (Fig. 35–7). While there is geographic variation in incidence rates, and while rates are most often higher in males, the absolute differences are generally small. Reported incidence rates are highest among persons living in the Canadian Northwest Territories and among Circumpolar Inuits for whom rates approach or exceed 4 in 100,000 for both males and females (Lanier and Alberts, 1996; Parkin et al., 2002). During the period approximating 1968–1972 to 1993–1997 and among those regions included in Cancer Incidence in Five Continents, volumes III through VIII (Waterhouse et al., 1976; Waterhouse et al., 1982; Muir et al., 1987; Parkin et al., 1992; Parkin et al., 1997; Parkin et al., 2002), age-adjusted incidence rates for SGCs in males showed a net decline of at least 50% in regions of Germany (Saarland) and the United Kingdom (West Midlands, Oxford), but increased over 50% in Mumbai (India), Zaragoza (Spain), Puerto Rico, and among Singaporean Chinese. Among females during the same period, incidence rates declined 50% or more in areas of Canada (Manitoba, Quebec, Saskatchewan), Germany (Saarland), and the United Kingdom (West Midlands, Oxford, South and Western Region), but rose by 50% or more in Mumbai (India), Miyagi (Japan), and Utah (US). Some portion of the net change in reported rates, however, may reflect modifications in ICD coding practices or the histopathologic classification of these tumors over time. In the United States (9 SEER registries)

688

PART IV: CANCER BY TISSUE OF ORIGIN Males

Females AMERICA US, SEER, White US, SEER, Black Canada Colombia, Cali EUROPE Croatia France, Haut-Rhin Austria, Tryol Italy, North East Czech Republic Estonia Slovenia Lithuania Ireland Poland, Lower Silesia Finland Denmark Germany, Saarland Sweden The Netherlands England ASIA Philippines, Manila Pakistan, South Karachi Hong Kong Israel, Jews India, Mumbai (Bombay) Japan, Osaka China, Tianjin OCEANIA Australia, Queensland New Zealand

2

1

0

1

2

Rate per 100,000 person-years

Figure 35–7. Age-adjusted (world) incidence rates for cancer of the major salivary glands, selected geographic regions, circa 1993–1997, all ages.

during the years 1973 through 2000, SGC incidence rates rose with an average percentage change of 0.9% for males and 0.2% for females (SEER, 2003b). In the United States, 5-year relative survival for SGC is moderately high and mortality relatively low. During 1992–1999 (9 SEER areas), the 5-year relative survival for SGC was 70% for males and 80% for females (Table 35–2) while mortality (1996–2000) was 0.4 and 0.2 in 100,000 for males and females, respectively (Ries et al., 2003; SEER, 2003a). Between 1969 and 2000, age-adjusted mortality rates for SGC in the United States declined with a statistically significant annual percent change of -1.3% for males and -1.8% for females (SEER, 2003a). Radiation is an established risk factor for SGCs, with elevated risks and dose-response relationships observed among atomic bomb survivors (Land et al., 1996; Saku et al., 1997) and persons exposed to prior therapeutic head or neck irradiation (Spitz et al., 1984; PrestonMartin et al., 1988; Preston-Martin, 1989; Spitz et al., 1990; HornRoss et al., 1997a; Modan et al., 1998; Rubino et al., 2003). SGC risk has also been associated with diagnostic radiation directed to the head or neck, most notably among persons exposed to frequent full-mouth dental X-rays, and particularly for those exposed prior to the 1960s when substantially higher doses were used (Preston-Martin et al., 1988; Preston-Martin and White, 1990; Zheng et al., 1996; Horn-Ross et al., 1997a). In one study, ultraviolet light treatments to the head and neck, used primarily to treat acne, were linked to an elevated risk of SGC, most notably among whites and particularly for exposures prior to 1955 (Horn-Ross et al., 1997a). SGC risk has also been linked to a previous history of UV-related, nonmelanoma skin cancer (Spitz et al., 1984; Teppo et al., 1985; Frisch and Melbye, 1995; Wassberg et al., 1999; Milan et al., 2000). However, studies have not consistently observed a clear relationship between average UVB intensity by geographic region and age-adjusted SGC incidence among whites living in the United States (Spitz et al., 1988; Sun et al., 1999), and mortality rates in the United States show little north-south gradient (Devesa et al., 1999).

There is evidence to link some occupational groups to increased risks of SGC. Elevated risks have been identified among women working in beauty shops (Swanson and Burns, 1997), male woodworkers employed in automobile plants (Swanson and Belle, 1982), rubber industry workers (Horn-Ross et al., 1997a; Mancuso and Brennan, 1970), persons occupationally exposed to nickel compounds or alloys (Horn-Ross et al., 1997a), radioactive materials (Horn-Ross et al., 1997a), or silica dust (Zheng et al., 1996), as well as among employees of an Australian underground colliery (Corbett and O’Neill, 1988) and residents of asbestos-mining counties in Quebec (Graham et al., 1977). While alcohol and tobacco use are clearly related to OCP cancer, their relationship with SGC is equivocal. Most case-control investigations have reported little or no consistent evidence in support of a tobacco-SGC association (Keller, 1969; Spitz et al., 1984; PrestonMartin et al., 1988; Spitz et al., 1990; Zheng et al., 1996; Swanson and Burns, 1997; Muscat and Wynder, 1998); however, cigarette smoking was a strong risk factor in a case-control study conducted in Puerto Rico (Hayes et al., 1999), and current smoking was associated with a twofold increase in SGC, and a nearly sevenfold increase in adenocarcinoma risk among men, but not women, in a study carried out in northern California (Horn-Ross et al., 1997a). Also consistent with a tobacco-SGC relationship are reports that find an excess risk of lung cancer among previous cases of SGC and vice versa (Boice and Fraumeni, 1985; Winn and Blot, 1985; Sun et al., 1999). With regard to alcohol consumption, one case-control study found that drinking doubled SGC risk among women, but not men (Spitz et al., 1990), while another investigation reported an OR of 2.5 for heavy drinking and an associated dose-response trend for SGC among men, but not women (Horn-Ross et al., 1997a). Other studies, however, provide little additional support for a link between alcohol consumption and SGC (Keller, 1969; Spitz et al., 1984; Zheng et al., 1996; Muscat and Wynder, 1998; Hayes et al., 1999). Diet has received little attention in relation to SGC risk; however, in a Chinese study, the consumption of dark yellow vegetables and liver was inversely related to risk (Zheng et al., 1996), while a US study found protective effects associated with a high, relative to low, intake of fiber from bean sources and total vitamin C, but an increased risk with high cholesterol consumption (Horn-Ross et al., 1997b). In general, viruses have also received relatively little consideration in relation to SGC risk although a number of case reports and series suggest a strong link between Epstein-Barr virus (EBV) and lymphoepithelial carcinomas of the salivary glands, cancers observed primarily in Eskimos and the southern Chinese (Hamilton-Dutoit et al., 1991; Lanier et al., 1991; Chan et al., 1994; Sheen et al., 1997). The possible etiologic role of HPV for SGC has been discounted by the observation that HPV-related cancers (anal, cervical) do not occur in excess after SGC (Sun et al., 1999). A possible hormonal link to SGC was supported by early reports of an excess risk of secondary breast cancer after SGC and vice versa (Berg et al., 1968; Prior and Waterhouse, 1977); however, most subsequent studies involving larger cohorts revealed little, if any, increased breast cancer risk (Schou et al., 1985; Winn and Blot, 1985; Sun et al., 1999). Although parity was not associated with SGC in one US-based case-control study (Spitz et al., 1984), age at menarche, number of births, age at first full-term delivery, and length of oral contraceptive use were all inversely related to SGC risk in another (Horn-Ross et al., 1999). Other potential risk factors for SGC have been identified; however, the epidemiological evidence is often limited to one investigation; these factors include the use of hair dye (women only) (Spitz et al., 1990), mouthwash use (Spitz et al., 1990), kerosene used as a cooking fuel in Shanghai (Zheng et al., 1996), and familial clustering of SGC in Greenland (Merrick et al., 1986). There is little evidence of an association between SGC and the use of chewing tobacco or snuff (Preston-Martin et al., 1988; Spitz et al., 1990), cellular telephones (Johansen et al., 2001; Auvinen et al., 2002), or diagnostic X-rays to the chest or limbs (Preston-Martin et al., 1988; Zheng et al., 1996). One study suggests that the pattern of allelic losses from chromosomes in SGC differs from that described for the progression model

Cancers of the Oral Cavity and Pharynx in OCP cancers and that allelic loss patterns for salivary gland subtypes—pleomorphic adenomas, adenoid cystic carcinomas, and mucoepidermoid cancers—may differ from each other (Johns et al., 1996).

FUTURE DIRECTIONS Effective prevention and control of OCP cancers in the future will require research strategies and public health resources to overcome the following key challenges: 1. Increasing rates of smoking in many countries and within segments of the US population require continued attention. Smokeless tobacco use is especially common in Asia (Gupta, 1996; Gupta and Warnakulasuriya, 2002). Also, new tobacco products are being developed (Slade et al., 2002), with unknown long-term effects on the risk of OCP cancers. 2. Despite major improvements in the past decade in identifying smoking and alcohol prevention and cessation interventions that are evidence-based, overall quit rates are still relatively low, and for persons with heavy tobacco and alcohol behaviors, cessation rates are even lower (US DHHS, 2003). 3. Serious disparities among race and ethnic groups in risk persist and require concerted attention. 4. Effective and non-toxic chemopreventive agents for oral precancerous lesions or for secondary prevention of OCP cancers have yet to be identified and verified. 5. Future molecular epidemiologic studies aimed at clarifying the role of genetic susceptibility must recognize and address key limitations in the current literature, including: assessment of the effect of genetic polymorphisms on risk may be affected by population stratification; lack of matching on alcohol and tobacco; small sample sizes; failure to consider HPV status; risk variation within subsites of the OCP; and effect of polymorphisms on survival (Olshan et al., 2000). Future progress in these five areas is critical to reduce the morbidity and mortality associated with these cancers. References Ahlbom HE. 1937. Pradisponierende faktoren fur plattenepithelkarzinom in mund, hals und speiserohre. Eine statistische untersuchung am material des radiumhemmets, Stockholm. Acta Radiol (Stockholm) 18:163–185. Alpha-Tocopherol Beta-Carotene Cancer Prevention Study Group. 1994. The effect of vitamin E and beta carotene on the incidence of lung cancer and other cancers in male smokers. The Alpha-Tocopherol, Beta Carotene Cancer Prevention Study Group [comment]. N Engl J Med 330(15): 1029–1035. Auvinen A, Hietanen M, Luukkonen R, Koskela RS. 2002. Brain tumors and salivary gland cancers among cellular telephone users. Epidemiol 13(3):356–359. Axell T, Pindborg JJ, Smith CJ, van der Waal I. 1996. Oral white lesions with special reference to precancerous and tobacco-related lesions: Conclusions of an international symposium held in Uppsala, Sweden, May 18–21, 1994. International Collaborative Group on Oral White Lesions. J Oral Pathol Med 25(2):49–54. Bagnardi V, Blangiardo M, La Vecchia C, Corrao G. 2001. A meta-analysis of alcohol drinking and cancer risk. Br J Cancer 85(11):1700–1705. Balaram P, Sridhar H, Rajkumar T, et al. 2002. Oral cancer in southern India: The influence of smoking, drinking, paan-chewing and oral hygiene. Int J Cancer 98(3):440–445. Banoczy J, Gintner Z, Dombi C. 2001. Tobacco use and oral leukoplakia. J Dent Educ 65(4):322–327. Barbone F, Franceschi S, Talamini R, et al. 1996. A follow-up study of determinants of second tumor and metastasis among subjects with cancer of the oral cavity, pharynx, and larynx. J Clin Epidemiol 49(3):367–372. Baric JM, Alman JE, Feldman RS, Chauncey HH. 1982. Influence of cigarette, pipe, and cigar smoking, removable partial dentures, and age on oral leukoplakia. Oral Surg Oral Med Oral Pathol 54(4):424–429. Bartsch H, Nair U, Risch A, Rojas M, Wikman H, Alexandrov K. 2000. Genetic polymorphism of CYP genes, alone or in combination, as a risk modifier of tobacco-related cancers. Cancer Epidemiol Biomarkers Prev 9(1): 3–28.

689

Bedi GC, Westra WH, Gabrielson E, Koch W, Sidransky D. 1996. Multiple head and neck tumors: Evidence for a common clonal origin. Cancer Res 56(11):2484–2487. Berg JW, Hutter RV, Foote FW, Jr. 1968. The unique association between salivary gland cancer and breast cancer. J Am Med Assoc 204(9): 771–774. Bernier JL, Clark ML. 1951. Squamous cell carcinoma of the lip: A statistical and morphologic analysis of 835 cases. Milit Surg 109:379–397. Berwick M, Vineis P. 2000a. Markers of DNA repair and susceptibility to cancer in humans: An epidemiologic review. J Natl Cancer Inst 92(11):874–897. Berwick M, Vineis P. 2000b. RESPONSE re: markers of DNA repair and susceptibility to cancer in humans: An epidemiologic review. J Natl Cancer Inst 92(18):1537. Blot WJ, Winn DM, Fraumeni JF, Jr. 1983. Oral cancer and mouthwash. J Natl Cancer Inst 70(2):251–253. Blot WJ, McLaughlin JK, Winn DM, et al. 1988. Smoking and drinking in relation to oral and pharyngeal cancer. Cancer Res 48(11):3282–3287. Boffetta P, Mashberg A, Winkelmann R, Garfinkel L. 1992. Carcinogenic effect of tobacco smoking and alcohol drinking on anatomic sites of the oral cavity and oropharynx. Int J Cancer 52(4):530–533. Boffetta P, Pershagen G, Jockel K-H, et al. 1999. Cigar and pipe smoking and lung cancer risk: A multicenter study from Europe. JNCI Cancer Spectrum 91(8):697–701. Boffetta P, Gridley G, Gustavsson P, et al. 2000. Employment as butcher and cancer risk in a record-linkage study from Sweden. Cancer Causes Control 11(7):627–633. Boffetta P, Gridley G, Lindelof B. 2001a. Cancer risk in a population-based cohort of patients hospitalized for psoriasis in Sweden. J Invest Derm 117(6):1531–1537. Boffetta P, Ye W, Adami HO, Mucci LA, Nyren O. 2001b. Risk of cancers of the lung, head and neck in patients hospitalized for alcoholism in Sweden. Br J Cancer 85(5):678–682. Boice JD, Jr., Fraumeni JF, Jr. 1985. Second cancer following cancer of the respiratory system in Connecticut, 1935–1982. Natl Cancer Inst Monogr 68:83–98. Bolla M, Lefur R, Ton Van J, et al. 1994. Prevention of second primary tumours with etretinate in squamous cell carcinoma of the oral cavity and oropharynx. Results of a multicentric double-blind randomised study [comment]. Eur J Cancer 30A(6):767–772. Bondy ML, Spitz MR, Halabi S, et al. 1993. Association between family history of cancer and mutagen sensitivity in upper aerodigestive tract cancer patients. Cancer Epidemiol Biomarkers Prev 2(2):103–106. Bosetti C, Negri E, Franceschi S, et al. 2000. Risk factors for oral and pharyngeal cancer in women: A study from Italy and Switzerland. Br J Cancer 82(1):204–207. Bouchardy C, Coutelle C, Ward PJ, Dayer P, Benhamou S. 1998. Re: Alcohol dehydrogenase 3 genotype and risk of oral cavity and pharyngeal cancers. J Natl Cancer Inst 90(12):937–939. Bouchardy C, Hirvonen A, Coutelle C, Ward PJ, Dayer P, Benhamou S. 2000. Role of alcohol dehydrogenase 3 and cytochrome P-4502E1 genotypes in susceptibility to cancers of the upper aerodigestive tract. Int J Cancer 87(5):734–740. Broders AC. 1920. Squamous-cell epithelioma of the lip: A study of five hundred and thirty-seven cases. J Am Med Assoc 74(10):656–664. Brown LM, Gridley G, Diehl SR, et al. 2001. Family cancer history and susceptibility to oral carcinoma in Puerto Rico. Cancer 92(8):2102– 2108. Buch SC, Notani PN, Bhisey RA. 2002. Polymorphism at GSTM1, GSTM3 and GSTT1 gene loci and susceptibility to oral cancer in an Indian population. Carcinogenesis 23(5):803–807. Bundgaard T, Wildt J, Frydenberg M, Elbrond O, Nielsen JE. 1995. Casecontrol study of squamous cell cancer of the oral cavity in Denmark. Cancer Causes Control 6(1):57–67. Califano J, van der Riet P, Westra W, et al. 1996. Genetic progression model for head and neck cancer: Implications for field cancerization. Cancer Res 56(11):2488–2492. Califano J, Leong PL, Koch WM, Eisenberger CF, Sidransky D, Westra WH. 1999. Second esophageal tumors in patients with head and neck squamous cell carcinoma: An assessment of clonal relationships. Clin Cancer Res 5(7):1862–1867. Canto MT, Devesa SS. 2002. Oral cavity and pharynx cancer incidence rates in the United States, 1975–1998. Oral Oncol 38(6):610–617. Castellsague X, Munoz N. 1999. Re: Cancer of the oral cavity and pharynx in nonsmokers who drink alcohol and in nondrinkers who smoke tobacco. J Natl Cancer Inst 91(15):1336–1338. Chainani-Wu. 2002. Diet and oral, pharyngeal, and esophageal cancer. Nutr Cancer 44(2):104–126.

690

PART IV: CANCER BY TISSUE OF ORIGIN

Chan JK, Yip TT, Tsang WY, Poon YF, Wong CS, Ma VW. 1994. Specific association of Epstein-Barr virus with lymphoepithelial carcinoma among tumors and tumorlike lesions of the salivary gland. Arch Pathol Lab Med 118(10):994–997. Chen C, Ricks S, Doody DR, Fitzgibbons ED, Porter PL, Schwartz SM. 2001. N-Acetyltransferase 2 polymorphisms, cigarette smoking and alcohol consumption, and oral squamous cell cancer risk. Carcinogenesis 22(12):1993–1999. Cheng L, Eicher SA, Guo Z, Hong WK, Spitz MR, Wei Q. 1998. Reduced DNA repair capacity in head and neck cancer patients. Cancer Epidemiol Biomarkers Prev 7(6):465–468. Cheng L, Sturgis EM, Eicher SA, Spitz MR, Wei Q. 2002. Expression of nucleotide excision repair genes and the risk for squamous cell carcinoma of the head and neck. Cancer 94(2):393–397. Cheng LS, Carmelli D, Swan GE, Pfefferbaum A. 2000a. Segregation analysis of drinking problem in elderly men and their first-degree relatives from the Western Collaborative Group Study. Ann Epidemiol 10(5):309–315. Cheng LS, Swan GE, Carmelli D. 2000b. A genetic analysis of smoking behavior in family members of older adult males. Addiction 95(3):427–435. Cianfriglia F, Di Gregorio DA, Manieri A. 1999. Multiple primary tumours in patients with oral squamous cell carcinoma. Oral Oncol 35(2):157–163. Cloos J, Reid CB, Snow GB, Braakhuis BJ. 1996a. Mutagen sensitivity: Enhanced risk assessment of squamous cell carcinoma. Eur J Cancer B Oral Oncol 32B(6):367–372. Cloos J, Spitz MR, Schantz SP, et al. 1996b. Genetic susceptibility to head and neck squamous cell carcinoma. J Natl Cancer Inst 88(8):530–535. Coble JB, Brown LM, Hayes RB, et al. 2003. Sugarcane farming, occupational solvent exposures, and the risk of oral cancer in Puerto Rico. J Occup Environ Med 45(8):869–874. Copper MP, Jovanovic A, Nauta JJ, et al. 1995. Role of genetic factors in the etiology of squamous cell carcinoma of the head and neck. Arch Otolaryngol Head Neck Surg 121(2):157–160. Corbett S, O’Neill BJ. 1988. A cluster of cases of lymphoma in an underground colliery. Med J Aust 149(4):178–179, 181–182, 184–185. Cowan CG, Gregg TA, Napier SS, McKenna SM, Kee F. 2001. Potentially malignant oral lesions in northern Ireland: A 20-year population-based perspective of malignant transformation. Oral Dis 7(1):18–24. Cox B, Taylor K, Treasure E. 1995. Trends in oral cancer by subsite in New Zealand. Eur J Cancer B Oral Oncol 31B(2):113–117. Cox M, Maitland N, Scully C. 1993. Human herpes simplex-1 and papillomavirus type 16 homologous DNA sequences in normal, potentially malignant and malignant oral mucosa. Eur J Cancer B Oral Oncol 29B(3):215–219. Curtis RE, Boice JD, Jr., Kleinerman RA, Flannery JT, Fraumeni JF, Jr. 1985. Summary: Multiple primary cancers in Connecticut, 1935–82. Natl Cancer Inst Monogr 68:219–242. Dal Maso L, La Vecchia C, Polesel J, et al. 2002. Alcohol drinking outside meals and cancers of the upper aero-digestive tract. Int J Cancer 10(24):435–437. Dallongeville J, Marecaux N, Fruchart JC, Amouyel P. 1998. Cigarette smoking is associated with unhealthy patterns of nutrient intake: A meta-analysis. J Nutr 128(9):1450–1457. Dardanoni L, Gafa L, Paterno R, Pavone G. 1984. A case-control study on lip cancer risk factors in Ragusa (Sicily). Int J Cancer 34(3):335– 337. Davis S, Severson RK. 1987. Increasing incidence of cancer of the tongue in the United States among young adults. Lancet 2(8564):910–911. Day GL, Blot WJ, McLaughlin JK, Fraumeni JF. 1994a. Carcinogenic risk of dark vs light liquor. Int J Cancer 59(3):319–321. Day GL, Blot WJ, Shore RE, et al. 1994b. Second cancers following oral and pharyngeal cancers: Role of tobacco and alcohol. J Natl Cancer Inst 86(2):131–137. Day GL, Shore RE, Blot WJ, et al. 1994c. Dietary factors and second primary cancers: A follow-up of oral and pharyngeal cancer patients. Nutr Cancer 21(3):223–232. De Andrade M, Amos CI, Foulkes WD 1998. Segregation analysis of squamous cell carcinoma of the head and neck: Evidence for a major gene determining risk. Ann Hum Genet 62(6):505–510. De Stefani E, Boffetta P, Oreggia F, Mendilaharsu M, Deneo-Pellegrini H. 1998. Smoking patterns and cancer of the oral cavity and pharynx: A casecontrol study in Uruguay. Oral Oncol 34(5):340–346. De Stefani E, Deneo-Pellegrini H, Mendilaharsu M, Ronco A. 1999. Diet and risk of cancer of the upper aerodigestive tract—I. Foods. Oral Oncol 35(1):17–21. de Visscher JG, Schaapveld M, Otter R, Visser O, van der Waal I. 1998. Epidemiology of cancer of the lip in The Netherlands. Oral Oncol 34(5):421–426.

Decoufle P. 1979. Cancer risks associated with employment in the leather and leather products industry. Arch Env Health 34(1):33–37. Depue RH. 1986. Rising mortality from cancer of the tongue in young white males. N Engl J Med 315(10):647. Devesa SS, Blot WJ, Stone BJ, Miller BA, Tarone RE, Fraumeni JF, Jr. 1995. Recent cancer trends in the United States. J Natl Cancer Inst 87(3):175–182. Devesa SS, Blot WJ, Fraumeni JF, Jr. 1998. Changing patterns in the incidence of esophageal and gastric carcinoma in the United States. Cancer 83(10):2049–2053. Devesa SS, Grauman DG, Blot WJ, Pennello G, Hoover RN, Fraumeni JF, Jr. 1999. Atlas of cancer mortality in the United States, 1950–94. Washington DC: US Government Printing Office. Dikshit RP, Kanhere S. 2000. Tobacco habits and risk of lung, oropharyngeal and oral cavity cancer: A population-based case-control study in Bhopal, India. Int J Epidemiol 29(4):609–614. Djordjevic MV, Brunnemann KD, Hoffmann D. 1993. The need for regulation of carcinogenic N-nitrosamines in oral snuff. Food Chem Toxicol 31(7):497–501. Doll R. 1991. Urban and rural factors in the aetiology of cancer. Int J Cancer 47(6):803–810. Ebenius B. 1943. Cancer of the lip: A clinical study of 778 cases with particular regard to predisposing factors and radium therapy. Acta Radiol (Stockholm) 48:1–232. Einhorn J, Wersall J. 1967. Incidence of oral carcinoma in patients with leukoplakia of the oral mucosa. Cancer 20(12):2189–2193. Elahi A, Zheng Z, Park J, Eyring K, McCaffrey T, Lazarus P. 2002. The human OGG1 DNA repair enzyme and its association with orolaryngeal cancer risk. Carcinogenesis 23(7):1229–1234. Elwood JM, Pearson JC, Skippen DH, Jackson SM. 1984. Alcohol, smoking, social and occupational factors in the aetiology of cancer of the oral cavity, pharynx and larynx. Int J Cancer 34(5):603–612. Epstein JB, Wong FL, Millner A, Le ND. 1994. Topical bleomycin treatment of oral leukoplakia: A randomized double-blind clinical trial. Head Neck 16(6):539–44. Erdmann J. 2003. Recent studies attempt to clarify relationship between oral cancer and human papillomavirus. J Natl Cancer Inst 95(9):638–639. Evstifeeva TV, Zaridze DG. 1992. Nass use, cigarette smoking, alcohol consumption and risk of oral and oesophageal precancer. Eur J Cancer B Oral Oncol 28B(1):29–35. Fisher MJ, Virmani AK, Wu L, et al. 2001. Nucleotide substitution in the ectodomain of trail receptor DR4 is associated with lung cancer and head and neck cancer. Clin Cancer Res 7(6):1688–1697. Foulkes WD, Brunet JS, Kowalski LP, Narod SA, Franco EL. 1995. Family history of cancer is a risk factor for squamous cell carcinoma of the head and neck in Brazil: A case-control study. Int J Cancer 63(6):769–773. Foulkes WD, Brunet JS, Sieh W, Black MJ, Shenouda G, Narod SA. 1996. Familial risks of squamous cell carcinoma of the head and neck: Retrospective case-control study. Br Med J 313(7059):716–721. Franceschi S, Talamini R, Barra S, et al. 1990. Smoking and drinking in relation to cancers of the oral cavity, pharynx, larynx, and esophagus in northern Italy. Cancer Res 50(20):6502–6507. Franceschi S, Levi F, Lucchini F, Negri E, Boyle P, La Vecchia C. 1994. Trends in cancer mortality in young adults in Europe, 1955–1989. Eur J Cancer 30A(14):2096–2118. Franceschi S, Levi F, Conti E, et al. 1999a. Energy intake and dietary pattern in cancer of the oral cavity and pharynx. Cancer Causes Control 10(5):439–444. Franceschi S, Levi F, La Vecchia C, et al. 1999b. Comparison of the effect of smoking and alcohol drinking between oral and pharyngeal cancer. Int J Cancer 83(1):1–4. Franceschi S, Levi F, Dal Maso L, et al. 2000. Cessation of alcohol drinking and risk of cancer of the oral cavity and pharynx. Int J Cancer 85(6):787–790. Franco EL, Kowalski LP, Oliveira BV, et al. 1989. Risk factors for oral cancer in Brazil: A case-control study. Int J Cancer 43(6):992–1000. Frisch M, Melbye M. 1995. New primary cancers after squamous cell skin cancer. Am J Epidemiol 141(10):916–922. Frisch M, Biggar RJ. 1999. Aetiological parallel between tonsillar and anogenital squamous-cell carcinomas. Lancet 354(9188):1442–1443. Fritz A, Percy C, Jack A, Shanmugaratnam K, Sobin L, Parkin DM, Whelan S. 2000. International Classification of Diseases for Oncology. Geneva: World Health Organization. Garrote LF, Herrero R, Reyes RM, et al. 2001. Risk factors for cancer of the oral cavity and oro-pharynx in Cuba. Br J Cancer 85(1):46–54. Geisler SA, Olshan AF. 2001. GSTM1, GSTT1, and the risk of squamous cell carcinoma of the head and neck: A mini-HuGE review. Am J Epidemiol 154(2):95–105.

Cancers of the Oral Cavity and Pharynx Gerlach KK, Cumings KM, Hyland A, Gilpin EA, Johnson MD, Pierce JP. 1998. Trends in cigar consumption and smoking prevalence. In: National Cancer Institute. Cigars: Health Effects and Trends. Smoking Tob Control Monog 9. Bethesda, MD: National Cancer Institute. NIH Publ. No. 98-4302. Ghosh S, Shukla HS, Mohapatra SC, Shukla PK. 1996. Keeping chewing tobacco in the cheek pouch overnight (night quid) increases risk of cheek carcinoma. Eur J Surg Oncol 22(4):359–360. Gillison ML, Koch WM, Capone RB, et al. 2000. Evidence for a causal association between human papillomavirus and a subset of head and neck cancers [comment]. J Natl Cancer Inst 92(9):709–720. Gillison ML, Shah KV. 2001. Human papillomavirus-associated head and neck squamous cell carcinoma: Mounting evidence for an etiologic role for human papillomavirus in a subset of head and neck cancers. Curr Opin Oncol 13(3):183–188. Goldgar DE, Easton DF, Cannon-Albright LA, Skolnick MH. 1994. Systematic population-based assessment of cancer risk in first-degree relatives of cancer probands. J Natl Cancer Inst 86(21):1600–1608. Goldstein AM, Blot WJ, Greenberg RS, et al. 1994. Familial risk in oral and pharyngeal cancer. Eur J Cancer B Oral Oncol 30B(5):319–322. Goode EL, Ulrich CM, Potter JD. 2002. Polymorphisms in DNA repair genes and associations with cancer risk. Cancer Epidemiol Biomarkers Prev 11(12):1513–1530. Grady D, Greene J, Daniels TE, et al. 1990. Oral mucosal lesions found in smokeless tobacco users. J Am Dent Assoc 121(1):117–123. Graham S, Blanchet M, Rohrer T. 1977. Cancer in asbestos-mining and other areas of Quebec. J Natl Cancer Inst 59(4):1139–1145. Greenberg RS, Haber MJ, Clark WS, et al. 1991. The relation of socioeconomic status to oral and pharyngeal cancer. Epidemiol 2(3):194–200. Gridley G, McLaughlin JK, Block G, et al. 1990. Diet and oral and pharyngeal cancer among blacks. Nutr Cancer 14(3–4):219–225. Gupta PC. 1984a. Epidemiologic study of the association between alcohol habits and oral leukoplakia. Commun Dent Oral Epidemiol 12(1):47– 50. Gupta PC. 1984b. A study of dose-response relationship between tobacco habits and oral leukoplakia. Br J Cancer 50(4):527–531. Gupta PC, Bhonsle RB, Murti PR, Daftary DK, Mehta FS, Pindborg JJ. 1989. An epidemiologic assessment of cancer risk in oral precancerous lesions in India with special reference to nodular leukoplakia. Cancer 63(11):2247–2252. Gupta PC, Murti PR, Bhonsle RB, Mehta FS, Pindborg JJ. 1995. Effect of cessation of tobacco use on the incidence of oral mucosal lesions in a 10-yr follow-up study of 12,212 users. Oral Dis 1(1):54–58. Gupta PC. 1996. Survey of sociodemographic characteristics of tobacco use among 99,598 individuals in Bombay, India using handheld computers. Tob Control 5:114–120. Gupta PC, Hebert JR, Bhonsle RB, Sinor PN, Mehta H, Mehta FS. 1998a. Dietary factors in oral leukoplakia and submucous fibrosis in a population-based case control study in Gujarat, India. Oral Dis 4(3):200–206. Gupta PC, Sinor PN, Bhonsle RB, Pawar VS, Mehta HC. 1998b. Oral submucous fibrosis in India: A new epidemic? Natl Med J India 11(3): 113–116. Gupta PC. 1999. Mouth cancer in India: A new epidemic? J Indian Med Assoc 97(9):370–373. Gupta PC, Hebert JR, Bhonsle RB, Murti PR, Mehta H, Mehta FS. 1999. Influence of dietary factors on oral precancerous lesions in a population-based case-control study in Kerala, India. Cancer 85(9):1885–1893. Gupta PC, Warnakulasuriya S. 2002. Global epidemiology of areca nut usage. Addict Biol 7(1):77–83. Ha PK, Tong BC, Westra WH, et al. 2002. Mitochondrial C-tract alteration in premalignant lesions of the head and neck: a marker for progression and clonal proliferation. Clin Cancer Res 8(7):2260–2265. Hakansson N, Floderus B, Gustavsson P, Feychting M, Hallin N. 2001. Occupational sunlight exposure and cancer incidence among Swedish construction workers. Epidemiol 12(5):552–557. Hamilton-Dutoit SJ, Therkildsen MH, Neilsen NH, Jensen H, Hansen JP, Pallesen G. 1991. Undifferentiated carcinoma of the salivary gland in Greenlandic Eskimos: Demonstration of Epstein-Barr virus DNA by in situ nucleic acid hybridization. Hum Pathol 22(8):811–815. Harty LC, Caporaso NE, Hayes RB, et al. 1997. Alcohol dehydrogenase 3 genotype and risk of oral cavity and pharyngeal cancers [see comments]. J Natl Cancer Inst 89(22):1698–1705. Hashibe M, Mathew B, Kuruvilla B, et al. 2000a. Chewing tobacco, alcohol, and the risk of erythroplakia. Cancer Epidemiol Biomarkers Prev 9(7):639–645. Hashibe M, Sankaranarayanan R, Thomas G, et al. 2000b. Alcohol drinking, body mass index and the risk of oral leukoplakia in an Indian population. Int J Cancer 88(1):129–134.

691

Hashibe M, Sankaranarayanan R, Thomas G, et al. 2002. Body mass index, tobacco chewing, alcohol drinking and the risk of oral submucous fibrosis in Kerala, India. Cancer Causes Control 13(1):55–64. Hayes RB, Bravo-Otero E, Kleinman DV, et al. 1999. Tobacco and alcohol use and oral cancer in Puerto Rico. Cancer Causes Control 10(1):27–33. Hazare VK, Goel RR, Gupta PC. 1998. Oral submucous fibrosis, areca nut and pan masala use: A case-control study. Natl Med J India 11(6):299. Hemminki K, Dong C, Frisch M. 2000. Tonsillar and other upper aerodigestive tract cancers among cervical cancer patients and their husbands. Eur J Cancer Prev 9(6):433–437. Herrero R, Castellsague X, Pawlita M, et al. 2003. Human papillomavirus and oral cancer: the International Agency for Reseach on Cancer Multicenter Study. J Natl Cancer Inst 95(23):1772–1783. Hindle I, Downer MC, Speight PM. 1996. The epidemiology of oral cancer. Br J Oral Maxillofac Surg 34(5):471–476. Hindle I, Downer MC, Moles DR, Speight PM. 2000. Is alcohol responsible for more intra-oral cancer? Oral Oncol 36(4):328–333. Hirayama T. 1985. A large scale cohort study on cancer risks by diet—with special reference to the risk reducing effects of green-yellow vegetable consumption. Princess Takamatsu Symposia 16:41–53. Ho PS, Ko YC, Yang YH, Shieh TY, Tsai CC. 2002. The incidence of oropharyngeal cancer in Taiwan: An endemic betel quid chewing area. J Oral Pathol Med 31(4):213–219. Hoffmann D, Harley NH, Fisenne I, Adams JD, Brunnemann KD. 1986. Carcinogenic agents in snuff. J Natl Cancer Inst 76(3):435–437. Homann N, Jousimies-Somer H, Jokelainen K, Heine R, Salaspuro M. 1997. High acetaldehyde levels in saliva after ethanol consumption: Methodological aspects and pathogenetic implications. Carcinogenesis 18(9): 1739–1743. Hong WK, Lippman SM, Itri LM, et al. 1990. Prevention of second primary tumors with isotretinoin in squamous-cell carcinoma of the head and neck [comment]. N Engl J Med 323(12):795–801. Horn-Ross PL, Ljung BM, Morrow M. 1997a. Environmental factors and the risk of salivary gland cancer. Epidemiol 8(4):414–419. Horn-Ross PL, Morrow M, Ljung BM. 1997b. Diet and the risk of salivary gland cancer. Am J Epidemiol 146(2):171–176. Horn-Ross PL, Morrow M, Ljung BM. 1999. Menstrual and reproductive factors for salivary gland cancer risk in women. Epidemiol 10(5): 528–530. Hsieh LL, Chien HT, Chen IH, et al. 2003. The XRCC1 399Gln polymorphism and the frequency of p53 mutations in Taiwanese oral squamous cell carcinomas. Cancer Epidemiol Biomarkers Prev 12(5):439–443. Hu YC, Sidransky D, Ahrendt SA. 2002. Molecular detection approaches for smoking associated tumors. Oncogene 21(48):7289–7297. Huang WY, Winn DM, Brown LM, et al. 2003. Alcohol concentration and risk of oral cancer in Puerto Rico. Am J Epidemiol 157(10):881–887. Huebner WW, Schoenberg JB, Kelsey JL, et al. 1992. Oral and pharyngeal cancer and occupation: A case-control study. Epidemiol 3(4):300–309. Hung HC, Chuang J, Chien YC, et al. 1997. Genetic polymorphisms of CYP2E1, GSTM1, and GSTT1; environmental factors and risk of oral cancer. Cancer Epidemiol Biomarkers Prev 6(11):901–905. IARC Working Group on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, International Agency for Research on Cancer, World Health Organization. 1999. Volume 71. Re-Evaluation of Some Organic Chemicals, Hydrazine and Hydrogen Peroxide. Lyon, France: International Agency for Research on Cancer. IARC Working Group on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, International Agency for Research on Cancer, World Health Organization. 1985. Volume 37. Tobacco habits other than smoking, betel-quid and areca-nut chewing, and some related nitrosamines. Lyon, France: International Agency for Research on Cancer. IARC Working Group on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, International Agency for Research on Cancer, World Health Organization. 1988. Volume 44. Alcohol drinking. Lyon, France: International Agency for Research on Cancer. IARC Working Group on the Evaluation of the Carcinogenic Risk of Chemicals to Humans. 2004. Volume 85. Betel-quid and areca-nut chewing and some areca-nut related nitrosamines. Lyon, France: International Agency for Research on Cancer. Idris AM, Ahmed HM, Malik MO. 1995. Toombak dipping and cancer of the oral cavity in the Sudan: A case-control study. Int J Cancer 63(4):477–480. Idris AM, Ibrahim SO, Vasstrand EN, et al. 1998. The Swedish snus and the Sudanese toombak: Are they different? Oral Oncol 34(6):558–566. International Agency for Research on Cancer. 1998. IARC Handbooks of Cancer Prevention, Volume 2, Carotenoids. Lyon, France: International Agency for Research on Cancer.

692

PART IV: CANCER BY TISSUE OF ORIGIN

Johansen C, Boice J, Jr., McLaughlin J, Olsen J. 2001. Cellular telephones and cancer—a nationwide cohort study in Denmark. J Natl Cancer Inst 93(3):203–207. Johns MM, 3rd, Westra WH, Califano JA, Eisele D, Koch WM, Sidransky D. 1996. Allelotype of salivary gland tumors. Cancer Res 56(5):1151– 1154. Joint WHO/FAO Expert Consultation. 2003. Diet, Nutrition and the Prevention of Chronic Diseases. WHO Technical Report Series 916. Geneva: World Health Organization. Jourenkova-Mironova N, Mitrunen K, Bouchardy C, Dayer P, Benhamou S, Hirvonen A. 2000. High-activity microsomal epoxide hydrolase genotypes and the risk of oral, pharynx, and larynx cancers. Cancer Res 60(3):534–536. Jovanovic A, Kostense PJ, Schulten EA, Snow GB, van der Waal I. 1992. Delay in diagnosis of oral squamous cell carcinoma; a report from The Netherlands. Eur J Cancer B Oral Oncol 28B(1):37–38. Jovanovic A, Schulten EA, Kostense PJ, Snow GB, van der Waal I. 1993a. Tobacco and alcohol related to the anatomical site of oral squamous cell carcinoma. J Oral Pathol Med 22(10):459–462. Jovanovic A, Schulten EA, Kostense PJ, Snow GB, van der Waal I. 1993b. Squamous cell carcinoma of the lip and oral cavity in The Netherlands; an epidemiological study of 740 patients. J Craniomaxillofac Surg 21(4):149–152. Jovanovic A, van der Tol IG, Kostense PJ, et al. 1994a. Second respiratory and upper digestive tract cancer following oral squamous cell carcinoma. Eur J Cancer B Oral Oncol 30B(4):225–229. Jovanovic A, van der Tol IG, Schulten EA, et al. 1994b. Risk of multiple primary tumors following oral squamous-cell carcinoma. Int J Cancer 56(3):320–323. Ju DM. 1973. On the etiology of cancer of the lower lip. Plast Reconstr Surg 52(2):151–154. Kabat GC, Hebert JR, Wynder EL. 1989. Risk factors for oral cancer in women. Cancer Res 49(10):2803–2806. Kabat GC, Chang CJ, Wynder EL. 1994. The role of tobacco, alcohol use, and body mass index in oral and pharyngeal cancer. Int J Epidemiol 23(6): 1137–1144. Kasum CM, Jacobs DR, Jr., Nicodemus K, Folsom AR. 2002. Dietary risk factors for upper aerodigestive tract cancers. Int J Cancer 99(2):267–272. Katoh T, Kaneko S, Boissy R, Watson M, Ikemura K, Bell DA. 1998. A pilot study testing the association between N-acetyltransferases 1 and 2 and risk of oral squamous cell carcinoma in Japanese people. Carcinogenesis 19(10):1803–1807. Katoh T, Kaneko S, Kohshi K, et al. 1999. Genetic polymorphisms of tobaccoand alcohol-related metabolizing enzymes and oral cavity cancer. Int J Cancer 83(5):606–609. Kaugars GE, Pillion T, Svirsky JA, Page DG, Burns JC, Abbey LM. 1999. Actinic cheilitis: A review of 152 cases. Oral Surg Oral Med Oral Pathol Oral Radiol Endod 88(2):181–186. Keller AZ. 1969. Residence, age, race and related factors in the survival and associations with salivary tumors. Am J Epidemiol 90(4):269–277. Keller AZ. 1970. Cellular types, survival, race, nativity, occupations, habits and associated diseases in the pathogenesis of lip cancers. Am J Epidemiol 91(5):486–499. Keller AZ. 1983. Selected factors in the risks of upper alimentary cancers. Prev Med 12(4):541–553. Khan FA, Robinson PG, Warnakulasuriya KA, Newton JT, Gelbier S, Gibbons DE. 2000. Predictors of tobacco and alcohol consumption and their relevance to oral cancer control amongst people from minority ethnic communities in the South Thames health region, England. J Oral Pathol Med 29(5):214–219. Khuder SA. 1999. Etiologic clues to lip cancer from epidemiologic studies on farmers. Scand J Work Environ Health 25(2):125–130. Khuri FR, Kim ES, Lee JJ, et al. 2001. The impact of smoking status, disease stage, and index tumor site on second primary tumor incidence and tumor recurrence in the head and neck retinoid chemoprevention trial. Cancer Epidemiol Biomarkers Prev 10(8):823–829. Khuri FR, Lee JJ, Lippman SM, et al. 2003. Isotretinoin effects on head and neck cancer recurrence and second primary tumors. Proc Am Soc Clin Oncol 22(90), abstr 359. Kjaerheim K, Gaard M, Andersen A. 1998. The role of alcohol, tobacco, and dietary factors in upper aerogastric tract cancers: A prospective study of 10,900 Norwegian men. Cancer Causes Control 9(1):99–108. Ko YC, Huang YL, Lee CH, Chen MJ, Lin LM, Tsai CC. 1995. Betel quid chewing, cigarette-smoking and alcohol-consumption related to oralcancer in Taiwan. J Oral Path Med 24(10):450–453. Kutler DI, Singh B, Satagopan J, et al. 2003. A 20-year perspective on the International Fanconi Anemia Registry (IFAR). Blood 101(4):1249– 1256.

Land CE, Saku T, Hayashi Y, et al. 1996. Incidence of salivary gland tumors among atomic bomb survivors, 1950–1987. Evaluation of radiationrelated risk. Radiat Res 146(1):28–36. Lanier AP, Clift SR, Bornkamm G, Henle W, Goepfert H, Raab-Traub N. 1991. Epstein-Barr virus and malignant lymphoepithelial lesions of the salivary gland. Arctic Med Res 50(2):55–61. Lanier AP, Alberts SR. 1996. Cancers of the buccal cavity and pharynx in Circumpolar Inuit. Acta Oncol 35(5):545–552. Lay KM, Sein K, Myint A, Ko SK, Pindborg JJ. 1982. Epidemiologic study of 6000 villagers of oral precancerous lesions in Bilugyun: preliminary report. Commun Dent Oral Epidemiol 10(3):152–155. Lazarus P, Park JY. 2000. Metabolizing enzyme genotype and risk for upper aerodigestive tract cancer. Oral Oncol 36(5):421–431. Le Marchand L, Hankin JH, Carter FS, et al. 1994. A pilot study on the use of plasma carotenoids and ascorbic acid as markers of compliance to a high fruit and vegetable dietary intervention. Cancer Epidemiol Biomarkers Prev 3(3):245–251. Lee CH, Ko YC, Huang HL, et al. 2003. The precancer risk of betel quid chewing, tobacco use and alcohol consumption in oral leukoplakia and oral submucous fibrosis in southern Taiwan. Br J Cancer 88(3): 366–372. Levi F, La Vecchia C, Randimbison L, Te VC. 1995. Cancer incidence and mortality in young adults in Vaud, Switzerland, 1974–1992. Int J Cancer 61(5):606–610. Levi F, Pasche C, La Vecchia C, Lucchini F, Franceschi S, Monnier P. 1998. Food groups and risk of oral and pharyngeal cancer. Int J Cancer 77(5):705–709. Levin ML, Goldstein H, Gerhardt PR. 1950. Cancer and tobacco smoking. J Am Med Assoc 143(4):336–338. Lewin F, Norell SE, Johansson H, et al. 1998. Smoking tobacco, oral snuff, and alcohol in the etiology of squamous cell carcinoma of the head and neck: A population-based case-referent study in Sweden. Cancer 82(7):1367–1375. Li D, Firozi PF, Chang P, et al. 2001. In vitro BPDE-induced DNA adducts in peripheral lymphocytes as a risk factor for squamous cell carcinoma of the head and neck. Int J Cancer 93(3):436–440. Li N, Sun Z, Han C, Chen J. 1999. The chemopreventive effects of tea on human oral precancerous mucosa lesions. Proc Soc Exp Biol Med 220(4):218–224. Li X, Hemminki K. 2003. Familial upper aerodigestive tract cancers: Incidence trends, familial clustering and subsequent cancers. Oral Oncol 39(3): 232–239. Lindqvist C, Teppo L. 1978. Epidemiological evaluation of sunlight as a risk factor of lip cancer. Cancer 37(6):983–989. Lindqvist C. 1979. Risk factors in lip cancer: a questionnaire survey. Am J Epidemiol 109(5):521–530. Lippman SM, Hong WK. 1989. Second malignant tumors in head and neck squamous cell carcinoma: The overshadowing threat for patients with early-stage disease. Int J Radiation Oncol Biol Physics 17(3):691–694. Lippman SM, Lee JJ, Karp DD, et al. 2001. Randomized phase III intergroup trial of isotretinoin to prevent second primary tumors in stage I non-smallcell lung cancer. J Natl Cancer Inst 93(8):605–618. Lissowska J, Pilarska A, Pilarski P, et al. 2003. Smoking, alcohol, diet, dentition and sexual practices in the epidemiology of oral cancer in Poland. Eur J Cancer Prev 12(1):25–33. Liu S, Park JY, Schantz SP, Stern JC, Lazarus P. 2001. Elucidation of CYP2E1 5¢ regulatory RsaI/Pstl allelic variants and their role in risk for oral cancer. Oral Oncol 37(5):437–445. Loning T, Ikenberg H, Becker J, Gissmann L, Hoepfer I, zur Hausen H. 1985. Analysis of oral papillomas, leukoplakias, and invasive carcinomas for human papillomavirus type related DNA. J Invest Derm 84(5):417– 420. Macfarlane GJ, Boyle P, Scully C. 1992. Oral cancer in Scotland: Changing incidence and mortality. Br Med J 305(6862):1121–1123. Macfarlane GJ, Zheng T, Marshall JR, et al. 1995. Alcohol, tobacco, diet and the risk of oral-cancer—a pooled analysis of 3 case-control studies. Oral Oncol Eur J Cancer B 31B(3):181–187. Macigo FG, Mwaniki DL, Guthua SW. 1995. The association between oral leukoplakia and use of tobacco, alcohol and khat based on relative risks assessment in Kenya. Eur J Oral Sci 103(5):268–273. Maden C, Beckmann AM, Thomas DB, et al. 1992. Human papillomaviruses, herpes simplex viruses, and the risk of oral cancer in men. Am J Epidemiol 135(10):1093–1102. Maher R, Lee AJ, Warnakulasuriya KA, Lewis JA, Johnson NW. 1994. Role of areca nut in the causation of oral submucous fibrosis: A case-control study in Pakistan. J Oral Pathol Med 23(2):65–69. Maier H, de Vries N, Weidauer H. 1990. [Occupation and cancer of the oral cavity, pharynx and larynx]. [German]. HNO 38(8):271–278.

Cancers of the Oral Cavity and Pharynx Mancuso TF, Brennan MJ. 1970. Epidemiological considerations of cancer of the gallbladder, bile ducts and salivary glands in the rubber industry. J Occup Med 12(9):333–341. Mao L, Lee JS, Fan YH, et al. 1996. Frequent microsatellite alterations at chromosomes 9p21 and 3p14 in oral premalignant lesions and their value in cancer risk assessment. Nat Med 2(6):682–685. Marshall JR, Graham S, Haughey BP, et al. 1992. Smoking, alcohol, dentition and diet in the epidemiology of oral cancer. Eur J Cancer B Oral Oncol 28B(1):9–15. Martin GC, Brown JP, Eifler CW, Houston GD. 1999. Oral leukoplakia status six weeks after cessation of smokeless tobacco use. J Am Dent Assoc 130(7):945–954. Mashberg A. 1978. Erythroplasia: the earliest sign of asymptomatic oral cancer. J Am Dent Assoc 96(4):615–620. Mashberg A, Boffetta P, Winkelman R, Garfinkel L. 1993. Tobacco smoking, alcohol drinking, and cancer of the oral cavity and oropharynx among U.S. veterans. Cancer 72(4):1369–1375. Mathew B, Sankaranarayanan R, Nair PP, et al. 1995. Evaluation of chemoprevention of oral cancer with Spirulina fusiformis. Nutr Cancer 24(2):197–202. Matthias C, Bockmuhl U, Jahnke V, et al. 1998. Polymorphism in cytochrome P450 CYP2D6, CYP1A1, CYP2E1 and glutathione S-transferase, GSTM1, GSTM3, GSTT1 and susceptibility to tobacco-related cancers: Studies in upper aerodigestive tract cancers. Pharmacogenetics 8(2):91–100. Mayne ST, Cartmel B, Baum M, et al. 2001. Randomized trial of supplemental beta-carotene to prevent second head and neck cancer. Cancer Res 61(4):1457–1463. Mayne ST, Cartmel B, Morse DE. 2003. Chemoprevention of oral premalignant lesions. In: Head and Neck Cancer: Emerging Perspectives. JF Ensley, JS Gutkind, JR Jacobs and SM Lippman editors. New York: Academic Press, pp. 261–269. McCredie M, Coates M, Grulich A. 1994. Cancer incidence in migrants to New South Wales (Australia) from the Middle East, 1972–91. Cancer Causes Control 5(5):414–421. Mehta FS, Gupta PC, Pindborg JJ. 1981. Chewing and smoking habits in relation to precancer and oral cancer. J Cancer Res Clin Oncol 99(1–2):35–39. Merchant A, Husain SS, Hosain M, et al. 2000. Paan without tobacco: An independent risk factor for oral cancer. Int J Cancer 86(1):128–131. Merletti F, Boffetta P, Ciccone G, Mashberg A, Terracini B. 1989. Role of tobacco and alcoholic beverages in the etiology of cancer of the oral cavity/oropharynx in Torino, Italy. Cancer Res 49(17):4919–4924. Merletti F, Boffetta P, Ferro G, Pisani P, Terracini B. 1991. Occupation and cancer of the oral cavity or oropharynx in Turin, Italy. Scand J Work Environ Health 17(4):248–254. Merrick Y, Albeck H, Nielsen NH, Hansen HS. 1986. Familial clustering of salivary gland carcinoma in Greenland. Cancer 57(10):2097–2102. Milan T, Pukkala E, Verkasalo PK, et al. 2000. Subsequent primary cancers after basal-cell carcinoma: A nationwide study in Finland from 1953 to 1995. Int J Cancer 87(2):283–288. Modan B, Chetrit A, Alfandary E, et al. 1998. Increased risk of salivary gland tumors after low-dose irradiation. Laryngoscope 108(7):1095–1097. Moller H. 1989. Changing incidence of cancer of the tongue, oral cavity, and pharynx in Denmark. J Oral Pathol Med 18(4):224–229. Morita S, Yano M, Tsujinaka T, et al. 1999. Genetic polymorphisms of drugmetabolizing enzymes and susceptibility to head-and-neck squamous-cell carcinoma. Int J Cancer 80(5):685–688. Mork J, Moller B, Glattre E. 1999. Familial risk in head and neck squamous cell carcinoma diagnosed before the age of 45: A population-based study. Oral Oncol 35(4):360–367. Mork J, Lie AK, Glattre E, et al. 2001. Human papillomavirus infection as a risk factor for squamous-cell carcinoma of the head and neck [comment]. N Engl J Med 344(15):1125–1131. Morse DE, Katz RV, Pendrys DG, et al. 1996. Smoking and drinking in relation to oral epithelial dysplasia. Cancer Epidemiol Biomarkers Prev 5(10):769–777. Morse DE, Katz RV, Pendrys DG, et al. 1997. Mouthwash use and dentures in relation to oral epithelial dysplasia. Oral Oncol 33(5):338–343. Morse DE, Pendrys DG, Neely AL, Psoter WJ. 1999. Trends in the incidence of lip, oral, and pharyngeal cancer: Connecticut, 1935–94. Oral Oncol 35(1):1–8. Morse DE, Pendrys DG, Katz RV, et al. 2000. Food group intake and the risk of oral epithelial dysplasia in a United States population. Cancer Causes Control 11(8):713–720. Muir C, Waterhouse J, Mack T, Powell J, Whelan S, editors. 1987. Cancer Incidence in Five Continents, Volume V. Lyon, France: International Agency for Research on Cancer. IARC Scientific Publications No. 88.

693

Murti PR, Bhonsle RB, Pindborg JJ, Daftary DK, Gupta PC, Mehta FS. 1985. Malignant transformation rate in oral submucous fibrosis over a 17-year period. Commun Dent Oral Epidemiol 13(6):340–341. Muscat JE, Wynder EL. 1998. A case/control study of risk factors for major salivary gland cancer. Otolaryngol Head Neck Surg 118(2):195–198. Muto M, Hitomi Y, Ohtsu A, et al. 2000. Acetaldehyde production by nonpathogenic Neisseria in human oral microflora: implications for carcinogenesis in upper aerodigestive tract. Int J Cancer 88(3):342–350. National Cancer Institute. 1997. Changes in cigarette-related disease risks and their implication for prevention and control. Smoking Tob Control Monog 8. Bethesda, MD: National Cancer Insitute. (NIH Publ. No. 97–4213). National Cancer Institute 2002. Smokeless Tobacco Fact Sheets (http://dccps.nci.nih.gov/tcrb/stfact_sheet_combined10–23–02.pdf) accessed December 2003. Negri E, La Vecchia C, Franceschi S, Tavani A. 1993. Attributable risk for oral cancer in northern Italy. Cancer Epidemiol Biomarkers Prev 2(3):189–193. Negri E, Franceschi S, Bosetti C, et al. 2000. Selected micronutrients and oral and pharyngeal cancer. Int J Cancer 86(1):122–127. Nelson DE, Davis RM, Chrismon JH, Giovino GA. 1996. Pipe smoking in the United States, 1965–1991: Prevalence and attributable mortality. Prev Med 25(2):91–99. Nicolau SG, Balus L. 1964. Chronic actinic cheilitis and cancer of the lower lip. Br J Dermatol 76:278–289. Nomura T, Noma H, Shibahara T, Yokoyama A, Muramatusu T, Ohmori T. 2000. Aldehyde dehydrogenase 2 and glutathione S-transferase M 1 polymorphisms in relation to the risk for oral cancer in Japanese drinkers. Oral Oncol 36(1):42–46. O’Hanlon S, Forster DP, Lowry RJ. 1997. Oral cancer in the North-East of England: Incidence, mortality trends and the link with material deprivation. Commun Dent Oral Epidemiol 25(5):371–376. Olshan AF, Weissler MC, Watson MA, Bell DA. 2000. GSTM1, GSTT1, GSTP1, CYP1A1, and NAT1 polymorphisms, tobacco use, and the risk of head and neck cancer. Cancer Epidemiol Biomarkers Prev 9(2):185–191. Olshan AF, Weissler MC, Watson MA, Bell DA. 2001. Risk of head and neck cancer and the alcohol dehydrogenase 3 genotype. Carcinogenesis 22(1):57–61. Olshan AF, Watson MA, Weissler MC, Bell DA. 2002. XRCC1 polymorphisms and head and neck cancer. Cancer Lett 178(2):181–186. Omenn GS, Goodman GE, Thornquist MD, et al. 1996. Effects of a combination of beta carotene and vitamin A on lung cancer and cardiovascular disease. N Engl J Med 334(18):1150–1155. Ostman J, Anneroth G, Gustafsson H, Tavelin B. 1997. Malignant salivary gland tumours in Sweden 1960–1989—an epidemiological study. Oral Oncol 33(3):169–176. Parkin DM, Muir CS, Whelan SL, Gao Y-T, Ferlay J, Powell J, eds. 1992. Cancer Incidence in Five Continents, Volume VI. Lyon, France: International Agency for Research on Cancer. IARC Scientific Publications No. 120. Parkin DM, Whelan SL, Ferlay J, Raymond L, Young J, eds. 1997. Cancer Incidence in Five Continents, Volume VII. Lyon, France: International Agency for Research on Cancer. IARC Scientific Publications No. 143. Parkin DM, Bray FI, Devesa SS. 2001. Cancer burden in the year 2000. The global picture. Eur J Cancer 37 Suppl 8:S4–66. Parkin DM, Whelan SL, Ferlay J, Teppo L, Thomas D, eds. 2002. Cancer Incidence in Five Continents, Volume VIII. Lyon, France: International Agency for Research on Cancer. IARC Scientific Publications No. 155. Partridge M, Pateromichelakis S, Phillips E, Emilion G, Langdon J. 2001. Profiling clonality and progression in multiple premalignant and malignant oral lesions identifies a subgroup of cases with a distinct presentation of squamous cell carcinoma. Clin Cancer Res 7(7):1860–1866. Petridou E, Zavras AI, Lefatzis D, et al. 2002. The role of diet and specific micronutrients in the etiology of oral carcinoma. Cancer 94(11): 2981–2988. Pindborg JJ, Reichart PA, Smith CJ, van der Waal I, in collaboration with Sobin LH, eds. 1997. Histological Typing of Cancer and Precancer of the Oral Mucosa. New York: Springer. Plesko I, Macfarlane GJ, Evstifeeva TV, Obsitnikova A, Kramarova E. 1994. Oral and pharyngeal cancer incidence in Slovakia 1968–1989. Int J Cancer 56(4):481–486. Pogoda JM, Preston-Martin S. 1996. Solar radiation, lip protection, and lip cancer risk in Los Angeles County women (California, United States). Cancer Causes Control 7(4):458–463. Preston-Martin S, Thomas DC, White SC, Cohen D. 1988. Prior exposure to medical and dental x-rays related to tumors of the parotid gland. J Natl Cancer Inst 80(12):943–949.

694

PART IV: CANCER BY TISSUE OF ORIGIN

Preston-Martin S. 1989. Prior x-ray therapy for acne related to tumors of the parotid gland. Arch Dermatol 125(7):921–924. Preston-Martin S, White SC. 1990. Brain and salivary gland tumors related to prior dental radiography: Implications for current practice. J Am Dent Assoc 120(2):151–158. Prior P, Waterhouse JA. 1977. Second primary cancers in patients with tumours of the salivary glands. Br J Cancer 36(3):362–368. Psoter WJ, Morse DE. 2001. Annual per capita apparent consumption of tobacco products in the United States: 1900–1990. Prev Med 32(1):1–9. Pukkala E, Soderholm AL, Lindqvist C. 1994. Cancers of the lip and oropharynx in different social and occupational groups in Finland. Eur J Cancer B Oral Oncol 30B(3):209–215. Rao DN, Ganesh B, Rao RS, Desai PB. 1994. Risk assessment of tobacco, alcohol and diet in oral cancer–a case-control study. Int J Cancer 58(4):469–473. Rao DN, Desai PB. 1998. Risk assessment of tobacco, alcohol and diet in cancers of base tongue and oral tongue—a case control study. Indian J Cancer 35(2):65–72. Reddy CR, Kameswari VR, Chandramouli KB, Prahlad D, Ramulu C. 1975. Evaluation of oral, pharyngeal, laryngeal and esophageal cancer risk in reverse smokers of chuttas. Int Surg 60(5):266–269. Reed AL, Califano J, Cairns P, et al. 1996. High frequency of p16 (CDKN2/MTS-1/INK4A) inactivation in head and neck squamous cell carcinoma. Cancer Res 56(16):3630–3633. Ries LAG, Eisner MP, Kosary CL, et al. 2003. SEER Cancer Statistics Review, 1975–2000: National Cancer Institute. Bethesda, MD. Robinson KL, Macfarlane GJ. 2003. Oropharyngeal cancer incidence and mortality in Scotland: are rates still increasing? Oral Oncol 39(1): 31–36. Roed-Petersen B. 1982. Effect on oral leukoplakia of reducing or ceasing tobacco smoking. Acta Derm Venereol 62(2):164–167. Rosas SL, Koch W, da Costa Carvalho MG, et al. 2001. Promoter hypermethylation patterns of p16, O6-methylguanine-DNA-methyltransferase, and death-associated protein kinase in tumors and saliva of head and neck cancer patients. Cancer Res 61(3):939–942. Rosenberg PS, Greene MH, Alter BP. 2003. Cancer incidence in persons with Fanconi anemia. Blood 101(3):822–826. Rosin MP, Cheng X, Poh C, et al. 2000. Use of allelic loss to predict malignant risk for low-grade oral epithelial dysplasia. Clin Cancer Res 6(2):357–362. Rubino C, De Vathaire F, Dottorini ME, et al. 2003. Second primary malignancies in thyroid cancer patients. Br J Cancer 89(9):1638– 1644. Saku T, Hayashi Y, Takahara O, et al. 1997. Salivary gland tumors among atomic bomb survivors, 1950–1987. Cancer 79(8):1465–1475. Sanchez MJ, Martinez C, Nieto A, et al. 2003. Oral and oropharyngeal cancer in Spain: Influence of dietary patterns. Eur J Cancer Prev 12(1):49– 56. Sankaranarayanan R, Duffy SW, Padmakumary G, Day NE, Padmanabhan TK. 1989. Tobacco chewing, alcohol and nasal snuff in cancer of the gingiva in Kerala, India. Br J Cancer 60(4):638–643. Sankaranarayanan R, Duffy SW, Padmakumary G, Day NE, Krishan Nair M. 1990. Risk factors for cancer of the buccal and labial mucosa in Kerala, southern India. J Epidemiol Commun Health 44(4):286–292. Sato M, Sato T, Izumo T, Amagasa T. 2000. Genetically high susceptibility to oral squamous cell carcinoma in terms of combined genotyping of CYP1A1 and GSTM1 genes. Oral Oncol 36(3):267–271. Schantz SP, Ostroff JS. 1997. Novel approaches to the prevention of head and neck cancer. Proc Soc Exp Biol Med 216(2):275–282. Schantz SP, Yu GP. 2002. Head and neck cancer incidence trends in young Americans, 1973–1997, with a special analysis for tongue cancer. Arch Otolaryngol Head Neck Surg 128(3):268–274. Schepman KP, van der Meij EH, Smeele LE, van der Waal I. 1998. Malignant transformation of oral leukoplakia: A follow-up study of a hospital-based population of 166 patients with oral leukoplakia from The Netherlands. Oral Oncol 34(4):270–275. Schildt EB, Eriksson M, Hardell L, Magnuson A. 1998. Oral snuff, smoking habits and alcohol consumption in relation to oral cancer in a Swedish case-control study. Int J Cancer 77(3):341–346. Schlecht NF, Franco EL, Pintos J, Kowalski LP. 1999. Effect of smoking cessation and tobacco type on the risk of cancers of the upper aero-digestive tract in Brazil. Epidemiol 10(4):412–418. Schou G, Storm HH, Jensen OM. 1985. Second cancer following cancers of the buccal cavity and pharynx in Denmark, 1943–80. Natl Cancer Inst Monogr 68:253–276. Schouten LJ, Meijer H, Huveneers JA, Kiemeney LA. 1996. Urban-rural differences in cancer incidence in The Netherlands 1989–1991. Int J Epidemiol 25(4):729–736.

Schwartz SM, Daling JR, Doody DR, et al. 1998. Oral cancer risk in relation to sexual history and evidence of human papillomavirus infection. J Natl Cancer Inst 90(21):1626–1636. Schwartz SM, Doody DR, Fitzgibbons ED, Ricks S, Porter PL, Chen C. 2001a. Oral squamous cell cancer risk in relation to alcohol consumption and alcohol dehydrogenase-3 genotypes. Cancer Epidemiol Biomarkers Prev 10(11):1137–1144. Schwartz SR, Yueh B, McDougall JK, Daling JR, Schwartz SM. 2001b. Human papillomavirus infection and survival in oral squamous cell cancer: A population-based study. Otolaryngol Head Neck Surg 125(1):1–9. Sciubba JJ. 1995. Oral leukoplakia. Crit Rev Oral Biol Med 6(2):147–160. Scully C. 2002. Oral squamous cell carcinoma; from an hypothesis about a virus, to concern about possible sexual transmission. Oral Oncol 38(3):227–234. SEER. 2002. Program Public-use Data (1973–1999): National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch. SEER. 2003a. (www.seer.cancer. gov) SEER*Stat Database: Mortality—All COD, Public-Use With State, Total U.S. for Expanded Races/Hispanics (1990–2000): National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2003. Underlying mortality data provided by NCHS (www.cdc.gov/nchs). SEER. 2003b. (www.seer.cancer.gov) SEER*Stat Database: Incidence—SEER 9 Regs Public-Use, Nov 2002 Submission (1973–2000), released April 2003: National Cancer Institute, DCCPS Surveillance Program, Cancer Statistics Branch. SEER. 2003c. (www.seer.cancer.gov) SEER*Stat Database: Incidence—SEER 12 Regs Public-Use, Nov 2002 Submission (1973–2000), released April 2003: National Cancer Institute, DCCPS Surveillance Program, Cancer Statistics Branch. SEER. 2003d. (www.seer.cancer.gov) SEER*Stat Database: Incidence—SEER 11 Regs Public-Use, Nov 2002 Submission for Hispanics (1992–2000), released April 2003: National Cancer Institute, DCCPS Surveillance Program, Cancer Statistics Branch. Shafer WG, Waldron CA. 1975. Erythroplakia of the oral cavity. Cancer 36(3):1021–1028. Shah KV. 1998. Do human papillomavirus infections cause oral cancer? J Natl Cancer Inst 90(21):1585–1586. Shah N, Sharma PP. 1998. Role of chewing and smoking habits in the etiology of oral submucous fibrosis (OSF): A case-control study. J Oral Pathol Med 27(10):475–479. Shapiro JA, Jacobs EJ, Thun MJ. 2000. Cigar smoking in men and risk of death from tobacco-related cancers. J Natl Cancer Inst 92(4):333– 337. Sheen TS, Tsai CC, Ko JY, Chang YL, Hsu MM. 1997. Undifferentiated carcinoma of the major salivary glands. Cancer 80(3):357–363. Shen H, Sturgis EM, Khan SG, et al. 2001. An intronic poly (AT) polymorphism of the DNA repair gene XPC and risk of squamous cell carcinoma of the head and neck: a case-control study. Cancer Res 61(8):3321–3325. Shiu MN, Chen TH, Chang SH, Hahn LJ. 2000. Risk factors for leukoplakia and malignant transformation to oral carcinoma: a leukoplakia cohort in Taiwan. Br J Cancer 82(11):1871–1874. Shohaimi S, Luben R, Wareham N, et al. 2003. Residential area deprivation predicts smoking habit independently of individual educational level and occupational social class. A cross sectional study in the Norfolk cohort of the European Investigation into Cancer (EPIC-Norfolk). J Epidemiol Commun Health 57(4):270–276. Sidransky D. 2002. Emerging molecular markers of cancer. Nat Rev Cancer 2(3):210–219. Silverman S, Jr., Gorsky M, Lozada F. 1984. Oral leukoplakia and malignant transformation. A follow-up study of 257 patients. Cancer 53(3):563– 568. Singh B, Wreesmann VB, Pfister D, et al. 2002. Chromosomal aberrations in patients with head and neck squamous cell carcinoma do not vary based on severity of tobacco/alcohol exposure. BMC Genet 3(1):22. Slade J, Connolly GN, Lymperis D. 2002. Eclipse: Does it live up to its health claims? Tob Control 11(Suppl 2):ii64–70. Soderholm AL, Pukkala E, Lindqvist C, Teppo L. 1994. Risk of new primary cancer in patients with oropharyngeal cancer. Br J Cancer 69(4):784– 787. Soni S, Kaur J, Kumar A, Chakravarti N, Mathur M, Bahadur S, Shukla NK, Deo SV, Ralhan R. 2005. Alterations of rb pathway components are frequent events in patients with oral epithelial dysplasia and predict clinical outcome in patients with squamous cell carcinoma. Oncol 68(4–6): 314–325. Spitz MR, Tilley BC, Batsakis JG, Gibeau JM, Newell GR. 1984. Risk factors for major salivary gland carcinoma. A case-comparison study. Cancer 54(9):1854–1859.

Cancers of the Oral Cavity and Pharynx Spitz MR, Sider JG, Newell GR, Batsakis JG. 1988. Incidence of salivary gland cancer in the United States relative to ultraviolet radiation exposure. Head Neck Surg 10(5):305–308. Spitz MR, Fueger JJ, Beddingfield NA, et al. 1989. Chromosome sensitivity to bleomycin-induced mutagenesis, an independent risk factor for upper aerodigestive tract cancers. Cancer Res 49(16):4626–4628. Spitz MR, Fueger JJ, Goepfert H, Newell GR. 1990. Salivary gland cancer. A case-control investigation of risk factors. Arch Otolaryngol Head Neck Surg 116(10):1163–1166. Spitz MR, Sider JG, Schantz SP, Newell GR. 1992. Association between malignancies of the upper aerodigestive tract and uterine cervix. Head Neck 14(5):347–351. Spitz MR, Fueger JJ, Halabi S, Schantz SP, Sample D, Hsu TC. 1993. Mutagen sensitivity in upper aerodigestive tract cancer: A case-control analysis. Cancer Epidemiol Biomarkers Prev 2(4):329–33. Spitzer WO, Hill GB, Chambers LW, Helliwell BE, Murphy HB. 1975. The occupation of fishing as a risk factor in cancer of the lip. N Engl J Med 293(9):419–424. Sreelekha TT, Ramadas K, Pandey M, Thomas G, Nalinakumari KR, Pillai MR. 2001. Genetic polymorphism of CYP1A1, GSTM1 and GSTT1 genes in Indian oral cancer. Oral Oncol 37(7):593–598. Stables GJ, Subar AF, Patterson BH, et al. 2002. Changes in vegetable and fruit consumption and awareness among US adults: Results of the 1991 and 1997 5 A Day for Better Health Program surveys. J Am Dietetic Assoc 102(6):809–817. Sturgis EM, Castillo EJ, Li L, et al. 1999. Polymorphisms of DNA repair gene XRCC1 in squamous cell carcinoma of the head and neck. Carcinogenesis 20(11):2125–2129. Sturgis EM, Zheng R, Li L, et al. 2000. XPD/ERCC2 polymorphisms and risk of head and neck cancer: a case-control analysis. Carcinogenesis 21(12): 2219–2223. Sturgis EM, Dahlstrom KR, Guan Y, et al. 2001. Alcohol dehydrogenase 3 genotype is not associated with risk of squamous cell carcinoma of the oral cavity and pharynx. Cancer Epidemiol Biomarkers Prev 10(3):273–275. Sturgis EM, Castillo EJ, Li L, et al. 2002. XPD/ERCC2 EXON 8 Polymorphisms: Rarity and lack of significance in risk of squamous cell carcinoma of the head and neck. Oral Oncol 38(5):475–477. Sun EC, Curtis R, Melbye M, Goedert JJ. 1999. Salivary gland cancer in the United States. Cancer Epidemiol Biomarkers Prev 8(12):1095–1100. Swan GE, Carmelli D, Cardon LR. 1997. Heavy consumption of cigarettes, alcohol and coffee in male twins. J Stud Alcohol 58(2):182–190. Swango PA. 1996. Cancers of the oral cavity and pharynx in the United States: An epidemiologic overview. J Public Health Dent 56(6):309–318. Swanson GM, Belle SH. 1982. Cancer morbidity among woodworkers in the U.S. automotive industry. J Occup Med 24(4):315–319. Swanson GM, Burns PB. 1997. Cancers of the salivary gland: workplace risks among women and men. Ann Epidemiol 7(6):369–374. Swerdlow AJ, Marmot MG, Grulich AE, Head J. 1995. Cancer mortality in Indian and British ethnic immigrants from the Indian subcontinent to England and Wales. Br J Cancer 72(5):1312–1319. Szpak CA, Stone MJ, Frenkel EP. 1977. Some observations concerning the demographic and geographic incidence of carcinoma of the lip and buccal cavity. Cancer 40(1):343–348. Takezaki T, Hirose K, Inoue M, et al. 1996. Tobacco, alcohol and dietary factors associated with the risk of oral cancer among Japanese. Japanese J Cancer Res 87(6):555–562. Takezaki T, Shinoda M, Hatooka S, et al. 2000. Subsite-specific risk factors for hypopharyngeal and esophageal cancer (Japan). Cancer Causes Control 11(7):597–608. Talamini R, Vaccarella S, Barbone F, et al. 2000. Oral hygiene, dentition, sexual habits and risk of oral cancer. Br J Cancer 83(9):1238–1242. Tanimoto K, Hayashi S, Yoshiga K, Ichikawa T. 1999. Polymorphisms of the CYP1A1 and GSTM1 gene involved in oral squamous cell carcinoma in association with a cigarette dose. Oral Oncol 35(2):191–196. Tassi E, Wellstein A. 2003. Angiogenesis, basic mechanisms, and role in head and neck squamous cell carcinoma. In: Head and Neck Cancer: Emerging Perspectives. JF Ensley, JS Gutkind, JR Jacobs, SM Lippman eds. New York: Academic Press, pp. 81–100. Teppo L, Pukkala E, Saxen E. 1985. Multiple cancer—an epidemiologic exercise in Finland. J Natl Cancer Inst 75(2):207–217. Tillonen J, Homann N, Rautio M, Jousimies-Somer H, Salaspuro M. 1999. Role of yeasts in the salivary acetaldehyde production from ethanol among risk groups for ethanol-associated oral cavity cancer. Alcohol Clin Exp Res 23(8):1409–1415. Tischkowitz MD, Hodgson SV. 2003. Fanconi anaemia. J Med Genet 40(1):1–10. Trivedy CR, Craig G, Warnakulasuriya S. 2002. The oral health consequences of chewing areca nut. Addict Biol 7(1):115–125.

695

U.S. DHHS. 2003. Those who continue to smoke. Is achieving abstinence harder and do we need to change our interventions? Smoking and Tobacco Control Monograph 15. Bethesda, MD: Public Health Service, National Institutes of Health, National Cancer Institute. U.S. Preventive Services Task Force. 1996. Screening for Oral Cancer. In: Guide to Clinical Preventive Services, 2nd ed: National Library of Medicine. Uzcudun AE, Retolaza IR, Fernandez PB, et al. 2002. Nutrition and pharyngeal cancer: Results from a case-control study in Spain. Head Neck 24(9):830–840. van Zandwijk N, Dalesio O, Pastorino U, de Vries N, van Tinteren H. 2000. EUROSCAN, a randomized trial of vitamin A and N-acetylcysteine in patients with head and neck cancer or lung cancer. For the European Organization for Research and Treatment of Cancer Head and Neck and Lung Cancer Cooperative Groups. J Natl Cancer Inst 92(12):977–986. Velly AM, Franco EL, Schlecht N, et al. 1998. Relationship between dental factors and risk of upper aerodigestive tract cancer. Oral Oncol 34(4):284–291. Waldron CA, Shafer WG. 1975. Leukoplakia revisited. A clinicopathologic study 3256 oral leukoplakias. Cancer 36(4):1386–1392. Wang LE, Sturgis EM, Eicher SA, Spitz MR, Hong WK, Wei Q. 1998. Mutagen sensitivity to benzo(a)pyrene diol epoxide and the risk of squamous cell carcinoma of the head and neck. Clin Cancer Res 4(7):1773–1778. Warnakulasuriya KA, Johnson NW, Linklater KM, Bell J. 1999. Cancer of mouth, pharynx and nasopharynx in Asian and Chinese immigrants resident in Thames regions. Oral Oncol 35(5):471–475. Warnakulasuriya S. 2002. Areca nut use following migration and its consequences. Addict Biol 7(1):127–132. Wasnik KS, Ughade SN, Zodpey SP, Ingole DL. 1998. Tobacco consumption practices and risk of oro-pharyngeal cancer: A case-control study in Central India. Southeast Asian J Trop Med Public Health 29(4):827– 834. Wassberg C, Thorn M, Yuen J, Ringborg U, Hakulinen T. 1999. Second primary cancers in patients with squamous cell carcinoma of the skin: A population-based study in Sweden. Int J Cancer 80(4):511–515. Waterhouse J, Muir C, Correa P, Powell J, eds. 1976. Cancer Incidence in Five Continents, Volume III. Lyon: IARC Scientific Publications No. 15. Waterhouse J, Muir C, Shanmugaratnam K, Powell J, eds. 1982. Cancer Incidence in Five Continents, Volume IV. Lyon, France: International Agency for Research on Cancer. IARC Scientific Publications No. 42. Weaver A, Fleming SM, Smith DB. 1979. Mouthwash and oral cancer: Carcinogen or coincidence? J Oral Surg 37(4):250–253. Wei Q, Eicher SA, Guan Y, et al. 1998. Reduced expression of hMLH1 and hGTBP/hMSH6: A risk factor for head and neck cancer. Cancer Epidemiol Biomarkers Prev 7(4):309–314. Weinstein SJ, Gridley G, Harty LC, et al. 2002. Folate intake, serum homocysteine and methylenetetrahydrofolate reductase (MTHFR) C677T genotype are not associated with oral cancer risk in Puerto Rico. J Nutr 132(4):762–767. Whitaker CJ, Moss E, Lee WR, Cunliffe S. 1979. Oral and pharyngeal cancer in the North-west and West Yorkshire regions of England, and occupation. Br J Industrial Med 36(4):292–298. Wight AJ, Ogden GR. 1998. Possible mechanisms by which alcohol may influence the development of oral cancer—a review. Oral Oncol 34(6):441–447. Wigle DT, Mao Y, Grace M. 1980. Relative importance of smoking as a risk factor for selected cancers. Can J Public Health 71(4):269–275. Williams RR, Horm JW. 1977. Association of cancer sites with tobacco and alcohol consumption and socioeconomic status of patients: Interview study from the Third National Cancer Survey. 58(3):525–547. Winn DM, Blot WJ, Shy CM, Pickle LW, Toledo A, Fraumeni JF, Jr. 1981. Snuff dipping and oral cancer among women in the southern United States. N Engl J Med 304(13):745–749. Winn DM, Blot WJ, Shy CM, Fraumeni JF, Jr. 1982. Occupation and oralcancer among women in the South. Am J Industrial Med 3(2):161–167. Winn DM, Blot WJ. 1985. Second cancer following cancers of the buccal cavity and pharynx in Connecticut, 1935–1982. Natl Cancer Inst Monogr 68:25–48. Winn DM, Blot WJ, McLaughlin JK, et al. 1991. Mouthwash use and oral conditions in the risk of oral and pharyngeal cancer. Cancer Res 51(11): 3044–3047. Winn DM, Diehl SR, Brown LM, et al. 2001. Mouthwash in the etiology of oral cancer in Puerto Rico. Cancer Causes Control 12(5):419–429. World Cancer Research Fund/American Institute for Cancer Research. 1997. Food, Nutrition and the Prevention of Cancer: A Global Perspective. Washington, D.C.: American Institute for Cancer Research. Wu X, Gu J, Hong WK, et al. 1998. Benzo[a]pyrene diol epoxide and bleomycin sensitivity and susceptibility to cancer of upper aerodigestive tract. J Natl Cancer Inst 90(18):1393–1399.

696

PART IV: CANCER BY TISSUE OF ORIGIN

Wu X, Lippman SM, Lee JJ, et al. 2002. Chromosome instability in lymphocytes: A potential indicator of predisposition to oral premalignant lesions. Cancer Res 62(10):2813–2818. Wynder EL, Bross IJ, Feldman RM. 1957. A study of the etiological factors in cancer of the mouth. Cancer 10(6):1300–1323. Wynder EL, Kabat G, Rosenberg S, Levenstein M. 1983. Oral cancer and mouthwash use. J Natl Cancer Inst 70(2):255–260. Yang YH, Lee HY, Tung S, Shieh TY. 2001. Epidemiological survey of oral submucous fibrosis and leukoplakia in aborigines of Taiwan. J Oral Pathol Med 30(4):213–219. Yu GP, Zhang ZF, Hsu TC, Spitz MR, Schantz SP. 1999. Family history of cancer, mutagen sensitivity, and increased risk of head and neck cancer. Cancer Lett 146(1):93–101. Zain RB, Ikeda N, Razak IA, et al. 1997. A national epidemiological survey of oral mucosal lesions in Malaysia. Commun Dent Oral Epidemiol 25(5):377–383. Zaridze DG, Blettner M, Trapeznikov NN, et al. 1985. Survey of a population with a high incidence of oral and oesophageal cancer. Int J Cancer 36(2):153–158. Zaridze DG, Safaev RD, Belitsky GA, Brunnemann KD, Hoffmann D. 1991. Carcinogenic substances in Soviet tobacco products. IARC Sci Publ 105:485–488. Zavras AI, Wu T, Laskaris G, et al. 2002. Interaction between a single nucleotide polymorphism in the alcohol dehydrogenase 3 gene, alcohol consumption and oral cancer risk. Int J Cancer 97(4):526–530. Zhang Z-F, Morgenstern H, Spitz MR, et al. 1999. Marijuana use and increased risk of squamous cell carcinoma of the head and neck. Cancer Epidemiol Biomarkers Prev 8(12):1071–1078.

Zhang Z-F, Morgenstern H, Spitz MR, et al. 2000. Environmental tobacco smoking, mutagen sensitivity, and head and neck squamous cell carcinoma. Cancer Epidemiol Biomarkers Prev 9(10):1043–1049. Zheng T, Boyle P, Willett WC, et al. 1993a. A case-control study of oral cancer in Beijing, People’s Republic of China. Associations with nutrient intakes, foods and food groups. Eur J Cancer B Oral Oncol 29B(1):45– 55. Zheng TZ, Boyle P, Hu HF, et al. 1990a. Dentition, oral hygiene, and risk of oral cancer: A case-control study in Beijing, People’s Republic of China. Cancer Causes Control 1(3):235–241. Zheng TZ, Boyle P, Hu HF, et al. 1990b. Tobacco smoking, alcohol consumption, and risk of oral cancer: A case-control study in Beijing, People’s Republic of China. Cancer Causes Control 1(2):173–179. Zheng W, Blot WJ, Diamond EL, et al. 1993b. Serum micronutrients and the subsequent risk of oral and pharyngeal cancer. Cancer Res 53(4):795– 798. Zheng W, Shu XO, Ji BT, Gao YT. 1996. Diet and other risk factors for cancer of the salivary glands: A population-based case-control study. Int J Cancer 67(2):194–198. Zheng Y, Li L, Shen H, Sturgis EM, et al. 2001a. Polymorphic hCHK2/hCds1 codon 84 allele and risk of squamous cell carcinoma of the head and neck—a case-control analysis. Carcinogenesis 22(12):2005–2008. Zheng Y, Shen H, Sturgis EM, et al. 2001b. Cyclin D1 polymorphism and risk for squamous cell carcinoma of the head and neck: A case-control study. Carcinogenesis 22(8):1195–1199.

36

Esophageal Cancer WILLIAM J. BLOT, JOSEPH K. MCLAUGHLIN, AND JOSEPH F. FRAUMENI, JR.

E

sophageal cancer exhibits epidemiologic patterns distinct from all other cancers. Most striking are the geographic variations in incidence of esophageal cancer around the world, and the temporal changes in histologic presentation whereby adenocarcinoma has surpassed squamous cell carcinoma as the leading cell type in most western societies. The known and suspected causes of esophageal cancer are described in this chapter, with emphasis on the differences in risk factors between adenocarcinoma and squamous cell carcinoma of the esophagus. Since treatment successes have been limited and esophageal cancers still are usually fatal, regardless of cell type, we note how the epidemiologic patterns may offer clues to prevention.

PATHOLOGY, ANATOMY, AND NATURAL HISTORY Cancers of the esophagus typically occur in one of two forms, squamous cell carcinomas arising from the stratified squamous epithelial lining of the organ, and adenocarcinomas affecting columnar glandular cells that replace the squamous epithelium. Rarely other histologic types occur, such as sarcomas and small cell carcinomas, but these generally represent less than 1%–2% of all esophageal cancers (Young et al., 1981; SEER, 2003). Squamous cell carcinomas most often occur in the middle third of the esophagus, followed by the lower third and the upper third, while the large majority of esophageal adenocarcinomas arise in the lower third. The natural histories of squamous cell carcinomas and adenocarcinomas of the esophagus appear to differ substantially. For squamous cell cancers, especially in high-risk populations, transition models have described squamous epithelium undergoing inflammatory changes that progress to dysplasia and in situ malignant change (Oettle et al., 1986; Jacob et al., 1993; Dawsey et al., 1994; Mandard et al., 2000). Most adenocarcinomas, however, tend to arise in the distal esophagus from columnar-lined metaplastic epithelium, commonly known as Barrett esophagus (Wild and Hardie, 2003), which replaces the squamous epithelium during the healing phase of reflux esophagitis and may progress to dysplasia. Keys to the causes and prevention of esophageal cancer are likely to come from a better understanding of these precursor lesions.

PATTERNS OF OCCURRENCE Geographic Variation The incidence rates for esophageal cancer vary internationally more than any other cancer, with over 50-fold differences between high-rate and low-rate areas (Table 36–1) (Parkin et al., 2003). In most parts of the world, esophageal cancer is relatively uncommon, ranking tenth in numbers of cases in the United States in 2003 (Parkin et al., 2003; Jemal et al., 2003). The annual incidence rates in most developed countries are less than 10 per 100,000, and often less than 5 per 100,000. In some developing regions, such as parts of north central China and Iran, it is the most common cancer, with squamous cell cancers of the esophagus accounting for nearly one-fourth of all deaths. These populations have incidence and mortality rates for esophageal cancer exceeding 100 cases per 100,000 population per year. In addition, pockets of elevated rates, although not as high, have been reported in South Africa, parts of France, and South America,

and even in the United States, notably among black men in coastal areas of South Carolina (Brown et al., 1988; Munoz and Day, 1996). In these high-incidence areas, the dominant cell type is squamous cell carcinoma.

Age, Sex, and Race Rates of esophageal cancer rise progressively with age, with an exponential increase until about age 60 and a slower increase thereafter (Fig. 36–1). The cancers rarely occur in children or young adults. In most parts of the world, esophageal cancer occurs more frequently among men than women, with the male-to-female ratio often surpassing threefold, as it does in the United States (Table 36–1, Fig. 36–1). The male excess is seen with adenocarcinoma and squamous cell carcinoma, except in those few areas of the world where the rates are extraordinarily high for squamous cell cancers in both sexes. Esophageal cancer also displays remarkable racial and ethnic differences in incidence and mortality according to tumor type. In the United States, rates of esophageal adenocarcinoma are higher among white than black men by a more than threefold margin, while the reverse is true for squamous cell carcinoma, with rates being more than five times higher among black than white men.

Time Trends In recent times the rates of esophageal cancer have changed dramatically in many western countries, with a rapid increase in the incidence of adenocarcinoma of the esophagus, while the incidence of squamous cell carcinoma has been relatively flat or even declining (Blot et al., 1991; Botterweck et al., 2000). Adenocarcinomas used to be relatively rare throughout the world, accounting for less than 15% of all esophageal cancers. This is still the situation in many countries, particularly in developing parts of the world, but adenocarcinomas now equal or exceed squamous cell cancers in a number of western countries. Table 36–1 shows the high proportion of squamous cell carcinomas reported in cancer registries outside North America and Europe during the 1990s. However, among white men in the United States, the proportion of adenocarcinomas rose to 34% by the mid-1980s (Blot et al., 1991), to nearly 60% by 1992–1994 (Devesa et al., 1998), to over 70% of all esophageal cancers by the year 2000 (Fig. 36–2). Rates of adenocarcinoma have increased also among women, although incidence is much lower. Among black men, adenocarcinoma incidence has also risen while rates for squamous cell carcinoma have declined, yet squamous cell tumors still outnumber adenocarcinomas by a 10 to 1 margin. Adenocarcinoma also has surpassed squamous cell carcinoma in the United Kingdom, which has the highest rates of adenocarcinoma reported worldwide (Parkin et al., 2003). The rising incidence of esophageal adenocarcinoma may be related in part to the way cancers are reported. Since early (pre-1970) coding rules called for tumors at the junction of the esophagus and stomach to be classified as stomach (cardia) cancers, changes in diagnostic or recording practices could account for some of the increases in esophageal cancer incidence rates. Arguing against this possibility, however, is the concomitant rise in the incidence of gastric cardia cancer, whereas a decline would have been expected had there been simply a shift in classification (Blot et al., 1991). In addition, improvements in endoscopic and imaging technology may have led to more precise histologic as well as anatomic classification, but the parallel trends for esophageal and gastric cardia adenocarcinomas suggest that

697

698

PART IV: CANCER BY TISSUE OF ORIGIN

Table 36–1. Esophageal Cancer Incidence in Selected Areas of the World

45

Age-Adjusted Cancer Incidence Rate per 100,000*

12.2 8.8

89 98

4.0 123.1 2.5 4.3 1.6 1.0

90 99 89 84 90 94

1.7 1.8 3.8 0.7 2.2 0.9 3.5

52 85 56 76 51 69 41

1.3 1.2 3.5 2.5

55 44 92 84

2.5 2.1

56 47

35 30 25 20 15

*Annual rate per 100,000 population adjusted to the world standard population; †Among cancers with specified cell type. Source: Data from Parkin et al., 2003.

10 5 0

85+

80–84

75–79

70–74

65–69

60–64

55–59

50–54

45–49

40–44

35–39

30–34

25–29

20–24

15–19

10–14

Age at diagnosis 05–09

Africa Uganda 13.2 Zimbabwe 19.3 Asia Beijing, China 10.2 Cixian, China 183.8 Hong Kong, China 11.7 Delhi, India 6.1 Osaka, Japan 10.0 Bangkok, Thailand 4.4 Europe Denmark 5.8 Calvados, France 17.2 Ireland 8.0 Florence, Italy 2.2 Netherlands 6.2 Sweden 3.1 UK, England 8.4 North, South, Central America Canada 4.2 US, SEER white 4.7 US, SEER black 10.7 Uruguay, Montevideo 10.7 Oceania Victoria, Australia 5.4 New Zealand 5.7

Female

00–04

Male

40

Percent Squamous Cell Carcinoma†

Rate per 100,000

Area

50

Male Female

Figure 36–1. Average annual esophageal cancer incidence (per 100,000) in the United States by age and gender, 1996–2000 (SEER, 2003).

to be higher among women than men and among whites than blacks, but do not differ substantially by cell type. Outside of North America, 5-year survival rates below 10% are typically reported (Nyren and Adami, 2002). The poor survival from esophageal cancer is related to the late stages at which the cancers are often diagnosed. In the United States, for example, only about one-third of the tumors are localized at initial diagnosis, whereas two-thirds have regional or distant spread (SEER, 2003).

RISK FACTORS the increasing incidence of both tumors is real and reflects shared causal factors.

Survival Times

Rate Per 100,000 Persons-Years

Rates of esophageal cancer mortality and incidence tend to be similar, indicating that survival from esophageal cancer is low. In the United States (SEER, 2003), 5-year relative survival in the 1990s averaged slightly above 10%, only a modest absolute improvement over survival rates that averaged slightly below 10% in the 1960s and 1970s (Axtell et al., 1976). Relative survival rates of esophageal cancer tend

The markedly different patterns of occurrence for adenocarcinoma vs. squamous cell carcinoma indicate that the two forms of esophageal cancer have distinct etiologic profiles. Prior to the advent of the adenocarcinoma epidemic, the dominant causes of esophageal cancer (i.e., squamous cell carcinomas) were tobacco smoking and alcohol consumption. The recent changes in histologic distribution, however, signal a shift in etiologic factors. For example, while tobacco smoking is a risk factor for both cell types, the link with alcohol intake is primarily for squamous cell cancers.

Figure 36–2. Trends in age-adjusted incidence rates of esophageal cancer by cell type among white and black males in the United States. (Source: Data from SEER 2003.)

Squamous Cell Carcinoma Black Males Adenocarcinoma White Males

Squamous Cell Carcinoma White Males

Adenocarcinoma Black Males

Year

699

Esophageal Cancer Table 36–2. Relative Risks of Esophageal Cancer (Predominantly Squamous Cell Carcinoma) Associated with Alcohol Drinking and Tobacco Smoking in Brittany, France

Table 36–3. Relative Risks of Esophageal Cancer, by Cell Type, Among Former and Current Cigarette Smokers* Smoking Status

Squamous Cell Carcinoma

Adenocarcinoma

1.0† 2.8 (1.5–4.9)‡ 5.1 (2.8–9.2)

1.0† 2.0 (1.4–2.9)‡ 2.2 (1.4–2.3)

Tobacco (g/day) Alcohol (g/day)

0–9

10–19

20–29

≥30

0–40 41–80 81–120 ≥121

1.0* 7.3 11.8 49.6

3.4 8.4 13.6 65.9

3.2 8.8 12.6 137.6

7.8 35.0 83.0 155.6

*Reference category. Source: Munoz and Day (1996). Adapted from Tuyns et al. (1977).

Tobacco It has long been known from cohort and case-control studies that the major risk factors for squamous cell carcinoma of the esophagus in most countries are cigarette smoking and alcoholic beverage consumption (IARC, 1986; IARC, 1988). In cohort studies, the risk of esophageal cancer has been about five times higher among cigarette smokers than nonsmokers, with the increase reaching 10-fold among heavy smokers. The cell types of esophageal cancer were typically not distinguished in cohort studies, but it is likely that squamous cell carcinomas predominated. Although smokers tend to be heavier users of alcohol, adjustment for alcohol intake did not eliminate the effect of smoking and excess risks were found among smokers who did not consume alcohol (Tuyns, 1983; La Vecchia and Negri, 1989; Cheng et al., 1995a; Tavani et al., 1996; Castellsague et al., 1999; Znaor et al., 2003). In addition, the risk of esophageal cancer has been consistently shown to increase with increasing amounts smoked, and to decrease following smoking cessation (IARC, 1986, 1988). The effects of tobacco are greatly enhanced by alcohol consumption, and vice versa. As shown in Table 36–2, the combined effects of smoking and drinking on squamous cell carcinomas of the esophagus are nearly multiplicative, based on data from a case-control study in France (Tuyns et al., 1977; Tuyns, 1983). While the relative risks were higher than found in most other studies, the dose-response gradient with increasing levels of tobacco and alcohol intake is typical. While some reports have related smokeless tobacco and other chews, including betel nut, to squamous cell esophageal cancer, especially in certain Asian countries (Znaor et al., 2003), the use of snuff or chewing tobacco is not a major risk factor (in contrast to oral cavity cancer) in the United States or western Europe (Surgeon General, 1986; Lagergren et al., 2000a). However, all types of smoked tobacco have been implicated in esophageal cancer risk. It has been suggested that the more frequent use of mentholated cigarettes in the black population of the United States contributes to their higher rates of esophageal cancer, but studies assessing the risks from menthol vs. non-menthol cigarettes have been unremarkable (Hebert and Kabat, 1989). In an analysis of data from five case-control studies of squamous cell esophageal cancer in South America, smoking of black tobaccos was associated with a twofold risk compared with blond tobaccos (Castellsague et al., 1999). A differential effect between dark and light tobaccos has been observed also in France (Launoy et al., 2000). Data on cigar and pipe smoking are limited, but the available evidence suggests that their effects on the risk of squamous cancers of the esophagus resemble those associated with cigarette smoking (National Cancer Institute, 1998). Although tobacco smoking is an established risk factor for esophageal adenocarcinoma, the relative increases in risk are not nearly as high as for squamous cell carcinoma. Table 36–3 presents data from a multicenter investigation in three areas of the United States (Gammon et al., 1997), one of the largest studies to distinguish risk factors for esophageal adenocarcinoma vs. squamous cell carcinoma. The risk for esophageal adenocarcinoma was twice as high among smokers compared with nonsmokers, whereas the risk for squamous cell tumors was five times higher. Similar findings have been reported

Nonsmokers Former Current

*Relative risks adjusted for age, sex, race, income, alcohol. †Reference category; ‡95% confidence interval. Source: Gammon et al., 1997.

in other studies (Kabat et al., 1993; Brown et al., 1994; Vaughan et al., 1995; Zhang et al., 1996; Lagergren et al., 2000a; Wu et al., 2001). Noteworthy in the multicenter study of esophageal adenocarcinoma was the persistent elevation of risk up to 30 years after smoking cessation. Similar long-lasting effects of smoking were seen in a large case-control study in Los Angeles (Wu et al., 2001). In contrast, the risk of squamous cell cancer was found to decline within a decade of smoking cessation (Gammon et al., 1997). These findings suggest that smoking exerts an early-stage effect in the initiation of adenocarcinoma, while it acts during later stages as a promoter in the development of squamous cell carcinoma. It is noteworthy that the rising incidence of esophageal adenocarcinoma in US males over the past two to three decades has occurred in the face of a declining prevalence of smoking among men, perhaps because the risk for ex-smokers does not drop until 30 or more years after smoking cessation. In contrast, the slight downward trends in rates for squamous cell carcinoma of the esophagus are consistent with the decrease in smoking prevalence in recent decades.

Alcohol As illustrated in Table 36–4, consumption of alcoholic beverages is a potent risk factor for squamous cell carcinoma of the esophagus, with a clear dose-response relationship (IARC, 1988; Blot, 1992; Gammon et al., 1997; Launoy et al., 1997; Castellsague et al., 1999). Among heavy drinkers, the excess risks are as high as 50-fold in some populations such as in northern France (Tuyns et al., 1977). As shown in Table 36–2, the effects of alcohol are stronger than tobacco, but the combined effects are nearly multiplicative, with the risk among heavy drinkers and heavy smokers being 100 times that of abstainers or very light consumers of alcohol and tobacco in northern France (IARC, 1988; Zambon et al., 2000). An effect of alcohol is apparent even among nonsmokers (IARC, 1988; La Vecchia and Negri, 1989; Blot, 1992; Cheng et al., 1995a; Castellsague et al., 1999), and cessation of drinking is associated with a fairly rapid decline in risk (Cheng et al., 1995b). The higher levels of alcohol consumption appear to contribute to the higher rates of squamous cell cancers among black men in the United States, but the greater risks at each alcohol consumption level among black men than white men suggest that susceptibility factors may be involved as well (Brown et al., 2001). A special feature of squamous cell esophageal cancer is the clustering of elevated rates associated with specific alcoholic beverages

Table 36–4. Relative Risks of Esophageal Cancer, by Cell Type, According to Alcoholic Beverage Intake* Alcohol intake (drinks/week) None <5 5–11 12–30 >30

Squamous Cell Carcinoma †

1.0 0.8 (0.4–1.6)‡ 1.8 (0.9–3.5) 2.9 (1.5–5.4) 7.4 (4.0–13.7)

Adenocarcinoma 1.0† 0.7 (0.4–1.0)‡ 0.6 (0.4–0.9) 0.7 (0.4–1.1) 0.9 (0.5–1.4)

*Relative risks adjusted for age, sex, race, income, smoking; †Reference category; ‡95% confidence interval. Source: Gammon et al., 1997.

700

PART IV: CANCER BY TISSUE OF ORIGIN

preferentially consumed in certain populations, such as apple brandies in northern France, maize beer in the South African Transkei, rum and sugar cane distilled beverages in Puerto Rico and South America, and moonshine whiskeys in coastal South Carolina (Blot, 1994). These findings implicate an ingredient common to all beverages, ethanol, although it is possible that other components or contaminants may enhance cancer risk. In the United States, squamous cell cancers of the esophagus have been linked to all forms of alcoholic beverages, but the risks are greatest for drinkers of hard liquor, consistent with evidence that the concentration of ethanol plays an important role in alcohol-related tumors of the upper aerodigestive tract (Blot, 1992; Huang et al., 2003). Since ethanol does not induce cancer in laboratory animals, it may act through its major metabolite, acetaldehyde, a known carcinogen in animal models. This enzymatic process appears to be facilitated by bacterial flora and alcohol-metabolizing genes, but alcohol may also exert a promoting effect by solubilizing tobaccospecific carcinogens or enhancing their penetration into the esophageal mucosa, by nutritional deficiencies associated with heavy drinking, or by other mechanisms. In contrast, alcohol’s effect on esophageal adenocarcinoma is much less certain. In the US multicenter study, there was no increase in risk of adenocarcinoma associated with alcohol consumption (Table 36–4), consistent with several other investigations (Kabat et al., 1993; Brown et al., 1994; Vaughan et al., 1995; Zhang et al., 1996; Lagergren et al., 2000a). In a few studies, however, slight excesses of adenocarcinoma have been described among drinkers. It is noteworthy that in areas of the world with the highest esophageal cancer rates, notably in China and Iran, alcohol and tobacco appear to play only minor roles despite the fact that nearly all the cancers are squamous cell in origin (Munoz and Day, 1996).

Diet and Nutrition Despite the dominant effects of alcohol and tobacco, dietary factors have long been suspected to play a causal role in squamous cancers of the esophagus. In the 1950s, an excess occurrence was noted among women in northern Sweden with Plummer-Vinson syndrome, a condition associated with deficiencies of iron and other micronutrients (Larsson et al., 1975). Subsequently, an increased risk was found in celiac disease (Holmes et al., 1976), a malabsorption disorder of the small intestine leading to deficiencies of various nutrients, and in pernicious anemia, an established risk factor for stomach cancer (Ye and Nyren, 2003). In areas of China and the Caspian littoral of Iran with extremely high rates of esophageal cancer, the populations are characterized as having poor nutritional status with a diet of corn, wheat, or maize as the dietary staple (Munoz and Day, 1996). Since the prevalence of tobacco smoking and alcohol consumption is not very high in these areas, attention has focused on the role of indigenous dietary factors (Blot et al., 1993; Cheng and Day, 1996). In western countries, epidemiologic studies have consistently related dietary insufficiencies to squamous cell tumors of the esophagus, but emerging data from case-control studies indicate that similar dietary factors are associated with esophageal adenocarcinomas (Steinmetz and Potter, 1991; Kabat et al., 1993; Brown et al., 1995; Cheng and Day, 1996; Tzonou et al., 1996; Castellsague et al., 2000; Mayne and Navarro, 2002; Nyren and Adami, 2002). In both tumor types, a protective effect of fruits and vegetables has been demonstrated, with up to twofold increases in risk associated with low intake of these food groups after adjusting for smoking and alcohol consumption. The protective ingredients in fruits and vegetables are unclear, but some studies have implicated antioxidants, other micronutrients, and fiber (Terry et al., 2000b; Mayne et al., 2001). The inverse relation between squamous cell cancers of the esophagus and intake of fruits and vegetables has prompted intervention trials in areas with exceptionally high incidence rates to see whether supplements of micronutrients might inhibit tumor development. One of the largest trials ever conducted involved administration of specific combinations of vitamins and minerals in Linxian, China (Blot et al., 1993; Li et al., 1993). The Linxian general population trial involved nearly 30,000 participants randomized to receive one of four daily vitamin/mineral supplement combinations. After 5 years of supple-

mentation, death rates from cancer overall were significantly lower (by 13%) among those who received a combination of beta-carotene, vitamin E, and selenium, with the reduction being more pronounced for stomach than esophageal cancer (Blot et al., 1993). Subanalyses indicated that subjects with high pre-trial levels of selenium and alphatocopherol, but not beta-carotene, had lower esophageal cancer mortality (Mark et al., 2000; Taylor et al., 2003; Abnet et al., 2003). No significantly reduced cancer occurrence was seen for the other three combinations evaluated in the trial (retinol plus zinc, riboflavin plus niacin, and vitamin C plus molybdenum), although the incidence of esophageal cancer was somewhat lower among those receiving the B vitamins, riboflavin and niacin. In a concomitant randomized trial in the same area involving nearly 3000 persons with esophageal dysplasia, an immediate precursor to esophageal malignancy, there was a slight reduction in total and cancer mortality among those who received a supplement containing multiple vitamins and minerals. On repeat endoscopy, this group was found to have a higher rate of regression of dysplasia (Mark et al., 1994). However, in a nearby county in China, weekly supplements of retinol, zinc, and riboflavin in the general population did not result in a lowered prevalence of esophagitis (Munoz et al., 1985). In Uzbekistan, another area with high rates of esophageal cancer, a 20-month randomized trial yielded a nonsignificant reduction in esophagitis among persons receiving a combination of beta-carotene, retinol and vitamin E, with no effect of riboflavin (Zaridze et al., 1993). The generalizability of these findings from high-risk populations in Asia to the relatively well-nourished, low-risk populations in the United States or western Europe is uncertain, but the epidemiologic findings taken together suggest that steps to improve nutritional status may help lower the incidence of squamous cell esophageal cancer around the world. In a US trial to prevent recurrences among 1300 basal or squamous cell skin cancer patients (Clark et al., 1996), daily supplementation with 200 mg of selenium over an average period of 4.5 years was found to lower the incidence of esophageal cancer in the treated group, but the number of cases was small and confidence limits wide. The possible role of selenium in esophageal adenocarcinoma was suggested by a recent study linking low levels of serum selenium to the risk of Barrett esophagus and dysplasia (Rudolph et al., 2003). Early studies suggested that tea drinking, especially herbal teas with high content of tannins, might contribute to the high rates of esophageal cancer in certain populations, such as in Japan, parts of South America, Curacao, and among blacks in coastal South Carolina (Blot, 1994). More recent data, however, suggest that tea drinking per se may be protective, while the consumption of tea and other beverages consumed at exceptionally high temperatures may be carcinogenic (Chow et al., 1999). Most striking are data from Shanghai, China, where heavy tea drinking was associated with over 40% reductions in esophageal cancer risk (Gao et al., 1994). The effect was most pronounced for green and oolong teas, perhaps due to high levels of flavonoids, isothiocyanates, phenols, and other compounds shown to inhibit esophageal cancer in laboratory animals (Lambert and Yang, 2003). The relationship between drinking hot beverages and squamous cell cancers of the esophagus is best seen in studies from South America, where twofold to fourfold increases in risk have been documented among drinkers of maté and other beverages consumed at high temperatures, whereas tea drinking was associated with a significantly reduced risk (Castellsague et al., 2000). Other aspects of diet may also influence esophageal cancer risk, such as cooking meats at high temperatures that result in the formation of heterocyclic amines, which are carcinogenic in experimental systems (Sinha and Rothman, 1999). An excess risk of squamous cell tumors of the esophagus has been associated with higher levels of meat intake in Sweden (Terry et al., 2003) and Uruguay (De Stefani et al., 1998), and with barbecuing of meat in Nebraska (Ward et al., 1997). However, other studies have shown no clear relationship to meat intake (Nyren and Adami, 2002). Despite suggestions that foods that relax the lower esophageal sphincter may increase reflux and the subsequent development of esophageal adenocarcinoma, no such relationships have been detected in epidemiologic studies of this tumor (Terry et al., 2000a).

Esophageal Cancer Table 36–5. Relative Risks of Esophageal Cancer, by Cell Type, Associated with Increases in Body Mass Index* BMI quartile I (low) II III IV

Squamous Cell Carcinoma

Adenocarcinoma

1.0 0.5 (0.3–0.9) 0.8 (0.5–1.3) 0.6 (0.3–1.0)

1.0 1.3 (0.8–2.2) 2.0 (1.3–3.3) 2.9 (1.8–4.7)

BMI, body mass index. *Relative risks adjusted for age, sex, race, smoking. Source: Chow et al., 1998b.

Obesity Rising caloric intake without corresponding increases in energy expenditure has resulted in a steadily increasing prevalence of obesity in the United States and other western countries (Kuczmarski et al., 1994; Li and Mobarhan, 2000). Indeed, at the start of the 21st century, over 50% of American adults were considered to be overweight for their height, with 21% considered obese, where obesity is defined as a body mass index (BMI, weight in kg divided by height in m squared) of 30 or above (Mokdad et al., 2003). In epidemiologic studies of esophageal adenocarcinoma, elevated BMI has been consistently shown to be a significant risk factor (Brown et al., 1995; Vaughan et al., 1995; Chow et al., 1998b; Lagergren et al., 1999b; Chang et al., 2000; Li and Mobarhan, 2000; Calle et al., 2003). In the US multicenter study (Chow et al., 1998b), the relative risks of esophageal adenocarcinoma rose steadily with increasing BMI, reaching a threefold excess among those in the heaviest quartile (Table 36–5). In Sweden, a greater than 10-fold differential in cancer risk was observed between high and low BMI (Lagergren et al., 1999b). In contrast, studies of squamous cell cancer of the esophagus have consistently shown an inverse relationship with BMI (Chow et al., 1998b). While these data are derived from case-control studies that could be affected by biased recall, similar results have emerged from a prospective follow-up of more than 900,000 US adults, whose rates of death from esophageal cancer rose with increasing BMI in both men and women (Calle et al., 2003). It is likely that the increases in obesity prevalence in the population have contributed to the parallel trends in esophageal adenocarcinoma. The precise mechanisms are unclear, but central obesity may increase intra-abdominal pressure to promote gastroesophageal reflux disease (GERD) and its transition to Barrett esophagus. Other mechanisms are probably involved as well, since adjustment for symptomatic reflux disease has not greatly reduced the magnitude of the association between obesity and adenocarcinoma risk (Chow et al., 1998b; Wilson et al., 1999; Lagergren et al., 2000b; Nilsson et al., 2003). On the other hand, the inverse association between BMI and the risk of squamous cell esophageal cancer is probably related to residual confounding from alcohol consumption and poor nutritional status.

Medical Conditions and Medications Nutritional deficiencies probably account for the elevated risk of squamous esophageal cancers associated with Plummer-Vinson syndrome, celiac disease, and pernicious anemia. Susceptibility to these tumors also occurs with achalasia, in which the gastroesophageal sphincter fails to relax due to disturbance of the autonomic nervous system (Sandler et al., 1995). While only a small proportion of squamous cell tumors are associated with preexisting medical conditions other than the immediate precursor, esophageal dysplasia, the majority of esophageal adenocarcinoma cases are thought to arise from Barrett epithelium (Wild and Hardie, 2003; McManus et al., 2004). Although the determinants of Barrett esophagus are not clearly defined, this condition shares demographic characteristics and risk factors with esophageal adenocarcinoma (Conio et al., 2002; Shaheen and Ransohoff, 2002; Wild and Hardie, 2003). Most remarkable are the upward trends reported in the frequency of Barrett esophagus, which are partly due to the rise in

701

endoscopic procedures in recent years, but also result from increases in the prevalence of GERD (Conio et al., 2001). Although Barrett esophagus is a common precursor to esophageal adenocarcinoma, the probability of malignant transformation is not well established. In a nationwide survey of patients undergoing esophageal biopsies in Northern Ireland from 1993–1999, nearly 3000 persons with Barrett esophagus were identified (Murray et al., 2003). In an average follow-up of 4 years, the annual incidence of esophageal adenocarcinoma was 0.3% among Barrett patients overall, but nearly 5% among those with associated severe dysplasia and 1% with mild or moderate dysplasia. The consensus estimate of the annual transformation rate from Barrett esophagus to esophageal adenocarcinoma averages around 0.5% (Enzinger and Mayer, 2003). Hence the lifetime risk of developing esophageal adenocarcinoma in a middle-aged person diagnosed with Barrett esophagus would be about 10%–15%. Most cases of Barrett esophagus arise in the context of long-standing GERD (Wild and Hardie, 2003), caused by esophageal reflux of gastric acid and probably also bile salts and alkaline duodenal contents (Campos et al., 2001). Epidemiologic studies have consistently reported increases in the risk of esophageal adenocarcinoma associated with a diagnosis of symptomatic GERD and hiatal hernia (Chow et al., 1995; Lagergren et al., 1999a; Farrow et al., 2000; Ye et al., 2001b; Avidan et al., 2002; Wu et al., 2003b). The central role of GERD in esophageal adenocarcinoma is underscored by a Swedish study reporting an overall eightfold increased risk rising to 40-fold among individuals with the most severe and long-lasting symptoms (Lagergren et al., 1999a). Similarly, an eightfold increased risk was observed in California among those with both GERD and hiatal hernia (Wu et al., 2003b), and a fivefold increase among those with daily reflux symptoms in the US multicenter study (Farrow et al., 2000). In the multicenter study, about 30% of the cases were related to symptomatic GERD, a percentage somewhat less than attributed to smoking (40%) or high BMI (40%), but ascertainment of GERD was likely to be incomplete in that study. Because of the increasing use of medications to manage GERD by lowering stomach acidity, hypotheses have emerged that H2 receptor antagonists, proton-pump inhibitors, and even over-the-counter antacids may increase the risk of esophageal adenocarcinoma. Also under suspicion are those drugs that relax the gastroesophageal sphincter and may promote reflux, such as tricyclic antidepressants, calcium channel blockers, asthma drugs containing theophylline or ß-agonists, and anticholinergics. However, the available evidence has pointed to the effects of underlying conditions rather than the medications used. For example, in a nested case-control study in a US prepaid health plan, a fourfold elevated risk of esophageal adenocarcinoma risk among users of H2 blockers was nearly eliminated after adjusting for GERD (Chow et al., 1995). Although an association with anti-asthmatic drugs has been reported (Vaughan et al., 1998), it is more likely a result of the underlying condition, asthma, which appears to predispose to esophageal adenocarcinoma through its interrelationship with GERD (Ye et al., 2001a). There is special interest in medications that may be protective against esophageal cancer, such as the nonsteroidal anti-inflammatory drugs (NSAIDs). In the US multicenter study, there were 50%–60% reductions in the risk of both types of esophageal cancer among current users of aspirin and other NSAIDs (Farrow et al., 1998). The reductions in adenocarcinoma, but not squamous tumors, appeared restricted to patients with over-expression of the cell cycle control gene, cyclin D1 (Gammon et al., 2004). These findings have raised hopes that NSAIDs, including selective COX-2 inhibitors, may protect against esophageal cancer among patients diagnosed with Barrett esophagus (DuBois, 2002; Gupta and DuBois, 2002; Yu et al., 2003).

Infectious Agents Most studies of infectious agents in squamous cell esophageal cancer have focused on human papillomaviruses (HPV). Although over 100 HPV subtypes exist, elevated prevalences of the oncogenic types HPV 16 and 18 have been detected in some but not all studies in endemic areas of China (Hille et al., 1985; Chang et al., 2000; Peixoto et al., 2001; Zhou et al., 2003). In western countries the association has been

702

PART IV: CANCER BY TISSUE OF ORIGIN

less evident. A cohort study in Norway suggested a relationship of HPV 16 serology to esophageal cancer (Bjorge et al., 1997), while a case-control study in Sweden showed no association between either type of esophageal cancer and HPV 16 or 18 serology (Lagergren et al., 1999c). While the role of HPV infection remains to be confirmed in squamous cell tumors of the esophagus, it seems clear that HPV is found less commonly than in oral cancer (Herrero et al., 2003; Smith et al., 2004). For esophageal adenocarcinoma, recent interest has centered on the possible protective effect of Helicobacter pylori infection, the main cause of gastric cancer around the world. Several case-control studies have reported reduced risk among those serologically positive for H. pylori infection, particularly the Cag A subtype (Chow et al., 1998a; Vieth et al., 2000; Henrik et al., 2001), although not all studies have found this effect (Wu et al., 2003a). An inverse relation with H. pylori infection has also been reported for GERD and Barrett esophagus (Blaser, 1999; Cremonini et al., 2003; Graham, 2003). The mechanism by which H. pylori might protect against esophageal adenocarcinoma and its precursor states is not clear; colonization of the gastric mucosa may reduce the amount of gastric acid that is refluxed into the esophagus, although a recent study in Sweden did not find an excess risk associated with gastric atrophy (Ye et al., 2004). Whatever the mechanism, it is possible that the rising incidence of esophageal adenocarcinoma may be related in part to the declining prevalence of H. pylori infection due to improvements in sanitation and the widespread use of antibiotics (Blaser, 1999).

Socioeconomic Status and Occupation Low socioeconomic status has been related to an increased risk of squamous cell carcinoma of the esophagus, even after adjustment for known risk factors, which may reflect as yet unrecognized exposures (Brown et al., 2001). In contrast, socioeconomic status has not been consistently linked to esophageal adenocarcinoma (Brown et al., 1994; Gammon et al., 1997). Although a number of reports have suggested excess risks of esophageal cancer among various occupational groups (Nyren and Adami, 2002), the associations are neither consistent nor strong, nor adjusted for social class. Most have derived from exploratory record-linkage studies assessing multiple occupations to generate hypotheses, or from cohort studies assessing multiple outcomes, so that the possibility of chance associations looms large. Perhaps the strongest occupational association reported has been the nearly twofold excess of squamous cell esophageal cancer among insulation workers heavily exposed to asbestos (Selikoff et al., 1979), but a meta-analysis showed little or no increase in esophageal cancer across 17 groups of asbestos-exposed workers (the summary relative risk was 1.02) (Goodman et al., 1999). Excesses of esophageal cancer have been reported with chlorinated solvent exposures such as perchloroethylene (PCE) (Weiss, 1995) and trichloroethylene (TCE) (Hansen et al., 2001; Raaschou-Nielsen et al., 2003), and with combustion products (Norell et al., 1983; Gustavsson et al., 1993; Gustavsson et al., 1998), but causal relationships have not been established. Few studies have examined occupational associations with esophageal adenocarcinoma separately from squamous cell cancers. The US multicenter study found slightly elevated risks for esophageal adenocarcinoma among several white collar occupations, but no evidence suggesting workplace hazards (Engel et al., 2002). In a systematic review of occupational carcinogens recognized by the International Agency for Research on Cancer, none were related to esophageal cancer (Boffetta et al., 1995).

Radiation Esophageal cancer is among the long list of malignancies that can be induced by ionizing radiation. About twice as many squamous cell esophageal cancers as expected have occurred among atom bomb survivors in Japan exposed to 200 rad or more as compared with less than 10 rad (RERF, 1981). A similar excess of squamous cell esophageal cancers has been observed among patients who received X-ray therapy for ankylosing spondylitis (Smith, 1984). Both cell types of esophageal cancer have been reported in excess among women given radiotherapy for breast cancer (Ahsan and Neugut, 1998).

Genetic Susceptibility For both types of esophageal cancer, clinical reports of familial aggregation are uncommon, with only limited epidemiologic evidence of a familial tendency in western countries (Tavani et al., 1994). In the US multicenter study, there was no increased risk of either cell type of esophageal cancer among persons with a family history of digestive cancers, but a link was suggested between esophageal adenocarcinoma and familial occurrence of breast cancer (Dhillon et al., 2001). In high-incidence areas of China with squamous cell cancers of the esophagus, the disease is so common that most extended families have multiple cases. Despite difficulties in distinguishing genetic susceptibility from shared environmental exposures, an autosomal recessive trait has been suggested by segregation analysis of affected families (Carter et al., 1992; Zhang et al., 2000). Familial tendencies have been noted also in high-risk areas of Iran (Munoz and Day, 1996) and Japan (Morita et al., 1998). A rare genetic syndrome predisposing to squamous cell cancer of the esophagus is tylosis, featuring hyperkeratoses of the palms and soles, an autosomal recessive trait with a genetic locus identified in a small region of chromosome 17q25 (Ellis et al., 1994; Iwaya et al., 1998; Risk et al., 1999). This region is often deleted in sporadic esophageal cancer, providing clues to a potential tumor suppressor gene (Risk et al., 2002). Inherited susceptibility has been suggested also in a fraction of cases with GERD and Barrett esophagus, based on reports of familial clustering. A study of twins indicated that approximately 30% of GERD cases may be due to inherited susceptibility (Cameron et al., 2002), and an autosomal dominant pattern has been reported in a subset of Barrett esophagus (Romero et al., 2002; Drovdlic et al., 2003). Risk of esophageal cancer may be influenced by individual variation in the body’s response to tobacco, alcohol, or other exposures. Polymorphisms in genes involved in the metabolic activation of compounds in tobacco or alcoholic beverages to carcinogenic forms, or in their detoxification, may contribute to risk at the individual or population level. Mixed results have emerged from assessment of genetic polymorphisms for alcohol dehydrogenase and CYP2E1, enzymes involved in the metabolism of ethanol to acetaldehyde (Brennan et al., 2004). However, polymorphisms of the acetaldehyde dehydrogenase gene that blocks the metabolism of acetaldehyde to acetate has been implicated in alcohol-related esophageal cancer in Japan, providing supporting evidence that acetaldehyde is a carcinogenic metabolite of ethanol (Yokoyama et al., 2002). The relationship between esophageal cancer and polymorphisms in genes for CYP1A1, GSTM1, NAT1, and NAT2, enzymes involved in tobacco metabolism, is unclear (Nyren and Adami, 2002). Somatic genetic changes, particularly p53 mutations, are common in esophageal cancer, especially for adenocarcinomas where these mutations are reported in up to 90% of all tumors (Carter et al., 1992; Eng et al., 1993; Rosen, 1994; Mandard et al., 2000; Lam, 2000). Wide variation has been reported in the frequency of other abnormalities, including absence of expression of p21 and mutations in p15, p16, and Rb and other tumor suppressor genes, among patients with esophageal cancer, particularly squamous cell tumors (Lam, 2000; Taniere et al., 2001; Sepehr et al., 2001). Since both cell types of esophageal cancer are the end result of multi-stage processes, recent work has been directed at detecting molecular changes associated with precursor lesions and their progression (Reid et al., 2003). Changes affecting numerous pathways, including cell cycle, signaling, chromosomal loss or gain, hypermethylation, and apoptosis are under study (Wild and Hardie, 2003). The growing availability of genomic, proteomic, and other molecular probes will enrich the prospects for future research into the causes and prevention of esophageal cancer, but strategies will need to be developed to process and analyze the flood of data that may generate false as well as promising leads.

PREVENTIVE MEASURES The keys to prevention of esophageal cancer vary by cell type. For squamous cell cancers, reduction or elimination of tobacco and alcohol

Esophageal Cancer consumption provide the best means to reduce the incidence of this cancer in western societies. The rapid reduction in risk of squamous cell carcinoma following cessation of smoking indicates that rates will fall within a decade of successful smoking reduction programs, as is being observed currently in the United States. It can be anticipated that the downward incidence of squamous cell carcinomas will persist as smoking prevalences continue to decline. In most parts of the world the excess risk associated with alcohol intake is primarily among heavy continuous drinkers, so that moderating consumption will aid in prevention. However, in those areas with extraordinary incidence rates unrelated to alcohol or tobacco intake, such as parts of China, further study is needed to identify causal factors. The alarming increase in the incidence of esophageal adenocarcinoma in western countries appears related to increases in the prevalence of obesity and GERD. The lack of an appreciable decline in risk among ex-smokers suggests that smoking cessation will have no immediate impact, although never starting to smoke would reduce risk by half compared with smokers. Hence, it will be important to target antismoking campaigns to children and young adults. Although no trials have demonstrated that weight reduction will lower esophageal adenocarcinoma risk, a beneficial effect seems very likely. Similarly, no clinical trials have shown that treating GERD will reduce the risk of progression to adenocarcinoma, but again it seems likely that vigorous medical treatment would be helpful. The optimal forms of such preventive measures remain to be clarified, as do strategies aimed at the detection and management of Barrett esophagus. It also remains to be seen whether chemoprevention with selenium and other micronutrients, or with NSAIDs including selective COX-2 inhibitors, will reduce the incidence of esophageal cancer in high-risk populations. Since dietary factors affect the risk of both types of esophageal cancer, the most prudent recommendation is to maintain adequate intake of fruits and vegetables as part of a balanced diet with multiple nutrients, including vitamin E and selenium. Whether tea intake at non-burning temperatures may lower risk is still an open question, but the protective effects suggested in some epidemiologic and experimental studies deserve further evaluation. Since both types of esophageal cancers generally arise from a multistage sequence of cellular events, detection of precursor lesions and strategies for medical surveillance should permit earlier diagnosis and treatment. It will be important to evaluate the impact of periodic endoscopic screening of Barrett esophagus patients on mortality from esophageal adenocarcinoma (Conio et al., 2003), although only about 5% of adenocarcinoma patients have been diagnosed previously with Barrett esophagus (Dulai et al., 2002). Further work is needed to develop optimal preventive, diagnostic, and therapeutic measures targeted to specific cellular and molecular stages in the carcinogenic process. In summary, despite uncertainties in our understanding of the causes and mechanistic pathways of esophageal cancer, there is sufficient evidence to take effective steps to prevent the majority of squamous cell carcinomas in western countries, while more information is needed to curb the epidemic increase in adenocarcinoma. By integrating epidemiologic, clinical, and molecular approaches in the study of high-risk populations and precursor states (e.g., Barrett esophagus), it should be possible to derive new insights into the carcinogenic process and thus expand the repertoire of preventive strategies. References Abnet CC, Qiao YL, Dawsey SM, et al. 2003. Prospective study of serum retinol, beta-carotene, beta-cryptoxanthin, and lutein/zeaxanthin and esophageal and gastric cancers in China. Cancer Causes Control 14: 645–655. Ahsan H, Neugut AI. 1998. Radiation therapy for breast cancer and increased risk for esophageal carcinoma. Ann Intern Med 128:114–117. Avidan B, Sonnenberg A, Schnell TG, Chejfec G, Metz A, Sontag SJ. 2002. Hiatal hernia size, Barrett’s length, and severity of acid reflux are all risk factors for esophageal adenocarcinoma. Am J Gastroenterol 97:1930– 1936. Axtell LM, Asire AJ, Myers MH. 1976. Cancer patient survival report number 5. DHEW Publication No. 77–992. Bethesda, MD: National Institute of Health.

703

Bjorge T, Hakulinen T, Engeland A, et al. 1997. A prospective, seroepidemiological study of the role of human papillomavirus in esophageal cancer in Norway. Cancer Res 57:3989–3992. Blaser MJ. 1999. Hypothesis: The changing relationships of Helicobacter pylori and humans: Implications for health and disease. J Infect Dis 179:1523–1530. Blot WJ. 1992. Alcohol and cancer. Cancer Res 52:2119s–2123s. Blot WJ. 1994. Esophageal cancer trends and risk factors. Semin Oncol 21:403–410. Blot WJ, Devesa SS, Kneller RW, Fraumeni JF, Jr. 1991. Rising incidence of adenocarcinoma of the esophagus and gastric cardia. JAMA 265: 1287–1289. Blot WJ, Li JY, Taylor PR, et al. 1993. Nutrition intervention trials in Linxian, China: Supplementation with specific vitamin/mineral combinations, cancer incidence, and disease-specific mortality in the general population. J Natl Cancer Inst 85:1483–1492. Boffetta P, Kogevinas M, Simonato L, Wilbourn J, Saracci R. 1995. Current perspectives on occupational cancer risks. Int J Occup Environ Health 1:315–325. Botterweck AA, Schouten LJ, Volovics A, Dorant E, van Den Brandt PA. 2000. Trends in incidence of adenocarcinoma of the oesophagus and gastric cardia in ten European countries. Int J Epidemiol 29:645–654. Brennan P, Lewis S, Hashibe M, et al. 2004. Pooled analysis of alcohol dehydrogenase genotypes and head and neck cancer: A HuGE review. Am J Epidemiol 159:1–16. Brown LM, Blot WJ, Schuman SH, et al. 1988. Environmental factors and high risk of esophageal cancer among men in coastal South Carolina. J Natl Cancer Inst 80:1620–1625. Brown LM, Hoover R, Silverman D, et al. 2001. Excess incidence of squamous cell esophageal cancer among US Black men: role of social class and other risk factors. Am J Epidemiol 153:114–122. Brown LM, Silverman DT, Pottern LM, et al. 1994. Adenocarcinoma of the esophagus and esophagogastric junction in white men in the United States: Alcohol, tobacco, and socioeconomic factors. Cancer Causes Control 5:333–340. Brown LM, Swanson CA, Gridley G, et al. 1995. Adenocarcinoma of the esophagus: Role of obesity and diet. J Natl Cancer Inst 87:104–109. Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ. 2003. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of US adults. N Engl J Med 348:1625–1638. Cameron AJ, Lagergren J, Henriksson C, Nyren O, Locke GR, III, Pedersen NL. 2002. Gastroesophageal reflux disease in monozygotic and dizygotic twins. Gastroenterology 122:55–59. Campos GM, DeMeester SR, Peters JH, et al. 2001. Predictive factors of Barrett esophagus: Multivariate analysis of 502 patients with gastroesophageal reflux disease. Arch Surg 136:1267–1273. Carter CL, Hu N, Wu M, Lin PZ, Murigande C, Bonney GE. 1992. Segregation analysis of esophageal cancer in 221 high-risk Chinese families. J Natl Cancer Inst 84:771–776. Castellsague X, Munoz N, De Stefani E, Victora CG, Castelletto R, Rolon PA. 2000. Influence of mate drinking, hot beverages and diet on esophageal cancer risk in South America. Int J Cancer 88:658–664. Castellsague X, Munoz N, De Stefani E, et al. 1999. Independent and joint effects of tobacco smoking and alcohol drinking on the risk of esophageal cancer in men and women. Int J Cancer 82:657–664. Chang F, Syrjanen S, Shen Q, et al. 2000. Human papillomavirus involvement in esophageal carcinogenesis in the high-incidence area of China. A study of 700 cases by screening and type-specific in situ hybridization. Scand J Gastroenterol 35:123–130. Cheng KK, Day NE. 1996. Nutrition and esophageal cancer. Cancer Causes Control 7:33–40. Cheng KK, Duffy SW, Day NE, Lam TH. 1995a. Oesophageal cancer in neversmokers and never-drinkers. Int J Cancer 60:820–822. Cheng KK, Duffy SW, Day NE, Lam TH, Chung SF, Badrinath P. 1995b. Stopping drinking and risk of oesophageal cancer. BMJ 310:1094– 1097. Chow WH, Blaser MJ, Blot WJ, et al. 1998a. An inverse relation between cagA+ strains of Helicobacter pylori infection and risk of esophageal and gastric cardia adenocarcinoma. Cancer Res 58:588–590. Chow WH, Blot WJ, McLaughlin JK. 1999. Tea drinking and cancer risk: Epidemiologic evidence. Proc Soc Exp Biol Med 220:197. Chow WH, Blot WJ, Vaughan TL, et al. 1998b. Body mass index and risk of adenocarcinomas of the esophagus and gastric cardia. J Natl Cancer Inst 90:150–155. Chow WH, Finkle WD, McLaughlin JK, Frankl H, Ziel HK, Fraumeni JF, Jr. 1995. The relation of gastroesophageal reflux disease and its treatment to adenocarcinomas of the esophagus and gastric cardia. JAMA 274:474– 477.

704

PART IV: CANCER BY TISSUE OF ORIGIN

Clark LC, Combs GF, Jr, Turnbull BW, et al. 1996. Effects of selenium supplementation for cancer prevention in patients with carcinoma of the skin. A randomized controlled trial. Nutritional Prevention of Cancer Study Group. JAMA 276:1957–1963. Conio M, Blanchi S, Lapertosa G, et al. 2003. Long-term endoscopic surveillance of patients with Barrett’s esophagus. Incidence of dysplasia and adenocarcinoma: A prospective study. Am J Gastroenterol 98:1931– 1939. Conio M, Cameron AJ, Romero Y, et al. 2001. Secular trends in the epidemiology and outcome of Barrett’s oesophagus in Olmsted County, Minnesota. Gut 48:304–309. Conio M, Filiberti R, Blanchi S, et al. 2002. Risk factors for Barrett’s esophagus: A case-control study. Int J Cancer 97:225–229. Cremonini F, Di Caro S, Delgado-Aros S, et al. 2003. Meta-analysis: The relationship between Helicobacter pylori infection and gastro-oesophageal reflux disease. Aliment Pharmacol Ther 18:279–289. Dawsey SM, Lewin KJ, Wang GQ, et al. 1994. Squamous esophageal histology and subsequent risk of squamous cell carcinoma of the esophagus. A prospective follow-up study from Linxian, China. Cancer 74:1686– 1692. De Stefani E, Ronco A, Mendilaharsu M, Deneo-Pellegrini H. 1998. Casecontrol study on the role of heterocyclic amines in the etiology of upper aerodigestive cancers in Uruguay. Nutr Cancer 32:43–48. Devesa SS, Blot WJ, Fraumeni JF, Jr. 1998. Changing patterns in the incidence of esophageal and gastric carcinoma in the United States. Cancer 83: 2049–2053. Dhillon PK, Farrow DC, Vaughan TL, et al. 2001. Family history of cancer and risk of esophageal and gastric cancers in the United States. Int J Cancer 93:148–152. Drovdlic CM, Goddard KA, Chak A, et al. 2003. Demographic and phenotypic features of 70 families segregating Barrett’s oesophagus and oesophageal adenocarcinoma. J Med Genet 40:651–656. DuBois RN. 2002. New agents for cancer prevention. J Natl Cancer Inst 94:1732–1733. Dulai GS, Guha S, Kahn KL, Gornbein J, Weinstein WM. 2002. Preoperative prevalence of Barrett’s esophagus in esophageal adenocarcinoma: A systematic review. Gastroenterology 122:26–33. Ellis A, Field JK, Field EA, et al. 1994. Tylosis associated with carcinoma of the oesophagus and oral leukoplakia in a large Liverpool family—a review of six generations. Eur J Cancer B Oral Oncol 30B:102–112. Eng C, Spechler SJ, Ruben R, Li FP. 1993. Familial Barrett esophagus and adenocarcinoma of the gastroesophageal junction. Cancer Epidemiol Biomarkers Prev 2:397–399. Engel LS, Vaughan TL, Gammon MD, et al. 2002. Occupation and risk of esophageal and gastric cardia adenocarcinoma. Am J Ind Med 42:11–22. Enzinger PC, Mayer RJ. 2003. Esophageal cancer. N Engl J Med 349:2241– 2252. Farrow DC, Vaughan TL, Hansten PD, et al. 1998. Use of aspirin and other nonsteroidal anti-inflammatory drugs and risk of esophageal and gastric cancer. Cancer Epidemiol Biomarkers Prev 7:97–102. Farrow DC, Vaughan TL, Sweeney C, et al. 2000. Gastroesophageal reflux disease, use of H2 receptor antagonists, and risk of esophageal and gastric cancer. Cancer Causes Control 11:231–238. Gammon MD, Schoenberg JB, Ahsan H, et al. 1997. Tobacco, alcohol, and socioeconomic status and adenocarcinomas of the esophagus and gastric cardia. J Natl Cancer Inst 89:1277–1284. Gammon MD, Terry MB, Arber N, et al. 2004. Nonsteroidal anti-inflammatory drug use associated with reduced incidence of adenocarcinomas of the esophagus and gastric cardia that overexpress cyclin D1: A populationbased study. Cancer Epidemiol Biomarkers Prev 13:34–39. Gao YT, McLaughlin JK, Blot WJ, Ji BT, Dai Q, Fraumeni JF, Jr. 1994. Reduced risk of esophageal cancer associated with green tea consumption. J Natl Cancer Inst 86:855–858. Goodman M, Morgan RW, Ray R, Malloy CD, Zhao K. 1999. Cancer in asbestos-exposed occupational cohorts: A meta-analysis. Cancer Causes Control 10:453–465. Graham DY. 2003. The changing epidemiology of GERD: Geography and Helicobacter pylori. Am J Gastroenterol 98:1462–1470. Gupta RA, DuBois RN. 2002. Cyclooxygenase-2 inhibitor therapy for the prevention of esophageal adenocarcinoma in Barrett’s esophagus. J Natl Cancer Inst 94:406–407. Gustavsson P, Evanoff B, Hogstedt C. 1993. Increased risk of esophageal cancer among workers exposed to combustion products. Arch Environ Health 48:243–245. Gustavsson P, Jakobsson R, Johansson H, Lewin F, Norell S, Rutkvist LE. 1998. Occupational exposures and squamous cell carcinoma of the oral cavity, pharynx, larynx, and oesophagus: A case-control study in Sweden. Occup Environ Med 55:393–400.

Hansen J, Raaschou-Nielsen O, Christensen JM, et al. 2001. Cancer incidence among Danish workers exposed to trichloroethylene. J Occup Environ Med 43:133–139. Hebert JR, Kabat GC. 1989. Menthol cigarette smoking and oesophageal cancer. Int J Epidemiol 18:37–44. Henrik SJ, Forsgren A, Berglund G, Floren CH. 2001. Helicobacter pylori infection is associated with a decreased risk of developing oesophageal neoplasms. Helicobacter 6:310–316. Herrero R, Castellsague X, Pawlita M, et al. 2003. Human papillomavirus and oral cancer: The International Agency for Research on Cancer multicenter study. J Natl Cancer Inst 95:1772–1783. Hille JJ, Markowitz S, Margolius KA, Isaacson C. 1985. Human papillomavirus and carcinoma of the esophagus. N Engl J Med 312:1707. Holmes GK, Stokes PL, Sorahan TM, Prior P, Waterhouse JA, Cooke WT. 1976. Coeliac disease, gluten-free diet, and malignancy. Gut 17:612–619. Huang WY, Winn DM, Brown LM, et al. 2003. Alcohol concentration and risk of oral cancer in Puerto Rico. Am J Epidemiol 157:881–887. International Agency for Research on Cancer. 1986. Tobacco smoking. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. Vol 38. Lyon, France. International Agency for Research on Cancer. 1988. Alcohol drinking. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. Lyon, France. Iwaya T, Maesawa C, Ogasawara S, Tamura G. 1998. Tylosis esophageal cancer locus on chromosome 17q25.1 is commonly deleted in sporadic human esophageal cancer. Gastroenterology 114:1206–1210. Jacob JH, Riviere A, Mandard AM, et al. 1993. Prevalence survey of precancerous lesions of the oesophagus in a high-risk population for oesophageal cancer in France. Eur J Cancer Prev 2:53–59. Jemal A, Murray T, Samuels A, Ghafoor A, Ward E, Thun MJ. 2003. Cancer statistics, 2003. CA Cancer J Clin 53:5–26. Kabat GC, Ng SK, Wynder EL. 1993. Tobacco, alcohol intake, and diet in relation to adenocarcinoma of the esophagus and gastric cardia. Cancer Causes Control 4:123–132. Kuczmarski RJ, Flegal KM, Campbell SM, Johnson CL. 1994. Increasing prevalence of overweight among US adults. The National Health and Nutrition Examination Surveys, 1960 to 1991. JAMA 272:205–211. La Vecchia C and Negri E. 1989. The role of alcohol in oesophageal cancer in non-smokers, and of tobacco in non-drinkers. Int J Cancer 43:784–785. Lagergren J, Bergstrom R, Lindgren A, Nyren O. 1999a. Symptomatic gastroesophageal reflux as a risk factor for esophageal adenocarcinoma. N Engl J Med 340:825–831. Lagergren J, Bergstrom R, Lindgren A, Nyren O. 2000a. The role of tobacco, snuff and alcohol use in the aetiology of cancer of the oesophagus and gastric cardia. Int J Cancer 85:340–346. Lagergren J, Bergstrom R, Nyren O. 1999b. Association between body mass and adenocarcinoma of the esophagus and gastric cardia. Ann Intern Med 130:883–890. Lagergren J, Bergstrom R, Nyren O. 2000b. No relation between body mass and gastro-oesophageal reflux symptoms in a Swedish population based study. Gut 47:26–29. Lagergren J, Wang Z, Bergstrom R, Dillner J, Nyren O. 1999c. Human papillomavirus infection and esophageal cancer: A nationwide seroepidemiologic case-control study in Sweden. J Natl Cancer Inst 91:156–162. Lam AK. 2000. Molecular biology of esophageal squamous cell carcinoma. Crit Rev Oncol Hematol 33:71–90. Lambert JD and Yang CS. 2003. Mechanisms of cancer prevention by tea constituents. J Nutr 133:3262S–3267S. Larsson LG, Sandstrom A, Westling P. 1975. Relationship of Plummer Vinson disease to cancer of upper alimentary tract in Sweden. Cancer Res 35:3308–3316. Launoy G, Milan C, Day NE, Faivre J, Pienkowski P, Gignoux M. 1997. Oesophageal cancer in France: Potential importance of hot alcoholic drinks. Int J Cancer 71:917–923. Launoy G, Milan C, Faivre J, Pienkowski P, Gignoux M. 2000. Tobacco type and risk of squamous cell cancer of the oesophagus in males: A French multicentre case-control study. Int J Epidemiol 29:36–42. Li JY, Taylor PR, Li B, et al. 1993. Nutrition intervention trials in Linxian, China: Multiple vitamin/mineral supplementation, cancer incidence, and disease-specific mortality among adults with esophageal dysplasia. J Natl Cancer Inst 85:1492–1498. Li SD, Mobarhan S. 2000. Association between body mass index and adenocarcinoma of the esophagus and gastric cardia. Nutr Rev 58:54–56. Mandard AM, Hainaut P, Hollstein M. 2000. Genetic steps in the development of squamous cell carcinoma of the esophagus. Mutat Res 462:335–342. Mark SD, Liu SF, Li JY, et al. 1994. The effect of vitamin and mineral supplementation on esophageal cytology: Results from the Linxian Dysplasia Trial. Int J Cancer 57:162–166.

Esophageal Cancer Mark SD, Qiao YL, Dawsey SM, et al. 2000. Prospective study of serum selenium levels and incident esophageal and gastric cancers. J Natl Cancer Inst 92:1753–1763. Mayne ST, Navarro SA. 2002. Diet, obesity and reflux in the etiology of adenocarcinomas of the esophagus and gastric cardia in humans. J Nutr 132:3467S–3470S. Mayne ST, Risch HA, Dubrow R, et al. 2001. Nutrient intake and risk of subtypes of esophageal and gastric cancer. Cancer Epidemiol Biomarkers Prev 10:1055–1062. McManus DT, Olaru A, Meltzer SJ. 2004. Biomarkers of esophageal adenocarcinoma and Barrett’s esophagus. Cancer Res 64:1561–1569. Mokdad AH, Ford ES, Bowman BA, et al. 2003. Prevalence of obesity, diabetes, and obesity-related health risk factors, 2001. JAMA 289:76–79. Morita M, Kuwano H, Nakashima T, et al. 1998. Family aggregation of carcinoma of the hypopharynx and cervical esophagus: Special reference to multiplicity of cancer in upper aerodigestive tract. Int J Cancer 76: 468–471. Munoz N, Day NE. 1996. Esophageal cancer. In: Schottenfeld D, Fraumeni J, eds. Cancer Epidemiology and Prevention, 2nd ed. New York: Oxford University Press, pp. 681–706. Munoz N, Wahrendorf J, Bang LJ, et al. 1985. No effect of riboflavine, retinol, and zinc on prevalence of precancerous lesions of oesophagus. Randomised double-blind intervention study in high-risk population of China. Lancet 2:111–114. Murray L, Watson P, Johnston B, Sloan J, Mainie IM, Gavin A. 2003. Risk of adenocarcinoma in Barrett’s oesophagus: Population based study. BMJ 327:534–535. National Cancer Institute. Cigars: Health Effects and Trends. 1998. Smoking and Tobacco Control Monograph No. 9. Bethesda, MD: National Institutes of Health, National Cancer Institute. Nilsson M, Johnsen R, Ye W, Hveem K, Lagergren J. 2003. Obesity and estrogen as risk factors for gastroesophageal reflux symptoms. JAMA 290:66–72. Norell S, Ahlbom A, Lipping H, Osterblom L. 1983. Oesophageal cancer and vulcanisation work. Lancet 1:462–463. Nyren O and Adami HO. 2002. Esophageal cancer. In: Adami HO, Hunter D, Trichopoulos D, eds. Textbook of Cancer Epidemiology. New York: Oxford University Press, pp. 137–161. Oettle GJ, Paterson AC, Leiman G, Segal I. 1986. Esophagitis in a population at risk for esophageal carcinoma. Cancer 57:2222–2229. Parkin DM, Whelan SL, Ferlay D, Teppo L, Thomas D. 2003. Cancer Incidence in Five Continents. IARC Sci Pub No. 155. Lyon, France: International Agency for Research on Cancer. Peixoto GD, Hsin LS, Snijderse P, et al. 2001. Absence of association between HPV DNA, TP53 codon 72 polymorphism, and risk of oesophageal cancer in a high-risk area of China. Cancer Lett 162:231–235. Raaschou-Nielsen O, Hansen J, McLaughlin JK, et al. 2003. Cancer risk among workers at Danish companies using trichloroethylene: A cohort study. Am J Epidemiol 158:1182–1192. Reid BJ, Blount PL, Rabinovitch PS. 2003. Biomarkers in Barrett’s esophagus. Gastrointest Endosc Clin N Am 13:369–397. Radiation Effects Research Foundation. 1981. Life span study report 9, 1950–1978. Supplementary tables. Hiroshima: RERF. Risk JM, Evans KE, Jones J, et al. 2002. Characterization of a 500 kb region on 17q25 and the exclusion of candidate genes as the familial Tylosis Oesophageal Cancer (TOC) locus. Oncogene 21:6395–6402. Risk JM, Mills HS, Garde J, et al. 1999. The tylosis esophageal cancer (TOC) locus: More than just a familial cancer gene. Dis Esophagus 12:173– 176. Romero Y, Cameron AJ, Schaid DJ, et al. 2002. Barrett’s esophagus: Prevalence in symptomatic relatives. Am J Gastroenterol 97:1127– 1132. Rosen N. 1994. The molecular basis for cellular transformation: Implications for esophageal carcinogenesis. Semin Oncol 21:416–424. Rudolph RE, Vaughan TL, Kristal AR, et al. 2003. Serum selenium levels in relation to markers of neoplastic progression among persons with Barrett’s esophagus. J Natl Cancer Inst 95:750–757. Sandler RS, Nyren O, Ekbom A, Eisen GM, Yuen J, Josefsson S. 1995. The risk of esophageal cancer in patients with achalasia. A population-based study. JAMA 274:1359–1362. Surveillance, Epidemiology and End Results (SEER) Program 2003 SEER & Stat Database (www.seer.cancer.gov). Selikoff IJ, Hammond EC, Seidman H. 1979. Mortality experience of insulation workers in the United States and Canada, 1943–1976. Ann N Y Acad Sci 330:91–116. Sepehr A, Taniere P, Martel-Planche G, et al. 2001. Distinct pattern of TP53 mutations in squamous cell carcinoma of the esophagus in Iran. Oncogene 20:7368–7374.

705

Shaheen N, Ransohoff DF. 2002. Gastroesophageal reflux, Barrett esophagus, and esophageal cancer: Scientific review. JAMA 287:1972–1981. Sinha R, Rothman N. 1999. Role of well-done, grilled red meat, heterocyclic amines (HCAs) in the etiology of human cancer. Cancer Lett 143: 189–194. Smith EM, Ritchie JM, Summersgill KF, et al. 2004. Human papillomavirus in oral exfoliated cells and risk of head and neck cancer. J Natl Cancer Inst 96:449–455. Smith PG. 1984. Late effects of x-ray treatment of ankylosing spondylitis. In: Boice J, Fraumeni J, eds. Radiation Carcinogenesis. New York: Raven Press, pp. 107–118. Steinmetz KA, Potter JD. 1991. Vegetables, fruit, and cancer. I. Epidemiology. Cancer Causes Control 2:325–357. U.S. Department of Health and Human Services. 1986. The Health Consequences of Using Smokeless Tobacco. Bethesda, Maryland. Taniere P, Martel-Planche G, Maurici D, et al. 2001. Molecular and clinical differences between adenocarcinomas of the esophagus and of the gastric cardia. Am J Pathol 158:33–40. Tavani A, Negri E, Franceschi S, La Vecchia C. 1994. Risk factors for esophageal cancer in lifelong nonsmokers. Cancer Epidemiol Biomarkers Prev 3:387–392. Tavani A, Negri E, Franceschi S, La Vecchia C. 1996. Tobacco and other risk factors for oesophageal cancer in alcohol non-drinkers. Eur J Cancer Prev 5:313–318. Taylor PR, Qiao YL, Abnet CC, et al. 2003. Prospective study of serum vitamin E levels and esophageal and gastric cancers. J Natl Cancer Inst 95: 1414–1416. Terry P, Lagergren J, Wolk A, Nyren O. 2000a. Reflux-inducing dietary factors and risk of adenocarcinoma of the esophagus and gastric cardia. Nutr Cancer 38:186–191. Terry P, Lagergren J, Ye W, Nyren O, Wolk A. 2000b. Antioxidants and cancers of the esophagus and gastric cardia. Int J Cancer 87:750–754. Terry PD, Lagergren J, Wolk A, Steineck G, Nyren O. 2003. Dietary intake of heterocyclic amines and cancers of the esophagus and gastric cardia. Cancer Epidemiol Biomarkers Prev 12:940–944. Tuyns AJ. 1983. Oesophageal cancer in non-smoking drinkers and in nondrinking smokers. Int J Cancer 32:443–444. Tuyns AJ, Pequignot G, Jensen OM. 1977. [Esophageal cancer in Ille-etVilaine in relation to levels of alcohol and tobacco consumption. Risks are multiplying]. Bull Cancer 64:45–60. Tzonou A, Lipworth L, Garidou A, et al. 1996. Diet and risk of esophageal cancer by histologic type in a low-risk population. Int J Cancer 68: 300–304. Vaughan TL, Davis S, Kristal A, Thomas DB. 1995. Obesity, alcohol, and tobacco as risk factors for cancers of the esophagus and gastric cardia: Adenocarcinoma versus squamous cell carcinoma. Cancer Epidemiol Biomarkers Prev 4:85–92. Vaughan TL, Farrow DC, Hansten PD, et al. 1998. Risk of esophageal and gastric adenocarcinomas in relation to use of calcium channel blockers, asthma drugs, and other medications that promote gastroesophageal reflux. Cancer Epidemiol Biomarkers Prev 7:749–756. Vieth M, Masoud B, Meining A, Stolte M. 2000. Helicobacter pylori infection: protection against Barrett’s mucosa and neoplasia? Digestion 62:225– 231. Ward MH, Sinha R, Heineman EF, et al. 1997. Risk of adenocarcinoma of the stomach and esophagus with meat cooking method and doneness preference. Int J Cancer 71:14–19. Weiss NS. 1995. Cancer in relation to occupational exposure to perchloroethylene. Cancer Causes Control 6:257–266. Wild CP, Hardie LJ. 2003. Reflux, Barrett’s oesophagus and adenocarcinoma: Burning questions. Nat Rev Cancer 3:676–684. Wilson LJ, Ma W, Hirschowitz BI. 1999. Association of obesity with hiatal hernia and esophagitis. Am J Gastroenterol 94:2840–2844. Wu AH, Crabtree JE, Bernstein L, et al. 2003a. Role of Helicobacter pylori CagA+ strains and risk of adenocarcinoma of the stomach and esophagus. Int J Cancer 103:815–821. Wu AH, Tseng CC, Bernstein L. 2003b. Hiatal hernia, reflux symptoms, body size, and risk of esophageal and gastric adenocarcinoma. Cancer 98: 940–948. Wu AH, Wan P, Bernstein L. 2001. A multiethnic population-based study of smoking, alcohol and body size and risk of adenocarcinomas of the stomach and esophagus (United States). Cancer Causes Control 12: 721–732. Ye W, Chow WH, Lagergren J, et al. 2001a. Risk of adenocarcinomas of the oesophagus and gastric cardia in patients hospitalized for asthma. Br J Cancer 85:1317–1321. Ye W, Chow WH, Lagergren J, Yin L, Nyren O. 2001b. Risk of adenocarcinomas of the esophagus and gastric cardia in patients with gastroe-

706

PART IV: CANCER BY TISSUE OF ORIGIN

sophageal reflux diseases and after antireflux surgery. Gastroenterology 121:1286–1293. Ye W, Held M, Lagergren J, et al. 2004. Helicobacter pylori infection and gastric atrophy: Risk of adenocarcinoma and squamous-cell carcinoma of the esophagus and adenocarcinoma of the gastric cardia. J Natl Cancer Inst 96:388–396. Ye W, Nyren O. 2003. Risk of cancers of the oesophagus and stomach by histology or subsite in patients hospitalised for pernicious anaemia. Gut 52:938–941. Yokoyama A, Kato H, Yokoyama T, et al. 2002. Genetic polymorphisms of alcohol and aldehyde dehydrogenases and glutathione S-transferase M1 and drinking, smoking, and diet in Japanese men with esophageal squamous cell carcinoma. Carcinogenesis 23:1851–1859. Young JL, Jr., Percy CL, Asire AJ. 1981. Surveillance, Epidemiology and End Results: Incidence and mortality, 1973–77. Natl Cancer Inst Monogr 57:56–61. Yu HP, Xu SQ, Liu L, et al. 2003. Cyclooxygenase-2 expression in squamous dysplasia and squamous cell carcinoma of the esophagus. Cancer Lett 198:193–201.

Zambon P, Talamini R, La Vecchia C, et al. 2000. Smoking, type of alcoholic beverage and squamous-cell oesophageal cancer in northern Italy. Int J Cancer 86:144–149. Zaridze D, Evstifeeva T, Boyle P. 1993. Chemoprevention of oral leukoplakia and chronic esophagitis in an area of high incidence of oral and esophageal cancer. Ann Epidemiol 3:225–234. Zhang W, Bailey-Wilson JE, Li W, et al. 2000. Segregation analysis of esophageal cancer in a moderately high-incidence area of northern China. Am J Hum Genet 67:110–119. Zhang ZF, Kurtz RC, Sun M, Karpeh M, Jr, et al. 1996. Adenocarcinomas of the esophagus and gastric cardia: Medical conditions, tobacco, alcohol, and socioeconomic factors. Cancer Epidemiol Biomarkers Prev 5:761–768. Zhou XB, Guo M, Quan LP, et al. 2003. Detection of human papillomavirus in Chinese esophageal squamous cell carcinoma and its adjacent normal epithelium. World J Gastroenterol 9:1170–1173. Znaor A, Brennan P, Gajalakshmi V, et al. 2003. Independent and combined effects of tobacco smoking, chewing and alcohol drinking on the risk of oral, pharyngeal and esophageal cancers in Indian men. Int J Cancer 105:681–686.

37

Stomach Cancer ATSUKO SHIBATA AND JULIE PARSONNET

A

ccording to the estimates by the World Health Organization (WHO) (World Health Organization, 2002), approximately 850,000 people (522,000 men and 328,000 women) died of stomach cancer in 2001. This makes stomach cancer the second leading cause of cancer death, accounting for 12% of cancer deaths (13% for men and 10% for women), behind only lung cancer. Also, in the global cancer incidence estimates for 1990 (Parkin et al., 1999), stomach cancer was the second most frequent malignant neoplasm after lung cancer, accounting for 9.9% (798,000 cases—511,000 men and 287,000 women) of newly diagnosed cancer cases. In the United States, stomach cancer was the leading cause of cancer death in men and the third leading cause in women until around 1940 (Howson et al., 1986). However, both incidence and mortality rates have declined steadily since (Jemal et al., 2003; Ries et al., 2003). The American Cancer Society (Jemal et al., 2005) estimated that there would be 21,860 new cases (13,510 men and 8350 women) of stomach cancer and 11,550 deaths (6770 men and 4780 women) due to stomach cancer in 2005 in the United States. Although age-standardized incidence and mortality rates have been declining for the past few decades even in high-risk countries (Parsonnet, 1999), stomach cancer still accounts for a large proportion of cancer cases among both men and women in Eastern Asia, the Andean regions of South America, and Eastern Europe (Hamilton and Aaltonen, 2000b; Parkin et al., 1999). Despite the “unplanned triumph” (Howson et al., 1986) of decreasing incidence and mortality, stomach cancer remains a major health problem because of its poor prognosis and the aging of the world’s population. Over 90% of stomach cancer cases are adenocarcinomas arising from the gastric glands (Coleman et al., 1993; Hamilton and Aaltonen, 2000b). Other histologic types of epithelial stomach cancer include squamous cell carcinoma, adenosquamous carcinoma, small cell carcinoma, and carcinoid (well-differentiated endocrine neoplasm). Gastric tumors of non-epithelial origin include leiomyoma, stromal tumor, and malignant lymphomas (Hamilton and Aaltonen, 2000b). Cancer that arises from the cardia (proximal) portion of the stomach and the junction between the esophagus and stomach presents distinct epidemiologic features in comparison with cancer of more distal parts of the stomach (i.e., fundus, body, and antrum) (Brown and Devesa, 2002). This chapter focuses on adenocarcinoma of the stomach, including that of the gastric cardia and gastoesophageal junction.

CLASSIFICATION Anatomic Subsite Grossly, the stomach has four regions: cardia, fundus, body, and antrum (including the pylorus). The fundus and body are identical in microscopic structure and differ from the antrum and the cardia (Junqueira et al., 1998). The anatomical region where the tubular esophagus joins the stomach is called the gastroesophageal (or esophagogastric) junction (Hamilton and Aaltonen, 2000a). The gastric cardia, a narrow circular band that is 1.5–3 cm wide, begins at the gastroesophageal (GE) junction, but its distal extent is poorly defined (Hamilton and Aaltonen, 2000a). Although the GE junction is observable endoscopically, it is often difficult to assign the origin of a cancer located in the GE junction area to the lower esophagus or to

the gastric cardia (Misumi et al., 1989). The most frequent site of subcardial stomach cancer is the antro-pyloric region (Hamilton and Aaltonen, 2000b). In the United States, age-standardized incidence rates (per 100,000 population per year, standardized to the 1970 US population) of gastric cardia adenocarcinoma have increased gradually between 1974 and 1994 in both white (2.1–3.3) and black (1.0–1.9) men (Devesa et al., 1998). During the same period, age-standardized incidence rates (per 100,000 population per year) for adenocarcinoma arising in other specified parts of the stomach did not change among black men, but declined from 5.1–3.7 among white men (Devesa et al., 1998). Similar upward trends of gastric cardia adenocarcinoma have been observed in women as well, but their rates remained much lower than those of men (Devesa et al., 1998). Wayman et al. (2001) observed similar time trends in subsite distribution in the United Kingdom. The majority of stomach cancer cases reported to cancer registries, however, had no detailed information on subsite recorded (Wayman et al., 2001). Improved completeness of registry data is needed to monitor changes in subsite distribution.

Histologic Type The WHO’s Histological Classification of Gastro-Esophageal Tumors (Hamilton and Aaltonen, 2000b; Watanabe et al., 1990) recognizes a variety of histologic types of gastric tumors (Table 37–1). Adenocarcinoma, a malignant tumor of glandular epithelium, is by far the most common type of malignant gastric tumor (Coleman et al., 1993; Hamilton and Aaltonen, 2000b). Laurén (1965) classified adenocarcinoma of the stomach into two histologic types—intestinal and diffuse—according to morphological features of tumor. Intestinal-type carcinomas form recognizable glands whereas diffuse-type counterparts consist of poorly cohesive cells diffusely infiltrating the gastric wall with little or no gland formation (Hamilton and Aaltonen, 2000b). The Laurén classification has proven useful in evaluating the natural history of gastric carcinoma, especially with regard to its incidence trends and precursors (Hamilton and Aaltonen, 2000b; Lewin and Appelman, 1995). Laurén himself observed a stronger male preponderance and older age at diagnosis for intestinal-type cancer than for diffuse-type cancer (Laurén, 1965). Hanai et al. (1982) examined time trends of histologic distribution of stomach cancer in Osaka, Japan, between 1966 and 1977, and corroborated Laurén’s observation. The decrease in intestinal-type tumors largely accounts for the declining trends of stomach cancer incidence whereas the incidence of diffuse-type cancer has not changed as significantly (Correa and Chen, 1994; Hanai et al., 1982). This finding suggests a strong influence of environmental or lifestyle risk factors on intestinal-type cancers. However, distinct time trends for these two Laurén subtypes have not been demonstrated unequivocally (Lundegardh et al., 1991). Preneoplastic lesions (see below) are often associated with the intestinal type but infrequently with the diffuse type (Correa, 1992).

Tumor Grade and Stage The grading of gastric adenocarcinoma, which applies primarily to tubular carcinomas, is based on the degree of differentiation (Hamilton and Aaltonen, 2000b). Well-differentiated tumors have well-formed glands that often resemble metaplastic intestinal

707

708

PART IV: CANCER BY TISSUE OF ORIGIN

Table 37–1. Histologic Classification of Gastric Tumors (World Health Organization) Tumor Origin and Type

Histologic Type

Intraepithelial neoplasia Adenoma Epithelial cancer Adenocarcinoma

Subtype

Intestinal type Diffuse type

Papillary adenocarcinoma Tubular adenocarcinoma Mucinous adenocarcinoma Signet-ring cell carcinoma Adenosquamous carcinoma Squamous cell carcinoma Undifferentiated carcinoma Others Carcinoid (well-differentiated endocrine neoplasm) Non-epithelial tumors Leiomyoma Schwannoma Granular cell tumor Glomus tumor Leiomyosarcoma GI stromal tumor Benign Uncertain malignant potential Malignant Kaposi sarcoma Others Malignant lymphomas Marginal zone B-cell lymphoma of MALT-type Mantle cell lymphoma Diffuse large B-cell lymphoma Others Source: Adapted from Hamilton and Aaltonen (2000).

epithelium, whereas poorly differentiated tumors are composed of highly irregular glands that are recognized with difficulty or single cells that remain isolated or arranged in small or large clusters. Moderately differentiated tumors display patterns intermediate between well and poorly differentiated. The TNM classification of gastric carcinomas for clinical and pathologic staging considers the level of invasiveness of primary tumor (T), involvement of regional lymph nodes (N), and presence or absence of distant metastasis (M) (American Joint Committee on Cancer, 2002). The staging system used by the SEER program classifies cancer cases into three categories: localized, regional, and distant. In the United States from 1995–2000, 24% of stomach cancer cases were diagnosed at the localized stage, 31% at the regional stage, and 32% at the distant stage, with the remaining 13% recorded as unstaged (Jemal et al., 2005).

Molecular Genetic Markers A large number of molecular and genetic markers have been examined in gastric adenocarcinoma for their relation to tumor development and progression as well as their potential as prognostic markers. According to a review of 212 articles published between 1979 and 1997 (Allgayer et al., 1997), some of the markers—including plasminogen activator inhibitor 1 (PAI-1) and c-erbB-2—have high potential as prognostic factors for gastric cancer. More recently, a systematic search for over- and under-expressed genes in gastric cancer tissue with the use of a microarray technique identified an association between low expression of PLA2G2A—a gene previously implicated as a modifier of colonic polyp risk in the mouse—in gastric adenocarcinoma and poor prognosis (Leung et al., 2002). Most of these molecular genetic studies have limited sample size, and their promising findings must be validated in larger, well-designed studies (Simon et al., 2003).

PRENEOPLASIA Some histopathological changes in gastric mucosa have been identified as precursors for intestinal-type adenocarcinoma of the stomach (Correa, 2002). No clearly defined and universally accepted precursors have been reported for diffuse-type cancer. Correa and colleagues (1976, 1992, 2002) have proposed a multistep premalignant process for the intestinal-type gastric carcinoma, which consists of sequential histopathological changes in the gastric mucosa—chronic gastritis, intestinal metaplasia, and dysplasia. In populations at high risk of gastric cancer, chronic gastritis is found predominantly in the antrum and is associated with multifocal gland loss (multifocal atrophic gastritis) (Correa, 2002). There are two main types of intestinal metaplasia: “complete” (also called small intestinal type or type I) and “incomplete” (also called colonic type or types II and III) (Hamilton and Aaltonen, 2000b). Incomplete metaplasia is frequently associated with frank dysplasia and early carcinoma (Correa, 2002). Dysplasia (also called intraepithelial neoplasia), which arises in either the native gastric or intestinalized gastric epithelia, is characterized by partial or complete loss of differentiation (Correa, 2002; Hamilton and Aaltonen, 2000b). There have been efforts (e.g., Padova International Classification) to rectify inconsistencies and semantic misunderstandings across countries regarding the terminology of the morphological spectrum of preneoplastic lesions and early invasive cancer (Correa, 2002; Hamilton and Aaltonen, 2000b; Rugge et al., 2000). Because it is not possible to follow histopathological changes of a single lesion over time without intervention, we cannot prove unequivocally that cells of an alleged preneoplastic condition, such as intestinal metaplasia, progress and expand into a full-blown cancer. Tatematsu et al. (2003) have proposed that the occurrence of intestinal metaplasia and gastric cancer are independent events with common causes. Another possible mechanism, by which preneoplasia and gastric carcinoma could be linked, is through physiologic changes in gastric environment caused by underlying atrophy, favoring the growth of a bacteria capable of producing endogenous mutagens (Stemmermann, 1994).

DEMOGRAPHIC PATTERNS Incidence and Mortality in the United States According to the data reported by the nine population-based cancer registries (San Francisco-Oakland, Connecticut, Detroit, Hawaii, Iowa, New Mexico, Seattle-Puget Sound, Utah, and Atlanta) in the US National Cancer Institute’s Surveillance, Epidemiology and End Results Program (SEER) (Ries et al., 2003), average annual incidence rates (per 100,000 population) of stomach cancer for all races combined during the period 1996–2000, standardized to the 2000 US standard population, were 12.2 (95% confidence interval (CI): 11.9–12.5) for men and 5.7 (95% CI: 5.5–5.8) for women. Age-standardized incidence rates were higher in blacks than in whites for both men (20.5 vs. 10.5) and women (9.8 vs. 4.6). Among the same nine SEER registries (Ries et al., 2003), agestandardized incidence rates (per 100,000 population) for white men ranged from 7.3 in Utah to 12.7 in Detroit, and those for white women ranged from 3.1 in Iowa to 6.5 in Hawaii. Incidence rates were approximately twofold higher in blacks than in whites in the registries with a black population large enough for statistically stable estimates. Male rates were approximately twofold higher than female rates across the registries for both blacks and whites. According to the age-standardized rates (per 100,000 population per year) for 1995–1999 from the 11 SEER registries (the abovementioned nine registries plus San Jose-Monterey and Los Angeles County) (Ries et al., 2002), Asian or Pacific Islander males had the highest incidence (23.9; 95% CI: 22.7–25.2), followed in descending order by black (18.7; 95% CI: 17.5–20.0), Hispanic (16.2; 95% CI: 15.2–17.3), white (11.3; 95% CI: 11.0–11.5), and American Indian/Alaska Native men (9.9; 95% CI: 7.0–13.7). Female rates showed a similar pattern with the highest incidence among Asian or Pacific Islanders (12.9; 95% CI: 12.1–13.7), followed by blacks (10.2;

709

Stomach Cancer

White Male 35

Rate per 100,000

30

Black Male 35

Rate per 100,000

White Female 15

Rate per 100,000

Black Female 15

30 12

12

9

9

APC = –1.1* 25

25

20

20

15

Rate per 100,000

APC = –1.8*

APC = –1.2*

APC = –2.6* APC = –1.9*

10

5

APC = –3.9*

5

APC = –0.8*

APC = –2.1*

15

10

APC = –0.8*

Incidence

APC = –2.2*

6

6

APC = –1.2*

3

3

Delay-Adjusted Incidence

APC = –2.6*

Mortality

0 1975 1980 1985 1990 1995 2000

0 1975 1980 1985 1990 1995 2000

0 1975 1980 1985 1990 1995 2000

0 1975 1980 1985 1990 1995 2000

Year of Diagnosis/Death

Year of Diagnosis/Death

Year of Diagnosis/Death

Year of Diagnosis/Death

Figure 37–1. Time trends of SEER incidence, delay-adjusted incidence, and US death rates of stomach cancer by race and sex. APC: Annual

Percent Change; SEER: Surveillance, Epidemiology, and End Results. (Source: Reproduced with permission from Ries et al., 2003.)

95% CI: 9.5–11.0), Hispanics (9.0; 95% CI: 8.4–9.6), American Indians/Alaska Natives (6.6; 95% CI: 4.7–9.2), and whites (5.1; 95% CI: 4.9–5.2). Age-standardized rates of stomach cancer incidence have steadily declined from 1975 to 1999 in white men and women (Fig. 37–1). Rates for blacks show a large fluctuation because of the smaller population size, but still indicate a long-range downward trend with a less steep slope for women (Fig. 37–1). Mortality from stomach cancer in the United States, when expressed in age-standardized rates, has also declined steadily over time. Mortality data, which had been collected long before incidence data from population-based cancer registries became available, indicate that the earliest declining trend of stomach cancer mortality in the United States was evident in 1926 (Howson et al., 1986). The downward time trends in stomach cancer mortality have been observed in both men and women and in both whites and blacks. Demographic patterns of stomach cancer mortality rates with regard to sex and race/ethnicity parallel those of incidence rates. In the United States between 1996 and 2000 (Ries et al., 2003), average annual age-standardized rates (per 100,000 population, standardized to the 2000 US standard population) of stomach cancer mortality for all races combined were 6.9 (95% CI: 6.8–6.9) for men and 3.4 (95% CI: 3.3–3.4) for women. Among men, age-standardized mortality rate was the highest in blacks (14.0; 95% CI: 13.6–14.4), followed by Asians or Pacific Islanders (12.5; 95% CI: 11.9–13.2), Hispanics (10.0; 95% CI: 9.6–10.4), and American Indians/Alaska Natives (7.0; 95% CI: 5.9–8.2), and the lowest in whites (6.1; 95% CI: 6.0–6.1). Among women, the highest mortality rate was observed among Asians or Pacific Islanders (7.4; 95% CI: 7.0–7.9), followed by blacks (6.5; 95% CI: 6.4–6.7), Hispanics (5.4; 95% CI: 5.2–5.7), and American Indians/Alaska Natives (4.2; 95% CI: 3.6–5.0), and the lowest in whites (2.9; 95% CI: 2.9–3.0). Based on the 1998–2000 data from the SEER 12 (11 registries mentioned above and the Alaska Native Tumor Registry), a lifetime risk of being diagnosed with stomach cancer was estimated to be 1.09% for white men, 1.32% for black men, 0.66% for white women, and 1.03% for black women (Ries et al., 2003). With the mortality statistics for the total United States (1998–2000), a lifetime risk of dying from stomach cancer was estimated to be 0.56% for white men, 0.92%

for black men, 0.37% for white women, and 0.69% for black women (Ries et al., 2003). Both incidence and mortality rates of stomach cancer increase with advancing age in both sexes and all racial/ethnic groups (Ries et al., 2002).

International Trends in Incidence and Mortality Age-standardized incidence rates of stomach cancer vary considerably among countries. According to the global estimates for 1990 (Parkin et al., 1999) (Fig. 37–2), Japan had the highest rates (per 100,000 population per year, standardized to the world standard population) of 77.9 in men and 33.3 in women. High rates were also observed in both sexes in Eastern Asia, China, Eastern Europe, and tropical South America. In contrast, low incidence rates were observed in Eastern and Northern Africa, North America, and South and Southeast Asia with age-standardized rates of 5.9–9.0 in men and 2.6–5.3 in women. The approximately twofold difference in stomach cancer incidence between men and women holds for all the regions of the world (Parkin et al., 1999). Trends in incidence in most countries are closely mirrored by trends in mortality (Coleman et al., 1993). Age-standardized rates of stomach cancer incidence and mortality have shown a steady decline in most countries (Coleman et al., 1993; Parkin et al., 1999). Because of the declining trend of age-adjusted incidence rates (4%–5% decline between 1985 and 1990), the world estimate of the number of new stomach cancer cases in 1990 was only 6% greater than that in 1985 despite the population increase and aging (Parkin et al., 1999). Age-standardized mortality rates have declined worldwide at a rate of 10%–20% per decade (Parsonnet 1999). Given the little improvement in survival after diagnosis, this decreasing trend seems to indicate a global decline in stomach cancer incidence rather than improvements in treatment.

Migration The examination of cancer risk, measured as incidence and mortality rates, among people who migrated from a low-risk to a high-risk population, or vice versa, can provide an insight about relative contributions of genetic (heritable) and environmental (potentially modifiable) factors to cancer risk (see Chapter 11 for a more detailed discussion

710

PART IV: CANCER BY TISSUE OF ORIGIN

(a)

Figure 37–2. Age-standardized incidence rates for stomach cancer (number of cases per 100,000 population, standardized to the world standard population) of stomach cancer by world region among (a) males and (b) females. (Source: Reproduced with permission from Parkin et al., 1999.)

(b)

of migrant studies). In a study of Japanese migrants to Hawaii and their descendants (Kolonel et al., 1986), age-adjusted incidence rates of stomach cancer in the 1970s for both men and women were lower in Japanese-born migrants to Hawaii (“Issei” or first generation) than in the Japanese in Japan. The rates were even lower in the Hawaiianborn Japanese (“Nisei” or second generation), which were still higher than the Caucasian rates. Migrants to Australia from seven European countries with higher stomach cancer rates than Australia (England, Scotland, Ireland, Poland, Yugoslavia, Greece, and Italy) showed a risk reduction with increased duration of residence in Australia (McMichael et al., 1980). A study of migrant populations within Italy (Fascioli et al., 1995) suggested that place of birth was a stronger predictor of stomach cancer risk than current place of residence. Mortality data from England and Wales also showed a closer relation of stomach cancer risk to county of birth than county of death (Coggon et al., 1990), suggesting the significance of the environment in earlier life.

pancreatic cancers, in most regions of the world (Parkin et al., 1999). Japan was an exception with a moderately good 5-year survival of 57%. This favorable statistic can be explained by the larger proportion of earlystage, curable cancers that are diagnosed through the intensive stomach cancer screening programs in Japan. On average, developed countries had better 5-year survival (28%) than developing countries (18%). Among the stomach cancer patients diagnosed in the United States between 1992 and 1998, the overall survival at 5 years for all disease stages combined was 22% (Jemal et al., 2003). Stage-specific survival was the highest for localized disease (59%), followed by that for regional disease (22%), and the lowest for distant disease (2%) (Jemal et al., 2003). The survival seems to be lower for patients with cancer in the proximal stomach, with only 10%–15% survival at 5 years even with apparently localized disease (National Cancer Institute, 2002a).

ENVIRONMENTAL FACTORS Survival Estimated 5-year survival of stomach cancer diagnosed around 1990 was among the lowest of all cancer sites, only better than that of lung and

The strongest risk factor for stomach cancer identified to date is chronic bacterial infection with Helicobacter pylori (H. pylori). Because H. pylori infection was established as a risk factor for

Stomach Cancer stomach cancer only recently (early to mid-1990s) (IARC Working Group on the Evaluation of Carcinogenic Risks to Humans, 1994), studies conducted earlier had evaluated other potential risk factors and protective factors in the absence of information on H. pylori infection. Even more recent case-control studies, unless nested in prospective cohort studies, have limited ability to examine other risk factors in conjunction with H. pylori infection because blood samples obtained at stomach cancer diagnosis are of lesser value. Serological markers of the infection tend to be attenuated as the infected stomach mucosa progresses through preneoplastic and neoplastic changes (Genta and Graham, 1993; Masci et al., 1996; Osawa et al., 1996). Some risk factors and protective factors, such as smoking and certain dietary components, may be correlated with H. pylori infection, leading to confounding. Alternatively, some of the non-H. pylori risk factors and protective factors may modify the risk due to H. pylori infection, which is very likely since only a small minority of people infected with H. pylori ever develop stomach cancer. Therefore, the findings discussed below must be interpreted with caution, when H. pylori infection and other risk factors are not both measured and simultaneously considered in the same study.

H. pylori Infection Long before the discovery of H. pylori, gastric adenocarcinoma was known to typically arise within areas of gastritis. The type of gastritis associated with cancer—termed chronic type B gastritis or chronic active gastritis because of the presence of both lymphocytes and neutrophils—was extremely common in elderly populations and was thought to be a natural consequence of aging. In detailed studies from northern Europe and Latin America, the majority of individuals over the age of 50 years was found to have chronic type B gastritis (Siurala et al., 1985; Correa et al., 1976). With the rediscovery of H. pylori in the early 1980s [H. pylori had been described almost a century (Kreinitz, 1906) before Marshall and Warren’s groundbreaking work (Warren and Marshall, 1983)], the idea that gastric inflammation preceded cancer took on new meaning. Experimental ingestions and clinical trials of H. pylori eradication all demonstrated that H. pylori was the preeminent cause of type B gastritis (Dixon et al., 1996). Simultaneously, many began to rethink traditional dietary and genetic theories of gastric carcinogenesis. If H. pylori caused gastritis and gastritis was a precursor to malignancy in a large proportion of cases, H. pylori was likely to be a critical factor in carcinogenesis. During the last two decades of the 20th century, epidemiologic studies from investigators around the world have variably linked H. pylori to gastric cancer. Although results are heterogeneous, meta-analyses of these many studies indicate that H. pylori increases the risk of cancer (Huang et al., 1998). Case-control studies yield the smallest risk estimate (OR ‡1.8), nested case-control studies yield a higher risk estimate (OR ‡ 3.0), nested case-control studies with greater than 10 years of follow-up yield an even higher estimate (OR ‡5.9), and the single prospective study conducted to date yielded an infinite risk associated with infection (Huang et al., 1998; Helicobacter and Cancer Collaborative Group, 2001; Uemura et al., 2001). In the latter study (Uemura et al., 2001), which was conducted in Japan, 36 of 1246 infected subjects developed gastric cancer over a mean span of 7.8 years compared with none of 280 uninfected subjects. It is now widely accepted that H. pylori causes gastric malignancy. Yet, infection is extraordinarily common, infecting 50% of the world’s population. Less than 5% of infected hosts will develop cancer. Reconciling these two facts has been the goal of much of the more recent work on H. pylori. To date, factors that appear to influence the outcome of infection include bacterial genetics, host genetics, age of infection acquisition, and environmental factors. H. pylori readily loses and acquires DNA fragments and undergoes various forms of genetic change, such as point mutations and chromosomal rearrangements (Blaser and Berg, 2001). As a consequence, H. pylori isolates have an extraordinary degree of genetic variability between and even within individual hosts (Israel et al., 2001; Aras et al., 2002). One variable factor among strains is the presence of a

711

virulence cassette of genes termed the Cag (for cytotoxin-associated gene) pathogenicity island (PAI). The PAI encodes for a type four secretion system that injects the CagA protein into the host cell where it is phosphorylated and binds to the Shp-2 tyrosine phosphatase (Stein et al., 2002; Yamazaki et al., 2003). The host cell consequently elongates and acquires a growth-factor-like response (Segal et al., 1997). Approximately 60% of H. pylori isolates harbor the PAI. These PAIcontaining organisms cause greater inflammation and are more closely associated with intestinal-type malignancy than are strains without the PAI (Parsonnet et al.,1997). In contrast, diffuse-type cancers are similarly linked to both PAI-positive and PAI-negative isolates. Other polymorphic genes—such as the vacuolating cytotoxin and babA-2 adherence gene—have been linked to variability in outcome but none as strongly as the PAI. The host also plays an important role in H. pylori outcome. For example, El-Omar et al. (2000) reported that H. pylori-infected subjects who developed gastric cancer were more likely to have specific genotypes of interleukin (IL)-1b or the IL-1b receptor antagonist. Interestingly, the adverse genotypes of IL-1b not only induce more inflammation than the less deleterious genotypes but also cause enhanced suppression of gastric acid secretion, supporting the pathogenic model devised by Correa. Moreover, IL-1b genotype was not a risk factor for cancer in the absence of infection. Subsequent work has identified similar, though less strong, interactions between H. pylori and TNF-a (Machado, 2003). Other putative host factors that are being explored include p53 polymorphisms and HLA genotype. Environmental factors may also enhance or diminish H. pylori’s deleterious effects. Unfortunately, however, few environmental factors have been studied systematically in conjunction with H. pylori infection. Dietary studies are particularly sparse. A prospective Scandinavian cohort demonstrated a protective effect of ascorbic acid (vitamin C) and beta-carotene in H. pylori-infected subjects but not in uninfected subjects (Ekstrom et al., 2000). A study by Correa and colleagues on gastric dysplasia and other preneoplastic conditions supported this finding. They observed that both H. pylori eradication therapy and dietary antioxidants (ascorbic acid and beta-carotene) prevented preneoplastic progression. Combining antioxidants and H. pylori eradication therapy, however, did not provide added benefit, suggesting that the benefit of antioxidants is limited to infected hosts. In animals, dietary salt magnifies H. pylori-associated gastric carcinogenesis (Fox et al., 2003); this finding has not been substantiated in humans. Non-dietary exposures that may alter H. pylori outcome include aspirin and non-steroidal anti-inflammatory drugs, which appear to protect against gastric cancer only in infected subjects (Zaridze et al., 1999). Smoking, on the other hand, appears to have no interactive effect with H. pylori; cigarettes increase the risk of stomach cancer equally in infected and uninfected subjects (Sivan et al., 2001). However, a recent study reported a stronger association between smoking and stomach cancer mortality among men with a history of gastroduodenal ulcer (Chao et al., 2002). One other factor that may mediate infection outcome is age at which H. pylori infection is acquired. It has been thought that infection at an early age is required for gastric cancer to develop. Only indirect data actually support the hypothesis (Blaser et al., 1995). The theory remains widely supported, however, both because of its biological plausibility (increased duration of chronic inflammation increases risk for genetic mutation in the host’s gastric epithelium over time) and the epidemiological patterns of disease (cancer occurs more frequently where childhood infection is common). Even among infected children, however, differences in gastric response to infection exist between high-risk and low-risk regions (Bedoya et al., 2003), suggesting that age at acquisition cannot completely explain important features of outcome variability. Although H. pylori infection clearly increases risk for cancers of the body and antrum, the same has not been established for cancers of the cardia. Some data indicate that these tumors are clinically and epidemiologically closer to esophageal than to gastric adenocarcinoma in that they occur more frequently in white males (Yang and Davis, 1988) and may occur in the setting of gastroesophageal reflux disease (GERD). Because H. pylori appears to protect against GERD

712

PART IV: CANCER BY TISSUE OF ORIGIN

(Raghunath et al., 2003) and possibly adenocarcinoma of the esophagus (Peek, Jr. et al., 1999; Hansen et al., 1999; Henrik et al., 2001), some have speculated that H. pylori protects against cardia tumors as well. However, a meta-analysis of the few nested case-control studies that specifically addressed this issue failed to confirm either an increase or a decrease in risk (Helicobacter and Cancer Collaborative Group, 2001). Therefore, the role of H. pylori in cardia tumors remains to be elucidated.

Diet and Alcohol In 1997, the World Cancer Research Fund (WCRF) and the American Institute for Cancer Research (AICR) published a comprehensive literature review on food, nutrition, and cancer risk (World Cancer Research Fund/American Institute for Cancer Research, 1997a). The project involved an expert panel of epidemiologists and other scientists from around the world, who evaluated scientific evidence from epidemiological studies, nutritional assessment, and biological and experimental studies to make judgments on the association between various components of diet (food groups, nutrients, beverages, and food processing) and cancer risk. The expert panel’s judgments regarding stomach cancer (World Cancer Research Fund/American Institute for Cancer Research, 1997b) are summarized in Table 37–2 and described further below. Other reviews (Kono and Hirohata, 1996; Terry et al., 2002) have come to conclusions similar to those of the WCRF/AICR panel. The evidence that diets high in vegetables and fruits protect against stomach cancer is convincing with highly consistent findings from numerous cohort and case-control studies. Another factor with convincing evidence for stomach cancer risk reduction is refrigeration, which operates indirectly by reducing the use of salt in food preservation and the risk of contamination of food with carcinogenic compounds. Consistently demonstrated protective effects of diets rich in vegetables and fruits have led to scientific investigations for identifying specific dietary constituents that account for the risk reduction and could be translated into chemopreventive agents. High consumption of several nutrients (vitamin C, carotenoids, and allium compounds) and food or beverage items (whole grain cereals and green tea) also seem to decrease stomach cancer risk. Of these likely protective factors, the evidence for vitamin C is the strongest, although it still did not reach the level of “convincing” according to the WCRF/AICR panel’s criteria. Vitamin C inhibits the intra-gastric formation of carcinogenic N-nitroso compounds and thus decreases the mutagenicity of gastric juice (Singh and Gaby, 1991). Betacarotene has received much attention as a potential chemopreventive agent because of its anti-oxidant property (Henderson et al., 1992), but the results of intervention trials have been disappointing (see below).

Allium compounds, found in such plants as onions, garlic, leeks, chives, and scallions, may affect stomach cancer risk by increasing activity of glutathione transferase (Sparnins et al., 1986) and cytochrome P450 enzymes (Guyonnet et al., 2000), which are involved in detoxification of carcinogenic compounds in diet. Allium compounds may also exert their effects through their antibacterial properties (Sivam et al., 1997) and by inhibiting the bacterial conversion of nitrate to nitrite in the stomach and thus reducing nitrosamine formation (World Cancer Research Fund/American Institute for Cancer Research, 1997b). Polyphenol extracts of green tea, notably epigallocatechin, have been shown to have anticarcinogenic effects in animals (Yang and Wang, 1993). The evidence for a possible risk reduction due to high consumption of dietary fiber, selenium, and garlic was judged to be insufficient. Among the dietary constituents that may increase stomach cancer risk, high consumption of salt and salted foods was judged as a probable risk factor by the WCRF/AICR panel (World Cancer Research Fund/American Institute for Cancer Research, 1997b). Salt itself is not carcinogenic, but intake of salt and salted foods causes mucosal damage in the stomach and induces inflammatory regenerative response, increased DNA synthesis, and cell proliferation (Charnley and Tannenbaum, 1985; Ames and Gold, 1990), which in turn facilitate a tumorigenic process caused by carcinogens. High consumption of starch possibly increases stomach cancer risk; this may be due to a direct effect of starch on the gastric mucosa or concomitant low protein intake causing increased nitrosation and decreased mucous production (Mirvish, 1983). Diets high in meat and fish that has been grilled or barbecued possibly increase stomach cancer risk. Despite a plausible biological mechanism suggested for the association (i.e., the formation of carcinogenic heterocyclic amines in meat and fish cooked at high temperature), specific data that link heterocyclic amines to gastric cancer is lacking. Although laboratory studies support a link between N-nitroso compounds and stomach cancer, the WCRF/AICR panel judged that N-nitroso compounds, found in cured meats and salted foods, may increase stomach cancer risk but that the evidence was insufficient. As mentioned above, high consumption of green tea possibly decreases stomach cancer risk. Other beverages, such as coffee and black tea, probably have no relationship with stomach cancer risk. Alcohol consumption probably does not affect overall risk of stomach cancer, but there is some evidence that alcohol intake may increase the risk of gastric cardia cancer (Terry et al., 2002).

Smoking Tobacco smoking has long been suspected for its association with increased stomach cancer risk. When the Working Group of the

Table 37–2. A Summary of Evidence for Association of Food and Nutrition with Stomach Cancer Evidence Convincing

Probable

Possible

Insufficient

Decreases Risk Vegetables and fruits Refrigeration (indirectly by reducing the use of salt and the risk of contamination) Vitamin C

Carotenoids Allium compounds Whole grain cereals Green tea Fiber Selenium Garlic

No Relationship

Alcohol (probably not related to stomach cancer as a whole, but may possibly increase risk of gastric cardia cancer) Coffee Black tea Nitrates from vegetables Sugar Vitamin E Retinol

Source: Adapted from World Cancer Research Fund/American Institute for Cancer Research (1997a).

Increases Risk

Salt Salting

Starch Grilled/barbecued meat and fish

Cured meats N-nitrosamines

Stomach Cancer International Agency for Research on Cancer (IARC) evaluated tobacco smoking as a human carcinogen in 1985 (IARC Working Group on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, 1986), the available data were not sufficient to conclude that the association between tobacco smoking and stomach cancer was causal. The IARC Working Group was convened again in 2002 and revisited tobacco smoking and cancer risk (IARC Working Group on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, 2002). With additional data accumulated since the evaluation in 1985, the IARC Working Group this time concluded that tobacco smoking increased stomach cancer risk and that confounding by other factors, such as alcohol consumption, H. pylori infection, and dietary factors, could be reasonably ruled out. The argument for the causal association, rather than confounding, was supported by a dose-response relationship of stomach cancer risk with duration of smoking and number of cigarettes smoked, as well as risk reduction with increasing duration of successful quitting. The findings from the American Cancer Society’s Cancer Prevention Study II (Chao et al., 2002), published after the 2002 IARC review, further support the conclusion of the IARC Working Group. In a 14-year follow-up of 467,788 men and 588,053 women observed from 1982 through 1996, cigarette smoking and use of other tobacco products were significantly associated with stomach cancer mortality (Chao et al., 2002). Among smoking men, a relative risk for stomach cancer death, when compared with those who had never used any tobacco products and adjusted for age, race, education, family history of stomach cancer, dietary habits, and aspirin intake, was 2.29 (95% CI: 1.49–3.51) for current cigar smokers and 2.16 (95% CI: 1.75–2.67) for current cigarette smokers, with larger relative risks observed for increasing smoking duration. The magnitude of association between cigarette smoking and stomach cancer mortality was smaller in women than in men but the risk increase was still statistically significant. Serological data on H. pylori infection were not available in this study, but the association between tobacco use and stomach mortality among men was stronger for those with a history of chronic indigestion or gastroduodenal ulcer. Smoking seems to increase the risk of both cardia and noncardia cancers (Gammon et al., 1997; Sasazuki et al., 2002). Some studies support a stronger association with smoking for gastric cardia cancer than for noncardia cancer, but this heterogeneity in smoking-related risk has not been consistent across studies (Terry et al., 2002). The 2002 IARC Working Group evaluated a possible effect of involuntary smoking (second-hand or environmental tobacco smoke) on cancer risk as well, and concluded that there was sufficient evidence for involuntary smoking causing lung cancer in humans but not for the risk of other cancers, including stomach cancer.

Medications Much interest has been manifest in the role of acid-inhibitory therapy—histamine antagonists and proton pump inhibitors (PPIs)— in gastric carcinogenesis. This interest stems from the hypothesis that bacterial overgrowth in the hypochlorydric stomach increases formation of N-nitrosoamines, a finding that has been confirmed in vivo (Vermeer et al., 2001). PPIs have been of additional concern since their use has been associated with extension of H. pylori from antrum to corpus and development of atrophic gastritis, a cancer precursor (Kuipers et al., 1996). Moreover, in animals, PPIs have been linked to development of gastric carcinoid tumors (Gillen and McColl, 2001). To date, however, there is no concrete evidence that acid inhibition does increase gastric cancer risk. Several studies indicate an increased risk of histamine antagonist use in the years proximate to cancer diagnosis but not if the medications were taken 5 years or longer prior to diagnosis (La Vecchia and Tavani, 2002). Thus, the observed relationship between tumors and histamine antagonists probably results from the medication use for symptoms related to the tumor rather than histamine antagonists being carcinogenic. There have been no longitudinal studies of PPIs and gastric cancer in humans. Despite this lack of evidence, many investigators now suggest H. pylori eradication prior to institution of PPI therapy.

713

Radiation A recent analysis of cancer incidence data from atomic bomb survivors in Japan, using a two-stage model of carcinogenesis, showed the highest excess relative risk of stomach cancer among those exposed as children (Kai et al., 1997). The model predicted that the excess relative risk declines with increasing years since exposure (Kai et al., 1997). Limited evidence exists for radiation and increased stomach cancer risk from other populations (Brown and Doll, 1965; Kohn and Fry, 1984).

Socioeconomic Status Higher rates of stomach cancer incidence have been observed among the lower socioeconomic groups (Neugut et al., 1996). This may be largely explained by risk factors strongly associated with low socioeconomic status; for example, poor sanitary conditions leading to acquisition of H. pylori infection, especially during childhood; higher smoking prevalence; and higher consumption of preserved foods and lower consumption of fresh vegetables and fruits. In the United States and other western populations, where stomach cancer screening is not recommended, the impact of delayed detection due to unequal access to medical care on stomach cancer mortality is likely to be small.

Epstein-Barr Virus A small percentage (probably less than 1%) of all gastric cancers fall histopathologically into the category of lymphoepithelioma-like carcinomas (LELCs)—epithelial tumors with intense lymphoid infiltration in the stroma. Between 80% and 100% of gastric LELCs, which are histopathologically similar in appearance to nasopharnygeal carcinomas, contain monoclonally integrated Epstein-Barr virus (EBV) (Herrmann and Niedobitek, 2003; Wu et al., 2000). Although the virus is not replicating within these tumors, it does express latent genes for nuclear RNAs (EBER1 and EBERs), a nuclear antigen (EBNA1), and the BARFO gene (Takada, 2000). In addition to their unique appearance, LELC gastric tumors have distinct oncogene expression compared with non-EBV tumors; for example, over-expression of p53 and under-expression of c-erb2 and E-cadherin (Wu et al., 2000). Given the consistency with which EBV has been found in gastric LELC, as well as the distinct histologic appearance and gene expression patterns, a strong consensus has emerged that EBV plays a causal role in this subset of gastric malignancies. The relationship between EBV and non-LELC gastric tumors is more speculative. Plasmids of monoclonal EBV have been identified in approximately 10% of non-LELC gastric adenocarcinomas. This proportion differs globally with a higher percentage (16%–18%) in the United States and Germany (Takada, 2000) and a lower proportion (2%–6%) in the United Kingdom and China (Burgess et al., 2002; Takada, 2000). Whereas LELC tumors have different gene expression from other gastric tumors, non-LELC tumors related to EBV have similar gene expression to gastric tumors without EBV. Moreover, EBV antigens can be expressed in normal gastric tissue as well as in tumors. In favor of a causal rather than incidental association, however, EBERs are expressed only in tumor cells, not normal epithelial cells, stromal cells, or lymphocytes (Yanai et al., 1997). Furthermore, EBV-associated tumors have different epidemiological features than non-EBV tumors, occurring more frequently in males and in younger patients. EBV-related tumors also typically localize in the gastric body or cardia rather than in the antrum (Takada, 2000) and are common in gastric stumps following gastric resection. Thus, although EBV appears to play a role in some gastric tumors, the attributable proportion is low despite the very high prevalence of EBV in the world’s populations. To date, no risk factors have been identified that contribute to the development of EBV-related tumors. In particular, H. pylori is equally common in cases of non-LELC gastric cancer with and without EBV (Wu et al., 2000).

714

PART IV: CANCER BY TISSUE OF ORIGIN

HOST FACTORS Genetic Susceptibility and Familial Risk A moderate to strong elevation in stomach cancer risk has been observed among those with family history of stomach cancer (Terry et al., 2002). Since many of the established and suspected risk factors for stomach cancer, such as H. pylori infection, dietary habits, and smoking, tend to aggregate among family members, excess risk associated with family history does not necessarily indicate heritable susceptibility. Nonetheless, prominent familial clustering of stomach cancer has been recognized for several decades, including a kindred of Maori ethnicity in New Zealand that presented a pattern of stomach cancer occurrence consistent with the dominant inheritance of a susceptibility gene with incomplete penetrance (Guilford et al., 1998). A linkage analysis of this and two other Maori pedigrees with early-onset, diffuse-type stomach cancer and subsequent molecular genetic studies led to a discovery that these pedigrees were segregating germ-line mutations in the E-cadherin/CDH1 gene (Guilford et al., 1998). Subsequently, germ-line mutations in the E-cadherin gene were also found in stomach cancer families of European (Gayther et al., 1998; Guilford et al., 1999), Japanese (Shinmura et al., 1999; Yabuta et al., 2002), and Korean (Yoon et al., 1999) descent. In addition to the critical role of E-cadherin in cell-cell adhesion, the biological plausibility for the involvement of the E-cadherin gene in diffuse-type stomach cancer is supported by frequent somatic mutations in the E-cadherin gene in diffuse-type but not in intestinal-type stomach cancer (Becker et al., 1994; Muta et al., 1996). Among the carriers of germ-line E-cadherin gene mutations included in the International Gastric Cancer Linkage Consortium, cumulative risk of stomach cancer by age 80 was 67% (95% CI: 39%–99%) for men and 83% (95% CI: 58%–99%) for women (Pharoah et al., 2001). Stomach cancer has also been observed as part of other hereditary cancer predisposition syndromes, including hereditary non-polyposis colorectal cancer, Li-Fraumeni syndrome, familial adenomatous polyposis, and Peutz-Jeghers syndrome (Caldas et al., 1999). A significantly increased risk of stomach cancer (standardized incidence ratio = 2.78, 95% CI: 1.59–4.52) was observed in Swedish hereditary prostate cancer families (Gronberg et al., 2000), but germ-line Ecadherin mutations do not seem to contribute to the elevated stomach cancer risk in these families (Jonsson et al., 2002). Little data are available for the population attributable risk of familial aggregation in stomach cancer (Caldas et al., 1999). Contributions of genetic makeup of the host, as opposed to environment including microbial agents, to stomach cancer risk have been investigated even in the absence of strong familial aggregation. For instance, higher incidence of stomach cancer in blood type A individuals than in those with blood type O was noticed as early as in the 1960s (Wynder et al., 1963). Prevalence of chronic atrophic gastritis, intestinal metaplasia, and dysplasia is also higher in subjects with blood type A than in type O individuals (Haenszel et al., 1976; Kneller et al., 1992). Recent studies found that adherence of H. pylori to the human gastric epithelium can be mediated by the blood-group antigenbinding adhesin (BabA) produced by H. pylori that targets human fucosylated blood group antigens H type I (type O substance) and Lewis b (Leb) (Gerhard et al., 1999; Ilver et al., 1998). The presence of the babA2 gene, encoding for BabA, in the H. pylori genome is crucial for H. pylori-related pathogenesis (Gerhard et al., 1999) and correlates with the activity of gastritis in the infected stomach (Prinz et al., 2001). This newly discovered connection between the ABO blood group and H. pylori may explain, at least in part, the old observation regarding the ABO blood group and stomach cancer risk. The establishment of H. pylori infection as a risk factor, along with advances in molecular genetic techniques, has facilitated studies of interaction between H. pylori and host genetic factors with regard to stomach cancer risk. Despite the consistent association between H. pylori infection and stomach cancer risk, only a small fraction of the infected develop stomach cancer (Parsonnet, 1999). It is plausible to hypothesize that some individuals are more susceptible to acquire persistent infection when exposed to H. pylori and to develop pre-

neoplastic lesions and eventually cancer once infection persists. Such variation in susceptibility may be due to inter-individual variability regarding response to and interaction with H. pylori. El-Omar et al. (2000) reported that IL-1 gene cluster polymorphisms, suspected of enhancing IL-1b (an important pro-inflammatory cytokine and a powerful inhibitor of gastric acid secretion), were associated with gastric cancer. Many studies have investigated a possible association of DNA sequence variants in various genes with gastric cancer risk [see a comprehensive review by Gonzalez et al. (2002)]. Besides the IL-1b gene discussed above, candidate genes examined to date include Ecadherin/CDH1 (Humar et al., 2002); MTHFR in folate metabolism (Miao et al., 2002); MUC1 in mucosal protection (Carvalho et al., 1997); CYP2E1 (Cai et al., 2001a; Gao et al., 2002), GSTM1 and T1 as metabolic enzymes (Cai et al., 2001b); XRCC1 in DNA repair (Lee et al., 2002; Shen et al., 2000); and TNFa in inflammatory response (Wu et al., 2002). Findings of these studies, however, should be interpreted with caution because of the methodological limitations, such as selection bias, limited sample size, and failure to adjust for confounders, in many of the studies (Gonzalez et al., 2002).

Body Size Obesity has been shown to increase the risk of gastric cardia cancer in some studies (IARC Working Group on the Evaluation of Cancer Preventive Strategies, 2002a), although the association seems to be weaker than that with the risk of esophageal adenocarcinoma. Body mass index (BMI) has been used as a marker of obesity in many of these studies. In a Swedish study (Lagergren et al., 1999b), BMI at 20 years before interview was statistically significantly associated with increased risk of gastric cardia cancer (OR for highest vs. lowest quartile = 2.3; 95% CI: 1.5–3.6 for men and women combined). In a US study (Chow et al., 1998), a statistically significant association was found in men (OR = 1.8; 95% CI: 1.1–2.9), but not in women (OR = 1.3; 95% CI: 0.4–4.2). Similarly, a study in China (Ji et al., 1997) found an increased risk associated with higher BMI in men only; OR = 3.0 (95% CI: 1.7–5.4) in men and OR = 1.4 (95% CI: 0.5–4.1) in women. In these two studies that estimated sex-specific ORs, the number of female subjects with gastric cardia cancer was small; 38 women in Chow et al. (1998) and 37 women in Ji et al. (1997). Therefore, the apparent heterogeneity in OR between men and women may be due to lack of statistical power for sex-specific analysis. Nevertheless, the sex difference in fat distribution in overweight and obese subjects (IARC Working Group on the Evaluation of Cancer Preventive Strategies, 2002b) may explain the possible effect modification by sex. Obesity does not seem to affect the risk of non-cardia stomach cancer (Terry et al., 2002).

Predisposing Diseases and Medical Conditions Pernicious anemia, characterized with severe atrophic gastritis and intrinsic factor deficiency, has been linked to an increased risk of gastric adenocarcinoma and carcinoid tumors (Hsing et al., 1993; Kokkola et al., 1998). A cohort study of 4517 patients with pernicious anemia in Uppsala, Sweden (Hsing et al., 1993) showed a 2.9-fold excess (95% CI: 2.4–3.5) in stomach cancer risk. Risk of cancer in the gastric remnant after resection for benign diseases has been investigated in many studies. Stalnikowicz and Benbassat (1990) reviewed 58 published studies, including case series, uncontrolled surveys, case-control studies, and cohort studies. These authors concluded that, despite study design problems in many of the reports, individuals who survived 15 years or longer after partial gastrectomy had twofold to fourfold increase in stomach cancer risk. A meta-analysis by Tersmette et al. (1990) corroborated the conclusion, although the magnitude of the association was smaller. According to the same meta-analysis (Tersmette et al., 1990), the elevated cancer risk was limited to those who had undergone partial gastrectomy for gastric ulcer (RR = 2.12; 95% CI: 1.73–2.59), whereas those whose surgery had been for duodenal ulcer did not experience the risk increase (RR = 0.84; 95% CI: 0.66–1.05).

Stomach Cancer Gastroesophageal acid reflux disease (GERD) is emerging as one of the strongest risk factors for adenocarcinoma of the esophagus, whereas reflux seems to be only weakly associated with risk of adenocarcinoma of the gastric cardia (Mayne and Navarro, 2002). A Swedish case-control study (Lagergren et al., 1999a) reported that those with recurrent symptoms of reflux had twofold risk of gastric cardia cancer (95% CI: 1.4–2.9) compared with those with no reflux symptom. Individuals with long-standing and severe symptoms of reflux had even higher risk: OR = 4.4; 95% CI: 1.7–11.0 (Lagergren et al., 1999a). A multi-institutional case-control study in the United States (Farrow et al., 2000) found no association between GERD symptoms and gastric cardia cancer, although a 5.5-fold increased risk of esophageal adenocarcinoma was observed among those with GERD symptoms. Another study (Chow et al., 1995) found a 2.1-fold increased risk (95% CI: 1.2–3.6) of adenocarcinoma of the esophagus and gastric cardia combined among those with a history of esophageal reflux.

PREVENTIVE MEASURES Primary Prevention There is considerable interest in preventing gastric cancer through treatment of H. pylori infection. The appeal of such a strategy is undeniable; H. pylori infection is curable with relatively simple antibiotic regimens. Several cost-effectiveness analyses indicate that a screenand-treat strategy could be cost-effective even if only a minority of H. pylori-associated tumors were prevented (Parsonnet et al., 1996; Fendrick et al., 1999; Roderick et al., 2003). Yet, although many studies in human subjects show that never having H. pylori prevents gastric adenocarcinoma, no randomized trial to date has shown that treating infection yields similar benefit. Initiation of such prevention programs hinges on large clinical trials, several of which are underway. In the interim, shorter-term studies have focused on the effects of H. pylori eradication on intermediate markers of gastric cancer (i.e., the preneoplastic conditions that precede malignancy) (Sung et al., 2000; Kokkola et al., 2002; Zhou et al., 2003; Correa et al., 2000). Data from these investigations have indicated a regression of preneoplasia in some patients, but the magnitude of the benefit has been insufficient to warrant broad treatment recommendations. Moreover, the age at which treatment would be most cost-effective has not been delineated. It is assumed that benefit would be greatest in people treated early in the course of progression to disease. The younger the age at which H. pylori eradication is instituted to prevent cancer, however, the less cost-effective the strategy is likely to be. Therefore, although some guidelines recommend treatment of high-risk patients, including those with a family history or with preneoplastic conditions (Malfertheiner et al., 2002), there is no general consensus about who should be treated and when. The successes of hepatitis B vaccine in preventing hepatoma have lent credibility to the prospect of vaccines against infectious diseases as effective cancer prevention tools. Yet, the quest for an H. pylori vaccine has been challenging since correlates of protective immunity are not established (Del Giudice et al., 2001). Nevertheless, such a vaccine could be cost-effective or even cost-saving should it be made successfully (Rupnow et al., 1999, 2001). According to the WCRF/AICR panel’s review of food, nutrition, and cancer risk (World Cancer Research Fund/American Institute for Cancer Research, 1997b), the most effective means of preventing stomach cancer is consumption of diets high in vegetables and fruits and low in salt, as well as industrial and domestic uses of refrigeration for perishable foods. Weight control, by reducing acid reflux and consequent mucosal damage and pathological changes, would reduce the risk of cancer of the gastric cardia and GE junction.

Chemoprevention Of the four intervention trials reviewed by the WCRF/AICR panel, one trial in a Chinese population, known to be micronutrient deficient, showed that a combination supplementation of beta-carotene, vitamin

715

E, and selenium was associated with a statistically significant reduction in stomach cancer mortality: RR = 0.87; 95% CI: 0.64–0.99 (Blot et al., 1993). The protective effect of this supplementation regimen in a population that is not deficient for these micronutrients is not clear. The other trials in China (Li et al., 1993), Finland (Malila et al., 2002), and the United States (Hennekens et al., 1996) showed no significant protective effect of nutrient supplementation on stomach cancer risk. Lack of support from these intervention trials for protective effects of specific micronutrients is in sharp contrast with convincing evidence for reduced stomach cancer risk associated with high consumption of vegetables and fruits (World Cancer Research Fund/American Institute for Cancer Research, 1997b). As a cancer preventive measure, encouraging consumption of vegetables and fruits seems more effective than depending on specific nutrient supplementation. As has been claimed about several gastrointestinal malignancies, long-term use of non-steroidal anti-inflammatory drugs (NSAIDs) may protect against gastric adenocarcinoma. In a case-control study conducted in Sweden, increasing use of aspirin provided increasing protection against gastric malignancy (overall OR = 0.7; 95% CI: 0.6–1.0); this finding was most evident in people with H. pylori infection (Akre et al., 2001). In a second case-control study, greater than 5 years of NSAID use strongly protected against gastric malignancy: OR = 0.2; 95% CI: 0.1–0.7 (Coogan et al., 2000). Despite the strength of these findings, however, these studies could not exclude the possibility of confounding, since dyspepsia and underlying gastric disease could have diminished NSAID use in cancer cases. Yet, a case-control study that restricted analysis to subjects without dyspepsia and with use of NSAIDs greater than 5 years prior to diagnosis, yielded similar protection for NSAIDs, supporting a true association (Farrow et al., 1998); this protection was observed only for non-cardia tumors. The only cohort study specifically examining the role of NSAIDs in gastric tumors found a somewhat reduced incidence of cancer in heavy NSAID users, although the finding was not statistically significant (Friis et al., 2003). Based on these findings, use of NSAIDs to prevent gastric cancer cannot be recommended currently and must await the results of randomized clinical trials.

Screening and Early Detection Most data on stomach cancer screening come from Japan, where stomach cancer incidence and mortality have been the highest in the world. The first mass screening program for early detection of stomach cancer in Japan started in 1960 (Watanabe and Fukao, 2001). By 1975, nearly all municipalities in Japan had been offering stomach cancer screening to their residents. The most commonly used screening method has been the double-contrast barium X-ray technique, currently performed with at least seven standard views. More recently, endoscopy, serological tests of pepsinogen concentrations, and anti-H. pylori antibodies have been introduced as additional or alternative modalities of stomach cancer screening. Studies to evaluate the effect of mass screening on reduction in stomach cancer mortality have been limited to those on the barium Xray test that has the longest history (Watanabe and Fukao, 2001). Because mass screening programs for stomach cancer have already been widespread in Japan, it has been impossible to evaluate their effectiveness through randomized clinical trials. Observational studies— including time-trend analysis, cohort studies, case-control studies, and ecological analysis of municipalities with variable rates of screening program participation—generally support the conclusion that the widespread application of screening in Japan contributed to a fall in stomach cancer mortality (Chamberlain et al., 1986; Miller et al., 1990). Moreover, a non-randomized trial has suggested that treatment of H. pylori infection in patients with early gastric cancer can prevent tumor recurrence after mucosal resection (Uemura et al., 1997). However, the possibility of self-selection bias (i.e., people who undergo screening tend to be at lower risk of dying from stomach cancer than those who are not screened) cannot be ruled out (Chamberlain et al., 1986). A recent study in Japan, based in regional hospital cancer registries, showed that the sensitivity of gastroscopy for early stomach cancer detection was only 81% (Hosokawa et al., 1998). Whether endoscopy

716

PART IV: CANCER BY TISSUE OF ORIGIN

Table 37–3. Recommendations by the Stomach/Esophageal Cancers Progress Review Group (US National Cancer Institute) for Future Research Priorities Target Area Population studies* Prevention* Patient/provider education Therapy Therapeutic targets Markers and molecular profiling* Outcomes* Host/environment interactions* Technologies for screening/surveillance* Preclinical models

Recommendation Establish collaborations for conducting interdisciplinary, population-based, endoscopic, multi-institutional studies to identify populations at greatest risk and to determine the prevalence and natural history of preneoplastic lesions. Develop prevention strategies based on the mechanisms of host/environment interaction that lead to metaplasia and neoplasia. Evaluate their effectiveness in at-risk populations. Educate patients and their families, health care professionals, and the public regarding risk factors, risk reduction, and treatment options and outcomes for cancers and their precursor states. Develop and test novel therapeutics, and optimize existing treatments for cancers and their precursors, based on the identification and understanding of molecular pathways involved in oncogenesis, tumor response, and resistance. Define host and molecular/biologic tumor characteristics that will help customize treatment and best predict recurrence and/or survival. Profile the molecular, cellular, and epidemiologic features of tumors and their precursor lesions to identify diagnostic, prognostic, predictive, preventive, and therapeutic targets. Develop and refine disease-specific, patient-oriented methods to assess quality of life, quality of care, and cost effectiveness of treatment in patients with cancers and their precursors through all stages of disease and treatment, and include these instruments in clinical trials and observational studies. Identify, develop, and validate genetic, biochemical, and biological markers that will help uncover host/environment interactions in carcinogenesis. Develop noninvasive and minimally invasive technologies (e.g., serum markers and imaging techniques) for screening and surveillance of premalignant and malignant lesions. Establish models to understand the biology of cancers and their precursor lesions and to stimulate novel prevention, diagnostic, and treatment strategies.

Source: Adapted from National Cancer Institute (2002b). *Target areas relevant to epidemiology and prevention research.

is used as a primary screening modality or for a follow-up of positive findings in other screening tests, data on screening efficacy in Japan must be interpreted with caution, because the apparently high rate of early stomach cancer in Japan is due in part to the different classification methods used by Japanese and Western pathologists (Lambert, 1998). The immunological test with the fetal sulphoglycoprotein antigen (FSA) used in Finland has a higher specificity than the barium Xray test (91% vs. 86.5%) but a lower sensitivity (73% vs. 85%) (Chamberlain et al., 1986). The effect of the FSA test and other screening modalities on mortality reduction remains to be evaluated. Populations with high stomach cancer incidence have high prevalence of H. pylori infection, and serological tests for the infection will not be specific enough as a cancer screening method. The UICC Project on the Evaluation of Screening for Cancer (Miller et al., 1990) concluded that screening programs should continue in those regions with high stomach cancer incidence where the programs are already being implemented, but that stomach cancer screening could not be recommended as public health policy in other countries. Neither the American Cancer Society (Smith et al., 2003) nor the National Cancer Institute (2002c) currently recommends stomach cancer screening in the United States.

FUTURE DIRECTIONS Stomach cancer is a moving target for research because of its declining trends in incidence and mortality. Nonetheless, the aging of populations would make up for the downward trends in age-standardized rates with respect to the absolute number of stomach cancer cases and deaths for the time being. Poor prognosis of stomach cancer reinforces the need to not only understand its etiology and pathogenesis but also develop and implement effective prevention and treatment strategies. Several lines of epidemiologic research, ranging from descriptive to analytical to intervention studies, would be important. In the area of descriptive epidemiology, the increasing trends of GE junction and gastric cardia cancers, most noticeable among US white men at this time, should be monitored over time. Analytical epidemiology through case-control and cohort studies should further elucidate the roles of obesity/overweight and H. pylori infection in the development of GE junction and gastric cardia cancers. For noncardia gastric cancer, it is important to investigate possible interaction between host factors,

including genetic makeup, and H. pylori infection. The effects of dietary and other risk factors, with a simultaneous consideration of H. pylori infection, would be best examined in nested case-control studies. Such studies could be conducted efficiently only in high-risk populations due to the sample size constraints. As a part of a series of progress reviews for various cancer sites, the US National Cancer Institute (2002b) convened a progress review group (PRG) for stomach and esophageal cancers. The charge of the PRG was to develop a national plan for stomach and esophageal cancer research over the next 5 years based on a review of progress and opportunities. Major challenges that the PRG identified for future research of stomach cancer (and esophageal cancer) were low incidence, lack of a coordinated multi-institutional network system for collaborative research, and lack of awareness by the public and some physicians. The PRG made general recommendations regarding future research needs in 10 target areas several of which are relevent to epidemiology and prevention (Table 37–3). For the purpose of observing pathological or molecular genetic changes of the mucosa longitudinally, the stomach provides a unique opportunity for research because repeat access to preneoplastic lesions for endoscopic visualization and biopsy is possible; those precancerous lesions, with an exception of dysplasia, are not removed as adenomatous polyps of the colon would be. This feature should be exploited in not only investigating biological processes involved in carcinogenesis but also evaluating the effect of preventive and therapeutic intervention to delay progression of preneoplasia to invasive cancer. References Akre K, Ekstrom AM, Signorello LB, Hansson LE, Nyren O. 2001. Aspirin and risk for gastric cancer: a population-based case-control study in Sweden. Br J Cancer 84:965–968. Allgayer H, Heiss MM, Schildberg FW. 1997. Prognostic factors in gastric cancer. Br J Surg 84:1651–1664. American Joint Committee on Cancer. 2002. Stomach Cancer. In: American Joint Committee on Cancer, ed. AJCC Cancer Staging Manual. New York: Springer-Verlag, pp. 99–106. Ames BN, Gold LS. 1990. Too many rodent carcinogens: Mitogenesis increases mutagenesis [published erratum appears in Science (1990) 249:1487]. Science 249:970–971. Aras RA, Small AJ, Ando T, Blaser MJ. 2002. Helicobacter pylori interstrain restriction-modification diversity prevents genome subversion by chromosomal DNA from competing strains. Nucleic Acids Research 30:5391–5397.

Stomach Cancer Becker K-F, Atkinson MJ, Reich U, et al. 1994. E-cadherin gene mutations provide clues to diffuse type gastric carcinomas. Cancer Res 54: 3845–3852. Bedoya A, Garay J, Sanzon F, Bravo LE, Bravo JC, Correa H, et al. 2003. Histopathology of gastritis in Helicobacter pylori-infected children from populations at high and low gastric cancer risk. Human Pathology 34:206–213. Blaser MJ, Berg DE. 2001. Helicobacter pylori genetic diversity and risk of human disease. J Clin Invest 107:767–773. Blaser MJ, Chyou PH, Nomura A. 1995. Age at establishment of Helicobacter pylori infection and gastric carcinoma, gastric ulcer, and duodenal ulcer risk. Cancer Res 55:562–565. Blot WJ, Li JY, Taylor PR, et al. 1993. Nutrition intervention trials in Linxian, China: supplementation with specific vitamin/mineral combinations, cancer incidence, and disease-specific mortality in the general population. J Natl Cancer Inst 85:1483–1492. Brown LM, Devesa SS. 2002. Epidemiologic trends in esophageal and gastric cancer in the United States. Surg Oncol Clin N Am 11:235–256. Brown WM, Doll R. 1965. Mortality from cancer and other causes after radiotherapy for ankylosing spondylitis. Br Med J 5474:1327–1332. Burgess DE, Woodman CB, Flavell KJ, Rowlands DC, Crocker J, Scott K, et al. 2002. Low prevalence of Epstein-Barr virus in incident gastric adenocarcinomas from the United Kingdom. Br J Cancer 86:702–704. Cai L, Yu SZ, Zhan ZF. 2001a. Cytochrome P450 2E1 genetic polymorphism and gastric cancer in Changle, Fujian Province. World J Gastroenterol 7:792–795. Cai L, Yu SZ, Zhang ZF. 2001b. Glutathione S-transferases M1, T1 genotypes and the risk of gastric cancer: A case-control study. World J Gastroenterol 7:506–509. Caldas C, Carneiro F, Lynch HT, et al. 1999. Familial gastric cancer: Overview and guidelines for management. J Med Genet 36:873–880. Carvalho F, Seruca R, David L, et al. 1997. MUC1 gene polymorphism and gastric cancer—an epidemiological study. Glycoconj J 14:107–111. Chamberlain J, Day NE, Hakama M, Miller AB, Prorok PC. 1986. UICC workshop of the Project on Evaluation of Screening Programmes for Gastrointestinal Cancer. Int J Cancer 37:329–334. Chao A, Thun MJ, Henley SJ, Jacobs EJ, McCullough ML, Calle EE. 2002. Cigarette smoking, use of other tobacco products and stomach cancer mortality in US adults: The Cancer Prevention Study II. Int J Cancer 101:380–389. Charnley G, Tannenbaum SR. 1985. Flow cytometric analysis of the effect of sodium chloride on gastric cancer risk in the rat. Cancer Res 45:5608–5616. Chow WH, Blot WJ, Vaughan TL, et al. 1998. Body mass index and risk of adenocarcinomas of the esophagus and gastric cardia. J Natl Cancer Inst 90:150–155. Chow WH, Finkle WD, McLaughlin JK, Frankl H, Ziel HK, Fraumeni JF Jr. 1995. The relation of gastroesophageal reflux disease and its treatment to adenocarcinomas of the esophagus and gastric cardia. JAMA 274: 474–477. Coggon D, Osmond C, Barker DJ. 1990. Stomach cancer and migration within England and Wales. Br J Cancer 61:573–574. Coleman MP, Esteve J, Damiecki P, Arslan A, Renard H. 1993. Stomach. In: Coleman MP, Esteve J, Damiecki P, Arslan A, Renard H, eds. Trends in Cancer Incidence and Mortality. Lyon: International Agency for Research on Cancer, pp. 193–224. Coogan PF, Rosenberg L, Palmer JR, Strom BL, Zauber AG, Stolley PD, et al. 2000. Nonsteroidal anti-inflammatory drugs and risk of digestive cancers at sites other than the large bowel. Cancer Epidemiol Biomark Prev 9:119–123. Correa P. 1992. Human gastric carcinogenesis: A multistep and multifactorial process—First American Cancer Society Award Lecture on Cancer Epidemiology and Prevention. Cancer Res 52:6735–6740. Correa P. 2002. Stomach. In: Franco EL, Rohan TE, eds. Cancer precursors: Epidemiology, detection, and prevention. New York: Springer-Verlag, pp. 117–126. Correa P, Chen VW. 1994. Gastric Cancer. In: Doll R, Fraumeni JF, Jr., Muir CS, eds. Trends in Cancer Incidence and Mortality. Plainview, NY: Cold Spring Harbor Laboratory Press, pp. 55–75. Correa P, Cuello C, Duque E, et al. 1976. Gastric cancer in Colombia. III. Natural history of precursor lesions. J Natl Cancer Inst 57:1027–1035. Correa P, Fontham ET, Bravo JC, Bravo LE, Ruiz B, Zarama G, et al. 2000. Chemoprevention of gastric dysplasia: randomized trial of antioxidant supplements and anti-helicobacter pylori therapy. J Natl Cancer Inst 92:1881–1888. Del Giudice G, Covacci A, Telford JL, Montecucco C, Rappuoli R. 2001. The design of vaccines against Helicobacter pylori and their development. Annual Rev Immunol 19:523–563.

717

Devesa SS, Blot WJ, Fraumeni JF Jr. 1998. Changing patterns in the incidence of esophageal and gastric carcinoma in the United States. Cancer 83:2049–2053. Dixon MF, Genta RM, Yardley JH, Correa P. 1996. Classification and grading of gastritis. The updated Sydney System. International Workshop on the Histopathology of Gastritis, Houston 1994. Am J Surg Pathol 20:1161–1181. Ekstrom AM, Serafini M, Nyren O, Hansson LE, Ye W, Wolk A. 2000. Dietary antioxidant intake and the risk of cardia cancer and noncardia cancer of the intestinal and diffuse types: a population-based case-control study in Sweden. Int J Cancer 87:133–140. El-Omar EM, Carrington M, Chow WH, et al. 2000. Interleukin-1 polymorphisms associated with increased risk of gastric cancer. Nature 404:398–402. Farrow DC, Vaughan TL, Hansten PD, Stanford JL, Risch HA, Gammon MD, et al. 1998. Use of aspirin and other nonsteroidal anti-inflammatory drugs and risk of esophageal and gastric cancer. Cancer Epidemiol Biomark Prev 7:97–102. Farrow DC, Vaughan TL, Sweeney C, et al. 2000. Gastroesophageal reflux disease, use of H2 receptor antagonists, and risk of esophageal and gastric cancer. Cancer Causes Control 11:231–238. Fascioli S, Capocaccia R, Mariotti S. 1995. Cancer mortality in migrant populations within Italy. Int J Epidemiol 24:8–18. Fendrick AM, Chernew ME, Hirth RA, Bloom BS, Bandekar RR, Scheiman JM. 1999. Clinical and economic effects of population-based Helicobacter pylori screening to prevent gastric cancer. Arch Intern Med 159:142–148. Fox JG, Rogers AB, Ihrig M, Taylor NS, Whary MT, Dockray G, et al. 2003. Helicobacter pylori-associated gastric cancer in INS-GAS mice is gender specific. Cancer Res 63:942–950. Friis S, Sorensen HT, McLaughlin JK, Johnsen SP, Blot WJ, Olsen JH. 2003. A population-based cohort study of the risk of colorectal and other cancers among users of low-dose aspirin. Br J Cancer 88:684–688. Gammon MD, Schoenberg JB, Ahsan H, et al. 1997. Tobacco, alcohol, and socioeconomic status and adenocarcinomas of the esophagus and gastric cardia. J Natl Cancer Inst 89:1277–1284. Gao C, Takezaki T, Wu J, et al. 2002. Interaction between cytochrome P-450 2E1 polymorphisms and environmental factors with risk of esophageal and stomach cancers in Chinese. Cancer Epidemiol Biomarkers Prev 11:29–34. Gayther SA, Gorringe KL, Ramus SJ, et al. 1998. Identification of germ-line E-cadherin mutations in gastric cancer families of European origin. Cancer Res 58:4086–4089. Genta RM, Graham DY. 1993. Intestinal metaplasia, not atrophy or achlorhydria, creates a hostile environment for Helicobacter pylori. Scand J Gastroenterol 28:924–928. Gerhard M, Lehn N, Neumayer N, et al. 1999. Clinical relevance of the Helicobacter pylori gene for blood-group antigen-binding adhesin. Proc Natl Acad Sci USA 96:12778–12783. Gillen D, McColl KE. 2001. Problems associated with the clinical use of proton pump inhibitors. Pharmacol Toxicol 89:281–286. Gonzalez CA, Sala N, Capella G. 2002. Genetic susceptibility and gastric cancer risk. Int. J Cancer 100:249–260. Gronberg H, Bergh A, Damber JE, Emanuelsson M. 2000. Cancer risk in families with hereditary prostate carcinoma. Cancer 89:1315–1321. Guilford P, Hopkins J, Harraway J, et al. 1998. E-cadherin germline mutations in familial gastric cancer. Nature 392:402–405. Guilford PJ, Hopkins JB, Grady WM, et al. 1999. E-cadherin germline mutations define an inherited cancer syndrome dominated by diffuse gastric cancer. Hum Mutat 14:249–255. Guyonnet D, Belloir C, Suschetet M, Siess MH, Le Bon AM. 2000. Liver subcellular fractions from rats treated by organosulfur compounds from Allium modulate mutagen activation. Mutat Res 466:17–26. Haenszel W, Correa P, Cuello C, et al. 1976. Gastric cancer in Colombia. II. Case–control epidemiologic study of precursor lesions. J Natl Cancer Inst 57:1021–1026. Hamilton SR, Aaltonen LA. 2000a. Tumours of the oesophagogastric junction. In: Hamilton SR, Aaltonen LA, eds. World Health Organization Classification of Tumours: Pathology of Tumours of the Digestive System. Lyon: IARC Press, pp. 31–36. Hamilton SR, Aaltonen LA. 2000b. Tumours of the stomach. In: Hamilton SR, Aaltonen LA, eds. World Health Organization Classification of Tumours: Pathology of Tumours of the Digestive System. Lyon: IARC Press, pp. 37–66. Hanai A, Fujimoto I, Taniguchi H. 1982. Trends of stomach cancer incidence and histologic types in Osaka. In: Magnus K, ed. Trends in Cancer Incidence: Causes and Practical Implications. Washington: Hemisphere Publishing Corporation, pp. 143–154.

718

PART IV: CANCER BY TISSUE OF ORIGIN

Hansen S, Melby KK, Aase S, Jellum E, Vollset SE. 1999. Helicobacter pylori infection and risk of cardia cancer and non-cardia gastric cancer. A nested case-control study. Scand J Gastroenterol 34:353–360. Helicobacter and Cancer Collaborative Group. 2001. Gastric cancer and Helicobacter pylori: a combined analysis of 12 case control studies nested within prospective cohorts. Gut 49:347–353. Henderson BE, Ross RK, Shibata A, Paganini-Hill A, Yu MC. 1992. Environmental carcinogens and anticarcinogens. In: Wattenberg L, Lipkin M, Boone CW, eds. Cancer chemoprevention. Boca Raton: CRC Press, pp. 3–17. Hennekens CH, Buring JE, Manson JE, et al. 1996. Lack of effect of long-term supplementation with beta carotene on the incidence of malignant neoplasms and cardiovascular disease. N Engl J Med 334:1145–1149. Henrik SJ, Forsgren A, Berglund G, Floren CH. 2001. Helicobacter pylori infection is associated with a decreased risk of developing oesophageal neoplasms. Helicobacter 6:310–316. Herrmann K, Niedobitek G. 2003. Epstein-Barr virus-associated carcinomas: facts and fiction. J Pathol 199:140–145. Hosokawa O, Tsuda S, Kidani E, et al. 1998. Diagnosis of gastric cancer up to three years after negative upper gastrointestinal endoscopy. Endoscopy 30:669–674. Howson CP, Hiyama T, Wynder EL. 1986. The decline in gastric cancer: Epidemiology of an unplanned triumph. Epidemiol Rev 8:1–27. Hsing AW, Hansson LE, McLaughlin JK, et al. 1993. Pernicious anemia and subsequent cancer. A population-based cohort study. Cancer 71: 745–750. Huang JQ, Sridhar S, Chen Y, Hunt RH. 1998. Meta-analysis of the relationship between Helicobacter pylori seropositivity and gastric cancer. Gastroenterology 114:1169–1179. Humar B, Graziano F, Cascinu S, et al. 2002. Association of CDH1 haplotypes with susceptibility to sporadic diffuse gastric cancer. Oncogene 21:8192–8195. IARC Working Group on the Evaluation of Cancer Preventive Strategies. 2002a. Cancer-preventive effects. In: Vainio H, Bianchini F, eds. IARC Handbook of Cancer Prevention: Weight Control and Physical Activity. Lyon: IARC Press, pp. 83–199. IARC Working Group on the Evaluation of Cancer Preventive Strategies. 2002b. Characteristics of weight control and physical activity. In: Vainio H, Bianchini F, eds. IARC Handbook of Cancer Prevention: Weight Control and Physical Activity. Lyon: IARC Press, pp. 1–12. IARC Working Group on the Evaluation of Carcinogenic Risks to Humans. 1994. Infection with Helicobacter Pylori. In: International Agency for Research on Cancer (IARC), ed. Schistosomes, Liver Flukes, and Helicobactor pylori. Lyon: International Agency for Research on Cancer, pp. 177–241. IARC Working Group on the Evaluation of the Carcinogenic Risk of Chemicals to Humans. 1986. Tobacco Smoking. International Agency for Research on Cancer, Lyon. IARC Working Group on the Evaluation of the Carcinogenic Risk of Chemicals to Humans. 2002. Tobacco Smoking and Tobacco Smoke. International Agency for Research on Cancer, Lyon. Ilver D, Arnqvist A, Ogren J, et al. 1998. Helicobacter pylori adhesin binding fucosylated histo-blood group antigens revealed by retagging. Science 279:373–377. Israel DA, Salama N, Arnold CN, Moss SF, Ando T, Wirth HP, et al. 2001. Helicobacter pylori strain-specific differences in genetic content, identified by microarray, influence host inflammatory responses. J Clin Invest 107:611–620. Jemal A, Murray T, Ward E, et al. 2005. Cancer statistics, 2003. CA Cancer J Clin 55:11. Ji BT, Chow WH, Yang G, et al. 1997. Body mass index and the risk of cancers of the gastric cardia and distal stomach in Shanghai, China. Cancer Epidemiol Biomarkers Prev 6:481–485. Jonsson BA, Bergh A, Stattin P, Emmanuelsson M, Gronberg H. 2002. Germline mutations in E-cadherin do not explain association of hereditary prostate cancer, gastric cancer and breast cancer. Int J Cancer 98:838–843. Junqueira LC, Carneiro J, Kelley RO. 1998. Digestive tract. In: Junqueira LC, Carneiro J, Kelley RO, eds. Basic Histology. Stamford, CT: Appleton & Lange, pp. 272–303. Kai M, Luebeck EG, Moolgavkar SH. 1997. Analysis of the incidence of solid cancer among atomic bomb survivors using a two-stage model of carcinogenesis. Radiat Res 148:348–358. Kneller RW, You WC, Chang YS, et al. 1992. Cigarette smoking and other risk factors for progression of precancerous stomach lesions. J Natl Cancer Inst 84:1261–1266. Kohn HI, Fry RJ. 1984. Radiation carcinogenesis. N Engl J Med 310:504– 511.

Kokkola A, Sipponen P, Rautelin H, Harkonen M, Kosunen TU, Haapiainen R, et al. 2002. The effect of Helicobacter pylori eradication on the natural course of atrophic gastritis with dysplasia. Aliment Pharmacol Ther 16:515–520. Kokkola A, Sjoblom SM, Haapiainen R, Sipponen P, Puolakkainen P, Jarvinen H. 1998. The risk of gastric carcinoma and carcinoid tumours in patients with pernicious anaemia. A prospective follow-up study. Scand J Gastroenterol 33:88–92. Kolonel LN, Hankin JH, Nomura AM. 1986. Multiethnic studies of diet, nutrition, and cancer in Hawaii. In: Hayashi Y, Nagao M, Sugimura T, Takayama S, Tomatis L, Wattenberg LW, Wogan GN, eds. Diet, Nutrition, and Cancer: Proceedings of the 16th International Symposium of the Princess Takamatsu Cancer Research Fund. Japan Scientific Societies Press/VNU Science Press BV, Tokyo/Utrecht, pp. 29–40. Kono S, Hirohata T. 1996. Nutrition and stomach cancer. Cancer Causes Control 7:41–55. Kreinitz W. 1906. Ueber das Auftreten von Spirochaeten verschiedener Form im Mageninhalt bei Carcinoma ventriculi. Deutsche Medizinische Wochenschrift 32:872. Kuipers EJ, Lundell L, Klinkenberg-Knol EC, Havu N, Festen HP, Liedman B, et al. 1996. Atrophic gastritis and Helicobacter pylori infection in patients with reflux esophagitis treated with omeprazole or fundoplication. N Engl J Med 334:1018–1022. La Vecchia C, Tavani A. 2002. A review of epidemiological studies on cancer in relation to the use of anti-ulcer drugs. European Journal of Cancer Prevention 11:117–123. Lagergren J, Bergstrom R, Lindgren A, Nyren O. 1999a. Symptomatic gastroesophageal reflux as a risk factor for esophageal adenocarcinoma. N Engl J Med 340:825–831. Lagergren J, Bergstrom R, Nyren O. 1999b. Association between body mass and adenocarcinoma of the esophagus and gastric cardia. Ann Intern Med 130:883–890. Lambert R. 1998. Mass screening programs in Japan: What can we learn in the West? Endoscopy 30:721–723. Laurén P. 1965. The two histological main types of gastric carcinoma: Diffuse and so-called intestinal-type carcinoma. Acta Path et Microbiol Scandinav 64:31–49. Lee SG, Kim B, Choi J, Kim C, Lee I, Song K. 2002. Genetic polymorphisms of XRCC1 and risk of gastric cancer. Cancer Lett 187:53–60. Leung SY, Chen X, Chu KM, et al. 2002. Phospholipase A2 group IIA expression in gastric adenocarcinoma is associated with prolonged survival and less frequent metastasis. Proc Natl Acad Sci USA 99: 16203–1628. Lewin KJ, Appelman HD. 1995. Atlas of Tumor Pathology. Washington, DC: Armed Forces Institute of Pathology. Li JY, Taylor PR, Li B, et al. 1993. Nutrition intervention trials in Linxian, China: Multiple vitamin/mineral supplementation, cancer incidence, and disease-specific mortality among adults with esophageal dysplasia. J Natl Cancer Inst 85:1492–1498. Lundegardh G, Lindgren A, Rohul A, et al. 1991. Intestinal and diffuse types of gastric cancer: secular trends in Sweden since 1951. Br J Cancer 64:1182–1186. Machado JC, Figueiredo C, Canedo P, et al. 2003. A proinflammatory genetic profile increases the risk for chronic atrophic gastritis and gastric carcinoma. Gastroenterology 125:364–371. Malfertheiner P, Megraud F, O’Morain C, Hungin AP, Jones R, Axon A, et al. 2002. Current concepts in the management of Helicobacter pylori infection—the Maastricht 2–2000 Consensus Report. Aliment Pharmacol Ther 16:167–180. Malila N, Taylor PR, Virtanen MJ, et al. 2002. Effects of alpha-tocopherol and beta-carotene supplementation on gastric cancer incidence in male smokers (ATBC Study, Finland). Cancer Causes Control 13:617–623. Masci E, Viale E, Freschi M, Porcellati M, Tittobello A. 1996. Precancerous gastric lesions and Helicobacter pylori. Hepatogastroenterology 43:854–858. Mayne ST, Navarro SA. 2002. Diet, obesity and reflux in the etiology of adenocarcinomas of the esophagus and gastric cardia in humans. J Nutr 132:3467S–3470S. McMichael AJ, McCall MG, Hartshorne JM, Woodings TL. 1980. Patterns of gastro-intestinal cancer in European migrants to Australia: The role of dietary change. Int J Cancer 25:431–437. Miao X, Xing D, Tan W, Qi J, Lu W, Lin D. 2002. Susceptibility to gastric cardia adenocarcinoma and genetic polymorphisms in methylenetetrahydrofolate reductase in an at-risk chinese population. Cancer Epidemiol Biomarkers Prev 11:1454–1458. Miller AB, Chamberlain J, Day NE, Hakama M, Prorok PC. 1990. Report on a Workshop of the UICC Project on Evaluation of Screening for Cancer. Int J Cancer 46:761–769.

Stomach Cancer Mirvish SS. 1983. The etiology of gastric cancer. Intragastric nitrosamide formation and other theories. J Natl Cancer Inst 71:629–647. Misumi A, Murakami A, Harada K, Baba K, Akagi M. 1989. Definition of carcinoma of the gastric cardia. Langenbecks Arch Chir 374:221– 226. Muta H, Noguchi M, Kanai Y, Ochiai A, Nawata H, Hirohashi S. 1996. Ecadherin gene mutations in signet ring cell carcinoma of the stomach. Jpn J Cancer Res 87:843–848. National Cancer Institute. 2002a. Gastric cancer (PDQ®): Treatment (health professional version). Available at: http://www.nci.nih.gov/cancerinfo/ pdq/treatment/gastric/healthprofessional/ Accessed 10–9-2002. National Cancer Institute. 2002b. Report of the stomach/esophageal cancers progress review group. Bethesda, MD: National Cancer Institute. National Cancer Institute. 2002c. Gastric cancer (PDQ®): Screening (health professional version). Available at: http://www.nci.nih.gov/cancerinfo/ pdq/screening/gastric/healthprofessional/ Accessed 10–10–2002. Neugut AI, Hayek M, Howe G. 1996. Epidemiology of gastric cancer. Semin Oncol 23:281–291. Osawa H, Inoue F, Yoshida Y. 1996. Inverse relation of serum Helicobacter pylori antibody titres and extent of intestinal metaplasia. J Clin Pathol 49:112–115. Parkin DM, Pisani P, Ferlay J. 1999. Global cancer statistics. CA Cancer J Clin 49:33–64. Parsonnet J. 1999. Helicobacter pylori and gastric adenocarcinoma. In: Parsonnet J, ed. Microbes and Malignancy: Infection as a Cause of Human Cancer. New York, Oxford University Press, pp. 372–408. Parsonnet J, Friedman GD, Orentreich N, Vogelman JH. 1997. Risk for gastric cancer in persons with CagA positive and CagA negative Helicobacter pylori infection. Gut 40:297–301. Parsonnet J, Harris RA, Hack HM, Owens DK. 1996. Modelling costeffectiveness of Helicobacter pylori screening to prevent gastric cancer: a mandate for clinical trials. Lancet 348:150–154. Peek RM, Jr, Vaezi MF, Falk GW, Goldblum JR, Perez-Perez GI, Richter JE, et al. 1999. Role of Helicobacter pylori cagA(+) strains and specific host immune responses on the development of premalignant and malignant lesions in the gastric cardia. Int J Cancer 82:520–524. Pharoah PD, Guilford P, Caldas C. 2001. Incidence of gastric cancer and breast cancer in CDH1 (E-cadherin) mutation carriers from hereditary diffuse gastric cancer families. Gastroenterology 121:1348–1353. Prinz C, Schoniger M, Rad R, et al. 2001. Key importance of the Helicobacter pylori adherence factor blood group antigen binding adhesin during chronic gastric inflammation. Cancer Res 61:1903–1909. Raghunath A, Hungin AP, Wooff D, Childs S. 2003. Prevalence of Helicobacter pylori in patients with gastro-oesophageal reflux disease: systematic review. BMJ 326:737–743. Ries LAG, Eisner MP, Kosary CL, et al. 2002. SEER Cancer Statistics Review, 1973–1999. Bethesda, MD: National Cancer Institute. Ries LAG, Eisner MP, Kosary CL, et al. 2003. SEER Cancer Statistics Review, 1975–2000, National Cancer Institute. Bethesda, MD. Available at http://seer.cancer.gov/csr/1975_2000. Roderick P, Davies R, Raftery J, Crabbe D, Pearce R, Bhandari P, et al. 2003. The cost-effectiveness of screening for Helicobacter pylori to reduce mortality and morbidity from gastric cancer and peptic ulcer disease: a discrete-event simulation model. Health Technology Assessment (Winchester, England) 7:1–86. Rugge M, Correa P, Dixon MF, et al. 2000. Gastric dysplasia: The Padova international classification. Am J Surg Pathol 24:167–176. Rupnow MF, Owens DK, Shachter R, Parsonnet J. 1999. Helicobacter pylori vaccine development and use: a cost-effectiveness analysis using the Institute of Medicine Methodology. Helicobacter 4:272–280. Rupnow MFT, Shachter RD, Owens DK, Parsonnet J. 2001. A Vaccine for Ulcers and Gastric Cancer: A Rationale for an H. pylori Vaccine. The Fourth Annual Conference on Vaccine Research. Arlington, VA. Arlington, VA. http://www.nfid.org/conferences/vaccine01/abstracts/ abs9-12.pdf. Sasazuki S, Sasaki S, Tsugane S. 2002. Cigarette smoking, alcohol consumption and subsequent gastric cancer risk by subsite and histologic type. Int. J Cancer 101:560–566. Segal ED, Lange C, Covacci A, Tompkins LS, Falkow S. 1997. Induction of host signal transduction pathways by Helicobacter pylori. Proc Natl Acad Sci USA 94:7595–7599. Shen H, Xu Y, Qian Y, et al. 2000. Polymorphisms of the DNA repair gene XRCC1 and risk of gastric cancer in a Chinese population. Int J Cancer 88:601–606. Shinmura K, Kohno T, Takahashi M, et al. 1999. Familial gastric cancer: Clinicopathological characteristics, RER phenotype and germline p53 and E-cadherin mutations. Carcinogenesis 20:1127–1131.

719

Siman JH, Forsgren A, Berglund G, Floren CH. 2001. Tobacco smoking increases the risk for gastric adenocarcinoma among Helicobacter pyloriinfected individuals. Scand J Gastroenterol 36:208–213. Simon R, Radmacher MD, Dobbin K, McShane LM. 2003. Pitfalls in the use of DNA microarray data for diagnostic and prognostic classification. J Natl Cancer Inst 95:14–18. Singh VN, Gaby SK. 1991. Premalignant lesions: Role of antioxidant vitamins and beta-carotene in risk reduction and prevention of malignant transformation. Am J Clin Nutr 53:386S–390S. Siurala M, Sipponen P, Kekki M. 1985. Chronic gastritis: dynamic and clinical aspects. Scand J Gastroenterol (suppl) 109:69–76. Sivam GP, Lampe JW, Ulness B, Swanzy SR, Potter JD. 1997. Helicobacter pylori—in vitro susceptibility to garlic (Allium sativum) extract. Nutr Cancer 27:118–121. Smith RA, Cokkinides V, Eyre HJ. 2003. American Cancer Society guidelines for the early detection of cancer, 2003. CA Cancer J Clin 53:27–43. Sparnins VL, Mott AW, Barany G, Wattenberg LW. 1986. Effects of allyl methyl trisulfide on glutathione S-transferase activity and BP-induced neoplasia in the mouse. Nutr Cancer 8:211–215. Stalnikowicz R, Benbassat J. 1990. Risk of gastric cancer after gastric surgery for benign disorders. Arch Intern Med 150:2022–2026. Stein M, Bagnoli F, Halenbeck R, Rappuoli R, Fantl WJ, Covacci A. 2002. c-Src/Lyn kinases activate Helicobacter pylori CagA through tyrosine phosphorylation of the EPIYA motifs. Mol Microbiol 43:971–980. Stemmermann GN. 1994. Intestinal metaplasia of the stomach. A status report. Cancer 74:556–564. Sung JJ, Lin SR, Ching JY, Zhou LY, To KF, Wang RT, et al. 2000. Atrophy and intestinal metaplasia one year after cure of H. pylori infection: a prospective, randomized study. Gastroenterology 119:7–14. Takada K. 2000. Epstein-Barr virus and gastric carcinoma. Mol Pathol 53:255–261. Tatematsu M, Tsukamoto T, Inada K. 2003. Stem cells and gastric cancer: Role of gastric and intestinal mixed intestinal metaplasia. Cancer Sci. 94:135–141. Terry MB, Gaudet MM, Gammon MD. 2002. The epidemiology of gastric cancer. Semin Radiat Oncol 12:111–127. Tersmette AC, Offerhaus GJ, Tersmette KW, et al. 1990. Meta-analysis of the risk of gastric stump cancer: Detection of high risk patient subsets for stomach cancer after remote partial gastrectomy for benign conditions. Cancer Res 50:6486–6489. Uemura N, Mukai T, Okamoto S, Yamaguchi S, Mashiba H, Taniyama K, et al. 1997. Effect of Helicobacter pylori eradication on subsequent development of cancer after endoscopic resection of early gastric cancer. Cancer Epidemiol Biomark Prev 6:639–642. Uemura N, Okamoto S, Yamamoto S, Matsumura N, Yamaguchi S, Yamakido M, et al. 2001. Helicobacter pylori infection and the development of gastric cancer. N Engl J Med 345:784–789. Vermeer IT, Engels LG, Pachen DM, Dallinga JW, Kleinjans JC, van Maanen JM. 2001. Intragastric volatile N-nitrosamines, nitrite, pH, and Helicobacter pylori during long-term treatment with omeprazole. Gastroenterology 121:517–525. Warren JR, Marshall BJ. 1983. Unidentified curved bacilli in gastric epithelium in active chronic gastritis. Lancet 1:1273–1275. Watanabe H, Jass JR, Sobin LH. 1990. Histological Typing of Oesophageal and Gastric Tumours. Berlin: Springer-Verlag. Watanabe Y, Fukao A. 2001. Stomach cancer screening. In: Hisamichi S, ed. Report on the evaluation of the efficacy of new cancer screening methods. Tohoku University School of Medicine, Sendai, Japan, pp 81–120 (in Japanese). Wayman J, Forman D, Griffin SM. 2001. Monitoring the changing pattern of esophago-gastric cancer: Data from a UK regional cancer registry. Cancer Causes Control 12:943–949. World Cancer Research Fund/American Institute for Cancer Research. 1997a. Food, Nutrition and the Prevention of Cancer: A Global Perspective. American Institute for Cancer Research, Washington, DC. World Cancer Research Fund/American Institute for Cancer Research. 1997b. Stomach. In: World Cancer Research Fund/American Institute for Cancer Research, ed. Food, Nutrition and the Prevention of Cancer: A Global Perspective. Washington, DC: American Institute for Cancer Research, pp. 148–175. World Health Organization. 2002. The World Health Report 2002. World Health Organization, Geneva, Switzerland. Wu MS, Huang SP, Chang YT, et al. 2002. Tumor necrosis factor-alpha and interleukin-10 promoter polymorphisms in Epstein-Barr virus-associated gastric carcinoma. J Infect Dis 185:106–109. Wu MS, Shun CT, Wu CC, Hsu TY, Lin MT, Chang MC, et al. 2000. EpsteinBarr virus-associated gastric carcinomas: relation to H. pylori infection and genetic alterations. Gastroenterology 118:1031–1038.

720

PART IV: CANCER BY TISSUE OF ORIGIN

Wynder EL, Kmet J, Dungal N, Segi M. 1963. An epidemiological investigation of gastric cancer. Cancer 16:1461–1496. Yabuta T, Shinmura K, Tani M, et al. 2002. E-cadherin gene variants in gastric cancer families whose probands are diagnosed with diffuse gastric cancer. Int J Cancer 101:434–441. Yamazaki S, Yamakawa A, Ito Y, Ohtani M, Higashi H, Hatakeyama M, et al. 2003. The CagA protein of Helicobacter pylori is translocated into epithelial cells and binds to SHP-2 in human gastric mucosa. J infect Dis 187:334–337. Yanai H, Takada K, Shimizu N, Mizugaki Y, Tada M, Okita K. 1997. EpsteinBarr virus infection in non-carcinomatous gastric epithelium. Journal of Pathology 183:293–298. Yang PC, Davis S. 1998. Epidemiological characteristics of adenocarcinoma of the gastric cardia and distal stomach in the United States, 1973–1982. Int J Epidemiol 17:293–297.

Yang CS, Wang ZY. 1993. Tea and cancer. J Natl Cancer Inst 85:1038–1049. Yoon KA, Ku JL, Yang HK, Kim WH, Park SY, Park JG. 1999. Germline mutations of E-cadherin gene in Korean familial gastric cancer patients. J Hum Genet 44:177–180. Zaridze D, Borisova E, Maximovitch D, Chkhikvadze V. 1999. Aspirin protects against gastric cancer: results of a case-control study from Moscow, Russia. Int J Cancer 82:473–476. Zhou L, Sung JJ, Lin S, Jin Z, Ding S, Huang X, et al. 2003. A five-year followup study on the pathological changes of gastric mucosa after H. pylori eradication. Chinese Med J 116:11–14.

38

Cancer of the Pancreas KRISTIN E. ANDERSON, THOMAS M. MACK, AND DEBRA T. SILVERMAN

C

ancer of the pancreas remains a serious medical and public health problem because of difficulties in early diagnosis, aggressive behavior, resistance to therapy, and limited opportunities for prevention. About 95% of the cancers arise from the exocrine portion of the pancreas, which is composed of acinar cells that produce and discharge enzymatic secretions into the pancreatic ductal system. Much less frequent are endocrine tumors arising from the islets of Langerhans. In the United States, about 32,000 cancers of the exocrine pancreas were diagnosed in 2005. Because of its high lethality, with less than 4% of cases surviving 5 years after diagnosis, pancreatic cancer is estimated to account for about 5% of all cancer deaths (Jemal et al., 2005).

PATHOLOGY Classification

(FNA), have improved physicians’ ability to correctly diagnose and stage the disease over the past two decades (O’Meara, 2002). However, because the organ is inconveniently located, and because of the morbidity associated with biopsy, pancreatic cancer continues to have among the lowest proportion of histologically verified cases of any major cancer. The average proportion of incident male cases that have been verified histopathologically prior to notification to various registries is given in Table 38–1 (Parkin et al., 2002). The accuracy of pancreatic cancer incidence estimates depends heavily on enumeration of cases with and without histologic examination, which varies throughout the world.

DESCRIPTIVE EPIDEMIOLOGY

Most cancers of the exocrine pancreas, adenocarcinoma, are believed to arise from ductal cells, and feature an infiltrative growth pattern that extends beyond the identifiable gross tumor, an intense collagenous stromal reaction, and entrapped normal-appearing pancreatic cells (Klöppel and Maillet, 1991; O’Meara, 2002; Bardeesy and DePinho, 2002). Approximately two-thirds of pancreas tumors occur in the head of the organ with the remainder in the body or tail (Kumar, 2004). Tissue adjacent to the tumor often contains precancerous lesions with varying grades of dysplasia and intraepithelial neoplasia. In these lesions the presence of genetic alterations, such as mutated K-ras oncogenes, appears to signal the risk of progression to invasive ductal cancer. Several variants of ductal adenocarcinoma have been described, including signet ring cell, mucinous or colloid, adenosquamous, giant cell, medullary, and anaplastic carcinomas (Klöppel and Maillet, 1991). Other exocrine tumors of the pancreas include mucinous cystic adenocarcinoma (which may be associated with invasive ductal adenocarcinoma), acinar carcinoma, and pancreaticoblastoma (a tumor mainly of childhood with ductal, mesenchymal, and endocrine elements). Islet cell tumors of the endocrine pancreas, whether benign or malignant, are uncommon and generally secrete high levels of specific peptide hormones, such as insulinomas or gastrinomas (O’Meara, 2002). These tumors are often a feature of the hereditary syndrome, multiple endocrine neoplasia type 1. Other rare cancers of the pancreas include sarcomas and lymphomas. Recently, researchers have characterized the natural history of early duct lesions that are thought to give rise to pancreatic malignancies. These early lesions have been designated “Pancreatic Intraepithelial Neoplasia” (PanIN) (Hruban et al., 2001; Biankin et al., 2003).

It is important to note that some, perhaps much, of the variation in both mortality and incidence patterns may be due to variation in methods and completeness of case ascertainment.

Diagnosis

Survival rates for pancreatic cancer are among the lowest of any cancer. Median survival is about 3 months. The 5-year relative survival rate (for all races and both sexes combined) for cases diagnosed between 1995 and 2000 was almost 4% which was a modest improvement over a 2.6% rate for cases diagnosed between 1974 and 1976 (Jemal et al., 2004). When survival trends are examined by stage, there appears to be improved survival from 1975–2000 in patients diagnosed with local stage disease; however, this could be an artifact resulting from improved staging techniques and stage migration, rather than real gains in survival (Feinstein et al., 1985; Jemal et al.,

There is no screening test available for early detection of pancreatic cancer. At diagnosis, less than 10% of patients present with local disease while 47% have distant metastases (Jemal et al., 2004). Patients commonly present with symptoms of weight loss, pain, or jaundice (O’Meara, 2002). Current diagnostic imaging techniques include computed tomography (CT), ultrasound imaging (US), endoscopic retrograde cholangiopancreatography (ERCP), and others. These, together with use of percutaneous fine-needle aspiration

Incidence Pancreatic cancer is the eleventh and ninth most common cause of cancer in men and women, respectively, in the United States (Jemal et al., 2004) with a similar ranking in the United Kingdom and Sweden (Parkin et al., 2002). Jemal et al., estimated that there would be 32,000 new cases in the United States in 2005, accounting for approximately 2.3% of cancer diagnoses. The age-adjusted (US 2000 standard population) incidence rates per 100,000 person-years of observation (pyo) for pancreatic cancer among males and females (SEER, all races) in the United States in 1992–2001 were 12.7 and 9.9, respectively (Jemal et al., 2004).

Mortality Carcinoma of the pancreas is the fourth leading cause of cancer deaths in the United States (Jemal et al., 2004). In 2005, it was estimated that there would be 31,800 deaths from this disease, or about 5% of all cancer deaths (Jemal et al., 2005). For 1992–2001 the annual ageadjusted (US 2000 standard population) mortality rates (SEER) per 100,000 pyo for pancreatic cancer in the United States were as follows: 12.0 for white males, 8.9 for white females, 16.6 for black males, 12.9 for black females; 9.5 for Hispanic/Latino males, and 7.5 for Hispanic/Latino females (Jemal et al., 2004).

Survival

721

722

PART IV: CANCER BY TISSUE OF ORIGIN

Table 38–1. Percentages of Incident Cancer of the Pancreas Confirmed Histopathologically, in Men Area

Number of Registries

Mean (%)

Range

6 4 12 11 6 12 5 14

29 36 34 45 40 59 73 78

0–57 24–42 24–49 10–96 29–56 37–74 64–85 68–86

Africa Poland United Kingdom Latin America Japan Canada Scandinavia United States Source: Parkin et al. (2002).

2004). In the minority of patients for whom surgery is an option, the 5-year survival rate is better than the rate for all stages but is still between 10%–15% (Bornman and Beckingham, 2001). None of the standard treatment options—surgery, radiation therapy, and chemotherapy—have been shown to significantly influence the high mortality rate of the majority (75%–85%) of cases that present with relatively advanced disease and unresectable tumors; therefore palliation is the most common therapy.

Time Trends in the United States Pancreatic cancer incidence and mortality rates increased in the United States during the 20th century, but have plateaued since about 1970. Between 1920 and 1965, the age-adjusted mortality rate increased approximately threefold (Krain, 1970). Mortality and incidence rates have plateaued since about 1970 with some slight declines (SEER, 2004a; SEER, 2004b). The age-adjusted (US 2000 standard population) incidence rate per 100,000 pyo for both sexes and all races in 1973 was 12.3, while in 2000, the comparable rate was 11.2 (SEER, 2004a). In 1973, the age-adjusted (US 2000 standard population) mortality rate for both sexes and all races was 10.7 and in 2000, it was 10.5/100,000 pyo, (SEER, 2004b). The slight declines in incidence and mortality rates over the past 15–25 years are mainly due to changes

in rates for white males. For black males, rates in both incidence and mortality have changed little between the 1970s and 2000. For black and white females, incidence and mortality rates rose from 1970 until about the mid-1980s and have remained relatively flat through 2000 (SEER, 2004a; SEER, 2004b). Reasons for the increase earlier in this century are not known, but it is believed to be due, in part, to improvements in diagnostic procedures (IARC, 1986). Another possible explanation is a lagged correlation between the prevalence rates of cigarette use and cancer mortality rates for men and women in the United States (Weiss and Bernarde, 1983). Given the strong epidemiologic data that implicate cigarettes as a risk factor for pancreatic cancer, trends in cigarette use might explain both the increase seen until about 1970 (Fontham and Correa, 1989) and the recent decrease in white males (CDC, 2004).

Race, Ethnicity, and International Patterns Among males and females, respectively, there is an approximately 10to 15-fold variation in pancreas cancer mortality around the world (Figs. 38–1, 38–2) (Ferlay et al., 2004; WHO, 2004). In a comparison of age-standardized (world) annual incidence rates per 100,000 pyo for the period 1993–1997, the highest rates are found among US black populations (e.g., in Connecticut the rates for blacks are 14.7 for males and 9.5 for females, while the corresponding rates in whites are 8.1 for males and 6.1 for females). In general, the incidence rates in blacks are 1.5 to 1.9 times those of whites throughout the United States. In contrast to the high rates in African Americans, the incidence rates in Africa (although based on very limited data) are low (Parkin et al., 2002). Silverman et al. (2003) found evidence to support the view that the disparity between rates in black and white Americans is due to prevalence differences in risk factors, and effect modification by race. Worldwide, incidence and mortality rates are generally high among North Americans, Europeans, and Japanese. Lower rates are reported from Africa, South East Asia, India, and parts of the Middle East. Among the lowest incidence rates reported are those from regional population-based registries of India (e.g., Ahmedabad (1.1, males; 0.8, females)). Low rates are also found in Thailand, Kuwait, and Algeria. Intermediate rates are seen in Central and South America (e.g., in Cali, Colombia (4.6, males; 4.4, females)) (Parkin et al., 2002).

Figure 38–1. Age-standardized (World, 2000) mortality rate per 100,000 pyo for pancreas cancer among males. (Source: Ferlay et al., 2004.)

Cancer of the Pancreas

723

Figure 38–2. Age-standardized (World, 2000) mortality rate per 100,000 pyo for pancreas cancer among females. (Source: Ferlay et al., 2004.)

A decade ago, the highest mortality rates (Aoki et al., 1992) occurred in northern Europe, including Britain, and in countries populated by migration from those areas. The corresponding rates in central and southern Europe were generally lower. Recently, this pattern has shifted and the highest mortality rates in Europe occur in males in Eastern Europe, though this trend is currently leveling off (Levi et al., 2003). Though relatively low 30 years ago (Segi and Kurihara, 1972), more recently, mortality rates (age-standardized, world) in Japan have been similar to rates in Western countries (WHO, 2004). Current data for regions of Europe can be illustrated with examples of mortality rates per 100,000 pyo, age-adjusted (world): Finland, 8.2 for males and 6.8 for females; United Kingdom, 6.6 for males and 5.5 for females; Greece, 6.2 for males and 3.6 for females; Czech Republic, 10.4 for males and 7.2 for females (WHO, 2004; Ferlay et al., 2004). Incidence patterns for the period 1993–1997 (Parkin et al., 2002) are similar to those seen for mortality (age-standardized, world) (e.g., incidence rates per 100,000 pyo in Finland are 8.8 for males and 6.3 for females; in England and Wales, 6.7 and 4.8; in the white population of United States (SEER), 7.3 and 5.5; and in New Zealand, 6.4 and 5.1). The Maori population of New Zealand has incidence rates that are higher than their white counterparts, particularly the women (Phillips et al., 2002). Rates similar to those in the United States, New Zealand, or Britain are reported in, Italy, Genoa (8.1 and 6.2); France, Haut-Rhin (7.5 and 4.5); and Spain, Granada (5.8 and 3.0). High rates can be found in Italy, Venetia (10.2 and 6.2) and Parma (10.3 and 5.8), and in Eastern Europe (e.g., Estonia) (11.4 and 5.3). Incidence rates are relatively low in Asia (e.g., among the Malay of Singapore (3.0 and 2.3)). Again, Japan is an exception where the rates are not very different from those in US whites (9.4 and 5.5, in Osaka). Incidence rates among Americans of Asian descent have generally, but not always, been between those of Asians in Asia and those of Americans of European origin (Mack and Paganini Hill, 1981; Parkin et al., 1992; Parkin, 2002). In South America, incidence rates are generally lower than rates in Western Europe and the United States (e.g., in Campinas, Brazil, (3.4 and 2.6, for men and women, respectively)) (Parkin et al., 2002), while Hispanic populations in the United States have rates that are compa-

rable to those of non-Hispanic white Americans. The New Mexico Tumor Registry reported the average annual age-adjusted (world) incidence rates per 100,000 pyo for the Hispanic population as 7.5 and 5.8, for men and women, respectively. The corresponding rates in the nonHispanic white population for the same period (1993–1997) were 6.5 and 5.0 (Parkin et al., 2002). In California, the rates for Hispanics and Non-Hispanic Whites were also similar (8.1 vs. 7.7 and 5.9 vs. 5.8) for males and females, respectively (Parkin et al., 2002). Since “Hispanic” is not a well-defined category, these patterns should be interpreted cautiously. Available estimates of incidence for American Indians are based on small numbers. In the past, studies of mortality for both American Indians (Creagan and Fraumeni, 1972) and Alaskan natives (Blot et al., 1975; Lanier et al., 1976) have shown no consistent variation from the patterns among whites. Between 1993 and 1997, the average annual age-adjusted (world) incidence rates for Native American males and females in New Mexico were 4.5 and 6.6, respectively (Parkin et al., 2002). Average annual (1992–2001) incidence rates for American Indian/Alaska Native (SEER) are slightly lower than rates in whites for both males and females (SEER, 2004a).

Migration Examination of mortality in relation to place of birth has not produced patterns that are easily interpretable. In the late 1950s and early 1960s, US mortality rates were twofold to threefold higher than those in Japan, but the rates among those of Japanese origin in the United States were actually higher than those in US whites (Haenszel and Kurihara, 1968). A similar observation has been made in Australia during the 1960s and 1970s where migrants from lower-risk southern Europe and Poland, after 16 or more years, experienced higher pancreatic cancer risk than native-born Australians (McMichael et al., 1980). Mortality in European immigrants to the United States has been reported to be somewhat higher in second or subsequent generations as well (Haenszel, 1961); mortality has also been reported to be higher in US counties populated by a high percentage of residents of northern European descent (Blot et al., 1978). Israelis born in Europe or America experience higher rates than those born in Africa or Asia (Parkin et al., 2002). These differences are not as great as racial

724

PART IV: CANCER BY TISSUE OF ORIGIN

differences, however, and the association with place of birth has not been a consistent finding.

The most reliable and important known predictor of pancreas cancer incidence and mortality is age. In the first three decades of life, this cancer is extremely uncommon, although it can occur even in childhood (Moynan et al., 1964; Tsukimoto et al., 1973; Taxy, 1976; Grosfeld et al., 1990). After the age of 30, as in most epithelial cancers, rates increase with age in approximately a log-linear fashion. The slopes of these curves are quite similar from population to population, regardless of race, geographic area, or absolute level of age-adjusted mortality. Individuals aged 80 years and above have mortality rates approximately 40 times higher than those for individuals aged 40 years. The median age at diagnosis in the United States is 72 years (SEER, 2004a) and the majority of cases occur between ages 65 and 79. The age-specific mortality curves in males and females (Figs. 38–3 and 38–4) increase steadily, even through the last age category (85+) in most regions as evidenced by recent data from Hungary, Sweden, United States, and Hong Kong (WHO, 2004). In the last age category (age 85+), rates decline or level off in Hungary (men and women) and in Hong Kong (women); such declines are consistent with an artifact of diagnosis ascertainment. Alternatively, there could be a genuine elimination of susceptibles or a higher proportion of unexposed (for example, fewer smokers) in this oldest age group in some populations.

Sex Pancreas cancer is about 50% more common in men than in women, varying somewhat by race, geographical location, and histological type. The sex ratio is remarkably constant as long as comparisons are based on data from relatively complete population-based registries (Parkin et al., 2002). In the United States, the sex ratio ranged from 1.54 : 1.0 for ages 20–54, to 1.13 : 1.0 for ages 75 and over (SEER, 2001). Among US whites, the sex ratios for both incidence and mortality peaked around 1970 with a 60% male excess and have declined continuously since (Devesa et al., 1987; Muir et al., 1987; Parkin et al., 2002). These changes over time are consistent both with changes in exposure (particularly increased smoking among women) and with possible differences between the sexes in the changes that have occurred in diagnostic work-up, cancer treatment, and reporting of the disease over time.

Mortality per 100,000 person-years

140

120

Hungary Sweden

100

United States Hong Kong

80

Mortality per 100,000 person-years

Age

140

120

Hungary Sweden

100

United States Hong Kong

80

60

40

20

0 20- 25- 30- 35- 40- 45- 50- 55- 60- 65- 70- 75- 80- 85+ Age

Figure 38–4. Average annual age-specific mortality for pancreas cancer among females for selected populations, 1990–1995 (WHO, 2004).

Socioeconomic Status The relation between pancreas cancer and socioeconomic (SES) status has been inconsistent over time. In England and Wales in the early 1930s, mortality from the disease, at least in women, was somewhat more common in lower socioeconomic status households (OPCS, 1938). In the next report of the Registrar General (OPCS, 1954), the trends were inconsistent, but mortality tended to be higher for persons in the upper social classes. In the most recent report (Logan, 1982), no trends were discernible. Similarly in the United States, observations from earlier periods suggested either no social class association (Cohart and Muller, 1955; Graham et al., 1960) or higher rates in the lower socioeconomic categories (Dorn and Cutler, 1958; Seidman, 1970; Krain, 1971a, b; Young et al., 1975). In more recent periods, some have observed no association (Moldow and Connelly, 1968; Wynder, et al., 1973a; 1983; Williams and Horm, 1977; Blot et al., 1978), while others have observed an inconsistent association with higher social class (Levin et al., 1981; Levin and Connelly, 1973; Lin and Kessler, 1981; Mack and Paganini-Hill, 1981; Falk et al., 1988; Ferraroni et al., 1989). Recently, in the large American Cancer Society CPSII cohort study, education was not a predictor of pancreas cancer mortality (Coughlin et al., 2000). In contrast, a recent population-based case-control study found that low income was associated with an 80% increased risk in white men and a 170% increased risk in black men after adjustment for smoking, dietary factors, and heavy alcohol drinking, but no significant SES relation was observed among women (Silverman et al., 2003). Whether the observed trends have been judged positive, inverse, or absent, however, the magnitude of any effect is usually small, and risk in both the lowest and highest categories has usually been elevated in comparison with intermediate categories.

Religion

60

40

20

0 20- 25- 30- 35- 40- 45- 50- 55- 60- 65- 70- 75- 80- 85+ Age

Figure 38–3. Average annual age-specific mortality for pancreas cancer among males for selected populations, 1990–1995 (WHO, 2004).

Certain religious groups have experienced either higher or lower than average susceptibility of developing pancreatic cancer. Increased mortality from pancreas cancer in New York Jews could not be explained by age or nativity (MacMahon, 1960; Newill, 1961; King et al., 1965), or socioeconomic status based on place of residence (Seidman, 1970). An excess among Jewish men, but not women, was also reported in a multicenter case-control study (Wynder, 1973a) and is evident in data from New York (Greenwald et al., 1975). Cohort studies of Seventh-Day Adventists have reported a deficit of deaths from pancreas cancer (Phillips, 1975; Phillips et al., 1980; Mills et al., 1988), and studies of both incidence (Lyon et al., 1976, 1980)

Cancer of the Pancreas and mortality (Enstrom, 1978; 1980) in members of the Church of Jesus Christ of the Latter Day Saints (LDS)(often called Mormons) have indicated a deficit of pancreas cancer of similar magnitude. In Los Angeles, Mack and Paganini-Hill (1981) observed a slight deficit of incident cases among LDS and a slight excess among Jews.

RISK FACTORS Smoking The most consistent risk factor for pancreatic cancer is cigarette smoking. In 2002, the International Agency for Research on Cancer (IARC, 2004) concluded the following: Cancer of the pancreas is causally associated with cigarette smoking. The risk increases with duration of smoking and number of cigarettes smoked daily. The risk remains elevated after allowing for potential confounding factors such as alcohol consumption. The relative risk decreased with increasing time since quitting smoking. An IARC working group that evaluated the carcinogenic risk of tobacco had previously concluded that cigarette smoking is an important cause of pancreatic cancer (IARC, 1986, p. 313). The conclusion was based on the evaluation of findings from 9 cohort studies (Doll and Peto, 1976; Kahn, 1966; Lossing et al., 1966; Hammond and Horn, 1958; Hammond, 1966; Cederlöf et al., 1975; Weir and Dunn, 1970; Hirayama, 1981; Heuch et al., 1983) and 8 case-control studies (Wynder et al., 1973a; Lin and Kessler, 1981; MacMahon et al., 1981; Wynder et al., 1983; Whittemore et al., 1983; Durbec et al., 1983; Kinlen and McPherson, 1984; Polissar et al., 1984). All of the cohort studies and all but one of the case-control studies showed increased risks for smokers and most had evidence of a positive dose-response. Relative risk estimates are around twofold, with some as high as sixfold in association with high exposure. Studies subsequent to the IARC report support its conclusions, including at least 10 cohort studies (Hirayama, 1988; Hiatt et al., 1988; Mills et al., 1988; Zheng et al., 1993; Shibata et al., 1994; Fuchs et al., 1996; Harnack et al., 1997; Coughlin et al., 2000; Lin et al., 2002a; Nöthlings et al., 2005) and 18 case-control studies (Mack et al., 1986; Norell et al., 1986a; Hsieh et al., 1986; Wynder et al., 1986; La Vecchia et al., 1987, Falk et al., 1988, Olsen et al., 1989; Cuzick and Babiker, 1989; Farrow and Davis, 1990a; Lyon et al., 1992a; Mizuno et al., 1992; Friedman and van den Eeden, 1993; Kalapothaki et al., 1993a; Silverman et al., 1994; Ji et al., 1995a; Boyle et al., 1996; Soler et al., 1998; Anderson et al., 2002a). One hospital-based case-control study reported finding no association (Clavel et al., 1989); however, the odds ratios reported are consistent with an increased risk with cigarette smoking. Howe et al. (1991) reported a rapid decrease in risk with time since quitting. Fuchs et al. (1996), in a large prospective analysis, found that compared with individuals who continued to smoke, former smokers had a 48% reduction in pancreas cancer risk with 2 years of quitting and the relative risk of the cancer among former smokers approached that for never-smokers after less than 10 years of quitting. Many studies report a relative risk estimate in former smokers that is lower than in current smokers; indeed, in some studies the difference in risk between never-smokers and ex-smokers is not statistically significant (Falk et al., 1988; Farrow and Davis, 1990a; Bueno de Mesquita et al., 1991a; Howe et al., 1991; Ghadirian et al., 1991a; Friedman and van den Eeden, 1993; Silverman et al., 1994; Harnack et al., 1997; Lin et al., 2002a). A number of studies (Mack et al., 1986; Howe et al., 1991; Farrow and Davis, 1990a; Cuzick and Babiker, 1989; Silverman et al., 1994; Lin et al., 2002a) found an inverse trend in risk of pancreas cancer with increasing years of cessation. Silverman et al. (1994) found that individuals who quit for more than 10 years experienced a reduction in risk of about 30% relative to current smokers, while those who quit for less than 10 years did not appear to be at reduced risk. Two other studies found reductions in risk within the first 10 years of smoking cessation (Mack et al., 1991; Howe et al., 1991). A population-based case-control analysis of noncigarette tobacco use among nonsmokers of cigarettes (Alguacil and Silverman, 2004)

725

revealed increased odds ratios associated with cigar smoking (ranging from 1.7–1.9) and smokeless tobacco use (ranging from 1.5–3.4), but no increased risk associated with pipe smoking. A recent cohort study that included 15,263 current or former pipe smokers found that the risk of pancreatic cancer was generally smaller than those associated with cigarette smoking and similar to or larger than those associated with cigar smoking (Henley et al., 2004). Wynder et al. (1983) reported a twofold increased risk associated with pipe and cigar smoking, but other earlier studies found no association (Kahn, 1966; Williams and Horm, 1977; Mack et al., 1986, Falk et al., 1988; Farrow and Davis, 1990a; Bueno de Mesquita, 1991a; Howe et al., 1991). Despite the strong evidence that smoking is a cause of pancreatic cancer, no unequivocal biological mechanism has been demonstrated for smoking-induced pancreatic carcinogenesis to date. There are numerous chemical carcinogens in cigarette smoke that could play a role. Wynder proposed that carcinogens absorbed from tobacco smoke may reach the pancreas through the blood, or alternatively, through refluxed bile (Wynder et al., 1973b). There have been arguments made about when smoking exerts its effects (i.e., early or late in the carcinogenic process) (Mack et al., 1986; Howe et al., 1991; Silverman et al., 1994). Data showing a decreasing risk with time since smoking cessation might support a late-stage effect; however, we do not know the minimum time required for the carcinogenic process to occur for this disease. It could take place in a relatively short amount of time. Several studies have reported carcinogen-DNA-adduct levels to be higher and profiles to be different in the pancreas of smokers compared with nonsmokers (Cuzick et al., 1990; Kaderlik et al., 1993; Wang et al., 1998). Such studies indicate that tobacco carcinogens reach the pancreas, but the methods used are not specific enough to implicate particular carcinogens. Among the possible human pancreatic carcinogens are nitrosamines. Nitrosamines induce pancreatic cancers in several animal models (Standop et al., 2001; Wei et al., 2003) and they are present at high levels in cigarettes (Hecht and Hoffmann, 1991; Wogan et al., 2004). Over a decade ago, the Surgeon General’s report (US DHHS, 1989) deemed the N-nitrosamine, NNK, as a putative human pancreatic carcinogen in cigarette smoke; the major metabolite of NNK, 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanol (NNAL) is also a likely human pancreatic carcinogen (Wogan et al., 2004). Tobacco contains many other possible pancreatic carcinogens (Patrianakos, 1979; CAEPA, 1997; Swauger et al., 2002; IARC, 2004), including aromatic amines (Kadlubar et al., 1992, 1993; Anderson et al., 1997), polycyclic aromatic hydrocarbons (Wogan et al., 2004), and metals such as cadmium (Schwartz and Reis, 2000).

Energy Balance Energy balance or imbalance reflects the difference between dietary caloric intake and energy expenditure. Caloric intake, physical activity, metabolism, and obesity may have independent effects on cancer etiology, but given the methods currently available for epidemiologic studies, it is difficult to disentangle their effects. Experimental studies are increasing our understanding of the biology of physical activity and obesity and how these may affect cancer etiology. Epidemiologists will be well served by more precise measurement tools (e.g., biomarkers or better measures of dietary intake) to achieve a clear understanding of the role of these factors in pancreas cancer etiology.

Caloric Intake Twelve studies have examined the relationship between energy intake and pancreatic cancer risk. The associations reported have been positive (Silverman et al., 1998; Howe et al., 1990; Ghadirian et al., 1991b; Olsen et al., 1991; Howe et al., 1992), null (Harnack et al., 1997; Kalapothaki et al., 1993b), inverse (Stolzenberg-Solomon et al., 2002; Farrow and Davis, 1990b; Durbec et al., 1983), and mixed (Hanley et al., 2001; Bueno de Mesquita et al., 1990). Harnack and colleagues (1997) analyzed data from a large population-based cohort of postmenopausal women and found no association between increasing calorie intake and pancreatic cancer risk. In contrast, Silverman and colleagues (1998) in a population-based

726

PART IV: CANCER BY TISSUE OF ORIGIN

case-control study based on direct interviews, found a relationship in both men and women. In a combined analysis, increasing calorie intake increased the risk of pancreatic cancer with odds ratios of 1.0, 1.2, 1.4, and 1.7 for increasing quartiles (P (trend) = 0.001). They also observed a significant interaction between caloric intake and body mass index (BMI) that was consistent by sex and race, and proposed a role for energy balance in pancreatic carcinogenesis. Hanley and colleagues (2001) found no association between calorie intake and pancreatic cancer for women in their case-control study, but did find that men with a high weekly calorie intake had an increased risk of pancreatic cancer. Stolzenberg-Solomon et al. (2002) analyzed data from a large prospective cohort of male smokers. They found that increasing energy intake decreased the risk of pancreatic cancer, but the generalizability of their results is unclear. There is support for the hypothesis that increased energy intake could increase the risk of pancreas cancer. Caloric restriction has been shown to decrease tumor incidence in a variety of tissues and animal models (Tannenbaum, 1959; Pariza, 1987). More specifically, reduced caloric intake protects against pancreatic carcinogenesis in rats (Roebuck et al., 1993), though not in hamsters (Birt et al., 1997). The possible inhibitory mechanism is not known, but may involve reduced trophic stimuli to the pancreas, reduced levels of carcinogenactivating enzymes within the pancreas, or other physiologic differences (Longnecker, 1990; Pariza, 1987). The conflicting results of the epidemiologic studies emphasize the need for more research in the area of calorie intake and pancreatic cancer risk.

Physical Activity At least 9 studies have examined the effect of physical activity on the risk of pancreatic cancer; three found an inverse association (Inoue et al., 2003; Isaksson et al., 2002; Michaud et al., 2001), four found no association (Garfinkel and Stellman, 1988; Waterbor et al., 1988; Brownson et al., 1991; Lee et al., 2003), and two found mixed results (Nilsen and Vatten, 2000; Hanley, 2001). Lee and colleagues (2003) analyzed data from a large prospective cohort in the United States with serial measurements of physical activity. They assessed participation in light, moderate, and vigorous activity and asked questions regarding distance walked and stairs climbed. None of these measures was associated with risk of subsequent pancreas cancer; increasing quartiles of physical activity did not decrease risk of pancreatic cancer. Michaud and colleagues (2001) reported an inverse association between total physical activity and pancreatic cancer risk within two large prospective cohorts in the United States, the Nurses Health Study and the Health Professionals Follow-up Study. The protective effect was significant among the nurses, but only marginally significant among the male health professionals. Vigorous activity was not associated with decreased risk, but moderate physical activity did significantly decrease risk. The authors also found that increased amounts of walking or hiking decreased pancreatic cancer risk and they observed that the protective effect of physical activity was most apparent in overweight subjects—in nonobese (BMI <25 kg/m2) subjects there was no physical activity effect. Insulin resistance and resulting hyperinsulinemia have been suggested as mechanisms by which adult onset diabetes and obesity could enhance pancreatic cancer development (Everhart et al., 1995; Silverman et al., 1999; Wang et al., 2003). Since insulin resistance improves with routine physical activity (Short, 2003; Goodpaster et al., 2003), reduced insulin resistance may explain findings of lower risk associated with increasing levels of activity for this disease. Physical activity can also reduce iron stores (Liu et al., 2003) and excessive iron has been implicated in human cancers (Toyokuni, 2002). As we increase our knowledge of the biologic effects of regular exercise, future epidemiologic research can be targeted towards specific hypotheses incorporating methods to obtain comprehensive data on amount, intensity, and duration of activity.

Obesity Overall, epidemiologic evidence suggests that overweight and obesity may cause a small to modest increased risk of pancreas cancer. At least

23 studies have examined this relationship with 9 finding a positive association (Friedman and van den Eeden, 1993; Moller et al., 1994; Ji et al., 1996; Silverman, 1998; Wolk et al., 2001; Michaud et al., 2001; Isaksson et al., 2002; Calle et al., 2003; Pan et al., 2004), 10 finding no association (Mack et al., 1986; Olsen et al., 1991; Howe et al., 1992; Kalapothaki et al., 1993a; Lyon et al., 1993; Shibata et al., 1994; Nilsen and Vatten, 2000; Stolzenberg-Solomon et al., 2002; Inoue et al., 2003; Lee et al., 2003), and 4 finding mixed results (Gapstur et al., 2000; Hanley et al., 2001; Wolk et al., 2001; Samanic et al., 2004). Lee and colleagues (2003) in the large prospective study of college alumni in the United States found that BMI did not significantly predict pancreatic cancer mortality with multivariate-adjusted relative risks for quartiles of BMI of 1.00, 0.84, 1.08, and 0.99, respectively. In contrast, Michaud and colleagues (2001) analyzed another large prospective study in the United States and found that individuals with a BMI of at least 30 kg/m2 (obese) had an increased risk of pancreatic cancer compared with those with a BMI of less than 23 kg/m2 (RR = 1.72). Silverman and colleagues (1998) analyzed data from a populationbased case-control study based on direct interviews and found that men and women in the highest quartile of BMI experienced a statistically significant 60% increase in pancreatic cancer risk with a significant trend with increasing BMI (P = 0.003). Samanic and colleagues (2004) analyzed data on obesity from a large cohort of male US veterans and found that, among white veterans, risk of pancreatic cancer was significantly elevated with a relative risk of 1.2 (1.07–1.33) for obese subjects, but no increased risk was apparent for black veterans (RR = 1.07(0.86–1.34)). Three other recent studies found that obesity was associated with an increased risk of pancreatic cancer for men, but not women (Hanley et al., 2001; Wolk et al., 2001; Gapstur et al., 2000). Berrington de Gonzales and colleagues (2003), in a meta-analysis of obesity and risk of pancreatic cancer, estimated a small per unit increase in relative risk per increase in unit of body mass. This translated to a 19% increase in RR for obese individuals (body mass index greater than 30 kg per m2) compared with those of normal body weight (22 kg per m2). Calle and Kaaks (2004), using slightly higher estimates of the relative risks associated with obesity (1.7) and overweight (1.3), based on a summary of selected literature, estimated the population attributable fraction percentages of pancreas cancer due to these conditions for Europe and the United States as 19.3% and 26.9%, respectively. Energy imbalance and consequent obesity result in hyperinsulinemia/insulin resistance and related physiologic effects that could lead to an increased risk of this cancer. Such mechanisms have been postulated to explain the link between diabetes and subsequent pancreas cancer (see “Diabetes” Section). Alternative mechanisms are also possible, including the altered metabolism in adipocytes and/or adipose tissue that occurs with obesity. Adipocytes secrete a variety of molecules with endocrine, paracrine, and autocrine effects (Rose et al., 2004). Obesity can increase production of adipocytokines, cytokines produced by adipocytes, that are biologically relevant to carcinogenesis in a number of tissues including pancreas. For example, TNF-alpha, hepatocyte growth factor, heparin-binding epidermal growth factor like-growth factor, and intereukin-6, which stimulate angiogenesis, are all increased in adipose tissues by obesity (Rose et al., 2004). Some of these same cytokines (e.g., TNF-alpha and intereukin-6) also have proinflammatory effects. Recent laboratory studies demonstrated that macrophage numbers increase in adipose tissue as a result of obesity, and macrophages are an important source of cytokines and nitric oxide (Weisberg et al., 2003). The damaging oxidative effects on cellular DNA may result in mutations that lead to pancreatic cancer. The epidemiologic literature suggests a small to modest increased risk of pancreatic cancer with increasing obesity. However, measurement of overweight and obesity is complex and error prone. The magnitude of the risks observed does not allow us to preclude the possibility that chance, bias, or confounding could cause the apparent increased risks. More studies, based on the measures of anthropome-

Cancer of the Pancreas try recommended by the World Health Organization (WHO, 1995), and ideally including biomarkers, are required to determine whether this association is, in fact, causal.

Diet The pancreas is, of course, intimately related to digestion and absorption. Since diet is implicated in the cause and modulation of cancer at other gastrointestinal sites, it seems reasonable to place diet high among the possible causal elements for pancreatic carcinoma (Lowenfels and Maisonneuve, 2002; Longnecker et al., 1990; Howe and Burch, 1996; Potter and Steinmetz, 1996; WCRF, 1997). In considering the relation of risk with specific dietary constituents, however, it is worth considering that, unlike every other part of the gastrointestinal tract, the pancreas is never exposed either directly (mouth to anus) or indirectly (liver) to ingested foods or their modified, digested, and absorbed products. Accordingly, the effects of diet on pancreatic carcinogenesis must be via changes in the internal metabolic environment of that organ and/or exposure to blood-borne carcinogens. The empirical approach, via ecologic studies, to the question of the relation between diet and pancreas cancer has raised interest in fat (Lea, 1967; Segi and Kurihara, 1972; Ghadirian et al., 1991c), eggs (Lea, 1967; Armstrong and Doll, 1975; Ghadirian et al., 1991c), animal protein (Lea, 1967; Armstrong and Doll, 1975; Ghadirian et al., 1991c), milk (Lea, 1967; Ghadirian et al., 1991c), sugar (Yanai et al., 1979), and plant foods/protein (Ohba et al., 1996). The World Cancer Research Fund (WCRF) and the American Institute for Cancer Research (AICR) published a report in 1997: Food, Nutrition and the Prevention of Cancer: A Global Perspective (WCRF, 1997), reflecting the consensus of an expert panel that evaluated evidence for dietary and related factors and disease associations and deemed these as convincing, probable, possible, or insufficient. For pancreatic cancer, no factors had evidence that reached the level of convincing. A decreased risk associated with vegetable and fruit consumption was considered probable. It was also considered probable that there was no increased risk for pancreatic cancer associated with either alcohol or coffee consumption. The panel considered it possible that non-starch polysaccharides (the principal and readily measurable component of dietary fiber) and vitamin C were associated with decreased risks and that high intake of energy, cholesterol, and meat were possibly associated with increased risks. The consensus for high body mass (intimately related to diet) and high tea intake was that there was possibly no relationship. Finally, the panel concluded there was insufficient evidence to make judgments for sugar, eggs, cured and smoked meat, and fish. A number of individual-level studies have examined dietary habits in relation to pancreatic cancer risk subsequent to the WCRF/AICR report, and the overall conclusions remain much the same, with the exception of the obesity effect reported in eight positive studies since 1997. Importantly, there have been at least seven case-control studies (Ji et al., 1995b; Fernandez et al., 1996; Soler et al., 1998; Silverman et al., 1998; Anderson et al., 2005; Tavani et al., 2000; Inoue et al., 2003) and five cohorts studies of pancreatic cancer that have assessed some aspects of diet subsequent to the WCRF/AICR report. They include the following: The Iowa Women’s Health Study (IWHS) cohort that included approximately 34,000 post-menopausal women in Iowa (US) (Zheng et al., 1996; Harnack et al., 1997); the American Cancer Society’s (ACS) Cancer Prevention Study II (CPS II) cohort that included approximately one million ACS volunteers (and over 3700 deaths from pancreas cancer) (Couglin et al., 2000); the Nurses’ Health Study cohort that included approximately 98,500 female registered nurses (Michaud et al., 2001, 2002, 2003); the AlphaTocopherol, Beta-Carotene (ATBC) Cancer Prevention Study in Finland (Stolzenberg-Solomon et al., 2002), a cohort of about 27,100 male smokers; and the Multiethnic Cohort Study in Hawaii and Los Angeles with over 190,000 subjects (Nöthlings et al., 2005). Here we present an overview of the epidemiologic data on diet and human pancreatic cancer (Table 38–2). Some possible carcinogenic and anticarcinogenic mechanisms of dietary components will be discussed.

727

Meat, Poultry, Fish The association of meat intake with pancreatic cancer has been examined in at least seven cohort studies (Mills et al., 1988; Hirayama, 1989; Zheng et al., 1993; Coughlin et al., 2000; Stolzenberg-Solomon et al., 2002; Michaud et al., 2003; Nöthlings et al., 2005) and 19 casecontrol studies (Gold et al., 1985; Mack et al., 1986; Norell et al., 1986a; Raymond et al., 1987; Falk et al., 1988; Olsen et al., 1989; Farrow and Davis, 1990b; La Vecchia et al., 1990; Bueno de Mesquita et al., 1991b; Baghurst et al., 1991; Mizuno et al., 1992; Lyon et al., 1993; Ji et al., 1995b; Ghadirian et al., 1995, Fernandez et al., 1996; Soler et al., 1998; Silverman et al., 1998; Tavani et al., 2000; Anderson et al., 2002a) (Table 38–2). Evidence for positive associations has been reported for the following: beef and bacon (Mack et al., 1986); fried and grilled meats (Norell et al., 1986a; Anderson et al., 2002a); pork products and beef (Falk et al., 1988); beef and pork (Olsen et al., 1989); daily meat consumption (Hirayama, 1981; 1989); beef, chicken, and pork (Farrow and Davis, 1990b); ham and meat (La Vecchia et al., 1990); all meat and red meat (Zheng et al., 1993; Fernandez et al., 1996; Tavani et al., 2000); nitrated meats (bacon, sausages, and hotdogs) in men (Lyon et al., 1993); liver, ham, and sausages (Soler et al., 1998); and processed meat, pork and total redmeat (Nöthlings et al., 2005). Other studies have reported generally null associations for various meats and fish including the following: bologna, salami, processed meats (Silverman et al., 1998), smoked/barbequed beef, bacon/sausage, beef (Gold, et al., 1985), red meats (Silverman et al., 1998; Coughlin et al., 2000; Michaud et al., 2003), meats (Baghurst et al., 1991; Mizuno et al., 1992; Mills et al., 1988), red meat or total meat, poultry, and fish (Stolzenberg-Solomon et al., 2002; Michaud et al., 2003). Inverse associations have also been found for fish (Mizuno et al., 1992; Soler et al., 1998; La Vecchia et al., 1990), and poultry and fish in women, but not men (Silverman et al., 1998). Raymond et al. (1987) found that meat intake did not alter the risk except for lean pork that was associated with a decreased risk.

Meat Preparation. Populations and individuals vary greatly in meat cooking practices and doneness preferences. The levels of several potential human carcinogens also vary with methods of preparation and doneness preferences (Sugimura et al., 2000) and failure to consider these, in addition to meat consumption, may result in misclassification of the relevant carcinogens and masking of true associations (Sinha and Rothman, 1999). Some of the inconsistent findings with respect to meat, poultry, and fish intake may result from the fact that few studies have assessed meat and fish preparation, as well as intake, with respect to risk of pancreatic cancer. Positive associations were reported for greater consumption of fried and grilled meat and fish, but not with meat prepared in other ways in a Swedish case-control study (Norell et al., 1986a). A general pattern of increased risks, though not always statistically significant, have been associated with deep fried, grilled, cured, and smoked foods in a case-control study from Shanghai (Ji et al., 1995b); salted/smoked meat or fish consumption in a US cohort (Zheng et al., 1993); and smoked meat and fried food in the Francophone community of Montreal, Canada (Ghadirian et al., 1995). Null results were reported for fried meat intake in the Finnish cohort of male smokers (StolzenbergSolomon et al., 2002). In a population-based case-control study of pancreatic cancer in the United States (Anderson et al., 2002a, 2005), detailed information on cooking practices and doneness levels was collected for specific types of commonly consumed meats. Mean levels of total meat and red meat consumption were higher in cases than controls, but neither were strong or statistically significant predictors of risk. However, higher consumption of grilled/barbequed red meat was associated with a statistically significant positive trend (P < 0.001). Fried meat intake was also positively associated, though the findings were not statistically significant. Overall, the data suggest an increased risk for pancreas cancer associated with meat consumption, and this elevation may be due, in part, to particular cooking or processing methods (WCRF, 1997). Numerous mutagens and carcinogens are formed in meat and fish cooked at

Table 38–2. Studies of Diet and Pancreas Cancer Authors (year)

Study Type and Size*

vegetables and fruit Gold et al. (1985)

Mack et al. (1986)

Norell et al. (1986a)

Comparison†

Relative Risk Estimate by Control Category

Hospital-based case-control 201/402

raw fruits and vegetables

hospital

population

Ever vs. never ≥ vs. <2/week ≥ vs. <5/week

0.5 0.5‡ 0.6‡

0.2‡ 0.7 0.6‡

Population-based case-control 490/490

fresh fruits or vegetables

direct interview all

Population-based case-control 99/301

vegetables

£5/week ≥5/week

<1/week Every week Almost daily

1.0 0.8

1.0 0.7‡

hospital

population

1.0 1.0 0.5

1.0 1.0 0.8

1.0 0.5** 0.5

1.0 0.7 0.6

1.0 0.6 0.3**

1.0 1.0 0.5**

1.0 0.9 0.6

1.0 0.6 0.6

raw vegetables <1/week Every week Almost daily

citrus fruits <1/week Every week Almost daily

fruit juices <1/week Every week Almost daily Voirol et al. (1987), Raymond (1987)

Population-based case-control 88/336

vegetables

population

<1110 g/week 1110–1609 g/week ≥1610 g/week

1.0 0.87 0.47**

fruits

<280 g/week 280–619 g/week ≥620 g/week Falk et al. (1988)

Mills et al. (1988)

Hospital-based case-control 363/1234

Cohort 40/193,483

1.0 1.1 0.56

fruits and juices

<25/month 25–63/month ≥64/month

beans, lentils, peas <1/week ≥3/week

male

female

1.0 0.6‡ 0.4‡

1.0 0.6 0.5‡

non-cases 1.0 0.4

dried fruits <1/month ≥3/week

1.0 0.4‡

vegetarian protein products

Hirayama (1989)

Cohort 679/265,118†† ¥17y follow-up

Olsen (1989)

Mortality-based case-control 212/220

<1/week ≥3/week

1.0 0.4‡

green salad cooked green vegetables green yellow vegetables

>1.0 >1.0

cruciferous vegetables £2 month ≥9/month

No difference between cases and non-cases

population 1.0 0.6

non-cruciferous vegetables £16 month 17–31/month ≥32/month

1.0 0.55† 0.95

fruit and juices Farrow and Davis (1990b)

728

Population-based case-control 148/188

£21/month ≥53/month

1.0 0.88

vegetables

population

U vs. L quartile

No association

Table 38–2. (cont.) Authors (year)

Study Type and Size*

Comparison†

Relative Risk Estimate by Control Category

raw vegetables U vs. L quartile

No association

green and yellow vegetables U vs. L quartile

No association

fruit U vs. L quartile

No association

citrus fruit U vs. L quartile La Vecchia (1990)

Hospital-based case-control 247/1089

No association

fresh fruit (tertiles)

population

1 3

1.0 0.68‡

green vegetables (tertiles) 1 3 Bueno de Mesquita (1991b)

Population-based case-control 164/480

1.0 0.84

vegetables (quintiles)

direct interview

1 5

1.0 0.22§

1.0 0.34§

1.0 0.32§

1.0 0.35§

1.0 0.30§

1.0 0.42§

— —

1.0 1.1

— —

1.0 0.77

cooked vegetables (quintiles) 1 5

raw vegetables (quintiles) 1 5

fruits total (quintiles) 1 5

fruit juices (quintiles) 1 5 Baghurst et al. (1991)

Population-based case-control 104/253

case consuming less than controls Tomato Dried grapes

Mizuno et al. (1992)

Hospital-based case-control 124/124

green, yellow vegetables

hospital

Daily vs. less often

1.2

male

female

0.01 < P < 0.05 P < 0.01

0.01 < P < 0.05 P < 0.01

other vegetables Daily vs. less often

0.71

fruit Daily vs. less often Lyon et al. (1993)

Population-based case-control 149/363

0.62

vegetables

male

female

Low Medium High

1.0 0.97 0.99

1.0 0.67 0.32‡§

1.0 1.05 0.81

1.0 0.88 0.37‡§

fruits Low Medium High Shibata et al. (1994)

Cohort 65/13,979

fruits

non-cases

Low Medium High

1.0 0.90 0.89

vegetables Low Medium High

1.0 0.96 0.82

dark green vegetables Low Medium High

1.0 0.70 1.18

yellow vegetables Low Medium High

1.0 0.64 0.60

(continued)

729

Table 38–2. (cont.) Authors (year)

Study Type and Size*

Comparison†

Ji et al. (1995)

Population-based case-control 451/1552

all fruits (quartiles)

male

female

1 4

1.0 0.66§

1.0 0.58

1.0 0.65‡§

1.0 0.58‡

1.0 0.58‡§

1.0 0.45‡§

1.0 0.69

1.0 0.59

1.0 0.68

1.0 0.85

1.0 1.37

1.0 3.07‡§

1.0 0.68§

1.0 0.67

1.0 0.84

1.0 0.96

1.0 0.79

1.0 1.07

1.0 0.94

1.0 0.55‡§

1.0 0.90

1.0 0.79

Relative Risk Estimate by Control Category

oranges (quartiles) 1 4

bananas (quartiles) 1 4

apples (quartiles) 1 4

other fruits (quartiles) 1 4

preserved vegetables (quartiles) 1 4

all vegetables (quartiles) 1 4

dark green leafy vegetables (quartiles) 1 4

cruciferous vegetables (quartiles) 1 4

legumes (quartiles) 1 4

soybean products (quartiles) 1 4 Fernandez et al. (1996)

Hospital-based case-control study 362/1408

fruit intake

male

female

High Intermediate Low

1.0 1.3 1.7‡

1.0 1.0 1.6

Soler et al. (1998)

Hospital-based case-control 362/1552

carrots (tertiles)

hospital

1 3

1.0 0.97

fresh fruit (tertiles) 1 3

1.0 0.59‡§

green vegetables (tertiles) 1 3 Silverman et al. (1998)

Population-based case-control 436/2003

1.0 0.87

fruit (quartiles)

male

female

1 4

1.0 0.9

1.0 1.1

1.0 0.8§

1.0 0.9

1.0 0.9

1.0 1.2

1.0 0.7

1.0 0.8

1.0 1.0

1.0 1.1

1.0 0.6‡§

1.0 0.9

raw fruit (quartiles) 1 4

citrus (quartiles) 1 4

non-citrus (quartiles) 1 4

fruits rich in vitamin a (quartiles) 1 4

vegetables (quartiles) 1 4

730

Table 38–2. (cont.) Authors (year)

Study Type and Size*

Comparison†

Relative Risk Estimate by Control Category

cruciferous Vegetables 1 4

1.0 0.5‡§

1.0 0.4‡§

1.0 0.8

1.0 0.6‡§

1.0 0.6†

1.0 1.0

1.0 0.8

1.0 0.8

1 4

1.0 0.8

1.0 0.4‡§

citrus fruit/juice consumption (quartiles)

male

female

1 4

1.0 1.0

1.0 0.9

1 4

1.0 0.9

1.0 0.9

vegetables (quintiles)

male

1 5

1.0 0.77

dark green (quartiles) 1 4

dark yellow (quartiles) 1 4

legumes (quartiles) 1 4

raw (quartiles) Coughlin et al. (2000)

American Cancer Society, Cancer Prevention Study II, Cohort 3751/483,109 men 619,199 women

Stolzenberg-Solomon Alpha-Tocopherol, Betaet al. (2002) Carotene Cancer Prevention Study Cohort 163/27,111 Male Smokers

vegetable consumption (quartiles)

fresh vegetables (quintiles) 1 5

1.0 0.96

cooked vegetables (quintiles) 1 5

1.0 1.13

cruciferous vegetables (quintiles) 1 5

1.0 0.82

root vegetables (quintiles) 1 5

1.0 0.69

legumes (quintiles) 1 5

1.0 0.89

all fruits and berries (quintiles) 1 5

1.0 0.85

citrus fruits (quintiles) 1 5

1.0 0.79

berries (quintiles) 1 5

1.0 0.72

vegetables and legumes (quintiles) 1 5

1.0 0.72

vegetables, fruits, and legumes (quintiles) 1 5 Inoue et al. (2003)

Nested case-control 200/2000

1.0 0.74

raw vegetables

male

female

Less Everyday

1.0 0.65

1.0 0.78

(continued)

731

Table 38–2. (cont.) Authors (year)

Study Type and Size*

meat, eggs, and dairy products Gold et al. (1985)

Mack et al. (1986)

Hospital-based case-control 201/402

Population-based case-control 490/490

Comparison†

Relative Risk Estimate by Control Category

butter

hospital

population

2.4‡ <1.0 <1.0

1.1

<2 vs. >2/week

deep fried foods beef beef <5/week ≥5/week

interview 1.0 2.1‡

1.0 1.2

1.0 0.7

1.0 0.8

1.0 1.4

1.0 0.9

eggs

<5/week ≥5/week

fried bacon/ham <5/week ≥5/week Norell et al. (1986a)

Voirol et al. (1987), Raymond (1987)

Population-based base-control 99/301

fried/grilled meat

Population-based case-control 88/336

red meat

<1/week Every week Almost daily

hospital

population

1.0 1.5 4.6**

1.0 1.7** 13.4**

population

<480 g/week ≥480 g/week

1.0 0.77

poultry

<18 g/week 18–59 g/week ≥60 g/week

1.0 1.6 1.1

lean pork 0 g/week <150 g/week >150 g/week

1.0 0.44 0.62

butter

<108 g/week 108–194 g/week ≥195 g/week

1.0 1.4 2.0**

margarine Any vs. none

0.35**

whole milk

Falk et al. (1988)

Hospital-based case-control 363/1234

0 ml/week <630 ml/week ≥630 ml/week

1.0 1.2 1.5

beef

male

female

1.0 1.2 1.1

1.0 0.8 0.7

1.0 1.4 1.7‡

1.0 1.6 1.3

1.0 1.6 2.2‡

1.0 1.6 1.0

1.0 1.0 1.0

1.0 1.2 1.9

<6/month 6–15 month ≥16/month

pork

<9/month 9–30/month ≥31/month

dairy

<34/month 34–67/month ≥68/month

seafood <2/month 2–7/month ≥8/month Mills et al. (1988)

Cohort 40/193,483

current use of meat, poultry, fish <1/week 1–2/week ≥3/week

non-cases 1.0 0.8 2.2

current use of eggs <1/week 1–2/week ≥3/week

butter milk

732

1.0 1.5 2.5‡ ~1.0 ~1.0

Table 38–2. (cont.) Authors (year)

Study Type and Size*

Comparison†

Relative Risk Estimate by Control Category

meat, eggs, and dairy products (cont.) Hirayama (1989)

Olsen et al. (1989)

Cohort 679/265,118†† ¥ 17y follow-up

meat¶

Mortality-based case-control 212/220

beef

Never Occasionally Daily

1.0 1.4 1.8

population

£8/month ≥18/month

1.0 1.8

pork

£2/month ≥9/month

1.0 1.9

poultry Farrow and Davis (1990b)

Population-based case-control 148/188

£2/month ≥6/month

1.0 0.95

beef

population

<1.6/month 2.5–4.0/week >4.0/week

1.0 2.4‡ 2.8‡

chicken

<0.71/week 1.1–2.0/week >2.0/week

1.0 1.8 2.5‡

pork

<0.21/week 0.51–1.0/week >1.0/week La Vecchia et al. (1990)

Hospital-based case-control 247/1089

1.0 1.3 1.7

meat (tertiles)

hospital

1 3

1.0 1.1

ham (tertiles) 1 3

1.0 1.4

fish (tertiles) 1 3

1.0 0.74

eggs (tertiles) 1 3

1.0 1.4

milk (tertiles) 1 3 Bueno de Mesquita (1991b)

Population-based case-control 164/480

1.0 0.67

pork (quintiles)

direct interview all

1 5

— —

1.0 1.4

— —

1.0 0.51

1.0 2.5§

1.0 1.8

1.0 0.44

1.0 1.81

1.0 1.6

1.0 2.2§

— —

1.0 0.78

1.0 —

1.0 1.1

beef (quintiles) 1 5

fish (quintiles) 1 5

cheese (quintiles) 1 5

eggs (quintiles) 1 5

milk and milk products (quintiles) 1 5

oil and fats (quintiles) 1 5

(continued)

733

Table 38–2. (cont.) Authors (year)

Study Type and Size*

Baghurst et al. (1991) Population-based case-control 104/254

Comparison†

Relative Risk Estimate by Control Category

cases consuming more than controls Omelet Boiled egg

cases consuming less than controls Fried fish Mizuno et al. (1992)

Hospital-based case-control 124/124

meat

hospital

Every day vs. less often

1.18

male

female

0.01 < P < 0.05 P < 0.01

P < 0.01 P < 0.01

0.01 < P < 0.05

No association

fish Every day vs. less often

0.56‡

milk Every day vs. less often Lyon et al. (1993)

Population-based case-control 149/363

0.41‡

red meats

male

female

Low Medium High

1.0 0.64 1.41

1.0 1.05 1.44

1.0 1.28 2.77‡§

1.0 0.85 1.08

nitrated meats Low Medium High Zheng et al. (1993)

Cohort 57/17,633

all meat (quartiles)

male

1 4

1.0 3.0‡§

red meat (quartiles) 1 4

1.0 2.4‡§

chicken (quartiles) 1 4

1.0 1.9

all fish (quartiles) 1 4

1.0 1.4

eggs and dairy (quartiles) 1 4

1.0 0.7

salted/smoked meat and fish (quartiles) 1 4 Ji et al. (1995b)

Population-based case-control 451/1552

1.0 1.5

fresh red meats (quartiles)

Male

Female

1 4

1.0 0.73

1.0 1.24

1.0 0.82

1.0 1.21

1.0 0.93

1.0 1.06

1.0 1.29

1.0 0.94

1.0 0.67

1.0 0.46‡§

1.0 1.37 1.37

1.0 1.16 1.30

1.0 1.13 4.08‡§

1.0 1.23 0.58

organ meats (quartiles) 1 4

poultry (quartiles) 1 4

fish (quartiles) 1 4

eggs (quartiles) 1 4

deep fried foods Never/seldom Sometimes Frequently

grilled foods Never/seldom Sometimes Frequently

734

Table 38–2. (cont.) Authors (year)

Study Type and Size*

Comparison†

Relative Risk Estimate by Control Category

cured foods Never/seldom Sometimes Frequently

1.0 1.18 1.40

1.0 1.77‡ 2.14‡§

1.0 0.85 1.67

1.0 0.63 4.86

smoked foods Never/seldom Sometimes Frequently Ghadirian et al. (1995)

Population-based case-control 179/239

smoked food (meat)

population

Never Rarely Once in a while Often

1.0 0.84 1.68 4.68‡§

fried food Never Rarely Once in a while Often Very often

1.0 1.92‡ 2.28‡ 3.84‡ 16.71‡§

Fernandez et al. (1996)

Hospital-based case-control 362/1408

meat intake

male

female

Low Medium High

1.0 1.1 1.3

1.0 1.5 1.3

Soler et al. (1998)

Hospital-based case-control 362/1552

milk (tertiles)

hospital

1 3

1.0 1.05

meat (tertiles) 1 3

1.0 1.43‡

liver (tertiles) 1 3

1.0 1.43‡

eggs (tertiles) 1 3

1.0 1.13

ham and sausages (tertiles) 1 3

1.0 1.64‡§

fish (tertiles) 1 3

1.0 0.65‡§

cheese (tertiles) 1 3

1.0 1.21

butter (tertiles) 1 3

1.0 0.90

margarine (dichotomous) 1 2

1.0 1.21

oil (tertiles) 1 3 Silverman et al. (1998)

Population-based case-control 436/2003

1.0 0.58‡§

meat, poultry, and fish (quartiles)

male

female

1 4

1.0 0.9

1.0 0.6§

1.0 1.5

1.0 0.5‡§

1.0 0.7

1.0 1.0

1.0 0.8

1.0 0.7

poultry and fish (quartiles) 1 4

red meat (quartiles) 1 4

processed meat (quartiles) 1 4

(continued)

735

Table 38–2. (cont.) Authors (year)

Study Type and Size*

Comparison†

Relative Risk Estimate by Control Category

dairy products (quartiles) Coughlin et al. (2000)

Tavani et al. (2000)

1 4

1.0 0.8

1.0 0.9

American Cancer Society, Cancer Prevention Study II, Cohort 375/1,102,308

red meat consumption (quartiles)

male

female

1 4

1.0 1.1

1.0 0.9

Series of coordinated hospital-based case-control 362/7990

red meat

hospital

Low Intermediate High

1.0 1.2 1.6‡§

increment of 1 portion/day milk products (quintiles)

1.5‡

male

Stolzenberg-Solomon Alpha-Tocopherol, Betaet al. (2002) Carotene Cancer Prevention 1 Study Cohort 5 163/27,111 Male Smokers

1.0 1.08

1 5

1.0 1.47

milk (whole and low-fat) (quintiles) sour milk products (quintiles) 1 5

1.0 0.76

cheese (quintiles) 1 5

1.0 0.73

cream (quintiles) 1 5

1.0 1.50

butter (quintiles) 1 5

1.0 1.4§

vegetable oils (quartiles) 1 4

1.0 0.92

red meat (quintiles) 1 5

1.0 0.95

beef (quintiles) 1 5

1.0 1.30

poultry (quartiles) 1 4

1.0 1.25

fish (quintiles) 1 5

1.0 0.91

pork (quintiles) 1 5

1.0 1.01

fried meat (quintiles) 1 5

1.0 0.98

processed meat (quintiles) 1 5

1.0 1.04

processed fish (quintiles) 1 5

1.0 1.22

organ meats (quartiles) 1 4

1.0 0.71

eggs (quintiles) 1 5

736

1.0 0.86

Table 38–2. (cont.) Authors (year)

Study Type and Size*

Comparison†

Anderson et al. (2002)

Population-based case-control 193/674

grilled/barbecued meat (quintiles)

population

1 and 2 5

1.0 2.2‡§

Relative Risk Estimate by Control Category

fried red meat (quintiles) 1 and 2 5

1.0 1.4

broiled red meat (quintiles) 1 and 2 5 Michaud et al. (2003)

Cohort (Nurse’s Health Study) total 172/88,802 1 5

1.0 0.7

meat (quintiles)

female 1.0 0.94

red meat (quintiles) 1 5

1.0 0.87

dairy products (quintiles) 1 5

1.0 1.04

beef, pork or lamb (main dish)

<3/month ≥5/week

1.0 0.75

beef, pork, or lamb (sandwich) 0 ≥2/week

1.0 0.95

processed meats 0 ≥2/week

1.0 1.28

bacon 0 ≥2/week

1.0 1.05

hamburger 0 ≥2/week

1.0 1.03

hot dogs 0 ≥2/week

1.0 0.69

chicken without skin 0 ≥2/week

1.0 1.05

chicken with skin 0 ≥2/week

1.0 1.27

skim milk

<4/month ≥5/week

1.0 1.0

hard cheese

<4/month ≥5/week

1.0 1.08

butter 0 ≥5/week

1.0 0.89

eggs

<2/week ≥5/week

1.0 1.25

fish

<4/month ≥2/week Nöthlings et al. (2005)

Multiethnic Cohort Study 482/190,545

1.0 1.30

fish (quintiles)

non-cases

1 5

1.0 0.91

poultry (quintiles) 1 5

1.0 1.01

(continued)

737

738

PART IV: CANCER BY TISSUE OF ORIGIN

Table 38–2. (cont.) Authors (year)

Study Type and Size*

Comparison†

Relative Risk Estimate by Control Category

beef (quintiles) 1 5

1.0 1.21§

pork (quintiles) 1 5

1.0 1.53‡§

red meat (quintiles) 1 5

1.0 1.45‡§

processed meat (quintiles) 1 5

1.0 1.68‡§

dairy products (quintiles) 1 5

1.0 1.05

eggs (quintiles) 1 5

1.0 0.94

g, grams; L, lowermost; U, uppermost. *Study size as cases/controls or cases/person-years; †Frequency of consumption or quantiles; ‡ 95% CI excludes 1.0; **90% CI excludes 1.0; ††Initial cohort size; ‡‡All interviews were indirect; § P (trend) <0.05; ¶Also shows there is an interaction with smoking.

high temperatures (Sugimura et al., 2000; Knize et al., 1999; Felton et al., 1995), including heterocyclic amines (HCAs) and polycyclic aromatic hydrocarbons (PAHs). Formation of these compounds depends on cooking method, temperature, and the degree of doneness (Knize et al., 1999; Layton et al., 1995, Sinha and Rothman, 1999). While baked, stewed, and microwaved meats do not contain these compounds, well-done pan-fried and barbecued/grilled meats typically contain high levels (Sinha and Rothman, 1999). HCAs can be found in the urine of individuals eating normal meat-containing diets (Ushiyama et al., 1991), and in controlled feeding studies of meat cooked at different temperatures, the urinary mutagenicity was correlated with mutagenicity of the meats (Peters et al., 2004). Polycyclic aromatic hydrocarbons (e.g., benzo(a)pyrene) and the HCAs are reasonable candidates for human pancreatic carcinogens (Weisburger, 2000; Anderson et al., 2005), and they represent a substantial portion of meat-derived mutagens/carcinogens in their respective classes (Layton et al., 1995; Kazerouni et al., 2001). In rodent models, PhIP, the most mass abundant of the HCAs, forms high levels of DNA adducts in the pancreas (Kaderlik et al., 1994; Fretland et al., 2003) and is preferentially taken up by pancreatic acini (Butcher et al., 1996); such data suggest that HCAs reach the pancreas and bind to DNA. Dietary N-nitroso compounds, activated in the liver and bloodborne to the pancreas, may play a role in human pancreatic cancer (Risch, 2003). N-Nitroso compounds are present in the human diet, particularly in cured and smoked meat and fish, as well as in other pickled and salt-preserved foods, such as vegetables, and in beer and some cheeses (Tricker and Perussmann, 1991; IARC, 1984). Nitrosamines and nitrosamides can also be formed endogenously via the combination of nitrosating compounds (such as nitrite) with amines or amides, respectively. Preserved meats can contain nitrites and some drinking water contains both nitrites and nitrates, but the primary source of nitrates in the diet is usually vegetables and fruits (Coss et al., 2004; IARC, 1984). Dietary nitrate can be converted in the digestive tract to nitrite (IARC, 1984). Coss and colleagues (Coss et al., 2004), in a study in Iowa (US), found that nitrite from animal sources was associated with a statistically significant increase of pancreas cancer in both men and women, but that dietary nitrate, mostly of vegetable origin was not associated with an increased risk (nor was there an association for nitrates from drinking water).

In human feeding studies, levels of excreted N-nitroso compounds were significantly greater on high vs. low red meat diets (Cross et al., 2003). In addition, endogenous N-nitrosation was increased when diets were supplemented with heme iron, but not with inorganic iron. Aside from HCAs, PAHs, and N-nitroso compounds, there are other components of meat, such as iron, salt (in preserved meats), and fat (discussed below) that may be relevant to carcinogenesis in the pancreas (Sugimura, 2000; Weisburger, 2000). The use of biomarkers for specific agents of interest will allow for more refined hypothesis testing in future epidemiologic studies.

Dairy Products and Eggs A number of studies have examined dairy products (Table 38–2) and at least eight, including three prospective studies, have found no major associations (Lyon et al., 1993; Zheng et al. 1993; Soler et al., 1998; Silverman et al. 1998; Michaud et al., 2003; Nöthlings et al., 2005). Among male smokers, Stolzenberg-Solomon et al. (2002) found an elevated risk associated with increasing energy-adjusted butter consumption that was statistically significant, but a null association for milk. In an analysis from the large multiethnic cohort, results were null for intake of dairy products (Nöthlings et al., 2005). Results have been mixed in three case-control studies (Bueno de Mesquita et al., 1991b; Falk et al., 1988; Olsen et al., 1989), and inverse associations were reported in another (La Vecchia et al., 1990) for a variety of dairy products. In general, the associations for this food group, particularly among more recent studies, are null or weak. At least 13 studies have looked at the relationship between eggs and pancreatic cancer with three studies finding positive associations with higher intake of eggs or omelets (Bueno de Mesquita et al., 1991b; Mills et al., 1988; Baghurst et al., 1991), one finding an inverse association (Ji et al., 1995b), and nine finding weak or null associations (Gold et al., 1985; Mack et al., 1986; Farrow and Davis, 1990b; La Vecchia et al., 1990; Soler et al., 1998; Silverman et al., 1998; Stolzenberg-Solomon et al., 2002; Michaud et al., 2003; Nöthlings et al., 2005). Overall, there is little evidence to support an association between egg intake and pancreas cancer. Eggs are a source of cholesterol and findings with respect to cholesterol are considered below.

Vegetables and Fruits Vegetable and fruit intake in relation to pancreas cancer risk has been examined in at least 23 studies (Table 38–2) in a variety of ways

Cancer of the Pancreas including: fruits and vegetables—separately and combined, in groups (e.g., cruciferous vegetables or citrus fruits), and by individual type of vegetable or fruit (e.g., peas or bananas). Of six cohort studies examining the association between pancreatic cancer and vegetables and fruit, one found evidence for an inverse association that was statistically significant (Mills et al., 1988). Zheng et al. (1993) reported no consistent associations and Shibata et al. (1994) found weak inverse associations that were not statistically significant. The ATBC cohort found reduced risks for greater consumption of both vegetables and fruits, but these were not statistically significant (Stolzenberg-Solomon et al., 2002). The ACS CPSII cohort found no significant association with high intake of either fruits or vegetables in men or women with RRs ranging from 0.9–1.0 (Coughlin et al., 2000). The results in a Japanese cohort were also null (Hirayama, 1989). Nearly all case-control studies (an exception being Farrow and Davis, 1990b) showed evidence, though not always statistically significant, of an inverse association for fruits and or vegetable consumption categorized in a variety of ways (Gold et al., 1985; Mack et al., 1986; Norell et al., 1986a; Voirol et al., 1987; Falk et al., 1988; Olsen et al., 1989; La Vecchia et al., 1990; Baghurst et al., 1991; Bueno de Mesquita et al., 1991b; Mizuno et al., 1992; Inoue et al., 2003; Lyon et al., 1993; Ji et al., 1995b; Fernandez et al., 1996; Soler et al., 1998; Silverman et al., 1998). Among the more recent studies that have examined associations for vegetables and fruits, either by individual type or group, inverse associations were found for high vs. low intakes of yellow vegetables (Shibata et al., 1994); oranges, bananas, and apples in both genders (Ji et al., 1995b); cruciferous vegetables for men, and legumes for women (Ji et al., 1995b). Silverman et al. (1998) reported significant inverse trends for higher consumption of cruciferous vegetables in men and women. The ACS CPSII study found no relationship between increased consumption of citrus fruits/juice and pancreatic cancer (Coughlin et al., 2000). Although the ATBC cohort generally found inverse associations for greater intakes of individual fruits and vegetables, they found no significant trends (Stolzenberg-Solomon et al., 2002). In the hospital-based case-control study reported by Soler et al. (1998), a significant inverse association was found in the third compared to first tertile of fresh fruit consumption, while no significant associations were found for either higher intake of carrots or green vegetables. Finally, in a nested case-control study in Japan, inverse associations were observed for daily consumption of raw vegetables for both men and women (Inoue et al., 2003). Overall, the preponderance of evidence indicates a lower risk of pancreas cancer associated with a lifestyle that includes a higher intake of plant foods. This association does not appear to be confined to a specific group or groups of plant food. Of note, case-control studies appear to show stronger effects than do cohort studies and it will be important to distinguish causal effects from potential bias (McCollough and Giovannucci, 2004).

Possible Mechanisms. Evidence on mechanisms by which vegetables and fruit might be protective against cancer in general has been reviewed (Stenimetz and Potter, 1991; McCullough and Giovannucci, 2004). Vegetables and fruit contain many compounds that are potential anticarcinogens including: carotenoids, vitamins C and E, isoflavones, phenols, protease inhibitors, isothiocyanates, flavonoids, dithiolthiones, indoles, glucosinolates, selenium, plant sterols, allium compounds, and limonene. A variety of protective mechanisms—many complex—by which vegetables and fruit may lower the risk of cancer have been postulated (Wattenberg, 1985; Adlercreutz, 2002; Steinmetz and Potter, 1996; McCullough and Giovannucci, 2004). Several mechanisms of possible importance to pancreatic cancer will be discussed briefly. A number of compounds present in plants can induce glutathione S-transferase and increase levels of glutathione (GSH). These include isothiocyanates and dithiolthiones, which are present in cruciferous vegetables, and limonene, the major component of citrus oil. There is evidence that GSH, with glutathione S-transferase, can detoxify activated metabolites of a number of possible pancreatic car-

739

cinogens including HCAs and PAHs (Coles et al., 2001; Steinkellner et al., 2001). Ascorbic acid (vitamin C), alpha-tocopherol, and certain polyphenols can interfere with the process of nitrosamine formation by inhibiting the conversion of dietary nitrate to nitrite by bacteria in the saliva and stomach (Steinmetz and Potter, 1996; IARC, 1984; Mirvish, 1986). Isothiocyanates, present in cruciferous vegetables, are thought to inhibit activation of carcinogens, such as tobacco-specific nitrosamines, through inhibition of cytochrome-P450 enzymes (Hecht, 2000; Bartsch and Frank, 1996). Naturally occurring coumarins, which are found in citrus fruits, can inhibit carcinogen-metabolizing enzymes and block the activation of benzo[a] pyrene in cultured human cells (Kleiner et al., 2003).

Cereals Some authors have reported increased risks associated with carbohydrate food groups (e.g., rice, breads, and cereals) (Falk et al., 1988); white, but not whole grain, bread (Gold et al., 1985); breads and cereals (Olsen et al., 1989); breads, grains, and rice in men but not women (Silverman et al., 1998), and pasta (Raymond et al., 1987). On the other hand, others have found no increased or decreased risks associated with these foods. Inverse associations have been reported for whole grain products in some (Mack et al., 1986; Chatenoud et al., 1998), but not other studies (Soler et al., 1998). No substantive associations were reported in the Lutheran Brotherhood cohort in the United States for bread or cereal products (Zheng et al., 1993) or in the ATBC study for either wheat or rye products (StolzenbergSolomon et al., 2002). No consistent observations are apparent for cereal and grain intake. The WCRF/AICR expert panel did not make a judgment on this food category, citing one, of many, complexities: that refined grains may differ in their effects from whole grains (WCRF, 1997).

Nutrients Epidemiologic evidence for the association between pancreatic cancer and nutrients has been reviewed (WCRF, 1997; Howe and Burch, 1996). Strong associations are lacking.

Carbohydrates and Glycemic Index/Load. All but one of the centers in a multicenter, population-based case-control study of pancreas cancer conducted in Canada, Europe, and Australia showed an increased risk of pancreatic cancer associated with a higher intake of carbohydrates, but not all of the positive associations were statistically significant (Howe et al., 1992; Howe and Burch, 1996). In the combined analysis, relative risks were estimated with highest vs. lowest quartiles of consumption and adjusted by the residuals method (Willett and Stampfer, 1986); increased risks were observed for higher intake of carbohydrates (OR = 1.74, 95% CI: 1.26–2.40) (Howe and Burch, 1996). In addition, a monotonically increasing significant trend in risk across the quartiles was apparent (P < 0.01). Further, the positive association with total carbohydrate intake (both simple and complex) largely explained the strong positive relationship between risk and total energy intake. Elevated relative risk estimates for pancreas cancer were also associated with the high compared to low category of carbohydrate consumption for men and women in two other studies (Lyon et al., 1993; Silverman et al., 1998). Analyses from other case-control studies reporting on carbohydrates found no significant associations (Howe and Burch, 1996; Durbec et al., 1983; Kalapothaki et al., 1993b; Ji et al., 1995). In contrast, in the ATBC cohort of male smokers, Stolzenberg-Solomon (2002) found a hazards ratio (HR) of less than one for the fifth relative to the first quintile of consumption of carbohydrates (HR = 0.62, 95% CI: 0.37–1.03), with a consistently decreasing risk across the quintiles (P (trend) = 0.02). Some types of carbohydrates increase glucose and insulin levels more than others (Augustin et al., 2001). Glycemic index and load are measures designed to take into account these differences. A high dietary glycemic index/load has been shown to be associated with an increased risk of several chronic diseases due to the adverse effects of high postprandial glucose levels and resulting increased insulin demands (Augustin et al., 2001; Michaud et al., 2002; Salmeron et al.,

740

PART IV: CANCER BY TISSUE OF ORIGIN

1997; Brand-Miller et al., 2003). High postload glucose levels have been found to be associated with the risk of pancreatic cancer in some (Gapstur et al., 2000; Levine et al., 1990), but not all (Smith et al., 1992) studies. Insulin has been shown to promote pancreatic cancer cell growth in vitro (Fisher et al., 1996). Therefore, conditions that cause excess insulin production may be associated with an increased risk. In one prospective study there was no association between dietary glycemic load/index with pancreatic cancer incidence (Johnson et al., 2005). In another, no overall association was found in subjects with high compared with low glycemic load/index diets (Michaud et al., 2002). However, in overweight and sedentary subjects, high dietary glycemic load/index was positively associated with risk. The association between pancreatic cancer and diets that cause postprandial hyperglycemia needs further exploration.

Fiber. A number of case-control studies have examined fiber intake. Four of the five centers in the SEARCH study found inverse associations for pancreas cancer and fiber intake and most were statistically significant. The pooled results, adjusted by the residuals method, showed an inverse association with a significant trend (P (trend) <0.01) and the OR for the highest compared with the lowest quartile was 0.42 (95% CI: 0.30–0.58) (Howe and Burch, 1996). Inverse associations for fiber have also been reported in other casecontrol studies as well (Kalapothaki et al., 1993b; Lyon et al., 1993; Ji et al., 1995b; Silverman et al., 1998). Lyon et al. (1993) found an inverse association with respect to fiber for women, but not men, while Silverman et al. found an inverse association (particularly from fruit or vegetable fiber) for men, but not women. Stolzenberg-Solomon (2002) reported results on fiber and fiber subtypes. No significant associations were found for total fiber, soluble fiber, or insoluble fiber. Dietary fiber may dilute or bind carcinogens in the digestive tract and thereby decrease their absorption (Steinmetz and Potter, 1996; Adlercreutz, 1990), but fiber may simply be correlated with foods (e.g., vegetables intake) that decrease risk by other mechanisms. It is unclear whether fiber decreases the risk for pancreatic cancer. Fat. Many studies that have examined total fat intake in relation to pancreatic cancer have reported null findings (Farrow and Davis, 1990b; Howe et al., 1992; Olsen et al., 1993; Kalaphothaki, 1993b; Silverman et al., 1998). Most of these studies found null results for subtypes of fat as well. Three case-control studies (Ghadirian et al., 1991b; Lyon et al., 1993; Durbec et al., 1993) and one cohort study (Stolzenberg-Solomon et al., 2002) found positive associations between fat and pancreatic cancer. A case-control study in Shanghai reported an unexpected inverse association (Ji et al., 1995b). Most case-control studies have not found an association between higher saturated fat intake and pancreas cancer, including those of Howe et al. (1992), Farrow and Davis (1990b), Kalaphothaki et al., (1993b), Ji et al. (1995b), and Silverman et al. (1998). Both Ghadirian et al. (1991b) and Olsen et al. (1993) found elevated relative risk estimates, although in the estimates in the latter study they were not statistically significant. One of three large cohort studies reported a positive association between pancreas cancer incidence and saturated fat (Stolzenberg-Solomon et al., 2002), while another reported no association for pancreas cancer and the following categories of fat: total, animal, vegetable saturated, trans-, polyunsaturated, and monounsaturated (Michaud et al., 2003). Overall, Nöthlings et al. (2005) found no associations of pancreas cancer with intake of total or saturated fat. Fatty acids, including oleic, linoleic, stearic, and linolenic acid, and omega-3 fish oils have been examined in five studies and generally there have been no significant associations (Farrow and Davis, 1990b; Olsen et al., 1993; Silverman et al., 1998; Stolzenberg-Solomon et al., 2002; Michaud et al., 2003). Dietary fat is strongly correlated with meat intake, and this could, in part, explain findings from some studies that implicate fat intake. Although several plausible mechanisms have been proposed and animal studies have supported a role for fat in pancreatic cancer (Birt et al., 1990; Roebuck, 1992; Woutersen et al., 1999), epidemiological studies have not provided consistent evidence that high fat intake is

related to risk. The WCRF/AICR report (1997) also concluded that there is insufficient evidence for an association between pancreatic cancer and total or saturated fat or other types of fat (i.e., animal fat, monosaturated and polyunsaturated fat, and omega-3 fatty acids).

Cholesterol. The Australian center of the SEARCH study reported an elevated relative risk estimate for the fourth to first quartiles of intake of cholesterol, OR = 5.09 (95% CI: 2.37–10.9) (Baghurst et al., 1991). Three other centers found ORs that were elevated, but not statistically significant: the highest was in the Polish center (Zatonski et al., 1991), with an OR of 3.26 (95% CI: 8.7–12.2) and the lowest was in the Toronto center with an OR of 1.07 (95% CI: 0.64–1.81) (Howe et al., 1990). The Montreal center reported an OR close to one (Ghadirian et al., 1991b). The combined analysis yielded a significantly elevated odds ratio (OR = 1.47, 95% CI: 1.10–1.96) for the highest vs. lowest quartile of dietary cholesterol (Howe and Burch, 1996). Three other case-control studies (Farrow and Davis, 1990b; Kalaphothaki et al., 1993b; Silverman et al., 1998) and three cohort studies (Stolzenberg-Solomon et al., 2002; Michaud et al., 2003; Nöthlings et al., 2005) found no statistically significant trends in disease risk associated with cholesterol intake. In the Nurse’s Health Study, the HR for the upper to lower quintile was 1.11 (95% CI: 0.67, 1.83). In the multiethnic cohort study, neither absolute nor energyadjusted cholesterol intake was statistically significantly related to risk of pancreatic cancer (Nöthlings et al., 2005). To date, evidence that diets high in cholesterol increase the risk of pancreatic cancer is not convincing. Protein. At least 8 case-control studies have examined protein intake and associations with pancreas cancer. Most studies have reported finding no association between risk and protein intake (Howe et al., 1992, Durbec et al., 1983; Kalapothaki et al., 1993b; Silverman et al., 1998). Two have found increased risks, (Farrow and Davis, 1990b; Lyon et al., 1993) though in the latter, the positive association was limited to men. Ji et al. (1995b) reported a significant inverse association for higher protein intake in both men and women. Analysis by type of protein (i.e., total, animal, vegetable) yielded null results in a large case-control study in the United States (Silverman et al., 1998) and in the ATBC cohort in Finland (Stolzenberg-Solomon et al., 2002). The findings for protein in another cohort were also null (Michaud et al., 2003). Most studies have not supported an increased risk of pancreatic cancer associated with high intake of protein. Micronutrients. In a case-control study nested in a cohort in Washington County, Maryland, Burney et al. (1989) compared prediagnostic serum micronutrient levels in individuals who subsequently developed cancer of the pancreas and their matched controls. The samples were assayed for several carotenoids, vitamin E, and selenium. The authors reported statistically significant lower levels of the carotenoid, lycopene (a major source is tomatoes), and selenium (men only), in cases than controls. In a combined analysis from the centers in the SEARCH study, results for beta-carotene were null (Howe and Burch, 1996). However, analysis of vitamin C showed a monotonically decreasing association from the lowest to highest intake quartile with an OR of 0.60 in the fourth quartile (95% CI: 0.43–0.83). Other case-control studies have reported on associations between pancreatic cancer and micro-nutrients. Falk et al. (1988) found a significant inverse trend for vitamin C in both men and women. Farrow and Davis (1990b) found null associations with intake of vitamins A or C; they also reported an inverse association between calcium intake and risk. Olsen et al. (1991) reported inverse trends in risk with increasing intake of vitamin C and beta-carotene. Positive trends were associated with consumption of retinol and riboflavin. No significant associations were found for calcium. Kalapothaki et al. (1993b) reported ORs of less than 1.0 for increasing intakes of both vitamins C and A, but these were not statistically significant; there were no notable findings for calcium or riboflavin. Ji and colleagues (1995b) reported inverse trends in risk with increasing intakes of carotene and retinol that were statistically significant (all P < 0.02) in both men and women in a case-control study

Cancer of the Pancreas in Shanghai. Odds ratios for the upper compared with the lowest quartile ranged from 0.38 to 0.61. There were also strong inverse associations between pancreas cancer and consumption of vitamins C and E, which were statistically significant in men, but not women. A casecontrol study conducted in northern Italy (Soler et al., 1998) found a significant increased risk of pancreas cancer in the third compared to first tertile for retinol (OR = 1.37, 95% CI: 1.01–1.85), but no association was found for beta-carotene. Silverman et al. (1998) reported results on vitamins and micronutrients, which included: vitamin C (total, from fruit, and from vegetable), vitamin A (total, from fruit, from vegetable, and from animal), lutein, xanthin, alpha-carotene, beta-carotene, cryptoxanthin, lycopene, provitamin A, retinol, B vitamins (folate, thiamine, riboflavin, niacin), and calcium, phosphorus, iron, sodium, and potassium. Increasing intake of vitamin C from vegetable sources was associated with a statistically significant inverse trend in risk in men (P (trend) = 0.008) and a nonsignificant decreased risk in women. Subjects in the highest compared with the lowest quartile experienced 40%–50% reductions in risk. When cruciferous vegetables were removed from the nutrient index for vitamin C from vegetables, there was no longer a vitamin C effect. In a cohort of older Californians (Shibata et al., 1994), a comparison of upper to lower tertiles yielded RRs below unity for dietary vitamin C (OR = 0.79, 95% CI: 0.44–1.43) and beta-carotene (OR = 0.78, 95% CI: 0.44–1.37). Micronutrients were analyzed in data from the cohort of male smokers (i.e., vitamin A, carotenoids, beta-carotene, lycopene, alpha-tocopherol, vitamin C, vitamin E, vitamin D, calcium, selenium, nitrite, nitrate, sodium, methionine, B vitamins (6 and 12), and folate) (Stolzenberg-Solomon et al., 2001; 2002). There were no significant associations for the highest compared with the lowest quintiles of intake except for dietary folate consumption (HR = 0.52, 95%CI: 0.31–0.87). In sum, data from nutritional epidemiologic studies suggest that fiber and vitamin C may lower the risk of pancreatic cancer. Data on some nutrients, such as folate and selenium, though intriguing, are too limited for conclusions. The data on fat are inconsistent and data on cholesterol and protein do not provide evidence for causal associations. Overall, the data on nutrients and pancreatic cancer do not yield greater insights than the conclusions we can draw from food analyses. Moreover, interpreting nutrient data in epidemiologic studies is problematic as discussed by Potter (2002). He argued that since nutrient levels are an abstraction—a score derived from data on food intake and a nutrient table—they do not provide enough data on real exposures. At most, they inform us about the foods that are the major contributors to the nutrient score. Analyses based on food items themselves may be a better approach, and we may need to find better ways to understand the relationship between eating patterns, more than nutrient intakes, to chronic diseases such as pancreas cancer. To move beyond the current knowledge regarding diet and pancreas cancer, future nutritional epidemiology studies will arguably require improved methods for dietary assessment, ideally with biomarkers, cohort studies in populations with better reliability and a wider range of dietary intake, and attention to potential effect modifiers, including genetic factors (Schatzkin and Kipnis, 2004).

Beverages Coffee. The hypothesis that coffee consumption might be related to pancreatic cancer was first suggested by correlations (Stocks, 1970) between cigarette and beverage consumption and cancer mortality. Stocks found an inconsistent correlation between coffee consumption and mortality from pancreas cancer, based largely on the coincidence of the two in Northern Europe. No correlation was found with use of cigarettes or tea. In 1981, a twofold to threefold increased risk was reported for coffee drinkers of three cups per day in a case-control study; the effect was apparent after control for smoking (MacMahon et al., 1981). The high correlation between coffee drinking and smoking in the United States made it difficult to rule out residual confounding by smoking. This report generated a great deal of publicity and controversy.

741

Concern developed regarding the inclusion of controls with dietaltering conditions such as gastrointestinal disease that may have ceased coffee drinking because of these conditions (Silverman et al., 1983). In 1990, Gordis reviewed 30 epidemiologic studies addressing the relation between coffee drinking and pancreatic cancer; with 14 of 17 case-control studies and 6 of 6 prospective studies showing no statistically significant increased risk, he concluded that the epidemiologic evidence did not support the hypothesis that coffee consumption increases the risk of pancreatic cancer. An IARC working group on the evaluation of carcinogenic risks to humans also assessed the epidemiologic evidence for a coffee drinking/pancreatic cancer association (IARC, 1991a). Evaluating 21 casecontrol studies and 6 cohort studies (a largely overlapping subset of that examined by Gordis et al. (1990)), the authors concluded that the data as a whole were suggestive of a weak relation between high levels of coffee consumption and pancreatic cancer, but that bias or confounding might account for the association. They also reported that, although the results with decaffeinated coffee were less comprehensive, they were generally null. For coffee consumption, most case-control studies subsequent to the IARC review (1991a) have found null or nonsignificant associations (Farrow and Davis, 1990a; Jain et al., 1991; Baghurst et al., 1991; Bueno de Mesquita, 1992a; Mizuno et al., 1992; Friedman and van den Eeden, 1993; Zatonski et al., 1993; Kalapothaki et al. 1993a; Partanen et al., 1995; Soler et al., 1998; Silverman et al., 1998). Null associations were reported in 5 cohort studies as well, including a study of elderly subjects in California (US)(Shibata et al., 1994); a cohort of 17,633 white males in the United States (Zheng et al., 1993); the American Cancer Society’s CPS II study (Couglin et al., 2000); the Nurses’ study (Michaud et al., 2001); and the ATBC cohort (Stolzenberg-Solomon et al., 2002). At least 5 studies have shown positive associations ((Hirayama, 1989; Lyon et al., 1992a; Gullo et al., 1995; Harnack et al., 1997; Lin et al., 2002b), and one reported a statistically nonsignificant inverse association with risk when comparing coffee drinkers with nondrinkers (Ghadirian et al., 1991a). The possibility that coffee is associated with an increased risk of pancreatic cancer seems unlikely. Reports of an increased risk associated with coffee drinking likely result from residual confounding from cigarette smoking, and possibly from other sources of confounding or bias.

Tea. The same IARC working group examined the evidence on the relation between tea consumption and pancreatic cancer; 3 of 4 cohort studies found no association and one reported a small inverse association. Only one of 6 case-control studies designed to evaluate the relation found an increased risk. The WCRF/AICR review (1997) summarized the data for tea consumption, stating that 4 of 6 cohort studies and 8 of 11 case-control studies showed no relationship between tea consumption and risk of pancreatic cancer. The remainder showed mixed results, most showing some inverse association, which could have arisen due to residual confounding by fruit and vegetable consumption or smoking. Ji et al. (1997) examined the association between green tea intake and risk of pancreatic cancer in a large population-based case-control study in Shanghai. They reported inverse associations with increasing intake of tea with significant trends in both men and women. There is some biologic plausibility for a protective effect of tea, since both black and green teas contain polyphenolic compounds, which have anticarcinogenic effects (e.g., they inhibit activation of the heterocyclic amine, PhIP, in a rodent model (Lin et al., 2003)). Alcohol. Dörken (1964) first suggested that alcohol abuse might increase the risk of pancreatic adenocarcinoma based on findings of a case-series. In an ecologic study, Hinds et al. (1980), found an association with beer consumption in Hawaii that is largely explainable by a single value. In several other ecologic studies, alcohol consumption was not correlated with pancreatic cancer incidence in the United States (Blot et al., 1978), Japan (Kono and Ikeda, 1979), or internationally (Sarles, 1974).

742

PART IV: CANCER BY TISSUE OF ORIGIN

Among the first analytic studies, case-control studies using hospital controls showed positive associations (Burch and Ansari, 1968; Ishii et al., 1968; Lin and Kessler, 1981), but the lack of detail about control selection in each instance reduces confidence in the findings. One similarly designed study was null (Wynder, 1973a). A casecontrol study using consenting cases from the Third National Cancer Survey, and other cancer patients as a comparison, found a small excess risk of pancreas cancer in male drinkers, but the effect disappeared after controlling for smoking (Williams and Horm, 1977). Durbec et al. (1983) reported an increased risk associated with beer and wine, but not spirits. In the first set of cohort studies, results were generally null, although in some studies the results were inconclusive because of small numbers. Of 4370 Finnish alcoholics observed for less than 4 years (Hakulinen et al., 1974), 4 cases of pancreatic cancer were observed compared with 2 expected. A proportional mortality analysis of the 909 deaths identified among 1382 known alcoholics revealed 3 pancreatic cancer deaths compared with 5.2 expected (Monson and Lyon, 1975). Klatsky et al. (1981) found 6 deaths among individuals with reported consumption of 6 or more alcoholic drinks per day compared with 2 among matched non-drinkers in a follow-up study of members of a health plan (RR = 3.0, 95% CI: 1.1–6.5). In a cohort study of nearly 17,000 Norwegians, Heuch et al. (1983) found a relative risk of 2.7 (a 95% CI: 1.2–6.4 was calculated by Velema et al. (1986)) among those who drank at least 14 times per month compared with those who did not drink or who drank infrequently. In a cohort study of over 14,000 Danish brewery workers (Jensen, 1979), the observed number of cases did not differ significantly from the number expected. In 1986, Velema et al. (1986) reviewed the literature and argued that evidence for a causal association was weak, though chronic high intake may increase the risk, and that neither total alcohol consumption nor any specific type of alcohol showed a strong and consistent relation. Most subsequent case-control studies have produced findings that are in line with these conclusions. Using a pooled analysis of three case-control studies from Italy, France, and Switzerland (La Vecchia et al., 1987; Raymond et al., 1987; Clavel et al., 1989), Bouchardy et al. (1990) reported on 494 cases and 1704 controls. The authors found no consistent association with consumption of wine, beer, or spirits, nor any evidence of a dose-response. Other case-control studies with generally null findings for alcohol include those of Mack et al. (1986), Norell et al. (1986a), Falk et al. (1988), Farrow and Davis (1990a), Jain et al. (1991), Ghadirian et al. (1991a), Bueno de Mesquita et al. (1992a), Lyon et al. (1992b), Mizuno et al. (1992), Friedman and van den Eeden (1993), Kalapothaki et al. (1993a), Tavani et al. (1997), and Soler et al. (1998). The possible association with heavy consumption of alcohol has been supported in some studies, however (Cuzick and Babiker, 1989; Olsen et al., 1989; Silverman et al., 1995). Recent prospective studies have examined alcohol intake and pancreas cancer and have reported associations that were null (Shibata et al., 1994; Coughlin et al., 2000; Stolzenberg-Solomon et al., 2001; Lin et al., 2002b), inconsistent (Hirayama, 1989), and positive (Zheng et al., 1993; Harnack et al., 1997). The epidemiologic evidence for a role of alcohol in the genesis of pancreatic cancer is weak. If alcohol has any role in the etiology of pancreatic cancer, it is likely to be among heavy drinkers. Together with a high-fat diet, alcohol is an important determinant of chronic calcifying pancreatitis (Johnson and Zintel, 1963; Ishii et al., 1973; Sarles and Tiscornia, 1974; Dufour, 2003), the form of pancreatitis most strongly suspected of constituting a risk factor for pancreatic carcinoma (Ishii et al., 1973). The mechanism whereby heavy alcohol increases the risk for pancreatic cancer, should this prove not to be a spurious association, could be via pancreatitis. However, in several studies where pancreatitis was associated with an increased risk for pancreatic cancer, the increase was not specific to alcohol-induced pancreatitis (Amman et al., 1984; Amman and Schueler, 1984; Lowenfels, 1984; Lowenfels et al., 1985; 1993). In some animal models, alcohol has been shown to promote carcinogenesis (Kuratsune et al.,

1971), but other studies do not support a role for alcohol (Pour et al., 1983).

Aspirin and Other Non-Steroidal Anti-Inflammatory Drugs Non-steroidal anti-inflammatory drugs (NSAIDs) are widely used for pain relief. Epidemiologic and laboratory studies indicate that these drugs are chemopreventive for colon and possibly other cancers (Baron 2003; IARC, 1997). It is plausible that aspirin and other NSAIDs could inhibit pancreatic carcinogenesis; experiments in animal models and cell culture provide evidence to support this claim (Molina et al., 1999; Perugini et al. 2000; Takahashi et al., 1990). The epidemiologic data, however, are inconsistent. In experimental studies, NSAIDs have been shown to induce apoptosis, block angiogenesis, and decrease carcinogen activation via cyclooxygenase (Rao and Reddy, 2004; Chu et al., 2003). In studies of hamsters treated with N-nitrosobis (2-oxopropyl) amine (Takahashi et al., 1990), NSAIDs reduced the number of animals with pancreatic tumors and the number of tumors per animal. In addition, NSAIDs have been shown to inhibit growth of human pancreatic cancer cell lines (Molina et al., 1999; Perugini et al., 2000; Kokawa et al., 2001). The chemopreventive potential of these drugs is thought to be, in part, through the inhibition of cyclooxygenase 1 and 2 (COX1 and 2), the rate-limiting enzymes in the conversion of arachidonic acid to prostaglandins (IARC, 1997; Kokawa et al., 2001). Prostaglandins are important mediators of signal transduction pathways that modify cellular adhesion, growth, and differentiation. Elevated prostaglandin synthesis may promote carcinogenesis through mechanisms that include direct mutagenesis, increased cell proliferation, immune suppression, tumor promotion, and facilitated metastasis (Molina et al., 1999; Perugini et al., 2000). The anti-proliferative effects in cell culture indicate the chemopreventive potential of NSAIDs is not limited to COX 2 inhibition and that these agents vary in their relative potency for particular targets (IARC, 1997; Molina et al., 1999; Koshiba et al., 1999; Perugini et al., 2000; Soh and Weinstein, 2003; Bottone et al., 2003; 2004). Epidemiologists cannot assume that all NSAIDs are equivalent with respect to chemopreventive potential in humans; nor that the timing of any inhibitory effects on the carcinogenic process is similar from one cancer to another. Seven cohort studies (Schreinemachers and Everson, 1994; Gridley et al., 1993; Anderson et al., 2002b; Sorensen et al., 2003; Friis et al., 2003; Schernhammer et al., 2004; Jacobs et al., 2004) and three casecontrol studies (Coogan et al., 2000; Langman et al., 2000; Menezes et al., 2002) have addressed the topic of NSAID use and pancreatic cancer. Schreinmachers and Everson (1994) found an inverse association (RR = 0.67) between reported use of aspirin in the past 30 days and pancreatic cancer incidence in the National Health and Nutrition Examination Study I cohort. However, the results were based on only 30 cases and the confidence interval included 1.0. A large cohort study of rheumatoid arthritis patients in Sweden reported findings based on 32 pancreatic cancer cases (Gridley et al., 1993). The use of NSAIDs was inferred by the condition, but not directly assessed. The authors reported standardized incidence ratios, not adjusted for other risk factors for pancreatic cancer, of 1.12 for men, 0.68 for women, and 0.83 for both sexes. In a prospective study of 28,283 postmenopausal women in the Iowa Women’s Health Study, any current use of aspirin vs. no use was associated with a lower risk of pancreatic cancer (n = 80 cases); the RR was 0.57 (95% CI: 0.36–0.90). In addition, there was an inverse trend of decreasing risk with increasing intake of aspirin that was statistically significant (P (trend) = 0.005); daily users had one-third of the risk of non-users (Anderson et al., 2002b). The study was not able to assess potential effects of duration of aspirin use. No association was observed with reported use of non-aspirin NSAIDs nor did adjustment for their use alter the findings for aspirin. In the Nurses Health Study, a large US cohort study, aspirin use was assessed among 88,378 women at baseline in 1980 and biennially thereafter except for 1986 (Schernhammer et al., 2004). During 18

Cancer of the Pancreas years of follow-up, there were 161 cases of pancreas cancer. An increasing risk was observed with increasing duration of aspirin use. In women reporting regular aspirin use (2 or more tablets per week) with long duration (20 years or more), the multivariate adjusted RR was 1.58 (95% CI: 1.03–2.43). There was no association for current use of non-aspirin NSAIDs with pancreas cancer risk (RR = 1.20, 95% CI: 0.79–1.80), but duration of use was not available. The authors reported that findings for aspirin were essentially unchanged when analyses were restricted to women who reported no use of other NSAIDS. Friis et al. (2003) linked a prescription database in Jutland, Denmark, with the Danish Cancer Registry. Cancer incidence was calculated over a 9-year follow-up period in a cohort of 29,470 adults who were prescribed low-dose aspirin between 1989 and 1995. The observed incidence rates, based on 62 cases, were compared with those expected if country-specific rates applied, and yielded a standardized incidence ratio (SIR) of 1.1 (95% CI: 0.8–1.3). The study did not assess higher doses or non-aspirin NSAIDs, but in a related analysis using the same prescription database (Sorensen et al., 2003), the findings were null (SIR = 1.1, 95% CI: 0.9–1.2) when incidence of pancreas cancer in 172,057 adults prescribed any non-aspirin NSAIDs (from 1989 through 1995) was compared with expected incidence based on country-specific rates. The analysis included 149 cases in a 9-year follow-up period. There was also no evidence of a difference in risk among individuals who had received 10 or more prescriptions for non-aspirin NSAIDs (SIR = 0.9, 95% CI: 0.6–1.3). In the largest prospective study to date, no association was found between aspirin use and pancreatic cancer mortality when frequency or duration of use was examined. The cohort of 987,590 adult volunteers (the Cancer Prevention Study II) was recruited by the American Cancer Society (Jacobs et al., 2004). There were nearly 4600 deaths from pancreas cancer over the 18-year follow-up. Those who used aspirin 30 or more times per month for 20 or more years experienced no reduction in risk compared with nonusers (RR = 0.96, 95% CI: 0.69–1.33). Non-aspirin NSAID use was not collected, but a subset of the cohort were resurveyed for this and other information (n = 156,895) and no effect was seen in the smaller cohort. In a case-control study of medication use and cancer risk, Coogan et al. (2000) analyzed data on 491 pancreatic cases and 5833 controls. The authors reported a non-significant inverse association between pancreas cancer and self-reported use of NSAIDs (mostly aspirin) (OR = 0.8, 95% CI: 0.5–1.1). Non-aspirin NSAIDs were not analyzed separately. In the United Kingdom, Langman et al. (2000) conducted a casecontrol study of various cancers, including pancreatic cancer, in relation to the number of NSAID prescriptions 13–36 months prior to diagnosis. The data were adjusted for age and smoking status (ever/never). Patients who received at least seven prescriptions for NSAIDs in the 13–36 months prior to diagnosis had an odds ratio of 1.49 (95% CI: 1.02–2.18), which could reflect treatment of symptoms prior to diagnosis among pancreatic cancer cases (Moertel et al., 1971; Ebbehøj et al., 1985). Non-prescription NSAID use was not assessed in that study, nor was aspirin use considered separately. A hospital-based case-control study in Buffalo, New York (194 cases and 582 controls) found no statistically significant associations for either regular aspirin use (defined as at least one aspirin per week for at least 6 months) or duration of regular use for 11 or more years, compared with no use. The authors did not report on, or adjust for, use of other NSAIDS (Menzes et al., 2002). Accurate assessment of medication use—even regular medication use—is difficult to achieve in epidemiologic studies. All of the studies to date have limitations. For example, not all could adjust for potential confounding factors. In many studies, aspirin use was not collected separately from other non-aspirin NSAIDs, which is problematic, as they are not necessarily equivalent in their effects. Case-control studies could be biased by the health of the respondents either because pre-clinical cases might decrease NSAID use because of nausea or increase use for pain relief (Moertel et al., 1971; Ebbehøj et al., 1985). Duration of NSAID use was not collected in all studies and without this measure or multiple assessment of use (cohort studies) there is the

743

potential for considerable misclassification. Dose of NSAIDs was also not collected in a number of studies. Studies that link prescription databases with cancer registries may also suffer from misclassification where they lack information on over-the-counter use. Frequency of current use may not reflect long-term use; a substantial proportion of the population uses these drugs only sporadically. Even regular users may change their patterns of use as they age (e.g., they may increase use with development of arthritis or decrease use due to stomach irritation); older individuals may start using low-dose aspirin to prevent heart attack and stroke. Surrogate measures, such as arthritis, may be better measures of long-term NSAID use than self-reported data (Gridley et al., 1993). However, they do not allow for a separate analysis of aspirin vs. other NSAIDs. Although the general population is reasonably familiar with aspirin, they may not be aware that there are a number of combination medications that contain aspirin, and therefore participants may not accurately recall their use. Non-aspirin NSAIDs present additional challenges. This broad category includes many drugs; new agents are being introduced frequently, and the same drug may go by more than one name. The availability of particular agents—by prescription vs. over the counter—varies by time and country. Because of these issues, it may be difficult for subjects to accurately recall their use of these drugs; even if they can accurately recall their use, the category of non-aspirin NSAIDs is not necessarily equivalent from one study to another. Such differences may, in part, explain inconsistencies between studies (and misclassification within studies) of pancreas cancer and NSAID use. In short, it is difficult to evaluate aspirin or NSAIDs as risk factors for pancreas cancer incidence. The ideal study design to sort out some of the issues described above would be a clinical trial, but the relatively low incidence of this cancer would render such a trial very expensive. Further observational studies of pancreatic cancer in relation to aspirin and NSAID use that include careful assessment of the dose, duration, and type of drugs being used, should be considered.

Occupation and Industry By virtue of its obscure etiology and the clear role of smoking (i.e., exposure to the products of incomplete combustion), pancreas cancer has been extensively examined in relation to occupation and workplace exposures. Descriptive epidemiology does not offer support for the importance of occupational causes. In Northern Europe, where access to care is universal, and diagnostic treatment facilities are uniform and of good quality, age-adjusted rates of pancreas cancer in the larger industrialized countries are no higher than those of relatively agrarian Iceland (Parkin et al., 2002) or the Faroe Islands (Jacobsen et al., 1985), moderating suspicions of important workplace carcinogens. Moreover, there is currently little difference between rates in rural (Iowa) and in industrialized urban (Connecticut, Los Angeles) environments in the United States. Neither Los Angeles, California (Mack, 2004) nor Olmsted County, Minnesota (Riela et al., 1992) has demonstrated an increase in the rate of incidence over time. Finally, the pattern of occurrence of pancreas cancer in relation to social class in both wealthy and poor countries fails to suggest substantially elevated risk among working-class men. Employment is not random, and workers differ from non-workers in many ways, including health status, social class, past workplace exposures, and smoking history, collectively raising concerns about the healthy worker bias as well as important sources of confounding. Because pancreas cancer is not common, no single occupation or industry accounts for more than a small proportion of cases in a population, making the power of studies to test a single occupational hypothesis low. It is inevitable that many associations between single occupations and single outcomes will be observed, and that many of these will be the result of chance.

Study Designs In the absence of any highly credible specific hypothesis, multiple hypotheses are often tested at the same time. The most common and

744

PART IV: CANCER BY TISSUE OF ORIGIN

extreme method of screening for evidence of a link between pancreas cancer and the workplace has been to examine the pattern of occupations and conditions listed on all death certificates or cancer reports collected for a defined population over a defined period. Such matrices have been developed for the United States (Guralnick, 1963; Williams et al., 1977), the United Kingdom (Logan, 1982; OPCS, 1978), Denmark (Lynge and Thygesen, 1990; Olsen and Jensen, 1987), Sweden (Alguacil et al., 2003), Finland (Partanen et al., 1994), Norway (Anonymous, 1976), Spain (Alguacil et al., 2000a), the states of Washington (Milham, 1976), California (Peterson and Milham, 1980), and Illinois (Mallin et al., 1989), an aggregate of 24 US states (Kernan et al., 1999), Quebec (Siemiatycki, 1991); three English counties (Magnani et al., 1987), and the cities of New Orleans (Pickle and Gottlieb, 1980), Los Angeles (Mack and Paganini-Hill, 1981), and Tokyo (Ishii et al., 1968). Similar analyses have been based on hospital patients from New York City (Cubilla and Fitzgerald, 1978), Buffalo (Bross et al., 1978; DeCouflé and Stanislawczyk, 1977), and Rochester, Minnesota (Maruchi et al., 1979), and from representative groups defined in other ways from Canada (Best, 1966) and Japan (Hirayama, 1975). The liabilities of these studies are several. Occupational classification is necessarily crude, usually based on death certificate information; any hypothesis so derived must be based on considerable speculation about the nature and intensity of actual exposure. This is even more likely when risks are assessed according to industry, where multiple workplaces are usually involved. Concern must also be generated about the validity of the pancreas cancer diagnoses, because they usually are derived from secondary sources. Perhaps most importantly, the job-exposure cells in such a matrix are inevitably numerous and many apparent increases in risk can be expected to appear by chance. Commonly, 50 or more occupational categories will be selected, and even if interest is restricted to neoplasms, 40 or more different outcomes will be examined. In such a circumstance, positive findings are difficult to interpret. Consider a recent comparatively well-conducted study, based on the Swedish population (Alguacil et al., 2003). Nearly 32 million male person-years were screened (among whom about 112,000 cancers would have occurred) and 4420 pancreatic cancers were identified; 1322 potentially informative associations (increased risk based on at least 5 cases) were distributed among well over 60 occupational categories. An increase in risk of greater than 50% was seen in 12 categories (with an average number of 6 excess cases), and five of these increases were incompatible with chance at the 95% level. Four of the five (technical assistant, travel agent, freight handler, and waiter) had low biologic credibility. The final occupation (metal workers) showed a significant increase, but based on 7 expected and 13 observed cases. A difference as great could be expected by chance (according to the Poisson distribution) once in every 67 occupational groups of comparable size. Somewhat more informative are studies in which the occupations screened have been categorized a priori according to those specific workplace exposures usually experienced by persons in that occupation (job-exposure matrices). Such screening has been carried out for workers in Finland (Kauppinen et al., 1995; Weiderpass et al., 2003), Great Britain (Pannett et al., 1985), Spain (Alguacil et al., 2000a), and the United States (Kernan et al., 1999). Because jobs usually involve multiple exposures, and because the exposures of a workplace are often highly correlated, the benefits of such additional exposure classification for the study of cancer have been modest. Also more informative are the studies designed to test multiple hypotheses that focus either on the diverse outcomes occurring as a result of a specific workplace exposure (cohort studies defined by employer, occupation, or union membership) or on the diverse causes of pancreas cancer (case-control studies). Investigators adopting the occupational cohort approach are usually in a position to characterize the exposure with more precision, but with some exceptions (Garabrant et al., 1992; Hanis et al., 1982), usually cannot adjust on an individual basis for confounding factors such as smoking. Investigations that employ case-control methodology can make such adjustments for confounding, but usually must accept the subject’s

recollection of the workplace exposure. Population-based, casecontrol studies do have the advantage of describing the impact of occupational exposure upon the community (Falk et al., 1988; Ji et al., 1999; Lin and Kessler, 1981; Mack et al., 1985; Norell et al., 1986b; Pietri et al., 1990). However, only when case-control studies are nested within well-characterized occupational cohorts (Bardin et al., 1976; Mikoczy et al., 1996; Li and Yu, 2002) can both accurate descriptions of the workplace exposure and adequate control of confounding be accomplished. Unfortunately, the numbers of available subjects for such ideal studies are generally small. Finally, there have been several attempts to systematically examine a collective experience using a process termed meta-analysis, but in actuality only superficially based on a commonality of definitions or referent group. Eligibility for such exercises can be based on the exposure (Calvert et al., 1998; Rachet et al., 2000), on pancreas cancer as an outcome (Ojajarvi et al., 2000), or on both outcome and specific exposure (Collins et al., 2001; Ojajarvi et al., 2001). When observed, the elevations in risk are generally small (as is the usual number of subjects in each comparison) and the confidence intervals are usually large. We have therefore elected to itemize them on a dichotomous (increased risk vs. no increased risk) basis rather than by a detailed citation of specific numerical risk ratios.

Sedentary Occupations Among those occupations commonly found to convey a low but statistically significant elevation in risk of pancreas cancer, are white collar workers such as managers, salespersons, clerks, civil servants, and travel agents (Lin and Kessler, 1981; Falk et al., 1990; Siematycki, 1991; Kernan et al., 1999; Aljuacil et al., 2000a; Weiderpass et al., 2003). Such positive associations sometimes stand out because the number of individuals in those categories is large, making statistical significance more attainable. Moreover, the links between pancreas cancer and sedentary occupations necessarily have been based in part on cases diagnosed years ago, leading to speculation that middle-class persons may have had access to more exhaustive methods of diagnosis in the years before universal availability of scanning technology. Parenthetically, sedentary occupations repeatedly have been linked to colon cancer, with a mechanism that is obscure. It is possible that pancreas cancer could result from the same unknown mechanism. A few of the studies suggesting that exercise conveys protection to the colon also suggest that it grants some protection to the pancreas (Paffenbarger et al., 1987; Weiderpass et al., 2003), but opportunities for such parallel scrutiny have been few.

Ionizing Radiation Initial analyses of the occurrence of pancreatic cancer among British radiologists indicated a possible excess, but the study of US radiologists (Matanoski et al., 1975) and a more complete follow-up of the British cohort (Smith and Doll, 1981) have suggested otherwise. Similarly, when atomic workers at the Hanford plant in Washington State were initially examined (Mancuso et al., 1977) and re-examined (Gilbert and Marks, 1979; Hutchison et al., 1979), it was concluded that an excess risk had occurred in the highest of six radiation dose categories, but subsequent follow-up failed to confirm the excess (Tolley et al., 1983). Workers exposed to thorium processing were also found to have increased risk (Polednak et al., 1983), as were Finnish workers with exposure to medical and other sources of radiation as entered into a job-exposure matrix (Kauppinen et al., 1995). However, no clear evidence of a link between pancreas cancer and occupational exposure to radiation was identified after a very intensive review (Committee Beoir, 1990).

Asbestos One observation suggestive of an increased risk of pancreas cancer among workers exposed to crocidolite has been reported (Newhouse et al., 1988). In addition, more cases than controls reported past asbestos exposure in one case-control study (Falk et al., 1990), and high risk has been reported among auto mechanics who also experienced exposure to asbestos (Hansen, 1989). A recent report that Japanese shipboard asbestos workers exposed to chrysotile and

Cancer of the Pancreas amosite experienced an excess of pancreas cancer was based on two cases (Kuromatani et al., 1999). In other cohorts of asbestos workers, no evidence of an excess of pancreas cancer has been found (Enterline et al., 1987; Puntoni et al., 1979; Selikoff and Seidman, 1981), and a meta-analysis of all such studies failed to identify a significant excess risk (Ojajarvi et al., 2000).

Metals Metal workers have been repeatedly designated a group at high risk, but the presumption is that the most pertinent exposure is to the polycyclic aromatic hydrocarbons (PAHs) and nitrosamines in the oils and fluids used for cooling and lubrication (see below) rather than the metal dust itself. Concern was raised about metal exposure after a comprehensive meta-analysis of occupational exposures and pancreas cancer (Ojajarvi et al., 2000) reported elevated risks following exposure to nickel and nickel compounds (meta-risk ratio 1.9, 95% CI: 1.2–3.2) and to chromium and chromium compounds (1.4; 95% CI: 0.9–2.3). It was subsequently noted (Seilkop, 2001) however that two studies showing no such link to nickel (Arena et al., 1998; Shannon et al., 1991) had been omitted from the meta-analysis. An association between chromium exposure and pancreas cancer has been reported among Finnish female workers (Weiderpass et al., 2003). Cadmium has been postulated to be a cause of pancreas cancer (Schwartz and Reis, 2000), although no support for links to cadmium (and none for links to iron or lead) emerged from the above-mentioned metaanalysis (Ojajarvi et al., 2000).

Polycyclic Aromatic Hydrocarbons and Nitrosamines Knowledge of the link between cigarette smoking and pancreas cancer suggests the possibility that other organic materials, especially when subjected to incomplete combustion might also cause pancreas cancer. Population-based screening of mortality in relation to stated occupation identified apparent risks in association with fossil fuel production (Falk et al., 1990; Hanis et al., 1982; Milham, 1976; Pickle and Gottlieb, 1980; Redmond et al., 1976; Siemiatycki, 1991; Thomas et al., 1980; Thomas et al., 1982), vehicle manufacturing and repair (Eisen et al., 1992; Hansen, 1989; Peterson and Milham, 1980; Vena et al., 1985; Viadana et al., 1976), and with those forms of metalwork requiring the use of cooling and lubricating mineral oils or ethanolamine-based liquids (Alguacil et al., 2003; Guralnick, 1963; Mack and Paganini-Hill, 1981; Mallin et al., 1989; Silverstein et al., 1988; Tolbert et al., 1992). Some case-control studies, based on small numbers, have produced associations consistent with these findings (Bardin et al., 1976; Falk et al., 1988; Ji et al., 1999; Maruchi et al., 1979; Norell et al., 1986b). Mineral oils used in the manufacture of transformers have also been associated with increased risk (Yassi et al., 2003). A systematic review of risk following occupational exposure to metalworking fluids (Calvert et al., 1998) concluded that despite some inconsistencies within and between studies, metalworking fluids containing complex mixtures of PAHs as well as nitrosamines may increase risk of pancreas cancer, especially synthetic fluids used in grinding and straight oils used in machining. Consistent with these results are the observed increases in risk that have been recorded among aluminum mill and reduction plant workers (Milham, 1976; Rockette and Arena, 1983; Ronneberg et al., 1999), printers and lithographers (Mallin et al., 1989; Williams et al., 1977), and vehicle drivers (Mallin et al., 1989; Partanen et al., 1994; Viadana et al., 1976). However, some studies have produced inconsistent finding in relation to race and age. Some have found the link between PAH exposure and pancreas cancer to be stronger among African Americans (Eisen et al., 1992; Vena et al., 1985). Press workers have also been studied with inconsistent results; a minor increase in risk has been found among commercial pressmen (Zoloth et al., 1986), but among a group of printing pressmen (Lloyd et al., 1977), the increase was found only in the youngest age group. Early studies of rubber workers produced evidence of both decreased (McMichael et al., 1974) and increased (Monson and Fine, 1978) risk. More recent results from the latter study (Delzell et al., 1981; Delzell and Monson, 1985) show a

745

low-level elevation in risk, but still based on a small number of observed cases. Similar increases in pancreas cancer among workers exposed to the curing of rubber have been reported from China (Li and Yu, 2002) and Russia (Solenova, 1992). The particular agents that might be responsible are not known, but solvent-based lubricants or nitrosamines may be involved. Moreover, no excess risk was found among British gas workers (Doll et al., 1972), and observations of refinery workers have shown generally minor elevations. With detailed analysis, the subgroups with higher risk have no special exposures to petroleum products, and no relation was seen between length of exposure and risk (Divine and Barron, 1986; Divine et al., 1985; Kaplan, 1986; Nelson et al., 1987; Rushton and Alderson, 1981; Schottenfeld et al., 1981; Theriault and Provencher, 1987; Wen et al., 1983; Wong et al., 1986). No risk was found among those members of the American Cancer Society cohort who had been exposed to diesel exhaust (Boffetta et al., 1988), nor was any observed among a large cohort of stationary engineers and firemen (DeCouflé et al., 1977), a group that has been singled out several times in screening studies. A population-based case-control study based on a job-exposure matrix in Finland resulted in a non-significant modest elevation in risk following exposure to PAHs and other products of incomplete combustion (Kauppinen et al., 1995), as did the comprehensive meta-analysis of pancreas cancer in relation to occupational exposures (Ojajarvi et al., 2000).

Chlorinated Hydrocarbons Prominent among the findings of the early screening studies are increased risks for those working in occupations requiring contact with halogenated hydrocarbons, including airplane mechanics (Peterson and Milham, 1980), dry cleaners (Lin and Kessler, 1981), and users of cleaning agents (Norell et al., 1986b). In addition, a nearly twofold risk increase was reported among those exposed to methylene chloride during the manufacture of photographic film (Hearne et al., 1987), and an even larger risk increase among workers involved in the production of chlorohydrin (Greenberg et al., 1990). Subsequently, additional follow-up of workers engaged in chlorohydrin production (Benson and Teta, 1993) has also identified a substantial increase in risk of pancreatic cancer (8 observed cases), with the conclusion that the most likely causative agent was ethylene dichloride. An excess number of cases (1 expected, 8 observed) was also reported from a plant processing polyvinyl chloride in the absence of exposure to vinyl chloride monomer (Selenskas et al., 1995). In the recent meta-analysis of occupational exposures and pancreatic cancer (Ojajarvi et al., 2000), a significant meta-risk ratio of 1.4 for exposure to chlorinated hydrocarbons was identified. A year later the same authors, by covering all pertinent publications from Europe, North America, and Asia during the period 1969–1998, implemented a more comprehensive and detailed meta-analysis of chlorinated hydrocarbons (Ojajarvi et al., 2001), and found significant increases in relative risk (of 2.0 or lower) for those engaged in metal degreasing and dry cleaning. With respect to specific agents, suggestive increases were seen for trichloroethlene, polychlorinated biphenyls, methylene chloride, vinyl chloride, and tetrachloroethylene, but not carbon tetrachloride. No exposure-response relationships could be documented, and confounding by other workplace and exposures, smoking could not be excluded. In addition, when a case-control study was initiated because high pancreas cancer mortality appeared among chemical production workers, a substantial excess risk (OR = 4.8) was observed among those ever exposed to DDT, with a higher risk (OR = 7.4) associated with those exposed, on average, for 4 years (Garabrant et al., 1992). Furthermore, DDT derivatives were independently linked to risk, and adjustment for individual exposure to other risk factors failed to alter the findings. No similar excess was found among workers exposed to DDT production elsewhere (Wong et al., 1984), although excess risk was observed among those applying DDT in Michigan (not significant) by Fryzek et al. (1997) and Australia (significant) by (Beard et al., 2003). The recent meta-analysis of occupational exposures and pancreas cancer found a significant elevated risk attributable to organochlorine insecticide exposure (Ojajarvi et al., 2000), and the

746

PART IV: CANCER BY TISSUE OF ORIGIN

pattern of pancreas cancer in rural California corresponds to the pattern of heavy organochlorine pesticide usage (Clary and Ritz, 2003). Anecdotally, a worker long engaged in the production of chlorinated insecticides and his wife developed pancreas cancer after residing many years on the grounds of the insecticide factory (Rubio and Rodriguez, 1989). As production of DDT in the United States diminished (around 1970), most of that produced was used in cotton fields (IARC, 1991b). Although no excess risk of pancreas cancer was associated with farming as an occupation in Louisiana (Falk et al., 1990), an excess risk was seen in association with cotton dust exposure.

Other Pesticides In a nested case-control study of 450 cases, performed in a prepaid care setting in California (Friedman and Van den Enden, 1993), an elevation in risk associated with exposure to insect or plant sprays was found. A multicenter US case-control study focusing on pesticide exposure found modest elevations in risk among those occupationally exposed to moderate/high levels of fungicides and herbicides (Ji et al., 2001). Finnish gardeners were noted to be at elevated risk (Partanen et al., 1994), although no such increase was observed among farmers in Iowa (Burmeister, 1990), Wisconsin (Saftlas et al., 1987), or farmers worldwide (Blair et al., 1985; Franceschi et al., 1993). Workers in companies engaged in the manufacture of phenoxyacid herbicides (Coggon et al., 1986) and brominated chemical pesticides, including DCBP, TRIS, and brominated organic and inorganic compounds (Wong et al., 1984) have been followed without finding any excess risk of pancreas cancer. In Seveso, Italy, those heavily exposed in 1976 to spilled dioxin, which is closely related to phenoxyacid herbicides, have been followed to 2001, and no excess of pancreas cancer has been observed (Bertazzi et al., 2001). No increased risk or doseresponse relation was found among Swedish agricultural workers (Wiklund and Holm, 1986) or pesticide applicators and flight instructors (Wiklund et al., 1989), although a few excess cases have been reported among United States (Cantor and Booze, 1991) and Spanish (Alguacil et al., 2000a) pesticide applicators.

Other Chemicals Early studies of chemists (Li et al., 1969) and chemical workers (Mancuso and El-Attar, 1967) found an excess of pancreas cancer, although based on small numbers. Subsequent similar studies have been inconsistent: finding such a risk (Bond et al., 1985) or failing to do so (Hoar and Pell, 1981). A study of British chemists (Searle et al., 1981) also found no evidence of excess risk. Aromatic and aliphatic (non-chlorinated) solvents were found to be linked to pancreatic cancer in a population-based case-control study based on a job-exposure matrix (Kauppinen et al., 1995) as well as in the recent meta-analysis of pancreas cancer and occupation (Ojajarvi et al., 2000), although one recent study based on exposure assessment by industrial hygienists failed to confirm the association (Alguacil et al., 2000a). Minor increases in risk have been attributed to occupations with exposure to formaldehyde (Kernan et al., 1999), but when examined in more detail (Collins et al., 2001), those allegedly at high risk did not include those more heavily exposed in the process of formaldehyde manufacturing. Long-term follow-up of workers exposed to acrylamide in four plants revealed an apparent twofold increase in risk, but without any exposure-response confirmation (Marsh et al., 1999). Exposure to dyes and pigments has also been implicated in screening studies (Alguacil et al., 2000b; Mack et al., 1985).

Miscellaneous Industries Electric. Increases in risk have been reported among electrical workers exposed to electromagnetic fields in Finland (Weiderpass et al., 2003), as well as among workers in the electric industry in Sweden (Alguacil et al., 2003) and China (Ji et al., 1999). In two of these studies, electromagnetic field (EMF) exposure was considered a possible explanation.

Leather. Cohort studies of workers in the tanning industries of Tuscany (Costantini et al., 1989) and Sweden (Edling et al., 1986)

revealed increases in risk of pancreas cancer, and a study of chromate pigment workers (chrome exposure occurs in the tanning process) found a few excess cases (Sheffet et al., 1982). More recently, a casecontrol study in Sweden (Mikoczy et al., 1996) attributed an increased risk to leather dust. However, population-based screening of industrial categories has not consistently identified leather manufacturing as a high-risk industry.

Pulp and Paper. In plants employing the sulfite process, increases in pancreas cancer have been reported from Finland (Jappinen et al., 1987), Sweden (Wingren et al., 1991), Denmark (Rix et al., 1997), and the United States (Henneberger et al., 1989; Pickle and Gottlieb, 1980; Robinson et al., 1986). No specific exposure has been cited. Food and Drink. Increases in risk were reported in English butchers and fishmongers (Magnani et al., 1987) and in Swedish brewery workers (Carstensen et al., 1990). Biological Research. In a review of seven reports of cancer experience among workers in research laboratories, increased risk from pancreas cancer was noted in two (Rachet et al., 2000). Textile. Elevated risk has been reported among female textile workers in Finland (Partanen et al., 1994), China (Ji et al., 1999), Spain (Alguacil et al., 2000b), and Denmark (Olsen and Jensen, 1987). No specific exposures were consistently suspected. Woodworking. Workers in the woodworking trade were found to have elevated risk in Finland (Partanen et al., 1994), but not in New Zealand (Kawachi et al., 1989). Stone Quarrying. Miners in Finnish quarries were observed to have elevated risk (Partanen et al., 1994), and a meta-analysis resulted in a significant excess risk attributed to exposure to silica dust (Ojajarvi et al., 2000). Negative Findings for Other Exposures. The most inclusive meta-analysis to date (Ojajarvi et al., 2000) found no support for increased risks associated with exposures to acrylonitrile, asbestos, diesel engine exhaust, EMF, formaldehyde, flour dust, cadmium and cadmium compounds, gasoline, herbicides, iron and iron compounds, lead and lead compounds, manmade vitreous fibers, oil mist, or wood dust. An unknown, but large number of cohort studies have been published without specific reference to pancreas cancer, implying that such results were negative. Such findings are often not disseminated except as adjuncts to other, more positive results. In summary, assessment of environmental causality depends upon the strength of association, the link between the magnitude of exposure and that of risk, the internal and external consistency of study results, and the strict exclusion of alternative explanations, including chance, bias, and confounding. None of these criteria can be definitively assessed for the workplace exposures described above. In general, the workplace experience that has been linked to pancreas cancer is highly varied, without any unifying element. The most suggestive and seemingly consistent associations include those between pancreas cancer and chlorinated hydrocarbons, the PAH/nitrosamine complex, and the pulp and paper industry. However, even these actually involve varied exposures to a wide range of molecules and mixtures. Few strong associations have emanated from the studies of occupational mortality or morbidity that are based on large samples, and most positive studies involve relatively rare occupational experiences, comparisons with inadequate referent groups, and/or comparisons that are potentially confounded, especially by smoking. Observations to date of persons according to occupation and workplace exposure have found no consistently strong and pervasive association with pancreas cancer and no convincing evidence that any common occupation or exposure is responsible for more than a small

Cancer of the Pancreas fraction of pancreas cancer occurrence. The proportion (etiologic fraction) of pancreas cancer cases that can be attributed to workplace exposures is small, generously estimated to be 12% (Ojajarvi et al., 2000). Indeed, in one examination of occupational mortality in Britain (Logan, 1982), no occupation met the rather liberal criteria for inclusion as a high-risk occupation. This should emphasize that the positive observations described above must be revisited in depth to ensure that they are not due to chance, bias, or confounding.

Host Factors Genetic Predisposition It is currently estimated that up to 10% of pancreatic cancers are due to inherited genetic disorders (Tersmette et al., 2001). For recent reviews of genetic predisposition to pancreas cancer, see Bardeesy and DePinho (2002); Cowgill and Muscarella (2003); Jaffee et al. (2002); and Lynch et al. (2004). There is evidence for a major gene or genes involved in familial pancreatic cancer, but as yet none have been identified (Eberle et al., 2002; Klein et al., 2002). Modeling, through segregation analysis, has indicated that the mode of inheritance for at least one gene is autosomal-dominant (Klein et al., 2002). Harboring a mutation in such a gene is estimated to increase risk of the disease 18-fold (Tersmette et al., 2001). Hereditary pancreatitis is caused by a mutation in the cationic trypsinogen gene (Whitcomb et al., 1996). The mutation results in a protein that cannot inactivate trypsin and the overabundance of trypsin leads to autodigestion of the pancreas. Individuals with this defect have an estimated cumulative risk of pancreatic cancer (to age 70) of approximately 40% (Lowenfels et al., 1997). Pancreatic cancer risk has been associated with familial hereditary non-polyposis colon cancer (HNPCC) syndrome (Silverman et al., 1999). The Lynch variant II of HNPCC is associated with mutations in DNA mismatch repair genes including MLH1 and MSH2 (Lynch et al., 2004). Pancreas tumors that arise in individuals from HNPCC kindreds typically have histologic features and somatic mutations that are distinct from most other pancreas cancers. BRCA2 germ-line mutations, which are known to greatly increase the risk of breast and ovarian cancer, have also been associated with increased risk of pancreas cancer (Naderi and Couch, 2002). The BRCA2 gene product is involved in regulation of double-stranded DNA repair (Hruban et al., 2001). The Peutz-Jeghers syndrome (LKB1/STK11) is another inherited disorder that includes an elevated incidence of pancreas cancers (Giardiello et al., 2000). Pancreas carcinomas are among several tumor types, most notably melanoma, which occur at an elevated incidence in a subgroup of families with familial atypical multiple mole melanoma syndrome (FAMMM) (Lynch et al., 2002). Germ-line mutations that give rise to this syndrome occur in the gene that codes for the cyclin-dependentkinase inhibitor-2 (CDK2NA)(p16) (Lynch et al., 2002; Goldstein et al., 1995). Germ-line mutations in the p53 tumor-suppressor genes are associated with increased incidence in a number of cancer types characterized collectively as the Li-Fraumeni syndrome. These include cancers of the breast, brain, lung, and adrenal cortex in addition to soft-tissue sarcomas and leukemias (Lynch et al., 2004). There is too little evidence to say definitively that pancreas cancers are a part of this syndrome, but a moderately increased incidence of pancreas carcinoma was reported in at least one study—a cohort of p53 mutation carriers (Birch et al., 2001).

Associated Medical Conditions Diabetes Mellitus. An association between diabetes and pancreatic cancer has been evaluated in more than 30 studies (Calle et al., 1998; Chow et al., 1995; Everhart and Wright, 1995; Silverman et al., 1999; Wideroff et al., 1997). In the United States, geographic correlations have been reported between pancreas cancer and diabetes, notably in women (Blot et al., 1978). Chart review of patients with diabetes has always resulted in an excess of cases with pancreas carcinoma (Marble, 1934; Ellinger and Landsman, 1944; Bell, 1957).

747

While pancreas cancer appearing in persons with a long history of diabetes has been documented (Cohen, 1965), some pancreas cancer patients diagnosed with diabetes mellitus received that diagnosis during the course of their cancer or in the period immediately preceding it (Green et al., 1958; Clark and Mitchell, 1961; Karmody and Kyle, 1969). The question therefore is whether established diabetics are really at higher subsequent risk of pancreas cancer or whether the risks of cancer and diabetes are related in some other way. Kessler followed 21,000 diabetics seen at a large Boston clinic between 1930 and 1956 for evidence of cancer on death certificates through 1959 (Kessler, 1970). Eligibility criteria included survival for at least 1 year after diagnosis, and a large number of pancreas cancer cases (20% of all deaths) were ineligible by this criterion, reaffirming that the diagnosis of cancer coincident with the diagnosis of diabetes is commonplace. An additional 11 of the 78 observed cases were diagnosed but did not die within 1 year following the diagnosis of diabetes. Based on overall mortality in Massachusetts, the standard mortality ratios for males and females were 1.5 and 2.1, and these fell to 1.3 and 1.8 after excluding the 11 cases diagnosed within a year after diagnosis of diabetes. A potential bias in this study exists because the physicians at the diabetes clinic had a strong interest in diseases of the pancreas and may have determined pancreas cancer as a cause of death more often than other Massachusetts physicians. In Olmsted County, Minnesota (Maruchi et al., 1979), no excess of pancreas cancer in diabetics was observed after cases diagnosed more than 2 years after the initial diagnosis of diabetes were excluded. Ragozzino et al. (1982) followed a population-based cohort of 1135 diabetics in Minnesota diagnosed between 1945–1969. There were 9 observed pancreatic cancers and 2.1 expected. When five cancer cases that were diagnosed in the first year following the onset of diabetes were excluded, the SMR was 2.6 (95% CI: 0.9–6.1) with similar findings for both sexes. Green and Jensen (1985) followed a cohort of 1499 insulindependent diabetics in Denmark. Based on age- and sex-specific Danish cancer incidence rates, 2.4 pancreas cancer cases were expected among the cohort during the 8.5 years of follow-up. Two cases in whom cancer was diagnosed less than 5 years from the onset of diabetes were excluded from the analysis, four other cases of pancreatic cancer were observed (O/E = 1.69, p = 0.29); the four included in the analysis were diagnosed with cancer 6, 7, 8, and 37 years after the onset of diabetes. The association between preexisting diabetes and pancreatic cancer has been examined in other prospective studies. Whittemore et al. (1983) found a statistically significant increased risk of approximately fivefold for the development of pancreas cancer among diabetics in a large prospective study of college men. Mills et al. (1988) reported a relative risk of 3.4 (95% CI: 1.7–8.3) for pancreatic cancer among self-reported diabetics in a cohort of Seventh-Day Adventists; the analysis was restricted to individuals who survived at least 1 year after entry into the study. In a large prospective study in Japan, selfreported male diabetics had a relative risk of pancreatic cancer death of 2.12 (95% CI: 1.19–3.77) while in female diabetics, there was a 50% increase in the risk that was not statistically significant (Lin et al., 2002b). The relative risk of incident pancreas cancer among postmenopausal women in Iowa (US) who self-reported diabetes compared with those who did not was 2.9 (95% CI: 1.61–5.23)(Anderson et al., 2002b). An association with preexisting diabetes was observed in two prospective studies of members of a prepaid health plan in the San Francisco Bay area; Hiatt et al. (1988) reported a relative risk of 2.1 for subjects with a history of diabetes among a cohort of 122,894 individuals; all of the cancer cases were diagnosed with diabetes at a minimum of 5.7 years prior to the diagnosis of cancer. A prospective study (nested case-control study) conducted within this same health plan found a twofold increase in risk of pancreatic cancer among subjects who had previously responded yes to questions on diabetes history during their first multiphasic health checkup (Friedman and van den Eeden, 1993). The relative risk was virtually unchanged when cases whose cancer developed within 5 years of the first check-up were excluded.

748

PART IV: CANCER BY TISSUE OF ORIGIN

More recent cohort studies of diabetics have focused on risk in relation to length of follow-up. Chow et al. (1995) observed that Swedish diabetics experienced a 70% increased risk with 10 or more years of follow-up. Wideroff et al. (1997) found that Danish diabetics also had an elevated risk of pancreatic cancer, but the elevation declined to an SIR of 1.3 (95% CI: 1.1–1.6) after 5–9 years of follow-up. In contrast, Calle et al. (1998) reported a significant, 40% increased risk in 9–12 years of follow-up based on the American Cancer Society cohort of 1,089,586 Americans. Among the case-control studies that have examined preexisting diabetes as an independent variable, several have included other hospital patients as controls (Wynder, 1973a; Lin and Kessler, 1981; MacMahon et al., 1981; Gold et al., 1985; LaVecchia et al., 1990). Only two of these studies (Wynder, 1973a; LaVecchia et al., 1990) found an association between diabetes and pancreas carcinoma: one noted the association with long-standing diabetes and then, only in women (Wynder, 1973a); the other found an increased risk, but the authors did not indicate if recently diagnosed diabetics were excluded from the analysis. A simple interpretation of these various results is difficult because of the inclusion of hospital patients in the control series: such a series may contain an excess of subjects with diabetes-related conditions relative to the general population. Cuzick and Babiker (1989) addressed this issue in their report of results from a case-control study, which included both hospital and general practice controls. The authors noted that there were no material differences in their results when only one of the two control groups was used. After excluding patients diagnosed with diabetes less than a year prior to diagnosis of cancer, they reported a fourfold risk ratio associated with diabetes, which did not change with the further exclusion of diabetics diagnosed within 2–5 years of their cancer. Among six population-based case-control studies that have examined this issue, at least four reported positive associations. Norell et al. (1986c) reported an OR of 2.4 (90% CI: 0.6–9.7); Farrow and Davis (1990a) found an OR of 6.7 (95% CI: 1.8–24.9). In both studies, subjects diagnosed with cancer less than 5 or 3 years, respectively, after the onset of diabetes, were excluded from analysis. Jain et al. (1991) found an OR of 2.1 (95% CI: 0.7–6.2) for subjects diagnosed with diabetes between 5 to 10 years prior to their cancer. More recently, Silverman et al. (1999) reported a significant positive trend in risk with increasing years prior to diagnosis of pancreatic cancer, with diabetics diagnosed at least 10 years prior to pancreatic cancer having a significant 50% increased risk. This observation provided support for the hypothesis that long-standing diabetes is a risk factor for pancreatic cancer, as well as a possible consequence of the tumor. In 1995, Everhart and Wright reported results of a meta-analysis of 20 pancreatic cancer studies and concluded that long-standing diabetes was an independent risk factor for pancreatic cancer. Diabetics diagnosed at least 5 years prior to the diagnosis of cancer had a relative risk (RR) of 2.0 (95% CI: 1.2–3.2). The pooled estimate from cohort studies (RR = 2.6) was higher than that from case-control studies (OR = 1.8). In 2005, Huxley et al. conducted a meta-analysis of 36 studies that confirmed the results of the earlier meta-analysis, providing additional support for a causal association between long-standing diabetes and pancreatic cancer risk. Although various mechanisms by which long-standing diabetes causes pancreatic cancer have been proposed, the most favored mechanism involves hyperinsulinemia. In 1995, Everhart and Wright suggested that hyperinsulinemia might play a role in diabetes-related pancreatic cancer. They speculated that the exocrine pancreas is exposed to a high concentration of pancreatic islet cell hormones because it receives much of its blood supply through the islets, providing plausibility for the hyperinsulinemia hypothesis. Three additional experimental/epidemiologic observations also supported a mechanistic role for hyperinsulinemia. First, experimental studies indicated that insulin has a dose-dependent growth-promoting effect in human pancreatic cell lines (Fisher et al., 1996). Second, Silverman et al. (1999) observed that although diabetes-related pancreatic cancer was not influenced by obesity, the risk associated with increasing BMI was seen only in non-diabetics. Among non-diabetics, resistance to insulin action typically increases with increasing BMI, resulting in

hyperinsulinemia. Among diabetics, insulin levels are not strongly correlated with BMI, but instead depend mainly on the degree of impaired B-cell function and hyperglycemia. Third, metformin treatment, which improves insulin resistance and reduces hyperinsulinemia, prevented pancreatic adenocarcinomas in hamsters fed high-fat diets and treated with the pancreatic carcinogen N-nitrosobis-(2-oxopropl)amine (Schneider et al., 2001). Most recently, results of four cohort studies (Gapstur et al., 2000; Michaud et al., 2002; Batty et al., 2004; Jee et al., 2005) were also consistent with the hyperinsulinemia hypothesis. In an expansion of the Chicago Heart Association cohort study, Gapstur et al. (2000) showed mortality from pancreatic cancer increased with increasing postload plasma glucose level. This association was virtually unchanged after excluding patients who died within the first 5 years of follow-up, indicating that the finding was not due to detection bias. These findings have recently been confirmed in an expansion of the Whitehall Study (Batty et al., 2004) and in a large Korean cohort (Jee et al., 2005). Gapstur et al. (2000) suggested that high concentrations of insulin might bind to and activate the insulin-like growth factor 1 (IGF-1) receptor, which has growth promotion effects including modulation of cell cycle progression. Excess insulin also may operate indirectly through down-regulation of insulin-like growth factor-binding protein 1 (IGFBP-1), which could increase bioavailability of IGF-1, a stimulator of pancreatic cell proliferation in vitro. Based on 180 pancreatic cancer patients in the Nurses’ Health Study, Michaud et al. (2002) found that a diet high in glycemic load may increase the risk of pancreatic cancer in women who are both overweight and sedentary, conditions associated with hyperinsulinemia and insulin resistance.

Pancreatitis. Pathologists have observed that pancreatitis occurs in a large proportion of autopsied cases of pancreatic carcinoma (Mikal and Campbell, 1950), and clinical signs of pancreatitis have been observed in many patients with pancreatic cancer (Gambill, 1971). A hereditary form of pancreatitis has been described where affected family members appear at high risk of pancreatic adenocarcinoma (Castleman et al., 1972; Kattwinkel et al., 1973). It has been demonstrated that, in some instances, chronic pancreatitis clearly precedes the diagnosis of carcinoma of the pancreas (Bartholomew et al., 1958; Robinson et al., 1970; Lundh and Nordenstam, 1970; Ammann et al., 1984; Lowenfels, 1984; Ammann and Schueler, 1984), as well as other carcinomas (Ammann et al., 1980). The hypothesis that chronic pancreatitis conveys an excess risk of pancreas cancer has been examined in a number of case-control studies, and evidence for a positive association (Lin and Kessler, 1981; Mack et al., 1986, Farrow and Davis 1990a; La Vecchia, et al., 1990; Jain et al., 1991; Kalapothaki et al., 1993a) and for no association (Wynder et al., 1973a; Gold et al., 1985; Bueno de Mesquita, 1992b) has been reported. However, the data from such studies are based on relatively few individuals with evidence of prior pancreatitis. Lowenfels et al. (1993) reported on a multicenter historical cohort study of 2015 patients with chronic pancreatitis. The risk of pancreatic cancer was significantly elevated. Even after excluding individuals with only 2 or 5 years of follow-up, the standardized incidence ratios were 16.5 (95% CI: 11.1–23.7) and 14.4 (95% CI: 8.5–22.8), respectively. These ratios were derived from the observed number of cases and the expected number based on country-specific incidence data. The increased risk was observed in men and women, was associated with both alcoholic and non-alcoholic pancreatitis, and was observed in the individual cohorts from all six participating countries. The center-specific relative risks were all statistically significant and ranged from 9.7–19.9 for the analysis restricted to individuals with 2 or more years of follow-up; all but one of the relative risk estimates were statistically significant when the analysis was restricted to individuals with 5 or more years of follow-up, with a range of 3.6–24.3. An increased risk associated with chronic pancreatitis has been observed in at least six additional studies in the past decade (Bansal and Sonnenberg, 1995; Chari et al., 1994; Ekbom et al., 1994; Fernandez et al., 1995a; Malka et al., 2002; Talamini et al., 1999).

Cancer of the Pancreas Elevated pancreatic cancer risk has also been observed for patients with two rare forms of pancreatitis, hereditary pancreatitis and tropical pancreatitis. In fact, patients with hereditary pancreatitis have been reported to have a 50- to 70-fold risk (Lowenfels et al., 1997). Although no interaction between smoking and hereditary pancreatitis was observed, pancreatic cancer developed about 20 years earlier in smokers than in nonsmokers with hereditary pancreatitis (Lowenfels et al., 2001). There is increasing evidence to support the hypothesis that patients with chronic pancreatitis, independent of pancreatitis risk factors, are at increased risk for pancreatic cancer. It is also true, however, that even if chronic pancreatitis is a risk factor for pancreatic cancer, it could explain only a small fraction of pancreatic cancers (Gold and Cameron, 1993). The significance of such a link is that it may provide clues about the etiology of this enigmatic cancer. Chronic pancreatic inflammation may play an important role in the elevated pancreatic cancer risk experienced by patients with chronic pancreatitis (Farrow and Evers, 2002). Inflammation associated with hereditary pancreatitis, in particular, is persistent and involves progressive glandular destruction over a long time, which may be related to the unusually high pancreatic cancer risk reported in hereditary pancreatitis patients (Lowenfels et al., 2001). In a European hereditary pancreatitis cohort, nearly 40% of patients developed pancreatic cancer by age 70 (Lowenfels et al., 1997), a risk so high that it is unlikely to be explained by detection bias. The mechanism by which chronic inflammation increases susceptibility to pancreatic cancer is unclear, however. Several mechanisms have been suggested including genomic instability, accelerated accumulation of cancer-causing mutations, and aberrant methylation (Rosty et al., 2003; Whitcomb and Pogue-Geile, 2002).

Gastric Surgery. A number of studies have shown higher risks for pancreas cancer following gastrectomy or peptic ulcer surgery. McLean-Ross and colleagues (1982) observed 11 cases of pancreatic cancer (3.9 expected) when they followed-up nearly 780 members of a Scottish gastrectomy cohort. Caygill et al. (1985) reported a 3.2-fold excess of pancreatic cancers among approximately 4200 gastrectomy patients after a 20-year latent period. Increased risks for other cancers were also reported in the cohort, including tumors of the large bowel, bronchus, biliary tract, and other sites. In contrast Maringhini et al. (1987) followed 336 gastrectomy patients in a cohort study in Olmsted County, Minnesota, and found 1 pancreatic cancer where 1.6 were expected. In a cohort of 4224 gastrectomy patients, Eide et al. (1991) reported O/E ratios of 1.7 for males (95% CI: 1.2–2.4) and 0.2 (95% CI: 0.1–1.4) for females. A marginally increased risk for pancreatic cancer, though not statistically significant, was also observed in another cohort of over 4000 gastrectomy patients (Møller and Toftgaard, 1991). In 2633 gastrectomy patients followed for up to 59 years after surgery, Tascilar et al. (2002) found an O/E ratio of 3.5 at 35 years or more years after surgery. The authors argue that this longterm effect is unlikely to be due to smoking because the pattern of lung cancer risk after gastrectomy does not increase with increasing years after surgery as it does for pancreatic cancer. It is difficult, however, to dismiss smoking as a confounder because no information on smoking in the cohort is available. Offerhaus et al. (1987) conducted a case-control autopsy study and included three control groups, each with a high prevalence of cigarette smoking, to adjust indirectly for the possible confounding effect of smoking. Increased risks of pancreatic cancer were found in comparisons with each control group. When all controls were included in the analysis, an OR of 3.1 (95% CI: 1.2–8.0) was found. Mack et al. (1986) also adjusted for smoking and reported a positive association for pancreatic cancer and gastrectomy (OR = 5.3, 95% CI: 1.6–21.5), as did Farrow and Davis (1990a) (OR = 2.3, 95% CI: 0.4–12.7), and Mills et al. (1988) for peptic ulcer surgery (OR = 2.6, 95% CI: 1.0–6.9). Studies reporting no strong association include those of Wynder et al. (1973a), La Vecchia et al. (1990), Bueno de Mesquita et al. (1992b), and Silverman et al., (1999). Several plausible mechanisms for an increased risk of pancreatic cancer among gastrectomy patients have been described (Mack et al.,

749

1986). Gastric secretion mediates both hormonal and neurological regulation of the pancreas, and gastrectomy could alter homeostatic regulation of hormones that are trophic for the pancreas. Detoxification of endogenous and exogenous substances in the small intestine might be less efficient after gastrectomy. Finally, gastrectomy can increase the endogenous production of N-nitroso compounds (Schlag et al., 1980), possibly as a result of pH-induced changes in bacterial growth, and is a plausible mechanism to explain the observed increased risk of pancreas cancer in gastrectomy patients (Caygill et al., 1985; Mack et al., 1986; Tascilar et al., 2002).

Cholecystitis. Clinicians (Berk, 1941) and pathologists (Mikal and Campbell, 1950) have noted that cholecystitis is commonly present in patients with pancreas carcinoma, but findings of an association between pancreatic cancer and cholecystectomy are inconsistent. Of 23 studies conducted to date, 14 have been positive (Wynder et al., 1973a; Lin and Kessler, 1981; Norell et al., 1986c; Hyvarinen and Partanen, 1987; Cuzick and Babiker, 1989; La Vecchia et al., 1990; Kalapothaki et al., 1993a; Shibata et al., 1994; Ekbom et al., 1996; Johansen et al., 1996; Chow et al., 1998; Silverman et al., 1999; Coughlin et al., 2000; Lin et al., 2002b). Few studies, however, have considered timing of the cholecystectomy/gallbladder disease in relation to pancreatic cancer. Of the 14 studies indicating a positive association, five demonstrated an increased pancreatic cancer risk 5 or more years prior to cancer diagnosis (Norell et al., 1986c; Hyvarinen and Partanen, 1987; Shibata et al., 1994; Chow et al., 1998; Silverman et al., 1999). Haddock and Carter (1990) proposed an explanation for a possible predisposition for cancer among subjects with cholelithiasis. They point out that cholecystectomy in hamsters can lead to an increase in circulating cholecystokinin (CCK) levels. CCK is a major regulator of pancreatic growth (Silverman et al., 1999) and elevated CCK levels can lead to hypertrophy and hyperplasia in the pancreas. Receptors for CCK have been found in human pancreatic cancer cells and CCK is a promoter of pancreatic carcinogenesis in rodents (Howatson and Carter, 1985; Smith et al., 1990). Thus, a role for cholecystectomy as a promoting factor for pancreatic cancer, perhaps mediated by the trophic effects of CCK is plausible, though not established. Allergies. The association of prior allergies with pancreatic cancer has been examined in a number of studies (Lin and Kessler, 1981; Gold et al., 1985; Mack et al., 1986; Mills et al., 1988; Dai et al., 1995; Farrow and Davis, 1990a; La Vecchia et al., 1990; Jain et al., 1991; Bueno de Mesquita, 1992b; Kalapothaki et al., 1993a; Silverman et al., 1999; Holly et al., 2003). Although the data are somewhat equivocal, the preponderance of evidence suggests that a history of allergies may be inversely associated with the risk of subsequently developing pancreas cancer. A cohort of asthmatic veterans was followed for 25–30 years; their cause-specific mortality was compared with national expectations and with the mortality occurring in a cohort of similar size and age, but with an initial diagnosis of acute nasopharyngitis (Robinette and Fraumeni, 1978). In relation to both, the expected deaths from all digestive system cancers (based on national mortality) and the observed pancreas cancer deaths in the comparison cohort, a substantial excess of pancreas cancer mortality occurred among the asthmatics. The authors expressed concern that this effect might have been related to aminophylline therapy for asthma, on the basis of an effect upon the DNA binding of carcinogens that has been observed in cultured cells (Huberman and Sachs, 1977). Others have examined the relation between these conditions, and found no increase in pancreas cancer (Farrow and Davis, 1990a; La Vecchia et al., 1990); most have observed an overall reduced risk of cancer in asthmatics (Shapiro et al., 1971; Meers, 1973; Alderson, 1974; Mack et al., 1986; Jain et al., 1991; Kalapothaki et al., 1993a; Dai et al., 1995), which is consistent with the findings for a history of allergies noted above. Because most case-control studies of pancreatic cancer have been based primarily on interviews with next of kin, the reliability of detailed information on allergic disorders may be questionable (Silverman et al., 1999). However, two case-control studies have been

750

PART IV: CANCER BY TISSUE OF ORIGIN

based solely on direct interviews (Silverman et al., 1999; Holly et al., 2003). Overall results of both studies were remarkably consistent; a prior history of any allergic condition was associated with a 20%– 30% reduced risk. Moreover, Holly et al. (2003) observed significant inverse trends in risk with increasing number of allergies and severity of symptoms, providing support for the plausibility that immune function in relation to allergies may play a role in pancreatic cancer etiology. Mack et al. (1986) proposed a mechanism by which trophic hormones mediate a protective effect of allergies.

Other Conditions. Lin and Kessler (1981), Gold et al. (1985), and Farrow and Davis (1990a) have all reported finding prior tonsillectomy associated with a significantly reduced risk of pancreatic cancer. Mills et al. (1988) reported a slight, though nonsignificant inverse association, as did Mack et al. (1986). Bueno de Mesquita et al. (1992b) found no association. No increased risk was found in a large cohort of patients with rheumatoid arthritis studied in Finland (Isomaki et al., 1978); nor was an excess risk found in patients with celiac disease (Holmes et al., 1976). An increased incidence of pancreatic neoplasia (O/E = 3.8, P < 0.02) in pernicious anemia patients vs. the general population of Sweden has been reported (Borch et al., 1988; Hsing et al., 1993). Increased endogenous production of N-nitroso compounds has been suggested as a possible mechanism to explain the increased incidence of pancreas cancer observed in pernicious anemia patients (Borch et al., 1988; Tascilar et al., 2002). Reproductive and Hormonal Factors Pancreatic cancer is more common among males than females, in both humans and in animal models (Andren-Sandberg et al., 1999). Steroid hormones play a role in normal pancreatic function and thus it is plausible that they modify pancreatic carcinogenesis (Andren-Sandberg et al., 1999; Robles-Diaz and Duarte-Rojo, 2001). Aspects of reproductive history and exogenous hormone use have been associated with female breast and other cancers; their role in pancreatic cancer has been hypothesized. At least two cohort (Kvale et al., 1994; Skinner et al., 2003) and eight case-control (Bueno de Mesquita et al., 1992c; Kalapothaki et al., 1993a; La Vecchia et al., 1993; Fernandez et al., 1995b; Ji et al., 1996; Karlson et al., 1998; Kreiger et al., 2001; Duell and Holly, 2005) studies have reported on reproductive factors and pancreatic cancer risk. Overall, the data are suggestive for a role of some reproductive factors, notably parity, but there are caveats. Duell and Holly (2005), in a population-based case-control in California, analyzed data from 241 female cases and 81.8 controls. They found no consistent pattern for reproductive factors and risk of pancreatic cancer. Three often case-control studies had less than 100 cases and therefore may have lacked power to detect significant differences (Bueno de Mesquita et al., 1992c; Kalapothaki et al., 1993a; Krieger et al., 2001). Three others (La Vecchia et al., 1993; Kvale et al., 1994; and Karlson et al., 1998) did not control for smoking, which is a well-established risk factor for pancreatic cancer and influences estrogen levels (Westhoff et al., 1996) and some did not adjust for other reproductive factors. Finally, many of the case-control studies used surrogate interviews, which likely increases reporting error of some reproductive factors such as those related to menstruation. Skinner and colleagues (2003) analyzed data from a large prospective study in the United States that included adjustment for confounding variables. Their strongest association was for parity. In a multivariate-adjusted analysis, the relative risk of pancreatic cancer was 0.86 for women with 1–2 births, 0.75 in those with 3–4 births, and 0.58 for those with 5 or more births compared with nulliparous women. These RRs were not statistically significant, but an analysis for linear trend, indicating a 10% reduction with each birth, was (P = 0.008). Seven of the other above-mentioned studies also examined the effect of parity and/or number of pregnancies/births on pancreatic cancer risk. The associations reported were inverse (Krieger et al., 2001; Fernandez et al., 1995b; Kalapothaki et al., 1993a), null (La Vecchia et al., 1993; Karlson et al., 1998; Duell and Holly, 2005), pos-

itive (Kvale et al., 1994; Ji et al., 1996), and mixed (Bueno de Mequita et al., 1992c). Age at first birth was examined in nine studies and investigators found inverse (Krieger et al., 2001; Karlson et al., 1998; Fernandez et al., 1995b; Skinner et al., 2003; Bueno de Mesquita et al., 1992c), null (Kalapothaki et al., 1993a; La Vecchia et al., 1993; Duell and Holly, 2005), and positive associations (Ji et al., 1996). In the one study to report on breast-feeding, the authors found a non-statistically inverse association with pancreatic cancer risk (Skinner et al., 2003). Early age at menarche was positively associated with risk in three studies (Bueno de Mesquita et al., 1992c; Fernandez et al., 1995b; Kalapothaki et al., 1993a) and null in three others (Ji et al., 1996; Krieger et al., 2001; Skinner et al., 2003) while data on late age at menopause and related variables were inconsistent across studies. There was no effect of spontaneous or induced abortions in the studies that analyzed these variables (Bueno de Mesquita et al., 1992c; Kalapothaki et al., 1993a; La Vecchia et al., 1993; Fernandez et al., 1995b; Ji et al., 1996). A number of other reproductive and hormonal variables were also considered in relation to pancreatic cancer risk in these studies, including use of exogenous estrogens—oral contraceptives and estrogen replacement therapy; a clear pattern in the findings does not emerge, however. The presence of estrogen and androgen receptors has been demonstrated in normal human pancreas and in some carcinomas of the pancreas (Greenway et al., 1981; Corbishley et al., 1986). This fact, together with experimental data on animal and human tissues (Benz et al., 1986; Greenway et al., 1982; Sperti et al., 1992; Tulassay et al., 1988), provided the rationale for clinical trials with both antiestrogens and anti-androgens and other drugs, which alter sex hormone levels. These agents, however, have not been shown to have appreciable effects on pancreatic cancer patient survival to date (Andren-Sandberg, 1999; Corrie et al., 2002). Overall, epidemiologic studies do not provide consistent evidence for an association between reproductive factors, exogenous hormone use and pancreatic cancer risk. Further research with biomarkers that would allow for more specific hypotheses may be useful.

MOLECULAR PATHOGENESIS For pancreas cancer, as with other human neoplasms, there is now steadily accumulating knowledge regarding the molecular events involved in carcinogenesis. Models of the molecular pathogenesis of pancreas cancer—analogous to those developed for colon cancer (Vogelstein et al., 1988)—have been developed (Hruban et al., 2000). These models are being refined as scientists more fully characterize the sequence of somatic molecular alterations and their specific biological consequences on cellular pathways and processes that occur in cancer progression. Gross chromosomal abnormalities have been observed in some malignant and precursor lesions (PanIN). These are of interest for their role in etiology and progression of this cancer and include translocations, amplifications, deletions, microsatellite instability, altered methylation, and telomere dysfunction (Hansel et al., 2003; Hine et al., 2003). The list of somatic changes in specific genes considered relevant to pancreas carcinogenesis is growing (Jaffee et al., 2002). Targets include cellular proto-oncogenes, tumor suppressor genes, growth factors and their receptors, cytokines, matrix metalloproteinases, and DNA mismatch repair genes. The documented changes include mutations, allele loss, intragenic mutation or hypermethylation in promoter regions (leading to loss of function), and overexpression of growth factors and their receptors (Hansel et al., 2003). Recent reviews detail the molecular biology and genetics of pancreas cancer (Hruban et al., 2001; Reddy, 2001; Bardeesy and DePinho, 2002; Cowgill and Muscarella, 2003; Jaffee et al., 2002; and Lynch et al., 2004). Our understanding of the pathogenic mechanisms of human pancreas cancer is informed in important ways by animal models and readers are referred to several excellent reviews covering that field (Standop et al., 2001; Wei et al., 2003).

Cancer of the Pancreas In a previous section, we described genetic predisposition to pancreas cancer. Below, we describe the most common somatic alterations observed in relation to human pancreas tumors to highlight progress in this important area and underline the significance for epidemiologic research on prevention and etiology. We briefly cite some investigations of genetic susceptibility and gene-environment interactions as well.

Somatic Mutations Although somatic mutations in at least 15 genes have been reported to occur in low frequencies in pancreatic cancers, alterations in four genes: the c-Kirsten-ras (K-ras) oncogene, and the p16, p53, and DPC4/SMAD4 tumor suppressor genes have been reported in a high proportion (over 50%) of pancreatic tumors (Jaffee et al., 2002), and are commonly seen together (Rozenblum et al., 1997). These mutations disable key regulatory pathways in cell growth and DNA repair. Scientists have begun to see how cells that become neoplastic have acquired an amalgam of defective pathways that allow for dysfunctional cells to evade growth control, DNA repair, and/or programmed cell death. The ras proto-oncogene family encodes membrane-bound proteins involved in GTP-mediated signal transduction across the cell membrane. The proteins are important regulators of cell growth and differentiation (Shields et al., 2000). Point mutations in these genes can render the protein constitutively activated and such mutations probably occur early in carcinogenesis for many cancers (Fearon and Vogelstein, 1990). It is believed that K-ras mutations are early and important events in the pathogenesis of pancreatic cancer. Mutations in K-ras genes occur in many, though not all, experimental pancreatic cancers (van Kranen et al., 1991; Cerny et al., 1992). The highest frequency of K-ras mutations in humans has been found in case series of adenocarcinomas of the pancreas (90%–100%) with the great majority of these mutations being found in codon 12 (Alomoquera et al., 1988; Tada et al., 1990; Rozenblum et al., 1997; Slebos et al., 2000). The K-ras gene is amplified in some malignant tumors and it is mutated in some preinvasive lesions and in both primary and metastatic foci (Shields et al., 2000; Bardeesy and DePinho, 2002). Testing for mutations in K-ras (together with those occurring in p53, and DPC4) has been employed as a diagnostic aid (van Heek et al., 2002) where tumor cells recovered from pancreatic juice or fine needle aspirates are analyzed. However, because K-ras mutations have been found in some normal pancreatic cells and in benign pancreatic disease, the mutation is not diagnostic for malignancy. The precise meaning and importance of these mutations is still a matter of debate (Lohr et al., 2000). Inactivation of the p16 (also known as CDKN2A) tumorsuppressor gene product occurs in a high proportion (about 95%) of adenocarcinomas of the pancreas (Hruban et al., 2001). Loss of function in the alleles can occur by a combination of events such as hypermethylation of the promoter region, allele loss, and intragenic mutation. When functioning normally, the p16 protein blocks a cyclindependent kinase from forming complexes necessary to phosphorylate the RB protein. Phosphorylating the RB protein lifts its inhibitory effects on the transition from G1 to S phase. Mutations in the p16 gene result in phosphorylated Rb proteins, thereby disabling an important checkpoint in the cell cycle (Hurban et al., 2001). DPC4 (also known as SMAD4 or MADH4) is a tumor-suppressor gene inactivated in about 55% of pancreas carcinomas (Jaffee et al., 2002). The gene product, the Smad4 protein plays a role in growth inhibitory signal transduction for the transforming growth factor-beta superfamily of cytokines (Hahn et al., 1996; Cowgill and Muscarella, 2003). Evidence also suggests that the loss of function for Smad4 may upregulate angiogenesis (Schwarte-Waldhoff et al., 2000). Alterations in the p53 tumor suppressor gene have been reported in 50%–75% of pancreas tumors (Jaffee et al., 2002). The p53 protein is a nuclear transcription factor with a pivotal role in cell cycle control and apoptosis and, as stated above, germ-line mutations in p53 are the underlying cause of the Li-Fraumeni familial cancer syndrome.

751

Growth Factors and Their Receptors It is well established that growth factors are important in normal pancreatic development and function as well as in the pathogenesis of pancreatic cancer (Friess et al., 1999; Kiehne et al., 2001; Reddy, 2001; Hruban et al., 2001). Research in this area may clarify our understanding of the disease biology and identify therapeutic targets. In normal cells, growth factors bind to plasma membrane receptors triggering particular signal cascades that cause cell growth, differentiation, and/or cell death (Reddy, 2001). The simultaneous overexpression of receptors and their ligands during cancer development is thought to result in autocrine and/or paracrine stimulation in these pathways (Friess et al., 1999; Korc et al., 1990). The epidermal growth factor receptor-ligand system can illustrate one type of alteration important in pancreatic malignancies. There are several members of the epidermal growth factor family of receptors (currently numbered 1 through 4) and they each go by several names, for example, EGF receptor 1, c-erb-B1, ErbB1, and HER-1 (Reddy, 2001). These transmembrane proteins are tyrosine kinases. Their ligands include epidermal growth factor (EGF) and transforming growth factor-alpha (TGF-alpha) (Reddy, 2001). Both the receptors and ligands are expressed in a limited manner in normal pancreas tissue (Korc et al., 1992). Upon binding a ligand, receptors form dimers with their sibling receptors in the ErbB superfamily (Reddy, 2001). The types of dimers formed then direct the particular signaling cascade and cellular response. In experimental systems, overexpression of the EGF receptor results in malignant transformation and both EGF and TGF-alpha can further stimulate proliferation in transformed cells (Friess et al., 1999). Overexpression of receptors for the various EGF ligands, such as erbB-2 and erbB–3 proteins, has been found in a majority of pancreatic carcinomas investigated (Lemoine et al., 1992). The overexpression results from increased mRNA transcription rather than gene amplification (Friess et al., 1999). The ligands for the EGF receptors: EGF, TGF-alpha, amphiregulin, and others, are also overexpressed in a high proportion of the cancers compared with normal cells. Evidence indicates that pancreatic cells are stimulated to grow by an autocrine loop in which excess TGF-alpha binds to excess EGF receptors expressed by those same cells that produce the TGF-alpha. Thus a growth advantage is provided to cells that concomitantly produce —or overproduce—TGF-alpha and overexpress the EGF receptor that binds it (Friess et al., 1999; Korc, 1990). This autocrine loop is thought to afford a particular advantage when coupled with a mutation in another portion of the signaling cascade; for TGF-alpha, this occurs when overexpression of the ligands and receptors occurs in conjunction with a mutated ras protein (Keihne et al., 2001). Many pancreatic cancers exhibit dysfunction in other important signal transduction pathways such as those involving the transforming growth factor-beta (TGF-beta), Fibroblast growth factor, Hepatocyte growth factor, and Tumor Necrosis factor (Keihne et al., 2001). TGF-beta has important inhibitory effects on growth, extracellular matrix events, cell differentiation, and angiogenesis (Friess et al., 1999), but most pancreas cancers lose their ability to receive the inhibitory message from TGF-beta (Keihne et al., 2001). This is thought to be due to down stream defects in the signaling cascade (i.e., a mutation in a SMAD protein4 (DPC4)), which are mutated in nearly half of pancreas carcinomas (Jaffee et al., 2002). The significance to pancreatic neoplasia of abnormalities in many other cell functions is being investigated; these include overexpression of cytokines (Friess et al., 1999) and molecules involved in tumor invasion (e.g., matrix metalloproteinases (Bloomston et al., 2002)), and molecules integral to angiogenesis (e.g., cell surface adhesion molecules (Perugini et al., 1998)).

Carcinogens and Genetic Polymorphisms Relatively few epidemiologic risk factors have been identified with certainty. Of these, cigarette smoking is the most established cause of human pancreatic cancer. Potential pancreatic carcinogens that are present in tobacco smoke (as well as in the diet and environmental

752

PART IV: CANCER BY TISSUE OF ORIGIN

exposures) include N-nitroso compounds, polycyclic aromatic hydrocarbons, and aromatic amines (IARC, 2004b) (see section on “Smoking” above). N-nitroso compounds can form DNA-adducts and induce pancreas tumors in rodent models (Hecht, 1998). DNA adducts from heterocyclic amines, aromatic amines, and PAHs have been detected in human pancreatic tissue (Li and Jiao, 2003; Thompson et al., 1999), which confirms their role as plausible human pancreatic carcinogens since these adducts may lead to mutations. Damage to DNA can occur from endogenous sources as well (e.g., from reactive oxygen species formed during inflammatory processes like pancreatitis or as a physiological consequence of obesity) (Farrow and Evers, 2002; Rose et al., 2004). As described above, mutations in codon 12 of the K-ras are frequent in human pancreatic cancer; preferential DNA damage and poor repair relative to other sites are thought to explain, in part, this high mutation frequency (Hu et al., 2003). The most common mutations in codon 12 are G to T and G to C transversions and G to A transitions. Carcinogens that can cause such mutations include PAHs (Hu et al., 2003), nitrosamines (Hecht, 1998), aromatic amines (Beland and Kadlubar, 1985), and heterocyclic amines (US NTP, 2004). To form DNA-adducts, PAHs, nitrosamines, and aromatic amines require metabolic activation by drug-metabolizing enzymes. Individuals vary in their ability to activate and inactivate these chemical carcinogens (Kadlubar et al., 1993; Bartsch et al., 1998; Fretland et al., 2003; Perera and Weinstein, 2000; Malats, 2001). Genetic polymorphisms in genes that encode enzymes that activate carcinogens or repair DNA may be effect modifiers for causal agents. We know relatively little about gene-environment interaction in pancreas cancer etiology, as there have been few genetic association studies. Genes that have been investigated include those that encode P450 1A1, 2D6, and 2E1; N-acetyltransferases; glutathione Stransferase M1, T1; NAD(P)H:quinone oxidoreductase; and UDP glucuronosyltransferase (Lee et al., 1997; Bartsch et al., 1998; Duell et al., 2002a; Ockenga et al., 2003). For the most part, the results have been null. Modest elevations in risk have been seen for a few polymorphisms, but the findings were based on small numbers of cases and the functional importance of the polymorphisms has not necessarily been established. Duell et al. (2000b) examined the Arg399Gln polymorphism in the XRCC1 gene, which encodes a DNA base excision repair enzyme. The polymorphism was not independently associated with pancreatic cancer, but smokers with the variant allele were at higher risk for cancer than were smokers with the wild type allele (Duell et al., 2002b). Subjects numbering in the thousands will be required to detect modest RRs for interactions between suspected risk factors and genetic polymorphisms. Future research based on collaborative endeavors with new methods to analyze large and complex amounts of information may be helpful towards this end (Maojo and MartinSanchez, 2004).

PREVENTION Risk of exocrine pancreas cancer increases dramatically with age, and is generally higher in men and in certain racial groups. Differences in rates worldwide may be explained, in part, by behavioral differences, particularly differences in smoking and nutritional patterns. Smoking, obesity, high caloric intake, diabetes, and excessive consumption of red or processed meats appear to be predictive of increased risk while high fruit and vegetable consumption may decrease risk. Occupational studies have not established the significance of specific industrial chemical exposures. Less than 10% of pancreatic cancers are attributed to germ-line mutations. The most significant step that could be taken to prevent pancreas cancer would be the elimination of cigarette smoking (American Cancer Society, 2004), and evidence indicates that the risk for this cancer declines as time since smoking cessation increases. Though the precise nature of the decline in risk with years of cessation is not

known, it appears that those who quit for more than 20 years are not at increased risk over nonsmokers. As smoking is presumed to account for approximately 25%–30% of all cases of pancreatic cancer in the United States, preventing smoking could hypothetically eliminate that same portion of the cases, though in the real world that proportion would depend on the success of smoking prevention and cessation programs. To prevent pancreatic cancer and other chronic disease as well, individuals should maintain a healthful weight throughout life; balance caloric intake with physical activity; eat five or more servings of a variety of vegetables and fruits each day; choose whole grains in preference to processed (refined) grains and sugars; limit consumption of red meats, especially high-fat and processed meats (American Cancer Society, 2004). Secondary prevention of pancreas cancer, through population screening and early detection are not currently feasible. Individuals at extremely high risk may benefit by screening with imaging modalities, such as endoscopic ultrasonography or endoscopic retrograde cholangiopancreatography (Goggins et al., 2000). New techniques, such as serial analysis of gene expression, cDNA and DNA microarrays, and proteomics may ultimately help in designing methods for early diagnosis of pancreas cancer, but various challenges must be overcome before these provide information useful in the clinic (Becich, 2000; Iacobuzio-Dohahue et al., 2003; Hermeking, 2003; Garber, 2004). Chemotherapies for various stages of the disease are being evaluated in clinical trials (McBride, 2004; Yang et al., 2005). Currently, surgery is available for only 10%–15% of patients. Palliative therapy and supportive end-of-life care should be available to all patients when needed (Alter, 1996).

FUTURE DIRECTIONS Our growing knowledge of molecularly defined subgroups of cancer and precursor lesions can be employed in epidemiology to discover more about the etiology of pancreatic cancer (Bardeesy and DePinho, 2002). Use of serial analysis of gene expression, cDNA and DNA microarray technologies, and proteomics methods are adding to our understanding of the alterations that occur in carcinogenesis—not just in individual genes and proteins, but in cell regulatory pathways critical to cancer development. These technologies have the potential to increase our understanding of pancreas cancer etiology and enable us to diagnose the disease early enough to intervene with molecularly targeted therapies; though there is the promise of benefit, more research is desperately needed. A challenge for these studies is the acquisition of tumor tissue from patients with pancreatic cancer and precursor lesions. Molecular genetic research has pinpointed genes that are frequently altered in this cancer, through mutation or epigenetic events such as hypermethylation or hypomethylation of promoter regions of the genome. Further studies to elucidate the normal function and the consequences of somatic and germ-line mutations in genes such as K-ras, p16, p53, DPC4/SMAD4, and the pathways damaged by these mutations will reveal insights into the pathogenesis of pancreatic cancer. This understanding may provide strategies for prevention and possible targets for growth inhibition of neoplasia (Hruban et al., 2000; Bardeesy and DePinho, 2002). The accuracy of diagnosis remains a significant methodologic issue for epidemiologic studies of pancreatic cancer in some parts of the world. Case-control studies that have used proxy respondents increase the chance of misclassifying exposures (Lyon et al., 1992a). Hospitalbased studies have the advantage of a higher proportion of direct vs. surrogate interviews, yet are burdened with other problems related to the selection of the control series. More population-based case-control studies, based on direct interviews with patients (and thus requiring interviews with cases as early as possible) may help to identity some pancreatic cancer risk factors. In epidemiologic studies, the use of biomarkers may allow for more precise and specific measures of factors that modify the carcinogenic

Cancer of the Pancreas process—either positively or negatively (Schatzkin and Kipnis, 2004; Perera and Weinstein, 2000; Marshall, 2003). To establish clearly the gene-environment interactions that characterize most cancers, large cohort studies with baseline blood samples would be ideal. Large numbers of pancreatic cancer patients will be required to detect modest RRs for interactions between suspected risk factors and genetic polymorphisms (Vineis et al., 2004). New methodologies that deal with measures of uncertainty will receive more attention so that epidemiologists can provide the most valid interpretation of their findings. These include methods that deal with full and proper disclosure of uncertainty in study results; identification and control of confounding; and quantifying the effect of random error on study results (Maldonado and Phillips, 2004). Epidemiologists will also need new statistical tools to analyze large amounts of information such as the relationship between multiple genetic loci and disease susceptibility, and molecular pathology and disease progression (Maojo and Martin-Sanchez, 2004; Becich, 2000). Finally, multidisciplinary collaborations involving individuals in basic science, medicine, and public health will be needed to provide the range of skills and knowledge required to control this formidable disease. References Adlercruetz H. 1990. Western diet and western diseases: some hormonal and biochemical mechanisms and associations. Scan J Clin Lab Invest 50 (Suppl 201):3–23. Adlercruetz H. 2002. Phyto-oestrogens and cancer. Lancet Oncol 3:364–373. Alderson M. 1974. Mortality from malignant disease in patients with asthma. Lancet 2:1475–1477. Alguacil J, Kauppinen T, Porta M, et al. 2000a. Risk of pancreatic cancer and occupational exposures in Spain. PANKRAS II Study Group. Ann Occup Hyg 44:391–403. Alguacil J, Pollan M, Gustafsson P. 2003. Occupations with increased risk of pancreatic cancer in the Swedish population. Occup Environ Med 60:570–576. Alguacil J, Porta M, Benevides F, et al. 2000b. Occupation and pancreatic cancer in Spain: A case-control study based on job titles. Int J Epidemiol 29:1004–1013. Alguacil J, Silverman DT. 2004. Smokeless and other noncigarette tobacco use and pancreatic cancer: A case-control study based on direct interviews. Cancer Epidemiol Biomarkers Prev 13:55–58. Alomoquera C, Shibata D, Forrester K, et al. 1988. Most human carcinomas of the exocrine pancreas contain mutant c-K-ras genes. Cell 53:549–554. Alter CL. 1996. Palliative and supportive care of patients with pancreatic cancer. Semin Oncol 23:229–240. Ammann RW, Akovbiamtz A, Largiader F, et al. 1984. Course and outcome of chronic pancreatitis: Longitudinal study of a mixed medical-surgical series of 245 patients. Gastroenterology 86:820–828. Ammann RW, Knoblauch M, Möhr P, et al. 1980. High incidence of extrapancreatic carcinoma in chronic pancreatitis. Scand J Gastroenterol 15:395–399. Ammann RW, Scheuler G. 1984. Chronic pancreatitis, pancreatic cancer, alcohol, and smoking (letter). Gastroenterology 87:744–745. Anderson KE, Hammons GJ, Kadlubar FF, et al. 1997. Metabolic activation of aromatic amines by human pancreas. Carcinogenesis 18:1085–1092. Anderson KE, Johnson TW, Lazovich D, et al. 2002b. Association between nonsteroidal anti-inflammatory drug use and the incidence of pancreatic cancer. J Natl Cancer Inst 94:1168–1171. Anderson KE, Kadlubar FF, Kulldorff M, et al. 2005. Dietary intake of heterocyclic amines and benzo(a)pyrene: Associations with pancreatic cancer. Cancer Epidemiol Biomarkers Prev 14:2261–2265. Anderson KE, Sinha R, Kulldorff M, et al. 2002a. Meat intake and cooking techniques: Associations with pancreatic cancer. Mutat Res 506–507: 225–231. Andren-Sandberg A, Hoem D, Backman PL. 1999. Other risk factors for pancreatic cancer: Hormonal aspects. Ann Oncol 10 Suppl 4:131–135. Anonymous. 1976. [Yrke og dodelighed 1970–73]. Oslo: Central Bureau of Statistics. Aoli K, Hayakawa N, Kurihara M, et al, editors. 1992. Death Rates for Malignant Neoplasms for Selected Sites by Sex and Five-Year Age Group in 33 Countries 1953–57 to 1983–87. Nagoya, Japan: The University of Nagoya Coop Press. Arena V, Sussman N, Redman C, et al. 1998. Using alternative comparison populations to assess occupation-related mortality risk: Results for the high nickel alloys workers cohort. Occup Environ Med 40:907–916.

753

Armstrong B, Doll R. 1975. Environmental factors and cancer incidence and mortality in different countries with special reference to dietary practises. Int J Cancer 15:617–631. Augustin LS, Dal Maso L, La Vecchia C, et al. 2001. Dietary glycemic index and glycemic load, and breast cancer risk: A case-control study. Ann Oncol 12:1533–1538. Baghurst PA, McMichael AJ, Slavotinek AH, et al. 1991. A case-control study of diet and cancer of the pancreas. Am J Epidemiol 134:167–179. Bansal P, Sonnenberg A. 1995. Pancreatitis is a risk factor for pancreatic cancer. Gastroenterology 109:247–251. Bardeesy N, DePinho RA. 2002. Pancreatic cancer biology and genetics. Nat Rev Cancer 2:897–909. Bardin J, Eisen E, Tolbert P, et al. 1976. Mortality studies of machining fulied exposure in the automobile industry V: A case-control study of pancreatic cancer. Am J Ind Med 32:240–247. Baron JA. 2003. Epidemiology of non-steroidal anti-inflammatory drugs and cancer. Prog Exp Tumor Res 37:1–24. Bartholomew LG, et al. 1958. Carcinoma of the pancreas associated with chronic relapsing pancreatitis. Gastroenterology 35:473–477. Bartsch H, Frank N. 1996. Blocking the endogenous formation of N-nitroso compounds and related carcinogens. IARC Sci Publ:189–201. Bartsch H, Malaveille C, Lowenfels AB, et al. 1998. Genetic polymorphism of N-acetyltransferases, glutathione S-transferase M1 and NAD(P)H:quinone oxidoreductase in relation to malignant and benign pancreatic disease risk. The International Pancreatic Disease Study Group. Eur J Cancer Prev 7:215–223. Batty GD, Shipley MJ, Marmot M, et al. 2004. Diabetes status and post-load plasma glucose concentration in relation to site-specific cancer mortality: findings from the original Whitehall study. Cancer Causes Control 15:873–881. Beard J, Sladen T, Morgan G, et al. 2003. Health impacts of pesticide exposure in a cohort of outdoor workers. Env Health Perspect 111:724–730. Becich MJ. 2000. The role of the pathologist as tissue refiner and data miner: The impact of functional genomics on the modern pathology laboratory and the critical roles of pathology informatics and bioinformatics. Mol Diagn 5:287–299. Beland FA, Kadlubar FF. 1985. Formation and persistence of arylamine DNA adducts in vivo. Environ Health Perspect 62:19–30. Bell ET. 1957. Carcinoma of the pancreas. Am J Pathol 33:499–523. Benson LO, Teta MJ. 1993. Mortality due to pancreatic and lymphopoietic cancers in chlorohydrin production workers. Br J Ind Med 50:710– 716. Benz C, Hollander C, Miller B. 1986. Endocrine-responsive pancreatic carcinoma: Steroid binding and cytotoxicity studies in human tumor cell lines. Cancer Res 46:2276–2281. Berk JE. 1941. Diagnosis of carcinoma of the pancreas. Arch Int Med 68:525–559. Berrington de Gonzalez A, Sweetland S, Spencer E. 2003. A meta-analysis of obesity and the risk of pancreatic cancer. Br J Cancer 89:519–523. Best EWR. 1966. A Canadian Study of Smoking and Health Ottawa, Canada: Dept Natl Health and Welfare. Biankin AV, Kench JG, Dijkman FP, et al. 2003. Molecular pathogenesis of precursor lesions of pancreatic ductal adenocarcinoma. Pathology 35:14–24. Birch JM, Alston RD, McNally RJ, et al. 2001. Relative frequency and morphology of cancers in carriers of germline TP53 mutations. Oncogene 20:4621–4628. Birt DF, Julius AD, Dwork E, et al. 1990. Comparison of the effects of dietary beef tallow and corn oil on pancreatic caarcinogenesis in the hamster model. Carcinogenesis 11:745–748. Birt DF, Pour PM, Nagel DL, et al. 1997. Dietary energy restriction does not inhibit pancreatic carcinogenesis by N-nitrosobis-2-(oxopropyl)amine in the Syrian hamster. Carcinogenesis 18:2107–2111. Blair A, Malker H, Cantor KP, et al. 1985. Cancer among farmers. A review. Scand J Work Environ Health 11:397–407. Bloomston M, Zervos EE, Rosemurgy AS, 2nd. 2002. Matrix metalloproteinases and their role in pancreatic cancer: a review of preclinical studies and clinical trials. Ann Surg Oncol 9:668–674. Blot WJ, Fraumeni JF, Jr., Stone BJ. 1978. Geographic correlates of pancreas cancer in the United States. Cancer 42:373–380. Blot WJ, Lanier A, Fraumeni JF, Jr., et al. 1975. Cancer mortality among Alaskan natives, 1960–69. JNCI 55:547–554. Boffetta P, Stellman SD, Garfinkel L. 1988. Diesel exhaust exposure and mortality among males in the American Cancer Society prospective study. Am J Ind Med 14:403–415. Borch K, Kullman E, Hallhagen S, et al. 1988. Increased incidence of pancreatic neoplasia in pernicious anaemia. World J Surg 12:866–870.

754

PART IV: CANCER BY TISSUE OF ORIGIN

Borman PC, Beckingham IJ. 2001. Chronic pancreatitis. Br Med J 322:660. Bottone FG, Jr., Martinez JM, Alston-Mills B, et al. 2004. Gene modulation by Cox-1 and Cox-2 specific inhibitors in human colorectal carcinoma cancer cells. Carcinogenesis 25:349–357. Bottone FG, Jr., Martinez JM, Collins JB, et al. 2003. Gene modulation by the cyclooxygenase inhibitor, sulindac sulfide, in human colorectal carcinoma cells: possible link to apoptosis. J Biol Chem 278:25790– 25801. Bouchardy C, Clavel F, La VC, et al. 1990. Alcohol, beer and cancer of the pancreas. Int J Cancer 45:842–846. Boyle P, Maisonneuve P, Bueno de Mesquita B, et al. 1996. Cigarette smoking and pancreas cancer: A case control study of the search programme of the IARC. Int J Cancer 67:63–71. Brand-Miller JC. 2003. Glycemic load and chronic disease. Nutr Rev 61:S49–55. Bross ID, Viadana E, Hooten L. 1978. Occupational cancer in men exposed to dust and other environmental hazards. Arch Environ Health 33:300–307. Brownson RC, Chang JC, Davis JR, et al. 1991. Physical activity on the job and cancer in Missouri. Am J Public Health 81:639–642. Bueno de Mesquita HB, Maisonneuve P, Moerman CJ, et al. 1991a. Life-time history of smoking and exocrine carcinoma of the pancreas: A population-based case-control study in The Netherlands. Int J Cancer 49: 816–822. Bueno de Mesquita HB, Maisonneuve P, Moerman CJ, et al. 1991b. Intake of foods and nutrients and exocrine carcinoma of the pancreas: A population-based case-control study in The Netherlands. Int J Cancer 48: 540–549. Bueno de Mesquita HB, Maisonneuve P, Moerman CJ, et al. 1992b. Aspects of medical history and exocrine carcinoma of the pancreas: A populationbased case-control study in The Netherlands. Int J Cancer 52:17–23. Bueno de Mesquita HB, Maisonneuve P, Moerman CJ, et al. 1992a. Lifetime consumption of alcoholic beverages, tea and coffee and exocrine carcinoma of the pancreas: A population-based case-control study in The Netherlands. Int J Cancer 50:514–522. Bueno de Mesquita HB, Maisonneuve P, Moerman CJ, et al. 1992c. Anthropometric and reproductive variables and exocrine carcinoma of the pancreas: A population-based case-control study in The Netherlands. Int J Cancer 52:24–29. Bueno de Mesquita HB, Moerman CJ, Runia S, et al. 1990. Are energy and energy-providing nutrients related to exocrine carcinoma of the pancreas? Int J Cancer 46:435–444. Burch GE, Ansari A. 1968. Chronic alcoholism and carcinoma of the pancreas. Arch Int Med 122:273–275. Burmeister LF. 1990. Cancer in Iowa farmers: Recent results. Am J Ind Med 18:295–301. Burney PGJ, Comstock GW, Morris JS. 1989. Serologic precursors of cancer: Serum micronutrients and the subsequent risk of pancreatic cancer. Am J Clin Nutr 49:895–900. Butcher NJ, Minchin RF, Kadlubar FF, et al. 1996. Uptake of the food-derived heterocyclic amine carcinogen 2-amino-1-methyl-6-phenylimidazo[4,5b]pyridine and its N-hydroxy metabolite into rat pancreatic acini and hepatocytes in vitro. Carcinogenesis 17:889–892. Calle EE, Kaaks R. 2004. Overweight, obesity and cancer: Epidemiological evidence and proposed mechanisms. Nat Rev Cancer 4:579–591. Calle EE, Murphy TK, Rodriguez C, et al. 1998. Diabetes mellitus and pancreatic cancer mortality in a prospective cohort of United States adults. Cancer Causes Control 9:403–410. Calle EE, Rodriguez C, Walker-Thurmond K, et al. 2003. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med 348:1625–1638. Calvert G, Ward E, Schnorr T, et al. 1998. Cancer risks among workers exposed to metalworking fluids: A systematic review. Am J Ind Med 33:282–292. Cantor KP, Booze CF, JR. 1991. Mortality among aerial applicators and flight instructors: A reprint. Arch Environ Health 46:110–116. Carstensen J, Bygren L, Hatschek T. 1990. Cancer incidence among Swedish brewery workers. Int J Cancer 45:393–396. Castleman B, Scully RE, McNeely BU. 1972. Case records of the Massachusetts General Hospital. N Engl J Med 286:1353–1359. Caygill CPG, Hill MJ, Hall N, et al. 1985. Gastric surgery as a risk factor in human carcinogenesis. Gastroenterology 88:1344. CDC. 2004. Cigarette smoking among adults—United States, 2002. MMWR 53:428–431. Cederlöf R, Friberg L, Hrubec Z, et al. 1975. The Relationship of Smoking and Some Social Covariables to Mortality and Cancer Morbidity Stockholm: The Karolinska Institute. Cerny WL, Mangold KA, Scarpelli DG. 1992. K-ras mutation is an early event in pancreatic duct carcinogenesis in the Syrian golden hamster. Cancer Res 52:4507–4513.

Chari ST, Mohan V, Pitchumoni CS, et al. 1994. Risk of pancreatic carcinoma in tropical calcifying pancreatitis: An epidemiologic study. Pancreas 9:62–66. Chatenoud L, Tavani A, La Vecchia C, et al. 1998. Whole grain food intake and cancer risk. Int J Cancer 77:24–28. Chow WH, Gridley G, Nyren O, et al. 1995. Risk of pancreatic cancer following diabetes mellitus: A nationwide cohort study in Sweden. J Natl Cancer Inst 87:930–931. Chu J, Lloyd FL, Trifan OC, et al. 2003. Potential involvement of the cyclooxygenase-2 pathway in the regulation of tumor-associated angiogenesis and growth in pancreatic cancer. Mol Cancer Ther 2:1–7. Clark CG, Mitchell PEG. 1961. Diabetes melitus and primary carcinoma of the pancreas. Br Med J 2:1259–1262. Clary T, Ritz B. 2003. Pancreatic cancer mortality and organocholorine pesticide exposure in California, 1989–1996. Am J Ind Med 43:306–313. Clavel F, Benhamou E, Auquier A, et al. 1989. Coffee, alcohol, smoking and cancer of the pancreas: A case-control study. Int J Cancer 43:17–21. Coggon D, Pannett B, Winter PD, et al. 1986. Mortality of workers exposed to 2 methyl-4-chlorophenoxyacetic acid. Scan J Work Environ Health 12:448–454. Cohart EM, Muller C. 1955. Socioeconomic distribution of cancer of the gastrointestinal tract in New Haven. Cancer 8:1126–1129. Cohen GF. 1965. Early diagnosis of pancreatic neoplasms in diabetics. Lancet 2:267–269. Coles B, Nowell SA, MacLeod SL, et al. 2001. The role of human glutathione S-transferases (hGSTs) in the detoxification of the food-derived carcinogen metabolite N-acetoxy-PhIP, and the effect of a polymorphism in hGSTA1 on colorectal cancer risk. Mutat Res 482:3–10. Collins J, Esmen N, Hall T. 2001. A review and meta-analysis of formaldehyde exposure and pancreatic cancer. Am J Ind Med 39:336–345. Committee BEIR. 1990. Health effects of exposure to low levels of ionizing radiation: BEIR V Washington DC: National Academy Press. Coogan PF, Rosenberg L, Palmer JR, et al. 2000. Nonsteroidal anti-inflammatory drugs and risk of digestive cancers at sites other than the large bowel. Cancer Epidemiol Biomarkers Prev 9:119–123. Corbishley TP, Iqbal MJ, Wilkinson ML, et al. 1986. Androgen receptor in human normal and malignant pancreatic tissue and cell lines. Cancer 57:1992–1995. Corrie P, Mayer A, Shaw J, et al. 2002. Phase II study to evaluate combining gemcitabine with flutamide in advanced pancreatic cancer patients. Br J Cancer 87:716–719. Coss A, Cantor KP, Reif JS, et al. 2004. Pancreatic cancer and drinking water and dietary sources of nitrate and nitrite. Am J Epidemiol 159:693–701. Costantini A, Paci E, Miligi L, et al. 1989. Cancer mortality among workers in the Tuscan tanning industry. Br J Ind Med 46:384–388. Coughlin SS, Calle EE, Patel AV, et al. 2000. Predictors of pancreatic cancer mortality among a large cohort of United States adults. Cancer Causes Control 11:915–923. Cowgill SM, Muscarella P. 2003. The genetics of pancreatic cancer. Am J Surg 186:279–286. Creagan ET, Fraumeni JF, JR. 1972. Cancer mortality among American Indians, 1950–67. J Natl Cancer Inst 49:959–967. Cross AJ, Pollock JR, Bingham SA. 2003. Haem, not protein or inorganic iron, is responsible for endogenous intestinal N-nitrosation arising from red meat. Cancer Res 63:2358–2360. Cubilla AL, Fitzgerald PJ. 1978. Pancreas cancer (non-endocrine): A review— part II. Clin Bull 8:143–155. Cuzick J, Babiker AG. 1989. Pancreatic cancer, alcohol, diabetes mellitus and gall-bladder disease. Int J Cancer 43:415–421. Dai Q, Zheng W, Ji BT, et al. 1995. Prior immunity-related medical conditions and pancreatic-cancer risk in Shanghai. Int J Cancer 63:337–340. DeCouflé P, Lloyd JW, Salvin LG. 1977. Mortality by cause among stationary engineers and stationary firemen. J Occup Med 19:679–682. DeCouflé P, Stanislawczyk K. 1977. A retrospective survey of cancer in relation to occupation Washington, DC: U.S. Government Printing Office. Delzell E, Louik C, Lewis J, et al. 1981. Mortality and cancer morbidity among workers in the rubber tire industry. Am J Ind Med 2:209–216. Delzell E, Monson RR. 1985. Mortality among rubber workers: IX. Curing workers. Am J Ind Med 8:537–544. Devesa SS, Silverman DT, Young JL, Jr., et al. 1987. Cancer incidence and mortality trends among whites in the United States, 1947–84. J Natl Cancer Inst 79:701–770. Divine BJ, Barron V. 1986. Texaco Mortality Study: II. Patterns of mortality among white males by specific job groups. Am J Ind Med 10:371– 381. Divine BJ, Barron V, Kaplan SD. 1985. Texaco mortality study. I. Mortality among refinery, petrochemical, and research workers. J Occup Med 27:445–447.

Cancer of the Pancreas Doll R, Peto R. 1976. Mortality in relation to smoking: 20 years’ observation on male British doctors. Br Med J 2:1525–1536. Doll R, Vessey MP, Beasley RW, et al. 1972. Mortality of gasworkers—final report of a prospective study. Br J Industr Med 29:394–406. Dörken VH. 1964. [Einige daten bei 280 Patienten mit Pankreaskrebs]. Gastroenterologia 102:47–77. Dorn HF, Cutler SJ. 1958. Morbidity from cancer in the United States, Part II: Trend in morbidity association with income and stage at diagnosis. Public Health Monogr 56. Duell EJ, Holly EA. 2005. Reproductive and menstrual risk factors for pancreatic cancer: A population-based study of San Francisco Bay Area Women. Am J Epidemiol 161:741–747. Duell EJ, Holly EA, Bracci PM, et al. 2002a. A population-based, case-control study of polymorphisms in carcinogen-metabolizing genes, smoking, and pancreatic adenocarcinoma risk. J Natl Cancer Inst 94:297–306. Duell EJ, Holly EA, Bracci PM, et al. 2002b. A population-based study of the Arg399Gln polymorphism in X-ray repair cross- complementing group 1 (XRCC1) and risk of pancreatic adenocarcinoma. Cancer Res 62: 4630–4636. Dufour MC, Adamson MD. 2003. The epidemiology of alcohol-induced pancreatitis. Pancreas 27:286–290. Durbec JP, Chevillotte G, Bidart JM, et al. 1983. Diet, alcohol, tobacco and risk of cancer of the pancreas: A case-control study. Br J Cancer 47:463–470. Ebbehøj N, Friis j, Svendsen LB, et al. 1985. Indomethacin treatment of acute pancreatitis: a controlled double-blind trial. Scand J Gastroenterol 20:798–800. Eberle MA, Pfutzer R, Pogue-Geile KL, et al. 2002. A new susceptibility locus for autosomal dominant pancreatic cancer maps to chromosome 4q32–34. Am J Hum Genet 70:1044–1048. Edling C, Kling H, Flodin U, et al. 1986. Cancer mortality among leather tanners. Br J Ind Med 43:494–496. Eide TJ, Viste A, Andersen A, et al. 1991. The risk of cancer at all sites following gastric operation for benign disease: A cohort study of 4,224 patients. Int J Cancer 48:333–339. Eisen EA, Tolbert PE, Monson RR, et al. 1992. Mortality studies of machining fluid exposure in the automobile industry. l A standardized mortality ratio analysis. Am J Ind Med 22:809–824. Ekbom A, McLaughlin JK, Karlsson BM, et al. 1994. Pancreatitis and pancreatic cancer: A population-based study. J Natl Cancer Inst 86: 625–627. Ekbom A, Yuen J, Karlsson BM, et al. 1996. Risk of pancreatic and periampullar cancer following cholecystectomy: A population-based cohort study. Dig Dis Sci 41:387–391. Ellinger F, Landsman H. 1944. Frequency and course of cancer in diabetics. New York State J Med 44:259–265. Enstrom JE. 1978. Cancer and total mortality among active Mormons. Cancer 42:1943–1951. Enstrom JE. 1980. Banbury Report 4. Cancer Incidence in Defined populations, Health and dietary practices and cancer mortality among California Mormons. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY. Enterline P, Hartley J, Henderson V. 1987. Asbestos and cancer: A cohort followed up to death. Brit J Ind Med 44:396–401. Everhart J, Wright D. 1995. Diabetes mellitus as a risk factor for pancreatic cancer. A meta-analysis. J Am Med Assoc 273:1605–1609. Falk RT, Pickle LW, Fontham ET, et al. 1990. Occupation and pancreatic cancer risk in Louisiana. Am J Ind Med 18:565–576. Falk RT, Pickle LW, Fontham ET, et al. 1988. Life-style risk factors for pancreatic cancer in Louisiana: A case-control study. Am J Epidemiol 128:324–336. Farrow B, Evers BM. 2002. Inflammation and the development of pancreatic cancer. Surg Oncol 10:153–169. Farrow DC, Davis S. 1990a. Risk of pancreatic cancer in relation to medical history and the use of tobacco, alcohol and coffee. Int J Cancer 45:816–820. Farrow DC, Davis S. 1990b. Diet and the risk of pancreatic cancer in men. Am J Epidemiol 132:423–431. Fearon ER, Vogelstein B. 1990. A genetic model for colorectal tumorigenesis. Cell 61:759–767. Feinstein AR, Sosin DM, Wells CK. 1985. The Will Rogers phenomenon. Stage migration and new diagnostic techniques as a source of misleading statistics for survival in cancer. N Engl J Med 312:1604–1608. Felton JS, Wu R, Knize MG, et al. 1995. Heterocyclic amine mutagenicity/ carcinogenicity: influence of repair, metabolism, and structure. Princess Takamatsu Symp 23:50–58. Ferlay J, Bray F, Pisani P, et al. 2004. Globocan 2002: Cancer Incidence, Mortality and Prevalence Worldwide, Version 2.0. IARC Cancer Base No. 5. Lyon, IARC Press, 2004: WHO IARC.

755

Fernandez E, La Vecchia C, D’Avanzo B, et al. 1995b. Menstrual and reproductive factors and pancreatic cancer risk in women. Int J Cancer 62:11–14. Fernandez E, La Vecchia C, Decarli A. 1996. Attributable risks for pancreatic cancer in northern Italy. Cancer Epidemiol Biomarkers Prev 5:23–27. Fernandez E, La Vecchia C, Porta M, et al. 1995a. Pancreatitis and the risk of pancreatic cancer. Pancreas 11:185–189. Ferraroni M, Negri E, La Vecchia C, et al. 1989. Socioeconomic indicators, tobacco and alcohol in the aetiology of digestive tract neoplasms. Int J Epidemiol 18:556–562. Fisher WE, Boros LG, Schirmer WJ. 1996. Insulin promotes pancreatic cancer: Evidence for endocrine influence on exocrine pancreatic tumors. J Surg Res 63:310–313. Fontham ETH, Correa P. 1989. Epidemiology of pancreatic cancer. Surg Clin North Am 69:551–567. Franceschi S, Barbone F, Bidoli E, et al. 1993. Cancer risk in farmers: Results from a multi-site case-(hospital) control study in North-eastern Italy. Int J Cancer 53:740–745. Fretland AJ, Devanaboyina US, Doll MA, et al. 2003. Metabolic activation of 2-hydroxyamino-1-methyl-6-phenylimidazo[4,5-b]pyridine in Syrian hamsters congenic at the N-acetyltransferase 2 (NAT2) locus. Toxicol Sci 74:253–259. Friedman GD, Van den Eeden SK. 1993. Risk factors for pancreatic cancer: An exploratory study. Int J Epidemiol 22:30–37. Freiss H, Guo XZ, Nan BC, et al. 1999. Growth factors and cytokines in pancreatic carcinogenesis. Ann N Y Acad Sci 880:110–121. Friis S, Sorensen HT, McLaughlin JK, et al. 2003. A population-based cohort study of the risk of colorectal and other cancers among users of low-dose aspirin. Br J Cancer 88:684–688. Fuchs CS, Colditz GA, Stampfer MJ, et al. 1996. A prospective study of cigarette smoking and the risk of pancreatic cancer. Arch Intern Med 156:2255–2260. Gambill EE. 1971. Pancreatitis associated with pancreatic carcinoma. A study of 26 cases. Mayo Clin Proc 46:174–177. Gapstur SM, Gann PH, Lowe W, et al. 2000. Abnormal glucose metabolism and pancreatic cancer mortality. J Am Med Assoc 283:2552–2558. Garabrant DH, Held J, Langholz B, et al. 1992. DDT and related compounds and risk of pancreatic cancer. J Natl Cancer Inst 84:764–771. Garber K. 2004. Debate rages over proteomic patterns. J Natl Cancer Inst 96:816–818. Garfinkel L, Stellman SD. 1988. Mortality by relative weight and exercise. Cancer 62:1844–1850. Ghadirian P, Baillargeon J, Simard A, et al. 1995. Food habits and pancreatic cancer: A case-control study of the Francophone community in Montreal, Canada. Cancer Epidemiol Biomarkers Prev 4:895–899. Ghadirian P, Simard A, Baillargeon J. 1991a. Tobacco, alcohol, and coffee and cancer of the pancreas. A population-based, case-control study in Quebec, Canada. Cancer 67:2664–2670. Ghadirian P, Simard A, Baillargeon J, et al. 1991b. Nutritional factors and pancreatic cancer in the francophone community in Montreal, Canada. Int J Cancer 47:1–6. Ghadirian P, Thouez JP, Petitclerc C. 1991c. International comparisons of nutrition and mortality from pancreatic cancer. Cancer Detect Prev 15:357–362. Giardiello FM, Brensinger JD, Tersmette AC, et al. 2000. Very high risk of cancer in familial Peutz-Jeghers syndrome. Gastroenterology 119: 1447–1453. Gilbert ES, Marks S. 1979. An analysis of the mortality of workers in a nuclear facility. Radiat Res 79:122–128. Goggins M, Canto M, Hruban R. 2000. Can we screen high-risk individuals to detect early pancreatic carcinoma? J Surg Oncol 74:243–248. Gold EB, Cameron JL. 1993. Chronic pancreatitis and pancreatic cancer. N Engl J Med 328:1485–1486. Gold EB, Gordis L, Diener MD, et al. 1985. Diet and other risk factors for cancer of the pancreas. Cancer 55:460–467. Goldstein AM, Fraser MC, Struewing JP, et al. 1995. Increased risk of pancreatic cancer in melanoma-prone kindreds with p16INK4 mutations. N Engl J Med 333:970–974. Goodpaster BH, Katsiaras A, Kelley DE. 2003. Enhanced fat oxidation through physical activity is associated with improvements in insulin sensitivity in obesity. Diabetes 52:2191–2197. Gordis L. 1990. Consumption of methylxanthine-containing beverages and risk of pancreatic cancer. Cancer Lett 52:1–12. Graham S, Levin M, Lilienfeld AM. 1960. The socioeconomic distribution of cancer of various sites in Buffalo, NY: 1948–1952. Cancer 13:180– 190. Green A, Jensen OM. 1985. Frequency of cancer among insulin-treated diabetic patients in Denmark. Diabetologia 28:128–130.

756

PART IV: CANCER BY TISSUE OF ORIGIN

Green RC, Jr., Baggenstoss AH, Sprague RG. 1958. Diabetes mellitus in association with primary carcinoma of the pancreas. Diabetes 7:308– 311. Greenberg HL, Ott MG, Shore RE. 1990. Men assigned to ethylene oxide production or other ethylene oxide related chemical manufacturing: A mortality study. Brit J Ind Med 47:221–230. Greenwald P, Korns RF, Nasca PL, et al. 1975. Cancer in United States Jews. Cancer Res 35:3507–3512. Greenway B, Duke D, Pym B, et al. 1982. The control of human pancreatic adenocarcinoma xenogratfts in nude mice by hormone therapy. Br J Surg 69:595–597. Greenway B, Iqbal MJ, Johnson PJ, et al. 1981. Oestrogen receptor proteins in malignant and fetal pancreas. Br Med J (Clin Res Ed) 283:751–753. Gridley G, McLaughlin JK, Ekbom A, et al. 1993. Incidence of cancer among patients with rheumatoid arthritis. J Natl Cancer Inst 85:307–311. Grosfeld JL, Vane DW, Rescorla FJ, et al. 1990. Pancreatic tumors in childhood: Analysis of 13 cases. J Pediatr Surg 25:1057–1062. Gullo L, Pezzilli R, Morselli-Labate AM. 1995. Coffee and cancer of the pancreas: an Italian multicenter study. The Italian Pancreatic Cancer Study Group. Pancreas 11:223–229. Guralnick L. 1963. Mortality by Occupation Level and Cause of Death Among Men 20 to 64 Years of Age: U.S., 1950. Vital Statistics Special Report 53 Washington, DC: U.S. Government Printing Office. 439–446. Haddock G, Carter DC. 1990. Aetiology of pancreatic cancer. Br J Surg 77:1159–1166. Haenszel W. 1961. Cancer mortality among the foreign-born in the United States. J Natl Cancer Inst 26:37–132. Haenszel W, Kurihara M. 1968. Studies of Japanese migrants. I. Mortality from cancer and other diseases among Japanese in the United States. J Natl Cancer Inst 40:43–68. Hahn SA, Schutte M, Hoque AT, et al. 1996. DPC4, a candidate tumor suppressor gene at human chromosome 18q21.1. Science 271:350– 353. Hakulinen T, Lehtimaki L, Lehtonen M, et al. 1974. Cancer morbidity among two male cohorts with increased alcohol consumption in Finland. J Natl Cancer Inst 52:1711–1714. Hammond EC. 1966. Smoking in relation to the death rates of one million men and women. Natl Cancer Inst Monogr 19:127–204. Hammond EC, Horn D. 1958. Smoking and death rates; report on forty-four months of follow-up of 187,783 men. II. Death rates by cause. J Am Med Assoc 166:1294–1308. Hanis NM, Holmes TM, Shallenberger LG, et al. 1982. Epidemiologic study of refinery and chemical plant workers. J Occup Med 24:203–212. Hanley AJ, Johnson KC, Villeneuve PJ, et al. 2001. Physical activity, anthropometric factors and risk of pancreatic cancer: Results from the Canadian enhanced cancer surveillance system. Int J Cancer 94:140–147. Hansel DE, Kern SE, Hruban RH. 2003. Molecular pathogenesis of pancreatic cancer. Annu Rev Genomics Hum Genet 4:237–256. Hansel DE, Kern SK, Hruban RH. 2003. Molecular pathogenesis of pancreatic cancer. Annu Rev Genomics Hum Genet 4:237–256. Hansen ES. 1989. Mortality of auto mechanics. A ten-year follow-up. Scand J Work Environ Health 15:43–46. Harnack LJ, Anderson KE, Zheng W, et al. 1997. Smoking, alcohol, coffee, and tea intake and incidence of cancer of the exocrine pancreas: The Iowa Women’s Health Study. Cancer Epidemiol Biomarkers Prev 6:1081–1086. Hearne FT, Grose F, Pifer JW, et al. 1987. Methylene chloride mortality study: Dose-response characterization and animal model comparison. J Occup Med 29:217–228. Hecht SS. 1998. Biochemistry, biology, and carcinogenicity of tobacco-specific N-nitrosamines. Chem Res Toxicol 11:559–603. Hecht SS. 2000. Inhibition of carcinogenesis by isothiocyanates. Drug Metab Rev 32:395–411. Hecht SS, Hoffmann D. 1991. Relevance to Human Cancer of N-Nitroso Compounds, Tobacco Smoke and Mycotoxins. IARC Sci Pub No 105, N-nitroso compounds and tobacco-induced cancers in man. International Agency for Research on Cancer, Lyon, France, 54–61. Henley SJ, Thun MJ, Chao A, et al. 2004. Association between exclusive pipe smoking and mortality from cancer and other diseases. J Natl Cancer Inst 96:853–861. Hennenberger P, Ferris B, Monson R. 1989. Mortality among pulp and paper workers in Berlin, New Hampshire. Brit J Ind Med 46:658–664. Hermeking H. 2003. Serial analysis of gene expression and cancer. Curr Opin Oncol 15:44–49. Heuch I, Kvale G, Jacobsen BK, et al. 1983. Use of alcohol, tobacco and coffee, and risk of pancreatic cancer. Br J Cancer 48:637–643. Hiatt RA, Klatsky AL, Armstrong MA. 1988. Pancreatic cancer, blood glucose and beverage consumption. Int J Cancer 41:794–797.

Hinds MW, Kolonel IN, Lee J, et al. 1980. Association between cancer incidence and alcohol/cigarette consumption among five ethnic groups in Hawaii. Br J Cancer 41:929–940. Hine RJ, Srivastava S, Milner JS, et al. 2003. Nutritional links to plausible mechanisms underlying pancreatic cancer: a conference report. Pancreas 27:356–366. Hirayama T. 1975. Prospective studies on cancer epidemiology based on census population in Japan Florence, Italy. 26–35 p. Hirayama T. 1981. A large-scale cohort study on the relationship between diet and selected cancers of digestive organs. In: Bruce WR, Correa P, Lipkin M, et al., eds. Gastrointestinal Cancer: Endogenous Factors, Banbury Report 7. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory, pp. 409–426. Hirayama T. 1988. Epidemiology of pancreatic cancer in Japan. Int J Pancreatol 3:S203–204. Hirayama T. 1989. Epidemiology of pancreatic cancer in Japan. Jpn J Clin Oncol 19:208–215. Holly EA, Eberle CA, Bracci PM. 2003. Prior history of allergies and pancreatic cancer in the San Francisco Bay area. Am J Epidemiol 158:432–441. Holmes GKT, Stokes PL, Sorahan TM, et al. 1976. Coeliac disease, gluten free diet, and malignancy. Gut 17:612–619. Howatson AG, Carter DC. 1985. Pancreatic carcinogensis—enhancement by cholecystokinin in the hamster–nitrosamine model. Br J Cancer 51:107–114. Howatson AG, Carter DC. 1985. Pancreatic carcinogenesis—enhancement by cholecystokinin in the hamster-nitrosamine model. Br J Cancer 51:107–114. Howe GR, Burch JD. 1996. Nutrition and pancreatic cancer. Cancer Causes Control 7:69–82. Howe GR, Ghadirian P, Bueno De Mesquita HB, et al. 1992. A collaborative case-control study of nutrient intake and pancreatic cancer within the search programme. Int J Cancer 51:365–372. Howe GR, Jain M, Burch JD, et al. 1991. Cigarette smoking and cancer of the pancreas: Evidence from a population-based case-control study in Toronto, Canada. Int J Cancer 47:323–328. Howe GR, Jain M, Miller AB. 1990. Dietary factors and risk of pancreatic cancer: Results of a Canadian population-based case-control study. Int J Cancer 45:604–608. Hruban RH, Goggins M, Parsons J, et al. 2000. Progression model for pancreatic cancer. Clin Cancer Res 6:2969–2972. Hruban RH, Iacobuzio-Donahue C, Wilentz RE, et al. 2001. Molecular pathology of pancreatic cancer. Cancer J 7:251–258. Hsieh CC, MacMahon B, Yen S, et al. 1986. Coffee and pancreatic cancer. (letter). N Engl J Med 315:587–589. Hsing AW, Hansson LE, McLaughlin JK, et al. 1993. Pernicious anemia and subsequent cancer. A population-based cohort study. Cancer 71:745–750. Hu W, Feng Z, Tang MS. 2003. Preferential carcinogen-DNA adduct formation at codons 12 and 14 in the human K-ras gene and their possible mechanisms. Biochemistry 42:10012–10023. Huberman E, Sachs L. 1977. DNA binding and its relationship to carcinogenesis by different polycyclic hydrocarbons. Int J Cancer 19:122–127. Hutchison GB, MacMahon B, Jablon S, et al. 1979. Review of report by Mancusco, Stewart and Kneale of radiation exposure of Hanford workers. Health Physics 37:207–220. Huxley R, Ansary-Moghaddam A, Berrington de Gonzalez A, et al. 2005. TypeII diabetes and pancreatic cancer: a meta-analysis of 36 studies. Br J Med 92:2076–2083. Hyvarinen H, Partanen S. 1987. Association of cholecystectomy with abdominal cancers. Hepatogastroenterology 34:280–284. IARC (International Agency for Research on Cancer). 1984. N. Nitroso Compounds: Occurrence, Biological, Effects, and Relevance to Human Cancer. Lyon: International Agency for Research on Cancer. IARC. 1986. Tobacco Smoking. International Agency for Research on Cancer. Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to Humans. Lyon: IARC. IARC. 1991a. Coffee, Tea, Mate, Methylxanthines and Methylglyoxal. Lyon: International Agency for Research on Cancer. IARC. 1991b. DDT and Associated Compounds. Lyon: International Agency for Research on Cancer. IARC. 2004. Tobacco Smoke and Involuntary Smoking. International Agency for Research on Cancer Monographs on the Evaluation of Carcinogenic Risks to Humans. Lyon: IARC. IARC. 1997. Non-steroidal Anti-inflammatory Drugs. IARC Handbooks of Cancer Prevention. Lyon: International Agency for Research on Cancer. Inoue M, Tajima K, Takezaki T, et al. 2003. Epidemiology of pancreatic cancer in Japan: A nested case-control study from the Hospital-based Epidemiologic Research Program at Aichi Cancer Center (HERPACC). Int J Epidemiol 32:257–262.

Cancer of the Pancreas Isaksson B, Jonsson F, Pedersen NL, et al. 2002. Lifestyle factors and pancreatic cancer risk: A cohort study from the Swedish Twin Registry. Int J Cancer 98:480–482. Ishii K, Nakamura K, Ozaki H, et al. 1968. Epidemiological problems of pancreas cancer. Jpn J Clin Med 26:1839–1842. Ishii K, Nakamura K, Takeuchi T, et al. 1973. Chronic calcifying pancreatatis and pancreatic carcinoma in Japan. Digestion 9:429–437. Isomaki HA, Hakulinen T, Joutsenlahti V. 1978. Excess risk of lymphomas, leukemia and myeloma in patients with rheumatoid arthritis. J Chron Dis 31:691–696. Jacobs EJ, Connell CJ, Rodriguez C, et al. 2004. Aspirin use and pancreatic cancer mortality in a large United States cohort. J Natl Cancer Inst 96:524–528. Jacobsen O, Winther OS, Nielsen NA. 1985. Pancreatic cancer in the Faroe Islands. Scand J Gastroenterol 20:1142–1146. Jaffee EM, Hruban RH, Canto M, et al. 2002. Focus on pancreas cancer. Cancer Cell 2:25–28. Jain M, Howe GR, St. Louis P, et al. 1991. Coffee and alcohol as determinants of risk of pancreas cancer: A case-control study from Toronto. Int J Cancer 47:384–389. Jappinen P, Hakulinen T, Pukkala E, et al. 1987. Cancer incidence of workers in the Finnish pulp and paper industry. Scan J Clin Lab Invest 13: 197–202. Jee SH, Ohrr H, Sull JW, et al. 2005. Fasting serum glucose level and cancer risk in Korean men and women. JAMA 293:194–202. Jemal A, Clegg LX, Ward E, et al. 2004. Annual report to the nation on the status of cancer, 1975–2001, with a special feature regarding survival. Cancer 101:3–27. Jemal A, Taylor M, Ward A, et al. 2005. Cancer statistics. CA Cancer J 55:10–30. Jensen OM. 1979. Cancer morbidity and causes of death among Danish brewery workers. Int J Cancer 23:454–463. Ji B-T, Silverman D, Dosemeci M, et al. 1999. Occupation and pancreatic cancer risk in Shanghai, China. Am J Ind Med 35:70–81. Ji B-T, Chow WH, Dai Q, et al. 1995a. Cigarette smoking and alcohol consumption and the risk of pancreatic cancer: A case-control study in Shanghai, China. Cancer Causes Control 6:369–376. Ji B-T, Chow WH, Gridley G, et al. 1995b. Dietary factors and the risk of pancreatic cancer: A case-control study in Shanghai China. Cancer Epidemiol Biomarkers Prev 4:885–893. Ji B-T, Chow WH, Hsing AW, et al. 1997. Green tea consumption and the risk of pancreatic and colorectal cancers. Int J Cancer 70:255–258. Ji B-T, Hatch MC, Chow WH, et al. 1996. Anthropometric and reproductive factors and the risk of pancreatic cancer: A case-control study in Shanghai, China. Int J Cancer 66:432–437. Ji B-T, Silverman DT, Stewart PA, et al. 2001. Occupational exposure to pesticides and pancreatic cancer. Am J Ind Med 39:92–99. Johansen C, Chow WH, Jorgensen T, et al. 1996. Risk of colorectal cancer and other cancers in patients with gall stones. Gut 39:439–443. Johnson JR, Zintel HA. 1963. Pancreatic calcification and cancer of the pancreas. Surg Gyn Obst 117:585–588. Johnson KJ, Anderson KE, Harnack L, et al. 2005. No association between dietary glycemic index or load and pancreatic cancer incidence in postmenopausal women. Cancer Epidemiol Biomarkers Prev 14:1574–1575. Kaderlik KR, Lin D-X, Friesen MD, et al. 1993. Human biomonitoring of arylamine-DNA adducts in colon and pancreas tissues. Proc Am Assoc Cancer Research 34:152. Kaderlik KR, Minchin RF, Mulder GJ, et al. 1994. Metabolic activation pathway for the formation of DNA adducts of the carcinogen 2-amino-1methyl-6-phenylimidazo[4,5-b]pyridine (PhIP) in rat extrahepatic tissues. Carcinogenesis 15:1703–1709. Kadlubar FF, Butler MA, Kaderlik KR, et al. 1992. Polymorphisms for aromatic amine metabolism in humans: Relevance for human carcinogenesis. Environ Health Perspec 98:69–74. Kadlubar FF, Kaderlik KR, Lin D-X, et al. 1993. Detection of aromatic amineDNA adducts in human populations. Proc Am Assoc Cancer Research 34:606. Kahn HA. 1966. The Dorn study of smoking and mortality among U.S. veterans: Report on eight and one-half years of observation. Natl Cancer Inst Monogr 19:1–125. Kalapothaki V, Tzonou A, Hsieh CC, et al. 1993b. Nutrient intake and cancer of the pancreas: A case-control study in Athens, Greece. Cancer Causes Control 4:383–389. Kalapothaki V, Tzonou A, Hsieh CC, et al. 1993a. Tobacco, ethanol, coffee, pancreatitis, diabetes mellitus, and cholelithiasis as risk factors for pancreatic carcinoma. Cancer Causes Control 4:375–382. Kaplan SD. 1986. Update of a mortality study of workers in petroleum refineries. J Occup Med 28:514–516.

757

Karlson BM, Wuu J, Hsieh CC, et al. 1998. Parity and the risk of pancreatic cancer: A nested case-control study. Int J Cancer 77:224–227. Karmody AJ, Kyle J. 1969. The association between carcinoma of the pancreas and diabetes mellitus. Br J Sur 56:362–364. Kattwinkel J, Lopey A, DiSant-Agnese PA, et al. 1973. Hereditary pancreatitis: Three new kindreds and a critical review of the literature. Pediatrics 51:55–69. Kauppinen T, Partanen T, Degerth R, et al. 1995. Pancreatic cancer and occupational exposures. Epidemiology 6:498–502. Kawachi I, Pearce N, Fraser J. 1989. A New Zealand cancer registry-based study of cancer in wood workers. Cancer 64:2609–2613. Kazerouni N, Sinha R, Hsu CH, et al. 2001. Analysis of 200 food items for benzo[a]pyrene and estimation of its intake in an epidemiologic study. Food Chem Toxicol 39:423–436. Kernan G, Ji B-T, Dosemeci M, et al. 1999. Occupational risk factors for pancreastic cancer: A case-control study based on death certificates from 24 U.S. States. Am J Ind Med 36:260–270. Kessler II. 1970. Cancer mortality among diabetes. J Natl Cancer Inst 44: 673–686. Kiehne K, Otte JM, Folsch UR, et al. 2001. Growth factors in development and diseases of the exocrine pancreas. Pancreatology 1:15–23. King H, Diamond E, Bailar JC, III. 1965. Cancer mortality and religious preference: A suggested method in research. Milbank Mem Fund Quart 43:349–357. Kinlen LJ, McPherson K. 1984. Pancreas cancer and coffee and tea consumption: A case-control study. Br J Cancer 49:93–96. Klatsky AL, Friedman GD, Siegelaub AB. 1981. Alcohol and mortality: A ten-year Kaiser-Permanente experience. Ann Intern Med 95:139– 145. Klein AP, Beaty TH, Bailey-Wilson JE, et al. 2002. Evidence for a major gene influencing risk of pancreatic cancer. Genet Epidemiol 23:133–149. Kleiner HE, Reed MJ, Digiovanni J. 2003. Naturally occurring coumarins inhibit human cytochromes P450 and block benzo[a]pyrene and 7,12dimethylbenz[a]anthracene DNA adduct formation in MCF-7 cells. Chem Res Toxicol 16:415–422. Klöppel G, Maillet B. 1991. Histological typing of pancreatic and periampullary carcinoma. Eur J Surg Oncol 17:139–152. Knize MG, Salmon CP, Pais P, et al. 1999. Food heating and the formation of heterocyclic aromatic amine and polycyclic aromatic hydrocarbon mutagens/carcinogens. Adv Exp Med Biol 459:179–193. Kokawa A, Kondo H, Gotoda T, et al. 2001. Increased expression of cyclooxygenase-2 in human pancreatic neoplasms and potential for chemoprevention by cyclooxygenase inhibitors. Cancer 91:333–338. Kono S, Ikeda M. 1979. Correlation between cancer mortality and alcoholic beverage in Japan. Br J Cancer 40:449–455. Korc M. 1990. Potential role of the epidermal growth factor receptor in human pancreatic cancer. Int J Pancreatol 7:71–81. Korc M, Chandrasekar B, Yamanaka Y, et al. 1992. Overexpression of the epidermal growth factor receptor in human pancreatic cancer is associated with concomitant increases in the levels of epidermal growth factor and transforming growth factor alpha. J Clin Invest 90:1352–1360. Koshiba T, Hosotani R, Miyamoto Y, et al. 1999. Immunohistochemical analysis of cyclooxygenase-2 expression in pancreatic tumors. Int J Pancreatol 26:69–76. Krain LS. 1970. The rising incidence of carcinoma of the pancreas—real or apparent? J Surg Oncol 2:115–124. Krain LS. 1971a. The rising incidence of cancer of the pancreas. Further epidemiologic studies. J Chron Dis 23:685–690. Krain LS. 1971b. Cancer of the pancreas in California, 1942–1967. Calif Med 115:38–41. Kreiger N, Lacroix J, Sloan M. 2001. Hormonal factors and pancreatic cancer in women. Ann Epidemiol 11:563–567. Kumar V. 2004. Robbins and Cotran Pathologic Basis of Disease. 7th ed. Philadelphia: Elsevier Science. Kuratsune M, Kochi S, Horie A, et al. 1971. Test of alcoholic beverages and ethanol solutions for carcinogenicity and tumor promoting activity. Gann 62:395. Kuromatani N, Natori Y, Mizutani R, et al. 1999. A historical cohort mortality study of workers exposed to asbestos in a refitting shipyard. Industrial Health 3 7:9–17. Kvale G, Heuch I, Nilssen S. 1994. Parity in relation to mortality and cancer incidence: A prospective study of Norwegian women. Int J Epidemiol 23:691–699. La Vecchia C, Liati P, Decarli A, et al. 1987. Coffee consumption and risk of pancreatic cancer. Int J Cancer 40:309–313. La Vecchia C, Negri E, D’Avanzo B, et al. 1990. Medical history, diet and pancreatic cancer. Oncology 47:463–466.

758

PART IV: CANCER BY TISSUE OF ORIGIN

La Vecchia C, Negri E, Franceschi S, et al. 1993. Long-term impact of reproductive factors on cancer risk. Int J Cancer 53:215–219. Langman MJ, Cheng KK, Gilman EA, et al. 2000. Effect of anti-inflammatory drugs on overall risk of common cancer: Case-control study in general practice research database. BMJ 320:1642–1646. Lanier AP, Bender TR, Blot WJ, et al. 1976. Cancer incidence in Alaska natives. Int J Cancer 18:409–412. Layton DW, Bogen KT, Knize MG, et al. 1995. Cancer risk of heterocyclic amines in cooked foods: An analysis and implications for research. Carcinogenesis 16:39–52. Lea AJ. 1967. Neoplasms and environmental factors. Ann Royal Coll Surg Engl 41:432–438. Lee HC, Yoon YB, Kim CY. 1997. Association between genetic polymorphisms of the cytochromes P-450 (1A1, 2D6, and 2E1) and the susceptibility to pancreatic cancer. Korean J Intern Med 12:128–136. Lee IM, Sesso HD, Oguma Y, et al. 2003. Physical activity, body weight, and pancreatic cancer mortality. Br J Cancer 88:679–683. Lemoine NR, Lobresco M, Leung H, et al. 1992. The c-erbB-3 gene in human pancreatic cancer. J Pathol 168:269–273. Levi F, Lucchini F, Negri E, et al. 2003. Pancreatic cancer mortality in Europe: The leveling of an epidemic. Pancreas 27:139–142. Levin DL, Connelly RR. 1973. Cancer of the pancreas. Cancer 31:1231– 1236. Levin DL, Connelly RR, Devesa SS. 1981. Demographic characteristics of cancer of the pancreas: Mortality, incidence, and survival. Cancer 47:1456–1468. Levine W, Dyer AR, Shekelle RB, et al. 1990. Post-load plasma glucose and cancer mortality in middle-aged men and women. 12-year follow-up findings of the Chicago Heart Association Detection Project in Industry. Am J Epidemiol 131:254–262. Li K, Yu S. 2002. A nested case-control study on risk of pancreatic cancer among workers in the rubber industry. Pancreas 24:417–418. Lin DX, Thompson PA, Teitel C, et al. 2003. Direct reduction of N-acetoxyPhIP by tea polyphenols: A possible mechanism for chemoprevention against PhIP-DNA adduct formation. Mutat Res 523–524:193–200. Lin RS, Kessler II. 1981. A multifactorial model for pancreatic cancer in man. J Am Med Assoc 245:147–152. Lin Y, Tamakoshi A, Kawamura T, et al. 2002a. A prospective cohort study of cigarette smoking and pancreatic cancer in Japan. Cancer Causes Control 13:249–254. Lin Y, Tamakoshi A, Kawamura T, et al. 2002b. Risk of pancreatic cancer in relation to alcohol drinking, coffee consumption and medical history: Findings from the Japan collaborative cohort study for evaluation of cancer risk. Int J Cancer 99:742–746. Liu JM, Hankinson SE, Stampfer MJ, et al. 2003. Body iron stores and their determinants in healthy postmenopausal US women. Am J Clin Nutr 78:1160–1167. Lloyd JW, Decoufle P, Salvin LG. 1977. Unusual mortality experience of printing pressmen. J Occup Med 19:543–550. Logan WPD. 1982. Cancer Mortality by occupation and social class 1851–1971 London, Lyon: HM Stat Off, International Agency for Research on Cancer. Lohr M, Maisonneuve P, Lowenfels AB. 2000. K-Ras mutations and benign pancreatic disease. Int J Pancreatol 27:93–103. Longnecker D. 1990. Experimental pancreatic cancer: Role of species, sex and diet. Bull Cancer 77:27–37. Lossing EH, Best EWT, McGregor JT, et al. 1966. A Canadian study of smoking and Health Ottawa: Department of National Health and Welfare. Lowenfels AB. 1984. Chronic pancreatitis, pancreatic cancer, alcohol, and smoking (letter). Gastroenterology 87:744. Lowenfels AB, Maisonneuve P. 2002. Epidemiologic and etiologic factors of pancreatic cancer. Hematol Oncol Clin North Am 16:1–16. Lowenfels AB, Maisonneuve P, Cavallini G, et al. 1993. Pancreatitis and the risk of pancreatic cancer. New Engl J Med 328:1433–1437. Lowenfels AB, Maisonneuve P, Dimagno EP, et al. 1997. Hereditary pancreatitis and the risk of pancreatic cancer. International Hereditary Pancreatitis Study Group. J Natl Cancer Inst 89:442–446. Lowenfels AB, Maisonneuve P, Whitcomb DC, et al. 2001. Cigarette smoking as a risk factor for pancreatic cancer in patients with hereditary pancreatitis. J Am Med Assoc 286:169–170. Lowenfels AB, Patel VP, Pitchumoni CS. 1985. Chronic calcific pancreatitis and pancreatic cancer. Dig Dis Sci 30:982. Lundh G, Nordenstam H. 1970. Pancreas calcification and pancreas cancer. Acta Chir Scand 136:493–496. Lynch HT, Brand RE, Hogg D, et al. 2002. Phenotypic variation in eight extended CDKN2A germline mutation familial atypical multiple mole melanoma-pancreatic carcinoma-prone families: The familial atypical mole melanoma-pancreatic carcinoma syndrome. Cancer 94:84–96.

Lynch HT, Shaw TG, Lynch JF. 2004. Inherited predisposition to cancer: A historical overview. Am J Med Genet 129C:5–22. Lynge E, Thygesen L. 1990. Primary liver cancer among women in laundry and dry-cleaning work in Denmark. Scand J Work Environ Health 16:108–112. Lyon JL, Egger MJ, Robison LM, et al. 1992b. Misclassification of exposure in a case-control study: the effects of different types of exposure and different proxy respondents in a study of pancreatic cancer. Epidemiology 3:223–231. Lyon JL, Gardener JW, West DW. 1980. Banbury Report 4. Cancer Incidence in Defined Populations, Cancer risk and life-style: Cancer among Mormons from 1967 to 1975. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, 3–30. Lyon JL, Klauber MR, Gardner JW, et al. 1976. Cancer incidence in Mormons and non-Mormons in Utah, 1966–1970. N Engl J Med 294:129–133. Lyon JL, Mahoney AW, French TK, et al. 1992a. Coffee consumption and the risk of cancer of the exocrine pancreas: a case-control study in a low-risk population. Epidemiology 3:164–170. Lyon JL, Slattery ML, Mahoney AW, et al. 1993. Dietary intake as a risk factor for cancer of the exocrine pancreas. Cancer Epidemiol Biomarkers Prev 2:513–518. Mack TM. 2004. Cancers in the Urban Environment. Boston: Academic Press. Mack TM, Paganini-Hill A. 1981. Epidemiology of pancreas cancer in Los Angeles. Cancer 47:1474–1483. Mack TM, Peters JM, Yu MC, et al. 1985. Pancreas cancer is unrelated to the workplace in Los Angeles. Am J Ind Med 7:253–266. Mack TM, Yu MC, Hanisch R. 1986. Pancreas cancer and smoking, beverage consumption and past medical history. J Natl Cancer Inst 76:49–60. MacMahon B. 1960. The ethnic distribution of cancer mortality in New York City, 1955. Acta Un Int Contra Cancer 16:1716–1724. MacMahon B, Yen S, Trichopoulos C, et al. 1981. Coffee and cancer of the pancreas. N Engl J Med 304:630–633. Magnani C, Coggon D, Osmond C, et al. 1987. Occupation and five cancers: A case-control study using death certificates. Br J Ind Med 44:769–776. Malats N. 2001. Gene-environment interactions in pancreatic cancer. Pancreatology 1:472–476. Maldonado G, Phillips CV. 2004. Wishful thinking (editorial). Epidemiol Perspect Innov 1:2. Malka D, Hammel P, Maire F, et al. 2002. Risk of pancreatic adenocarcinoma in chronic pancreatitis. Gut 51:849–852. Mallin K, Rubin M, Joo E. 1989. Occupational cancer mortality in Illinois white and black males, 1979–1984 for seven cancer sites. Am J Ind Med 15:699–717. Mancuso TF, Stewart A, Kneale G. 1977. Radiation exposures of Hanford workers dying from cancer and other causes. Health Physics 33:369–385. Maojo V, Martin-Sanchez F. 2004. Bioinformatics: Towards new directions for public health. Methods Inf Med 43:208–214. Marble A. 1934. Diabetes and cancer. N Engl J Med 211:339–349. Maringhini A, Thiruvengadam R, Melton LJ, et al. 1987. Pancreatic cancer risk following gastric surgery. Cancer 60:245–247. Marshall JR. 2003. Methodologic and statistical considerations regarding use of biomarkers of nutritional exposure in epidemiology. J Nutr 133 Suppl 3:881S–887S. Maruchi N, Brian D, Ludwig J, et al. 1979. Cancer of the pancreas in Olmsted County, Minnesota, 1935–1974. Mayo Clin Proc 54:245–249. Matanoski GM, Selster R, Sartwell PE, et al. 1975. The current mortality rates of radiologists and other physician specialists: Specific causes of death. Am J Epidemiol 101:199–210. McBride G. 2004. Researchers optimistic about targeted drugs for pancreatic cancer. J Natl Cancer Inst 96:1570–1572. McCullough ML, Giovannucci EL. 2004. Diet and cancer prevention. Oncogene 23:6349–6364. McLean-Ross AH, Smith MA, Anderson JR, et al. 1982. Late mortality after surgery for peptic ulcer. N Engl J Med 307:519–522. McMichael AJ, McCall MG, Hartshorne JM, et al. 1980. Patterns of gastrointestinal cancer in European migrants to Australia: The role of dietary change. Int J Cancer 25:431–437. McMichael AJ, Spirtas R, Kupper LL. 1974. An epidemiologic study of mortality within a cohort of rubber workers, 1964–72. Occup Med 16:458–464. Meers PD. 1973. Allergy and cancer. The Lancet 1:884–885. Menezes RJ, Huber KR, Mahoney MC, et al. 2002. Regular use of aspirin and pancreatic cancer risk. BMC Public Health 2:18. Michaud DS, Giovannucci E, Willett WC, et al. 2003. Dietary meat, dairy products, fat, and cholesterol and pancreatic cancer risk in a prospective study. Am J Epidemiol 157:1115–1125. Michaud DS, Giovannucci E, Willett WC, et al. 2001. Physical activity, obesity, height, and the risk of pancreatic cancer. J Am Med Assoc 286:921–929.

Cancer of the Pancreas Michaud DS, Liu S, Giovannucci E, et al. 2002. Dietary sugar, glycemic load, and pancreatic cancer risk in a prospective study. J Natl Cancer Inst 94:1293–1300. Mikal S, Campbell AJA. 1950. Carcinoma of the pancreas. Surgery 28: 963–969. Mikoczy Z, Schutz A, Stromberg U, et al. 1996. Cancer and specific occupational exposures in the Swedish leather tanning industry: A cohort based case-control study. Occup Environ Med 53:463–467. Milham S, Jr. 1976. Occupational Mortality in Washington State 1950–1971. Washington, DC: US Government Printing Office. Mills PK, Beeson L, Abbey DE, et al. 1988. Dietary habits and past medical history as related to fatal pancreas cancer risk among Adventists. Cancer 61:2578–2585. Mirvish SS. 1986. Effects of vitamins C and E on N-nitroso compound formation, carcinogenesis, and, cancer. Cancer 58 (Suppl 8):1842–1850. Mizuno S, Watanabe S, Nakamura K, et al. 1992. A multi-institute case-control study on the risk factors of developing pancreatic cancer. Jpn J Clin Oncol 22:286–291. Moertel CG, Ahmann DL, Taylor WF, et al. 1971. Aspirin and pancreatic cancer pain. Gastroenterology 60:552–553. Moldow RE, Connelly RR. 1968. Epidemiology of pancreatic cancer in Connecticut. Gastroenterology 55:667–686. Molina MA, Sitja-Arnau M, Lemoine MG, et al. 1999. Increased cyclooxygenase-2 expression in human pancreatic carcinomas and cell lines: growth inhibition by nonsteroidal anti-inflammatory drugs. Cancer Res 59: 4356–4362. Moller H, Mellemgaard A, Lindvig K, et al. 1994. Obesity and cancer risk: A Danish record-linkage study. Eur J Cancer 30A:344–350. Møller H, Toftgaard C. 1991. Cancer occurrence in a cohort of patinets surgically treated for peptic ulcer. Gut 32:740–744. Monson RR, Fine LA. 1978. Cancer mortality and morbidity among rubber workers. J Natl Cancer Inst 61:1047–1053. Monson RR, Lyon JL. 1975. Proportional mortality among alcoholics. Cancer 36:1077–1079. Moynan RW, Neerhout RC, Johnson TS. 1964. Pancreatic carcinoma in childhood: Case report and review. J Pediatrics 65:711–719. Muir C, Waterhouse J, Mack T, et al, editors. 1987. Cancer Incidence in Five Continents, volume V. Lyon, France: International Agency for Research on Cancer. Naderi A, Couch FJ. 2002. BRCA2 and pancreatic cancer. Int J Gastrointest Cancer 31:99–106. Nelson NA, Van Peenen PF, Blanchard AG. 1987. Mortality in a recent oil refinery cohort. J Occup Med 29:610–612. Newhouse ML, Mathews G, Sheikh K, et al. 1988. Mortality of workers at acetylene production plants. Br J Ind Med 45:63–69. Newill VA. 1961. Distribution of cancer mortality among ethnic subgroups of the white population of New York City. J Natl Cancer Inst 26:405–417. Nilsen TI, Vatten LJ. 2000. A prospective study of lifestyle factors and the risk of pancreatic cancer in Nord-Trondelag, Norway. Cancer Causes Control 11:645–652. Norell SE, Ahlbom A, Erwald R, et al. 1986a. Diet and pancreatic cancer: a case-control study. Am J Epidemiol 124:894–902. Norell S, Ahlbom A, Olin R, et al. 1986b. Occupational factors and pancreatic cancer. Br J Ind Med 43:775–778. Norell S, Ahlbom A, Erwald R, et al. 1986c. Diabetes, gall stone disease, and pancreatic cancer. Br J Cancer 54:377–378. Nöthlings U, Wilkens LR, Murphy SP, et al. 2005. Meat and fat intake as risk factors for pancreatic cancer: the multiethnic cohort study. J Natl Cancer Inst 97:1458–1465. O’Meara AT. 2002. Pancreatic cancer: Evidence-based diagnosis and treatment. Clin Obstet Gynecol 45:855–865. Ockenga J, Vogel A, Teich N, et al. 2003. UDP glucuronosyltransferase (UGT1A7) gene polymorphisms increase the risk of chronic pancreatitis and pancreatic cancer. Gastroenterology 124:1802–1808. Offerhaus JGA, Giardiello FM, Moore GW, et al. 1987. Partial gastrectromy: A risk factor for carcinoma of the pancreas? Human Pathol 18:285– 288. Ohba S, Nishi M, Miyake H. 1996. Eating habits and pancreas cancer. Int J Pancreatol 20:37–42. Ojajarvi A, Partanen T, Ahlbom A, et al. 2000. Occupational exposures and pancreatic cancer: A meta-analysis. Occup Environ Med 57:316– 324. Jajarvi A, Partanen T, Ahlbom A, et al. 2001. Risk of pancreatic cancer in workers exposed to chlorinated hydrocarbon solvents and related compounds: A meta-analysis. Am J Epidemiol 153:841–850. Olsen GW, Mandel JS, Gibson RW, et al. 1989. A case-control study of pancreatic cancer and cigarettes, alcohol, coffee and diet. Am J Publ Hlth 79:1016–1019.

759

Olsen GW, Mandel JS, Gibson RW, et al. 1991. Nutrients and pancreatic cancer: A population-based case-control study. Cancer Causes Control 2:291–297. Olsen JH, Jensen OM. 1987. Occupation and risk of cancer in Denmark: An analysis of 93,810 cancer cases, 1970–1979. Scan J Work Environ Health 13 (Suppl):7–91. OPCS (Office of Population Census Surveys). 1938. Occupational mortality: The Registrar General’s decennial supplement for England and Wales, 1931. London: HMSO (Her Majesty’s Stationery Office). OPCS. 1954. Occupational mortality: Registrar General’s Decennial Supplement for England and Wales, 1951. London: HMSO. OPCS. 1978. Occupational mortality: Registrar General’s Decennial Supplement for England and Wales, 1970–1972. London: HMSO. OPCS. 1978. Occupational mortality: The Registrar General’s decennial supplement for England and Wales, 1970–72 London: HMSO. Paffenbarger RJ, Hyde R, Wing A. 1987. Physical activity and incidence of cancer in diverse populations: a preliminary report. Am J Clin Nutr 45(Suppl):312–317. Pan SY, Johnson KC, Ugnat AM, et al. 2004. Association of obesity and cancer risk in Canada. Am J Epidemiol 159:259–268. Pannett B, Coggon D, Acheson E. 1985. A job-exposure matrix for use in population based studies in England and Wales. Brit J Ind Med 42:777–783. Pariza MW. 1987. Dietary fat, calorie restriction, ad libitum feeding and cancer risk. Nutr Rev 45:1–7. Parkin DM, Muir CS, Whelan SL, et al. 1992. Cancer Incidence in Five Continents, Vol VI. Lyon, France: International Agency for Research on Cancer. Parkin DM, Whelan SL, Ferlay J, et al, editors. 2002. Cancer Incidence in Five Continents, Vol VIII. Lyon, France: International Agency for Research on Cancer. Partanen T, Hemminki K, Vainio H, et al. 1995. Coffee consumption not associated with risk of pancreas cancer in Finland. Prev Med 24:213–216. Partanen T, Kauppinen T, Degerth R, et al. 1994. Pancreatic cancer in industrial branches and occupations in Finland. Am J Ind Med 25: 851–866. Perera FP, Weinstein IB. 2000. Molecular epidemiology: recent advances and future directions. Carcinogenesis 21:517–524. Perugini RA, McDade TP, Vittimberga FJ, Jr., et al. 1998. The molecular and cellular biology of pancreatic cancer. Crit Rev Eukaryot Gene Expr 8:377–393. Perugini RA, McDade TP, Vittimberga FJ, Jr., et al. 2000. Sodium salicylate inhibits proliferation and induces G1 cell cycle arrest in human pancreatic cancer cell lines. J Gastrointest Surg 4:24–32, discussion 32–33. Peters U, Sinha R, Bell DA, et al. 2004. Urinary mutagenesis and fried red meat intake: Influence of cooking temperature, phenotype, and genotype of metabolizing enzymes in a controlled feeding study. Environ Mol Mutagen 43:53–74. Peterson GR, Milham S, Jr. 1980. Occupational Mortality in California 1959–61 Washington, DC: U.S. Government Printing Office. Phillips AR, Lawes CM, Cooper GJ, et al. 2002. Ethnic disparity of pancreatic cancer in New Zealand. Int J Gastrointest Cancer 31:137–145. Phillips RL. 1975. Role of life-style and dietary habits in risk of cancer among Seventh-Day Adventists. Cancer Res 35:3513–3522. Phillips RL, Garfinkel L, Kuzma JW, et al. 1980. Mortality among California Seventh-Day Adventists for selected cancer sites. J Natl Cancer Inst 65:1097–1107. Pickle LW, Gottlieb MS. 1980. Pancreatic cancer mortality in Louisiana. Am J Public Health 70:256–259. Pietri F, Clavel F, Auquier A, et al. 1990. Occupational risk factors for cancer of the pancreas: A case-control study. Br J Ind Med 47:425–428. Polednak AP, Stehney AF, Lucas HF. 1983. Mortality among male workers at a thorium processing plant. Health Phys 44 (Suppl 1):239–251. Polissar L, Severson RK, Boatman ES. 1984. A case-control study of asbestos in drinking water and cancer risk. Am J Epidemiol 119:456–471. Potter JD. 2002. Pancreas cancer—we know about smoking, but do we know anything else? Am J Epidemiol 155:793–795, discussion 796–797. Potter JD, Steinmetz K. 1996. Vegetables, fruit and phytoestrogens as preventive agents. IARC Sci Pub l:61–90. Pour PM, Reber HA, Stepan K. 1983. Modification of pancreatic carcinogenesis in the hamster model. XII. Dose-related effect of ethanol. J Natl Cancer Inst 71:1085–1087. Puntoni R, Vercelli M, Merlo F, et al. 1979. Mortality among shipyard workers in Genoa, Italy. Ann NY Acad Sci 330:353–377. Rachet B, Partanen T, Kauppinen T, et al. 2000. Cancer risk in laboratory workers: An emphasis on biological research. Am J Ind Med 38:651–655. Ragozzino M, Melton LJ, Chu CP, et al. 1982. Subsequent cancer risk in the incidence cohort of Rochester, Minnesota, residents with diabetes mellitus. J Chron Dis 35:13–19.

760

PART IV: CANCER BY TISSUE OF ORIGIN

Rao CV, Reddy BS. 2004. NSAIDs and chemoprevention. Curr Cancer Drug Targets 4:29–42. Raymond L, Infante F, Tuyns AJ, et al. 1987. [Alimentation et cancer du pancreas.] Gatroenterol Clin Biol 11:488–492. Reddy SA. 2001. Signaling pathways in pancreatic cancer. Cancer J 7:274–286. Redmond CK, Strobino BR, Cypress RH. 1976. Cancer experience among coke by-product workers. Ann NY Acad Sci 271:102–115. Riela A, Zinsmeister AR, Melton LJ, 3rd, et al. 1992. Increasing incidence of pancreatic cancer among women in Olmsted County, Minnesota, 1940 through 1988. Mayo Clin Proc 67:839–845. Risch HA. 2003. Etiology of pancreatic cancer, with a hypothesis concerning the role of N-nitroso compounds and excess gastric acidity. J Natl Cancer Inst 95:948–960. Rix BA, Villadsen E, Lynge E. 1997. Cancer incidence of sulfite pulp workers in Denmark. Scand J Work Environ Health 23:458–461. Robinette CD, Fraumeni JF, Jr. 1978. Asthma and subsequent mortality in World War II veterans. J Chron Dis 31:619–624. Robinson AB, et al. 1970. The occurrence of carcinoma of the pancreas in chronic pancreatitis. Radiology 94:289–290. Robinson C, Waxweiler R, Fowler D. 1986. Mortality among production workers in pulp and paper mills. Scan J Clin Lab Invest 12:552–560. Robles-Diaz G, Duarte-Rojo A. 2001. Pancreas: A sex steroid-dependent tissue. Isr Med Assoc J 3:364–368. Rockette HE, Arena VC. 1983. Mortality studies of aluminium reduction plant workers: Portroom and carbon department. J Occup Med 25:549–557. Roebuck BD. 1992. Dietary fat and the development of pancreatic cancer. Lipids 27:804–806. Roebuck BD, Baumgartner KJ, MacMillan DL. 1993. Caloric restriction and intervention in pancreatic carcinogenesis in the rat. Cancer Res 53:46–52. Ronneberg A, Haldorsen T, Romundstad P, et al. 1999. Occupational exposure and cancer incidence among workers from an aluminum smelter in western Norway. Scand J Work Environ Health 25:207–214. Rose DP, Komninou D, Stephenson GD. 2004. Obesity, adipocytokines, and insulin resistance in breast cancer. Obes Rev 5:153–165. Rosty C, Geradts J, Sato N, et al. 2003. p16 Inactivation in pancreatic intraepithelial neoplasias (PanINs) arising in patients with chronic pancreatitis. Am J Surg Pathol 27:1495–1501. Rozenblum E, Schutte M, Goggins M, et al. 1997. Tumor-suppressive pathways in pancreatic carcinoma. Cancer Res 57:1731–1734. Rubio V, Rodriguez JI. 1989. Near-simultaneous adenocarcinoma of the pancreas in husband and wife. Lancet 1:166–167. Rushton L, Alderson MR. 1981. An epidemiological survey of eight oil refineries in Britain. Br J Ind Med 38:225–234. Saftlas AF, Blair A, Cantor KP, et al. 1987. Cancer and other causes of death among Wisconsin farmers. Am J Ind Med 11:119–129. Salmeron J, Manson JE, Stampfer MJ, et al. 1997. Dietary fiber, glycemic load, and risk of non-insulin-dependent diabetes mellitus in women. J Am Med Assoc 277:472–477. Samanic C, Gridley G, Chow WH, et al. 2004. Obesity and cancer risk among white and black United States veterans. Cancer Causes Control 15:35–43. Sarles H, Figarella C, Tiscornia O, et al. 1980. The Pancreas, Chronic calcifying pancreatitis (CCP), mechanism of formation of the lesions, New data and critical study. Williams and Wilkins, Baltimore, 48–66. Sarles H, Tiscornia O. 1974. Ethanol and chronic calcifying pancreatitis. Med Clin North Am 58:1333–1346. Schatzkin A, Kipnis V. 2004. Could exposure assessment problems give us wrong answers to nutrition and cancer questions? J Natl Cancer Inst 96:1564–1565. Schernhammer ES, Kang JH, Chan AT, et al. 2004. A prospective study of aspirin use and the risk of pancreatic cancer in women. J Natl Cancer Inst 96:22–28. Schlag P, Ulrich H, Nerkle P, et al. 1980. Are nitrite and N-nitroso compounds in gastric juice risk factors for carcinoma in the operated stomach? Lancet 1:727–729. Schneider MB, Matsuzaki H, Haorah J, et al. 2001. Prevention of pancreatic cancer induction in hamsters by metformin. Gastroenterology 120: 1263–1270. Schottenfeld D, Warshauer ME, Zauber AG, et al. 1981. Quantification of Occupational Cancer. Banbury Report. A prospective study of morbidity and mortality in petroleum industry employees in the United States. A preliminary report. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory, pp. 247–265. Schreinemachers DM, Everson RB. 1994. Aspirin use and lung, colon, and breast cancer incidence in a prospective study. Epidemiology 5:138– 146. Schwarte-Waldhoff I, Volpert OV, Bouck NP, et al. 2000. Smad4/DPC4mediated tumor suppression through suppression of angiogenesis. Proc Natl Acad Sci USA 97:9624–9629.

Schwartz GG, Reis IM. 2000. Is cadmium a cause of human pancreatic cancer? Cancer Epidemiol Biomarkers Prev 9:139–145. SEER. 2004a. Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database: Incidence–SEER 9 Regs Public-Use, Nov 2003 Sub (1973–2001), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2004, based on the November 2003 submission. SEER. 2004b. Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database: Mortality–All COD, Public-Use With State, Total U.S. (1969–2001), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2004. Underlying mortality data provided by NCHS (www.cdc.gov/nchs). . Segi M, Kurihara M. 1972. Cancer Mortality for Selected Sites in 24 Countries. No 6 (1966–1967). Nagoya, Japan: Japan Cancer Society. Seidman H. 1970. Cancer death rates by site and sex for religious and socioeconomic groups in New York City. Environ Res 3:234–250. Seilkop S. 2001. Occupational exposures and pancreatic cancer: A metaanalysis. Occup Environ Med 58:63–64. Selenskas S, Teta M, Vitale J. 1995. Pancreatic cancer among workers processing synthetic resins. Am J Ind Med 28:385–398. Selikoff IJ, Seidman H. 1981. Cancer of the pancreas among asbestos insulation workers. Cancer 47:1469–1473. Shannon H, Walsh C, Jadon N, et al. 1991. Mortality of 11,500 nickel workers—extended follow up and relationship to environmental conditions. Toxicol Indust Health 7:277–294. Shapiro S, Heinonen OP, Siskind V. 1971. Cancer and allergy. Cancer 28:396–400. Sheffet A, Thind I, Miller AM, et al. 1982. Cancer mortality in a pigment plant utilizing lead and zinc chromates. Arch Environ Health 37:44–52. Shibata A, Mack TM, Paganini-Hill A, et al. 1994. A prospective study of pancreatic cancer in the elderly. Int J Cancer 58:46–49. Shields JM, Pruitt K, McFall A, et al. 2000. Understanding Ras: ‘It ain’t over ’til it’s over’. Trends Cell Biol 10:147–154. Siemiatycki J. 1991. Risk factors for cancer in the workplace. Boca Raton: CRC Press. Silverman DT, Brown LM, Hoover RN, et al. 1995. Alcohol and pancreatic cancer in blacks and whites in the United States. Cancer Res 55: 4899–4905. Silverman DT, Dunn JA, Hoover RN, et al. 1994. Cigarette smoking and pancreas cancer: A case-control study based on direct interviews. J Natl Cancer Inst 86:1510–1516. Silverman DT, Hoover RN, Brown LM, et al. 2003. Why do Black Americans have a higher risk of pancreatic cancer than White Americans? Epidemiology 14:45–54. Silverman DT, Hoover RN, Swanson GM, et al. 1983. The prevalence of coffee drinking among hospitalized and population-based control groups. J Am Med Assoc 249:1877–1880. Silverman DT, Schiffman M, Everhart J, et al. 1999. Diabetes mellitus, other medical conditions and familial history of cancer as risk factors for pancreatic cancer. Br J Cancer 80:1830–1837. Silverman DT, Swanson CA, Gridley G, et al. 1998. Dietary and nutritional factors and pancreatic cancer: a case-control study based on direct interviews. J Natl Cancer Inst 90:1710–1719. Silverstein M, Park R, Marmor M, et al. 1988. Mortality among bearing plant workers exposed to metal-working fluids and abrasives. J Occup Med 30:706–714. Sinha R, Rothman N. 1999. Role of well-done, grilled red meat, heterocyclic amines (HCAs) in the etiology of human cancer. Cancer Lett 143: 189–194. Skinner HG, Michaud DS, Colditz GA, et al. 2003. Parity, reproductive factors, and the risk of pancreatic cancer in women. Cancer Epidemiol Biomarkers Prev 12:433–438. Slebos RJ, Hoppin JA, Tolbert PE, et al. 2000. K-ras and p53 in pancreatic cancer: association with medical history, histopathology, and environmental exposures in a population-based study. Cancer Epidemiol Biomarkers Prev 9:1223–1232. Smith GD, Egger M, Shipley MJ, et al. 1992. Post-challenge glucose concentration, impaired glucose tolerance, diabetes, and cancer mortality in men. Am J Epidemiol 136:1110–1114. Smith JP, Solomon TE, Bagheri S, et al. 1990. Cholecystokinin stimulates growth of human pancreatic adenocarcinoma SW-1990. Dig Dis Sci 35:1377–1384. Smith PG, Doll R. 1981. Mortality from cancer and all causes among British radiologists. Br J Radiol 54:187–194. Soh JW, Weinstein IB. 2003. Role of COX-independent targets of NSAIDs and related compounds in cancer prevention and treatment. Prog Exp Tumor Res 37:261–285.

Cancer of the Pancreas Solenova LG. 1992. [The cancer morbidity of workers in rubber manufacture.] Vopr Onkol 38:1174–1182. Soler M, Chatenoud L, La Vecchia C, et al. 1998. Diet, alcohol, coffee and pancreatic cancer: final results from an Italian study. Eur J Cancer Prev 7:455–460. Sorensen Ht, Friis S, Norgard B, et al. 2003. Risk of cancer in a large cohort of nonaspirin NSAID users: A population–based study. Br J Cancer 88:1687–1692. Sperti C, Bonadimani B, Guolo P, et al. 1992. Androgen profile in patients with pancreatic carcinoma. Ital J Gastroenterol 24:328–331. Standop J, Schneider MB, Ulrich A, et al. 2001. Experimental animal models in pancreatic carcinogenesis: lessons for human pancreatic cancer. Dig Dis 19:24–31. Steinkellner H, Rabot S, Freywald C, et al. 2001. Effects of cruciferous vegetables and their constituents on drug metabolizing enzymes involved in the bioactivation of DNA-reactive dietary carcinogens. Mutat Res 480–481:285–297. Steinmetz KA, Potter JD. 1996. Vegetables, fruit, and cancer prevention: A review. J Am Diet Assoc 96:1027–1039. Stocks P. 1970. Cancer mortality in relation to national consumption of cigarettes solid fuel, tea and coffee. Br J Cancer 24:215–225. Stolzenberg-Solomon RZ, Pietinen P, Barrett MJ, et al. 2001. Dietary and other methyl-group availability factors and pancreatic cancer risk in a cohort of male smokers. Am J Epidemiol 153:680–687. Stolzenberg-Solomon RZ, Pietinen P, Taylor PR, et al. 2002. Prospective study of diet and pancreatic cancer in male smokers. Am J Epidemiol 155:783–792. Sugimura T. 2000. Nutrition and dietary carcinogens. Carcinogenesis 21:387–395. Sugimura T, Nagao M, Wakabayashi K. 2000. How we should deal with unavoidable exposure of man to environmental mutagens: Cooked food mutagen discovery, facts and lessons for cancer prevention. Mutat Res 447:15–25. Swauger JE, Steichen TJ, Murphy PA, et al. 2002. An analysis of the mainstream smoke chemistry of samples of the U.S. cigarette market acquired between 1995 and 2000. Regul Toxicol Pharmacol 35:142–156. Tada M, Yososuka O, Omata M, et al. 1990. Analysis of ras gene mutations in biliary and pancreatic tumors by polymerase chain reaction and direct sequencing. Cancer 66:930–935. Takahashi M, Furukawa F, Toyoda K, et al. 1990. Effects of various prostaglandin synthesis inhibitors on pancreatic carcinogenesis in hamsters after initiation with N-nitrosobis(2-oxopropyl)amine. Carcinogenesis 11:393–395. Talamini G, Bassi C, Falconi M, et al. 1999. Alcohol and smoking as risk factors in chronic pancreatitis and pancreatic cancer. Dig Dis Sci 44:1303–1311. Tannenbaum A. 1959. The Physiopathology of Cancer, Nutrition and Cancer. New York: Hoeber-Harper, pp. 517–562. Tascilar M, Van Rees BP, Sturm PD, et al. 2002. Pancreatic cancer after remote peptic ulcer surgery. J Clin Pathol 55:340–345. Tavani A, La Vecchia C, Gallus S, et al. 2000. Red meat intake and cancer risk: A study in Italy. Int J Cancer 86:425–428. Tavani A, Pregnolato A, Negri E, et al. 1997. Alcohol consumption and risk of pancreatic cancer. Nutr Cancer 27:157–161. Taxy JB. 1976. Adenocarcinoma of the pancreas in childhood. Cancer 37:1508–1518. Tersmette AC, Petersen GM, Offerhaus GJ, et al. 2001. Increased risk of incident pancreatic cancer among first-degree relatives of patients with familial pancreatic cancer. Clin Cancer Res 7:738–744. Theriault G, Provencher S. 1987. Mortality study of oil refinery workers: Five year follow-up. J Occup Med 29:357–360. Thomas TL, Decouflé P, Moure-Eraso R. 1980. Mortality among workers employed in petroleum refining and petrochemical plants. J Occup Med 22:97–103. Thomas TL, Maxweiler RJ, Moure-Erason R, et al. 1982. Mortality patterns among workers in three Texas oil refineries. J Occup Med 24:135– 141. Thompson PA, Seyedi F, Lang NP, et al. 1999. Comparison of DNA adduct levels associated with exogenous and endogenous exposures in human pancreas in relation to metabolic genotype. Mutat Res 424:263–274. Tolbert P, Eisen E, Pothier L, et al. 1992. Mortality studies of machining fluid exposure in the automobile industry. II Risks associated with specific fluid types. Scand J Work Environ Health 18:351–360. Tolley HD, Marks S, Buchanan JA, et al. 1983. A further update of the analysis of the mortality of workers in a nuclear facility. Radiat Res 95:211–213. Toyokuni S. 2002. Iron and carcinogenesis: from Fenton reaction to target genes. Redox Rep 7:189–197.

761

Tricker AR, Preussmann R. 1991. Carcinogenic N-nitrosamines in the diet: Occurrence, formation, mechanisms and carcinogenic potential. Mutat Res 259:277–289. Tsukimoto I, Watanabe K, Lin JB, et al. 1973. Pancreatic carcinoma in children in Japan. Cancer 31:1203–1207. Tulassay Z, Sandor Z, Bodrogi L, et al. 1988. An endocrine marker for pancreatic cancer. BMJ 297:1447–1448. U.S. 1989. Reducing the Health Consequences of Smoking: 25 Years of Progress, A Report of the Surgeon General. USDHS, PHS, CDC, CCDPHP, Office on Smoking and Health. U.S. 1990. Cancer Statistics Review 1973–1987: USDHHS, PHS, NCI. U.S. 2004. National Toxicology Program, 11th report on carcinogens, 2004. Full report. Research Triangle Park, NC: USDHHS, PHS. Ushiyama H, Wakabayashi K, Hirose M, et al. 1991. Presence of carcinogenic heterocyclic amines in urine of healthy volunteers eating normal diet, but not of inpatients receiving parenteral alimentation. Carcinogenesis 12:1417–1422. Van Heek T, Rader AE, Offerhaus GJ, et al. 2002. K-ras, p53, and DPC4 (MAD4) alterations in fine-needle aspirates of the pancreas: A molecular panel correlates with and supplements cytologic diagnosis. Am J Clin Pathol 117:755–765. Van Kranen HJ, Vermeulen E, Schoren L, et al. 1991. Activation of c-K-ras is frequent in pancreatic carcinomas of Syrian hamsters, but is absent in pancreatic tumors of rats. Carcinogenesis 12:1477–1482. Velema JP, Walker AM, Gold EB. 1986. Alcohol and pancreatic cancer. Insufficient epidemiologic evidence for a causal relationship. Epidemiol Rev 8:28–41. Vena JE, Sultz HA, Fiedler RC, et al. 1985. Mortality of workers in an automobile engine and parts manufacturing complex. Br J Ind Med 42:85–93. Viadana E, Bross IDJ, Houten L. 1976. Cancer experience of men exposed to inhalation of chemicals or to combustion products. J Occup Med 18:787–792. Vineis P, Schulte PA, Carreon T, et al. 2004. Mechanisms of Carcinogenesis: Contributions of Molecular Epidemiology IARC Scientific Publication No. 157, Issues of design and analysis in studies of gene environment interactions. Lyon, France: International Agency for Research on Cancer. Vogelstein B, Fearon ER, Hamilton SR, et al. 1988. Genetic alterations during colorectal-tumor development. N Engl J Med 319:525–532. Voirol M, Infante F, Raymond L, et al. 1987. [Nutrional profile of patients with cancer of the pancreas.] Schweiz med Wochenshr 117:1101– 1104. Wang F, Herrington M, Larsson J, et al. 2003. The relationship between diabetes and pancreatic cancer. Mol Cancer 2:4. Wang M, Abbruzzese JL, Friess H, et al. 1998. DNA adducts in human pancreatic tissues and their potential role in carcinogenesis. Cancer Res 58:38–41. Waterbor J, Cole P, Delzell E, et al. 1988. The mortality experience of majorleague baseball players. N Engl J Med 318:1278–1280. Wattenberg LW. 1985. Chemoprevention of cancer. Cancer Res 45:1–8. WCRF (World Cancer Research Fund). 1997. Food, Nutrition, and the Prevention of Cancer: A Global Perspective. Washington, D.C.: World Cancer Research Fund/American Institute for Cancer Research. Wei D, Xiong HQ, Abbruzzese JL, et al. 2003. Experimental animal models of pancreatic carcinogenesis and metastasis. Int J Gastrointest Cancer 33:43–60. Weiderpass E, Vainio H, Kauppinen T, et al. 2003. Occupational exposures and gastrointestinal cancers among Finnish women. J Occup Environ Med 45:305–315. Weir JM, Dunn JE, Jr. 1970. Smoking and mortality: A prospective study. Cancer 25:105–112. Weisberg SP, McCann D, Desai M, et al. 2003. Obesity is associated with macrophage accumulation in adipose tissue. J Clin Invest 112:1796– 1808. Weusburger JH. 2000. Eat to live, not live to eat. Nutrition 16:767–773. Weiss W, Bernarde MA. 1983. The temporal relation between cigarette smoking and pancreatic cancer. Am J Public Health 73:1403–1404. Wen CP, Tsai SP, McClellan WA, et al. 1983. Long-term mortality study of oil refinery workers. I. Mortality of hourly and salaried workers. Am J Epidemiol 118:526–542. Westhoff C, Gentile G, Lee J, et al. 1996. Predictors of ovarian steroid secretion in reproductive-age women. Am J Epidemiol 144:381–388. Whitcomb DC, Gorry MC, Preston RA, et al. 1996. Hereditary pancreatitis is caused by a mutation in the cationic trypsinogen gene. Nat Genet 14:141–145. Whitcomb DC, Pogue-Geile K. 2002. Pancreatitis as a risk for pancreatic cancer. Gastroenterol Clin North Am 31:663–678.

762

PART IV: CANCER BY TISSUE OF ORIGIN

Whittemore AS, Paffenbarger RS, Anderson K, et al. 1983. Early precursors of pancreatic cancer in college men. J Chron Dis 36:251–256. WHO (World Health Organisation). 1995. Physical status: The use and interpretation of anthropometry. Report of a WHO Expert Committee. WHO Technical Report Series, No. 854 Geneva: World Health Organization, pp. 1–452. WHO. 2004. World Health Organisation (WHO) Databank: WHO. http://www.dep.iarc.fr/ and http://www.who.int/whosis/ Wideroff L, Gridley G, Mellemkjaer L, et al. 1997. Cancer incidence in a population-based cohort of patients hospitalized with diabetes mellitus in Denmark. J Natl Cancer Inst 89:1360–1365. Wiklund K, Dich J, Holm LE, et al. 1989. Risk of cancer in pesticide applicators in Swedish agriculture. Br J Ind Med 46:809–814. Wiklund K, Holm LE. 1986. Trends in cancer risks among Swedish agricultural workers. J Natl Cancer Inst 77:657–664. Willett WC, Stampfer MJ. 1986. Total energy intake: Implications for epidemiologic analyses. Am J Epidemiol 124:17–27. Williams RR, Horm JW. 1977. Association of cancer sites with tobacco and alcohol consumption and socioeconomic status of patients: Interview study from the Third National Cancer Survey. J Natl Cancer Inst 58:525–547. Williams RR, Stegens NL, Goldsmith JR. 1977. Associations of cancer site and type with occupation and industry from the Third National Cancer Survey Interview. J Natl Cancer Inst 59:1147–1185. Wingren G, Persson B, Thoren K, et al. 1991. Mortality pattern among pulp and paper mill workers in Sweden: A case-referent study. Am J Ind Med 20:769–774. Wogan GN, Hecht SS, Felton JS, et al. 2004. Environmental and chemical carcinogenesis. Semin Cancer Biol 14:473–486. Wolk A, Gridley G, Svensson M, et al. 2001. A prospective study of obesity and cancer risk (Sweden). Cancer Causes Control 12:13– 21. Wong O, Brocker W, Davis HV, et al. 1984. Mortality of workers potentially exposed to organic and inorganic brominated chemicals, DBCP, TRIS, PBB, and DDT. Br J Ind Med 41:15–24. Wong O, Morgan RW, Bailey WJ, et al. 1986. An epidemiological study of petroleum refinery employees. Br J Ind Med 43:6–17.

WCRF (World Cancer Research Fund). 1997. Food, Nutrition, and the Prevention of Cancer: A Global Perspective. Washington, D.C.: World Cancer Research Fund/American Institute for Cancer Research. Woutersen RA, Appel MJ, Van Garderen-Hoetmer A, et al. 1999. Dietary fat and carcinogenesis. Mutat Res 443:111–127. Wynder EL, Dieck GS, Hall NEL. 1986. Case-control study of decaffeinated coffee consumption and pancreatic cancer. Cancer Res 46:5360–5363. Wynder EL, Hall NEL, Polansky M. 1983. Epidemiology of coffee and pancreatic cancer. Cancer Res 43:3900–3906. Wynder EL Mabuchi K, Maruchi N, et al. 1973a. A case control study of cancer of the pancreas. Cancer 31:641–648. Wynder EL Mabuchi K, Maruchi N, et al. 1973b. Epidemiology of cancer of the pancreas. J Natl Cancer Inst 50:645–667. Yang GY, Wagner TD, Fuss M, Thomas CR, Jr. 2005. Multimodality approaches for pancreatic cancer. CA Cancer J Clin 55:352–367. Yanai H, Inaba Y, Takagi H, et al. 1979. Multivariate analysis of cancer mortalities for selected sites in 24 countries. Environ Health Perspect 32:83–101. Yassi A, Tate R, Routedge M. 2003. Cancer incidence and mortality in workers employed at a transformer manufacturing plant: update to a cohort study. Am J Ind Med 44:58–62. Young J, Devesa S, Culter SJ. 1975. Incidence of cancer in United States blacks. Cancer Res 35:3523–3526. Zatonski W, Boyle P, Przewozniak K, et al. 1993. Cigarette smoking, alcohol, tea and coffee consumption and pancreas cancer risk: a case-control study from Opole, Poland. Int J Cancer 53:601–607. Zatonski W, Przewozniak K, Howe GR, et al. 1991. Nutritional factors and pancreatic cancer: A case-control study from south-west Poland. Int J Cancer 48:390–394. Zheng W, Doyle TJ, Kushi LH, et al. 1996. Tea consumption and cancer incidence in a prospective cohort study of postmenopausal women. Am J |Epidemiol 144:175–182. Zheng W, McLaughlin JK, Gridley G, et al. 1993. A cohort study of smoking, alcohol consumption, and dietary factors for pancreatic cancer (United States). Cancer Causes Control 4:477–482. Zoloth SR, Michaels DM, Villalbi JR, et al. 1986. Patterns of mortality among commercial pressmen. JNCI 76:1047–1051.

39

Liver Cancer W. THOMAS LONDON AND KATHERINE A. MCGLYNN

T

his chapter is concerned with cancers that originate in the liver, not with cancers that begin in other sites and metastasize to the liver. Liver cancers, 75%–90% of which are hepatocellular carcinomas (HCC), caused more than 600,000 deaths in 2002 (WHO, 2002). They are the third most common cause of cancer deaths among men and sixth most common among women (Parkin, 2001a). Approximately 80% of HCCs and the resulting deaths occur in the developing countries of Asia and Africa. In the United States, liver cancer incidence is increasing, but it is still relatively uncommon (El-Serag and Mason, 1999). Hepatocellular carcinomas are malignant tumors of liver parenchymal cells (hepatocytes). The other primary liver cancer that occurs with a much lower, but still significant frequency is intrahepatic cholangiocarcinoma (ICC), a tumor of cells lining bile ducts. ICC is associated with chronic inflammation and injury of bile duct cells (cholangitis). Eighty to 95% of HCCs are associated with chronic infection of hepatocytes with either of two viruses, hepatitis B virus (HBV) or hepatitis C virus (HCV). HBV infections account for 75%–80% of virus-assoiated HCCs, while HCV is responsible for 10%–20%. HBV infection is preventable by immunization and HCV is largely preventable by public health measures. Therefore, if the opportunities for intervention were acted upon, HCC could become a minor cause of cancer mortality in the foreseeable future.

CLASSIFICATION Anatomic Distribution The liver is a common site of metastasis for tumors originating in other organs. Death certificates and hospital charts cannot be relied on to distinguish primary from secondary tumors. Statistics based on criteria other than histopathology are suspect. Nevertheless, because of the hazards associated with biopsying liver tumors, much of the literature, particularly from developing countries, is not derived from histologically verified diagnoses. Both benign and malignant tumors occur in the liver, but most benign tumors are quite rare. Table 39–1 provides a simplified classification of liver tumors, with a brief description of their histopathology, etiology, and epidemiology. Hepatocellular adenomas are usually subcapsular, well circumscribed, 2–20 cm in diameter, and located predominantly in the right lobe of the liver. Because they are composed entirely of hepatocytes, the blood supply to the center of the tumor is limited, which makes these neoplasms prone to central necrosis and hemorrhage. They may present with acute abdominal pain resulting from bleeding into the peritoneal cavity (Flora and Berk, 2002). Bile duct adenomas are also generally subcapsular, but they are generally trivial lesions (1 cm in diameter or less) found at autopsy. Hemangiomas are usually solitary lesions less than 4 cm in diameter that can occur anywhere in the liver. They are often detected as incidental findings on ultrasonogram (US) or computerized tomogram (CT) of the liver. Giant hemangiomas, defined as larger than 10 cm in diameter, are usually subcapsular. They are usually solitary, but occasionally are multiple. They may spontaneously involute or they may more rarely give rise to angiosarcomas. Infantile hemangioendotheliomas may also occur anywhere in the liver, but they may be solitary or multicentric and vary in size from 1–15 cm in diameter. They are usually detected during the first 6

months of life. Despite their size, they almost always spontaneously involute. Mesenchymal hamartomas are uncommon tumors that are usually diagnosed before age 2 years. They range in size from 15–30 cm in diameter and may occur anywhere in the liver. Hepatoblastomas, the most common childhood hepatic tumors, vary in size at diagnosis from 5–17 cm and usually present at diagnosis as a large, well-circumscribed solitary mass. The right lobe is the site of 57% of tumors, the left lobe of 15%, and both lobes of 27% (Stocker and Ishak, 1987). Angiosarcomas are most often multifocal, involving the entire liver (Ishak, 1987). Cholangiocarcinomas (also known as cholangiocellular carcinomas) are tumors arising in the epithelial lining of the bile duct. Intraheptic cholangiocarcinomas (ICC) arise in the intrahepatic bile duct and are classified by the International Classification of Diseases as primary liver cancers. Extrahepatic cholangiocarcinomas arise in the extrahepatic bile duct and are classified as primary biliary tract cancers (see Chapter 40 for discussion of extrahepatic bile duct cancer). Two types of ICC exist: those arising in the periphery of the liver and those arising more central to the major bile ducts. These tumors are more often diffuse and multicentric, but they may be solitary or multinodular. The diffuse types may be densely fibrotic (de Groen et al., 1999; Edmondson, 1958). Prospective studies of persons chronically infected with hepatitis B virus in China, Japan, and Alaska demonstrate that hepatocellular carcinomas begin as solitary nodules that can occur anywhere in the liver. The right lobe is more frequently involved than the left, but this may be related to its larger size. A diffuse form has been described in patients in Africa, but screening studies of hepatitis B carriers have not been done to exclude the possibility that these tumors also begin as single nodules. Hepatocellular carcinomas commonly invade the portal and hepatic venous systems, producing tumor thrombi in the portal vein and its tributaries. Tumor thrombi undoubtedly occur in hepatic veins, but are more difficult to identify (Kojiro and Nakashima, 1987).

Precursor Neoplastic Lesions of Hepatocellular Carcinoma The time from initial infection with hepatitis B or C viruses to the development of hepatocellular carcinoma is two to eight decades. Iron overload associated with hemochromatosis begins early in life, but HCC generally does not develop until the fifth to seventh decade. During this long interval of exposure many changes may occur in the liver including chronic inflammation, fibrosis, cirrhosis, increased hepatocyte death rates, and regeneration. Studies of experimental hepatocarcinogenesis in animals reveal a sequence of events beginning with foci of phenotypically altered hepatocytes, proceeding to dysplastic foci and nodules (Farber and Cameron, 1980). This sequence has been confirmed in woodchucks chronically infected with the woodchuck hepatitis virus (Toshkov et al., 1990) and in humans chronically infected with hepatitis B or C viruses (Takaishi et al., 2000; Tornillo et al., 2002). The continuous cycle of cell death and regeneration may eventually result in the proliferation of hepatic stem cells morphologically recognizable as oval cells (Vessey and de la Hall, 2001; Fausto and Campbell, 2003), although mature hepatocytes are the major source of cell replacement in the damaged liver (Summers et al., 2003). Oval cells may give rise to the phenotypically

763

764

PART IV: CANCER BY TISSUE OF ORIGIN

Table 39–1. Classification of Primary Liver Tumors Liver Tumors

Histology

Etiology/Epidemiology

benign Simple hepatic cyst Hepatocellular adenoma

Hemangioma Infantile hemangioendothelioma Epithelioid hemangioendothelioma Bile duct hamartoma Mesenchymal hamartoma

Unilocular lesion lined by simple cuboidal epithelium Composed entirely of well-differentiated hepatocytes with abundant eosinophilic cytoplasm, arranged in sheets, cords. No lobules, portal areas, bile ducts. No capsule, but clearly demarcated from surrounding liver. Differentiated endothelial cells Type I: Multiple small vascular channels. Type II pleomorphic endothelial cells. Dendritic and epithelioid cells. Not necessarily benign. May invade hepatic veins, metastasize. Usually multiple lesions; dense stroma containing dilated bile ducts Multicystic masses of loose mesenchymal tissue containing small bile ducts

80% in women Almost all cases in women. Exposure to oral contraceptives, androgens, clomiphene. Associated with glycogen storage disease Type I, tyrosinemia, galactosemia. Most common benign tumor; Prevalence 1%–5% of adults; Giant hemangiomas (>10 cm) more common in women Diagnosed at birth to 6 months. Spontaneously involute.

malignant Hepatoblastoma

Epithelial type (56%)—fetal and/or embryonal hepatocytes, rarely macrotrabecular or anaplastic. Mixed type (44%)— epithelial and mesenchymal components (osteoid, cartilage, other spindle cells).

Cholangiocarcinoma

Adenocarcinoma of bile duct epithelial cells (90%) Squamous cell tumors (10%)

Angiosarcoma

Multicentric ill-defined hemorrhagic nodules composed of anaplastic endothelial-derived spindle cells.

Hepatocellular carcinoma

Well, moderately, or poorly differentiated parenchymal cells. Trabeculae 2–8 cells thick, separated by sinusoids. Clear cell variant contains excess glycogen. Fibrolamellar variant— Large polygonal, eosinophilic cells circumscribed by bundles of acellular, creating large tumor islands or trabeculae.

Most common childhood hepatic tumor. Incidence = 1.5/106/yr; Wh : Bl = 5 : 1; M : F = 1.5–2.0 : 1. Associated with Beckwith-Wiedemann syndrome, hemihypertrophy, familial adenomatous polyposis, precocious puberty. US incidence = 1/105/yr; M : F = 1.5 : 1. International incidence = 2–6/105/yr. S. E. Asia associated with chronic infestation with liver flukes Clonorchis sinensis, Opisthorcis viverrini. Associated with inflammatory bowel disease, primary sclerosing cholangitis, a 1-antitrypsin deficiency, thorotrast exposure. US incidence = 25 cases/yr; M : F = 3 : 1. In adults, associated with exposure to thorotrast, arsenicals, vinyl chloride monomer. M : F = 3–5 : 1 Chronic infection with hepatitis B or C viruses Exposure to mycotoxins, alcohol, androgens

tumor-like growths Focal nodular hyperplasia Nodular regenerative hyperplasia

Solitary, circumscribed, unencapsulated, nodule with fibrotic central (stellate) scar with radiating septae. Contains disorganized hepatocytes, bile ducts, Kuppfer cells among fibrous bands. Multiple nodules without fibrosis containing normal hepatocytes, bile ducts, Kuppfer cells

8% of primary liver “tumors” in US M:F = 1:2 Not associated with oral contraceptives; estrogens stimulate growth. Rare disease, M < F. Associated with renal transplantation, auto-immune diseases.

Source: See De Groen et al., 1999; Edmondson, 1958; Flora and Berk, 2002; Kojiro and Nakashima, 1987; Ishak, 1987; Stocker and Ishak, 1987.

altered foci, but whether the oval cell is the precursor cell of HCC is unproven. Many phenotypically altered foci and dysplastic nodules are composed of monoclonal populations of hepatocytes (Takaishi et al., 2000).

Molecular Genetic Characteristics of Hepatocellular Carcinoma Hepatocellular carcinomas are characterized by a heterogeneous array of epigenetic and genomic changes, no one of which molecularly defines these tumors. Comparative genomic hybridization (CGH) studies, for example, have revealed that almost all HCCs have at least one allelic chromosome deletion, but these are scattered across the genome with deletions on 11 chromosome arms in 30%–60% of HCCs. CGH has also demonstrated gains on four chromosome arms in 25%–55% of tumors. Most HCCs have multiple allelic deletions and gains and losses on several chromosome arms (Thorgeirsson and Grisham, 2002). Epigenetic changes are also common in HCCs and apparently precede detectable structural changes in genes and chromosomes since they are present in a high proportion of dysplastic and cirrhotic nodules. Telomerase expression, with concomitant telomere shorten-

ing, is increased in more than 50% of dysplastic nodules and more than 80% of HCCs. Aberrant methylation, both hypomethylation (gene activation) and hypermethylation (gene silencing), occur in 10%–30% of livers with chronic hepatitis and cirrhosis and in 40%–75% of HCCs. DNA methyltransferases (DNMTs), which catalyze methylation and demethylation reactions, are up-regulated in HCCs, as are Sadenosylmethionine synthase and glycine N-methyltransferase, which increase the cellular pools of methyl groups available for methylation reactions (Thorgeirsson and Grisham, 2002). In summary, a large proportion of hepatocytes are affected by one or more epigenetic or genetic alterations during the long period of HCC development. Because HCCs are clonal and usually single tumors, only rare altered hepatocytes give rise to a malignant neoplasm. The combination of abnormalities that gives rise to the neoplastic phenotype is unknown.

DEMOGRAPHIC PATTERNS Incidence, Mortality and Survival Around the World Primary liver cancer is the fifth most common cancer in the world and the third most common cause of cancer mortality (Parkin, 2001b).

Liver Cancer

765

Figure 39–1. Global incidence of liver cancer in males, 2000. (Source: Parkin, 2001.)

There is wide geographical variability in incidence (Fig. 39–1). The great majority of liver cancer (>80%) occurs in either sub-Saharan Africa or in Eastern Asia, with one country alone, China, accounting for 50% of the cases. In most countries, liver cancers are primarily HCCs (75%–90% of cases) and secondarily ICCs (10%–25% of cases) (Okuda et al., 2002). Because of this, trends in total liver cancer incidence and mortality tend to reflect trends in HCC incidence and mortality. An exception is northeast Thailand, which has one of the highest rates of liver cancer in the world due to the exceptionally high incidence of ICC (male incidence = 88/100,000; female incidence = 35.4/100,000). In contrast, the cancer registry reporting the highest liver cancer rates due to HCC is Qidong, China, where male incidence is 72.1/100,000 and female incidence is 29.6/100,000. North America, South America, Northern Europe, and Australia tend to be low-rate areas for primary liver cancer. Typical of rates in these areas are those of Canada (male: 3.2/100,000; female: 1.1/100,000), the United Kingdom (male: 2.2/100,000; female: 1.1/100,000), and Australia (male: 3.6/100,000; female: 1.0/100,000). Spain (male: 7.5/100,000; female: 5.5/100,000), Italy (male: 13.5/100,000; female: 4.6/100,000), and Greece (male: 12.1/100,000; female: 4.6/100,000), by contrast, have rates typical of medium-rate countries (Ferlay et al., 2001). Although the majority of primary liver cancers occur in Eastern Asia and sub-Saharan Africa, declining trends in incidence have been seen in some of these high-rate areas (McGlynn et al., 2001). Between 1978–1982 and 1988–1992, cancer registries in Hong Kong, Shanghai, and Singapore reported decreases in incidence. In contrast, registries in a number of low-rate areas have reported increases in incidence including Australia, the United States, Canada, and the United Kingdom. Reasons for both the decreased incidence in highrate areas and increased incidence in low-rate areas are not yet clear, but the increased incidence in low-rate areas may be related to increased prevalence of HCV infection (El-Serag et al., 2003). Because ICC incidence rates are frequently not reported separately from all liver cancer, ICC incidence trends in many countries are not available. Mortality rates of ICC, however, have been reported to be increasing in a number of countries (Taylor-Robinson et al., 2001;

Khan et al., 2002; Patel, 2002). Whether the rates represent increased incidence, better diagnosis, better survival from predisposing medical conditions, or diagnostic misclassification, is not clear. Survival rates of primary liver cancer are uniformly poor in both high-rate and low-rate areas. The International Agency for Research on Cancer (IARC) estimates that the age-standardized worldwide incidence rate of primary liver cancer among males is 17.4/100,000 in underdeveloped countries and 8.7/100,000 in developed countries. The corresponding mortality rates are 16.8/100,000 and 8.1/100,000, indicating very little difference in survival in the contrasting areas (Ferlay et al., 2001).

Incidence, Mortality, and Survival in the United States Primary liver cancers are uncommon in the United States. Liver cancers account for approximately 1.3% of new cancer cases and 2.6% of cancer deaths (Jemal et al., 2003). In the United States approximately 65% of liver tumors are HCCs, 15% are ICCs, and 20% are either rare tumor types, such as angiosarcomas, or types too poorly defined to be accurately classified (Table 39–2) (SEER Program, 2003). In the Surveillance, Epidemiology, and End Results (SEER) registries during the period 1992–1999, the highest age-adjusted incidence rate of liver cancer among males occurred in Asians/Pacific Islanders (21.08/100,000), followed by African Americans (9.61/100,000), American Indians/Alaska Natives (7.69/100,000), and whites (6.43/100,000). Among females, Asians/Pacific Islanders again had the highest rate (7.79/100,000), followed by American Indians/Alaska Natives (5.74/100,000), African Americans (3.54/100,000), and whites (2.52/100,000). Temporal data are limited for Asians/Pacific Islanders and American Indians/Alaska Natives, but are available for whites and African Americans for the years 1973 through 2000. During this time period, African Americans experienced consistently higher incidence rates than did whites. All four sex/ethnic groups, however, experienced increases in incidence with white males seeing the greatest percentage increase (107.3% between the years 1973–1978 and 1996–2000). In comparison, African American male

766

PART IV: CANCER BY TISSUE OF ORIGIN

Table 39–2. Age-Adjusted Incidence Rates (US 2000 Standard) of Primary Hepatic Tumors in 9 SEER Registries, 1973–1978 and 1996–2000 White Males

All tumors All liver tumors HCC ICC Hepatoblastoma Angiosarcoma

White Females

Black Males

Black Females

1973– 1978 Rate

1996– 2000 Rate

% Increase

1973– 1978 Rate

1996– 2000 Rate

% Increase

1973– 1978 Rate

1996– 2000 Rate

% Increase

1973– 1978 Rate

1996– 2000 Rate

% Increase

479.4 3.30 1.99 0.38 0.04 0.01

589.0 6.82 4.62 1.08 0.05 0.01

22.8 107.3 132.1 184.2 25.0 0

418.8 1.52 0.63 0.26 0.02 0.03

485.8 2.59 1.22 0.67 0.05 0.01

16.0 70.4 93.6 157.7 150.00 -66.7

551.8 6.36 4.89 0.29 0.01 0.06

715.6 11.0 8.75 1.03 0.04 0.00

29.7 73.4 78.9 255.2 300.0 -100.0

427.4 2.35 1.49 0.20 0.01 0.04

443.5 3.90 2.56 0.42 0.03 0.00

3.8 66.0 71.8 110.0 200.0 -100.0

HCC, hepatocellular carcinoma; ICC, intrahepatic cholangiocarcinoma. All tumors: all sites combined. All liver tumors: all primary tumors of liver and intrahepatic bile duct. *All rates per 100,000 per year.

rates increased 73.4%, white female rates increased 70.4%, and African American female rates increased 66.0% (SEER Program, 2003). Unlike HCC, incidence rates of ICC in the United States are slightly higher among whites (M: 0.38/100,000; F: 0.26/100,000) than among blacks (M: 0.29/100,000; F: 0.20/100,000). Among all four sex-ethnic groups, the highest rates occur among persons aged 75 years and older. Both HCC and ICC have increased in incidence. Between 1973–1978 and 1996–2000, HCC incidence increased 132.1% in white males, 93.6% in white females, 78.9% in black males, and 71.8% in black females. During the same interval, ICC incidence increased 184.2% in white males, 157.7% in white females, 255.2% in black males, and 110.0% in black females (SEER Program, 2003). Survival rates for both HCC and ICC are exceedingly poor. During the period 1992–1999, the 5-year relative survival rate for all liver cancer was only 6.9%, compared with 63.0% for all types of cancer. Unfortunately, the 5-year survival rates for ICC are even poorer than those for HCC, with only 4.7% of males and 3.0% of females surviving for 5 years (Ries et al., 2003). Because survival is so uniformly dismal, incidence and mortality rates are virtually identical. During the period 1996–2000, mortality among white males was 6.0/100,000, with corresponding rates of 2.7/100,000 among white females, 9.3/100,000 among black males, and 3.7/100,000 among black females. Similar to incidence rates, mortality rates increased among all four sex/ethnic groups during the time period between 1973–1976 and 1998–2000. White males experienced the greatest percentage increase (69.5%) and black females the lowest percentage increase (34.4%). In contrast, African American and white females saw percentage increases of 43.8% and 36.0%, respectively (Ries et al., 2003).

Age In most of the world’s populations, liver cancer incidence is highly correlated with age (Parkin et al., 1997). This is particularly true among low-risk populations (e.g., United States, Canada, Australia, United Kingdom) where the highest age-specific rates occur among persons aged 80 years and greater. Female rates in high-risk Asian populations (e.g., Hong Kong, Shanghai, Singapore) also follow this age pattern. In contrast, male rates in high-risk Asian populations tend to peak about 10 years younger, at age 70. Incidence rates among African populations (Bamako, Harare, Kyadondo County) peak at 10 years younger, yet, at ages 55–60 years. Exceptions to these age patterns occur among the populations of Japan and Qidong, China. In Japan, particularly among males, incidence rates peak at age 60 and then plateau. In very high-rate Qidong, China, the age-specific male incidence rates rise until age 50 and then decline. Among females, the incidence rates rise until age 60, and then decline.

The different age patterns of risk are most likely related to the dominant hepatitis virus in the population, the age at viral infection, and the existence of other risk factors. In Japan, the dominant virus is HCV, while in Qidong, China, it is HBV. Almost all HBV viral carriers in Qidong and elsewhere have been infected since birth. Most persons infected with HCV, however, became infected as adults.

Sex Worldwide, the male : female ratio in liver cancer incidence is biased toward males, with most regions seeing ratios between 2 : 1 and 4 : 1. Whereas it has previously been reported that the greatest discrepancies between male and female rates occurred in high-risk populations (Parkin, 2001b), this no longer appears to be the case (McGlynn et al., 2001). At present, the largest discrepancies in rates are found in low-risk populations of Europe. Typical among these ratios are the ones reported from Calvados, France (8.8 : 1), Geneva, Switzerland (7 : 1), and Trieste, Italy (4.8 : 1). In contrast, typical ratios currently seen in high-risk populations are those of Qidong, China (3.7 : 1), Osaka, Japan (4.0 : 1), Kangwha, Korea (3.6 : 1), and Hanoi, Vietnam (4.1 : 1). The only registries in the world that report ratios at or near 1 : 1 are in South America (Cali, Colombia; Quito, Ecuador; and Lima, Peru). The more pronounced male : female ratios currently found in low-rate areas may be related to the changing rates in these regions, with male rates increasing somewhat faster than female rates in low-risk areas while decreasing somewhat faster than female rates in high-risk areas. The reasons that males have higher rates of liver cancer than females are not completely understood, but may be partly explained by the sex-specific prevalence of risk factors. Males are more likely to be infected with HBV and HCV, consume alcohol, smoke cigarettes, and have increased iron stores. Whether androgenic hormones or increased genetic susceptibility influence gender differences in rates is not clear.

Ethnicity Liver cancer incidence rates vary not only by country, but also by ethnic group within individual countries. For example, as shown in Table 39–3, Korean men living in Los Angeles have incidence rates more than 5 times higher than white men. Similarly, ethnic Chinese men in Singapore have rates 2.7 times greater than Indian men, while Chinese men in San Francisco have rates 4.2 times the rate of San Francisco white men. The differences in rates among men of various ethnic groups are equally true of rates among women. The ethnic variation in rates almost certainly reflects differences in the likelihood of infection with HBV and HCV, although genetic susceptibility and different patterns of exposure to other risk factors may also play a role.

Liver Cancer Table 39–3. Primary Liver Cancer Incidence Rates by Ethnicity Within Selected Registries (1993–1997) Male Rate*

Female

Rate/Base

Rate*

Rate/Base

Los Angeles 3.9 6.5 5.5 10.6 10.9 16.2 20.7

1.0 1.7 1.4 2.7 2.8 4.2 5.3

White Black Japanese Hispanic Filipino Chinese Korean

1.6 2.0 4.3 3.8 2.4 5.0 10.4

1.0 1.3 2.7 2.4 1.5 3.1 6.5

1.0 1.2 1.4 2.0 2.0

White Japanese Chinese Filipino Hawaiian

1.5 3.4 4.8 2.0 3.1

1.0 2.3 3.2 1.3 2.1

1.0 2.6 3.8

White Hispanic American Indian

1.4 3.3 5.3

1.0 2.4 3.8

1.0 2.0 2.7

Indian Malay Chinese

1.8 3.3 5.1

1.0 1.8 2.8

Hawaii 5.0 6.2 7.2 9.9 9.9

New Mexico 3.2 8.4 12

Singapore 7.9 16.0 21.2

*All rates are age-adjusted to the world standard population and are calculated per 100,000 person-years.

Figure 39–2. Geographic distribution of chronic hepatitis-B infection, 2000.

767

ENVIRONMENTAL FACTORS Hepatitis B Virus The biology and epidemiology of the hepatitis viruses, HBV and HCV, are described in Chapter 26. In this and the following section, we will expand on the relationships of chronic HBV and HCV infections to hepatocellular carcinoma. Several pathologists first proposed in the 1950s the possibility that chronic virus infections of the liver could lead to liver cancer (Edmondson and Steiner, 1954; Edmondson, 1958; Higginson et al., 1957). They noted that at least 80% of liver cancers occurred in cirrhotic livers. Edmondson (1958) pointed out that in the United States and western Europe 3%–10% of male patients with cirrhosis developed HCC, whereas in Asia and Africa 15%–50% of men with cirrhosis developed liver cancer. They proposed that cirrhosis in western countries was mainly due to alcohol abuse, but in Asia and Africa where alcoholism was uncommon, some other agent caused cirrhosis and HCC. They included nutritional deficiencies, exposure to toxins, and chronic viral infections among the possible candidates. We now know that chronic infection with HBV or HCV is causally associated with 80%–95% of all HCCs in the world. In the regions of the world where HCC is most common, HBV infection is associated with most cases of cirrhosis and 80% or more of the cases of HCC. The evidence supporting the causal association of HBV with HCC is summarized below: 1. Areas of the world with high incidence and mortality rates for HCC have high prevalences of chronic HBV infection. The reverse is also true. Countries with prevalences of chronic HBV infection of greater than 2% have increased incidence and mortality rates for HCC (Figs. 39–1 and 39–2) (Parkin, 2001a; CDC, 2002). 2. Cirrhosis is closely associated with chronic HBV infection in the regions of the world where HCC mortality is high (Hann et al., 1982; Yang and Tang, 1985). 3. Case-control studies in all regions of the world have consistently shown that chronic HBV infection (seropositivity for hepatitis B surface antigen, HBsAg) is much more common among cases than controls. Odds ratios range from 5 : 1–65 : 1 (IARC Monographs, 1994a).

768

PART IV: CANCER BY TISSUE OF ORIGIN

4. Prospective studies of persons chronically infected with HBV have demonstrated very high relative risks for HCC, ranging from 5 to 103 (Beasley et al., 1981; Ijima et al., 1984; Heyward et al., 1985; IARC Monographs, 1994a; Evans et al., 2002). In Beasley’s study of male government workers in Taiwan, the age-adjusted annual incidence of HCC was 474 per 100,000 in HBsAg(+) men compared with 6 per 100,000 in HBsAg(-) men. A study by Evans et al. (2002) shows the magnitude of the effects of chronic HBV infection. The study enrolled an unselected general population of 58,545 men (15.0% HBV carriers) and 25,340 women (10.7% HBV carriers) in Haimen City, China in 1992–1993. After 8 years of followup, HCC was the major cause of death in the cohort, accounting for 977 of the 3487 deaths in the population. Among men, the cumulative risks for death from HCC were: 0.5% for HBV uninfected (HBsAg(-) at entry) and 8% for HBV carriers (HBsAg(+)); among women, the cumulative risks were 0.1% for HBV uninfected and 2.0% for HBV carriers. 5. In areas of the world with high incidences of both HCC and chronic HBV infection, about 70% of HBV infections are acquired in the perinatal period or in early childhood (Stevens et al., 1975; Okada et al., 1976; Marinier et al., 1985; CDC, 2003b). Thus, among HBV carriers in endemic areas, those born to HBV-infected mothers are likely to have been infected longest and are at higher risk of HCC than HBV carriers with HBsAg(-) mothers (Larouze et al., 1976; Hann et al., 1982; Beasley, 1988). 6. HBV is present in the liver tissues of almost all patients with HCC who are seropositive for HBsAg. In older studies using immunohistopathology, HBsAg was detected in up to 90% of the nonneoplastic liver tissues of such cases (Nayak et al., 1977; Tan et al., 1977). With the development of sensitive and specific assays for HBV DNA, all patients who are seropositive for HBsAg have detectable HBV DNA. Investigators have also detected HBV DNA sequences in 10%–20% of HCC tumors from patients who were seronegative for HBsAg, but positive for antibodies to HBsAg or HB core antigen (HBcAg) (Shafritz et al., 1981; Brechot et al., 1985; Ming et al., 2002). 7. Viruses belonging to the same family as HBV, hepadnaviruses, cause HCC in their natural hosts. The woodchuck hepatitis virus (WHV) in woodchucks (Summers et al., 1978), and the ground squirrel hepatitis virus (GSHV) in ground squirrels are the best studied examples (Marion et al., 1986). In a defining experiment, inoculation of newborn woodchucks with WHV resulted in chronic infection with the virus and HCC within 3 years (Popper et al., 1987). 8. Prevention of infection with HBV reduces the risk of subsequent HCC. Following universal vaccination of infants in Taiwan, the incidence of HCC among children declined from 0.7 to 0.36 per 100,000 (Chang et al., 1997). IARC has classified HBV as carcinogenic to humans (IARC Monographs, 1994a). Currently, about 5% of the world’s population (350 million people) is chronically infected with HBV. The lifetime risk of HCC in these individuals is estimated to be 10%–25%. The World Health Organization (WHO) and the Centers for Disease Control and Prevention (CDC) project that about 600,000 chronically infected people will die annually from HCC and chronic liver disease (CDC, 2003b); eventually 35–87 million of the 350 million prevalent HBV carriers with die of HCC (WHO, 2002).

Hepatitis C Virus Hepatitis C virus (HCV) like HBV may cause a chronic infection of very long duration accompanied by a slowly evolving liver disease. Unlike HBV, HCV is rarely acquired in childhood and, in adults, acute HCV infections are usually silent. Even among patients who eventually develop cirrhosis and HCC, the illness may not become clinically manifest for many years. Hepatitis C virus was identified in 1989, 20 years after the identification of HBV (Choo et al., 1989), and reliable tests for anti-HCV antibodies and HCV RNA became available over the succeeding several years. Therefore, the causal link of chronic

HCV infection to HCC has not been established as precisely as for HBV infections, but it is nevertheless a reliable relationship. IARC classified HCV as carcinogenic to humans in 1994 (IARC Monographs, 1994a), well before the most compelling epidemiologic studies were reported. The strongest evidence for a causal link of HCV infection with HCC comes from Japan. In fact, before the identification of HCV, Okuda and colleagues (1987) hypothesized that a non-A, non-B (NANB) virus was the cause of a significant proportion of HCC cases in Japan (Liver Cancer Study Group of Japan, 1984). They based their hypothesis on two observations: (1) the incidence of HCC (as recorded in the Osaka Cancer Registry) had risen from 16.3 in 1966–1968 to 35.6 in 1983; (2) over the same time interval the proportion of cases associated with HBV infection had fallen from 50% to 30%. They further proposed that the then unidentified NANB virus was acquired from a liberal use of blood transfusions to treat anemias of all causes after World War II. As soon as the first serological assay for anti-HCV antibodies became available, Kiyosawa et al. (1990) tested a group of 231 patients who had been diagnosed with chronic NANB hepatitis, including 96 with chronic hepatitis, 81 with cirrhosis, and 54 with HCC. Overall, while 90% of these patients tested positive for antiHCV antibodies, 94% of the HCC patients were positive. In addition, 86% of the HCV-associated HCC cases had cirrhosis. Serial liver biopsies were available from some patients and they demonstrated a progression from increasingly severe inflammation and fibrosis to cirrhosis to HCC. In this report, cirrhosis always preceded HCC. As predicted by Okuda et al. (1987), 42% of the HCV-associated HCC patients had received blood transfusions in the past whereas none of 29 HBV-associated HCC patients had been transfused. More information on risk of HCC can be derived from a recent prospective study from Japan of 2890 patients with hepatitis C who had a liver biopsy. Of this group, 490 were untreated and 59 untreated patients developed HCC over a 4.3-year interval. The annual incidence of HCC in patients with HCV and cirrhosis was 7.9%, and 0.5% in patients with little or no fibrosis (Yoshida et al., 1999). Studies of patients with chronic hepatitis and/or cirrhosis referred to liver disease clinics have confirmed that HCC can be an outcome of long-term HCV infection. The proportion of patients who develop HCC, however, varies by country of report, length of follow-up, and prevalence of cirrhosis (Table 39–4). Cirrhosis almost always precedes the diagnosis of HCC. These studies are informative for clinicians who most often are asked to treat patients with established chronic hepatitis and only rarely follow healthy individuals from initial exposure to HCV to subsequent disease outcomes. However, several cohort studies of this type have been reported in recent years and they project a much less dire outcome for most HCV-infected persons than do the clinicbased studies (Table 39–4). Four studies have been published of individuals whose time of exposure to HCV is known and whose HCV status near the time of exposure was documented. Seef et al. (1992) ascertained long-term mortality among five separate studies of transfusion-associated hepatitis conducted from 1967–1980. An average of 18 years after transfusion, all-cause mortality was the same, 51%, in both those patients who had developed NANB hepatitis and those patients who had not. Liver-related deaths occurred in 2.3% of the hepatitis cases and 1.3% of the controls. Cirrhosis was cited as the cause of death in only 1.9% of cases and 1.0% of controls. One patient with hepatitis (0.2%) and two controls (0.2%) died of HCC. After 25 years of follow-up in the three study groups that had archived serum samples, all-cause mortality remained equal between cases and controls, but liver-related deaths increased to 4.1% for the hepatitis cases and 1.3% for the controls. Nine liver-related deaths occurred among HCV-positive cases, 3 from HCC, 3 from cirrhosis and 3 from less-specific causes. Five liverrelated deaths occurred among controls; four from cirrhosis and none from HCC. Among cases and controls in these transfused populations, alcohol abuse was the strongest factor related to liver disease deaths. Among the same multicenter cohorts, long-term morbidity in a subset of 103 HCV(+) cases with elevated serum transferases (ALT) was ascertained. Thirty percent of these individuals had cirrhosis by

769

Liver Cancer Table 39–4. Long-term Outcome of Hepatitis C Virus Infection According to Study Design Study Design †

Retrospective

Prospective‡

Cohort

Author (Reference)

Country

No. Patients

Exposure Interval (yrs) Mean or Range of Means*

Kiyosawa et al. (1990) Tong et al. (1995) Yano et al. (1996) Niederau et al. (1998) Gordon et al. (1993) Gordon et al. (1993)** DiBisceglie et al. (1991) Koretz et al. (1993) Mattson et al. (1993) Tremolada et al. (1992) Seeff et al. (2001) †† Seeff et al. (2000) Kenny-Walsh et al. (1999)‡‡ Wiese et al. (2000)‡‡

Japan US Japan Germany US US US US Sweden Italy US US Ireland Germany

231 131 70 838 215 195 65 80 61 135 222 17 376 1018

10–29 14–28 NR 9–22 19 20 9.7 16.0 13.0 7.6 25 45–50 17 20

Cirrhosis (%) 35.1 51.0 50.0 16.8 55.0 21.0 12.3 7.0 8.0 15.6 15† 5.9 2.0 0.4

HCC (%) 23.4 10.6 NR 2.0 3.7 1.0 0 1.3 NR 0.7 1.9 0.0 0.0 0.0

Source: Adapted from Alter and Seef (2000). *Based on interval from transfusion or initial use of intravenous drugs when that date was known. † Based on referrals to tertiary care centers for persons with established liver disease. ‡ Exposure through transfusion. **Exposure through intravenous drug use. †† Long-term follow-up of persons studied from the time of acute infection. Recall of patients diagnosed with acute hepatitis in prior prospective transfusion studies followed by renewed prospective follow-up with non-hepatitis controls. ‡‡ Exposure to contaminated immune globulin.

histological criteria. The investigators assumed that no more than 5% of the subjects with normal ALT levels would have cirrhosis and estimated that in the total HCV-infected population less than 15% would develop cirrhosis in 20 years. In a second study by Seef and his colleagues (2000), frozen serum samples from 8568 Air Force recruits collected between 1948 and 1954 were tested for evidence of HCV infection. Only 17 were antiHCV positive; 11 were HCV RNA positive. After 50 years, mortality was higher among the 17 HCV-infected persons (41%) than among the HCV-negative population (26%). Nevertheless, only one of the seven deaths among the HCV infected group was from liver disease and none died from HCC. Two other studies, from Ireland and Germany, in which the time and source of exposure to HCV is known, contribute further to our understanding of the long-term risks of HCV infection. Both resulted form the inadvertent administration of HCV-contaminated Rh immune globulin. In Ireland, 704 women who had been exposed in 1977 (at a mean age of 28 years) tested positive for anti-HCV 17 years later; 390 were HCV RNA positive. Liver biopsies were performed on 363 women. Although most women showed mild to moderate degrees of inflammation and fibrosis, only 2% had definite cirrhosis and none developed HCC (Kenny-Walsh et al., 1999). In Germany, 1018 similarly exposed HCV-infected women were observed for 20 years and yielded similar results. Seven of these women developed overt cirrhosis (0.4%); none developed HCC. Liver biopsies of 44% of the HCV RNA-positive women showed some inflammation in 96% and mild degrees of fibrosis in 50% (Wiese et al., 2000). Combining the information derived from the long-term follow-up of patients in clinic-based studies with the cohort studies of individuals from time of initial infection, Alter and Seef (2000) proposed an algorithm of eventual outcomes of acute HCV infection. They suggested that among 100 persons acutely infected, 20% would spontaneously recover and 80% would develop a persistent infection. Of the 80 patients with chronic infection about 30% (24 patients) would develop severe progressive hepatitis. The other 70% (56 patients) would develop either stable chronic hepatitis or a very slowly progressive liver disease. These projections are similar to those of the WHO (2000), which proposed that 20% of the individuals initially infected with HCV would develop cirrhosis and 2%–4% HCC. Several factors affect the rate of progression of chronic liver disease among HCV-infected individuals. Like HBV, males appear to be at higher risk of HCC than females (Poynard et al., 1997). Males are more likely to engage in high-risk behaviors for acquiring HCV and

for acquiring co-infections with HBV and HIV, which may also increase risk of HCC. In contrast to HBV, older age (greater than 40 years) at infection is associated with an increased risk of HCC (Poynard et al., 1997; Colombo, 1998; Niederau et al., 1998). Age at infection, however, is not a strong risk factor and may be a surrogate for more important co-factors. Alcohol intake has been consistently found to increase risk of HCC (Miyakawa et al., 1996; Poynard et al., 1997; Corrao and Arico, 1998; Bellentani et al., 1999). In an Italian HCC case-control study of 305 HCC cases and 610 controls, in which the odds ratio for HCV infection was 35.6 when HBV markers were negative, odds ratios for alcohol intake (0–40 g/day as referent) increased 2.4-fold for daily intakes of 40–80 g and four-fold with intakes greater than 80 g/day controlling for HCV RNA positivity (Tagger et al., 1999). Globally, about 170 million people (3% of the world’s population) are chronically infected with HCV and 3–4 million are newly infected each year (WHO, 2002). The world distribution of chronic HCV infection (Fig. 39–3) is quite different from that of either HCC mortality or chronic HBV prevalence (Figs. 39–1 and 39–2). In the United States, the CDC estimates that 3.9 million people have been infected, of whom 2.7 million are chronically infected (Alter, 1999). The CDC (1998) projects that 8000–10,000 HCV-infected people will die annually in the United States from chronic liver disease or HCC. In the world, 7–9 million of the 170 million currently HCV-infected people will likely die eventually of HCC. Recent evidence, much of it reported from Japan, suggests that HCV infection, particularly in conjunction with alcohol, may also be related to ICC (Tomimatsu et al., 1993; Yamamoto et al., 1998; Yin and Chen, 1998; Tanaka et al., 1998a; Nagano et al., 2000; Suriawinata et al., 2000; Kobayashi et al., 2000; Donato et al., 2001; Polizos et al., 2003).

Aflatoxin Aflatoxin is a mycotoxin produced by Aspergillus species. Storage of corn or peanuts in warm, humid environments may lead to overgrowth of Aspergillus and heavy contamination with aflatoxin. Experiments in fish, poultry, and rodents have shown that aflatoxin B1 (AFB1) is a powerful hepatic carcinogen (Blount, 1961; Asplin and Carnaghan, 1961; Halver, 1965). The production of liver tumors in a variety of animal species led IARC to conclude that there was sufficient evidence to classify aflatoxin as a carcinogen (IARC Monographs, 1987a). Although there are four principal aflatoxins, B1, B2, G1, and G2, aflatoxin B1 (AFB1) is the most potent in animal studies (IARC

770

PART IV: CANCER BY TISSUE OF ORIGIN

Figure 39–3. Global distribution of chronic hepatitis-C infection, 2001 (WHO, 2001).

Monographs, 1987a). Many ecological studies of AFB1 contamination of food conducted in the 1970s and 1980s were compatible with a role for the carcinogen in human HCC, but assigning causality from such studies is problematic. However, person-specific epidemiological studies performed in the past 15 years have yielded stronger evidence of AFB1 as an etiologic factor or co-factor in HCC. These studies were permitted by the development of assays for aflatoxin metabolites in urine, AFB1-albumin adducts in serum, and detection of a signature aflatoxin DNA mutation in tissues. An interaction of AFB1 exposure with chronic HBV infection was revealed in short-term prospective studies in Shanghai, China. Urinary excretion of aflatoxin metabolites increased the risk of HCC four-fold, HBV infection increased the risk seven-fold, but individuals who both excreted AFB1 metabolites and were HBV carriers had a 60-fold increased risk of HCC (Ross et al., 1992; Qian et al., 1994). p53 is the most frequently mutated gene in human cancer. Usually such mutations are distributed throughout the coding regions of the gene. Among HCCs, however, a mutational hotspot at the third base of codon 249 (p53 249ser mutation) has been observed in 30%–60% of tumors arising in persons living in environments thought to be rich in aflatoxin (Ozturk, 1991; Bressac et al., 1991; Hsu et al., 1991). G to T transversion at codon 249 is postulated to result from the reaction of the 8,9 epoxide activated form of AFB1 with the N-7 guanine in DNA. Recently, Ming et al. (2002) analyzed a consecutive series of 181 cases of HCC from Qidong, China, a site of many previous studies of the relationship of HBV and AFB1 exposure to HCC. All 181 cases were positive for one or more markers of HBV infection. Serum samples from 119 cases were tested for antibodies to HCV; only six were positive. The p53 249ser mutation was detected in 54% of liver tumor tissues among all cases. Concomitantly, 42 age- and sexmatched HCC cases from Beijing, where AFB1 exposure is rare, were analyzed and none had a detectable 249ser mutation. The investigators then tested the liver tissues from 7 of 31 cases of HCC that had developed among a group of 145 men with chronic HBV infection and who had aflatoxin exposure documented in urine samples in 1987 and 1988. The 249ser mutation was found in all seven cases. Using the urinary aflatoxin measurements at the time of enrollment, the investi-

gators estimated the relative risk for AFB1 exposure as 3.5 (95% CI: 1.5–8.1). By and large, in areas of the world where AFB1 exposure continues to be an environmental problem, chronic HBV infection is highly prevalent. The fact that AFB1 is primarily a carcinogen in HBV carriers, however, should not be taken as a source of comfort, but a problem that must be addressed.

Alcohol The evidence in support of a positive association between alcohol consumption and HCC led IARC to conclude in 1988 that there was a causal relationship (IARC Monographs, 1988). Nevertheless, a number of questions about the alcohol-HCC relationship remain, including: Does alcohol have an effect in the absence of hepatitis viral infections? Does the combination of alcohol and HCV result in a greater risk than the combination of alcohol and HBV? Is the risk associated with alcohol consumption different for men and women? What is the mechanism by which alcohol increases the risk of HCC? Earlier studies of HCC found alcohol to be a more significant risk factor in low-incidence areas than in high-incidence areas. This may have been due to lower mean alcohol consumption in high-risk than in low-risk populations and/or due to the dominant effect of chronic HBV infection in high-incidence areas masking any additional risk of alcohol consumption. While some studies in HBV-endemic populations have shown a positive association between alcohol consumption and HCC (Oshima et al., 1984; Chen et al., 1991), other studies have not (Lam et al., 1982; Lu et al., 1988; Goodman et al., 1995; Yu et al., 1997a). In comparison, most studies in populations where HCV is more common than HBV have found alcohol to be a significant risk factor (Shibata et al., 1986; Kono et al., 1987; Tanaka et al., 1988; Hirayama et al., 1989; Tsukuma et al., 1990; Roudot-Thoraval et al., 1997; Ikeda et al., 1998; Corrao et al., 1999; Aizawa et al., 2000). Two recent studies, however, have reported equal risks when contrasting the combination of HBV and alcohol with the risk of HCV and alcohol (Kuper et al., 2000a; Donato et al., 2002). Donato et al. (2002) reported that alcohol consumption in Brescia, Italy was associated with HCC in the absence of either HBV or HCV

Liver Cancer infection, though the data suggest that higher levels of consumption are required for HCC in the absence of viral infection. Whether alcohol is more strongly associated with HCC in women than in men has been difficult to study given that women are less likely to be heavy drinkers and less likely to develop HCC than men. A greater effect of alcohol on women has been hypothesized based on differences in alcohol dehydrogenase activity (Frezza et al., 1990) and evidence of a greater association between alcohol and cirrhosis among women (Tuyns and Pequignot, 1984; Corrao et al., 1997). No substantial sex difference in risk of HCC with alcohol consumption, however, was reported in the Brescia study (Donato et al., 2002). The mechanism by which alcohol increases HCC risk is not entirely clear. Animal and human studies provide little evidence that ethanol is a carcinogen (Ketcham et al., 1963; Schottenfeld, 1979; McCoy et al., 1981). Some of the mechanisms by which alcohol might increase risk include the production of acetaldehyde and free radicals during alcohol metabolism, cytochrome P4502E1 induction, modulation of cell regeneration, promotion or exacerbation of nutritional deficiencies, and alterations of the immune system (Seitz et al., 1998). It is certain that alcohol induces cirrhosis, a factor in 60%–90% of HCCs. Whether alcohol is related to HCC independent of cirrhosis is less clear, though likely to be of little relevance in terms of attributable risk.

Tobacco The effect of cigarette smoking on risk of HCC has been extensively studied, yet the results remain inconclusive. In almost every population that has been examined in more than one study, there are conflicting conclusions concerning the effect of tobacco exposure. Only in Italy (three studies finding no association) have the study results been consistent (Filippazzo et al., 1985; La Vecchia et al., 1988; Chiesa et al., 2000). Positive associations have been reported from Greece (Trichopoulos et al., 1980; 1987; Kuper et al., 2000a; Tzonou et al., 1991) Taiwan (Chen et al., 1991), Japan (Oshima et al., 1984; Tanaka et al., 1988; 1995; Hirayama, 1989; Tsukuma et al., 1990; Goodman et al., 1995; Mukaiya et al., 1998; Mori et al., 2000; Mizoue et al., 2000), China (Lam et al., 1982; Tu et al., 1985; Evans et al., 2002), Egypt (Badawi and Michael, 1999) and United States (Yu et al., 1983, 1988a; Hsing et al., 1990). In addition to Italy, lack of association between smoking and HCC has been reported in Greece (Hadziyannis et al., 1995), Taiwan (Lu et al., 1988), Japan (Shibata et al., 1986; Kono et al., 1987; Tanaka et al., 1992, 1998b; Fukuda et al., 1993), China (Evans et al., 2002), Korea (Pyong et al., 1994; Shin et al., 1996), South Africa (Kew et al., 1985), Spain (Vall Mayans et al., 1990), Egypt (Hassan et al., 2001), Sweden (Hardell et al., 1984), Nigeria (Olubuyide and Bamgboye, 1990) and the United States (Stemhagen et al., 1983; Austin et al., 1986). Of at least 43 studies published between 1983 and 2002, 22 studies reported a positive association, 20 reported no association, and 1 study reported a negative association (Lin et al., 1991). Among the positive studies, some investigators found the association only among HBV(-) persons (Lam et al., 1982; Trichopoulos et al., 1980, 1987; Kuper et al., 2000a), although positive associations have also been reported in HBV(+) cohorts (Tu et al., 1985). Interactions between tobacco and HCV infection have also been reported (Tzonou et al., 1991; Mori et al., 2000). In addition, some reports indicate an effect of smoking only in subsets defined by genetic polymorphisms or levels of other exposures (Yu et al., 1995, 1999b, 1999c, 2000a; Chen et al., 2002). Of two studies that specifically examined smoking in females, both reported positive associations (Tanaka et al., 1995; Evans et al., 2002). The accumulated evidence is compatible with a weak association between smoking and HCC. This association is most likely limited to subsets of the general population that are, as yet, not well defined.

Iron The evidence that higher body iron stores may increase the risk of HCC comes from several sources. Persons with hemochromatosis, an

771

inherited metabolic disorder characterized by hepatic iron loading, have a greatly increased risk of HCC (Niederau et al., 1985). Individuals with another inherited metabolic disorder, porphyria cutanea tarda, also are subject to hepatic iron overload and are at increased risk of HCC, though the risk is not as high as in hemochromatosis (Linet et al., 1999). Studies of persons at high risk of HCC due to exposure to HBV, HCV, and alcohol have also suggested that iron may be a co-factor in HCC. Stevens et al. (1986) reported a positive association between increased iron stores and HCC in an HBV(+) cohort in Taiwan. Risk of iron-related HCC among HCV carriers may be even greater than among HBV carriers in that iron levels may be higher in HCV infection (Kato et al., 2001; Cardin et al., 2002). Attempts to treat HCV infection with phlebotomy have resulted in improvements in alanine aminotransferase levels (Hayashi et al., 1994; Piperno et al., 1996), but not in increased responses to a-interferon therapy (Piperno et al., 1996). However, treatment of HCV carriers with a combination of phlebotomy and low-iron diet has resulted in decreased hepatic 8hydroxy-2¢-deoxyguanosine levels, a marker of DNA damage, with concomitant improvement of hepatitis severity even though HCV titers were unaffected (Kato et al., 2001). Among persons with alcoholic cirrhosis, Ganne-Carrie et al. (2000) reported a significant correlation between hepatic iron content and HCC mortality. African iron overload, first described by Strachan (1929), has been associated with an increased risk of HCC (Gordeuk et al., 1996; Mandishona et al., 1998; Moyo et al., 1998). The consumption of traditional iron-rich beer is a major risk factor for African iron overload, but there is also a genetic component to risk that is unrelated to mutations in the hemochromatosis (HFE) gene (Gordeuk et al., 1992; McNamara et al., 1998).

Exogenous and Endogenous Hormones A relationship between oral contraceptive (OC) use and benign hepatic adenoma is well established, with exposure estimated to increase risk by a factor of between 100 and 500 (Edmondson et al., 1976; Klatskin, 1977; Rooks et al., 1979; Christopherson et al., 1980; Mettlin and Natarajan, 1981; Mays and Christopherson, 1984). The cumulative evidence suggests that there is also a relationship between OC use and HCC. Although cohort studies have not reported significant risks (Colditz et al., 1994; Hannaford et al., 1997), case-control studies have found increased risks with long-term OC use (>5 years) (Henderson et al., 1983; Neuberger et al., 1986; Forman et al., 1986; Palmer et al., 1989; Vall Mayans et al., 1990; Yu et al., 1991a; Hsing et al., 1992a; Tavani et al., 1993; Collaborative MILTS Project Team, 1997). Among the studies that have also examined co-existing viral infection, most have found that risk is limited to OC use in the absence of viral infections (Henderson et al., 1983; Neuberger et al., 1986; Vall Mayans et al., 1990; Collaborative MILTS Project Team, 1997). Consistent with this finding is the failure to find increased risk with OC use in countries where viral infections are endemic (Kew et al., 1990; WHO Collaborative Study of Neoplasia and Steroid Contraceptives, 1989). Based on the cumulative evidence of increased risk of HCC with OC use, IARC concluded that there is sufficient evidence that OCs are carcinogenic in humans (IARC Monographs, 1999). Whether newer, lower-dose OC formulations carry the same risk as older formulations is not known. Studies of intrahepatic cholangiocarcinoma have not revealed an increased risk associated with OC use (Forman et al., 1986; Palmer et al., 1989; WHO Collaborative Study of Neoplasia and Steroid Contraceptives, 1989; Hsing et al., 1992a). Limited investigations of the risk of HCC with injectable progestogen-only contraceptives have been reported. Two studies, conducted in HBV-endemic areas (Kew et al., 1990; WHO Collaborative Study of Neoplasia and Steroid Contraceptives, 1991), found no increased risk of HCC with injectable progestogen use. Studies of post-menopausal estrogen-replacement therapy (Yu et al., 1991a; Tavani et al., 1993; Goodman et al., 1995; Persson et al., 1996) and estrogen-progestogen therapy (Persson et al., 1996) have not revealed increased risks of HCC.

772

PART IV: CANCER BY TISSUE OF ORIGIN

Anabolic androgenic steroids are derivatives of testosterone that were developed to decrease the androgenic side effects of the parent compound. Medically accepted uses for anabolic steroids include treatment of certain types of anemia to stimulate erythropoesis, androgen replacement therapy for hypogonadism, and treatment of osteoporosis (Snyder, 2001). Non-therapeutic use of anabolic steroids was first reported in World War II, when they were administered to German troops to foster aggressiveness (Wade, 1972). Use by athletes began in the 1950s among Russian competitors and has since become widespread in sports (Wade, 1972). Evidence from case reports suggests that anabolic steroids can cause the development of peliosis hepatitis, subcellular changes in hepatocytes, hepatocellular hyperplasia, hepatocellular adenomas, and HCC (Soe et al., 1992). No epidemiologic study of anabolic steroids and HCC has been reported, but the risk of HCC may be dependent on long-term administration (Kosaka et al., 1996). At least one case of liver cancer (combined HCC and cholangiocarcinoma) has been reported in an athlete (Overly et al., 1984). Because the relationship between steroid use and HCC has not yet been examined in an analytic epidemiologic study, IARC has classified the evidence for carcinogenicity in humans as “limited” (IARC Monographs, 1987b). A role for endogenous hormones in the etiology of HCC has been proposed to explain the male excess of HCC in almost all countries (McGlynn et al., 2001). Experimental animal studies have supported a role for androgens in hepatocarcinogenesis (De Maria et al., 2002), but whether androgens are also a risk factor in humans is not clear. In studies conducted in Taiwan, Yu et al. reported a positive association between increased testosterone levels and HCC among men chronically infected with HBV (Yu and Chen, 1993; Yu et al., 2000b, 2001). In contrast, a study carried out in Shanghai found that HBsAg(+) men had significantly higher testosterone levels than HBsAg(-) men and that this relationship explained the testosterone-HCC relationship (Yuan et al., 1995). Prospective data from other geographic areas have not been reported. Among women, several studies found that HCC risk increased with increasing parity (Tzonou et al., 1992; Stanford et al., 1992; La Vecchia et al., 1992, 1993; Hsing et al., 1992b). HBV status may affect the HCC-parity relationship. Tzonou et al. (1992) reported a significant result only among HBV(+) women and the study of Stanford et al. was conducted in HBV-endemic areas. In contrast, one study, conducted in Sweden, reported no association (Lambe et al., 1993). The mechanism by which increased parity would increase HCC risk is not clear, but may be related to the altered estrogen profile in pregnancy (Cole et al., 1976), or to the promotional effect of estrogens on a liver that is chronically infected with HBV. In support of a role for estrogen are recent data showing an increased risk of HCC among women with younger ages at menarche and older ages at menopause (Mucci et al., 2001).

Diet The two diet-related items most consistently associated with increased risk of HCC are alcohol and consumption of foods contaminated with AFB1. In addition, there is some evidence that high iron intake may increase the risk of HCC. Although few other dietary items have been extensively examined in human studies, decreased risks of HCC have been reported in association with selenium, tea, and vegetable consumption. Selenium (Se) has been most widely studied in Qidong, China, a high-risk HCC area. A geographic correlation between high HCC rates and low serum Se levels (Yu et al., 1988b) prompted implementation of several intervention studies. Though methodologic detail is lacking, three Se-supplementation trials reported a decrease in both HCC and the HBV carrier rate in Qidong (Yu et al., 1991b). In support of these results, a cohort study in Taiwan reported that men who developed HCC had lower plasma selenium levels at study enrollment than did men who did not develop HCC (Yu et al., 1999a). In contrast, an ecologic study in China reported no geographic correlation between plasma selenium levels and liver cancer mortality (Hsing et al., 1991).

Several studies in animals have shown that green tea consumption decreased the risk of liver cancer (Klaunig, 1992; Gong et al., 2000; Qin et al., 1997, 2000). Few human studies have been reported (Heilbrun et al., 1986; Ye et al., 1994), but the combined results suggest that green tea may be protective, particularly among persons who are also exposed to alcohol and tobacco (Mu et al., 2003). An inverse association between vegetable consumption and risk of HCC has been supported by at least seven studies (Lam et al., 1982; La Vecchia et al., 1988; Hirayama, 1990; Srivatanakul et al., 1991; Yu and Chen, 1993; Braga et al., 1997; Sauvaget et al., 2003) conducted in Japan, Taiwan, Hong Kong, Thailand, and Italy. Conversely, three studies, two of which were conducted in Greece, reported no association (Fukuda et al., 1993; Hadziyannis et al., 1995; Kuper et al., 2000b). The combined evidence led the World Cancer Research Fund and the American Institute of Cancer Research to conclude that diets high in vegetables probably reduce the risk of HCC (World Cancer Research Fund, 1997).

Schistosomiasis Schistosomiasis, caused by infestation with trematode blood flukes, is endemic in tropical areas of Africa, South America, Asia, and the Caribbean. Five species of schistosomes develop in humans, Schistosoma mansoni, S. japonicum, S. haematobium, S. mekongi, and S. intercalatum, but only three species, S. mansoni, S. japonicum, S. mekongi preferentially infect the liver. The vast majority of schistosomal liver disease is associated with either S. mansoni in eastern Africa and South America, or S. japonicum in Asia. Schistosomal liver disease is due to the host’s response to the deposition of worm eggs in the portal venules, which can progress to advanced schistosomal hepatic fibrosis (Dunn, 2003). The fibrosis results from a chronic granulomatous inflammatory reaction that can block the venules and may alter carcinogen metabolism (Ishii et al., 1994). The evidence linking S. japonicum to HCC led IARC to conclude in 1994 that infection with S. japonicum is “possibly carcinogenic to humans” (IARC Monographs, 1994b). The assessment was based on positive correlation, case-control and cohort studies conducted in Japan, as the data from China have only inconsistently linked schistosomiasis to HCC. More recently reported studies have supported a role for S. japonicum infection in HCC as a cofactor with HBV and HCV infections rather than as a primary hepatocarcinogen (Takemura et al., 1998; Iida et al., 1999; Chou et al., 2003). Although S. mansoni is one of the most common causes of chronic liver disease in eastern Africa (Ghaffar et al., 1991), few studies of its relationship to liver cancer have been reported. Three recent studies concluded that schistosomiasis increases the risk of HCC in persons infected with HBV (Badawi and Michael, 1999); increases the risk in persons infected with HCV (Hassan et al., 2001); and does not increase the risk after adjustment for confounding factors (El-Zayadi et al., 2001). Previous evidence suggested that persons infected with S. mansoni have higher rates of infection with HBV and HCV (Lyra et al., 1976; Halim et al., 1999). Higher viral infection rates may be explained by viral transmission during blood transfusions or parenteral schistosomal therapy (Darwish et al., 1993; Madwar et al., 1989; Frank et al., 2000). Alternatively, persons with schistosomiasis may be more likely to develop chronic viral infections due to their well-documented depressed cell-mediated immune response (Wahib et al., 1998). As a result, persons with schistosomiasis have a tendency to be infected with HBV and HCV for longer periods (Ghaffar et al., 1991), which may increase their risk of developing HCC.

Liver Flukes Human liver flukes of the class Trematoda have a complex life-cycle that requires several intermediate hosts, including snails and fish (IARC Monographs, 1994c). In areas where trematodes are common, individuals are likely to become chronically infected by consuming raw freshwater fish, which contain the infective stage of the flukes. The flukes then mature and live within the smaller intrahepatic bile ducts of their human hosts for up to 10 years (Chapman, 1999). Three

Liver Cancer members of the Trematoda, Opisthorchis viverrini, Clonorchis sinensis, and Opisthorchis felineus have been associated, by varying degrees, with ICC. O. viverrini infects approximately 9 million individuals in Thailand, Laos, and Cambodia, and has been dubbed carcinogenic to humans by IARC (IARC Monographs, 1994c). In high-rate ICC areas, such as the Khon Kaen region of Thailand, infestation with O. viverrini affects one-third of the population and is the major ICC risk factor (Sithithaworn et al., 1994). ICC in this region occurs at a younger age than in low-rate areas and has a more unbalanced sex ratio (M : F = 2.5 : 1) (Srivatanakul, 2001). C. sinensis infects 7 million people in China, Korea, Taiwan, and Vietnam, and has been designated as a probable cause of cancer by IARC (IARC Monographs, 1994c). O. felineus infects 1.5 million individuals in Siberia, Kazakhstan, and Ukraine. Though several studies have linked O. felineus to ICC, the evidence in support of its carcinogenicity remains weaker than for O. viverrini or C. sinensis (IARC Monographs, 1994c). The mechanism by which trematode infestation causes ICC is not well understood, but possible mechanisms include chronic irritation and inflammation of the bile duct epithelium, endogenous nitrosation, activation of drug-metabolizing enzymes, and increased nitric oxide production (Watanapa and Watanapa, 2002).

Thorotrast Thorotrast, a colloidal solution of thorium dioxide (ThO2) and dextrin, was the trade name of an X-ray contrast medium used for cerebral angiography and liver-spleen scans. Between the years 1930 and 1960 as many as 100,000 persons may have been exposed (Srinivasan et al., 1997). After injection, Thorotrast accumulates in the reticuloendothelial system where it continues to emit radioactive alpha-particles at an annual dose of about 25 rad. Fifty-nine percent of an intravenous dose is sequestered in the liver (Kaul, 1995). Thus, Thorotrast-exposed persons receive chronic, low-level, internal exposure throughout the liver, primarily to alpha-particle radiation. In contrast to gamma radiation and X-rays, alpha radiation is weakly penetrating and mainly affects cells very close to the source of radiation. The first report of a link between Thorotrast and angiosarcoma of the liver (ASL) was published in 1947 (MacMahon et al., 1947). Subsequently, follow-up of Thorotrast-exposed cohorts in Germany (van Kaick et al., 1999), Japan (Mori et al., 1999a,b), Portugal (dos Santos Silva et al., 1999), the United States (Travis et al., 2001), Denmark (Andersson et al., 1995), and Sweden (Nyberg et al., 2002) reported a 120-fold increased risk of primary liver cancer, largely due to risks of ASL and ICC. The greatly elevated risks of these tumors may be higher than the risk of HCC because Thorotrast concentrates in Kupffer cells that congregate at the portal triad (Lipshutz et al., 2002). While the risk of developing HCC is not as great, evidence from Japan suggests that the proportion of Thorotrast-associated liver tumors that are HCCs has increased over time (Mori et al., 1999a). The latency period from exposure to development varies between 16 and more than 45 years, depending on the cumulative dose to the liver (Travis et al., 1992). Based on data gathered from the occupational cohorts, it has been estimated that the excess lifetime risk due to Thorotrast is between 260–300 liver cancers per 104 person-Gy (National Academy of Sciences, 1988). Thorotrast also is associated with other liver diseases (notably, cirrhosis) and cancer at other sites (notably gallbladder, extrahepatic bile duct, peritoneum, bone marrow, lung) though to a lesser extent than liver cancer (Lipshutz et al., 2002). With a nearly 50% excess risk of cancer mortality in exposed populations, Thorotrast is one of the most carcinogenic substances in humans studied to date (Travis et al., 2001).

Vinyl Chloride Vinyl chloride (VC), commercially available since the 1920s, has been used since the 1930s to manufacture polyvinyl choride (PVC) resin. The polymerization process to make PVC takes place in a reactor vessel that must be cleaned periodically. In the past, this cleaning

773

was done manually by workers who thereby sustained exposures to VC as high as 1000 ppm (2600 mg/m3). It was among these workers that the association between vinyl chloride exposure and ASL was first reported (Creech and Johnson, 1974; Tabershaw and Gaffey, 1974; Monson et al., 1974; Nicholson et al., 1975; Ott et al., 1975; Duck et al., 1975). Based on these reports, many countries reduced the allowable exposure levels in 1975 to <1–5 ppm (<2.6–13 mg/m3). Follow-up of occupationally exposed cohorts reported 197 cases of VC-associated ASL by 1999 (Kielhorn et al., 2000). Based on the available data, it appears that extremely high levels of exposure are necessary for tumor induction, that the excess risk is associated with some but not all factories under study, that ASL is a rare tumor even among the occupational cohorts, and that no relationship has been reported in the nonoccupational setting. Follow-up studies have indicated that the average latency between exposure and development of ASL is approximately 22 years (Lelbach, 1996) and that VC exposure prior to age 30 may further increase the risk of developing a tumor (Wong et al., 2002). While ASL is the dominant tumor associated with VC exposure (McLaughlin and Lipworth, 1999), some studies have also reported associations with ICC (Forman et al., 1985) and HCC (Evans et al., 1983, Ward et al., 2001, Wong et al., 2002). The associations with these other forms of liver cancer are much weaker than between VC and ASL and may require the presence of additional risk factors, such as HBV (Wong et al., 2002).

Arsenic The association between inorganic arsenic and ASL was first reported in 1957 (Roth, 1957). In a series of 47 autopsies performed on Moselle vintners, Roth reported an unusually high rate of ASL coexisting with nonalcoholic liver cirrhosis. The German vineyard workers had been exposed to arsenic in the 1930s and 1940s via application to grapes of arsenical insecticides and through the consumption of a “house drink” made from the grape skins. Later reports from the United States (Regelson, 1968; Lander et al., 1975; Falk et al., 1981) and Chile (Rennke et al., 1971) reported ASL in association with long-term intake of the arsenic-containing medicinal, Fowler’s solution, and with arsenic in drinking water, respectively. Whether arsenic is also a risk factor for HCC is less clear. A series of ecologic studies done in the Blackfoot Disease (BFD) endemic area of Taiwan reported increased liver cancer rates (Chen et al., 1985; 1986; 1988a; 1988b; 1990; 1992; Wu et al., 1989; Lu and Chen, 1991; Chiou et al., 1995). The very high levels of arsenic in the artesian wells of the BFD endemic area were correlated also with higher rates of cancers of the bladder, kidney, skin, lung and colon. Although the histology of the liver tumors was not determined, the investigators assumed them to be HCCs (Chen et al., 1985). Studies in the BFD endemic area did not account for known HCC risk factors, such as HBV infection, however, and reports from other areas have not found elevated risks of HCC in association with arsenic-contaminated drinking water (Tsuda et al., 1995; Hopenhayn-Rich et al., 1998; Lewis et al., 1999). Although arsenic may promote HCC among HBV- or HCVinfected individuals (Lu and Chen, 1991), that hypothesis has not been tested.

HOST FACTORS Cirrhosis As noted earlier in this chapter, cirrhosis predisposes to HCC. Chronic liver disease of all etiologies is characterized by varying degrees of inflammation and fibrosis. The degree or stage of fibrosis appears to correlate best with prognosis. Although various classifications of fibrosis have been proposed over the past 30 years, there is general agreement that cirrhosis is the most advanced stage of fibrosis. It is characterized histopathologically by nodules (pseudolobules) of hepatocytes surrounded by dense bands of fibrous tissue (Brunt, 2000). In addition, cirrhosis is classified by five variables—serum bilirubin,

774

PART IV: CANCER BY TISSUE OF ORIGIN

serum albumin, ascites, neurological impairment (encephalopathy), and prothrombin time—into three degrees of clinical compensation (Child-Turcotte-Pugh score) A, B, and C (Pugh et al., 1973). Patients with Child A cirrhosis have normal or near normal values for the five criteria and are considered well compensated, those with Child B have moderately abnormal values, and those with Child C are severely impaired. The prospective studies that have been done are usually restricted to Child A, or A and B disease. Because cirrhosis is diagnosed definitively by liver biopsy, relatively few population-based or unbiased prospective studies have been done to demonstrate the rate at which HCC develops in persons with cirrhosis. In lieu of biopsy, various clinical, laboratory, and ultrasonographic criteria have been substituted (Gaiani et al., 1997; Oberti et al., 1997; Imbert-Bismut et al., 2001; Colli et al., 2003), but whether they have sufficient sensitivity or specificity to separate cirrhosis from lower stages of fibrosis is uncertain. An exception is the Dionysos study, in which 70% of the population (6917 people) of two towns in northern Italy between 12 and 65 years of age were screened for the presence of liver disease by questionnaire including medical history, diet, and detailed alcohol intake, by physical examination, and by assay of blood samples for HBV and HCV markers and biochemical evidence of liver damage. Liver biopsies were performed on individuals with evidence of liver disease (Bellentani et al., 1994). The screened population was then followed for 9 years. Overall, 78 people, 1.1% of the population, had cirrhosis; 8 developed HCC during follow-up (Bellentani and Tiribelli, 2001; Tiribelli, personal communication, 2003). At entry, 81 persons (1.2%) were HbsAg positive; 7 had cirrhosis, all Child A. After 9 years of follow-up, none of these individuals developed HCC. In contrast, at entry, 199 people (3.2%) had anti-HCV antibodies; 22 had cirrhosis and 1 had HCC. Over the 9 years of follow-up, 5 of the 22 anti-HCV (+) patients with cirrhosis developed HCC. Two of the five individuals were HCV RNA positive at entry and consumed more than 90 g of alcohol daily (Bellentani et al., 1999); two consumed more than 30 g of alcohol daily; and one person drank more than 30 g of alcohol per day and also had hemochromatosis; 1349 people without HBV or HCV consumed more than 30 g of alcohol per day (21% of the population); 30 had cirrhosis and 1 had HCC. During follow-up, two of these individuals developed HCC. The largest study of outcomes of cirrhosis was conducted in Denmark using the Danish National Registry of Patients to identify greater than 1-year survivors of cirrhosis between 1977 and 1989, linkage with the Danish Cancer Registry to identify cancers that had occurred through 1993 (Sorensen et al., 1998). More than 60% of the cases of cirrhosis had been diagnosed as alcoholic. Among 11,605 individuals with cirrhosis, 199 HCCs were diagnosed compared with 3.3 expected, resulting in a standardized incidence rate of 60. Velazquez et al. (2003) prospectively followed 463 patients with Child A (72%) or B (28%) cirrhosis for the development of HCC. During a mean follow-up of 3.2 years, 38 patients were diagnosed with HCC, yielding a mean annual incidence of 2.95 percent. Other investigators (Colombo et al., 1991; Tsukuma et al., 1993; Ikeda et al., 1993) came to a similar conclusion, that the yearly risk of HCC among patients with cirrhosis is about 3%. Contrary to early reports that cirrhosis was unrelated to ICC, even in the absence of other risk factors, cirrhosis also confers an increased risk of ICC (Sorensen et al., 1998).

Immune Function For the past 20 years, a natural experiment of the effects of immune deficiency on cancer incidence has taken place. The HIV (human immunodeficiency virus) / AIDS (acquired immunodeficiency syndrome) epidemic began in the early 1980s and continues until the present. During this interval many persons became co-infected with HIV and HBV and/or HCV. With the introduction of HAART (highly active anti-retroviral therapy) in the mid-1990s, the degree of immunodeficiency was reduced and survival time lengthened, introducing another variable into the co-infection studies. Using several study

designs, including registry matching (Frisch et al., 2001) and prospective cohort studies (Thio et al., 2002; Darby et al., 1997), the hypothesis that immunodeficiency increases the risk of liver cancer was tested. Frisch et al. (2001) analyzed 302,834 individuals in linked population-based AIDS and cancer registries from 11 geographic sites in the United States and concluded that liver cancer incidence was not increased by HIV/AIDS. Their interpretation, however, requires satisfying three criteria: elevated relative risk (RR) in the period from 60 months before to 27 months after the diagnosis of AIDS (interval 1), elevated RR in the interval from 4 to 27 months post-AIDS (interval 2), and an increasing trend in RR from before to after AIDS onset. HCC was increased in interval 1 (87 cases, RR = 7.7) and interval 2 (23 cases, RR = 3.1), but there was not an increasing trend of RRs from pre- to post-AIDS. The investigators assume that immunodeficiency is more severe in interval 2 and that the degree of immunodeficiency should correlate with cancer incidence. Whether there is a linear relationship of immunodeficiency with cancer is unknown. Therefore, this study cannot be taken as a rejection of the immunodeficiency hypothesis. A cohort study of mortality from liver cancer and liver disease in 4865 males with hemophilia treated with blood products between 1969 and 1985 and followed-up to 1993, revealed a 16.7-fold increased mortality for liver disease and 5.6-fold increase for liver cancer (5 deaths) compared with the general population of the United Kingdom (Darby et al., 1997). Many of these men were co-infected with HIV and HCV, but they could not be reliably identified from death certificates. Other cohort studies reported increased deaths from liver disease among HIV/HBV and HIV/HCV co-infected hemophilia patients or injecting drug users (Yee et al., 2000; Cacoub et al., 2001; Thio et al., 2002), but did not specifically identify HCC or liver cancer among the causes of death. HIV increases viral load and accelerates the progression of liver fibrosis to cirrhosis in both HBV- and HCV-infected individuals (Colin et al., 1999; Gilson et al., 1997; Thio, 2003; Benhamou et al., 1999; Goedert et al., 2001; Gonzalez and Talal, 2003). Both the increase in viral load and the accelerated progression of fibrosis are inversely associated with CD4 T-cell counts. Since cirrhosis is the precursor to HCC in most cases and since survival time has increased greatly in the HAART era, it is possible that co-infection with HIV will have a major impact on HCC risk in the future.

Genetic Susceptibility Genetic susceptibility studies have most often focused on genes that encode enzymes in the aflatoxin B1 (AFB1) detoxification pathways, principally the glutathione-S-transferases (GST) and epoxide hydrolases (EPHX). Although the GSTs have somewhat overlapping specificities, in vitro studies of human hepatocytes have suggested that glutathione-S-transferase mu 1 (GSTM1) is the most critical of GSTs in AFB1 conjugation in humans. The results of studies of GSTM1 genotype and HCC, however, have been inconsistent, with an almost equal number of studies supporting (Chen et al., 1996; Yu et al., 1999b; Bian et al., 2000; Omer et al., 2001; Deng et al., 2001) and not supporting (Yu et al., 1995; Hsieh et al., 1996; Sun et al., 2001; McGlynn et al., 2003) the association. The collective results suggest that if there is an association between GSTM1 and HCC, it is weak. Glutathione-S-transferase theta 1 (GSTT1) and glutathione-Stransferase pi 1 (GSTP1) have been examined less frequently than GSTM1, but also with mixed results. Three studies in Taiwan have reported increased risk of HCC in association with a GSTT1 null genotype (Chen et al., 1996; Yu et al., 1999b; Sun et al., 2001), but Tiemersma et al. (2001) in Sudan and McGlynn et al. (2003) in China reported no association. Similarly, GSTP1 was found to be associated with HCC in Taiwan (Chen et al., 2002), but not in China (McGlynn et al., 2003). Among the other GSTs (GSTM2, GSTM3, GST12, GSTA1, GSTA4, GSTT2, GSTA4) examined in the study of McGlynn et al. (2003), only glutathione-S-transferase alpha 4 (GSTA4) was found to be significantly associated with HCC risk. Microsomal epoxide hydrolase, encoded by EPHX1, catalyzes the hydrolysis of reactive epoxide intermediates, thereby favoring their

Liver Cancer elimination. Polymorphisms in exons 3 and 4 have been examined in HCC. McGlynn et al. (1995) reported a significant association between HCC and the 113His allele of the exon 3 polymorphism in a Chinese population. This finding was not supported by Wong et al. (2000), studying a Scottish population. The investigators also found no relationship between HCC and an EPHX1 polymorphism at residue 139 of exon 4. Tiemersma et al. (2001) did not find a significant association between either polymorphism and HCC in a Sudanese population. Similarly, no association was found between either polymorphism and risk in a subsequent study of McGlynn et al. (2003). Though the EPHX1 studies are not numerous, the combined results do not support a role for EPHX1 and HCC. Another member of the epoxide hydrolase family, EPHX2, has only been examined in one study (McGlynn et al., 2003). The results of that study, adjusted for multiple comparisons, support a role for EPHX2 in determining risk of HCC. N-acetyltransferase (NAT) is involved in both phase I and phase II metabolism of carcinogenic aromatic amines derived from a number of sources including tobacco smoke and cooked meat. Agundez et al. (1996), Yu et al. (2000a), and Huang et al. (2003) examined polymorphisms in NAT1 and NAT2 in HCC. Studying a Spanish population, Agundez et al. (1996) reported associations between NAT2 slow acetylation genotypes and HCC among patients who were not infected with either HBV or HCV. In contrast, Yu et al. (2000a) in Taiwan reported an association between NAT2 rapid acetylation genotypes and HCC among HBV(+) persons who smoked. Also examining a Taiwanese population, Huang et al. (2003) reported a trend of increasing risk of HCC among NAT2 rapid acetylators who consumed red meat. No support for a relationship between NAT1 and HCC was reported by Yu et al. (2000a). Because alcohol is a recognized risk factor for HCC, polymorphisms in the alcohol metabolizing enzymes, alcohol dehydrogenase 2 (ADH2) and aldehyde deydrogenase 2 (ALDH2), have been examined for association with HCC (Shibata et al., 1998; Takeshita et al., 2000). Thus far, no association has been reported. A number of cytochrome p450 (CYP) enzymes have also been examined in relationship to HCC, though rarely in more than a few studies. Among the CYP enzymes, CYP2E1 has received the most attention because it metabolizes alcohol in the non-alcohol dehydrogenase pathway. Of five studies examining CYP2E1, three supported an association with at least one polymorphism (Yu et al., 1995; Ladero et al., 1996; Silvestri et al., 2003) and two did not (Lee et al., 1997; Wong et al., 2000). Due to the male predominance in risk of HCC, investigators have examined polymorphisms in hormone-related enzymes encoded by androgen-receptor (AR), 5-alpha reductase (SRD5A2), and cytochrome p450c 17 alpha (CYP17) (Yu et al., 2000b; Yu et al., 2001; Yeh et al., 2002). Significant associations were reported for all three loci and HCC, suggesting that variability in androgen-signaling may be associated with the risk of HCC.

Hemochromatosis The majority of iron-overload disease in persons of Northern European ancestry is due to hereditary, or type 1, hemochromatosis (HH). Most HH is associated with mutations in the HFE gene on chromosome 6. Two less common forms of hemochromatosis also exist. Type 2, or juvenile, hemochromatosis, is a more severe disorder than HH and is characterized by rapid iron-loading and disease presentation in the second decade of life. The disorder has been linked to a region on chromosome 1q that has been named HFE2 (Roetto et al., 1999). Type 3 hemochromotosis has a very similar presentation to HH but is associated with mutations in the transferrin receptor 2 gene (TFR2) on chromosome 7 (Camaschella et al., 2000). Hemochromatosis is an autosomal recessive disorder that is characterized by excessive dietary iron absorption and subsequent deposition in the parenchymal cells of the liver, pancreas, heart, joints, and pituitary gland (Powell et al., 1994). Prior to the cloning of the HFE gene (Feder et al., 1996), studies had linked the condition to the HLA region on chromosome 6 (Simon et al., 1976) and had estimated the

775

carrier rate at 10% among populations of Northern European ancestry. With the identification of the HFE gene came the demonstration that two missense mutations, C282Y and H63D, accounted for the great majority of HH (Feder et al., 1996; Jouanolle et al., 1996; Jazwinska et al., 1996; Beutler et al., 1996; The UK Haemochromatosis Consortium, 1997). Homozygosity for the C282Y mutation is responsible for between 50% and 80% of HH, consistent with the observation that the C282Y mutation causes a greater loss of HFE protein function than does the H63D mutation (Feder et al., 1997). Approximately 5% of HH occurs among H63D heterozygotes, another 5% among those with the C292Y/H63D compound heterozygous state, and 7% among individuals who carry neither mutation (Hanson et al., 2001). Numerous other allelic variants of HFE have been reported, at least nine of which are missense mutations that result in amino acid substitutions (Pointon et al., 2000). Only the C282Y and H63D mutations, however, have been consistently associated with clinically manifested HH. The relative risk of liver cancer in persons with HH has been estimated to be 200 (Bradbear et al., 1985; Niederau et al., 1985). Although the risk is greatest for HCC, persons with HH are also at increased risk of ICC and combined hepatocholangiocarcinoma (Morcos et al., 2001). The risk of HCC in HH is not uniformally distributed however, but is increased in the presence of a variety of cofactors including male sex, age greater than 50 years, drinking, smoking, and HBV and HCV infections (Deugnier et al., 1993; Fargion et al., 1994). Even though the vast majority of HCC in persons with HH develops in cirrhotic livers, HCC in non-cirrhotic livers has been reported (Goh et al., 1999). Despite these observations, the majority of patients who are diagnosed and treated with reduction of iron stores prior to suffering irreversible liver damage have a normal life expectancy (Niederau et al., 1996). Individuals who are heterozygous for the HFE mutations have higher iron stores than individuals who carry no mutant HFE alleles (Borecki et al., 1989; Bulaj et al., 1996; Datz et al., 1998). Whether heterozygous individuals are at increased risk of HCC is not known (Nelson et al., 1995).

Other Inherited Metabolic Diseases In addition to hemochromatosis, several other inherited metabolic diseases predispose to hepatocellular carcinoma. Notable among these disorders are a1-antitrypsin deficiency, tyrosinemia, and several porphyrias. Porphyrias are the result of enzyme deficiencies in the heme biosynthesis pathway. Two types of porphyria, prophyria cutanea tarda (PCT) and acute intermittent porphyria (AIP), have been associated with increased risk of HCC (Hardell et al., 1984; Bjersing et al., 1996; Linet et al., 1999). The more common porphyria, PCT, is the result of deficient activity of hepatic uroporphyrinogen decarboxylase and is characterized by mild to moderate iron overload, which is alleviated by phlebotomy (Lundvall and Weinfeld, 1968). PCT has been associated with HCV infection (Tsukazaki et al., 1998; El-Serag et al., 2002) and heterozygosity for either of the two major mutations in the hemochromatosis (HFE) gene, C282Y and H63D, depending on the population studied (Roberts et al., 1997; Bonkovsky et al., 1998; Sampietro et al., 1998; Hift et al., 2002). AIP, common in Sweden, is characterized by deficient activity of porphobilinogen deaminase (PBGD) due to a G593A mutation in the PBGD gene (Bjersing et al., 1996). A large prospective study of porphyrias in Sweden and Denmark reported the risk of HCC to be higher among persons with AIP (SIR = 70.4) than among persons with PCT (SIR = 21.2), though the confidence intervals overlapped (Linet et al., 1999). a1-antitrypsin (AAT) is the main proteinase inhibitor (Pi) in serum and is encoded by the AAT gene on chromosome 14 (Billingsley et al., 1993). A GÆA substitution in exon 6, referred to as the Z mutation, results in the replacement of a glutamic acid with a lysine (Glu342Lys) and causes a conformational change in the AAT molecule. As a consequence, only 15% of AAT can move from hepatocytes into the circulation (Qu et al., 1997). Homozygosity for the Z mutation (PiZZ) is most common among persons of Northern European

776

PART IV: CANCER BY TISSUE OF ORIGIN

descent, among whom the incidence is between 1 in 1600 to 1 in 2000 (Sveger, 1976). Homozygous a1-antitrypsin deficiency is the most common genetic cause of liver disease in children (Perlmutter, 1998). PiZZ adults, particularly males, are at increased risk of both cirrhosis (OR = 7.8) and HCC (OR = 20) (Eriksson et al., 1986). Recent evidence has suggested that individuals with only one Z allele (PiZ) are also are at increased risk of liver cancer (Zhou et al., 2000). Hereditary tyrosinemia type I is a pediatric autosomal recessive condition that occurs due to a deficiency of fumarylacetoacetase, an enzyme in the tyrosine degradation pathway. As a result of the deficiency, hepatotoxic metabolites accumulate and may lead to HCC in over a third of affected children (Rezvani, 2000).

Diabetes Mellitus Among 20 studies of diabetes mellitus and HCC reported between 1970 and 2002, 6 of 8 cohort studies and 9 of 12 case-control studies reported a positive association. Most of the studies were conducted in low-risk areas for HCC, although studies from Japan have also reported a positive association (Sasaki et al., 1996; Shibata et al., 1998). Many of the studies in individuals with diabetes also noted a relationship between diabetes and cirrhosis (Adami et al., 1991; Sasaki et al., 1994; 1996; Koskinen et al., 1998; Weiderpass et al., 2001). As insulin resistance is known to be associated with cirrhosis, it is possible that the associations are a consequence of the cirrhotic process. Cohort studies, which have found increased risks of HCC among diabetics and persons with hyperinsulinemia, suggest that diabetes usually precedes the development of cirrhosis and HCC (Adami et al., 1996; Wideroff et al., 1997, Balkau et al., 2001). In support of these observations are studies demonstrating that hepatic steatosis (fatty liver) is common among persons with type II diabetes (Kalk, 1960). Similarly, it has been suggested that the diabetes-HCC relationship is a result of HCV infection (Mason et al., 1999), which may impair glucose and insulin metabolism (Petrides et al., 1989). HCV status was determined in four studies that examined the diabetes-HCC relationship. Two studies reported that diabetes was not independently related to HCC (Hadziyannis et al., 1995; El-Serag et al., 2001), while two reported independent effects of diabetes on HCC (Lagiou et al., 2000; Hassan et al., 2002). Further study of the relationship will be needed as the incidence of diabetes continues to grow in most developed countries of the world.

Nonalcoholic Steatohepatitis In 1980, Ludwig et al. coined the term nonalcoholic steatohepatitis (NASH) to describe a condition among nondrinkers, characterized by morphologic evidence of fatty changes in the liver with lobular hepatitis (Ludwig et al., 1980). Though subsequent definitions have varied, Brunt et al. (1999) proposed that NASH be defined by the presence of steatosis, inflammation, hepatocellular degenerative changes, and variable fibrosis. Now recognized as the most severe form of nonalcoholic fatty liver disease (NAFLD), NASH is estimated to be the third most common liver disorder in North America (Byron and Minuk, 1996) and the most common in Australia and New Zealand (Farrell, 2003). While the majority of the patients described in the initial report of Ludwig et al. (1980) were female, subsequent reports have found that NASH occurs equally among males and females (Bacon et al., 1994). Conditions frequently found in association with NASH include insulin resistance, impaired glucose tolerance, type II diabetes mellitus, hypertriglyceridemia, age greater than 45 years, and obesity, particularly central obesity (Farrell, 2003). In addition, elevated body iron stores have been reported to be common among NASH patients (Ludwig et al., 1997; Bacon et al., 1994) and may be related to mutations in the hemochromatosis gene (George et al., 1998; Bonkovsky et al., 1999). Evidence for a possible genetic component to risk has come from a family study that found an unexpectedly high occurrence of NASH-related conditions in relatives of NASH probands (Struben et al., 2000). Although some early reports suggested that NASH was a nonprogressive disorder, it is now recognized that severe fibrosis occurs

in 15%–50% of NASH patients and cirrhosis in 7%–25% (Bugianesi et al., 2002). It has also been suggested that “burned out” NASH is the cause of many cases of cryptogenic cirrhosis because many of the same co-morbid conditions are equally present in NASH and cryptogenic cirrhosis (Powell et al., 1990; Caldwell et al., 1999; Poonawala et al., 2000). While the incidence of HCC is increased in most forms of cirrhosis (Schafer and Sorrell, 1999), the risk of HCC among patients with NASH is not yet clearly defined. However, several case reports and case-series of HCCs arising in NASH patients have been reported (Cotrim et al., 2000; Zen et al., 2001; Shimada et al., 2002). In addition, a case-control study in Italy (Bugianesi et al., 2002) linked HCC to NASH-like cryptogenic cirrhosis and a record-linkage study in Denmark (2003) reported that patients with fatty liver disease were at significantly increased risk of developing liver cancer (Sorensen et al., 2003). A more definitive estimate of HCC risk awaits prospective studies of NASH patients.

Primary Sclerosing Cholangitis Primary sclerosing cholangitis (PSC) is a chronic cholestatic liver disease characterized by inflammation, destruction, and fibrosis of the bile ducts, which frequently results in biliary cirrhosis (Lindor and Larusso, 2003). Liver transplantation is the sole curative treatment option (Boberg and Schrumpt, 1998). PSC more commonly affects men than women and very often occurs in individuals with coexisting inflammatory bowel disease, usually ulcerative colitis (Chapman, 1999). A recent study in a largely white population in Minnesota reported the incidence of PSC to be approximately 0.9/100,000 person-years and found that the incidence had increased between the years 1976 and 2000 (Bambha et al., 2003). The authors note, however, that the increase is likely due to improved diagnosis by endoscopic retrograde cholangiopancreatography (ERCP). In developed countries, PSC is the single best described risk factor for ICC, conferring a risk 30 times greater than that of the general population (Rosen et al., 1991). Estimates of the prevalence of ICC among PSC patients have varied, but a recent large Scandinavian study reported the prevalence to be 13% (Bergquist et al., 2002). The true prevalence may be even greater, as many ICCs are diagnosed only at laparotomy for liver transplantation (Chapman, 1999). The identification of PSC patients who are at high risk of developing ICC has proven to be difficult, however, particularly as the duration of disease does not correlate with risk (Chapman, 1999). Factors that may increase the risk of ICC in PSC patients include ulcerative colitis, cirrhosis, and smoking (Lindor and Larusso, 2003). PSC has also been reported to be associated with the development of HCC (Harnois et al., 1997), though the risk of HCC is considerably less than that of ICC.

PATHOGENESIS The molecular pathogenesis of HCC, defined as the specific genomic alterations that drive its development, is not understood (Thorgeirsson and Grisham, 2002; Tornillo et al., 2002). Intracellular replication of either HBV or HCV does not kill hepatocytes. Liver damage is caused primarily by attack of the host immune system on virus-infected hepatocytes resulting in cell death by apoptosis and necrosis. When hepatocytes die, division of other mature hepatocytes replaces them. In a remarkable study, Summers et al. (2003) demonstrated that in an acute infection with woodchuck hepatitis virus (WHV), regeneration is accomplished by division of infected, not uninfected, hepatocytes. Furthermore, they calculated that between 70% and 100% of the liver is regenerated during recovery from an acute, transient infection. With chronic infection, the immune response is inadequate to clear the infection and there are continuing rounds of cell death and regeneration. This process, in turn, results in multiple opportunities for mutation. Hepatitis B virus DNA integrates into the host cell genome and integrated HBV sequences are present in a clonal pattern in human HCCs, indicating that integration is a very early event in tumor formation (Shafritz et al., 1981). One hypothesis that was exhaustively tested was that HBV acted as an insertional mutagen of a specific host gene.

Liver Cancer Intense searches for specific HBV DNA insertion sites revealed the opposite, random integrations throughout the genome (Seeger and Mason, 2000). The hypothesis that HBV contained a gene (oncogene) that was capable of transforming hepatocytes into a malignant phenotype has also been tested. Three of the four HBV genes are clearly not transforming, but the HBV X gene in some experimental systems is transforming (Feitelson et al., 2002). Nevertheless, the majority view is that the X gene product may have an important promotional, but not a critical role in HBV-induced hepatocarcinogenesis (Seeger and Mason, 2000). Hepatitis C virus is an RNA virus that is not copied into DNA and does not integrate. It is possible that one or more HCV gene products promote genomic changes that lead to malignant transformation, but this is unclear (Block et al., 2003). One direction of current research is to search for genomic lesions in high-grade dysplastic nodules, which may be the immediate precursors of HCC. Such studies have found many fewer chromosomal aberrations than in HCC (Thorgeirsson and Grisham, 2002). Tornillo et al. (2002) found only deletions of chromosome 8p and gains of 1q in high-grade dysplastic nodules. Mutations of p53 are primarily late events, even those associated with AFB1 exposure (Seeger and Mason, 2000). Epigenetic changes in dysplastic nodules (precursor neoplastic lesions) are likely to be important in the carcinogenic process, but the specific molecular events critical to the process are unknown.

777

It is unclear how much HCC incidence would be affected by reducing exposure to AFB1. In China, the government has implemented a strategy of replacing corn with rice as the primary staple food in high HCC incidence areas; HCC incidence has declined modestly, but whether this is due to the dietary change or some other factor is uncertain. Furthermore, the attributable risk of AFB1 exposure is much less than that of HBV infection, indicating that vaccination against HBV would have a more profound effect than dietary manipulation (Ming et al., 2002). Limiting fungal contamination of crops either pre- or post-harvest can reduce AFB1 exposure. Approaches may involve low technology post-harvest measures to limit fungal growth or genetic engineering of crops to be resistant to fungal infection or toxin biosynthesis (Gong et al., 2003). The Dionysos study defined a cutoff level of daily alcohol consumption of 30 g, above which the risk of cirrhosis rose sharply (Bellentani and Tiribelli, 2001). This and other studies cited earlier demonstrated a strong interaction of HCV infection with alcohol consumption. Although treatment of alcohol abuse is notoriously difficult, physicians’ advice and self-help booklets can cause problem drinkers to reduce their consumption by 25% (Jones, 2001). Therefore, a public health strategy of reducing alcohol intake to one to two drinks a day rather than abstention might significantly reduce the risk of cirrhosis and HCC.

Screening and Early Detection PREVENTIVE MEASURES Primary Prevention Chronic HBV infection, which is the cause of the largest proportion of HCCs in the world, is 90% preventable with proper use of the hepatitis B vaccine (Kane, 2003). The vaccine, approved in the United States in 1982, has remarkably few adverse effects. Studies from Taiwan, where universal immunization of newborns was introduced in 1984, have shown at least a 50% reduction in the incidence of HCC among adolescents (Chang et al., 1997; Lee and Ko, 1997; Lee et al., 2003). As the vaccinated population moves into their late teens and early 20s, it will be important to see whether the magnitude of protection is maintained. The challenge now is to expand hepatitis B vaccination coverage to the populations at greatest risk of HBV infection and HCC. In 1992, the WHO set a goal for all countries to integrate hepatitis B vaccination into their universal childhood vaccination programs by 1997. That goal was not achieved, but significant progress is being made. By 2001, 126 of the 191 WHO member states had universal infant or childhood hepatitis B vaccination programs (CDC, 2003b). Through the efforts of the Global Alliance for Vaccines and Immunization (GAVI), the cost of vaccine has been reduced from $100 to $1.00 per pediatric dose. The developed countries of the world can already afford universal vaccination of newborns and now vaccination programs are being introduced into the 72 poorest countries (Kane, 2003). From studies done in several high-risk countries, we can predict that vaccination programs will reduce the prevalence of chronic HBV infection from between 8% and 20% to less than 2% (Mahoney and Kane, 1999) and this should lead to major decreases in the incidence of HCC in the coming decades. As of this writing, there is no vaccine available for the prevention of infection with HCV. Nevertheless, the incidence of HCV infection has been declining in the United States and other developing countries since 1992 because of screening of blood and organ donors and reduction of transmission among injecting drug users (IDUs). During the 1980s, an estimated 230,000 new HCV infections occurred each year. Since then, the annual number of new infections has declined by more than 80% to about 30,000 in 2000 (CDC, 2003a). With the virtual elimination of contaminated transfused blood, injecting drug use is the major source of new HCV infections in the United States, accounting for 60% of such infections in 2000. The risk of transmission of HCV can be greatly reduced by counseling IDUs about the risks of sharing drug solutions, syringes, and needles; providing sterile disposable needles with instructions for their destruction following use; and enrolling IDUs in substance abuse treatment programs (CDC, 2003a).

Clinicians commonly screen patients who are chronically infected with HBV or HCV, with or without cirrhosis, with annual or semiannual serum alpha-fetoprotein (AFP) level detection and ultrasonography (US) of the liver to detect small, incident HCCs. Whether such screening confers a survival advantage is still uncertain. Randomized clinical trials are difficult to justify ethically and those that have been done have either not shown a mortality benefit or have had methodological problems. There are several components to screening for HCC that limit its value including: 1. sensitivity and specificity of AFP 2. quality and maintenance level of the US equipment, particularly in developing countries 3. skill of the US operator in identifying small tumors 4. availability of high quality surgery 5. cost (Lok and McMahon, 2001; Gebo et al., 2002). As with all tests, the sensitivity of AFP can be increased to 100% if the cut-off level is set low enough, but at the price of loss of specificity. The most commonly used cut-off level for HBV carriers is 20 ng/ml, which yields a sensitivity between 50% and 75% and a specificity of 90% (Lok and McMahon, 2001). For patients with chronic hepatitis C infection, threshold values of 10–19 ng/ml have been used and have yielded sensitivities of 45%–100% and specificities of 70%–95% (Gebo et al., 2002). The sensitivity of US to detect HCCs less than 3 cm in diameter among HBV-infected individuals has ranged from 68%–87%, but the specificity has been 80% or less. Studies of US among persons with chronic HCV have shown higher specificities (>95%), but lower sensitivities (Gebo et al., 2002). The results of surgery for HCC are highly dependent on the experience and skill of the surgeon and the quality of surgical facilities. Even in the best of hands, however, the recurrence rate and/or the development of new tumors are high. Among highly selected patients, however, 5-year survival rates higher than 50% have been reported (Llovet et al., 2003). Orthotopic liver transplantation (OLT) has been used successfully to treat patients with decompensated cirrhosis caused by HBV, HCV, or alcohol and small solitary HCCs (Figueras et al., 1997). One-year survival is about 80% and 5-year survival 65%–70% (Wright, 2002; Fontana and Lok, 2003; Chopra and Bonis 2003, Curley et al., 2003). The problem is that the number of patients who want or need an OLT far exceeds the supply of donor organs. Presently, about 18,000 patients are on waiting lists, but only 5000 patients are transplanted each year (Wright, 2002). Furthermore, the first-year cost of an OLT is $80,000 to $200,000. Therefore, OLT as a preventive measure can only be narrowly applied. Thus, periodic

778

PART IV: CANCER BY TISSUE OF ORIGIN

screening with AFP and US is the best available means for secondary prevention, but it is of limited value.

Chemoprevention A number of chemopreventive agents have been examined in HCC, though many have not moved beyond animal studies. In general, the aim of most chemoprevention has been to prevent cirrhosis because prevention of HCC in non-HCV–related cirrhosis has been unachievable. Of the agents that have moved into clinical trials, a-interferon has been the most extensively investigated. Among HCV-infected individuals, the efficacy of a-interferon in reducing HCC risk has been clearly demonstrated among drug responders (Yoshida et al., 1999; Nishiguchi et al., 2001; Ikeda et al., 2001; Toyoda et al., 2001). Encouragingly, the risk of HCC has also decreased among ainterferon–treated patients who have not attained a sustained response when compared with untreated individuals (Kasahara et al., 1998; Papatheodoridis et al., 2001; Hayashi et al., 2002). Unfortunately, ainterferon appears to have little effect in diminishing HCC risk among HBV-infected individuals (Camma et al., 2001; Yuen et al., 2001). Whether therapy with nucleoside inhibitors like Lamivudine and Adefovir dipivoxil will reduce the risk of HCC among HBV-infected patients is unknown, but the improvement in liver histopathology with prolonged treatment with these drugs is encouraging (Lok and McMahon, 2002). In addition to a-interferon, glycyrrhizin, an aqueous extract of licorice root, has been reported to decrease the risk of HCC in HCVinfected individuals (Arase et al., 1997; Kumada 2002). Unlike ainterferon, glycyrrhizin appears to be an immune modulator rather than an antiviral agent (Bean, 2002). Medicinal ginseng (GinsengHCC Chemopreventive Study Osaka Group, 2001) and acyclic retinoid (Okuno et al., 2002) are also being tested for HCCpreventive capability among HCV-infected Japanese patients. Two agents that are targeted against aflatoxin-related liver damage are oltipraz and chlorophyllin. Oltipraz works to alter phase I and II metabolism of aflatoxin, while chlorophyllin acts by forming complexes with aflatoxin, which limit aflatoxin bioavailabilty (Guyton and Kensler, 2002). Though both have demonstrated promising results in reducing the levels of aflatoxin markers, neither has yet been demonstrated to decrease the risk of HCC in aflatoxin-exposed populations. Chemopreventive agents that have not yet reached HCC clinical trials include cyclooxygenase-2 inhibitors (Hu, 2002), S-adenosyl-Lmethionine (Pascale et al., 1995), curcumin (Chuang et al., 2000), a 5a-reductase inhibitor (Maruyama et al., 2001), vitamin E (Kakizaki et al., 2001), vitamin D (Basak et al., 2001), green tea (Qin et al., 2000), selenium (Yu et al., 1997b), and a number of herbal extracts.

FUTURE DIRECTIONS Because the major etiologic factors for HCC have been identified, and universal hepatitis B vaccination is being implemented globally, the unsolved problems are more limited than they were a few years ago. A vaccine to prevent infection with HCV is not available. A new strategy is needed to circumvent the antigenic variability of HCV between individuals and within the same individual. Even though multiple genomic changes have been identified in malignant and pre-malignant hepatocytes, the molecular pathogenesis of HCC is not understood. The new technologies of micro-arrays and proteomics must be applied to small HCCs and dysplastic nodules to define the molecular pathways to malignant transformation of hepatocytes. Identification of such pathways could reveal new targets for chemoprevention of HCC in high-risk individuals and possibly therapies for established HCCs. References Adami HO, Chow WH, Nyren O, et al. 1996. Excess risk of primary liver cancer in patients with diabetes mellitus. J Natl Cancer Inst 88:1472–1477. Adami HO, McLaughlin J, Ekbom A, et al. 1991. Cancer risk in patients with diabetes mellitus. Cancer Causes Control 2:307–314.

Agundez JA, Olivera M, Ladero JM, et al. 1996. Increased risk for hepatocellular carcinoma in NAT2-slow acetylators and CYP2D6-rapid metabolizers. Pharmacogenetics 6:501–512. Ahrendt SA, Nakeeb A, Pitt HA. 2001. Cholangiocarcinoma. Clin Liver Dis 5:191–218. Aizawa Y, Shibamoto Y, Takagi I, Zeniya M, Toda G. 2000. Analysis of factors affecting the appearance of hepatocellular carcinoma in patients with chronic hepatitis C. Cancer 89:53–59. Alter HJ, Seef LB. 2000. Recovery, persistence, and sequelae in hepatitis C infection: A perspective on long-term outcome. Semin Liv Dis 20:17–35. Alter MJ. 1999. Hepatitis C virus infection in the United States. J Hepatol 31: S-1:88–91. Andersson M, Carstensen B, Storm HH. 1995. Mortality and cancer incidence after cerebral arteriography with or without Thorotrast. Radiat Res 142:305–320. Arase Y, Ikeda K, Murashima N, et al. 1997. The long term efficacy of glycyrrhizin in chronic hepatitis C patients. Cancer 79:1494–1500. Asplin F, Carnaghan R. 1961. The toxicity of certain groundnut meals for poultry with special references to their effect on ducklings and chickens. Vet Rec 73:1215–1219. Austin H, Delzell E, Grufferman S, et al. 1986. A case-control study of hepatocellular carcinoma and the hepatitis B virus, cigarette smoking and alcohol consumption. Cancer Res 46:962–966. Bacon BR, Farahvash MJ, Janney CG, Neuschwander-Tetri BA. 1994. Nonalcoholic steatohepatitis: An expanded clinical entity. Gastroenterology 107:1103–1109. Badawi AF, Michael MS. 1999. Risk factors for hepatocellular carcinoma in Egypt: The role of hepatitis-B viral infection and schistosomiasis. Anticancer Res 19:4565–4569. Balkau B, Kahn HS, Courbon D, Eschwege E, Ducimetiere P. 2001. Hyperinsulinemia predicts fatal liver cancer but is inversely associated with fatal cancer at some other sites: The Paris Prospective Study. Diabetes Care 24:583–849. Bambha K, Kim WR, Talwalkar J, et al. 2003. Incidence, clinical spectrum, and outcomes of primary sclerosing cholangitis in a United States community. Gastroenterology 125:1364–1369. Basak R, Bhattacharya R, Chatterjee M. 2001. 1 alpha, 25-Dihydroxyvitamin D3 inhibits rat liver ultrastructural changes in diethylnitrosamine-initiated and phenobarbital promoted rat hepatocarcinogenesis. J Cell Biochem 81:357–367. Bean P. 2002. The use of alternative medicine in the treatment of hepatitis C. Am Clin Lab 21:19–21. Beasley RP. 1988. Hepatitis B virus, the major etiology of hepatocellular carcinoma. Cancer 61:1942–1956. Beasley RP, Lin CC, Hwang LY, Chien CS. 1981. Hepatocellular carcinoma and hepatitis B virus. A prospective study of 22, 707 men in Taiwan. Lancet 2:1129–1133. Bellentani S, Pozatto G, Saccoccio G, et al. 1999. Clinical course and risk factors of hepatitis C related liver disease in the general population: Report from the Dionysos study. Gut 44:874–880. Bellentani S, Tiribelli C. 2001. The spectrum of liver disease in the general population: Lesson from the Dionysos study. J Hepatol 35:531–537. Bellentani S, Tiribelli C, Saccoccio G, et al. 1994. Prevalence of chronic liver disease in the general population of northern Italy: The Dionysos Study. Hepatology 20:1442–1449. Benhamou Y, Bochet M, Di Martino V, et al. 1999. Liver fibrosis progression in human immunodeficiency virus and hepatitis C virus coinfected patients. The Multivirc Group. Hepatology 30:1054–1058. Bergquist A, Ekbom A, Olsson R, et al. 2002. Hepatic and extrahepatic malignancies in primary sclerosing cholangitis. J Hepatol 36:321– 327. Beutler E, Gelbart T, West C, et al. 1996. Mutation analysis in hereditary hemochromatosis. Blood Cells Mol Dis 22:187–194. Bian JC, Shen FM, Shen L, et al. 2000. Susceptibility to hepatocellular carcinoma associated with null genotypes of GSTM1 and GSTT1. World J Gastroenterol 6:228–230. Billingsley GD, Walter MA, Hammond GL, Cox DW. 1993. Physical mapping of four serpin genes: alpha-1-antitrypsin, alpha-1-antichymotrypsin, corticosteroid-binding globulin, and protein C inhibitor, within a 280-kb region on chromosome 14q32.1. Am J Hum Genet 52:343–353. Bjersing L, Andersson C, Lithner F. 1996. Hepatocellular carcinoma in patients from northern Sweden with acute intermittent porphyria: morphology and mutations. Cancer Epidemiol Biomarkers Prev 5:393–397. Block TM, Mehta AS, Fimmel CJ, Jordan R. 2003. Molecular viral oncology of hepatocellular carcinoma. Oncogene 22:5093–5107. Blount W. 1961. Turkey “X” disease. Turkeys (Journal of the British Turkey Federation) 9:55–58.

Liver Cancer Boberg KM, Schrumpf E. 1998. Treatment of primary sclerosing cholangitis. In: Autoimmune Liver Diseases. 2nd ed. Krawitt EL, Wiesner RH, Nishioka M, eds. Amsterdam: Elsevier Science B.V, pp. 529–551. Bonkovsky HL, Jawaid Q, Tortorelli K, et al. 1999. Non-alcoholic steatohepatitis and iron: Increased prevalence of mutations of the HFE gene in nonalcoholic steatohepatitis. J Hepatol 31:421–429. Bonkovsky HL, Poh-Fitzpatrick M, Pimstone N, et al. 1998. Porphyria cutanea tarda, hepatitis C, and HFE gene mutations in North America. Hepatology 27:1661–1669. Borecki IB, Rao DC, Le Mignon L, Yaouanq J, Simon M, Lalouel JM. 1989. Genetic hemochromatosis: Distribution analysis of six laboratory measures of iron metabolism. Am J Med Genet 34:435–441. Bradbear RA, Bain C, Siskind V, et al. 1985. Cohort study of internal malignancy in genetic hemochromatosis and other chronic non-alcoholic liver diseases. J Natl Cancer Inst 75:81–84. Braga C, La Vecchia C, Negri E, Franceschi S. 1997. Attributable risks of hepatocellular carcinoma in northern Italy. Eur J Cancer 33:629–634. Brechot C, Degos F, Lugassy C, et al. 1985. Hepatitis B virus DNA in patients with chronic liver disease and negative tests for hepatitis B surface antigen. N Engl J Med 312:270–276. Bressac B, Kew M, Wands J, Ozturk M. 1991. Selective G to T mutations of p53 gene in hepatocellular carcinoma from southern Africa. Nature 350:429–431. Brunt EM. 2000. Grading and staging the histopathological lesions of chronic hepatitis: the Knodell histology activity index and beyond. Hepatology. 31:241–246. Brunt EM, Janney CG, Di Bisceglie AM, Neuschwander-Tetri BA, Bacon BR. 1999. Nonalcoholic steatohepatitis: A proposal for grading and staging the histologic lesions. Am J Gastroenterol 94:2467–2474. Bugianesi E, Leone N, Vanni E, et al. 2002. Expanding the natural history of nonalcoholic steatohepatitis: from crytogenic cirrhosis to hepatocellular carcinoma. Gastroenterology 123:134–140. Bulaj ZJ, Griffen LM, Jorde LB, Edwards CQ, Kushner JP. 1996. Clinical and biochemical abnormalities in people heterozygous for hemochromatosis. N Engl J Med 355:1799–1805. Byron D, Minuk GY. 1996. Clinical hepatology: Profile of an urban, hospitalbased practice. Hepatology 24:813–815. Cacoub P, Geffray L, Rosenthal E, et al. 2001. Mortality among human immunodeficiency virus-infected patients with cirrhosis or hepatocellular carcinoma due to hepatitis C virus in French Departments of Internal Medicine/Infectious Diseases, in 1995 and 1997. Clin Infect Dis 32:1207–1214. Caldwell SH, Oelsner DH, Iezzoni JC, Hespenheide EE, Battle EH, Driscoll CJ. 1999. Cryptogenic cirrhosis; clinical characterization and risk factors for underlying disease. Hepatology 29:664–669. Camaschella C, Roetto A, Cali A, et al. 2000. The gene TFR2 is mutated in a new type of haemochromatosis mapping to 7q22. Nat Genet 25:14– 15. Camma C, Giunta M, Andreone P, Craxi A. 2001. Interferon and prevention of hepatocellular carcinoma in viral cirrhosis: An evidence-based approach. J Hepatol 34:593–602. Cardin R, D’Errico A, Fiorentino M, Cecchetto A, Naccarato R, Farinati F. 2002. Hepatocyte proliferation and apoptosis in relation to oxidative damage in alcohol-related liver disease. Alcohol 37:43–48. CDC. 1998. Recommendations for prevention and control of hepatitis C virus (HCV) infection and HCV-related chronic disease. MMWR 47:1–33. CDC. 2002. Geographic distribution of chronic HBV infection. Hepatitis B slide set (Internet). CDC. 2003a. Prevention and control of infections with hepatitis viruses in correctional settings. MMWR 52:1–17. CDC. 2003b. Global progress toward universal childhood hepatitis B vaccination, 2003. MMWR 52:868–869. Chang MH, Chen CJ, Lai MS, et al. 1997. Universal hepatitis B vaccination in Taiwan and the incidence of hepatocellular carcinoma in children. N Engl J Med 336:1855–1859. Chapman RW. 1999. Risk factors for biliary tract carcinogenesis. Ann Oncol 10 (Suppl 4): 308–311. Chen CJ, Chen CW, Wu MM, Kuo TL. 1992. Cancer potential in liver, lung, bladder and kidney due to ingested inorganic arsenic in drinking water. Br J Cancer 66:888–892. Chen CJ, Chuang YC, Lin TM, Wu HY. 1985. Malignant neoplasms among residents of a blackfoot disease-endemic area in Taiwan: High-arsenic artesian well water and cancers. Cancer Res 45:5895–5899. Chen CJ, Chuang YC, You SL, Lin TM, Wu HY. 1986. A retrospective study on malignant neoplasms of bladder, lung and liver in blackfoot disease endemic area in Taiwan. Br J Cancer 53:399–405. Chen CJ, Kuo TL, Wu M-M. 1988a. Arsenic and cancers (letter). Lancet 1:414–415.

779

Chen CJ, Liang KY, Chang AS, et al. 1991. Effects of hepatitis B virus, alcohol drinking, cigarette smoking and familial tendency on hepatocellular carcinoma. Hepatology 13:398–406. Chen CJ, Wang CJ. 1990. Ecological correlation between arsenic level in well water and age-adjusted mortality from malignant neoplasms. Cancer Res 50:5470–5474. Chen CJ, Wu MM, Lee SS, Wang JD, Cheng SH, Wu HY. 1988b. Atherogenicity and carcinogenicity of high-arsenic artesian well water. Multiple risk factors and related malignant neoplasms of blackfoot disease. Arteriosclerosis 8:452–460. Chen CJ, Yu MW, Liaw Y-F, et al. 1996. Chronic hepatitis B carriers with null genotypes of glutathione s-transferase M1 and T1 polymorphisms who are exposed to aflatoxin are at increased risk of hepatocellular carcinoma. Am J Hum Genet 59:128–134. Chen SY, Wang LY, Lunn RM, et al. 2002. Polycyclic aromatic hydrocarbonDNA adducts in liver tissues of hepatocellular carcinoma patients and controls. Int J Cancer 99:14–21. Chiesa R, Donato F, Tagger A, et al. 2000. Etiology of hepatocellular carcinoma in Italian patients with and without cirrhosis. Cancer Epidemiol Biomarkers Prev 9:213–216. Chiou H-Y, Hsueh Y-M, Liaw K-F, et al. 1995. Incidence of internal cancers and ingested inorganic arsenic: A seven-year follow-up study in Taiwan. Cancer Res 55:1296–1300. Choo QL, Kuo G, Weiner J, et al. 1989. Isolation of a cDNA clone derived from a bloodborne non-A, non-B viral hepatitis genome. Science 244:359–362. Chopra S, Bonis PA. 2003. Liver transplantation for hepatitis C virus infection. In: UpToDate Online. Rose BD, ed. Wellesley, MA: UptoDate. Chou YH, Chiou HJ, Tiu CM, et al. 2003. Duplex doppler ultrasound of hepatic schistosomiasis japonica: A study of 47 patients. Am J Trop Med Hyg 68:18–23. Christopherson W, Mays E, Barrows G. 1980. Liver tumours in young women. In: Progress in Surgical Pathology. Vol. 2. Fenoglio CM, Wolff M, eds. New York: Masson Publishing UK, pp. 187–205. Chuang SE, Kuo ML, Hsu CH, et al. 2000. Curcumin-containing diet inhibits diethylnitrosamine-induced murine hepatocarcinogenesis. Carcinogenesis 21:331–335. Colditz GA for the Nurses’ Health Study Research Group. 1994. Oral contraceptive use and mortality during 12 years of follow-up: The Nurses’ Health Study. Ann Intern Med 120:821–826. Cole P, Macmahon B, Brown JB. 1976. Oestrogen profiles of parous and nulliparous women. Lancet 2:596–599. Colin JF, Cazals-Hatem D, Loriot MA, et al. 1999. Influence of human immunodeficiency virus infection on chronic hepatitis B in homosexual men. Hepatology 29:1306–1310. Collaborative MILTS Project Team. 1997. Oral contraceptives and liver cancer. Contraception 56:275–284. Colli A, Fraquelli M, Andreoletti M, et al. 2003. Severe liver fibrosis or cirrhosis: Accuracy of US for detection—analysis of 300 cases. Radiology 227:89–94. Colombo M. 1998. The role of hepatitis C virus in hepatocellular carcinoma. Recent Results Cancer Res 154:337–344. Colombo M, De Franchis R, Del Ninno E, et al. 1991. Hepatocellular carcinoma in Italian patients with cirrhosis. N Engl J Med 325:675– 680. Corrao G, Arico S. 1998. Independent and combined action of hepatitis C virus infection and alcohol consumption on the risk of symptomatic liver cirrhosis. Hepatology 27:914–919. Corrao G, Arico S, Zambon A, Torchio P, Di Orio F. 1997. Female sex and the risk of liver cirrhosis. Collaborative Groups for the Study of Liver Diseases in Italy. Scand J Gastroenterol 32:1174–1180. Corrao G, Bagnardi V, Zambon A, Arico S. 1999. Exploring the dose-response relationship between alcohol consumption and the risk of several alcoholrelated conditions: A meta-analysis. Addition 94:1551–1573. Cotrim HP, Parana R, Braga E, Lyra L. 2000. Nonalcoholic steatohepatitis and hepatocellular carcinoma: natural history? Am J Gastroenterol 95:3018–3019. Creech J, Johnson M. 1974. Angiosarcoma of the liver in the manufacture of PVC. J Occup Med 13:150–151. Curley SA, Barnett CC, Abdalla EK. 2003. Liver transplantation of hepatocellular carcinoma. In: UpToDate Online. Rose BD, ed. Wellesley, MA, UptoDate. Darby SC, Ewart DW, Giangrande PL. 1997. Mortality from liver cancer and liver disease in haemophilic men and boys in UK given blood products contaminated with hepatitis C. UK Haemophilia Centre Directors’ Organisation. Lancet 350:1425–1431. Darwish MA, Raouf TA, Rushdy P, Constantine NT, Rao MR, Edelman R. 1993. Risk factors associated with a high seroprevalence of hepatitis

780

PART IV: CANCER BY TISSUE OF ORIGIN

C virus infection in Egyptian blood donors. Am J Trop Med Hyg 49:440–447. Datz C, Haas T, Rinner H, Sandhofer F, Patsch W, Paulweber B. 1998. Heterozygosity for the C292Y mutation in the hemochromatosis gene is associated with increased serum iron, transferrin saturation, and hemoglobin in young women: A protective role against iron deficiency? Clin Chem 44:2429–2432. De Groen PC, Gores GJ, Larusso NF, Gunderson LL, Nagorney DM. 1999. Biliary tract cancers. N Engl J Med. 341:1368–1378. De Maria N, Manno M, Villa E. 2002. Sex hormones and liver cancer. Mol Cell Endocrinol 193:59–63. Deng Z, Wei Y, Ma Y. 2001. Glutathione-S-transferase M1 genotype in patients with hepatocellular carcinoma. Zhonghua Zhong Liu Za Zhi 23:477–479. Deugnier YM, Guyader D, Crantock L, et al. 1993. Primary liver cancer in genetic hemochromatosis: A clinical, pathological and pathogenetic study of 54 cases. Gastroenterology 104:228–234. Di Bisceglie AM, Goodman ZD, Ishak KG, et al. 1991. Long-term clinical and histopathological follow-up of chronic posttransfusion hepatitis. Hepatology. 14:969–974. Donato F, Gelatti U, Tagger A. 2001. Intrahepatic cholangiocarcinoma and hepatitis C and B virus infection, alcohol intake, and hepatolithiasis: A casecontrol study in Italy. Cancer Causes Control 12:959–964. Donato F, Tagger A, Gelatti U, et al. 2002. Alcohol and hepatocellular carcinoma: The effect of lifetime intake and hepatitis virus infections in men and women. Am J Epidemiol 155:323–331. Dos Santos Silva I, Jones M, Malveiro F, Swerdlow A. 1999. Mortality in the Portuguese Thorotrast study. Radiat Res 152:S88–S93. Duck BW, Carter JT, Coombes EJ. 1975. Mortality study of workers in a polyvinyl-chloride production plant. Lancet 13:1197–1199. Dunn MA. 2003. Parasitic diseases. In: Schiff ER, Sorrell MF, Maddrey WC, eds. Schiff’s Diseases of the Liver. Philadelphia: Lippincott Williams & Wilkins, pp. 1509–1527. Edmondson HA. 1958. In: Tumors of the Liver and Intrahepatic Bile Ducts. Washington: Armed Forces Institute of Pathology, pp. 32–33. Edmondson HA, Henderson B, Benton B. 1976. Liver cell adenomas associated with the use of the oral contraceptive. N Engl J Med 294:470–472. Edmondson HA, Steiner PE. 1954. Primary carcinoma of the liver. A study of 100 cases among 48,000 autopsies. Cancer 7:462–503. El-Serag HB, Davila JA, Petersen NJ, McGlynn KA. 2003. The continuing increase in the incidence of hepatocellular carcinoma in the United States: An update. Ann Intern Med 139:817–823. El-Serag HB, Hampel H, Yeh C, Rabeneck L. 2002. Extrahepatic manifestations of hepatitis C among United States male veterans. Hepatology 36:1439–1445. El-Serag HB, Mason AC. 1999. Rising incidence of hepatocellular carcinoma in the United States. N Engl J Med 340:745–750. El-Serag HB, Richardson PA, Everhart JE. 2001. The role of diabetes in hepatocellular carcinoma: A case-control study among United States veterans. Am J Gastroenterol 96:2462–2467. El-Zayadi AR, Abaza H, Shawky S, Mohamed MK, Selim OE, Badran HM. 2001. Prevalence and epidemiological features of hepatocellular carcinoma in Egypt—a single center experience. Hepatol Res 19:170–179. Eriksson S, Carlson J, Velez R. 1986. Risk of cirrhosis and primary liver cancer in alpha 1-antitrypsin deficiency. N Engl J Med 314:736–739. Evans AA, Chen G, Ross EA, et al. 2002. Eight-year follow-up of the 90,000person Haimen City Cohort: I. Hepatocellular carcinoma mortality, risk factors, and gender differences. Cancer Epidemiol Biomarkers Prev 11:369–376. Evans DM, Williams WJ, Kung IT. 1983. Angiosarcoma and hepatocellular carcinoma in vinyl chloride workers. Histopathology 7:377–388. Falk H, Caldwell GG, Ishak KG, Thomas LB, Popper H. 1981. Arsenic-related hepatic angiosarcoma. Am J Ind Med 2:43–50. Farber E, Cameron R. 1980. The sequential analysis of cancer development. Adv in Cancer Research 31:125–226. Fargion S, Fracanzani AL, Piperno A, et al. 1994. Prognostic factors for hepatocellular carcinoma in genetic hemochromatosis. Hepatology 20:1426–1431. Farrell GC. 2003. Non-alcoholic steatohepatitis: what is it and why is it important in the Asia-Pacific region? J Gastroenterol Hepatol 18:124–138. Fausto N, Campbell JS. 2003. The role of hepatocytes and oval cells in liver regeneration and repopulation. Mechanisms of Development 120: 117–130. Feder JN, Gnirke A, Thomas W, et al. 1996. A novel MHC class-I like gene is mutated in patients with hereditary haemochromatosis. Nat Genet 13: 399–408. Feder JN, Tsuchihashi Z, Irrinki A, et al. 1997. The hemochromatosis founder mutation in HLA-H disrupts beta2-microglobulin interaction and cell surface expression. J Biol Chem 272:14025–14028.

Feitelson MA, Sun B, Satiroglu Tufan NL, et al. 2002. Genetic mechanisms of hepatocarcinogenesis. Oncogene 21:2593–2604. Ferlay J, Bray J, Pisani P, Parkin DM. 2001. Globocan 2000—cancer incidence, mortality and prevalence worldwide. Lyon: IARCPress. Figueras J, Jaurrieta E, Valls C, et al. 1997. Survival after liver transplantation in cirrhotic patients with and without hepatocellular carcinoma: A comparative study. Hepatology 25:1485–1489. Filippazzo M, Aragona E, Cottone M, et al. 1985. Assessment of some risk factors for hepatocellular carcinoma: A case-control study. State Med 4:345–351. Flora KD, Berk B. 2002. Hepatocellular adenoma. In: Patel T, Talavera F, Achord JL, Mechaber AJ, Katz J, eds. eMedicine, Gastroenterology (Internet). Fontana RJ, Lok AS. 2003. Liver transplantation for chronic hepatitis B infection. In: UpToDate Online. Rose BD, ed. Wellesley, MA: UpToDate. Forman D, Bennett B, Stafford J, Doll R. 1985. Exposure to vinyl chloride and angiosarcoma of the liver: A report of the register of cases. Br J Ind Med 42:750–753. Forman D, Vincent TJ, Doll R. 1986. Cancer of the liver and the use of oral contraceptives. Br Med J 292:1357–1361. Frank C, Mohamed MK, Strickland GT, et al. 2000. The role of parenteral antischistosomal therapy in the spread of hepatitis C in Egypt. Lancet 355:887–891. Frezza M, Di Padova C, Pozzato G, Perin M, Baraona E, Lieber CS. 1990. High blood alcohol levels in women. The role of decreased gastric alcohol dehydrogenase activity and first-pass metabolism. N Engl J Med 322: 95–99. Frisch M, Biggar RJ, Engels EA, Goedert JJ; Aids-Cancer Match Registry Study Group. 2001. Association of cancer with AIDS-related immunosuppression in adults. JAMA 285:1736–1745. Fukuda K, Shibata A, Hirohata I, Tanikawa K, Yamaguchi G, Ishii M. 1993. A hospital based case-control study on hepatocellular carcinoma in Fukuoka and Saga prefectures, Northern Kyushu, Japan. Jpn J Cancer Res 84:708–714. Gaiani S, Gramantieri L, Venturoli N, et al. 1997. What is the criterion for differentiating chronic hepatitis from compensated cirrhosis? A prospective study comparing ultrasonography and percutaneous liver biopsy. J Hepatol 27:979–985. Ganne-Carrie N, Christidis C, Chastang C, et al. 2000. Liver iron is predictive of death in alcoholic cirrhosis: A multivariate study of 229 consecutive patients with alcoholic and/or hepatitis C virus cirrhosis: A prospective follow-up study. Gut 46:277–282. Gebo KA, Chander G, Jenckes MW, et al. 2002. Screening tests for hepatocellular carcinoma in patients with chronic hepatitis C: A systematic review. Hepatology 36:S84–92. George DK, Goldwurm S, Macdonald GA, et al. 1998. Increased hepatic iron concentration in nonalcoholic steatohepatitis is associated with increased fibrosis. Gastroenterology 114:311–318. Ghaffar YA, Fattah SA, Kamel M, Badr RM, Mahomed FF, Strickland GT. 1991. The impact of endemic schistosomiasis on acute viral hepatitis. Am J Trop Med Hyg 45:743–750. Gilson RJ, Hawkins AE, Beecham MR, et al. 1997. Interactions between HIV and hepatitis B virus in homosexual men: Effects on the natural history of infection. AIDS 11:597–606. Ginseng-HCC Chemopreventive Study Osaka Group. 2001. Study on chemoprevention of hepatocellular carcinoma by ginseng: An introduction to the protocol. J Korean Med Sci 16 Suppl: S70–S74. Goedert JJ, Hatzakis A, Sherman KE, Eyster ME, Multicenter Hemophilia Cohort Study. 2001. Lack of association of hepatitis C virus load and genotype with risk of end-stage liver disease in patients with human immunodeficiency virus coinfection. J Infect Dis 184:1202– 1205. Goh J, Callagy G, McEntee G, O’Keane J, Bomford A, Crowe J. 1999. Hepatocellular carcinoma arising in the absence of cirrhosis in genetic haemochromatosis: Three case reports and a review of the literature. Eur J Gastroenterol Hepatol 11:159–191. Gong Y, Han C, Chen J. 2000. Effect of tea polyphenols and tea pigments on the inhibition of precancerous liver lesions in rats. Nutr Cancer 38:81–86. Gong YY, Egal S, Hounsa A, et al. 2003. Determinants of aflatoxin exposure in young children from Benin and Togo, West Africa: The critical role of weaning. Int J Epidemiol 32:556–562. Gonzalez SA, Talal AH. 2003. Hepatitis C virus in human immunodeficiency virus-infected individuals: An emerging comorbidity with significant implications. Gut 47:845–851. Goodman MT, Moriwaki H, Vaeth M, Akiba S, Hayabuchi H, Mabuchi K. 1995. Prospective cohort study of risk factors for primary liver cancer in Hiroshima and Nagasaki, Japan. Epidemiology 6:36–41.

Liver Cancer Gordeuk, V, Mukiibi J, Hasstedt SJ, et al. 1992. Iron overload in Africa: Interaction between a gene and dietary iron content. N Engl J Med 326:95–100. Gordeuk VR, McLaren CE, MacPhail AP, Deichsel G, Bothwell TH. 1996. Associations of iron overload in Africa with hepatocellular carcinoma and tuberculosis: Strachan’s 1929 thesis revisited. Blood 87:3470–3476. Guyton KZ, Kensler TW. 2002. Prevention of liver cancer. Curr Oncol Rep 4:464–470. Hadziyannis S, Tabor E, Kaklamani E, et al. 1995. A case-control study of hepatitis B and C virus infections in the etiology of hepatocellular carcinoma. Int J Cancer 60:627–631. Halim AB, Gary RF, Dash S, Gerber MA. 1999. Effect of schistosomiasis and hepatitis on liver disease. Am J Trop Med Hyg 60:915–920. Halver J. 1965. Hepatomas in fish. In: Burdette W, eds. Primary Hepatoma. Utah: Utah Press, pp. 103–112. Hann HL, Kim C, London WT, et al. 1982. Hepatitis B virus and primary hepatocellular carcinoma: Family studies in Korea. Int J Cancer 30:47–51. Hannaford PC, Kay CR, Vessey MP, Painter R, Mant J. 1997. Combined oral contraceptives and liver disease. Contraception 55:145–151. Hanson EH, Imperatore G, Burke W. 2001. HFE gene and hereditary hemochromatosis: A HuGE review. Am J Epidemiol 154:193–206. Hardell L, Bengtsson NO, Jonsson U, Eriksson S, Larsson LG. 1984. Aetiological aspects on primary liver cancer with special regard to alcohol, organic solvents and acute intermittent porphyria—an epidemiological investigation. Br J Cancer 50:389–397. Harnois DM, Gores GJ, Ludwig J, Steers JL, Larusso NF, Wiesner RH. 1997. Are patients with cirrhotic stage primary sclerosing cholangitis at risk for the development of hepatocellular cancer? J Hepatol 27:512–516. Hassan MM, Hwang L-Y, Hatten CJ, et al. 2002. Risk factors for hepatocellular carcinoma; synergism of alcohol with viral hepatitis and diabetes mellitus. Hepatology 36:1206–1213. Hassan MM, Zaghloul AS, El-Serag HB, et al. 2001. The role of hepatitis C in hepatocellular carcinoma. A case-control study among Egyptian patients. J Clin Gastroenterol 33:123–126. Hayashi H, Takikawa T, Nishimura N, Yano M, Isomura T, Sakamoto N. 1994. Improvements of serum aminotransferase levels after phlebotomy in patients with chronic active hepatitis C and excess hepatic iron. Am J Gastroenterol 89:986–988. Hayashi K, Kumada T, Nakano S, et al. 2002. Incidence of hepatocellular carcinoma in chronic hepatitis C after interferon therapy. Hepatogastroenterology 49:508–512. Heilbrun LK, Nomura A, Stemmermann GN. 1986. Black tea consumption and cancer risk: A prospective study. Br J Cancer 54:677–683. Henderson BE, Preston-Martin S, Edmondson HA, Peters RL, Pike MC. 1983. Hepatocellular carcinoma and oral contraceptives. Br J Cancer 48: 437–440. Heyward WL, Lanier AP, McMahon BJ, et al. 1985. Early detection of primary hepatocellular carcinoma. Screening for primary hepatocellular carcinoma among persons infected with hepatitis B virus. JAMA 254: 3052–3054. Hift RF, Corrigall AV, Hancock V, Kannemeyer J, Kirsch RE, Meissner PN. 2002. Porphyria cutanea tarda: the etiological importance of mutations in the HFE gene and viral infection is population-dependent. Cell Mol Biol 48:853–859. Higginson J, Grobellar BG, Walker ARP. 1957. Hepatic fibrosis and cirrhosis in man in relation to malnutrition. Am J Pathol 33:29–54. Hirayama T. 1989. A large-scale cohort study on risk factors for primary liver cancer, with special reference to the role of cigarette smoking. Cancer Chemother Pharmacol 23:S114–S117. Hirayama T. 1990. Life-Style and Mortality: A Large-Scale Census-Based Cohort Study in Japan. Vol. 6. Contributions to Epidemiology and Biostatistics. Basel, Karger. Hopenhayn-Rich C, Biggs ML, Smith AM. 1988. Lung and kidney cancer mortality associated with arsenic in drinking water in Cordoba, Argentina. Int J Epidemiol 27:561–569. Hsieh LL, Huang RC, Yu MW, Chen CJ, Liaw YF. 1996. L-myc, GSTM1 genetic polymorphism and hepatocellular carcinoma risk among chronic hepatitis B carriers. Cancer Lett 103:171–176. Hsing AW, Gwo W, Chen J, Stone BJ, Blot WJ, Fraumeni JF, Jr. 1991. Correlates of liver cancer mortality in China. Int J Epidemiol 20:54–59. Hsing AW, Hoover RN, McLaughlin JK, et al. 1992a. Oral contraceptives and primary liver cancer among young women. Cancer Causes Control 3:43–48. Hsing AW, McLaughlin JK, Hoover RN, Co Chien HT, Blot WJ, Fraumeni JF, Jr. 1992b. Parity and primary liver cancer among young women. J Natl Cancer Inst 84:1118–1119. Hsing AW, McLaughlin JK, Hrubec Z, et al. 1990. Cigarette smoking and liver cancer among US veterans. Cancer Causes Control 1:217–221.

781

Hsu IC, Metcalf RA, Sun T, et al. 1991. Mutational hotspot in the p53 gene in human hepatocellular carcinomas. Nature 350:427–428. Hu KQ. 2002. Rationale and feasibility of chemoprevention of hepatocellular carcinoma by cyclooxygenase-2 inhibitors. J Lab Clin Med 139:234– 243. Huang YS, Chern HD, Wu JC, et al. 2003. Polymorphism of the Nacetyltransferase 2 gene, red meat intake, and the susceptibility of hepatocellular carcinoma. Am J Gastroenterol 98:1417–1422. IARC Monographs. 1987a. Overall evaluations of carcinogenicity: An updating of IARC monographs volumes 1–42. Suppl 7. Lyon: IARCPress, pp. 83–87. IARC Monographs. 1987b. Overall evaluations of carcinogenicity: An updating of IARC monographs volumes 1–42. Suppl 7. Lyon: IARCPress, pp. 96–98. IARC Monographs. 1988. Alcohol drinking. Vol. 44. Lyon: IARCPress, pp. 255–259. IARC Monographs. 1994a. Hepatitis viruses. Vol. 59. Lyon: IARCPress, pp. 72–97. IARC Monographs. 1994b. Schistosomes, liver flukes and helicobacter pylori. Vol. 61. Lyon: IARCPress, pp. 45–119. IARC Monographs. 1994c. Schistosomes, liver flukes and helicobacter pylori. Vol. 61. Lyon: IARCPress, pp. 121–175. IARC Monographs. 1999. Hormonal contraception and post-menopausal hormonal therapy. Vol. 72. Lyon: IARCPress, pp. 293. Iida F, Iida R, Kamijo H, et al. 1999. Chronic Japanese schistosomiasis and hepatocellular carcinoma: Ten years of follow-up in Tamanashi Prefecture, Japan. Bull World Health Organ 77:573–581. Ijima T, Saitoh N, Nobotumo K, et al. 1984. A prospective cohort study of hepatitis B surface antigen carriers in a working population. Gann 75:571–573. Ikeda K, Saitoh S, Kobayashi M, et al. 2001. Long-term interferon therapy for 1 year or longer reduces the hepatocellular carcinogenesis rate in patients with liver cirrhosis caused by hepatitis C virus: A pilot study. J Gastroenterol Hepatol 16:406–415. Ikeda K, Saitoh S, Koida I, et al. 1993. A multivariate analysis of risk factors for hepatocellular carcinogenesis: A prospective observation of 795 patients with viral and alcoholic cirrhosis. Hepatology 18:47– 53. Ikeda K, Saitoh S, Suzuki Y, et al. 1998. Disease progression and hepatocellular carcinogenesis in patients with chronic viral hepatitis: A prospective observation of 2215 patients. J Hepatol 28:930–938. Imbert-Bismut F, Ratziu V, Pieroni L, et al. 2001. Biochemical markers of liver fibrosis in patients with hepatitis C virus infection: A prospective study. Lancet. 357:1069–1075. Ishak KG. 1987. Malignant mesenchymal tumors of the liver. In: Okuda K, Ishak KG, eds. Neoplastic Diseases of the Liver. Berlin: Springer-Verlag, pp. 159–176. Ishii A, Matsuoka H, Aji T, et al. 1994. Parasite infection and cancer: With special emphasis on Schistosoma japonicum infections (Trematoda). A review. Mutat Res 305:273–281. Jazwinska EC, Cullen LM, Busfield F, et al. 1996. Haemochromatosis and HLA-H. Nat Genet 14:249–251. Jemal A, Murray T, Samuels A, Ghafoor A, Ward E, Thun MJ. 2003. Cancer Statistics, 2003. CA Cancer J Clin 53:5–26. Jones J. 2001. New alcohol strategy being developed for England. BMJ 322:636. Jouanolle AM, Gandon G, Jezequel P, et al. 1996. Haemochromatosis and HLA H. Nat Genet 14:251–252. Kakizaki S, Takagi H, Fukusato T, et al. 2001. Effect of alpha-tocopherol on hepatocarcinogenesis in transforming growth factor-alpha (TGF-alpha) transgenic mice treated with diethylnitrosamine. Int J Vitam Nutr Res 71:261–267. Kalk H. 1960. The relationship between fatty liver and diabetes mellitus. Ger Med Mon 5:81–84. Kane MA. 2003. Global control of primary hepatocellular carcinoma with hepatitis B vaccine: The contributions of research in Taiwan. Cancer Epidemiol Biomarkers Prev 12:2–3. Kasahara A, Hayashi N, Mochizuki K, et al. 1998. Risk factors for hepatocellular carcinoma and its incidence after interferon treatment in patients with chronic hepatitis C. Hepatology 27:1394–1402. Kato J, Kobune M, Nakamura T, et al. 2001. Normalization of elevated hepatic 8-hydroxy-2¢-deoxyguanosine levels in chronic hepatitis C patients by phlebotomy and low iron diet. Cancer Res 61:8697– 8702. Kaul A. 1995. Biokinetic models and data. In: van Kaick G, Karaoglou A, Kellerer AM, eds. Health Effects of Internally Deposited Radionuclides: Emphasis on Radium and Thorium. Singapore: World Scientific Publishing Co., pp. 53–68.

782

PART IV: CANCER BY TISSUE OF ORIGIN

Kenny-Walsh E for the Irish Hepatology Research Group. 1999. Clinical outcomes after hepatitis infection from contaminated anti-globulin. N Engl J Med 340:1228–1233. Ketcham AS, Wexler H, Mantel N. 1963. Effect of alcohol in mouse neoplasia. Cancer Res 23:667–670. Kew M, Di Bisceclie A, Paterson A, et al. 1985. Smoking as a risk factor in hepatocellular carcinoma. A case-control study in southern African blacks. Cancer 56:2315–2317. Kew MC, Song E, Mohammed A, Hodkinson J. 1990. Contraceptive steroids as a risk factor for hepatocellular carcinoma: A case-control study in South African black women. Hepatology 11:298–302. Khan SA, Taylor-Robinson SD, Toledano MB, Beck A, Elliott P, Thomas HC. 2002. Changing international trends in mortality rates for liver, biliary and pancreatic tumours. J Hepatol 37:806–813. Kielhorn J, Melber C, Wahnschaffe U, Aitto A, Mangelsdorf I. 2000. Vinyl chloride: still a cause for concern. Environ Health Perspect 108:579–588. Kiyosawa K, Sodeyama T, Tanaka E, et al. 1990. Interrelationship of blood transfusion, non-A, non-B hepatitis and hepatocellular carcinoma: Analysis by detection of antibody to hepatitis C virus. Hepatology 12:671–675. Klatskin G. 1977. Hepatic tumors—possible relationship to use of oral contraceptives. Gastroenterology 73:386–394. Klaunig JE. 1992. Chemopreventive effects of green tea components on hepatic carcinogenesis. Prev Med 21:510–519. Kobayashi M, Ikeda K, Saitoh S, et al. 2000. Incidence of primary cholangiocellular carcinoma of the liver in Japanese patients with hepatitis C virusrelated cirrhosis. Cancer 88:2471–2477. Kojiro M, Nakashima T. 1987. Pathology of hepatocellular carcinoma. In: Okuda K, Ishak KG, eds. Neoplastic Diseases of the Liver. Berlin: Springer-Verlag, pp. 81–104. Kono S, Ikeda M, Tokudome S, et al. 1987. Cigarette smoking, alcohol and cancer mortality: A cohort study of male Japanese physicians. Jpn J Cancer Res 78:1323–1328. Koretz RL, Abbey H, Coleman E, Gitnick G. 1993. Non-A, non-B posttransfusion hepatitis. Looking back in the second decade. Ann Intern Med 19:110–115. Kosaka A, Takahashi H, Yajima Y, et al. 1996. Hepatocellular carcinoma associated with anabolic steroid therapy: Report of a case and review of the Japanese literature. J Gastroenterol 31:450–454. Koskinen SV, Reunanen AR, Martelin TP, Valkonen T. 1998. Mortality in a large population-based cohort of patients with drug-treated diabetes mellitus. Am J Public Health 88:765–770. Kumada H. 2002. Long-term treatment of chronic hepatitis C with glycyrrhizin [stronger neo-minophagen C (SNMC)] for preventing liver cirrhosis and hepatocellular carcinoma. Oncology 62 Suppl 1:94–100. Kuper H, Tzonou A, Kaklamani E, et al. 2000a. Tobacco smoking, alcohol consumption and their interaction in the causation of hepatocellular carcinoma. Int J Cancer 85:498–502. Kuper H, Tzonou A, Lagiou P, et al. 2000b. Diet and hepatocellular carcinoma: A case-control study in Greece. Nutr Cancer 38:6–12. La Vecchia C, Negri E, Decarli A, D’Avanzo B, Franceschi S. 1988. Risk factors for hepatocellular carcinoma in northern Italy. Int J Cancer 42:872–876. La Vecchia C, Negri E, Franceschi S, D’Avanzo B. 1992. Reproductive factors and the risk of hepatocellular carcinoma in women. Int J Cancer 52:351–354. La Vecchia C, Negri E, Franceschi S, Parazzini F. 1993. Long-term impact of reproductive factors on cancer risk. Int J Cancer 53:215–219. Ladero JM, Agundez JA, Rodriguez-Lescure A, Diaz-Rubio M, Benitez J. 1996. RsaI polymorphism at the cytochrome P4502E1 locus and risk of hepatocellular carcinoma. Gut 39:330–333. Lagiou P, Kuper H, Stuver SO, Tzonou A, Trichopoulos D, Adami H-O. 2000. Role of diabetes mellitus in the etiology of hepatocellular carcinoma. J Natl Cancer Inst 92:1096–1099. Lam KC, Yu MC, Leung JW, Henderson BE. 1982. Hepatitis B virus and cigarette smoking: Risk factors for hepatocellular carcinoma in Hong Kong. Cancer Res 42:5246–5248. Lambe M, Trichopoulos D, Hsieh CC, Ekbom A, Pavia M. 1993. Parity and hepatocellular carcinoma. A population-based study in Sweden. Int J Cancer 55:745–747. Lander JJ, Stanley RJ, Sumner HW, Boswell DC, Aach RD. 1975. Angiosarcoma of the liver associated with Fowler’s solution (potassium arsenite). Gastroenterology 68:1582–1586. Larouze B, Saimot G, Lustbader ED, et al. 1976. Host responses to hepatitisB infection in patients with primary hepatic carcinoma and their families. A case/control study in Senegal, West Africa. Lancet 2:534–538. Lee CL, Hsieh KS, Ko YC. 2003. Trends in the incidence of hepatocellular carcinoma in boys and girls in Taiwan after large-scale hepatitis B vaccination. Cancer Epidemiol Biomarkers Prev 12:57–59.

Lee CL, Ko YC. 1997. Hepatitis B vaccination and hepatocellular carcinoma in Taiwan. Pediatrics 99:351–353. Lee HS, Yoon JH, Kamimura S, Iwata K, Watanabe H, Kim CY. 1997. Lack of association of cytochrome P450 2E1 genetic polymorphisms with the risk of human hepatocellular carcinoma. Int J Cancer 71:737–740. Lelbach WK. 1996. A 25-year follow-up study of heavily exposed vinyl chloride workers in Germany. Am J Ind Med 29:446–458. Lewis DR, Southwick JW, Ouellet-Hellstrom R, Rench J, Calderon RL. 1999. Drinking water arsenic in Utah: A cohot mortality study. Environ Health Perspect 107:359–365. Lin L, Yang F, Ye Z, et al. 1991. Case-control study of cigarette smoking and primary hepatoma in an aflatoxin-endemic region of China: A protective effect. Pharmacogenetics 1:79–85. Lindor KD, Larusso NF. 2003. Primary sclerosing cholangitis. In: Schiff ER, Sorrell MF, Maddrey WC, eds. Schiff’s Diseases of the Liver. Philadelphia: Lippincott Williams & Wilkings, pp. 673–684. Linet MS, Gridley G, Nyren O, et al. 1999. Primary liver cancer, other malignancies, and mortality risks following porphyria: A cohort study in Denmark and Sweden. Am J Epidemiol 149:1010–1015. Lipshutz GS, Brennan TV, Warren RS. 2002. Thorotrast-induced liver neoplasia: A collective review. J Am Coll Surg 195:713–718. Liver Cancer Study Group of Japan. 1984. Primary liver cancer in Japan. Cancer 54:1747–1755. Llovet JM, Burroughs A, Bruix J. 2003. Hepatocellular carcinoma. Lancet 362:1907–1917. Lok AS, McMahon BJ; Practice Guidelines Committee, American Association for the Study of Liver Diseases. 2001. Chronic hepatitis B. Hepatology 34:1225–1241. Lu JN, Chen CJ. 1991. Prevalence of hepatitis B surface antigen carrier status among residents in the endemic area of chronic arsenicism in Taiwan. Anticancer Res 11:229–234. Lu SN, Lin TM, Chen CJ, et al. 1988. A case-control study of primary hepatocellular carcinoma in Taiwan. Cancer 62:2051–2055. Ludwig J, McGill DB, Lindor KD. 1997. Review: Nonalcoholic steatohepatitis. J Gastroenterol Hepatol 12:398–403. Ludwig J, Viggiano TR, McGill DB, Ott BJ. 1980. Nonalcoholic steatohepatitis. Mayo Clinic experiences with a hitherto unnamed disease. Mayo Clin Proc 55:434–438. Lundvall O, Weinfeld A. 1968. Studies of the clinical and metabolic effects of phlebotomy treatment in porphyria cutanea tarda. Acta Med Scand 184:191–199. Lyra LG, Reboucas G, Andrade ZA. 1976. Hepatitis B surface antigen carrier state in hepatosplenic schistosomiasis. Gastroenterology 71: 641–645. MacMahon HE, Murphy AS, Bates MI. 1947. Endothelial-cell sarcoma of the liver following Thorotrast injections. Am J Path 23:585–611. Madwar MA, El Tahawy M, Strickand GT. 1989. The relationship between uncomplicated schistosomiasis and hepatitis B infection. Trans R Soc Trop Med Hyg 83:233–236. Mahoney FJ, Kane MA. 1999. Hepatitis B vaccine. In: Plotkin SA, Orenstein WA, eds. Vaccine. ed 3. Philadelphia: WB Saunders, pp. 158–182. Mandishona E, MacPhail AP, Gordeuk VR, et al. 1998. Dietary iron overload as a risk factor for hepatocellular carcinoma in Black Africans. Hepatology 27:1563–1566. Marinier E, Barrois V, Larouze B, et al. 1985. Lack of perinatal transmission of hepatitis B virus in Senegal, West Africa. J Pediatr 106:843–849. Marion PL, Van Davelaar MJ, Knights SS, et al. 1986. Hepatocellular carcinoma in ground squirrels persistently infected with ground squirrel hepatitis virus. Proc Natl Acad Sci USA 83:4543–4546. Maruyama S, Nagasue N, Dhar DK, et al. 2001. Preventive effect of FK143, a 5a-reductase inhibitor, on chemical hepatocarcinogenesis in rats. Clin Cancer Res 7:2096–2104. Mason AL, Lau JY, Hoang N, et al. 1999. Association of diabetes mellitus and chronic hepatitis C virus infection. Hepatology 29:328–333. Mattsson L, Sonnerborg A, Weiland O. 1993. Outcome of acute symptomatic non-A, non-B hepatitis: A 13-year follow-up study of hepatitis C virus markers. Liver 13:274–278. Mays E, Christopherson W. 1984. Hepatic tumors induced by sex steroids. Semin Liver Dis 4:147–157. McCoy GD, Hecht SS, Katayama S, Wynder EL. 1981. Differential effect of chronic ethanol consumption on the carcinogenicity of N-nitrosopyrrolidine and N¢-nitrosonornicotine in male Syrian golden hamsters. Cancer Res 41:2849–2854. McGlynn KA, Hunter K, Levoyer T, et al. 2003. Susceptibility to aflatoxin B1related primary hepatocellular carcinoma in mice and humans. Cancer Res 63:4594–4601. McGlynn KA, Rosvold EA, Lustbader ED, et al. 1995. Susceptibility to hepatocellular carcinoma is associated with genetic variation in the

Liver Cancer detoxification of aflatoxin B1. Proc Natl Acad Sci USA 92:2384– 2387. McGlynn KA, Tsao L, Hsing AW, Devesa SS, Fraumeni JF, Jr. 2001. International trends and patterns of primary liver cancer. Int J Cancer 94: 290–296. McLaughlin JK, Lipworth L. 1999. A critical review of the epidemiologic literature on health effects of occupational exposure to vinyl chloride. J Epidemiol Biostat 4:253–275. McNamara L, MacPhail AP, Gordeuk VR, Hasstedt SJ, Rouault T. 1998. Is there a link between African iron overload and the described mutations of the hereditary hemochromatosis gene? Br J Haematol 102:1176– 1178. Mettlin C, Natarajan N. 1981. Studies on the role of oral contraceptive use in the etiology of benign and malignant liver tumors. J Surg Oncol 18:73–85. Ming L, Thorgeirsson SS, Gail MH, et al. 2002. Dominant role of hepatitis B virus and cofactor role of aflatoxin in hepatocarcinogenesis in Qidong, China. Hepatology 36:1214–1220. Miyakawa H, Izumi N, Marumo F, Sato C. 1996. Roles of alcohol, hepatitis virus infection, and gender in the development of hepatocellular carcinoma in patients with liver cirrhosis. Alcohol Clin Exp Res 20 (1S): 91A–94A. Mizoue T, Totui N, Nishisaka K, et al. 2000. Prospective study on the relation of cigarette smoking with cancer of the liver and stomach in an endemic region. Int J Epidemiol 29:232–237. Monson RR, Peters JM, Johnson MN. 1974. Proportional mortality among vinyl-chloride workers. Lancet 2:397–398. Morcos M, Dubois S, Bralet MP, Belghiti J, Degott C, Terris B. 2001. Primary liver cancer in genetic hemochromatosis reveals a broad histologic spectrum. Am J Clin Pathol 116:738–743. Mori M, Hara M, Wada I, et al. 2000. Prospective study of hepatitis B and C viral infections, cigarette smoking, alcohol consumption, and other factors associated with hepatocellular carcinoma risk in Japan. Am J Epidemiol 151:131–139. Mori T, Fukutomi K, Kato Y, et al. 1999a. 1998 results of the first series of follow-up studies on Japanese Thorotrast patients and their relationships to an autopsy series. Radiat Res 152:S72–S80. Mori T, Kido C, Fukutomi K, et al. 1999b. Summary of entire Japanese Thorotrast follow-up study: Updated 1998. Radiat Res 152:S84–S87. Moyo VM, Makunike R, Gangaidzo IT, et al. 1998. African iron overload and hepatocellular carcinoma. Eur J Haematol 60:28–34. Mu LN, Zhou XF, Ding BG, et al. 2003. A case-control study on drinking green tea and decreasing risk of cancers in the alimentary canal among cigarette smokers and alcohol drinkers. Zhonghua Liu Xing Bing Xue Za Zhi 24:192–195. Mucci LA, Kuper HE, Tamimi R, Lagiou P, Spanos E, Trichopoulos D. 2001. Age at menarche and age at menopause in relation to hepatocellular carcinoma in women. BJOG 108:291–294. Mukaiya M, Nishi M, Miyake H, et al. 1998. Chronic liver diseases for the risk of hepatocellular carcinoma: A case-control study in Japan. Etiologic association of alcohol consumption, cigarette smoking and the development of chronic liver diseases. Hepatogastroenterology 45:2328– 2332. Nagano K, Fukuda Y, Nakano I, et al. 2000. A case of the development of two hepatocellular carcinomas and a cholangiocarcinoma with cirrhosis after elimination of serum hepatitis C virus RNA with interferon therapy. Hepatogastroenterology 47:1436–1438. National Academy of Sciences. 1988. Health risks of radon and other internally deposited alpha-emitters. Washington, DC: National Academy Press, Bier IV Report. Nayak NC, Dhar A, Sachdeva R. 1977. Association of human hepatocellular carcinoma and cirrhosis with hepatitis B surface and core antigens in the liver. Int J Cancer 20:643–654. Nelson RL, Davis FG, Persky V, Becker E. 1995. Risk of neoplastic and other disease among people with heterozygosity for hereditary hemochromatosis. Cancer 76:875–879. Neuberger J, Forman D, Doll R, Williams R. 1986. Oral contraceptives and hepatocellular carcinoma. Br Med J 292:1355–1357. Nicholson WJ, Hammond EC, Seidman H, Selikoff IJ. 1975. Mortality experience of a cohort of vinyl chloride-polyvinyl chloride workers. Ann NY Acad Sci 246:225–230. Niederau C, Fischer R, Purschel A, Stremmel W, Saussinger D, Strohmeyer G. 1996. Long-term survival in patients with hereditary hemochromatosis. Gastroenterology 110:1107–1119. Niederau C, Fischer R, Sonnenberg A, Stremmel W, Trampisch HJ, Strohmeyer G. 1985. Survival and cause of death in cirrhotic and in non-cirrhotic patients with primary hemochromatosis. New Engl J Med 131:1256– 1261.

783

Niederau C, Lange S, Heintges T, et al. 1998. Prognosis of chronic hepatitis C: Results of a large prospective cohort study. Hepatology 28:1687–1695. Nishiguchi S, Shiomi S, Nakatani S, et al. 2001. Prevention of hepatocellular carcinoma in patients with chronic active hepatitis C and cirrhosis. Lancet 357:196–197. Nyberg U, Nilsson B, Travis LB, Holm L-E, Hall P. 2002. Cancer incidence among Swedish patients exposed to radioactive Thorotrast: A forty-year follow-up survey. Radiat Res 157:419–425. Oberti F, Valsesia E, Pilette C, et al. 1997. Noninvasive diagnosis of hepatic fibrosis or cirrhosis. Gastroenterology 113:1609–1616. Okada K, Kamiyama I, Inomata M, Imai M, Miyakawa Y. 1976. e Antigen and anti-e in the serum of a symptomatic carrier mothers as indicators of positive and negative transmission of hepatitis B to their infants. N Engl J Med 294:746–749. Okuda K, Fujimoto I, Hanai A, Urano Y. 1987. Changing incidence of hepatocellular carcinoma in Japan. Cancer Res 47:4967–4972. Okuda K, Nakanuma Y, Miyazaki M. 2002. Cholangiocarcinoma: Recent progress. Part I: epidemiology and etiology. J Gastroenterol Hepatol 17:1049–1055. Okuno M, Kojima S, Akita K, et al. 2002. Retinoids in liver fibrosis and cancer. Front Biosci 7:d204–d218. Olubuyide IO, Bamgboye EA. 1990. A case-controlled study of the current role of cigarette smoking and alcohol consumption in primary liver cell carcinoma in Nigeria. Afr J Med Sci 19:191–194. Omer RE, Verhoef L, Van’t Veer P, et al. 2001. Peanut butter intake, GSTM1 genotype and hepatocellular carcinoma: A case-control study in Sudan. Cancer Causes Control 12:23–32. Oshima A, Tsukuma H, Hiyama T, Fujimoto I, Yamano H, Tanaka M. 1984. Follow-up study of HBsAg-positive blood donors with special reference to effect of drinking and smoking on development of liver cancer. Int J Cancer 34:775–779. Ott MG, Langner RR, Holder BH. 1975. Vinyl chloride exposure in a controlled industrial environment. Arch Environ Health 30:333–339. Overly WL, Dankoff JA, Wang BK, Singh UD. 1984. Androgens and hepatocellular carcinoma in an athlete (letter). Ann Intern Med 100:158–159. Ozturk M. 1991. p53 mutation in hepatocellular carcinoma after aflatoxin exposure. Lancet 338:1356–1359. Palmer JR, Rosenberg L, Kaufman DW, Warshauer ME, Stolley P, Shapiro S. 1989. Oral contraceptive use and liver cancer. Am J Epidemiol 130:878–882. Papatheodoridis GV, Papadimitropoulos VC, Hadziyannis SJ. 2001. Effect of interferon therapy on the development of hepatocellular carcinoma in patients with hepatitis C virus-related cirrhosis: A meta-analysis. Aliment Pharmacol Ther 15:689–698. Parkin DM. 2001b. Global cancer statistics in the year 2000. Lancet Oncol 2:533–543. Parkin DM, Bray FI, Devesa SS. 2001a. Cancer burden in the year 2000. The global picture. Eur J Cancer 37: Suppl 8:S4–S66. Parkin DM, Whelan SL, Ferlay J, Raymond L, Young J, eds. 1997. Cancer Incidence in Five Continents, Vol. VII. IARC Scientific Publications No. 143. Lyon: IARCPress. Pascale RM, Simile MM, De Miglio MR, et al. 1995. Chemoprevention by S-adenosyl-L-methionine of rat liver carcinogenesis initiated by 1,2-dimethylhydrazine and promoted by orotic acid. Carcinogenesis 16:427–430. Patel T. 2002. Worldwide trends in mortality from biliary tract malignancies. BMC Cancer 2:10. Epub 2002 May 03. Perlmutter DH. 1998. Alpha-1-antitrypsin deficiency. Semin Liver Dis 18:217–225. Persson I, Yuen J, Bergkvist L, Schairer C. 1996. Cancer incidence and mortality in women receiving estrogen and estrogen-progestin replacement therapy—long-term follow-up of a Swedish cohort. Int J Cancer 67:327–332. Petrides AS, Defronzo RA. 1989. Glucose and insulin metabolism in cirrhosis. J Hepatol 8:107–114. Piperno A, Sampietro M, D’Alba R, et al. 1996. Iron stores, response to alphainterferon therapy, and effects of iron depletion in chronic hepatitis C. Liver 16:248–254. Pointon JJ, Wallace D, Merryweather-Clarke AT, Robson KJ. 2000. Uncommon mutations and polymorphisms in the hemochromatosis gene. Genet Test 4:151–161. Polizos A, Kelekis N, Sinani C, Patsiaoura K, Papadamou G, Dalekos GN. 2003. Advanced intrahepatic cholangiocarcinomas in hepatitis C virusrelated decompensated cirrhosis: Case-report and review of the literature. Eur J Gastroenterol Hepatol 15:331–334. Poonawala A, Nair SP, Thuluvath PJ. 2000. Prevalence of obesity and diabetes in patients with cryptogenic cirrhosis: A case-control study. Hepatology 32:689–692.

784

PART IV: CANCER BY TISSUE OF ORIGIN

Popper H, Roth RL, Purcell RH, et al. 1987. Hepatocarcinogenicity of the woodchuck hepatitis virus. Proc Natl Acad Sci USA 84:866–870. Powell EE, Cooksley WG, Hanson R, Searle J, Halliday JW, Powell LW. 1990. The natural history of nonalcoholic steatohepatitis: A follow-up study of forty-two patients for up to 21 years. Hepatology 11:74–80. Powell LW, Jazwinska EC, Halliday JW. 1994. Primary iron overload. In: Brock JH, Halliday JW, Pippard MJ, Powell LW, eds. Iron Metabolism in Health and Disease. London: WB Saunders, pp. 228–270. Poynard T, Bedossa P, Opolon P for the OBSVIRC, METAVIR, CLINIVIR, and DOSVIRC groups. 1997. Natural history of liver fibrosis progression in patients with chronic hepatitis C. Lancet 349:825–832. Pugh RN, Murray-Lyon IM, Dawson JL, Pietroni MC, Williams R. 1973. Transection of the oesophagus for bleeding oesophageal varices. Br J Surg 60:646–649. Pyong SJ, Tsukuma H, Hiyama T. 1994. Case-control study of hepatocellular carcinoma among Koreans living in Osaka, Japan. Jpn J Cancer Res 85:674–679. Qian GS, Ross RK, Yu MC, et al. 1994. A follow-up study of urinary markers of aflatoxin exposure and liver cancer risk in Shanghai, People’s Republic of China. Cancer Epidemiol Biomarkers Prev 3:3–10. Qin G, Gopalan-Kriczky P, Su J, Ning Y, Lotlikar PD. 1997. Inhibition of aflatoxin B1-induced initiation of hepatocarcinogenesis in the rat by green tea. Cancer Lett 112:149–154. Qin G, Ning Y, Lotlikar PD. 2000. Chemoprevention of aflatoxin B1-initiated and carbon tetrachloride-promoted hepatocarcinogenesis in the rat by green tea. Nutr Cancer 38:215–222. Qu D, Teckman JH, Perlmutter DH. 1997. Metabolic liver diseases. Review: alpha 1-antitrypsin deficiency associated liver disease. J Gastroenterol Hepatol 12:404–416. Regelson W. 1968. Hemangioendothelial sarcoma of the liver from chronic arsenic intoxication by Fowler’s solution. Cancer 21:514–522. Rennke H, Prat G, Etcheverry R, Katz R, Donoso S. 1971. Hemangioendotelioma maligno del higado y arsenicismo cronico. Rev Med Chile 99:664–668. Rezvani I. 2000. Defects in metabolism of amino acids. In: Behrman RE, Kliegman RM, Jenson HB, eds. Nelson Textbook of Pediatrics. Philadelphia: WB Saunders, pp. 345–377. Ries LAG, Eisner MP, Kosary CL, et al., eds. 2003. SEER Cancer statistics review, 1975–2000. National Cancer Institute, Bethesda, MD. Available at: http://seer.cancer.gov/csr/1975_2000. Roberts AG, Whatley SD, Morgan RR, Worwood M, Elder GH. 1997. Increased frequency of the haemochromatosis Cys282Tyr mutation in sporadic porphyria cutanea tarda. Lancet 349:321–323. Roetto A, Totaro A, Cazzola M, et al. 1999. Juvenile hemochromatosis locus maps to chromosome lq. Am J Hum Genet 64:1388–1393. Rooks JB, Ory HW, Ishak KG, et al. 1979. Epidemiology of hepatocellular adenoma. JAMA 242:644–648. Rosen CB, Najorney DM, Wiesner RH, Coffey RJ Jr., Larusso NF. 1991. Cholangiocarcinoma complicating primary sclerosing cholangitis. Ann Surg 213:21–25. Ross RK, Yuan JM, Yu MC, et al. 1992. Urinary aflatoxin biomarkers and risk of hepatocellular carcinoma. Lancet 339:943–946. Roth F. 1957. The sequelae of chronic arsenic poisoning in Moselle vintners. German Med Month 2:172–175. Roudot-Thoraval F, Bastie A, Pawlotsky JM, Dhumeaux D. 1997. Epidemiological factors affecting the severity of hepatitis C virus-related liver disease: A French survey of 6,664 patients. Hepatology 26:485–490. Sampietro M, Piperno A, Lupica L. 1998. High prevalence of the His63Asp HFE mutation in Italian patients with porphyria cutanea tarda. Hepatology 27:181–184. Sasaki A. 1994. Mortality and causes of death in patients with diabetes mellitus in Japan. Diabetes Res Clin Pract 24 Suppl:S299–306. Sasaki A, Uehara M, Horiuchi N, Hasegawa K, Shimizu T. 1996. A 15 year follow-up study of patients with non-insulin dependent diabetes mellitus (NIDDM) in Osaka, Japan. Long-term prognosis and causes of death. Diabetes Res Clin Pract 34:47–55. Sauvaget C, Nagano J, Hayashi M, Spencer E, Shimizu Y, Allen N. 2003. Vegetables and fruit intake and cancer mortality in the Hiroshima/Nagasaki Life Span Study. Br J Cancer 88:689–694. Schafer DF, Sorrell MF. 1999. Hepatocellular carcinoma. Lancet 353:1253– 1257. Schottenfeld D. 1979. Alcohol in the etiology of cancer. Cancer 43:1962– 1966. Seef LB, Buskell-Bales Z, Wright EC, et al. 1992. Long term mortality after transfusion associated non-A, non-B hepatitis. N Engl J Med 327:1906– 1911. Seef LB, Hollinger FB, Alter HJ, et al. 2001. Long-term mortality and morbidity of transfusion-associated non-A, non-B, and type C hepatitis: A

National Heart, Lung, and Blood Institute collaborative study. Hepatology 33:455–463. Seef LB, Miller RN, Rabkin CS, et al. 2000. 45-year follow-up of hepatitis C virus infection in healthy young adults. Ann Int Med 132:105–111. Seeger C, Mason WS. 2000. Hepatitis B virus biology. Microbial Mol Biol Rev 64:51–68. SEER (Surveillance, Epidemiology, and End Results) Program. 2003 (www.seer.cancer.gov) SEER*Stat Database: Incidence—SEER 9 Regs, Nov 2002 Sub (1973–2000), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2003, based on the November 2002 submission. Seitz HK, Poschl G, Simanowski UA. Alcohol and cancer. 1998. Recent Development Alcoholism 14:67–95. Shafritz DA, Shouval D, Sherman HI, et al. 1981. Integration of hepatitis B virus DNA into the genome of liver cells in chronic liver disease and hepatocellular carcinoma. N Engl J Med 305:1067–1073. Shibata A, Fukuda K, Nishiyori A, Ogimoto I, Sakata R, Tanikawa K. 1998. A case-control study on male hepatocellular carcinoma based on hospital and community controls. J Epidemiol 8:1–5. Shibata A, Hirohata T, Toshima H, Tashiro H. 1986. The role of drinking and cigarette smoking in the excess deaths from liver cancer. Jpn J Cancer Res 77:287–295. Shimada M, Hashimoto E, Taniai M, et al. 2002. Hepatocellular carcinoma in patients with non-alcoholic steatohepatitis. J Hepatol 37:154–160. Shin HR, Lee CU, Park HJ, et al. 1996. Hepatitis B and C virus, clonorchis sinensis for the risk of liver cancer: A case-control study in Pusan, Korea. Int J Epidemiol 25:933–940. Silvestri L, Lonzogni L, De Silvestri A, et al. 2003. CYP enzyme polymorphisms and susceptibility to HCV-related chronic liver disease and liver cancer. Int J Cancer 104:310–317. Simon M, Bourel M, Fauchet R, Genetet B. 1976. Association of HLA-A3 and HLA-B14 antigens with idiopathic hemochromatosis. Gut 17:332– 334. Sithithaworn P, Haswell-Elkins MR, Mairiang P, et al. 1994. Parasiteassociated morbidity: Liver fluke infection and bile duct cancer in northeast Thailand. Int J Parasitol 24:833–843. Snyder PJ. 2001. Androgens. In: Goodman & Gilman’s The Pharmalogical Basis of Therapeutics. Hardman JG, Limbird LE, eds. New York: McGraw-Hill, pp. 1635–1648. Soe KL, Soe M, Gluud CN. 1992. Liver pathology associated with the use of anabolic-androgenic steroids. Liver 12:73–79. Sorensen HT, Friis S, Olsen JH, et al. 1998. Risk of liver and other types of cancer in patients with cirrhosis: A nationwide cohort study in Denmark. Hepatology 28:921–925. Sorensen HT, Mellemkjaer L, Jepsen P, et al. 2003. Risk of cancer in patients hospitalized with fatty liver. J Clin Gastroenterol 36:356–359. Srinivasan R, Dean HA. 1997. Thorotrast and the liver—revisited. J Toxicol Clin Toxicol 35:199–202. Srivatanakul P. 2001. Epidemiology of liver cancer in Thailand. Asian Pac J Cancer Prev 2:117–121. Srivatanakul P, Parkin DM, Khlat M, et al. 1991. Liver cancer in Thailand. II. A case-control study of hepatocellular carcinoma. Int J Cancer 48:329–332. Stanford JL, Thomas DB, WHO Collaborative Study of Neoplasia and Steroid Contraceptives. 1992. Reproductive factors in the etiology of hepatocellular carcinoma. Cancer Causes Control 3:37–42. Stemhagen A, Slade J, Altman R, et al. 1983. Occupational risk factors and liver cancer: A retrospective case-control study of primary liver cancer in New Jersey. Am J Epidemiol 117:443–454. Stevens CE, Beasley RP, Tsui J, Lee WC. 1975. Vertical transmission of hepatitis B antigen in Taiwan. N Engl J Med 292:771–774. Stevens RG, Beasley RP, Blumberg BS. 1986. Iron-binding proteins and risk of cancer in Taiwan. J Natl Cancer Inst 76:605–610. Stocker JT, Ishak KG. 1987. Hepatoblastoma. In: Okuda K, Ishak KG, eds. Neoplastic Diseases of the Liver. Berlin: Springer-Verlag, pp. 127–136. Strachan AS. 1929. Haemosiderosis and haemochromatosis in South African natives with a comment on the etiology of haemochromatosis. MD Thesis: University of Glasgow, Glasgow, Scotland. Struben VM, Hespenheide EE, Caldwell SH. 2000. Nonalcoholic steatohepatitis and cryptogeneic cirrhosis within kindreds. Am J Med 108:9–13. Summers J, Jilbert AR, Yang W, et al. 2003. Hepatocyte turnover during resolution of a transient hepadnaviral infection. Proc Natl Acad Sci USA 100:11652–11659. Summers J, Smolec JM, Snyder R. 1978. A virus similar to hepatitis B associated with hepatitis and hepatoma in woodchucks. Proc Natl Acad Sci USA 75:4533–4537. Sun C-A, Wang L-Y, Chen C-J, et al. 2001. Genetic polymorphisms of glutathione-s-transferases M1 and T1 associated with susceptibility to

Liver Cancer aflatoxin-related hepatocarcinogenesis among chronic hepatitis B carriers: A nested case-control study in Taiwan. Carcinogenesis 22:1289–1294. Suriawinata A, Ivanov K, Ben Haim M, Schwartz ME. 2000. A 67-year-old man with hepatitis C virus infection and a liver tumor. Semin Liver Dis 20:227–231. Sveger T. 1976. Liver disease in a1-antitrypsin deficiency detected by screening of 200,000 infants. N Engl J Med 294:1316–1321. Tabershaw IR, Gaffey WR. 1974. Mortality study of workers in the manufacture of vinyl chloride and its polymers. J Occup Med 16:509–518. Tagger A, Donato F, Ribero ML, et al. 1999. Case-control study on hepatitis C virus (HCV) as a risk factor for hepatocellular carcinoma: The role of HCV genotypes and the synergism with hepatitis B virus and alcohol. Brescia HCC Study. Int J Cancer 81:695–699. Takaishi H, Kitamoto M, Takahashi S, et al. 2000. Precancerous hepatic nodules had significant levels of telomerase activity determined by sensitive quantitation using a hybridization protection assay. Cancer 88:312–317. Takemura Y, Kikuchi S, Inaba Y. 1998. Epidemiologic study of the relationship between schistosomiasis due to Schistosoma japonicum and liver cancer/cirrhosis. Am J Trop Med Hyg 59:551–556. Takeshita T, Yang X, Inoue Y, Sato S, Morimoto K. 2000. Relationship between alcohol drinking, ADH2 and ALDH2 genotypes, and risk for hepatocellular carcinoma in Japanese. Cancer Lett 149:69–76. Tan AY, Lao CH, Lee YS. 1977. Hepatitis B antigen in the liver cells in cirrhosis and hepatocellular carcinoma. Pathology 9:57–64. Tanaka K, Hirohata T, Fukuda K, et al. 1995. Risk factors for hepatocellular carcinoma among Japanese women. Cancer Causes Control 6:91–98. Tanaka K, Hirohata T, Takashita S. 1988. Blood transfusion, alcohol consumption, and cigarette smoking in causation of hepatocellular carcinoma: A case-control study in Fukuoka, Japan. Jpn J Cancer Res (Gann) 79:1075–1082. Tanaka K, Hirohata T, Takashita, et al. 1992. Hepatitis B virus, cigarette smoking and alcohol consumption in the development of hepatocellular carcinoma: A case-control study in Fukuoka, Japan. Int J Cancer 51: 509–514. Tanaka T, Imamura A, Hayashi S, et al. 1998a. Minute mixed hepatoma with two components: Hepatocellular and cholangiocarcinoma, which developed on liver cirrhosis with HCV. Hepatogastroenterology 45:220–223. Tanaka K, Sakai H, Hashizume M, et al. 1998b. A long-term follow-up study on risk factors for hepatocellular carcinoma among Japanese patients with liver cirrhosis. Jpn J Cancer Res 89:1241–1250. Tavani A, Negri E, Parazzini F, Franceschi S, La Vecchia C. 1993. Female hormone utilisation and risk of hepatocellular carcinoma. Br J Cancer 67:635–637. Taylor-Robinson SD, Toledano MB, Arora S, et al. 2001. Increase in mortality rates for intrahepatic cholangiocarcinomas in England and Wales, 1968–1998. Gut 48:816–820. The UK Haemochromatosis Consortium. 1997. A simple genetic test identifies 90% of UK patients with haemochromatosis. Gut 41:841–844. Thio CL. 2003. Hepatitis B in the human immunodeficiency virus-infected patient: Epidemiology, natural history, and treatment. Semin Liver Dis 23:125–136. Thio CL, Seaberg EC, Skolasky, et al. 2002. HIV-1, hepatitis B virus, and risk of liver-related mortality in the Multicenter Cohort Study (MACS). Lancet 360:1921–1926. Thorgeirsson SS, Grisham JW. 2002. Molecular pathogenesis of human hepatocellular carcinoma. Nature Genetics 31:339–346. Tiemersma EW, Omer RE, Bunschoten A, et al. 2001. Role of genetic polymorphism of glutathione-S-transferase T1 and microsomal epoxide hydrolase in aflatoxin-associated hepatocellular carcinoma. Cancer Epidemiol Biomarkers Prev 10:785–791. Tomimatsu M, Ishiguro N, Taniai M, et al. 1993. Hepatitis C virus antibody in patients with primary liver cancer (hepatocellular carcinoma, cholangiocarcinoma, and combined hepatocellular-cholangiocarcinoma). Cancer 72:683–688. Tong MJ, El-Farra NS, Reikes AR, Co RL. 1995. Clinical outcomes after transfusion-associated hepatitis C. N Engl J Med 332:1463–1466. Tornillo L, Carafa V, Sauter G, et al. 2002. Chromosomal alterations in hepatocellular nodules by comparative genomic hybridization: High-grade dysplastic nodules represent early stages of hepatocellular carcinoma. Lab Invest 82:547–553. Toshkov I, Hacker HJ, Roggendorf M, Bannasch PJ. 1990. Phenotypic patterns of preneoplastic and neoplastic hepatic lesions in woodchucks infected with woodchuck hepatitis virus. Cancer Res Clin Oncol 116:581–590. Toyoda H, Kumada T, Nakano S, et al. 2001. Effect of the dose and duration of interferon-alpha therapy on the incidence of hepatocellular carcinoma in noncirrhotic patients with a nonsustained response to interferon for chronic hepatitis C. Oncology 61:134–142.

785

Travis LB, Kathren RL, Boice JD. 1992. Cancer risk following exposure to Thorotrast: Overview in relation to a case report. Health Phys 63:89–97. Travis LB, Land CE, Andersson M, et al. 2001. Mortality after cerebral angiography with or without radioactive Thorotrast: An international cohort of 3,143 two-year survivors. Radiat Res 156:136–150. Tremolada F, Casarin C, Alberti A, et al. 1992. Long-term follow-up of nonA, non-B (type C) post-transfusion hepatitis. J Hepatol 16:273–281. Trichopoulos D, Day N, Kaklamani E, et al. 1987. Hepatitis B virus, tobacco smoking and ethanol consumption in the etiology of hepatocellular carcinoma. Int J Cancer 39:45–49. Trichopoulos D, MacMahon B, Sparros L, et al. 1980. Smoking and hepatitis B-negative primary hepatocellular carcinoma. J Natl Cancer Inst 65: 111–114. Tsuda T, Babazono A, Yamamoto E, et al. 1995. Ingested arsenic and internal cancer: A historical cohort study followed for 33 years. Am J Epidemiol 141:198–209. Tsukazaki N, Watanabe M, Irifune H. 1998. Porphyria cutanea tarda and hepatitis C virus infection. Br J Dermatol 138:1015–1017. Tsukuma H, Hiyama T, Oshima A, et al. 1990. A case-control study of hepatocellular carcinoma in Osaka, Japan. Int J Cancer 45:231–236. Tsukuma H, Hiyama T, Tanaka S, et al. 1993. Risk factors for hepatocellular carcinoma among patients with chronic liver disease. N Engl J Med 328:1797–1801. Tu J, Gao R, Zhang D, et al. 1985. Hepatitis B virus and primary liver cancer on Chongming Island, People’s Republic of China. Natl Cancer Inst Monogr 69:213–215. Tuyns AJ, Pequignot G. 1984. Greater risk of ascitic cirrhosis in women in relation to alcohol consumption. Int J Epidemiol 13:53–57. Tzonou A, Trichopoulos D, Kaklamani E, et al. 1991. Epidemiologic assessment of interactions of hepatitis-C virus with seromarkers of hepatitis-B and -D viruses, cirrhosis and tobacco smoking in hepatocellular carcinoma. Int J Cancer 49:377–380. Tzonou A, Zavitsanos X, Hsieh CC, Trichopoulos D. 1992. Liveborn children and risk of hepatocellular carcinoma. Cancer Causes Control 3:171–174. Vall Mayans M, Calvet X, Bruix J, et al. 1990. Risk factors for hepatocellular carcinoma in Catalonia, Spain. Int J Cancer 46:378–381. Van Kaick G, Dalheimer A, Hornik S, et al. 1999. The German Thorotrast study: Recent results and assessment of risks. Radiat Res 152:S64–S71. Velazquez RF, Rodriguez M, Navascues CA, et al. 2003. Prospective analysis of risk factors for hepatocellular carcinoma in patients with liver cirrhosis. Hepatology 37:520–527. Vessey CJ, de la Hall PM. 2001. Hepatic stem cells: a review. Pathology 33:130–141. Wade N. 1972. Anabolic steroids (news). Science 176:1399–1403. Wahib AA, Masoud AA, Halem AA, et al. 1998. Cell mediated immune response in chronic liver diseases: Schistosomal, viral and neoplastic. J Egypt Soc Parasitol 28:929–939. Ward E, Boffetta P, Andersen A, et al. 2001. Update of the follow-up of mortality and cancer incidence among European workers employed in the vinyl chloride industry. Epidemiology 12:710–718. Watanapa P, Watanapa WB. 2002. Liver fluke-associated cholangiocarcinoma. Br J Surg 89:962–970. Weiderpass E, Gridley G, Nyren O, Pennello G, Landstrom AS, Ekbom A. 2001. Cause-specific mortality in a cohort of patients with diabetes mellitus: A population-based study in Sweden. J Clin Epidemiol 54:802–809. WHO. 2000. Hepatitis C Fact Sheet. WHO fact sheet no. 164. (internet). WHO. 2002. Global burden of disease. In: The World Health Report 2002, Reducing risks, promoting healthy life, Geneva. WHO Collaborative Study of Neoplasia and Steroid Contraceptives. 1989. Combined oral contraceptives and liver cancer. Int J Cancer 43:254–259. WHO Collaborative Study of Neoplasia and Steroid Contraceptives. 1991. Depot-medroxyprogesterone acetate (DMPA) and risk of liver cancer. Int J Cancer 49:182–185. Wideroff L, Gridley G, Mellemkjaer L, et al. 1997. Cancer incidence in a population-based cohort of patients hospitalized with diabetes mellitus in Denmark. J Natl Cancer Inst 89:1360–1365. Wiese M, Berr F, Lafrenz M, et al. 2000. Low frequency of cirrhosis in a hepatitis C (genotype 1b) single-source outbreak in Germany: A 20-year multicenter study. Hepatology 32:91–96. Wong NA, Rae F, Bathgate A, Smith CA, Harrison DJ. 2000. Polymorphisms of the gene for microsomal epoxide hydrolase and susceptibility to alcoholic liver disease and hepatocellular carcinoma in a Caucasian population. Toxicol Lett 115:17–22. Wong RH, Chen PC, Du CL, Wang JD, Cheng TJ. 2002. An increased standardized mortality ratio for liver cancer among polyvinyl chloride workers in Taiwan. Occup Environ Med 59:405–409. World Cancer Research Fund, American Institute for Cancer Research. 1997. Liver. In Food, Nutrition and the Prevention of Cancer: A global

786

PART IV: CANCER BY TISSUE OF ORIGIN

perspective. Washington, D.C., American Institute for Cancer Research, pp. 202–214. Wright TL. 2002. Treatment of patients with hepatitis C and cirrhosis. Hepatology 36:S185–194. Wu MM, Kuo TL, Hwang YH, Chen CJ. 1989. Dose-response relation between arsenic concentration in well water and mortality from cancers and vascular diseases. Am J Epidemiol 130:1123–1132. Yamamoto M, Takasaki K, Nakano M, Saito A. 1998. Minute nodular intrahepatic cholangiocarcinoma. Cancer 82:2145–2149. Yang BH, Tang ZY. 1985. Hepatitis b virus and hepatocellular carcinoma— clinical and serological aspects. In: Tang ZY, ed. Sub-Clinical Hepatocellular Carcinoma. Berlin: Springer-Verlag, pp. 212–217. Yano M, Kumada H, Kage M, et al. 1996. The long-term pathological evolution of chronic hepatitis C. Hepatology 23:1334–1340. Ye B, Shen J, Xu Y. 1994. Etiologic study on the relationship between HBV, HCV and HCC. Zhonghua Liu Xing Bing Xue Za Zhi 15:131–134. Yee TT, Griffioen A, Sabin CA, Dusheiko G, Lee CA. 2000. The natural history of HCV in a cohort of haemophilic patients infected between 1961 and 1985. Semin Liver Dis 23:149–166. Yeh SH, Chang CF, Shau WY, et al. 2002. Dominance of functional androgen receptor allele with longer CAG repeat in hepatitis B virus-related female hepatocarcinogenesis. Cancer Res 62:4346–4351. Yin F, Chen B. 1998. Detection of hepatitis C virus RNA sequences in hepatic portal cholangiocarcinoma tissue by reverse transcription polymerase chain reaction. Chin Med J 111:1068–1070. Yoshida H, Shiratori Y, Moriyama M, et al. 1999. Interferon therapy reduces the risk for hepatocellular carcinoma: National surveillance program of cirrhotic and noncirrhotic patients with chronic hepatitis C in Japan. Ann Intern Med 131:174–181. Yu H, Harris R, Kabat G, et al. 1988a. Cigarette smoking, alcohol consumption and primary liver cancer: A case-control study in the U.S.A. Int J Cancer 42:325–328. Yu MC, Mack T, Hanish R, et al. 1983. Hepatitis, alcohol consumption, cigarette smoking and hepatocellular carcinoma in Los Angeles. Cancer Res 42:6077–6079. Yu MC, Tong MJ, Govindarajan S, Henderson BE. 1991a. Nonviral risk factors for hepatocellular carcinoma in a low-risk population, the non-Asians of Los Angeles County, California. J Natl Cancer Inst 83:1820–1826. Yu MW, Chen CJ. 1993. Elevated serum testosterone levels and risk of hepatocellular carcinoma. Cancer Res 53:790–794. Yu MW, Cheng SW, Lin MW, et al. 2000b. Androgen-receptor gene CAG repeats, plasma testosterone levels, and risk of hepatitis B-related hepatocellular carcinoma. J Natl Cancer Inst 92:2023–2028. Yu MW, Chiu YH, Chiang YC, et al. 1999b. Plasma carotenoids, glutathione s-transferase M1 and T1 genetic polymorphisms, and risk of hepato-

cellular carcinoma: Independent and interactive effects. Am J Epidemiol 149:621–629. Yu MW, Chiu YH, Yang SY, et al. 1999c. Cytochrome p4501A1 genetic polymorphisms and risk of hepatocellular carcinoma among chronic hepatitis B carriers. Br J Cancer 80:598–603. Yu MW, Gladek-Yarborough A, Chiamprasert S, Santella RM, Liaw YF, Chen CJ. 1995. Cytochrome p450 2E1 and glutathione s-transferase M1 polymorphisms and susceptibility to hepatocellular carcinoma. Gastroenterology 109:1266–1273. Yu MW, Horng IS, Hsu KH, Chiang YC, Liaw YF, Chen CJ. 1999a. Plasma selenium levels and risk of hepatocellular carcinoma among men with chronic hepatitis infection. Am J Epidemiol 150:367–374. Yu MW, Hsu FC, Sheen IS, et al. 1997a. Prospective study of hepatocellular carcinoma and liver cirrhosis in asymptomatic chronic hepatitis B virus carriers. Am J Epidemiol 145:1039–1047. Yu MW, Pai CI, Yang SY, et al. 2000a. Role of N-acetyltransferase polymorphisms in hepatitis B related hepatocellular carcinoma: Impact of smoking on risk. Gut 47:703–709. Yu MW, Yang YC, Yang SY, et al. 2001. Hormonal markers and hepatitis B virus-related hepatocellular carcinoma risk: A nested case-control study among men. J Natl Cancer Inst 93:1644–1651. Yu SY, Chu YJ, Li WG. 1988b. Selenium chemoprevention of liver cancer in animals and possible human applications. Biol Trace Elem Res 15:231–241. Yu SY, Zhu YJ, Li WG. 1997b. Protective role of selenium against hepatitis B virus and primary liver cancer in Qidong. Biol Trace Elem Res 56:117–124. Yu SY, Zhu YJ, Li WG, et al. 1991b. A preliminary report on the intervention trials of primary liver cancer in high-risk populations with nutritional supplementation of selenium in China. Biol Trace Elem Res 29:289– 294. Yuan JM, Ross, RK, Stanczyk FZ, et al. 1995. A cohort study of serum testosterone and hepatocellular carcinoma in Shanghai, China. Int J Cancer 63:491–493. Yuen MF, Hui CK, Cheng CC, Wu CH, Lai YP, Lai CL. 2001. Long-term follow-up of interferon alfa treatment in Chinese patients with chronic hepatitis B infection: The effect on hepatitis B e antigen seroconversion and the development of cirrhosis-related complications. Hepatology 34:139–145. Zen Y, Katayanagi K, Tsuneyama K, Harada K, Araki I, Nakanuma Y. 2001. Hepatocellular carcinoma arising in non-alcoholic steatohepatitis. Pathol Int 51:127–131. Zhou H, Ortiz-Pallardo ME, Ko Y, Fischer H-P. 2000. Is heterozygous alpha1-antitrysin deficiency type PiZ a risk factor for primary liver cancer? Cancer 88:2668–2676.

40

Biliary Tract Cancer ANN W. HSING, ASIF RASHID, SUSAN S. DEVESA, AND JOSEPH F. FRAUMENI, JR.

I

n the United States, cancer of the biliary tract, encompassing tumors arising from the gallbladder, extrahepatic bile ducts, and ampulla of Vater (ICD10 = C23, C24) (World Health Organization, 1992), is the fifth most common malignant neoplasm of the digestive tract (Ries and Devesa, Chapter 9, this volume). Tumors of the intrahepatic bile ducts (ICD10 = C22.1) are classified with primary liver cancer (ICD10 = C22) and are not included in this category. Biliary tract cancers account for about 3500 deaths per year in the United States (Jemal et al., 2004). Gallbladder cancer occurs twice as often in women, while other biliary tract tumors are 50% more common in men (Ries and Devesa, Chapter 9, this volume). Because most patients have regional or distant disease at diagnosis, the prognosis is generally poor, with 5year relative survival ranging from 15%–19%. According to 1992–2000 data from the Surveillance, Epidemiology, and End Results (SEER) program, gallbladder cancer accounts for 46% of the biliary tract cancers in the United States (Ries et al., 2003). Of the other biliary tract tumors, about one half arise in the extrahepatic bile ducts and 40% in the ampulla of Vater, while the remaining 10% have an unspecified subsite (Ries et al., 2003). Histologic confirmation of diagnoses in the SEER program occurs in about 93% of gallbladder cancer cases and about 86% of other biliary subsites. Many etiologic leads for biliary tract cancer have come from clinical observations, autopsy series, and descriptive epidemiologic studies. While a significant fraction of these tumors are related to gallstones (cholelithiasis), information on other risk factors is limited, due to the rarity of the tumors, the often rapidly fatal course, and the small number of epidemiologic studies conducted to date. Because the three anatomic categories of biliary tract cancer have distinct epidemiologic patterns and molecular changes, including somatic mutations and loss of heterozygosity (LOH), it has been suggested that the causal factors vary by subsite.

CLASSIFICATION Anatomic Distribution Anatomically, the biliary tract includes the gallbladder, extrahepatic bile ducts, and ampulla of Vater (Fig. 40–1). The extrahepatic bile ducts are further subdivided into the left and right hepatic ducts, common hepatic duct, cystic duct, and common bile duct. Most gallbladder carcinomas (70%) originate in the fundus of the gallbladder, with 20% in the body and 10% in the neck (Lazcano-Ponce et al., 2001). Most extrahepatic bile duct tumors occur in the upper third of the biliary tree (including the left and right hepatic ducts and the common hepatic duct), where they are often termed proximal or hilar cholangiocarcinomas (Klatskin tumors). The ampulla of Vater is located at the distal portion of the bile duct where it meets the junction of the pancreatic duct and the duodenum at the sphincter of Oddi. Clinically, it is difficult to differentiate ampullary cancers from tumors arising in the distal bile duct, head of the pancreas, or second part of the duodenum.

Histopathology The gallbladder, bile ducts, and ampulla of Vater give rise to a variety of epithelial tumors, with gland-forming adenocarcinomas occurring in at least 75% of cases (Albores-Saavedra et al., 2000). Next most

frequent are papillary and mucinous (colloid) adenocarcinomas, squamous cell, and adenosquamous tumors, each accounting for 5%–10% of all biliary tract cancers. Compared with other histologic types, patients with papillary adenocarcinomas have a better prognosis (Henson et al., 1992).

Precursor Lesions Although little is known about the natural history of precursor lesions, examination of cholecystectomy specimens removed from patients with gallstones, cholecystitis, and cancer has suggested a series of epithelial changes evolving through hyperplasia, atypical hyperplasia, metaplasia, dysplasia, and carcinoma in situ (Albores-Saavedra et al., 1980; Dowling and Kelly, 1986). Recent data have further shown that most gallbladder and bile duct cancers arise from dysplasia, with the metaplasia-dysplasia-carcinoma sequence as the major pathway (Kurashina et al., 1988; Wistuba et al., 1999), while most ampullary cancers arise from adenomas (Yamaguchi and Enjoji, 1987).

MOLECULAR CHARACTERISTICS Biliary tract cancers have been associated with a number of molecular changes, including point mutations of K-ras and b-catenin protooncogenes as well as inactivation of p53, p16, adenomatous polyposis coli (APC), and deleted in pancreatic cancer (DPC4) tumor suppressor genes (Rashid et al., 2002). Although the precise sequence of molecular changes leading to neoplastic transformation is unknown, the causal mechanisms appear to vary by anatomic subsite, histologic type, and stage. Recent studies using genome-wide allelotyping analysis of biliary tract cancer have identified 3p, 8p, 9q, and 22q as chromosomal regions with frequent LOH (Nakayama et al., 2001; Wistuba et al., 2001). K-ras is the most extensively studied oncogene, with activating K-ras mutations present in 10%–50% of gallbladder cancer, 5%–100% of extrahepatic bile duct cancer, and 31%–61% of ampullary cancer (Rashid, 2002). K-ras mutations appear to be an unfavorable prognostic factor for bile duct cancer (Malats et al., 1995; Rashid et al., 2002) but not for tumors of other subsites. Mutations or LOH of the p53 gene are frequently found in all three subsites of biliary tract cancer, with a prevalence ranging from 22% to over 90% (Rashid et al., 2002). Similarly, alterations of the p16 gene are present in 55%–71% of all subsites and are unfavorable prognostic factors (Ueki et al., 2004). Dysregulation of b-catenin signaling due to inactivation of the APC gene or b-catenin mutations is an important event in a variety of tumors, but such changes are less frequent in biliary tract tumors. The prevalence and pattern of APC gene alterations or b-catenin mutations vary according to the three subsites of biliary cancer, with ampullary tumors having more frequent allelic loss of chromosome 5q (chromosomal location of the APC gene) or APC gene mutations (Achille et al., 1996; Imai et al., 1997). Point mutations of the b-catenin gene are more commonly found in papillary carcinoma of the gallbladder and ampullary carcinoma but are absent in other types of biliary cancer (Rashid et al., 2001; Yanagisawa et al., 2001). The distinctive patterns of molecular events in subsites and cell types of biliary tract cancer are consistent with etiologic heterogeneity.

787

788

PART IV: CANCER BY TISSUE OF ORIGIN Liver Intrahepatic bile duct Gallbladder Left and right hepatic ducts Common hepatic duct Cystic duct

Extrahepatic bile ducts

The Biliary Tract

Common bile duct

Ampulla of Vater Pancreas

Duodenum Pancreatic duct

Figure 40–1. Anatomy of the biliary tract.

Molecular changes also vary by precursor lesion. For example, both dysplasia and carcinoma of the gallbladder are often accompanied by alterations of the p53 gene but lack K-ras mutations (Wistuba et al., 1996; Rashid et al., 2002), whereas gallbladder adenoma has frequent K-ras mutations (Wistuba et al., 1999).

DESCRIPTIVE PATTERNS International Variation Based on data from Volume 8 of Cancer Incidence in Five Continents (Parkin et al., 2003), Figure 40–2 displays biliary tract cancer rates by subsite, ranked by gallbladder cancer rates among women within each continent. The population-based registries shown were selected from registries with high rates of histologic confirmation, a sufficient number of cases, and stable risk estimates. In most parts of the world, the incidence is low (<10 per 100,000 person-years for gallbladder cancer and <5 for other sites). In most populations, the most common subsite is the gallbladder, which is among the few forms of cancer that occur more often in women than men. Among women, the highest rate in the world is reported in Delhi, India (8.8 per 100,000 person-years). Rates for gallbladder cancer also are elevated in countries such as Ecuador, Colombia, and Uruguay in Latin America; among Hispanics in Los Angeles; in Korea and Japan in Asia; and the Czech Republic, Slovakia, and Poland in Eastern Europe. Notably high rates have been reported also in Chile and Peru (Parkin et al., 1992) but not in Puerto Rico, suggesting that the high rates observed in Latin America are primarily in Hispanic populations with Indian heritage. Among men, the geographic differences in gallbladder cancer are less striking, although countries with high rates for women generally have high rates for men. The highest rates for men are seen in Asia, particularly Korea and Japan, and in Eastern Europe, while the rates in Latin America are much less remarkable than the female pattern. Cancer of the extrahepatic bile ducts generally is less common than gallbladder cancer. In contrast to gallbladder cancer, the rates for bile duct cancer are usually higher in men than in women and display less geographic variation. Around the world, the rates in men range from 0.1–4.2, with Daegu, Korea and Osaka, Japan reporting the highest rates. Although the rates for cancer of the bile ducts in women are generally less than 1 per 100,000, rates in Japan and Korea are much higher (>2 per 100,000 person-years). Cancer of the ampulla of Vater is the least common of the three subsites in all populations, with rates generally lower than 1.0, except in Korea and Ecuador. In both men and women, elevated rates are generally found in countries with high rates of gallbladder and bile duct cancer, including Korea, Japan, Ecuador, and Poland.

Rates for cancer of the biliary tract with “subsite not specified” vary internationally and appear to reflect the pattern of gallbladder cancer, the most common of the biliary tumors. This may be the case for registries in Colombia, Ecuador, Uruguay, Spain, and Korea, where more than 80% of biliary tract cancers other than gallbladder represent tumors of unspecified subsite. Of special interest are the unusually high rates of biliary tract cancer in parts of Asia, such as Korea and India. In northern and eastern parts of India, gallbladder cancer is the third most common cancer in women (Sen et al., 2002; Kapoor and McMichael, 2003), and the high rates extend to Indian immigrants in the United Kingdom (Dhir and Mohandas, 1999). In Shanghai, a remarkable increase in the incidence of biliary tract cancer has been reported in recent times (Hsing et al., 1998), prompting a large-scale population-based study that is under analysis (Rashid et al., 2002; Liu et al., 2002).

US Rates In the United States, incidence data from the SEER program indicate substantial variation by racial/ethnic group and gender (Table 40–1). For gallbladder cancer, rates are especially high among American Indians/Alaska Natives and among the white Hispanic population, with a female excess in all ethnic groups. Figure 40–3 shows the age-specific incidence curves for gallbladder cancer among US men and women. Compared with Hispanic whites and American Indians/Alaska Natives, the rates among nonHispanic whites and blacks increase with age at a slower rate, with non-Hispanic whites having the lowest rates in almost every age group. The rates for gallbladder cancer are higher among women than men at virtually all ages, with the gender difference decreasing slightly with increasing age. For other subsites of biliary tract cancer (not shown), the male excess in all ethnic groups is apparent by age 45 and persists throughout life. Evaluation of time trends for biliary tract cancer has been impaired by failure of the International Classification of Diseases to distinguish between cancers of the liver and biliary system until the seventh revision in 1958 (World Health Organization, 1957), whereas cancers of the gallbladder and other biliary tract were not separately reported until the eighth revision in 1967 (Public Health Service, 1967). From the mid 1970s through the late 1990s, gallbladder cancer incidence and mortality decreased among white men and women, with mortality declining more steeply (42%–50%) than incidence (33%–34%) (Fig. 40–4). The incidence of extrahepatic bile duct cancer decreased less rapidly, but mortality rates declined faster. The incidence of cancers of ampulla of Vater changed little, in contrast with decreasing mortality. Incidence of and mortality from biliary tract cancer with subsite not specified remained fairly constant. Of the deaths attributed

789

Biliary Tract Cancer

Gallbladder

Extrahepatic Bile Ducts

Ampulla of Vater

Subsite not Specified

5 0 5 Female Male

5 0 5 Female Male

AMERICA, CENTRAL AND SOUTH Ecuador, Quito Colombia, Cali Uruguay, Montevideo Costa Rica USA, Puerto Rico AMERICA, NORTH USA, Los Angeles: Hispanic White Canada USA, SEER: Black USA, Los Angeles: Non-Hispanic White USA, SEER: White ASIA India, Delhi Korea, Busan Korea, Daegu Japan, Hiroshima Japan, Osaka Prefecture Japan, Miyagi Prefecture Korea, Seoul China, Shanghai India, Mumbai (Bombay) Israel: Non-Jews Israel: Jews China, Beijing Singapore: Chinese Thailand, Bangkok Taiwan EUROPE Czech Republic Slovakia Poland, Warsaw City Yugoslavia, Vojvodina Germany, Saarland Slovenia Italy, Varese Province France, Bas-Rhin Russia, St Petersburg Italy, Florence Spain, Zaragoza Switzerland, Zurich The Netherlands Denmark UK, Scotland UK, England Norway OCEANIA Australia, Queensland Australia, New South Wales New Zealand

10

0 Female

Male

10 5 0 5 Female Male

Incidence rates per 100,000 person-years Figure 40–2. International variation in biliary tract cancer incidence rates by subsite and gender (per 100,000 person-years, age-standardized to the world population): gallbladder cancer, extrahepatic bile duct cancer,

ampulla of Vater, and subsite not specified, circa 1993–1997. (Source: Parkin et al., 2003.)

to biliary subsites other than gallbladder cancer, 60%–70% were due to extrahepatic bile duct cancer, about 15% to ampulla of Vater cancer, and the remainder to subsite not specified. It is noteworthy that incidence rates in western countries are increasing for intrahepatic bile duct cancer (cholangiocarcinoma) as well as hepatocellular carcinoma (HCC) (McGlynn et al., 2001), in contrast to the downward trends for biliary tract cancer. Reasons for the increase in incidence of cholangiocarcinoma and HCC in western countries are unclear, but may be due to the rising prevalence of hepatitis C infection and obesity (McGlynn et al., 2001). In contrast, the decline in biliary tract cancer appears related at least partly to the increasing number of cholecystectomies performed for gallbladder disease in the United States (Blanken, 1976), which reduces the number of gallbladders at risk of developing cancer. A similar trend has been noted in other countries, including the United Kingdom (Khan et al., 2002; Levi et al., 2003; Wood et al., 2003), France (Manfredi et al., 2000), and Czechoslovakia (Plesko et al., 1985). The reduced risk after cholecystectomy extends to bile duct cancer as well as gallbladder cancer, which is consistent with the role of gallstones in both tumors (Ekbom et al., 1993). In Sweden, where the number of cholecystectomies has actually declined over time (Ahlberg et al., 1978), gallbladder and bile duct cancer rates have increased (Diehl and

Beral, 1981), particularly among older women (Broden et al., 1978). Similar trends have been reported in Chile (Serra et al., 1990). During the 1970s and 1980s, cholecystectomy rates in the United States increased mainly among persons aged 65 and older, whereas rates among younger persons showed little change (Diehl, 1987). The introduction of gallstone lithotripsy in 1986 may have temporarily reduced cholecystectomy rates; but during the late 1980s, laparoscopic cholecystectomy became increasingly popular, due to its lower morbidity when compared with open cholecystectomy (Johnston and Kaplan, 1993). Increases in cholecystectomy rates in the United States have been documented at all ages (Legorreta et al., 1993; Diehl et al., 1994), so that the rates for biliary tract cancer should continue to decline.

US Mortality Maps The mapping of US mortality data by state economic area has revealed a fivefold variation in rates of gallbladder cancer among whites (Fig. 40–5). Gallbladder cancer rates are elevated in portions of Appalachia, the midwestern and north-central regions, and parts of the Southwest, with generally low rates across the southeastern and far western states. The pattern is more pronounced among women, whose rates for

790

PART IV: CANCER BY TISSUE OF ORIGIN

Table 40–1. Incidence Rates* for Biliary Tract Cancer by Subsite, Gender, and Ethnic Group in the United States, 12 SEER areas†, 1992–2000 Males

Females

No.

Rate

No.

Rate

M : F ratio

780 531 134 83 24 160

0.8 0.7 1.4 0.8 3.3 1.4

2195 1368 563 218 48 250

1.6 1.2 4.2 1.6 4.1 1.8

0.36 0.58 0.33 0.50 0.80 0.78

0.9 0.8 1.1 0.8 1.9 1.5

878 628 119 67 16 107

0.6 0.6 0.9 0.5 1.4 0.8

1.50 1.33 1.22 1.60 1.36 1.88

759 558 117 52 6 107

0.8 0.7 1.2 0.6 0.8 0.9

627 461 94 62 3 120

0.4 0.4 0.7 0.4 0.2 0.8

2.00 1.75 1.71 1.40 4.00 1.13

136 96 25 16 0 22

0.1 0.1 0.3 0.2 0.0 0.2

205 151 28 13 7 26

0.1 0.1 0.2 0.1 0.6 0.2

1.00 1.00 1.50 2.00 0.00 1.00

gallbladder cancer White Non-Hispanic Hispanic Black American Indian/Alaska Native Asian/Pacific Islander

extrahepatic bile duct cancer White Non-Hispanic Hispanic Black American Indian/Alaska Native Asian/Pacific Islander

902 644 100 74 15 165

ampulla of vater cancer White Non-Hispanic Hispanic Black American Indian/Alaska Native Asian/Pacific Islander

subsite not specified White Non-Hispanic Hispanic Black American Indian/Alaska Native Asian/Pacific Islander

Source: SEER*Stat, 2003. *Per 100,000 person-years, age-standarized to the 2000 US population. † Data for non-Hispanics vs Hispanics are not available from Detroit or Hawaii; data for Alaska Natives available only from Alaska.

Rate per 100,000 person-years

100

Figure 40–3. Age-specific incidence rates for gallbladder cancer by gender and racial/ethnic group in the United States, 12 SEER areas, 1992–2000. (Source: SEER*Stat, 2003.)

various gallbladder diseases, including gallstones and cancer, are higher than those of men. The geographic distribution of “other biliary tract cancers” (including bile duct and ampulla of Vater) resembles that for gallbladder cancer, especially in females (Devesa et al., 1999). However, the variation is less pronounced, probably because gallstones are less conspicuous as a risk factor for these tumors. Despite the rarity of mortality from gallstones, the geographic patterns resemble those for gallbladder cancer (Fig. 40–5). The elevated rates of gallstones and gallbladder cancer among Hispanics largely account for the excessive mortality seen in the southwestern states, while a high incidence of both conditions has been reported in areas of Appalachia with low socioeconomic levels (Richardson et al., 1973). The elevated mortality observed in the north-central areas may be related to the concentration of high-risk ethnic groups from eastern Europe (Fraumeni et al., 1996), but further studies are needed. The concomitant variation in the incidence of gallbladder cancer and gallstones extends to many populations. At one extreme, for example, the elevated risk of gallbladder cancer among American Pima Indians appears to be part of a spectrum of gallstone-related disease endemic in this population (Morris et al., 1978). Also prone to both conditions are Alaskan natives, including Inuits (Nutting et al., 1993), and Hispanic populations with Indian admixture (Miquel et al., 1998; Everhart, 2001; Lazcano-Ponce et al., 2001) as well as Central and Eastern Europeans (Angelico et al., 1997; Zatonski et al., 1997). In Israel, the rates for gallbladder cancer and gallstones are higher among women born in Europe than among those born in Asia or Africa (Hart et al., 1971). At the other extreme, in sub-Saharan Africa, both gallbladder cancer and gallstones are rarely reported (Heaton, 1973).

GALLSTONES The presence of gallstones is the dominant risk factor for cancers of the biliary tract, and particularly gallbladder cancer. In the United States, 10%–15% of men and 25% of women over age 50 have gallstones, but the absolute risk of developing biliary tract cancer is very low (<1%) (de Groen et al., 1999; Lazcano-Ponce et al., 2001). In various surveys of gallbladder cancer, stones have been reported in 60%–90% of cases (Zatonski et al., 1997; Caroli-Bosc et al., 1999; de

100

Males

10

10

1

1

0.1

Females

0.1 20

40

60

80

100

20

Age American Indian / Alaska Native Asian / Pacific Islander Black

40

60

Age White Hispanic White non-Hispanic

80

100

791

Biliary Tract Cancer Incidence

Rate per 100,000 person-years

10

10

Males

1

Mortality

Females

1

Gallbladder

10

Males

1

0.1 0.1 1975 1980 1985 1990 1995 2000 1975 1980 1985 1990 1995 2000

Year

10

Year

Extrahepatic bile ducts

Females

1

0.1 0.1 1975 1980 1985 1990 1995 2000 1975 1980 1985 1990 1995 2000

Year

Ampulla of Vater

Groen et al., 1999; Everhart, 2001; Gurleyik et al., 2002; Rizvi and Zuberi, 2003; Singh and Choudhuri, 2003), a rate several times higher than expected, while in bile duct cancer the percentage of cases with stones is usually less than 50% (Csendes et al., 2000; Vitetta et al., 2000). Based mainly on autopsy surveys and case-control studies, a relationship has been established between gallstones and the subsequent occurrence of gallbladder cancer (Zatonski et al., 1997; Elnemr et al., 2001; Lazcano-Ponce et al., 2001; Tazuma and Kajiyama, 2001; Gurleyik et al., 2002). In a cohort study of gallstone patients who did not have cholecystectomy, the cumulative incidence of gallbladder cancer after 20 years was estimated to be about 1%, which represents a threefold elevation in relative risk (Maringhini et al., 1987). In a pooled analysis from Australia, Canada, Netherlands, and Poland, a history of gallbladder symptoms, mostly from gallstones, was associated with a fourfold excess risk of gallbladder cancer (Zatonski et al., 1997). Even higher risks (sixfold) were found among subjects reporting biliary symptoms for more than 20 years. In several studies, the risk of gallbladder cancer has been related to the number and size of gallstones as well as the duration of symptoms (Zatonski et al., 1997; Zou and Zhang, 2000; Serra et al., 2002). For example, subjects with stones over 3 cm in diameter have a much higher risk than those with stones less than 1 cm (Diehl, 1983; Lowenfels et al., 1989; Csendes et al., 2000; Vitetta et al., 2000; Zou and Zhang, 2000). Although less consistent, case-control studies have also noted an elevated risk of bile duct and ampullary cancers subsequent to benign biliary disease, including stones (Yen et al., 1987; Chow et al., 1994). In western countries, the risk of biliary tract cancer has been linked primarily to cholesterol gallstones. Individuals prone to these stones tend to have “lithogenic” bile that is supersaturated with cholesterol, due to increased hepatic secretion of cholesterol or diminished secretion of bile salts and phospholipids that maintain the solubility of cholesterol (Amigo et al., 1999; Dowling, 2000; Portincasa et al., 2003). The formation of stones is also enhanced by stasis and destabilization of bile in the gallbladder. Because of the close relationship between cholesterol stones and biliary tract cancer, a better understanding of the origins of gallstones should help elucidate the causes of biliary tract cancer. Most of the risk factors for cholesterol stones are associated with hypersecretion and saturation of cholesterol in the bile. These factors include the

Year

Subsite not specified

Figure 40–4. Trends in incidence (9 SEER areas) and mortality (US) rates (per 100,000 person-years, agestandardized to the 2000 US population) for cancers of the gallbladder, extrahepatic bile ducts, ampulla of Vater, and subsite not specified by gender among whites, 1976–1980 to 1996–2000. (Source: SEER*Stat, 2003.)

steady increase in incidence with age, predominance in females, and the associations observed with obesity, physical inactivity, multiple pregnancies, and use of exogenous estrogens (Misciagna et al., 1996; Attili et al., 1997; Zhuang and Li, 2000; Lazcano-Ponce et al., 2001; Nakeeb et al., 2002). Although inconsistent, there is some evidence that high serum triglycerides, apolipoprotein B and E, and low levels of high-density lipoprotein cholesterol levels may be associated with an elevated risk of gallstones (Tang, 1996; Fu et al., 1997; Singh et al., 1997). Genetic susceptibility appears to contribute to the very high rates of gallstones occurring among American Indians and certain Hispanic populations as well as the familial tendency to develop cholesterol stones noted in some studies (Misciagna et al., 1996; Mittal and Mittal, 2002; Nakeeb et al., 2002). Data from animal models have also been informative. In inbred mice, a cholesterol gallstone susceptibility gene, the Lith gene, has been identified, with the encoded protein increasing hepatic secretion of biliary lipids (Lammert et al., 1999). A number of susceptibility genes related to lipid metabolism and cholesterol transport, such as polymorphisms of apolipoproteins A, B, and E, cholecystokinin (CCK ), lipoprotein lipase (LPL), and low-density lipoprotein receptor (LDLR) genes, have been linked to gallstones as well as high serum levels of cholesterol or triglyceride (Feng et al., 1998; Lin et al., 1999; Niemi et al., 1999; Suo et al., 1999; Han et al., 2000; Kosters et al., 2003; Jiang et al., 2004; Singh et al., 2004). Also of interest is the role of the cholesterol 7 alpha-hydroxylase (CYP7A) gene, a microsomal cytochrome P450 mapped to chromosome 8q11q12 that catalyzes the first stage of bile acid synthesis (Cohen et al., 1992), although its relation to gallstones is unclear. While gallbladder cancer and cholesterol gallstones share many epidemiologic characteristics, the carcinogenic mechanisms are poorly understood. It has not yet been possible to distinguish a direct effect of stones from an underlying property of bile that may lead to the formation of both stones and tumors. An effect of gallbladder hypomotility and stasis on stones is suggested by the elevated risk associated with pregnancy (Tavani et al., 1996; Zatonski et al., 1997; Scott et al., 1999; Rizvi and Zuberi, 2003), fasting (Attili et al., 1997), thiazide use (Kakar et al., 1986), and gastric resection (Kodama et al., 1996), but gallbladder hypomotility and stasis also increase the likelihood of prolonged contact with biliary carcinogens. A role of chronic inflammation has been suggested by precancerous changes of the gallbladder mucosa, including metaplasia and dysplasia, in areas adjacent

792

PART IV: CANCER BY TISSUE OF ORIGIN

Gallbladder Cancer

Gallstone Disease White Females

* Sparse data (<12 observed deaths; 24 SEAs; 0.42% of deaths)

US = 1.32/100,000 1.66 - 3.14 1.34 - 1.65 1.15 - 1.33 and sparse data * 0.94 - 1.14 0.37 - 0.93

* Sparse data (<12 observed deaths; 62 SEAs; 2.12% of deaths)

US = 0.65/100,000 0.86 - 1.85 0.71 - 0.85 0.67 - 0.70 and sparse data * 0.56 - 0.66 0.33 - 0.55

White Males

* Sparse data (<12 observed deaths; 156 SEAs; 6.56% of deaths)

US = 0.71/100,000 0.96 - 1.56 0.77 - 0.95 0.65 - 0.76 and sparse data * 0.55 - 0.64 0.29 - 0.54

* Sparse data (<12 observed deaths; 101 SE As; 4.18% of deaths)

US = 0.81/100,000 1.10 - 2.52 0.94 - 1.09 0.82 - 0.93 and sparse data * 0.67 - .081 0.30 - 0.66

Figure 40–5. Geographic variation in mortality rates for gallbladder cancer and gallstones (per 100,000 person-years, age-standardized to the 2000 US population) among US whites by state economic area, 1970–1999. Gallbladder cancer mortality among women, Gallstone mor-

tality among women, Gallbladder cancer mortality among men, and Gallstone mortality among men. (Source: Based on data updated from Devesa et al., 1999. Available at http://www3.cancer.gov/atlaspub1.)

to stones as well as tumors (Portincasa et al., 1997; Billo et al., 2000; Tazuma and Kajiyama, 2001). In contrast to cholesterol stones, the relationship between pigment stones and biliary tract cancer is unclear (Diehl et al., 1995; Ho et al., 1995; Kim et al., 1999; Stewart et al., 2002). Pigment stones, which occur more frequently in Asian populations, are further classified into black stones (commonly associated with excessive bilirubin secretion) and brown stones (related to biliary stasis and bacterial infection). The risk factors for pigment stones include advancing age, bacterial infections of the biliary tract, low-protein diet, hemolytic anemia, and liver cirrhosis (Diehl et al., 1995; Kurtin et al., 2000).

consider other risk factors. Their influence is suggested by evidence that only 1% of gallstone patients eventually develop biliary tract cancer, and that some tumors may arise in the absence of stones. However, information on specific risk factors is limited by the relatively small number of epidemiologic studies of biliary tract cancer and by difficulty in separating the determinants of biliary tract cancer from those of the underlying gallstones.

OTHER RISK FACTORS Although cholesterol gallstones represent the major determinant of biliary tract cancer, especially the gallbladder, it is important to

Inflammation and Infection It seems likely that chronic inflammation of the biliary tree, in association with gallstones, infectious agents, or congenital anomalies, plays an important role in the pathogenesis of biliary tract cancer. In the presence or absence of stones, dysplastic and neoplastic changes in the biliary tract have been associated with a variety of inflammatory conditions, including cholecystitis, primary sclerosing cholangitis, fibrotic reactions, and strictures (Blendis and Lurie, 2002). The

Biliary Tract Cancer mechanisms are unclear but may, for example, involve nitric oxide (NO) production in inflamed tissue, leading to oxidative stress and DNA damage (Jaiswal et al., 2000; Tazuma and Kajiyama, 2001).

Cholecystitis Patients with gallstones who develop biliary tract cancer often have chronic cholecystitis, with the risk of cancer being greatest in the late atrophic stages of cholecystitis (Assisi et al., 1998; Kanoh et al., 2001). In addition, incidental tumors are often diagnosed in resected tissue of patients with acute or chronic cholecystitis undergoing cholecystectomy (Kanoh et al., 2001). The risk of biliary tract cancer appears to be especially high among subjects with both gallstones and a history of chronic cholecystitis (Liu et al., 2002).

Porcelain Gallbladder Porcelain gallbladder, also called calcifying cholecystitis due to extensive calcification of the gallbladder wall, has been associated with gallbladder carcinoma in 12%–60% of patients (Stephen and Berger, 2001; Towfigh et al., 2001), with the risk being greater among patients with selective compared with diffuse mucosal calcification (Stephen and Berger, 2001). Porcelain gallbladder occurs in less than 1% of the population, with most cases (>90%) associated with gallstones. Like gallstones and gallbladder cancer, porcelain gallbladder is more common in women than in men. Although the mechanism predisposing to gallbladder cancer is unclear, an inflammatory process seems likely (Hoover et al., 1992).

Primary Sclerosing Cholangitis Primary sclerosing cholangitis (PSC), a progressive autoimmune disease featuring inflammation and fibrotic changes of the periductal tissues, is an established risk factor for biliary tract cancer, primarily of the bile ducts (Boberg et al., 2002; Buckles et al., 2002; Yamamoto et al., 2003; Burak et al., 2004). In follow-up studies, between 5% and 15% of patients with PSC have developed bile duct cancer, with an average incidence of 1.5% per year. In one study, 90% of individuals with PSC-associated tumors had a concomitant diagnosis of inflammatory bowel disease, usually ulcerative colitis. In a recent study of 116 patients with PSC, the risk of bile duct cancer was increased about 1500-fold, compared with incidence rates in the general population (Burak et al., 2004). The tumors are often multicentric, due to the extensive inflammatory process involving the bile ducts.

Ulcerative Colitis Although patients with ulcerative colitis are prone to PSC, a strong risk factor for bile duct cancer, an excess risk of these tumors has been associated with inflammatory bowel disease without clinical evidence of PSC or gallstones (Karlen et al., 1999; Bernstein et al., 2001; Bergquist et al., 2002), suggesting a subtle autoimmune mechanism affecting the bile ducts.

Liver Flukes In China and Thailand, bile duct cancers, particularly involving intrahepatic ducts, occur excessively in populations infested with liver flukes, including Clonorchis sinensis, Opisthorchis viverrini, and Opisthorchis felineus (Watanapa and Watanapa, 2002). Similar tumors have been seen in infected cats and dogs and in hamsters exposed to metacercaria of O. viverrini plus dimethylnitrosamine (Thamavit et al., 1978; Chaimuangraj et al., 2003). Possible carcinogenic mechanisms include chronic irritation and inflammation, nitric oxide formation, intrinsic nitrosation, and activation of metabolizing genes. In northern Thailand, a high-incidence area for cholangiocarcinoma and endemic infestation with O. viverrini, the infection is accompanied by a severe inflammatory and fibrotic process involving the intrahepatic and extrahepatic bile ducts (Sripa et al., 2003). It has been difficult to quantify the risk of cancer associated with liver flukes because of the lack of reliable serologic assays to detect past infection.

Typhoid Carrier State Chronic carriage of typhoid fever has been linked to an elevated risk of gallbladder cancer (Dutta et al., 2000; Shukla et al., 2000). Typhoid

793

fever, an infectious disease caused by Salmonella typhi, affects 17 million people worldwide each year. Among untreated cases, 2%–5% become lifetime carriers. Several clinical surveys in developing countries have suggested an excess risk of biliary tract cancer among chronic typhoid and paratyphoid carriers. In India, the prevalence of S. typhi and S. paratyphi in bile specimen cultures was higher in patients with gallbladder cancer than in those with gallstones alone (Nath et al., 1997), whereas typhoid carriage identified by Vi antigen serology was associated with a risk of biliary cancer 8–14 times higher than expected (Dutta et al., 2000; Shukla et al., 2000). In Egypt, patients with bile duct cancer had a higher prevalence of chronic Salmonella carriage than controls (39% vs. 2%) (el Zayadi et al., 1991). The mechanisms by which S. typhi or S. paratyphi infections promote biliary cancer are unknown, but the carrier state is often accompanied by gallstones, which may contribute to the excess risk.

Helicobacter Infection Recent observations have raised the possibility that bacterial infection with Helicobacter species may play an etiologic role in biliary cancer. Helicobacter pylori, one of the enteric species of Helicobacter, is classified by the International Agency for Research on Cancer (IARC) as a class I carcinogen because of its major causal role in gastric cancer. Although H. pylori prefers an acidic environment, such as the stomach, and may not survive in alkaline bile, molecular studies using polymerase chain reaction assays have detected bacterial DNA of the Helicobacter species, including H. bilis and H. pylori, in bile fluid and biliary tissue from patients with gallstones and biliary tract tumors (Fox et al., 1998; Lee, 1999; Bulajic et al., 2002; Fukuda et al., 2002; Leong and Sung, 2002; Matsukura et al., 2002; Monstein et al., 2002; Chen et al., 2003), suggesting that H. pylori may also be involved in the pathogenesis of biliary tract cancer. Recent laboratory evidence further suggests that colonization by H. pylori and H. bilis may promote the development of cholesterol stones as well as inflammatory and proliferative changes leading to chronic cholangitis (Roe et al., 1999; Swidsinski and Lee, 2001; Bulajic et al., 2002; Fukuda et al., 2002; Monstein et al., 2002; Fox et al., 1998).

Hepatitis B and C Viruses Although major causes of HCC around the world, these agents have not been generally considered as risk factors for biliary tract cancer. However, recent molecular studies from China have detected HBV DNA and HCV RNA in specimens of cancers arising from the intrahepatic and extrahepatic bile ducts (Wang et al., 1998; Yin and Chen, 1998; Chen et al., 2000). These preliminary observations warrant further investigation in developing countries where HBV is endemic and in developed countries including the United States where about 2.7 million individuals are chronically infected with HCV (Fleckenstein, 2004).

Cholecystectomy In two large population-based follow-up studies in Sweden and Denmark, patients with cholecystectomy had a higher risk of cancer of the ampulla of Vater (Ekbom et al., 1996; Chow et al., 1999), in contrast with an earlier study in Sweden that reported a lower risk of bile duct cancer following this surgery (Ekbom et al., 1993). Reasons for the excess of ampullary cancers following cholecystectomy are unclear but may be related to residual inflammatory changes, past history of gallstones, and refluxed bile or duodenal juice.

Partial Gastrectomy An excess risk of biliary tract cancer has been reported after partial gastrectomy (Tersmette et al., 1990; Tersmette et al., 1991; Caygill et al., 1991a). This finding may be related to the high frequency of gallstones reported after gastrectomy or to biliary excretion of N-nitroso compounds that are formed by bacterial action in the gastric remnant (Caygill et al., 1988; Caygill et al., 1991b).

Reproductive and Hormonal Factors A strong positive relation has been observed between the number of pregnancies and the risk of gallstones as well as gallbladder cancer

794

PART IV: CANCER BY TISSUE OF ORIGIN

(Moerman et al., 1997; Zatonski et al., 1997; Scott et al., 1999; Pandey and Shukla, 2003; Rizvi and Zuberi, 2003). In addition, an association with biliary tract cancer has been reported for early age at menarche, early age at first pregnancy, older age at menopause, and prolonged fertility (Moerman et al., 1994a; Tavani et al., 1996; Zatonski et al., 1997; Khan et al., 1999). These observations suggest the role of endogenous estrogens in promoting gallstones, increasing biliary cholesterol, and impairing biliary motility (Kritz-Silverstein et al., 1990; Everson et al., 1991), or through a direct carcinogenic effect on the biliary epithelium. Hormonal factors may also contribute to the slow intestinal transit time reported among women with gallstones, thus increasing the absorption of secondary bile salts and enhancing the hepatobiliary secretion of cholesterol (Lewis et al., 1997). Epidemiologic studies of biliary tract tumors have shown no clear association with exogenous hormones (World Health Organization, 1989; Chow et al., 1994; Moerman et al., 1994a), despite reports that oral contraceptive use and supplemental estrogens accelerate the development of gallstones (Uhler et al., 2000; Kartal et al., 2001; Simon et al., 2001). A recent study in Italy suggested an excess risk of gallbladder cancer following hormone replacement therapy, but it was based on small numbers (Gallus et al., 2002; Fernandez et al., 2003). Although a possible role of gonadotropins has been suggested by an excess risk of pancreatic and biliary cancers in the mothers of dizygotic twins (Wyshak et al., 1983), data on the relationship between serum levels of gonadotropins and biliary tract cancer are limited.

Obesity A strong and consistent association has been observed between the risk of gallbladder cancer and high relative weight (usually measured by body mass index), particularly among women (Zatonski et al., 1997; Bergstrom et al., 2001; Wolk et al., 2001; Calle et al., 2003). Obesity has also been linked to an elevated risk of bile duct cancer in both sexes but not cancer of the ampulla of Vater (Chow et al., 1994), due perhaps to the small number of ampullary tumors in these studies. The high risks of biliary tract cancer associated with obesity and overweight may reflect well-established associations with antecedent gallstones (Grundy, 2000; Pi-Sunyer, 2002), or independent effects mediated by high levels of endogenous estrogens (Kritz-Silverstein et al., 1990), serum insulin, insulin-like growth factors, or leptin in obese subjects, as well as increases in the hepatobiliary secretion of cholesterol (Giovannucci, 2003; Kaaks et al., 2003) and hypomotility of the gallbladder associated with obesity (Petroni, 2000; Chuang et al., 2001). Obesity may also contribute to biliary carcinogenesis through an inflammation pathway, since fat cells secrete a large number of inflammatory mediators, including transforming growth factor beta, tumor necrosis factor alpha, and interleukins (Hukshorn et al., 2004). In addition to overall obesity, recent data suggest that abdominal obesity and metabolic syndrome may be important risk factors for gallstones (Kodama et al., 1999; Torgerson et al., 2003; Tsai et al., 2004a). Further work is needed to determine the mechanisms by which obesity elevates the risk of biliary tract cancer and to disentangle the effects of obesity from physical inactivity and caloric intake, which also increase the risk of gallstones (Leitzmann et al., 1998; Misciagna et al., 1999; Chuang et al., 2001; Tsai et al., 2004a).

Dietary Factors The effects of diet on biliary tract cancer have been difficult to identify because the symptoms associated with biliary disease may induce dietary changes. However, epidemiologic studies have suggested a positive association with increased intake of calories, sugar, and fat, along with protective effects from dietary fiber and micronutrients in vegetables and fruits (Moerman et al., 1994b; Zatonski et al., 1997; Chatenoud et al., 1998; Tavani et al., 2000; Pandey and Shukla, 2002; Pandey, 2003; Rizvi and Zuberi, 2003). Similar findings have been reported for gallstones, but the available evidence suggests that dietary components play a relatively small role compared with obesity, physical activity, and energy balance (Maclure et al., 1990; Diehl, 1991;

Leitzmann et al., 1999; Tsai et al., 2004b; Tsai et al., 2004c; Tsai et al., 2004d). Recent reports of an elevated risk of gallbladder cancer associated with consumption of red chili pepper have suggested a potential carcinogenic ingredient such as capsaicin or a fungal contaminant (Endoh et al., 1997; Pandey and Shukla, 2002; Serra et al., 2002).

Tobacco and Alcohol The effects of tobacco smoking and alcohol consumption are unclear. A few studies have reported excess risks of gallbladder and bile duct cancers among cigarette smokers (Chow et al., 1994; Khan et al., 1999; Scott et al., 1999; Jee et al., 2004), while data on gallstones suggest a slight positive association (Logan and Skelly, 2000). A large prospective study reported that moderate intake of alcohol may lower the risk of symptomatic gallstone disease (Leitzmann et al., 1999). Further studies are needed to clarify the role of smoking and drinking in biliary diseases, including cancer.

Occupational Exposures Although the overall impact of occupational exposures seems limited, excess risks of biliary tract cancer have been reported in several occupations, including chemical workers (Bond et al., 1990), painters (Guberan et al., 1989), pesticide manufacturers (Brown, 1992), vinyl chloride workers (Wong et al., 1991), munitions workers exposed to dinitrotoluene (Stayner et al., 1993), textile workers (Chow et al., 1996; Kuzmickiene et al., 2004), those involved in cellulose triacetate fiber manufacturing (Lanes et al., 1990; Goldberg and Theriault, 1994), and workers in petroleum refining, paper mills, chemical processing, shoemaking and repairing, and asbestos-related occupations (Malker et al., 1986). Elevated risks have also been associated with heavy exposure to solvents such as methylene chloride (Lynge et al., 1997) and trichlorinated hydrocarbons (Zarchy, 1996).

Water Pollution Organochlorine Pesticides Evidence for a role of water pollution in biliary tract cancer is limited, but some clues have come from studies in developing countries. In a small study in India, where the rates of biliary tract cancer are high, the levels of organochlorines in bile, including dichlorodiphenyltrichloroethane (DDT), benzene hexachloride, aldrin, and endosulfan, were significantly higher in patients with gallbladder cancer than in those with gallstones (Shukla et al., 2001). This observation, though preliminary, is interesting since high rates of gallbladder cancer have been reported in regions with high levels of organochlorines contaminating the river water used for drinking (Dua et al., 1998). In a mortality survey in Italy, workers exposed to DDT had a significantly increased risk of liver and biliary tract cancer (Cocco et al., 1997).

Heavy Metals Another small study in India involving 96 patients with biliary disease, including 38 with gallbladder cancer and 58 with gallstones, found that the levels of cadmium, chromium, and lead in bile were significantly higher in patients with gallbladder cancer than in those with gallstones (Shukla et al., 1998). The concentration of these metals in drinking water was high in regions of India with elevated rates of gallbladder cancer.

Radiation There is little evidence that ionizing radiation causes biliary tract tumors, except for the elevated risk reported among patients injected with the radiographic contrast medium Thorotrast (van Kaick et al., 1999; Travis et al., 2001; Nyberg et al., 2002; Travis et al., 2003). A possible excess of biliary tract cancer among underground miners exposed to radon has also been reported (Tomasek et al., 1993).

Biliary Tract Cancer

Congenital Defects Anomalous Junction of Pancreaticobiliary Duct This anomaly has been consistently linked to gallbladder and bile duct cancers (Sasatomi et al., 2000; Nakayama et al., 2001), with an excess approaching 20-fold in some surveys, especially in Asian countries (Chijiiwa and Tanaka, 1996; Sasatomi et al., 2000; Nakayama et al., 2001; Hu et al., 2003). In anomalous junction of pancreaticobiliary duct (AJPBD), the junction between the common bile duct and the pancreatic duct is located outside the sphincter of Oddi, so that pancreatic fluid refluxes directly into the bile duct to cause extensive inflammatory changes. Many of the tumors are papillary adenocarcinomas with multicentric involvement of the biliary and pancreatic ducts (Kimura et al., 1985; Yamauchi et al., 1987; Iwase et al., 1997; Sugiyama et al., 2000; Tazuma and Kajiyama, 2001). Although cases are usually sporadic, familial aggregation of AJPBD associated with biliary neoplasia has been reported (Iwafuchi et al., 1990). Despite convincing clinical data, epidemiologic studies have been limited by the need for imaging techniques to detect AJPBD to evaluate cancer risks associated with this anomaly.

Choledochal Cysts An excess of biliary tract tumors, especially of the bile ducts, has been reported in patients with congenital cystic disease of the biliary tree, often referred to as Caroli’s disease (Todani and Toki, 1996; Ohtsuka et al., 2001; Jan et al., 2002). This malformation is usually accompanied by cholangitis, pancreatitis, or biliary calculi (Visser et al., 2004), while some cases are associated with AJPBD. It seems likely that biliary stasis, inflammatory processes, and pancreatic reflux contribute to the excess risk of biliary cancer associated with choledochal cysts (Benjamin, 2003).

Heritable Factors Multiple Cancer Syndromes Biliary cancer may be a component of certain genetic syndromes that predispose to multiple forms of cancer. In hereditary polyposis of the colon, including Gardner’s syndrome, up to 12% of patients develop neoplasms of the ampulla of Vater (Harned et al., 1991; Doko et al., 2003), with a smaller proportion developing bile duct cancer (Walsh et al., 1987). In addition, periampullary tumors, mainly of the carcinoid type, seem to be excessive among patients with neurofibromatosis type 1 (Klein et al., 1989). Furthermore, patients with colorectal cancer have an excess risk of developing ampullary cancer (Su et al., 1999; Das et al., 2004), which appears to be part of the spectrum of tumors in the syndrome of hereditary non-polyposis colon cancer.

Familial Occurrence A familial tendency to biliary tract cancer has been suggested by several case reports (Garber and Shipley, 1989) and by a recent nationwide survey in Sweden (Hemminki and Li, 2003), although the available data are insufficient for precise risk estimates. Several of the familial occurrences have been associated with gallstones, suggesting an opportunity to search for underlying genetic and metabolic defects in high-risk families as well as in high-risk populations, such as American Indians.

Susceptibility Genes Despite promising leads from studies of gallstone patients, little attention has been paid to common low-penetrant susceptibility genes in biliary tract cancer. A small study of gallbladder cancer from Japan reported an excess risk associated with polymorphism of the cytochrome P450 1A1 gene (CYP1A1), which encodes a protein involved in catalyzing the synthesis of cholesterol and other lipids (Tsuchiya et al., 2002). Another study of gallbladder cancer from India revealed an association with polymorphism of apolipoprotein B (apoB) XbaI (Singh et al., 2004), which affects lipid metabolism and is associated with predisposition to gallstones, suggesting a role for lipid-metabolizing genes in the development of gallstones and biliary tract cancers.

795

PREVENTIVE MEASURES Although biliary tract tumors are relatively uncommon in the United States, elevated rates are seen in certain population groups, including American Indians and Hispanics. In Mexico and other Latin American countries, gallbladder cancer accounts for 5% or more of malignant neoplasms in women and is surpassed only by cancers of the cervix, breast, and stomach (Lazcano-Ponce et al., 2001). Similarly, in parts of India, gallbladder cancer is surpassed only by breast and cervical cancers (Dhir and Mohandas, 1999). High rates are also seen in some Central and Eastern European countries, but it is unclear whether elevated risks extend to corresponding ethnic groups in the United States. Since the absolute risk for developing biliary tract cancer among patients with gallstones is low (<1%), further studies are needed to clarify the carcinogenic risk following gallstones, to identify the causal mechanisms and co-factors involved, and to develop appropriate diagnostic and preventive measures, especially in susceptible populations. Advances in genomics, proteomics, and metabolomics will play an important role in future studies. In recent years, non-surgical approaches to treat gallstones such as dissolution therapy or shock-wave lithotripsy have generally been replaced by the availability of laparoscopic cholecystectomy, except for high-risk patients who may not tolerate laparoscopy. While treatment of stones has been recommended mainly for symptomatic patients (American College of Physicians, 1993), cholecystectomy may be indicated for asymptomatic (silent) stones when the cancer risk is substantial (e.g., large stones, calcified gallbladder). There is little question that the risk of biliary tract cancer can be reduced by controlling the lifestyle risk factors for gallstones, including overweight and physical inactivity. The prospects for chemoprevention are unclear, since no micronutrients or other protective agents have been clearly linked to a decreased risk of gallstones or biliary tract cancer.

FUTURE DIRECTIONS Further leads to the prevention of biliary tract cancer are likely to come from molecular epidemiologic research into nutritional, metabolic, immunologic, and genetic determinants of predisposing conditions such as cholesterol stones and primary sclerosing cholangitis, and their relationship to biliary tract cancer. These studies should carefully distinguish between cancers of the gallbladder, bile duct, and ampulla of Vater, since epidemiologic and molecular data point to distinct etiologic pathways. A major epidemiologic challenge is to distinguish between the risk factors for biliary tract cancer and gallstones, since only 1% of the gallstone patients develop biliary tract cancer. In case-control studies of biliary tract cancer, a parallel group of patients with symptomatic gallstones who are about to undergo cholecystectomy provides a unique opportunity for comparison of risk factors and for access to biologic samples to evaluate etiologic biomarkers at the target site. This approach should help to clarify the separate and combined effects of gallstones and other risk factors in biliary tract cancer. Further insights should also come from the assessment of susceptibility genes in the pathogenesis of gallstones and biliary cancer, including those in the lipid metabolism and inflammation pathways. By applying emerging technologies in genomics, proteomics, and metabolomics to population studies, it should be possible to fully identify the risk factors and mechanisms of biliary carcinogenesis and to intervene with the most effective preventive strategies. References Achille A, Scupoli MT, Magalini AR, et al. 1996. APC gene mutations and allelic losses in sporadic ampullary tumours: evidence of genetic difference from tumours associated with familial adenomatous polyposis. Int J Cancer 68:305–312. Ahlberg J, Ewerth S, Hellers G, Holmstrom B. 1978. Decreasing frequency of cholecystectomies in the counties of Stockholm and Uppsala, Sweden. Acta Chir Scand Suppl 482:21–23.

796

PART IV: CANCER BY TISSUE OF ORIGIN

Albores-Saavedra J, Alcantra-Vazquez A, Cruz-Ortiz H, Herrera-Goepfert R. 1980. The precursor lesions of invasive gallbladder carcinoma. Hyperplasia, atypical hyperplasia and carcinoma in situ. Cancer 45: 919–927. Albores-Saavedra J, Henson DE, Klimstra DS. 2000. Tumors of the Gallbladder and Extrahepatic Bile Ducts. Series 3, Fascicle 27. Washington, DC, Armed Forces Institute of Pathology; 1–265. American College of Physicians. 1993. Guidelines for the treatment of gallstones. American College of Physicians. Ann Intern Med 119:620–622. Amigo L, Zanlungo S, Mendoza H, Miquel JF, Nervi F. 1999. Risk factors and pathogenesis of cholesterol gallstones: State of the art. Eur Rev Med Pharmacol Sci 3:241–246. Angelico F, Del Ben M, Barbato A, Conti R, Urbinati G. 1997. Ten-year incidence and natural history of gallstone disease in a rural population of women in central Italy. The Rome Group for the Epidemiology and Prevention of Cholelithiasis (GREPCO). Ital J Gastroenterol Hepatol 29:249–254. Assisi A, Papotti R, Negri P, Bordoli M, Nardone A, Nespoli A. 1998. [Acute cholecystitis and gallbladder neoplasms: The experience of 2215 cases]. G Chir 19:391–394. Attili AF, Capocaccia R, Carulli N, et al. 1997. Factors associated with gallstone disease in the MICOL experience. Multicenter Italian Study on Epidemiology of Cholelithiasis. Hepatology 26:809–818. Benjamin IS. 2003. Biliary cystic disease: The risk of cancer. J Hepatobiliary Pancreat Surg 10:335–339. Bergquist A, Ekbom A, Olsson R, et al. 2002. Hepatic and extrahepatic malignancies in primary sclerosing cholangitis. J Hepatol 36:321–327. Bergstrom A, Pisani P, Tenet V, Wolk A, Adami HO. 2001. Overweight as an avoidable cause of cancer in Europe. Int J Cancer 91:421–430. Bernstein CN, Blanchard JF, Kliewer E, Wajda A. 2001. Cancer risk in patients with inflammatory bowel disease: A population-based study. Cancer 91:854–862. Billo P, Marchegiani C, Capella C, Sessa F. 2000. [Expression of p53 in gallbladder carcinoma and in dysplastic and metaplastic lesions of the surrounding mucosa]. Pathologica 92:249–256. Blanken A. 1976. Hospital discharges and length of stay: Short-stay hospitals. United States—1972. Vital Health Stat 1 Series 10:1–66. Blendis L, Lurie Y. 2002. Primary sclerosing cholangitis: A premalignant condition. Gastroenterology 123:647–648. Boberg KM, Bergquist A, Mitchell S, et al. 2002. Cholangiocarcinoma in primary sclerosing cholangitis: Risk factors and clinical presentation. Scand J Gastroenterol 37:1205–1211. Bond GG, McLaren EA, Sable FL, Bodner KM, Lipps TE, Cook RR. 1990. Liver and biliary tract cancer among chemical workers. Am J Ind Med 18:19–24. Broden G, Ahlberg J, Bengtsson L, Hellers G. 1978. The incidence of carcinoma of the gallbladder and bile ducts in Sweden 1958 to 1972. Acta Chir Scand Suppl 482:24–25. Brown DP. 1992. Mortality of workers employed at organochlorine pesticide manufacturing plants—an update. Scand J Work Environ Health 18:155–161. Buckles DC, Lindor KD, Larusso NF, Petrovic LM, Gores GJ. 2002. In primary sclerosing cholangitis, gallbladder polyps are frequently malignant. Am J Gastroenterol 97:1138–1142. Bulajic M, Maisonneuve P, Schneider-Brachert W, et al. 2002. Helicobacter pylori and the risk of benign and malignant biliary tract disease. Cancer 95:1946–1953. Burak K, Angulo P, Pasha TM, Egan K, Petz J, Lindor KD. 2004. Incidence and risk factors for cholangiocarcinoma in primary sclerosing cholangitis. Am J Gastroenterol 99:523–526. Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ. 2003. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med 348:1625–1638. Caroli-Bosc FX, Pugliese P, Peten EP, et al. 1999. Gallbladder volume in adults and its relationship to age, sex, body mass index, body surface area and gallstones. An epidemiologic study in a nonselected population in France. Digestion 60:344–348. Caygill C, Hill M, Kirkham J, Northfield TC. 1988. Increased risk of biliary tract cancer following gastric surgery. Br J Cancer 57:434–436. Caygill CP, Knowles RL, Hall R. 1991a. Increased risk of cancer mortality after vagotomy for peptic ulcer: A preliminary analysis. Eur J Cancer Prev 1:35–37. Caygill CP, Leach SA, Fernandez F, et al. 1991b. Endogenous N-nitrosation and cancer of the biliary tract. IARC Sci Publ 137–138. Chaimuangraj S, Thamavit W, Tsuda H, Moore MA. 2003. Experimental investigation of opisthorchiasis-associated cholangiocarcinoma induction in the Syrian hamster—pointers for control of the human disease. Asian Pac J Cancer Prev 4:87–93.

Chatenoud L, Tavani A, La Vecchia C, et al. 1998. Whole grain food intake and cancer risk. Int J Cancer 77:24–28. Chen R, Zou S, Zhao X. 2000. The expression of hepatitis C virus in hilar cholangiocarcinoma and implication. Chin J Exp Surg 17:223–225. Chen W, Li D, Cannan RJ, Stubbs RS. 2003. Common presence of Helicobacter DNA in the gallbladder of patients with gallstone diseases and controls. Dig Liver Dis 35:237–243. Chijiiwa, K, Tanaka M. 1996. [Carcinoma of the gallbladder in anomalous pancreaticobiliary ductal junction]. Nippon Geka Gakkai Zasshi 97:599–605. Chow WH, Ji BT, Dosemeci M, McLaughlin JK, Gao YT, Fraumeni, JF, Jr. 1996. Biliary tract cancers among textile and other workers in Shanghai, China. Am J Ind Med 30:36–40. Chow WH, Johansen C, Gridley G, Mellemkjaer L, Olsen JH, Fraumeni JF, Jr. 1999. Gallstones, cholecystectomy and risk of cancers of the liver, biliary tract and pancreas. Br J Cancer 79:640–644. Chow WH, McLaughlin JK, Menck HR, Mack TM. 1994. Risk factors for extrahepatic bile duct cancers: Los Angeles County, California (USA). Cancer Causes Control 5:267–272. Chuang CZ, Martin LF, LeGardeur BY, Lopez A. 2001. Physical activity, biliary lipids, and gallstones in obese subjects. Am J Gastroenterol 96:1860–1865. Cocco P, Blair A, Congia P, et al. 1997. Long-term health effects of the occupational exposure to DDT. A preliminary report. Ann N Y Acad Sci 837:246–256. Cohen G, Davion T, Capron D, et al. 1992. [The estrogen-androgen profile is unchanged in men with cholelithiasis]. Gastroenterol Clin Biol 16:299–301. Csendes A, Becerra M, Rojas J, Medina E. 2000. Number and size of stones in patients with asymptomatic and symptomatic gallstones and gallbladder carcinoma: A prospective study of 592 cases. J Gastrointest Surg 4:481–485. Das A, Neugut AI, Cooper GS, Chak A. 2004. Association of ampullary and colorectal malignancies. Cancer 100:524–530. de Groen PC, Gores GJ, Larusso NF, Gunderson LL, Nagorney DM. 1999. Billiary tract cancers. N Engl J Med 341:1368–1378. Devesa SS, Grauman DG, Blot WJ, Pennello G, Hoover RN, Fraumeni JF, Jr. 1999. Atlas of cancer mortality in the United States, 1950–1994. Washington, DC, U.S. Gov. Printing Office. NIH Publ No. (NIH) 99-4564. Dhir V, Mohandas KM. 1999. Epidemiology of digestive tract cancers in India IV. Gall bladder and pancreas. Indian J Gastroenterol 18:24–28. Diehl AK. 1983. Gallstone size and the risk of gallbladder cancer. JAMA 250:2323–2326. Diehl AK. 1987. Trends in cholecystectomy rates in the United States. Lancet 2:683. Diehl AK. 1991. Epidemiology and natural history of gallstone disease. Gastroenterol Clin North Am 20:1–19. Diehl AK, Beral V. 1981. Cholecystectomy and changing mortality from gallbladder cancer. Lancet 2:187–189. Diehl AK, Schwesinger WH, Holleman DR Jr., Chapman JB, Kurtin WE. 1994. Gallstone characteristics in Mexican Americans and non-Hispanic whites. Dig Dis Sci 39:2223–2228. Diehl AK, Schwesinger WH, Holleman DR, Jr., Chapman JB, Kurtin WE. 1995. Clinical correlates of gallstone composition: Distinguishing pigment from cholesterol stones. Am J Gastroenterol 90:967–972. Doko M, Zovak M, Glavan E, Kopljar M, Tomas D. 2003.Synchronous primary carcinomas of the ampulla of Vater and ascending colon in a patient with multiple flat adenomas. Int J Gastrointest Cancer 33:117–121. Dowling GP, Kelly JK. 1986. The histogenesis of adenocarcinoma of the gallbladder. Cancer 58:1702–1708. Dowling RH. 2000. Review: Pathogenesis of gallstones. Aliment Pharmacol Ther 14 Suppl 2:39–47. Dua VK, Kumari R, Johri RK, Ojha VP, Shukla RP, Sharma VP. 1998. Organochlorine insecticide residues in water from five lakes of Nainital (U. P.), India. Bull Environ Contam Toxicol 60:209–215. Dutta U, Garg PK, Kumar R, Tandon RK. 2000. Typhoid carriers among patients with gallstones are at increased risk for carcinoma of the gallbladder. Am J Gastroenterol 95:784–787. Ekbom A, Hsieh CC, Yuen J, et al. 1993. Risk of extrahepatic bileduct cancer after cholecystectomy. Lancet 342:1262–1265. Ekbom A, Yuen J, Karlsson BM, McLaughlin JK, Adami HO. 1996. Risk of pancreatic and periampullar cancer following cholecystectomy: A population-based cohort study. Dig Dis Sci 41:387–391. el Zayadi A, Ghoneim M, Kabil SM, el Tawil A, Sherif A, Selim O. 1991. Bile duct carcinoma in Egypt: Possible etiological factors. Hepatogastroenterology 38:337–340. Elnemr A, Ohta T, Kayahara M, et al. 2001. Anomalous pancreaticobiliary ductal junction without bile duct dilatation in gallbladder cancer. Hepatogastroenterology 48:382–386.

Biliary Tract Cancer Endoh K, Nakadaira H, Yamazaki O, et al. 1997. [Risk factors for gallbladder cancer in Chilean females]. Nippon Koshu Eisei Zasshi 44:113–122. Everhart JE. 2001. Gallstones and ethnicity in the Americas. J Assoc Acad Minor Phys 12:137–143. Everson GT, McKinley C, Kern F Jr. 1991. Mechanisms of gallstone formation in women. Effects of exogenous estrogen (Premarin) and dietary cholesterol on hepatic lipid metabolism. J Clin Invest 87:237–246. Feng D, Han T, Chen S. 1998. [Polymorphism at the LDL receptor gene locus in patients with cholesterol gallstone disease]. Zhonghua Yi Xue Za Zhi 78:63–65. Fernandez E, Gallus S, Bosetti C, Franceschi S, Negri E, La Vecchia C. 2003. Hormone replacement therapy and cancer risk: A systematic analysis from a network of case-control studies. Int J Cancer 105:408–412. Fleckenstein J. 2004. Chronic hepatitis C in African Americans and other minority groups. Curr Gastroenterol Rep 6:66–70. Fox JG, Dewhirst FE, Shen Z, et al. 1998. Hepatic Helicobacter species identified in bile and gallbladder tissue from Chileans with chronic cholecystitis. Gastroenterology 114:755–763. Fraumeni JF, Devesa SS, McLaughlin JK, Stanford JL. 1996. Biliary tract cancer. In: Schottenfeld D, Fraumeni JF, eds. Cancer epidemiology and prevention. 2nd ed. New York: Oxford University Press, pp. 794–805. Fu X, Gong K, Shen T, et al. 1997. Gallstones and their chemical types in relation to serum lipids and apolipoprotein levels. Chin Med J (Engl) 110:384–387. Fukuda K, Kuroki T, Tajima Y, et al. 2002. Comparative analysis of Helicobacter DNAs and biliary pathology in patients with and without hepatobiliary cancer. Carcinogenesis 23:1927–1931. Gallus S, Negri E, Chatenoud L, Bosetti C, Franceschi S, La Vecchia C. 2002. Post-menopausal hormonal therapy and gallbladder cancer risk. Int J Cancer 99:762–763. Garber JE, Shipley W. 1989. Carcinoma of the gall bladder in three members of a family. Cancer Genet Cytogenet 39:141–142. Giovannucci E. 2003. Nutrition, insulin, insulin-like growth factors and cancer. Horm Metab Res 35:694–704. Goldberg MS, Theriault G. 1994. Retrospective cohort study of workers of a synthetic textiles plant in Quebec: I. General mortality. Am J Ind Med 25:889–907. Grundy SM. 2000. Metabolic complications of obesity. Endocrine 13:155– 165. Guberan E, Usel M, Raymond L, Tissot R, Sweetnam PM. 1989. Disability, mortality, and incidence of cancer among Geneva painters and electricians: A historical prospective study. Br J Ind Med 46:16–23. Gurleyik G, Gurleyik E, Ozturk A, Unalmiser S. 2002. Gallbladder carcinoma associated with gallstones. Acta Chir Belg 102:203–206. Han T, Jiang Z, Suo G, Zhang S. 2000. Apolipoprotein B-100 gene Xba I polymorphism and cholesterol gallstone disease. Clin Genet 57:304– 308. Harned RK, Buck JL, Olmsted WW, Moser RP, Ros PR. 1991. Extracolonic manifestations of the familial adenomatous polyposis syndromes. AJR Am J Roentgenol 156:481–485. Hart J, Modan B, Shani M. 1971. Cholelithiasis in the aetiology of gallbladder neoplasms. Lancet 1:1151–1153. Heaton KW. 1973. The epidemiology of gallstones and suggested aetiology. Clin Gastroenterol 2:67–83. Hemminki K, Li X. 2003. Familial liver and gall bladder cancer: A nationwide epidemiological study from Sweden. Gut 52:592–596. Henson DE, Albores-Saavedra J, Corle D. 1992. Carcinoma of the gallbladder. Histologic types, stage of disease, grade, and survival rates. Cancer 70:1493–1497. Ho KJ, Lin XZ, Yu SC, Chen JS, Wu CZ. 1995. Cholelithiasis in Taiwan. Gallstone characteristics, surgical incidence, bile lipid composition, and role of beta-glucuronidase. Dig Dis Sci 40:1963–1973. Hoover EL, Natesha RK, Cooperman M, Hassett J. 1992. Right upper quadrant calcification: Porcelain gallbladder disease. Am Fam Physician 45:2171–2174. Hsing AW, Gao YT, Devesa SS, Jin F, Fraumeni JF Jr. 1998. Rising incidence of biliary tract cancers in Shanghai, China. Int J Cancer 75:368–370. Hu B, Gong B, Zhou DY. 2003. Association of anomalous pancreaticobiliary ductal junction with gallbladder carcinoma in Chinese patients: An ERCP study. Gastrointest Endosc 57:541–545. Hukshorn CJ, Lindeman JH, Toet KH, et al. 2004. Leptin and the proinflammatory state associated with human obesity, J Clin Endocrinol Metab 89:1773–1778. Imai Y, Oda H, Tsurutani N, Nakatsuru Y, Inoue T, Ishikawa T. 1997. Frequent somatic mutations of the APC and p53 genes in sporadic ampullary carcinomas. Jpn J Cancer Res 88:846–854. Iwafuchi M, Ohsawa Y, Naito S, Naito M, Maruta Y, Saito H. 1990. Familial occurrence of congenital bile duct dilatation. J Pediatr Surg 25:353–355.

797

Iwase T, Nakazawa S, Yamao K, et al. 1997. Ras gene point mutations in gallbladder lesions associated with anomalous connection of pancreatobiliary ducts. Hepatogastroenterology 44:1457–1462. Jaiswal M, Larusso NF, Burgart LJ, Gores GJ. 2000. Inflammatory cytokines induce DNA damage and inhibit DNA repair in cholangiocarcinoma cells by a nitric oxide-dependent mechanism. Cancer Res 60:184–190. Jan YY, Chen HM, Chen MF. 2002. Malignancy in choledochal cysts. Hepatogastroenterology 49:100–103. Jee SH, Samet JM, Ohrr H, et al. 2004. Smoking and cancer risk in Korean men and women. Cancer Causes Control 15:341–348. Jemal A, Tiwari RC, Murray T, et al. 2004. Cancer statistics, 2004. CA Cancer J Clin 54:8–29. Jiang ZY, Han TQ, Suo GJ, et al. 2004. Polymorphisms at cholesterol 7alphahydroxylase, apolipoproteins B and E and low density lipoprotein receptor genes in patients with gallbladder stone disease. World J Gastroenterol 10:1508–1512. Johnston DE, Kaplan MM. 1993. Pathogenesis and treatment of gallstones. N Engl J Med 328:412–421. Kaaks R, Lukanova A, Rinaldi S, et al. 2003. Interrelationships between plasma testosterone, SHBG, IGF-I, insulin and leptin in prostate cancer cases and controls. Eur J Cancer Prev 12:309–315. Kakar F, Weiss NS, Strite SA. 1986. Thiazide use and the risk of cholecystectomy in women. Am J Epidemiol 124:428–433. Kanoh K, Shimura T, Tsutsumi S, et al. 2001. Significance of contracted cholecystitis lesions as high risk for gallbladder carcinogenesis. Cancer Lett 169:7–14. Kapoor VK, McMichael AJ. 2003. Gallbladder cancer: an ‘Indian’ disease. Natl Med J India 16:209–213. Karlen P, Lofberg R, Brostrom O, Leijonmarck CE, Hellers G, Persson PG. 1999. Increased risk of cancer in ulcerative colitis: A population-based cohort study. Am J Gastroenterol 94:1047–1052. Kartal A, Aksoy F, Vatansev C, et al. 2001. Does estrogen cause low conversion rates in laparoscopic cholecystectomies for acute and chronic cholecystitis in women? JSLS 5:309–312. Khan SA, Taylor-Robinson SD, Toledano MB, Beck A, Elliott P, Thomas HC. 2002. Changing international trends in mortality rates for liver, biliary and pancreatic tumours. J Hepatol 37:806–813. Khan ZR, Neugut AI, Ahsan H, Chabot JA. 1999. Risk factors for biliary tract cancers. Am J Gastroenterol 94:149–152. Kim MH, Lim BC, Myung SJ, et al. 1999. Epidemiological study on Korean gallstone disease: A nationwide cooperative study. Dig Dis Sci 44: 1674–1683. Kimura K, Ohto M, Saisho H, et al. 1985. Association of gallbladder carcinoma and anomalous pancreaticobiliary ductal union. Gastroenterology 89:1258–1265. Klein A, Clemens J, Cameron J. 1989. Periampullary neoplasms in von Recklinghausen’s disease. Surgery 106:815–819. Kodama H, Kono S, Todoroki I, et al. 1999. Gallstone disease risk in relation to body mass index and waist-to-hip ratio in Japanese men. Int J Obes Relat Metab Disord 23:211–216. Kodama I, Yoshida C, Kofuji K, Ohta J, Aoyagi K, Takeda J. 1996. Gallstones and gallbladder disorder after gastrectomy for gastric cancer. Int Surg 81:36–39. Kosters A, Jirsa M, Groen AK. 2003. Genetic background of cholesterol gallstone disease. Biochim Biophys Acta 1637:1–19. Kritz-Silverstein D, Barrett-Connor E, Wingard DL. 1990. The relationship between reproductive history and cholecystectomy in older women. J Clin Epidemiol 43:687–692. Kurashina M, Kozuka S, Nakasima N, Hirabayasi N, Ito M. 1988. Relationship of intrahepatic bile duct hyperplasia to cholangiocellular carcinoma. Cancer 61:2469–2474. Kurtin WE, Schwesinger WH, Diehl AK. 2000. Age-related changes in the chemical composition of gallstones. Int J Surg Investig 2:299–307. Kuzmickiene I, Didziapetris R, Stukonis M. 2004. Cancer incidence in the workers cohort of textile manufacturing factory in Alytus, Lithuania. J Occup Environ Med 46:147–153. Lammert F, Wang DQ, Paigen B, Carey MC. 1999. Phenotypic characterization of Lith genes that determine susceptibility to cholesterol cholelithiasis in inbred mice: Integrated activities of hepatic lipid regulatory enzymes. J Lipid Res 40:2080–2090. Lanes SF, Cohen A, Rothman KJ, Dreyer NA, Soden KJ. 1990. Mortality of cellulose fiber production workers. Scand J Work Environ Health 16:247–251. Lazcano-Ponce EC, Miquel JF, Munoz N, et al. 2001. Epidemiology and molecular pathology of gallbladder cancer. CA Cancer J Clin 51:349– 364. Lee A. 1999. Animal models of Helicobacter infection. Mol Med Today 5:500–501.

798

PART IV: CANCER BY TISSUE OF ORIGIN

Legorreta AP, Silber JH, Costantino GN, Kobylinski RW, Zatz SL. 1993. Increased cholecystectomy rate after the introduction of laparoscopic cholecystectomy. JAMA 270:1429–1432. Leitzmann MF, Giovannucci EL, Rimm EB, et al. 1998. The relation of physical activity to risk for symptomatic gallstone disease in men. Ann Intern Med 128:417–425. Leitzmann MF, Giovannucci EL, Stampfer MJ, et al. 1999. Prospective study of alcohol consumption patterns in relation to symptomatic gallstone disease in men. Alcohol Clin Exp Res 23:835–841. Leong RW, Sung JJ. 2002. Review article: Helicobacter species and hepatobiliary diseases. Aliment Pharmacol Ther 16:1037–1045. Levi F, Lucchini F, Negri E, La Vecchia C. 2003. The recent decline in gallbladder cancer mortality in Europe. Eur J Cancer Prev 12:265– 267. Lewis SJ, Heaton KW, Oakey RE, McGarrigle HH. 1997. Lower serum oestrogen concentrations associated with faster intestinal transit. Br J Cancer 76:395–400. Lin QY, Du JP, Zhang MY, et al. 1999. Effect of apolipoprotein E gene Hha I restricting fragment length polymorphism on serum lipids in cholecystolithiasis. World J Gastroenterol 5:228–230. Liu EJ, Gao YT, Deng J, et al. 2002. Medical history and risk of biliary tract cancers: A population-based study. Tumor 22:453–456. Logan RF, Skelly MM. 2000. Smoking and hepato-biliary disease. Eur J Gastroenterol Hepatol 12:863–867. Lowenfels AB, Walker AM, Althaus DP, Townsend G, Domellof L. 1989. Gallstone growth, size, and risk of gallbladder cancer: An interracial study. Int J Epidemiol 18:50–54. Lynge E, Anttila A, Hemminki K. 1997. Organic solvents and cancer. Cancer Causes Control 8:406–419. Maclure KM, Hayes KC, Colditz GA, Stampfer MJ, Willett WC. 1990. Dietary predictors of symptom-associated gallstones in middle-aged women. Am J Clin Nutr 52:916–922. Malats N, Porta M, Pinol JL, Corominas JM, Rifa J, Real FX. 1995. Ki-ras mutations as a prognostic factor in extrahepatic bile system cancer. PANK-ras I Project Investigators. J Clin Oncol 13:1679–1686. Malker HS, McLaughlin JK, Malker BK, et al. 1986. Biliary tract cancer and occupation in Sweden. Br J Ind Med 43:257–262. Manfredi S, Benhamiche AM, Isambert N, Prost P, Jouve JL, Faivre J. 2000. Trends in incidence and management of gallbladder carcinoma: A population-based study in France. Cancer 89:757–762. Maringhini A, Moreau JA, Melton LJ III, Hench VS, Zinsmeister AR, DiMagno EP. 1987. Gallstones, gallbladder cancer, and other gastrointestinal malignancies. An epidemiologic study in Rochester, Minnesota. Ann Intern Med 107:30–35. Matsukura N, Yokomuro S, Yamada S, et al. 2002. Association between Helicobacter bilis in bile and biliary tract malignancies: H. bilis in bile from Japanese and Thai patients with benign and malignant diseases in the biliary tract. Jpn J Cancer Res 93:842–847. McGlynn KA, Tsao L, Hsing AW, Devesa SS, Fraumeni JF Jr. 2001. International trends and patterns of primary liver cancer. Int J Cancer 94:290–296. Miquel JF, Covarrubias C, Villaroel L, et al. 1998. Genetic epidemiology of cholesterol cholelithiasis among Chilean Hispanics, Amerindians, and Maoris. Gastroenterology 115:937–946. Misciagna G, Centonze S, Leoci C, et al. 1999. Diet, physical activity, and gallstones—a population-based, case-control study in southern Italy. Am J Clin Nutr 69:120–126. Misciagna G, Leoci C, Guerra V, et al. 1996. Epidemiology of cholelithiasis in southern Italy. Part II: Risk factors. Eur J Gastroenterol Hepatol 8: 585–593. Mittal B, Mittal RD. 2002. Genetics of gallstone disease. J Postgrad Med 48:149–152. Moerman CJ, Berns MP, Bueno de Mesquita HB, Runia S. 1994a. Reproductive history and cancer of the biliary tract in women. Int J Cancer 57:146–153. Moerman CJ, Bueno de Mesquita HB, Runia S. 1994b. Smoking, alcohol consumption and the risk of cancer of the biliary tract; a population-based case-control study in The Netherlands. Eur J Cancer Prev 3:427–436. Moerman CJ, Bueno de Mesquita HB, Smeets FW, Runia S. 1997. Lifestyle factors including diet and cancer of the gallbladder and bile duct: A population-based case-control study in The Netherlands. Eur J Cancer Prev 6:139–142. Monstein HJ, Jonsson Y, Zdolsek J, Svanvik J. 2002. Identification of Helicobacter pylori DNA in human cholesterol gallstones. Scand J Gastroenterol 37:112–119. Morris DL, Buechley RW, Key CR, Morgan MV. 1978. Gallbladder disease and gallbladder cancer among American Indians in tricultural New Mexico. Cancer 42:2472–2477.

Nakayama K, Konno M, Kanzaki A, et al. 2001. Allelotype analysis of gallbladder carcinoma associated with anomalous junction of pancreaticobiliary duct. Cancer Lett 166:135–141. Nakeeb A, Comuzzie AG, Martin L, et al. 2002. Gallstones: Genetics versus environment. Ann Surg 235:842–849. Nath G, Singh H, Shukla VK. 1997. Chronic typhoid carriage and carcinoma of the gallbladder. Eur J Cancer Prev 6:557–559. Niemi M, Kervinen K, Rantala A, et al. 1999. The role of apolipoprotein E and glucose intolerance in gallstone disease in middle aged subjects. Gut 44:557–562. Nutting PA, Freeman WL, Risser DR, et al. 1993. Cancer incidence among American Indians and Alaska Natives, 1980 through 1987. Am J Public Health 83:1589–1598. Nyberg U, Nilsson B, Travis LB, Holm LE, Hall P. 2002. Cancer incidence among Swedish patients exposed to radioactive thorotrast: A forty-year follow-up survey. Radiat Res 157:419–425. Ohtsuka T, Inoue K, Ohuchida J, et al. 2001. Carcinoma arising in choledochocele. Endoscopy 33:614–619. Pandey M. 2003. Risk factors for gallbladder cancer: A reappraisal. Eur J Cancer Prev 12:15–24. Pandey M, Shukla VK. 2002. Diet and gallbladder cancer: A case-control study. Eur J Cancer Prev 11:365–368. Pandey M, Shukla VK. 2003. Lifestyle, parity, menstrual and reproductive factors and risk of gallbladder cancer. Eur J Cancer Prev 12:269–272. Parkin DM, Whelan SL, Ferlay J, Raymond L, Young JL Jr. 1992. Cancer Incidence in Five Continents. Volume VII. No. 143. Lyon, IARC Scientific Publication. Parkin DM, Whelan SL, Ferlay J, Teppo L, Thomas DB. 2003. Cancer Incidence in Five Continents. Volume VIII. No. 155. Lyon, IARC Sci Publ. 1–781. Petroni ML. 2000. Review article: gall-bladder motor function in obesity. Aliment Pharmacol Ther 14 Suppl 2:48–50. Pi-Sunyer FX. 2002. The medical risks of obesity. Obes Surg 12 Suppl 1:6S–11S. Plesko I, Somogyi J, Dimitrova E, Kiss J, Vlasak V. 1985. Epidemiological features of biliary tract cancer incidence in Slovakia. Neoplasma 32:125–134. Portincasa P, Moschetta A, Calamita G, Margari A, Palasciano G. 2003. Pathobiology of cholesterol gallstone disease: From equilibrium ternary phase diagram to agents preventing cholesterol crystallization and stone formation. Curr Drug Targets Immune Endocr Metabol Disord 3:67–81. Portincasa P, van de MP, van Erpecum KJ, Palasciano G, VanBergeHenegouwen GP. 1997. An update on the pathogenesis and treatment of cholesterol gallstones. Scand J Gastroenterol Suppl 223:60–69. Public Health Service. 1967. International statistical classification of disease, adapted for use in the United States. Vol. 1. 8th Revision. Washington, DC, U.S. Gov. Printing Office. PHS Publ. No. 1963. Rashid A. 2002. Cellular and molecular biology of biliary tract cancers. Surg Oncol Clin N Am 11:995–1009. Rashid A, Gao YT, Bhakta S, et al. 2001. Beta-catenin mutations in biliary tract cancers: A population-based study in China. Cancer Res 61:3406–3409. Rashid A, Ueki T, Gao YT, et al. 2002. K-ras mutation, p53 overexpression, and microsatellite instability in biliary tract cancers: A population-based study in China. Clin Cancer Res 8:3156–3163. Richardson JD, Scutchfield FD, Proudfoot WH, Benenson AS. 1973. Epidemiology of gallbladder disease in an Appalachian community. Health Serv Rep 88:241–246. Ries LA, Devesa SS. In press. Cancer incidence and mortality patterns in the United States. In: Schottenfeld D, Fraumeni JF Jr, eds. Cancer Epidemiology and Prevention, 3rd ed. New York: Oxford University Press. Ries LA, Eisner MP, Kosary CL, et al. 2003. SEER Cancer Statistics Review, 1975–2000. Rizvi TJ, Zuberi SJ. 2003. Risk factors for gall bladder cancer in Karachi. J Ayub Med Coll Abbottabad 15:16–18. Roe IH, Kim JT, Lee HS, Lee JH. 1999. Detection of Helicobacter DNA in bile from bile duct diseases. J Korean Med Sci 14:182–186. Sasatomi E, Tokunaga O, Miyazaki K. 2000. Precancerous conditions of gallbladder carcinoma: Overview of histopathologic characteristics and molecular genetic findings. J Hepatobiliary Pancreat Surg 7:556–567. Scott TE, Carroll M, Cogliano FD, Smith BF. Lamorte WW. 1999. A casecontrol assessment of risk factors for gallbladder carcinoma. Dig Dis Sci 44:1619–1625. Sen U, Sankaranarayanan R, Mandal S, Ramanakumar AV, Parkin DM, Siddiqi M. 2002. Cancer patterns in eastern India: The first report of the Kolkata cancer registry. Int J Cancer 100:86–91. Serra I, Calvo A, Maturana M, Medina E, Sharp A. 1990. Changing trends of gall bladder cancer in Chile, a high-risk area. Int J Cancer 45:376–377.

Biliary Tract Cancer Serra I, Yamamoto M, Calvo A, et al. 2002. Association of chili pepper consumption, low socioeconomic status and longstanding gallstones with gallbladder cancer in a Chilean population. Int J Cancer 102:407– 411. Shukla VK, Prakash A, Tripathi BD, Reddy DC, Singh S. 1998. Biliary heavy metal concentrations in carcinoma of the gall bladder: Case-control study. BMJ 317:1288–1289. Shukla VK, Rastogi AN, Adukia TK, Raizada RB, Reddy DC, Singh S. 2001. Organochlorine pesticides in carcinoma of the gallbladder: A case-control study. Eur J Cancer Prev 10:153–156. Shukla VK, Singh H, Pandey M, Upadhyay SK, Nath G. 2000. Carcinoma of the gallbladder—is it a sequel of typhoid? Dig Dis Sci 45:900–903. Simon JA, Hunninghake DB, Agarwal SK, et al. 2001. Effect of estrogen plus progestin on risk for biliary tract surgery in postmenopausal women with coronary artery disease. The Heart and Estrogen/progestin Replacement Study. Ann Intern Med 135:493–501. Singh MK, Choudhuri G. 2003. Re: Dutta et al.—gallbladder cancer. Am J Gastroenterol 98:936–937. Singh MK, Pandey UB, Ghoshal UC, et al. 2004. Apolipoprotein B-100 XbaI gene polymorphism in gallbladder cancer. Hum Genet 114:280–283. Singh V, Zaidi SA, Singh VS. 1997. Lipids in biliary lithogenesis. J Pak Med Assoc 47:253–255. Sripa B, Haswell-Elkins MR, Sinawat P. 2003. Histological analysis of gallbladder diseases in relation to opisthorchiasis in endemic areas of Thailand. Acta Trop 88:239–246. Stayner LT, Dannenberg AL, Bloom T, Thun M. 1993. Excess hepatobiliary cancer mortality among munitions workers exposed to dinitrotoluene. J Occup Med 35:291–296. Stephen AE, Berger DL. 2001. Carcinoma in the porcelain gallbladder: A relationship revisited. Surgery 129:699–703. Stewart L, Oesterle AL, Erdan I, Griffiss JM, Way LW. 2002. Pathogenesis of pigment gallstones in Western societies: The central role of bacteria. J Gastrointest Surg 6:891–903. Su Y, Ahsan H, Neugut AI. 1999. The association between biliary tract cancers and cancers of other sites. Am J Gastroenterol 94:2256–2262. Sugiyama Y, Kobori H, Hakamada K, Seito D, Sasaki M. 2000. Altered bile composition in the gallbladder and common bile duct of patients with anomalous pancreaticobiliary ductal junction. World J Surg 24: 17–20. Suo G, Han T, Feng D. 1999. [Association of polymorphisms of apolipoprotein B gene with cholesterol gallstone disease]. Zhonghua Yi Xue Za Zhi 79:673–675. Swidsinski A, Lee SP. 2001. The role of bacteria in gallstone pathogenesis. Front Biosci 6:E93–103. Tang WH. 1996. Serum and bile lipid levels in patients with and without gallstones. J Gastroenterol 31:823–827. Tavani A, La Vecchia C, Gallus S, et al. 2000. Red meat intake and cancer risk: A study in Italy. Int J Cancer 86:425–428. Tavani A, Negri E, La Vecchia C. 1996. Menstrual and reproductive factors and biliary tract cancers. Eur J Cancer Prev 5:241–247. Tazuma S, Kajiyama G. 2001. Carcinogenesis of malignant lesions of the gall bladder. The impact of chronic inflammation and gallstones. Langenbecks Arch Surg 386:224–229. Tersmette AC, Goodman SN, Offerhaus GJ, et al. 1991. Multivariate analysis of the risk of stomach cancer after ulcer surgery in an Amsterdam cohort of postgastrectomy patients. Am J Epidemiol 134:14–21. Tersmette AC, Offerhaus GJ, Giardiello FM, Tersmette KW, Vandenbroucke JP, Tytgat GN. 1990. Occurrence of non-gastric cancer in the digestive tract after remote partial gastrectomy: Analysis of an Amsterdam cohort. Int J Cancer 46:792–795. Thamavit W, Bhamarapravati N, Sahaphong S, Vajrasthira S, Angsubhakorn S. 1978. Effects of dimethylnitrosamine on induction of cholangiocarcinoma in Opisthorchis viverrini-infected Syrian golden hamsters. Cancer Res 38:4634–4639. Todani T, Toki A. 1996. [Cancer arising in choledochal cyst and management]. Nippon Geka Gakkai Zasshi 97:594–598. Tomasek L, Darby SC, Swerdlow AJ, Placek V, Kunz E. 1993. Radon exposure and cancers other than lung cancer among uranium miners in West Bohemia. Lancet 341:919–923. Torgerson JS, Lindroos AK, Naslund I, Peltonen M. 2003. Gallstones, gallbladder disease, and pancreatitis: Cross-sectional and 2-year data from the Swedish Obese Subjects (SOS) and SOS reference studies. Am J Gastroenterol 98:1032–1041. Towfigh S, McFadden DW, Cortina GR, et al. 2001. Porcelain gallbladder is not associated with gallbladder carcinoma. Am Surg 67:7–10. Travis LB, Hauptmann M, Gaul LK, et al. 2003. Site-specific cancer incidence and mortality after cerebral angiography with radioactive thorotrast. Radiat Res 160:691–706.

799

Travis LB, Land CE, Andersson M, et al. 2001. Mortality after cerebral angiography with or without radioactive Thorotrast: An international cohort of 3,143 two-year survivors. Radiat Res 156:136–150. Tsai CJ, Leitzmann MF, Willett WC, Giovannucci EL. 2004a. Prospective study of abdominal adiposity and gallstone disease in US men. Am J Clin Nutr 80:38–44. Tsai CJ, Leitzmann MF, Hu FB, Willett WC, Giovannucci EL. 2004b. Frequent nut consumption and decreased risk of cholecystectomy in women. Am J Clin Nutr 80:76–81. Tsai CJ, Leitzmann MF, Willett WC, Giovannucci EL. 2004c. Dietary protein and the risk of cholecystectomy in a cohort of US women: The Nurses’ Health Study. Am J Epidemiol 160:11–18. Tsai CJ, Leitzmann MF, Willett WC, Giovannucci EL. 2004d. Long-term intake of dietary fiber and decreased risk of cholecystectomy in women. Am J Gastroenterol 99:1364–1370. Tsuchiya Y, Sato T, Kiyohara C, et al. 2002. Genetic polymorphisms of cytochrome P450 1A1 and risk of gallbladder cancer. J Exp Clin Cancer Res 21:119–124. Ueki T, Hsing AW, Gao YT, et al. 2004. Alterations of p16 and prognosis in biliary tract cancers from a population-based study in China. Clin Cancer Res 10:1717–1725. Uhler ML, Marks JW, Judd HL. 2000. Estrogen replacement therapy and gallbladder disease in postmenopausal women. Menopause 7:162– 167. van Kaick G, Dalheimer A, Hornik S, et al. 1999. The German Thorotrast Study: Recent results and assessment of risks. Radiat Res 152:S64–S71. Visser BC, Suh I, Way LW, Kang SM. 2004. Congenital choledochal cysts in adults. Arch Surg 139:855–860. Vitetta L, Sali A, Little P, Mrazek L. 2000. Gallstones and gall bladder carcinoma. Aust N Z J Surg 70:667–673. Walsh N, Qizilbash A, Banerjee R, Waugh GA. 1987. Biliary neoplasia in Gardner’s syndrome. Arch Pathol Lab Med 111:76–77. Wang WL, Gu GY, Hu M. 1998. Expression and significance of HBV genes and their antigens in human primary intrahepatic cholangiocarcinoma. World J Gastroenterol 4:392–396. Watanapa P, Watanapa WB. 2002. Liver fluke-associated cholangiocarcinoma. Br J Surg 89:962–970. Wistuba II, Gazdar AF, Roa I, Albores-Saavedra J. 1996. p53 protein overexpression in gallbladder carcinoma and its precursor lesions: An immunohistochemical study. Hum Pathol 27:360–365. Wistuba II, Miquel JF, Gazdar AF, Albores-Saavedra J. 1999. Gallbladder adenomas have molecular abnormalities different from those present in gallbladder carcinomas. Hum Pathol 30:21–25. Wistuba II, Tang M, Maitra A, et al. 2001. Genome-wide allelotyping analysis reveals multiple sites of allelic loss in gallbladder carcinoma. Cancer Res 61:3795–3800. Wolk A, Gridley G, Svensson M, et al. 2001. A prospective study of obesity and cancer risk (Sweden). Cancer Causes Control 12:13–21. Wong O, Whorton MD, Foliart DE, Ragland D. 1991. An industry-wide epidemiologic study of vinyl chloride workers, 1942–1982. Am J Ind Med 20:317–334. Wood R, Fraser LA, Brewster DH, Garden OJ. 2003. Epidemiology of gallbladder cancer and trends in cholecystectomy rates in Scotland, 1968–1998. Eur J Cancer 39:2080–2086. World Health Organization. 1957. Manual of the international statistical classification of diseases, injuries, and causes of death. Vol. 1. 7th Revision. Geneva, World Health Organization. World Health Organization. 1989. Combined oral contraceptives and gallbladder cancer. The WHO Collaborative Study of Neoplasia and Steroid Contraceptives. Int J Epidemiol 18:309–314. World Health Organization. 1992. International Statistical Classification of Diseases and Related Health Problems. 1989 Revision. Geneva, World Health Organization. Wyshak G, Honeyman MS, Flannery JT, Beck AS. 1983. Cancer in mothers of dizygotic twins. J Natl Cancer Inst 70:593–599. Yamaguchi K, Enjoji M. 1987. Carcinoma of the ampulla of vater. A clinicopathologic study and pathologic staging of 109 cases of carcinoma and 5 cases of adenoma. Cancer 59:506–515. Yamanoto T, Uki K, Takeuchi K, et al. 2003. Early gallbladder carcinoma associated with primary sclerosing cholangitis and ulcerative colitis. J Gastroenterol 38:704–706. Yamauchi S, Koga A, Matsumoto S, Tanaka M, Nakayama F. 1987. Anomalous junction of pancreaticobiliary duct without congenital choledochal cyst: A possible risk factor for gallbladder cancer. Am J Gastroenterol 82: 20–24. Yanagisawa N, Mikami T, Saegusa M, Okayasu I. 2001. More frequent betacatenin exon 3 mutations in gallbladder adenomas than in carcinomas indicate different lineages. Cancer Res 61:19–22.

800

PART IV: CANCER BY TISSUE OF ORIGIN

Yen S, Hsieh CC, MacMahon B. 1987. Extrahepatic bile duct cancer and smoking, beverage consumption, past medical history, and oralcontraceptive use. Cancer 59:2112–2116. Yin F, Chen B. 1998. Detection of hepatitis C virus RNA sequences in hepatic portal cholangiocarcinoma tissue by reverse transcription polymerase chain reaction. Chin Med J (Engl) 111:1068–1070. Zarchy TM. 1996. Chlorinated hydrocarbon solvents and biliary-pancreatic cancer: Report of three cases. Am J Ind Med 30:341–342.

Zatonski WA, Lowenfels AB, Boyle P, et al. 1997. Epidemiologic aspects of gallbladder cancer: A case-control study of the SEARCH Program of the International Agency for Research on Cancer. J Natl Cancer Inst 89:1132–1138. Zhuang X, Li L. 2000. [A case-control study of gallstone disease in female population in Taicang]. Zhonghua Liu Xing Bing Xue Za Zhi 21:44–47. Zou S, Zhang L. 2000. Relative risk factors analysis of 3,922 cases of gallbladder cancer. Zhonghua Wai Ke Za Zhi 38:805–808.

41

Cancers of the Small Intestine JENNIFER L. BEEBE-DIMMER AND DAVID SCHOTTENFELD

M

alignant neoplasms of the small intestine are relatively rare in the United States with an estimated 5420 new cases diagnosed in 2005, which is less than 0.5% of the total number of new cancers diagnosed in this country. The projected number of new cases among men, 2840, is similar to that in women, 2580. Worldwide, the agestandardized incidence rates for malignant neoplasms of the small intestine varied in males from 2.6 per 100,000 among blacks in the United States (Connecticut) to 0.1–0.3 per 100,000 in India, Kuwait, and Vietnam (Parkin et al., 2002). Approximately 1000 deaths each year in the United States are attributed to cancers of the small intestine; the age-adjusted mortality per 100,000 (2000 US standard) during 1996–2000 was 0.5 in males and 0.3 in females (Ries et al., 2003). The small intestine, consisting of the duodenum, jejunum, and ileum, represents 75% of the length and 90% of the absorptive surface area of the entire gastrointestinal tract, but remarkably, less than 3% of all gastrointestinal cancers arise in the small intestine. The average annual age-adjusted incidence, 1.7 per 100,000, is about 1/50th the incidence of colorectal cancer in the United States. Several hypotheses have been advanced to account for the dramatic differences in cancer incidence between the large and small intestine (Chow et al., 1993). The apparent protection afforded to the small bowel, in contrast with the large bowel, may be attributed to diminished concentrations of anaerobic bacteria resulting in decreased metabolic conversion of bile acids and other procarcinogenic substrates to mutagenic and mitogenic agents; rapid transit of potential carcinogenic agents limiting exposure of the bowel mucosa; and higher concentrations of mucosal enzymes that detoxify potentially harmful endogenous biochemical and exogenous ingested agents.

CLASSIFICATION Anatomic Distribution The adult small intestine is an elongated tube about 20 feet (610 cm) long, its length depending in part on the tone of the muscular wall. The duodenum constitutes the proximal 25–30 cm, adhering to the posterior abdominal wall and encircling the head of the pancreas in a horseshoe fashion. Bile and pancreatic secretions enter the duodenum in the region of the ampulla of Vater. Unlike the fixed retroperitoneal duodenum, the jejunum and ileum are suspended by an extensive mesentery and have considerable mobility within the abdominal cavity. The jejunum, constitutes the next segment and is approximately 2.5 meters (about 8 feet) in length. No specific anatomic structure delineates the end of the jejunum and the beginning of the ileum, which constitutes about the distal 3.5 meters (about 11 feet) of the small intestine. The diameter of the proximal jejunum is almost twice that of the distal ileum, and its wall considerably thicker due to the circumferential mucosal and submucosal folds, the plicae circulares. The distal half of the ileum contains prominent elliptical aggregates of lymphoid follicles (Peyer patches).

Histopathology The innermost tissue layer of the small intestine is the mucosa consisting of absorptive cells and intestinal glands that line the crypts and villi. The crypt epithelium functions in cell proliferation and renewal and is comprised of Paneth, goblet, undifferentiated, and endocrine

cells; the villi normally do not participate in cell renewal and proliferation but function as absorptive cells (Madara and Trier, 1987). The intestinal glandular cells give rise to adenocarcinomas, and the enteroendocrine cells, also called enterochromaffin or argentaffin cells, are the cells of origin for benign and malignant carcinoid tumors (Ashley and Wells, 1988). Beneath the mucosa is the submucosa, consisting of dense connective tissue, sparsely infiltrated by cells, including fibroblasts, lymphocytes, macrophages, mast cells, and plasma cells. The submucosa contains lymphatic and venous plexuses, and an extensive network of arterioles, ganglion cells, and nerve fibers. The submucosa of the proximal half of the duodenum contains small islands of Brunner glands, which contain both mucous and serous secretory cells. Beneath and within the submucosa, lymphoid nodules throughout the small intestine can give rise to malignant lymphomas. The muscularis, located beneath the serosal outer layer, is the tissue of origin for leiomyosarcomas.

Adenocarcinoma About 30%–40% of all small intestinal cancers are adenocarcinomas, with 75%–80% of the glandular neoplasms distributed in the duodenum and adjacent jejunum. The distal jejunum and proximal ileum, for unexplained reasons, are generally spared. The average annual incidence rate per 100,000 (adjusted to the 2000 US population) is 0.61 in males and 0.43 in females (Ries et al., 2002).

Carcinoid The carcinoid tumor, or the argentaffinoma, accounts for approximately 35% of all malignant neoplasms in the small intestine; 90% of the carcinoid tumors of the small intestine originate in the ileum, and appear to be multicentric in at least one-third of reported cases (Gabos et al., 1993). The average annual incidence rate per 100,000 (adjusted to the 2000 US population) is 0.81 in males and 0.56 in females. Black males have the highest rates with an annual incidence of 1.47 per 100,000 (Ries et al., 2002). However, carcinoid tumors are often indolent and not discovered until autopsy, which suggests that reported measures of incidence are likely underestimates of true disease frequency. The production of serotonin by the neoplastic cells, and the urinary excretion of its metabolite, 5-hydroxyindoleacetic acid, represent an important biologic marker for recurrent tumor growth and metastatic disease. The carcinoid tumor is classified as a neuroendocrine tumor with the capability of producing biogenic amines and polypeptide hormones.

Lymphoma Lymphomas of the small intestine, which are less common than gastric lymphomas, must be evaluated carefully to determine if the tumor has originated in the small intestine or whether the small intestine is involved secondarily in conjunction with disseminated disease. Approximately 15%–20% of small bowel cancers are classified as lymphomas. Tumors are most frequently found in the terminal ileum, followed by jejunum (Weiss and Yang, 1987; Chow et al., 1996; Howe et al., 2001). The average annual incidence per 100,000 is higher in

801

802

PART IV: CANCER BY TISSUE OF ORIGIN

males (0.47) than females (0.21), and in whites (0.34) compared with blacks (0.18) (Ries et al., 2002). The small intestine is the most common site of origin for extranodal immunoproliferative lymphoma, which has been described in the Arabs and Jews in the Middle East (Mediterranean lymphoma) and North Africa, and in South African blacks. Mediterranean lymphoma, an immunoblastic B-cell lymphoma, is commonly associated with clonal expression of Ig A heavy chain immunoglobulin, macroscopic presentation in the jejunum or duodenum, and the occurrence in patients who are 10–29 years of age (Al-Saleem and Al-Bahrani, 1973; Isaacson et al., 1979; Fine and Stone, 1999). The pathologic classification of small intestinal non-Hodgkin lymphomas (NHL) consists of diffuse histiocytic (B-cell diffuse large cell lymphoma more commonly than immunoblastic), lymphocytic, and mixed types. Primary Hodgkin lymphoma of the small intestine accounts for less than 3% of all small bowel lymphomas (Freeman et al., 1977).

Sarcoma Sarcomas of the small intestine are extremely rare, accounting for just 10%–15% of small bowel tumors (Chow et al., 1996; Howe et al., 2001). Current diagnostic techniques have further classified tumors as gastrointestinal stromal tumors (GIST), gastrointestinal autonomic nerve tumors (GANT), or a mixture of both depending upon whether tumors arise primarily from smooth muscle or neural tissue (Erlandson et al., 1996). Leiomyosarcoma is the most common histologic subtype, diagnosed in over 80% of cases, followed by Kaposi and spindle cell sarcomas (Howe et al., 2001). The average annual incidence of leiomyosarcoma is 0.11 per 100,000 (Ries et al., 2002).

Premalignant Lesions Adenoma Adenomas of the small intestine are the most common asymptomatic premalignant lesions (Sellner, 1990; Gill et al., 2001). Three types of adenomas are found in the small intestine: Brunner gland in the submucosa of the duodenum, islet-cell, and polypoid. Most adenomatous polyps occur in the periampullary region of the duodenum, where bile and pancreatic secretions enter the small intestine. As in the colon, the risk of adenocarcinoma increases in relation to increasing polyp size, the predominance of villous features, and the extent of epithelial dysplasia.

of b-catenin may result in uncontrolled oncogenic signaling. The most common mutation observed was on exon 3 of the b-catenin gene, substituting the amino acid alanine by serine at codon 37 (36% of cases).

DEMOGRAPHIC PATTERNS The estimated annual incidence (1996–2000) of cancer of the small intestine in the United States was 1.7 per 100,000 (adjusted to 2000 US population). The age-adjusted incidence per 100,000 during 1996–2000 was 1.9 in males and 1.4 in females. Rates in this country have increased steadily since 1973. From 1973 to 1994, it was estimated that age-adjusted incidence increased nearly 50% or an annual increase of approximately 2.5% per year. This trend reflected the increase in incidence registered for adenocarcinomas, carcinoids, and lymphomas in males and carcinoids and lymphomas in females. The incidence of sarcomas for both sexes has remained relatively stable since 1973 (Chow et al., 1996). The age-specific incidence rates, after age 30, increased proportionately with increasing age for adenocarcinomas, carcinoids, and lymphomas, while sarcomas appeared to level off after age 70 (Figure 41–1). The age-adjusted mortality 0.6 per 100,000 and incidence 3.0 per 100,000 in US blacks reflected higher risks compared with US whites (Table 41–1). An examination by histologic subtype reveals that age-adjusted incidence of adenocarcinoma and carcinoids were higher in blacks compared with whites, but the reverse is true for sarcomas and lymphomas (Ries et al., 2002). The patterns of age-standardized international incidence rates (1993–1997) suggested that the risks were generally higher in men than in women for each racial and ethnic group comparison, and somewhat higher in North American and western European countries (Figure 41–2). Lowenfels (1973) indicated that international incidence rates for cancer of the small intestine were positively correlated with colon cancer incidence rates; no correlation existed between stomach cancer and small intestinal cancer incidence rates. In Israel, the lowest rates for men were described for the Jews born in Israel, while for women rates were similar between Jews born in Israel and Jewish migrants. Among the Japanese residing in the United States, the agestandardized incidence rates in males and females (Hawaii and Los Angeles) were closer to those registered in Japan (Miyagi) than in US whites and blacks (Table 41–2).

HOST FACTORS Leiomyoma Leiomyomas are the most common symptomatic benign lesion and originate in the smooth muscle wall of the muscularis mucosae or muscularis propria. Leiomyomas occur most frequently in the jejunum and are typically well-differentiated tumors. Atypical cells in mitosis and distorted architecture are hallmark features indicative of leiomyosarcoma (Gill et al., 2001).

Molecular Genetic Characteristics There is evidence to suggest that the genes involved in cell cycle regulation that have been implicated in colorectal carcinogenesis play similar roles in small intestinal adenocarcinoma (Arber et al., 1996; Sutter et al., 1997; Rashid and Hamilton, 1997). Arber (1999) observed increased expression of the protein products of cell cycle controlling or “gate-keeper” genes such as cyclin D1, cyclin E, p53, p16, p21, and p27 in a sample of patients with small bowel adenomas or adenocarcinomas. The most common genetic alteration was an overexpression of p16 protein in over 90% of adenomatous polyps and adenocarcinomas. A significant difference was reported in the overexpression of the p53 protein product between adenomas (47%) and adenocarcinomas (65%) indicating a role for p53 in tumor progression. Little is known about the genetic epidemiology of gastrointestinal carcinoid tumors. Fujimori (2001) examined the potential role of the APC gene and expression of b-catenin, which functions in intercellular adhesion and signaling in carcinoid tumors. Overexpression

Inflammatory bowel disease (IBD) including Crohn Disease (CD), familial adenomatous polyposis (FAP), Peutz-Jeghers syndrome (PSJ), and acromegaly have been reported to be associated with adenocarcinoma of the small bowel, whereas patients with celiac sprue and acquired immune deficiency syndrome (AIDS) are at an increased risk for small intestinal lymphomas. Gastrointestinal polyposis refers to the presence of multiple polypoid lesions throughout the gastrointestinal tract. Familial adenomatous polyposis (FAP) and Gardner syndrome are attributed to a germ line mutation or deletion in the 5q21-q22 chromosomal region (Groden et al., 1991). Colorectal cancer is generally considered an inevitable consequence in the natural history of FAP, appearing 10–15 years after the clinical diagnosis of polyposis. Polyps are present in the upper gastrointestinal tract in nearly all FAP patients. The duodenum, and in particular the periampullary region, may contain multiple adenomas; in such patients, the prevalence of adenocarcinoma may be as high as 10%–15% (MacDonald et al., 1967; Melmed and Bouchier, 1972; Qizilbash, 1976; Bjork et al., 2001; Wieman et al., 2003). Peutz-Jeghers syndrome (PJS) is an example of a familial hamartomatous polyposis syndrome in which there is an increased risk of adenocarcinoma of the small intestine (Dozois et al., 1969). PJS consists of mucocutaneous pigmentation and gastrointestinal polyposis. The syndrome appears to be inherited as a single pleiotropic autosomal dominant gene with variable penetrance. The PJS gene has recently been identified and confirmed in the original family as a muta-

803

Cancers of the Small Intestine

Table 41–1. Age-Adjusted* Incidence and Mortality Rates Per 100,000 Population for Cancer of Small Intestine, Males and Females, Whites and Blacks, United States, 1996–2000

4 3.5 3

Whites

2.5

Small Intestine

2

Mortality Incidence

1.5

Blacks

Males

Females

Males

Females

0.5 1.9

0.3 1.4

0.7 3.6

0.5 2.6

Source: Surveillance, Epidemiology End Results (SEER). Ries et al., 2003. *Age-adjusted to 2000 US standard population.

1 0.5 0

to

to

to

to

+

85

75

65

55

45

to

to

84

74

64

54

44

34

24

14

to

to

35

25

15

5

<5

Age Adenocarcinoma Carcinoid Sarcoma Lymphoma

Figure 41–1. Annual age-specific incidence rates per 100,000 for cancer of the small intestine by histologic subtype, United States, 1990–1999. (Source: Ries et al., 2002.)

tion in the LKB1 or STK11 gene on the short arm of chromosome 19, which causes a frameshift mutation resulting in the inactivation of its product, serine threonine kinase (Jenne et al., 1998; Hemminki et al., 1998; Westerman et al., 1999). The histological framework for the

Males

Peutz-Jeghers polyp consists of a central core of broad bands of smooth muscle that are contiguous with the muscularis mucosae. The melanin deposits in the syndrome are distributed around the mouth, nose, lips, buccal mucosa, hands, feet, and the anogenital region. While the Peutz-Jeghers polyps are not true adenomas, carcinomas in the small intestine may arise from foci of adenomatous epithelium. The cumulative lifetime incidence of small intestinal adenocarcinomas in Peutz-Jeghers patients has been estimated as 2.4%. It is thought that the PJS gene acts as a tumor suppressor with the ablation of its function involved in the neoplastic conversion of hamartomas to adenocarcinomas (Gruber et al., 1998). In addition, the PJS gene may have pleiotropic effects in that it has been reported to be associated with increased risks of breast cancer and ovarian germ cell tumors in young women; Sertoli cell testicular tumors and gynecomastia in young boys; and polyps and carcinomas in the biliary tree, gallbladder, pancreas, esophagus, colon, stomach, and lung (Giardello et al., 1987; Konishi et al., 1989; Spigelman et al., 1990; Boardman et al., 1998; Giardello et al., 2000). The risk of a metachronous primary adenocarcinoma in the large intestine was found to be significantly increased in patients with an

Females NORTH AMERICA US SEER White

1.2

0.9

US SEER Black

2.4

1.8

CANADA

1.0

0.6

ICELAND

1.5

1.8

EUROPE DENMARK

0.6 0.5

0.5 0.4 0.4 0.5 0.5 0.5 0.4 0.4

ESTONIA FRANCE:Bas Rhin

1.0

FRANCE:Calvados

1.3

GERMANY:Saarland

0.6

ITALY:Florence

0.9

ITALY:Ragusa

0.4

LATVIA

0.6

NORWAY

1.1

0.7

POLAND:Cracow

0.2

0.5

POLAND:Warsaw

0.4

0.1

SWEDEN

1.4

1.0

SWITZERLAND

1.6 0.6 0.7

1.4 0.4 0.5

UK:England UK:Scotland ASIA CHINA:Shanghai

0.6 0.5

0.5 0.4

CHINA:Hong Kong

0.3

0.2

INDIA:Bombay

0.6 0.7

0.4

JAPAN:Miyagi

0.6

SINGAPORE:Chinese

0.3

0.3

SINGAPORE:Malay AFRICA

0.4

ALGERIA:Algiers

0.1

0.3 0.0

MALI:Bamako OCEANIA

1.0

3

2.5

2

1.5

0.5 0.6

NEW ZEALAND

0.8 1

AUSTRALIA

0.5

0

0

0.5

1

1.5

2

2.5

3

Figure 41–2. Age-adjusted (World Population) incidence rates per 100,000 population for cancer of the small intestine, males and females, selected countries, 1993–1997. (Source: Parkin et al., 2002.)

804

PART IV: CANCER BY TISSUE OF ORIGIN

Table 41–2. Age-adjusted (world population) incidence rates per 100,000 population for cancer of the small intestine, males and females by ethnicity in selected countries, 1993–1997 Population United States Los Angeles White Hispanic Black Japanese Chinese Filipino Korean Hawaii White Japanese Hawaiian Filipino Chinese Israel All Jews Born in Israel European/American Africa/Asia Non-Jews

Males

Females

1.0 1.1 1.8 0.6 1.0 1.5 1.0

0.8 0.6 0.9 0.5 0.2 0.6 0.3

2.1 1.0 1.5 0.9 0.0

1.1 0.2 1.0 0.8 0.2

0.8 0.6 0.9 0.8 0.9

0.4 0.5 0.4 0.4 0.5

Source: Cancer Incidence in Five Continents (volume VIII), 2003.

index primary adenocarcinoma in the small intestine. The Standardized Incidence Ratio (SIR) was 5.0 (95% C.I.: 2.3–9.4) in men, and 3.7 (95% C.I.: 1.3–8.0) in women. The relative risks were mutually increased; in men with colorectal cancer, the SIR for adenocarcinoma in the small intestine was 7.1 (95% C.I.: 4.7–10.3), and in women with colorectal cancer, the SIR for adenocarcinoma in the small intestine was 9.0 (95% C.I.: 6.0–12.9). There were no apparent associations between adenocarcinoma in the small intestine and second primary cancers in the stomach, female breast, ovary, or endometrium (Neuget and Santos, 1993). Adenocarcinoma of the small intestine has been reported in members of families with Lynch syndrome II, or hereditary nonpolyposis colorectal cancer (Lynch et al., 1989). Acromegaly is a rare disorder characterized by a prolonged, excessive secretion of growth hormone (GH) from the pituitary gland. It is estimated that in 90% of acromegaly cases, overproduction of growth hormone is caused by benign tumors or adenomas of the pituitary. While small adenomas of the pituitary gland are quite common, most tumors do not cause symptoms. It has been estimated that the annual incidence of acromegaly in the United States is 3 per million. Studies have shown increased risks of colorectal cancer (SIR = 2.6; 95% C.I.: 1.6–3.8) and small intestinal cancer (SIR = 6.0; 95% C.I.: 1.2–17.4) in patients with acromegaly. A plausible mechanism may be the overproduction of insulin-like growth factor (IGF-1) as a result of increased GH or somatotropin concentrations (Pines et al., 1985; Brunner et al., 1990; Ron et al., 1991; Baris et al., 2002). Increased concentration of bioavailable IGF-1 has been linked to increased risks of colorectal, lung, breast, and prostate cancer, for example, and is associated with augmented mitogenic and anti-apoptotic properties (Chan et al., 1998; Manousos et al., 1999; Ma et al., 1999; Giovannucci et al., 2000). Inflammatory bowel disease (IBD) includes two related but clinically and histologically separate entities: ulcerative colitis (UC) and Crohn disease (CD), also classified as regional enteritis or enterocolitis. CD is a chronic, recurrent transmural inflammation of the alimentary tract that usually affects the ileum, colon, and/or perianal region. The transmural inflammatory process commonly begins with focal microscopic mucosal crypt abscesses. The focal lesions are accompanied by an inflammatory response and the aggregation of macrophages, lymphocytes, and plasma cells. Progression of the disease is accompanied by segmental ulcerations, fissures, and fistulae, and subsequently by fibrotic thickening of the intestinal wall and its mesentery with regional narrowing of the lumen.

Although CD may affect any segment of the gastrointestinal tract, the most common anatomic distributions are small intestine alone (30%–40%), both small and large intestine (40%–55%), and large intestine alone (15%–25%). For those patients with CD involving the small intestine, the terminal ileum will be affected in at least 90%. The determination of the true incidence of CD may be problematic because of the delay between onset of symptoms and the date of clinical diagnosis, the absence of population-based registration data, and the potential for misclassification or in distinguishing CD from other inflammatory or noninfectious diseases of the small bowel. Both CD and UC incidence rates are highest in the United States, Canada, northern and western European countries, American- or European-born Jews rather than Israeli-born Jews, and populations of European origin living in Australia, New Zealand, and South Africa. The estimated agestandardized annual incidence of CD ranges from 1–15 per 100,000; the peak age-specific interval at initial diagnosis in males and females is 15–35 years. Many family studies have shown a significantly elevated risk of CD among the relatives of patients with IBD. Studies of familial aggregation, in monozygotic twin pairs, and of genetic immunologic markers, underscore genetic influences in CD. Siblings of patients with CD are 20–30 times more likely to develop CD than siblings of the general population. The first gene to confer susceptibility to CD was identified on chromosome 16 (Hugot et al., 2001; Ogura et al., 2001). The gene, NOD2, has been shown to have a role in mediating apoptosis (Beutler, 2001). Individuals homozygous for the NOD2 mutation are more than 20 times more likely to develop CD (Cuthbert et al., 2002; Lesage et al., 2002). Yet, less than 20% of patients with CD are homozygous for NOD2 variants. The NOD2 protein in CD is expressed in macrophages, regulating the activation of nuclear factor kB (NF-kB), causing disruption of normal cytokine pathways. The result is impairment of apoptosis and accumulation of activated type 1 helper T-lymphocyte cells and production of potent cytokines including TNF-a, IL-6, IL-12, which enhance inflammation and tissue damage (Shanahan, 2002; Podolsky, 2002). Many of the extraintestinal manifestations of IBD, (i.e., sclerosing cholangitis, ankylosing spondylitis—especially in conjunction with the HLA-B27 haplotype, and erythema nodosum) suggest that immune mechanisms are important in disease pathogenesis. The early events in the mucosa and submucosa of CD appear to represent an impaired ability to regulate immune mechanisms that accompany the inflammatory response to the luminal stream of microbial, dietary, and other exogenous antigens (Sands, 2002). Adenocarcinomas in the small intestine and colon occur with increased frequency in patients with CD (Fielding et al., 1972; Greenstein et al., 1980; Hawker et al., 1982; Richards et al., 1989). In comparing de novo adenocarcinomas in the small intestine, carcinomas associated with Crohn disease occur at a younger mean age, or after an average interval of follow-up of 15–20 years, occur predominantly in the terminal ileum or generally in areas where CD has been active for many years, tend to be multifocal, and in approximately 30%–40% of the total of reported cases the carcinomas developed in surgically excluded or bypassed segments of small intestine. Although the incidence of adenocarcinoma of the small intestine is rare, the relative risk for patients with CD is increased more than 10-fold (Lightdale et al., 1975; Nesbitt et al., 1976; Korelitz, 1983; Mellemkjaer et al., 2000; Bernstein et al., 2001). An increased incidence of non-Hodgkin lymphoma (NHL) has been observed among patients with acquired immunodeficiency syndrome (AIDS), which may be manifested as a systemic disease or as a primary extranodal lymphoma, as, for example, gastrointestinal lymphoma. NHL in AIDS patients is usually classified as poorly differentiated, diffuse B-cell lymphoma. The molecular basis for AIDS-associated lymphoma, in a number of instances, is distinguished by rearrangement of the [cf2]c-myc[cf1] proto-oncogene and DNA integration of the lymphotropic Epstein-Barr virus (Carbone, 2002). Based on a highly specific immunofluorescent serologic test of antibodies against endomysium, the global prevalence of celiac sprue is about 0.5%. The autoantigen for endomysial antibodies is the enzyme, tissue transglutaminase. The endomysium is the sheath of reticular

Cancers of the Small Intestine fibrils surrounding each muscle fiber (Green and Jabri, 2003). Celiac sprue or gluten-sensitive enteropathy is characterized by blunting and clubbing of the villous mucosal folds of the small intestine, lymphoid and plasma cell infiltration of the lamina propria, and hyperplasia of the crypt epithelium. These architectural distortions decrease the mucosal surface area and functional activity of mucosal enzymes that are normally available for digestion and absorption. Celiac sprue usually affects the proximal small intestine to a greater degree than the ileum. The interaction of the water-insoluble, alcohol-soluble, gluten fractions of proteins of cereal grains, namely, of wheat, barley, and rye, with the mucosa of the small intestine is fundamental in the pathogenesis of celiac sprue. Possible mechanisms for the clinical and pathologic sequelae of the gluten interaction with the intestinal mucosa are enzyme deficiency with incomplete digestion of toxic products (prolamins) of grains such as the gliadins in wheat that are damaging to a susceptible mucosa; toxicity mediated by a lectin-like interaction between toxic products of gluten and intestinal cells due to abnormal glycosylation of epithelial cell apical membrane glycoproteins; and genetically determined, antibody and cell-mediated aberrant inflammatory and immune responses resulting in the overproduction of cytokines and activated T-lymphocytes in response to gluten. In most populations studies, up to 97% of patients with celiac sprue disease express the HLA-DQ2 or HLA-DQ8 genes encoding for a class II major histocompatibility complex molecule (Mowat, 2003). Celiac disease appears to exhibit a heritable component as there is evidence of familial aggregation. Siblings who share the HLA-DQw2 haplotype possess 40% concordance for celiac disease, and concordance between monozygotic twins is 70%. The latter observation has been interpreted to signify that there are likely other undetected genes conferring risk in conjunction with environmental, lifestyle, and immunologic factors (Murray, 1999). It has been hypothesized that a specific “trigger” may be an infectious agent, with homology to the antigenic determinants in gluten (Trier, 1993; Green and Jabri, 2003). Patients with celiac sprue are at increased risk of T-cell lymphomas (or “enteropathy-associated T-cell lymphoma” (EATL)), and possibly adenocarcinomas, in the small intestine (Petreshock et al., 1975; Holmes et al., 1976; Brandt et al., 1978; Swanson et al., 1983; Trier, 1993; Catassi et al., 2002). Green (2003) evaluated a hospital-based cohort of 381 patients with celiac sprue and observed a significant increase in occurrence of non-Hodgkin lymphomas. A precise estimate of risk could not be inferred, because cancers occurring before or simultaneously with the diagnosis of celiac disease were included with those occurring after the diagnosis. However, of the nine cancer cases occurring after diagnosis of celiac disease (mean follow-up interval = 6 years ± 11 years), three were identified as primary non-Hodgkin lymphomas of the small intestine.

ENVIRONMENTAL FACTORS Because cancer of the small intestine is a rare occurrence, little is known about potential environmental determinants. Smoking tobacco has consistently been shown to be associated with small bowel carcinoid tumors, but not adenocarcinomas (Chow et al., 1993; Chen et al., 1994; Negri et al., 1999; Kaerlev et al., 2000; Kaerlev et al., 2002). In a population-based case-control study conducted in Europe, Kaerlev (2002) reported smoking tobacco was associated with small bowel carcinoid tumors (OR = 1.9; 95% C. I.: 1.1–3.2), and adenocarcinomas (OR = 1.5; 95% C. I.: 0.8–2.8). In a small US case-control study, odds ratios of greater than 4.0 were reported for both carcinoids and adenocarcinomas of the small bowel (Chen et al., 1994). Consumption of alcoholic beverages has also been linked with small intestinal cancer. Specifically, intake of beer or spirits (≥24 g/day), but not wine consumption, was associated with an approximate 3.5-fold increase in risk for small bowel adenocarcinoma (Kaerlev et al., 2000). It has been hypothesized that alcohol may have depressive effects on the intestinal immune response; however, this does not explain the differential effects of beer or liquor on bowel mucosal immunity as compared to wine.

805

Dietary factors have also been suspected to play a role in the etiology of small intestinal adenocarcinoma given the fact that the international patterns of disease frequency between cancers of the large and small bowel parallel one another and are positively correlated with per capita consumption of dietary fat and red meat. Negri (1999) reported increased risk of adenocarcinoma of the small intestine among highest consumers of red meat (OR = 4.57; 95% C. I.: 1.01–20.81), refined carbohydrates (OR = 3.76; 95% C. I.: 1.13–12.50), and sugar (OR = 2.88; 95% C. I.: 0.93–8.92). Increased consumption of vegetables and fish were associated with a decreased risk. Increased body mass has also been shown to be associated with a higher incidence of small bowel adenocarcinoma (SIR = 2.8; 95% C. I.: 1.6–4.5) (Wolk et al., 2001). However, earlier studies have observed either no association (Chow et al., 1993) or an inverse relationship with body mass (Negri et al., 1999).

SMALL INTESTINAL CARCINOGENESIS: THE INFREQUENCY OF ADENOCARCINOMAS Although adenocarcinoma is the most common histopathologic subgroup of the malignant neoplasms occurring in the small intestine, the incidence in Western industrialized nations is about 1/50th the incidence of colorectal cancer. International patterns of age-standardized incidence rates indicate that the risks of small intestinal and colorectal cancers are positively correlated (Lowenfels, 1973). Per capita consumption of animal protein and fat are positively correlated with international age-standardized small intestinal and colorectal cancer incidence rates (Chow et al., 1993). The risks of second primary adenocarcinomas of the small intestine and of the large intestine are mutually increased (Neuget and Santos, 1993). These observations would suggest that there are common causal determinants for adenocarcinomas, and for the precursor lesions (e.g., adenomas and dysplasias), throughout the intestinal tract. In advancing a unifying hypothesis that accounts for contrasting mucosal susceptibilities, the biochemistry and microbial flora of the fecal stream, enterohepatic bile acid metabolism, mucosal cell proliferation kinetics, and intestinal motility will be reviewed. The contents of the small intestine remain liquid throughout, and then beyond the ileum, the luminal contents become increasingly concentrated. Transit time through the small intestine is generally much faster per unit length than through the large intestine, which may further limit the exposure of the lining epithelium of the small intestine to ingested mutagens and carcinogens. The peristaltic “ring” contractions in the small intestine occur with greater frequency per unit time in the duodenum, and then progressively diminish in the jejunum and ileum. Overall transit time in the large intestine is slow. It may require as long as 1 week for solid radiopaque markers to pass through the colon. The principal points of delay are in the cecum and ascending colon, and in the sigmoid colon and rectum (Cook and Brookes, 2002). The epithelial cells lining the mucosal crypts of the small and large intestine are rapidly proliferating. Mitoses are not seen normally in the absorptive cells of small intestinal villi or colorectal surface epithelium. As new progenitor cells are formed within the crypts, they migrate to and terminally differentiate at the level of villous tips in the small intestine, or onto the flat absorptive surface in the colon. Because of this significant regenerative potential, the intestinal epithelium is susceptible to damage by genotoxic agents. Given the similarities in the dynamics of cell renewal in the small and large intestine, the contrasting cancer incidence patterns require consideration of other microecologic and metabolic differences (Lipkin and Higgins, 1988). The small intestine normally contains relatively small numbers of bacteria. When bacteria are present, they are usually lactobacilli and enterococci, gram-positive aerobes, or facultative anaerobes that are present in concentrations not exceeding 10,000 viable organisms per gram of jejunal contents. Anaerobic bacteroides organisms do not reside in the proximal small intestine. The pH in the jejunum is around 6.0–7.0, but is maintained above 7.0 in the distal ileum. The pH of the cecum is generally around 6.8–7.3. In the ileum, the concentration of

806

PART IV: CANCER BY TISSUE OF ORIGIN

microorganisms increases up to levels of one billion organisms per gram of contents. Strict anaerobes, which normally cannot survive in the jejunum, frequently colonize the ileum at concentrations around one million organisms per gram of contents. Beyond the ileocecal valve, the bacterial flora of the large intestine increase up to one million-fold and are composed of anaerobic organisms such as bacteroides and clostridia, which outnumber aerobic and facultative organisms by 10,000 : 1 (Wilkins and Van Tassell, 1983; Venitt, 1988). The intestinal anaerobic microorganisms generate various metabolic enzymes, such as beta-glucuronidases, sulfatases, reductases, and decarboxylases, which act on various substrates, for example, on cholesterol, bile acids, fatty acids, and steroid hormones. Dietary fat increases the metabolic activity of gut bacteria and bile acid secretion (Reddy and Ohmori, 1981; Reddy et al., 1996). The conceptualized pathogenic mechanism of metabolic activation of ingested chemicals by the indigenous intestinal bacterial flora, and their conversion to mutagens or electrophilic molecules, has been derived mainly from animal studies. In such studies, the rate of chemically induced tumor development in germ-free and antibiotic-treated rodents is compared with that in untreated rodents. For example, cycasin, methylazoxymethanol (MAM) glucoside, found in Cycadeceae plants, is not tumorigenic when given parenterally or orally to germ-free rats (Laquer, 1964). However, when fed orally to rats with normal intestinal microflora, less than 50% of the conjugated compound was recovered in the feces and urine, and adenocarcinomas appeared in the large intestine, liver, biliary passages, and kidney (Bull et al., 1979). Microbial beta-glucosidase converts cycasin to MAM, an active mutagen that is tumorigenic. Tumor induction by dimethylhydrazine (DMH) and 2, 3 dimethyl 4-aminobiphenyl is enhanced by the betaglucuronidase enzymatic activity of intestinal bacteria (Weisburger and Fiala, 1983). The increased risk of adenocarcinoma of the small intestine in patients with Crohn disease occurring in segments of surgically bypassed loops of chronically inflamed ileum is attributed to stagnation of intestinal contents accompanied by bacterial multiplication (Sands, 2002). Bile acids are the water-soluble end products of cholesterol metabolism that are essential for lipid digestion and absorption in the proximal small intestine. More than 90% of the bile acids discharged into the duodenum are absorbed in the terminal segment of the ileum, transported in the portal venous blood back to the liver, extracted by the liver, and then resecreted into bile (Hofmann, 1977; Dawson, 2002). The active transport mechanism in the terminal ileum is specific and efficient for unconjugated bile acids but is inefficient in the absorption of glucuronide, sulfate, and glycine conjugates of bile acids. The primary bile acids formed from cholesterol in the liver are the 3, 7 dihydroxy derivative, chenodeoxycholic acid, and cholic acid (3, 7, 12 trihydroxy bile acid). The bile acids that are not absorbed in the terminal ileum enter the cecum by passing through the ileocecal valve. In the cecum, the anaerobic bacterial enzymes dehydrogenate the hydroxy substituents at the 3, 7, and 12 positions on the primary bile acids. The 7-dehydroxylation of cholic acid results in the formation of deoxycholic acid, and 7-dehydroxylation of chenodeoxycholic acid results in the formation of lithocholic acid. These secondary bile acids formed by the action of colonic bacteria may be passively absorbed from the colon, to some degree, and recycled with the primary bile acids. However, bacterial deconjugation in the colon of residual bile acids results in the fecal passage of unconjugated secondary bile acids. In experimental animal models, the unconjugated secondary bile acids act as potent promoters of colonic carcinogenesis (Hill, 1990; Ross et al., 1991). These pathophysiologic interrelationships will be considered in more detail in the chapter on the large intestine.

PREVENTIVE MEASURES Because the absolute risk of developing small intestinal cancer in the general population is quite small, strategies for the primary prevention of the disease are limited to those at high risk, namely individuals with the predisposing hereditary and acquired conditions mentioned

earlier in this chapter, including inflammatory bowel disease and celiac sprue. The efficacy of potential chemopreventive agents in inducing remission and reducing morbidity due to Crohn disease has been tested in a number of trials. Infliximab is an anti-TNF-a monoclonal antibody that has been shown to induce remission in approximately one-third of patients with CD; however, most patients relapse within months. Monoclonal antibodies act by binding with TNF-a, presumably leading to a reduction in activated immune cells via the complement cascade or apoptosis (Scallon et al., 1995). However, it has been shown that certain genetic polymorphisms of TNF-a and TNF receptor (TNF-R) genes predict non-response to treatment with Infliximab (Mascheretti et al., 2002). Natalizumab is another monoclonal antibody that binds to a4 integrins, and has resulted in remissions of limited duration (Ghosh et al., 2003). The a4 integrins are heterodimeric receptors that activate leukocytes within the vascular endothelium and are accompanied by inflammatory responses. Most patients with celiac disease improve by avoiding glutencontaining grains. It has been shown that patients who adhere to a gluten-free diet experience reduced risk of cancer, including enteropathy-associated T-cell lymphoma. However, of special concern are those individuals with refractory disease who do not respond to dietary intervention. Patients with refractory sprue are often treated with immunosuppressive agents.

FUTURE DIRECTIONS The apparent similarities in the molecular genetic events of physiologic cell renewal and apoptosis, and the events of malignant cell transformation in the small and large intestines, contrast dramatically with global cancer incidence patterns. As stated previously, current research focuses on the biochemistry and microbial flora of the fecal stream, enterohepatic bile acid metabolism, mucosal cell proliferation kinetics, and intestinal motility. Future research should focus on the environmental and host factor components of the causal pathway in the adenoma-carcinoma sequences and its pattern of anatomic distribution in the small intestine. Chronic inflammation and perturbations in immune mechanisms of response should be investigated in studies of primary non-Hodgkin lymphoma of the small intestine and in the pathogenesis of adenocarcinoma of the small and large intestines. References Al-Saleem T, Al-Bahrani Z. 1973. Malignant lymphoma of the small intestine in Iraq (Middle East Lymphoma). Cancer 31:291–294. Arber N, Hibshoosh H, Moss SF, et al. 1996. Increased expression of cyclin Dl is an early event in multistage colorectal carcinogenesis. Gastroenterology 110:669–674. Arber N, Hibshoosh H, Yasui W, et al. 1999. Abnormalities in the expression of cell cycle-related proteins in tumors of the small bowel. Cancer Epidemiol Biomarkers Prev 8:1101–1105. Ashley SW, Wells SAJ. 1988. Tumors of the small intestine. Semin Oncol 15:115–128. Baris D, Gridley G, Ron E, et al. 2002. Acromegaly and cancer risk: A cohort study in Sweden and Denmark. Cancer Causes Control 13: 395–400. Bernstein CN, Blanchard JF, Kliewer E, Wajda A. 2001. Cancer risk in patients with inflammatory bowel disease. Cancer 91:854–862. Beutler B. 2001. Autoimmunity and apoptosis: The Crohn’s connection. Immunity 15:5–14. Bjork J, Akerbrant H, Iselius L, et al. 2001. Periampullary adenomas and adenocarcinomas in familial adenomatous polyposis: Cumulative risks and APC gene mutations. Gastroenterology 121:1246–1248. Boardman LA, Thibodeau SN, Shaid DJ, et al. 1998. Increased risk for cancer in patients with the Peutz-Jeghers syndrome. Ann Intern Med 128: 896–899. Brandt L, Hagander B, Norden A. 1978. Lymphoma of the small intestine in adult celiac disease. Acta Med Scandinavica 204:467–470. Brunner JE, Johnson CC, Zafar S, et al. 1990. Colon cancer and polyps in acromegaly: Increased risk associated with family history of colon cancer. Clin Endocrinol 32:65–71. Bull AW, Burd AD, Nigro ND. 1979. Promotion of azoxymethane-induced intestinal cancer by high fat diets in rats. Cancer Res 39:4956–4959.

Cancers of the Small Intestine Carbone A. 2002. AIDS-related non-Hodgkin’s lymphomas: From pathology and molecular pathogenesis to treatment. Hum Pathol 33:392–404. Catassi C, Fabiani E, Corrao G, et al. 2002. Risk of non-hodgkin lymphoma in celiac disease. JAMA 287:1413–1419. Chan JE, Stampfer MJ, Giovannucci E. 1998. Plasma insulin-like growth factor-I and prostate cancer risk: A prospective study. Science 279: 563–566. Chen CC, Neugut AI, Rotterdam H. 1994. Risk factors for adneocarcinomas and malignant carcinoids of the small intestine: Preliminary findings. Cancer Epidemiol Biomarkers Prev 3:205–207. Chow JS, Chen CC, Ahsan H, Neugut AI. 1996. A population-based study of the incidence of malignant small bowel tumours: SEER, 1973–1990. Int J Epidemiol 25:722–728. Chow W-H, Linet MS, McLaughlin JK. 1993. Risk factors for small intestine cancer. Cancer Causes Control 4:163–169. Cook IJ, Brookes SJ. 2002. Motility of the Large Intestine. In: Feldman M, Friedman LS, Sleisenger MH, eds. Gastrointesteral and Liver Disease. Pathophysiology Diagnosis Management 7th ed. Philadelphia: WB Saunders Co, pp. 1679–1691. Cuthbert AP, Fisher SA, Mirza MM. 2002. The contribution of NOD2 gene mutations to the risk and site of disease in inflammatory bowel disease. Gastroenterology 122:867–874. Dawson PA. 2002. Bile Secretion and the Enterohepatic Circulation of Bile Acids. In: Feldman M, Friedman LS, Sleisenger MH, eds. Gastrointestinal and Liver Disease. Pathophysiology Diagnosis Management 7th ed. Philadelphia: WB Saunders Co, pp. 1051–1064. Dozois RR, Judd ES, Dahlin DC. 1969. The Peutz-Jeghers syndrome. Is there a predisposition to the development of intestinal malignancy? Arch Surg 98:509–515. Erlandson RA, Klimstra DS, Woodruff JM. 1996. Subclassification of gastrointestinal stromal tumors based on evaluation by electron microscopy and immunohistochemistry. Ultrastruct Pathol 20:373–393. Fielding JR, Prior P, Waterhouse JA. 1972. Malignancy in Crohn’s disease. Scand J Gasteroenterol 7:3–7. Fine KD, Stone MJ. 1999. a-heavy chain disease, mediterranean lymphoma, and immunoproliferative small intestinal disease: A review of clinicopathological features, pathogenesis and differential diagnosis. Am J Gastroenterol 94:1139–1152. Freeman HJ, Weinstein WM, Shnitka TK. 1977. Primary abdominal lymphoma. Am J Med 63:585–594. Fujimori M, Ikeda S, Shimizu Y, Okajima M, Asahara T. 2001. Accumulation of b-catenin protein and mutations in exon 3 of b-catenin gene in gastrointestinal carcinoid tumor. Cancer Res 61:6656–6659. Gabos S, Berkel J, Band P. 1993. Small bowel cancer in western Canada. Int J Epidemiol 22:198–206. Ghosh S, Goldin E, Gordon FH, et al. 2003. Natalizumab for active Crohn’s disease. New Engl J Med 348:24–32. Giardello FM, Brensinger JD, Tersmette AC, et al. 2000. Very high risk of cancer in Familial Peutz-Jeghers syndrome. Gastroenterology 119: 1447–1453. Giardello FM, Welsh SB, Hamilton SD. 1987. Increased risk of cancer in the Peutz-Jeghers syndrome. New Engl J Med 316:1511–1514. Gill SS, Heuman DM, Mihas AA. 2001. Small intestinal neoplasms. J Clin Gastroenterol 33:267–282. Giovannucci E, Pollak MN, Platz EA. 2000. A prospective study of plasma insulin-like growth factor-1 and binding-protein-3 and risk of colorectal neoplasia in women. Cancer Epidemiol Biomarkers Prev 9:345–349. Green PHR, Fleischauer AT, Bhagat G, et al. 2003. Risk of malignancy in patients with celiac disease. Am J Med 115:191–195. Green PHR, Jabri B. 2003. Coeliac disease. Lancet 362:383–391. Greenstein AJ, Sachar DB, Smith H. 1980. Patterns of neoplasia in Crohn’s disease and ulcerative colitis. Cancer 46:403–407. Groden J, Thliveris A, Samowitz W. 1991. Identification and characterization of the familial adenomatous polyposis coli gene. Cell 66:589–600. Gruber SB, Entius MM, Petersen GM, et al. 1998. Pathogenesis of adenocarcinoma in Peutz-Jeghers syndrome. Cancer Res 58:5267–5270. Hawker PC, Gyde SN, Thompson H. 1982. Adenocarcinoma of the small intestine complicating Crohn’s disease. Gut 23:188–193. Hemminki A, Markie D, Tomlinson I, et al. 1998. A serine/threonine kinase gene defective in Peutz-Jeghers syndrome. Nature 391:184–187. Hill MJ. 1990. Bile flow and colon cancer. Mutat Res 238:3130–3320. Hofmann AF. 1977. The enterohepatic circulation of bile acids in man. Clin Gastroenterol 6:3–24. Holmes GKT, Stokes PL, Sorahan TM. 1976. Celiac disease, gluten-free diet, and malignancy. Gut 17:612–619. Howe JR, Karnell LH, Scott-Conner C. 2001. Small bowel sarcoma: Analysis of survival from the national cancer data base. Ann Surg Oncol 8:496–508.

807

Hugot JP, Chamaillard M, Zouali H. 2001. Association of NOD2 leucine-rich repeat variants with susceptibility to Crohn’s disease. Nature 411:599–603. Isaacson P, Wright DH, Judd MA. 1979. Primary gastrointestinal lymphomas. Cancer 43:1805–1819. Jenne DE, Reimann H, Nezu J, et al. 1998. Peutz-Jeghers syndrome is caused by mutations in a novel serine threonine kinase. Nat Genet 18:38–43. Kaerlev L, Teglbjaerg PS, Sabroe S, et al. 2002. The importance of smoking and medical history for development of small bowel carcinoid tumor: A European population-based case-control study. Cancer Causes Control 13:27–34. Kaerlev L, Teglbjaerg PS, Sabroe S, et al. 2000. Is there an association between alcohol intake or smoking and small bowel adenocarcinoma? Results from a European multi-center case-control study. Cancer Causes Control 11:791–797. Konishi F, Wyse NE, Muto T. 1989. Peutz-Jeghers polyposis associated with carcinoma of the digestive organs. Dis Colon Rectum 30:790–799. Korelitz BI. 1983. Carcinoma of the intestinal tract in Crohn’s disease: Results of a survey conducted by the National Foundation for Ileitis and Colitis. Am J Gastroenterol 78:44–46. Laquer GL. 1964. Carcinogenic effects of cycad meal and cycasin, methylazoxymethanol glycoside, in rats and effects of cycasin in germfree rats. Fed Proc 23:1386–1387. Lesage S, Zouali H, Cezard JP. 2002. CARD15/NOD2 mutational analysis and genotype-phenotype correlation in 612 patients with inflammatory bowel disease. Am J Hum Genet 70:845–857. Lightdale CJ, Sternberg SS, Posner G. 1975. Carcinoma complicating Crohn’s disease. Report of seven cases and a review of the literature. Am J Med 59:262–268. Lipkin M, Higgins P. 1988. Biological markers of cell proliferation and differentiation in human gastrointestinal diseases. Adv Cancer Res 50:1–24. Lowenfels AB. 1973. Why are small bowel tumors so rare? Lancet 1:24–26. Lynch HT, Smyrk TC, Lynch PM. 1989. Adenocarcinoma of the small bowel in Lynch syndrome II. Cancer 64:2178–2183. Ma J, Pollak MN, Giovannucci E. 1999. Prospective study of colorectal cancer risk in men and plasma levels if insulin-like growth factor (IGF)-I and IGF-binding protein-3. J Natl Cancer Inst 91:620–625. MacDonald JM, Davis WC, Crago HR. 1967. Gardner’s syndrome and periampullary malignancy. Am J Surg 113:425–430. Madara JL, Trier JS. 1987. Functional morphology of the mucosa of the small intestine. In: Johnson LR, ed. Physiology of the Gastrointestinal Tract. 2nd ed. New York: Raven Press, pp. 1209–1249. Manousos O, Souglakos J, Bosetti C. 1999. IGF-I and IGF-II in relation to colorectal cancer. Int J Cancer 83:15–17. Mascheretti S, Hampe J, Kuhbacher T, et al. 2002. Pharmacogenetic investigation of the TNF/TNF-receptor system in patients with chronic active Crohn’s disease treated with infliximab. The Pharmacogenomics Journal 2:127–136. Mellemkjaer L, Johansen C, Gridley G, et al. 2000. Crohn’s disease and cancer risk (Denmark). Cancer Causes Control 11:145–150. Melmed RN, Bouchier IAD. 1972. Duodenal involvement in Gardner’s syndrome. Gut 13:524–527. Mowat AM. 2003. Coeliac disease—a meeting point for genetics, immunology, and protein chemistry. Lancet 361:1290–1292. Murray JA. 1999. The widening spectrum of celiac disease. Am J Clin Nutr 69:354–365. Negri E, Bosetti C, La Vecchia C, et al. 1999. Risk factors for adenocarcinoma of the small intestine. Int J Cancer 82:171–174. Nesbitt RR, Elbadaw NA, Morton JH. 1976. Carcinoma of the small bowel. A complication of regional enteritis. Cancer 37:2948–2959. Neuget AI, Santos J. 1993. The associations between cancers of the small and large bowel. Cancer Epidemiol Biomarkers Prev 2:551–553. Ogura Y, Bonen DK, Inohara N. 2001. A frameshift mutation in NOD2 associated with susceptibility to Crohn’s disease. Nature 411:603–606. Parkin DM, Whelan SL, Ferlay J, Teppo L, Thomas DB. 2002. Cancer Incidence in Five Continents, Volume VIII. Petreshock EP, Pessah M, Menachemi E. 1975. Adenocarcinoma of the jejunum associated with nontropical sprue. Am J Dig Dis 20:796–802. Pines A, Rozen P, Ron E, Gilat T. 1985. Gastrointestinal tumors in acromegalic patients. Am J Gastroenterol 80:266–269. Podolsky DK. 2002. Inflammatory bowel disease. New Engl J Med 347: 417–429. Qizilbash AH. 1976. Familial polyposis col and periampullary carcinoma. Can J Surg 19:166–168. Rashid A, Hamilton SR. 1997. Genetic alterations in sporadic and Crohn’sassociated adenocarcinomas of the small intestine. Gastroenterology 113:127–135.

808

PART IV: CANCER BY TISSUE OF ORIGIN

Reddy BS, Ohmori T. 1981. Effect of intestinal microflora and dietary fat on 3,2¢-dimethyl-4-aminobiphenyl-induced colon carcinogenesis in F344 rats. Cancer Res 41:1363–1367. Reddy BS, Simi B, Patel N, Aliaga C, Rao CV. 1996. Effect of amount and types of dietary fat on intestinal bacterial 7 alpha-dehydroxylase and phosphatidylinositol-specific phospholipase C and colonic mucosal diacylglycerol kinase and PKC activities during stages of colon tumor promotion. Cancer Res 56:2314–2320. Richards ME, Rickert RR, Nance FC. 1989. Crohn’s disease-associated carcinoma. Ann Surg 209:764–773. Ries LAG, Eisner MP, Kosary CL, et al. 2002. SEER Cancer Statistics Review 1975–2000. Ries LAG, Eisner MP, Kosary CL, et al. 2003. SEER Cancer Statistics Review 1975–2000. Ron E, Gridley G, Hrubec Z, et al. 1991. Acromegaly and gastrointestinal cancer. Cancer 68:1673–1677. Ross RK, Hartnett NM, Bernstein L. 1991. Epidemiology of adenocarcinomas of the small intestine: Is bile a small bowel carcinogen? Br J Cancer 63:143–145. Sands BE. 2002. Crohn’s Disease. In: Feldman M, Friedman LS, Sleisenger MH, eds. Gastrointestinal and Liver Disease. Pathophysiology Diagnosis Management 7th ed. Philadelphia: WB Saunders Co, pp. 2005–2038. Scallon BJ, Moore MA, Trinh H, Knight DM, Ghrayeb J. 1995. Chimeric antiTNF-alpha monoclonal antibody cA2 binds recombinant transmembrane TNF alpha and activates immune effector functions. Cytokine 7:251–259. Sellner F. 1990. Investigations on the significance of the adenoma-carcinoma sequence in the small bowel. Cancer 66:702–715. Shanahan F. 2002. Crohn’s disease. Lancet 359:62–69.

Spigelman AD, Murday V, Phillips PKS. 1990. Cancer and the Peutz-Jeghers syndrome. Gut 30:1588–1590. Sutter T, Doi S, Carnevale KA, Arber N, Weinstein IB. 1997. Expression of cyclins Dl and E in human colon adenocarcinomas. J Natl Cancer Inst 28:285–309. Swanson CM, Slavin G, Coles EC. 1983. Celiac disease and malignancy. Lancet 1:111–115. Trier JS. 1993. Celiac sprue. In: Sleisenger MH, Fordtran JS, eds. Gastrointestinal Disease Pathophysiology/Diagnosis/Management. 5th ed. Philadelphia: WB Saunders Co., pp. 1078–1096. Venitt S. 1988. Mutagens in human feces and cancer of the large bowel. In: Venitt S, ed. Role of the Gut Flora in Toxicity and Cancer. New York: Academic Press, pp. 399–460. Weisburger JH, Fiala ES. 1983. Experimental colon carcinogens and their mode of action. In: Autrup H, Williams GM, eds. Experimental Colon Carcinogenesis. Boca Raton: CRC Press, pp. 27–50. Weiss NS, Yang CT. 1987. Incidence of histologic types of cancer of the small intestine. J Natl Cancer Inst 78:653–656. Westerman AM, Entius MM, de Baar E, et al. 1999. Peutz-Jeghers syndrome: 78 year follow-up of the original family. Lancet 353:1211–1215. Wieman TJ, Taber SW, Reed DN Jr., Vitale GC. 2003. Upper gastrointestinal cancer risk in familial adenomatous polyposis. J Ky Med Assoc 101:142–145. Wilkins TD, Van Tassell RL. 1983. Bacterial mutagens in the human colon. In: Hentges DJ, ed. Human Intestinal Microflora in Health and Disease. New York: Academic Press, pp. 265–288. Wolk A, Gridley G, Svensson M, et al. 2001. A prospective study of obesity and cancer risk (Sweden). Cancer Causes Control 12:13–21.

42

Cancers of the Colon and Rectum EDWARD GIOVANNUCCI AND KANA WU

I

n the United States, colorectal cancer is the third leading cause of cancer deaths in each sex and second overall (Landis et al., 1999). At current rates, approximately 6% of individuals in the United States will develop a cancer of the colon or rectum within their lifetime. An estimated 104,950 colon and 40,340 rectal cancer incident cases and approximately 56,290 deaths from colorectal cancer are expected to occur in 2005 (Jemal et al., 2005). The world-wide number of incident colorectal cancer cases diagnosed in 2000 was 944,717 (Parkin et al., 2001). The incidence of colorectal cancer increases with age; when diagnosed before age 40, the causes are usually relatively rare, highly penetrant genetic syndromes. Although the age-standardized rates are higher in men, the number of new cases and deaths is approximately equal between the sexes because of the tendency for women to live longer than men. Colorectal cancer was relatively uncommon prior to 1900, but following economic development, its incidence rose dramatically. In high-incidence countries, cancers of the large bowel are among the most common malignancy, while in low-risk areas, they are quite uncommon, indicating that the majority of colorectal cancers are potentially preventable. In addition, the existence of an accessible and removable precursor lesion, the adenoma, renders this malignancy susceptible to secondary prevention. Colorectal cancers discovered and treated at a relatively early stage have a good prognosis, while those detected at more advanced stages are often incurable. This chapter will review the epidemiology of colorectal cancer, focusing on adenocarcinomas, which comprise the vast majority of the cases. The etiologies of colon and rectal cancer overlap, but the differences between the two will be addressed when relevant.

CLASSIFICATION Anatomic Distribution The large bowel can be separated into the cecum, ascending colon, hepatic flexure, transverse colon, splenic flexure, descending colon, sigmoid colon, and the rectum. The parts of the colon up to the midtransversum are considered the right-sided or proximal colon, whereas the parts of the colon after the mid-transversum are considered the leftsided or distal colon (Haubrich, 1995). Incidence rates for colorectal cancer differ by subsites (see discussion below). Colon cancer develops most frequently in the sigmoid colon (around 25%), followed by cecum (around 20%), transverse colon (including the two flexures: around 15%), and ascending colon (around 10%). Cancers arising from the rectosigmoid junction account for about 10% and rectal cancers account for approximately 20% of all colorectal cancers (Wu et al., 2001).

Histopathology The vast majority of colorectal cancers are adenocarcinomas (Levin and Raijman, 1995), which are preceded by adenomas or adenomatous polyps in most cases. Only around 10% of adenomas will develop into cancers, and this progression can take at least 10 years (Lev, 1990). Three different histologic groups of adenomas have been defined: tubular, tubulovillous, and villous adenomas. The majority of adenomas are tubular adenomas (between 75%–90%) and the least common are villous adenomas, which constitute around 3%–10%

(Lee, 1995). Factors that predict a higher likelihood that an adenoma will progress into cancer include villous histology, larger size (≥1 cm in surface diameter), multiplicity, and degree of dysplasia (Morson, 1984; Cotton et al., 1996). These characteristics are correlated to each other (Lee, 1995). Other types of polyps that have been identified included hyperplastic polyps, which are the most frequently detected colorectal polyps. Hyperplastic polyps had long been considered benign lesions with a low likelihood to progress to cancer (Lee, 1995), but more recent data suggest otherwise. For example, in some hyperplastic polyps, various mutations including K-ras mutations (Otori et al., 1997) or microsatellite instability (see section on molecular genetic characteristics) (Iino et al., 1999) have been identified. Microsatellite instability and other genetic mutations have also been found in another type of polyp, the serrated adenoma (Iino et al., 1999; Fogt et al., 2002; Torlakovic et al., 2003), which combines the serrated morphological appearance of an hyperplastic polyp with the dysplastic features of an adenoma (Longacre and Fenoglio-Preiser, 1990). Mounting evidence suggests that serrated adenomas may also be precursors of colorectal cancers (Torlakovic et al., 2003).

Molecular Genetic Characteristics A model has been formed by Fearon and Vogelstein (1990) describing the progression from adenoma to carcinoma as a process characterized by several steps, each involving genetic mutations of oncogenes (ras gene mutations) and tumor suppressor genes. In most instances, mutations of the tumor suppressor genes affect the chromosomes 5q, 18q, and 17p. The steps involved include the progression from normal epithelium to hyperproliferative epithelium, early, intermediate, late adenoma, and subsequently carcinoma and metastasis. An early event in this multistage process is the mutation of the APC gene (adenomatous polyposis coli gene), which is a tumor suppressor gene located on chromosome 5q. Other alterations include hypomethylation of DNA (DNA methylation may affect the regulation of DNA transcription), mutations of the oncogene K-ras located on chromosome 12p, and loss of the DCC gene (“deleted in colorectal cancer gene”) located on chromosome 18q (Fearon and Vogelstein, 1990). Mutations of the tumor suppressor gene p53, which are commonly found in human cancers (Greenblatt et al., 1994; May and May, 1999) usually occur very late in the adenoma-carcinoma sequence. However, the accumulation of genetic alterations may be more important than their sequence (Fearon et al., 1990). Advances in molecular technology have since then provided consistent support for the existence of such an adenoma-carcinoma sequence, but alternative carcinogenic mechanisms are suggested for some colorectal cancers, such as mutations in mismatch repair genes (Buermeyer et al., 1999; Wheeler et al., 2000a; Peltomaki, 2001; Leslie et al., 2002). In addition, recent research has led to the discovery of several additional cancer-related genes (Leslie et al., 2002). Based on those recent data, Potter (1999) illustrates various pathways depicting the progression from normal cell to cancer cell in a recent review (Fig. 42–1, modified from Potter, 1999). These pathways are described in more detail below:

APC-b-catenin-Tcf-MYC pathway This pathway is an extension of the adenoma-carcinoma model previously proposed by Fearon and Vogelstein (1990). Most colorectal

809

Figure 42–1. Three pathways to colorectal cancer. (Source: Adapted from Potter et al., J Natl Cancer Inst, Vol. 91 (11), 1999, pp. 916–932.)

810

Cancers of the Colon and Rectum cancers are initiated by a mutation of a tumor suppressor gene, the APC gene (either somatic or inherited), which may result in the development of adenoma. If this mutation is inherited (autosomal dominant), it will result in the familial adenomatous polyposis syndrome (FAP), which is characterized by the development of multiple (up to thousands) colorectal adenomas. By age 50, penetrance of colorectal adenomas is 100%. If adenomas are not removed, the risk of one or more becoming malignant is very high (Lynch, 1995). The FAP syndrome is discussed in more detail in the section on genetic and familial susceptibility below. The APC gene encodes the APC protein, which is involved in the regulation of b-catenin. Mutation of the APC gene thus results in increased concentrations of b-catenin, which after adhering to the T-cell factor 4 (Tcf4), mediates transcription of certain genes including the oncogene c-myc (He et al., 1998). Some colorectal cancers without mutations in the APC gene have recently been found to have mutations in the b-catenin gene, providing some support that the pathway leading from normal cell to adenoma may be mediated by the regulation of b-catenin (Morin et al., 1997). Progression from adenoma to carcinoma is then dependent on accumulation of other genetic and epigenetic alterations, such as DNA hypomethylation, and mutations of the K-ras and p53 genes (Fearon and Vogelstein, 1990). Other tumor suppressor genes located on chromosome 18q may be involved in the progression from normal cell to carcinoma; examples include mutations of the SMAD2 or SMAD4 gene (DPC4 gene), which are genes mediating transcription of the transforming growth factor b (TGF b); one of several functions of the protein TGF-b includes inhibition of cell growth (Thiagalingam et al., 1996; Duff and Clarke, 1998; Leslie et al., 2002).

Hereditary Nonpolyposis Colorectal Cancer Pathway Another carcinogenic mechanism involves mutations in DNA mismatch repair genes (Buermeyer et al., 1999; Wheeler et al., 2000a; Peltomaki, 2001). Microsatellites are characterized by 15–30 repeated sets of base pairs, each set consisting of 1–5 base pairs. Generally, deletion or insertion of those repeated sets during DNA replication are repaired. Mutations in mismatch repair genes can negatively affect DNA repair (Wheeler et al., 2000a). A manifestation of mutations in mismatch repair genes (replication error positive (RER+) ) is found in approximately 90% of hereditary nonpolyposis colorectal cancer (HNPCC) (Wheeler et al., 2000a) and in up to 15% of sporadic colorectal cancers (Jass et al., 1998). HNPCC is discussed in more detail in the section on genetic and familial susceptibility below. Several mismatch repair genes have been identified: hMLH1, hMSH2, hPMS1, hPMS2, hMSH6, mutations in the hMLH1 and hMSH2 genes are most commonly found in HNPCC (Peltomaki et al., 2001). In the majority of sporadic colorectal cancers associated with microsatellite instability, hypermethylation of the promoter region of the hMLH1 gene, which inhibits gene transcription, has been found (Herman et al., 1998; Kuismanen et al., 2000; Wheeler et al., 2000b; Peltomaki et al., 2001). Genes other than the APC, K-ras, and p53 genes may be more important in colorectal cancers with microsatellite instability (Olschwang et al., 1997; Peltomaki et al., 2001) because tumors with microsatellite instability occur more frequently in the proximal colon (Thibodeau et al., 1993; Watatani et al., 1996; Breivik et al., 1997; Lleonart et al., 1998; Lynch and de la Chappelle, 1999), which do not commonly have mutations in chromosomes 5q, 17p, and 18q (Delattre et al., 1989). Alterations of two genes have commonly been identified in colorectal cancers with microsatellite instability: a frameshift mutation of the BAX gene, a gene that promotes apoptosis (Rampino et al., 1997) and mutations of the TGF-b type II receptor gene (TGF-b), a tumor suppressor gene (see above) (Markowitz et al., 1995).

The Ulcerative-Colitis Dysplasia Carcinoma Sequence A separate carcinogenesis pathway has been suggested for subjects with ulcerative colitis, an inflammatory bowel disease, which is associated with an approximately 20-fold increased risk of colorectal cancer (Wong and Harrison, 2001; Murthy et al., 2002). Ulcerative colitis is discussed in more detail in the section on predisposing diseases. In patients with ulcerative colitis, chronic inflammation can

811

result in genetic alterations, which can progress to dysplasia and subsequently to cancer (Wong and Harrison, 2001; Murthy et al., 2002). Some evidence indicates that genetic alterations found in ulcerative colitis-associated cancer may differ from those found in hereditary or sporadic colorectal cancer (Murthy et al., 2002); for example, p53 mutations associated with ulcerative colitis tend to occur earlier than in sporadic colorectal cancers (Brentnall et al., 1994).

Evidence for Differences in Genetic Etiology by Cancer Subsites Growing evidence suggests etiologic differences exist for proximal colon cancer, distal colon cancer, and rectal cancer (McMichael and Potter, 1985; Ponz de Leon et al., 1990; Inoue et al., 1995; Kanazawa et al., 1996; Todoroki et al., 1999; Yoo et al., 1999; Knekt et al., 2000). Support for this hypothesis comes from observations that proximal and distal parts of the large bowel differ with regard to their embryologic source and their physiological function, consequently affecting bile-acid metabolism, fecal composition, and transit time (McMichael and Potter, 1985; Bufill, 1990); a gender difference for incidence of colon cancer subsites exists, with women more likely to develop proximal colon cancer than men (McMichael et al., 1985; Bufill, 1990; Dornschneider, 1990; Lampe et al., 1993) (also see below in section regarding demographic characteristics); and certain risk factors for colon cancer such as alcohol intake, calcium intake, and physical activity may be more important risk factors for distal colon cancers (Gapstur et al., 1994; Thune and Furberg, 2001; Wu et al., 2002), whereas cholecystectomy may more likely influence the risk of developing proximal colon cancers (McMichael and Potter, 1985; Todoroki et al., 1999). Subsite differences are also observed for genetic factors. Tumors with microsatellite instability occur more frequently in the proximal colon, whereas in distal colon cancers mutations in p53 and chromosomes 5q, 17q, and 18q are more commonly found (Delattre et al., 1989; Bufill, 1990; Breivik et al., 1997; Lleonart et al., 1998; Lynch and de la Chappelle, 1999; Soong et al., 2000).

DEMOGRAPHIC PATTERNS Overall Mortality and Incidence in the United States According to the American Cancer Society (based on data from the US Surveillance, Epidemiology, and End Results Program (SEER) and the National Center for Health Statistics), the estimated number of colorectal cancer cases in the United States in 2003 is 147,500 (105,500 colon and 42,000 rectum) and the estimated number of deaths from colorectal cancer is 57,100 (Jemal et al., 2003). In 2003, colorectal cancer is expected to be the third overall leading incident cancer and to tie with prostate cancer for the second leading cause of cancer death among men. Among women, colorectal cancer is expected to be the third leading incident cancer and the third leading cause of cancer death. The lifetime probability of developing colon or rectal cancer is 1 in 17 (5.9%) among men and 1 in 18 (5.6%) among women (Jemal et al., 2003).

Incidence Trends Between 1973 and 1985 age-standardized incidence rates of colorectal cancer in both men and women increased but then started to gradually decrease in 1985 (Fig. 42–2). In white males, incidence rates decreased sharply between 1991–1995, then increased slightly after 1995, and since 1997 incidence rates have been declining again. In white females, rates have been declining consistently since 1985, though the rate of decrease declined after 1996. The estimated annual percentage change (EAPC) between 1973 and 1999 is -0.53% per year for white males and -0.76% for white females. Between 1990–1999 though the decline in EAPC is more pronounced in white males (-1.57% per year) than in white females (-0.79% per year). In both black males and females, incidence rates increased dramatically between 1973 and 1980. After 1980, however, rates for blacks remained stable, but were still above those for whites.

812

PART IV: CANCER BY TISSUE OF ORIGIN

Figure 42–2. SEER Standardized Colorectal Cancer Incidence Rates Per 100,000 by Sex and Race, 1973–1999 (2000 US Standard Population).

Incidence rates over time also varied by cancer subsites. Between 1977–1997 incidence rates for the transverse, splenic flexure, descending, sigmoid colon, and the rectosigmoid and rectum have been decreasing, whereas cancer incidence rates for the cecum, appendix, ascending, and hepatic flexure (i.e., right-sided colon) have been rising (Ries et al., 2000). One possible reason for the observed shift in the incidence rates with regard to colon cancer subsites may be that “average risk” individuals usually receive a sigmoidoscopy, which can detect and remove distal adenomas, but not proximal adenomas (Wu et al., 2001). Other possible explanations include findings that risk factors for proximal and distal cancers may differ (see above).

Mortality Between 1973 and 1978 mortality rates for colorectal cancer among white males remained stable, but after 1978 mortality rates started to decrease, initially slightly, with sharper decreases after 1986. Between 1973–1999 mortality rates decreased for white females, especially between 1984–1999. Among black males mortality increased gradually until 1989, then between 1989–1996 the mortality rates remained stable and after 1996 decreased slightly. Among black females the mortality rates have slightly increased between 1973–1985, after 1985 mortality rates declined albeit at a slow rate (Fig. 42–3). Between 1990–1999 decreases in mortality rates were highest in white men (EAPC = -2.23% per year), followed by white females (EAPC =

Figure 42–3. SEER Standardized Colorectal Cancer Mortality Rates per 100,000 by Sex and Race, 1969–1999 (2000 U.S. Standard Population).

-1.68% per year), black males (EAPC = -0.98% per year), and black females (EAPC = -0.56 per year).

Survival According to the most recent SEER data, the overall 5-year survival rate for colorectal cancer is 61.9% in both men and women (among those diagnosed between 1992–1998). Survival rates, however, differ by race, age, and by distribution of disease. Five-year survival rates are lower among blacks than among whites (62.6% vs. 52.8%), but survival rates are similar between black men and black women. Survival rates are highest for localized colorectal cancers (90.1%), followed by regional disease (65.2%), unstaged disease (36.2%), and distant colorectal cancer (8.8%). Overall survival rates are similar for both colon and rectal cancer (61.9% vs. 61.7%), but differ by stage of disease (localized: colon: 91.6%; rectal: 87.1%; regional: colon: 67.7%, rectal: 58.4%; distant: colon: 9.0%, rectal: 8%; unstaged: colon: 32.8%, rectal: 41.4%) (SEER database). Between 1992–1998, 35% of all diagnosed colon cancers were localized and 39% of all colon cancers were considered regional; the respective percentages for rectal cancers were 43% (localized) and 35% (regional).

Age Incidence rates of colorectal cancer start to increase after age 35 years with a rapid increase after age 50 years, when more than 90% of all

813

Cancers of the Colon and Rectum

Figure 42–4. SEER Crude Colorectal Cancer Incidence Rates per 100,000 by Age at Diagnosis, 1955–1999.

blacks and whites, and are lowest for American Indians/Alaska Natives. Mortality rates due to colorectal cancers are highest among blacks, followed by non-Hispanic whites (Table 42–2). Five-year survival are also lower among blacks compared with whites (1992–1998: blacks: 52.8%, whites: 62.6%), partly because blacks tend to be diagnosed at a later stage than whites (1992–1998 SEER): distant disease: blacks: 24%, whites: 19%. Among whites both incidence and mortality due to colorectal cancer have steadily decreased after 1985; blacks, however, did not experience such a decline in incidence and mortality over time (also see above time trends, Figs. 42–2 and 42–3). There are also racial differences regarding anatomic location of colon cancer. Based on recent data from 28 population-based cancer registries (1992–1997), blacks tend to have higher rates for cancers of the descending colon (black to white rate ratio: males: 1.40, females: 1.63) followed by cancers of the transverse colon (including two flexures) (black to white rate ratio: males: 1.22, females: 1.24), cecum (black to white rate ratio: males: 1.14, females: 1.18), and ascending colon (black to white rate ratio: 1.11 for both males and females) than whites. On the other hand, whites have higher rates of cancers of the rectum and rectosigmoid junction compared with blacks (Wu et al., 2001).

Socioeconomic Status colorectal cancers develop. Between 1995–1999, the incidence rates among those aged 30–34 years was 3.6 per 100,000, at age 50–59 years the rate was 48.9 per 100,000, at age 70–79 years the rate was 267.6 per 100,000, and among those over age 85 years the rate was 450 per 100,000. In both black men and women, incidence rates decline after age 75. Black men have higher incidence rates between ages 50–64 years than white men; after age 75 white men have higher incidence rates than black men. Between ages 50–80 years black women have higher incidence rates than white women; after age 80 white women have higher rates than black women (Fig. 42–4).

Gender Between 1995–1999 the overall age-standardized incidence rates for colorectal cancer were 65.1 per 100,000 for men and 47.6 per 100,000 for women (male-female ratio = 1.37). At age 40–44 the male to female ratio was 1.19, but the ratio rose to 1.37 at age 50–54 (1995–1999). Mortality rates due to colorectal cancer were also higher in men than in women (1999: 25.4 per 100,000 in men, 18.0 per 100,000 in women). Men were more likely to develop cancer in the left colon whereas women were more likely to develop cancer in the right colon; the male to female ratio between 1992–1997 among whites was highest for rectal cancer (1.76), followed by rectosigmoid junction (1.62), descending colon (1.56), sigmoid colon (1.51), transverse colon (and 2 flexures) (1.31), and ascending colon (1.18). They were lowest for cecum (1.12) (Wu et al., 2001). Among blacks, the male to female ratios were 1.57 for rectal cancer, 1.35 for descending colon, 1.34 for rectosigmoid colon, 1.28 for transverse colon (and 2 flexures), 1.18 for ascending colon, and 1.08 for cecum (Wu et al., 2001). The reasons why colorectal cancer locations may differ by sex are not clear, but differences in sex hormones, which consequently may influence bile acid metabolism, fecal transit time and composition, may in part be responsible (McMichael and Potter, 1985; Lampe et al., 1993). Differences in sex hormones may also help explain why the overall male to female ratio for colorectal cancer is lower before age 50–54 years (i.e., for premenopausal women) than after age 50–54 years (i.e., for postmenopausal women) (McMichael and Potter, 1985; Bufill, 1990; Dornschneider et al., 1990; Lampe et al., 1993). A recent study suggests that estrogens protect against MSI+ cancers (Slattery et al., 2001).

Race and Ethnicity Table 42–1 shows age-standardized incidence rates by race/ethnicity (SEER database) between 1992–1999. Age-standardized incidence rates for colon cancer are highest for blacks (both males and females), followed by non-Hispanic whites. Colon cancer rates are lowest for Hispanics and American Indians/Alaska natives. On the other hand, age-standardized incidence rates for rectal cancer are similar for

Implementing data from the US census, Singh et al. (2002b) devised a method creating an area socioeconomic index, which was then used to group 3097 US counties into five different socioeconomic (SES) categories. After combining those SES data with mortality data from 1950–1998 in those counties, annual mortality rates were calculated for those five socioeconomic groups and trends in mortality rates over time (1950–1998) were examined separately by socioeconomic status. In women, colorectal cancer mortality in all five socioeconomic groups decreased over time. In men, however, colorectal cancer mortality in the low SES groups increased, whereas mortality in the high SES groups decreased over time. Among women, high SES was associated with higher colorectal cancer mortality rates, but this positive association decreased markedly over time. In younger women (25–64

Table 42–1. Colon Cancer (Invasive) and Rectum Cancer (Invasive), AgeAdjusted SEER Incidence by Race/Ethnicity and Gender Rate 1992–1999 Rate per 100,000 Persons Total

Males

Females

39.4 39.1 25.6 40.3 47.8 32.5 25.3 24.5

45.3 45.3 30.2 46.3 52.9 37.9 28.5 28.8

35.0 34.5 22.3 35.8 44.3 28.3 22.8 21.3

colon incidence Race/Ethnicity All Races White White Hispanic White Non-Hispanic Black Asian/Pacific Islander American Indian/Alaska Native Hispanic

Rate 1992–1999 Rate per 100,000 Persons Total

Males

Females

14.9 14.8 11.8 14.9 14.2 15.4 9.9 11.2

19.3 19.1 15.9 19.0 17.8 20.8 12.2 15.1

11.6 11.6 8.8 11.7 11.5 11.1 8.0 8.4

rectum incidence Race/Ethnicity All Races White White Hispanic White Non-Hispanic Black Asian/Pacific Islander American Indian/Alaska Native Hispanic

814

PART IV: CANCER BY TISSUE OF ORIGIN

Table 42–2. Colon and Rectum Cancer (Invasive), US Death Rates Rate 1992–1999 Rate per 100,000 Persons Mortality

Total

Males

Females

Race/Ethnicity All Races White White Hispanic White Non-Hispanic Black Asian/Pacific Islander American Indian/Alaska Native Hispanic

22.3 21.9 13.7 22.0 29.1 13.7 12.8 13.2

27.1 26.7 17.4 26.9 34.8 16.5 14.6 16.6

18.9 18.4 11.0 18.6 25.4 11.6 11.3 10.6

Source: U.S. Surveillance, Epidemiology, and End Results Program (SEER).

years) mortality rates after 1990 were even higher among women with low SES than in women with high SES. In men aged 25–64 years, a higher SES also appeared to be associated with higher colorectal cancer mortality rates, but the positive association decreased over time and in the 1990s the relationship was reversed. In older men (>65 years), the positive association between SES and colorectal cancer mortality also decreased markedly over time, but did not reverse in the 1990s (Singh et al., 2002a). Results from some other studies that have investigated the associations between colorectal cancer incidence and mortality and SES (or markers of SES such as income) in the United States also do not provide consistent support for a positive association between SES and colorectal cancer mortality and incidence (Ziegler et al., 1986; Mackillop et al., 2000). However, urbanization may be associated with a higher risk of colorectal cancer (Ziegler et al., 1986).

International Patterns Colorectal cancer is a major public health problem throughout the world. The worldwide number of incident colorectal cancer cases in 2000 was estimated to be 944,717 (498,754 in men and 445,963 in women) (Parkin et al., 2001). Incidence rates of colon or rectal cancer vary considerably by country (Parkin et al., 1997), suggesting that environmental and lifestyle factors such as diet may influence colorectal carcinogenesis (Armstrong and Doll, 1975). According to recent cancer incidence data obtained from five continents around 1990 (Fig. 42–5), high colon cancer rates are found in certain areas and ethnic groups in the United States, Canada, Japan, and New Zealand, whereas low colon cancer rates are found in countries such as India or Algeria. High rectal cancer rates are found in certain areas and ethnic groups in Canada, parts of Europe, New Zealand, Israel, and Australia and low rectal cancer rates are found in Algeria and India (Fig. 42–5) (Parkin et al., 1997). Worldwide more variation exists for colon cancer incidence than for rectal cancer incidence, and in most countries men are more likely to develop rectal cancer than women. Incidence of rectal cancer is higher than that of colon cancer in some countries that have low colon cancer incidence rates, such as India and Algeria (Parkin et al., 1997). The majority of colorectal cancers still arise in the industrialized countries. With the introduction of a Western lifestyle, incidence rates for colon and rectal cancer usually start to rise (Waterhouse et al., 1982; Muir et al., 1987; Parkin et al., 1992; Parkin et al., 1997). For example in Japan (Miyagi), the age-adjusted incidence rate for colon cancer in males was 8.3 per 100,000 between 1973–1977, 9.8 per 100,000 between 1978–1981, 17.1 per 100,000 between 1983–1987 and 24.9 per 100,000 between 1988–1992. A more recent example is China (Shanghai), where colon cancer incidence rates in males steadily increased from 6.7 per 100,000 in 1975 to 12.2 between 1988–1992. Rises in incidence of colon cancer incidence over time have also been observed in several Eastern European countries such as Slovenia (colon cancer incidence in males: 1978–1981: 8.7 per 100,000, 1982–1987: 10.1 per 100,000, 1988–1992: 15.7 per 100,000).

Migration Results from migrant studies provide support that lifestyle and environment may influence colorectal carcinogenesis (Ziegler et al., 1986). In a classic study by Haenszel and Kurihara (1968), colorectal cancer mortality rates between 1949–1952 and between 1959–1962 among Japanese who had immigrated the United States were higher than among Japanese residing in Japan, but lower than mortality rates in US whites. Between 1949–1952 and 1959–1962, the difference in colorectal cancer mortality rates between Japanese Americans and US whites had also decreased over time (Haenszel and Kurihara, 1968). In a later study by Locke and King (1980), colon cancer mortality rates among Japanese Americans in 1970 had shifted nearer towards those of US whites (males: SMR = 75, females: SMR = 79) and mortality from rectal cancer in men had reached those of US whites (males: SMR = 113, females: SMR = 87) (Locke and King, 1980). Wynder et al. (1991) also noted a higher occurrence of leftsided cancers, especially sigmoid colon, compared with right-sided cancers among Japanese natives over time, indicating that Japanese may also be shifting nearer towards US whites with regard to colon cancer subsites. In a more recent study (Flood et al., 2000), colorectal cancer incidence rates for men between 1973–1986 among Japanese Americans born in the United States were about 60% higher than those for US whites, and for Japanese Americans born in Japan, rates were about 20% lower than the rates for US whites. In women, incidence rates among Japanese Americans born in the United States were about 40% higher than those for US whites, and for Japanese Americans born in Japan, rates were about the same as those for US whites (Flood et al., 2000). The higher colorectal cancer incidence rate among the Japanese men and women born in the United States was mainly evident for rectal (including the rectosigmoid junction) and distal colon cancers. Migrant studies have also been conducted in Chinese Americans. In the 1970s, foreign-born Chinese Americans (mainly immigrants from the Guangzhou province in China) had higher colorectal cancer mortality than Chinese living in Guangzhou (King et al., 1985). A more recent study (Flood et al., 2000) found that in men, colorectal cancer incidence rates among Chinese Americans born in the United States were about 30% lower than those for US whites, and for Chinese Americans born in China, rates were about the same as those for US whites. In women, incidence rates among Chinese Americans born in the United States were about 40% lower than those for US whites, and for Chinese Americans born in China, rates were about 30% lower than those for US whites (Flood et al., 2000). Trends in colorectal cancer mortality rates have also been studied for migrant populations from European countries. Lilienfeld et al. (1972a; 1972b) noted that around 1960 the mortality rates from colorectal cancer among migrants to the United States from several European countries were close to or for some countries even higher (particularly among male immigrants from Ireland) than those for USborn whites. However, for most countries, mortality rates for residents living in their native country were lower than those for US-born whites. Another study examined cancer mortality and incident rates in European immigrants to Australia. McCredie et al. (1999) examined trends in colorectal cancer mortality according to duration of residence in New South Wales, Australia, and found that among immigrants from the Great Britain and southern Europe, the relative risk of dying from colorectal cancer (compared with Australian-born) increased with increasing duration.

ENVIRONMENTAL FACTORS Nutrition Fruit, Vegetables, and Fiber. The concept that a diet high in fiber, especially from fruits and vegetables, lowers risk of colorectal cancer has been in existence for over four decades, following the observation of the relative rarity of colorectal cancers in populations in Africa that have a high-fiber diet (Burkitt, 1971). Subsequently, the relation between fiber, as well as fruits and vegetables in general, and colorectal cancer risk has been evaluated in case-control and cohort

815

Cancers of the Colon and Rectum

Colon Rectum

A

Colon Rectum

B

Figure 42–5. (A) Age-standardized colorectal cancer incidence rates per 100,000 in men for 20 selected countries around 1990. (B) Age-

standardized colorectal cancer incidence rates per 100,000 in women for 20 selected countries around 1990. (Source: Parkin et al., 1997.)

studies. The majority of case-control studies showed an association between higher intake of vegetables, and possibly fruits, and lower colon cancer risk. Earlier on, Trock et al. (1990) conducted a metaanalysis of six case-control studies and found that a high intake of vegetables was associated with approximately half the risk for colon cancer, and high fiber intake was associated with about a 40% reduction in risk. A pooled analysis that included 13 case-control studies (Howe et al., 1992) reported an approximately 50% lower risk of colon cancer associated with higher fiber intake. Just as the evidence for the fiber hypothesis appeared to be consolidating, the results from large prospective cohort studies began emerging over the past decade. These studies, in contrast to the earlier case-control studies, have tended to show a weak or nonexistent inverse association for fiber and risk of colon cancer (Willett et al., 1990; Bostick et al., 1994; Giovannucci et al., 1994d; Goldbohm et al., 1994; Fuchs et al., 1999), although a recent report from the EPIC study suggests otherwise (Bingham et al., 2003). Similarly null

results were observed for total fruits and vegetables (Michels et al., 2000). In a recent comprehensive prospective study examining the role of fiber and its components on risk of colorectal neoplasms, Fuchs et al. (1999) found that a high-fiber diet did not protect against colorectal cancer or adenoma. Furthermore, no important associations were observed when analyses were conducted for cereal, fruit, or vegetable fiber. Even in a Finnish population that has a wide range of fiber intakes (16.0–34.1 g/day between bottom and top quartiles), fiber intake was not related to risk of colorectal cancer (Pietinen et al., 1999). Neither were soluble and insoluble fiber, whole grain cereal and rye products, and vegetables and fruits. However, a recent report from the large EPIC study found an inverse association with fiber (Bingham et al., 2003). The reasons for the apparent inconsistencies between the casecontrol and most of the recent cohort studies are not clear. In general, case-control studies are more prone to bias because dietary information is collected after the diagnosis of cancer. Moreover, recent

816

PART IV: CANCER BY TISSUE OF ORIGIN

randomized intervention studies using recurrent adenomas as the endpoint have not supported a role of fruits and vegetables (Schatzkin et al., 2000), wheat bran fiber (Alberts et al., 2000), or ispaghula fiber (Bonithon-Kopp et al., 2000). These studies are discussed in more detail in the Chemoprevention section. Another relevant factor may be that in the United States, multivitamins and fortified breakfast cereals may provide potentially beneficial factors (such as B-vitamins including folate). The acquisition of these factors from multivitamins and fortified cereals may have attenuated the effect of fruits and vegetables, which may be the primary sources of these vitamins in other populations. While a modest-sized influence for fruits, vegetables, and fiber may be possible, the recent data suggest that these effects, if present at all, are likely to be much weaker than had been previously believed. Diets high in fruits, vegetables, and fiber may have many health benefits, but a substantial reduction in colon cancer risk may not be among one of the benefits.

Folate. Higher intakes or higher circulating levels of folate are relatively consistently related to a lower risk of colorectal adenomas (Benito et al., 1991; Giovannucci et al., 1993b; Bird et al., 1995; Paspatis et al., 1995; Boutron-Ruault et al., 1996; Tseng et al., 1996), and colorectal cancer in case-control studies (Benito et al., 1991; Freudenheim et al., 1991; Meyer and White, 1993; Ferraroni et al., 1994; White et al., 1997) and prospective studies (Giovannucci et al., 1995d; Glynn et al., 1996; Ma et al., 1997; Giovannucci et al., 1998; Kato et al., 1999; Su and Arab, 2001; Terry et al., 2002a). A few studies have provided equivocal results (Boutron-Ruault et al., 1996; Slattery et al., 1997c), but the majority of the studies are supportive. The risk reductions in these studies have ranged from approximately 20%–50%. These findings were confirmed in an analysis from the Pooling Project of Prospective Studies, which is based on combined data from nine cohort studies in North America and Europe (Kim et al., 2001). In 503,237 men and women, among whom 4824 incident cases of colorectal cancer were diagnosed in up to 13 years of followup, persons in the top quintile of total folate intake were at significantly lower risk of colorectal cancer (multivariate RR = 0.79, 95% CI: 0.70–0.89, Ptrend = 0.002). Both dietary and supplementary sources of folate were protective in that analysis. Complementary to the studies of folate are those of alcohol, which is a potent antagonist of folate metabolism, (see Alcohol section). Although the inverse association between folate and colorectal neoplasia risk is relatively modest, the results are more striking when folate and alcohol intakes are considered jointly. In general, a twofold to fivefold elevation in colorectal cancer or adenoma risk is observed among individuals with high intakes of alcohol and low intakes of folate (Freudenheim et al., 1991; Giovannucci et al., 1993b; Giovannucci et al., 1995d; Glynn et al., 1996; Baron et al., 1998; Giovannucci et al., 1998; Kato et al., 1999; Su and Arab, 2001). The association is more consistent in men. Also relevant to the folate hypothesis is the role of methylenetetrahydrofolate reductase (MTHFR), an enzyme at a critical metabolic branch point that directs the folate pool towards remethylation of homocysteine to methionine, at the expense of thymidylate synthesis (Fig. 42–6). The enzyme irreversibly converts 5,10 methylenetetrahydrofolate to 5-methyltetrahydrofolate. Kang et al. (1992) reported a thermolabile variant of MTHFR, associated with increased plasma total homocysteine and a relative deficiency of cellular methionine. Later this variant was found to be related to a single nucleotide polymorphism of the MTHFR gene (677C Æ T), which was associated with an alanine-to-valine substitution (Frosst et al., 1995). Because mechanisms involving either 5-methyltetrahydrofolate (DNA methylation) or 5,10-methylenetetrahydrofolate (DNA synthesis and repair) could influence colorectal neoplasia, polymorphic forms of MTHFR may interact with folate and alcohol in influencing colorectal carcinogenesis. Four studies have reported on interactions between folate or alcohol and risk of colorectal cancer (Chen et al., 1996; Ma et al., 1997; Slattery et al., 1999; Le Marchand et al., 2002) and six have reported on adenoma risk (Chen et al., 1998; Ulrich et al., 1999; Levine et al., 2000; Marugame et al., 2000; Ulvik et al., 2001; Giovannucci et al., 2003). Thus far, the data strongly suggest

the following pattern: among individuals with the CC or CT genotypes, only modest associations are observed between alcohol and folate and risk of colorectal neoplasia, but individuals with the TT genotype appear to be “hyper-responders” to folate or alcohol. Specifically, TT homozygotes are at relatively low risk (compared with those with CC or CT genotypes) if they have a “low-risk” diet (high folate, low alcohol), but have no apparent protection, or even elevated risks, if they have a “high-risk” diet (high alcohol, low folate) (Fig. 42–6). This pattern is most striking for alcohol, for which statistically significant interactions with this MTHFR genotype have been observed in three studies (Chen et al., 1996; Ma et al., 1997; Levine et al., 2000).

Calcium. Calcium has been hypothesized to reduce colon cancer risk by binding secondary bile acids and ionized fatty acids and forming insoluble soaps of these potentially toxic compounds in the lumen of the colon (Newmark et al., 1984; van der Meer and de Vrieg, 1985), or by directly lowering proliferation of the colonic mucosa (Lipkin and Newmark, 1985). Large prospective studies relatively consistently show a modest, typically non-statistically significant inverse association, but the dose-response relationship is not clear (Martinez and Willett, 1998). An analysis of cases from both the Nurses’ Health Study and the Health Professionals Follow-Up Study found that total, dietary, and supplemental calcium reduced the risk of distal colon cancer but not of proximal colon cancer (Wu et al., 2002). The study also found that most of the benefit is achieved by reaching intakes of 700–800 mg/day, suggestive of a threshold effect. In the Physicians’ Health Study, higher calcium intake was related to a lower risk primarily in men who have relatively high circulating concentrations of IGF-1 (Ma et al., 2001). The epidemiologic studies are now supported by results from randomized trials using adenoma recurrence as the end point (see Chemoprevention section). The data strongly suggest that the avoidance of low intakes of calcium may minimize risk of colon cancer. Whether very high intakes of calcium beyond a certain threshold would further lower risk requires more study. Vitamin D may also play a role, but the epidemiologic data are inadequate at this point to allow a firm conclusion.

Figure 42–6. Under conditions of low folate/high alcohol, the MTHFR 677CÆT polymorphism may increase cancer risk through methylation pathway aberrations; when folate is high and alcohol is low, this polymorphism may decrease risk by increasing the pool of 5,10-methylene tetrahydrofolate, which is required for DNA synthesis.

Cancers of the Colon and Rectum

Macronutrients: Fat, Carbohydrates and Protein. Excess energy intake and insufficient physical activity, which in combination lead to obesity, increase the risk of colon cancer, as summarized elsewhere in this chapter. Because of the relationship between BMI and colon cancer, excessive intake of any of the important energy-supplying macronutrient components of the diet could contribute potentially to higher risk of colon cancer. Thus, concerning macronutrients, the relevant question becomes whether the individual energy-supplying macronutrients, independent of their contribution to total energy intake, are related to colon cancer risk. For example, dietary fat could possibly increase risk of colon cancer because it contributes to energy intake, but would an individual’s risk be lowered if fat is replaced with an equivalent caloric content of carbohydrate? Moreover, does it matter whether the fat is predominantly saturated, mono-unsaturated, or poly-unsaturated fat, and whether the carbohydrates are highly refined or not? In general, neither case-control nor prospective studies have supported the hypothesis that replacing fat isocalorically with carbohydrates would lower risk (Willett et al., 1990; Howe, 1993; Bostick et al., 1994; Giovannucci et al., 1994d; Goldbohm et al., 1994). The relationship between BMI and colon cancer suggests that overconsumption of either fat or carbohydrates, resulting in obesity, could increase risk of colon cancer. A growing body of evidence even suggests that diets rich in highly refined carbohydrates may be an independent risk factor of colon cancer, possibly through their hyperinsulinemic effects (Giovannucci, 2001a). Within populations, total protein intake has not been consistently related to higher risk of colorectal cancer. However, the different sources of protein may have different effects. Red meat intake has been examined in many epidemiologic studies (Bjelke, 1980; Phillips and Snowdon, 1983; Stemmermann et al., 1984; Garland et al., 1985; Hirayama, 1986; Gerhardsson et al., 1988; Bostick et al., 1994; Giovannucci et al., 1994d; Goldbohm et al., 1994), and most, though not all, suggest an increase in colon cancer risk associated with higher red meat intake. Recent meta-analyses of the epidemiologic studies showed that high intake of processed meats, as opposed to fresh meats, increases risk (Sandhu et al., 2001; Norat et al., 2002). In contrast, sources of animal protein other than red meat including low-fat dairy products, fish, and poultry, have been associated with lower risk of colon cancer or adenoma in most studies (Hirayama, 1986; Willett et al., 1989; Kato et al., 1990; Giovannucci et al., 1992; Benito et al., 1993; Neugut et al., 1993; Sandler et al., 1993; Bostick et al., 1994; Giovannucci et al., 1994d; Goldbohm et al., 1994). These results do not support an adverse effect of protein, and even suggest a benefit. The underlying mechanism for this potential benefit is unknown, but these foods are good sources of methionine, which may be beneficial regarding DNA methylation (see section on Folate). The reason why red meat, as opposed to other sources of protein, tends to be associated with increased risk of colon cancer remains unclear. Some of the relevant hypotheses have stated that red meat is a major source of total fat, saturated fat, protein, carcinogens, or heme iron. Results of some studies suggest that risk of colon cancer may be increased among meat eaters who consume meat with a heavily browned surface, but not among those who consume meat with a medium or lightly browned surface (Lee et al., 1989; Gerhardsson de Verdier et al., 1991). When meat undergoes prolonged frying, grilling, or broiling at high temperatures, mutagenic heterocyclic amines are formed from creatinine reacting with amino acids (Sugimura and Sato, 1983; Sugimura, 1985; Wakabayashi et al., 1992; Norat et al., 2002). Regardless of the mechanism, which remains obscure, the literature provides support for the intake of sources of protein other than red meat, both animal and vegetable, to lower colon cancer risk.

Energy, Obesity, and Physical Activity Body Mass and Fat Distribution. Energy intakes and expenditures cannot be measured precisely enough in individuals to allow for a direct calculation of energy balance in epidemiologic studies. However, measures of body mass provide an estimate of longterm energy balance, with excessive energy intake leading to obesity (although individual metabolic variation in the response to high energy

817

intakes undoubtedly exists). Case-control and prospective studies show that a higher BMI is associated with an increased risk of colon cancer (Lew and Garfinkel, 1979; Waaler, 1984; Garland et al., 1985; Phillips and Snowdon, 1985; Wu et al., 1987; Klatsky et al., 1988; Chute et al., 1991b; Lee et al., 1991; Le Marchand et al., 1992; Must et al., 1992; Bostick et al., 1994; Giovannucci et al., 1995b; Martinez et al., 1997). The association is typically linear and approximately a twofold higher risk is observed in individuals who are overweight or obese. Some more limited evidence indicates that the tendency for the central distribution of adipose (visceral or central) adiposity increases risk of colon cancer independently of BMI. For example, in one prospective study (Giovannucci et al., 1995b), men in the upper 20% of the distribution for waist-to-hip ratio had about a three and a halffold higher risk of colon cancer than men in the lower 20%. A somewhat puzzling finding has been that the association between BMI and colon cancer risk appears to be stronger and more consistently observed for men than for women. Nonetheless, most studies support an increased risk in women with higher BMI (Lew and Garfinkel, 1979; Phillips and Snowdon, 1985; Wu et al., 1987; Chute et al., 1991b; Martinez et al., 1996; Slattery et al., 1997a; Ford, 1999). Some recent studies indicate that the reason for this gender difference is that the relationship between BMI and colon cancer risk becomes attenuated as women age (Terry et al., 2001; Terry et al., 2002b), suggesting that obesity in women may be more deleterious for colon cancer in a high-estrogen environment (pre-menopausal) than in a low-estrogen (post-menopausal) environment. Reasons why obesity may increase risk for colon cancer are not established, but insulin resistance and the resulting hyperinsulinemia, because of the growthpromoting effects of insulin, have been proposed as the underlying mechanism, as discussed elsewhere in this chapter (Giovannucci, 1995a).

Physical Activity. One of the most consistent relationships observed for colon cancer has been an inverse association with physical activity (reviewed by (Colditz et al., 1997) ). Both prospective cohort and case-control studies have consistently found that individuals who are more physically active have a decreased risk of colon cancer. In addition, higher levels of physical activity have been associated with a reduced risk of colon adenoma, particularly large adenomas (Kono et al., 1991; Little et al., 1993; Giovannucci et al., 1995b; Giovannucci et al., 1996; Neugut et al., 1996). In general, a relationship with rectal cancer has not been observed. The inverse association between higher physical activity level and lower colon cancer risk is likely to be causal because of its consistency in many studies of various designs, its presence for (large) adenomas and cancer, for men and women, in diverse populations, for both leisuretime and occupational activity, and because of the apparent lack of confounding by dietary and other lifestyle factors (Thun et al., 1992; Giovannucci et al., 1995b; Martinez et al., 1996; Slattery et al., 1997a). While physical activity may lower risk in part by reducing the occurrence of obesity, some benefits also appear to be independent of BMI. The highest risk of colon cancer has been observed among persons who are both physically inactive and who have high BMIs (Giovannucci et al., 1995b; Slattery et al., 1997a). In a review of the literature, Colditz et al. (1997) concluded that the incidence of colon cancer was reduced by approximately 50% among the most active individuals compared with the least active. Overall, the results from these studies are suggestive of a dose-response relation with risk reduction present across a wide range of physical activity frequency and intensity. While greater risk reductions are possible with even higher levels of activity, even moderate levels of physical activity (e.g., brisk walking for 3–4 hours/week) are associated with substantial benefits. The mechanism of preventive action for physical activity is not known, but may involve in part the strong influence of activity in reducing insulin levels (discussed elsewhere in this chapter) (Lindgarde and Saltin, 1981; Wang et al., 1989; Dowse et al., 1991; Regensteiner et al., 1991). Physical activity may also increase colonic motility, though colonic motility has not been definitely linked to colon cancer risk.

818

PART IV: CANCER BY TISSUE OF ORIGIN

OTHER LIFESTYLE FACTORS Alcohol The relationship between alcohol and colorectal cancer risk has been controversial, but the weight of evidence now indicates that high intake of alcohol increases risk of colorectal cancer (Kune and Vitetta, 1992b). Not all studies are supportive, but an association between alcohol intake and colon cancer risk has been observed in many prospective studies and case-control studies, and studies of adenomas (reviewed by (Kune and Vitetta, 1992b) ). A recent meta-analysis found that high consumers of alcohol had an elevated risk of colorectal cancer (Bagnardi et al., 2001). In the Pooling Project of Prospective Studies, a pooled multivariate risk of colorectal cancer of 1.24 (95% CI: 1.07–1.42) was observed for consumption of ≥30 g/day of alcohol compared to low intake (Cho et al., 2002). This association, observed in many studies, is not confounded by other known risk factors of colorectal cancer, and is generally observed for both rectal and colon cancers. Whether an association exists for intakes of alcohol lower than 30 g/day (about 2 drinks/day) is unclear. The mechanism of action is unknown, but the antagonist effect of alcohol on folate metabolism (Finkelstein et al., 1974; Barak et al., 1987; Shaw et al., 1989) has received interest in recent years (see section on Folate).

Tobacco Our understanding of the role of tobacco in colorectal carcinogenesis has evolved in the past decade. The early classic studies of smoking and cancer did not find cigarette smoking associated with an increased risk of colorectal cancer (Hammond and Horn, 1958; Hammond and Horn, 1966; Kahn, 1966; Weir and Dunn, 1970; Doll and Peto, 1976; Doll et al., 1980; Rogot and Murray, 1980). The time period of risk covered by these studies, conducted primarily in men, was predominantly in the 1950s and 1960s. However, when risk factors for colorectal adenomas, which had become established as cancer precursors (Morson, 1974; Lev, 1990), were first studied in the 1980s and 1990s, smokers were found to have an elevated risk in over 20 published studies (reviewed by (Giovannucci, 2001b) ). In general, individuals who smoke one to two packs of cigarettes (approximately 20–40 cigarettes per day) or who have a 20 to 40 cigarette pack-year history have about a twofold to threefold increased risk of adenoma. Based on the null results for cancers, opposed to the strikingly positive studies for adenomas, Giovannucci et al. hypothesized that smoking may act as an initiator of colorectal neoplasia in the large bowel, but that an extraordinarily long induction period was required to observe an increase in cancer incidence. (Giovannucci et al., 1994a; Giovannucci et al., 1994b). Because US men generally started smoking substantially in the 1920s, the lack of an association for cancer in the studies that covered the 1950s and 1960s suggested the existence of at least a 30–40-year induction period. Indeed, studies of US men conducted after 1970 have almost consistently supported an association (Wu et al., 1987; Sandler et al., 1988; Slattery, 1990; Giovannucci et al., 1994b; Heineman et al., 1994; Chyou et al., 1996; Le Marchand et al., 1997; Slattery et al., 1997b; Hsing et al., 1998; Chao et al., 2000; Stürmer et al., 2000). Women in the United States began smoking significantly only during the late 1940s and 1950s (Pierce et al., 1989); thus, assuming a 40 or so year induction period, a rise in the incidence of colorectal cancer would not have been expected until the late 1980s. This expectation has been confirmed in recent studies (Chute et al., 1991a; Giovannucci et al., 1994a; Newcomb et al., 1995; Le Marchand et al., 1997; Slattery et al., 1997b; Chao et al., 2000). An association between smoking and colorectal neoplasia should not be surprising because the burning of tobacco generates a wide range of genotoxic compounds, including polynuclear aromatic hydrocarbons, heterocyclic amines, nitrosamines, and aromatic amines (IARC Working Group on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, 1986), which reach the large bowel mucosa either through the circulatory system (Yamasaki and Ames, 1977) or perhaps through direct ingestion (Kune et al., 1992a). An important remaining issue is how quickly risk drops after one quits smoking. Most data, but with exceptions (Chao et al., 2000), indicate that some of the excess risk persists indefinitely in past smokers (Wu et al., 1987;

Giovannucci et al., 1994a; Giovannucci et al., 1994b; Slattery et al., 1997b; Stürmer et al., 2000). Another question is whether the association is limited to rectal cancers, or whether it also exists for colon cancer. Generally, associations have been observed for both colon and rectal cancers, but in some studies, the association has been appreciably stronger for rectal cancer (Doll et al., 1994; Giovannucci et al., 1994a; Heineman et al., 1994; Newcomb et al., 1995) or perhaps limited to rectal cancer (Inoue et al., 1995; Chyou et al., 1996; Nyrén et al., 1996). In regards to the population attributable risk due to smoking in the United States, estimates have been 21% of colorectal cancer in men (Giovannucci et al., 1994b), 16% of colon cancer, and 22% of rectal cancer in men (Heineman et al., 1994); 12% of colorectal cancer in men and women (Chao et al., 2000), and 11% of colon cancer and 17% of rectal cancer in women (Newcomb et al., 1995).

HOST FACTORS Genetic Susceptibility Syndromes Genetic Susceptibility. There are two known main inherited syndromes that predispose to colorectal cancer: familial adenomatous polyposis (FAP) (Lynch, 1995), and hereditary nonpolyposis colorectal cancer (HNPCC) (Lynch and Smyrk, 1998).

Familial Adenomatous Polyposis. Familial adenomatous polyposis, also called familial polyposis coli or adenomatous polyposis of the colorectum, is a very rare, inherited, autosomal dominant syndrome. The frequency of this syndrome in the population is estimated to be 1 in 8000 (Bisgaard et al., 1994). Clinically, this syndrome is characterized by the occurrence of multiple colorectal adenomas (usually a minimum of 100 to up to several thousands) in individuals in their 20s and 30s. If lesions are not treated (i.e., via colectomy or proctocolectomy), colorectal cancer will very likely develop (Lynch, 1995; Fearnhead et al., 2001), and the risk appears to be positively related to the number of adenomas (Debinski et al., 1996). The genetic mutation underlying this disorder is the inherited form of a mutation of the APC gene (Lynch, 1995; Fearnhead et al., 2001). A somatic mutation of APC is an important early event in the development of sporadic colorectal cancer (see previous section on Molecular Genetic Characteristics) (Fearon and Vogelstein, 1990). By age 50, penetrance is almost 100%, but substantial differences in phenotypic expression have also been reported (Fearnhead et al., 2001). Some individuals develop a modified form of FAP, characterized by the combination of colorectal adenomas and extracolonic tumors. In Turcot syndrome, additional tumors occur in the central nervous system (Hamilton et al., 1995; Lynch, 1995), and in Gardner syndrome, affected individuals may also develop osteomas, epidermal cysts, or abnormalities of the teeth (Lynch, 1995). Some FAP patients may also develop desmoid tumors (fibrous tumors), commonly in the abdominal wall or retroperitoneum, while others may also present with other gastrointestinal tumors, such as periampullary carcinomas (Fearnhead et al., 2001; Houlston et al., 2001). A milder form of FAP, called attenuated FAP (AAPC) and characterized by fewer adenomas, occurs when mutations of APC occur closer to the 5¢ end of the APC gene (Spirio, 1993; Houlston, 2001) than those found in the typical form of FAP. In addition, some animal data support the existence of so-called modifier genes, which may at least in part explain differences in the phenotypic expression of the FAP syndrome; however, human data so far have not been conclusive on this issue (Houlston, 2001). Individuals with a family history of FAP should be routinely screened for colorectal adenomas starting in their teens (Lynch, 1995). Hereditary Nonpolyposis Colorectal Cancer. The HNPCC syndrome, also called Lynch Syndrome, is another autosomal dominant inherited disorder that predisposes to colorectal cancer, and may account for approximately 1%–5% of colorectal cancer cases (Lynch and Smyrk, 1996). Cancer usually develops at an early age (approximately mid-forties), and about two-thirds of cancers occur in the proximal colon (Lynch and Smyrk, 1996; Vasen et al., 1999).

Cancers of the Colon and Rectum

819

HNPCC is more difficult to diagnose than FAP because the occurrence of adenomas is not common. However, if adenomas are present, they are more likely to be villous adenomas and have a higher grade of dysplasia than sporadic adenomas (Lynch and Smyrk, 1996; Vasen et al., 1999). In addition to colorectal cancers (Lynch I Syndrome: only colorectal cancers), some patients with HNPCC can also develop cancers outside the colorectum, including tumors of the endometrium, ovary, stomach, pancreas, small bowel, hepatobiliary tract, ureter, and renal pelvis (Lynch II Syndrome) (Wheeler et al., 2000a). However, molecular findings indicated that Lynch I and II Syndromes are not likely to correspond to two distinct diseases (Wheeler et al., 2000a). Increased microsatellite instability has been found in approximately 90% of HNPCC (Wheeler et al., 2000a). Genetic alterations have been identified in several mismatch repair genes including hMSH2, hMLH1, hPMS1, hPMS2, and hMSH6, and mutations in the hMLH1 and hMSH2 genes are most commonly found in HNPCC (approximately 70% of patients with HNPCC) (Wheeler et al., 2000a; Peltomaki, 2001). Penetrance of this inherited disorder has been estimated to be 80%–85% (Wheeler et al., 2000a), but penetrance of the mutations in hMLH1/hMSH2 genes appears to be higher in men (about 80%) than in women (about 40%) (Mitchell et al., 2002).

slow acetylators (Hein, 2000). Slow NAT2 acetylators include around 50%–60% of Caucasians (Gertig and Hunter, 1998). Recent studies have also implicated several alleles for the NAT1 gene, but the functional importance of those genotypes must be determined (Brockton et al., 2000). Epidemiological evidence as to whether genetic polymorphisms in the NAT1 or NAT2 genes may affect colorectal cancer risk is not consistent, but results depended on whether associations between NAT genotypes or phenotypes and colorectal cancers were investigated (de Jong et al., 2002). In the recent pooled analysis by de Jong et al. (2002), neither NAT1 fast genotype or NAT2 fast genotype and colorectal cancer were associated with risk of colorectal cancer. On the other hand, NAT2 fast phenotype was significantly associated with an approximately 70% increased risk of colorectal cancer. Polycyclic aromatic hydrocarbons are metabolized to a carcinogenic compound through a reaction that requires the enzyme ary1 hydrocarbon hydroxylase, which is encoded by the P450 CYP1A1 gene (Gertig and Hunter, 1998). The m1 and m2 polymorphisms of this gene have been studied with regard to colorectal cancer risk, but the pooled analysis did not find any evidence for an association between any of these two polymorphisms and risk of colorectal cancer (de Jong et al., 2002).

Family History of Colorectal Cancer in First-Degree Relative

GSTm1, GSTq1. The glutathione-S transferase enzymes (GSTs) belong to a group of biotransformation enzymes responsible for deactivating carcinogenic compounds, such as PAH. Of the four GST isoenzymes that have been found in humans (a, m, p and q), two (m and q) have been examined more closely with regard to their relationship with colorectal cancer. Those two isoenzymes are encoded by the glutathione-S transferase m1 (GSTm1) and the glutathione-S transferase q1 (GSTq1) genes, which are polymorphic (Cotton et al., 2000). Homozygous deletion of these genes (null genotype) has been identified and is associated with lack of enzymatic activity (Cotton et al., 2000; Grubben et al., 2001). The frequency of GSTm1 null genotype in the US population varies with ethnicity. In blacks, frequencies between 23%–41% and in whites frequencies between 35%–62% have been reported; the respective frequencies regarding the GSTq1 null genotype are between 15%–27% for whites and 22%–29% for blacks (Cotton et al., 2000). Heterozygous forms do not appear to be of any functional significance (Cotton et al., 2000). In the pooled analysis by de Jong et al., no association between GSTm1 null genotype and colorectal cancer was found, whereas the GSTq1 null genotype was associated with an approximately 40% increased risk of colorectal cancer (de Jong et al., 2002).

Family history of colorectal cancer in one or more first-degree relatives is associated with increased risk of colorectal cancer. In a recent meta-analysis of 27 case-control and prospective studies that investigated the association between family history and risk of colorectal cancers (Johns and Houlston, 2001), the pooled relative risk of colorectal cancer was quite high in those with more than one relative with colorectal cancer (RR = 4.25; 95% CI: 3.01–6.08). The relative risk of developing colorectal cancer was 2.25 (95% CI: 2.00–2.53) for those with one first-degree relative of colorectal cancer and the risk was higher if the first-degree relative was diagnosed with colon cancer (RR = 2.42; 95% CI: 2.20–2.65) than if the relative was diagnosed with rectal cancer (1.89; 95% CI: 1.62–2.21) (Johns and Houlston, 2001).

Low-Penetrance Genes and Susceptibility to Colorectal Cancer In contrast to high penetrance genes, such as APC, hMLH1, and hMLH2, a number of low penetrance genes that may also increase susceptibility to develop colorectal cancer have been identified (de Jong et al., 2002). Some investigations have suggested that interactions between these low penetrance genes and environmental factors (geneenvironment interaction) may play a role in the development of sporadic colorectal cancer (Gertig and Hunter, 1998). A recent extensive pooled analysis by de Jong et al. (2002) investigated associations in 30 polymorphisms in 20 genes that have been reported in at least two colorectal adenoma or cancer studies. In this pooled analysis seven of the 30 polymorphisms (GSTq1, NAT2 (phenotype), HRAS1, ALDH2, MTHFR and Tp53 (intron3) and TNF) were associated with either increased or decreased risk of colorectal cancer. Some of these polymorphisms are discussed here, while MTHFR is discussed in more detail in the section on Folate.

NAT1, NAT2, and CYP1A1. Meat intake and smoking have been hypothesized to play a role in colorectal carcinogenesis. Among the mechanisms suggested is the carcinogenic effect of aromatic and heterocyclic amines (HAA), and polycyclic aromatic hydrocarbons (PAH), found in cooked meat and cigarette smoke (IARC Working Group on the Evaluation of the Carcinogenic Risk of Chemicals to Humans, 1986; Manabe et al., 1991; Alexandrov et al., 1996; Hoffmann and Hoffmann, 1997; Norat et al., 2002; Pisani and Mitton, 2002). Heterocyclic amines are activated by a reaction that is catalyzed by an enzyme called N-acetyltransferase (NAT). In humans, two distinct NAT genes, both located on chromosome 8, have been identified: NAT1 and NAT2 (Minchin et al., 1993). A number of allelic variants have been identified for the NAT2 gene. Based on their phenotypes, these have been grouped into three categories: rapid, intermediate, and

ALDH2. The enzyme aldehyde dehydrogenase, encoded by the gene ALDH2, is involved in alcohol catabolism. Alcohol is first broken down to acetaldehyde by alcohol dehydrogenase. ALDH2 catalyzes the oxidation of acetaldehyde, which is considered a carcinogenic compound (Yokoyama et al., 1998). Individuals with a mutant allele of ALDH2 have decreased activity of this enzyme. This polymorphism is highly prevalent among Asians and rare in other populations (Yokoyama et al., 1998; de Jong et al., 2002). Heterozygotes with ALDH2 have higher levels of acetaldehyde if they drink alcohol, though on average they drink less due to flushing. Homozygotes tend to avoid alcohol (Yokoyama et al., 1998). According to the pooled analysis by de Jong et al. (2002), individuals who were either homozygous or heterozygous for this variant allele appeared to be at higher risk for colorectal cancer, consistent with a carcinogenic effect of acetaldehyde. This finding also supports a causal role of alcohol. Growth Factors and Related Factors Circulating Insulin-Like Growth Factor(IGF)-1 and Insulin-Like Growth Factor Binding Protein (IGFBP)-3 An increasing body of evidence suggests that individuals with normal but relatively high circulating levels of IGF-1, which has mitogenic and anti-apoptotic properties, are at elevated risk of colorectal cancer. As discussed elsewhere in this chapter, an increased risk of colorectal neoplasia is observed in acromegalics, who have abnormal elevations in IGF-1. In the Physicians’ Health Study (Ma et al., 1999), men in the top quintile of circulating IGF-1 had a 2.5-fold elevated risk of

820

PART IV: CANCER BY TISSUE OF ORIGIN

colorectal cancer compared with those in the bottom quintile. For those in the bottom quintile of IGFBP-3, a greater than threefold higher risk of colorectal was observed relative to those in the top quintile. Similar results were observed for colorectal cancers and “high-risk” adenomas (e.g., ≥1 cm in diameter, those with a villous component or high dysplasia, three or more adenomas, in situ cancers) in the Nurses’ Health Study and the Flexi-Scope Trial, but not for “low-risk” adenomas (small, tubular adenomas) (Giovannucci et al., 2000; Renehan et al., 2001). These results suggest that IGF-1 and IGFBP-3 influence tumor progression rather than tumor formation or initiation. Other studies also confirm that higher levels of IGF-1 are associated with an increased risk of colorectal cancer, though the role of IGFBP-3 has been less consistent (Manousos et al., 1999; Kaaks et al., 2000).

Hyperinsulinemia As discussed elsewhere in this chapter, a large body of evidence concerning determinants (i.e., obesity, central obesity, physical inactivity) or markers (NIDDM) of insulin resistance and hyperinsulinemia indirectly support a role of high insulin levels in the promotion of colon cancer. Data directly relating pre-diagnostic circulating insulin to risk of colon cancer are limited at this point. One study, the Cardiovascular Health Study (Schoen et al., 1999), followed about 6000 men and women aged 65 or older and found that fasting and 2-hour glucose and 2-hour insulin were significantly and linearly related to higher risk, but that fasting insulin was not. This study was followed by another study based on prospectively collected serum samples from 14,275 women in New York and 95 cases of colorectal cancer (Kaaks et al., 2000). In that study, colon cancer risk was increased fourfold in those in the highest quintile of C-peptide, a marker for insulin secretion, relative to those in the lowest quintile. In addition, several studies have examined colorectal neoplasia in relation to hypertriglyceridemia, a marker of insulin resistance (Topping and Mayes, 1972; Tobey et al., 1981). These studies indicate that higher levels of plasma triglycerides are associated with an increased risk of colorectal carcinoma in situ (Yamada et al., 1998) or adenoma (Bayerdorffer et al., 1993a; Bird et al., 1996; Manus et al., 1997). These limited data support an association between hyperinsulinemia or insulin resistance with colon cancer risk, but more definitive data are needed for firmer conclusions to be drawn.

Height Tallness has been observed to be an independent risk factor for colon cancer, but probably not rectal cancer, in most studies that have assessed this association (Albanes et al., 1998; Chute et al., 1991a; Giovannucci et al., 1995b; Hebert et al., 1997; Ghadirian et al., 1998; Robsahm and Tretli, 1999). Tallness may be acting as a surrogate for some factor related to determinants of height. An interesting candidate for this factor is IGF-1. As discussed elsewhere in this chapter, higher concentrations of circulating IGF-1 appear to be a risk factor for colorectal cancer. Childhood and adolescent levels of IGF-1 influence linear growth (particularly leg length) and correlate with attained height (Juul et al., 1994). Thus, the relationship between tallness and colon cancer risk is consistent with the hypothesis that high concentrations of circulating IGF-1 during growth predisposes to higher risk of colon cancer.

Related Medical Conditions Inflammatory Bowel Disease. Inflammatory bowel disease (IBD) includes ulcerative colitis and Crohn disease (granulomatous colitis). These relatively rare conditions have distinct clinical and pathologic features. Patients with IBD are at substantially elevated risk of colorectal cancer; however, because of the relative rarity of these conditions and treatment (coloectomy), only about 1% of patients with colorectal cancer have a previous history of inflammatory bowel disease. The overall increased risk has been estimated to be fourfold to 20-fold. The magnitude of the cancer-associated risk increases in relation to early onset of the disease, the extent of involvement

of mucosa, the duration of the symptoms, and the presence of multicentric foci of dysplasia. The malignant pathway in these patients usually does not involve an adenoma precursor. For ulcerative colitis, the cumulative incidence for those with disease after 25–35 years has been reported to approach 50% (Podolsky, 2002). Crohn disease affects the ileum and sometimes the large bowel. The risk of colorectal cancer is increased in Crohn patients, but to a much lesser degree than for ulcerative colitis patients. Patients with IBD require intensive surveillance.

Gallstones and Cholecystectomy. A large number of studies have considered the relationship between cholecystectomy and risk of colorectal cancer. A smaller number has considered the relationship for gallstone disease. The interest in cholecystectomy is based primarily on the fact that individuals with their gallbladders removed experience a continuous rather than periodic excretion of bile acids into the intestine. The constant exposure of bile acids to the colonic bacteria may increase the proportion of potentially toxic secondary bile acids. A previous meta-analysis of the literature suggested a moderate increase in risk of proximal colon cancer from case-control studies (RR = 1.34) but not from prospective studies (Giovannucci et al., 1993a). Moreover, the evidence was strongest for case-control studies that used hospital-based controls rather than population-based controls. The results from some studies suggested that risk increased over time since the cholecystectomy. An additional concern is that many of the potential risk factors for gallstones may overlap with those for colon cancer (e.g., obesity, physical inactivity, insulin resistance, diabetes); thus, the relationship may possibly not be causal but may be confounded by these other factors. Few studies have taken these into account. Overall, the data are suggestive of a modestly increased risk of proximal colorectal cancer in individuals who have had a cholecystectomy. However, the inconsistencies across studies and by study design, the relatively modest magnitude of the association, and the potential for uncontrolled confounding raise questions regarding the causal nature of this association. The data are not compelling in supporting that individuals that have had a cholecystectomy should receive enhanced screening. Glucose Intolerance, Non-Insulin Dependent Diabetes Mellitus. A commonality in the etiology of colon cancer and non-insulin dependent diabetes mellitus (NIDDM) is suggested by their similar geographic patterns. Specifically, both diseases were relatively rare before industrialization, and their incidence increases in regions undergoing economic development. Besides genetics, the major determinants of NIDDM include excess adiposity, central obesity, physical inactivity, excessive caloric intake, and dietary patterns that stimulate insulin secretion (Tuomilehto et al., 1992), factors remarkably similar to risk factors for colon cancer. In recent years, the risk of colon cancer has been studied in individuals with NIDDM. NIDDM has been associated with a higher risk of colon cancer in casecontrol studies (La Vecchia et al., 1991; Le Marchand et al., 1997), adenoma studies (Kono et al., 1998), and prospective studies (Will et al., 1998; Hu et al., 1999; Nilsen and Vatten, 2001). The temporal relationship between NIDDM and colon cancer risk may be complex because early in the natural history of NIDDM, hyperinsulinemia exists, whereas in later stages, pancreatic ß-cell depletion leads to hypoinsulinemia (DeFronzo et al., 1992). This pattern between insulin levels and stage of NIDDM may explain why in one prospective study, the association between NIDDM and colon cancer was strongest 11–15 years after the diagnosis (approximately threefold increase in risk) but became attenuated >15 years since diagnosis (Hu et al., 1999).

Acromegaly. Acromegaly is a clinical condition characterized by excessive production of growth hormone and IGF-1. Acromegalic patients experience an increase in epithelial cell proliferation and an extension of the proliferative zone in the colonic mucosa, and growth hormone and IGF-1 levels correlate with cell proliferation rate (Cats et al., 1996; Jenkins, 1999). Moreover, a large number of small studies indicate that acromegalics are at elevated risk of developing both

Cancers of the Colon and Rectum benign and malignant colorectal tumors (Klein et al., 1982; Ituarte et al., 1984; Pines et al., 1985; Ritter et al., 1987; Ziel and Peters, 1988; Brunner et al., 1990; Barzilay et al., 1991; Ron et al., 1991; Terzolo et al., 1994; Vasen et al., 1994; Cheung and Boyages, 1997; Jenkins et al., 1997; Orme et al., 1998; Bari et al., 2002 #8736). In one study of 129 acromegalics, the prevalence of colorectal neoplasia found during a colonoscopic examination was 13-fold higher than that expected based on published rates for asymptomatic screened controls (Jenkins et al., 1997). In the same study extended to 222 patients, serum IGF-1 levels were significantly higher in patients with a recurrent adenoma than in those without any (mean, 390 mg/l vs. 244 mg/l; P < 0.005) (Jenkins, 1999). In a multicenter retrospective cohort study of 1362 patients with acromegaly, colon cancer mortality was increased (standardized mortality ratio, 2.47; 95% CI: 1.31–4.22); mortality from other cancers was not elevated (Orme et al., 1998). In a nationwide, registry-based cohort study of 1634 patients hospitalized for acromegaly in Denmark and Sweden, the standardized incidence ratio was 2.6 (95% CI: 1.6–3.8) for colon cancer and 2.15 (95% CI: 1.3–4.2) for rectal cancer (Bari et al., 2002).

821

genotoxic events in early-related genes, such as APC and b-catenin. In addition to the parallel in timing of events for exposures and molecular events, some links are additionally supported by mechanism. For example, folate and alcohol may be related to colorectal cancer through changes in DNA methylation, and NSAIDs may be more protective in tumors with COX-2 overexpression. To date, there are minimal data relating specific etiologic factors to molecular events in tumors, but this is a very active area of study by many investigators. Finally, as suggested in the section on genetic polymorphisms, genetic susceptibility may modify the influence of environmental factors. For example, MTHFR appears to influence the role of folate and alcohol, and ALDH2 influences the effect of alcohol. Potentially, carcinogens from tobacco and cooked meat may be influenced by NAT and GSTq. In the future, examining genes that influence insulin sensitivity and IGF-1 determinants will be of interest.

PREVENTIVE MEASURES Primary Prevention

PATHOGENESIS Over the past two decades, our understanding of colorectal carcinogenesis has progressed substantially from a morphologic, molecular, and epidemiologic basis. We are now able to begin to link environmental and genetic susceptibility to molecular and morphologic events in tumors. Figure 42–7 shows a hypothetical model taking into account environmental factors, genetic susceptibility, and molecular alterations. From a morphologic perspective, most colorectal cancers progress from normal cells to cells with increased and disordered growth, to small adenomas with less dysplasia to large dysplastic adenomas, and finally to cancer. Some of the important molecular events have been mapped to the various stages of colorectal cancer. For example, APC/b-catenin mutations tend to occur early (initiators) whereas p53 mutations tend to occur late in the process. Environmental factors also tend to be relevant at distinct stages. For example, tobacco appears to act early, possibly through genotoxic effects in genes that are relevant early. The timing is inferred because tobacco is strongly related to risk of small adenomas, but only related to cancer risk after a time lag of several decades. In contrast, some factors such as obesity, physical activity, and circulating IGF-1 tend to be associated with large adenoma and cancer risk, but not to risk of small adenoma. Thus, it is likely that these factors act on adenoma progression but probably not in their initiation. Some links between environmental and genetic factors can be suggested by the timing of events and sometimes by mechanism. For example, it seems reasonable to hypothesize that tobacco is related to

The potential for primary prevention for colorectal cancer is substantial. Unquestionably, some aspects of the Western diet and sedentary lifestyle lead to an increased incidence of this malignancy. In high-risk countries, cancers of the large bowel are among the most common malignancy, whereas in low-risk areas, they are relatively uncommon. Thus, while some controversy remains concerning what are the specific etiologic factors, clearly the majority of colorectal cancers are potentially preventable. The evidence is quite strong that prevention of smoking during adolescence and early adulthood, the prevention of weight gain, and maintenance of a reasonable level of physical activity in adulthood can prevent a substantial proportion of colorectal cancers. The avoidance of excessive alcohol (3 or more drinks per day) may also reduce occurrence of some colorectal cancers. Avoidance of low calcium intakes may also be beneficial. These factors influence many other major health-related conditions in addition to colorectal cancer. While controversy exists regarding the role of specific dietary factors, consideration of the dietary pattern as a whole may be useful for formulating recommendations. For example, several studies show that high intakes of red and processed meats, high-fat dairy products, highly refined grains and starches, and sugars are related to a higher risk of colon cancer. Thus, replacing these factors with poultry, fish, and plant sources as the primary protein sources, mono-unsaturated and poly-unsaturated fats as the primary fat sources, and unrefined grains, legumes, and fruits as the primary carbohydrate sources is likely to lower risk of colorectal cancer. This benefit is likely to occur even though neither the independent benefit of each component, nor the precise mechanism, is established. The strong concordance of many of these factors to those that influence coronary disease and adult-onset diabetes suggest a common underlying mechanism, such as insulin resistance and hyperinsulinemia. Increasing evidence also suggests that the use of a multivitamin with RDA levels of vitamins and possibly minerals may be beneficial, especially in alcohol drinkers, although it is advisable currently to avoid “mega-doses”. The magnitude of the potential benefit of primary prevention was demonstrated in an analysis of colon cancer in male health professionals. In that study, as least 70% of colon cancers were found to be potentially preventable by moderate changes in diet and lifestyle (Platz et al., 2000). Secondary prevention, through screening by sigmoidoscopy and colonoscopy, is also critically important to prevent mortality from colorectal cancer. However, many of the diet and lifestyle risk factors for colorectal cancers are similar for cardiovascular disease and for some other cancers, so focusing on the modifiable risk factors for colorectal cancer is likely to have many additional benefits beyond this cancer.

Chemoprevention Figure 42–7. Hypothetical model that takes into account environmental factors, genetic susceptibility, and molecular alterations. (-) signifies an inverse association

Chemoprevention is based on the concept that a number of years may elapse in the process of carcinogenesis before a precursor lesion becomes invasive, allowing the opportunity to intervene with agents ranging from vitamins, minerals, micronutrients, phytochemicals, or

822

PART IV: CANCER BY TISSUE OF ORIGIN

synthetic compounds. For individuals at average risk of colorectal cancer, chemopreventive agents with virtually no side effects must be identified. If the probability of important side effects is large enough, some potential chemopreventive agents may be unacceptable for use in low- to moderate-risk individuals. However, in those with a high risk of colorectal cancer, such as those with FAP, the use of such agents may be justified. Potential cancer-inhibitory effects of specific agents can be tested in a randomized clinical trial setting. A major challenge for chemoprevention trials is the large numbers of participants that would have to be enrolled to allow detection of an important effect within a reasonable time. For cancer as the end point, the required numbers are large (tens of thousands) and the number of years may be long. Thus, the most common approach used for colorectal cancer chemoprevention has been to test specific agents on the risk of new adenomas among individuals with a history of adenomas. Typically, patients who have undergone removal of adenoma(s) are randomized to intervention or placebo, and then followed for a period of approximately 3 or more years, at which time they undergo a colonoscopic examination for detection of new adenomas. Although the adenoma model is a useful part of our armamentarium, and has been used with some success, some limitations must be considered. Most importantly, the influence of agents that are active at earlier or later stages in the colorectal carcinogenesis sequence are not assessed using this design. Some factors (e.g., tobacco) may be acting well before the appearance of an adenoma, and some factors (e.g., obesity, physical inactivity) seem to be more important for adenoma progression than for adenoma initiation. The influence of such factors (or countering agents) could be missed using this design. Moreover, while it may be a reasonable assumption, whether the factors that influence the occurrence of new (initial) adenomas are identical to those that occur in later-occurring adenomas (the ones assessed in trials) is not entirely clear. Another point to consider is that because the follow-up time in these trials may be 3 years or less, falsenegative findings may result from an insufficient follow-up period. Finally, not all adenomas may be equally susceptible to progress further to cancer, so the magnitude of the influence of an intervention on cancer may differ from the result on adenomas. Despite these limitations, chemoprevention trials using recurrent adenomas are important in providing randomized data that avoids potential confounding, which may occur in observational studies. Ideally, results from intervention trials are integrated with those of other sources of data, including mechanistic, animal, and observational data. Several agents have been tested or are being evaluated for chemopreventive potential, specifically against colorectal neoplasia or composite end points that include colorectal cancer. They include folic acid, ursodeoxycholic acid (a bile acid modifier), wheat bran and other types of fiber, calcium, low fat-high vegetable diet, NSAIDs (e.g., aspirin, sulindac, sulindac sulfone, celecoxib), and estrogen replacement in women. The following is a review of agents, to date, that have been investigated or show promise as interventions for colorectal cancer prevention.

Bile Acid Modifiers. For almost half a century bile acids have been suspected to be important colorectal carcinogens or promoters. The primary bile acids, cholic and chenodeoxycholic acids, are synthesized in the liver and then secreted into the intestines through the bile duct. In the colon, the bile salts that have not been reabsorbed through the small intestine are metabolized by anaerobic bacteria into deoxycholic and lithocholic acids, called secondary bile acids. Small amounts of chenodeoxycholic acid are also converted by bacterial enzymes into ursodeoxycholic acid (Federowski et al., 1979). Deoxycholic acid is hypothesized to enhance colorectal carcinogenesis based on experimental (Rafter et al., 1987; Lapre and Van der Meer, 1992) and observational (Stadler et al., 1988; Bayerdorffer et al., 1993b) studies. However, the role of deoxycholic acid in human colorectal cancer remains controversial. Oral supplementation of ursodeoxycholic acid, which is generally found in trace amounts in humans, decreases the concentration of deoxycholic acid and the formation of tumors in rodent models (Earnest et al., 1994). The potential benefit

of 600 mg daily ursodeoxycholic acid is being tested in a double-blind, placebo-controlled trial on the 3 year adenoma rates among 1200 subjects with a history of adenomas.

Fiber. As discussed elsewhere in this chapter, recent prospective epidemiologic data have cast doubts for the fiber-colorectal cancer hypothesis. In the past several years, this hypothesis has been tested in randomized trials using new adenomas in adenoma patients as the end point. Alberts et al. (2000) conducted a randomized trial based on dietary supplementation with wheat-bran fiber. In that study, 1429 men and women 40 to 80 years of age and who had had one or more colorectal adenomas removed within 3 months before recruitment were randomized to receive either 13.5 g per day or 2 g per day of wheatbran fiber. By the time of the last follow-up colonoscopy at about 3 years, at least one new adenoma had been identified in 47% of the high-fiber group and 51.2% of the low-fiber group. The multivariate adjusted odds ratio for new adenoma in the high-fiber group, as compared with the low-fiber group, was 0.88 (95% CI: 0.70–1.11; P = 0.28), and the number of new adenomas did not differ across the two groups (P = 0.93). Schatzkin et al. (2000) randomly assigned 2079 men and women 35 years of age or older who had colorectal adenomas removed within 6 months before randomization to one of two groups: an intervention group given intensive counseling and assigned to follow a diet that was low in fat (20% of total calories) and high in fiber (18 g of dietary fiber per 1000 kcal) and fruits and vegetables (3.5 servings per 1000 kcal), and a control group given a standard brochure on healthy eating and assigned to follow their usual diet; 1905 of the randomized subjects (91.6%) completed the 4-year study; 39.7% of the intervention group and 39.5% of the control group had at least one new adenoma; the unadjusted risk ratio was 1.00 (95% CI: 0.90–1.12). Also, number, size, or degree of dysplasia did not differ between the two groups. In another study of fiber from the ispaghula husk (Bonithon-Kopp et al., 2000), 665 patients with a history of colorectal adenomas were randomized to three treatment groups, in a parallel design: 2 g elemental calcium daily, fiber (3.5 g ispaghula husk), or placebo. Among the participants who completed the follow-up examination, at least one adenoma developed in 28 of 176 patients (15.9%) in the calcium group, 58 of 198 (29.3%) in the fiber group, and 36 of 178 (20.2%) in the placebo group. The adjusted odds ratio for adenoma recurrence for fiber treatment was 1.67 (95% CI: 1.01–2.76, p = 0.042). This increased risk associated with this type of fiber was unexpected, and combined with the other trials and prospective epidemiologic studies, raises questions on the effectiveness of fiber alone to reduce risk of colorectal cancer. Calcium. As discussed elsewhere in this chapter, epidemiologic, animal, and mechanistic studies support a potential benefit of calcium against colorectal neoplasia. These data led to the design and completion of randomized trials testing the potential benefits of calcium. Results of an intervention trial of calcium supplementation (1200 mg of elemental calcium vs. placebo) among 913 participants found a moderate but statistically significant reduction in risk of adenoma recurrence (Baron et al., 1999). The rate of new adenomas was 31% in the calcium group and 38% in the placebo group (RR = 0.76; 95% CI: 0.60–0.96). Similar results were observed in the European Calcium Fibre Polyp Prevention trial (RR = 0.66 (95% CI: 0.38–1.17; p = 0.16)) (Faivre et al., 1999) for 2 g elemental calcium daily vs. placebo, although the result was not statistically significant in this relatively small study. These data strongly support the increase of calcium in diets, particularly in those who are deficient. Recent epidemiologic studies suggest a possible threshold effect (Wu et al., 2002). NSAIDs. Initial interest in NSAIDs as a means of chemoprevention was stimulated by animal (Pollard and Luckert, 1980; Narisawa et al., 1981; Pollard and Luckert, 1981a; Pollard and Luckert, 1981b; Barnes and Lee, 1998) and by epidemiologic studies (Kune et al., 1988; Rosenberg et al., 1991; Thun et al., 1991; Suh et al., 1993; Giovannucci et al., 1994c; Muscat et al., 1994; Peleg et al., 1994; Schreinemachers and Everson, 1994; Giovannucci et al., 1995c; La

Cancers of the Colon and Rectum Vecchia et al., 1997; Freedman et al., 1998) that generally found about a 30%–50% reduction in colon and rectal cancer risk among long-term aspirin users. In addition, studies of adenomas also suggested a 30%–50% risk reduction among aspirin users (Giovannucci et al., 1994c). The results from these studies led to strong interest for aspirin, as well as piroxicam, sulindac, sulindac sulfone, and celecoxib as potential colorectal cancer chemopreventive agents. Results from some randomized trials are now available for aspirin. In the Physicians’ Health study (Gann et al., 1993), 22,071 healthy volunteers were randomized to either 325 mg of aspirin every other day or placebo to examine the impact of aspirin on cardiovascular disease risk. After 5 years of follow-up, there was no reduction in risk of colorectal cancer for the aspirin group compared with placebo (RR = 1.15; 95% CI: 0.80–1.65). The null results in the Physicians’ Health Study could have been due to short duration in follow-up or low dose of aspirin. Investigators at Dartmouth University conducted a multicenter clinical trial of aspirin (325 mg/day vs. 80 mg/day vs. placebo) in patients with a history of colorectal adenomatous polyps to examine the new adenomas as the outcome. In 1084 patients, the incidence of one or more new adenomas was 47% in the placebo group, 38% in the group given 81 mg aspirin/day, and 45% in those receiving 325 mg/day (Baron et al., 2003). Interestingly, the effect was stronger for tubulovillous or villous adenomas, lesions that have a higher probability to progress to malignancy than do tubular adenomas. In a simultaneously published study of 517 randomized patients who had colonoscopies after a previous diagnosis of colorectal cancer, a statistically significant 35% reduction in risk of a new adenoma was observed in those randomized to 325 mg of aspirin per day compared with those receiving placebo (Sandler et al., 2003). The results of these intervention trials provide strong confirmation that aspirin use reduces colorectal neoplasia occurrence, although the question of the appropriate or optimal dose remains open. Whether NSAIDs, including aspirin, will be suitable candidates for chemoprevention remains a question because of their significant toxicities, including increased risk of gastrointestinal bleeding, ulcers, and kidney damage (Pennisi, 1998). The action of most NSAIDs involves their inhibition of the enzyme cyclooxygenase (Earnest et al., 1992). There are two cyclooxygenases: the first, and the most abundant (COX-1), is involved in normal maintenance of cellular function; the second (COX-2) is induced by injury or inflammation. The recent focus of research on cyclooxygenase inhibitors has been the identification of agents that block COX-2. Celecoxib, almost exclusively a COX-2 inhibitor, is presently one of the most promising of the modern-day NSAIDs. In experimental systems, Celecoxib has been shown to lessen aberrant crypt formation (Reddy et al., 1996), and the incidence, multiplicity, and weight burden of induced tumors (Kawamori et al., 1998) with low toxicity. Chemoprevention trials of celecoxib are presently underway.

Folic Acid. As described elsewhere in this chapter, increasing epidemiologic, genetic, and experimental evidence support a preventive role of folic acid on colorectal carcinogenesis. The generally nontoxic nature of this agent, a vitamin, makes it an ideal candidate for chemoprevention. Ongoing trials are assessing the role of folic acid on new adenomas in adenoma patients. Antioxidants and Other Micronutrients. Some chemoprevention trials have tested the anti-carcinogenic potential of antioxidant nutrients, such as selenium, carotenoids, and vitamins A, C, and E on colorectal cancer incidence (The Alpha-Tocopherol BetaCarotene Cancer Prevention Study Group, 1994; Clark et al., 1996) or new adenoma in adenoma patients (Paganelli et al., 1992; Greenberg et al., 1994; MacLennan et al., 1995). The results have been inconclusive. A remarkable statistically significant 50% reduction in colorectal cancer incidence was shown for a selenium (in the form of brewer’s yeast) intervention in the Nutritional Prevention of Skin Cancer (Clark et al., 1996). The study was conducted in seleniumdeficient areas in the United States. Because these results were based on secondary end point data, additional large trials are needed for confirmation.

823

Vegetables and fruits are the major sources of most dietary antioxidants so the weak results in recent observational studies (discussed elsewhere in this chapter), as well as a recent large randomized trial (Schatzkin et al., 2000), raise doubts as to the importance of dietary antioxidants for colorectal cancer. Vitamins A, C, and E were also shown not to confer protection in a randomized trial of recurrent adenomas (Greenberg et al., 1994). At present, the data do not support a strong role of antioxidant vitamins to combat colorectal cancer.

Post-Menopausal Hormone Use. Women who use postmenopausal hormones have about a 30%–40% decreased risk of colon or colorectal cancer in almost all case-control and cohort studies that assessed the relationship (Grodstein et al., 1999), and recently the influence of post-menopausal estrogens on colorectal cancer risk has been demonstrated in a randomized trial setting (Rossouw et al., 2002). This benefit could reflect a direct benefit of estrogens on estrogen receptors or estrogen may act through indirect mechanisms. For example, post-menopausal hormone use substantially lowers circulating IGF-1 levels (Posaci et al., 2001). One study found that women were less likely to have microsatellite instability in tumors (MSI+) associated with a young age, but post-menopausal women experienced an increased risk of MSI+ tumors only if they were not on estrogen replacement therapy (Slattery et al., 2001). These results suggest a benefit of estrogens on MSI+ tumors. Although estrogens are likely to lower colorectal cancer risk, the public health and clinical implications must account for other potential benefits and risks of hormonal replacement, as well as alternative means of preventing colorectal cancer. Screening Methods and Recommended Frequency The early detection and removal of colorectal adenomas reduces risk of colorectal cancer (Jessup et al., 1997; Markowitz and Winawer, 1997), and the early diagnosis and treatment of colorectal cancer likely increase survival. Several approaches to early detection tests are available. The fecal occult blood test (FOBT) involves testing for the presence of blood in the stool, which may indicate either a colorectal adenoma or cancer. The specificity of this test is relatively low because many factors can produce false-positive results. In addition, many adenomas do not bleed, reducing the sensitivity of this test. Sigmoidoscopy involves cleansing the descending and sigmoid colon and directly visualizing the lower half or third of the colorectum using a 60-cm endoscope. Colonscopy involves cleansing and directly visualizing the entire colon and rectum using a colonoscope. Sigmoidoscopies are more feasible to use than colonoscopy, but do not evaluate the proximal regions of the large bowel. An additional option is the double-contrast barium enema (DCBE), which involves barium and air to visualize the colorectal mucosal profile using X-rays. Colorectal cancer screening is widely recommended for adults beginning at age 50 by major professional medical societies and organizations that issue guidelines. These recommendations are based on the results of several randomized controlled trials that show reductions in colorectal cancer mortality among individuals who undergo periodic screening with FOBT and on case-control studies that indicate a screening benefit of flexible sigmoidoscopy and colonoscopy (Byes et al., 1997). The American Cancer Society (ACS) recommendations for colorectal screening encompass three risk level groups (Byers et al., 1997). The guidelines for individuals at average risk and who are older than 50 years include either FOBT annually and sigmoidoscopy every 5 years, or colonoscopy every 10 years, or DCBE every 5–10 years, along with a digital rectal examination at the time of endoscopies or DCBE. Individuals with a family history of FAP or HNPPC, or who have inflammatory bowel disease, including ulcerative colitis and Crohn disease, are considered to be at high risk. Individuals at moderate risk are those with a previous small or large adenoma, or who have a family history of colorectal adenoma or cancer in first-degree relatives. Recommendations for high and moderate risk individuals have been tailored to their expected risk of colorectal cancer, and vary in age at which to begin screening, frequency of screening, and screening methods used (Byers et al., 1997). Patients who have undergone resection for colorectal cancer should have either a colonoscopy or

824

PART IV: CANCER BY TISSUE OF ORIGIN

DCBE within 1 year of resection, and if normal in 3 years, and if normal again, then in 5 years.

FUTURE DIRECTIONS Many characteristics of colorectal cancer make it prime for future research. In principle, this malignancy is largely avoidable either through primary or secondary prevention. There are a number of important leads regarding dietary and lifestyle patterns. Some of these factors appear to converge on hormonal pathways, including insulin and the IGF axis. Estrogens also appear to be important. Because influences on some of these pathways, such as insulin resistance, are likely to have the largest impact on overall health, research on modifiable factors centered on these pathways should be vigorously pursued. For example, a better understanding on how nutritional factors influence insulin and the IGF axis may shed some insights into how nutritional factors, particularly energy intake and macronutrients, influence colorectal carcinogenesis. Secondly, the role of micronutrients, such as calcium and folate, must be better understood. Although their overall influence may possibly be not as great as that of macronutrients and energy balance, the intake of micronutrients is more easily altered. Micronutrients can also be more readily studied in the context of randomized, double-blind clinical trials and toxicity is likely to be relatively low. Other agents, such as NSAIDs, also offer promise but their toxicities must be carefully evaluated. A third area of emphasis for research should be the incorporation of genetic susceptibility into studies. In identifying relevant, highly prevalent, low penetrance genes, the most likely benefit will be in enhancing our understanding of potentially modifiable environmental factors rather than in defining high-risk individuals who require enhanced screening. For example, the identification of consistent interactions between xenobiotic metabolizing genes and heterocyclic amines would strengthen our conclusions about the potential role of heterocyclic amines. A good example has been for the interaction between MTHFR and ALDH2 polymorphisms and folate and alcohol intakes, which add support to a causal role of folate and alcohol. Finally, the role of all potential factors must be considered in light of the different molecular pathways for colorectal cancer. If we can subtype tumors type by molecular pathway, we may better understand mechanisms that are impacted by environmental factors; in addition, we might attain stronger more conclusive results from epidemiologic studies than when all colorectal cancers, with heterogeneous etiologies, are pooled together for analysis. Our understanding of the epidemiology of colorectal cancer must grow alongside our increasing knowledge of molecular mechanisms. References Albanes D, Jones DY, Schatzkin A, et al. 1998. Adult stature and risk of cancer. Cancer Res 48:1658–1662. Alberts DS, Martinez ME, Roe DJ, et al. 2000. Lack of effect of a high-fiber cereal supplement on the recurrence of colorectal adenomas. Phoenix Colon Cancer Prevention Physicians’ Network. N Engl J Med 342:1156–1162. Alexandrov K, Rojas M, Kadlubar FF, et al. 1996. Evidence of antibenzo[a]pyrene diolepoxide-DNA adduct formation in human colon mucosa. Carcinogenesis 17:2081–2083. Armstrong B, Doll R. 1975. Environmental factors and cancer incidence and mortality in different countries, with special reference to dietary practices. Int J Cancer 15:617–631. Bagnardi V, Blangiardo M, Vecchia CL, et al. 2001. A meta-analysis of alcohol drinking and cancer risk. Br J Cancer 85(11):1700–1705. Barak AJ, Beckenhauer HC, Tuma DJ, et al. 1987. Effects of prolonged ethanol feeding on methionine metabolism in rat liver. Biochem Cell Biology 65:230–233. Bari D, Gridley G, Ron E, et al. 2002. Acromegaly and cancer risk: A cohort study in Sweden and Denmark. Cancer Causes Control 13:395–400. Barnes CJ, Lee M. 1998. Chemoprevention of spontaneous intestinal adenomas in the adenomatous polyposis coli Min mouse model with aspirin. Gastroenterology 114:873–877. Baron JA, Beach M, Mandel JS, et al. 1999. Calcium supplements for the prevention of colorectal adenomas. N Engl J Med 340:101–107.

Baron JA, Cole BF, Sandler RS, et al. 2003. A randomized trial of aspirin to prevent colorectal cancer. N Engl J Med 348:891–899. Baron JA, Sandler RS, Haile RW, et al. 1998. Folate intake, alcohol consumption, cigarette smoking, and risk of colorectal adenomas. J Natl Cancer Inst 90:57–62. Barzilay J, Heatley G, Cushing G. 1991. Benign and malignant tumors in patients with acromegaly. Arch Intern Med 151:1629–1632. Bayerdorffer E, Mannes GA, Richter WO, et al. 1993a. Decreased high-density lipoprotein cholesterol and increased low-density cholesterol levels in patients with colorectal adenomas. Ann Inter Med 118:481–487. Bayerdorffer E, Mannes GA, Richter WO, et al. 1993b. Increased serum deoxycholic acid levels in men with colorectal adenomas. Gastroenterology 104:141–151. Benito E, Cabeza E, Moreno V, et al. 1993. Diet and colorectal adenomas: A case-control study in Majorca. Int J Cancer 55:213–219. Benito E, Stiggelbout A, Bosch FX, et al. 1991. Nutritional factors in colorectal cancer risk: A case-control study in Majorca. Int J Cancer 49: 161–167. Bingham SA, Day NE, Luben R, et al. 2003. Dietary fibre in food and protection against colorectal cancer in the European Prospective Investigation into Cancer and Nutrition (EPIC): An observational study. Lancet 361(9368):1496–1501. Bird CL, Ingles SA, Frankl HD, et al. 1996. Serum lipids and adenomas of the left colon and rectum. Cancer Epidemiol Biomarkers Prev 5:607– 612. Bird CL, Swendseid ME, Witte JS, et al. 1995. Red cell and plasma folate, folate consumption, and the risk of colorectal adenomatous polyps. Cancer Epidemiol Biomarkers Prev 4:709–714. Bisgaard ML, Fenger K, Bulow S, et al. 1994. Familial adenomatous polyposis (FAP): Frequency, penetrance, and mutation rate. Hum Mutat 3:121–125. Bjelke E, 1980. Epidemiology of colorectal cancer, with emphasis on diet. In: Davis W, Harrup KR, Stathopoulos G, eds. Human Cancer. Its Characterization and Treatment. Amsterdam, Congress Series No. 484: Exerpta Medica, Int. pp. 158–174. Bonithon-Kopp C, Kronborg O, Giacosa A, et al. 2000. Calcium and fibre supplementation in prevention of colorectal adenoma recurrence: A randomized intervention trial. Lancet 356:1300–1306. Bostick RM, Potter JD, Kushi LH, et al. 1994. Sugar, meat, and fat intake, and non-dietary risk factors for colon cancer incidence in Iowa women (United States). Cancer Causes Control 5:38–52. Boutron-Ruault M-C, Senesse P, Faivre J, et al. 1996. Folate and alcohol intakes: Related or independent roles in the adenoma-carcinoma sequence? Nutr Cancer 26:337–346. Breivik J, Lothe RA, Meling GI, et al. 1997. Different genetic pathways to proximal and distal colorectal cancer influenced by sex-related factors. Int J Cancer 74:664–669. Brentnall TA, Crispin DA, Rabinovitch PS, et al. 1994. Mutations in the p53 gene: An early marker of neoplastic progression in ulcerative colitis. Gastroenterology 107:369–378. Brockton N, Little J, Sharp L, et al. 2000. N-acetyltransferase polymorphisms and colorectal cancer: A HuGE review. Am J Epidemiol 151:846–861. Brunner J, Johnson C, Zafar S, et al. 1990. Colon cancer and polyps in acromegaly: Increased risk associated with family history of colon cancer. Clin Endocrinol (Oxf) 32:65–71. Buermeyer AB, Deschenes SM, Baker SM, et al. 1999. Mammalian DNA mismatch repair. Annu Rev Genet 33:533–564. Bufill JA. 1990. Colorectal cancer: Evidence for distinct genetic categories based on proximal or distal tumor location. Ann Intern Med 113:779–788. Burkitt DP. 1971. Epidemiology of cancer of the colon and rectum. Cancer 28:3–13. Byers T, Levin B, Rothenberger JD, et al. 1997. American Cancer Society guidelines for screening and surveillance for early detection of colorectal polyps and cancer: Update 1997. CA Cancer J Clin 47:154–167. Cats A, Dullaart R, Kleibeuker J, et al. 1996. Increased epithelial cell proliferation in the colon of patients with acromegaly. Cancer Res 56:523–526. Chao A, Thun MJ, Jacobs EJ, et al. 2000. Cigarette smoking and colorectal cancer mortality in the Cancer Prevention Study II. J Natl Cancer Inst 92:1888–1896. Chen J, Giovannucci E, Hankinson SE, et al. 1998. A prospective study of methylenetetrahydrofolate reductase and methionine synthase gene polymorphisms, and risk of colorectal adenoma. Carcinogenesis 19: 2129–2132. Chen J, Giovannucci E, Kelsey K, et al. 1996. A methylenetetrahydrofolate reductase polymorphism and the risk of colorectal cancer. Cancer Res 56:4862–4864. Cheung NW, Boyages SC. 1997. Increased incidence of neoplasia in females with acromegaly. Clin Endocrinol (Oxf) 47:323–327.

Cancers of the Colon and Rectum Cho E, Smith-Warner SA, Hunter DJ. 2002. Alcohol and colorectal cancer: A pooled analysis of cohort studies [abstract]. Am J Epidemiol 155(11):S106. Chute CG, Willett WC, Colditz GA, et al. 1991a. A prospective study of body mass, height, and smoking on the risk of colorectal cancer in women. Cancer Causes Control 2:117–124. Chute CG, Willett WC, Colditz GA, et al. 1991b. A prospective study of reproductive history and exogenous estrogens on the risk of colorectal cancer in women. Epidemiology 2:201–207. Chyou P-H, Nomura AMY, Stemmermann GN. 1996. A prospective study of colon and rectal cancer among Hawaii Japanese men. Ann Epidemiol 6:276–282. Clark LC, Combs GF, Jr., Turnbull BW, et al. 1996. Effects of selenium supplementation for cancer prevention in patients with carcinoma of the skin. A randomized controlled trial. Nutritional Prevention of Cancer Study Group. JAMA 276:1957–1963. Colditz G, Cannuscio C, Frazier A. 1997. Physical activity and reduced risk of colon cancer: implications for prevention. Cancer Causes Control 8:649–667. Cotton S, Sharp L, Little J. 1996. The adenoma-carcinoma sequence and prospects for the prevention of colorectal neoplasia. Crit Rev Oncog 7:293–342. Cotton SC, Sharp L, Little J, et al. 2000. Glutathione S-transferase polymorphisms and colorectal cancer: A HuGE review [comment]. Am J Epidemiol 151:7–32. de Jong MM, Nolte IM, te Meerman GJ, et al. 2002. Low-penetrance genes and their involvement in colorectal cancer susceptibility. Cancer Epidemiol Biomarkers Prev 11:1332–1352. Debinski HS, Love S, Spigelman AD, et al. 1996. Colorectal polyp counts and cancer risk in familial adenomatous polyposis. Gastroenterology 110: 1028–1030. DeFronzo RA, Bonadonna RC, Ferrannini E. 1992. Pathogenesis of NIDDM. A balanced overview. Diabetes Care 15:318–368. Delattre O, Olschwang S, Law DJ, et al. 1989. Multiple genetic alterations in distal and proximal colorectal cancer. Lancet 2(8659):353–356. Doll R, Gray R, Hafner B, et al. 1980. Mortality in relation to smoking: 22 years’ observations on female British doctors. Br Med J 280:967–971. Doll R, Peto R. 1976. Mortality in relation to smoking: 20 years’ observations on male British doctors. Br Med J 2:1525–1536. Doll R, Peto R, Wheatley K, et al. 1994. Mortality in relation to smoking: 40 years’ observations on male British doctors. Br Med J 309:901–911. Dornschneider G, Izbicki JR, Wilker DK, et al. 1990. The effects of sex steroids on colon carcinogenesis. Anticancer Drugs 1:15–21. Dowse GK, Zimmet PZ, Gareeboo H, et al. 1991. Abdominal obesity and physical inactivity as risk factors for NIDDM and impaired glucose tolerance in Indian, Creole, and Chinese Mauritians. Diabetes Care 14:271–282. Duff EK, Clarke AR. 1998. Smad4 (DPC4)—a potent tumour suppressor? Br J Cancer 78:1615–1619. Earnest DL, Hixon LJ, Alberts DS. 1992. Piroxicam and other cyclooxygenase inhibitors: Potential for cancer chemoprevention. J Cell Biochem Suppl 161:156–166. Earnest DL, Holubec H, Wali RK, et al. 1994. Chemoprevention of azoxymethane-induced colonic carcinogenesis by supplemental dietary ursodeoxycholic acid. Cancer Res 54:5071–5074. Faivre J, Bonithon-Kopp C. 1999. A randomized trial of calcium and fiber supplementation in the prevention of recurrence of colorectal adenomas. American Gastroenterological Association, Orlando, FL. Fearnhead NS, Britton MP, Bodmer WF. 2001. The ABC of APC. Hum Mol Genet 10:721–733. Fearon ER, Vogelstein B. 1990. A genetic model for colorectal tumorigenesis. Cell 61:759–767. Federowski T, Salen G, Tint GS, et al. 1979. Transformation of chenodeoxycholic acid and ursodeoxycholic acid by human intestinal bacteria. Gastroenterology 77:1068–1073. Ferraroni M, La Vecchia C, D’Avanzo B, et al. 1994. Selected micronutrient intake and the risk of colorectal cancer. Br J Cancer 70:1150–1155. Finkelstein JD, Cello JP, Kyle WE. 1974. Ethanol-induced changes in methionine metabolism in rat liver. Biochem Biophys Res Commun 61:525–531. Flood DM, Weiss NS, Cook LS, et al. 2000. Colorectal cancer incidence in Asian migrants to the United States and their descendants. Cancer Causes Control 11:403–411. Fogt F, Brien T, Brown CA, et al. 2002. Genetic alterations in serrated adenomas: comparison to conventional adenomas and hyperplastic polyps. Hum Pathol 33:87–91. Ford ES. 1999. Body mass index and colon cancer in a national sample of adult US men and women. Am J Epidemiol 150:390–399.

825

Freedman AN, Michalek AM, Weiss HA, et al. 1998. Aspirin use and p53 expression in colorectal cancer. Cancer Detect Prev 22:213–218. Freudenheim JL, Graham S, Marshall JR, et al. 1991. Folate intake and carcinogenesis of the colon and rectum. Int J Epidemiol 20:368–374. Frosst P, Blom HJ, Milos R, et al. 1995. A candidate genetic risk factor for vascular disease: A common mutation in methylenetetrahydrofolate reductase [letter]. Nature Genetics 10:111–113. Fuchs CS, Colditz GA, Stampfer MJ, et al. 1999. Dietary fiber and the risk of colorectal cancer and adenoma in women. New Engl J Med 340:169–176. Gann PH, Manson JE, Glynn RJ, et al. 1993. Low-dose aspirin and incidence of colorectal tumors in a randomized trial. J Natl Cancer Inst 85: 1220–1224. Gapstur SM, Potter JD, Folsom AR. 1994. Alcohol consumption and colon and rectal cancer in postmenopausal women. Int J Epidemiol 23:50–57. Garland C, Shekelle RB, Barrett-Conner E, et al. 1985. Dietary vitamin D and calcium and risk of colorectal cancer: A 19-year prospective study in men. Lancet 1:307–309. Gerhardsson de Verdier M, Hagman U, Peters RK, et al. 1991. Meat, cooking methods and colorectal cancer: A case-referent study in Stockholm. Int J Cancer 49:520–525. Gerhardsson M, Floderus B, Norell SE. 1988. Physical activity and colon cancer risk. Int J Epidemiol 17:743–746. Gertig DM, Hunter DJ. 1998. Genes and environment in the etiology of colorectal cancer. Semin Cancer Biol 8:285–298. Ghadirian P, Maisonneuve P, Perret C, et al. 1998. Epidemiology of sociodemographic characteristics, lifestyle, medical history, and colon cancer: A case-control study among French Canadians in Montreal. Cancer Detect Prev 22:396–404. Giovannucci E. 1995. Insulin and colon cancer. Cancer Causes Control 6:164–179. Giovannucci E. 2001a. Insulin, insulin-like growth factors and colon cancer: A review of the evidence. J Nutr 131:3109S–3120S. Giovannucci E. 2001b. An updated review of the epidemiological evidence that cigarette smoking increases risk of colorectal cancer. Cancer Epidemiol Biomarkers Prev 10:725–731. Giovannucci E, Ascherio A, Rimm EB, et al. 1995b. Physical activity, obesity, and risk for colon cancer and adenoma in men. Ann Intern Med 122:327–334. Giovannucci E, Chen J, Smith-Warner SA, et al. 2003. Methylenetetrahydrofolate reductase, alcohol dehydrogenase, diet and risk of colorectal adenomas. Cancer Epidemiol Biomarkers Prev 12:970–999. Giovannucci E, Colditz GA, Stampfer MJ. 1993a. A meta-analysis of cholecystectomy and risk of colorectal cancer. Gastroenterology 105: 130–141. Giovannucci E, Colditz GA, Stampfer MJ, et al. 1994a. A prospective study of cigarette smoking and risk of colorectal adenoma and colorectal cancer in U.S. women. J Natl Cancer Inst 86:192–199. Giovannucci E, Colditz GA, Stampfer MJ, et al. 1996. Physical activity, obesity, and risk of colorectal adenoma in women (United States). Cancer Causes Control 7:253–263. Giovannucci E, Egan KM, Hunter DJ, et al. 1995c. Aspirin and the risk of colorectal cancer in women. N Engl J Med 333:609–614. Giovannucci E, Pollak MN, Platz EA, et al. 2000. A prospective study of plasma insulin-like growth factor-I and binding protein-3 and risk of colorectal neoplasia in women. Cancer Epidemiol Biomarkers Prev 9:345–349. Giovannucci E, Rimm EB, Ascherio A, et al. 1995d. Alcohol, low-methioninelow-folate diets, and risk of colon cancer in men. J Natl Cancer Inst 87:265–273. Giovannucci E, Rimm EB, Stampfer MJ, et al. 1994b. A prospective study of cigarette smoking and risk of colorectal adenoma and colorectal cancer in U.S. men. J Natl Cancer Inst 86:183–191. Giovannucci E, Rimm EB, Stampfer MJ, et al. 1994c. Aspirin use and the risk for colorectal cancer and adenoma in male health professionals. Ann Intern Med 121:241–246. Giovannucci E, Rimm EB, Stampfer MJ, et al. 1994d. Intake of fat, meat, and fiber in relation to risk of colon cancer in men. Cancer Res 54:2390–2397. Giovannucci E, Stampfer MJ, Colditz GA, et al. 1998. Multivitamin use, folate, and colon cancer in women in the Nurses’ Health Study. Ann Intern Med 129:517–524. Giovannucci E, Stampfer MJ, Colditz GA, et al. 1993b. Folate, methionine, and alcohol intake and risk of colorectal adenoma. J Natl Cancer Inst 85:875–884. Giovannucci E, Stampfer MJ, Colditz GA, et al. 1992. Relationship of diet to risk of colorectal adenoma in men. J Natl Cancer Inst 84:91–98. Glynn SA, Albanes D, Pietinen P, et al. 1996. Colorectal cancer and folate status: A nested case-control study among male smokers. Cancer Epidemiol Biomarkers Prev 5:487–494.

826

PART IV: CANCER BY TISSUE OF ORIGIN

Goldbohm RA, van den Brandt PA, van’t Veer P, et al. 1994. A prospective cohort study on the relation between meat consumption and the risk of colon cancer. Cancer Res 54:718–723. Greenberg ER, Baron JA, Tosteson TD, et al. 1994. A clinical trial of antioxidant vitamins to prevent colorectal adenoma. N Engl J Med 331:141–147. Greenblatt MS, Bennett WP, Hollstein M, et al. 1994. Mutations in the p53 tumor suppressor gene: clues to cancer etiology and molecular pathogenesis. Cancer Res 54:4855–4878. Grodstein F, Newcomb PA, Stampfer MJ. 1999. Postmenopausal hormone therapy and the risk of colorectal cancer: A review and meta-analysis. Am J Med 106:574–582. Grubben MJ, Nagengast FM, Katan MB, et al. 2001. The glutathione biotransformation system and colorectal cancer risk in humans. Scand J Gastroenterol Suppl 234:68–76. Haenszel W, Kurihara M. 1968. Studies of Japanese migrants. I. Mortality from cancer and other diseases among Japanese in the United States. J Natl Cancer Inst 40:43–68. Hamilton SR, Liu B, Parsons RE, et al. 1995. The molecular basis of Turcot’s syndrome. N Engl J Med 332:839–847. Hammond CE, Horn D. 1966. Smoking in relation to the death rates of one million men and women. In: Haenszel W, ed. National Cancer Institute Monograph 19. Bethesda, MD: National Cancer Institute, pp. 127–204. Hammond EC, Horn D. 1958. Smoking and death rates: Report on forty-four months of followup of 187,783 men. II. Death rates by causality. JAMA 166:1294–1308. Haubrich WS. 1995. Anatomy of the colon. In: Haubrich WS, Schaffner F, Berk J, eds. Bockus Gastroenterology. 5th ed. Philadelphia: Saunders, pp. 1744–1772. He TC, Sparks AB, Rago C, et al. 1998. Identification of c-MYC as a target of the APC pathway. Science 281(5382):1509–1512. Hebert PR, Ajani U, Cook NR, et al. 1997. Adult height and incidence of cancer in male physicians (United States). Cancer Causes Control 8:591–597. Hein DW. 2000. N-Acetyltransferase genetics and their role in predisposition to aromatic and heterocyclic amine-induced carcinogenesis. Toxicol Lett 112–113:349–356. Heineman EF, Zahm SH, McLaughlin JK, et al. 1994. Increased risk of colorectal cancer among smokers: Results of a 26-year follow-up of US veterans and a review. Int J Cancer 59:728–738. Herman JG, Umar A, Polyak K, et al. 1998. Incidence and functional consequences of hMLH1 promoter hypermethylation in colorectal carcinoma. Proc Natl Acad Sci U S A 95:6870–6875. Hirayama T. 1986. A large-scale study on cancer risks by diet—with special reference to the risk reducing effects of green-yellow vegetable consumption. In: Hayashi Y, Magao M, Sugimura T, et al., eds. Diet, Nutrition, and Cancer. Tokyo: Japan Scientific Societies Press, pp. 41–53. Hoffmann D, Hoffmann I. 1997. The changing cigarette, 1950–1995. J Toxicol Environ Health 50:307–364. Houlston R, Crabtree M, Phillips R, et al. 2001. Explaining differences in the severity of familial adenomatous polyposis and the search for modifier genes. Gut 48:1–5. Howe GR. Meeting Presentation. Advances in the biology and therapy of colorectal cancer. 1993. M.D. Anderson Cancer Center, Houston, Texas. Howe GR, Benito E, Castelleto R, et al. 1992. Dietary intake of fiber and decreased risk of cancers of the colon and rectum: Evidence from the combined analysis of 13 case-control studies. J Natl Cancer Inst 84:1887–1896. Hsing AW, McLaughlin JK, Chow W-H, et al. 1998. Risk factors for colorectal cancer in a prospective study among U.S. white men. Int J Cancer 77:549–553. Hu FB, Manson JE, Liu S, et al. 1999. Prospective study of adult onset diabetes mellitus (Type 2) and risk of colorectal cancer in women. J Natl Cancer Inst 91:542–547. iARC Working Group on the Evaluation of the Carcinogenic Risk of Chemicals to Humans. 1986. Tobacco smoking. Lyon, France: World Health Organization, International Agency for Research on Cancer, 397 p. Iino H, Jass JR, Simms LA, et al. 1999. DNA microsatellite instability in hyperplasic polyps, serrated adenomas, and mixed polyps: A mild mutator pathway for colorectal cancer? J Clin Pathol 52:5–9. Inoue M, Tajima K, Hirose K, et al. 1995. Subsite-specific risk factors for colorectal cancer: A hospital-based case-control study in Japan. Cancer Causes Control 6:14–22. Ituarte E, Petrini J, Hershman J. 1984. Acromegaly and colon cancer. Ann Intern Med 101:627–628. Jass JR, Do KA, Simms LA, et al. 1998. Morphology of sporadic colorectal cancer with DNA replication errors. Gut 42:673–679. Jemal A, Murray T, Samuels A, et al. 2003. Cancer statistics, 2003. CA Cancer J Clin 53:5–26.

Jenkins P, Fairclough P, Richards T, et al. 1997. Acromegaly, colonic polyps and carcinoma. Clin Endocrinol (Oxf) 47:17–22. Jenkins PJ. 2000. Acromegaly and Colon Cancer. Growth Horm IFG Res 10 Suppl 1A:S35–36. Jessup JM, Menck MR, Fremgen A, et al. 1997. Diagnosing colorectal carcinoma: Clinical and molecular approaches. CA Cancer J Clin 47: 70–92. Johns LE, Houlston RS. 2001. A systematic review and meta-analysis of familial colorectal cancer risk. Am J Gastroenterol 96:2992–3003. Juul A, Bang P, Hertel NT, et al. 1994. Serum insulin-like growth factor-I in 1030 healthy children, adolescents, and adults: Relation to age, sex, stage of puberty, testicular size, and body mass index. J Clin Endocrinol Metab 78:744–752. Kaaks R, Toniolo P, Akhmedkhanov A, et al. 2000. Serum C-peptide, insulinlike growth factor (IGF)-1, IGF-binding proteins, and colorectal cancer risk in women. J Natl Cancer Inst 92(19):1592–1600. Kahn HA. 1966. The Dorn study of smoking and mortality among US Veterans on 8 years of observation. In: Haenszel W, ed. National Cancer Institute Monograph 19. Bethesda, MD: National Cancer Institute, pp. 1–125. Kanazawa K, Konishi F, Mitsuoka T, et al. 1996. Factors influencing the development of sigmoid colon cancer. Bacteriologic and biochemical studies. Cancer 77(8 Suppl):1701–1706. Kang S-S, Wong PWK, Malinow MR. 1992. Hyperhomocyst(e)inemia as a risk factor for occlusive vascular disease. Annu Rev Nutr 12:279–298. Kato I, Dnistrian AM, Schwartz M, et al. 1999. Serum folate, homocysteine and colorectal cancer risk in women: A nested case-control study. Br J Cancer 79:1917–1921. Kato I, Tominaga S, Matsuura A, et al. 1990. A comparative case-control study of colorectal cancer and adenoma. Jpn J Cancer Res 81:1101–1108. Kawamori T, Rao CV, Seibert K, et al. 1998. Chemopreventive activity of celecoxib, a specific cyclooxygenase-2 inhibitor, against colon carcinogenesis. Cancer Res 58:409–412. Kim DH, Smith-Warner SA, Hunter DJ. 2001. Pooled analysis of prospective cohort studies on folate and colorectal cancer. Pooling Project of Diet and Cancer Investigators [abstract]. Am J Epidemiol Suppl 153(11):S118. King H, Li JY, Locke FB, et al. 1985. Patterns of site-specific displacement in cancer mortality among migrants: The Chinese in the United States. Am J Public Health 75:237–242. Klatsky AL, Armstrong MA, Friedman GD, et al. 1988. The relations of alcoholic beverage use to colon and rectal cancer. Am J Epidemiol 128: 1007–1015. Klein I, Parveen G, Gavaler J, et al. 1982. Colonic polyps in patients with acromegaly. Ann Intern Med 97:27–30. Knekt P, Hakulinen T, Leino A, et al. 2000. Serum albumin and colorectal cancer risk. Eur J Clin Nutr 54:460–462. Kono S, Honjo S, Todoroki I, et al. 1998. Glucose intolerance and adenomas of the sigmoid colon in Japanese men (Japan). Cancer Causes Control 9:441–446. Kono S, Shinchi K, Ikeda N, et al. 1991. Physical activity, dietary habits and adenomatous polyps of the sigmoid colon: A study of self-defense officials in Japan. J Clin Epidemiol 44(11):1255–1261. Kuismanen SA, Holmberg MT, Salovaara R, et al. 2000. Genetic and epigenetic modification of MLH1 accounts for a major share of microsatelliteunstable colorectal cancers. Am J Pathol 156:1773–1779. Kune GA, Kune S, Vitetta L, et al. 1992a. Smoking and colorectal cancer risk: data from the Melbourne Colorectal Cancer Study and brief review of literature. Int J Cancer 50:369–372. Kune GA, Kune S, Watson LF. 1988. Colorectal cancer risk, chronic illnesses, operations, and medications: Case control results from the Melbourne Colorectal Cancer Study. Cancer Res 48:4399–4404. Kune GA, Vitetta L. 1992b. Alcohol consumption and the etiology of colorectal cancer: A review of the scientific evidence from 1957 to 1991. Nutr Cancer 18:97–111. La Vecchia C, D’Avanzo B, Negri E, et al. 1991. History of selected diseases and the risk of colorectal cancer. Eur J Cancer 27:582–586. La Vecchia C, Negri E, Franceschi S, et al. 1997. Aspirin and colorectal cancer. Br J Cancer 76:675–677. Lampe JW, Fredstrom SB, Slavin JL, et al. 1993. Sex differences in colonic function: A randomised trial. Gut 34:531–536. Landis SH, Murray T, Bolden S, et al. 1999. Cancer Statistics, 1999. CA Cancer J Clin 49:8–31. Lapre JA, Van der Meer R. 1992. Diet induced increase in colonic bile acids stimulates lytic activity in fecal water and proliferation of colonic cells. Carcinogenesis 13:41–44. Le Marchand L, Donlon T, Hankin JH, et al. 2002. B-vitamin intake, metabolic genes, and colorectal cancer risk (United States). Cancer Causes Control 13(3):239–248.

Cancers of the Colon and Rectum Le Marchand L, Wilkens LR, Kolonel LN, et al. 1997. Associations of sedentary lifestyle, obesity, smoking, alcohol use, and diabetes with the risk of colorectal cancer. Cancer Res 57:4787–4794. Le Marchand L, Wilkins LR, Mi MP. 1992. Obesity in youth and middle age and risk of colorectal cancer in men. Cancer Causes Control 3: 349–354. Lee HP, Gourley L, Duffy SW, et al. 1989. Colorectal cancer and diet in an Asian population—A case-control study among Singapore Chinese. Int J Cancer 43:1007–1016. Lee IM, Paffenbarger RS, Jr., Hsieh CC. 1991. Physical activity and risk of developing colorectal cancer among college alumni. J Natl Cancer Inst 83:1324–1329. Lee RG. 1995. Benign tumors of the colon. In: Haubrich WS, Schaffner F, Berk J, eds. Bockus Gastroenterology. 5th ed. Philadelphia: Saunders, pp. 1715–1730. Leslie A, Carey FA, Pratt NR, et al. 2002. The colorectal adenoma-carcinoma sequence. Br J Surg 89:845–860. Lev R. 1990. Adenomatous Polyps of the Colon. New York: Springer-Verlag, 61 p. Levin B, Raijman I. 1995. Malignant tumors of the colon and rectum. In: Haubrich WS, Schaffner F, Berk J, eds. Bockus Gastroenterology. 5th ed. Philadelphia: Saunders, pp. 1744–1772. Levine AJ, Siegmund KD, Ervin CM, et al. 2000. The methylenetetrahydrofolate reductase 677C Æ T polymorphism and distal colorectal adenoma risk. Cancer Epidemiol Biomarkers Prev 9:657–663. Lew EA, Garfinkel L. 1979. Variations in mortality by weight among 750,000 men and women. J Chronic Dis 32:563–576. Lilienfeld AM, Levin ML, Kessler II. 1972a. Mortality among the foreign born and their countries of origin. Cancer in the United States. Cambridge, MA: Harvard University Press, pp. 233–278. Lilienfeld AM, Levin ML, Kessler II. 1972b. Summary of mortality data for selected sites. Cancer in the United States. Cambridge, MA: Harvard University Press, pp. 279–297. Lindgarde F, Saltin B. 1981. Daily physical activity, work capacity and glucose tolerance in lean and obese normoglycaemic middle-aged men. Diabetologia 20:134–138. Lipkin M, Newmark H. 1985. Effect of added dietary calcium on colonic epithelial-cell proliferation in subjects at high risk for familial colonic cancer. N Engl J Med 313:1381–1384. Little J, Logan RF, Hawtin PG, et al. 1993. Colorectal adenomas and energy intake, body size and physical activity: A case-control study of subjects participating in the Nottingham faecal occult blood screening programme. Br J Cancer 67(1):172–176. Lleonart ME, Garcia-Foncillas J, Sanchez-Prieto R, et al. 1998. Microsatellite instability and p53 mutations in sporadic right and left colon carcinoma: different clinical and molecular implications. Cancer 83:889–895. Locke FB, King H. 1980. Cancer mortality risk among Japanese in the United States. J Natl Cancer Inst 65:1149–1156. Longacre TA, Fenoglio-Preiser CM. 1990. Mixed hyperplastic adenomatous polyps/serrated adenomas. A distinct form of colorectal neoplasia. Am J Surg Pathol 14:524–537. Lynch HT, de la Chapelle A. 1999. Genetic susceptibility to non-polyposis colorectal cancer. J Med Genet 36:801-818. Lynch HT, Smyrk T. 1996. Hereditary nonpolyposis colorectal cancer (Lynch syndrome). An updated review. Cancer 78:1149–1167. Lynch HT, Smyrk T. 1998. An update on Lynch syndrome. Curr Opin Oncol 10:349–356. Lynch PM. 1995. Polyposis syndromes. In: Haubrich WS, Schaffner F, Berk J, eds. Bockus Gastroenterology. 5th ed. Philadelphia: Saunders, pp. 1731–1743. Ma J, Giovannucci E, Pollak M, et al. 2001. Milk intake, circulating levels of insulin-like growth factor-I, and risk of colorectal cancer in men. J Natl Cancer Inst 93:1330–1336. Ma J, Pollak MN, Giovannucci E, et al. 1999. Prospective study of colorectal cancer risk in men and plasma levels of insulin-like growth factor (IGF)1 and IGF-binding protein-3. J Natl Cancer Inst 91:620–625. Ma J, Stampfer MJ, Giovannucci E, et al. 1997. Methylenetetrahydrofolate reductase polymorphism, dietary interactions, and risk of colorectal cancer. Cancer Res 57:1098–1102. Mackillop WJ, Zhang-Salomons J, Boyd CJ, et al. 2000. Associations between community income and cancer incidence in Canada and the United States. Cancer 89:901–912. MacLennan R, Macrae F, Bain C, et al. 1995. Randomized trial of intake of fat, fiber, and beta carotene to prevent colorectal adenomas. J Natl Cancer Inst 87:1760–1766. Manabe S, Tohyama K, Wada O, et al. 1991. Detection of a carcinogen, 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP), in cigarette smoke condensate. Carcinogenesis 12:1945–1947.

827

Manousos O, Souglakos J, Bosetti C, et al. 1999. IGF-I and IGF-II in relation to colorectal cancer. Int J Cancer 83:15–17. Manus B, Adang RP, Ambergen AW, et al. 1997. The risk factor profile of rectosigmoid adenomas: A prospective screening study of 665 patients in a clinical rehabilitation centre. Eur J Cancer Prev 6:38–43. Markowitz AJ, Winawer SJ. 1997. Management of colorectal polyps. CA Cancer J Clin 47:93–112. Markowitz S, Wang J, Myeroff L, et al. 1995. Inactivation of the type II TGFbeta receptor in colon cancer cells with microsatellite instability. Science 268(5215):1336–1338. Martinez ME, Giovannucci E, Spiegelman D, et al. 1997. Leisure-time physical activity, body size, and colon cancer in women. Nurses’ Health Study Research Group. J Natl Cancer Inst 89:948–955. Martinez ME, Giovannucci E, Spiegelman D, et al. 1996. Physical activity, body size, and colorectal cancer in women (abstract). Am J Epidemiol 143:S73. Martinez ME, Willett WC. 1998. Calcium, vitamin D, and colorectal cancer: A review of the epidemiologic evidence. Cancer Epidemiol Biomarkers Prev 7:163–168. Marugame T, Tsuji E, Inoue H, et al. 2000. Methylenetetrahydrofolate reductase polymorphism and risk of colorectal adenomas. Cancer Lett 151:181–186. May P, May E. 1999. Twenty years of p53 research: structural and functional aspects of the p53 protein. Oncogene 18:7621–7636. McCredie M, Williams S, Coates M. 1999. Cancer mortality in migrants from the British Isles and continental Europe to New South Wales, Australia, 1975–1995. Int J Cancer 83:179–185. McMichael AJ, Potter JD. 1985. Host factors in carcinogenesis: certain bileacid metabolic profiles that selectively increase the risk of proximal colon cancer. J Natl Cancer Inst 75:185–191. Meyer F, White E. 1993. Alcohol and nutrients in relation to colon cancer in middle-aged adults. Am J Epidemiol 138:225–226. Michels KB, Giovannucci E, Joshipura KJ, et al. 2000. Prospective study of fruit and vegetable consumption and incidence of colon and rectal cancers. J Natl Cancer Inst 92:1740–1752. Minchin RF, Kadlubar FF, Ilett KF. 1993. Role of acetylation in colorectal cancer. Mutat Res 290:35–42. Mitchell RJ, Farrington SM, Dunlop MG, et al. 2002. Mismatch repair genes hMLH1 and hMSH2 and colorectal cancer: A HuGE review. Am J Epidemiol 156:885–902. Morin PJ, Sparks AB, Korinek V, et al. 1997. Activation of beta-catenin-Tcf signaling in colon cancer by mutations in beta-catenin or APC. Science 275(5307):1787–1790. Morson BC. 1974. Evolution of cancer of the colon and rectum. Cancer 34(suppl):845–849. Morson BC. 1984. The evolution of colorectal carcinoma. Clin Radiol 35:425–431. Muir C, Waterhouse J, Mack T, et al. 1987. Cancer Incidence in Five Continents. Vol. V. Lyon: IARC Scientific Publications, no. 88. Murthy S, Flanigan A, Clearfield H. 2002. Colorectal cancer in inflammatory bowel disease: Molecular and clinical features. Gastroenterol Clin North Am 31:551–564. Muscat JE, Stellman SD, Wynder EL. 1994. Nonsteroidal antiinflammatory drugs and colorectal cancer. Cancer 74:1847–1854. Must A, Jacques PF, Dallal GE, et al. 1992. Long-term morbidity and mortality of overweight adolescents. A follow-up of the Harvard Growth Study of 1922 to 1935. N Engl J Med 327:1350–1355. Narisawa T, Sato M, Tani M, et al. 1981. Inhibition of development of methylnitrosourea-induced rat colon tumors by indomethacin treatment. Cancer Res 41:1945–1957. Neugut AI, Garbowski GC, Lee WC, et al. 1993. Dietary risk factors for the incedence and recurrence of colorectal adenomatous polyps: A casecontrol study. Ann Int Med 118:91–95. Neugut AI, Terry MB, Hocking G, et al. 1996. Leisure and occupational physical activity and risk of colorectal adenomatous polyps. Int J Cancer 68:744–748. Newcomb PA, Storer BE, Marcus PM. 1995. Cigarette smoking in relation to risk of large bowel cancer in women. Cancer Res 55: 4906–4909. Newmark HL, Wargovich MJ, Bruce WR. 1984. Colon cancer and dietary fat, phosphate, and calcium: A hypothesis. J Natl Cancer Inst 72:1323– 1325. Nilsen TI, Vatten LJ. 2001. Prospective study of colorectal cancer risk and physical activity, diabetes, blood glucose and BMI: Exploring the hyperinsulinaemia hypothesis. Br J Cancer 84(3):417–422. Norat T, Lukanova A, Ferrari P, et al. 2002. Meat consumption and colorectal cancer risk: Dose-response meta-analysis of epidemiological studies. Int J Cancer 98:241–256.

828

PART IV: CANCER BY TISSUE OF ORIGIN

Nyrén O, Bergström R, Nyström L, et al. 1996. Smoking and colorectal cancer: A 20-Year follow-up study of Swedish construction workers. J Natl Cancer Inst 88:1302–1307. Olschwang S, Hamelin R, Laurent-Puig P, et al. 1997. Alternative genetic pathways in colorectal carcinogenesis. Proc Natl Acad Sci U S A 94:12122–12127. Orme SM, McNally RJ, Cartwright RA, et al. 1998. Mortality and cancer incidence in acromegaly: A retrospective cohort study. United Kingdom Acromegaly Study Group. J Clin Endocrinol Metab 83:2730–2734. Oshima M, Dinchuk JE, Kargman SL, et al. 1996. Suppression of intestinal polyposis in Apc delta716 knockout mice by inhibition of cyclooxygenase 2 (COX-2). Cell 87:803–809. Otori K, Oda Y, Sugiyama K, et al. 1997. High frequency of K-ras mutations in human colorectal hyperplastic polyps. Gut 40:660–663. Paganelli GM, Biasco G, Brandi G, et al. 1992. Effect of vitamin A, C, and E supplementation on rectal cell proliferation in patients with colorectal adenomas. J Natl Cancer Inst 84:47–51. Parkin DM, Bray F, Ferlay J, et al. 2001. Estimating the world cancer burden: Globocan 2000. Int J Cancer 94:153–156. Parkin DM, Muir CS, Whelan SL, et al. 1992. Cancer Incedence in Five Continents. Vol. VI. Lyon: IARC Scientific Publications, no. 120. Parkin DM, Whelan SL, Ferlay J, et al. 1997. Cancer Incidence in Five Continents. Vol. VII. Lyon: IARC Scientific Publications, no. 143. Paspatis GA, Kalafatis E, Oros L, et al. 1995. Folate status and adenomatous colonic polyps. A colonoscopically controlled study. Dis Colon Rectum 38:64–68. Peleg II, Maibach HT, Brown SH, et al. 1994. Aspirin and nonsteroidal antiinflammatory drug use and the risk of subsequent colorectal cancer. Arch Intern Med 154:394–399. Peltomaki P. 2001. Deficient DNA mismatch repair: A Common etiologic factor for colon cancer. Hum Mol Genet 10:735–740. Pennisi E. 1998. Building a better aspirin. Science 280:1191–1192. Phillips RL, Snowdon DA. 1983. Association of meat and coffee use with cancers of the large bowel, breast, and prostate among Seventh-Day Adventists: Preliminary results. Cancer Res 43(suppl):2403S–2408S. Phillips RL, Snowdon DA. 1985. Dietary relationships with fatal colorectal cancer among Seventh-Day Adventists. J Natl Cancer Inst 74:307–317. Pierce JP, Fiore MC, Novotny TE, et al. 1989. Trends in cigarette smoking in the United States. Projections to the year 2000. JAMA 261:61–65. Pietinen P, Malila N, Virtanen M, et al. 1999. Diet and risk of colorectal cancer in a cohort of Finnish men. Cancer Causes Control 10:387–396. Pines A, Rozen P, Ron E, et al. 1985. Gastrointestinal tumors in acromegalic patients. Am J Gastroenterol 80:266–269. Pisani P, Mitton N. 2002. Cooking methods, metabolic polymorphisms and colorectal cancer. Eur J Cancer Prev 11:75–84. Platz EA, Willett WC, Colditz GA, et al. 2000. Proportion of colon cancer risk that might be preventable in a cohort of middle-aged US men. Cancer Causes Control 11:579–588. Podolsky DK. 2002. Inflammatory bowel disease. N Engl J Med 347:417– 429. Pollard M, Luckert PH. 1980. Indomethacin teatment of rats with dimethylhydrazine-induced intestinal tumors. Cancer Treat Rep 64:1323– 1327. Pollard M, Luckert PH. 1981a. Effect of indomethacin on intestinal tumors induced in rats by the acetate derivative of dimethylnitrosamine. Science 214:558–559. Pollard M, Luckert PH. 1981b. Treatment of chemically-induced intestinal cancers with indomethacin. Proc Soc Exp Biol Med 167:161–164. Ponz de Leon M, Sacchetti C, Sassatelli R, et al. 1990. Evidence for the existence of different types of large bowel tumor: Suggestions from the clinical data of a population-based registry. J Surg Oncol 44:35–43. Posaci C, Altunyurt S, Islekel H, et al. 2001. Effects of HRT on serum levels of IGF-I in postmenopausal women. Maturitas 40:69–74. Potter JD. 1999. Colorectal cancer: Molecules and populations [comment]. J Natl Cancer Inst 91:916–932. Rafter JJ, Child P, Anderson AM, et al. 1987. Cellular toxicity of fecal water depends on diet. Am J Clin Nutr 45:559–563. Rampino N, Yamamoto H, Ionov Y, et al. 1997. Somatic frameshift mutations in the BAX gene in colon cancers of the microsatellite mutator phenotype. Science 275(5302):967–969. Reddy BS, Rao CV, Seibert K. 1996. Evaluation of cyclooxygenase inhibitor for potential chemopreventive properties in colon carcinogenesis. Cancer Res 56:4566–4569. Regensteiner JG, Mayer EJ, Shetterly SM, et al. 1991. Relationship between habitual physical activity and insulin levels among nondiabetic men and women. San Luis Valley Diabetes Study. Diabetes Care 14:1066–1074. Renehan AG, Painter JE, Atkin WS, et al. 2001. High-risk colorectal adenomas and serum insulin-like growth factors. Br J Surg 88:107–113.

Ries LA, Wingo PA, Miller DS, et al. 2000. The annual report to the nation on the status of cancer, 1973–1997, with a special section on colorectal cancer. Cancer 88:2398–2424. Ritter M, Richter W, Schwandt P. 1987. Acromegaly and colon cancer. Ann Intern Med 106:636–637. Robsahm TE, Tretli S. 1999. Height, weight and gastrointestinal cancer: A follow-up study in Norway. Eur J Cancer Prev 8:105–113. Rogot E, Murray JL. 1980. Smoking and causes of death among US Veterans: 16 years of observation. Public Health Rep 95:213–322. Ron E, Gridley G, Hrubec Z, et al. 1991. Acromegaly and gastrointestinal cancer. Cancer 68:1673–1677. Rosenberg L, Palmer J, Zauber A, et al. 1991. A hypothesis: nonsteroidal antiinflammatory drugs reduce the incidence of large-bowel cancer. J Natl Cancer Inst 83:355–358. Rossouw JE, Anderson GL, Prentice RL, et al. 2002. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: Principal results from the Women’s Health Initiative randomized controlled trial. JAMA 288(3):321–333. Sandhu MS, White IR, McPherson K. 2001. Systematic review of the prospective cohort studies on meat consumption and colorectal cancer risk: A meta-analytic approach. Cancer Epidemiol Biomarkers Prev 10: 439–446. Sandler RS, Halabi S, Baron JA, et al. 2003. A randomized trial of aspirin to prevent colorectal adenomas in patients with previous colorectal cancer. N Engl J Med 348:883–890. Sandler RS, Lyles CM, Peipins LA, et al. 1993. Diet and risk of colorectal adenomas: Macronutrients, cholesterol and fiber. J Natl Cancer Inst 85:884–891. Sandler RS, Sandler DP, Comstock GW, et al. 1988. Cigarette smoking and the risk of colorectal cancer in women. J Natl Cancer Inst 80:1329– 1333. Schatzkin A, Lanza E, Corle D, et al. 2000. Lack of effect of a low-fat, highfiber diet on the recurrence of colorectal adenomas. Polyp Prevention Trial Study Group. N Engl J Med 342:1149–1155. Schoen RE, Tangen CM, Kuller LH, et al. 1999. Increased blood glucose and insulin, body size, and incident colorectal cancer. J Natl Cancer Inst 91:1147–1154. Schreinemachers D, Everson R. 1994. Aspirin use and lung, colon, and breast cancer incidence in a prospective study. Epidemiology 5:138–146. Shaw S, Jayatilleke E, Herbert V, et al. 1989. Cleavage of folates during ethanol metabolism: Role of acetaldehyde/xanthine oxidase-generated superoxide. Biochem J 257:277–280. Singh GK, Miller BA, Hankey BF. 2002a. Changing area socioeconomic patterns in U.S. cancer mortality, 1950–1998: Part II—lung and colorectal cancers. J Natl Cancer Inst 94:916–925. Singh GK, Miller BA, Hankey BF, et al. 2002b. Changing area socioeconomic patterns in U.S. cancer mortality, 1950–1998: Part I—all cancers among men. J Natl Cancer Inst 94:904–915. Slattery ML, Potter J, Caan B, et al. 1997a. Energy balance and colon cancer— beyond physical activity. Cancer Res 57:75–80. Slattery ML, Potter JD, Curtin K, et al. 2001. Estogens reduce and withdrawal of estrogens increase risk of microsatellite instability-positive colon cancer. Cancer Res 61:126–130. Slattery ML, Potter JD, Friedman GD, et al. 1997b. Tobacco use and colon cancer. Int J Cancer 70:259–264. Slattery ML, Potter JD, Samowitz W, et al. 1999. Methylenetetrahydrofolate reductase, diet, and risk of colon cancer. Cancer Epidemiol Biomarkers Prev 8:513–518. Slattery ML, Schaffer D, Edwards SL, et al. 1997c. Are dietary factors involved in DNA methylation associated with colon cancer? Nutr Cancer 28:52–62. Slattery ML, West DW, Robison LM, et al. 1990. Tobacco, alcohol, coffee, and caffeine as risk factors for colon cancer in a low-risk population. Epidemiology 1:141–145. Soong R, Powell B, Elsaleh H, et al. 2000. Prognostic significance of TP53 gene mutation in 995 cases of colorectal carcinoma. Influence of tumour site, stage, adjuvant chemotherapy and type of mutation. Eur J Cancer 36:2053–2060. Spirio L, Olschwang S, Groden J, et al. 1993. Alleles of the APC gene: An attenuated form of familial polyposis. Cell 75:951–957. Stadler J, Yeung KS, Furrer R, et al. 1988. Proliferative activity of rectal mucosa and soluble fecal bile acids in patients with normal colons and in patients with colonic polyps or cancer. Cancer Lett 38:315–320. Stemmermann GN, Nomura AM, Heilbrun LK. 1984. Dietary fat and the risk of colorectal cancer. Cancer Res 44:4633–4637. Stürmer T, Glynn RJ, Lee I-M, et al. 2000. Lifetime cigarette smoking and colorectal cancer incidence in the Physicians’ Health Study I. JNCI 92:1178–1181.

Cancers of the Colon and Rectum Su LJ, Arab L. 2001. Nutritional status of folate and colon cancer risk: Evidence from NHANES I epidemiologic follow-up study. Ann Epidemiol 11:65–72. Sugimura T. 1985. Carcinogenicity of mutagenic heterocyclic amines formed during the cooking process. Mutation Res 150:33–41. Sugimura T, Sato S. 1983. Mutagens-carcinogens in foods. Cancer Res 43:2415S–2421S. Suh O, Mettlin C, Petrelli NJ. 1993. Aspirin use, cancer, and polyps of the large bowel. Cancer 72:1171–1177. Terry P, Giovannucci E, Bergkvist L, et al. 2001. Body weight and colorectal cancer risk in a cohort of Swedish women: Relation varies by age and cancer site. Br J Cancer 85:346–349. Terry P, Jain M, Miller AB, et al. 2002a. Dietary intake of folic acid and colorectal cancer risk in a cohort of women. Int J Cancer 97:864–867. Terry PD, Miller AB, Rohan TE. 2002b. Obesity and colorectal cancer in women. Gut 51:191–194. Terzolo M, Tappero G, Borretta G, et al. 1994. High prevalence of colonic polyps in patients with acromegaly. Influence of sex and age. Arch Intern Med 154:1272–1276. The Alpha-Tocopherol Beta-Carotene Cancer Prevention Study Group. 1994. The effect of vitamin E and beta carotene on the incidence of lung cancer and other cancers in male smokers. N Engl J Med 330:1029–1035. Thiagalingam S, Lengauer C, Leach FS, et al. 1996. Evaluation of candidate tumour suppressor genes on chromosome 18 in colorectal cancers. Nat Genet 13:343–346. Thibodeau SN, Bren G, Schaid D. 1993. Microsatellite instability in cancer of the proximal colon. Science 260(5109):816–819. Thun MJ, Calle EE, Namboodiri MM, et al. 1992. Risk factors for fatal colon cancer in a large prospective study. J Natl Cancer Inst 84:1491–1500. Thun MJ, Namboodiri MM, Heath CW, Jr. 1991. Aspirin use and reduced risk of fatal colon cancer. N Engl J Med 325:1593–1596. Thune I, Furberg AS. 2001. Physical activity and cancer risk: Dose-response and cancer, all sites and site-specific. Med Sci Sports Exerc 33(6 Suppl):S530–S550. Tobey TA, Greenfield M, Kraemer F, et al. 1981. Relationship between insulin resistance, insulin secretion, very low density lipoprotein kinetics, and plasma triglyceride levels in normotriglyceridemic man. Metabolism 30:165–171. Todoroki I, Friedman GD, Slattery ML, et al. 1999. Cholecystectomy and the risk of colon cancer. Am J Gastroenterol 94:41–46. Topping DL, Mayes PA. 1972. The immediate effects of insulin and fructose on the metabolism of the perfused liver. Changes in lipoprotein secretion, fatty acid oxidation and esterification, lipogenesis and carbohydrate metabolism. Biochem J 126:295–311. Torlakovic E, Skovlund E, Snover DC, et al. 2003. Morphologic reappraisal of serrated colorectal polyps. Am J Surg Pathol 27:65–81. Trock B, Lanza E, Greenwald P. 1990. Dietary fiber, vegetables, and colon cancer: critical review and meta-analyses of the epidemiologic evidence. J Natl Cancer Inst 82:650–661. Tseng M, Murray SC, Kupper LL, et al. 1996. Micronutrients and the risk of colorectal adenomas. Am J Epidemiol 144:1005–1014. Tuomilehto J, Knowler WC, Zimmet P. 1992. Primary prevention of non-insulin-dependent diabetes mellitus. Diabetes Metab Rev 8:339–353. Ulrich CM, Kampman E, Bigler J, et al. 1999. Colorectal adenomas and the C677T MTHFR polymorphism: evidence for gene-environment interaction? Cancer Epidemiol Biomarkers Prev 8:659–668. Ulvik A, Evensen ET, Lien EA, et al. 2001. Smoking, folate and methylenetetrahydrofolate reductase status as interactive determinants of adenomatous and hyperplastic polyps of colorectum. Am J Med Genet 101: 246–254. van der Meer R, de Vries HT. 1985. Differential binding of glycine- and taurineconjugated bile acids to insoluble calcium phosphate. Biochem J 229:265–268. Vasen HF, van Erpecum KJ, Roelfsema F, et al. 1994. Increased prevalence of colonic adenomas in patients with acromegaly. Eur J Endocrinol 131:235–237.

829

Vasen HF, Watson P, Mecklin JP, et al. 1999. New clinical criteria for hereditary nonpolyposis colorectal cancer (HNPCC, Lynch syndrome) proposed by the International Collaborative group on HNPCC. Gastroenterology 116:1453–1456. Waaler HT. 1984. Height, weight and mortality. The Norwegian experience. Acta Med Scand Suppl 679:1–56. Wakabayashi K, Nagao M, Esumi H, et al. 1992. Food-derived mutagens and carcinogens. Cancer Res 52:2092s–2098s. Wang JT, Ho LT, Tang KT, et al. 1989. Effect of habitual physical activity on age-related glucose intolerance. J Am Geriatr Soc 37:203–209. Watatani M, Yoshida T, Kuroda K, et al. 1996. Allelic loss of chromosome 17p, mutation of the p53 gene, and microsatellite instability in right- and leftsided colorectal cancer. Cancer 77(8 Suppl):1688–1693. Waterhouse J, Muir C, Shanmugaratnam K, et al. 1982. Cancer Incidence in Five Continents. Vol. IV. Lyon: IARC Scientific Publications, no. 42. Weir JM, Dunn JE, Jr. 1970. Smoking and mortality: A prospective study. Cancer 25:105–112. Wheeler JM, Bodmer WF, Mortensen NJ. 2000a. DNA mismatch repair genes and colorectal cancer. Gut 47:148–153. Wheeler JM, Loukola A, Aaltonen LA, et al. 2000b. The role of hypermethylation of the hMLH1 promoter region in HNPCC versus MSI+ sporadic colorectal cancers. J Med Genet 37:588–592. White E, Shannon JS, Patterson RE. 1997. Relationship between vitamin and calcium supplement use and colon cancer. Cancer Epidemiol Biomarkers Prev 6:769–774. Will JC, Galuska DA, Vinicor F, et al. 1998. Colorectal cancer: another complication of diabetes mellitus? Am J Epidemiol 147:816–825. Willett WC, Stampfer MJ, Colditz GA, et al. 1989. Relation of meat, fat and fiber intake to colon cancer risk in a prospective study among women [abstract]. Am J Epidemiol 130:820. Willett WC, Stampfer MJ, Colditz GA, et al. 1990. Relation of meat, fat, and fiber intake to the risk of colon cancer in a prospective study among women. N Engl J Med 323:1664–1672. Wong NA, Harrison DJ. 2001. Colorectal neoplasia in ulcerative colitis-recent advances. Histopathology 39:221–234. Wu AH, Paganini-Hill A, Ross RK, et al. 1987. Alcohol, physical activity and other risk factors for colorectal cancer: A prospective study. Br J Cancer 55:687–694. Wu K, Willett WC, Fuchs CS, et al. 2002. Calcium intake and risk of colon cancer in women and men. J Natl Cancer Inst 94:437–446. Wu XC, Chen VW, Steele B, et al. 2001. Subsite-specific incidence rate and stage of disease in colorectal cancer by race, gender, and age group in the United States, 1992–1997. Cancer 92:2547–2554. Wynder EL, Fujita Y, Harris RE, et al. 1991. Comparative epidemiology of cancer between the United States and Japan. A second look. Cancer 67:746–763. Yamada K, Araki S, Tamura M, et al. 1998. Relation of serum total cholesterol, serum triglycerides and fasting plasma glucose to colorectal carcinoma in situ. Int J Epidemiol 27:794–798. Yamasaki E, Ames BN. 1977. Concentration of mutagens from urine by absorption with the nonpolar resin XAD-2: Cigarette smokers have mutagenic urine. Proc Natl Acad Sci U S A 74:3555–3559. Yokoyama A, Muramatsu T, Ohmori T, et al. 1998. Alcohol-related cancers and aldehyde dehydrogenase-2 in Japanese alcoholics. Carcinogenesis 19:1383–1387. Yoo KY, Tajima K, Inoue M, et al. 1999. Reproductive factors related to the risk of colorectal cancer by subsite: A case-control analysis. Br J Cancer 79:1901–1906. Ziegler RG, Devesa SS, Fraumeni JFJ. 1986. Epidemiologic patterns of colorectal cancer. In: DeVita VT, Hellman S, Rosenberg SA, eds. Important Advances in Oncology. 5th ed. Philadelphia: J.B. Lppincott, pp. 209–232. Ziel F, Peters A. 1988. Acromegaly and gastrointestinal adenocarcinomas. Ann Intern Med 109:514–515.

43

Anal Cancer MORTEN FRISCH AND MADS MELBYE

V

ariants of invasive squamous carcinoma of the anal canal and perianal skin, referred to collectively as anal cancers in this chapter, are rare malignancies constituting approximately 2% of all intestinal cancers. Of a crude estimate of 35,000–40,000 new anal cancers worldwide, approximately 3300 cases are diagnosed in the United States each year. In developed countries, a marked and sustained increase in incidence has taken place over the past half century, which has been more pronounced in urban than rural areas and more so in women than men, except homosexual men who constitute a particular high-risk group. As in cervical cancer, convincing molecular biologic evidence has documented that most anal cancers are caused by sexually transmissible human papillomaviruses (HPVs). Recent trials of prophylactic vaccination against ano-genital HPV infections warrant cautious optimism with regard to primary anal cancer prevention in the future.

CLASSIFICATION Anatomy and Histology The anal region comprises the anal canal and the perianal skin. Clinically, the anal canal begins at the upper surface of the anorectal ring, from which it passes through the pelvic floor and ends at the external aperture of the alimentary tract (anus). The anal canal is divided in three histologically distinct, but anatomically highly variable zones (Fenger, 1987; Fenger, 1997). The upper zone (colorectal zone) of the anal canal is covered by colorectal type mucosa indistinguishable from rectal mucosa. The middle zone (anal transitional zone) is covered by specialized transitional epithelium and usually extends from the dentate line and around 0.5–1.0 cm upwards where it gradually merges with the epithelium of the colorectal zone. The lower zone (squamous zone) of the anal canal is covered by squamous epithelium and extends from the dentate line and downwards to the anus, where it gradually merges with the perianal skin, which in turn is defined histologically by the appearance of skin appendages. There is no generally accepted definition of the limit between perianal skin and buttock skin (Fenger et al., 2000b). The anal mucosa gives rise to a variety of distinct tumor types. Combined anatomic and histologic criteria should be used to avoid important misclassification that will otherwise affect incidence rate estimates, blur or accentuate temporal changes, hamper international comparisons, and bias risk factor associations in etiologic studies. Specifically, squamous histologies constitute less than 30% of all anally located invasive cancers in countries like Israel, the Phillippines, Vietnam, and Lithuania, whereas in Scandinavia, Switzerland, and the United States variants of squamous cell carcinoma account for 70% or more of cancers in the anal region (International Agency for Research on Cancer, 2002). Therefore, adenocarcinomas should be disregarded, or analyzed separately, in etiologic anal cancer studies. Similarly, common basal cell carcinomas located in the ill-demarcated perianal skin region should not be included. Genuine anal gland adenocarcinomas and other rare cancers (including anally located Paget disease, malignant melanoma, lymphomas, carcinoids, and neuroendocrine, mesenchymal, and neurogenic tumors) are so uncommon and presumably etiologically diverse that they, too, should be disregarded. Due to considerable interobserver and intraobserver variability in the assessment of detailed histologic subtypes of anal squamous carci-

830

noma (Fenger et al., 2000a), anal cancer studies with histopathological data collection should restrict the focus to variants of squamous carcinoma that originate in the anal canal or perianal skin. Registerbased epidemiological anal cancer studies using routinely recorded cancer data should include not only anal cancers coded as variants of anal squamous carcinoma but, when possible, also those recorded as variants of squamous carcinoma of the rectum, because these most likely represent imprecisely coded proximal anal canal cancers.

DEMOGRAPHIC PATTERNS Incidence in the United States Anal cancers, defined as all invasive cancers (behavior code 3) under topography codes 209–218 (anus, anal canal, and rectum) and histology codes 8050–8076, 8094, or 8120–8124 (variants of squamous cell carcinoma), according to the second edition of the International Classification of Diseases for Oncology, ICD-02 (World Health Organization, 1990), account for approximately 0.2% of all invasive cancers in the United States. Based on data from the Surveillance, Epidemiology, and End Results (SEER) program of the United States, age-standardized (United States 2000 standard population) incidence rates of anal cancer for the period 1996–2000 were 1.1 and 1.5 per 100,000 person-years among white and black men, respectively, and 1.4 and 1.2 per 100,000 person-years among white and black women, respectively.

Temporal and Demographic Variation in Incidence Few systematic studies have described temporal trends and demographic patterns for anal cancer in well-defined, unselected populations (Goldman et al., 1989; Frisch et al., 1993; Melbye et al., 1994b; Fisher et al., 1997; Frisch and Goodman, 2000; Cress and Holly, 2003). Based on data from cancer registries in Denmark and Connecticut the incidence of anal cancer remained fairly stable up to around 1960 (Frisch et al., 1993; Melbye et al., 1994b). Since then, the annual incidence rate in Denmark has increased fourfold in women and doubled in men (Frisch, 2002). In Connecticut, incidence rates of anal cancer doubled in both women and men between 1940 and 1988 (Melbye et al., 1994b). SEER data show continuously increasing incidence rates of anal cancer throughout the period 1973–2000 (Fig. 43–1). Rates per 100,000 person-years increased between 1973–1977 and 1998–2000 from 0.5 to 1.1 among white men, from 0.9 to 1.5 among white women, from 0.6 to 1.6 among black men, and from 1.0 to 1.3 among black women. Significant differences in incidence have been observed between populations in urban and rural areas. In Denmark, the increase in incidence for both men and women was steeper among people living in or near the capital of Copenhagen compared with people living in rural areas (Frisch et al., 1993). In Sweden, an almost fivefold variation in incidence was seen between urban (highest) and rural (lowest) areas (Goldman et al., 1989). In the United States, the increase in incidence has been particularly pronounced in metropolitan areas with large populations of homosexual and bisexual men (Melbye et al., 1994b; Cress and Holly, 2003). During 1973–2000 the incidence rate among white men in San Francisco/Oakland increased from 0.7 to 2.4 per 100,000 person-years (Fig. 43–2).

RATE PER 100,000 PERSON-YEARS

Anal Cancer

Cancer, 2002). Because a non-trivial proportion of anal cancers are coded under ICD-02 topography code 209 (rectum), incidence figures serve mostly to highlight major differences in incidence between geographic regions, rather than as exact estimates of anal cancer incidence. Low rates are seen in Israel and Japan (0.1 per 100,000 person-years) and high rates are seen among black and white men in California (1.0–1.8 per 100,000 person-years) and among women in Switzerland (1.2 per 100,000 person-years). In Western Europe and Central and South America, women generally have higher incidence rates than men, whereas in North America, the Middle and Far East, and Oceania differences between the sexes are generally small.

2.0 WHITE MALES WHITE FEMALES BLACK MALES BLACK FEMALES

1.0

Age, Sex, and Race

0.5 1975

1980

1985

1990

1995

2000

YEAR

Figure 43–1. Age-standardized (US 2000 population) incidence rates 1973–2000 of invasive anal cancer per 100,000 person-years among US white and black men and women, in 5-year calendar periods 1973–1977 through 1993–1997, and 1998–2000. Anal cancer defined as ICD-O2 topography codes 209–218 and histology codes 8050–8076, 8094, or 8120–8124. (Source: Data from SEER program 2003.)

International Patterns Due to important classification issues that may differ between countries, caution is warranted when comparing anal cancer incidence rates between geographic regions. Major differences exist between registries in the proportion of recorded anal cancers that is histologically examined. Moreover, in some countries only a minority of those anally located cancers that are histologically examined represent genuine anal cancers of squamous carcinoma or basaloid/cloacogenic carcinoma histology (13% in Lithuania, 17% in the Philippines, and 28% in Israel). In other countries, corresponding proportions are considerably higher (72% in Norway, 74% in Denmark, 76% among US blacks, 79% among US whites, and 87% in Sweden). Figure 43–3 presents international incidence rates for those anal cancers that are coded under ICD-02 topography code 210–218 (anus and anal canal) and restricted histologically to verified squamous carcinomas or basaloid/cloacogenic carcinomas (International Agency for Research on

RATE PER 100,000 PERSON-YEARS

831

2.0

SAN FRANCISCO/ OAKLAND

Incidence rates increase sharply with age, as illustrated by SEER data for the period 1973–2000 (Fig. 43–4). In the period 1996–2000, median age at diagnosis was 58 years in white men, 64 years in white women, 50 years in black men, and 57 years in black women. While the median age at diagnosis decreased over time in men it remained relatively stable among women. Accordingly, in both the United States and Denmark the increase in incidence over time has been most pronounced among young men whereas incidence rates have increased among all age groups in women. Anal cancer incidence rates among black men have remained higher than among white men throughout 1973–2000, whereas in recent years the incidence among white women has become higher than among black women (Fig. 43–1). In the San Francisco Bay area the incidence rate in black men is above 3 per 100,000 person-years. In contrast to most other populations and ethnic groups, black men have higher incidence rates than black women (female/male incidence rate ratio 0.8). Asian-Pacific Islanders have markedly lower incidence rates than whites and blacks (Frisch and Goodman, 2000; Cress and Holly, 2003).

Marital Status Comparison of marital status distributions among patients with anal cancer and control patients with non-anogenital malignancies (e.g., stomach cancer or colon cancer) has been undertaken to evaluate the association between anal cancer and unmarried status (Austin, 1982; Daling et al., 1982; Peters and Mack, 1983; Frisch et al., 1993; Melbye et al., 1994b). Across studies, no remarkable association with marital status was seen among women, but Danish men with anal cancer were 2–3 times more likely to have remained unmarried than controls during 1943–1987 (Frisch et al., 1993). In contrast, men with anal cancer diagnosed in Atlanta, Detroit, Seattle, or San Francisco/ Oakland during 1985–1989 were around 10 times more likely to have never married than were controls (Melbye et al., 1994b). Because never married is only an insensitive and unspecific correlate of homosexuality, underlying associations with male homosexuality are underestimated.

1.0 CONNECTICUT

Survival 0.5

IOWA

1975

1980

1985

1990

1995

2000

YEAR

Figure 43–2. Age-standardized (US 2000 population) incidence rates 1973–2000 of invasive anal cancer per 100,000 person-years among white men in San Francisco/Oakland, Connecticut, and Iowa, in 5-year calendar periods 1973–1977 through 1993–1997, and 1998–2000. Anal cancer defined as ICD-O2 topography codes 209–218 and histology codes 8050–8076, 8094, or 8120–8124. (Source: Data from SEER program 2003.)

Among anal cancer patients in the United States, 5-year relative survival rates range from 80% for patients with localized anal cancer to 20% for patients with metastatic disease (data from SEER Program, 2003). Among patients whose tumor stage is localized or regionally spread at the time of diagnosis, whites have better prognosis than blacks, whereas survival varies little among subgroups of patients with metastatic anal cancer. For both whites and blacks, relative survival is higher for females than males at any given stage.

LIFESTYLE AND ENVIRONMENTAL FACTORS Male Homosexuality More than half a century ago, some men with anal cancer were “suspected of having abnormal sexual habits” (Binkley and Derrick, 1945). Decades later, around the onset of the acquired immunodeficiency

832

PART IV: CANCER BY TISSUE OF ORIGIN

NORTH AMERICA United States, SEER San Francisco

Los Angeles

White Black White Black Hispanic White Black Hispanic

Canada

CENTRAL AND SOUTH AMERICA Puerto Rico Colombia Uruguay

MIDDLE AND FAR EAST India, Delhi Israel, Jews Japan, Osaka Korea, Seoul Philippines, Manila Taiwan Vietnam, Ho Chi Minh City

EUROPE Czech Republic Denmark Ireland Italy, Torino Lithuania The Netherlands Norway Sweden Switzerland, Zurich Uk, England Uk, Scotland

OCEANIA Australia, New South Wales New Zealand

Figure 43–3. International incidence rates (age-standardized to the world standard population) of invasive anal cancer per 100,000 person-years among men and women. Incidence rates are calculated as rates for all invasive cancers under ICD-O2 topography codes 210–218 multiplied by registry-specific proportions of histologically verified squamous cell

carcinomas or basaloid/cloacogenic carcinomas at this site. Anal cancers recorded under ICD-O2 topography code 209 (rectum) are not included, see text data from Cancer Incidence in Five Continents, Vol. VIII, 2002. (Source: International Agency for Research on Cancer, 2002.)

syndrome (AIDS) epidemic, cases of in situ or invasive anal cancer were reported among young homosexual men (Cooper et al., 1979; Li et al., 1982; Croxson et al., 1984; Howard et al., 1986; Wexner et al., 1987) and register-based observations showed that men who had remained unmarried were at increased risk of anal cancer (Austin, 1982; Daling et al., 1982; Peters and Mack, 1983). The perception of a strong positive association with male homosexuality was supported in subsequent studies. Observations in the late 1980s (Daling et al., 1987; Holly et al., 1989) and in the 1990s (Scholefield et al., 1990b; Frisch et al., 1993; Melbye et al., 1994b; Biggar and Melbye, 1996; Frisch et al., 1997; Frisch and Goodman, 2000) substantiated the close link between unmarried status among men, but not women, and the risk of anal cancer. Unmarried status in men was found in both Denmark and Connecticut to be associated with a twofold or higher anal cancer risk, and in both settings the association dated back at least

to the 1940s. In the state of Hawaii, a high-incidence area for anal cancer among men, 43% of male patients diagnosed between 1973 and 1996 were unmarried at the time of diagnosis as contrasted by 9% of male patients with stomach cancer (odds ratio 8.1) (Frisch and Goodman, 2000). With the onset of the AIDS epidemic, the high risk of anal cancer in homosexual men was documented more directly (Melbye et al., 1994a; Goedert et al., 1998; Frisch et al., 2000). Among 309,365 patients with AIDS, 221 cases of invasive anal cancer and 107 cases of in situ anal cancer were diagnosed in the period from 5 years before to 2 years after the onset of AIDS. Among all human immunodeficiency virus (HIV) exposure categories, the highest relative risk was seen in homosexual men for both invasive anal cancer (standardized incidence ratio 59.5) and in situ anal cancer (standardized incidence ratio 99.8) (Frisch et al., 2000). However, the explanation for these

Anal Cancer 5.0

MALES

RATE PER 100,000 PERSON-YEARS

1.0 0.5

0.1

5.0 FEMALES

1.0 0.5

0.1

30-39 40-49 50-59 1975

1980

1985

1990

60-69 70-79 80+ 1995

2000

Figure 43–4. Age-specific incidence rates of invasive anal cancer per 100,000 person-years among US white and black men and women, in 5-year calendar periods 1973–1977 through 1993–1997, and 1998–2000. Anal cancer defined as ICD-O2 topography codes 209–218 and histology codes 8050–8076, 8094, or 8120–8124. (Sources: Data from SEER program 2003.)

high relative risks remains incompletely understood. Koblin et al. (1996) calculated incidence rates of anal cancer according to HIV status among 2300 homosexual men in San Francisco and found rather similar anal cancer incidence estimates among HIV-positive (13.4 per 100,000 person-years) and HIV-negative (16.6 per 100,000 personyears) homosexual men, thus failing to support HIV infection as an independent risk factor. In the AIDS-Cancer Match Registry Study in the United States, the annual post-AIDS incidence of anal cancer was estimated to be 23.9 per 100,000 homosexual men with AIDS in the period 1978–1996 (Frisch et al., 2000), an estimate that is comfortably embedded in the range of between 12.5 and 36.9 per 100,000 person-years believed to apply to the general homosexual male population in the United States before the onset of the AIDS epidemic (Daling et al., 1982). Homosexual men have also been reported to be at increased risk of anal cancer outside the AIDS setting in the United States. In a Scandinavian case-control study, men with anal cancer were significantly more likely than controls to report homosexual experience. Indeed, while 15% of 92 men with invasive or in situ anal cancer reported any homosexual experience, none of 392 male control subjects reported such experience (Frisch et al., 1997). A population-based study in Denmark showed that homosexual men in registered homosexual partnerships were at 31-fold increased risk of anal cancer compared with the general male population (Frisch et al., 2003). Taken together, these studies corroborate the view that homosexual men are at high relative risk of anal cancer compared with the general male population.

Sexual Factors Other Than Male Homosexuality Combined evidence from case-control studies in the United States (Daling et al., 1987; Holmes et al., 1988; Holly et al., 1989) and

833

Scandinavia (Frisch et al., 1997) show that sexual factors are also centrally involved in the etiology of anal cancer among women and heterosexual men. In the Scandinavian study, statistically significant associations were reported for a variety of sexually transmitted diseases, prior cervical neoplasia, and among women, measures of sexual promiscuity in the male partner. In both sexes, the risk of anal cancer was found to increase proportionally with the lifetime number of partners of the opposite sex (Frisch et al., 1997). A cohort study of husbands of 135,500 Swedish women with in situ or invasive cervical cancer showed an almost twofold increased risk of anal cancer in the husbands, thus adding further to the presumed heterosexual transmission of HPV in the etiology of some cases of anal cancer in men (Hemminki and Dong, 2000). Three case-control studies have reported an increased risk of anal cancer among women who practice receptive anal intercourse, notably among women starting this sexual practice at a young age (Daling et al., 1987; Holly et al., 1989; Frisch et al., 1997). Further support for a role of anal intercourse as a mode of transmission of the presumed etiologic agent to the anal region came from a cohort study of 410 young women in San Francisco. Receptive anal intercourse was found to be associated with an almost sevenfold increased risk of abnormal anal cytology (Moscicki et al., 1999). Penile HPV status of the male partners with whom women engage in unprotected anal intercourse is likely to determine whether this sexual activity is risky or not. The Danish-Swedish case-control study yielded some support for such a “male factor” in the development of anal cancer similar to that described for cervical cancer (Kjær et al., 1991). Women who reported having a spouse with a lifetime number of three or more other sex partners or a spouse with a history of sexually transmitted diseases were at higher risk of anal cancer than women who reported having a spouse with no other sex partners and those women reporting no sexually transmitted disease history in the spouse (Frisch et al., 1997). Underreporting of sensitive matters like anal intercourse is likely to have occurred to some extent in all studies. If underreporting occurred to a similar extent among cases and controls, as suggested by similar finding using two different control groups in the Scandinavian casecontrol study (Frisch et al., 1997), the strength of the underlying true association with anal intercourse is underestimated. Consequently, anal intercourse may be a more important route of transmission of the presumed etiologic factor to the anal region than suggested by published data. Overall, recent studies corroborate the view that measures of sexually extroverted lifestyles, regardless of sexual preference and specific sexual practices, are centrally linked to the risk of anal cancer. Most likely these links reflect an underlying close association between such measures and the risk of anal HPV infection.

Human Papillomaviruses Around 1976, HPV was hypothesized to be involved in the etiology of cervical and other anogenital cancers, including cancer of the perianal skin (zur Hausen, 1976), a hypothesis that has been supported ever since. Early observations of epidemiological parallels between anal and lower female genital cancers (Cabrera et al., 1966; Stern and Kaplan, 1969) were followed by studies evaluating the possible role of HPV in anal cancer more directly. Molecular biologic changes associated with HPV infection have been most extensively studied for lesions of the uterine cervix. However, interactions between cellular proteins and viral oncoproteins from HPV types associated with high risk of cervical cancer (hrHPVs), such as HPV16, are likely to be similar, if not identical, in cancers of neighboring, including anal, epithelia (see Pathogenesis section below). Paraffin-embedded anal cancer tissues have been examined for the presence of HPV in several studies. Depending largely on the sensitivity of the methods used, prevalence estimates for HPV positivity have varied from 0% in a series of 13 anal cancers examined by DNA in situ hybridization for HPV types 6, 11, 16 and 18 (Duggan et al., 1989) to 100% in a series of 21 anal cancers tested by the polymerase chain reaction (PCR) technique for HPV types 16 and 18 (Youk et al., 2001). Table 43–1 shows the results from studies that used PCR to

834

PART IV: CANCER BY TISSUE OF ORIGIN

Table 43–1. Studies Using the Polymerase Chain Reaction Technique to Detect High-Risk Human Papillomavirus (hrHPV) DNA in Anal Cancer Specimens Proportion hrHPV-Positive Study

hrHPV Types Tested

Crook et al., 1991 Palefsky et al., 1991 Daling et al., 1992

16,18 16,18,31,33 16,18

Zaki et al., 1992 Holm et al., 1994

16,18,33 16

Shroyer et al., 1995 Noffsinger et al., 1995 Vincent-Salomon et al., 1996

16,18,33 16,18 16,18,33

Poletti et al., 1998 Indinnimeo et al., 1999 Frisch et al., 1999a Carter et al., 2001

16,18,31,33,52,58 16,18,31,33 16,18,31,33,35,39,45,51, 52,56,58,59,66,68 16

Youk et al., 2001

16,18

Anal Cancers 40/50 (80%) 11/13 (85%) Women: 65/93 (70%) Men: 24/36 (67%) 3/11 (27%) Women: 62/74 (84%) Men: 13/25 (56%) 22/27 (81%) 20/50 (40%) Women: 16/19 (84%) Men: 3/8 (38%) 12/33 (36%) 9/14 (64%) Women: 228/253 (90%) Men: 49/78 (63%) Women: 37/45 (82%) Men: 29/38 (76%) Women: 7/7 (100%) Men: 14/14 (100%)

Control Tissues 0/12 hemorrhoid tissues NA NA NA 0/13 normal anal epithelium and 0/16 rectal adenocarcinoma NA NA NA NA NA 0/20 rectal adenocarcinomas NA 0/13 normal anal epithelium from cases, 0/21 hemorrhoid tissues

NA, information not available.

determine hrHPV status in anal cancer tissues. Generally, anal cancers in women are hrHPV-positive more often than in men. Among patients in those studies that provided hrHPV status by gender, 85% of anal cancers among 491 women and 66% of anal cancers among 199 men were hrHPV-positive. Anal cancers in homosexual men were found in one large study to be significantly more often hrHPV-positive than anal cancers in men without homosexual experience (100% vs. 58%) (Frisch et al., 1999a). As in cervical cancer, HPV16 is by far the single most common hrHPV type found in anal cancer tissues from various parts of the world. HPV16 was detected by PCR in 73% of anal cancer specimens in Denmark and Sweden (Frisch et al., 1997), in 76% of specimens in Norway (Holm et al., 1994), in 80% of specimens from the United States (Carter et al., 2001), and in 100% of specimens from South Korea (Youk et al., 2001). Few attempts have been made to identify possible histologic differences and differences in the association with HPV between squamous carcinoma variants of the anal canal and those of the perianal skin. Limited numbers of patients combined with difficulties for the clinician to determine the exact anatomic origin for cancers in this region have made such attempts difficult. In one Scandinavian study anal canal cancers were clearly more often hrHPV-positive than cancers restricted to the perianal skin. Indeed, while hrHPVs were present in 92% of anal canal cancers, the proportion of hrHPVpositive perianal skin cancers was 64% (odds ratio 7.5). Histologic characteristics also differed between cancers of the anal canal and those of the perianal skin. Anal canal cancers were more often dominated by small- or medium-sized tumor cells, and they more often exhibited basaloid features and lacked keratinization. Perianal cancers were more often dominated by large tumor cells, absence of basaloid features, and presence of various degrees of keratinization (Frisch et al., 1999a). Thus, anal cancers may be divided in two categories based on the most proximal localization of the tumor in the anal region; one major group consisting of almost invariably hrHPV-positive cancers of the anal canal, and another smaller group of perianal skin cancers comprising both hrHPV-positive and hrHPV-negative cancers.

Infectious Agents Other than Human Papillomaviruses As for cervical cancer, infection with herpes simplex virus has been examined as a possible viral cofactor in HPV-associated anal cancer. Case-control studies reported serologic evidence of prior herpes simplex virus type 2 infection (Daling et al., 1987; Holmes et al., 1988;

Holly et al., 1989) or self-reported history of genital herpes to be more common in female anal cancer patients than controls, while findings in men have been less consistent (Daling et al., 1987; Frisch et al., 1997). Among 17 patients with high-grade anal squamous intraepithelial lesion (SIL) or invasive anal cancer, a DNA fragment common to herpes simplex virus types 1 and 2 was amplified by PCR in eight patients, and the authors suggested that herpes simplex virus infection might play a role in disease progression (Palefsky et al., 1991). To address the issue of whether herpes simplex virus infection could act as an etiologic cofactor in HPV-associated anal cancers, tumor tissues from 92 patients with hrHPV-positive anal cancer in the DanishSwedish case-control study were examined for herpes simplex virus DNA by PCR. Herpes simplex virus DNA was not detected in any of the tumors (Frisch, 2002), thus failing to support herpes simplex virus as an etiologic cofactor in anal cancer. There is also little evidence to suggest that viral DNA from other herpes viruses would be involved in the etiology of anal cancer (Palefsky et al., 1991; Mougin et al., 1995). The role of HIV infection as a possible cofactor in the etiology of anal cancer is incompletely understood. Prospective studies have shown that HIV infection is a major determinant of whether HPVinfected anal mucosa will remain histologically normal or undergo neoplastic transformation. Among HPV-infected individuals the incidence rate of anal SIL is several times higher in HIV-positive than in HIV-negative individuals and the rate is inversely related to the level of cellular immune competence (Critchlow et al., 1995; Palefsky et al., 1998c; Ellerbrock et al., 2000). So far, the logical extension that HIV infection would also lead to increased rates of invasive anal cancer has not been supported by data. In the AIDS-Cancer Match Registry Study in the United States there was no indication of an increasing trend in invasive anal cancer incidence with progression of HIV-related immune deficiency (Frisch et al., 2000). However, if more prolonged periods of immune dysregulation are required for invasive anal cancer to develop, as in the setting of transplantation-related immuno-suppression, improved survival among HIV-positive individuals might be accompanied by increasing incidence rates of invasive anal cancer in the years to come. Preliminary data suggest that combination antiretroviral therapies do not markedly alter the risk of anal cancer in HIV-positive individuals (Chin-Hong and Palefsky, 2002). Anal cancer patients more frequently report a history of sexually transmitted infections such as gonorrhea and syphilis than controls (Daling et al., 1987; Holmes et al., 1988; Holly et al., 1989; Frisch

Anal Cancer et al., 1997). However, these findings are likely to reflect sexual behaviors that are simultaneously associated with risk of anal HPV infection, and a direct etiologic role of these bacteria appears unlikely.

Tobacco Smoking The observation of a high smoker prevalence among patients with anal cancer was first reported in 1985 (Daniell, 1985). Subsequent casecontrol studies found higher proportions of smokers among anal cancer patients than controls (Daling et al., 1987; Holmes et al., 1988; Holly et al., 1989). These findings were interpreted in favor of a causal role for tobacco smoking in anal carcinogenesis, although incomplete control for confounding by factors related to sexual behavior could not be excluded. In a large Danish-Swedish case-control study, smoking was only significantly associated with anal cancer risk among premenopausal women (Frisch et al., 1999b). After adjustment for several important confounding variables, including sexual, venereal, and partner-related factors, the odds ratio associated with current smoking was 5.6 among premenopausal women. Corresponding multivariate odds ratios, 1.3 among postmenopausal women and 1.6 among men, were considerably smaller and not significantly different from unity. Additional observations question the possible role of smoking in the etiology of anal cancer among men. Firstly, compared with lifelong non-smokers, men who were former smokers in the Danish-Swedish study were at 50% reduced risk of anal cancer compared with never smokers, a finding that is unlikely to reflect a beneficial effect of prior smoking. Secondly, by restricting the case group to men whose anal cancers were hrHPV-positive, a role of smoking became further unlikely. The multivariate odds ratio associated with current smoking (vs. lifelong non-smoking) among men dropped from 1.6 in the hrHPV-unrestricted analysis to 1.0 in the hrHPV-restricted analysis (Frisch et al., 1999b). A cohort study in San Francisco showed that, after control for several potential confounders, smoking was not related to the risk of developing anal SIL in homosexual men (Palefsky et al., 1998a), a finding that further questions the role of smoking in the etiology of anal cancer among men. Several biological mechanisms have been proposed for the association of smoking with anogenital cancer risk (Winkelstein, 1990). Suggested explanations include, among others, the direct carcinogenic action of chemical compounds in tobacco smoke and the induction of immunological changes by constituents in smoke. However, such mechanisms are not easily compatible with a role of tobacco smoking restricted to premenopausal women. An alternative hypothesis has been suggested that is compatible with the recently documented existence of sex hormone receptors in the anal mucosa (Oettling and Franz, 1998); smoking might act in anal carcinogenesis, and possibly in anogenital squamous carcinogenesis in general, through antiestrogenic actions of tobacco smoke (Frisch et al., 1999b). Overall, current evidence suggests that smoking is a risk factor for anal cancer in women. The recent observation of no similar association in men and the observation among women of an effect of smoking that may be restricted to premenopausal women need replication in additional studies with adequate adjustment for potentially confounding sexual factors.

HOST FACTORS Familial and Genetic Susceptibility Published data do not suggest aggregation of anal cancer in firstdegree relatives. Likewise, the association between specific human leucocyte antigen class I or II alleles and the risk of anal cancer has not been addressed, and attempts to identify other genetic susceptibility markers for anal cancer have been unsuccessful (Chen et al., 1996; Chen et al., 1999; Zucchini et al., 2000).

Immune Function Cross-sectional and prospective studies in the settings of organ transplantation and HIV/AIDS have made it clear that the ability to

835

maintain healthy anogenital epithelia upon papillomaviral challenge depends on the competence of the immune system. The ability to mount adequate cell-mediated anti-viral immue responses is hampered in HIV-infected individuals (Clerici et al., 1997) and a general Th1to-Th2 shift in cytokine production profile has been suggested to accompany the progression from HPV infection to SIL (Wu and Kurman, 1997). Immunosuppressed individuals, whether related to organ transplantation or HIV infection, have higher rates of detectable anogenital HPV infection (Halpert et al., 1986; Alloub et al., 1989; Melbye et al., 1990; Ogunbiyi et al., 1994; Critchlow et al., 1995; Chiasson et al., 1997; Unger et al., 1997), presumably due to higher viral copy numbers (Melbye et al., 1996; Friedman et al., 1998). Additionally, in the HPV-infected individual, immunosuppression is associated with higher progression rates from normal epithelium to low-grade SILs and from low-grade to high-grade SILs (Critchlow et al., 1995; Palefsky et al., 1998c; Lacey et al., 1999; Ellerbrock et al., 2000) and with lower rates of HPV clearance and regression from neoplastic to normal epithelium (Critchlow et al., 1998; Delmas et al., 2000). In studies from San Francisco, significantly more HIV-positive (93%) than HIV-negative (61%) homosexual men had detectable HPV infection in the anal region, and detection of multiple hrHPV types was more common among HIV-positive men (Palefsky et al., 1998b). HIVpositive homosexual men with normal anal Papanicolaou smears at baseline had significantly higher incidence rates of high-grade anal SIL over a 4-year period than HIV-negative homosexual men, and the risk was inversely associated with the baseline CD4+ lymphocyte concentration (Palefsky et al., 1998c). The association of immunosuppression with risk of anogenital carcinoma in situ, the presumed immediate precursor for most HPVassociated anogenital cancers, has been only sparsely studied. Anogenital carcinoma in situ lesions are frequently categorized together with dysplastic lesions of lesser malignant potential in the broad group of high-grade SILs. In the setting of HIV/AIDS, relative risks for cervical, vulvar/vaginal, and penile carcinoma in situ, but not for anal carcinoma in situ, increased significantly with time in relation to AIDS onset, a correlate of the level of immune dysfunction (Frisch et al., 2000). However, the data quality for in situ cancers routinely reported to cancer registries may be questioned, so cautious interpretation is warranted. The possible role of immunosuppression in the final step from carcinoma in situ to invasive anal cancer is also unclear. Among renal transplant patients, invasive HPV-associated anogenital cancers have been reported to occur in marked excess (Blohmé and Brynger, 1985; Penn, 1986; Fairley et al., 1994; Birkeland et al., 1995). A cohort analysis among 8215 renal transplant patients and 7605 patients undergoing dialysis in Australia suggested that highly elevated risks of anal, vulvar, penile, and cervical cancers were due to transplantationassociated immunosuppression (Fairley et al., 1994). While established immunosuppression-associated cancers such as Kaposi sarcoma and non-Hodgkin lymphoma require only relatively short periods of immune dysregulation to occur in measurable excess, longer followup is required to identify less conspicuous increases for other cancers. Indeed, considerably longer intervals from the time of renal transplantation to the time of diagnosis have been noted for anogenital cancers than for other cancers occurring in excess among transplant patients. Specifically, while the average interval from transplantation to cancer was 23 months for Kaposi sarcoma, 36 months for lymphomas, and 61 months for non-anogenital cancers, it was 88 months for anogenital cancers, even though the latter cancer category included one-third carcinoma in situ lesions (Penn, 1986). Like in transplant recipients, excesses of invasive HPV-associated anogenital cancers have been well documented in patients with HIVmediated immunosuppression. However, the underlying reason for the excess of invasive anal, cervical, and other anogenital cancers in patients with HIV infection and AIDS (Frisch et al., 2000) has not been convincingly shown to be due to immunosuppression. While reported high relative risks show that HIV-infected individuals are more likely to develop HPV-associated cancers, the association appears to be due at least in part to a higher prevalence of HPV infection and possible

836

PART IV: CANCER BY TISSUE OF ORIGIN

cofactors among individuals with HIV infection. In the HIV/AIDS setting, the possible role of immunosuppression has remained difficult to assess, because cancer rates among comparable HIV-negative populations are not readily available. Early in the AIDS epidemic, homosexual male patients were perceived to be at exceptional risk of anal cancer (Li et al., 1982; Howard et al., 1986; Lorenz et al., 1991). This impression was confirmed in register-based cohort studies of HIV-infected individuals in the United States (Melbye et al., 1994a; Goedert et al., 1998; Frisch et al., 2000) and Denmark (Frisch et al., 2003). However, in other studies of AIDS patients, whether dominated by homosexual men (Grulich et al., 1999) or by patients in other HIV exposure categories (Sitas et al., 1997; Franceschi et al., 1998), anal cancer risks were not similarly elevated. Cancer trends in Uganda, where HIV is predominantly heterosexually transmitted, suggest no major changes in the incidence of HPVassociated cancers during the AIDS epidemic (Parkin et al., 1999). In the population-based AIDS-Cancer Match Registry Study, which linked cancer data to records of 309,365 persons with AIDS in the United States, anal cancer was estimated to occur at an incidence of 23.9 per 100,000 person-years among homosexual men in the first 2 years after AIDS onset (Frisch et al., 2000). This incidence estimate is well within the range of between 12.5 and 36.9 per 100,000 personyears believed to apply to the general US population of homosexual men in the pre-AIDS era (Daling et al., 1982). Consequently, the observed high relative risk of anal cancer in homosexual men with AIDS in the first years after the AIDS diagnosis appears to be explained mainly by a high risk of anal cancer among some homosexual men and not by a marked and rapid effect of HIV/AIDS-related immunosuppression. The partial restoration of the immune system and the resulting dramatic improvement in survival among AIDS patients given highly active antiretroviral therapy since 1996 (Anonymous, 2000) will expectedly increase the population of HIV-infected long-term survivors considerably. Among these HIV-positive individuals, a substantial proportion will have chronic or repeated anal HPV infection due to common modes of acquisition of HIV and HPV. If more than a few years are required for these persons to develop invasive anogenital HPV-associated cancers, as seen for organ transplant patients (Penn, 1986), the future may carry a paradoxical long-term increase in incidence of in situ and invasive anogenital cancers along with the overall improvement in survival.

Anthropometric Measures Although bodily measures have not been linked to anal cancer risk in men, premorbid body mass index was found in one large study to be inversely associated with the risk of anal cancer in women. After adjustment for several potential confounders, the risk associated with having a premorbid body mass index below the median for female population controls was almost twofold increased (Frisch et al., 1999b).

Sex Hormones Receptors for both female and male sex hormones have been demonstrated in the normal anal mucosa (Oettling and Franz, 1998), suggesting a physiological role of sex hormones in the maintenance of healthy anal epithelia. The absence of detectable sex hormone receptors in tumor cells from anal carcinomas in one study (Goldman et al., 1992) may reflect conformational changes or loss of hormone receptors during the cancerous process.

Benign Anal Lesions For decades, non-cancerous lesions of the anal region and accompanying inflammatory responses were considered to be predisposing factors for the development of anal cancer (Brofeldt, 1927; Rosser, 1931; Keyes, 1937; Buckwalter and Jurayj, 1957; Kline et al., 1964). A case-control study in California found histories of anal fissure, anal fistula, or hemorrhoids to be significantly more common among

patients with anal cancer than population controls (Holly et al., 1989), although some benign anal lesions may have been early signs of anal cancer. Clinical difficulties in distinguishing incipient cases of anal cancer from the considerably more frequent benign anal lesions are well known and may delay cancer diagnosis (Jensen et al., 1987). Two other case-control studies from the United States failed to report significant associations of anal cancer with fissures, fistulae, hemorrhoids, and anal abscesses (Daling et al., 1987; Holmes et al., 1988). A cohort study of 68,549 patients hospitalized with hemorrhoids or anal fissures, fistulae, or abscesses confirmed the strong statistical link between benign anal lesions and subsequent risk of anal cancer. Overall, a 4.4-fold increased risk was observed, but a closer look at the temporality of the association suggested a non-causal relationship. For each individual anal lesion-specific patient cohort, the relative risk of anal cancer was markedly increased only in the first year following the initial benign anal lesion, but the risk rapidly diminished and was not significantly elevated in any cohort more than 5 years after the initial cohort-defining lesion (Frisch et al., 1994a). These findings were confirmed in a subsequent cohort study (Lin et al., 1995). A report from the Danish-Swedish case-control study focused on benign anal lesions occurring 5 years or more before the diagnosis of anal cancer among case subjects and a comparable point in time among controls (Frisch et al., 1998). Associations with benign anal lesions differed between men and women. Anal fissure or fistula, hemorrhoids, anorectal prolapse, and unspecific anal irritation were all more frequent among men, but not women, with anal cancer compared with controls, whereas a history of anorectal abscess was positively associated with the risk only among women. This lack of consistency between the sexes does not fit well with a causal effect of benign anal lesions, because if epithelial repair mechanisms following anal inflammation result in genetic error and thus a higher risk of malignant transformation, a similar risk association would be expected in men and women. Increased risks associated with benign anal lesions were seen predominantly among men, which might be due, at least partly, to incomplete adjustment for confounding by one major determinant of anal cancer risk among men. To the extent unreported male homosexuality was present in the study, associations with less sensitive and thus more readily assessable exposures that are common in this group, such as benign anal lesions (Kazal et al., 1976; Owen, 1980; Scholefield et al., 1990a), may be spuriously linked to anal cancer risk as the result of confounding. Anal inflammation is a frequent complication in inflammatory bowel diseases, notably Crohn disease (Marks, 1990), and several case reports and small patient series have suggested an excess of anal cancers in young patients with Crohn disease (Slater et al., 1984; Walgenbach et al., 1984; Lumley and Stitz, 1991; Connell et al., 1994; Ky et al., 1998; Garcia et al., 2000). A cohort study of 2723 patients with Crohn disease and 6334 patients with ulcerative colitis was carried out to evaluate whether inflammatory bowel diseases confer an increased risk of anal cancer (Frisch and Johansen, 2000). Although an increased relative risk of anal cancer could not be excluded from these data, the absolute risk among inflammatory bowel disease patients appears to be small.

Metachronous Cancers Early case reports and small series of patients with synchronous or metachronous occurrence of anal cancer and cancers of the lower female genital tract, i.e. the uterine cervix, vagina or vulva (Cabrera et al., 1966; Stern and Kaplan, 1969) or their presumed precursor lesions (Scholefield et al., 1989) suggested a shared etiology for cancers at these contiguous sites. Later, the increased withinindividual occurrence of anal and lower gynecologic malignancies, notably cervical and vulvar cancers, has been substantiated statistically (Melbye and Sprøgel, 1991; Rabkin et al., 1992; Frisch et al., 1994b; Hemminki and Dong, 2000). Recent studies suggest that a proportion of oral and pharyngeal cancers, notably tonsillar cancers, may be etiologically linked to the same HPV types detected in anogenital cancers (Paz et al., 1997; Schwartz et al., 1998; Wilczynski et al., 1998). Follow-up of cohorts of patients with anal and genital cancers

Anal Cancer showed a particular excess of new squamous carcinomas in the palatine tonsils among these patients (Frisch and Biggar, 1999). This strengthens the belief that a common transmissible agent, most likely certain types of HPV, may be involved in anogenital and some oropharyngeal cancers, notably tonsillar cancers.

PATHOGENESIS Loss of functional tumor suppressor protein p53 is a crucial step in the development of most anogenital cancers, including anal cancer (Crook et al., 1991; Heselmeyer et al., 1997; Gupta and Sharma, 1998). In the absence of HPV, point mutations in the gene coding for p53 may lead to the production of inactive p53 or, more rarely, chromosomal deletions may be responsible. However, p53 inactivation in most anogenital carcinomas occurs at the protein level through formation of complexes between p53 and viral oncoproteins. The HPV genome has approximately 8000 base pairs and contains early genes termed E6, E7, E1, E2, E4, and E5, which code for proteins with regulatory functions, and late genes termed L1 and L2, whose gene products are viral capsid proteins (International Agency for Research on Cancer, 1995). The E6 protein of hrHPVs forms a complex with a cellular protein, the E6-associated protein. Upon binding to p53, this complex leads to proteolytic degradation of p53 (Werness et al., 1990). In the event of DNA damage, the cell then lacks p53 to either induce cell cycle arrest or apoptosis. The E7 protein of hrHPVs binds to another cell cycle regulator, the retinoblastoma protein pRb (Dyson et al., 1989). Hereby, signals that normally restrict cellular proliferation to basal epithelial layers are disrupted. When exposed to DNAdamaging stimuli the higher numbers of proliferating epithelial cells increase the likelihood of malignant transformation. The combination of increased cell proliferation (resulting from pRb inactivation) and impaired ability to induce cell cycle arrest or apoptosis following DNA damage (resulting from p53 inactivation) are two central mechanisms through which hrHPVs are believed to increase the risk of anogenital SILs (Alani and Münger, 1998). However, the genetic errors responsible for the final step toward invasion are not believed to be a direct consequence of the action of papillomaviral proteins (Galloway and McDougall, 1996).

Natural History of Anal Human Papillomaviruses Infections and Anal Neoplasia Anal HPV infections occur exceedingly rarely prior to sexual debut (Koch et al., 1997). Transmission of HPV from one person to another requires intimate physical contact, specifically a high number of sexual partners and the practice of anal intercourse are associated with anal HPV detection (Moscicki et al., 1999). However, anal intercourse seems not to be a requirement for the acquisition of anal HPV infection. Such infections are not infrequently encountered in heterosexual men and in women reporting no experience with anal intercourse (International Agency for Research on Cancer, 1995). Therefore, intimate contact other than penile-anal intercourse is likely to lead to transmission of HPV to the anal region. Women who have cervical HPV infection are more likely than other women to also have detectable HPV infection in the anal region. Indeed, studies have found rates of HPV detection to be similar or even greater in the anus compared with the uterine cervix (Melbye et al., 1996; Palefsky et al., 2001b). Anal SILs, histologically similar to those of the transformation zone of the uterine cervix, are divided into low-grade anal SIL, characterized by proliferation restricted to basal and parabasal epithelial cell layers, and high-grade anal SIL, characterized by proliferation of immature basal cells throughout the full thickness of the epithelium. Studies of HPV in anal SILs have been undertaken primarily among persons at high risk of HIV infection, including homosexual and bisexual men, intravenous drug users, and venereal disease clinic attenders (International Agency for Research on Cancer, 1995; Melbye et al., 1996). Whereas most low-grade SILs contain either low-risk or highrisk types of HPV, high-grade SILs are usually hrHPV-positive, notably to HPV16.

837

Prospective studies on the development of anal SILs are few (Critchlow et al., 1998; Xi et al., 1998; Palefsky et al., 1998c; Lacey et al., 1999). In a San Francisco study, homosexual men who were cytologically normal at baseline were stratified according to HIV status and CD4 level at baseline. Among HIV-positive men with CD4 counts below 500/ul, the 4-year projected incidence of highgrade anal SIL was around 50% as compared with 17% among HIV-negative men (Palefsky et al., 1998c). A follow-up study of 589 homosexual men in Seattle detected in situ anal cancer in 2% of 384 men who were consistently HPV-negative during the study, in 7% of 183 men who had evidence of anal infection with the prototype-like variant of HPV16 at baseline, and in 18% of 22 men with evidence of anal infection with non-prototype-like variants of HPV16 at baseline (Xi et al., 1998). These data have led to the suggestion that HPV16 strain variation may be a determining factor in the natural history of anogenital neoplasia (Xi et al., 1998; Youk et al., 2001; Da Costa et al., 2002). Whether HPV-encoded proteins play a role in the progression from low-grade to high-grade anal SIL and eventually from high-grade anal SIL to invasive anal cancer is unclear but several scenarios have been discussed (International Agency for Research on Cancer, 1995; Palefsky, 1998). It appears that as long as systemic and local immune responses remain intact, HPV replication and gene expression are kept under control and anal SILs are unlikely to occur. Lack of immunological control may result in increased HPV replication and expression of HPV-transforming genes that may eventually lead to the development of anal SIL.

PREVENTIVE MEASURES AND FUTURE DIRECTIONS Over the past half century, incidence rates of anal cancer have increased markedly in both men and women. Like for cervical cancer, it has been shown that most cases of anal cancer are caused by potentially preventable HPV infections. Additional factors are required for incident anal HPV infections to become chronic and, in a small proportion, to give rise to anal SIL and ultimately, in a small proportion of infected individuals, to invasive anal cancer. Further study may address the possible role of tobacco smoking and sex hormones. Also, the role of the immune system in maintaining healthy anogenital epithelia upon papillomaviral challenge requires study. In the not-too-distant future, however, additional insight into the cofactors that act with HPV to cause anal and genital cancers may become a matter of mere academic interest. With cautious optimism, groups around the world are currently testing prophylactic HPV vaccines (Sherman et al., 1998; Harro et al., 2001; Koutsky et al., 2002). Prophylactic strategies currently under investigation focus on the induction of humoral immune responses elicited by synthetic papillomavirus-like particles. If successful, such vaccines—aimed primarily at reducing morbidity and mortality from cervical HPV infections and cervical HPV-associated neoplasias—will expectedly also provide primary protection against most cases of anal SIL and anal cancer and against a proportion of other genital, conjunctival, and oropharyngeal cancers as well. For anal cancer, the theoretical impact of a prophylactic HPV vaccine is substantial. Using a range of realistic prevalence estimates of anal HPV infection in different subsets of the population, the proportion of all anal cancers that is theoretically preventable through vaccination was estimated to be 86%–89% in women, 40%–53% in heterosexual men, and 95% or more in HIV-positive homosexual men (Frisch, 2002). Attempts to develop therapeutic vaccines for the treatment of individuals already infected with HPV are also underway using strategies that aim at improving cellular immunity against HPV antigens (Murakami et al., 1999). Although several decades will elapse before the full impact of a successful HPV vaccination program can be appreciated, combined attempts at primary prevention and more effective therapy of already established HPV infections through vaccination create realistic hopes for a drastic long-term reduction in the incidence of HPV infections, SILs, and invasive cancers of anal, cervical, and other susceptible epithelia.

838

PART IV: CANCER BY TISSUE OF ORIGIN

Meanwhile, the recently suggested introduction of screening programs for anal SIL in populations at high risk of anal cancer needs careful consideration. In some urban areas anal cancer incidence among homosexual men may be as high as 40 per 100,000 personyears, which is comparable with the incidence of cervical cancer in the general female population prior to the introduction of cervical cytology screening (Chin-Hong and Palefsky, 2002). Goldie et al. (1999, 2000) have modeled the clinical and cost-related effectiveness of screening for high-grade anal SIL in homosexual and bisexual men. With the use of anal Papanicolaou smears annually or every 2 years among HIV-positive men and every 2–3 years among HIV-negative men, these researchers concluded that quality-adjusted years of life could be saved at a cost accepted for other preventive interventions. However, even if these conclusions are confirmed by other researchers in other settings, economic and political feasibility is not the only major determinant of whether screening for anal SILs should be implemented or not. The optimal treatment regimen for high-grade anal SIL, including adequate postoperative pain management, remains to be established (Chang et al., 2002). Current therapies include the destruction of neoplastic anal mucosa by procedures that may leave patients with painful wounds for weeks, a problem that may result in low patient compliance in a screening program. Through successful vaccine strategies the 21st century may provide both primary prevention and more efficient treatment strategies for anogenital lesions caused by HPV, including anal SIL and anal cancer. However, trends in anal cancer incidence will expectedly continue to increase in the foreseeable future, reflecting that successively more sexually experienced generations will reach high-risk ages of anal cancer in the years to come. The increased prevalence over time of anogenital HPV infections in sexually active subsets of the population combined with greater sexual experimentation, including the increased practice of anal intercourse (Jæger et al., 2000), suggest that a growing proportion of the general population will be at risk of long-standing or repeated anal HPV infection. Another possible contributor to the expected future increase in anal cancer incidence is the increased life expectancy among HIV-infected individuals on highly active antiretroviral therapy. Of note, preliminary data suggest that the impact of antiretroviral therapies on the natural history of anal SIL is limited (Palefsky et al., 2001a). Consequently, if prolonged periods of immune dysregulation are required for the occurrence of in situ and invasive anogenital HPV-associated carcinomas, the increased life expectancy among HIV-infected individuals may add to the expected increase in anal cancer incidence. References Alani RM, Münger K. 1998. Human papillomaviruses and associated malignancies. J Clin Oncol 16:330–337. Alloub MI, Barr BB, McLaren KM, et al. 1989. Human papillomavirus infection and cervical intraepithelial neoplasia in women with renal allografts. Br Med J 298:153–156. Anonymous. 2000. Survival after introduction of HAART in people with known duration of HIV-1 infection. The CASCADE Collaboration. Concerted Action on SeroConversion to AIDS and Death in Europe. Lancet 355:1158–1159. Austin DF. 1982. Etiological clues from descriptive epidemiology: Squamous carcinoma of the rectum or anus. Natl Cancer Inst Monogr 62:89–90. Biggar RJ, Melbye M. 1996. Marital status in relation to Kaposi’s sarcoma, non-Hodgkin’s lymphoma, and anal cancer in the pre-AIDS era. J AIDS Hum Retrovirol 11:178–182. Binkley GE, Derrick WA. 1945. The association of squamous cancer with anal manifestations of lymphogranuloma venereum. Am J Dig Dis 12:46–47. Birkeland SA, Storm HH, Lamm LU, et al. 1995. Cancer risk after renal transplantation in the Nordic countries, 1964–1986. Int J Cancer 60: 183–189. Blohmé I, Brynger H. 1985. Malignant disease in renal transplant patients. Transplantation 39:23–25. Brofeldt SA. 1927. Zur Pathogenese des Plattenepithelkrebses der Pars analis recti. Acta Soc Med Fenn Duodecim 8:3–15. Buckwalter JA, Jurayj MN. 1957. Relationship of chronic anorectal disease to carcinoma. Arch Surg 75:352–361. Cabrera A, Tsukada Y, Pickren JW, Moore R, Bross ID. 1966. Development of lower genital carcinomas in patients with anal carcinoma. A more than casual relationship. Cancer 19:470–480.

Carter JJ, Madeleine MM, Shera K, et al. 2001. Human papillomavirus 16 and 18 L1 serology compared across anogenital cancer sites. Cancer Res 61:1934–1940. Chang GJ, Berry JM, Jay N, Palefsky JM, Welton ML. 2002. Surgical treatment of high-grade anal squamous intraepithelial lesions: A prospective study. Dis Colon Rectum 45:453–458. Chen C, Cook LS, Li XY, et al. 1999. CYP2D6 genotype and the incidence of anal and vulvar cancer. Cancer Epidemiol Biomarkers Prev 8:317–321. Chen C, Madeleine MM, Lubinski C, et al. 1996. Glutathione S-transferase M1 genotypes and the risk of anal cancer: A population-based case-control study. Cancer Epidemiol Biomarkers Prev 5:985–991. Chiasson MA, Ellerbrock TV, Bush TJ, Sun XW, Wright TC, Jr. 1997. Increased prevalence of vulvovaginal condyloma and vulvar intraepithelial neoplasia in women infected with the human immunodeficiency virus. Obstet Gynecol 89:690–694. Chin-Hong PV, Palefsky JM. 2002. Natural history and clinical management of anal human papillomavirus disease in men and women infected with human immunodeficiency virus. Clin Infect Dis 35:1127–1134. Clerici M, Fusi ML, Ruzzante S, et al. 1997. Type 1 and type 2 cytokines in HIV infection—a possible role in apoptosis and disease progression. Ann Med 29:185–188. Connell WR, Sheffield JP, Kamm MA, et al. 1994. Lower gastrointestinal malignancy in Crohn’s disease. Gut 35:347–352. Cooper HS, Patchefsky AS, Marks G. 1979. Cloacogenic carcinoma of the anorectum in homosexual men: An observation of four cases. Dis Colon Rectum 22:557–558. Cress RD, Holly EA. 2003. Incidence of anal cancer in California: Increased incidence among men in San Francisco, 1973–1999. Prev Med 36: 555–560. Critchlow CW, Hawes SE, Kuypers JM, et al. 1998. Effect of HIV infection on the natural history of anal human papillomavirus infection. AIDS 12:1177–1184. Critchlow CW, Surawicz CM, Holmes KK, et al. 1995. Prospective study of high grade anal squamous intraepithelial neoplasia in a cohort of homosexual men: Influence of HIV infection, immunosuppression and human papillomavirus infection. AIDS 9:1255–1262. Crook T, Wrede D, Tidy J, et al. 1991. Status of c-myc, p53 and retinoblastoma genes in human papillomavirus positive and negative squamous cell carcinomas of the anus. Oncogene 6:1251–1257. Croxson T, Chabon AB, Rorat E, Barash IM. 1984. Intraepithelial carcinoma of the anus in homosexual men. Dis Colon Rectum 27:325–330. Da Costa MM, Hogeboom CJ, Holly EA, Palefsky JM. 2002. Increased risk of high-grade anal neoplasia associated with a human papillomavirus type 16 E6 sequence variant. J Infect Dis 185:1229–1237. Daling JR, Sherman KJ, Hislop TG, et al. 1992. Cigarette smoking and the risk of anogenital cancer. Am J Epidemiol 135:180–189. Daling JR, Weiss NS, Hislop TG, et al. 1987. Sexual practices, sexually transmitted diseases, and the incidence of anal cancer. N Engl J Med 317:973–977. Daling JR, Weiss NS, Klopfenstein LL, et al. 1982. Correlates of homosexual behavior and the incidence of anal cancer. JAMA 247:1988–1990. Daniell HW. 1985. Causes of anal carcinoma [letter]. JAMA 254:358. Delmas MC, Larsen C, van Benthem B, et al. 2000. Cervical squamous intraepithelial lesions in HIV-infected women: Prevalence, incidence and regression. European Study Group on Natural History of HIV Infection in Women. AIDS 14:1775–1784. Duggan MA, Boras VF, Inoue M, McGregor SE, Robertson DI. 1989. Human papillomavirus DNA determination of anal condylomata, dysplasias, and squamous carcinomas with in situ hybridization. Am J Clin Pathol 92:16–21. Dyson N, Howley PM, Munger K, Harlow E. 1989. The human papilloma virus-16 E7 oncoprotein is able to bind to the retinoblastoma gene product. Science 243:934–937. Ellerbrock TV, Chiasson MA, Bush TJ, et al. 2000. Incidence of cervical squamous intraepithelial lesions in HIV-infected women. JAMA 283: 1031–1037. Fairley CK, Sheil AG, McNeil JJ, et al. 1994. The risk of anogenital malignancies in dialysis and transplant patients. Clin Nephrol 41:101–105. Fenger C. 1987. The anal transitional zone. Acta Pathol Microbiol Immunol Scand Suppl 289:1–42. Fenger C. 1997. Anal canal. In: Sternberg SS, ed. Histology for Pathologists. Philadelphia: Lippincott-Raven Publishers, pp. 551–571. Fenger C, Frisch M, Jass JJ, Williams GT, Hilden J. 2000a. Anal cancer subtype reproducibility study. Virchows Arch 436:229–233. Fenger C, Frisch M, Marti MC, Parc R. 2000b. Tumours of the Anal Canal. In: Hamilton SR, Aaltonen LA, eds. World Health Organization Classification of Tumours. Pathology and Genetics of Tumours of the Digestive System. Lyon: IARC Press, pp. 147–155.

Anal Cancer Fisher G, Harlow SD, Schottenfeld D. 1997. Cumulative risk of second primary cancers in women with index primary cancers of uterine cervix and incidence of lower anogenital tract cancers, Michigan, 1985–1992. Gynecol Oncol 64:213–223. Franceschi S, Dal Maso L, Arniani S, et al. 1998. Risk of cancer other than Kaposi’s sarcoma and non-Hodgkin’s lymphoma in persons with AIDS in Italy. Cancer and AIDS Registry Linkage Study. Br J Cancer 78: 966–970. Friedman HB, Saah AJ, Sherman ME, et al. 1998. Human papillomavirus, anal squamous intraepithelial lesions, and human immunodeficiency virus in a cohort of gay men. J Infect Dis 178:45–52. Frisch M. 2002. On the etiology of anal squamous carcinoma. Dan Med Bull 49:194–209. Frisch M, Biggar RJ. 1999. Aetiological parallel between tonsillar and anogenital squamous-cell carcinomas. Lancet 354:1442–1443. Frisch M, Biggar RJ, Goedert JJ. 2000. Human papillomavirus-associated cancers in patients with human immunodeficiency virus infection and acquired immunodeficiency syndrome. J Natl Cancer Inst 92: 1500–1510. Frisch M, Fenger C, van den Brule AJ, et al. 1999a. Variants of squamous cell carcinoma of the anal canal and perianal skin and their relation to human papillomaviruses. Cancer Res 59:753–757. Frisch M, Glimelius B, van den Brule AJ, et al. 1998. Benign anal lesions, inflammatory bowel disease and risk for high-risk human papillomaviruspositive and -negative anal carcinoma. Br J Cancer 78:1534–1538. Frisch M, Glimelius B, van den Brule AJ, et al. 1997. Sexually transmitted infection as a cause of anal cancer. N Engl J Med 337:1350–1358. Frisch M, Glimelius B, Wohlfahrt J, Adami HO, Melbye M. 1999b. Tobacco smoking as a risk factor in anal carcinoma: An anti-estrogenic mechanism? J Natl Cancer Inst 91:708–715. Frisch M, Goodman MT. 2000. Human papillomavirus-associated carcinomas in Hawaii and the mainland U.S. Cancer 88:1464–1469. Frisch M, Johansen C. 2000. Anal carcinoma in inflammatory bowel disease. Br J Cancer 83:89–90. Frisch M, Melbye M, Møller H. 1993. Trends in incidence of anal cancer in Denmark. Br Med J 306:419–422. Frisch M, Olsen JH, Bautz A, Melbye M. 1994a. Benign anal lesions and the risk of anal cancer. N Engl J Med 331:300–302. Frisch M, Olsen JH, Melbye M. 1994b. Malignancies that occur before and after anal cancer: Clues to their etiology. Am J Epidemiol 140: 12–19. Frisch M, Smith E, Grulich A, Johansen C. 2003. Cancer in a population-based cohort of men and women in registered homosexual partnerships. Am J Epidemiol 157:966–972. Galloway DA, McDougall JK. 1996. The disruption of cell cycle checkpoints by papillomavirus oncoproteins contributes to anogenital neoplasia. Semin Cancer Biol 7:309–315. Garcia DE, Castro M, Romero GM, et al. 2000. Carcinoma anal en la enfermedad de Crohn. Gastroenterol Hepatol 23:150–151. Goedert JJ, Cote TR, Virgo P, et al. 1998. Spectrum of AIDS-associated malignant disorders. Lancet 351:1833–1839. Goldie SJ, Kuntz KM, Weinstein MC, Freedberg KA, Palefsky JM. 2000. Costeffectiveness of screening for anal squamous intraepithelial lesions and anal cancer in human immunodeficiency virus-negative homosexual and bisexual men. Am J Med 108:634–641. Goldie SJ, Kuntz KM, Weinstein MC, et al. 1999. The clinical effectiveness and cost-effectiveness of screening for anal squamous intraepithelial lesions in homosexual and bisexual HIV-positive men. JAMA 281:1822– 1829. Goldman S, Glimelius B, Nilsson B, Påhlman L. 1989. Incidence of anal epidermoid carcinoma in Sweden 1970–1984. Acta Chir Scand 155:191– 197. Goldman S, Skoog L, Wilking N. 1992. Immunocytochemical analysis of receptors for estrogen and progesterone in fine needle aspirates from anal epidermoid carcinoma. Dis Colon Rectum 35:163–165. Grulich AE, Wan X, Law MG, Coates M, Kaldor JM. 1999. Risk of cancer in people with AIDS. AIDS 13:839–843. Gupta S, Sharma BK. 1998. Immunoexpression of mutant p53 and human papillomavirus related E6 oncoprotein in anal malignancies. Indian J Exp Biol 36:875–878. Halpert R, Fruchter RG, Sedlis A, et al. 1986. Human papillomavirus and lower genital neoplasia in renal transplant patients. Obstet Gynecol 68: 251–258. Harro CD, Pang YY, Roden RB, et al. 2001. Safety and immunogenicity trial in adult volunteers of a human papillomavirus 16 L1 virus-like particle vaccine. J Natl Cancer Inst 93:284–292. Hemminki K, Dong C. 2000. Cancer in husbands of cervical cancer patients. Epidemiology 11:347–349.

839

Heselmeyer K, du Manoir S, Blegen H, et al. 1997. A recurrent pattern of chromosomal aberrations and immunophenotypic appearance defines anal squamous cell carcinomas. Br J Cancer 76:1271–1278. Holly EA, Whittemore AS, Aston DA, et al. 1989. Anal cancer incidence: Genital warts, anal fissure or fistula, hemorrhoids, and smoking. J Natl Cancer Inst 81:1726–1731. Holm R, Tanum G, Karlsen F, Nesland JM. 1994. Prevalence and physical state of human papillomavirus DNA in anal carcinomas. Mod Pathol 7:449–453. Holmes F, Borek D, Owen-Kummer M, et al. 1988. Anal cancer in women. Gastroenterology 95:107–111. Howard LC, Paterson-Brown S, Weber JN, et al. 1986. Squamous carcinoma of the anus in young homosexual men with T helper cell depletion. Genitourin Med 62:393–395. Indinnimeo M, Cicchini C, Stazi A, et al. 1999. Human papillomavirus infection and p53 nuclear overexpression in anal canal carcinoma. J Exp Clin Cancer Res 18:47–52. International Agency for Research on Cancer. 1995. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Vol. 64. Human papillomaviruses. 1st ed. Lyon, IARC. International Agency for Research on Cancer. 2002. Cancer incidence in five continents, Volume VIII. Lyon, IARC. Jæger AB, Gramkow A, Sørensen P, et al. 2000. Correlates of heterosexual behavior among 23–87 year olds in Denmark and Sweden, 1992–1998. Arch Sex Behav 29:91–106. Jensen SL, Hagen K, Shokouh-Amiri MH, Nielsen OV. 1987. Does an erroneous diagnosis of squamous-cell carcinoma of the anal canal and anal margin at first physician visit influence prognosis? Dis Colon Rectum 30:345–351. Kazal HL, Sohn N, Carrasco JI, Robilotti JG, Delaney WE. 1976. The gay bowel syndrome: Clinico-pathologic correlation in 260 cases. Ann Clin Lab Sci 6:184–192. Keyes EL. 1937. Squamous cell carcinoma of the lower rectum and anus. Ann Surg 106:1046–1058. Kjær SK, de Villiers EM, Dahl C, et al. 1991. Case-control study of risk factors for cervical neoplasia in Denmark. I: Role of the “male factor” in women with one lifetime sexual partner. Int J Cancer 48:39–44. Kline RJ, Spencer RJ, Harrison EG. 1964. Carcinoma associated with fistulain-ano. Arch Surg 89:989–994. Koblin BA, Hessol NA, Zauber AG, et al. 1996. Increased incidence of cancer among homosexual men, New York City and San Francisco, 1978–1990. Am J Epidemiol 144:916–923. Koch A, Hansen SV, Nielsen NM, Palefsky J, Melbye M. 1997. HPV detection in children prior to sexual debut. Int J Cancer 73:621–624. Koutsky LA, Ault KA, Wheeler CM, et al. 2002. A controlled trial of a human papillomavirus type 16 vaccine. N Engl J Med 347:1645– 1651. Ky A, Sohn N, Weinstein MA, Korelitz BI. 1998. Carcinoma arising in anorectal fistulas of Crohn’s disease. Dis Colon Rectum 41:992–996. Lacey HB, Wilson GE, Tilston P, et al. 1999. A study of anal intraepithelial neoplasia in HIV positive homosexual men. Sex Transm Infect 75:172–177. Li FP, Osborn D, Cronin CM. 1982. Anorectal squamous carcinoma in two homosexual men. Lancet 2:391. Lin AY, Gridley G, Tucker M. 1995. Benign anal lesions and anal cancer. N Engl J Med 332:190–191. Lorenz HP, Wilson W, Leigh B, Crombleholme T, Schecter W. 1991. Squamous cell carcinoma of the anus and HIV infection. Dis Colon Rectum 34:336–338. Lumley JW, Stitz RW. 1991. Crohn’s disease and anal carcinoma: An association? A case report and review of the literature. Aust N Zeal J Surg 61:76–77. Marks CG. 1990. Anal lesions in Crohn’s disease. Ann R Coll Surg Engl 72:158–159. Melbye M, Cote TR, Kessler L, Gail M, Biggar RJ. 1994a. High incidence of anal cancer among AIDS patients. The AIDS/Cancer Working Group. Lancet 343:636–639. Melbye M, Palefsky J, Gonzales J, et al. 1990. Immune status as a determinant of human papillomavirus detection and its association with anal epithelial abnormalities. Int J Cancer 46:203–206. Melbye M, Rabkin C, Frisch M, Biggar RJ. 1994b. Changing patterns of anal cancer incidence in the United States, 1940–1989. Am J Epidemiol 139:772–780. Melbye M, Smith E, Wohlfahrt J, et al. 1996. Anal and cervical abnormality in women—prediction by human papillomavirus tests. Int J Cancer 68:559– 564. Melbye M, Sprøgel P. 1991. Aetiological parallel between anal cancer and cervical cancer. Lancet 338:657–659.

840

PART IV: CANCER BY TISSUE OF ORIGIN

Moscicki AB, Hills NK, Shiboski S, et al. 1999. Risk factors for abnormal anal cytology in young heterosexual women. Cancer Epidemiol Biomarkers Prev 8:173–178. Mougin C, Bernard C, Schaal JP, et al. 1995. Detection of human cytomegalovirus in female and male anogenital lesions with papillomavirus-associated dysplastic changes. Cancer Detect Prev 19:316–324. Murakami M, Gurski KJ, Steller MA. 1999. Human papillomavirus vaccines for cervical cancer. J Immunother 22:212–218. Noffsinger AE, Hui YZ, Suzuk L, et al. 1995. The relationship of human papillomavirus to proliferation and ploidy in carcinoma of the anus. Cancer 75:958–967. Oettling G, Franz HB. 1998. Mapping of androgen, estrogen and progesterone receptors in the anal continence organ. Eur J Obstet Gynecol Reprod Biol 77:211–216. Ogunbiyi OA, Scholefield JH, Raftery AT, et al. 1994. Prevalence of anal human papillomavirus infection and intraepithelial neoplasia in renal allograft recipients. Br J Surg 81:365–367. Owen WFJR. 1980. Sexually transmitted diseases and traumatic problems in homosexual men. Ann Intern Med 92:805–808. Palefsky JM. 1998. Human papillomavirus infection and anogenital neoplasia in human immunodeficiency virus-positive men and women. J Natl Cancer Inst Monogr:15–20. Palefsky JM, Holly EA, Gonzales J, et al. 1991. Detection of human papillomavirus DNA in anal intraepithelial neoplasia and anal cancer. Cancer Res 51:1014–1019. Palefsky JM, Holly EA, Ralston ML, et al. 1998a. Anal squamous intraepithelial lesions in HIV-positive and HIV-negative homosexual and bisexual men: Prevalence and risk factors. J AIDS Hum Retrovirol 17:320–326. Palefsky JM, Holly EA, Ralston ML, et al. 2001a. Effect of highly active antiretroviral therapy on the natural history of anal squamous intraepithelial lesions and anal human papillomavirus infection. J AIDS 28: 422–428. Palefsky JM, Holly EA, Ralston ML, da Costa M, Greenblatt RM. 2001b. Prevalence and risk factors for anal human papillomavirus infection in human immunodeficiency virus (HIV)-positive and high-risk HIVnegative women. J Infect Dis 183:383–391. Palefsky JM, Holly EA, Ralston ML, Jay N. 1998b. Prevalence and risk factors for human papillomavirus infection of the anal canal in human immunodeficiency virus (HIV)-positive and HIV-negative homosexual men. J Infect Dis 177:361–367. Palefsky JM, Holly EA, Ralston ML, et al. 1998c. High incidence of anal highgrade squamous intra-epithelial lesions among HIV-positive and HIVnegative homosexual and bisexual men. AIDS 12:495–503. Parkin DM, Wabinga H, Nambooze S, Wabwire-Mangen F. 1999. AIDS-related cancers in Africa: Maturation of the epidemic in Uganda. AIDS 13:2563–2570. Paz IB, Cook N, Odom-Maryon T, Xie Y, Wilczynski SP. 1997. Human papillomavirus (HPV) in head and neck cancer. An association of HPV 16 with squamous cell carcinoma of Waldeyer’s tonsillar ring. Cancer 79:595–604. Penn I. 1986. Cancers of the anogenital region in renal transplant recipients. Analysis of 65 cases. Cancer 58:611–616. Peters RK, Mack TM. 1983. Patterns of anal carcinoma by gender and marital status in Los Angeles County. Br J Cancer 48:629–636. Poletti PA, Halfon A, Marti MC. 1998. Papillomavirus and anal carcinoma. Int J Colorect Dis 13:108–111. Rabkin CS, Biggar RJ, Melbye M, Curtis RE. 1992. Second primary cancers following anal and cervical carcinoma: Evidence of shared etiologic factors. Am J Epidemiol 136:54–58. Rosser C. 1931. The etiology of anal cancer. Am J Surg 11:328–333.

Scholefield JH, Northover JM, Carr ND. 1990a. Male homosexuality, HIV infection and colorectal surgery. Br J Surg 77:493–496. Scholefield JH, Sonnex C, Talbot IC, et al. 1989. Anal and cervical intraepithelial neoplasia: Possible parallel. Lancet 2:765–769. Scholefield JH, Thornton JH, Cuzick J, Northover JM. 1990b. Anal cancer and marital status. Br J Cancer 62:286–288. Schwartz SM, Daling JR, Doody DR, et al. 1998. Oral cancer risk in relation to sexual history and evidence of human papillomavirus infection. J Natl Cancer Inst 90:1626–1636. Sherman ME, Schiffman MH, Strickler H, Hildesheim A. 1998. Prospects for a prophylactic HPV vaccine: Rationale and future implications for cervical cancer screening. Diagn Cytopathol 18:5–9. Shroyer KR, Brookes CG, Markham NE, Shroyer AL. 1995. Detection of human papillomavirus in anorectal squamous cell carcinoma. Correlation with basaloid pattern of differentiation. Am J Clin Pathol 104:299–305. Sitas F, Bezwoda WR, Levin V, et al. 1997. Association between human immunodeficiency virus type 1 infection and cancer in the black population of Johannesburg and Soweto, South Africa. Br J Cancer 75:1704–1707. Slater G, Greenstein A, Aufses AH, Jr. 1984. Anal carcinoma in patients with Crohn’s disease. Ann Surg 199:348–350. Stern BD, Kaplan L. 1969. Multicentric foci of carcinomas arising in structures of cloacal origin. Am J Obstet Gynecol 104:255–266. Unger ER, Vernon SD, Lee DR, et al. 1997. Human papillomavirus type in anal epithelial lesions is influenced by human immunodeficiency virus. Arch Pathol Lab Med 121:820–824. Vincent-Salomon A, de la Rochefordière A, Salmon R, et al. 1996. Frequent association of human papillomavirus 16 and 18 DNA with anal squamous cell and basaloid carcinoma. Mod Pathol 9:614–620. Walgenbach S, Junginger T, Muller JM. 1984. Plattenepithelkarzinom des anorektalen Überganges bei Morbus Crohn mit langjährigen analem Fistelleiden. Dtsch Med Wochenschr 109:618–621. Werness BA, Levine AJ, Howley PM. 1990. Association of human papillomavirus types 16 and 18 E6 proteins with p53. Science 248:76–79. Wexner SD, Milsom JW, Dailey TH. 1987. The demographics of anal cancers are changing. Identification of a high-risk population. Dis Colon Rectum 30:942–946. Wilczynski SP, Lin BT, Xie Y, Paz IB. 1998. Detection of human papillomavirus DNA and oncoprotein overexpression are associated with distinct morphological patterns of tonsillar squamous cell carcinoma. Am J Pathol 152:145–156. Winkelstein W, Jr. 1990. Smoking and cervical cancer—current status: A review. Am J Epidemiol 131:945–957. World Health Organization. 1990. International Classification of Diseases for Oncology. 2nd ed. Geneva, Switzerland. Wu TC, Kurman RJ. 1997. Analysis of cytokine profiles in patients with human papillomavirus-associated neoplasms. J Natl Cancer Inst 89:185–187. Xi LF, Critchlow CW, Wheeler CM, et al. 1998. Risk of anal carcinoma in situ in relation to human papillomavirus type 16 variants. Cancer Res 58:3839–3844. Youk EG, Ku JL, Park JG. 2001. Detection and typing of human papillomavirus in anal epidermoid carcinomas: Sequence variation in the E7 gene of human papillomavirus type 16. Dis Colon Rectum 44:236–242. Zaki SR, Judd R, Coffield LM, et al. 1992. Human papillomavirus infection and anal carcinoma. Retrospective analysis by in situ hybridization and the polymerase chain reaction. Am J Pathol 140:1345–1355. Zucchini C, Strippoli P, Rosati G, et al. 2000. Expression analysis and mutational screening of the epithelium-specific ets gene-1 (ESE-1) in patients with squamous anal cancer. Int J Oncol 17:265–270. zur Hausen H. 1976. Condylomata acuminata and human genital cancer. Cancer Res 36:794.

44

The Leukemias MARTHA S. LINET, SUSAN S. DEVESA, AND GARETH J. MORGAN

T

he leukemias arise from malignant transformation of hematopoietic stem or progenitor cells that originate in the bone marrow, lymph nodes, and/or other lymphoid tissue with immune function. A small pool of stem cells, which persists throughout an individual’s lifetime, differentiates to early precursors, then divides into multiple subtypes, and ultimately produces large numbers of end-stage cells of myeloid and lymphoid lineage progeny. Because the effector or end-stage cells of each lineage have a finite lifespan and, therefore, cannot transmit mutations, all leukemias are the consequence of derangements of growth and differentiation of the pluripotential early precursors of myeloid or lymphoid progeny (Hoffman et al., 2000). The rarity of individual leukemia subtypes and limitations of earlier classifications have hampered elucidation of risk factors despite rapid advances in the molecular and immunologic underpinnings of leukemogenesis. Earlier reviews extensively discussed the history of leukemia classification and the results of epidemiologic studies prior to the 1990s (Linet, 1985; Finch and Linet, 1992; Linet and Cartwright, 1996; Sgambati et al., 2001; Linet and Devesa, 2002). The landmark French-American-British (FAB) classification (Bennett et al., 1976, 1982, 1985, 1989, 1994) achieved international consensus on morphologic criteria. Subsequent efforts to incorporate developmental and functional aspects of hematopoiesis according to lineage as well as key aspects of pathogenesis, and cytogenetic and immunophenotypic characteristics (Head, 1996; Harris, 1999; McKenna, 2000; Bennett, 2000) culminated in the World Health Organization (WHO) classification of neoplastic and related hematopoietic and lymphoid tissue diseases (Jaffe et al., 2001). The WHO classification was incorporated in the third revision of the International Classification for Diseases in Oncology (ICD-O-3) (Fritz et al., 2000). Initial descriptions of leukemia mortality and incidence were restricted to all forms of leukemia combined (hereafter designated total leukemia). Because death certificates and early population-based cancer registries often did not designate leukemia subtype, evaluation and comparison of long-term trends in leukemia occurrence across populations is restricted to total leukemia. Internationally, the highest age-standardized total leukemia mortality rates occurred in populations of Western Europe, Oceania, North America, Israel, and Costa Rica, where rates generally ranged from 4.8 to 7.4/100,000 personyears for males and from 3.2 to 4.6/100,000 for females (Aoki et al., 1992). Lower rates (ranging from 3.7 to 4.5 for males and from 2.8 to 3.5 for females) characterized populations in Latin America and Asia. In the United States, the highest age-standardized total leukemia mortality rates have consistently been apparent in the north and southcentral regions for whites of both sexes from the 1950s to the present (Devesa et al., 1999) (see more recent data in Figs. 44–1A and 44–1B). Mortality data are too sparse in the north and south-central regions of the United States to evaluate patterns for African Americans. While age-standardized total leukemia mortality has modestly declined among whites and increased slightly among nonwhites (the latter comprising African Americans, Asian-Pacific Islanders, and Native Americans) since the early 1950s (data not shown), the patterns have differed by age (Fig. 44–2). Rates have consistently declined since the early 1960s for children and adolescents (aged 0–19) of both racial groups, declined less rapidly among young adults (ages 20–44 years), changed little among middle-aged adults (aged 45–64), and increased

among the elderly (aged 65 and older), although the rate of increase slowed substantially after 1960, perhaps reflecting improvements in medical care in the elderly. The notable decline in mortality for children and adolescents and the increase for the elderly that leveled off during the 1960s and 1970s were also evident in other Caucasian populations in developed countries (Kinlen, 1994). Internationally, there is a distinct racial and geographic gradient for total leukemia incidence, with highest rates in white and Hispanic white populations of North America and in Oceania, Italy, and the United Kingdom (Fig. 44–3A). Mid-level rates are seen among African Americans, Israeli Jews, and populations in Denmark, France, Sweden, Spain, and Switzerland. Lowest rates occur in descending order in Japanese in Osaka, Chinese in Shanghai, and Indians in Bombay. Internationally, total leukemia incidence rates have been relatively stable since the 1960s for children, adolescents, young adults, and middle-aged persons, but rates rose among the elderly from the 1960s to the early 1970s (Draper et al., 1994; Kinlen, 1994). The pattern in the elderly may be consistent with increasing diagnostic ascertainment. In the geographic regions covered by the US Surveillance, Epidemiology and End Results (SEER) program, agestandardized rates for total leukemia are higher for males of both races than for females, and within each gender higher for whites than for African Americans (Fig. 44–4). Among whites, incidence changed little between 1973–1979 and 1994–2000. Incidence among African Americans rose between 1973–1979 and 1980–1986, and then declined slowly (Ries et al., 2003). Age-specific rates slightly increased for children under age 15, mostly due to an unexplained abrupt jump in rates between 1983 and 1984 (data not shown) (Linet et al., 1999). For leukemia patients younger than age 65 years at diagnosis, there was little overall change for whites during 1973–2000 and a small decline for African Americans (the latter greater for females than males). Declines were seen during 1973–2000 among leukemia patients diagnosed at age 65 years or older (Ries et al., 2003). The four major categories of leukemia considered in this chapter are acute myeloid (including acute monocytic), chronic myeloid, acute lymphocytic, and chronic lymphocytic leukemia. To interpret variation in incidence according to subtype, it is important to consider leukemia subtypes with other designations and leukemias with subtype not specified as a proportion of total leukemia. For the registries shown in Figure 44–3A, other and not otherwise specified leukemias comprise 7.0%–31.5% of total leukemia. Populations with lowest rates of other and not otherwise specified (e.g., <10%) include Denmark and Italy (Varese); populations with mid-level rates (e.g., 10%–19%) include all in North America, India (Bombay), all European populations shown except Denmark, and the Oceania registries; and populations with 20% or more of other and not otherwise specified include Israel (Jews), China (Shanghai) and Japan (Osaka). In the United States, other and not othewise specified leukemia currently comprises 7.8% of the total, more than half of which (4.7%) are other acute leukemias. Rates for other and not otherwise specified leukemias, similar to total leukemia rates in the United States, are higher for males than for females of both races, higher for white than for African American males, but similar for white and African American females. Other and not otherwise specified leukemias declined for all four race-gender groups during 1973–1979 through 1994–2000, with an acceleration of the rate of decline between 1987–1993 and 1994–2000 for whites and a steeper decline between 1980–1986 and 1987–1993 for African-

841

842

PART IV: CANCER BY TISSUE OF ORIGIN

US= 10.74/100,000 12.51 - 16.43 11.37 - 12.50 10.47 - 11.36 9.43 - 10.46 7.38 - 9.42 (a) White Males

Figure 44–1. United States leukemia mortality rates for white males (a) and females (b) by state economic area (age-adjusted 2000 US population), 1980–1999. (Source: Devesa, et al. Atlas of Cancer Mortality in the

United States. Bethesda, MD: National Institutes of Health, National Cancer Institute, NIH Publication No. 99-4564, 1999. Updated data: http://www3.cancer.gov/atlasplus/.)

American females (Fig. 44–4). This may reflect improving specificity of the diagnoses, which should be kept in mind when discussing the type-specific patterns. This review is organized into two major components corresponding to the myeloid and lymphoid lineage origin of the leukemias. Although most epidemiological studies are retrospective and use earlier classifications, where possible, this chapter tries to present epidemiological data in a context that broadly matches more current classifications.

may all transform to AML. Although morphologically similar in appearance, AML may arise de novo or following a myelodysplastic or myeloproliferative state. The antecedent history of AML is taken into acount in the WHO classification in the designations of secondary AML (sAML), which arises following a prior MDS phase, and that of therapy-related AML (tAML), which arises following treatment with leukemogenic agents. Therapy-related MDS and tAML are part of the same disease spectrum (Hayashi, 2000).

MYELOID LEUKEMIAS AND RELATED DISORDERS Origin, Progeny, Subtypes, and Inter-Relationships of Myeloid Leukemias and Related Disorders Myeloid leukemias and related disorders originate in pluripotential precursor cells that normally give rise to red blood cells, polymorphonuclear neutrophils, monocytes, and platelets. Disruptions of the normal hierarchy of myeloid maturation result in hematologic disorders characterized by either excesses or deficiencies of the mature effector cells (Knowles et al., 1992). The disorders of myeloid origin include acute myeloid leukemia (AML), myelodysplastic syndromes (MDS), and myeloproliferative disorders (MPD) (the latter including chronic myeloid leukemia (CML). When maturation is blocked by genetic changes, the result is transformation to AML. MDS and MPD, in contrast, are the clinical consequences of disordered, but relatively complete maturation (Hayashi, 2000). MDS, CML, and the other MPD

Characteristics and Pathogenesis of Myeloid Leukemias and Related Disorders Characteristic Features Acute myeloid leukemias are classified by the World Health Organization (WHO) (Jaffe et al., 2001) into four categories: AML with recurrent cytogenetic abnormalities (including t(8;21), t(15;17), or inv(16) ); AML with multilineage dysplasia (prior history of MDS or de novo AML including the cytogenetic variants 5q- and 7q-; AML/MDS therapy- or occupation-related (including those with the acquired chromosomal abnormalities of 5q-, 7q-, and 11q23, t(8;21) rearrangements and even t(15;17) ); and AML not otherwise categorized. MDS-related AML includes most AML in the elderly, AML following MDS, AML among individuals with Fanconi anemia, and a small fraction (10%–15%) in younger patients. The extent of differentiation determines the morphologic appearance of the AML subtypes.

843

The Leukemias

US = 6.21/100,000 6.95 - 8.52 6.37 - 6.92 5.87 - 6.36 5.01 - 5.86 3.60 - 5.00 (b) White Females

Figure 44–1. (cont.)

Myelodysplastic syndromes are characterized by bone marrow hyperplasia and peripheral cytopenias, and morphologically recognizable abnormal differentiation. The FAB classification recognized five distinct categories of MDS in adults (Bennett et al., 1982), with definitions (for both MDS and AML) based on the presence of trilineage dysplasia and the percentage of blasts. The WHO classification for MDS overlaps with the FAB classification, but also includes an MDS category characterized by an important cytogenetic abnormality (5q deletion) and therapy-related MDS (Jaffe et al., 2001; Nosslinger et al., 2001). Childhood MDS subtypes are more heterogeneous, and thus more difficult to classify (Groupe Francais de Cytogenetique Hematologique, 1997; Mandel et al., 2002). Myeloproliferative disorders are associated with bone marrow hyperplasia and an excess of differentiated progeny, and include polycythemia vera (composed of progenitors of red blood cells), primary (essential) thrombocythemia (the progeny of platelets), chronic myeloid leukemia (the progeny of myeloid cells), and chronic myelomonocytic leukemia (the progeny of monocytes) (Gordon et al., 1999; Hoffman et al., 2000). CML results from the transformation of a hemopoietic stem cell that retains the capacity to differentiate, but differentiates abnormally, with an excess of committed precursors and their differentiated progeny. After a median interval of about 4 years, CML transforms to an acute leukemic phase, designated “blast crisis.”

Subtypes Defined by Cytogenetic, Numerical, or Other Recurrent Karyotypic Aberrations Recurrent Cytogenetic Abnormalities. Non-randomly occurring cytogenetic changes identified in AML, MDS, and CML

include the Philadelphia chromosome (Ph, a consequence of a reciprocal translocation in which the ABL oncogene from chromosome 9 is transposed to chromosome 22 within the breakpoint cluster region of a gene, BCR), characterizing more than 90% of CML (Hoffman et al., 2000), 30% of ALL (Faderl et al., 2000), and 1%–4% of AML (Paietta et al., 1998). The majority of patients with AML and MDS have acquired chromosome aberrations (Aul et al., 1995; Mrozek et al., 1997; Willman, 1999), including balanced translocations (such as t(15;17), t(8;21), and inv(16) ); partial deletions or loss of whole chromosomes (such as 5q or 7q); and the numerical forms (such as trisomy 21 and trisomy 8). For the most part, the timing of the occurrence of the acquired translocation is unknown, although recent evidence supports the prenatal origin of t(8;21) AML1-ETO translocations in childhood AML, based on demonstration of this translocation in neonatal Guthrie blood spots (Wiemels et al., 2002a). The molecular mechanisms responsible for translocations may give important clues about environmental exposures, for example, the binding sites for topoisomerase II inhibitors at the 11q23 breakpoints (Broeker et al., 1996).

Numerical and Other Molecular Abnormalities. Trisomy 21, seen in Down syndrome, the most common congenital abnormality affecting newborns, is one of the most commonly observed numerical cytogenetic abnormalities associated with AML. Children with Down syndrome are at 10- to 40-fold increased risk of developing acute leukemia (Neglia and Robison, 1988; Hill et al., 2003). More recently, GATA1 mutations have been described in leukemias occurring with Down syndrome (Mundschau et al., 2003). Acquired trisomy 21 may occur in patients with AML or ALL who do

844

PART IV: CANCER BY TISSUE OF ORIGIN based specialist registry of hematologic diseases in the United Kingdom; however, significant decreases, estimated at 3% per year, occurred among the elderly (Cartwright et al., 1997). This decline in AML among the elderly paralleled a highly significant increase of 6% per year for myelodysplastic syndromes among older persons. It is likely that the increasing recognition of MDS during the same period may have resulted in less misclassification of these disorders as AML in this region of the United Kingdom. Age-standardized incidence rates for CML are higher in males than in females of both races in the United States, with rates higher in African Americans than in whites within each gender, unique among the leukemias (Fig. 44–4). Incidence among whites has remained fairly stable over time (Fig. 44–4). Among African Americans, agestandardized incidence rose between 1973–1979 and 1987–1993 for males and between 1973–1979 and 1980–1986 for females, then declined more rapidly among females than among males.

Age-Specific Patterns by Gender and Race Figure 44–2. United States trends in leukemia mortality (age-adjusted, world standard) by age group among whites and nonwhites, 1950–1954 to 1995–2000. (Source: Based on data from the US National Center for Health Statistics.)

not have Down syndrome (Rowley, 2000). Trisomy 8, linked with MDS, may predispose to a higher risk of acute leukemic transformation (Sole et al., 2000), perhaps suggesting overlapping pathogenesis for MDS and AML. Other molecular abnormalities seen in AML include RAS mutations (Santon et al., 1995), mutations of P53 (Nakano et al., 2000), microsatellite instability (Zhu et al., 1999), and endoduplications of the Flt 3 receptor (Kohler et al., 2000).

The Descriptive Epidemiology of Myeloid Leukemias and Related Disorders International Patterns In general, for both AML and CML, similar to the pattern for total leukemia and for other and not otherwise specified leukemias, ageadjusted incidence rates are lower for females than males. The geographic patterns for AML are often similar for males and females, and rates vary internationally threefold (Fig. 44–3B). Highest AML rates occur in both sexes in white populations in North America, Oceania, northern and western Europe, and in Hispanics in Los Angeles. Consistent with the elevated incidence rate in Los Angeles Hispanics is an excess AML characterized by the t(15;17) translocation, which has been described among adults of Hispanic origin (Tomas et al., 1996; Douer et al., 1996). Mid-level AML rates are observed in both sexes among African Americans, Israeli Jews, and populations in France (Isere), Spain (Taragona), and Sweden; among males in Japan; and in females among Los Angeles Hispanic whites, Italians (Romagna and Varese), and Swiss (Zurich). Lowest AML rates are seen in Asians, particularly Chinese in Shanghai and Indians in Bombay (Fig. 44–3B). Chronic myeloid leukemia incidence rates generally are lower than AML rates, and they vary internationally threefold among men and twofold among women (Fig. 44–3B). Highest CML rates are in Italian males, and in Australians, African Americans, and Los Angeles Hispanics of both sexes; lowest rates are in Asians (Fig. 44–3B). Limited population-based incidence data for MDS and MPD preclude international comparisons.

Trends in Age-Adjusted Patterns by Gender and Race Age-standardized rates for AML (including acute monocytic leukemia) are modestly higher in males of both races than in females; within gender, rates generally are higher in whites than in African Americans (Fig. 44–4). Incidence was fairly stable during 1973–1979 through 1987–1993, but rose somewhat in all four groups between 1987–1993 and 1994–2000. A stable incidence trend was seen for adults aged 15–64 years old in the regions covered by a population-

In the United States, myeloid and monocytic leukemias comprise 44.7% of total leukemia, with acute myeloid and monocytic leukemia accounting for 28.4%, chronic myeloid leukemia for 13.8%, and other myeloid and monocytic leukemia for 2.4%. As shown in Figure 44–5, age-specific AML incidence rates in the United States peak slightly in infancy, then decline until age 10 after which incidence rates increase. After age 40, incidence rates rise more rapidly until approximately age 70; among older persons, rates increase more slowly. During infancy, AML rates are higher in females than in males, and higher in whites than in African Americans. In childhood through early adulthood, AML incidence rates are similar in both sexes, but higher in African Americans than in whites. Among middle-aged persons, AML rates are similar by sex and race, whereas among the elderly incidence rates increase more rapidly in males than in females of both races, and in whites than African Americans for persons of both sexes (Fig. 44–5). Population-based data from the United Kingdom reveal low and flat incidence for MDS, all types combined, during childhood and early adolescence, followed by a linear increase during later adolescence, then an exponential rise after ages 50–60 (data not shown) (Cartwright et al., 1997). MDS incidence rates are higher for males than for females during childhood, and higher for females than males from adolescence to age 50; after age 50, rates for males rise exponentially, surpassing rates in females at older ages. CML is rare among children except for a peak in very early childhood among white boys (Fig. 44–5). At age 10, CML incidence begins to rise log linearly to age 30; subsequently, the rate of increase slows substantially with increasing age. CML rates are consistently higher in males than in females throughout life, and higher in African Americans than whites of both sexes until age 70. Among the elderly, rates for African Americans begin to flatten with increasing age, whereas rates for elderly whites continue to increase linearly. For persons 80 years and older, CML incidence is highest among white males, whereas rates converge for African Americans of both sexes and white females (Fig. 44–5). In the United Kingdom, rates for all MPD combined are very low and stable until about age 10, then increase exponentially, followed by a slowing of the increase at age 30; subsequently, rates continue to rise linearly with increasing age (data not shown) (Cartwright et al., 1997). MPD rates are higher in males than females during childhood, similar in both sexes from adolescence until age 50, and slightly higher in males than in females from middle age onward. There are few descriptive epidemiological investigations focusing on cytogenetically defined subtypes of the myeloid malignancies and related disorders, but some data from the United Kingdom suggest that leukemias characterized by deletions increase notably with age, whereas leukemias characterized by balanced translocations remain constant with increasing age (Rossi et al., 2000; Moorman et al., 2001; Moorman et al., 2002a).

Survival: US Population-Based Data For patients with AML, prognosis is poorer than for those with other leukemia subtypes (Table 44–1). Based on US SEER program data,

The Leukemias

845

(A)

Figure 44–3. A, International variation in leukemia incidence rates per 100,000 (age-adjusted, world standard) by continent and sex: Total leukemia and other and unspecified leukemia, circa 1993–1997. (Source. Parkin, et al. Cancer Incidence in Five Continents, vol. 8. Lyon, France: IARC Scientific Publication Number 155, 2002.) B, International variation

in leukemia incidence rates per 100,000 (age-adjusted, world standard) by continent and sex: Four specific cell types, circa 1993–1997. (Source: Parkin, et al. Cancer Incidence in Five Continents, vol. 8. Lyon, France: IARC Scientific Publication Number 155, 2002.)

5-year relative survival for patients with AML improved significantly from 6.0% in 1974–1976 to 18.7% in 1992–1999. Similar improvements in survival were apparent for males and females, and for whites and African Americans. When survival data are evaluated according to age at diagnosis, improvements in 5-year relative survival are most apparent for AML in children, increasing from 14.3% in 1974–1976 to 47.3% in 1992–1999. For childhood AML patients, higher survival is apparent for children ages 5–9 at diagnosis, compared with that for either younger or older children, and for females than males (Ries et al., 2003). Compared with the 18.7% 5-year relative survival observed in 1992–1999 among all patients with AML, 5-year relative survival was substantially higher among younger patients than among the elderly (3.9%), but similar among males and females, and among whites and African Americans of all ages (Table 44–1). CML is often indolent initially, with rising blood cell counts sometimes resulting in early mortality from congestive heart failure or stroke due to hyperviscosity. Between 1974–1976 and 1992–1999, 5year relative survival rose significantly from 22.6% to 34.9% (Table

44–1). Five-year relative survival was similar among white males and females, somewhat lower for African American males, but somewhat higher for African American females, although the rates should be interpreted cautiously due to relatively small numbers of patients (Table 44–1). Survival rates are higher among patients diagnosed before age 65 (45.9%) compared with those ages 65 or older at diagnosis (21.2%).

Analytical Epidemiological Studies Because MDS and AML are often evaluated together or sometimes considered as a single entity, and there are few epidemiological studies focusing solely on MDS, the analytical studies for both conditions are summarized together in the following section. The MPD most comprehensively evaluated in epidemiological studies is CML; there have been relatively few investigations of polycythemia vera, primary thrombocythemia, or chronic myelomonocytic leukemia. Therefore, the analytical studies for these related disorders are described below in the section on CML and Other Myeloproliferative Disorders.

846

PART IV: CANCER BY TISSUE OF ORIGIN

(B)

Figure 44–3. (cont.)

ACUTE MYELOID LEUKEMIA AND MYELODYSPLASTIC SYNDROMES Physical Agents—Ionizing Radiation Mechanism of Leukemogenesis Ionizing radiation, the most well-studied risk factor for the leukemias, is a clastogen that deposits energy at random in tissues, and induces DNA strand breaks (Little, 1993). Strand breaks, which may occur directly from radiation or indirectly from generation of oxygen free radicals, can lead to chromosomal translocations and deletions. Susceptibility to radiation-induced malignancy appears to vary widely among different subsets of marrow-derived cells, thus highlighting the need to consider lineage-specific developmental processes as well as radiation dose, type, and relative biologic effectiveness (e.g., the ability to produce a given adverse effect) (Boice et al., 1996). There is no goldstandard biological measure of radiation dose, but quantitative measurement of chromosome aberrations in peripheral blood lymphocytes using fluorescent in situ hybridization (FISH) may be a correlate for the initiating radiogenic lesions leading to radiation-induced leukemia (Tucker et al., 1997). Studies of the Japanese atomic bomb survivors using FISH have shown several abnormalities that may represent mechanisms by which radiation causes MDS or AML, including significantly higher incidence of subclones with monosomy 7 and deletion of the 20q13.2 region among AML/MDS cases exposed to ≥1 Gray (1 Gy = 100 rads) compared with non-exposed de novo AML cases (Nakanishi

et al., 1999). There was also a high prevalence of point mutations of the AML1 gene, a transcriptional activator essential for normal hematopoietic development, which has been posited as a target for radiationinduced AML (Hromas et al., 2000; Harada et al., 2003). Atomic bomb survivors who developed MDS had increased mutations in TP53 (Imamura et al., 2002). In addition, clinically asymptomatic atomic bomb survivors had a high prevalence of neutropenia, due to NK or NK-like T-cell proliferative disorders (Imamura and Kimura, 2000). Leukemia outcomes reported in epidemiological studies of radiation-exposed populations often include all leukemias combined for mortality because subtypes are frequently not provided on death certificates. The leukemia category most commonly associated with many forms of moderate-to-high radiation exposure is AML (Boice et al., 1996). Results from the follow-up study of atomic bomb survivors (Preston et al., 1994), and, to a lesser extent, epidemiological data from other radiation-exposed populations (Boice et al., 1996), also reveal increased risks of CML and acute lymphocytic leukemia (ALL) associated with moderate-to-high levels of radiation exposure. The epidemiological findings for all leukemias other than chronic lymphocytic leukemia (hereafter designated non-CLL) are described in this section because AML, CML, and ALL are often considered as a combined entity and radiation-related risks are not ascertained or quantified for the individual leukemia entities; the age-standardized incidence of AML is higher than that for CML or ALL, and AML has been more frequently associated with ionizing radiation exposure than the other leukemia subtypes.

The Leukemias

847

Figure 44–4. United States trends in leukemia incidence (age-adjusted, world standard) by race and sex for total leukemia and by cell type in nine SEER areas, 1973–1979 to 1994–2000. (Source: Ries, et al. SEER Cancer

Statistics Review, 1973–2000. Bethesda, MD: National Cancer Institute. NIH Pub. No. 03-2789, 2003.)

Figure 44–5. United States age-specific leukemia incidence rates according to cell type by race and sex in nine SEER areas, 1973–2000.

(Source: Ries, et al. SEER Cancer Statistics Review, 1973–2000. Bethesda, MD: National Cancer Institute. NIH Pub. No. 03-2789, 2003.)

848

PART IV: CANCER BY TISSUE OF ORIGIN

Table 44–1. Five-Year Relative Survival Rates for Leukemia in the United States Surveillance Epidemiology and End Results Program, by Time Period, Face, Sex, Age, and Cell Type Leukemia Cell Type Period, Age, and Race/Sex Group

Total (%)

AML (%)

CML (%)

ALL (%)

CLL (%)

1974–1976, all ages, all races and sexes 1992–1999, all ages, all races and sexes White males White females A-A males A-A females Ages 0–14 Ages 0–64 Ages 65+

34.4

6.0

22.6

38.3

68.2

46.3*

18.7*

34.9*

63.5*

73.5*

48.3* 46.6* 38.5 40.1 78.1 56.0 35.2

16.8* 19.8* 22.0* 19.0 47.3 30.8 3.9

35.0* 34.8 31.4 41.1 N/C 45.9 21.2

61.4* 66.2* 56.6 63.0 85.1 68.1 8.0

74.1* 75.5 55.3 61.4 N/C 83.5 68.1

Source: Ries LAG, Eisner MP, Kosary CL, et al., eds. SEER Cancer Statistics Review, 1975–2000, National Cancer Institute. Bethesda, MD. Available at: http://seer.cancer.gov/csr/1975-2000, 2003. *p < 0.05 for 1992–1999 vs. 1974–1976. A-A, African American; ALL, acute lymphoblastic leukemia; AML, acute myeloid leukemia; CLL, chronic lymphocytic leukemia; CML, chronic myeloid leukemia; N/C, not calculated.

Atomic Bomb Survivors The most important epidemiological study of radiation-exposed populations for the quantitative risk assessment of the leukemias and other cancers is the follow-up investigation of the Japanese atomic bomb survivors. The dose-response pattern observed for both leukemia mortality (Shimizu et al., 1990; UNSCEAR, 1994; Little and Muirhead, 1998) and AML incidence among the survivors (Preston et al., 1994) was a highly statistically significant upward curvature consistent with a linear-quadratic curve. In incidence analyses of AML, the estimated excess absolute risk was 1.1 per 104 person-years Sv when averaged over the follow-up period of 1950–1987, and the estimated average excess relative risk at 1 Sv exposure was 3.3 (Preston et al., 1994). Relative risks were similar for males and females but males had twofold higher absolute excess risks, while survivors who were under age 10 at exposure had substantially higher average absolute excess risks compared with persons who were older at exposure. Following a detailed histopathological review of myeloid malignancy cases among the survivors in the late 1980s, investigators became aware that a substantial proportion of these patients actually had MDS (Matsuo et al., 1988). Quantitative risk assessment has shown a significant dose-response for MDS mortality, and an excess relative risk several times greater than that seen for all solid cancers combined (Shimizu et al., 1999). In a pooled analysis of Japanese atomic bomb survivors, women treated for cervical cancer, and patients irradiated for ankylosing spondylitis, Little et al. (1999) found similar dose-response and similar patterns of risk as a function of age at exposure and time since exposure, and no difference among the three populations in the relative risk of AML (or of CML or ALL, each considered separately).

Other Environmental Sources of Radiation Although an ecological study suggested correlations between radon or other natural background radiation sources and myeloid leukemia (Henshaw et al., 1990), there was little evidence of a link between measured levels of radon or gamma radiation in subjects’ homes and risk of AML in adults (Forastiere et al., 1998; Law et al., 2000) or in children (Steinbuch et al., 1999; UK Childhood Cancer Study Investigators, 2002).

Radiation Therapy and Radiomimetic Agents for Malignant Conditions Patients treated for non-Hodgkin lymphoma, breast cancer, uterine cervix cancer, uterine corpus cancer, and Ewing sarcoma have gener-

ally shown moderately elevated risks of AML as a second primary cancer (i.e., sAML), likely due to treatment with radiation therapy (Li et al., 1983; Greene, 1984; Boice et al., 1987; Curtis et al., 1992, 1994; Travis et al., 1994, 1996; Kuttesch et al., 1996; Inskip, 1999). In contrast, the elevated risks of sAML associated with Hodgkin lymphoma and ovarian cancer are likely due to alkylating agents or other chemotherapy (Kaldor et al., 1990a,b; Travis et al., 1999). Secondary AML following treatment of testicular cancer has been linked with both radiotherapy and platinum chemotherapy (Travis et al., 2000). Similar to the temporal pattern for occurrence of AML among the Japanese atomic bomb survivors, sAML following radiotherapy often appears within 5 years with most cases occurring within 10–15 years (Inskip, 1999). Excess risks of AML following radiotherapy have been associated with estimated bone marrow doses ranging from 1–15 Gy for adults (and often higher for children), and appear to be greater when large volumes of bone marrow are exposed to lower doses or dose fractions. The low risk of sAML associated with high partial-body radiation exposures may reflect radiation-induced apoptosis in exposed cells (Inskip, 1999). Radioactive phosphorus (32P), used to manage MPD disorders (CML, PV, and ET) in earlier decades was also linked with increased subsequent t-AML (Modan, 1965).

Radiation Therapy for Benign Conditions Increased risk of AML has been linked with radiation treatment for ankylosing spondylitis (associated with a relative risk of 7.0 for leukemia in the period 1–25 years after exposure to a uniform dose of 1 Gy) in earlier decades (Weiss et al., 1995). Somewhat lower AML excesses (relative risks ranging from 1.2–3.0) have been associated with radiation treatments for benign gynecologic disorders (Inskip et al., 1993), menorrhagia not associated with malignancy (Smith and Doll, 1976; Inskip et al., 1990), peptic ulcer (Griem et al., 1994), and tinea capitis (Shore et al., 1976; Ron et al., 1988).

Diagnostic Radiation Exposures There is debate about whether the evidence linking prenatal or postnatal diagnostic X-rays with risk of childhood or adult AML (Gunz and Atkinson, 1964; UNSCEAR, 2000; Naumberg et al., 2001) may be due not to a causal relationship, but to underlying nonmalignant medical conditions or to early clinical manifestations of leukemia (Evans et al., 1986; Boice et al., 1991). Chronic low-dose alpha-particle radiation from injections with the radiographic contrast medium Thorotrast, used in earlier decades for cerebral angiography, has been consistently associated with increased risk of AML/MDS (dos Santos Silva et al., 1999; Martling et al., 1999; van Kaick et al., 1999), with cumulative risk estimated as 173 cases/104 persons per Gy (Andersson et al., 1993).

Occupational Radiation Exposures Historically, medical radiation workers (radiologists and radiologic technologists), nuclear industry workers, radium dial workers, miners (uranium and tin), flight crew, and military servicemen exposed to above-ground nuclear tests are the major categories of workers exposed to ionizing radiation. Radiologists and X-ray technicians employed in the first half of the 20th century experienced notably elevated leukemia mortality, with risks ranging from 6.0 to 8.8-fold excesses among British (Smith and Doll, 1981) and US (Seltser and Sartwell, 1965; Matanoski et al., 1984) radiologists joining specialty societies during 1897–1921 and 1920–1929, respectively. Risk was elevated 3.4-fold among US radiologists joining during 1930–1939. No significant excesses occurred in British radiologists entering after 1921 or in US radiologists entering after 1939. Leukemia mortality was modestly increased among US (Mohan et al., 2003) and Japanese (Yoshinaga et al., 1999) radiologic technologists who first worked before 1950 or 1960, respectively, but not during later periods. No excesses of leukemia were found among US Army radiographic technicians followed up during 1946–1974 (Jablon and Miller, 1978). Non-CLL leukemia incidence was significantly elevated among 27,911 diagnostic X-ray workers followed up during 1950–1980 in

The Leukemias China (Wang et al., 1990). Although many individual studies of nuclear industry workers have shown no significant excesses of leukemia, combined analyses of large cohorts of nuclear workers have revealed small excess relative risks (Cardis et al., 1995; Muirhead et al., 1999). There was no significant excess of leukemia among US or UK radium dial workers (Spiers et al., 1983; Stebbings et al., 1984; Baverstock and Papworth, 1986), nor among uranium miners (National Research Council, 1999), but a borderline significant increase of myeloid leukemia occurred among UK tin miners who had worked 10–20 years underground (Hodgson and Jones, 1990). Excesses of myeloid leukemia or leukemia not otherwise specified were reported among some US (Caldwell et al., 1983), New Zealand (Pearce et al., 1990), and British (Darby et al., 1988) servicemen, but not others (Robinette et al., 1985) exposed to atomic weapons tests. Excess risks of AML have been described in Canadian airline pilots (Band et al., 1996) and Danish cockpit crew flying more than 5000 hours (Gundestrup and Storm, 1999).

Physical Agents—Non-Ionizing Radiation Residential Exposures In population-based studies, adults residing in homes with measured exposures to power-frequency magnetic field levels ≥0.2 microtesla in western Washington state (Severson et al., 1988) or Sweden (Feychting et al., 1994) did not have significantly elevated risks of acute myeloid leukemia.

Occupational Exposures Investigators using job title as a proxy measure of extremely low frequency (ELF) magnetic fields have often reported increased risks of AML among workers considered as having high ELF exposure (e.g., power linemen, utilities workers, and electronics workers) (Milham, 1982; Portier and Wolfe, 1998). A meta-analysis estimated an overall 40% increase in relative risk (Kheifets et al., 1997). However, occupational studies with measurements of ELF magnetic fields have shown inconsistent findings (Floderus et al., 1993; Matanoski et al., 1993; Sahl et al., 1993; London et al., 1994; Theriault et al., 1994; Savitz and Loomis, 1995; Miller et al., 1996; Feychting et al., 1997; Johansen and Olsen, 1998).

Chemical Exposures—Manufacturing, Farming, Medications Benzene Benzene is the oldest (Delore and Borgomano, 1928) and best-known chemical leukemogen. Originally, workplaces were the primary source of exposure, but recognition of benzene leukemogenicity led to a notable reduction in workplace use. Sources of the low levels of benzene consistently identified in blood samples of the general population have not been completely identified, but include cigarette smoke and unleaded gasoline (Brugnone et al., 1994; Kok and Ong, 1994; Melikian et al., 1993; Ong and Lee, 1994). The contribution of internally generated sources from food or other non-occupational sources has not been well delineated. Benzene-exposed painters and printers, and workers employed in chemical, rubber, Pliofilm, and shoe manufacturing and in petroleum refining (Ott et al., 1978; Rinsky et al., 1987; Wong, 1987; Yin et al., 1987; Hayes et al., 1997; Rinsky et al., 2002) experienced leukemia (mostly AML) risks that were 1.9- to 10.0-fold increased. Benzene has been strongly associated with AML and aplastic anemia, but some studies have linked CML, ALL, MDS, and non-Hodgkin lymphoma with exposure to this chemical (Yin et al., 1987; Rinsky et al., 1987; Schnatter et al., 1992, 1996a,b; Hayes et al., 1997; Huebner et al., 1997; Ireland et al., 1997; Rushton and Romaniuk, 1997; Savitz and Andrews, 1997). The risks of AML at low doses, the dose-response pattern, the relevant exposure metrics (e.g., average exposure level, duration of exposure, cumulative exposure, or exposure peaks), latency, and the specific mechanism of leukemogenicity have been much debated (Paustenbach et al., 1992; Utterback and Rinsky, 1995; Crump, 1996; Irons, 2000; Oikawa et al.,

849

2001; Eastmond et al., 2001; Zhang et al., 2002). AML/MDS was associated with recent, but not distant exposure among Chinese benzene-exposed workers (Hayes et al., 1997). Data from the longterm follow-up of a cohort of US pliofilm workers suggests that the excess risk diminishes with time since exposure (Rinsky et al., 2002). Peripheral lymphocyte and neutrophil levels appear to represent sensitive markers of benzene exposure (Ward et al., 1996; Rothman et al., 1996; Qu et al., 2002), even at very low levels of benzene (Qu et al., 2002). Decreased lymphocyte counts may be a useful biomarker for validating estimates of benzene exposure levels (Dosemeci et al., 1996). Benzene exposure in Chinese workers was associated with significant increases in the long arm deletion of chromosomes 5 and 7 (Zhang et al., 1998), with hyperdiploidy of chromosomes 8 and 21, and with translocations between 8 and 21 (Smith et al., 1998) in a dose-dependent fashion; the deletions of the long arm of chromosomes 5 and 7, and the translocation t(8; 21) may be useful markers of early biological effect in a population at increased risk of leukemia. How these findings relate to effects on hemopoietic progenitor cells is uncertain.

Other Chemicals Increased risk of AML and/or MDS have been reported in painters (Matanoski et al., 1986; Chen and Seaton, 1998); plant and machine operators and assemblers (Nisse et al., 1995); coal miners (Nisse et al., 1995); embalmers (Hayes et al., 1990; Linos et al., 1990) and other formaldehyde-exposed workers (Walrath and Fraumeni, 1983, 1984; Blair et al., 1986, 1990, 1992; IARC, 1995; Shaham et al., 1996; Merk et al., 1998); garage and transport workers (Hunting et al., 1995; Hotz and Lauwerys, 1997); shoe workers (Bulbulyan et al., 1998); hairdressers and cosmetologists (Lynge, 1994; Miligi et al., 1999); seaman on tankers (Nilsson et al., 1998); clinical laboratory and science technicians (Burnett et al., 1999); and other occupational and industrial populations (Linet and Cartwright, 1996). Myeloid leukemia was increased among 20,000 persons residing in Seveso within 10 years after an industrial accident caused contamination of the region with 2, 3, 7, 8-tetrachlorobibenzo-p-dioxin and who were less than 19 years old at the time of the accident (Pesatori et al., 1993).

Farming, Agricultural, and Related Exposures Some studies of farmers and farm workers have shown modest excesses of AML (Pearce et al., 1988) as well as virtually all other subtypes of leukemia (risks ranging from 1.1-fold to 1.4-fold elevated), whereas others have shown no increase in risk of AML (Blair and Zahm, 1995; Keller-Byrne et al., 1995; Zahm et al., 1997). International variation in risks may reflect differences in agriculture-related exposures (Wiklund and Holm, 1986; Dean, 1994; Amadori et al., 1995; Kristensen et al., 1996b; Nanni et al., 1996; Avnon et al., 1998), such as pesticides (particularly animal insecticides and herbicides), fertilizers, diesel fuel and exhaust, or infectious agents (Blair et al., 1992). Few earlier studies evaluated specific pesticide exposures in relation to AML (Brown et al., 1990; Sathiakumar and Delzell, 1997), but recent studies are increasingly incorporating newer exposure assessment approaches and biological measurements (Alavanja et al., 1996; Stewart et al., 1996, 1999). Excess AML among New Zealand abattoir workers (Pearce et al., 1988) and US veterinarians (Blair and Hayes, 1982) may be consistent with a viral etiology (Pearce and Reif, 1990).

Cytotoxic Chemotherapy Treatment with alkylating agents has been associated with increased risks of MDS and tAML (Leone et al., 1999); a preceding MDS is observed in over 70% of patients who develop therapy-related AML (Giles and Koeffler, 1994). Typically, t-AML occurs 5–7 years following treatment and risk is related to cumulative alkylating drug dose. t-AML associated with alkylating agents is frequently characterized by a preleukemic phase, tri-lineage dysplasia, and cytogenetic abnormalities involving partial deletions of chromosomes 5 and 7. Therapy-related MDS and t-AML have been reported subsequent to management of Hodgkin lymphoma, non-Hodgkin lymphoma,

850

PART IV: CANCER BY TISSUE OF ORIGIN

multiple myeloma, polycythemia vera, and breast, ovarian, and testicular cancers with alkylating agents, although certain drugs (such as melphalan) pose higher risks than others (such as cyclophosphamide) (Kaldor et al., 1990a, 1990b; Curtis et al., 1992; van Leeuwen et al., 1994; Travis et al., 1994, 1996, 2000; Krishnan et al., 2000). Topoisomerase II inhibitors (specifically epipodophyllotoxins) have also been linked with elevated risk of t-AML, although the resultant t-AML is generally not preceded by a preleukemic phase, develops after a shorter latency period (typically 2 years), is not related to cumulative dose of epipodophyllotoxins (Smith et al., 1999), and has different cytogenetic abnormalities, in particular balanced translocations involving 11q23 (Felix, 1998; Andersen et al., 2000). Increasing doses of platinum-based chemotherapy for ovarian (Travis et al., 1999) and testicular cancers (Travis et al., 2000) have been quantitatively associated with increasing risks for t-AML. A 10-fold higher risk of tMDS/AML has been observed in breast cancer patients treated with mitoxantrone and methotrexate or methotrexate and mitomycin C (Saso et al., 2000). t-MDS/AML has also been associated with the intensity of pretransplantation chemotherapy (e.g., mechlorethamine (Metayer et al., 2003) and/or conditioning treatments (e.g., total body irradiation (Pedersen-Bjergaard et al., 2000), particularly at doses >12 Gy, or VP-16 (Krishnan et al., 2000) in preparation for autologous stem cell transplantation for lymphoma and other malignant diseases (Metayer et al., 2003). Efforts are currently directed at identifying host-related genetic variables that influence risk of developing treatment-related AML (Woo et al., 2000; Perentesis, 2001).

Other Medications A population-based case-control study of childhood leukemia in Shanghai linked use of the antibiotic chloramphenicol and syntomycin, which is pharmacologically related to chloramphenicol, with increased risk of childhood AML and ALL (Shu et al., 1987). Previous clinical surveys have suggested that patients with chloramphenicol-induced aplastic anemia may be at elevated risk of AML (Fraumeni, 1967; Adamson and Sieber, 1981). The evidence has not borne out concerns based on clinical reports that phenylbutazone is linked with adult AML (Friedman and Ury, 1980; Friedman, 1982).

Smoking, Diet, Alcohol, and Hair Dyes Several large studies (Garfinkel and Boffetta, 1990; Sandler et al., 1993; Kane et al., 1999), but not all (Engeland et al., 1996; Adami et al., 1998) have associated cigarette smoking with small excesses of AML. Reviews suggest that risks may be 20–50% increased (Brownson et al., 1993; Doll, 1996). Whether the smoking-related excess of adult AML is restricted to the subgroup with the t(8;21) (q22:22) translocation (Moorman et al., 2002b) will require confirmation in additional studies. Limited data link cigarette smoking with MDS (Bjork et al., 2000). Neither maternal smoking during pregnancy (Magnani et al., 1990; van Duijn et al., 1994; Brondum et al., 1999; Schuz et al., 1999; Sasco and Vaino, 1999) nor paternal smoking during the preconception period (Magnani et al., 1990; Sorahan et al., 1997) appear to be linked to subsequent risk of childhood AML in offspring, although a study in China has implicated paternal preconception smoking with risk of childhood AML (Ji et al., 1997). Overall, there have been relatively few studies of the possible role of diet in AML (or other leukemias) (Hursting et al., 1993; Kwiatkowski, 1993). Infant AML, but not ALL, was linked with mother’s increasing consumption during pregnancy of DNA topoisomerase II inhibitorcontaining foods in a small US study (Ross et al., 1996). In vivo and in vitro experiments suggest that maternal pregnancy-related ingestion of bioflavonoids may induce chromosomal translocations of the MLL gene in utero, potentially leading to infant leukemia (Strick et al., 2000). Because phenol and hydroquinone have been strongly implicated in occurrence of leukemia associated with benzene exposure, it has been hypothesized that perhaps phenol and hydroquinone produced by direct dietary ingestion, catabolism of tyrosine and other substrates by gut bacteria, ingestion of arbutin-containing foods, and some over-the-counter medications may causally be related to de novo AML in the general population (McDonald et al., 2001). High birth weight

has been linked with childhood AML in a record-linkage study in Denmark (Westergaard et al., 1997) and a medical record-based casecontrol study in the United Kingdom (Roman et al., 1997). Experimental data suggest that caloric restriction may mitigate the leukemogenic effects of exposure to single, high-dose total body radiation (Yoshida et al., 1997). Studies to date do not support a role for alcohol consumption in the etiology of adult AML (Williams and Horm, 1977; Jensen, 1979; Blackwelder et al., 1980; Hinds et al., 1980; Schmidt and Popham, 1981; Carstensen et al., 1990; Brown et al., 1992a) or MDS (Ido et al., 1996), but two US investigations have linked maternal alcohol consumption during pregnancy with elevated risks of AML among infants and very young children (Severson et al., 1993; Shu et al., 1996). Hair dye use, particularly dark hair dyes, has been weakly linked with adult AML and MDS (Mele et al., 1994; Nagata et al., 1999; Correa et al., 2000b).

Infectious Agents Few studies have assessed the potential role of infectious agents in the etiology of AML or MDS, and the evidence is not particularly compelling.

Molecular Epidemiology of AML and Related Disorders Because most hematological malignancies, and other malignant and non-malignant chronic disorders are thought to result from multiple and interacting environmental and genetic components, common genetic variants, including single nucleotide polymorphisms, may influence susceptibility. Increasingly, the role of DNA polymorphisms has been evaluated in relation to myeloid disorders, but the literature is still modest. Unfortunately, fewer than 200 gene variant-disease outcomes have been studied more than three times among the thousands of studies that have examined the relationship between common gene variants and numerous serious diseases, and only a small fraction of the reported associations have been consistently replicated (Hirschhorn et al., 2002). Lack of replication may reflect problems with study design, the approach used for genetic and statistical analysis, inadequate sample size, and other potential sources of bias (Rothman et al., 2001). Large, well-designed investigations are needed that consider genetic pathways, rather than single or unrelated genes, and that also consider relevant exposures (Goode et al., 2002). As sound methodological approaches are increasingly incorporated, inconsistencies due to false positives and false negatives should diminish and consistent results should, hopefully, become apparent in meta-analyses or pooled analyses (Lohmueller et al., 2003).

Glutathione S Transferases (GST) The metabolizer enzymes GST T1, GST M1, and GST P1 are related to the individual’s ability to detoxify a range of environmental carcinogens, including potential carcinogens associated with AML/MDS, such as chemicals in cigarette smoke, ethylene oxide, and certain cytotoxic drugs (Chen et al., 1996; Basu et al., 1997; Atoyebi et al., 1997; Strange and Fryer, 1999). The GSTs also protect DNA from damage as a consequence of oxidation stress. It was postulated that persons with homozygosity for null alleles for one or more forms of the GSTs might be at increased risk for developing AML/MDS, due to inability to detoxify specific leukemogens (Sasai et al., 1999; Davies et al., 2000). Persons with null alleles for GST T1 were at modestly increased risk of developing de novo MDS or AML in some (Chen et al., 1996; Sasai et al., 1999; Arruda et al., 2001; Rollinson et al., 2000), but not all (Crump et al., 2000; Naoe et al., 2000) case-control studies. In contrast to the inconsistencies in results for de novo AML/MDS, studies have shown that null alleles for GST T1 are not related to risk of tAML (Crump et al., 2000; Naoe et al., 2000; Allan et al., 2001). Individuals with null alleles for GST M1 were at elevated risk of developing de novo AML (Rollinson et al., 2000; Arruda et al., 2001) or MDS (Tsabouri et al., 2000), but not t-AML (Naoe et al., 2000; Allan et al., 2001), whereas persons with null alleles for both GST T1 and GST M1 experienced an excess risk of developing t-AML (Haase et al., 2002). Risks of t-AML were increased among patients previ-

The Leukemias ously treated with chemotherapy agents that are known substrates of GST P1, and the GST P1 codon 105 Val allele occurred more often in the t-AML cases than in those with de novo AML in that population (Allan et al., 2001). The null genotype of GST T1 was not associated with increased risk of developing t-AML among children treated with epipodophyllotoxins for ALL (Woo et al., 2000).

N-Acetyl Transferases (NAT) A potential role of slow acetylator status in leukemogenesis was suggested by a report of increased DNA adduct levels in peripheral blood lymphocytes (Millikan et al., 1998), but adult AML was not linked with NAT2 metabolizer status in a large population-based case-control study in the United Kingdom (Rollinson et al., 2001).

Cytochrome P450 (CYP) Altered expression from genetic polymorphism at the CYP2D6 gene locus is responsible for pronounced inter-individual variation in the metabolism of many clinically important drugs and small molecules. CYP2D6 poor metabolizers (PM) had increased risks of heterogeneous groups of leukemias in adults in two studies (Wolf et al., 1992; Lemos et al., 1999). Significantly increased risks were seen for t-AML and de novo AML in adults aged 40 and older among persons with the CYP2D6 PM phenotype in a large population-based study in the United Kingdom (Roddam et al., 2000), but there was no interaction between the CYP2D6 PM phenotype and smoking status, despite CYP2D6’s recognized involvement in the metabolism of tobacco components, such as nicotine and 4-(methylnitrosmaine)-1-(3-pyridyl)-1butanone) (Roddam et al., 2000). Roddam et al. (2000, 2003) also found that the CYP2C19 PM phenotype, but not the CYP1A1*3, was associated with an increased risk of developing both AML and sAML; there were no interactions with age, gender, or smoking status for either of these alleles. Among 447 patients with abnormal karyotype treated in clinical trials of AML in the United Kingdom, the CYP1A1*2B (Val) variant allele was over-represented in patients with NRAS mutation compared with no mutation in both the entire population of AML cases and in the poor-risk karyotype group of AML patients with partial or complete deletion of chromosomes 5 or 7 or abnormalities of chromosome 3 (Bowen et al., 2003). There were no differences in the frequencies of the CYP3A5*3 and the CYP3A4*1B alleles in children with ALL who developed t-AML, and in children with ALL who did not develop t-AML (Blanco et al., 2002). NAD(P)H:quinone oxidoreductase (NQO1) is an enzyme that detoxifies quinones, derivatives, and other natural and synthetic compounds (Joseph et al., 1994; Traver et al., 1992). Many topoisomerase II-inhibiting drugs as well as benzene, dietary flavonoids, and other chemicals share the same structural feature (Ross et al., 1994; Smith, 1999; Strick et al., 2000; McDonald et al., 2001). NQO1 is induced by synthetic antioxidants and cruciferous vegetables, and protects cells against oxidative stress. Individuals who are homozygous for the variant allele completely lack NQO1 activity, whereas heterozygotes have low-to-intermediate activity compared with individuals with the wild-type alleles. Disruption of the NQO1 gene in mice has been shown to cause myeloid hyperplasia of bone marrow (Long et al., 2002). NQO1-deficient mice were much more sensitive to benzene than wildtype NQO1 mice, but the specific patterns of toxicity differed between male and female mice (Bauer et al., 2003). Benzene hematotoxicity, which was strongly associated with development of hematopoietic neoplasms in Chinese workers, was linked with polymorphisms in genotypes of enzymes that activate (i.e., CYP2E1) and detoxify (i.e., NQO1) benzene and its metabolites (Rothman et al., 1997). The NQO1 variant allele appears to be significantly over-represented in adults with therapy-related myeloid leukemias (Larson et al., 1999; Naoe et al., 2000), and in those whose myeloid leukemia was characterized by deletions of chromosomes 5 or 7 (Larson et al., 1999). Null alleles for NQO1 predisposed to increased risk of de novo AML in adults (Smith et al., 2001). Infants with leukemia characterized by MLL gene rearrangements were eightfold more likely to have low NQO1 function than healthy children or childhood leukemia patients with TELAML1 gene fusions or hyperdiploidy (Wiemels et al., 1999b). Patients

851

with childhood ALL who developed t-AML following chemotherapy were not more likely to have low NQO1 function than those who did not develop t-AML (Blanco et al., 2002).

DNA Repair Pathways DNA damage is the final common pathway that mediates occurrence of malignancy (Hansen and Kelley, 2000). There is limited information on the possible relationship of single nucleotide polymorphisms in genes within DNA repair pathways (which include direct repair of damaged bases, base excision repair, nucleotide excision repair, mismatch repair, double-strand break repair, and single-strand break repair) with pathogenesis of myeloid disorders. Primary myeloid leukemia cells have shown an increased frequency of misrepair compared with normal control cells (Gaymes et al., 2002), and cell lines with microsatellite instability were defective in strand-specific mismatch repair (Gu et al., 2000). Inherited variation in genes involved in mismatch repair, including MSH2 and hMCH1, have been suggested to predispose to chemotherapy-induced AML (Worrillow et al., 2003; Allan et al., 2003). Detailed study of four adult AML cases with the t(8;16(p11;13) translocation has shown deletions, duplications, and insertions in the breakpoint regions of the MOX and CBP genes, suggesting a damage-repair mechanism in the origin of this translocation (Panagopoulos et al., 2003). At least one copy of the variant allele XRCC1 399Glu conferred a protective effect against t-AML in a small case-control study (Seedhouse et al., 2002). Families with individuals homozygous for mutations in mismatch repair genes are at increased risk of developing hematological malignancies and/or neurofibromatous 1 at an early age (Ricciardone et al., 1999; Whiteside et al., 2002). Patients with Fanconi anemia, a condition characterized by cells that are sensitive to DNA cross-linking, are at increased risk of developing AML (Folias et al., 2002; Tamary et al., 2002). The potential role of variant alleles in DNA repair genes in the etiology of sporadic de novo myeloid disorders has not been evaluated.

Familial and Genetic Factors in AML and Related Disorders Familial Aggregation Clinical reports of familial occurrence of AML, MDS, or both are rare, but data support the contribution of highly penetrant mutations in leukemia susceptibility genes. Familial AML appears to be a heterogeneous disease, however, in that there does not appear to be a single karyotypic abnormality, molecular defect or typical age at onset common to affected persons in all such families (Horwitz et al., 1997). Some familial AML is characterized by monosomy 7 (Gilchrist et al., 1990; Kwong et al., 2000). Other families demonstrate loss of the long arm of chromosome 5 (Grimwade et al., 1993; Olopade et al., 1996), while still others have other or no karyotypic abnormalities (Mandla et al., 1998). The molecular mechanism(s) underlying occurrence of familial AML (even those with apparent autosomal dominant transmission) have not been elucidated, although anticipation has been described in familial AML (Horwitz et al., 1996; DeLord et al., 1998). Familial platelet disorder (associated with mutations in AML1, the gene deregulated by the t(8;21) translocation) (Song et al., 1999), Kostmann syndrome (severe chronic neutropenia, linked with mutations in the G-CSF receptor (Freedman and Alter, 2002), sideroblastic anemia (Kardos et al., 1996), polyposis coli (Greenberg et al., 1981), and neurofibromatosis, type I (Shannon et al., 1992) appear to increase risk of familial MDS/AML. Although childhood onset of MDS was believed to occur in the setting of familial MDS, a Danish population-based linked registry study did not support this conclusion (Hasle and Olsen, 1997).

Genetic Syndromes Predisposing to Myeloid Disorders It has been estimated that approximately 5% of AML/MDS may be associated with inherited genetic syndromes (Taylor and Birch, 1996). Children with Down syndrome are at increased risk of developing acute leukemia, particularly AML M7, perhaps due to a functional role of mutant P53 in the evolution from a transient form of leukemia to

852

PART IV: CANCER BY TISSUE OF ORIGIN

acute megakaryoblastic leukemia (Malkin et al., 2000; Hasle et al., 2000). More recently, the role of mutations in GATA1 has been highlighted (Wechsler et al., 2002; Mundschau et al., 2003; Groet et al., 2003). Bone marrow failure syndromes linked with AML/MDS include Fanconi anemia (Alter, 2003); Bloom syndrome (Bloom et al., 1966; German, 1997), Schwachman-Diamond syndrome (Dror and Freedman, 2001), amegakaryocytic thrombocytopenia, dyskeratosis congenita, and Kostmann syndrome (Alter, 1996).

CHRONIC MYELOID LEUKEMIA AND OTHER MYELOPROLIFERATIVE DISORDERS Physical Agents—Ionizing Radiation Environmental Sources

Chemical Exposures—Manufacturing, Farming, and Medications Chemical Exposures Chronic myeloid leukemia has been reported among benzene-exposed workers in China (Yin et al., 1989; Yin et al., 1994) and the United States (Rinsky et al., 1987), but the small numbers of cases preclude precise quantification of risk and the association requires confirmation in other benzene-exposed populations. Excesses of myelomonocytic leukemia (Paganini-Hill et al., 1980) and myelofibrosis (Zoloth et al., 1986) have been identified among pressmen and printers. Farming has been associated with elevated risk of CML in a few studies (Blair and White, 1981). Farming and agricultural exposures that may be linked with CML are the same as those described for AML.

Excesses of CML have been documented in the Japanese atomic bomb survivors. Occurrence of CML peaked within 5 years of exposure based on mortality data, regardless of age at the time of the bombing, but excess CML declined rapidly with time since exposure (Shimuzu et al., 1989; Pierce et al., 1996). The relative risk of developing CML was greater in persons who were younger compared with those who were older at exposure (Finch and Linet, 1992). CML was much less common among unexposed survivors in Nagasaki than Hiroshima, but there was no significant difference in relative risk of radiation-induced CML between the two cities (Preston and Pierce, 1988; Preston et al., 1994). In contrast to the nonlinear dose-response function for AML, there was no evidence against linearity for CML (Preston et al., 1994). When averaged over the follow-up period (1950–1987), the excess absolute risk for CML was estimated as 0.9 per 104 person-years Sv, while the estimated average excess relative risk at 1 Sv was 6.2 (Preston et al., 1994). Male survivors had greater excess absolute risks of CML and a high initial peak that subsequently declined rapidly with time, whereas female survivors had lower excess absolute risks that remained constant over the years. CML has not been linked with residential radon, background gamma radiation, or other environmental radiation exposures.

Medications

Radiotherapy for Malignant and Benign Conditions

There is little literature on the molecular epidemiology of CML or related disorders.

Although much less common than AML, case reports (Iurlo et al., 1989; Beaty et al., 1995) and some epidemiological studies (Inskip et al., 1990; Little et al., 1999) provide evidence that CML can occur after radiation therapy for benign or malignant conditions. AML may occur substantially more often than CML following radiotherapy, however, because ionizing radiation generates the AML1-ETO fusion gene, associated with the t(8:21) translocation of AML, significantly more often than the BCR-ABL fusion gene, associated with the t(9;22) translocation of CML (Deininger et al., 1998). Treatment of certain disorders with radiotherapy may, nevertheless, result in radiationrelated CML more often than AML. Radiation-treated disorders associated with secondary CML include histiocytosis X (Chap and Nimer, 1994), uterine bleeding with intra-uterine radiation (Inskip et al., 1990), and metastatic papillary and follicular thyroid cancer with lowdose 131I (Shimon et al., 1995). It is possible that underlying genetic factors or chance events may play a role, since no clinically or molecularly appreciable differences were seen between treatment-related and de novo CML (Aguiar, 1998). In a pooled analysis of Japanese atomic bomb survivors, women treated for cervical cancer, and patients irradiated for ankylosing spondylitis, Little et al. (1999) found that relative risk of CML rose within a few years after exposure, then decreased with increasing time after exposure, similar to the pattern for AML.

Diagnostic Radiation Exposures Excesses of CML have been linked with diagnostic X-rays (PrestonMartin et al., 1989), particularly large numbers of X-rays (Gibson et al., 1972), although additional studies are needed to confirm the relationship. Increased CML has been linked with Thorotrast, used for cerebral angiography in earlier decades (Visfeldt and Andersson, 1995).

A growing literature has described CML as a late effect following cytotoxic therapy (Aguiar, 1998; Waller et al., 1999), but no appreciable differences or molecular abnormalities distinguished treatmentrelated from de novo CML.

Smoking, Diet, Alcohol, and Hair Dyes Cigarette smoking is known to induce leukocytosis and increased genetic instability but only limited data are available on associations of CML with the frequency or duration of smoking (Williams and Horm, 1977; Brown et al., 1992a). Neither cigarette smoking nor diet, alcohol, or hair dyes have been adequately studied in relation to CML.

Infectious Agents Few studies have evaluated the relationship of specific infectious agents with CML.

Molecular Epidemiology of CML and Related Disorders

Familial and Genetic Factors in CML and Related Disorders There are a few clinical reports of familial CML, but systematic studies of familial aggregation provide little evidence that this occurs much more commonly than by chance (Gunz et al., 1978; Harnden, 1985).

LYMPHOID LEUKEMIAS Origin, Progeny, and Subtypes Lymphoid malignancies originate in cells committed to differentiation along lymphoid lineage. The WHO classification (Jaffe et al., 2001), which is based on current understanding of the development and functions of the immune system, recognizes two broad categories of lymphoid disorders: those composed of hemopoietic stem cells or immature precursor cells that are in the process of rearranging their immunoglobulin or T-cell receptor genes (hereafter abbreviated as precursor diseases, which include pediatric and adult forms of acute lymphocytic leukemia or ALL) and those composed of functional peripheral B-cells or T-cells in which the rearrangement of immunoglobulins or T-cell receptor genes is complete (hereafter abbreviated as peripheral diseases, which include chronic lymphocytic leukemia (CLL) as well as non-Hodgkin lymphoma, Hodgkin lymphoma, and multiple myeloma (Harris, 1999; Fritz et al., 2000; Jaffe et al., 2001). Peripheral diseases are classified according to B-cell or T-cell lineage, with those of B-cell lineage further categorized by stage of differentiation of the cells compared with the germinal center. The heterogeneity of subtypes comprising the major categories of the lym-

The Leukemias phoid malignancies is increasingly recognized (Greaves, 1993; Child et al., 1998; Harris, 1999; Hamblin et al., 2000).

Characteristics and Pathogenesis of Lymphoid Leukemias Characteristics To understand the etiology of lymphoid disorders, it is important to recognize the characteristic genetic instability and highly variable history within the normal life cycle of lymphocytes, which undergo genetic recombination and mutation to generate high affinity antibodies (Kelsoe et al., 1998; Vanasse et al., 1999). The immunoglobulin heavy chain and the T-cell receptor genes provide the diversity necessary for the recognition of many types of infective organisms by antibody-producing B-cells, with T-cell help. Recombination and insertion or deletion of multiple germline V, D, and J regions generate the primary sequence of the gene coding for the antigen receptor in each lymphocyte. Chromosomal translocations involving the immunoglobulin heavy chain and T-cell receptor genes are common and can deregulate several oncogenes controlled by these genes (Willis and Dyer, 2000). Exposure to specific antigens leads to clonal expansion and effector cell differentiation.

Precursor Diseases—Characteristics of Initiation of ALL Acute lymphocytic leukemia (ALL) is characterized by a malignant proliferation of lymphoid blast cells in the bone marrow. The target cell in childhood ALL is a committed lymphoid precursor cell that carries the machinery to undergo apoptosis, but instead is transformed (Greaves, 1993), whereas the target cell in adult ALL is thought to be a stem cell that differentiates along the lymphoid lineage and is inherently resistant to apoptosis (Greaves, 1999). In the WHO Classification, three main subtypes of ALL are recognized: precursor B-cell ALL; precursor T-cell ALL; and Burkitt-cell leukemia (Jaffe et al., 2001). Molecular subgroups differ between children and adults, with the translocation t(12;21) predominating in childhood ALL and rare in adult ALL, whereas cases with t(9;22) translocation increase in frequency with increasing age (Secker-Walker et al., 1991). The translocation t(11q23), seen in up to 85% of infant ALL patients, deregulates the mixed lineage leukemia (or MLL) gene, a transcription factor with trithorax homology that may play an important role early in the development of the hematopoietic system (Ernst et al., 2002). There appear to be differences in the distribution of chromosomal breakpoints within the MLL breakpoint cluster region for infant ALL with the MLL (t[4;11]) translocation compared with children older than 1 year of age and adults with ALL characterized by the t(4;11) translocation (Reichel et al., 2001). Similar translocations occur in leukemias resulting from chemotherapies that target DNA topoisomerase II (Felix, 1998). Following translocation, the MLL gene combines with one of numerous different genes to produce an array of fusion proteins whose role in leukemogenesis is not well understood. The translocation t(1;19) is the commonest abnormality in pre-B ALL, but is rarely seen in other types of childhood ALL.

ALL: Timing of Initiation. Accumulating evidence suggests that the key translocation event occurs prenatally for several specific molecularly defined ALL subtypes. For infant leukemia characterized by the MLL-AF4 (t[4;11]) translocation (Ford et al., 1993; Gale et al., 1997) and childhood ALL characterized by the TEL-AML1 (t[12;21]) translocation (Wiemels et al., 1999a), as well as childhood AML characterized by the AML1-ETO (t[8;21])) translocation (Wiemels et al., 2002a), the initiating chromosome translocation event likely occurs in utero, whereas for childhood ALL characterized by the E2A-PBX1 (t[1;19]) translocation, evidence supports a postnatal pre-B–cell origin (Wiemels et al., 2002b). Evidence suggesting that postnatal exposures and/or additional genetic events are required subsequent to the initiating translocation event in utero includes: the 100-fold higher occurrence of TEL-AML1 or AML1-ETO fusion genes in cord blood than the occurrence of the corresponding childhood leukemias (Mori et al., 2002), the modest concordance rate for childhood leukemia in identi-

853

cal twins (Zuelzer and Cox, 1969; Buckley et al., 1996) and a study of triplets who all developed childhood ALL (Maia et al., 2001), and the long and variable latency periods before diagnosis of ALL in children (Wiemels et al., 1999a, 1999b). The most plausible explanation for the same unique, clonotypic, but nonconstitutive fusion sequences found in identical twins with infant leukemia characterized by the MLL-AF4 (Ford et al., 1993) or the TEL-AML1 (Ford et al., 1998) fusion genes is a single cell origin of the fusion in one fetus in utero, followed by intraplacental ‘metastasis’ of clonal progeny to the other twin (Clarkson and Boyse, 1971). Among the triplets, Maia et al. (2001) confirmed that different postnatal genetic events occurred in the two monozygotic members than in the third dizygotic triplet, despite all three triplets demonstrating clonotypic TEL-AML1 fusion sequences in their blood spots.

Peripheral Diseases—Characteristics of CLL Chronic lymphocytic leukemia (CLL) is classified with the lymphomas as one of the peripheral lymphoproliferative diseases in the WHO Classification. We include CLL in this chapter on the leukemias because until recently CLL was frequently designated as one of the four major categories of the leukemias. Morphologically CLL appears as a homogeneous disease of small lymphocytes, but the increasingly wide application of flow cytometry and molecular analysis reveals that CLL is not homogeneous, but can be classified into two major subtypes based on the pattern of immunoglobulin gene mutations of preand post-germinal center CLL (Hamblin et al., 2000). No disease gene has been identified for CLL, and the molecular pathogenesis remains largely unknown. The commonest cytogenetic abnormalities include: interstitial deletions of 13q and trisomy 12 (which are most likely involved in tumor progression), deletions of 11q at the ATM gene locus, 6p and 6q rearrangements, and p53 mutations, which may play a role in transformation of CLL (Brown et al., 1993; Liu et al., 1993; Gahn et al., 1997; Corcoran et al., 1998; Stankovic et al., 1999; Sgambati et al., 2001).

CLL: Timing of Initiation. The likely timing of initiation of CLL is unknown, although studies of possible precursors may be useful. Using flow cytometry, monoclonal B lymphocytes with the CLL phenotype were identified in 2.1% of outpatients ages 40–59 and in 5.0% of outpatients ages 60 and older with normal complete blood cell counts (Rawstron et al., 2002a), and in 13.5% of healthy firstdegree relatives of patients with CLL (Rawstron et al., 2002b). The Descriptive Epidemiology of Lymphoid Leukemias International Patterns Age-adjusted incidence rates for ALL among persons of both sexes are highest in Hispanics in Los Angeles, and in Spain (Taragona), northern Italy, and in whites in New Zealand. Lowest rates are observed among African Americans and Asians (Fig. 44–3). Childhood ALL patterns are similar, with highest incidence in Costa Rica and Hispanics in Los Angeles, and lowest rates in African Americans, the Middle East, and in India (data not shown) (Parkin et al., 1998). CLL shows by far greater international variation in age-adjusted incidence rates than other leukemias. A 40-fold difference is seen between the highest (Italy, Romagna) and lowest (Japan, Osaka) rates for men, and a 38-fold difference between the highest (New Zealand) and lowest (Japan, Osaka) rates for women (Fig. 44–3B). Overall, rates are consistently higher in males than in females, although the male : female ratio varies from 1.3 in France (Isere) to 3.1 in China (Shanghai). Highest age-adjusted incidence rates for CLL are seen also in New Zealand, North America, and Denmark. Mid-level rates for CLL occur in African Americans and in southern Europe. Rates are somewhat lower in US Hispanics, and lowest in all Asian populations (Fig. 44–3B). Adult T-cell leukemia/lymphoma (ATLL) has been classified as a form of non-Hodgkin lymphoma in ICD-O-3, and therefore will not be considered in detail in this chapter. Recent reviews can be found

854

PART IV: CANCER BY TISSUE OF ORIGIN

elsewhere (Blattner, 1999; Edlich et al., 2000; Gotuzzo et al., 2000; Van Kricken and Hoeve, 2000; Siegel et al., 2001). Briefly, ATLL is rare in most populations, except in regions endemic for the human Tlymphotropic virus type 1 (a retrovirus), which include southwestern parts of Japan, Jamaica and Trinidad, parts of West and Central Africa, some regions in South America, and other restricted geographic areas (Ferreira et al., 1997; Manns et al., 1999). ATLL occurs after a long latent period, but most individuals infected with the human Tlymphotropic virus type 1 do not develop ATLL. It is estimated that ATLL occurs in only 1 per 1,000 carriers per year, for a total of an estimated 2500–3000 cases worldwide (Blattner, 1999). The agestandardized incidence rate of ATLL is estimated to range from 1.9 to 2.9 per 100,000 person-years, based on data from Jamaica and Trinidad (Cleghorn et al., 1995), and 2.6 to 4.0 based on data from Kyushu, Japan (Tajima, 1990), with highest risk among individuals who acquire the retrovirus during childhood. Those infected before the age of 20 years are estimated to have a cumulative lifetime risk of 4.0% for males and 4.2% for females (Murphy et al., 1989).

Trends in Age-Adjusted Patterns by Gender and Race For ALL, age-standardized incidence was higher for whites of both sexes than for African Americans, but within each racial group there was a consistently higher rate for males than females among whites, but only a slightly higher incidence for males than females among African Americans until 1994–2000, when a greater gender difference became apparent (Fig. 44–4). Incidence of ALL rose over time in all four race-sex groups, but the rate of the initial increase was greater in African Americans than in whites. During 1973–1979 to 1987–1993, the rate of the increase was similar in white males and females; between 1987–1993 and 1994–2000 the rise in incidence leveled off in white males, but continued to increase in white females. Among African Americans, a dramatic increase was apparent for both male and females between 1973–1979 and 1980–1986, while increases occurred at a slower rate between 1980–1986 and 1987–1993. Between 1987–1993 and 1994–2000, a slower rate of increase was also apparent in African American males, but the incidence rate declined rapidly in African American females. For CLL, rates for males of both races were substantially higher than rates for females of each race, but within each gender, rates were higher for whites than for African Americans (Fig. 44–4). The pattern was generally similar for all four race-sex groups, with a modest increase between 1973–1979 and 1987–1993, followed by a notable decline between 1987–1993 and 1994–2000 that was more marked for African Americans than for whites.

Age-Specific Patterns by Gender and Race Lymphoid leukemias comprise 47.5% of total leukemias in the United States, including 11.3% ALL, 32.2% CLL, and 4.0% other lymphoid leukemia. Among children under age 14, 77.8% of leukemias are lymphocytic, virtually all of which is ALL. Among persons aged 65 or older, 46.1% of leukemias are lymphocytic, most of which (40.1%) are CLL and small fractions are ALL (2.4%) or other lymphocytic leukemia (3.6%). ALL is the most common cancer occurring in childhood, comprising about 30% of all childhood cancers in most populations internationally except in Africa and the Middle East (Parkin et al., 2002). The age-specific incidence pattern for ALL is quite distinctive, with a peak at ages 2–4, followed by a declining incidence rate throughout the remainder of childhood, adolescence, and early adulthood to a nadir at age 40; subsequently, incidence of ALL rises with increasing age to a second, albeit lower peak among the elderly (Fig. 44–5). Pediatric ALL is notably higher in US whites than in African Americans of both sexes, and consistently higher in males than in females. CLL is uncommon in persons younger than age 30, but increases exponentially from age 30 until age 60, when the rate of the increase becomes slower. At all ages, rates are higher in males than females. CLL incidence rates are similar in US whites and African Americans until age 50, when the white : African American ratio increases (Fig. 44–4).

Survival: US Population-Based Data ALL patients have shown a substantial and significant improvement in prognosis, with 5-year relative survival increasing from 38.3% for all age, gender, and racial groups in 1974–1976 to 63.5% in 1992–1999 (Ries et al., 2003), a continuation of the dramatic improvement in survival for pediatric ALL that began in the 1960s. Between 1974–1976 and 1992–1999, 5-year relative survival among children diagnosed with ALL under age 15 increased from 53.2% to 85.1%, respectively. Survival for children with ALL is strongly associated with age at diagnosis, with highest 5-year relative survival among those ages 1–4 (87.7%) and 5–9 (82.3%) compared with lower survival among those ages 10–14 (69.9%) or 15–19 (54.9%). For children with ALL, 5-year survival rates were slightly better for females (45.0%) than males (36.8%) (Ries et al., 2003). The dramatic improvements in survival were limited to children and young adults. Five-year relative survival in 1992–1999 ranged from 22.3% for those aged 45–54 at diagnosis to 18.7% for those aged 55–64, and the elderly had a poor prognosis (8.5% at ages 65 and older, and 2.5% at ages 75 and older). Only modest improvements in 5-year relative survival were apparent for those with CLL, increasing from 68.2% in 1974–1976 to 73.5% in 1992–1999 (Ries et al., 2003). The increases were limited to white males and females, because overall there has been little change during this same time interval in African Americans. Younger patients demonstrate better 5-year relative survival (e.g., 83.5% for those under age 65) than older patients (e.g., 58.1% for those aged 75 and older). The results for African Americans according to age at diagnosis are more difficult to interpret due to small numbers, but there is a similar large difference in survival for those under age 65 compared with those aged 75 or older at diagnosis, although the relative survival rates are substantially lower than those for whites for each corresponding age group.

Analytical Epidemiological Studies Despite growing evidence of important biologic and/or prognostic heterogeneity among different subtypes of ALL, few descriptive (Groves et al., 2000) or analytical (Shu et al., 2002b) epidemiological studies have considered patterns of occurrence or postulated risk factors for ALL by histologic, immunophenotypic, cytogenetic, molecular, or other form of biologically defined subtypes, thus precluding detailed evaluation using these categories. The growing recognition that CLL is heterogeneous has not been incorporated to date in epidemiological studies.

ACUTE LYMPHOCYTIC LEUKEMIA Physical Agents—Ionizing Radiation Japanese Atomic Bomb Survivors Risks of childhood leukemia (or other childhood cancers) were not increased among the Japanese atomic bomb survivors exposed in utero (Delongchamp et al., 1997), although the number of leukemias occurring in this population was very small. Among the Japanese atomic bomb survivors, incidence follow-up data for 1950–1987 revealed that those exposed under age 10 had the highest excess absolute risks for ALL, which peaked at less than 10 years since exposure and decreased rapidly at about 14% per year (although the estimates were based on a small number of cases) (Preston et al., 1994). The absolute risk for incidence was estimated to decrease by about 5% for each year’s increase in age, and risk for females was only about 40% of the risk of ALL in males. Compared with Japanese atomic bomb survivors exposed during childhood and adolescence, those exposed in adulthood had lower estimated absolute risks of ALL, but the pattern of ALL following adult exposure was similar to that following childhood exposure, with risk peaking at less than 10 years since exposure, a declining risk for each year’s increase in age at exposure, and risks for women less than half those for men (Preston et al., 1994). For all ages at exposure, the estimated relative risk for ALL at 1 Sv exposure based on incidence data for the period 1950–1987 was 10.3, and the corre-

The Leukemias 4

sponding estimated absolute risk was 0.57 cases per 10 person-year Sv.

Proximity to Nuclear Plants Small clusters of childhood leukemia observed in geographic proximity to nuclear plants in the United Kingdom in the mid-1980s prompted several large surveys. Mortality from leukemia and lymphoma in persons under age 25 was increased among populations living near nuclear fuel reprocessing or weapons production plants during 1962–1983 (particularly Sellafield and Dounreay) and persisted in proximity to Sellafield through 1990 (Draper et al., 1993), but there was no excess among populations residing close to nuclear plants generating electricity (Cook-Mozaffari et al., 1989; Forman et al., 1987; Roman et al., 1987; Darby and Doll, 1987). Excesses of leukemia and lymphoma occurred in young people residing in geographic regions where construction of nuclear installations had been considered or had occurred at a later date (potential sites) in addition to the increases in areas of existing sites of nuclear facilities (Cook-Mozaffari et al., 1989). Environmental radiation levels measured in proximity to Sellafield and other nuclear facilities were considered too low to ascribe the childhood leukemia excesses to radiation exposures from these plants (Black, 1984; Darby and Doll, 1987).

Residential Radon Although residential exposure to radon was linked with elevated risk of ALL in an ecological study (Henshaw et al., 1990), large casecontrol studies with residential radon measurements have shown no excess risk of childhood ALL in the United States (Lubin et al., 1998), Germany (Kaletsch et al., 1999), or the United Kingdom (UK Childhood Cancer Study Investigators, 2002), but a modest excess was observed in Sweden among children living in homes built from uranium-containing alum shale concrete (Axelsson et al., 2002).

Parental Occupational Radiation Exposures Excesses of childhood ALL and non-Hodgkin lymphoma were reported among offspring of male nuclear workers at Sellafield with an estimated cumulative radiation dose of >100 mSv (Gardner et al., 1990). In a large national study of UK nuclear workers monitored for occupationally related exposures to ionizing radiation, there was a modest increase of childhood leukemia among offspring, but this was based on only three cases (Roman et al., 1993). Risks of childhood ALL were not increased among offspring of nuclear workers in Scotland (Kinlen et al., 1993a) or Canada (McLaughlin et al., 1993b).

Medically Related Radiation Exposures Numerous case-control studies (including some in which exposures were validated) have linked prenatal diagnostic X-ray exposures of mothers, particularly during the last trimester of pregnancy, to 40%–50% elevated risks of childhood ALL and other types of childhood cancers (Doll and Wakeford, 1997; UNSCEAR 1988, 1994, 2000). There was no association of prenatal diagnostic X-rays with risk of childhood ALL in the few cohort studies (Court Brown et al., 1960; Diamond et al., 1973), but the statistical power of these cohort studies was limited, given the rarity of childhood leukemia. In general, offspring of long-term survivors of childhood or adolescent cancer have not experienced elevated risks or only very modest increases of childhood ALL (Li et al., 1979; Mulvihill et al., 1987; Green et al., 1997), with the latter most often representing recognized patterns of familial cancer or perhaps new cancer family syndromes. Children treated with radiation therapy for certain malignant or benign conditions have shown elevated risks of AML, and occasionally ALL, although the leukemia subtypes were not characterized for many of the cases (Hempelmann et al., 1975; Shore et al., 1976; Hildreth et al., 1985; Ron et al., 1988). Among the studies evaluating the relation of postnatal diagnostic X-ray exposures with risk of childhood leukemia, only one (Graham et al., 1966) of several investigations validating interview-derived reports of diagnostic X-ray exposures with medical records (Stewart et al., 1958; Polhemus and Koch, 1959; Ager et al., 1965; Graham et al., 1966) demonstrated statistically significant positive associations.

855

Physical Agents—Non-Ionizing Radiation Early reports of twofold to threefold excesses of childhood leukemia (and other forms of cancer) in children residing in homes with high levels of 60-Hertz magnetic field exposures induced by nearby power lines (Wertheimer and Leeper, 1979; Savitz et al., 1988; Feychting and Ahlbom, 1993) were assessed further in large studies with more extensive and direct measures of children’s exposures (Michaelis et al., 1997; Linet et al., 1997; McBride et al., 1999; UK Childhood Cancer Study Investigators, 1999) and in pooled analyses (Ahlbom et al., 2000; Greenland et al., 2000). Risks of leukemia were not increased among children residing in homes with power frequency magnetic field exposures under 0.3 microtesla (which included more than 99% of residences internationally), but were estimated to be approximately twofold elevated among the less than 1% of children living in homes with magnetic field exposure levels of 0.3 microtesla or higher. Reasons for the increase remain unexplained, and experimental studies have not linked power frequency magnetic field exposures with carcinogenesis (Portier and Wolfe, 1998; Boorman et al., 2000). Measurement studies of children’s exposure to electrical appliances (Kaune et al., 2000a, 2000b; Kaune et al., 2002) suggest that the elevated risks for childhood leukemia associated somewhat inconsistently with interview reports of postnatal exposures to hair dryers, television sets, and perhaps other electrical appliances in the few epidemiological studies assessing these exposures (Savitz et al., 1990; London et al., 1991; Hatch et al., 1998; Dockerty et al., 1999) may not be related to magnetic field exposures per se. Childhood leukemia was not related to magnetic fields inside infant incubators (Soderberg et al., 2002).

Chemical Exposures—Occupational, Residential, Farming, and Medications Parental Exposure to Pesticides Interview studies have linked maternal occupational exposure to pesticides during pregnancy with elevated risks of childhood ALL in offspring in China (Shu et al., 1988), mothers’ and fathers’ residential indoor and outdoor use of pesticides and herbicides in southern California (Lowengart et al., 1987; London et al., 1991), and use of a professional pest control service and indoor pesticides, particularly during pregnancy, with excess risk of childhood leukemia in northern California (Ma et al., 2002). Childhood leukemia in Denver was associated with indoor use of pest strips during the last 3 months of pregnancy and postnatally (Leiss and Savitz, 1995). Use of pesticides on farms and preconception paternal exposure to pesticides, respectively, were linked with modest increases in childhood leukemia in Germany (Meinert et al., 2000) and Quebec (Infante-Rivard and Sinnett, 1999). Frequent prenatal use of pesticides in the garden and on interior plants was associated with elevated risk of ALL among children with the CYP1A1m1 and CYP1A1m2 mutations in Quebec (Infante-Rivard et al., 1999). In contrast, childhood ALL was not linked with maternal or paternal occupational exposure to pesticides, herbicides, or insecticides during pregnancy in the Netherlands (Van Steensel-Moll et al., 1985a), with paternal employment in jobs considered to have high pesticide exposures at conception in Sweden (Feychting et al., 2001), or with paternal exposure to chlorophenate fungicides in British Columbia sawmills (Heacock et al., 2000), but paternal employment in woodworking jobs at conception was linked to a twofold risk of childhood leukemia in Sweden (Feychting et al., 2001).

Parental Exposure to Hydrocarbons or Other Solvents Subsequent evaluation of a report linking paternal hydrocarbon exposures with childhood ALL (Fabia and Thuy, 1974) has revealed inconsistent findings (Sanders et al., 1981; Gold et al., 1982; Shaw et al., 1984; Vianna et al., 1984; Lowengart et al., 1987; Van Steensel-Moll et al., 1985; Feingold et al., 1992; Shu et al., 1999b), as have a smaller number of studies assessing maternal hydrocarbon exposures (Van Steensel-Moll et al., 1985a; Shu et al., 1988; Infante-Rivard et al., 1991). Although risk of childhood leukemia was modestly increased

856

PART IV: CANCER BY TISSUE OF ORIGIN

among fathers occupationally exposed to solvents prior to conception (McKinney et al., 1991), neither childhood ALL (Infante-Rivard et al., 1991; Buckley et al., 1994) nor total childhood leukemia (Lowengart et al., 1987; McKinney et al., 1991) were significantly associated with mothers’ occupational exposures to solvents during pregnancy. Reynolds et al. (2002) found no correlation of occurrence of childhood leukemia from California cancer registry data with estimated level of pesticide application to residences based on California pesticide application and usage data. Exposure of mothers to Baygon (an anti-mosquito pesticide) during pregnancy was associated with a significantly elevated risk of infant leukemia with the MLL translocation, but not infant leukemia without the MLL translocation in a case-control study (Alexander et al., 2001).

Postnatal Exposure to Pesticides, Solvents, and Other Chemicals Compared with numerous epidemiological studies of prenatal exposures to pesticides, there have been only a few investigations of the potential role of postnatal pesticide, solvent, and chemical exposures in the etiology of childhood leukemia (Buckley et al., 1989; Meinert et al., 2000; Infante-Rivard et al., 2001; Freedman et al., 2001; Ma et al., 2002). Only one epidemiological study has evaluated the potential relationship of gene-pesticide exposure interaction in the etiology of childhood leukemia (Infante-Rivard et al., 1999). Inconsistent findings and absence of validation have led to debate about the role of maternal and paternal occupational exposures in the etiology of childhood ALL. To address lack of reproducibility of findings, improved data sources and more precise exposure assessment methods are being evaluated (Kristensen et al., 1996a; Roman et al., 1999). In addition, investigators are assessing parental occupational exposures suspected of being linked with childhood leukemia by evaluating these exposures according to specific windows of exposure (e.g., preconception, prenatal, or postnatal time periods (Shu et al., 1999b; Anderson et al., 2000).

Farming Some data link occupation as a farmer with increased risk of acute lymphocytic leukemia (Burmeister, 1990; Blair et al., 1992; Blair and Zahm, 1995), but the evidence is not conclusive.

Medications A population-based case-control study of childhood leukemia in Shanghai related the use of the antibiotics chloramphenicol and syntomycin, which is pharmacologically related to chloramphenicol, with increased risk of childhood ALL as well as AML (Shu et al., 1987). The evidence has not borne out concerns based on clinical reports that medications taken during preganancy by the mother (including narcotic analgesics, sedatives, tranquilizers, antinausea drugs) or growth hormones are linked with increased risks of childhood ALL (Linet and Cartwright, 1996; Little, 1999; McKinney et al., 1999; Wen et al., 2002).

Parental Smoking, Alcohol Consumption, Diet, and Infant Feeding Maternal smoking during pregnancy has not been linked with childhood ALL in most studies (Van Steensel-Moll et al., 1985b; Buckley et al., 1986; Magnani et al., 1990; Cnattingius et al., 1995; Sorahan et al., 1995; Shu et al., 1996). Paternal smoking during the preconception period was associated with significantly elevated risks of childhood ALL in Shanghai, China (Ji et al., 1997) and in two investigations in the United Kingdom (Sorahan et al., 1997; Sorahan et al., 2001), and with infant leukemia in the United States and Canada (Shu et al., 1996), but not in Italy (Magnani et al., 1990). Maternal alcohol consumption during pregnancy has not been shown to increase the risk of childhood ALL in offspring (Van Steensel-Moll et al., 1985b; Nishi and Miyaki et al., 1989; Shu et al., 1996; Schuz et al., 1999; InfanteRivard et al., 2002). Most investigations of breastfeeding have shown 20%–30% reduced risks of ALL among children who were breastfed during infancy (Shu et al., 1990a; Bener et al., 2000; Infante-Rivard

et al., 2000; Perillat et al., 2002), with a declining risk observed for prolonged breastfeeding in two investigations (Shu et al., 1999a; Perillat et al., 2002). Other studies have shown a smaller reduction in risk (UK Childhood Cancer Study Investigators, 2001) or no clear relationship (Schuz et al., 1999). There has been little study of the role of maternal diet during pregnancy in the etiology of childhood leukemia, although mother’s consumption of hamburger was associated with a twofold non-significant excess risk of childhood ALL in the offspring (Sarasua and Savitz, 1996). In experimental studies assessing effects of 20 bioflavonoids on primary progenitor hematopoietic cells from healthy newborns, adults, and animals, Strick et al. (2000) found that bioflavonoids caused cleavage of DNA at the same site within the MLL breakpoint cluster region as occurs with epipodophyllotoxins and doxorubicin. A protective effect for childhood ALL was linked with maternal folate or iron supplementation during pregnancy (Thompson et al., 2001). High birth weight has been linked with childhood ALL in most studies (Fasal et al., 1971; Daling et al., 1984; Shu et al., 1988; Buckley et al., 1994; Cnattingius et al., 1995; Ross et al., 1997; Westergaard et al., 1997; Murray et al., 2002; Shu et al., 2002), although a few studies did not find an association (Eisenberg and Sorahan, 1987; Robison et al., 1987; Kaye et al., 1991; Reynolds et al., 2002b).

Maternal Reproductive History and Birth Characteristics Small increases in risk of childhood leukemia were linked with previous miscarriage in some studies (Stewart et al., 1958; Graham et al., 1966; Ross et al., 1997), but not others (MacMahon and Newill, 1962; Van Steensel-Moll et al., 1985b; Cnattingius et al., 1995). A weak positive association was reported for advancing maternal age and risk of childhood leukemia in a few studies (Stewart et al., 1958; MacMahon and Newill, 1962; Stark and Mantel, 1966; Dockerty et al., 2001), but not most (Graham et al., 1966; Shu et al., 1988; Ross et al., 1997; Van Steensel-Moll et al., 1985b; Cnattingius et al., 1995; Shu et al., 2002a). Similarly, being first-born was linked with elevated risks in a few investigations (Stewart et al., 1958; MacMahon and Newill, 1962; Stark and Mantel, 1986; Dockerty et al., 2001), but not the majority (Shaw et al., 1984; Kaye et al., 1991; Cnattingius et al., 1995; McKinney et al., 1999; Shu et al., 1999a; Infante-Rivard et al., 2000; Neglia et al., 2000; Perrillat et al., 2002; Shu et al., 2002a).

Infectious Disorders and Immune Function A large literature on childhood leukemia clusters (reviewed in Linet, 1985; Little, 1999) has suggested an infectious etiology for childhood ALL. Inconsistent results have been reported from studies evaluating indirect measures linking infectious agents with childhood ALL including interview-derived data on history of maternal infection during pregnancy (Till et al., 1979; Van Steensel-Moll et al., 1985b; Roman et al., 1997) or vaccination history (Comstock et al., 1971; Comstock et al., 1975; Snider et al., 1978; Groves et al., 1999; Auvinen et al., 2000; Groves et al., 2002). Reported daycare attendance (Petridou et al., 1997; Dockerty et al., 1999; Neglia et al., 2000; Rosenbaum et al., 2000; Infante-Rivard et al., 2000; Perrillat et al., 2002; Ma et al., 2002) is associated with reduced risks. Household crowding (Murray et al., 2002), medical conditions suggesting poor hygiene (Smith et al., 1998), and household pets (Petridou et al., 1997; Swensen et al., 2001) have not been adequately studied. There is some evidence that measures of higher socioeconomic status are linked with increased risk of childhood ALL (Chow et al., 1996; Draper et al., 1991; Dockerty et al., 2001). A growing number of studies have shown elevated risks of childhood ALL in geographic regions with high levels of population mixing (e.g., previously isolated areas in which there was a recent increase in population density, areas of population growth, and regions with population movements during wartime, increased social contact during commuting, or mass tourism (Kinlen, 1988; Kinlen et al., 1990; Kinlen and Hudson, 1991; Petridou et al., 1991; Kinlen et al., 1993b; Kinlen and John, 1994; Kinlen and Petridou, 1995; Stiller and Boyle, 1996; Alexander et al., 1997). Screening

The Leukemias for four lymphotropic herpesvirus genomes in leukemic samples using conventional molecular techniques and sensitive real-time PCR revealed no novel herpesvirus genomes (MacKenzie et al., 2001). Samples from children with ALL demonstrated no evidence of genomes of the JC and BK polyoma viruses (Smith et al., 1999; MacKenzie et al., 1999). A history of one or more allergic disorders was linked with a significantly reduced risk of childhood ALL (Wen et al., 2000).

CHRONIC LYMPHOCYTIC LEUKEMIA Physical Agents—Ionizing and Non-Ionizing Radiation Chronic lymphocytic leukemia has not been associated with exposure to ionizing radiation (Boice et al., 1996; UNSCEAR, 2000). A comprehensive assessment of the epidemiological literature on health effects from exposure to power-line frequency electric and magnetic field radiation concluded that there was limited evidence of a relationship between this exposure and CLL (Portier and Wolfe, 1998). The conclusion derived from the modestly elevated risks of CLL based on small numbers of cases with high measured exposures from three investigations (Floderus et al., 1993; Theriault et al., 1994; Feychting et al., 1997), and a meta-analysis of 38 studies employing job title as the primary exposure measure that found a significantly elevated 60% increase in risk for CLL (Kheifets et al., 1997).

Chemical Exposures—Manufacturing, Farming, and Medications Benzene While some case-control studies have suggested a link with CLL (Arp et al., 1983), large cohort studies in the United States (Rinsky et al., 1987; Utterback et al., 1995) and China (Hayes et al., 1997) have shown little evidence of increased risk of CLL associated with benzene exposure, although the finding must be interpreted in light of the rarity of CLL in Asians (Parkin et al., 2002).

Petroleum Industry Workers A few studies have described excesses of lymphocytic leukemia or CLL (Bertazzi et al., 1989; Wongsrichanalai et al., 1989) among petroleum industry workers, but no CLL excess was found in other studies (Marsh et al., 1991; Rushton, 1993; Honda et al., 1995; Sathiakuman et al., 1995; Schnatter et al., 1996b; Divine et al., 1999) or in a leukemia type-specific meta-analysis of 208,000 workers with potential exposure to benzene (Raabe and Wong, 1996). Gasoline service station workers also had no excess of CLL (Lynge et al., 1997).

857

1987) were followed by inconsistent results for workers exposed to monomeric styrene and butadiene (Ott et al., 1980; Wong, 1990; Matanoski et al., 1990; Cole et al., 1993). A retrospective cohort study of 40,683 workers (from Denmark, Finland, Norway, Sweden, Italy, and the United Kingdom) employed in the reinforced plastics industry revealed no excesses of neoplasms of the lymphatic and hematopoietic system overall or evidence of increasing risk with longer duration of exposure, but mortality from leukemia and lymphoma rose twofold 20 years after first exposure (Kogevinas et al., 1993). Small excesses of lymphohematopoietic cancers were reported in some studies of 1,3-butadiene production facilities (Divine et al., 1993), and elevated risks were found in workers employed in the 1960s in companies producing reinforced plastics in Denmark (Kolstad et al., 1993).

Ethylene Oxide Although specific types of lymphoid cancers were often not reported in studies of workers exposed to ethylene oxide, excesses of lymphatic leukemia and lymphoma were observed in three small Swedish cohorts (Hogstedt et al., 1979a; 1979b; Hogstedt et al., 1986) and in US workers (Stayner et al., 1993). The study of US workers showed a dose-response pattern, but only non-Hodgkin lymphoma was significantly elevated among workers from 14 plants producing sterilized medical supplies (Steenland et al., 1991). Findings from other studies were mixed, with elevated leukemia linked with production of ethylene or propylene chlorhydrin rather than ethylene oxide at two US plants (Greenberg et al., 1990; Benson et al., 1993), but no excesses were seen in German (Theiss et al., 1981) or other US plants (Morgan et al., 1981; Teta et al., 1993). It is possible that CLL may be linked with ethylene oxide in view of the excess risks seen for non-Hodgkin lymphoma and lymphoid leukemias in the positive studies, but investigations focusing on total leukemia mortality (because death certificates often lack specificity of cell type) could fail to identify excess risks for CLL, even if a causal association with ethylene oxide exists.

Chemists Small excesses of lymphopoietic malignancies (often including CLL and non-Hodgkin lymphoma, and sometimes Hodgkin lymphoma) have been reported among chemists (Hunter et al., 1993), workers in biomedical research (Cordier et al., 1995), science technicians (Burnett et al., 1999) and clinical laboratory technicians working in pathology, cytology, and other laboratories (Gustavsson et al., 1999). Excess risks of lymphopoietic tumors have been linked with a high probability to chemical exposures, but other possibilities could not be ruled out, and risks for CLL have not been separately reported in several of these studies.

Agricultural Exposures Rubber Workers Benzene was used extensively in the rubber industry in earlier years, but replaced with other chemicals (some contaminated with small amounts of benzene) in recent times. Excess risks of lymphocytic leukemia observed in large cohorts of rubber manufacturing workers in the United Kingdom (Fox and Collier, 1976; Parkes et al., 1982) and the United States (Mancuso, 1975; Delzell and Monson, 1981) were followed by more detailed investigations that linked solvent exposures (McMichael et al., 1976), and specifically benzene, carbon disulfide, carbon tetrachloride, xylene, and others with increased risks of lymphocytic leukemia based on small numbers of cases (Arp et al., 1983; Wilcosky et al., 1984; Checkoway et al., 1984). More recent studies in rubber workers in the United Kingdom have shown no excess of CLL or lymphocytic leukemia (Sorahan and Cooke, 1989), but a small increase was reported in a detailed review of 12 cohort studies (Kogevinas et al., 1998).

Styrene and Butadiene Earlier studies reporting excesses of leukemia and lymphoma in small cohorts of workers in plants producing, polymerizing, and/or processing styrene monomers and butadiene (McMichael et al., 1976; Monson and Nakano, 1976; Hodgson and Jones, 1985; Downs et al.,

Some studies have implicated farming and related exposures with elevated risks for CLL (Blair and White, 1985; Brown et al., 1990; Kelleher et al., 1988). Although based on small studies, specific exposures have included DDT (Flodin et al., 1988), animal breeding (Amadori et al., 1995), and employment in flour mills (Alavanja et al., 1990). Insecticides, carbamates, and phosphates were linked with CLL in a highly agricultural area in Italy (Nanni et al., 1996).

Other Occupational Exposures Chronic lymphocytic leukemia has also been associated with working in underground coal mining (Gilman et al., 1985), carpet manufacturing (Cartwright et al., 1987a; O’Brien and Decoufle, 1988), sawmills or industries using lumber products (Burkhardt, 1982), and with employment as a barber or hairdresser (Spinelli et al., 1984; Teta et al., 1984; Miligi et al., 1999), or as a vehicle mechanic (Hunting et al., 1995).

Smoking, Diet, Alcohol, and Hair Dyes Three cohort investigations (Kinlen and Rogot, 1988; Garfinkel and Boffetta, 1990; Linet et al., 1991) and one case-control study (Brown et al., 1992b) have noted elevated risks of lymphocytic leukemia

858

PART IV: CANCER BY TISSUE OF ORIGIN

among smokers, but recent large cohort studies (Friedman, 1993; Adami et al., 1998) have reported no association. There has been little effort to evaluate diet or alcohol as potential risk factors for CLL. CLL (as well as AML, and myelodysplastic syndromes) have been linked with employment in cosmetology in some investigations (reviewed in Linet and Cartwright, 1996; Miligi et al., 1999). Risk of CLL was not elevated among men or women using hair dyes in a population-based case-control study in Nebraska (Zahm et al., 1992), nor in large cohorts of women (Grodstein et al., 1994) or men and women (Thun et al., 1994; Altekruse et al., 1999).

Immune Dysfunction and Infections It has been hypothesized that chronic infections, allergic disorders, and autoimmune diseases may result in long-lasting effects on the immune system, particularly on B-cell lymphocytes. The chronic immune dysregulation and related perturbations of lymphocytes associated with autoimmune diseases may also predispose to CLL. Small increased risks of CLL have been linked with chronic infectious disorders (syphilis, tuberculosis, chronic urinary tract infections, chronic ear infections, bronchitis) (Cartwright et al., 1987b; Rosenblatt et al., 1991), autoimmune diseases such as rheumatoid arthritis and hyperthyroidism (Zheng et al., 1993), allergic disorders such as asthma and eczema (Zheng et al., 1993), and appendectomy (Rosenblatt et al., 1991; Zheng et al., 1993) in some studies, although results have not been consistent. Other investigations have shown opposite effects, with reduced risks of CLL linked with rheumatoid arthritis (Doody et al., 1992), allergic disorders (Linet et al., 1986), and surgical ablation of lymphoid tissue (Linet et al., 1986). A recent history of herpes zoster has been associated with CLL (Cartwright et al., 1987b), but it is unclear whether this represented an early clinical manifestation of CLL or was etiologically related. Overall, results from studies examining prior medical conditions, surgical, and other treatments and risk of CLL have shown no clear, consistent relationships (reviewed in Sgambati et al., 2001).

Molecular Epidemiology of Lymphoid Disorders There are a number ways in which genetic variation may impact the risk of developing lymphoproliferative disease. A brief general description of xenobiotic metabolism is provided above in the section on myeloid disorders. A few small studies have examined the role of polymorphisms of the metabolizer enzymes on the risk of adult and childhood ALL, but these are not discussed further because data are insufficient to draw firm conclusions. Instead, we focus on the role played by folic acid metabolism in protecting against ALL, and how genetic variation may affect immune response and risk of lymphoproliferative diseases. The impact of double-strand break repair on risk of lymphoma is briefly considered, because CLL overlaps with welldifferentiated lymphomas.

Folic Acid Metabolism Individuals with specific polymorphisms in the methylenetetrahydrofolate reductase gene (MTHFR) were observed to be at reduced risk of developing adult ALL (Skibola et al., 1999). Folic acid is essential in the transfer of methyl groups to various biochemical targets in mammalian tissues involved in amino acid metabolism and in the synthesis of the purine and pyrimidine components of DNA and RNA (Selhub et al., 1992). Depleted folic acid levels lead to elevated uracil incorporation into DNA (Pogribny et al., 1997; Blount et al., 1997) and diminished DNA repair capacity (Choi et al., 1998), resulting in DNA strand breaks (Kim et al., 1997) and chromosomal damage (Fenech et al., 1997). Impaired folate status modulates the process of neoplastic transformation in selected epithelial tissues (including colon, cervix, lung, and esophagus) (Glynn et al., 1996; Kwasneiewska et al., 1997; Mason and Levesque, 1996). Folic acid deficiency has also been associated with neural tube defects in newborns (Bower et al., 1989), and elevated plasma homocysteine levels linked with increased cardiovascular disease (Jacob et al., 1998; Ubbink et al., 1998). The 5,10-methylenetetrahydrofolate reductase

(MTHFR) enzyme, critical in the regulation of folate and methionine metabolism, catalyzes the reduction of 5,10-methyleneTHF (required for purine and pyrimidine synthesis) to 5-methylTHF. A common polymorphism (677CÆT) in the MTHFR gene is the homozygous (677TT) allelic variant, in which up to 15% of individuals have reduced enzymatic activity, thus affecting folate metabolism (de Franchis et al., 1995; Frosst et al., 1995; Nishio et al., 1996). This polymorphism is also implicated in neural tube defects, hyperhomocysteinemia, and occlusive vascular disease (van der Put et al., 1995; Ou et al., 1996; Ma et al., 1996; Engbersen et al., 1995). A second common polymorphism (glutamate to alanine (AÆC) transversion at position 1298 of the MTHFR gene) has also been implicated in neural tube defects (van der Put et al., 1995). Allele frequencies occurring in as many as 33% of persons (van der Put et al., 1995; Weisberg et al., 1998) may lead to a reduction in enzyme activity. Previous studies have shown that consumption of adequate dietary folate and methionine by individuals homozygous for this mutation may reduce incidence of colorectal cancer (Chen et al., 1996; Ma et al., 1997).

Cytokine and Chemokine Polymorphisms Cytokines are fundamental to both structural and functional aspects of the immune system (Mire-Sluis, 1999). Cytokines are humoral immunomodulatory proteins or glycoproteins that control or modulate the activities of target cells within hematopoietic and other systems, by binding to specific cytokine receptor ligands, and initiating signal transduction and second messenger pathways (Anhuf et al., 2000). The cytokine network is highly complex, containing interactive cascades of gene activation and suppression (Townsend and McKenzie, 2000). In the interaction of pro-inflammatory (designated as TH-1 functions) and anti-inflammatory (designated as TH-2 functions) processes in specific malignant, immune, or infectious diseases, one of these immune functional activities predominates. The relationship between TH1 and TH2 responses may be established early in life as a consequence of environmental exposures and has been implicated in ALL (Greaves, 1997), asthma (Huang et al., 2001), and autoimmune diseases (Youn et al., 2000). Most cytokine and cytokine receptor genes are polymorphic (Shieh et al., 2000; Wang et al., 2000), but the types and frequencies of polymorphisms at different alleles and the potential etiological relationship of specific polymorphisms to lymphoproliferative disorders (or most other diseases) has not been examined systematically (Demeter et al., 1997). High producer haplotypes have been evaluated in relation to multiple myeloma (Feugas et al., 1997; Hjelmstrom et al., 1998; Date et al., 1999; Higham et al., 2000; Davies et al., 2000) and malignant lymphoma (Warzocha et al., 1998). The CCR5-delta32 allele has been linked with lower risk of developing nonHodgkin lymphoma among patients with AIDS (Dean et al., 1999).

Non-Homologous End Joining (NHEJ) Double-strand DNA breaks (hereafter abbreviated as DNA DSBs) are the precursor lesions of chromosome translocations (Taccioloi et al., 1993). In mammalian cells, DNA DSBs trigger several damage response mechanisms including cell cycle checkpoint arrest, apoptosis, or DNA repair processes (Carney et al., 1998). The major repair mechanism in mammalian cells is non-homologous end joining (NHEJ), in which the chromosome breaks are rearranged during the V(D)J recombination process (Takata et al., 1998). Several genes (including ATM, NBS, hMrell, and Rad 50) function as early sensors of DNA DSBs and trigger signal transduction pathways via a phosphorylation cascade that can result in cell cycle arrest at G1/S, S, and G2M (Melton et al., 1998; Inbar and Kupiec, 2000). P53 is involved in some, but not all, of these processes (Gurley et al., 1998; Aubrecht et al., 1999). Inherited disorders involving these genes, such as ataxia telangiectasia (involving mutations in ATM), a variant form called ataxia telangiectasia-like disorder (involving Mrell), and Nijmegen breakage syndrome (involving the NBS1 gene), are associated with immunodeficiency and increased risk of occurrence of lymphoproliferative malignancies (Girard et al., 2000; Seidemann et al., 2000). Experimental studies in human cell lines and in mice defective in NHEJ have shown inability to carry out V(D)J recombination in the

The Leukemias former and severe combined immunodeficiency phenotypes (SCID) in the latter (Zhu et al., 1996; Xu et al., 1998; Bishop et al., 2000; Bogue et al., 1997).

Familial Occurrences Familial aggregation of ALL is rare, but has been reported in kindreds in which consanguinity and intermarriage were prevalent (Kende et al., 1994) and in siblings carrying a heterozygous frameshift mutation in the Fanconi anemia C gene (Rischewski et al., 2000). Familial tendency to CLL, recognized for more than 50 years (Vidabaek, 1947), is one of the strongest risk factors for development of CLL (Sgambati et al., 2001). First-degree relatives with leukemia were more frequent in families of CLL cases than in CML cases (Gunz et al., 1978), but only a small percent of CLL cases have affected close family members (Linet et al., 1989; Horowitz, 1997), although the proportion with familial CLL may be higher in Ashkenazi Jews of Eastern European or Russian descent (Cuttner et al., 1992). Within families with two or more cases, leukemia subtypes are generally concordant, particularly for CLL (Pottern et al., 1991; Linet and Pottern, 1992). Recent studies of familial CLL have shown anticipation (Yuille et al., 1998; Goldin et al., 1999). Postulated genetic mechanisms include inherited germline mutations (not supported in studies of monozygotic twins), primary immunologic alterations, sharing of common haplotypes, and/or consanguinity. Environmental influences may also be important either as primary causes, or in conjunction with specific, but unidentified polymorphisms.

Genetic Syndromes Congenital Immunodeficiency Disorders Genetically determined immunodeficiency disorders, including congenital X-linked immunodeficiency, ataxia telangiectasia, and Wiskott-Aldrich syndrome (Filipovich et al., 1992), are associated with dramatically elevated relative risks of non-Hodgkin lymphoma and related to lymphoproliferative disorders, with a cumulative incidence ranging from 1%–25% (Filipovich et al., 1994), and half of 491 cancers occurring in patients reported to an international registry of immunodeficiency disorders were non-Hodgkin lymphoma (Filipovich et al., 1987), but risks of non-Hodgkin lymphoma have not been described according to subtype. Risks of non-Hodgkin lymphoma ranged from 30- to 100-fold elevated in patients with primary hypogammaglobulineia (mostly common variable immunodeficiency) (Kinlen et al., 1985; Sneller et al., 1993; Cunningham-Rundles and Bodian, 1999).

Other Genetic Syndromes Approximately 5% of ALL and AML have been associated with inherited genetic syndromes, often involving genes functionally linked to DNA repair or other aspects of genomic stability (Taylor and Birch, 1996; Birch, 1999). Children with Down syndrome (trisomy 21) are at increased risk of developing acute leukemia, particularly the M7 (megakaryoblastic) variant of AML as well as ALL (Hasle et al., 2000; Malkin et al., 2000, Hill et al., 2003). The pathogenesis may involve a mutant p53 gene playing a critical role in the transformation of the transient form of leukemia seen in patients with Down syndrome to full-blown acute megakaryoblastic leukemia, although recently the role of GATA1 mutations has been recognized. The Li-Fraumeni cancer family syndrome is highly penetrant and includes childhood leukemia, sarcomas, breast cancer, brain tumors, adrenocortical carcinoma, and multiple primary cancers (Hisada et al., 1998). Most LiFraumeni families have inherited germline mutations in the p53 gene, whereas the tumor suppressor hCHK2 gene, whose activation prevents cellular entry into mitosis, has been recently implicated (Bell et al., 1999). To date, investigators have found no evidence of germline p53 mutations among most familial leukemia pedigrees (Felix et al., 1992). In 91 families with P53 germline mutations, ALL and Hodgkin lymphoma comprised only 4.2% of the 475 tumors occurring in affected members (Kleihues et al., 1997). The p73 gene, which encodes a protein homologous to the p53 protein, is aberrantly methylated in

859

about 30% of primary ALL and nonendemic Burkitt lymphomas suggesting a possible role (Corn et al., 1999). Patients with ataxia telangiectasia are characterized by inherited mutations within the ATM gene and an increased risk of developing lymphoproliferative disorders, which are frequently of T-cell origin and often involve T-cell receptor rearrangement (Hecht and Hecht, 1990; Khanna, 2000). A study of 24 families with familial B-cell CLL found little evidence that ATM was likely to make a significant contribution to the threefold to fourfold excess risk of CLL in close relatives of affected CLL cases (Bevan et al, 1999). Bloom’s syndrome, associated with both ALL and AML, is a disorder in which the specific chromosome bands non-randomly affected by spontaneous chromosomal aberrations are also significantly correlated with the fragile sites, breakpoints, and rearrangements characteristic of AML. Other than these and other rare defined germline syndromes, the existing literature suggests no major role for germ cell mutations in parents of children with ALL (Rinaldi et al., 1999).

PREVENTION AND FUTURE DIRECTIONS Despite advances in molecular biology and genetic and mechanistic aspects of leukemogenesis, progress has continued to be slow in identifying etiologic factors. Because of the rarity of the leukemias and the biologic and epidemiologic evidence of distinct subtypes, many of the etiologic leads described in this review should be pursued in large case-control and nested studies in carefully chosen cohorts in conjunction with detailed molecular, cytogenetic, genetic, and immunophenotypic characterization of cases. Prudence suggests that efforts be made to limit occupational, environmental, and medical exposures to radiation and chemicals implicated in leukemia. Physicians should continue to weigh the leukemia risks against the benefits of diagnostic and therapeutic radiation and cytotoxic chemotherapy agents. Doses of diagnostic radiation should be decreased to the extent feasible, while maintaining an acceptable quality of image resolution. Efforts should be implemented to limit radiation scatter and organ dose, while maintaining treatment efficacy. The number, dose, and duration of cytotoxic therapies should be minimized to the extent feasible and consistent with good treatment outcomes, particularly in children and older patients. Although further research is needed to confirm associations of leukemia with drugs other than cytotoxic agents, physicians should limit prescriptions of chloramphenicol in all forms (including ophthalmic preparations) unless no other options are available. Although most efforts to reduce benzene-related exposures have been directed at the workplace, other postulated sources should be limited as these sources are confirmed. Further studies of benzeneexposed groups are needed to ascertain risks of leukemia types other than AML and aplastic anemia, and to determine whether other hematopoietic or lymphoproliferative malignancies occur excessively. Efforts should be undertaken to reduce the percent of benzene in gasoline, paints, solvents, and other sources to limit exposures. Devices should be installed and other strategies developed for minimizing exposure to benzene in gasoline at service stations. Clarification of possible leukemogenic exposures in pesticides, among farmers, and in the petrochemical, rubber, styrene, butadiene, formaldehyde, and other implicated industries and occupations is necessary. To evaluate leukemogenic risk of these and other suspected leukemogens, populations with valid exposure measurements should be sought or the best available measurements should be incorporated. Molecular and genetic studies should be carefully and thoughtfully incorporated to enable investigation of gene-environment interactions. While cessation of cigarette smoking is the overriding goal to reduce smoking-related excesses of many types of cancer, it is important to identify and eliminate leukemogenic agent(s) in cigarettes. It is also important to direct additional efforts at smoking cessation among men before conception. Important priorities for prevention of childhood ALL include identification of the agent(s) related to increased risks associated with population mixing, and reasons for the elevated risks associated with high

860

PART IV: CANCER BY TISSUE OF ORIGIN

birthweight. It is important to conduct additional studies to evaluate whether mother’s consumption during pregnancy of DNA topoisomerase II inhibitor-containing foods or exposures to specific pesticides are related to risk of infant leukemias. As data continues to accumulate that several forms of childhood ALL and AML may be preceded by prenatal translocations, additional efforts should be undertaken to ascertain whether suspected leukemogenic agents may be responsible, and to follow up those children whose cord blood or Guthrie cards indicated presence at birth of specific translocations linked with childhood leukemia. Familial aggregation of leukemias, other hematologic malignancies, autoimmune conditions, and immunologic abnormalities suggests the need for detailed genetic studies along with evaluation of suspected leukemogenic exposures. International registries should be established of persons with various congenital disorders, particularly those characterized by abnormalities of genes functionally linked to DNA repair or other aspects of genomic instability or immunodeficiency to generate sufficient numbers of cases to pursue detailed epidemiologic investigations in conjunction with state-of-the-art molecular genetic studies. The discovery of a retrovirus as a cause of adult T-cell leukemia/lymphoma resulted in some initial efforts to look for viral genomes in childhood leukemia. While these efforts have not yet resulted in identification of additional leukemogenic viruses, it is important to pursue strategies for identifying agents related to clustering of childhood ALL and excess risks associated with population mixing. Acknowledgments The authors express their appreciation to John Lahey of Information Management Systems, Inc. for preparation of the figures, and their gratitude to the staff of the NCI SEER registries for the high quality of data collection and preparation.

References Adami J, Nyren O, Bergstrom R, et al. 1998. Smoking and the risk of leukemia, lymphoma, and multiple myeloma (Sweden). Cancer Causes Control 9:49–56. Adamson RH, Sieber SM. 1980. Carcinogenic potential of drugs. J Environ Pathol Toxicol 3 (4 Spec No):377–385. Ager EA, Schuman LM, Wallace HM, et al. 1965. An epidemiological study of childhood leukaemia. J Chron Dis 18:113–132. Aguiar RC. 1998. Therapy-related chronic myeloid leukemia: An epidemiological, clinical and pathogenetic appraisal. Leuk Lymphoma 29:17–26. Ahlbom A, Day N, Feychting M, et al. 2000. A pooled analysis of magnetic fields and childhood leukaemia. Br J Cancer 83:692–698. Alavanja MC, Blair A, Masters MN. 1990. Cancer mortality in the U.S. flour industry. J Natl Cancer Inst 82:840–848. Alavanja MC, Sandler DP, McMaster SB, et al. 1996. The Agricultural Health Study. Environ Health Perspect 104:362–369. Alexander FE, Chan LC, Lam TH, et al. 1997. Clustering of childhood leukaemia in Hong Kong: Association with the childhood peak and common acute lymphoblastic leukaemia and with population mixing. Br J Cancer 75:457–463. Alexander FE, Patheal S, Biondi A, et al. 2001. Transplacental chemical exposure and risk of infant leukaemia with MLL gene fusion. Cancer Res 61:2542–2546. Allan JM, Wild CP, Rollinson S, et al. 2001. Polymorphism in glutathione Stransferase P1 is associated with susceptibility to chemotherapy-induced leukaemia. Proc Natl Acad Sci USA 98:11592–11597. Altekruse SF, Henley SJ, Thun MJ. 1999. Deaths from hematopoietic and other cancers in relation to permanent hair dye use in a large, prospective study (United States). Cancer Causes Control 10:617–625. Alter BP. 1996. Fanconi’s anemia and malignancies. Am J Hematol 53:99–110. Alter BP. 2003. Cancer in Fanconi anemia, 1927–2001. Cancer 97:425–440. Anderson LM, Diwan BA, Fear NT, et al. 2000. Critical windows of exposure for children’s health: Cancer in human epidemiological studies and neoplasms in experimental animal models. Environ Health Perspect 08:573–594. Andersen MK, Pedersen-Bjergaard J. 2000. Increased frequency of dicentric chromosomes in therapy-related MDS and AML compared to de novo disease is significantly related to previous treatment with alkylating agents and suggests a specific susceptibility to chromosome breakage at the centromere. Leukemia 14:105–111.

Andersson M, Carstensen B, Visfeldt J. 1993. Leukemia and other related hematological disorders among Danish patients exposed to Thorotrast. Radiat Res 134:224–233. Anhuf D, Weissenback M, Schmitz J, et al. 2000. Signal transduction of IL-6, leukemia-inhibitory factor, and oncostatin M: Structural receptor requirements for signal attenuation. J Immunol 165:2535–2543. Aoki K, Kurihara M, Hayakawa N, et al. 1992. Death rates for malignant neoplasms for selected sites by sex and five year age group in 33 countries 1953–57 through 1983–87. Nagoya, Japan, University of Nagoya Press, pp. 503–536. Arp EW, Wolfe PH, Checkoway H. 1983. Lymphocytic leukemia and exposures to benzene and other solvents in the rubber industry. J Occup Med 25:598–602. Arruda VR, Lima CS, Grignoli CR, et al. 2001. Increased risk for acute myeloid leukaemia in individuals with glutathione S-transferase mu 1 (GSTM1) and theta 1 (GSTT1) gene defects. Eur J Haematol 66:383–388. Atoyebi W, Kusec R, Fidler C, et al. 1997. Glutathione S-transferase gene deletions in myelodysplasia. Lancet 349:1450–1451. Aubrecht J, Secretan MB, Bishop AJ, et al. 1999. Involvement of p53 in X-ray induced intrachromosomal recombination in mice. Carcinogenesis 20:2229–2236. Aul C, Gattermann N, Schneider W. 1995. Epidemiological and etiological aspects of myelodysplastic syndromes. Leuk Lymphoma 16:247–262. Auvinen A, Hakulinen T, Groves F. 2000. Haemophilus influenzae type B vaccination and risk of childhood leukaemia in a vaccine trial in Finland. Br J Cancer 83:956–958. Avnon L, Oryan I, Kordysh E, et al. 1998. Cancer incidence and risks in selected agricultural settlements in the Negev of Israel. Arch Environ Health 53:336–343. Axelsson U, Fredrikson M, Akerblom G, et al. 2002. Leukemia in childhood and adolescence and exposure to ionising radiation in homes built from uranium-containing alum shale concrete. Epidemiology 13:146–150. Band PR, Le ND, Fang R, et al. 1996. Cohort study of Air Canada pilots: Mortality, cancer incidence, and leukaemia risk. Am J Epidemiol 143:137–143. Basu T, Gale RE, Langabeer S, et al. 1997. Glutathione S-transferase theta 1 (GSTT1) gene defect in myelodysplasia and acute myeloid leukaemia. Lancet 349:1450. Bauer AK, Faiola B, Abernethy DJ, et al. 2003. Genetic susceptibility to benzene-induced toxicity: Role of NADPH: quinone oxidoreductase-1. Cancer Res 63:929–935. Baverstock KF, Papworth DG. 1986. The UK radium luminiser survey. In: Gossner W, Gerber GB, eds. The Radiobiology of Radium. Munich, Urban pp. 22–28. Beaty O III, Hudson MM, Greenwald C, et al. 1995. Subsequent malignancies in children and adolescents after treatment for Hodgkin’s disease. J Clin Oncol 13:603–609. Bell DW, Varley JM, Szydlo TE, et al. 1999. Heterozygous germ line hCHK2 mutations in Li-Fraumeni syndrome. Science 286:2528–2531. Bener A, Denic S, Galadari S. 2001. Longer breast-feeding and protection against childhood leukaemia and lymphomas. Eur J Cancer 37:234– 238. Bennett JM. 2000. World Health Organization classification of the acute leukemia-myelodysplastic syndrome. Int J Hematol 72:131–133. Bennett JM, Catovsky D, Daniel MT, et al. 1976. Proposals for the classification of the acute leukemias. Br J Hematol 33:451–458. Bennett JM, Catovsky D, Daniel MT, et al. 1982. Proposals for the classification of the myelodysplastic syndromes. Br J Haematol 51:189–199. Bennett JM, Catovsky D, Daniel MT, et al. 1985. Criteria for the diagnosis of acute leukemia of megakaryocyte lineage (M7). A report of the French-American-British Cooperative Group. Ann Intern Med 103: 460–462. Bennett JM, Catovsky D, Daniel MT, et al. 1989. Proposal for the classification of chronic (mature) B and T lymphoid leukemias. J Clin Pathol 42:567–584. Bennett JM, Catovsky D, Daniel MT, et al. 1994. The chronic myeloid leukaemias: Guidelines for distinguishing chronic granulocytic, atypical chronic myeloid and chronic myelomonocytic leukaemia. Proposals by the French-American-British Cooperative Leukaemia Group. Br J Haematol 87:746–754. Benson LO, Teta MJ. 1993. Mortality due to pancreatic and lymphopoietic cancers in chlorhydrin production workers. Br J Ind Med 50:710–716. Bertazzi PA, Pesatori AC, Zocchetti C, et al. 1989. Mortality study of cancer risk among oil refinery workers. Int Arch Occup Environ Health 61:261–270. Bevan S, Catovsky D, Marossy A, et al. 1999. Linkage analysis for ATM in familial B cell chronic lymphocytic leukaemia. Leukemia 13:1497–1500. Birch JM. 1999. Genes and cancer. Arch Dis Child 80:1–3.

The Leukemias Bishop AJR, Barlow C, Wynshaw-Boris AJ, et al. 2000. ATM deficiency causes an increased frequency of intrachromosomal homologous recombination in mice. Cancer Res 60:395–399. Bjork J, Albin M, Mauritzson N, et al. 2000. Smoking and myelodysplastic syndromes. Epidemiology 11:285–291. Black D. 1984. Investigation of the Possible Increased Incidence of Cancer in West Cumbria. London, Her Majesty’s Stationery Office. Blackwelder WC, Yano K, Rhoads GG, et al. 1980. Alcohol and mortality: The Honolulu Heart Study. Am J Med 68:164–169. Blair A, Hayes HM Jr. 1982. Mortality patterns among US veterinarians, 1947–1977: An expanded study. Int J Epidemiol 11:391–397. Blair A, Stewart PA. 1992. Do quantitative exposure assessments improve risk estimates in occupational studies of cancer? Am J Ind Med 21:53– 63. Blair A, Stewart P, O’Berg M, et al. 1986. Mortality among industrial workers exposed to formaldehyde. J Natl Cancer Inst 76:1971–1984. Blair A, Stewart PA, Hoover RN. 1990. Mortality from lung cancer among workers employed in formaldehyde industries. Am J Ind Med 17:683– 699. Blair A, White D. 1985. Leukemia cell types and agricultural practices in Nebraska. Arch Environ Health 40:211–214. Blair A, Zahm SH, Pearce NE, et al. 1992. Clues to cancer etiology from studies of farmers. Scand J Work Environ Health 18:209–215. Blair A, Zahm SH. 1995. Agricultural exposures and cancer. Environ Health Perspect 103:205–208. Blanco JG, Edick MJ, Hancock ML, et al. 2002. Genetic polymorphisms in CYP3A5, CYP3A4 and NQO1 in children who developed therapy-related myeloid malignancies. Pharmacogenetics 12:605–611. Blattner WA. 1999. Human retroviruses: Their role in cancer. Proc Assoc Am Physicians 111:563–572. Bloom GE, Warner S, Gerland PS, et al. 1966. Chromosome abnormalities in constitutional aplastic anemia. N Engl J Med 274:8–14. Blount BC, Mack MM, Wehr CM, et al. 1997. Folate deficiency causes uracil misincorporation into human DNA and chromosome breakage: Implications for cancer and neuronal damage. Proc Natl Acad Sci USA 94:3290– 3295. Bogue MA, Wang C, Zhu C, et al. 1997. V(D)J recombination in Ku86deficient mice: distinct effects on coding, signal, and hybrid joint formation. Immunity 7:37–47. Boice JD Jr, Blettner M, Kleinerman RA, et al. 1987. Radiation dose and leukemia risk in patients treated for cancer of the cervix. J Natl Cancer Inst 79:1295–1311. Boice JD Jr, Morin MM, Glass AG, et al. 1991. Diagnostic x-ray procedures and risk of leukemia, lymphoma, and multiple myeloma. J Am Med Assoc 265:1290–1294. Boorman GA, Owen RD, Lotz WG, et al. 2000a. Evaluation of in vitro effects of 50 and 60 Hz magnetic fields in regional EMF exposure facilities. Radiat Res 153:648–657. Boorman GA, McCormick DL, Ward JM, et al. 2000b. Magnetic fields and mammary cancer in rodents: A critical review and evaluation of published literature. Radiat Res 153:617–626. Bowen DT, Frew ME, Rollinson S, et al. 2003. CYP1A1*sB (Val) allele is overrepresented in a subgroup of acute myeloid leukemia patients with poor-risk karyotype associated with NRAS mutation, but not associated with FLT3 internal tandem duplication. Blood 101:2770–2774. Bower C, Stanley FJ. 1989. Dietary folate as a risk factor for neural-tube defects: Evidence from a case-control study in Western Australia. Med J Aust 150:613–619. Broeker PL, Super HG, Thirman MJ, et al. 1996. Distribution of 11q23 breakpoints within the MLL breakpoint cluster region in de novo acute leukemia and in treatment-related acute myeloid leukemia: Correlation with scaffold attachment regions and topoisomerase II consensus binding sites. Blood 87:1912–1922. Brondum J, Shu XO, Steinbuch M, et al. 1999. Parental cigarette smoking and the risk of acute leukemia in children. Cancer 85:1380–1388. Brown AG, Ross FM, Dunne EM, et al. 1993. Evidence for a new tumor suppressor locus (DBM) in human B-cell neoplasia telomeric to the retinoblastoma gene. Nat Genet 3:67–72. Brown LM, Blair A, Gibson R, et al. 1990. Pesticide exposures and other agricultural risk factors for leukemia among men in Iowa and Minnesota. Cancer Res 50:6585–6591. Brown LM, Gibson R, Blair A, et al. 1992b. Smoking and risk of leukemia. Am J Epidemiol 135:763–768. Brown LM, Gibson R, Burmeister LF, et al. 1992a. Alcohol consumption and risk of leukemia, non-Hodgkin’s lymphoma, and multiple myeloma. Leuk Res 16:979–984. Brownson RC, Novotny TE, Perry MC. 1993. Cigarette smoking and adult leukemia. A meta-analysis. Arch Intern Med 153:469–475.

861

Brugnone F, Perbellini L, Giuliari C, et al. 1994. Blood and urine concentrations of chemical pollutants in the general population. Med Lav 85: 370–389. Buckley JD, Buckley CM, Breslow NE, et al. 1996. Concordance for childhood cancer in twins. Med Pediatr Oncol 26:223–229. Buckley JD, Buckley CM, Ruccione K, et al. 1994. Epidemiological characteristics of childhood acute lymphocytic leukaemia. Analysis by immunophenotype. Leukemia 8:856–864. Buckley JD, Hobbie WL, Ruccione K, et al. 1986. Maternal smoking during pregnancy and the risk of childhood cancer [letter]. Lancet ii:519–520. Buckley JD, Robison LL, Swotinsky R, et al. 1989. Occupational exposures of parents of children with acute nonlymphocytic leukemia: A report from the Children’s Cancer Study Group. Cancer Res 49:4030–4037. Bulbulyan MA, Changuina OV, Zaridze DG, et al. 1998. Cancer mortality among Moscow shoe workers exposed to chloroprene (Russia). Cancer Causes Control 9:381–387. Burkhardt JA. 1982. Leukemia in hospital patients with occupational exposure to the sawmill industry. West J Med 137:440–441. Burmeister LF. 1990. Cancer in Iowa farmers: Recent results. Am J Ind Med 18:295–301. Burnett C, Robison C, Walker J. 1999. Cancer mortality in health and science technicians. Am J Ind Med 36:155–158. Caldwell GG, Kelley D, Heath CW Jr. 1983. Mortality and cancer frequency among military nuclear test (Smoky) participants 1957 through 1979. JAMA 250:620–624. Cardis E, Gilbert ES, Carpenter L, et al. 1995. Effects of low doses and low dose rates of external ionizing radiation: Cancer mortality among nuclear industry workers in three countries. Radiat Res 142:117–132. Carney JP, Maser RS, Olivares H, et al. 1998. The hMrell/hRad50 protein complex and Nijmegen breakage syndrome: Linkage of double-strand break repair to the cellular DNA damage response. Cell 93:477–486. Carstensen JM, Bygren LO, Hatschek T. 1990. Cancer incidence among Swedish brewery workers. Int J Cancer 45:393–396. Cartwright RA, McNally RJQ, Rowland DJ, et al. 1997. The Descriptive Epidemiology of Leukaemia and Related Conditions in Parts of the United Kingdom, 1984–1993. London, Leukaemia Research Fund. Cartwright RA, Miller JG, Scarisbrick DA. 1987a. Leukaemia in a carpet factory: An epidemiological investigation. J Soc Occup Med 37:42–43. Chap L, Nimer SD. 1994. Chronic myelogenous leukemia following repeated radiation therapy for histiocytosis X. Leuk Lymphoma 12:315–316. Checkoway H, Wilcosky T, Wolf P, et al. 1984. An evaluation of the associations of leukemia and rubber industry solvent exposures. Am J Ind Med 5:239–249. Chen H, Sandler DP, Taylor JA, et al. 1996. Increased risk for myelodysplastic syndromes in individuals with glutathione transferase theta 1 (GSTT1) gene defect. Lancet 347:295–297. Chen J, Giovannucci E, Kelsey K, et al. 1996. A methylenetetrahydrofolate reductase polymorphism and the risk of colorectal cancer. Cancer Res 56:4862–4864. Chen R, Seaton A. 1998. A meta-analysis of painting exposure and cancer mortality. Cancer Detect Prev 22:533–539. Child JA, Jack AS, Morgan GJ. 1998. Lymphoproliferative Disorders: Handbook of Diagnosis, Investigation, and Management. London: Chapman & Hall. Choi SW, Kim YI, Weitzel JN, et al. 1998. Folate depletion impairs DNA excision repair in the colon of the rat. Gut 43:93–99. Chow WH, Linet MS, Liff JM, et al. 1996. Cancers in children. In: Schottenfeld D, Praumeni JF Jr., eds. Cancer Epidemiology and Prevention. 2nd ed. New York: Oxford University Press, pp. 1331–1369. Clarkson BD, Boyse EA. 1971. Possible explanation of the high concoddance for acute leukaemia in monozygotic twins. Lancet i:699–701. Cleghorn FR, Manns A, Falk R, et al. 1995. Effect of human T-lymphotropic virus type I infection on non-Hodgkin’s lymphoma incidence. J Natl Cancer Inst 87:1009–1014. Cnattingius S, Zack MM, Ekborn A, et al. 1995. Prenatal and neonatal risk factors for childhood lymphatic leukemia. J Natl Cancer Inst 87:908– 914. Cole P, Delzell E, Acquavella J. 1993. Exposure to butadiene and lymphatic and hematopoietic cancer. Epidemiology 4:96–103. Comstock GW, Livesay VT, Webster RG. 1971. Leukaemia and B.C.G.—a controlled trial. Lancet ii:1062–1063. Comstock GW, Martinez I, Livesay VT. 1975. Efficacy of BCG vaccination in prevention of cancer. J Natl Cancer Inst 54:835–839. Cook-Mozaffari PJ, Darby SC, Doll R, et al. 1989. Cancer near potential sites of nuclear installations. Lancet ii:1145–1147. Corcoran MM, Rasool O, Liu Y, et al. 1998. Detailed molecular delineation of 13q14.3 loss in B cell chronic lymphocytic leukemia. Blood 91: 1382–1390.

862

PART IV: CANCER BY TISSUE OF ORIGIN

Cordier S, Mousel ML, Le Goaster C, et al. 1995. Cancer risk among workers in biomedical research. Scand J Work Environ Health 21:450–459. Corn PG, Kuerbitz SJ, van Noesel MM, et al. 1999. Transcriptional silencing of the p73 gene in acute lymphoblastic leukemia and Burkitt’s lymphoma is associated with 5¢ CpG island methylation. Cancer Res 59:3352– 3356. Correa A, Mohan A, Jackson L, et al. 2000b. Use of hair dyes, hematopoietic neoplasms, and lymphomas: A literature review. I. Leukemias and myelodysplastic syndromes. Cancer Invest 18:366–380. Court Brown WM, Doll R, Bradford Hill A. 1960. Incidence of leukaemia after exposure to diagnostic radiation in utero. Br Med J ii:1539–1544. Crump C, Chen C, Appelbaum FR, et al. 2000. Glutathione S-transferase theta 1 gene deletion and risk of acute myeloid leukemia. Cancer Epidemiol Biomarkers Prev 9:457–460. Crump KS. 1996. Risk of benzene-induced leukemia predicted from the Pliofilm cohort. Environ Health Perspect 104(Suppl 6):1437–1441. Cunningham-Rundles C, Bodian C. 1999. Common variable immunodeficiency: Clinical and immunological features of 248 patients. Clin Immunol 92:34–48. Curtis RE, Boice JD, Stovall M, et al. 1992. Risk of leukemia after chemotherapy and radiation treatment for breast cancer. N Engl J Med 326: 1745–1751. Curtis RE, Boice JD, Stovall M, et al. 1994. Relation of leukemia risk of radiation dose after cancer of the uterine corpus. J Natl Cancer Inst 86:1315–1324. Cuttner J. 1992. Increased incidence of hematologic malignancies in firstdegree relatives of patients with chronic lymphocytic leukemia. Cancer Invest 10:103–109. Daling JR, Starzyk P, Olshan AF, et al. 1984. Birth weight and the incidence of childhood cancer. J Natl Cancer Inst 72:1039–1041. Darby SC, Doll R. 1987. Fallout, radiation doses near Dounreay, and childhood leukaemia. Br Med J 294:603–607. Darby SC, Kendall GM, Fell TP, et al. 1988. A summary of mortality and incidence of cancer in men from the United Kingdom who participated in the United Kingdom’s atmospheric nuclear weapons tests and experimental programmes. Br Med J 296:332–338. Date Y, Seki N, Kamizono S, et al. 1999. Identification of a genetic risk factor for systemic juvenile rheumatoid arthritis in the flanking region of the TNF alphagene and HLA genes. Arthritis and Rheumatism 42(12):2577–2582. Davies FE, Rollinson SJ, Rawstron AC, et al. 2000. High-producer haplotypes of tumor necrosis factor alpha and lymphotoxin alpha are associated with an increased risk of myeloma and have an improved progression-free survival after treatment. J Clin Oncol 18:2843–2851. Davies SM, Robison LL, Buckley JD, et al. 2000. Glutathione S-transferase polymorphisms in children with myeloid leukemia: A Children’s Cancer Group study. Cancer Epidemiol Biomarkers Prev 9:563–566. Dean G. 1994. Deaths from primary brain cancers, lymphatic and haematopoietic cancers in agricultural workers in the Republic of Ireland. J Epidemiol Comm Health 48:364–368. Dean M, Jacobson LP, McFarlane G, et al. 1999. Reduced risk of AIDS lymphoma in individuals heterozygous for the CCR5-delta32 mutation. Cancer Res 59:3561–3564. de Franchis R, Sebastio G, Mandato C, et al. 1995. Spina bifida, 677ÆC mutation, and role of folate. Lancet 346:1703. Deininger MW, Bose S, Gora-Tybor J, et al. 1998. Selective induction of leukemia-associated fusion genes by high-dose ionizing radiation. Cancer Res 58:421–425. Delongchamp RR, Mabuchi K, Yoshimoto Y, et al. 1997. Cancer mortality among atomic bomb survivors exposed in utero or as young children, October 1950–Nov 1992. Radiat Res 147:385–395. DeLord C, Powles R, Mehta J, et al. 1998. Familial acute myeloid leukaemia: Four male members of a single family over three consecutive generations exhibiting anticipation. Br J Haematol 100:557–560. Delore P, Borgomano C. 1928. Leukemia aigue un cour de l’intoxication benzenique, sur l’origine toxique de certains leukemias aigues et leurs relations avec les anemies graves. J Med Lyon 9:227–233. Delzell E, Monson RR. 1981. Mortality among rubber workers. II. Causespecific mortality, 1940–1978. J Occup Med 23:677–684. Demeter J, Porzsolt F, Ramisch S, et al. 1997. Polymorphism of the tumor necrosis factor-alpha and lymphotoxin-alpha genes in chronic lymphocytic leukaemia. Br J Haematol 97(1):107–112. Devesa SS, Grauman DJ, Blot WJ, et al. 1999. Atlas of cancer mortality in the United States. 1950–94. Bethesda, Maryland NIH, NCI, NIH Publ. No. 99-4564. Diamond EL, Schmerler H, Lilienfeld AM. 1973. The relationship of intrauterine radiation to subsequent mortality and development of leukemia in children. A prospective study. Am J Epidemiol 7:283–313.

Divine BJ, Hartman CM, Wendt JK. 1999. Update of the Texaco mortality study 1947–93: Part II. Analyses of specific causes of death for white men employed in refining, research, and petrochemicals. Occup Environ Med 56:174–180. Divine BJ, Wendt JK, Hartman CM. 1993. Cancer mortality among workers at a butadience production facility. IARC Scientific Publication No. 127. Lyon, France: International Agency for Research on Cancer 345–362. Dockerty JD, Draper G, Vincent T, et al. 2001. Case-control study of parental age, parity, and socio-economic level in relation to childhood cancers. Int J Epidemiol 30:1428–1437. Dockerty JD, Skegg DC, Elwood JM, et al. 1999. Infections, vaccinations, and the risk of childhood leukaemia. Br J Cancer 80:1483–1489. Doll R, Wakeford R. 1997. Risk of childhood cancer from fetal irradiation. Br J Radiol 70:130–139. Doll R. 1996. Cancers weakly related to smoking. Br Med Bull 52:35–49. Doody MM, Linet MS, Glass AG, et al. 1992. Leukemia, lymphoma, and multiple myeloma following selected medical conditions. Cancer Causes Control 3:449–456. dos Santos Silva I, Jones M, Malveiro F, et al. 1999. Mortality in the Portuguese thorotrast study. Radiat Res 152(6 Suppl):S88–92. Dosemeci M, Yin SN, Linet M, et al. 1996. Indirect validation of benzene exposure assessment by association with benzene poisoning. Environ Health Perspect 104(Suppl 6):1343–1347. Douer D, Preston-Martin S, Chang E, et al. 1996. High frequency of acute promyelocytic leukemia among Latinos with acute myeloid leukemia. Blood 87:308–313. Downs TD, Crane MM, Kim KW. 1987. Mortality among workers at a butadiene facility. Am J Ind Med 12:311–329. Draper GJ, Kroll ME, Stiller CA. 1994. Childhood cancer. Cancer Surv 19–20:493–517. Draper GJ, Little MP, Sorahan T, et al. 1997. Cancer in the offspring of radiation workers: A record linkage study. BMJ 315:1181–1188. Draper GJ, Stiller CA, Cartwright RA, et al. 1993. Cancer in Cumbria and in the vicinity of the Sellafield nuclear installation 1963–90. Br Med J 306:89–94. Dror Y, Freedman MH. 2001. Shwachman-Diamond syndrome marrow cells show abnormally increased apoptosis mediated through the Fas pathway. Blood 97:3011–3016. Eastmond DA, Schuler M, Frantz C, et al. 2001. Characterization and mechanisms of chromosomal alterations induced by benzene in mice and humans. Res Rep Health Eff Inst 103:69–80. Edlich RF, Arnette JA, Williams FM. 2000. Global epidemic of human T-cell lymphotropic virus type-I (HTLV-I). J Emerg Med 18:109–119. Egdahl RD, Coppock E, Hornik R. 1991. Mortality experience at a hydrometallurgical nickel refinery in Fort Saskatchewan, Alberta between 1954 and 1984. J Soc Occup Med 4:29–31. Eisenberg DE, Sorahan T. 1987. Birth weight and childhood cancer deaths. J Natl Cancer Inst 78:1095–1100. Engbersen AM, Franken DG, Boers GH, et al. 1995. Thermolabile 5,10methylenetetrahydrofolate reductase as a cause of mild hyperhomocysteinemia. Am J Hum Genet 56:142–150. Engeland A, Andersen A, Haldorsen T, et al. 1996. Smoking habits and risk of cancers other than lung cancer: 28 years’ follow-up of 25,000 Norwegian men and women. Cancer Causes Control 7:497–506. Ernst P, Wang J, Korsmeyer SJ. 2002. The role of MLL in hematopiesis and leukaemia. Curr Opin Hematol 9:282–287. Evans JS, Wennberg JE, McNell BJ. 1986. The influence of diagnostic radiography on the incidence of breast cancer and leukemia. N Engl J Med 315:810–815. Fabia J, Thuy TD. 1974. Occupation of father at time of birth of children dying of malignant diseases. Br J Prev Soc Med 28:98–100. Faderl S, Kantarjian HM, Thomas DA, et al. 2000. Outcome of Philadelphia chromosome-positive adult acute lymphoblastic leukemia. Leuk Lymphoma 36:263–273. Fasal E, Jackson EW, Klauber MR. 1971. Birth characteristics and leukemia in childhood. J Natl Cancer Inst 47:501–509. Feingold L, Savitz DA, John EM. 1992. Use of a job-exposure matrix to evaluate parental occupation and childhood cancer. Cancer Causes Control 3:161–169. Felix CA. 1998. Secondary leukemias induced by topoisomerase-targeted drugs. Biochim Biophys Acta 1400:233–255. Felix CA, D’Amico D, Mitsudomi T, et al. 1992. Absence of hereditary p53 mutations in 10 familial leukemia pedigrees. J Clin Invest 90:653–658. Fenech MF, Dreosti IE, Rinaldi JR. 1997. Folate, vitamin B12, homocysteine status and chromosome damage rate in lymphocytes of older men. Carcinogenesis 18:1329–1336. Fernandez E, Chatenoud L, La Vecchia C, et al. 1999. Fish consumption and cancer risk. Am J Clin Nutr 70:85–90.

The Leukemias Ferreira OLC Jr, Planelles V, Rosenblatt JD. 1997. Human T-cell leukemia viruses: Epidemiology, biology, and pathogenesis. Blood Rev 11:91–104. Feugeas JP, Caillens H, Poirier JC, et al. 1997. Influence of metabolic and genetic factor on tumour necrosis factor-alpha and lymphotoxin-alpha production in insulin-dependent diabetes mellitus. Diabetes Metabolism 23(4):295–301. Feychting M, Ahlbom A. 1993. Magnetic fields and cancer in children residing near Swedish high-voltage power lines. Am J Epidemiol 138:467–481. Feychting M, Ahlbom A. 1994. Magnetic fields, leukemia, and central nervous system tumors in Swedish adults residing near high-voltage power lines. Epidemiology 5:501–509. Feychting M, Forssen U, Floderus B. 1997. Occupational and residential magnetic field exposure and leukemia and central nervous system tumors. Epidemiology 8:384–389. Feychting M, Plato N, Nise G, et al. 2001. Paternal occupational exposures and childhood cancer. Environ Health Perspect 109:193–196. Filipovich AH, Heinitz KJ, Robison LL, et al. 1987. The immunodeficiency cancer registry: A research resource. Am J Pediatr Hematol Oncol 9: 83–84. Filipovich AH, Mathur A, Kamat D, et al. 1994. Lymphoproliferative disorders and other tumors complicating immunodeficiencies. Immunodeficiency 5:91–112. Filipovich AH, Mathur A, Kamat D, et al. 1992. Shapiro RS. Primary immunodeficiencies: Genetic risk factors for lymphoma. Cancer Res 52(19 Suppl):5465s–5467s. Finch SC, Linet MS. 1992. Chronic leukemias. In: Fleming AF, ed. Epidemiology of Haematological Disease. Part I. London: Bailliere Tindal, pp. 27–56. Floderus B, Persson T, Stenlund C, et al. 1993. Occupational exposure to electromagnetic fields in relation to leukemia and brain tumors: A casecontrol study in Sweden. Cancer Causes Control 4:465–476. Flodin U, Fredriksson M, Persson, et al. 1988. Chronic lymphatic leukemia and engine exhausts, fresh wood, and DDT: A case-referent study. Br J Ind Med 45:33–38. Folias A, Matkovic M, Brunn D, et al. 2002. BRCA1 interacts directly with the Fanconi anemia protein FANCA. Hum Mol Genet 11:2591–2597. Forastiere F, Sperati A, Cherubini G, et al. 1998. Adult myeloid leukaemia, geology, and domestic exposure to radon and gamma radiation: A case control study in central Italy. Occup Environ Med 55:106–110. Ford AM, Bennett CA, Price CM, et al. 1998. Fetal origins of the TEL-AML1 fusion gene in identical twins with leukaemia. Proc Natl Acad Sci USA 14:4584–4588. Ford AM, Ridge SA, Cabrera ME, et al. 1993. In utero rearrangements in the trithorax-related oncogene in infant leukemias. Nature 363:358–360. Forman D, Cook-Mozaffari P, Darby S, et al. 1987. Cancer near nuclear installations. Nature 329:499–505. Fox AJ, Collier PF. 1976. A survey of occupational cancer in the rubber and cable-making industries: Analysis of deaths occurring in 1972–74. Br J Ind Med 33:249–264. Fraument JF Jr. 1969. Clinical epidemiology of leukemia. Semin Hematol 6:250–260. Freedman DM, Stewart P, Kleinerman RA, et al. 2001. Household solvent exposures and childhood acute lymphoblastic leukaemia. Am J Public Health 91:564–567. Freedman MH, Alter BP. 2002. Risk of myelodysplastic syndrome and acute myeloid leukemia in congenital neutropenias. Semin Hematol 39: 128–133. Friedman GD. 1982. Phenlybutazone, musculoskeletal disease, and leukemia. J Chronic Dis 35:233–243. Friedman GD, Ury HK. 1980. Initial screening for carcinogenicity of commonly used drugs. J Natl Cancer Inst 65:723–733. Friedman GD. 1993. Cigarette smoking, leukemia, and multiple myeloma. Ann Epidemiol 3:425–428. Fritz A, Percy C, Jack A, et al. 2000. In: Whelen et al., ed. International Classification of Diseases for Oncology, 3rd ed. U.S. Interim Version 2000. World Health Organization, Geneva. Frosst P, Blom HJ, Milos R, et al. 1995. A candidate genetic risk factor for vascular disease: A common mutation in methylenetetrahydrofolate reductase. Nat Genet 10:111–113. Gahn B, Schafer C, Neef J, et al. 1997. Detection of trisomy 12 and Rbdeletion in CD34+ cells of patients with B-cell chronic lymphocytic leukemia. Blood 89:4275–4281. Gale KB, Ford AM, Repp R, et al. 1997. Backtracking leukemia to birth: Identification of clonotypic genefusion sequences in neonatal blood spots. Proc Natl Acad Sci USA 94:13950–13954. Gardner MJ, Snee MP, Hall AJ, et al. 1990. Results of case-control study of leukaemia and lymphoma among young people near Sellafield nuclear plant in West Cumbria. BMJ 300:423–429.

863

Garfinkel L, Boffetta P. 1990. Association between smoking and leukemia in two American Cancer Society prospective studies. Cancer 65:2356– 2360. Gaymes TJ, Mufti GJ, Rassool FV. 2002. Myeloid leukemias have increased activity of the nonhomologous end-joining pathway and concomitant DNA misrepair that is dependent on the Ku70/86 heterodimer. Cancer Res 62:2791–2797. German J. 1997. Bloom’s syndrome. XX. The first 100 cancers. Cancer Genet Cytogenet 93:100–106. Gibson R, Graham S, Lilienfield A, et al. 1972. Irradiation in the epidemiology of leukemia among adults. J Natl Cancer Inst 48:301–311. Giles FJ, Koeffler HP. 1994. Secondary myelodysplastic syndromes and leukemias. Curr Opin Hematol 1:256–260. Gilman PA, Ames RG, McCawley MA. 1985. Leukemia risk among U.S. white male coal miners. J Occup Med 27:669–671. Girard PM, Foray N, Stumm M, et al. 2000. Radiosensitivity in Nimegen Breakage Syndrome cells is attributable to a repair defect and not cell cycle checkpoint defects. Cancer Res 60:4881–4888. Glynn SA, Albanes D, Pietinen P, et al. 1996. Colorectal cancer and folate status: A nested case-control study among male smokers. Cancer Epidemiol Biomarkers Prev 5:487–494. Gold EB, Diener MD, Szklo M. 1982. Parental occupations and cancer in children—a case-control study and review of the methodologic issues. J Occup Med 24:578–584. Goldin LR, Sgambati M, Marti GE, et al. 1999. Anticipation in familial CLL. Am J Hum Genet 65:265–269. Goode EL, Ulrich CM, Potter JD. 2002. Polymorphisms in DNA repair genes and associations with cancer risk. Cancer Epidemiol Biomarkers Prev 11:1513–1530. Gordon MY, Dazzi F, Marley SB, et al. 1999. Cell biology of CML cells. Leukemia 13:S65–S71. Gotuzzo E, Arango C, de Queiroz-Campos A. 2000. Human T-cell lymphotropic virus-I in Latin America. Infect Dis Clin North Am 14:211–239. Graham S, Levin ML, Lilienfeld AM, et al. 1966. Preconception, intrauterine, and postnatal irradiation as related to leukemia. Natl Cancer Inst Monogr 19:347–371. Greaves M. 1999. Molecular genetics, natural history and the demise of childhood leukaemia. Eur J Cancer 34:1941–1953. Greaves MF. 1997. Aetiology of acute leukaemia. Lancet 349:344–349. Greaves MF. 1993. Stem cell origins of leukaemia and curability. Br J Cancer 67:413–423. Green DM, Fiorello A, Zevon MA, et al. 1997. Birth defects and childhood cancer in offspring of survivors of childhood cancer. Arch Pediatr Adolesc Med 151:379–383. Greenberg HL, Ott MG, Shore RE. 1990. Men assigned to ethylene oxide or other ethylene oxide related chemical manufacturing. A mortality study. Br J Ind Med 47:221–230. Greenberg MS, Anderson KC, Marchetto DJ, et al. 1981. Acute myelocytic leukemia in two brothers with polyposis coli and carcinoma of the colon. Ann Intern Med 95:702–703. Greenland S, Sheppard AR, Kaune WT, et al. 2000. A pooled analysis of magnetic fields, wire codes, and childhood leukemia. Childhood LeukemiaEMF Study Group. Epidemiology 11:624–634. Griem ML, Kleinerman RA, Boice JD Jr, et al. 1994. Cancer following radiotherapy for peptic ulcer. J Natl Cancer Inst 86:842–849. Grimwade DJ, Stephenson J, De Silva C, et al. 1993. Familial MDS with 5qabnormality. Br J Haematol 84:536–538. Grodstein F, Hennekens CH, Colditz GA, et al. 1994. A prospective study of permanent hair dye use and hematopoietic cancer. J Natl Cancer Inst 86:1466–1470. Groet J, Mcelwaine S, Spinelli M, et al. 2003. Acquired mutations in GATA1 in neonates with Down syndrome with transient myeloid disorder. Lancet 362:1617–1620. Groupe Francais de Cytogenetioue Hematologioue. 1997. Forty-four cases of childhood myelodysplasia with cytogenetics, documented by the Groupe Francais de Cytogenetique Hematologique. Leukemia 11:1478– 1485. Groves FD, Gridley G, Wacholder S. 1999. Infant vaccinations and risk of childhood acute lymphoblastic leukaemia in the USA. Br J Cancer 81:175–178. Groves FD, Linet MS, Travis LB, et al. 2000. Cancer surveillance series: NonHodgkin’s lymphoma incidence by histologic subtype in the United States, 1978–1995. J Natl Cancer Inst 92:1240–1251. Groves FD, Linet MS, Devesa SS. 1995. Patterns of occurrence of the leukaemias. Eur J Cancer 31A:941–949. Gu L, Cline-Brown B, Zhang F, et al. 2002. Mismatch repair deficiency in haematological malignancies with microsatellite instability. Oncogene 21:5758–5764.

864

PART IV: CANCER BY TISSUE OF ORIGIN

Gundestrup M, Storm HH. 1999. Radiation-induced acute myeloid leukaemia and other cancers in commercial jet cockpit crew: A population-based cohort study. Lancet 354:2029–2031. Gunz FW, Atkinson H. 1964. Medical radiation and leukemia: A retrospective survey. Br Med J 5380:389–396. Gunz FW, Gunz JP, Vincent PC, et al. 1978. Thirteen cases of leukemia in a family. J Natl Cancer Inst 60:1243–1250. Gurley KE, Vo K, Kemp CJ. 1998. DNA double-strand breaks, p53, and apoptosis during lymphomagenesis in scid/scid mice. Cancer Res 58: 3111–3115. Gustavsson P, Reuterwall C, Sadigh J, et al. 1999. Mortality and cancer incidence among laboratory technicians in medical research and routine laboratories (Sweden). Cancer Causes Control 10:59–64. Haase D, Binder C, Bunger J, et al. 2002. Increased risk of therapy-associated hematologic malignancies in patients with carcinoma of the breast and combined homozygoud deletions of glutathione transferases M1 and T1. Leuk Res 26:249–254. Hamblin TJ, Orchard JA, Gardiner A, et al. 2000. Immunoglobulin V genes and CD38 expression in CLL. Blood 95:2455–2457. Hansen WK, Kelley MR. 2000. Review of mammalian DNA repair and translational implications. J Pharmacol Exp Ther 295:1–9. Harada H, Harada Y, Tanaka H, et al. 2003. Implications of somatic mutations in the AML1 gene in radiation-associated and therapy-related myelodysplastic syndrome/acute myeloid leukemia. Blood 101:673–380. Harnden DG. 1985. Inherited factors in leukaemia and lymphoma. Leuk Res 9:705–707. Harris NL. 1999. Hodgkin’s lymphomas: Classification, diagnosis, and grading. Semin Hematol 36:220–232. Hasle H, Clemmensen IH, Mikkelsen M. 2000. Risks of leukaemia and solid tumours in individuals with Down’s syndrome. Lancet 355:165–169. Hasle H, Olsen JH. 1997. Cancer in relatives of children with myelodysplastic syndrome, acute and chronic myeloid leukaemia. Br J Haematol 97:127–131. Hatch EE, Linet MS, Kleinerman RA, et al. 1998. Association between childhood acute lymphoblastic leukemia and use of electrical appliances during pregnancy and childhood. Epidemiol 9:234–245. Hayashi Y. 2000. The molecular genetics of recurring chromosome abnormality in acute myeloid leukemia. Semin Hematol 37(4):368–380. Hayes RB, Blair A, Stewart PA, et al. 1990. Mortality of U.S. embalmers and funeral directors. Am J Ind Med 18:641–652. Hayes RB, Yin SN, Dosemeci M, et al. 1997. Benzene and the dose-related incidence of hematologic neoplasms in China. Chinese Academy of Preventive Medicine-National Cancer Institute Benzene Study Group. J Natl Cancer Inst 89:1065–1071. Heacock H, Hertzman C, Demers PA, et al. 2000. Childhood cancer in the offspring of male sawmill workers occupationally exposed to chlorophenate fungicides. Environ Health Perspect 108:499–503. Head DR. 1996. Revised classification of acute myeloid leukemia. Leukemia 10:1826–1831. Hecht F, Hecht BK. 1990. Cancer in ataxia telangectasia patients. Cancer Genet Cytogenet 46:9–19. Hempelmann L, Hall W, Phillips M, et al. 1975. Neoplasms in persons treated with x-ray in infancy: Fourth survey in 20 years. J Natl Cancer Inst 55:519–530. Henshaw DL, Eatough JP, Richardson RB. 1990. Radon as a causative factor in induction of myeloid leukaemia and other cancers. Lancet 335:1008–1012. Higham MA, Pride NB, Alikhan A, et al. 2000. Tumour necrosis factor-alpha gene promoter polymorphism in chronic obstructive pulmonary disease. Eur Respir J 15(2):281–284. Hildreth NG, Shore RE, Hempelmann LH, et al. 1985. Risk of extrathyroid tumors following radiation treatment in infancy for thymic enlargement. Radiat Res 102:378–391. Hill DA, Gridley G, Cnattingius S, et al. 2003. Mortality and cancer incidence among individuals with Down syndrome. Arch Intern Med 163:705–711. Hinds MW, Kolonel LN, Lee T, et al. 1980. Associations between cancer incidence and alcohol and cigarette consumption among five ethnic groups in Hawaii. Br J Cancer 41:929–940. Hirschhorn JN, Lohmueller K, Byrne E, et al. 2002. A comprehensive review of genetic association studies. Genet Med 4:45–61. Hisada M, Garber JE, Fung CY, et al. 1998. Multiple primary cancers in families with Li-Fraumeni syndrome. J Natl Cancer Inst 90:606–611. Hjelmstrom P, Peacock CS, Giscombe R, et al. 1998. Polymorphism in tumor necrosis factor genes associated with myasthenia gravis. J Neuroimmunol 88(1–2):137–143. Hodgson JT, Jones RD. 1985. Mortality of styrene production, polymerization and processing workers at a site in northwest England. Scand J Work Environ Health 11:347–352.

Hodgson JT, Jones RD. 1990. Mortality of a cohort of tin miners 1941–86. Br J Ind Med 47:655–676. Hoffman R, Benz EJ Jr, Shattil SJ, et al. 2000. Hematology. Basic Principles and Practice. 3rd ed. New York: Church Livingstone, pp. 120–143, 979–99, 1006–16, 1018–24, 1052–69, 1071–78, 1089–932, 1103–14, 1125–37, 1140–45, 1151–63, 1167–71, 1188–91, 1194–96, 1260–93, 1339–42, 1347–49, 1350–53, 1360–63, 1388–1407, 1410–16. Hogstedt C, Malinqvist N, Wadman B. 1979a. Leukemia in workers exposed to ethylene oxide. JAMA 241:1132–1133. Hogstedt C, Arringer L, Gustavsson A. 1986. Epidemiologic support for ethylene oxide as a cancer-causing agent. JAMA 255:1575–1578. Hogstedt C, Rohlen O, Berndtsson BS, et al. 1979b. A cohort study of mortality and cancer incidence in ethylene oxide production workers. Br J Ind Med 36:276–280. Honda Y, Delzell E, Cole P. 1995. An updated study of mortality among workers at a petroleum manufacturing plant. J Occup Environ Med 37:194–200. Horwitz M. 1997. The genetics of familial leukemia. Leukemia 11:1347–1359. Horwitz M, Benson KF, Li FQ, et al. 1997. Genetic heterogeneity in familial acute myelogenous leukemia: evidence for a second locus at chromosome 16q21–23.2. Am J Hum Genet 61:873–881. Horwitz M, Goode EL, Jarvik GP. 1996b. Anticipation in familial leukemia. Am J Hum Genet 59:990–998. Hotz P, Lauwerys RR. 1997. Hematopoietic and lymphatic malignancies in vehicle mechanics. Crit Rev Toxicol 27:443–494. Hromas R, Shopnick R, Jumean HG, et al. 2000. A novel syndrome of radiation-associated acute myeloid leukemia involving AML1 gene translocations. Blood 95:4011–4013. Huang TJ, MacAry PA, Eynott P, et al. 2001. Allergen-specific Th1 cells counteract efferent Th2 cell-dependent bronchial hyperresponsiveness and eosinophilic inflammation partly via IFN-gamma. J Immunol 166: 207–217. Huebner WW, Schnatter AR, Nicolich MJ, et al. 1997. Mortality experience of a young petrochemical industry cohort. 1979–1992 follow-up study of US-based employees. J Occup Environ Med 39:970–982. Hunter WF, Henman BA, Bartlett DM, et al. 1993. Mortality of professional chemists in England and Wales, 1965–1989. Am J Ind Med 23:615– 627. Hunting KL, Longbottom H, Kalavar SS, et al. 1995. Haematopoietic cancer mortality among vehicle mechanics. Occup Environ Med 52:673–678. Hursting SD, Margolin BH, Switzer BR. 1993. Diet and human leukemia: an analysis of international data. Prev Med 22:409–422. IARC Working Group on the Evaluation of Carcinogenic Risks to Humans. 1995. Monographs on the Evaluation of the Carcinogenic Risks to Humans. Wood dust and formaldehyde. Volume 62. Lyon, France: IARC. Ido M, Nagata C, Kawakami N, et al. 1996. A case-control study of myelodysplastic syndromes among Japanese men and women. Leuk Res 20: 727–731. Imamura N, Abe K, Oguma, N. 2002. High incidence of point mutations of p53 suppressor oncogene in patients with myelodysplastic syndrome among atomic-bomb survivors: A 10-year follow up. Leukemia 16:154–156. Imamura N, Kimura A. 2000. Neutropenia among survivors of atomic bomb explosion. Lancet 355:117. Inbar O, Kupiec M. 2000. Recombination between divergent sequences leads to cell death in a mismatch-repair-independent manner. Curr Genet 38:23–32. Infante-Rivard C, Fortier I, Olson E. 2000. Markers of infection, breast-feeding and childhood acute lymphoblastic leukemia. Br J Cancer 83:1559–1564. Infante-Rivard C, Krajinovic M, Labuda D, et al. 2002. Childhood acute lymphoblastic leukemia associated with parental alchohol consumption and polymorphisms of carcinogen-metabolizing genes. Epidemiology 13:277–281. Infante-Rivard C, Olson E, Jacques L, et al. 2001. Drinking water contaminants and childhood leukaemia. Epidemiology 12:13–19. Infante-Rivard C, Labuda D, Krajinovic M, et al. 1999. Risk of childhood leukemia associated with exposure to pesticides and with gene polymorphisms. Epidemiology 10:481–487. Infante-Rivard C, Mur P, Armstrong B, et al. 1991. Acute lymphoblastic leukaemia among Spanish children and mothers’ occupation: A casecontrol study. J Epidemiol Community Health 45:11–15. Infante-Rivard C, Sinnett D. 1999. Preconceptional paternal exposure to pesticides and increased risk of childhood leukaemia. Lancet 354: 1819. Inskip PD. 1999. Second cancers following radiotherapy. In: Neugut AI, Meadows AT, eds. Multiple Primary Cancers. Philadelphia: Lippincott Williams & Wilkins, pp. 91–135.

The Leukemias Inskip PD, Kleinerman RA, Stovall M, et al. 1993. Leukemia, lymphoma, and multiple myeloma after pelvic radiotherapy for benign diseases. Radiat Res 135:108–124. Inskip PD, Monson RR, Wagoner JK, et al. 1990. Leukemia following radiotherapy for uterine bleeding. Radiat Res 122:107–119. Ireland B, Collins JJ, Buckley CF, et al. 1997. Cancer mortality among workers with benzene exposure. Epidemiology 8:318–320. Irons RD. 2000. Molecular models of benezene leukemogenesis. J Toxicol Environ Health A 61:391–397. Iurlo A, Foa P, Maiolo AT, et al. 1989. Chronic myelogenous leukemia following radiation therapy for testicular seminomas. Blut 59:503– 504. Jablon S, Hrubec Z, Stone BJ, et al. 1990. Cancer in Populations Living Near Nuclear Facilities. NTH Publ 90–874. Washington, DC U.S. Government Printing Office. Jacob RA, Gretz DM, Taylor PC, et al. 1998. Moderate folate depletion increases plasma homocysteine and decreases lymphocyte DNA methylation in postmenopausal women. J Nutr 128:1204–1212. Jaffe ES, Harris NL, Stein H, et al., eds. 2001. Pathology and Genetics of Tumours of Haematopoietic and Lymphoid Tissues. Lyon, France: IARC Press. Ji BT, Shu XO, Linet MS, et al. 1997. Paternal cigarette smoking and the risk of childhood cancer among offspring of nonsmoking mothers. J Natl Cancer Inst 89:238–244. Johansen C, Olsen JH. 1998. Risk of cancer among Danish utility workers—a nationwide cohort study. Am J Epidemiol 147:548–555. Joseph P, Xie T, Xu Y, et al. 1994. NAD(P)H:quinone oxidoreductase 1 (DTdiaphorase): Expression, regulation, and role in cancer. Oncol Res 6:525. Kaldor JM, Day NE, Clarke EA, et al. 1990a. Leukemia following Hodgkin’s disease. N Engl J Med 322:7–13. Kaldor JM, Day NE, Pettersson F, et al. 1990b. Leukemia following chemotherapy for ovarian cancer. N Engl J Med 322:1–6. Kaletsch U, Kaatsch P, Meinert R, et al. 1999. Childhood cancer and residential radon exposure—results of a population-based case-control study in Lower Saxony (Germany). Radiat Environ Biophys 38:211–225. Kane EV, Roman E, Cartwright R, et al. 1999. Tobacco and the risk of acute leukaemia in adults. Br J Cancer 81:1228–1233. Kardos G, Veerman AJ, de Waal FC, et al. 1996. Familial sideroblastic anemia with emergence of monosomy 5 and myelodysplastic syndrome. Med Pediatr Oncol 26:54–56. Kaune WT, Miller MC, Linet MS, et al. 2000. Children’s exposure to magnetic fields produced by U.S. television sets used for viewing programs and playing video games. Bioelectromagnetics 21:214–227. Kaune WT, Miller MC, Linet MS, et al. 2002. Magnetic fields produced by hand held hair dryers, stereo headsets, home sewing machines, and electric clocks. Bioelectromagnetics 23:14–25. Kaye SA, Robison LL, Smithson WA, et al. 1991. Maternal reproductive history and birth characteristics in childhood acute lymphoblastic leukemia. Cancer 68:1351–1355. Kelleher C, Newell J, MacDonagh-White C, et al. 1998. Incidence and occupational patterns of leukaemias, lymphomas, and testicular tumors in western Ireland over an 11 year period. J Epidemiol Comm Health 52:651–656. Keller-Byrne JE, Khuder SA, Schaub EA. 1995. Meta-analysis of leukemia and farming. Environ Res 71:1–10. Kelsoe G. 1998. V(D)J hypermutation and DNA mismatch repair: Vexed by fixation. Proc Natl Acad Sci USA 95:6576–6577. Kende G, Toren A, Mandel M, et al. 1994. Familial leukemia: Description of two kindreds and a review of the genetic aspects of the disease. Acta Haematol 92:208–211. Keshava-Prasad HS, Prangnell DR, Adelman M. 1996. Multiple myeloma in spouses. Clin Lab Haematol 18:61–62. Khanna KK. 2000. Cancer risk and the ATM gene: A continuing debate. J Natl Cancer Inst 92:795–802. Kheifets LI, Afifi AA, Buffler P, et al. 1997. Occupational electric and magnetic field exposure and leukemia. J Occup Environ Health 39: 1074–1091. Kim YI, Pogribny IP, Basnakian AG, et al. 1997. Folate deficiency in rats induces DNA strand breaks and hypomethylation within the p53 tumor suppressor gene. Am J Clin Nutr 65:46–52. Kinlen LJ. 1994. Leukemia. In: Doll R, Fraumeni JF, Muir CS, eds. Trends in Cancer Incidence and Mortality. Cancer Surv 19/20:475–491. Kinlen LJ, Clarke K, Balkwill A. 1993a. Paternal preconceptional radiation exposure in the nuclear industry and leukaemia and non-Hodgkin’s lymphoma in young people in Scotland. BMJ 306:1153–1158. Kinlen LJ, Clarke K, Hudson C. 1990. Evidence from population mixing in British New Towns 1946–85 of an infective basis for childhood leukaemia. Lancet 336:577–582.

865

Kinlen LJ, Hudson C. 1991. Childhood leukaemia and poliomyelitis in relation to military encampments in England and Wales in the period of national military service, 1950–63. BMJ 303:1357–1362. Kinlen LJ, John SM. 1994. Wartime evacuation and mortality from childhood leukaemia in England and Wales in 1945–9. BMJ 309:1197–1202. Kinlen LJ, O’Brien F, Clarke K, et al. 1993b. Rural population mixing and childhood leukaemia: Effects of the North Sea oil industry in Scotland, including the area near Dounreay nuclear site. BMJ 306: 743–748. Kinlen LJ, Petridou E. 1995. Childhood leukemia and rural population movements: Greece, Italy, and other countries. Cancer Causes Control 6: 445–450. Kinlen LJ, Rogot E. 1988. Leukaemia and smoking habits among United States veterans. Br Med J 297:657–659. Kinlen LJ, Webster AD, Bird AG, et al. 1985. Prospective study of cancer in patients with hypogammaglobulinaemia. Lancet i:263–266. Kleihues P, Schauble B, zur Hausen A, et al. 1997. Tumors associated with p53 germline mutations: A synopsis of 91 families. Am J Pathol 150:1–13. Knowles DM, ed. 1992. Neoplastic Hematopathology. Baltimore, MD: Williams & Wilkins. Kogevinas M, Ferro G, Saracci R, et al. 1993. Cancer mortality in an international cohort of workers exposed to styrene. IARC Scientific Publications No. 127. Lyon, France: International Agency for Research on Cancer, pp. 289–300. Kogevinas M, Sala M, Boffetta P, et al. 1998. Cancer risk in the rubber industry: A review of the recent epidemiological evidence. Occup Environ Med 55:1–12. Kohler T, Plettig R, Wetzstein W, et al. 2000. Cytokine-driven differentiation of blasts for patients with acute meylogenous and lymphoblastic leukemia into dendritic cells. Stem Cells 18:139–147. Kok PW, Ong CN. 1994. Blood and urinary benzene determined by headspace gas chromatography with photoionization detection: Application in biological monitoring of low-level nonoccupational exposure. Int Arch Occcup Environ Health 66:195–201. Kolstad HA, Lynge E, Olsen J. 1993. Cancer incidence in the Danish reinforced plastics industry. IARC Scientific Publications No. 127. Lyon, France: International Agency for Research on Cancer, pp. 301–308. Krishnan A, Bhatia S, Slovak ML. et al. 2000. Predictors of therapy-related leukemia and myelodysplasia following autologous transplantation for lymphoma: An assessment of risk factors. Blood 95:1588–1593. Kristensen P, Andersen A, Irgens LM, et al. 1996a. Cancer in offspring of parents engaged in agricultural activities in Norway: Incidence and risk factors in the farm environment. Int J Cancer 65:39–50. Kristensen P, Andersen A, Irgens LM, et al. 1996b. Incidence and risk factors of cancer among men and women in Norwegian agriculture. Scand J Work Environ Health 22:14–26. Kuttesch JF Jr, Wexler LH, Marcus RB, et al. 1996. Second malignancies after Ewing’s sarcoma: Radiation dose dependency of secondary sarcomas. J Clin Oncol 14:2818–2825. Kwasniewska A, Tukendorf A, Semczuk M. 1997. Folate deficiency and cervical intraepithelial neoplasia. Eur J Cynaecol Oncol 18:526–530. Kwiatkowski A. 1993. Dietary and other environmental risk factors in acute leukemias. A case-control study of 119 patients. Eur J Cancer Prev 2:139–146. Kwong YL, Ng MH, Ma SK. 2000. Familial acute myeloid leukemia with monosomy 7: Late onset and involvement of a multipotential progenitor cell. Cancer Genet Cytogenet 116:170–173. Larson RA, Wang Y, Banerjee M, et al. 1999. Prevalence of the inactivating 609C (T polymorphism in the NAD(P)H:quinone oxidoreductase (NQO1) gene in patients with primary and therapy-related myeloid leukemia. Blood 94:803–807. Law GR, Kane EV, Roman E, et al. 2000. Residental radon exposure and adult acute leukemia. Lancet 355:1888. Leiss JK, Savitz DA. 1995. Home pesticide use and childhood cancer: a casecontrol study. Am J Public Health 85:249–252. Lemos MC, Cabrita FJ, Silva HA, et al. 1999. Genetic polymorphism of CYP2D6, GSTM1 and NAT2 and susceptibility to haematological neoplasias. Carcinogenesis 20, 1225–1229. Leone G, Mele L, Pulsoni A, et al. 1999. The incidence of secondary leukemias. Haematologica 84:937–945. Li FP, Fine W, Jaffe N, et al. 1979. Offspring of patients treated for cancer in childhood. J Natl Cancer Inst 62:1193–1197. Linet MS. 1985. The Leukemias: Epidemiologic Aspects. New York: Oxford University Press. Linet MS, Cartwright RA. 1996. The leukemias. In: Schottenfeld DM, Fraumeni JF, eds. Cancer Epidemiology and Prevention 2nd ed. New York: Oxford University Press, pp. 841–892.

866

PART IV: CANCER BY TISSUE OF ORIGIN

Linet MS, Devesa SS. 2002. Epidemiology of leukemia: Overview and patterns of occurrence. In: Henderson E, Lister TA, Greaves MF, eds. Leukemia 7th ed. Philadelphia: WB Saunders, pp. 131–151. Linet MS, Hatch EE, Kleinerman RA, et al. 1997. Residential exposure to magnetic fields and acute lymphoblastic leukemia in children. N Engl J Med 337:1–7. Linet MS, McCaffrey LD, Humphrey RL, et al. 1986. Chronic lymphocytic leukemia and acquired disorders affecting the immune system: A casecontrol study. J Natl Cancer Inst 77:371–378. Linet MS, Pottern LM. 1992. Familial aggregation of hematopoietic malignancies and risk of non-Hodgkin’s lymphoma. Cancer Res 52 (19 Suppl): 5468s–5473s. Linet MS, Smith MA, Ries LAG, et al. 1999. Cancer surveillance series: Recent trends in childhood cancer incidence and mortality in the United States. J Natl Cancer Inst 91:1051–1058. Linet MS, Van Natta ML, Brookmeyer R, et al. 1989. Familial cancer history and chronic lymphocytic leukemia. A case-control study. Am J Epidemiol 130:655–664. Linos A, Blair A, Cantor KP, et al. 1990. Leukemia and non-Hodgkin’s lymphoma among embalmers and funeral directors. J Natl Cancer Inst 82:66. Little J. 1999. Epidemiology of Childhood Cancer. IARC Scientific Publications No. 149. Lyon, France: International Agency for research on Cancer. Little JB. 1993. Cellular, molecular, and carcinogenic effects of radiation. Hematol Oncol Clin N Am 7:337–352. Little MP, Muirhead CR. 1998. Curvature in the cancer mortality dose response in Japanese atomic bomb survivors: Absence of evidence of threshold. Int J Radiat Biol 74:471–480. Little MP, Weiss HA, Boice JD Jr, et al. 1999. Risks of leukemia in Japanese atomic bomb survivors, in women treated for cervical cancer, and in patients treated for ankylosing spondylitis. Radiat Res 152: 280–292. Liu Y, Szekely L, Drander D, et al. 1993. Chronic lymphocytic leukemia cells with allelic deletion at 13q14 commonly have one intact RBI gene: Evidence for a role of an adjacent locus. Proc Natl Acad Sci USA 90: 8697–8701. Lohmueller KE, Pearce CL, Pike M, et al. 2003. Meta-analysis of genetic association studies supports a contribution of common variants to susceptibility to common disease. Nat Genet 33:177–182. London SJ, Bowman JD, Sobel E, et al. 1994. Exposure to magnetic fields among electrical workers in relation to leukemia risk in Los Angeles County. Am J Ind Med 26:47–60. London SJ, Thomas DC, Bowman JD, et al. 1991. Exposure to residential electric and magnetic fields and risk of childhood leukemia. Am J Epidemiol 134:923–937. Long DJ 2nd, Gaikwad A, Multani A, et al. 2002. Disruption of the NAD(P)H: quinone oxidoreductase 1 (NQD1) gene in mice causes myleogenous hyperplasia. Cancer Res 62:3030–3036. Lowengart RA, Peters JM, Cicioni C, et al. 1987. Childhood leukemia and parents’ occupational and home exposures. J Natl Cancer Inst 79: 39–46. Lubin JH, Linet MS, Boice JD Jr, et al. 1998. Case-control study of childhood acute lymphoblastic leukemia and residential radon exposure. J Natl Cancer Inst 90:294–300. Lynge E. 1994. Danish Cancer Registry as a resource for occupational research. J Occup Med 36:1169–1173. Lynge E, Andersen A, Nilsson R, et al. 1997. Risk of cancer and exposure to gasoline vapors. Am J Epidemiol 145:449–458. Ma J, Stampfer MJ, Giovannucci E, et al. 1997. Methylenetetrahydrofolate reductase polymorphism, dietary interactions, and risk of colorectal cancer. Cancer Res 57:1098–1102. Ma J. Stampfer MJ, Hennekens CH, et al. 1996. Methylenetetrahydrofolate reductase polymorphism, plasma folate, homocysteine, and risk of myocardial infarction in US physicians. Circulation 94:2410–2416. Ma X, Buffler PA, Gunier RB, et al. 2002. Critical windows of exposure to household pesticides and risk of childhood leukemia. Environ Health Perspect 110:955–960. Ma X, Buffler PA, Selvin S, et al. 2002. Daycare attendance and risk of childhood acute lymphoblastic leukaemia. Br J Cancer 86:1419–1424. MacKenzie J, Gallagher A, Clayton RA, et al. 2001. Screening for herpesvirus genomes in common acute lymphoblastic leukaemia. Leukemia 15: 415–421. MacKenzie J, Perry J, Ford AM, et al. 1999. JC and BK virus sequences are not detectable in leukemic samples from children with common acute lymphoblastic leukaemia. Br J Cancer 81:898–899. MacMahon B. 1962. Prenatal X-ray exposure and childhood cancer. J Natl Cancer Inst 28:1173–1191. MacMahon B, Newill VA. 1962. Birth characteristics of children dying of malignant neoplasms. J Natl Cancer Inst 28:231–244.

Magnani C, Pastore G, Luzzatto L, et al. 1990. Parental occupation and other environmental factors in the etiology of leukemias and non-Hodgkin’s lymphomas in childhood: A case-control study. Tumori 76:413–419. Maia AT, Ford AM, Jalali GR, et al. 2001. Molecular tracking of leukemogenesis in a triplet pregnancy. Blood 98:478–482. Malkin D, Brown EJ, Zipursky A. 2000. The role of p53 in megakaryocyte differentiation and the megakaryocytic leukemias of Down syndrome. Cancer Genet Cytogenet 116:1–5. Mancuso TF. 1975. Epidemiological investigation of occupational cancers in the rubber industry. In: Levinson C, ed. The New Multinational Health Hazards. Geneva: ICF, pp. 80–136. Mandel K, Dror Y, Poon A, et al. 2002. A Practical, comprehensive classification for pediatric myelodysplastic syndromes: The CCC system. J Pediatr Hematol Oncol 24:596–605. Mandla SG, Goobie S, Kumar RT, et al. 1998. Genetic analysis of familial myelodysplastic syndrome: Absence of linkage to chromosomes 5q31 and 7q22. Cancer Genet Cytogenet 105:113–118. Manns A, Hisada M, La Grenade L. 1999. Human T-lymphotropic virus type I infection. Lancet 353:1951–1958. Marsh GM, Enterline PE, McGraw D. 1991. Mortality patterns among petroleum refinery and chemical plant workers. Am J Ind Med 19: 29–42. Martling U, Mattsson A, Travis LB, et al. 1999. Mortality after long-term exposure to radioactive Thorotrast: A forty-year follow-up survey in Sweden. Radiat Res 151:293–299. Mason JB, Levesque T. 1996. Folate: Effects on carcinogenesis and the potential for cancer chemoprevention. Oncology 10:1727–1736. Matanoski GM, Elliott E, Breysse PN, et al. 1993. Leukemia in telephone linemen. Am J Epidemiol 137:609–619. Matanoski GM, Santos-Burgoa C, Schwartz L. 1990. Mortality of a cohort of workers in the styrene-butadiene polymer manufacturing industry (1943– 1982). Environ Health Perspect 86:107–117. Matanoski GM, Sartwell P, Elliott E, et al. 1984. Cancer risks in radiologists and radiation workers. In: Boice JD Jr., Fraumeni. JF Jr, eds. Radiatoin Carcinogenesis: Epidemiology and Biological Significance. New York: Raven Press, pp. 83–96. Matanoski GM, Stockwell HG, Diamond EL, et al. 1986. A cohort mortality study of painters and allied tradesmen. Scand J Work Environ Health 12:16–21. Matsuo T, Tomonaga M, Bennett JM, et al. 1988. Reclassification of leukemia among A-bomb survivors in Nagasaki using French-AmericanBritish (FAB) classification for acute leukemia. Jpn J Clin Oncol 18:91–96. McBride ML, Gallagher RP, Theriault G, et al. 1999. Power-frequency electric and magnetic fields and risk of childhood leukemia in Canada. Am J Epidemiol 149:831–842. McCunney RJ. 1999. Hodgkin’s disease, work, and the environment. A review. J Occup Environ Med 41:36–46. McDonald TA, Holland NT, Skikbola C. et al. 2001. Hypothesis: phenol and hydroquinone derived mainly from diet and gastrointestinal flora activity are causal factors in leukemia. Leukemia 15:10–20 McKenna RW. 2000. Multifaceted approach to the diagnosis and classification of acute leukemias. Clin Chem 468 (Pt 2):1252–1259. McKinney PA, Alexander FE, Cartwright RA, et al. 1991. Parental occupations of children with leukaemia in west Cumbria, north Humberside, and Gateshead. BMJ 302:681–687. McKinney PA, Juszczak E, Findlay E, et al. 1999. Pre- and perinatal risk factors for childhood leukaemia and other malignancies: A Scottish case control study. Br J Cancer 80:1844–1851. McLaughlin JR, King WD, Anderson TW, et al. 1993b. Paternal radiation exposure and leukaemia in offspring: The Ontario case-control study. BMJ 307:959–966. McMichael AJ, Spirtas R, Gamble JF, et al. 1976. Mortality among rubber workers: Relationship to specific jobs. J Occup Med 18:178–185. Meinert R, Schuz J, Kaletsch U, et al. 2000. Leukemia and non-Hodgkin’s lymphoma in childhood and exposure to pesticides: Results of a registerbased case-control study in Germany. Am J Epidemiol 151:639–646. Mele A, Szklo M, Visani G, et al. 1994. Hair dye use and other risk factors for leukemia and pre-leukemia: A case-control study. Italian Leukemia Study Group. Am J Epidemiol 139:609–619. Melikian AA, Prahalad AK, Hoffmann D. 1993. Urinary trans, Ftransmuconic acid as an indicator of exposure to benzene in cigarette smokers. Cancer Epidemiol Biomarkers Prev 2:47–51. Melton D, Ketchen AM, Nuntilde F, et al. 1998. Cells from ERCC1-deficient mice show increased genome instability and a reduced frequency of Sphase genome instability and a reduced frequency of S-phase-dependent illegitimate chromosome exchange but a normal frequency of homologous recombination. J Cell Sci 111:395–404.

The Leukemias Merk O, Speit G. 1998. Significancer of formaldehyde-induced DNA-protein crosslinks for mutagenesis. Environ Mol Mutagen 32:260–268. Metayer C, Curtis RE, Vose J, et al. 2003. Myelodysplastic syndrome and acute myeloid leukemia after autotransplantation for lymphoma: A multicenter case-control study. Blood 101:2015–2023. Michaelis J, Schuz J, Meinert R, et al. 1997. Childhood leukemia and electromagnetic fields: Results of a population-based case-control study in Germany. Cancer Causes Control 8:167–174. Milham S. 1982. Mortality from leukemia in workers exposed to electrical and magnetic fields [letter to the editor]. N Engl J Med 307:249. Miligi L, Seniori Costantini A, Crosignani P, et al. 1999. Occupational, environmental and life-style factors associated with the risk of hematolymphopoietic malignancies in women, Am J Ind Med 36:60–69. Miller AB, To T, Agnew DA, et al. 1996. Leukemia following occupational exposure to 60 Hz electric and magnetic fields among Ontario electricity utility workers. Am J Epidemiol 144:150–160. Millikan RC, Pittman GS, Newman B, et al. 1998. Cigarette smoking, Nacetyltransferases 1 and 2, and breast cancer risk. Cancer Epidemiol Biomark Prev 7:371–378. Mire-Sluis A. 1999. Cytokine and growth factor standardization: the need for research to answer modern issues. Dev Biol Stand 100:83–93. Modan B. 1965. An epidemiological study of polycythemia vera. Blood 26:657–667. Mohan AK, Hauptmann M, Freedman DM, et al. 2003. Cancer and other causes of mortality among radiologic technologists in the United States. Int J Cancer 103:259–267. Monson RR, Nakano KK. 1976. Mortality among rubber workers. I. White male union employees in Akron, Ohio. Am J Epidemiol 103:284–296. Moorman A, Roman E, Cartwright RA, et al. 2002a. Age-specific incidence rates of cytogenetically-defined subtypes of acute myeloid leukemia. Br J Cancer 86:1061–1063. Moorman A, Roman E, Cartwright RA, et al. 2002b. Smoking and the risk of acute myeloid leukemia and cytogenetic subgroups. Br J Cancer 86: 60–62. Moorman A, Roman E, Willett EV, et al. 2001. Karyotype and age in acute myeloid leukemia. Are they linked? Cancer Genet Cytogenet 126: 155–161. Morgan RW, Claxton KW, Divine BJ, et al. 1981. Mortality among ethylene oxide workers. J Occup Med 25:767–770. Mori H, Colman SM, Xiao Z, et al. 2002. Chromosome translocations and covert leukenic clones are generated during normal fetal development. Proc Natl Acad Sci USA 99:8242–8247. Mrozek K, Heinonen K, de la Chapelle A, et al. 1997. Clinical significance of cytogenetics in acute myeloid leukemia. Semin Oncol 24:17–31. Muirhead CR, Goodill AA, Haylock RG, et al. 1999. Occupational radiation exposure and mortality: Second analysis of the National Registry for Radiation Workers. J Radiol Prot 19:3–26. Mulvihill JJ, Myers MH, Connelly RR, et al. 1987. Cancer in offspring of long-term survivors of childhood and adolescent cancer. Lancet ii: 813–817. Mundschau G, Gurbunxani S, Gaines AS, et al. 2003. Mutagenesis of GATA1 is an initiating event in Down syndrome leukemogenesis. Blood 101: 4298–4300. Murphy EL. Hanchard B, Figueroa JP, et al. 1989. Int J Cancer 43:250–253. Murray L, McCarron P, Bailie K, et al. 2002. Association of early life factors and acute lymphoblastic leukaemia in childhood: Histroical cohort study. Br J Cancer 86:356–361. Nagata C, Shimizu H, Hirashima K, et al. 1999. Hair dye use and occupational exposure to organic solvents as risk factors for myelodysplastic syndrome. Leuk Res 23:57–62. Nakanishi M, Tanaka K, Shintani T, et al. 1999. Chromosomal instability in acute myelocytic leukemia and myelodysplastic syndrome patients among atomic bomb survivors. J Radiat Res (Tokyo) 40:159–167. Nakano Y, Naoe T, Kiyoi H, et al. 2000. Poor clinical significance of p53 gene polymorphism in acute myeloid leukemia. Leuk Res 24:349–352. Nanni O, Amadori D, Lugaresi C, et al. 1996. Chronic lymphocytic leukaemias and non-Hodgkin’s lymphomas by histological type in farming-animal breeding workers: A population case-control study based on a prior exposure matrices. Occup Environ Med 53:652–657. Naoe T, Takeyama K, Yokozawa T, et al. 2000. Analysis of genetic polymorphism in NQ01, GST-M1, GST-T1, and CYP3A4 in 469 Japanese patients with therapy-related leukemia/myelodysplastic syndrome and de nova acute myeloid leukemia. Clin Cancer Res 6:4091–4095. NRC (National Research Council) 1999. Health Effects of Exposure to Radon. BEIR VI. Washington, DC: National Academy of sciences, pp. 18–20. Naumburg E, Bellocco R, Cnattingius S, et al. 2001. Intrauterine exposure to diagnostic x-rays and risk of childhood leukemia subtypes. Radiat Res 156:718–723.

867

Neglia JP, Linet MS, Shu XO, et al. 2000. Patterns of infection and day care utilization and risk of childhood acute lymphoblastic leukaemia. Br J Cancer 82:234–240. Neglia JP, Robison LL. 1988. Epidemiology of childhood leukemias. Pediatr Clin North Am 35:675–692. Nilsson RI, Nordlinder R, Horte LG, et al. 1998. Leukaemia, lymphoma, and multiple myeloma in seamen on tankers. Occup Environ Med 55: 517–521. Nishi M, Miyake H. 1989. A case-control study of non-T cell acute lymphoblastic leukaemia of children in Hokkaido, Japan. J Epidemiol Community Health 43:352–355. Nishio H, Lee MJ, Fujii M, et al. 1996. A common mutation in methylenetetrahydrofolate reductase gene among the Japanese population. Jpn J Hum Genet 41:247–251. Nisse C, Lorthois C, Dorp V, et al. 1995. Exposure to occupational and environmental factors in myelodysplastic syndromes. Preliminary results of a case-control study. Leukemia 9:693–699. Nosslinger T, Reisner R, Koller E, et al. 2001. Myelodysplastic syndromes, from French-American-British to World Health Organization: Comparison of classifications on 431 unselected patients from a single institution. Blood 98:2935–2941. Obrams GI, Grufferman S. 1991. Epidemiology of HIV associated nonHodgkin’s lymphoma. Cancer Surv 10:91–102. O’Brien TR, Decoufle P. 1988. Cancer mortality among northern Georgia carpet and textile workers. Am J Ind Med 14:15–24. Oikawa S, Hirosawa I, Hirakawa K, et al. 2001. Site specificity and mechanism of oxidative DNA damage induced by carcinogenic catechol. Carcinogenesis 22:1239–1245. Olopade OI, Roulston D, Baker T, et al. 1996. Familial myeloid leukemia associated with loss of the long arm of chromosome 5. Leukemia 10: 669–674. Ong CN Lee BL. 1994. Determination of benzene and its metabolites: application in biological monitoring of environmental and occupational exposure to benzene. J Chromatogr B Biomed Appl 660:1–22. Ott MG, Kolesar RC, Scharnweber HC, et al. 1980. A mortality survey of employees engaged in the development or manufacture of styrene-based products. J Occup Med 22:445–460. Ott MG, Townsend JC, Fishbeck WA, et al. 1978. Mortality among individuals occupationally exposed to benzene. Arch Environ Med 33:3–10. Ou CY, Stevenson RE, Brown VK, et al. 1996. 5, 10 Methylenetetrahydrofolate reductase genetic polymorphism as a risk factor for neural tube defects. Am J Med Genet 63:610–614. Paganini-Hill A, Glazer E, Henderson BE, et al. 1980. Cause-specific mortality among newspaper web pressman. J Occup Med 22:542–544. Paietta E, Racevskis J, Bennett JM, et al. 1998. Biologic heterogeneity in Philadelphia chromosome-positive acute leukemia with myeloid morphology: The Eastern Cooperative Oncology Group experience. Leukemia 12:1881–1885. Panagopoulos I, Isaksson M, Lindvall C, et al. 2003. Genomic characterization of MOZ/CBP and CBP/MOZ chimeras in acute myeloid leukemia suggests the involvement of a damage-repair mechanism in the origin of the t (8;16) (p11;p13). Genes Chromosomes Cancer 36:90–98. Parkes HG, Veys CA, Waterhouse JAH, et al. 1982. Cancer mortality in the British rubber industry. Br J Ind Med 39:209–220. Parkin DM, Whelan SL, Ferlay J, et al. 2002. Cancer Incidence in Five Continents. Volume VIII. IARC Scientific Publications No. 155. Lyon, France: International Agency for Research on Cancer. Paustenbach DJ, Price PS, Ollison W, et al. 1992. Reevaluation of benezene exposure for the Pliofilm (rubberworker) cohort (1936–1976). J Toxicol Environ Health 36:177–231. Pearce N, Porta M. 1986. Association of non-Hodgkin’s lymphoma with rheumatoid arthritis. Am J Med 81:747–748. Pearce N, Reif JS. 1990. Epidemiologic studies of cancer in agricultural workers. Am J Ind Med 18:133–148. Pearce N, Smith AH, Reif JS. 1988. Increased risks of soft tissue sarcoma, malignant lymphoma, and acute myeloid leukemia in abattoir workers. Am J Ind Med 14:63–72. Pedersen-Bjergaard J, Andersen MK, Christiansen DH. 2000. Therapyrelated acute myeloid leukemia and myelodysplasia after high-dose chemotheapy and autologous stem cell transplantation. Blood 95: 3273–3279. Perentesis JP. 2001. Genetic predisposition and treatment-related leukemia. Med Pediatr Oncol 36:541–548. Perrillat F, Clavel J, Auclerc MF, et al. 2002. Day-care, early common infections and childhood acute leukemia: A multicentre French case-control study. Br J Cancer 86:1064–1069. Pesatori AC, Consonni D, Tironi A, et al. 1993. Cancer in a young population in a dioxin-contaminated area. Int J Epidemiol 22:1010–1013.

868

PART IV: CANCER BY TISSUE OF ORIGIN

Petridou E, Hseh CC, Kotsifakis G, et al. 1991. Absence of laukaemia clustering on Greek islands [letter to the editor]. Lancet 338:1204–1205. Petridou E, Trichopoulis D, Kalapothaki V, et al. 1997. The risk profile of childhood leukaemia in Greece: A nationwide case-control study. Br J Cancer 76:1241–1247. Pierce DA, Shimizu Y, Preston DL, et al. 1996. Studies of the mortality of atomic bomb survivors. Report 12, Part I. Cancer: 1950–1990. Radiat Res 146:1–27. Pogribny IP, Muskhelishvili L, Miller BJ, et al. 1997. Presence and consequence of uracil in preneoplastic DNA from folate/methyldeficient rats. Carcinogenesis 18:2071–2076. Polhemus DW, Koch R. 1959. Leukemia and Medical radiation. Pediatrics 23:453–461. Portier CJ, Wolfe MS, eds. 1998. Assessment of Health Effects from Exposure to Power-Line Frequency Electric and Magnetic Fields. NIEHS, NIH Publ No 98-3981. Research Triangle Park, NC: NIEHS, pp. 108–133. Pottern LM, Linet M, Blair A, et al. 1991. Familial cancers associated with subtypes of leukemia and non-Hodgkin’s lymphoma. Leuk Res 15: 305–314. Preston DL, Kusumi S, Tomonaga M, et al. 1994. Cancer incidence in atomic bomb survivors. Part III. Leukemia, lymphoma and multiple myeloma, 1950–1987. Radiat Res 137:S68–97. Preston DL, Pierce DA. 1988. The effect of changes in dosimetry on cancer mortality risk estimates in the atomic bomb survivors. Radiat Res 114:437–466. Preston-Martin S, Thomas DC, Yu MC, et al. 1989. Diagnostic radiography as a risk factor for chronic myeloid and monocytic leukaemia (CML). Br J Cancer 59:649–644. Qu Q, Shore R, Li G, et al. 2002. Hematological changes among Chinese workers with a broad range of benzene exposures. Am J Med 42:275–285. Raabe GK, Wong O. 1996. Leukemia mortality by cell type in petroleum workers with potential exposure to benzene. Environ Health Perspect 104:1381–1392. Rawstron AC, Green MJ, Kuznicki A, et al. 2002a. Monoclonal B lymphocytes with the characteristics of “indolent” chronic lymphocytic leukaemia are present in 3.5% of adults with normal blood counts. Blood 100: 635–639. Rawstron AC, Yuille MR, Fuller J, et al. 2002b. Inherited predisposition to CLL is detectable as subclinical monoclonal B-lymphocyte expansion. Blood 100:2289–2290. Reichel M, Gillert E, Angermuller S, et al. 2001. Biased distribution of chromosomal breakpoints involving the MLL gene in infants versus children and adults with t(4; 11) acute lymphoblastic leukemia. Oncogene 20: 2900–2907. Reynolds P, von Behren J, Elkin EP. 2002b. Birth characteristics and leukemia in young children. Am J Epidemiol 155:603–613. Reynolds P, von Behren J, Gunier RB, et al. 2002a. Childhood cancer and agricultural pesticide use: An ecologic study in California. Environ Health Perspect 110:319–324. Riccardone MD, Ozcelik T, Cevher B, et al. 1999. Human MLH1 deficiency predisposes to haematological malignancy and neurofibromatosis type 1. Cancer Res 59:290–293. Ries LAG, Eisner MP, Kosary CL, et al. 2003. SEER Cancer Statistics Review, 1975–2000. Hankey BF, Miller BA, Clegg L, Mariotto A, Fay MP, Feuer EJ, Edwards BK, eds. Bethesda, MD: National Cancer Institute. Available at: http://seer.cancer.gov/csr/1975–2000. Rinaldi F, Mairs RJ, Wheldon TE, et al. 1999. Clonality analysis suggests that early-onset acute lymphoblastic leukaemia is of single-cell origin and implies no major role for germ cell mutations in parents. Br J Cancer 80:909–913. Rinsky RA, Hornung RW, Silver SR, et al. 2002. Benzene exposure and hematopoietic mortality: A long-term epidemiologic risk assessment. Am J Ind Med 42:474–480. Rinsky RA, Smith AB, Hornung R, et al. 1987. Benzene and leukemia. An epidemiologic risk assessment. N Engl J Med 316:1044–1050. Rischewski JR, Clausen H, Leber V, et al. 2000. A heterozygous frameshift mutation in the Fanconi anemia C gene in familial T-ALL and secondary malignancy. Klin Paediatr 212:174–176. Robinette CD, Jablon S, Preston TL. 1985. Mortality of Nuclear Weapons Test Participants. Medical Follow-Up Agency. National Research Council. Washington, DC: National Academy Press. Robison LL, Buckley JD, Daigle AE, et al. 1989. Maternal drug use and risk of childhood nonlymphoblastic leukemia among offspring. An epidemiologic investigation implicating marijuana (a report from the Children’s Cancer Study Group). Cancer 63:1904–1911. Robison LL, Codd M, Gunderson P, et al. 1987. Birth weight as a risk factor for childhood acute lymphoblastic leukemia. Pediatr Hematol Oncol 4:63–72.

Roddam PL, Allan JM, Rollinson S, et al. 2003. Poor metabolizer status at the cytochrome P450 2C9 and 2D6 loci does not modulate susceptibility to therapy-related acute myeloid leukaemia. Br J Haematol 121:192– 194. Roddam PL, Rollinson S, Kane E, et al. 2000. Poor metabolizers at the cytochrome P450 2D6 and 2C19 loci are at increased risk of developing adult acute leukaemia. Pharmacogenetics 10:605–615. Rollinson S, Roddam P, Kane E, et al. 2000. Polymorphic variation within the glutathione S-transferase genes and risk of adult acute leukaemia. Carcinogenesis 21:43–47. Rollinson S, Roddam P, Willett E, et al. 2001. NAT2 acetylator genotypes confer no effect on the risk of developing adult acute leukemia: A casecontrol study. Cancer Epidemiol Biomarkers Prev 10:567–568. Roman E, Ansell P, Bull D. 1997. Leukaemia and non-Hodgkin’s lymphoma in children and young adults: Are prenatal and neonatal factors important determinants of disease? Br J Cancer 76:406–415. Roman E, Beral V, Carpenter L, et al. 1987. Childhood leukaemia in the West Berkshire and Basingstoke and North Hampshire District Health Authorities in relation to nuclear establishments in the vicinity. Br Med J 294:597–602. Roman E, Doyle P, Maconochie N, et al. 1999. Cancer in children of nuclear industry employees: Report on children aged under 25 years from nuclear industry family study. BMJ 318:1443–1450. Roman E, Watson A, Beral V, et al. 1993. Case-control study of leukaemia and non-Hodgkin’s lymphoma among children aged 0–4 years living in west Berkshire and north Hampshire health districts. BMJ 306:615–621. Ron E, Modan B, Boice JD Jr. 1988. Mortality after radiotherapy for ringworm of the scalp. Am J Epidemiol 127:713–725. Rosenblatt KA, Koepsell TD, Daling JR, et al. 1991. Antigenic stimulation and the occurrence of chronic lymphocytic leukemia. Am J Epidemiol 134:22–28. Ross JA, Potter JD, Reaman GH, et al. 1996. Maternal exposure to potential inhibitors of DNA topoisomerase II and infant leukemia (United States): A report from the Children’s Cancer Group. Cancer Causes Control 7:581–590. Ross JA, Potter JD, Robison LL. 1994. Infant leukemia, topoisomerase II inhibitors, and the MLL gene. J Natl Cancer Inst 86:1678–1680. Ross JA, Potter JD, Shu XO, et al. 1997. Evaluating the relationships among maternal reproductive history, birth characteristics, and infant leukemia: A report from the Children’s Cancer Group. Ann Epidemiol 7:172–179. Rossi G, Pelizzari AM, Bellotti D, et al. 2000. Cytogenetic analogy between myelodysplastic syndrome and acute myeloid leukemia of elderly patients. Leukemia 14:636–641. Rothman N, Li GL, Dosemeci M, et al. 1996. Hematotoxicity among Chinese workers heavily exposed to benzene. Am J Ind Med 29:236–246. Rothman N, Smith MT, Hayes RB, et al. 1997a. Benzene poisoning, a risk factor for haematological malignancy, is associated with the NQO1 609 CÆT mutation and rapid fractional excretion of chlorzoxazone. Cancer Res 57:2839–2842. Rothman N, Wacholder S, Caporaso NE, et al. 2001. The use of common genetic polymorphisms to enhance the epidemiologic study of environmental carcinogens. Biochim Biophys Acta 147:C1–10. Rowley JD. 2000. Molecular genetics in acute leukemia. Leukemia 14: 513–517. Rushton L. 1993. A 39-year follow-up of the U.K. oil refinery and distribution center studies: Results for kidney cancer and leukaemia. Environ Health Perspect 101:77–84. Rushton L, Romaniuk H. 1997. A case-control study to investigate the risk of leukaemia associated with exposure to benzene in petroleum marketing and distribution workers in the United Kingdom. Occup Environ Med 54:152–166. Sahl JD, Kelsh MA, Greenland S. 1993. Cohort and nested case-control studies of hematopoietic cancers and brain cancer among electric utility workers. Epidemiology 4:104 –114. Sanders BM, White GC, Draper GJ. 1981. Occupations of fathers of children dying from neoplasms. J Epidemiol Community Health 35:245–250. Sandler DP, Shore DL, Anderson JR, et al. 1993. Cigarette smoking and risk of acute leukemia: Associations with morphology and cytogenetic abnormalities in bone marrow. J Natl Cancer Inst 85:1994–2003. Santon A, Arranz E, Robledo M, et al. 1995. N-Ras mutations: Incidence and clinical relevance in meylodysplastic syndrome and acute myelogenous leukaemia. Am J Hematol 50(4):315–316. Sarasua S, Savitz DA. 1994. Cured and broiled meat consumption in relation to childhood cancer: Denver, Colorado. Cancer Causes Control 5: 141–148. Sasai Y, Horiike S, Misawa S, et al. 1999. Genotype of glutathione Stransferase and other genetic configurations in myelodysplasia. Leuk Res 23:975–981.

The Leukemias Sasco AJ, Vainio H. 1999. From in utero and childhood exposure to parental smoking to childhood cancer: A possible link and the need for action. Hum Exp Toxicol 18:192–201. Saso R, Kulkarni S, Mitchell P, et al. 2000. Secondary myelodysplastic syndrome/acute myeloid leukaemia following mitoxantrone-based therapy for breast carcinoma. Br J Cancer 83:91–94. Sathiakumar N, Delzell E. 1997. A review of epidemiologic studies of triazine herbicides and cancer. Crit Rev Toxicol 27:599–612. Sathiakumar N, Delzell E. 1998. Mortality from cancer and other causes of death among synthetic rubber workers. Occup Environ Med 55: 230–235. Sathiakumar N, Delzell E, Cole P, et al. 1995. A case-control study of leukemia among petroleum workers. J Occup Environ Med 37:1269–1277. Savitz DA, Andrews KW. 1997. Review of epidemiologic evidence on benzene and lymphatic and hematopoietic cancers. Am J Ind Med 31:287–295. Savitz DA, John EM, Kleckner RC. 1990. Magnetic field exposure from electric appliances and childhood cancer. Am J Epidemiol 131:763–773. Savitz DA, Loomis DP. 1995. Magnetic field exposure in relation to leukemia and brain cancer mortality among electric utility workers. Am J Epidemiol 141:123–134. Savitz DA, Wachtel H, Barnes FA, et al. 1988. Case-control study of childhood cancer and exposure to 60-Hz magnetic fields. Am J Epidemiol 128: 21–38. Schnatter AR, Armstrong TW, Nicolich MJ, et al. 1996a. Lymphohaematopoietic malignancies and quantitative estimates of exposure to benzene in Canadian petroleum distribution workers. Occup Environ Med 53: 773–781. Schnatter AR, Armstrong TW, Thompson LS, et al. 1996b. The relationship between low-level benzene exposure and leukemia in Canadian petroleum distribution workers. Environ Health Perspect 104:1375–1379. Schnatter AR, Theriault G, Katz AM, et al. 1992. Thompson FS, Doraleski D, Murray N. A retrospective mortality study within operating segments of a petroleum company. Am J Ind Med 22:209–229. Schuz J, Kaatsch P, Kaletsch U, et al. 1999. Meinert R, Michaelis J. Association of childhood cancer with factors related to pregnancy and birth. Int J Epidemiol 28:631–639. Schuz J, Kaletsch U, Meinert R, et al. 1999. Association of childhood leukemia with factors related to the immune system. Br J Cancer 80:585–590. Secker-Walker LM. 1991. Distribution of Philadelphia positive acute lymphoblastic leukemia: Geographical heterogeneity or age related incidence? Genes Chromosomes Cancer 3:320–321. Seedhouse C, Bainton R, Lewis M, et al. 2002. The genotype distribution of the XRCC1 gene indicates a role for base excision repair in the development of therapy-related acute myeloblastic leukemia. Blood 100:3761– 3766. Seidemann K, Henze G, Beck JD, et al. 2000. Non-Hodgkin’s lymphoma in pediatric patients with chromosomal breakage syndromes (AT and NBS): Experience from the BFM trials. Ann Oncol 11 Suppl 1:141–145. Selhub J. 1994. Folate binding proteins. Mechanisms for placental and intestinal uptake. Adv Exp Med Biol 352:141–149. Seltser R, Sartwell PE. 1965. The influence of occupational exposure on the mortality of American radiologists and other medical specialists. Am J Epidemiol 101:188–198. Severson RK, Buckley JD, Woods WG, et al. 1993. Cigarette smoking and alcohol consumption by parents of children with acute myeloid leukemia: An analysis within morphological subgroups—a report from the Children’s Cancer Group. Cancer Epidemiol Biomarkers Prev 2:433–439. Severson RK, Stevens RG, Kaune WT, et al. 1988. Acute nonlymphocytic leukemia and residential exposure to power frequency magnetic fields. Am J Epidemiol 128:10–20. Sgambati M, Linet M, Devesa S. 2001. Chronic lymphocytic leukemia: Epidemiological, familial, and genetic aspects. In: Cheson B, ed. Chronic Lymphocytic Leukemia. 2nd ed. New York: Marcel Dekker, Inc., pp. 33–62. Shaham J, Bornstein Y, Meltzer A, et al. 1996. DNA–protein crosslinks, a biomarker of exposure to formaldehyde—in vitro and in vivo studies. Carcinogenesis 17:121–125. Shannon KM, Watterson J, Johnson P, et al. 1992. Monosomy 7 myeloproliferative disease in children with neurofibromatosis, type 1: Epidemiology and molecular analysis. Blood 79:1311–1318. Shaw G, Lavey R, Jackson R, et al. 1984. Association of childhood leukemia with maternal age, birth order, and paternal occupation. A case-control study. Am J Epidemiol 119:788–795. Shieh B, Liau YE, Hsieh PS, et al. 2000. Influence of nucleotide polymorphisms in the CCR2 gene and the CCR5 promoter on the expression of cell surface CCR5 and CXCR4. Int Immunol 12:1311–1318. Shimizu Y, Kato H, Schull W, et al. 1989. Studies of the mortality of A-bomb survivors. Mortality, 1950–1985. Part 1. Comparison of the risk coeffi-

869

cients for site-specific cancer mortality based on DS86 and T65DR shielded kerma and organ doses. Radiat Res 118:502–524. Shimizu Y, Pierce DA, Preston DL, et al. 1999. Studies of the mortality of atomic bomb survivors. Report 12, part II. Noncancer mortality: 1950–1990. Radiat Res 152:374–389. Shimizu Y, Schull WJ, Kato H. 1990. Cancer risk among atomic bomb survivors. The RERF Life Span Study. Radiation Effects Research Foundation. JAMA 264:601–604. Shimon I, Kneller A, Olchovsky D. 1995. Chronic myeloid leukaemia following 1311 treatment for thyroid carcinoma: A report of two cases and review of the literature. Clin Endocrinol (Oxf) 43:651–654. Shore RE, Albert RE, Pasternack BS. 1976. Follow-up study of patients treated by X-ray epilation for tinea capitis; resurvey of post-treatment illness and mortality experience. Arch Environ Health 31:21–28. Shu X-O, Gao Y-T, Brinton LA, et al. 1988. A population-based case-control study of childhood leukemia in Shanghai. Cancer 62:635–644. Shu X-O, Gao Y-T, Linet MS, et al. 1987. Chloramphenicol use and childhood leukaemia in Shanghai. Lancet ii:934–937. Shu X-O, Han D, Severson RK, et al. 2002a. Birth certificates, maternal reproductive history, hormone use during pregnancy, and risk of childhood acute lymphocytic leukemia by immunophenotyp (Unites States). Cancer Causes Control 13:15–25. Shu X-O, Linet MS, Steinbuch M, et al. 1999a. Breast-feeding and risk of childhood acute leukemia. J Natl Cancer Inst 91:1765–1772. Shu X-O, Potter JD, Linet MS, et al. 2002b. Diagnostic x-rays and ultrasound exposure and risk of childhood acute lymphoblastic leukemia by immunophenotype. Cancer Epidemiol Biomarkers Prev 11:177–185. Shu X-O, Ross JA, Pendergrass TW, et al. 1996. Parental alcohol consumption, cigarette smoking, and risk of infant leukemia: A Children’s Cancer Group study. J Natl Cancer Inst 88:24–31. Shu X-O, Stewart P, Wen W-Q, et al. 1999b. Parental occupational exposure to hydrocarbons and risk of acute lymphocytic leukemia in offspring. Cancer Epidemiol Biomarkers Prev 8:783–791. Shugart YY. 1998. Anticipation in familial Hodgkin lymphoma. Am J Hum Genet 63:270–272. Siegel R, Gartenhaus R, Kuzel T. 2001. HTLV-I associated leukemia/ lymphoma: Epidemiology, biology, and treatment. Cancer Treat Res 104: 75–88. Skibola CF, Smith MT, Kane E, et al. 1999. Roman E, Rollinson S, Cartwright RA, Morgan G. Polymorphisms in the methylenetetrahydrofolate reductase gene are associated with susceptibility to acute leukemia in adults. Proc Natl Acad Sci USA 96:12810–12815. Smith MA, Rubinstein L, Anderson JR, et al. 1999. Secondary leukemia or myelodysplastic syndrome after treatment with epipodophyllotoxins. J Clin Oncol 17:569–577. Smith MA, Simon R, Strickler HD, et al. 1998. Evidence that childhood acute lymphoblastic leukemia is associated with an infectious agent linked to hygiene conditions. Cancer Causes Control 9:285–298. Smith MA, Strickler HD, Granovsky M, et al. 1999. Investigation of leukemia cells from children with common acute lymphoblastic leukemia for genomic sequences of the primate ployomaviruses JC virus, BK virus, and simian virus 40. Med Pediatr Oncol 33:441– 443. Smith MT. 1999. Benzene, NQO1, and genetic susceptibility to cancer. Proc Natl Acad Sci USA 96:7624–7626. Smith MT, Wang YX, Kane E, et al. 2001. Low NAD (P) H quinone oxidoreductase 1 (NQO1) activity is associated with increased risk of acute leukemia in adults. Blood 97:1422–1426. Smith MT, Zhang L, Jeng M, et al. 2000. Hydroquinone, a benzene metabolite, increases the level of aneusomy of chromosomes 7 and 8 in human CD34-positive blood progenitor cells. Carcinogenesis 21:1485–1490. Smith MT, Zhang L, Wang Y, et al. 1998. Increased translocations and aneusomy in chromosomes 8 and 21 among workers exposed to benzene. Cancer Res 58:2176–2181. Smith PG, Doll R. 1976. Late effects of X irradiation in patients treated for metropathia hemorrhagica. Br J Radiol 49:224–232. Smith PG, Doll R. 1981. Mortality from cancer and all causes among British radiologists. Br J Radiol 54:187–194. Sneller MC, Strober W, Eisenstein E, et al. 1993. NIH conference. New insights into common variable immunodeficiency. Ann Intern Med 118:720–730. Snider DE, Comstock GW, Martinez I, et al. 1978. Efficacy of BCG vaccination in prevention of cancer: An update. J Natl Cancer Inst 60:785– 788. Soderberg KC, Naumburg E, Anger G, et al. 2002. Childhood leukemia and magnetic fields in infant incubators. Epidemiology 13:45–49. Sole F, Espinet B, Sanz GF, et al. 2000. Incidence, characterization and prognostic significance of chromosomal abnormalities in 640 patients with primary myelodysplastic syndromes. Grupo Cooperativo Espanol de Citogenetica Hematologica. Br J Haematol 108:346–356.

870

PART IV: CANCER BY TISSUE OF ORIGIN

Song WJ, Sullivan MG, Legare RD, et al. 1999. Haploinsuffiiciency of CBFA2 causes familial thrombocytopenia with propensity to develop acute myelogenous leukaemia. Nat Genet 23:166–175. Sorahan T, Cooke MA. 1989. Cancer mortality in a cohort of United Kingdom steel foundry workers, 1946–85. Br J Ind Med 46:74–81. Sorahan T, Lancashire RJ, Temperton DH, et al. 1995. Childhood cancer and paternal exposure to ionizing radiation: A second report from the Oxford Survey of Childhood Cancers. Am J Ind Med 28:71–78. Sorahan T, McKinney PA, Mann JR, et al. 2001. Lancashire RM, Stiller CA, Birch JM, Dodd HE, Cartwright RA. Childhood cancer and parental use of tobacco: Findings from the inter-regional epidemiological study of childhood cancer (IRESCC). Br J Cancer 84:141–146. Sorahan T, Prior P, Lancashire RJ, et al. 1997. Childhood cancer and parental use of tobacco: Deaths from 1971 to 1976. Br J Cancer 76:1525–1531. Spiers FW, Lucas HF, Rundo J, et al. 1983. Leukemia incidence in the U.S. dial workers. Health Physics 44:65–72. Spinelli JJ, Gallagher RP, Band PR, et al. 1984. Multiple myeloma, leukemia, and cancer of the ovary in cosmetologists and hairdressers. Am J Ind Med 6:97–102. Stark CR, Mantel N. 1966. Effects of maternal age and birth order on the risk of mongolism and leukaemia. J Natl Cancer Inst 37:687–698. Stankovic T, Weber P, Stewart C, et al. 1999. Inactivation of ataxia telangiectasia mutated gene in B-cell chronic lymphocytic leukaemia. Lancet 353:26–29. Stayner L, Steenland K, Greife A, et al. 1993. Exposure-response analysis of cancer mortality in a cohort of workers exposed to ethylene oxide. Am J Epidemiol 138:787–798. Stebbings JH, Lucas HF, Stenney AF. 1984. Mortality from cancers of major sites in female radium dial workers. Am J Ind Med 5:435–459. Steenland K, Stayner L, Griefe A, et al. 1991. Mortality among workers exposed to ethylene oxide. N Engl J Med 324:1402–1407. Steinbuch M, Weinberg CR, Buckley JD, et al. 1999. Indoor residential radon exposure and risk of childhood acute myeloid leukaemia. Br J Cancer 81:900–906. Stewart A, Webb J, Hewitt D. 1958. A survey of childhood malignancies. Br Med J i:1495–1508. Stiller CA, Boyle PJ. 1996. Effect of population mixing and socioeconomic status in England and Wales, 1979–85, on lymphoblastic leukaemia in children. BMJ 313:1297–1300. Strick R, Strissel PL, Borgers S, et al. 2000. Dietary bioflavonoids induce cleavage in the MLL gene and may contribute to infant leukemia. Proc Natl Acad Sci USA 97:4790–4795. Strange RC, Fryer AA. 1999. The glutathione S-transferases: influence of polymorphism on cancer susceptibility. IARC Scientific Publication 148. Lyon, France: International Agency for Research on Cancer, pp. 231–249. Swensen AR, Ross JA, Shu Xo, et al. 2001. Pet ownership and childhood acute leukemia (USA and Canada). Cancer Causes Control 12:301–303. Swerdlow AJ, Higgins CD, Hunt BJ, et al. 2000. Risk of lymphoid neoplasia after cardiothoracic transplantation a cohort study of the relation to Epstein-Barr virus. Transplantation 69:897–904. Taccioli GE, Rathbun G, Oltz E, et al. 1993. Impairment of V (D) J recombination in double-strand break repair mutants. Science 260:207–210. Tajima K. 1990. The 4th nation-wide study of adult T-cell leukemia/lymphoma (ATL) in Japan: Estimates of risk of ATL and its geographical and clinical features. The T- and B-cell Malignancy Study Group. Int J Cancer 45:237–243. Takata M, Sasaki MS, Sonoda E, et al. 1998. Homologous recombination and non-homologous end-joining pathways of DNA double-strand break repair have overlapping roles in the maintenance of chromosomal integrity in vertebrate cells. EMBO J 17:5497–5508. Tamary H, Bar-Yam R, Zemach M, et al. 2002. The molecular biology of Fanconi anemia. Isr Med Assoc J 4:819–823. Taylor GM, Birch JM. 1996. The hereditary basis of human leukemia. In: Henderson ES, Lister TA, Greaves MJ, eds. Leukemia. 6th ed. Philadelphia: WB Saunders, pp. 210–245. Teta MJ, Benson LO, Vitale JN. 1993. Mortality study of ethylene oxide workers in chemical manufacturing: A 10-year update. Br J Ind Med 50:704–709. Teta MJ, Walrath J, Meigs WJ, et al. 1984. Cancer incidence among cosmetologists. J Natl Cancer Inst 72:1051–1057. Theiss AM, Frentzel-Beyme R, Link R. 1981. Mortality study on employees exposed to aklylene oxides (ethylene oxide/propylene oxide) and their derivatives. In: Prevention of Occupational Cancer-International Symposium. Occupational and Health Series No. 46. Geneva: International Labour Office, pp. 249–259. Theriault G, Goldberg M, Miller AB, et al. 1994. Cancer risks associated with occupational exposure to magnetic fields among electric utility workers

in Ontario and Quebec, Canada, and France: 1970–89. Am J Epidemiol 139:550–572. Thompson JR, Gerald PF, Willoughby ML, et al. 2001. Maternal folate supplementation in pregnancy and protection against acute lympoblastic leukaemia in childhood: A case-control study. Lancet 358:1935–1940. Thun MJ, Altekruse SF, Namboordiri MM, et al. 1994. Hair dye use and risk of fatal cancers in U.S. women. J Natl Cancer Inst 86:210–215. Till M, Rapson N, Smith PG. 1979. Family studies in acute leukaemia in childhood: A possible association with autoimmune disease. Br J Cancer 40:62–71. Tomas JF, Fernandez-Ranada JM. 1996. About the increased frequency of acute promyelocytic leukemia among Latinos: The experience from a center in Spain. Blood 88:2357–2358. Townsend MJ, McKenzie AN. 2000. Unravelling the net? cytokines and diseases. J Cell Sci 113 (Pt 20):3549–3550. Traver RD, Horikoshi T, Danenberg KD, et al. 1992. NAD (P) H: quinone oxidoreductase gene expression in human colon carcinoma cells: Characterization of a mutation which modulates DT-diaphorase activity and mitomycin sensitivity. Cancer Res 52:797–802. Travis LB, Andersson M, Gospodarowicz M, et al. 2000. Treatment-associated leukemia following testicular cancer. J Natl Cancer Inst 92:1165–1171. Travis LB, Curtis RE, Stovall M, et al. 1994. Risk of leukemia following treatment for non-Hodgkin’s lymphoma. J Natl Cancer Inst 86: 1450–1457. Travis LB, Holowaty EJ, Bergfeldt K, et al. 1999. Risk of leukaemia after platinum-based chemotherapy for ovarian cancer. N Engl J Med 340: 351–357. Travis LB, Weeks J, Curtis RE, et al. 1996. Leukemia following low-dose totalbody irradiation and chemotherapy for non-Hodgkin’s lymphoma. J Clin Oncol 14:565–571. Tsabouri SE, Georgiou I, Alamanos I, et al. 2000. Increased prevalence of GSTM (1) null genotype in patients with myelodysplastic syndrome: A case-control study. Acta Haematol 104:169–173. Tucker JD, Eastmond DA, Littlefield LG. 1997. Cytogenetic end-points as biological dosimeters and predictors of risk in epidemiological studies. IARC Sci Publ 142:185–200. Ubbink JB, Fehily AM, Pickering J, et al. 1998. Homocysteine and ischaemic heart disease in the Caerphilly cohort. Atherosclerosis 140:349–356. UK Childhood Cancer Study Investigators. 2001. Breastfeeding and childhood cancer. Br J Cancer 85:1685–1694. UK Childhood Cancer Study Investigators. 1999. Exposure to power-frequency magnetic fields and the risk of childhood cancer. Lancet 354:1925– 1931. UK Childhood Cancer Study Investigators. 2002. the United Kingdom Childhood Cancer Study of exposure to domestic sources of ionising radiation: 1: Radon gas. Br J Cancer 86:1721–1726. Ullrich A, Fischbach W, Blettner M. 2002. Incidence of gastric B-cell lymphomas: A population-based study in Germany. Ann Oncol 13:1120– 1127. UNSCEAR (United Nations Scientific Committee on the Effects of Atomic Radiation). 1988. Sources, Effects and Risks of Ionizing Radiation. 1988 Report to the General Assembly with Annexes. New York: United Nations, pp. 24–374. UNSCEAR (United Nations Scientific Committee on the Effects of Atomic Radiation). 1994. Sources and effects of ionizing radiation. United Nations Publ No E. 94. IX. 11. Annex A. Epidemiological studies of radiation carcinogenesis. New York: United Nations Scientific Committee on the Effects of Atomic Radiation, pp. 11–183. UNSCEAR (United Nations Scientific Committee on the Effects of Atomic Radiation). 2000. Sources and effects of ionizing radiation. United Nations Publ No E. 00. IX. 3. Annex I. Epidemiological evaluation of radiation-induced cancer. New York: United Nations Scientific Committee on the Effects of Atomic Radiation, pp. 297–431. Utterback DF, Rinsky RA. 1995. Benzene exposure assessment in rubber hydrochloride workers: A critical evaluation of previous estimates. Am J Ind Med 27:661–676. Van der Put NM, Steegers-Theunissen RP, Frosst P, et al. 1995. Mutated methylenetetrahydrofolate reductase as a risk factor for spina bifida. Lancet 346:1070–1071. Van Duijn SM, Van Steensel-Moll HA, Coebergh JW, et al. 1994. Risk factors for childhood acute non-lymphocytic leukemia: An association with maternal alcohol consumption during pregnancy? Cancer Epidemiol Biomarkers Prev 3:457–460. Van Krieken JH, Hoeve MA. 2000. Epidemiological and prognostic aspects of gastric MALT-lymphoma. Recent Results Cancer Res 156:3–8. Van Leeuwen FE, Chorus AMJ, van den Belt-Dusebout AW, et al. 1994. Leukemia risk following Hodgkin’s disease: Relation to cumulative dose of alkylating agents, treatment with teniposide combinations, number of

The Leukemias episodes of chemotherapy, and bone marrow damage. J Clin Oncol 12:1063–1073. Van Steensel-Moll HA, Valkenburg HA, van Zanen GE, et al. 1985a. Childhood leukemia and parental occupation. A register-based case-control study. Am J Epidemiol 121:216–224. Van Steensel-Moll HA, Valkenburg HA, Vandenbroucke JP, et al. 1985b. Are maternal fertility problems related to childhood leukaemia? Int J Epidemiol 14:555–559. Vanasse GJ, Concannon P, Willerford DM. 1999. Regulated genomic instability and neoplasia in the lymphoid lineage. Blood 94:3997–4010. Vardiman JW, Harris NL, Brunning RD. 2002. The World Health Organization (WHO) classification of the myeloid neoplasms. Blood 100:2292– 2302. Vianna NJ, Kovasznay B, Polan A, et al. 1984. Infant leukemia and paternal exposure to motor vehicle exhaust fumes. J Occup Med 26:679– 682. Vidabaek A. 1947. Familial leukemia. A preliminary report. Acta Med Scand 127:26–52. Visfeldt J, Anderson M. 1995. Pathoanatomical aspects of malignant haematological disorders among Danish patients exposed to thorium dioxide. APMIS 103:29–36. Waller CF, Fetscher S, Lange W. 1999. Treatment-related chronic myelogenous leukemia. Ann Hematol 78:341–354. Walrath J, Fraumeni JF Jr. 1983. Mortality patterns among embalmers. Int J Cancer 31:407–411. Walrath J, Fraumeni JF Jr. 1984. Cancer and other causes of death among embalmers. Cancer Res 44:4638–4641. Wang D, Kanuma T, Mizunuma H, et al. 2000. Analysis of specific gene mutations in the transforming growth factor-beta signal transduction pathway in human ovarian cancer. Cancer Res 60:4507–4512. Wang J-X, Inskip PD, Boice JD Jr, et al. 1990. Cancer incidence among medical diagnostic x-ray workers in China, 1950 to 1985. Int J Cancer 45: 889–895. Ward E, Hornung R, Morris J, et al. 1996. Risk of low red or white blood cell count related to estimated benzene exposure in a rubberworker cohort (1940–1975). Am J Ind Med 29:247–257. Warzocha K, Ribeiro P, Bienvenu J, et al. 1998. Genetic polymorphisms in the tumor necrosis factor locus influence non-Hodgkin’s lymphoma outcome. Blood 91:3574–3581. Wechsler J, Greene ME, McDevitt MA, et al. 2002. Acquired mutations in GATA1 in the megakaryoblastic leukemia of Down syndrome. Nat Genet 32:148–152. Weisberg I, Tran P, Christensen B, et al. 1998. A second genetic polymorphism in methylenetetrahydrofolate reductase (MTHFR) associated with decreased enzyme activity. Mol Genet Metab 64:169–172. Weiss HA, Darby SC, Fearn T, et al. 1995. Leukemia mortality after X-ray treatment for ankylosing spondylitis. Radiat Res 142:1–11. Wen W, Shu XO, Linet MS, et al. 2000. Allergic disorders and the risk of childhood acute lymphoblastic leukemia (United States). Cancer Causes Control 11:303–307. Wen W, Shu XO, Potter JD, et al. 2002. Parental medication use and risk of childhood acute lymphoblastic leukemia. Cancer 95:1786–1794. Wertheimer N, Leeper E. 1979. Electrical wiring configurations and childhood cancer. Am J Epidemiol 109:273–284. Westergaard T, Andersen PK, Pedersen JB, et al. 1997. Birth characteristics, sibling patterns, and acute leukemia risk in childhood: A population-based cohort study. J Natl Cancer Inst 89:939–947. Whiteside D, McLeod R, Graham G, et al. 2002. A homozygous germ-line mutation in the human MSH2 gene predisposes to hematological malignancy and multiple café-au-lait spots. Cancer Res 62:359–362. Wiemels JL, Cazzaniga G, Daniotti M, et al. 1999b. Prenatal origin of acute lymphoblastic leukaemia in children. Lancet 354:1499–1503. Wiemels JL, Ford AM, Van Wering ER, et al. 1999c. Protracted and variable latency of acute lymphoblastic leukemia after TEL-AML1 gene fusion in utero. Blood 94:1057–1062. Wiemels JL, Leonard BC, Wang Y, et al. 2002b. Site-specific translocation and evidence of postnatal origin of the t(1;19) E2A-PBX1 fusion in childhood acute lymphoblastic leukaemia. Proc Natl Acad Sci USA 99:15101– 15106. Wiemels JL, Pagnamenta A, Taylor GM, et al. 1999a. A lack of functional NAD(P)H:quinone oxidoreductase allele is selectively with pediatric leukemias that have MLL fusions. United Kingdom Childhood Cancer Study Investigators. Cancer Res 59:4095–4099. Wiemels JL, Xiao Z, Buffler PA, et al. 2002a. In utero origin of t(8;21) AML1ETO translocations in childhood acute myeloid leukaemia. Blood 99:3801–3805.

871

Wiklund K, Holm L-H. 1986. Trends in cancer risks among Swedish agricultural workers. J Natl Cancer Inst 77:657–664. Wilcosky TC, Checkoway H, Marshall EG, et al. 1984. Cancer mortality and solvent exposures in the rubber industry. Am Ind Hyg Assoc J 45: 809–811. Williams RR, Horm JW. 1977. Association of cancer sites with tobacco and alcohol consumption and socioeconomic status of patients: Interview study from the Third National Cancer Survey. J Natl Cancer Inst 58:525–547. Willis TG, Dyer MJ. 2000. The role of immunoglobulin translocations in the pathogenesis of B-cell malignancies. Blood 96:808–822. Willman CL. 1999. Molecular evaluation of acute myeloid leukemias. Semin Hematol 36:390–400. Wolf CR, Smith G, Gough AC, et al. 1992. Relationship between the debrisoquine hydroxylase polymorphism and cancer susceptibility. Carcinogenesis 13:1935–1938. Wong O. 1987. An industry-wide mortality study of chemical workers occupationally exposed to benzene. II. Dose response analyses. Br J Ind Med 44:382–395. Wong O. 1990. A cohort mortality study and a case-control study of workers potentially exposed to styrene in the reinforced plastics and composites industry. Br J Ind Med 47:753–762. Wongsrichanalai C, Delzell E, Cole P. 1989. Mortality from leukemia and other diseases among workers at a petroleum refinery. J Occup Med 31:106–111. Woo MH, Shuster JJ, Chen C, et al. 2000. Glutathione S-transferase genotypes in children who develop treatment-related acute myeloid malignancies. Leukemia 14:232–237. Worrilow LJ, Travis LB, Smith AG, et al. 2003. An intron splice acceptor polymorphism in hMSH2 and risk of leukemia after treatment with chemotherapeutic alkylating agnets. Clin Cancer Res 9:3012–3020. Xu Y. 1999. ATM in lymphoid development and tumorigenesis. Adv Immunol 72:179–189. Yin S-N, Li G-L, Tain F-D, et al. 1987. Leukaemia in benzene workers: A retrospective cohort study. Br J Ind Med 44:124–128. Yin S-N, Li G-L, Tain F-D, et al. 1989. A retrospective cohort study of leukemia and other cancers in benzene workers. Environ Health Perspect 82:207–213. Yin S-N, Linet MS, Hayes RB, et al. 1994. Cohort study among workers exposed to benzene in China: I. General methods and resources. Am J Ind Med 26:383–400. Yoshida K, Inoue T, Nojima K, et al. 1997. Calorie restriction reduces the incidence of myeloid leukemia induced by a single whole-body radiation in C3H/He mice. Proc Natl Acad Sci USA 94:2615–2619. Yoshinaga S, Aoyama T, Yoshimoto Y, et al. 1999. Cancer mortality among radiological technologists in Japan: Updated analysis of follow-up data from 1969 to 1993. J Epidemiol 9:61–72. Youn J, Hwang SH, Cho CS, et al. 2000. Association of the interleukin-4 receptor alpha variant Q576R with Th1/Th2 imbalance in connective tissue disease. Immunogenetics 51:743–746. Yuille MR, Houlston RS, Catovsky D. 1998. Anticipation in familial chronic lymphocytic leukaemia. Leukemia 12:1696–1698. Zahm SH, Ward MH, Blair A. 1997. Pesticides and cancer. Occup Med 12:269–289. Zahm SH, Weisenburger DD, Babbitt PA, et al. 1992. Use of hair coloring products and the risk of lymphoma, multiple myeloma, and chronic lymphocytic leukemia. Am J Pub Health 82:990–997. Zhang L, Eastmond DA, Smith MT. 2002. The nature of chromosomal aberrations detected in humans exposed to benezene. Crit Rev Toxicol 32:1–42. Zhang L, Rothman N, Wang Y, et al. 1998. Increased aneusomy and long arm deletion of chromosomes 5 and 7 in the lymphocytes of Chinese workers exposed to benzene. Carcinogenesis 19:1955–1961. Zheng W, Linet MS, Shu X-O, et al. 1993. Prior medical conditions and the risk of adult leukemia in Shanghai, People’s Republic of China. Cancer Causes Control 4:361–368. Zhu C, Bogue MA, Lim DS, et al. 1996. Ku86-deficient mice exhibit severe combined immunodeficiency and defective processing of V(D)J recombination intermediates. Cell 86:379–389. Zhu YM, Das-Gupta EP, Russell NH. 1999. Microsatellite instability and p53 mutations are associated with abnormal expression of the MSH2 gene in adult acute leukemia. Blood 94:733–740. Zoloth SR, Michaels DM, Villalbi JR, et al. 1986. Patterns of mortality among commercial pressmen. J Natl Cancer Inst 76:1947–1951. Zuelzer WW, Cox DE. 1969. Genetic aspects of leukemia. Semin Hematol 6:228–249.

45

Hodgkin Lymphoma NANCY E. MUELLER AND SEYMOUR GRUFFERMAN

H

odgkin lymphoma (HL) is a malignant disease involving the lymph nodes, spleen, and other lymphoid tissue. In many ways a curious and fascinating disease, HL has stimulated the interest of epidemiologists, basic scientists, and clinicians alike. This interest reflects its unusual biology, epidemiology, and the remarkable success that has been achieved in treatment. Its epidemiology and pathology have long suggested that it has an infectious etiology, and it is now established that the Epstein Barr virus (EBV) is a likely cause of some—but not all—HL cases. With the tools of molecular biology now in hand, along with a more precise pathologic definition of the disease, transdisciplinary teams of epidemiologists, virologists, immunologists, and other relevant scientists have the opportunity to finally unravel the web of causation of HL. Hodgkin lymphoma is distinguished by its singular pathology and its unusual epidemiology. In affected lymph nodes, the malignant Reed-Sternberg cells (RSC) typically account for only 1%–2% of the involved tissue. These giant multinucleated cells are surrounded by a multitude of activated immune cells. The micro-environment of the RSC is laced with a mixture of B-cell stimulator and immune reactive and suppressive factors (Kadin and Liebowitz, 1999; Weiss et al., 1999; Marshall et al., 2004; Poppema, 2005), suggestive of a chronic inflammatory process. The origin of the RSC in classical HL is now recognized as that of a crippled germinal center B-cell (Kuppers et al., 1999). This origin is deduced from the presence of rearranged Ig heavy and light chains, although the normal B-cell surface markers are usually not expressed. This places the genesis of HL in the “high risk” environment of B-stimulatory molecule-induced DNA modification that occurs during the generation of antibody specificity (MartinezMaza and Breen, 2002). Epidemiologically, HL has a unique bimodal (sometimes trimodal) age incidence curve, the magnitude of which and its relation to age varies by socioeconomic conditions. In economically advantaged populations the initial peak occurs in young adulthood, whereas in disadvantaged populations the first peak is evident in early childhood, especially among boys (Correa and O’Conor, 1971).

HISTORICAL BACKGROUND The history of HL is extremely rich and colorful. Its early history is comprehensively reviewed in a two-part paper by Hoster and Dratman (1948) who reviewed 572 papers published from the late 19th to mid20th century. The disease was named after Thomas Hodgkin, a leading physician of his time in England and an extremely brilliant and colorful character (Hellman, 1999). Hodgkin was a devout Quaker (when it was quite unfashionable to be one) and was a social activist on behalf of the underprivileged. He particularly was concerned with the “Aborigines and Jews” and was a good friend of Sir Moses Montefiore (Kass, 1984). It was on a trip with Montefiore to what is now Israel that Hodgkin died and where he was buried in Jaffa. His very colorful and productive life has been chronicled in several biographies (Aterman, 1986; Kass and Kass, 1988; Rosenfeld, 1993). Although Hodgkin made many important medical contributions, such as being the first to describe aortic insufficiency and to introduce the stethoscope to England, he is best known for his classic paper in 1832, which described the gross pathology of a series of cases of what appeared to be ‘morbid’ enlargement of the lymph nodes and spleen (Hodgkin,

872

1832). Several of these original cases proved later not to be HL (Kaplan, 1980). Another British physician, Wilkes, later assigned Hodgkin’s name to the lymphoma in 1865. Although the gross pathologic description of the disease proved sufficient to generate a body of knowledge on the natural history of HL, it was not until the late 19th century when scientists were able to use microscopic examination of tumor specimens to more precisely define the disease. It is now more than 100 years since the first definitive descriptions of the characteristic tumor cells (RSC) were observed in HL by Carl Sternberg (1898) in Germany, and Dorothy Reed (1902) in the United States. The RSC is central to a pathologic diagnosis of HL today. Hodgkin lymphoma has a long history of being suspected of having an infectious etiology. Early in the 20th century, a number of leading American physicians published a curious body of research papers on suspected bacterial causes of HL (Hoster and Dratman, 1948). Leading journals reported papers on such agents as Corynybacterium granulomatis maligni and Bacillus hodgkini (Yates and Bunting, 1915; Cunningham, 1917). Although such reports may now seem naïve, these leading investigators of their time were most likely led in this direction of inquiry by the unusual clinical picture of the disease. The involvement of lymph nodes, the frequent occurrence of fever often in recurrent cycles (Pel-Ebstein fever), and night sweats in patients suggested that it might be an infectious disease. There were other unusual features of the disease, such as pain at the site of disease involvement after alcohol was ingested, even though lymph nodes contain no pain receptors, which piqued the curiosity of early investigators (Callahan et al., 1994). With hindsight, it can now be seen that the instincts of the early students of HL were correct, but that they lacked the technology to search for a viral etiologic agent. Another source of interest in the disease stems from the work of MacMahon (1966) on the fascinating descriptive epidemiology of HL. MacMahon noted the bimodal age-incidence pattern for the disease and found very different characteristics among cases in the two age peaks, which suggested to him that they may represent two different diseases. He hypothesized that the young-adult-onset disease may be an infectious disease whereas the old-adult disease was likely to be a malignant neoplasm. His interest in the disease proved “infectious” and several of his students (including the co-authors of this chapter) went on to research careers focused on HL. The epidemiologic features of the disease are analogous to paralytic poliomyelitis, particularly the young-adult-onset disease (Newell, 1970; Abramson, 1974; Gutensohn and Cole, 1977). More specifically, we suggested that late age of first infection with a presumed common infectious agent was a cause of young-adult disease (Gutensohn and Cole, 1980, 1981). This hypothesis was based on the observations of Newell in 1970 and Abramson in 1974 of the similarity between the geographical distributions of HL with that of a highly prevalent infection for which early infection was protective and generally mild, much like poliomyelitis. When infections that are normally encountered in childhood are instead experienced in adulthood, they may be clinically more severe. For a herpesvirus infection, such as the EBV, this may result in less efficient immunologic control of the subsequent viral latency. Many analytic epidemiologic studies have supported this notion, including Paffenbarger et al. (1977) and Alexander et al. (1991b), as discussed below. Interest in the possibly infectious etiology of HL was further stimulated by a 1973 report of possible person-to-person transmission of

Hodgkin Lymphoma

873

the disease in high schools (Vianna and Polan, 1973). Although this report could not be confirmed, it led to a surge of research on the suspected infectious nature of the disease (Smith et al., 1977; Grufferman et al., 1979). More recently, the search for an infectious etiology of HL received added impetus from the development of molecular techniques capable of identifying the presence of the EBV in HL tumor specimens (Weiss et al., 1987). Before this, it had been observed that persons with infectious mononucleosis (IM), which is caused by the EBV, had a threefold increased risk of developing HL (Grufferman, 1982). Subsequently, it was found that patients with HL more frequently had elevated serum EBV antibody titers than did controls, years in advance of the diagnosis of the disease (Evans and Comstock, 1981; Mueller et al., 1989). This evidence, plus the numerous studies documenting the presence of EBV genetic material (which has been shown to be clonal) in the malignant cells of HL tumors in 25%–50% of cases, makes it highly likely that the EBV is a causal factor in a substantial proportion of HL cases (Mueller and Grufferman, 1999). In fact, the International Agency for Research on Cancer recently classified the EBV as a Class I human carcinogen (i.e., “EBV is carcinogenic to humans”) for HL (IARC, 1997). It is truly remarkable that the early clinical investigators of HL (almost 100 years ago) have been proven prescient in their pursuit of infectious etiologic agents. In addition, attention also has been focused on this relatively uncommon cancer due to the evolution of highly successful therapies. HL was largely incurable before the midpoint of the 20th century, and today it is highly curable with 5-year survival rates in excess of 90% for localized disease. At mid-20th century the introduction of highdose, extended field radiation therapy was found to result in permanent disease remission. The history of radiation therapy of the disease is well documented in the monograph by Kaplan (1980), who was one of the important contributors to the development of this treatment approach. Survival from HL was further advanced by the use of multiagent chemotherapy (MOPP) in its treatment, first introduced in the 1960s (De Vita and Bonadonna, 1999). The treatment successes unfortunately have come with a price. Second malignant neoplasms are now the leading cause of death in long-term HL survivors, and these appear to be the results of therapy—following both radiation therapy and chemotherapy (Sankila et al., 1996; Hudson et al., 1998; Travis et al., 2003). The current challenge in treatment of the disease is finding minimally harmful, yet effective, treatment approaches.

(It should be recognized that newer treatments are less harmful.) Among all childhood cancers, the risk of second cancers is greatest for HL survivors. In the Childhood Cancer Survivor Study cohort of 5-year survivors of childhood HL, an observed/expected ratio of 9.7 (95% confidence interval: 8.1–11.6) was found for second malignant neoplasms, with a cumulative incidence at 20 years of 7.6%. The standardized mortality ratio for death from secondary cancers in this cohort was 24.0 (19.2–29.7) (Neglia et al., 2001; Mertens et al., 2001). Ng et al. (2002) followed 1319 patients of all ages with HL for a median of 12 years. They found a relative risk (RR) of second malignancy of 4.6 and an absolute excess risk of 89.3/100 000 person-years. The risk was greater in HL patients receiving chemotherapy and radiation (RR = 6.1) than with radiation alone (RR = 4.0). After 15 and 20 years there were 2.3% and 4.0% excess risks respectively of second cancers per person/year. Dores et al. (2002) followed 32, 591 HL patients, of whom 1111 were 25-year survivors from 16 populationbased cancer registries in North America and Europe. Overall, they found an observed to expected ratio of 2.3 (2.2–2.4) for second cancers. The actuarial risk of developing a “solid tumor” after 25 years was 21.9%. The RR of solid neoplasms generally decreased with increasing age at diagnosis of HL. HL survivors also have increased mortality from heart disease and infection (Hudson et al., 1998). Thus, although most patients are cured of their HL, there is a continuing burden on their health. This underlines the public health significance of this disease and the importance of research that leads to understanding of its pathogenesis and the possibility of prevention. In summary, the immunologic and histopathologic features of HL and its origin in the “risky” environment of the germinal center argue that its origin involves an infectious agent. The descriptive epidemiology suggests that HL is a disease complex that results from two or more age-dependent etiologic pathways that are influenced by social and economic environment (MacMahon, 1957). It appears that these associations with socioeconomic conditions reflect their influence on age of childhood infections and perhaps immune competency. Clonal EBV genome is present in a substantial proportion (30%–40%) of HL cases, providing strong evidence of a causal role of the virus in these cases (IARC, 1997). However, whether other infectious agents or factors play a parallel role in cases lacking evidence of EBVpositivity is unknown. The theme of this chapter is to evaluate the evidence concerning factors that influence the risk of HL, in light of its pathogenesis within the environment of the germinal center.

INCIDENCE AND MORTALITY

CLASSIFICATION

It is currently estimated that about 7350 (54% males) cases of HL are diagnosed with 1400 (55% males) deaths each year in the United States (American Cancer Society, 2005). The lifetime risk in the United States of developing HL is about 1 in 455 and that of dying, 1 in 2000 (Ries et al., 2002). In comparison to non-Hodgkin lymphoma (NHL), for each case of HL, almost eight cases of NHL are diagnosed, and for each death from HL, 14 cases die from NHL (American Cancer Society, 2005). Because nearly two-thirds of HL cases in the United States are diagnosed before age 45 years, and the great majority of patients will survive their disease, the social impact of the disease is greater than that for many other malignancies. Important concerns among survivors include the risk of secondary cancers, cardiac damage, impaired fertility, and diminished general health. In terms of second malignancies, HL survivors suffer a variety of other cancers ranging from leukemia and NHL to breast and thyroid cancer (Sankila et al., 1996; Dores et al., 2002; Travis et al., 2003). Increased adverse late effects occur in HL patients treated with chemotherapy alone, radiation therapy alone, or combined modality therapy. Many of the second malignant neoplasms in HL occur in the paths of radiation aimed at mediastinal and other involved lymph nodes. The magnitude of the increased risk of second malignant neoplasms in HL survivors varies with age at diagnosis, type of therapy, and duration of survival. Most of the observed increased risks are quite large.

Pathologic diagnosis of HL is based on the presence in tumor specimens of an abnormal, very large, multilobed and often multinucleated cell, the RSC (or its mononucleated variants termed “Hodgkin cells”) in a characteristic cellular milieu (Weiss et al., 1999). The majority of epidemiologic research on HL has treated the disease as if it were a unified diagnostic entity classified under the ICD code number 201. However, an important question is whether HL is a single disease or a group of related diseases. Historically, the disease was subcategorized by pathologic examination of fixed biopsy specimens under the light microscope. Early investigators such as Jackson and Parker (1944a,b) attempted to correlate histologic patterns of tumor specimens with the course and prognosis of the disease. Increasingly sophisticated subclassifications of HL have evolved, which still rely primarily on histopathology, but with consideration of immunophenotyping or other molecular biologic characterization; most notably, the REAL classification scheme (Harris et al., 1994). The latest WHO classification of HL has two broad categories— “nodular lymphocyte predominance (NLP) Hodgkin lymphoma” and “classical Hodgkin lymphoma” (Anagnostopoulos et al., 2000; Jaffe et al., 2001) (Table 45–1). The former category is separated out from lymphocyte predominance (LP) because of its distinct clinical features and histology. The “classical” category encompasses the previous nosologic subcategories from the Rye classification, of nodular sclerosis (NS), mixed cellularity (MC), and lymphocyte depletion (LD), plus a new category of lymphocyte-rich HL (Lukes and Butler, 1966).

874

PART IV: CANCER BY TISSUE OF ORIGIN

Table 45–1. The WHO Classification of Hodgkin Lymphoma Nodular lymphocyte predominance Hodgkin lymphoma Classical Hodgkin lymphoma Nodular sclerosis Hodgkin lymphoma Lymphocyte depletion Hodgkin lymphoma Mixed cellularity Hodgkin lymphoma Lymphocyte-rich Hodgkin lymphoma

to avoid misclassification. However, because LP represents only about 10% of HL cases, earlier research that included all types of HL was probably only minimally affected by potential disease misclassification. Subclassification of HL based on presence or absence of EBV in tumors will add to the cost of research because EBV assays are not currently done as part of routine medical care.

Source: Jaffe et al., 2001.

The lymphocyte-rich category includes the residual portion of the LP category. In subdividing the Rye classification’s LP category, immunohistologic studies are used for differentiating the two new categories. In the new diagnostic category of NLP-HL, the diagnostic cells are typically multilobated variants of the RSC that are mononuclear blasts and have been termed “L & H cells” due to the predominant cellular background of lymphocytes and histiocytes (Stein et al., 1999). It is generally accepted that the L & H cells are all of B-cell origin and do not carry the EBV genome. The NLP variant of HL is thought to arise from germinal center B cells at the centroblastic stage of development (Jaffe et al., 2001). In contrast in classical HL, the great majority of Hodgkin and RSC cells are B-cell type; a major subset of cases are EBV-positive and are recognized as being a crippled germinal center B-cell, although a very small proportion are T-cell in origin (Stein et al., 1999; Kuppers et al., 1999). Although these various histologic categories correlate with clinical course and prognosis, they do not necessarily determine the choice of therapy. In classical HL, clinical stage (i.e., the extent of involvement of different body regions) plays a greater role in choice of therapy than does histology. This is in contrast with the NHL where, increasingly, therapeutic choices are tailored to the immunologically classified lymphoma types. The important concern these findings raise is that on clinical grounds, classical HL appears to be a relatively uniform disease that responds to similar treatment regimens, albeit in different dosages and/or combinations. However, on epidemiologic grounds, HL appears to be more heterogeneous. For example, the histology of the disease varies by age, and the EBV is found with varying frequencies in different histologic subcategories. Yet, there is still overlap between categories. Not all MC-HL tumors are found to contain the EBV, and some NS-HL tumors contain the EBV. If, in epidemiologic research, all HL cases are considered to be a single entity, the question arises of whether research findings are hampered by the possible misclassification of different diseases into a single category, likely bringing estimates of associations towards the null value. Several recent studies have attempted to assess various HL risk factors in analyses stratified by histologic subtype of the disease. A newer axis of subclassification of the disease is by the presence or absence of EBV in patients’ tumors. In this sense the EBV serves as a “biomarker” of the EBV itself or other distinct causal factors. A general problem posed by these approaches is that because the disease is relatively uncommon, it is difficult to accrue sufficient numbers of cases in each subcategory to have adequate statistical power for analyses. In general, it does not appear that there are major differences in risk factors among the various (old) Rye classification subcategories of HL per se, once age is taken into account, and a great deal of prior research relied on that nosology. Some promising new research results suggest that risk factors and epidemiologic features are different for those subjects with EBV-positive and EBV-negative tumors. Studies of EBV in tumors are more “biologically” based in that they measure a potential cause of the disease directly in malignant tissues. It is perhaps on the basis of these differences that attempts to define different etiologic entities appear to be more successful for EBV categorization of tumors than for pathologic classifications (Jarrett et al., 2003). What does all this mean for future epidemiologic research on HL? First, diagnoses of LP subtypes of the disease will need to be further subdivided. This will require either immunophenotyping or careful pathologic review. In analyzing results of future studies of HL, those cases with NLP should either be excluded or analyzed separately

DEMOGRAPHIC PATTERNS Incidence and Mortality in the United States The 1999 age-adjusted incidence rate of HL in the United States was 2.8 per 100,000 person-years and that for mortality, 0.5 per 100,000 person-years, based on the SEER program (Ries et al., 2002). In 1998, the estimated 5-year relative survival of HL cases was 84%. The incidence rates for males and females in 1999 were 3.0 and 2.6 per 100,000 person-years, respectively. The estimated 5-year relative survival for 1992–1998 is higher for women (86%) compared with men (82%), mirroring the fact that women more commonly present with the NS subtype of HL and a more favorable stage. The incidence rates for whites in 1999 for males and females were 3.1 and 2.9 per 100,000 person-years, respectively. Those for blacks are now quite similar: 3.1 and 1.9 per 100,000 person-years, respectively. The incidence of HL among Hispanics is more similar to that of United States blacks than whites (Miller et al., 1996). The current lifetime risk of developing HL for whites is 0.25% (1 in 400) and 0.23% (1 in 435) for males and females, respectively. That for blacks is 0.17% (1 in 588) and 0.15% (1 in 667) for males and females, respectively. The relative differences between blacks and whites in their cumulative risk of developing HL are greater than for the age-adjusted rates because of their differences in life expectancy. The probability of survival among HL cases indicates disparities between whites and blacks in the United States. The estimated 5-year relative survival in 1992–1998 for whites is 83% for males and 87% for females, with that for blacks being 74% and 80%, respectively. In terms of geographic variation within the United States, HL mortality rates are generally higher in the eastern two-thirds of the country, particularly in the northeast, with an apparent north-south gradient among young adults (Devesa et al., 1999). Between 1973 and 1994, the incidence of HL decreased for both United States blacks (-2.9%) and whites (-13.1%). Most of this decrease occurred among the population aged 65 years and older—a total of 37.2% (Ries et al., 1997). Glaser and Swartz (1990) analyzed the national data from the Repository Center for Lymphoma Clinical Studies for the period 1969–1980, with correction for diagnostic error based on time, age, and histology specific confirmation rates. Upon adjustment, they found that the incidence rates for older adults were lower than previously observed and showed no secular trend. Further, they found a slight increase for NS-HL among young adults. An analysis of time trends and age-period-cohort patterns for the incidence in HL in Connecticut between 1935 and 1992 concluded that the incidence has increased among young adults aged 20–44 years. This increase was greater for women and primarily associated with the NS subtype (Chen et al., 1997). Since the mid 1990s, the incidence of HL has slightly decreased among males, and slightly increased among females (Ries et al., 2002). An interesting feature of the descriptive epidemiology of HL is the variation in gender ratio by age. As Glaser (1994) has pointed out, there appears to be a deficit of cases among women in their late thirties and forties in recent data from the United States and elsewhere. She proposes that this may represent a protective effect from childbearing, perhaps mediated by estrogens. The greatest male excess in incidence occurs in young children as discussed below.

Economic Development, Westernization, and International Variation Internationally, the overall incidence of HL varies with the level of “westernized” style of economic development. This variation is most

875

Hodgkin Lymphoma

4 –7 70

4 –6 60

50

–5

4

4 –4 40

30

–3

4

4 –2 20

4 –1 10

0–

4

Females

Figure 45–1. Age-specific average annual incidence rates of Hodgkin lymphoma for all races, by gender, per 100,000 population, 1996–2000, in the United States SEER program. (Source: Parkin et al., 2002b.)

evident in the first peak of incidence; that is, in the peak age and magnitude of HL incidence rates during the first five decades of life. The distinguishing epidemiological feature of HL—its bimodal ageincidence curve—is characteristically seen among “developed” or, more precisely, westernized, economically advantaged populations (Fig. 45–1). In such populations there are very few cases occurring among children, a rapid increase of incidence among teenagers that peaks about age 25, and then a decrease to a plateau through middleage, after which rates increase with advancing age to the second peak. MacMahon (1957) proposed that the bimodality results from the overlap of at least two disease distributions with differing age peaks. Specifically, he proposed that among young adults, HL is caused by a biological agent of low infectivity, and that among the elderly, the cause is probably similar to those of the other lymphomas (MacMahon, 1966). In 1971, Correa and O’Conor noted that among economically disadvantaged populations, a different age pattern is evident. In such populations, there is an initial peak in childhood only for boys (Parkin et al., 1988), relatively low rates among young adults, followed by a late peak among those of advanced age. An example of this pattern is apparent in incidence data for Cali, Colombia in the early 1960s (Fig. 45–2). By 1970, the pattern for Cali had evolved from the developing pattern to what Correa and O’Conor termed an “intermediate” pattern. A similar shift is evident in data from Puerto Rico and Bombay for males between 1970 and 1985 (Hartge et al., 1994). A notable exception to these observations has been Japan, where HL had been extremely rare before age 50; however, the disease has become more frequent among young adults in recent decades (Aozasa et al., 1986). It was reported in the seminal paper by Correa and O’Conor (1971) that the relative frequencies of HL incidence in children and young adults varied internationally and was related to a country’s level of development, which generally translates to a country’s level of public hygiene. This intriguing set of findings was reassessed by Macfarlane et al. (1995) using more recent data from the same source. They found that the inverse relationship between childhood and young adult HL was no longer consistently evident. Incidence rates for young adult HL had risen in less developed countries while remaining fairly constant in more developed countries. These newer findings suggest that environmental factors might be changing in Third World countries to becoming more like those in the rest of the world, and that birth cohort effects may modify the age incidence curves for HL. Essentially all of the predominant populations in Europe and North America have a well-defined “developed” pattern of HL incidence at the present time. The height of peak occurrence in young adulthood varies within this set of countries, being high in Canada, the United

10 CONNECTICUT

CALI 8 6 4 2

INCIDENCE PER 100,000

Males

States, Scotland, and Denmark and lower in southern Europe. The pattern within the former Eastern Block countries has been merging into a developed pattern. In contrast, the pattern in Asia and Africa is generally intermediate or developing (Parkin et al., 1992). In Africa, the rates of HL are quite low; however, for those countries with incidence data, there is evidence of the early peak occurring in childhood, typical of developing populations (Parkin et al., 2002a). Representative age incidence curves are shown in Figure 45–3. Within the same population, the age pattern will differ between population subgroups with major differences in socioeconomic level. An example of this was observed in Norway in the early 1960s in relation to urbanization, where the pattern in rural areas was intermediate and in urban areas, developed (Fig. 45–2). Examination of longitudinal data from the Connecticut Cancer Registry reveals the evolution from an intermediate to a typical developed pattern between 1935 and 1980 (Fig. 45–4). Note the suggestion of a childhood peak among boys in the late 1930s. Clemmesen (1981) had noted the existence of a “secondary peak” among middle-aged males in some economically advantaged populations. That secondary peak is apparent in Figure 45–4. The predominant incidence rate pattern among blacks in America was “intermediate-developed” from 1973 to 1981 (Horm et al., 1985), and has moved toward a developed profile (Glaser, 1991), with somewhat lower rates among young adults compared with whites (Ries et al., 2002). As reviewed by Glaser (1990), the secular changes in HL mortality between blacks and whites in America are consistent with socioeconomic conditions. Spitz et al. (1986), who evaluated the ageincidence SEER data for white, black, and Hispanic American children and adolescents, found the data are generally in keeping with the pattern expected by general socioeconomic level. However, Wilkinson et al. (2001) recently noted that the incidence of HL was significantly higher among Florida’s Hispanic children compared with whites. The incidence pattern for Asian Americans as a group appears to be intermediate-developed (Glaser and Hsu, 2002). From social class-specific incidence data, there is evidence that rates of HL are higher among adults from higher social class; however, the data have not always been consistent. When the data are examined separately by age group corresponding to the two major age peaks, a clearer picture emerges. Glaser (1987) evaluated United States

10 08 06 Males Females

04 02

8 6 4 2 10 08 06 04 RURAL

02

0

10

20

URBAN

NORWAY

30

40

50

60

70

10

20

NORWAY

30

40

50

60

70

AGE

Figure 45–2. Age-specific average annual incidence rates per 100,000 population of Hodgkin lymphoma by gender in (a) Cali, Colombia (1962–1966); (b) Connecticut, United States (1960–1962); (c) rural Norway (1964–1966); and (d) urban Norway (1964–1966). (Source: Correa and O’Conor, 1971.)

876

PART IV: CANCER BY TISSUE OF ORIGIN with risk increased if infection is delayed until adolescence or young adulthood or in early childhood with early infectious exposure, primarily among boys living under generally poor conditions. In addition, the pathogenesis of HL may well be modified by the underlying stage of immune competency. The maturation of “type 1” (“Th1”) response—which includes cytotoxic T-cell response—is induced in the newborn by exposure to infectious agents (Christiansen, 2000). There is now a sizable body of evidence that early life exposure to other children—either within the family or in daycare settings—is generally protective against allergy and asthma, which are associated with a predominantly “type 2” (“Th2”) immune response (von Mutius et al., 1994, 1998; Strachan et al., 1996; Kramer et al., 1999; Ball et al., 2000; Infante-Rivard et al., 2001). The general socioeconomic characteristics of young adults who develop HL are consistent with delayed infectious exposure in early childhood, suggesting that their underlying state of immunity may be skewed toward a type 2 response or that they have

70

–7

4

4 –6 60

–5 50

4 –4 40

4

France Czech Republic

4 –3 30

4 –2 20

–1 10

0–

4

4

US (SEER) Japan (Osaka)

14

Females: Cases per 100,000 person-years

incidence data from 1969 to 1980. She found that rates for young adults were correlated with community-level social class indicators. She also reported that the incidence of NS-HL, the most common histologic type diagnosed in young adults, increased with regional social class indices. Henderson et al. (1979) computed histologic-specific incidence rates for HL in Los Angeles County from 1972 to 1975 by social class. They reported that the incidence of NS-HL was directly related to social class, but there was no consistent association for the other histologic types. These data were extended through 1985 by Cozen et al. (1992), who confirmed the initial observation. In addition, they found that the risk pattern for the MC subtype was quite distinct and was negatively associated with social class, suggesting the two histologic subtypes had separate etiologies. Of note, in the SEER data from Los Angeles County, there was a secular increase in the NS form of HL but not in other subtypes between 1971 and 1985 (Cozen et al., 1992). Similarly, Hjalgrim et al. (2001) reported a significant increase in HL among young adults in Denmark, Finland, Norway, and Sweden between 1978 and 1997, which was primarily of the NS-HL subtype. As part of a large population-based leukemia/lymphoma registry covering about half of the United Kingdom over a 5-year period, Alexander and colleagues evaluated the characteristics of over 1800 HL cases by area-based socioeconomic and population density indices (Alexander et al., 1989). They reported that of the 486 cases diagnosed under age 25, the rates were significantly greater in the high socioeconomic areas, and there was a significant trend in increase in areas closer to “built-up areas” with mutual adjustment (Alexander et al., 1991b). They also reported that for ages less than 35 years at diagnosis, there was a significantly positive association for social class; while a negative association was found for cases grouped by ages 35–49 and 50–79, with the trend for the latter being significant (Alexander et al., 1991a). A general observation is that HL cases, occurring in economically developing populations (Mueller, 1987) and among lower social class groups in developed populations (Hu et al., 1988), are predominantly of the MC and LD subtypes. These subtypes are associated with advanced stage of disease and with EBV-positivity. The social class differences that are seen in the histologic presentation of HL may reflect a mix of age-dependent host responses related to social environmental exposures.

Time periods 1935–39

12

1950–59 1970–79

10

8

6

4

2

0 0–

10–

20–

30–

40–

50–

60–

70–

50–

60–

70–

Age Group (Years)

14

Males: Cases per 100,000 person-years

Figure 45–3. Male age-specific average annual incidence rates of Hodgkin lymphoma in various countries, per 100,000 population, 1993–1997. (Source: Parkin et al., 2002b.)

Time periods 1935–39

12

1950–59 1970–79

10

8

6

4

2

0 0–

10–

20–

30–

40–

Age Group (Years)

Summary Taken together, these observations are consistent with the hypothesis that HL may develop as a rare consequence of a common infection,

Figure 45–4. Age-specific incidence rates for Hodgkin lymphoma in the state of Connecticut (US) for three time periods, 1935 through 1979, according to gender. (Source: Cusano and Young, 1986.)

Hodgkin Lymphoma some impairment in type 1 response. Because the immunologic control of the EBV and many other viral agents resides in the cytotoxic T-cell response, such individuals may be at increased risk of oncogenic effects of such infections encountered after childhood. This hypothesis is likely an over-simplification as the roles of innate immunity and T regulatory cells need also to be considered (Matricardi and Yazdanbakhsh, 2003), and both type 1 and type 2 diseases can occur simultaneously (Simpson et al., 2002). However, if generally true, this paradigm provides coherence to the descriptive epidemiology of HL through mid-life. Whether early life events and relative immune competency affect risk of developing HL after age 50 (the second major peak) is unclear.

CHILDHOOD SOCIAL ENVIRONMENT Findings from analytic studies confirm the observations based on the descriptive epidemiology of HL. Factors associated with susceptibility to delayed childhood infections pertain to risk of HL occurring from early childhood through middle age; that is, within the first incidence peak. However, among patients who are diagnosed in their fifties or later, there is no consistent pattern of association with indicators of childhood social class.

Adults In early studies among younger adults (generally defined as those aged from the mid-teens through the thirties), the occurrence of HL was consistently associated with characteristics of social environment that foster escape from EBV and similar infections in childhood (Gutensohn and Cole, 1977). In this age group there was generally a twofold or more risk in persons with a higher social class and educational level (Cohen et al., 1964; Gutensohn and Cole, 1977, 1981; Abramson et al., 1978; Bernard et al., 1987; Serraino et al., 1991). More interestingly, early case-control studies generally found that there was an inverse association of risk with sibship size, with the risk among persons from larger families only half that of persons from the smallest (Gutensohn and Cole, 1977, 1981; Bernard et al., 1987; Bonelli et al., 1990). In addition, those persons in the later birth-order positions of large families were at lower risk than those born earlier (Gutensohn and Cole, 1977, 1981). In a population-based case-control study conducted in eastern Massachusetts in the late 1970s, we evaluated whether the inverse sibship size association was explainable by its mixture with other related risk factors. These factors included higher maternal education and parental social class, lower housing density in childhood, Jewish religion, fewer playmates, and self-reported history of IM (Gutensohn and Cole, 1981). However, adjustment for these correlated factors had little effect on the association with sibship size, indicating that reduced exposure to infectious agents within the family was associated with increased risk of HL as a young adult. Similarly, living in a single family house during childhood, as opposed to multiple family housing, was a primary risk factor for the malignancy. These findings on sibship size and birth order were replicated in a population-based cohort study involving over two million Danes whose mothers were born in Denmark since 1935 (Westergaard et al., 1997). In this cohort, 72 children and 306 young adults were diagnosed with HL, and their characteristics (sibship size, birth order, parental age at birth, and calendar time) were contrasted to that of the remaining cohort. For those cases diagnosed as young adults, there was a decrease in risk with increasing sibship size (Table 45–2). This decrease was not statistically significant, primarily because of the lower risk observed among single children vs. those with one sibling. A possible explanation for this heterogeneity could be that single children in Denmark are more likely to attend daycare than those with siblings. However, for those in the largest families (≥5 children), the RR of HL was 0.57 (0.30–1.1) compared with those with only one sibling. There is a parallel graded relationship with birth order; p value for trend is 0.07. In contrast, these associations were reversed for the children who developed HL, where being from a large family carried a RR of 3.3 (1.4–8.0). Chang et al.

877

(2004) reported that having older siblings was protective for HL in young adults. Similarly, Chatenoid et al. (2005) reported that having older siblings was protective for HL in adults. In our early Massachusetts study (Gutensohn, 1982), increased risk of HL among middle-aged persons (generally 40–54 years) was also found to be associated with factors that reflect susceptibility to late infections. However, the mix of these factors differed somewhat from that seen among the young adult subjects. The most important risk factor in this middle-aged group was maternal education. Risk among persons whose mothers had more than a high school education was five times that of people whose mothers had not attended high school. Although maternal education was clearly associated with the other social class factors, its association with disease risk was quite primary. Higher paternal social class was also associated with HL risk; however, it had no independent effect once maternal education was taken into account. The findings from a similar study in Israel (Abramson et al., 1978) that provided data for this age group were consistent with the observation that middle-aged cases appear to be individuals whose childhood provided some protection from early infection. These observations suggest that these may be susceptible individuals who were infected as adults, perhaps by their own children. Recently, Glaser et al. (2002) reported an analysis of these identical factors in a population-based study of HL in 204 women in northern California who were diagnosed with HL between the ages of 19 and 44 from 1988 to 1994. For comparison, they interviewed 254 women from the same population. Surprisingly, they found only weak associations with these factors. The authors postulate that this discordance in findings from our similarly conducted early Massachusetts study may be due to secular changes in child rearing practices, particularly the increase in use of daycare and nursery school. We recently completed another population-based study in the greater Boston area and the state of Connecticut (Chang et al., 2004a). This study included 470 cases aged 15–54 years when diagnosed with HL between late 1997 through 2001. Their interview information was compared with that of 557 population controls. In our analysis we found no associations with the childhood social factors found 20 years earlier in much of the same population. What we did find was a significantly protective effect for having attended daycare or nursery school for ≥2 years; odds ratio (OR) = 0.64 (0.45–0.92). These findings are consistent with those of Glaser et al. (2002) and underscore the protective effect of early exposure to other children—and their infections—in relation to subsequent risk of HL among younger adults. Among the oldest persons (≥55 years) in the 1970s Massachusetts study, risk was not directly associated with socioeconomic status (SES) (Gutensohn, 1982). If anything, patients came from a somewhat lower SES than controls. However, within the Israeli population (Abramson et al., 1978), older cases appeared to come from somewhat higher SES. Whether this latter observation is confounded by the apparent general increased risk of HL among Jewish adults is unknown. Glaser et al. (2001) analyzed risk factors for 37 women aged ≥55 years recently diagnosed in the northern California study. They found no consistent associations with childhood social class factors. In our new study, we compared 95 cases aged 55–79 years at diagnosis to 122 comparably aged population controls for childhood risk factors. We found a significant protective effect of low maternal education, OR = 0.52 (0.32–0.83), and low paternal education, OR = 0.64 (0.44–0.95).

Childhood Relatively little is known about environmental risk factors for HL in children younger than 15 years of age at diagnosis. This reflects the rarity of this diagnosis in children, particularly those diagnosed under age 10 years. The annual incidence of HL in United States children aged 0–14 years at diagnosis is about 5.6 per million. Most of what is known about environmental risk factors for HL in children relates to their socioeconomic environment. We conducted an early case-control study in children in the greater Boston area in the late 1970s (Gutensohn and Shapiro, 1982). This population-based study used limited demographic data collected in

878

PART IV: CANCER BY TISSUE OF ORIGIN HL that are characteristically seen in populations living under poor conditions (Fig. 45–2). The second is the marked male excess in very young cases. This was first reported by MacMahon (1966). Fraumeni and Li (1969) found a 3 : 1 male excess in children with HL. This male excess is most marked in the very young cases; in children diagnosed before age four, there is a 19-fold excess of affected males, and in another series a high gender ratio (M/F) of 4.6 : 1 was found for children diagnosed before age seven (White et al., 1983; Kung, 1991). In the ongoing COG study by Grufferman, the highest male excess (9.5 : 1) was found for children aged 0–4.9 years at diagnosis, an intermediate male excess was found for 5–9.9 year olds (2.5 : 1), and the lowest excess was for those aged 10–14.9 years at diagnosis (1.3 : 1). Similarly, Cartwright et al. (2002) found gender ratios of 12.0 for 0–4 year olds, 3.7 for 5–9 year olds, and 2.6 for 10–14-year-old cases in a population-based study in the United Kingdom. These findings are consistent with an infectious etiology for the disease as it is well known that young boys are far more susceptible to infections than are young girls (Washburn et al., 1965; Schlegel and Bellanti, 1969; Melnick, 1997; Green, 1992). This is true for bacterial, viral, and even parasitic infections, and the male excess is greatest in very early childhood. As can be seen from the COG data, the male excess in childhood HL diminishes with age, particularly after age 10. These data suggest that the very early childhood form of the disease is quite different from older forms of the disease, and appears to be related to very early infectious exposure. In summary, children—particularly boys—living under relatively poor living conditions are at greater risk of HL than other children. For both young adult and middle-aged persons, there is evidence that HL is strongly influenced by susceptibility to late infections. The knowledge on hand concerning the role of SES and opportunities for early infections strongly supports the poliomyelitis model of HL causation in these age groups. However, a role for age of infection in childhood is not consistently apparent for risk in the oldest ages, but the data are sparse.

annual town registers. We found that the 14 very young children (<10 years at diagnosis) with HL came from lower SES backgrounds than did population controls. However, for the 52 children aged 10–14 years at diagnosis no SES differences were found. This suggests a transition in SES of HL patients from early to later childhood, in parallel with their likely age at infection with the EBV and other similarly transmitted infections. A case-control study of childhood HL is currently being conducted through the Children’s Oncology Group (COG) and has collected data from over 570 cases of HL in children under age 15 years from numerous treatment centers in the United States and Canada. The study focuses on environmental, lifestyle, and hereditary factors and has data of the EBV status in most of the children’s tumors. The first published finding is a statistically significant, 40% protective effect of breast feeding (Grufferman et al., 1998b). This finding confirms the earlier findings of Davis et al. (1988) and Schwartzbaum et al. (1991). It cannot be determined whether it is exposure of the child to viruses or other infectious agents transmitted by mother’s milk or due to a protective effect of antibodies, cytokines, or other substances in breast milk that protect the child against HL. The association between breast feeding and HL is not modified by EBV status of the patients’ tumors. Perhaps related, is the report of Kusuhara et al. (1997) that there were no observed differences in the acquisition of EBV infection between breast-fed and non-breast-fed Japanese infants. A second important environmental finding in the current casecontrol study is that of an inverse relationship between HL risk in childhood and several markers of SES (S Grufferman, unpublished). This analysis was based on a case-control interview study of 506 HL cases, aged 0–14 years at diagnosis and 763 individually matched community controls. We found inverse associations between HL risk and family income at birth (a 2.6-fold gradient from lowest to highest category, trend p < 0.0001), and at time of interview (1.6-fold gradient, p = 0.01), and with mothers’ educational level (2.9-fold gradient, p < 0.0001). Sibship size was associated positively with HL (a 2.4fold direct gradient, p = 0.13). As noted previously (Table 45–2), in the Danish birth cohort study the risk of childhood HL was greatest among children from large families (Westergaard et al., 1997). Together, the findings regarding social environment for children with HL are the mirror images of what is generally seen for the young adult HL cases. There have also been several descriptive epidemiologic observations of HL in children that are of note. The first is the higher rates of

INFECTIOUS EXPOSURES Clustering Clusters of diseases have been frequently reported in both the medical and lay literature. In light of the long-standing notion that HL might have an infectious etiology, it is not surprising that much attention has been directed to possible clustering of HL. In fact, this research has

Table 45–2. The Relative Risk of Developing Hodgkin Lymphoma in a Danish Cohort of Children by Sibship Size and Birth Order for Diagnosis in Childhood and Young Adulthood Age at Diagnosis <15 years RR (95% CI)

≥15 years Trend (95% CI)

RR (95% CI)

Trend (95% CI)

sibship size* 1 2 3 4 5+

0.71 (0.31–1.61) 1.00 ref 0.94 (0.53–1.68) 1.11 (0.46–2.66) 3.31 (1.36–8.02)

1.28 (1.00–1.63) P = 0.06

0.80 (0.50–1.28) 1.00 ref 0.94 (0.73–1.22) 0.74 (0.50–1.09) 0.57 (0.30–1.08)

0.91 (0.81–1.03) P = 0.12

birth order† 1 2 3 4+

1.00 ref 0.93 (0.50–1.70) 2.04 (0.97–4.26) 1.50 (0.42–5.33)

1.26 (0.92–1.73) P = 0.17

1.00 ref 0.98 (0.74–1.28) 0.78 (0.50–1.22) 0.30 (0.10–0.97)

0.85 (0.71–1.01) P = 0.07

Source: Adapted from Westergaard et al., 1997. RR, relative risk; CI, confidence interval; Trend, the relative increase in risk of Hodgkin lymphoma per increase in sibship size or birth order. *Adjusted for age, gender, calendar period, and maternal age at birth of children. † Adjusted for age, gender, calendar period, maternal age at birth of child, and number of younger siblings.

Hodgkin Lymphoma led to the development of several useful statistical methods and strategies for assessing disease clustering in general. Clustering of a disease can be defined as the occurrence of a group of cases of a disease at the time of their diagnosis in a defined geographic area and time period that is in excess of normal expectancy. This definition is analogous to the public health definition of an epidemic: “The occurrence in a community or region of cases of an illness (or an outbreak) with a frequency clearly in excess of normal expectancy” (Benenson, 1995). Clusters typically are defined on the basis of the location (in time and place) of cases of a disease at the time of their diagnosis. This is to be contrasted with the term “aggregation of exposures” used to denote the closeness of cases in the time and place of their exposure to a possible etiologic agent. Clustering is considered at the time of cases’ diagnoses and aggregation at the time of their exposure to a suspected cause of their disease (Grufferman, 1977). When the latency period between a causal exposure and development of a disease is relatively short, as in most infectious diseases, studies of clustering and aggregation of exposures should yield similar results. When disease latency periods are long or variable, the results of the two types of studies would be expected to be quite different. In the case of HL, the disease latency period may be long and variable. Thus, studies of clustering of HL cases are likely to prove of limited value. Some general principles of clustering are worth addressing. First, although case clustering is a characteristic of many infectious diseases, clusters could also be due to common-source exposure to noninfectious environmental agents. Second, many apparent clusters of HL (and other cancer) cases are chance occurrences that come to attention because of community concerns about possible causal environmental hazards. Although overall patterns of occurrence of a disease across a country may be randomly distributed, small chance clusters can be observed at more local levels. The frequency of such small area random clusters has been shown to be surprisingly high (Smith and Pike, 1974). Third, the issue of small area random variation is further compounded by well-meaning investigators who use fallacious methodology to assess community clusters. In early cluster investigations, many people used what has been termed the “Texas sharpshooter” approach (Grufferman, 1977). This term is based on the joke about a traveler in rural Texas who notes that on all of the barns in the region there is a painted target with a single bullet hole exactly in the center of the bull’s eye. When the traveler stopped for gas, he inquired about the incredible sharpshooter. He was informed that it was the work of a man named “Old Joe” who would first shoot at the side of the barn and then paint a target over the bullet hole so that it was always in the center of the bull’s eye. In many early investigations of apparent clusters, the geographic spread and time period of the cluster were defined in a similar post hoc manner. The geographic area for study was often selected on the basis of which cases lived farthest apart, and the time frame as the period between the diagnoses of the first and last case. When statistical analyses are then done to compare the observed frequency of cases in the cluster with that expected, there was almost always a strikingly higher than expected incidence of the disease. Whereas this fallacy has become well recognized and is less frequently used, two other fallacious approaches to cluster investigation have been increasingly used, particularly by concerned lay community members and the media. Many community investigations of clusters fall victim to the “Rain Dancers’ Fallacy” or as used in logic textbooks, the “post hoc, ergo propter hoc” (after this, therefore because of this) fallacy. In this method of cluster investigation, following observation of a suspected cluster, potential antecedent environmental carcinogenic exposures are sought. If the investigator finds some suspected causal exposure, it is concluded that it must have caused the disease because cases were exposed to it before they developed the disease, and there is a cluster in relation to the suspected cause. This post hoc approach is analogous to the rain dancers’ fallacy. Every year, just before spring, an aboriginal tribe performs an important ceremony, the rain dance. It is believed that unless the dance is performed, the spring rains vital to their agriculture will not come. Almost invariably, the dance works its magic and the rains come. A final frequently used fallacious approach

879

to cluster investigation can be termed “The Usual Suspects Approach” (or Fallacy). This is based on the famous line of the late Claude Raines in the movie “Casablanca”: “Round up the usual suspects.” It is a frequent companion to the rain dancers’ fallacy in many cluster investigations. After the observation of an apparent cluster, the investigator looks about to find a potential environmental cause to explain the cluster. In the usual suspects approach, particular efforts are made to identify environmental factors that are known or suspected carcinogens. It does not matter in this approach whether the suspected carcinogen has ever been linked to the specific cancer in the cluster, so long as it has been linked to any cancer. The investigator then concludes that the clustered cases have been exposed to an environmental exposure that is a suspected or real cause of another cancer; therefore, it is likely to have caused the different cancer in the cluster. In some instances, the fallacious reasoning is compounded by the use of circular reasoning. In such a situation, the ill-informed investigator finds a link between a suspected environmental carcinogen and a disease cluster and then reasons that the cluster must be real (i.e., not due to chance). Although well-informed scientists should not fall victim to these fallacies in cluster investigation, many concerned, well-meaning community groups and media reporters do. Most of the early statistical methods for assessing space-time interactions of disease were developed to study leukemia. Knox (1963) first reported the method of “all possible pairs” in 1963. This was followed by the statistical methods for assessing clustering of Ederer, Myers, and Mantel (1964), that of Barton et al. (1966), and that of Mantel (1967). The first report in the literature of case clustering of HL was that of George et al. (1965), who reported two college classmates with the disease. Using the methodology developed for leukemia clusters, several investigations of HL clusters were carried out in the 1970s and 1980s in the United Kingdom and the United States (Alderson and Nayak, 1971, 1972; Kryscio et al., 1973; Greenberg et al., 1983; Grufferman and Delzell, 1984). More recently, Gilman et al. (1999), assessed space-time clustering in the United Kingdom during 1984–1993. In assessing the results of these studies, the most salient feature is the lack of consistency of findings among the studies. All the studies assessed HL clustering using many stratified analyses. All look at clustering in different age groups, and most looked at different time periods. Analyses were also stratified by religion, gender, and histologic subtype. For example, in the study by Gilman et al. (1999), analyses were stratified by age at diagnosis, time of diagnosis, gender of the case, and histologic subtype. They found no space-time clustering for the 0–14 year age group. For the age group 15–34 years, they found little evidence of clustering for the non-NS subtype during both time periods. For this same age group with NS-HL, many significant excesses of case-case pairs were found during the first time period (1984–1988) but not during the second time period (1989–1993). For the 15–34 year NS subtype, 1984–1988 subgroup, the evidence of clustering was almost entirely for females. Among the older cases aged 35–79 years at diagnosis, there was no significant evidence of clustering for the non-NS cases. For the NS-HL cases, there were no significant excesses of pairs during the first time period. During the second time period the authors found “weak evidence of clustering among cases of NS disease in older adults, particularly males diagnosed in 1989–1993” (Gilman et al., 1999). The lack of consistency of results within studies and between studies is the hallmark of studies of space-time interactions in HL. It would take a highly convoluted hypothesis to reconcile the findings in the various studies. For example, how does one explain observations of clustering for females, but not males, in one time period, but not in the other in the Gilman et al. (1999) study? Statistical significance (even when conservatively computed) cannot substitute for the application of biological common sense in assessing epidemiologic observations. The issue of possible person-to-person transmission of HL was further advanced by two studies of HL at schools in New York State in the early 1970s. In 1971, Vianna et al. reported a remarkable aggregation of HL cases centered about the 1954 graduation class of Albany High School (Vianna et al., 1971). There were 31 HL cases that were

880

PART IV: CANCER BY TISSUE OF ORIGIN

linked to each other socially or by family relation over a 23-year period. Nine of these links were of cases to other cases and 22 were indirect case–contact–case linkages. The investigators used two comparison groups to evaluate the significance of the linkages observed. However, the cases were identified on the basis of linkages having been observed, and the comparison groups were identified on other bases; for example, one consisted of patients in a burn registry. Thus, the significance of the Albany report cannot be evaluated. Nevertheless, this was an important report for several reasons. First, the aggregate of cases reported is one of the largest reported for any cancer. Second, the authors used methods of infectious disease epidemiology to investigate a cancer. For example, they sought case–contact–case linkages for HL. Third, this study was the first to address the issue of aggregation of exposures for a cancer rather than clustering. Due to the limitations of the Albany study and having identified high schools as a possible locus for transmission of HL, Vianna and Polan (1973) then reported a study of HL at high schools on Long Island, New York. They used a well-designed study based on the methods of infectious disease epidemiology that employed two different approaches. One approach assessed the HL risk to students and teachers who overlapped with a case for at least one year. They found an RR of two for students and eight for teachers who were thus “exposed” to HL cases. The second approach divided the 10-year study period into two 5-year periods. Those high schools with a case in attendance during the first 5 years were considered as positive and those without as negative schools. The two groups of schools were compared with regard to the occurrence of HL cases during the second 5-year period. They found that students at positive schools were 15 times more likely to be diagnosed with HL during the second time period than were students at negative schools. At both the individual and school levels of exposure, there was striking evidence suggesting person-to-person transmission of HL at schools. The Long Island school study findings created a great deal of scientific and public interest. However, there were several criticisms of the study, primarily that there appeared to be incomplete case finding (Pike et al., 1974; Smith and Pike, 1974). We then undertook an extensive replication of the Long Island study in the greater Boston area (Grufferman et al., 1979). This large population-based study did not confirm the Long Island findings. There was no evidence of increased risk of HL in Boston students exposed to cases in high school. Other studies also examined aggregation of HL cases at schools. Smith et al. (1977) examined case to case linkage among HL patients less than 40 years of age in Oxford, England. Their study was essentially a re-evaluation of the Albany study, although the British study only assessed direct case-to-case links. They found no excess of overlapping case pairs in school attendance compared with control pairs. Zack et al. (1977) conducted a similar study using HL cases identified through the Connecticut State Tumor Registry. High school attendance could be determined for 83% of cases. They used simulated controls proportionally drawn for Connecticut high schools. Both groups were assessed for overlapping attendance at the same schools. Some lowmagnitude positive associations were found. Heath et al. (1971) did a survey of HL in Atlanta and found no excess of overlap in cases at high schools. Paffenbarger et al. (1977) searched for overlapping school attendance and social contacts between HL cases in a cohort of 50,000 male college students and found no evidence of aggregation. There were also several case-control studies of HL that searched for school contacts between cases (Schimpff et al., 1976; Greenwald et al., 1979; Isager and Larsen, 1980; Scherr et al., 1984; Davis, 1986). The findings of these studies were variable; some found evidence of case-case linkage and others did not. None of the positive associations was very high. On balance, the accumulated study results suggest there is no evidence of increased risk of HL to persons who attend schools where cases of HL are in attendance. But how does one explain the very strong findings of the Long Island school study? The most likely explanation was that there was serious under-ascertainment of cases that introduced a strong bias. The HL incidence rate based on the Long Island study data was lower than the mortality rate for HL reported for the area during the same time period (Grufferman et al., 1979). Additionally, the Long Island study

excluded cases diagnosed in nearby New York City, which could have led to selective loss of cases among residents of Long Island living adjacent to the city. This could result in artifactual aggregation of cases at schools in areas of Long Island remote from the city. Perhaps as a testament to the strength of the hypothesis that HL has an infectious etiology, there has been continued research on clustering in HL. In the late 1980s new statistical methods were developed for assessing spatial clustering of cases as opposed to time-space clustering (Alexander et al., 1989; Alexander, 1990; Cuzick and Edwards, 1990). These approaches test for spatial clustering of cases using a “nearest neighbor” test and use both cases and controls from the same specified region. Basically, it tests for the frequency of nearest neighbors who are cases in the two groups. Dockerty et al. (1999) used the Cuzick-Edwards test, with an extension of the approach by Jacquez (1994a,b), to assess spatial clustering of HL (and leukemia and NHL) in persons less than 25 years in New Zealand based on the national cancer registry. Because the study was focused primarily on acute leukemia, clustering was assessed using the cases’ residence at birth rather than at diagnosis. No significant clustering for HL overall or in the age groups 0–14 and 15–24 years was found. More recently, Alexander et al. (1995) evaluated the spatial clustering of HL in relation to the EBV genome status of tumors and to herpes virus serology. They used the nearest neighbors’ test of Alexander et al. (1989) to assess spatial clustering. Their underlying hypothesis was that the EBV status and/or elevated antibody titers might be associated with “. . . aggregation of exposures during schoolage and adolescence and hence with clustering by residence at diagnosis or at ages 5–18 years.” This study builds on an earlier one in the same region (the Yorkshire Health Region) for 1984–1987 (Alexander et al., 1989). The first study found that for young people (£34 years at diagnosis) the percentage of clustered cases was 17% whereas only 8% were expected by chance (p =< 0.05). They also looked at spatial clustering among all NS-HL cases at all ages for the same period and found significant clustering. The second study appears to overlap with the first study in that it covers the same registry of cases for 1985–1989. The second study also found significant spatial clustering for young onset cases, but did not confirm the earlier observation of significant clustering for NS cases overall. Using a case-control design, cases were classified as clustered, peripheral, and random by a system that would be difficult to apply elsewhere. The proportion of cases with EBV-positive tumors was highest in the clustered cases, intermediate in the peripheral group, and lowest in the random group, and the trend was statistically significant (p = 0.02). However, the number of cases with EBV-positive tumors in each group was extremely small—5, 2, and 1, respectively. Given the very small numbers and the complicated study design, the results are difficult to interpret. However, if the results are correct, they would imply that there is significantly greater clustering in EBV-positive cases than in EBV-negative cases. It is important to note that this study was the first to integrate clustering, EBV genome status, and EBV antibody profiles into a unified investigation (Grufferman, 1995). A curious addendum to the issue of clustering is the paucity of reported spouse pairs with HL in the literature. If the disease was caused by person-to-person transmission, spouses of cases would be expected to be at increased risk. However, person-to-person transmission of HL might only occur in the very young. Several studies have found some evidence of clustering in the age groups 0–34 or 15–34 and these would have included many married people. Because HL is very rare in the 0–14 age group, the use of an age group of 0–34 would be heavily weighted towards cases with onset at ages 20–34. This apparent lack of increased risk to spouses of cases argues against person-to-person transmission of an infectious agent or suggests that such transmission only occurs during early life. In summary, the numerous studies of aggregation of exposures, other than those of Vianna et al. (1971) and Vianna and Polan (1973), have been negative or inconclusive. At this juncture, it is safe to say that there is no confirmed person-to-person transmission of HL at schools. The numerous studies of time-space clustering in HL have been negative, inconsistent, and occasionally weakly positive. It must

881

Hodgkin Lymphoma be emphasized that almost all of these studies assessed clustering at the time of diagnosis when the putative causal exposures might have occurred years in advance. Similarly, the more recent studies of spatial clustering in HL have produced inconsistent and confusing results. These methods also primarily use data of diagnosis as the defining point for a case’s residence. The only suggestion of clustering emerging from all these investigations is the sporadic finding of evidence for clustering among young adults. But the associations are not striking. On balance, it is safe to say that there is no persuasive evidence to suggest that there is meaningful time-space clustering of HL. It is unlikely that further clustering studies of HL would yield any meaningful new knowledge. Without knowledge of the latency period between the causal exposure(s) and HL diagnosis, and its degree of variability, new studies will likely continue to provide unclear results. Additionally, given the emerging new knowledge about the role of EBV in HL etiology, the issue of clustering in support of an infectious etiology of HL may become moot.

Epstein Barr Virus The ubiquitous herpesvirus EBV has long been the leading candidate as an etiologic infectious agent for HL (Evans, 1971). (See Chapter 26). This association was first suggested by the ability of the virus to transform B-lymphocytes, the finding of RSC in lymphoid tissue of patients with IM (Lukes et al., 1969), and the recognition of the epidemiologic similarities between HL and IM. Subsequent epidemiologic, serologic, and molecular biologic data have confirmed the association. There have been several case reports of HL developing in close association with serologically documented EBV primary infection (Henle and Henle, 1973; Green et al., 1979; Veltri et al., 1983). Further, in seven cohort studies involving approximately 80,000 young adults with serologically confirmed IM who were followed for subsequent HL, a consistent twofold to threefold excess of HL incidence has been found (Miller and Beebe, 1973; Rosdahl et al., 1974; Connelly and Christine, 1974; Carter et al., 1977; Munoz et al., 1978; Kvale et al., 1979; Hjalgrim et al., 2000). The recent study of Hjalgrim et al. (2000) included over 38,000 IM cases in population-based cohorts identified in Denmark and Sweden. To address whether their finding was due to confounding by social class, the authors evaluated whether the first-degree relatives of the 17,000 Danish IM patients had themselves an elevated risk of HL. No excess HL incidence was found among these relatives, strengthening the validity of the elevated risk with prior IM itself (Hjalgrim et al., 2002). The association of IM with HL finding has been generally replicated in case-control studies (Gutensohn and Cole, 1981; Evans and Gutensohn, 1984; Bernard et al., 1987; Serraino et al., 1991; Levine et al., 1998; Vineis et al., 2000; Alexander et al., 2000). However, in the recent population-based case-control study that we conducted in the state of Connecticut and the greater Boston area, we did not find an association between history of IM and HL either in comparison to population controls (Chang et al., 2004a), or in case-case analysis (Chang et al., 2004c). As found in other EBV-associated cancers, HL patients as a group characteristically have an altered EBV antibody pattern. In studies of more than 1900 HL cases of all ages, the proportion who have IgG antibody to the EBV viral capsid antigen (VCA), indicative of prior infection, have been quite similar to that of controls. However, the cases consistently have higher mean antibody titers than controls. Further, more of the cases have antibodies (as well as higher titers) against the early antigen (EA) of EBV, indicative of viral replication (Mueller, 1987). In the early Massachusetts case-control study conducted in the late 1970s, we compared the antibody profiles of 304 cases to those of 276 of their siblings (Evans and Gutensohn, 1984). Of particular interest was the observation that those young adult cases with a history of IM had a significantly higher geometric mean titer (GMT) of antibodies against VCA (239), than cases who did not report a history (159). Similarly, sibling controls in this age group with a history of IM had a

somewhat higher GMT of 71, compared with a GMT of 57 for the siblings without a history of IM. Because all these antibody results were based on blood specimens collected after the diagnosis of the disease, the findings may simply reflect the reactivation of latent EBV as a result of the immune dysfunction characteristically seen in HL. To clarify the temporal relationship between altered antibody patterns against the EBV and HL, we undertook a collaborative serologic case-control study in the 1980s, consolidating the resources of five serum banks with specimens from over 240,000 persons (Mueller et al., 1989). In these populations, we identified 43 patients from whom blood had been drawn and stored an average of 50.5 months before diagnosis. For comparison, we selected 96 matched controls who had blood drawn at the same time. All blood specimens were assayed for antibodies against the VCA (IgG, IgA, IgM), the EA (both the diffuse (EA-D) and restricted form (EA-R), and the nuclear antigen complex (EBNA). As shown in Table 45–3, the previous finding of elevated IgG titers against the VCA, RR = 2.6, and EA (RR for EA-D = 2.6 and EA-R = 1.9) among the cases were confirmed. In addition, a greater proportion of cases had elevated titers of IgA against the VCA than controls, RR = 3.7, and substantially fewer had IgM antibody, RR = 0.22. When all antibodies were simultaneously controlled, the most significant findings were that a higher prevalence of high titers for antibody against the EBNA complex (RR = 6.7) and a low prevalence of VCAIgM (RR = 0.07) predicted subsequent HL. These findings for altered EBV antibody patterns before the diagnosis of HL were generally stronger in blood specimens drawn at least three years before diagnosis than in those tested closer to the time of diagnosis. Lehtinen et al. (1993) conducted a similar study in a cohort of 39,000 healthy Finnish adults followed for 12 years. Of these, six were subsequently diagnosed with HL. Although the data were not shown, the authors report that risk for HL was associated with increased antibody response to the EBNA complex and to the EA, consistent with our findings. The combination of both elevated anti-EA and elevated anti-EBNA is unusual based on our knowledge of normative host control of EBV (Rickinson and Kieff, 1996). The presence of anti-EA is thought to reflect virus replication and is negatively associated with the frequency of EBV-specific cytotoxic T-cells (CTL). Elevated anti-EBNA complex is thought to reflect the level of in vivo destruction of virusinfected cells by T-cells, and is positively associated with the frequency of EBV-specific CTL (Kusunoki et al., 1994). The observations from these two pre-diagnosis serologic studies implied that for a subset of cases, the development of HL is preceded over an extended period of time by endogenous EBV activation, likely coupled with an unusual host type 2 immune response. The traditional assay for anti-EBNA measures antibodies against a complex of EBNA antigens that are expressed in cells latently infected

Table 45–3. The Relative Risk of Hodgkin Lymphoma Associated with Elevated Titers of Antibodies against Epstein Barr Virus in Blood Samples Drawn Prior to Diagnosis Relative Risk (90% Confidence Interval) Epstein Barr Virus Antibody VCA IgG (≥1 : 320)† IgA (≥1 : 20) IgM (≥1 : 5) EBNA (≥1 : 80) EA Diffuse (≥1 : 5) Restricted (≥1 : 40)

Adjusted for IgM

Full Model*

2.6 (1.1–6.1) 3.7 (1.4–9.3) 0.22 (0.04–1.3)

1.7 (0.52–5.4) 4.1 (1.3–12.9) 0.07 (0.01–0.53)

4.0 (1.4–11.4)

6.7 (1.8–24.5)

2.6 (1.1–6.1) 1.9 (0.90–4.0)

1.5 (0.55–4.2) 1.2 (0.41–3.4)

Source: Mueller et al., 1989. VCA, viral capsid antigen; EBNA, nuclear antigen; EA, early antigen. *Adjusted for all other antibodies against EBV. † Referent category for all antibodies less than elevated.

882

PART IV: CANCER BY TISSUE OF ORIGIN

with the EBV. Anti-EBNA2 is the predominant form of the antibody during acute EBV infection. With the resolution of the acute infection, there is a shift from anti-EBNA2 to anti-EBNA1. Thus, anti-EBNA1 is the predominant form of the antibody to EBNA complex in normal carriers of the infection (Rickinson and Kieff, 1996). Empirically, a persistent anti-EBNA1/anti-EBNA2 ratio of £1.0 has been considered to be abnormal (Henle et al., 1987). Whether the elevated anti-EBNA titers found in HL patients were predominantly anti-EBNA1 or EBNA2 was unknown. To address this question and to determine whether anti-EBNA2 was also elevated in healthy persons with a history of IM, we undertook a study based on the case and sibling serum samples from the early Massachusetts case-control study. In this serologic study we selected the 35 HL cases and 23 siblings of cases who had reported a history of IM. For each of these persons, three controls without a history of IM were selected, matched for age, gender, year of blood draw, and case/sibling status. The levels of anti-EBNA1 and anti-EBNA2 were measured in the blood specimens. In brief, we found that among the HL cases, there was an increased prevalence of antibodies against EBNA2; the RR associated with an anti-EBNA1/EBNA2 ratio £1.0 was 2.6 (1.1–6.0). The prevalence of the low ratio was also elevated with history of IM for both HL cases and the sibling controls; the RR for the combined groups was 1.6 (0.7–3.5). These findings suggest that defective host control of the EBV underlies the risk of HL and may be more common following IM (N. Mueller, unpublished).

Molecular Evidence of Epstein Barr Virus With the advent of highly sensitive molecular probes, assays for the direct detection of viral genome or of viral-encoded proteins or transcripts have demonstrated the presence of the EBV in the tumors of many HL patients (Ambinder and Weiss, 1999; Gulley et al., 2002). The reports by Weiss et al. (1987, 1991) provided the first concrete evidence that monoclonal episomal EBV was detectable in HL tissue and was localized to the RSC in 4 of 21 specimens tested. This discovery has been confirmed in a large number of subsequent reports (reviewed by Mueller, 1996). The detection rate of EBV genes or gene products in HL biopsies is higher with the use of more sensitive methods, such as that detecting the presence of the abundant virally encoded small nonpolyadenylated, noncoding RNA transcripts that are expressed in EBV latently infected cells, the so-called “EBERs” (Ambinder and Mann, 1994). Overall, the findings indicate that 30%–40% of HL tumors are EBV-positive. Pallesen et al. (1991a) further demonstrated that the EBV-positive RSC express an altered latent infection phenotype of latent membrane protein (LMP1)-positive but EBNA2-negative, both viral gene products being normally co-expressed in latently infected lymphocytes (Kieff and Liebowitz, 1990). Subsequent investigators confirmed this important finding (reviewed by Mueller, 1996). This has also been reported in a small proportion of T-cell NHL (Hamilton-Dutoit and Pallesen, 1992). The latent phenotype of EBV found in HL is similar to that found in EBV-positive nasopharyngeal carcinoma (Knecht et al., 1993a). These two malignancies also share similar EBV transcriptional programs (Deacon et al., 1993) and similar serologic patterns of elevated anti-EA-D and IgA antibodies against the VCA (Evans and Mueller, 1990; IARC, 1997). The EBV status of HL tumors appears to be consistent at multiple involved sites in a given patient (Boiocchi et al., 1993; Vasef et al., 1995; Herling et al., 2003). EBV status is also generally stable over time. Coates et al. (1991) evaluated sequential biopsies in three EBVpositive patients (range 2–10 years after diagnosis) and found all remained positive, with about the same level of viral genomes detected as in the initial biopsy. Delsol et al. (1992) reported that EBV status was consistent in subsequent biopsies at relapse (range 14–126 months) in 12 cases, of which seven were initially positive. They reported that two of the EBV-positive cases showed substantial reduction or loss of LMP1 staining in their later biopsy. Brousset et al. (1994) evaluated the initial and relapse biopsies of 12 HL cases. Five cases were EBV negative in both samples. Of the seven initially EBV-positive tumors, all were EBV positive at relapse. For each of two of these latter cases for which DNA analysis was possible, the

authors reported that the identical virus was found in the sequential biopsies. In terms of EBV strains found in HL, most are the predominant type circulating in the population at question (Jarrett et al., 1991; Ambinder et al., 1993; Boyle et al., 1993; Lin et al., 1993; Preciado et al., 1995). In general, the number of virus episomes per cell is low (Staal et al., 1989). When adjusted for the number of RSC, Gulley et al. (1994) estimates that at least 50 copies of viral DNA per RSC. The question of the specificity of the EBV positivity in HL was raised by reports of EBV-positive (but LMP1-negative) small lymphocytes in both EBV-positive and EBV-negative HL (Masih et al., 1991; Weiss et al., 1991), which are also present at a low frequency in normal lymph nodes (Niedobitek et al., 1992). Further, clonal EBV has been detected in reactive hyperplasia biopsies (Libetta et al., 1990; Masih et al., 1991). However, in a recent report of three HL cases, single-cell analysis of the clones found in small EBV-infected cells, and the EBV-positive HL cells was performed. In two of these cases, all small EBV-infected cells were clonally unrelated to the tumor cells. In the third case, 2 of 29 small EBV-infected lymphocytes carried the same clone as the tumor cells (Spieker et al., 2000). In summary, the consistency of the finding of clonal EBV of a unique phenotype expressing the oncogenic viral protein LMP1 (Knecht et al., 1993b) in a substantial number of HL cases, in many patient populations throughout the world, argues strongly for a causal role in these cases (Herbst and Niedobitek, 1993). EBNA1, the virus genome maintenance protein, is also expressed in EBV-positive HL tumors (Deacon et al., 1993; Young, 1994; Grässer et al., 1994). The expression of LMP2A in 22 of 42 EBV-positive cases has also been reported (Niedobitek et al., 1997). The expression of LMP2A appears to lead to an activated proliferative state (Portis et al., 2003). Another viral protein, BARF0, is also expressed in EBV-positive HL (Oudejans et al., 1995). Pallesen et al. (1991b) evaluated whether the BZLF1 gene product ZEBRA was expressed in 47 HLpositive biopsies. This product induces the switch from latency to the lytic cycle and virus replication. They found that it was rarely expressed—only three cases were positive—and no structural viral proteins were detected. Brousset et al. (1993) and Bibeau et al. (1994) replicated this finding. This observation suggested to these investigators that the latent state of the EBV in HL is not severely impaired; rather, the infrequent activation of replication is impaired, resulting in an abortive viral productive cycle. Thus, the apparent mechanism of EBV in HL is not related to viral replication per se, but more likely to the transforming and transactivating properties of LMP1—in particular, the constitutive upregulation of the nuclear factor kB (NFkB) pathway (Hinz et al., 2002). The lack of expression of EBNA2, a target for cytotoxic T-cells, likely aids the escape of the RSC from immune surveillance.

Variation of Histology and Demographic Factors by EBV Status The data concerning histology and EBV status of presumably HIVnegative HL cases from representative studies from 1989–1999 (with at least 30 specimens determined by EBER and/or LMP1 staining), are summarized in Table 45–4. Because the sensitivity of assays varies between studies, the most valid comparison is within studies. Of the two most common subtypes of classical HL, MC-HL cases generally had higher rates than the NS-HL cases, with a range 10%–100% in studies with at least 10 MC-HL cases tested; however, the rates among NS-HL cases were also sometimes high and covered the same range. In studies involving cases from economically developing populations, in which patients generally present with more advanced disease, the positivity rate is notably high. Gulley et al. (1994) compared 125 cases from the United States, Mexico, and Costa Rica and found in multivariate analysis that Hispanic ethnicity per se was an independent predictor (RR = 4.3) of EBV positivity. Flavell et al. (2000), based on a large population case series from the United Kingdom, reported that histology—rather than age per se—predicted EBV positivity. Early childhood cases (<10 years), particularly in developing populations, tend to have high rates of EBV positivity (Ambinder et al., 1993; Razzouk et al., 1997; Kusuda et al., 1998; Flavell et al., 1999, 2001; Zhou et al., 2001).

883

Hodgkin Lymphoma Table 45–4. Prevalence of Epstein-Barr Virus Genome or Gene Products Detected in Tissue of Presumably HIV-Negative Hodgkin Lymphoma Cases by Histologic Classification

Reference Herbst et al. (1989) Uhara et al. (1990) Herbst et al. (1990) Libetta et al. (1990) Pallasen et al. (1991a) Brousset et al. (1991) Herbst et al. (1991) Masih et al. (1991) Jarrett et al. (1991) Gledhill et al. (1991)* Weiss et al. (1991) Coates et al. (1991) Brocksmith et al. (1991) Delsol et al. (1992) Khan et al. (1992) Herbst et al. (1992) Fellbaum et al. (1992) Murray et al. (1992) Ambinder et al. (1993)† Honduras United States Chang et al. (1993)‡ Carbone et al. (1993) Niedobitek et al. (1993) Khan et al. (1993) Gulley et al. (1994) United States Mexico Costa Rico Preciado et al. (1995)† Weinreb et al. (1996)† United Kingdom Greece Middle East Africa Australia Costa Rico Razzouk et al. (1997)† United States Brazil Enblad et al. (1999)

Lymphocyte Predominance % + (No.) 33 (3) 25 (4) 50 (12) 60 (5) 10 (10) 0 (5) 0 (1) 42 (12) — 100 (2) 0 (14) 0 (7) 60 (5) 0 (10) 0 (2) 100 (2) 31 (13) 8 (12)

Nodular Sclerosis % + (No.) 15 (26) 24 (17) 61 (109) 41 (22) 32 (50) 18 (22) 70 (27) 52 (21) 40 (45) 27 (26) 33 (12) 29 (24) 54 (41) 10 (40) 15 (26) 42 (24) 28 (98) 50 (24)

Mixed Cellularity % + (No.)

Lymphocyte Depletion % + (No.)

0 (10) 33 (9) 58 (64) 50(6) 96 (24) 46 (26) 72 (18) 69 (13) 52 (25) 17 (6) 75 (8) 13 (16) 73 (11) 60 (55) 67 (3) 56 (18) 50 (68) 86 (7)

— 0 (1) 25 (4) 0 (1) — — 0 (1) 83 (6) — 100 (1) 50 (2) 0 (8) — — 0 (2) 50 (2) 12 (8) 100 (3)

100 (1) 0 (2) — 5 (0) 11 (9) 0 (10)

100 (3) 13 (15) 100 (7) 10 (20) 27 (75) 24 (38)

100 (6) 86 (7) 100 (20) 64 (11) 39 (31) 68 (22)

— — 60 (5) 100 (3) 0 (1) 14 (7)

— — — 33 (6)

25 (53) 50 (8) 20 (10) 0 (9)

67 (21) 87 (23) 75 (4) 76 (25)

— — — 100 (1)

36 (14) 100 (2) 50 (8) 100 (3) 0 (1) 33 (3)

41 (37) 92 (12) 40 (10) 93 (30) 73 (11) 79 (24)

84 (19) 86 (7) 63 (19) 86 (28) 100 (3) 100 (10)

0 (1) 0 (1) 20 (5)

33 (6) 50 (6) 23 (87)

71 (17) 59 (17) 35 (23)

40 (5) — 50 (2) 82 (11) — 80 (5) 50 (2) 100 (2) 100 (3)

*Subset of above. † Pediatric cases. ‡ Cases from Peru.

Glaser et al. (1997) conducted a combined analysis of data from 14 studies of EBV positivity and demographic and histologic features of HL, involving a total of 1546 patients. Using multivariate analysis, they found that patients with EBV-positive HL were significantly more likely to have MC vs. NS histology; the RR varied by age group—7.3 (3.8–14.2) for children, 13.4 (9.0–19.9) for ages 15–49, and 4.9 (2.8–8.7) for those older. Cases with EBV-positive disease were more likely to be male among young adults, RR = 2.5 (1.7–3.3) and also among children from less economically developed areas, RR = 6.0 (2.0–18.0). In addition, patients of Hispanic background were more likely to have EBV-positive disease, RR = 4.1 (1.8–9.6), consistent with the findings of Gulley et al. (1994). These findings are consistent with those of Flavell et al. (2001) in the United Kingdom. In this study Townsend scores were computed to estimate “material deprivation”, a measure of SES of subjects’ residence, because no direct information on cases’ SES was available. They found a very strong association between EBV-positive HL and being of South Asian origin in cases aged 0–14 at diagnosis (OR = 17.1 (2.0–146.7) ). They also found that the “most deprived” quartile of cases had a 7.0-fold (1.10–44.6) increased risk of having an EBVpositive tumor than the least deprived quartile. When ethnicity of sub-

jects was adjusted for in logistic regression analyses, there remained a 3.1-fold difference between most deprived and least deprived groups, although this was not statistically significant. Because essentially all HL cases have antibodies against the EBV, these risk factors relate not to whether a case has been infected with EBV, but more likely, the interaction of virus and host. (Some EBVnegative cases are seronagative (Alexander, 2001; Chang et al., 2004c.) Thus, given that a patient has HL, the likelihood that the tumor DNA itself contains the EBV genome appears to be related to factors indicative of somewhat poorer host response; namely, very young or old age, male gender, living under somewhat poorer conditions, having MC or LD histology. Whether the increased risk among Hispanics is attributable to poorer living conditions or to genetic factors is unknown. There is some evidence that patients with EBV-positive HL exhibit impaired immune control of the virus. The most striking observation is by Frisan et al. (1995). In this study, cultures of EBV-specific CTL from nine patients were derived from tumor-infiltrating lymphocytes and expanded in interleukin (IL)-2 conditioned medium. These CTL were then tested for killing of autologous EBV-transformed lymphoblastic cell lines. None of six cultures from EBV-positive patients

884

PART IV: CANCER BY TISSUE OF ORIGIN

exhibited human leukocyte antigen (HLA) class 1-restricted EBVspecific cytotoxicity to several EBNA epitopes, whereas all of three cultures from EBV-negative patients did. It has also been reported that levels of expression of IL-10—a cytokine that down-regulates cellular immunity—are significantly higher in EBV-positive RSC than in EBV-negative RSC (Herbst et al., 1996; Dukers et al., 2000; Herling et al., 2003). Higher levels of IL-6, another type 2 cytokine, also have been reported among EBV-positive HL cases (Herbst et al., 1997).

Association of Epstein Barr Virus Genome Status with IM History and Epstein Barr Virus Serology We conducted an analysis of a prevalent series of cases on the association of IM history and the EBV status of HL cases (Sleckman et al., 1998). This study included 83 cases whose biopsies could be typed for EBV status, and who completed a questionnaire. These cases were 15–55 years at diagnosis and had been diagnosed between 1980 and 1994 (median, 1998). Sixteen were found to have had EBV-positive HL. The two types of cases were compared to determine whether the association of various risk factors differed between the two (Begg and Zhang, 1994). For IM history, no difference was apparent; the OR was 1.2 (0.28–4.9). Alexander et al. (2000) evaluated the association of IM history in relation to EBV status of HL in a population-based case-control study conducted among young adults (aged 16–24 years) in the United Kingdom in 1991–1995. A total of 103 HL cases with EBV typing were interviewed; of these, 19 cases were EBV-positive. Two population controls were identified for each with matching for gender, age, and administrative area of residence. The OR for “definite” prior IM history was significantly elevated, 9.2 (1.1–78.3) among the EBVpositive cases in comparison with their matched controls. However, among the 84 EBV-negative cases, there was no significant association, OR = 1.6 (0.63–4.1). We recently completed a population-based study in eastern Massachusetts and Connecticut involving HL cases diagnosed between August 1, 1997, and December 31, 2001 (Chang et al., 2004c). In this study, we compared 96 EBV-positive HL cases with 311 EBVnegative cases for IM history. The adjusted OR was 1.1 (0.48–2.2). In a population-based case-control study conducted in Scotland and the Northern Region of England, Alexander et al. (2003) evaluated the association of IM history and risk of HL by EBV status. The study included 408 cases of classic HL, aged 16–74 and 513 controls. The cases included 113 EBV-positive and 243 EBV-negative cases. The authors found a positive association with both EBV-positive and EBVnegative HL with IM history. Among subjects aged 16–34 years, IM history was associated with EBV-positive HL, OR = 2.9 (1.1–8.0), and with EBV-negative HL, OR = 1.9 (0.85–4.1). Among subjects aged ≥35 years, the ORs were 2.2 (0.71–6.7) and 2.5 (0.93–6.6), respectively. On case-case analysis, the OR for EBV-positive HL was 1.8 (0.63–4.9) for the younger group and on the null for the older. The study also evaluated whether familial (based on first-degree relatives) history of IM was predictive of HL. Among the younger group, the OR associated with EBV-positive HL was 5.2 (2.2–12.7) and 1.8 (0.85–4.0) with EBV-negative HL; on case-case analysis, there was a significant difference, OR = 3.4 (1.3–8.4). However, among older subjects, there was no evidence of any association with HL with familial IM history. Finally, Hjalgrim et al. (2000) evaluated the role of EBV status for HL cases developing in their previously published cohort study of patients with IM in Denmark and Sweden. In their new study (Hjalgrim et al., 2003), the authors were able to characterize EBV status on biopsies from 29 of 40 cases of HL diagnosed ≥2 years following the diagnosis of IM. The RR of EBV-positive HL (N = 16) following IM was 2.8 (1.7–4.6), and for EBV-negative HL (N = 13), 1.1 (0.7–2.0). The estimated median latency from the diagnosis of IM to that of EBV-positive HL was 4.1 years. Thus the data regarding the predictive role of IM history with EBV positivity for HL is mixed in case-control studies. In the single cohort study, the association appears to be much more specific. One explanation for these inconsistencies may be that history of IM is a marker of two related phenomenon: the occurrence of late primary EBV infec-

tions, and susceptibility to late infections for agents that share similar transmission patterns with EBV. A next question concerns the relationship of EBV serology and the EBV status of tumors. In studies based on serologic profiles at or following diagnosis of HL, the findings are somewhat mixed. Ohshima et al. (1990) reported a case of an 8-year-old boy with EBV-positive HL. The detailed EBV serologic pattern for this child was consistent with an active EBV infection. In an overlapping series of 107 cases, Brousset et al. (1991) and Delsol et al. (1992) concluded there was no association with EBV positivity and a serologic pattern of reactivation that they defined as anti-VCA >1 : 640, anti-EA >1 : 40, anti-EBNA >1 : 160. However, only 1 of 35 EBV-negative cases and none of 16 EBVpositive cases had this extreme pattern. Levine et al. (1994) assessed the relation of EBV status for 39 cases with previously published serology. In this series, there were no differences between the EBVpositive and EBV-negative cases for either anti-VCA or anti-EA levels. Enblad et al. (1997) compared detailed EBV serology between 27 EBV-positive HL cases and 80 EBV-negative HL cases. They reported only one significant difference: the EBV-positive HL cases were more likely to have IgG antibodies against the EA-R. Axdorph et al. (1999) analyzed the EBV serology from 15 EBVpositive HL cases and 32 EBV-negative cases diagnosed in Stockholm. The EBV-positive cases had significantly higher titers against the EAR. Similarly, Alexander et al. (2001) evaluated EBV serology of nine EBV-positive HL cases and 39 EBV-negative cases. All of the EBVpositive HL cases were seropositive, whereas 27% of the EBVnegative cases were seronegative. They reported significantly higher titers against VCA in the EBV-positive cases, as well as higher titers against the EA. In our recent HL study, we conducted a case-case analysis to evaluate whether the EBV antibody profiles differed between 89 EBVpositive and 290 EBV-negative HL cases (Chang et al., 2004c). The two sets of patients had substantially different antibody profiles. Of note, all EBV-positive HL cases were seropositive, compared with 94% of the EBV-negative cases, in agreement with the report of Alexander et al. (2001). In comparison with the EBV-negative cases, the EBV-positive cases had significantly higher titers against VCA (IgG and IgA), EA, EA-D-IgA, EBNA2, and an anti-EBNA1/EBNA2 ratio £1.0. When all antibodies were considered simultaneously, the OR for elevated anti-VCA IgG was 4.7 (1.5–9.1), and for the low anti-EBNA ratio the OR was 3.1 (1.1–8.6). Our finding on a higher prevalence of a low anti-EBNA1/EBNA2 ratio in EBV-positive HL was validated by a companion study we conducted using pre-diagnosis specimens for HL patients identified in the Department of Defense Serum Repository (L. Levin, unpublished). This repository holds over 17 million specimens that have been collected from approximately 7.5 million uniformed service personnel since 1990 (Rubertone and Brundage, 2002). In this study, 139 HL cases were identified who had a stored serum specimen drawn prior to their diagnosis, and whose diagnostic tissue was tested for the presence of EBV. Of these, 40 cases were EBV positive. We attempted to match three controls to each case by age, gender, race-ethnicity, and date of index blood collection (±30 days). A total of 401 controls were identified. On average, the index specimens were collected 36 months (range 0 months to 8 years) prior to diagnosis of HL. The mean age at diagnosis was 24 years; 90% of the subjects were male. The ORs associated with elevated levels of antibodies against the VCA, EA, EBNA2, and a low anti-EBNA1/EBNA2 ratio were elevated for the EBV-positive HL cases in comparison with the controls. The OR of 4.7 for the anti-EBNA1/EBNA2 ratio was significant. In contrast, the ORs for the antibodies for the EBV-negative HL cases were not statistically significant. In addition, the detection of circulating free EBV DNA correlates with EBV status in HL. This was first reported by Gallagher et al. (1999), who found that 91% of 33 patients with EBV-positive HL had detectable serum EBV DNA, significantly more than the 23% of 26 patients with EBV-negative HL. They found that the serologic EBV DNA was resistant to treatment by DNase, and demonstrated it was free DNA not derived from circulating virions. Berger et al. (2001) found that all of three children with EBV-positive HL had detectable

Hodgkin Lymphoma serum levels of EBV DNA, whereas a child with EBV-negative HL had no detectable serum EBV DNA. We have confirmed this association in preliminary findings among a series of untreated patients. Of 18 EBV-positive HL cases, 72% had detectable virion-free serum EBV DNA; this was significantly higher than 7% of 30 EBV-negative cases (Lin et al., 2004). We found no significant difference in cellular EBV DNA levels. We also found that by testing for the presence of aberrant promoter methylation of p14 (a key gene alteration often present in tumors), that three of eight HL cases had such altered DNA in their circulation. These findings are consistent with the notion that at least part of the circulating EBV DNA in HL comes from tumor cells. In our recent prospective serologic study of HL conducted in the Department of Defense Serum Repository, we tested sequential prediagnostic serum EBV DNA in samples for 39 EBV-positive and 98 EBV-negative HL cases against that of matched controls. For the EBVpositive cases, there was a significant increase in the detection of serum EBV DNA as diagnosis was approached, OR = 8.6 (1.8–42.0). However, for EBV-negative HL cases, there was no significant increase, OR = 1.3 (0.84–2.1) (R. Ambinder, personal communication). These studies provide evidence that alterations in EBV serology and circulating EBV DNA generally track with the viral status of the HL tumor itself. Together, the data argue against our hypothesis that the EBV plays an etiologic role in all forms of the HL, but the viral genome is selectively lost from the tumor in some patients with better immunity (Mueller, 1997).

Other Factors Associated with Epstein Barr Virus Status The role of social environment and other factors in relation to EBV status of HL has been evaluated in several studies. In the prevalent series of 16 EBV-positive and 67 EBV-negative HL cases, Sleckman et al. (1998) compared the two groups for their childhood social environment characteristics; no significant differences were found. Flavell et al. (1999) compared 47 EBV-positive and 176 EBV-negative HL cases from a population-based study conducted in the United Kingdom for each group’s Townsend scores—a measure of material deprivation based on postal codes. They found that the EBV-positive cases had a higher mean score compared with the EBV-negative HL cases, p = 0.12. Alexander et al. (2000), in their study of 19 EBV-positive and 84 EBV-negative British HL cases aged 16–24 years, evaluated selfreported history of early infections. Both groups of cases reported significantly fewer early infections when compared with controls. The OR for total infections at age two or younger was 0.18 (0.03–0.95) for the EBV-positive cases and 0.43 (0.21–0.86) for the EBV-negative cases. In our recent case-control study conducted in eastern Massachusetts and Connecticut, we compared the risk factor data between 96 EBVpositive and 311 EBV-negative HL cases (Chang et al., 2004c). In our study, we found that there were no significant differences between the two groups of cases for childhood social factors—including nursery school and daycare experience. The only significant difference was for cigarette smoking; the OR for EBV-positive HL was 1.9 (1.2–3.2) in comparison with EBV-negative HL cases. Overall, it appears that EBV-positive and EBV-negative HL cases are generally similar in associations with factors related to delayed infectious exposure, at least among younger adults. This suggests that another infectious agent may be involved in EBV-negative HL (Jarrett et al., 1996). However, despite fairly extensive testing for known agents (Evans and Gutensohn, 1984; Levine et al., 1992; Cozen et al., 1998; Armstrong et al., 1998; Schmidt et al., 2000; Shiramizu et al., 2001) or novel herpesviruses (Gallagher et al., 2002), no other known virus has been implicated by serologic or molecular evidence. In terms of oncogenic pathways, there appears to be little difference in the expression of several major oncogenes between EBV-positive and EBV-negative HL (Ambinder and Weiss, 1999). Most importantly, the constitutive activation of the NFkB pathway, a key element in the pathogenesis and malignant behavior of HL of all subtypes (Izban et al., 2001), is independent of EBV status (Bargou et al., 1997; Hinz et al., 2001). NFkB is a potent transcription factor that activates a

885

cascade of inflammatory molecules (Barnes and Karin, 1997). The expression of NFkB-induced cytokine production characterizes the microenvironment of the RSC. It is also central to the proliferation and survival of RSC (Bargou et al., 1997). The upregulation of NFkB in EBV-positive HL is readily explained by the induction of NFkB by the EBV-LMP1 (Izumi and Kieff, 1997). How this transcriptional pathway is constitutively upregulated in EBV-negative HL is unknown. However, Jungnickel et al. (2000) have reported deleterious clonal somatic mutations in the IkBa gene in two of three EBVnegative HL tumors by single cell analysis. Such mutations could lead to constitutive activation of NFkB. No similar mutations were detected in two EBV-positive HL tumors. It is conceivable that within the genetically risky environment of the germinal center, Bcells could stochastically acquire such mutations by the cumulative antigenic stimulation of multiple infectious agents. It is also conceivable that the probability of these mutations may be enhanced by a dominant systemic type 2 cytokine profile at the time of a crucial infection.

OCCUPATIONAL AND ENVIRONMENTAL EXPOSURES There is very little evidence linking risk of HL to any occupation, specific occupational exposures, or environmental chemical exposures. The absence of any clear-cut occupational and environmental links to HL risk is not because it has been insufficiently studied. Numerous case-control studies have reported weak and/or inconsistent associations. The problem with these studies is that the disease is relatively uncommon and case-control studies in the past have been of limited size. There are many more reports of cohort studies of occupational exposures and risk of HL. However, these are limited by the low incidence of HL. Given the disease incidence of 2.8/100,000 person-years, 357,000 person-years of observation would yield only 10 cases of HL if there were no increased risk of the disease. Thus, a large cohort study of 10,000 workers followed for at least 35 years would yield a very small number of HL cases; most reported cohort studies are smaller than this. Further, the use of mortality from HL as an outcome measure in cohort studies likely limits the HL cases identified to those with poor prognosis, those without health care insurance, or those who are elderly. In earlier reviews, we (Grufferman and Delzell, 1984; Mueller, 1996) as well as McCunney (1999) concluded that of all occupation exposures studied, the findings are most consistent for an association between woodworking occupations and an increased risk of HL. However the findings have not been consistent in all studies (Table 45–5). In those studies demonstrating an increased risk of HL in woodworking occupations, the associations have generally been weak, and there are no well-described dose-response relationships. In 1998, Demers and Boffetta authored an International Agency for Research on Cancer (IARC) technical report on occupational exposure to wood dust based on a pooled analysis of epidemiologic studies. They reviewed data from six cohort studies of woodworkers in the United States and the United Kingdom to assess risk of a variety of causes of death including HL. They found an SMR of 0.63 (0.32–1.10) for HL in the pooled cohort, based on 12 observed vs. 19.2 expected deaths. They concluded that the evidence for an association between exposure to wood dust and HL was somewhat more suggestive, in that some case-control studies showed moderately high risks, but these results were not substantiated by the results of cohort studies or some of the well-designed case-control studies. (Demers and Boffetta, 1998). It is unlikely that there will be further revealing studies of the wood dust-HL association for several reasons. First, the production of wood products has undergone major changes, at least in developed countries, with much cleaner workplaces (and lower exposures) for both economic and industrial hygiene reasons. For example, the use of vacuum hoods over wood processing machines allows the capture of wood particulates for particle board (large particles), and as potential power sources (finer particles). The second most frequently considered occupational cause of HL is workplace exposure to “chemicals” (reviewed by Grufferman and

886

PART IV: CANCER BY TISSUE OF ORIGIN

Table 45–5. Studies of Hodgkin Lymphoma and Wood-Related Exposures

Study Area and Reference

Study Design (Time Period)

Upstate New York Milham and Hesser (1967)

Case-control (1940–1964)

Washington State Petersen and Milham (1974)

Case-control (1950–1971)

United States Milham (1974)

Mortality survey (1969–1970)

Boston Grufferman et al. (1976)

Subjects 1,549 white male HL deaths; 1,549 controls with other causes of death 707 male HL deaths; 707 controls with other causes of death Members of the Brotherhood of Carpenters and Joiners of America Union (16,443 deaths); referent rate: US age- and sex-specific mortality rates in 1968 1,577 incident HL cases; referent rate; census data on population 397 HL cases; 397 general population controls

Israel Abramson et al. (1978)

Incidence survey (1959–1973) Case-control (1960–1972)

Nortth Carolina Greene et al. (1978)

Case-control (1956–1974)

Denmark Olsen and Sabroe (1979)

Retrospective cohort (1971–1976)

Brazil Kirchoff et al. (1980)

Case-control (1976)

Italy Fonte et al. (1982)

Case-control (1972–1979)

Buckinghamshire, UK Acheson et al. (1984)

Retrospective cohort (••–1982)

Yorkshire, UK Bernard et al. (1987)

Case-control (1979–1984)

5,108 male furniture workers, mortality; referent rates; areacorrected England and Wales, age- and sex-specific mortality rates, 1968–1978 248 incident cases; 489 hospital controls

Boston-Worcester metropolitan area Matte and Mueller (unpublished)

Case-control (1973–1977)

181 incident male cases; 345 population controls

Milan area, Italy LaVecchia et al. (1989)

Case-control (1983–1988)

69 cases; 396 hospital controls

Connecticut and Eastern Massachusetts Chang et al. (2004a)

Case-control (1997–2001)

565 cases; 679 population controls

167 white male HL deaths; 334 controls with other causes of death 40,428 members of the Carpenters/Cabinet Makers’ Trade Union; referent rate: mortality of all Danish men 38 ever-employed HL cases; 42 ever-employed hospital controls with neoplastic diseases 387 HL cases; 771 hospital controls

(a) Exposure Definition (b) Source of Data

Relative Risk

P Value or 95% Confidence Interval

(a) Employment in wood-working or wood-related industries (b) Death certificates (a) Employment in wood-related industries (b) Death certificates, next-of-kin interviews for deaths in 1965– 1970 (a) Union membership (b) Union records

2.3

<0.001

1.8

<0.05

(a) Employment in wood-related industries (b) Medical records (a) Work with wood or trees (b) Interviews (a) Occupations with wood or paper exposure (b) Death certificates (a) Union membership (b) Union records

1.8

All Ages (0.16–1.2) ≥60 + yr (1.2–2.6)

1.6

(0.9–2.6)

0.9

1.1

Not stated Mixed Cell 5.2 0.0005 1.4 (0.8–2.5) Carpentry/Lumber 4.2 (1.4–15.0) 0.6 Not stated

(a) Occupational exposure to wood or wood products (b) Interviews

2.9

(0.7–12.0)

(a) Occupation in the wood industry (b) Medical records (a) Occupation within industry (b) Industry records

7.2

(2.3–22.2)

(a) Occupation as woodworker Industry employment as woodworker Self-reported contact with wood dust (b) Interviews (a) Occupation in paper and wood industry Self-reported wood-related occupation Self-reported exposure to dust or sawdust (b) Interviews (a) Self-reported exposure to wood dust for >10 years (b) Interviews (a) Self-reported occupational exposure to wood or wood products (b) Interviews

0.3 Not stated Other/unspecified lymphoma 3.5 (1.5–6.9) 1.1 0.8 1.0

Not stated Not stated Not stated

1.0 0.9 1.8

(0.5–1.9) (0.6–1.6) (1.2–2.7)

0.46

Not stated

Ages 15–54 yr (1.0–2.0) Ages 55–79 yr 2.3 (0.87–6.1) 1.4

Source: Adapted from Grufferman and Delzell, 1984.

Delzell, 1984; Mueller, 1996; McCunney, 1999). The reported findings are inconsistent in the specific chemicals, groupings of chemicals (e.g. “solvents”), and occupations involved, as well as in the strength and statistical significance of the reported associations. In conclusion, occupational exposures as causes of HL have been studied extensively and none has emerged as an established risk factor. If valid causal associations exist of substantial magnitude, they should

have been uncovered by now, given the large number of case-control and occupational cohort studies that have been done. Thus, new approaches are needed. Given the substantial evidence that underlying immune status is important in the etiology of this disease, De Sanjose et al. (2004) have recently presented an analysis from a large case-control study conducted in Spain, which takes a creative new approach. In their study,

Hodgkin Lymphoma they categorized lifetime occupational history on the basis of exposure to high molecular weight agents that are associated with asthma and act predominantly through an IgE response. These include occupational exposures such as latex, flour, and animal proteins. They found that the OR associated with these exposures for HL was OR = 2.9 (1.2–6.6), with evidence for a dose response based on length of occupational exposure. If confirmed in other populations, these findings provide new insight into this disease. In future studies, it would be important to evaluate the role of EBV status of the tumors as well. Finally Smedby et al. (2005) evaluated whether HL was associated with exposure to ultraviolet (UV) light via self-reported sunlight exposure. They found that various measures of exposure to UV light were associated with a significantly lower risk of HL. For example, the RR among those who sunbathed four or more times per week five to ten years ago, compared to those who never sunbathed, was 0.7 (0.5–1.0) and that for individuals sunburned at least two times per year five to ten years ago, compared to those who were never sunburned, was also 0.7 (0.4–1.0). An analysis of cancer incidence among nearly 324,000 Swedish men who worked in the construction industry similarly found a significantly reduced incidence of HL among those men with high sunlight exposure, RR = 0.3 (0.1–0.9) based on four such exposed cases (Hakansson et al., 2001). Paradoxically, it has been reported that HL occurs more than expected among individuals with a prior nonmelanoma skin cancer (Levi et al., 1997, Kahn et al., 1998). Similarly paradoxical findings have been found for NHL. For HL, these associations need further validation, but if true, point to multiple carcinogenic properties of UV light exposure. The inverse association between sun exposure and HL risk could be mediated through an anticarcinogenic effect of vitamin D synthesized in the skin in response to UV radiation (Egan et al., 2005).

HOST FACTORS Familial Aggregation and Genetics Familial aggregation and genetic susceptibility appear to play important roles in the etiology of HL. Razis et al. (1959) were the first to try to assess the increased risk of HL in first-degree relatives of HL cases. They estimated about a threefold increased risk of HL in firstdegree relatives. We recently re-analyzed the Razis data and found an RR of 4.8 (2.1–11.3) for HL occurrence in first-degree relatives (Mueller and Grufferman, 1999). In Israel, Haim et al. (1982) estimated the risk of HL in first-degree relatives of HL cases was increased ninefold in a study with limited information on family size. Paltiel et al. (2000), in a registry-linked study in Israel, reported a standardized incidence ratio (SIR) of 2.6 (0.84–6.1) for HL in first-degree relatives of probands. They also found an SIR of 3.1 (1.0–7.3) for siblings of cases. This study also had incomplete ascertainment of probands’ relatives. Our earlier Boston study (Grufferman et al., 1977) examined the risk of HL in siblings of cases and found a relative risk of 7.1 (1.6–30) for siblings of cases diagnosed before 45 years of age. We found no increased risk for siblings of cases diagnosed at age 45 years or later. In the series of 13 HL sibling pairs, we found a ninefold RR for like-gender sibling pairs and a fivefold RR for discordant gender pairs. This pattern of excess concordance in gender of sibling pairs with HL is difficult to explain on a simple genetic basis. It is more likely that this increased risk could be due to shared environmental exposures. It is likely that same-gender siblings are more apt to have shared bedrooms, friends, and to have spent more time together. Although this finding of an excess of same-gender sibling pairs in HL has not been confirmed by others (Chakravarti et al., 1986), similar significant excesses of like-gender sibling pairs have been reported for sarcoidosis, Behcet disease, and multiple sclerosis, a group of diseases that are suspected of having infectious etiologies (Grufferman et al., 1987). Horowitz and Wiernek (1999) propose the interesting hypothesis that segregation patterns for pseudoautosomal region linked genes of the X and Y chromosomes are predicted to result in affected siblings tending to be of the same gender. This mechanism

887

might underline the apparent excess gender concordance observed for HL. Strong evidence of genetically determined familial aggregation of HL was provided by the 1995 report of Mack et al. They found that among 179 monozygotic twin pairs with at least one twin affected with HL, there were 10 pairs who became concordant for HL, SIR = 99 (48–182). None of 187 pairs of dyzygotic twins with one affected member became concordant for HL. The authors suggested that the absence of an increased HL risk in dyzygotic twins of cases was due to much lower genetic predisposition and to a low risk of HL in their relatively small population of twins. However, it is curious that such a greatly increased risk was found for monozygotic twins but no increased risk for dyzygotic twins, with comparable numbers of the two-type of twin pairs. This study employed an unusual approach to identify subjects that could have introduced bias. They relied on advertisements in newspapers and magazines in the United States and Canada to identify “twins with cancer”, rather than use any systematic population-based method of ascertainment. Whether this subject identification process could have led to selective reporting of doubly affected monozygotic twin pairs cannot be evaluated. Nevertheless, it is unlikely that bias alone could have produced the remarkably high SIR observed. This same group recently published a follow-up study among the discordantly affected monozygotic twins, looking at the effect of IL6 genotype on HL risk (Cozen et al., 2004). Using the 88 unaffected twins (“surrogate cases”) or young adult HL patients, they determined their genotypes for the IL-6 174 G > C promoter polymorphisms, the CC genotype being the low IL-6 secreting variant. These genotypes were compared with that of 87 matched controls, who were non-blood relatives of the twins. They found that risk of HL decreased with an increasing number of C alleles; the OR for the CC genotype vs. the GG high-secreting genotype was 0.29 (0.10–0.87). This finding suggests that genetically determined higher levels of IL-6 (type 2) cytokine may increase susceptibility of developing HL. A more recent large study of cancer in twin pairs was done using twin registry data for Sweden, Denmark and Finland (Lichtenstein et al., 2000). Despite the very large size of this study, only 32 monozygotic twin pairs with one twin affected by HL were found. There were no concordant pairs for HL; based on the proportion observed by Mack et al. (1995) (10/179), about two concordant pairs would have been expected. There were no concordant HL pairs among 31 dyzygotic twin pairs with one affected twin. Although the numbers of affected twin pairs are very small, this study did not confirm the Mack et al. (1995) findings. Preliminary results from a large case-control study of childhood HL (diagnosis < 15 years of age) currently in progress, show that firstdegree relatives of children with HL had a 2.7-fold increased risk of all cancers (Grufferman et al., 1998a). In a series of 464 HL cases and 699 individually matched controls, 29 cases and 17 controls had a firstdegree relative with cancer. Four cases but no controls had parents with HL. There also appeared to be an increased risk associated with all lymphoreticular malignancies, melanoma, and testicular cancer in case families. A previous study by Olsen et al. (1995) used cancer registry data in Denmark to evaluate cancer risks in parents of childhood cancer patients. They found no increased occurrence of cancer in case parents overall or for HL. In our recent case-control study conducted in the greater Boston area and the state of Connecticut, family history of hematologic malignancy appeared to be associated with an increased risk of HL among younger adults (15–54 years), OR = 2.1 (1.1–3.9), and those older, OR = 1.8 (0.58–5.4) (Chang et al., 2004a). In summary, there have been frequent reports of familial aggregation of HL and other cancers. Such reports have generally been of selected kindreds with no precise quantification of risks of other cancers in HL relatives. Thus far, the only estimate of overall cancer risk in close relatives of HL cases is from our childhood HL casecontrol study of Grufferman et al. (1998a) discussed above. While we have found a 2.7-fold increased risk of all cancers in first-degree relatives, removing the four cases of HL in cases’ relatives (there was none in the control families) results in an odds ratio of 2.3. The most frequently reported cancer, other than HL, in close relatives of HL

888

PART IV: CANCER BY TISSUE OF ORIGIN

cases has been NHL (Bjerrum et al., 1986; Linet and Pottern, 1992; Lynch et al., 1992; Shpilberg et al., 1994; Siebert et al., 1997; Paltiel et al., 2000). Although many of the family links between HL and NHL are the result of studies that focused on all lymphomas for analysis and thus may be biased, nevertheless, it is likely that close relatives of HL patients are at increased risk of other lymphomas. Sporadic case reports have also linked HL with a variety of other cancers. These data suggest that it would be extremely valuable to collect information on the occurrence of all cancers in relatives in future studies of familial aggregation in HL. The various studies of familial aggregation of HL have found significantly elevated risk in close family members. The magnitude of the increased risk varies with the age at diagnosis and the type of family relationship. The risk is highest for a monozygotic twin of a case (99fold increase), but may not be elevated for a dyzygotic twin. Non-twin siblings of cases appear to have between a threefold to sevenfold increased risk of HL with the higher estimate being for siblings diagnosed while under 45 years of age. For all first-degree relatives of cases the level of increased HL risk varies from 2.6- to 8-fold. For first-degree relatives of childhood HL cases the risk is probably quite high but is presently undefined. As an estimate of the overall risk of HL in first-degree relatives of HL probands, we believe that it is on the order of threefold, as was initially reported by Razis et al. (1959). These findings of familial aggregation underline the notion that genetic factors may play a role in the etiology of HL. Of note, as first reported by MacMahon (1957), Jewish people tend to have a higher incidence of HL. This was confirmed in our early Boston-area casecontrol study (Gutensohn and Cole, 1981), and our more recent study that also included Connecticut (Chang et al., 2004a). It has also been reported in large case-control studies in Great Britain (Bernard et al., 1984, 1987) and in Los Angeles (Cozen et al., 1992). Although there is some heterogeneity among these studies by age group, it appears that Jews are at somewhat greater risk of HL. There have been few reports on the genetics of HL. Many formal genetic approaches focus on multiply affected kindreds with a disease as the basis of their research. HL is an uncommon disease and the occurrence of multiply affected families is even more uncommon. Thus, it becomes a major undertaking to assemble a large number of families with multiple occurrences of HL. The studies of HL genetics reported include those on linkage and complex segregation, genetic anticipation, and HLA genotypes. Chakravarti et al. (1986) performed linkage analysis on 41 pedigrees with multiple HL occurrences. They found strong evidence of a recessive susceptibility gene that was linked closely to the HLA complex and was responsible for 60% of cases in multiplex families. They attributed the residual 40% to other familial and/or environmental factors. They did not find sibling sex concordance in their families but found increased concordance for histologic subtype of HL. Concordance for histologic subtype might be related to the repeated observation that most of the reported multiply affected families involve young adults—a group in which the majority of cases have the NS subtype. The authors concluded that there is etiologic heterogeneity in HLA with at least three independent determinants: an HLAlinked gene, an HLA-unlinked factor, and an environmental/genetic factor determining concordance in histologic subtype. A genetic analysis of HL was done by Yang et al. (1993) using data from the COG study of childhood HL epidemiology. Subjects were 336 HL probands, aged less than 15 years at diagnosis. There were data on 4781 family members, among whom there were three first-degree and seven second-degree relatives with HL. Complex segregation analysis was performed and three hypotheses were evaluated: 1. The occurrence of HL in relatives is sporadic 2. Genetic predisposition accounts for familial occurrence 3. Environmental factors underlie the observed familial aggregation. They found that hypothesis three best explained the observed data. Other data showed that the occurrence of HL “at younger vs. older age may be determined by different combinations of environmental and/or genetic factors.” This preliminary study was based on a relatively large

series of HL cases, yet the number of multiply affected families was quite small, illustrating the difficulty of doing genetic studies of HL. There have been many studies of the relationship between HLA genotypes and HL risk; such studies are of interest because HLA types are related to immune responsiveness. Hors and Dausset (1983) reviewed such early studies. They concluded there was a slightly increased risk of HL associated with the HLA antigens Al, B5, and B18. Persons with these genotypes had a RR ranging from 1.3–1.5. Later studies of HLA and HL have produced a confusing array of results. Some studies found HLA associations with only certain histologic subtypes of HL and others did not. However, there appears to be consistency in the findings of an association with HLA-A1 and to a lesser degree with HLA-B5, -B8, and -B18. Oza et al. (1994) reported the results of a pooled analysis of 741 HL cases and 686 controls from 17 centers. Using approaches allowing them to define specific alleles, they found an RR of 1.95 (p < 0.01) for the association with HLA-DPB1 *0301 in white patients. However, there were significant reductions in the frequency of HLA-DPB1 *0401 in patients from Japan and Taiwan (RR = 0.15, p < 0.01). (They also found shorter treatment remissions in patients with HLA-DPB1 *0901 overall (p < 0.05), but particularly in Japan and Taiwan (p = 0.02) where this type is most prevalent.) Harty et al. (2002) evaluated the relationship of HLA and the transporter associated with antigen processing (TAP) loci in familial HL. Their subjects included 100 members of 16 families, in which at least two members had confirmed HL. Using the transmission disequilibrium test, they found linkage disequilibrium with familial HL, particularly for the NS subtype, for the DRB1*1501-DQA1*0102DQB1*0602 haplotype, the TAP 1 allele encoding I1e at residue 333, and the DRB5-0101 allele. There have also been studies to determine whether the EBV status of patients’ tumors correlated with HLA types. This was initiated because HLA-A *0201 is known to be associated with cytotoxic T-cell responses to the LMP2 protein of the virus in healthy seropositive people (Bryden et al., 1997). However, no associations between EBV status and HLA-A2 were found in two studies (Poppema and Visser, 1994; Bryden et al., 1997). However, recently Diepstra et al. (2005) undertook a large population-based study to evaluate the role of HLA in risk of HL. The study population included 200 HL cases diagnosed from 1987–2000 in northern Netherlands with first degree relatives and spouses as controls. They genotyped 33 microsatellite markers spanning the entire HLA region on chromosome 6. They found that an HLA class I region defined by two consecutive markers differed significantly between the 54 EBV-positive cases as compared to the controls, as well as to the EBV-negative cases. Since EBV-positive RSC are much more likely to express HLA class I than EBV-negative RSC, their finding suggests that antigenic presentation of EBV peptides play a role in the pathogenesis of EBV-positive HL. They also identified an HLA class III region haplotype that was significantly more common in EBVnegative HL cases. If these findings are confirmed in other populations, it could lead to a clarification of the role of HLA-related genes in HL. Genetic anticipation has also been reported in HL multiply affected families. Anticipation is a phenomenon in which an inherited disease is diagnosed in each successive generation of a family at earlier ages and with more severe symptoms. It has been found in some inherited neurologic diseases as well as in NHL and chronic lymphocytic leukemia (Lynch, 2000; Wiernik et al., 2000, 2001). Shugart (1998) reported a study of 30 parent-child pairs with HL. In all but one of the pairs, HL was diagnosed at a younger age than it was in the parent. The mean age at diagnosis was 21 years in children and 46 years in parents, which was statistically significant (p < 0.0001). Shugart et al. (2000) used a more systematic approach based in the Swedish cancer registry, which confirmed genetic anticipation in HL. They estimated that about 28% of HL in the Swedish population was due to genetic factors. The underlying biologic basis of anticipation in HL is unclear, if it is real. Reliance on published pedigrees of HL for studying anticipation might be inappropriate due to many opportunities for report-

Hodgkin Lymphoma ing bias. Further evaluation of anticipation in large linked cancer registries is needed. Another approach to the study of HL genetics was reported by Horwitz and Wiernik (1999). Reports of the molecular cloning of a rare disease, Leri-Weill dyschondrosteosis (LWD), which had been linked anecdotally with HL in a family, and also had exhibited sibling gender concordance in affected pairs, led to this study. The gene responsible for LWD localizes to the short-arm pseudoautosomal region (PAR) of the X and Y chromosomes. A unique segregation pattern for PAR-linked genes has been predicted (i.e., that affected siblings will tend to be of the same gender). The authors tested the hypothesis that there is a PAR-localized gene for HL. They also reevaluated evidence for HLA linkage in HL using genotyped sibling pairs. They found that “the putative PAR- and HLA-linked loci account for 29% and 40% respectively of the heritability of HL in an American population.” These interesting observations need confirmation by other studies. Two studies have examined the relationship between the EBV and familial HL (Schlaifer et al., 1994; Lin et al., 1996). In their study the prevalence of EBV positivity of familial cases was the same as that in general case series. No evidence of concordance for EBV positivity in affected relative pairs was found. Thus, there appears to be no excess (or deficit) of EBV positivity in familial HL cases’ tumors and no concordance of positivity in affected pairs. In summary, relatively little is known about the genetics of HL, as opposed to familial aggregation. Although it is quite evident that relatives of HL cases are at modestly to highly increased risk of the disease, the underlying mechanisms are unknown. The great difficulty in studying the genetics of HL is the extreme rarity of multiply affected families. There is great need for large studies, using systematically ascertained families to avoid biases, to address the question of why close relatives of HL cases are at such high risk of the disease.

Immune Function Unlike most other lymphomas, HL tissues characteristically express high levels of a range of cytokines, chemokines, and their corresponding receptors (Skinnider and Mak, 2002). Many of these molecules are believed to act as autocrine growth factors that enhance RSC growth and/or survival, as well as attract the reactive cellular infiltrate that surrounds RSCs. Studies in HL cell lines and primary HL tissues have identified several cytokines that appear to be consistently elevated in HL compared with normal or non-HL cells and specimens. Most of the cytokines found in HL are associated with a type 2 immune response (Skinnider and Mak, 2002). The major cytokines found in HL include IL-10, an immunosuppressive cytokine that inhibits T-cell growth, blocks Th1 cell cytokine production, down regulates pro-inflammatory cytokine production, stimulates B-cell growth and differentiation (Moore et al., 2001), and is increased more often in EBV-positive than EBV-negative HL cases (Ohshima et al., 1995; Herbst et al., 1996; Wolf et al., 1996). IL-13, a type 2 cytokine that stimulates B-cell proliferation and survival, and triggers Ig class switching to IgG4 and IgE (Skinnider et al., 2002b), is expressed along with its receptor in almost all HL cells and tissues examined. It appears to be an important autocrine growth factor for RSCs (Kapp et al., 1999; Ohshima et al., 2001; Oshima and Puri, 2001; Skinnider et al., 2001, 2002a). IL-5, a type 2 cytokine necessary for eosinophil proliferation, differentiation, and activation (Sanderson, 1992), is also expressed in RSC lines and tumors, particularly in association with tissue eosinophilia (Samoszuk and Nansen, 1990; Klein et al., 1992; Kapp et al., 1999). IL-6, another type 2 cytokine that induces plasma cell differentiation, stem cell hematopoiesis, and acute phase reactions (Kishimoto et al., 1995), is also elevated, along with components of the IL-6 receptor complex, in the majority of HL cell lines and primary HL samples (Jucker et al., 1991; Merz et al., 1991; Hsu et al., 1992; Klein et al., 1992; Foss et al., 1993; Wolf et al., 1996; Herbst et al., 1997; Beck et al., 2001). Herbst et al. (1997) reported that IL-6 expression was significantly higher in EBV-positive HL cases than EBV-negative cases.

889

IL-12, a type 1 cytokine that promotes Th1-cell differentiation (Gately et al., 1998), has also been found in a majority of HL tissues, although it appears to be expressed by reactive cells rather than RSCs themselves. Again, IL-12 levels are more commonly elevated in EBVpositive than EBV-negative HL cases (Schwaller et al., 1995). Finally, IgE, an antibody isotype associated with allergy, atopy, and type 2 immunity, is increased in the serum and tumor tissue of HL cases compared with controls and non-HL patients (Amlot and Green, 1978; Amlot and Slaney, 1981; Samoszuk and Ramzi, 1993; Di Biagio et al., 1996). In summary, it is clear that a type 2 immune environment predominates in HL lesions as well as systemically in diagnosed patients. This is particularly seen in EBV-positive tumors. However, it is unknown whether this state precedes and influences the development of HL or is simply a correlate of the lymphoma.

Associated Medical Conditions There are a number of medical conditions associated with an increased risk of HL. These include certain primary immunodeficiences, including severe combined immunodeficiency, hypogammaglobulinemia, and hyperimmunoglobulin M syndrome (Filipovich et al., 1992; Mueller, 1999). HL is also associated with prior solid organ and allogeneic bone marrow transplantation (Bierman et al., 1996; Garnier et al., 1996; Rowlings et al., 1999; Baker et al., 2003). Most notably, HL is now recognized as occurring excessively among HIV-I infected patients (Goedert, 2000; Frisch et al., 2001; Dal Masco and Franceschi, 2003). The majority of HL occurring in immunodeficiency is EBV-positive. Hodgkin lymphoma is also increased among patients with autoimmune diseases including ulcerative colitis (Palli et al., 2000), systemic lupus erythematosis (Sultan et al., 2000), and rheumatoid arthritis patients (Mariette et al., 2002); and it has been reported to occur in kindreds with autoimmune lymphoproliferative syndrome at an excess rate (Straus et al., 2001). That all of these conditions involve some level of immune dysregulation provides evidence to the assertion that immune dysregulation plays a central role in the etiology of HL. The question of whether tonsillectomy is a risk factor for HL was addressed in a number of early studies. Tonsillectomy is a risk factor for two diseases that share epidemiological characteristics with HL; namely, poliomyelitis (Paffenbarger and Wilson, 1955) and multiple sclerosis (Poskanzer, 1965). In the past, tonsillectomy rates in the United States were found to vary inversely with family size and other correlates of social class (Gutensohn et al., 1975). In four studies that used sibling controls to control these potential confounding factors, the ORs ranged from 0.8 to 3.6 among young adults (Johnson and Johnson, 1972; Cole et al., 1973; Vianna et al., 1974; Gutensohn et al., 1975; Mueller et al., 1987). These estimates have varied by sibship size. Taken together, the variability of findings, the geographic variability of the practice of tonsillectomy that does not correlate with HL rates, and the potential confounding with known risk factors for HL, argue against a causal association. Similarly, appendectomy has been hypothesized to lead to an increased risk of HL and other lymphomas because it involves removal of the appendix and surrounding lymphoid tissue usually at a young age. Several studies of this possible association were done, leading to inconsistent results. This hypothesis was recently re-evaluated in a large record-linkage cohort study from Sweden (Cope et al., 2003). Among all children undergoing appendectomy (N = 106,763) there were 25 cases of HL for an SIR of 0.96 (0.6–1.4). Thus, it is also unlikely that appendectomy affects the risk of HL.

Other Factors Although cigarette smoking has been causally linked to many cancers, the possible association with HL risk has received limited scientific attention. Noting that results of early studies have been mixed, Briggs et al. (2002) evaluated the possible association in a case-control study based on the Selected Cancers Cooperative Study data. They identified 308 HL cases in the study cohort and frequency matched them to

890

PART IV: CANCER BY TISSUE OF ORIGIN

1796 controls selected by random digit telephone dialing. They found an OR of 1.3 (1.0–1.8) for ever smokers and 1.8 (1.4–2.5) for current smokers. Of note, they found a much stronger association for MC-HL, OR = 2.2 (1.2–4.1), than for NS-HL, OR-1.2 (0.8–1.6). There was a very striking trend with increasing pack-years of smoking (p < 0.001) that was stronger for the MC-HL. Since EBV positivity is higher in MC-HL, this finding is consistent with our report that smoking is a risk factor of EBV-positive HL noted above (Chang et al., 2004c), and underlines the utility of evaluating risk factors in relation to EBV status of tumors. Potential dietary risk factors for HL have been assessed, particularly in a series of hospital-based case-control studies from Italy (Negri et al., 1991; Tavani et al., 1997; 2000; Chatenoud et al., 1998). A variety of weak negative and/or inconsistent positive associations between multiple dietary components and HL risk were found, suggesting that these are perhaps chance findings. Wolk et al. (2001) conducted a prospective study of obesity and cancer incidence in Sweden. They found an increased risk among men for HL, RR = 3.3 (1.4–6.3) based on eight cases; however, no increase was seen among women, RR = 0.9 based on four cases. Unlike most malignancies, HL is unusual in that there is no apparent association with ionizing radiation exposure (Halnan, 1988; Boice et al., 1996). This finding underlines the notion that the pathogenesis of HL may be unique. Glaser (1994) has proposed that the deficit of HL cases among women of childbearing age may reflect a protective effect of parity. A population-based study in Norway reported a decreased risk (54%) among women with at least three births, compared with those with fewer, and was not modified by control of SES (Kravdal and Hansen, 1993). However, this finding was not confirmed in other studies (Franceschi et al., 1994; Zwitter et al., 1996). In terms of medication use, aspirin may be protective against HL because it selectively inhibits NFkB (Kopp and Ghosh, 1994; Grilli et al., 1996; Yin et al., 1998). To test this hypothesis, we investigated self-reported analgesic use among HL cases and population controls in our recent study in the greater Boston area and the state of Connecticut (Chang et al., 2003b). Regular use of non-steroidal anti-inflammatory drugs (NSAIDs) is associated with reduced risk of several cancers. NSAIDS may prevent cancer development by blocking the synthesis of pro-inflammatory prostaglandins by the cyclooxygenase (COX) enzymes. Aspirin is unique among the NSAIDs in that its binding to COX is irreversible (Van der Ouderaa et al., 1980), in addition to its specific inhibition of NFkB transcription. We defined regular use of aspirin, non-aspirin NSAIDs, and acetaminophen as ≥2 tablets per week over the last 5 years. We found that regular aspirin use was significantly associated with a reduced risk of HL, OR = 0.60 (0.42–0.85). In contrast, non-aspirin NSAIDs use was not associated with disease risk, OR = 0.97 (0.73–1.3). However, regular use of acetaminophen was associated with a significantly increased risk of HL, OR = 1.72 (1.3–2.3). These ORs are mutually standardized, and did not vary significantly across levels of age, gender, EBV status of tumor, the presence of B symptoms, or the interval of time between diagnosis and interview among cases. These findings must be confirmed in other investigations.

PATHOGENESIS The crucible in which the pathogenesis of HL occurs is the germinal center of lymph nodes and spleen (Küppers et al., 1999). Here, naïve B cells that have recognized an antigen undergo vigorous proliferation and genetic modification in the differentiation of specificity of the B-cell receptors, driven largely by B-cell stimulatory factors (type 2). The processes that induce these changes include somatic hypermutation and class switching, creating a “risky” environment for B-cells. Classic HL derives from “crippled” germinal center B-cells that had acquired somatic Ig gene mutations, but fail to undergo apoptosis (Küppers, 2003). What appears to trigger the failure of apoptosis in these genetically damaged HL-precursor cells is the constitutive expression of the NFkB transcription pathway. NFkB activation likely

contributes to or accounts for the bizarre cytokine and cellular microenvironment of the subsequent RSC. The presence of EBV and the expression of the LMP1 protein can explain NFkB activation in EBV-positive HL. What other factors have a hand at the switch—so to speak—in EBV-negative HL are unknown. In terms of epidemiology, a consistent thread for much of HL risk, at least for those cases diagnosed through the middle years of life, is the effect of social and cultural factors that influence the probability of early infectious exposure. The profile of these factors varies between generations and populations, but the effect is apparently the same; that is, other than for the youngest children, early infectious exposure is generally protective against HL, with subsequent susceptibility to “late” infections increasing risk of HL. This finding, if causal, likely operates at two levels of pathogenesis. The first is the effect on the systemic immune cytokine balance, biasing toward a type 2 profile or otherwise impairing the system. the second is the delayed triggering infection with EBV or other oncogenic agents itself, the effects of which may be enhanced by the state of systemic immunity. As in other cancers, the induction of HL is a stochastic process of acquired genetic changes. It is likely that other factors that influence the biology of the germinal center may play a role. These include genetic factors that influence the level of cytokine production or genes that may modify the fidelity of DNA repair or methylation. Unhealthy diet and physical inactivity may also play a minor role. The use of aspirin and other drugs that downregulate the NFkB pathway may prove to be important but do not appear to be candidates for population interventions.

FUTURE DIRECTIONS Future research on HL should rely on the new WHO classification of HL and exclude cases of LP-HL, which are more properly classified with NHL. Another axis for classification of HL for research analyses should be the EBV status of cases’ tumors; this will necessitate larger studies to have adequate statistical power for stratified analyses and assessment of interactions. The most important challenge facing HL researchers is determining the cause(s) of EBV-negative tumors. The epidemiologic evidence suggests that an infectious agent may well be involved, and many known agents have been eliminated as suspects. The application of more novel molecular methods to identify new agents, which was successfully used to discover the Kaposi sarcoma herpesvirus (Chang et al., 1994) should be pursued. Genetic susceptibility and attendant familial aggregation likely play important roles in HL etiology. Multicenter pooled studies of multiply affected families that have sufficient numbers of subjects for genetic analyses are needed. Such studies should also examine familial aggregation of HL with other lymphoreticular cancers. Characterization of the immune milieu that precedes and accompanies the occurrence of HL should aid the clarification of the role of immunity itself. Development and validation of serologic biomarkers of systemic immune status would facilitate epidemiologic research in this area. An area of growing importance, as survival in HL patients continues to improve, is that of late effects of therapy, particularly second malignancies. Cured HL patients have remarkably high risks of second neoplasms, which are often fatal. Such studies should consider the risk factors for HL itself in relation to second cancer occurrence. Although much progress has been made in understanding this fascinating lymphoma, major questions persist. Given the rarity of the disease, and the need to consider heterogeneity of risk factors by EBV status and age, collaboration among investigators is a necessity. A venue for such collaboration is offered by Interlymph, a consortium of epidemiologic studies on lymphoma, which has been organized by IARC and the National Cancer Institute. Acknowledgments We are indebted to Drs. Richard Ambinder, Otoniel Martinez-Maza, and Ellen Chang for their thoughtful contributions to our understanding of this disease.

Hodgkin Lymphoma We are also indebted to Ms. Mary Fronk and Judith Kadosh for their outstanding editorial assistance. This work was supported by the following grants from The National Cancer Institute: P01CA069266 and R01CA47473.

References Abramson JH. 1974. Childhood experience and Hodgkin’s disease in adults: An interpretation of incidence data. Br J Med Sci 10:1365–1370. Abramson JH, Pridan H, Sacks MI, Avitzour M, Peritz E. 1978. A case-control study of Hodgkin’s disease in Israel. J Natl Cancer Inst 61:307–314. Acheson ED, Pippard EC, Winter PD. 1984. Mortality of English furniture makers. Scand J Work Environ Health 10:211–217. Alderson MR, Nayak R. 1971. A study of space-time clustering in Hodgkin’s disease in the Manchester region. Br J Prev Soc Med 25:168–173. Alderson MR, Nayak R. 1972. Epidemiology of Hodgkin’s disease. J Chron Dis 25:253–259. Alexander FE. 1990. Clustering and Hodgkin’s disease. Br J Cancer 62:708–711. Alexander FE, Daniel CP, Armstrong AA, et al. 1995. Case clustering, EpsteinBarr virus Reed-Sternberg cell status and herpes virus serology in Hodgkin’s disease: Results of a case-control study. Eur J Cancer 31A:1479–1486. Alexander FE, Jarrett RF, Cartwright RA, et al. 2001. Epstein-Barr virus and HLA-DPB1-*301 in young adult Hodgkin’s disease: Evidence for inherited susceptibility to Epstein-Barr virus in cases that are EBV+ve. Cancer Epidemiol Biomarkers Prev 10:705–709. Alexander FE, Jarrett RF, Lawrence D, et al. 2000. Risk factors for Hodgkin’s disease by Epstein-Barr virus (EBV) status: Prior infection by EBV and other agents. Br J Cancer 82:1117–1121. Alexander FE, Lawrence DJ, Freeland J, et al. 2003. An epidemiologic study of index and family infectious mononucleosis and adult Hodgkin’s disease (HD): Evidence for a specific association with EBV+ve HD in young adults. Int J Cancer 107:298–302. Alexander FE, McKinney PA, Williams J, Ricketts TJ, Cartwright RA. 1991a. Epidemiological evidence for the two-disease hypothesis in Hodgkin’s disease. Int J Epidemiol 20:354–361. Alexander FE, Ricketts TJ, McKinney PA, Cartwright RA. 1991b. Community lifestyle characteristics and incidence of Hodgkin’s disease in young people. Int J Cancer 48:10–14. Alexander FE, Williams J, McKinney PA, Ricketts TJ, Cartwright RA. 1989. A specialist leukaemia/lymphoma registry in the UK. Part 2. Clustering of Hodgkin’s disease. Br J Cancer 60:948–952. Ambinder RF, Browning PJ, Lorenzana I, et al. 1993. Epstein-Barr virus and childhood Hodgkin’s disease in Honduras and the United States. Blood 81:462–467. Ambinder RF, Mann R. 1994. Epstein-Barr-encoded RNA in situ hybridization. Hum Pathol 25:602–605. Ambinder RF, Weiss LM. 1999. Association of Epstein-Barr virus with Hodgkin’s disease. In: Mauch PM, Armitage JO, Diehl V, Hoppe RT, Weiss LM, eds. Hodgkin’s Disease. Philadelphia: Lippincott Williams and Wilkins, pp. 79–98. American Cancer Society. 2005. Cancer Facts and Figures. Atlanta: American Cancer Society. Amlot PL, Green LA. 1978. Atopy and immunoglobulin E concentrations in Hodgkin’s disease and other lymphomas. Br Med J 1:327–329. Amlot PL, Slaney J. 1981. Hypergammaglobulinaemia E in Hodgkin’s disease and its relationship to atopy or a familial predisposition to atopy. Int Arch Allergy Appl Immunol 64:138–145. Anagnostopoulos I, Hansmann M-L, Franssila K, et al. 2000. European Task Force on Lymphoma project on lymphoctye predominance Hodgkin disease: Histologic and immunohistologic analysis of submitted cases reveals two types of Hodgkin disease with a nodular growth pattern and abundant lymphocytes. Blood 96:1889–1899. Aozasa K, Ueda T, Tamai M, Tsujimura T. 1986. Hodgkin’s disease in Osaka, Japan (1964–1985). Eur J Cancer Clin Oncol 22:1117–1119. Armstrong AA, Shield L, Gallagher A, Jarrett RF. 1998. Lack of involvement of known oncogenic DNA viruses in Epstein-Barr virus-negative Hodgkin’s disease. Br J Cancer 77:1045–1047. Aterman K. 1986. Thomas Hodgkin (1798–1866). Am J Dermatopathol 8:157–167. Axdorph U, Porwit-MacDonald A, Sjöberg J, et al. 1999. Epstein-Barr virus expression in Hodgkin’s disease in relation to patient characteristics, serum factors, and blood lymphocyte function. Br J Cancer 81:1182–1187. Baker KS, Defor TE, Burns LJ, Ramsey NKC, Neglia JP, Robison LL. 2003. New malignancies after blood or marrow stem-cell transplantation in children and adults: Incidence and risk factors. J Clin Oncol 21: 1352–1358.

891

Ball TM, Castro-Rodriguez JA, Griffith KA, Holberg CJ, Martinez FD, Wright AL. 2000. Siblings, day-care attendance, and the risk of asthma and wheezing during childhood. N Engl J Med 343:538–543. Bargou RC, Emmerich F, Krappmann D, et al. 1997. Constitutive nuclear factor-kB-RelA activation is required for proliferation and survival of Hodgkin’s disease tumor cells. J Clin Invest 100:2961–2969. Barnes PJ, Karin M. 1997. Nuclear factor-kB- A pivotal transcription factor in chronic inflammatory diseases. N Engl J Med 336:1066–1071. Barton DE, David FN, Fix E. 1966. Tests for space-time interaction and a power function. Proceedings of the 5th Berkeley Symposium on Mathematical Statistics and Probability. Berkeley, University of California Press. Beck A, Pazolt D, Grabenbauer GG, et al. 2001. Expression of cytokine and chemokine genes in Epstein-Barr virus-associated nasopharyngeal carcinoma: Comparison with Hodgkin’s disease. J Pathol 194:145–151. Begg C, Zhang Z. 1994. Statistical analysis of molecular epidemiology studies employing case-series. Cancer Epidemiol Biomarkers Prev 3:173–175. Benenson AS, ed. 1995. Control of Communicable Diseases Manual. 16th ed. Washington, DC: American Public Health Association, pp. 535–536. Berger C, Day P, Meier G, Zingg W, Bossart W, Nadal D. 2001. Dynamics of Epstein-Barr virus DNA levels in serum during EBV-associated disease. J Med Virol 64:505–512. Bernard SM, Cartwright RA, Bird CC, Richards ID, Lauder I, Roberts BE. 1984. Aetiologic factors in lymphoid malignancies: A case-control epidemiological study. Leuk Res 8:681–689. Bernard SM, Cartwright RA, Darwin CM, et al. 1987. Hodgkin’s disease: Case control epidemiological study in Yorkshire. Br J Cancer 55:85–90. Bibeau F, Brousset P, Knecht H, et al. 1994. Epstein-Barr virus replication in Hodgkin’s disease. Bull Cancer 81:114–118. Bierman PJ, Vose JM, Langnas AN, et al. 1996. Hodgkin’s disease following solid organ transplantation. Ann Oncol 7:265–270. Bjerrum OW, Hasselbalch HC, Drivsholm A, Nissen NI. 1986. Non-Hodgkin malignant lymphomas and Hodgkin’s disease in first-degree relatives. Evidence for a mutual genetic predisposition? Scand J Haematol 36:398–401. Boice JD Jr., Land CE, Preston DL. 1996. Ionizing radiation. In: Schottenfeld D, Fraumeni JF Jr., eds. Cancer Epidemiology and Prevention. 2nd ed. New York: Oxford University Press, pp. 319–354. Boiocchi M, Dolcetti R, de Re V, Gloghini A, Carbone A. 1993. Demonstration of a unique Epstein-Barr virus-positive cellular clone in metachronous multiple localizations of Hodgkin’s disease. Am J Pathol 142:33–38. Bonelli L, Vitale V, Bistolfi F, Landucci M, Bruzzi P. 1990. Hodgkin’s disease in adults: Association with social factors and age at tonsillectomy. A casecontrol study. Int J Cancer 45:423–427. Boyle MJ, Vasak E, Tschuchnigg M, et al. 1993. Subtypes of Epstein-Barr virus (EBV) in Hodgkin’s disease: Association between B-type EBV and immunocompromise. Blood 81:468–474. Briggs NC, Hall HI, Brann EA, Moriarty CJ, Levine RS. 2002. Cigarette smoking and risk of Hodgkin’s disease: A population-based case-control study. Am J Epidemiol 156:1011–1020. Brocksmith D, Angel CA, Pringle JH, Lauder J. 1991. Epstein-Barr viral DNA in Hodgkin’s disease: Amplification and detection using the polymerase chain reaction. J Pathol 165:11–15. Brousset P, Chittal S, Schlaifer D, et al. 1991. Detection of Epstein-Barr virus messenger RNA in Reed-Sternberg cells of Hodgkin’s disease by in situ hybridization with biotinylated probes on specially modified acetone methyl benzoate xylene (ModAmex) sections. Blood 77:1781–1786. Brousset P, Knecht H, Rubin B, et al. 1993. Demonstration of Epstein-Barr virus replication in Reed-Sternberg cells of Hodgkin’s disease. Blood 82:872–876. Brousset P, Schlaifer D, Meggetto F, et al. 1994. Persistence of the same viral strain in early and late relapse of Epstein-Barr virus-associated Hodgkin’s disease. Blood 84:2447–2451. Bryden H, MacKenzie J, Andrew L, et al. 1997. Determination of HLA-A*02 antigen status in Hodgkin’s disease and analysis of an HLA-A*02 restricted epitope of the Epstein-Barr virus LMP-2 protein. Int J Cancer 72:614–618. Callahan BC, Coe R, Place HM. 1994. Hodgkin disease of the spine presenting as alcohol-related pain. A case report and review of the literature. J Bone Joint Surg Am 76:119–121. Carbone A, Gloghini A, Zanette I, Canal B, Rizzo A, Volpe R. 1993. Coexpression of Epstein-Barr virus latent membrane protein and vimentin in “aggressive” histological subtypes of Hodgkin’s disease. Virchows Arch A Pathol Anat Histopathol 422:39–45. Carter CD, Brown TM Jr, Herbert JT, Heath CW Jr. 1977. Cancer incidence following infectious mononucleosis. Am J Epidemiol 105:30–36. Cartwright RA, Gurney KA, Moorman AV. 2002. Sex ratios and the risks of haematological malignancies. Br J Haematol 118:1071–1077.

892

PART IV: CANCER BY TISSUE OF ORIGIN

Chakravarti A, Halloran SL, Bale SJ, Tucker MA. 1986. Etiological heterogeneity in Hodgkin’s disease: HLA linked and unlinked determinants of susceptibility independent of histological concordance. Genet Epidemiol 3:407–415. Chang ET, Montgomery SM, Richiardi L, Ehlin A, Ekbom A, Lambe M. 2004. Number of siblings and risk of Hodgkin’s lymphoma. Cancer Epidemiol Biomarkers Prev 13:1235–1243. Chang ET, Zheng T, Lennette ET, et al. 2004c. Heterogeneity of risk factors and antibody profiles in Epstein-Barr virus genome-positive and genomenegative Hodgkin’s lymphoma. J Infect Dis 189:2271–2281. Chang ET, Zheng T, Weir E, Borowitz M, Mann R, Mueller N. 2004b. Aspirin and the risk of Hodgkin’s lymphoma in a population-based case-control study. J Natl Cancer Inst 96:305–315. Chang ET, Zheng T, Weir E, et al. 2004a. Childhood social environment and Hodgkin’s lymphoma in a population-based case-control study. Cancer Epidemiol Biomarkers Prev 13:1361–1370. Chang KL, Albujar PF, Chen Y-Y, Johnson RM, Weiss LM. 1993. High prevalence of Epstein-Barr virus in the Reed-Sternberg cells of Hodgkin’s disease occurring in Peru. Blood 81:496–501. Chang Y, Cesarman E, Pessin MS, et al. 1994. Identification of herpesviruslike DNA sequences in AIDS-associated Kaposi’s sarcoma. Science 266:1865–1869. Chatenoud L, Gallus S, Altieri A, et al. 2005. Number of siblings and risk of Hodgkin’s and other lymphoid neoplasmas. Cancer Epidemliol Biomarkers Prev 14:552. Letter. Chatenoud L, Tavani A, LaVecchia C, et al. 1998. Whole grain intake and cancer risk. Int J Cancer 77:24–28. Chen YT, Zheng T, Chou MC, Boyle P, Holford TR. 1997. The increase of Hodgkin’s disease incidence among young adults. Experience in Connecticut, 1935–1992. Cancer 79:2209–2218. Christiansen SC. 2000. Day care, siblings, and asthma—please, sneeze on my child. N Engl J Med 343:574–575. Clemmesen J. 1981. To the epidemiology of Hodgkin’s lymphogranulomatosis. J Belge Radiol 3:263–271. Coates PJ, Slavin G, D’Ardenne AJ. 1991. Persistence of Epstein-Barr virus in Reed-Sternberg cells throughout the course of Hodgkin’s disease. J Pathol 164:291–297. Cohen BM, Smetana HE, Miller RW. 1964. Hodgkin’s disease: Long survival in a study of 388 World War II army cases. Cancer 17:856. Cole P, Mack T, Rothman K, Henderson B, Newell G. 1973. Tonsillectomy and Hodgkin’s disease. N Engl J Med 288:634–635. Connelly RR, Christine BW. 1974. A cohort study of cancer following infectious mononucleosis. Cancer Res 34:1172–1178. Cope JU, Askling J, Gridley G, et al. 2003. Appendectomy during childhood and adolescence and the subsequent risk of cancer in Sweden. Pediatrics 111:1343–1350. Correa P, O’Conor GT. 1971. Epidemiologic patterns of Hodgkin’s disease. Int J Cancer 8:192–201. Cozen W, Gill PS, Ingles SA, et al. 2004. IL-6 levels and genotype are associated with risk of young adult Hodgkin’s lymphoma. Blood 103:3216– 3221. Cozen W, Katz J, Mack T. 1992. Risk patterns of Hodgkin’s disease in Los Angeles vary by cell type. Cancer Epidemiol Biomarkers Prev 1:261–268. Cozen W, Masood R, Mack T, Gill PS, Ablashi DV. 1998. Seroprevalence of Kaposi’s sarcoma-associated herpes virus antibody in young adult Hodgkin’s disease. Blood 91:724–728. Cunningham WF. 1917. The status of diphtheroids with special reference to Hodgkin’s disease. Am J Med Sci 153:406–412. Cusano MM, Young JL Jr. 1986. Forty-five years of Cancer Incidence in Connecticut: 1935–1979. NIH publication no. 86-2652. Bethesda, MD: National Cancer Institute, U.S. Department of Health and Human Service. Cuzick J, Edwards R. 1990. Spatial clustering for inhomogeneous populations. JR Stat Soc B 52:73–104. Dal Maso L, Franceschi S. 2003. Epidemiology of non-Hodgkin’s lymphomas and other haemolymphopoietic neoplasms in people with AIDS. Lancet Oncol 4:110–119. Davis MK, Savitz DA, Graubard BI. 1988. Infant feeding and childhood cancer. Lancet 2:365–368. Davis S. 1986. Case aggregation in young adult Hodgkin’s disease. Etiologic evidence from a population experience. Cancer 51:1602–1612. de Sanjose S, Zock J-P, Cavalle N, et al. 2004. Occupational exposure to immunologically active agents and risk for lymphoma. Cancer Epidemiol Biomarkers Prev 13:1814–1818. De Vita VT, Bonadonna G. 1999. The history of the chemotherapy of Hodgkin’s disease. In: Mauch PM, Armitage JO, Diehl V, Hoppe RT, Weiss LM, eds. Hodgkin’s Disease. Philadelphia: Lippincott Williams and Wilkins, pp. 9–22.

Deacon EM, Pallesen G, Niedobitek G, et al. 1993. Epstein-Barr virus and Hodgkin’s disease: Transcriptional analysis of virus latency in the malignant cells. J Exp Med 177:339–349. Delsol G, Brousset P, Chittal S, Rigal-Huguet F. 1992. Correlation of the expression of Epstein-Barr virus latent membrane protein and in situ hybridization with biotinylated Bam HI-W probes in Hodgkin’s disease. Am J Pathol 140:247–253. Demers PA, Boffetta P. 1998. Cancer risk from occupational exposure to wood dust, a pooled analysis of an epidemiological study. IARC Technical Report No. 30. Lyon, France: International Agency for Research on Cancer. Devesa SS, Grauman DJ, Blot WJ, Pennello GA, Hoover RN, Fraumeni JF Jr. 1999. Atlas of Cancer Mortality in the United States: 1950–1994. NIH publication no. 99-4564. Bethesda, MD: National Cancer Institute. Di Biagio E, Sanchez-Borges M, Desenne JJ, Suarez-Chacon R, Somoza R, Acquatella G. 1996. Eosinophilia in Hodgkin’s disease: A role for interleukin 5. Int Arch Allergy Immunol 110:244–251. Diepstra A, Niens M, Vellenga E, et al. 2005. Association with HLA class I in Epstein-Barr-virus positive and with HLA class III in Epstein-Barr-virusnegative Hodgkin’s lymphoma. Lancet 365:2216–2224. Dockerty JD, Sharples KJ, Borman B. 1999. An assessment of spatial clustering of leukemias and lymphomas among young people in New Zealand. J Epidemiol Community Health 53:154–158. Dores GM, Metayer C, Curtis RE, et al. 2002. Second malignant neoplasms among long-term survivors of Hodgkin’s disease: A population-based evaluation over 25 years. J Clin Oncol 20:3484–3494. Dukers DF, Jaspars LH, Vos W, et al. 2000. Quantitative immunohistochemical analysis of cytokine profiles in Epstein-Barr virus-positive and -negative cases of Hodgkin’s disease. J Pathol 190:143–149. Ederer F, Myers MH, Mantel N. 1964. A statistical problem in space and time: Do leukemia cases come in clusters? Biometrics 20:626. Egan KM, Sosman JA, Blot WJ. 2005. Sunlight and reduced risk of cancer: Is the real story vitamin D? J Natl Cancer Inst 97:161–163. Enblad G, Sandvej K, Lennette E, et al. 1997. Lack of correlation between EBV serology and presence of EBV in the Hodgkin and Reed-Sternberg cells of patients with Hodgkin’s disease. Int J Cancer 72:394–397. Enblad G, Sandvej K, Sundström C, Pallesen G, Glimelius B. 1999. EpsteinBarr virus distribution in Hodgkin’s disease in an unselected Swedish population. Acta Oncolgica 38:425–429. Evans AS. 1971. The spectrum of infections with Epstein-Barr virus: A hypothesis. J Infect Dis 124:330–337. Evans AS, Comstock GW. 1981. Presence of elevated antibody titers to Epstein-Barr virus before Hodgkin’s disease. Lancet 1:1183–1186. Evans AS, Gutensohn (Mueller) N. 1984. A population-based case-control study of EBV and other viral antibodies among persons with Hodgkin’s disease and their siblings. Int J Cancer 34:149–157. Evans AS, Mueller N. 1990. Viruses and cancer: Causal associations. Ann Epidemiol 1:71–92. Fellbaum C, Hansmann M-L, Niedermeyer H, et al. 1992. Influence of EpsteinBarr virus genomes on patient survival in Hodgkin’s disease. Am J Clin Pathol 98:319–323. Filipovich AH, Mathur A, Kamat D, Shapiro RS. 1992. Primary immunodeficiencies: Genetic risk factors for lymphoma. Cancer Res 52 (Suppl): 5465–5467s. Flavell KJ, Biddulph JP, Powell JE, et al. 2001. South Asian ethnicity and material deprivation increase the risk of Epstein-Barr virus infection in childhood Hodgkin’s disease. Br J Cancer 85:350–356. Flavell KJ, Biddulph JP, Constandinou CM, et al. 2000. Variation in the frequency of Epstein-Barr virus-associated Hodgkin’s disease with age. Leukemia 14:748–753. Flavell K, Constandinou C, Lowe D, et al. 1999. Effect of material deprivation on Epstein-Barr infection in Hodgkin’s disease in the West Midlands. Br J Cancer 80:604–608. Fonte R, Grigis L, Grigis P, Franco G. 1982. Chemicals and Hodgkin’s disease. Lancet 2:50. Foss HD, Herbst H, Oelmann E, et al. 1993. Lymphotoxin, tumour necrosis factor and interleukin-6 gene transcripts are present in Hodgkin and Reed-Sternberg cells of most Hodgkin’s disease cases. Br J Haematol 84:627–635. Franceschi S, Bidoli E, La Vecchia C. 1994. Pregnancy and Hodgkin’s disease. Int J Cancer 58:465–466. Fraumeni JF Jr., Li FP. 1969. Hodgkin’s disease in childhood: An epidemiologic study. J Natl Cancer Inst 42:681–691. Frisan T, Sjoberg J, Dolcetti R, et al. 1995. Local suppression of Epstein-Barr virus (EBV)—specifically cytotoxicity in biopsies of EBV-positive Hodgkin’s disease. Blood 86:1493–1501.

Hodgkin Lymphoma Frisch M, Biggar RJ, Engels EA, Goedert JJ for the AIDS-Cancer Match Registry Study Group. 2001. Association of cancer with AIDS-related immunosuppression in adults. JAMA 285:1736–1745. Gallagher A, Armstrong AA, MacKenzie J, et al. 1999. Detection of EpsteinBarr virus (EBV) genomes in the serum of patients with EBV-associated Hodgkin’s disease. Int J Cancer 84:442–448. Gallagher A, Perry J, Shield L, Freeland J, MacKenzie J, Jarrett RF. 2002. Viruses and Hodgkin’s disease: No evidence of novel herpes viruses in non-EBV-associated lesions. Int J Cancer 101:259–264. Garnier JL, Lebranchu Y, Dantal J, et al. 1996. Hodgkin’s disease after transplantation. Transplantation 61:71–76. Gately MK, Renzetti LM, Magram J, et al. 1998. The interleukin-12/ interleukin-12-receptor system: Role in normal and pathologic immune responses. Ann Rev Immunol 16:495–521. George WK, Miller M, George WD. 1965. Hodgkin’s disease in roommates. J Am Coll Health Assoc 13:399–402. Gilman EA, McNally RJQ, Cartwright RA. 1999. Space-time clustering of Hodgkin’s disease in parts of the UK, 1984–1993. Leuk Lymphoma 36:85–100. Glaser SL. 1987. Regional variation in Hodgkin’s disease incidence by histologic subtype in the US. Cancer 60:2841–2847. Glaser SL. 1990. Hodgkin’s disease in black populations: A review of the epidemiologic literature. Semin Oncol 17:643–659. Glaser SL. 1991. Black-white differences in Hodgkin’s disease incidence in the United States by age, sex, histology subtype and time. Int J Epidemiology 20:68–74. Glaser SL. 1994. Reproductive factors in Hodgkin’s disease in women: A review. Am J Epidemiol 139:237–246. Glaser SL, Clarke CA, Nugent RA, Stearns CB, Dorfman RF. 2002. Social class and risk of Hodgkin’s disease in young-adult women in 1988–94. Int J Cancer 98:110–117. Glaser SL, Clarke CA, Stearns CB, Dorfman RF. 2001. Age variation in Hodgkin’s disease risk factors in older women: Evidence from a population-based case-control study. Leuk Lymphoma 42:997–1004. Glaser SL, Hsu JL. 2002. Hodgkin’s disease in Asians: Incidence patterns and risk factors in population-based data. Leukemia Res 26:261–269. Glaser SL, Lin RJ, Stewart SL, et al. 1997. Epstein-Barr virus-associated Hodgkin’s disease: Epidemiologic characteristics in international data. Int J Cancer 70:375–382. Glaser SL, Swartz WG. 1990. Time trends in Hodgkin’s disease incidence: The role of diagnostic accuracy. Cancer 66:2196–2204. Gledhill S, Gallagher A, Jones DB, et al. 1991. Viral involvement in Hodgkin’s disease: Detection of clonal type A Epstein-Barr virus genomes in tumour samples. Br J Cancer 64:227–232. Goedert JJ. 2000. The epidemiology of acquired immunodeficiency syndrome malignancies. Semin Oncol 27:390–401. Grässer FA, Murray PG, Kremmer E, et al. 1994. Monoclonal antibodies directed against the Epstein-Barr virus-encoded nuclear antigen 1 (EBNA1): Immunohistologic detection of EBNA1 in the malignant cells of Hodgkin’s disease. Blood 84:3792–3798. Green JA, Dawson AA, Valerio D. 1979. Hodgkin’s disease and infectious mononucleosis. Scand J Haematol 23:313–314. Green MS. 1992. The male predominance in the incidence of infectious diseases in children: A postulated explanation for disparities in the literature. Int J Epidemiol 21:381–386. Greenberg RS, Grufferman S, Cole P. 1983. An evaluation of space-time clustering in Hodgkin’s disease. J Chronic Dis 36:257–262. Greene MH, Brinton LA, Fraumeni JH Jr., d’Amico R. 1978. Familial and sporadic Hodgkin’s disease associated with occupational wood exposure. Lancet 2:626–627. Greenwald P, Rose JS, Daitch PB. 1979. Acquaintance networks among leukemia and lymphoma patients. Am J Epidemiol 110:162–177. Grilli M, Pizzi M, Memo M, Spano P. 1996. Neuroprotection by aspirin and sodium salicylate through blockade of NFkB activation. Science 274:1383–1385. Grufferman S. 1977. Clustering and aggregation of exposures in Hodgkin’s disease. Cancer 39(4Suppl):1829–1833. Grufferman S, Cole P, Smith PG, Lukes RJ. 1977. Hodgkin’s disease in siblings. N Engl J Med 296:248–250. Grufferman S. 1982. Hodgkin’s disease. In: Schottenfeld D, Fraumeni JF, Jr., eds. Cancer Epidemiology and Prevention. Philadelphia, WB Saunders Co., pp. 739–753. Grufferman S. 1995. Is Hodgkin’s disease infectious? Eur J Cancer 31A:1388–1389. Grufferman S, Ambinder RF, Shugart YY, Brecher ML, Gilchrist GS. 1998a. Increased cancer risk in families of children with Hodgkin’s disease [abstract]. Am J Epidemiol 147:S8.

893

Grufferman S, Barton JW 3rd, Eby NL. 1987. Increased sex concordance of sibling pairs with Behcet’s disease, Hodgkin’s disease, multiple sclerosis and sarcoidosis. Am J Epidemiol 126:365–369. Grufferman S, Cole P, Levitan TR. 1979. Evidence against transmission of Hodgkin’s disease in high schools. N Engl J Med 300:1006–1011. Grufferman S, Davis MK, Ambinder RF, Shugart YY, Gilchrist GS, Brecher ML. 1998b. A protective effect of breast feeding on risk of Hodgkin’s disease in childhood. Paediatr Perinat Epidemiol 12:A13. Grufferman S, Delzell E. 1984. Epidemiology of Hodgkin’s disease. Epidemiol Rev 6:76–106. Grufferman S, Duong T, Cole P. 1976. Occupation and Hodgkin’s disease. J Natl Cancer Inst 57:1193–1195. Gulley ML, Eagen PA, Quintanilla-Martinez L, et al. 1994. Epstein-Barr virus DNA is abundant and monoclonal in the Reed-Sternberg cells of Hodgkin’s disease: Association with mixed cellularity subtype and Hispanic American ethnicity. Blood 83:1595–1602. Gulley ML, Glaser SL, Craig FE, et al. 2002. Guidelines for interpreting EBER in situ hybridization and LMP1 immunohistochemical tests for detecting Epstein-Barr virus in Hodgkin lymphoma. Am J Clin Pathol 117:259–267. Gutensohn (Mueller) N. 1982. Social class and age at diagnosis of Hodgkin’s disease: New epidemiologic evidence on the two-disease hypothesis. Cancer Treat Rep 66:689–695. Gutensohn (Mueller) N, Cole P. 1977. Epidemiology of Hodgkin’s disease in the young. Int J Cancer 19:595–604. Gutensohn (Mueller) N, Cole P. 1980. Epidemiology of Hodgkin’s disease. Semin Oncol 7:92–102. Gutensohn (Mueller) N, Cole P. 1981. Childhood social environment and Hodgkin’s disease. N Engl J Med 304:135–140. Gutensohn (Mueller) N, Li FP, Johnson RE, Cole P. 1975. Hodgkin’s disease, tonsillectomy and family size. N Engl J Med 292:22–25. Gutensohn (Mueller) N, Shapiro D. 1982. Social class risk factors among children with Hodgkin’s disease. Int J Cancer 30:433–435. Haim N, Cohen Y, Robinson E. 1982. Malignant lymphoma in first-degree blood relatives. Cancer 49:2197–2200. Hakansson N, Floderus B, Gustavsson P, Feychting M, Hallin N. 2001. Occupational sunlight exposure and cancer incidence among Swedish construction workers. Epidemiol 12:552–557. Halnan KE. 1988. Failure to substantiate two cases of alleged occupational radiation carcinogenesis. Lancet i:639. Hamilton-Dutoit SJ, Pallesen G. 1992. A survey of Epstein-Barr virus gene expression in sporadic non-Hodgkin’s lymphoma. Am J Pathol 140:1315–1325. Harris NL, Jaffe ES, Stein H, et al. 1994. A revised European-American classification of lymphoid neoplasms: A proposal from the International Lymphoma Study Group. Blood 84:1361–1392. Hartge P, Devesa SS, Fraumeni JF Jr. 1994. Hodgkin’s disease and nonHodgkin’s lymphoma. Cancer Surv 19/20:423–453. Harty LC, Lin AY, Goldstein AM, et al. 2002. HLA-DR, HLA-DQ, and TAP genes in familial Hodgkin disease. Blood 99:690–693. Heath CW, Rosenstock JG, Lobdell G. 1971. An epidemic of Hodgkin’s disease? Lancet ii:426–427. Hellman S. 1999. A brief consideration of Thomas Hodgkin and his times. In: Mauch PM, Armitage JO, Diehl V, Hoppe RT, Weiss LM, eds. Hodgkin’s Disease. Philadelphia: Lippincott Williams and Wilkins, pp. 3–7. Henderson BE, Dworsky R, Pike MC, et al. 1979. Risk factors for modular sclerosis and other types of Hodgkin’s disease. Cancer Res 39:4507– 4511. Henle W, Henle G. 1973. Epstein-Barr virus-related serology in Hodgkin’s disease. Natl Cancer Inst Monog 36:79–84. Henle W, Henle G, Andersson J, et al. 1987. Antibody responses to EpsteinBarr virus-determined nuclear antigen (EBNA)-1 and EBNA-2 in acute and chronic Epstein-Barr virus infection. Proc Natl Acad Sci USA 84:570–574. Herbst H, Dallenbach F, Hummel M, et al. 1991. Epstein-Barr virus latent membrane protein expression in Hodgkin and Reed-Sternberg cells. Proc Natl Acad Sci USA 88:4766–4770. Herbst H, Foss HD, Samol J, et al. 1996. Frequent expression of interleukin10 by Epstein-Barr virus-harboring tumor cells of Hodgkin’s disease. Blood 87:2918–2929. Herbst H, Niedobitek G. 1993. Epstein-Barr virus and Hodgkin’s disease. Int J Clin Lab Res 23:13–16. Herbst H, Niedobitek G, Kneba M, et al. 1990. High incidence of Epstein-Barr virus genomes in Hodgkin’s disease. Am J Pathol 137:13–18. Herbst H, Samol J, Foss HD, Raff T, Niedobitek G. 1997. Modulations of interleukin-6 expression in Hodgkin and Reed-Sternberg cells by EpsteinBarr virus. J Pathol 182:299–306.

894

PART IV: CANCER BY TISSUE OF ORIGIN

Herbst H, Steinbrecher E, Niedobitek G, et al. 1992. Distribution and phenotype of Epstein-Barr virus-harboring cells in Hodgkin’s disease. Blood 80:484–491. Herbst H, Tippelmann G, Aragnostopoulos I, et al. 1989. Immunoglobulin and T-cell receptor gene rearrangements in Hodgkin’s disease and ki-1positive anaplastic large cell lymphoma: Dissociation between phenotype and genotype. Leuk Res 13:103–116. Herling M, Rassidakis GZ, Medeiros J, et al. 2003. Expression of Epstein-Barr virus latent membrane protein-1 in Hodgkin and Reed-Sternberg cells of classical Hodgkin’s lymphoma: Associations with presenting features, serum interleukin 10 levels, and clinical outcome. Clin Cancer Res 9:2114–2120. Hinz M, Lemke P, Anagnostopoulos I, et al. 2002. Nuclear factor kBdependent gene expression profiling of Hodgkin’s disease tumor cells, pathogenetic significance, and link to constitutive signal transducer and activator of transcription 5a activity. J Exp Med 196:605–617. Hinz M, Löser P, Mathas S, Krappmann D, Dörken B, Scheidereit C. 2001. Constitutive NFkB maintains high expression of a characteristic gene network, including CD40, CD86, and a set of anti-apoptotic genes in Hodgkin/Reed-Sternberg cells. Blood 97:2798–2807. Hjalgrim H, Askling J, Pukkala E, Hansen S, Munksgaard L, Frisch M. 2001. Incidence of Hodgkin’s disease in Nordic countries. Lancet 358: 297–298. Hjalgrim H, Askling J, Rostgaard K, et al. 2003. Characteristics of Hodgkin’s lymphoma after infectious mononucleosis. N Engl J Med 349: 1324–1332. Hjalgrim H, Askling J, Sorensen P, et al. 2000. Risk of Hodgkin’s disease and other cancers after infectious mononucleosis. J Natl Cancer Inst 92:1522–1528. Hjalgrim H, Rostgaard K, Askling J, et al. 2002. Hematopoietic and lymphatic cancers in relatives of patients with infectious mononucleosis. J Natl Cancer Inst 94:678–681. Hodgkin T. 1832. On some morbid appearances of the absorbent glands and spleen. Med Chir Trans 17:68–114. Horm JW, Asire AJ, Young JL, et al. 1985. SEER Program: Cancer Incidence and Mortality in the United States, 1973–1981. NIH publication no. 85–1837. Bethesda, MD: National Cancer Institute, US Department of Health and Human Services. Hors J, Dausset J. 1983. HLA and susceptibility in Hodgkin’s disease [review]. Immunol Rev 70:167–192. Horwitz M, Wiernik PH. 1999. Pseudoautosomal linkage of Hodgkin disease. Am J Hum Genet 65:1413–1422. Hoster HA, Dratman MV, Craver LF, Rolnick HA. 1948. Hodgkin’s disease 1832–1947. Cancer Res 8:Part I 1–48; Part II 49–78. Hsu SM, Zie SS, Hsu PL, Waldron JA Jr. 1992. Interleukin-6, but not interleukin-4, is expressed by Reed-Sternberg cells in Hodgkin’s disease with or without histologic features of Castleman’s disease. Am J Pathol 141:129–138. Hu E, Hufford S, Lukes R, et al. 1988. Third-world Hodgkin’s disease in Los Angeles County-University of Southern California Medical Center. J Clin Oncol 6:1285–1292. Hudson MM, Poquette CA, Lee J, et al. 1998. Increased mortality after successful treatment for Hodgkin’s disease. J Clin Oncol 16:3592–3600. IARC. 1997. Monographs on the Evaluation of Carcinogenic Risks to Humans. Vol 70. Epstein-Barr Virus and Kaposi’s Sarcoma Herpesvirus/Human Herpesvirus 8. Lyon, France: International Agency for Research on Cancer, pp. 47–373. Infante-Rivard C, Amre D, Gautrin D, Malo J-L. 2001. Family size, day-care attendance and breastfeeding in relation to the incidence of childhood asthma. Am J Epidemiol 153:653–658. Isager H, Larsen S. 1980. Pre-morbid factors in Hodgkin’s disease. III. School contact among patients. Scan J Haematol 25:158–164. Izban KF, Ergin M, Huang Q, et al. 2001. Characterization of NF-kappaB expression in Hodgkin’s disease: Inhibition of constitutively expressed NF-kappaB results in spontaneous caspase-independent apoptosis in Hodgkin and Reed-Sternberg cells. Mod Pathol 14:297–310. Izumi KM, Kieff ED. 1997. The Epstein-Barr virus oncogene product latent membrane protein 1 engages the tumor necrosis factor receptorassociated death domain protein to mediate B lymphocyte growth transformation and activate NFkB. Proc Natl Acad Sci USA 94:12592–12597. Jackson H Jr., Parker F Jr. 1944a. Hodgkin’s disease I. General considerations. N Engl J Med 230:1–8. Jackson H Jr., Parker F Jr. 1944b. Hodgkin’s disease II. Pathology. N Engl J Med 231:35–44. Jacquez GM. 1994a. Cuzick and Edwards’ test when exact locations are unknown. Am J Epidemiol 140:58–64. Jacquez GM. 1994b. STAT! Statistical software for the clustering of health events. Ann Arbor, MI: BioMedware.

Jaffe ES, Harris NL, Stein H, Vardiman JW, eds. 2001. Pathology & Genetics of Tumours of Haematopoietic and Lymphoid Tissues. World Health Organization Classification of Tumours. Lyon, France: International Agency for Research on Cancer. Jarrett RF, Armstrong AA, Alexander E. 1996. Epidemiology of EBV and Hodgkin’s lymphoma. Ann Oncol 7:(Suppl4):S5–S10. Jarrett RF, Gallagher A, Jones DB, et al. 1991. Detection of Epstein-Barr virus genomes in Hodgkin’s disease: Relation to age. J Clin Pathol 44: 844–848. Jarrett RF, Krajewski AS, Angus B, et al. 2003. The Scotland and Newcastle epidemiological study of Hodgkin’s disease: Impact of histopathological review and EBV status on incidence estimates. J Clin Pathol 56:811–816. Johnson SK, Johnson RE. 1972. Tonsillectomy history in Hodgkin’s disease. N Engl J Med 287:1122–1125. Jucker M, Abts H, Li W, et al. 1991. Expression of interleukin-6 and interleukin-6 receptor in Hodgkin’s disease. Blood 77:2413–2418. Jungnickel B, Staratschek-Jox A, Bräuninger A, et al. 2000. Clonal deleterious mutations in the IkBa gene in the malignant cells in Hodgkin’s lymphoma. J Exp Med 191:395–401. Kadin ME, Liebowitz DN. 1999. Cytokines and cytokine receptors in Hodgkin’s disease. In: Mauch PM, Armitage JO, Diehl V, Hoppe RT, Weiss LM, eds. Hodgkin’s Disease. Philadelphia: Williams Wilkins, pp. 139–157. Kahn KS, Tatham LM, Patel AV, Thun MJ, Heath Jr. CW. 1998. Increased cancer mortality following a history of nonmelanoma skin cancer. JAMA 280:910–912. Kaplan HS. 1980. Hodgkin’s Disease, 2nd ed. Cambridge, MA: Harvard University Press. Kapp U, Yeh WC, Patrerson B, et al. 1999. Interleukin 13 is secreted by and stimulates the growth of Hodgkin and Reed-Sternberg cells. J Exp Med 189:1939–1946. Kass AM. 1984. Doctors afield: Dr. Thomas Hodgkin’s friendship with Sir Moses Montefiore. N Engl J Med 310:401–404. Kass AM, Kass EH. 1988. Perfecting the World: The Life and Times of Dr. Thomas Hodgkin 1798–1866. New York: Harcourt Brace Javanovich. Khan G, Coates PJ, Gupta RK, Kangro HO, Slavin G. 1992. Presence of Epstein-Barr virus in Hodgkin’s disease is not exclusive to ReedSternberg cells. Am J Pathol 140:757–762. Khan G, Norton AJ, Slavin G. 1993. Epstein-Barr virus in Hodgkin’s disease: Relation to age and subtype. Cancer 71:3124–3129. Kieff E, Liebowitz D. 1990. Epstein-Barr virus and its replication. In: Fields B, Knipe DM, eds. Virology. 2nd ed. New York: Raven Press, pp. 1889–1920. Kirchoff LV, Evans AS, McClelland KE, Carvalho RP, Pannuti CS. 1980. A case-control study of Hodgkin’s disease in Brazil. I. Epidemiologic aspects. Am J Epidemiol 112:595–608. Kishimoto T, Akira S, Narazaki M, Taga T. 1995. Interleukin-6 family of cytokines and gp130. Blood 86:1243–1254. Klein S, Jucker M, Diehl V, Tesch H. 1992. Production of multiple cytodines by Hodgkin’s disease derived cell lines. Hematol Oncol 10:319–329. Knecht H, Bachmann E, Brousset P, et al. 1993a. Deletions within the LMP1 oncogene of Epstein-Barr virus are clustered in Hodgkin’s disease and identical to those observed in nasopharyngeal carcinoma. Blood 82:2037–2042. Knecht H, Brousset P, Bachmann E, Sandvej K, Odermatt BF. 1993b. Latent membrane protein 1: A key oncogene in EBV-related carcinogenesis? Acta Haematol 90:167–171. Knox G. 1963. Detection of low intensity epidemicity: Application to cleft lip and palate. Br J Prev Soc Med 17:121–127. Kopp E, Ghosh S. 1994. Inhibition of NFkB by sodium salicylate and aspirin. Science 265:956–959. Kramer U, Heinrich J, Wjst M, Wichmann HE. 1999. Age of entry to day nursery and allergy in later childhood. Lancet 353:450–454. Kravdal O, Hansen S. 1993. Hodgkin’s disease: The protective effect of childbearing. Int J Cancer 55:909–914. Kryscio RJ, Myers MH, Prusiner ST, Heise HW, Christine BW. 1973. The space-time distribution of Hodgkin’s disease in Connecticut, 1940–1969. J Natl Cancer Inst 50:1107–1110. Kung FH. 1991. Hodgkin’s disease in children four years of age or younger. Cancer 67:1428–1430. Kuppers R. 2003. Somatic hypermutation and B cell receptor selection in normal and transformed human B cells. Ann NY Acad Sci 987: 173–179. Küppers R, Klein U, Hansmann M-L, Rajewsky K. 1999. Cellular origin of human B-cell lymphomas. N Engl J Med 341:1520–1529. Kusuda M, Toriyama K, Kamidigo NO, Itakura H. 1998. A comparison of epidemiologic, histologic, and virologic studies on Hodgkin’s disease in western Kenya and Nagasaki, Japan. Am J Trop Med Hyg 59:801–807.

Hodgkin Lymphoma Kusuhara K, Takabayashi A, Ueda K, et al. 1997. Breast milk is not a significant source for early Epstein-Barr virus or human herpesvirus 6 infection in infants; a seroepidemiologic study in two endemic areas of human T-cell lymphotropic virus type I in Japan. Microbiol Immunol 41:309–312. Kusunoki Y, Kyoizumi S, Fukuda Y, et al. 1994. Immune responses to EpsteinBarr virus in atomic bomb survivors: Study of precursor frequency of cytotoxic lymphoctyes and titer levels of anti-Epstein-Barr virus-related antibodies. Radiat Res 138:127–132. Kvale G, Hoiby EA, Pedersen E. 1979. Hodgkin’s disease in patients with previous infectious mononucleosis. Int J Cancer 23:593–597. LaVecchia C, Negri E, D’Aranzo B, Franceschi S. 1989. Occupation and lymphoid neoplasms. Br J Cancer 60:385–388. Lehtinen T, Lumio J, Dillner J, et al. 1993. Increased risk of malignant lymphoma indicated by elevated Epstein-Barr virus antibodies—a prospective study. Cancer Causes Control 4:187–193. Levi F, Randimbison L, LaVecchia C, Erler G, Te V-C. 1997. Incidence of invasive cancers following squamous cell skin cancer. Am J Epidemiol 146:734–739. Levine PH, Ebbesen P, Ablashi DV, Saxinger WC, Nordentoft A, Connelly RR. 1992. Antibodies to human herpes virus-6 and clinical course in patients with Hodgkin’s disease. Int J Cancer 51:53–57. Levine P, Pallesen G, Ebbesen P, Harris N, Evans AS, Mueller N. 1994. Evaluation of Epstein-Barr virus antibody patterns and detection of viral markers in the biopsies of patients with Hodgkin’s disease. Int J Cancer 59:48–50. Levine R, Zhu K, Gu Y, et al. 1998. Self-reported infectious mononucleosis and 6 cancers: A population-based, case-control study. Scand J Infect Dis 30:211–214. Libetta CM, Pringle JH, Angel CA, Craft AW, Malcolm AJ, Lauder I. 1990. Demonstration of Epstein-Barr viral DNA informalin-fixed paraffinembedded samples Hodgkin’s disease. J Pathol 161:255–260. Lichtenstein P, Holm NV, Verkasalo PK, et al. 2000. Environmental and heritable factors in the causation of cancer—analyses of cohorts of twins from Sweden, Denmark and Finland. N Engl J Med 343:78–85. Lin AY, Kingma DW, Lennette ET, et al. 1996. Epstein-Barr virus and familial Hodgkin’s disease. Blood 88:3160–3165. Lin J-C, Lin S-C, De BK, Chan W-P, Evatt BL. 1993. Precision of genotyping of Epstein-Barr virus by polymerase chain reaction using three gene loci (EBNA-2, EBNA-3C, and EBER): Predominance of type A virus associated with Hodgkin’s disease. Blood 81:3372–3381. Linet MS, Pottern LM. 1992. Familial aggregation of hematopoietic malignancies and risk of non-Hodgkin’s lymphoma. Cancer Res 52(19 Suppl): 5458s–5473s. Lukes RJ, Butler JJ. 1966. The pathology and nomenclature of Hodgkin’s disease. Cancer Res 26:1063–1081. Lukes RJ, Tindle BH, Parker JW. 1969. Reed-Sternberg-like cells in infectious mononucleosis. Lancet 2:1003–1004. Lynch HT. 2000. Anticipation in familial Hodgkin’s lymphoma. Hum Genet 107:290–293. Lynch HT, Marcus JN, Lynch JF. 1992. Genetics of Hodgkin’s and nonHodgkin’s lymphoma: A review. Cancer Investigation 10:247–256. MacFarlane GJ, Evstifeeva TV, Boyle P, Grufferman S. 1995. International patterns in the occurrence of Hodgkin’s disease in children and young adult males. Int J Cancer 61:165–167. Mack TM, Cozen W, Shibata DK, et al. 1995. Concordance for Hodgkin’s disease in identical twins suggesting genetic susceptibility to the youngadult form of the disease. N Engl J Med 332:413–418. MacMahon B. 1957. Epidemiological evidence on the nature of Hodgkin’s disease. Cancer 10:1045–1054. MacMahon B. 1966. Epidemiology of Hodgkin’s disease. Cancer Res 26:1189–1201. Mantel N. 1967. The detection of disease clustering and a generalized regression approach. Cancer Res 27:209–220. Mariette X, Cazals-Hatem D, Warszawki J, Liote F, Balandraud N, Sibilia J. Investigators of the Club Rhumatismes et Inflammation. 2002. Lymphomas in rheumatoid arthritis patients treated with methotrexate: A three-year prospective study in France. Blood 99:3909–3915. Marshall NA, Christie LE, Munro LR, et al. 2004. Immunosuppressive regulatory T cells are abundant in the reactive lymphocytes of Hodgkin’s lymphoma. Blood 103:1755–1762. Martinez-Maza O, Breen EC. 2002. B-cell activation and lymphoma in patients with HIV. Curr Opin Oncol 14:528–532. Masih A, Weisenburger D, Duggan M, et al. 1991. Epstein-Barr viral genome in lymph nodes from patients with Hodgkin’s disease may not be specific to Reed-Sternberg cells. Am J Pathol 139:37–43. Matricardi PM, Yazdanbakhsh M. 2003. Mycobacteria and atopy, 6 years later: A fascinating, still unfinished, business. Clin Exp Allergy 33: 717–720.

895

McCunney RJ. 1999. Hodgkin’s disease, work and the environment: A review. J Occup Environ Med 41:36–46. Melnick JL. 1997. Poliovirus and other enteroviruses. In: Evans AS, Kaslow RA, eds. Viral Infections of Humans: Epidemiology and Control. 4th ed. New York: Plenum, p. 609. Mertens AC, Yasui Y, Neglia JP, et al. 2001. Late mortality experience in fiveyear survivors of childhood and adolescent cancer: The Childhood Cancer Survivor Study. J Clin Oncol 19:3163–3172. Merz H, Fleidner A, Orscheschek K, et al. 1991. Cytokine expression in T-cell lymphomas and Hodgkin’s disease: Its possible implication in autocrine or paracrine production as a potential basis for neoplastic growth. Am J Pathol 139:1173–1180. Milham S Jr., Hesser JE. 1967. Hodgkin’s disease in woodworkers. Lancet 2:136–137. Milham S Jr. 1974. Mortality experience of the AFL-CIO United Brotherhood of Carpenters and Joiners of America, 1969–1970. DHEW publication no. (NIOSH) 174–129. Cinncinnati, Ohio, Office of Technical Publications. Miller BA, Kolonel LN, Bernstein L, et al. 1996. Racial/ethnic patterns of cancer in the United States 1988–1992. NIH publication no. 96-4104. Bethesda, MD: National Cancer Institute. Miller RW, Beebe GW. 1973. Infectious mononucleosis and the empirical risk of cancer. J Natl Cancer Inst 50:315–321. Moore KW, De Waal Malelyt R, Coffman RRL, O’Garra A. 2001. Interleukin10 and the interleukin-10 receptor. Ann Rev Immunol 19:683–765. Mueller NE. 1987. The epidemiology of Hodgkin’s disease. In: Selby D, McElwain TJ, eds. Hodgkin’s Disease. London: Blackwell Scientific Publications, pp. 68–93. Mueller NE. 1996. Hodgkin’s disease. In: Schottenfeld D, Fraumenti JF Jr, eds. Cancer Epidemiology and Prevention, 2nd ed. New York: Oxford University Press, pp. 893–919. Mueller NE. 1997. Epstein-Barr virus and Hodgkin’s disease: An epidemiological paradox. EBV Report 4:1–2. Mueller NE. 1999. Overview of the epidemiology of malignancy in immune deficiency. J Acquir Immune Defic Syndr 21(Suppl):S5–10. Mueller N, Evans A, Harris NL, et al. 1989. Hodgkin’s disease and EpsteinBarr virus. Altered antibody Pattern before diagnosis. N Engl J Med 320:689–695. Mueller NE, Grufferman S. 1999. The epidemiology of Hodgkin’s disease. In: Mauch PM, Armitage JO, Diehl V, Hoppe RT, Weiss LM, eds. Hodgkin’s Disease. Philadelphia: Lippincott Williams and Wilkins, pp. 61–77. Mueller N, Swanson GM, Hsieh C-C, Cole P. 1987. Tonsillectomy and Hodgkin’s disease: Results from companion population-based studies. J Natl Cancer Inst 78:1–5. Munoz N, Davidson RJL, Witthoff B, Ericsson JE, De-The G. 1978. Infectious mononucleosis and Hodgkin’s disease. Int J Cancer 22:10–13. Murray PG, Young LS, Rowe M, Crocker J. 1992. Immunohistochemical demonstration of the Epstein-Barr virus-encoded latent membrane protein in paraffin sections of Hodgkin’s disease. J Pathol 166:1–5. Neglia JP, Friedman DL, Yasui Y, et al. 2001. Second malignant neoplasms in five-year survivors of childhood cancer: The Childhood Cancer Survivor Study. J Natl Cancer Inst 93:618–629. Negri E, La Vecchia C, Franceschi S, D’Avanzo B, Parazzini F. 1991. Vegetable and fruit consumption and cancer risk. Int J Cancer 48:350–354. Newell GR. 1970. Etiology of multiple sclerosis and Hodgkin’s disease. Am J Epidemiol 91:119–122. Ng AK, Bernard MV, Weller E, et al. 2002. Second malignancy after Hodgkin disease treated with radiation therapy with or without chemotherapy: Long-term risks and risk factors. Blood 100:1989–1996. Niedobitek G, Herbst H, Young LS, et al. 1992. Patterns of Epstein-Barr virus infection in non-neoplastic lymphoid tissue. Blood 79:2520–2526. Niedobitek G, Kremmer E, Herbst H, et al. 1997. Immunohistochemical detection of the Epstein-Barr virus-encoded latent membrane 2A in Hodgkin’s disease and infectious mononucleosis. Blood 90:1664–1672. Niedobitek G, Rowlands DC, Young LS, et al. 1993. Overexpression of p53 in Hodgkin’s disease: Lack of correlation with Epstein-Barr virus infection. J Pathol 169:207–212. Ohshima K, Akaiwa M, Umeshita R, Suzumiya J, Izuhara K, Kikuchi M. 2001. Interleukin-13 and interleukin-13 receptor in Hodgkin’s disease: Possible autocrine mechanism and involvement in fibrosis. Histopathology 38:368–375. Ohshima K, Kikuchi M, Eguchi F, et al. 1990. Analysis of Epstein-Barr viral genomes in lymphoid malignancy using southern blotting, polymerase chain reaction and in situ hybridization. Virchows Arch B Cell Pathol Incl Mol Pathol 59:383–390. Ohshima K, Suzumiya J, Akamatu M, Takeshita M, Kikuchi M. 1995. Human and viral interleukin-10 in Hodgkin’s disease, and its influence on CD4+ and CD8+ T lymphocytes. Int J Cancer 62:5–10.

896

PART IV: CANCER BY TISSUE OF ORIGIN

Olsen JH, Boice JD, Seersholm N, Bautz A, Fraumeni JF Jr. 1995. Cancer in the parents of children with cancer. N Engl J Med 333:1594–1599. Olsen J, Sabroe S. 1979. A follow-up study of non-retired and retired members of the Danish Carpenters/Cabinet Marker’s Trade Union. Int J Epidemiol 8:375–382. Oshima Y, Puri PK. 2001. Suppression of an IL-13 autocrine growth loop in a human Hodgkin/Reed-Sternberg tumor cell line by a novel IL-13 antagonist. Cell Immunol 211:37–42. Oudejans JJ, van den Brule AJC, Jiwa NM, et al. 1995. BHRFl, the EpsteinBarr virus (EBV) homologue of the BCL-2 proto-oncogene, is transcribed in EBV-associated B-cell lymphomas and in reactive lymphocytes. Blood 86:1883–1902. Oza AM, Tonks S, Lim J, Fleetwood MA, Lister TA, Dodmer JG and Collaborating Centers. 1994. A clinical epidemiological study of human leukocyte antigen-DPB alleles in Hodgkin’s disease. Cancer Res 54: 5101–5105. Paffenbarger RS Jr., Wilson VO. 1955. Previous tonsillectomy and current pregnancy as they affect risk of poliomyelitis attack. Ann NY Acad Sci 61:856–868. Paffenbarger RS Jr., Wing AL, Hyde RT. 1977. Characteristics in youth indicative of adult onset Hodgkin’s disease. J Natl Cancer Inst 58:1489–1491. Pallesen G, Hamilton-Dutoit SJ, Rowe M, Young LS. 1991a. Expression of Epstein-Barr virus latent gene products in tumour cells of Hodgkin’s disease. Lancet 337:320–322. Pallesen G, Sandvej K, Hamilton-Dutoit SJ, Rowe M, Young LS. 1991b. Activation of Epstein-Barr virus replication in Hodgkin and Reed-Sternberg cells. Blood 78:1162–1165. Palli D, Trallori G, Bagnoli S, et al. 2000. Hodgkin’s disease risk is increased in patients with ulcerative colitis. Gastroenterology 119:647–653. Paltiel O, Schmit T, Adler B, et al. 2000. The incidence of lymphoma in first-degree relatives of patients with Hodgkin disease and non-Hodgkin lymphoma. Results and limitations of a registry-linked study. Cancer 88:2357–2366. Parkin DM, Ferlay J, Hamdi-Cherif M, et al., eds. 2002a. Cancer in Africa: Epidemiology and Prevention. IARC Scientific Publication No. 153. Lyon, France: International Agency for Research on Cancer. Parkin DM, Muir CS, Whelan SL, Gao YT, Ferlay J, Powell J, eds. 1992. Cancer Incidence in Five Continents, Vol VI. IARC Scientific Publication No. 120. Lyon, France: International Agency for Research on Cancer. Parkin DM, Stiller CA, Draper GJ, Bieber CA. 1988. The international incidence of childhood cancer. Int J Cancer 41:511–520. Parkin DM, Whelan SL, Ferlay J, Teppo L, Thomas D, eds. 2002b. Cancer Incidence in Five Continents. Vol VIII. IARC Scientific Publication No. 155. Lyon, France: International Agency for Research on Cancer. Petersen GR, Milham S Jr. 1974. Hodgkin’s disease mortality and occupational exposure to wood. J Natl Cancer Inst 53:957–958. Pike MC, Henderson BE, Casagrande J, et al. 1974. Infectious aspects of Hodgkin’s disease. [letter] N Engl J Med 290:341. Poppema S. 2005. Immunobiology and pathophysiology of Hodgkin lymphoma. Hematology (Am Soc Hematol Educ Program) p. 239–244. Poppema S, Visser L. 1994. Epstein-Barr virus positivity in Hodgkin’s disease does not correlate with an HLA A2-negative phenotype. Cancer 73:3059–3063. Portis T, Dyck P, Longnecker R. 2003. Epstein-Barr (EBV) LMP2A induces alterations in gene transcription similar to those observed in ReedSternberg cells in Hodgkin lymphoma. Blood 102:4166–4178. Poskanzer DC. 1965. Tonsillectomy and multiple sclerosis. Lancet ii:1264– 1266. Preciado MV, De Matteo E, Diez B, Menárguez J, Grinstein S. 1995. Presence of Epstein-Barr virus and strain type assignment in Argentine childhood Hodgkin’s disease. Blood 86:3922–3929. Razis DV, Diamond HD, Craver LF. 1959. Familial Hodgkin’s disease: Its significance and implications. Ann Intern Med 51:933–971. Razzouk BI, Gan YJ, Mendonça C, et al. 1997. Epstein-Barr virus in pediatric Hodgkin disease: Age and histiotype are more predictive than geographic region. Med Peiatr Oncol 28:248–254. Reed DM. 1902. On the pathological changes in Hodgkin’s disease with especial reference to its relation to tuberculosis. Johns Hopkins Hospital Report 10:133–196. Rickinson AB, Kieff E. 1996. Epstein-Barr virus. In: Fields BN, Knipe DM, Howley PM, et al., eds. Fields Virology. Vol 2, 3rd ed. Philadelphia: Lippincott-Raven Publishers, pp. 2397–2446. Ries LAG, Eisner MP, Kosary CL, et al., eds. 2002. SEER Cancer Statistics Review: 1973–1999. National Cancer Institute, Bethesda, MD. Available at: http://seer.cancer.gov/csr/1973 1999/. Ries LAG, Kosary CL, Hankey BF, Miller BA, Harras A, Edwards BK, eds. 1997. SEER Cancer Statistics Review: 1973–1994. NIH publication no. 97–2789. National Cancer Institute, Bethesda, MD.

Rosdahl N, Larsen SO, Clemmesen J. 1974. Hodgkin’s disease in patients with previous infectious mononucleosis: 30 years’ experience. Br Med J 2:253–256. Rosenfeld L. 1993. Thomas Hodgkin: Morbid Anatomist and Social Activist. Lanham, MD, Madison. Rowlings PA, Curtis RE, Passweg Jr. et al. 1999. Increased incidence of Hodgkin’s disease after allogeneic bone marrow transplantation. J Clin Oncol 17:3122–3127. Rubertone MV, Brundage JF. 2002. The defense medical surveillance system and the Department of Defense Serum Repository: Glimpses of the future of comprehensive public health surveillance. Am J Public Health 92:1900–1904. Samoszuk M, Nansen L. 1990. Detection of interleukin-5 messenger RNA in Reed-Sternberg cells of Hodgkin’s disease with eosinophilia. Blood 75:13–16. Samoszuk M, Ramzi E. 1993. IgE, Reed-Sternberg cells, and eosinophilia in Hodgkin’s disease. Leuk Lymphoma 9:315–319. Sanderson CJ. 1992. Interleukin-5, eosinophils, and disease. Blood 79:3101– 3109. Sankila R, Garwicz S, Olsen JH, et al. 1996. Risk of subsequent malignant neoplasms among 1,641 Hodgkin’s disease patients diagnosed in childhood and adolescence: A population-based cohort study in the five Nordic countries. J Clin Oncol 14:1442–1446. Scherr P, Gutensohn (Mueller) N, Cole P. 1984. School contact among persons with Hodgkin’s disease. Am J Epidemiol 120:29–38. Schimpff SC, Brager DM, Schimpff CR, Comstock GW, Wiernik PH. 1976. Leukemia and lymphoma patients linked by prior social contact. Evaluation using a case-control approach. Ann Intern Med 84:547–550. Schlaifer D, Rigal-Huguet F, Robert A, et al. 1994. Epstein-Barr virus in familial Hodgkin’s disease. Br J Haematol 88:636–638. Schlegel RJ, Bellanti JA. 1969. Increased susceptibility of males to infection. Lancet 2:826–827. Schmidt CA, Oettle H, Peng R, et al. 2000. Presence of human b- and g-herpes virus DNA in Hodgkin’s disease. Leuk Res 24:865–870. Schwaller J, Tobler A, Niklaus G, et al. 1995. Interleukin-12 expression in human lymphomas and non-neoplastic lymphoid disorders. Blood 85:2182–2188. Schwartzbaum JA, George SL, Pratt CB, David B. 1991. An exploratory study of environmental medical factors potentially related to childhood cancer. Med Pediatr Oncol 19:115–121. Serraino D, Franceschi S, Talamini R, et al. 1991. Socio-economic indicators, infectious diseases and Hodgkin’s disease. Int J Cancer 47:352–357. Shiramizu B, Chang CW, Cairo MS. 2001. Absence of human herpes-6 genome by polymerase chain reaction in children with Hodgkin disease: A Children’s Cancer Group Lymphoma Biology Study. J Pediatr Hematol/Oncol 23:282–285. Shpilberg O, Modan M, Modan B, Chetrit A, Fuchs Z, Ramot B. 1994. Familial aggregation of haematological neoplasms: A controlled study. Br J Haematol 87:75–80. Shugart YY. 1998. Anticipation in familial Hodgkin lymphoma. Am J Hum Genet 63:270–272. Shugart YY, Hemminki K, Vaittinen P, Kingman A, Dong C. 2000. A genetic study of Hodgkin’s lymphoma: An estimate of heritability and anticipation based on the familial cancer database in Sweden. Hum Genet 106:553–556. Siebert R, Louie D, Lacher M, Schlager A, Offit K. 1997. Familial Hodgkin’s and non-Hodgkin’s lymphoma: Different patterns in first-degree relatives. Leuk Lymphoma 27:503–507. Simpson CR, Anderson WJA, Helms PJ, et al. 2002. Coincidence of immunemediated diseases driven by Th1 and Th2 subsets suggests a common aetiology. A population-based study using computerized General Practice data. Clin Exp All 32:37–42. Skinnider BF, Elia AJ, Gascoyne RD, et al. 2001. Interleukin 13 and interleukin 13 receptor are frequently expressed by Hodgkin and Reed-Sternberg cells of Hodgkin’s lymphoma. Blood 97:250–255. Skinnider BF, Elia AJ, Gascoyne RD, et al. 2002a. Signal transducer and activator of transcription 6 is frequently activated in Hodgkin and ReedSternberg cells of Hodgkin’s lymphoma. Blood 99:618–626. Skinnider BF, Kapp U, Mak TW. 2002b. The role of interleukin-13 in classical Hodgkin’s lymphoma. Leuk Lymphoma 43:1203–1210. Skinnider BF, Mak TW. 2002. The role of cytokines in classical Hodgkin’s lymphoma. Blood 99:4283–4297. Sleckman B, Mauch PM, Ambinder RF, et al. 1998. Epstein-Barr virus in Hodgkin’s disease: Correlation of risk factors and disease characteristics with molecular evidence of viral infection. Cancer Epidemiol Biomarkers Prev 7:1117–1121. Smedby KE, Hjalgrim H, Melbye M, et al. 2005. Ultraviolet radiation exposure and risk of malignant lymphomas. J Natl Cancer Inst 97:199–209.

Hodgkin Lymphoma Smith PG, Pike MC. 1974. Case clustering in Hodgkin’s disease: A brief review of the present position and report of current work in Oxford. Cancer Res 34:1156–1160. Smith PG, Pike MC, Kinlen LJ, Jones A, Harris R. 1977. Contacts between young patients with Hodgkin’s disease. A case-control study. Lancet 2:59–62. Spieker T, Kurth J, Küppers R, Rajewsky K, Bräuninger A, Hansmann M-L. 2000. Molecular single-cell analysis of the clonal relationship of small Epstein-Barr virus-infected cells and Epstein-Barr virus-harboring Hodgkin and Reed-Sternberg cells in Hodgkin’s disease. Blood 96: 3133–3138. Spitz MR, Sider JG, Johnson CC, Butler JJ, Pollack ES, Newell GR. 1986. Ethnic patterns of Hodgkin’s disease incidence among children and adolescents in the United States, 1973–1982. J Natl Cancer Inst 76: 235–239. Staal SP, Ambinder R, Beschorner WE, Hayward GS, Mann R. 1989. A survey of Epstein-Barr virus DNA in lymphoid tissue: Frequent detection in Hodgkin’s disease. Am J Clin Pathol 91:1–5. Stein H, Diehl V, Marafioti T, Jox A, Wolf J, Hummel M. 1999. The nature of Reed-Sternberg cells, lymphocytic and histiocytic cells. In: Mauch PM, Armitage JO, Diehl V, Hoppe RT, Weiss LM, eds. Hodgkin’s Disease. Philadelphia: Lippincott Williams and Wilkins, pp. 121–138. Sternberg C. 1898. Uber eine eigenartige unter dem Bilde der Pseudoleukamie verlaufende Tuberculose des lymphatischen apparates. Zeitschrift fur Heilkunde 19:21–91. Strachan DP, Taylor EM, Carpenter RG. 1996. Family structure, neonatal infection, and hay fever in adolescence. Arch Dis Child 74:422–426. Straus SE, Jaffe ES, Puck JM, et al. 2001. The development of lymphomas in families with autoimmune lymphoproliferative syndrome with germline Fas mutations and defective lymphocyte apoptosis. Blood 98:194–200. Sultan SM, Ioannou Y, Isenberg DA. 2000. Is there an association of malignancy with systemic lupus erythematosis? An analysis of 276 patients under long-term review. Rheumatology 39:1147–1152. Tavani A, La Vecchia C, Gallus S, et al. 2000. Red meat intake and cancer risk: A study in Italy. Int J Cancer 86:425–428. Tavani A, Pregnolato A, Negri E, et al. 1997. Diet and risk of lymphoid neoplasms and soft tissue sarcomas. Nutr Cancer 27:256–260. Travis LB, Hill DA, Dores GM, et al. 2003. Breast cancer following radiotherapy and chemotherapy among young women with Hodgkin disease. JAMA 290:465–475. Uhara H, Sato Y, Mukai K, et al. 1990. Detection of Epstein-Barr virus DNA in Reed-Sternberg cells of Hodgkin’s disease using the polymerase chain reaction and in situ hybridization. Jpn J Cancer Res 81:272–278. Van Der Ouderaa FJ, Buytenhek M, Nugteren DH, Van Dorp DA. 1980. Acetylation of prostaglandin endoperoxide synthetase with acetylsalicylic acid. Eur J Biochem 109:1–8. Vasef MA, Kamel OW, Chen Y-Y, Medeiros LJ, Weiss LM. 1995. Detection of Epstein-Barr virus in multiple sites involved by Hodgkin’s disease. Am J Pathol 147:1408–1415. Veltri RW, Shan SH, McClung JE, Klingberg WD, Sprinkle PM. 1983. EpsteinBarr virus, fatal infectious mononucleosis, and Hodgkin’s disease in siblings. Cancer 51:509–520. Vianna NJ, Greenwald P, Davies JN. 1971. Extended epidemic of Hodgkin’s disease in high school students. Lancet 1:1209–1211. Vianna NJ, Greenwald P, Polan A, Koeogh MD, Davies JNP. 1974. Tonsillectomy and Hodgkin’s disease. Lancet ii:168–169. Vianna NJ, Polan AK. 1973. Epidemiologic evidence for transmission of Hodgkin’s disease. N Engl J Med 289:499–502.

897

Vineis P, Crosignaru P, Sacerdote C, et al. 2000. Hematopoietic cancer and medical history: A multicentre case control study. J Epidemiol Community Health 54:431–436. Von Mutius E, Martinez FD, Fritzsch C, Nicolai T, Reitmeir P, Thiemann H-H. 1994. Skin test-reactivity and number of siblings. Br Med J 308: 692–695. Von Mutius E, Weiland SK, Fritzsch C, Duhme H, Keil U. 1998. Increasing prevalence of hay fever and atopy among children in Leipzig, East Germany. Lancet 351:862–866. Washburn TC, Medearis DN, Childs B. 1965. Sex differences in susceptibility to infections. Pediatrics 35:57–64. Weinreb M, Day PJR, Niggli F, et al. 1996. The role of Epstein-Barr virus in Hodgkin’s disease from different geographical area. Arch Dis Child 74:27–31. Weiss LM, Chan JKC, MacLennan K, Warnke RA. 1999. Pathology of classical Hodgkin’s disease. In: Mauch PM, Armitage JO, Diehl V, Hoppe RT, Weiss LM, eds. Hodgkin’s Disease. Philadelphia: Lippincott Williams and Wilkins, pp. 101–120. Weiss LM, Chen Y-Y, Liu X-F, Shibata D. 1991. Epstein-Barr virus and Hodgkin’s disease: A correlative in situ hybridization and polymerase chain reaction study. Am J Pathol 139:1259–1265. Weiss LM, Strickler JG, Warnke RA, Purtilo DT, Sklar J. 1987. Epstein-Barr virus DNA in tissues of Hodgkin’s disease. Am J Pathol 129:86–91. Westergaard T, Melbye M, Pedersen JB, Frisch M, Olsen JH, Andersen PK. 1997. Birth order, sibship size and risk of Hodgkin’s disease in children and young adults: A population-based study of 31 million person-years. Int J Cancer 72:977–981. White L, McCourt BA, Isaacs H, Siegel SE, Stowe SM, Higgins GR. 1983. Patterns of Hodgkin’s disease at diagnosis in young children. Am J Pediatr Hematol Oncol 5:251–257. Wiernik PH, Ashwin M, Hu XP, Paietta E, Brown K. 2001. Anticipation in familial chronic lymphocytic leukemia. Br J Haematol 113:407–414. Wiernik PH, Wang SQ, Hu XP, Marino P, Paietta E. 2000. Age of onset evidence for anticipation in familial non-Hodgkin’s lymphoma. Br J Haematol 108:72–79. Wilkinson JD, Fleming LE, MacKinnon J, et al. 2001. Lymphoma and lymphoid leukemia incidence in Florida children: Ethnic and racial distribution. Cancer 91:1402–1408. Wolf J, Kapp U, Bohlen H, et al. 1996. Peripheral blood mononuclear cells of a patient with advanced Hodgkin’s lymphoma give rise to permanently growing Hodgkin-Reed-Sternberg cells. Blood 87:3418–3428. Wolk A, Gridley G, Svensson M, et al. 2001. A prospective study of obesity and cancer risk. Cancer Causes Control 12:13–21. Yang P, Grufferman S, Schwartz AG, et al. 1993. A genetic analysis of Hodgkin’s disease. Am J Hum Genet 53:A391. Yates JL, Bunting CH. 1915. The rational treatment of Hodgkin’s disease. JAMA 64:1953–1961. Yin MJ, Yamamoto Y, Gaynor RB. 1998. The anti-inflammatory agents aspirin and salicylate inhibit the activity of IkB kinase-b. Nature 396:77–80. Young LS. 1994. Ludwig/CRC EBV Meeting. EBV Report 1:10–12. Zack MM Jr, Heath CW Jr., Andrews MD, Grivas AS Jr., Christine BW. 1977. High school contact among persons with leukemia and lymphoma. J Natl Cancer Inst 59:1343–1349. Zhou XG, Sandvej K, Li PJ, et al. 2001. Epstein-Barr virus (EBV) in Chinese pediatric Hodgkin’s disease: Hodgkin’s disease in young children is an EBV-related lymphoma. Cancer 92:1621–1631. Zwitter M, Zakelj MP, Kosmelj K. 1996. A case-control study of Hodgkin’s disease and pregnancy. Br J Cancer 73:246–251.

46

Non-Hodgkin Lymphoma PATRICIA HARTGE, SOPHIA S. WANG, PAIGE M. BRACCI, SUSAN S. DEVESA, AND ELIZABETH A. HOLLY

N

on-Hodgkin lymphomas (NHL) are expected to account for 4.2% of cancer diagnoses and 3.3% of cancer deaths in the United States (US) in 2006 (Jemal et al., 2006). With approximately 58,870 cases diagnosed and 18,840 deaths from NHL expected in 2006 (Jemal et al., 2006), this group of malignancies constitutes a serious public health problem in the United States as it does in most developed countries. The lifetime cumulative incidence is about 2% in the US population. Non-Hodgkin lymphomas are strikingly heterogeneous in clinical course and in etiology because NHLs encompass a variety of histologically distinct forms. For several specific lymphomas, the causes or major contributing factors have been found. Burkitt lymphoma, mucosa-associated lymphoid tissue (MALT) lymphoma, primary effusion lymphoma, and adult T-cell leukemia/lymphoma each have revealed distinctive etiologies. On the other hand, human immunodeficiency virus (HIV) infection and several other risk factors increase the risk of multiple histologic types, demonstrating that common etiology does accompany histologic diversity of NHL. However, the majority of cases of all lymphoma and the majority of individual histologic types remain unexplained. Non-Hodgkin lymphoma occurrence has risen steadily and dramatically in most parts of the world for several decades. For example, incidence rates in the United States have almost doubled since the early 1970s (Devesa and Fears, 1992; Ries et al., 2003), and changes in detection and diagnosis of lymphoma fail to account for the rise (Hartge and Devasa, 1992). In many populations, an epidemic of lymphoma caused by HIV infections was added to the unexplained longterm rise of lymphoma (Groves et al., 2000). As the increase related to HIV infection subsides with successful HIV therapies in developed countries, it is unclear whether the long-term trend will reach a plateau, begin a sustained decline, or resume its increase (Eltom et al., 2002).

CLASSIFICATION Anatomic Distribution Lymphomas originate in lymphatic organs during progressive stages of lymphoid development. B and T lymphocytes develop and mature in the primary, or central, lymphoid organs, the B-cells in the bone marrow and the T-cells in the thymus. The mature lymphocytes then migrate to the secondary, or peripheral, lymphoid organs including the lymph nodes, spleen, blood, tonsils, and other lymphoid tissues located in the gastrointestinal and respiratory tract (Harris et al., 2001). It is estimated that 55%–75% of lymphomas worldwide present in the lymph nodes (Evans and Hancock, 2003). In the Surveillance, Epidemiology, and End Results (SEER) database, nodal disease in recent years accounted for approximately 65% of all lymphomas in the United States, as shown in Table 46–1. Recently, the incidence of extranodal lymphoma has risen especially rapidly, with the most common sites of origin in the skin, the gastrointestinal tract, and the central nervous system.

HISTOLOGIC CLASSIFICATION The World Health Organization (WHO) classification of lymphoid neoplasms has become the international standard (Fritz et al., 2000;

898

Jaffe et al., 2001). The nomenclature for each distinctive disease entity, shown in Table 46–2, now reflects the postulated cell lineage and stage of differentiation. In this chapter, NHL refers to ICDO-2 codes 9590–9595, 9670–9677, 9680–9688, 9690–9698, 9700–9717, 9823, and 9827 (Percy et al., 1990). The WHO scheme was accepted as the standard for classification of NHL following the demonstration of its reproducibility and clinical relevance (Chan, 2001; Non-Hodgkin’s Lymphoma Classification Project, 1997). In the past, the great heterogeneity of NHL created enormous challenges for developing a consistent and reliable disease classification system, resulting in several substantial revisions between 1966 and 1999 (Harris et al., 1994, 1999; National Cancer Institute, 1982; Rappaport, 1966). The WHO classification combines morphologic, immunophenotypic, genetic, and clinical features to group lymphomas into three major categories: B-cell neoplasms, T- and NK-cell neoplasms, and Hodgkin lymphoma. Within these categories, lymphomatous and leukemic entities are recognized as different manifestations of the same disease, for example, B-cell chronic lymphocytic leukemia and B-cell small lymphocytic lymphoma. Within the B- and T-cell neoplasms, the WHO scheme distinguishes between precursor and mature neoplasms. The precursor neoplasms correspond to the earliest stages of differentiation; the mature neoplasms correspond to the later differentiated stages. Worldwide, it is estimated that mature B-cell lymphomas account for over 90% of all lymphomas, with the largest two subsets, diffuse large B-cell lymphoma (DLBCL) and follicular lymphoma, accounting for approximately 31% and 22% of all lymphomas, respectively (Jaffe et al., 2001; Non-Hodgkin’s Lymphoma Classification Project, 1997). T-cell and NK neoplasms are relatively uncommon and account for roughly 12% of all NHL (Non-Hodgkin’s Lymphoma Classification Project, 1997). Patterns and time trends vary markedly by histologic subtype (Morton et al., 2005a).

MOLECULAR CHARACTERISTICS Recurring chromosomal abnormalities are a hallmark of NHL. Characteristic chromosomal translocations are strongly associated with specific histologic subtypes of NHL, and gene deletions, amplifications, and somatic mutations have been demonstrated (Chaganti et al., 2000). The consequences of these various chromosomal aberrations include the activation or overexpression of oncogenes, the inactivation of tumor suppressor genes, or the production of chimeric proteins (Dalla-Favera and Gaidano, 2001; Hauke and Armitage, 2000). Several of the numerous chromosomal aberrations that have been reported for NHL have been extensively documented, as shown in Table 46–3. These include one of the first translocations identified, t(8;14) (q24; q32), now known to occur in 80%–85% of Burkitt lymphoma specimens (Vega and Mederos, 2003). Other wellcharacterized translocations include t(2;5) (p23; q35) in anaplastic large T-cell lymphomas, t(11;14) (q13;q32) in chronic lymphocytic leukemia/small lymphocytic lymphoma (CLL/SLL), t(14;18) (q32;q21) in diffuse large B-cell lymphoma (DLBCL) and in the majority of follicular lymphomas, t(9;14) (p13;q32) in lymphoplasmacytic lymphoma, t(11;14) (q13;q32) in most mantle-cell lymphomas, and t(11;18) (q21;q21) in MALT lymphomas (Vega and Mederos, 2003). Genes affected by these translocations include, for

Non-Hodgkin Lymphoma Table 46–1. Non-Hodgkin Lymphoma Cases and Incidence Rates 1978–1989 and 1990–2000 in the Nine SEER Areas by Site of Origin* 1978–1989

1990–2000

1990–2000 : 1978–1989

Site of Origin

Cases

Rate

Cases

Rate

Rate Ratio

All Sites Nodal Extranodal—Total Skin Stomach Brain Small Intestine Lung Soft Tissue Colon Eye Thyroid

34452 25736 (D) 1606 1537 682 675 270 300 395 267 295

14.8 11.0 3.8 0.7 0.7 0.3 0.3 0.1 0.1 0.2 0.1 0.1

49955 32947 (D) 3009 2292 1934 1014 655 726 700 630 398

19.2 12.7 6.5 1.2 0.9 0.7 0.4 0.2 0.3 0.3 0.2 0.2

1.3 1.2 1.7 1.7 1.3 2.3 1.3 2.0 3.0 1.5 2.0 2.0

Source: National Cancer Institute, 2003. *Per 100,000 person-years, age-adjusted to the 2000 US standard.

example, c-MYC overexpression from t(8;14) (q21;q32), BCL-2 overexpression from t(14;8) (q32;q21), cyclin D1/BCL-1 overexpression in t(11;14) (q13;q32), PAX-5 deregulation in t(9;14) (p13;q32), and BCL-6 overexpression in t(3;14) (q27;q32). Fused genes leading to chimeric proteins also are observed, such as with t(11;18) (q21;q21) and t(2;5) (p23;q35), resulting in API2 oncogene overexpression and ALK deregulation, respectively (Evans and Handcock, 2003). Chromosomal gains, losses, and specific somatic mutations also independently lead to the deregulation of gene expression (Chaganti et al., 2000; Misra et al., 2000; Tiirikainen et al., 2001). Gene amplification occurs commonly in NHL, resulting in the overexpression of REL (2p12), c-MYC (8q24), BCL-2 (18q21), and other protooncogenes (Rao et al., 1998). Gene deletions have been less extensively studied but analyses by loss of heterozygosity have detected deletions on chromosomes 6, 7, and 13. Lymphomas can be distinguished from one another not just by their chromosomal abnormalities, but also by their gene expression profiles

899

(Staudt, 2002). Two distinct molecular profiles have been determined for DLBCL and were found to be derived from different stages of Bcell differentiation, germinal center B-like DLBCL, and activated Blike DLBCL. Survival appears to be better for patients with germinal center B-like DLBCL (Alizadeh et al., 2000). In addition, germinal center B-like DLBCL is distinctly associated with t(14;18) (q32;q21), BCL-2 overexpression and REL amplification, whereas activated Blike DLBCL shows none of these (Staudt, 2003). The rapidly evolving technologies for measuring gene and protein expression are likely to prove useful for future epidemiologic studies of risk.

PATTERNS OF OCCURRENCE Time Trends The descriptive epidemiology of NHL exhibits several striking features. Most notably, incidence and mortality rates reported in many areas increased dramatically during the last half of the twentieth century. The 50-year increase in US mortality rates from NHL, shown in Figure 46–1, occurred in men and women. Mortality rates are subject to improvements in survival, but they may be less sensitive than incidence rates to improvements in diagnostic practices. Thus, increasing mortality rates provide strong evidence of a real increase in the occurrence of lymphoma. Artifacts of diagnosis and prevalence of known causes do not explain the increase in NHL (Hartge and Devesa, 1992). The long-term increase in overall lymphoma rates has included especially steep increases in lymphoma incidence and mortality among older people. As shown in Figure 46–2, for example, among white men aged 75–84, the mortality rate rose from 18.9 to 78.4 deaths per 100,000 person-years. By contrast, rates in white men aged 35–44 rose from 2.5 to 2.8. The pattern of more rapid increase at older ages appears in incidence and mortality data and among both women and men. Using the same data, one can also compare age-specific rates among different birth cohorts. The cohorts born before 1900 show the most dramatic increase in age-specific rates over time, and those born after 1930 show little or no increase. In the United States, Canada, and the United Kingdom, the calendar year of diagnosis appears to have

Table 46–2. World Health Organization Classification of Lymphoid Neoplasms B-cell neoplasms

T-cell and NK-neoplasms

precursor b-cell neoplasms

precursor t-cell neoplasms

Precursor B-lymphoblastic leukemia/lymphoma

mature (peripheral) b-cell neoplasms Chronic lymphocytic leukemia Small lymphocytic lymphoma B-cell prolymphocytic leukemia Lymphoplasmacytic lymphoma Splenic marginal zone lymphoma Hairy cell leukemia Plasma cell myeloma Solitary plasmacytoma of bone Extraosseous plasmacytoma Extranodal marginal zone B-cell lymphoma of mucosaassociated lymphoid tissue (MALT-lymphoma) Nodal marginal zone B-cell lymphoma Follicular lymphoma Mantle cell lymphoma Diffuse large B-cell lymphoma Mediastinal (thymic) large B-cell lymphoma Intravascular large B-cell lymphoma Primary effusion lymphoma Burkitt lymphoma/leukemia

b-cell proliferations of uncertain malignant potential Lymphomatoid granulomatosis Post-transplant lymphoproliferative disorder, polymorphic Source: Fritz et al., 2000.

Precursor T lymphoblastic leukemia/lymphoma Blastic NK-cell lymphoma

mature t-cell and nk-cell neoplasms T-cell prolymphocytic leukemia T-cell large granular lymphocytic leukemia Aggressive NK-cell leukemia Adult T-cell leukemia/lymphoma Extranodal NK/T-cell lymphoma, nasal type Enteropathy-type T-cell lymphoma Hepatosplenic T-cell lymphoma Subcutaneous panniculitis-like T-cell lymphoma Mycosis fungoides Sezary syndrome Primary cutaneous anaplastic large cell lymphoma Peripheral T-cell lymphoma, unspecified Angioimmunoblastic T-cell lymphoma Anaplastic large cell lymphoma

t-cell proliferation of uncertain malignant potential Lymphomatoid papulosis

900

PART IV: CANCER BY TISSUE OF ORIGIN

Table 46–3. Examples of Chromosomal Translocations Associated with Selected Non-Hodgkin Lymphoma Subtypes Lymphoma Subtype Anaplastic large T-cell lymphoma Burkitt lymphoma Chronic lymphocytic leukemia/ small lymphocytic lymphoma Diffuse large B-cell lymphoma

Follicular lymphoma

Lymphoplasmacytic lymphoma Mantle cell lymphoma Mucosa-associated lymphoid tissue type lymphoma

Chromosomal Translocation

Percentage Reported in Cases

Gene Involved

t(2;5)(p23;q35) t(8;14)(q24;q32) t(2;8)(p11;q24) t(8;22)(q24;q11) t(11;14)(q13;q32) t(14;19)(q32;q13) t(14;18)(q32;q21) t(8;14)(q24;q32) t(3;14)(q27;q32); t(3q27) t(10;14)(q24;q32) t(1;14)(q21;q32) t(14;18)(q32;q21) t(2;18)(p11;q21) t(18;22)(q21;q11) t(3;14)(q27;q32) t(3q27) t(9;14)(p13;q32) t(11;14)(q13;q32) t(11;18)(q21;q21) t(1;14)(p22;q32)

60% 80% 15% 5%

NPM, ALK c-MYC BCL-1 BCL-3 BCL-2 c-MYC BCL-6 NFKB2 MUC1 BCL-2

20% 10%–30% 5%–10% 90%

BCL-6 6%–26% 50% 70% 50%

PAX-5 BCL-1 API2, MLT BCL-10

Adapted with permission from Harris et al., 2001. Adapted from Chaganti et al., 2000.

played a larger role, but the birth cohort effect also contributed (Liu et al., 2003; McNally et al., 1999). Time trends have differed by histology and by site of origin, but the use of older classification schemes makes it impossible to dissect completely the long-term time trends of different forms of lymphoma. Between the late 1970s and about 2000, incidence of many types of NHL rose. High-grade NHL incidence increased the most; incidence

of both follicular small-cell and diffuse small-cell declined (Groves et al., 2000). These increases largely reflect the influence of the Acquired Immunodeficiency Syndrome (AIDS) epidemic on lymphoma rates. The superimposed HIV-NHL epidemic appeared most starkly in areas with a high prevalence of HIV infection, for example, San Francisco County (Eltom et al., 2002). Similarly, when Highly Active Antiretroviral Therapy (HAART) became widely available, the 1000

100

Males

Rates per 100,000 person-years

Rates per 100,000 person-years

Females

10

100

Age 85+ 75-84 65-74 55-64 45-54 35-44 25-34 10

1996–00

1991–95

1981–85 1986–90

1971–75 1976–80

1961–65 1966–70

Year of death

1

1951–55 1956–60

1996–00

1991–95

1981–85

1986–90

1976–80

1971–75

1966–70

1956–60

1961–65

1951–55

1

Year of death

Figure 46–1. Trends in US non-Hodgkin lymphoma mortality rates (age-adjusted to 2000 US Standard) among men and women from 1951–1955 to 1996–2000. (Source: Mortality data provided by NCHS (http://www.cdc.gov/nchs.)

Figure 46–2. Trends in age-specific non-Hodgkin lymphoma mortality among US white men, 1951–1955 to 1990–2000. (Source: Mortality data provided by NCHS (http://www.cdc.gov/nchs.)

901

Non-Hodgkin Lymphoma

Figure 46–3. Trends in NHL incidence rates (age-adjusted to 2000 US Standard) among white men in 9 SEER areas by year of diagnosis and selected histology type, 1978–2000. (Source: SEER*Stat, 2003.)

contribution of HIV-related lymphoma decreased rapidly (Clarke and Glaser, 2001). Eltom et al. demonstrated that the emergence of HIV and then the introduction of HAART created a classic epidemic curve for several lymphomas caused by uncontrolled HIV infection, especially those that were otherwise rare, for example, immunoblastic, Burkitt, and central nervous system (CNS) lymphomas. Figure 46–3 contrasts the time trends for DLBCL, immunoblastic lymphoma, and Burkitt lymphoma. The incidence of NHL diagnosed at extranodal sites increased faster than the incidence of nodal lymphomas both in the United States, as shown in Table 46–1, and elsewhere around the world (Newton et al., 1997). CNS lymphomas have drawn attention because of the increasing incidence of other brain malignancies and the possibility of distinct etiology (Cote et al., 1996; Kadan-Lottick et al., 2002; Olson et al., 2002b). Incidence of CNS lymphoma rose faster than incidence of other lymphomas both among people with AIDS and among the rest of the population. CNS lymphoma incidence increased by 12% per year between 1973 and 1984, the era when diagnosis relied on computed tomography scans, then increased less rapidly during the subsequent magnetic resonance imaging era (Olson et al., 2002b). For the most common forms of lymphoma, which may have been caused by HIV infection or by other factors, it is harder to disentangle the more recent AIDS epidemic from the longer unexplained rise. From the beginning of the SEER Program in the early 1970s, the incidence rates rose in all sex-race groups and accelerated somewhat during the beginning of the AIDS epidemic, especially among white men, as shown in Figure 46–4. During the late 1990s, the rates in whites and in black women plateaued, and the rates in black men fell. The plateau in rates is welcome news, but it is too soon to forecast whether the long-term increase will resume.

Age, Sex, Ethnicity Advancing age is a major risk factor for lymphoma, as it is for most cancers. Among men and women and in all ethnic groups, risk rises exponentially with age. For example, among white men in the United States today, the annual incidence is 3 per 100,000 in the age range 20–24, in contrast to 135 in the age range 80–84 (National Cancer Institute, 2003). The age-specific incidence curves for a few lymphomas deviate notably from the pattern seen overall. For instance,

100

10

White Men White Women Black Men Black Women

1

97–98

Year of diagnosis

91–92

1999

1996

1993

1990

1987

1984

1981

1978

0.1

85–86

1

79–80

All NHL Diffuse Large B-cell Immunoblastic Burkitt

73–74

10

SLL and CLL, which are combined in the WHO scheme, both show a particularly striking increase in incidence with older age. Lymphoblastic lymphoma incidence is high in childhood and early adulthood, falls to a trough in midlife, and rises again later in life (Groves et al., 2000). Risk of immunoblastic lymphoma rises until 40, plateaus until 60, and rises somewhat thereafter. Men are more likely than women to develop NHL, generally by a 3 : 2 margin. Table 46–4 summarizes the patterns of incidence, mortality, and survival in the United States during the period 1996–2000 for all NHL subtypes combined (Ries et al., 2003). The overall incidence rate during that period, 19.1 cases diagnosed per 100,000 person-years, reflected higher rates in white men (24.5) and black men (18.6) and lower rates in white women (16.4) and black women (11.6). There was little racial difference in the rates at ages under 65 : 12.4 among both black and white men and ranging between 7.1 and 7.7 among women. However, at older ages, rates were considerably higher among whites than blacks: 108.5 vs. 61.2 among men, and 76.4 vs. 42.6 among women, respectively. Mortality rates generally were about one-half the corresponding incidence rates, with similar patterns apparent. Overall 5-year relative survival among patients diagnosed with NHL was 56.1%, ranging from a high of 60.9% among white women to a low of 42.3% among black men. Survival rates were higher among patients diagnosed at younger compared with older ages and exceeded 72% among white women diagnosed under age 65 years. The larger size of the population of elderly women counterbalanced the lower rates in women, so about 45% of newly diagnosed non-Hodgkin lymphoma patients were women during this period (SEER*Stat, 2003). Similarly, cumulative lifetime incidence was nearly as high in white women (1.8%) as in white men (2.1%) (Ries et al., 2003). SEER incidence rates by NHL histology diagnosed between 1996 and 2000 reveal the clear predominance of mature B-cell lymphomas, as shown in Table 46–5. DLBCL accounted for the largest group of NHL in the United States, with rates of 7.5 for men and 5.0 for women. Follicular lymphomas were the second most common subtype, with rates of 3.5 and 2.9 in men and women, respectively. For most other

Rates per 100,000 population

Rates per 100,000 person-years

100

Year of diagnosis

Figure 46–4. Trends in non-Hodgkin lymphoma incidence rates (ageadjusted to 2000 US Standard) in 9 SEER areas by race and sex, 1973–1974 to 1999–2000. (Source: SEER*Stat, 2003.)

902

PART IV: CANCER BY TISSUE OF ORIGIN

Table 46–4. Non-Hodgkin Lymphoma Incidence, Mortality and Survival Rates by Age, Race, and Sex Whites

Incidence* Under 65 65 and over All ages Mortality† Under 65 65 and over All ages Survival‡ Under 65 65 and over All ages

Blacks

All Races

Males

Females

Total

Males

Females

Total

Males

Females

Total

12.4 108.5 24.5

7.7 76.4 16.4

10.1 89.1 20.1

12.4 61.2 18.6

7.1 42.6 11.6

9.6 49.5 14.6

12.0 103.0 23.5

7.4 72.1 15.6

9.7 84.4 19.1

3.6 62.3 11.1

2.1 43.2 7.3

2.9 50.6 8.9

3.9 33.4 7.6

2.0 22.4 4.6

2.9 26.5 5.9

3.6 59.6 10.7

2.1 41.1 7.0

2.8 48.3 8.6

56.8 49.8 54.1

72.3 51.6 60.9

62.6 50.8 57.2

42.8 40.9 42.3

61.3 40.4 54.9

49.1 40.5 47.1

55.3 48.7 52.9

70.8 50.6 60.1

61.1 49.8 56.1

Source: Ries et al., 2003. *1996–2000 SEER 12 areas. Rates are per 100,000 person-years and are age-adjusted to the 2000 US standard population. † 1996–2000 United States. NCHS public use data file. Rates are per 100,000 person-years and are age-adjusted to the 2000 US standard population. ‡ 5-year relative survival rates (percent), cases diagnosed during 1992–1999 and followed up through 2000. Rates are from the SEER 9 areas.

types of B-cell lymphoma, incidence was less than 1.0. T- and NKcell lymphomas were also rare. Male predominance in incidence was evident for most histologic subtypes, with rates generally at least 50% higher among men, as shown in Table 46–5. Exceptions were follicular lymphoma, marginal zone lymphoma, and angioimmunoblastic lymphoma, with male : female rate ratios of 1.2, 1.0, and 1.0, respectively. Burkitt and mantle cell lymphomas exhibited the most marked excess among men, with male : female ratios of 4 and 3. Among US subpopulations defined by race or ethnicity, as shown in Figure 46–5, NHL incidence and mortality rates varied substantially. Incidence and mortality were highest in the white, nonHispanic group and lowest in American Indians. Intermediate rates were recorded for Hispanic, Black, and Asian/Pacific Islander populations.

Geographic Variation Geographic variation in lymphoma rates suggests the importance of environmental effects, some known, such as Human T-cell Lym-

photrophic Virus type I (HTLV-I) in the Caribbean and Japan, or HIV in San Francisco in the early 1980s, but most remain unknown. In the United States, deaths per 100,000 person-years among white men varied from about five deaths in the south to 10 or more in the northcentral areas, as shown in Figure 46–6. Among white women, a similar geographic pattern appears, suggesting that the underlying environmental or behavioral factors are likely to be common to men and women. The geographic pattern of lymphoma mortality within the United States contrasts markedly with those for melanoma and non-melanoma skin cancer (Hartge et al., 1996), providing evidence against a strong association between ultraviolet radiation and lymphoma risk. Some of the areas with the highest mortality rates are heavily agricultural, so pesticides or other farm-related exposures may play a role. International variation in lymphoma risk, shown in Figure 46–7, exceeds the diversity within the United States (Parkin et al., 2003). In general, incidence rates are lowest in Asia and highest in Western Europe and North America. From the mid-1970s to the mid-1990s,

Table 46–5. Non-Hodgkin Lymphoma Incidence Rates in 12 SEER Areas during 1996–2000 by Sex and Four-Digit Histology Codes* Males

All histologies combined† Mature B-cell lymphomas Small B lymphocytic (9670, 9823)‡ Lymphoplasmacytic (9671) Mantle cell (9673, 9674, 9677) Diffuse large B-cell (9680, 9681, 9682, 9683, 9688, 9712) Diffuse large B-cell, immunoblastic (9684) Burkitt (9687) Marginal zone B-cell (9715, 9711, 9710) Follicular lymphoma (9690–9698) Mature T-cell and NK-cell lymphomas Mycosis fungoides (9700) Mature (peripheral) T-cell (9702, 9703, 9704, 9706, 9707) Angioimmunoblastic (9705) Cutaneous (9709) Anaplastic (9714) Other and unspecified T-cell and NK-cell (9708, 9713,9701, 9716, 9717, 9827) Precursor B-cell neoplasms Precursor B-cell lymphoblastic lymphoma (9685–9686) Lymphoma, not otherwise specified (9672, 9675, 9676, 9590–9595) Source: National Cancer Institute, 2003; Percy et al., 1990. *Rates are per 100,000 person-years and age-adjusted to the 2000 US standard. † Does not equal the sum due to rounding. ‡ Numbers in parentheses indicate the ICDO-2 codes included.

Females

Count

Rate

Count

Rate

Male/Female Ratio

18028 12796 1212 291 660 5750 784 384 1004 2711 1442 499 253 54 276 307 53 466 466 3324

23.5 16.7 1.7 0.4 0.9 7.5 1.0 0.4 1.3 3.5 1.9 0.6 0.3 0.1 0.4 0.4 0.1 0.5 0.5 4.4

14923 10906 953 226 324 4825 438 136 1237 2767 1080 401 173 51 211 207 37 224 224 2713

15.6† 11.3 1.0 0.2 0.3 5.0 0.5 0.1 1.3 2.9 1.1 0.4 0.2 0.1 0.2 0.2 0.0 0.2 0.2 2.8

1.5 1.5 1.7 2.0 3.0 1.5 2.0 4.0 1.0 1.2 1.7 1.5 1.5 1.0 2.0 2.0 • 2.5 2.5 1.6

903

Non-Hodgkin Lymphoma Men

25.1

Women White/nonHispanic

11.2 19.0 18.6

6.8 5.9 10

5

0

Incidence

11.0 4.2

American Indian

8.1

15

11.6 4.6

Asian/Pacific Islander

16.8

20

5.5

Black

7.6

25

13.2

Hispanic

8.4

30

16.6 7.4

Mortality 7.2

4.2 0

5

10

15

20

25

30

Figure 46–5. Non-Hodgkin lymphoma incidence and mortality rates by race/ethnicity and sex from 1996–2000. (Source: Ries et al., 2003.)

rates increased worldwide. Among men, the increases were most rapid in the United States and Italy, exceeding 85%. Among women, all rates shown rose by at least 50%. In all except the United States, the proportional increases among women were equal to or greater than among men. Migrants to the United States from China had lower incidence than US-born whites, as did their descendants (Herrinton, 1998). Similarly, migrants from Japan had lower incidence rates than US-born whites, and rates did not increase consistently among the migrants’ descendants. Incidence rates also were decreased for Filipino-Americans in either the first or later generations compared with US-born whites. For follicular lymphoma but not other types, there was some evidence of the incidence gap closing with the later generations of Asian Americans. Descriptive small-area studies of lymphoma have been conducted to detect clustering that might reveal either an infection or a chemical point-source of exposure. An Italian study (Masala et al., 1996), reported slight elevations in an agricultural area and in a heavily industrialized area. A cluster investigation of leukemia and lymphoma in children in Northern England reported no direct contact among cases, yet closer than expected distance to neighbor cases prompted the authors to suggest transmissible agents in the population of the child’s contacts (Alexander et al., 1992). A later study of childhood leukemia and lymphoma near large rural construction sites showed an increase in incidence in the year following construction and the associated influx of population and introduction of new transmissible agents (Kinlen et al., 1995).

Environment Retroviruses Human T-cell Lymphotrophic Virus Type I (HTLV-I) was the first retrovirus established as a cause of lymphoma, specifically of adult Tcell leukemia/lymphoma (ATL) (Cleghorn et al., 1995; Gallo et al., 1981; Manns et al., 1993; Mueller and Blattner, 1997). HTLV-I infection is endemic in southern Japan and in the Caribbean. It also is found in isolated parts of Africa, the Middle East, South America, Melanesia, and Papua New Guinea, but otherwise is rare, typically with seroprevalence lower than 1.0%. HTLV-I can be transmitted from mother to infant, by sexual activity, with easier transmission from men to women; through injection drug use; and through transfusion of whole blood, but not plasma. Seroprevalence in endemic populations increases with age. There is a 20–40-year incubation period for ATL, and the age peak for T-cell NHL varies by geographic region (Mueller and Blattner, 1997). Although fewer than 5% of those infected with HTLV-I before 20 years of age develop ATL, the prognosis generally is poor (Siegel et al., 2001). In two HTLV-I endemic Caribbean populations, 56% and 78% of ATL cases were attributable to HTLV-I (Manns et al., 1993), much higher proportions than in the United States (Poiesz et al., 2001).

Risk of developing ATL is most strongly associated with HTLV-I infection early in life (Bartholomew et al., 1998; Cleghorn et al., 1995; Manns et al., 1993; Wilks et al., 1996) and has shown familial aggregation (Mueller and Blattner, 1997). Coinfection with strongyloides stercoralis, a gastrointestinal parasite endemic in the same geographical regions as HTLV-I, may increase the risk for ATL in those infected with HTLV-I (Mortreux et al., 2003). Another retrovirus, human immunodeficiency virus (HIV), also is associated with NHL (Beral et al., 1991; Levine, 1992; Ziegler et al., 1984). The US Centers for Disease Control (1982) included lymphoma of the brain as an AIDS-defining illness but subsequently expanded the case definition in 1985 and again in 1987 to include other and more specific NHL subtypes (Council of State and Territorial Epidemiologists, 1985, 1987). NHL accounts for 3%–5% of all initial AIDS diagnoses (Dal Maso and Franceschi, 2003). AIDS-related lymphomas typically are B-cell and disproportionately high grade and extranodal. Small noncleaved-cell Burkitt and Burkitt-like, immunoblastic large-cell and DLBCL subtypes comprise the majority of AIDS-related NHL. Lymphomas specific to HIV infection include primary effusion lymphoma (PEL) and plasmablastic lymphoma of the oral cavity that are rare even among those who are HIV positive (Raphael et al., 2001). Systemic NHL accounted for approximately 80% of AIDS-related lymphomas (Knowles, 1996). Prior to the widespread use of HAART that began in 1996, primary CNS lymphomas were common and accounted for a preponderance of extranodal NHL (Besson et al., 2001; International Collaboration on HIV and Cancer, 2000; Jones et al., 1999; Kirk et al., 2001). PreHAART median survival time for patients with AIDS-related NHL was less than 1 year, much shorter than for non-AIDS NHL patients (Cote et al., 1997). Although first noticed among homosexual men with HIV/AIDS (Ziegler et al., 1982; Ziegler et al., 1984), NHL occurred at similar rates in all HIV/AIDS risk groups including homosexual/bisexual men, intravenous drug users, heterosexual contact, and others (Biggar, 2000). Non-Hodgkin lymphoma risk is extraordinarily high in HIVinfected people, with relative risk estimates commonly in the range of 50–100 (Beral et al., 1991; Biggar et al., 2001; Cote et al., 1997; Goedert et al., 1998; Holly et al., 1997; Holly et al., 2002). For individual NHL histologic subtypes, relative risks vary from 1000 for Burkitt and Burkitt-like lymphoma (Beral et al., 1991) to 145 for DLBCL (Cote et al., 1997). Even for T-cell lymphomas and Working Formulation low-grade lymphomas, relative risks of 15 have been reported (Biggar et al., 2001; Cote et al., 1997). Demographic factors for AIDS-related NHL resemble those for non-AIDS NHL, with higher incidence among men, among whites, and at older ages (Holly and Lele, 1997; Holly et al., 1999). Because HIV infection is an overwhelming risk factor for NHL and certain lifestyles associated with risk for HIV infection may be confounding factors, there are few published studies that assess other environmental risk factors for AIDSrelated NHL. One population-based epidemiologic investigation

Men

9.87 - 12.18 9.40 - 9.86 8.96 - 9.39 8.36 - 8.95 5.27 - 8.35

Women

6.79 - 8.83 6.41 - 6.78 6.13 - 6.40 5.68 - 6.12 3.92 - 5.67

Figure 46–6. Non-Hodgkin lymphoma mortality rates by state economic area among white men and women during 1970–2000. (Source: Cancer Mortality Maps & Graphs Website, National Cancer Institute.)

Men

Women U.S. SEER, Whites

16.8 9.1 15.4 13.2 10.1 7.3

3.8 3.2 8

6

4

2

0

1973–77

3 .6 1 .9

Bombay, India

4.5

1993–97*

6 .0 2 .8

Osaka, Japan

6.3

10

6 .9 4 .3

Cali, Colombia

6.0

12

9 .6 5 .1

Sweden

6.3

14

3 .9

Va rese, Italy

7.1

16

7 .3

U.S. SEER, Blacks

6.0

18

1 0 .6 6 .9

3 .3 1 .7 0

2

4

6

8

10

12

Rates per 100,000 person-years

Figure 46–7. Non-Hodgkin lymphoma incidence rates (age-adjusted to world standard) by sex, 1993–1997 and 1973–1977. (Source: Waterhouse et al., 1982; Parkin et al., 2003.)

904

14

16

18

Non-Hodgkin Lymphoma reported similar risk factors for AIDS-related and non-AIDS–related lymphoma among HIV/AIDS patients (Holly and Lele, 1997; Holly et al., 1999). How HIV increases the risk for NHL is not completely understood. The virus infects T-cells, but is not integrated into NHL tumors (Carbone et al., 2001; Herndier et al., 1994; Shiramizu et al., 1994). Instead, HIV increases risk through its ability to disrupt immune surveillance (Gaidano et al., 1992; Grulich et al., 2000; Knowles, 1996). Extent of immune suppression measured by CD4 cell depletion and length of time with HIV infection, and B-cell stimulation measured by serum globulin and HIV p24 antigenemia have been reported to be independent risk factors for NHL among those with AIDS (Grulich et al., 2000). Clinical and laboratory data indicate that the cytokine environment resulting from HIV infection is likely to be an important component of AIDS-related B-cell lymphomagenesis. Genetic polymorphisms may influence HIV-related lymphomagenesis through their effects on B-cell proliferation, activation, and response, through control of HIV-replication or through associated cytokine release and production. AIDS-related NHL is a heterogeneous group of tumors that arise from different B-cell subtypes, germinal and post-germinal center B-cells. Burkitt and Burkitt-like lymphoma tend to occur earlier in the course of HIV-infection when CD4+ cell counts are higher. These tumors are characterized by Epstein-Barr virus (EBV) infection (30%–50%), p53 inactivation (50%–60%), and c-MYC activation (100%). The incidence of AIDS-related large-cell and immunoblastic lymphomas increases with immunosuppression and mostly occurs at later stages of HIV infection when CD4+ cell counts are low. HIV infection is the primary unifying characteristic that defines these diverse tumors. EBV infection without latent membrane protein-1 expression (30%–40%) and BCL-6 proto-oncogene expression (20%) are characteristic of the diffuse large-cell lymphomas, whereas EBV infection (90%; 100% if CNS) with LMP-1 expression (65%–75%; 90% if CNS) characterizes immunoblastic large cell lymphomas (Carbone, 2003; Cesarman et al., 1999; Gaidano et al., 1996, 1998; Knowles, 1996). Risk of NHL, especially CNS lymphoma, also is elevated after diagnosis with Kaposi sarcoma, a cancer strongly associated with human herpesvirus 8 (HHV-8) infection (Ahsan and Neugut, 1996; Ridolfo et al., 1996). The rare PEL tumors are 100% associated with HHV-8 infection and also are highly likely to be EBV infected (90%–100%) (Boulanger et al., 2001; Cesarman et al., 1995; Nador et al., 1996). Widespread use and availability of HAART has improved length of survival among those infected with HIV. Reduced incidence of CNS lymphoma and other NHL subtypes associated with HIV has been demonstrated in SEER data (Clarke and Glaser, 2001; Eltom et al., 2002). However, some HIV/AIDS cohort studies in the United States and France have shown improved survival and remission for patients with AIDS-related lymphoma but no differences in pre- and postHAART incidence of NHL (Gerard et al., 2002; Tam et al., 2002). Australian AIDS data indicated an increase in NHL as an AIDS-defining illness and no change in pre- and post-HAART median survival with NHL (Dore et al., 2002). Methodologic differences between investigations are likely to have contributed to the inconsistency in the results reported by individual studies. A meta-analysis that includes data from 23 prospective studies showed a decline in the incidence of AIDSrelated NHL with substantial decreases in CNS and immunoblastic lymphomas but not Burkitt lymphoma (International Collaboration on HIV and Cancer, 2000). In sum, HAART has been most effective in decreasing the development of cancers associated with greater immunosuppression. Continued surveillance of HIV-infected populations is warranted to examine the long-term incidence of NHL and other cancers.

Herpesviruses Among the herpesviruses, both EBV and HHV-8 infection are highly specific for Burkitt lymphoma and PEL, respectively. EBV quite probably plays a role in other NHL, but is not yet fully understood despite intensive investigation. First identified as the cause of Kaposi sarcoma and termed KSHV, HHV-8 is endemic in the Mediterranean basin with

905

a prevalence that corresponds to the prevalence of Kaposi sarcoma before the AIDS epidemic. It also is endemic in sub-Saharan Africa but is uncommon in other regions of the world (Ascoli et al., 2002; Schulz, 2000). Although HHV-8 seroprevalence is low in the general US population, its prevalence among male homosexuals ranges from 20%–40% (Schulz, 2000). HHV-8 was linked to a unique AIDSrelated NHL in 1995 (Cesarman et al., 1995) that subsequently was designated as PEL (Nador et al., 1996). HHV-8 has been reported in all PEL tumors (Cesarman et al., 1995; Karcher and Alkan, 1997) with 90% also infected with EBV (International Agency for Research on Cancer, 1997). HHV-8 also is associated with multicentric Castleman disease-plasmablastic lymphoma (Oksenhendler et al., 2002). These HHV-8–related NHL subtypes are associated almost exclusively with HIV infection and appear to develop in settings of profound immunosuppression, particularly PEL (Knowles and Cesarman, 1997). Although the number of patients with NHL related to HHV-8 infection accounts for only a small fraction of all NHL, a clear understanding of this remarkable association may help to elucidate the mechanisms of NHL pathogenesis (Cohen et al., 2001). Epstein-Barr virus presents a more complex picture. The International Agency for Research on cancer has determined that there is sufficient evidence to classify EBV as a human carcinogen (International Agency for Research on Cancer, 1997). First identified in children with endemic Burkitt lymphoma in equatorial Africa (Burkitt, 1958), EBV infection is ubiquitous, infecting perhaps more than 90% of the world’s population. Infection usually occurs early in life in developing countries where endemic Burkitt lymphoma is common (Niedobitek et al., 2001), and in later childhood in developed countries with better hygiene (Hsu and Glaser, 2000). In areas where Burkitt lymphomas are endemic, nearly all are associated with EBV infection. In Western countries, only 20%–34% of lymphomas harbor EBV (Hummel et al., 1995; zur Hausen et al., 1970). Epstein Barr virus DNA is found in lymphomas other than Burkitt but possibly as a passenger virus in many of those lymphomas. The strength of the association between EBV and NHL differs by NHL subtype and correlates closely with degree of immunosuppression. EBV occurs in 33%–67% of AIDS-related lymphomas, depending on the detection method (Cohen and Scadden, 2001), but in few highgrade HIV-negative lymphomas. Similarly, EBV appears in virtually all post-transplant lymphomas and CNS lymphomas in immunocompromised patients. EBV infection also occurs in nearly all NHL that develops in patients with Wiskott-Aldrich syndrome (WAS) and Xlinked lymphoproliferative disease (XLP), two hereditary disorders marked by severe immunosuppression (Levine, 1994; Mueller, 1999). WAS patients have an increased frequency of EBV-driven lymphomatoid granulomatosis that can develop into EBV-positive DLBCL (Borisch et al., 2001). Children with XLP have an increased susceptibility to EBV infection that leads to the development of NHL in approximately 30% of those who survive their primary EBV infection (Levine, 1994; Mueller, 1999; Nagy et al., 2002). The mechanism for EBV-related NHL pathogenesis is complex and not completely understood, although epidemiologic and molecular studies have elucidated some of EBV’s role and serum bank studies have helped to determine the sequence of events. A cohort study of 240,000 persons in Norway and the United States reported twofold to threefold increased risk for NHL associated with increasing EBV antibody titers (Mueller et al., 1991). Similarly, a Finnish cohort followed for 12 years revealed that NHL risk increased threefold in people with elevated EBV antibodies (Lehtinen et al., 1993; Scherr et al., 1996). The modest serology effects, the variation in viral detection in tumors, and the studies of immunocompromised cohorts implicate EBV as a cofactor but not the major agent in many forms of NHL. EBV can produce lymphoma in an immunodeficient host, but immunosuppression appears to be required (International Agency for Research on Cancer, 1997). Other members of the herpesviruses family, including Human T-cell Lymphotrophic Virus type II, Human Herpes Virus-6, and cytomegalovirus also have been investigated in association with NHL. However, none of these herpesviruses appears to predispose to NHL (Knowles, 1999).

906

PART IV: CANCER BY TISSUE OF ORIGIN

Other Viruses Several studies have shown an association between hepatitis C virus (HCV) and NHL, but the issue is not fully resolved (Engels et al., 2003a). HCV is lymphotrophic and associated with mixed cryoglobulinemia, a B-lymphocyte proliferative disorder (Cacoub et al., 1994). A 1994 Italian study on NHL and HCV found seroprevalence of 34% in 50 patients (Ferri et al., 1994). The majority of subsequent Italian studies have reported high HCV seroprevalence in NHL patients, but most US studies have not, with HCV prevalence rates ranging from 1.4% to 22% in US populations (King et al., 1998; Zuckerman et al., 1997). Some studies supported an association with low-grade lymphoplasmacytoid/immunocytic lymphomas (Germanidis et al., 1999; Mazzaro et al., 1996; Silvestri et al., 1996; Vallisa et al., 1999) whereas others reported associations with a wider range of subtypes (Cucuianu et al., 1999; Mele et al., 2003; Pioltelli et al., 2000; Zucca et al., 2000). In general, the evidence for an association with specific HCV genotypes has been inconsistent (De Rosa et al., 1997; King et al., 1998; Luppi et al., 1998; Silvestri et al., 1997; Zignego et al., 1997). Several studies from Northern Europe, France, the Netherlands, Germany, and Canada have reported no HCV association with NHL. Geographic variability in HCV seroprevalence may account for some of the inconsistency in reported results. Although the epidemiologic evidence is conflicting, several clinical and laboratory studies have provided support for an association with specific NHL subtypes. In a small series of patients with splenic lymphoma with villous lymphocytes, complete tumor remission was attained in HCV-infected patients whose infection was successfully managed with antiviral medications, whereas tumors did not regress in HCV-negative patients who received the same antiviral regimen (Hermine et al., 2002). This effect is similar to that observed for gastric MALT lymphomas that regress after successful treatment of Helicobacter pylori (H. pylori) infection. In laboratory studies, E2 antigens bind to some HCV lymphoma immunoglobins but not to those from non-HCV lymphomas. It appears that some HCV-lymphomas originate from B-cells activated by HCV-E2 antigen (Quinn et al., 2001). Others have shown that t(14; 18) translocations are more frequent in patients with HCV infection, especially those with mixed cryoglobulinemia, and that some translocations disappear after antiviral therapy, implicating HCV-related anti-apoptotic effects on B-cells (Zignego et al., 2002; Zuckerman et al., 2001). These clinical and laboratory findings may indicate that HCV promotes lymphomagenesis via an antigen-driven process (Weng and Levy, 2003). Further research is needed to determine the relevant antigens’ to identify the specific NHL subtypes, and to evaluate cofactors such as infection with hepatitis B virus (Hausfater et al., 2000). Simian virus 40 (SV40) has been proposed as a factor in NHL incidence. In 2002, two laboratories reported the detection of SV40 DNA sequences in a large proportion of lymphomas (Shivapurkar et al., 2002; Vilchez et al., 2002). Experimental studies in laboratory animals previously had suggested the link (Diamandopoulos, 1972). SV40 associations have been reported but not confirmed for other cancers shown to be induced by SV40 infection in laboratory animals, including mesothelioma, osteosarcoma and other bone tumors, brain tumors, leukemia, and testicular cancer (Carbone et al., 2003; CarrollPankhurst et al., 2001; Engels et al., 2003b; Fisher et al., 1999; Shah et al., 2000; Strickler et al., 1998). SV40 is a polyomavirus similar to the human polyomaviruses BK virus and JC virus, whose natural host is the rhesus macaque. Discovered in 1960 during safety testing of the poliovirus vaccine (Sweet and Hilleman, 1960), SV40 shortly thereafter was found to be oncogenic in laboratory animals. Large T antigen has been identified as the major SV40 oncoprotein (Sullivan and Pipas, 2002), but mechanisms relevant to tumorigenesis in humans are unknown (Malkin, 2002). Human exposure has occurred mainly via immunization with contaminated poliovirus vaccine that was used throughout the United States, Canada, parts of Europe, Japan, Mexico, and Central and South America between 1955 and 1963 (Vilchez et al., 2003). Tens of millions of people worldwide may have been exposed to live SV40 via the contaminated vaccines, but the extent of infection is unknown because not all batches were contaminated, the concentration of live

virus varied by batch, there was unequal geographic distribution of batches of polio vaccine that were known to have been contaminated, and vaccine manufacturing and distribution were not adequately tracked (Gazdar et al., 2002). The high proportion of SV40 in NHL tumors reported by some studies (David et al., 2001; Shivapurkar et al., 2002; Vilchez et al., 2002) has not been reported consistently (Daibata et al., 2003; MacKenzie et al., 2003; Rizzo et al., 1999). Studies that used sera to evaluate SV40 exposure measured SV40 in DNA extracted from blood (David et al., 2001; MacKenzie et al., 2003), measured SV40 neutralizing antibodies, and measured antibodies to SV40 using a virus-like-particle-based enzyme immunoassay (de Sanjose et al., 2003; Engels et al., 2004). Despite the different methods for SV40 detection in sera, seroprevalence of SV40 consistently was low (David et al., 2001; de Sanjose et al., 2003; MacKenzie et al., 2003; Minor et al., 2003), and in two case-control studies unrelated to NHL risk (de Sanjose et al., 2003; Engels et al., 2004). Several epidemiologic studies that have evaluated the association between NHL and vaccination with contaminated polio vaccine also have found no association (Carroll-Pankhurst et al., 2001; Engels et al., 2003b; Engels et al., 2003c; Holly et al., 1999; Holly and Bracci, 2003). These studies were based on the estimated exposure rate of individuals who were highly likely to have been immunized with contaminated vaccine between 1955 and 1963. Interestingly, SV40 also has been detected in tumor samples obtained from patients who could not have been exposed to the contaminated poliovirus vaccine. A recent overview of the scientific evidence by the Institute of Medicine of the U.S. National Academies concluded that there is strong evidence that SV40 is a transforming virus but the evidence is moderate that under natural conditions, SV40 can lead to cancers in humans including lymphomas (Immunization Safety Review, 2002). Future studies that accurately identify those exposed to SV40, determine the molecular mechanism through which SV40 may promote tumorigenesis in humans, and determine transmissibility of SV40 in humans are required to determine whether SV40 plays a role in the etiology of NHL.

Other Infections Increased risk of gastric B-cell lymphoma, especially low-grade MALT lymphoma, has been associated with H. pylori infection in several epidemiologic studies (Cuttner et al., 2001; Doglioni et al., 1992; Nakamura et al., 1997; Parsonnet et al., 1994; Ullrich et al., 2002; Wotherspoon et al., 1991; Zucca et al., 2000). H. pylori infection is more likely to occur in lower SES groups, younger age groups, certain geographic regions, and ethnic groups (Brown, 2000). Infection rates are high in developing countries and decreasing in developed countries. The most likely mode of transmission is person-to-person transmission via fecal-oral and oral-oral routes. Infection can cause chronic gastritis although most people with H. pylori infection are asymptomatic. There is some evidence that more virulent strains of H. pylori are associated with greater risk for symptomatic disease (Parsonnet, 1998). In the United States, it has been suggested that the prevalence of infection coincides with age (e.g., approximately 40% of 40 year olds and 50% of 50 year olds are infected) (Nomura et al., 2002). A prospective study of populations in the United States and Norway reported that risk for gastric NHL was increased sixfold among individuals with detectible H. pylori IgG antibodies (Parsonnet et al., 1994), with most of these patients diagnosed with DLBCL. Investigators who examined a series of Japanese gastric lymphoma patients (Nakamura et al., 1997) noted that MALT lymphoma was the most strongly linked. This observation fits with the epidemiologic studies of peptic ulcer and gastric lymphoma that demonstrate a stronger association for MALT lymphoma (Vineis et al., 1999). A large proportion of gastric lymphomas are classified as high-grade DLBCL, but it is unclear what proportion is the result of progression of low-grade MALT lymphoma vs. de novo DLBCL. Further, many studies have shown sustained regression of low-grade gastric B-cell MALT lymphoma after treatment to eradicate H. pylori infection (Neubauer et al., 1997; Ruskone-Fourmestraux et al., 2001; Wotherspoon et al., 1993),

Non-Hodgkin Lymphoma but regression among high-grade lymphomas has been less consistent (Wundisch et al., 2003). Successful treatment of gastric B-cell lymphoma through eradication of H. pylori infection appears to depend upon depth of infiltration and also may be associated with specific genetic translocations (Alpen et al., 2000; Liu et al., 2001). Meanwhile, the etiology of H. pylori-negative gastric MALT lymphomas remains unclear (Wotherspoon et al., 2002). Other infectious causes of NHL in children have been suggested by hypotheses related to population mixing (Dickinson and Parker, 2002; Kinlen et al., 1995; Law et al., 2003). The hypothesis proposes that children who live in isolated areas and have little exposure to infectious agents are at increased risk for leukemia and lymphoma through population mixing that results in unexposed children being exposed to infected persons (Kinlen et al., 1995). Delayed exposure to common infectious agents among children who live in geographically isolated regions also has been proposed as a factor in childhood leukemia and NHL development (Stagnaro et al., 2001). Recent results from a large UK cohort study of childhood cancer and population mixing showed that low population diversity but not volume of migration was associated with risk for childhood leukemia and lymphoma (Law et al., 2003). The authors also reviewed 17 other published studies of childhood cancer and population mixing and concluded that, consistent with their results, large population migration was not necessary for the observed increased risks but that patterns of childhood exposure to infectious diseases were likely to play a role in the etiology of these childhood hematopoietic cancers (Law et al., 2003). Further detailed investigation is warranted to differentiate patterns of childhood exposure to infectious disease from exposures that are the result of population mixing.

Radiation Exposure to high doses of ionizing radiation can lead to the development of NHL, but not as rapidly or as markedly as it does to leukemia and many solid tumors. Among atomic bomb survivors, excess deaths from leukemia but not lymphoma appeared shortly after exposure (Shimizu et al., 1990). A later report found a slight excess, with the estimated relative risk for lymphoma as 1.3 at the level of 1 SV, compared with 4.9 for leukemia (Preston et al., 1994). Studies of occupational exposure to ionizing radiation generally have reported no excess NHL risk, possibly because the exposure levels were low or measured imprecisely. Nuclear workers mostly have shown no excess NHL risk (Cardis et al., 1995; Gilbert et al., 1993; Kendall et al., 1992), but nonsignificant excesses in NHL mortality were observed in UK atomic energy employees (Beral et al., 1985; Carpenter et al., 1994). Similarly, several studies of technologists and radiologists have reported no excess NHL risk (Mohan et al., 2003; Wang et al., 2002), but there is some uncertainty about the effects of exposure before 1950 when doses were higher (Berrington et al., 2001; Jablon and Miller, 1978; Mohan et al., 2003; Wang et al., 2002). A few studies have seen an increase in risk, notably from those time periods when radiation exposure to workers was higher. A cohort of uranium millers showed excess mortality from lymphoma (Archer et al., 1973). However, a recent large case-control study that imputed radiation exposure based on lifetime job history showed no association with NHL risk (Eheman et al., 2000). A study of childhood leukemia and lymphoma (combined) implicated paternal exposure to ionizing radiation at a nuclear installation in the United Kingdom (Gardner et al., 1990), but several subsequent studies (Kinlen et al., 1993; McLaughlin et al., 1993; Urquhart et al., 1991) showed no such association in other nuclear worker populations. Studies of therapeutic irradiation, typically at much higher doses than workers experience, have demonstrated increased NHL risk. Women irradiated for infertility developed more lymphomas than expected (Ron et al., 1994). A total of 37 patients treated for ankylosing spondylitis developed lymphoma, compared with 21 lymphomas expected in this population (Weiss et al., 1994). No excess NHL risk was found in patients irradiated for scalp ringworm (Ron et al., 1988) or benign locomotor conditions (Damber et al., 1995) nor in twins frequently X-rayed in utero (Inskip et al., 1993). Non-ionizing radiation from electromagnetic fields surrounding power lines was suggested to increase risk of childhood lymphoma in

907

some early studies (Savitz et al., 1988). Among Canadian electric utility workers, lymphoma risk was associated with exposure to electromagnetic fields, but with no dose-response gradient (Schroeder and Savitz, 1997). A weak positive association also was reported from a cohort of US electric utility workers (Schroeder and Savitz, 1997). A critical review of the literature in 2001 concluded that the evidence suggests no effect of EMF on childhood lymphoma and insufficient data on adult lymphoma risk (Ahlbom et al., 2001). Ultraviolet radiation, principally from sunlight, has been suggested as a risk factor for NHL partly based on co-occurrence of skin cancers and lymphomas. In SEER registry data, melanoma survivors showed a relative risk of 1.42 for NHL, and survivors of NHL showed a relative risk of 1.75 for skin melanoma (Goggins et al., 2001). This pattern of second primary co-occurrence of skin cancer and NHL also has been reported in population cancer registries and case series in other geographic locations, with higher risks for melanoma after NHL (Adami et al., 1995; Hemminki et al., 2003; Levi et al., 1996). The pattern of co-occurrence is now established, but the relative contributions of therapy, genetic susceptibility, and environmental exposures remain to be discerned. The concomitant increases in NHL and melanoma rates in the past several decades raise the possibility that sunlight exposures may contribute to both trends. Geographic correlation studies of NHL and ultraviolet radiation show inconclusive results. In Swedish men (Adami et al., 1999), residence at latitude 55°–56° N, compared to latitude 63°–69° N, was associated with a relative risk of 1.2 for NHL, 2.4 for multiple myeloma, and 2.1 for squamous cell carcinoma. In England and Wales, NHL risk in counties in the highest quartile of estimated mean annual radiation was about 30% greater than in counties in the lowest quartile of annual radiation (Bentham, 1996). Using data from registries around the world to increase greatly the variation in latitude, McMichael and Giles (1996) estimated a correlation coefficient of 0.50 between UV dose and NHL incidence, with a confirmatory strong correlation for melanoma. By contrast, NHL rates in the United States are weakly but inversly correlated with UV exposure based on geographic location (Freedman et al., 1997; Hartge et al., 1996), whereas skin cancers show the expected strong positive correlation. Occupational sunlight exposure has been explored in several populations. A large study of US death certificates showed a slight reduced risk (Freedman et al., 1997), a cohort study of Swedish construction workers showed a slight increased risk (Hakansson et al., 2001), and a US cohort study of electric utility workers showed no association (van Wijngaarden and Savitz, 2001). Recent case-control data from Australia (Hughes et al., 2004) and from Sweden and Denmark (Smedby et al., 2005) show a modest inverse association between level of UV exposure and NHL risk. This apparent protective association agrees with the gradient in risk in the United States (Hartge et al., 1996) and could indicate a beneficial effect of vitamin D (Egan et al., 2005). More epidemiologic data on sunlight exposures and on promising biological markers of those exposures are needed before a firm conclusion can be drawn on the role of ultraviolet radiation in NHL development.

Occupational Exposures Certain occupations and industries have been associated with increased lymphoma risk in multiple studies, notably farming, forestry, paper and pulp production, rubber manufacture, woodworking, dry cleaning, and metal work (Scherr and Mueller, 1996). A large cohort study in the United States used death certificates to identify excess NHL mortality among farm managers, fire fighters, aircraft mechanics, electronic repairers, mining machine operators, and crane and tower operators (Figgs et al., 1995). In a Canadian case-control study with occupational exposures queried and assessed by a team of chemists and industrial hygienists, NHL risk was elevated in farmers, horticulturists, artists, chefs, leather workers, textile workers, rubber workers, and janitors (Fritschi et al., 1996). Jobs that involve exposures to animals warrant attention because of the plausibility of transmission of zoonotic infections. Abattoir workers have shown increased risk for NHL in some studies (Johnson et al., 1995, 1997; Metayer et al., 1998), as have workers involved in

908

PART IV: CANCER BY TISSUE OF ORIGIN

farm animal breeding and cattle raising (Fritschi et al., 2002; McDuffie et al., 2002; Nanni et al., 1996). If the animal exposure hypothesis is confirmed, investigators are likely to examine tumor specimens to search for novel viruses or other infectious agents. Barbers and hairdressers have increased lymphoma risk in some studies (Boffetta et al., 1994; Lamba et al., 2001), as have chemists and lab technicians (Burnett et al., 1999; Hoar et al., 1981). The responsible agents are unknown, but the occupational association among barbers and hairdressers has been cited to support the hypothesis that exposure to hair dyes increases NHL risk. Asbestos exposure was linked to lymphomas in an early report, but subsequent studies have found no association (Becker et al., 2001). Several chemical exposures have been pursued in depth, including pesticides, dioxins, polychlorinated biphenyls (PCBs), benzene, and other organic solvents.

Pesticides and Organochlorines Farmers have shown an increased risk of NHL in many studies, but not all, and the specific practices associated with risk vary. In a rural farming community in Michigan, NHL rates were elevated and correlated with county-level indices of agricultural activity (Waterhouse et al., 1996). Although no association with agricultural work was observed in a large Italian hospital-based case-control study (Settimi et al., 2001), a small study in an agricultural region of Italy reported an association with animal breeding, use of insecticides, and childhood farm exposures (Nanni et al., 1996). A cohort study of Norwegian farmers associated elevated rates of acute leukemia and multiple myeloma, but not NHL, with several types of farming and farm-related exposures (Kristensen et al., 1996). It seems likely that the variation in findings among studies of agriculture and NHL stems from small study size, from real differences in exposures by crop and by country, and from challenges in assessing exposure. A thorough review of the literature on pesticides and cancer as of 1997 noted that associations have been reported between risk of NHL and phenoxyacetic acid herbicides, including 2, 4-D; arsenical insecticides; organochlorine insecticides; and organophosphate (Zahm et al., 1997a). Since the 1980s, exposures to herbicides and other pesticides have been linked to increased lymphoma risk in case-control and cohort studies conducted in the United States and elsewhere (Hardell et al., 1981; Hoar et al., 1986; Wingle et al., 1990; Zahm et al., 1990; Zahm, 1997). Chlorophenols in particular (Kogevinas et al., 1995) have been linked to moderately increased lymphoma risk in a few studies (Hardell et al., 1994; Hardell and Eriksson, 1999; Kogevinas et al., 1995) and to slightly increased risk in others (Garabedian et al., 1999). Chlorophenols are used as fungicides, wood preservatives, and in the production of phenoxyacetic acid herbicides. To increase the statistical power to test hypotheses about specific agents, investigators have conducted large multicenter studies and pooled case-control studies. These analyses have generally but not uniformly (Burns et al., 2001) reported results of an association with phenoxyacetic acid herbicides and also have implicated carbamate insecticides, carbamate herbicides, organophosphates, and chlorophenols, with relative risk estimates in the range of 1.2 to 1.8 (De Roos et al., 2003; Hardell et al., 2002; McDuffie et al., 2001; Waddell et al., 2001; Zheng et al., 2001). Serum and adipose tissue samples are especially valuable indicators of prior exposure to fat-soluble pesticides and other organic compounds, particularly if they were collected before NHL was diagnosed. In one case-control study, NHL patients had higher tissue levels of chlordane and PCBs than did controls (Hardell et al., 1996a,b). In other case-control serologic studies associations were reported for dioxin and PCBs (Hardell et al., 2001a,b), or for furans and PCBs (de Roos et al., 2005). In pre-diagnostic sera from a prospective cohort study, PCB but not DDT levels were related to risk of NHL (Rothman et al., 1997), and an interaction with EBV was suggested. By contrast, NHL cases had levels of chlordane, lindane, dieldrin, and other organochlorines that were indistinguishable from the rest of the cohort (Cantor et al., 2003). More research is needed to confirm which specific organochlorines affect NHL risk.

Industrial Solvents Organic industrial solvents that are used in a variety of occupations have been associated with lymphoma risk in numerous studies. Trichloroethylene and tetrachloroethylene, also known as “perchlorethylene,” are degreasing agents that have been used in dry cleaning and other industries and associated with lymphomas. In a comprehensive review, the International Agency for Research on Cancer (1995) reported “limited evidence for the carcinogenicity of trichloroethylene in humans” in an evaluation that included studies on lymphoma in humans. In the same evaluation the authors concluded that tetrachloroethylene “is probably carcinogenic to humans” and again noted the studies on lymphoma. One study of aircraft maintenance workers employed between 1952 and 1956 showed an increased risk of developing lymphomas in those who were exposed to solvents (Blair et al., 1998), whereas a cohort mortality study of workers employed at an aircraft manufacturing facility after 1959 showed little or no increased risk of NHL (Boice et al., 1999). A study of Danish workers exposed to trichloroethylene revealed a greater relative risk of NHL (Hansen et al., 2001), whereas a cohort mortality study of US aerospace workers found no elevated risk among those exposed to trichloroethylene (Morgan et al., 1998). Results from a pooled casecontrol analysis showed increased odds ratios for lymphoma among workers who handled aviation gasoline, white spirits, paint thinner, and other solvents (Persson and Fredrikson, 1999). A Canadian casecontrol study reported increased risk to workers exposed to benzidine (Mao et al., 2000). Studies of laboratory workers and chemists who are exposed to many solvents, have revealed a moderate excess of lymphoma incidence and mortality (Burnett et al., 1999; Hoar and Pell, 1981; Kauppinen et al., 2003). Benzene exposure, a clear risk factor for leukemia, has not been shown to play a large role in lymphoma risk (Savitz and Andrews, 1997). A large meta-analysis of 308,000 workers from 26 petroleum industry cohorts found no excess mortality from NHL (Wong and Raabe, 2000), but the study lacked specific information on exposure levels. Low doses in most settings, inadequate measures of exposure, and confounding by other industrial exposures may obscure lymphoma risk (O’Connor et al., 1999). In a cohort of Chinese workers, hematotoxicity was observed in association with low levels of benzene exposure (Lan et al., 2004).

Environmental Contaminants Outside of the workplace, levels of organochlorines, pesticides, solvents, and other compounds in the general environment may be related to lymphomas, but the effects of these environmental exposures have been little studied. Diet can be a source of exposure to PCBs and other organic compounds, though usually at low concentration, and one observation of elevated NHL risk in fisherman suggested dietary PCBs as the cause (Svensson et al., 1995). Dioxins released from incinerators result in ground level contamination, and one case-control study reported an odds ratio of 2.3 for residents of the areas of high contamination (Floret et al., 2003). Using a similar approach to infer exposures to benzene and nitrogen oxide from auto traffic, investigators studying childhood NHL estimated that a doubling of benzene level corresponded to a relative risk of 1.25 (Raaschou-Nielsen et al., 2001). These observations all need confirmation in other studies—studies that will become more common as geographic information systems expand. Drinking water can be a source of exposure to many organic compounds. An ecological study found slightly elevated rates of lymphomas in New Jersey municipalities with high levels of trichloroethylene and tetrachloroethylene (Cohn et al., 1994). Measured nitrate levels in drinking water were associated with moderately increased risk in a case-control study in Nebraska (Ward et al., 1996) and an ecological study in Slovakia (Gulis et al., 2002). On the other hand, a cohort study of Iowa women (Weyer et al., 2001) and a casecontrol study in Minnesota (Freedman et al., 2000) did not find an association. An ecologic study in Italy found an effect of nitrate levels in men but not women with a wide range of exposure levels (Cocco et al., 2003). Further studies should help to resolve this issue.

Non-Hodgkin Lymphoma

Personal Exposures Several studies have examined whether people who use hair color products have an increased risk of lymphoma (Correa et al., 2000). Several case-control studies suggested a doubling of risk in some groups of users, especially those who used darker colors or permanent dye (Cantor et al., 1988; Zahm et al., 1992), whereas others have reported slight (Miligi et al., 1999) or no (Holly et al., 1998) elevation in risk. Two sizeable cohort studies showed no effect overall among long-term users of permanent dyes (Altekruse et al., 1999; Grodstein et al., 1994). The cohort studies had the advantage of collecting data before the lymphomas occurred but the disadvantage of less detail on the type and timing of exposure. The issue is not fully resolved, but the body of current evidence suggests no increased lymphoma risk, certainly none to the large majority of users. Tobacco has been studied in relation to NHL in multiple investigations. In most case-control (Holly and Lele, 1997; Holly et al., 1999; Nelson et al., 1997; Tavani et al., 1994a) and cohort studies (Adami et al., 1998; Herrinton and Friedman, 1998; McLaughlin et al., 1995), smokers had about the same risk of NHL as did non-smokers, yet several positive associations have been reported from subgroups within large and rigorous cohort and case-control studies. Associations with follicular but not other lymphomas have been noted in several reports (Herrinton and Friedmen, 1998; Parker et al., 2000; Stagnaro et al., 2001), and one study implicated smoking as a risk factor for diffuse or small lymphocytic lymphoma (Zahm et al., 1997b). A pooled analysis of nine case-control studies showed a modest association with risk of NHL, primarily for follicular subtypes (Morton et al., 2005b). In one study, lymphomas with t(14 : 18) translocations were associated with smoking (Schroeder et al., 2002). It remains to be determined whether there are susceptible subgroups or additional histologies associated with smoking.

Dietary Patterns Diet, exercise, and energy balance might influence the risk of developing lymphoma through a variety of pathways (Skibola et al., 2005b) and dietary patterns. These patterns have changed markedly during the years of the long-term increase in NHL. The current data do not support firm conclusions, but several studies have reported excess lymphoma risk in association with diets high in fat and protein or low in vegetables and fruits, typically with relative risks between 1.5 and 2.0. In a cohort of 88,000 nurses, those with high consumption of fat, especially of trans-unsaturated fat, had about twice the risk of developing lymphoma as did those with low consumption (Zhang et al., 1999), and those with high intake of vegetables, and fruits to a lesser extent, had 40%–50% reduced risk of lymphoma (Zhang et al., 2000), with each effect measured holding the other one constant. In a cohort of 35,000 Iowa women (Chiu et al., 1996), NHL risk was higher in those who consumed more animal fat and lower in those consuming more fruits. Case-control studies in Italy (Tavani et al., 2001) and Nebraska (Ward et al., 1994) found somewhat similar patterns. Current drinkers have shown a slightly but consistently reduced risk of NHL in a pooled analysis (Morton et al., 2005c) of nine case-control studies with information on alcohol risk, some of them previously published (Morton et al., 2003; Tavani et al., 2001; Holly and Lele, 1997; Holly et al., 1999; Nelson et al., 1997; Chang et al., 2004; Willett et al., 2004; Stagnaro et al., 2001). Other case-control studies have shown either an inverse association (Chiu et al., 2002) or no association (Brown et al., 1992). In addition, some cohort studies (Briggs et al., 2002; Chiu et al., 1999) but not all (Kato et al., 1992) support an inverse association of alcohol and NHL. Whether the association is causal and whether there are susceptible subgroups remains to be determined. Coffee and tea consumption appear not to be related to NHL risk (Tavani et al., 1994b; Zheng et al., 2001). Energy balance has been studied very little in relation to risk of lymphoma. In a large death-certificate study, jobs with high physical activity were unrelated to NHL risk (Zahm et al., 1999). In the cohort of Iowa women, risk of developing NHL did not vary with height, weight, obesity, or exercise (Cerhan et al., 2002a). In the American Cancer Society cohort of 95,000 men and women, NHL mortality was

909

high in overweight men (RR = 1.5) and in overweight women (RR = 2.0) (Calle et al., 2003). Among 28,000 patients discharged from hospitals with a diagnosis of obesity, NHL risk was elevated in obese women (RR = 1.6) but not in obese men (Wolk et al., 2001). Increased BMI also has been related to NHL in a case-control study that showed an increased risk for all NHL among obese women (OR = 2.6) and obese men (OR = 2.8) with a trend of increasing risk with increasing body mass index (Holly et al., 1999; Skibola et al., 2005). Increased odds ratios and increasing trends also were reported for NHL subtypes—DLBCL among obese men and follicular lymphoma among obese women (Skibola et al., 2004). This important question has not been fully explored at this writing.

Reproductive and Hormonal Factors Pregnancy produces transient changes in immunity that are complex and not completely understood (Luppi, 2003). Several studies have assessed whether number or timing of pregnancies or births affects the risk of developing NHL. Despite a few early reports of an association (Miller et al., 1980; Olsson et al., 1990) subsequent studies have shown pregnancy and reproductive factors not to be strongly associated with NHL (Adami et al., 1997; Cerhan et al., 2002a; Nelson et al., 2001; Tavani et al., 1997). However, one small study showed increased risk estimates for NHL among women younger than 50 years of age who had given birth within the past 10 years compared with nulliparous women (Tavani et al., 1997). Investigation of other reproductive health factors including menarche and menopause also have shown little or no association with NHL, although history of endmetriosis has been associated with nearly twofold risk estimates in two studies (Brinton et al., 1997; Olson et al., 2002a). Exogenous estrogens have been evaluated in case-control and cohort studies with inconsistent results that vary by type of exposure and NHL subtype (Cerhan et al., 2002b; Nelson et al., 2001; Schiff et al., 1998). Two studies reported that use of oral contraceptives was associated with decreased risk for NHL (Nelson et al., 2001; Schiff et al., 1998), including lower risk with longer duration and lowest risk for OC used before 1970 (Nelson et al., 2001). Decreased risk also was reported for use of lactation suppressants (Nelson et al., 2001). The positive association between NHL and hormone replacement therapy (HRT) has been less consistent (Cerhan et al., 2002b; Nelson et al., 2001; Schiff et al., 1998). A cohort study among older women revealed an increased risk associated with current use of HRT, especially for nodal and DLBCL, and an increased risk for follicular lymphomas with any history of HRT use (Cerhan et al., 2002b). Other researchers have reported no association between HRT and NHL risk (Nelson et al., 2001; Schiff et al., 1998). Polymorphisms in genes in estrogen-related pathways provide support of a role for prolaction and estrogen in NHL pathogenesis (Skibola et al., 2005a). Reproductive hormones affect immune function, and estrogen and estrogen-like compounds influence B-cell development (Medina et al., 2000); further research is needed to clarify the role of exogenous hormones on risk of NHL. Less information exists on childhood NHL, but one study has suggested no effect of maternal age or parity but an increased risk to children delivered by Cesarean section (Adami et al., 1996).

HOST FACTORS Personal and Family History of Cancer An important indicator of the risk for developing a primary NHL is a history of a previous primary NHL. The estimated relative risk is 2.4, and the excess risk may persist for 10 years or more after the first diagnosis (Dong and Hemminki, 2001b). Strikingly similar relative risk estimates are associated with NHL in the immediate family, based on data from registries that capture all cancers in a population over several generations (Dong and Hemminki, 2001a; Goldgar et al., 1994). Slightly higher estimates have come from case-control studies, probably because of differential recall and reporting in cases compared with controls (Chiu et al., 2002; Holly et al., 1999; Zhu et al., 2001). Although familial aggregation of NHL is well established, the mode

910

PART IV: CANCER BY TISSUE OF ORIGIN

of inheritance is not. An observation in one study that siblings had higher risk than parents or children suggests X-linked or recessive inheritance (Dong et al., 2001a). Familial aggregation may be stronger for some histologic types (Zhu et al., 1998), but larger studies are required for clarification. Consanguinity was unrelated to childhood NHL in a small study in the United Arab Emirates (Bener et al., 2001). Hodgkin lymphoma, lymphoid leukemia, myeloid leukemia, and multiple myeloma cluster with NHL in many studies of second primary cancers and of cancers in close relatives. In Swedish registry data, for example, the relative risk for hematopoetic and lymphatic malignancies combined following a first diagnosis of NHL was 2.4 (Dong and Hemminki, 2001b). Familial aggregation of these malignancies has been reported in registry data (Dong and Hemminki, 2001a) and in many case-control studies (Linet and Pottern, 1992; Pottern et al., 1991; Zhu et al., 1998). Conversely, survivors of childhood Hodgkin lymphoma have increased risk for NHL, estimated to be sevenfold in one series (Metayer et al., 2000). The estimates of relative risk of developing NHL associated with a history of myeloma or Hodgkin lymphoma in the family vary widely among studies, but typically are slightly lower than the estimates for a history of NHL in the family. The development of skin cancers following NHL and vice versa has been studied in several populations, as noted in relation to the ultraviolet radiation hypothesis. Some of the correlation may reflect effects of treatment and some may reflect unidentified common etiology, possibly a defect in immune response. Few other cancers have been reported consistently to be in excess either as additional primaries in NHL patients or in their families.

Organ Transplantation Lymphoma can arise quickly in organ transplant patients receiving immunosuppressive therapy (Hoover and Fraumenti, 1973; Knowles, 1999; Levine, 1994; Penn et al., 1969). The strength of the association is correlated with the type of transplantation, the size of the organ transplanted, and the degree of immunosuppression needed to prevent rejection (Hoover and Fraumenti, 1973). Data from a large, multicenter study in Europe and North America showed that lymphomas occurred in 0.2% of kidney transplant recipients and in 1.2% of heart transplant recipients during the first year, rates that were 20 and 120 times higher than those seen in the general population (Opelz and Henderson, 1993; Opelz et al., 1994). Post-transplant lymphoma occurred in 2% of a large series of renal transplant patients, usually extranodally, often within a few months (Bates et al., 2003). In a large cohort of Swedish kidney transplant patients, relative risks for NHL were 9.9 in the first year and 3.2 thereafter. Recipients of other organs had relative risks of 38.0 and 24.3, respectively (Adami et al., 2003). Similarly, post-transplant NHL risk has been linked to the use of specific medications, including cyclosporine, monoclonal antibodies for T-cell depletion, prednisone, and azathioprine. In the Collaborative Transplant Study, heart transplant recipients received higher doses of cyclosporine and azathioprine than did kidney transplant recipients, and their lymphoma risk was three times as high (Swinnen et al., 1990). Furthermore, individuals needing cyclosporine, azathioprine, and steroids simultaneously experienced the highest lymphoma incidence (Opelz et al., 1994). Two other features of transplantation lymphomas warrant comment. Cessation of the immunosuppressive therapy can sometimes lead to regression of the lymphoma (Bayerdorffer et al., 1995). Second, in virtually all tumors from transplantation patients, EBV DNA can be detected (Swinnen, 2000). Risk estimates are increased for EBVnegative transplant recipients especially within the first year posttransplant (Swerdlow et al., 2000) and have been associated with low levels of EBV T-cells and high EBV viral load (Smets et al., 2002). Post-transplant lymphomas that occur more than 1 year after organ transplant appear to be distinctly different from those that occur within the first year. Compared with lymphomas with earlier onset, later onset lymphomas are more likely to occur among adults, are more likely to be EBV-negative tumors, and are less likely to respond positively to reductions in immunosuppressive therapies (Penn, 2000).

Primary Immune Deficiencies For patients with primary or inherited immune deficiencies, the risk for lymphoproliferative diseases, including lymphomas, is greatly increased (Knowles, 1999). These primary immune deficiencies include Wiskott-Aldrich syndrome (WAS), X-linked lymphoproliferative disease (XLP), ataxia-telangiectasia (AT), and common variable immunodeficiency (Filipovich et al., 1994; Knowles, 1999). The excess lymphomas are independent of other risk factors, and the absolute risks are very high. These disorders entail a large degree of immune deficiency, and the associated lymphomas tend to be clinically aggressive, high-grade malignancies. As many as one-quarter of all patients with congenital immune deficiencies develop cancers during their lifetime, half of them lymphomas (Filipovich et al., 1992). Individuals with AT have a 50-fold to 150-fold risk of cancer in general; approximately 10% of children with AT develop lymphoma (Levine, 1994). This autosomal recessive disorder is caused by mutations of both alleles in the ATM gene on the long arm of chromosome 11 (11q22.3), which encodes for an enzyme that regulates cell division following DNA damage. AT is a genetic neurodegenerative disorder with early onset, oculocutaneous telangiectasia, and cellular and humoral immunodeficiency. The frequency of AT ranges from 1 per 40,000 to 1 per 300,000 births (Olsen et al., 2001). Lymphomas developing in individuals with AT are largely B-cell in type, and contrary to other primary immune deficiency-related lymphomas, EBV usually is not present (Levine, 1994). Wiskott-Aldrich syndrome is an X-linked recessive disorder caused by mutations in the long arm of the WAS protein gene, Xp11.23p11.22, with symptoms that include immunodeficiency, eczema, thrombocytopenia, and recurrent infections. WAS occurs in 4 per 1 million live male births in the United States. Fourteen percent of boys with WAS develop lymphomas, typically with EBV present in the tumors (Levine, 1994). These lymphomas are primarily large-cell immunoblastic B-cell lymphomas. Many are extranodal, with CNS involvement in one-quarter (Levine, 1994). Common variable immunodeficiency refers to a heterogeneous collection of rare genetic disorders characterized by abnormalities in immune maturation. Patients suffer from antibody deficiency, hypogammaglobulinemia (e.g., reduction in immunoglobulins in serum), and recurrent bacterial infections. Probably an autosomal recessive disorder, common variable immunodeficiency affects approximately 1 in 10,000 to 100,000 men and women. It has been estimated that 1.4%–7% of affected people develop lymphoproliferative disease and lymphoma (Elenitoba-Johnson et al., 1997). These lymphomas are primarily extranodal B-cell tumors (Filipovich et al., 1992). X-linked lymphoproliferative disease is caused by mutations in the SAP/SH2D1A gene. The EBV genome is present in virtually all tissue samples from XLP lymphomas (Levine, 1994), and the syndrome itself includes defective immune response to EBV infection. It is estimated that 20%–35% of boys with XLP develop lymphoma (Levine, 1994; Morra, 1999). XLP-associated lymphomas are usually extranodal and B-cell in type, including small cleaved and immunoblastic lymphomas (Levine, 1994). Other congenital immunodeficiencies associated with higher frequency of lymphomas include severe combined immunodeficiency, hyper-IgM syndrome, Chediak-Higashi syndrome, B-cell proliferative syndrome, and Bruton agammaglobulinemia.

Medical Conditions Rheumatoid arthritis, a common inflammatory connective tissue disease, has been linked to both B-cell and T-cell lymphomas in several cohort studies. Finnish investigators demonstrated a twofold risk of lymphoma in a registry-based study (Ehrenfeld et al., 2001; Hakulinen et al., 1985; Kauppi and Hakala, 1994). Swedish and Danish investigators reported similar excess risk (Gridley et al., 1993; Mellemkjaer et al., 1996). For Felty syndrome, a rare complication affecting 1% of rheumatoid arthritis patients, an eightfold higher risk of lymphoma was reported based on a retrospective cohort study of US men (Gridley et al., 1994).

Non-Hodgkin Lymphoma The effects of medication complicate the interpretation of findings on autoimmune diseases. A Finnish study using registry data demonstrated excess risk of lymphoma for rheumatoid arthritis patients receiving cyclophosphamide (Hakulinen et al., 1985) whereas a 3-year prospective French study found no increased risk among rheumatoid arthritis patients treated with methotrexate (Mariette et al., 2002). In a review of autoimmune rheumatic diseases, rheumatoid arthritis patients who were on immunosuppressive therapies (e.g., aziothioprine, cyclophosphamide, and chlorambucil) had 10 times the risk for lymphoma, with longer use and higher doses correlating with level of risk (Leandro and Isenberg, 2001). Other clinical and laboratory evidence supports a biologically plausible association between EBVrelated NHL and rheumatoid arthritis (Ollier, 2000) perhaps due to immunosuppressive therapies, although no association was found in a case-control study of NHL (Kamel et al., 1999). A Swedish cohort study reported an increased frequency of DLBCL among rheumatoid arthritis patients, possibly related to the extent of rheumatoid arthritis inflammatory activity (Baecklund et al., 2003). Results from this study could not confirm an association with EBV-related NHL but showed that of the five EBV-related lymphomas, four occurred among patients with high inflammatory activity who also had taken immunosuppressive therapies (Baecklund et al., 2003). Sjögren syndrome is characterized by the progressive destruction of salivary and lacrimal glands and affects 2 to 4 million individuals in the United States, predominantly women. An increased risk for developing B-cell lymphoma has been demonstrated, with very high risk estimates in case series (Bloch et al., 1992; Cohen and Scadden, 2001; Kassan et al., 1978; Valesini et al., 1997). Recent data from the Finnish Cancer Registry showed fourfold lymphoma risk with secondary Sjogren syndrome and ninefold with primary Sjogren syndrome (Kauppi et al., 1997). Systemic lupus erythematosus patients have substantially increased risk for B-cell lymphoma (Mellemkjaer et al., 1997; Pettersson et al., 1993). Animal models provide additional evidence of a causal association (Mellors, 1966). A fivefold risk of lymphomas appeared in Danish Cancer Registry data (Mellemkjaer et al., 1997). A populationbased case-control study in Italy estimated an eightfold risk for developing lymphoma in systemic lupus erythematosus patients (Vineis et al., 2000). Lymphoma risk increased to sevenfold in a Canadian cohort (Cibere et al., 2001). Celiac disease patients also exhibit elevated rates of lymphoma. In an Italian cohort, risk of NHL mortality was very high (SMR = 69) (Corrao et al., 2001). In a Swedish population registry study, sixfold risk was seen (Askling et al., 2002). In a US cohort, an overall relative risk of 9.1 was reported, with a relative risk of 6.2 even for those on a gluten-free diet (Green et al., 2003). While both B-cell and T-cell lymphomas have been reported in celiac disease patients, there appears to be a predominance of extranodal T-cell lymphomas. In blood transfusion recipients, elevated risk of developing lymphomas has been reported in some studies but not in others (Chow and Holly, 2002a). In particular, three cohort studies found positive associations (Blomberg et al., 1993; Cerhan et al., 2001; Memon and Doll, 1994), but case-control studies mostly have not, including a large study in San Francisco (Chow and Holly, 2002b). Finally, various common medical conditions and medications have been associated with lymphomas in one or two studies, but not established as risk factors. People with asthma, allergies, and eczema have been inconsistently reported as showing decreased risk of developing lymphomas (Doody et al., 1992; Eriksson et al., 1995; Holly and Lele, 1997; Mills et al., 1992). Vaccinations (Holly and Lele, 1997; Holly and Bracci, 2003) and injection drug use (Bernstein and Ross, 1992) may alter risk. One study found increased risk with use of amphetamines (Doody et al., 1996), another with endometriosis (Olsen et al., 2001) and another with adult-onset diabetes (Cerhan et al., 1997). Consideration of distinct lymphomas or of biomarkers of immunity, as in a pooled genetic study (Rothman et al., 2006), may clarify this area. Epidemiologic studies will continue to investigate leads on this wide range of medical conditions and medications, with a focus on conditions with moderately large potential for modulating immune response.

911

FUTURE DIRECTIONS The hundreds of investigations of NHL etiology have yielded some established associations and numerous leads to pursue, but they have not explained most of the burden of the disease. The epidemic of NHL traceable to HIV infection was uncovered quickly and has abated in many geographic locations, while the longer and slower epidemic of the last half of the 20th century remains poorly understood. Most investigations have considered NHL as a whole, even though particular agents, especially infections, often produce lymphomas that are clinically or immunohistochemically distinctive. Many epidemiologists believe that classifying the lymphomas according to their molecular or genetic characteristics will uncover more etiologically distinctive subsets. From the histologically based WHO classification to the lymphoma tissue expression arrays that hold such promise for predicting and improving survival, new ways to consider NHL are emerging. There is good reason to believe that distinguishing among the lymphomas will reveal etiology by histologic subtype. A related approach is to divide the aggregate population into subpopulations according to genetic susceptibility. Investigative teams of geneticists and epidemiologists are evaluating hundreds of genetic polymorphisms, especially in the pathways related to infection and inflammation. For some exposures, researchers are pursuing more accurate measures of agents in the environment or as absorbed in the body. These and other biologically oriented approaches, singly or in combination, may illuminate the causal pathways. Striking individual case reports may arise and yield new clues. Epidemiologists will continue to examine the development of lymphoma in AIDS and transplant patients to understand co-factors and late natural history of lymphoma development. They will study agricultural workers, populations exposed to particular viruses, and other especially informative groups to measure their risks of developing lymphoma. As data from investigations conducted in the general population are divided into various subsets to discover etiology, meta-analyses, cooperative parallel studies, and consortia will be used more widely. Biospecimens collected before diagnosis offer special opportunities to assess pathogenesis. Because of the relative rarity of lymphomas in the general population, even the largest cohorts with stored blood samples typically yield a few hundred lymphoma cases at most. Studies with these precious pre-diagnosis biological specimens will be used primarily to confirm and clarify risks already reported in casecontrol studies and elsewhere. Advances in the understanding of Hodgkin lymphoma, multiple myeloma, chronic lymphocytic and other leukemia and non-malignant immune conditions also will be examined for possible relevance to lymphoma. Increasingly sophisticated models of the immune system and its controlling genetic pathways, in addition to advancing technology, will present an enormous array of options for analyzing data gathered in epidemiologic studies. As a clearer understanding of the causal pathways emerges, the prospects for intervening to prevent the occurrence of NHL will brighten. Acknowledgment We thank Geoffrey Tobias for editorial assistance and graphical presentation.

References Adami HO, Tsaih S, Lambe M, et al. 1997. Pregnancy and risk of nonHodgkin’s lymphoma: A prospective study. Int J Cancer 70:155–158. Adami J, Frisch M, Yuen J, Glimelius B, Melbye M. 1995. Evidence of an association between non-Hodgkin’s lymphoma and skin cancer. BMJ 310:1491–1495. Adami J, Gabel H, Lindelof B, et al. 2003. Cancer risk following organ transplantation: A nationwide cohort study in Sweden. Br J Cancer 89:1221– 1227. Adami J, Glimelius B, Cnattingius S, et al. 1996. Maternal and perinatal factors associated with non-Hodgkin’s lymphoma among children. Int J Cancer 65:774–777. Adami J, Gridley G, Nyren O, et al. 1999. Sunlight and non-Hodgkin’s lymphoma: A population-based cohort study in Sweden. Int J Cancer 80:641–645.

912

PART IV: CANCER BY TISSUE OF ORIGIN

Adami J, Nyren O, Bergstrom R, et al. 1998. Smoking and the risk of leukemia, lymphoma, and multiple myeloma (Sweden). Cancer Causes Control 9:49–56. Ahlbom IC, Cardis E, Green A, Linet M, Savitz D, Swerdlow A. 2001. Review of the epidemiologic literature on EMF and Health. Environ Health Perspect 109 Suppl 6:911–933. Ahsan H, Neugut AI. 1996. High risk of Kaposi’s sarcoma and central nervous system lymphoma in the same individuals: A finding related to acquired immunodeficiency syndrome. Int J Cancer 66:176–178. Alexander FE, McKinney PA, Moncrieff KC, Cartwright RA. 1992. Residential proximity of children with leukaemia and non-Hodgkin’s lymphoma in three areas of northern England. Br J Cancer 65:583–588. Alizadeh AA, Eisen MB, Davis RE, et al. 2000. Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 403:503–511. Alpen B, Neubauer A, Dierlamm J, et al. 2000. Translocation t(11;18) absent in early gastric marginal zone B-cell lymphoma of MALT type responding to eradication of Helicobacter pylori infection. Blood 95:4014–4015. Altekruse SF, Henley SJ, Thun MJ. 1999. Deaths from hematopoietic and other cancers in relation to permanent hair dye use in a large prospective study (United States). Cancer Causes Control 10:617–625. Archer VE, Wagoner JK, Lundin FE, Jr. 1973. Cancer mortality among uranium mill workers. J Occup Med 15:11–14. Ascoli V, Lo CF, Torelli G, et al. 2002. Human herpesvirus 8-associated primary effusion lymphoma in HIV—patients: A clinicopidemiologic variant resembling classic Kaposi’s sarcoma. Haematologica 87:339–343. Askling J, Linet M, Gridley G, Halstensen TS, Ekstrom K, Ekbom A. 2002. Cancer incidence in a population-based cohort of individuals hospitalized with celiac disease or dermatitis herpetiformis. Gastroenterology 123:1428–1435. Baecklund E, Sundstrom C, Ekbom A, et al. 2003. Lymphoma subtypes in patients with rheumatoid arthritis: Increased proportion of diffuse large B cell lymphoma. Arthritis Rheum 48:1543–1550. Bartholomew C, Jack N, Edwards J, et al. 1998. HTLV-I serostatus of mothers of patients with adult T-cell leukemia and HTLV-I-associated myelopathy/tropical spastic paraparesis. J Hum Virol 1:302–305. Bates WD, Gray DW, Dada MA, et al. 2003. Lymphoproliferative disorders in Oxford renal transplant recipients. J Clin Pathol 56:439–446. Bayerdorffer E, Neubauer A, Rudolph B, et al. 1995. Regression of primary gastric lymphoma of mucosa-associated lymphoid tissue type after cure of Helicobacter pylori infection. MALT Lymphoma Study Group. Lancet 345:1591–1594. Becker N, Berger J, Bolm-Audorff U. 2001. Asbestos exposure and malignant lymphomas—a review of the epidemiological literature. Int Arch Occup Environ Health 74:459–469. Bener A, Denic S, Al Mazrouei M. 2001. Consanguinity and family history of cancer in children with leukemia and lymphomas. Cancer 92:1–6. Bentham G. 1996. Association between incidence of non-Hodgkin’s lymphoma and solar ultraviolet radiation in England and Wales. BMJ 312:1128–1131. Beral V, Inskip H, Fraser P, Booth M, Coleman D, Rose G. 1985. Mortality of employees of the United Kingdom Atomic Energy Authority, 1946–1979. Br Med J (Clin Res Ed) 291:440–447. Beral V, Peterman T, Berkelman R, Jaffe H. 1991. AIDS-associated nonHodgkin lymphoma. Lancet 337:805–809. Bernstein L, Ross RK. 1992. Prior medication use and health history as risk factors for non-Hodgkin’s lymphoma: Preliminary results from a casecontrol study in Los Angeles County. Cancer Res 52:5510s–5515s. Berrington A, Darby SC, Weiss HA, Doll R. 2001. 100 years of observation on British radiologists: Mortality from cancer and other causes 1897–1997. Br J Radiol 74:507–519. Besson C, Goubar A, Gabarre J, et al. 2001. Changes in AIDS-related lymphoma since the era of highly active antiretroviral therapy. Blood 98:2339–2344. Biggar RJ. 2000. Epidemiology of malignancies in HIV/AIDS. In: Feigal EG, Levine AM, Biggar RJ, eds. AIDS-Related Cancers and Their Treatment. New York: M. Dekker, pp. 25–58. Biggar RJ, Engels EA, Frisch M, Goedert JJ. 2001. Risk of T-cell lymphomas in persons with AIDS. J Acquir Immune Defic Syndr 26:371–376. Blair A, Hartge P, Stewart PA, McAdams M, Lubin J. 1998. Mortality and cancer incidence of aircraft maintenance workers exposed to trichloroethylene and other organic solvents and chemicals: Extended follow up. Occup Environ Med 55:161–171. Bloch KJ, Buchanan WW, Wohl MJ, Bunim JJ. 1992. Sjogren’s syndrome. A clinical, pathological, and serological study of sixty-two cases. 1965. Medicine (Baltimore) 71:386–401. Blomberg J, Moller T, Olsson H, Anderson H, Jonsson M. 1993. Cancer morbidity in blood recipients—results of a cohort study. Eur J Cancer 29A:2101–2105.

Boffetta P, Andersen A, Lynge E, Barlow L, Pukkala E. 1994. Employment as hairdresser and risk of ovarian cancer and non-Hodgkin’s lymphomas among women. J Occup Med 36:61–65. Boice JD, Jr., Marano DE, Fryzek JP, Sadler CJ, McLaughlin JK. 1999. Mortality among aircraft manufacturing workers. Occup Environ Med 56:581–597. Borisch B, Raphael M, Swerdlow SH, Jaffe E. 2001. Lymphoproliferative diseases associated with primary immune disorders. In: Jaffe E, Harris NL, Stein H, Variman J, eds. Pathology and Genetics of Tumours of Haematopoietic and Lymphoid Tissues. Lyon, France: IARC Press, pp. 257–259. Boulanger E, Agbalika F, Maarek O, et al. 2001. A clinical, molecular and cytogenetic study of 12 cases of human herpesvirus 8 associated primary effusion lymphoma in HIV-infected patients. Hematol J 2:172–179. Briggs NC, Levine RS, Bobo LD, Haliburton WP, Brann EA, Hennekens CH. 2002. Wine drinking and risk of non-Hodgkin’s lymphoma among men in the United States: A population-based case-control study. Am J Epidemiol 156:454–462. Brinton LA, Gridley G, Persson I, Baron J, Bergqvist A. 1997. Cancer risk after a hospital discharge diagnosis of endometriosis. Am J Obstet Gynecol 176:572–579. Brown LM. 2000. Helicobacter pylori: Epidemiology and routes of transmission. Epidemiol Rev 22:283–297. Brown LM, Gibson R, Burmeister LF, Schuman LM, Everett GD, Blair A. 1992. Alcohol consumption and risk of leukemia, non-Hodgkin’s lymphoma, and multiple myeloma. Leuk Res 16:979–984. Burkitt D. 1958. A sarcoma involving the jaw in African children. Br J Radiol 1958:218–223. Burnett C, Robinson C, Walker J. 1999. Cancer mortality in health and science technicians. Am J Ind Med 36:155–158. Burns CJ, Beard KK, Cartmill JB. 2001. Mortality in chemical workers potentially exposed to 2,4-dichlorophenoxyacetic acid (2,4-D) 1945–94: An update. Occup Environ Med 58:24–30. Cacoub P, Fabiani FL, Musset L, et al. 1994. Mixed cryoglobulinemia and hepatitis C virus. Am J Med 96:124–132. Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ. 2003. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med 348:1625–1638. Cantor KP, Blair A, Everett G, et al. 1988. Hair dye use and risk of leukemia and lymphoma. Am J Public Health 78:570–571. Cantor KP, Strickland PT, Brock JW, et al. 2003. Risk of nonHodgkin’s lymphoma and prediagnostic serum organochlorines: betahexachlorocyclohexane, chlordane/heptachlor-related compounds, dieldrin, and hexachlorobenzene. Environ Health Perspect 111:179–183. Carbone A. 2003. Emerging pathways in the development of AIDS-related lymphomas. Lancet Oncol 4:22–29. Carbone A, Gloghini A, Capello D, Gaidano G. 2001. Genetic pathways and histogenetic models of AIDS-related lymphomas. Eur J Cancer 37:1270–1275. Carbone M, Pass HI, Miele L, Bocchetta M. 2003. New developments about the association of SV40 with human mesothelioma. Oncogene 22:5173–5180. Cardis E, Gilbert ES, Carpenter L, et al. 1995. Effects of low doses and low dose rates of external ionizing radiation: Cancer mortality among nuclear industry workers in three countries. Radiat Res 142: 117–132. Carpenter L, Higgins C, Douglas A, Fraser P, Beral V, Smith P. 1994. Combined analysis of mortality in three United Kingdom nuclear industry workforces, 1946–1988. Radiat Res 138:224–238. Carroll-Pankhurst C, Engels EA, Strickler HD, Goedert JJ, Wagner J, Mortimer EA, Jr. 2001. Thirty-five year mortality following receipt of. Br J Cancer 85:1295–1297. Centers for Disease Control (CDC). 1982. Update on acquired immune deficiency syndrome (AIDS)—United States. MMWR Morb Mortal Wkly Rep 31:507–504. Cerhan JR, Habermann TM, Vachon CM, et al. 2002a. Menstrual and reproductive factors and risk of non-Hodgkin lymphoma: The lowa women’s health study (United States). Cancer Causes Control 13: 131–136. Cerhan JR, Vachon CM, Habermann TM, et al. 2002b. Hormone replacement therapy and risk of non-hodgkin lymphoma and chronic lymphocytic leukemia. Cancer Epidemiol Biomarkers Prev 11:1466–1471. Cerhan JR, Wallace RB, Dick F, et al. 2001. Blood transfusions and risk of non-Hodgkin’s lymphoma subtypes and chronic lymphocytic leukemia. Cancer Epidemiol Biomarkers Prev 10:361–368. Cerhan JR, Wallace RB, Folsom AR, et al. 1997. Medical history risk factors for non-Hodgkin’s lymphoma in older women. J Natl Cancer Inst 89:314–318.

Non-Hodgkin Lymphoma Cesarman E, Chang Y, Moore PS, Said JW, Knowles DM. 1995. Kaposi’s sarcoma-associated herpesvirus-like DNA sequences in AIDS-related body-cavity-based lymphomas. N Engl J Med 332:1186–1191. Cesarman E, Mesri EA. 1999. Virus-associated lymphomas. Curr Opin Oncol 11:322–332. Chaganti RS, Nanjangud G, Schmidt H, Teruya-Feldstein J. 2000. Recurring chromosomal abnormalities in non-Hodgkin’s lymphoma: Biologic and clinical significance. Semin Hematol 37:396–411. Chan JK. 2001. The new World Health Organization classification of lymphomas: The past, the present and the future. Hematol Oncol 19:129–150. Chiu BC, Cerhan JR, Folsom AR, et al. 1996. Diet and risk of non-Hodgkin lymphoma in older women. JAMA 275:1315–1321. Chiu BC, Weisenburger DD, Cantor KP, et al. 2002. Alcohol consumption, family history of hematolymphoproliferative cancer, and the risk of nonHodgkin’s lymphoma in men. Ann Epidemiol 12:309–315. Chow EJ, Holly EA. 2002a. Blood transfusions and non-Hodgkin’s lymphoma. Epidemiol Rev 24:269–279. Chow EJ, Holly EA. 2002b. Blood transfusions as a risk factor for nonHodgkin’s lymphoma in the San Francisco Bay Area: A population-based study. Am J Epidemiol 155:725–731. Cibere J, Sibley J, Haga M. 2001. Systemic lupus erythematosus and the risk of malignancy. Lupus 10:394–400. Clarke CA, Glaser SL. 2001. Epidemiologic trends in HIV-associated lymphomas. Curr Opin Oncol 13:354–359. Cleghorn FR, Manns A, Falk R, et al. 1995. Effect of human T-lymphotropic virus type I infection on non-Hodgkin’s lymphoma incidence. J Natl Cancer Inst 87:1009–1014. Cocco P, Broccia G, Aru G, et al. 2003. Nitrate in community water supplies and incidence of non-Hodgkin’s lymphoma in Sardinia, Italy. J Epidemiol Community Health 57:510–511. Cohen K, Scadden DT. 2001. Non-Hodgkin’s lymphoma: pathogenesis, clinical presentation, and treatment, Cancer Treat Res 104:201–230. Cohn P, Klotz J, Bove F, Berkowitz M, Fagliano J. 1994. Drinking water contamination and the incidence of leukemia and non-Hodgkin’s lymphoma. Environ Health Perspect 102:556–561. Corrao G, Corazza GR, Bagnardi V, et al. 2001. Mortality in patients with coeliac disease and their relatives: A cohort study. Lancet 358: 356–361. Correa A, Jackson L, Mohan A, Perry H, Helzlsouer K. 2000. Use of hair dyes, hematopoietic neoplasms, and lymphomas: A literature review. II. Lymphomas and multiple myeloma. Cancer Invest 18:467–479. Cote TR, Biggar RJ, Rosenberg PS, et al. 1997. Non-Hodgkin’s lymphoma among people with AIDS: Incidence, presentation and public health burden. AIDS/Cancer Study Group. Int J Cancer 73:645–650. Cote TR, Manns A, Hardy CR, Yellin FJ, Hartge P. 1996. Epidemiology of brain lymphoma among people with or without acquired immunodeficiency syndrome. AIDS/Cancer Study Group. J Natl Cancer Inst 88: 675–679. Council of State and Territorial Epidemiologists. 1985. Revision of the case definition of acquired immunodeficiency syndrome for national reporting—United States. MMWR Morb Mortal Wkly Rep 34:373–375. Council of State and Territorial Epidemiologists. 1987. Revision of the CDC surveillance case definition for acquired immunodeficiency syndrome. Council of State and Territorial Epidemiologists; AIDS Program, Center for Infectious Diseases. MMWR Morb Mortal Wkly Rep 36 Suppl 1:1S–15S. Cucuianu A, Patiu M, Duma M, et al. 1999. Hepatitis B and C virus infection in Romanian non-Hodgkin’s lymphoma patients. Br J Haematol 107:353–356. Cuttner J, Werther JL, McGlynn P, et al. 2001. Seroprevalence of Helicobacter pylori infection in patients with lymphoma. Leuk Lymphoma 40:591–597. Daibata M, Nemoto Y, Kamioka M, Imai S, Taguchi H. 2003. Simian virus 40 in Japanese patients with lymphoproliferative disorders. Br J Haematol 121:190–191. dal Maso L, Franceschi S. 2003. Epidemiology of non-Hodgkin lymphomas and other haemolymphopoietic neoplasms in people with AIDS. Lancet Oncol 4:110–119. Dalla-Favera R, Gaidano G. 2001. Molecular biology of lymphomas. In: DeVita VTJ, Hellman S, Rosenberg SA, eds. Cancer: Principles and Practice of Oncology. Philadelphia: Lippincott Williams, and Wilkins, pp. 2215–2234. Damber L, Larsson LG, Johansson L, Norin T. 1995. A cohort study with regard to the risk of haematological malignancies in patients treated with x-rays for benign lesions in the locomotor system. I. Epidemiological analyses. Acta Oncol 34:713–719. David H, Mendoza S, Konishi T, Miller CW. 2001. Simian virus 40 is present in human lymphomas and normal blood. Cancer Lett 162:57–64.

913

De Roos AJ, Zahm SH, Cantor KP, et al. 2003. Integrative assessment of multiple pesticides as risk factors for non-Hodgkin’s lymphoma among men. Occup Environ Med 60:E11. De Rosa G, Gobbo ML, De Renzo A, et al. 1997. High prevalence of hepatitis C virus infection in patients with B-cell lymphoproliferative disorders in Italy. Am J Hematol 55:77–82. de Sanjose S, Shah KV, Domingo-Domenech E, et al. 2003. Lack of serological evidence for an association between simian virus 40 and lymphoma. Int J Cancer 104:522–524. Devesa SS, Fears T. 1992. Non-Hodgkin’s lymphoma time trends: United States and international data. Cancer Res 52:5432s–5440s. Diamandopoulos GT. 1972. Leukemia, lymphoma, and osteosarcoma induced in the Syrian golden hamster by simian virus 40. Science 176:173–175. Dickinson HO, Parker L. 2002. Leukaemia and non-Hodgkin’s lymphoma in children of male Sellafield radiation workers. Int J Cancer 99:437–444. Doglioni C, Wotherspoon AC, Moschini A, De Boni M, Isaacson PG. 1992. High incidence of primary gastric lymphoma in northeastern Italy. Lancet 339:834–835. Dong C, Hemminki K. 2001a. Modification of cancer risks in offspring by sibling and parental cancers from 2,112,616 nuclear families. Int J Cancer 92:144–150. Dong C, Hemminki K. 2001b. Second primary neoplasms among 53 159 haematolymphoproliferative malignancy patients in Sweden, 1958–1996: A search for common mechanisms. Br J Cancer 85:997–1005. Doody MM, Linet MS, Glass AG, et al. 1996. Risks of non-Hodgkin’s lymphoma, multiple myeloma, and leukemia associated with common medication. Epidemiology 7:131–139. Doody MM, Linet MS, Glass AG, et al. 1992. Leukemia, lymphoma, and multiple myeloma following selected medical conditions. Cancer Causes Control 3:449–456. Dore GJ, Li Y, McDonald A, Ree H, Kaldor JM, Kaldo JM. 2002. Impact of highly active antiretroviral therapy on individual AIDS-defining illness incidence and survival in Australia. J Acquir Immune Defic Syndr 29:388–395. Egan KM, Sosman JA, Blot WJ. 2005. Sunlight and reduced risk of cancer: is the real story vitamin D? J Natl Cancer Inst Feb 2;97(3):161–163. Eheman CR, Tolbert PE, Coates RJ, Devine O, Eley JW. 2000. Case-control assessment of the association between non-Hodgkin’s lymphoma and occupational radiation with doses assessed using a job exposure matrix. Am J Ind Med 38:19–27. Ehrenfeld M, Abu-Shakra M, Buskila D, Shoenfeld Y. 2001. The dual association between lymphoma and autoimmunity. Blood Cells Mol Dis 27:750–756. Elenitoba-Johnson KS, Jaffe ES. 1997. Lymphoproliferative disorders associated with congenital immunodeficiencies. Semin Diagn Pathol 14:35–47. Eltom MA, Jemal A, Mbulaiteye SM, Devesa SS, Biggar RJ. 2002. Trends in Kaposi’s sarcoma and non-Hodgkin’s lymphoma incidence in the United States from 1973 through 1998. J Natl Cancer Inst 94:1204–1210. Engels EA, Chatterjee N, Cerhan JR, et al. 2003a. Hepatitis C virus infection and non-Hodgkin lymphoma: Results of the NCI-SEER multi-center case-control study. Int J Cancer 111:76–80. Engels EA, Katki HA, Nielsen NM, et al. 2003b. Cancer incidence in Denmark following exposure to poliovirus vaccine contaminated with simian virus 40. J Natl Cancer Inst 95:532–539. Engels EA, Rodman LH, Frisch M, Goedert JJ, Biggar RJ. 2003c. Childhood exposure to simian virus 40-contaminated poliovirus vaccine and risk of AIDS-associated non-Hodgkin’s lymphoma. Int J Cancer 106:283–287. Engels EA, Viscidi RP, Galloway DA, et al. 2004. Case-control study of simian virus 40 and non-Hodgkin lymphoma in the United States. JNCI 6(18):1368–1374. Engels EA, Chen J, Hartge P, et al. 2005. Antibody responses to simian virus 40 T antigen: a case-control study of non-Hodgkin lymphoma. Cancer Epi Bio Prev Feb;14(2):521–524. Eriksson NE, Holmen A, Hogstedt B, Mikoczy Z, Hagmar L. 1995. A prospective study of cancer incidence in a cohort examined for allergy. Allergy 50:718–722. Evans LS, Hancock BW. 2003. Non-Hodgkin lymphoma. Lancet 362:139–146. Ferri C, Caracciolo F, Zignego AL, et al. 1994. Hepatitis C virus infection in patients with non-Hodgkin’s lymphoma. Br J Haematol 88:392–394. Figgs LW, Dosemeci M, Blair A. 1995. United States non-Hodgkin’s lymphoma surveillance by occupation 1984–1989: A twenty-four state death certificate study. Am J Ind Med 27:817–835. Filipovich AH, Mathur A, Kamat D, Kersey JH, Shapiro RS. 1994. Lymphoproliferative disorders and other tumors complicating immunodeficiencies. Immunodeficiency 5:91–112. Filipovich AH, Mathur A, Kamat D, Shapiro RS. 1992. Primary immunodeficiencies: Genetic risk factors for lymphoma. Cancer Res 52:5465s– 5467s.

914

PART IV: CANCER BY TISSUE OF ORIGIN

Fisher SG, Weber L, Carbone M. 1999. Cancer risk associated with simian virus 40 contaminated polio vaccine. Anticancer Res 19:2173–2180. Floret N, Mauny F, Challier B, Arveux P, Cahn JY, Viel JF. 2003. Dioxin emissions from a solid waste incinerator and risk of non-Hodgkin lymphoma. Epidemiology 14:392–398. Freedman DM, Cantor KP, Ward MH, Helzlsouer KJ. 2000. A case-control study of nitrate in drinking water and non-Hodgkin’s lymphoma in Minnesota. Arch Environ Health 55:326–329. Freedman DM, Zahm SH, Dosemeci M. 1997. Residential and occupational exposure to sunlight and mortality from non-Hodgkin’s lymphoma: composite (threefold) case-control study. BMJ 314:1451–1455. Fritschi L, Johnson KC, Kliewer EV, Fry R. 2002. Animal-related occupations and the risk of leukemia, myeloma, and non-Hodgkin’s lymphoma in Canada. Cancer Causes Control 13:563–571. Fritschi L, Siemiatycki J. 1996. Lymphoma, myeloma and occupation: results of a case-control study. Int J Cancer 67:498–503. Fritz A, Percy C, Jack A, Shanmugaratnam K, Sobin L, Parkin DM, Whelan S. 2000. International Classification of Diseases for Oncology. Geneva: World Health Organization. Gaidano G, Carbone A, Dalla-Favera R. 1998. Genetic basis of acquired immunodeficiency syndrome-related lymphomagenesis. J Natl Cancer Inst Monogr 95–100. Gaidano G, Dalla-Favera R. 1992. Biologic aspects of human immunodeficiency virus-related lymphoma. Curr Opin Oncol 4:900–906. Gaidano G, Pastore C, Gloghini A, et al. 1996. AIDS-related non-Hodgkin’s lymphomas: Molecular genetics, viral infection and cytokine deregulation. Acta Haematol 95:193–198. Gallo RC, Poiesz BJ, Ruscetti FW. 1981. Regulation of human T-cell proliferation: T-cell growth factor and isolation of a new class of type-C retroviruses from human T-cells. Haematol Blood Transfus 26:502–514. Garabedian MJ, Hoppin JA, Tolbert PE, Herrick RF, Brann EA. 1999. Occupational chlorophenol exposure and non-Hodgkin’s lymphoma. J Occup Environ Med 41:267–272. Gardner MJ, Snee MP, Hall AJ, Powell CA, Downes S, Terrell JD. 1990. Results of case-control study of leukaemia and lymphoma among young people near Sellafield nuclear plant in West Cumbria. BMJ 300:423– 429. Gazdar AF, Butel JS, Carbone M. 2002. SV40 and human tumours: Myth, association or causality? Nat Rev Cancer 2:957–964. Gerard L, Galicier L, Maillard A, et al. 2002. Systemic non-Hodgkin lymphoma in HIV-infected patients with effective suppression of HIV replication: Persistent occurrence but improved survival. J Acquir Immune Defic Syndr 30:478–484. Germanidis G, Haioun C, Dhumeaux D, Reyes F, Pawlotsky JM. 1999. Hepatitis C virus infection, mixed cryoglobulinemia, and B-cell non-Hodgkin’s lymphoma. Hepatology 30:822–823. Gilbert ES, Omohundro E, Buchanan JA, Holter NA. 1993. Mortality of workers at the Hanford site: 1945–1986. Health Phys 64:577–590. Goedert JJ, Cote TR, Virgo P, et al. 1998. Spectrum of AIDS-associated malignant disorders. Lancet 351:1833–1839. Goggins WB, Finkelstein DM, Tsao H. 2001. Evidence for an association between cutaneous melanoma and non-Hodgkin lymphoma. Cancer 91:874–880. Goldgar DE, Easton DF, Cannon-Albright LA, Skolnick MH. 1994. Systematic population-based assessment of cancer risk in first-degree relatives of cancer probands. J Natl Cancer Inst 86:1600–1608. Green PH, Fleischauer AT, Bhagat G, Goyal R, Jabri B, Neugut AI. 2003. Risk of malignancy in patients with celiac disease. Am J Med 115:191– 195. Gridley G, Klippel JH, Hoover RN, Fraumeni JF, Jr. 1994. Incidence of cancer among men with the Felty syndrome. Ann Intern Med 120:35–39. Gridley G, McLaughlin JK, Ekbom A, et al. 1993. Incidence of cancer among patients with rheumatoid arthritis. J Natl Cancer Inst 85:307–311. Grodstein F, Hennekens CH, Colditz GA, Hunter DJ, Stampfer MJ. 1994. A prospective study of permanent hair dye use and hematopoietic cancer. J Natl Cancer Inst 86:1466–1470. Groves FD, Linet MS, Travis LB, Devesa SS. 2000. Cancer surveillance series: Non-Hodgkin’s lymphoma incidence by histologic subtype in the United States from 1978 through 1995. J Natl Cancer Inst 92:1240–1251. Grulich AE, Wan X, Law MG, et al. 2000. B-cell stimulation and prolonged immune deficiency are risk factors for non-Hodgkin’s lymphoma in people with AIDS. AIDS 14:133–140. Gulis G, Czompolyova M, Cerhan JR. 2002. An ecologic study of nitrate in municipal drinking water and cancer incidence in Trnava District, Slovakia. Environ Res 88:182–187. Hakansson N, Floderus B, Gustavsson P, Feychting M, Hallin N. 2001. Occupational sunlight exposure and cancer incidence among Swedish construction workers. Epidemiology 12:552–557.

Hakulinen T, Isomaki H, Knekt P. 1985. Rheumatoid arthritis and cancer studies based on linking nationwide registries in Finland. Am J Med 78:29–32. Hansen J, Raaschou-Nielsen O, Christensen JM, et al. 2001. Cancer incidence among Danish workers exposed to trichloroethylene. J Occup Environ Med 43:133–139. Hardell L, Eriksson M. 1999. A case-control study of non-Hodgkin lymphoma and exposure to pesticides. Cancer 85:1353–1360. Hardell L, Eriksson M, Degerman A. 1994. Exposure to phenoxyacetic acids, chlorophenols, or organic solvents in relation to histopathology, stage, and anatomical localization of non-Hodgkin’s lymphoma. Cancer Res 54:2386–2389. Hardell L, Eriksson M, Lenner P, Lundgren E. 1981. Maligant lymphoma and exposure to chemicals, especially organic solvents, chlorophenols and phenoxy acids: A case-control study. Br J Cancer 43:169–176. Hardell L, Eriksson M, Lindstrom G, et al. 2001a. Case-control study on concentrations of organohalogen compounds and titers of antibodies to Epstein-Barr virus antigens in the etiology of non-Hodgkin’s lymphoma. Leuk Lymphoma 42:619–629. Hardell L, Eriksson M, Nordstrom M. 2002. Exposure to pesticides as risk factor for non-Hodgkin’s lymphoma and hairy cell leukemia: Pooled analysis of two Swedish case-control studies. Leuk Lymphoma 43: 1043–1049. Hardell L, Liljegren G, Lindstrom G, et al. 1996a. Increased concentrations of chlordane in adipose tissue from non-Hodgkin’s lymphoma patients compared with controls without a malignant disease. Int J Oncol 9: 1139–1142. Hardell L, Lindstrom G, Van Bavel B, et al. 2001b. Adipose tissue concentrations of dioxins and dibenzofurans, titers of antibodies to Epstein-Barr virus early antigen and the risk for non-Hodgkin lymphoma. Environ Res 87:99–107. Hardell L, Van Bavel B, Lindstrom G, et al. 1996b. Higher concentrations of specific polychlorinated biphenyl congeners in adipose tissue from nonHodgkin’s lymphoma patients compared with controls without a malignant disease. Int J Oncol 9:603–608. Harris NL, Jaffe ES, Diebold J, et al. 1999. World Health Organization classification of neoplastic diseases of the hematopoietic and lymphoid tissues: Report of the Clinical Advisory Committee Meeting—Airlie House, Virginia, November 1997. J Clin Oncol 17:3835–3849. Harris NL, Jaffe ES, Stein H, et al. 1994. A revised European-American classification of lymphoid neoplasms: A proposal from the International Lymphoma Study Group. Blood 84:1361–1392. Harris NL, Stein H, Coupland SE, et al. 2001. New approaches to lymphoma diagnosis. Hematology (Am Soc Hematol Educ Program): 194–220. Hartge P, Devesa SS. 1992. Quantification of the impact of known risk factors on time trends in non-Hodgkin’s lymphoma incidence. Cancer Res 52:5566s–5569s. Hartge P, Devesa SS, Grauman D, Fears TR, Fraumeni JF, Jr. 1996. NonHodgkin’s lymphoma and sunlight. J Natl Cancer Inst 88:298–300. Hauke RJ, Armitage JO. 2000. A new approach to non-Hodgkin’s lymphoma. Intern Med 39:197–208. Hausfater P, Rosenthal E, Cacoub P. 2000. Lymphoproliferative diseases and hepatitis C virus infection. Ann Med Interne (Paris) 151:53–57. Hemminki K, Jiang Y, Steineck G. 2003. Skin cancer and non-Hodgkin’s lymphoma as second malignancies markers of impaired immune function? Eur J Cancer 39:223–229. Hermine O, Lefrere F, Bronowicki JP, et al. 2002. Regression of splenic lymphoma with villous lymphocytes after treatment of hepatitis C virus infection. N Engl J Med 347:89–94. Herndier BG, Kaplan LD, McGrath MS. 1994. Pathogenesis of AIDS lymphomas. AIDS 8:1025–1049. Herrinton LJ. 1998. Epidemiology of the Revised European-American Lymphoma Classification subtypes. Epidemiol Rev 20:187–203. Herrinton LJ, Friedman GD. 1998. Cigarette smoking and risk of nonHodgkin’s lymphoma subtypes. Cancer Epidemiol Biomarkers Prev 7:25–28. Hoar SK, Blair A, Holmes FF, et al. 1986. Agricultural herbicide use and risk of lymphoma and soft-tissue sarcoma. J Am Med Assn 256: 1141–1147. Hoar SK, Pell S. 1981. A retrospective cohort study of mortality and cancer incidence among chemist. J Occup Med 23:485–494. Holly EA, Bracci PM. 2003. Population-based study of non-Hodgkin lymphoma, histology, and medical history among human immunodeficiency virus-negative participants in San Francisco. Am J Epidemiol 158: 316–327. Holly EA, Gautam M, Bracci PM. 2002. Comparison of interviewed and noninterviewed non-Hodgkin’s lymphoma (NHL) patients in the San Francisco Bay Area. Ann Epidemiol 12:419–425.

Non-Hodgkin Lymphoma Holly EA, Lele C. 1997. Non-Hodgkin’s lymphoma in HIV-positive and HIVnegative homosexual men in the San Francisco Bay Area: Allergies, prior medication use, and sexual practices. J Acquir Immune Defic Syndr Hum Retrovirol 15:211–222. Holly EA, Lele C, Bracci PM. 1998. Hair-color products and risk for nonHodgkin’s lymphoma: A population-based study in the San Francisco bay area. Am J Public Health 88:1767–1773. Holly EA, Lele C, Bracci PM, McGrath MS. 1999. Case-control study of nonHodgkin’s lymphoma among women and heterosexual men in the San Francisco Bay Area, California. Am J Epidemiol 150:375–389. Hoover R, Fraumeni JFJ. 1973. Risk of cancer in renal-transplant recipients. Lancet 2:55–57. Hsu JL, Glaser SL. 2000. Epstein-barr virus-associated malignancies: epidemiologic patterns and etiologic implications. Crit Rev Oncol Hematol 34:27–53. Hughes AM, Armstrong BK, Vajdic CM, et al. 2004. Sun exposure may protect against non-Hodgkin lymphoma: a case-control study. Int J Cancer Dec 10;112(5):865–871. Hummel M, Anagnostopoulos I, Korbjuhn P, Stein H. 1995. Epstein-Barr virus in B-cell non-Hodgkin’s lymphomas: Unexpected infection patterns and different infection incidence in low- and high-grade types. J Pathol 175:263–271. Inskip PD, Kleinerman RA, Stovall M, et al. 1993. Leukemia, lymphoma, and multiple myeloma after pelvic radiotherapy for benign disease. Radiat Res 135:108–124. Immunization Safety Review Institute of Medicine. 2002. SV40 contamination of polio vaccine and cancer. Washington, DC: The National Academies Press. International Agency for Research on Cancer. 1995. IARC Monographs on the evaluation of carcinogenic risks to humans: Dry cleaning, some chlorinated solvents and other industrial chemicals. Lyon, France: International Agency for Research on Cancer. International Agency for Research on Cancer. 1997. Proceedings of the IARC Working Group on the Evaluation of Carcinogenic Risks to Humans. Epstein-Barr Virus and Kaposi’s Sarcoma Herpesvirus/Human Herpesvirus 8. Lyon, France, 17–24 June 1997. IARC Monogr Eval Carcinog Risks Hum 70:1–492. International Collaboration on HIV and Cancer. 2000. Highly active antiretroviral therapy and incidence of cancer in human immunodeficiency virusinfected adults. J Natl Cancer Inst 92:1823–1830. Jablon S, Miller RW. 1978. Army technologists: 29-year follow up for cause of death. Radiology 126:677–679. Jaffe E, Harris NL, Stein H, Vardiman J. 2001. World Health Organization Classification of Tumours: Pathology and Genetics of Tumours of Hematopoietic and Lymphoid Tissues. Lyon: IARC. Jemal A, Siegel R, Ward EM, Thun MJ. 2006. Cancer Facts & Figures 2006. CA Cancer J Clin 56:1–52. Johnson ES, Dalmas D, Noss J, Matanoski GM. 1995. Cancer mortality among workers in abattoirs and meatpacking plants: An update. Am J Ind Med 27:389–403. Johnson ES, Shorter C, Rider B, Jiles R. 1997. Mortality from cancer and other diseases in poultry slaughtering/processing plants. Int J Epidemiol 26:1142–1150. Jones JL, Hanson DL, Dworkin MS, Ward JW, Jaffe HW. 1999. Effect of antiretroviral therapy on recent trends in selected cancers among HIVinfected persons. Adult/Adolescent Spectrum of HIV Disease Project Group. J Acquir Immune Defic Syndr 21 Suppl 1:S11–S17. Kadan-Lottick NS, Skluzacek MC, Gurney JG. 2002. Decreasing incidence rates of primary central nervous system lymphoma. Cancer 95:193– 202. Kamel OW, Holly EA, van de Rijn M, Lele C, Sah A. 1999. A population based, case control study of non-Hodgkin’s lymphoma in patients with rheumatoid arthritis. J Rheumatol 26:1676–1680. Karcher DS, Alkan S. 1997. Human herpesvirus-8-associated body cavitybased lymphoma in human immunodeficiency virus-infected patients: A unique B-cell neoplasm. Hum Pathol 28:801–808. Kassan SS, Thomas TL, Moutsopoulos HM, et al. 1978. Increased risk of lymphoma in sicca syndrome. Ann Intern Med 89:888–892. Kauppi M, Hakala M. 1994. Prevalence of cervical spine subluxations and dislocations in a community-based rheumatoid arthritis population. Scand J Rheumatol 23:133–136. Kauppi M, Pukkala E, Isomaki H. 1997. Elevated incidence of hematologic malignancies in patients with Sjogren’s syndrome compared with patients with rheumatoid arthritis (Finland). Cancer Causes Control 8:201– 204. Kauppinen T, Pukkala E, Saalo A, Sasco AJ. 2003. Exposure to chemical carcinogens and risk of cancer among Finnish laboratory workers. Am J Ind Med 44:343–350.

915

Kendall GM, Muirhead CR, MacGibbon BH, et al. 1992. Mortality and occupational exposure to radiation: First analysis of the National Registry for Radiation Workers. BMJ 304:220–225. King PD, Wilkes JD, Diaz-Arias AA. 1998. Hepatitis C virus infection in nonHodgkin’s lymphoma. Clin Lab Haematol 20:107–110. Kinlen LJ, Clarke K, Balkwill A. 1993. Paternal preconceptional radiation exposure in the nuclear industry and leukaemia and non-Hodgkin’s lymphoma in young people in Scotland. BMJ 306:1153–1158. Kinlen LJ, Dickson M, Stiller CA. 1995. Childhood leukaemia and nonHodgkin’s lymphoma near large rural construction sites, with a comparison with Sellafield nuclear site. BMJ 310:763–768. Kirk O, Pedersen C, Cozzi-Lepri A, et al. 2001. Non-Hodgkin lymphoma in HIV-infected patients in the era of highly active antiretroviral therapy. Blood 98:3406–3412. Knowles DM. 1996. Etiology and pathogenesis of AIDS-related non-Hodgkin’s lymphoma. Hematol Oncol Clin North Am 10:1081–1109. Knowles DM. 1999. Immunodeficiency-associated lymphoproliferative disorders. Mod Pathol 12:200–217. Knowles DM, Cesarman E. 1997. The Kaposi’s sarcoma-associated herpesvirus (human herpesvirus-8) in Kaposi’s sarcoma, malignant lymphoma, and other diseases. Ann Oncol 8 Suppl 2:123–129. Kogevinas M, Kauppinen T, Winkelmann R, et al. 1995. Soft tissue sarcoma and non-Hodgkin’s lymphoma in workers exposed to phenoxy herbicides, chlorophenols, and dioxins: Two nested case-control studies. Epidemiology 6:396–402. Kristensen P, Andersen A, Irgens LM, Laake P, Bye AS. 1996. Incidence and risk factors of cancer among men and women in Norwegian agriculture. Scand J Work Environ Health 22:14–26. Lamba AB, Ward MH, Weeks JL, Dosemeci M. 2001. Cancer mortality patterns among hairdressers and barbers in 24 US states, 1984 to 1995. J Occup Environ Med 43:250–258. Lan Q, Zhang L, Li G, et al. 2004. Hematotoxicity in workers exposed to low levels of benzene. Science Dec 3;306(5702):1774–1776. Law GR, Parslow RC, Roman E. 2003. Childhood cancer and population mixing. Am J Epidemiol 158:328–336. Leandro MJ, Isenberg DA. 2001. Rheumatic diseases and malignancy—is there an association? Scand J Rheumatol 30:185–188. Lehtinen T, Lumio J, Dillner J, et al. 1993. Increased risk of malignant lymphoma indicated by elevated Epstein-Barr virus antibodies—a prospective study. Cancer Causes Control 4:187–193. Levi F, Randimbison L, Te VC, La Vecchia C. 1996. Non-Hodgkin’s lymphomas, chronic lymphocytic leukaemias and skin cancers. Br J Cancer 74:1847–1850. Levine AM. 1992. Acquired immunodeficiency syndrome-related lymphoma. Blood 80:8–20. Levine AM. 1994. Lymphoma complicating immunodeficiency disorders. Ann Oncol 5 Suppl 2:29–35. Linet MS, Pottern LM. 1992. Familial aggregation of hematopoietic malignancies and risk of non-Hodgkin’s lymphoma. Cancer Res 52:5468s–5473s. Liu H, Ruskon-Fourmestraux A, Lavergne-Slove A, et al. 2001. Resistance of t(11;18) positive gastric mucosa-associated lymphoid tissue lymphoma to Helicobacter pylori eradication therapy. Lancet 357:39–40. Liu S, Semenciw R, Mao Y. 2003. Increasing incidence of non-Hodgkin’s lymphoma in Canada, 1970–1996: Age-period-cohort analysis. Hematol Oncol 21:57–66. Luppi M, Ferrari MG, Torelli G. 1998. Hepatitis C virus genotype distribution in B-cell non-Hodgkin lymphoma. Ann Intern Med 128:602. Luppi P. 2003. How immune mechanisms are affected by pregnancy. Vaccine 21:3352–3357. MacKenzie J, Wilson KS, Perry J, Gallagher A, Jarrett RF. 2003. Association between simian virus 40 DNA and lymphoma in the United kingdom. J Natl Cancer Inst 95:1001–1003. Malkin D. 2002. Simian virus 40 and non-Hodgkin lymphoma. Lancet 359:812–813. Manns A, Cleghorn FR, Falk RT, et al. 1993. Role of HTLV-I in development of non-Hodgkin lymphoma in Jamaica and Trinidad and Tobago. The HTLV Lymphoma Study Group. Lancet 342:1447–1450. Mao Y, Hu J, Ugnat AM, White K. 2000. Non-Hodgkin’s lymphoma and occupational exposure to chemicals in Canada. Canadian Cancer Registries Epidemiology Research Group. Ann Oncol 11 Suppl 1:69–73. Mariette X, Cazals-Hatem D, Warszawki J, Liote F, Balandraud N, Sibilia J. 2002. Lymphomas in rheumatoid arthritis patients treated with methotrexate: A 3-year prospective study in France. Blood 99:3909–3915. Masala G, Di Lollo S, Picoco C, et al. 1996. Incidence rates of leukemias, lymphomas and myelomas in Italy: Geographic distribution and NHL histotypes. Int J Cancer 68:156–159. Mazzaro C, Zagonel V, Monfardini S, et al. 1996. Hepatitis C virus and nonHodgkin’s lymphomas. Br J Haematol 94:544–550.

916

PART IV: CANCER BY TISSUE OF ORIGIN

McDuffie HH, Pahwa P, McLaughlin JR, et al. 2001. Non-Hodgkin’s lymphoma and specific pesticide exposures in men: Cross-Canada study of pesticides and health. Cancer Epidemiol Biomarkers Prev 10:1155– 1163. McDuffie HH, Pahwa P, Spinelli JJ, et al. 2002. Canadian male farm residents, pesticide safety handling practices, exposure to animals and nonHodgkin’s lymphoma (NHL). Am J Ind Med Suppl 2:54–61. McLaughlin JK, Hrubec Z, Blot WJ, Fraumeni JF, Jr. 1995. Smoking and cancer mortality among U.S. veterans: A 26-year follow-up. Int J Cancer 60:190–193. McLaughlin JR, King WD, Anderson TW, Clarke EA, Ashmore JP. 1993. Paternal radiation exposure and leukaemia in offspring: The Ontario casecontrol study. BMJ 307:959–966. McMichael AJ, Giles GG. 1996. Have increases in solar ultraviolet exposure contributed to the rise in incidence of non-Hodgkin’s lymphoma? Br J Cancer 73:945–950. McNally RJ, Roman E, Cartwright RA. 1999. Leukemias and lymphomas: time trends in the UK, 1984–93. Cancer Causes Control 10:35–42. Medina KL, Strasser A, Kincade PW. 2000. Estrogen influences the differentiation, proliferation, and survival of early B-lineage precursors. Blood 95:2059–2067. Mele A, Pulsoni A, Bianco E, et al. 2003. Hepatitis C virus and B-cell nonHodgkin lymphomas: An Italian multicenter case-control study. Blood 102:996–999. Mellemkjaer L, Andersen V, Linet MS, Gridley G, Hoover R, Olsen JH. 1997. Non-Hodgkin’s lymphoma and other cancers among a cohort of patients with systemic lupus erythematosus. Arthritis Rheum 40:761– 768. Mellemkjaer L, Linet MS, Gridley G, Frisch M, Moller H, Olsen JH. 1996. Rheumatoid arthritis and cancer risk. Eur J Cancer 32A:1753–1757. Mellors RC. 1966. Autoimmune and immunoproliferative diseases of NZB/BI mice and hybrids. Int Rev Exp Pathol 5:217–252. Memon A, Doll R. 1994. A search for unknown blood-borne oncogenic viruses. Int J Cancer 58:366–368. Metayer C, Johnson ES, Rice JC. 1998. Nested case-control study of tumors of the hemopoietic and lymphatic systems among workers in the meat industry. Am J Epidemiol 147:727–738. Metayer C, Lynch CF, Clarke EA, et al. 2000. Second cancers among longterm survivors of Hodgkin’s disease diagnosed in childhood and adolescence. J Clin Oncol 18:2435–2443. Miligi L, Seniori CA, Crosignani P, et al. 1999. Occupational, environmental, and life-style factors associated with the risk of hematolymphopoietic malignancies in women. Am J Ind Med 36:60–69. Miller AB, Barclay TH, Choi NW, et al. 1980. A study of cancer, parity and age at first pregnancy. J Chronic Dis 33:595–605. Mills PK, Beeson WL, Fraser GE, Phillips RL. 1992. Allergy and cancer: Organ site-specific results from the Adventist Health Study. Am J Epidemiol 136:287–295. Minor P, Pipkin P, Jarzebek Z, Knowles W. 2003. Studies of neutralising antibodies to SV40 in human sera. J Med Virol 70:490–495. Misra RR, Pinsky PF, Srivastava S. 2000. Prognostic factors for hematologic cancers. Hematol Oncol Clin North Am 14:907–924. Mohan AK, Hauptmann M, Freedman DM, et al. 2003. Cancer and other causes of mortality among radiologic technologists in the United States. Int J Cancer 103:259–267. Moller P, Wallin H, Holst E, Knudsen LE. 2002. Sunlight-induced DNA damage in human mononuclear cells. FASEB J 16:45–53. Morgan RW, Kelsh MA, Zhao K, Heringer S. 1998. Mortality of aerospace workers exposed to trichloroethylene. Epidemiology 9:424–431. Morra E. 1999. The biological markers of non-Hodgkin’s lymphomas: Their role in diagnosis, prognostic assessment and therapeutic strategy. Int J Biol Markers 14:149–153. Morton LM, Holford TR, Leaderer B, et al. 2003. Alcohol use and risk of nonHodgkin’s lymphoma among Connecticut women (United States). Cancer Causes Control 14:687–694. Morton LM, Wang SS, Devesa SS, et al. 2005a. Lymphoma incidence patterns by WHO subtype in the United States, 1992–2001. Blood. [Epub ahead of print]. Morton LM, Hartge P, Holford T, et al. 2005b. Cigarette smoking and risk of non-Hodgkin lymphoma: a pooled analysis from the InterLymph consortium. Cancer Epi Bio Prev 14:925–933. Morton LM, Zheng T, Holford T, et al. 2005c. Alcohol consumption and risk of non-Hodgkin lymphoma: a pooled analysis from the InterLymph consortium. Lancet Oncol Jul;6(7):469–476. Mortreux F, Gabet AS, Wattel E. 2003. Molecular and cellular aspects of HTLV-1 associated leukemogenesis in vivo. Leukemia 17:26–38. Mueller N. 1999. Overview of the epidemiology of malignancy in immune deficiency. J Acquir Immune Defic Synd 21:S5–S10.

Mueller N, Mohar A, Evans A, et al. 1991. Epstein-Barr virus antibody patterns preceding the diagnosis of non-Hodgkin’s lymphoma. Int J Cancer 49:387–393. Mueller NE, Blattner WA. 1997. Retroviruses—Human T-cell lymphitropic virus. In: Evans AS, Kaslow RA, eds. Viral Infections of humans: Epidemiology and control. New York: Plenum Publishing Corporation, pp. 785–813. Nador RG, Cesarman E, Chadburn A, et al. 1996. Primary effusion lymphoma: A distinct clinicopathologic entity associated with the Kaposi’s sarcomaassociated herpes virus. Blood 88:645–656. Nagy N, Mattsson K, Maeda A, Liu A, Szekely L, Klein E. 2002. The X-linked lymphoproliferative disease gene product SAP is expressed in activated T and NK cells. Immunol Lett 82:141–147. Nakamura S, Yao T, Aoyagi K, Iida M, Fujishima M, Tsuneyoshi M. 1997. Helicobacter pylori and primary gastric lymphoma. A histopathologic and immunohistochemical analysis of 237 patients. Cancer 79:3–11. Nanni O, Amadori D, Lugaresi C, et al. 1996. Chronic lymphocytic leukaemias and non-Hodgkin’s lymphomas by histological type in farming-animal breeding workers: A population case-control study based on a priori exposure matrices. Occup Environ Med 53:652–657. National Cancer Institute. 2003. Surveillance research program. SEER*Stat, 2003 version 5.0.19. National Cancer Institute (NCI). 1982. National Cancer Institute sponsored study of classifications of non-Hodgkin’s lymphomas: Summary and description of a working formulation for clinical usage. The NonHodgkin’s Lymphoma Pathologic Classification Project. Cancer 49: 2112–2135. Nelson RA, Levine AM, Bernstein L. 2001. Reproductive factors and risk of intermediate- or high-grade B-Cell non-Hodgkin’s lymphoma in women. J Clin Oncol 19:1381–1387. Nelson RA, Levine AM, Marks G, Bernstein L. 1997. Alcohol, tobacco and recreational drug use and the risk of non-Hodgkin’s lymphoma. Br J Cancer 76:1532–1537. Neubauer A, Thiede C, Morgner A, et al. 1997. Cure of Helicobacter pylori infection and duration of remission of low-grade gastric mucosa-associated lymphoid tissue lymphoma. J Natl Cancer Inst 89:1350–1355. Newton R, Ferlay J, Beral V, Devesa SS. 1997. The epidemiology of nonHodgkin’s lymphoma: Comparison of nodal and extra-nodal sites. Int J Cancer 72:923–930. Niedobitek G, Meru N, Delecluse HJ. 2001. Epstein-Barr virus infection and human malignancies. Int J Exp Pathol 82:149–170. Nomura AM, Perez-Perez GI, Lee J, Stemmermann G, Blaser MJ. 2002. Relation between Helicobacter pylori cagA status and risk of peptic ulcer disease. Am J Epidemiol 155:1054–1059. Non-Hodgkin’s Lymphoma Classification Project. 1997. A clinical evaluation of the International Lymphoma Study Group classification of nonHodgkin’s lymphoma. Blood 89:3909–3918. O’Connor SR, Farmer PB, Lauder I. 1999. Benzene and non-Hodgkin’s lymphoma. J Pathol 189:448–453. Oksenhendler E, Boulanger E, Galicier L, et al. 2002. High incidence of Kaposi sarcoma-associated herpesvirus-related non-Hodgkin lymphoma in patients with HIV infection and multicentric Castleman disease. Blood 99:2331–2336. Ollier W. 2000. Rheumatoid arthritis and Epstein-Barr virus: A case of living with the enemy? Ann Rheum Dis 59:497–499. Olsen JH, Hahnemann JM, Borresen-Dale AL, et al. 2001. Cancer in patients with ataxia-telangiectasia and in their relatives in the Nordic countries. J Natl Cancer Inst 93:121–127. Olson JE, Cerhan JR, Janney CA, Anderson KE, Vachon CM, Sellers TA. 2002a. Postmenopausal cancer risk after self-reported endometriosis diagnosis in the lowa Women’s Health Study. Cancer 94:1612– 1618. Olson JE, Janney CA, Rao RD, et al. 2002b. The continuing increase in the incidence of primary central nervous system non-Hodgkin lymphoma: A surveillance, epidemiology, and end results analysis. Cancer 95: 1504–1510. Olsson H, Olsson ML, Ranstam J. 1990. Late age at first full-term pregnancy as a risk factor for women with malignant lymphoma. Neoplasma 37:185–190. Opelz G, Henderson R. 1993. Incidence of non-Hodgkin lymphoma in kidney and heart transplant recipients. Lancet 342:1514–1516. Opelz G, Schwarz V, Henderson R, Kneifel G, Ruhenstroth A. 1994. NonHodgkin’s lymphoma after kidney or heart transplantation: Frequency of occurrence during the first posttransplant year. Transpl Int 7 Suppl 1:S353–S356. Parker AS, Cerhan JR, Dick F, et al. 2000. Smoking and risk of non-Hodgkin lymphoma subtypes in a cohort of older women. Leuk Lymphoma 37:341–349.

Non-Hodgkin Lymphoma Parkin DM, Whelan S, Ferlay J, Teppo L, Thomas D. 2003. Cancer incidence in five continents, Vol. VIII. Lyon: IARC Press. Parsonnet J. 1998. Helicobacter pylori. Infect Dis Clin North Am 12:185–197. Parsonnet J, Hansen S, Rodriguez L, et al. 1994. Helicobacter pylori infection and gastric lymphoma. N Engl J Med 330:1267–1271. Penn I. 2000. Some problems with posttransplant lymphoproliferative disease. Transplantation 69:705–706. Penn I, Hammond W, Brettschneider L, Starzl TE. 1969. Malignant lymphomas in transplantation patients. Transplant Proc 1:106–112. Percy C, Van Holten V, Muir C. 1990. International Classification of Diseases for Oncology. Geneva: World Health Organization. Persson B, Fredrikson M. 1999. Some risk factors for non-Hodgkin’s lymphoma. Int J Occup Med Environ Health 12:135–142. Pettersson T, Rosenlof K, Friman C, Mickos A, Teppo AM, Fyhrquist F. 1993. Successful treatment of the anemia of rheumatoid arthritis with subcutaneously administered recombinant human erythropoietin. Slower response in patients with more severe inflammation. Scand J Rheumatol 22:188–193. Pioltelli P, Gargantini L, Cassi E, et al. 2000. Hepatitis C virus in nonHodgkin’s lymphoma. A reappraisal after a prospective case-control study of 300 patients. Lombart Study Group of HCV-Lymphoma. Am J Hematol 64:95–100. Poiesz BJ, Papsidero LD, Ehrlich G, et al. 2001. Prevalence of HTLV-Iassociated T-cell lymphoma. Am J Hematol 66:32–38. Pottern LM, Linet M, Blair A, et al. 1991. Familial cancers associated with subtypes of leukemia and non-Hodgkin’s lymphoma. Leuk Res 15:305–314. Preston DL, Kusumi S, Tomonaga M, et al. 1994. Cancer incidence in atomic bomb survivors. Part III. Leukemia, lymphoma and multiple myeloma, 1950–1987. Radiat Res 137:S68–S97. Quinn ER, Chan CH, Hadlock KG, Foung SK, Flint M, Levy S. 2001. The B-cell receptor of a hepatitis C virus (HCV)-associated non-Hodgkin lymphoma binds the viral E2 envelope protein, implicating HCV in lymphomagenesis. Blood 98:3745–3749. Raaschou-Nielsen O, Hertel O, Thomsen BL, Olsen JH. 2001. Air pollution from traffic at the residence of children with cancer. Am J Epidemiol 153:433–443. Rao PH, Houldsworth J, Dyomina K, et al. 1998. Chromosomal and gene amplification in diffuse large B-cell lymphoma. Blood 92:234–240. Raphael M, Borisch B, Jaffe E. 2001. Lymphomas associated with infection by the human immune deficiency virus (HIV). In: Jaffe E, Harris NL, Stein H, Vardiman J, eds. Pathology and Genetics of Tumours of Haematopioetic and Lymphoid Tissues. Lyon, France: IARC Press, pp. 260–263. Rappaport H. 1966. Tumors of the hematopoietic system. Atlas of Tumor Pathology. Washington, DC: Armed Forces Institute of Pathology. Ridolfo AL, Santambrogio S, Mainini F, et al. 1996. High frequency of nonHodgkin’s lymphoma in patients with HIV-associated Kaposi’s sarcoma. AIDS 10:181–185. Ries LA, Eisner MP, Kosary CL, et al. 2003. SEER Cancer Statistics Review 1975–2000. SEER Cancer Statistics Review 1975–2000. Rizzo P, Carbone M, Fisher SG, et al. 1999. Simian virus 40 is present in most United States human mesotheliomas, but it is rarely present in nonHodgkin’s lymphoma. Chest 116:470S–473S. Ron E, Boice JD, Jr., Hamburger S, Stovall M. 1994. Mortality following radiation treatment for infertility of hormonal origin or amenorrhoea. Int J Epidemiol 23:1165–1173. Ron E, Modan B, Boice JD, Jr. 1988. Mortality after radiotherapy for ringworm of the scalp. Am J Epidemiol 127:713–725. Rothman N, Cantor KP, Blair A, et al. 1997. A nested case-control study of non-Hodgkin lymphoma and serum organochlorine residues. Lancet 350:240–244. Rothman N, Skibola C, Wang SS, et al. 2006. Genetic variation in TNF and IL10 and risk of non-Hodgkin lymphoma: a report from the InterLymph Consortium. Lancet Oncol 7:27–38. Ruskone-Fourmestraux A, Lavergne A, Aegerter PH, et al. 2001. Predictive factors for regression of gastric MALT lymphoma after anti-Helicobacter pylori treatment. Gut 48:297–303. Savitz DA, Andrews KW. 1997. Review of epidemiologic evidence on benzene and lymphatic and hematopoietic cancers. Am J Ind Med 31:287–295. Savitz DA, Wachtel H, Barnes FA, John EM, Tvrdik JG. 1988. Case-control study of childhood cancer and exposure to 60-Hz magnetic fields. Am J Epidemiol 128:21–38. Scherr PA, Mueller NE. 1996. Non-Hodgkin’s lymphoma. In: Schottenfeld D, Fraumeni JF, eds. Cancer Epidemiology and Prevention 2nd edition. New York: Oxford University Press, pp. 920–945. Schiff D, Suman VJ, Yang P, Rocca WA, O’Neill BP. 1998. Risk factors for primary central nervous system lymphoma: A case-control study. Cancer 82:975–982.

917

Schroeder JC, Olshan AF, Dent RB, et al. 2002. A case-control study of tobacco use and other non-occupational risk factors for t(14; 18) subtypes of non-Hodgkin’s lymphoma (United States). Cancer Causes Control 13:159–168. Schroeder JC, Savitz DA. 1997. Lymphoma and multiple myeloma mortality in relation to magnetic field exposure among electric utility workers. Am J Ind Med 32:392–402. Schulz TF. 2000. Kaposi’s sarcoma-associated herpesvirus (human herpesvirus 8): Epidemiology and pathogenesis. J Antimicrob Chemother 45 Suppl T3:15–27. SEER*STAT. 2003. NCI software package to access SEER public-use file: 1973–2000. Available at: http://www.seer.cancer.gov/. Settimi L, Comba P, Bosia S, et al. 2001. Cancer risk among male farmers: A multi-site case-control study. Int J Occup Med Environ Health 14:339–347. Shah KV. 2000. Does SV40 infection contribute to the development of human cancers? Rev Med Virol 10:31–43. Shimizu Y, Kato H, Schull WJ. 1990. Studies of the mortality of A-bomb survivors. 9. Mortality, 1950–1985: Part 2. Cancer mortality based on the recently revised doses (DS86). Radiat Res 121:120–141. Shiramizu B, Herndier BG, McGrath MS. 1994. Identification of a common clonal human immunodeficiency virus integration site in human immunodeficiency virus-associated lymphomas. Cancer Res 54:2069–2072. Shivapurkar N, Harada K, Reddy J, et al. 2002. Presence of simian virus 40 DNA sequences in human lymphomas. Lancet 359:851–852. Siegel R, Gartenhaus R, Kuzel T. 2001. HTLV-I associated leukemia/ lymphoma: Epidemiology, biology, and treatment. Cancer Treat Res 104:75–88. Silvestri F, Barillari G, Fanin R, et al. 1997. The genotype of the hepatitis C virus in patients with HCV-related B cell non-Hodgkin’s lymphoma. Leukemia 11:2157–2161. Silvestri F, Pipan C, Barillari G, et al. 1996. Prevalence of hepatitis C virus infection in patients with lymphoproliferative disorders. Blood 87:4296–4301. Skibola CF, Bracci PM, Paynter RA, et al. 2005a. Polymorphisms and haplotypes in the cytochrome P450 17A1, prolactin, and catechol-Omethyltransferase genes and non-Hodgkin lymphoma risk. Cancer Epidemiol Biomarkers Prev 14:2391–2401. Skibola CF, Forrest MS, Coppede F, et al. 2005b. Polymorphisms and haplotypes in folate-metabolizing genes and risk of non-Hodgkin lymphoma. Blood 105:906–907. Skibola CF, Holly EA, Forrest MS, et al. 2004. Body mass index, leptin and leptin receptor polymorphisms and non-Hodgkin lymphoma. Cancer Epidemiol Biomarkers Prev. Cancer Epi Bio Prev 13:779–786. Skibola DR, Smith MT, Bracci PM, et al. 2005. Polymorphisms in ghrelin and neuropeptide Y genes are associated with non-Hodgkin lymphoma. 2005. Cancer Epidemiol Biomarkers Prev 14:1251–1256. Smedby KE, Hjalgrim H, Melbye M, et al. 2005. Ultraviolet radiation exposure and risk of malignant lymphomas. J Natl Cancer Inst Feb 2;97:199–209. Smets F, Latinne D, Bazin H, et al. 2002. Ratio between Epstein-Barr viral load and anti-Epstein-Barr virus specific T-cell response as a predictive marker of posttransplant lymphoproliferative disease. Transplantation 73:1603–1610. Stagnaro E, Ramazzotti V, Crosignani P, et al. 2001. Smoking and hematolymphopoietic malignancies. Cancer Causes Control 12:325–334. Staudt LM. 2002. Gene expression profiling of lymphoid malignancies. Annu Rev Med 53:303–318. Staudt LM. 2003. Molecular diagnosis of the hematologic cancers. N Engl J Med 348:1777–1785. Strickler HD, Rosenberg PS, Devesa SS, Hertel J, Fraumeni JF, Jr., Goedert JJ. 1998. Contamination of poliovirus vaccines with simian virus 40 (1955–1963) and subsequent cancer rates. JAMA 279:292–295. Sullivan CS, Pipas JM. 2002. T antigens of simian virus 40: Molecular chaperones for viral replication and tumorigenesis. Microbiol Mol Biol Rev 66:179–202. Svensson BG, Mikoczy Z, Stromberg U, Hagmar L. 1995. Mortality and cancer incidence among Swedish fishermen with a high dietary intake of persistent organochlorine compounds. Scand J Work Environ Health 21:106–115. Sweet BH, Hilleman MR. 1960. The vacuolating virus SV40. Proc Soc Exp Biol Med 1960:1–12. Swerdlow AJ, Higgins CD, Hunt BJ, et al. 2000. Risk of lymphoid neoplasia after cardiothoracic transplantation. A cohort study of the relation to Epstein-Barr virus. Transplantation 69:897–904. Swinnen LJ. 2000. Transplantation-related lymphoproliferative disorder: A model for human immunodeficiency virus-related lymphomas. Semin Oncol 27:402–408.

918

PART IV: CANCER BY TISSUE OF ORIGIN

Swinnen LJ, Costanzo-Nordin MR, Fisher SG, et al. 1990. Increased incidence of lymphoproliferative disorder after immunosuppression with the monoclonal antibody OKT3 in cardiac-transplant recipients. N Engl J Med 323:1723–1728. Tam HK, Zhang ZF, Jacobson LP, et al. 2002. Effect of highly active antiretroviral therapy on survival among HIV-infected men with Kaposi sarcoma or non-Hodgkin lymphoma. Int J Cancer 98:916–922. Tavani A, Gallus S, La Vecchia C, Franceschi S. 2001. Alcohol drinking and risk of non-Hodgkin’s lymphoma. Eur J Clin Nutr 55:824–826. Tavani A, Negri E, Franceschi S, Serraino D, La Vecchia C. 1994a. Smoking habits and non-Hodgkin’s lymphoma: A case-control study in northern Italy. Prev Med 23:447–452. Tavani A, Negri E, Franceschi S, Talamini R, La Vecchia C. 1994b. Coffee consumption and risk of non-Hodgkin’s lymphoma. Eur J Cancer Prev 3:351–356. Tavani A, Pregnolato A, La Vecchia C, Franceschi S. 1997. A case-control study of reproductive factors and risk of lymphomas and myelomas. Leuk Res 21:885–888. Tiirikainen MI, Mullaney BP, Holly EA, Pallavicini MG, Jensen RH. 2001. DNA copy number alterations in HIV-positive and HIV-negative patients with diffuse large-cell lymphomas. J Acquir Immune Defic Syndr 27: 272–276. Ullrich A, Fischbach W, Blettner M. 2002. Incidence of gastric B-cell lymphomas: A population-based study in Germany. Ann Oncol 13: 1120–1127. Urquhart JD, Black RJ, Muirhead MJ, et al. 1991. Case-control study of leukaemia and non-Hodgkin’s lymphoma in children in Caithness near the Dounreay nuclear installation. BMJ 302:687–692. Valesini G, Priori R, Bavoillot D, et al. 1997. Differential risk of non-Hodgkin’s lymphoma in Italian patients with primary Sjogren’s syndrome. J Rheumatol 24:2376–2380. Vallisa D, Berte R, Rocca A, et al. 1999. Association between hepatitis C virus and non-Hodgkin’s lymphoma, and effects of viral infection on histologic subtype and clinical course. Am J Med 106:556–560. Van Wijngaarden E, Savitz DA. 2001. Occupational sunlight exposure and mortality from non-Hodgkin lymphoma among electric utility workers. J Occup Environ Med 43:548–553. Vega F, Medeiros LJ. 2003. Chromosomal translocations involved in nonHodgkin lymphomas. Arch Pathol Lab Med 127:1148–1160. Vilchez RA, Kozinetz CA, Arrington AS, Madden CR, Butel JS. 2003. Simian virus 40 in human cancers. Am J Med 114:675–684. Vilchez RA, Madden CR, Kozinetz CA, et al. 2002. Association between simian virus 40 and non-Hodgkin lymphoma. Lancet 359:817–823. Vineis P, Crosignani P, Sacerdote C, et al. 1999. Hematopoietic cancer and peptic ulcer: A multicenter case-control study. Carcinogenesis 20: 1459–1463. Vineis P, Crosignani P, Sacerdote C, et al. 2000. Haematopoietic cancer and medical history: A multicentre case control study. J Epidemiol Community Health 54:431–436. Waddell BL, Zahm SH, Baris D, et al. 2001. Agricultural use of organophosphate pesticides and the risk of non-Hodgkin’s lymphoma among male farmers (United States). Cancer Causes Control 12:509–517. Wang JX, Zhang LA, Li BX, et al. 2002. Cancer incidence and risk estimation among medical x-ray workers in China, 1950–1995. Health Phys 82: 455–466. Ward MH, Mark SD, Cantor KP, Weisenburger DD, Correa-Villasenor A, Zahm SH. 1996. Drinking water nitrate and the risk of non-Hodgkin’s lymphoma. Epidemiology 7:465–471. Ward MH, Zahm SH, Weisenburger DD, et al. 1994. Dietary factors and nonHodgkin’s lymphoma in Nebraska (United States). Cancer Causes Control 5:422–432. Waterhouse D, Carman WJ, Schottenfeld D, Gridley G, McLean S. 1996. Cancer incidence in the rural community of Tecumseh, Michigan: A pattern of increased lymphopoietic neoplasms. Cancer 77:763– 770. Weiss HA, Darby SC, Doll R. 1994. Cancer mortality following X-ray treatment for ankylosing spondylitis. Int J Cancer 59:327–338. Weng WK, Levy S. 2003. Hepatitis C virus (HCV) and lymphomagenesis. Leuk Lymphoma 44:1113–1120. Weyer PJ, Cerhan JR, Kross BC, et al. 2001. Municipal drinking water nitrate level and cancer risk in older women: The lowa Women’s Health Study. Epidemiology 12:327–338. Wilks R, Hanchard B, Morgan O, et al. 1996. Patterns of HTLV-I infection among family members of patients with adult T-cell leukemia/lymphoma and HTLV-I associated myelopathy/tropical spastic paraparesis. Int J Cancer 65:272–273.

Wingle DT, Semenciw RM, Wilkins K, et al. 1990. Mortality study of Canadian male farm operators: Non-Hodgkin’s lymphoma mortality and agricultural practices in Saskatchewan. J Natl Cancer Inst 82: 575–582. Wolk A, Gridley G, Svensson M, et al. 2001. A prospective study of obesity and cancer risk (Sweden). Cancer Causes Control 12:13–21. Wong O, Raabe GK. 2000. Non-Hodgkin’s lymphoma and exposure to benzene in a multinational cohort of more than 308,000 petroleum workers, 1937 to 1996. J Occup Environ Med 42:554–568. Wotherspoon AC, Dogan A, Du MQ. 2002. Mucosa-associated lymphoid tissue lymphoma. Curr Opin Hematol 9:50–55. Wotherspoon AC, Doglioni C, Diss TC, et al. 1993. Regression of primary lowgrade B-cell gastric lymphoma of mucosa-associated lymphoid tissue type after eradication of Helicobacter pylori. Lancet 342:575–577. Wotherspoon AC, Ortiz-Hidalgo C, Falzon MR, Isaacson PG. 1991. Helicobacter pylori-associated gastritis and primary B-cell gastric lymphoma. Lancet 338:1175–1176. Wundisch T, Kim TD, Thiede C, et al. 2003. Etiology and therapy of Helicobacter pylori-associated gastric lymphomas. Ann Hematol 82:535–545. Zahm SH. 1997. Mortality study of pesticide applicators and other employees of a lawn care service company. J Occup Environ Med 39:1055–1067. Zahm SH, Hoffman-Goetz L, Dosemeci M, Cantor KP, Blair A. 1999. Occupational physical activity and non-Hodgkin’s lymphoma. Med Sci Sports Exerc 31:566–571. Zahm SH, Ward MH, Blair A. 1997a. Pesticides and cancer. Occup Med 12:269–289. Zahm SH, Weisenburger DD, Babbitt PA, Saal RC, Vaught JB, Blair A. 1990. A case-control study of non-Hodgkin’s lymphoma and the herbicide 2,4dichlorophenoxyacetic acid (2,4-D) in eastern Nebraska. Epidemiology 1:349–356. Zahm SH, Weisenburger DD, Babbitt PA, Saal RC, Vaught JB, Blair A. 1992. Use of hair coloring products and the risk of lymphoma, multiple myeloma, and chronic lymphocytic leukemia. Am J Public Health 82:990–997. Zahm SH, Weisenburger DD, Holmes FF, Cantor KP, Blair A. 1997b. Tobacco and non-Hodgkin’s lymphoma: Combined analysis of three case-control studies (United States). Cancer Causes Control 8:159–166. Zhang S, Hunter DJ, Rosner BA, et al. 1999. Dietary fat and protein in relation to risk of non-Hodgkin’s lymphoma among women. J Natl Cancer Inst 91:1751–1758. Zhang SM, Hunter DJ, Rosner BA, et al. 2000. Intakes of fruits, vegetables, and related nutrients and the risk of non-Hodgkin’s lymphoma among women. Cancer Epidemiol Biomarkers Prev 9:477–485. Zheng T, Zahm SH, Cantor KP, Weisenburger DD, Zhang Y, Blair A. 2001. Agricultural exposure to carbamate pesticides and risk of non-Hodgkin lymphoma. J Occup Environ Med 43:641–649. Zhu K, Levine RS, Brann EA, et al. 2001. Risk factors for non-Hodgkin’s lymphoma according to family history of haematolymphoproliferative malignancies. Int J Epidemiol 30:818–824. Zhu K, Levine RS, Gu Y, et al. 1998. Non-Hodgkin’s lymphoma and family history of malignant tumors in a case-control study (United States). Cancer Causes Control 9:77–82. Ziegler JL, Beckstead JA, Volberding PA, et al. 1984. Non-Hodgkin’s lymphoma in 90 homosexual men. Relation to generalized lymphadenopathy and the acquired immunodeficiency syndrome. N Engl J Med 311: 565–570. Ziegler JL, Drew WL, Miner RC, et al. 1982. Outbreak of Burkitt’s-like lymphoma in homosexual men. Lancet 2:631–633. Zignego AL, Ferri C, Giannelli F, et al. 2002. Prevalence of bcl-2 rearrangement in patients with hepatitis C virus-related mixed cryoglobulinemia with or without B-cell lymphomas. Ann Intern Med 137:571–580. Zignego AL, Ferri C, Giannini C, et al. 1997. Hepatitis C virus infection in mixed cryoglobulinemia and B-cell non-Hodgkin’s lymphoma: Evidence for a pathogenetic role. Arch Virol 142:545–555. Zucca E, Roggero E, Maggi-Solca N, et al. 2000. Prevalence of Helicobacter pylori and hepatitis C virus infections among non-Hodgkin’s lymphoma patients in Southern Switzerland. Haematologica 85:147–153. Zuckerman E, Zuckerman T, Levine AM, et al. 1997. Hepatitis C virus infection in patients with B-cell non-Hodgkin lymphoma. Ann Intern Med 127:423–428. Zuckerman E, Zuckerman T, Sahar D, et al. 2001. bcl-2 and immunoglobulin gene rearrangement in patients with hepatitis C virus infection. Br J Haematol 112:364–369. zur Hausen H, Schulte-Holthausen H, Klein G, et al. 1970. EBV DNA in biopsies of Burkitt tumours and anaplastic carcinomas of the nasopharynx. Nature 228:1056–1058.

47

Multiple Myeloma ANNECLAIRE J. DE ROOS, DALSU BARIS, NOEL S. WEISS, AND LISA J. HERRINTON

P

lasma cells, the final products of B-cell differentiation, synthesize and release immunoglobulin (Ig) and Ig subunits (light and heavy chains). Plasma cell malignancies, which are characterized by the presence of an elevated number of plasma cells in bone marrow and, very often, elevated levels of monoclonal protein in serum and urine, include the following: myeloma, in which there is production of IgA, IgD, IgE, IgG, or light chains; Waldenström macroglobulinemia, in which there is production of IgM; and the heavy chain diseases, in which there is production of the heavy chains (gamma, mu, and delta) (Osserman et al., 1987). Because they occur relatively infrequently, neither descriptive nor analytical studies of Waldenstrom macroglobulinemia or the heavy chain diseases are reported in this review. Myeloma presently accounts for almost 10% of all hematologic malignancies and 1% of cancer deaths in Western countries (Kastrinakis et al., 2000). Although myeloma is a rare malignancy, there is relatively high mortality, with a 5-year survival of 28% (Dalton et al., 2001). Nevertheless, high-dose treatment followed by single or double autologous stem cell transplantation and novel targeted therapies have improved disease control over the past decade (Barlogie et al., 2004; Kyle and Rajkumar, 2004; Fassas et al., 2005). Recent advances in the understanding of molecular phenotypes of the disease, especially by use of gene expression profiling, may prove useful in developing more effective and targeted therapies (Barille-Nion et al., 2003; Fassas et al., 2005).

CLASSIFICATION It is likely that myeloma develops through a multistep transformation process (Hallek et al., 1998). There is some evidence that the initial transformations causing myeloma occur at the early B cell or lymphoid stem cell phase (Barlogie et al., 1989). B-cell precursors express specific antigens during discrete stages of maturation, and early B-cell and T-cell antigens have been identified from tumor cells (Barlogie et al., 1989). Furthermore, cytogenetic abnormalities of transformed plasma cells from one myeloma patient were also evident in the patient’s early B cells (MacKenzie and Lewis., 1985). Nevertheless, similarities between myeloma and plasma cell phenotypes suggest that the immortalizing oncogenic events occur after or do not interfere with the normal plasma cell maturation process (Hallek et al., 1998). The malignantly transformed B-cell clone proliferates and accumulates in the bone marrow. Progression of the malignancy is then steered by various genetic events and factors regulating the growth of plasma cells. The manifestations of myeloma are variable and the disease can be difficult to diagnose. Excluding nonsecretory myeloma, the diagnosis is based on the presence of monoclonal protein in serum or urine and at least 10% atypical plasma cells in bone marrow (Angtuaco et al., 2004) with additional symptoms, which may include osteolytic lesions, renal insufficiency, hypercalcemia, anemia, and increased susceptibility to infections. Agarose gel electrophoresis is the preferred method to screen for monoclonal protein, with followup by immunofixation for confirmation and to determine the heavy and light chain types (International Myeloma Working Group, 2003). Nearly all myeloma cell lines harbor chromosome translocations involving the IgH locus (Avet-Loiseau et al., 1998; Fonseca et al., 2002b; Kuehl and Bergsagel, 2002), of which t (11; 14) (q13; q32) is

the most common, resulting in upregulation of the cyclin D1 oncogene (Fonseca et al., 2002b). Patients with this translocation exhibit an aggressive clinical course (Kastrinakis et al., 2000). Other common translocations involve several loci, including 4p16.3, 16q23, 6p21, and 6p25 generally resulting in dysregulation of putative oncogenes including 6p25 FGFR3 and MMSET (on 4p16.3), c-maf (on 16q23) cyclin D3 (on 6p21), and MUM-1 (on 6p25) (Fonseca et al., 2002b; Barille-Nion et al., 2003). Deletion of chromosome 13q is found in the majority of plasma cells in myeloma (Avet-Loiseau et al., 1999), and has been shown to be strongly associated with translocation t (4;14) (p16.3;q32) (Fonseca et al., 2001). Research into the role of methylation in myeloma risk may provide some clues into chromosomal alterations unique to myeloma etiology. Methylation of chromosome p16 was common in a group of myeloma patients (42%), and was associated with shorter overall and progression-free survival, primarily due to its association with an increased proliferative rate of plasma cells (Mateos et al., 2002). Another study found high prevalences of both p16 (75%) and p15 (65%) methylation in a small group of myeloma patients (Ng et al., 1997). Silencing of the p73 tumor suppressor gene by hypermethylation has also been observed in myeloma and other lymphoid neoplasms (Katusic et al., 1985).

INCIDENCE AND MORTALITY Demographic Patterns Internationally, the reported incidence of myeloma varies substantially, as shown in Table 47–1 for the years 1993 to 1995. The availability and use of serum protein electrophoresis, immunoelectrophoresis, and immunofixation, all relatively sensitive diagnostic methods, vary by location. This variation probably accounts for part of the geographic variability in the reported incidence of and mortality from the disease. The highest incidence rates have been reported for blacks; Europeans and North American Caucasians have intermediate rates; whereas generally low rates have been reported for Asians living in Asia and the United States. A nationwide evaluation of US myeloma mortality found higher rates in urban vs. rural areas (Blattner et al., 1981), as did a study of myeloma incidence in Israel (Shapira and Carter, 1986), but a comprehensive review of studies from Europe, Japan, and the United States concluded that the urban and rural rates were similar (Doll, 1991). Age-, sex-, and race-specific incidence rates from the US Surveillance, Epidemiology, and End Results (SEER) program are shown in Figure 47–1 for the time period 1975 to 1999, and corresponding graphs for mortality rates are shown in Figure 47–2 for 1970 to 1999. Myeloma incidence and mortality increased steeply with age during these time periods. Men had higher rates than women; this sex difference has been observed consistently in international comparisons (Cartwright et al., 2002; Levi et al., 1992). Blacks had higher rates than whites; the cumulative incidence (ages 0–74 years) in black men during the period from 1973 to 1990 was approximately 10 per 100,000; in black women, white men, and white women, the corresponding rates were seven, four, and three per 100,000, respectively (Miller et al., 1993). Myeloma incidence among Asian Americans was lower still (Table 47–1). Racial differences in myeloma rates have also been observed internationally. British migrants from West Africa, East

919

Table 47–1. Annual Age-standardizeda incidence (per 100,000) Rates by Geographic Location (1993–1995) Location

Race/Ethnicity

Men

Women

Location

america, central and south

Ethnicity

Men

Women

2.4–2.7 1.4 3.4–4.3 2.1 2.0 3.3 1.7 2.9 2.3–3.9 3.0 2.4 4.2 2.7–6.2 1.7 2.1 3.3–4.1 3.5 2.0–2.6 1.2 2.4 2.2 2.2–3.6 3.6 3.0–4.9 2.7–4.1 1.1

2.3–2.9 1.1 2.8–2.9 1.6 1.3 2.1 1.1 2.5 1.7–3.3 2.5 3.3 2.7 2.0–3.7 1.5 1.8 2.3–2.4 2.1 1.2–1.6 1.3 2.2 1.8 1.7–3.9 2.4 1.6–3.0 2.0–2.9 0.8

2.1–4.0 4.2 3.9 4.7 4.7 1.3

1.9–3.2 3.0 1.4 3.8 3.8 0.8

europe

Argentina Brazil Colombia, Cali Costa-Rica Cuba, Villa Clara Ecuador, Quito France, Martinique USA, Puerto Rico Uruguay, Montevideo

3.0 1.1–1.9 1.4 1.5 1.2 1.6 2.3 3.3 1.3

2.1 0.9–1.5 1.1 1.3 1.4 1.9 1.7 2.0 1.9

4.0 3.9–4.0 4.2–5.6 7.1–7.4 1.7 3.3 2.9 2.0 3.9 8.8

2.7 2.5–2.6 2.1–3.3 6.4–7.0 0.7 2.3 0.8 1.7 2.5 6.8

0.2–1.9 0.7–1.9 3.3 2.7 3.2 3.3 2.7 1.5–2.1 0.9–1.4 2.4 2.1 0.7–1.0 1.1–2.4 0.4–0.7

0.2–1.3 0.5–1.1 2.5 2.4 2.1 2.9 1.9 1.0–1.3 0.7–1.1 0.8 0.5 0.9–1.1 0.9–1.9 0.3–0.5

Austria Belarus Belgium Croatia Czech Republic Denmark Estonia Finland France Germany, Saarland Iceland Ireland Italy Latvia Lithuania The Netherlands Norway Poland Russia, St Petersburg Slovakia Slovenia Spain Sweden Switzerland United Kingdom Yugoslavia

america, north Canada USA, California

Non-Hispanic White Hispanic White Black Chinese Filipino Japanese Korean White Black

USA, SEER

asia China India Israel

Jews Jews born in Israel Jews born in Europe or America Jews born in Africa or Asia Non-Jews

Japan Korea Kuwait

Kuwaitis Non-Kuwaitis

Philippines Singapore Thailand

oceania Australia New Zealand USA, Hawaii

White Filipino Hawaiian Japanese

Source: Adapted from Parkin et al., 2002. Complete citation: Cancer Incidence in Five Continents, Volume VIII, Eds: D.M. Parkin, S.L. Whelan, J. Ferlay, L. Teppo and D.B. Thomas. IARC Scientific Publications No. 155, Lyon, France, 2002. a Standardized to the age distribution of the world population, ages 0–74.

100

Rate per 100,000 person-years

Rate per 100,000 person-years

100

10

1

10

1

Black male Black female White male White female 0.1

0.1 0

20

40 60 Age at diagnosis

80

100

Figure 47–1. Age-specific multiple myeloma incidence rates in the U.S. (SEER areas) by race and sex, 1975–1999.

920

Black male Black female White male White female

0

20

40

60

80

100

Age at death

Figure 47–2. Age-specific multiple myeloma mortality rates in the U.S. by race and sex, 1970–1999.

921

Multiple Myeloma Africa, and the Caribbean had higher rates of mortality from myeloma than native-born British residents between 1970 and 1985 (relative risk [RR] Caribbean males, RR = 2.2, 95% CI:1.7–2.7; Caribbean females, RR = 2.0, 95% CI:1.5–2.7; East African males, RR = 1.9, 95% CI:1.0–3.7; East African females, RR = 1.7, 95% CI:0.7–4.1); West African males RR = 4.1, 95% CI:2.2–7.6; West African females, RR = 1.6; 95% CI:0.2–11.5) (Grulich et al., 1992). An epidemiologic study of healthy black and white Americans observed that the ratio of B to T cells was higher among blacks, and that blacks had higher proportions of HLA-DR–positive cells and activated T cells (Tollerud et al., 1991); the investigators hypothesized that these differences may be related to the observed difference in myeloma incidence between these two groups.

20

Black Female

Rate per 100,000 person-years

White Male

Time Trends

White Female

1

0.2 1970

1980

1990

2000

Year of diagnosis

Figure 47–3. Age-adjusted (1970 U.S. standard) multiple myeloma incidence trends in the U.S. (SEER areas) by race and sex, 1975–1979 to 1995–99.

20

10

Rate per 100,000 person-years

Trends in reported rates of myeloma incidence and mortality over time are likely to be misleading for a number of reasons related to improvements in case ascertainment and reporting (Velez et al., 1982). The availability and use of serum protein electrophoresis, immunoelectrophoresis, and immunofixation, all relatively sensitive diagnostic methods, have increased over time, likely resulting in more widespread and accurate identification of cases in later years. A unique International Classification of Diseases (ICD) code for multiple myeloma was first established in 1948, resulting in changes in reporting of the disease. Other factors influencing rates are changes in diagnostic criteria for myeloma, autopsy rates, and increased access to medical care across social and racial groups over time. Increases in reported myeloma incidence in the United States have been observed for time periods from the mid 1900s through the 1970s, with little change in later years. Secular changes in myeloma incidence among white residents of four geographic areas of the United States who had participated in national cancer surveys or had established tumor registries (Atlanta, Connecticut, Detroit, San Francisco/ Oakland) were examined by Devesa et al. for the period of 1947 to 1984 (Devesa et al., 1987). The reported annual incidence rates increased between 1947 and 1975 by about 150% to 3.8 per 100,000 in men and 2.6 per 100,000 in women (1950 US standard). However, no significant increases were observed between 1975 and 1984. The age-adjusted incidence of myeloma among blacks also did not change appreciably from 1973 to 1990 (Miller et al., 1993). Race- and sexspecific, age-adjusted, incidence rates from the US SEER program for the years 1975 to 1999 are shown in Figure 47–3, and corresponding graphs for mortality rates for 1950 to 1999 are shown in Figure 47–4. Both incidence and mortality rates appear to have increased slightly in each race- and sex- specific group between 1975 and 1990, leveling off in the most recent years after 1995, except for a continued increase in mortality among non-white females. There have also been reports of increasing incidence through the 1970s in Denmark (Hansen et al., 1989), Australia (Nandakumar et al., 1988), New Zealand (Pearce et al., 1985), and Israel (Shapira and Carter, 1986). In Denmark, the reported annual incidence increased between 1943 and 1962 from 1.3 to 3.3 per 100,000 in men and from 1.2 to 2.5 per 100,000 in women (European standard); however, no increase was observed between 1963 and 1982. Increases in myeloma incidence in northeast Scotland were reported for a later time period, from 1973 to 1987; however, the observation that the increase was observed for registries with the lowest initial rates, but not among those with higher initial rates, argues for the role of improved surveillance in the increase (Soutar et al., 1996). Three studies of time trends in myeloma incidence were conducted in areas where a particularly high level of case ascertainment would have been expected for the entire surveillance period: one in Olmsted County, Minnesota, where the Mayo Clinic is located and all major diagnoses are documented in a records-linkage system; another in Malmö, Sweden, where the relatively stable population has had access to comprehensive medical care since the 1950s; and the other in the Canton of Vaud, Switzerland, where close surveillance of incident cancers has occurred since the 1970s. In Olmsted County, the annual age-standardized rate (1950 US standard) of myeloma increased from

Black Male

10

1

Black male Non-white male Black female Non-white female White male White female 0.2 1950

1960

1970

1980

1990

2000

Year of death

Figure 47–4. Age-adjusted (1970 U.S. standard) multiple myeloma mortality trends in the U.S. by race and sex, 1950–1954 to 1995–1999.

3.3 per 100,000 during 1945 through 1964, to 4.1 per 100,000 in 1978 through 1990, but the difference was not appreciable (Kyle et al., 1994). Incidence rates over the 46-year time period from 1945 through 1990 were approximately 4.5 per 100,000 among men and 2.4 per 100,000 among women (Kyle et al., 1994). In Malmö, the incidence rate in men increased by 60% between 1950 and 1979 to an annual rate of 4.6 per 100,000 (1950 US standard) and all age groups were affected (Turesson et al., 1984). In contrast, no increase was observed in the rate among women (average annual incidence

922

PART IV: CANCER BY TISSUE OF ORIGIN

was 2.7 per 100,000) (Turesson et al., 1984). In the Canton of Vaud, no changes in incidence were noted between 1978 and 1987, with average annual incidence rates of 4.8 per 100,000 in men and 2.7 per 100,000 in women, European standard) (Levi and La Vecchia, 1990). Several lines of evidence suggest that the reported secular increase in myeloma incidence was predominantly the result of changes in ascertainment of the disease, including the following: 1. There was no appreciable increase in Olmsted County or the Canton of Vaud, areas with good surveillance for myeloma 2. In countries observing increases, rates generally stabilized since the 1970s 3. When reported, increases were greatest among the elderly, a group in whom a laboratory diagnosis might have less commonly been sought in the past than in the present. Thus, changes in case ascertainment, including the availability and use of serum protein electrophoresis certainly account for much of the reported increases in incidence and mortality in some locations. Exceptions such as the finding of an increase in myeloma incidence in men, but not women, of all ages who resided in Malmö is consistent with the possible introduction of an occupational agent causing myeloma (Turesson et al., 1984).

Socioeconomic Status Positive associations of socioeconomic status indicators with myeloma were reported in several studies of myeloma mortality (Blattner et al., 1981; Cuzick et al., 1983; MacMahon, 1966; Velez et al., 1982). In a hospital-based case-control study conducted in North Carolina during the period from 1976 to 1982, there was a 60% increased myeloma incidence associated with home ownership (OR = 1.6, 95% CI:1.0–2.6), and there was a suggested trend with increasing occupational rank; however, family income and education were unrelated to risk of the disease (Johnston et al., 1985). Conversely, a population-based case-control study found inverse associations between myeloma risk and occupation-based socioeconomic status, income, and education among both black and white subjects (Baris et al., 2000). Several additional studies that were conducted in Europe, the United States, and Australia, with various study periods set between 1961 and 1986, observed little association between myeloma incidence and various indices of socioeconomic status (including occupational and educational level, income, and social class) (Boffetta et al., 1986; Cuzick and De Stavola, 1988; McWhorter et al., 1989; Miligi et al., 1999; Nandakumar et al., 1986; Nandakumar et al., 1988; Vagero and Persson, 1986). The positive association of socioeconomic status with myeloma that was reported in earlier studies of mortality (Blattner et al., 1981; MacMahon, 1966; Velez et al., 1982) was possibly the result of better access to sensitive diagnostic methods by individuals of higher socioeconomic status.

HOST FACTORS Monoclonal Gammopathy of Undetermined Significance Monoclonal gammopathy of undetermined significance (MGUS) is an asymptomatic disorder characterized by production of monoclonal protein (M-component) and proliferation of plasma cells in persons without evidence of a plasma cell proliferative disorder. Individuals with MGUS are predisposed to developing myeloma. Sixty-four individuals with M-component, identified in a population-based survey in the Värmland district of Sweden, were observed for up to 20 years; two (3%) were diagnosed with myeloma and died from the disease (Axelsson, 1986), relating to an annual mortality rate of 2.1 per 1000 compared with approximately 0.02 per 1000 expected in similarly aged persons in Sweden during this time period (Cuzick et al., 1983). In a New Zealand rural community, six subjects (54%) were diagnosed with myeloma or Waldenström macroglobulinemia from a series of 11 with M-component who were followed

over a 31-year period, and five (46%) eventually died from their disease (Colls, 1999), relating to an annual mortality rate of approximately 42 per 1000. The higher rate of myeloma in the New Zealand study compared with the Swedish study may be attributable to the manner in which samples were stored and tested (New Zealand study: cellulose acetate electrophoresis on samples stored frozen for 3 years; Swedish study: paper electrophoresis with presumably little or no storage time), possibly resulting in differences in the type of Mcomponent identified (Colls, 1999). A rate of malignant transformation in persons with MGUS similar to that in the New Zealand study has also been observed in studies of patient populations (Kyle, 1993; Kyle et al., 2002; Pasqualetti et al., 1997; Van De Donk et al., 2001). This increased risk appears to continue even after 25 years or more of stable MGUS (Kyle et al., 2002). In a study of 241 MGUS patients followed from 1960 to 1994 in the 11 counties of Minnesota, the risk of progression to a malignant plasma cell disorder was 1% per year (Kyle and Rajkumar, 2005). The prevalence of M-component has been characterized in several large, population-based studies: the original Värmland study (Axelsson et al., 1966), and in studies conducted in the town of Thief River Falls, Minnesota (Kyle et al., 1972), the Finistère region of France (Saleun et al., 1982), and in a New Zealand rural population (Colls, 1999). The studies were consistent in revealing an approximate 0.5%–1.0% prevalence of M-component among adults, and an agerelated increase in prevalence, which reached 4%–5% among persons aged 80 years and over. There was also a sex differential, with men being 1.5–2.0 times more likely to have the condition. Each of these studies used either electrophoresis on paper or cellulose acetate, relatively insensitive methods to detect monoclonal proteins (Sinclair et al., 1986); the true prevalence may be higher by as much as 20% (Axelsson, 1986). The distribution of M-component in a French population with MGUS was as follows: IgG, 68%; IgM, 24%; IgA, 6%; more than one component, 2%; light chains only, 1% (Saleun et al., 1982). The age-specific prevalence of M-component in 1864 black veterans was higher than that in 857 white veterans (all ages: 5% in blacks compared with 2% in whites; age 70 years and older: 7% compared with 5%) (Schechter et al., 1991). The age-adjusted prevalence ratio of MGUS in blacks compared to whites was 3.0 with a similar risk for progression (17% in blacks and 15% in whites) during the first 10 years of follow up of 4 million US veterans admitted to Veterans Affairs hospitals (Landgren et al., 2006a). There was also a difference in the prevalence of M-component between black and white elderly in North Carolina, using the more sensitive laboratory method of agarose gel electrophoresis and immunofixation (age >70 years: 8% in blacks compared with 4% in whites) (Cohen et al., 1998). The prevalence of M-component among adults aged 70 to 80 years was lower among Japanese presenting at a community center for a health screening (3%) than among white American residents of a retirement community (10%) (Bowden et al., 1993). Several studies have found associations between clinical features of MGUS and progression to malignancy. The most consistent observation is that of the initial concentration of M-component at the time of MGUS diagnosis as a predictor of malignant progression (Baldini et al., 1996; Gregersen et al., 2001b; Gregersen et al., 2001a; Van De Donk et al., 2001); a recent study found a direct relation between the concentration of monoclonal protein in the serum and risk of malignant transformation (Kyle et al., 2002). Patients with IgA type Mcomponent were found to be at increased risk of malignant transformation in some studies (Blade et al., 1992; Gregersen et al., 2001b; Kyle et al., 2002; Ogmundsdottir et al., 2002), but not others (Baldini et al., 1996; Kyle, 1993). Associations with other clinical features reported in some, but not all studies, include the plasma cell percentage (Van De Donk et al., 2001), total uninvolved immunoglobulin (Kyle et al., 2002), k light chains (Van De Donk et al., 2001; van de Poel et al., 1995), hypogammaglobulinemia (Baldini et al., 1996; Gregersen et al., 2001b), and gammaglobulin (van de Poel et al., 1995). Similar translocations are found in both MGUS and myeloma (Fonseca et al., 2002a), although aberrations involving two or more

923

Multiple Myeloma chromosomes are more common in myeloma (Feinman et al., 1997). Deletion of chromosome 13q is also common in MGUS, indicating a potential early event leading to malignancy, although its role for eventual progression to myeloma remains to be determined prospectively (Avet-Loiseau et al., 1999; Konigsberg et al., 2000). Comparison of MGUS and multiple myeloma patients showed that 42% of the myeloma patients had methylation of the p16 tumor suppressor gene, whereas none of the MGUS patients did, suggesting that methylation could be a relevant oncogenic event (Mateos et al., 2001).

Prior Medical Conditions and Treatments Certain prior medical conditions and treatments have been suspected to increase the risk of myeloma, either through chronic immune stimulation or through another biologic mechanism. When first put forth, the hypothesis that chronic immune stimulation could cause myeloma was based on the assumption that antigenic stimulation could increase the likelihood of a malignant transformation in a mature B cell. Since then, evidence has accumulated that the malignant transformation in myeloma occurs at the level of a pre-B or stem cell (Wolvekamp and Marquet, 1990); these cells are not stimulated by antigen. Nonetheless, chronic immune stimulation could have a promotional effect on myeloma, although there is no experimental evidence to support this hypothesis. Table 47–2 lists the epidemiologic studies of the association between myeloma risk and prior medical conditions. Results from those studies are listed in Table 47–3. Interpretation of the results shown in Table 47–3 is somewhat difficult for several reasons: 1. Interviews of unknown and differing sensitivity and specificity were used in the various studies to measure prior conditions and treatments 2. Because of the large numbers of factors examined, many comparisons were made, increasing the likelihood that spurious associations were identified 3. Various studies categorized prior conditions and treatments in different ways, so it is not possible to make direct comparisons between studies 4. Myeloma may have a prolonged prodromal period, and most studies did not account for such a period in their analyses. So, for example, an association between a history of recent infection and myeloma could more plausibly be the result of myeloma leading to diminished immune competence than to the result of the infection in myeloma etiology.

Autoimmune Disorders A prospective analysis of the National Health and Nutrition Examination Survey (NHANES) I cohort found that myeloma risk increased with the number of autoimmune conditions reported, increasing to a 2.5-fold risk among those reporting two or more autoimmune conditions on the baseline questionnaire (Bourguet and Logue, 1993); however, this association was based on only 18 myeloma cases, and no association was observed when excluding myeloma cases diagnosed within 5 years following the NHANES interview. Two hospital-based case-control studies reported no association of any autoimmune disease with myeloma (Gramenzi et al., 1991; Linet et al., 1987); however, selecting controls from hospitalized patients may have introduced a bias that obscured an actual association. A population-based U.S. case-control study observed a modest increased myeloma risk with any autoimmune disease among blacks, but not whites (Lewis et al., 1994). A large population registry-based case-control study from Sweden did not support the association between personal/familial autoimmune diseases and risk of myeloma (Landgren et al., 2006b). Elevated myeloma incidence (Isomaki et al., 1978; Katusic et al., 1985; Kauppi et al., 1997) and elevated incidence of all hematologic malignancies (Matteson et al., 1991) were observed in several cohorts of rheumatoid arthritis patients, with relative risks varying from 1.2–8. In other cohorts, there were small increases in risk associated with a history of rheumatoid arthritis diagnosis, but a higher risk was

Table 47–2. Design Characteristics of Studies That Have Assessed the Risk of Myeloma in Relation to Prior Medical Conditions and Treatments Study Bjornadal et al., 2002 Boffetta et al., 1989 Bourguet and Logue, 1993 Brinton et al., 1989 Cuzick and De Stavola, 1988 Doody et al., 1992 Eriksson, 1993 Gallagher et al., 1983 Gramenzi et al., 1991 Goedert et al., 1998 Gridley et al., 1993 Grulich et al., 1999 Isomaki et al., 1978 Katusic et al., 1985 Kauppi et al., 1997 Koepsell et al., 1987 Landgren et al., 2006b Lewis et al., 1994 Linet et al., 1987, 1988 Mellemkjaer et al., 1996a Mellemkjaer et al., 1996b Mills et al., 1992 Montella et al., 2001 Pearce et al., 1986 Pickard et al., 2002 Vesterinen et al., 1993 Vineis et al., 2000

Design Cohort of patients with systemic lupus erythematosus Nested casecontrol, mortality Population-based case-control Cohort of patients with pernicious anemia Hospital-based case-control Population-based case-control Population-based case-control Hospital-based case-control* Hospital-based case-control Cohort of patients with AIDS Population-based cohort Cohort of patients with AIDS Population-based cohort Cohort of patients with rheumatoid arthritis Cohorts of patients with RA and Sjogren syndrome Population-based case-control Population-based case-control Population-based case-control Hospital-based case-control Cohort of patients with pernicious anemia Cohort of patients with rheumatoid arthritis Population-based cohort Hospital-based case-control Tumor-registrybased casecontrol* Population-based cohort Cohort of patients with asthma Population-based case-control

Population

Period

Sweden

1964–1995

American Cancer Society NHANES I, US

1982–1986 1971–1986

Veterans Administration institutions in US Six areas in England and Wales Kaiser Permanente, Oregon and California, US Northern Sweden

1969–1985

1982–1986

Vancouver, BC

1972–1981

Greater Milan area, northern Italy US and Puerto Rico

1983–1989 1985–1989

Sweden

1965–1984

New South Wales, Australia Finland

1980–1993

1978–1984 1956–1982

1967–1973

Rochester, Minnesota, US

1950–1974

Finland

1970–1991

Four SEER areas, US Sweden

1977–1981 1958–1998

Three metropolitan areas, US Baltimore, MD, US

1986–1989 1975–1982

Denmark

1977–1991

Denmark

1977–1987

California, US

1977–1982

Naples, Italy

1997–1999

New Zealand

1977–1981

Denmark

1977–1993

Finland

1970–1987

Italy

1990–1993

*Controls were registered with cancers other than multiple myeloma.

associated with indications of more severe rheumatoid arthritis, namely hospital visits or hospitalization for the condition (Gridley et al., 1993; Tennis et al., 1993). The results from Mellemkjaer et al. showed increased myeloma incidence only for the time period within

Table 47–3. Summary of Studies That Have Assessed Risk of Multiple Myeloma in Relation to Prior Medical Conditions and Treatments* Category

Study

Prevalence of Exposure in Controls (%)

Number of Exposed Cases†

Effect Estimate (95% Confidence Interval)

autoimmune diseases Autoimmune diseases

Number of autoimmune conditions 0 1 ≥2 Addison’s disease Amyotrophic lateral sclerosis Ankylosing spondylitis Autoimmune hemolytic anemia Crohn’s disease Grave’s disease Hashimoto’s disease Idiopathic thrombocytopenic purpura Multiple sclerosis Myasthenia gravis Pernicious anemia

Polyarteritis nodosa Polymyalgia rheumatica Polymyositis/dermatomyositis Psoriasis

Rheumatoid arthritis

Rheumatoid arthritis patients with >2 hospital visits for RA Rheumatoid arthritis patients who were hospitalized 1–4 years since hospitalization 5–15 years since hospitalization Sarcoidosis Scleroderma Sjogren syndrome Systemic lupus erythematosus

Systemic sclerosis Thyroiditis Ulcerative colitis Wegener’s granulomatosis

924

Gramenzi et al., 1991 Lewis et al., 1994, whites Lewis et al., 1994, blacks Linet et al., 1987 Bourguet and Logue, 1993

9.2 4.5 2.4 8.0‡

17 15 9 5‡

1.3 (0.7–2.3) 1.0 (0.5–1.8) 1.7 (0.7–3.7) 1.0 (0.3–3.6)

Landgren et al., 2006b Landgren et al., 2006b Landgren et al., 2006b Landgren et al., 2006b Landgren et al., 2006b Landgren et al., 2006b Lewis et al. 1994, whites Landgren et al., 2006b Lewis et al. 1994, whites Landgren et al., 2006b Landgren et al., 2006b Landgren et al., 2006b Brinton et al., 1989 Landgren et al., 2006b Lewis et al. 1994, whites Lewis et al. 1994, blacks Mellemkjaer et al., 1996a Landgren et al., 2006b Landgren et al., 2006b Landgren et al., 2006b Landgren et al., 2006b Lewis et al. 1994, whites Lewis et al. 1994, blacks Vineis et al., 2000 Cuzick and De Stavola, 1989 Doody et al., 1992 Eriksson, 1993 Gridley et al., 1993 Isomaki et al., 1978, males Isomaki et al., 1978, females Katusic et al., 1985 Kauppi et al., 1997 Landgren et al., 2006b Lewis et al., 1994, blacks Vineis et al., 2000 Pearce et al., 1986 Gridley et al., 1993

— — — 0.0 0.1 0.1 0.0 0.1 0.1 0.08 0.0 0.08 0.0 0.2 0.0 — 0.2 1.6 1.2 — 0.0 0.3 0.3 0.3 2.1 0.8 — 5.8 — 0.9 — — — — — 1.2 0.8 — 2.2 —

7 / 14538 PY 7 / 10464 PY 4 / 3717 PY 2 3 13 2 7 4 1 3 2 2 7 2 9 O / 4.3 E 67 6 5 7 O / 5.3 E 2 45 34 30 6 1 — 22 — 9 16 7 O / 3.3 E 21 O / 9.5 E 4 O/ 0.8 E 8 87 1 — 4 7

1.0 1.4 (1.0–2.0) 2.5 (0.9–6.6) 0.8 (0.2–4.1) 0.3 (0.1–1.0) 1.5 (0.7–3.1) 1.3 (0.2–8.0) 0.7 (0.3–1.8) 0.5 (0.2–1.6) 5.3 (0.2–114) 1.5 (0.3–6.7) 6.9 (0.5–88.6) 1.3 (0.2–8.0) 0.5 (0.2–1.1) 2.0 (0.3–4.1) 2.1 (1.0–4.0) 3.3 (2.2–4.8) 0.8 (0.3–2.2) 1.5 (0.5–4.5) 1.3 (0.5–2.7) 1.0 (0.2–5.4) 1.8 (1.2–2.8) 1.4 (0.8–2.9) 1.4 (0.9.2.2) 0.8 (0.3–2.0) 0.5 (0.1–4.2) 0.4 (0.1–1.6) 1.0d 1.2 8.0 (1.4–46.1) 1.2 (0.7–1.9) 2.1d 2.2d 5.0 (1.4–12.8) 1.2 (0.5–2.3) 0.9 (0.7–1.2) 0.5 (0.1–4.2) 0.4 (0.1–1.6) 2.3 (0.6–8.0) 1.6 (0.6–3.3)

Mellemkjaer et al., 1996b

—

21 O / 19.0 E

1.1 (0.7–1.7)

Landgren et al., 2006b Vineis et al., 2000 Kauppi et al., 1997 Landgren et al., 2006b Bjornadal et al., 2002 Cuzick and De Stavola, 1989 Landgren et al., 2006b Lewis et al., 1994, blacks Landgren et al., 2006b Cuzick and De Stavola, 1989 Landgren et al., 2006b Landgren et al., 2006b

— — 0.1 — — 0.0 — 0.0 0.1 0.2 0.1 6.0 0.0 0.0

16 O / 9.5 E 5 O / 9.5 E 14 — 2 2 7 O / 5.9 E 2 8 1 4 17 21 1

1.7 (1.0–2.7) 0.5 (0.2–1.2) 1.4 (0.7–2.8) 1.4 (0.2–11.9) 3.4 (0.4–12.4) 0.8 (0.2–4.1) 1.2 (0.5–2.5) •** 0.9 (0.4–2.2) 2.1 (0.2–24.2) 0.5 (0.2–1.6) 0.7** 1.2 (0.7–2.1) 0.4 (0.1–3.4)

Table 47–3. (cont.) Number of Exposed Cases†

Effect Estimate (95% Confidence Interval)

4.1 6.8 5.4 5.5 — 9.2 — — 6.0

5 20 22 16 — 9 22 16 8

1.0 (0.3–2.7) 0.7 1.2 (0.7–2.1) 1.3 (0.7–2.3) 0.8 (0.1–6.1) 1.3 (0.6–2.9) 1.1 (0.7–1.6) 0.5 (0.3–0.9) 1.8 (0.7–4.3)

31.3 7.3 11.3 20.5 47.3 31.9 17‡ 47.5 —

142 20 24 17 153 75 21‡ — —

1.1 1.1 (0.5–2.3) 3.1 (1.6–6.3) 0.6 (0.3–1.0) 1.1 (0.8–1.4) 1.2 (0.9–1.7) 1.0 (0.5–2.3) 1.7 (0.7–4.0) 1.1 (0.6–1.8)

11 / 20667 PY 5 2

1.0** 1.7 (0.3–10.6) 3.4 (0.1–286)

38.4 5.2 3.7

135 11 7

1.1 (0.8–1.4) 0.8 (0.4–1.6) 0.7 (0.3–1.6)

28.3 2.6 1.0 10.7 5.9 7.4 4.0 8.2 20.6 17.7 18.6 — 7.4 7.6 3.8 1.6

66 6 3 35 18 14 16 18 54 40 — — 43 18 8 0

1.2 (0.9–1.7) 1.1 (0.4–2.8) 1.5 (0.4–5.9) 0.9 (0.6–1.4) 1.1 (0.8–1.5) 1.6 (0.8–2.9) 1.0** 1.3 (0.6–2.7) 0.9 (0.7–2.0) 1.1 (0.7–1.6) 1.5 (0.5–4.6) 0.8 (0.4–1.5) 0.8 (0.5–1.2) 0.7 (0.4–1.3) 0.9 (0.4–2.0) —

Gallagher et al., 1983 Mills et al., 1992 Lewis et al., 1994, whites Lewis et al., 1994, blacks Mills et al., 1992 Pearce et al., 1986 Lewis et al., 1994, whites Lewis et al., 1994, blacks Cuzick and De Stavola, 1988 Doody et al., 1992 Lewis et al., 1994, whites Lewis et al., 1994, blacks Pearce et al., 1986 Vineis et al., 2000 Cuzick and De Stavola, 1988 Pearce et al., 1986 Lewis et al., 1994, whites Lewis et al., 1994, blacks

1.5 2.0 4.2 5.3 13.8 5.4 17.0 6.0 5.2 3.5 4.3 6.8 4.9 2.6 5.7 — 20.1 1.9 1.8 1.6

6 14 11 — 53 18 — 4 13 9 15 2 19 6 4 — 93 3 21 40

0.4 (0.1–1.1) 0.9 (0.5–1.9) 3.5 (1.1–10.6) 1.8 (0.4–7.7) 1.1 (0.8–1.5) 1.6 (0.9–2.9) 1.2 (0.4–3.6) 0.8 (0.3–2.5) 0.8 (0.4–1.5) 1.2 (0.5–2.5) 0.8** 2.0 (1.1–4.0) 1.2 (0.7–2.0) 1.1 (0.4–2.7) 0.9 (0.3–2.8) 0.6 (0.3–1.3) 1.1** 2.1 (0.5–8.5) 0.9 (0.3–2.4) 1.1 (0.7–1.6)

Gramenzi et al., 1991 Lewis et al., 1994, whites Lewis et al., 1994, blacks Gramenzi et al., 1991 Lewis et al., 1994, whites Lewis et al., 1994, blacks Linet et al., 1987

36.3 50.0 36.3 13.6 2.8 4.5 11‡

47 149 71 26 10 5 16‡

1.2 (0.8–1.8) 0.9 (0.7–1.1) 1.0 (0.7–1.4) 1.8 (1.1–2.8) 1.1 (0.5–2.3) 0.6 (0.2–1.4) 0.8 (0.3–1.8)

Category

Study

Prevalence of Exposure in Controls (%)

asthma Asthma

Asthma medication

Boffetta et al., 1989 Cuzick and De Stavola, 1988 Lewis et al., 1994, whites Lewis et al., 1994, blacks Mills et al., 1992 Pearce et al., 1986 Vesterinen et al., 1993, men Vesterinen et al., 1993, women Pearce et al., 1986

allergies and allergy treatments Allergies

Number of allergies 0 1 ≥2 Number of allergies 1–2 vs. 0 3–4 vs. 0 ≥5 vs. 0 Number of allergies 1–2 vs. 0 3–4 vs. 0 ≥5 vs. 0 Severe allergic reaction Hay fever

Allergy shots

Number of allergy desensitization shots 1–99 vs. 0 ≥100 vs. 0 Allergy treatment Bee sting allergy Drug allergies

Dust allergy Eczema

Non-eczema skin allergy Food allergies Household products allergies

Cuzick and De Stavola, 1988 Eriksson, 1993 Gallagher et al., 1983 Gramenzi et al., 1991 Lewis et al., 1994, whites Lewis et al., 1994, blacks Linet et al., 1987 Mills et al., 1992 Vineis et al., 2000 Bourguet and Logue, 1993

— 5406 2646 Lewis et al., 1994, whites

Lewis et al., 1994, blacks

Lewis et al., 1994, whites Lewis et al., 1994, blacks Boffetta et al., 1989 Cuzick and De Stavola, 1988 Doody et al., 1992 Lewis et al., 1994, whites Lewis et al., 1994, blacks Mills et al., 1992 Vineis et al., 2000 Koepsell et al., 1987 Lewis et al., 1994, whites Lewis et al., 1994, blacks Pearce et al., 1986 Koepsell et al., 1987

bacterial infections Acute bacterial infections Chronic bacterial infections

(continued)

925

Table 47–3. (cont.) Category Number of bacterial infections 1 vs. 0 2 vs. 0 >2 vs. 0 Number of bacterial infections 1 vs. 0 ≥2 vs. 0 Number of bacterial infections 1 vs. 0 2 vs. 0 ≥3 vs. 0 Number of acute bacterial infections 1–2 vs. 0 3–5 vs. 0 ≥6 vs. 0 Number of acute bacterial infections 1–2 vs. 0 3–5 vs. 0 ≥6 vs. 0 Bronchitis Chronic bronchitis Ear infection Gonorrhea Osteomyelitis Pancreatitis Pneumonia Pyelonephritis Rheumatic fever

Scarlet fever

Sinus infection Strep throat Syphilis Throat, tonsil, or ear infection Tonsillitis Tooth abscess Tuberculosis

Typhus Whooping cough Urinary tract infection

Study

Prevalence of Exposure in Controls (%)

Number of Exposed Cases†

Effect Estimate (95% Confidence Interval)

35.2 8.4 1.5

36 15 7

0.8 (0.5–1.6) 1.5 (0.7–3.0) 3.8 (1.3–10.8)

2 0

0.4 (0.1–1.5) 0.1 (0.01–2.3)

Gramenzi et al., 1991

Bourguet and Logue, 1993 4734 2152 Koepsell et al., 1987 44.6 18.9 6.6

220 96 41

0.9 (0.7–1.1) 1.1 (0.8–1.5) 1.1 (0.7–1.8)

20.1 8.7 20.8

71 31 47

1.0 (0.7–1.3) 1.0 (0.6–1.6) 0.7 (0.5–1.0)

Doody et al., 1992 Gramenzi et al., 1991 Koepsell et al., 1987 Lewis et al., 1994, whites Lewis et al., 1994, blacks Koepsell et al., 1987 Lewis et al., 1994, whites Lewis et al., 1994, blacks Koepsell et al., 1987 Lewis et al., 1994, whites Lewis et al., 1994, blacks Lewis et al., 1994, whites Lewis et al., 1994, blacks Koepsell et al., 1987 Lewis et al., 1994, whites Lewis et al., 1994, blacks Gramenzi et al., 1991 Doody et al., 1992 Gramenzi et al., 1991 Koepsell et al., 1987 Landgren et al., 2006b Lewis et al., 1994, whites Lewis et al., 1994, blacks Cuzick and De Stavola, 1988 Gramenzi et al., 1991 Koepsell et al., 1987 Lewis et al., 1994, whites Lewis et al., 1994, blacks Koepsell et al., 1987 Lewis et al., 1994, whites Lewis et al., 1994, blacks Lewis et al., 1994, whites Lewis et al., 1994, blacks Koepsell et al., 1987 Lewis et al., 1994, whites Lewis et al., 1994, blacks Koepsell et al., 1987 Lewis et al., 1994, whites Lewis et al., 1994, blacks Koepsell et al., 1987 Doody et al., 1992 Eriksson, 1993 Gramenzi et al., 1991 Koepsell et al., 1987 Vineis et al., 2000 Gramenzi et al., 1991 Cuzick and De Stavola, 1988 Gramenzi et al., 1991 Koepsell et al., 1987 Lewis et al., 1994, white men Lewis et al., 1994, black men Lewis et al., 1994, white women Lewis et al., 1994, black women

19.8 6.8 9.7 10.0 10.7 10.3 8.9 2.3 4.1 1.8 12.0 1.3 0.4 0.6 0.9 0.8 0.3 2.2 0.7 0.8 0.9 4.8 3.1 0.1 2.7 4.0 20.8 8.2 14.1 9.7 3.0 2.5 12.3 7.3 6.8 2.6 0.8 0.09 2.0 1.6 10.2 5.3 49.1 5.0 4.5 2.7 2.1 2.0 5.9 40.3 28.3 23.9 19.9 9.3 17.3 7.6

46 12 13 27 15 70 22 2 21 7 2 17 3 3 5 2 4 9 4 3 3 9 32 9 15 7 87 17 68 36 6 22 32 12 19 1 12 2 6 18 25 7 341 11 12 8 20 5 9 176 35 188 37 8 24 15

Gramenzi et al., 1991

75.5

81

Lewis et al., 1994, whites

Lewis et al., 1994, blacks 1.3 (0.9–1.9) 0.9 (0.5–1.7) 0.6 (0.3–1.1) 2.0 (1.0–3.9) 1.3 (0.7–2.4) 1.0 (0.7–1.3) 0.7 (0.4–1.2) 0.4 (0.1–1.6) 0.6 (0.4–1.0) 1.6 (0.6–3.9) 0.4 (0.1–1.6) 1.5 (0.7–3.0) 1.7 (0.4–7.5) 2.3 (0.6–9.4) 0.8 (0.3–2.2) 0.3 (0.1–1.3) 1.3 (0.3–5.1) 1.1 (0.5–2.4) 2.6 (0.7–9.2) 2.0 (0.5–8.4) 1.5 (0.3–9.0) 1.4 (0.6–3.2) 1.7 (1.1–2.8) 1.2 (0.5–2.7) 1.4 (0.7–2.7) 0.7 (0.3–1.6) 1.0** 2.0 (1.1–3.9) 0.8 (0.6–1.0) 0.9 (0.6–1.4) 1.0 (0.4–2.4) 1.7 (1.0–3.1) 0.8 (0.5–1.2) 0.6 (0.3–1.2) 0.9 (0.5–1.5) 0.2 (0.0–1.3) 1.9 (0.8–4.7) 11.6 (0.8–173) 1.5 (0.6–4.0) 1.6 (0.8–3.2) 0.8 (0.5–1.2) 0.6 (0.3–1.3) 1.0 (0.8–1.2) 2.0 (0.8–5.5) 1.1 (0.4–2.8) 2.3 (0.9–5.7) 1.3 (0.7–2.5) 0.7 (0.3–1.8) 1.2 (0.6–2.6) 1.1** 1.1 (0.7–1.7) 1.2 (1.0–1.5) 1.0 (0.7–1.5) 1.0 (0.4–2.3) 0.9 (0.5–1.5) 2.0 (1.1–3.8)

viral infections Viral infections

926

0.8 (0.5–1.3)

Table 47–3. (cont.) Category Number of viral diseases 1 vs. 0 2 vs. 0 3 vs. 0 4 vs. 0 AIDS Chicken pox

Hepatitis

Hepatitis C virus detected in blood Herpes fever blisters Herpes genitalis Herpes labialis Infectious mononucleosis

Measles German measles Mumps

Polio Rubella Shingles (herpes zoster)

Shingles, by years before myeloma diagnosis ≥10 years 5–10 years 3–5 years 1–3 years

Study

Prevalence of Exposure in Controls (%)

Number of Exposed Cases†

Effect Estimate (95% Confidence Interval)

8.4 19.0 37.8 31.2 — — — 73.9 47.4 69.7 69.8 67.6 — 2.4 3.6 1.0 6.3 8.0 42.1 — — 5.8 0.6 0.8 2.2 88.7 59.7 89.3 45.4 65.4 40.9 73.2 63.9 69.3 1.2 1.6 14.3 19.3 3.2 6.1 8.3 9.7 3.2

25 76 169 161 7 O / 0.95 E 3 3 312 42 490 235 235 — 7 8 3 23 13 322 — — 34 2 62 6 340 69 606 192 278 45 522 221 134 5 5 15 114 5 13 71 39 11

0.7 (0.3–1.6) 0.9 (0.5–1.9) 1.1 (0.6–1.9) 1.2 (0.6–2.3) 7.4 (3.0–15.2) 4.5 (0.9–13.2) 12.1 (2.5–35) 1.0** 0.7 (0.5–1.1) 1.0 (0.8–1.3) 0.8 (0.6–1.1) 0.9 (0.6–1.2) 0.8 (0.5–1.2) 0.4 (0.2–1.0) 0.7 (0.3–1.6) 1.5 (0.4–5.4) 1.5 (0.9–2.4) 4.5 (1.9–10.7) 1.2 (1.0–1.4) 2.3 (0.8–6.5) 0.8 (0.6–1.2) 1.4** 0.4 (0.1–2.6) 0.8 (0.2–3.1) 1.2 (0.5–3.0) 0.9** 1.0 (0.6–1.5) 1.0 (0.7–1.4) 1.0** 1.0** 1.0 (0.7–1.6) 1.1 (0.9–1.5) 1.0 (0.8–1.3) 0.8 (0.6–1.2) 0.7 (0.2–2.0) 0.9 (0.3–2.5) 0.8 (0.5–1.4) 1.4** 0.7 (0.2–2.3) 1.8 (0.9–3.5) 1.2 (0.9–1.7) 1.0 (0.7–1.5) 1.7 (0.8–3.6)

7.5 1.5 0.8 1.5

36 16 6 14

1.2** 2.7** 1.9** 2.3**

11.9

21

1.6 (0.9–3.1)

— — — 10.0 0.5 4.5 32.0 4.4 6.4 3.9 0.7 1.4 5.9 5.2 1.7

10 / 21912 PY 1 / 3984 PY 7 / 2823 PY 37 17 6 39 85 21 8 2 1 9 18 1

7.4 14.1 8.5 16.4

17 22 38 36

1.0** 2.0 (1.2–3.3) 4.3 (1.5–12.4) 1.9 (1.0–3.5) 0.5 (0.3–0.8) 0.8 (0.3–2.1) 0.9 (0.6–1.5) 1.4 (0.9–2.2) 0.8 (0.5–1.4) 1.0 (0.5–2.4) 0.8 (0.2–4.1) 0.3 (0.0–2.6) 0.6 (0.2–1.5) 0.9 (0.5–1.5) 0.3 (0.1–2.1) 0.8 (0.3–2.1) 1.9 (1.1–3.8) 0.7 (0.3–1.4) 1.1 (0.7–1.7) 1.0 (0.6–1.4) 0.7 (0.4–1.1)

Koepsell et al., 1987

Fordyce et al., 2000 Goedert et al., 1998 Grulich et al., 1999 Cuzick and De Stavola, 1988 Gramenzi et al., 1991 Koepsell et al., 1987 Lewis et al., 1994, whites Lewis et al., 1994, blacks Vineis et al., 2000 Koepsell et al., 1987 Lewis et al., 1994, whites Lewis et al., 1994, blacks Vineis et al., 2000 Montella et al., 2001 Koepsell et al., 1987 Vineis et al., 2000 Vineis et al., 2000 Cuzick and De Stavola, 1988 Gramenzi et al., 1991 Koepsell et al., 1987 Lewis et al., 1994, whites Cuzick and De Stavola, 1988 Gramenzi et al., 1991 Koepsell et al., 1987 Cuzick and De Stavola, 1988 Cuzick and De Stavola, 1988 Gramenzi et al., 1991 Koepsell et al., 1987 Lewis et al., 1994, whites Lewis et al., 1994, blacks Koepsell et al., 1987 Lewis et al., 1994, whites Gramenzi et al., 1991 Cuzick and De Stavola, 1988 Eriksson, 1993 Gramenzi et al., 1991 Koepsell et al., 1987 Lewis et al., 1994, whites Lewis et al., 1994, blacks Cuzick and De Stavola, 1988

other chronic conditions Chronic inflammatory conditions Number of inflammatory conditions 0 1 ≥2 Rheumatic disease Angina pectoris Arthritis Chronic lung disease Cirrhosis Colitis or inflammatory bowel disease

Diabetes

Gramenzi et al., 1991 Bourguet and Logue, 1993

Eriksson, 1993 Landgren et al., 2006b Eriksson, 1993 Boffetta et al., 1989 Doody et al., 1992 Lewis et al., 1994, whites Lewis et al., 1994, blacks Lewis et al., 1994, whites Lewis et al., 1994, blacks Eriksson, 1993 Lewis et al., 1994, whites Lewis et al., 1994, blacks Vineis et al., 2000 Boffetta et al., 1989 Eriksson, 1993 Lewis et al., 1994, whites Lewis et al., 1994, blacks Vineis et al., 2000

(continued)

927

Table 47–3. (cont.) Category Disc and other muculoskeletal disease Embedded shrapnel Gallbladder disease Gastric ulcer Goiter Hyperlipidemia Hypertension Hyperparathyroidism Hypothyroidism Hyperthyroidism and/or myxedema Kidney disease Malaria Medical implant Metabolic disorder Nervous complaints Skin infection Thrombosis

Prevalence of Exposure in Controls (%)

Number of Exposed Cases†

Effect Estimate (95% Confidence Interval)

Doody et al., 1992 Koepsell et al., 1987 Lewis et al., 1994, whites Lewis et al., 1994, blacks Eriksson, 1993 Eriksson, 1993 Lewis et al., 1994, whites Lewis et al., 1994, blacks Eriksson, 1993 Eriksson, 1993 Pickard et al., 2002 Lewis et al., 1994, whites Lewis et al., 1994, blacks Gallagher et al., 1983 Lewis et al., 1994, whites Lewis et al., 1994, blacks Boffetta et al., 1989 Gramenzi et al., 1991 Koepsell et al., 1987 Vineis et al., 2000 Koepsell et al., 1987 Eriksson, 1993 Eriksson, 1993 Koepsell et al., 1987 Eriksson, 1993

12.3 1.7 6.2 4.1 5.0 4.5 0.4 0.2 4.5 29.5 — 2.2 1.8 1.2 4.4 1.6 2.7 3.6 0.8 4.3 7.0 5.0 9.1 1.5 9.0

36 22 16 3 10 14 1 2 10 62 4 / 14703 PY 5 7 6 18 2 5 7 62 21 46 8 24 9 169

2.3 (1.2–4.1) 2.0 (1.1–3.5) 0.8 (0.4–1.4) 0.4 (0.1–1.3) 1.0 (0.4–2.7) 1.5 (0.6–3.7) 0.9 (0.1–8.5) 3.2 (0.4–23.0) 1.1 (0.4–3.0) 0.9 (0.6–1.4) 2.2 (0.6–5.5) 0.4 (0.2–1.1) 1.5 (0.6–3.7) 5.0 (1.0–25.7) 1.1 (0.6–1.9) 0.5 (0.1–2.3) 1.4 (0.5–4.0) 1.8 (0.7–4.5) 0.9 (0.3–2.5) 1.9** 1.0 (0.7–1.4) 0.6 (0.3–1.6) 1.2 (0.6–2.3) 1.2 (0.5–2.5) 0.8 (0.4–1.6)

Vineis et al., 2000 Cuzick and De Stavola, 1988 Koepsell et al., 1987 Lewis et al., 1994, whites Lewis et al., 1994, blacks Koepsell et al., 1987 Koepsell et al., 1987 Linet et al., 1987 Koepsell et al., 1987 Cuzick and De Stavola, 1988 Lewis et al., 1994, whites Lewis et al., 1994, blacks Vineis et al., 2000

10.1 20.0 25.1 18.6 18.8 3.8 79.4 19.0‡ 2.8 29.6 54.1 22.3 25.5

21 65 182 83 45 25 556 25‡ 10 99 166 38 45

1.2 (0.7–1.9) 0.8** 1.1 (0.8–1.3) 1.1 (0.8–1.5) 1.0 (0.7–1.5) 1.0 (0.6–1.8) 0.9 (0.7–1.2) 1.2 (0.6–2.6) 0.5 (0.3–1.1) 0.8** 0.8 (0.6–1.0) 0.8 (0.5–1.1) 0.8 (0.6–1.2)

Lewis et al., 1994, whites Lewis et al., 1994, blacks Koepsell et al., 1987

98.1 96.3

339 186

0.7 (0.3–1.8) 0.8 (0.4–1.6)

14.0 24.0 32.9 26.2 6.5 5.0 9.5 4.0 14.3 60.2 59.1 51.2 13.0 23.7 53.9 62.5 45.7 0.8 62.2 8.3 83.4 90.6 76.0 66.0 46.6 78.0 59.2 75.7 56.6

99 137 153 127 28 11 46 29 10 393 220 100 49 28 322 213 82 7 261 31 91 617 277 138 182 70 386 262 119

0.8 (0.4–1.6) 0.7 (0.3–1.3) 0.7 (0.4–1.3) 0.8 (0.4–1.5) 1.0** 3.0 (1.4–6.4) 1.2** 1.8** 0.9 (0.4–1.8) 0.9 (0.7–1.1) 1.1 (0.8–1.4) 0.9 (0.7–1.2) 0.9** 0.9 (0.6–1.4) 1.0 (0.8–1.3) 1.0 (0.8–1.4) 0.8 (0.6–1.1) 2.3** 1.0** 0.9** 0.7 (0.4–1.3) 0.9 (0.7–1.3) 1.0 (0.7–1.5) 1.1 (0.7–1.4) 1.0** 0.6 (0.4–1.0) 0.9 (0.7–1.1) 1.0 (0.8–1.4) 1.2 (0.8–1.7)

Study

other acute conditions Adenoidectomy Blood transfusion

Horse serum injections Insect sting Lymphoid tissue surgery Snakebite Tonsillectomy

immunizations Childhood illness/vaccines Number of diseases for which subject was immunized 1 vs. 0 2 vs. 0 3 vs. 0 4 vs. 0 BCG Cholera Diphtheria Influenza Polio

Scarlet fever Smallpox

Tetanus

928

Cuzick and De Stavola, 1988 Gramenzi et al., 1991 Cuzick and De Stavola, 1988 Cuzick and De Stavola, 1988 Gramenzi et al., 1991 Koepsell et al., 1987 Lewis et al., 1994, whites Lewis et al., 1994, blacks Cuzick and De Stavola, 1988 Gramenzi et al., 1991 Koepsell et al., 1987 Lewis et al., 1994, whites Lewis et al., 1994, blacks Cuzick and De Stavola, 1988 Cuzick and De Stavola, 1988 Cuzick and De Stavola, 1988 Gramenzi et al., 1991 Koepsell et al., 1987 Lewis et al., 1994, whites Lewis et al., 1994, blacks Cuzick and De Stavola, 1988 Gramenzi et al., 1991 Koepsell et al., 1987 Lewis et al., 1994, whites Lewis et al., 1994, blacks

929

Multiple Myeloma Table 47–3. (cont.) Category Tetanus ≥4 times Typhoid Typhus Whooping cough Yellow fever

Study Cuzick and De Stavola, 1988 Cuzick and De Stavola, 1988 Cuzick and De Stavola, 1988 Cuzick and De Stavola, 1988 Cuzick and De Stavola, 1988

Prevalence of Exposure in Controls (%)

Number of Exposed Cases†

8.8 14.8 6.3 1.8 7.0

33 63 29 12 36

Effect Estimate (95% Confidence Interval) 0.9** 1.0** 1.1** 1.7** 1.3**

*Comparison is ever vs. never unless otherwise specified. † For cohort studies, the number of person-years (PY), or observed (O) and expected (E) cases are presented, where available. ‡ Based on the number of discordant pairs only. **Unadjusted odds ratio.

4 years since hospitalization for rheumatoid arthritis, indicating a possible bias resulting from a prodromal period for myeloma (Mellemkjaer et al., 1996b). Case-control studies conducted in the United States (Doody et al., 1992), New Zealand (Pearce et al., 1986), and Sweden (Eriksson, 1993) found 1.2-fold to 8-fold increased myeloma risk associated with rheumatoid arthritis. Three other case-control studies found no association of myeloma with rheumatoid arthritis (Cuzick and De Stavola, 1989; Landgren et al., 2006b; Lewis et al., 1994). Cytotoxic medications frequently used in managing rheumatoid arthritis, in particular azathioprine and cyclophosphamide, have been associated with increased risk of hematologic malignancy (Baker et al., 1987; Beauparlant et al., 1999; Georgescu et al., 1997; Matteson et al., 1991); nevertheless, several studies observed increased risk of hematologic malignancies in rheumatoid arthritis patients that could not be completely explained by the use of cytotoxic medications (Cibere et al., 1997; Symmons, 1985; Tennis et al., 1993). Several other autoimmune diseases have been studied in relation to myeloma incidence. In a large general population registry-based case-control study from Sweden, myeloma risk was significantly elevated among subjects with polymyalgia rheumatica (OR=1.8, 95% CI:1.2–2.8.) (Landgren et al., 2006b). Three cohorts of pernicious anemia patients, from Denmark (Mellemkjaer et al., 1996a), Sweden (Hsing et al., 1993), and the United States (Brinton et al., 1989) had an approximate twofold increased risk of myeloma. However, in Denmark the excess risk was restricted to within 5 years of follow-up, indicating a possible bias (Mellemkjaer et al., 1996a). A modest association of prior pernicious anemia with myeloma was observed in black (OR = 1.5), but not white (OR = 0.8) persons in a population-based casecontrol study set in three metropolitan areas of the United States (Lewis et al., 1994). Myeloma risk was significantly elevated among subjects with a history of pernicious anemia (OR=3.3, 95% CI:2.2–4.8) in a large Swedish case-control study (Landgren et al., 2006b). Sjogren syndrome was associated with a 3.4-fold increased risk of myeloma (95% CI:0.4–12.4) in a study of Finnish patients (Kauppi et al., 1997). In a cohort of patients with systemic lupus erythematosus, a 4.1-fold increased risk of hematologic cancers (including one case of Waldenstrom macroglobulinemia) was observed (Abu-Shakra et al., 1996). A relatively large cohort study of lupus patients in Sweden (n = 5715) found that myeloma risk was only slightly elevated (OR = 1.2), based on 7 cases (Bjornadal et al., 2002) but a large Swedisle case-control study did not show any association (Landgren et al., 2006b). A twofold elevation in myeloma risk associated with lupus was also observed in a case-control study, but the estimate was very imprecise (Lewis et al., 1994). Some cohort studies of lupus patients did not observe any case of myeloma (Pettersson et al., 1992; Sweeney et al., 1995), but these studies were too small to reliably identify even a moderate increase in risk for such a rare malignancy. There has been some suggestion of increased myeloma risk associated with rare autoimmune conditions including Graves disease and Hashimoto’s disease (Lewis et al., 1994) but these associations were not observed in a large Swedish casecontrol study (Landgren et al., 2006b).

Allergies and Asthma There appears not to be an association between myeloma risk and a history of asthma (Boffetta et al., 1989; Lewis et al., 1994; Pearce

et al., 1986; Vesterinen et al., 1993), nor with a history of allergies or allergy treatments (Boffetta et al., 1989; Cuzick and De Stavola, 1988; Doody et al., 1992; Eriksson, 1993; Gramenzi et al., 1991; Koepsell et al., 1987; Lewis et al., 1994; Linet et al., 1987; Mills et al., 1992; Pearce et al., 1986; Vineis et al., 2000). A prospective analysis of the NHANES I cohort found that myeloma risk increased in relation to the number of allergies reported in the interview, but the estimates were very imprecise (Bourguet and Logue, 1993), and this pattern was not observed in another study population (Lewis et al., 1994). In one study that found an overall association between a history of allergies and the risk of myeloma (OR = 3.1) (Gallagher et al., 1983), the authors noted that the character of reported allergies differed between cases and controls. Nearly half of the allergies reported by controls were described as breathing difficulties, while fewer than 20% of myeloma patients with allergies reported this symptom; myeloma patients described their symptoms mainly as skin rashes, swelling, and hives.

Bacterial Infections Three studies found no association of myeloma risk with the number of lifetime acute or chronic bacterial infections (Bourguet and Logue, 1993; Koepsell et al., 1987; Lewis et al., 1994). However, risk increased in relation to the number of previous acute or chronic bacterial infections in a hospital-based case-control study in Italy, with 3.8-fold increase in risk for those reporting more than two infections (Gramenzi et al., 1991). Osteomyelitis was associated with an increased risk of myeloma in three study populations, with odds ratios ranging between 1.5 and 2.3 (Koepsell et al., 1987; Lewis et al., 1994). Five studies found an association between a history of rheumatic fever and myeloma risk, with odds ratios of 1.2 to 1.7 (Doody et al., 1992; Gramenzi et al., 1991; Koepsell et al., 1987; Landgren et al., 2006b; Lewis et al., 1994). Other bacterial infections that have been observed to be associated with myeloma include bronchitis (Doody et al., 1992), pneumonia (Lewis et al., 1994), pyelonephritis (Gramenzi et al., 1991), scarlet fever (Gramenzi et al., 1991), sinus infection (Koepsell et al., 1987), syphilis (Lewis et al., 1994), tuberculosis (Doody et al., 1992; Gramenzi et al., 1991), and urinary tract infection (Lewis et al., 1994). Most of these associations were not confirmed in a second report.

Viral Infections Specific viral infections may play a role in myeloma risk. Strong associations between myeloma incidence and AIDS (RR ranging from 4.5–12.1) have been observed in studies linking AIDS patient registries to population cancer registries in the United States and Puerto Rico (Goedert et al., 1998), New York City (Fordyce et al., 2000), and Australia (Grulich et al., 1999). It is not known whether this excess of myeloma is attributable to AIDS-related immunosuppression, or results from infection with viruses associated with other AIDS-related cancers, such as Kaposi sarcoma-associated herpesvirus, also known as human herpesvirus-8 (HHV-8) (Beral and Newton, 1998). HHV-8 viral sequences have been found in cultured nonmalignant bone marrow dendritic cells from myeloma patients and bone marrow biopsies (Beksac et al., 2001; Brousset et al., 1997; MacKenzie et al., 1997; Marcelin et al., 1997; Rettig et al., 1997; Said et al., 1997; Tedeschi

930

PART IV: CANCER BY TISSUE OF ORIGIN

et al., 2001). However, the role of HHV-8 in the pathophysiology of myeloma is controversial, and most investigators have failed to detect increased prevalence of HHV-8 in samples of blood, bone marrow, or bone marrow stromal cells from myeloma patients (Ablashi et al., 2000; Cathomas et al., 1998; Cull et al., 1999; Dominici et al., 2000; Drabick et al., 2000; Olsen et al., 1998; Parravicini et al., 1997; Patel et al., 2001; Rask et al., 2000; Sitas et al., 1999; Tarte et al., 1998; Tisdale et al., 1998; Yi et al., 1998). It has been suggested that HHV8 sequence variation and differences in technique could account for discrepancies in results; for example, some studies found HHV-8 DNA in bone marrow stromal cells or biopsies, but no serologic responses to HHV-8 were present (Agbalika et al., 1998; Chauhan et al., 1999). In three of five studies that ascertained a history of shingles, or herpes zoster, this condition was noted to be more prevalent among myeloma cases than controls (Cuzick and De Stavola, 1988; Gramenzi et al., 1991; Lewis et al., 1994). Some investigators suggested that because of the temporal proximity of the infection to the recognition of myeloma it most likely was a manifestation of the as-yetundiagnosed malignancy; indeed, one analysis showing an overall increased myeloma risk associated with shingles indicated little association with shingles occurring 10 or more years prior to myeloma diagnosis (Cuzick and De Stavola, 1988). Difficulties in recall may obscure associations between viral history and disease. Self-reported history of hepatitis infection has not been associated with myeloma risk (Koepsell et al., 1987; Lewis et al., 1994; Vineis et al., 2000), whereas hepatitis C virus (HCV) detected in blood was strongly associated (OR = 4.5) (Montella et al., 2001). A small seroprevalence study supported this finding, with an observed 11% anti-HCV positivity among myeloma cases, and 0% among a rheumatoid arthritis patient control group (Gharagozloo et al., 2001).

Other Chronic and Acute Conditions A prospective analysis of the NHANES I cohort found that myeloma risk increased by the number of inflammatory conditions reported in the interview (including gout, gallstones, pleurisy, and recurrent or chronic enteritis), with the risk increasing from those reporting one condition (OR = 2.0, compared with those with 0 conditions) to those reporting two or more conditions (OR = 4.3); there was also a statistically significant increase in risk with longer time since first exposure (RR = 1.6 for each additional 10 years since the start of the inflammatory condition) (Bourguet and Logue, 1993). Similarly, increased myeloma risk was associated with a history of chronic inflammatory conditions in a case-control study in Italy (Gramenzi et al., 1991). However, a different collection of diseases were included in their grouping (including rheumatic fever, ulcerative colitis, multiple sclerosis, glomerulonephritis, peptic ulcer, and Raynaud disease) than in the NHANES study. No consistent patterns emerge for specific chronic or acute conditions or treatments (Boffetta et al., 1989; Cuzick and De Stavola, 1988; Doody et al., 1992; Eriksson, 1993; Gallagher et al, 1983; Gramenzi et al., 1991; Koepsell et al., 1987; Lewis et al., 1994; Linet et al., 1987; Pickard et al., 2002; Vineis et al., 2000). Where elevated risks have been observed, they were usually not replicated in a subsequent study. There was some overlap of specific conditions examined in various studies with the groupings of inflammatory conditions examined in the studies by Bourguet et al. (1993) and Gramenzi et al. (1991), with sometimes inconsistent results. For example, no association was observed between myeloma incidence and colitis or other inflammatory bowel disease in three studies (Eriksson, 1993; Lewis et al., 1994; Vineis et al., 2000). There were increased risks observed for undefined rheumatic conditions (OR = 1.9) (Eriksson, 1993), gastric ulcer (OR = 1.5) (Eriksson, 1993), disc- and other musculoskeletal disease (OR = 2.3) (Doody et al., 1992), and kidney disease (OR = 1.5) (Boffetta et al., 1989).

Summary The presence of some autoimmune diseases appears to increase myeloma risk, as do certain chronic inflammatory conditions. Specific viruses, particularly those that cause immunosuppression, may also contribute to the etiology of myeloma. A possible underlying pathologic basis for the relation of certain autoimmune diseases, viral infections, or other conditions with

myeloma risk concerns the cytokine interleukin-6 (IL-6). IL-6 is a potent stimulator of B-cell differentiation and a promoter of myeloma cell growth (Hirano, 1991; Wolvekamp and Marquet, 1990). IL-6 is produced by a wide variety of cell types in response to viruses, bacterial products, trauma, and other stimuli (Wolvekamp and Marquet, 1990). Hirano notes that IL-6 gene deregulation may occur from insertion of viral DNA in the promoter region of the IL-6 gene, by a cytokine cascade induced by an inflammatory reaction, or by other mechanisms (Hirano, 1991). IL-6 production has been noted to increase substantially in association with myeloma, rheumatoid arthritis, systemic lupus erythematosus, acute infectious neural diseases, trauma, cardiac myxoma, and transplantation (Wolvekamp and Marquet, 1990). A second pathway by which certain medical conditions could contribute to myeloma risk is through an increased incidence of MGUS. The production of M-component as a response to chronic disease and infection has been studied epidemiologically only to a very limited extent. Autoimmune diseases are common in patients with monoclonal gammopathy (Youinou et al., 1996), and M-components have been observed to have autoimmune activity against self-antigens (Wang et al., 1992). Monoclonal and oligoclonal immunoglobulins were found in 24 of 27 AIDS patients with Kaposi sarcoma, and 2 of 15 AIDS patients with opportunistic infections, suggesting that B-cell activation may be operative in malignant proliferation among AIDS patients (Papadopoulos et al., 1985). A cohort of patients with infections found increased levels of M-component to be present in persons with leishmaniasis (80%) and cytomegalovirus (44%), but not in those with echinococcosis (6%), nor infectious mononucleosis (0%), nor among healthy control subjects (3%) (Haas et al., 1990). A second study also found high prevalence of M-component in patients with cytomegalovirus (40%) compared with patients with Epstein Barr virus (0%) (Buhler et al., 2002).

Familial Aggregation There have been numerous reports of myeloma occurring in two or more members of a family and of myeloma occurring with MGUS in families (Maldonado and Kyle, 1974; Landgren et al., 2006b; Shoenfeld et al., 1982). Several case-control studies observed a relation between a first-degree family history of multiple myeloma with myeloma occurrence (OR = 2.3, 95% CI:0.5–10.1) (Bourguet et al., 1985); (OR = 3.7, 95% CI:1.2–2.0) (Brown et al., 1999); (OR = 5.6, 90% CI:1.2–28) (Eriksson and Hallberg, 1992); (OR = 1.7, 95% CI:1.0–2.7) (Landgren et al., 2006b). The increased risk of myeloma associated with family history of myeloma has been observed to be stronger in blacks (OR = 17.4, 95% CI:2.4–348) than whites (OR = 1.5, 95% CI:0.3–6.4), based on very small numbers (Brown et al., 1999). Such clustering of myeloma in families might arise from shared genetic factors or common environmental exposures, and it is plausible that these tumors might differ from non-familial cases. Olshan (1991) compared the distributions of sex, age, immunoglobulin classes, and kappa-to-lambda ratios, for familial myeloma cases reported in the literature with data from the SEER program and the Mayo Clinic. Kappa-to-lambda immunoglobulin ratios were higher in familial cases than in other cases, but other characteristics did not differ. Family aggregation of myeloma may be partially explained by familial aggregation of other medical conditions such as autoimmune diseases. Family members of myeloma patients have elevated levels of immunoglobulins, rheumatoid factor, and autoantibodies in blood (Festen et al., 1977; Linet et al., 1988; Youinou et al., 1996). Linet et al. investigated the relation between a history of autoimmune diseases in first-degree relatives and the occurrence of myeloma; they observed an odds ratio of 3.0 (95% CI:1.3–7.1) (Linet et al., 1988). Familial aggregation of myeloma with degenerative central nervous system diseases in first-degree relatives has also been reported (OR = 4.4, 95% CI:1.9–10.3) (Grufferman et al., 1989).

Genetic Susceptibility There have been several studies of common gene variants in the general population in relation to myeloma risk. Lines of investigation

Multiple Myeloma have primarily pursued polymorphisms in genes regulating immune response and inflammation. Variations in human leukocyte antigen (HLA) were studied for many years prior to the ready availability of genotyping. Several studies have found associations between myeloma incidence and the B and C locus antigens. An analysis of pooled data on HLA-A and -B antigens showed a positive association between the presence of HLA B5 and the risk of myeloma (RR = 1.7, p < 0.05) (Ludwig and Mayr, 1982). An association of myeloma with HLA B18 antigen was observed in both the pooled analysis (OR = 1.4, p > 0.05) (Ludwig and Mayr, 1982) and a more recent study (OR = 6.3, 95% CI:1.0–39.7) (Patel et al., 2002). A case-control study of 46 black men and 85 white men found a strong association between the HLA-Cw2 antigen and myeloma incidence in both racial groups (blacks, OR = 5.7, 95% CI:1.0–7.2; whites, OR = 2.6, 95% CI: 1.0–7.2) (Pottern et al., 1992a), while two other studies noted associations with HLA-Cw2 that were more modest among African blacks (OR = 1.7, p > 0.05) (Patel et al., 2002) and European whites (OR = 1.5, p > 0.05) (Ludwig and Mayr, 1982). HLA-Cw5 and Cw6 antigens were associated with increased myeloma risk in blacks (Cw5, OR = 15.1, p = 0.001; Cw6, OR = 6.5, p = 0.007) (Leech et al., 1983). However, no association with Cw5 was observed among whites in another study (Ludwig and Mayr, 1982), Several polymorphisms in proinflammatory cytokine genes have been investigated, IL-6 is an essential growth and survival factor for myeloma cells, and a polymorphism at position -174 is known to be functionally significant (Jeffery and Mitchison, 2001). However, two studies of IL-6 (-174) polymorphism have found no differences in genotype frequencies between cases and controls (Dring et al., 2001; Zheng et al., 2000). TNF-a and IL-1 are inducers of IL-6 production, and IL-1b is mainly responsible for IL-6 production in the tumor environment. No differences have been observed in allele frequencies of the TNF-a (-308) (Zheng et al., 2000), IL-1b TaqI (Zheng et al., 2000), or IL-1Ra variable number tandem repeat (VNTR) polymorphisms (Demeter et al., 1996; Zheng et al., 2000) between myeloma cases and controls, or between MGUS patients and controls (Zheng et al., 2000). One study found that the haplotype of high-producer alleles TNF-a (-308) and lymphotoxin a (LT5.5/10.5) was associated with a twofold increased risk of myeloma (OR = 2.0, 95% CI:1.3–3.4) (Davies et al., 2000); similar patterns were observed for MGUS patients compared with controls. Myeloma and MGUS patients were compared with an ethnically matched control group for polymorphisms in the IL-10 gene, which is a gene implicated in growth and differentiation of normal B-cells, and in proliferation of myeloma cells (Zheng et al., 2001b). IL-10 production by stimulated peripheral blood mononuclear cells was significantly higher in subjects who were heterozygous or homozygous for the IL10.G allele 136 (Zheng et al., 2001b), demonstrating the functional significance of the variant. Myeloma risk was positively associated with the IL10.G genotype 136/136 (OR = 6.9, 95% CI:2.6–18.2), and the IL10.R genotype 112/114 (OR = 3.1, 95% CI:1.6–6.3), and negatively associated with the IL10.R genotype 114/116 (OR = 0.2, 95% CI:0.1–0.5). Similar patterns were observed in comparisons of MGUS patients with controls. Cytotoxic T lymphocyte antigen-4 (CTLA-4), which is involved in the regulation of immune responses including mediating inhibitory signals to activated T cells, has a microsatellite polymorphism in the 3¢ untranslated region of exon 3, which has been associated with an increased risk of certain autoimmune diseases (Kotsa et al., 1997). Of the multiple CTLA-4 genotypes, the only association observed was a presumably unpredicted inverse association of genotype 86/86 with a decreased risk of myeloma (OR = 0.5, 95% CI:0.2, 1.0) and MGUS (OR = 0.1, 95% CI:0.03, 0.6) (Zheng et al., 2001a). NF-kB proteins act as transcription factors for many genes, the products of which have important roles in cell proliferation, immune response, and inflammation. NF-kB activation is normally transient in most cells; during interim periods, NF-kB proteins remain bound to IkBa in the cytosol, and are thus unable to act as transcription factors. Increased NF-kB activity has been observed in myeloma cell lines and is much higher in chemo-resistant than chemo-sensitive cell lines. Because the presence of IkBa somewhat regulates NF-kB activity,

931

polymorphisms in the IkBa gene have been investigated for their functional significance and association with malignancies. Certain variants of the IkBa gene have been shown to render the protein incapable of interacting with NF-kB (Cabannes et al., 1999). In a small, hospital-based case-control study, increased risk of myeloma was associated with polymorphisms in the IkBa gene, including sites 104149 (A Æ T) (OR = 7.8, 95% CI:1.7–35.4), 101799 (T Æ C) (OR = 6.4, 95% CI:1.2–34.0), and 101675 (A Æ G) (OR = 11, 95% CI:1.9–64.6) (Parker et al., 2002). Because chromosomal abnormalities are common in lymphoproliferative disorders, another line of investigation has pursued variants in genes involved in maintaining error-free DNA. Given that folate availability is critical to DNA integrity, Gonzalez Ordonez et al. (2000) compared the frequency of polymorphisms in methylene tetrahydrofolatereductase (MTHFR), which plays a critical step in folate synthesis, between myeloma cases and controls. Variant alleles at nucleotides 677 and 1298 have been associated with a decrease in MTHFR activity (Skibola et al., 1999). The common MTHFR CC genotype, with homozygosity at nucleotide 677, was associated with a decreased risk of myeloma overall (19% vs. 46% in controls, OR = 0.3, 95% CI:0.1–0.8) (Gonzalez Ordonez et al., 2000). A second study observed a similar association with the CC genotype at MTHFR nucleotide 667 (34% in myeloma cases vs. 48% in controls, p = 0.18), but no strong association with the AA genotype at nucleotide 1298 (36% in cases vs. 43% in controls, p = 0.61) (Gonzalez-Fraile et al., 2002).

ENVIRONMENTAL FACTORS Ionizing Radiation Two studies have been published of myeloma risk in Japanese survivors of the atomic bombs detonated in Hiroshima and Nagasaki in 1945 (Table 47–4). The study of Shimizu et al. (1990) examined myeloma mortality during the period 1950 through 1985, whereas that of Preston et al. (1994) considered incidence from 1950 through 1987. The design and methods differed appreciably between the two studies, as did the findings. Shimizu et al. (1990) ascertained 36 persons whose primary cause of death was myeloma and for whom DS86 revised doses had been estimated among the approximately 75,000 persons who were in the cities at the time of the bomb. They observed a slope in the relation of radiation dose and myeloma risk such that the RR was 3.3 (90% CI:1.7–6.3) following exposure to 1 Gy bone marrow dose (1990). The mean bone marrow dose in the cohort was 0.14 Gy (corresponding RR of 1.18). Preston et al. (1994) identified 59 persons whose first cancer diagnosis was myeloma and whose DS86 kerma doses were estimated to be less than 4 Gy. After adjusting for age at diagnosis and age at the time of the bomb, they observed a relative risk of 1.3 (p > 0.05) per Sievert (Sv). Including persons whose myeloma was a second primary or whose doses exceeded 4 Gy increased the RR to 1.9 (p = 0.02). More than 99% of the radiation from the atomic bomb was reported as gamma radiation, so the dose in Sv would be approximately equal to the dose in Gy in this study. Radiation-exposed workers at three US nuclear weapons plants were found to be at increased risk of death from myeloma relative to their unexposed peers (Gilbert et al., 1989). A direct relation between dose and mortality was observed, with workers who received 0.050–0.099, 0.010–0.199, and 0.200 Sv or more of external radiation being at 3.3-fold (n = 1), 5.0-fold (n =1), and 33-fold (n = 1) increased risk, respectively, relative to the general population. A case-control study of multiple myeloma at four US nuclear facilities did not show an association with lifetime cumulative whole-body ionizing radiation and myeloma risk; however, there was an association between multiple myeloma and doses received at older ages (Wing et al., 2000). British radiation workers at major sites of the nuclear industry were at lower risk of death from myeloma than the general population, with a standardized mortality ratio (SMR) of 0.7 (number of exposed cases = 40), but there was a trend (p = 0.06) of increasing risk with increasing level of exposure such that workers exposed to 0.400 or more Sv (n = 2) were at 2.5-fold higher risk than those exposed to less than

Table 47–4. Summary of Studies That Have Assessed Risk of Multiple Myeloma in Relation to Ionizing Radiation Study

Study Location and Design

Study Period

Comparison

Number of Exposed Cases

Effect Estimate

95% Confidence Interval

a-bomb survivors Shimizu et al., 1990 Preston et al., 1994

Japan, cohort mortality

1950–1985

Japan, cohort incidence

1950–1987

nuclear workers Gilbert et al., 1989

Cardis et al., 1995

Muirhead et al., 1999

Kendall et al., 1992 Omar et al., 1999

Smith and Douglas, 1986

Carpenter et al., 1994

Three US nuclear weapons plants, cohort mortality

Canada, UK, US study, nuclear industry workers, cohort mortality

1943–1981

1944–1988

National Registry for Radiation 1976–1992 Workers, UK, cohort mortality

National Registry for Radiation 1976–1983 Workers, UK, cohort mortality Sellafield plant, UK, cohort, 1947–1992 mortality

Sellafield plant, UK, cohort, mortality

Combined analyses of three UK nuclear industry workforces, cohort mortality

1946–1988

UK Atomic Energy Authority, cohort mortality

1946–1979

Pearce et al., 1990

New Zealand, cohort mortality

1957–1987

Iwasaki et al., 2003

Nuclear industry workers in Japan

1986–1997

Three US nuclear weapons plants, case-control

1979–1990

932

23

3.3

1.7–6.3

59

1.3

—

8 1 1 1 1

0.9 0.4 3.3 5.0 33.0

— — — — —

28 3 1 5 3 2 2 40

1.1 0.6 0.2 1.9 1.4 1.1 2.5 0.7

0.7–1.5 0.1–1.7 0.0–1.2 0.6–4.3 0.3–1.2 0.1–3.8 0.3–9.0 0.5–1.0

20 4 3 8 0 3 2 17

1.0 0.8 0.4 2.3 0.0 1.8 2.5 0.7

0.6–1.5 0.2–2.1 0.1–1.3 1.0–4.6 0.0–1.5 0.4–5.2 0.3–9.1 —

0 0 2 3 1 0 2

0.0 0.8 1.3 2.5 0.9 0.0 2.0

0.0–16.1 0.0–4.6 0.2–4.8 0.5–7.3 0.0–5.1 0.0–3.4 0.2–7.2

0 0 2 2 1 0 2

0.0 0.0 1.5 1.8 1.0 0.0 2.2

0.0–3.7 0.0–6.1 0.2–5.6 0.2–6.6 0.0–5.6 0.0–27.6 0.3–8.0

6 1 2 4 2 0 2 8

1.0 0.5 0.6 1.9 1.4 0.0 1.9 0.8

— — — — — — — 0.4–1.6

0

0.0

0.0–3.1†

6 0 0 1 1

1.0 0.0 0.0 3.6 4.2

0.4–2.2 0.0–4.7 0.0–4.6 0.0–20.0 0.2–40.2

External dose ≥10 years earlier, Sv 0.001–0.009 vs. general population 0.010–0.049 0.050–0.099 0.100–0.199 ≥0.200 Nuclear industry workers vs. general population, cumulative dose, mSv 0–9 10–19 20–49 50–99 100–199 200–399 400+ Registered persons vs. general population, mSv <10 10–19 20–49 50–99 100–199 200–399 400+ Registered persons vs. general population External dose, mSV

<10 10–19 20–49 50–99 100–199 200–399 400+ 1947–1983 External dose, mSv

Beral et al., 1985

Wing et al., 2000

Increased risk per 1 Gy bone marrow dose Increased risk per Sv bone marrow dose

<10 10–19 20–49 50–99 100–199 200–399 400+ Employed by three UK nuclear industry workforces vs. general population, cumulative whole-body dose, mSv <10 10–19 20–49 50–99 100–199 200–399 400+ Employed by AEA vs. general population, median external dose <0.010 Sv Nuclear weapons test workers vs. unexposed workers; doses not estimated Cumulative dose categories, mSv <10 10–20 20–50 50–100 100+ Cumulative dose categories, cumulative dose at ages 45 and above, 5-year lag, mSv

Table 47–4. (cont.) Study

Study Location and Design

Darby et al., 1993

UK, case-control

Boffetta et al., 1989

American Cancer Society, US, nested case-control

Study Period

Comparison

<10 10–49 50–99 >100 1952–1990 Nuclear weapons test workers vs. unexposed workers; doses not estimated 1982–1986 Ever vs. never occupationally exposed to X-rays or radioactive materials

Number of Exposed Cases 83 5 3 7 6

Effect Estimate

95% Confidence Interval

1.0 (ref.) 0.8 3.6 5.2 1.5

— — — — 0.6–4.3

7

1.9

0.8–4.8

0

0.0

0.0–6.0

5

5.0

1.6–12.0

radiology workers Wang et al., 1988

China, cohort incidence

Lewis, 1963

US radiologists, cohort mortality US radiologic technlogists

Linet et al., 2005

1950–1980

Diagnostic X-ray workers vs. general population 1948–1961 Employed as radiologist vs. general population 1926–1998 Number of years worked as radiologic technologist <10 10–19 ≥20

therapeutic irradiation Boice et al., 1985

International, cohort incidence

—

Darby et al., 1987

England and N. Ireland ankylosing spondylitits patients, cohort mortality International, case-control study of cervical cancer patients

1935–1983

Boice et al., 1988

Boffetta et al., 1989 Darby et al., 1994 Eriksson et al., 1993 Flodin et al., 1987 Friedman, 1986

See above Scotland, metropathia hemorrhagica patients cohort mortality Northern Sweden, populationbased case-control Southeast Sweden, populationbased case-control Kaiser Permanente, northern California, population-based case-control

—

Exposed to cervical radiation ≥15 yr vs. unexposed; average bone marrow dose approximately 10 Gy Single course of X-ray treatment ≥5 yr earlier vs. none; skeletal dose approximately 3 Gy Bone marrow dose:

2–4 vs. <2 Gy 5–9 vs. <2 Gy ≥10 vs. <2 Gy X-ray treatment vs. none 1940–1986 Radiographic treatment 5 years earlier vs. general population, mean bone marrow dose = 1.3 Gy 1982–1986 Ever vs. never received radiotherapy 1973–1983

X-ray treatment vs. none

7 9 10

1.0 (ref.) 1.3 0.8

0.5–3.6 0.4–2.0

33

2.0

1.1–3.2

8

1.7

—

12 23 11 14 9

0.3 0.2 0.6 1.6 2.6

0.0–2.6† 0.0–1.4† 0.1–5.2† 0.8–3.0 1.2–4.9

10

0.7

0.3–1.8

4

0.9

0.3–2.7

‡

1969–1982

Ever vs. never exposed to X-ray therapy

9

1.9

0.9–4.2

Throtrast injection vs. general population Ever vs. never received X-ray fluoroscopy examination; mean dose = 0.09 Gy Diagnostic X-rays, above vs. below median number, median not reported Diagnostic X-rays <5 5–9 10–19 20+ Bone marrow dose >2 years earlier, exposed to ≥0.01 Gy 0.01 vs. 0 Gy 0.02 vs. 0 Gy 0.03 vs. 0 Gy 0.04 vs. 0 Gy Number of diagnostic X-rays:

4

4.6

1.2–12.0

2

0.4

0.1–1.8

62

0.9

0.6–1.4

1.0 (ref.) 0.9 1.0 0.9

— 0.7–1.2 0.7–1.3 0.7–1.2

diagnostic radiation Andersson and Storm, 1992 Davis et al., 1989

Danish neurology patients, cohort mortality Massachusetts tuberculosis patients, cohort mortality

1946–1988

Boffetta et al., 1989

See above

See above

Hatcher et al., 2001

Three areas of US

1986–1989

Boice et al., 1991

Cuzick and De Stavola, 1988

1925–1986

Kaiser Permanente, population- 1956–1982 based case-control

England & Wales, hospitalbased case-control

1978–1984

Eriksson, 1993

See above

See above

Flodin et al., 1987

See above

See above

1–4 vs. 0 5–8 vs. 0 ≥9 vs. 0 5 vs. £4 6–10 vs. £4 11–20 vs. £4 ≥21 vs. £4 Heavy vs. light X-ray examination 10–30 years earlier

106 104 133 137 198

1.3 1.5 1.3 3.9 86 79 181 65 29 20 16 2

0.8** 0.7** 0.7** 0.6 0.5 0.8 0.9 2.9

— — — — — — — 0.3–11 0.3–0.9 0.4–1.5 0.4–2.1 0.4–19.0

*Studies of ionizing radiation use various units of dose, including Sievert (Sv) and Gray (Gy), which are Standard International units, and rem (1 Sv = 100 rem) and rad (1 Gy = 100 rad). Sv is related to Gy as follows: Sv = Gy ¥ Q, where Q is the “quality factor” (i.e., the biological potency of the specific type of radiation relative to orthovoltage X-rays); although not standardized, gamma rays are often assigned a Q of 1, and fast neutrons a Q of 10. Dose can be expressed as external radiation dose (whole-body dose, shielded kerma) or organ dose. † 90% confidence interval. ‡ No. of discordant pairs in which only the case was exposed. **Unadjusted odds ratio.

933

934

PART IV: CANCER BY TISSUE OF ORIGIN

0.010 Sv (n = 20) (Kendall et al., 1992; Muirhead et al., 1999). A trend with dose for multiple myeloma was also observed in the IARC Canada/UK/US study (one-sided p value: 0.037) (Cardis et al., 1995). A study conducted at the Sellafield nuclear fuels plant, in which there were seven deaths from myeloma, also showed a direct relation between dose and mortality, with workers receiving an external dose of 0.400 mSv or more being at 2.0-fold (n = 2) risk relative to the general population (Smith and Douglas, 1986). Same findings were observed when the follow-up of the study population was extended (Omar et al., 1999). A study of persons exposed occupationally to considerably lower levels of radiation (<0.01 Sv) observed no excess risk (Beral et al., 1985). A combined analyses of mortality in three UK nuclear industry workforce showed a modest association between external radiation dose and risk of multiple myeloma (Carpenter et al., 1994). British workers who participated in atmospheric nuclear weapons testing were at 50% increased risk of death from myeloma (90% CI:0.55–4.26); their radiation doses were not estimated (Tomasek et al., 1993). New Zealand subjects in the same testing were not at increased risk (SIR = 0 vs. 5.56 in unexposed workers, 90% CI:0.0–3.1) (Pearce and Reif, 1990). Czech uranium miners with cumulative radon exposures of 210–329 and 330 or more “working level months” were at 1.9-fold (n = 1) and 4.4-fold (n = 2) increased risk of myeloma (Tomasek et al., 1993). In addition, an American Cancer Society nested case-control study noted a 1.9-fold increased risk of mortality (95% CI:0.8–4.8) following occupational exposure to X-rays and radioactive materials (Boffetta et al., 1989). Three cohort studies of medical radiology workers have been conducted. In China, no myeloma cases were diagnosed, though only 0.5 were expected, among 27,000 radiology workers (Wang et al., 1988). In the United States, physicians listed by the American Board of Radiologists in 1950 or 1960 were at a five-fold increased risk of dying from myeloma (95% CI:1.6–12) (Lewis, 1963). In a study of US radiologic technologists, although no obvious overall association was observed with myeloma risk and duration of employment, nonsignificant 40–90% increases of myeloma were found among radiologists working five or more years before 1950, those first employed at age 20 or younger, and those in other subgroups with greater possibility of exposure (e.g., those who did not use a lead apron when first started working) (Linet et al., 2005). Doses received during therapeutic irradiation can be high. However, both therapeutic and diagnostic irradiation are directed to a focal area, whereas the doses received by the atomic bomb survivors and by occupational cohorts were relatively uniform over the body. Three cohort studies and five case-control studies of the relation between myeloma and therapeutic irradiation have been reported. Risk of myeloma was increased among women who were estimated to have received a mean bone marrow dose of 10 Gy during therapy for cervical cancer; after 15 years, there was a 2.0-fold increased risk (95% CI:1.1–3.2) (Bocie, Jr. et al., 1985). Boice et al., (1998) also conducted a nested casecontrol study using the cervical cancer cohort described above, along with additional subjects, to obtain more precise information on radiation exposures. In the case-control study, no excess risk was noted among women who received an average marrow dose of 7.1 Gy compared with those who received less than 2.0 Gy (OR = 0.3, 90% CI:0.1–1.4); no relation was observed even after excluding women who had been followed for less than 15 years. Furthermore, the investigators found no evidence that risk increased with increasing radiation dose above 2 Gy. The authors stated that the only differences between the case-control and cohort studies were: 1. The number of women with myeloma—34 in the cohort and 46 in the case-control study 2. the comparison group—population rates were used for the cohort study, whereas cervical cancer patients who did not undergo radiotherapy were used for the case-control study. Compared with individuals who had never undergone X-ray therapy, myeloma mortality was modestly elevated in a cohort of patients receiving a single course of X-ray treatment for ankylosing spondylitis (RR = 1.7, n = 8) (Darby et al., 1987) or metropathia hemorrhagica (RR = 2.6, 95% CI:1.2–4.9) (Darby et al., 1994), among American Cancer Society subjects who had received X-ray treatment

for any condition (OR = 1.6, 95% CI:0.8–3.0) (Boffetta et al., 1989), and among Kaiser Permanente members who had received X-ray treatment for any condition (OR = 1.9, 95% CI:0.9–4.2) (Friedman, 1986), but not among subjects who reported a history of radiotherapy for any condition in two Swedish population-based case-control studies, however (OR = 0.9, 95% CI:0.3–2.7) (Flodin et al., 1987); OR = 0.7, 95% CI:0.3–1.8 (Eriksson, 1993). Compared with the general population, myeloma incidence was elevated 4.6-fold (95% CI:1.2–12) among 1095 Danish patients exposed to alpha-emitting Thorotrast used with cerebral arteriography (Andersson and Storm, 1992). A single, diagnostic X-ray received by Massachusetts tuberculosis patients did not increase their risk of myeloma (RR = 0.4, 95% CI:0.1–1.8) (Davis et al., 1989). Having a relatively large number of diagnostic X-rays (more than the median) did not increase myeloma risk among American Cancer Society members (OR = 0.9, 95% CI:0.6–1.4) (Boffetta et al., 1989), nor did having nine or more diagnostic X-rays in a UK case-control study (unadjusted OR = 0.7 relative to having no exams, no. of exposed individuals = 81) (Cuzick and De Stavola, 1988). In a Swedish case-control study conducted between 1973 and 1983, individuals who received “heavy” levels of X-ray examinations were at 2.9-fold increased risk of myeloma (95% CI:0.4–19) relative to individuals who received “light” levels (Flodin et al., 1987). In a later study set in northern Sweden risk was reduced among persons who reported a history of five or more X-rays compared with those who reported a history of four or fewer; however, risk increased with increasing numbers of examinations such that exposure to 21 or more resulted in an odds ratio of 0.9 (95% CI:0.4–2.1), compared with the odds ratio of 0.5 (95% CI:0.3–0.9) for exposure to 6 to 10 examinations (Eriksson, 1993). At Kaiser Permanente, bone marrow doses of 0.04 Gy or more were related to a 3.9-fold risk of myeloma, and there was some evidence that increasing doses were related to increasing risk (Table 47–4) (Boice, Jr. et al., 1991). The authors remarked that categorizing the data on the basis of number of procedures resulted in substantial misclassification; for example, five or fewer X-ray procedures led to a range in the bone marrow dose of 0.00001–0.03 Gy, whereas the range for 15 or more X-ray procedures was 0.001–0.23 Gy. The discordance between the number of prior X-ray examinations and bone marrow dose was caused by differences in procedures; for example, an upper gastrointestinal procedure contributed 60 times more radiation to the bone marrow than a chest roentgenogram (Boice, Jr. et al., 1991). Failure to take into account this source of misclassification may account for the negative findings in some studies (Boffetta et al., 1989; Cuzick and De Stavola, 1988; Eriksson, 1993). In a population-based case-control study in three US areas, there was no significant association between myeloma and the total number of reported X-rays of any type (OR = 2.0, 95% CI:0.7–1.2) (Hatcher et al., 2001). There was no evidence of excess risk of myeloma among individuals who reported exposure to 10 or more diagnostic X-rays with high bone marrow dose, compared with individuals reporting no such exposures (OR = 0.7, 95% CI:0.4–1.3). There is evidence from studies of atomic bomb survivors, occupational groups with exposure to approximately 0.05 Sv or more, persons exposed to therapeutic radiation, and from a carefully conducted study of diagnostic procedures (Boice, Jr. et al., 1991) to suggest that ionizing radiation causes myeloma. It has also been reported that administration of ionizing radiation to rhesus monkeys was followed by increased incidence of MGUS and myeloma (Radl et al., 1991). However, there are inconsistencies in the body of evidence, for example from the two studies of the atomic bomb survivors and from the two studies of women who received radiotherapy for cervical cancer. There are also differences in the reported dose-response relations, but this could be related to differences in the quality and timing of exposure.

Occupational Exposures Because multiple myeloma is a relatively rare cancer, cohort studies of the relation between occupational exposures and myeloma have generally provided limited information. The only cohort studies in which there were large numbers of myeloma cases have been those

Multiple Myeloma using publicly-available databases such as death certificates; in such studies job title was generally the only occupational variable that could be reliably evaluated. Respondents in most interview-based case-control studies were asked whether they had ever been exposed to specific chemical and physical agents. However, it is possible that they knew the agents by different names, and self-reported occupational exposures are also subject to recall bias. The case-control studies were generally of limited power as well, in that only a small number of subjects had worked in the occupations and industries of interest. An exception is the case-control study reported by Heineman et al. (1992) and Pottern et al. (1992b); it was based on work histories recorded in a national pension plan and included approximately 800 men and 600 women with myeloma. Characteristics of case-control studies that have evaluated the relation between occupation and myeloma are summarized in Table 47–5. Several studies have reported on cosmetologists and hairdressers as well; they are discussed in a later section concerning personal and occupational exposure to hair coloring products.

Agricultural Work and Pesticides Numerous cohort (Cerhan et al., 1998; Lee et al., 2002; Nandakumar et al., 1988; Pukkala and Notkola, 1997; Stark et al., 1990; Steineck and Wiklund, 1986) and case-control studies (Alavanja et al., 1988; Baris et al., 2004; Boffetta et al., 1989; Brownson et al., 1989; Burmeister et al., 1983; Cantor and Blair, 1984; Costantini et al., 2001; Cuzick and De Stavola, 1988; Demers et al., 1993; Eriksson and Karlsson, 1992; Figgs et al., 1994; Flodin et al., 1987; Franceschi et al., 1993; Gallagher et al., 1983; Heineman et al., 1992; La Vecchia et al., 1989; Mester et al., 2006; Milham, 1971; Miligi et al., 1999; Nandakumar et al., 1986; Pahwa et al., 2003; Pasqualetti et al., 1990; Pearce and Howard, 1986; Pottern et al., 1992b; Reif et al., 1989a; Svec et al., 2005; Tollerud et al., 1985) of agricultural work in relation to myeloma occurrence have been conducted and most have observed relative risks exceeding 1.1 (Table 47–6), although some to only a very limited extent. One study that did not observe a positive relation was that of Tollerud et al. (1985), which was one of several that ascertained occupational information from death certificates. A study in Italy found no increased risk for farmers, but did find a slight increased risk (OR = 1.3) for agricultural and animal husbandry workers (Costantini et al., 2001; Miligi et al., 1999). No increased risk was associated with farming occupation among women in Denmark (Pottern et al., 1992b). However, any work in the agricultural products industry was associated with myeloma (OR = 1.5). Several studies considered duration of employment as a farmer. A trend of increasing relative risk with duration was noted by Demers et al. (1993) (less than 10 years, OR = 1.1; 10 years or longer, OR = 1.3), Alavanja et al. (1998) (less than 15 years, OR = 0.2; 15 years or longer, OR = 2.6), and Boffetta et al. (1989) (20 years or less, OR = 0.0; 21–40 years, OR = 1.7; longer than 40 years, OR = 4.3), but not by Heineman et al. (1992) (employment in agricultural industry: 5 years or less, OR = 1.1; more than 5 years, OR = 0.8) or Mester et al. (10 years or less, OR = 10.4; longer than 10 years, OR = 8.6) (Mester et al., 2006). Following the observation that agricultural workers were at increased risk of myeloma, exposure to pesticides (i.e., insecticides, herbicides, fungicides) was hypothesized as the basis for the association. Two case-control studies that sought to estimate the risk of myeloma in relation to agricultural work and pesticide exposure as independent factors, observed that myeloma risk was greater for individuals exposed to both factors than for individuals exposed to either factor alone. Boffetta et al. (1989) reported the following results for comparisons using persons with neither exposure as the referent: pesticides alone, OR = 1.0; farming alone, OR = 1.7; both pesticides and farming, OR = 4.3. The results of Demers et al. (1993) were as follows: pesticides alone, OR = 2.1; farming alone, OR = 1.4; both pesticides and farming, OR = 7.9 (1993). At an ecologic level, Burmeister et al. (1983) found that farming occupation was more strongly associated with myeloma mortality among residents of 33 Iowa counties with the highest herbicide and insecticide usage, compared with the association among the lowest usage counties (for mortality from myeloma associated with farming occupation for men born after the year 1990:

935

highest herbicide usage counties, OR = 2.4, p < 0.05; lowest herbicide usage counties, OR = 1.2; highest insecticide usage counties, OR = 2.0, p < 0.05, lowest insecticide usage counties, OR =1.2). Most of the studies of myeloma risk in relation to pesticide exposure had very limited exposure assessment, giving few leads to specific pesticides as potential risk factors. In a cohort study covering the period from 1965 to 1982 in which licensed Swedish agricultural pesticide applicators were compared with agricultural workers who had not been exposed to pesticides, Wiklund et al. (1989) did not find an increased risk of myeloma (OR = 1.0, 95% CI:0.5–1.9). Conversely, myeloma mortality was elevated by a factor of 8.2 (95% CI:1.6–23) in a cohort of Dutch licensed applicators who only applied herbicides (Swaen et al., 1992). The most heavily used herbicides to which the applicators were exposed during the time period of interest were simazine, chlorothiamide, dalapon, dichlorbenil, and diuron. Myeloma mortality was significantly increased in a cohort of workers who applied the insecticide DDT during anti-malarial campaigns in Italy between 1956 and 1992 (PMR = 3.4; 95% CI:1.1–8.0) (Cocco et al., 1997), although no association was found in an ecological study of myeloma mortality and DDT residues in adipose tissue among the general US population from 22 states (Cocco et al., 2000). One study estimated the risk of myeloma associated with specific pesticide classes in a multivariable model, while adjusting for other pesticides and farm animal exposures (Eriksson and Karlsson, 1992). Myeloma risk was associated with exposure to pheoxyacetic acids (OR = 1.9, 95% CI:0.8–4.4) and DDT (OR = 1.4, 95% CI:0.9–2.3), but not chlorophenols (OR = 0.9, 95% CI:0.4–1.8). An association with phenoxyherbicides was also observed among New Zealand phenoxyherbicide producers and sprayers who had elevated myeloma mortality (SMR = 5.5, 95% CI:1.1–16.1) (’t Mannetje et al., 2005). Evidence for a positive association between pesticide exposure and myeloma was also provided by several case-control studies that did not take into account potential confounding by other aspects of agricultural work (Baris et al., 2004; Burmeister, 1990; Flodin et al., 1987; Morris et al., 1986; Pasqualetti et al., 1990; Pottern et al., 1992b). Besides the cohort study of Wiklund et al. (1989), other studies that observed no relation between a history of pesticide exposure and myeloma risk were those of La Vecchia et al. (1989) (unadjusted OR = 0.9; no. of exposed cases = 4) Pearce et al. (1986) (phenoxyherbicides, OR = 1.3, 95% CI:0.8–2.5; chlorophenols, OR = 1.1, 95% CI:0.4–2.7), Heineman et al. (1992) (OR = 0.9, 95% CI:0.5–1.5) and Pahwa et al. (herbicides, OR = 0.9, 95% CI:0.6–1.4; insecticides, OR = 1.2, 95% CI:0.7–2.2) (Pahwa et al., 2003). When interpreting results from these studies, it should be noted that dioxin (2,3,7,8-tetrachlorodibenzo-p-dioxin) has been a chemical contaminant in certain formulations of several commonly used herbicides, including phenoxyherbicides. In Seveso, Italy, where an industrial accident in 1976 exposed the local population to high levels of dioxin, myeloma incidence in the subsequent 10 years was elevated in both men (RR = 3.2, 95% CI:0.8–13.3) and women (RR = 5.2, 95% CI:1.2–22.6) (Bertazzi et al., 1993). Because dioxin has been a chemical contaminant of herbicides, and has been associated with myeloma risk, epidemiologic studies of myeloma in relation to herbicide exposures should consider the possibility of dioxin as an etiologic agent, in addition to the herbicide active ingredients. Besides pesticides, other specific exposures common among farmers include paints, solvent, wood-treatment chemicals used for fencing, engine exhaust, welding fumes, dusts, animals, zoonotic infections, and pollen (Blair et al., 1985; Pearce and Reif, 1990). Efforts to identify myeloma risk in relation to animal exposures have been made in several studies (Baris et al., 2004; Eriksson and Karlsson, 1992; Pearce et al., 1986; Reif et al., 1989a; Svec et al., 2005). In an analysis of death certificate data, the risk of myeloma mortality associated with livestock industry farming (SMR = 1.5, 95% CI:1.4–1.7) significantly exceeded that of crop industry farming (SMR = 1.2, 95% CI:1.1–1.3) (Svec et al., 2005). In the study of Pearce et al. (1986) conducted in New Zealand, in which 36% of the controls reported having a history of employment as a farmer, the magnitude of the relation between farming and myeloma varied somewhat by type of farming (categories not mutually exclusive), as follows: orchard farmers (OR = 2.8), produce farmers (OR = 2.0), sheep

Table 47–5. Design Characteristics of Case-Control Studies That Have Assessed Risk of Multiple Myeloma in Relation to Job Title, Industry, or Specific Chemical and Physical Agents (Excluding Occupations and Exposures Related to Ionizing Radiation) Study Alavanja et al., 1988 Baris et al., 2004

Bethwaite et al., 1990 Boffetta et al., 1989

Brownson et al., 1989

Study Population and Period Death certificates of persons who were employed at the US Department of Agriculture, 1970–1979 General population, three SEER areas, US, 1986–1989

Persons registered with the New Zealand tumor registry, 1981–1984 American Cancer Society cohort, US, 1982–1986

Burmeister et al., 1983

Persons registered with the Missouri tumor registry, US, 1984–1988 Death certificates, Iowa, US, 1964–1978

Cantor and Blair, 1984

Death certificates, Wisconsin, US, 1968–1976

Costantini et al., 2001

General population, Italy, 1991–1993

Cuzick and De Stavoa, 1988

Hospital patients, six area in England and Wales, 1978–1984

Demers et al., 1993

General population, four SEER areas, US, 1977–1981

Eriksson and Karlsson, 1992

General population, northern Sweden, 1982–1986

Flodin et al., 1987

General population, southeast Sweden, 1973–1983

Franceschi et al., 1993

Hospital patients, northeast Italy

Friedman, 1986

Kaiser Permanente members, northern California, US, 1969–1982

Fritschi et al., 2002

Myeloma cases recorded in the National Enhanced Cancer Surveillance System; General population controls, US, 1994–1998 Patients admitted to a Vancouver, BC cancer center, 1972–1981 General population, Denmark, 1970–1984

Gallagher et al., 1983 Heineman et al., 1992

936

Ascertainment of Exposure

Study

Employment as an Kawachi agricultural extension et al., 1989 agent ascertained from work records La Vecchia Work history categorized by et al., 1989 occupational title and industry, and farming exposures assigned using a job exposure matrix Linet et al., Most recent occupation at 1987 time of registration Most recent occupation, occupation held for the longest period, and occupational and leisure-time exposure to 12 groups of agents Most recent occupation at time of registration Most recent occupation as recorded on the death certificate Most recent occupation as recorded on the death certificate Occupational history ascertained during interview Occupational history and occupational exposure to various agents ascertained during interview Work history categorized by job title and industry ascertained during interview Occupational history and occupational exposure to various agents ascertained during interview Occupational history and occupational exposure to various agents ascertained from a mailed questionnaire Occupational history ascertained during interview Occupation recorded on the medical chart 6 months or earlier prior to the reference date Occupational history ascertained by mailed questionnaire

Occupational history ascertained during interview Industrial history recorded since 1964 in the nationwide pension fund program; occupation recorded on the most recent tax records; job-exposure matrix developed by an industrial hygienist

Study Population and Period Persons registered with the New Zealand tumor registry, 1980–1984 Hospital patients, Milan, 1983–1988

Hospital patients, Baltimore, Maryland, US, 1975–1982

Mester et al., 2006

General population, six regions of Germany, 1999–2002

Milham, 1971

Death certificates, Washington and Oregon, US, 1950–1967 General population, Italy, 1991–1993

Miligi et al., 1999 Morris et al., 1986

General population, four SEER areas, US, 1977–1981

Nandakumar et al., 1986

Death certificates, Western Australia, 1975– 1984 General population myeloma cases and controls identified from public records, Canada, 1971–1991 Hospital patients resident in referral basin of Aquila and Avezzano, Italy, 1970–1988 Persons registered with the New Zealand tumor registry, 1977–1981

Pahwa et al., 2003 Pasqualetti et al., 1990 Pearce et al., 1986 Pottern et al., 1992

Women from the general population, Denmark, 1970–1984

Reif et al., 1989a

Persons registered with the New Zealand tumor registry, 1980–1984 General population, four SEER areas, US, 1977–1981

Schwartz et al., 1988

Svec et al., 2005

Death certificates from 24 US States, 1984–1998

Tollerud et al., 1985

Death certificates, North Carolina, US, 1956–1980

Wong et al., 1995

Members of a cohort of petroleum workers, US

Ascertainment of Exposure Most recent occupation at time of registration Occupational history and occupational exposures to various agents ascertained during interview Occupational history and occupational exposure to various agents ascertained during interview Occupational history including job tasks, obtained by interview Most recent occupation as recorded on the death certificate Occupational history ascertained during interview Formation of ever/never categories of exposure to various agents from responses to specific questions about a history of “high” exposures Most recent occupation as recorded on the death certificate Occupational history and specific exposures obtained by questionnaire Occupational history and ever exposure to classes of chemicals ascertained during interview Occupational history and occupational exposure to pesticides ascertained during interview Industrial history recorded since 1964 in the nationwide pension fund program; occupation recorded on the most recent tax records; job-exposure matrix developed by an industrial hygienist Most recent occupation at time of registration Probability and intensity of exposure to low, medium, and high levels of asbestos based on a job-exposure matrix from an occupational history Usual occupation as recorded on the death certificate, and use of a JEM to code exposures to animals and the public Most recent occupation as recorded o the death certificate Estimated exposure to total hydrocarbons based on historical records

937

Multiple Myeloma Table 47–6. Summary of Studies That Have Assessed Risk of Multiple Myeloma in Relation to Agricultural Work Study

Measure of Exposure

Prevalence of Exposure in Controls (%)

Number of Exposed Cases*

—

—

Odds Ratio or Relative Risk (95% Confidence Interval)

cohort studies Cerhan et al., 1988 Lee et al., 2002

Nandakumar et al., 1988 Pukkala and Notkola, 1997 Stark et al., 1990 Steineck and Wiklund, 1986 Vagero and Persson, 1986

Farmer as usual occupation listed on the death certificate, Iowa residents Farmer occupation in the crop or liverstock industry as listed in the National Occupational Mortality Surveillance data system Farming occupation as recorded in the cancer registry Farmers registered in the Farm Register of Finland Membership in New York State Farm Bureau Agricultural occupations Self-employed in agriculture

—

746 (crop) 186 (livestock)

—

15

—

—

— — —

11 O/ 9 E 568 / 4330717 PY 347

— 0.8 (farmers) 3.4 (farmworkers) 5.5 11 — 21.7 — — 4.5 10.3 3.3 8.2 10.8 47.3

7 7 18 16 24 550 110 22 30 28 72 26 57 19 151 5 29 20 31 84 98 4 3 5 45 10 — 5 60 12 21 15 21 95 44 43 18 23 11 5 1 2 14 14 11 54 — — — — 1147 360

1.2 (1.0–1.4) 1.1 (1.0–1.2) 1.2 (1.0–1.4) 1.3 (0.9–1.9) 1.0 (0.8–1.1), men 0.9 (0.7–1.2), women 1.1† 1.2 (1.1–1.3) 1.2 (1.0–1.3)

case-control studies Alavanja et al., 1988 Baris et al., 2004

Agricultural extension agents Farmers and farmworkers, general

Boffetta et al., 1989 Brownson et al., 1989 Burmeister et al., 1983 Cantor and Blair, 1984 Costantini et al., 2001

Farmers Farmers Farmers Farmers Farmers Agricultural and animal husbandry workers Farmers Food processing/agricultural industry Farmers Farmworkers and gardeners Agricultural industry Farmers Farm worker supervisor Farmers Farmers Farmworkers Farm owners Tenant farmers or agricultural garden workers Herdsmen Nurserymen Other specialty farmworkers Agricultural industry Truck garden/orchard/nursery industry Agricultural occupations Farmers Farmers Farmers Agricultural and animal husbandry workers Orchard, vineyard, and related workers Farmers Farmers Agricultural workers Farmers Sheep farmers Dairy farmers Livestock farmers Crop farmers Poultry farmers Orchard farmers Farm/land owner, farmer Agricultural products industry Orchards/nurseries industry Farmers General farmers Dairy farmers Livestock farmers Orchard and crop farmers Farmer occupation in the crop or livestock industry as listed on US death certificates from 24 states Farmers

Cuzick and De Stavola, 1988 Demers et al., 1993 Eriksson and Karlsson, 1992 Figgs et al., 1994 Flodin et al., 1987 Franceschi et al., 1993 Gallagher et al., 1983 Heineman et al., 1992

La Vecchia et al., 1989 Mester et al., 2006 Milham, 1971 Miligi et al., 1999 Nandakumar et al., 1986 Pahwa et al., 2003 Pasqualetti et al., 1990 Pearce et al., 1986

Pottern et al., 1992 Reif et al., 1989a

Svec et al., 2005 Tollerud et al., 1985

12.3 16.4 22.6 10.0 10.4 0.3 0.3 0.1 5.3 1.0 — — 7.3 — — — 11.4 15.2 18.8 35.9 5.7 7.3 3.5 1.0 0.3 0.6 0.7 2.7 2.0 15.1 — — — — 3.1 (crop) 0.8 (livestock) 19

*For cohort studies, the number of person-years (PY), or observed (O) and expected (E) cases are presented, where available. † Unadjusted odds ratio.

39

1.1 (0.4–2.9) 1.9 (0.8–4.6) 1.4 (0.8–2.3) 3.4 (1.5–7.5) 1.4 (0.9–2.2) 1.5† 1.4 (1.0–1.8) 0.7 (0.5–1.2) 1.3 (0.8–2.2) 1.6† 1.8† 1.2 (0.8–2.5) 1.0 (0.8–1.6) 1.2 (1.0–1.9) 1.7 (1.2–2.3) 2.5 (0.8–7.3) 1.4 (0.8–2.5) 1.3 (0.7–2.3) 2.2 (1.2–4.0) 1.0 (0.8–1.3) 1.1 (0.9–1.5) 1.8 (0.5–6.6) 1.3 (0.3–5.5) 4.6 (1.1–20.3) 1.1 (0.8–1.5) 1.3 (0.6–2.8) 2.0 (1.1–3.5) 9.2 (2.3–33.1) 1.8† 0.7 (0.4–1.5) 1.3 (0.7–2.3) 1.8 (0.9–3.5) 1.4 (0.8–2.5) 1.4 (1.0–1.9) 2.7 (1.9–4.4) 1.7 (1.0–2.9) 1.9 (1.0–3.6) 1.4 (0.8–2.5) 1.3 (0.6–2.6) 2.0 (0.6–6.0) 0.9 (0.1–8.4) 2.8 (0.5–16.9) 0.9 (0.3–3.0) 1.5 (0.8–2.8) 1.5 (0.7–3.2) 1.2 (0.9–1.6) 1.3 (0.9–1.7) 1.5 (0.6–3.3) 0.9 (0.3–2.8) 0.5 (0.1–1.9) 1.2 (1.1–1.3) 1.5 (1.4–1.6) 0.6†

938

PART IV: CANCER BY TISSUE OF ORIGIN

farmers (OR = 1.9), mixed sheep and beef farmers (OR = 1.3), dairy farmers (OR = 1.4), and poultry farmers (OR = 0.9). Rather different results were obtained in another study conducted in New Zealand, for orchard and crop farmers (OR = 0.5), livestock farmers (OR = 0.9), and dairy farmers (OR = 1.5) (Reif et al., 1989a). Baris et al. (2004) found the strongest association among sheep farm residents or workers (OR = 1.7, 95% CI: 1.0–2.7). Several studies estimated risk associated with self-reported exposure to specific types of animals (despite occupational classification) (Pearce et al., 1986, Eriksson and Karlsson, 1992, Pahwa et al., 2003, Fritschi et al., 2002). Results from these studies were mixed, indicating that further investigation is needed to disentangle the contribution of animal exposures to the increased risk of myeloma in agricultural occupations.

Benzene and Other Organic Solvents Rinsky et al. (2002) developed a job-exposure matrix for benzene using industrial hygiene data from an Ohio rubber plant and found a 2.5-fold increased risk of death from myeloma (95% CI:0.7–4.8) associated with employment in benzene-exposed jobs. All of the myeloma cases were diagnosed 20 years or later after first exposure, but no patterns of risk with cumulative level of benzene exposure were apparent. Similar results of increased myeloma mortality (SMR = 2.3; 95% CI:0.7–5.3) were observed in a cohort of chemical plant workers with low levels of benzene exposure, but there was no observed increased risk for maintenance workers suspected to have the highest benzene exposures (Ireland et al., 1997). No excess in myeloma incidence was observed in a large cohort of benzene-exposed workers in China (Hayes et al., 1997). There was also no association between occupational benzene exposure and myeloma occurrence in two case-control studies (Heineman et al., 1992; Linet et al., 1987), with exposures assigned based on self-report or according to occupational title or industry. The broad grouping of organic solvents, which includes benzene, has been investigated in two meta-analyses of cancer risk. In a metaanalysis of 55 cohort mortality studies across different industries with probable exposure to organic solvents, there was only a slightly increased risk for death from myeloma (SMR = 1.1, 95% CI:0.8–1.6) (Chen and Seaton, 1996). A meta-analysis of case-control studies published from 1986 through 1994 found a slightly decreased risk of myeloma associated with exposure to organic solvents (OR = 0.7; 95% CI:0.6–0.9) (Sonoda et al., 2001).

Petroleum Refining and Distribution Known carcinogens to which petroleum workers are exposed include polycyclic aromatic hydrocarbons and various solvents, which in the past may have included benzene. A meta-analysis of 22 cohort studies found that petroleum workers were at no increased risk of death from myeloma (SMR = 0.9; 95% CI:0.8–1.1) (Wong, 1995). A cancer incidence study of a cohort of Australian petroleum industry workers found a 2.2-fold (95% CI:0.6–5.6) increased risk for all men employed 5 years or longer in any department (Christie et al., 1991). The case-control study of Cuzick and De Stavola (1988) noted that four cases but no controls had worked in the petroleum industry; however, three other case-control studies found no evidence for a relation (Demers et al., 1993; Eriksson and Karlsson, 1992; La Vecchia et al., 1989). Self-reported occupational exposure to petroleum was associated with increased myeloma risk in one case-control study (OR = 3.7, 95% CI:1.3–10.3) (Linet et al., 1987).

Rubber and Plastics Manufacturing Rubber workers can be exposed to organic solvents, plastic monomers, rubber additives, and asbestos, and in the past, exposure to benzene was relatively high. There was a greater than expected number of myeloma deaths in several cohorts of rubber workers, but these elevations were based on small numbers of cases (Andjelkovich et al., 1978; Divine and Hartman, 2001; Gustavsson et al., 1986; Wilczynska et al., 2001), and no association was found in a cohort of rubber workers in Great Britain followed during 1946 to 1985 (Sorahan and Cooke, 1989) or in US and Canadian rubber workers followed from 1943 to 1998 (Sathiakumar N et al., 2006). Similarly, a meta-analysis of six case-control studies that evaluated myeloma risk associated with

employment in workplaces producing rubber and/or plastic products (Demers et al., 1993; Figgs et al., 1994; Flodin et al., 1987; Heineman et al., 1992; Morris et al., 1986; Pottern et al., 1992b) found no association (OR = 1.1, 95% CI:0.9–1.3) (Sonoda et al., 2001).

Paint-Related Occupations Painters are exposed to dyes and pigments, aromatic and aliphatic hydrocarbons, and low molecular weight solvents such as trichloroethylene and methylethyl ketone (Bethwaite et al., 1990). The majority of epidemiologic studies that have assessed it provide evidence for a relation between a history of paint-related occupation and myeloma risk. A cohort study of Swedish production workers in nine paint manufacturing companies who were employed for 5 years or longer from 1955 to 1975 observed 5.5-fold increased risk (95% CI:1.1–16) of myeloma; all of the workers who developed myeloma had received “high” exposures (RR = 10, n = 3) (Lundberg, 1986). A second Swedish cohort study, using census information on occupation, observed a 1.7-fold increased risk (n = 7) (McLaughlin et al., 1988). In six case-control studies, odds ratios of 1.6 to 3 were reported (Bethwaite et al., 1990; Cuzick and De Stavola, 1988; Demers et al., 1993; Friedman, 1986; Morris et al., 1986; Pasqualetti et al., 1990), but no association was observed in other studies conducted in Italy (unadjusted OR = 0.9, no. of exposed cases = 5) (La Vecchia et al., 1989) or Denmark (women, OR = 0.7, 95% CI:0.3–1.7) (Pottern et al., 1992b) (men, OR = 1.0, 95% CI:0.5–2.1) (Heineman et al., 1992). One study reported that myeloma risk increased with duration of employment as a painter (less than 10 years, OR = 1.4, 95% CI:0.6–2.8; 10 years or longer, OR = 4.1, 95% CI:1.8–10) and was stronger for individuals who reported relatively high exposure to paints or solvents (compared with low leisure-time exposure: high leisure-time exposure, OR = 1.6, 95% CI:1.0–3.2; low occupational exposure, OR = 1.9, 95% CI:0.6–5.5; high occupational exposure, OR = 3.1, 95% CI:1.5–7.5) (Demers et al., 1993). In terms of specific components of paint, a history of exposure to dyes and inks was associated with myeloma risk in studies conducted in the United States by Boffetta et al. (OR = 2.7, 95% CI:0.9–8.6) (Boffetta et al., 1989) and Morris et al. (OR = 1.9, 95% CI:0.7–5.3) (Morris et al., 1986), but not in studies conducted in the United Kingdom (Cuzick and De Stavola, 1988), Italy (La Vecchia et al., 1989), or Denmark (Heineman et al., 1992; Pottern et al., 1992b).

Wood Products Industries This category includes manufacturing of lumber, wood products, furniture and fixtures, as well as the forestry industries, and may involve exposure to wood dust, chemicals used to treat wood, adhesives, paint, and stains. A pooled analysis of cohort studies in wood-related industries (Miller et al., 1989; Miller et al., 1994) suggested an increased risk of death from myeloma (SMR = 1.3, 95% CI:0.9–1.9) (Demers et al., 1995) and a retrospective cohort mortality study of US woodtreating plant employees found elevated myeloma mortality, based on 6 cases (SMR = 40.1, 95% CI:14.7–87.3) (Wong and Harris, 2005). However, case-control studies of myeloma suggest little or no altered risk among workers in wood products industries (Cuzick and De Stavola, 1988; Kawachi et al., 1989; Figgs et al., 1994; La Vecchia et al., 1989; McLaughlin et al., 1988; Nandakumar et al., 1986; Pottern et al., 1992b; Reif et al., 1989b; Tollerud et al., 1985; Heineman et al., 1992), with some exceptions. Demers et al. (1993) reported a 2.5-fold (95% CI:1.1–12) increased risk associated with forestry and logging occupations, whereas Eriksson and Karlsson (1992) reported a 1.5-fold increased risk (95% CI:1.0–2.3) for lumberjacks. Myeloma mortality was associated with a 1.7-fold increased risk (95% CI:0.8–3.9) for men in woodworking occupations in western Auestralia (Nandakumar et al., 1988). Some studies reported associations between myeloma risk and occupational exposure to wood dust (OR = 1.8, 95% CI:0.7– 4.3) (Bofetta et al., 1989); (OR = 1.9, 95% CI:0.4–8.4) (Pottern et al., 1992b), and fresh wood (OR = 3.9, 95% CI:1.9–7.6) (Flodin et al., 1987); (OR = 1.5, 95% CI:1.0–2.4) (Eriksson and Karlsson, 1992).

Asbestos Asbestos was postulated as a lymphoid system carcinogen following the publication of case reports of men with both asbestosis and B-cell

Multiple Myeloma neoplasms, as well as animal and human studies that demonstrated increased production of M-component in relation to asbestos exposure and asbestosis (Kagan, 1985). A Danish cohort of all employees in an asbestos-cement plant, followed from their employment as early as 1928 through 1984, noted a 1.7-fold increased risk of myeloma (95% CI:0.7–3.3) (Raffn et al., 1989). A meta-analysis of case-control studies conducted in the United States and Europe (Boffetta et al., 1989; Cuzick and De Stavola, 1988; Eriksson and Karlsson, 1992; Figgs et al., 1994; La Vecchia et al., 1989; Linet et al., 1987; Morris et al., 1986; Pasqualetti et al., 1990; Schwartz et al., 1988) found a slightly increased myeloma risk associated with asbestos exposure (OR = 1.2; 95% CI:1.0–1.4) (Becker et al., 2001).

Engine Exhaust Myeloma risk was related to a history of occupational exposure to diesel or engine exhaust in one cohort and six case-control studies. Self-reported occupation as a truck driver in the 1970 Swedish census was related to myeloma mortality in the subsequent 10 years (standardized mortality ratio = 4.4, 95% CI:1.4–10.2) (Hansen, 1993). In case-control studies, odds ratios for the association between myeloma risk and occupational exposure to engine exhaust ranged from 1.4 to 2.1 (Boffetta et al., 1989; Eriksson and Karlsson, 1992; Flodin et al., 1987; Heineman et al., 1992; Pahwa et al., 2003; Pottern et al., 1992b; Van den Eeden and Friedman, 1993). Boffetta et al. (1989) found an association for self-reported exposure to diesel exhaust (OR = 1.4, 95% CI:0.7–2.7) but not gasoline exhaust (OR = 0.9, 95% CI:0.5–1.6). Myeloma risk was similarly elevated in association with diesel exposure (RR = 1.3, 95% CI:1.0–1.7) in a cohort of Swedish construction workers (Lee et al., 2003). Risk was related to prior carbon monoxide exposure in two studies, with odds ratios of 1.9 (95% CI:1.1–3.2) (Morris et al., 1986) and 1.5 (95% CI:1.2–2.2) (Pasqualetti et al., 1990). A meta-analysis found a slightly elevated summary odds ratio for the association between engine exhaust and myeloma risk (OR = 1.3, 95% CI:1.1–1.6), based on self-reported exposure or exposure inferred from job titles (Sonoda et al., 2001).

939

did not report results for myeloma specifically, but found no increase of all hematologic malignancies combined (Beaumont et al., 1991; Vena and Fiedler, 1987). Occupation as a welder was associated with myeloma incidence in several case-control studies (OR = 2.0, 95% CI:0.6–5.7) (Heineman et al., 1992); (OR = 3.3; 95% CI:1.3–8.5) (Costantini et al., 2001); (OR = 1.2; 95% CI:0.7–2.0) (Demers et al., 1993), and extremely low frequency magnetic field exposure among welders was associated with myeloma incidence among women (high vs. low exposure, RR = 3.8; 95% CI:0.9–15.6) (Hakansson et al., 2002). Myeloma risk was associated with occupation as an embalmer and funeral director (PMR = 1.4, 95% CI:0.8–2.1) (Hayes et al., 1990) and formaldehyde exposure (OR = 1.8, 95% CI:0.6–5.7) (Boffetta et al., 1989), but no association with formaldehyde was found in another case-control study (Heineman et al., 1992). Myeloma incidence and mortality were elevated in two cohorts of fishermen (RR = 2.5, 95% CI:0.7–6.4) (Hagmar et al., 1992) (east coast of Sweden, SMR = 3.1, 95% CI:1.2–6.4; west coast of Sweden, SMR = 3.2, 95% CI:1.2–8.7) (Svensson et al., 1995), and a case-control study found a slight elevation associated with occupation as a fisherman (OR = 1.3, p. 0.05) (McLaughlin et al., 1988).

Summary Multiple studies have provided evidence that agricultural work is associated with myeloma risk. Pesticide use may account for part of the association of agricultural work with myeloma, but there is also some evidence that other agricultural exposures including exposure to animals, may increase risk. Unfortunately, there are few data concerning the specific pesticides that should be targeted for further investigation. There is fairly consistent evidence that exposures in paint-related occupations increase myeloma risk. Whether this increased risk results from dyes and pigments, or from solvents used in paint formulations, has not been discerned. Limited evidence from cohort studies that occupational exposure to benzene is associated with myeloma incidence is contradicted by largely negative results from studies of petroleum and rubber workers, arguing against a role of benzene as a strong risk factor for myeloma.

Metals In a cohort study, Teta and Ott (1988) noted that male hourly workers employed at a New York metal fabrication facility and followed during 1946 to 1981 were at slightly increased risk of myeloma (RR = 1.4, no. of exposed cases = 3). A cohort study of metal workers found increased myeloma risk among machinists (proportionate mortality ratio = 209) (Gallagher and Threlfall, 1983). In two case-control studies, a history of exposure to metals (not otherwise specified) was related to myeloma risk, with odds ratios of 1.8 (95% CI:1.1–2.9) (Morris et al., 1986) and 1.5 (no. of exposed cases = 81) (Cuzick and De Stavola, 1988). In four other reports, there was no relation (Heineman et al., 1992; La Vecchia et al., 1989; Pasqualetti et al., 1990; Pottern et al., 1992b). There is very limited information about exposure to specific metals in relation to the disease (Egedahl et al., 1993; Linet et al., 1987). One study found a significantly increased risk of myeloma mortality among workers of a nickel refinery in Canada (SMR = 12.6, 95% CI:2.5–36.8) (Egedahl et al., 1993). Although no other study has specifically addressed nickel exposure, a study of workers in jobs grinding steel observed no excess mortality from lymphoma and myeloma combined (SMR = 0.7, 95% CI:0.2–2.2), except for a slight increase among those employed 5 years or longer with more than 20 years latency (SMR = 1.4, 95% CI:0.2–5.4) (Svensson et al., 1989).

Other Occupations and Exposures Other occupations and exposures have not been extensively investigated, but have been associated with myeloma in several studies. Cohort studies of firefighters observed increased myeloma mortality in Philadelphia, United States (for 10–19 years employment: OR = 1.5, 95% CI:0.5–4.7; for ≥ 20 years employment: OR = 2.3, 95% CI:1.0–5.2) (Baris et al., 2001) in Seattle, United States (for ≥ 20 years employment: SMR = 9.9, 2 cases) (Heyer et al., 1990), and in Canada (SMR = 10.0, 95% CI:1.2–36.1) (Howe and Burch, 1990), and several proportionate mortality analyses in the United States have found similar results (Howe and Burch, 1990). Other studies of firefighters

Lifestyle Factors Alcohol Intake and Tobacco Use Studies of the relation between alcohol intake and risk of myeloma have found no apparent association (Boffetta et al., 1989; Brown et al., 1992c; Linet et al., 1987; Tavani et al., 1997; Williams and Horm, 1977). Of many studies that have investigated tobacco use in relation to myeloma risk (Adami et al., 1998; Boffetta et al., 1989; Brown et al., 1992b; Brownson 1991; Doll and Peto, 1976; Flodin et al., 1987; Friedman, 1993; Gallagher et al., 1983; Heineman et al., 1992; Herrinton et al., 1992; Linet et al., 1987; Miligi et al., 1999; Rogot and Murray, 1980; Stagnaro et al., 2001; Williams and Horm, 1977), only one found consistent evidence for an increased risk among smokers (Mills et al., 1990).

Diet and Obesity Excess risk of myeloma has been linked to obesity in a number of casecontrol and cohort studies. In a large Canadian case-control study (Pan et al., 2004), obese [body-mass index (BMI) of ≥30 kg/m2] men and women had an elevated risk of multiple myeloma (OR = 2.0, 95% CI:1.5–2.9) compared with people with normal BMI (£25 kg/m2). Obesity and overweight were also associated with increased risk of myeloma in both whites (obese vs. normal BMI, OR = 1.9, 95% CI:1.2–3.1) and blacks (obese vs. normal BMI, OR = 1.5, 95% CI:0.9–2.4) in a US case-control study (Brown et al., 2001). A significant positive linear trend was observed in myeloma death rates with increasing BMI among both men and women in a large US prospective cohort study (Calle et al., 2003). In a cohort of male US veterans, there was a significantly elevated risk of myeloma among obese white and black men (RR = 1.2 and 1.3, respectively) (Samanic et al., 2004). In a women’s health study in Iowa, there was 1.5 to 2 fold increased risk of myeloma for women in the highest category of measure by body-mass index, weight, waist/hip ratio, and waist and hip circumferences compared to women in the lowest category (Blair et al., 2005).

940

PART IV: CANCER BY TISSUE OF ORIGIN

Dietary factors have been examined in relation to myeloma risk in only a handful of studies, with some consistent results. Decreased risk of myeloma was observed in relation to consumption of certain vegetables, including cruciferous vegetables (highest vs. lowest fourth, OR = 0.7, p < 0.05) (Brown et al., 2001) and green vegetables (highest third, OR = 0.4, p < 0.01) (Tavani et al., 1997). An inverse association was also observed for high consumption of whole grain foods (>3 days/week, OR = 0.5, 95% CI:0.2–1.1) (Chatenoud et al., 1998). In a US study, increased myeloma risk was associated with fruit and juice intake among blacks (highest fourth, OR = 1.5, p for trend across fourths <0.05), but not whites (Brown et al., 2001). Intake of several meats and fats were associated with increased risk of myeloma, including liver (highest vs. lowest third, OR = 2.0, p < 0.01), butter (highest third, OR = 2.8, p < 0.01), and total seasoning fats (highest third, OR = 2.4, p < 0.01) (Tavani et al., 1997), whereas fish intake was associated with decreased risk in two studies (highest vs. lowest fourth, OR = 0.6, p < 0.05 (Brown et al., 2001); ≥2 servings/week, OR = 0.5, 95% CI:0.3–0.9 (Fernandez et al., 1999). Few specific nutrients have been evaluated in relation to myeloma. In a US study, vitamin C supplementation for 5 years or longer was associated with decreased risk among whites (OR = 0.6, 95% CI:0.4–0.9), but not blacks (Brown et al., 2001). In an Italian study, b-carotene was inversely associated (highest third, OR = 0.06, p < 0.01), and retinol intake was positively associated (highest third, OR = 2.3, p < 0.01) with myeloma risk (Tavani et al., 1997), but neither nutrient was associated with myeloma in blacks or whites in the US study (Brown et al., 2001).

Hair-Coloring Products Some initial reports of an excess risk of myeloma among women working as cosmetologists or hairdressers (Guidotti et al., 1982; Spinelli et al., 1984) led to a hypothesis of an association between exposure to hair-coloring products and myeloma risk. Observations in other study populations gave further indication of increased risk in occupations with exposure to hair-coloring products (Flodin et al., 1987; Herrinton et al., 1994; Miligi et al., 1999), although not all studies found such an association (Eriksson and Karlsson, 1992; Pottern et al., 1992b; Teta et al., 1984). Personal use of permanent hair dyes for 20 years or longer was associated with a slightly increased risk of myeloma in the American Cancer Society cohort (RR = 1.4, 95% CI:0.9–2.3) that was most apparent in women who had used black dyes (RR = 4.4, 95% CI:1.1–18.3) (Thun et al., 1994). In case-control studies, personal use of hair dyes was associated with myeloma risk in both women (OR = 1.8, 95% CI:0.9–3.7) (Zahm et al., 1992) and men (OR = 1.8, 95% CI:0.5–5.7) (Zahm et al., 1992); OR = 1.9, 95% CI:1.0–3.6 (Brown et al., 1992a); OR = 1.3, 95% CI:0.7–2.3 (Herrinton et al., 1994). No association between permanent hair dye use and myeloma risk was observed among women in a multicenter US study (Herrinton et al., 1994), or in a case-control study conducted in Italy (Miligi et al., 1999). In the Zahm et al. study, there was some indication of a higher increased risk among those using permanent hair-coloring products compared with those using semi- or non-permanent hair-coloring products or products that change hair color gradually (Zahm et al., 1992).

Medication Use There are scant data on the use of specific medications in relation to myeloma risk. A study of members of the Kaiser Permanente Medical Care Program using computerized pharmacy records found increased myeloma risk associated with erythromycin use (14 observed cases, 5.1 expected; SMR = 2.7, p < 0.002) (Selby et al., 1989). Diphenylhydantoin use was in excess among myeloma cases compared with controls in a hospital-based case-control study (3 vs. 0 discordant pairs) (Linet et al., 1987), but not in the Kaiser Permanente study (Selby et al., 1989). The same case-control study found a positive association with the use of laxatives (OR = 3.5; 95% CI:1.1–11.1), and imprecise increased risks (OR ≥ 2.0) associated with use of diazepam, ibuprofen, and diet pills or stimulants (Linet et al., 1987). A second case-control study found suggestive positive associations with corticosteroids (OR = 2.3, 95% CI:0.60–8.7) and salicylates or paracetamol (OR = 2.0, 95% CI:0.7–5.7) (Eriksson, 1993).

FUTURE DIRECTIONS From the research completed to date, it appears that the question, “What causes myeloma?” will not be easy to answer. In areas where there are indicators, for example, associations with exposure to radiation and pesticides or a history of farming—the difficulty of more precisely measuring potential etiologic factors has limited our understanding of the nature of the associations. One promising avenue for future research is based on the recognition of MGUS as a strong predictor of myeloma risk, and would involve studying the causes of MGUS, and the factors associated with malignant transformation of MGUS to myeloma. Advances in the field of molecular biology have made research into genetic susceptibility feasible and affordable. Investigation into the role of common genetic polymorphisms in myeloma risk has just begun, with few promising leads so far. However, this line of research should be pursued, because an association with family history and racial differences in incidence strongly suggest a role of genetic factors. Genes of interest include those involved in immune function, growth factor genes (e.g., insulin growth factor (IGF) ) because of their influence on plasma cell growth, and metabolism genes responsible for biotransformation of toxic chemicals (e.g., cytochrome P-450 (CYP) gene family) because of the possible effect of exogenous chemicals on myeloma risk. A promising area of research is investigation into risk factors for molecular changes (chromosomal translocations, gene silencing by methylation) that are commonly present in myeloma tumors, although the small numbers in patient subgroups will limit this research. It may be useful to look to differences between MGUS and myeloma for clues about other genes implicated in malignant transformation; for example, virtually all myeloma patients (95%) are positive for IL-1b production by monoclonal plasma cells, whereas the majority of patients with MGUS are negative for IL-1b expression (25%) (Lacy et al., 1999). Emerging technologies such as proteomics and DNA arrays may also be useful in further identifying proteins or genes that are overexpressed in myeloma cells (De Vos et al., 2001). Targeting research toward demographic groups at increased risk of myeloma, such as blacks or farmers, might prove useful to investigate early biologic effects relevant to the development of myeloma. Investigations into the prevalence of MGUS, other immune system effects, or chromosomal alterations in these susceptible subgroups may provide clues into reasons for racial disparities and occupational risk factors among farmers. References Ablashi DV, Chatlynne L, Thomas D, et al. 2000. Lack of serologic association of human herpesvirus-8 (KSHV) in patients with monoclonal gammopathy of undetermined significance with and without progression to multiple myeloma. Blood 96:2304–2306. Abu-Shakra M, Gladman DD, Urowitz MB. 1996. Malignancy in systemic lupus erythematosus. Arthritis Rheum 39:1050–1054. Adami J, Nyren O, Bergstrom R, et al. 1998. Smoking and the risk of leukemia, lymphoma, and multiple myeloma (Sweden). Cancer Causes Control 9:49–56. Agbalika F, Mariette X, Marolleau JP, Fermand JP, Brouet JC. 1998. Detection of human herpesvirus-8 DNA in bone marrow biopsies from patients with multiple myeloma and Waldenstrom’s macroglobulinemia. Blood 91:4393–4394. Alavanja MC, Blair A, Merkle S, Teske J, Eaton B. 1988. Mortality among agricultural extension agents. Am J Ind Med 14:167–176. Andersson M, Storm HH. 1992. Cancer incidence among Danish Thorotrastexposed patients. J Natl Cancer Inst 84:1318–1325. Andjelkovich D, Taulbee J, Blum S. 1978. Mortality of female workers in rubber manufacturing plant. J Occup Med 20:409–413. Angtuaco EJC, Fassas ABT, Walker R, Sethi R, Barlogie B. 2004. Multiple myeloma: clinical review and diagnostic imaging. Radiology 231:11–23. Avet-Loiseau H, Li JY, Facon T, et al. 1998. High incidence of translocations t(11;14)(q13;q32) and t(4;14)(p16;q32) in patients with plasma cell malignancies. Cancer Res 58:5640–5645. Avet-Loiseau H, Li JY, Morineau N, et al. 1999. Monosomy 13 is associated with the transition of monoclonal gammopathy of undetermined significance to multiple myeloma. Intergroupe Francophone du Myelome. Blood 94:2583–2589.

Multiple Myeloma Axelsson U. 1986. A 20-year follow-up study of 64 subjects with Mcomponents. Acta Med Scand 219:519–522. Axelsson U, Bachmann R, Hallen J. 1966. Frequency of pathological proteins (M-components) om 6,995 sera from an adult population. Acta Med Scand 179:235–247. Baker GL, Kahl LE, Zee BC, Stolzer BL, Agarwal AK, Medsger TA, Jr. 1987. Malignancy following treatment of rheumatoid arthritis with cyclophosphamide. Long-term case-control follow-up study. Am J Med 83:1–9. Baldini L, Guffanti A, Cesana BM, et al. 1996. Role of different hematologic variables in defining the risk of malignant transformation in monoclonal gammopathy. Blood 87:912–918. Barille-Nion S, Barlogie B, Bataille R, et al. 2003. Advances in biology and therapy of multiple myeloma. Hematology 248–278. Baris D, Brown LM, Silverman DT, et al. 2000. Socioeconomic status and multiple myeloma among US blacks and whites. Am J Public Health 90:1277–1281. Baris D, Garrity T, Telles JL, Heineman EF, Olshan A, Zahm. SH. 2001. A cohort mortality study of Philadelphia firefighters. Am J Ind Med 39:463–476. Baris D, Silverman DT, Brown LM, et al. 2004. Occupation, pesticide exposure and risk of multiple myeloma. Scand J Work Environ Health 30:215–222. Barlogie B, Epstein J, Selvanayagam P, Alexanian R. 1989. Plasma cell myeloma—new biological insights and advances in therapy. Blood 73:865–879. Barlogie B, Shaughnessy J, Tricot G, et al. 2004. Treatment of multiple myeloma. Blood 103:20–32. Beaumont JJ, Chu GS, Jones JR, et al. 1991. An epidemiologic study of cancer and other causes of mortality in San Francisco firefighters. Am J Ind Med 19:357–372. Beauparlant P, Papp K, Haraoui B. 1999. The incidence of cancer associated with the treatment of rheumatoid arthritis. Semin Arthritis Rheum 29:148–158. Becker N, Berger J, Bolm-Audorff U. 2001. Asbestos exposure and malignant lymphomas—a review of the epodemiological literature. Int Arch Occup Environ Health 74:459–469. Beksac M, Ma M, Akyerli C, et al. 2001. Frequent demonstration of human herpesvirus 8 (HHV-8) in bone marrow biopsy samples from Turkish patients with multiple myeloma (MM). Leukemia 15:1268– 1273. Beral V, Inskip H, Fraser P, Booth M, Coleman D, Rose G. 1985. Mortality of employees of the United Kingdom Atomic Energy Authority, 1946–1979. Br Med J (Clin Res Ed) 291:440–447. Beral V, Newton R. 1998. Overview of the epidemiology of immunodeficiencyassociated cancers. J Natl Cancer Inst Monogr 23:1–6. Bertazzi A, Pesatori AC, Consonni D, Tironi A, Landi MT, Zocchetti C. 1993. Cancer incidence in a population accidentally exposed to 2, 3, 7, 8-tetrachlorodibenzo-para-dioxin. Epidemiology 4:398–406. Bethwaite PB, Pearce N, Fraser J. 1990. Cancer risks in painters: Study based on the New Zealand Cancer Registry. Br J Ind Med 47:742–746. Bjornadal L, Lofstrom B, Yin L, Lundberg IE, Ekbom A. 2002. Increased cancer incidence in a Swedish cohort of patients with systemic lupus erythematosus. Scand J Rheumatol 31:66–71. Blade J, Lopez-Guillermo A, Rozman C, et al. 1992. Malignant transformation and life expectancy in monoclonal gammopathy of undetermined significance. Br J Haematol 81:391–394. Blair A, Malker H, Cantor KP, Burmeister L, Wiklund K. 1985. Cancer among farmers. A review. Scand J Work Environ Health 11:397–407. Blair CK, Cerhan JR, Folsom AR, Ross JA. 2005. Anthropometric characteristics and risk of multiple myeloma. Epidemiology 16:691–494. Blattner WA, Blair A, Mason TJ. 1981. Multiple myeloma in the United States, 1950–1975. Cancer 48:2547–2554. Boffetta P, Stellman SD, Garfinkel L. 1989. A case-control study of multiple myeloma nested in the American Cancer Society prospective study. Int J Cancer 43:554–559. Boice JD, Jr., Day NE, Andersen A, et al. 1985. Second cancers following radiation treatment for cervical cancer. An international collaboration among cancer registries. J Natl Cancer Inst 74:955–975. Boice JD, Jr., Engholm G, Kleinerman RA, et al. 1988. Radiation dose and second cancer risk in patients treated for cancer of the cervix. Radiat Res 116:3–55. Boice JD, Jr., Morin MM, Glass AG, et al. 1991. Diagnostic x-ray procedures and risk of leukemia, lymphoma, and multiple myeloma. JAMA 265:1290–1294. Bourguet CC, Grufferman S, Delzell E, Delong ER, Cohen HJ. 1985. Multiple myeloma and family history of cancer. A case-control study. Cancer 56:2133–2139. Bourguet CC, Logue EE. 1993. Antigenic stimulation and multiple myeloma. A prospective study. Cancer 72:2148–2154.

941

Bowden M, Crawford J, Cohen HJ, Noyama O. 1993. A comparative study of monoclonal gammopathies and immunoglobulin levels in Japanese and United States elderly. J Am Geriatr Soc 41:11–14. Brinton LA, Gridley G, Hrubec Z, Hoover R, Fraumeni JF, Jr. 1989. Cancer risk following pernicious anaemia. Br J Cancer 59:810–813. Brousset P, Meggetto F, Attal M, Delsol G. 1997. Kaposi’s sarcoma-associated herpesvirus infection and multiple myeloma. Science 278:1972–1973. Brown LM, Everett GD, Burmeister LF, Blair A. 1992a. Hair dye use and multiple myeloma in white men. Am J Public Health 82:1673–1674. Brown LM, Everett GD, Gibson R, Burmeister LF, Schuman LM, Blair A. 1992b. Smoking and risk of non-Hodgkin’s lymphoma and multiple myeloma. Cancer Causes Control 3:49–55. Brown LM, Gibson R, Burmeister LF, Schuman LM, Everett GD, Blair A. 1992c. Alcohol consumption and risk of leukemia, non-Hodgkin’s lymphoma, and multiple myeloma. Leuk Res 16:979–984. Brown LM, Gridley G, Pottern LM, et al. 2001. Diet and nutrition as risk factors for multiple myeloma among blacks and whites in the United States. Cancer Causes Control 12:117–125. Brown LM, Linet MS, Greenberg RS, et al. 1999. Multiple myeloma and family history of cancer among blacks and whites in the U.S. Cancer 85:2385–2390. Brownson RC. 1991. Cigarette smoking and risk of myeloma. J Natl Cancer Inst 83:1036–1037. Brownson RC, Reif JS, Chang JC, Davis JR. 1989. Cancer risks among Missouri farmers. Cancer 64:2381–2386. Buhler S, Laitinen K, Holthofer H, Jarvinen A, Schauman KO, Hedman K. 2002. High rate of monoclonal gammopathy among immunocompetent subjects with primary cytomegaloviurs infection. Clin Infect Dis 35: 1430–1433. Burmeister LF. 1990. Cancer in Iowa farmers: Recent results. Am J Ind Med 18:295–301. Burmeister LF, Everett GD, Van Lier SF, Isacson P. 1983. Selected cancer mortality and farm practices in Iowa. Am J Epidemiol 118:72–77. Cabannes E, Khan G, Aillet F, Jarrett RF, Hay RT. 1999. Mutations in the IkBa gene in Hodgkin’s disease suggest a tumour suppressor role for IkappaBalpha. Oncogene 18:3063–3070. Cantor KP, Blair A. 1984. Farming and mortality from multiple myeloma: A case-control study with the use of death certificates. J Natl Cancer Inst 72:251–255. Cardis E, Gilbert ES, Carpenter L, et al. 1995. Effects of low doses and low dose rates of external ionizing radiation: Cancer mortality among nuclear industry workers in three countries. Radiat Res 142:117–132. Carpenter L, Higgins C, Douglas A, Fraser P, Beral V, Smith P. 1994. Combined analysis of mortality in three United Kingdom nuclear industry workforces, 1946–1988. Radiat Res 138:224–238. Cartwright RA, Gurney KA, Moorman AV. 2002. Sex ratios and the risks of haematological malignancies. Br J Haematol 118:1071–1077. Cathomas G, Stalder A, Kurrer MO, Regamey N, Erb P, Joller-Jemelka HI. 1998. Multiple myeloma and HHV8 infection. Blood 91:4391–4393. Cerhan JR, Cantor KP, Williamson K, Lynch CF, Torner JC, Burmeister LF. 1998. Cancer mortality among Iowa farmers: Recent results, time trends, and lifestyle factors (United States). Cancer Causes Control 9:311–319. Chatenoud L, Tavani A, La Vecchia C, et al. 1998. Whole grain food intake and cancer risk. Int J Cancer 77:24–28. Chauhan D, Bharti A, Raje N, et al. 1999. Detection of Kaposi’s sarcoma herpesvirus DNA sequences in multiple myeloma bone marrow stromal cells. Blood 93:1482–1486. Chen R, Seaton A. 1996. A meta-analysis of mortality among workers exposed to organic solvents. Occup Med (Lond) 46:337–344. Christie D, Robinson K, Gordon I, Bisby J. 1991. A prospective study in the Australian petroleum industry. II. Incidence of cancer. Br J Ind Med 48:511–514. Cibere J, Sibley J, Haga M. 1997. Rheumatoid arthritis and the risk of malignancy. Arthritis Rheum 40:1580–1586. Cocco P, Blair A, Congia P, et al. 1997. Proportional mortality of dichlorodiphenyl-trichloroethane (DDT) workers: A preliminary report. Arch Environ Health 52:299–303. Cocco P, Kazerouni N, Zahm SH. 2000. Cancer mortality and environmental exposure to DDE in the United States. Environ Health Perspect 108:1–4. Cohen HJ, Crawford J, Rao MK, Pieper CF, Currie MS. 1998. Racial differences in the prevalence of monoclonal gammopathy in a communitybased sample of the elderly. Am J Med 104:439– 444. Colls BM. 1999. Monoclonal gammopathy of undetermined significance (MGUS)—31 year follow up of a community study. Aust N Z J Med 29:500–504. Costantini AS, Miligi L, Kriebel D, et al. 2001. A multicenter case-control study in Italy on hematolymphopoietic neoplasms and occupation. Epidemiology 12:78–87.

942

PART IV: CANCER BY TISSUE OF ORIGIN

Cull GM, Carter GI, Timms JM, Thomson BJ, Russell NH, Haynes AP. 1999. Low incidence of human herpesvirus 8 in stem cell collections from myeloma patients. Bone Marrow Transplant 23:759–761. Cuzick J, De Stavola B. 1988. Multiple myeloma—a case-control study. Br J Cancer 57:516–520. Cuzick J, De Stavola BL. 1989. Autoimmune disorders and multiple myeloma. Int J Epidemiol 18:283. Cuzick J, Velez R, Doll R. 1983. International variations and temporal trends in mortality from multiple myeloma. Int J Cancer 32:13–19. Dalton WS, Bergsagel PL, Kuehl WM, Anderson KC, Harousseau JL. 2001. Multiple myeloma. Hematology (Am Soc Hematol Educ Program) 157–177. Darby SC, Doll R, Gill SK, Smith PG. 1987. Long term mortality after a single treatment course with X-rays in patients treated for ankylosing spondylitis. Br J Cancer 55:179–190. Darby SC, Reeves G, Key T, Doll R, Stovall M. 1994. Mortality in a cohort of women given X-ray therapy for metropathia haemorrhagica. Int J Cancer 56:793–801. Davies FE, Rollinson SJ, Rawstron AC, et al. 2000. High-producer haplotypes of tumor necrosis factor alpha and lymphotoxin alpha are associated with an increased risk of myeloma and have an improved progression-free survival after treatment. J Clin Oncol 18:2843–2851. Davis FG, Boice JD, Hrubec Z, Monson RR. 1989. Cancer mortality in a radiation-exposed cohort of Massachusetts tuberculosis patients. Cancer Res 49:6130–6136. De Vos J, Couderc G, Tarte K, et al. 2001. Identifying intercellular signaling genes expressed in malignant plasma cells by using complementary DNA arrays. Blood 98:771–780. Demers PA, Boffetta P, Kogevinas M, et al. 1995. Pooled reanalysis of cancer mortality among five cohorts of workers in wood-related industries. Scand J Work Environ Health 21:179–190. Demers PA, Vaughan TL, Koepsell TD, et al. 1993. A case-control study of multipe myeloma and occupation. Am J Ind Med 23:629–639. Demeter J, Messer G, Ramisch S, Mee JB, Di Giovine FS, Schmid M, Herrmann F, Porzsolt F. 1996. Polymorphism within the second intron of the IL-1 receptor antagonist gene in patients with hematopoietic malignancies. Cytokines Mol Ther 2:239–242. Devesa SS, Silverman DT, Young JLJ, et al. 1987. Cancer incidence and mortality trends among whites in the United States, 1947–84. J Natl Cancer Inst 79:701–770. Divine BJ, Hartman CM. 2001. A cohort mortality study among workers at a 1,3 butadiene facility. Chem Biol Interact 135–136:535–553. Doll R. 1991. Urban and rural factors in the aetiology of cancer. Int J Cancer 47:803–810. Doll R, Peto R. 1976. Mortality in relation to smoking: 20 years’ observations on male British doctors. Br Med J 2:1525–1536. Dominici M, Luppi M, Campioni D, et al. 2000. PCR with degenerate primers for highly conserved DNA polymerase gene of the herpesvirus family shows neither human herpesvirus 8 nor a related variant in bone marrow stromal cells from multiple myeloma patients. Int J Cancer 86:76–82. Doody MM, Linet MS, Glass AG, et al. 1992. Leukemia, lymphoma, and multiple myeloma following selected medical conditions. Cancer Causes Control 3:449–456. Drabick JJ, Davis BJ, Lichy JH, Flynn J, Byrd JC. 2002. Human herpesvirus 8 genome is not found in whole bone marrow core biopsy specimens of patients with plasma cell dyscrasias. Ann Hematol 81:304–307. Dring AM, Davies FE, Rollinson SJ, et al. 2001. Interleukin 6, tumour necrosis factor alpha and lymphotoxin alpha polymorphisms in monoclonal gammopathy of uncertain significance and multiple myeloma. Br J Haematol 112:249–251. Egedahl RD, Carpenter M, Homik R. 1993. An update of an epidemiology study at a hydrometallurgical nickel refinery in Fort Saskatchewan, Alberta. Health Rep 5:291–302. Eriksson M. 1993. Rheumatoid arthritis as a risk factor for multiple myeloma: A case-control study. Eur J Cancer 29A:259–263. Eriksson M, Hallberg B. 1992. Familial occurrence of hematologic malignancies and other diseases in multiple myeloma: A case-control study. Cancer Causes Control 3:63–67. Eriksson M, Karlsson M. 1992. Occupational and other environmental factors and multiple myeloma: A population based case-control study. Br J Ind Med 49:95–103. Fassas A, Shaughnessy J, Barlogie B. 2005. Cure of myeloma: hype or reality? Bone Marrow Transplantation 35:215–24. Feinman R, Sawyer J, Hardin J, Tricot G. 1997. Cytogenetics and molecular genetics in multiple myeloma. Hematol Oncol Clin North Am 11:1–25. Fernandez E, Chatenoud L, La Vecchia C, Negri E, Franceschi S. 1999. Fish consumption and cancer risk. Am J Clin Nutr 70:85–90. Festen JJ, Marrink J, Waard-Kuiper EH, Mandema E. 1977. Immunoglobulins in families of myeloma patients. Scand J Immunol 6:887–896.

Figgs LW, Dosemeci M, Blair A. 1994. Risk of multiple myeloma by occupation and industry among men and women: A 24-state death certificate study. J Occup Med 36:1210–1221. Flodin U, Fredriksson M, Persson B. 1987. Multiple myeloma and engine exhausts, fresh wood, and creosote: A case-referent study. Am J Ind Med 12:519–529. Fonseca R, Bailey RJ, Ahmann GJ, et al. 2002a. Genomic abnormalities in monoclonal gammopathy of undetermined significance. Blood 100:1417–1424. Fonseca R, Blood EA, Oken MM, et al. 2002b. Myeloma and the t(11;14)(q13;q32); evidence for a biologically defined unique subset of patients. Blood 99:3735–3741. Fonseca R, Oken MM, Greipp PR. 2001. The t(4;14)(p16.3;q32) is strongly associated with chromosome 13 abnormalities in both multiple myeloma and monoclonal gammopathy of undetermined significance. Blood 98:1271. Fordyce EJ, Wang Z, Kahn AR, et al. 2000. Risk of cancer among women with AIDS in New York City. AIDS Public Policy J 15:95–104. Franceschi S, Barbone F, Bidoli E, et al. 1993. Cancer risk in farmers: Results from a multi-site case-control study in north-eastern Italy. Int J Cancer 53:740–745. Friedman GD. 1986. Multiple myeloma: Relation to propoxyphene and other drugs, radiation and occupation. In J Epidemiol 15:424–426. Friedman GD. 1993. Cigarette smoking, leukemia, and multiple myeloma. Ann Epidemiol 3:425–428. Friedman GD. Herrinton LJ. 1994. Obesity and multiple myeloma. Cancer Causes Control 5:479–483. Fritschi L, Johnson KC, Kliewer EV, Fry R. 2002. Animal-related occupations and the risk of leukemia, myeloma, and non-Hodgkin’s lymphoma in Canada. Cancer Causes Control 13:563–571. Gallagher RP, Spinelli JJ, Elwood JM, Skippen DH. 1983. Allergies and agricultural exposure as risk factors for multiple myeloma. Br J Cancer 48:853–857. Gallagher RP, Threlfall WJ. 1983. Cancer mortality in metal workers. Can Med Assoc J 129:1191–1194. Georgescu L, Quinn GC, Schwartzman S, Paget SA. 1997. Lymphoma in patients with rheumatoid arthritis: Association with the disease state or methotrexate treatment. Semin Arthritis Rheum 26:794–804. Gharagozloo S, Khoshnoodi J, Shokri F. 2001. Hepatitis C virus infection in patients with essential mixed cryoglobulinemia, multiple myeloma and chronic lymphocytic leukemia. Pathol Oncol Res 7:135–139. Gilbert ES, Fry SA, Wiggs LD, Voelz GL, Cragle DL, Petersen GR. 1989. Analyses of combined mortality data on workers at the Hanford Site, Oak Ridge National Laboratory, and Rocky Flats Nuclear Weapons Plant. Radiat Res 120:19–35. Goedert JJ, Cote TR, Virgo P, et al. 1998. Spectrum of AIDS-associated malignant disorders. Lancet 351:1833–1839. Gonzalez-Fraile MI, Garcia-Sanz R, Mateos MV, et al. 2002. Methylenetetrahydrofolate reductase genotype does not play a role in multiple myeloma pathogenesis. Br J Haematol 117:890–892. Gonzalez Ordonez AJ, Fernandez Carreira JM, Fernandez Alvarez CR, et al. 2000. Normal frequencies of the C677T genotypes on the methylenetetrahydrofolate reductase (MTHFR) gene among lymphoproliferative disorders but not in multiple myeloma. Leuk Lymphoma 39:607–612. Gramenzi A, Buttino I, D’Avanzo B, Negri E, Franceschi S, La Vecchia C. 1991. Medical history and the risk of multiple myeloma. Br J Cancer 63:769–772. Gregersen H, Ibsen J, Mellemkjoer L, Dahlerup J, Olsen J, Sorensen H. 2001a. Mortality and causes of death in patients with monoclonal gammopathy of undetermined significance. Br J Haematol 112:353–357. Gregersen H, Mellemkjaer L, Ibsen JS, Dahlerup JF, Thomassen L, Sorensen HT. 2001b. The impact of M-component type and immunoglobulin concentration on the risk of malignant transformation in patients with monoclonal gammopathy of undetermined significance. Haematologica 86:1172–1179. Gridley G, McLaughlin JK, Ekbom A. et al. 1993. Incidence of cancer among patients with rheumatoid arthritis. J Natl Cancer Inst 85:307–311. Grufferman S, Cohen HJ, Delzell ES, Morrison MC, Schold SCJ, Moore JO. 1989. Familial aggregation of multiple myeloma and central nervous system diseases. J Am Geriatr Soc 37:303–309. Grulich AE, Swerdlow AJ, Head J, Marmot MG. 1992. Cancer mortality in African and Caribbean migrants to England and Wales. Br J Cancer 66:905–911. Grulich AE, Wan X, Law MG, Coates M, Kaldor JM. 1999. Risk of cancer in people with AIDS. AIDS 13:839–843. Guidotti S, Wright WE, Peters JM. 1982. Multiple myeloma in cosmetologists. Am J Ind Med 3:169–171. Gustavsson P, Hogstedt C, Holmberg B. 1986. Mortality and incidence of cancer among Swedish rubber workers, 1952–1981. Scand J Work Environ Health 12:538–544.

Multiple Myeloma Haas H, Anders S, Bornkamm GW, et al. 1990. Do infections induce monoclonal immunoglobulin components? Clin Exp Immunol 81:435–440. Hagmar L, Linden K, Nilsson A, et al. 1992. Cancer incidence and mortality among Swedish Baltic Sea fishermen. Scand J Work Environ Health 18:217–224. Hakansson N, Floderus B, Gustavsson P, Johansen C, Olsen JH. 2002. Cancer incidence and magnetic field exposure in industries using resistance welding in Sweden. Occup Environ Med 59:481–486. Hallek M, Bergsagel PL, Anderson KC. 1998. Multiple myeloma: increasing evidence for a multistep transformation process. Blood 91:3–21. Hansen ES. 1993. A follow-up study on the mortality of truck drivers. Am J Ind Med 23:811–821. Hansen NE, Karle H, Olsen JH. 1989. Trends in the incidence of multiple myeloma in Denmark 1943–1982: A study of 5500 patients. Eur J Haematol 42:72–76. Harousseau JL, Shaughnessy J Jr, Richardson P. 2004. Multiple myeloma. Hematology (Am Soc Hematol Educ Program) 237–256. Hatcher JL, Baris D, Olshan AF, et al. 2001. Diagnostic radiation and the risk of multiple myeloma (United States). Cancer Causes Control 12:755–761. Hayes RB, Blair A, Stewart PA, Herrick RF, Mahar H. 1990. Mortality of U.S. embalmers and funeral directors. Am J Ind Med 18:641–652. Hayes RB, Yin SN, Dosemeci M, et al. 1997. Benzene and the dose-related incidence of hematologic neoplasms in China. Chinese Academy of Preventive Medicine—National Cancer Institute Benzene Study Group. J Natl Cancer Inst 89:1065–1071. Heineman EF, Olsen JH, Pottern LM, Gomez M, Raffn E, Blair A. 1992. Occupational risk factors for multiple myeloma among Danish men. Cancer Causes Control 3:555–568. Herrinton LJ, Koepsell TD, Weiss NS. 1992. Smoking and multiple myeloma. Cancer Causes Control 3:391–392. Herrinton LJ, Weiss NS, Koepsell TD, et al. 1994. Exposure to hair-coloring products and the risk of multiple myeloma. Am J Public Health 84:1142–1144. Heyer N, Weiss NS, Demers P, Rosenstock L. 1990. Cohort mortality study of Seattle fire fighters: 1945–1983. Am J Ind Med 17:493–504. Hirano T. 1991. Interleukin 6 (IL-6) and its receptor: Their role in plasma cell neoplasias. Int J Cell Cloning 9:166–184. Howe Gr, Burch JD. 1990. Fire fighters and risk of cancer: An assessment and overview of the epidemiologic evidence. Am J Epidemiol 132:1039–1050. Hsing AW, Hansson LE, McLaughlin JK, et al. 1993. Pernicious anemia and subsequent cancer. A population-based cohort study. Cancer 71:745–750. International Myeloma Working Group. Criteria for classification of monoclonal gammopathies, multiple myeloma and related disorders: a report of the International Myeloma Working Group. British J Haematol 2003;121:749–757. Ireland B, Collins JJ, Buckley CF, Riordan SG. 1997. Cancer mortality among workers with benzene exposure. Epidemiology 8:318–320. Isomaki HA, Hakulinen T, Joutsenlahti U. 1978. Excess risk of lymphomas, leukemia and myeloma in patients with rheumatoid arthritis. J Chronic Dis 31:691–696. Iwasaki T, Murata M, Ohshima S, et al. 2003. Second analysis of mortality of nuclear industry workers in Japan, 1986–1997. Radiat Res 159:228–238. Jeffery R, Mitchison NA. 2001. IL-6 polymorphism, anti-IL-6 therapy and animal models of multiple myeloma. Cytokine 16:87. Johnston JM, Grufferman S, Bourguet CC, Delzell E, Delong ER, Cohen HJ. 1985. Socioeconomic status and risk of multiple myeloma. J Epidemiol Community Health 39:175–178. Kagan E. 1985. Current perspectives in asbestosis. Ann Allergy 54:464–473. Kastrinakis NG, Gorgoulis VG, Foukas PG, Dimopoulos MA, Kittas C. 2000. Molecular aspects of multiple myeloma. Ann Oncol 11:1217–1228. Katusic S, Beard CM, Kurland LT, Weis JW, Bergstralh E. 1985. Occurrence of malignant neoplasms in the Rochester, Minnesota, rheumatoid arthritis cohort. Am J Med 78:50–55. Kauppi M, Pukkala E, Isomaki H. 1997. Elevated incidence of hematologic malignancies in patients with Sjogren’s syndrome compared with patients with rheumatoid arthritis (Finland). Cancer Causes Control 8:201–204. Kawachi I, Pearce N, Fraser J. 1989. A New Zealand Cancer Registry-based study of cancer in wood workers. Cancer 64:2609–2613. Kendall GM, Muirhead CR, MacGibbon BH, et al. 1992. Mortality and occupational exposure to radiation: First analysis of the National Registry for Radiation Workers. BMJ 304:220–225. Koepsell TD, Daling JR, Weiss NS, et al. 1987. Antigenic stimulation and the occurrence of multiple myeloma. Am J Epidemiol 126:1051–1062. Konigsberg R, Ackermann J, Kaufmann H, et al. 2000. Deletions of chromosome 13q in monoclonal gammopathy of undetermined significance. Leukemia 14:1975–1979. Kotsa K, Watson PF, Weetman AP. 1997. A CTLA-4 gene polymorphism is associated with both Graves disease and autoimmune hypothyroidism. Clin Endocrinol (Oxf) 46:551–554.

943

Kuehl WM, Bergsagel PL. 2002. Multiple myeloma. Evolving genetic events and host interactions. Nat Rev Cancer 2:175–187. Kyle RA. 1993. Monoclonal gammopathy—after 20 to 35 years of follow-up. Mayo Clin Proc 68:26–36. Kyle RA, Beard CM, O’Fallon WM, Kupland LT. 1994. Incidence of multiple myeloma in Olmsted County, Minnesota: 1978 through 1990, with a review of the trend since 1945. J Clin Oncol 12:1577–1583. Kyle RA, Finkelstein S, Elveback LR, Kurland LT. 1972. Incidence of monoclonal proteins in a Minnesota community with a cluster of multiple myeloma. Blood 40:719–724. Kyle RA, Rajkumar SV. 2004. Multiple Myeloma. New Engl J Med 351:1860–1873. Kyle RA, Rajkumar SV. 2005. Monoclonal gammopathies of undetermined significance. Best Pract Res Clin Haematol 18:689-7-7. Kyle RA, Therneau TM, Rajkumar SV, et al. 2002. A long-term study of prognosis in monoclonal gammopathy of undetermined significance. N Engl J Med 346:564–569. Landgren O, Gridley G, Turesson I, et al. 2006a. Risk of monoclonal gammopathy of undetermined significance (MGUS) and subsequent multiple myeloma among African American and white veterans in the United States. Blood 107:904–906. Landgren O, Linet MS, McMaster ML, Gridley G, Hemminki K, Goldin LR. 2006b. Familial characteristics of autoimmune and hematologic disorders in 8,406 multiple myeloma patients: A population-based case-control study. Int J Cancer Jan 4 [Epub ahead of print]. La Vecchia C, Negri E, D’Avanzo B, Franceschi S. 1989. Occupation and lymphoid neoplasms. Br J Cancer 60:385–388. Lacy MQ, Donovan KA, Heimbach JK, Ahmann GJ, Lust JA. 1999. Comparison of interleukin-1 beta expression by in situ hybridization in monoclonal gammopathy of undetermined significance and multiple myeloma. Blood 93:300–305. Lee E, Burnett CA, Lalich N, Cameron LL, Sestito JP. 2002. Proportionate mortality of crop and livestock farmers in the United States, 1984–1993. Am J Ind Med 42:410–420. Lee WJ, Baris D, Jarvholm B, Silverman DT, Bergdahl IA, Blair A. 2003. Multiple myeloma and diesel and other occupational exposures in Swedish construction workers. Int J Cancer 107:134–138. Leech SH, Bryan CF, Elston RC, Rainey J, Bickers JN, Pelias MZ. 1983. Genetic studies in multiple myeloma. 1. Association with HLA-Cw5. Cancer 51:1408–1411. Levi F, La Vecchia C. 1990. Trends in multiple myeloma. Int J Cancer 46:755–756. Levi F, La Vecchia C, Lucchini F, Negri E. 1992. Trends in cancer mortality sex ratios in Europe, 1950–1989. World Health Stat Q 45:117–164. Lewis DR, Pottern LM, Brown LM, et al. 1994. Multiple myeloma among blacks and whites in the United States: The role of chronic antigenic stimulation. Cancer Causes Control 5:529–539. Lewis EB. 1963. Leukemia, multiple myeloma and aplastic anemia in American radiologists. Science 142:1492–1499. Linet MS, Freedman DM, Mohan AK, et al. 2005. Incidence of haematopoietic malignancies in US radiologic technologists. Occup Environ Med 62:861–867. Linet MS, Harlow SD, McLaughlin JK. 1987. A case-control study of multiple myeloma in whites: Chronic antigenic stimulation, occuption, and drug use. Cancer Res 47:2978–2981. Linet MS, McLaughlin JK, Harlow SD, Fraumeni JF. 1988. Family history of autoimmune disorders and cancer in multiple myeloma. Int J Epidemiol 17:512–513. Ludwig H, Mayr W. 1982. Genetic aspects of susceptibility to multiple myeloma. Blood 59:1286–1291. Lundberg I. 1986. Mortality and cancer incidence among Swedish paint industry workers with long-term exposure to organic solvents. Scand J Work Environ Health 12:108–113. MacKenzie J, Sheldon J, Morgan G, Cook G, Schulz TF, Jarrett RF. 1997. HHV-8 and multiple myeloma in the UK. Lancet 350:1144–1145. MacKenzie MR, Lewis JP. 1985. Cytogenetic evidence that the malignant event in multiple myeloma occurs in a precursor lymphocyte. Cancer Genet Cytogenet 17:13–20. MacMahon B. 1966. Epidemiology of Hodgkin’s disease. Cancer Res 26:1189–1201. Maldonado JE, Kyle RA. 1974. Familial myeloma. Report of eight families and a study of serum proteins in their relatives. Am J Med 57:875–884. Marcelin AG, Dupin N, Bouscary D, et al. 1997. HHV-8 and multiple myeloma in France. Lancet 350:1144. Mateos MV, Garcia-Sanz R, Lopez-Perez R, et al. 2001. p16/INK4a gene inactivation by hypermethylation is associated with aggressive variants of monoclonal gammopathies. Hematol J 2:146–149. Mateos MV, Garcia-Sanz R, Lopez-Perez R, et al. 2002. Methylation is an inactivating mechanism of the p16 gene in multiple myeloma associated with

944

PART IV: CANCER BY TISSUE OF ORIGIN

high plasma cell proliferation and short survival. Br J Haematol 118: 1034–1040. Matteson EL, Hickey AR, Maguire L, Tilson HH, Urowitz MB. 1991. Occurrence of neoplasia in patients with rheumatoid arthritis enrolled in a DMARD Registry. Rheumatoid Arthritis Azathioprine Registry Steering Committee. J Rheumatol 18:809–814. McLaughlin JK, Malker HS, Linet MS, et al. 1988. Multiple myeloma and occupation in Sweden. Arch Environ Health 43:7–10. McWhorter WP, Schatzkin AG, Horm JW, Brown CC. 1989. Contribution of socioeconomic status to black/white differences in cancer incidence. Cancer 63:982–987. Mellemkjaer L, Gridley G, Moller H, et al. 1996a. Pernicious anaemia and cancer risk in Denmark. Br J Cancer 73:998–1000. Mellemkjaer L, Linet MS, Gridley G, Frisch M, Moller H, Olsen JH. 1996b. Rheumatoid arthritis and cancer risk. Eur J Cancer 32A:1753–1757. Mester B, Nieters A, Deeg E, Elsner G, Becker N, Seidler A. 2006. Occupation and malignant lymphoma: a population based case control study in Germany. Occup Environ Med 63:17–26. Milham SJ. 1971. Leukemia and multiple myeloma in farmers. Am J Epidemiol 94:507–510. Miligi L, Seniori CA, Crosignani P, et al. 1999. Occupational, environmental, and life-style factors associated with the risk of hematolymphopoietic malignancies in women. Am J Ind Med 36:60–69. Miller BA, Blair A, Reed EJ. 1994. Extended mortality follow-up among men and women in a U.S. furniture workers union. Am J Ind Med 25:537–549. Miller BA, Blair AE, Raynor HL, Stewart PA, Zahm SH, Fraumeni JFJ. 1989. Cancer and other mortality patterns among United States furniture workers. Br J Ind Med 46:508–515. Miller BA, Ries LAG, Hankey BF, et al. 1993. SEER Cancer Statistics: Review 1973–1990. NIH Pub. No. 93–2789. National Cancer Institute. Mills PK, Beeson WL, Fraser GE, Phillips RL. 1992. Allergy and cancer: Organ site-specific results from the Adventist Health Study. Am J Epidemiol 136:287–295. Mills PK, Newell GR, Beeson WL, Fraser GE, Phillips RL. 1990. History of cigarette smoking and risk of leukemia and myeloma: Results from the Adventist health study. J Natl Cancer Inst 82:1832–1836. Montella M, Crispo A, De Bellis G, et al. 2001. HCV and cancer: A case-control study in a high-endemic area. Liver 21:335–341. Morris PD, Koepsell TD, Daling JR, et al. 1986. Toxic substance exposure and multiple myeloma: A case-control study. J Natl Cancer Inst 76:987–994. Muirhead CR, Goodill AA, Haylock RG, et al. 1999. Occupational radiation exposure and mortality: Second analysis of the National Registry for Radiation Workers. J Radiol Prot 19:3–26. Nandakumar A, Armstrong BK, De Klerk NH. 1986. Multiple myeloma in Western Australia: A case-control study in relation to occupation, father’s occupation, socioeconomic status and country of birth. Int J Cancer 37:223–226. Nandakumar A, English DR, Dougan LE, Armstrong BK. 1988. Incidence and outcome of multiple myeloma in Western Australia, 1960 to 1984. Aust N Z J Med 18:774–779. Ng MH, Chung YF, Lo KW, Wickham NW, Lee JC, Huang DP. 1997. Frequent hypermethylation of p16 and p15 genes in multiple myeloma. Blood 89:2500–2506. Ogmundsdottir HM, Haraldsdottir V, Johannesson M, et al. 2002. Monoclonal gammopathy in Iceland: A population-based registry and follow-up. Br J Haematol 118:166–173. Olsen SJ, Tarte K, Sherman W, et al. 1998. Evidence against KSHV infection in the pathogenesis of multiple myeloma. Virus Res 57:197–202. Olshan AF. 1991. Familial and genetic associations. In: Obrams GI, Potter M, eds. Epidemiology and Biology of Multiple Myeloma. New York: Springer, pp. 31–39. Omar RZ, Barber JA, Smith PG. 1999. Cancer mortality and morbidity among plutonium workers at the Sellafield plant of British Nuclear Fuels. Br J Cancer 79:1288–1301. Osserman EF, Merlini G, Butler VP, J. 1987. Multiple myeloma and related plasma cell dyscrasias. JAMA 258:2930–2937. Papadopoulos NM, Lane HC, Costello R, et al. 1985. Oligoclonal immunoglobulins in patients with the acquired immunodeficiency syndrome. Clin Immunol Immunopathol 35:43– 46. Pahwa P, McDuffie HH, Dosman JA, et al. 2003. Exposure to animals and selected risk factors among Canadian farm residents with Hodgkin’s disease, multiple myeloma, or soft tissue sarcoma. J Occup Environ Med 45:857–868. Pan SY, Johnson KC, Ugnat AM, et al. 2004. Association of obesity and cancer risk in Canada. Am J Epidemiol 159:259–268. Parker KM, Ma MH, Manyak S, et al. 2002. Identification of polymorphisms of the IkappaBalpha gene associated with an increased risk of multiple myeloma. Cancer Genet Cytogenet 137:43– 48.

Parravicini C, Lauri E, Baldini L, et al. 1997. Kaposi’s sarcoma-associated herpesvirus infection and multiple myeloma. Science 278:1969–1970. Pasqualetti P, Casale R, Collacciani A, Colantonio D. 1990. Work activities and the risk of multiple myeloma. A case-control study. Med Lav 81:308–319. Pasqualetti P, Festuccia V, Collacciani A, Casale R. 1997. The natural history of monoclonal gammopathy of undetermined significance. A 5- to 20-year follow-up of 263 cases. Acta Haematol 97:174–179. Patel M, Mahlangu J, Patel J, et al. 2001. Kaposi sarcoma-associate herpesvirus/human herpesvirus 8 and multiple myeloma in South Africa. Diagn Mol Pathol 10:95–99. Patel M, Wadee AA, Galpin J, et al. 2002. HLA class I and class II antigens associated with multiple myeloma in southern Africa. Clin Lab Haematol 24:215–219. Pearce N, Reif JS. 1990. Epidemiologic studies of cancer in agricultural workers. Am J Ind Med 18:133–148. Pearce NE, Howard JK. 1986. Occupation, social class and male cancer mortality in New Zealand, 1974–78. Int J Epidemiol 15:456– 462. Pearce NE, Smith AH, Fisher DO. 1985. Malignant lymphoma and multiple myeloma linked with agricultural occupations in a New Zealand Cancer Registry-based study. Am J Epidemiol 121:225–237. Pearce NE, Smith AH, Howard JK, Sheppard RA, Giles HJ, Teague CA. 1986. Case-control study of multiple myeloma and farming. Br J Cancer 54:493–500. Pettersson T, Pukkala E, Teppo L, Friman C. 1992. Increased risk of cancer in patients with systemic lupus erythematosus. Ann Rheum Dis 51:437–439. Pickard AL, Gridley G, Mellemkjae L, et al. 2002. Hyperparathyroidism and subsequent cancer risk in Denmark. Cancer 95:1611–1617. Pottern LM, Gart JJ, Nam JM, et al. 1992a. HLA and multiple myeloma among black and white men: Evidence of a genetic association. Cancer Epidemiol Biomarkers Prev 1:177–182. Pottern LM, Heineman EF, Olsen JH, Raffn E, Blair A. 1992b. Muliple myeloma among Danish women: Employment history and workplace exposures. Cancer Causes Control 3:427–432. Preston DL, Kusumi S, Tomonaga M, et al. 1994. Cancer incidence in atomic bomb survivors. Part III. Leukemia, lymphoma and multiple myeloma, 1950–1987. Radiat Res 137:S68–S97. Pukkala E, Notkola V. 1997. Cancer incidence among Finnish farmers, 1979–93. Cancer Causes Control 8:25–33. Radl J, Liu M, Hoogeveen CM, et al. 1991. Monoclonal gammapathies in longterm surviving rhesus monkeys after lethal irradiation and bone marrow transplantation. Clin Immunol Immunopathol 60:305–309. Raffn E, Lynge E, Juel K, Korsgaard B. 1989. Incidence of cancer and mortality among employees in the asbestos cement industry in Denmark. Br J Ind Med 46:90–96. Rask C, Kelsen J, Olesen G, Nielsen JL, Obel N, Abildgaard N. 2000. Danish patients with untreated multiple myeloma do not harbour human herpesvirus 8. Br J Haematol 108:96–98. Reif J, Pearce N, Fraser J. 1989a. Cancer risks in New Zealand farmers. Int J Epidemiol 18:768–774. Reif J, Pearce N, Kawachi I, Fraser J. 1989b. Soft-tissue sarcoma, nonHodgkin’s lymphoma and other cancers in New Zealand forestry workers. Int J Cancer 43:49–54. Rettig MB, Ma HJ, Vescio RA, et al. 1997. Kaposi’s sarcoma-associated herpesvirus infection of bone marrow dendritic cells from multiple myeloma patients. Science 276:1851–1854. Rinsky RA, Hornung RW, Silver SR, Tseng CY. 2002. Benzene exposure and hematopoietic mortality: A long-term epidemiologic risk assessment. Am J Ind Med 42:474–480. Rogot E, Murray JL. 1980. Smoking and causes of death among U.S. veterans:16 years of observation. Public Health Rep 95:213–222. Said JW, Rettig MR, Heppner K, et al. 1997. Localization of Kaposi’s sarcomaassociated herpesvirus in bone marrow biopsy samples from patients with multiple myeloma. Blood 90:4278–4282. Samanic C, Gridley G, Chow WH, Lubin J, Hoover RN, Fraumeni JF Jr. 2004. Obesity and cancer risk among white and black United States veterans. Cancer Causes Control 15:35–43. Saleun JP, Vicariot M, Deroff P, Morin JF. 1982. Monoclonal gammopathies in the adult population of Finistere, France. J Clin Pathol 35:63–68. Sathiakumar N, Graff J, Macaluso M, Maldonado G, Matthews R, Delzell E. 2005. An updated study of mortality among North American synthetic rubber industry workers. Occup Environ Med 62:822–829. Schechter GP, Shoff N, Chan C, McManus CD, Hawley HP. 1991. The frequency of monoclonal gammopathy of unknown significance in Black and Caucasian veterans in a hospital population. In: Obrams GI, Potter M, eds. Epidemiology and Biology of Multiple Myeloma. New York Springer-Verlag, pp. 83–85. Schwartz DA, Vaughan TL, Heyer NJ, et al. 1988. B cell neoplasms and occupational asbestos exposure. Am J Ind Med 14:661–671.

Multiple Myeloma Selby JV, Friedman GD, Fireman BH. 1989. Screening prescription drugs for possible carcinogenicity: Eleven to fifteen years of follow-up. Cancer Res 49:5736–5747. Shapira R, Carter A. 1986. Multiple myeloma in Northern Israel, 1970–1979. Cancer 58:206–209. Shimizu Y, Kato H, Schull WJ. 1990. Studies of the mortality of A-bomb survivors. 9. Mortality, 1950–1985; Part 2. Cancer mortality based on the recently revised doses (DS86). Radiat Res 121:120–141. Shoenfeld Y, Berliner S, Shaklai M, Gallant LA, Pinkhas J. 1982. Familial multiple myeloma. Areview of thirty-seven families. Postgrad Med J 58:12–16. Sinclair D, Sheehan T, Parrott DM, Stott DI. 1986. The incidence of monoclonal gammopathy in a population over 45 years old determined by isoelectric focusing. Br J Haematol 64:745–750. Sitas F, Carrara H, Beral V, et al. 1999. Antibodies against human herpesvirus 8 in black South African patients with cancer. N Engl J Med 340:1863–1871. Skibola CF, Smith MT, Kane E, et al. 1999. Polymorphisms in the methylenetetrahydrofolate reductase gene are associated with susceptibility to acute leukemia in adults. Proc Natl Acad Sci U S A 96:12810–12815. Smith PG, Douglas AJ. 1986. Mortality of workers at the Sellafield plant of British Nuclear Fuels. Br Med J (Clin Res Ed) 293:845–854. Sonoda T, Nagata Y, Mori M, Ishida T, Imai K. 2001. Meta-analysis of multiple myeloma and benzene exposure. J Epidemiol 11:249–254. Sorahan T, Cooke MA. 1989. Cancer mortality in a cohort of United Kingdom steel foundry workers: 1946–85. Br J Ind Med 46:74–81. Soutar RL, Dawson AA, Wilson BJ. 1996. Multiple myeloma in north east Scotland: A review of incidence and survival over three decades. Health Bull (Edinb) 54:232–240. Spinelli JJ, Gallagher RP, Band PR, Threlfall WJ. 1984. Multiple myeloma, leukemia, and cancer of the ovary in cosmetologists and hairdressers. Am J Ind Med 6:97–102. Stagnaro E, Ramazzotti V, Crosignani P, et al. 2001. Smoking and hematolymphopoietic malignancies. Cancer Causes Control 12:325–334. Stark AD, Chang HG, Fitzgerald EF, Riccardi K, Stone RR. 1990. A retrospective cohort study of cancer incidence among New York State Farm Bureau members. Arch Environ Health 45:155–162. Steineck G, Wiklund K. 1986. Multiple myeloma in Swedish agricultural workers. Int J Epidemiol 15:321–325. Svec MA, Ward MH, Dosemeci M, Checkoway H, De Roos AJ. 2005. Risk of lymphatic or haematopoietic cancer mortality with occupational exposure to animals or the public. Occup Environ Med 62:726–735. Svensson BG, Englander V, Akesson B, et al. 1989. Deaths and tumors among workers grinidng stainless steel. Am J Ind Med 15:51–59. Svensson BG, Mikoczy Z, Stromberg U, Hagmar L. 1995. Mortality and cancer incidence among Swedish fishermen with a high dietary intake of persistent organochlorine compounds. Scand J Work Environ Health 21:106–115. Swaen GM Van Vliet C, Slangen JJ, Sturmans F. 1992. Cancer mortality among licensed herbicide applicators. Scand J Work Environ Health 18:201–204. Sweeney DM, Manzi S, Janosky J, et al. 1995. Risk of malignancy in women with systemic lupus erythematosus. J Rheumatol 22:1478–1482. Symmons DP. 1985. Neoplasms of the immune system in rheumatoid arthritis. Am J Med 78:22–28. Tarte K, Olsen SJ, Rossi JF, et al. 1998. Kaposi’s sarcoma-associated herpesvirus is not detected with immunosuppression in multiple myeloma. Blood 92:2186–2188. Tavani A, Pregnolato A, Negri E, et al. 1997. Diet and risk of lymphoid neoplasms and soft tissue sarcomas. Nutr Cancer 27:256–260. Tedeschi R, Kvarnung M, Knekt P, et al. 2001. A prospective seroepidemiological study of human herpesvirus-8 infection and the risk of multiple myeloma. Br J Cancer 84:122–125. Tennis P, Andrews E, Bombardier C, et al. 1993. Record linkage to conduct an epidemiologic study on the association of rheumatoid arthritis and lymphoma in the Province of Saskatchewan, Canada. J Clin Epidemiol 46:685–695. Teta MJ, Ott MG. 1988. A mortality study of a research, engineering, and metal fabrica tion facility in western New York State. Am J Epidemiol 127:540–551. Teta MJ, Walrath J, Meigs JW, Flannery JT. 1984. Cancer incidence among cosmetologists. J Natl Cancer Inst 72:1051–1057. Thun MJ, Altekruse SF, Namboodiri MM, Calle EE, Myers DG, Heath CW, Jr. 1994. Hair dye use and risk of fatal cancers in U.S. women. J Natl Cancer Inst 86:210–215. Tisdale JF, Stewart AK, Dickstein B, et al. 1998. Molecular and serological examination of the relationship of human herpesvirus 8 to multiple myeloma: orf 26 sequences in bone marrow stroma are not restricted to myeloma patients and other regions of the genome are not detected. Blood 92:2681–2687. ’t Mannetje A, McLean D, Cheng S, Boffetta P, Colin D, Pearce N. 2005. Mortality in New Zealand workers exposed to phenoxy herbicides and dioxins. Occup Environ Med 62:34–40.

945

Tollerud DJ, Brinton LA, Stone BJ, Tobacman JK, Blattner WA. 1985. Mortality from multiple myeloma among North Carolina furniture workers. J Natl Cancer Inst 74:799–801. Tollerud DJ, Brown LM, Blattner WA, Hoover RN. 1991. The influence of race on T-cell subset distributions. In: Obrams GI, Potter M, eds. Epidemiology and Biology of Multiple Myeloma. New York: Springer-Verlag, pp. 45–49. Tomasek L, Darby SC, Swerdlow AJ, Placek V, Kunz E. 1993. Radon exposure an cancers other than lung cancer among uranium miners in West Bohemia. Lancet 341:919–923. Turesson I, Zettervall O, Cuzick J, Waldenstrom JG, Velez R. 1984. Comparison of trends in the incidence of multiple myeloma in Malmo, Sweden, and other countries, 1950–1979. N Engl J Med 310:421–424. Vagero D, Persson G. 1986. Occurrence of cancer in socioeconomic groups in Sweden. An analysis based on the Swedish Cancer Environment Registry. Scand J Soc Med 14:151–160. van de Donk N, de Weerdt O, Eurelings M, Bloem A, Lokhorst H. 2001. Malignant transformation of monoclonal gammopathy of undetermined significance: Cumulative incidence and prognostic factors. Leuk Lymphoma 42:609–618. van de Poel MH, Coebergh JW, Hillen HF. 1995. Malignant transformation of monoclonal gammopathy of undetermined significance among outpatients of a community hospital in southeastern Netherlands. Br J Haematol 91:121–125. Van Den Eeden SK, Friedman GD. 1993. Exposure to engine exhaust and risk of subsequent cancer. J Occup Med 35:307–311. Velez R, Beral V, Cuzick J. 1982. Increasing trends of multiple myeloma mortality in England and Wales; 1950–79: Are the changes real? J Natl Cancer Inst 69:387–392. Vena JE, Fiedler RC. 1987. Mortality of a municipal-worker cohort: IV. Fire fighters. Am J Ind Med 11:671–684. Vesterinen E, Pukkala E, Timonen T, Aromaa A. 1993. Cancer incidence among 78,000 asthmatic patients. Int J Epidemiol 22:976–982. Vineis P, Crosignani P, Sacerdote C, et al. 2000. Haematopoietic cancer and medical history: A multicentre case control study. J Epidemiol Community Health 54:431–436. Wang CR, Chuang CY, Lin KT, et al. 1992. Monoclonal gammopathies and the related autoimmune manifestations in Taiwan. Asian Pac J Allergy Immunol 10:123–128. Wang JX, Boice JD, Jr., Li BX, Zhang JY, Fraumeni JF, Jr. 1988. Cancer among medical diagnostic x-ray workers in China. J Natl Cancer Inst 80:344– 350. Wiklund K, Dich J, Holm LE, Eklund G. 1989. Risk of cancer in pesticide applicators in Swedish agriculture. Br J Ind Med 46:809–814. Wilczynska U, Szadkowska-Stanczyk I, Szeszenia-Dabrowska N, Sobala W, Strzelecka A. 2001. Cancer mortality in rubber tire workers in Poland. Int J Occup Med Environ Health 14:115–125. Williams RR, Horm JW. 1977. Association of cancer sites with tobacco and alcohol consumption and socioeconomic status of patients: Interview study from the Third National Cancer Sruvey. J Natl Cancer Inst 58:525–547. Wing S, Richardson D, Wolf S, Mihlan G, Crawford-Brown D, Wood J. 2000. A case control study of multiple myeloma at four nuclear facilities. Ann Epidemiol 10:144–153. Wolvekamp MC, Marquet RL. 1990. Interleukin-6: Historical background, genetics and biological significance. Immunol Lett 24:1–9. Wong O. 1995. Risk of acute myeloid leukaemia and multiple myeloma in workers exposed to benzene. Occup Environ Med 52:380–384. Wong O, Harris F. 2005. Retrospective cohort mortality study and nested casecontrol study of workers exposed to creosote at 11 wood-treating plants in the United States. 47:683–697. Yi Q, Ekman M, Anton D, et al. 1998. Blood dendritic cells from myeloma patients are not infected with Kaposi’s sarcoma-associated herpesvirus (KSHV/HHV-8). Blood 92:402–404. Youinou P, Le Corre R, Dueymes M. 1996. Autoimmune diseases and monoclonal gammopathies. Clin Exp Rheumatol 14 Suppl 14:S55–S58. Zahm SH, Weisenburger DD, Babbitt PA, Saal RC, Vaught JB, Blair A. 1992. Use of hair coloring products and the risk of lymphoma, multiple myeloma, and chronic lymphocytic leukemia. Am J Public Health 82:990–997. Zheng C, Huang D, Liu L, et al. 2001a. Cytotoxic T-lymphocyte antigen-4 microsatellite polymorphism is associated with multiple myeloma. Br J Haematol 112:216–218. Zheng C, Huang D, Liu L, et al. 2001b. Interleukin-10 gene promoter polymorphisms in multiple myeloma. Int J Cancer 95:184–188. Zheng C, Huang DR, Bergenbrant S, et al. 2000. Interleukin 6, tumour necrosis factor alpha, interleukin 1beta and interleukin 1 receptor antagonist promoter or coding gene polymorphisms in multiple myeloma. Br J Haematol 109:39–45.

48

Bone Cancer ROBERT W. MILLER, JOHN D. BOICE, JR., AND ROCHELLE E. CURTIS

C

ancers that arise from bone or cartilage account for about 0.5% of all malignant neoplasms in humans. Great progress has been made recently in understanding the genesis of these cancers, beginning with clues from clinical observations and epidemiology.

DEMOGRAPHY Descriptive studies in the past have been handicapped because all cell types were combined under “Bone Cancer” in the International Classification of Diseases (ICD). When classified by cell type one can see that they differ epidemiologically, which reflects differences in etiology and pathogenesis. The three main types are osteosarcoma, which arises most often from the growing ends of long bones; chondrosarcoma, which develops in cartilage; and Ewing sarcoma, most commonly in the axial skeleton. Histologic diagnoses are required, as from population-based cancer registries. Of particular value in this regard are data from the Surveillance, Epidemiology and End Results (SEER) Program of the National Cancer Institute (Percy et al., 1995), which has covered about 10% of the US population since 1973. Ninety-seven percent of bone cancers were histologically confirmed. The locations of SEER registries used here are shown in the footnote to Table 48–1. There were 3634 primary bone cancers among whites and 356 among blacks registered in the SEER Program from 1980 through 2000. Among whites, osteosarcoma was reported in 32%, chondrosarcoma in 30%, and Ewing sarcoma in 16%. Age-adjusted rates by histologic type are presented in Figure 48–1. Osteosarcoma has a bimodal age-distribution with peaks in adolescence and late in life (Fig. 48–2). Study by single year of age revealed that the rate for males exceeded that for females at 13 years of age as it rose to a peak at 15–19 years (Miller, 1981). In 1958 Price first noted the apparent relationship between bone cancer and the adolescent growth spurt. The peak late in life especially for males (Fig. 48–2) is attributed to Paget disease. There is no peak among Japanese, who rarely develop this disease (Ishikawa et al., 1996). It has also long been known that giant breeds of dogs have much higher relative risks of osteosarcoma than do small dogs (Priester and Mantel, 1971). The relative risks of bone cancer in St. Bernards and Great Danes were 8.8 and 5.7, respectively, compared with 0.2 for toy poodles and 0.6 for mixed breeds (Priester and McKay, 1980). Osteosarcomas account for 80% of canine bone cancers. Chondrosarcoma is rare in childhood and rises with advancing age for unknown reasons. Studies of substantial case series have been made of clinical course and treatment (e.g., Bjornsson et al., 1998), but not for associated disease (past history, physical examination, or subsequent disease). The age distribution of Ewing sarcoma resembles that of osteosarcoma early in life, but it rarely develops over 35 years of age (Fig. 48–2). In 1986 Dehner proposed that it be designated a primitive neuroectodermal tumor of bone (PNET). It is derived from the neural crest, and identified by a distinctive chromosomal translocation and immunohistochemical staining (Dagher et al., 2001; Ginsberg et al., 2002). The translocation is in somatic cells, not the germ line, and is therefore not transmitted from parent to child. Ewing sarcoma now refers to the much more common undifferentiated form of the neoplasm, and PNET refers to the differentiated form. The SEER

946

registries have only nine cases since the term was introduced in 1990, through 2000. In a case-control study of 208 cases of Ewing sarcoma no clues to etiology were found (Winn et al., 1992). The study included data on demography, maternal reproduction, child health, parental occupation, and family history. Chordoma, thought to arise from vestigeal notochord, was the diagnosis for 400 cases registered by SEER, 1973–1995 (McMaster et al., 2001). The rates rose steadily to a peak at 70–79 years of age. The anatomic distribution was about equal for the cranium, spine, and sacrum. The median age for cranial occurrence, however, was 49 years as contrasted with 69 years for sacral chordomas. Table 48–1 shows three chordomas in blacks, when about 13 were expected if the ratio of blacks to whites is the same as it is for osteosarcoma. Among blacks there is also a rarity of giant cell and vascular tumors (Table 48–1). Fibrosarcoma and malignant fibrous histiocytomas are rare malignant bone tumors arising from the fibrous elements of the medullary cavity of bones. In the SEER program 188 were listed, 1980–2000 (Table 48–1). The numbers are changing as new entities are being defined through molecular biology (Sandberg and Bridge, 2003). There is a male predominance of each major form of bone cancer among whites and blacks (Fig. 48–1). The two races have similar incidence rates for childhood osteosarcoma, but blacks have almost no cases of Ewing sarcoma, either in the United States (Fig. 48–1) or Africa (Parkin et al., 1999). Rates of Ewing sarcoma are also low among Asians, but less so than in blacks (Table 48–2). These racial differences indicate that gene mutation for osteosarcoma occurs in both races, but in Ewing sarcoma mutation is rare in blacks and Asians. SEER data show that chondrosarcoma is substantially less common in blacks than in whites (Fig. 48–1). So too is fibrosarcoma of bone. Mortality rates, 1969–2000, for all forms of bone cancer combined reached a plateau in 1985 (Fig. 48–3). Five-year relative survival rates for each of the three main types of bone cancer were substantially better for females than males, 1985–2000, as shown in Fig. 4A and 4B. The survival rates for childhood osteosarcoma, Ewing sarcoma, and chondrosarcoma in seven European countries, sexes combined, had averages of 70%, 59%, and 63% in 1985–1989, compared with 60%, 56%, and 59% in 1978–1989 (Stiller et al., 2001). Neither the United States (all ages) nor Europe (children) has shown much improvement in survival over the past 15 years.

ENVIRONMENTAL FACTERS Ionizing Radiation The main environmental exposures that induce osteosarcomas are ionizing radiation (Table 48–3) and/or certain types of chemotherapy. High-dose exposures to radionuclides occurred in radium dial painters (US), patients treated with intravenous 224-radium for bone tuberculosis or ankylosing spondylitis (Germany), others given intravenous Thorotrast for arteriography (various countries), and occupational and environmental exposure to plutonium in and around a nuclear weapons factory near Siberia. Table 48–3 shows the numbers of people exposed, how many are known to have developed bone sarcomas, the

Table 48–1. Number of Patients with Primary Bone Cancer Among Whites and Blacks According to Histologic Type, 9 SEER Cancer Registries, 1980–2000* All Races Histology Bones and joints, malignant Fibrosarcoma (8810–8814, 8830–8831) Fibrous histiocytoma, malignant (8830–8831) Other (8810–8814) Osteosarcoma (9180–9200) Osteosarcoma, NOS (9180) Chondroblastic osteosarcoma (9181) Fibroblastic osteosarcoma (9182) Telangiectatic osteosarcoma (9183) Osteosarcoma in Paget disease of bone (9184) Juxtacortical osteosarcoma (9190) Other osteosarcoma (9185, 9200) Chondrosarcoma (9229–9240) Ewing sarcoma (9260, 9364) Giant cell sarcoma (9250) Adamantinoma of long bones (9261) Hemangiosarcoma and malignant hemangio endothelioma (9120–9133) Chordoma (9370) Sarcoma, NOS (8800–8803) Unspecified (8000–8004) All other types

Frequency †

4304 188 136 52 1492 1117 139 76 34 35 83 8 1215 627 63 22 55 241 116 116 169

Whites

Blacks

%

Frequency

%

Frequency

%

100.0 4.4 3.2 1.2 34.7 26.0 3.2 1.8 0.8 0.8 1.9 0.2 28.2 14.6 1.5 0.5 1.3 5.6 2.7 2.7 3.9

3634 164 123 41 1163 865 110 60 26 33 62 7 1082 580 42 14 47 218 95 96 133

100.0 4.5 3.4 1.1 32.0 23.8 3.0 1.7 0.7 0.9 1.7 0.2 29.8 16.0 1.2 0.4 1.3 6.0 2.6 2.6 3.7

356 17 9 8 187 144 18 9 3 1 11 1 70 8 11 5 6 3 11 15 23

100.0 4.8 2.5 2.3 52.5 40.5 5.1 2.5 0.8 0.3 3.1 0.3 19.7 2.3 3.1 1.4 1.7 0.8 3.1 4.2 6.5

*Data from 9 SEER registries, including the states of Connecticut, Hawaii, Iowa, New Mexico, Utah and the metropolitan areas of Detroit, Atlanta, Seattle, and San Francisco–Oakland. Histologic categories based on Dorfman and Czerniak, 1995. † In addition to 3634 whites and 356 blacks, total includes 277 bone cancers for other races and 37 for unknown race.

Figure 48–1. Age-adjusted incidence rates for bone cancer by sex, race (whites, blacks), and histologic type (12 SEER areas, 1992–2000). Rates are per 100,000 and age-adjusted to the 2000 US standard population by 5-year age groups. Age-adjusted incidence rates not shown in graph: Hispanic females: osteosarcoma, 0.31; chondrosarcoma, 0.19; Ewing sarcoma, 0.069. Hispanic males: osteosarcoma, 0.31; chondrosarcoma, 0.22; Ewing sarcoma, 0.085. Hispanic is not mutually exclusive from whites, blacks, Asian/Pacific Islanders. Incidence rates for Hispanics exclude registries from Detroit, Hawaii, and Alaska native registry. Ageadjusted incidence rates for Asian/Pacific Islanders are available for osteosarcoma: females, 0.28; and males, 0.34.

Figure 48–2. Age- and sex-specific incidence rates for three major types of bone cancer (9 SEER areas, 1980–2000, all races combined).

947

948

PART IV: CANCER BY TISSUE OF ORIGIN

Table 48–2. Geographic Variation in Age-Standardized Incidence Rates (ASR) among Children (Ages 0–14 years) for Osteosarcoma and Ewing Sarcoma, Both Sexes Combined, All Races* Osteosarcoma Location US SEER, Whites Canada England & Wales Australia Germany France, Lorraine Czech Republic Spain Puerto Rico Cuba Los Angeles, Hispanic India, Bangalore India, Bombay India, Madras India, Poona Japan Hong Kong Korea, Seoul Singapore, Chinese Thailand US SEER, Blacks Namibia

Ewing Sarcoma

Cases

ASR

Cases

ASR

136 160 284 94 216 58 70 46 29 20 27 18 65 21 19 113 48 25 15 20 23 15

3.3 2.9 2.6 2.2 3.0 2.7 2.6 2.7 2.7 2.0 3.8 1.2 1.6 1.4 1.8 2.3 2.6 2.7 2.9 1.1 3.3 2.7

111 132 235 116 174 62 65 56 24 21 15 18 62 22 15 44 6 4 4 4 2 3

2.7 2.4 2.2 2.9 2.4 3.0 2.5 3.6 1.8 1.6 2.1 1.2 1.6 1.6 1.5 1.0 0.4 0.5 0.8 0.2 0.3 0.5

used to enhance the glow. The women used their lips to make the brushes come to a point. The daily intake before 1925 was estimated to be 3–43 mg. Dr. T. Blum, a dentist in Orange, NJ, is credited with calling attention to a dial painter with osteomyelitis of the jaw in 1923. Two years later, Martland, the chief medical examiner for Essex County, NJ, wrote the first of a series of reports on the autopsies of dial painters who died of illnesses related to their work. The first two cases of osteosarcoma were found in 1924 and 1928 (Martland et al., 1925; Fry, 1998). Then, in 1938 the first malignancy of the epithelial lining of paranasal or mastoid sinuses was reported. Centers for the study of radium-induced tumors were established by the Atomic Energy Commission at the New Jersey Department of Health, MIT and the Argonne National Laboratory. Among 2403 persons who had a measured exposure occupationally or therapeutically, 60 developed osteosarcomas and 32 developed carcinoma of paranasal or mastoid sinuses (Rowland, 1995; Carnes et al., 1997; Stebbings, 2001). Radium decays by emitting high-LET alpha particles, and the resulting bone dose can be very high. Among 759 women whose bone doses have been determined (1700 cGy average), an S-shaped dose-effect curve was found. A linear relationship could be rejected. The downturn at very high doses may be due to cell killing or an otherwise inability of cells to divide. Interestingly, no osteosarcomas occurred below a dose of about 1000 cGy (Priest, 2001). It has been suggested that a practical threshold could exist for cancer induced by radium isotopes (NAS, 1972), which is consistent with recent evaluations of the US radium dial study (Rowland, 1995). A threshold would explain why only 1 osteosarcoma was reported among 1203 radium

Source: Parkin, et al. International Incidence of Childhood Cancer, Volume II, IARC Scientific Publication No. 144. IARC, Lyon, France: 1998. *Age-standardized rate per million, standardized to World standard. Time periods covered varied by registry, but generally included the calendar years 1980–1989 or portions of this period.

average bone dose, and relative risk. Note that radiotherapy for cancer has induced osteosarcoma, but A-bomb exposure in Japan and occupational exposures in other countries have not. Ewing sarcoma is not induced by ionizing radiation.

226,228-Radium In about 1913 radium was first painted on dials of watches and clocks in the United States to make them luminous. The history, as described by Fry (1998), revealed that the industry hired 4318 women, 1915–1979. 226-Radium and later 228-Radium (mesothorium) were (a)

Figure 48–3. Age-adjusted mortality rates for bone cancer for white males and females, 1969–2000. Points represent 4-year calendar-year groupings. Rates are per 100,000 and age adjusted to the 2000 US standard population by 5-year age groups. Break in graph indicates change in mortality coding in 1980 with the introduction of ICD-9, which reclassified bone cancers not specified as primary to metastatic.

(b)

Figure 48–4. Relative survival rates (%) for males (a) and females (b) with bone cancer, by histologic type (all races, 9 SEER areas, 1985–1999, follow-up for vital status to 2000).

Table 48–3. Epidemiologic Studies of Radiation-Induced Bone Cancer Study

Type of Exposure

Number Exposed

Duration of Person-Years Follow-up (years) at Risk

Average Bone Dose (cGy)

Relative Risk (O/E)

adult–external radiation 1. Cervical cancer (Boice et al., 1985a, 1988) 2. Cervical cancer (Kleinerman et al., 1995) 3. Bone marrow transplanta (Curtis et al., 1997) 4. Hodgkin’s diseasea (Dores et al., 2002) 5. Ankylosing spondylitis (Weiss et al., 1994) 6. Benign gynecological disease (Inskip et al., 1990) 7. Metropathia hemorrhagica (Darby et al., 1994) 8. Atomic bomb survivors (Thompson et al., 1994; UNSCEAR, 2000) 9. Radiation workers—3 countries (Cardis et al., 1995) 10. Radiological technologists—US (Doody et al., 1998)

External radiotherapy Brachytherapy External radiotherapy Brachytherapy External radiotherapy

82,616

1–35

547,222

2,200

11/5.7

49,828

1–30+

532,740

2,200

17/5.7

~10,200

0–10+

~33,000

~1,000

5/0.4

External radiotherapy

13,793

1–25+

115,794

~3,500

23/0.5

External radiotherapy

14,556

5–25+

355,350

454–1,856

9/2.7

Brachytherapy (intrauterine 226Radium) External radiotherapy

4,153

1–60

109,911

~150

0/2.7

2,067

1–30+

~58,000

~310

0/0.9

Gamma rays, neutrons

79,972

13–42

1,950,567

~23

16/12.1

Gamma rays

95,673

1–30+

2,124,526

4

11/11i

143,429

2–40+

2,839,466

NA

4/10.3

External radiotherapy

9,170

2–29

50,609

0–7,000

48/0.36

External radiotherapy

13,175

3–35+

~141,000

~1,000

54/1.3

External radiotherapy

13,581

5–25

140,792

NA

28/1.5

External radiotherapy

4,257

3–48

63,900

0–7,000

32/0.32

X-rays, gamma rays

childhood/fetal—external radiation 11. Childhood cancer—LESGb,c (Tucker et al., 1984) 12. Childhood cancer—UKb,c (Hawkins et al., 1996) 13. Childhood cancer—CCSSb,c (Neglia et al., 2001) 14. Childhood cancer—France, UKb (Le Vu et al., 1998) 15. Retinoblastoma—NYb (Abramson et al., 1984) 16. Retinoblastoma—UKb (Hawkins et al., 1996; Draper et al., 1986) 17. Retinoblastoma—NY, MAb,d (Wong et al., 1997) 18. Childhood cancer—LESG (Tucker et al., 1987) 19. Hemangioma—Stockholm (Fürst et al., 1988) 20. Hemangioma—Göteborg (Lindberg et al., 1995) 21. Oxford survey (Bithell and Stewart, 1975)

External radiotherapy Radium plaque External radiotherapy Radium plaque

688

1–32

~5,000

3,000–15,000

45/small

504b

3–35+

~9,000

~2,000

24/0.13

External radiotherapy Radium plaque External radiotherapy

961b

1–50

~30,000

3,280

65/0.15

2,690

2.7

226

Radium applicator Orthovoltage X-ray 226 Radium applicator Prenatal X-ray

b

84% of 64 cases 73% of 209 controls 15,336 11,807 10.7% of 244 cases 9% of 244 controls

NA

NA

1–62

323,895

~1–4,000

3/2.7

1–59

370,517

~1–4,000

2/2.8

<10

1.11

NA

NA

radionuclides—internal radiation 22. Radium-dial painterse (Rowland et al., 1978; Polednak et al., 1978) 23. Bone disease—Germanyf (Spiess et al., 1989; Nekolla et al., 2000) 24. Ankylosing spondylitis—Germanyf (Wick and Gössner, 1989; Wick et al., 1999) 25. Thorotrast patients—Portugalg (dos Santos Silva et al., 2003) 26. Thorotrast patients—Germanyg (van Kaick et al., 1989; 1999) 27. Thorotrast patients—Denmarkg (Faber, 1979; Andersson et al., 1995) 28. Thorotrast patients—Swedeng,h (Nyberg et al., 2002) 29. Mayak workers—Plutonium (Koshirnakova et al., 2000) 30. UK plutonium workers (Omar et al., 1999) 31. Uranium workers (CRS, 2001)j

a

226

Radium, 228Radium

757

1–56

33,597

1,700

38/0.4

224

Radium

899

1–43

21,600+

~3,600f

56/<1

224

Radium

1,577

1–38

~32,800

500f

4/1.3

1,096

1–50+

16,717

26.3 mlg

5/0.7

2,326

1–45

>20,000

26.3 mlg

4/1.2

999

1–50+

19,365

~20 mlg

0/0.2

g

16.1 ml

2/1.2

232

Thorium (translocating 224 Ra, 228Th, 228Ra) As above As above As above 239

Plutonium, External gamma 239 Plutonium, External gamma 238 U, 235U

432

10–40+

7,284

11,000

1–40+

427,500

4–14,400

24/13.3

5,203

1–45

134,817

~3j

0/1.1

120,000

1–50

NA

30/32.3

NA

Includes children. Includes genetic retinoblastoma; these children are genetically predisposed to develop osteosarcoma. Radiotherapy status not reported. d 643 non-hereditary retinoblastoma cases were similarly followed and the O/E/ was 0/0.11. e Only data for women with known doses are presented. Overall 61 bone sarcomas have occurred in 1,474 women employed prior to 1930. f Bone surface dose. Skeletal dose is factor of 9 lower (Nekolla et al., 2000). g Amount of Thorotrast administered. h Includes soft tissue sarcoma. i Negative dose response. j A radiation weighting factor of 20 was used (~630 mSv was average equivalent dose). NA = Not applicable or not available. ~ = Estimated or approximate. b c

950

PART IV: CANCER BY TISSUE OF ORIGIN

luminizers studied in the United Kingdom who ingested much lower amounts of 226,228-Ra than US dial painters did (Baverstock and Papworth, 1989).

224-Radium Peteosthor, a colloidal drug containing 224-Ra, platinum, and eosin dye was used in Germany after World War II to treat bone tuberculosis and ankylosing spondylitis (Spiess, 2002). The short half-life (3.8 days) indicates that radiation dose to bone would have been received over a period of about 1 month as opposed to over a lifetime for 226Ra. Among 899 patients injected with 224-Ra, 56 (6%) developed bone cancer, particularly osteosarcoma of the fibroblastic or fibrohistiocytic type (Gossner, 1999), with latent periods from 4 to 22 years (Spiess et al., 1989). The average dose to bone surface was approximately 3600 cGy. Subsequent follow-up revealed 56 malignant bone tumors as compared with 1 case expected (Nekolla et al., 2000). The pattern of incidence over time for 224-Ra–induced bone sarcomas was generally similar to that observed for leukemia in A-bomb survivors, spondylitis patients given radiotherapy, and cervical cancer patients treated with radiation. Excess bone sarcomas appeared within 5 years after injection, peaked between 6 to 8 years, and decreased to normal levels after about 33 years. This wave of induced cancers is contrary to observations on other forms of solid cancer, which have a sustained rise. The lowest skeletal dose was 90 cGy; however, the dose to bone surface, where most of the energy was deposited, was approximately 800 cGy (Nekolla et al., 2000). Four bone cancers compared to 1.3 expected were found in a series of 1577 patients treated with lower doses of 224-Ra (Wick and Gossman, 1989; Wick et al., 1999). Dose calculations and pathology have been recently re-evaluated (Leenhouts and Brugmans, 2000); excesses of fibrosarcoma of bone were observed in patients who received high doses of 226/228-Ra, 224-Ra, or external radiotherapy (Gossner, 1999). The difference in carcinogenic effectiveness, as well as in latency period and shape of the dose-effect curve, appears to be related to the different radium isotopes involved. The isotope 226-Ra has a long half-life (1600 years) and deposits its energy throughout the entire bone, whereas 224-Ra deposits its energy almost entirely on the bone surfaces. The long latent periods observed are probably related to continuous irradiation of bone by 226-Ra.

Plutonium In 1948 the Soviet Union established the Mayak Production Association to make plutonium nuclear weapons. The Mayak facility is located in the Southern Urals on the Techa River. From 1949–1956 about 3 million curies of liquid waste were discharged by the factory into the river (Dicus, 1997; Anspaugh et al., 2002). The primary exposure to the population living downstream was to external gamma rays from short-lived fusion products and to internal beta particles from ingested 90-strontium (Kellerer, 2002). Between 1950–1989, 12 deaths from bone malignancies were observed in a cohort of 26,485 residents in the Techa River region (Kossenko et al., 1997). Risk estimates are difficult at the moment because of incomplete information on vital status and uncertainties in the dose estimates (UNSCEAR, 2000). Other contamination occurred. In 1957 chemicals exploded in a high-level storage tank, and about 20 million curies of radioactivity were thrown from the tank, and 2 million curies were carried downwind. About 270,000 people were exposed to the fallout. In 1967 the last major accident occurred when a reservoir used to store waste evaporated after a dry hot summer, and windstorms carried radioactive dust (600,000 curies) over 2700 square kilometers. These and other misadventures could have brought the total to billion curies (Dicus, 1997). The radiation level of the discharge at the Mayak facility in 1951 was 180 rem per hour (5.4 rem per hour downstream). Over time about 21,500 workers were exposed to radiation (highest dose, greater than 10 Gy, mean dose 0.8 Gy) (Shilnikova et al., 2003). Nearly 11,000 workers were exposed to 239-plutonium with bone doses ranging from 0.04 to 14.4 Gy (Koshurnikova et al., 2000). The half-life of 239plutonium is 24,000 years.

Plutonium concentrates in liver, lungs, and bone where its alpha particles bombard tissue, and at sufficiently high doses, can raise cancer rates in these organs. In a study of 11,000 workers who began working at Mayak in 1948–1958 (exposed to external radiation and particles internally), 16 osteosarcomas and 8 chondrosarcomas occurred. Only one osteosarcoma developed in a worker hired after 1958 (Koshurnikova et al., 2000). Also there were nine soft tissue sarcomas adjacent to bone. Most of these cancers occurred at least 20 years after the date of first hire. Statistical analyses showed an RR of 7.9 for the four cases with body-burdens of 7.4 kBq or greater, and 4.1 for the seven other workers who were exposed but not monitored (Koshurnikova et al., 2000). In addition, some of these workers had external doses greater than 1 Gy. In these analyses, the data on 16 osteosarcomas and 3 chondrosarcomas were combined with 8 soft tissue sarcomas. The plutonium exposures received by the Mayak workers were enormous and much larger than experienced anywhere else in the world (IARC, 2001). Other studies of workers exposed to substantially lower plutonium doses have found no excess bone cancers (Omar et al., 1999; IARC, 2001; Voltz et al., 1997).

Thorotrast Another disaster began in 1932 when Thorotrast was introduced in radiology as a contrast medium for arteriography. It was given by intravenous injection and had essentially no acute side effects. The radioactive particles were, however, permanently retained in reticuloendothelial cells, particularly in the liver and spleen. Thorotrast is a colloidal suspension of millimicron-size thorium oxide particles. Thousands of patients have been followed up for cause of death. The main sites of cancer were the liver and bile ducts, lung, and other organs where the concentration of Thorotrast was high or prolonged (e.g., the brain and urinary bladder). Bone cancer in Thorotrast recipients was found in five Portuguese (dos Santos Silva et al., 2003), four Germans (van Kaick et al., 1999), two Japanese (Mori et al., 1999), and two Swedes (Nyberg et al., 2002), but no Danes (Andersson et al., 1995). The cell types, given only for the Japanese, were osteosarcoma and hemangiosarcoma. By 1979, seven case reports of osteosarcoma had also been published on patients after Thorotrast (Harrist et al., 1979). One was a 24year-old male who had been injected with the contrast medium at 2 years of age. At autopsy his vertebrae showed dense sclerosis of the innermost bone (exposed at 2 years of age), surrounded by normal bone laid down in the 22 years since then (Fig. 48–5A). Autoradiography of a lymph node was teeming with alpha radiation tracks (Fig. 5B) (Miller, 1985). Sindelar et al. (1978), who first reported this case, did not publish the Figure. Thorium has a half-life of 14 billion years. In its natural state, thorium is excreted rather rapidly from the body, but the colloid Thorotrast remained essentially for life. Estimated dose to the skeleton from 25 ml of Thorotrast is about 4 Gy (Kathren and Hill, 1992).

Radiotherapy Although primary cancers of bone have been associated with external high-dose radiation, used especially in the therapy of various cancers, the fraction of bone cancers that result from this exposure appears to be small (i.e., less than 0.05%–0.02% of patients treated) (Boice et al., 1985a, 1985b; Robinson et al., 1988). Between 1935 and 1982 in Connecticut, 30 bone cancers developed vs. 17 expected in 253,536 patients treated for cancer (30.8% of whom received radiotherapy) (Boice et al., 1985b). Among 379,941 cancer patients in Denmark treated between 1946 and 1980, 43 bone cancers occurred vs. 23 expected. Among 82,616 patients with cervical cancer treated with high-dose pelvic radiation, only 11 bone cancers were reported vs. 5.7 expected (Boice et al., 1985a); the dose to exposed bone was estimated to be 2200 cGy (rad) on average (Boice et al., 1988). In each of these three studies, 0.01% of the patients given radiotherapy developed bone sarcoma—about twice the expected frequency. Excess bone cancers have also been reported following radiotherapy for breast cancer (Doherty et al., 1986), Hodgkin disease (Woodard et al., 1988; Dores et al., 2002), cervical cancer (Kleinerman et al., 1995), and childhood cancer (Hawkins et al., 1996; Neglia et al., 2001).

Bone Cancer

951

A

B

Children with certain cancers, however, seem to be particularly susceptible to radiogenic bone cancer. High risks have been reported following treatment, primarily with radiation, for retinoblastoma (12 vs. 0.01 expected), Wilms tumor (6 vs. 0.05), Hodgkin disease (5 vs. 0.05), neuroblastoma (4 vs. 0.03), and Ewing sarcoma (7 vs. 0.01) (Tucker et al., 1984). A radiation dose response from the Late Effects Study Group was based on 64 secondary osteosarcomas from 9120 two-year childhood cancer survivors as compared with 209 matched controls (Tucker et al., 1987). A recent study was made of 32 secondary osteosarcomas among 4400 three-year survivors of childhood cancer in France and Britain, compared with 160 matched controls for dose-response analyses (Le Vu et al., 1998). In comparison with general population rates, the observed vs. expected numbers of primary tumors were excessive after soft tissue sarcomas (11 vs. 0.04), Ewing sarcoma (8 vs. 0.01), and bilateral retinoblastoma (5 vs. 0.00). Only Ewing sarcoma and bilateral retinoblastoma were excessive in both studies, seemingly due to an interaction of environment and host susceptibility. Several other large series of patients treated for heritable retinoblastoma reveal very large increases in osteosarcoma (Abramson et al., 1984; Draper et al., 1986; Hawkins et al., 1996; Wong et al., 1997).

Figure 48–5. Thorotrast effects, 24-year-old male. (A) Transverse section of a vertebral body showing sclerosis of the innermost portion due to a-irradiation since 2 years of age. Normal bone was laid down in the subsequent 22 years. (B) Lymph node, autoradiograph, showing short linear alpha radiation tracks.

Dose-response data rarely find increases below 5–9 Gy for heritable retinoblastoma treatments (Tucker et al., 1987; Wong et al., 1997). Perhaps the greatest risk of radiogenic cancer is in marrow transplant patients after the administration of immunosuppressive drugs and high exposure to total-body irradiation. The cancers include osteosarcoma reported in eight children after allogeneic marrow transplantation, four of them studied by Bielack et al. (2003). All four had received chemotherapy with alkylators, and three had totalbody irradiation. The cancers were diagnosed 32–80 months after transplantation. In a study of nearly 20,000 patients who received bone marrow transplantation in the United States, 1964–1992, one (included in the previously mentioned case series) had osteosarcoma, and three (not mentioned there) developed chondrosarcomas (Curtis et al., 1997). This finding led us to look for chondrosarcomas in other series of radiogenic cancers. We found that Gossner (1999) had made a comprehensive review that provided new information on chondrosarcomas and fibrosarcomas in various groups who were heavily exposed to ionizing radiation. Table 48–4 shows the relative frequencies of osteosarcomas, chondrosarcomas, and fibrosarcomas after various types of radiation exposure. Osteosarcomas after heavy radiation exposures

952

PART IV: CANCER BY TISSUE OF ORIGIN Table 48–4. Bone Cancer Cell Types: Numbers and Percentages by Sources of Radiation Exposure (Three Cell Types Add to 100%) Osteosarcoma Source 224-Radium 226/228-Radium Plutonium Marrow implant External irradiation General pop, white

Chondrosarcoma

Fibrosarcoma

Number

%

Number

%

Number

%

References

22 32 16 8 155 1163

53 70 80 73 63 48

6 0 3 3 9 1082

14 0 15 27 4 45

14 14 1 0 82 164

33 30 5 0 33 7

Gossner Gossner Koshurnikova Bielack/Curtis Gossner Table 48–1

were more common than they are in the general population (53–80% vs. 48%). Chondrosarcoma has not been reported in radium-dial painters. It was far less common among the other irradiated groups than in the general population (4–27% vs. 45%). Fibrosarcoma of bone was not described as an entity until 1972 and the first radiogenic case was recognized in 1977 (reviewed by Gossner, 1999). Most radiation-induced bone tumors were studied before that date, but reclassification of the pathology has been made in recent years based on stored specimens (Gossner, 1999). In the SEER Program the frequency of fibrosarcoma in the general US white population was only 7% compared with 30–33% among those exposed to 224-Ra, 226/228-Ra, or external radiation (Table 48–4). Soft tissue sarcomas are frequently induced by radiotherapy, especially in children with heritable retinoblastoma, so it is not surprising that fibrosarcoma is induced, because this soft tissue is widely distributed in bone. It is noteworthy that hereditary retinoblastoma has a shorter latent period than usual for radiotherapy-induced osteosarcoma and soft tissue sarcomas of the face (Chauveinc et al., 2001). Radiation-induced bone cancer appears, it seems, only at very high doses (IARC, 2000; UNSCEAR, 2000), and is rarely reported at doses under 5 Gy. Patients treated for benign conditions with low-dose radiotherapy have not shown increased bone cancers (Inskip et al., 1990; Darby et al., 1994; Fürst et al., 1988; Lindberg et al., 1995), nor have studies of prenatal X-ray (Bithell and Stewart, 1975) or occupational exposures (Cardis et al., 1995; Darby et al., 1998; Council of the Royal Society, 2001).

Chemicals Treatment of childhood cancer with alkylating agents has been linked to a 4.7-fold risk of bone cancer, with risk increasing as cumulative drug exposures rose (Tucker et al., 1987; Kingston et al., 1987; Newton et al., 1991; Hawkins et al., 1996; Le Vu et al., 1998; Neglia et al., 2001). Half of the 64 children who developed bone cancer as a second malignancy received chemotherapeutic agents, most commonly cyclophosphamide, triethylene-melamine, and chlorambucil. Radiotherapy and hereditary retinoblastoma were ruled out as potential confounders. There is no clear evidence that systemic exposure to other chemical agents are related to bone cancers. Concerns about the role of fluoridated drinking water and osteosarcoma based on animal studies were dispelled by epidemiological studies (Hoover et al., 1991).

Viruses Osteosarcomas and cancers of connective tissue have been induced by certain viruses in experimental animals since Rous’s work in 1912. There is no epidemiological evidence of horizontal transmission (clustering) of bone tumors. The British have made a tremendous effort to no avail to find statistical support for such transmission of solid childhood tumors other than lymphoma (e.g., McNally et al., 2003; Parslow et al., 2002; Gilman and Knox, 1995). Silcocks and Murrells (1987) focused on osteosarcoma and found no geographic clustering. Occasionally clusters of bone cancer have been reported, but virtually all forms of neoplasia, no matter how rare, may cluster geographically by chance, given that in the United States alone, for example, there are

29,000 towns (Neutra et al., 1990) and various other ways to group people, as in neighborhoods, schools, churches, clubs, and date of birth. As noted below, an adenovirus has been suggested in the formation of the EWS/FLI1 translocation found in Ewing sarcoma.

Implants and Other Foreign Bodies Occasional case reports have led to the suspicion that implants (metal, ceramic, polymer) can cause bone or soft tissue sarcomas. The International Agency for Research on Cancer (IARC, 1999) has issued an exhaustive review of the literature (from humans, domestic animals, experimental animals, and laboratory tests for carcinogenicity) and concluded that the evidence for inducing cancer fell into IARC category 2B, which is borderline at best regarding exposures to polymeric implants prepared as thin smooth films, metallic implants prepared as thin smooth films, implanted foreign bodies of metallic cobalt, metallic nickel, and an alloy powder containing 66%–67% nickel, 13%–16% chromium and 7% iron. In IARC category 3, not classifiable as to their carcinogenicity to humans, are organic polymeric materials as a group, orthopedic implants of complex composition, implanted foreign bodies of metallic chromium or titanium and of cobalt-based, chromium-based, and titanium-based alloys, stainless steel and depleted uranium, and dental materials. A number of large-scale epidemiologic studies of cancer risk among patients with implants have been conducted since the publication of the 1999 IARC monograph. These studies provide further evidence of the lack of association between bone cancer and metallic implants in humans (Olsen et al., 1999; Signorello et al., 2001; Fryzek et al., 2002). Millions of people throughout the world have received implants, and by 1999, when the IARC report was published, a total of 35 cases had been reported of malignant neoplasms arising from the bone or the soft tissue in the region of an implant. Prior to the IARC study there had been 10 reports of non-metallic foreign bodies followed by various bone tumors. At the site of osseous metallic foreign bodies (mainly bullets and shrapnel fragments), there were 23 sarcomas and 23 carcinomas. No conclusions can be drawn from these case reports. The same is true of osteosarcomas reported at the sites of previous fractures. An excess of osteosarcomas at the ankle in Werner premature aging syndrome, however, may be due to the trauma of walking, at a site predisposed by the severe connective tissue disorder in the syndrome (see below) (Ishikawa et al., 2000).

GENETIC SUSCEPTIBILITY Osteosarcoma after Retinoblastoma In 1970 an excess of osteosarcoma after retinoblastoma was first reported (Jensen and Miller, 1971). In 1986 the gene for retinoblastoma (OMIM 180200) was isolated (Friend et al., 1986), and it was thought that mutation of the RB1 gene was involved in the genesis of both cancers. Most of the osteosarcomas have been in the field of radiotherapy, but 29% occur near the knee, well outside the field of radiation (Wong et al., 1997). Radiogenic osteosarcomas were reported to develop only when external beam radiation was given before 12 months of age (Abramson and Frank, 1998).

Bone Cancer Study of childhood osteosarcomas showed that 13 of 47 tumors had point mutations and 5 of 37 had gross rearrangements of TP53 (Miller et al., 1996). RB1 was either rearranged or deleted in 7 of 37 osteosarcomas, five of which also had TP53 mutations. Osteosarcomas in dogs have similar histology to those in humans, but study of the tumor in 21 dogs revealed normal RB1 in all. Thirty-eight percent had TP53 mutations (Mendoza et al., 1998; Loukopoulos et al., 2003). It should be noted that dogs do not develop retinoblastoma (Priester and McKay, 1980). RB1 was the first recognized tumor suppressor gene, and led the way to understanding a class of genes that are involved in the origins of a wide variety of common cancers (Knudson, 2000, 2002). Patients with this neoplasm have a germline mutation in about 40% of cases and a somatic cell mutation in the rest. RB1 plays a role in the development of other cancers, including the breast, lung, soft tissue, and genitourinary tract (Eng et al., 1993). In addition, various cancers occur excessively in patients with hereditary retinoblastoma. Followup study of retinoblastoma patients revealed that the cumulative incidence of a second cancer at age 50 was 51% when retinoblastoma was hereditary, and 5% when it was non-hereditary (Wong et al., 1997). A substantial number were in the radiation field, indicating a gene–environment interaction.

Li-Fraumeni Syndrome, TP53 A spectacular example of familial aggregation of cancer is found in Li-Fraumeni syndrome (OMIM 191170) (Li and Fraumeni, 1969; Friend et al., 1986; Li et al., 1988). In the syndrome, a variety of cancers occur in excess singly or in combination in children or young adults, notably osteosarcoma, soft tissue sarcoma, breast carcinoma, brain cancer, adrenocortical carcinoma, and leukemia. Study of patients with multiple primary cancers in Li-Fraumeni syndrome has shown that 8 of 27 new cancers occurred in the radiation field for treatment of a previous cancer (Hisada et al., 1998)—again, a possible gene–environment interaction. Molecular studies showed germline mutation of the tumorsuppressor gene, TP53, located at chromosome 17p13 (Malkin et al., 1990). There are now nearly 250 independent germline TP53 mutations in Li-Fraumeni syndrome described in over 100 publications according to Varley (2003), whose report is in a supplement of Human Mutation devoted to TP53. Varley described the spectrum of mutations, the methods for their detection, and the associated tumors. The mutated gene can be sought to diagnose the syndrome if the aggregation of cancer is limited by small family size, or if an individual develops multiple primary cancers of types found in LiFraumeni syndrome. The TP53 gene, like the RB1 gene, regulates normal growth of specific organs; when the tumor-suppressor function is inactivated, as by chromosomal deletion or degradation, growth becomes abnormal and cancerous. This class of genes, first recognized in rare cancers of childhood, now has counterparts in adult cancers that are far more common, and may lead to new strategies for cancer prevention, early detection, and treatment.

DNA Helicase Mutations Three DNA helicase mutations unwind duplex DNA, which results in three separate syndromes that increase genomic instability and thus predispose to cancer and premature aging. These mutations occur in what is now known as the RecQ family of helicases (Nakayama, 2002). In an overview, Furuichi (2001) described the expression of the genes in the tissues, highly correlated with the phenotypes, and tissuespecific genomic instability, as, for example, in the pancreas (diabetes mellitus), ovary, and testis (hypogonadism) in Werner syndrome. A comparison of the molecular and clinical findings in the three syndromes has been published (Lindor et al., 2000).

953

cancers are soft tissue sarcomas, acral lentiginous melanomas, myeloid disorders, thyroid carcinoma, and benign meningiomas. Due to genotype differences, non-Japanese have shown no excess of melanoma or thyroid cancer (Goto et al., 1996). In Japanese, the ratio of carcinomas to sarcomas was 1 : 1 instead of the usual 10 : 1 (Miller and Myers, 1981). The syndrome, an autosomal recessive trait, is characterized by premature aging, graying of the hair, diabetes mellitus, atrophy of the skin, connective tissue disorders, arteriosclerosis, unusual cancers, and death on average at 46 years. The syndrome is far more common in Japan than elsewhere because of the mutant gene in cousin marriages. The osteosarcomas are atypical in that they cluster at the ankle at 35–57 years of age (Ishikawa et al., 2000). In this age range, patients with the syndrome have poor circulation and severe wasting of the soft tissue of the lower legs that concentrates weight bearing at the ankle. The acral lentiginous melanomas have the same distribution on the feet as were reported in a series of Caucasian patients with no underlying disease. The authors noted that the distribution matched that of weight-bearing sites (Feibleman et al., 1980). In Werner syndrome a DNA helicase gene mutation has been found at chromosome 8p12-11.2. Through the use of a monoclonal antibody directed against the DNA helicase gene-product a molecular defect could be found without the need for more complex mutational analysis (Furuichi, 2001; Shimizu et al., 2002). Moser et al. (1999) have reviewed the possible mechanistic links between loss of WRN protein function and the pathogenesis of clinical and cellular abnormalities.

Rothmund-Thomson Syndrome (OMIM 268400) This syndrome, an autosomal recessive trait, is also attributed to a DNA helicase mutation (RecQL4) (Kitao et al., 1999). The disease is characterized by a sun-sensitive rash usually at 3–6 months of age, short stature, poikiloderma (marbled dermal atrophy) among other skin disorders, and 75% had skeletal dysplasias on X-ray examination. In a series of 41 cases, 13 (32%) developed osteosarcoma at 3–41 years (median 11.6 years). Twenty-two patients were still under age 15 years so more osteosarcomas can occur among the 41 in this series (Wang et al., 2001). None had soft tissue sarcomas, in contrast to Werner syndrome. Osteosarcoma was associated with a distinctive pattern of mutations in the RecQL4 gene (Wang et al., 2003). In a series of 11 cases, all had at least 1 of 19 truncating mutations. Ten patients with the syndrome who did not have truncating mutation did not have osteosarcoma. Thus, molecular diagnosis may identify children with the syndrome who are at high risk of osteosarcoma. At chromosome 8q24.3 a gene was found that belonged in the RecQ family of helicases, but its effect on health was not known (Kitao et al., 1999). Because Rothmund-Thomson syndrome, first described in 1868, had features of premature aging, Werner in 1904 had to differentiate it from the premature aging syndrome he was describing (Martin, 2001). The overlap in clinical findings led to the testing of two Mayo Clinic patients (brothers) with Rothmund-Thomson syndrome and osteosarcoma (Lindor et al., 1996, 2000; Kitao et al., 1999). The brothers had the newly discovered mutated gene. The syndrome has since been shown to have substantial genetic heterogeneity (Wang et al., 2003). Certain genotypes have been linked by these authors to high risk of osteosarcoma, thus providing the physician with the opportunity for early detection and treatment. RecQ5 is another member of the helicase family found along with RecQL4, now known as the RTS gene, but it has not yet been linked to any disease. RTS is an excellent tool for diagnosing Rothmund-Thomson syndrome and studying its molecular basis. The diagnosis can be made by an immunoblot technique. Further details about the clinical and laboratory features and pathophysiology are well described by Kitao et al. (2003).

Bloom Syndrome (OMIM 210900) Werner Syndrome (OMIM 27770) Osteosarcoma is one of six neoplasms that occur excessively in Werner syndrome (Goto et al., 1996; Goto and Miller, 2001). The other

Bloom syndrome, due to a third helicase mutation, located on chromosome 15q26.1, is characterized by dwarfism, a sun-sensitive rash, immunodeficiency, male infertility, and an excess of cancers—mainly

954

PART IV: CANCER BY TISSUE OF ORIGIN

acute leukemia and non-Hodgkin lymphoma in childhood, and carcinomas in adulthood. Males are sterile. The gene is expressed in the testes but not in the ovaries (Furuichi, 2001). In the first 100 cancers in the Bloom Syndrome Registry, there were only two patients with osteosarcomas and none with soft tissue sarcomas (German, 1997), markedly different from the syndromes of Werner and RothmundThomson.

Osteosarcoma in Other Syndromes Case reports also describe osteosarcoma and fibrosarcoma with polyostotic fibrous dysplasia (Yabut et al., 1988), and osteogenesis imperfecta (Lasson et al., 1978). Osteosarcoma has also been described with Hutchinson-Gilford progeria (King et al., 1978) and with incompletely defined syndromes involving growth disturbances (Parry et al., 1978). Cancers of various types have been observed on the walls of bone cysts, and have arisen from benign giant cell tumors, osteomas, osteoblastomas, bone infarcts, fibrous dysplasia, and chronic osteomyelitic sinuses (Unni and Dahlin, 1979). Long before these observations were made, Johnson (1953) suggested that bone disorders with prolonged periods of excessive cell activity are prone to neoplastic change.

Genetics of Ewing Sarcoma Ewing sarcoma (OMIM 133450) rarely occurs in siblings (Joyce et al., 1984; Zamora et al., 1986). As noted previously, it is rare in blacks and Asians, suggesting a genetic influence. Ewing sarcoma occurs excessively only after retinoblastoma, substantially outnumbered by osteosarcoma in this regard. A case series, assembled through a search of PubMed, revealed 10 cases of Ewing tumor following retinoblastoma, when only one or two cases were expected based on the percentage distribution of Ewing sarcoma among cancers in the general population (Cope et al., 2001). A somatic cell mutation, t (11,22), found in almost all cases, forms an EWS-FLI1 fusion product, in which the Ewing sarcoma gene, EWS, on chromosome 22 is a powerful transcriptional activator; FLI1 (friend leukemia insertion [OMIM 193067]) on chromosome 11, is a gene in the ets family. These translocations are limited to somatic cells. Sanchez-Prieto et al. (1999) noted the close resemblance between Ewing sarcoma cells and tumor cells expressing the adenovirus E1A gene. They postulated that this gene induces the translocation and the oncogenic fusion protein. They tested the hypothesis and demonstrated the fusion product after E1A expression. An accompanying commentary (Kirn and Hermiston, 1999) hailed the finding for its novelty in suggesting a link between a virus and this human cancer. They said it may open a new area for study of carcinogenesis, which if it transpires, may provide a basis for developing antiviral vaccines. After failures by two other investigators to confirm the finding, the original authors stated that they found E1A sequences in all three Ewing tumor lines they had studied and in 14 of 32 Ewing tumors from patients (deAlava et al., 2000).

Genetics of Chordoma Chordoma (OMIM 215400), when familial, is transmitted as an autosomal dominant trait. A genome-wide linkage study of 11 members of a family with chordoma plus five affected members of two unrelated families mapped the gene locus to chromosome 7q33 (Kelley et al., 2001). Another study of 16 chordomas, using genetic hybridization and cytogenetics, suggested that an oncogene at 7q36 might be involved (Scheil et al., 2001).

Genetics of Chondrosarcoma Although chondrosarcoma is the second most common bone cancer, it has not yielded much information about its genetic origins. It is not familial. It is less common in blacks than whites, and it does occur in syndromes due to malignant degeneration of benign cartilaginous

tumors. Multiple exostoses (diaphyseal aclasia) are osteochondromas on the surfaces of growing bone. This dominantly inherited condition may produce severe deformities, and transformation to chondrosarcoma has been reported in 5%–11% of patients. However, no malignant change was seen in in the follow-up of 43 patients, 20 of whom had a family history of the disorder (Pierz et al., 2002). Chondrosarcoma also occurs excessively with enchondromatosis (Ollier syndrome) or with the combination of enchondromatosis and skin hemangiomas (Maffucci syndrome), but neither syndrome is inherited in a simple Mendelian fashion (Schwartz et al., 1987). Cytogenetic studies have to date shown that chondrosarcomas, except when they are extraskeletal, are not associated with specific translocations (Avery, 2002). See also Avery and Bridge (2003) for an extensive review of the literature and their own work on cartilaginous tumors.

Paget Disease Paget disease (osteitis deformans) (OMIM 167250) predisposes mainly to osteosarcoma, but also to fibrosarcoma, chondrosarcoma, and giant cell tumor (Haibach et al., 1985). Localized bone destruction occurs for an unknown reason, and it makes the bone susceptible to the effects of stress. Repair occurs almost simultaneously with resorption, and the bone enlarges (Fallon and Schwamm, 1989). The skull is commonly involved, and one sign of the disease is the need for a larger hat size than before. Of 101 osteosarcomas in a series of patients older than 60 years of age, 55% were associated with Paget disease (Huvos, 1986). In the United Kingdom, a survey was made of abdominal radiographs of patients 55+ years old, 1970–1977 (Cooper et al., 1999). About 1000 radiographs were examined from each of 31 towns for evidence of Paget disease. Each film showed both femoral heads and all the lumbar vertebrae. There was a steep increase in frequency with advancing age, from 2% in men at 50–59 years to 20% in those at 85+ years. The rates in women at these ages rose from 1% to about 7%. The prevalence in British towns ranged from 2.7% to 5.6%, except for the highest rates, 6.3% to 8.3%, in a cluster around Lancashire (Barker et al., 1980). A repeat survey of 10 towns, 1993–1995, indicated that the prevalence had diminished in the population from 5% to 2%, and the Lancashire cluster was less prominent (Cooper et al., 1999). Through the use of a large record-linkage resource (the General Practice Research Database) in England and Wales, 2465 patients with a diagnosis of Paget disease of bone were ascertained, 1988–1999 (van Staa et al., 2002). The incidence of clinically diagnosed Paget disease declined from 1.1 to 0.7 per 100,000 person-years. A decrease was also noted in the Mayo Clinic’s population-based study of residents in Olmstead County MN (Tiegs et al., 2000), and in a Spanish hospitalbased study (Morales-Piga et al., 2002), where, as in the British experience, the disease appeared to be less severe. This decline and within-country variations in rates indicate an environmental influence, as do variations from country to country. Among British migrants to Australia, the rates were midway between those in Britain and nativeborn Australians (Gardner et al., 1978). In a case-control study, a history of Paget disease in parents or siblings was given by 12.3% of 788 cases compared with 2.1% of 387 spouse controls (Siris et al., 1991). Among relatives of cases, the cumulative risk was highest when the case was diagnosed under 55 years of age and had bone deformity, an indication of severity of the disease. Familial Paget disease is consistent with autosomal dominant transmission (McKusick, 1994). A recent genome-wide search for the gene involved 319 individuals in 62 kindreds predominantly of British descent (Hocking et al., 2001). Several susceptibility loci were identified, the strongest of which is on chromosome 5q35.

Multiple Neoplasms Osteosarcoma and retinoblastoma tend to aggregate individually among close relatives. Such observations, including those in Li-

Bone Cancer Fraumeni syndrome, have prompted the generalization that cancers occurring excessively as double primaries may also aggregate excessively in families (Hansen and Cavenee, 1987). These clinical observations provide clues for laboratory scientists in their search for genes of diverse cancers that are familial or occur as multiple primaries.

PREVENTION High exposure to alpha-emitting radionuclides is not a problem at present, nor is unrestrained use of radioisotopes after the 1950s (Advisory Committee on Human Radiation Experiments, 1995). Radiotherapy is supposed to be used only when the benefits outweigh the risk of carcinogenesis. The possibility that a vaccine may be developed against Ewing sarcoma has been raised by the recent report that a specific adenovirus may induce the translocation and oncogenic fusion protein characteristic of this neoplasm. Wang et al. (2003) have reported that osteosarcoma in Rothmund-Thomson syndrome may be anticipated by testing for a particular genotype. Early detection and treatment of osteosarcoma are facilitated by medical surveillance of individuals with predisposing genetic disorders. Paget disease, which predisposes to osteosarcoma in the elderly, has been decreasing in frequency especially in Great Britain. The British have first-class record systems for following these trends, a great advantage in seeking the explanation, which when found, may lead to improved prevention and earlier detection.

FUTURE DIRECTIONS Progress in bone cancer etiology in the past 5 years has been immense. It comes from subtyping by pathology, linkage of subtypes to genetic syndromes (clinical and molecular epidemiologic studies), and radiation accidents (fallout from Chernobyl and the contamination of air, water and soil with plutonium at the Mayak Complex in Chelyabinsk). Most of these developments could not have been foreseen. What lies in the near future now?

Demography Data should be evaluated by subtype insofar as possible. Mortality and survival rates have been unchanged since the 1980s. Therapy may improve based on advances in molecular genetics and immunology.

Genetics As described previously, there are many genotypes in RothmundThomson syndrome, only one of which has been linked to osteosarcoma in the syndrome. The presence of this genotype, detected by laboratory study, makes possible early detection and treatment. This observation illustrates an approach to identifying those at highest risk of this and other cancers. In the past 5 years DNA helicase gene mutations have been found in two conditions with an excess of osteosarcoma, Werner syndrome and Rothmund-Thomson syndrome, and one in which there is no such excess, Bloom syndrome. There is a fourth helicase gene mutation to which no disease has yet been linked, RecQL-5. It may be a syndrome in which premature aging is a feature (as it was in linking RTS to RecQ4 [Kitao et al., 1999]). Mulvihill-Smith syndrome is a possibility (Mulvihill and Smith, 1975), but no specimens for study have been available. Bone cancers have also been central in the new understanding of two other major findings in cancer genetics. Clinical and epidemiological findings concerning retinoblastoma (and osteosarcoma) provided data that led Knudson to develop the concept of tumor suppressor genes. The Li-Fraumeni syndrome of multiple familial cancers including osteosarcoma implicated TP53 in their genesis. The individual genes, TP53 and RB1, play major roles in protecting the genome from damage that can lead to cancer and aging. Molecular

955

epidemiology of osteosarcoma will continue to be a fertile field for new laboratory and clinical developments.

Environmental Exposures Longer-term study of accidental exposures in the manufacture of alpha-particle emitters in the Russian Federation should add to information already published. Apart from radiation, the changing rates of Paget disease, with its elevated risk of osteosarcoma, may be tracked to an environmental cause. References Abramson DH, Frank CM. 1998. Second non-ocular tumors in survivors of bilateral retinoblastoma: A possible age effect on radiation-related risk. Ophthalmology 105:573–579. Advisory Committee on Human Radiation Experiments, Final Report (stock number 061-000-00-848-9). U.S. Government Printing Office P.O. Box 371954 Pittsburgh, PA 15250-7954. Andersson M, Carstensen B, Storm HH. 1995. Mortality and cancer incidence after cerebral arteriography with or without Thorotrast. Radiat Res 142:305–320. Anspaugh LR, Degteva MO, Vasilenko EK. 2002. Mayak Production Association: Introduction. Radiat Environ Biophys 41:19–22. Barker DJ, Chamberlain AT, Guyer PB, et al. 1980. Paget’s disease of bone: The Lancashire focus. Br Med J 280:1105–1107. Baverstock KF, Papworth DG. 1989. The UK radium luminiser survey. In: Taylor DM, Mays CW, Gerber GB, Thomas RG, eds. Risks from Radium and Thorotrast. BIR Report 21. London: Br Inst Radiol, pp. 72–76. Bielack SS, Rerin JS, Dickerhoff R, et al. 2003. Osteosarcoma after allogeneic bone marrow transplantation. A report of four cases from the Cooperative Osteosarcoma Study Group (COSS). Bone Marrow Transplant 315:353–359. Bjornsson J, McLeod RA, Unni KK, Ilstrup DM, Pritchard DJ. 1998. Primary chondrosarcoma of long bones and limb girdles. Cancer 83:2105–2119. Boice JD Jr., Day NE, Andersen A, et al. 1985a. Second cancers following radiation treatment for cervical cancer. An international collaboration among cancer registries. JNCI 74:955–975. Boice JD Jr., Engholm G, Kleinerman RA, et al. 1988. Radiation dose and second cancer risk in patients treated for cancer of the cervix. Radiat Res 116:3–55. Boice JD Jr., Storm HH, Curtis RE, et al., eds. 1985b. Multiple Primary Cancers in Connecticut and Denmark. Natl Cancer Inst Monogr 68. Washington DC, US Govt Print Off, pp. 1–437. Campisi J. 2003. Cancer and ageing: rival demons? Nat Rev Cancer 3:339–349. Cardis E, Gilbert ES, Carpenter L, et al. 1995. Effects of low doses and low dose rates of external ionizing radiation: Cancer mortality among nuclear industry workers in three countries. Radiat Res 142:117–132. Carnes BA, Groer PG, Kotek TJ. 1997. Radium dial workers: Issues concerning dose response and modeling. Radiat Res 147:707–714. Chauveinc L, Mosseri V, Quintana E, et al. 2001. Osteosarcoma following retinoblastoma: Age at onset and latency period. Ophthal Genet 22:77–88. Cooper C, Dennison E, Schafheutle K, Kellingray S, Guyer P, Barker D. 1999. Epidemiology of Paget’s disease of bone. Bone 24(5 Suppl):3S–5S. Cope JU, Tsokos M, Miller RW. 2001. Ewing sarcoma and sinonasal neuroectodermal tumors as second malignant tumors after retinoblastoma and other neoplasms. Med Pediatr Oncol 36:290–294. Council of the Royal Society. 2001. The Health Hazards of Depleted Uranium Munitions. Part 1. Science Advice Section. London: The Royal Society. Curtis RE, Rowlings PA, Deeg HJ, et al. 1997. Solid cancers after bone marrow transplantation. N Engl J Med 336:897–904. Dagher R, Pham TA, Sorbara L, et al. 2001. Molecular confirmation of Ewing sarcoma. J Pediatr Hematol Oncol 23:221–224. Darby SC, Reeves G, Key T, Doll R, Stovall M. 1994. Mortality in a cohort of women given X-ray therapy for metropathia haemorrhagica. Int J Cancer 56:793–801. de Alava E, Sanchez-Prieto R, Ramon Y, Cajal S. 2000. Adenovirus ElA and Ewing tumors. Nat Med 6:4. Dehner LP. 1986. Peripheral and central primitive neuroectodermal tumors. A nosologic concept seeking a consensus. Arch Pathol Lab Med 110:997–1005. de Vathaire F, Francois P, Hill C, et al. 1989. Role of radiotherapy and chemotherapy in the risk of second malignant neoplasms after cancer in childhood. Br J Cancer 59:792–796. Dicus GJ. Joint American-Russian radiation health effects research. U.S. Nuclear Regulatory Commission. Presentation to the joint meeting of American Nuclear Society, Washington, DC Section and Health Physics

956

PART IV: CANCER BY TISSUE OF ORIGIN

Society, Baltimore-Washington Chapter, January 16, 1997. Available at: http://www.nrc.gov/reading-rm/doc-collections/commission/speeches/ 1997/s97-04.html Doody MM, Mandel JS, Lubin JH, Boice JD Jr. 1998. Mortality among United States radiologic technologists, 1926–90. Cancer Causes Control 9:67–75. Dores GM, Metayer C, Curtis RE, et al. 2002. Second malignant neoplasms among long-term survivors of Hodgkin’s disease: A population-based evaluation over 25 years. J Clin Oncol 20:3484–3494. Dorfman HD, Czerniak B. 1995. Bone cancers. Cancer 75:203–210. Dos Santos Silva I, Malveiro F, Jones ME, Swerdlow AJ. 2003. Mortality after radiological investigation with radioactive Thorotrast: A follow-up study of up to fifty years in Portugal. Radiat Res 159:521–534. Draper GJ, Sanders BM, Kingston JE. 1986. Second primary neoplasms in patients with retinoblastoma. Br J Cancer 53:661–671. Eng C, Li FP, Abramson DH, Ellsworth RM, et al. 1993. Mortality from second tumors among long-term survivors of retinoblastoma. J Natl Cancer Inst 85:1121–1128. Fallon MD, Schwamm HA. 1989. Paget’s disease of bone. An update on the pathogenesis, pathophysiology, and treatment of osteitis deformans. Pathol Annu 24 Pt 1:115–159. Feibleman CE, Stoll H, Maize JC. 1980. Melanomas of the palm, sole, and nailbed: A clinicopathologic study. Cancer 46:2492–2504. Friend SH, Bernards R, Rogelj S, et al. 1986. A human DNA segment with properties of the gene that predisposes to retinoblastoma and osteosarcoma. Nature 323:643–646. Fryzek JP, Ye W, Signorello LB, Lipworth L, Blot WJ, McLaughlin JK, Nyren O. 2002. Incidence of cancer among patients with knee implants in Sweden, 1980–1994. Cancer 94:3057–3062. Fürst CJ, Lundell M, Holm LE, Silfversward C. 1988. Cancer incidence after radiotherapy for skin hemangioma: A retrospective cohort study in Sweden. J Natl Cancer Inst 80:1387–1392. Furuichi Y. 2001. Premature aging and predisposition to cancers caused by mutations in RecQ family helicases. Ann N Y Acad Sci 928:121–131. Gardner MJ, Guyer PB, Barker DJP. 1978. Radiological prevalence of Paget’s disease of bone in British migrants to Australia. Br Med J 1:1655–1657. Garwicz S, Anderson H, Olsen JH, et al. 2000. Second malignant neoplasms after cancer in childhood and adolescence: A population-based casecontrol study in the 5 Nordic countries. The Nordic Society for Pediatric Hematology and Oncology. The Association of the Nordic Cancer Registries. Int J Cancer 88:672–678. German J. 1997. Bloom’s syndrome. The first hundred cancers. Cancer Genet Cytogenet 93:100–106. Gilbert ES. 1998. The risk of cancer from exposure to plutonium. Radiat Res 158:783–784. Gilman EA, Knox EG. 1995. Childhood cancers: Space-time distribution in Britain. J Epidemiol Community Health 49:158–163. Ginsberg JP, Woo SY, Johnson ME, Hicks MJ, Horowitz ME. 2001. Ewing’s Sarcoma Family of Tumors: Ewing’s sarcoma of bone and soft tissue and the peripheral primitive neuroectodermal tumors. In: Pizzo PA, Poplack DG, eds. Principles and Practice Oncology, 4th ed. Philadelphia: Lippincott Williams & Wilkins, pp. 1351–1364. Gossner W. 1999. Pathology of radium-induced bone tumors: New aspects of histopathology and histogenesis. Radiat Res 152(6 Suppl):S12–S15. Goto M, Miller RW, Ishikawa Y, Sugano H. 1996. Excess of rare cancers in Werner syndrome (adult progeria). Cancer Epidemiol Biomarkers Prev 5:239–246. Goto M, Miller RW, eds. 2001. From Premature Gray Hair to Helicase-Werner Syndrome: Implications for Aging and Cancer (Gann Monograph on Cancer Research, 49) 1–178. Haibach H, Farrell C, Dittrich FJ. 1985. Neoplasms arising in Paget’s disease of bone: A study of 82 cases. Am J Clin Pathol 83:594–600. Hansen MF, Cavenee WK. 1987. Retinoblastoma and osteosarcoma: The prototypic cancer family. Acta Paediatr Jpn 29:526–533. Harrist TJ, Schiller AL, Trelstad RL, Mankin HJ, Mays CW. 1979. Thorotrastassociated sarcoma of bone: A case report and review of the literature. Cancer 44:2049–2058. Hawkins MM, Draper GJ, Kingston JE. 1987. Incidence of second primary tumours among childhood cancer survivors. Br J Cancer 56:339–347. Hawkins MM, Wilson LM, Burton HS, et al. 1996. Radiotherapy, alkylating agents, and risk of bone cancer after childhood cancer. J Natl Cancer Inst 88:270–278. Hisada M, Garber JE, Fung CY, Fraumeni JF Jr., Li FP. 1998. Multiple primary cancers in families with Li-Fraumeni syndrome. J Natl Cancer Inst 90:606–611. Hocking LJ, Herbert CA, Nicholls RK, et al. 2001. Genomewide search in familial Paget disease of bone shows evidence of genetic heterogeneity

with candidate loci on chromosomes 2q36, 10p13, and 5q35. Am J Hum Genet 69:1055–1061. Hoover RN, Devesa SS, Cantor KP, et al. 1991. Fluoridation of drinking water and subsequent cancer incidence and mortality. In: Report of the ad hoc Subcommittee on Fluoride of the Committee to Coordinate Environmental Health and Related Programs: Review on Fluoride, Benefits and Risks. DHHS PHS February, Apps. E and F. Human Mutation 2003. Focus on p53 and cancer. Hum Mut 21:173–326. Huvos AG. 1986. Osteogenic sarcoma of bones and soft tissues in older persons. A clinicopathologic analysis of 117 patients older than 60 years. Cancer 57:1442–1449. IARC. 1999. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. Surgical Implants and Other Foreign Bodies. Lyon, France. v. 74. IARC. 2000. IARC monographs on the evaluation of carcinogenic risks to humans. Ionizing Radiation, Part 1: X- and Gamma (••)-Radiation, and Neutrons. Lyon, France. v. 75. IARC. 2001. IARC monographs on the evaluation of carcinogenic risks to humans. Ionizing Radiation, Part 2: Some Internally Deposited Radionuclides. Lyon, France. v. 78. Inskip PD, Monson RR, Wagoner JK, et al. 1990. Cancer mortality following radium treatment for uterine bleeding. Radiat Res 123:331–344. Erratum in: Radiat Res 1991 128:326. Ishikawa Y, Miller RW, Machinami R, Sugano H, Goto M. 2000. Atypical osteosarcomas in Werner Syndrome (adult progeria). Jpn J Cancer Res 91:1345–1349. Ishikawa Y, Tsukuma H, Miller RW. 1996. Low rates of Paget’s disease of bone and osteosarcoma in elderly Japanese. Lancet (9014)347:1559. Ishikawa Y, Wada I, Fukumoto M. 2001. Alpha-particle carcinogenesis in Thorotrast patients: Epidemiology, dosimetry, pathology, and molecular analysis. J Environ Pathol Toxicol Oncol 20:311–315. Jensen RD, Miller RW. 1971. Retinoblastoma: Epidemiologic characteristics. N Engl J Med 285:307–311. Johnson LC. 1953. A general theory of bone tumors. Bull NY Acad Med 164–171. Joyce MJ, Harmon DC, Mankin HJ, Suit HD, Schiller AL, Truman JT. 1984. Ewing’s sarcoma in female siblings. A clinical report and review of the literature. Cancer 53:1959–1962. Kathren RL, Hill RL. 1992. Distribution and dosimetry of Thorotrast in USUR case 1001. Health Phys 63:72–88. Kellerer AM. 2002. The Southern Urals radiation studies A reappraisal of the current status. Radiat Environ Biophys 41:307–316. Kelley MJ, Korczak JF, Sheridan E, Yang X, Goldstein AM, Parry DM. 2001. Familial chordoma, a tumor of notochordal remnants, is linked to chromosome 7q33. Am J Hum Genet 69:454–460. Kirn D, Hermiston T. 1999. Induction of an oncogenic fusion protein by a viral gene—a new chapter in an old story. Nat Med 5:991–992. King CR, Lemmer J, Campbell JR, et al. 1978. Osteosarcoma in a patient with Hutchinson-Gilford progeria. J Med Genet 15:481–484. Kitao S, Shimamoto A, Furuichi Y. 2003. Molecular biology of RothmundThomson syndrome. In: Hisama FM, Weissman SM, Martin GM, eds. Chromosomal Instability and Aging. New York: Marcel Dekker, pp. 223–244. Kitao S, Shimamoto A, Goto M, et al. 1999. Mutations in RECQL4 cause a subset of cases of Rothmund-Thomson syndrome. Nat Genet 22:82–84. Kleinerman RA, Boice JD Jr., Storm HH, et al. 1995. Second primary cancer after treatment for cervical cancer. An international cancer registries study. Cancer 76:442–452. Knudson AG. 2000. Chasing the cancer demon. Annu Rev Genet 34:1–19. Knudson AG. 2002. Cancer genetics. Am J Med Genet 111:96–102. Koshurnikova NA, Gilbert ES, Sokolnikov M, et al. 2000. Bone cancers in Mayak workers. Radiat Res 154:237–245. Kossenko MM, Degteva MO, Vyushkova OV, et al. 1997. Issues in the comparison of risk estimates for the population in the Techa River region and atomic bomb survivors. Radiat Res 148:54–63. Leenhouts HP, Brugmans MJ. 2000. An analysis of bone and head sinus cancers in radium dial painters using a two-mutation carcinogenesis model. J Radiol Prot 20:169–188. Le Vu B, De Vathaire F, Shamsaldin A, et al. 1998. Radiation dose, chemotherapy and risk of osteosarcoma after solid tumours during childhood. Int J Cancer 77:370–377. Li FP, Fraumeni JF Jr. 1969. Soft-tissue sarcomas, breast cancer, and other neoplasms. A familial syndrome? Ann Intern Med 71:747–752. Li FP, Fraumeni JF Jr., Mulvihill JJ, et al. 1988. A cancer family syndrome in twenty-four kindreds. Cancer Res 48:5358–5362. Lindberg S, Karlsson P, Arvidsson B, Holmberg E, Lunberg LM, Wallgren A. 1995. Cancer incidence after radiotherapy for skin haemangioma during infancy. Acta Oncol 34:735–740.

Bone Cancer Lindor NM, Devries EM, Michels VV, et al. 1996. Rothmund-Thomson syndrome in siblings: Evidence for acquired in vivo mosaicism. Clin Genet 49:124–129. Lindor NM, Furuichi Y, Kitao S, Shimamoto A, Arndt C, Jalal S. 2000. Rothmund-Thomson syndrome due to RECQ4 helicase mutations: Report and clinical and molecular comparisons with Bloom syndrome and Werner syndrome. Am J Med Genet 90:223–228. Loukopoulos P, Thornton JR, Robinson WF. 2003. Clinical and pathologic relevance of p53 index in canine osseous tumors. Vet Pathol 40:237–248. Malkin D, Li FP, Strong LC, et al. 1990. Germ line p53 mutations in a family syndrome of breast cancer, sarcomas and other neoplasms. Science 250:1233–1238. Martin GM. 2001. A brief history of research on the Werner syndrome. In: Goto M, Miler RW, eds. From Premature Gray Hair to Helicase-Werner Syndrome: Implications for Aging and Cancer (Gann Monograph on Cancer Research, 49) pp. 1–10. Martland HS, Conlon P, Knef JP. 1925. Some unrecognized dangers in the use and handling of radioactive substances. JAMA 85:1769–1776. McNairn JD, Damron TA, Landas SK, Ambrose JL, Shrimpton AE. 2001. Inheritance of osteosarcoma and Paget’s disease of bone: A familial loss of heterozygosity study. J Mol Diagn 3:171–177. McNally RJ, Kelsey AM, Eden OB, Alexander FE, Cairns DP, Birch JM. 2003. Space-time clustering patterns in childhood solid tumours other than central nervous system tumours. Int J Cancer 103:253–258. McKusick VA. 1972. Paget’s disease of the bone. In: McKusick VA, ed. Heritable Disorders of Connective Tissue. 4th ed. St. Louis: CV Mosby Co, pp. 718–723. McKusick VA. 1994. Mendelian Inheritance in Man. 11th ed. Baltimore: Johns Hopkins University Press. McMaster ML, Goldstein AM, Bromley CM, Ishibe N, Parry DM. 2001. Chordoma: Incidence and survival patterns in the United States, 1973–1995. Cancer Causes Control 12:1–11. Mendoza S, Konishi T, Dernell WS, Withrow SJ, Miller CW. 1998. Status of the p53, RB-1 and MDM2 genes in canine osteosarcoma. Anticancer Res 18(6A):4449–4453. Metayer C, Lynch CF, Clarke EA, et al. 2000. Second cancers among longterm survivors of Hodgkin’s disease diagnosed in childhood and adolescence. J Clin Oncol 18:2435–2443. Miller CW, Aslo A, Won A, Tan M, Lampkin B, Koeffler HP. 1996. Alterations of the p53, Rb and MDM2 genes in osteosarcoma. J Cancer Res Clin Oncol 122:559–565. Miller RW. 1981. Contrasting epidemiology of childhood osteosarcoma, Ewing’s tumor, and rhabdomyosarcoma. Natl Cancer Inst Monogr 56:9–14. Miller RW. 1985. Detection of environmental effects through anatomic pathology. Monogr Pathol 26:11–18. Miller RW, Myers MH. 1983. Age distribution of epithelial and non-epithelial cancers. Lancet 2(8361):1250. Morales-Piga AA, Bachiller-Corral FJ, Abraira V, Beltran J, Rapado A. 2002. Is clinical expressiveness of Paget’s disease of bone decreasing? Bone 30:399–403. Mori T, Kido C, Fukutomi K, Kato Y, 1999. Summary of entire Japanese thorotrast follow-up study: Updated 1998. Radiat Res 152(6 Suppl):S84–87. Moser MJ, Oshima J, Monnat RJ Jr. 1999. WRN mutations in Werner syndrome. Hum Mutat 13:271–279. Nakayama H. 2002. RecQ family helicases: Roles as tumor suppressor proteins. Oncogene 21:9008–9021. NAS (National Academy of Sciences). 1972. Advisory Committee on the Biological Effects of Ionizing Radiations (The BEIR Report): The Effects on Populations of Exposure to Low Levels of Ionizing Radiation. Washington DC: US Government Printing Office. Neglia JP, Friedman DL, Yasui Y, et al. 2001. Second malignant neoplasms in five-year survivors of childhood cancer: Childhood cancer survivor study. J Natl Cancer Inst 93:618–629. Nekolla EA, Kreisheimer M, Kellerer AM, Kuse-Isingschulte M, Gossner W, Spiess H. 2000. Induction of malignant bone tumors in radium-224 patients: Risk estimates based on the improved dosimetry. Radiat Res 153:93–103. Neutra RR, Swan S, Freedman D, et al. 1990. Clusters galore. Arch Environ Health 45:314. Newton WA, Meadows AT, Shimada H, et al. 1991. Bone sarcomas as second malignant neoplasms following childhood cancer. Cancer 67:193–201. Nyberg U, Nilsson B, Travis LB, Holm LE, Hall P. 2002. Cancer incidence among Swedish patients exposed to radioactive thorotrast: A forty-year follow-up survey. Radiat Res 157:419–425. Olsen JH, et al. 1999. Hip and knee implantations among patients with osteoarthritis and risk of cancer: A record-linkage study from Denmark. Int J Cancer 81:719–722.

957

Omar RZ, Barber JA, Smith PG. 1999. Cancer mortality and morbidity among plutonium workers at the Sellafield plant of British Nuclear Fuels. Br J Cancer 79:1288–1301. Parkin DM, Kramarova E, Draper GJ, et al. 1999. International Incidence of Childhood Cancer IARC Sci Publ 144:1–391. Parry DM, Safyer AW, Mulvihill JJ. 1978. Waardenburg-like features with cataracts, small head size, joint abnormalities, hypogonadism, and osteosarcoma. J Med Genet 15:66–69. Parslow RC, Law GR, Feltbower R, Kinsey SE, McKinney PA. 2002. Population mixing, childhood leukaemia, CNS tumours and other childhood cancers in Yorkshire. Eur J Cancer 38:2033–2040. Percy C, Young JL Jr., Muir C, et al. 1995. Histology of cancer: SEER population-based data. Cancer (suppl) 75:139–146. Pierz KA, Stieber JR, Kusumi K, Dormans JP. 2002. Hereditary multiple exostoses: One center’s experience and review of etiology. Clin Orthop 401:49–59. Polednak AP. 1978. Bone cancer among female radium workers. Latency periods and incidence rates by time after exposure. J Natl Cancer Inst 60:77–82. Price CHG. 1958. Primary bone-forming tumours and their relationship to skeletal growth. J Bone Joint Surg 40B:574–593. Priest ND. 2001. Toxicity of depleted uranium. Lancet 357:244–246. Priester WA, Mantel N. 1971. Occurrence of tumors in domestic animals. Data from 12 United States and Canadian colleges of veterinary medicine. J Natl Cancer Inst 47:1333–1344. Priester WA, McKay FW. 1980. The occurrence of tumors in domestic animals. Natl Cancer Inst Monogr 54:1–210. Robinson E, Neugut AI, Wylie P. 1998. Clinical aspects of post-irradiation sarcomas. J Natl Cancer Inst 80:233–240. Rowland RE. 1995. Dose-response relationships for female radium dial workers: A new look. In: van Kaick G, Karaoglou A, Kellerer AM, eds. Health Effects of Internally Deposited Radionuclides: Emphasis on Radium and Thorium. Singapore: World Scientific, pp. 135–143. Rowland RE, Stehney AF, Lucas HF Jr. 1978. Dose-response relationships for female radium dial workers. Radiat Res 76:368–383. Sanchez-Prieto R, de Alava E, Palomino T, et al. 1999. An association between viral genes and human oncogenic alterations: The adenovirus E1A induces the Ewing tumor fusion transcript EWS-FLI1. Nat Med 5: 1076–1079. Sandberg AA. 2002. Cytogenetics and molecular genetics of bone and softtissue tumors. Am J Med Genet 115:189–193. Sandberg AA, Bridge JA. 2003. Updates on the cytogenetics and molecular genetics of bone and soft tissue tumors: Chondrosarcoma and other cartilaginous neoplasms. Cancer Genet Cytogenet 143:1–31. Scheil S, Bruderlein S, Liehr T, et al. 2001. Genome-wide analysis of sixteen chordomas by comparative genomic hybridization and cytogenetics of the first human chordoma cell line, U-CH1. Genes Chromosomes Cancer 32:203–211. Schwartz HS, Zimmerman NB, Simon MA, et al. 1987. The malignant potential of enchondromatosis. J Bone Joint Surg 69A:269–274. Shilnikova NS, Preston DL, Ron E, et al. 2003. Cancer mortality risk among workers at the Mayak Nuclear Complex. Radiat Res 2003 159:787–798. Shimizu T, Tateishi Y, Furuichi Y, et al. 2002. Diagnosis of Werner syndrome by immunoblot analysis. Clin Exp Dermatol 271:157–159. Shimizu Y, Kato H, Schull WJ. 1990. Studies of the mortality of A-bomb survivors. 9. Mortality, 1950–1985: Part 2. Cancer mortality based on the recently revised doses (DS86). Radiat Res 121:120–141. Signorello LB, Ye W, et al. 2001. Nationwide study of cancer risk among hip replacement patients in Sweden. J Natl Cancer Inst 93:1405–1410. Silcocks PBS, Murrells T. 1987. Space-time clustering and bone turnours: Application of Knox’s method to data from a population-based cancer registry. Int J Cancer 40:769–771. Sindelar WF, Costa J, Ketcham AS. 1978. Osteosarcoma associated with Thorotrast administration: Report of two cases and literature review. Cancer 42:2604–2609. Siris ES, Ottman R, Flaster E, Kelsey JL. 1991. Familial aggregation of Paget’s disease. J Bone Miner Res 6:495–500. Spiess H, Mays CW, Chmelevsky D. 1989. Malignancies in patients injected with radium 224. In: Taylor DM, Mays CW, Gerber GB, eds. Risks from Radium and Thorotrast. BIR Report 21. London: Br Inst Radio, pp. 7–12. Spiess H. 2002. Peteosthor—a medical disaster due to Radium-224a personal recollection. Radiat Environ Biophys 41:163–172. Sterbbings JH. 2001. Health risks from radium in workplaces: An unfinished story. Occup Med 16:259–270. Stiller CA, Craft AW, Corazziari I; Eurocare Working Group. 2001. Survival of children with bone sarcoma in Europe since 1978: Results from the EUROCARE study. Eur J Cancer 2001 37:760–766.

958

PART IV: CANCER BY TISSUE OF ORIGIN

Thompson DE, Mabuchi K, Ron E, et al. 1994. Cancer incidence in atomic bomb survivors. Part II: Solid tumors, 1958–1987. Radiat Res 137(2 Suppl):S17–S67. Tiegs RD, Lohse CM, Wollan PC, Melton LJ. 2000. Long-term trends in the incidence of Paget’s disease of bone. Bone 27:423–427. Tucker MA, Meadows AT, Boice JD Jr., et al. 1984. Cancer risk following treatment of childhood cancer. In: Boice JD Jr., Fraumeni JF Jr., eds. Radiation Carcinogenesis: Epidemiology and Biological Significance. New York: Raven Press, pp. 211–224. Tucker MA, D’Angio GJ, Boice JD Jr., et al. 1987. Bone sarcoma linked to radiotherapy and chemotherapy in children. N Engl J Med 317:588–593. United Nations Scientific Committee on the Effects of Atomic Radiation. 2000. UNSCEAR 2000 Report to the General Assembly, with Scientific Annexes. Sources and Effects of Ionizing Radiation. E.00.IX.4. New York: United Nations. Unni KK, Dahlin DC. 1979. Premalignant tumors and conditions of bone. Am J Surg Pathol 3:47–60. Van Kaick G, Dalheimer A, Hornik S, et al. 1999. The German thorotrast study: Recent results and assessment of risks. Radiat Res 152(6 Suppl):S64–71. Van Staa TP, Selby P, Leufkens HG, Lyles K, Sprafka JM, Cooper C. 2002. Incidence and natural history of Paget’s disease of bone in England and Wales. J Bone Miner Res 17:465–471. Varley JM. 2003. Germline TP53 mutations and Li-Fraumeni syndrome. Hum Mutat 21:313–320. Voelz GL, Lawrence JN, Johnson ER. 1997. Fifty years of plutonium exposure to the Manhattan Project plutonium workers: An update. Health Phys 73:611–619.

Wang LL, Levy ML, Lewis RA, et al. 2001. Clinical manifestations in a cohort of 41 Rothmund-Thomson syndrome patients. Am J Med Genet 102:11–17. Wang LL, Gannavarapu A, Kozinetz CA, et al. 2003. Association between osteosarcoma and deleterious mutations in the RECQL4 gene in Rothmund-Thomson syndrome. J Natl Cancer Inst 95:669–674. Weaver JE III. 1994, 1995. A Brief Chronology of Radiation and Protection Health Physics. Available at: http://www.umich.edu/~radinfo/introduction/chrono.html Weiss HA, Darby SC, Doll R. 1994. Cancer mortality following X-ray treatment for ankylosing spondylitis. Int J Cancer 59:327–338. Wick RR, Gossner W. 1989. Recent results of the follow-up of radium-224treated ankylosing spondylitis patients. In: Taylor DM, Mays CW, Gerber GB, eds. Risks from Radium and Thorotrast. BIR Report 21. London: Br Inst Radiol, pp. 25–28. Wick RR, Nekolla EA, Gossner W, Kellerer AM. 1999. Late effects in ankylosing spondylitis patients treated with 224Ra. Radiat Res 152(6 Suppl):S8–S11. Winn DM, Li FP, Robison LL, et al. 1992. A case-control study of the etiology of Ewing’s tumor. Cancer Epidemiol Biomarkers Prev 1: 525–532. Wong FL, Boice JD Jr., Abramson DH, et al. 1997. Cancer incidence after retinoblastoma. Radiation dose and sarcoma risk. JAMA 278:1262–1267. Woodard HQ, Huvos AG, Smith J. 1988. Radiation-induced malignant tumors of bone in patients with Hodgkin’s disease. Health Phys 55:615–620. Zamora P, Garcia de Paredes ML, Gonzalez Baron M, et al. 1986. Ewing’s tumour in brothers: An unusual observation. Am J Clin Oncol 9:358–360.

49

Soft Tissue Sarcoma MARIANNE BERWICK

S

oft tissue sarcoma (STS) is a rare tumor, occurring in approximately 1 to 2 of every 100,000 individuals worldwide, and constitutes about 0.6% of all cancer cases and 0.7% of all cancer deaths (Howe et al., 2001). Prognosis is generally poor with a relative survival rate of approximately 67% at 5 years, with little difference by race. Approximately 8300 cases of STS will be diagnosed in the United States in 2003 and approximately 4000 deaths from STS will occur (Jemal et al., 2003). The etiology of STS is still poorly understood with multiple investigations into the role of environmental chemical exposures and a few investigations into host characteristics that predispose to STS. With the advent of genetic techniques, and particularly as the technology and computational methods for understanding the role of low-penetrance genes in a variety of pathogenic pathways progress, it is likely that advances will be made in understanding the causes of this cancer. The successful development of treatment for gastrointestinal stromal tumors (GIST) with a specific mutation is a paradigm for future molecular discoveries in the etiology and management of sarcoma and all cancers.

CLASSIFICATION Anatomic Distribution Soft tissue sarcoma is a disease of mesenchymal tissue other than bone and cartilage: connective, blood, lymph, adipose, nerves, and muscle tissue; it occurs throughout the body. Approximately half of soft tissue sarcomas occur in the limbs and the other half on the trunk and within the retroperitoneum. Sarcomas generally develop as deeply located masses and infrequently as superficial lesions (Enzinger and Weiss, 1995). The terms “connective tissue tumor” and “soft tissue tumor” are often used interchangeably. Many studies have investigated clinical predictors for survival with soft tissue sarcoma, often focusing on institutional series of specific histologic subtypes or anatomic sites (e.g., Tsujimoto et al., 1988; Lewis et al., 1998). Kattan et al. (2002) derived a “nomogram” for 12year sarcoma-specific death, based on clinical characteristics of 2327 adult patients treated at Memorial Sloan-Kettering Cancer Center. Variables associated with poor prognosis included older age at diagnosis, larger tumor size, high tumor grade, the histologic subtypes MPNT and Synovial, tumor depth, and the anatomic sites head, neck, and retrointra-abdominal. Similar predictive models based on large series can be used to assist the clinician in counseling patients and in determining optimal treatment.

Histopathology Soft tissue sarcoma is a particularly complex tumor to study. It is relatively rare throughout the world, and the histopathologic classification and the cancer registry coding are inconsistent. In addition, histopathologists frequently disagree on histologic subtype (Harris et al., 1991), thus diluting the ability of epidemiologists to investigate the etiology. The Surveillance, Epidemiology and End Result (SEER) registry and International Association for Research on Cancer (IARC) classify STS both by anatomic and histologic site, due to the diverse nature of soft tissue tumors and their distribution throughout the body. This dual classification system generates confusion because certain tumors will be coded by tissue type whereas others will be coded by

anatomic site (Lynge et al., 1987). For example, “connective tissue neoplasm” excludes soft tissue neoplasms occurring at organ sites. Most leiomyosarcomas occur in the visceral organs and can be entirely missed by classifications that rely on site-oriented codes (Zahm et al., 1996). Therefore, a soft tissue sarcoma arising in the wall of the stomach would be classified under stomach rather than “malignant neoplasms of soft and other connective tissue” (Suruda et al., 1993). This convention probably leads to an underestimate of the true rates of STS (Zahm et al., 1996). Embryologically, the soft tissue is derived principally from the mesoderm, with some contribution from the neuroectoderm (Enzinger and Weiss, 1995). There are numerous histologic subtypes that vary in definition depending on the pathologist. Tables 49–1 and 49–2 present the major types of soft tissue sarcoma found in adults and children (NCI, 2002). Based on a large clinical series at Memorial Sloan-Kettering Cancer Center, the four most common histologic subtypes of STS in adults, after Kaposi sarcoma, are liposarcoma (19.8%), leiomyosarcoma (18.2%), and fibrosarcoma (10.3%), which now includes the pleomorphic form of fibrosarcoma MFH, (17.6%) and gastro-intestinal stromal tumor, or GIST (6%). Among adults the risk for soft tissue sarcoma increases with age and is generally somewhat higher among men. Rates of Kaposi sarcoma were included with soft tissue sarcomas prior to the beginning of the AIDS epidemic, although now its rates are usually considered separately (Table 49–3). Tumor registries have been keeping statistics separately since the 1980s, but the increased incidence of Kaposi sarcoma, particularly in Africa, may still lead to overestimates of the incidence trends for soft tissue sarcoma (Zahm and Fraumeni, 1997; Levi et al., 1999; Froehner and Wirth, 2001). Kaposi sarcoma (KS), the most common soft tissue sarcoma, had been rarely seen in Western countries prior to the acquired immunodeficiency syndrome (AIDS) epidemic. It usually occurred in men of Italian or Jewish descent and followed a relatively mild clinical course. The classical form of KS is not common in African Americans but accounts for a significant portion of cancer in some African countries, such as Uganda. Surprisingly, leiomyosarcomas have been recently found in association with AIDS in children in Uganda (Biggar and Frisch, 2000; Granovsky, 1998; McClain et al., 1996). The next most common sarcomas are those of the fibrous tissue, including fibrosarcoma that occurs at all anatomic sites, malignant fibrous histiocytoma (MFH) that usually occurs on the legs, and dermatofibrosarcoma that usually occurs on the trunk. Sarcomas developing in the fatty tissue are liposarcomas and are composed of adult or embryonal fat cells or lipoblasts (Pack and Pierson, 1954). Although these can develop anywhere in the body, they are most commonly found in the thigh, the retroperitoneum, and the inguinal region and are frequently larger than most other tumors. The peak incidence is between 40 and 60 years (Enzinger and Weiss, 1995). Those sarcomas developing in muscle tissue consist of rhabdomyosarcomas that develop in the striated muscles of the extremities and leiomyosarcomas that develop in the smooth muscles, usually in the uterus and the digestive tract. Rhabdomyosarcomas are the most common childhood sarcomas with a peak incidence under the age of 5 and a smaller peak at 15 to 19 years (Mahour et al., 1967); they are rare in individuals older than 45. In general, leiomyosarcomas are more common in women than in men and appear to coincide with

959

960

PART IV: CANCER BY TISSUE OF ORIGIN pregnancy as well as estrogenic stimulation (Enzinger and Weiss, 1995). Uterine leiomyosarcoma rates are higher for African American women than for whites (Harlow, 1986; Polednak, 1986). Leiomyosarcomas are rare in children, except for those appearing in association with AIDS and Epstein Barr virus (McClain et al., 1995).

Table 49–1. Major Types of Soft Tissue Sarcomas in Adults Tissue of Origin

Usual Location in the Body

Type of Cancer

Fibrous tissue

Fat Muscle Striated muscle Smooth muscle Blood vessels Lymph vessels Synovial tissue (linings of joint cavities, tendon sheaths) Peripheral nerves Cartilage and boneforming tissue

Fibrosarcoma Malignant fibrous histiocytoma Dermatofibrosarcoma Liposarcoma

Arms, legs, trunk Legs

Rhabdomyosarcoma Leiomyosarcoma Hemangiosarcoma Kaposi sarcoma Lymphangiosarcoma Synovial sarcoma

Arms, legs Uterus, digestive tract Arms, legs, trunk Legs, trunk Arms Legs

Neurofibrosarcoma Extraskeletal chondrosarcoma Extraskeletal osteosarcomas

Arms, legs, trunk Legs

Precursor Lesions

Trunk Arms, legs, trunk

There are no known precursor lesions for soft tissue sarcoma, but there are benign forms of soft tissue neoplasms that arise from similar tissues. Benign tumors of soft tissues outnumber sarcomas by 100-fold (Enzinger and Weiss, 1995). These do not invade locally, resemble normal tissue, and are unlikely to recur after excision. Malignant change of benign sarcomas occurs rarely (King et al., 1979), and molecular studies have shown that these benign forms do not develop into malignant forms. For example, leiomyomas and leiomyosarcomas develop from the lining of the digestive tract and the uterus, but leiomyosarcomas do not arise from benign leiomyomas. Lipomas and liposarcoma develop from fatty tissue. Evidence that lipomas do not usually develop into liposarcomas can be found in cytogenetic studies. Most of the chromosomal aberrations found in benign lipomas involve the region 12q14-15, for example, and are not found in liposarcomas where myxoid and round cell sarcoma rearrangements have been sublocalized to 12q13.3) (Enzinger and Weiss, 1995).

Legs, trunk (not involving the bone)

Source: NCI, 2002.

Table 49–2. Major Types of Soft Tissue Sarcomas in Children Tissue of Origin

Type of Cancer

Muscle Striated muscle

Rhabdomyosarcoma Embryonal Alveolar

Smooth muscle Fibrous tissue Fat Blood vessels

Leiomyosarcoma Fibrosarcoma Malignant fibrous histiocytoma Dermatofibrosarcoma Liposarcoma Infantile hemangiopericytoma

Synovial tissue

Synovial sarcoma

Peripheral nerves

Malignant peripheral nerve sheaf tumors (also called neurofibrosarcomas, malignant schwannomas, and neurogenic sarcomas) Alveolar soft part sarcoma Extraskeletal myxoid chondrosarcoma

Muscular nerves Cartilage and bone-forming tissue

Extraskeletal mesenchymal

Usual Location in Body

Most Common Ages

Head and neck, genitourinary tract Arms, legs, head, and neck Trunk Arms and legs Legs Trunk Arms and Legs Arms, legs, trunk, head, and neck Legs, arms, and trunk Arms, legs, and trunk

Infant–4

Arms and legs Legs Legs

Infant–19 15–19 15–19 15–19 15–19 15–19 Infant–4 15–19 15–19

Infant–19 10–14 10–14

Source: NCI, 2002.

Table 49–3. Average Annual Age-Adjusted Mortality Rates (per 100,000) for Connective Tissue Cancer in the United States by Calendar Year, Gender, and Race, 1950–1998*

White men White women Non-white men Non-white women

1950–1954

1955–1959

1960–1964

1965–1969

1970–1974

1975–1979

1980–1984

1985–1989

1990–1994

1994–1998

0.49 0.36 0.40 0.27

0.60 0.44 0.52 0.45

0.71 0.52 0.61 0.58

0.81 0.59 0.79 0.64

0.85 0.62 0.80 0.74

0.85 0.68 0.81 0.86

1.19 0.96 1.21 1.17

1.23 1.00 1.27 1.27

1.3 1.1 1.4 1.4

1.4 1.2 1.5 1.7

*Mortality rates are adjusted to the 1970 United States standard population.

961

Soft Tissue Sarcoma

Molecular Genetic Characteristics of Tumor

Time Trends

Correct classification of soft tissue sarcoma is based on integration of clinical, radiographic, and pathological findings by an experienced pathologist. Genetic factors are often considered as well. The great majority of STS throughout the world is now diagnosed or confirmed microscopically (Parkin et al., 2002). Tissue culture, histochemical stains, and cytogenetic analysis often supplement classification. A major impediment to consistent and appropriate classification is the lack of specificity of the known markers and antibodies (Hibshoosh and Lattes, 1997) although newer molecular classification techniques have been helpful in characterizing certain sarcomas, such as synovial sarcoma (Biermann and Baker, 1999). Molecular genetic characterization of STS has been developed for diagnostic purposes as well as prognostic inference. The frequency of tumor-specific karyotypic alterations in sarcomas is far higher than in any other group of solid human tumors (Busam and Fletcher, 1997). Many STSs show nonrandom, recurrent chromosomal abnormalities and a subset shows potentially diagnostic reciprocal chromosomal translocations, typical of hematopoietic malignancy. Analyses of normal function and expression patterns of genes involved could improve etiologic insights and lead to new therapies (Ladanyi, 1995). Some investigators have found a molecular genetics approach for diagnosis of STS problematic (Pfeifer et al., 2000), but definitive diagnosis of STS should be based on a combination of approaches— molecular, morphological, and clinical (Fletcher et al., 2001). A promising avenue of research into the etiology of STS is identification of specific mutations in the tumors that are associated with particular exposures (Jones et al., 1991; Hussain and Harris, 1998). For example, ras mutations occur frequently in soft tissue sarcomas. In human liver angiosarcomas, a GC Æ AT transition that leads to a GGC Æ GAC mutation at codon 13 in k-ras (Marion et al., 1996) is an important markers of vinyl chloride exposure.

DEMOGRAPHIC PATTERNS Mortality and Incidence in United States Mortality

Between 1973 and 1990, soft tissue sarcoma incidence increased very slowly at an annual percentage increase of 0.5%; however, between 1990 and 1998, the average annual increase in incidence was 2.4%. Mortality due to soft tissue sarcoma increased more rapidly, 12.7% annually between 1976 and 1979 and then declined to 1.3% annually between 1979 and 1998 (Howe, 2001). The apparent increases in incidence and mortality suggest environmental influences, although it is not possible to exclude the role of diagnostic and reporting practices.

Age There is a bimodal distribution of soft tissue sarcoma with a peak in early childhood and a much larger peak in middle age, with a median age at diagnosis of approximately 56 years. The incidence of MFH and liposarcoma appear to increase steadily with age. Specific histologies are more common among children, such as rhabdomyosarcoma, which represents approximately 50% of soft tissue sarcomas for children and adolescents (Gurney et al., 1999). STS accounts for 7.4% of childhood cancer, ranking fifth in cancer incidence in children under the age of 15 after leukemia, CNS tumors, lymphomas, and sympathetic nervous system cancers (Ries et al., 1999). The incidence of STS among those younger than 20 years has changed little between 1975–1979 (1.02 per 100,000) and 1990–1995 (1.13 per 100,000). Age-specific incidence rates are similar for blacks and whites, even though the incidence rates differ (Table 49–4). The age-specific incidence rates for STS begin rising at approximately 40 years of age and reach a steep increase after 60. The median age at diagnosis is 58 for white males, 57 for white females, 44 for black males, and 50 for black females. The median age at death is 66 for white males, 68 for white females, 52 for black males, and 61 for black females.

International Trends

In the United States, current mortality rate for STS for white males is 1.4 per 100,000, for white females 1.2 per 100,000, for black males 1.5 per 100,000, and for black females 1.7 per 100,000 (Table 49–3). When the mortality rates for 1995–1999 were age adjusted to the 2000 United States standard population instead of the 1970 United States standard population, they increased slightly. For example, the mortality rate among white men was 1.4 for the period 1995–1998, adjusted to the 1970 United States standard population; it increased to 1.6 for the period 1995–1999 when adjusted to the 2000 United States standard population.

Incidence Incidence statistics for STS are somewhat uncertain due to the relatively low incidence that is very difficult to define histologically and the poor reproducibility of histologic subtypes. Overall, the incidence of STS among all races in the US was 2.7 per 100,000 in 1990–1999 (Table 49–4). Among white males the incidence was 2.9 per 100,000 and among black males 3.4 per 100,000; among white females the incidence was 1.4 per 100,000 and among black females 1.7 per 100,000 (Ries et al., 2003).

The highest rates for connective tissue tumors in the world are in South African white males and females; rates for males are twice those in other countries—5.0 per 100,000, with Ugandan females a close second among females (Table 49–5). Apparent discrepancies between reports of incidence rates may lie in the population used for standardization of rates. In the Netherlands for example, Nijhuis et al. (1999) reported population-based rates for STS during 1989–1995 (excluding Kaposi, urogenital, and gastrointestinal sarcomas) of 3.5 per 100,000 in contrast to IARC, which reported rates for an overlapping period, 1993–1997, in the Netherlands as 2.3 in males and 1.7 in females. Nijhuis’ series contained 50% of STS patients older than 65 years, whereas the world population has a much younger age distribution, so standardization to that younger distribution would tend to lower the rates of those cancers occurring in older people. The incidence of Kaposi sarcoma in Uganda is extremely high, 37.7 per 100,000 among males and 20.5 per 100,000 among females (Table 49–6). The United States is a distant second with 6.8 per 100,000 among black males and 4.4 per 100,000 among white males. Clearly HIV infection, or some HIV-associated condition, such as HIV-8 infection, is driving the rates of KS in these areas. It is possible that lower rates in other countries may be a function of incomplete ascertainment rather than a true reflection of KS incidence.

Table 49–4. Average Annual Age-Adjusted Incidence Rates (per 100,000) for Connective Tissue Cancer in Connecticut by Calendar Period and Gender, 1935–1989, and for United States by Calendar Period and Gender, 1990–98*

Men Women

1935–1939

1940–1944

1945–1949

1950–1954

1955–1959

1960–1964

1965–1969

1970–1974

1975–1979

1980–1984

1985–1989

1990–1994

1994–1998

1.6 1.5

1.8 1.7

1.9 1.4

2.0 1.7

2.4 1.8

3.1 2.0

2.8 1.7

2.5 1.7

2.7 1.5

2.8 2.1

2.6 1.6

2.6 1.9

2.9 2.2

*Incidence rates for Connecticut and for the United States for 1990–1998 are adjusted to the 1970 United States standard population. Case selection was based on ICDO site code 171 only, not histology. Soft tissue sarcoma of the heart is excluded. Soft tissue sarcoma of the heart is included in the United States figures.

962

PART IV: CANCER BY TISSUE OF ORIGIN Table 49–5. Rates of Soft Tissue Sarcoma in Selected Areas of the World, by Sex and Race Male Country Africa

Race

Algeria Nigeria Congo Uganda South Africa Black White

Europe North America

France Netherlands England United States

Black White

Years

Number

Female Rate

Number

Rate

1993–1997 1998–1999 1996–1999 1993–1997

51 18 24 35

1.1 1.4 2.8 2.2

48 7 10 46

0.9 0.6 1.0 3.1

1989–1992 1989–1992 1993–1997 1993–1997 1995–1997 1993–1997 1993–1997

843 543 356 1123 2811 174 1543

2.4 5.0 2.0 2.3 1.7 2.7 2.6

721 388 291 940 2371 163 1121

1.7 3.2 1.4 1.7 1.2 2.0 1.8

Source: Parkin et al., 2002a.

ENVIRONMENTAL FACTORS A major problem with all studies of the etiology of soft tissue sarcoma is the generally low statistical power to find an association between a risk factor and STS, a cancer with relatively low incidence and inherent misclassification of histology. Partially due to the rarity of this tumor, different histologic subtypes are often grouped together in studies, resulting in a loss of specificity and possibly masking the behavior of the different subtypes. This problem is further complicated in studies of environmental causes by the difficulty of making accurate environmental measurements. It is most often not possible, except in some unusual occupational cohorts, to have an accurate history of prior environmental exposure. Thus many investigators have relied on self-reported occupational histories as well as self-reported exposures, which may suffer from misclassification. In some cases, more precise estimates of exposure can be made with the assistance of an expert industrial hygienist (Piacitelli et al., 2000).

Phenoxy Herbicides, Dioxin, and Pesticides A controversial risk factor most often investigated in relation to the development of STS is dioxin (2, 3, 7, 8-tetrachlorodibenzo-p-dioxin, or TCDD), a contaminant of industrial processes. Numerous epidemiologic studies have examined occupational groups and those accidentally exposed to putative risk substances, such as dioxin, to substantiate or refute earlier reports of associations with soft tissue sarcoma (Hardell, 1977). As dioxin is a “contaminant” of industrial processes, it is difficult to measure indirectly by means of exposures to those processes. Therefore, it is not surprising that there continues to be conflicting evidence of an association between exposure to dioxin and soft tissue sarcoma (Table 49–7). The question is, however,

important, because there is still a poor understanding of the etiology of soft tissue sarcoma. The polychlorinated dibenzo-p-dioxin (2, 3, 7, 8-TCDD) consists of a group of 75 congeners. Dioxin is stored in the lipids in the body and can be measured in the lipid fraction of serum (Schecter, 1994). The measured background values of dioxin were 5 parts per trillion (ppt) and 19,110 ppt in children accidentally exposed to dioxin emissions in Seveso, Italy in 1976 (Landi et al., 1997). In most but not all areas of the world dioxin levels as measured in serum have declined. Contaminated feed clearly increases the concentration of dioxin in the fat of chickens and pigs (Neuberger et al., 2000). In Spain blood levels are steadily increasing and could reflect an increase of exposure from foods or other unidentified sources (Gonzales et al., 2001). The US Environmental Protection Agency has revised upwards the estimated risks of cancer, particularly soft tissue sarcoma, due to dioxin in the food chain (Pohl et al., 2002) even though emissions of such chemicals and blood levels have decreased dramatically in the past 10 to 15 years (Ayleward and Hays, 2002). The presumed mechanism of action in humans and animals is the binding of the dioxin-like compound to the Ah receptor; the relative potency of the dioxin compound seems to depend on how well it fits the receptor. Polychlorinated biphenyls fit more weakly but their relative abundance makes them biologically important. TCDD and PCDD (the polychlorinated dibenzo-p-dioxins) and PCDF (polychlorinated dibenzofurans) are unintended byproducts of the production of herbicides, wood preservatives, and contaminants of PCBs. PCDD and PCDF are also byproducts of waste incineration and metal processing (Schecter, 1994). Since the 1970s epidemiologic studies have shown relationships between exposure to dioxin and the development of soft tissue sarcoma. Case-control studies have often found positive associations

Table 49–6. Rates of Kaposi Sarcoma in Selected Areas in the World, by Sex and Race Male Country Africa

Race

Algeria Nigeria Congo Uganda South Africa Black White

Europe North America

France Netherlands England United States

Source: Parkin et al., 2002b.

Black White

Female

Years

Number

Rate

Number

Rate

1993–1997 1998–1999 1996–1999 1993–1997

16 3 37 843

0.4 0.2 4.3 37.7

3 1 14 533

0.1 0.1 1.4 20.5

1989–1992 1989–1992 1993–1997 1993–1997 1995–1997 1993–1997 1993–1997

232 68 204 425 672 560 2788

0.7 0.6 1.2 0.8 0.5 6.8 4.4

69 17 26 25 68 32 70

0.2 0.1 0.1 0.0 0.0 0.3 0.1

Table 49–7. Summary of Selected Studies of Environmental Exposures and Soft Tissue Sarcoma

Exposure

Study Design

Dioxins, Matched chlorophenols, casephenoxyacetic control acids and herbicides

Author, Yr, Place

Number of Subjects

Hardell, 1979, Sweden

52 cases (21 living, 31 deceased), 4 matched controls– 208 sex, age, residence, or dead controls 110 cases (38 deceased), and 220 referents

Group

Estimate of Risk

95% Confidence Interval

With previous exposure, age range 26–80 years

OR = 6.2 RR = 5.7

Patients with STS that had been diagnosed and reported to the National Social Welfare Board between 1974–1978, who lived in Swedish counties at time of diagnosis Chemical sprayers/ applicators in Finland

OR = 5.1 OR = 4.7

Increase in the risk for STS after exposure, but this risk related also to exposure to phenoxy acids free from impurities.

No cases

No increase in cancer mortality was detected, and the distribution of cancer was unremarkable.

(2.9–11.3)

Case-referent study

Eriksson, 1981, Sweden

Ongoing prospective cohort study

Riihimaki, 1982, Finland

Case-control

Smith, 1984 82 cases, 92 New Zealand controls Greenwald, 281 cases (151 1984, US living, 130 deceased), 281 living controls

National Cancer Registry subjects Draftable men of New York City during the time of the Vietnam War

OR = 1.3

(0.6–2.7)

OR = 1.0

(0.0–0.0)

Coggon, 1986, UK

5784 male workers (1 death from STS)

Occupational exposure to male employees at an MCPA manufacturing company

SMR = 0.6

(0.02–3.47)

HospitalKang, 1986, based caseVietnam comparison Study

234 cases, 13,496 comparison patients

OR = 0.83

(0.63–1.09)

Case-control

217 cases from the Armed Forces Institute of Pathology, 599 controls

Vietnam era-veteran patients who served in the US military between 1964–1975 and were treated in one of the 172 VA hospitals between 1969 and 1983 Men who were of draftable age during the Vietnam conflict

OR = 0.85

(0.54–1.36)

Case-control

Cohort study, mortality

Kang, 1987, Vietnam

1926 men

All newly diagnosed cases OR = 1.0 of STS, HD, and NHL among white male Kansas residents, aged 21+, from 1976–1982, identified through Kansas population registry, and white men from Kansas population. Female rice weeders RR = 2.7 in a northern Italy region

PopulationHoar, 1986, based caseUS control

133 STS cases, 948 STS controls

PopulationVineis, 1986, based caseItaly referent study

68 cases (31 women) and 158 referents (71 women)

PopulationWoods, 1987, based caseUS control

128 cases, 694 controls, aged 20–79

Occupational exposure to male population of western Washington State

Population study, mortality

30,703 subjects, aged 20–74 years, separated into 3 different zones

Contaminated populated area, near Seveso, Italy.

Bertazzi, 1989, Italy

Comments

OR = 0.99

(0.7–1.6)

Approximately six fold increase in the risk for this type of tumor.

No significant associations were found for “Agent Orange” or other variables that might be related to herbicide exposure. The findings suggest that any risk of STS is less than that indicated by earlier studies of 2,4,5-T and 2,4,5trichlorophenol and is small in absolute terms. No significant association of STS and previous military service in Vietnam was observed. Possibly latency period too short. Vietnam veterans in general did not have an increased risk of STS when compared to those men who had never been in Vietnam. STS was not associated with pesticide exposure.

(0.59–12.37) Further increased when attention was restricted to women exposed in the whole 1950–1955 period. (0.7–1.5) No risk associated with overall duration or intensity of chemical exposure or exposure to any specific phenoxyherbicide. Mortality from several cancers was elevated. However, no definite patterns related to exposure classification were apparent.

(continued)

Table 49–7. (cont.)

Exposure

Study Design

Author, Yr, Place

Number of Subjects

PopulationEriksson, 1990, 237 cases, 237 based caseSweden controls control

Case-control

Wingren 1990, Sweden

96 cases, 450 population-based controls, 200 cancer controls

Cohort study

Zober, 1990, Germany

247 partially heavily exposed, (78 of which died), divided into 3 cohorts 86 cases

PopulationCiccone, based case1991, Italy control

964

Retrospective cohort study of mortality

Fingerhut, 1991, US

5172 workers (4 deaths from STS)

Cohort, mortality

Manz, 1991, Germany

1583 workers (1184 men, 399 women)

Historical cohort study of mortality

Saracci, 1991, International

Cohort Mortality study

Green, 1991, Canada

18,910 production workers or sprayers (4 observed deaths due to STS) 1222 men with 25,274 person years

Cohort study, mortality

Coggon, 1991, UK

2239 men, employed during 1963–1985

HospitalFranceschi, based case1992, Italy control

93 STS cases, and 721 controls

Cohort study

2310 workers

Bas Bueno de Mesquita, 1993, Netherlands

Group Male patients, aged 25–80 years, who were diagnosed from January 1978–June 1986, and were reported to the Regional Cancer Registry Gardeners Railroad workers Construction workers Asbestos Pressure impregnating agents Unspecified chemical workers

BASF employees exposed to TCDD after an accident that occurred in Germany in 1953 Residents from the provinces of Novara, Vercelli, and Alessandria Occupational exposure of workers of 12 plants in the US that produced chemicals contaminated with TCDD Employed in a chemical plant in Germany that produced herbicides, TCDD International register of workers or sprayers from ten countries Forestry workers at a public electrical utility who had worked for 6 months or more during 1950–1982 and had routine exposure Four British cohorts of chemical manufacturers, traced through the National Health Service Central Register and the National Insurance Index Aviano Cancer Center patients in the FriuliVenezia Giulia region of Northeast Italy

Workers from 2 plants in the Netherlands, who had occupational exposure

Estimate of Risk

95% Confidence Interval

RR = 1.80

(1.02–3.18)

OR = 4.1 OR = 3.1 OR = 2.3 OR = 1.8 OR = 1.7 OR = 1.6

Comments Statistically significant increased rate ratio.

Unexpected and unclear This grouping of jobs seems to confirm the associations reported for pheoxy herbicides and chlorophenols. Reported little evidence of recall bias, unlikely confounding by smoking. 34-year follow-up. Among those with chloracne, for 20+ yrs after exposure.

No cases

The overall proportion of patients alive 3 years after a diagnosis of STS was 57%. Mortality from STS was increased, but not significantly.

No cases

Mortality follow-up. No death due to STS (ICD 171) was observed.

SMR = 196 (53–502)

Non-significant twofold excess risk was noted for STS.

No cases

33-year follow-up. Overall, no excess mortality was found in this cohort relative to reference population.

SMR = 0

0–2087

No cases of soft tissue sarcoma were recorded.

OR = 0.4

(0.1–1.2)

Workers who reported exposure to chemical agents or other solvents for more than 10 years had, respectively, a 1.8-fold and a 2.2-fold higher risk of developing STS. No cases of STS were encountered.

No cases

Table 49–7. (cont.)

Exposure

Study Design

Number of Subjects

Cohort study, update

Lynge, 1993, Denmark

940 workers (5 deaths due to STS)

Manufacturing and packaging workers from Danish manufacturing plants

SIR = 2.3 SIR = 6.4

(0.6–5.8) (1.3–18.7)

Cohort study

Flesch-Janys, 1995, Germany

1189 male workers

Male workers from a Herbicide-producing plant in Hamburg, Germany

RR = 3.30

(2.05–5.31)

Nested case control study Cohort, mortality

Kogevinas, 1995, International Ott, 1996, Germany

11 cases, 55 controls

OR = 10.3

(1.2–91)

Cohort study, mortality

Becher, 1996, Germany

1493 subjects, (960 of which, died during time of study)

Cohort

95,126 person-yrs

Bertazzi, 1997, Italy Rix, 1997, Denmark

20,943 men, 4415 women

Employees exposed to high doses of TCDD after a 1953 reactor accident Occupational exposure to workers from four plants in Germany

SMR = 101 (92–111)

Increased overall cancer mortality after longterm exposure to these agents. Slightly increased SMR was found for some cancers, including STS, but statistically insignificant. 15-year follow-up

Rice growers in Novara Province, northern Italy, during an observation period between 1957–1992 Hi exposure—Seveso

No cases

Paper mill employees

RR - 2.7

(1.5–4.5)

(0.8–4.4)

SMR = 2.0

2 deaths from STS

SMR = 1.4

562 subjects (549 males, 13 females) 294 cases, 1908 controls

Exposed workers in a Netherlands chemical factory Self-reported use

Case-control

Hoppin, 1999, US

Self-reported herbicide use, MFH

Cohort study, mortality

Fleming, 1999, US

251 living men 30– 60; 1908 living controls 33,658 (10% female)

Licensed pesticide applicators of Florida

Cohort

Bertazzi, 2000, Italy

6745 individuals

Hi exposure—Seveso

Ecological

Costani 2000, Italy

20 STS compared to expected

Spatial cluster design Case-control

Viel, 2000, France

45 cases

Case-control Cohort study, mortality

Hoppin, 1998, US Hoppin, 1998, US Ward, 2001, international

12,700 men working in the vinyl chloride industry

Municipal solid waste incinerator 10+ years Work with plywood, Leiomyosarcoma Work with wood or sawdust, Leiomyosarcoma Angiosarcoma

Based on small numbers, it adds to the evidence for a possible association between phenoxy-herbicide exposure and risk of STS. Strong dose-dependent relation between mortality due to cancer or ischemic heart diseases and exposure to PCDD/F.

No cases of STS have been found to date.

6 deaths from STS

Retrospective cohort Study Case-control

Comments

No cases of STS

Kogevinas, 1997, International Kogevinas, 1997, International Hooiveld, 1998, Netherlands Hoppin, 1998, US

Cohort study, mortality Cohort, mortality

Vinyl chloride

484 deaths from STS

Cohort, study, Gambini, mortality 1997, Italy

Cohort

Wood and products

243 men

Group

Estimate of Risk

95% Confidence Interval

Author, Yr, Place

RR = 1.8

(1.2–2.5)

OR = 1.8

(1.1–2.9)

OR = 2.9

(1.1–7.3)

No cases SIR = 2.3 SIR = 2.6

(1.3–3.5) (1.5–4.0)

SIR = 1.4

P < 0.01

OR = 2.8 OR = 2.2

(2.5–24.7) (1.3–3.9)

OR = 2.4

(1.4–4.2)

SMR = 2.9

(1.9–4.3)

Among women, no increase among men who had little contact with paper. Risk increased from time since first exposure.

Results support evidence of a high cancer risk due to exposure. Phenoxyherbicide use was not associated with STS although it did have sufficient statistical power. A prevalence study, probable multiple comparisons is an issue. The pesticide applicators were consistently and significantly healthier than the general population of Florida. 20-year follow-up, Power still low, no mention of low zone. Comparison—Varese province Comparison— pooled registries Emissions = 16.3 ng TEF/m3.

Marked exposure-response relationship.

(continued)

965

966

PART IV: CANCER BY TISSUE OF ORIGIN

Table 49–7. (cont.)

Exposure Radiation therapy

Study Design

Number of Subjects

Group

Case study

Kim, 1978, US

20 cases of STS

STS arising from irradiated tissues

Retrospective study

Brady, 1992, US

565 previous MSKCC patients

Patients with sarcoma and a second malignancy seen at MSKCC between 1943–1989

Registry follow-up of breast cancer

Karlson, 1998

Case-control study

Cozen, 1999, Los Angeles County SEER program

Cohort study

122,991 br ca w/subs STS 116 (3% total STS) 908,580 person years Cases with angiosarcomas of the upper extremity (n = 20) and of the chest (n = 48) 10 patients sarcomas—5 STS

Multiinstitutional data base to review second cancers after Ewing sarcoma Huang, 2001, 194,798 women US with invasive breast cancer and subsequent STS, SEER

Cohort study, incidence

Virus—HIV

Author, Yr, Place

Retrospective Review

Blanchard, 2002, US

34 women with breast sarcoma/ carcinoma

AIDS cancer registry

Biggar, 2000

4954 children with AIDS

Linked AIDS, Frisch, 2001, cancer US registry data

302,834 adults 15– 69 with HIV/ AIDS

Sarcoma after breast cancer

Retrospective review from 1975–2001 of women with a sarcoma in an irradiated field after treatment for breast cancer 124 with cancer before, at, or after diagnosis, 4 with leiomyosarcoma Immunosuppression and cancer

between STS and a variety of risk factors, particularly dioxin, phenoxyacetic acids, and pesticides. However, disentangling the various contributions of chlorophenols, phenoxyherbicides, and dioxins is complicated because chlorophenols are used in the production of phenoxyherbicides and dioxins have been found in both. Hoppin et al. (1998) studied occupational exposures among 294 men, aged 30–60 from eight population-based registries and 1908 controls obtained by random-digit dialing. An industrial hygienist rated occupations according to their intensity of exposure to chlorophenols, in those jobs involving wood preservatives, cutting oils, sawmills, leather tanning, or shoe dust. Risk for STS was 1.8 (95% CI: 1.1–2.9) for “ever high-intensity chlorophenol exposure”. There was a significant duration-response trend among more highly exposed subjects. With 10 or more years’ exposure, the risk increased to 2.8 (95% CI: 2.5–24.7). In a follow-up study of the same cases and controls to assess histologic subtypes, they found that self-reported herbicide use was associated with MFH histology (OR = 2.9, 95% CI: 1.1–7.3) (Hoppin et al., 1999). Chlorophenol exposure and cutting oil expo-

Estimate of Risk

SIR = 1.9 OR = 9.5

95% Confidence Interval

(1.5–2.2) (3.2–28.0)

Comments Latent period from first irradiation to the discovery of sarcoma varied from 3 to 40 years, with a median of 12 years. 5-year survival for STS of the extremity vs. other soft-tissue sites was 41% vs. 41%. Angiosarcoma and Lympheclema (non RT) RT w/other ¥ angio.

OR = 59.3 OR = 11.6 OR = 3.3

(21.9–152.8) Upper extremity (4.3–26.1) angiosarcomas (1.1–1.7) Chest and breast angiosarcomas Other sarcomas of the chest and breast. Cumulative (SD = 2.4%) Cumulative incidence incidence rate of secondary = 6.5% sarcoma was dose dependent (P = 0.002). No sarcomas developed in patients who had received <48 Gy. SIR = 26.2 (16.5–41.4) Radiation-treated, SIR = 2.5 (1.8–3.5) angiosarcomas SIR = 2.1 (1.1–4.4) Radiation-treated, other SIR = 1.3 (1.0–1.7) STS No radiation-treatment, angiosarcomas No radiation-treatment, other STS. Radiation-induced sarcoma is a late complication of definite treatment for breast carcinoma. RR > 1900

RR = 3.3

(2.6–4.1)

Suggests that profound immunosuppression is necessary for these tumors in children. Possible misclassification of KS with STS.

sures were associated with MFH and leiomyosarcoma although not significantly. No occupational risk factor was associated with liposarcoma. It should not be surprising that some cohort studies have not found associations with STS because even large studies with relatively long follow-up do not often have the statistical power necessary to find an association (Boroush and Gough, 1994). The low incidence of STS, misclassification of outcome or exposure, and other biases all combine to reduce power. Possibly the most convincing cohort study was conducted in 5172 men at 12 different chemical plants in the United States (Fingerhut et al., 1991) where the group most highly exposed to dioxins had an almost 50% excess risk of dying from cancer, mostly STS. A followup of this cohort in 1999 (Steenland et al., 1999) identified no new STS deaths but the risk remained high. Other studies have had mixed results. Other studies have found excesses of STS among manufacturing workers putatively exposed to dioxin (Zack and Suskind, 1980; Cook et al., 1980; Cook, 1981; Johnson et al., 1981).

Soft Tissue Sarcoma A number of investigations have focused on Vietnam veterans who may have been exposed to dioxins due to the extensive herbicide use during the Vietnam war. Kogan and Clapp (1988) found an elevated PMR of 880 and an elevated odds ratio for mortality (OR = 5.6, 95% CI: 2.4–11.1) for STS among Vietnam era veterans, as well as nonveteran males. Other investigators have also reported increased risk for STS in Vietnam veterans (Sarma and Jacobs, 1983; Anderson et al., 1986; Holmes et al., 1986). On the other hand, Greenwald et al. (1984) and others (Lawrence et al., 1985; Kang et al., 1986, 1987; Breslin et al., 1988; Goun and Kuller, 1988; Selected Cancers Cooperative Study Group, 1990; Lathrop et al., 1984) were not able to find any evidence of an association between service in Vietnam and STS risk (OR = 0.53, 95% CI: 0.21–1.31). Kogevinas et al. (1995), in an international cohort, first reported an excess risk of STS associated with exposure to any phenoxy herbicide (OR = 10.3, 95% CI: 1.2–91), and to each of the three major classes of phenoxy herbicides (2,4-dichlorophenoxyacetic acid, 2,4,5trichlorophenoxyacetic acid and 4-chloro-2-methylphenoxyacetic acid), to any polychlorinated dibenzodioxin or furan (OR 5.6, 95% CI: 1.1–28) and to 2,3,7,8-tetrachlorodibenzo-p-dioxon—OR = 5.2), 95% CI: 0.85–32). STS was not associated with exposure to raw materials or other process chemicals. However, Becher et al. (1996) found no cases of soft tissue sarcoma in a cohort of 2479 male German workers exposed to phenoxy herbicides and dioxins with 54,063 person-years of observation. Rix and Lynge (1997) reported no overall increase in cancer in a cohort of 20,953 men and 4415 women who worked in Danish paper mills although female workers had an increased risk of STS (14 cases, STR–2.7, 95% CI: 1.5–4.5). There was no excess among males who had little direct contact with paper, which was the likely source of contaminants. In a follow-up study on a large international cohort of 21,863 male and female workers (Kogevinas et al., 1997), among workers exposed to phenoxyherbicides the SMR for STS was 2.03 (95% CI: 0.75–4.43) and higher than for all malignant neoplasms, non-Hodgkin lymphoma, and lung cancer. This association was strengthened by the fact that risks increased from time from first exposure. Hooiveld et al. (1998) followed a Dutch cohort of workers exposed to phenoxy herbicides, chlorophenols, and contaminants with 13,634 person years of follow-up. Although they found an overall cancer mortality of 4.1 (95% CI: 1.8–9.0), there were no soft tissue sarcomas diagnosed. Likewise, even though overall mortality was increased. Fleming et al. (1999) studied a cohort of 33,658 pesticide applicators in Florida and found they were actually healthier than the general population, with no cases of STS (January 1, 1975–December 31, 1993). Italy was the focus of much concern about dioxin exposure—an accident that occurred in Seveso in 1976 has been a “natural experiment”. Several follow-up studies have only shown “suggestive” increases in STS. Bertazzi et al. (2001) in a 20-year follow-up found no STS in the two zones nearest the accident, but in a third area further out but exposed there was a significant increase in sarcoma incidence (Puntoni et al., 1986). Smaller studies with good measures of exposure have not found any cases of STS (Millham, 1982; Smith et al., 1982, 1983, 1984; Gallagher and Threlfall, 1984; Wiklund and Holm, 1986; Woods et al., 1987; Wiklund et al., 1988; Hoar et al., 1986; Green, 1991; Serraino et al., 1992; Ott and Zober, 1996). Published ecological studies add some weight to the evidence for an association between environmental contaminants and soft tissue sarcoma, although these studies may be prone to publication bias. An ecological study Viel et al. (2000) studied the spatial distribution of cancers around a French municipal solid waster incinerator with high emission levels of dioxin (16.3 ng international toxic equivalency factor/m3) and found an SIR for STS of 1.4 (focused test p value = 0.0004) for those within the electoral wards of the area of Doubs, France that was stronger closer to the incinerator (SIR = 3.4, p = 0.008). In another ecological assessment, Costani et al. (2000) observed an SIR for soft tissue sarcoma of 2.3 (1.3–3.5) and 2.6 (1.6–4.0) in two registries near the industrial city of Mantua in North-

967

ern Italy. They have hypothesized that the excess may result from industrial emissions, which likely included dioxin. In sum, the effects of dioxin and similar compounds associated with risk for the development or death from STS have ranged from “no effect” to a 10-fold increased risk. It is critical to continue to examine this potential association as “human activities result unavoidably in population exposure” (Samet 1996).

Vinyl Chloride Vinyl Chloride appears to have very specific effects, and has been strongly associated with the development of angiosarcomas (Creech and Johnson, 1974; Popper et al., 1978). Ward et al. (2001) updated a study of mortality and cancer incidence among 12,700 Europeans working in the vinyl chloride industry. Excess liver cancer was found as in the earlier study. The first study found a 2.9-fold SMR (1.9–4.3). The results of the newer study are consistent with the first, reporting an SMR for liver cancer of 2.4 (95% CI: 1.8–3.1) and soft tissue sarcoma of 1.9 (95% CI: 0.7–4.1) as well as an SIR for liver of 3.9 (95% CI: 2.7–5.7). Strong exposure-response relations are reported for angiosarcomas where time since first employment, duration of employment, and cumulative exposure were all associated with extremely high risks ranging from 7.9 (95% CI: 1.7–37.3), to 15.7 (95% CI: 5.6–44.0), and 88.2 (26.4–295), respectively. Smaller studies, such as Rhomberg (1998), report on case series, with similar conclusions. Of 21 cases of angiosarcomas, four had clear exposure to vinyl chloride.

Radiation The role of therapeutic radiation in inducing soft tissue sarcomas is well established although the incidence is low. Several populationbased registry studies have supplied dose-response estimates for the role of radiation in the development of sarcoma (Table 49–7). Most studies, however, are case studies or institutionally based (Kim et al., 1974; Brady et al., 1992; Brady et al., 1994; Salloum et al., 1996; Lagrange et al., 2000; Blanchard et al., 2002). Using cancer registry data, several investigators have identified women with breast cancer who were treated with radiation and followed for subsequent diagnoses of sarcoma (Harvey and Brinton, 1985; Curtis et al., 1985; Karsson et al., 1990; Taghian et al., 1991; Cozen et al., 1999; Huang and Mackillop, 2001; Yap et al., 2001). It should be noted that the risk estimates are likely smaller than the reported estimates due to the probable incomplete exposure information collected by the SEER registries. Huang et al. (2001) followed 194,798 women diagnosed with invasive breast cancer and found 135 with a subsequent diagnosis of soft tissue sarcoma: 54 had been treated with radiation and 81 had not. Of those treated with radiation, angiosarcomas predominated (SIR 26.2, 95% CI: 16.5–41.4), but other radiation-treated sarcomas were also significantly increased (SIR 2.5, 95% CI: 1.8–3.5). Even among the breast cancers that were not treated by radiation, significant increases in sarcoma incidence were noted (SIR 1.4, 95% CI: 1.1–1.7). It is likely that some underlying low-penetrant genetic factors are responsible for the association between breast cancer and sarcoma. Yap et al. (2002) reviewed 274,572 cases of invasive breast cancer from the SEER registry and found 263 subsequent sarcomas; 87 of these had received radiation therapy and 176 none. Angiosarcoma accounted for 56.8% and MFH accounted for 15.9% of the sarcomas occurring within the field of radiation. Although survival was poor for sarcoma cases after radiation, it was not significantly different (p = 0.60) between patients receiving or not receiving radiation therapy for their primary breast cancer. Five-year survival rates were 27.7% for STS within the radiation field and 35% for those outside the radiation field. Another variant of this type of study was conducted in Los Angeles (Cozen et al., 1999). Females diagnosed with angiosarcomas of the upper extremity (n = 20) or chest (n = 48) were compared with females diagnosed with other cancers during the same time period (n = 266,444). A comparison of prior history of invasive breast cancer showed an increased risk of 59.3 (95% CI: 21.9–152.8) for upper extremity angiosarcomas as well as

968

PART IV: CANCER BY TISSUE OF ORIGIN

chest angiosarcomas, OR = 11.6 (95% CI: 4.3–26.1) and other sarcomas of the chest and breast (OR = 3.3, 95% CI: 1.1–1.7). These results are highly consistent among studies. Other models, such as radiationinduced sarcoma of the head and neck (Patel et al., 1999), have been investigated. Sarcomas induced by radiation after cancer of the head and neck were more often MFH histologically and had a median latency of 17 years. The genetic defect is enhanced by radiation therapy among survivors of retinoblastoma (Wong et al., 1997). In a cohort of 961 hereditary Rb patients followed for 20 years, 114 soft tissue sarcomas of all histologies developed. A stepwise increase occurred at all dose categories and was significant at 10–29.9 Gy and 30–59.9 Gy. Risk from radiation began at 5 Gy, rising to 10.7–fold at 60 Gy+ (P < 0.05). The use of thorotrast (colloidal thorium dioxide) as cancer treatment has been discontinued after several different cancer types were associated with its use, particularly hepatic angiosarcomas (Falk, 1979a; Locker et al., 1979; Van Kaick et al., 1983, Kato and Kido, 1987; Mays, 1988).

Viruses Viruses have long been suspected as causal in the development of soft tissue sarcoma and the association of KS with AIDS is somewhat supportive of this hypothesis. The risk of AIDS-related KS differs by HIV exposure category and has changed over time, as well. At the beginning of the epidemic, just prior to 1980, approximately 40%–50% of homosexual men with AIDs developed KS, dropping to less than 15% by the late 1980s. The prevalence of AIDS-related KS has been much lower in IV drug users (10%), hemophiliacs (4%), and children with AIDS (3%), (Beral et al., 1990). In 1994 a novel herpes virus, HHV8, was identified as necessary but not sufficient for the development of KS (Chang et al., 1994). Clearly there are other important factors as only a fraction of HHV-8–infected people develop KS. HHV-8 is transmitted by saliva and it is not known whether blood-borne transmission occurs (Pauk et al., 2000). New technologies, such as gene expression arrays, may help determine if there is a viral gene-host interaction and identify molecular targets for treatment and prevention (Paulose-Murphy et al., 2001; Mikovits et al., 2001).

Tobacco Use Tobacco use is inconsistently associated with soft tissue sarcoma. An early case-control study of soft tissue sarcoma by Zahm et al. (1989) found a significant increase in risk with the use of smokeless tobacco (OR = 1.8, 95% CI: 1.1–2.9). This finding was not replicated in their cohort mortality study of US veterans (Zahm et al., 1992) although that study found an excess risk for cigarette smokers (RR = 1.8, 95% CI: 1.1–2.9). Other studies have not replicated either the association of smokeless tobacco with risk for sarcoma (Franceschi and Serraino, 1992) or the association of smoking with sarcoma (Serraino et al., 1991; Hardell and Sandstrom, 1979; Gebauer, 1982, Kang et al., 1987; Vineis et al., 1987; Woods et al., 1987; Hardell and Eriksson, 1988; Eriksson et al., 1990; Schwartz et al., 1996).

Exogenous Hormonal Factors Several lines of evidence suggest that exogenous hormone use might be associated with soft tissue sarcoma. In the first place, high levels of body fat seem to be associated with increased risk for STS (Tavani et al., 2000; Schwartz et al., 1996; Zahm et al., 1989) although not in all studies (Serraino et al., 1991); and in the second place some clinical observations suggest an association: oophorectomy seems to have led to regression of sarcoma metastases (Abu-Rustum et al., 1997); steroids were associated with the development of angiosarcomas of the liver (Falk et al., 1979b). Schwartz et al. (1996) found an excess of uterine leiomyosarcoma (OR = 1.7, 95% CI: 0.7–4.1) that appeared to have a weak dose-response association in women taking oral contraceptives. Endogenous hormonal factors have also been associated with risk (see below).

Diet Diet has been little studied, but Serraino did find positive associations with dairy products (<0.01) and oil (<0.05) and negative association with whole grain bread and pasta (0.4, 0.2–0.9) as did Tavani et al. (1997) (OR = 0.20, p < 0.01). Rhabdomyosarcoma in childhood has been associated with diets rich in organ meats (Grufferman et al., 1982). Considerable debate has focused on the suggestion that dioxins might concentrate in the food supply and therefore some individuals may have higher exposures to dioxins than otherwise anticipated (Paustenbach, 2002). It is not certain that epidemiologic studies can actually answer this issue.

Other Factors A case-control study of HIV-seronegative people in Uganda that examined risk factors for Kaposi sarcoma found that specific tribal groups had higher risk for KS. Those with higher household income, those owning goats and pigs, and those who rarely or never wore shoes were at highest risk for KS (Ziegler et al., 2003). In an Italian case-control study, exposure to solvents increased risk 2.2-fold (95% CI: 0.9–5.5) non-significantly (Franceschi and Serraino, 1992). Other chemicals, such as formaldehyde, have been associated with excess risk for STS in some studies (Stayner et al., 1988) and not in others (Acheson et al., 1984; Blair et al., 1986). An interesting association between employment in an abattoir and STS was reported by Pearce et al. (1988). Such workers might be exposed to chemicals in plastics used to wrap meat or in the treatment of pelts or exposed to zoonotic viruses. However, a larger study in Sweden was not able to reproduce this finding (Boffetta et al., 2000).

HOST FACTORS Knowledge of host factors associated with soft tissue sarcoma is limited due to the difficulty of conducting epidemiologic studies in a rare tumor with multiple histologic subtypes.

Associated Cancers Another approach taken to understand the etiology of soft tissue sarcoma is to conduct studies of the relationship between STS and other cancers using information from tumor registries. In these studies, individuals diagnosed with soft tissue sarcoma are evaluated for the incidence of cancers prior to or subsequent to STS. The underlying hypothesis of such studies is that strong associations might yield clues as to common environmental or genetic factors. Several studies of LiFraumeni syndrome families have been conducted with similar goals. Follow-up studies of childhood cancers have found a high incidence of soft tissue sarcoma (Smith et al., 1993; Wong et al., 1997; Neglia et al., 2001; Bhatia et al., 2002; Hwang et al., 2003). Even among adults STS is a relatively common second cancer (Keohan and Antman, 1999). Bhatia et al. (2002) investigated the incidence of second cancers among 8831 children diagnosed with acute lymphoblastic leukemia between 1983 and 1995 (54,883 person-years of follow-up). The SIR for all cancer was 7.2 (95% CI: 5.5–9.1), and the SIR for soft tissue sarcoma was 9.1 (2.4–20.2). They found a higher risk among females and no correlation with risk for STS due to therapeutic radiation even though there was a dose-response relationship between radiation and other tumors. However, the follow-up times were short in relation to other studies of STS that have found radiation effects. In the relatives of probands who form an unselected cohort of 402 breast cancer patients, Bennet et al. (2002) found an increased incidence of STS in mothers only (RR 15.4, P = 0.001). Berking and Brady (1997) reported on a large number of sarcomas in association with melanoma that were notable for a family history of cancer (50%). Hemminki and Li (2001) reviewed soft tissue sarcomas in Swedish registries in parents and by histology. Stomach cancer and endocrine gland cancer in parents were associated with fibrosarcoma (SIR = 3.19,

Soft Tissue Sarcoma

969

95% CI: 1.69–5.17; SIR = 3.66, 95% CI: 1.41–9.23, respectively). Breast cancer in parents was associated with leiomyosarcoma (SIR = 2.04, 95% CI: 1.08–3.30). Second cancers following soft tissue sarcomas were notable for additional soft tissue sarcomas (SIR 1.94, 95% CI: 0.77–3.65). In an Israeli series of patients with soft tissue tumors, malignant fibrous histiocytoma (MFH) histology has been associated with the development of renal cell carcinoma (Merimsky et al., 2001). Mothers of young adults with soft tissue sarcoma (Hartley et al., 1990) had no overall excess risk for cancer, but mothers of patients with synovial sarcoma had more breast cancer than expected.

rosarcoma or fibrosarcoma. Zoller et al. (1997) reported that sarcoma was over-represented among a series of adult Swedish patients with NF1. Werner syndrome, a rare autosomal recessive disorder characterized by premature aging, may be a connective tissue disease (Usui et al., 1984) and clinical reports have linked the nevoid basal cell carcinoma syndrome to fibrosarcoma and rhabdomyosarcoma, tuberous sclerosis to rhabdomyosarcoma, and familial hydronephrosis to congenital renal sarcomas (Mulvihill, 1975).

Genetic Factors and Familial Susceptibility

The role of immunosuppression as an etiologic factor in soft tissue sarcoma has been long suspected (Vineis and Zahm, 1988; Vineis, 1999). Dioxin and similar environmental chemicals are immunotoxic (Vos et al., 1997/98), so the extent to which dioxin is associated with the development of STS may suggest a role for immunosuppression by other means. Earlier studies reported an excess of STS in patients receiving immunosuppression for renal transplantation and other conditions (Hoover and Fraumeni, 1975; Kinlen, 1979). Recently strong associations have been noted between leiomyosarcoma and AIDS, mostly in children (Mbulaiteye et al., 2003). After non-Hodgkin lymphoma, the second leading cancer in children is soft tissue sarcoma. In fact, Biggar et al. (2000) reported a risk of more than 1900 for leiomyosarcoma in children with AIDS in Uganda. Within 2 years of an AIDS diagnosis in all ages (Frisch et al., 2001), the risk for soft tissue sarcoma is 3.6 (2.5–5.3), which the authors see as evidence of immunosuppression.

Tumor suppressor genes are important in the pathogenesis of soft tissue sarcoma—most prominent among them are p53 and Rb, the hereditary retinoblastoma gene. Germline p53 mutations in families with multiple tumors are diagnostic of Li-Fraumeni syndrome (Malkin et al., 1990). In the first place, the Li-Fraumeni syndrome was first recognized in four families with an autosomal dominant pattern of STS, breast cancer, and other tumors among young individuals (Li and Fraumeni, 1969). Families have typically been identified through an increased number of both bone and soft tissue sarcoma cases in conjunction with a specific constellation of other tumors, such as breast, brain, adrenal cortical carcinoma, Wilms’ tumor, and leukemia. Among Li-Fraumeni families, approximately 20% develop soft tissue sarcoma and 28% breast cancer. In those families with p53 mutations, tumors tend to appear in very young individuals—85% of the sarcomas (including bone) occurred under the age of 20 and 79% of the breast cancer under the age of 40 (Varley et al., 1997). Clusters of families occur that appear to be LiFraumeni syndrome but do not have germline mutations and thus suggest other syndromes or underlying genetic etiology (Li et al., 1997; Bell, 1999). The range of cancers has been enlarged to include a number of other tumors associated with the Li-Fraumeni syndrome less often, such as stomach, ovary, colorectal, lymphoma, melanoma, endometrial, thyroid, pancreas, prostate, and cervix (Nichols et al., 2001). In addition, individuals within such families tend to develop multiple primary tumors (Strong et al., 1987; Hisada et al., 1998), particularly soft tissue sarcomas arising in a radiation field. The identification of the Li-Fraumeni syndrome has been critically important for understanding the etiology of soft tissue sarcoma and cancer in general. However, germline p53 mutations do not account for a large proportion of soft tissue sarcomas, having been found in approximately 4% of unselected soft tissue sarcoma cases (Malkin, 1998). Both point mutations and deletions and insertions in p53 have been identified in soft tissue tumors themselves. In MFH, leiomyosarcoma, liposarcoma, and rhabdomyosarcoma, abnormalities in p53 are found in 20%–30% of the cases (Meltzer, 1995). Studies of families of children with STS have demonstrated the probable genetic basis of such early cancer incidence (Strong, 1987, 1989; Birch, 1990; Hwang et al., 2003). In individuals with hereditary retinoblastoma, there is a significantly increased risk for second cancers, particularly sarcomas and osteosarcomas (Sanders, 1989; Fontanesi et al., 1995). Germline mutations in the retinoblastoma gene (Rb-1) on chromosome 13q14 seem to be responsible for hereditary retinoblastoma as well as sarcomas (Hansen et al., 1985). Retinoblastoma is typically treated with radiation and there is a dose-response relationship between radiation dose and sarcoma risk, evident at doses about 5 Gy (Wong, 1997), and increasing to a relative risk of 10.7 at doses of 60 Gy or higher. Mutations in the Rb gene have been found in sporadic tumors as well (Cance et al., 1990; Zahm and Fraumeni, 1997). P53 and RB1 abnormalities often occur together (Stratton et al., 1990) indicating that coincident inactivation of more than one tumor suppressor gene may be required for tumor development. Interestingly, the type of p53 mutation in a tumor seems to be related to prognosis; non-frameshift mutations have significantly poorer outcome (Taubert et al., 1996). Other genetic syndromes have been associated with soft tissue sarcoma. Individuals with NF1 (the neurofibromatosis type 1 gene) have a 7%–14% lifetime risk of developing a sarcoma, usually neu-

Immunosuppression

Endogenous Hormonal Factors Hormonal factors may play a role in the etiology of soft tissue sarcoma. Fioretti et al. (2000) studied 104 women with incident STS and 505 controls admitted to the same hospitals for acute, nonneoplastic, nongynecologic, and non-immune–related conditions. They found late age at first pregnancy and birth increased risk for STS but there was no association for menstrual cycle pattern, parity, age at menopause, or abortion history. Molife et al. (2001) examined the role of gender in survival from solid tumors including sarcoma. A strong survival advantage, 42%, was seen for women with sarcoma. The authors suggested that this advantage is most often present in earlystage disease but lost in advanced disease and possibly over time. The mechanism by which female gender might improve survival in earlystage disease should be explored further. Additional clues to a hormonal etiology in rhabdomyosarcoma are noted in studies by Ghali et al. (1992), where childhood rhabdomyosarcoma was associated with a maternal history of stillbirth and by dos Santos Silva and Swerdlow (1993), who noted that rhabdomyosarcoma increases in incidence during puberty.

Body Mass Index Several recent studies have estimated the risk for developing soft tissue sarcoma associated with excess weight. Zahm et al. (1989) reported increasing risk for sarcoma with increasing body weight; with a referent group of 141 pounds, the risk increased with increasing weight to an odds ratio of 2.0 (p for trend = 0.03) for those who weighed more than 200 pounds. In Italy (Tavani et al., 2000), the risk for sarcoma was also increased among those with a BMI of >30 kg/m2 compared with those with a BMI = 20 kg/m2 (males OR = 3.5, 95% CI: 1.1–11.6, females OR = 3.3, 95% CI: 1.3–8.4). Similar results were obtained in a population-based study (Schwartz et al., 1996) where the women in the highest quantile of body mass index (>27.5 kg/m2) were at increased risk for leiomyosarcoma (OR = 2.5, 95% CI: 1.1–5.7), mixed mullerian tumors (OR = 2.9, 95% CI: 1.3, 6.7), and stromal sarcoma (OR = 3.5, 95% CI: 1.1, 10.9).

Other Risk Factors Certain medical conditions and medications seem to be associated with risk for STS, particularly some viral infections, such as herpes

970

PART IV: CANCER BY TISSUE OF ORIGIN

zoster (OR = 2.4, 1.1–5.3), chicken pox (OR—2.2, 1.2–4.3), and mumps (OR = 2.0, 1.1–3.9) (Serraino et al., 1991; Franceschi and Serraino, 1992). Previously, Zahm et al. (1989) found no association with history of conditions such as allergies, chicken pox, chloracne, diabetes, eczema, hepatitis, hypertension, kidney stones, or trauma.

Chronic Repair Processes Froehner and Wirth (2001) have suggested that chronic repair processes may increase risk for STS. A decreased DNA repair capacity was associated with the development of soft tissue sarcoma in a small case-control study (Berwick et al., 2000). Other evidence for repair associations is seen in studies using animal models that developed MFH after implants with a wide variety of substances (Rhomberg, 1998; Kirkpatrick et al., 2000). Several early reports found sarcomas in association with chronic skin ulcers (Routh et al., 1985; Fletcher, 1987), but a recent ecological study did not find evidence that breast implantation was associated with breast sarcoma (Engel et al., 1995). Although trauma has often been suggested as causal in STS, no studies have found positive associations.

PATHOGENESIS Although our knowledge of the genome has increased exponentially and technological advances have been impressive, tumor biologists are still limited in their ability to study alterations in rare tumors. A major problem for epidemiology is how to study the etiology of a tumor that has not yet been well defined histologically. Genome-wide expression profiling may hold some important clues as to the pathogenesis of soft tissue sarcoma. Studies by Nielsen et al. (2002), Lee et al. (2003), and Allander et al. (2001) are being expanded to include more samples and more replicates. It is now abundantly clear that GIST tumors are remarkably distinct and present a uniform expression profile (Allander et al., 2001). Because a therapy (STI571 or Gleevec) for those GIST tumors with a mutated proto-oncogene, KIT, has been extremely successful, molecular characterization of soft tissue sarcomas will probably continue to be high priority research. Gene expression profiling studies (Nielson, 2002) have found distinctive subsets of GIST and synovial sarcoma as well as a morphologically heterogeneous group of tumors with fibrous and histiocytic features. It should be no surprise that some of the genes identified overlap with those genes identified in breast cancer because individuals with Li-Fraumeni syndrome have a high risk for breast cancer and the families of STS probands without the syndrome who have a high risk for breast cancer. MFH, liposarcoma, and some leiomyosarcomas were characterized by genes expressed by macrophages, genes of the interferon responsive cluster, genes associated with other inflammatory processes, genes for collagen, collagen metabolism, and constituents of the extracellular matrix and angiogenesis. Overall these overlap with stromal/fibroblast and endothelial gene clusters identified in breast tumors. Lee et al. (2003) report divergent gene expression findings from Nielsen et al. (2002). The lack of congruence points out the need for repeated and larger studies as well as continued examination of methods for gene expression profiling before we come to a solid understanding of the molecular classes of STS or its pathogenesis. The recent discovery that GIST tumors with c-Kit mutations are responsive to therapy has been a paradigm for the potentially exciting role for molecular pathology to elucidate mechanisms of carcinogenesis as well as additional targets for therapy.

PREVENTIVE MEASURES Unfortunately, preventive measures cannot yet be undertaken except among the high-risk groups for STS that include Li-Fraumeni families, childhood survivors of cancer, and individuals who have received

more that 5 Gy radiations. These groups must continue to be followed clinically for the development of additional cancers.

FUTURE DIRECTIONS The future looks optimistic for the incorporation of molecular techniques in epidemiologic studies of STS. Once the tumor subtypes can be identified consistently, then patients can be appropriately categorized for etiologic research. It is possible that tumor-specific mutations will be an adjunct to exposure measures, pathologic characterization, and genetic susceptibility factors that will more accurately and reproducibly define the pathways to tumor development and thus provide targets for preventive measures as well as therapeutic intervention. References Abu-Rustum NR, Curtin JP, Burt M, Jones WB. 1997. Regression of uterine low-grade smooth-muscle tumors metastic to the lung after oophorectomy. Obstet Gynecol 89:850–852. Acheson ED, Barnes HR, Gardner MJ, et al. 1984. Formaldehyde in the British chemical industry: An occupational cohort study. Lancet 1:611–616. Allander SV, Nipponen NN, Ringer M, et al. 2001. Gastrointestinal stromal tumors with KIT mutations exhibit a remarkably homogenous gene expression profile. Cancer Res 61:8624–8628. Aylward LL, Hays SM. 2002. Temporal trends in human TCDD body burden: Decreases over three decades and implications for exposure levels. J Exposure Anal Environ Epidemiol 12:319–328. Bas Bueno de Mesquita H, Doornbos G, van der Kuip DAM, Kogevinas M, Winkelmann R. 1993. Occupational exposure to phenoxy herbicides and chlorophenols and cancer mortality in the Netherlands. Am J Indust Med 23:289–300. Becher H, Flesch-Janys D, Kauppinen T, et al. 1996. Cancer mortality in German male workers exposed to phenoxy herbicides and dioxins. Cancer Causes Control 7:312–321. Bell DW, Varley JM, Szydlo TE, et al. 1999. Heterozygous germ line hCHK2 mutations in Li-Fraumeni syndrome. Science 286:2528–2531. Bennett KE, Howell A, Evans DGR, Birch JM. 2002. A follow-up study of breast and other cancers in families of an unselected series of breast cancer patients. Br J Cancer 86:718–722. Beral V, Peterman TA, Berkelman RC, Jaffe HW. 1990. Kaposi’s sarcoma among persons with AIDS: A sexually transmitted infection? Lancet 335:123–128. Berking C, Brady MS. 1997. Cutaneous melanoma in patients with sarcoma. Cancer 79:843–846. Bertazzi PA, Zocchetti C, Pesatori AC, et al. 1989. Ten-year mortality study of the population involved in the Seveso incident in 1976. Am J Epidemiol 129:1187–1199. Bertazzi PA, Zocchetti C, Guercilena S, et al. 1997. Dioxin exposure and cancer risk: A 15-year mortality study after the “Seveso accident”. Epidemiology 8:646–652. Bertazzi PA, Consonni D, Bachetti S, et al. 2001. Health effects of dioxin exposure: A 20-year mortality study. Am J Epidemiol 153:1031–1044. Berwick M, Song Y, Jordan R, Brady MS, Orlow I. 2001. Mutagen sensitivity as an indicator of soft tissue sarcoma risk. Environ Mol Mutagen 38:223–226. Bhatia S, Sather HN, Pabustan OB, Trigg ME, Gaynon PS, Robison LL. 2001. Low incidence of second neoplasms among children diagnosed with acute lymphoblastic leukemia after 1983. Blood 99:4257–4264. Biermann JS, Baker LH. 1999. Soft tissue sarcoma: Opening the door to advances in diagnosis and treatment—editorial. Cancer 85:2497– 2498. Biggar BJ, Frisch M, Goedert JJ. 2000. Risk of cancer in children with AIDS. JAMA 284:205–209. Birch JM, Hartley AL, Marsden HB, Harris M, Swindell R. 1984. Excess risk of breast cancer in the mothers of children with soft tissue sarcomas. Br J Cancer 49:325–331. Birch JM, Hartley AL, Blair V, et al. 1990. Cancer in the families of children with soft tissue sarcoma. Cancer 66:2239–2248. Blair A, Stewart P, O’Berg M, et al. 1986. Mortality among industrial workers exposed to formaldehyde. J Natl Cancer Inst 57:487–492. Blanchard DK, Reynolds C, Grant CS, Farley DR, Donohue JH. 2002. Radiation-induced breast sarcoma. Am J Surgery 184:356–358. Boffeta P, Gridley R, Gustavsson P, et al. 2000. Employment as butcher and cancer risk in a record-linkage study from Sweden. Cancer Causes Control 11:627–633.

Soft Tissue Sarcoma Boroush M, Gough M. 1994. Can cohort studies detect any human cancer excess that may result from exposure to dioxin? Maybe. Regul Toxicol Pharmacol 20:198–210. Brady MS, Gaynor JJ, Brennan MF. 1992. Radiation-associated sarcoma of bone and soft tissue. Arch Surg 127:1379–1385. Brady MS, Garfein CF, Petrek JA, Brennan MF. 1994. Post-treatment sarcoma in breast cancer patients. Ann Surg Oncol 1:66–72. Breslin P, Kang HK, Lee Y, et al. 1988. Proportionate mortality study of US Army and US Marine Corps veterans of the Vietnam war. J Occup Med 30:412–419. Burke E, Li FP, Janov AJ, Batter S, Grier H, Goorin A. 1991. Cancer in relatives of survivors of childhood sarcoma. Cancer 67:1476–1469. Busam KJ, Fletcher CDM. 1997. The clinical role of molecular genetics in soft tissue tumor pathology. Cancer Metastasis Rev 16:207–227. Cance WG, Brennan MF, Dudas ME, Huang CM, Cordon-Cardo C. 1990. Altered expression of the retinoblastoma gene product in human sarcomas. N Engl J Med 323:1457–1462. Chang Y, Cesuman E, Pessin MS, et al. 1994. Identification of herpesvirus-like DNA sequences in AIDS-associated Kaposi’s sarcoma. Science 266: 1865–1869. Coggon D, Pannett B, Winter P, Acheson ED, Bonsall J. 1986. Mortality of workers exposed to 2 methyl-4 chlorophenoxyacetic acid. Scand J Work Environ Health 12:448–454. Coggon D, Pannett B, Winter P. 1991. Mortality and incidence of cancer at four factories making phenoxy herbicides. Br J of Ind Med 48:173–178. Cook RR. 1981. Dioxin, chloracne, and soft tissue sarcoma. Lancet 1:618–619. Cook RR, Townsend JC, Ott MG, et al. 1989. Mortality experience of employees exposed to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). J Occup Med 22:530–532. Costani G, Rabitti P, Mambrini A, Bai E, Berrino F. 2000. Soft tissue sarcomas in the general population living near a chemical plant in Northern Italy. Tumori 5:381–383. Cozen W, Bernstein L, Wang F, Press MF, Mack TM. 1999. The risk of angiosarcoma following primary breast cancer. Br J Cancer 81:532–536. Creech JL Jr., Johnson MN. 1974. Angiosarcoma of liver in the manufacture of polyvinyl chloride. J Occup Med 16:150–151. Curtis RE, Hoover RN, Kleinerman RA, et al. 1985. Second cancer following cancer of the female genital system. Natl Cancer Inst Monogr 68: 113–137. dos Santos Silva I, Swerdlow AJ. 1993. Sex differences in the risks of hormonedependent cancer. Am J Epidemiol 130:10–28. Engel A, Lamm SH, Lai SH. 1995. Human breast sarcoma and human breast implantation: A time trend analysis based on SEER data (1973–1990). J Clin Epidemiol 48:539–544. Enzinger FM, Weiss SW. 1996. Soft Tissue Tumors. 3rd ed. St. Louis, MO: Mosby—Year Book, Inc. Ericksson M, Hardell L, Berg NO, et al. 1981. Soft tissue sarcoma and exposure to chemical substances: A case-referent study. Br J Ind Med 38:27–33. Ericksson M, Hardell L, Adami H-O. 1990. Exposure to dioxins as a risk factor for soft tissue sarcoma: A population-based case-control study. J Natl Cancer Inst 82:486–490. Falk H, Telles NC, Ishak KG, Thomas LB, Popper H. 1979a. Epidemiology of thorotrast-induced hepatic angiosarcoma in the United States. Environ Res 18:65–73. Falk H, Thomas LB, Popper H, Ishak KG. 1979b. Hepatic angiosarcomas associated with androgenic-anabolic steroids. Lancet 2:1120–1123. Fingerhut MA, Halperin WE, Honchar PA, Smith AB, Groth DH, Russell WO. 1984. An evaluation of reports of dioxin exposure and soft tissue sarcoma pathology among chemical workers in the United States. Scand J Work Environ Health 10:299–303. Fingerhut MA, Halperin WE, Marlow DA, et al. 1991. Cancer mortality in workers exposed to 2,3,7,8-tetrachlorodibenzo-p-dioxin. N Engl J Med 324:212–218. Fioretti F, Tavani A, Gallus S, Negri E, Franceschi S, La Vecchia C. 2000. Menstrual and reproductive factors and risk of soft tissue sarcomas. Cancer 88:786–789. Fleming LE, Bean JA, Rudolph M, Hamilton K. 1999. Mortality in a cohort of licensed pesticide applicators in Florida. Occup Environ Med 56:14–21. Flesch-Janys D, Berger J, Gurn P, et al. 1995. Exposure to polychlorinated dioxins and furans (PCDD/F) and mortality in a cohort of workers from a herbicide-producing plant in Hamburg, Federal Republic of Germany. Am J Epidemiol 142:1165–1175. Fletcher CDM. 1987. Soft tissue sarcomas apparently arising in chronic tropical ulcers. Histopathology 11:501–510. Fletcher CDM, Fletcher JA, Cin PD, Ladanyi M, woodruff JM. 2001. Diagnostic gold standard for soft-tissue tumours: Morphology or molecular genetics? Histopathology 39:100–101.

971

Fontanesi J, Parkham DM, Pratt C, Meyer D. 1995. Second malignant neoplasms in children with retinoblastoma: The St. Jude Children’s Research Hospital experience. Opthalmic Genet 16:105–108. Franceschi S, Serraino D. 1992. Risk factors for adult soft tissue sarcoma in northern Italy. Ann Oncol (Suppl22) 2:S85–S88. Frisch M, Biggar RJ, Engels EA, Joedert JJ for the AIDS-Cancer Match Registry Study Group. 2001. Association of cancer with AIDS-related immunosuppression in adults. JAMA 285:1736–1745. Froehner M, Wirth MP. 2001. Etiologic factors in soft tissue sarcoma. Onkologie 24:139–142. Gallagher RP, Threlfall WJ. 1984. Cancer and occupational exposure to chlorophenols. Lancet 2:48. Gambini GF, Pira E, Piolatto PG, Negri E, Manotvani C. 1997. Cancer mortality among rice growers in Novara Province, northern Italy. Am J Ind Med 31:435–441. Gebauer C. 1982. Primary pulmonary sarcomas. Etiology, clinical assessment and prognosis with a comparison to pulmonary carcinomas—A review of 41 cases and 394 other cases of the literature. Jpn J Surg 12:148– 159. Ghali MH, Yoo K-Y, Flannery JT, et al. 1992. Association between childhood rhabdomyosarcoma and maternal history of stillbirths. Int J Cancer 50:148–159. Gonzales CA, Kogevinas M, Gadea E, et al. 2001. Increase of dioxin blood levels over the last 4 years in the general population in Spain. Epidemiology 12:365. Goun B, Kuller L. 1988. Vietnam military service and risk of selected cancers. Am J Epidemiol 128:901. Granovsky MO, Mueller BU, Nicholson HS, Rosenberg PS, Rabkin CS. 1998. Cancer in human immunodeficiency virus-infected children: A case series from the Children’s Cancer Group and the National Cancer Institute. J Clin Oncol 16:1729–1735. Green LM. 1991. A cohort mortality study of forestry workers exposed to phenoxy acid herbicides. Br J Ind Med 48:234–238. Greenwald P, Kovasznay B, Collins DN, Therriault G. 1984. Sarcomas of soft tissues after Vietnam service. J Natl Cancer Inst 73:1107–1109. Gurney JG, Young JL Jr., Roffers SD, Smith MA, Bunin GR. 1999. Soft tissue sarcomas. In: Ries LAG, Smith MA, Gurney JG, Linet M, Tamra T, Young JL, Bunin GR, eds. Cancer Incidence and Survival among Children and Adolescents: United States SEER Program 1975–1995. Bethesda, MD: National Cancer Institute, SEER Program, NIH Pub. No. 99-4649. Hansen MF, Koufos A, Gallie BL, et al. 1985. Osteosarcoma and retinoblastoma: A shared chromosomal mechanism revealing recessive predisposition. Proc Natl Acad Sci USA 82:6216–6220. Hardell L. 1977. Malignant mesenchymal tumours and exposure to phenoxy acids—a clinical observation. Lakartidningen 74:2753–2754. Hardell L, Sandstrom A. 1979. Case-control study: Soft-tissue sarcomas and exposure to phenoxyacetic acids or chlorophenols. Br J Cancer 39:711–717. Hardell L, Eriksson M. 1988. The association between soft tissue sarcomas and exposure to phenoxyacetic acids: A new case-referent study. Cancer 62:652–656. Harlow BL, Weiss NS, Lofton S. 1986. The epidemiology of sarcomas of the uterus. J Natl Cancer Inst 76:399–402. Harris M, Hartley AL, Blair V, et al. 1991. Sarcomas in North West England: I Histopathological peer review. Br J Cancer 64:315–320. Hartley AL, Birch JM, Blair V. 1990. Malignant disease in the mothers of a population-based series of young adults with bone and soft tissue sarcomas. Br J Cancer 63:416–419. Harvey EB, Brinton LA. 1985. Second cancer following cancer of the breast in Connecticut, 1935–82. Natl Cancer Inst Monogr 68:99–112. Hemminki K, Li X. 2001. A population-based study of familial soft tissue tumors. J Clin Epidemiol 54:411–416. Hibshoosh H, Lattes R. 1997. Immunohistochemical and molecular genetic approaches to soft tissue tumor diagnosis: A primer. Semin Oncol 24:515–525. Hisada M, Garber JE, Fung CY, et al. 1998. Multiple primary cancers in families with Li-Fraumeni syndrome. J Natl Cancer Inst 90:606–611. Hoar SK, Blair A, Holmes FF, et al. 1986. Agricultural herbicide use and risk of lymphoma and soft-tissue sarcoma. JAMA 256:1141–1147. Hooiveld M, Heederik DJJ, Kogevinas M, et al. 1998. Second follow-up of a Dutch cohort occupationally exposed to phenoxy herbicides, chlorophenols, and contaminants. Am J Epidemiol 147:891–901. Hoover R, Fraumeni JF Jr. 1973. Risk of cancer in renal transplant recipients. Lancet 2:55–57. Hoppin JA, Tolbert PE, Herrick RF, et al. 1998. Occupational chlorophenol exposure and soft tissue sarcoma risk among men aged 30–60 years. Am J Epidemiol 148:693–701.

972

PART IV: CANCER BY TISSUE OF ORIGIN

Hoppin JA, Tolbert PE, Flanders WD, et al. 1999. Occupational risk factors for sarcoma subtypes. Epidemiology 10:300–306. Howe HL, Wingo PA, Thun MJ, et al. 2001. Annual report to the nation on the status of cancer (1973 through 1998), featuring cancers with recent increasing trends. J Natl cancer Inst 93:824–842. Huang J, Mackillop WJ. 2001. Increased risk of soft tissue sarcoma after radiotherapy in women with breast carcinoma. Cancer 92:172–180. Hussain SP, Harris CC. 1998. Molecular epidemiology of human cancer: Contribution of mutation spectra studies of tumor suppressor genes. Cancer Res 58:4023–4037. Hwang S-J, Lozano G, Amos CI, Strong LC. 2003. Germline p53 mutations in a cohort with childhood sarcoma: Sex differences in cancer risk. Am J Hum Genet 72:975–983. Jemal A, Murray T, Samuels A, Ghafoor A, Ward E, Thun MJ. 2003. Cancer Statistics, 2003. CA Cancer J Clin 53:5–26. Johnson FE, Kugler AM, Brown SM. 1981. Soft tissue sarcomas and chlorinated phenols. Lancet 2:40. Jones PA, Buckley JD, Henderson BE, Ross RK, Pike MC. 1991. From gene to carcinogen: A rapidly evolving field in molecular epidemiology. Cancer Res 51:3617–3620. Kang HK, Weatherbee L, Breslin PP, Lee Y, Shepard BM. 1986. Soft tissue sarcomas and military service in Vietnam: A case comparison group analysis of hospital patients. J Occup Med 28:1215–1218. Kang H, Einziger F, Breslin P, Feil M, Lee Y, Shepard B. 1987. Soft-tissue sarcoma and military service in Vietnam: A case-control study. J Natl Cancer Inst 79:693–699. Karlsson P, Holmberg E, Samuelsson A, Johansson K-A, Wallgren A. 1998. Soft tissue sarcoma after treatment for breast cancer—a Swedish population-based study. Eur J Cancer 34:2068–2067. Kato I, Kido C. 1987. Increased risk of death in thorotrast-exposed patients during the late follow up period. Jpn J Cancer Res 78:1187–1192. Kattan MW, Leung DHY, Brennan MF. 2002. Postoperative nomogram for 12-year sarcoma-specific death. J Clin Oncol 20:791–796. Keohan ML, Antman KH. 1999. Sarcomas. In: Neugat AI, Meadows AT, Robinson E, eds. Multiple Primary Cancers Philadelphia: Lippincott Williams & Williams, pp. 417–430. Kim JH, Chu FC, Woodard HQ, Melamed MR, Huvos A, Cantin J. 1978. Radiation-induced soft-tissue and bone sarcoma. Radiology 129:501– 508. King DT, Duffy DM, Hirose FM, et al. 1979. Lymphangiosarcoma arising from lymphangioma circumscriptum. Arch Dermatol 115:959–972. Kinlen LJ, Sheil AGR, Peto J, et al. 1979. Collaborative United Kingdom– Australian study of cancer in patients treated with immunosuppressive drugs. BMJ 2:1461–1466. Kirkpatrick CJ, Alves A, Kohler H, et al. 2000. Biomaterial-induced sarcoma: A novel model tos tudy preneoplastic change. Am J Pathol 156:1455– 1467. Kogan MD, Clapp RW. 1988. Soft tissue sarcoma mortality among Vietnam veterans in Massachusetts, 1972 to 1983. Int J Epidemiol 17:39–43. Kogevinas M, Kauppinen T, Winkelmann R, et al. 1995. Soft tissue sarcoma and non-Hodgkin’s lymphoma in workers exposed to phenoxy herbicides, chlorophenols, and dioxins: Two nested case-control studies. Epidemiology 6:396–402. Kogevinas M BH, Benn T, Bertazzi PA, et al. 1997. Cancer mortality in workers exposed to phenoxy herbicides, chlorophenols, and dioxins. Am J Epidemiol 145:1061–1075. Ladanyi M. 1995. The emerging molecular genetics of sarcoma translocations. Diagn Mol Pathol 4:162–173. Lagrange JL, Ramaioli A, Chateau MC, et al. 2000. Sarcoma after radiation therapy: Retrospective multi-institutional study of 80 histologically confirmed cases. Radiology 216:197–205. Landi MT, Needham LL, Lucier G, et al. 1997. Concentrations of dioxin 20 years after Seveso. Lancet 349:1811. Lawrence CE, Reilla AA, Quickenton P, et al. 1985. Mortality patterns of New York State Vietnam veterans. Am J Public Health 75:277–279. Lee Y-F, John M, Edwards S, et al. 2003. Molecular classification of synovial sarcomas, leiomyosarcomas and malignant fibrous histiocytomas by gene expression profiling. Br J Cancer 88:510–515. Levi F, La Vecchia C, Randimbison L, et al. 1999. Descriptive epidemiology of soft tissue sarcomas in Vaud, Switzerland. Eur J Cancer 35:1711–1716. Lewis JJ, Leung D, Woodruff JM, Brennan MF. 1998. Retroperitoneal softtissue sarcoma: Analysis of 500 patients treated and followed at a single institution. Ann Surg 228:355–365. Li FP, Fraumeni JF, Jr. 1969. Soft-tissue sarcomas, breast cancer, and other neoplasms: A familial syndrome? Ann Intern Med 71:747–752. Li FP, Dreyfuss M, Russell TL, Verselis SJ, Hutchinson RJ, Fraumeni JF, Jr. 1997. Molecular epidemiology study of a suspected community cluster of childhood cancers. Med Ped Oncol 28:243–247.

Locker GY, Doroshow JH, Swelling LA, et al. 1979. The clinical features of hepatic angiosarcoma: A report of four cases and a review of the English literature. Medicine 58:43–64. Lynge E, Storm HH, Jensen OM. 1987. The evaluation of trends in soft tissue sarcoma according to diagnostic criteria and consumption of phenoxy herbicides. Cancer 60:1896–1901. Lynge E. 1993. Cancer in phenoxy herbicide manufacturing workers in Denmark, 1947–87—an update. Cancer Causes Control 4:261–272. Mahour GH, Soule EH, Mills SD, et al. 1967. Rhabdomyosarcoma in infants and children: A clinicopathologic study of 75 cases. J Pediatr Surg 2:402–409. Malkin D, Li F, Strong LC, et al. 1990. Germ line p53 mutations in a familial syndrome of breast cancer, sarcomas, and other neoplasms. Science 250:1233–1238. Malkin D. 1998. Li-Fraumeni syndrome. In: Vogelstein B, Kinzler KW, eds. The Genetic Basis of Human Cancer. New York: McGraw-Hill, pp. 393–408. Manz A, Berger J, Dywer JH, et al. 1991. Cancer mortality among workers in chemical plant contaminated with dioxin. Lancet 338:959–964. Marion MJ, DeVivo I, Smith S, Luo JC, Brandt-Rauf PW. 1996. The molecular epidemiology of occupational carcinogenesis in vinyl chloride exposed workers. Int Arch Occup Environ Health 68:394–398. Mays CW. 1988. Alpha-particle-induced cancer in humans. Health Phys 55:637–652. Mbulaiteye SM, Biggar RJ, Goedert JJ, Engels EA. 2003a. Immune deficiency and risk for malignancy among persons with AIDS. J Acquire Immune Defic Syndr 32:527–533. Mbulaiteye SM, Parkin DM, Rabkin CS. 2003b. Epidemiology of AIDS-related malignancies: An international perspective. Hematol Oncol Clin North Am 17:673–696. McClain KL, et al. 1995. Association of Epstein-Barr virus with leiomyosarcomas in children with AIDS. N Engl J Med 332:55–57. McClain KL, Joshi VV, Murphy SB. 1996. Cancers in children with HIV infection. Hematol Oncol Clin North Am 20:1189–1201. Meltzer P. 1996. Molecular biology of soft tissue tumors. In: Eingziger FM, Weiss SW, eds. Soft Tissue Tumors 3rd ed., New York: Mosby, pp. 89–104. Merimsky O, Kollender Y, Issakov J, et al. 2001. Multiple primary malignancies in association with soft-tissue sarcomas. Cancer 91: 1363–1371. Mikovits J, Ruscetti F, Zhu W, Bagni R, Dorjsuren D, Shoemaker R. 2001. Potential cellular signatures of viral infections in human hematopoietic cells. Dis Markers 17:173–178. Milham S Jr. 1982. herbicides, occupation, and cancer. Lancet 1:1464–1465. Molife R, Lorigan P, MacNeil S. 2001. Gender and survival in malignant tumours. Cancer Treatment Rev 27:201–209. Mulvihill JJ. 1975. Congenital and genetic diseases. In: Fraumeni JF Jr., ed. Persons at High Risk of Cancer: An Approach to Cancer Etiology and Control. New York: Academic Press, pp. 3–39. NCI. 2002. National Cancer Institute. Cancer Facts. Available at: cis.nci.nih.gov/fact/6_12.htm. Neglia JP, Friedman DL, Yasui Y, et al. 2001. Second malignant neoplasms in five-year survivors of childhood cancer: Childhood cancer survivor study. J Natl Cancer Inst 93:618—629. Neuberger M, Grossgut R, Gyimothy J, et al. 2000. Dioxin contamination of feed and food. Lancet 355:1883. Nichols KE, Malkin D, Garber JE, et al. 2001. Germ-line p53 mutations predispose to a wide spectrum of early-onset cancers. Cancer Epidemiol Biomarkers Prev 10:83–87. Nielsen TO, West RB, Linn SC, et al. 2002. Molecular characterization of soft tissue tumours: A gene expression study. Lancet 359:1301– 1307. Nijhuis PHA, Schaapveld M, Otter R, et al. 1999. Epidemiological aspects of soft tissue sarcomas (STS)—consequences for the design of clinical STS trials. Eur J Cancer 35:1705–1710. Ott MG, Zober A. 1996. Cause specific mortality and cancer incidence among employees exposed to 2,3,7,8-TCDD after a 1953 reactor accident. J Occup Environ Med 53:606–612. Pack GT, Pierson JC. 1954. Liposarcoma: A study of 105 cases. Surgery 36:687–712. Parkin DM, Whelan SL, Ferlay J, Teppo L, Thomas D, eds. 2002a. Cancer incidence in five continents. Volume VIII. Lyon, France: IARC Scientific Publication No. 155. IARC. Parkin DM, Ferlay J, Hamdi-Cherif M, et al., eds. 2002b. Cancer in Africa: Epidemiology and Prevention. Lyon, France: IARC Scientific Publication No. 153. IARC. Pauk J, Huang ML, Brudie SJ, et al. 2000. Mucosal shedding of human herpesvirus 8 in men. N Engl J Med 343:1369–1377.

Soft Tissue Sarcoma Paulose-Murphy M, Ha NK, Xiang C, et al. 2001. Transcription program of human herpesvirus 8 (Kaposi’s sarcoma-associated herpesvirus). J Virol 75:4843–4853. Patel SG, See ACH, Williamson PA, Archer DJ, Rhys Evans PH. 1999. Radiation induced sarcoma of the head and neck. Head Neck 21:346– 354. Paustenbach DJ. 2002. The US EPA science advisory board evaluation (2001) of the EPA dioxin reassessment. Regul Toxicol Pharmacol 36:211–219. Pearce N, Smith AH, Reif JS. 1988. Increased risks of soft tissue sarcoma, malignant lymphoma, and acute myeloid leukemia in abattoir workers. Am J Ind Med 14:63–72. Pfeifer JD, Hill DA, O’Sullivan JM, Dehner LP. 2000. Diagnostic gold standard for soft tissue tumors: Morphology or molecular genetics? Histopathology 37:485–500. Piacitelli L, Marlow D, Fingerhut M, Steeland K, Sweeney MH. 2000. A retrospective job exposure matrix for estimating exposure to 2,3,7,8tetrachlorodibenzo-p-dioxin. Am J Ind Med 38:28–39. Pohl HR, Hicks HE, Jones DE, Hansen H, DeRosa CT. 2002. Public health perspectives on dioxin risks: Two decades of evaluations. Hum Ecol Risk Assess 8: 233–250. Polednak AP. 1986. Incidence of soft-tissue cancers in blacks and whites in New York State. Int J Cancer 38:21–26. Popper H, Thomas LB, Telles NC, et al. 1978. Development of hepatic angiosarcoma in man induced by vinyl chloride, thorotrast, and arsenic. Am J Pathol 92:349–376. Puntoni R, Merlo F, Fini A, et al. 1986. Soft tissue sarcomas in Seveso. Lancet 2:525. Reis LAG, Eisner MP, Kosary CL, et al. 2003. SEER Cancer Statistics Review, 1975–2000. Bethesda, MD: National Cancer Institute. Rhomberg W. 1998. Exposure to polymeric materials in vascular soft-tissue sarcomas. Int Arch Occup Environ Health 71:343–347. Riihimaki V AS, Hernberg S. 1982. Mortality of 2,4-dichlorophenoxyacetic acid and 2,4,5-trichlorophenoxyacetic acid herbicide applicators in Finland. Scand J Work Environ Health 8:37–42. Rix BA, Lynge E. 1997. Soft tissue sarcomas among female paper sorters in Denmark. Cancer Causes Control 8:259. Routh A, Hickman BT, Johnson WW. 1985. Malignant fibrous histiocytoma arising from chronic ulcer. Arch Dermatol 121:529–531. Salloum E, Doria R, Schubert W, et al. 1996. Second solid tumors in patients with Hodgkin’s disease cured after radiation or chemotherapy plus adjuvant low-dose radiation. J Clin Oncol 14:2435–2443. Samet L. 1996. Dioxin and cancer: The never-ending story [editorial]. Cancer Causes Control 7:302–304. Sanders BM, Jay M, Draper GJ, et al. 1989. Non-ocular cancer in relatives of retinoblastoma patients. Br J Cancer 60:358–365. Saracci R, Kogevina M, Bertazzi PA, et al. 1991. Cancer mortality in workers exposed to chlorophenoxy herbicides an chlorophenols. Lancet 338: 1027–1032. Sarma PR, Jacobs J. 1982. Thoracic soft-tissue sarcoma in Vietnam veterans exposed to Agent Orange. N Engl J Med 306:1109. Selected Cancers Cooperative Study Group. 1990. The association of selected cancers with service in the US military in Vietnam. II. Soft tissue and other sarcomas. Arch Intern Med 150:2485–2492. Serraino D, Franceschi S, Talamini R, et al. 1991. Non-occupational risk factors for adult soft-tissue sarcoma in northern Italy. Cancer Causes Control 2:157–164. Serraino D, Franceschi S, LaVecchia C, et al. 1992. Occupation and soft-tissue sarcoma in northeastern Italy. Cancer Causes Control 3:25–30. Schecter A, ed. 1994. Dioxins and Health. New York: Plenum Press. Schwartz SM, Weiss NS, Daling JR, et al. 1996. Exogenous sex hormone use, correlates of endogenous hormone levels, and the incidence of histologic types of sarcoma of the uterus. Cancer 77:717–724. Smith AH, Fisher DO, Pearce N, Teague CA. 1982. Do agricultural chemicals cause soft tissue sarcoma? Initial findings of a case-control study in New Zealand. Community Health Stud 6:114–119. Smith AH, Fisher DO, Giles HJ, Pearce NE. 1983. The New Zealand soft tissue sarcoma case-control study: Interview findings concerning phenoxyacetic acid exposure. Chemosphere 12:565–571. Smith AH, Pearce NE, Fisher DO, Giles HJ, Teague CA, Howard JK. 1984. Soft tissue sarcoma and exposure to phenoxyherbicides and chlorophenols in New Zealand. J Natl Cancer Inst 73:1111–1117. Smith MB, Xue H, Strong L, Takahashi H, Jaffe N, Ried H, Zietz H, Andrassy RJ. 1993. Forty-year experience with second malignancies after treatment of childhood cancer: Analysis of outcome following the development of the second malignancy. J Pediatr Surg 28:1342–1348. Steenland K, Piacitelli L, Deddens J, Fingerhut M, Chang LI. 1999. Cancer, heart disease, and diabetes in workers exposed to 2,3,7,8tetrachlorodibenzo-p-dioxin. J Natl Cancer Inst 91:779–786.

973

Stratton MR, Moss S, Warren W, et al. 1990. Mutation of the p53 gene in human soft tissue sarcomas: Association with abnormalities of the RBl gene. Oncogene 5:1297–1301. Strong LC, Stine M, Norsted TL. 1987. Cancer in survivors of childhood soft tissue sarcoma and their relatives. J Natl Cancer Inst 79:1213–1220. Strong LC, Williams WR, Ferrell RE, Tainsky MA. 1989. Genetic analysis of childhood sarcoma. Princess Takamatsu Symp 20:51–157. Suruda AJ, Ward EM, Fingerhut MA. 1993. Identification of soft tissue sarcoma deaths in cohorts. Epidemiology 4:14–19. Taghian A, de Vathaire F, Terrier P, et al. 1991. Long-term risk of sarcoma following radiation treatment for breast cancer. Int J Radiat Oncol Biol Phys 21:361–367. Taubert H, Meye A, Wurl P. 1996. Prognosis is correlated with p53 mutation type for soft tissue sarcoma patients. Cancer Res 56:4134–4136. Tavani A, Pregnolato A, Negri E et al. 1997. Diet and risk of lymphoid neoplasms and soft tissue sarcomas. Nutrition and Cancer 27:256– 260. Tavani A, Soler M, LaVecchia C, Negri E, Gallus S, Franceschi S. 2000. Body weight and risk of soft-tissue sarcomas. Br J Cancer 81:890–892. Tsujimoto M, Aozasa K, Ueda T, Morimura Y, Komatsubara Y, Doi T. 1988. Multivariate analysis for histologic prognostic factors in soft tissue sarcomas. Cancer 62:994–998. Usui M, Ishii S, Yamawaki S, Hirayama T. 1984. The occurrence of soft tissue sarcomas in three siblings with Werner’s syndrome. Cancer 54:2580– 2586. Van Kaick G, Lieberman D, Lorenz D, et al. 1983. Recent results of the German thorotrast study–Epidemiologic results and dose effect relationships in thorotrast patients. Health Phys 44:299–306. Varley JM MG, Thorncroft M, Santibanez-Koref MF. 1997. Germ-line mutations of TP53 in Li-Fraumeni families: An extended study of 39 families. Cancer Res 57:3245–3252. Varley JM, Evans DGR, Birch JM. 1997. Li-Fraumeni syndrome—a molecular and clinical review. Br J Cancer 76:1–14. Viel JF, Arveux P, Baverel J, Chan JY. 2000. Soft-tissue sarcomea and nonHodgkin’s lymphoma clusters around a municipal solid waste incinerator with high dioxin emission levels. Am J Epidemiol 152:13–19. Vineis P, Terracini B, Ciccone G, et al. 1987. Phenoxy herbicides and soft tissue sarcomas in female rice weeders: A population-based case-referent study. Scand J Work Environ Health 13:9–17. Vineis P, Zahm SH. 1988. Immunosuppressive effects of dioxin in the development of Kaposi’s sarcoma and non-Hodgkin’s lymphoma. Lancet 1:55. Vineis P. 1999. Occupational factors for sarcoma subtypes. Epidemiology 10:788. Vos JG, DeHeer C, Van Loveren. 1997/1998. Immunotoxic effects of TCDD and toxic equivalency factors. Teratogenesis Carcinog Mutagen 17: 275–284. Ward E BP, Anderson A, Colin D, et al. 2001. Update of the follow-up of mortality and cancer incidence among European workers employed in the vinyl chloride industry. Epidemiology 12:710–718. Wiklund K, Holm L-E. 1986. Soft tissue sarcoma risk in Swedish agricultural and forestry workers. J Natl Cancer Inst 76:229–234. Wiklund K, Dich J, Holm L-E. 1988. Soft tissue sarcoma risk in Swedish licensed pesticide applicators. J Occup Med 30:801–804. Wingren G, Frederickson M, Brage HN, et al. 1990. Soft tissue sarcoma and occupational exposures. Cancer 66:806–811. Wong FL, et al. 1997. Cancer incidence after retinoblastoma: Radiation dose and sarcoma risk. JAMA 278:1262–1267. Woods JS, Polissar L, Severson RK, et al. 1987. Soft tissue sarcoma and non-Hodgkin’s lymphoma in relation to phenoxyherbicide and chlorinated phenol exposure in Western Washington. J Natl Cancer Inst 78:899–910. Yap J, Chuba PJ, Thomas R, et al. 2002. Sarcoma as a second malignancy after treatment for breast cancer. Int J Radiat Oncol Biol Phys 52:1231–1237. Zack JA, Suskind RR. 1980. The mortality experience of workers exposed to tetrachlorodibenzodioxin in a trichlorophenol process accident. J Occup Med 22:11–14. Zahm SH, Blair A, Holmes FF, et al. 1989. A case-control study of soft-tissue sarcoma. Am J Epidemiol 130:665–674. Zahm SH, Blair A, Holmes FF, et al. 1988. A case-referent study of soft-tissue sarcoma and Hodgkin’s disease: Farming and insecticide use. Scand J Work Environ Health 14:224–230. Zahm SH, Heineman EF, Vaught JB. 1992. Soft tissue sarcoma and tobacco use: Data from a prospective cohort study of United States veterans. Cancer Causes Control 3:371–376. Zahm SH, Tucker MA, Fraumeni JF, Jr. 1996. Soft tissue sarcomas. In: Schottenfeld D, Fraumeni JF, Jr., eds. Cancer Epidemiology and Prevention. New York, Oxford University Press. pp. 984–999.

974

PART IV: CANCER BY TISSUE OF ORIGIN

Zahm SH, Fraumeni JF, Jr. 1997. The epidemiology of soft tissue sarcoma. Semin Oncol 24:504–514. Ziegler J, Newton R, Bourboulia D, et al. 2003. Risk factors for Kaposi’s sarcoma: A case-control study of HIV-seronegative people in Uganda. Int J Cancer 103:233–240. Zober A, Messerer P, Huber P. 1990. Thirty-four year mortality follow-up of BASF employees exposed to 2,3,7,8-TCDD after the 1953 accident. Int Arch Occup Environ Health 62:139–157.

Zoller MET, Rembeck B, Oden A, Samuelsson M, Angervall L. 1997. Malignant and benign tumors in patients with neurofibromatosis type 1 in a defined Swedish population. Cancer 79:2125–2131.

50

Thyroid Cancer ELAINE RON AND ARTHUR B. SCHNEIDER

C

ancer of the thyroid is relatively uncommon, although it is the most common malignancy of the endocrine system and is the eighth ranking cancer among women. The prognosis is extremely good for papillary and follicular carcinoma and extremely poor for anaplastic carcinoma. Other than ionizing radiation, the causes of thyroid cancer remain relatively obscure. The study of thyroid cancer is complicated by the fact that the histologic types vary in their demographic, etiologic, and survival patterns. This chapter reviews the epidemiology of thyroid cancer, taking histology into account whenever possible; however, the small number of cases included in many studies does not allow systematic analysis by histology.

HISTOPATHOLOGY Almost all (about 95%) thyroid cancers originate from cells derived from the follicular epithelium and are divided into papillary, follicular, and anaplastic forms. For clinical and etiologic purposes they often are divided into well-differentiated (papillary and follicular) and poorly differentiated (anaplastic) cancers. Medullary thyroid cancers also arise from epithelial cells, but from the calcitonin-producing parafollicular or C cells. Non-epithelial thyroid tumors are infrequent and are grouped together, except for malignant lymphomas, which form a separate class. The histopathologic classification system for thyroid cancer has evolved over time. The changes must be taken into account when analyzing time trends for thyroid cancer. Most importantly in 1988, the World Health Organization (WHO) revised its classification when it was recognized that nuclear features (particularly nuclear inclusions and nuclear folds) are more important than architectural patterns in classifying thyroid cancer (Hedinger et al., 1988, 1989). As a result, many cancers previously classified as follicular are now categorized as follicular variants of papillary cancer (LiVolsi and Asa, 1994). In addition, small cell carcinomas originally thought to be variants of anaplastic cancer are now recognized to be malignant lymphomas of the thyroid. The most common histologic type of thyroid cancer is, by far, the papillary type (Table 50–1). The proportion of differentiated thyroid cancers represented by the papillary type has increased over time to 70% or more in many countries including the United States. This results, in part, from the 1988 change in diagnostic criteria mentioned above. In addition, when iodine supplementation occurs in iodinedeficient regions, the proportion of papillary thyroid cancers often increases (Harach et al., 2002; Huszno et al., 2003; Lind et al., 2002). Finally, the increasing ability to diagnose smaller papillary cancers, primarily by ultrasonography, may contribute to this trend (Hegedus, 2001). In Geneva, Switzerland, where iodine supplementation has been practiced since the 1920s, a more recent shift toward the papillary form has been ascribed to the revised classification and to earlier diagnosis (Verkooijen et al., 2003). The relative frequency of follicular carcinoma ranges from about 10% to 40% depending on the iodine status of the region. Because it may be difficult to distinguish between benign and malignant follicular neoplasms, there may be some under or over-reporting of follicular thyroid cancer. On the other hand, it is relatively easy to diagnose papillary carcinoma and ascertainment is more complete. Anaplastic and medullary carcinomas usually account for about 5% of thyroid cancers.

Several subtypes of well-differentiated thyroid cancer are generally recognized, but data about them are limited because of their relative rarity. A subtype of papillary thyroid cancer of epidemiologic interest is the solid variant, which has an unusually high frequency among the childhood cases occurring in the Chernobyl area (Williams, 2002; Nikiforov et al., 1997). It remains to be determined whether this represents a particular feature of radiation-related thyroid cancer or other factors such as young age or iodine deficiency. In addition, two subtypes are reported to be associated with aggressive behavior—the tall cell variant and columnar cell variant of thyroid cancer (Baloch et al., 2001). Hürthle cell neoplasms have, until recently, been categorized as variants of follicular neoplasms. The Hürthle cell is filled with mitochondria, which results in distinctive staining properties, and the tumors are difficult to treat because they have lost the ability to concentrate iodine. Recent studies of somatic mutations indicate that ret gene rearrangements are found both in papillary thyroid cancers and in many Hürthle cell tumors (Cheung et al., 2000), but not in follicular cancers. Initial findings suggested that ret-positive cases are characterized by local spread, similar to papillary cancers, while ret-negative cases are prone to hematogenous distant spread, characteristic of follicular cancers. Similarly, PAX8 rearrangements are characteristic of follicular cancers, but are not seen in Hürthle cell cancers (Nikiforova et al., 2003). The female/male ratio for papillary and follicular thyroid cancer is approximately 3 : 1, whereas medullary carcinoma occurs about equally in males and females. The mean age at diagnosis is mid-40s to early 50s for papillary carcinoma, the 50s for follicular and medullary cancer, and the 60s and older for anaplastic cancer. Children and adolescents usually have papillary cancer, with 70%–90% of childhood thyroid cancers being of this type (Poth, 2000). Medullary thyroid cancer occurs as a sporadic tumor or as a part of the multiple endocrine neoplasia type 2 (MEN-2) syndrome, caused by germline mutations of the ret gene.

DIAGNOSIS With the introduction of new methods of diagnosing thyroid cancer, the incidence and mortality patterns of this malignancy have changed. Initially, palpation was the predominant means of finding thyroid nodules. When they became available, isotopic scans were used to confirm and/or characterize these nodules. Surgery was necessary to diagnose thyroid cancer. Later, technical improvements (new forms of tracers and pin hole collimation) enabled thyroid scans to become more sensitive than palpation (Ryo et al., 1976). As a result, in some settings, scans were used for screening (Schneider et al., 1980). In most instances an abnormality found on a thyroid scan would signify a benign or malignant thyroid nodule generally larger than 1 cm. The most recent advance in thyroid imaging is the use of high-definition ultrasonography. Nodules of only a few millimeters are readily detected. Importantly, nodules larger than 1 cm that are not evident by palpation or scanning are found by ultrasound (Inskip et al., 1997; Schneider et al., 1997). Whether thyroid ultrasound should be used for screening in any circumstance has been discussed extensively (Eden et al., 1999, 2001). Although its sensitivity is impressive, its specificity is not, since most

975

976

PART IV: CANCER BY TISSUE OF ORIGIN

Table 50–1. Distribution of Thyroid Cancer by Histologic Type in Selected Countries, 1993–1997 Histology %* Country, Region/Race

No. of Cases†

Papillary

Follicular

Medullary

Anaplastic

Algeria, Algiers South Africa, Black South Africa, White Colombia, Cali‡ Canada US (SEER) Black US (SEER) White India, Bombay Israel, Jews Japan, Hiroshima** Singapore, Chinese Austria, Tyrol Denmark Finland Iceland Italy, North East Netherlands Norway Sweden UK, Scotland Australia, NSW

227 647 737 300 7065 404 6296 469 1559 452 505 267 580 1725 135 481 1629 851 1526 574 1530

46.3 26.1 59.3 80.0 79.9 70.8 81.5 46.5 77.5 92.3 68.9 56.9 55.2 78.4 83.7 67.8 54.8 70.4 63.7 59.1 74.1

25.6 34.8 18.7 8.3 12.0 21.3 12.5 29.0 12.5 6.0 26.7 33.7 23.3 10.0 11.9 13.3 23.1 12.6 16.6 25.4 17.4

5.3 3.4 3.7 1.3 2.9 3.5 2.4 8.5 3.3 0.7 2.0 1.9 9.7 3.3 3.0 6.0 10.8 4.5 5.6 5.9 4.2

0.9 4.9 2.4 4.3 2.1 1.0 1.7 3.0 1.9 0.2 2.0 4.5 7.1 — 0.7 2.9 5.3 3.3 8.8 3.5 1.6

Source: Data from Parkin et al., 2002a. *Total does not add to 100 because there are cancers of other and unspecified histology. † Cases include only those that are microscopically verified. ‡ Period of follow-up 1992–1996. **Period of follow-up 1991–1995.

thyroid nodules are benign and very small cancers may not be clinically significant. Nevertheless, to the extent that thyroid ultrasound is used, an increase in thyroid cancer incidence is to be expected. Thyroid ultrasound has been used to screen a large number of individuals living in the Chernobyl area and to screen other radiationexposed groups (Davis et al., 2004a; Stezhko et al., 2004). In addition, ultrasound is used to visualize other structures in the neck (e.g., the carotid arteries) and thyroid nodules may be discovered as an unanticipated consequence (Burguera and Gharib, 2000). When a thyroid nodule larger than 1 cm is discovered and it is not evident whether it is malignant or not, as is true in most cases, it is subjected to fine needle aspiration (FNA). How the introduction of FNA has influenced epidemiologic patterns is not certain. In one series from a large referral center, the frequency of thyroid surgery for nodules decreased dramatically, while the number of thyroid cancers recognized at surgery remained constant (i.e., the specificity of surgery increased) (Gharib and Goellner, 1993). Still, it seems likely that some nodules, particularly those within a multinodular goiter, which previously might not have been observed, are now identified as cancers and removed. On the other hand, before FNA, small thyroid cancers (socalled microcarcinomas) were discovered as an incidental finding when a larger benign nodule was removed. Now some of these microcarcinomas may be missed because FNA of a large nodule may forestall surgery. As discussed below, the ascertainment of familial forms of medullary thyroid cancer has changed dramatically. Children with an affected parent can be tested genetically. Prophylactic thyroidectomy prevents thyroid cancer in those who test positive and should, in time, be reflected in epidemiologic trends.

PROGNOSIS AND TREATMENT The prognosis for well-differentiated thyroid cancer is very good, and the clinical challenge is identifying the minority of patients at high risk. This has been accomplished using multivariate analyses and cancer recurrence as an endpoint. A variety of scoring systems have been developed to estimate the risk of recurrence, and a summary has

been published by the British Thyroid Association and the Royal College of Medicine (2002). There is considerable agreement among the scoring systems about the predominant risk factors. Factors associated with increased risk are gross invasion through the thyroid capsule, distant metastases, larger size, and older age. One widely used scoring system is based on the TNM system of describing thyroid cancer and uses four stages. Emphasizing the importance of age at presentation, cases occurring before age 45 can only be classified in the lowest two stages. All patients younger than age 45 are considered stage one, except for those with distant metastases who are classified as stage two (American Joint Committee on Cancer, 1997). Above age 45 years at diagnosis, all four stages are used. The principal modalities for managing thyroid cancer are surgery and radioactive iodine (131I) ablation of thyroid tissue. Many unresolved questions remain about the optimum treatment because the excellent prognosis of most cases makes it virtually impossible to carry out controlled studies. General agreement emerges for high-risk cases in which the goal is to ablate all thyroid tissue by the combination of surgery and radioactive iodine. Because thyroglobulin is a protein made exclusively by thyroid follicular cells, measuring its level in the blood is now the most important measure of successful ablation. Even when ablation is not completely successful, the outlook remains favorable in many cases, but differences emerge regarding the best treatment approach.

DEMOGRAPHIC PATTERNS Incidence and Mortality in the United States The thyroid gland is one of the less cancer-prone organs in the body, representing only 0.84% of cancers occurring among US males and 2.5% among US females (Jemal et al., 2003). The lifetime risk of developing thyroid cancer is less than 1%. The 1996–2000 agestandardized incidence rates for thyroid cancer in the US SEER cancer registries were 3.6 and 9.9 per 100,000 males and females, respectively (Ries et al., 2003). Because long-term survival is exceptionally good for the large majority of thyroid cancer patients, only 0.21% of male and 0.30% of female cancer mortality is due to thyroid cancer. The number of incident cases is about 15 times higher than that of deaths: 22,000 new cases and 1400 deaths were projected for 2003 (Jemal et al., 2003). The ratio of incidence to mortality decreases, however, with age. This is partially because anaplastic (undifferentiated) thyroid cancer, which has a high fatality rate, occurs among older patients. Overall, the 1996–2000 age-adjusted US mortality rates were 0.4 and 0.5 per 100,000 males and females, respectively. Due to its excellent prognosis, the prevalence of thyroid cancer is higher than would be expected based on incidence rates. In 2000, there were about 300,000 people in the United States living with thyroid cancer (Ries et al., 2003) (i.e., about 3% of all cancer cases). Thyroid cancer incidence and mortality rates measure different types of disease. Incidence rates are heavily weighted with young women with papillary or follicular carcinoma, whereas mortality statistics reflect elderly males and females with a relatively high proportion of undifferentiated carcinomas.

International Patterns Approximately 123,000 new thyroid cancers are diagnosed annually worldwide. Similar to the US data, thyroid cancer comprises somewhat less than 2% of all cancers (Parkin et al., 2002b). With the exception of a few regions, where thyroid cancer incidence is particularly high, incidence rates (world adjusted) from much of the world range between 1 and 2 cases per 100,000 males and between 2 and 8 per 100,000 females. Low rates occur both in developed and developing countries, although high rates are generally confined to developed nations with good medical systems (Table 50–2). Based on data from 23 world areas defined by the United Nations as developed or developing, the number of thyroid cancer deaths among women was higher in developing than in developed regions (Parkin et al., 1999). In a

977

Thyroid Cancer

Incidence* Continent Asia

Europe

North America South America

Oceania

Country

Area

Male

Female

China† China† India Israel Japan Kuwait† Philippines† Singapore Denmark Estonia Finland France Germany Iceland Norway Italy Poland Slovenia Spain Sweden Switzerland UK† Canada US US Colombia† Costa Rica Cuba† Ecuador† Australia New Zealand Hawaii

Hong Kong Shanghai Mumbai (Bombay) All (Jews) Osaka Prefecture All (Kuwaitis) Rizal Province All (Chinese) All All All Bas-Rhin Saarland All All Lombardy Region Cracow City All Murcia Province All Vaud England All SEER (White) SEER (Black) Cali All All Quito New South Wales All Filipino Hawaiian

1.9 1.1 0.8 3.5 1.2 1.8 2.4 1.9 1.0 1.1 2.3 1.9 2.2 4.3 1.6

7.1 3.9 2.0 9.7 3.8 7.6 7.5 5.9 2.3 3.5 7.8 4.8 4.8 12.6 4.3

1.7 1.3 1.5 1.4 1.4 0.8 2.2 2.8 1.4 1.7 1.1 1.2 2.2 2.2 1.7 5.0 4.6

4.4 3.6 5.2 3.5 4.9 1.9 6.4 7.7 4.0 6.7 6.4 7.2 8.0 6.3 3.5 19.4 11.0

Source: Data from Parkin et al., 2002a. *Annual rate per 100,000 population standardized to the World Standard Population. † Quality indicators suggest that data from these registries may be incomplete.

recent publication of cancer incidence in 15 African registries, the rates of thyroid cancer incidence in parts of Algeria, Tunisia, Zimbabwe, and South Africa were similar to those found in low-incidence European countries such as the United Kingdom and Netherlands (Parkin et al., 2002b). Although data are sparse, no significant difference in incidence has been observed between urban and rural areas in Europe; but thyroid cancer is more common in some regions within countries (e.g., Tarn or Cote d’Or in France) or in population subgroups (e.g., Filipinos living in California) (Parkin et al., 2002a). Proposed explanations for geographic variation include a possible association with endemic goiter, volcanic lava, dietary patterns, or differences in reporting or case ascertainment. International comparisons for thyroid cancer are particularly subject to a wide diversity in diagnostic, treatment, and reporting practices. Furthermore, because thyroid cancer is a rare malignancy, many of the reported sex-specific rates in low-incidence areas are based on few cases, sometimes less than 10, making them unstable. In addition, quality indices suggest that the data from several of the Asian, African, and South American registries may not be complete (Parkin et al., 2002a,b). Thyroid cancer incidence in the five Nordic countries varies considerably, with a fivefold difference between highest rate (Iceland) and the lowest rate (Denmark) (Fig. 50–1). These difference are notable because similar standards of medical care and reporting in the Nordic countries reduce the limitations associated with other intercountry comparisons. Differences between countries were apparent for both sexes, all age groups, and all histologic types, although the variation was widest for papillary thyroid cancer (Parkin et al., 2002a). Disparity in thyroid cancer incidence rates in the Nordic countries also was

observed when study pathologists reclassified diagnoses using uniform diagnostic criteria (Franssila et al., 1981). The reasons for the variation in incidence among the Nordic countries are presently unknown, although differences in iodine intake may play some role. In contrast to the low incidence and mortality rates of thyroid cancer, the prevalence of occult thyroid cancer found at autopsy is typically quite high and the epidemiologic patterns differ. Among autopsied cases, the sex ratio is more nearly equal and geographic variation can be extreme. The prevalence of occult papillary thyroid cancer at autopsy was 36% in Finland (Harach et al., 1985), 35% for Japanese living in Japan or Hawaii (Fukunaga and Lockett, 1971; Fukunaga and Yatani, 1975), 3%–13% in Canada, the United States, Poland, Israel, and Argentina (Fukunaga and Yatani, 1975; Ottino et al., 1989; Sampson et al., 1974; Siegal and Modan, 1981; Silverberg and Vidone, 1966), and less than 1% in Switzerland (Heitz et al., 1976). Although some variation can be attributed to the extent of thyroid examination, geographic differences also were found when comparisons were made by the same pathologists (Fukunaga and Yatani, 1975). Franssila and Harach (1986) observed a high prevalence of occult thyroid cancer during puberty and postulated that pubertal hormonal factors may act as promoters. With the exception of Switzerland (Heitz et al., 1976), papillary carcinoma is the predominant histologic type in most autopsy series, probably reflecting its slow-growing behavior.

Time Trends In contrast to the recent statistically significant decrease in overall US cancer incidence, thyroid cancer incidence is increasing sharply (Fig. 50–2). The increase has been observed for both whites and blacks and in many geographic regions (Mulla and Margo, 2000; Zheng et al., 1996). In the period 1975–2001, thyroid cancer incidence went from 6.3 to 12.4 per 100,000 white females (Fig. 50–3). During the same time period, the rate also rose for white men (from 2.8 to 4.4), black women (from 4.2 to 6.0), and black men (from 1.5 to 2.7). The trend, however, was not monotonic; rates rose significantly between 1973 and 1977, declined between 1977 and 1980, and then increased again from 1980 to 2000, with a notable acceleration after 1988 (Howe et al., 2001). Based on data from the 12 SEER cancer registries, which cover 14% of the US population, there was a 3.8% annual increase in age-adjusted (2000 US standard population) thyroid cancer incidence between 1992 and 2000 (Ries et al., 2004). The percent increase was second only to liver and intrahepatic bile duct cancer, and it was first among women (4%) and third among men (1.8%).

14 12 Thyroid Cancer Incidence

Table 50–2. Age-Standardized* Incidence of Thyroid Cancer in Selected Populations by Gender, 1993–1997

10 8 6 4 2 0 Denmark

Sweden

Norway

Finland

Iceland Male Female

Figure 50–1. Annual, sex-specific, age-standardized (world) thyroid cancer incidence rates in the Nordic countries, 1993–1997. Rates are per 100,000 population. Source: Parkin et al., 2002a.

978

PART IV: CANCER BY TISSUE OF ORIGIN

In other countries, increasing incidences also have been observed for men and women over the past two to three decades (Bacher-Stier et al., 1997; Burgess, 2002; Colonna et al., 2002; Fahey et al., 1995; Liu et al., 2001; Lundgren et al., 2003; Remontot et al., 2003; Verkoi-

jen et al., 2003). In general, the striking increases are due to steadily climbing rates of papillary carcinoma and appear to be higher among women than men (Burgess, 2002; Colonna et al., 2002; Lundgren et al., 2003; Verkoijen et al., 2003). During a similar time period, follicular thyroid cancer increased slightly in some, but not all countries (Burgess, 2002; Colonna et al., 2002). Time trends for anaplastic and medullary thyroid cancers are not very clear, but the incidence of medullary cancer seems to be fairly stable while anaplastic cancer appears to be decreasing slightly. In Australia, where health care practices across the country are considered to be uniform, the annual increase in papillary thyroid carcinoma was 10.7% for females and 8.3% for males during the period 1982–1997 (Burgess, 2002). The increase was seen primarily in geographic areas with iodine deficiency, which is inconsistent with the generally reported association of papillary thyroid carcinoma with iodine sufficiency and follicular thyroid carcinoma with iodine deficiency. To help explain the unusual pattern, Burgess (2002) hypothesized that radioactive fallout from nuclear weapons testing in the South Pacific in the 1950s and 1960s may have increased the incidence of papillary thyroid cancer in the iodine-deficient areas because iodine deficiency causes greater thyroidal uptake of radioiodines. Recent data from Chernobyl suggest that iodine deficiency increases the development of thyroid cancer following radioiodine exposure, possibly as much as three-fold (Cardis et al., 2005; Shakhtarin et al., 2003). Two important reasons for the overall rise in thyroid cancer incidence over the past three decades have been identified: improved detection and radiation exposure. The introduction of better diagnostic tools since the 1970s has resulted in an upsurge in incidence. In those countries where medical irradiation was used to treat benign diseases from the 1920s to the 1960s, thyroid cancer incidence increased among persons born during, or soon after, those years (Pottern et al., 1980; Zheng et al., 1996). Radioactive fallout from nuclear testing (Burgess, 2002; Lund et al., 1999) or from the Chernobyl accident (Cotterill et al., 2001; Mangano, 1996) may also contribute to the upward trend. However, while radioactive fallout from weapons testing may have increased the risk for thyroid cancer in some population subgroups in the United States (Gilbert et al., 1998; Kerber et

Figure 50–3. Age-standardized US thyroid cancer incidence and mortality trends by gender and race, 1975–2001. US SEER (9 registries) and national mortality rates. Rates are per 100,000 population. Regression lines are calculated using Jointpoint Regression Program.

(Source: Ries LAG, Eisner MP, Kosary CL, Hankey BF, Miller BA, Clegg L, Mariotto A, Feuer EJ, Edwards BK, eds. SEER Cancer Statistics Review, 1975–2001, National Cancer Institute. Bethesda, MD. Available at: http://seer.cancer.gov/csr/1975_2001/, 2004.)

5 4.5 4

Trend over time

3.5 3 2.5 2 1.5 1 0.5 0

1975–2001

1997–2001

1992–2001

Year of diagnosis Male and female Male Female

Figure 50–2. Annual percent change in US age-standardized thyroid cancer incidence rates by gender, 1975–2001. US SEER (9 registries). Rates are per 100,000 population. Source: Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database: Incidence–SEER 9 Registries, Public Use, Nov 2003 Sub (1973–2001), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2004, based on the November 2003 submission.

979

Thyroid Cancer al., 1993), reports of significant increases in other parts of the world are not common. In Belarus, Ukraine and parts of Russia, there has been a striking increase in papillary thyroid cancer incidence due to childhood exposure from Chernobyl (United Nations Scientific Committee on the Effects of Atomic Radiation [UNSCEAR], 2000), but outside of these highly contaminated areas, a clear association with radioactivity has not been demonstrated. Mortality, on the other hand, has decreased in many countries (Coleman et al., 1993; Franceschi and La Vecchia, 1994; La Vecchia et al., 1992; Levi et al., 1999). Over the 40-year period of 1950 through 1991, US mortality rates for white females decreased more than 50% from 0.8 per 100,000 in 1950–1954 to 0.4 per 100,000 in 1976–1980 to 0.3 per 100,000 in 1988–1991. Among males there was also a decline, but it was less dramatic—from 0.5 per 100,000 to 0.3 per 100,000 during the same time period (Pickle et al., 1987; Ries et al., 1994). Since 1992, however, the overall mortality rate for thyroid cancer was one of the few to increase in the United States. Although the mortality rate remained very low and the increase was small (a non-significant 0.4% annual percent change from 1992–2000), this is a disturbing trend (Howe et al., 2001; Ries et al., 2003). As seen in Figure 50–4, from 1990–2000, thyroid cancer mortality increased for all groups except white females, and the increase for white males and black females was greater in the past few years. Because survival is higher for females, the sex differential in mortality rates is much smaller than in incidence rates. An estimated 840 women and 620 men are projected to die from thyroid cancer in 2004 (Jemal et al., 2004).

Age and Sex Thyroid cancer has an unusual age distribution that is somewhat between the pattern for adult and childhood cancers (Fig. 50–5). About 25% of all thyroid cancer occurs in patients less than 35 years of age and incidence rises comparatively slowly with age. Although the median age at diagnosis for cancers of all sites combined is 67 years, it is only 46 years for thyroid cancer (Ries et al., 2003).

Females consistently have a higher incidence of thyroid cancer than males (Parkin et al., 2002a). The female-to-male ratio is high (>3) after puberty and during the reproductive years (Fig. 50–5). At the time of menopause, female incidence rates begin to decline, whereas the rates among males continue to increase until age 70 years, at which time the female-to-male ratio drops to about 1.2. Although the incidence of thyroid cancer is roughly 2–3 times greater among females, mortality rates are nearly equal until about age 60. After about 60 years, mortality begins to increase more rapidly among females than males.

Race, Ethnicity, and Religion Patterns relating to race, ethnicity, and religion generally are inconsistent, although the incidence of thyroid cancer is about two times greater among US whites than blacks (Table 50–3), and is threefold to fourfold higher among South African whites compared with blacks. In the United States, the higher incidence of thyroid cancer among whites is confined to papillary carcinoma (Correa and Chen, 1995; Ries et al., 2003), suggesting that it may be related to better medical care. On the other hand, in a very large nationwide survey, whites had higher serum TSH concentrations than blacks (Hollowell et al., 2002), which might enhance their thyroid cancer risk. It also is notable that persons of Asian origin, especially from the Philippines and the South Pacific Islands, have elevated rates compared with persons of other ethnic backgrounds (Ballivet et al., 1995; Bernstein et al., 1995; Blot et al., 1997; Haselkorn et al., 2003; Iribarren et al., 2001; Mishra, 1996; Rossing et al., 1995). Thyroid cancer incidence rates are elevated in Hawaii, and Hawaii residents often have higher rates than members of the same ethnic group living elsewhere (Goodman et al., 1988). The substantially enhanced rates of thyroid cancer among Filipinos living outside of their home country are striking. Filipino women living in Hawaii, San Francisco, and Los Angeles have incidence rates that are two to four times higher than women in the Philippines (Bernstein et al., 1995; Haselkorn et al., 2000; Parkin et al., 2002a; Rossing et al., 1995). Thyroid cancer incidence is elevated also in several other Pacific

4.5 4 3.5 3

Trend over time

2.5 2 1.5 1 0.5 0 –0.5 –1 –1.5 –2

1969-2001

1997-2001

1992-2001

Year of death White / Male White / Female Black / Male Black / Female

Figure 50–4. Annual percent change in US age-standardized thyroid cancer mortality rates by gender, 1969–2001. US SEER (9 registries). Rates are per 100,000 population. Source: Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat

Database: Mortality–All COD, Public Use with state, total US (1969–2001), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2004. Underlying mortality data provided by NCHS (www.cdc.gov/nchs.)

980

PART IV: CANCER BY TISSUE OF ORIGIN 17.5

15

Rate per 100,000

12.5

10

7.5

5

2.5

Figure 50–5. US age-specific thyroid cancer incidence rates by gender, 1992–2001. (Source: Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Databases: Incidence–SEER 11 Regs + AK Public Use, Nov 2003 Sub for Expanded Races (1992–2001) and Incidence–SEER 11 Regs Public Use, Nov 2003 Sub for Hispanics (1992– 2001), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2004, based on the November 2003 submission.)

0

Area Atlanta, Georgia Alameda, California Bay Area, California Detroit, Michigan Los Angeles, California New Orleans, Louisiana

85+

80–84

75–79

70–74

65–69

60–64

55–59

50–54

45–49

40–44

35–39

30–34

25–29

20–24

15–19

Male Female

Table 50–3. Age-Standardized Incidence* of Thyroid Cancer in Selected US Areas by Race and Gender, 1983–1987 White : Black Ratio

Black

10–14

Male and Female

Island populations including French Polynesia, New Caledonia, Fiji, Vanuatu, and American Samoa (Ballivet et al., 1995; Blot et al., 1997; de Vathaire et al., 2000; Henderson et al., 1985; Mishra et al., 1996; Paksoy et al., 1990, 1991). The reasons for the exceptional rates of thyroid cancer in the Pacific Islands are unclear, but dietary factors, especially high iodine intake, may play a role. Indeed, in New Caledonia and French Polynesia, high levels of urinary iodine have been reported (Le Marchand et al., 1995). In addition, Haselkorn et al. (2003) noted that the high incidence of thyroid cancer among Southeast Asian women living in the San Francisco area is associated with a history of goiter or thyroid nodules. The hypothesis that Jews are more susceptible to both naturally occurring (Bross et al., 1971) and radiation-associated (Hempelmann et al., 1975; Shore et al., 1985, 1993) thyroid cancer has prompted other investigators to examine religion as a risk factor. Most studies have not indicated a significantly increased risk among Jews compared with Catholics or Protestants (Greenwald et al., 1975; McTiernan et al., 1984b; Pottern et al., 1990; Ron et al., 1987); however, an elevated incidence of thyroid cancer among Jews was observed in Los Angeles (Preston-Martin and Menck, 1979). Jews in Israel have a high incidence of thyroid cancer, twice as high as non-Jewish Israelis (Parkin et al., 2002a), but this difference may be due to incomplete case ascertainment and reporting in the non-Jewish population.

White

05–09

00–04

Age at diagnosis

Female

Male

Female

Male

Female

Male

5.4 5.1 5.6 6.7 5.9

2.0 2.5 2.6 2.9 2.3

2.4 2.7 3.3 3.5 3.2

0.7† 1.2† 0.9 1.4 1.2

2.2 1.9 1.7 1.9 1.8

2.9 2.1 2.9 2.1 1.9

5.4

2.2

3.9

1.6

1.4

1.4

Source: Data from Parkin et al., 2002a. *Annual rate per 100,000 population standardized to the World Standard Population. † Based on <10 cases.

Socioeconomic Status Neither social class nor education are clear risk factors for thyroid cancer (Franceschi et al., 1989; Frentzel-Beyme and Helmert, 2000; Kolonel et al., 1990; McTiernan et al., 1984b, Ron et al., 1987; Vågerö and Persson, 1986; Zivaljevic et al., 2003). However, in California and Kuwait the incidence of thyroid cancer was elevated among persons with a high school or greater education or with high socioeconomic status (Haselkorn et al., 2000, 2003; Iribarren et al., 2001; Memon et al., 2002b; Preston-Martin and Menck, 1979), and in Sweden it was lower among male blue collar workers than other men (Carstensen et al., 1990). In Switzerland, thyroid cancer cases were better educated and more often belonged to a higher socioeconomic class than controls (Levi et al., 1991). The positive associations in some populations suggest the role of medical surveillance and detection bias. In contrast, several studies have found higher incidence rates among populations with low socioeconomic levels (Goodman et al., 1988; Memon et al., 2002a; Trapido et al., 1990).

Survival With the exception of non-melanoma skin cancer, survival for thyroid cancer is better than for any other cancer site. In the United States, the 1992–1999 all-stage relative 5-year survival rate was 96%, which is significantly better than it was in 1974–1976 (Ries et al., 2003). One possible reason for the improved survival is the rise in incidence of histologic types (papillary and follicular) with good prognosis and more effective treatment. On the other hand, improved prognosis may be due, in large part, to the increased diagnosis of very small cancers that have little risk of progression. Statistics from 1992–1999 indicate that 5-year survival was somewhat better for women (97%) than for men (92%) and the rate was 99% for localized thyroid cancer. Although thyroid cancer with distant metastases had poor 5-year survival (60%), only 6% of thyroid cancers were diagnosed at this stage, compared with 36% at the regional stage, and 56% at the local stage (Ries et al., 2003). At 10 years post diagnosis, overall prognosis remains excellent for differentiated thyroid cancer (Gilliland et al., 1997; Lundgren et al., 2003; Mazzaferri, 1994). After 20 years, the relative survival rate decreased, and the poorer survival for follicular thyroid carcinoma and for males became more pronounced, as well as the strong trend for decreased survival with increasing age at diagnosis (Lundgren et al., 2003).

Thyroid Cancer Survival rates are highly correlated with histologic type. Differentiated thyroid cancers have good survival rates. Papillary carcinoma has the best prognosis of all histopathologic types. Follicular carcinoma also has a good prognosis, except when distant metastases are present. Papillary carcinoma spreads locally through lymphatic vessels, whereas follicular carcinoma metastasizes widely via the blood stream. Survival is extremely poor for undifferentiated or anaplastic cancer and the grim outlook for these patients has not improved substantially over time. For medullary thyroid cancer, with the exception of early stage tumors that are cured by the initial surgery, prognosis is not favorable, but survival may extend over many years. Taking histology into account, survival is inversely correlated with age at diagnosis, size of tumor, degree of local invasion, and extent of distant metastases (Akslen et al., 1991; Gilliland et al., 1997; Levi et al., 1990; Mazzaferri and Jhiang, 1994; Tanaka et al., 1999). Although thyroid cancer prognosis is excellent for most patients, patients aged 70 and over at diagnosis appear to have more aggressive cancers and have considerably poorer survival than younger patients (Vini et al., 2003). In contrast, although young patients (<18 years old) are diagnosed with neck and distant metastases more often than adults, their long-term survival does not appear to be reduced (Kowalski et al., 2003).

RISK FACTORS A major source of current information on thyroid cancer risk factors, excluding radiation, is an international pooled analysis of 14 casecontrol studies (Negri et al., 1999a; Preston-Martin et al., 2003). Because of the statistical stability gained by pooling data from multiple studies, especially for a relatively uncommon disease such as thyroid cancer, and the breadth of the topics addressed, the pooled data form the basis for much of the following discussion of host and environmental risk factors.

Host Factors Thyroid Stimulating Hormone Thyroid stimulating hormone (TSH) is produced by the pituitary gland and it in turn prompts the thyroid to produce thyroid hormone. In humans, elevated levels of TSH are associated with thyroid growth and are thought to increase the risk of thyroid carcinoma, but the relationship is not clearly established (Derwahl et al., 1999; Henderson et al., 1982; Williams, 1990). In a US survey, women were found to have higher TSH levels than men (Hollowell et al., 2002). In rodents, increased TSH secretion induces thyroid tumors (Hempelmann and Furth, 1978; Williams, 1995), but because rodents appear to be more sensitive to the carcinogenic effects of elevated TSH, its role in humans remains uncertain (Hill et al., 1998; Hard 1998). Several possible mechanisms by which TSH interacts with other risk factors have been considered. TSH secretion is elevated in women during puberty, pregnancy, delivery, and oral contraceptives use (Glinoer 1999; Malkesian and Mayberry, 1970; Pacchiarotti et al., 1986; Weeke and Hanson, 1975). TSH levels may also rise following partial thyroidectomy for benign thyroid disease (Nmec et al., 1980), goitrogen intake (Paynter et al., 1988), and radiation to the neck (Williams, 1990).

Reproduction and Hormones As discussed earlier, thyroid cancer is one of the few non-gender specific cancers that is substantially more common among women than men. Women also have a higher incidence of other thyroid diseases. Although the causes for these gender differences are unknown, a role for sex hormones, especially estrogens, is suspected and an interaction with TSH has been hypothesized. In women, the peak occurrence of thyroid cancer is during the reproductive years. Thus, it is noteworthy that the thyroid gland becomes enlarged during puberty, menstruation, and pregnancy (Krassas, 2000) and that estrogen receptors are highly expressed in human and animal thyroid neoplasms (Takeichi et al., 1991; Yane et al., 1994). Estradiol metabolites were found

981

to be elevated in premenopausal thyroid cancer patients before surgery compared with normal women (Lee et al., 2003). Experiments in animals and cell lines also support a role for reproductive hormones (Gross et al., 1993; Manole et al., 2001; Takagi et al., 2002). Estradiol was found to have a growth-stimulating effect on both benign and malignant thyroid cells associated with increased expression of cyclin D1, suggesting that estradiol promotes thyroid tumorigenesis in females (Manole et al., 2001). In male rats, castration reduces the incidence of thyroid cancer (Paloyan et al., 1982), whereas testosterone promotes radiation-induced thyroid tumors by increasing TSH levels (Hofmann et al., 1986). In human thyroid cancer cells, testosterone and estradiol increased the proliferation of papillary carcinoma cell lines. Testosterone also increased proliferation in follicular cancer cell lines, but estradiol had an antiproliferative effect (Banu et al., 2001). Although many studies have reported an association between thyroid cancer and the number of pregnancies and/or livebirths (Franceschi et al., 1990, Galanti et al., 1995; Hallquist et al., 1994; Kravdal et al., 1991; Levi et al., 1993; McTiernan et al., 1984b; Preston-Martin et al., 1987; Ron et al., 1987; Takezaki et al., 1996; Wingren et al., 1993), a nearly equal number have not found an association (Akslen et al., 1992; Iribarren et al., 2001; Kolonel et al., 1990; Mack et al., 1999; Memon et al., 2002a; Picchi et al., 2001; Rossing et al., 2000). In the pooled analysis, later age at menarche conferred a small (10%–20%) elevated risk of thyroid cancer (Table 50–4). Parity also was associated with a higher risk, but risk did not increase with increasing number of births and was only marginally associated with increasing age at first birth. Women who had a history of a miscarriage or induced abortion did not have an overall increased risk of thyroid cancer, but miscarriage as the outcome of a first pregnancy conferred an elevated risk (Negri et al., 1999b), suggesting that changes early in reproductive life may be important. A marginally elevated risk of thyroid cancer among infertile women has been reported (Brinton et al., 1989; Hallquist et al., 1994; Modan et al., 1998; Negri et al., 1999b) and, in a case-control study, fertility drugs were used significantly more often by cases than by controls (Kolonel et al., 1990). In the pooled analysis, which included the data from Kolonel et al. (1990), the odds ratios for a self-reported history of infertility and infertility treatment were 1.2 (95% CI: 0.9–1.6) and 1.6 (95% CI: 0.9–2.9), respectively (Negri et al., 1999b; La Vecchia et al., 1999). If the infertility were due to habitual abortion, these data would be consistent with studies implicating miscarriage in thyroid cancer. Artificial menopause was linked to thyroid cancer in some studies (Levi et al., 1993; Luota et al., 2003; Negri et al., 1999b) possibly related to underlying conditions such as a bleeding disorder (Luota et al., 2003). The use of oral contraceptives has been associated with thyroid cancer in several studies (La Vecchia et al., 1999; McTiernan et al., 1984b; Picchi et al., 2001; Preston-Martin et al., 1987, 1993; Zivaljevic et al., 2003), but not all (Memon et al., 2002a; Rossing et al., 1998; Sakoda et al., 2002). The association with oral contraceptives was strongest for current use, suggesting a promotional rather than an initiating effect. A 50% rise in thyroid cancer incidence was observed among women who took drugs for lactation suppression, but no increase was found for hormone replacement therapy (La Vecchia et al., 1999). Although the current data suggest an influence of hormonal and reproductive factors in the etiology of thyroid cancer, the effects have been relatively weak and not always consistent. The few associations that have been seen appear somewhat stronger among patients who were young at the time of thyroid cancer diagnosis (La Vecchia et al., 1999; Negri et al., 1999b; Preston-Martin et al., 2003).

Predisposing Thyroid Disease As shown in Table 50–5, a strong link between benign thyroid nodules or goiter and thyroid cancer has been demonstrated in numerous epidemiologic studies (D’Avanzo et al., 1995; Franceschi et al., 1999; From et al., 2000; Iribarren et al., 2001; Levi et al., 1991; Mack et al., 1999; Memon et al., 2002b; Soki et al., 1994; Takazaki et al., 1996; Zivaljevic et al., 2003). In most case-control studies, the odds ratios

982

PART IV: CANCER BY TISSUE OF ORIGIN

Table 50–4. Thyroid Cancer Risk Associated with Selected Hormonal and Reproductive Factors, International Pooled Thyroid Cancer Case-Control Study* Factor

Cases %

Controls %

Odds Ratio

95% CI†

31 44 25

35 43 22

1‡ 1.1 1.2

1.0–1.3 1.0–1.4

26 74 18 43 13

27 73 16 44 14

1 1.2 1.3 1.2 1.2

1.0–1.4 1.0–1.6 1.0–1.4 1.0–1.6

8 32 24 9

9 35 22 7

1 0.9 1.1 1.3

0.7–1.1 0.8–1.4 1.0–1.8

26 62 6 6

1 1.3 1.3 1.8

1.0–1.8 0.9–2.0 1.2–2.6

67 23 10

1 1.3 1.8

1.0–1.8 1.4–2.4

61 39

1 1.2

1.0–1.4

71 5 16 9

1 1.5 1.1 1.1

1.0–2.1 0.9–1.6 0.8–1.4

98 2

1 1.6

0.9–2.9

76 24

1 1.5

1.1–2.1

age at menarche <13 13–14 15+

number of births** Nulliparous Parous 1 2–3 4+

age at first birth†† <20 20–24 15–29 30+

outcome of first pregnancy†† Nulligravidae Birth Induced abortion Miscarriage

23 62 7 8

menopausal status‡‡ Pre Post, natural Post, artificial

68 20 12

oral contraceptive use§ Never Ever

62 38

time since last use (years)§ Never used Current 1–10 >10

70 6 17 8

infertility treatment†† Never Ever

97 3

lactation suppressants

¶

Never Ever

74 26

hormone replacement therapy|| Never Ever

92 8

91 9

1 0.8

0.6–1.1

Source: Data from Negri et al., 1999; La Vecchia et al., 1999. *Risk estimates from conditional logistic regression conditioned on study and age, and adjusted for history of radiation and age. † CI, confidence interval. ‡ Reference category. **Also adjusted for oral contraceptive use. †† Also adjusted for parity. ‡‡ Also adjusted for hormone replacement therapy. § Also adjusted for parity, type of menopause and education. ¶ Parous women only, also adjusted for parity and breastfeeding. || Also adjusted for menopausal status and type of menopause.

history of thyroid adenoma, nodule, or goiter confers an enhanced risk of thyroid cancer, Indeed, other than childhood radiation exposure, a history of benign thyroid nodules is the strongest known risk factor for thyroid cancer. This association may reflect a true causal relation, a precursor lesion, an effect of treatment for the benign disease, similar risk factors exerting independent effects, close medical surveillance, or misdiagnosis of the earlier disease. But the strong epidemiologic evidence of a relationship between thyroid nodules and subsequent thyroid cancer underscores the need to understand the mechanisms involved. Somatic genetic mutations seem to be associated with thyroid tumor progression (Fagin, 1994; Farid et al., 1994; Segev et al., 2003). Mutations in ras oncogenes appear to play a role in the progression of normal follicular cells to follicular adenomas to follicular carcinomas. The pathway to papillary carcinoma is less well understood, and no benign precursor lesion has been identified, although progression from a normal follicular cell to microcarcinoma to papillary carcinoma has been related to ret oncogene rearrangements (Williams, 1995; Segev et al., 2003). Further research in molecular genetics is needed to better understand the pathogenic events that may link benign and malignant thyroid diseases. Despite numerous studies, the role of hyperthyroidism in the etiology of thyroid cancer remains uncertain. A history of hyperthyroidism was linked to thyroid cancer in the pooled analysis, but risks were reduced when adjustment was made for a history of goiter (Franceschi et al., 1999). Risks were highest within 2 years after diagnosis of hyperthyroidism and they decreased with increasing age at diagnosis. Other studies have also reported elevated, but often non-significant, risks of thyroid cancer associated with a history of hyperthyroidism (From et al., 2000; Mack et al., 1999; Mellemgaard et al., 1998; Memon et al., 2002b; Takezaki et al., 1996). Hypothyroidism has not been related to thyroid cancer (Franceschi et al., 1999; Iribarren et al., 2001; Mack et al., 1999; Memon et al., 2002b). Autoimmune thyroiditis (Hashimoto thyroiditis) is frequently accompanied by years of subclinical hypothyroidism with elevated TSH levels, but it has been difficult to evaluate reports of thyroid cancer coexisting with autoimmune thyroiditis. However, a significant excess risk of thyroid lymphomas only was reported in two prospective studies of thyroiditis patients (Aozasa, 1990; Holm et al., 1985), whereas no association was suggested in case-control studies of thyroid cancer (Crile and Hazard, 1962; Maceri et al., 1986; Memon et al., 2002b).

Associated Neoplasms and Other Diseases Several studies have reported an excess of thyroid cancer among patients with breast cancer (Harvey and Brinton, 1985; Iwasa et al., 1986; Ron et al., 1984; Schenker et al., 1984; Schottenfeld and Berg, 1971; Teppo et al., 1985), as well as an excess of breast cancer among women with thyroid cancer (Adjadj et al., 2003; Chen et al., 2001; Ron et al., 1984; Ronckers et al., 2005; Rubino et al., 2002; Simon et al., 2002; Vassilopoulou-Sellin et al., 1999). These results, however,

Table 50–5. Relative Risk of Thyroid Cancer Associated with Prior Benign Thyroid Disease, International Pooled Thyroid Cancer CaseControl Study* Females

associated with a history of benign thyroid nodules and goiter were over 10 and 5, respectively, and the association was found for both papillary and follicular carcinomas. Population attributable risks of 17% and 26% have been reported (Memon et al., 2002b; Ron et al., 1987). In the pooled analysis, risks were extremely high 2–4 years after a diagnosis of benign thyroid nodule, but remained elevated 10 years later, indicating that active medical follow-up is not the sole reason for the increased risk (Francheschi et al., 1999). Although casecontrol studies are always subject to recall bias, the strength of the association, the consistency of the results over all studies, and the lack of an association with other thyroid diseases, support the view that a

Disease Nodule Goiter Hyperthyroidism Hypothyroidism

Males †‡

Cases/Controls*

OR (95% CI)

Cases/Controls

OR (95% CI)

148/8 178/56 58/60 82/141

29.9 (14.5–62.0) 5.9 (4.2–8.1) 1.4 (1.0–2.1) 0.9 (0.7–1.3)

18/0 20/1 9/5 3/2

4 (9.2–4)** 38.3 (5.0–291) 3.1 (1.9–9.8) 1.7 (0.3–11.7)

Source: Data from Franceschi et al., 1999. *Number of cases and controls who have the disease. † Risk estimates from conditional logistic regression conditioned on study and age, and adjusted for history of radiation and age. ‡ OR, odds ratio CI, confidence interval. **CI based on Fisher exact test.

Thyroid Cancer have not been replicated in other studies (Hall et al., 1990; Hemminki and Jiang, 2001; Sadetzki et al., 2003). Although thyroid and breast cancers do not share many of the same risk factors, both sites are very susceptible to the carcinogenic effects of radiation, as well as to hormonal variations. Thyroid cancer patients have been reported to have a high risk of secondary malignancies, particularly among young patients (Rubino et al., 2003). Excess risks have been noted for leukemia, non-Hodgkin lymphoma, melanoma, and cancers of the salivary gland, digestive tract, kidney, bone and connective tissue, male genital organs, brain and nervous system, and endocrine glands (Hall et al., 1990; Ronckers et al., 2005; Rubino et al., 2003; Teppo et al., 1985). The mechanisms are not entirely clear, but 131I treatment may play a role in development of leukemia and certain other cancers (Ronckers et al., 2005; Rubino et al., 2003). In the opposite direction, elevated risks of thyroid cancer have been reported among patients with childhood cancer (Inskip, 2001; Sigurdson et al., 2005; Tucker et al., 1991), renal cancer (Teppo et al., 1985) and Hodgkin disease (Greene and Wilson, 1985; Hancock et al., 1991), which appear to be due at least partly to radiation treatment for the initial cancer. Hyperparathyroidism has been linked to thyroid cancer in some studies (Al-Jurf et al., 1979; Kaplan et al., 1971; LiVolsi and Feind, 1976), perhaps related to the fact that radiation exposure can induce both conditions. Associations have been reported also with acromegaly (Balkany et al., 1995; Barzilay et al., 1991; Baris et al., 2002), familial polyposis coli (Bülow et al., 1988; Hizawa et al., 1997), and ataxia-telangiectasia (Narita and Takagi, 1984; Ohta et al., 1986). Previous tonsillectomy, allergy, and skin disorders have been suggested as risk factors (Bross et al., 1971), but not substantiated in other studies (Kolonel et al., 1990; Levi et al., 1991; Ron et al., 1987).

Familial Aggregation and Heritable Conditions Medullary Thyroid Cancer. Approximately 20% of all medullary thyroid carcinomas are familial (Block et al., 1967). There is an autosomal dominant pattern of inheritance with a penetrance of nearly 100%. The tumor can occur alone or as part of the multiple endocrine neoplasia (MEN) syndrome type 2. MEN2A features medullary thyroid cancer with hyperparathyroidism, and results from inherited mutations at one of several specific cysteine residues of the ret gene. When familial medullary thyroid cancer occurs alone, the condition is caused by the same ret mutations. MEN2B is a multisystem syndrome including medullary thyroid cancer and mucosal neuromas on the lips and throughout the gastrointestinal tract. Almost all cases are associated with a specific cysteine mutation in the ret gene. Based on the aggressive behavior of the medullary cancer in MEN2B, treatment guidelines suggest very early thyroidectomy for these cases (Brandi et al., 2001). With the identification of the causative ret mutations, it is now possible to accurately determine whether or not medullary thyroid cancer cases are components of a heritable syndrome, which tends to occur at an earlier age than sporadic cases and to be multicentric in origin (Jackson et al., 1979).

Non-Medullary Thyroid Cancer. Papillary and follicular thyroid cancers are most often sporadic, but in about 3% to 7% of cases familial or genetic factors are evident (Goldgar et al., 1994; Galanti et al., 1997; Hemminki and Dong, 2000; Ozaki et al., 1988; Pal et al., 2001; Ron et al., 1991; Stoffer et al., 1986). Familial nonmedullary thyroid cancer tends to be more aggressive than sporadic tumors, with higher reported rates of distant metastases and recurrences (Grossman et al., 1995; Lupoli et al., 1999). Papillary thyroid cancer may occur as part of the familial adenomatous polyposis (FAP) syndrome, a dominantly inherited disease caused by germline mutations in the APC gene (Bülow et al., 1988; Hizawa et al., 1997; Plail et al., 1987). Excess risks for thyroid cancer of over 100 have been noted among FAP patients (Bülow et al., 1988; Hizawa et al., 1997). Thyroid cancers associated with this syndrome tend to occur before age 35, are often multicentric, and are more frequent among women. A sporadic form of papillary thyroid tumors that

983

has morphological features similar to FAP-associated thyroid cancer has been described (Cameselle-Teijeiro and Chan, 1999). Somatic mutations in the adenomatous polyposis coli (APC) gene have been found in some of these tumors (Ng et al., 2003; Kameyama and Mukai, 2004). In addition, benign and malignant thyroid tumors are associated with Cowden disease, a rare autosomal dominant disorder that predisposes also to breast cancer (McIver et al., 2002), and results from mutations in the tumor suppressor gene PTEN.

Environmental Factors Radiation Radiation is one of the few risk factors clearly associated with thyroid cancer. Between 1920 and 1960, radiotherapy was widely used to manage benign conditions, such as tinea capitis, enlargement of the thymus gland, cervical lymphadenopathy, tonsillar hypertrophy, and skin disorders. These study populations, along with patients treated for a first primary cancer, and survivors of the atomic bombings in Hiroshima and Nagasaki, have contributed much to what is known about radiation-induced thyroid cancer. Although the studies employed different methodologies and included populations from many countries exposed to a broad range of doses, all have demonstrated significantly increased risks of thyroid carcinomas following radiation exposure during childhood. In contrast, exposure during adulthood, to either external radiation or internal 131I, has not been linked convincingly to thyroid cancer (UNSCEAR, 2000).

Acute External Exposures. Based on a small case series, an association between radiation and thyroid cancer was first suggested by Duffy and Fitzgerald (1950). Soon after, Clark (1955) found that in a series of 15 children with thyroid cancer all had previous irradiation to their head or chest. In 1970 Winship and Rosvoll (1970) summarized the literature on childhood thyroid cancer and found that over 70% of the patients had a history of radiation exposure. The first comprehensive study of this subject was started in the 1950s by Hempelmann and colleagues (Hempelmann et al., 1975; Shore et al., 1985). Since then, numerous epidemiologic studies have shown that the thyroid gland of children is highly sensitive to the carcinogenic effects of acute, external low-LET ionizing radiation (Committee on the Biological Effects of Ionizing Radiations [BEIR], 1990; Shore, 1992; UNSCEAR, 2000). It has been suggested that the extreme radiosensitivity of the thyroid is related to its superficial site, its high degree of oxygenation, and the high rate of cell division (Barnes, 1988). A pooled analysis of seven studies of external radiation and thyroid cancer was undertaken to better quantify the radiation risk and to evaluate modifying influences on risk (Ron et al., 1995). Seven studies contributing about 700 thyroid cancers and almost 120,000 persons were included in the analyses (Boice et al., 1988; Pottern et al., 1990; Ron et al., 1989; Schneider et al., 1993; Shore et al., 1985; Thompson et al., 1994; Tucker et al., 1991). A significant dose-response with childhood exposure was observed in each study. At doses as low as 0.10 gray (Gy) the dose-response relationship was consistent with linearity, and only at doses above 10 Gy, where cell killing may be significant, did risk appear to level off. The excess relative risks (ERR) per Gy in these studies ranged from 1.1 (95% CI: 0.4, 29.4) following high-dose radiotherapy for childhood cancer, to 32 (95% CI: 14.0, 57.1) following relatively low-dose treatment for tinea capitis. When data for childhood exposure from the five cohort studies were combined, the pooled ERR/Gy was 7.7 (95% CI: 2.1, 28.7). Persons exposed as young children (<5 years) had an ERR more than two times higher than children exposed between 5 and 15 years of age. The pooled ERR/Gy was two times higher for women than for men (p = 0.07), but the gender results from the individual studies were inconsistent. The ERR was highest 15–29 years after exposure, but continued to be elevated at 40 or more years. For childhood exposure, the pooled excess absolute risk was 4.4 per 104 person-year (PY) Gy (95% CI: 1.9, 10.1) and the attributable risk per Gy was 88%. In contrast, the ERRs/Gy were not significantly elevated in the two studies of adult exposure (Fig. 50–6). More recent studies and reviews tend to confirm results from the pooled analysis (Acharya et al., 2003; De Vathaire et

984

PART IV: CANCER BY TISSUE OF ORIGIN

Figure 50–6. Pooled fitted thyroid cancer dose-response curves from five cohort studies of childhood exposure (<15 years) and two studies of adult exposure (≥15 years). (Source: Adapted from Ron et al., 1995.)

al., 1999; Inskip et al., 2001; Lundell et al., 1994; Rubino et al., 2002). Papillary carcinomas continue to be the main histologic type found after radiation exposure, risk remains high throughout life but may stabilize or decline after reaching a peak at about 15–30 years (Shore and Xue, 1999), females have a higher radiation risk than males, and based on limited data, the risk of thyroid cancer may be reduced slightly when exposure is fractionated (De Vathaire et al., 1999; Ron et al., 1995). At very high doses, such as those used to treat childhood cancer, there is a levelling off of risk which is consistent with an effect of cell killing (Sigurdon et al., 2005; Tucker et al., 1991).

Protracted Exposures. The widespread medical use of 131I, as well as concerns about the long-term health effects from the 1986 Chernobyl nuclear reactor accident, radioactive fallout from nuclear bomb testing, and the releases of 131I from the Hanford nuclear facility, have focused attention on radioactive iodides. In the United States, 131 I is the treatment of choice for hyperthyroidism, and therefore evaluating the associated health effects has clinical importance. In animals, experiments using high doses have indicated that 131I is much less effective than X-rays in the induction of thyroid tumors (NCRP, 1985). However, a large study of 3000 female rats suggested that at doses of about 3–4 Gy, the risks from external X-rays and 131I may be roughly equal (Lee et al., 1982). Human studies comparing 131 I to external gamma or X-ray exposure have been limited and equivocal. In the older literature, 131I was estimated to be from one-fifteenth to one-half as effective in inducing thyroid cancer as external Xradiation (NCRP, 1985). The difference in the carcinogenic effects of the two types of radiation was thought to be due to the lower dose rate of 131I, which allows time for repair of radiation damage, but more recent epidemiologic data raise doubts about this conclusion. In a follow-up of about 35,000 hyperthyroid patients treated between 1946 and 1964, Dobyns et al. (1974) reported that through 1968 there was no significant excess of thyroid cancer among the 131Itreated patients. Additional follow-up of the original hyperthyroid patient cohort and a new study of hyperthyroid patients suggested a small, but significant, increased risk of thyroid cancer among 131Itreated patients (Franklyn et al., 1998, 1999; Ron et al., 1998), whereas another from Sweden found that the risk of thyroid cancer was not significantly elevated 10 or more years after treatment (Holm et al., 1991; Hall et al., 1992). Because these studies are composed of persons with underlying thyroid disease and do not include childhood exposure, their results can not easily be extrapolated to children with normal thyroids. Continued follow-up of these cohorts is needed. To date, no significant excess risk of thyroid cancer has been demonstrated in patients examined with 131I for suspected thyroid or other disorders, even though the mean thyroid doses are on the order

of 0.5–1 Gy (Dickman et al., 2003; Hahn et al., 2001; Holm et al., 1988; Glöbel et al., 1984). However, despite the fact that it is this dose range that is most important from a public health standpoint, the risks of thyroid cancer following diagnostic doses of 131I have not been studied extensively and very few people exposed as children have been studied. Although the data on therapeutic and diagnostic 131I exposure suggest that in humans 131I has less of a carcinogenic effect than external X-rays, there are no definitive risk estimates. The fact that most persons exposed to 131I were adults with thyroid disease, whereas the majority of persons exposed to X-rays were children with normal thyroids, complicates the comparison. Large accidental releases of 131I have occurred as a result of nuclear weapons testing, periodic emissions releases from nuclear weapons plants, and accidents at nuclear power facilities. During the 1950s and 1960s, a large number of above-ground nuclear weapons tests were conducted in the United States. The fallout contaminated the entire country at various levels, with exposures particularly high in Utah. The latest follow-up (Bouville et al., 2002; Kerber et al., 1993) of children living at the time of the tests in highly contaminated areas in Utah and non-exposed neighboring regions observed a significant excess for benign and malignant thyroid neoplasms combined, but the dose response for malignant tumors alone did not reach statistical significance. In 1954, one of the tests in the Pacific Ocean accidentally exposed people living in the Marshall Islands to large doses of both short- and long-lived radioisotopes and gamma radiation (Conard, 1984; Hamilton et al., 1987; Robbins and Adams, 1989; Takahashi et al., 2003). The Marshall Islanders experienced an excess of thyroid abnormalities including cancer at thyroid doses estimated as between 7 and 14 Gy for young children and approximately 20 Gy for infants. Because 80%–90% of the dose absorbed by the thyroid was from the short-lived radioisotopes, this study does not provide specific information on the risks from 131I, but large increases in thyroid cancer occurred among exposed Islanders. The carcinogenic potential of continuous low-dose radionuclide exposure has been studied among persons living near nuclear plants and waste sites. The largest evaluation of such exposures is of people living near the Hanford nuclear site (Davis et al., 2004; Kopecky et al., 2004). In total, 3441 people born between 1940 and 1946 in seven counties in eastern Washington State participated in the study. Thyroid doses were estimated for the 3193 study participants who lived near Hanford during the time of the radioiodine releases, whereas the remaining 248 participants had moved from the Hanford area and received little or no exposure. Doses for the people who continued to live near Hanford ranged from 0–2.84 Gy (median 0.10 Gy), with only a small percentage having doses in the higher end of the range. No evidence of a dose-response relationship was found for malignant or benign nodules. A review of the study by a committee of the National Academy of Sciences (NAS) highlighted the large uncertainties in the dose estimates (NAS, 2000), but taking these into account does not appear to change the negative results materially. Following the Chernobyl reactor accident in 1986, radioactive nuclides were released into the atmosphere resulting in millions of people being exposed to substantial internal and external radiation to the thyroid in the most contaminated areas of the former Soviet Union (Hatch et al., 2005; Likhtarov et al., 2005; Moysich et al., 2002; UNSCEAR, 2000; Yamashita et al., 2002). Less than 5 years after the accident, childhood thyroid cancer, particularly of the papillary type, increased dramatically in Belarus and Ukraine (Abelin et al., 1994; Baverstock et al., 1992; Kazakov et al., 1992; Prisyazhiuk et al., 1991). In the following decade, studies of children exposed to radiation from Chernobyl provided compelling, but largely circumstantial, evidence of an association between pediatric thyroid cancer incidence and 131I dose to the thyroid (Astakhova et al., 1998; Buglova et al., 1996; Likhtarev et al., 1995; Stsjazhko et al., 1995; Tronko et al., 1996). Excess thyroid cancer continues to occur among people living in Belarus, Ukraine, and the contaminated region of Bryansk, Russia (Ivanov et al., 2004; Kenigsberg et al., 2002; Shibata et al., 2001; Tronko et al., 1999), but the magnitude and patterns of the dose-

Thyroid Cancer response relationship have not yet been firmly established. This is due in part to the difficulty in reconstructing radiation doses for individuals, but recent international efforts have resulted in improved dose estimates that can be used for risk assessment (Gavrilin et al., 2004; Likhtarov et al., 2005; Stepanenko et al., 2004). Although findings from recent investigations suggest that age at exposure influences risk, the role of attained age, time since exposure, gender and iodine deficiency, and ethnic differences is unclear. Jacob et al. (2002) examined the modifying effects of age at exposure and age at the time of thyroid surgery, and found that they were different in Belarus and Ukraine. It is not yet known whether these disparities stem from variations in radiation or other exposures, radiation sensitivity, or diagnosis and reporting. Data on adult exposure are severely limited; one study conducted in Bryansk, Russia, reported a small increase in risk, which was not dose-related (Ivanov et al., 2003). The future level of risk is a concern because of the apparently large risk involving a large population. Assuming that the pattern is similar to that for external radiation (Ron et al., 1995; Shore and Xue, 1999), excess cancers will continue to occur for several more decades, but whether the excess relative risk will increase or stabilize over time must be determined. A radiation-associated increase in thyroid cancer incidence has been shown for women and men, with no significant difference in excess relative risk. This means that 2–3 times more thyroid cancers should be expected in women due to their higher incidence of spontaneous cancers (Tronko et al., 1999; Jacob et al., 1999). Because iodine deficiency influences both thyroid dose and function, it has been hypothesized that people living in iodine-deficient regions, such as the vicinity of Chernobyl, would be more susceptible to radiation carcinogenesis than persons living in iodine-sufficient regions (Gembiecki et al., 1997). This relationship has been examined in two recent studies and 2–3 fold risks were observed in highly deficient areas compared with other regions (Cordis et al., 2005; Shakhtarin et al., 2003). Although variation in thyroid cancer incidence related to ethnic differences has been postulated, data regarding Chernobyl are lacking. Long-term follow-up of the health effects of Chernobyl will help fill the many gaps in our knowledge regarding 131 I. Studies of residents living in areas of high natural background radiation may provide an opportunity to evaluate risks from protracted, low-dose radiation. Results to date have been negative, partly because of their low statistical power and the small differences in dose between the high and normal background areas. Among people living in high background areas in India and China, there was no evidence of an enhanced risk of thyroid cancer (Pillaj et al., 1976; Wang et al., 1990b). Low-dose fractionated occupational exposures are relevant for setting radiation exposure standards. Unfortunately, data from studies of nuclear workers have yielded little information in regard to thyroid cancer for several reasons: the rarity of the disease, the predominance of male workers studied, and the reliance on mortality data, which are of limited value in evaluating thyroid cancer. Given these limitations, the majority of studies have not found a statistically significant relation between occupational exposure and thyroid cancer. Studies of medical radiation workers have yielded new information on occupational risks from ionizing radiation that should be further investigated. Based on only eight cases, Chinese medial radiation workers who were employed before 1960 had a non-significant excess incidence of thyroid cancer (RR = 1.7; 95% CI: 0.6–4.7) (Wang et al., 1990a). Among US radiologic technologists, thyroid cancer incidence was higher than expected compared with US population rates (Sigurdson et al., 2003). Similarly, in a hypothesis-generating recordlinkage study, a significant twofold risk of thyroid cancer was seen among Swedish X-ray operators and laboratory assistants compared with the general population (Carstensen et al., 1990). The results of these studies should be interpreted with caution because individual doses were not available, multiple comparisons were made, comparing workers to a general population is not ideal, and the risks were not always significant. Chernobyl clean-up workers were exposed to varying levels of external gamma radiation depending on their specific job and how soon and how long after the accident they worked. To date, studies of

985

these workers have not found significantly elevated risks of thyroid cancer (Inskip et al., 1997; Ivanov, 2002). Even though radiation-induced tumors are more often bilateral (Samaan et al., 1987; Schneider et al., 1985, 1986a; Turrin et al., 1985) and multicentric (Schneider et al., 1985, 1986a), the clinical course of external radiation-associated cancers is similar to spontaneous cancers (Schneider et al., 1986b; Rubino et al., 2002). Well-differentiated papillary cancer is the principal cell type induced by radiation, but an increase in follicular and anaplastic carcinoma may occur as exposed persons reach the natural ages to develop these types (DeGroot et al., 1983; Ron et al., 1989, 1995; Samaan et al., 1987; Schneider et al., 1985, 1993). Although it is too soon to be sure, the clinical outlook may be different for the cases occurring in the Chernobyl region. They tend to present with advanced features (Reiners et al., 2002) and may be more difficult to ablate with radioactive iodine (Oliynyk et al., 2001; Reiners et al., 2002). These characteristics may not be features of 131I exposure, but rather of the young ages of the cases and the iodine deficiency that is prevalent in the area.

Diet Over the past few decades, diet has become a focus of epidemiologic investigations of thyroid cancer. Early studies suggested that dietary iodine (Williams et al., 1977), goitrogens (Hempelmann and Furth, 1978), calcium and vitamin D (Williams, 1979), and alcohol (Breslow and Enstrom, 1974; Williams, 1976) were possible risk factors. More recent studies have used fish and seafood as proxy measures of iodine intake and have looked for potentially protective effects of vegetables and soy products. In the recent pooled analysis of thyroid cancer casecontrol studies, 13 of the 14 studies had collected some data on diet. Reports, based on 2497 cases and 4337 controls, on the consumption of fish and shellfish (Bosetti et al., 2001) and fruits and vegetables (Bosetti et al., 2002) were published.

Iodine. Iodine, used to synthesize thyroid hormones, is essential in the regulation of normal thyroid hormone metabolism. Iodine deficiency is associated with a broad spectrum of benign thyroid diseases, but its role in thyroid carcinogenesis is uncertain. Rodents fed iodinedeficient diets have shown elevated rates of thyroid cancer, accompanied by chromosomal aberrations (Feldt-Rasmussen, 2001), and iodine deficiency has been shown to act as both a carcinogen and a tumor promoter in laboratory animals (Isler, 1959; Ohshima and Ward, 1986; Ward and Ohshima, 1986). In humans, the role of iodine is not so clear. Following an early report (Wegelin, 1928) that the prevalence of thyroid cancer at autopsy was 10 times higher in Bern, an endemic goiter area, than in Berlin, it was suggested that endemic goiter and/or iodine deficiency may be risk factors for thyroid cancer. However, studies of the correlation between iodine levels (or endemic goiter areas) and thyroid cancer incidence or mortality rates demonstrated elevated rates also in iodine-rich regions where goiters were uncommon (Feldt-Rasmussen, 2001; Franceschi et al., 1989; Williams et al., 1977). These contradictory findings may stem from differences in tumor histology. In rats, an iodine-deficient diet was associated with follicular cancers, whereas iodine excess was related to papillary cancers (Axelrod and Leblond, 1955). In humans, endemic goiter areas often are associated with an elevated risk of follicular and, perhaps, anaplastic thyroid cancers, whereas iodine-rich areas have been linked to papillary carcinoma (Bacher-Stier et al., 1997; Doniach, 1969, 1971; Feldt-Rasmussen, 2001; Franceschi, 1998; Lind et al., 1998; Pettersson et al., 1996; Williams et al., 1977). In Transvaal, South Africa, 58% of the thyroid cancer diagnosed among blacks living in rural areas was follicular, whereas papillary carcinoma was most frequent among blacks and whites living in urban areas (Kalk et al., 1997). Differences in iodine status were thought to account for this pattern. Similarly, the ratio of papillary to follicular carcinoma increased after the introduction of dietary iodine supplementation in Argentina and Austria (Harach et al., 2002; Lind et al., 2002). In Poland, thyroid cancer incidence rose following the cessation of iodine supplementation. After restarting iodine supplementation, the overall incidence of thyroid cancer remained stable, but the ratio of papillary to follicular

986

PART IV: CANCER BY TISSUE OF ORIGIN

cancers increased significantly (Huszno et al., 2003; Szybinski, 2003). Thus, a change in iodine intake may increase the frequency of one cell type while decreasing the frequency of the other, resulting in little or no net change in incidence. In addition, iodine supplementation may be associated with better medical services and, therefore, an improved chance to diagnose papillary carcinoma.

Dietary Iodine, Fish, and Seafood. Although some epidemiologic studies have tried to assess dietary intake of iodine in thyroid cancer etiology, others studies have used frequency of fish and shellfish consumption as a proxy measure of exposure. Using a detailed quantitative dietary history questionnaire, Kolonel et al. (1990) reported a non-significant positive association between iodine intake and thyroid cancer, while Horn-Ross et al. (2001) reported a diminution in papillary thyroid cancer risk associated with iodine intake, largely due to the greater use of iodine-containing multivitamins by controls than cases. In the same study, a 20% reduction in risk was observed among individuals in the highest quintile of iodine measured in toenail clippings. As expected based on correlational data, the mean iodine level as measured in toenails, was significantly lower among the cases with follicular carcinoma than the controls, but no difference was seen based on reported dietary intake. The inconsistency in these findings may reflect the difficulties in quantifying iodine intake. In the pooled analysis (Bosetti et al., 2001), total fish consumption overall was not associated with thyroid cancer risk, even taking histologic type and gender into account. There was, however, heterogeneity (P = 0.16) among studies, with a suggestion of a protective effect for fish consumption in studies conducted in endemic goiter areas (ORs = 0.91; 95% CI: 0.75, 1.1 and 0.65; 95% CI: 0.48, 0.88 for medium and high fish consumption compared with low fish consumption, respectively). In studies conducted in iodine-sufficient areas, no appreciable effect was seen. Risks were close to unity for high consumption of iodine-rich salt-water fish and shellfish. These data indicate that high intake of fish is not a major contributor to overall thyroid cancer risk and may even be protective in endemic goiter regions. In a recent case-control investigation conducted in Kuwait, fresh fish, which are eaten frequently in this coastal country, was somewhat protective, but canned or frozen fish conferred a threefold risk (Memon et al., 2002b). The authors suggested that the elevated risk associated with processed fish may be related to higher iodine levels in processed fish or to altered iodine uptake from additives or the processing itself. Vegetables, Fruits, and Grains. High vegetable consumption has been reported to have a protective effect on a wide range of tumors (Negri et al., 1991) and this effect may extend to the thyroid gland. Special attention has been given to the potential role of cruciferous vegetables. On the one hand, vegetables contain thioglycosides that may be degraded to form goitrogens with marked antithyroid activity. These goitrogens have been implicated in the pathogenesis of endemic goiter (Gaitan, 1997). In laboratory animals, goitrogens and cruciferous vegetables promote thyroid cancer (Kanno et al., 1990; Birt et al., 1987). Cruciferous vegetables, however, also contain flavones, isothiocyanates, and phenols that may inhibit the development of certain cancers by increasing the activity of enzymes involved in the detoxification of carcinogens (Steinmetz and Potter, 1996; Wattenberg, 1983). Furthermore, carotenoids and other antioxidants are found in cruciferous vegetables (Steinmetz and Potter, 1996). In the pooled thyroid cancer analysis (Bosetti et al., 2002), cases consumed slightly less cruciferous vegetables than controls, but there was heterogeneity among studies largely due to the opposite effect found in Japan. These findings were similar in iodine-deficient and iodinesufficient areas. For high consumption of non-cruciferous vegetables, a significant protective effect was observed (odds ratio for highest vs. lowest intake = 0.82; 95% CI: 0.69, 0.98). Results concerning vegetable intake have not been consistent and some studies have reported no, or even elevated, risks associated with cruciferous or other vegetables (FrentzelBeyme and Helmert, 2000; Memon et al., 2002b). In a study that found

significantly increased risks associated with high consumption of tomatoes and red peppers, the authors suggested that pesticide exposures may be related to risk (Frentzel-Beyme and Helmert, 2000). In China, a lower risk of thyroid nodules was associated with high intake of allium vegetables such as garlic and onions (Wang et al., 1990b). Flavonoids are compounds that occur naturally in a wide variety of fruits, vegetables, and grains. They can inhibit enzymes that are important in thyroid hormone synthesis and deiodination (Gaitan, 1997), but they also have antiproliferative effects on tumors (Yin et al., 1999). Structurally, certain flavonoids are similar to estrogenic steroids whereas some appear to have antiestrogenic activity, similar to tamoxifen. In human thyroid cell lines, flavonoids have shown strong antiproliferative activity, but the degree of activity varied by thyroid cell type and specific flavonoids (Yin et al., 1999). Phytoestrogens are natural estrogenic compounds that also have antiestrogenic effects and may have chemopreventive properties (Adlercreutz et al., 1995). Tofu and other soy-based products are the major components of phytoestrogens, but these products also are goitrogens. In Asians and vegetarians, phytoestrogen-rich foods are an important part of the diet. In a multi-ethnic case-control study of thyroid cancer among women in San Francisco, consumption of high soy-based foods lowered the risk of thyroid cancer among both white and Asian women (Haselkorn et al., 2003; Horn-Ross et al., 2002). These results are consistent with in vitro studies of the larger class of flavonoids (Yin et al., 1999), but since the human data are so limited, further studies are needed to better understand this relationship.

Trace Metals. Selenium is important in thyroid hormone metabolism because it is essential for some of the (deiodinase) enzymes that remove iodine from thyroxine and its metabolites. It also helps prevent damage to the thyroid gland from too much iodine (Zimmermann and Kohrle, 2002). Glattre et al. (1989) studied 43 thyroid cancer cases and 129 controls selected from a large Norwegian prospective study. Serum selenium was significantly higher among the controls, but this effect was limited to selenium levels measured less than 7 years before diagnosis of thyroid cancer. Dietary information was available for a subset of the study population, but no difference in selenium intake was observed between cases and controls. A protective role of selenium also was suggested by the report of an elevated thyroid cancer risk among hemodialysis patients, who are known to have impaired selenium status (Zimmermann and Kohrle, 2002). Vitamins. In a case-control study of thyroid cancers the use of vitamin D, but not calcium supplements, was significantly linked with medullary carcinoma, which arises from calcitonin-secreting C-cells (Ron et al., 1987). Although serum calcium levels influence C-cell production of calcitonin, studies in laboratory animals showed that dietary calcium did not elevate the incidence of medullary-like tumors, whereas vitamin D increased both calcitonin production and C-cell hyperplasia, a potential precursor lesion (Williams et al., 1977; Williams, 1979). Smoking, Alcohol, and Coffee Cigarette smoking has multiple effects on the thyroid gland and appears to act differently depending on thyroid status (Utiger, 1995). In 12 of the 14 studies in the pooled analysis, an inverse association between ever smoking and papillary and follicular thyroid cancers was demonstrated, and the risk was significantly lower among current smokers compared with never smokers (OR = 0.6; 95% CI: 0.6, 0.7) (Mack et al., 2003). A reduced risk was seen for both men and women, and a clear inverse dose-response was evident; thyroid cancer risk decreased with increasing pack-years, number of cigarettes per day, and years smoked. Reduced thyroid cancer risks were also found for smokers in two other recent studies (Kreiger and Parkes, 2000; Rossing et al., 2000), although results have not been entirely consistent (Iribarren et al., 2001; Memon et al., 2000b). Three biological mechanisms have been proposed for the inverse association with smoking: a smoking-related reduction in TSH, a smoking-related reduction in body weight, which appears to confer a small protective

Thyroid Cancer effect (Dal Maso et al., 2000), and anti-estrogenic effects of smoking. Although these hypotheses merit further investigation, the currently available data do not convincingly support any of these pathways. A geographic correlation between US thyroid cancer mortality and alcohol consumption was first reported by Breslow and Enstrom (1974) and subsequently supported by an analysis of interview data from the Third National Cancer Survey, which suggested a positive association between thyroid cancer and alcohol intake. Williams (1976) further suggested that alcohol could promote thyroid cancer by inducing TSH secretion. In recent epidemiologic studies, however, results have been inconsistent (Frentzel-Beyme and Helmart, 2000; Iribarren et al., 2001). In a Greek study of 61 thyroid cancer cases and an equal number of controls, controls drank significantly more coffee than cases (Linos et al., 1989). There also was some evidence of a protective effect of coffee from a pooled Italian and Swiss analysis (Franceschi et al., 1991), and from studies in Germany (Frentzel-Beyme and Helmart, 2000) and Japan (Takezaki et al., 1996). In the German study, a greater consumption of decaffeinated coffee was reported by more cases than controls.

Occupation and Chemical Exposures Other than radiation, the role of occupational determinants of thyroid cancer etiology has received little attention. In a large Canadian study (1272 cases and 2666 controls), employment in wood processing, pulp and papermaking was associated with a 2.5-fold risk (95% CI: 1.11–5.83), and working in sales and service occupations had a slightly elevated risk (OR = 1.19; 95% CI: 1.00–1.41) compared with other occupations (Fincham et al., 2000). In addition, some studies of thyroid cancer have suggested an effect of occupational exposure to hydrocarbons (Carstensen et al., 1990; Divine et al., 1987), silica (Fillmore et al., 1999), agricultural chemicals (Soki et al., 1994) and other chemicals (Hallquist et al., 1993; Soki et al., 1994; Zivaljevic et al., 2003). However, no significant associations between thyroid cancer and 11 occupational exposures were found in a large US cohort study (Iribarren et al., 2001). Although certain chemicals can disrupt the thyroid-pituitary axis, increase TSH secretion, and be directly mutagenic or tumor promoters in experimental systems, no chemical has been clearly identified as a thyroid carcinogen in epidemiologic studies.

Medications Epidemiologic surveys of prescription drug use have reported statistically significant associations between thyroid cancer and pentobarbital (barbiturate), meclizine (antihistamine used for motion sickness), diphenoxylate (antidiarrheal), dicyclomine (antispasmodic used for gastrointestinal disorders), griseofulvin (antifungal), bisacodyl (cathartic), and senna (cathartic) (Friedman and Ury, 1980, 1983; Selby et al., 1989). These associations have not been confirmed in subsequent analytic studies (Kolonel et al., 1990; Olsen et al., 1989; Ron et al., 1987), although use of spironolactone conveyed a fourfold risk of borderline significance in one study (Ron et al., 1987).

MOLECULAR PATHOGENESIS Somatic mutations that initiate benign and malignant thyroid neoplasms have been identified in many, but not all tumors. However, even when mutations are found, it is likely that additional genetic changes occur before a neoplasm develops fully. Rearrangements of the ret gene play a key role in the pathogenesis of papillary thyroid cancer (Kroll, 2002; Segev et al., 2003). The normal product of the ret gene is a transmembrane tyrosine kinase. Its extracellular domain forms part of a receptor complex for glial cellderived neurotropic growth factor, whereas its intracellular domain has tyrosine kinase activity. The ret gene has a promoter that is not active in thyroid follicular cells, so that ret is not expressed in normal thyroid follicular cells. The rearranged gene is referred to as PTC for papil-

987

lary thyroid cancer. The two most common rearrangement partners, H4 (to form PTC1) and ELE1 (to form PTC3 and PTC3 variants) are located on the same chromosome as ret. Less common rearrangements (PTC2 and others) occur between chromosomes. For all of the rearrangements the pattern is the same. The tyrosine kinase domain of ret is separated from the transmembrane and extracellular domains, and becomes attached to the promoter and N-terminal portion of the rearrangement partner. The promoter is active in thyroid cells and allows the ret tyrosine kinase to be expressed, where it is found intracellularly. The Nterminal portion provides a site for dimerization, necessary for activation of the kinase activity. Unopposed activation of the ret tyrosine kinase is sufficient to cause thyroid cancer, as demonstrated by the cancers produced when rearranged ret genes are introduced into the thyroid glands of rats (Jhiang et al., 1996). The fraction of papillary cancers with ret mutations has varied widely in reports from different areas (Segev et al., 2003). Although some surveys have found more than 50%, most fall into the range of 10%–30%. In children the frequency appears to be higher. In radiation-related thyroid cancer, particularly the papillary tumors in the Chernobyl region, the frequency of ret rearrangements is very high (Thomas et al., 1999). This is not surprising, as double-strand breaks in DNA, the predominant lesion caused by radiation, are likely to predispose to gene rearrangements. The proportion of rearrangements of the PTC3 type is higher than in other cases, although this may be a result of the young age of the patients. PTC3 has been associated with the solid variant of papillary cancer and may account for the aggressive nature of many of the cancers in the region (Nikiforov et al., 1997). Also, in the Chernobyl area cases, the spectrum of ret rearrangements is greater than in cases not related to radiation. Two research groups have investigated the exact DNA breakpoints in a small number of radiation-related cases of papillary cancer, in an attempt to deduce the mechanism of recombination (Nikiforov et al., 1999; Klugbauer et al., 2001). Unfortunately, results have been discordant. Studying the exact DNA breakpoints has led to an attractive hypothesis for explaining the sensitivity of the thyroid to radiation and the specificity of ret rearrangements to thyroid cancer. Based on sequencing data it is proposed that the DNA strands at the time of rearrangement have an antiparallel alignment (Nikiforov et al., 1999). Further, based on visualization studies, it has been proposed that ret and its rearrangement partners (or at least ELE1) are held in proximity to each other by the chromosomal suprastructure of the thyroid nucleus (Nikiforova et al., 2000). Mutations in other tyrosine kinase receptors have been found in papillary thyroid cancers. Recently, point mutations in the BRAF gene, part of the signaling pathway initiated by ret, have been found in papillary cancers (Cohen et al., 2003; Kimura et al., 2003). However, it should be stressed that most papillary cancers have not shown ret rearrangements or other identifiable mutations. Point mutations in one of the ras genes are found in follicular tumors, and appear to contribute to chromosomal instability in these tumors. The mutations are not thought to be related to radiation. Activation of ras (H-ras, K-ras, and N-ras) oncogenes has been found in follicular adenomas, follicular cancers and anaplastic cancers (Lemoine et al., 1988, 1989; Namba et al., 1990), suggesting that the ras oncogene affects the early stage of tumor development. Ras mutations are found in up to 50% of thyroid tumors, most commonly in follicular carcinomas (Karga et al., 1991; Lemoine et al., 1988, 1989; Suárez et al., 1990; Wright et al., 1989). In one study, the mutation rate of the ras oncogene in follicular tumors was substantially higher in an iodine-deficient area than an iodine-rich area, suggesting that iodine may influence oncogene activation (Shi et al., 1991). Recently, a genetic rearrangement has been found in many follicular neoplasms, particularly cancers, that juxtaposes the PAX8 gene, which plays an important role in thyroid gland development, with the PPAR gamma 1 gene (Kroll et al., 2000; Nikiforova et al., 2003). The role of this rearrangement in thyroid cancer and its relation to radiation are not clear as yet. Poorly differentiated and anaplastic thyroid cancers, as well as anaplastic thyroid cell culture lines, often harbor mutations in the p53

988

PART IV: CANCER BY TISSUE OF ORIGIN

gene (Ito et al., 1992; Fagin et al., 1993). Although p53 mutations occur in a wide variety of malignancies, its role in thyroid cancer, at what appears to be a late stage, remains to be clarified. Germline mutations of the ret proto-oncogene on chromosome 10 have been described in individuals with MEN2A, MEN2B, and familial medullary thyroid cancer (Donis-Keller et al., 1993; Hofstra et al., 1994; Mulligan et al., 1993). In families with these syndromes, detection of specific germline mutations of ret allows early identification of gene carriers and is used to screen at-risk children (Lips et al., 1994). Performing thyroidectomy on gene carriers at a very early age has been recommended as a method to prevent medullary thyroid cancer (Brandi et al., 2001).

FUTURE DIRECTIONS Despite recent improvements in understanding the etiology of thyroid cancer, several areas deserve further research. Although radiation has been studied extensively, the risks associated with adult exposure to external radiation and childhood exposure to 131I are not well described. Moreover, little is known about dose-rate effects in humans and additional data on the effects of radiation on gene regulation are needed. For epidemiologic studies, it would be useful to distinguish thyroid cancer cases related to radiation from those developing spontaneously, and molecular characteristics may eventually allow us to do so. A history of benign thyroid adenomas, nodules, and goiters has been linked to thyroid cancer and may contribute to the higher incidence of thyroid cancer among women, but the mechanisms involved in tumor induction and progression are unknown. Prospective studies will be needed to clarify the relationship between precursor lesions and thyroid cancer. Iodine is essential in regulating thyroid hormone metabolism, but its etiologic role in thyroid cancer is not clear, even among cases associated with goiter. Based on epidemiologic and experimental studies, other dietary factors may play a role, including a weak protective effect of cruciferous vegetables and an association with seafood intake in some studies; however, at present, dietary data are limited. Since thyroid cancer and other thyroid diseases are 2–3 times more frequent in women than men, it is perplexing that hormonal and reproductive factors seem to have only a small role in thyroid cancer etiology based on epidemiologic studies. The incorporation of metabolic profiles in epidemiologic studies should help clarify the influence of pituitary and sex hormones, as well as growth factors, on the development of benign and malignant thyroid lesions. On experimental grounds, high levels of TSH appear to increase the risk of thyroid cancer, but further studies in humans are needed. Epidemiologic research on specific histologic types of thyroid cancer has been limited because of relatively small numbers and unstable risk estimates. Large-scale studies with pooling of data sets are needed to identify risks for specific histologies, and special attention should be given to high-incidence populations in various parts of the world. Finally, genetic susceptibility may be a more important determinant of thyroid cancer, especially the papillary type, than once thought, and family-based studies should be helpful in this area. Recent studies of somatic mutations have also advanced our knowledge of thyroid carcinogenesis and the multistage mechanisms involved in tumor induction and progression. The development of molecular diagnostics for underlying genetic syndromes has led to early screening and treatment for affected family members, and should also be useful in further categorizing various subtypes of thyroid cancer. Continuing advances in developing molecular and metabolic probes that can be incorporated in population studies should illuminate our understanding of the origins of thyroid cancer and provide insights into preventive, diagnostic, and therapeutic interventions. References Abelin T, Egger M, Ruchti C. 1994. Belarus increase was probably caused by Chernobyl. Br Med J 309:1298.

Acharya S, Sarafoglou K, LaQuaglia M, et al. 2003. Thyroid neoplasms after therapeutic radiation for malignancies during childhood or adolescence. Cancer 97:2397–2403. Adjadj E, Rubino C, Shamsaldim A, et al. 2003. The risk of multiple promary breast and thyroid carcinomas—role of the radiation dose. Cancer 98:1309–1317. Adlercreutz CH, Goldin BR, Gorbach SL, et al. 1995. Soybean phytoestrogen intake and cancer risk. J Nutr 125:757S–770S. Akslen LA, Haldorsen T, Thoresen SØ, et al. 1991. Survival and causes of death in thyroid cancer: A population-based study of 2479 cases from Norway. Cancer Res 51:1234–1241. Akslen LA, Nilssen S, Kvåle G. 1992. Reproductive factors and risk of thyroid cancer. A prospective study of 63,090 women from Norway. Br J Cancer 65:772–774. Al-Jurf A, Esselstyn CB, Crile G Jr. 1979. Thyroid lesions in patients with hyperparathyroidism. Int Surg 64:33–36. American Joint Committee on Cancer. 1997. Thyroid Gland. In: American Joint Committee on Cancer: AJCC Cancer Staging Manual. 5th ed. Philadelphia: Lippincott-Raven Publishers, pp. 59–64. Aozasa K. 1990. Hashimoto’s thyroiditis as a risk factor of thyroid lymphoma. Acta Pathol Jpn 40:459–468. Astakhova LN, Anspaugh LR, Beebe G, et al. 1998. Chernobyl-related thyroid cancer in children of Belarus: A case-control study. Radiat Res 150:349–356. Axelrod AA, Leblond CP. 1955. Induction of thyroid tumors in rats by a low iodine diet. Cancer 8:339–367. Bacher-Stier C, Riccabona G, Totsch M, et al. 1997. Incidence and clinical characteristics of thyroid carcinoma after iodine prophylaxis in an endemic goiter country. Thyroid 7:733–741. Balkany C, Cushing GW. 1995. An association between acromegaly and thyroid carcinoma. Thyroid 5:47–50. Baloch ZW, Livolsi VA. 2001. Prognostic factors in well-differentiated follicular-derived carcinoma and medullary thyroid carcinoma. Thyroid 11:637–645. Ballivet S, Salmi LR, Dubourdieu D, et al. 1995. Incidence of thyroid cancer in New Caledonia, South Pacific, during 1985–1992. Am J Epidemiol 141:741–746. Banu KS, Govindarajulu P, Aruldhas MM. 2001. Testosterone and estradiol have specific differential modulatory effect on the proliferation of human thyroid papillary and follicular carcinoma cell lines independent of TSH action. Endocr Pathol 12:315–327. Baris D, Gridley G, Ron E, et al. 2002. Acromegaly and cancer risk: A linkage study in Sweden and Denmark. Cancer Causes Control 13: 395–400. Barnes ND. 1988. Effects of external irradiation on the thyroid gland in childhood. Horm Res 30:84–89. Baverstock K, Egloff B, Pinchera A, et al. 1992. Thyroid cancer after Chernobyl. Nature 359:21–22. Barzilay J, Heatley GJ, Cushing GW. 1991. Benign and malignant tumors in patients with acromegaly. Arch Intern Med 151:1629– 1632. BEIR. 1990 (Committee on the Biological Effects of Ionizing Radiations): Health effects of exposure to low levels of ionizing radiation (BEIR V). Washington, DC, National Academy Press. Bernstein L, Miu A, Monroe K, et al. 1995. Cancer incidence among Filipinos in Los Angeles County, 1972–1991. Int J Cancer 63:345– 348. Birt DF, Pelling JC, Pour PM, et al. 1987. Enhanced pancreatic and skin tumorigenesis in cabbage-fed hamsters and mice. Carcinogenesis 8: 913–917. Block MA, Horn RC, Miller JM, et al. 1967. Familial medullary carcinoma of the thyroid. Ann Surg 166:403–411. Blot WJ, Le Marchand L, Boice JD Jr., et al. 1997. Thyroid cancer in the Pacific. J Natl Cancer Inst 89:90–91. Boice JD Jr, Engholm G, Kleinerman RA, et al. 1988. Radiation dose and second cancer risk in patients treated for cancer of the cervix. Radiat Res 116:3–55. Bosetti C, Kolonel L, Negri E, et al. 2001. A pooled analysis of case-control studies of thyroid cancer. VI. Fish and shellfish consumption. Cancer Causes Control 12:375–382. Bosetti C, Negri E, Kolonel L, et al. 2002. A pooled analysis of case-control studies of thyroid cancer. VII. Cruciferous and other vegetables (International). Cancer Causes Control 13:765–775. Bouville A, Simon SL, Miller CW, et al. 2002. Estimates of doses from global fallout. Health Phys 82:69–705. Brandi ML, Gagel RF, Angeli A, et al. 2001. Guidelines for diagnosis and therapy of MEN type 1 and type 2. J Clin Endocrinol Metab 86:5658–5671.

Thyroid Cancer Breslow NE, Enstrom JE. 1974. Geographic correlations between cancer mortality rates and alcohol-tobacco consumption in the United States. JNCI 53:631–639. Brinton LA, Melton J, Malkasian GD Jr, et al. 1989. Cancer risk after evaluation for infertility. Am J Epidemiol 129:712–722. British Thyroid Association and the Royal College of Medicine. 2002. Guidelines for the Management of Thyroid Cancer in Adults. London: Royal College of Physicians. Bross ID, Shimaoka K, Tidings J. 1971. Some epidemiological clues in thyroid cancer. Arch Intern Med 128:755–760. Buglova EE, Kenigsberg JE, Sergeeva NV. 1996. Cancer risk estimation in Belarussian children due to thyroid irradiation as a consequence of the Chernobyl nuclear accident. Health Phys 71:45–49. Bülow S, Holm NV, Mellamgaard A. 1988. Papillary thyroid carcinoma in Danish patients with familial adenomatous polyposis. Int J Colorect Dis 3:29–31. Burgess JR. 2002. Temporal trends for thyroid carcinoma in Australia: An increasing incidence of papillary thyroid carcinoma (1982–1997). Thyroid 12:141–149. Burguera B, Gharib H. 2000. Thyroid incidentalomas—Prevalence, diagnosis, significance, and management. Endocrinol Metab Clin North Am 29: 187–203. Camesslle-Teijeiro J and Chan JKC. 1999. Cribriform-morular variant of papillary carcinoma: A distinctive variant representing the sporadic counterpart of familial adenomatous polyposis-associated thyroid carcinoma? Mod Pathol 12:400–411. Cardis E, Kesminiene A, Ivanov V, et al. 2005. Risk of thyroid cancer after exposure to 131I in childhood. J Natl Cancer Inst 97:724–732. Carstensen JM, Wingren G, Hatschek T, et al. 1990. Occupational risks of thyroid cancer: Data from the Swedish Cancer Environment Register, 1961–1979. Am J Ind Med 18:535–540. Chen AY, Levy L, Goepfert H, et al. 2001. The development of breast carcinoma in women with thyroid carcinoma. Cancer 92:225–231. Clark DE. 1955. Association of irradiation with cancer of the thyroid in children and adolescents. JAMA 159:1007–1009. Cohen J, Xing MZ, Mambo E, et al. 2003. BRAF mutation in papillary thyroid carcinoma. J Nat Cancer Inst 95:625–627. Coleman MP, Esteve J, Damiecki P, et al. 1993. Trends in cancer incidence and mortality. IARC, Publication No. 121. Lyon, France: International Agency for Research on Cancer, pp. 609–640. Colonna M, Grosclaude P, Remontet L, et al. 2002. Incidence of thyroid cancer in adults recorded by French cancer registries (1978–1997). Eur J Cancer 38:1762–1768. Conard RA. 1984. Late radiation effects in Marshall Islanders exposed to fallout 28 years ago. In: Boice JD Jr., Fraumeni JF Jr., eds. Radiation Carcinogenesis: Epidemiology and Biological Significance. New York Raven Press, pp. 57–71. Correa P, Chen VW. 1995. Endocrine gland cancer. Cancer 75:338–352. Cotterill SJ, Pearce MS and Parker L. 2001. Thyroid cancer in children and young adults in the North of England. Is increasing incidence related to the Chernobyl accident? Eur J Cancer 37:1020–1026. Crile G Jr, Hazard JB. 1962. Incidence of cancer in struma lymphomatosa. Surg Gynecol Obstet 115:101–103. Dal Maso L, La Vecchia C, Franceschi S, et al. 2000. A pooled analysis of thyroid cancer studies. V. Anthropometric factors. Cancer Causes Control 11:137–144. D’Avanzo B, La Vecchia C, Franceschi S, et al. 1995. History of thyroid diseases and subsequent thyroid cancer risk. Cancer Epidemiol Biomarkers Prev 4:193–199. Davis S, Kopecky KJ, Hamilton TE, et al. 2004a. Thyroid neoplasia, autoimmune thyroiditis, and hypothyroidism in persons exposed to iodine 131 from the hanford nuclear site. JAMA 292:2600–1613. Davis S, Stepanenko V, Rivkind N, et al. 2004b. Risk of thyroid cancer in the Bryansk Oblast of the Russian Federation after the Chernobyl Power Station accident. Radiat Res 162:241–248. de Vathaire F, Hardiman C, Shamsaldin A, et al. 1999. Thyroid carcinomas after irradiation for a first cancer during childhood. Arch Intern Med 159:2713–2719. de Vathaire F, Le VB, Vathaire CC. 2000. Thyroid cancer in French Polynesia between 1985 and 1995: Influence of atmospheric nuclear bomb tests performed at Mururoa and Fangataufa between 1966 and 1974. Cancer Causes Control 11:59–63. DeGroot LJ, Reilly M, Pinnameneni K, et al. 1983. Retrospective and prospective study of radiation-induced thyroid disease. Am J Med 74: 852–862. Derwahl M, Broecker M, Karaiem Z. 1999. Clinical review 101: Thyrotropin may not be the dominant growth factor in benign and malignant thyroid tumors. J Clin Endocrinol Metab 84:829–834.

989

Dickman PW, Holm LE, Lundell G, et al. 2003. Thyroid cancer risk after thyroid examination with I131: A population-based cohort study in Sweden. Int J Cancer 106:580–587. Divine BJ, Barron V. 1987. Texaco mortality study: III. A cohort study of producing and pipeline workers. Am J Ind Med 11:189–202. Dobyns BM, Sheline GE, Workman JB, et al. 1974. Malignant and benign neoplasms of the thyroid in patients treated for hyperthyroidism: A report of the Cooperative Thyrotoxicosis Therapy Follow-up Study. J Clin Endocrinol Metab 38:976–998. Doniach I. 1969. The thyroid: Epidemiology of thyroid cancer. J Roy Coll Surg Edinb 14:261–262. Doniach I. 1971. Aetiological considerations of thyroid cancer. Br J Radiol 44:819. Donis-Keller H, Dou S, Chi D, et al. 1993. Mutations in the RET proto-oncogene are associated with MEN 2A and FMTC. Hum Mol Genet 2:851–856. Duffy BJ Jr, Fitzgerald PJ. 1950. Cancer of the thyroid in children: A report of 28 cases. J Clin Endocrinol Metab 10:1296–1308. Eden K, Helfand M, Mahon S. 1999. Screening for thyroid cancer: background paper. In: National Research Council and Institute of Medicine, ed. Exposure of the American People to Iodine-131 From Nevada Nuclear-bomb Tests: Review of the National Cancer Institute Report and Public Health Implications. Washington DC: National Academy Press, pp. 221–263. Eden K, Mahon S, Helfand M. 2001. Screening high-risk populations for thyroid cancer. Med Pediatr Oncol 36:583–591. Fagin JA, Matsuo K, Karmarker A, et al. 1993. High prevalence of mutations of the p53 gene in poorly differentiated human thyroid carcinomas. J Clin Invest 91:179–184. Fagin JA. 1994. Molecular pathogenesis of human thyroid neoplasms. Thyroid Today 17:1–7. Fahey TJ, Reeve TS, Delbridge L. 1995. Increasing incidence and changing presentation of thyroid cancer over a 30-year period. Br J surg 82:518–520. Farid NR, Shi Y, Zou M. 1994. Molecular basis of thyroid cancer. Endocrine Rev 15:202–232. Feldt-Rasmussen U. 2001. Iodine and cancer. Thyroid 11:483–486. Fillmore CM, Petralia SA, Dosemeci M. 1999. Cancer mortality in women with probable exposure to silica: A death certificate study in 24 states of the U.S. Am J Ind Med 36:122–128. Fincham SM, Ugnat AM, Hill GB, et al. 2000. Is occupation a risk factor for thyroid cancer? Canadian cancer registries epidemiology Research Group. J Occup Environ Med 42:318–322. Franceschi S, Fassina A, Talamini R, et al. 1989. Risk factors for thyroid cancer in Northern-Italy. Int J Epidemiol 18:578–584. Franceschi S, Fassina A, Talamini R, et al. 1990. The influence of reproductive and hormonal factors on thyroid cancer in women. Rev Epidemiol Santé Publique 38:27–34. Franceschi S, La Vecchia C. 1994. Cancer of the thyroid. In: Doll R, Fraumeni JF Jr., Muir CS, eds. Trends in cancer incidence and mortality. Vol. 19/20. Cold Spring Harbor: Cold Spring Harbor Press, pp. 393–424. Franceschi S, Levi F, Negri E, et al. 1991. Diet and thyroid cancer: A pooled analysis of four European case-control studies. Int J Cancer 48:395–398. Franceschi S, Preston-Martin S, Dal Maso L, et al. 1999. A pooled analysis of case-control studies of thyroid cancer. IV. Benign thyroid diseases. Cancer Causes Control 10:583–595. Franceschi S. 1998. Iodine intake and thyroid carcinoma—a potential risk factor. Exp Clin Endocrinol Diabetes 106(suppl 3):38–44. Franklyn JA, Maisonneuve P, Sheppard M, et al. 1999. Cancer incidence and mortality after radioiodine treatment for hyperthyroidism: A populationbased cohort study. Lancet 353:2111-2115. Franklyn JA, Maisonneuve P, Sheppard MC, et al. 1998. Mortality after the treatment of hyperthyroidism with radioactive iodine. N Engl J Med 338:712–718. Franssila K, Saxén E, Teppo L, et al. 1981. Incidence of different morphological types of thyroid cancer in the Nordic countries. Acta Pathol Microbiol Scand 89:49–55. Franssila KO, Harach HR. 1986. Occult papillary carcinoma of the thyroid in children and young adults: A systemic autopsy study in Finland. Cancer 58:715–719. Frentzel-Beyme R, Helmert U. 2000. Association between malignant tumors of the thyroid gland and exposure to environmental protective and risk factors. Rev Environ Health 15:337–358. Friedman G, Ury HK. 1980. Initial screening for carcinogenicity of commonly used drugs. JNCI 65:723–733. Friedman GD, Ury HK. 1983. Screening for possible drug carcinogenicity: Second report of findings. JNCI 71:1165–1175. From G, Mellemgaard A, Knudsen N, et al. 2000. Review of thyroid cancer cases among patients with previous benign thyroid disorders. Thyroid 10:697–700.

990

PART IV: CANCER BY TISSUE OF ORIGIN

Fukunaga FH, Lockett LJ. 1971. Thyroid carcinoma in the Japanese in Hawaii. Arch Pathol 92:6–13. Fukunaga FH, Yatani R. 1975. Geographic pathology of occult thyroid carcinomas. Cancer 36:1095–1099. Gaitan E. 1997. Environmental goitrogens. In: Braverman LE, ed. Diseases of the Thyroid. Totowa, New Jersey: Humana Press, pp. 331–348. Galanti MR, Ekbom A, Grimelius L, et al. 1997. Parental cancer and risk of papillary and follicular thyroid carcinoma. Br J Cancer 75:451–456. Galanti MR, Lambe M, Ekbom A, et al. 1995. Parity and risk of thyroid cancers: A nested case-control study of a nationwide Swedish cohort. Cancer Causes Control 6:37–44. Gembicki M, Stozharov AN, Arinchin AN, et al. 1997. Iodine deficiency in Belarusian children as a possible factor stimulating the irradiation of the thyroid gland during the Chernobyl catastrophe. Environ Health Perspect 105:1487–1490. Gharib H, Goellner JR. 1993. Fine-needle aspiration biopsy of the thyroid— An appraisal. Ann Intern Med 118:282–289. Gilbert ES, Tarone R, Bouville A, et al. 1998. Thyroid cancer rates and I-131 doses from Nevada atmospheric nuclear bomb tests. J Natl Cancer Inst 90:1654–1660. Gilliland FD, Hunt WC, Morris DM, et al. 1997. Prognostic factors for thyroid carcinoma. A population-based study of 15,698 cases from the Surveillance, Epidemiology and End Results (SEER) program 1973–1991. Cancer 79:564–573. Glattre E, Thomassen Y, Thoresen SØ, et al. 1989. Prediagnostic serum selenium in a case-control study of thyroid cancer. Int J Epidemiol 18:45–49. Glöbel B, Glöbel H, Oberhausen E. 1984. Epidemiologic studies on patients with iodine-131 diagnostic and therapy. In: Kaul A, Neider R, Pensko J, et al., eds. Radiation-risk-protection, vol II. International Radiation Protection Association. Köln: Fachverband für Strahlenschutz e.V., 565–568. Goldgar DE, Easton DF, Cannon-Albright LA et al. 1994. Systematic population-based assessment of cancer risk in first-degree relatives of cancer probands. J Natl Cancer Inst 86: 1600–1608. Goodman MT, Yoshizawa CN, Kolonel LN. 1988. Descriptive epidemiology of thyroid cancer in Hawaii. Cancer 61:1272–1281. Greene MH, Wilson J. 1985. Second cancer following lymphatic and hematopoietic cancers in Connecticut, 1935–82. Natl Cancer Inst Monogr 68:191–218. Greenwald P, Korns RF, Nasca PC, et al. 1975. Cancer in United States Jews. Cancer Res 35:3507–3512. Gross C, Yu M, Van Herle AJ, et al. 1993. Presence of a specific antiestrogen binding site on human follicular thyroid carcinoma cell line (UCLA RO 82 W-1): Inhibition by an endogenous ligand present in human serum. J Clin Endocrinol Metab 77:1361–1366. Grossman RF, Tu SH, Duh QY, et al. 1995. Familial nonmedullary thyroid cancer: An emerging entity that warrants aggressive treatment. Arch Surg 130:892–899. Hahn K, Schnell-Inderst P, Grosche B, et al. 2001. Thyroid cancer after diagnostic administration of iodine-131 in childhood. Radiat Res 156:61–70. Hall P, Berg G, Bjelkengren G, et al. 1992. Cancer mortality after iodine-131 therapy for hyperthyroidism. Int J Cancer 50:886–890. Hall P, Holm L-E, Lundell G. 1990. Second primary tumors following thyroid cancer. A Swedish record-linkage study. Acta Oncol 29:869–873. Hallquist A, Hardell L, Degerman A, et al. 1993. Occupational exposures and thyroid cancer: Results of a case-control study. Eur J Cancer Prev 2:345–349. Hallquist A, Hardell L, Degerman A, et al. 1994. Thyroid cancer: Reproductive factors, previous thyroid diseases, drug intake, family history and diet. A case-control study. Eur J Cancer Prev 3:481. Hamilton TE, van Belle G, LoGerfo JP. 1987. Thyroid neoplasia in Marshall Islanders exposed to nuclear fallout. JAMA 258:629–636. Hancock SL, Cox RS, McDougall IE. 1991. Thyroid diseases after treatment of Hodgkin’s disease. N Engl J Med 325:599–605. Harach HR, Escalante DA, Day ES. 2002. Thyroid cancer and thyroiditis in Salta, Argentina: A 40-yr study in relation to iodine prophylaxis. Endocr Pathol 13:175–181. Harach HR, Franssila KO, Wasenius V-M. 1985. Occult papillary carcinoma of the thyroid. Cancer 56:531–538. Hard GC. 1998. Recent developments in the investigation of thyroid regulation a thyroid carcinogenesis. Environ Health Perspect 106:427–436. Harvey EB, Brinton LA. 1985. Second cancer following cancer of the breast in Connecticut, 1935–82. Natl Cancer Inst Monogr 68:99–112. Haselkorn T, Bernstein L, Preston-Martin S, et al. 2000. Descriptive epidemiology of thyroid cancer in Los Angeles County, 1972–1995. Cancer Causes Control 11:163–170. Haselkorn T, Stewart SL, Horn-Ross PL. 2003. Why are thyroid cancer rates so high in southeast asian women living in the United States? The

bay area thyroid cancer study. Cancer Epidemiol Biomarkers Prev 12:144–150. Hatch M, Ron E, Bouville A, et al. 2005. Cancer following the accident at the Chernobyl nuclear power plant. Epidemiol Rev 27:56–66. Hedinger C, Williams ED, Sobin LH. 1988. Histological Typing of Thyroid Tumours. 2nd ed. International Histological Classification of Tumours. World Health Organization. No. 11. Berlin, Springer-Verlag. Hedinger C, Williams ED, Sobin LH. 1989. The WHO histological classification of thyroid tumors: A commentary on the second edition. Cancer 63:908–911. Hegedus L. 2001. Thyroid ultrasound. Endocrinol Metab Clin North Am 30:339–360. Heitz P, Moser H, Staub JJ. 1976. Thyroid cancer: A study of 573 thyroid tumors and 161 autopsy cases observed over a thirty-year period. Cancer 37:2329–2337. Hemminki K, Dong C. 2000. Familial relationships in thyroid cancer by histopathological type. Int J Cancer 85:201–205. Hemminki K, Jiang Y. 2001. Second primary neoplasms after 19281 endocrine gland tumors: Aetiological links? Eur J Cancer 37:1886–1894. Hempelmann LH, Furth J. 1978. Etiology of thyroid cancer. In: Greenfield LD, ed. Thyroid cancer. West Palm Beach, FL: CRC Press, pp. 37–49. Hempelmann LH, Hall WJ, Phillips M, et al. 1975. Neoplasms in persons treated with x-rays in infancy: Fourth survey in 20 years. JNCI 55:519–530. Henderson BE, Kolonel LN, Dworsky R, et al. 1985. Cancer incidence in the islands of the Pacific. Natl Cancer Inst Monogr 69:73–81. Henderson BE, Ross RK, Pike MC, et al. 1982. Endogenous hormones as a major factor in human cancer. Cancer Res 42:3232–3239. Hill RN, Crisp TM, Hurley PM, et al. 1998. Risk assessment of thyroid follicular cell tumors. Environ Health Perspect 106:447–457. Hizawa K, Iida M, Aoyagi K et al. 1997. Thyroid neoplasia and familial adenomatous polyposis/Gardner’s syndrome. J Gastroenterol 32:196– 199. Hofmann C, Oslapas R, Nayyar R, et al. 1986. Testosterone enhancement of thyroid carcinoma in rats: The role of TSH. Surgery 100:1078–1085. Hofstra RM, Landsvater RM, Ceccherini I. 1994. A mutation in the RET protooncogene associated with multiple endocrine neoplasia type 2B and sporadic medullary thyroid carcinoma. Nature 367:375–376. Hollowell JG, Staehling NW, Flanders WD, et al. 2002. Serum TSH, T4, and thyroid antibodies in the United States population (1988 to 1994): National Health and Nutrition Examination Survey (NHANES III). J Clin Endocrinol Metab 87:489–499. Holm L-E, Blomgren H, Löwhagen T. 1985. Cancer risks in patients with chronic lymphocytic thyroiditis. N Engl J Med 312:601–604. Holm L-E, Hall P, Wiklund KE, et al. 1991. Cancer risk after iodine-131 therapy for hyperthyroidism. J Natl Cancer Inst 83:1072–1077. Holm L-E, Wiklund KE, Lundell GE, et al. 1988. Thyroid cancer after diagnostic doses of Iodine-131: A retrospective cohort study. J Natl Cancer Inst 80:1132–1138. Horn-Ross PL, Hoggatt KJ, Lee MM. 2002. Phytoestrogens and thyroid cancer risk: The San Francisco Bay area thyroid cancer study. Cancer Epidemiol Biomarkers Prev 11:43–49. Horn-Ross PL, Morris JS, Lee M, et al. 2001. Iodine and thyroid cancer risk among women in a multiethnic population: The bay area thyroid cancer study. Cancer Epidemiol Biomarkers Prev 10:979–985. Howe HL, Wingo PA, Thun MJ. 2001. Annual report to the Nation on the status of cancer (1973 through 1998), featuring cancers with recent increasing trends. J Natl Cancer Inst 93:824–842. Huszno B, Szybinski Z, Przybylik-Mazurek E, et al. 2003. Influence of iodine deficiency and iodine prophylaxis on thyroid cancer histotypes and incidence in endemic goiter area. J Endocrinol Invest 26(suppl 2):71–76. Inskip PD, Hartshorne MF, Tekkel M, et al. 1997. Thyroid nodularity and cancer among Chernobyl cleanup workers from Estonia. Radiat Res 147:225–235. Inskip PD. 2001. Thyroid cancer after radiotherapy for childhood cancer. Med Pediatr Oncol 36:568–573. Iribarren C, Haselkorn T, Tekawa IS, et al. 2001. Cohort study of thyroid cancer in a San Francisco Bay area population. Int J Cancer 93:745–750. Isler H. 1959. Effect of iodine on thyroid tumors induced in the rat by a lowiodine diet. J Natl Cancer Inst 23:675–693. Ito T, Seyama T, Mizuno T, et al. 1992. Unique association of p53 mutations with undifferentiated but not with differentiated carcinomas of the thyroid gland. Cancer Res 52:1369–1371. Ivanov V, Gorski A, Maksyutov M, et al. 2003. Thyroid cancer incidence among adolescents and adults in Bryansk region of Russia following Chernobyl accident. Health Phys 84:46–60. Ivanov VK, Tsyb AF, Petrov AV, et al. 2002. Thyroid cancer incidence among liquidators of the Chernobyl accident: Absence of dependence of radia-

Thyroid Cancer tion risks on external radiation dose. Radiat Environ Biophys 41: 195–198. Ivanov VK, Tsyb AF, Gorsky AI, et al. 2004. Radiological health effects of the Chernobyl accident: Estimation of radiation risks. IRPA 11 Madrid. Iwasa Z, Jinnai D, Koyama H, et al. 1986. Second primary cancer following adjuvant chemotherapy, radiotherapy and endocrine therapy for breast cancer: A nationwide survey on 47,005 Japanese patients who underwent mastectomy from 1963–1982. Jpn J Surg 16:262–271. Jackson CE, Block MA, Greenwald KA, et al. 1979. The two mutational-event theory in modelling thyroid cancer. Am J Hum Genet 31:704–710. Jacob P, Bogdanova TI, Buglova EE, et al. 2002. Comparison of thyroid cancer incidence after the Chernobyl accident in Belarus and Ukraine. In: Yamashita S, Shibata M, Hoshi K, et al., eds. Chernobyl: Message for the 21st Century. Amsterdam: Elsevier, pp. 215–220. Jacob P, Kenigsberg Y, Zvonova I, et al. 1999. Childhood exposure due to the Chernobyl accident and thyroid cancer risk in contaminated areas of Belarus and Russia. Br J Cancer 80:1461–1469. Jemal A, Murray T, Samuels A, et al. 2003. Cancer statistics, 2003. CA Cancer J Clin 53:5–26. Jemal A, Tiwari RC, Murray T, et al. 2004. Cancer statistics, 2004. CA Cancer J Clin 54:8–29. Jhiang SM, Sagartz JE, Tong Q, et al. 1996. Targeted expression of the ret/PTC1 oncogene induces papillary thyroid carcinomas. Endocrinology 137:375–378. Kalk WJ, Sitas F, Patterson AC. 1997. Thyroid cancer in South Africa—an indicator of regional iodine deficiency. S Afr Med J 87:735–738. Kameyama K, Mukai, M. 2004. Cribriform-morular variant of papillary carcinoma: Ultrastructural study and somatic/germline mutation analysis of the APC gene. Ultrastructural Pathol 28:97–102. Kanno J, Matsuoka C, Furuta K, et al. 1990. Tumor promoting effect of goitrogens on the rat thyroid. Toxicol Pathol 18:239–246. Kaplan L, Katz AD, Ben-Isaac C, et al. 1971. Malignant neoplasms and parathyroid adenoma. Cancer 28:401–407. Karga H, Lee J-K, Vickery AL, et al. 1991. Ras oncogene mutations in benign and malinant thyroid tumors. J Clin Endocrinol Metab 73:832–836. Kazakov VS, Demidchik EP, Astakhova LN. 1992. Thyroid cancer after Chernobyl. Nature 359:21. Kenigsberg J, Buglova E, Kruk J, et al. 2002. Thyroid cancer among children and adolescents of Belarus exposed due to the Chernobyl accident: dose and risk assessment. In: Yamashita S, Shibata Y, Hoshi M, Fujimura K, eds. Chernobyl: Message for the 21st Century. International Congress Series 1234, pp. 293–300. Kerber RA, Till JE, Simon SL, et al. 1993. A cohort study of thyroid disease in relation to fallout from nuclear weapons testing. JAMA 270: 2076–2082. Kimura ET, Nikiforova MN, Zhu Z, et al. 2003. High prevalence of BRAF mutations in thyroid cancer: genetic evidence for constitutive activation of the RET/PTC-RAS-BRAF signaling pathway in papillary thyroid carcinoma. Cancer Res 63:1454–1457. Klugbauer S, Pfeiffer P, Gassenhuber H, et al. 2001. RET rearrangements in radiation-induced papillary thyroid carcinomas: High prevalence of topoisomerase I sites at breakpoints and microhomology-mediated end joining in ELE1 and RET chimeric genes. Genomics 73:149–160. Kolonel LN, Hankin JH, Wilkens LR, et al. 1990. An epidemiologic study of thyroid cancer in Hawaii. Cancer Causes Control 1:223–234. Kopecky KJ, Davis S, Hamilton TE, et al. 2004. Estimation of thyroid radiation doses for the hanford thyroid disease study: results and implications for statistical power of the epidemiological analyses. Health Phys 87:15–32. Kowalski LP, Goncalves Filho J, Pinto CA, et al. 2003. Long-term survival rates in young patients with thyroid carcinoma. Arch Otolaryngol Head Neck Surg 129:746–749. Krassas GE. 2000. Thyroid disease and female reproduction. Fertil Steril 74:1063–1070. Kravdal Ø, Glattre E, Haldorsen T. 1991. Positive correlation between parity and incidence of thyroid cancer: New evidence based on complete Norwegian birth cohorts. Int J Cancer 49:831–836. Kreiger N, Parkes R. 2000. Cigarette smoking and the risk of thyroid cancer. Eur J Cancer 36:1969–1973. Kroll TG. 2002. Molecular rearrangements and morphology in thyroid cancer. Am J Pathol 160:1941–1944. Kroll TG, Sarraf P, Pecciarini L, Chen, et al. 2000. PAX8-PPARgamma1 fusion oncogene in human thyroid carcinoma [corrected]. Science 289(5483):1357–1360. La Vecchia C, Lucchini F, Negri E, et al. 1992. Trends of cancer mortality in Europe, 1955–1989: IV, urinary tract, eye, brain and nerves, and thyroid. Eur J Cancer 28A:1210–1281.

991

La Vecchia C, Ron E, Franceschi S, et al. 1999. A pooled analysis of casecontrol studies of thyroid cancer. III. Oral contraceptives, menopausal replacement therapy and other female hormones. Cancer Causes Control 10:157–166. Lee SH, Kim KM, Jung BH, et al. 2003. Estrogens in female thyroid cancer: Alteration of urinary profiles in pre- and post-operative cases. Cancer Lett 189:27–32. Lee W, Chiacchierini RP, Shleien B, et al. 1982. Thyroid tumors following 131I or localized X irradiation to the thyroid and pituitary glands in rats. Radiat Res 92:307–319. Le Marchand L, Hankin J, Bach F, et al. 1995. An ecological study of diet and lung cancer in the South Pacific. Int J Cancer 63:18–23. Lemoine NR, Mayall ES, Wyllie FS, et al. 1988. Activated ras oncogenes in human cancers. Cancer Res 48:4459–4463. Lemoine NR, Mayall ES, Wyllie FS, et al. 1999. High frequency of ras oncogene activation in all stages of human thyroid tumorigenesis. Oncogene 4:159–164. Levi F, Franceschi S, Gulie C, Negri E, et al. 1993. Female thyroid cancer: The role of reproductive and hormonal factors in Switzerland. Oncology 50:309–315. Levi F, Franceschi S, La Vecchia C, et al. 1991. Previous thyroid disease and risk of thyroid cancer in Switzerland. Eur J Cancer 27:85–88. Levi F, Franceschi S, Te VC, et al. 1990. Descriptive epidemiology of thyroid cancer in the Swiss canton of Vaud. J Cancer Res Clin Oncol 116: 639–647. Levi F, Lucchini F, Negri E, et al. 1999. Cancer mortality in Europe, 1990–1994, and an overview of trends from 1955 to 1994. Eur J Cancer 35:1477–1516. Likhtarov I, Kovgan S, Vavilov M, et al. 2005. Post-Chornobyl thyroid cancers in Ukraine. Report 1: Estimation of thyroid doses. Radiat Res 163:125–136. Likhtarev IA, Chumack VV, Repin VS. 1994. Retrospective reconstruction of individual and collective external gamma doses of population evacuated after the Chernobyl accident. Health Phys 66:643–652. Likhtarev IA, Sobolev BG, Kairo IA, et al. 1995. Thyroid cancer in the Ukraine. Nature 375:365. Lind P, Kumnig G, Heinisch M, et al. 2002. Iodine supplementation in Austria: methods and results. Thyroid 12:903–907. Lind P, Langsteger W, Molnar M. 1998. Epidemiology of thyroid diseases in iodine sufficiency. Thyroid 8:1179–1183. Linos A, Linos DA, Vgotza N, et al. 1989. Does coffee consumption protect against thyroid disease? Acta Chir Scand 155:317–320. Lips CJM, Landsvaater RM, Höppener JWM, et al. 1994. Clinical screening as compared with DNA analysis in families with multiple endocrine neoplasia type 2A. N Engl J Med 331:828–835. Liu S, Semenciw R, Ugnat AM, et al. 2001. Increasing thyroid cancer incidence in Canada, 1970–1996: Time trends and age-period-cohort effects. Br J Cancer 85:1335–1339. LiVolsi VA, Asa SL. 1994. The demise of follicular carcinoma of the thyroid gland. Thyroid 4:233–236. LiVolsi VA, Feind CR. 1976. Parathyroid adenoma and non-medullary thyroid carcinoma. Cancer 38:1391–1393. Lund E, Galanti MR. 1999. Incidence of thyroid cancer in Scandinavia following fallout from atomic bomb testing: An analysis of birth cohorts. Cancer Causes Control 10:181–187. Lundell M, Hakulinen T, Holm L-E. 1994. Thyroid cancer after radiotherapy for skin hemangioma in infancy. Radiat Res 140:334–339. Lundgren CI, Hall P, Ekbom A, et al. 2003. Incidence and survival of Swedish patients with differentiated thyroid cancer. Int J Cancer 106:569–573. Lupoli G, Vitale G, Caraglia M, et al. 1999. Familial papillary thyroid microcarcinoma: A new clinical entity. Lancet 353:637–639. Maceri DR, Sullivan MJ, McClatchney KD. 1986. Autoimmune thyroiditis: Pathophysiology and relationship to thyroid cancer. Laryngoscope 96:82–86. Mack WJ, Preston-Martin S, Bernstein L, et al. 1999. Reproductive and hormonal risk factors for thyroid cancer in Los Angeles County females. Cancer Epidemiol Biomarkers Prev 8:991–997. Mack WJ, Preston-Martin S, Dal Maso L, et al. 2003. A pooled analysis of case-control studies of thyroid cancer: Cigarette smoking and consumption of alcohol, coffee, and tea. Cancer Causes Control 14:773–785. Malkasian GD, Mayberry WE. 1970. Serum total and free thyroxine and thyrotropin in normal and pregnant women, neonates, and women receiving progestogens. Am J Obstet Gynec 108:1234–1238. Mangano JJ. 1996. A post-Chernobyl rise in thyroid cancer in Connecticut, USA. Eur J Cancer Prev 5:75–81. Manole D, Schildknecht B, Gosnell B, et al. 2001. Estrogen promotes growth of human thyroid tumor cells by different molecular mechanisms. J Clin Endocrinol Metab 86:1072–1077.

992

PART IV: CANCER BY TISSUE OF ORIGIN

Mazzaferri EL, Jhiang SM. 1994. Long-term impact of initial surgical and medical therapy on papillary and follicular thyroid cancer. Am J Med 97:418–428. McIver B, Eberhardt NL. 2002. Cowden disease and the PTEN/MMAC1 gene. Genetics of Complex Thyroid Diseases. 151–175. McTiernan AM, Weiss NS, Daling JR. 1984b. Incidence of thyroid cancer in women in relation to reproductive and hormonal factors. Am J Epidemiol 120:423–435. Mellemgaard A, From G, Jorgensen T et al. 1998. Cancer risk in individuals with benign thyroid disorders. Thyroid 8:751–754. Memon A, Darif M, Al Saleh K, et al. 2002. Epidemiology of reproductive and hormonal factors in thyroid cancer: Evidence from a case-control study in the Middle East. Int J Cancer 97:82–89. Memon A, Varghese A, Suresh A. 2002. Benign thyroid disease and dietary factors in thyroid cancer: A case-control study in Kuwait. Br J Cancer 86:1745–1750. Mishra SI, Luce-Aoelua PH, Wilkens LR. 1996. Cancer among indigenous populations. The experience of American Samoans. Cancer 78:1553– 1557. Modan B, Ron E, Lerner-Geva L, et al. 1998. Cancer incidence in a cohort of infertile women. Am J Epidemiol 147:1038–1042. Moysich K, Menezes R, Michalek A. 2002. Chernobyl-related ionising radiation exposure and cancer risk: An epidemiological review. Lancet Oncol 5:269–279. Mulla ZD, Margo CE. 2000. Primary malignancies of the thyroid: epidemiologic analysis of the Florida cancer data system registry. Ann Epidemiol 10:24–30. Mulligan LM, Kwok JBJ, Healey CS, et al. 1993. Germ-line mutations of the RET proto-oncogene in multiple endocrine neoplasia type 2A. Nature 363:458–460. Namba H, Rubin SA, Fagin JA. Point mutations of ras oncoenes are an early event in thyroid tumorigenesis. Mol Endocrinol 4:1474–1479. Narita T, Takagi K. 1984. Ataxia-telangiectasia with dysgerminoma of right ovary, papillary carcinoma of thyroid, and adenocarcinoma of pancreas. Cancer 54:1113–1116. NAS (National Academy of Sciences). 2000. Review of the Hanford thyroid disease study draft final report. Washington, DC: National Academy Press. NCRP (National Council on Radiation Protection and Measurement). 1985. Induction of thyroid cancer by ionizing radiation. Report No. 80. Bethesda, MD: NCRP. Negri E, Dal Maso L, Ron E, et al. 1999b. A pooled analysis of case-control studies of thyroid cancer. II. Menstrual and reproductive factors. Cancer Causes Control 10:143–155. Negri E, La Vecchia C, Franceschi S, et al. 1991. Vegetable and fruit consumption and cancer risk. Int J Cancer 48:350–354. Negri E, Ron E, Franceschi S, et al. 1999a. A pooled analysis of case-control studies of thyroid cancer—I. Methods. Cancer Causes Control 10:131– 142. Nmec J, Zamrazil V, Pohunková D, et al. 1980. Thyroidectomy as a pathogenic factor in the evolution of thyroid cancer. Neoplasma 27:595–599. Ng SB, Sittampalam K, Goh YH, et al. 2003. Cribriform-morular variant of papillary carcinoma: The sporadic counterpart of familial adenomatous polyposis-associated thyroid carcinoma. A case report with clinical and molecular genetic correlation. Pathology 35:42–46. Nikiforov YE, Koshoffer A, Nikiforova M, et al. 1999. Chromosomal breakpoint positions suggest a direct role for radiation in inducing illegitimate recombination between the ELE1 and RET genes in radiation-induced thyroid carcinomas. Oncogene 18:6330–6334. Nikiforov YE, Rowland JM, Bove KE, et al. 1997. Distinct pattern of ret oncogene rearrangements in morphological variants of radiation-induced and sporadic thyroid papillary carcinomas in children. Cancer Res 57: 1690–1694. Nikiforova MN, Lynch RA, Biddinger PW, et al. 2003. RAS point mutations and PAX8-PPAR gamma rearrangement in thyroid tumors: Evidence for distinct molecular pathways in thyroid follicular carcinoma. J Clin Endocrinol Metab 88:2318–2326. Nikiforova MN, Stringer JB, Blough R, et al. 2000. Proximity of chromosomal loci that participate in radiation-induced rearrangements in human cells. Science 290:138–141. Ohshima M, Ward JM. 1986. Dietary iodine deficiency as a tumor promoter and carcinogen in male F344/NCR rats. Cancer Res 46:877–883. Ohta S, Katsura T, Shimada M, et al. 1986. Ataxia-telangiectasia with papillary carcinoma of the thyroid. Am J Pediatr Hemat/Oncol 8:255–268. Oliynyk V, Epshtein O, Sovenko T, et al. 2001. Post-surgical ablation of thyroid residues with radioiodine in Ukrainian children and adolescents affected by post-Chernobyl differentiated thyroid cancer. J Endocrinol Invest 24:445–447.

Olsen JH, Boice JD Jr., Jensen JPA, et al. 1989. Cancer among epileptics exposed to anti-convulsant drugs. J Natl Cancer Inst 81:803–808. Ottino A, Pianzola HM, Castelletto RH. 1989. Occult papillary thyroid carcinoma at autopsy in La Plata, Argentina. Cancer 64:547–551. Ozaki O, Ito K, Kobayashi K, et al. 1988. Familial occurrence of differentiated, nonmedullary thyroid carcinoma. World J Surg 12:565–571. Pacchiarotti A, Martino E, Bartalena L, et al. 1986. Serum thyrotropin by ultrasensitive immunoradiometric assay and serum free thyroid hormones in pregnancy. J Endocrinol Invest 9:185–189. Paksoy N, Bouchardy C, Parkin DM. 1991. Cancer incidence in Western Samoa. Int J Epidemiol 20:634–641. Paksoy N, Montaville B, McCarthy SW. 1990. Cancer occurrence in Vanuatu in the South Pacific, 1980–86. Trop Geogr Med 42:157–161. Pal T, Vogl FD, Chappuis PO, et al. 2001. Increased risk for nonmedullary thyroid cancer in the first degree relatives of prevalent cases of nonmedullary thyroid cancer: A hospital-based study. J Clin Endocrinol Metab 86:5307–5312. Paloyan E, Hofmann C, Prinz R, et al. 1982. Castration induces a marked reduction in the incidence of thyroid cancers. Surgery 92:839–848. Parkin DM, Ferlay J, Hamdi-Cherif M, et al. 2002a. Cancer in Africa: Epidemiology and Prevention. IARC Scientific Publications No. 153. Lyon, France: International Agency for Research on Cancer. Parkin DM, Pisani P, Ferlay J. 1999. Global cancer statistics. CA Cancer Clin 49:33–64. Parkin DM, Whelan SL, Ferlay J, et al. 2002b. Cancer Incidence in Five Continents Vol. VIII. Lyon, France: IARC Scientific Publications. Paynter OE, Burin GJ, Jaeger RB, et al. 1988. Goitrogens and thyroid follicular cell neoplasia: Evidence for a threshold process. Regul Toxicol Pharmocol 8:102–119. Pettersson B, Coleman MP, Ron E, et al. 1996. Iodine supplementation in Sweden and regional trends in thyroid cancer incidence by histopathologic type. Int J Cancer 65:13–19. Picchi P, Faloci C, Salabe GB. 2001. Reproductive history, contraceptives and cigarette smoke as risk factors for cancer of the thyroid in women. Casecontrol study. Minerva Endocrinol 26:53–57. Pickle LW, Mason TJ, Howard N, et al. 1987. Atlas of U.S. mortality among whites: 1930–1980. Washington, DC: NIH Publication No. 87-2900, US Govt. Printing Office. Pillaj NK, Thangavelu M, Ramalingaswami V. 1976. Nodular lesions of the thyroid in an area of high background radiation in coastal Kerala, India. Indian J Med Res 64:537–544. Plail RO, Bussey HJR, Glazer G, et al. 1987. Adenomatous polyposis: An association with carcinoma of the thyroid. Br J Surg 74:377– 380. Poth M. 2000. Thyroid cancer in children and adolescents, in L Wartofsky “Thyroid Cancer: A Comprehensive Guide to Clinical Management”. Totowa, New Jersey: Humana Press, pp. 121–128. Pottern LM, Kaplan MM, Larsen PR, et al. 1990. Thyroid nodularity after irradiation for lymphoid hyperplasia: A comparison of questionnaire and clinical findings. J Clin Epidemiol 43:449–460. Pottern LM, Stone BJ, Day NE, et al. 1980. Thyroid cancer in Connecticut, 1935–1975: An analysis by cell type. Am J Epidemiol 112:764–774. Preston-Martin S, Bernstein L, Pike MC, et al. 1987. Thyroid cancer among young women related to prior thyroid disease. Br J Cancer 55:191–195. Preston-Martin S, Franceschi S, Ron E, et al. 2003. Thyroid cancer pooled analysis from 14 case-control studies: What have we learned? Cancer Causes Control 14:787–790. Preston-Martin S, Jin F, Duda MJ, et al. 1993. A case-control study of thyroid cancer in women under age 55 in Shanghai (People’s Republic of China). Cancer Causes Control 4:431–440. Preston-Martin S, Menck HR. 1979. The epidemiology of thyroid cancer in Los Angeles County. West J Med 131:369–372. Prisyazhiuk A, Pjatak OA, Buzanov VA, et al. 1991. Cancer in the Ukraine, post-Chernobyl. Lancet 338:1334–1335. Reiners C, Biko J, Demidchik EP, et al. 2002. Results of radioactive iodine treatment in children from Belarus with advanced stages of thyroid cancer after the Chernobyl accident. Chernobyl: Message for the 21st Century. 205–214. Remontet L, Esteve J, Bouvier AM, et al. 2003. Cancer incidence and mortality in France over the period 1978–2000. Rev Epidemiol Sante Publique 51(1 Pt 1):3–30. Ries LAG, Eisner MP, Kosary CL, et al., eds. 2004. SEER Cancer Statistics Review, 1975–2001, National Cancer Institute. Bethesda, MD. Available at: http://seer.cancer.gov/csr/1975_2001/. Ries LAG, Miller BA, Hankey BF, et al., eds. 1994. SEER cancer statistics review, 1973–1991: Tables and graphs. NIH publ. no. 94-2789, Bethesda, MD: National Cancer Institute.

Thyroid Cancer Robbins J, Adams WH. 1989. Radiation effects in the Marshall Islands. In: Nagataki S, ed. Radiation and the Thyroid. Amsterdam: Excerpta Medica, pp. 11–24. Ron E, Curtis R, Hoffman DA, et al. 1984. Multiple primary breast and thyroid cancer. Br J Cancer 49:87–92. Ron E, Doody MM, Becker D, et al. 1998. Cancer mortality following treatment for hyperthyroidism. JAMA 280:347–355. Ron E, Kleinerman RA, Boice JD Jr., et al. 1987. A population-based casecontrol study of thyroid cancer. JNCI 79:1–12. Ron E, Kleinerman RA, LiVolsi VA, et al. 1991. Familial nonmedullary thyroid cancer. Oncology 48:309–311. Ron E, Lubin JH, Shore RE, et al. 1995. Thyroid cancer following exposure to external radiation: A pooled analysis of seven studies. Radiat Res 141:255–273. Ron E, Modan B, Preston D, et al. 1989. Thyroid neoplasia following low-dose radiation in childhood. Radiat Res 120:516–531. Ronckers CM, McCarron P, Ron E. 2005. Thyroid cancer and multiple primary tumors in the SEER cancer registries. Int J Cancer 117:281–288. Rossing MA, Cushing KL, Voigt LF, et al. 2000. Risk of papillary thyroid cancer in women in relation to smoking and alcohol consumption. Epidemiology 11:49–54. Rossing MA, Schwartz SM, Weiss NS. 1995. Thyroid cancer incidence in Asian migrants to the United States and their descendants. Cancer Causes Control 6:439–444. Rossing MA, Voigt LF, Wicklund KG, Williams M, Daling JR. 1998. Use of exogenous hormones and risk of papillary thyroid cancer (Washington, United States). Cancer Causes Control 9:341–349. Rubino C, Cailleux AF, de Vathaire F, et al. 2002. Thyroid cancer after radiation exposure. Eur J Cancer 38:645–647. Rubino C, de Vathaire F, Dottorini ME, et al. 2003. Second primary malignancies in thyroid cancer patients. Br J Cancer 89:1638–1644. Ryo UY, Arnold J, Colman M, et al. 1976. Thyroid scintigram. Sensitivity with sodium pertechnetate tc 99 gamma camera with pinhole collimator. JAMA 235:1235–1238. Sadetzki S, Calderon-Margalit R, Peretz C, et al. 2003. Second primary breast and thyroid cancers (Israel). Cancer Causes Control 14:367–375. Sakoda LC, Horn-Ross PL. 2002. Reproductive and menstrual history and papillary thyroid cancer risk: The San Francisco Bay Area thyroid cancer study. Cancer Epidemiol Biomarkers Prev 11:51–57. Samaan N, Schultz P, Ordonez N, et al. 1987. A comparison of thyroid carcinoma in those who have and have not had head and neck irradiation in childhood. J Clin Endocrinol Metab 64:219–223. Sampson RJ, Woolner LB, Bahn RC, et al. 1974. Occult thyroid carcinoma in Olmstead County, Minnesota. Prevalence at autopsy compared with that in Hiroshima and Nagasaki, Japan. Cancer 34:2072–2076. Schenker JC, Levinsky R, Ohel G. 1984. Multiple primary malignant neoplasms in breast cancer patients in Israel. Cancer 54:145–150. Schneider AB, Recant W, Pinsky S, et al. 1986b. Radiation-induced thyroid carcinoma: Clinical course and results of therapy in 298 patients. Ann Intern Med 105:405–412. Schneider AB, Bekerman C, Leland J, et al. 1997. Thyroid nodules in the follow-up of irradiated individuals: Comparison of thyroid ultrasound with scanning and palpation. J Clin Endocrinol Metab 82:4020–4027. Schneider AB, Pinsky S, Bekerman C, et al. 1980. Characteristics of 108 thyroid cancers detected by screening in a population with a history of head and neck irradiation. Cancer 46:1218–1227. Schneider AB, Ron E, Lubin J, et al. 1993. Dose-response relationships for radiation-induced thyroid cancer and thyroid nodules: Evidence for the prolonged effects of radiation on the thyroid. J Clin Endocrinol Metab 77:362–369. Schneider AB, Shore-Freedman E, Ryo UY, et al. 1985. Radiation-induced tumors of the head and neck following childhood irradiation. Medicine 64:1–15. Schneider AB, Shore-Freedman E, Weinstein RA. 1986a. Radiation-induced thyroid and other head and neck tumors: Occurrence of multiple tumors and analysis of risk factors. J Clin Endocrinol Metab 63:107–112. Schottenfeld D, Berg JW. 1971. Incidence of multiple primary cancers. IV. Cancers of the female breast and genital organs. J Natl Cancer Inst 46:161. Segev DL, Umbricht C, Zeiger MA. 2003. Molecular pathogenesis of thyroid cancer. Surg Oncol 12:69–90. Selby JV, Friedman GD, Fireman BH. 1989. Screening prescription drugs for possible carcinogenicity: Eleven to fifteen years of follow-up. Cancer Res 49:5736–5747. Shakhtarin VV, Tsyb AF, Stepanenko VF, et al. 2003. Iodine deficiency, radiation dose, and the risk of thyroid cancer among children and adolescents in the Bryansk region of Russia following the Chernobyl power station accident. Int J Epidemiol 32:584–591.

993

Shi Y, Zou M, Schmidt H, et al. 1991. High rates of ras codon 61 mutation in thyroid tumors in an iodide-deficient area. Cancer Res 51:2690–2693. Shibata Y, Yamashita S, Masyakin V, et al. 2001. 15 years after Chernobyl: New evidence of thyroid cancer. Lancet 358:1965–1966. Simon MS, Tang MT, Bernstein L, et al. 2002. Do thyroid disorders increase the risk of breast cancer? Cancer Epidemiol Biomarkers Prev 11:1574– 1578. Shore RE. 1992. Issues and epidemiological evidence regarding radiationinduced thyroid cancer. Radiat Res 131:98–111. Shore RE, Hempelmann LH, Woodward ED. 1986. Carcinogenic effects of radiation on the human thyroid gland. In: Upton AC, Albert RE, Burns FJ, et al., eds. Radiation Carcinogenesis. New York: Elsevier, pp. 293–310. Shore RE, Hildreth N, Dvoretsky E, et al. 1993. Thyroid cancer among persons given X-ray treatment in infancy for an enlarged thymus gland. Am J Epidemiol 137:1068–1080. Shore RE, Woodard E, Hildreth N, et al. 1985. Thyroid tumors following thymus irradiation. JNCI 74:1177–1184. Shore RE, Xue X-N. 1999. Comparative thyroid cancer risk of childhood and adult radiation exposure and estimation of lifetime risk. In: Thomas D, Karaoglou A, Williams ED, eds. Radiation and Thyroid Cancer. Singapore, World Scientific Publishing, pp. 491–498. Siegal A, Modan M. 1981. Latent carcinoma of thyroid in Israel: a study of 260 autopsies. Isr J Med Sci 17:249–253. Sigurdson AJ, Doody MM, Rao RS, et al. 2003. Cancer incidence in the US radiologic technologists health study, 1983–1998. Cancer 15:3080–3089. Sigurdson AJ, Ronckers CM, Mertens AC, et al. 2005. Primary thyroid cancer after a first tumour in childhood (the Childhood Cancer Survivor Study): a nested case-control study. Lancet 365:2014–2023. Silverberg SG, Vidone RA. 1966. Carcinoma of the thyroid in surgical and postmortem material. Analysis of 300 cases at autopsy and literature review. Ann Surg 164:291–299. Simon MS, Tang MT, Bernstein L, et al. 2002. Do thyroid disorders increase the risk of breast cancer? Cancer Epidemiol Biomarkers Prev 12: 684–685. Sokic SI, Adanja BJ, Vlajinac HD, et al. 1994. Risk factors for thyroid cancer. Neoplasma 41:371–374. Steinmetz KA, Potter JD. 1996. Vegetables, fruit, and cancer prevention: A review. J Am Diet Assoc 96:1027–1039. Stepanenko VF, Voilleque PG, Gavrilin YI, et al. 2004. Estimating individual thyroid doses for a case-control study of childhood thyroid cancer in Bryansk Oblast, Russia. Radiat Prot Dosimetry 108:1431–1460. Stezhko VA, Buglova EE, Danilova LI, et al. 2004. A cohort study of thyroid cancer and other thyroid diseases after the Chornobyl accident: objectives, design and methods. Radiat Res 161:481–492. Stoffer SS, Van Dyke DL, Bach JV, et al. 1986. Familial papillary carcinoma of the thyroid. Am J Med Genet 25:775–782. Stsjazhko VA, Tsyb AF, Tronko ND, et al. 1995. Childhood thyroid cancer since accident at Chernobyl. Br Med J 310:801. Suárez HG, du Villard JA, Severino M, et al. 1990. Presence of mutations in all three ras genes in human thyroid tumors. Oncogene 5:565–570. Szybinski Z, Huszno B, Zemla B, et al. 2003. Incidence of thyroid cancer in the selected areas of iodine deficiency in Poland. J Endocrinol Invest 26:63–70. Takagi H, Mitsumori K, Onodera H, et al. Improvement of a two-stage carcinogenesis model to detect modifying effects of endocrine disrupting chemicals on thyroid carcinogenesis in rats. Cancer Lett 178:1–9. Takahashi T, Schoemaker MJ, Trott KR, et al. 2003. The relationship of thyroid cancer with radiation exposure from nuclear weapon testing in the Marshall Islands. J Epidemiol 13:99–107. Takeichi N, Ito H, Haruta R, et al. 1991. Relation between estrogen receptor and malignancy of thyroid cancer. Jpn J Cancer Res 82:19–22. Takezaki T, Hirose K, Inoue M, et al. 1996. Risk factors of thyroid cancer among women in Tokai, Japan. J Epidemiol 6:140–147. Tanaka K, Sonoo H, Yamamoto Y, et al. 1999. Analyses of the outcome of locally invasive papillary thyroid carcinomas. Thyroid 9:1017–1022. Teppo L, Pukkala E, Saxén E. 1985. Multiple cancer—an epidemiologic exercise in Finland. JNCI 75:207–217. Thomas GA, Bunnell H, Cook HA, et al. 1999. High prevalence of RET/PTC rearrangements in Ukrainian and Belarussian post-Chernobyl thyroid papillary carcinomas: A strong correlation between RET/PTC3 and the solid-follicular variant. J Clin Endocrinol Metab 84:4232–4238. Thompson D, Mabuchi K, Ron E, et al. 1994. Cancer incidence in atomic bomb survivors. Part II: Solid tumors, 1958–87. Radiat Res 137:S17–S67. Trapido EJ, McCoy CB, Stein NS, et al. 1990. The epidemiology of cancer among Hispanic women. Cancer 66:2435–2441. Tronko M, Bogdanova T, Komissarenko I, et al. 1999. Thyroid carcinoma in children and adolescents in Ukraine after the Chernobyl nuclear accident:

994

PART IV: CANCER BY TISSUE OF ORIGIN

Statistical data and clinicomorphologic characteristics. Cancer 86: 149–156. Tronko M, Bogdanova T, Komissarenko I, et al. 1996. Thyroid carcinoma in children and adolescents in Ukraine after the Chernobyl nuclear accident (1986–1995). In: Karaoglou A, Desmet G, Kelly GN, et al., eds. The Radiological Consequences of the Chernobyl Accident. Luxembourg: EUR 16544 EN, pp. 683–690. Tucker MA, Morris Jones PH, Boice JD Jr., et al. 1991. Therapeutic radiation at a young age is linked to secondary thyroid cancer. Cancer Res 51:2885–2888. Turrin A, Pilotti S, Ricci SB. 1985. Characteristics of thyroid cancer following irradiation. Int J Radiat Oncol Biol Phys 11:2149–2154. UNSCEAR (United Nations Scientific Committee on the Effects of Radiation). 2000. Sources and effects of ionizing radiation. New York: United Nations. Utiger RD. 1995. Cigarette smoking and the thyroid. N Engl J Med 333: 1001–1002. Utiger RD. 1994. Medullary thyroid carcinoma, genes, and the prevention of cancer. N Engl J Med 331:870–871. Vågerö D, Persson G. 1986. Occurrence of cancer in socioeconomic groups in Sweden. An analysis based on the Swedish Cancer Environment Registry. Scand J Soc Med 14:151–160. Vassilopoulou-Sellin R, Palmer L, Taylor S, et al. 1999. Incidence of breast carcinoma in women with thyroid carcinoma. Cancer 85:696–705. Verkooijen HM, Fioretta G, Pache J-C, et al. 2003. Diagnostic changes as a reason for the increase in papillary thyroid cancer incidence in Geneva, Switzerland. Cancer Causes Control 14:13–17. Vini L, Hyer SL, Marshall J, et al. 2003. Long-term results in elderly patients with differentiated thyroid carcinoma. Cancer 97:2736–2742. Wang J-X, Inskip PD, Boice JD Jr., et al. 1990a. Cancer incidence among medical diagnostic x-ray workers in China, 1950 to 1985. Int J Cancer 45:889–895. Wang Z, Boice JD Jr., Luxin W, et al. 1990b. Thyroid nodularity and chromosome aberrations among women in areas of high background radiation in China. J Natl Cancer Inst 82:478–485. Ward JM, Ohshima M. 1986. The role of iodine in carcinogenesis. Adv Exp Med Biol 206:529–542. Wattenberg LW. 1983. Inhibition of neoplasia by minor dietary constituents. Cancer Res 43(suppl):2448S–2453S.

Weeke J, Hansen AP. 1975. Serum TSH and serum T3 levels during normal menstrual cycles and during cycles on oral contraceptives. Acta Endocrinol 79:431–438. Wegelin C. 1928. Malignant disease of the thyroid gland and its relation to goitre in man and animal. Cancer Rev 3:297–313. Williams D. 2002. Cancer after nuclear fallout: Lessons from the Chernobyl accident. Nat Rev Cancer 2:543–549. Williams ED, Doniach I, Bjarnason O, et al. 1977. Thyroid cancer in an iodinerich area. A histopathological study. Cancer 39:215–222. Williams ED. 1995. Mechanisms and pathogenesis of thyroid cancer in animals and man. Mutat Res 333:123–129. Williams ED. 1979. The aetiology of thyroid tumors. Clin Endocrinol Metab 8:193–207. Williams ED. 1990. TSH and thyroid cancer. Horm Metab Res Suppl 23:72–75. Williams RR. 1976. Breast and thyroid cancer and malignant melanoma promoted by alcohol-induced pituitary secretion of prolactin, TSH and MSH. Lancet 1:996–999. Wingren G, Hatschek T, Axelson O. 1993. Determinants of papillary cancer of the thyroid. Am J Epidemiol 138:482–491. Winship T, Rosvoll RV. 1970. Thyroid carcinoma in childhood: Final report on a 20 year study. Clin Proc Children’s Hosp 26:327–348. Wright PA, Lemoine NR, Mayall ES, et al. 1989. Papillary and follicular thyroid carcinomas show a different pattern of ras oncogene mutation. Br J Cancer 60:576–577. Yamashita S, Shibata Y, Hoshi M, et al., eds. Chernobyl: Message for the 21st Century. Amsterdan, Elsevier, 2002. Yane K, Kitahori Y, Konishi N, et al. 1994. Expression of the estrogen receptor in human thyroid neoplasms. Cancer Lett 84:59–66. Yin F, Giuliano AE, Van Herle AJ. 1999. Growth inhibitory effects of flavonoids in human thyroid cancer cell lines. Thyroid 9:369–376. Zheng TZ, Holford TR, Chen YT, et al. 1996. Time trend and age-period-cohort effect on incidence of thyroid cancer in Connecticut, 1935–1992. Int J Cancer 67:504–509. Zimmermann MB, Kohrle J. 2002. The impact of iron and selenium deficiencies on iodine and thyroid metabolism: Biochemistry and relevance to public health. Thyroid 12:867–878. Zivaljevic V, Vlajinac H, Jankovic R, et al. 2003. Case-control study of female thyroid cancer—menstrual, reproductive and hormonal factors. Eur J Cancer Prev 12:63–66.

51

Breast Cancer GRAHAM A. COLDITZ, HEATHER J. BAER, AND RULLA M. TAMIMI

B

reast cancer is the most common cancer diagnosis and the second leading cause of cancer death among women in the United States. Over 211,000 women and 1600 men in this country are diagnosed with breast cancer each year, and 40,000 Americans die of this disease annually (ACS, 2005). In terms of its global burden, breast cancer ranks second in incidence and fifth in mortality among all cancers, with approximately 1.05 million new cases and 373,000 deaths in the year 2000 (Parkin et al., 2001). Due to the dominance of breast cancer among women, this chapter focuses on female breast cancer. Furthermore, our primary focus is invasive breast cancer, as it has been studied more extensively than in situ disease; studies that have evaluated risk factors for both types, however, suggest similar etiologic factors (Longnecker et al., 1996; Trentham-Dietz et al., 2000). Incidence rates of breast cancer increased in the United States during most of the 20th century. Over the past 50 years, incidence rates have also been rising in many other regions of the world, with the most notable increases in traditionally low-incidence Asian countries (Nagata, Kawakami, and Shimizu, 1997; Seow et al., 1996). These international trends may reflect secular changes in reproductive patterns and lifestyle factors that affect breast cancer risk, which will be discussed later in this chapter. Trends in breast cancer mortality have largely paralleled trends in incidence, although mortality rates in some groups have begun to decline since the early 1990s, which may be attributable to improvements in screening practices and treatment effectiveness. Ovarian hormones, and estrogens in particular, play an important role in breast cancer etiology. Both endogenous and exogenous hormones increase cellular proliferation in the breast, thereby increasing the likelihood of random genetic errors during cell division (Henderson and Feigelson, 2000). Many of the established risk factors—including early menarche, late menopause, and postmenopausal obesity—are believed to reflect the cumulative “dose” of estrogen to the breast epithelium (Henderson and Feigelson, 2000). Although understanding of the causes of breast cancer has improved substantially over the past 50 years, few of the known risk factors are modifiable in light of existing societal norms. Hence, to reduce the global burden of breast cancer during the 21st century, it will be necessary to identify and implement novel strategies for prevention as well as to uncover other important environmental and genetic factors involved in the etiology of this disease.

CLASSIFICATION Anatomic Distribution The human breast is composed of stromal and epithelial elements. The stromal elements of the breast consist of adipose, connective tissue, blood and lymphatic vessels and give the breast its structure. The epithelial elements represent the functional units of the mammary gland and are composed of a number of branching ducts, which connect the lobules to the nipple-areola complex.

*Drs. Colditz and Tamimi are supported by Grants CA87969 and CA46475. Dr. Colditz is an American Cancer Society Clinical Research Professor. Dr. Baer is supported by grant DAMD17-00-1-0165 from the Department of Defense and by grant T32 CA 09001-28 from the National Institutes of Health.

At birth, the mammary gland is made up of only elementary ducts, which grow and divide slowly until puberty. At the initiation of menses, the growth and branching of the ducts increases and the terminal end buds of the ducts begin to give rise to lobules. During the menstrual cycle, epithelial and stromal proliferation and regression occur in a cyclic fashion. Lobule formation continues from menarche and increases until about age 25. Only through pregnancy and lactation does complete differentiation of the lobular tissue take place.

Histopathology More than 95% of breast cancers originate from the epithelial elements of the mammary gland and are classified as adenocarcinomas. There are two major classifications of breast malignancies: in situ or noninvasive cancers, and invasive cancer. In situ breast cancers are morphologically similar to invasive cancers, but they remain confined to the duct (ductal carcinoma in situ, DCIS) or lobule (lobular carcinoma in situ, LCIS) and show no evidence of infiltrating the surrounding stroma. DCIS increases the risk of ipsilateral breast cancer, although it is still controversial whether these lesions represent obligate precursors (Going, 2003; Collins et al., 2005). LCIS is frequently noted to be multifocal and bilateral (Lakhani, 2001). In contrast to DCIS, LCIS is associated with an increased risk of breast cancer which is bilateral, suggesting that LCIS may be a marker of high risk rather than a precursor lesion (Schnitt and Morrow, 1999). Invasive cancers are a heterogeneous group of lesions characterized by tumor cells that invade the breast stroma. Seventy-five to 80% of invasive cancers are infiltrating ductal carcinomas, while infiltrating lobular carcinomas make up the second most common type of invasive cancers (5%–10%). Other types of breast cancer include medullary, adenoid cystic, mucinous, and tubular; these types of tumors make up a very small proportion of breast cancers and usually have an excellent prognosis. Pathologic features of breast tumors provide important indications for predicting response to treatment and overall outcome. Axillary lymph node status of breast cancer is the number one prognostic indicator of disease survival, with lower survival associated with increasing numbers of lymph nodes involved. There are three histologic grades used to characterize invasive breast cancers: well differentiated (Grade I), moderately differentiated (Grade II), and poorly differentiated (Grade III) (Schnitt, 2000). Higher-grade tumors are associated with a poorer prognosis. Tumor size, histologic type, nuclear grade, proliferative rate, and hormonal status are other important predictors of treatment response and prognosis.

Precursor Neoplastic Lesions Benign breast disease (BBD) is an umbrella term that encompasses a number of heterogeneous breast abnormalities. These unique benign conditions vary in their morphologic and pathologic features and, most importantly, in their subsequent breast cancer risk. A classification system originally proposed by Dupont and Page and later supported by the College of American Pathologists provides a general scheme to categorize BBD into three clinically relevant groups: nonproliferative, proliferative without atypia, and proliferative with atypia (Dupont and Page, 1985).

995

996

PART IV: CANCER BY TISSUE OF ORIGIN

Non-proliferative lesions include cysts, apocrine metaplasia, and mild hyperplasia of usual type. Women with these lesions are at the same risk of breast cancer as women without a breast biopsy (Dupont and Page, 1985). Proliferative lesions without atypia (e.g., intraductal papilloma, sclerosing adenosis, moderate hyperplasia of usual type) are associated with a 1.5-fold to twofold increased risk of breast cancer compared with non-proliferative lesions (Schnitt and Connolly, 2000). Atypical ductal (ADH) and lobular (ALH) hyperplasias make up the group of proliferative lesions with atypia. Atypical hyperplasias are similar to in situ carcinomas in that they are both characterized by proliferation of epithelial cells, but they do not share all of the morphologic and pathologic features. These lesions are associated with a 3.5-fold to sixfold increased risk of subsequent breast cancer (Schnitt and Connolly, 2004). When benign breast lesions are considered as one group, the increased risk of breast cancer subsequent to the BBD diagnosis is likely to be the same in the contralateral breast as it is in the ipsilateral breast. Although earlier studies suggested that BBD is a general marker of increased risk and not a precursor lesion, more recent evidence examining specific benign lesions indicates that this may not be true for all benign lesions. There is evidence that risk of breast cancer in women with previous ALH is three times more likely to occur in the same breast than in the opposite breast, suggesting that ALH may in fact be a precursor (not necessarily obligate) as well as a risk indicator (Lakhani, 2003; Page et al., 2003).

Molecular Genetic Characteristics of Tumors Hormone Receptor Status The effects of estrogen and progesterone on cell growth and development are mediated through hormone receptors. A large portion of breast cancer tumors express estrogen (ER) and progesterone (PR) receptors. Tumors that have receptor levels above a specific cutoff are considered to be ER or PR positive, respectively. The ER and PR status of the cancer is important for two reasons. First, tumors that express these receptors at high levels tend to be more differentiated, and these patients are likely to have a better prognosis (Rosen, 2001). Second, ER and PR expression is strongly predictive of the tumor’s response to hormonal or anti-estrogen therapies (Rosen, 2001). Risk factor patterns differ according to receptor status and indicate that the receptors are markers of different tumor types rather than stages of a single disease with a single disease pathway (Colditz et al., 2003).

HER2/neu HER2/neu is a protooncogene that is amplified in approximately 20% of breast cancers. Most studies have shown that amplification of this gene is associated with poorer prognosis and response to chemotherapy, primarily among node positive cancers (Schnitt, 2001). HER2/neu status of tumor cells is important because it predicts response to Herceptin, a humanized anti-HER2 antibody. Herceptin significantly improves the response to chemotherapy in tumors overexpressing HER2/neu (Pegram, Pauletti, and Slamon, 1998).

DEMOGRAPHIC PATTERNS Incidence and Mortality in the United States The United States has one of the highest incidence rates of breast cancer in the world. Between 1998 and 2002, the Surveillance, Epidemiology, and End Results (SEER) age-adjusted incidence rate was 134 per 100,000 and the corresponding mortality rate was 26 per 100,000 (Ries et al., 2005). There is some geographic variation in breast cancer incidence and mortality within this country, although it is small in comparison with the degree of international variation. The age-adjusted incidence of breast cancer is above the national average among women in parts of the Northwest, northern California, and the Northeast, whereas Utah and New Mexico have the lowest breast cancer incidence among the 11 SEER registries (Ries et al., 2005). Geographic differences in breast cancer mortality are similar to those in incidence (Ries et al., 2005). Recent reports suggest that much of

this variation in both incidence and mortality can be explained by regional differences in the prevalence of known risk factors, including parity, age at first full-term pregnancy, age at menarche, and age at menopause (Centers for Disease Control and Prevention, 1992; Sturgeon et al., 1995; Laden et al., 1997; Robbins, Brescianini, and Kelsey, 1997).

Secular Trends in the United States Incidence rates of breast cancer have been increasing in the United States since formal record keeping began in the 1930s. Between 1940 and 1982, the age-adjusted incidence rate rose by an average of 1.2% per year in Connecticut (White, Lee, and Kristal, 1990), representing a cumulative increase of about 65% over the 42 years. Data from the SEER program, which began collecting data from different registries across the country in 1973, show that incidence rates rose more sharply during the 1980s (Ries et al., 2005). Several studies have examined whether the increase in breast cancer incidence in the United States has been due to the increasing use of mammography (White, Lee, and Kristal, 1990; Lantz, Remington, and Newcomb, 1991; Liff et al., 1991; Miller, Feuer, and Hankey, 1991; Feuer and Wun, 1992; Miller, Feuer, and Hankey, 1993). Because screening causes at most a transient increase in incidence, and because its use was limited before the 1980s, it can explain little of the longterm increase between the 1930s and the 1980s. However, during the 1980s the increased incidence was almost entirely due to an increase in localized disease and in tumors measuring less than 2 centimeters in diameter. According to the Wisconsin Cancer Reporting System, between 1980 and 1988 the incidence rate of in situ breast cancer and localized disease increased by 328% and 37%, respectively, whereas the incidence rate of regional disease increased by only 7% and the rate of distant disease decreased by 57% (Newcomb and Lantz, 1993). These findings suggest that the increase in use of screening mammography accounts for part of the recent increase in breast cancer incidence (Miller, Feuer, and Hankey, 1991; Chu, Tarone, and Kessler, 1996). Trends in breast cancer mortality are of major public health interest, but their interpretation is complex because they reflect the combined effects of trends in underlying risk of breast cancer, changes in screening practices, and effectiveness of treatment. Further, mortality rates lag behind changes in breast cancer risk, screening, and treatment by at least 5–10 years (Chevarley and White, 1997). Breast cancer death rates increased in the United States in the 1930s, remained relatively stable from the 1940s through the 1970s, increased again in the 1980s, and then decreased in the 1990s; the recent decline may be related to the increased use of adjuvant therapies and the incorporation of breast cancer screening into routine medical care (Wingo et al., 2003). These overall trends, however, obscure important variations in mortality by age and race. Since the 1970s, mortality rates have fallen for younger white women, and this decline has accelerated since the late 1980s. In contrast, mortality rates for white women aged 60 years and older increased slowly during the 1970s and 1980s, though mortality has also begun to decline in this group since the late 1980s (Chu, Tarone, and Kessler, 1996; Chevarley and White, 1997). Mortality rates have increased for black women in all age groups since the 1970s, and there is no clear evidence of a recent decline in mortality as has been seen for white women (Chevarley and White, 1997).

Age Breast cancer incidence rates increase sharply with age, a pattern that is most evident among premenopausal women, regardless of the underlying population incidence rates (Moolgavkar, Day, and Stevens, 1980). Between 1998 and 2002, the SEER incidence rates were 62, 120, and 195 per 100,000 for women ages 35–39, 40–44, and 45–49, respectively (Ries et al., 2005). The rate of increase in breast cancer incidence continues throughout life but slows at menopause (Fig. 51–1), strongly suggesting the involvement of reproductive hormones in breast cancer etiology, as non-hormone-dependent cancers do not exhibit this change in slope (Pike et al., 1993). Mortality rates also

Breast Cancer 1000

100

10

75

65

55

45

35

25

15

1

Age (years) San Francisco, USA Japan China

Figure 51–1. Average annual incidence of breast cancer per 100,000 women by age group, 1982. (Source: Data from Waterhouse et al., 1982.)

increase with age, ranging from 9 per 100,000 for women ages 35–39 to 98 per 100,000 for women ages 70–74 (Ries et al., 2005).

Race, Ethnicity, and Socioeconomic Status There are important racial and ethnic differences in breast cancer incidence and mortality in the United States. In 2002, the SEER ageadjusted incidence rates of breast cancer were 138 per 100,000 for white women compared with 120 per 100,000 for black women (Ries et al., 2005). However, these age-adjusted figures conceal a crossover pattern, in which the risk of breast cancer at younger ages is modestly higher among black women than white women, whereas at ages 50 and older, incidence rates among white women are substantially higher than among black women. Unlike most other illnesses, lifetime risk of breast cancer increases with higher socioeconomic status. This association may be explained, at least in part, by established reproductive risk factors for breast cancer (Heck and Pamuk, 1997). For instance, women in lower socioeconomic strata are more likely to have a greater number of children and to have them at younger ages than women in higher socioeconomic strata, and higher parity and earlier age at first full-term pregnancy both are associated with lower breast cancer risk (Kelsey, Gammon, and John, 1993). It is likely that much, if not all, of the black/white differences in breast cancer incidence reflect racial differences in social class (Krieger, 1990) and, thus, in the distribution of established reproductive risk factors. Despite the higher overall incidence of breast cancer among white women, however, black women have higher mortality rates and, as mentioned earlier, there has been no obvious decline in these rates as seen for white women. Breast cancer incidence and mortality rates among Asian, Hispanic, and American Indian women in the United States are considerably lower than those of (non-Hispanic) white and black women (Ries et al., 2005).

Survival The overall 5-year breast cancer survival rate in the United States was 88% between 1995 and 2001. Survival varies substantially according to stage, with the highest survival rates for localized disease (98%) and the lowest for distant disease (26%). Age at diagnosis is also related to breast cancer survival, with slightly lower 5-year survival rates among women diagnosed at young ages (Ries et al., 2005).

997

Black women have poorer breast cancer survival rates at all ages of diagnosis compared with white women; between 1995 and 2001, the 5-year survival rate was 90% for white women and 76% for black women (Ries et al., 2005). This poorer survival can be attributed, in part, to the tendency of black women to be diagnosed at later stages of disease. However, whites also have higher survival rates than blacks at each stage of disease, suggesting that there may be racial differences in prognosis or treatment (Campbell, 2002). Some studies indicate that black women diagnosed with breast cancer between the ages of 25 and 40 are more likely than white women to have tumors with biologically aggressive characteristics (Eley et al., 1994; Stanford and Greenberg, 1989). According to data from the Florida state tumor registry, racial differences in survival remain apparent even after socioeconomic status, insurance payer, and stage at diagnosis are taken into account (Roetzheim et al., 2000). A meta-analysis of 14 studies also found that African-American ethnicity was a predictor of worst breast cancer outcome, independent of socioeconomic status (Newman et al., 2002). A more recent analysis among postmenopausal women, however, showed no difference in stage-specific survival rates for blacks and whites with access to care through Medicare (Chu, Lamar, and Freeman, 2003). Hence, the evidence remains inconclusive as to the extent of this difference being biologic versus socioeconomic.

International Patterns of Incidence The incidence of female breast cancer varies by approximately fivefold between countries, being highest in the United States and Northern Europe, intermediate in Southern and Eastern Europe and South America, and lowest in Asia (Parkin et al., 1992). The marked difference between age-specific incidence rates among women in San Francisco and in China and Japan is evident in Figure 51–1. Since the 1950s, breast cancer incidence has been increasing in many traditionally lower-risk countries as well as in high-risk Western countries. For example, breast cancer incidence rates have nearly doubled in recent decades in Japan (Tominaga et al., 1994; Nagata, Kawakami, and Shimizu, 1997), Singapore (Seow et al., 1996), and urban areas of China (Jin et al., 1993). Dramatic changes in lifestyle in such regions brought about by growing economies, increasing affluence, and increases in the proportion of women in the industrial workforce have had an impact on the population distribution of breast cancer risk factors, resulting in a convergence toward the risk factor profile of Western countries and a narrowing of the international gap in breast cancer incidence (Hoover, 1996).

Migration In the United States, the magnitude of the difference in incidence rates among various ethnic groups often depends on migrant status. Between 1973 and 1986, breast cancer incidence was about 50% lower for Chinese-American and Japanese-American women born in Asia and about 25% lower for those born in the United States compared with US-born white women (Stanford et al., 1995). Compared with Chinese women living in the mainland, Singapore, and Hong Kong, Asian-born Chinese women living in the United States had an almost twofold higher annual rate of breast cancer, and US-born Chinese women had a higher rate still; the pattern for Japanese women was similar (Stanford et al., 1995). According to data from a multi-ethnic cohort in Hawaii and Los Angeles, Japanese-American women had almost the same risk of breast cancer between 1993 and 1999 as white women living in the same areas, and US rates of postmenopausal breast cancer were very high compared with the low rates previously observed in traditional Japanese women and early migrants to this country (Pike et al., 2002). A large body of literature shows increases in breast cancer incidence following migration from a low-risk country to the United States (Dunn, 1975; Kolonel, 1980; Tominaga, 1985; Shimizu et al., 1991; Yu et al., 1991; Ziegler et al., 1993). These findings strongly suggest that factors associated with the lifestyle or environment of the destination country influence breast cancer risk and are consistent with a positive relationship between length of time in the destination country and adoption of that country’s lifestyle.

998

PART IV: CANCER BY TISSUE OF ORIGIN

ENVIRONMENTAL FACTORS Reproductive Factors Age at Menarche and Characteristics of the Menstrual Cycle Menarche represents the development of the mature hormonal environment for a young woman and the onset of monthly cycling of hormones that induce ovulation, menstruation, and cell proliferation within the breast and endometrium. Earlier age at menarche has been consistently associated with increased risk of breast cancer, and most studies suggest that age at menarche is related to both premenopausal and postmenopausal breast cancer (Kelsey, Gammon, and John, 1993). Breast cancer risk generally decreases by 10%–24% with each 1-year delay in menarche (Bernstein, 2002). Although menarche is most clearly related to the onset of ovulation, some, but not all, studies suggest that hormone levels may be higher through the reproductive years among women who have early menarche (MacMahon, Trichopoulos, and Brown, 1982). In addition, early menarche may be associated with more rapid onset of regular, ovulatory menstrual cycles and hence greater lifetime exposure to endogenous hormones (Bernstein, 2002). In one study, women whose menstrual cycles became regular within 1 year of their first menstrual period had more than twofold greater breast cancer risk compared with women with 5 or more years between menarche and the onset of regular cycles (Henderson, Pike, and Casagrande, 1981). Shorter cycle length is also consistently related to greater risk of breast cancer, likely due to a greater number of cycles and more time spent in the luteal phase, when both estrogen and progesterone levels are high and proliferative activity in the breast appears to be greatest (Kelsey, Gammon, and John, 1993).

Parity, Age at First Full-Term Pregnancy, and Lactation Nulliparous women are at greater risk of breast cancer compared with parous women. This increased risk is evident for breast cancer diagnosed after age 40–45 years, but not for breast cancer occurring at younger ages. A younger age at first full-term pregnancy predicts a lower lifetime risk of breast cancer (Kelsey, Gammon, and John, 1993), reflecting the final maturation of the breast with hormonal exposures during first pregnancy and preparation for lactation. The reduction in risk following pregnancy is not immediate, but rather takes approximately 10–15 years to manifest (Bruzzi, Negri, and La Vecchia, 1988). In fact, risk of breast cancer is increased for the first decade following first pregnancy, with a greater adverse effect the longer the interval from menarche to first birth (Pike et al., 1983; Rosner, Colditz, and Willett, 1994; Rosner and Colditz, 1996). A higher number of births is also consistently related to lower risk of breast cancer; each additional birth beyond the first birth reduces long-term risk of breast cancer. In addition to a protective effect of higher parity, several studies now indicate that more closely spaced births are associated with lower lifetime risk of breast cancer (Rosner, Colditz, and Willett, 1994; Trichopoulos et al., 1983). This may be due to the breast having less time to accumulate DNA damage before it attains maximal differentiation by repeated pregnancies. As early as 1926, it was proposed that a breast never used for lactation is more liable to become cancerous (Lane-Claypon, 1926). The overall evidence supports a reduction in risk with longer duration of breastfeeding, but the findings have varied substantially in the level of risk reduction (Lipworth, Bailey, and Trichopoulos, 2000). The strongest results support at least a 50% reduction in risk for women who have breastfed for 2 or more years (Romieu et al., 1996), but this was in the setting of extremely high parity. The combined evidence from the Oxford collaborative group reanalysis of case-control and cohort studies indicates that lactation is consistently related to reduced risk (Collaborative Group on Hormonal Factors in Breast Cancer, 2002). According to results from the reanalysis, the relative risk of breast cancer decreases by 4.3% for every 12 months of breastfeeding, in addition to a 7% reduction for each birth.

Spontaneous and Induced Abortion A number of studies have examined the relation between spontaneous and induced abortion and breast cancer risk, as some animal experiments suggest that incomplete differentiation of the mammary gland during the first trimester may render the breast more susceptible to carcinogenesis (Russo and Russo, 1980). Results from epidemiologic studies, however, have been inconsistent (Kelsey, Gammon, and John, 1993). By far the strongest study to date on the association between breast cancer and abortion was a population-based cohort study made up of 1.5 million Danish women born between 1935 and 1978 (Melbye, Wohlfahrt, and Olsen, 1997). Of these women, 18.4% had had one or more induced abortions. After adjusting for potential confounders, the risk of breast cancer for women with a history of induced abortion was the same as the risk for women who had no history of induced abortion. A statistically significant increase in risk was found among the very small number of women with a history of second-trimester abortion. Results from this population-based prospective cohort provide strong evidence against an increase in risk of breast cancer among women with a history of induced abortion during the first trimester. Taken as a whole, the available evidence does not support any important relation between induced abortion and risk of breast cancer.

Age at Menopause As noted earlier, the rate of increase in breast cancer incidence slows at menopause (Fig. 51–1), which marks the termination of the monthly cycling of hormones that induce regular breast cell proliferation. Early studies of age at menopause showed that women who undergo bilateral oophorectomy at a young age have a greatly reduced risk of breast cancer (Trichopoulos, MacMahon, and Cole, 1972; Lilienfeld, 1956). On average, the risk of breast cancer increases by some 3% per year of delay in menopause (Colditz and Rosner, 2000). The effect of artificial menopause by either bilateral oophorectomy or pelvic irradiation appears to be somewhat greater than the effect of natural menopause, due to the immediate cessation of ovarian function rather than a gradual decline over months or years (Bernstein, 2002). Though some studies suggest the effect of age at menopause decreases with advancing age at breast cancer diagnosis (Collaborative Group on Hormonal Factors in Breast Cancer, 1996), this is likely due to greater error in recall of age at menopause as women are further removed from the event (Colditz et al., 1987). Adjustment for error in recall removes this apparent decrease in the effect of menopause with advancing age.

Exogenous Hormones Oral Contraceptives It was originally believed that oral contraceptives might increase breast cancer risk, since they contain concentrations of estrogen and progestin that could be greater than the levels of these hormones produced by a woman during a normal ovulatory cycle (Bernstein, 2002). Results of more than 50 studies have provided considerable reassurance that there is little, if any, increase in risk with oral contraceptive use in general, even among women who have used oral contraceptives for 10 or more years. However, current users and recent users (fewer than 10 years since last use) have a modest elevation in risk compared with never users. In a combined reanalysis including more than 53,000 cases of breast cancer, the relative risk for current users compared with never users was 1.24, whereas the relative risks for women 1–4 years after stopping and 5–9 years after stopping were 1.16 and 1.07, respectively (Collaborative Group on Hormonal Factors in Breast Cancer, 1996). Because most women taking oral contraceptives are young and, therefore, are at low absolute risk, these modest increases in risk will result in few additional cases of breast cancer. Nevertheless, the apparent increased risk among current and recent users should be considered in the overall decision whether to use oral contraceptives.

Postmenopausal Hormones The possible relation between postmenopausal estrogen use and risk of breast cancer has been investigated in many epidemiologic studies over the past 20 years. Overall, ever users of postmenopausal estro-

Breast Cancer gens have little or no increase in risk of breast cancer compared with women who have never used this therapy (Collaborative Group on Hormonal Factors in Breast Cancer, 1997). However, increased risk has been observed in two important subgroups: users of long duration and current users. In a large, pooled reanalysis that combined data from 51 epidemiologic studies, the investigators observed a statistically significant association between current or recent use of postmenopausal hormones and risk of breast cancer, with the strongest positive association among those with the longest duration of use (Collaborative Group on Hormonal Factors in Breast Cancer, 1997). Among women who had used postmenopausal hormones within the previous 5 years (compared with never users of postmenopausal hormones), the relative risks for duration of use were 1.1 for 1–4 years, 1.3 for 5–9 years, 1.2 for 10–14 years, and 1.6 for 15 years or more of use. No significant increase in breast cancer risk was noted for women who had quit using postmenopausal hormones 5 or more years in the past, regardless of their duration of use. Women with an uncertain age at menopause (e.g., women with simple hysterectomies) were excluded from this analysis, as inadequate accounting for age at menopause in the analysis can lead to substantial attenuation of the observed relationships between postmenopausal hormone use and breast cancer risk (Rockhill, Colditz, and Rosner, 2000). Data on recency of use have been sparse because many studies did not distinguish current from past users. In the report from the Nurses’ Health Study cohort (Colditz et al., 1995), an excess risk of breast cancer was limited to women with current or very recent use of postmenopausal hormones. In the Breast Cancer Detection Demonstration Project cohort (BCDDP), a positive association with invasive breast cancer was noted among current users of 5–15 or more years duration (Schairer et al., 1994). An underlying concern is that these data are not independent of duration of use; at any age, past users will have accumulated a shorter duration of use of postmenopausal hormones than continuing current users. The pooled reanalysis did not observe significant differences in risk according to either the type of estrogen used (conjugated estrogen vs. other) or the estrogen dose (less than 0.625 mg vs. 1.25 mg or greater), although some modest differences in estimates suggested that further evaluation is warranted (Collaborative Group on Hormonal Factors in Breast Cancer, 1997). Risk also did not appear to vary according to reproductive history, alcohol intake, smoking history, or family history of breast cancer, but the relative risk associated with 5 or more years of postmenopausal hormone use was highest among the leanest women; this interaction is consistent with findings from other studies (Huang et al., 1997; Schairer et al., 2000). The addition of a progestin to estrogen regimens has become increasingly common as it minimizes or eliminates the increased risk of endometrial hyperplasia and cancer associated with using unopposed estrogens. The impact of an added progestin on the risk of breast cancer has been evaluated only recently. Two of the first studies to assess this relationship suggested that the addition of a progestin could decrease breast cancer risk (Nachtigall et al., 1979; Gambrell, Maier, and Sanders, 1983). However, these studies were small, and potentially important confounders (e.g., age, parity) were not accounted for in the analyses. Since this time, numerous additional studies have assessed this relationship and together indicate that a protective effect of typical doses used in postmenopausal hormone therapy can be ruled out (Collaborative Group on Hormonal Factors in Breast Cancer, 1997; Ross et al., 2000; Schairer et al., 2000; Newcomb et al., 2002). Furthermore, some data suggest that the increase in breast cancer risk associated with estrogen plus progestin use may be greater than that for use of estrogen alone. In the Breast Cancer Detection Demonstration Project, the risk of breast cancer went up by about 1% for every year that women took estrogen alone and about 8% for every year that they took estrogen plus progestin (Schairer et al., 2000). According to a recent report from the Million Women Study in the United Kingdom, the relative risk of breast cancer for current users of estrogen only preparations compared with never users was 1.30 (95% CI: 1.22–1.38), whereas the relative risk for current users of estrogen plus progestin combinations was 2.00 (95% CI: 1.91–2.09); this observed difference in the magnitudes of the associated risk was highly signif-

999

icant (Million Women Health Study Collaborators, 2003). In the Women’s Health Initiative, a randomized controlled trial, even relatively short-term use of estrogen plus progestin increased the incidence of breast cancer relative to placebo, and breast cancers among women using estrogen plus progestin were diagnosed at a more advanced stage (Chlebowski et al., 2003). Because widespread use of estrogen plus progestin is so recent, few data are currently available to evaluate the effect of different formulations, doses, or schedules of use of progestin on risk of breast cancer. The results from the Million Women Study, however, indicate little variation in risk based on specific doses or regimens (Million Women Health Study Collaborators, 2003).

Nutritional Factors Dietary Fat The relation between fat intake and breast cancer risk has been the focus of a large number of studies and has received substantial public attention. High-fat diets have long been known to increase the occurrence of mammary tumors in rodents, but fat consumption has been confounded by energy intake in many animal experiments, rendering the interpretation of these data difficult. The dietary fat hypothesis is largely based on the observation that national per capita fat consumption is highly correlated with breast cancer mortality rates (Armstrong and Doll, 1975). A serious problem with ecologic comparisons of diet and breast cancer, however, is the potential for confounding by known and suspected breast cancer risk factors (e.g., low parity, late age at first birth) that have vastly different distributions among regions of the world. Case-control studies that have accounted for confounding by total energy intake and other risk factors also have indicated a weak positive association between fat intake and breast cancer risk. Howe et al. (1990) conducted a meta-analysis to summarize the results from 12 smaller case-control studies comprising a total of 4312 cases and 5978 controls. The overall pooled relative risk for a 100-gram increase in daily total fat intake was 1.35, and the risk was somewhat stronger for postmenopausal than premenopausal women. However, because the average total fat consumption is about 70 grams per day for US women, a reduction in fat intake as large as 100 grams would be impossible for almost all women. Furthermore, relative risks of this magnitude in case-control studies may easily be due to selection bias (the controls are drawn from a population with a different distribution of fat intake than the distribution in the population that gave rise to the cases) or recall bias (the cases, knowing their diagnosis, differentially misreport their pre-diagnosis diet) (Giovannucci et al., 1993). Prospective cohort studies should not be subject to these biases because the population that gives rise to the cases is known and dietary information is collected before knowledge of disease. A pooled reanalysis has been conducted of all the prospective studies, including a total of 4980 cases of breast cancer among 337,819 women (Hunter et al., 1996). Overall, no association was observed between intake of total, saturated, monounsaturated, or polyunsaturated fat and risk of breast cancer, and no reduction in risk was seen even for fat intakes as low as 20% of energy. In the Women’s Health Initiative, a large randomized controlled trial, a low-fat dietary pattern did not lead to a significant reduction in risk of invasive breast cancer among women in the intervention group (Prentice et al., 2006). Although the results from the pooled analysis do not support this, some other findings have indicated that specific types of fat could differentially affect risk of breast cancer. In most animal studies, diets high in polyunsaturated fat (linoleic acid), but typically at levels beyond human exposure, have clearly increased the occurrence of mammary tumors, but a positive association has not been found in prospective epidemiologic studies (Hunter et al., 1996). In contrast, high intake of omega-3 fatty acids from marine oils has inhibited the occurrence of mammary tumors in animals, but case-control and cohort studies generally have found little relation between intake of omega-3 fatty acids or fish (the major source of extra long chain omega-3 fatty acids) and risk of breast cancer (Willett, 1997). Some animal studies have suggested that monounsaturated fat, in the form of olive oil, may be protective relative to other sources of energy (Cohen et al., 1991), and several epidemiologic studies have supported

1000

PART IV: CANCER BY TISSUE OF ORIGIN

these findings. For example, in a Spanish study specifically undertaken because of the high consumption of olive oil and low breast cancer rates in this population, no association was observed with total fat intake, but higher intake of olive oil was associated with reduced risk of breast cancer; women in the highest quartile of consumption had approximately 35% lower risk compared with women in the lowest quartile (Martin-Moreno et al., 1994). Similar inverse associations with olive oil or monounsaturated fat were seen in case-control studies in Greece, Italy, and elsewhere in Spain (Willett, 1997). In a recent report from the Nurses’ Health Study II, high intake of animal fat, but not vegetable fat, in early adulthood was associated with elevated breast cancer risk (Cho et al., 2003). As noted by the investigators, however, a biologic mechanism to explain this observed association remains to be elucidated, and other components in food containing animal fat (e.g., heterocyclic amines, fat-soluble hormones, or growth factors) could be responsible.

Fiber Fiber has been hypothesized to lower breast cancer risk. Fiber inhibits re-absorption of estrogens in the gastrointestinal tract (Goldin et al., 1982), which may lead to lower circulating levels of estrogens, and a high-fiber diet has been associated with reduced incidence of mammary tumors in animals (Cohen et al., 1991). Case-control studies originally suggested a moderate protective effect of fiber. In their meta-analysis of 12 case-control studies, Howe et al. (1990) observed a 15% reduction in reach for a 20-gram-per-day increase in fiber intake. Prospective studies, however, have shown little or no association between fiber intake and breast cancer risk (Willett et al., 1992; Verhoeven et al., 1997; Rohan et al., 1993). It is possible that certain subfractions of fiber are related to breast cancer risk, but this remains to be determined (Willett, 2001).

Micronutrients and Fruits and Vegetables Vitamins A, C, and E and carotenoids have been examined in relation to breast cancer risk. These nutrients function as antioxidants, neutralizing free radicals that can cause DNA damage. There is little evidence of an association of retinol (preformed vitamin A) with risk, with the exception of a possible effect of intake from supplements. For b-carotene intake, most but not all studies have found that risk decreases with increasing intakes (World Cancer Research Fund and American Institute for Cancer Research, 1997), and studies of blood levels of carotenoids also suggest decreasing risk with higher levels (Dorgan et al., 1998; Toniolo et al., 2001; Tamimi et al., 2005). Higher intakes of vitamins C and E, on the other hand, do not appear to be protective (Willett, 2001). Increasing evidence indicates that higher intake of folate is associated with reduced breast cancer risk (Freudenheim et al., 1999; Zhang et al., 1999). Furthermore, women with higher folate intake appear to be protected from the increase in risk observed with alcohol (Zhang et al., 2003), discussed in the next section. Fruits and vegetables are the major sources of intake for many of these nutrients, although fortified breakfast cereal and vitamin supplements are increasing as sources. There is some evidence that intake of fruits and vegetables may be protective against breast cancer. One review examined 70 different associations regarding particular fruits and vegetables and groups of fruits and vegetables in 21 epidemiologic studies. Most of those associations suggested some risk reduction (World Cancer Research Fund and American Institute for Cancer Research, 1997). A combined reanalysis of data from eight prospective cohort studies that included more than 350,000 women, however, observed no evidence that intake of either fruits or vegetables reduces the risk of breast cancer (Smith-Warner et al., 2001). The effect of fruit and vegetable intake on risk, therefore, remains inconclusive.

Alcohol The association between alcohol consumption and breast cancer risk has been evaluated in more than 100 investigations that now clearly support a causal relation. In an early meta-analysis of 38 case-control and cohort studies, Longnecker et al. (1994) estimated relative risks of 1.1 (95% CI: 1.1–1.2) for one drink per day, 1.2 (1.1–1.3) for two

drinks per day, and 1.4 (1.2–1.6) for three drinks per day. In a pooled analysis of the six cohort studies with data on alcohol and dietary factors that included 200 or more cases (Smith-Warner et al., 1998), the risk of breast cancer increased monotonically with increasing intake of alcohol, with no statistical evidence of heterogeneity among studies; the multivariate relative risk for a 10-gram-per-day increase in alcohol was 1.09 (95% CI: 1.04–1.13). Beer, wine, and liquor all contribute to the positive association (World Cancer Research Fund and American Institute for Cancer Research, 1997; Smith-Warner et al., 1998), strongly suggesting that alcohol per se is responsible for the increased risk. One study has shown that recent adult drinking may be more important than drinking patterns earlier in life and that reductions in consumption in midlife should reduce risks of breast cancer (Longnecker et al., 1995). In intervention studies, consumption of approximately two alcoholic drinks per day increased total and bioavailable estrogen levels in both premenopausal and postmenopausal women (Reichman et al., 1993; Dorgan et al., 2001), and single doses of alcohol acutely increased plasma estradiol levels in postmenopausal women (Ginsbury et al., 1995), suggesting a mechanism by which alcohol may increase breast cancer risk. In prospective analyses, high intake of folic acid and high plasma folate levels appear to mitigate completely the excess risk of breast cancer associated with alcohol intake (Zhang et al., 1999; Zhang et al., 2003). Because alcohol metabolites inactivate folic acid, and low folate levels are associated with increased misincorporation of uracil into DNA, this finding suggests another mechanism for the adverse effects of alcohol.

Soy and Phytoestrogens Much public interest currently focuses on the potential for phytoestrogens to reduce the risk of breast cancer. Phytoestrogens are naturally occurring plant compounds that may alter estrogen metabolism away from genotoxic metabolites (Xu et al., 2000). However, several intervention studies show no evidence to support a protective role for phytoestrogens from soy. For example, in a study in which women consumed 38 grams of soy protein daily for 5 months, premenopausal women experienced elevated plasma estradiol concentrations and no change in progesterone (Petrakis et al., 1996). Of concern, however, was that 29% of the women had epithelial hyperplasia on nipple aspirate during the months they were consuming soy protein. Growing evidence suggests that hyperplasia in nipple aspirate may be a useful marker for risk of breast cancer. In another study, women with benign or malignant breast disease who were randomized to a 60-gram soy supplement showed a significant increase in the proliferation rate of breast cells on biopsy, another potential marker of breast cancer risk, after only 14 days of soy supplementation (McMichael-Phillips et al., 1998), and similar results were seen in the normal breast tissue of premenopausal women (Hargreaves et al., 1999). In contrast, a large prospective study in Japan with 427 cases of incident breast cancer demonstrated no relation between the intake of soy products in 1970 and the risk of subsequent breast cancer during approximately 500,000 person-years of follow-up (Key et al., 1999). Given the potential for adverse effects, a priority must be to clarify the relation between phytoestrogen intake and breast cancer risk.

Anthropometric Factors Height Epidemiologic studies in a variety of populations have found that height is positively related to breast cancer risk. In a pooled analysis of seven prospective cohort studies (van den Brandt et al., 2000), the relative risk for each 5-centimeter increase in height, after controlling for other breast cancer risk factors, was 1.07 for all women (95% CI: 1.02–1.11). The relative risk for women 1.75 meters (approximately 69 inches) or taller compared with those 1.60 meters (about 63 inches) or shorter was 1.42 for premenopausal women and 1.28 for postmenopausal women. Attained height is determined by a mixture of genetic and environmental factors, with one environmental determinant being childhood energy intake (Hunter and Willett, 1993). The association between height and breast cancer risk appears to be

Breast Cancer stronger in populations where childhood growth was limited by energy deprivation, which suggests that energy intake early in life may play a role in breast carcinogenesis (Friedenreich, 2001). In addition to the association that has been observed for attained height, two casecontrol studies have suggested that women who reach their maximum height at a later age may have a reduced risk of breast cancer compared with those who stop growing at an earlier age; later age at attained height may be a marker of a later pubertal growth spurt and delayed exposure of maturing breast tissue to high levels of growth hormone and insulin-like growth factor (Li et al., 1997; Li, Stanford, and Daling, 2000). A recent record linkage study conducted in Denmark also showed that early age at peak growth and high growth rate around the time of puberty are associated with increased breast cancer risk (Ahlgren et al., 2004).

Weight and Weight Change during Adulthood Attained weight and weight change in adults summarize the balance between long-term energy intake and expenditure. The relation between adiposity and breast cancer depends on menopausal status: in affluent Western populations with high rates of breast cancer, measures of body fatness are inversely related to risk of premenopausal breast cancer, and body fatness is positively related to postmenopausal breast cancer risk. A modest inverse relation between body weight (typically used as body mass index, BMI, calculated as weight in kilograms divided by height in meters2, to account for variation in height) and incidence of premenopausal breast cancer has been consistently observed in both case-control and cohort studies (Ursin et al., 1995). Heavier premenopausal women, even at the upper limits of what are considered to be healthy weights, have more irregular menstrual cycles and increased rates of anovulatory infertility (Rich-Edwards et al., 1994), suggesting that their lower risk may be due to fewer ovulatory cycles and less exposure to ovarian hormones. In both case-control and prospective studies conducted in affluent Western countries, the association between BMI and risk of breast cancer among postmenopausal women has been only weakly positive (Howe et al., 1990; Hunter and Willett, 1993). The lack of a stronger association has been surprising because obese postmenopausal women have plasma levels of endogenous estrogens nearly twice as high as lean women. However, an elevated body mass index in a postmenopausal woman represents two opposing risks: a protective effect due to the correlation between early weight and postmenopausal weight, and an adverse effect due to elevated estrogens after menopause. For this reason, weight gain from early adult life to after menopause should be more strongly related to postmenopausal breast cancer risk than attained weight, and this has been consistently supported by both case-control (Ziegler et al., 1996) and prospective studies (Harvie et al., 2005; Huang et al., 1997; Le Marchand et al., 1988). Another reason for failing to appreciate a greater adverse effect of excessive weight or weight gain on risk of postmenopausal breast cancer is that the use of postmenopausal hormones obscures the variation in endogenous estrogens due to adiposity and elevates breast cancer risk regardless of body weight. Among women who never used postmenopausal hormones in the Nurses’ Health Study, those who gained 20 kilograms or more after age 18 had double the risk of breast cancer compared with women who maintained their weight within 2 kilograms (Huang et al., 1997). The relation between body weight and breast cancer risk in lower risk, mainly non-Western, countries has been observed to be somewhat different than in higher risk countries (Pathak and Whittemore, 1992). In general, the inverse relation between weight and premenopausal breast cancer risk has not been observed, and the association between weight and postmenopausal risk has been stronger. This difference is likely to be due to the lower prevalence of overweight among premenopausal women in these low risk countries; few women are likely to be sufficiently overweight so as to cause anovulation and a reduction in premenopausal breast cancer risk. As a result, BMI after menopause would only reflect the adverse effects of high endogenous estrogens, unopposed by a residual protective effect due to correlation with overweight in early adult life.

1001

Physical Activity The relation of physical activity to risk of breast cancer has been assessed by the International Agency for Research on Cancer, which concluded that, although studies have not been entirely consistent, the overall results support a reduction in risk with higher levels of activity (International Agency for Research on Cancer, 2002). Evidence for a dose-response effect was found in most of the studies that examined the trend. The majority of studies have focused on postmenopausal breast cancer, although there is also some evidence for a protective effect of physical activity on premenopausal disease. Upon review of more than 30 studies, it was noted that the most important time periods in life for activity are not currently known. Furthermore, which activities may offer the greatest protection has not been systematically examined. Three main biologic mechanisms have been hypothesized for how physical activity may prevent or delay breast carcinogenesis. Physical activity may influence breast cancer risk by reducing the frequency of ovulatory cycles and decreasing estrogen levels, by decreasing obesity in postmenopausal women, and by enhancing innate immune defenses against cancer cells (Brinton, Bernstein, and Colditz, 1998).

Other Environmental Factors Biologically persistent organochlorines have received considerable attention as possible causes of breast cancer. These compounds include pesticides (e.g., DDT), industrial chemicals (e.g., PCBs), and dioxins produced as combustion products of PCBs or contaminants of pesticides. Although several small studies have evaluated possible relations, the pooled analysis of data from five large studies in the northeastern United States has found no association between PCBs and DDE levels and breast cancer risk (Laden et al., 2001). Overall, recent studies have not found evidence of increased risk of breast cancer, and organochlorines appear unlikely to be major breast cancer risk factors. The relation between active cigarette smoking and risk of breast cancer has been extensively evaluated; collectively, the data provide strong evidence against any major overall relationship. Initiation of smoking early in adolescence, when breast tissue may be maximally sensitive to carcinogenic influences, has been associated with an increased risk in a large case-control study (Palmer et al., 1991). However, in the largest case-control study to date, no association with smoking at an early age was observed (Baron et al., 1996). Passive smoking has also been suggested to be an important risk for breast cancer, in part because sidestream smoke contains more carcinogenic activity per milligram than does mainstream smoke. In a cohort study of cancer mortality among Japanese women exposed to passive smoke at home (Hirayama, 1984), a slight and insignificant risk elevation was seen (crude RR = 1.3; 95% CI: 0.8–2.0). In several case-control studies, increases in risk of breast cancer have been seen, but usually without evidence of a dose-response. Despite these positive associations in several studies (Hirayama, 1984; Wells, 1991), it is difficult to reconcile the absence of an effect of heavy smoking for decades with an effect of exposure to much lower amounts of environmental smoke. A likely explanation for the positive association seen in case-control studies is methodologic bias related to the selection of controls or the retrospective recall of exposure to passive smoke. In a large prospective study, neither active nor passive smoking was associated with any appreciable risk (Egan et al., 2002). Although much popular attention is focused on lifestyle factors such as use of underarm deodorant or antiperspirant, which may contribute to higher risk of breast cancer in westernized societies, a rigorous study of this topic showed no association (Mirick, Davis, and Thomas, 2002). Numerous studies evaluating a possible relation between silicone breast implants and risk of breast cancer have failed to show any positive association. In fact, most observational studies have reported lower rates of breast cancer among women with implants (Brinton and Brown, 1997; Deapen, Bernstein, and Brody, 1997; Brinton et al., 1996; Bryant and Brasher, 1995). Overall, these data provide strong evidence that breast implants do not lead to increased risk of breast cancer.

1002

PART IV: CANCER BY TISSUE OF ORIGIN

Risk Factors in Early Life and Adolescence The majority of research on determinants of breast cancer risk has focused on risk factors in adulthood, but animal data and epidemiologic evidence now suggest that exposures in earlier periods of life may have important effects on risk. Mammary gland tissue exists in a partially undifferentiated state throughout the perinatal period, rendering it susceptible to carcinogenesis (Adami, Signorello, and Trichopoulos, 1998). Trichopoulos proposed that high concentrations of maternal estrogens during pregnancy in humans may increase the probability of breast cancer in daughters by creating a “fertile soil” for subsequent cancer initiation (Trichopoulos, 1990). This hypothesis has been supported by epidemiologic studies showing moderate positive associations between indicators of high prenatal estrogen levels—such as birthweight, maternal age, and twin pregnancies—and adult breast cancer risk; in contrast, pre-eclampsia and eclampsia, indicators of low pregnancy estrogen levels, appear to be inversely associated with risk (Potischman and Troisi, 1999). Furthermore, although little research has been conducted on exposures shortly after birth, several casecontrol studies have observed significant reductions in risk among women who were breastfed as infants (Brinton, Hoover, and Fraumeni, 1983; Freudenheim et al., 1994; Weiss et al., 1997). Findings from studies of in utero and perinatal exposures are inconsistent, however, and specific biologic mechanisms to explain the apparent associations remain unclear. Growing evidence indicates that the years between menarche and first birth are important in establishing future breast cancer risk (Colditz and Frazier, 1995). During this time period, undifferentiated cells of the breast are proliferating rapidly in response to ovarian hormones. In rats, pregnancy and lactation induce terminal differentiation of cells, which leads to lengthening of their average cell-cycling time and more time for DNA repair; exposure to carcinogens after the first pregnancy results in very few tumors (Russo et al., 1990). Studies of atomic bomb survivors in Hiroshima have shown that exposure to ionizing radiation is associated with increased breast cancer risk and that the magnitude of the increase is dependent on age at exposure as well as on dose; the younger women were at the time of the bombing, the greater their excess risk (Tokunaga et al., 1994). Among girls who were treated with repeated fluoroscopy for tuberculosis or with ionizing radiation for Hodgkin disease, younger age at exposure to radiation also confers greater breast cancer risk (Miller et al., 1989; Hancock, Tucker, and Hoppe, 1993). Lifestyle factors at young ages may also affect breast cancer risk. Greater body fatness during childhood and adolescence has been associated with reduced breast cancer risk (Baer et al., 2005; Berkey et al., 1999), which may be due to increased frequency of menstrual irregularities and anovulatory cycles among overweight girls (Stoll, 1998). An early cross-sectional study observed a lower prevalence of breast cancer and benign tumors of the breast among former college athletes (Frisch et al., 1985), and several case-control studies have reported significant decreases in risk among women who participated in moderate or strenuous physical activity during adolescence (Mittendorf et al., 1995; Bernstein et al., 1994; Marcus et al., 1999). Physical activity at early ages may delay menarche, increase the frequency of anovulatory cycles, and decrease levels of estrogens (Hoffman-Goetz et al., 1998). Certain dietary factors during adolescence also may affect risk of breast cancer (Frazier et al., 2003) and benign breast disease (Baer et al., 2003). Although further research in this area is necessary, these findings suggest that adolescence may constitute an important time period for breast cancer prevention (Colditz and Frazier, 1995).

HOST FACTORS Endogenous Sex Hormones Variation in the rate of breast cancer before and after menopause, together with the role of several reproductive variables, point to a central role for sex hormones in the etiology of breast cancer. After menopause, adipose tissue is the major source of estrogen, and obese

postmenopausal women have both higher levels of endogenous estrogens and a higher risk of breast cancer (Harris et al., 1992; Huang et al., 1997). In animals, estrogens and progesterone promote mammary tumors. Also, hormonal manipulations such as anti-estrogens (e.g., tamoxifen and raloxifene) reduce breast cancer incidence (Fisher et al., 1998; Cummings et al., 1999). Despite this strong body of evidence suggesting a central role of endogenous estrogens, studies directly relating blood or urinary hormone levels to risk of breast cancer have been viewed as inconsistent, perhaps due to methodologic issues and laboratory variability. The recent availability of data from prospective studies has facilitated a far clearer picture. The combined evidence from these prospective studies when reanalyzed using consistent analytic approaches shows that estrogen and testosterone each contribute significantly to risk of breast cancer (The Endogenous Hormones and Breast Cancer Collaborative Group, 2002).

Estrogens and Estrogen Metabolites Estradiol, which circulates in blood either unbound (“free”) or bound to sex hormone binding globulin (SHBG) or albumin, is considered the most biologically active endogenous estrogen. Among postmenopausal women, the most consistent finding is a positive relationship between total estradiol and risk of breast cancer. In a recent reanalysis that combined data from nine prospective studies and included over 650 incident cases of breast cancer, the relative risks for women with increasing quintiles of estradiol concentrations were 1.4, 1.2, 1.8, and 2.0, respectively, compared with the lowest quintile (The Endogenous Hormones and Breast Cancer Collaborative Group, 2002). Free or bioavailable (free plus albumin bound) estradiol is thought to be readily available to breast tissue and thus may be more strongly related to risk than total estradiol. In the combined reanalysis, the relative risk for women with the highest quintile of free estradiol concentration, compared with the lowest quintile, was 2.6 (The Endogenous Hormones and Breast Cancer Collaborative Group, 2002). In a nested case-control analysis within the Nurses’ Health Study that was published since the combined reanalysis, circulating estradiol levels were most strongly associated with risk of ER+/PR+ tumors, with more than a threefold increase in risk for the highest vs. the lowest quartile (Missmer et al., 2004). Data on premenopausal estrogen levels and breast cancer risk are more limited, in large part because of the complexities related to sampling during the menstrual cycle. Data from several case-control studies, but not others, suggest that high levels of estradiol in premenopausal women increase the risk of breast cancer (Bernstein and Ross, 1993). In a prospective analysis involving 61 cases, estradiol levels were 12% higher in women who developed breast cancer than in controls (Thomas et al., 1997). A second prospective study with 22 cases reported a nonsignificant increased risk for women with higher estrone and estradiol levels in the follicular phase but not the luteal phase of the menstrual cycle (Helzlsouer et al., 1994). In the largest prospective analysis to date, which included 285 cases and 555 matched controls within the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort, there was no clear relation between breast cancer risk and premenopausal serum levels of estrone or estradiol (Kaaks et al., 2005). To date, few studies have evaluated free or bioavailable estradiol levels in premenopausal women (Bernstein and Ross, 1993). A woman’s pattern of estrogen metabolism also has been hypothesized to influence her breast cancer risk. Estradiol and estrone can be metabolized through one of two mutually exclusive pathways, the 16a and 2-hydroxy pathways (Yager and Liehr, 1996). Products of these two pathways have markedly different biological properties, and opposing hypotheses have been proposed concerning their influence on risk (Yager and Liehr, 1996). The epidemiologic studies that have evaluated associations between estrogen metabolites and breast cancer risk have been small and results are inconsistent. Given the limited data to address these hypotheses, no conclusions can yet be drawn.

Androgens Androgens have been hypothesized to increase breast cancer risk either directly, by increasing the growth and proliferation of breast

Breast Cancer

1003

cancer cells, or indirectly, by their conversion to estrogen (Bernstein and Ross, 1993). In humans, dehydroepiandrosterone (DHEA), an adrenal androgen, may act like an antiestrogen premenopausally, but an estrogen postmenopausally in stimulating cell growth (Ebeling and Koivisto, 1994), in part because its metabolite, 5-androstene-3b,17bdiol, also can bind to the estrogen receptor (Seymour-Munn and Adams, 1983). In postmenopausal women, testosterone has been positively associated with breast cancer in most (Thomas et al., 1997; ZeleniuchJacquotte et al., 1997; Dorgan et al., 1996; Berrino et al., 1996) but not all (Wysowski et al., 1987) studies. However, the association has tended to be attenuated after controlling for plasma estrogen (Bernstein and Ross, 1993; Thomas et al., 1997). In the combined reanalysis of prospective studies, there was approximately a 40% increase in risk associated with a doubling of testosterone levels; the magnitude of this association was reduced after adjustment for estradiol, but remained statistically significant (The Endogenous Hormones and Breast Cancer Collaborative Group, 2002). In the recent analysis from the Nurses’ Health Study, circulating levels of testosterone and other androgens among postmenopausal women were most strongly associated with risk of ER+/PR+ breast tumors, as was seen for estradiol (Missmer et al., 2004). In the EPIC cohort, higher levels of testosterone and androstenedione were also associated with increased risk of breast cancer in premenopausal women (Kaaks et al., 2005).

(Michels et al., 1996) and height (Hunter and Willett, 1993), which are positively correlated with IGF-I levels (Lassarre et al., 1991; Juul et al., 1994). Relationships between blood levels of IGF-I, its major binding protein, IGFBP-3, and breast cancer risk have been evaluated in several recent epidemiologic studies. A prospective case-control study nested within the Nurses’ Health Study, including 800 incident cases of breast cancer, found a modest positive association between plasma concentrations of IGF-I and risk of breast cancer in premenopausal women, with a relative risk of 1.6 for the top vs. the bottom tertile; this association was stronger among premenopausal women who were younger than age 50, with a relative risk of 2.5 for the top vs. the bottom tertile (Schernhammer et al., 2005). In contrast, there was no association between IGF-I concentrations and breast cancer risk in postmenopausal women. A similar prospective analysis in a cohort of women from New York City, including 287 cases of breast cancer, also found that breast cancer risk increased with increasing concentrations of serum IGF-I among premenopausal, but not postmenopausal women (Toniolo et al., 2000). The difference between the observed associations for premenopausal and postmenopausal women may reflect an effect of IGF-I that occurs early in life, soon after breast development (Hankinson, Willett, Colditz et al., 1998).

Progesterone

A family history of breast cancer is a well-established risk factor for breast cancer; women with a first-degree relative with the disease have a 1.5-fold to threefold increased risk of breast cancer compared to women without. On average, 5%–10% of breast cancers are due to inherited genetic mutations (Bennett, Gattas, and Teh, 1999). Hereditary breast cancers are often characterized by an early age of disease onset and an excess of bilateral breast cancer. Two to five percent of all breast cancers are estimated to be attributable to germline mutations in BRCA1 and/or BRCA2 (Easton, Ford, and Peto, 1993; Oesterreich and Fuqua, 1999). Recent research has indicated that these two genes are involved in genome stability, DNA repair, and cell cycle checkpoint control (Unger and Weber, 2000). The lifetime risk of breast cancer among mutation carriers is reported to be as high as 80%, although early estimates were based on high-risk families and are likely to overestimate the overall penetrance (Struewing et al., 1997; Thorlacius et al., 1997; Ford et al., 1998; Easton, Ford, and Bishop, 1995). Women of Ashkenazi Jewish ancestry or from Iceland or Poland are more likely to harbor mutations in the BRCA genes (Narod, 2002). p53 is a tumor suppressor gene associated with hereditary breast cancer. Li-Fraumeni syndrome is a rare cancer syndrome linked to mutations in p53. Individuals with this syndrome are at increased risk of leukemias and cancers of the lung, brain, and breast. Germline mutations in p53 are relatively rare and thus do not contribute to a large portion of breast cancers. Mutations in the PTEN gene are responsible for Cowden disease, a syndrome characterized by hamartomas and benign lesions of the skin and oral cavity along with an increased risk of breast cancer. Thirty percent to 50% of women with Cowden disease are estimated to develop breast cancer by the age of 50 (Foulkes, Rosenblatt, and Chappuis, 2001). Ataxia telangiectasia (AT) is an autosomal recessive disease characterized by neurodegeneration, cerebral ataxia, oculo-cutaneous telangiectasia, sensitivity to radiation, and a 100-fold increased risk of developing cancer compared with the general population (Swift et al., 1991). The most common cancers among AT patients are lymphomas and leukemias, although solid tumors including breast cancer are included. Women heterozygous for mutations in the ataxia telangiectasia mutated (ATM) gene, estimated to be approximately 1% of the population, are reported to have a fourfold to fivefold increased risk of breast cancer compared with non-carriers of the mutations (Swift et al., 1991; Inskip et al., 1999; Janin et al., 1999), although not all studies have confirmed this association (FitzGerald et al., 1997; Chen et al., 1998).

Progesterone exerts powerful influences on breast physiology and can influence tumor development in rodents (Kelsey, 1979). In one prospective study, serum progesterone was 95% higher in cases compared with controls in the luteal phase and 20% higher in the follicular phase (Helzlsouer et al., 1994); however, the study included only 22 cases. A nonsignificant inverse association between progesterone and breast cancer has been reported in another prospective analysis (Thomas et al., 1997). In the Nurses’ Health Study, which included 322 cases, there was no statistically significant association between progesterone levels and risk of breast cancer in postmenopausal women (Missmer et al., 2004). Additional, larger studies are needed to address this relationship in detail.

Prolactin Indirect evidence suggests that prolactin could play a role in breast carcinogenesis. Prolactin receptors have been found on more than 50% of breast tumors (Partridge and Hahnel, 1979), and prolactin increases the growth of both normal and malignant breast cells in vitro (Malarkey et al., 1983). To date, only two prospective studies of the relation between prolactin levels and breast cancer risk have been conducted. In the first, with 40 postmenopausal breast cancer cases (Wang et al., 1992), women in the top quintile of prolactin levels had a nonsignificant 63% higher risk of breast cancer compared with those in the bottom quintile. In a prospective analysis of prolactin and breast cancer risk from the Nurses’ Health Study that included 306 postmenopausal cases and 448 postmenopausal controls, a significant positive association was seen (top vs. bottom quartile comparison: RR = 2.0; 95% CI: 1.2–3.3, p trend = 0.01) (Hankinson, Willett, Michaud et al., 1998). An updated analysis that included 851 cases of postmenopausal cancer observed that the positive association between prolactin and breast cancer risk was limited to invasive cancers and was particularly strong for tumors that were ER+ (Tworoger et al., 2004).

Insulin-Like Growth Factor Insulin-like growth factor I (IGF-I) is a polypeptide hormone with structural homology to insulin, and it is regulated primarily by growth hormone (Zapf and Froesch, 1986). There is increasing evidence that the growth hormone-IGF-I axis stimulates proliferation of breast cancer (Yang et al., 1996; Pollak et al., 1990) and normal breast epithelial cells (Ruan, Newman, and Kleinberg, 1992). Positive associations have been observed between breast cancer and both birthweight

Genetic and Familial Susceptibility

1004

PART IV: CANCER BY TISSUE OF ORIGIN

Table 51–1. Summary of Low-Penetrance Genetic Polymorphisms and Breast Cancer Gene

Polymorphism

Hypothesized Function

Association with Breast Cancer

Association with/Modifiers of Breast Cancer Risk Factors

estrogen biosynthesis genes CYP17

T Æ C in 5¢UTR

Creates SP1 type promoter, hypothesized to increase transcription

No overall association

CYP19

(TTTA)n repeat in intron 5

May alter splice site, unknown

[TTTA]10 allele may be associated with ≠ risk

17 b-HSD

C Æ T in exon 10 A Æ G in exon 6 (Ser312Gly)

A2 allele may be associated with ≠ E2 levels Ø risk associated with later age at first menarche, may be restricted to A1/A1 women

Unknown

estrogen metabolism genes CYP1A1

COMT

T Æ C in 3¢ UTR (Msp I) A Æ G in exon 7 (Ile462Val) G Æ A in exon 4 (Val158Met)

Hypothesized higher enzyme activity

No overall association

Higher enzyme activity

No association

Reduced methylation activity

No overall association

Variant allele may be associated with ≠ risk among smokers

carcinogen metabolism genes GSTM1 NAT1 NAT2

Deletion of gene (null) NAT1*10 NAT2*4

No enzyme activity

No overall association

Slow acetylator Slow acetylator

No association No association

Low-Penetrance Genes

Cholesterol

There is a great deal of evidence to suggest that other genes with low penetrance may also affect breast cancer susceptibility. Low penetrance genes are expected to confer only a small amount of risk, but because the variation is likely to be more common, the population attributable risk for these genetic polymorphisms, alone or in combination with other risk factors, is likely to be high. Candidate gene studies have primarily focused on major classes of genes involved in hormone synthesis and metabolism, carcinogen metabolism, and more recently, DNA damage and repair. Although a number of studies have examined variation in candidate genes, there has been very little conclusive evidence that any of these lowpenetrance genes are associated with overall breast cancer risk. Table 51–1 summarizes the major classes of genes and polymorphisms that have been well studied with respect to breast cancer risk.

Pregnenelone

CYP17

17 a-Hydroxypregnenelone

Progesterone

CYP17

Dehydroepiandrosterone

Estrogen Biosynthesis and Metabolism Genes A number of enzymes are involved in the biosynthesis of estradiol (E2) from cholesterol (Fig. 51–2). Candidate gene studies have focused on polymorphisms within these genes, for which the variation is hypothesized to alter the functional activity of the enzyme. CYP17 plays a major role in the synthesis of androgens, whereas CYP19 and 17bHSD (17b-hydroxysteroid dehydrogenase) catalyze the final steps of converting androgens to E2. In postmenopausal women, aromatization of androstenedione to E2 in peripheral adipose tissue is the primary source of estrogen. A polymorphism in the promoter region of the CYP17 gene has been reported to create an additional SpI promoter site, which is hypothesized to increase promoter activity and ultimately estradiol levels. Two studies reported increased estradiol levels among women with the A2 allele, suggesting that the polymorphism is functional (Haiman et al., 1999; Feigelson et al., 1998). Despite these observed differences in hormone levels, however, there does not appear to be any difference in breast cancer risk associated with the A2 allele (Mitrunen and Hirvonen, 2003; Kristensen and Borresen-Dale, 2000). A polymorphic tetranucleotide repeat [TTTA]n in intron 5 of the CYP19 gene has been evaluated in relation to breast cancer risk with inconsistent results. Seven alleles have been reported at this locus, with repeats ranging from 7 to 13. The [TTTA]12 allele was associated with increased risk of breast cancer in Swedish and Norwegian women

Testosterone

Androstenedione CYP19

CYP19 17 b-HSD Estrone

Estradiol CYP1A1

CYP1B1

2OH-estradiol COMT

4OH-estradiol COMT

Figure 51–2. Estradiol synthesis and metabolism pathway. (Source: Adapted from Yager and Liehr, 1996.)

Breast Cancer (Kristensen et al., 1998), while the [TTTA]10 allele (Haiman et al., 2000) and the [TTTA]7 allele (Siegelmann-Danieli and Buetow, 1999) have both been reported to be associated with risk in US Caucasian women in different studies. A recent meta-analysis of three studies suggests that the [TTTA]10 allele may be associated with a more than twofold increased risk of breast cancer (OR = 2.33, 95% CI: 1.36–4.17) (Dunning et al., 1999). Only a few studies have examined the role of the Ser312Gly polymorphism in the 17b-HSD gene. There is no evidence that this polymorphism is associated with breast cancer risk (Feigelson et al., 2001; Mannermaa et al., 1994).

1005

evidence that women with the greatest mammographic densities are at a fourfold to sixfold increased risk of breast cancer compared to women with little or no density (Oza and Boyd, 1993; Boyd et al., 1995; Byrne et al., 1995; Byrne, 1997), making mammographic density one of the strongest independent risk factors for breast cancer. It is unclear what the biologic mechanism is for this relationship, although it has been hypothesized that mammographic density is a marker for cellular proliferation in the breast tissue (Spicer et al., 1994).

PATHOGENESIS Carcinogen Metabolism Genes The cytochrome P450 family of enzymes are Phase I enzymes and are involved in the activation of carcinogens. CYP1A1 encodes for aryl hydrocarbon hydroxylase (AHH) and is involved in the metabolic activation of polycyclic aromatic hydrocarbon and catalyzes the 2hydroxylation of E2 (Fig. 51–2). Four polymorphisms in CYP1A1 have been identified, with association studies primarily focusing on the T Æ C SNP in the 3¢ UTR (Msp1) and the A Æ G SNP in exon 7 (Ile462Val). Polymorphisms in CYP1A1 do not appear to have an overall effect on breast cancer risk, although there is evidence that variant allele carriers exposed to cigarette smoke may be at an increased risk (Ambrosone et al., 1995; Ishibe et al., 1998). The N-acetyl transferase (NAT) and glutathione-S-transferase (GST) families are both Phase II enzymes, which are involved in the inactivation of carcinogens. Both NAT1 (NAT1*10) and NAT2 (NAT2*4) have functional polymorphisms resulting in slow and fast acetylator phenotypes (Hein, 2002). The overall evidence does not support a role of these polymorphisms in breast cancer risk (Dunning et al., 1999; de Jong et al., 2002). Deletion variants in GSTM1 result in a lack of enzyme function. Individuals homozygous for these null mutations may therefore be unable to eliminate specific carcinogens effectively and may be at increased risk of cancer. Association studies have reported conflicting results for GSTM1 polymorphisms and risk of breast cancer. A pooled analysis including more than 2700 genotyped cases and controls suggests that GSTM1 is not associated with substantial increased risk of breast cancer (de Jong et al., 2002). COMT, another Phase II enzyme, is involved in the inactivation of catechol estrogens. The G Æ A polymorphism in exon 4 has been associated with a threefold to fourfold reduction in enzyme activity (Lachman et al., 1996). Overall, there does not appear to be an increased risk of breast cancer associated with this polymorphism, although risk may be modified by body mass index (Mitrunen et al., 2001; Thompson et al., 1998; Lavigne et al., 1997) and use of hormone replacement therapy (Mitrunen et al., 2001).

DNA Damage and Repair Genes Endogenous and exogenous mutagens can cause DNA damage. If the damage goes unrepaired, the cell may either undergo apoptosis or may grow in an unregulated manner and eventually lead to cancer. A few studies have examined candidate genes in DNA repair pathways including XRCC1, XRCC2, XRCC3, LIG4, RAD51, BRCA2. At this time, there is little data on the functionality and the significance of these polymorphisms in breast cancer. Although a number of studies have examined variation in candidate genes, there has been very little conclusive evidence that any of these low-penetrance genes are associated with substantial breast cancer risk. It is possible that these polymorphisms have only a small effect on risk, which cannot be detected when examined individually.

Mammographic Density The radiographic appearance of the breast on a mammogram varies depending on the composition of the individual breast. Fat is radiolucent and appears dark on mammogram, whereas epithelial cells and connective tissue are radiodense and appear light. Mammographic density can be measured continuously as the overall percentage of dense tissue in the breast or with a categorical rating system. There is

Biomathematical models relating epidemiologic risk factors to breast cancer incidence can provide a structure to view the process of carcinogenesis. In addition, such models summarize the impact of multiple variables and provide a means to identify areas that require more research (Moolgavkar, 1990). Pike and colleagues (1983) reviewed the epidemiologic evidence in the early 1980s and proposed a model of tissue aging that accounted for the relation between reproductive risk factors and breast cancer incidence. This model, which built on earlier work by Moolgavkar and Knudson (1981), was based on the observed age-incidence curve and on the known relations of ages at menarche, first birth, and menopause to the risk of breast cancer (Pike et al., 1983). The original Pike model related breast cancer rates to the growth of the breast. The model allowed a short-term increase in risk with first pregnancy followed by a subsequent decrease in risk, as well as an additional reduction in risk with the onset of perimenopause, when the breast begins an involutional process that is thought to reflect a decrease in cell turnover and eventual disappearance of epithelium. The original Pike model, however, did not include terms for the second or subsequent births or for the spacing of pregnancies, nor did it easily accommodate pregnancies after age 40 years. An extension of the Pike model utilized prospective data from the Nurses’ Health Study and added a term to summarize the spacing of births (Rosner, Colditz, and Willett, 1994; Rosner and Colditz, 1996; Colditz and Rosner, 2000). In this model, the incidence of breast cancer prior to menopause increased by 1.7% for each 1-year increase in age at first birth. Closer spacing of births was related to significantly reduced risk of breast cancer. For each additional year of delay between the first and second births, for example, the risk of breast cancer increased by 0.4%. The increase in risk with first pregnancy originally observed with this modified Pike model has since been documented in a prospective study from Sweden (Lambe et al., 1994) and in an analysis from an international case-control study (Hsieh, Pavia, and Lambe, 1994). According to the extended Pike model, a parous woman with a single birth at age 35 years has a 34% increase in breast cancer incidence at the time of the birth relative to a nulliparous woman. The excess risk goes down very slowly over time. Even at age 70 years, such a woman has excess risk compared with a nulliparous woman. In sum, the cumulative risk to age 70 is 16% greater than that of a nulliparous woman. Conversely, a parous woman with an early age at first birth (20 years of age) and multiple births conceived at a young age has a slight excess risk immediately after the first birth relative to the nulliparous woman (RR = 1.10), which slowly diminishes over time, reaching equality at age 32 years and continuing to decline until menopause (age 50) at which time the RR is 0.82. Since the relationship between breast cancer incidence and reproductive history changes with age, cumulative incidence rather than age-specific incidence is a useful summary. A woman with one birth at age 35 years has a 16% excess risk over the age period 30–70 years, whereas a woman with births at ages 20, 23, and 26 years has a 27% decrease in risk over the similar age period (Colditz and Rosner, 2000), compared with a nulliparous woman. In the original Pike model (Pike et al., 1983), factors associated with reduced risk of breast cancer were each considered to slow the rate of “breast tissue aging,” which correlates with the accumulation of molecular damage in the pathway to breast cancer. In the Rosner et al.

1006

PART IV: CANCER BY TISSUE OF ORIGIN

extension of the Pike model (Rosner and Colditz, 1996; Colditz and Rosner, 2000), the rate of tissue aging was highest between menarche and first birth, consistent with the hypothesis that this is the period when the breast is most vulnerable to mutagenesis. The transient increase in the risk of breast cancer associated with the first pregnancy is followed by a 20% decrease in the rate of breast tissue aging (Rosner and Colditz, 1996). This observation helps explain the “crossover” effect in certain subgroups of women: around menopause, rates of one subgroup that were initially higher drop below rates of a second subgroup. For instance, using data from New York State, Janerich and Hoff (1982) showed a crossover in breast cancer incidence between single and married women at age 42, such that married women had a higher incidence before this age and lower mortality thereafter. As mentioned earlier, a similar crossover of incidence also has been reported for black and white women in the United States (Gray, Henderson, and Pike, 1980; Krieger, 1990), consistent with the distribution of age at first birth by race (US Bureau of the Census, 1983). The first pregnancy is associated with permanent changes in the glandular epithelium and with changes in the biologic properties of the mammary cells. After the differentiation of pregnancy, epithelial cells have a longer cell cycle and spend more time in G1, the phase that allows for DNA repair (Russo, Tay, and Russo, 1982). The longer the interval from menarche to first pregnancy, the greater the adverse effect of the first pregnancy (Rosner and Colditz, 1996). The later the age at first full-term pregnancy, the more likely that DNA mistakes have accumulated and are propagated with the proliferation of mammary cells during pregnancy. The susceptibility of mammary tissue to carcinogens decreases after the first pregnancy, reflecting the differentiation of the mammary gland. This is seen in the agedependent susceptibility of the breast to radiation (Colditz and Frazier, 1995). The proliferation of breast cells during the first pregnancy results in differentiation into mature breast cells prepared for lactation; this may also lead to growth of mutated cells and excess risk over the next decade. Epidemiologic evidence for the transient excess risk after the first pregnancy is consistent. Less clear is the presence of a transient increase in risk after subsequent pregnancies; some studies suggest an adverse effect (Lambe et al., 1994) but others do not (Rosner, Colditz, and Willett, 1994). The reduction in risk of breast cancer with early menopause is likely due to the cessation of breast cell division with the termination of menstrual cycles, and the decline in endogenous hormone levels, which become substantially lower than during the premenopausal years. In sum, the role of hormonal factors, cell division, and the accumulation of DNA damage support the hypothesized model of cell proliferation as a key component of breast cancer etiology.

PREVENTIVE MEASURES Primary Prevention Approaches for the primary prevention of breast cancer according to period of life are discussed here briefly and are considered in more detail elsewhere (Colditz and Frazier, 1995). Unfortunately, current knowledge of breast cancer etiology does not necessarily translate easily into strategies for breast cancer prevention. Many of the established and suspected risk factors for breast cancer are given in Table 51–2. Some risk factors (such as age at menarche) are well established but difficult to modify; some (such as postmenopausal hormone use) are well established and carry documented risks and benefits; and others (such as replacing saturated fat with monounsaturated fat and the addition of a daily multivitamin to raise folate levels) are unproven but have other strong benefits that justify the strategy, with reduction in breast cancer being a possible additional benefit. Also, known risk factors for breast cancer are modest in magnitude; relative risks are usually in the range of 1.3–1.8 for attainable changes. Although these relative risks are far less dramatic than that between smoking and lung cancer, they are still important. To provide perspective, the relative risk of death from breast cancer for women who do not have mammography compared with those who receive regular mammograms is about 1.3. As we give great importance and resources to the provision of mammography, the avoidance of a risk factor with a similar magnitude of effect should have even higher priority because this prevents both the occurrence and need for treatment of breast cancer as well as death. When considering primary prevention, it is important to remember that even small changes at the individual level can produce substantial changes in the population rates of disease (Rose, 1981). Early onset of menarche in the United States and other affluent countries is largely the result of rapid growth and weight gain of children related to an abundant food supply, excellent sanitation, and low levels of physical activity (including sitting in school). Much of this is desirable for many reasons, and there is no reasonable expectation that we could, or would want, to increase the average age at menarche to 17 years, as has been typical in rural China (Chen et al., 1990). Still, generally desirable increases in physical activity, such as greater recreational activities, have been associated with modest delays in age at menarche (Merzenich, Boeing, and Wahrendorf, 1993; Mosian, Meyer, and Gingras, 1991) and should thus contribute to reductions in breast cancer. The amount of time spent television watching is a major determinant of excessive weight gain by children (Gortmaker, Dietz, and Cheung, 1990; Berkey et al., 2003) and thus an appropriate focus for reducing risk of breast cancer as well as future cardiovascular disease and diabetes. Society, through the provision of daily physical activity in schools and safe environments for recreational activity, must play a major role in these efforts.

Table 51–2. Established and Suspected Risk Factors for Breast Cancer and Approximate Strength of Association Reproductive Factors Age at menarche (≥15 vs. 11) Age at first birth (≥35 vs. £20) Number of births (0 or 1 child) Age at menopause (5-year increment) Lactation (>1 yr vs. none)

Nutritional and Anthropometric Factors

Hormonal Factors ++ + + -

Estrogen replacement (<5 year vs. none) Oral contraceptive use (current vs. none) High blood estrogens (postmenopausal) High blood IGF-I (premenopausal) High blood prolactin

Other Factors

+

Monounsaturated fat

-

+

Saturated fat

+

+++

Alcohol (>1 drink/day vs. none) Height (>5¢7≤)

+

+++ ++

Source: Adapted from Colditz et al., 2000. +, Relative Risk (RR) = 1.1–1.4; ++, RR = 1.5–2.9; +++, RR = 3.0–6.9; -, RR = 0.7–0.8.

Premenopausal obesity (>27 vs. <21 BMI) Postmenopausal obesity (>27 vs. <21 BMI) Physical activity (≥3 hour/week)

+ + -

Family history (firstdegree relative) Benign breast disease (MD-diagnosed) Ionizing radiation (yes/no)

+++ ++ +

Breast Cancer Early age at first birth will substantially reduce breast cancer incidence, but societal trends are in the opposite direction. Furthermore, unplanned early pregnancies have undesirable social and ecological consequences. Nevertheless, a social norm that encouraged carefully planned first pregnancies at the beginning of advanced education and career development would reduce breast cancer rates. This would require major behavioral changes and social supports, such as for childcare, to be practical on a widespread basis. At least 6 months of lactation is recommended for optimal infant health (Committee on Nutrition and American Academy of Pediatrics, 1993), and evidence suggests this will modestly reduce risk of breast cancer, particularly among premenopausal women. Improved physician counseling (Freed et al., 1995) can encourage this practice, but changes at workplaces to allow breastfeeding and longer maternity leaves will also be needed for many women to adopt this practice. Alcohol consumption has a complex mix of desirable and adverse health effects, one being an increase in breast cancer risk. Individuals should make decisions considering all the risks and benefits, but for a middle-aged woman who drinks alcohol on a daily basis, reducing intake is one of relatively few behavioral changes that is likely to reduce risk of breast cancer. Taking a multivitamin containing folic acid greatly reduces risks of neural tube defects and may prevent coronary heart disease (Rimm et al., 1996) and colon cancer (Giovannucci et al., 1998), and growing evidence suggests this may mitigate the excess risk of breast cancer due to alcohol (Zhang et al., 2003). Thus, taking a multivitamin appears sensible for women who do elect to drink regularly. Postmenopausal hormone use, like alcohol consumption, involves a complex trade-off of benefits and risks. From the standpoint of breast cancer risk, the optimal strategy would be to use estrogens not at all, or at most for a few years to relieve menopausal symptoms. The range of options, however, is rapidly increasing with the availability of selective estrogen receptor modulators, such as tamoxifen and raloxifene, that can prevent the progression of osteoporosis and simultaneously reduce risk of breast cancer. Physicians will need to play a key role in advising women in this rapidly evolving field. Avoiding weight gain during adult life can importantly reduce risk of postmenopausal breast cancer as well as cardiovascular disease and many other important conditions. Individual women can reduce weight gain by exercising regularly and moderately restraining caloric intake. Health care providers play an important role in counseling patients throughout adult life about the importance of weight control. However, the incorporation of greater physical activity into daily life will be difficult for many persons unless governments provide a safer and more accessible environment for pedestrians and bicycle riders. The provision of worksite and community exercise facilities can also contribute importantly. Specific aspects of diet that influence risk of breast cancer are not yet well established, but available evidence generally suggests that increasing folate intake can modestly reduce risk and that replacing saturated and trans fat with monounsaturated fat may reduce risk. These are reasonable strategies to pursue because this will reduce risk of cardiovascular and other diseases, and reduced risk of breast cancer may be an added benefit. Physicians can assess dietary habits and provide guidance, and governmental policies influence diets in many ways. Providing the best current information on diet and health is one such role. With demonstration that tamoxifen, and probably other selective estrogen receptor modulators, can be effective in the primary prevention of breast cancer (Fisher et al., 1998; Cummings et al., 1999), chemoprevention has become an option for women at elevated risk. Many other pharmacologic agents are being evaluated at present and are likely to increase the alternatives. The availability of effective chemopreventive agents raises many questions about the optimal criteria for use of these drugs. Evaluation of an individual woman’s risk of breast cancer has become much more important because this risk can now be modified. Until recently, risk has been primarily based on an evaluation of family and reproductive history and history of benign breast disease. New information on risk based on genotype, detailed histologic characteristics of benign breast disease (Jacobs et al., 1999),

1007

and serum hormone levels (The Endogenous Hormones and Breast Cancer Collaborative Group, 2002) now allows a much more powerful prediction of risk for an individual woman. The efficacy of raloxifene among women with higher baseline estrogen levels (Cummings et al., 2002) adds to the need to refine strategies to identify women with elevated hormones and other lifestyle factors that may combine to place them at increased risk of breast cancer to better tailor chemoprevention efforts and maximize the risk-benefit trade off. Screening for elevated estrogen levels in postmenopausal women to help identify those who would most benefit from an estrogen antagonist, as is done for serum cholesterol, may become part of medical practice. In summary, available evidence provides a basis for a number of strategies that can reduce risk of breast cancer, although some of these represent complex decision making. Attainable objectives can make an important impact on individual risk of breast cancer. However, the collective implementation of all lifestyle strategies will not reduce population rates of breast cancer to the very low levels of traditional poor societies because the magnitude of the necessary changes is unrealistic or undesirable.

Screening There is general consensus that appropriate regular screening is effective in reducing breast cancer mortality. Mammography and clinical breast exam are each reasonably sensitive (detecting as many cases as possible), but their sensitivity improves when they are used in combination. There is no debate about the life-saving importance of mammography for women over age 50. This recommendation is based on scientific evidence that has been accumulating since the 1960s from randomized controlled trials conducted in the United States, Canada, and Europe. In these trials, women who were screened annually with mammography had a 25%–30% lower risk of dying from breast cancer than their peers who were not screened annually (Smith et al., 2003). A number of randomized controlled trials have shown that mammography in women ages 40–49 can lower the risk of dying from breast cancer. However, some debate still remains about the exact magnitude and timing of this benefit. In several of the trials, a reduction in mortality did not occur until at least 10 years into the study, when the women had moved into the 50–69-year-old age group. Hence, it is unclear whether the reduction in mortality was due to screening that was performed while the women were in the 40–49year-old age group or the 50–69-year-old age group. Unfortunately, most trials to date may not have included enough women under age 50 to be able to answer this question with certainty. It may be that regular mammography can reduce breast cancer deaths in younger women but that most of the studies were not large enough to show this. Looking at the evidence as a whole, annual mammograms in women ages 40–49 probably lower the risk of dying from breast cancer by about 20% (Kerlikowske, Grady, and Ernster, 1995; Antman and Shea, 1999). The reason that mammography might be less beneficial in younger women than in older women is not yet clear. One explanation is that mammographic technology is not sensitive enough to discern abnormal tissue from normal tissue in the breasts of many younger women. This is because breast tissue tends to be dense in younger women, and on a mammogram, dense tissue resembles an abnormality.

Conclusion for Prevention Given the evidence reviewed above, a role will exist for hormonal and other chemopreventive interventions that may be appropriate for women at particularly high risk and, potentially, for wide segments of the population, as few women can be considered to have very low risk. Together, the modification of nutritional and lifestyle risk factors and the judicious use of chemopreventive agents can have a major impact on incidence of this important disease. Such strategies will complement early detection through screening mammography programs to reduce the mortality burden from breast cancer.

1008

PART IV: CANCER BY TISSUE OF ORIGIN

FUTURE DIRECTIONS To reduce the global burden of breast cancer during the 21st century, it will be necessary to identify and implement novel strategies for prevention as well as to uncover new important environmental and genetic factors involved in the etiology of this disease. With increasing ability to synthesize risk factors into prediction models, it will be necessary to refine the ability to relate risk to individual women, identifying those who will benefit from chemopreventive measures and those who will not. Among the areas that require more study is the spectrum of endogenous hormones and their relation to risk. Few data are available for progesterone, for example, and data are also sparse among premenopausal women, where timing of blood collection within the menstrual cycle has important implications for interpretation of data. Because there is redundancy in the genome and because the majority of the genes evaluated in epidemiologic studies to date have their effects in pathways, it will be necessary to examine multigene and pathway effects rather than isolated single nucleotide polymorphisms. Finally, from the life course perspective of risk accumulation, greater understanding of lifestyle from menarche to first pregnancy (and particularly during adolescence) in relation to breast cancer risk will add substantially to our ability to understand the natural history of this disease and to develop effective prevention strategies at both the individual and population levels. References ACS. 2005. Breast Cancer Facts & Figures, 2005. Atlanta, GA: American Cancer Society. Adami HO, Signorello LB, Trichopoulos D. 1998. Towards an understanding of breast cancer etiology. Semin Cancer Biol 8(4):255–262. Ahlgren M, Melbye M, Wohlfahrt J, Sorensen TI. 2004. Growth patterns and the risk of breast cancer in women. N Engl J Med 351(16):1619–1626. Ambrosone CB, Freudenheim JL, Graham S, et al. 1995. Cytochrome P4501A1 and glutathione S-transferase (M1) genetic polymorphisms and postmenopausal breast cancer risk. Cancer Res 55(16):3483–3485. American Academy of Pediatrics Committee on Nutrition. 1993. Pediatric Nutrition Handbook. 3rd ed. Elk Grove Village, IL: American Academy of Pediatrics. Antman K, Shea S. 1999. Screening mammography under age 50. JAMA 281(16):1470–1472. Armstrong B, Doll R. 1975. Environmental factors and cancer incidence and mortality in different countries, with special reference to dietary practices. Int J Cancer 15:617–631. Baer HJ, Colditz GA, Rosner B, et al. 2005. Body fatness during childhood and adolescence and incidence of breast cancer in premenopausal women: a prospective cohort study. Breast Cancer Res 7(3):R314–R325. Baer HJ, Schnitt SJ, Connolly JL, et al. 2003. Adolescent diet and incidence of proliferative benign breast disease. Cancer Epidemiol Biomarkers Prev 12(11 Pt 1):1159–1167. Baron JA, Newcomb PA, Longnecker MP, et al. 1996. Cigarette smoking and breast cancer. Cancer Epidemiol Biomarkers Prev 5:399–403. Bennett IC, Gattas M, Teh BT. 1999. The genetic basis of breast cancer and its clinical implications. Aust N Z J Surg 69(2):95–105. Berkey CS, Frazier AL, Gardner JD, Colditz GA. 1999. Adolescence and breast carcinoma risk. Cancer 85(11):2400–2409. Berkey CS, Rockett HR, Gillman MW, Colditz GA. 2003. One-year changes in activity and in inactivity among 10- to 15-year-old boys and girls: Relationship to change in body mass index. Pediatrics 111(4 Pt 1):836–843. Bernstein L. 2002. Epidemiology of endocrine-related risk factors for breast cancer. J Mammary Gland Biol Neoplasia 7(1):3–15. Bernstein L, Henderson BE, Hanisch R, Sullivan-Halley J, Ross RK. 1994. Physical exercise and reduced risk of breast cancer in young women. J Natl Cancer Inst 86:1403–1408. Bernstein L, Ross R. 1993. Endogenous hormones and breast cancer risk. Epidemiol Rev 15:48–62. Berrino F, Muti P, Micheli A, et al. 1996. Serum sex hormone levels after menopause and subsequent breast cancer. J Natl Cancer Inst 88:291–296. Boyd NF, Byng JW, Jong RA, et al. 1995. Quantitative classification of mammographic densities and breast cancer risk: Results from the Canadian National Breast Screening Study. J Natl Cancer Inst 87(9):670–675. Brinton L, Bernstein L, Colditz G. 1998. Summary of workshop on physical activity and breast cancer, November 13–14, 1997. Cancer 83 (suppl): 595–599. Brinton LA, Brown SL. 1997. Breast implants and cancer. J Natl Cancer Inst 89:1341–1349.

Brinton LA, Malone KE, Coates RJ, et al. 1996. Breast enlargement and reduction: Results from a breast cancer case-control study. Plast Reconstr Surg 97:269–275. Brinton LA, Hoover R, Fraumeni JF. 1983. Reproductive factors in the aetiology of breast cancer. Br J Cancer 47:757–762. Bruzzi P, Negri E, La Vecchia C. 1988. Short term increase in risk of breast cancer after full term pregnancy. Br Med J 297:1096–1098. Bryant H, Brasher P. 1995. Breast implants and breast cancer—reanalysis of a linkage study. N Engl J Med 332:1535–1539. Byrne C. 1997. Studying mammographic density: Implications for understanding breast cancer. J Natl Cancer Inst 89(8):531–533. Byrne C, Schairer C, Wolfe J, et al. 1995. Mammographic features and breast cancer risk: Effects with time, age, and menopause status. J Natl Cancer Inst 87(21):1622–1629. Campbell JB. 2002. Breast cancer-race, ethnicity, and survival: A literature review. Breast Cancer Res Treat 74(2):187–192. Centers for Disease Control and Prevention. 1992. Breast Cancer on Long Island, New York. Washington, DC: US Department of Health and Human Services. Chen J, Birkholtz GG, Lindblom P, Rubio C, Lindblom A. 1998. The role of ataxia-telangiectasia heterozygotes in familial breast cancer. Cancer Res 58(7):1376–1379. Chen J, Campbell TC, Junyao L, Peto R. 1990. Diet, Life-Style, and Mortality in China: A Study of the Characteristics of 65 Chinese Counties. Oxford, England: Oxford University Press. Chevarley F, White E. 1997. Recent trends in breast cancer mortality among white and black US women. Am J Public Health 87:775–781. Chlebowski RT, Hendrix SL, Langer RD, et al. 2003. Influence of estrogen plus progestin on breast cancer and mammography in healthy postmenopausal women: the Women’s Health Initiative Randomized Trial. JAMA 289(24):3243–3253. Cho E, Spiegelman D, Hunter DJ, et al. 2003. Premenopausal fat intake and risk of breast cancer. J Natl Cancer Inst 95(14): 1079–1085. Chu KC, Lamar CA, Freeman HP. 2003. Racial disparities in breast carcinoma survival rates: Separating factors that affect diagnosis from factors that affect treatment. Cancer 97(11):2853–2860. Chu K, Tarone R, Kessler L. 1996. Recent trends in US breast cancer incidence, survival, and mortality rates. J Natl Cancer Inst 88:1571–1579. Cohen LA, Kendall ME, Zang E, Meschter C, Rose DP. 1991. Modulation of N-nitrosomethylurea-induced mammary tumor promotion by dietary fiber and fat. J Natl Cancer Inst 83:496–501. Colditz GA, Atwood KA, Emmons K, et al. 2000. Harvard report on cancer prevention volume 4: Harvard Cancer Risk Index. Risk Index Working Group, Harvard Center for Cancer Prevention. Cancer Causes Control 11(6):477–488. Colditz GA, Stampfer MJ, Willett WC, et al. 1987. Reproducibility and validity of self-reported menopausal status in a prospective cohort study. Am J Epidemiol 126:319–325. Colditz, G, Rosner B, Chen WY, Holmes M, Hankinson SE. 2004. Risk factors for breast cancer: Model fitting and application of polychotomous logistic regression to evaluate risk factors for breast cancer according to estrogen and progesterone receptor status. J Natl Cancer Inst 96: 218–228. Colditz GA, Hankinson SE, Hunter DJ, et al. 1995. The use of estrogens and progestins and the risk of breast cancer in postmenopausal women. N Engl J Med 332:1589–1593. Colditz G, Frazier A. 1995. Models of breast cancer show that risk is set by events of early life: Prevention efforts must shift focus. Cancer Epidemiol Biomarkers Prev 4:567–571. Colditz GA, Rosner BA. 2000. Cumulative risk of breast cancer to age 70 years according to risk factor status: Data from the Nurses’ Health Study. Am J Epidemiol 152(10):950–964. Collaborative Group on Hormonal Factors in Breast Cancer. 1996. Breast and hormonal contraceptives: Collaborative reanalysis of individual data on 53,297 women with breast cancer and 100,239 women without breast cancer from 54 epidemiological studies. Lancet 347:1713–1727. Collaborative Group on Hormonal Factors in Breast Cancer. 1997. Breast cancer and hormone replacement therapy: Collaborative reanalysis of data from 51 epidemiologic studies of 52,705 women with breast cancer and 108,411 women without breast cancer. Lancet 350:1047–1059. Collaborative Group on Hormonal Factors in Breast Cancer. 2002. Breast cancer and breastfeeding: Collaborative reanalysis of individual data from 47 epidemiological studies in 30 countries, including 50302 women with breast cancer and 96973 women without the disease. Lancet 360(9328):187–195. Collins LC, Tamimi RM, Baer HJ, et al. 2005. Outcome of patients with ductal carcinoma in situ untreated after diagnostic biopsy: results from the Nurses’ Health Study. Cancer 103(9):1778–1784.

Breast Cancer Cummings SR, Duong T, Kenyon E, Cauley JA, Whitehead M, Krueger KA. 2002. Serum estradiol level and risk of breast cancer during treatment with raloxifene. JAMA 287(2):216–220. Cummings SR, Eckert S, Krueger KA, et al. 1999. The effect of raloxifene on risk of breast cancer in postmenopausal women. JAMA 281:2189– 2197. de Jong MM, Nolte IM, te Meerman GJ, et al. 2002. Genes other than BRCA1 and BRCA2 involved in breast cancer susceptibility. J Med Genet 39(4):225–242. Deapen DM, Bernstein L, Brody GS. 1997. Are breast implants anticarcinogenic? A 14-year follow-up of the Los Angeles Study. Plast Reconstr Surg 99:1346–1353. Dorgan JF, Longcope C, Stephenson HE, et al. 1996. Relation of prediagnostic serum estrogen and androgen levels to breast cancer risk. Cancer Epidemiol Biomarkers Prev 5:533–539. Dorgan JF, Sowell A, Swanson CA, et al. 1998. Relationships of serum carotenoids, retinol, a-tocopherol, and selenium with breast cancer risk: Results from a prospective study in Columbia, Missouri (United States). Cancer Causes Control 9:89–97. Dorgan JF, Baer DJ, Albert PS, et al. 2001. Serum hormones and the alcoholbreast cancer association in postmenopausal women. J Natl Cancer Inst 93:710–715. Dunn J. 1975. Cancer epidemiology in populations of the United States—with emphasis on Hawaii and California—and Japan. Cancer Res 35:3240– 3245. Dunning AM, Healey CS, Pharoah PD, Teare MD, Ponder BA, Easton DF. 1999. A systematic review of genetic polymorphisms and breast cancer risk. Cancer Epidemiol Biomarkers Prev 8(10):843–854. Dupont WD, Page DL. 1985. Risk factors for breast cancer in women with proliferative breast disease. N Engl J Med 312:145–151. Easton DF, Ford D, Bishop DT. 1995. Breast and ovarian cancer incidence in BRCA1-mutation carriers. Breast Cancer Linkage Consortium. Am J Hum Genet 56(1):265–271. Easton D, Ford D, Peto J. 1993. Inherited susceptibility to breast cancer. Cancer Surv 18:95–113. Ebeling P, Koivisto VA. 1994. Physiological importance of dehydroepiandrosterone. Lancet 343:1479–1481. Egan KM, Stampfer MJ, Hunter D, et al. 2002. Active and passive smoking in breast cancer: Prospective results from the Nurses’ Health Study. Epidemiology 13(2):138–145. Eley JW, Hill HA, Chen VW, et al. 1994. Racial differences in survival from breast cancer. Results of the National Cancer Institute Black/White Cancer Survival Study. JAMA 272(12):947–954. Feigelson HS, McKean-Cowdin R, Coetzee GA, Stram DO, Kolonel LN, Henderson BE. 2001. Building a multigenic model of breast cancer susceptibility: CYP17 and HSD17B1 are two important candidates. Cancer Res 61(2):785–789. Feigelson HS, Shames LS, Pike MC, Coetzee GA, Stanczyk FZ, Henderson BE. 1998. Cytochrome P450c17alpha gene (CYP17) polymorphism is associated with serum estrogen and progesterone concentrations. Cancer Res 58(4):585–587. Feuer E, Wun M. 1992. How much of the recent rise in breast cancer incidence can be explained by increases in mammography utilization: A dynamic population model approach. Am J Epidemiol 136:1423–1436. Fisher B, Costantino JP, Wickerham DL, et al. 1998. Tamoxifen for prevention of breast cancer—report of the National Surgical Adjuvant Breast and Bowel Project P-1. J Natl Cancer Inst 90:1371–1388. FitzGerald MG, Bean JM, Hegde SR, et al. 1997. Heterozygous ATM mutations do not contribute to early onset of breast cancer. Nat Genet 15(3):307–310. Ford D, Easton DF, Stratton M, et al. 1998. Genetic heterogeneity and penetrance analysis of the BRCA1 and BRCA2 genes in breast cancer families. The Breast Cancer Linkage Consortium. Am J Hum Genet 62(3):676–689. Foulkes WD, Rosenblatt J, Chappuis PO. 2001. The contribution of inherited factors to the clinicopathological features and behavior of breast cancer. J Mammary Gland Biol Neoplasia 6(4):453–465. Frazier AL, Ryan CT, Rockett H, Willett WC, Colditz GA. 2003. Adolescent diet and risk of breast cancer. Breast Cancer Res 5(3):R59–64. Freed GL, Clark SJ, Sorensen J, Lohr JA, Celalo R, Curtis P. 1995. National assessment of physicians’ breastfeeding knowledge, attitudes, training, and experience. JAMA 273:472–476. Freudenheim JL, Ambrosone CB, Moysich KB, et al. 1999. Alcohol dehydrogenase 3 genotype modification of the association of alcohol consumption with breast cancer risk. Cancer Causes Control 10:369–377. Freudenheim JL, Marshall JR, Graham S, et al. 1994. Exposure to breast milk in infancy and the risk of breast cancer. Epidemiology 5(3):324– 331.

1009

Friedenreich CM. 2001. Review of anthropometric factors and breast cancer risk. Eur J Cancer Prev 10(1):15–32. Frisch RE, Wyshak G, Albright NL, et al. 1985. Lower prevalence of breast cancer and cancers of the reproductive system among former college athletes compared to nonathletes. Br J Cancer 52: 885–891. Gambrell RD Jr., Maier RC, Sanders BI. 1983. Decreased incidence of breast cancer in postmenopausal estrogen-progestogen users. Obstet Gynecol 62:435–443. Ginsburg ES, Walsh BW, Gao XP, Gleason RE, Feltmate C, Barbieri RL. 1995. The effect of acute ethanol ingestion on estrogen levels in postmenopausal women using transdermal estradiol. J Soc Gynecol Invest 2:26–29. Giovannucci E, Stampfer MJ, Colditz GA, et al. 1998. Multivitamin use, folate, and colon cancer in women in the Nurses’ Health Study. Ann Intern Med 129:517–524. Giovannucci E, Stampfer MJ, Colditz GA, et al. 1993. A comparison of prospective and retrospective assessments of diet in the study of breast cancer. Am J Epidemiol 137:502–511. Going JJ. 2003. Stages on the way to breast cancer. J Pathol 199(1):1–3. Goldin BR, Aldercreutz H, Gorbach SL, et al. 1982. Estrogen excretion patterns and plasma levels in vegetarian and omnivorous women. N Engl J Med 307:1542–1547. Gortmaker SL, Dietz WH, Cheung LW. 1990. Inactivity, diet, and the fattening of America. J Am Diet Assoc 90:1247–1252. Gray GE, Henderson BE, Pike MC. 1980. Changing ratio of breast cancer incidence rates with age of black females compared with white females in the United States. J Natl Cancer Inst 64:461–463. Haiman CA, Hankinson SE, Spiegelman D, et al. 1999. The relationship between a polymorphism in CYP17 with plasma hormone levels and breast cancer. Cancer Res 59(5):1015–1020. Haiman CA, Hankinson SE, Spiegelman D, et al. 2000. A tetranucleotide repeat polymorphism in CYP19 and breast cancer risk. Int J Cancer 87(2): 204–210. Hancock S, Tucker M, Hoppe R. 1993. Breast cancer after treatment of Hodgkin’s disease. J Natl Cancer Inst 85:25–31. Hankinson SE, Willett WC, Colditz GA, et al. 1998. Circulating concentrations of insulin-like growth factor-I and risk of breast cancer. Lancet 351: 1393–1396. Hankinson SE, Willett WC, Michaud DS, et al. 1998. Plasma prolactin levels and subsequent risk of breast cancer in postmenopausal women. J Natl Cancer Inst 91:629–634. Hargreaves DF, Potten CS, Harding C, et al. 1999. Two-week dietary soy supplementation has an estrogenic effect on normal premenopausal breast. J Clin Endocrinol Metab 84:4017–4024. Harris JR, Lippman ME, Veronsei U, Willett WC. 1992. Breast cancer. N Engl J Med 327:319–328. Harvie M, Howell A, Vierkant RA, et al. 2005. Association of gain and loss of weight before and after menopause with risk of postmenopausal breast cancer in the Iowa women’s health study. Cancer Epidemiol Biomarkers Prev 14(3):656–661. Heck KE, Pamuk ER. 1997. Explaining the relation between education and postmenopausal breast cancer. Am J Epidemiol 145:366–372. Hein DW. 2002. Molecular genetics and function of NAT1 and NAT2: Role in aromatic amine metabolism and carcinogenesis. Mutat Res 506–507: 65–77. Helzlsouer KJ, Alberg AJ, Bush TL, Longcope C, Gordon GB, Comstock GW. 1994. A prospective study of endogenous hormones and breast cancer. Cancer Detect Prev 18:79–85. Henderson BE, Feigelson HS. 2000. Hormonal carcinogenesis. Carcinogenesis 21(3):427–433. Henderson BE, Pike MC, Casagrande JT. 1981. Breast cancer and the oestrogen window hypothesis. Lancet 2(8242):363–364. Hirayama T. 1984. Cancer mortality in nonsmoking women with smoking husbands based on a large-scale cohort study in Japan. Prev Med 13: 680–690. Hoffman-Goetz L, Apter D, Demark-Wahnefried W, Goran MI, McTiernan A, Reichman ME. 1998. Possible mechanisms mediating an association between physical activity and breast cancer. Cancer 83(3 Suppl): 621–628. Hoover R. 1996. Geographic, migrant, and time-trend patterns. In: Fortner J, Sharp P, eds. Accomplishments in Cancer Research. New York: Lippincott-Raven. Howe GR, Hirohata T, Hislop TG, et al. 1990. Dietary factors and risk of breast cancer: Combined analysis of 12 case-control studies. J Natl Cancer Inst 82:561–569. Hsieh C, Pavia M, Lambe M. 1994. Dual effect of parity on breast cancer risk. Eur J Cancer 30A:969–973.

1010

PART IV: CANCER BY TISSUE OF ORIGIN

Huang Z, Hankinson SE, Colditz GA, et al. 1997. Dual effects of weight and weight gain on breast cancer risk. JAMA 278:1407–1411. Hunter DJ, Spiegelman D, Adami HO, et al. 1996. Cohort studies of fat intake and the risk of breast cancer: A pooled analysis. N Engl J Med 334:356–361. Hunter DJ, Willett WC. 1993. Diet, body size, and breast cancer. Epidemiol Rev 15:110–132. Inskip HM, Kinlen LJ, Taylor AM, Woods CG, Arlett CF. 1999. Risk of breast cancer and other cancers in heterozygotes for ataxia-telangiectasia. Br J Cancer 79(7–8):1304–1307. International Agency for Research on Cancer. 2002. Weight Control and Physical Activity. IARC Handbooks on Cancer Prevention, Vol. 6. Lyon, France: IARC. Ishibe N, Hankinson SE, Colditz GA, et al. 1998. Cigarette smoking, cytochrome P450 1A1 polymorphisms, and breast cancer risk in the Nurses’ Health Study. Cancer Res 58(4):667–671. Jacobs TW, Byrne C, Colditz G, Connolly JL, Schnitt SJ. 1999. Radial scars in benign breast biopsy specimens and breast cancer risk: A case-control study. N Engl J Med 340:430–436. Janerich DT, Hoff MB. 1982. Evidence for a crossover in breast cancer risk factors. Am J Epidemiol 116:737–742. Janin N, Andrieu N, Ossian K, et al. 1999. Breast cancer risk in ataxia telangiectasia (AT) heterozygotes: Haplotype study in French AT families. Br J Cancer 80(7):1042–1045. Jin F, Shu X-O, Devesa S, Zheng W, Blot W, Gao Y-T. 1993. Incidence trends for cancers of the breast, ovary, and corpus uteri in urban Shanghai, 1972–1989. Cancer Causes Control 4:355–360. Juul A, Bang P, Hertel NT, et al. 1994. Serum insulin-like growth factor-I in 1030 healthy children, adolescents, and adults: Relation to age, sex, stage of puberty, testicular size, and body mass index. J Clin Endocrinol Metab 78:744–752. Kaaks R, Berrino F, Key T, et al. 2005. Serum sex steroids in premenopausal women and breast cancer risk within the European Prospective Investigation into Cancer and Nutrition (EPIC). J Natl Cancer Inst 97(10):755–765. Kelsey JL. 1979. A review of the epidemiology of human breast cancer. Epidemiol Rev 1:74–109. Kelsey JL, Gammon MD, John EM. 1993. Reproductive factors and breast cancer. Epidemiol Rev 15:36–47. Kerlikowske K, Grady D, Ernster V. 1995. Benefit of mammography screening in women ages 40–49 years: Current evidence from randomized controlled trials. Cancer 76(9):1679–1681. Key TJ, Sharp GB, Appleby PN, et al. 1999. Soya foods and breast cancer risk: A prospective study in Hiroshima and Nagasaki, Japan. Br J Cancer 81:1248–1256. Kolonel L. 1980. Cancer patterns of four ethnic groups in Hawaii. J Natl Cancer Inst 65:1127–1139. Krieger N. 1990. Social class and the black/white crossover in the age-specific incidence of breast cancer: A study linking census-derived data to population-based registry records. Am J Epidemiol 131:804–814. Kristensen VN, Andersen TI, Lindblom A, Erikstein B, Magnus P, BorresenDale AL. 1998. A rare CYP19 (aromatase) variant may increase the risk of breast cancer. Pharmacogenetics 8(1):43–48. Kristensen VN, Borresen-Dale AL. 2000. Molecular epidemiology of breast cancer: Genetic variation in steroid hormone metabolism. Mutat Res 462(2–3):323–333. Lachman HM, Papolos DF, Saito T, Yu YM, Szumlanski CL, Weinshilboum RM. 1996. Human catechol-O-methyltransferase pharmacogenetics: Description of a functional polymorphism and its potential application to neuropsychiatric disorders. Pharmacogenetics 6(3):243–250. Laden F, Spiegelman D, Neas LM, et al. 1997. Geographic variation in breast cancer incidence rates in a cohort of US women. J Natl Cancer Inst 89:1373–1378. Laden F, Collman G, Iwamoto K, et al. 2001. 1,1-Dichloro-2,2-bis(pchlorophenyl)ethylene and polychlorinated biphenyls and breast cancer: Combined analysis of five U.S. studies. J Natl Cancer Inst 93(10): 768–776. Lakhani SR. 2001. Molecular genetics of solid tumours: Translating research into clinical practice. What we could do now: breast cancer. Mol Pathol 54(5):281–284. Lakhani SR. 2003. In-situ lobular neoplasia: Time for an awakening. Lancet 361(9352):96. Lambe M, Hsieh CC, Trichopoulos D, Ekbom A, Pavia A, Adami HO. 1994. Transient increase in risk of breast cancer after giving birth. N Engl J Med 331:5–9. Lane-Claypon JE. 1962. A further report on cancer of the breast, with special reference to its associated antecedent conditions. London: Ministry of Health.

Lantz P, Remington P, Newcomb P. 1991. Mammography screening and increased incidence of breast cancer in Wisconsin. J Natl Cancer Inst 83:1540–1546. Lassarre C, Hardouin S, Daffos F, Forestier F, Frankenne F, Binoux M. 1991. Serum insulin-like growth factors and insulin-like growth factor binding proteins in the human fetus. Relationships with growth in normal subjects and in subjects with intrauterine growth retardation. Pediatr Res 29:219–225. Lavigne JA, Helzlsouer KJ, Huang HY, et al. 1997. An association between the allele coding for a low activity variant of catechol-O-methyltransferase and the risk for breast cancer. Cancer Res 57(24):5493–5497. Le Marchand L, Kolonel LN, Earle ME, Mi MP. 1988. Body size at different periods of life and breast cancer risk. Am J Epidemiol 128:137–152. Li CI, Malone KE, White E, Daling JR. 1997. Age when maximum height is reached as a risk factor for breast cancer among young U.S. women. Epidemiology 8:559–565. Li CI, Stanford JL, Daling JR. 2000. Anthropometric variables in relation to risk of breast cancer in middle-aged women. Int J Epidemiol 29(2):208–213. Liff J, Sung J, Chow W, Greenberg R, Flanders W. 1991. Does increased detection account for the rising incidence of breast cancer? Am J Public Health 81:462–465. Lilienfeld AM. 1956. The relationship of cancer of the female breast to artificial menopause and marital status. Cancer 9:927–934. Lipworth L, Bailey LR, Trichopoulos D. 2000. History of breast-feeding in relation to breast cancer risk: A review of the epidemiologic literature. J Natl Cancer Inst 92:302–312. Longnecker MP. 1994. Alcoholic beverage consumption in relation to risk of breast cancer: Meta-analysis and review. Cancer Causes Control 5:73–82. Longnecker MP, Bernstein L, Paganini-Hill A, Enger SM, Ross RK. 1996. Risk factors for in situ breast cancer. Cancer Epidemiol Biomarkers Prev 5(12):961–965. Longnecker MP, Newcomb PA, Mittendorf R, et al. 1995. Risk of breast cancer in relation to lifetime alcohol consumption. J Natl Cancer Inst 87: 923–929. MacMahon B, Trichopoulos D, Brown J. 1982. Age at menarche, urine estrogens and breast cancer risk. Int J Cancer 30:427–431. Malarkey WB, Kennedy M, Allred LE, Lo G. 1983. Physiological concentrations of prolactin can promote the growth of human breast tumor cells in culture. J Clin Endocrinol Metab 56:673–677. Mannermaa A, Peltoketo H, Winqvist R, et al. 1994. Human familial and sporadic breast cancer: Analysis of the coding regions of the 17 betahydroxysteroid dehydrogenase 2 gene (EDH17B2) using a single-strand conformation polymorphism assay. Hum Genet 93(3):319–324. Marcus PM, Newman B, Moorman PG, et al. 1999. Physical activity at age 12 and adult breast cancer risk (United States). Cancer Causes Control 10(4):293–302. Martin-Moreno JM, Willett WC, Gorgojo L, et al. 1994. Dietary fat, olive oil intake and breast cancer risk. Int J Cancer 58:774–780. McMichael-Phillips DF, Harding C, Morton M, et al. 1998. Effects of soyprotein supplementation on epithelial proliferation in the histologically normal human breast. Am J Clin Nutr 68:1431S–1435S. Melbye M, Wohlfahrt J, Olsen J. 1997. Induced abortion and the risk of breast cancer. N Engl J Med 336:81–85. Merzenich H, Boeing H, Wahrendorf J. 1993. Dietary fat and sports activity as determinants for age at menarche. Am J Epidemiol 138: 217–224. Michels KB, Trichopoulos D, Robins JM, et al. Birthweight as a risk factor for breast cancer. Lancet 348:1542–1546. Miller AB, Howe GR, Sherman GJ, et al. 1989. Mortality from breast cancer after irradiation during fluoroscopic examinations in patients being treated for tuberculosis. N Eng J Med 321:1285–1289. Miller BA, Feuer EJ, Hankey BF. 1991. The increasing incidence of breast cancer since 1982: Relevance of early detection. Cancer Causes Control 2:67–74. Miller BA, Feuer EJ, Hankey BF. 1993. Recent incidence trends for breast cancer in women and the relevance of early detection: An update. CA Cancer J Clin 43:27–41. Million Women Health Study Collaborators. 2003. Breast cancer and hormone-replacement therapy in the Million Women Study. Lancet 362:419–427. Mirick DK, Davis S, Thomas DB. 2002. Antiperspirant use and the risk of breast cancer. J Natl Cancer Inst 94(20):1578–1580. Missmer SA, Eliassen AH, Barbieri RL, Hankinson SE. 2004. Endogenous estrogen, androgen, and progesterone concentrations and breast cancer risk among postmenopausal women. J Natl Cancer Inst 96(24): 1856–1865.

Breast Cancer Mitrunen K, Hirvonen A. 2003. Molecular epidemiology of sporadic breast cancer. The role of polymorphic genes involved in oestrogen biosynthesis and metabolism. Mutat Res 544(1):9–41. Mitrunen K, Jourenkova N, Kataja V, et al. 2001. Polymorphic catechol-Omethyltransferase gene and breast cancer risk. Cancer Epidemiol Biomarkers Prev 10(6):635–640. Mittendorf R, Longnecker MP, Newcomb PA, et al. 1995. Strenuous physical activity in young adulthood and risk of breast cancer (United States). Cancer Causes Control 6:347–353. Moolgavkar SH, Day NE, Stevens RG. 1980. Two-stage model for carcinogenesis: Epidemiology of breast cancer in females. J Natl Cancer Inst 65:559–569. Moolgavkar S. 1990. Cancer models. Epidemiology 1:419–420. Moolgavkar S, Knudson A Jr. 1981. Mutation and cancer: A model for human carcinogenesis. J Natl Cancer Inst 66:1037–1052. Mosian J, Meyer F, Gingras S. 1991. Leisure, physical activity and age at menarche. Med Sci Sprts Exerc 23(10):1170–1175. Nachtigall LE, Nachtigall RH, Nachtigall RD, Beckman EM. 1979. Estrogen replacement therapy, II: A prospective study in the relationship to carcinoma and cardiovascular and metabolic problems. Obstet Gynecol 54:74–79. Nagata C, Kawakami N, Shimizu H. 1997. Trends in the incidence rate and risk factors for breast cancer in Japan. Breast Cancer Res Treat 44:75–82. Narod SA. 2002. Modifiers of risk of hereditary breast and ovarian cancer. Nat Rev Cancer 2(2):113–123. Newcomb PA, Lantz PM. 1993. Recent trends in breast cancer incidence, mortality, and mammography. Breast Cancer Res Treat 28(2):97–106. Newcomb PA, Titus-Ernstoff L, Egan KM, et al. 2002. Postmenopausal estrogen and progestin use in relation to breast cancer risk. Cancer Epidemiol Biol Prevention 11:593–600. Newman LA, Mason J, Cote Y, et al. 2002. African-American ethnicity, socioeconomic status, and breast cancer survival: a meta-analysis of 14 studies involving over 10,000 African-American and 40,000 White American patients with carcinoma of the breast. Cancer 94(11): 2844–2854. Oesterreich S, Fuqua SA. 1999. Tumor suppressor genes in breast cancer. Endocr Relat Cancer 6(3):405–419. Oza AM, Boyd NF. 1993. Mammographic parenchymal patterns: A marker of breast cancer risk. Epidemiol Rev 15(1):196–208. Page DL, Schuyler PA, Dupont WD, Jensen RA, Plummer WD Jr., Simpson JF. 2003. Atypical lobular hyperplasia as a unilateral predictor of breast cancer risk: A retrospective cohort study. Lancet 361(9352):125–129. Palmer JR, Rosenberg L, Clarke EA, et al. 1991. Breast cancer and cigarette smoking: A hypothesis. Am J Epidemiology 134(1):1–13. Parkin DM, Bray F, Ferlay J, Pisani P. 2001. Estimating the world cancer burden: Globocan 2000. Int J Cancer 94(2):153–156. Parkin DM, Muir CS, Whelan SL, Gao YT, Ferlay J, Powell J, eds. 1992. Cancer Incidence in Five Continents, Vol. VI. IARC Scientific Publications, No. 120. Lyon, France: IARC. Partridge RK, Hahnel R. 1979. Prolactin receptors in human breast carcinoma. Cancer 43:643–646. Pathak DR, Whittemore AS. 1992. Combined effects of body size, parity, and menstrual events on breast cancer incidence in seven countries. Am J Epidemiol 135:153–168. Pegram MD, Pauletti G, Slamon DJ. 1998. HER-2/neu as a predictive marker of response to breast cancer therapy. Breast Cancer Res Treat 52(1–3):65–77. Petrakis NL, Barnes S, King EB, et al. 1996. Stimulatory influence of soy protein on breast secretion inpre-and postmenopausal women. Cancer Epidemiol Biomarkers Prev 5:785–794. Pike MC, Krailo MD, Henderson BE, Casagrande JT, Hoel DG. 1983. “Hormonal” risk factors, “breast tissue age” and the age-incidence of breast cancer. Nature 303:767–770. Pike MC, Kolonel LN, Henderson BE, et al. 2002. Breast cancer in a multiethnic cohort in Hawaii and Los Angeles: Risk factor-adjusted incidence in Japanese equals and in Hawaiians exceeds that in whites. Cancer Epidemiol Biomarkers Prev 11(9):795–800. Pike MC, Spicer DV, Dahmoush L, Press MF. 1993. Estrogens, progestogens, normal breast cell proliferation, and breast cancer risk. Epidemiol Rev 15:48–65. Pollak M, Costantino J, Polychronakos C, et al. 1990. Effect of tamoxifen on serum insulinlike growth factor I levels in stage I breast cancer patients. J Natl Cancer Inst 82:1693–1697. Potischman N, Troisi R. 1999. In-utero and early life exposures in relation to risk of breast cancer. Cancer Causes Control 10(6):561–573. Prentice RL, Chlebowski RT, Patterson R, et al. 2006. Low-fat dietary pattern and risk of invasive breast cancer: the Women’s Health Initiative Randomized Controlled Dietary Modification Trial. JAMA 295(6):629–642.

1011

Reichman ME, Judd JT, Longcope C, et al. 1993. Effects of alcohol consumption on plasma and urinary hormone concentrations in premenopausal women. J Natl Cancer Inst 85:722–727. Rich-Edwards JW, Goldman MB, Willett WC, et al. 1994. Adolescent body mass index and ovulatory infertility. Am J Obstet Gynecol 171: 171–177. Ries LAG, Eisner MP, Kosary CL, et al. 2005. Seer Cancer Statistics Review, 1975–2002. Bethesda, MD: National Cancer Institute. Rimm E, Willett W, Manson J, Speizer F, Hennekens C, Stampfer M. 1996. Folate and vitamin B6 intake and risk of myocardial infarction among US women abstract. Am J Epidemiol 143(suppl):S36. Robbins A, Brescianini S, Kelsey J. 1997. Regional differences in known risk factors and the higher incidence of breast cancer in San Francisco. J Natl Cancer Inst 89:960–965. Rockhill B, Colditz GA, Rosner B. 2000. Bias in breast cancer analyses due to error in age at menopause. Am J Epidemiol 151:404–408. Roetzheim RG, Gonzalez EC, Ferrante JM, Pal N, Van Durme DJ, Krischer JP. 2000. Effects of health insurance and race on breast carcinoma treatments and outcomes. Cancer 89(11):2202–2213. Rohan TE, Howe GR, Friedenreich CM, Jain M, Miller AB. 1993. Dietary fiber, vitamins A, C, and E, and risk of breast cancer: A cohort study. Cancer Causes Control 4:29–37. Romieu I, Hernandez-Avila M, Lazcano E, Lopez L, Romero-Jaime R. 1996. Breast cancer and lactation history in Mexican women. Am J Epidemiol 132:17–26. Rose G. 1981. Strategy of prevention: Lessons from cardiovascular disease. Br Med J—Clin Res 282:1847–1851. Rosen P, ed. 2001. Rosen’s Breast Pathology. 2nd ed. Philadelphia: Lippincott Williams & Wilkins. Rosner B, Colditz G. 1996. Extended mathematical model of breast cancer incidence in the Nurses’ Health Study. J Natl Cancer Inst 88:359–364. Rosner B, Colditz GA, Willett WC. 1994. Reproductive risk factors in a prospective study of breast cancer: The Nurses’ Health Study. Am J Epidemiol 139:819–835. Ross RK, Paganini-Hill A, Wan PC, Pike MC. 2000. Effect of hormone replacement therapy on breast cancer: Estrogen versus estrogenplus progestin. JNCI 92:328–332. Ruan W, Newman C, Kleinberg D. 1992. Intact and amino-terminally shortened forms of insulin-like growth factor I induce mammary gland differentiation and development. Proc Natl Acad Sci USA 89: 10872–10876. Russo J, Gusterson BA, Rogers AE, Russo IH, Wellings SR, van Zwieten MJ. 1990. Biology of disease: Comparison study of human and rat mammary tumorigenesis. Lab Invest 62:244–278. Russo J, Russo IH. 1980. Susceptibility of the mammary gland to carcinogenesis. II. Pregnancy interruption as a risk factor in tumor incidence. Am J Pathol 100(2):497–512. Russo J, Tay LK, Russo IK. 1982. Differentiation of the mammary gland and susceptibility to carcinogenesis. Breast Cancer Res Treat 2:5–73. Schairer C, Byrne C, Keyl PM, Brinton LA, Sturgeon SR, Hoover RN. 1994. Menopausal estrogen and estrogen-progestin replacement therapy and risk of breast cancer (United States). Cancer Causes Control 5:491–500. Schairer C, Lubin J, Troisi R, Sturgeon S, Brinton L, Hoover R. 2000. Menopausal estrogen and estrogen-progestin replacement therapy and breast cancer risk. JAMA 283:485–491. Schernhammer ES, Holly JM, Pollak MN, Hankinson SE. 2005. Circulating levels of insulin-like growth factors, their binding proteins, and breast cancer risk. Cancer Epidemiol Biomarkers Prev 14(3):699–704. Schnitt SJ. 2001. Traditional and newer pathologic factors. J Natl Cancer Inst Monogr (30):22–26. Schnitt SJ, Morrow M. 1999. Lobular carcinoma in situ: Current concepts and controversies. Semin Diagn Pathol 16(3):209–223. Schnitt SJ. 2000. Pathology of breast cancer. In: Hayes DF, ed. Atlas of Breast Cancer. London: Harcourt Publishers Limited. Schnitt SJ, Connolly JL. 2004. Pathology of benign breast disorders. In: Harris JR, Lippman ME, Morrow M, Osborne CK, eds. Diseases of the Breast. Philadelphia: Lippincott Williams & Wilkins. Seow A, Duffy S, McGee M, Lee J, Lee H. 1996. Breast cancer in Singapore: Trends in incidence 1968–1992. Int J Epidemiol 25:40–45. Seymour-Munn K, Adams J. 1983. Estrogenic effects of 5-androstene-3beta, 17beta-diol at physiological concentrations and its possible implication in the etiology of breast cancer. Endocrinology 112:486–491. Shimizu H, Ross RK, Bernstein L, Yatani R, Henderson BE, Mack TM. 1991. Cancers of the prostate and breast among Japanese and white immigrants in Los Angeles County. Br J Cancer 63:963–966. Siegelmann-Danieli N, Buetow KH. 1999. Constitutional genetic variation at the human aromatase gene (Cyp19) and breast cancer risk. Br J Cancer 79(3–4):456–463.

1012

PART IV: CANCER BY TISSUE OF ORIGIN

Smith RA, Saslow D, Sawyer KA, et al. 2003. American Cancer Society guidelines for breast cancer screening: Update 2003. CA Cancer J Clin 53(3):141–169. Smith-Warner SA, Spiegelman D, Yaun SS, et al. 2001. Intake of fruits and vegetables and risk of breast cancer: A pooled analysis of cohort studies. JAMA 285:769–776. Smith-Warner SA, Spiegelman D, Yaun S-S, et al. Alcohol and breast cancer in women: A pooled analysis of cohort studies. JAMA 279:535–540 Spicer DV, Ursin G, Parisky YR, et al. 1994. Changes in mammographic densities induced by a hormonal contraceptive designed to reduce breast cancer risk. J Natl Cancer Inst 86(6):431–436. Stanford J, Herrinton L, Schwartz S, Weiss N. 1995. Breast cancer incidence in Asian migrants to the United States and their descendants. Epidemiology 6:1819–1827. Stanford JL, Greenberg RS. 1989. Breast cancer incidence in young women by estrogen receptor status and race. Am J Public Health 79(1):71–73. Stoll BA. 1998. Teenage obesity in relation to breast cancer risk. Int J Obes Relat Metab Disord 22(11):1035–1040. Struewing JP, Hartge P, Wacholder S, et al. 1997. The risk of cancer associated with specific mutations of BRCA1 and BRCA2 among Ashkenazi Jews. N Engl J Med 336(20):1401–1408. Sturgeon S, Schairer C, Gail M, McAdams M, Brinton L, Hoover R. 1995. Geographic variation in mortality from breast cancer among white women in the United States. J Natl Cancer Inst 87:1846–1853. Swift M, Morrell D, Massey RB, Chase CL. 1991. Incidence of cancer in 161 families affected by ataxia-telangiectasia. N Engl J Med 325(26): 1831–1836. Tamimi RM, Hankinson SE, Campos D, et al. 2005. Plasma carotenoids, retinol, and tocopherols and risk of breast cancer. Am J Epidemiol 161(2):153–160. The Endogenous Hormones and Breast Cancer Collaborative Group. 2002. Endogenous sex hormones and breast cancer in postmenopausal women: Reanalysis of nine prospective studies. J Natl Cancer Inst 94:606–616. Thomas HV, Key TJ, Allen DS, et al. 1997. A prospective study of endogenous serum hormone concentrations and breast cancer risk in premenopausal women on the island of Guernsey. Br J Cancer 75:1075–1079. Thompson PA, Shields PG, Freudenheim JL, et al. 1998. Genetic polymorphisms in catechol-O-methyltransferase, menopausal status, and breast cancer risk. Cancer Res 58(10):2107–2110. Thorlacius S, Sigurdsson S, Bjarnadottir H, et al. 1997. Study of a single BRCA2 mutation with high carrier frequency in a small population. Am J Hum Genet 60(5):1079–1084. Tokunaga M, Land C, Tokuoka S, Nishimori I, Soda M, Akiba S. 1994. Incidence of female breast cancer among atomic bomb survivors 1950–1985. Radiat Res 138:209–223. Tominaga S. 1985. Cancer incidence in Japanese in Japan, Hawaii, and western United States. NCI Monographs 69:83–92. Tominaga S, Aoki K, Fujimoto I, Kurihara M. 1994. Cancer Mortality and Morbidity Statistics: Japan and the World—1994. Tokyo: Japan Scientific Societies Press. Toniolo P, Van Kappel AL, Akhmedkhanov A, et al. 2001. Serum carotenoids and breast cancer. Am J Epidemiol 153(12):1142–1147. Toniolo P, Bruning PF, Akhmedkhanov A, et al. 2000. Serum insulin-like growth factor-I and breast cancer. Int J Cancer 88(5):828–832. Trentham-Dietz A, Newcomb PA, Storer BE, Remington PL. 2000. Risk factors for carcinoma in situ of the breast. Cancer Epidemiol Biomarkers Prev 9(7):697–703. Trichopoulos D. 1990. Is breast cancer initiated in utero? Epidemiology 1:95–96. Trichopoulos D, Hsieh C, MacMahon B, Lin T, Lowe C, Mirra A. 1983. Age at any birth and breast cancer risk. Int J Cancer 31:701–704. Trichopoulos D, MacMahon B, Cole P. 1972. Menopause and breast cancer risk. J Natl Cancer Inst 48:605–613. Tworoger SS, Eliassen AH, Rosner B, Sluss P, Hankinson SE. 2004. Plasma prolactin concentrations and risk of postmenopausal breast cancer. Cancer Res 64(18):6814–6819. U. S. Bureau of the Census. 1983. Fertility of American Women: June 1983, Current Population Report. Series P-20, Population Characteristics, No. 395. Washington, DC: U. S. Dept. of Commerce, Bureau of the Census.

Unger MA, Weber BL. 2000. Recent advances in breast cancer biology. Curr Opin Oncol 12(6):521–525. Ursin G, Longnecker MP, Halies RW, Greenland S. 1995. A meta-analysis of body mass index and risk of premenopausal breast cancer epidemiology. Epidemiology 6:137–141. van den Brandt PA, Spiegelman D, Yaun SS, et al. 2000. Pooled analysis of prospective cohort studies on height, weight, and breast cancer risk. Am J Epidemiol 152:514–527. Verhoeven DT H, Assen N, Goldbohm RA, et al. 1997. Vitamins C and E, retinol, beta-carotene and dietary fiber in relation to breast cancer risk: A prospective cohort study. Br J Cancer 75:149–155. Wang DY, De Stavola BL, Bulbrook RD, et al. 1992. Relationship of blood prolactin levels and the risk of subsequent breast cancer. Int J Epidemiol 21:214–221. Waterhouse JAH, Muir GS, Shanmugaratnam K, Powell J, eds. 1982. Cancer Incidence in Five Continents. Vol. IV. IARC Scientific Publication No. 42. Lyon, France: IARC. Weiss HA, Potischman NA, Brinton LA, et al. 1997. Prenatal and perinatal risk factors for breast cancer in young women. Epidemiol 8:181–187. Wells AJ. 1991. Breast cancer, cigarette smoking, and passive smoking. Am J Epidemiol 133(2):208–210. White E, Lee C, Kristal A. 1990. Evaluation of the increase in breast cancer incidence in relation to mammography use. J Natl Cancer Inst 82:1546–1552. Willett WC. 1997. Specific fatty acids and risks of breast and prostate cancer: Dietary intake. Am J Clin Nutr 66(suppl):1557s–1563s. Willett WC. 2001. Diet and breast cancer. J Intern Med 249:395–411. Willett WC, Hunter DJ, Stampfer MJ, et al. 1992. Dietary fat and fiber in relation to risk of breast cancer: An 8-year follow-up. JAMA 268:2037–2044. Wingo PA, Cardinez CJ, Landis SH, et al. 2003. Long-term trends in cancer mortality in the United States, 1930–1998. Cancer 97(12 suppl): 3133–3275. World Cancer Research Fund, and American Institute for Cancer Research. 1997. Food, Nutrition and the Prevention of Cancer: A Global Perspective. Washington, DC: American Institute for Cancer Research. Wysowski DK, Comstock GW, Helsing KJ, Lau HL. 1987. Sex hormone levels in serum in relation to the development of breast cancer. Am J Epidemiol 125:791–799. Xu X, Duncan AM, Wangen KE, Kurzer MS. 2000. Soy consumption alters endogenous estrogen metabolism in postmenopausal women. Cancer Epidemiol Biomarkers Prev 9(8):781–786. Yager JD, Liehr JG. 1996. Molecular mechanisms of estrogen carcinogenesis. Ann Rev Pharmacol Toxicol 36:203–232. Yang XF, Beamer W, Huynh HT, Pollak M. 1996. Reduced growth of human breast cancer xenografts in hosts homozygous for the lit mutation. Cancer Res 56:1509–1511. Yu H, Harris R, Gao Y, Gao R, Wynder E. 1991. Comparative epidemiology of cancers of the colon, rectum, prostate, and breast in Shanghai, China versus the United States. Int J Epidemiol 20:76–81. Zapf J, Froesch ER. 1986. Insulin-like growth factors/somatomedins: Structure, secretion, biological actions and physiological role. Hormone Res 24:121. Zeleniuch-Jacquotte A, Bruning PF, Bonfrer JMG, Koenig KL, Shore RE, Kim MY. 1997. Relation of serum levels of testosterone and dehydroepiandrosterone sulfate to risk of breast cancer in postmenopausal women. Am J Epidemiol 145:1030–1038. Zhang S, Hunter DJ, Hankinson SE, et al. 1999. A prospective study of folate intake and the risk of breast cancer. JAMA 281:1632–1637. Zhang SM, Willett WC, Selhub J, et al. 2003. Plasma folate, vitamin B6, vitamin B12, and homocysteine and risk of breast cancer. J Natl Cancer Inst 95:373–380. Ziegler RG, Hoover RN, Nomura AMY, et al. 1996. Relative weight, weight change, height, and breast cancer risk in Asian-American women. J Natl Cancer Inst 88:650–660. Ziegler RG, Hoover RN, Pike MC, et al. 1993. Migration patterns and breast cancer risk in Asian-American women. J Natl Cancer Inst 85:1819–1827.

52

Ovarian Cancer SUSAN E. HANKINSON AND KIM N. DANFORTH

O

varian cancer is the fifth most common cancer among women in the United States, accounting for 4% of cancer diagnoses, or about 25,400 new cases each year. It is also the fifth-leading cause of cancer-related mortality (ACS, 2003). Ovarian cancers can arise from the surface epithelium, germ cells, or stroma of the ovary. Cancers of the surface epithelium make up about 90% of invasive ovarian cancers, while germ cell and stromal tumors each make up 5% or less of the total (Crum, 1999). Therefore, most epidemiologic studies have focused on epithelial ovarian cancers, and our review will also concentrate on this subtype.

CLASSIFICATION Ovarian tumors are broadly classified either as being benign, of low malignant potential (or borderline), or malignant. Borderline tumors demonstrate some degree of nuclear atypia and are capable of metastasizing but do not invade the ovarian stroma (Chapman, 2001; Feeley and Wells, 2001). Although borderline tumors of all histologic types have been found, the majority are either serous or mucinous (Chapman, 2001). An important and unresolved question is whether borderline tumors are precursors to invasive ovarian cancer or entirely separate entities. Accumulating data suggest that serous borderline tumors rarely progress to become invasive. Several molecular marker studies found that certain genetic abnormalities, such as Ki-ras mutations and microsatellite instability, occurred much more often in serous borderline tumors than in serous invasive tumors, a finding that would not be expected if the invasive cancer arose from the borderline lesion (Feeley and Wells, 2001; Werness and Eltabbakh, 2001). In contrast, there is evidence that some borderline mucinous tumors may progress to malignancy. For example, foci of both benign and transitionalappearing epithelium are frequently observed within and bordering invasive mucinous tumors (Feeley and Wells, 2001). Thus, in contrast to a number of other cancers such as those of the colon and cervix, and with the possible exception of the mucinous subtype, no established precursor lesion exists for ovarian cancer. Ovarian epithelial tumors can be classified into a number of histologic types, the most common being serous, mucinous, endometrioid, and clear-cell. Within a tumor, mixtures of these cell types often occur. Although reports vary, in general, 45–50% of invasive epithelial tumors are found to be serous, 12% mucinous, 24% endometrioid, 8% clear cell, and 6–11% undifferentiated or other. (Crum, 1999) Distinctions between invasive and borderline tumors and between specific histologic types can be difficult. Accurate classification of invasive tumors is difficult in part because so many women are not diagnosed until the tumor is already at an advanced stage, i.e., Stage III or IV.

DEMOGRAPHIC PATTERNS Incidence and Mortality in the United States Five percent of cancer deaths in women are due to ovarian cancer, and approximately 14,300 women die from this disease each year (ACS, Funded by: NCI/NIH Grant CA 87969 (SH), PHS Training Grant CA 09001 (KD).

2003). On average, a woman has a 1.72% chance of being diagnosed with ovarian cancer in her lifetime and a 1.02% chance of dying from an ovarian tumor. From 1996 to 2000, the age-adjusted incidence rate for invasive ovarian cancer was 16.9 per 100,000 women, and the mortality rate was 8.8 per 100,000 women (Ries et al., 2003).

Time Trends From 1989 to 2000, ovarian cancer incidence rates decreased by 0.7% each year. During a similar time period, mortality rates also declined by 0.8% annually (Ries et al., 2003). However, temporal patterns vary by histologic type and race/ethnicity. Based on data from the Surveillance, Epidemiology, and End Results (SEER) Program of the National Cancer Institute, incidence rates for serous tumors significantly increased among white women from 1978–1998, while rates for mucinous, papillary not otherwise specified (NOS), and other epithelial tumors significantly decreased. Among black women, the only statistically significant trend was a decrease in papillary tumors; as with white women, the magnitude of the decrease was approximately 70% over the entire time period. Although differences in temporal trends by histologic type may reflect different etiologies, they could also reflect changes in tumor classification (Mink, Sherman, and Devesa, 2002). For example, tumors are no longer classified as “papillary NOS,” which is likely to explain at least part of the observed increase in serous tumors.

Survival When ovarian cancer is diagnosed at the localized stage, the five-year relative survival rate is 95%. Unfortunately, only 29% of cases are diagnosed this early in the disease process, and the overall survival rate for ovarian cancer is 53%. Nearly two-thirds of cases are diagnosed at an advanced stage or are unstaged at diagnosis; among these groups the five-year relative survival rates are about 30% (Ries et al., 2003).

Age Incidence and mortality rates generally increase exponentially with age, with a sharp increase beginning in the forties for incidence and the fifties for mortality. Although rates decrease among the oldest individuals (Fig. 52–1), the age-adjusted incidence and mortality rates for women under 65 years are only 11.2 and 3.7 per 100,000 women, respectively, compared with 56.3 and 44.1 for women age 65 and older. Survival is almost twice as high among younger women: the five-year relative survival rate is 65.8% for those under 65 years of age compared with 32.9% for those age 65 and older (Ries et al., 2003) (Fig. 52–2). The majority of ovarian tumors are epithelial, but non-epithelial tumors have a larger impact on younger women. Incidence rates for germ cell tumors peak among 15 to 19 year olds, with a rate of 1.22 per 100,000. Risk of sex-cord stromal tumors is highest among women in their fifties, with an approximate incidence rate of 37 per 100,000 (IARC, 2002a).

1013

1014

PART IV: CANCER BY TISSUE OF ORIGIN

70

120

60

100

< 65 yrs

>= 65 yrs

50 Percent Survival

Rate per 100,000

80 40

30

60

40 20 Incidence: White Incidence: Black Mortality: White Mortality: Black

10

0 10–14

20–24

30–34

40–44 50–54 Age Group

60–64

70–74

20

0 All Stages

80–84 85+

Localized

Regional

Distant

Unstaged

50% 70%

3% 9%

Stage at Diagnosis

Figure 52–1. Age-specific incidence and mortality rates for invasive ovarian cancer among whites and blacks: SEER data 1996–2000. (Source: Ries et al., 2003.)

Distribution of Stage at Diagnosis Age < 65 yrs 41% 15% Age ≥ 65 yrs

7% 5%

Figure 52–2. Five-year relative survival rates for invasive ovarian cancer by stage at diagnosis and age: SEER data, 1992–1999. (Source: Ries et al., 2003.)

Race and Ethnicity

Socioeconomic Status

Although both incidence and mortality are higher among white women than among black women, five-year relative survival rates are nearly identical, as is the distribution of stage at diagnosis (Ries et al., 2003). The age-adjusted incidence rate for ovarian cancer in non-Hispanic whites was 18.3 per 100,000 women, 14.0 for Hispanics, 12.2 for Asian/Pacific Islanders, 11.9 for blacks, and 10.7 for American Indians/Alaska Natives. Mortality rates were also highest for nonHispanic whites at 9.3 per 100,000 women. The mortality rate for black women was 7.4 per 100,000 women, 6.1 for Hispanics, 4.5 for American Indians/Alaska Natives, and 4.7 for Asian/Pacific Islanders (Ries et al., 2003). Rates also show some variation by race/ethnicity for specific subtypes of ovarian tumors (Fig. 52–3) (Mink, Sherman, and Devesa, 2002).

Data are conflicting regarding the association between socioeconomic status (SES) and ovarian cancer. Some studies have suggested that individuals of higher SES experience a greater burden of disease, but other studies have been inconclusive or failed to find an association (IARC, 1997).

International Patterns In 2000, an estimated 192,000 women were newly diagnosed with ovarian cancer, and 114,000 women died from the disease worldwide (Parkin, Bray, and Devesa 2001). Cancer incidence varies by geographic location: rates are highest in Denmark, Sweden, England, and the United States; intermediate in France, Spain, and Italy; and lowest in Japan, China, and Egypt (Fig. 52–4) (IARC, 2002a).

5 White

4.5

Black

Rate per 100,000 women

4

3.5

3

2.5

2

1.5 1

0.5

0 Mucinous

Serous

Endometrioid

Clear cell

Other epithelial Gonadal stromal

Germ cell

1.23 (1.13, 1.33)

0.90 (0.73, 1.10)

Histologic type Rate Ratios Comparing Whites and Blacks RR 95% Cl

1.2 (1.14, 1.47)

1.81 (1.67, 1.97)

2.10 (1.80, 2.45)

Age Adjusted to the 1970 United States Population

2.56 (1.92, 3.43)

0.53 (0.42, 0.67)

Figure 52–3. Age-adjusted incidence rates for invasive ovarian cancer by race: SEER data (9 regions), 1978–1998. (Source: Mink, Sherman, and Devesa, 2002.)

Ovarian Cancer 80

Rate per 100,000 women

70 60

Denmark U.S. France Japan

50 40 30 20 10 0 10–14

20–24

30–34

40–44 50–54 Age Group

60–64

70–74

80–84 85+

Figure 52–4. Age-adjusted ovarian cancer incidence rates for four countries: Denmark, United States (SEER regions), France (Isere), and Japan (Hiroshima). Rates for Denmark, United States, and France are for 1993–1997; rates for Japan are for 1991–1995. (IARC, 2002a)

Migration A study of Japanese migrants to the United States found that they and their descendents died from ovarian cancer at rates intermediate between those seen in the general population of the two countries (Haenszel and Kurihara, 1968). The study suggested that modifiable factors might alter mortality rates from ovarian cancer within a relatively short period of time.

ENVIRONMENTAL FACTORS Oral Contraceptive Use Use of oral contraceptives (OCs) represents one of the few modifiable risk factors that is established for ovarian cancer. In a 1992 metaanalysis of 20 studies, ever use of OCs was associated with a significant decrease in the risk of ovarian cancer (RR = 0.64, 95% CI: 0.57–0.73). The protection persisted for at least a decade and increased with duration of use (Hankinson et al., 1992). Whether the decrease in risk persists throughout life is as yet unknown. Early evidence suggests a decrease in risk for at least 15 years (Siskind et al., 2000). Research has shifted to examining whether the protection against ovarian cancer varies by different formulations of OCs, patterns of use (e.g., age at first use), or the presence of specific gene mutations. There is some suggestion that different doses of progestin or estrogen may result in different degrees of protection against ovarian cancer (Schildkraut et al., 2002), but again, further research is needed. The pattern of OC use has not been shown to have a large effect. In a 2002 collaborative re-analysis of six case-control studies, recent use, age at first use, and time since first use were not associated with ovarian cancer (Bosetti et al., 2002). Conflicting results have been reported regarding whether the protection offered by OC use extends to women who are at increased risk due to inherited mutations. While some studies have found that OC use is protective among BRCA1 and BRCA2 carriers, (Ness et al., 2001; Ness, et al. 2000; Narod, Sun, and Risch, 2001) at least one study failed to detect any protection. (Modan et al., 2001) Attempts to understand how use of OCs affects ovarian cancer risk have led to the development of broader etiologic hypotheses, such as the incessant ovulation theory and the gonadotropin hypothesis (Hankinson et al., 1992).

Postmenopausal Hormone Use Use of postmenopausal hormones (PMH) appears to increase the risk of ovarian cancer, although the association remains to be firmly established. The strongest evidence comes from two recent prospective

1015

cohort studies and a randomized trial. In the cohort studies, ever users of PMH were found to have an increased risk of developing (RR = 1.6, 95% CI: 1.2–2.0) (Lacey et al., 2002) and dying from ovarian cancer (RR = 1.23, 95% CI: 1.06–1.43) (Rodriguez et al., 2001); in both studies, the risk increased with increasing durations of use. In the trial, which included only 32 cases of ovarian cancer, a nonsignificant increase in the risk of ovarian cancer was associated with continuous combined estrogen plus progestin (RR = 1.58, 95% CI: 0.77–3.24) (Anderson et al., 2003). Other recent studies also support an association, (Riman et al., 2002; Tavani et al., 2000) although not all results were statistically significant. (Chiaffarino et al., 2001) However, the literature as a whole is not consistent. From 1998 to 2000, three reviews and a collaborative re-analysis reached different conclusions despite considering many of the same studies. In 1998, a meta-analysis of nine studies estimated that ever-use of PMH was associated with 1.15 times the odds of ovarian cancer of never-users (95% CI: 1.05–1.27). (Garg et al., 1998) A pooled analysis of four hospital-based, European case-control studies also found a significant increase in risk (OR = 1.28) (Negri et al., 1999; Bosetti et al., 2001). In contrast, a second meta-analysis of 15 case-control studies found no statistically significant association between PMH use and ovarian cancer, although the summary estimate was similar to that produced by the 1998 meta-analysis (OR = 1.1, 95% CI: 0.9–1.3). (Coughlin et al., 2000) A 1999 monograph by the International Agency for Research on Cancer (IARC) also concluded that there was no evidence of an association between PMH use and epithelial ovarian cancer. (IARC, 1999) Conflicting results may be a result of the dependence of the PMHovarian cancer relationship on tumor histology or the specific type of PMH therapy. For example, a relatively early study suggested that PMH use may specifically increase risk of endometrioid tumors (Weiss et al., 1982). Other studies support an increase for endometrioid tumors but suggest there may be an increase for some other histologic types as well (Risch, 1996; Purdie et al., 1999; Riman et al., 2002). Results from a recent cohort study suggested that estrogen only, and not estrogen-progestin therapy, increases risk (Lacey et al., 2002). One case-control study found an increased risk of ovarian cancer with sequential (OR = 1.98, 95% CI: 1.40–2.78) but not with continuous estrogen-progestin therapy (OR = 1.11, 95% CI: 0.71–1.74) (Riman et al., 2002). To date, most studies have had limited ability to examine such specific aspects of PMH use. Considered alone, theories regarding the pathogenesis of ovarian cancer might have predicted that PMH use would lower risk of ovarian cancer. For example, PMH use decreases gonadotropin levels, which have been proposed to increase ovarian cancer risk (Chiaffarino et al., 2001; Coughlin et al., 2000). If PMH use is associated with an increased risk of ovarian cancer, as currently seems likely, the mechanism(s) will need to be identified in future studies.

Analgesic Use Interest in the relationship between the use of analgesics and ovarian cancer emerged from the reported association between analgesics and other cancers. Research has focused on three main exposures: (1) nonsteroidal anti-inflammatories (NSAIDs) in general, (2) aspirin, and (3) acetaminophen. Two early studies reported an approximate halving of risk associated with use of analgesics (mostly salicylates) (Tzonou et al., 1993) and acetaminophen (Cramer et al., 1998). Subsequent studies generally have not supported an association, although some point estimates suggest a decrease in risk (Table 52–1). Many studies to date have been limited by small numbers of exposed cases. The potential biologic mechanism remains to be elucidated but could include anti-inflammatory effects (Ness and Cottreau, 1999) or an influence on hormone levels (Cramer et al., 1998).

Dietary Intake Coffee/Caffeine In 1991, IARC published a monograph that concluded there was inadequate evidence to classify coffee as a carcinogen for ovarian cancer.

1016

PART IV: CANCER BY TISSUE OF ORIGIN

Table 52–1. Analgesics and ovarian cancer Study Name

nsaids

Dose*: RR (95% CI)

Cramer et al. (1998)

Once a week for ≥6 mos. Analgesic: 0.9 (0.5–1.5) Ibuprofen: 1.0 (0.6–1.6)

Rosenberg et al. (2000)

Days per week for ≥6 mos. ≥1: 0.7 (0.5–1.0) ≥4: 0.7 (0.5–1.0)

Fairfield et al. (2002)

Days per month 1–4: 0.6 (0.3–1.2) 5–20: 0.6 (0.3–1.2) ≥20: 0.6 (0.3–1.2)

Meier et al. (2002)

No. of prescriptions 1–9: 1.1 (0.9–1.4) 10–19: 0.8 (0.5–1.2) 20–29: 0.7 (0.4–1.4) ≥30: 1.1 (0.7–1.8)

Duration in years*: RR (95% CI)

RR by dose category ≥5†: 0.7 (0.4–1.2) ≥5†: 0.5 (0.3–0.9)

aspirin Tzonou et al. (1993)

Qualitative use Infrequent: 0.8 (0.4–1.4) Frequent: 0.5 (0.3–1.0)

Cramer et al. (1998)

Once a week for ≥6 mos. 0.8 (0.5–1.1)

Rosenberg et al. (2000)

Days per week for ≥6 mos ≥4: 0.8 (0.5–1.2)

RR for dose category 0.5 (0.2–1.0) ≥5†:

Tavani et al. (2000)

<5: ≥5:

0.8 (0.4–1.6) 1.2 (0.5–2.9)

Akhmedkhanov et al. (2001)

<5: ≥5:

0.5 (0.2–1.8) 0.7 (0.2–2.0)

Moysich et al. (2001)

Tablets per week 1–6: 1.1 (0.7–1.7) ≥7: 0.9 (0.5–1.4)

0.5–10: 1.3 (0.8–2.3) ≥11: 0.9 (0.6–1.3)

Fairfield et al. (2002)

Tablets per week 1–2: 1.1 (0.8–1.4) 3–5: 1.0 (0.7–1.5) 6–14: 0.9 (0.6–1.4) ≥15: 1.0 (0.6–1.5)

<5: 5–9: 10–19: ≥20:

0.9 (0.7–1.2) 1.3 (0.9–1.8) 0.7 (0.4–1.4) 1.0 (0.7–1.4)

1.0 (0.8–1.1) 1.0 (0.8–1.3)

acetaminophen Cramer et al. (1998)

Once a week for ≥6 mos. 0.5 (0.3–0.9)

Rodriguez et al. (1998)**

Times per month 1–14: 1.0 (0.8–1.2) 15–29: 1.3 (0.8–2.2) ≥30: 0.6 (0.3–1.1)

<10: ≥10:

Rosenberg et al. (2000)

Days per week for ≥6 mos. ≥1: 1.0 (0.6–1.5) ≥4: 0.9 (0.5–1.6)

RR by dose category ≥5†: 1.1 (0.6–1.9) ≥5†: 1.2 (0.5–2.6)

Moysich et al. (2001)

Per week 1–6: ≥7:

0.5–10: 0.6 (0.3–1.1) ≥11: 0.5 (0.3–1.0)

0.6 (0.4–1.0) 0.3 (0.1–1.1)

Friis et al. (2002)

No. of prescriptions, SIR 1: 0.5 (0.1–1.7) 2–4: 1.5 (0.5–3.5) 5–9: 0.9 (0.1–3.3) ≥10: 0.3 (0.0–1.6)

Meier et al. (2002)

No. of prescriptions 1–9: 1.2 (1.0–1.6) 10–19: 1.4 (0.9–2.2) 20–29: 1.7 (1.0–2.9) ≥30: 1.0 (0.6–1.5)

Fairfield et al. (2002)

Days per month 1–4: 0.9 (0.6–1.5) ≥5: 0.8 (0.5–1.4)

*Comparison to non-users; **Outcome was ovarian cancer mortality. † RR’s are given for years of use at the dose specified in the left-hand column.

Despite IARC’s overall conclusion, the summary estimate suggested that coffee consumption (versus non-consumption) was associated with a small increase in ovarian cancer risk (RR = 1.3, 95% CI: 1.1–1.5). However, results were statistically significant in only two studies, and only one study found a significant dose response relationship (IARC, 1991). Four studies on coffee and ovarian cancer have been published since the monograph: two hospital case-control studies (Tavani, Gallus, Dal Maso, et al., 2001; Polychronopoulou et al., 1993), one population case-control study (Kuper et al., 2000), and one prospective cohort study (Stensvold and Jacobsen, 1994). Point estimates were generally consistent with the summary measure calculated by IARC, although none were statistically significant and most studies did not observe a dose-response relation. In one study, the association with coffee/caffeine consumption appeared particularly associated with ovarian cancer among premenopausal women or with serous-borderline and mucinous tumors (Kuper et al., 2000). Caffeine might increase ovarian cancer risk through several mechanisms. It has been hypothesized to affect not only the metabolism of DNA precursors, but the structure and function of DNA as well. In addition, it may affect hormone levels hypothesized to play an etiologic role in ovarian cancer (Kuper et al., 2000). While results have not often been significant, the estimated effect is fairly consistent across studies and seems to suggest a slight increase in risk. Given the widespread consumption of coffee and caffeine products, as well as the uncertain effects of potential selection and recall bias in the existing case-control studies, results from prospective studies will be important to clarify the association.

Alcohol Consumption of alcohol does not appear to be associated with ovarian cancer. While some studies report significant elevations (La Vecchia et al., 1992; Tzonou et al., 1984) or reductions in risk, (Goodman and Tung, 2003; Lagiou et al., 2001; Kato, Tominaga, and Terao, 1989) the majority of studies have failed to find any association (Tavani, Gallus, Dal Maso et al., 2001; Kuper et al., 2000; Nandkumar et al., 1995; Polychronopoulou et al., 1993; Hartge et al., 1989; Whittemore et al., 1988; IARC, 1988). In the one prospective study published to date, a significant inverse trend was observed between amount of alcohol consumed daily and ovarian cancer risk (Kushi et al., 1999). Further prospective analyses are needed and ideally would evaluate exposure at different points in the lifespan. While the current literature does not support an association, it has been shown that alcohol affects hormone levels, and thereby could influence risk of ovarian cancer (Hankinson et al., 1995).

Lactose/Galactose Evidence from case-control studies on the association between lactose/ galactose consumption and ovarian cancer risk has been conflicting (Cramer et al., 1984; Snowdon, 1985; Mori et al., 1988; Goodman et al., 2002b). An early study in which yogurt consumption was associated with ovarian cancer risk found that cases also had lower galactose transferase activity compared with controls (Cramer et al., 1989). However, subsequent studies failed to find associations between biologic measures, dietary measures, and ovarian cancer risk (Herrinton et al., 1995; Cramer et al., 2000). Furthermore, studies on polymorphisms related to the metabolism of lactose (N314D GALT gene) have not supported an association (Fung et al., 2003; Cozen et al., 2002; Goodman et al., 2002a), and results also are conflicting regarding lactose intolerance and ovarian cancer (Risch et al., 1994b; Meloni et al., 1999). However, more recently three prospective studies provided some support for an association between higher levels of lactose consumption and an increased risk of ovarian cancer. One prospective study found positive associations with consumption of dairy products in general, skim milk, and eggs, although results were only suggestive of an effect for lactose itself (Kushi et al., 1999). Two other prospective studies found positive associations between lactose intake and ovarian cancer, particularly for serous tumors (Fairfield et al., 2004; Larsson, Bergkvist, and Wolk, 2004). Finally, a meta-analysis con-

Ovarian Cancer cluded that prospective studies support a positive association between lactose intake and ovarian cancer risk although case-control studies do not (Larsson, Orsini, and Wolk, 2005). Dietary galactose and the enzyme galactose-l-phosphate uridyltransferase are linked to hypergonadotropic hypogonadism, and in turn, high secretions of gonadotropins have been hypothesized to cause ovarian cancer (Cramer et al., 1989). Further support of the proposed mechanism comes from studies in rats in which high galactose levels were found to decrease the number of oocytes (Herrinton et al., 1995). Given the existence of a plausible biologic mechanism but the inconsistency of the epidemiologic evidence, additional large prospective studies are needed to further examine this association. Furthermore, studies will ideally incorporate biologic as well as dietary measures when possible.

Antioxidants (Vitamin C, Vitamin E, Carotenoids) A protective effect of antioxidants against ovarian cancer has been suggested by some (Bidoli et al., 2001; Cramer et al., 2001; Fleischauer et al., 2001; McCann et al., 2003; Engle et al., 1991) but not all (La Vecchia et al., 1987; Tzonou et al., 1993) case-control studies. One study found a particularly strong association with consumption of raw carrots and tomato sauce, as well as carotene and lycopene (Cramer et al., 2001). In contrast, another study suggested that supplements were responsible for the protective effect of vitamins C and E (Fleischauer et al., 2001). Some suggestion has been made that overall dietary pattern, such as high consumption of fruits and vegetables, might be associated with reduced risk (McCann et al., 2003). However, results from two prospective cohort studies do not support an association. In the largest prospective study (with 301 cases), neither consumption of antioxidants (vitamins A, C, E, b-carotene) nor the foods that contain them (fruits and vegetables) was associated with risk. A modest inverse association with fruit and vegetable intake in adolescence was observed, suggesting that evaluation of diet at different points in the lifespan may be warranted (Fairfield et al., 2001). Similarly, despite observing a reduced risk of ovarian cancer with green leafy vegetable consumption (p for trend = 0.01), the Iowa Women’s Health Study did not find significant associations between specific antioxidants and ovarian cancer (Kushi et al., 1999).

Fat In 2001, a meta-analysis of seven case-control studies and one prospective study reported positive associations between ovarian cancer and total fat, saturated fat, and animal fat. Due to significant heterogeneity, the summary estimate comparing high and low consumption of total fat excluded the two hospital case-control studies (RR = 1.24, 95% CI: 1.07–1.43), although the point estimate seemed unaffected by the exclusion. Analyses of saturated fat and animal fat, each based on just three studies, found significant, positive associations of 1.20 and 1.70, respectively (Huncharek and Kupelnick, 2001). A case-control study in China also found a significant increase in risk comparing women in the highest versus lowest quartile of animal fat intake (Zhang et al., 2002). The few prospective cohort studies have not, however, supported an association (Bertone et al., 2002; Kushi et al., 1999). In the one prospective cohort study that performed detailed analyses by types of fat, point estimates were null for both broad categories of fat (e.g., saturated fat) and specific fatty acids (Bertone et al., 2002). Results from the second cohort study also did not consistently support an increased risk of ovarian cancer with increasing consumption of total, saturated, or animal fat (Kushi et al., 1999).

Fiber Only a few studies have examined the relationship between fiber and ovarian cancer. Two case-control studies found a 40–50% reduction in risk with high consumption of fiber (McCann et al., 2003) or vegetable fiber (Risch et al., 1994a). Two other studies found that consumption of whole grains (Chatenoud et al., 1998) or wholegrain bread and pasta (La Vecchia et al., 1987) was similarly protective against ovarian cancer. However, a prospective cohort study failed to find any

1017

evidence of an association (comparing highest to lowest tertiles, OR = 1.01, 95% CI: 0.61–1.68) (Kushi et al., 1999).

Body Size Although a growing body of research suggests that body mass index (BMI) and ovarian cancer are positively associated, a clear relationship remains to be established. In a recent review, the population casecontrol studies (n = 11) and cohort studies (n = 5) both yielded modest but statistically significant increases in risk with higher BMI (RR = 1.4 95% CI: 1.2–1.6 and RR = 1.2, 95% CI: 1.1–1.3, respectively). In contrast, results from the 13 hospital case-control studies were heterogeneous (p < 0.001), with positive, negative, and null point estimates (Purdie et al., 2001), likely due in part to the challenge of appropriate control selection in this setting. Subsequent to the review by Purdie et al. (2001), several studies found moderate positive associations between BMI and ovarian cancer, although some findings were restricted to a particular subgroup. (Schouten et al., 2003; Calle et al., 2003; Rodriguez et al., 2002; Dal Maso et al., 2002; Fairfield et al., 2002; Kuper et al., 2002) In two studies that stratified by menopausal status, the positive association with BMI appeared to be limited to premenopausal ovarian cancer (Fairfield et al., 2002; Kuper et al., 2002) In one of these studies, the effect was further restricted to BMI at age 18 (Fairfield et al., 2002). In another large prospective study of ovarian cancer mortality in postmenopausal women, a positive association was observed only among women who had never used estrogen replacement therapy (ERT) (Rodriguez et al., 2002). Few studies have had sufficient power to evaluate histologic subtypes separately, another source of potential heterogeneity. One study reported an increase in risk that was particularly strong among serous borderline tumors. (Kuper et al., 2002) A collaborative reanalysis also found that high BMI was associated with an increased risk of epithelial borderline tumors (Harris et al., 1992). The relationship between waist to hip ratio (WHR) and ovarian cancer has been assessed in only two prospective studies to date. In one, a significant increase in risk with abdominal adiposity was observed despite null results for BMI overall, (Mink et al., 1996) while no association was seen in the second (Fairfield et al., 2002) However, neither study was large, and further assessments are needed. A high BMI has been proposed to influence ovarian cancer through several mechanisms, including its effects on endogenous hormone levels and association with infertility (Fairfield et al., 2002). A high WHR is associated with high androgen and insulin levels (Mink et al., 1996). Attained height in adulthood may partially reflect hormonal and nutritional factors in early life, variables for which information is not often available (Schouten et al., 2003). Several of the more recent studies have observed a positive association between height and ovarian cancer (Schouten et al., 2003; Kuper et al., 2002; Rodriguez et al., 2002) although one study reported an inverse association (Dal Maso et al., 2002).

Physical Activity Research has suggested that physical activity may increase, decrease, or have no effect on ovarian cancer, prompting IARC to deem the data inconclusive in a 2002 review of five studies. (IARC, 2002b) The literature’s inconsistency may reflect differences in the effects of physical activity by frequency, duration, intensity, and point in the lifespan, as well as the infrequent examination of the combination of occupational and leisure-time activity. One study assessed both occupational and recreational physical activity, but results were still stratified by type of activity. Comparing the highest and lowest levels of activity, the data suggested a modest decreased risk of ovarian cancer for both occupational activity and recreational activity (Tavani et al., 2001). The majority of studies to date have focused on recreational physical activity. One case-control study assessed physical activity at six points in the lifespan. A significant decrease in ovarian cancer was observed comparing the highest and lowest lifetime physical activity groups (OR = 0.73, 95% CI: 0.56,

1018

PART IV: CANCER BY TISSUE OF ORIGIN

0.94), and there was a significant dose-response trend (p = 0.01) (Cottreau et al., 2000). A population case-control study did not find a significant association using total or metabolic equivalent task (MET) hours of recreational activity, but point estimates suggested a protective effect of vigorous physical activity (Bertone et al., 2002). In contrast, results from two large prospective cohort studies suggested a positive association between physical activity and ovarian cancer, particularly vigorous activity, although again results were not always statistically significant (Bertone et al., 2001; Mink et al., 1996). Studies comparing occupational categories to expected numbers of cases have also produced conflicting results (Pukkala et al., 1993; Zheng et al., 1993). Plausible mechanisms exist to explain how physical activity might increase or decrease risk. For instance, evidence suggests exercise can either increase or decrease frequency of ovulation, depending on the amount and intensity of the activity. Alternatively, physical activity might reduce risk by decreasing estrogen levels, or the decreased estrogen levels might increase risk through negative-feedback mechanisms, resulting in increased production of gonadotropins (Bertone et al., 2001). While clarification of the association between physical activity and ovarian cancer will not influence public health recommendations for exercise, a better understanding of the relationship might yield important information about the underlying etiology of ovarian cancer.

Tubal Ligation and Simple Hysterectomy Epidemiologic studies indicate that tubal ligation and hysterectomy are each associated with a decreased risk of ovarian cancer. In a 1993 review of 14 studies, women who had a tubal ligation were estimated to have 0.63 times the risk of ovarian cancer as women without a tubal ligation (95% CI: 0.44–0.91) (Hankinson et al., 1993). Subsequent to the review, several studies provided further support for an association (RR range: 0.25 to 0.71), with most point estimates close to the estimated summary RR (Modugno et al., 2001; Narod et al., 2001; Cornelison et al., 1997; Green et al., 1997; Miracle-McMahill et al., 1997; Kreiger et al., 1997; Rosenblatt and Thomas, 1996; Nandakumar et al., 1995) (Fig. 52–5). Hysterectomy has also been associated with a reduced risk of ovarian cancer, although the evidence is less convincing. In a collaborative analysis of eight case-control studies, both hos-

pital studies (RR = 0.66) and population studies (RR = 0.88) suggested a decreased risk, but nonsignificantly so for the population studies. (Whittemore et al., 1992) Results of subsequent studies were consistent with a decreased risk (Modugno et al., 2001; Green et al., 1997; Purdie et al., 1995; Kreiger et al., 1997; Loft et al., 1997; Rosenblatt and Thomas, 1996; Hankinson et al., 1993). Tubal ligation and hysterectomy have been hypothesized to decrease the risk of ovarian cancer by blocking carcinogens from reaching the ovaries, altering ovulation, or reducing exposure to hormones through decreased blood flow to the ovaries. Two main noncausal explanations for these associations have been proposed: that the association is confounded by parity, or that the decreased risk is the result of selective screening of the ovaries at the time of the procedure (Weiss and Harlow, 1986). However, in most studies, the inverse association has been observed when controlling carefully for parity and contraceptive use, tumors appear to be rarely diagnosed at the time of surgery, and the protective effect appears to persist for years after the surgery.

Chemical Agents Occupational cohort studies have suggested that asbestos increases the risk and mortality of ovarian cancer (Germani et al., 1999; VasamaNeuvonen et al., 1999; Berry et al., 2000). However, studies have had limited ability to control for confounding by factors other than age. (Germani et al. 1999) Even in a study that attempted to control for other potentially important confounders, mean levels for each job title were used instead of individual measurements (Vasama-Neuvonen et al., 1999). Supporting an association, asbestos fibers have been found in ovarian tissue (Heller et al., 1999). Interest in talc was stimulated by its similarity to asbestos. A summary estimate from 14 case-control studies found a significant increase in risk of ovarian cancer associated with perineal talc exposure (OR = 1.36, 95% CI: 1.24–1.49), although no dose-response relationship was evident (Cramer, 1999). A subsequent prospective analysis found a positive association of similar magnitude (RR = 1.40) specifically for serous tumors. (Gertig et al., 2000) While the association is plausibly causal and studies are broadly consistent, further understanding of the biology and the ability of talc to be transported to the ovaries is necessary (Harlow and Hartge 1995).

Modugno et al. (2001) Narod et al. (2001) Green et al. (1997) Cornelison et al. (1997) Miracle-McMahill et al. (1997)* Rosenblatt et al. (1996) Nandakumar et al. (1995) Hankinson (1993)**

0.5

1

1.5

Relative risk – log scale

Figure 52–5. Tubal ligation and ovarian cancer. *Study on mortality and ovarian cancer. **Summary estimate from Hankinson et al. (1993). Included results from Koch et al., 1984; Koch et al., 1988; Mori et al.,

1988; Booth et al. 1989; Shu et al., 1989; Whittemore et al., 1992; Hankinson et al., 1993.

Ovarian Cancer

Smoking Despite an initial report in 1980 that smoking was positively associated with ovarian cancer mortality, subsequent research has generally failed to support an association. However, more recent studies suggest that smoking may be associated with certain histologic types of ovarian cancer (Terry et al., 2003). In 2000, Marchbanks et al. updated the analysis of a population case-control study that had previously reported a null association between smoking and ovarian cancer. The updated analysis reported results by histologic type and found a positive association for mucinous tumors only (Marchbanks et al., 2000). In several subsequent reports, a positive association with smoking was observed either primarily or exclusively with the mucinous subtype (Kuper et al., 2000; Green et al., 2001; Modugno et al., 2002; Terry et al., 2003) Smoking of several decades duration also may increase the risk of ovarian cancer for nonmucinous tumors, but only two studies have examined this question (Terry et al., 2003; Kuper et al., 2000). Biologically, there is reason to believe that mucinous and non-mucinous tumors may have different etiologies (Terry et al., 2003). Researchers have noted that mucinous cells are similar to those of the cervix and intestine, sites where smoking has been implicated as a factor in cancer development (Marchbanks et al., 2000). Additionally, some mucinous ovarian tumors may actually be metastases from the gastrointestinal tract (Seidman et al., 2003). Nonetheless, despite the consistency of results from more recent studies, the small numbers of mucinous tumors in each study limited the authors’ ability to evaluate the dose-response relationship in detail.

Radiation Information on the relationship between ovarian cancer and radiation comes primarily from two types of studies: (1) follow-up of survivors of the atomic bomb, and (2) studies of women treated with radiation for other conditions. Over the period 1950 to 1990, women exposed to the atomic bomb had an excess relative risk of ovarian cancer of 0.87 per sievert, a small but statistically significant increase (Tavassoli et al., 1996). A previous report found that women exposed to 100 or more rads had 2.2 times the risk of women exposed to 0 rads. It had further estimated the latency between exposure to development of ovarian cancer to be approximately 15 to 20 years (Tokuoka et al., 1987). In 1995, the Hiroshima International Council for Medical Care of the Radiation-Exposed classified ovarian cancer as a type of cancer for which increased risk due to radiation was “suggested” (Shigematsu, 1995). In a study of women treated for cervical cancer, there was a short-term deficit in ovarian cancer (RR = 0.13) followed by a statistically non-significant increased risk among those who survived for 10 years (RR = 1.4, 90% CI: 0.3–5.6). The authors hypothesized that the radiation had initially killed premalignant cells (Boice et al., 1988). However, some studies have not found that individuals exposed to radiation have long-term increased mortality or risk (Darby et al., 1987; Wang et al., 1988).

HOST FACTORS Reproductive Factors Age at Menarche and Age at Menopause Neither age at menarche nor age at menopause appears to be strongly associated with ovarian cancer, although weak associations have been suggested. In an analysis of six population case-control studies, a slight but nonsignificant increase was observed with increasing age at natural menopause (RR = 1.09, 95% CI: 0.99–1.20) (Schildkraut et al., 2001). A pooled analysis of three hospital case-control studies likewise found no association between age at menarche and ovarian cancer. However, an approximate doubling of risk among women whose menopause was at age 53 and older compared to women with menopause before age 45 was observed (Franceschi et al., 1991). These results were not supported by a prospective cohort study, however, that did not find significant effects of age at menarche or age

1019

at menopause (Kvale et al., 1992) Another prospective cohort study also failed to find a trend associated with age at menarche or age at menopause, although point estimates suggested a potential decrease in risk among those with later menarche (Hankinson et al., 1995). It has been proposed that ovarian cancer results from repeated trauma of the ovary epithelium caused by ovulation. If so, younger age at menarche or later age at menopause might increase risk of ovarian cancer by increasing the number of ovulations. Supporting this, positive associations have been observed between estimates of a woman’s “lifetime number of ovulations” and ovarian cancer risk (Purdie et al., 2003).

Parity, Age at First Birth, Age at Last Birth Parity is well-established as a protective factor for ovarian cancer. A collaborative reanalysis of hospital and population case-control studies found that women with at least one live birth were at a substantially lower risk of ovarian cancer compared with nulliparous women (OR = 0.76, 95% CI: 0.63–0.93 and OR = 0.47, 95% CI: 0.40–0.56, respectively). Each additional pregnancy offered further protection (OR = 0.87 and 0.81 per birth, respectively, p < 0.001 for both) (Whittemore et al., 1992). Similar findings have been reported in several cohort studies. The roles of a woman’s age at first birth and last birth remain unclear. While there is some suggestion that older age at first and last birth may be protective (Whiteman et al., 2003; Cooper et al., 1999), other studies find no association (Hankinson et al., 1995). In 1994, Adami et al. hypothesized that pregnancy might be protective because it clears cells which have undergone malignant transformation from the ovaries. If true, this theory suggests that late age at pregnancy might exert greater protection than an earlier age, because the likelihood of malignant transformation increases with age and correspondingly the potential benefit of clearing cells increases (Adami et al., 1994).

Lactation Although conflicting results have been reported, lactation seems to decrease risk of ovarian cancer. However, longer durations of breastfeeding may not increase the protective effect (Whittemore et al., 1992; Rosenblatt and Thomas, 1993). Biologically, breastfeeding immediately postpartum might be expected to offer the most protection, because suppression of ovulation is greatest during this period (Siskind et al., 1997). In a collaborative re-analysis of two hospital and five population case-control studies, ever breastfeeding was associated with a decreased risk of ovarian cancer among parous women (OR = 0.73, 95% CI: 0.51–1.0 and OR = 0.81, 95% CI: 0.68–0.95, respectively). Each month of breastfeeding seemed to further reduce risk (RR = 0.99), but the effect of breastfeeding was also strongest within 6 months of delivery (Whittemore et al., 1992). A similar decrease in risk was observed in a population case-control study (OR = 0.7, 95% CI: 0.5–1.0). There was no significant trend with average or total duration of breastfeeding (Titus-Ernstoff et al., 2001). A multinational hospital based case-control study included several developing countries, allowing an assessment of long durations of breastfeeding (up to 48 months or more). A nonsignificant decrease in ovarian cancer was observed with 5–8 months of breastfeeding (OR = 0.82), and no further decrease was suggested with longer periods of lactation (Rosenblatt and Thomas, 1993). A study which assessed exclusive breastfeeding suggested a decrease in ovarian cancer among premenopausal women who breastfed, but results were not significant (Siskind et al., 1997).

Infertility and Fertility Drugs In 2002, a pooled analysis of eight population case-control studies evaluated the effect of fertility drugs on ovarian cancer risk among subfertile women. Subfertility was self-reported as at least two years of unsuccessful attempts to conceive or medical attention for infertility. Overall, fertility drugs showed no association with ovarian cancer risk (OR = 0.97, 95% CI: 0.76–1.25). Among women who were never pregnant, the point estimate was nonsignificantly increased (OR = 1.60, 95% CI: 0.90–2.87). Duration of use did not show any evidence

1020

PART IV: CANCER BY TISSUE OF ORIGIN

of a dose-response relationship. However, the study did observe a significant increase for borderline serous tumors among women who had never been pregnant (OR = 2.43, 95% CI: 1.01–5.88). (Ness et al., 2002) In analyses that included fertile and infertile women, infertility itself was associated with an increase in risk of ovarian cancer. Endometriosis was associated with a 1.73-fold increase in the odds of ovarian cancer (95% CI: 1.10–2.71), while unknown causes of fertility were associated with a more modest increase (OR = 1.19) (Ness et al., 2002). In contrast, in a pooled analysis of 12 case-control studies comparing infertile women using fertility drugs with fertile women, no evidence was found of increased risk with infertility, but a significant increase (OR = 2.8, 95% CI: 1.3–6.1) was associated with use of fertility drugs (Whittemore et al., 1992). Cohort studies have been suggestive of a role of fertility drugs (Rossing et al., 1994) or infertility (Rodriguez et al., 1998).

consistent findings have yet to emerge. Variants in several glutathione S-transferases, a family of phase II enzymes, have been assessed, and results have been generally null (Coughlin and Hall, 2002). As described above, several studies evaluated the N314D polymorphism in the galactose-1-phosphate uridyltransferase (GALT) gene, one of the genes involved in lactose metabolism, in relation to risk; here too, most studies have not detected an association. Further assessments in large studies with additional consideration of haplotypes and gene-gene interactions will be needed to move this area forward.

Somatic Alterations in Ovarian Cancers

Women who have a family history of ovarian cancer are at a substantially higher risk of developing ovarian cancer. In a meta-analysis that combined estimates from 13 case-control and two cohort studies (Stratton et al., 1998), the relative risk in women with a first-degree family history was 3.1 (95% CI: 2.6–3.7). In women less than 40 years of age with breast cancer, having a family history of breast or ovarian cancer has been observed to increase their risk of ovarian cancer 5.6fold (95% CI: 1.8–13.0) or 17-fold (95% CI: 3.5–50.0), respectively (Bergfeldt et al., 2002).

A large number of chromosomal alterations have been observed in ovarian tumors. The chromosomal regions most commonly affected include 3p, 6q, 11p, 17q, and 17p13 (Liu and Ganesan, 2002). A growing number of specific tumor suppressor genes, such as TP53 and PTEN, have been implicated in ovarian cancer pathogenesis based on mutations or deletions observed in tumors. Recent data suggest that epigenetic events, such as gene hypermethylation (resulting in decreased gene expression) also may play an important etiologic role (Ahluwalia et al., 2001). However, to date, few studies have evaluated epidemiologic risk factors in relation to ovarian tumors with specific gene alterations. In one of the first studies to do so, risk factors associated with increases in the number of ovulatory cycles were more strongly associated with tumors found to have abnormal P53 protein expression (Schildkraut et al., 1997); however, these findings were not confirmed in a subsequent study (Webb et al., 1998).

Genetic Predisposition/Major Gene Mutations

Predisposing Diseases

The several hereditary forms of ovarian cancer have been proposed to cause about 10% of all ovarian cancers. Mutations in BRCA1 and BRCA2 (which are also an important cause of hereditary breast cancer) are responsible for the vast majority of these cases, with additional small contributions made by hereditary non-polyposis colorectal cancer (HNPCC) syndrome and possibly other as-yet-unidentified mutations. The population frequency of BRCA1 and BRCA2 mutations varies substantially. In the general population, an estimated prevalence is 1 in 280 (range 1 in 112 to 1 in 441). In the Ashkenazi Jewish population, approximately 1 in 40 women are carriers; most of these women have one or more specific founder mutations in BRCA1 or BRCA2 (i.e., 185delAG and 5382insC in BRCA1 and 6174delT in BRCA2) (Boyd, 2003). Estimates of the probability of developing ovarian cancer among BRCA1 or BRCA2 mutation carriers range from 35–50% for BRCA1 and 10–30% for BRCA2 (Antoniou et al., 2003; Boyd, 2003).

In 1925, Sampson proposed a link between endometriosis and ovarian cancer. Subsequent research has largely supported this hypothesis, with an estimated 0.3% to 1.6% of endometriosis cases progressing to ovarian cancer, particularly endometrioid and clear cell histologies (Swiersz, 2002). Clinical studies and case series dominate the literature, but epidemiologic research also supports an association. In two studies, endometriosis was associated with an approximate doubling of risk (Brinton et al., 1997; Ness, 2003). Some researchers, accepting the association as well-established, have begun to focus on whether and how ovarian cancers associated with endometriosis might be different from those not associated with endometriosis (Erzen et al., 2001). Limited data exists on the relationship between other predisposing conditions and ovarian cancer. Mumps has been associated with ovarian cancer, and it was hypothesized that the infection might damage the ovaries, thereby impacting ovarian function. Animal studies suggest that the effect might occur through oocyte depletion (Cramer et al., 1983). Polycystic ovarian cancer has also been associated with a significant increase in risk in one study, possibly through increased androgen levels (Schildkraut et al., 1996). In contrast, a study of pelvic inflammatory disease did not find evidence of an increase in risk (Parazzini et al., 1996).

Family History

Gene Polymorphisms and Ovarian Cancer Risk It is likely that a number of more common gene polymorphisms also influence risk of ovarian cancer, although the magnitude of the associations would be expected to be much weaker. The identification and evaluation of potential candidate gene polymorphisms is a relatively recent phenomenon with ovarian cancer (at least compared with the large number of molecular epidemiology studies conducted on breast and colon cancers); hence much remains to be learned. A number of studies have evaluated a polymorphism in the progesterone receptor, the PROGINS allele, in relation to ovarian cancer risk. Study results have been inconsistent, although the larger studies have tended to observe no association (McKenna et al., 1995; Manolitsas et al., 1997; Lancaster et al., 1998; Runnebaum et al., 2001; Tong et al., 2001; Tong et al., 2002; Lancaster et al., 2003). Other functional SNPs have been found in the progesterone receptor (De Vivo, 2002) but have yet to be evaluated in relation to ovarian cancer. Several groups have evaluated polymorphisms in other genes involved in hormone metabolism or action such as CYP17 (central to the synthesis of the androgens, androstenedione and dihydroepiandrosterone), and the androgen receptor (Spurdle et al., 2000; Goodman et al., 2001; Garner et al., 2002). Although this is likely to be a fruitful area,

Endogenous Hormones Because of the potential influence of clinically evident disease on the hormonal axes, associations between circulating hormone levels and ovarian cancer are best assessed prospectively, although few such studies have been conducted to date. The association between androgens and ovarian cancer has been evaluated in three prospective studies. In one, with 12 cases, urinary DHEA levels were inversely related to ovarian cancer risk (Cuzick et al., 1983). In the second, with 31 cases, plasma androstenedione was positively related to risk (Helzlsouer et al., 1995). Mean DHEA levels also were significantly higher in cases than controls. In the largest study, a pooled analysis of three nested case-control studies with 132 cases and 264 controls, a non-significant positive association was noted with plasma testosterone, but no association was observed with DHEAS (Lukanova et al., 2003). When assessed by menopausal

Ovarian Cancer status, associations appeared stronger in the small subset of premenopausal women. In the only study to date of gonadotropin levels, mean follicle stimulating hormone levels were significantly lower in cases than in controls (p = 0.04), while no association was observed with luteinizing hormone (p = 0.39) (Helzlsouer et al., 1995). In the only prospective study to date, no significant relationship was observed overall for IGF-I or IGFBP-3 (Lukanova et al., 2002). However, among the small group of cases <55 years of age at diagnosis, a strong positive association was noted with IGF-I (RR 4.98 [95% CI 1.21–20.6] for the top to bottom tertile). A stronger association among young women has also been observed for breast cancer, (Hankinson et al., 1998; Muti et al., 2002; Toniolo et al., 2000) and thus is conceivable here. C-peptide, a marker of insulin secretion, was also evaluated in this study, but no significant association was observed (Lukanova et al., 2002).

Other Serologic Markers Recent advances in mass spectrometry may make feasible the simultaneous assessment of hundreds of proteins (“proteomics”) in relation to risk (Weinberger et al., 2002). New approaches to protein-profiling in blood include surface-enhanced laser desorption and ionization time-of-flight mass spectroscopy (SELDI-TOF) (Issaq et al., 2002). In one of the first evaluations of this technology, SELDI-TOF was used to compare sera from normal women and women with Stage I ovarian cancer (Petricoin et al., 2002). Petricoin et al. found a combination of five plasma proteins that, in a masked set of cases and controls, distinguished all 50 women with ovarian cancer and 63 of 66 patients without malignant disease. Although the findings have not yet been replicated, nor have the specific protein peaks been identified, this initial study suggests the considerable potential of a proteomics approach for identifying markers of risk or early tumor development.

PATHOGENESIS The human ovarian surface epithelium is made up of a single layer of squamous-to-cuboidal epithelial cells (Auersperg et al., 2001). It is now thought that most epithelial tumors arise from the epithelial cells that line ovarian cortical inclusion cysts rather than cells on the surface of the ovary. It is not known exactly why the cysts form, although they may begin when a small part of the surface epithelium becomes trapped in or close to ruptured follicles at the time of ovulation or as a result of inflammatory adhesions involving the ovary and peritoneum (Radisavljevic, 1977; Scully, 1995). Why ovarian cancers preferentially arise from epithelial cells in the inclusion cysts rather than the surface epithelium is also unknown. However, the epithelium in the cysts may be more heavily exposed to hormones and cytokines from the stromal microenvironment and through autocrine mechanisms due to the confined space of the cyst (Auersperg et al., 2001). A number of hypotheses exist regarding the etiology of ovarian cancer. One of the first to be proposed was that “incessant ovulation,” resulting in repeated injury and repair to the surface epithelium, would increase epithelial proliferation and thus the frequency of DNA mutations (Fathalla, 1971). Support for this mechanism includes the high ovarian cancer rates in domestic fowl that ovulated daily (Fathalla, 1971; Fredrickson, 1987). The inverse associations observed with factors that decrease a woman’s lifetime number of ovulations (i.e., parity, breast feeding, oral contraceptive use) also provide indirect support for this theory. However, the inverse association observed with one full-term pregnancy appears greater than would be expected by an equivalent period of anovulation alone, (Whittemore et al., 1992) and hence, at a minimum, additional factors are likely to play a role. A number of other hormonal hypotheses have been proposed. One of the first was the “gonadotropin” hypothesis, whereby high levels of gonadotropic hormones could either directly or indirectly (through increases in ovarian stromal estrogen production) increase proliferation and the possibility of malignant transformation of the epithe-

1021

lium in inclusion cysts (Cramer and Welch, 1983). FSH and LH receptors are present on normal and malignant ovarian epithelial cells (Auersperg et al., 2001). Both oral contraceptives and breastfeeding decrease gonadotropin levels; radiation exposure, leading to premature ovarian failure, increases levels. In the only study of circulating FSH and LH among 31 ovarian cancer cases, an inverse association was suggested with FSH (Helzlsouer et al., 1995). More recently, important roles for both androgens (in increasing risk) and progesterone (in decreasing risk) have been proposed (Risch, 1998). Factors supporting a role for progesterone include the potent induction of apoptosis by progestins in the primate ovarian epithelium (Rodriguez et al., 1998), the high progesterone levels during pregnancy (which might contribute to the protective effect of parity), and the inverse association with OCs, including limited data on formulations containing progestin alone (Rosenberg et al., 1994). OCs with high doses of progestins were more strongly associated with reduced risk than lower-dose formulations in one study but not in a second (Schildkraut et al., 2002; Ness et al., 2000). In regards to androgens, epithelial ovarian cells trapped in inclusion cysts are exposed to high levels of androgens (Risch, 1998), most normal surface epithelia express androgen receptors (Edmondson et al., 2002), and testosterone increases proliferation of both normal and malignant ovarian epithelium (Syed et al., 2001; Godwin et al., 1992; Ahonen et al., 2000). In several recent reports, obese premenopausal women, who tend to have higher endogenous androgen levels, were at higher risk of ovarian cancer than lean women (Harlass et al., 1984; Pasquali et al., 1989). OC use is also known to decrease ovarian testosterone levels (van der Vange et al., 1990; Kuhnz et al., 1991). An association with waist-tohip ratio or polycystic ovarian syndrome, which are also associated with higher androgen levels, would provide further support for this theory, but to date only limited or inconsistent evidence exists (Mink et al., 1996; Fairfield et al., 2002; Schildkraut et al., 1996). In the few studies of circulating androgens and risk, associations have tended to be positive, but were not always statistically significant. Other hormones or growth factors also may play a role in ovarian cancer pathogenesis. For example, the insulin-like growth factor (IGF) axis plays an important role in normal ovarian physiology (Takahashi et al., 1996). IGF-I is a potent mitogen for both normal and neoplastic cells (Giudice et al., 1996; LeRoith et al., 1995). IGF-1, its receptor, and several IGF binding proteins are all expressed in ovarian cancer cell lines, and in vitro studies demonstrate proliferation of ovarian cancer cell lines in the presence of higher IGF-1 concentrations (Beck et al., 1994; Yee et al., 1991). Several recent studies found a positive association between height and ovarian cancer risk, supporting a possible role for early-life IGF. In the only prospective study conducted to date, a positive association between circulating IGF-I level and risk was observed, primarily in the small subgroup of premenopausal women. Oxidative damage to cells, and to DNA specifically, has been hypothesized as a cause for many cancers, including ovarian cancer. Oxidative damage may occur in the ovary from frequent epithelial injury and repair due to ovulation (Casagrande et al., 1979) or from oxidative radicals released during steroidogenesis (Riley and Behrman, 1991). As such, antioxidants may decrease risk of the disease (Frei, 1994). To date, only a limited number of antioxidants have been evaluated (e.g., vitamins C and E), and results have not consistently supported an inverse association. The “pelvic contamination” theory suggests that contaminants such as talc or asbestos, which may reach the ovaries from the uterus or vagina, increase risk. (Whittemore et al., 1988). The consistent inverse associations of tubal ligation and simple hysterectomy with ovarian cancer, and the relatively consistent positive associations with talc use and endometriosis would all support this hypothesis. With the exception of talc use, these factors also may alter ovarian or uterine function and hence hormone or growth factor levels (Cramer and Xu, 1995), another possible mechanistic pathway underlying these exposures. Also related to contamination theories, inflammation of the ovarian epithelium has been proposed to play a key role in cancer etiology (Ness and Cottreau, 1999). Factors such as talc and endometriosis may cause cellular inflammation, leading to oxidative DNA

1022

PART IV: CANCER BY TISSUE OF ORIGIN

damage and carcinogenesis. Tubal ligation and hysterectomy may reduce ovarian exposure to such irritants. These etiologic theories are in no way mutually exclusive. In fact it is most likely that several work in concert or in parallel to influence risk. Clearly, there is some overlap among the pelvic contamination, inflammation, and oxidative damage hypotheses. In addition, while estrogen increases cell proliferation, certain of its metabolites may be directly genotoxic through oxidative mechanisms (Yager and Liehr 1996). These theories have served to point epidemiologists towards additional exposures to consider in their studies and suggest a wide range of potential environmental and gene-environment interactions to consider in future research.

PREVENTIVE MEASURES Primary Prevention Few modifiable factors have been established for ovarian cancer, thus limiting opportunities for primary prevention. Oral contraceptives and (possibly) tubal ligation are modifiable factors. Among women with children, breastfeeding may also reduce risk. Research on such factors as dietary intake and exercise is not sufficiently well confirmed to form primary prevention recommendations for ovarian cancer at this time.

Screening and Early Detection In 1996, the United States Preventive Services Task Force recommended against the use of ultrasound, serum tumor markers (e.g., CA 125), and pelvic examinations to screen average-risk women for ovarian cancer. For women at above-average risk, the Task Force concluded there was insufficient evidence to make a recommendation. Proposed screening tests have not been shown to have acceptable sensitivity (especially for early-stage disease), specificity, and positive predictive value. Because positive screening tests are followed up with invasive tests, false positives are of particular concern when screening for ovarian cancer (U.S. Preventive Services Task Force, 1996). CA125, a glycoprotein, is often used to monitor women with ovarian cancer and has been evaluated as a potential screening test. However, despite its frequent elevation among those with ovarian cancer (e.g., over 85% of advanced cases have elevated levels), it is not sufficiently specific to be used as a screening test alone because approximately 6% of non-ovarian cases also have elevated levels. More recent research on screening has focused on the potential use of multiple screening markers, such as M-CSF, a cytokine found in serum, or LPA, a plasma tumor marker (Urban, 2003). Two screening trials have been planned to evaluate combinations of existing tests. In the United States, the Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial will study the effectiveness of CA125 and transvaginal ultrasound in 37,000 women over a 23-year period (Prorok et al., 2000). The UK Collaborative Trial of Ovarian Cancer Screening has been planned as a three-armed randomized controlled trial that will include screening by ultrasound, ultrasound and CA125, or an unscreened control group (Menon and Jacobs, 2000).

FUTURE DIRECTIONS Over the last decade, substantial progress has been made both in developing new hypotheses regarding ovarian cancer etiology (e.g., the potential importance of inflammation) and in further delineating and identifying risk factors for ovarian cancer (e.g., tubal ligation, analgesic use). In Table 52–2, we classify risk factors as being confirmed, probable, or possible. Additional work is needed to confirm or discount those currently considered probable or possible risk factors. It is becoming increasingly apparent that risk factor / disease relationships may vary by histologic tumor type. Distinguishing invasive from borderline and accurately identifying serous, mucinous, endometrioid, and clear cell tumors may add substantial clarity to the

Table 52–2. Summary of Ovarian Cancer Risk Factors Direction and Magnitude of Association*

well-confirmed risk factors Oral contraceptive use Parity Tubal ligation Family history BRCA carrier Endometriosis

probable risk factors Postmenopausal hormone use Lactation Hysterectomy Asbestos Talc Radiation

data inconsistent Age at menarche (older) Age at menopause (older) Alcohol Lactose / galactose Coffee / caffeine Fat intake Antioxidant intake Body Mass Index Physical activity Smoking

limited data available Analgesic use Waist-to-hip ratio Height Infertility and fertility drugs Fiber

ØØ ØØ ØØ ≠≠ ≠≠≠ ≠≠ ≠ Ø Ø ≠≠ ≠ ≠≠ Ø ≠ ´ ≠ ≠ ≠Ø Ø ≠ ≠Ø ≠ ØØ ≠ ≠ ≠ Ø

*slight (≠); moderate (≠≠), or substantial (≠≠≠) increase in risk; slight (Ø); moderate (ØØ); or substantial (ØØØ) decrease in risk; ´ no change in risk. ≠Ø findings particularly inconsistent.

observed relationships (these categories are suggested with the recognition that there are more possible subcategories, such as the intestinal versus endocervical type of borderline mucinous tumors). In addition to using histopathologic techniques for tumor classification, assessing somatic mutations and gene expression profiling should ultimately allow for even more homogeneous groupings. Within the highrisk subgroup of BRCA1 and BRCA2 carriers, further assessments are needed of risk factors that have been studied primarily in general population samples to determine their influence on gene penetrance. Several recent studies of body weight, height and IGF suggest that additional attention to early-life events also may yield important information on etiology. Advances in genomics will allow the careful assessment of gene-environment interactions, both to help confirm specific exposure/cancer relationships and to identify women who may derive the most benefit from either increased surveillance or lifestyle change. Ongoing prospective studies of stored biologic samples are needed to evaluate serum markers that either predict disease risk or identify women with early disease. In all of these efforts, the rarity of some of these tumor types will increase the need for both large epidemiologic studies and the pooling of data across multiple studies. References Adami HO, Hsieh CC, Lambe M, et al. 1994. Parity, age at first childbirth, and risk of ovarian cancer. Lancet 344:1250–1254. Ahluwalia A, Yan P, Hurteau JA, et al. 2001. DNA methylation and ovarian cancer. I. Analysis of CpG island hypermethylation in human ovarian cancer using differential methylation hybridization. Gynecol Oncol 82:261–268. Ahonen MH, Zhuang YH, Aine R, Ylikomi T, Tuohimaa P. 2000. Androgen receptor and vitamin D receptor in human ovarian cancer: Growth stimulation and inhibition by ligands. Int J Cancer 86:40–46.

Ovarian Cancer Akhmedkhanov A, Toniolo P, Zeleniuch-Jacquotte A, Kato I, Koenig KL, Shore RE. 2001. Aspirin and epithelial ovarian cancer. Prev Med 33:682–687. American Cancer Society (ACS). 2003. Cancer Facts & Figures. Atlanta: American Cancer Society, Inc. Anderson GL, Judd HL, Kaunitz AM, et al. 2003. Effects of estrogen plus progestin on gynecologic cancers and associated diagnostic procedures: The Women’s Health Initiative randomized trial. JAMA 290:1739–1748. Antoniou A, Pharoah PD, Narod S, et al. 2003. Average risks of breast and ovarian cancer associated with BRCA1 or BRCA2 mutations detected in case Series unselected for family history: A combined analysis of 22 studies. Am J Hum Genet 72:1117–1130. Auersperg N, Wong AS, Choi KC, Kang SK, Leung PC. 2001. Ovarian surface epithelium: biology, endocrinology, and pathology. Endocr Rev 22:255–288. Beck EP, Russo P, Gliozzo B, et al. 1994. Identification of insulin and insulinlike growth factor I (IGF I) receptors in ovarian cancer tissue. Gynecol Oncol 53:196–201. Bergfeldt K, Rydh B, Granath F, et al. 2002. Risk of ovarian cancer in breastcancer patients with a family history of breast or ovarian cancer: A population-based cohort study. Lancet 360:891–894. Berry G, Newhouse ML, Wagner JC. 2000. Mortality form all cancers of asbestos factory workers in east London 1993–80. Occup Environ Med 57:782–785. Bertone ER, Newcomb PA, Willett WC, Stampfer MJ, Egan KM. 2002. Recreational physical activity and ovarian cancer in a population-based casecontrol study. Int J Cancer 99:431–436. Bertone ER, Rosner BA, Hunter DJ, et al. 2002. Dietary fat intake and ovarian cancer in a cohort of US women. Am J Epidemiol 156:22–31. Bertone ER, Willett WC, Rosner BA, et al. 2001. Prospective study of recreational physical activity and ovarian cancer. J Natl Cancer Inst 93: 942–948. Bidoli E, La Vecchia C, Talamini R, et al. 2001. Micronutrients and ovarian cancer: A case-control study in Italy. Ann Oncol 12:1589–1593. Boice JD Jr., Engholm G, Kleinerman RA, et al. 1988. Radiation dose and second cancer risk in patients treated for cancer of the cervix. Radiat Res 116:3–55. Bosetti C, Negri E, Franceschi S, Trichopoulos D, Beral V, La Vecchia C. 2001. Relationship between postmenopausal hormone replacement therapy and ovarian cancer. [comment]. JAMA 285:3089; discussion 3090. Bosetti C, Negri E, Trichopoulos D, et al. 2002. Long-term effects of oral contraceptives on ovarian cancer risk. Int J Cancer 102:262–265. Booth M, Beral V, Smith P. 1989. Risk factors for ovarian cancer: A casecontrol study. Br J Cancer 60:592–598. Boyd J. 2003. Specific keynote: hereditary ovarian cancer: What we know. Gynecol Oncol 88:S8–10; discussion S11–3. Brinton LA, Gridley G, Persson I, Baron J, Bergqvist A. 1997. Cancer risk after a hospital discharge diagnosis of endometriosis. Am J Obstet Gynecol 176:572–579. Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ. 2003. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med 348:1625–1638. Casagrande JT, Louie EW, Pike MC, Roy S, Ross RK, Henderson BE. 1979. “Incessant ovulation” and ovarian cancer. Lancet 2:170–173. Chapman WB. 2001. Developments in the pathology of ovarian tumours. Curr Opin Obstet Gynecol 13:53–59. Chatenoud L, Tavani A, La Vecchia C, et al. 1998. Whole grain food intake and cancer risk. Int J Cancer 77:24–28. Chiaffarino F, Pelucchi C, Parazzini F, et al. 2001. Reproductive and hormonal factors and ovarian cancer. Ann Oncol 12:337–341. Cooper GS, Schildkraut JM, Whittemore AS, Marchbanks PA. 1999. Pregnancy recency and risk of ovarian cancer. Cancer Causes Control 10:397– 402. Cornelison TL, Natarajan N, Piver MS, Mettlin CJ. 1997. Tubal ligation and the risk of ovarian carcinoma. Cancer Detect Prev 21:1–6. Cottreau CM, Ness RB, Kriska AM. 2000. Physical activity and reduced risk of ovarian cancer. Obstet Gynecol 96:609–614. Coughlin SS, Giustozzi A, Smith SJ, Lee NC. 2000. A meta-analysis of estrogen replacement therapy and risk of epithelial ovarian cancer. J Clin Epidemiol 53:367–375. Coughlin SS, Hall IJ. 2002. Glutathione S-transferase polymorphisms and risk of ovarian cancer: A HuGE review. Genet Med 4:250–257. Cozen W, Peters R, Reichardt JK, et al. 2002. Galactose-1-phosphate uridyl transferase (GALT) genotype and phenotype, galactose consumption, and the risk of borderline and invasive ovarian cancer (United States). Cancer Causes Control 13:113–120. Cramer DW. 1999. Perineal talc exposure and subsequent epithelial ovarian cancer: A case-control study. Obstet Gynecol 94:160–161. Cramer DW, Greenberg ER, Titus-Ernstoff L, Liberman RF, Welch WR, Li E, Ng WG. 2000. A case-control study of galactose consumption and metab-

1023

olism in relation to ovarian cancer. Cancer Epidemiol Biomarkers Prev 9:95–101. Cramer DW, Harlow BL, Titus-Ernstoff L, Bohlke K, Welch WR, Greenberg ER. 1998. Over-the-counter analgesics and risk of ovarian cancer. Lancet 351:104–107. Cramer DW, Harlow BL, Willett WC, et al. 1989. Galactose consumption and metabolism in relation to the risk of ovarian cancer. Lancet 2:66–71. Cramer DW, Kuper H, Harlow BL, Titus-Ernstoff L. 2001. Carotenoids, antioxidants and ovarian cancer risk in pre- and postmenopausal women. Int J Cancer 94:128–134. Cramer DW, Welch WR. 1983. Determinants of ovarian cancer risk. II. Inferences regarding pathogenesis. J Natl Cancer Inst 71:717–721. Cramer DW, Welch WR, Cassells S, Scully RE. 1983. Mumps, menarche, menopause, and ovarian cancer. Am J Obstet Gynecol 147:1–6. Cramer DW, Welch WR, Hutchison GB, Willett W, Scully RE. 1984. Dietary animal fat in relation to ovarian cancer risk. Obstet Gynecol 63:833–838. Cramer DW, Xu H. 1995. Epidemiologic evidence for uterine growth factors in the pathogenesis of ovarian cancer. Ann Epidemiol 5:310–314. Crum CP. 1999. Ovaries. In: Cotran RS, Tucker C, eds. Pathologic Basis of Disease, Philadelphia: W.B. Saunders Company. Cuzick J, Bulstrode JC, Stratton I, Thomas BS, Bulbrook RD, Hayward JL. 1983. A prospective study of urinary androgen levels and ovarian cancer. Int J Cancer 32:723–726. Dal Maso L, Franceschi S, Negri E, et al. 2002. Body size indices at different ages and epithelial ovarian cancer risk. Eur J Cancer 38:1769–1774. Darby SC, Doll R, Gill SK, Smith PG. 1987. Long term mortality after a single treatment course with X-rays in patients treated for ankylosing spondylitis. Br J Cancer 55:179–190. De Vivo I, Huggins GS, Hankinson SE, et al. 2002. A functional polymorphism in the promoter of the progesterone receptor gene associated with endometrial cancer risk. PNAS 99:12263–12268. Edmondson RJ, Monaghan JM, Davies BR. 2002. The human ovarian surface epithelium is an androgen responsive tissue. Br J Cancer 86:879–885. Engle A, Muscat JE, Harris RE. 1991. Nutritional risk factors and ovarian cancer. Nutrition Cancer 15:239–247. Erzen M, Rakar S, Klancnik B, Syrjanen K, Klancar B. 2001. Endometriosisassociated ovarian carcinoma (EAOC): An entity distinct from other ovarian carcinomas as suggested by a nested case-control study. Gynecol Oncol 83:100–108. Fairfield KM, Hankinson SE, Rosner BA, Hunter DJ, Colditz GA, Willett WC. 2001. Risk of ovarian carcinoma and consumption of vitamins A, C, and E and specific carotenoids: A prospective analysis. Cancer 92:2318–2326. Fairfield KM, Hunter DJ, Colditz GA, et al. 2004. A prospective study of dietary lactose and ovarian cancer. Int J Cancer 110:271–277. Fairfield KM, Willett WC, Rosner BA, Manson JE, Speizer FE, Hankinson SE. 2002. Obesity, weight gain, and ovarian cancer. Obstet Gynecol 100: 288–296. Fathalla MF. 1971. Incessant ovulation—a factor in ovarian neoplasia? Lancet 2:163. Feeley KM, Wells M. 2001. Precursor lesions of ovarian epithelial malignancy. Histopathology 38:87–95. Fleischauer AT, Olson SH, Mignone L, Simonsen N, Caputo TA, Harlap S. 2001. Dietary antioxidants, supplements, and risk of epithelial ovarian cancer. Nutr Cancer 40:92–98. Franceschi S, La Vecchia C, Booth M, et al. 1991. Pooled analysis of 3 European case-control studies of ovarian cancer: II. Age at menarche and at menopause. Int J Cancer 49:57–60. Fredrickson TN. 1987. Ovarian tumors of the hen. Environ Health Perspect 73:35–51. Frei B. 1994. Reactive oxygen species and antioxidant vitamins: Mechanisms of action. Am J Med 97:5S–13S; discussion 22S–28S. Friis S, Nielsen GL, Mellemkjaer L, et al. 2002. Cancer risk in persons receiving prescriptions for Paracetamol: A Danish cohort study. Int J Cancer 97:96–101. Fung WL, Risch H, McLaughlin J, et al. 2003. The N314D polymorphism of galactose-1-phosphate uridyl transferase does not modify the risk of ovarian cancer. Cancer Epidemiol Biomarkers Prev 12:678–680. Garg PP, Kerlikowske K, Subak L, Grady D. 1998. Hormone replacement therapy and the risk of epithelial ovarian carcinoma: a meta-analysis. Obstet Gynecol 92:472–479. Garner EI, Stokes EE, Berkowitz RS, Mok SC, Cramer DW. 2002. Polymorphisms of the estrogen-metabolizing genes CYP17 and catechol-Omethyltransferase and risk of epithelial ovarian cancer. Cancer Res 62:3058–3062. Germani D, Belli S, Bruno C, et al. 1999. Cohort mortality study of women compensated for asbestosis in Italy. Am J Ind Med 36:129–134. Gertig DM, Hunter DJ, Cramer DW, et al. 2000. prospective study of talc use and ovarian cancer. J Natl Cancer Inst 92:249–252.

1024

PART IV: CANCER BY TISSUE OF ORIGIN

Giudice LC, Cataldo NA, van Dessel HJ, Yap OW, Chandrasekher YA. 1996. Growth factors in normal ovarian follicle development. Semin Reprod Endocrinol 14:179–196. Godwin AK, Perez RP, Johnson SW, Hamaguchi K, Hamilton TC. 1992. Growth regulation of ovarian cancer. Hematol Oncol Clin North Am 6:829–841. Goodman MT, McDuffie K, Kolonel LN, et al. 2001. Case-control study of ovarian cancer and polymorphisms in genes involved in catecholestrogen formation and metabolism. Cancer Epidemiol Biomarkers Prev 10: 209–216. Goodman MT, Tung KH. 2003. Alcohol consumption and the risk of borderline and invasive ovarian cancer. Obstet Gynecol 101:1221–1228. Goodman MT, Wu AH, Tung KH, et al. 2002a. Association of galactose-1phospate uridyltransferase activity and N314D genotype with the risk of ovarian cancer. Am J Epidemiol 156:693–701. Goodman MT, Wu AH, Tung KH, et al. 2002b. Association of dairy products, lactose, and calcium with the risk of ovarian cancer. Am J Epidemiol 156:148–157. Green A, Purdie D, Bain C, et al. 1997. Tubal sterilisation, hysterectomy and decreased risk of ovarian cancer. Survey of Women’s Health Study Group. Int J Cancer 71:948–951. Green A, Purdie D, Bain C, Siskind V, Webb PM. 2001. Cigarette smoking and risk of epithelial ovarian cancer (Australia). Cancer Causes Control 12:713–719. Haenszel W, Kurihara M. 1968. Studies of Japanese migrants. I. Mortality from cancer and other diseases among Japanese in the United States. J Natl Cancer Inst 40:43–68. Hankinson SE, Colditz GA, Hunter DJ, Spencer TL, Rosner B, Stampfer MJ. 1992. A quantitative assessment of oral contraceptive use and risk of ovarian cancer. Obstet Gynecol 80:708–714. Hankinson SE, Colditz GA, Hunter DJ, et al. 1995. A prospective study of reproductive factors and risk of epithelial ovarian cancer. Cancer 76:284–290. Hankinson SE, Hunter DJ, Colditz GA, et al. 1993. Tubal ligation, hysterectomy, and risk of ovarian cancer. A prospective study. JAMA 270:2813–2818. Hankinson SE, Willett WC, Manson JE, et al. 1998. Plasma sex steroid hormone levels and risk of breast cancer in postmenopausal women. J Natl Cancer Inst 90:1292–1299. Harlass FE, Plymate SR, Fariss BL, Belts RP. 1984. Weight loss is associated with correction of gonadotropin and sex steroid abnormalities in the obese anovulatory female. Fertil Steril 42:649–652. Harlow BL, Hartge PA. 1995. A review of perineal talc exposure and risk of ovarian cancer. Regul Toxicol Pharmacol 21:254–260. Harris R, Whittemore AS, Itnyre J. 1992. Characteristics relating to ovarian cancer risk: collaborative analysis of 12 US case-control studies. III. Epithelial tumors of low malignant potential in white women. Collaborative Ovarian Cancer Group. Am J Epidemiol 136:1204–1211. Hartge P, Schiffman MH, Hoover R, McGowan L, Lesher L, Norris HJ. 1989. A case-control study of epithelial ovarian cancer. Am J Obstet Gynecol 161:10–16. Heller DS, Gordon RE, Katz N. 1999. Correlation of asbestos fiber burdens in fallopian tubes and ovarian tissue. Am J Obstet Gynecol 181:346–347. Helzlsouer KJ, Alberg AJ, Gordon GB, et al. 1995. Serum gonadotropins and steroid hormones and the development of ovarian cancer. JAMA 274:1926–1930. Herrinton LJ, Weiss NS, Beresford SA, et al. 1995. Lactose and galactose intake and metabolism in relation to the risk of epithelial ovarian cancer. Am J Epidemiol 141:407–416. Hiroshima International Council for Medical Care of the Radiation-Exposed, Shigematsu I, Ito C, Kamada N, Akiyama M, Sasaki H, eds., Harrison B translator. 1995. Effects of a-bomb radiation on the human body. Chur, Switzerland: Harwood Academic Publishers; Tokyo: Bunkodo Co., Ltd. Huncharek M, Kupelnick B. 2001. Dietary fat intake and risk of epithelial ovarian cancer: a meta-analysis of 6,689 subjects from 8 observational studies. Nutr Cancer 40:87–91. IARC. 1988. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans: Alcohol Drinking. Vol. 44. Lyon: International Agency for Research on Cancer. IARC. 1991. IARC Monographs on the Evaluation of Carcinogenoc Risks to Humans: Coffee, Tea, Mate, Methylxanthines and Methlyglyoxal. Vol. 51. Lyon: International Agency for Research on Cancer. IARC. 1997. IARC Scientific Publications: Social Inequalities and Cancer. Kogevinas M, Pearce N, Susser M, Boffetta P, eds. Vol. 138. Lyon: International Agency for Research on Cancer. IARC. 1999. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans: Hormonal Contraception and Postmenopausal Hormonal Therapy. Vol. 72. Lyon: International Agency for Research on Cancer.

IARC. 2002a. Cancer Incidence in Five Continents Vol. VIII. Parkin DM, Whelan SL, Ferlay J, Teppo L, Thomas DB, eds. Lyon: International Agency for Research on Cancer. IARC. 2002b. IARC Handbooks of Cancer Prevention: Weight Control and Physical Activity. Vainio H, Bianchini F, eds. Vol. 6. Lyon: International Agency for Research on Cancer. Issaq HJ, Veenstra TD, Conrads TP, Felschow D. 2002. The SELDI-TOF MS approach to proteomics: protein profiling and biomarker identification. Biochem Biophys Res Commun 292:587–592. Kato I, Tominaga S, Terao C. 1989. Alcohol Consumption and cancers of hormone-related organs in females. Jpn J Clin Oncol 19:202–207. Koch M, Jenkins H, Gaedke H. 1988. Risk factors of ovarian cancer of epithelial origin: a case control study. Cancer Detect Prev 13:131–136. Koch M, Starreveld AA, Hill GB, Jenkins H. 1984. The effect of tubal ligation on the incidence of epithelial cancer of the ovary. Cancer Detect Prev 7:241–245. Kreiger N, Sloan M, Cotterchio M, Parsons P. 1997. Surgical procedures associated with risk of ovarian cancer. Int J Epidemiol 26:710–715. Kuhnz W, Sostarek D, Gansau C, Louton T, Mahler M. 1991. Single and multiple administration of a new triphasic oral contraceptive to women: Pharmacokinetics of ethinyl estradiol and free and total testosterone levels in serum. Am J Obstet Gynecol 165:596–602. Kuper H, Cramer DW, Titus-Ernstoff L. 2002. Risk of ovarian cancer in the United States in relation to anthropometric measures: does the association depend on menopausal status? Cancer Causes Control 13:455–463. Kuper H, Titus-Ernstoff L, Harlow BL, Cramer DW. 2000. Population based study of coffee, alcohol and tobacco use and risk of ovarian cancer. Int J Cancer 88:313–318. Kushi LH, Mink PJ, Folsom AR, et al. 1999. Prospective study of diet and ovarian cancer. Am J Epidemiol 149:21–31. Kvale G, Heuch I, Nilssen S. 1992. Is the risk of ovarian cancer related to age at menarche and age at menopause? Int J Cancer 51:333–334. La Vecchia C, Decarli A, Negri E, et al. 1987. Dietary factors and the risk of epithelial ovarian cancer. J Natl Cancer Inst 79:663–669. La Vecchia C, Negri E, Franceschi S, Parazzini F, Gentile A, Fasoli M. 1992. Alcohol and epithelial ovarian cancer. J Clin Epidemiol 45:1025–1030. Lacey JV Jr, Mink PJ, Lubin JH, et al. 2002. Menopausal hormone replacement therapy and risk of ovarian cancer. [comment]. JAMA 288:334– 341. Lagiou P, Ye W, Wedren S, et al. 2001. Incidence of ovarian cancer among alcoholic women: A cohort study in Sweden. Int J Cancer 91:264– 266. Lancaster JM, Berchuck A, Carney ME, Wiseman R, Taylor JA. 1998. Progesterone receptor gene polymorphism and risk for breast and ovarian cancer. Br J Cancer 78:277. Lancaster JM, Wenham RM, Halabi S, et al. 2003. No relationship between ovarian cancer risk and progesterone receptor gene polymorphism in a population-based, case-control study in North Carolina. Cancer Epidemiol Biomarkers Prev 12:226–227. Larsson SC, Bergkvist L, Wolk A. 2004. Milk and lactose intakes and ovarian cancer risk in the Swedish Mammography Cohort. Am J Clin Nutr 80:1353–1357. Larsson SC, Orsini N, Wolk A. 2005. Milk, milk products and lactose intake and ovarian cancer risk: A meta-analysis of epidemiological studies. Int J Cancer 118:431–441. LeRoith D, Baserga R, Helman L, Roberts Jr CT. 1995. Insulin-like growth factors and cancer. Ann Intern Med 122:54–59. Liu Y, Ganesan TS. 2002. Tumour suppressor genes in sporadic epithelial ovarian cancer. Reproduction 123:341–353. Loft A, Lidegaard O, Tabor A. 1997. Incidence of ovarian cancer after hysterectomy: a nationwide controlled follow up. Br J Obstet Gynaecol 104:1296–1301. Lukanova A, Lundin E, Akhmedkhanov A, et al. 2003. Circulating levels of sex steroid hormones and risk of ovarian cancer. Int J Cancer 104: 636–642. Lukanova A, Lundin E, Toniolo P, et al. 2002. Circulating levels of insulin-like growth factor-I and risk of ovarian cancer. Int J Cancer 101:549–554. Manolitsas TP, Englefield P, Eccles DM, Campbell IG. 1997. No association of a 306-bp insertion polymorphism in the progesterone receptor gene with ovarian and breast cancer. Br J Cancer 75:1398–1399. Marchbanks PA, Wilson H, Bastos E, Cramer DW, Schildkraut JM, Peterson HB. 2000. Cigarette smoking and epithelial ovarian cancer by histologic type. Obstet Gynecol 95:255–260. McCann SE, Freudenheim JL, Marshall JR, Graham S. 2003. Risk of human ovarian cancer is related to dietary intake of selected nutrients, phytochemicals and food groups. J Nutr 133:1937–1942. McKenna NJ, Kieback DG, Carney DN, Fanning M, McLinden J, Headon DR. 1995. A germline TaqI restriction fragment length polymorphism in the

Ovarian Cancer progesterone receptor gene in ovarian carcinoma. Br J Cancer 71:451– 455. Meier CR, Schmitz S, Jick H. 2002. Association between acetaminophen or nonsteroidal antiinflammatory drugs and risk of developing ovarian, breast, or colon cancer. Pharmacotherapy 22:303–309. Meloni GF, Colombo C, La Vecchia C, et al. 1999. Lactose absorption in patients with ovarian cancer. Am J Epidemiol 150:183–186. Menon U, Jacobs I. 2000. Ovarian cancer screening in the general population. Ultrasound Obstet Gynecol 15:350–353. Mink PJ, Folsom AR, Sellers TA, Kushi LH. 1996. Physical activity, waist-tohip ratio, and other risk factors for ovarian cancer: a follow-up study of older women. Epidemiology 7:38–45. Mink PJ, Sherman ME, Devesa SS. 2002. Incidence patterns of invasive and borderline ovarian tumors among white women and black women in the United States. Results from the SEER Program, 1978–1998. Cancer. 95:2380–2389. Miracle-McMahill HL, Calle EE, Kosinski AS, et al. 1997. Tubal ligation and fatal ovarian cancer in a large prospective cohort study. Am J Epidemiol 145:349–357. Modan B, Hartge P, Hirsh-Yechezkel G, et al. 2001. Parity, oral contraceptives, and the risk of ovarian cancer among carriers and noncarriers of a BRCA1 or BRCA2 mutation. N Engl J Med 345:235–240. Modugno F, Ness RB, Cottreau CM. 2002. Cigarette smoking and the risk of mucinous and nonmucinous epithelial ovarian cancer. Epidemiology 13:467–471. Modugno F, Ness RB, Wheeler JE. 2001. Reproductive risk factors for epithelial ovarian cancer according to histologic type and invasiveness. Ann Epidemiol 11:568–574. Mori M, Harabuchi I, Miyake H, Casagrande JT, Henderson BE, Ross RK. 1988. Reproductive, genetic, and dietary risk factors for ovarian cancer. Am J Epidemiol 128:771–777. Moysich KB, Mettlin C, Piver MS, Natarajan N, Menezes RJ, Swede H. 2001. Regular use of analgesic drugs and ovarian cancer risk. Cancer Epidemiol Biomarkers Prev 10:903–906. Muti P, Quattrin T, Grant BJ, et al. 2002. Fasting glucose is a risk factor for breast cancer: a prospective study. Cancer Epidemiol Biomarkers Prev 11:1361–1368. Nandakumar A, Anantha N, Dhar M, et al. 1995. A case-control investigation on cancer of the ovary in Bangalore, India. Int J Cancer 63:361– 365. Narod SA, Sun P, Ghadirian P, et al. 2001. Tubal ligation and risk of ovarian cancer in carriers of BRCA1 or BRCA2 mutations: a case-control study. Lancet 357:1467–1470. Narod SA, Sun P, Risch HA. 2001. Ovarian cancer, oral contraceptives, and BRCA mutations. N Engl J Med 345:1706–1707. Negri E, Tzonou A, Beral V, et al. 1999. Hormonal therapy for menopause and ovarian cancer in a collaborative re-analysis of European studies. Internat J Cancer 80:848–851. Ness RB. 2003. Endometriosis and ovarian cancer: thoughts on shared pathophysiology. Am J Obstet Gynecol 189:280–294. Ness RB, Cottreau C. 1999. Possible role of ovarian epithelial inflammation in ovarian cancer. [comment]. J Natl Cancer Instit 91:1459–1467. Ness RB, Cramer DW, Goodman MT, et al. 2002. Infertility, fertility drugs, and ovarian cancer: a pooled analysis of case-control studies. Am J Epidemiol 155:217–224. Ness RB, Grisso JA, Klapper J, et al. 2000. Risk of ovarian cancer in relation to estrogen and progestin dose and use characteristics of oral contraceptives. SHARE Study Group. Steroid hormones and reproductions. Am J Epidemiol 152:233–241. Ness RB, Grisso JA, Vergona R, Klapper J, Morgan M, Wheeler JE. 2001. Oral contraceptives, other methods of contraception, and risk reduction for ovarian cancer. Epidemiology 12:307–312. Parazzini F, La Vecchia C, Negri E, Moroni S, dal Pino D, Fedele L. 1996. Pelvic inflammatory disease and risk of ovarian cancer. Cancer Epidemiol Biomarkers Prev 5:667–669. Parkin DM, Bray FI, Devesa SS. 2001. Cancer burden in the year 2000. The global picture. Eur J Cancer 37(Suppl 8):S4–S66. Pasquali R, Antenucci D, Casimirri F, et al. 1989. Clinical and hormonal characteristics of obese amenorrheic hyperandrogenic women before and after weight loss. J Clin Endocrinol Metab 68:173–179. Petricoin EF, Ardekani AM, Hitt BA, et al. 2002. Use of proteomic patterns in serum to identify ovarian cancer. Lancet 359:572–577. Polychronopoulou A, Tzonou A, Hsieh CC, et al. 1993. Reproductive variables, tobacco, ethanol, coffee and somatometry as risk factors for ovarian cancer. Int J Cancer 55:402–407. Prorok PC, Andriole GL, Bresalier RS, et al. 2000. Design of the prostate, lung, colorectal and ovarian (PLCO) cancer screening trial. Control Clin Trials 21(6 Suppl):273S–309S.

1025

Pukkala E, Poskiparta M, Apter D, Vihko V. 1993. Life-long physical activity and cancer risk among Finnish female teachers. Eur J Cancer Prev 2:369–376. Purdie D, Green A, Bain C, et al. 1995. Reproductive and other factors and risk of epithelial ovarian cancer: an Australian case-control study. Survey of Women’s Health Study Group. Int J Cancer 62:678–684. Purdie DM, Bain CJ, Siskind V, et al. 1999. Hormone replacement therapy and risk of epithelial ovarian cancer. Br J Cancer 81:559–563. Purdie DM, Bain CJ, Siskind V, Webb PM, Green AC. 2003. Ovulation and risk of epithelial ovarian cancer. Int J Cancer 104:228–232. Purdie DM, Bain CJ, Webb PM, Whiteman DC, Pirozzo S, Green AC. 2001. Body size and ovarian cancer: case-control study and systematic review (Australia). Cancer Causes Control 12:855–863. Radisavljevic SV. 1977. The pathogenesis of ovarian inclusion cysts and cystomas. Obstet Gynecol 49:424–429. Ries LAG, Eisner MP, Kosary CL, Hankey BF, Miller BA, Clegg L, Mariotto A, Fay MP, Feuer EJ, Edwards BK, eds. 2003. SEER Cancer Statistics Review, 1975–2000. National Cancer Institute, Bethesda, MD, http://seer.cancer.gov/csr/1975_2000/. Riley JC, Behrman HR. 1991. Oxygen radicals and reactive oxygen species in reproduction. Proc Soc Exp Biol Med 198:781–791. Riman T, Dickman PW, Nilsson S, et al. 2002. Hormone replacement therapy and the risk of invasive epithelial ovarian cancer in Swedish women. J Natl Cancer Inst 94:497–504. Risch HA. 1996. Estrogen replacement therapy and risk of epithelial ovarian cancer. Gynecol Oncol 63:254–257. Risch HA. 1998. Hormonal etiology of epithelial ovarian cancer, with a hypothesis concerning the role of androgens and progesterone. J Natl Cancer Inst 90:1774–1786. Risch HA, Jain M, Marrett LD, Howe GR. 1994a. Dietary fat intake and risk of epithelial ovarian cancer. J Natl Cancer Inst 86:1409–1415. Risch HA, Jain M, Marrett LD, Howe GR. 1994b. Dietary lactose intake, lactose intolerance, and the risk of epithelial ovarian cancer in southern Ontario (Canada). Cancer Causes Control 5:540–548. Rodriguez C, Calle EE, Fakhrabadi-Shokoohi D, Jacobs EJ, Thun MJ. 2002. Body mass index, height, and the risk of ovarian cancer mortality in a prospective cohort of postmenopausal women. Cancer Epidemiol Biomarkers Prev 11:822–828. Rodriguez C, Henley SJ, Calle EE, Thun MJ. 1998. Paracetamol and risk of ovarian cancer mortality in a prospective study of women in the USA. Lancet 352:1354–1355. Rodriguez C, Patel AV, Calle EE, Jacob EJ, Thun MJ. 2001. Estrogen replacement therapy and ovarian cancer mortality in a large prospective study of US women. JAMA 285:1460–1465. Rodriguez GC, Walmer DK, Cline M, et al. 1998. Effect of progestin on the ovarian epithelium of macaques: cancer prevention through apoptosis? J Soc Gynecol Investig 5:271–276. Rosenberg L, Palmer JE, Rao RS, et al. 2000. A case-control study of analgesic use and ovarian cancer. Cancer Epidemiol Biomarkers Prev 9:933– 937. Rosenberg L, Palmer JR, Zauber AG, et al. 1994. A case-control study of oral contraceptive use and invasive epithelial ovarian cancer. Am J Epidemiol 139:654–661. Rosenblatt KA, Thomas DB. 1993. Lactation and the risk of epithelial ovarian cancer. The WHO Collaborative Study of Neoplasia and Steroid Contraceptives. Int J Epidemiol 22:192–197. Rosenblatt KA, Thomas DB. 1996. Reduced risk of ovarian cancer in women with a tubal ligation or hysterectomy. The World Health Organization Collaborative Study of Neoplasia and Steroid Contraceptives. Cancer Epidemiol Biomarkers Prev 5:933–935. Rossing MA, Daling JR, Weiss NS, Moore DE, Self SG. 1994. Ovarian tumors in a cohort of infertile women. N Engl J Med 331:771–776. Runnebaum IB, Wang-Gohrke S, Vesprini D, et al. 2001. Progesterone receptor variant increases ovarian cancer risk in BRCA1 and BRCA2 mutation carriers who were never exposed to oral contraceptives. Pharmacogenetics 11:635–638. Schildkraut JM, Bastos E, Berchuck A. 1997. Relationship between lifetime ovulatory cycles and overexpression of mutant p53 in epithelial ovarian cancer. J Natl Cancer Inst 89:932–938. Schildkrant JM, Calingaert B, Marchbanks PA, Moorman PG, Rodriguez GC. 2002. Impact of progestin and estrogen potency in oral contraceptives on ovarian cancer risk. J Natl Cancer Inst 94:32–38. Schildkraut JM, Cooper GS, Halabi S, Calingaert B, Hartge P, Whittemore AS. 2001. Age at natural menopause and the risk of epithelial ovarian cancer. Obstet Gynecol 98:85–90. Schildkraut JM, Schwingl PJ, Bastos E, Evanoff A, Hughes C. 1996. Epithelial ovarian cancer risk among women with polycystic ovary syndrome. Obstet Gynecol 88:554–559.

1026

PART IV: CANCER BY TISSUE OF ORIGIN

Schouten LJ, Goldbohm RA, van den Brandt PA. 2003. Height, weight, weight change, and ovarian cancer risk in the Netherlands cohort study on diet and cancer. Am J Epidemiol 157:424–433. Scully RE. 1995. Pathology of ovarian cancer precursors. J Cell Biochem (Suppl) 23:208–218. Seidman JD, Kurman RJ, Ronnett BM. 2003. Primary and metastatic mucinous adenocarcinomas in the ovaries: incidence in routine practice with a new approach to improve intraoperative diagnosis. Am J Surg Pathol 27: 985–993. Shu XO, Brinton LA, Gao YT, Yuan JM. 1989. Population-based case-control study of ovarian cancer in Shanghai. Cancer Res 49:3670–3674. Siskind V, Green A, Bain C, Purdie D. 1997. Breastfeeding, menopause, and epithelial ovarian cancer. Epidemiology 8:188–191. Siskind V, Green A, Bain C, Purdie D. 2000. Beyond ovulation: oral contraceptives and epithelial ovarian cancer. Epidemiology 11:106–110. Snowdon DA. 1985. Diet and ovarian cancer. JAMA 254:356–357. Spurdle AB, Chen X, Abbazadegan M, et al. 2000. CYP17 promotor polymorphism and ovarian cancer risk. Int J Cancer 86:436–439. Stensvold I, Jacobsen BK. 1994. Coffee and cancer: a prospective study of 43,000 Norwegian men and women. Cancer Causes Control 5:401–408. Stratton JF, Pharoah P, Smith SK, Easton D, Ponder BA. 1998. A systematic review and meta-analysis of family history and risk of ovarian cancer. Br J Obstet Gynaecol 105:493–499. Swiersz LM. 2002. Role of endometriosis in cancer and tumor development. Ann N Y Acad Sci 955:281–292; discussion 293–295, 396–406. Syed V, Ulinski G, Mok SC, Yiu GK, Ho SM. 2001. Expression of gonadotropin receptor and growth responses to key reproductive hormones in normal and malignant human ovarian surface epithelial cells. Cancer Res 61:6768–6776. Takahashi H, Chiu HC, Bandera CA, et al. 1996. Mutations of the BRCA2 gene in ovarian carcinomas. Cancer Res 56:2738–2741. Tavani A, Gallus S, Dal Maso L, et al. 2001. Coffee and alcohol intake and risk of ovarian cancer: an Italian case-control study. Nutr Cancer 39:29–34. Tavani A, Gallus S, La Vecchia C, et al. 2001. Physical activity and risk of ovarian cancer: an Italian case-control study. Int J Cancer 91:407– 411. Tavani A, Gallus S, La Vecchia C, Conti E, Montella M, Franceschi S. 2000. Aspirin and ovarian cancer: an Italian case-control study. Ann Oncol 11:1171–1173. Tavani A, Ricci E, La Vecchia C, et al. 2000. Influence of menstrual and reproductive factors on ovarian cancer risk in women with and without family history of breast or ovarian cancer. Int J Epidemiol 29:799–802. Tavassoli M, Steingrimsdottir H, Pierce E, et al. 1996. Loss of heterozygosity on chromosome 5q in ovarian cancer is frequently accompanied by TP53 mutation and identifies a tumour suppressor gene locus at 5q13.1–21. Br J Cancer 74:115–119. Terry PD, Miller AB, Jones JG, Rohan TE. 2003. Cigarette smoking and the risk of invasive epithelial ovarian cancer in a prospective cohort study. Eur J Cancer 39:1157–1164. Titus-Ernstoff L, Perez K, Cramer DW, Harlow BL, Baron JA, Greenberg ER. 2001. Menstrual and reproductive factors in relation to ovarian cancer risk. Br J Cancer 84:714–721. Tokuoka S, Kawai K, Shimizu Y, et al. 1987. Malignant and benign ovarian neoplasms among atomic bomb survivors, Hiroshima and Nagasaki, 1950–80. J Nail Cancer Inst 79:47–57. Tong D, Fabjani G, Heinze G, Obermair A, Leodolter S, Zeillinger R. 2001. Analysis of the human progesterone receptor gene polymorphism progins in Austrian ovarian carcinoma patients. Int J Cancer 95:394–397. Tong D, Fabjani G, Heinze G, Zeillinger R. 2002. No association between the human progesterone receptor gene polymorphism PROGINS and risk for ovarian carcinoma in Austrian women. Int J Cancer 101:203. Toniolo P, Bruning PF, Akhmedkhanov A, et al. 2000. Serum insulin-like growth factor-I and breast cancer. Int J Cancer 88:828–832.

Tzonou A, Day NE, Trichopoulos D, et al. 1984. The epidemiology of ovarian cancer in Greece: a case-control study. Eur J Cancer Clin Oncol 20: 1045–1052. Tzonou A, Hsieh CC, Polychronopoulou A, et al. 1993. Diet and ovarian cancer: a case-control study in Greece. Int J Cancer 55:411–414. Tzonou A, Polychronopoulou A, Hsieh CC, Rebelakos A, Karakatsani A, Trichopoulos D. 1993. Hair dyes, analgesics, tranquilizers and perineal talc application as risk factors for ovarian cancer. Int J Cancer 55: 408–410. U.S. Preventive Services Task Force. 1996. Guide to Clinical Preventive Services. 2nd ed. Washington, DC: U.S. Department of Health and Human Services, Office of Disease Prevention and Health Promotion. Urban N. 2003. Specific keynote: ovarian cancer risk assessment and the potential for early detection. Gynecol Oncol 88:S75–S79; discussion S80–S83. van der Vange N, Blankenstein MA, Kloosterboer HJ, Haspels AA, Thijssen JH. 1990. Effects of seven low-dose combined oral contraceptives on sex hormone binding globulin, corticosteroid binding globulin, total and free testosterone. Contraception 41:345–352. Vasama-Neuvonen K, Pukkala E, Paakkulainen H, et al. 1999. Ovarian cancer and occupational exposures in Finland. Am J Ind Med 36:83–89. Wang JX, Boice JD Jr, Li BX, Zhang JY, Fraumeni JF, Jr. 1988. Cancer among medical diagnostic x-ray workers in China. J Natl Cancer Inst 80: 344–350. Webb PM, Green A, Cummings MC, Purdie DM, Walsh MD, Chenevix-Trench G. 1998. Relationship between number of ovulatory cycles and accumulation of mutant p53 in epithelial ovarian cancer. J Natl Cancer Inst 90:1729–1734. Weinberger SR, Dalmasso EA, Fung ET. 2002. Current achievements using ProteinChip Array technology. Curr Opin Chem Biol 6:86–91. Weiss NS, Harlow BL. 1986. Why does hysterectomy without bilateral oophorectomy influence the subsequent incidence of ovarian cancer? Am J Epidemiol 124:856–858. Weiss NS, Lyon JL, Krishnamurthy S, Dietert SE, Liff JM, Daling JR. 1982. Noncontraceptive estrogen use and the occurrence of ovarian cancer. J Natl Cancer Inst 68:95–98. Werness BA, Eltabbakh GH. 2001. Familial ovarian cancer and early ovarian cancer: biologic, pathologic, and clinical features. Int J Gynecol Pathol 20:48–63. Whiteman DC, Siskind V, Purdie DM, Green AC. 2003. Timing of pregnancy and the risk of epithelial ovarian cancer. Cancer Epidemiol Biomarkers Prev 12:42–46. Whittemore AS, Harris R, Itnyre J. 1992. Characteristics relating to ovarian cancer risk: collaborative analysis of 12 US case-control studies. II. Invasive epithelial ovarian cancers in white women. Collaborative Ovarian Cancer Group. Am J Epidemiol 136:1184–1203. Whittemore AS, Harris R, Itnyre J, Halpern J. 1992. Characteristics relating to ovarian cancer risk: collaborative analysis of 12 US case-control studies. I. Methods. Collaborative Ovarian Cancer Group. Am J Epidemiol 136:1175–1183. Whittemore AS, Wu ML, Paffenbarger RS, et al. 1988. Personal and environmental characteristics related to epithelial ovarian cancer. II. Exposures to talcum powder, tobacco, alcohol, and coffee. Am J Epidemiol 128: 1228–1240. Yager JD, Liehr JG. 1996. Molecular mechanisms of estrogen carcinogenesis. Annu Rev Pharmacol Toxicol 36:203–232. Yee D, Morales FR, Hamilton TC, Von Hoff DD. 1991. Expression of insulinlike growth factor I, its binding proteins, and its receptor in ovarian cancer. Cancer Res 51:5107–5112. Zhang M, Yang ZY, Binns CW, Lee AH. 2002. Diet and ovarian cancer risk: a case-control study in China. Br J Cancer 86:712–715. Zheng W, Shu XO, McLaughlin JK, Chow WH, Gao YT, Blot WJ. 1993. Occupational physical activity and the incidence of cancer of the breast, corpus uteri, and ovary in Shanghai. Cancer 71:3620–3624.

53

Endometrial Cancer LINDA S. COOK, NOEL S. WEISS, JENNIFER A. DOHERTY, AND CHU CHEN

E

ndometrial cancer is a relatively common gynecologic cancer and diagnosis generally occurs after abnormal uterine bleeding or spotting (Burke et al., 1996). The overall five-year relative survival following diagnosis is relatively high, varying from roughly 85% in the United States (Pisani et al., 1999; SEER, 2002), to 73% in China, 66% in India, and 63% in Eastern Europe (Pisani et al., 1999). Five-year relative survival is approximately 96% to 100% for women with tumors confined to the uterine body (i.e., Stage I) or highly differentiated tumors (i.e., low grade), but falls to 26% to 47% with tumors that have spread beyond the pelvis (i.e., Stage IV) or are poorly differentiated (i.e., high grade) (Kodama et al., 1996; SEER, 2002). Exposure of the endometrium to high circulating levels of estrogens increases the likelihood of developing this disease. In contrast, the presence of progestogens (both endogenous and exogenous) can mitigate against the occurrence of endometrial cancer. The actions of many of the other known or suspected factors that alter endometrial cancer risk, such as obesity, reproductive characteristics, certain medical conditions, and cigarette smoking, may be explained at least in part by their influence on estrogen and progestogen activity.

PATTERNS OF INCIDENCE AND MORTALITY The body (corpus) of the uterus contains several different types of tissue: the endometrium (the inner mucosal layer); the myometrium (the thick, middle muscular layer); and, the serosa (the thin external coat). The cervix is the lower portion of the uterus that lies below the uterine body. Especially in the past, the incidence of uterine corpus cancer was presented without further delineation of the location of the tumor within the uterine body. However, “corpus cancer” is a relatively good proxy for endometrial cancer; among African-Americans, 75% of cancers of the corpus uteri are endometrial in nature, and among White-Americans, the figure is more than 90% (SEER, 2002). Errors in documenting mortality from endometrial cancer have resulted from the practice of listing deaths from cancers of the uterine corpus or cervix simply as deaths from uterine cancer, not otherwise specified (NOS). Endometrial cancer mortality in any given year is probably best estimated by a combination of deaths from cancer of the corpus uteri and deaths from uterine cancer NOS, because only a small fraction of the latter is comprised of deaths from cervical cancer. In the United States, this fraction has fallen over time, due to fewer cervical cancers included in the uterine cancer, NOS category (Szekely et al., 1978) as well as to true decreases in cervical cancer mortality (SEER, 2002). Further complicating the interpretation of incidence and mortality rates of this disease is the lack of correction made in reports from cancer registries and vital records agencies for the percentage of women with hysterectomies who are no longer at risk for developing endometrial cancer. Nonetheless, even with these caveats, we can draw some conclusions about patterns and changes in endometrial cancer incidence and mortality.

Time Rapid changes have occurred in the incidence of endometrial cancer over the past five decades in the United States. The incidence began to rise in the 1960s and reached a peak in the mid-1970s (SEER, 2002; Weiss et al., 1976). The increase in incidence was experienced pri-

marily by postmenopausal women and was generally greater in the western than in the eastern part of the country (Table 53–1) (Department of Health, California, 1978; Marrett et al., 1978; SEER, 2002). The increase shown in the table is actually an underestimate of the true increase; the rate of hysterectomy for reasons other than cancer rose rapidly during the same time period and the rates presented have not been corrected for this change (Dicker et al., 1982; Howe, 1984; Koepsell et al., 1980; Lyon, Gardner, 1977). Following the peak in the mid-1970s, the incidence of endometrial cancer steadily declined in the United States until the 1990s (Table 53–1, 53–2) (Muir et al., 1987; Parkin et al., 1992; Parkin et al., 1997; SEER, 2002; Waterhouse et al., 1976; Waterhouse et al., 1982). In contrast to these changes, the incidence during the same time period outside of North America generally remained stable or has increased (Table 53–2). Since 1990, rates in the United States have generally remained stable (Tables 53–1 and 53–2) and little change has occurred in hysterectomy rates (Keshavarz et al., 2002). The increased incidence in the United States during the 1970s does not appear to be an artifact of changes in diagnostic criteria or classifying estrogen-induced hyperplasia as cancer due to similarities in morphology (Gordon, 1994). Based on histologic reviews of reported endometrial cancers that were conducted using conservative criteria to identify cancer, the incidence of endometrial cancer during the mid1970s was unequivocally greater than that of earlier years (Gordon et al., 1977; Szekely et al., 1978). The increased incidence in the 1960s to 1970s and the ensuing decline in rates do, however, parallel patterns of postmenopausal unopposed estrogen use (Austin, 1982). Replacement estrogens were introduced into medical practice in the 1930s, but they were not widely taken by postmenopausal women in the United States until the 1960s. Estrogen use was greater in the western United States than in other regions (Table 53–1) (Jick et al., 1980) and outside of North America use was generally uncommon (Doll et al., 1976). With increasing evidence of an association between estrogens and endometrial cancer, the U.S. Food and Drug Administration issued a warning to physicians in 1976 (Food and Drug Administration, 1976). A steady decline in the proportion of American women who used unopposed estrogens ensued, initially due to a reduction in use of postmenopausal hormone use altogether, and later to an increase in use of combined estrogen-progestogen hormone replacement therapy (Austin, 1982; Gruber, 1986; Kennedy et al., 1985; Ross et al., 1988; Standeven et al., 1986; Wysowski et al., 1995). In the United States, age-standardized mortality due to corpus cancer and uterine cancer NOS decreased by 60% between 1950 (8.9 per 100,000) and 1985 (3.6 per 100,000) (Division of Cancer Prevention and Control, National Cancer Institute, 1988) and has remained relatively stable since that time (SEER, 2002). The magnitude of this decrease is clearly an over-estimate of any true decrease in corpus cancer mortality; over the same time period, cervical cancer mortality decreased by 75% (Division of Cancer Prevention and Control, National Cancer Institute, 1988) and the hysterectomy rate rose by as much as 60% (U.S. Department of Health and Human Services, 1981; Koepsell et al., 1980; Lyon, 1977). After adjustment for the declining proportion of women with an intact uterus and the contribution of cervical cancer deaths included in uterine cancer NOS mortality, an estimated 14% decrease in corpus cancer mortality occurred between 1960 and 1970 in the United States (Weiss, 1978a). Likewise, age-standardized mortality rates from cancer of the body of

1027

1028

PART IV: CANCER BY TISSUE OF ORIGIN the uterus and cancer of uterus, part unspecified, also show a downward trend between 1955 and 1990 in England and Wales, Hungary, Italy, Japan, France, Canada, Norway, Australia, and Yugoslavia (Mant, Vessey, 1994). Adjustment for the rising hysterectomy rate (Coulter et al., 1988) diminished, but did not eliminate, the overall downward trend in morality in England and Wales (Villard, Murphy, 1990).

a

Table 53–1. Annual Incidence of Invasive Carcinoma of the Uterine Corpus by Age: Connecticut and Alameda County, California, 1960–1994 Connecticut

Alameda Countyb

Age (years)

Age (years)

Time period

30–49

50–69

30–49

50–69

1960–1964 1965–1969 1970–1971 1972–1973 1974–1975 1976–1977 1978–1979 1980–1981 1982–1983 1984–1985 1986–1987 1988–1989 1990–1991 1992–1993 1994–1995 1996–1997 1998–1999

15.4 13.7 11.6 16.7 15.7 14.0 13.6 13.3 10.3 12.1 11.8 10.5 11.4 13.5 13.2 11.1 12.4

66.5 67.9 77.1 84.4 98.8 90.8 82.8 78.1 77.2 71.0 69.0 68.8 75.4 77.8 85.6 86.3 83.8

11.0 15.7 17.0 23.2 19.9 13.5 6.9 10.4 9.5 8.3 9.5 9.1 11.3 7.2 12.6 12.2 12.4

70.6 109.8 135.6 195.4 191.4 173.3 122.5 103.7 90.3 96.4 88.5 83.6 73.9 60.9 73.5 69.3 85.2

Age Starting at about 50 years of age, uterine cancer mortality steadily increases with increasing age (Fig. 53–1) (SEER, 2002). In comparison, the incidence of uterine cancer rises rapidly in late reproductive life, peaks between 60 and 70 years of age, and plateaus or slightly declines in later life (Fig. 53–1). Exposure of the postmenopausal female population to unopposed estrogens affects the relationship between endometrial cancer incidence and age. The peak incidence is accentuated in populations where exogenous estrogen use is widespread among postmenopausal women (Table 53–1).

Geographic Area and Race/Ethnicity The incidence of uterine corpus cancer among whites exceeds that among women of other races, although United States women of various ethnicities experience the highest rates in the world (Table 53–2). Even prior to the estrogen-stimulated increase in the incidence of uterine corpus cancer, rates among white women in the United States were greater than rates among European women, as were the rates among African-American and Asian-American women relative to their counterparts in Africa and Asia, respectively (Table 53–2). In the United States differences in incidence between white- and AfricanAmericans depend on age (Fig. 53–1), with the greatest disparity in the postmenopausal years (50 to 75 years of age). In contrast, uterine

Source: Data complied from State of California, Department of Health, 1978; Marrett et al., 1978; SEER, 2002. a Rate per 100,000 adjusted, within the broad age groups shown, to a uniform standard (10 year ago groups for Connecticut, five-year age groups for Alameda County). b Whites only.

Table 53–2. Cancer Incidence of the Uterine Corpus in Selected Populations “Truncated” Incidencea Continent Africa

Area

Race

1969–1972b

1973–1977c

1978–1982d

1983–1987e

Ibadan Dakar All Harare Harare Sao Paulo Porto Alegre Detroit

All All All African European All All White Black Hispanic White Other White Am. Indian All Jews Non-Jews All All All All All All White Japanese Chinese Hawaiian Maori Non-Maori

4.2 — —

— 3.8 —

— — —

— — 2.6

Country Nigeria Senegal Gambia Zimbabwe

South America

Brazil

North America

USA

New Mexico Asia

Europe

Oceania

a

India Israel

Bombay All

China Japan Norway UK Spain Yugoslavia USA

Shanghai Miyagi All Oxford Navarra Slovenia Hawaii

New Zealand

All

9.0 20.5 18.1 — 44.1 20.8 17.9 39.0 11.4 3.1 22.9 3.1 — 3.2 23.2 20.7 — 21.2 71.4 40.7 49.0 63.8 58.1 22.8

25.7 — 56.5 18.6 17.5 53.7 9.2 3.0 21.4 1.7 10.4 4.6 25.3 22.7 27.6 23.4 76.7 46.7 66.9 75.3 33.4 24.3

20.3 15.5 38.9 15.6 18.8 30.9 9.7 3.9 16.6 6.0 6.5 5.9 23.9 18.2 23.7 23.1 36.7 31.4 37.1 51.7 31.5 17.6

Annual rate per 100,000, ages 35–64, standardized to the age distribution of the World Standard Population. (Waterhouse et al., 1976) Depending on population, rates apply to a part of the period 1969–1972. c (Waterhouse et al., 1982) Depending on population, rates apply to a part of the period 1973–1977. d (Muir et al., 1987) Depending on population, rates apply to a part of the period 1978–1982. e (Parkin et al., 1992) Depending on population, rates apply to a part of the period 1983–1987. f (Parkin et al., 1997) (truncated incidence calculated by the current authors) Depending on the population, rates apply to a part of the period 1988–1992. b

1988–1992f

— 13.2 42.4 18.6 — — — 4.6 21.0 6.8 7.4 8.4 28.2 21.9 23.3 25.2 34.8 28.2 32.6 38.2 38.9 18.7

9.4 37.6 21.1 17.9 31.2 21.9 5.1 21.9 11.1 9.3 10.1 26.6 20.1 26.8 26.3 31.9 34.4 44.0 30.0 18.2

Endometrial Cancer

Estrogen-secreting Ovarian Tumors

120

Among women with estrogen-secreting ovarian tumors (e.g., granulosa-theca cell tumors), the prevalence of endometrial carcinoma at the time of oophorectomy ranges from 6% to 21% (Diddle, 1952; Gusberg, Kardon, 1971; Larson, 1954; Salerno, 1962). It is likely that this observed frequency of endometrial tumors is considerably greater than expected even beyond possible biases in the selection of women and the diagnoses of uterine cancer.

100

80

Rate per

1029

60

Polycystic Ovaries

40

20

0 25-29 30-34 35-39 40-44 45-49 50-54 55-59 60-64 65-69 70-74 75-79 80-84 85+

Age

Incidence, whites Mortality, whites

Incidence, blacks Mortality, blacks

Figure 53–1. Average annual age-specific incidence and mortality: Cancer of the uterine corpus and uterus, NOS, United States, 1995–1999 (SEER, 2002).

cancer mortality is higher among African-Americans than whiteAmericans beginning at about 50 years of age. Maoris in New Zealand experience a rate of uterine corpus cancer approximately twice that of non-Maoris (Table 53–2) and HawaiianAmericans experience a rate that rivals that among white-Americans (Table 53–2). The high rate among Maoris is not due to an unusually high incidence of corpus cancer that is non-endometrial in nature (F. Foster, 1975, personal communication). This suggests that some feature of their shared Polynesian heritage or some common characteristic such as obesity in Maoris and Hawaiian women contributes to a relatively high incidence of endometrial cancer. Even with decreases in rates over the last 20 years, Chinese-Americans and JapaneseAmericans still experience uterine corpus cancer rates that rivals the rate among whites and that exceeds their “homeland” rates by 3-fold (Table 53–2). This suggests that some life-style features in the United States, such as obesity, may contribute to a relatively high incidence of endometrial cancer in these women.

HORMONAL RISK FACTORS Endogenous Estrogens Estrogens are the primary stimulants of endometrial proliferation. Because cellular proliferation is a prerequisite for carcinogenesis (Ruddon, 1995), and unchecked proliferation can lead to malignant transformation (Pitot, 1986), it follows that estrogens are a cause in the development of at least some endometrial carcinomas. Some casecontrol studies have assessed estrogen levels and endometrial cancer risk among postmenopausal women who were not using exogenous hormones, but blood samples were collected after cancer was known or suspected and estrogen levels in cases may have been influenced by the disease. Only one prospective study has made such an assessment; higher circulating levels of total estradiol (>8 pg/ml vs. <6 pg/ml), percent free estradiol (>1.28% vs. <1.1%), and estrone (>28 pg/ml vs. <20 pg/ml) were associated with 2- to 4-fold elevations in endometrial cancer risk after a median follow-up of 5 years (Zeleniuch-Jacquotte et al., 2001). Consistent with these findings, risk factors for endometrial cancer include medical conditions that are known to result in relatively high endogenous estrogen levels.

A number of case series have suggested that a high proportion of young women with endometrial cancer (Dockerty et a., 1951; Farhi et al. 1986) and hyperplasia (Chamlian, Taylor, 1970) have polycystic ovaries and amenorrhea or oligomenorrhea (the polycystic ovary syndrome or Stein-Leventhal syndrome). One follow-up study indicated that women with polycystic ovaries or a clinically diagnosed ovulatory disorder had a 3-fold elevated risk for endometrial cancer, although this was based on only 5 women who developed cancer (Coulam et al., 1983). Due to chronically elevated levels of luteinizing hormone, women with polycystic ovaries secrete abnormally large quantities of androstenedione. This is metabolized peripherally by the enzyme aromatase to estrone (a potent estrogen), and results in levels of estrone similar to those found at the peak of the normal ovulatory cycle (Siiteri, 1973).

Weight/Body Mass and Physical Activity Studies have consistently found that both pre- and postmenopausal women with endometrial cancer are more likely to be overweight than other women (Austin et al., 1991; Baanders-van Halewyn et al., 1996; Damon, 1960; Elwood et al., 1977; Ewertz et al., 1988; Folsom et al., 1989; Henderson et al., 1983; Hulka et al., 1980; Kelsey et al.,1982; Lawrence et al., 1987; Olson et al., 1995; Parazzini et al., 1991; Pettersson et al., 1985a; Rubin et al., 1990; Shu et al., 1991; Spengler et al., 1981; Stavraky et al., 1981; Swanson et al., 1993; Tornberg, Carstensen, 1994; Weiss et al., 1980; Wynder et al., 1966; Benshushan et al., 2001; Goodman et al., 1997a; Hachisuga et al., 1998; Hirose et al., 1996; Hirose et al., 1999; Kalandidi et al., 1996; Newcomer et al., 2001; Petridou et al., 2002a; Salazar-Martinez et al., 2000; Shoff, Newcomb, 1998; Terry et al., 1999; Weiderpass et al., 2000b; Zelmanowicz et al., 1998). Depending on the measures used to identify excess weight or body mass, the elevation in risk for overweight women relative to women of normal weight is roughly 2- to 10-fold. While some studies report that risk increases with incremental increases in weight or body mass (Baanders-van Halewyn et al., 1996; Ewertz et al., 1988; Goodman et al., 1997a; Henderson et al., 1983; Petridou et al., 2002a; Tornberg, Carstensen, 1994), others find a strong elevation in risk only among obese women (Elwood et al., 1977; Goodman et al., 1997a; Hirose et al., 1996; Kelsey et al., 1982; Newcomer et al., 2001; Rubin et al., 1990; Salazar-Martinez et al., 2000; Shoff, Newcomb, 1998; Weiderpass et al., 2000b; Weiss et al., 1980). Greater weight gains (such as >40 lbs, >12 kg, or >6 kg/m2) from early adulthood (16 to 25 years of age) to diagnosis may also increase risk (Olson et al., 1995; Shu et al., 1992; Swanson et al., 1993; Terry et al., 1999), although other studies find no such relation (Le Marchand et al., 1991; Weiderpass et al., 2000b). Obesity leads to a net increase in the amount of endogenous estrogens, both through increased conversion of androstenedione to estrogen (Edman, MacDonald, 1978; MacDonald et al., 1978; MacDonald, Siiteri, 1974) and decreased circulating levels of sex hormone-binding globulin (Davidson et al., 1981; Kaye et al., 1991; Nyholm et al., 1993). Given equivalent body mass, women with greater upper body fat as indicated by a waist-to-hip ratio of >1.14 vs £1.14 and an abdomen-to-thigh skin fold ratio of >0.82 vs £0.82 may be at a particularly high risk for endometrial cancer with relative risks of 15 (95%CI = 2.0–58.0) and 6.7 (95%CI = 1.5–27.0), respectively (Schapira et al., 1991). These results are consistent with the biologic mechanism noted above, because women with upper body fat localization have lower circulating levels of sex hormone-binding globulin and higher levels of nonprotein bound estrogen (Kirschner et al., 1990).

1030

PART IV: CANCER BY TISSUE OF ORIGIN

While the results of many studies suggest that there is an elevated risk of endometrial cancer among relatively inactive women compared to active women (Hirose et al., 1996; Levi et al., 1993c; Littman et al., 2001b; Moradi et al., 2000; Pukkala et al., 1993; Salazar-Martinez et al., 2000; Sturgeon et al., 1993; Terry et al., 1999), the results of other studies are negative (Dosemeci et al., 1993; Goodman et al., 1997a; Olson et al., 1997; Shu et al., 1993a; Zheng et al., 1993). Whether a low level of physical activity plays a role in obesity-related endometrial cancer is unclear at the present time; only a few studies have looked at physical activity in obese versus non-obese women and the results have been inconsistent (Levi et al., 1993c; Littman et al., 2001b; Moradi et al., 2000). A better understanding of the independent effects of obesity and physical activity are needed to target appropriate interventions for primary prevention.

ance of carcinoma (Deligdisch, Holinka, 1987; Gusberg, Hall, 1961; Pettersson et al., 1985b). A follow-up study of 170 women (mean follow-up: 13.4 years) with untreated hyperplasia found that 1.6% of the women with simple hyperplasia and 23% of the women with atypical hyperplasia developed endometrial cancer (Kurman et al., 1985).

Endogenous Progesterone

Duration/Recency

The surge of luteal progesterone secretion which begins just prior to ovulation each month in premenopausal women arrests endometrial proliferation, promotes secretory differentiation of the endometrium, and initiates endometrial sloughing in the absence of fertilization (King, Whitehead, 1983). These events decrease the likelihood of developing hyperplastic or neoplastic lesions, because cell differentiation alone can arrest or reverse neoplastic cellular transformation (Pitot, 1986). Thus, conditions that are characterized by low endogenous progesterone levels, particularly when coupled with relatively high estrogen levels, are generally associated with higher rates of endometrial cancer. This is a likely explanation for the peak incidence of endometrial cancer that occurs after the menopause (Fig. 53–1), when limited estrogen production continues (through peripheral conversion of adrenal androgens) without any cyclic progesterone production or substantial peripheral production of progesterone. Similarly, women with the polycystic ovary syndrome (discussed previously) not only have excessive production of estrogen, but also lack cyclic progesterone secretion (Farhi et al., 1986; Lucas, 1974; Yen, 1980). Lastly, the increased risk of endometrial cancer in premenopausal obese women may be related not only to a hyperestrogenic state, but to decreased progesterone levels as well. A study of six obese, oligomenorrheic premenopausal women found consistently subnormal levels of serum progesterone relative to ten nonobese controls, even though the luteal phase in all subjects lasted at least ten days (Sherman, Korenman, 1974). Estrogen levels did not differ between the two groups.

Most studies have found an elevation in risk beginning after 1 to 5 years of use with risk elevations increasing further with increasing duration of use (Antunes et al., 1979; Brinton et al., 1993b; Buring et al., 1986; Gray et al., 1977; Green et al., 1996; Hoogerland et al., 1978; Hulka et al., 1980; Jain et al., 2000b; Jelovsek et al., 1980; Jick et al., 1993a; Kelsey et al., 1982; La Vecchia et al., 1984; Mack et al., 1976; McDonald et al., 1977; Paganini-Hill et al., 1989; Persson et al., 1989; Pike et al., 1997; Rubin et al., 1990; Shapiro et al., 1980; Shapiro et al., 1985; Spengler et al., 1981; Stampfer et al., 1986; Stavraky et al., 1981; Weiderpass et al., 1999a; Ziel, Finkle, 1975). Relative to the incidence among current users, risk of endometrial cancer decreases after cessation of estrogen therapy (Brinton et al., 1993b; Finkle et al., 1995; Green et al., 1996; Hulka et al., 1980; Levi et al., 1993b; Mack et al., 1976; Pettersson et al., 1986b; Shapiro et al., 1980; Shoff, Newcomb, 1998; Stavraky et al., 1981; Weiss et al., 1979). In three studies there remained little or no residual excess above the rate in women who never used hormones after two drug-free years had elapsed (Finkle et al., 1995; Hulka et al., 1980; Pettersson et al., 1986b), whereas in other studies some excess risk persisted for up to 3–5 years (Buring et al., 1986; Jain et al., 2000b; Shapiro et al., 1980; Weiderpass et al., 1999a) and even 10 years (Levi et al., 1993b; Paganini-Hill et al., 1989; Shapiro et al., 1985). Following cessation, a particularly large excess risk may remain for women with the longest durations of estrogen therapy (Green et al., 1996; Rubin et al., 1990), except perhaps for women who used lower doses (<1.25 mg/day conjugated estrogens) (Cushing et al., 1998).

Endometrial Cancer Unopposed estrogen use is positively associated with an increase of endometrial cancer in nearly all epidemiologic studies that have reported on this association. Detailed reviews of unopposed estrogen and endometrial cancer can be found elsewhere (Grady D. et al., 1995; Grady, 1996; Herrington, 1993). In the interest of brevity, the following discussion will provide only a summary of selected aspects of estrogen use.

Type Exogenous Estrogens Postmenopausal Estrogens Hyperplasia. It is well established that women who take estrogens unopposed by progestogens develop adenomatous hyperplasia of the uterus more commonly than do other women (Gusberg, 1947; Pickar et al., 2001; Speroff et al., 1996; Writing Group for the PEPI Trial, 1995). For example, among the women assigned to receive unopposed estrogen therapy (conjugated equine estrogen, 0.625 mg/day) in the Postmenopausal Estrogen/Progestin Interventions (PEPI) trial, 22% developed adenomatous hyperplasia and 12% developed atypical hyperplasia after three years of follow-up (Writing Group for the PEPI Trial, 1995). Fewer than 1% of women in the placebo arm and 1% of other women who received various combinations of estrogen and progestogens in the trial developed hyperplasia. In another randomized trial, the Continuous Hormones as Replacement Therapy (CHART) Study, endometrial hyperplasia occurred in 55.6% of women receiving 10 ug of unopposed ethinyl estradiol daily compared with 1.7% of the women receiving placebo (Speroff et al., 1996). Some hyperplastic lesions regress when estrogens are discontinued (Kistner, 1973), and most hyperplastic lesions can be successfully treated with progestogens (Thom et al., 1979). One or more forms of hyperplasia probably represent an early stage in the development of endometrial carcinoma, as they frequently coexist with endometrial cancer in estrogen users and have been observed to precede the appear-

Use of conjugated estrogens, the type of estrogen most commonly prescribed in the United States, has consistently been related to an increased endometrial cancer risk of roughly 2- to 20-fold (Antunes et al., 1979; Buring et al., 1986; Cushing et al., 1998; Gray et al., 1977; Horwitz, 1978; Hulka et al., 1980; Mack et al., 1976; McDonald et al., 1977; Persson et al., 1989; Shapiro et al., 1980; Weiss et al., 1979; Ziel, 1975; Weiderpass et al., 1999a). Similar results have been reported for other non-conjugated estrogens (e.g., stilbestrol or ethinyl estradiol) in most studies (Antunes et al., 1979; Gray et al., 1977; Mack et al., 1976; Persson et al., 1989; Weiderpass et al., 1999a; Weiss et al., 1979), but not all (Shapiro et al., 1980). Among users of vaginal hormone creams, two studies reported a small elevation in endometrial cancer risk (Kelsey et al., 1982; Weiderpass et al., 1999c), whereas two others did not (Brinton et al., 1993b; Gray et al., 1977). Transdermal patches and intramuscular injections of estrogen have been evaluated in too few women in existing studies to adequately address their influence on risk.

Dosage No increase in the incidence of hyperplasia was noted among women treated with 0.3 mg of esterified estrogen relative to the incidence in women treated with placebo (Genant et al., 1997). However, the risk of endometrial cancer appears to be elevated with all commonly prescribed dosages of conjugated estrogens (0.3 mg to 1.25 mg per day or the equivalent amount of other estrogens), and may rise with increas-

Endometrial Cancer ing dose (Antunes et al., 1979; Buring et al., 1986; Cushing et al., 1998; Gray et al., 1977; Hulka et al., 1980; Jick et al., 1979; Stavraky et al., 1981; Weiderpass et al., 1999a; Weiss et al., 1979). Women who take unopposed estrogens orally usually do so either on a daily basis or with a break of 5 to 7 days each month. Each of these regimens are associated with an elevated risk for both endometrial hyperplasia (Schiff et al., 1982) and endometrial cancer (Antunes et al., 1979; Buring et al., 1986; Hulka et al., 1980; Mack et al., 1976; McDonald et al., 1977; Stavraky et al., 1981; Weiss et al., 1979).

Stage of Cancer While both higher and lower grades of endometrial cancer are found in excess among unopposed estrogen users, the highest risks are seen for the less invasive and more highly differentiated tumors (Buring et al., 1986; Hulka et al., 1980; Kelsey et al., 1982; Shapiro et al., 1998; Shapiro et al., 1985; Weiss et al., 1979). This may be due, in part, to atypical hyperplasia being misdiagnosed as early endometrial cancer in some instances (Gordon et al., 1977; Szekely et al., 1978) and to earlier detection of cancer in estrogen users than nonusers (Mack et al., 1976; Weiss, 1978b). Whatever the reason, postmenopausal estrogen users have a more favorable endometrial cancer survival experience than do nonusers (Chu et al., 1982; Collins et al., 1980; Elwood, 1980; Robboy, 1979).

Influence of Other Risk Factors on the Estrogen/Endometrial Cancer Association Five studies have noted an elevated relative risk of endometrial cancer among estrogen-users relative to non-users in both smokers and nonsmokers (Brinton et al., 1993b; Levi et al., 1993b; Rubin et al., 1990; Shields et al., 1999; Weiss et al., 1980), although another noted an attenuation of the increased risk with estrogen use among current smokers (Newcomer et al., 2001). Most studies have reported similar elevations in relative risk irrespective of parity among estrogen-users relative to non-users (Brinton et al., 1993b; Hoogerland et al., 1978; Jelovsek et al., 1980; Rubin et al., 1990; Shields et al., 1999), although one noted a particularly high relative risk in women of low parity (Weiss et al., 1980). In both users and non-users of oral contraceptives the use of estrogen replacement therapy was associated with an elevated risk of endometrial cancer in three studies (Brinton et al., 1993b; Levi et al., 1993b; Shields et al., 1999), whereas in another there was no elevation in risk among oral contraceptive users who later used estrogen (Rubin et al., 1990). Almost all studies have reported an elevation in relative risk among estrogen-users compared to non-users across all categories of weight or body mass (Ewertz et al., 1988; Hoogerland et al., 1978; Jelovsek et al., 1980; Kelsey et al., 1982; Levi et al., 1993b; Rubin et al., 1990; Weiss et al., 1980). The results of some studies suggest that the relative risk is greater among women of lower weight or body mass (Kelsey et al., 1982; Levi et al., 1993b) or even restricted to this group (Brinton et al., 1993b; Hulka et al., 1980). Another observed that the magnitude of the absolute increase in risk of endometrial cancer was greater in heavy women than in light women (Shields et al., 1999). In summary, the evidence to date indicates that women in all subgroups of known risk factors (i.e., smokers and non-smokers, parous and nulliparous women, etc.) have an elevated risk for endometrial cancer with unopposed estrogen use. No subgroup is free from some elevation in risk.

Other Use of Unopposed Estrogen Further information about the long-term effects of estrogen supplementation comes from a series of young women with gonadal agenesis (undeveloped ovaries) who received diethylstibestrol for prolonged periods. These women had an usually high incidence of endometrial cancer (Cutler et al., 1972). Additionally, a group of patients with breast cancer who had received estrogenic hormones (primarily stilbestrol) as a cancer treatment had about twice the incidence of endometrial cancer than that expected for other breast cancer patients (Hoover et al., 1976).

1031

Selected Estrogen Receptor Modulators Tamoxifen, a non-steroidal hormone used in the treatment of breast cancer, has been shown to be estrogenic in the human uterus (Gorodeski et al., 1992; Satyaswaroop et al., 1984) and to increase the risk of endometrial cancer, particularly when given at a relatively high daily dosage (30 to 40 mg) or for relatively long periods of time (5 or more years). In randomized controlled trials of breast cancer patients and women without cancer, relative risks of 6.4 (95% CI, 1.4–28), 7.5 (95% CI, 1.7–33), and 2.5 (95% CI, 1.4–5.03) for endometrial cancer have been associated with tamoxifen regimens of 40 mg/day for 2 years in the first trial and 20 mg/day for 5 years in the latter two (Fisher et al., 1994; Fisher et al., 1998; Fornander et al., 1989). Three casecontrol studies in breast cancer patients that adjusted for confounders such as BMI and HRT also found 4- to 10-fold elevations in risk with 5 or more years of tamoxifen use (Bergman et al., 2000; Bernstein et al., 1999; Pukkala et al., 2002). The results above contributed to the current therapy recommendations of a total of 5 years of tamoxifen use for breast cancer patients and for breast cancer prophylaxis (PDR, 2003). Although tamoxifen therapy for less than two years does not appear to elevate endometrial cancer risk (Bergman et al., 2000; Bernstein et al., 1999; Cook et al., 1995; Ribeiro, Swindell, 1992; Sasco et al., 1996) and therapy for five or more years clearly increases risk (as noted above), results are not consistent across studies for tamoxifen use for two to five years. As daily doses of tamoxifen can vary between countries and regions, risk associated with 2 to 5 years of use is probably best assessed with cumulative doses of therapy rather than total durations of therapy. Two studies report results for cumulative doses; in one study only cumulative doses greater than 15,000 mg (the equivalent of 20 mg/day for about 2.1 years) resulted in an elevated risk (Bernstein et al., 1999), whereas in the other study only cumulative doses greater than 26,060 mg (the equivalent of 20 mg/day for about 3.5 years) resulted in an elevated risk (Pukkala et al., 2002). Although there has been limited power to assess interactions in studies, the results of one study suggests that an elevated risk with two to five years of tamoxifen use may be limited to women with prior ERT use and a high BMI (>24.5 kg/m2) or to nonsmokers (Bernstein et al., 1999). While improved breast cancer prognosis with two to five years of tamoxifen therapy may outweigh any modest increases in endometrial cancer risk in breast cancer patients, such increases may be of concern for women using tamoxifen to prevent breast cancer and are worthy of further study. Another non-steroidal hormone, raloxifene, is used to prevent and treat osteoporosis in postmenopausal women (Jordan, 1995). In contrast to tamoxifen, raloxifene does not appear to stimulate proliferation of the endometrium (Boss et al., 1997) and early results suggest that it does not increase endometrial cancer risk (Cauley et al., 2001). A number of studies are underway to more fully assess the risk/benefit profile of raloxifene.

Exogenous Progestogens Exogenous progestogens mimic the actions of luteal progesterone in promoting differentiation and arresting proliferation of endometrial tissue. These beneficial effects are present even when continous low-dose progestogens are given with estrogens and no endometrial sloughing occurs (Magos et al., 1985; Mattsson, Samsioe, 1985; Staland, 1985).

Postmenopausal Estrogens-Progestogens Hyperplasia. As mentioned previously, several studies have found that estrogen-stimulated hyperplasia reverts to normal endometrium after administration of exogenous progestogens (Kistner, 1973; Thom et al., 1979; Whitehead et al., 1977). Several studies have found that an estrogen-progestogen regimen is associated with a lower occurrence of hyperplasia (0–4%) (Clisham et al., 1991; Paterson et al., 1980; Sturdee et al., 1978; Thom et al., 1979; Whitehead et al., 1979) than estrogen alone, particularly when the progestogens are used for more than ten days each month (Paterson

1032

PART IV: CANCER BY TISSUE OF ORIGIN

et al., 1980; Sturdee et al., 1978; Thom et al., 1979). More recently, the occurrence of hyperplasia in women randomized to progestogen supplementation, either continuous or for twelve days each month and either through oral or transdermal delivery, was found to be very low (<2%) relative to women receiving unopposed estrogen (8–56%) (Archer et al., 1999; Corson et al., 1999; Kurman et al., 2000; Shulman et al., 2002; Woodruff et al., 1994), and no different than that among women receiving placebo (Brynhildsen, 2002; Pickar et al., 2001; Speroff et al., 1996; Writing Group for the PEPI Trial, 1995). To counter the often unwanted monthly menses in women receiving progestogen supplementation of estrogen each month, a “long-cycle” regimen has been proposed where progestogen supplementation is given for 10 to 14 days once every three months. In one study, only one of the 219 women followed for one year and none of the 132 women followed for two years had evidence of hyperplasia with a long-cycle regimen that included 20 mg per day of medroxyprogesterone acetate for 14 days (Hirvonen et al., 1995). In another study that included 10 mg per day of medroxyprogesterone acetate for 14 days, 1.5% of the 214 women given a long-cycle regimen developed hyperplasia after one year of follow-up, similar to the baseline prevalence of 0.9% (Ettinger et al., 1994). However, a Scandinavian trial was recently discontinued after 3 years of follow-up due to an unacceptably high occurrence of hyperplasia in the long-cycle group (6%)

relative to a monthly-cycle group (<1%); the long-cycle group received 1 mg per day of norethindrone for 10 days (Cerin et al., 1996).

Endometrial Cancer Compared to the epidemiological information on unopposed estrogen therapy and endometrial cancer, information on combined estrogenprogestogen therapy and endometrial cancer is sparse. In a small, randomized controlled trial of institutionalized women, no cases of endometrial cancer were observed in the 84 women in the combined therapy group (2.5 mg estrogen/day and 10 mg progesterone for 7 days/month), and only one was observed in the 84 women in the placebo group after 10 years of follow-up (Nachtigall et al., 1979). A small cohort study also observed no cases of endometrial cancer among women who received estrogen-progestogen therapy, whereas three cases were observed among the non-users (Hammond et al., 1979). Other studies that have assessed the impact of progestogen supplementation of postmenopausal estrogens on the risk of endometrial cancer are summarized in Table 53–3. As predicted from the studies of hyperplasia, all observed a lower risk of endometrial cancer associated with the use of combined therapy than with unopposed estrogens. Yet, compared with women who did not use hormones, only one study found a reduction in endometrial cancer risk with any combined therapy use (Gambrell et al., 1979), whereas the remainder found no

Table 53–3. Summary of Studies Evaluating Postmenopausal Estrogen plus Progestin Therapy and Endometrial Cancer Study

Study Design

Gambrell et al. (1979)

cohort

Voigt et al. (1991)

case-control

Jick et al. (1993a)

case-control

Brinton et al. (1993b)

case-control

Beresford et al. (1997)

case-control

Pike et al. (1997)

case-control

Weiderpass et al. (1999a)

case-control

Persson et al. (1999)

cohort

g

Hill et al. (2000) Jain et al. (2000b)

case-control case-control

Pukkala et al. (2001)

cohort

Hulley et al. (2002) WHI Writing Group (2002)

RCTi RCT

Number of exposed cases

Type/Measure of Combined Therapy

RRa

95% CI

8 2b 54 18b 11b 7b 32 18b 6b 60 11b 324 67c 25c 25c 422d

rate in estrogen users vs nonusers rate in E + P* users vs nonusers estrogen users vs nonusers any use of E + P vs nonusers E + P with progestin for <10 days/month E + P with progestin for ≥10 days/month current/recent estrogen users vs nonusers current/recent use of E + P vs nonusers past use of E + P vs nonusers estrogen users vs nonusers any use of E + P vs nonusers estrogen users vs nonusers any use of E + P vs nonuser E + P with progestin for <10 days/month E + P with progestin for ≥10 days/month for each 5 years estrogen use vs nonusers for each 5 years E + P use vs nonusers with progestin <10 days/month with progestin ≥10 days/month with continuous progestin estrogen users vs nonusers any use of E + P vs nonuser with cyclic progestinf with continuous progestin rate with 6+ years estrogen only vs general population rate with 6+ years E + P vs general population continuous E + P vs nonusers estrogen users vs nonusers any use of E + P vs nonusers E + P cyclic vs nonusers E + P continuous vs nonusers rate with long cycle E + Ph vs general population rate with cyclic progestin E + Pf vs general population E + P continuous vs placebo E + P continuousj vs placebo

1.9 0.2 3.1 1.3 2.0 0.9 6.5 1.9 0.9 3.4 1.8 4.0 1.4 3.1 1.3 2.2e

NPa NP (1.6, 5.8) (0.6, 2.8) (0.7, 5.3) (0.3, 2.4) (3.1, 13.3) (0.9, 3.8) (0.3, 3.4) (1.8, 6.3) (0.6, 4.9) (3.1, 5.1) (1.0, 1.9) (1.7, 5.7) (0.8, 2.2) (1.9, 2.5)

1.9e 1.1e 1.1e 3.2 1.3 2.0 0.7 4.2 1.4 0.6 2.2 1.3 1.1 1.5 2.0 1.3 0.3 0.8

(1.3, 2.7) (0.8, 1.4) (0.8, 1.4) (2.4, 4.4) (1.0, 1.7) (1.4, 2.7) (0.4, 1.0) (2.1, 8.4) (0.6, 3.3) (0.3, 1.3) (1.5, 3.4) (0.9, 1.8) (0.7, 1.6) (0.7, 3.4) (1.6, 2.6) (1.1, 1.6) (0.1, 1.2) (0.5, 1.5)

74d 79d 94d 98 119 27 11 9 77 97 65 15 2 22

RR = relative risk; E + P = estrogen plus progestin; NP = information not provided in published paper. women with prior unopposed estrogen use included. c women with prior unopposed estrogen use excluded. d categories not mutually exclusive; 509 cases had used one or more types of hormone replacement therapy. e adjusted for use of other types of hormone replacement therapy. f any length of monthly cyclic progestin. g includes study population from (Beresford et al., 1997). h long cycle E + P = 70 days estradiol valerate, 14 days estradiol valerate plus medroxyprogesterone acetate, 7 days placebo. i RCT = randomized controlled trial. j includes women previously randomized to estrogen alone. a

b

Endometrial Cancer difference or a modest elevation in risk (Beresford et al., 1997; Brinton et al., 1993b; Hill et al., 2000; Hulley et al., 2002; Jain et al., 2000b; Jick et al., 1993a; Persson et al., 1989; Persson et al., 1999; Pike et al., 1997; Pukkala et al., 2001; Voigt et al., 1991; Weiderpass et al., 1999a; WHI Writing Group, 2002). Two studies report that endometrial cancer risk with combined hormone therapy is lower for all stages of disease relative to that for unopposed estrogen users (Pike et al., 1997; Shapiro et al., 1998). Studies that were large enough to address the question reported a 2 to 4-fold elevated risk associated with short cyclic progestogen use (i.e., for less than 10 days each month) (Beresford et al., 1997; Pike et al., 1997; Voigt et al., 1991), particularly when this regimen was used for five years or longer (Beresford et al., 1997; Pike et al., 1997). No elevation in risk has been noted with a regimen of longer cyclic progestogen use (i.e., for 10 or more days per month) when taken for less than five years (Beresford et al., 1997; Pike et al., 1997), whereas more than five years use of this regimen has been associated with an elevated risk in one study (Beresford et al., 1997) but not another (Pike et al., 1997). Six studies have evaluated continuous combined therapy. Three studies report no difference in risk between users and nonusers of these hormones (Jain et al., 2000b; Pike et al., 1997; WHI Writing Group, 2002), whereas three others observed a suggestion of a reduction in risk (Hill et al., 2000; Hulley et al., 2002; Weiderpass et al., 1999a), with the greatest reduction noted with longer durations of therapy (≥5 years) (Weiderpass et al., 1999a). The results of studies that have investigated the impact of using a relatively high progestogen dosage (e.g., 10 mg/day of methoxyprogesterone acetate) are not consistent. In one study women given a high dosage of progestogen for more than 10 days per month for 5 or more years still had an elevated endometrial cancer risk compared with nonusers of hormones (Beresford et al., 1997), whereas another study reported no elevation in risk associated with the same regimen (Pike et al., 1997). Another study reported that, among women who used combined hormone therapy, a lower proportion of women with endometrial cancer (17%) than other women (29%) had used highdosage progestogen (Jick et al., 1993a). In summary, while available evidence suggests that the occurrence of endometrial cancer is lower in combined estrogen-progestogen users than in estrogen-only users, and that there may be a reduction in risk with continuous combined therapy, the long-term risks associated with different dosages and monthly durations of progestogen supplementation still require evaluation.

Hormonal Contraception Women are also exposed to exogenous progestogens through use of hormonal contraceptive agents. Over the past 40 years this has included sequential and combination oral contraceptives, progestinonly pills, “minipills,” and the long-acting injectable or implanted progestogens (e.g., medroxyprogesterone). While a fair amount of study has been devoted to oral contraceptive use in relation to endometrial cancer risk, little information is available concerning the impact of the other hormone contraceptives. Although evidence is limited, the use of progestin-only pills or injected/implanted progestogens may substantially reduce endometrial cancer risk. A United States study reported no increase in uterine cancer incidence after 13 years of follow-up among African-American women receiving depot medroxyprogesterone acetate (DMPA) injections relative to AfricanAmerican women in general (Liang et al., 1983). A study in Thailand reported that endometrial cancer incidence was approximately 80% (95% CI = 0.1, 0.8) lower among DMPA users than among other women (WHO Collaborative Study of Neoplasia and Steroid Contraceptives, 1991). Furthermore, a Swedish case control study reported that DMPA use was only reported among control women, and also reported a suggested reduction (OR = 0.6, 95% CI = 0.2, 1.4) in risk with the use of progestogen-only pills (Weiderpass et al., 1999b). Oral contraceptive pills (OCPs) were available initially as sequential preparations (estrogen only followed by a short course of estrogen plus progestogen) or as combination preparations (concurrent estrogen and progestogen). Starting in the mid-1970s, there were reports

1033

of endometrial abnormalities (Kaufman et al., 1976; Lyon, Frisch, 1976) and an increased risk of endometrial cancer (Henderson et al., 1983; Lyon, Frisch, 1976; Silverberg et al., 1977; The Centers for Disease Control, Cancer and Steroid Hormone Study, 1983; Weiss, 1980) among women who used sequential preparations, particularly Oracon (Mead Johnson, Evansville, IL), which contained a relatively potent estrogen (ethinyl estradiol) followed by a weak progestogen (dimethisterone). Two studies observed no excess risk with use of sequential OCPs or Oracon (Kaufman et al., 1980a; Kelsey et al., 1982), but they included a very small number of women who used sequential OCPs. Sequential preparations were removed from the consumer market in the United States and Canada in 1976. In contrast, women who have taken combination OCPs have about one-half the risk of endometrial cancer as do nonusers (Armstrong et al., 1988; Benshushan et al., 2001; Beral et al., 1988; Hannaford, 1998; Henderson et al., 1983; Hulka et al., 1982; Jick et al., 1993b; Kaufman et al., 1980a; Kelsey et al., 1982; La Vecchia et al., 1986a; Levi et al., 1991; Parslov et al., 2000; Pettersson et al., 1986a; SalazarMartinez et al., 2000; Stanford et al., 1993; The Centers for Disease Control, Cancer and Steroid Hormone Study, 1987; Vessey, 1995; Voigt et al., 1994; Weiss, 1980; Zelmanowicz et al., 1998). The reduced risk appears to be evident within 2 to 5 years of initiation of use (Henderson et al., 1983; Hulka et al., 1982; Jick et al., 1993b; Kaufman et al., 1980a; Levi et al., 1991; Stanford et al., 1993; The Centers for Disease Control, Cancer and Steroid Hormone Study, 1987; Weiderpass et al., 1999b), and further decreases as the duration of OCP use increases (Henderson et al., 1983; Hulka et al., 1982; Kaufman et al., 1980a; Kelsey et al., 1982; Levi et al., 1991; Pettersson et al., 1986a; The Centers for Disease Control, Cancer and Steroid Hormone Study, 1987; Weiderpass et al., 1999b). In a study restricted to women less than 50 years of age, even short-term OCP use (<1 year) was associated with a reduced risk (Parslov et al., 2000). Some studies report a greater reduction in risk with relatively more recent use (Jick et al., 1993b; Levi et al., 1991; Stanford et al., 1993; Weiderpass et al., 1999b), but another study reported no difference according to the time since last use (The Centers for Disease Control, Cancer and Steroid Hormone Study, 1987). When duration and recency of use were evaluated jointly, longer durations of use (>5 years), but not shorter durations of use (<5 years), were associated with a reduced endometrial cancer risk irrespective of recency (Voigt et al., 1994). Some studies report that the reduction in risk may be greatest with OCPs in which progestogen effects predominate (Hulka et al., 1982) or that contain higher dosages of progestogen (Rosenblatt et al., 1991), but another study found that a longer duration of use (>5 years), and not progestogen dosage, was most predictive of a reduced risk (Voigt et al., 1994). In joint evaluation with other risk factors for endometrial cancer, no reduction in relative risk was found among combination OCP users in the heaviest category of body weight in two studies (Henderson et al., 1983; Stanford et al., 1993), whereas a reduced relative risk was found among OCP users regardless of weight or body mass in three others (Levi et al., 1991; The Centers for Disease Control, Cancer and Steroid Hormone study, 1987; Weiderpass et al., 1999b). Four studies found that relative risk reductions with combination OCP use were strongest among parous women (Levi et al., 1991; Weiderpass et al., 1999b) or women of higher parity (>5 births) (Armstrong et al., 1988; Stanford et al., 1993), but another noted a reduced relative risk for endometrial cancer only among OCP users who were nulliparous (The Centers for Disease Control, Cancer and Steroid Hormone Study, 1987). No reduction in relative risk among OCP users who later used estrogen replacement therapy for 3 or more years was noted in two studies (Stanford et al., 1993; Voigt et al., 1994), whereas a reduced relative risk among OCP users who ever used ERT was noted in five others (Hulka et al., 1982; Kaufman et al., 1980a; Levi et al., 1991; The Centers for Disease Control, Cancer and Steroid Hormone Study., 1987; Weiderpass et al., 1999b). The inclusion of women with brief use (less than 2 or 3 years) of ERT in the latter studies could have obscured the attenuation of the association between combined OCPs and endometrial cancer that appears to be present among women with longer durations of ERT use.

1034

PART IV: CANCER BY TISSUE OF ORIGIN

OTHER RISK FACTORS Events of Reproductive Life Some studies report a reduction in endometrial cancer risk for those with older ages (14–17 years of age) relative to those with younger ages (10–12 years of age) at menarche (Brinton et al., 1992; Elwood et al., 1977; Ewertz et al., 1988; Kalandidi et al., 1996; Kelsey et al., 1982; Koumantaki et al., 1989; Kvale et al., 1988; McPherson et al., 1996; Petridou et al., 2002b), but others do not (Hirose et al., 1996; Mack et al., 1976; Pettersson et al., 1986a; Salazar-Martinez et al., 1999; Spengler et al., 1981; Wynder et al., 1966). A more consistent finding is that women who experience a natural menopause at a relatively late age are at a modestly greater risk of endometrial cancer than are other women (Baanders-van Halewyn et al., 1996; Elwood et al., 1977; Ewertz et al., 1988; Kalandidi et al., 1996; Kelsey et al., 1982; Koumantaki et al., 1989; Kvale et al., 1988; McPherson et al., 1996; Petridou et al., 2002b; Pettersson et al., 1986a; Spengler et al., 1981). Nulliparity is consistently associated with an increased risk of endometrial cancer, and among women with at least one birth, the probability of developing endometrial cancer is inversely related to the number of births (Albrektsen et al., 1995; Baanders-van Halewyn et al., 1996; Brinton et al., 1992; Elwood et al., 1977; Henderson et al., 1983; Hirose et al., 1996; Inoue et al., 1994; Kalandidi et al., 1996; Kelsey et al., 1982; Koumantaki et al., 1989; Lambe et al., 1999; Mogren et al., 2001; Parazzini et al., 1998; Parslov et al., 2000; Petridou et al., 2002b; Pettersson et al., 1986a; Rubin et al., 1990; Salazar-Martinez et al., 1999; Shoff, Newcomb, 1998; Shu et al., 1991; Terry et al., 1999), even beyond five births (Hinkula et al., 2002). This parity-related reduction in risk appears to be unmodified by smoking status (Newcomer et al., 2001), but may be greater among women diagnosed at younger ages (<50 or <55 years) than among women diagnosed at older ages (Hachisuga et al., 1998; Parazzini et al., 1998). Nulliparious women also may have a poorer survival than parous women (Hachisuga et al., 2001; Salvesen et al., 1998), with each additional childbirth related to a 9.5% reduction in uterine cancer mortality (Lochen, Lund, 1997). The evidence for a reduction in endometrial cancer risk with one or more spontaneous or therapeutic abortions is mixed. Five studies report no association (Brinton et al., 1992; Elwood et al., 1977; Kalandidi et al., 1996; Salazar-Martinez et al., 1999; Shu et al., 1991), but two others report a reduction in risk (McPherson et al., 1996; Parazzini et al., 1998) with spontaneous abortions. Two studies report no association (Brinton et al., 1992; Kalandidi et al., 1996), but three others found a reduction in risk (Parazzini et al., 1998; Parslov et al., 2000; Shu et al., 1991) among women with a history of therapeutic abortions. No consistent alteration in endometrial cancer risk has been found in women who have undergone tubal sterilization (Castellsague et al., 1996; Lacey, Jr. et al., 2000). In contrast to findings on breast cancer, there is no consistent relationship between age at first birth and the incidence of endometrial cancer after adjustment for parity (Albrektsen et al., 1995; Brinton et al., 1992; Henderson et al., 1983; Kvale et al., 1988; Lambe et al., 1999; Lesko et al., 1991; McPherson et al., 1996; Mogren et al., 2001; Parazzini et al., 1998; Parslov et al., 2000; Salazar-Martinez et al., 1999). Evidence for a reduced risk associated with an older age at last pregnancy (>35 years of age) or a shorter time period between last pregnancy and diagnosis (<10 years) is more consistent; all but one study (Albrektsen et al., 1995) has found a reduction in risk (Hinkula et al., 2002; Kvale et al., 1988; Lambe et al., 1999; Lesko et al., 1991; McPherson et al., 1996; Parazzini et al., 1998; Salazar-Martinez et al., 1999). An international study (Rosenblatt, Thomas, 1995), as well as a study from Mexico (Salazar-Martinez et al., 1999), observed that lactation, especially longer durations of lactation, is associated with a lower risk of endometrial cancer. In contrast, studies from the United States and Japan found neither a difference in risk between women who had and had not breast fed (Brinton et al., 1992; Hirose et al., 1996; Newcomb, Trentham-Dietz, 2000) nor any evidence of a inverse risk with longer durations of breast feeding (Newcomb, TrenthamDietz, 2000). Only one study has investigated the use of lactationsuppressant hormones that usually are composed of estrogens; a

slight elevation in risk was noted among women using these hormones two or more times relative to non-users (Newcomb, Trentham-Dietz, 2000). After accounting for parity, impaired fertility measured in various ways (e.g., three or more years of unsuccessfully attempting pregnancy, seeking medical advice for infertility, or self-reported infertility) has been associated with an increased risk of endometrial cancer in most (Benshushan et al., 2001; Brinton et al., 1992; Escobedo et al., 1991; Henderson et al., 1983; The Centers for Disease Control, Cancer and Steroid Hormone Study, 1983), but not all (McPherson et al., 1996) studies. In addition, a rate roughly two to five times that in the general population is found in most (Modan et al., 1998; Venn et al., 1995; Venn et al., 1999), but not all (Doyle et al., 2002), studies investigating endometrial/uterine cancer among cohorts of women either referred for fertility treatment or diagnosed with infertility. In one of these cohorts, a nine-fold elevation in risk (95% CI = 5.0–16.0) was noted among women whose infertility was characterized by normal estrogen production but progesterone deficiency (Modan et al., 1998).

Intrauterine Contraceptive Devices Intrauterine contraceptive devices (IUDs) cause alterations in the surface morphology of cells (Shaw, 1979), as well as acute and chronic endometrial inflamation (Moyer, 1971), events that may influence the development of endometrial cancer. However, nine epidemiologic studies investigated this relationship and none of them have found an elevation in risk. Seven studies found a 40% to 60% reduction in endometrial cancer risk among women who ever used an IUD compared to never users (Benshushan et al., 2002; Castellsague et al., 1993; Hill et al., 1997; Parazzini et al., 1994a; Salazar-Martinez et al., 1999; Sturgeon et al., 1997; The Centers for Disease Control, Cancer and Steroid Hormone Study, 1987), with the reduced risk persisting for many years after use (Castellsague et al., 1993; Hill et al., 1997; Parazzini et al., 1994a; Sturgeon et al., 1997). The remaining two found no alteration in risk (Rosenblatt et al., 1996; Shu et al., 1991). Biologically, it is unclear why IUDs may reduce endometrial cancer risk. Theories include a reduction in abnormal, precancerous cells through the IUD-induced inflammatory process (Moyer, 1971), more complete shedding of the endometrial lining (Guillebraud et al., 1976), and a possible reduction in hormone receptors (De Castro, 1986).

Abnormal Glucose Tolerance and Diabetes Mellitus In two cohorts of women with diabetes mellitus, the incidence of endometrial cancer was 40% to 80% higher than that among women in general (Weiderpass et al., 1997; Wideroff et al., 1997), but two other cohort studies with limited statistical power to detect a small excess in risk found no elevation (Kessler, 1970; Ragozzino et al., 1982). Diabetes mellitus has been associated with endometrial cancer in most other observational studies (Brinton et al., 1992; Elwood et al., 1977; Gray et al., 1977; Hoogerland et al., 1978; Hulka et al., 1980; Inoue et al., 1994; Jelovsek et al., 1980; Rubin et al., 1990; Shapiro et al., 1980; Spengler et al., 1981) (Hachisuga et al., 1998; La Vecchia et al., 1994; Maatela et al., 1994; O’Mara et al., 1985; Olson et al., 1997; Parazzini et al., 1999; Salazar-Martinez et al., 2000; Shoff, 1998; Weiderpass et al., 2000b), but not all (Goodman et al., 1997a; Horwitz, 1978; Kelsey et al., 1982; Mack et al., 1976; Pettersson et al., 1985a). Obesity increases both the risk of endometrial cancer and the risk of diabetes making it possible that diabetes is merely a surrogate for obesity. However, most studies that account for body weight or body mass still find at least a 2-fold elevation in risk (Brinton et al., 1992; Elwood et al., 1977; Hachisuga et al., 1998; Inoue et al, 1994; La Vecchia et al., 1994; Parazzini et al., 1999; Salazar-Martinez et al., 2000), although a few did not (Goodman et al., 1997a; Kelsey et al., 1982). Furthermore, while the results of several studies suggest that an elevated risk or the highest elevation in risk may be limited to obese diabetic women (BMI roughly ≥30 kg/m2) (Anderson et al., 2001; Salazar-Martinez et al., 2000; Shoff, 1998; Wideroff et al., 1997), there are also reports of modest elevations in risk for non-

1035

Endometrial Cancer obese diabetic women (Salazar-Martinez et al., 2000; Shoff, 1998; Weiderpass et al., 1997; Weiderpass et al., 2000b; Wideroff et al., 1997). If there is a body mass-independent elevation in endometrial cancer risk among diabetic women, then this suggests that there is a biological mechanism related to diabetes, and not obesity per se, that increases risk. The high insulin levels that are characteristic of women with type 2 diabetes may increase endometrial mitogenesis directly (Nagamani, 1998). High insulin levels also may have an indirect effect by inhibiting expression of insulin-like growth factor binding protein 1 (the main binding protein that regulates activity of insulin-like growth factor-I [IGF-I] in the endometrium), resulting in increased IGF-I mitogenic effects (Rutanen et al., 1993). This is an area of ongoing research.

Hypertension and Gallbladder Disease Although a relationship between hypertension and endometrial cancer has long been hypothesized, results from epidemiologic studies have been inconsistent. Few studies have accounted for age and weight or BMI, factors associated both with endometrial cancer incidence and hypertension (Ascherio et al., 1996). Amongst such studies, five reported a 50 to 100% elevation in risk (Elwood et al., 1977; Hachisuga et al., 1998; Salazar-Martinez et al., 2000; Soler et al., 1999; Weiss, Sayvetz, 1980), two reported a suggested elevation (Inoue et al., 1994; Pettersson et al., 1985a), but four others found no alteration in risk (Brinton et al., 1992; Goodman et al., 1997b; Kelsey et al., 1982; Weiderpass et al., 2000b) for hypertensive relative to normotensive women. Although a history of gallbladder disease was related to an elevated risk among women with and without exposure to exogenous estrogens in one study (Mack et al., 1976), another study that adjusted for HRT and OC use, as well as proxy measures of endogenous estrogen exposure (i.e., weight and parity) (Brinton et al., 1992), found no elevation in risk. Thus, the association of estrogens with both conditions may confound the link between gallbladder disease and endometrial cancer.

Diet, Alcohol Consumption On average, women who reside in countries in which high-fat diets are prevalent generally have relatively high rates of corpus cancer (Armstrong, Doll, 1975), but evidence of such an association on an individual level is mixed. Using various indices of dietary fat intake, several studies report that endometrial cancer risk is increased with higher levels of fat intake after adjustment for weight or body mass (Goodman et al., 1997a; Jain et al., 2000a; La Vecchia et al., 1986b; Levi et al., 1993a; Littman et al., 2001a; Potischman et al., 1993; Shu et al., 1993b) whereas others do not (Barbone et al., 1993; Jain et al., 2000c; McCann et al., 2000; Tzonou et al., 1996; Zheng et al., 1995).

While most studies find at least the suggestion of a reduced risk with higher consumption of dietary fiber (Barbone et al., 1993; Goodman et al., 1997b; Jain et al., 2000a; Littman et al., 2001a; McCann et al., 2000), vegetables (Goodman et al., 1997b; Jain et al., 2000a; La Vecchia et al., 1986b; Levi et al., 1993a; Littman et al., 2001a; Tzonou et al., 1996) or fruits (Goodman et al., 1997a; La Vecchia et al., 1986b; Levi et al., 1993a; Littman et al., 2001a) or vegetable and fruit related micronutrients (Barbone et al., 1993; McCann et al., 2000), the results are not entirely consistent (Jain et al., 2000c; Kasum et al., 2001; Potischman et al., 1993; Shu et al., 1993b; Terry et al., 2002a). In general, other evidence for a relationship between types of food, macronutrients, and micronutrients in the diet and endometrial cancer risk is inconsistent and these relationships are presently unresolved. The European Prospective Investigation into Cancer and Nutrition (EPIC) study may help to clarify some of these potential relationships (International Agency for Research on Cancer, World Health Organization, 1997). In EPIC cancer incidence is being monitored in over 400,000 people in nine countries who are providing detailed information on diet as well as providing biological samples (plasma, serum, lymphocytes and erythrocytes) for future biochemical and molecular studies. Apart from one study that found an elevated risk for women who drank more than two alcoholic beverages a day relative to non-drinkers (Parazzini et al., 1995a), most studies find that alcohol consumption is unrelated to endometrial cancer risk, especially among postmenopausal women (Austin et al., 1993; Gapstur et al., 1993; Goodman et al., 1997a; Hirose et al., 1996; Jain et al., 2000c; Littman et al., 2001a; Newcomb et al., 1997; Swanson et al., 1993). The results of four studies that evaluated younger women (approximately 55 years of age or less) suggest that those with the highest alcohol consumption (ranging from 4 to >14 drinks/week) may have a modest reduction in risk compared with non-drinkers (Newcomb et al., 1997; Parazzini et al., 1995a; Swanson et al., 1993; Webster et al., 1989).

Cigarette Smoking Most studies of endometrial cancer that have ascertained cigarette smoking behavior note a reduced risk among smokers (Table 53–4 and (Austin et al., 1993; Baron et al., 1986; Elliott et al., 1990; Folsom et al., 1989; Hirose et al., 1996; Olson et al., 1997; Zelmanowicz et al., 1998)), especially among postmenopausal women. Additionally, one study reported that a longer duration of postmenopausal smoking, but not premenopausal smoking, was associated with a reduction in risk (Weiderpass, Baron, 2001). Thus, if cigarette smoking does prevent some women from developing endometrial cancer, it is likely to do so in ways other than by reducing ovarian production of estrogens. Although not entirely consistent (Newcomer et al., 2001; Parazzini et al., 1994a; Weir et al., 1994), the negative association is particularly strong in recent smokers (Austin et al., 1993; Brinton et al., 1993a;

Table 53–4. Epidemiologic Studies of Cigarette Smoking and Endometrial Cancer by Menopausal Status Relative risk Study Weiss, Sayvetz (1980) Smith et al. (1984) Tyler et al. (1985)b Franks et al. (1987)b Lesko et al. (1985) Stockwell (1987) Parazzini et al. (1995b) Lawrence et al. (1987) Koumantaki et al. (1989) Brinton et al. (1993a) Weir et al. (1994) Weiderpass, Baron (2001) Terry et al. (2002b)

Measure of Smoking Ever vs never Ever vs never Ever vs never Postmenopausal smoking vs never ≥25 cigarettes/day vs never >40 cigarettes/day vs never ≥20 cigarettes/day vs never Current vs never Current vs never Current vs never Current vs never Current vs never >20 cigarettes/day vs never

Premenopausal Women a

0.5 1.3 1.1 — 0.9 1.9c 0.6 0.6 2.3 0.5 2.8 — 0.7

Source: Noel Weiss and Lynda Voigt, personal communication. b Centers for Disease Control Cancer and Steroid Hormone Study; the study by (Franks et al., 1987) is a subset of the study of (Tyler et al., 1985). c Women <50 and ≥50 years of age.

Postmenopausal Women 0.4 0.4 0.8 0.5 0.5 0.4c 0.6 0.6 0.2 0.4 0.8 0.6 0.6

Table 53–5. Selected Results of Case-control Studies of Endometrial Cancer and Polymorphic Genes involved in Estrogen Biosynthesis, Catabolism, and Response Gene

Polymorphism

No. of Cases

No. of Controls

Comparison

T to C change in promoter region, “A2”

50 184

51 554

51

391

A2/A2 vs. any A1 A1/A2 vs. A1/A1 A2/A2 vs. A1/A1 A1/A2 vs. A1/A1

182 110

A2/A2 vs. A1/A1 (Among ever users of estrogens) A2/A2 vs. any A1 ≥1 12 repeat alleles vs. none

Odds Ratio (95% CI)

estrogen biosynthesis Cytochrome P-450 17 (CYP17)

Cytochrome P-450 19 (CYP19 or aromatase)

(TTTA)n in intron 4

114 85

estrogen catabolism Cytochrome P-450 1A1 (CYP1A1 or ary1 hydrocarbon hydroxylase)

80

60

≥1 m1 variant vs. no m1 variant

3.7 (1.2–13.3) (Esteller et al., 1997a)

80 43 80 113

60 36 60 100

≥1 m2 variant vs. no m2 variant ≥1 m2 variant vs. no m2 variant ≥1 m4 variant vs. no m4 variant ≥1 T vs. C/C

3.7 (1.2–13.3) (Esteller et al., 1997a) 0.5 (0.1–2.0) (Olson et al., 2000) 6.4 (2.0–26.5) (Esteller et al., 1997a) 1.0 (0.8, 1.7) (Sasaki et al., 2003)

113 113 65 113 74 113

100 100 81 100 76 100

≥1 Gly vs. Arg/Arg ≥1 Ser vs. Ala/Ala Ser/Ser vs. Ala/Ala ≥1 Val vs. Leu/Leu Val/Val vs. Leu/Leu ≥1 T vs. C/C

(0.8, 1.3) (Sasaki et al., 2003) 2.3 (1.6, 3.3) (Sasaki et al., 2003) 3.3 (1.4, 8.0) (Sasaki et al., 2003) 1.6 (1.1, 2.3) (Sasaki et al., 2003) 2.5 (1.1, 5.7) (Sasaki et al., 2003) 1.1 (0.9, 1.4) (Sasaki et al., 2003)

Ala119Ser in exon 2 Leu432Val in exon 3

57

58

4.7 (2.2, 9.7) (Sasaki et al., 2003)

Val158Met in exon 4

49

29

Any 119Ser and Any 432Val vs. 119Ala/119Ala and 432Leu/432Leu ≥1 Met vs. Val/Val

(TA)n in promoter region

261

380

PvuII RFLP, C to T change in intron 1, p allele contains restriction site

261

380

m1; T6235C in 3¢ non-coding region, “MspI ” m2; Ile462Val in exon 7

Cytochrome P-450 1B1 (CYP1B1)

m4; Thr461Asn in exon 7 C to T change in intron 1 Arg48Gly in exon 2 Ala119Ser in exon 2 Leu432Val in exon 3 Silent C to T change, codon 449 in exon 3

Combined: CYP1B1 Catechol-O-Methyltransferase (COMT)

estrogen response Estrogen Receptor Alpha (ERa)

≥1 11 repeat alleles vs. none

1.9 (0.5–6.9) (Olson et al., 1999b) 0.9 (0.6–1.3) (Haiman et al., 2001) 0.4 (0.2–0.8) (Haiman et al., 2001) 0.4 (0.1–1.1) (Feigelson et al., 2001; Haiman et al., 2001) 0 (Haiman et al., 2001; McKean-Cowdin et al., 2001) 0.5 (0.3–0.9) (Berstein et al., 2002) 5.6 (1.1–39.4)a (Berstein et al., 2001) 1.9 (1.0–3.6)a (Berstein et al., 2001)

XbaI RFLP in intron 1, x allele contains restriction site

113

200

261

380

<19/≥19 repeats vs. ≥19/≥19 repeats <19/<19 repeats vs. ≥19/≥19 repeats Pp vs. pp

1.2 (0.8–2.0) (Weiderpass et al., 2000c) 1.6 (0.9–2.7) (Weiderpass et al., 2000c) 1.0 (0.6–1.4) (Weiderpass et al., 2000c)

PP vs. pp

0.7 (0.4–1.2) (Weiderpass et al., 2000c) 1.0 (0.8–1.1) (Sasaki et al., 2002) 0.9 (0.6–1.4) (Sasaki et al., 2002) 0.8 (0.5–1.1) (Weiderpass et al., 2000c)

Pp vs. pp PP vs. pp Xx vs. xx XX vs. xx

Human Progesterone Receptor (hPR)

a

Silent T to C change, codon 10

113

200

T/C vs. T/T

0.6 (0.3–1.1) (Weiderpass et al., 2000c) 0.7 (0.5–0.9) (Sasaki et al., 2002)

Silent C to T change, codon 243 in exon 3

113

200

C/C vs. T/T C/T vs. C/C

0.4 (0.2–0.9) (Sasaki et al., 2002) 1.2 (0.5–2.8) (Sasaki et al., 2002)

Silent C to G change, codon 325 in exon 4

113

200

T/T vs. C/C C/G vs. C/C

— (Sasaki et al., 2002) 1.0 (0.8–1.2) (Sasaki et al., 2002)

Silent G to A change, codon 594 in exon 8

113

200

G/G vs. C/C G/A vs. G/G

0.9 (0.7–1.2) (Sasaki et al., 2002) 1.1 (0.8–1.6) (Sasaki et al., 2002)

+44 C to T change in promoter region +331 G to A change in promoter region

187

397

A/A vs. G/G ≥1 T vs. C/C

1.2 (0.6–2.4) (Sasaki et al., 2002) 0.9 (0.5–1.6) (De Vivo et al., 2002)

187

397

≥1 A vs. G/G

1.9 (1.1–3.3) (De Vivo et al., 2002)

Calculated by the present authors.

1036

1.6 (0.6–4.2) (Olson et al., 2000)

Endometrial Cancer Elliott et al., 1990; Folsom et al., 1989; Lawrence et al., 1987; Lesko et al., 1985; Terry et al., 2002b; Tyler et al., 1985; Weiderpass, 2001; Zelmanowicz et al., 1998). Most studies have not found differences in blood levels of estradiol or estrone associated with smoking (Austin et al., 1993; Friedman et al., 1987; Jensen et al., 1985), but there may be differential metabolism of these compounds among smokers that favors the 2-hydroxylation pathway producing a metabolite of relatively low estrogenic activity (Key et al., 1996; Michnovicz et al., 1986). Similarly, higher levels of circulating progesterone, as well as a higher ratio of progesterone to estrogen among smokers (Friedman et al., 1987), may also act to diminish net estrogenic effects. These findings are consistent with the observations that smoking is related to an increased risk of osteoporosis (Daniell, 1976; Williams et al., 1982) and an earlier age at menopause (Kaufman et al., 1980b; Willett et al., 1983), both of which are associated with decreased estrogenic activity.

Environmental Chemical Exposure No strong or consistent associations were noted in the two case control studies that have assessed lipid-adjusted serum levels of various organochlorines and endometrial cancer risk (Sturgeon et al., 1998; Weiderpass et al., 2000a).

Genetic Predisposition Only a handful of studies have evaluated the risk of endometrial cancer associated with a history of endometrial cancer in one or more firstdegree relatives. Two studies that included women up to 75 years of age reported modest (50 to 90%) elevations in risk with a positive family history (Kelsey et al., 1982; Parazzini et al., 1994b), although another found no elevation in risk (Olson et al., 1999a). The results of three other studies restricted to women less than 50 or 55 years of age at diagnosis observed a stronger association (Gruber, Thompson, 1996; Henderson et al., 1983; Parslov et al., 2000). At the present time, it is unknown if these associations are due more to shared genes or to the shared environment and behaviors among female family members. It also is possible that germline allelic variants in tumor-suppressor genes, genes encoding DNA repair/methylation enzymes, or genes encoding enzymes that metabolize carcinogens or procarcinogens (such as compounds found in food or tobacco smoke) may place women at greater or lesser risk of endometrial cancer. An elevated endometrial cancer risk has been reported for women with a family history of colorectal cancer (Gruber, 1996); the same germline mutations in genes encoding DNA mismatch repair enzymes important in the familial aggregation of colorectal cancer (Marra, 1995) may also influence the development of a small percentage of endometrial carcinomas. Although this area is just beginning to be explored, women carrying variant p53 alleles (Esteller et al., 1997a; Peller et al., 1999), variant alleles for a protein (WAF1) that is a downstream mediator of p53 (Hachiya et al., 1999), variant methylenetetrahydrofolate reductase alleles (Esteller et al., 1997b), or variant alleles for an enzyme that degrades extracellular matrix for tumor invasion (Nishioka et al., 2000) may have an elevated risk of endometrial cancer relative to women with the respective wildtype alleles. Although the functional significance of the examples indicated above is unknown at the present time, the investigation of allelic variants and their expression is sure to continue and expand as array technology for efficient testing is more fully developed. Given the strong evidence for a causal role of hormones in the development of endometrial cancer, it is plausible that allelic variants regulating enzymes involved in hormone biosynthesis and catabolism, as well as variants in hormone receptors, may be related to endometrial cancer. Results from studies that have been done to date are summarized in Table 53–5. The most consistent results, particularly in premenopausal women, are found for a variant cytochrome P-450 (CYP) 17 allele (Sharp et al., 2004). CYP17 is involved in the biosynthesis of precursors of estradiol and mediates conversion of progesterone to androstenedione. Several studies indicate that women who carry a variant CYP 17 allele have a decreased risk of endometrial cancer relative to women who carry no variant allele, especially among women who are homozygous for the variant (Berstein et al.,

1037

2002; Haiman et al., 2001; McKean-Cowdin et al., 2001). While the functional significance of the variant allele is not known with certainty, it occurs in the promoter region of the gene that modulates CYP17 expression (Feigelson et al., 1997). Women with endometrial cancer may also be more likely to carry a variant androgen receptor allele (with more CAG repeats in the N-terminus domain that reduces transcriptional activity); fewer receptors may lead to a reduction in the anti-proliferative effects that androgens have on endometrial cells (Sasaki et al., 2003; Yaron et al., 2001). Research on the activity and function of all these allelic variants in the development of endometrial cancer is in its infancy. Active research continues in this area and ultimately may help to identify a genetic basis for altered endometrial cancer susceptibility.

DISEASE PREVENTION Our knowledge of identified risk factors for endometrial cancer indicates that an impact can be made in the primary prevention of this disease by minimizing the use of unopposed estrogens. It is highly likely that the sharp decline in the use of this hormone regimen has been responsible for the sharp decline in the occurrence of endometrial cancer in the United States since the mid-1970s. In terms of other modifiable risk factors, a decrease in the prevalence of extreme obesity in women would also reduce the occurrence of endometrial cancer (in addition to having many other health benefits). No other interventions have been identified that would have a clear impact on reducing endometrial cancer risk.

FUTURE DIRECTIONS The reasons for the international variation in the occurrence of endometrial cancer among women remain unclear, apart from differences in the use of postmenopausal hormones and in the prevalence of obesity. A likely explanation is that lifestyle differences or genetic variations are responsible. A challenge for future investigations will be to determine if distinct aspects of lifestyle (e.g., dietary factors, exercise, or occupation) account for all or part of this variation beyond any influence they may have on the prevalence of obesity. The incidence of endometrial cancer associated with hormone replacement therapy among postmenopausal women will continue to be an active area of research as the means of providing replacement therapy changes over time. This research will include continued evaluation of estrogens supplemented with cyclic or continuous progestogens and the use of low dosage estrogens alone (e.g., 0.3 mg/day of esterified estrogen). Long-term evaluations of selective estrogen receptor modulators (SERMs), such as raloxifene, that may provide some of the benefits of estrogen replacement therapy without inducing endometrial proliferation are also needed. Nonetheless, it will remain a challenge to identify hormones and combinations of hormones that minimize any increase in the occurrence of endometrial cancer while also minimizing any increases in risk for other long-term health outcomes such as breast cancer (WHI Writing Group, 2002). Finally, research on genetic susceptibilities for endometrial cancer may ultimately allow us to identify those women who can safely take unopposed estrogens without sustaining an elevated risk of endometrial cancer. References Albrektsen G, Heuch I, Tretli S, Kvale G. 1995. Is the risk of cancer of the corpus uteri reduced by a recent pregnancy? A prospective study of 765,756 Norwegian women. Int J Cancer 61:485–490. Anderson KE, Anderson E, Mink PJ, et al. 2001. Diabetes and endometrial cancer in the Iowa women’s health study. Cancer Epidemiol Biomarkers Prev 10:611–616. Antunes CM, Strolley PD, Rosenshein NB, et al. 1979. Endometrial cancer and estrogen use. Report of a large case-control study. N Engl J Med 300:9–13. Archer DF, Furst K, Tipping D, Dain M-P, Vandepol C, for the CombiPatch Study Group. 1999. A randomized comparison of continuous combined

1038

PART IV: CANCER BY TISSUE OF ORIGIN

transdermal delivery of estradiol-norethindrone acetate and estradiol alone for menopause. Obstet Gynecol 94:498–503. Armstrong B, Doll R. 1975. Environmental factors and cancer incidence and mortality in different countries, with special reference to dietary practices. Int J Cancer 15:617–631. Armstrong BK, Ray RM, Thomas DB. 1988. Endometrial cancer and combined oral contraceptives. The Who Collaborative Study of Neoplasia and Steroid Contraceptives. Int J Epidemiol 17:263–269. Ascherio A, Hennekens C, Willet WC, et al. 1996. Prospective study of nutritional factors, blood pressure, and hypertension among US women. Hypertension 27:1065–1072. Austin DF, Roe KM. 1982. The decreasing incidence of endometrial cancer: public health implications. Am J Public Health 72:65–68. Austin H, Austin JM, Jr., Partridge EE, Hatch KD, Shingleton HM. 1991. Endometrial cancer, obesity, and body fat distribution. Cancer Res 51:568–572. Austin H, Drews C, Partridge EE. 1993. A case-control study of endometrial cancer in relation to cigarette smoking, serum estrogen levels, and alcohol use. Am J Obstet Gynecol 169:1086–1091. Baanders-van Halewyn EA, Blankenstein MA, Thijssen JH, de Ridder CM, de Waard F. 1996. A comparative study of risk factors for hyperplasia and cancer of the endometrium. Eur J Cancer Prev 5:105–112. Barbone F, Austin H, Partridge EE. 1993. Diet and endometrial cancer: a casecontrol study. Am J Epidemiol 137:393–403. Baron JA, Byers T, Greenberg ER, Cummings KM, Swanson M. 1986. Cigarette smoking in women with cancers of the breast and reproductive organs. J Natl Cancer Inst 77:677–680. Benshushan A, Paltiel O, Brzezinski A, et al. 2001. Ovulation induction and risk of endometrial cancer: a pilot study. Eur J Obstet Gynecol Reprod Biol 98:53–57. Benshushan A, Paltiel O, Rojansky N, Brzezinski A, Laufer N. 2002. IUD use and the risk of endometrial cancer. Eur J Obstet Gynecol Reprod Biol 105:166–169. Beral V, Hannaford P, Kay C. 1988. Oral contraceptive use and malignancies of the genital tract. Lancet 2:1331–1335. Beresford SAA, Weiss NS, Voigt LF, McKnight B. 1997. Risk of endometrial cancer in relation to use of oestrogen combined with cyclic progestagen in postmenopausal women. Lancet 349:458–461. Bergman L, Beelen ML, Gallee MP, Hollema H, Benraadt J, van Leeuwen EE. 2000. Risk and prognosis of endometrial cancer after tamoxifen for breast cancer. Lancet 356:881–887. Bernstein L, Deapen D, Cerhan JR, et al. 1999. Tamoxifen therapy for breast cancer and endometrial cancer risk. J Natl Cancer Inst 91: 1654–1662. Berstein LM, Imyanitov EN, Gamajunova VB, et al. 2002. CYP17 genetic polymorphism in endometrial cancer: are only steroids involved? Cancer Lett 180:47–53. Berstein LM, Imyanitov EN, Suspitsin EN, et al. 2001. CYP19 gene polymorphism in endometrial cancer patients. J Cancer Res Clin Oncol 127:135–138. Boss SM, Huster WJ, Neild JA, Glant MD, Eisenhut CC, Draper MW. 1997. Effects of raloxifene hydrochloride on the endometrium of postmenopausal women. Am J Obstet Gynecol 177:1458–1464. Brinton LA, Barrett RJ, Berman ML, Mortel R, Twiggs LB, Wilbanks GD. 1993a. Cigarette smoking and the risk of endometrial cancer. Am J Epidemiol 137:281–291. Brinton LA, Berman ML, Mortel R, et al. 1992. Reproductive, menstrual, and medical risk factors for endometerial cancer: results from a case-control study. Am J Obstet Gynecol 167:1317–1325. Brinton LA, Hoover RN, and The Endometrial Cancer Collaborative Group. 1993b. Estrogen replacement therapy and endometrial cancer risk: unresolved issues. Obstet Gynecol 81:265–271. Brynhildsen J, Hammar M. 2002. Low dose transdermal estradiol/norethisterone acetate treatment over 2 years does not cause endometrial proliferation in postmenopausal women. Menopause 9:137–144. Buring JE, Bain CJ, Ehrmann RL. 1986. Conjugated estrogen use and risk of endometrial cancer. Am J Epidemiol 124:434–441. Burke TW, Tortolero-Luna G, Malpica A, et al. 1996. Endometrial hyperplasia and endometrial cancer. Obstet Gynecol Clin North Am 23:411– 456. Castellsague X, Thompson WD, Dubrow R. 1993. Intra-uterine contraception and the risk of endometrial cancer. Int J Cancer 54:911–916. Castellsague X, Thompson WD, Dubrow R. 1996. Tubal sterilization and the risk of endometrial cancer. Int J Cancer 65:607–612. Cauley JA, Norton L, Lippman ME, et al. 2001. Continued breast cancer risk reduction in postmenopausal women treated with raloxifene: 4-year results from the MORE trial. Multiple outcomes of raloxifene evaluation. Breast Cancer Res Treat 65:125–134.

Cerin A, Heldaas K, Moeller B, the Scandinavian LongCycle Study Group. 1996. Adverse endometrial effects of long-cycle estrogen and progestogen replacement therapy (letter). N Engl J Med 334:668–669. Chamlian DL, Taylor HB. 1970. Endometrial hyperplasia in young women. Obstet Gynecol 36:659–666. Chu J, Schweid AI, Weiss NS. 1982. Survival among women with endometrial cancer: a comparison of estrogen users and nonusers. Am J Obstet Gynecol 143:569–573. Clisham PR, de Ziegler D, Lozano K, Judd HL. 1991. Comparison of continuous versus sequential estrogen and progestin therapy in postmenopausal women. Obstet Gynecol 77:241–246. Collins J, Donner A, Allen LH, Adams O. 1980. Oestrogen use and survival in endometrial cancer. Lancet 2:961–964. Cook LS, Weiss NS, Schwartz SM, et al. 1995. A population-based study of tamoxifen therapy and subsequent ovarian, endometrial, and breast cancers. J Natl Cancer Inst 87:1359–1364. Corson SL, Richart RM, Caubel P, Lim P. 1999. Effect of a unique constantestrogen, pulsed-progestin hormone replacement therapy containing 17Bestradiol and norgestimate on endometrial histology. Int J Fertil 44: 279–285. Coulam CB, Annegers JF, Kranz JS. 1983. Chronic anovulation syndrome and associated neoplasia. Obstet Gynecol 61:403–407. Coulter A, McPherson K, Vessey M. 1988. Do British women undergo too many or too few hysterectomies? Soc Sci Med 27:987–994. Cushing KL, Weiss NS, Voigt LF, McKnight B, Beresford SA. 1998. Risk of endometrial cancer in relation to use of low-dose, unopposed estrogens. Obstet Gynecol 91:35–39. Cutler BS, Forbes AP, Ingersoll FM, Scully RE. 1972. Endometrial carcinoma after stilbestrol therapy in gonadal dysgenesis. N Engl J Med 287: 628–631. Damon A. 1960. Host factors in cancer of the breast and uterine cervix and corpus. J Natl Cancer Inst 24:483–516. Daniell HW. 1976. Osteoporosis of the slender smoker. Vertebral compression fractures and loss of metacarpal cortex in relation to postmenopausal cigarette smoking and lack of obesity. Arch Intern Med 136:298–304. Davidson BJ, Gambone JC, Lagasse LD, et al. 1981. Free estradiol in postmenopausal women with and without endometrial cancer. J Clin Endocrinol Metab 52:404–408. De Castro A, Gonzalez-Gancedo P. 1986. The effect of copper ions in vivo on specific hormonal endometrial receptors. Adv Contracept 2:399– 404. De Vivo I, Huggins GS, Hankinson SE, et al. 2002. A functional polymorphism in the promoter of the progesterone receptor gene associated with endometrial cancer risk. Proc Natl Acad Sci U S A 99:12263–12268. Deligdisch L, Holinka CF. 1987. Endometrial carcinoma: two diseases? Cancer Detect Prev 10:237–246. Dicker RC, Scally MJ, Greenspan JR, et al. 1982. Hysterectomy among women of reproductive age: trends in the United States, 1970–1978. JAMA 248:323–327. Diddle AW. 1952. Granulosa and theca cell ovarian tumors: Prognosis. Cancer 5:215–228. Division of Cancer Prevention and Control, National Cancer Institute. 1988. 1987 Annual Cancer Statistics Review: including cancer trends 1950–1985. NIH Publication No. 88-2789. Dockerty MB, Lovelady SB, Foust GT, Jr. 1951. Carcinoma of the corpus uteri in young women. Am J Obstet Gynecol 61:966–981. Doll R, Kinlen LJ, Skegg DC. 1976. Letter: Incidence of endometrical carcinoma. Br Med J 1:1071–1072. Dosemeci M, Hayes RB, Vetter R, et al. 1993. Occupational physical activity, socioeconomic status, and risks of 15 cancer sites in Turkey. Cancer Causes Control 4:313–321. Doyle P, Maconochie N, Beral V, Swerdlow AJ, Tan SL. 2002. Cancer incidence following treatment for infertility at a clinic in the UK. Hum Reprod 17:2209–2213. Edman CD, MacDonald PC. 1978. Effect of obesity on conversion of plasma androstenedione to estrone in ovulatory and anovulator young women. Am J Obstet Gynecol 130:456–461. Elliott EA, Matanoski GM, Rosenshein NB, Grumbine FC, Diamond EL. 1990. Body fat patterning in women with endometrial cancer. Gynecol Oncol 39:253–258. Elwood JM, Boyes DA. 1980. Clinical and pathological features and survival of endometrial cancer patients in relation to prior use of estrogens. Gynecol Oncol 10:173–187. Elwood JM, Cole P, Rothman KJ, Kaplan SD. 1977. Epidemiology of endometrial cancer. J Natl Cancer Inst 59:1055–1060. Escobedo LG, Lee NC, Peterson HB, Wingo PA. 1991. Infertility-associated endometrial cancer risk may be limited to specific subgroups of infertile women. Obstet Gynecol 77:124–128.

Endometrial Cancer Esteller M, Garcia A, Martinez-Palones JM, Xercavins J, Reventos J. 1997a. Germ line polymorphisms in cytochrome-P450 1A1 (C4887 CYP1A1) and methylenetetrahydrofolate reductase (MTHFR) genes and endometrial cancer susceptibility. Carcinogenesis 18:2307–2311. Esteller M, Garcia A, Martinez-Palones JM, Xercavins J, Reventos J. 1997b. Susceptibility to endometrial cancer: influence of allelism at p53, glutathione S-transferase (GSTM1 and GSTT1) and cytochrome P-450 (CYP1A1) loci. Br J Cancer 75:1385–1388. Ettinger B, Selby J, Citron JT, Vangessel A, Ettinger VM, Hendrickson MR. 1994. Cyclic hormone replacement therapy using quarterly progestin. Obstet Gynecol 83:693–700. Ewertz M, Schou G, Boice JD. 1988. The joint effect of risk factors on endometrial cancer. Eur J Cancer Clin Oncol 24:189–194. Farhi DC, Nosanchuk J, Silverberg SG. 1986. Endometrial adenocarcinoma in women under 25 years of age. Obstet Gynecol 68:741–745. Feigelson HS, Coetzee GA, Kolonel LN, Ross RK, Henderson BE. 1997. A polymorphism in the CYP17 gene increases the risk of breast cancer. Cancer Res 57:1063–1065. Feigelson HS, McKean-Cowdin R, Coetzee GA, Stram DO, Kolonel LN, Henderson BE. 2001. Building a multigenic model of breast cancer susceptibility: CYP17 and HSD17B1 are two important candidates. Cancer Res 61:785–789. Finkle WD, Greenland S, Miettinen OS, Ziel HK. 1995. Endometrial cancer risk after discontinuing use of unopposed conjugated estrogens (California, United States). Cancer Causes Control 6:99–102. Fisher B, Costantino JP, Redmond CK, Fisher ER, Wickerham DL, Cronin WM. 1994. Endometrial cancer in tamoxifen-treated breast cancer patients: Findings from the National Surgical Adjuvant Breast and Bowel Project (NSABP) B-14. J Natl Cancer Instit 86:27–37. Fisher B, Costantino JP, Wickerham DL, et al. 1998. Tamoxifen for prevention of breast cancer: report of the National Surgical Adjuvant Breast and Bowel Project P-1 Study. J Natl Cancer Inst 90:1371–1388. Folsom AR, Kaye SA, Potter JD, Prineas RJ. 1989. Association on incident carcinoma of the endometrium with body weight and fat distribution in older women: early findings of the Iowa Women’s Health Study. Cancer Res 49:6828–6831. Food and Drug Administration. 1976. Estrogens and endometrial cancer. FDA Drug Bulletin 6:18–20. Fornander T, Rutqvist LE, Cedermark B, et al. 1989. Adjuvant tamoxifen in early breast cancer: occurrence of new primary cancers. Lancet 1: 117–120. Franks AL, Kendrick JS, Tyler CW, Jr. 1987. Postmenopausal smoking, estrogen replacement therapy, and the risk of endometrial cancer. Am J Obstet Gynecol 156:20–23. Friedman AJ, Ravnikar VA, Barbieri RL. 1987. Serum steroid hormone profiles in postmenopausal smokers and nonsmokers. Fertil Steril 47: 398–401. Gambrell RD, Massey FM, Castaneda TA, Ugenas AJ, Ricci CA. 1979. Reduced incidence of endometrial cancer among postmenopausal women treated with progestogens. J Am Geriatrics Soc XXVII:389–394. Gapstur SM, Potter JD, Sellers TA, Kushi LH, Folsom AR. 1993. Alcohol consumption and postmenopausal endometrial cancer: results from the Iowa Women’s Health Study. Cancer Causes Control 4:323–329. Genant HK, Lucas J, Weiss S, et al. 1997. Low-Dose Esterified Estrogen Therapy. Arch Intern Med 157:2609–2615. Goodman MT, Hankin JH, Wilkens LR, et al. 1997a. Diet, body size, physical activity, and the risk of endometrial cancer. Cancer Res 57: 5077–5085. Goodman MT, Wilkens LR, Hankin JH, Lyu LC, Wu AH, Kolonel LN. 1997b. Association of soy and fiber consumption with the risk of endometrial cancer. Am J Epidemiol 146:294–306. Gordon J, Reagan JW, Finkle WD, Ziel HK. 1977. Estrogen and endometrial carcinoma. An independent pathology review supporting original risk estimate. N Engl J Med 297:570–571. Gordon MD, Ireland K. 1994. Pathology of hyperplasia and carcinoma of the endometrium. Seminars Oncol 21:64–70. Gorodeski GI, Beery R, Lunenfeld B, Geier A. 1992. Tamoxifen increases plasma estrogen-binding equivalents and has estradiol agonistic effect on histologically normal premenopausal and postmenopausal endometrium. Fert Steril 57:320–327. Grady D, Gebretsadik T, Kerlikowske K, Ernster V, Petitti D. 1995. Hormone replacement therapy and endometrial cancer risk: a meta-analysis. Obstet Gynecol 85:304–313. Grady D, Ernster VL. 1996. Endometrial cancer. In Schottenfeld D, Fraumeni JF, eds. Cancer Epidemiology and Prevention, 2nd edition. New York, Oxford University Press, pp. 1058–1089. Gray LA, Christopherson WM, Hoover RN. 1997. Estrogens and endometrial carcinoma. Obstet Gynecol 49:385–389.

1039

Green PK, Weiss NS, McKnight B, Voigt LF, Beresford SA. 1996. Risk of endometrial cancer following cessation of menopausal hormone use (Washington, United States). Cancer Causes Control 7:575–580. Gruber JS, Luciani CT. 1986. Physicians changing postmenopausal sex hormone prescribing regimens. Prog Clin Biol Res 216:325–335. Gruber SB, Thompson WD. 1996. A population-based study of endometrial cancer and familial risk in younger women. Cancer and Steroid Hormone Study Group. Cancer Epidemiol Biomarkers Prev 5:411–417. Guillebraud J, Bonnar J, Morehead J, Matthews A. 1976. Menstrual blood-loss with intrauterine devices. Lancet 1:387–390. Gusberg SB. 1947. Precursors of corpus carcinoma estrogens and adenomatous hyperplasia. Am J Obstet Gynecol 54:905–927. Gusberg SB, Hall RE. 1961. Precursors of corpus cancer III. The appearance of cancer of the endometrium in estrogenically conditioned patients. Obstet Gynecol 17:397–412. Gusberg SB, Kardon P. 1971. Proliferative endometrial response to thecagranulosa cell tumors. Am J Obstet Gynecol 111:633–643. Hachisuga T, Fukuda K, Hirakawa T, Kawarabayashi T. 2001. The effect of nulliparity on survival in endometrial cancer at different ages. Gynecol Oncol 82:122–126. Hachisuga T, Kaetsu A, Sugimori H, Kamura T, Tsuneyoshi M, Kawarabayashi T. 1998. Risk factors for endometrial cancer in Japanese women. Int J Gynecol Cancer 8:292–297. Hachiya T, Kuriaki Y, Ueoka Y, Nishida J, Kato K, Wake N. 1999. WAF1 genotype and endometrial cancer susceptibility. Gynecol Oncol 72: 187–192. Haiman CA, Hankinson SE, Colditz GA, Hunter DJ, De Vivo I. 2001. A polymorphism in CYP17 and endometrial cancer risk. Cancer Res 61:3955–3960. Hammond CB, Jelovsek FR, Lee KL, Creasman WT, Parker RT. 1979. Effects of long-term estrogen replacement therapy. I. Metabolic effects. Am J Obstet Gynecol 133:525–536. Hannaford PC, Kay CR. 1998. The risk of serious illness among oral contraceptive users: evidence from the RCGP’s oral contraceptive study. Br J Gen Pract 48:1657–1662. Henderson BE, Casagrande JT, Pike MC, Mack T, Rosario I, Duke A. 1983. The epidemiology of endometrial cancer in young women. Br J Cancer 47:749–756. Herrington LJ, Weiss NS. 1993. Postmenopausal unopposed estrogens: characteristics of use in relation to the risk of endometrial carcinoma. Annal Epidemiol 3:308–318. Hill DA, Weiss NS, Beresford SA. 2000. Continuous combined hormone replacement therapy and risk of endometrial cancer. Am J Obstet Gynecol 183:1456–1461. Hill DA, Weiss NS, Voigt LF, Beresford SA. 1997. Endometrial cancer in relation to intra-uterine device use. Int J Cancer 70:278–281. Hinkula M, Pukkala E, Kyyronen P, Kauppila A. 2002. Grand multiparity and incidence of endometrial cancer: a population-based study in Finland. Int J Cancer 98:912–915. Hirose K, Tajima K, Hamajima N, et al. 1999. Comparative case-referent study of risk factors among hormone-related female cancers in Japan. Jpn J Cancer Res 90:255–261. Hirose K, Tajima K, Hamajima N, et al. 1996. Subsite (cervix/endometrium)specific risk and protective factors in uterus cancer. Jpn J Cancer Res 87:1001–1009. Hirvonen E, Salmi T, Puolakka J, et al. 1995. Can progestin be limited to very third month only in postmenopausal women taking eestrogen? Maturitas 21:39–44. Hoogerland DL, Buchler DA, Crowley JJ, Carr WF. 1978. Estrogen use—risk of endometrial carcinoma. Gynecol Oncol 6:451–458. Hoover R, Fraumeni JF, Everson R, Myers MH. 1976. Cancer of the uterine corpus after hormonal treatment for breast cancer. Lancet 1:885–887. Horwitz RI, Feinstein AR. 1978. Alternative analytic methods for case-control studies of estrogens and endometrial cancer. N Engl J Med 299: 1089–1094. Howe HL. 1984. Age-specific hysterectomy and oophorectomy prevalence rates and the risks for cancer of the reproductive system. Am J Public Health 74:560–563. Hulka BS, Chambless LE, Kaufman DG, Fowler WC, Greenberg BG. 1982. Protection against endometrial carcinoma by combination-product oral contraceptives. JAMA 247:475–477. Hulka BS, Fowler WC, Kaufman DG, et al. 1980. Estrogen and endometrial cancer: cases and two control groups from North Carolina. Am J Obstet Gynecol 137:92–101. Hulley S, Furberg C, Barrett-Connor E, et al. 2002. Noncardiovascular disease outcomes during 6.8 years of hormone therapy: Heart and Estrogen/ progestin Replacement Study follow-up (HERS II). JAMA 288:58– 66.

1040

PART IV: CANCER BY TISSUE OF ORIGIN

Inoue M, Okayama A, Fujita M, Enomoto T, Tanizawa O, Ueshima H. 1994. A case-control study on risk factors for uterine endometrial cancer in Japan. Jpn J Cancer Res 85:346–350. International Agency for Research on Cancer, World Health Organization. 1997. Biennial Report 1996/1997. Jain MG, Howe GR, Rohan TE. 2000a. Nutritional factors and endometrial cancer in Ontario, Canada. Cancer Control 7:288–296. Jain MG, Rohan TE, Howe GR. 2000b. Hormone replacement therapy and endometrial cancer in Ontario, Canada. J Clin Epidemiol 53:385–391. Jain MG, Rohan TE, Howe GR, Miller AB. 2000c. A cohort study of nutritional factors and endometrial cancer. Eur J Epidemiol 16:899–905. Jelovsek FR, Hammond CB, Woodard BH, et al. 1980. Risk of exogenous estrogen therapy and endometrial cancer. Am J Obstet Gynecol 137:85–91. Jensen J, Christiansen C, Rodbro P. 1985. Cigarette smoking, serum estrogens, and bone loss during hormone-replacement therapy early after menopause. N Engl J Med 313:973–975. Jick H, Walker AM, Rothman KJ. 1980. The epidemic of endometrial cancer: a commentary. Am J Public Health 70:264–267. Jick H, Watkins RN, Hunter JR, et al. 1979. Replacement estrogens and endometrial cancer. N Engl J Med 300:218–222. Jick SS, Walker AM, Jick H. 1993a. Estrogens, progesterone, and endometrial cancer. Epidemiol 4:20–24. Jick SS, Walker AM, Jick H. 1933b. Oral contraceptives and endometrial cancer. Obstet Gynecol 82:931–935. Jordan VC. 1995. Tamoxifen: toxicities and drug resistance during the treatment and prevention of breast cancer. Annual Rev Pharmacol Toxicol 35:195–211. Kalandidi A, Tzonou A, Lipworth L, Gamatsi I, Filippa D, Trichopoulos D. 1996. A case-control study of endometrial cancer in relation to reproductive, somatometric, and life-style variables. Oncol 53:354–359. Kasum CM, Nicodemus K, Harnack LJ, Jacobs DR, Jr., Folsom AR. 2001. Whole grain intake and incident endometrial cancer: the Iowa Women’s Health Study. Nutr Cancer 39:180–186. Kaufman DW, Shapiro S, Slone D, et al. 1980a. Decreased risk of endometrial cancer among oral-contraceptive users. N Engl J Med 303:1045–1047. Kaufman DW, Slone D, Rosenberg L, Miettinen OS, Shapiro S. 1980b. Cigarette smoking and age at natural menopause. Am J Public Health 70:420–422. Kaufman RH, Reeves KO, Dougherty CM. 1976. Severe atypical endometrial changes and sequential contraceptive use. JAMA 236:923–926. Kaye SA, Folsom AR, Soler JT, Prineas RJ, Potter JD. 1991. Associations of body mass and fat distribution with sex hormone concentrations in postmenopausal women. Int J Epidemiol 20:151–156. Kelsey JL, LiVolsi VA, Holford TR, et al. 1982. A case-control study of cancer of the endometrium. Am J Epidemiol 116:333–342. Kennedy DL, Baum C, Forbes MB. 1985. Noncontraceptive estrogens and progestins: use patterns over time. Obstet Gynecol 65:441–446. Keshavarz H, Hillis SD, Kieke BA, Marchbanks PA. 2002. Hysterectomy Surveillance—United States, 1994–1999. MMWR CDC Surveill Summ 51:1–8. Kessler II. 1970. Cancer mortality among diabetics. J Natl Cancer Inst 44:673–680. Key TJ, Pike MC, Brown JB, Hermon C, Allen DS, Wang DY. 1996. Cigarette smoking and urinary oestrogen excretion in premenopausal and postmenopausal women. Br J Cancer 74:1313–1316. King RJB, Whitehead MI. 1983. Estrogen and progestin effects on epithelium and stroma from pre- and postmenopausal endometria: application to clinical studies of the climacteric syndrome. In Jasonni VM, ed. Steroids and Endometrial Cancer. New York Raven Press. Kirschner MA, Samjlik E, Drejka M, Szmal E, Schneider G, Ertel N. 1990. Androgen-estrogen metabolism in women with upper body vs lower body obesity. J Clin Endocrinol Metab 70:473–479. Kistner RW. 1973. Endometrial alterations associated with estrogen and estrogen-progestin combinations. In Norris HJ, Hertig AT, Abell MR, eds. The Uterus. Baltimore, Williams & Wilkins, pp. 227–254. Kodama S, Kase H, Tanaka K, Matsui K. 1996. Multivariate analysis of prognostic factors in patients with endometrial cancer. Int J Gynaecol Obstet 53:23–30. Koepsell TD, Weiss NS, Thompson DJ, Martin DP. 1980. Prevalence of prior hysterectomy in the Seattle-Tacoma area. Am J Public Health 70:40– 47. Koumantaki Y, Tzonou A, Koumantakis E, Kaklamani E, Aravantinos D, Trichopoulos D. 1989. A case-control study of cancer of endometrium in Athens. Int J Cancer 43:795–799. Kurman RJ, Felix J, Archer DF, Nanavati N, Arce J-C, Moyer DL. 2000. Norethinedrone acetate and estradiol-induced endometrial hyperplasia. Obstet Gynecol 96:373–379.

Kurman RJ, Kaminski PF, Norris HJ. 1985. The behavior of endometrial hyperplasia. A long-term study of “untreated” hyperplasia in 170 patients. Cancer 56:403–412. Kvale G, Heuch I, Ursin G. 1988. Reporductive factors and risk of cancer of the uterine corpus: a prospective study. Cancer Res 8:6217–6221. La Vecchia C, Decarli A, Fasoli M, et al. 1986a. Oral contraceptives and cancers of the breast and of the female genital tract. Interim results from a case-control study. Br J Cancer 54:311–317. La Vecchia C, Decarli A, Fasoli M, Gentile A. 1986b. Nutrition and diet in the etiology of endometrial cancer. Cancer 57:1248–1253. La Vecchia C, Franceschi S, Decarli A. 1984. Risk factors for endometrial cancer at different ages. J Natl Cancer Inst 73:667–671. La Vecchia C, Negri E, Franceschi S, D’Avanzo B, Boyle P. 1994. A casecontrol study of diabetes mellitus and cancer risk. Br J Cancer 70:950–953. Lacey JV, Jr., Brinto LA, Mortel R, et al. 2000. Tubal sterilization and risk of cancer of the endometrium. Gynecol Oncol 79:482–484. Lambe M, Wuu J, Weiderpass E, Hsieh CC. 1999. Childbearing at older age and endometrial cancer risk (Sweden). Cancer Causes Control 10:43– 49. Larson JA. 1954. Estrogens and endometrial cancer. Obstet Gynecol 3:551–572. Lawrence C, Tessaro I, Durgerian S, et al. 1987. Smoking, body weight, and early-stage endometrial cancer. Cancer 59:1665–1669. Le Marchand L, Wilkens LR, Mi M-P. 1991. Early-age body size, adult weight gain and endometrial cancer risk. Int J Cancer 48:807–811. Lesko SM, Rosenberg L, Kaufman DW, et al. 1985. Cigarette smoking and the risk of endometrial cancer. N Engl J Med 313:593–596. Lesko SM, Rosenberg L, Kaufman DW, et al. 1991. Endometrial cancer and age at last delivery: evidence for an association. Am J Epidemiol 133:554–559. Levi F, Franceschi S, Negri E, La Vecchia C. 1993a. Dietary factors and the risk of endometrial cancer. Cancer 71:3575–3581. Levi F, La Vecchia C, Gulie C, Franceschi S, Negri E. 1993b. Oestrogen replacement treatment and the risk of endometrial cancer: an assessment of the role of covariates. Eur J Cancer 29A:1445–1449. Levi F, La Vecchia C, Gulie C, et al. 1991. Oral contraceptives and the risk of endometrial cancer. Cancer Causes Control 2:99–103. Levi F, La Vecchia C, Negri E, Franceschi S. 1993c. Selected physical activities and the risk of endometrial cancer. Br J Cancer 67:846–851. Liang AP, Levenson AG, Layde PM, et al. 1983. Risk of breast, uterine corpus, and ovarian cancer in women receiving medroxyprogesterone injections. JAMA 249:2909–2912. Littman AJ, Beresford SA, White E. 2001a. The association of dietary fat and plant foods with endometrial cancer (United States). Cancer Causes Control 12:691–702. Littman AJ, Voigt LF, Beresford SA, Weiss NS. 2001b. Recreational physical activity and endometrial cancer risk. Am J Epidemiol 154:924–933. Lochen ML, Lund E. 1997. Childbearing and mortality from cancer of the corpus uteri. Acta Obstet Gynecol Scand 76:373–377. Lucas WE. 1974. Causal relationships between endocrine-metabolic variables in patients with endometrial carcinoma. Obstet Gynecol Surv 29:507–528. Lyon FA, Frisch MJ. 1976. Endometrial abnormalities occurring in young women on long-term sequential oral contraceptives. Obstet Gynecol 47:639–643. Lyon JL, Gardner JW. 1977. The rising frequency of hysterectomy: its effect on uterine cancer rates. Am J Epidemiol 105:439–443. Maatela J, Aromaa A, Salmi T, Pohja M, Vuento M, Gronroos M. 1994. The risk of endometrial cancer in diabetic and hypertensive patients: a nationwide record-linkage study in Finland. Annales Chirurgiae et Gynaecologiae–Supplementum 208:20–24. MacDonald PC, Edman CD, Hemsell DL, Porter JC, Siiteri PK. 1978. Effect of obesity on conversion of plasma androstenedione to estrone in postmenopausal women with and without endometrial cancer. Am J Obstet Gynecol 130:448–455. MacDonald PC, Siiteri PK. 1974. The relationship between the extraglandular production of estrone and the occurrence of endometrial neoplasia. Gynecol Oncol 2:259–263. Mack TM, Pike MC, Henderson BE, et al. 1976. Estrogens and endometrial cancer in a retirement community. N Engl J Med 294:1262–1267. Magos AL, Brincat M, Studd JW, Wardle P, Schlesinger P, O’Dowd T, 1985. Amenorrhea and endometrial atrophy with continuous oral estrogen and progestogen therapy in postmenopausal women. Obstet Gynecol 65:496–499. Mant JW, Vessey MP. 1994. Ovarian and endometrial cancers. In Cancer Surveys Volume 19: Trends in Cancer Incidence and Mortality. Imperial Cancer Research Fund. London, pp. 287–307.

Endometrial Cancer Marra G, Boland CR. 1995. Hereditary nonpolyposis colorectal cancer: the syndrome, the genes, and historical perspectives. J Natl Cancer Inst 87:1114–11125. Marrett LD, Elwood JM, Meigs JW, et al. 1978. Recent trends in the incidence and mortality of cancer of the uterine corpus in Connecticut. Gynecol Oncol 6:183–195. Mattsson LA, Samsioe G. 1985. Estrogen-progestogen replacement in climacteric women, particularly as regards a new type of continuous regimen. Acta Obstet Gynecol Scand Suppl 130:53–58. McCann SE, Freudenheim JL, Marshall JR, Brasure JR, Swanson MK, Graham S. 2000. Diet in the epidemiology of endometrial cancer in western New York (United States). Cancer Causes Control 11:965–974. McDonald TW, Annegers JF, O’Fallon WM, Dockerty MB, Malkasian GD, Kurland LT. 1977. Exogenous estrogen and endometrial carcinoma: casecontrol and incidence study. Am J Obstet Gynecol 127:572–580. McKean-Cowdin R, Feigelson HS, Pike MC, Coetzee GA, Kolonel LN, Henderson BE. 2001. Risk of endometrial cancer and estrogen replacement therapy history by CYP17 genotype. Cancer Res 61: 848–849. McPherson CP, Sellers TA, Potter JD, Bostick RM, Folsom AR. 1996. Reproductive factors and risk of endometrial cancer. The Iowa Women’s Health Study. Am J Epidemiol 143:1195–1202. Michnovicz JJ, Hershcopf RJ, Naganuma H, Bradlow HL, Fishman J. 1986. Increased 2-hydroxylation of estradiol as a possible mechanism for the anti-estrogenic effect of cigarette smoking. N Engl J Med 315: 1305–1309. Modan B, Ron E, Lerner-Geva L, et al. 1998. Cancer incidence in a cohort of infertile women. Am J Epidemiol 147:1038–1042. Mogren I, Stenlund H, Hogberg U. 2001. Long-term impact of reproductive factors on the risk of cervical, endometrial, ovarian and breast cancer. Acta Oncol 40:849–854. Moradi T, Weiderpass E, Signorello LB, Persson I, Nyren O, Adami HO. 2000. Physical activity and postmenopausal endometrial cancer risk (Sweden). Cancer Causes Control 11:829–837. Moyer DL, Mishell DR. 1971. Reactions of human endometrium to the intrauterine foreign body. II. Long-term effects on the endometrial histology and cytology. Am J Obstet Gynecol 111:66–80. Muir C, Waterhouse J, Mack T, Powell J, Whelan S. 1987. Cancer Incidence in Five Continents (Vol 5). IARC, Lyon, France. Nachtigall LE, Nachtigall RH, Nachtigall RD. 1979. Estrogen replacement therapy. II. A prospective study in the relationship to carcinoma and cardiovascular and metabolic problems. Obstet Gynecol 54:74–79. Nagamani M, Stuart CA. 1998. Specific binding and growth-promoting activity of insulin in endometrial cancer cells in culture. Am J Obstet Gynecol 179:6–12. Newcomb PA, Trentham-Dietz A. 2000. Breast feeding practices in relation to endometrial cancer risk, USA. Cancer Causes Control 11:663–667. Newcomb PA, Trentham-Dietz A, Storer BE. 1997. Alcohol consumption in relation to endometrial cancer risk. Cancer Epidemiol Biomarkers Prevent 6:775–778. Newcomer LM, Newcomb PA, Trentham-Dietz A, Storer BE. 2001. Hormonal risk factors for endometrial cancer: modification by cigarette smoking (United States). Cancer Causes Control 12:829–835. Nishioka Y, Kobayashi K, Sagae S, et al. 2000. A single nucleotide polymorphism in the matrix metalloproteinase-1 promoter in endometrial carcinomas. Jpn J Cancer Res 91:612–615. Nyholm HC, Nielsen AL, Lyndrup J, Dreisler A, Hagen C, Haug E. 1993. Plasma oestrogens in postmenopausal women with endometrial cancer. Br J Obstet Gynaecol 100:1115–1119. O’Mara BA, Byers T, Schoenfeld E. 1985. diabetes mellitus and cancer risk: a multisite case-control study. J Chronic Dis 38:435–441. Olson JE, Sellers TA, Anderson KE, Folsom AR. 1999a. Does a family history of cancer increase the risk for postmenopausal endometrial carcinoma? A prospective cohort study and a nested case-control family study of older women. Cancer 85:2444–2449. Olsen SH, Elahi A, Roy P, Berwick M. 1999b. CYP17 and COMT genotypes in endometrial cancer (abstract #26). Am J Epidemiol 149: SER Abstracts S7, p. 57. Olson SH, Elahi A, Tang G, et al. 2000. Polymorphisms in CYP1A1 and CYP1B1 in endometrial cancer (abstract #149). Proc Am Assoc Cancer Res 41:23–23. Olson SH, Trevisan M, Marshall JR, et al. 1995. Body mass index, weight gain, and risk of endometrial cancer. Nutr Cancer 23:141–149. Olson SH, Vena JE, Dorn JP, et al. 1997. Exercise, occupational activity, and risk of endometrial cancer. Ann Epidemiol 7:46–53. Paganini-Hill A, Ross RK, Henderson BE. 1989. Endometrial cancer and patterns of use of oestrogen replacement therapy: a cohort study. Br J Cancer 59:445–447.

1041

Parazzini F, La Vecchia C, D’Avanzo B, Moroni S, Chatenoud L, Ricci E. 1995a. Alcohol and endometrial cancer risk: findings from an Italian casecontrol study. Nutrit Cancer 23:55–62. Parazzini F, La Vecchia C, Moroni S. 1994a. Intrauterine device use and risk of endometrial cancer. Br J Cancer 70:672–673. Parazzini F, La Vecchia C, Moroni S, Chatenoud L, Ricci E. 1994b. Family history and the risk of endometrial cancer. Int J Cancer 59:460–462. Parazzini F, La Vecchia C, Negri E, Fedele L, Balotta F. 1991. Reproductive factors and risk of endometrial cancer. Am J Obstet Gynecol 164:522–527. Parazzini F, La Vecchia C, Negri E, Moroni S, Chatenoud L. 1995b. Smoking and risk of endometrial cancer: results from an Italian case-control study. Gynecol Oncol 56:195–199. Parazzini F, La Vecchia C, Negri E, et al. 1999. Diabetes and endometrial cancer: an Italian case-control study. Int J Cancer 81:539–542. Parazzini F, Negri E, La Vecchia C, et al. 1998. Role of reproductive factors on the risk of endometrial cancer. Int J Cancer 76:784–786. Parkin DM, Muir C, Whelan S, Gao Y-T, Ferlay J, Powell J. 1992. Cancer Incidence in Five Continents (Vol. 6). IARC, Lyon, France. Parkin DM, Whelan SL, Ferlay J, Raymond L, Young J. 1997. Cancer Incidence in Five Continents Volume VII. IARC, Lyon, France. Parslov M, Lidegaard O, Klintorp S, et al. 2000. Risk factors among young women with endometrial cancer: a Danish case-control study. Am J Obstet Gynecol 182:23–29. Paterson ME, Wade-Evans T, Sturdee DW, Thom MH, Studd JW. 1980. Endometrial disease after treatment with oestrogens and progestogens in the climacteric. Br Med J 280:822–824. PDR. 2003. Physicians’ Desk Reference, 57th ed. Montvale, New Jersey. www.nolvadex.com Peller S, Halperin R, Schneider D, Kopilova Y, Rotter V. 1999. Polymorphisms of the p53 gene in women with overian or endometrial carcinoma. Oncol Rep 6:193–197. Persson I, Adami H, Bergkvist L, et al. 1989. Risk of endometrial cancer after treatment with oestrogens alone or in conjunction with progestogens: results of a prospective study. Br Med J 298:147–151. Persson I, Weiderpass E, Bergkvist L, Bergstrom R, Schairer C. 1999. Risks of breast and endometrial cancer after estrogen and estrogen-progestin replacement. Cancer Causes Control 10:253–260. Petridou E, Belechri M, Dessypris N, et al. 2002a. Leptin and body mass index in relation to endometrial cancer risk. Ann Nutr Metab 46:147–151. Petridou E, Koukoulomatis P, Dessypris N, Karalis D, Michalas S, Trichopoulos D. 2002b. Why is endometrial cancer less common in Greece than in other European Union countries? Eur J Cancer Prev 11:427–432. Pettersson B, Adami HO, Bergstrom R. 1985a. Obesity, hypertension and diabetes as risk factors for endometrial cancer. In Risk Factors for Endometrial Carcinoma. Uppsala, Sweden, Acta Universitatis Upsaliensis. Comprehensive Summaries of Upsala Dissertations from the Faculty of Medicine, Vol 6, 42 pages. Pettersson B, Adami HO, Bergstrom R, Johansson ED, 1986a. Menstruation span—a time-limited risk factor for endometrial carcinoma. Acta Obstet Gynecol Scand 65:247–255. Pettersson B, Adami HO, Lindgren a, Hesselius I. 1985b. Endometrial polyps and hyperplasia as risk factors for endometrial carcinoma. A case-control study of curettage specimens. Acta Obstet Gynecol Scand 64:653–659. Pettersson B, Adami HO, Persson I, Bergstrom R, Lindgren A, Johansson ED. 1986b. Climacteric symptoms and estrogen replacement therapy in women with endometrial carcinoma. Acta Obstet Gynecol Scand 65:81–87. Pickar JH, Yeh I-T, Wheeler JE, Cunnane MF, Speroff L. 2001. Endometrial effects of lower doses of conjugated equine estrogens and medroxyprogesterone acetate. Fertil Steril 76:25–31. Pike MC, Peters RK, Cozen W, et al. 1997. Estrogen-progestin replacement therapy and endometrial cancer. J Natl Cancer Instit 89:1110–1116. Pisani P, Parkin DM, Bray F, Ferlay J. 1999. Estimates of the worldwide mortality from 25 cancers in 1990. Int J Cancer 83:18–29. Pitot HC. 1986. Fundamentals of Oncology. New York, Marcel Dekker Inc. Potischman N, Swanson CA, Brinton La, et al. 1993. Dietary associations in a case-control study of endometrial cancer. Cancer Causes Control 4:239–250. Pukkala E, Kyyronen P, Sankila R, Holli K. 2002. Tamoxifen and toremifene treatment of breast cancer and risk of subsequent endometrial cancer: a population-based case-control study. Int J Cancer 100:337–341. Pukkala E, Poskiparta M, Apter D, Vihko V. 1993. Life-long physical activity and cancer risk among Finnish female teachers. Eur J Cancer Prev 2:369–376. Pukkala E, Tulenheimo-Silfvast A, Leminen A. 2001. Incidence of cancer among women using long versus monthly cycle hormonal replacement therapy, Finland 1994–1997. Cancer Causes Control 12:111–115.

1042

PART IV: CANCER BY TISSUE OF ORIGIN

Ragozzino M, Melton LJ, Chu CP, Palumbo PJ. 1982. Subsequent cancer risk in the incidence cohort of Rochester, Minnesota, residents with diabetes mellitus. J Chronic Dis 35:13–19. Ribeiro G, Swindell R. 1992. The Christie Hospital Adjuvant Tamoxifen Trial. J Natl Cancer Instit 11:121–125. SEER (Surveillance, Epidemiology, and End Results) Program (www.seer.cancer.gov) Public-Use Data (1973–1999), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2002, based on the November 2001 submission. Robboy SJ, Bradley R. 1979. Changing trends and prognostic features in endometrial cancer associated with exogenous estrogen therapy. Obstet Gynecol 54:269–277. Rosenblatt KA, Thomas DB. 1995. Prolonged lactation and endometrial cancer. WHO Collaborative Study of Neoplasia and Steroid Contraceptives. Int J Epidemiol 24:499–503. Rosenblatt KA, Thomas DB, The WHO Collaborative Study of Neoplasia and Steroid Contraceptives. 1991. Hormonal content of combined oral contraceptives in relation to the reduced risk of endometrial cancer. Int J Cancer 49:870–874. Rosenblatt KA, Thomas DB, WHO Collaborative Study of Neoplasia and Steroid Contraception. 1996. Intrauterine devices and endometrial cancer. Contraception 54:329–332. Ross RK, Paganini-Hill A, Roy S, Chao A, Henderson BE. 1988. Past and present preferred prescribing practices of hormone replacement therapy among Los Angeles gynecologists: possible implications for public health. Am J Public Health 78:516–519. Rubin GL, Peterson HB, Lee NC, Maes EF, Wingo PA, Becker S. 1990. Estrogen replacement therapy and the risk of endometrial cancer: remaining controversies. Am J Obstet Gynecol 162:148–154. Ruddon RW. 1995. Cancer Biology, 3rd ed. New York, Oxford University Press, pp. 240–251. Rutanen EM, Stenman S, Blum W, Karkkainen T, Lehtovirta P, Stenman UH. 1993. Relationship between carbohydrate metabolism and serum insulinlike growth factor system in postmenopausal women: comparison of endometrial cancer patients with healthy controls. J Clin Endocrinol Metab 77:199–204. Salazer-Martinez E, Lazcano-Ponce EC, Gonzalez Lira-Lira G, Escudero-De los Rios P, Salmeron-Castro J, Hernandez-Avila M. 1999. Reproductive factors of ovarian and endometrial cancer risk in a high fertility population in Mexico. Cancer Res 59:3658–3662. Salazar-Martinez E, Lazcano-Ponce EC, Lira-Lira GG, et al. 2000. Casecontrol study of diabetes, obesity, physical activity and risk of endometrial cancer among Mexican women. Cancer Causes Control 11:707–711. Salerno LJ. 1962. Feminizing mesenchymomas of the ovary—An analysis of 28 granulosa-theca cell tumors and their relationship to coexistent carcinoma. Am J Obstet Gynecol 84:731–738. Salvesen HB, Akslen LA, Albrektsen G, Iversen OE. 1998. Poorer survival of nulliparous women with endometrial carcinoma. Cancer 82:1328–1333. Sasaki M, Sakuragi N, Dahiya R. 2003. The CAG repeats in exon 1 of the androgen receptor gene are significantly longer in endometrial cancer patients. Biochem Biophysical Res Comm 35:1105–1108. Sasaki M, Tanaka Y, Kaneuchi M, Sakuragi N, Dahiya R. 2002. Polymorphisms of estrogen receptor alpha gene in endometrial cancer. Biochem Biophys Res Commun 297:558. Sasco AJ, Chaplin G, Amoros E, Saez S. 1996. Endometrial cancer following breast cancer: effect of tamoxifen and castration by radiotherapy. Epidemiol 7:9–13. Satyaswaroop PG, Zaino RJ, Mortel R. 1984. Estrogen-like effects of tamoxifen on human endometrial carcinoma transplanted into nude mice. Cancer Res 44:4006–4010. Schapira DV, Kumar RD, Lyman GH, Cavanagh D, Roberts WS, LaPolla J. 1991. Upper-Body Fat Distribution and Endometrial Cancer Risk. JAMA 266:1808–1811. Schiff I, Sela HK, Cramer D, Tulchinsky D, Ryan KJ. 1982. Endometrial hyperplasia in women on cyclic or continuous estrogen regimens. Fertil Steril 37:79–82. Shapiro JA, Weiss NS, Beresford SA, Voigt LF. 1998. Menopausal hormone use and endometrial cancer, by tumor grade and invasion. Epidemiol 9:99–101. Shapiro S, Kaufman DW, Slone D, et al. 1980. Recent and past use of conjugated estrogens in relation to adenocarcinoma of the endometrium. N Engl J Med 303:485–489. Shapiro S, Kelly JP, Rosenberg L, et al. 1985. Risk of localized and widespread endometrial cancer in relation to recent and discontinued use of conjugated estrogens. N Engl J Med 313:969–972. Sharp L, Cardy AH, Cotton SC, Little J. 2004. CYP17 gene polymorphisms: prevalence and associations with hormone levels and related factors. A Huge review. Amer J Epidemiol 160:729–740.

Shaw ST, Macaulay LK. 1979. Morphorlogic studies on IUD-induced metrorrhagia. II. Surface changes of the endometrium and microscopic localization of bleeding sites. Contraception 19:63–81. Sherman BM, Korenman SG. 1974. Measurement of serum LH, FSH, estradiol and progesterone in disorders of the human menstrual cycle: the inadequate luteal phase. J Clin Endocrinol Metab 39:145–149. Shields TS, Weiss NS, Voigt LF, Beresford SA. 1999. The additional risk of endometrial cancer associated with unopposed estrogen use in women with other risk factors. Epidemiology 10:733–738. Shoff SM, Newcomb PA. 1998. Diabetes, body size, and risk of endometrial cancer. Am J Epidemiol 148:234–240. Shu XO, Brinton LA, Zheng W, Gao YT, Fan J, Fraumeni JF, Jr. 1991. A population-based case-control study of endometrial cancer in Shanghai, China. Int J Cancer 49:38–43. Shu XO, Brinton LA, Zheng W, et al. 1992. Relation of obesity and body fat distribution to endometrial cancer in Shanghai, China. Cancer Res 52:3865–3870. Shu XO, Hatch MC, Zheng W, Gao YT, Brinton LA. 1993a. Physical activity and risk of endometrial cancer. Epidemiol 4:342–349. Shu XO, Zheng W, Potischman N, et al. 1993b. A population-based casecontrol study of dietary factors and endometrial cancer in Shanghai, People’s Republic of China. Am J Epidemiol 137:155–165. Shulman LP, Yankov V, Uhl K. 2002. Safety and efficacy of a continuous oncea-week 17B-estradiol/levonorgestrel transdermal system and its effects on vasomotor symptoms and endometrial safety in postmenopausal women: the results of two multicenter, double-blind, randomized, controlled trials. Menopause 9:195–207. Siiteri PK, MacDonald PC. 1973. The role of extra-glandular estrogen in human endocrinology. Handbook of Physiology 2:615. Silverberg SG, Makowski EL, Roche WD. 1977. Endometrial carcinoma in women under 40 years of age: comparison of cases in oral contraceptive users and non-users. Cancer 39:592–598. Smith EM, Sowers MF, Burns TL. 1984. Effects of smoking on the development of female reproductive cancers. J Natl Cancer Inst 73:371–376. Soler M, Chatenoud L, Negri E, Parazzini F, Franceschi S, La Vecchia C. 1999. Hypertension and hormone-related neoplasms in women. Hypertension 34:320–325. Spengler RF, Clarke EA, Woolever A, Newman AM, Osborn RW. 1981. Exogenous estrogens and endometrial cancer: A case-control study and assessment of potential bias. Am J Epidemiol 114:497–506. Speroff L, Rowan J, Symons J, Genant H, Wilborn W. 1996. The comparative effect on bone density, endometrium, and lipids of continuous hormones as replacement therapy (CHART Study). JAMA 276:1397–1403. Staland B. 1985. Continuous treatment with a combination of estrogen and gestagen—a way of avoiding endometrial stimulation. Clinical experiences with Kliogest. Acta Obstet Gynecol Scand Suppl 130:29– 35. Stampfer M, Colgitz G, Willet W, Rosner B, Hennekens C, Speizer F. 1986. A prospective study of exogenous hormones and risk of endometrial cancer (Abstract). Am J Epidemiol 124:520. Standeven M, Criqui MH, Klauber MR, Gabriel S, Barrett-Connor E. 1986. Correlates of change in postmenopausal estrogen use in a populationbased study. Am J Epidemiol 124:268–274. Stanford JL, Brinton LA, Berman ML, et al. 1993. Oral contraceptives and endometrial cancer: do other risk factors modify the association? Int J Cancer 54:243–248. State of California, Department of Health. 1978. Uterine Cancer Incidence. Vol. V, No. 1. Sacramento, California. Stavraky KM, Collins JA, Donner A, Wells GA. 1981. A comparison of estrogen use by women with endometrial cancer, gynecologic disorders, and other illnesses. Am J Obstet Gynecol 141:547–555. Stockwell HG, Lyman GH. 1987. Cigarette smoking and the risk of female reproductive cancer. Am J Obstet Gynecol 157:35–40. Sturdee DW, Wade-Evans T, Paterson ME, Thom M, Studd JW. 1978. Relations between bleeding pattern, endometrial histology, and oestrogen treatment in menopausal women. Br Med J 1:1575–1577. Sturgeon SR, Brinton LA, Berman ML, et al. 1993. Past and present physical activity and endometrial cancer risk. Br J Cancer 68:584–589. Sturgeon SR, Brinton LA, Berman ML, et al. 1997. Intrauterine device use and endometrial cancer risk. Int J Epidemiol 26:496–500. Sturgeon SR, Brock JW, Potischman N, et al. 1998. Serum concentrations of organochlorine compounds and endometrial cancer risk (United States). Cancer Causes Control 9:417–424. Swanson CA, Wilbanks GD, Twiggs LB, et al. 1993. Moderate alcohol consumption and the risk of endometrial cancer. Epidemiol 4:530–536. Szekely DR, Weiss NS, Schweid AI. 1978. Incidence of endometrial carcinoma in King County, Washington: a standardized histologic review. J Natl Cancer Inst 60:985–989.

Endometrial Cancer Terry P, Baron JA, Weiderpass E, Yuen J, Lichtenstein P, Nyren O. 1999. Lifestyle and endometrial cancer risk: a cohort study from the Swedish Twin Registry. Int J Cancer 82:38–42. Terry P, Vainio H, Wolk A, Weiderpass E. 2002a. Dietary factors in relation to endometrial cancer: a nationwide case-control study in Sweden. Nutr Cancer 42:25–32. Terry PD, Miller AB, Rohan TE. 2002b. A prospective cohort study of cigarette smoking and the risk of endometrial cancer. Br J Cancer 86:1430–1435. The Centers for Disease Control, Cancer and Steroid Hormone Study. 1983. Oral contraceptive use and the risk of endometrial cancer. JAMA 249:1600–1604. The Centers for Disease Control, Cancer and Steroid Hormone Study. 1987. Combination oral contraceptive use and the risk of endometrial cancer. JAMA 257:796–800. Thom MH, White PJ, Williams RM, et al. 1979. Prevention and treatment of endometrial disease in climacteric women receiving oestrogen therapy. Lancet 2:455–457. Tornberg SA, Carstensen JM. 1994. Relationship between Quetelet’s index and cancer of breast and female genital tract in 47,000 women followed for 25 years. Br J Cancer 69:358–361. Tyler CW, Webster LA, Ory HW, Rubin GL. 1985. Endometrial cancer: how does cigarette smoking influence the risk of women under age 55 years having this tumor? Am J Obstet Gynecol 151:899–905. Tzonou A, Lipworth L, Kalandidi A, et al. 1996. Dietary factors and the risk of endometrial cancer: a case—control study in Greece. Br J Cancer 73:1284–1290. U.S. Department of Health and Human Services. 1981. Hysterectomy in women aged 15–44, United States, 1970–1978. MMWR Morb Mortal Wkly Rep 30:173–176. Venn A, Watson L, Bruinsma F, Giles G, Healy D. 1999. Risk of cancer after use of fertility drugs with in-vitro fertilisation. Lancet 354:1586–1590. Venn A, Watson L, Lumley J, Giles G, King C, Healy D. 1995. Breast and ovarian cancer incidence after infertility and in vitro fertilization. Lancet 346:995–1000. Vessey MP, Painter R. 1995. Endometrial and ovarian cancer and oral contraceptives—findings in a large cohort study. Br J Cancer 71:1340–1342. Villard L, Murphy M. 1990. Endometrial cancer trends in England and Wales: a possible protective effect of oral contraceptives. Int J Epidemiol 19:255–258. Voigt LF, Deng Q, Weiss NS. 1994. Recency, duration, and progestin content of oral contraceptives in relation to the incidence of endometrial cancer (Washington, USA). Cancer, Causes & Control 5:227–233. Voigt LF, Weiss NS, Chu J, Daling JR, McKnight B, Van Belle G. 1991. Progestagen supplementation of exogenous oestrogens and risk of endometrial cancer. Lancet 338:274–277. Waterhouse J, Muri C, Correa P, Powell J. 1976. Cancer Incidence in Five Continents, Vol. 3. IARC, Lyon, France. Waterhouse J, Muri C, Shanmugaratnam K, Powell J. 1982. Cancer Incidence in Five Continents, Vol. 4. IARC, Lyon, France. Webster LA, Weiss NS, the Cancer and Steroid Hormone Study Group. 1989. Alcoholic beverage consumption and the risk of endometrial cancer. Int J Epidemiol 18:786–791. Weiderpass E, Adami HO, Baron JA, et al. 1999a. Risk of endometrial cancer following estrogen replacement with and without progestins. J Natl Cancer Inst 91:1131–1137. Weiderpass E, Adami HO, Baron JA, Magnusson C, Lindgren A, Persson I. 1999b. Use of oral contraceptives and endometrial cancer risk (Sweden). Cancer Causes Control 10:277–284. Weiderpass E, Adami HO, Baron JA, et al. 2000a. Organochlorines and endometrial cancer risk. Cancer Epidemiol Biomarkers Prev 9:487–493. Weiderpass E, Baron JA. 2001. Cigarette smoking, alcohol consumption, and endometrial cancer risk: a population-based study in Sweden. Cancer Causes Control 12:239–247. Weiderpass E, Baron JA, Adami HO, et al. 1999c. Low-potency oestrogen and risk of endometrial cancer: a case-control study. Lancet 353:1824–1828. Weiderpass E, Gridley G, Persson I, Nyren O, Ekbom A, Adami H-O. 1997. Risk of endometrial and breast cancer in patients with diabetes mellitus. Int J Cancer 71:360–363. Weiderpass E, Persson I, Adami HO, Magnusson C, Lindgren A, Baron JA. 2000b. Body size in different periods of life, diabetes mellitus, hypertension, and risk of postmenopausal endometrial cancer (Sweden). Cancer Causes Control 11:185–192.

1043

Weiderpass E, Persson I, Melhus H, Wedren S, Kindmark A, Baron JA. 2000c. Estrogen receptor alpha gene polymorphisms and endometrial cancer risk. Carcinogenesis 21:623–627. Weir HK, Sloan M, Kreiger N. 1994. The relationship between cigarette smoking and the risk of endometrial neoplasms. Int J Epidemiol 23:261–266. Weiss NS. 1978a. Assessing the risks from menopausal estrogen use: what can we learn from trends in mortality from uterine cancer. J Chron Dis 31:705–708. Weiss NS. 1978b. Noncontraceptive estrogens and abnormalities of endometrial proliferation. Ann Intern Med 88:410–412. Weiss NS, Farewall VT, Szekely DR, English DR, Kiviat N. 1980. Oestrogens and endometrial cancer: effect of other risk factors on the association. Maturitas 2:185–190. Weiss NS, Sayvetz TA. 1980. Incidence of endometrial cancer in relation to the use of oral contraceptives. N Engl J Med 302:551-554. Weiss NS, Szekely DR, Austin DF. 1976. Increasing incidence of endometrial cancer in the United States. N Engl J Med 294:1259–1262. Weiss NS, Szekely DR, English DR, Schweid AI. 1979. Endometrial cancer in relation to patterns of menopausal estrogen use. JAMA 242:261–264. WHI Writing Group. 2002. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results from the Women’s Health Initiative randomized controlled trial. JAMA 288:321–333. Whitehead MI, King RJ, McQueen J, Campbell S. 1979. Endometrial histology and biochemistry in climacteric women during oestrogen and oestrogen/progestogen therapy. J R Soc Med 72:322–327. Whitehead MI, McQueen J, Beard RJ, Minardi J, Campbell S. 1977. The effects of cyclical oestrogen therapy and sequential oestrogen/progestogen therapy on the endometrium of postmenopausal women. Acta Obstet Gynecol Scand Suppl 91–101. WHO Collaborative Study of Neoplasia and Steroid Contraceptives. 1991. Depot medroxyprogesterone acetate (DMPA) and risk of endometrial cancer. Int J Cancer 49:186–190. Wideroff L, Gridley G, Mellemkjaer L, et al. 1997. Cancer incidence in a population-based cohort of patients hospitalized with diabetes mellitus in Denmark. J Natl Cancer Inst 89:1360–1365. Willett W, Stampfer MJ, Bain C, et al. 1983. Cigarette smoking, relative weight, and menopause. Am J Epidemiol 117:651–658. Williams AR, Weiss NS, Ure CL, Ballard J, Daling JR. 1982. Effect of weight, smoking, and estrogen use on the risk of hip and forearm fractures in postmenopausal women. Obstet Gynecol 60:695–699. Woodruff JD, Pickar JH, and the Menopause Study Group. 1994. Incidence of endometrial hyperplasia in postmenopausal women taking conjugated estrogens (Premarin) with medroxyprogesterone acetate or conjugated estrogens alone. Am J Obstet Gynecol 170:1213–1223. Writing Group for the PEPI Trial. 1995. Effects of estrogen or estrogen/progestin regimens on heart disease risk factors in postmenopausal women: the Postmenopausal Estrogen/Progestin Interventions (PEPI) Trial. JAMA 273:199–208. Wynder EL, Escher GC, Mantel N. 1966. An epidemiological investigation of cancer of the endometrium. Cancer 19:489–520. Wysowski DK, Golden L, Burke L. 1995. Use of menopausal estrogens and medroxyprogesterone in the United States, 1982–1992. Obstet Gynecol 85:6–10. Yaron M, Levy T, Chetrit A, et al. 2001. The polymorphic CAG repeat in the androgen receptor gene in Jewish Israeli women with endometrial carcinoma. Cancer 92:1190–1194. Yen SS. 1980. The polycystic ovary syndrome. Clin Endocrinol 12:177– 207. Zeleniuch-Jacquotte A, Akhmedkhanov A, Kato I, et al. 2001. Postmenopausal endogenous oestrogens and risk of endometrial cancer: results of a prospective study. Br J Cancer 84:975–981. Zelmanowicz A, Hildesheim A, Sherman ME, et al. 1998. Evidence for a common etiology for endometrial carcinomas and malignant mixed mullerian tumors. Gynecol Oncol 69:253–257. Zheng W, Kushi LH, Potter JD, et al. 1995. Dietary intake of energy and animal foods and endometrial cancer incidence. The Iowa women’s health study. Am J Epidemiol 142:388–394. Zheng W, Shu XO, McLaughlin JK, Chow WH, Gao YT, Blot WJ. 1993. Occupational physical activity and the incidence of cancer of the breast, corpus uteri, and ovary in Shanghai. Cancer 71:3620–3624. Ziel HK, Finkle WD. 1975. Increased risk of endometrial carcinoma among users of conjugated estrogens. N Engl J Med 293:1167–1170.

54

Cervical Cancer MARK H. SCHIFFMAN AND ALLAN HILDESHEIM

C

ervical cancer is an important public health problem worldwide. It is the second most common cancer among women, ranking first in many developing countries (Parkin et al., 2001). In the United States and other countries with broad-coverage cervical cytologic screening (Pap test) programs, there has been a marked decline in cervical cancer incidence and mortality over recent decades (Beral et al., 1994; Gustafsson et al., 1997a; Wang et al., 2003b). The reduction has apparently been due to detection and treatment of precancerous lesions. Nonetheless, despite costly and laborious screening programs collectively costing billions of dollars annually, there were approximately 9710 cases of invasive cervical cancer in 2006 in the United States, with about 3700 deaths (Jemal et al., 2006). It is now known that virtually all cases of cervical cancer and preceding precancerous changes worldwide can be attributed to infection with one of the approximately 15 oncogenic genotypes of human papillomavirus (HPV) (Bosch et al., 1995; Bosch et al., 2003; Munoz et al., 2003). However, cervical infection with oncogenic types of HPV is extremely common compared to the relatively rare development of cervical cancer. Thus, additional etiologic factors are involved, including HPV type variants, variability in the host immunologic response, smoking, high parity, oral contraceptive use, co-infection with other viral or bacterial agents, and probably diet. Although challenges still exist in clarifying the multi-stage pathogenesis of cervical cancer, epidemiologic understanding now rivals our understanding of any other malignancy. As a result, it is possible, using a variety of molecular epidemiologic approaches, to define new prevention strategies even more effective than cervical cytologic screening alone. For example, HPV DNA testing is a useful adjunct to cytologic screening (2002; Saslow et al., 2002; Wright et al., 2003b). In the long term, the most exciting possibility is the primary prevention of cervical neoplasia via HPV immunization of the general population (Harper et al., 2004; Koutsky et al., 2002; Lowy et al., 2003; Villa et al., 2005a).

CLASSIFICATION Anatomic Distribution: The Cervical Transformation Zone The cervix is the cylindrically shaped lower third of the uterus extending into the vagina from the anterior vaginal wall (Fig. 54–1). The cervical epithelium is derived from two embryologically distinct sources. The part of the cervix that projects into the vagina, called the ectocervix or portio, is covered by non-keratinized stratified squamous epithelium similar to the neighboring lining of the vagina (Wright et al., 2002). The endocervical canal is covered by tall, mucus-secreting columnar cells of the same embryologic derivation as the uterine endometrium. At birth, the columnar epithelium extends out onto the ectocervix, but with age the position of the squamocolumnar junction recedes into the endocervical canal, as columnar epithelium is replaced by squamous epithelium in a process called “squamous metaplasia.” The metaplastic area adjacent to the receding squamocolumnar junction is called the transformation zone, and this area has a unique sensitivity for neoplastic events (Jacobson et al., 1999). Other transformation zones at the anus and oropharynx also are prone to HPV carcinogenesis. The reasons for the special susceptibility of the transformation zones are unknown. As a possibly related point, women with

1044

an increased area of squamous metaplasia, such as diethylstilbestrolexposed (DES) daughters, may have an increased incidence of cervical neoplasia (Robboy et al., 1984). The majority of cervical cancers are diagnosed in the anterior and posterior parts of the transformation zone in the so-called 12 o’clock and 6 o’clock positions (Guido, 2005). These anatomic positions might be most exposed to sexual trauma and transmission, more susceptible due to their peculiarly slow completion of squamous transformation, or might be immunologically compromised because blood flows into the cervix laterally along the uterine broad ligaments and terminates anteriorly and posteriorly. Alternatively, the differences by position might reflect artifacts in appearance or ease of biopsy (Guido, 2005). When precancerous lesions are found, regardless of their exact position, destruction of the entire transformation zone epithelium by the loop electrical excision procedure (LEEP) or, now less frequently, cryotherapy or laser is the mainstay of clinical intervention (Wright et al., 1992b).

Histopathology In the United States, approximately 80% of invasive cervical cancers are squamous cell carcinomas of epithelial origin (Wang et al., 2003b). The remaining 20% of tumors consist largely of adenocarcinomas (15%), with mixed adenosquamous cell carcinomas and other rare histological types accounting for the residual 5% of tumors (Wang et al., 2003b).

PRECURSOR LESIONS Thanks to rapid progress in understanding the natural history of cervical HPV infection and resultant neoplasia, molecular epidemiologists can now study the epidemiology of cervical carcinogenesis as a multistep process. It is possible to examine the risk factors for successive steps from sexual transmission of oncogenic HPV infection, to cervical precancer, to cancer itself. Because the cervix can be seen and sampled in the context of screening programs, cervical carcinogenesis can be measured macroscopically, microscopically, and molecularly. Epidemiologists study cervical cancer precursors and pathogenesis together, but this chapter will present the two aspects separately. The present section will describe the two major precursor stages: HPV infection and precancer. In contrast, the section on Pathogenesis will outline what is understood about the mechanisms of progression between the stages of cervical carcinogenesis shown in Figure 54–2.

HPV Infection Introduction Infection with one of the oncogenic types of HPV (including but possibly not limited to types 16, 18, 26, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, 68, 73, 82) is the central cause of cervical cancer (Bosch et al., 2003; Munoz et al., 2003). Therefore, detection of oncogenic types of HPV DNA in cervical tissue can be viewed as the earliest measure of tissue-specific exposure moving toward cancer. The epidemiology of oncogenic types of HPV as infectious agents is detailed more fully as part of Chapter 26. In addition to the oncogenic types of HPV,

Cervical Cancer

Figure 54–1. Aspects of cervical anatomy that are essential to understanding cervical carcinogenesis.

there are probably more than 100 other types, which cause common warty lesions or microscopically and macroscopically inapparent infections of the skin and mucosa, but they are outside of the scope of this discussion (Beutner, 2000; Broker et al., 2001). The reference standard of sensitive detection of current HPV infection is HPV type-specific DNA detection (Iftner et al., 2003). HPV serology using virus-like particles (VLP) as antigens is a useful adjunctive measurement that is specific but not sensitive in detecting exposure to individual types of HPV, including past exposure (Castle et al., 2002b; Castle et al., 2002c; Dillner, 1999; Wang et al., 2003a). Microscopic diagnoses (cytology from scraped cells and histology from biopsies) and macroscopic diagnoses (colposcopy) of the tissue effects of early HPV infection are critical to cervical cancer screening and prevention efforts (Sherman, 2003; Wright, 2003). However, early HPV effects on the cervix are unpredictable, pleiomorphic, and notoriously difficult to interpret reproducibly. There is also a plethora of vocabulary that the epidemiologist must know to read the literature. This chapter can only serve as an introduction.

1045

Microscopic Diagnosis of Early HPV Infection. The full success of cervical cancer screening was made possible by what can now be recognized as the microscopic diagnosis of HPV infection, cervical neoplasia, and their intimate inter-relationship (Koss, 1989). Originally, Georges Papanicolaou proposed a cytology classification meant to predict the risk of concurrent invasive cancer from examination of exfoliated cervical-vaginal cells. However, in addition to cancerous and precancerous changes, cytopathologists noted more subtle changes that they later came to interpret as even earlier cancer precursors. One of the early major advances in understanding the cytopathology of early HPV infection occurred in the mid-1950s, when Koss and Durfee described a morphologic abnormality of squamous cells, which they termed “koilocytotic atypia” (Koss, 1989). “Koilos” is a Greek word meaning “hollow.” Synonyms for koilocytotic atypia include “condylomatous atypia” and “warty atypia.” Koilocytes have hyperchromatic, enlarged, wrinkled, and sometimes multiple nuclei surrounded by perinuclear clear zones (halos). Twenty years later, Meisels and Fortin, and Purola and Savia, were the first groups to propose formally that flat cervical lesions demonstrating koilocytotic atypia were cervical equivalents of exophytic condyloma acuminatum (genital warts), already known by then to be caused by HPV. Electronic microscopic studies showed that koilocytotic cells contained crystalline arrays of assembled HPV virions in their abnormal nuclei. Of note, even in flat lesions, there is often widening of the middle, “spiny layer” of the epithelium, called acanthosis. Excessive keratin production (hyperkeratosis) or abnormal keratin deposition (parakeratosis) are also microscopic hallmarks of HPV infection. Independent of HPV, Richart and colleagues (1969) noted the continuity of cervical cancer precursor lesions and introduced the concept of “Cervical Intraepithelial Neoplasia (CIN).” CIN was defined as a continuum (CIN 1, CIN 2, and CIN 3) of progressively more severe pre-invasive changes. In normal epithelium, proliferation is restricted to the basal layer, the so-called germinal epithelium. The epithelial cells differentiate on a fixed and orderly schedule that leads to their programmed death and sloughing. In CIN 1, squamous differentiation in the more superficial layers of the epithelium becomes abnormal, but the cells continue to differentiate, undergo programmed cell death and slough, such that the proliferative (non-differentiated or “immortalized”) compartment remains less than one-third of the full thickness of the epithelium. Over the 1970s and 1980s, pathologists learned that the cytopathologic effects of HPV classified as koilocytotic atypia were impossible to distinguish from “pre-cancerous” changes called CIN 1. In fact, CIN 1 is now considered a histologic sign of HPV infection, linked more closely to viral exposure than cancer risk. This recognition led to the combination of HPV effects and CIN 1 for the purposes of cancer screening in the Bethesda System of cervical cytology (Kurman et al.,

Persistence (>1-2 Years) Infection HPV Progression Pre- Invasion Norm infected Cancer Cervix Clearance Cervix Regression cancer

Mild Cytologic Abnormalities and/or Seroconversion

Figure 54–2. An epidemiologic model of multi-stage cervical carcinogenesis.

1046

PART IV: CANCER BY TISSUE OF ORIGIN

1991). In the Bethesda System, which has been updated twice (Solomon et al., 2002), the category “low-grade squamous intraepithelial lesion (LSIL)” subsumes older cytologic terms such as CIN 1, mild dysplasia, mild dyskaryosis (British), koilocytotic atypia, condylomatous atypia, warty atypia, and HPV effect. The Bethesda System is used in the United States and several other countries for cytology screening, The CIN scale remains more accepted worldwide when discussing histopathology (i.e., biopsies). The cytologic interpretation is a reasonably specific and reproducible sign of HPV infections (Stoler et al., 2001), typically observed in approximately 2% of Pap tests in the U.S. screening population (Davey et al., 2000). In fact, if all HPV types are assayed and cytology is carefully reviewed, HPV-negative LSIL is rare (Zuna et al., 2005). However, equivocal interpretations outnumber classic presentations (up to 5% of Pap test interpretations). In other words, millions of women are told annually that they have equivocal cytologic results that may result from HPV infection. The most common of these equivocal interpretations is called “atypical squamous cells of undetermined significance (ASC-US).” Approximately one-third to onehalf of such cases represent subtle signs of HPV infection, primarily with oncogenic types of HPV (Solomon et al., 2001; Solomon, 2003). It is important to realize that the natural history of HPV that presents as ASC-US is not markedly different than infections that present as classical LSIL (Cox et al., 2003). The common presentation of HPV infection as ASC-US demonstrates the difficulty of using cytology to define HPV infection. In fact, only the minority of HPV infections exhibits a concurrent, routinely perceivable cytologic abnormality, as discussed under the section on Pathogenesis. Currently, HPV infections and resultant early lesions are not treated when first diagnosed (Guido et al., 2003; Wright et al., 2002a; Wright et al., 2003a), given their tendency to resolve without treatment (see Pathogenesis).

Macroscopic Diagnosis of Early HPV Infection. It is often possible for the examining clinician to recognize cervical HPV infections macroscopically (Wright, 2003). Although characteristically flat and nearly invisible without staining, early lesions due to oncogenic HPV infections may often be diagnosed by application of 5% acetic acid (vinegar) or Lugol’s iodine solution (Ferris, 1994). Lugol’s solution stains glycogen, which is often lacking in HPVassociated lesions. HPV-induced lesions are usually “acetowhite,” meaning they temporarily turn white after vinegar is applied. The mechanism of acetowhitening is not known for sure, and several theories exist. Acetowhitening is frequently used by clinicians examining the cervix, vagina, and vulva under magnification (colposcopy) following an abnormal cytologic result, in order to identify and biopsy lesions. Low-cost proxy versions of colposcopy also depend on acetowhitening or staining with Lugol’s solution to diagnose HPVinduced cervical lesions (Blumenthal et al., 2001; Denny et al., 2000; Sankaranarayanan et al., 2003). Novel, computer-assisted scanning technologies are attempting to exploit the same and additional tissue characteristics to permit diagnosis, perhaps without an expert clinician (Wright, 2003). Of note, no visual appearance, including acetowhitening, is specific for HPV infection. Many other types of epithelial conditions, particularly immature squamous metaplasia, can also be acetowhite, creating diagnostic uncertainty for colposcopy and related simpler alternatives or novel computer assisted techniques. Historically, colposcopically directed biopsies have been the clinical reference standard for defining the severity of anogenital HPVrelated disease, specifically, to distinguish between early infection and precancer requiring treatment. However, the choice of biopsy site and the histopathologic diagnosis of resultant biopsies tend to be variable and subjective (Stoler et al., 2001). Thus, the error inherent in colposcopy affects the more general process of diagnosing the state of HPV natural history and the severity of cervical cancer precursors.

genic HPV infection, with or without macroscopic/microscopic signs, is extremely common and usually benign. Cervical precancer is relatively rare by comparison, and represents a truly pre-malignant lesion in the most severe cases (carcinoma in situ). Although epidemiologists study precancer as a useful surrogate endpoint for cervical cancer risk, it remains a non-trivial task to define “precancer,” because there is still heterogeneity in the microscopic diagnoses. Some certainly represent acute HPV infections of particularly bad microscopic appearance that are destined to regress. Others are incipient precancer destined to persist with high risk of invasion. Non-oncogenic HPV infections are capable of producing lesions diagnosed as precancer, showing that this level of abnormality is not a perfect surrogate for cancer risk (Clifford et al., 2003b). Following the United States emphasis on safety and concern over loss to follow-up, treating precancer is a valid clinical strategy to provide a margin of safety, given that it is not yet possible to know which lesions pose a threat. Better accuracy based on molecular profiling is the goal.

Microscopic Diagnosis of Precancer. A morphologic continuum of changes exists at the microscopic level leading from HPV infection to precancer, without a current clear cut-point. The gradient from infection to precancer is characterized by increasing nuclear atypia and failure of cellular differentiation in progressively more superficial levels of epithelium. When non-differentiation extends beyond the basal third of the epithelium to the middle third of the epithelium, the diagnosis changes from CIN 1 to CIN 2 (“moderate dysplasia”). In turn, non-differentiation reaching the upper third is called CIN 3, which includes carcinoma in situ representing full thickness replacement with undifferentiated, immortalized, atypical cells. CIN 3 is the best standard of precancer for epidemiologic studies (ALTS Group, 2003). CIN 2, in contrast, encompasses much of the heterogeneity in the precancer stage, combining the worst appearing HPV infections still destined to regress with incipient precancer in a mixture that is not yet separable. Nonetheless, as a clinical endpoint of importance, United States (but not all European) clinicians treat CIN 2 lesions, with rare exceptions in young women among whom fertility is a major concern (Nobbenhuis et al., 1999; Wright et al., 2003a). In screening programs, less than 1% of screened women in the United States are diagnosed annually with precancer (Davey et al., 2000). The low prevalence of precancer, an order of magnitude less common than signs of HPV infection, is not as pronounced in regions with deficient cervical cancer screening and treatment where precancers can develop and accumulate. Macroscopic Diagnosis of Precancer. When HPV infections progress to precancer, the clonal expansion usually occurs at the proximal (endocervical) edge of the acetowhite HPV lesion. Precancer has a thicker, grayer appearance than early HPV infection (Ferris et al., 1994). As a function of neoplastic angiogenesis, abnormally large and unusual blood vessels supplying the precancer rise to the surface to create coarse mosaic patterns or even evident dead-end vessels. These changes, though highly predictive when they all occur, are variably present. Incipient precancerous lesions are usually very small and can easily be missed by the colposcopist on examination, especially during biopsy placement. Thus, the colposcopically guided biopsy tends to under-diagnosis of precancer (Ferris et al., in press). MOLECULAR CHARACTERISTICS OF CERVICAL CANCER The molecular biology of HPV in relation to cervical cancer as a model DNA tumor virus is under intensive study. The reader is referred to elsewhere since few details can be outlined here (Munger, 2002; zur Hausen, 2000).

Cervical Precancer Introduction

Oncogenic HPV DNA

The distinction between early HPV infection and precancer is quite important for epidemiologists studying cervical carcinogenesis. Onco-

The sine qua non of cervical cancer and precancer is the presence of transcriptionally active oncogenic HPV DNA, often integrated into the

Cervical Cancer

1047

Figure 54–3. The percentage of cervical cancers attributable to the major carcinogenic types in a large international case series (Source: Bosch et al., 2003).

host genome. Even in long-maintained cervical cancer cell culture lines, such as HeLa (which contains HPV 18) and CaSki (HPV 16), suppression of HPV RNA by antisense DNA hybridization techniques leads to apoptosis. The occurrence of HPV-negative cervical cancer is so rare that epidemiologists may not be able to accrue meaningful series. The main types of HPV found alone as etiologic agents of cervical cancer include 16, 26, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, 68, 73, and 82 (Munoz N et al., 2003; Clifford et al., 2005; Cogliano et al., 2005). However, the relative etiologic fraction of these types ranges greatly, from the great importance of HPV 16 which is a common and powerful carcinogen to the equivocal role of less frequent and often benign types that might not deserve to be designated as carcinogens. The contribution of the types is indicated in Figure 54–3. It seems that cancers might very rarely be caused by typically non-oncogenic HPV types, for yet unknown reasons related to either viral or host factors. Precancers contain virtually the same distribution of HPV types as cancers, with a slightly broader spectrum reflecting that CIN 3 is not a perfect surrogate for cancer risk (Clifford et al., 2003a; 2003b). By extension, CIN 2 tends to be associated in a substantial minority of cases with non-oncogenic types of HPV (Clifford et al., 2003b).

HPV DNA Integration In early infection and precancer, HPV is typically found in an episomal form as circular double-stranded DNA. However, invasive cancers often (but not always) contain linear HPV DNA that has integrated into the host genome (Wang et al., 2003a). Although it is unclear whether HPV integration is random relative to the host genome, it typically involves a characteristic break in the viral genome that disrupts regulatory regions while retaining viral oncogenes.

DEMOGRAPHIC PATTERNS Incidence and Mortality in the United States Incidence in the United States Although the main cause of cervical cancer is known to be HPV infection, which is common throughout the United States, epidemiologists still examine national demographic patterns in cervical cancer for prevention research. Despite the effectiveness of prevention programs incorporating repeated screening and treatment of precursor lesions, in 2006, an estimated 9710 cases of invasive cervical cancer were

diagnosed in the United States (Jemal et al., 2006). The lifetime cumulative probability of cervical cancer based on 1998–2000 data was 0.8% (http://seer.cancer.gov/csr/1975_2000). Previously striking regional differences in incidence, with excesses particularly in Appalachia, are now less visible although high-incidence areas such as the Mexican-United States border still exist. In the United States and many other developed nations, incidence rates of squamous cell carcinomas have declined steadily since the introduction of Pap smear screening, while those for adenocarcinomas have not (Alfsen et al., 2000; Beral et al., 1994; Herbert et al., 2001; Peters et al., 1986a; Sasieni et al., 2001; Schwarz et al., 1986; Smith et al., 2000; Vizcaino et al., 1998; Vizcaino et al., 2000). In fact, evidence suggests that rates of cervical adenocarcinomas have risen in the past two to three decades in various countries, both relative to rates of squamous cell carcinoma and in absolute numbers (Fig. 54–4). Various explanations have been proposed for the decline in rates of squamous cell tumors coupled with increases in rates of adenocarcinomas of the cervix. While squamous cell carcinomas are believed to arise from HPV infections that occur in the transformation zone of the cervix, adenocarcinoma typically arises in the endocervical canal, a region of the cervix that is less amenable to adequate Pap smear sampling. Second, while precursor states are known and well understood for squamous cell carcinoma, other than the rarely detected adenocarcinomas in situ, precursor states are less understood for adenocarcinomas, making early detection and treatment more difficult for this condition. Third, increasing recognition of the histopathological features of invasive adenocarcinomas in recent years has resulted in a decrease in the proportion of tumors classified as being of unknown histology and a concomitant increase in the proportion of tumors classified as cervical adenocarcinomas (Smith et al., 2000; Vizcaino et al., 1998). Possibly related, infection with an oncogenic HPV is a necessary cause of both squamous cell carcinomas and adenocarcinomas, but the distribution of oncogenic HPV types and variants detected in these two tumor types vary (Altekruse et al., 2003; Bosch et al., 2003; Burk et al., 2003; Castellsague et al., 2006; Pirog et al., 2000). HPV 18, which is curiously uncommon in squamous precancers found by screening, causes nearly as many cases of adenocarcinoma as HPV 16 does, suggesting that HPV 18 might be causing the poorly-detectable precursor lesions in the endocervical epithelium. Finally, increases over time in exposure to HPV co-factors like oral contraceptives, linked more strongly with the development of cervical adenocarcinomas, might partly explain why rates have risen in absolute terms (Altekruse et al., 2003; Castellsague et al., 2006; Kjaer et al., 1993; Lacey et al., 2001; Madeleine et al., 2001; Ursin et al., 1994; Ursin et al., 1996).

1048

PART IV: CANCER BY TISSUE OF ORIGIN 1000

1000

Rate per 100,000 person-years

(e) Malignant SCC: White

1000

1000 (g) Malignant adenocarcinoma: White

(f) Malignant SCC: Black

(h) Malignant adenocarcinoma: Black

100

100

100

100

10

10

10

10

1

1

1

1

0.1

0.1

0.1

0.1

0.01

0.01 1978 1988 1998 1976–80, 1981–85, 1986–90, 1991–95, 1996–2000

0.01

0.01 1978 1988 1998 1976–80, 1981–85, 1986–90, 1991–95, 1996–2000

1978 1988 1998 1976–80, 1981–85, 1986–90, 1991–95, 1996–2000

1978 1988 1998 1976–80, 1981–85, 1986–90, 1991–95, 1996–2000

Year of diagnosis 15-34 years 35-54 years 55-74 years 75+ years

Figure 54–4. Rates of squamous cell carcinoma (SCC) and adenocarcinoma in the United States for Whites and Blacks, by calendar time (1976–80, 1981–85, 1986–90, 1991–95, 1996–2000) and age group.

Black and white Hispanic women are at high risk of cervical cancer compared with the comparably aged White population (Table 54–1 from SEER). To a lesser degree, Asian/Pacific Islander women are also at increased risk compared with White non-Hispanic women.

Table 54–1. Age-adjusted SEER incidence and U.S. Death Rates per 100,000 Women, by Race/Ethnicity, 1992–1999 Race/Ethnicity All Races White White Hispanic White Non-Hispanic Black Asian/Pacific Islander American Indian/Alaskan Native Hispanics (not exclusive from other categories)

Incidence

Mortality

10.2 9.6 18.5 8.0 13.3 11.7 7.7 17.5

3.2 2.8 4.1 2.7 6.7 3.1 3.3 3.8

There are sufficient data to examine secular trends in Black and white incidence rates of precancer (CIN 3, particularly carcinoma in situ) and invasive cervical cancer, stratified by age and by histology (i.e., squamous cancer versus adenocarcinoma) (Wang et al., 2003b). As shown in Figure 54–4, the overall incidence of invasive squamous cell cancer among White women and Black women has declined substantially over the past 25 years. The pace of decline of incidence has been more pronounced in Black than White women. There was a sharp increase in the early 1990s in the incidence of squamous precancer, possibly due to improved cell sampling tools and changes in diagnostic criteria in the SEER registry. This increase in early detection would be consistent with decreased incidence of invasive squamous cancer, especially at later stages. In White women, adenocarcinoma in situ (AIS) incidence has also increased over time, particularly among young women, but without accompanying declines in adenocarcinoma. In Black women, although few AIS and invasive adenocarcinomas are detected, invasive adenocarcinoma rises linearly with age. In sum, the rise in adenocarcinoma and its precursors represents the most poorly understood and concerning trend in U.S. incidence rates for cervical cancer.

1049

Cervical Cancer Of note, women in their teens and early twenties very rarely develop cervical cancer (Hildesheim et al., 1999; Saslow et al., 2002), which is logical given the need for sexual HPV infection and subsequent neoplastic progression. However, the rare cases among young women are very important in their influence on screening policies such as choice of age at which population-wide cervical screening is initiated. Rates of precancer are no longer recorded in the United States, because the definition of precancer is not sufficiently uniform (e.g., carcinoma in situ versus CIN 3 versus CIN 2 or CIN 3), Earlier, it was mentioned that the incidence of precancer is <10% that of HPV infection. In turn, the incidence of precancer as estimated from large international screening series is approximately 10 times greater than the rate of cervical cancer if CIN 3 is included as precancer (CIN 2 and CIN 3 are equally prevalent in several large epidemiologic series) Bosch et al., 2003). This high ratio supports the belief that many cases of precancer, particularly CIN 2, would not invade but rather would regress if followed for many years. Time trend data regarding cervical HPV infection as measured by DNA in exfoliated cell specimens are not available because standardized specimen collection and HPV testing protocols have only recently been developed. The few studies that have attempted to assess cervical HPV DNA prevalence in archival pathology specimens have found a comparable proportion of HPV positivity in all time periods, suggesting that the proportion of cervical neoplasia associated with HPV infection has not changed (Thompson et al., 1992).

United States Mortality There were 3700 deaths from cervical cancer estimated for 2006 in the United States (Jemal et al., 2006). Because cervical cancer occurs earlier than most other cancers in women, it is one of the five leading causes of cancer death among women prior to 40 years of age. The projected cumulative lifetime risk of death for cervical cancer, for females born in 1998–2000, was estimated to be 0.3% (http://seer.cancer.gov/csr/1975_2000). Age-adjusted mortality by ethnicity for 1996–2000 is shown in Table 54–1. Mortality rates among Black women were more than twice the mortality rates among White non-Hispanic women. Hispanic women had intermediate elevations in mortality, with slighter elevations among other Asian/Pacific and American Indian/Alaskan Native groups. Mortality rates declined between 1992 and 2000 for all ethic/racial groups, with the steepest declines for American Indian/Alaskan Native and Black women. Recent advances in combined radiation and chemotherapy treatments for advanced cervical cancer will hopefully soon translate into improved survival (Waggoner, 2003).

International Patterns of Incidence and Mortality Cervical cancer accounted for approximately 233,000 deaths worldwide in the year 2000, or about one-tenth of the total number of female cancer deaths (Parkin et al., 2001). Cases are often detected at late stages due to non-existent or inadequate screening, and the standard treatment options requiring gynecologic oncologists and radiation oncologists are often absent or unaffordable to affected women. As to incidence, cervical cancer is the second most common cancer of women worldwide, with 471,000 incident cases estimated in 2000 (Parkin et al., 2001), and a 5-year prevalence of more than 1.4 million cases. The cancer burden (incidence and mortality) is disproportionately high in the developing world. In several developing countries, cervical cancer is the most prevalent and important female cancer, whose importance is accentuated even further by the relatively young average age at death. The incidence rate per 100,000 women-years for invasive cervical cancer in various geographic areas is shown in Table 54–3 (Bosch and De Sanjosé, 2003). The highest age-standardized rates, more than five times the rates in the United States and Canada, were reported from East Africa, Central America, and the Pacific Islands. The geographic distribution of HPV infection has been studied mainly in correlation with cervical cancer incidence rates, to determine whether variation in prevalences of HPV measured by DNA would be reflected in cancer rates. Ecologic studies are complicated by geographical variation in cervical cancer prevention due to screening and treatment of precursor lesions. However, in comparison of regions without extensive effective screening, geographic studies using sensitive DNA testing methods to detect the oncogenic HPV

Stage at Diagnosis and Survival The mortality rate from cervical cancer is highly dependent on stage at diagnosis. Younger women and White women are more likely than older women and Black women, respectively, to be diagnosed with localized cancer that carries a good prognosis (Table 54–2). Earlier diagnosis does not appear to account for the improved survival of White compared to Black women under 50 (http://seer.cancer.gov/csr/1975_2000) if one assumes that staging is equally accurate by race.

Impact of Socioeconomic Factors At least some of the racial/ethnic differences in demographic patterns can be explained by the strong inverse associations observed between socioeconomic indicators and the risk of invasive cervical cancer. Internationally, women defined as low social class were found in a recent meta-analysis to have twice the risk of cervical cancer compared to women defined as belonging to a high social class (Parikh et al., 2003). In the United States, the inverse relationships of risk with income and education prevail among both Whites and Blacks. In one analysis, when adjustment was made for socioeconomic variables, the relative risk of cervical cancer among Blacks compared to Whites was substantially reduced from greater than 1.7 to less than 1.3 (Devesa et al., 1980). Moreover, in a recent study among HPV-infected women, socioeconomic status but not race remained a risk factor for cervical precancer (Khan et al., 2005b).

Table 54–2. Survival Rates, by Race, Age, and Stage (United States 1992–1998), SEER All Races All

<50

White Females 50+

Black Females

All

<50

50+

All

<50

50+

5,663 56 30 8 6

3,367 67 23 5 5

2,296 40 41 11 8

1,097 46 35 9 10

606 54 31 6 9

491 36 41 11 12

82 96 56 25 71

57 85 47 12 36

60 87 41 8 50

63 86 38 13 56

55 89 44 3 40

stage distribution at diagnosis (%) No. Cases Localized Regional Distant Unstaged

7,594 54 32 8 7

5-year relative All 71 Localized 92 Regional 51 Distant 15 Unstaged 52

4,376 66 24 5 5

3,218 39 42 11 9

survival by stage (%) 80 94 54 22 67

58 87 48 11 36

72 93 51 16 53

1050

PART IV: CANCER BY TISSUE OF ORIGIN

Table 54–3. Numbers and Rates of Invasive Cervical Cancer, by Region Region

No. of Cases in 2000

ASRW*

470 606 91 451 379 153 67 078 30 206 6 947 10 479 5 541 13 903 92 136 6 670 21 596 49 025 14 845 64 928 35 482 6 049 10 116 13 282 245 670 51 266 33 373 11 681 6 212 39 648 151 297 3 458 2 156 1 078 1 077

16.1 11.3 18.7 27.3 44.3 25.1 16.8 30.3 20.3 21.0 35.8 40.3 30.9 7.9 13.0 16.8 9.8 10.2 10.4 14.9 6.4 5.2 11.1 15.3 18.3 26.5 4.8 12.6 40.3 7.7

World More developed Less developed Africa Eastern Middle Northern Southern Western America Caribbean Central South U.S. and Canada Europe Eastern Northern Southern Western Asia Eastern China Japan Other Southeastern Southcentral Western Oceania Melanesia, Micronesia, and Polynesia Australia/N. Zealand Source: Adapted from Bosch and De Sanjosé, 2003. *ASRW = age-standarized incidence rates per 100,000.

types have observed HPV prevalences to correlate with the population risks of cervical cancer. Notably low rates of cervical cancer have been reported in China, possibly due to the effects on HPV transmission of the strict sexual mores adopted after the establishment of the People’s Republic of China (Li et al., 2000). As another striking example, HPV DNA prevalence and precancers are also quite uncommon in North Vietnam compared with South Vietnam (Pham et al., 2003). A comparison of age-specific cervical cancer incidence between countries prior to introduction of screening has demonstrated that rates begin to increase around age 25, with an unusually early plateau or peak starting at age 40–50 (Gustafsson et al., 1997b). It is unusual for cancer rates to plateau or fall with increasing age, and this age structure reflects that cervical cancers originate from HPV infections transmitted mainly in late adolescence and early adulthood. Currently, the effective screening programs produce age-restricted declines in cervical cancer incidence confined primarily to women aged 30–70 years old, leaving rates among younger and older women relatively unchanged (Gustafsson et al., 1997a). Migrant studies of cervical cancer are now important mainly to epidemiologists interested in the historical spread of sexually transmitted agents or in molecular evolution (Ho et al., 1993; Ong et al., 1993; Van Ranst et al., 1992; Yamada et al., 1997). As mentioned in Chapter 26, the slowly-evolving variants are being studied to understand the sexual spread of oncogenic HPVs, which are all increasingly pandemic.

ENVIRONMENTAL RISK FACTORS FOR CERVICAL CANCER HPV Infection as the Major Cause of Cervical Cancer The epidemiologic association between HPV infection and cervical cancer fulfills all of the established epidemiologic criteria for causality (Bosch et al., 2002). The causal criteria include strength and con-

sistency of the epidemiologic association, time sequence, specificity of the association, and coherence with existing biologic and epidemiologic evidence. Major reviews have recently discussed the voluminous support for causality (Bosch et al., 2002), which will only be summarized here. The cross-sectional association between oncogenic HPV DNA detection and cervical cancer is remarkably strong and consistent, with virtually no negative studies and with odds ratios of the best population-based case-control studies in the range of 50 to 500 (Herrero et al., 2005; Munoz et al., 2003). Similar risks have been observed in studies of precancer which included expert microscopic review to minimize misclassification of cases. Thus, the great majority of women with cervical cancer and/or precancer have detectable oncogenic HPV DNA, compared to a consistently lower percentage of control women (Clifford et al., 2003a). In fact, in huge case series worldwide, virtually all squamous cervical cancers and adenocarcinomas have been found to contain HPV of the same types. The International Agency for Research against Cancer (IARC) has coordinated the most definitive studies of invasive cervical cancer from over 20 countries. Over 90% of cervical cancers from each country contained HPV DNA, with the inclusion of “possible” infections raising the proportions to virtually 100% (Bosch et al., 1995; Munoz et al., 2003; Walboomers et al., 1999). Accordingly, the cancer-associated group of genital HPV types is defined as those found alone with appreciable prevalence in invasive cervical cancers. Based on this definition, the current list of cancerassociated HPV types includes at least types 16, 18, 26, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, 68, 73 and 82 (Munoz et al., 2003). HPV 53 is occasionally found in cancer, but is such a common nononcogenic type that misclassification might account for this anomaly. HPV 16 is by far the most important cancer-associated type, accounting for half of cervical cancer cases worldwide. With regard to a logical time sequence, HPV infection (again optimally measured by DNA) precedes and predicts incident cervical precancer and cancer. Results from large prospective studies of cytologically normal women, either infected or not with oncogenic types of HPV, show substantially elevated relative (>10) and absolute (approximately 10%) risks of precancer and cancer within 5–10 years of viral DNA detection (Koutsky et al., 1992; Sherman et al., 2003a). HPV 16 and HPV 18 pose especially high absolute risks for precancer that approach 20% within 10 years after infection (Khan et al., 2005). Longer-term, prospective data from follow-up of cohorts established to study HPV and cervical cancer are not yet available but convincing prospective relative risks have already been provided by archival studies based on DNA testing of stored cytologic slides or serologic testing of archived blood (Dillner et al., 1997; Wallin et al., 1999). As to the causal criterion of specificity, HPV infection causes a specific set of carcinomas of muco-cutaneous epithelia (Daling et al., 1992; Schiffman and Kjaer, 2003), particularly anogenital tumors (cervical carcinoma, and some types of vulvar, vaginal, penile, and anal carcinomas). To a lesser extent, some subsets of head and neck carcinomas are associated with HPV (Herrero, 2003). Numerous anecdotal reports of associations with a wide variety of tumor types have not been confirmed, and serosurveys of patients with multiple cancers have supported the specificity of the causal associations (Gillison et al., 2003). The animal data and experimental evidence for HPV carcinogenicity are strong, satisfying the causal criterion of “coherence.” In fact, the potential for malignant transformation of papillomavirus-induced lesions has long been recognized. Cotton-tailed rabbit papillomavirus (CRPV) causes skin cancers in conjunction with exposure to coal tar, and bovine papillomavirus (BPV) causes alimentary tract cancers in cows ingesting the co-carcinogen bracken fern. Cellular and molecular biological evidence for the oncogenic potential of human papillomaviruses is especially compelling (zur Hausen, 1989). The oncogenic types of HPV that have been studied have been shown to transform human cell lines in in vitro systems. Protein products of HPV early genes (E6, E7) have been identified that interact with growth-regulatory proteins of the human cell (p53, pRb), providing part of the mechanism for the HPV oncogenic effect. In addi-

Cervical Cancer tion, the cancer-associated types are observed to be genetically related when “phylogenetic trees” are constructed that categorize HPV types by DNA sequence homology (de Villiers et al., 2004; Schiffman et al., 2005a). Finally, as shown throughout this chapter, HPV infection explains much of the established epidemiology of cervical cancer, meeting the criterion of “coherence with existing epidemiologic knowledge”. As a result, HPV is now accepted to be the central, necessary causal factor for virtually all cases of cervical cancer in the world.

Other Environmental and Behavioral Risk Factors for Cervical Cancer The recognition of the key etiologic role of HPV infection has profoundly altered the epidemiologic study of cervical cancer. It is increasingly clear which previously “established” epidemiologic risk factors for cervical cancer are correlates of HPV infection, which lead to infection, and which are HPV co-factors operating only in the presence of infection. Because HPV infection is a necessary cause of cervical cancer worldwide, no other risk factors are important in the absence of HPV, a somewhat startling conclusion that greatly affects usual epidemiologic approaches to effect modification and confounding. The several interesting methodologic issues of how to study HPV-related cervical carcinogenesis, including control selection in retrospective studies, misclassification of a powerful confounding variable, multi-stage carcinogenesis, and surrogate endpoints, are worth pursuing in other publications (Franco, 2000; Schiffman et al., 1994; Wacholder, 2003).

Sociodemographic Factors Cervical cancer predominantly affects women in lower social classes, as defined by levels of income and education (Brinton et al., 1986a; Parikh et al., 2003). The quality and population coverage of screening and treatment of precancer has certainly been a partial explanation for this association, but is probably not the only reason. Cervical HPV infections appear to have historically been more prevalent in women of lower educational and income levels (Hildesheim et al., 1993; Stone et al., 2002). Even controlling for exposure to HPV infection does not account for the SES association entirely (Khan et a., 2005b). The correlates of low socioeconomic level that are HPV co-factors for cervical cancer are not understood.

Religion Low cervical cancer risks were recorded historically among Catholic nuns, the Amish, Mormons, and Jews. It is probable that a reduced number of sexual partners and subsequently lowered risk of HPV infection among these groups accounts for their historically lower cancer risk (Brinton et al., 1986a). The low rates in Jewish women could also relate to circumcision, which a recent large study has firmly associated with lower prevalence of penile HPV infection (Castellsague et al., 2002a). However, no definitive study of religion and cervical cancer incorporating accurate HPV testing has been reported.

Sexual Factors HPV infection is sexually transmitted. Epidemiologic studies pointed the way to the identification of HPV as the sexually transmitted cause of cervical cancer, starting with the original observations by Rignoni Stern in 1842 (Brinton et al., 1986a). Virgins were observed by epidemiologists to be at virtually no risk of cervical cancer. Women having sexual relationships at early ages were at higher risk than women whose sexual experiences began later in life. In fact, the risk of cervical cancer was linked clearly in a dose-dependent fashion to the lifetime number of different sexual partners. In contrast, the numbers of sexual acts was not independently linked to risk (Herrero et al., 1990b). The lack of effect of frequency of intercourse on risk, independent of number of partners (Brinton et al., 1987; Herrero et al., 1990b) is concordant with the knowledge that HPV is easily transmitted during vaginal intercourse. The sexual risk factors for cervical cancer can now be seen to be the risk factors for HPV infection (Schiffman et al., 1993). When HPV

1051

infection is taken into account, the effect of lifetime number of partners is virtually eliminated. Age at first intercourse might be viewed logically as a useful proxy for time of HPV infection (Hildesheim et al., 2001), i.e., the start of viral “latency,” although this variable also tends to weaken as a risk factor once HPV infection is taken into account (Deacon et al., 2000; Schiffman et al., 1993). Some investigators have hypothesized that early ages at first intercourse might represent a “vulnerable period” of the cervix when the transforming effect of HPV is greatest. However, attempts to associate risk with the number of different sexual partners at early, specific ages have not yielded consistent support for this hypothesis (Herrero et al., 1990b; Peters et al., 1986b).

Characteristics of the Male Sexual Partner Geographic clusters of cervical and penile cancers (Cartwright et al., 1980), as well as elevated rates of cervical cancer among the wives of men with penile cancer (Hellberg et al., 1989; Smith et al., 1980), raised the suspicion that a “male factor” might be important. This notion was supported by a follow-up study in which the wives of men previously married to cervical cancer patients were found to have elevated rates of cervical neoplasia compared to control wives (Kessler, 1977). Further interest in the role of a male factor derived from findings that some female populations exhibit high incidence rates of cervical cancer, despite traditions of having few sexual contacts, e.g., many Latin American populations. In case-control studies (Brinton et al., 1989b; Kjaer et al., 1991), the husbands of cases were found to report significantly more sexual partners, high-risk sexual behavior, and histories of other sexually transmitted infections than husbands of controls. The few studies that included direct examination of HPV infection in male partners of cervical cancer cases versus controls have confirmed that case partners are more likely to be infected (and that circumcision reduces both HPV infection and female partner risk of cervical cancer) (Castellsague et al., 2002b).

Smoking A correlation between the distribution of cervical cancer and other smoking-related cancers prompted the suggestion (Winkelstein, 1977) that cigarette smoking may affect the risk of cervical cancer. The approximate doubling of risk associated with smoking was not explained by the confounding effects of sexual behavior among smokers versus non-smokers (Brinton et al., 1986c; La Vecchia et al., 1986; Peters et al., 1986b). More recently, case-control and cohort studies among groups of women infected with oncogenic HPV have shown that smokers are at increased risk compared with infected women who do not smoke (Castellsague et al., 2003; Castle et al., 2002d; Deacon et al., 2000; Hildesheim et al., 2001; McIntyreSeltman et al., 2005; Plummer et al., 2003; International Collaboration, 2006). Evidence of increasing risk with increasing intensity and duration of smoking has been found. Current smokers appear to be at higher risk than past smokers, although there is insufficient information regarding the relationship of risk to time from quitting. Several investigations have attempted to define possible mechanisms by which smoking might alter cervical epithelium. Tobaccoderived carcinogens are secreted into the cervix at levels higher than in serum. (McCann et al., 1992; Prokopczyk et al., 1997; Schiffman et al., 1987), suggesting possible genotoxicity. The immunosuppressive effects of smoking (Barton et al., 1988) might enhance the persistence of HPV infection (Burger et al., 1993). In a randomized clinical trial, quitting smoking was associated with increased regression rates of microscopically-identified HPV infections (Szarewski et al., 1996), possibly due to an effect on cell-mediated immunity (Szarewski et al., 2001).

Parity and Other Obstetrical and Gynecologic Events HPV infected women who have many live births are at increased risk of cervical cancer and precancer. The increased risk is only evident among women with several live births, and there is a dose-dependent increase in risk with many live births (e.g., >14 versus 5) (Brinton et al., 1987; Brinton et al., 1989a; Castellsague et al., 2003; Kjaer et al., 1992; Munoz et al., 2002).

1052

PART IV: CANCER BY TISSUE OF ORIGIN

Although the association of multiparity with risk of cervical precancer and cancer among HPV infected women is firmly established, the explanatory mechanism is not clear (Castle et al., 2005c). It has been suggested that pregnancy could influence HPV effects on the cervical epithelium through immunologic or hormone-dependent mechanisms. The virus itself is hormone responsive (de Villiers, 2003). Also, anecdotally, venereal warts may sometimes worsen during pregnancy then regress. Some but not all studies suggest that HPV detection rates may increase slightly during current pregnancies (Hildesheim et al., 2001; Nobbenhuis et al., 2002; Schneider et al., 1987). However, the prevalence of HPV infection is not clearly increased in multiparous women. Alternatively, the effect of pregnancy could reflect cervical trauma during parturition, supported by findings from two (but not all) studies of reduced cervical cancer risk associated with a caesarean section (Bosch et al., 1992). Parturition itself has a potentially genotoxic effect mediated by free radicals related to endogenous nitric oxide release (Gravitt et al., 2001). Finally, nutritional effects of reproduction deserve attention. There is no convincing evidence that the risk of cervical cancer is affected by age at menarche, age at menopause, or character of menses (Brinton et al., 1987). It is noteworthy that age at menopause is not a risk factor, given the age curve of cervical cancer incidence. The plateau in incidence probably does not reflect decline in hormonal levels but, rather, depletion of a susceptible group of women having been infected by an average age of around 20 and having progressed to precancer by an average age of 30.

Oral Contraceptives Studies examining the relationship of oral contraceptive use to cervical cancer risk among HPV infected women are especially complex, with questions arising about the potential for confounding, particularly by the duration of HPV infection and screening behavior (Beral et al., 1988; Brinton et al., 1986b; Brinton et al., 1990; Castle et al., 2005c; Castellsague et al., 2003; Moreno et al., 2002; Smith et al., 2003). Use of oral contraceptives could plausibly potentiate the carcinogenicity of HPV infection, because transcriptional regulatory regions of HPV DNA contain hormone-recognition elements and transformation of cells in vitro with viral DNA is enhanced by hormones (de Villiers, 2003). A recent large multicentric case-control study and meta-analysis found an elevated risk of invasive cervical cancer among HPV-positive women who used oral contraceptives for more than 10 years (Castellsague et al., 2003; Smith et al., 2003). Shorter durations of use were not associated with elevated risk. There is not yet prospective confirmation of the risk of precancer among HPV infected women taking oral contraceptives. Evidence linking oral contraceptives to cervical abnormalities has raised concern about long-acting steroid preparations, notably depotmedroxyprogesterone acetate (DMPA). Although these agents are widely used in many countries, studies evaluating their effects, particularly among HPV infected women are limited (Herrero et al., 1990a; Hildesheim et al., 2001). Additional studies of hormone replacement therapy and menopause in the context of HPV infection are also indicated.

Other Contraceptive Methods In a number of studies prior to measurements of HPV infection, users of barrier methods of contraception (diaphragm and condom) were found to have a low risk of cervical cancer. It is plausible that the diaphragm (like the condom) may protect the cervix from sexually transmitted agents like HPV. It has also been suggested that part of the protection associated with diaphragm use may reflect concurrent use of spermicides, which have anti-viral properties. However, the most common spermicide has no appreciable anti-HPV activity in vitro (Hermonat et al., 1992).

Infectious Agents Other than HPV In the 1970s, herpes simplex virus (HSV-2) was hypothesized to be the sexually transmitted cause of cervical cancer (Aurelian, 1991). Now, HPV infection is known to be the central, necessary cause of cervical cancer, but other sexually transmitted agents could increase

the risk of cervical cancer among HPV infected women. Of the other agents examined, most attention continues to focus on Chlamydia trachomatis. Although residual confounding by some aspect of HPV infection has not been completely ruled out, recent large case control studies comparing cervical cancer cases to oncogenic HPV DNApositive controls have observed Chlamydia trachomatis seropositive women to be at increased risk compared to seronegative women (Smith et al., 2001; Smith et al., 2004). The evidence for HSV-2 is weaker (Cogliano et al., 2005; Smith et al., 2002). Additional infections that have been studied to a lesser extent include syphilis, gonorrhea, cytomegalovirus, Epstein-Barr virus, and bacterial vaginosis. No consistent association with cervical cancer risk has been observed for any one of these agents (Watts et al., 2004). It is interesting that the mildly immunosuppressive retrovirus, HTLV-1, has not been linked to an increased risk of precancer and cancer among HPVinfected women (Castle et al., 2003b). One investigation (Schmauz et al., 1989) but not another (De Sanjose et al., 1994) noted a rise in risk of cervical cancer with multiple, concurrent infections. There is some epidemiologic evidence (Castle et al., 2001; Castle et al., 2003c) that chronic cervicovaginal inflammation regardless of causal agent might increase the oncogenicity of HPV infection. The relatively subtle inflammatory response caused by HPV itself is poorly understood. Some studies (Brinton et al., 1987; Herrero et al., 1990b; Peters et al., 1986b) but not all (Hildesheim et al., 2001) have found frequent vaginal douching, especially with other than vinegar or water, associated with increased cervical cancer risk. If real, a douching association could possibly relate to local tissue damage or to the destruction of normal vaginal flora.

Nutrients The influence of nutrient status on risk of cervical neoplasia has received substantial research attention (Garcia-Closas et al., 2005; Giuliano, 2000; Potischman, 1993). Most studies have utilized casecontrol approaches, assessing dietary intake or blood levels at the time of diagnosis, but some prospective studies and a few clinical trials have been completed. Few studies have been conducted among HPV infected women, using state-of-the-art assessments of nutrient status. Although epidemiologists continue to suspect that diet is important in cervical carcinogenesis, no firm associations between a specific aspect of nutritional status and HPV infection or cervical cancer risk have been established or completely discarded, possibly due to methodologic difficulties. Low serum levels of antioxidants have been associated with increased risk of HPV persistence, precancer, and cancer. However different studies have observed the protective effects of antioxidants to be associated with different, highly correlated micronutrients. Betacarotenoid, lycopene, alpha-carotenoid, and tocopherol (vitamin E) have all been linked to protection, but not consistently. Low folate levels or high homocysteine levels have been linked to risk of cervical cancer, leading to interest in markers of one-carbon metabolism and DNA repair (Weinstein et al., 2001b; Weinstein et al., 2001a). Unfortunately, seven Phase III chemoprevention trials giving folic acid or beta-carotene failed to significantly ameliorate HPV lesions and precancers, although topical administration of retinoic acid did lead to lesion regression, similar to its therapeutic utility on facial warts (Castle et al., 2003c).

Occupational Factors The only professions linked definitely to HPV infection are those associated with an increased risk of exposure to the virus. Thus, commercial sex workers are at an increased risk of clinical outcomes of genital HPV infection, although the point prevalence of HPV DNA in surveys of immunocompetent (HIV-uninfected) commercial sex workers is not necessarily elevated (Kjaer et al., 2000; Kreiss et al., 1992) possibly suggesting immunity in some women following intense exposure. Findings regarding the role of occupational factors in the etiology of cervical cancer have been limited to high rates among cleaners and food preparation workers (Savitz et al., 1995), and waitresses (Kjaerheim et al., 1994), although these studies did not take HPV infection into account.

Cervical Cancer Other occupational studies have centered on the male partner. A high rate of cervical cancer was observed among spouses of men whose work necessitated prolonged absences from home, leading to the suggestion that male extramarital affairs might be responsible (Beral, 1974). Male occupations linked to chemical exposures have not been consistently linked with risk to the female.

Second Cancers Following Cervical Cancer Evaluation of second primary tumors following successful treatment for a first primary tumor can often provide clues to both the genetic and environmental causes of tumors at a specific site. As one might expect, studies of second primary cancers following cervical cancer have indicated excesses in other tumors known or suspected to be linked to HPV infection (Hemminki et al., 2001a; Kleinerman et al., 2003). In a recent SEER-based study, women successfully treated for cervical cancer were at 4.8-fold increased risk of cancer at another HPV-related cancer site, defined as cancers of the anus, esophagus, oropharynx, tonsils, vagina, and vulva (Kleinerman et al., 2003). In this same study, a 2.2-fold increased risk of second primaries in smoking-related sites was observed following cervical cancer, supporting the role of cigarette smoking as a risk factor for the development of cervical cancer (Kleinerman et al., 2003). Other notable associations observed in studies of second primary cancer were attributable to effects of radiation treatment on risk of the second primary tumor.

Risk Factors by Histologic Type The distribution of HPV types and possibly of variants within these types differs by histology. While approximately 50% of both squamous cell carcinomas and cervical adenocarcinomas contain HPV 16, HPV 18 infections are represented in a much higher proportion of cervical adenocarcinomas than squamous cell carcinomas (35–40% versus 10–15%, respectively) (Bosch et al., 2003; Castellsague et al., 2006). Furthermore, recent evidence suggests that HPV 16 infections detected in women diagnosed with cervical adenocarcinoma might be more likely to be from variants of Asian American and/or African lineages than HPV 16 infections detected among women diagnosed with squamous cell tumors, which are more likely to be of the prototypic (originally-described) European lineage (Berumen et al., 2001; Burk et al., 2003; Lizano et al., 1997). These differences in HPV type and possibly variant distribution point to important etiological differences between these two distinct histological types of cervical cancer. As described above, established risk factors for squamous cell carcinomas in addition to HPV infections include cigarette smoking and high parity. A modest association between long-term oral contraceptive use and squamous cell carcinomas has also been demonstrated in populations without effective Pap smear screening programs. Conversely, for cervical adenocarcinomas, cigarette smoking (Berrington de Gonzalez et al., 2004; Green et al., 2003; Lacey et al., 2001) and parity (Altekruse et al., 2003; Munoz et al., 2002) have been shown either to not be strongly associated with or to protect against the development of cervical adenocarcinomas. On the other hand, hormonal factors, including oral contraceptive use (Lacey et al., 1999; Madeleine et al., 2001; Moreno et al., 2002; Ursin et al., 1994), hormone replacement therapy use (Lacey et al., 2000), and increased body mass index (Lacey et al., 2003) are associated with increased risk of cervical adenocarcinoma. In summary, both adenocarcinomas and squamous carcinomas require HPV infection as a necessary causal factor, but cervical adenocarcinomas share non-viral co-factors in common with endometrial adenocarcinoma (Green et al., 2003).

HOST FACTORS INFLUENCING RISK OF CERVICAL CANCER While it is accepted that environmental exposures are primary determinants of cervical cancer risk (Czene et al., 2002), evidence suggests that host factors may be important modulators of the impact of environmental exposures on cervical cancer risk. In particular, it has been suggested that host factors that modulate responses to HPV infections

1053

are likely to play a role in cervical cancer pathogenesis. In this section, we summarize evidence linking host factors to the development of cervical cancer.

Genetic and Familial Susceptibility Studies conducted largely in Scandinavian countries with established nationwide tumor, twin, and other family registries clearly indicate that cervical cancer aggregates in families (Ahlbom et al., 1997; Goldgar et al., 1994; Hemminki et al., 1999; Hemminki et al., 2001b; Lichtenstein et al., 2000; Magnusson et al., 1999; Zelmanowicz et al., 2005). In general, an approximate 2-fold increase in risk of precancer or invasive cervical cancer relative to general population risk is observed in family members of cervical cancer patients. While this elevation in risk among relatives of individuals affected with cervical cancer can be attributed to either shared environmental or genetic effects, there is some evidence that the elevation is explained, at least in part, by shared genetics among family members. Some studies of twins have observed increased risk of disease among monozygotic twins compared to dizygotic twins (Ahlbom et al., 1997). Also, in a study in which degree of genetic relatedness was considered, elevations in risk of precancer and cancer combined were highest for biological first degree full-blooded female relatives (mothers and full-sisters), intermediate for biological half-sisters, and not significantly elevated among adoptive female relatives (mothers or sisters), suggesting a real genetic effect rather than an effect of shared environment (Magnusson et al., 1999). As a caution, it should be noted that while studies of cervical precancer have fairly consistently shown evidence for familial aggregation that is attributable to shared genetics, studies restricted to invasive cervical cancer have not. For example, in an initial study conducted in Sweden that evaluated cervical precancer among twins, concordance among twins in the diagnosis of precancer was seen more often among monozygotic twins than among dizygotic twins, suggesting the presence of a true genetic component to cervical cancer (Ahlbom et al., 1997). In a follow-up to this initial study, registries from Sweden and other Scandinavian countries were combined to allow for an evaluation of invasive cervical cancer, and in this larger study familial aggregation appeared to be fully explained by shared environmental factors rather than shared genetics (Lichtenstein et al., 2000). Precancer can only be detected among women who undergo routine cytological screening and screening practices might correlate among relatives. Therefore familial aggregation studies of precancer might be more likely than studies of invasive cervical cancer to be affected by confounding.

Immunity Results from studies conducted among individuals who are immunocompromised have highlighted the importance of the host immune response as a determinant of disease susceptibility. Women infected with HIV have been shown to be at increased risk of HPV infection and cervical precancer (Moscicki et al., 2000; Palefsky et al., 2003; Strickler et al., 2005). Similarly, female renal transplant recipients have been observed to be at elevated risk of HPV and precancer (Fairley et al., 1994; Penn, 1991). These elevations can not plausibly be explained by increased exposure to HPV, and appear to reflect a true effect of immune dysfunction. In support of this, some studies have shown that prevalence of HPV infections are highest among individuals with the most severe immunosuppression, as reflected by low CD4 counts (Harris et al., 2005; Palefsky et al., 2003; Strickler et al., 2005; Sun et al., 1997). These findings are further supported by the reduction in immune infiltrates observed in HIV+ women with CIN2 or CIN3 compared to HIV-women with CIN2 or CIN3 (Kobayashi et al., 2004). Data on risk of invasive cervical cancer among immunocompromised women is less consistent, although several studies have now reported elevated rates of cervical cancer and other HPV-related anogenital cancers among HIV positive adult women and renal transplant recipients (Grigsby et al., 2001; Massad et al., 2003; Penn, 1991). With the advent of HAART treatment for HIV, studies have been able to evaluate the effect of immune reconstitution (as evidenced by

1054

PART IV: CANCER BY TISSUE OF ORIGIN

increases in CD4 counts) on HPV-associated lesions. Results have been mixed (Ahdieh-Grant et al., 2004; Heard et al., 2002; Heard et al., 2004; Minkoff et al., 2001; Palefsky et al., 2001). However, even in studies that have suggested that HAART treatment is beneficial for HPV-related lesions, lesion regression was not observed in the majority of HIV-positive women with precancer who were treated with HARRT. Taken together, these data suggest that the immune response to HPV is an important determinant of susceptibility to HPV infection and precancer, but that once precancer develops the immune response does not play a critical role in the transition to invasive cancer.

Other Host Factors Affecting Risk of Cervical Cancer Endogenous Hormones Parity and long-term oral contraceptive use have both been identified as co-factors for cervical cancer among HPV-infected women. Limited data also suggest that use of injectable contraceptives is associated with increased risk of cervical cancer. For all of these factors the increase in risk could relate to elevated levels of circulating hormones. This is supported by laboratory and animal studies demonstrating that estrogens and progesterone upregulate HPV transcription and that transgenic mice constitutively expressing HPV oncogenes are prone to the development of cervical tumors after chronic administration of estradiol. These findings have led to the hypothesis that elevated levels of endogenous hormones might be associated with cervical cancer. Very few studies to date have directly evaluated this possibility in humans. In the most extensive study to date, no significant association was observed between circulating levels of estradiol, estrone, estrone-sulfate, progesterone, sex hormone-binding globulin, and dehydroepiandrosterone sulfate and risk of precancer among pre- and post-menopausal women (Shields et al., 2003). However, this study suffered from a relatively small study size and from the measurement of circulating hormones at a single time point. Additional studies are required to more carefully address the potential association between endogenous hormones and cervical cancer.

Predisposing Diseases and Medicines Apart from immunosuppression related to organ transplantation, there is little evidence for medicines associated with risk of cervical cancer (Wang et al., 2005). As a possible exception with some epidemiologic support, psychosocial stress and depression-related immunosuppression have become topics of interest. (Coker et al., 2003).

PATHOGENESIS As shown in Figure 54–2 and described in the section on Precursors above, there are three major, necessary stages in cervical carcinogenesis that can be reproducibly distinguished, studied by molecular epidemiologists, and targeted in prevention programs. These include HPV infection, progression of infection to “precancer,” and invasion. This section on Pathogenesis outlines briefly what is known about the mechanism of transition between each of the stages. “Backward” steps occur also, namely clearance of HPV infection and the less frequent regression of precancer to normalcy. This molecular epidemiologic schema is already a departure from traditional views of cervical carcinogenesis based on histology, like the CIN scale, and is not universally accepted. Advancing biologic understanding is prompting a rapid re-assessment of time-honored means of describing cervical carcinogenesis. Conceivably, interdisciplinary teams could soon describe the pathogenesis of cervical cancer almost entirely in even more molecular terms (e.g., persistent oncogenic HPV infection with genomic integration), related to the interplay of viral and host biomarkers, rather than using the traditional microscopic or macroscopic terms.

HPV Transmission/Acquisition This topic is covered also in Chapter 26. Cervical HPV transmission, which is primarily sexual, is studied best at the molecular level, because types must be distinguished for natural history studies and

because most infections are not microscopically or macroscopically evident. More than 40 types of HPV can infect the cervix, with nononcogenic types more prevalent in aggregate than oncogenic types (Herrero et al., 2000; Herrero et al., 2005). The cervical cancer etiologist is interested in non-oncogenic HPV types as controls to understand why oncogenic types are oncogenic. Cancer prevention specialists consider the non-oncogenic types as equivalent to HPV negativity (Wright et al., 2003a). It is assumed but not proven that the initial site of cervical HPV infection is germinal cells in the basal layer of the epithelium. Infections of the introitus and vagina are as common as cervical infections (an important clue regarding the importance of the transformation zone to carcinogenesis) but vaginal infections virtually never result in cancer (Bauer et al., 1991; Castle et al., 2003a; Winer et al., 2003). Virus probably reaches the germinal cells secondary to minor epithelial injuries during sexual intercourse (Andersson-Ellstrom et al., 1994). HPV-induced lesions appear from scanty genetic data to be monoclonal, suggesting that each lesion derives from a single infected germinal cell. The number of HPV in an infected cell can vary widely, but the average is approximately 50 (Fehrmann et al., 2003). HPV genomes replicate in concert with the infected cell, and the HPV life cycle is tightly tied to the differentiation of the epithelium. Early viral transcripts are detectable in the basal and parabasal layers of the epithelium, whereas capsid production and virion assembly occur in the more superficial layers of the differentiated epithelium (Stoler, 1996). The protein products of the E5, E6, and E7 open reading frames of HPV appear to be principally responsible for HPV neoplastic effects (Fehrmann et al., 2003; Munger, 2002; Stoler, 1996; zur Hausen, 1989; zur Hausen, 2000). The E6 and E7 proteins have been shown to be cooperative transforming proteins in vitro. The E6 protein binds to and promotes the degradation of the p53 tumor suppressor protein by forming a complex requiring the cellular protein E6-AP. E6 proteins also activate telomerase (Fehrmann et al., 2003). The E7 protein is the principal transforming protein. It binds to and inactivates the retinoblastoma (Rb) tumor suppressor protein. The E5 protein is a weaker transforming protein with less clear activity and importance. The transforming proteins of cancer-associated HPV types such as types 16 and 18 have greater binding affinities for tumor suppressor proteins, and greater in vitro transforming abilities, than E6 and E7 proteins from non-oncogenic HPV types like HPV 6 and 11. The net effect of the transforming proteins is to extend the lifespan and DNA synthesis of infected cells, to permit HPV replication. Each HPV type should be considered a separate sexually transmitted infection. Because all oncogenic types are transmitted by the same sexual route, concurrent multiple-type infections are common. The available data, which are limited, seem to indicate that HPV types influence each other minimally (Liaw et al., 2001; Thomas et al., 2000). The typical age of cervical HPV infection is similar to other sexually transmitted infections, with a large peak following sexual initiation. Some, but not all, studies of highly exposed women such as commercial sex workers have shown a significant decrease in the HPV prevalence with age despite continuously high sexual activity, suggesting that loss of viral detection and the development of HPV typespecific immunity to re-infection occur. The most common oncogenic type, HPV 16, is also the most common type in the general population. However, several non-oncogenic types like HPV 53, 61, and 62 are also very common (Herrero et al., 2005).

Development of Microscopic Abnormalities HPV sometimes, but certainly not always, produces the pathognomonic (e.g., koilocytotic) or equivocal morphologic changes that can be diagnosed by light microscopic examination of routinely-prepared tissue biopsies or cytologic preparations (Castle et al., 2002e; Moscicki et al., 2001; Schlecht et al., 2003; Woodman et al., 2001). The warty expansion of the spiny layer of the epithelium characterizing HPV lesions, particularly exophytic ones, apparently results from altered cell differentiation and reduced squamous cell sloughing, rather than increased rate of cell turnover (Steinberg et al., 1985). As

Cervical Cancer a result, the proliferative compartment (the number of cells dividing and synthesizing DNA) is increased in HPV-infected epithelia, but the rate of cell division may not be. In cattle, experimental inoculation of BPV by scarification leads to microscopically evident warts in about 2–3 months. Comparable experimental data are not available for cervical HPV infection. However, some observational data on the pathogenesis of cervical HPV infections, at the cellular level, are now available from large cohorts under study. Among cytologically normal women who are HPV DNA positive at enrollment, the absolute risk of incident abnormal smears rises to a high level (approximately 25% of smears taken) at 1–2 years following enrollment and declines thereafter, returning to baseline (<5% of smears taken) at about 4 years (Castle et al., 2002b; Moscicki et al., 2001; Schlecht et al., 2003; Woodman et al., 2001). Microscopic and macroscopic diagnoses are prone to subjectivity, particularly when mild or equivocal changes are involved. Therefore, misclassification is always a concern when epidemiologists consider how best to relate HPV infection to microscopic diagnoses (including histology that depends on colposcopic recognition of abnormalities). HPV type is arguably more important than microscopic evidence of infection, because non-oncogenic HPV infections can cause microscopic and/or macroscopic abnormalities without implying high risk of cancer (Khan et al., 2005). Most mild cytologic abnormalities are caused by oncogenic types of HPV (Zuna et al., 2005), although clearance is still the usual outcome (Moscicki et al., 2004). As a result of these issues, it is still not clear whether microscopically or macroscopically evident mild abnormalities represent a separate natural history stage from HPV detected by DNA testing alone. Smoking has been observed to be a possible, weak risk factor for cytologic abnormality among HPV infected women, arguing that a difference does exist (Castle et al., 2002e; Moscicki et al., 2001). On the other hand, a large fraction of precancers arise from preceding HPV infections without microscopically diagnosed abnormalities, although the certainty of this statement is limited by the practical limitations of screening frequency in prospective studies. Because the detection of cytologic abnormalities rises with screening intensity, it is probable that rapidly progressing or equivocal microscopic abnormalities would be missed (while longer-duration precancer would be found).

HPV Clearance versus Persistence Natural History of Early HPV Infections Cervical HPV infections tend to clear, as do warts anywhere on the body. Cervical HPV infections remain detectable by PCR for a median of less than one year (Burk, 1999; Evander et al., 1995; Hildesheim et al., 1994; Ho et al., 1998; Richardson et al., 2003). The processes of HPV acquisition and clearance dynamically oppose each other in each cohort of women, to produce the characteristic age distributions as infections are transmitted sexually when women have new partners and then cleared. Persistence tends to increase at older ages (Castle et al., 2005a). The major unresolved question of HPV natural history relates to viral latency. In follow-up studies lasting up to 10 years, it is evident that virtually all HPV infections become non-detectable by sensitive HPV DNA tests, usually within 2 years, except for those that lead to precancer. But little else is known about latency, what might cause reemergence like that seen in renal transplant patients and HIVimmunosuppressed women, and what fractions of precancers or cancers arise following a period of latency. Answers to these questions could greatly affect prevention strategies reliant on HPV DNA detection. The ratio of HPV infections that are persistent vs. newly-acquired increases with age (Castle et al., 2005a).

Immune Response to HPV The immune response to HPV is an important determinant of viral clearance versus persistence and, by extension, a major determinant of cervical cancer risk (Stanley, 2005). Based on animal experiments and data on immunosuppressed individuals summarized earlier, it is assumed that the key immune responses involved in the clearance of

1055

HPV infections are cell-mediated. The two broad classes of cells thought to be involved in the cellular immune response to HPV are antigen-presenting cells and cytotoxic T-lymphocytes (Kadish et al., 2002; Szarewski et al., 2001). Humoral immune response, as measured by antibody levels, is a marker of exposure to HPV but has not been consistently shown to strongly modify risk of progression of viral infection (Ho et al., 2002; Viscidi et al., 2003). The cell-mediated immune response to HPV is clearly an important determinant of cervical cancer risk, but the specific immunological markers of immune protection and risk are not well understood. This is due partly to the complexity of the immune response itself, but also to the fact that the replicative cycle of HPV has evolved to avoid detection by the host immune system (Tindle, 2002). As a result, the natural immune responses to HPV are relatively modest and difficult to detect using currently available laboratory methods. Despite these problems, however, studies conducted to date have demonstrated an association between markers of host immune responses to HPV and cervical neoplasia. More specifically, some but not all studies have shown that the ability of lymphocytes from infected patients to proliferate in vitro after exposure to HPV correlates with regression of cervical lesions (Kadish et al., 1997; Kadish et al., 2002). Lymphocyte responses seen among HPV-infected women without evidence of precancer or cancer were shown to induce the production of cytokines indicative of a cellular immune response that activates T-helper cells involved in stimulating immune cells capable of targeting infected cells (known as a T-helper type 1 response) (de Jong et al., 2004; Tsukui et al., 1996). Finally, the ability of T-cells to specifically target and lyse HPV-infected cells was found to negatively correlate with persistence of HPV infection (Nakagawa et al., 2000). With respect to the specific viral components targeted by the immune response, studies have suggested that cellular immune responses against specific HPV proteins, including the E2 and E6 proteins of the virus, are important in immune recognition of the virus by immune cells, while immune responses against other components of the virus, are less likely to be induced by viral infection (Welters et al., 2003). In fact, evidence from animal and in vitro studies suggests that the E7 viral protein might be able to induce immune tolerance as one of the mechanisms used by the virus to evade the immune system (Tindle, 2002). Most of the studies conducted to date have evaluated systemic immune responses to HPV. However, HPV infections do not have a viremic phase and the infection is limited to epithelial surfaces. Therefore, immune responses occurring at the mucosal surfaces of the genital tract are likely to be better markers of protective and permissive immune responses to HPV than systemic measures of immune response to HPV. This has led more recent studies to focus on measurement of immune responses at the cervix. To date, studies have been largely methodological, and have attempted to understand the correlation (or lack thereof) between systemic and local immune responses and to develop robust collection and laboratory testing methods capable of measuring immune responses at the cervix (Castle et al., 2002a; Hildesheim et al., 2002a). These methods are now beginning to be used in analytical epidemiological studies and it is hoped that results from these studies will improve our understanding of the role of immune responses to HPV in the etiology of cervical cancer.

HLA Studies Perhaps the strongest evidence for a genetic determinant to the natural history of HPV infections and resultant cervical cancer risk comes from work done to relate risk of disease to specific human leukocyte antigen (HLA) alleles and haplotypes (Arias-Pulido et al., 2004; Engelmark et al., 2004; Hildesheim et al., 2002b). HLA genes encode molecules that are responsible for the presentation of viral and other exogenous pathogens to the immune system and are therefore believed to play an important role in the recognition and subsequent clearance of HPV-infected cells. Both Class I HLA genes (those that encode HLA molecules that are present in all nucleated cells) and Class II HLA genes (those that encode HLA molecules that are present in lymphocytes and other immune-related cells) are involved in immune presentation. To date, HLA Class II genes have been

1056

PART IV: CANCER BY TISSUE OF ORIGIN

more extensively studied than HLA Class I genes for their association with cervical cancer. Early studies suggested a possible association between specific HLA Class II alleles/haplotypes, including DQB1*03 alleles and the DRB1*1501-DQB1*0602 haplotype, and disease risk (Apple et al., 1994; Wank et al., 1991). Subsequent studies have been conflicting with respect to these HLA alleles/haplotypes, but have consistently demonstrated that HLA Class II DRB1*13 alleles and haplotypes containing these alleles are associated with decreased disease susceptibility (Hildesheim et al., 2002b). Further support for a protective effect of DRB1*13 alleles is provided by studies that have shown reduced HPV viral load and higher rates of regression of mild HPV histopathology among DRB1*13 carriers (Beskow et al., 2005; Sastre-Garau et al., 2004). To the extent that studies have had the ability to evaluate the HPV type specificity of findings, the associations have tended to be type-specific. Fewer studies have been conducted to evaluate the association between HLA Class I gene polymorphisms and risk of cervical cancer and its precursors. Some studies have observed an association between the HLA Class I B7 allele and cervical cancer, but this has not been seen in other studies (Hildesheim et al., 1998; Wang et al., 2001). Some data suggest that the HLA Class I C*0202 allele is associated with reduced risk of cervical cancer and its precursors (Wang et al., 2002). Since the C*0202 allele is involved not only in acquired immune response to viruses (i.e., the antigen-specific, memory immune responses associated with T-cells of the immune system) but also in the innate immune response to foreign pathogens (i.e., nonspecific, inflammatory immune responses), this finding has prompted the speculation that not only the HPV-type specific T-cell mediated immune responses but also non-specific innate immune responses might be important in the immune response to HPV infections and in cervical cancer pathogenesis (Carrington et al., 2005).

Other Genetic Susceptibility Factors In addition to HLA, genotypic polymorphisms in other genes involved in the immune response or in modulating other disease co-factors such as cigarette smoking have received limited attention. Results from the studies conducted to date have been conflicting. Some have observed disease associations with polymorphisms in immune-regulatory genes (such as TAP1, TAP2, IFNA17, TNF-A) phase I and II metabolism genes (such as GSTM1, GSTT1, CYP1A1), and genes involved in DNA repair and cell cycle control (XRCC1, CCND1), while others have not (Catarino et al., 2004; Chen et al., 1999; Deshpande et al., 2005; Duarte et al., 2005; Goodman et al., 2001; Gostout et al., 2003; Kim et al., 2000; Kim et al., 2003; Kirkpatrick et al., 2005; Niwa et al., 2005; Roh et al., 2001; Sierra-Torres et al., 2003; Ueda et al., 2005a; Ueda et al., 2005b). There is some indication for an association with cervical neoplasia of polymorphisms in the FAS and FAS ligand genes involved in the regulation of apoptosis (Lai et al., 2005; Sun et al., 2005). Careful studies that take into account HPV and the co-factors (e.g., smoking) hypothesized to be modulated by the genotype of interest are needed to resolve current discrepant findings.

Figure 54–5. Five-year persistence of HPV types in Guanacaste, Costa Rica (Source: Schiffman et al., 2005).

The average time of viral persistence that leads to precancer is not known, because long-term longitudinal data are still lacking. Women who are virginal can demonstrate possible precancers within a few years of initial HPV infection detected by DNA. This short time period represents the leading edge of what is typically a longer incidence curve of precancer in persistently HPV infected women. The average age of women with newly-diagnosed precancer is between 25 and 30 (Schiffman, 1992), approximately 10 years after the average peak ages of oncogenic HPV prevalence and associated LSIL in screening populations. Of note, the older literature relying on microscopy to estimate rates of progression to precancer, though voluminous (Ostor, 1993), is of limited use because microscopic and macroscopic data can not distinguish between persistence of an HPV infection and sequential infections with different HPV types (Schiffman et al., 2003). It is not yet known whether the interactions of HPV transforming proteins with tumor suppressor proteins explain the full spectrum of HPV growth-altering effects, from the production of warts and koilocytotic atypia to the rare development of precancer and cancer. Other early proteins of HPV, or other biochemical activities of E6 and E7, are probably also important (zur Hausen, 2000). For example, E5 may downregulate the HLA Class 1 antigen presentation system, perhaps as a mechanism of immune evasion. The degree of nuclear atypia apparently increases with the duration of cervical HPV infections, possibly due to the effects of HPV itself. To reiterate an important caution, although prospective studies have demonstrated that women with HPV infection are at increased risk of

Association of Persistent Oncogenic Infection with Precancer It is the persistence of one of the oncogenic HPV types that is strongly linked to precancer (Ho et al., 1995; Kjaer et al., 2002; Nobbenhuis et al., 1999; Schlecht et al., 2001). We would like to distinguish risk factors for oncogenic HPV persistence from determinants of progression to precancer given viral persistence. The two phenomena are not identical, but thus far they are so closely linked that epidemiologists are only beginning to disentangle them. We do know that HPV type greatly affects both the absolute risk of viral persistence, and of progression to precancer given viral persistence. As shown in Figures 54–5 and 54–6, HPV 16 is uniquely able to persist, and is also remarkably oncogenic given persistence (Castle et al., 2005b; Schiffman et al., 2005b). The other oncogenic types of HPV are not, in aggregate, much more likely to persist than non-oncogenic types of HPV studied in comparison. Non-oncogenic HPV types do not cause precancer or cancer even when they persist (Figure 54–6).

Figure 54–6. Progression to CIN3 or cancer among persistently infected women in Guanacaste, Costa Rica (Source: Schiffman et al., 2005).

Cervical Cancer developing precancer somewhere on the cervix, no one is sure that early HPV associated (CIN 1/LSIL) lesions progress to precancerous lesions. Precancers are usually found more proximal to the endocervical canal than early HPV-associated lesions (Saito et al., 1987), which are commonly further out on the ectocervix. It has been suggested, therefore, that precancers might arise from normal-appearing or atypical metaplastic epithelium at the internal margin of low-grade lesions (Kinney et al., 1998; Kiviat et al., 1992). Cofactors for progression also remain under study. The etiologic cofactors that appear to promote HPV progression, (multiparity, longterm oral contraceptive use, chronic inflammation, and possibly nutrition) have been established by case-control studies, but await prospective confirmation. Epidemiologists have been able to confirm prospectively that smoking is a risk factor for progression to precancer and cancer among HPV infected women (Castle et al., 2002d; Deacon et al., 2000).

Influence of HPV Type Variants on Risk of Persistence and Progression HPV types vary in their carcinogenicity and, within types, variants related to risk can be distinguished. To date, studies on viral variants have mainly focused on HPV 16 and HPV 18, with some work on a few other oncogenic types. Several studies but not all have documented an association between HPV-16 variants and the development of cervical cancer (Beskow et al., 2005; Wang et al., 2003a; Xi et al., 1997). The very limited data available for HPV types other than HPV 16 suggest that variants of HPV 18, 52, and 58 might be associated with risk of cervical cancer (Aho et al., 2004; Hecht et al., 1995; Lizano et al., 1997; Wang et al., 2003a). There might be a relationship between ethnicity, viral variants, and risk (Beskow et al., 2005; Schlecht et al., 2005). The intermediate oncogenicity of certain HPV types, might relate to heterogeneous oncogenic potential of their variants, a topic under active investigation.

1057

by or independent of HPV infection, are also required for progression to cervical cancer. Identification of such somatic alterations, whether they be alterations in tumor suppressor genes, oncogenes, or their products has become the focus of recent work in this area (Wang et al., 2003a). Work conducted to date has evaluated genes known to be affected by HPV (e.g., p53 and pRb) and candidate tumor suppressor genes and oncogenes previously associated with other tumors (Lazo, 1999). In addition, studies aimed at identifying chromosomal regions that might be important in cervical cancer pathogenesis using largely loss-ofheterozygosity (LOH) and comparative genetic hybridization (CGH) assays have identified numerous regions of interest, including chromosomal losses at 3p, 6, 11, 13, 16, 17, and 19, and chromosomal gains at 3q (Kersemaekers et al., 1999; Luft et al., 1999; Matthews et al., 2000). Identification of specific target genes affected in these regions are the focus of current research, as illustrated by work attempting to validate the importance of the FHIT (fragile histidine triad) as a tumor suppressor gene on the short arm of chromosome 3 (Helland et al., 2000; Terry et al., 2004). Interesting associations of risk have also been reported with biomarkers related to p16, telomerase, and aneuploidy (Duensing et al., 2003; Wang et al., 2004; Zhang et al., 2002). In addition to genetic alterations, epigenetic events that alter gene expression (phenotype) without a change in the DNA sequence (genotype) are likely to play a role in cervical cancer pathogenesis. These include hyper- or hypomethylation of viral oncogenes (e.g., addition or removal of a methyl group) with the potential implications for suppression or activation of viral oncogenic expression, respectively. Recent studies have identified silencing of tumor suppressor genes via promoter hypermethylation in HPV-infected host cells as a frequent human epigenetic event, a parallel but distinct event from viral gene methylation (Dong et al., 2001; Feng et al., 2005; Steenbergen et al., 2004).

Risk of Invasion Influence of Co-infection with Multiple HPV Types and Viral Load The amount of HPV DNA in the cervical epithelium is a complex sum of the number and size of the HPV-associated lesions and their pathologic state. Ultra-low viral loads are associated with microscopic normalcy and with low risk of subsequent precancer/cancer but in the clinical setting the prognostic importance of increasingly high viral loads is not at all established (Lorincz et al., 2002; Ylitalo et al., 2000). Some of the highest viral loads can be associated with ultimately resolving CIN 1/LSIL producing large amounts of virus, analogous to benign warts. The viral loads measured in women with precancers, unless the precancers are very large, are often not very high and are determined mainly by the extent of surrounding HPV-infected tissue (£CIN 1) (Sherman et al., 2003b). Cervical cancers do not produce large amounts of intact virus, probably linked to the disruption of the HPV virion that accompanies genomic integration. Further complicating the measurement of viral load is the common occurrence of multiple HPV types due to sexual co-transmission. Cervical cancer is typically a monoclonal event related to a single HPV type, however, the surrounding cervical epithelium can still be infected with other types. As measured by sensitive DNA detection methods, 20% to 30% of women with cervical infections have more than one type, regardless of stage of pathology (Herrero et al., 2005). Finally, there is some evidence that the association of viral load and cancer risk varies by HPV type. Because of great variability and uncertain interpretation, viral load despite its intuitive appeal is not a useful molecular marker of cervical cancer risk.

Somatic Genetic Changes Related to Progression to Precancer HPV persistence related to ineffective host immunologic response is a major determinant of progression to precancer and has already been discussed. Persistent infection with an oncogenic HPV expressing E6 and E7, which interact with pRb and p53, respectively, can be viewed as the “first hit.” Additional somatic genetic events, whether induced

Precancerous lesions tend not to regress over short-term follow-up, however, even among precancerous lesions, risk and timing of invasion versus eventual regression are matters of probability. Peterson (Peterson, 1956) noted a 33% progression rate after 9 years among 127 women with untreated precancer. Longer follow-up would presumably have led to even higher progression rates (Kinlen et al., 1978). Given the ethical constraint against additional cohort studies of this topic, the absolute risk of untreated precancer developing into invasive disease is still argued, with estimates averaging about 30% but ranging from 10% to 90%. The longer precancer persists, the higher the risk of invasion. In any case, the epidemiologic risk factors for invasive cervical cancer among HPV-infected women are the same as mentioned above for precancer, except for age. Screened detected cases of invasive cancer, on average, occur approximately 10 to 15 years later than for precancer suggesting a long sojourn time in the precancerous state. The median age moves toward even older ages as the quality of screening decreases, but the average stage of cancer at diagnosis also worsens. Invasion seems to be a stochastic process requiring additional, genotoxic events with few prominent risk factors appreciable by conventional epidemiology. For example, integration of HPV DNA into the host genome has been suggested to relate to invasive potential, although the evidence has derived from cross-sectional associations (Wang et al., 2003a). The HPV genome is maintained in the cell nucleus and is usually episomal. In contrast, in invasive cervical cancers, integration of HPV DNA into the host genome is found in the majority of cases, particularly in cancers containing HPV 18. Integration tends to occur throughout the cell genome. The frequency of site-specific integration is unclear and might be high. However, with reference to the viral genome, integration seen among cancers is definitely not random suggesting selection by growth advantage. Regulatory elements and the E6 and E7 open reading frames are preserved, with frequent disruption during integration of the E1 and E2 genes that normally inhibit E6 and E7. Thus, continuous production of E6 and E7 proteins appears to have a role

1058

PART IV: CANCER BY TISSUE OF ORIGIN

in HPV carcinogenicity and the E6 and E7 regions of the HPV genome are transcriptionally active in HPV-associated cervical cancers and derived cell lines.

PREVENTIVE MEASURES Primary Prevention Behavioral approaches to prevention of transmission of HPV infections might not be effective, given current patterns of sexual activity. Transmission to the introitus is possible without intromission. Condom use can not entirely prevent the spread of genital HPV infections (Hildesheim et al., 2001; Manhart et al., 2002; Winer et al., 2003), because genital HPV infections in the male are not limited to the penile skin. Viral spread from scrotal and perineal lesions is possible and can not be prevented by condom use. Ultimately, advances in HPV immunology might make most HPV infections and clinical sequelae vaccine-preventable, including genital warts and cervical cancer (Goldie et al., 2004b). Two general classes of vaccines are currently under development: primary prophylactic vaccines designed to prevent initial infection with the virus, and secondary prophylactic (therapeutic) vaccines designed to prevent infections from persisting and progressing to precancer and cancer by targeting HPV-infected cells for destruction by host immune cells. Primary prophylaxis efforts have focused largely on a virus-like particle (VLP) vaccine comprised of the main structural capsid protein of HPV (the L1 protein) multimerized into its correct three-dimensional capsid configuration (Frazer, 2004; Schiller et al., 2004). This capsid looks like real viral capsids, but is non-infectious because it does not contain any viral DNA that would allow for establishment of infection and replication. VLP-based animal papillomavirus vaccines have been shown in animal models to provide near complete protection against type-specific challenge subsequent to vaccination. Three animal models exist in cows, rabbits, and dogs (Breitburd et al., 1995; Kirnbauer et al., 1996; Suzich et al., 1995). Protection in animal studies has been shown to be due to type-specific neutralizing antibodies generated in response to vaccination and to require the threedimensional conformation of the VLP (Suzich et al., 1995). HPV 16 VLP vaccines have been tested in humans for its safety and immunogenicity. Systemic vaccination with HPV 16 VLPs has been shown to be well tolerated and to induce levels of anti-HPV 16 antibodies that are approximately 40 times higher than those observed following natural infection with HPV 16 (Harro et al., 2001). Presence of these antibodies at the cervix has been demonstrated following vaccination (Nardelli-Haefliger et al., 2003). Finally, proof-of-principle trials have shown that vaccination with HPV 16 and HPV 18 VLP-based vaccines confers nearly complete protection against persistent infection with those HPV types. Duration of protection has been demonstrated for up to 4 years with sustained antibody levels. Some type-type crossprotection is evident (Harper et al., 2004; Koutsky et al., 2002; Mao et al., 2006; Villa et al., 2005a).

Chemoprevention and Therapeutic Vaccines Until prophylactic vaccines with long-term durability are available, HPV infections and HPV-induced lesions will continue to be highly prevalent. Research into the proper treatment of HPV can be seen as research into the prevention of cervical cancer. In the United States, the standard response to the detection of HPV and its early associated microscopic signs is initial expectant management (monitoring to observe whether regression), followed by destruction of persistent lesions and the remaining cervical transformation zone by LEEP or cryosurgery (Wright et al., 1992a; Wright et al., 2003a). Some groups have attempted to treat HPV infections of the cervix chemically with topical, destructive chemical applications or Vitamin A analogues (Stanley, 2003). Trials of oral micronutrient supplementation to promote lesion regression are underway, but so far have been unsuccessful (Castle et al., 2003c).

In addition to the prophylactic HPV VLP vaccine, numerous secondary prevention vaccine candidates are under evaluation (Stanley, 2002). Unlike the VLP-based vaccine, these vaccines would not protect against the initial infection, but would target already infected cells and prevent the establishment of persistent and progressive infections. Unlike the VLP-based vaccines, these secondary prevention vaccines would target the induction of cell-mediated rather than humoral immunity. They would promote such responses against nonstructural proteins, including the E6 and E7 proteins that are known to be expressed in cervical cancers and to be important in tumorigenesis, and the E2 protein, which is known to be expressed throughout the natural history of HPV infection. Promising secondary vaccine candidates for which human studies have been published include an HPV 16 E7 protein vaccine in which E7 is fused to a potent adjuvant called heat shock protein Hsp65, and an HPV 16 E7-based encapsulated plasmid DNA vaccine, but numerous other candidates are currently under pre-clinical and clinical evaluation (Goldstone et al., 2002; Roden et al., 2004; Sheets et al., 2003; Stanley, 2002). Evidence to date suggest that these vaccine candidates are safe, that they induce cell-mediated immune responses in animals, and that vaccination of animals induces regression of HPV-induced lesions. However, their efficacy in humans remains to be documented (Roden et al., 2004; Stanley, 2002).

Screening and Early Detection Because of the years of HPV-associated changes preceding cancer, there is little doubt that cervical cytology screening (the Pap test), used to detect treatable cervical precancer and early stages of cancer, can have profound effects on incidence and mortality. The destructive treatment of HPV-induced associated lesions and precancers has resulted in large and convincing declines in cervical cancer rates in areas where screening has been widespread, high-quality, and prolonged (Gustafsson et al., 1997b; Morrison et al., 1996; Peto et al., 2004b). A number of case-control studies have indicated the effectiveness of cervical screening in preventing invasive cervical cancer (Eddy, 1990). However, there is still extensive discussion regarding the details of optimal implementation (Saslow et al., 2002; Sawaya et al., 2003). For example, the American Cancer Society currently recommends that all women who are, or who have been, sexually active for three years, or have reached 21 years of age, have an annual Pap test (Saslow et al., 2002). Very young women should not be over-treated if found to have minor abnormalities, which tend to regress (Sawaya, 2005). Vaginal HPV infection is common following hysterectomy for reasons unrelated to cervical cancer or CIN (Castle et al., 2004a). However, the risk of vaginal cancer is so low that routine screening post-hysterectomy is not recommended (Sirovich et al., 2004). Recommendations for interval of screening differ between professional groups, but all agree that following consecutive satisfactory normal annual examinations, screening may be performed less frequently at the discretion of the woman and her physician. After age 65 to 70, women with normal screening histories might not need further screening (Sawaya et al., 2001). The discovery that persistent, oncogenic HPV infection is a necessary cause for cervical cancer is revolutionizing cervical cancer screening because an obvious corollary is that absence of infection equals lack of cancer risk. HPV DNA testing is already being incorporated into screening programs that have relied previously on cytology. Based on a large randomized clinical trial (ALTS Group, 2003) and other supportive data (Arbyn et al., 2004; Kim et al., 2002; Kulasingam et al., 2006; Manos et al., 1999), HPV DNA testing is now the preferred approach for managing most women with equivocal cervical cytology findings (ASC-US) (Wright et al., 2002a). A single HPV test can identify virtually all women destined to have precancers diagnosed during the subsequent 24 months, while reassuring approximately half of women with equivocal cytology that they are not at appreciably elevated risk of cancer. In terms of cost utility, HPV triage is apparently more effective than two repeat Papanicolaou tests or referral of all women with equivocal cytology to colposcopy (Arbyn et al., 2004; Kim et al., 2002; Kulasingam et al., 2006).

Cervical Cancer

Novel Primary Screening Methods To address some problems with conventional Pap test screening liquid-based cytology and computer-assisted screening have been developed (Sherman, 2003). Liquid-based cytology is already the predominant method in use in the United States, having largely replaced the conventional Pap smear. It is more expensive than the conventional Pap smear, which concerns health planners regarding cost-utility. Computer-assisted screening is still under evaluation and the few FDA-approved methods have not been widely implemented to date. Ancillary screening methods to complement or even replace cytology are emerging. Visual screening denotes the detection of cervical lesions by inspection, sometimes aided by a magnifying eyepiece and tissue stains (Blumenthal et al., 2001; Sankaranarayanan et al., 2003; Wright, 2003). In some very poor regions, clinical inspection may be the only currently affordable strategy, although it tends to have low specificity when used alone by examiners without extensive training, and is insensitive for detecting lesions within the endocervical canal (such as many cases of precancer among older women). However, it might prove cost effective if made sufficiently sensitive (Goldie et al., 2005; Wright et al., 2005). Based on strong prospective evidence, HPV DNA testing has received Food and Drug Administration approval and organizational endorsement by the American Cancer Society and the American College of Obstetrics and Gynecology, but not the United States Preventive Services Task Force, for use in conjunction with cytology for primary cervical cancer screening of women 30 and older (2003a; Kulasingam SL et al., 2002; Mandelblatt et al., 2002; Petry et al., 2003; Saslow et al., 2002; Wright et al., 2003b) (http://www.ahrq.gov/ clinic/3rduspstf/cervcan/cervcanrr.htm). Because HPV DNA prevalence in cytologically normal women declines sharply with age, while HPV prevalence in women with cervical precancer and cancer remains very high regardless of age, the positive predictive value of finding HPV DNA rises with age (Cuzick et al., 1999). Moreover, the accuracy of the Pap test declines with age, due to inadequate sampling of the receding transformation zone (false negatives) and over-diagnosis of atrophy-related changes (false positives). Thus, the use of cytology and HPV testing in women 30 and older defines a subset of patients who remain at appreciable risk and could benefit from closer surveillance (Bulkmang et al., 2005; Cuzick et al., 2003; Franco EL, 2003; Goldie et al., 2004a; IARC, 2005; Khan et al., 2005a; Kim et al., 2005; Peto et al., 2004a; Schiffman and Castle, 2006; Sherman et al., 2003a; Wright et al., 2004). Older women who are cytologically normal and have a negative HPV DNA test are at very low risk for the development of cervical cancer. The increased sensitivity and negative predictive value achieved by using both tests together will permit the screening interval to be safely extended from the currently recommended 1 to 2 years for cytology to 3 or more years using the combined test (compensating for the increased numbers of women who would be classified as text-positive by the use of two tests rather than one). Effective, less-frequent screening would be especially important for under-screened populations at most risk of cervical cancer in the United States and elsewhere (Kuhn et al., 2000). The affordability of screening protocols that depend on more than a single test will depend on low-cost versions of each component technique (Ferreccio et al., 2003; Jeronimo et al., 2003).

FUTURE DIRECTIONS Etiologic epidemiologic research on HPV and cervical cancer will continue to exploit this remarkably useful model of viral carcinogenesis. Based on advances in understanding of natural history, translational researchers will play an important role in research into the next generation of risk biomarkers that define why a small fraction of women with oncogenic HPV infections experience viral persistence and progression toward cancer. There will be major opportunities to evaluate the accuracy and practicality of new tools, like vaccines, lowcost same-day HPV DNA tests, and computer-assisted visual scans. As the primary public health challenge, to bring the benefit of our

1059

increased understanding of HPV in cervical carcinogenesis to the target populations will require adaptation of sophisticated new biotechnology to low-resource settings. There will be numerous opportunities to interact with cost-utility researchers and policy makers. Each of these research areas will be briefly highlighted.

Etiologic Research The search for important HPV co-factors for progression to precancer is likely to dominate future etiologic, epidemiologic research on cervical cancer. Regarding HPV itself, intensive examinations of HPV types and variants related to natural history are underway. It will be broadly informative (not just to HPV research) to understand which viral genetic polymorphisms (and resultant protein products) are associated with neoplastic progression rather than viral clearance. Because the importance of HPV persistence in the pathogenesis of precancer and cancer has been verified, the proper measurement and risk factors for persistence must be clarified. There is some evidence that condom use decreases persistence in infected couples, for unclear reasons (Hogewoning et al., 2003). If cancers arise only from persistent infections, not from molecularly inapparent (completely “latent”) infections that suddenly reactivate, then prevention of cancers should be achievable by screening for virus at ages preceding the usual onset of cancer. The study of HPV persistence and latency requires repeated measurements and extremely reliable HPV testing. It will be necessary to distinguish variants of HPV types to permit, for example, the distinction of new HPV 16 infections from recurrent ones. Regarding host factors, the immune response and viral mechanisms to evade it will definitely be studied with great interest. The interactions of multiple HPV types in mixed cervical infections should be clarified as one pathway to understanding HPV immunity. The role of multiparity is also particularly interesting, because the elevation in risk is firmly established but not at all understood. Through continued descriptive analyses and cohort studies, the decrease in cervical HPV infection rates with increasing age (which is not observed everywhere) should be better understood (Thomas et al., 2004). The separate contributions to the age trend of cohort effects, re-exposure at older ages, and immunologic suppression must be distinguished, because any cohort effect of increasing HPV infection in currently younger women might predict further increases in invasive cervical cancer in the future. Finally, gene and protein expression studies using microarray technology should allow characterization of the tumor microenvironment and identification of gene families (e.g., immune genes) that are upregulated or downregulated at varying stages of cervical carcinogenesis (Alizadeh et al., 1999). This novel technology has etiologic significance but can also lead to the discovery of biomarkers of risk that might be used to triage women infected with HPV. Coupled with lasercapture microdissection, a method for procuring pure cell types from tissues, the assessment of gene expression in thousands of genes simultaneously should soon allow for careful examination of molecular events occurring in the cervical transformation zone compared to surrounding tissue at risk of infection but not carcinogenesis. Microarray analyses can be applied to identify differences in pathogenesis with regard to both host and viral factors. These include comparing genetic events at different stages of the pathogenic process (e.g., cancer versus precancer versus HPV infection), comparing cancer types (e.g., adenocarcinomas, squamous cell carcinomas), and comparing different HPV types and variants. There is broad need for studies to better define any risk factors for the rarely occurring adenocarcinomas and adenosquamous carcinomas of the cervix, whose incidence is increasing and whose epidemiology is less understood.

Translational Research Epidemiologists will participate in series of trials of prophylactic and therapeutic HPV vaccines. Vaccine products and strategies will be chosen by groups of commercial, government, and independent donor groups. Given the expense of vaccine trials, the choice of which vaccine candidates to study will be extremely important. It would

1060

PART IV: CANCER BY TISSUE OF ORIGIN

greatly expedite vaccine research if there were tests of successful immune responses that could be used as surrogate endpoints. As the HPV VLP-based vaccines currently in large, phase III efficacy trials are shown to be effective at preventing HPV infection and the development of precancer, a new era in cervical cancer prevention via the development and implementation of successively more effective vaccines will ensue. However, many challenges will still remain. These include the development of cheaper ways to produce HPV vaccines to reduce costs and increase availability to resource-poor regions of the world, the development of broader coverage, multivalent vaccines that protect against all oncogenic HPV types, the evaluation of alternative vaccine delivery methods that can either maximize efficacy or simplify delivery of the vaccine in low-resource regions of the world, and the evaluation of vaccine efficacy among men and immunocompromised subpopulation (e.g., HIV positive individuals). Future work will also likely attempt to incorporate a secondary prevention (therapeutic) component to the VLP-based prophylactic HPV vaccine to broaden the population that can be targeted by vaccination programs to include not only unexposed women prior to initiation of sexual intercourse but also those already infected with the virus. Regarding cervical cancer screening, epidemiologists will participate in the validation of successive generations of diagnostic assays designed to distinguish best those oncogenic infections destined to progress to precancer and cancer (Schiffman et al., 2005c; Schiffman et al., 2005d). The goal must be to develop an inexpensive test that can be applied in low-resource regions where the problem of cervical cancer is major and uncontrolled. The value of self-collected specimens should be defined, as one means of extending screening ease and coverage (Baldwin et al., 2005; Belinson et al., 2003; Dzuba et al., 2002; Nobbenhuis et al., 2002b). In the United States, the Pap test continues to be the major microscopic screening technique. Promising refinements are likely to include (1) new cell collection instruments designed to optimize sampling of the cervical transformation zone even in older women, (2) transport of cell specimens in improved liquid media permitting fuller automation of slide preparation, and (3) computer-assisted screening or re-screening of smears.

Prevention Research Cervical cytologic screening has reduced cervical cancer where screening has been feasible. Based on new advances in understanding HPV natural history and cervical carcinogenesis, there is an even more tangible and highly motivating opportunity to reduce suffering from cervical cancer worldwide. Prevention research will concentrate on two complementary tracks: screening and vaccines. The critical aspects of screening are wide population coverage, accuracy of screening, and correct management of precancers that are found. Deciding the optimal intervals of screening, exact tests used, and treatment choice (LEEP or cryotherapy) are the secondary issues for the many regions of the world where the goal is to introduce any effective and affordable screening, based on some combination of macroscopic, cytologic, and/or molecular (e.g., inexpensive, same day HPV test) methods. Cryotherapy is a low-cost and safe treatment; improving it would be very important. Forging workable programs of screening and treatment for low-resource regions would save tens of thousands of lives annually. A preventive approach to cervical cancer must be international in scope and sensitive to costs (Denny et al., 2005; Goldie, 2003; Goldie et al., 2005). In the United States and other wealthy regions with existent cervical cytologic screening programs, the cost-utility of new methods must be weighed against established methods (Goldie, 2003). The goal is to provide excellent preventive care for all women. Pap screening no longer varies as greatly by race and income as it once did. However, it is not clear which of the expensive, new technologies should be made universally available. As one very topical example, although there are very real benefits to be gained, the use of HPV DNA testing for primary screening offers considerable potential for overuse if clinical use does not take into account the natural history of HPV. Excessively frequent use of HPV

DNA testing could classify large numbers of women as being at high risk even though their infections are destined to resolve. This would cause considerable anxiety, costs, and over-treatment with possible morbidity. The possibility of overuse is particularly problematic for women in their teens and twenties who are most likely to have new, transient HPV infections, but unlikely to have cervical cancer because of the years- or even decades-long latency between the onset of HPV infection and cancer development. Regardless of how it is implemented, incorporating HPV DNA testing into primary screening will result in millions of women with normal Pap tests being told that they are at risk for having or developing cervical cancer (Schiffman et al., 2005). Colposcopically directed biopsy as routinely practiced for diagnosis is not optimally accurate (Ferris et al., 2006; Jeronimo and Schiffman, in press; Stoler et al., 2001). Combination of virologic, cytologic, and colposcopy-like technologies can provide more accurate risk prediction; affordability is a concern (Ferreccio et al., 2003; Jeronimo et al., 2003; Wang et al., 2005). Importantly, treatment can cause fertility problems in a small percentage of patients (Sadler et al., 2004). This provides the challenge to validate secondary biomarkers that indicate which women are truly at risk of cervical cancer (Wentzensen et al., 2005). The incorporation of HPV testing into the clinical management of women postcolposcopy to access need for treatment and post-treatment as a measure of cure (Arbyn et al., 2005; Schiffman and Castle, 2006; Zielinski et al., 2004) seems promising but will require careful epidemiologic study as well. Because HPV is a central cause of most cervical neoplasia, it is reasonable to consider HPV immunization as the ultimate primary preventive strategy for eliminating most cervical cancer (Villa 2005b). The issue of how to incorporate vaccination programs in regions where successful cervical cancer screening programs are already in place in a cost-effective manner will need to be addressed (Kulasingam et al., 2003; Schiffman et al., 2005). In summary, given adequate resources and emerging, improved prevention tools, cervical cancer should be largely preventable, and not just in wealthy nations. Acknowledgments We thank Jaclyn Dozier for technical assistance. We appreciate the useful suggestions made by our colleagues: Robert Burk, Philip Castle, Julia Gage, Michelle Khan, Diane Solomon, and Sophia Wang.

References ACOG practice bulletin. 2002. Diagnosis and treatment of cervical carcinomas. Number 35, May 2002. American College of Obstetricians and Gynecologists. Int J Gynaecol Obstet 78:79–91. ACOG practice bulletin: clinical management guidelines for obstetriciangynecologists. Number 45, August 2003. Cervical cytology screening (replaces committee opinion 152, March 1995). 2003a. Obstet Gynecol 102:417–427. Ahdieh-Grant L, Li R, Levine AM, et al. 2004. Highly active antiretroviral therapy and cervical squamous intraepithelial lesions in human immunodeficiency virus-positive women. J Natl Cancer Inst 96:1070–1076. Ahlbom A, Lichtenstein P, Malmstrom H, Feychting M, Hemminki K, Pedersen NL. 1997. Cancer in twins: genetic and nongenetic familial risk factors. J Natl Cancer Inst 89:287–293. Aho J, Hankins C, Tremblay C, et al. 2004. Genomic polymorphism of human papillomavirus type 52 predisposes toward persistent infection in sexually active women. J Infect Dis 190:46–52. Alfsen GC, Thoresen SO, Kristensen GB, Skovlund E, Abeler VM. 2000. Histopathologic subtyping of cervical adenocarcinoma reveals increasing incidence rates of endometrioid tumors in all age groups: a population based study with review of all nonsquamous cervical carcinomas in Norway from 1966 to 1970, 1976 to 1980, and 1986 to 1990. Cancer 89:1291–1299. Alizadeh A, Eisen M, Davis RE, et al. 1999. The lymphochip: a specialized cDNA microarray for the genomic-scale analysis of gene expression in normal and malignant lymphocytes. Cold Spring Harb Symp Quant Biol 64:71–78. Altekruse SF, Lacey JV, Jr., Brinton LA, et al. 2003. Comparison of human papillomavirus genotypes, sexual, and reproductive risk factors of cervical adenocarcinoma and squamous cell carcinoma: Northeastern United States. Am J Obstet Gynecol 188:657–663.

Cervical Cancer ALTS Group. 2003. Results of a randomized trial on the management of cytology interpretations of atypical squamous cells of undetermined significance. Am J Obstet Gynecol 188:1383–1392. Andersson-Ellstrom A. Dillner J, Hagmar B, Schiller J, Forssman L. 1994. No serological evidence for non-sexual spread of HPV16. Lancet 344:1435. Apple RJ, Erlich HA, Klitz W, Manos MM, Becker TM, Wheeler CM. 1994. HLA DR-DQ associations with cervical carcinoma show papillomavirustype specificity. Nat Genet 6:157–162. Arbyn M, Buntinx F, Van Ranst M, Paraskevaidis E, Martin-Hirsch P, Dillner J. 2004. Virologic versus cytologic triage of women with equivocal Pap smears: a meta-analysis of the accuracy to detect high-grade intraepithelial neoplasia. J Natl Cancer Inst 96:280–293. Arbyn M, Paraskevaidis E, Martin-Hirsh P, Prendiville W, Dillner J. 2005. Clinical utility of HPV-DNA detection: Triage of minor cervical lesions, follow-up of women treated for high-grade CIN: An update of pooled evidence. Gynecol Oncol (e-pub ahead of publication). Arias-Pulido H, Joste N, Wheeler CM. 2004. Loss of heterozygosity on chromosome 6 in HPV-16 positive cervical carcinomas carrying the DRB1*1501-DQB1*0602 haplotype. Genes Chromosomes Cancer 40:277–284. Aurelian L. 1991. Viruses and carcinoma of the cervix. Contrib Gynecol Obstet 18:54–70. Baldwin S, Santos C, Mendez Brown E, et al. 2005. Comparison of typespecific human papillomavirus data from self and clinician directed sampling. Gynecol Oncol 97:612–617. Barton SE, Maddox PH, Jenkins D, Edwards R, Cuzick J, Singer A. 1988. Effect of cigarette smoking on cervical epithelial immunity: a mechanism for neoplastic change? Lancet 2:652–654. Bauer HM, Ting Y, Greer CE, et al. 1991. Genital human papillomavirus infection in female university students as determined by a PCR-based method. JAMA 265:472–477. Belinson JL, Qiao YL, Pretorius RG, et al. 2003. Shanxi Province cervical cancer screening study II: self-sampling for high-risk human papillomavirus compared to direct sampling for human papillomavirus and liquid based cervical cytology. Int J Gynecol Cancer 13:819–826. Beral V. 1974. Cancer of the cervix: a sexually transmitted infection? Lancet 1:1037–1040. Beral V, Hannaford P, Kay C. 1988. Oral contraceptive use and malignancies of the genital tract. Results from the Royal College of General Practitioners’ Oral Contraception Study. Lancet 2:1331–1335. Beral V, Hermon C, Munoz N, Devesa SS. 1994. Cervical cancer. Cancer Surv 19–20:265–285. Berrington de Gonzalez A, Sweetland S, Green J. 2004. Comparison of risk factors for squamous cell and adenocarcinomas of the cervix: a metaanalysis. Br J Cancer 90:1787–1791. Berumen J, Ordonez RM, Lazcano E, et al. 2001. Asian-American variants of human papillomavirus 16 and risk for cervical cancer: a case-control study. J Natl Cancer Inst 93:1325–1330. Beskow AH, Engelmark MT, Magnusson JJ, Gyllensten UB. 2005. Interaction of host and viral risk factors for development of cervical carcinoma in situ. Int J Cancer 117:690–692. Beskow AH, Moberg M, Gyllensten UB. 2005. HLA class II allele control of HPV load in carcinoma in situ of the cervix uteri. Int J Cancer 117:510–514. Beutner KR. 2000. Nongenital human papillomavirus infections. Clin Lab Med 20:423–430. Blumenthal PD, Gaffikin L, Chirenje ZM, McGrath J, Womack S, Shah K. 2001. Adjunctive testing for cervical cancer in low resource settings with visual inspection, HPV, and the Pap smear. Int J Gynaecol Obstet 72:47–53. Bosch FX, De Sanjosé S. 2003. Chapter 1: Human papillomavirus and cervical cancer—burden and assessment of causality. J Natl Cancer Inst Monogr 31:3–13. Bosch FX, Lorincz A, Munoz N, Meijer CJ, Shah KV. 2002. The causal relation between human papillomavirus and cervical cancer. J Clin Pathol 55:244–265. Bosch FX, Manos MM, Munoz N, et al. 1995. Prevalence of human papillomavirus in cervical cancer: a worldwide perspective. International biological study on cervical cancer (IBSCC) Study Group. J Natl Cancer Inst 87:796–802. Bosch FX, Munoz N, Shah KV, Meheus A. 1992. Second International Workshop on the Epidemiology of Cervical Cancer and Human Papillomaviruses. Int J Cancer 52:171–173. Breitburd F, Kirnbauer R, Hubbert NL, et al. 1995. Immunization with viruslike particles from cottontail rabbit papillomavirus (CRPV) can protect against experimental CRPV infection. J Virol 69:3959–3963. Brinton LA, Fraumeni JF, Jr 1986a. Epidemiology of uterine cervical cancer. J Chronic Dis 39:1051–1065.

1061

Brinton LA, Hamman RF, Huggins GR, et al. 1987. Sexual and reproductive risk factors for invasive squamous cell cervical cancer. J Natl Cancer Inst 79:23–30. Brinton LA, Huggins GR, Lehman HF, et al. 1986b. Long-term use of oral contraceptives and risk of invasive cervical cancer. Int J Cancer 38:339–344. Brinton LA, Reeves WC, Brenes MM, et al. 1989a. Parity as a risk factor for cervical cancer. Am J Epidemiol 130:486–496. Brinton LA, Reeves WC, Brenes MM, et al. 1990. Oral contraceptive use and risk of invasive cervical cancer. Int J Epidemiol 19:4–11. Brinton LA, Reeves WC, Brenes MM, et al. 1989b, The male factor in the etiology of cervical cancer among sexually monogamous women. Int J Cancer 44:199–203. Brinton LA, Schairer C, Haenszel W, et al. 1986c. Cigarette smoking and invasive cervical cancer. JAMA 255:3265–3269. Broker TR, Jin G, Croom-Rivers A, et al. 2001. Viral latency—the papillomavirus model. Dev Biol (Basel) 106:443–451. Bulkmans NW, Rozendaal L, Voorhorst FJ, Snijders PJ, Meijer CJ. 2005. Longterm protective effect of high-risk human papillomavirus testing in population-based cervical screening. Br J Cancer 92:1800–1802. Burger MP, Hollema H, Gouw AS, Pieters WJ, Quint WG. 1993. Cigarette smoking and human papillomavirus in patients with reported cervical cytological abnormality. BMJ 306:749–752. Burk R, Terai M, Gravitt P, et al. 2003. HPV 16/18 variants and cervical adenocarcinomas. Cencer Res, submitted manuscript. Burk RD. 1999. Pernicious papillomavirus infection. N Engl J Med 341: 1687–1688. Carrington M, Wang S, Martin MP, et al. 2005. Hierarchy of resistance to cervical neoplasia mediated by combinations of killer immunoglobulin-like receptor and human leukocyte antigen loci. J Exp Med 201:1069–1075. Cartwright RA, Sinson JD. 1980. Carcinoma of penis and cervix. Lancet 1:97. Castellsague X, Bosch FX, Munoz N. 2002a. Environmental co-factors in HPV carcinogenesis. Virus Res 89:191–199. Castellsague X, Bosch FX, Munoz N, et al. 2002b. Male circumcision, penile human papillomavirus infection, and cervical cancer in female partners. N Engl J Med 346:1105–1112. Castellsague X, Diaz M, de Sanjose S, et al. 2006. Worldwide human papillomavirus etiology of cervical adenocarcinoma and its cofactors: implications for screening and prevention. J Natl Cancer Inst 98:303–315. Castellsague X, Munoz N. 2003. Chapter 3: Cofactors in human papillomavirus carcinogenesis—role of parity, oral contraceptives, and tobacco smoking. J Natl Cancer Inst Monogr 31:20–28. Castle PE, Schiffman M, Bratti MC, et al. 2003a. High prevalence and determinants of vaginal HPV infection in women following hysterectomy for benign indications: a population-based study in Guanacaste, Costa Rica. Castle PE, Escoffery C, Schachter J, et al. 2003b. Chlamydia trachomatis, herpes simplex virus 2, and human T-cell lymphotrophic virus type 1 are not associated with grade of cervical neoplasia in Jamaican colposcopy patients. Sex Transm Dis 30:575–580. Castle PE, Giuliano AR. 2003c. Chapter 4: Genital tract infections, cervical inflammation, and antioxidant nutrients—assessing their roles as human papillomavirus cofactors J Natl Cancer Inst Monogr 31:29–34. Castle PE, Hildesheim A, Bowman FP, et al. 2002a. Cervical concentrations of interleukin-10 and interleukin-12 do not correlate with plasma levels. J Clin Immunol 22:23–27. Castle PE, Hillier SL, Rabe LK, et al. 2001. An association of cervical inflammation with high-grade cervical neoplasia in women infected with oncogenic human papillomavirus (HPV). Cancer Epidemiol Biomarkers Prev 10:1021–1027. Castle PE, Lorincz AT, Mielzynska-Lohnas I, et al. 2002b. Results of human papillomavirus DNA testing with the hybrid capture 2 assay are reproducible. J Clin Microbiol 40:1088–1090. Castle PE, Schiffman M, Bratti MC, et al. 2004a. A population-based study of vaginal human papillomavirus infection in hysterectomized women. J Infect Dis 190:458–467. Castle PE, Schiffman M, Herrero R, et al. 2005a. A prospective study of age trends in cervical human papillomavirus acquisition and persistence in Guanacaste, Costa Rica. J Infect Dis 191:1808–1816. Castle PE, Shields T, Kirnbauer R, et al. 2002c. Sexual behavior, human papillomavirus type 16 (HPV 16) infection, and HPV 16 seropositivity. Sex Transm Dis 29:182–187. Castle PE, Solomon D, Schiffman M, Wheeler CM, ALTS Group. 2005b. Human papillomavirus type 16 infections and 2-year absolute risk of cervical precancer in women with equivocal or mild cytologic abnormalities. J Natl Cancer Inst 97:1066–1071. Castle PE, Wacholder S, Lorincz AT, et al. 2002d. A prospective study of highgrade cervical neoplasia risk among human papillomavirus-infected women. J Natl Cancer Inst 94:1406–1414.

1062

PART IV: CANCER BY TISSUE OF ORIGIN

Castle PE, Wacholder S, Sherman ME, et al. 2002e. Absolute risk of a subsequent abnormal pap among oncogenic human papillomavirus DNApositive, cytologically negative women. Cancer 95:2145–2151. Castle PE, Walker JL, Schiffman M, Wheeler CM, ALTS Group. 2005c. Hormonal contraceptive use, pregnancy and parity, and the risk of cervical intraepithelial neoplasia 3 among oncogenic HPV DNA-positive women with equivocal or mildly abnormal cytology. Int J Cancer 117:1007–1012. Castle PE, Wheeler CM, Solomon D, Schiffman M, Peyton CL, ALTS Group. 2004b. Interlaboratory reliability of Hybrid Capture 2. Am J Clin Pathol 122:238–245. Catarino R, Matos A, Pinto D, et al. 2005. Increased risk of cervical cancer associated with cyclin D1 gene A870G polymorphism. Cancer Genet Cytogenet 160:49–54. Chen C, Madeleine MM, Weiss NS, Daling JR. 1999. Glutathione S-transferase M1 genotypes and the risk of vulvar cancer: a population-based casecontrol study. Am J Epidemiol 150:437–442. Clifford GM, Gallus S, Herrero R, et al. 2005. Worldwide distribution of human papillomavirus types in cytologically normal women in the International Agency for Research on Cancer HPV prevalence surveys: a pooled analysis. Lancet 366:991–998. Clifford GM, Smith JS, Aguado T, Franceschi S. 2003a. Comparison of HPV type distribution in high-grade cervical lesions and cervical cancer: a meta-analysis. Br J Cancer 89:101–105. Clifford GM, Smith JS, Plummer M, Munoz N, Franceschi S. 2003b. Human papillomavirus types in invasive cervical cancer worldwide: a metaanalysis. Br J Cancer 88:63–73. Cogliano V, Baan R, Straif K, Grosse Y, Secretan B, El Ghissassi F. 2005. Carcinogenicity of human papillomaviruses. WHO International Agency for Research on Cancer. Lancet Oncol 6:204. Coker AL, Bond S, Madeleine MM, Luchok K, Pirisi L. 2003. Psychosocial stress and cervical neoplasia risk. Psychosom Med 65:644–651. Cox JT, Schiffman M, Solomon D. 2003. Prospective follow-up suggests similar risk of subsequent cervical intraepithelial neoplasia grade 2 or 3 among women with cervical intraepithelial neoplasia grade 1 or negative colposcopy and directed biopsy. Am J Obstet Gynecol 188:1406–1412. Cuzick J, Szarewski A, Cubie H, et al. 2003. Management of women who test positive for high-risk types of human papillomavirus: the HART study. Lancet 362:1871–1876. Cuzick J, Sasieni P, Davies P, et al. 1999. A systematic review of the role of human papillomavirus testing within a cervical screening programme. Health Technol Assess 3:i–196. Czene K, Lichtenstein P, Hemminki K. 2002. Environmental and heritable causes of cancer among 9.6 million individuals in the Swedish FamilyCancer Database. Int J Cancer 99:260–266. Daling JR, Sherman KJ. 1992. Relationship between human papillomavirus infection and tumours of anogenital sites other than the cervix. IARC Sci Publ223–241. Davey DD, Woodhouse S, Styer P, Stastny J, Mody D. 2000. Atypical epithelial cells and specimen adequacy: current laboratory practices of participants in the college of American pathologists interlaboratory comparison program in cervicovaginal cytology. Arch Pathol Lab Med 124:203–211. de Jong A, van Poelgeest MI, van der Hulst JM, et al. 2004. Human papillomavirus type 16-positive cervical cancer is associated with impaired CD4+ T-cell immunity against early antigens E2 and E6. Cancer Res 64:5449–5455. De Sanjose S, Munoz N, Bosch FX, et al. 1994. Sexually transmitted agents and cervical neoplasia in Colombia and Spain. Int J Cancer 56:358–363. de Villiers EM. 2003. Relationship between steroid hormone contraceptives and HPV, cervical intraepithelial neoplasia and cervical carcinoma. Int J Cancer 103:705–708. de Villiers EM, Fauquet C, Broker TR, Bernard HU, zur Hausen H. 2004. Classification of papillomaviruses. Virology 324:17–27. Deacon JM, EvanS CD, Yule R, et al. 2000. Sexual behaviour and smoking as determinants of cervical HPV infection and of CIN3 among those infected: a case-control study nested within the Manchester cohort. Br J Cancer 83:1565–1572. Denny L, Kuhn L, Pollack A, Wainwright H, Wright TC, Jr. 2000. Evaluation of alternative methods of cervical cancer screening for resource-poor settings. Cancer 89:826–833. Denny L, Kuhn L, De Souza M, Pollack AE, Dupree W, Wright TC Jr. 2005. Screen-and-treat approaches for cervical cancer prevention in lowresource settings: a randomized controlled trial. JAMA 294:2173–2181. Deshpande A, Nolan JP, White PS, et al. 2005. TNF-alpha promoter polymorphisms and susceptibility to human papillomavirus 16-associated cervical cancer. J Infect Dis 191:969–976. Devesa SS, Diamond EL. 1980. Association of breast cancer and cervical cancer incidence with income and education among whites and blacks. J Natl Cancer Inst 65:515–528.

Dillner J. 1999. The serological response to papillomaviruses. Semin Cancer Biol 9:423–430. Dillner J, Lehtinen M, Bjorge T, et al. 1997. Prospective seroepidemiologic study of human papillomavirus infection as a risk factor for invasive cervical cancer. J Natl Cancer Inst 89:1293–1299. Dong SM, Kim HS, Rha SH, Sidransky D. 2001. Promoter hypermethylation of multiple genes in carcinoma of the uterine cervix. Clin Cancer Res 7:1982–1986. Duarte I, Santos A, Sousa H, et al. 2005. G-308A TNF-alpha polymorphism is associated with an increased risk of invasive cervical cancer. Biochem Biophys Res Commun 334:588–592. Duensing S, Munger K. 2003. Centrosomes, genomic instability, and cervical carcinogenesis. Crit Rev Eukaryot Gene Expr 13:9–23. Dzuba IG, Diaz EY, Allen B, et al. 2002. The acceptability of self-collected samples for HPV testing vs. the pap test as alternatives in cervical cancer screening. J Womens Health Gend Based Med 11:265–275. Eddy DM. 1990. Screening for cervical cancer. Ann Intern Med 113:214–226. Engelmark M, Beskow A, Magnusson J, Erlich H, Gyllensten U. 2004. Affected sib-pair analysis of the contribution of HLA class I and class II loci to development of cervical cancer. Hum Mol Genet 13:1951–1958. Evander M, Edlund K, Gustafsson A, et al. 1995. Human papillomavirus infection is transient in young women: a population-based cohort study. J Infect Dis 171:1026–1030. Fairley CK, Chen S, Tabrizi SN, et al. 1994. Prevalence of HPV DNA in cervical specimens in women with renal transplants: a comparison with dialysis-dependent patients and patients with renal impairment. Nephrol Dial Transplant 9:416–420. Fehrmann F, Laimins LA. 2003. Human papillomaviruses: targeting differentiating epithelial cells for malignant transformation. Oncogene 22:5201–5207. Feng Q, Balasubramanian A, Hawes SE, et al. 2005. Detection of hypermethylated genes in women with and without cervical neoplasia. J Natl Cancer Inst 97:273–282. Ferreccio C, Bratti MC, Sherman ME, et al. 2003. A comparison of single and combined visual, cytologic, and virologic tests as cervical cancer screening methods in a region at high risk of cervical cancer. Cancer Epidemiol Biomarkers Prev 12:815–823. Ferris DG, Greenberg MD. 1994. Reid’s Colposcopic Index. J Fam Pract 39:65–70. Ferris DG, Litaker MS, ALTS Group. 2006. Prediction of cervical histologic results using an abbeviated Reid Colposcopic Index during ALTS. Am J Obstet Gynecol 194:704–710. Franco EL. 2000. Statistical issues in human papillomavirus testing and screening. Clin Lab Med 20:345–367. Franco EL. 2003. Chapter 13: Primary screening of cervical cancer with human papillomavirus tests. J Natl Cancer Inst Monogr 31:89–06. Frazer IH. 2004. Prevention of cervical cancer through papillomavirus vaccination. Nat Rev Immunol 4:46–54. Garcia-Closas R, Castellsague X, Bosch X, Gonzalez CA. 2005. The role of diet and nutrition in cervical carcinogenesis: a review of recent evidence. Int J Cancer 117:629–637. Gillison ML, Shah KV. 2003. Chapter 9: Role of mucosal human papillomavirus in nongenital cancers. J Natl Cancer Inst Monogr 31:57–65. Giuliano AR. 2000. The role of nutrients in the prevention of cervical dysplasia and cancer. Nutrition 16:570–573. Goldgar DE, Easton DF, Cannon-Albright LA, Skolnick MH. 1994. Systematic population-based assessment of cancer risk in first-degree relatives of cancer probands. J Natl Cancer Inst 86:1600–1608. Goldie SJ. 2003. Chapter 15: Public health policy and cost-effectiveness analysis. J Natl Cancer Inst Monogr 31:102–110. Goldie SJ, Gaffikin L, Goldhaber-Fiebert JD, et al. 2005. Cost-effectiveness of cervical-cancer screening in five developing countries. N Engl J Med 353:2158–2168. Goldie SJ, Kim JJ, Wright TC. 2004a. Cost-effectiveness of human papillomavirus DNA testing for cervical cancer screening in women aged 30 years or more. Obstet Gynecol 103:619–631. Goldie SJ, Kohli M, Grima D, et al. 2004b. Projected clinical benefits and costeffectiveness of a human papillomavirus 16/18 vaccine. J Natl Cancer Inst 96:604–615. Goldstone SE, Palefsky JM, Winnett MT, Neefe JR. 2002. Activity of HspE7, a novel immunotherapy, in patients with anogenital warts. Dis Colon Rectum 45:502–507. Goodman MT, McDuffie K, Hernandez B, et al. 2001. CYP1A1, GSTM1, and GSTT1 polymorphisms and the risk of cervical squamous intraepithelial lesions in a multiethnic population. Gynecol Oncol 81:263–269. Gostout BS, Poland GA, Calhoun ES, et al. 2003. TAP1, TAP2, and HLA-DR2 alleles are predictors of cervical cancer risk. Gynecol Oncol 88:326– 332.

Cervical Cancer Gravitt PE, Castle PE. 2001. Chlamydia trachomatis and cervical squamous cell carcinoma. JAMA 285:1703–1704. Green J, Berrington de Gonzalez A, Sweetland S, et al. 2003. Risk factors for adenocarcinoma and squamous cell carcinoma of the cervix in women aged 20–44 years: the UK National Case-Control Study of Cervical Cancer. Br J Cancer 89:2078–2086. Grigsby, PW Williamson JF, Clifford Chao KS, Perez CA. 2001. Cervical tumor control evaluated with ICRU 38 reference volumes and integrated reference air kerma. Radiother Oncol 58:19–23. Guido RS, Jeronimo J, Schiffman M, et al. 2005. The distribution of neoplasia arising on the cervix: results from the ALTS trial. Am J Obstet Gynecol 193:1331–1337. Guido R, Schiffman M, Solomon D, Burke L. 2003. Postcolposcopy management strategies for women referred with low-grade squamous intraepithelial lesions or human papillomavirus DNA-positive atypical squamous cells of undetermined significance: a two-year prospective study. Am J Obstet Gynecol 188:1401–1405. Gustafsson L, Ponten J, Bergstrom R, Adami HO. 1997a. International incidence rates of invasive cervical cancer before cytological screening. Int J Cancer 71:159–165. Gustafsson L, Ponten J, Zack M, Adami HO. 1997b. International incidence rates of invasive cervical cancer after introduction of cytological screening. Cancer Causes Control 8:755–763. Harper DM, Franco EL, Wheeler C, et al. 2004. Efficacy of a bivalent L1 viruslike particle vaccine in prevention of infection with human papillomavirus types 16 and 18 in young women: a randomized controlled trial. Lancet 364:1757–1765. Harro CD, Pang YY, Roden RB, et al. 2001. Safety and immunogenicity trial in adult volunteers of a human papillomavirus 16 L1 virus-like particle vaccine. J Natl Cancer Inst 93:284–292. Harris TG, Burk RD, Palefsky JM, et al. 2005. Incidence of cervical squamous intraepithelial lesions associated with HIV serostatus, CD4 cell counts, and human papillomavirus test results. JAMA 293:1471–1476. Heard I, Palefsky JM, Kazatchkine MD. 2004. The impact of HIV antiviral therapy on human papillomavirus (HPV) infections and HPV-related diseases. Antivir Ther 9:13–22. Heard I, Tassie JM, Kazatchkine MD, Orth G. 2002. Highly active antiretroviral therapy enhances regression of cervical intraepithelial neoplasia in HIV-seropositive women. AIDS 16:1799–1802. Hecht JL, Kadish AS, Jiang G, Burk RD. 1995. Genetic Characterization of the human papillomavirus (HPV) 18 E2 gene in clinical specimens suggests the presence of a subtype with decreased oncogenic potential. Int J Cancer 60:369–376. Helland A, Kraggerud SM, Kristensen GB, et al. 2000. Primary cervical carcinomas show 2 common regions of deletion at 3P, 1 within the FHIT gene: evaluation of allelic imbalance at FHIT, RB1 and TP53 in relation to survival. Int J cancer 88:217–222. Hellberg D, Nilsson S. 1989. Genital cancer among wives of men with penile cancer. A study between 1958 and 1982. Br J obstet Gynaecol 96:221–225. Hemminki K, Jiang Y, Dong C. 2001a, Second primary cancers after anogenital, skin, oral, esophageal and rectal cancers: etiological links? Int J Cancer 93:294–298. Hemminki K, Li X, Mutanen P. 2001b. Familial risks in invasive and in situ cervical cancer by histological type. Eur J Cancer Prev 10:83–89. Hemminki K, Vaittinen p. 1999. Familial cancers in a nationwide family cancer database: age distribution and prevalence. Eur J Cancer 35:1109–1117. Herbert A, Singh N, Smith JA. 2001. Adenocarcinoma of the uterine cervix compared with squamous cell carcinoma: a 12-year study in Southampton and South-west Hampshire. Cytopathology 12:26–36. Hermonat PL, Daniel RW, Shah KV. 1992. The spermicide nonoxynol-9 does not inactivate papillomavirus. Sex Transm Dis 19:203–205. Herrero R. 2003. Chapter 7: Human papillomavirus and cancer of the upper aerodigestive tract. J Natl Cancer Inst Monogr 31:47–51. Herrero R, Brinton LA, Reeves WC, et al. 1990a. Injectable contraceptives and risk of invasive cervical cancer: evidence of an association. Int J Cancer 46:5–7. Herrero R, Brinton LA, Reeves WC, et al. 1990b. Sexual behavior, venereal diseases, hygiene practices, and invasive cervical cancer in a high-risk population. Cancer 65:380–386. Herrero R, Castle PE, Schiffman M, et al. 2005. Epidemiologic profile of typespecific human papillomavirus infection and cervical neoplasia in Guanacaste, Costa Rica. J Infect Dis 191:1796–1807. Herrero R, Hildesheim A, Bratti C, et al. 2000. Population-based study of human papillomavirus infection and cervical neoplasia in rural Costa Rica. J Natl Cancer Inst 92:464–474. Hildesheim A, Gravitt P, Schiffman MH, et al. 1993. Determinants of genital human papillomavirus infection in low-income women in Washington, D.C. Sex Transm Dis 20:279–285.

1063

Hildesheim A, Hadjimichael O, Schwartz PE, et al. 1999. Risk factors for rapidonset cervical cancer. Am J Obstet Gynecol 180:571–577. Hildesheim A, Herrero R, Castle PE, et al. 2001. HPV co-factors related to the development of cervical cancer: results from a population-based study in Costa Rica. Br J Cancer 84:1219–1226. Hildesheim A, Ryan RL, Rinehart E, et al. 2002a. Simultaneous measurement of several cytokines using small volumes of biospecimens. Cancer Epidemiol Biomarkers Prev 11:1477–1484. Hildesheim A, Schiffman M, Scott DR, et al. 1998. Human leukocyte antigen class I/II alleles and development of human papillomavirus-related cervical neoplasia: results from a case-control study conducted in the United States. Cancer Epidemiol Biomarkers Prev 7:1035–1041. Hildesheim A, Schiffman MH, Gravitt PE, et al. 1994. Persistence of typespecific human papillomavirus infection among cytologically normal women. J Infect Dis 169:235–240. Hildesheim A, Wang SS. 2002b. Host and viral genetics and risk of cervical cancer: a review. Virus Res 89:229–240. Ho GY, Bierman R, Beardsley L, Chang CJ, Burk RD. 1998. Natural history of cervicovaginal papillomavirus infection in young women. N Engl J Med 338:423–428. Ho GY, Burk RD, Klein S, et al. 1995. Persistent genital human papillomavirus infection as a risk factor for persistent cervical dysplasia. J Natl Cancer Inst 87:1365–1371. Ho GY, Studentsov Y, Hall CB, et al. 2002. Risk factors for subsequent cervicovaginal human papillomavirus (HPV) infection and the protective role of antibodies to HPV-16 virus-like particles. J Infect Dis 186: 737–742. Ho L, Chan SY, Burk RD, et al. 1993. The genetic drift of human papillomavirus type 16 is a means of reconstructing prehistoric viral spread and the movement of ancient human populations. J Virol 67:6413–6423. Hogewoning CJ, Bleeker MC, van den Brule AJ, et al. 2003. Condom use promotes regression of cervical intraepithelial neoplasia and clearance of human papillomavirus: a randomized clinical trial. Int J Cancer 107:811–816. IARC Handbooks of Cancer Prevention Vol. 10: Cervix Cancer Screening. 2005. IARCPress, Lyon. Iftner T, Villa LL. 2003. Chapter 12: Human papillomavirus technologies. J Natl Cancer Inst Monogr 31:80–88. International Collaboration of Epidemiological Studies of Cervical Cancer. 2006. Carcinoma of the cervix and tobacco smoking: collaborative reanalysis of individual data on 13,541 women with carcinoma of the cervix and 23,017 women without carcinoma of the cervix from 23 epidemiological studies. Int J Cancer 118:1481–1495. Jacobson DL, Peralta L, Farmer M, Graham NM, Wright TC, Zenilman J. 1999. Cervical ectopy and the transformation zone measured by computerized planimetry in adolescents. Int J Gynaecol Obstet 66:7–17. Jemal A, Siegel R, Ward E, et al. 2006. Cancer Statistics, 2006. CA Cancer J Clin 56:106–130. Jeronimo J, Castle PE, Herrero R, Burk RD, Schiffman M. HPV testing and visual inspection for cervical cancer screening in resource-poor regions. 2003. Int J Gynaecol Obstet 83:311–313. Jeronimo J, Castle PE, Herrero R, Burk RD, Schiffman M. 2003. HPV testing and visual inspection for cervical cancer screening in resource-poor regions. Int J Gyn Obstet, in press. Kadish AS, Ho GY, Burk RD, et al. 1997. Lymphoproliferative responses to human papillomavirus (HPV) type 16 proteins E6 and E7: outcome of HPV infection and associated neoplasia. J Natl Cancer Inst 89:1285– 1293. Kadish AS, Timmins P, Wang Y, et al. 2002. Regression of cervical intraepithelial neoplasia and loss of human papillomavirus (HPV) infection is associated with cell-mediated immune responses to an HPV type 16 E7 peptide. Cancer Epidemiol Biomarkers Prev 11:483–488. Kersemaekers AM, Van De Vijver MJ, Kenter GG, Fleuren GJ. 1999. Genetic alterations during the progression of squamous cell carcinomas of the uterine cervix. Genes Chromosomes Cancer 26:346–354. Kessler II. 1977. Venereal factors in human cervical cancer: evidence from marital clusters. Cancer 39:1912–1919. Khan MJ, Castle PE, Lorincz AT, et al. 2005a. The elevated 10-year risk of cervical precancer and cancer in women with human papillomavirus (HPV) type 16 or 18 and the possible utility of type-specific HPV testing in clinical practice. J Natl Cancer Inst 97:1072–1079. Khan MJ, Partridge EE, Wang SS, Schiffman M, ALTS Group. 2005b. Socioeconomic status and the risk of cervical intraepithelial neoplasia grade 3 among oncogenic human papillomavirus DNA-positive women with equivocal or mildly abnormal cytology. Cancer 104:61–70. Kim JJ, Wright TC, Goldie SJ. 2002. Cost-effectiveness of alternative triage strategies for atypical squamous cells of undetermined significance. JAMA 287:2382–2390.

1064

PART IV: CANCER BY TISSUE OF ORIGIN

Kim JJ, Wright TC, Goldie SJ. 2005. Cost-effectiveness of human papillomavirus DNA testing in the United Kingdom, The Netherlands, France, and Italy. J Natl Cancer Inst 97:888–895. Kim JW, Lee CG, Park YG, et al. 2000. Combined analysis of germline polymorphisms of p53, GSTM1, GSTT1, CYP1A1, and CYP2E1: relation to the incidence rate of cervical carcinoma. Cancer 88:2082–2091. Kim JW, Roh JW, Park NH, Song YS, Kang SB, Lee HP. 2003. Interferon, alpha 17 (IFNA17) Ile184Arg polymorphism and cervical cancer risk. Cancer Lett 189:183–188. Kinlen LJ, Spriggs AI. 1978. Women with positive cervical smears but without surgical intervention. A follow-up study. Lancet 2:463–465. Kinney WK, Manos MM, Hurley LB, Ransley JE. 1998. Where’s the highgrade cervical neoplasia? The importance of minimally abnormal Papanicolaou diagnoses. Obstet Gynecol 91:973–976. Kirnbauer R, Chandrachud LM, O’Neil BW, et al. 1996. Virus-like particles of bovine papillomavirus type 4 in prophylactic and therapeutic immunization. Virology 219:37–44. Kirkpatrick A, Bidwell J, van den Brule AJ, Meijer CJ, Pawade J, Glew S. 2004. TNF alpha polymorphism frequencies in HPV-associated cervical dysplasia. Gynecol Oncol 92:675–679. Kiviat NB, Critchlow CW, Kurman RJ. 1992. Reassessment of the morphological continuum of cervical intraepithelial lesions: does it reflect different stages in the progression to cervical carcinoma? IARC Sci Publ 59–66. Kjaer SK, Brinton LA. 1993. Adenocarcinomas of the uterine cervix: the epidemiology of an increasing problem. Epidemiol Rev 15:486–498. Kjaer SK, Dahl C, Engholm G, Bock JE, Lynge E, Jensen OM. 1992. Casecontrol study of risk factors for cervical neoplasia in Denmark. II. Role of sexual activity, reproductive factors, and venereal infections. Cancer Causes Control 3:339–348. Kjaer SK, De Villiers EM, Dahl C, et al. 1991. Case-control study of risk factors for cervical neoplasia in Denmark. I: Role of the “male factor” in women with one lifetime sexual partner. Int J Cancer 48:39–44. Kjaer SK, Svare EI, Worm AM, Walboomers JM, Meijer CJ, Van Den Brule AJ. 2000. Human papillomavirus infection in Danish female sex workers. Decreasing prevalence with age despite continuously high sexual activity. Sex Transm Dis 27:438–445. Kjaer SK, Van Den Brule AJ, Paull G, et al. 2002. Type specific persistence of high risk human papillomavirus (HPV) as indicator of high grade cervical squamous intraepithelial lesions in young women: population based prospective follow up study. BMJ 325:572. Kjaerheim K, Andersen A. 1994. Cancer incidence among waitresses in Norway. Cancer Causes Control 5:31–37. Kleinerman R, Hildesheim A, Kosary C. 2006. New malignancies following cancer of the cervix, vagina, and vulva. J Natl Cancer Inst Monograph, in press. Kobayashi A, Greenblatt RM, Anastos K, et al. 2004. Functional attributes of mucosal immunity in cervical intraepithelial neoplasia and effects of HIV infection. Cancer Res 64:6766–6774. Koss LG. 1989. The Papanicolaou test for cervical cancer detection. A triumph and a tragedy. JAMA 261:737–743. Koutsky LA, Ault KA, Wheeler CM, et al. 2002. A controlled trial of a human papillomavirus type 16 vaccine. N Engl J Med 347:1645–1651. Koutsky LA, Holmes KK, Critchlow CW, et al. 1992. A cohort study of the risk of cervical intraepithelial neoplasia grade 2 or 3 in relation to papillomavirus infection. N Engl J Med 327:1272–1278. Kreiss JK, Kiviat NB, Plummer FA, et al. 1992. Human immunodeficiency virus, human papillomavirus, and cervical intraepithelial neoplasia in Nairobi prostitutes. Sex Transm Dis 19:54–59. Kuhn L, Denny L, Pollack A, Lorincz A, Richart RM, Wright TC. 2000. Human papillomavirus DNA testing for cervical cancer screening in low-resource settings. J Natl Cancer Inst 92:818–825. Kulasingam SL, Kim JJ, Lawrence WF, et al. 2006. A cost-effectiveness analysis based on the ASCUS and LSIL Triage Study (ALTS). J Natl Cancer Inst 98:92–100. Kulasingam SL, Hughes JP, Kiviat NB, et al. 2002. Evaluation of human papillomavirus testing in primary screening for cervical abnormalities: comparison of sensitivity, specificity, and frequency of referral. JAMA 288:1749–1759. Kulasingam SL, Myers ER. 2003. Potential health and economic impact of adding a human papillomavirus vaccine to screening programs. JAMA 290:781–789. Kurman RJ, Malkasian GD, Jr., Sedlis A, Solomon D. 1991. From Papanicolaou to Bethesda: the rationale for a new cervical cytologic classification. Obstet Gynecol 77:779–782. La Vecchia C, Franceschi S, Decarli A, Fasoli M, Gentile A, Tognoni G. 1986. Cigarette smoking and the risk of cervical neoplasia. Am J Epidemiol 123:22–29.

Lacey JV, Jr., Brinton LA, Abbas FM, et al. 1999. Oral contraceptives as risk factors for cervical adenocarcinomas and squamous cell carcinomas. Cancer Epidemiol Biomarkers Prev 8:1079–1085. Lacey JV, Jr., Brinton LA, Barnes WA, et al. 2000. Use of hormone replacement therapy and adenocarcinomas and squamous cell carcinomas of the uterine cervix. Gynecol Oncol 77:149–154. Lacey JV, Jr., Frisch M, Brinton LA, et al. 2001. Associations between smoking and adenocarcinomas and squamous cell carcinomas of the uterine cervix (United States). Cancer Causes Control 12:153–161. Lacey JV, Jr., Swanson CA, Brinton LA, et al. 2003. Obesity as a potential risk factor for adenocarcinomas and squamous cell carcinomas of the uterine cervix. Cancer 98:814–821. Lai HC, Lin WY, Lin YW, et al. 2005. Genetic polymorphisms of FAS and FASL (CD95/CD95L) genes in cervical carcinogenesis: An analysis of haplotype and gene-gene interaction. Gynecol Oncol 99:113–118. Lazo PA. 1999. The molecular genetics of cervical carcinoma. Br J Cancer 80:2008–2018. Li H, Jin S, Xu H, Thomas DB. 2000. The decline in the mortality rates of cervical cancer and a plausible explanation in Shandong, China. Int J Epidemiol 29:398–404. Liaw KL, Hildesheim A, Burk RD, et al. 2001. A prospective study of human papillomavirus (HPV) type 16 DNA detection by polymerase chain reaction and its association with acquisition and persistence of other HPV types. J Infect Dis 183:8–15. Lichtenstein P, Holm NV, Verkasalo PK, et al. 2000. Environmental and heritable factors in the causation of cancer–analyses of cohorts of twins from Sweden, Denmark, and Finland. N Engl J Med 343:78–85. Lizano M, Berumen J, Guido MC, Casas L, Garcia-Carranca A. 1997. Association between human papillomavirus type 18 variants and histopathology of cervical cancer. J Natl Cancer Inst 89:1227–1231. Lorincz AT, Castle PE, Sherman ME, et al. 2002. Viral load of human papillomavirus and risk of CIN3 or cervical cancer. Lancet 360:228–229. Lowy DR, Frazer IH. 2003. Chapter 16: Prophylactic human papillomavirus vaccines. J Natl Cancer Inst Monogr 31:111–116. Luft F, Gebert J, Schneider A, Melsheimer P, Von Knebel DM. 1999. Frequent allelic imbalance of tumor suppressor gene loci in cervical dysplasia. Int J Gynecol Pathol 18:374–380. Madeleine MM, Daling JR, Schwartz SM, et al. 2001. Human papillomavirus and long-term oral contraceptive use increase the risk of adenocarcinoma in situ of the cervix. Cancer Epidemiol Biomarkers Prev 10:171–177. Magnusson PK, Sparen P, Gyllensten UB. 1999. Genetic link to cervical tumours. Nature 400:29–30. Mandelblatt JS, Lawrence WF, Womack SM, et al. 2002. Benefits and costs of using HPV testing to screen for cervical cancer. JAMA 287:2372– 2381. Manhart LE, Koutsky LA. 2002. Do condoms prevent genital HPV infection, external genital warts, or cervical neoplasia? A meta-analysis. Sex Transm Dis 29:725–735. Manos MM, Kinney WK, Hurley LB, et al. 1999. Identifying women with cervical neoplasia: using human papillomavirus DNA testing for equivocal Papanicolaou results. JAMA 281:1605–1610. Mao C, Koutsky LA, Ault KA, et al. 2006. Efficacy of human papillomavirus16 vaccine to prevent cervical intraepithelial neoplasia: a randomized controlled trial. Obstet Gynecol 107:18–27. Massad LS, Behbakht K, Collins YC, Cejtin HE. 2003. Histologic findings from the cervix among older women with abnormal cervical cytology. Gynecol Oncol 88:340–344. Matthews CP, Shera KA, Mcdougall JK. 2000. Genomic changes and HPV type in cervical carcinoma. Proc Soc Exp Biol Med 223:316–321. Mccann MF, Irwin DE, Walton LA, Hulka BS, Morton JL, Axelrad CM. 1992. Nicotine and cotinine in the cervical mucus of smokers, passive smokers, and nonsmokers. Cancer Epidemiol Biomarkers Prev 1:125–129. McIntyre-Seltman K, Castle PE, Guido R, Schiffman M, Wheeler CM, ALTS Group. 2005. Smoking is a risk factor for cervical intraepithelial neoplasia grade 3 among oncogenic human papillomavirus DNA-positive women with equivocal or mildly abnormal cytology. Cancer Epidemiol Biomarkers Prev 14:1165–1170. Minkoff H, Ahdieh L, Massad LS, et al. 2001. The effect of highly active antiretroviral therapy on cervical cytologic changes associated with oncogenic HPV among HIV-infected women. AIDS 15:2157–2164. Moreno V, Bosch FX, Munoz N, et al. 2002. Effect of oral contraceptives on risk of cervical cancer in women with human papillomavirus infection: the IARC multicentric case-control study. Lancet 359:1085–1092. Morrison BJ, Coldman AJ, Boyes DA, Anderson GH. 1996. Forty years of repeated screening: the significance of carcinoma in situ. Br J Cancer 74:814–819. Moscicki AB, Ellenberg JH, Vermund SH, et al. 2000. Prevalence of and risks for cervical human papillomavirus infection and squamous intraepithelial

Cervical Cancer lesions in adolescent girls: impact of infection with human immunodeficiency virus. Arch Pediatr Adolesc Med 154:127–134. Moscicki AB, Hills N, Shiboski S, et al. 2001. Risks for incident human papillomavirus infection and low-grade squamous intraepithelial lesion development in young females. JAMA 285:2995–3002. Moscicki AB, Shiboski S, Hills NK, et al. 2004. Regression of low-grade squamous intra-epithelial lesions in young women. Lancet 364:1678– 1683. Munger K. 2002. The role of human papillomaviruses in human cancers. Front Biosci 7:d641–d649. Munoz N, Bosch FX, De Sanjose S, et al. 2003. Epidemiologic classification of human papillomavirus types associated with cervical cancer. N Engl J Med 348:518–527. Munoz N, Franceschi S, Bosetti C, et al. 2002. Role of parity and human papillomavirus in cervical cancer: the IARC multicentric case-control study. Lancet 359:1093–1101. Nakagawa M, Stites DP, Patel S, et al. 2000. Persistence of human papillomavirus type 16 infection is associated with lack of cytotoxic T lymphocyte response to the E6 antigens. J Infect Dis 182:595–598. Nardelli-Haefliger D, Wirthner D, Schiller JT, et al. 2003. Specific antibody levels at the cervix during the menstrual cycle of women vaccinated with human papillomavirus 16 virus-like particles. J Natl Cancer Inst 95:1128–1137. Niwa Y, Matsuo K, Ito H, et al. 2005. Association of XRCC1 Arg399Gln and OGG1 Ser326Cys polymorphisms with the risk of cervical cancer in Japanese subjects. Gynecol Oncol 99:43–49. Nobbenhuis MA, Helmerhorst TJ, Van Den Brule AJ, et al. 2002. High-risk human papillomavirus clearance in pregnant women: trends for lower clearance during pregnancy with a catch-up postpartum. Br J Cancer 87:75–80. Nobbenbuis et al. 2002b. Primary screening for high risk HPV by home obtained cervicovaginal lavage is an alternative screening tool for unscreened women. J Clin Pathol 55:435–439. Nobbenhuis MA, Walboomers JM, Helmerhorst TJ, et al. 1999. Relation of human papillomavirus status to cervical lesions and consequences for cervical-cancer screening: a prospective study. Lancet 354:20–25. Ong CK, Chan SY, Campo MS, et al. 1993. Evolution of human papillomavirus type 18: an ancient phylogenetic root in Africa and intratype diversity reflect coevolution with human ethnic groups. J Virol 67:6424–6431. Ostor AG. 1993. Natural history of cervical intraepithelial neoplasia: a critical review. Int J Gynecol Pathol 12:186–192. Palefsky JM, Holly EA. 2003. Chapter 6: Immunosuppression and co-infection with HIV. J Natl Cancer Inst Monogr 31:41–46. Palefsky JM, Holly EA, Efirdc JT, et al. 2005. Anal intraepithelial neoplasia in the highly active antiretroviral therapy era among HIV-positive men who have sex with men. AIDS 19:1407–1414. Palefsky JM, Holly EA, Ralston ML, et al. 2001. Effect of highly active antiretroviral therapy on the natural history of anal squamous intraepithelial lesions and anal human papillomavirus infection. J Acquir Immune Defic Syndr 28:422–428. Parikh S, Brennan P, Boffetta P. 2003. Meta-analysis of social inequality and the risk of cervical cancer. Int J Cancer 105:687–691. Parkin DM, Bray FI, Devesa SS. 2001. Cancer burden in the year 2000. The global picture. Eur J Cancer 37 Suppl 8:S4–66. Penn I. 1991. Cancer in the immunosuppressed organ recipient. Transplant Proc 23:1771–1772. Peters RK, Chao A, Mack TM, Thomas D, Bernstein L, Henderson BE. 1986a. Increased frequency of adenocarcinoma of the uterine cervix in young women in Los Angeles County. J Natl Cancer Inst 76:423–428. Peters RK, Thomas D, Hagan DG, Mack TM, Henderson BE. 1986b. Risk factors for invasive cervical cancer among Latinas and non-Latinas in Los Angeles County. J Natl Cancer Inst 77:1063–1077. Peterson O. 1956. Spontaneous course of cervical precancerous conditions. Am J Obstet Gynecol 72:1063–1071. Peto J, Gilham C, Deacon J, et al. 2004a. Cervical HPV infection and neoplasia in a large population-based prospective study: the Manchester cohort. Br J Cancer 91:942–953. Peto J, Gilham C, Fletcher O, Matthews FE. 2004b. The cervical cancer epidemic that screening has prevented in the UK. Lancet 364:249–256. Petry KU, Menton S, Menton M, et al. 2003. Inclusion of HPV testing in routine cervical cancer screening for women above 29 years in Germany: results for 8466 patients. Br J Cancer 88:1570–1577. Pham TH, Nguyen TH, Herrero R, et al. 2003. Human papillomavirus infection among women in South and North Vietnam. Int J Cancer 104:213–220. Pirog EC, Kleter B, Olgac S, et al. 2000. Prevalence of human papillomavirus DNA in different histological subtypes of cervical adenocarcinoma. Am J Pathol 157:1055–1062.

1065

Plummer M, Herrero R, Franceschi S, et al. 2003. Smoking and cervical cancer: pooled analysis of the IARC multi-centric case–control study. Cancer Causes Control 14:805–814. Potischman N. 1993. Nutritional epidemiology of cervical neoplasia. J Nutr 123:424–429. Prokopczyk B, Cox JE, Hoffmann D, Waggoner SE. 1997. Identification of tobacco-specific carcinogen in the cervical mucus of smokers and nonsmokers. J Natl Cancer Inst 89:868–873. Richardson H, Kelsall G, Tellier P, et al. 2003. The natural history of typespecific human papillomavirus infections in female university students. Cancer Epidemiol Biomarkers Prev 12:485–490. Richart RM, Barron BA. 1969. A follow-up study of patients with cervical dysplasia. Am J Obstet Gynecol 105:386–393. Robboy SJ, Noller KL, O’Brien P, et al. 1984. Increased incidence of cervical and vaginal dysplasia in 3,980 diethylstilbestrol-exposed young women. Experience of the National Collaborative Diethylstilbestrol Adenosis Project. JAMA 252:2979–2983. Roden RB, Ling M, Wu TC. 2004. Vaccination to prevent and treat cervical cancer. Hum Pathol 35:971–982. Roh J, Kim M, Kim J, et al. 2001. Polymorphisms in codon 31 of p21 and cervical cancer susceptibility in Korean women. Cancer Lett 165:59–62. Sadler L, Saftlas A, Wang W, Exeter M, Whittaker J, McCowan L. 2004. Treatment for cervical intraepithelial neoplasia and risk of preterm delivery. JAMA 291:2100–2106. Saito K, Saito A, Fu YS, Smotkin D, Gupta J, Shah K. 1987. Topographic study of cervical condyloma and intraepithelial neoplasia. Cancer 59:2064– 2070. Sankaranarayanan R, Wesley R, Thara S, et al. 2003. Test characteristics of visual inspection with 4% acetic acid (VIA) and Lugol’s iodine (VILI) in cervical cancer screening in Kerala, India. Int J Cancer 106:404– 408. Sasieni P, Adams J. 2001. Changing rates of adenocarcinoma and adenosquamous carcinoma of the cervix in England. Lancet 357:1490–1493. Saslow D, Runowicz CD, Solomon D, et al. 2002. American Cancer Society guideline for the early detection of cervical neoplasia and cancer. CA Cancer J Clin 52:342–362. Sastre-Garau X, Cartier I, Jourdan-Da Silva N, De Cremoux P, Lepage V, Charron D. 2004. Regression of low-grade cervical intraepithelial neoplasia in patients with HLA-DRB1*13 genotype. Obstet Gynecol 104:751–755. Savitz DA, Andrews KW, Brinton LA. 1995. Occupation and cervical cancer. J Occup Environ Med 37:357–361. Sawaya GF. 2005. A 21-year-old woman with atypical squamous cells of undetermined significance. JAMA 294:2210–2218. Sawaya GF, McConnell KJ, Kulasingam SL, et al. 2003. Risk of cervical cancer associated with extending the interval between cervical-cancer screenings. N Engl J Med 349:1501–1509. Sawaya GF, Sung HY, Kearney KA, et al. 2001. Advancing age and cervical cancer screening and prognosis. J Am Geriatr Soc 49:1499–1504. Schiffman M, Castle PE. 2005a. The promise of global cervical-cancer prevention. N Engl J Med 353:2101–2104. Schiffman M, Castle PE. 2006. When to test women for human papillomavirus. BMJ 332:61–62. Schiffman M, Herrero R, Desalle R, et al. 2005b. The carcinogenicity of human papillomavirus types reflects viral evolution. Virology 337:76–84. Schiffman M, Khan MJ, Solomon D, et al. 2005c. A study of the impact of adding HPV types to cervical cancer screening and triage tests. J Natl Cancer Inst 19:147–150. Schiffman M, Wheeler CM, Dasgupta A, Solomon D, Castle PE, The ALTS Group. 2005d. A comparison of a prototype PCR assay and hybrid capture 2 for detection of carcinogenic human papillomavirus DNA in women with equivocal or mildly abnormal papanicolaou smears. Am J Clin Pathol 124:722–732. Schiffman M, Kjaer SK. 2003. Chapter 2: Natural history of anogenital human papillomavirus infection and neoplasia. J Natl Cancer Inst Monogr 31:14–19. Schiffman MH. 1992. Recent progress in defining the epidemiology of human papillomavirus infection and cervical neoplasia. J Natl Cancer Inst 84:394–398. Schiffman MH, Bauer HM, Hoover RN, et al. 1993. Epidemiologic evidence showing that human papillomavirus infection causes most cervical intraepithelial neoplasia. J Natl Cancer Inst 85:958–964. Schiffman MH, Castle P. 2003. Epidemiologic studies of a necessary causal risk factor: human papillomavirus infection and cervical neoplasia. J Natl Cancer Inst 95:E2. Schiffman MH, Haley NJ, Felton JS, et al. 1987. Biochemical epidemiology of cervical neoplasia: measuring cigarette smoke constituents in the cervix. Cancer Res 47:3886–3888.

1066

PART IV: CANCER BY TISSUE OF ORIGIN

Schiffman MH, Schatzkin A. 1994. Test reliability is critically important to molecular epidemiology: an example from studies of human papillomavirus infection and cervical neoplasia. Cancer Res 54:1944s–1947s. Schiller JT, Davies P. 2004. Delivering on the promise: HPV vaccines and cervical cancer. Nat Rev Microbiol 2:343–347. Schlecht NF, Burk RD, Palefsky JM, et al. 2005. Variants of human papillomaviruses 16 and 18 and their natural history in human immunodeficiency virus-positive women. J Gen Virol 86:2709–2720. Schlecht NF, Kulaga S, Robitaille J, et al. 2001. Persistent human papillomavirus infection as a predictor of cervical intraepithelial neoplasia. JAMA 286:3106–3114. Schlecht NF, Platt RW, Duarte-Franco E, et al. 2003. Human papillomavirus infection and time to progression and regression of cervical intraepithelial neoplasia. J Natl Cancer Inst 95:1336–1343. Schmauz R, Okong P, De Villiers EM, et al. 1989. Multiple infections in cases of cervical cancer from a high-incidence area in tropical Africa. Int J Cancer 43:805–809. Schneider A, Hotz M, Gissmann L. 1987. Increased prevalence of human papillomaviruses in the lower genital tract of pregnant women. Int J Cancer 40:198–201. Schwartz SM, Weiss NS. 1986. Increased incidence of adenocarcinoma of the cervix in young women in the United States. Am J Epidemiol 124: 1045–1047. Sheets EE, Urban RG, Crum CP, et al. 2003. Immunotherapy of human cervical high-grade cervical intraepithelial neoplasia with microparticledelivered human papillomavirus 16 E7 plasmid DNA. Am J Obstet Gynecol 188:916–926. Sherman ME. 2003. Chapter 11: Future directions in cervical pathology. J Natl Cancer Inst Monogr 31:72–79. Sherman ME, Lorincz AT, Scott DR, et al. 2003a. Baseline cytology, human papillomavirus testing, and risk for cervical neoplasia: a 10-year cohort analysis. J Natl Cancer Inst 95:46–52. Sherman ME, Wang SS, Wheeler CM, Rich L, Gravitt PE, Schiffman M. 2003b. Determinants of human papillomavirus load among women with histologic CIN3: dominant impact of surrounding low-grade lesions. Cancer Epidemiol Biomarkers Prev, in press. Shields TS, Falk RT, Herrero R, et al. 2003. A case-control study of endogenous hormones and cervical cancer. Br J Cancer 90:146–152. Shields TS, Falk RT, Herrero R, et al. 2004. A case-control study of endogenous hormones and cervical cancer. Br J Cancer 90:146–152. Sierra-Torres CH, Au WW, Arrastia CD, et al. 2003. Polymorphisms for chemical metabolizing genes and risk for cervical neoplasia. Environ Mol Mutagen 41:69–76. Sirovich BE, Welch HG. 2004. Cervical cancer screening among women without a cervix. JAMA 291:2990–2993. Smith HO, Tiffany MF, Qualls CR, Key CR. 2000. The rising incidence of adenocarcinoma relative to squamous cell carcinoma of the uterine cervix in the United States—a 24-year population-based study. Gynecol Oncol 78:97–105. Smith JS, Bosetti C, Munoz N, et al. 2004. Chlamydia trachomatis and invasive cervical cancer: a pooled analysis of the IARC multicentric casecontrol study. Int J Cancer 111:431–439. Smith JS, Green J, Berrington DG, et al. 2003. Cervical cancer and use of hormonal contraceptives: a systematic review. Lancet 361:1159– 1167. Smith JS, Herrero R, Bosetti C, et al. 2002. Herpes simplex virus-2 as a human papillomavirus cofactor in the etiology of invasive cervical cancer. J Natl Cancer Inst 94:1604–1613. Smith JS, Munoz N, Franceschi S, Eluf-Neto J, Herrero R, Peeling RW. 2001. Chlamydia trachomatis and cervical squamous cell carcinoma. JAMA 285:1704–1706. Smith PG, Kinlen LJ, White GC, Adelstein AM, Fox AJ. 1980. Mortality of wives of men dying with cancer of the penis. Br J Cancer 41:422– 428. Solomon D. 2003. Chapter 14: Role of triage testing in cervical cancer screening. J Natl Cancer Inst Monogr 31:97–101. Solomon D, Davey D, Kurman R, et al. 2002. The 2001 Bethesda System: terminology for reporting results of cervical cytology. JAMA 287:2114– 2119. Solomon D, Schiffman M, Tarone R. 2001. Comparison of three management strategies for patients with atypical squamous cells of undetermined significance: baseline results from a randomized trial. J Natl Cancer Inst 93:293–299. Stanley M. 2003. Chapter 17: Genital human papillomavirus infections— current and prospective therapies. J Natl Cancer Inst Monogr 31:117–124. Stanley MA. 2002. Human papillomavirus vaccines. Curr Opin Mol Ther 4:15–22. Stanley M. 2005. Immune responses to human papillomavirus. Vaccine. [Epub ahead of print].

Steenbergen RD, Kramer D, Braakhuis BJ, et al. 2004. TSLC1 gene silencing in cervical cancer cell lines and cervical neoplasia. J Natl Cancer Inst 96:294–305. Steinberg BM, Abramson AL, Kahn LB, Hirschfield L, Freiberger I. 1985. Vocal cord polyps: biochemical and histologic evaluation. Laryngoscope 95:1327–1331. Stoler MH. 1996. A brief synopsis of the role of human papillomaviruses in cervical carcinogenesis. Am J Obstet Gynecol 175:1091–1098. Stoler MH, Schiffman M. 2001. Interobserver reproducibility of cervical cytologic and histologic interpretations: realistic estimates from the ASCUSLSIL Triage Study. JAMA 285:1500 –1505. Stone KM, Karem KL, Sternberg MR, et al. 2002. Seroprevalence of human papillomavirus type 16 infection in the United States. J Infect Dis 186:1396–1402. Strickler HD, Burk RD, Fazzari M, et al. 2005. Natural history and possible reactivation of human papillomavirus in human immunodeficiency viruspositive women. J Natl Cancer Inst 97:577–586. Sun T, Zhou Y, Li H, et al. 2005. FASL -844C polymorphism is associated with increased activation-induced T cell death and risk of cervical cancer. J Exp Med 202:967–974. Sun XW, Kuhn L, Ellerbrock TV, Chiasson MA, Bush TJ, Wright TC, Jr. 1997. Human papillomavirus infection in women infected with the human immunodeficiency virus. N Engl J Med 337:1343–1349. Suzich JA, Ghim SJ, Palmer-Hill FJ, et al. 1995. Systemic immunization with papillomavirus L1 protein completely prevents the development of viral mucosal papillomas. Proc Natl Acad Sci U S A 92:11553–11557. Szarewski A, Jarvis MJ, Sasient P, et al. 1996. Effect of smoking cessation on cervical lesion size. Lancet 347:941–943. Szarewski A, Maddox P, Royston P, et al. 2001. The effect of stopping smoking on cervical Langerhans’ cells and lymphocytes. BJOG 108:295–303. Terry G, Ho L, Londesborough P, Cross P, Lopes A, Monaghan J, Cuzick J. 2004. The role of human papillomavirus type 16 and the fragile histidine triad gene in the outcome of cervical neoplastic lesions. Br J Cancer 91:2056–2062. Thomas JO, Herrero R, Omigbodun AA, et al. 2004. Prevalence of papillomavirus infection in women in Ibadan, Nigeria: a population-based study. Br J Cancer 90:638–645. Thomas KK, Hughes JP, Kuypers JM, et al. 2000. Concurrent and sequential acquisition of different genital human papillomavirus types. J Infect Dis 182:1097–1102. Thompson CH, Rose BR, Cossart YE. 1992. Detection of HPV DNA in archival specimens of cervical cancer using in situ hybridisation and the polymerase chain reaction. J Med Virol 36:54–59. Tindle RW. 2002. Immune evasion in human papillomavirus-associated cervical cancer. Nat Rev Cancer 2:59–65. Tsukut T, Hildesheim A, Schiffman MH, et al. 1996. Interleukin 2 production in vitro by peripheral lymphocytes in response to human papillomavirusderived peptides: correlation with cervical pathology. Cancer Res 56: 3967–3974. Ueda M, Hung YC, Terai Y, et al. 2005a. Fas gene promoter -670 polymorphism (A/G) is associated with cervical carcinogenesis. Gynecol Oncol 98:129–133. Ueda M, Hung YC, Terai Y, et al. 2005b. Glutathione-S-transferase and p53 polymorphisms in cervical carcinogenesis. Gynecol Oncol 96:736–740. Ursin G, Peters RK, Henderson BE, D’Ablaing G, III, Monroe KR, Pike MC. 1994. Oral contraceptive use and adenocarcinoma of cervix. Lancet 344:1390–1394. Ursin G, Pike MC, Preston-Martin S, D’Ablaing G, III, Peters RK. 1996. Sexual, reproductive, and other risk factors for adenocarcinoma of the cervix: results from a population-based case-control study (California, United States). Cancer Causes Control 7:391–401. Van Ranst M, Kaplan JB, Burk RD. 1992. Phylogenetic classification of human papillomaviruses: correlation with clinical manifestations. J Gen Virol 73 (Pt 10):2653–2660. Villa LL, Costa RL, Petta CA, et al. 2005a. Prophylactic quadrivalent human papillomavirus (types 6, 11, 16, and 18) L1 virus-like particle vaccine in young women: a randomised double-blind placebo-controlled multicentre phase II efficacy trial. Lancet Oncol 6:271–278. Villa LL. 2005b. Prophylactic HPV vaccines: Reducing the burden of HPVrelated diseases. Vaccine. [Epub ahead of print] Viscidi R, Schiffman M, Herrero R, et al. 2003. Seropositivity to human papillomavirus (HPV) types 16, 18, or 31 and risk of subsequent HPV infection: results from a poluation-based study in Cosat Rica. Vizcaino AP, Moreno V, Bosch FX, et al. 2000. International trends in incidence of cervical cancer: II. Squamous-cell carcinoma. Int J Cancer 86:429–435. Vizcaino AP, Moreno V, Bosch FX, Munoz N, Barros-Dios XM, Parkin DM. 1998. International trends in the incidence of cervical cancer: I.

Cervical Cancer Adenocarcinoma and adenosquamous cell carcinomas. Int J Cancer 75: 536–545. Wacholder S. 2003. Chapter 18: Statistical issues in the design and analysis of studies of human papillomavirus and cervical neoplasia. J Natl Cancer Inst Monogr 31:125–130. Waggoner SE. 2003. Cervical cancer. Lancet 361:2217–2225. Walboomers JM, Jacobs MV, Manos MM, et al. 1999. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol 189:12–19. Wallin KL, Wiklund F, Angstrom T, et al. 1999. Type-specific persistence of human papillomavirus DNA before the development of invasive cervical cancer. N Engl J Med 341:1633–1638. Wang SS, Hildesheim A. 2003a. Chapter 5: Viral and host factors in human papillomavirus persistence and progression. J Natl Cancer Inst Monogr 31:35–40. Wang SS, Hildesheim A, Gao X, et al. 2002. Comprehensive analysis of human leukocyte antigen class I alleles and cervical neoplasia in 3 epidemiologic studies. J Infect Dis 186:598–605. Wang SS, Sherman ME, Hildesheim A, Lacey JV, Devesa S. 2003b. Incidence trends for cervical adenocarcinoma and squamous cell carcinoma among white women and black women in the United States from 1976 through 2000. Wang SS, Trunk M, Schiffman M, et al. 2004. Validation of p16INK4a as a marker of oncogenic human papillomavirus infection in cervical biopsies from a population-based cohort in Costa Rica. Cancer Epidemiol Biomarkers Prev 13:1355–1360. Wang SS, Schiffman M, for the ALTS Group. 2005. Medication use, medical conditions, and the risk of human papillomavirus infection and subsequent cervical intraepithelial neoplasia 3 among women with mild cytologic abnormalities. Cancer Epidemiol Biomarkers Prev 14:542–545. Wang SS, Walker JL, Schiffman M, Solomon D, ALTS Group. 2005. Evaluating the risk of cervical precancer with a combination of cytologic, virologic, and visual methods. Cancer Epidemiol Biomarkers Prev 14:2665–2668. Wang SS, Wheeler CM, Hildesheim A, et al. 2001. Human leukocyte antigen class I and II alleles and risk of cervical neoplasia: results from a population-based study in Costa Rica. J Infect Dis 184:1310–1314. Wank R, Thomssen C. 1991. High risk of squamous cell carcinoma of the cervix for women with HLA-DQw3. Nature 352:723–725. Watts DH, Fazzari M, Minkoff H, et al. 2004. Effects of bacterial vaginosis and other genital infections on the natural history of human papillomavirus infection in HIV-1-infected and high-risk HIV-1-uninfected women. J Infect Dis 191:1129–1139. Weinstein SJ, Ziegler RG, Frongillo EA, Jr., et al. 2001a. Low serum and red blood cell folate are moderately, but nonsignificantly associated with increased risk of invasive cervical cancer in U.S. women. J Nutr 131: 2040–2048. Weinstein SJ, Ziegler RG, Selhub J, et al. 2001b. Elevated serum homocysteine levels and increased risk of invasive cervical cancer in US women. Cancer Causes Control 12:317–324. Welters MJ, De Jong A, Van Den Eeden SJ, et al. 2003. Frequent display of human papillomavirus type 16 E6-specific memory t-Helper cells in the healthy population as witness of previous viral encounter. Cancer Res 63:636–641. Wentzensen N, Bergeron C, Cas F, Eschenbach D, Vinokurova S, von Knebel Doeberitz M. 2005. Evaluation of a nuclear score for p16(INK4a)-stained cervical squamous cells in liquid-based cytology samples. Cancer (Epub ahead of print). Winer RL, Lee SK, Hughes JP, Adam DE, Kiviat NB, Koutsky LA. 2003. Genital human papillomavirus infection: incidence and risk factors

1067

in a cohort of female university students. Am J Epidemiol 157:218– 226. Winkelstein W, Jr. 1977. Smoking and cancer of the uterine cervix: hypothesis. Am J Epidemiol 106:257–259. Woodman CB, Collins S, Winter H, et al. 2001. Natural history of cervical human papillomavirus infection in young women: a longitudinal cohort study. Lancet 357:1831–1836. Wright TC, Ferenczy A. 2002. Anatomy and histology of the cervix. In: Kurman, R. J. Blaustein’s Pathology of the Female Genital Tract. New York: Springer-Verlag.os Wright TC, Jr. 2003. Chapter 10: Cervical cancer screening using visualization techniques. J Natl Cancer Inst Monogr 31:66–71. Wright TC, Jr., Cox JT, Massad LS, Carlson J, Twiggs LB, Wilkinson EJ. 2003a. 2001 consensus guidelines for the management of women with cervical intraepithelial neoplasia. Am J Obstet Gynecol 189:295–304. Wright TC, Jr., Cox JT, Massad LS, Twiggs LB, Wilkinson EJ. 2002a. 2001 Consensus Guidelines for the management of women with cervical cytological abnormalities. JAMA 287:2120 –2129. Wright TC, Jr., Denny L, Kuhn L, Goldie S. 2002b. Use of visual screening methods for cervical cancer screening. Obstet Gynecol Clin North Am 29:701–734. Wright TC, Jr., Gagnon S, Richart RM, Ferenczy A. 1992a. Treatment of cervical intraepithelial neoplasia using the loop electrosurgical excision procedure. Obstet Gynecol 79:173–178. Wright TC, Jr., Richart RM, Ferenczy A, Koulos J. 1992b. Comparison of specimens removed by CO2 laser conization and the loop electrosurgical excision procedure. Obstet Gynecol 79:147–153. Wright TC, Jr., Schiffman M. 2003b. Adding a test for human papillomavirus DNA to cervical-cancer screening. N Engl J Med 348:489–490. Wright TC Jr, Schiffman M, Solomon D, et al. 2004. Interim guidance for the use of human papillomavirus DNA testing as an adjunct to cervical cytology for screening. Obstet Gynecol 103:304–309. Xi LF, Koutsky LA, Galloway DA, et al. 1997. Genomic variation of human papillomavirus type 16 and risk for high grade cervical intraepithelial neoplasia. J Natl Cancer Inst 89:796–802. Yamada T, Manos MM, Peto J, et al. 1997. Human papillomavirus type 16 sequence variation in cervical cancers: a worldwide perspective. J Virol 71:2463–2472. Ylitalo N, Sorensen P, Josefsson AM, et al. 2000. Consistent high viral load of human papillomavirus 16 and risk of cervical carcinoma in situ: a nested case-control study. Lancet 355:2194–2198. Zelmanowicz A de M, Schiffman M, Herrero R, et al. 2005. Family history as a co-factor for adenocarcinoma and squamous cell carcinoma of the uterine cervix: results from two studies conducted in Costa Rica and the United States. Int J Cancer 116:599–605. Zhang A, Zheng C, Hou M, et al. 2002. Amplification of the telomerase reverse transcriptase (hTERT) gene in cervical carcinomas. Genes Chromosomes Cancer 34:269–275. Zielinski GD, Bais AG, Helmerhorst TJ, et al. 2004. HPV testing and monitoring of women after treatment of CIN 3: review of the literature and meta-analysis. Obstet Gynecol Surv 59:543–553. Zuna RE, Wang SS, Rosenthal DL, et al. 2005. Determinants of human papillomavirus-negative, low-grade squamous intraepithelial lesions in the atypical squamous cells of undetermined significance/low-grade squamous intraepithelial lesions triage study (ALTS). Cancer 105:253– 262. zur Hausen H. 1989. Papillomaviruses in anogenital cancer as a model to understand the role of viruses in human cancers. Cancer Res 49:4677–4681. zur Hausen H. 2000. Papillomaviruses causing cancer: evasion from host-cell control in early events in carcinogenesis. J Natl Cancer Inst 92:690–698.

55

Cancers of the Vulva and Vagina MARGARET M. MADELEINE AND JANET R. DALING

V

ulvar and vaginal cancers are rare throughout the world. In Africa and Asia the age-adjusted incidence rates of vulvar and vaginal cancer vary from 0.1 to 1.4 per 100,000 women. Higher rates for invasive vulvar cancer were reported in Europe and the Americas. For example, a rate of 1.7 per 100,000 women was reported for the Czech Republic and for the Biella Province of Italy. In Central and South America the highest vulvar cancer rate, 2.5 per 100,000 women, was reported in Concordia, Argentina. In North America the vulvar cancer rates were somewhat lower: 1.3 in Canada and 1.6 in the United States (IARC, 2002a, 2002b). Vaginal cancer rates are lower than vulvar cancer rates worldwide, with the highest vaginal cancer rate, 1.4 per 100,000, reported for Goiania, Brazil and Vila Nova de Gaia, Portugal. Incidence rates below 1.0 per 100,000 women are reported for vaginal cancer in most parts of the world, such as 0.4 per 100,000 in the United Kingdom, 0.7 per 100,000 in Bombay, 0.4 per 100,000 in black South African women, 0.6 per 100,000 in Canada, and 0.5 per 100,00 in New Zealand (IARC, 2002a and 2002b). Approximately 6,000 women will be diagnosed with vulvar or vaginal cancer and 800 deaths will be attributed to each of these cancers in the United States in 2003 (ACS, 2003). The highest agespecific rates for invasive disease occur in older women (SEER, 2003). Representative details of incidence and survival rates presented in this chapter were obtained from the SEER registry, which has collected information from population-based cancer registries throughout the United States since the early 1970s. Vaginal and vulvar cancers share some etiologic features with cervical cancers, and they often occur in women with a prior history of cervical cancer (Sherman et al., 1991; Daling et al., 2002). When appropriate, findings from the cervical cancer literature that supplement our discussion of these diseases will be included. Human papillomavirus (HPV) plays a central role in the etiology of a majority of the squamous cell cancers (SCC) of the vagina and vulva, although some SCC tumors are not HPV-related (Crum, 1992; Kurman et al., 1992). Smoking is another key risk factor for vaginal and vulvar neoplasms. Although diethylstilbestrol (DES) is no longer prescribed for pregnant women, the role of in utero exposure to DES in the genesis of clear cell adenocarcinomas of the vagina continues to be of interest. Information on both in situ and invasive lesions of the vagina and vulva will be presented in this chapter. The term neoplasm will be used when referring to both in situ and invasive lesions.

CLASSIFICATION Anatomic Distribution All vaginal and the majority of vulvar neoplasms are reported to the SEER registry without an exact anatomic location. When a subsite is recorded for vulvar neoplasms, the labia majorum and labia minorum are most often specified (SEER, 2003).

Histopathology Most vulvar neoplasms reported to the SEER registry are squamous cell in origin (92%), with 2% basal cell, 2% melanoma, 2% Paget’s, and 2% other histologies. Approximately 86% of vaginal neoplasms

1068

have a squamous cell origin, with 8% adenocarcinomas (of which 2% are clear cell adenocarcinomas), 2% melanomas, and 4% other histologies. Squamous cell vulvar cancer is characterized by three patterns: warty, basaloid, or keratinizing. Warty and basaloid patterns can appear together, and are classified according to the predominant pattern or as mixed (Anderson, 1991). These neoplasms are more likely to be HPVrelated. HPV-related vulvar neoplasms are non-keratinizing, more often multifocal, and are found in women with a younger mean age at diagnosis. Keratinizing vulvar tumors with an undifferentiated (simplex) morphology occur in older women, are more likely to be HPV negative, and sometimes appear with lichen sclerosis or epithelial hyperplasia (Kurman et al., 1992; Crum et al., 1997).

Precursor Neoplastic Lesions The incidence of in situ neoplasms of the vulva (also referred to as vulvar intraepithelial neoplasia or VIN3) in the United States, at 2.0 per 100,000 women, is higher than the rate for invasive neoplasms, 1.6 per 100,000, when all occurrences of these neoplasms are recorded (Table 55–1). Vaginal neoplasms (vaginal intraepithelial neoplasia or VAIN3) do not share this distinction: the overall rate of in situ disease is 0.4 per 100,000 women compared with 0.6 per 100,000 women for invasive disease. For both vulvar and vaginal neoplasia there has been an increase in the number of in situ cancers reported to the registry, while the rate of invasive cancer has remained stable (Table 55–1). Progression from superficially invasive to invasive cancer of the vulva was described for 12 of 26 patients in a follow up study in Boston (Herod et al., 1996). A recent study (McNally et al., 2002) found that 3% of women with in situ lesions progressed to invasive disease post-treatment. Among women diagnosed with VIN3 in a case series in North Carolina, 37% had recurrent disease (Modesitt et al., 1998). Recurrence rates have been reported to be higher for vulvar compared to cervical or vaginal intraepithelial neoplasia (Wright, 1992). Over 50% of women with multifocal VIN were reported to have recurrences in two hospital-based follow-up studies (McNally et al., 2002; Modesitt, 1998). Reports of vulvar cancer precursors in the literature generally describe squamous cell vulvar intraepithelial neoplasia (VIN) without making a distinction between differentiated (simplex) and nondifferentiated VIN. Yang and Hart (2000) recently described a small case series of 12 simplex VIN. They concluded that underdiagnosis of simplex VIN or a shorter time to invasion may explain why it is less frequently detected compared to HPV-related VIN, the most common vulvar precursor lesion. Vaginal intraepithelial neoplasia (VAIN) is the precursor lesion for squamous cell invasive vaginal cancer. The natural history of VAIN is similar to that of cervical intraepithelial neoplasia (Gallup, 1975). Progression of VAIN to invasive disease has been described in a number of studies (Aho et al., 1991; Lenehan et al., 1986; Minucci et al., 1995).

MOLECULAR GENETIC CHARACTERISTICS Oncogenic HPV types have been documented in 99% of cervical tumor tissue collected in a worldwide, comprehensive study, with

1069

Cancers of the Vulva and Vagina Table 55–1. Incidence Rates of Vaginal and Vulvar Neoplasia in Four 7year Periods, US SEER data 1973–2000a Vulva

Table 55–2. Age-Specific Incidence Rates of Vaginal and Vulvar Neoplasms Over Time, U.S. SEER data 1973–2000a

Vagina

Diagnosis Years

In situ

Invasive

Total

In situ

Invasive

1973–1979 1980–1986 1987–1993 1994–2000 1973–2000

1.0 1.6 2.1 2.7 2.0

1.5 1.5 1.6 1.7 1.6

2.5 3.1 3.7 4.4 3.6

0.3 0.3 0.4 0.5 0.4

0.6 0.6 0.5 0.5 0.6

Vulva

Vagina

Total

Diagnosis Years

15–34

35–54

55–74

15–34

35–54

55–74

0.9 0.9 0.9 1.0 1.0

1973–1979 1980–1986 1987–1993 1994–2000 1973–2000

0.6 1.0 1.3 1.4 1.1

1.9 2.7 3.5 4.6 3.4

4.3 4.4 5.3 6.3 5.1

0.1 0.1 0.2 0.2 0.2

0.6 0.6 0.7 0.8 0.7

1.7 1.9 1.8 2.0 1.9

a

Incidence rates include all races, ages, and histologies and are from SEER (2003). The rates are per 100,000 women and age-adjusted to the 2000 U.S. standard.

a Incidence rates for in situ and invasive disease combined, includea all races and histologies (SEER, 2003). Rates are reported per 100,000 women and age-adjusted to the 2000 U.S. standard. Only women aged 15–74 years are included.

HPV-16, -18, -31, and -45 the most frequently occurring types in cervical tumor tissue (Bosch et al., 1995; Walboomers et al., 1999). In a recently published population-based study, the HPV DNA prevalence in tumor tissue from SCC anogenital cancers was similar to that for cervical cancer: 89.3% for invasive cervical cancer, 89.0% for vulvar neoplasias and 90.7% for vaginal neoplasias (Carter et al., 2001). HPV-related vulvar and vaginal cancers likely accumulate genetic changes induced by the expression of high-risk HPV E6 and E7 genes, similar to the changes that lead to cervical cancer (reviewed by zur Hausen, 2002). Different comparative genome hybridization profiles for HPVnegative and HPV-positive vulvar cancers were recently reported (Allen et al., 2002). Although the number of cases studied was small (n = 18), the findings reinforced the hypothesis that some vulvar cancers do not involve HPV infection. In another study, more loss of heterozygosity was detected among HPV negative compared with HPV positive VIN, suggesting that more genetic changes are necessary for oncogenesis in HPV negative VIN (Flowers et al., 1999).

The incidence rates for both tumor types did not differ substantially by race. In contrast to invasive cervical cancer rates that begin to plateau at ages 60 to 64, invasive vulvar and vaginal rates continue to increase with age without plateauing (Fig. 55–1). The most steeply rising incidence rate is seen for women 35 to 54 years old. The incidence rates of in situ vaginal and vulvar lesions have been increasing for all age groups over time (Table 55–2). This pattern has been especially acute for in situ vulvar cancer (Fig. 55–2).

DEMOGRAPHIC PATTERNS

Overall relative survival rates are higher for women with vulvar (81%) as compared with vaginal (55%) cancer (Table 55–3, SEER 2003). For vulvar cancer, stage at diagnosis was predictive of prognosis at 5 years: 90% for locally invasive disease and 58% for more advanced stages. There was also an increased risk of death associated with age over 55, irrespective of stage of disease. Race did not influence stagespecific survival. The overall 5-year relative survival rate was 72% for women with local vaginal cancer and decreased to 40% for women diagnosed with regional or distant disease. Survival decreased with increasing age for all stages. There is some indication that women diagnosed in more recent years have had lower survival rates for local disease and better survival for late stage disease. This may indicate a change in staging practices for vaginal cancer.

ENVIRONMENTAL FACTORS In addition to the initiating and promoting role that HPV oncoproteins play in many SCC vulvar and vaginal cancers, cofactors other than HPV must play an important part in cancer development. Several

25 20 15 10 5 0 20 -2 4 25 ye -2 ar s 9 30 ye -3 ar s 4 35 ye -3 ar s 9 40 ye -4 ar s 4 45 ye -4 ar s 9 50 ye a -5 4 rs 55 ye -5 ar s 9 60 ye a -6 4 rs 65 ye -6 ar s 9 70 ye -7 ar s 4 75 ye -7 ar s 9 80 ye -8 ar s 4 y 85 ear s + ye ar s

Incidence/100,000 women

Sturgeon et al. (1992) used SEER data to show that the incidence rate of squamous cell in situ vulvar cancer was increasing while the rate of invasive cancer was stable. Levi and co-workers (1998) drew similar conclusions using the cancer registry in Vaud, Switzerland, to examine vulvar cancer rates between 1974–1994. In recent SEER data (2003), the incidence rates for in situ and invasive vulvar cancer combined have increased from 2.5 per 100,000 women in 1973–1979 to 4.4 per 100,000 women in 1994–2000. In contrast, the incidence rate for vaginal neoplasias has remained steady, with rates of 0.9 per 100,000 in 1973–979 and 1.0 per 100,000 in 1994–2000, 20 years later (Table 55–1).

SURVIVAL

Age at diagnosis Cervix

Vulva

Vagina

Figure 55–1. Age-specific incidence rates per 100,000 women diagnosed with invasive cervical, vulvar, and vaginal cancer from 1973 to 2000 in the U.S. SEER data. All races and histologies are included, and the rates are age-adjusted to the 2000 U.S. standard (SEER, 2003).

1070

PART IV: CANCER BY TISSUE OF ORIGIN Incidence/100,000 women

8.0 7.0 6.0 5.0 4.0 3.0 2.0 1.0

Figure 55–2. Age-specific incidence rates per 100,000 women diagnosed with in situ vulvar cancer from 1973 to 2000 in the U.S. SEER data. All races and histologies are included, and the rates are age-adjusted to the 2000 U.S. standard (SEER, 2003).

Age at diagnosis 1973-1979

natural history studies have shown that most women are infected with HPV at onset of sexual activity, but only a subset of those infected develop persistent HPV-related neoplasia (Holowaty et al., 1999). For cervical cancer, the main cofactors of HPV are smoking, parity, hormone use, and immunosupression. Although the mechanisms are not well understood by which these factors promote the genetic insta-

Table 55–3. Relative 5-year Survival Among Women Diagnosed with Vaginal or Vulvar Cancer between 1973 and 2000 from U.S. SEER Data Invasive Vulvar Cancer Stage at Diagnosis Overall

Overall Race White Black Other Age 15–34 35–54 55–74 Period 1973–1979 1980–1986 1987–1993 1994–2000 Overall Race White Black Other Age 15–34 35–54 55–74 Period 1973–1979 1980–1986 1987–1993 1994–2000

Local

25

20

-2 4

ye -2 ars 9 30 ye -3 ar s 4 35 ye -3 ars 9 40 ye -4 ars 4 45 ye -4 ar s 9 50 ye a -5 4 rs 55 ye -5 ars 9 60 ye -6 ars 4 65 ye -6 ar s 9 70 ye -7 ars 4 75 ye -7 ars 9 80 ye -8 ar s 4 y 85 ear s + ye ar s

0.0

Regional and Distant

N

Rel. 5-yr

N

Rel. 5-yr

N

Rel. 5-yr

3338

0.81

2224

0.91

844

0.58

2947 277 114

0.81 0.77 0.85

1972 168 84

0.91 0.90 0.91

733 86 25

0.58 0.59 0.63

187 1042 2109

0.91 0.88 0.77

127 739 1358

0.96 0.94 0.89

28 225 591

0.69 0.69 0.53

710 729 890 1009 1018

0.81 0.80 0.82 0.83 0.55

441 483 613 687 432

0.92 0.89 0.90 0.92 0.72

167 175 222 280 448

0.60 0.52 0.60 0.59 0.40

815 156 47

0.56 0.50 0.55

355 62 —

0.73 0.67 —

354 68 26

0.41 0.31 0.45

84 275 659

0.75 0.59 0.50

44 123 265

0.88 0.73 0.68

26 117 305

0.53 0.40 0.38

255 278 247 238

0.54 0.58 0.53 0.55

124 123 95 90

0.73 0.76 0.69 0.66

94 113 118 123

0.33 0.40 0.41 0.49

Incidence rates for in situ and invasive cancer combined, includes all races and histologies (SEER, 2003). Rates are reported per 100,000 women and age-adjusted to the 2000 U.S. standard. Only women aged 15 to 74 years are included. —, less than 25 cases.

1980-1986

1987-1993

1994-2000

bility that leads to neoplastic progression, the role of cofactors is an area of intense research for molecular biologists and epidemiologists.

Human Papillomavirus Infection Two population-based case-control studies of vulvar cancer reported increased risk of neoplasia associated with genital warts, which are caused most commonly by non-oncogenic HPV types 6 and 11. In a study conducted in Chicago and upstate New York, 14.8% of cases and 1.4% of controls had a prior history of genital warts (Brinton et al., 1990a). In a larger study in Seattle, there were even more striking case-control differences in genital wart prevalence: 38.7% for in situ cases, 41.7% for invasive cases, and 4.5% for controls (4.5%) (Sherman et al., 1991). Similarly, the two population based casecontrol studies of vaginal cancer (Brinton et al., 1990b; Daling et al., 2002) found that more cases reported a history of genital warts (12.2% and 25.6%) compared to controls (4.1% and 7.2%). Since nononcogenic HPV types cause most genital warts, it may be that in some women a history of genital warts is a marker of a compromised ability to clear all types of HPV infections. In a study of anogenital cancers ascertained from a population-based cancer registry, the prevalence of HPV DNA was higher for in situ vulvar tumors (91.2%) then invasive vulvar tumors (79.0%), although HPV is clearly present in the majority of SCC invasive tumors (Carter et al., 2001). In a small subsample of 34 case subjects from a prior study (Madeleine et al., 1997), HPV16 DNA was found in 75.0% of 8 basaloid or warty carcinomas and 22.7% of keratinizing SCC. In a separate report from the same group that focused on vaginal neoplasias, HPV DNA was found in the biopsies of 82.4% of 74 in situ and 64% of 25 invasive case subjects with squamous cell vaginal cancer (Daling et al., 2002). The most common HPV type was HPV16, which was detected in over 50% of each case group. These results are consistent with those of Kiyahu et al. (1989) and Ikenberg et al. (1990), who also detected HPV DNA in over 50% of invasive tumors of women with vaginal cancer. Sugase et al. (1997) found 100% of vaginal intraepithelial neoplasias were positive for at least one HPV DNA type. They identified 15 different HPV types. The finding of numerous HPV types among women with vaginal carcinomas was also found in the case-control study of Daling et al. (2002), where 15 of 43 HPV DNA positive tumors harbored HPV types other than type 16, 18, 31, 33, or 35, the types most often found in cervical and vulvar tumors. Daling et al. (2002) found that 50% of women with vaginal neoplasia were positive for either HPV 16 or 18 L1 antibodies, in contrast to less than 20% positive among the controls. In a nested case-control study within a cohort study from Finland and Denmark, Bjorge et al. (1997) found that seropositivity for HPV16 was associated with an increased risk of developing vulvar and vaginal cancers

Cancers of the Vulva and Vagina (odds ratio 4.5, 95% CI 1.1–2.2). In case-control studies conducted in the Seattle area (Madeleine et al., 1997) and upstate New York and Chicago (Hildesheim et al., 1997), the risk of vulvar cancer was also found to be significantly elevated 3- to 5-fold with antibodies to HPV16. In the Seattle study (Madeleine et al., 1997) the risk was specific to HPV16 seropositivity, as there was no difference in the prevalence of HPV6 or HPV18 between cases and controls. The study by Hildesheim (1997) found 31.2% of all invasive vulvar cases were seropositive for HPV16, compared to only 11.1% of the controls. When the cases were subdivided between those that had warty or basaloid histology and those that were keratinizing, they found that 30.0% of the basaloid or warty cases but only 15.8% of the keratinizing cases were HPV16 seropositive. Although there were only 10 basaloid or warty cases and 19 keratinizing cases, these data support a less central role of HPV in the pathway to disease for keratinizing SCC of the vulva. The consistent finding of HPV DNA in vaginal and vulvar tumor tissue, increased prevalence of HPV16 seropositivity in cases compared to controls, and increased risk associated with higher numbers of sexual partners and earlier age at first intercourse indicates that squamous cell vulvar and vaginal neoplasms are for the most part sexually transmitted diseases.

Smoking and Alcohol Cigarette smoking, especially current smoking at diagnosis, has been a consistent risk factor for all squamous cell anogenital cancers (Daling, 1992). In the case-control study conducted by Brinton and colleagues (1990a), there was an elevated risk of in situ vulvar cancer associated with current smoking (OR, 4.7, 95% CI: 2.2–10.0) that was not found for invasive disease (OR, 1.2, 95% CI: 0.6–2.2). In a casecontrol study in Seattle, the relative risk of in situ vulvar cancer associated with current smoking was 6.4 (95% CI: 4.4 –9.3) and a lower but still significant relative risk associated with current smoking for women with invasive disease (OR 3.0, 95% CI: 1.7–5.3) (Madeleine et al., 1997). In the latter study, there was also an increased risk of HPV-DNA negative vulvar neoplasms associated with current smoking (OR 5.0, 95% CI 2.6–9.8). There was no excess risk of vulvar cancer associated with ever smoking in one hospital-based study by Parazzini et al. (1993), which did not separate former from current smoking. Another hospital-based study by Mabuchi et al. (1985) did report an unadjusted increased risk for current smoking of 1.5 (95% CI: 0.9–2.7). The failure to find significant risks (both in magnitude and statistical significance) associated with former smoking indicates that smoking is likely to act as a late-stage promoter in neoplastic development. The largest case-control study of squamous cell vaginal neoplasia, 156 cases, found that women who were current smokers at diagnosis of their disease were at over two times the risk of developing neoplasia (Daling et al., 2002). In this study, over 40% of the cases were current smokers compared to 23% of 2,041 controls. The risk associated with smoking was higher for invasive disease (OR 4.6, 95% CI 1.9–10.9) compared to in situ vaginal cancer (OR 1.8, 95% CI 1.4–3.1). Brinton (1990b) also found an increase of vaginal neoplasms associated with current smoking (OR 1.4, 95% CI 0.6–3.3); however, the confidence intervals were wide, likely due to the small number of cases in the study (n = 41). The mechanism by which current smoking increases vulvar and vaginal neoplasia risk is not clear. Nicotine may block apoptosis, an important mechanism in the regulation of cell growth (Heusch, 1998; Wright et al., 1993; Wright et al., 1994). In addition, cigarette smoking has been related to systemic immunosuppression, which could increase the incidence of malignancy (Poppe et al., 1995; Poppe et al., 1996; Hughes et al., 1985). Weiderpass et al. (2001) conducted a population-based cohort study to analyze the risk of vaginal and other cancers among women with a hospital discharge diagnosis of alcoholism. They identified ten women who subsequently developed vaginal cancer, for an SIR of 4.6 (95% CI 2.2–8.5). The investigators were unable to control for history of smoking, HPV infection, or sexual behavior, so the association with

1071

alcoholism could be due to these and other potentially confounding factors.

Hormonal Factors While hormonal contraceptives and parity have been linked to an increased risk of cervical cancer in many studies (Smith et al., 2003; Moreno et al., 2002; Munoz et al., 2002), there has been no increased risk of in situ or invasive vulvar cancer associated with hormones or reproductive factors (Sherman et al., 1994; Newcomb et al., 1984; Brinton et al., 1990a). A body mass index over 30 was related to invasive vulvar disease in two studies (Parazzini et al., 1993; Sherman et al., 1994), which may indicate a role for hormones in neoplastic development. Between 1947 and 1971 DES, a synthetic estrogen, was prescribed for pregnant women to prevent spontaneous abortions and premature delivery. Experimental and clinical evidence has since indicated that prenatal exposure to this estrogenic chemical during the critical period of organ genesis can interfere with the differentiation of estrogen target organs (Robboy et al., 1981; Swan, 2000). In 1971, use of DES was linked to clear cell adenocarcinoma of the vagina (Herbst et al., 1971), a rare tumor even among women exposed to DES. This association was confirmed by others (Greenwald et al., 1971; Noller et al., 1972). A registry of all cervical and vaginal clear cell adenocarcinomas was established, and in 1987 Melnick et al. (1987) estimated that one in every 1,000 women exposed to DES will develop a clear cell adenocarcinoma by her mid-30s. Among women exposed in utero, the development of cancer was more likely to occur among those exposed before the twelfth week of gestation (Herbst et al., 1986; Hatch et al., 2001). Reports of squamous cell carcinoma of the vagina in DESexposed women have also been published (Bornstein et al., 1987; Faber et al., 1990; Hatch et al., 2001). Palmer et al. (2000) conducted a case-control study of 244 DESexposed women diagnosed with clear cell adenocarcinoma and 244 DES-exposed controls. The goal of this study, the largest such study to date, was to assess the possible influence of hormonal postnatal exposures on the risk of developing a clear cell carcinoma. Consistent with a previous case-control study (Sharp and Cole, 1991), pregnancy and oral contraceptive use did not increase the risk of DES associated adenocarcinoma; however, there was some indication that oral contraceptive use and pregnancy might advance the time of clear cell adenocarcinoma diagnosis. Palmer et al. were not able to replicate a previous finding that season of birth was related to risk of clear cell carcinoma (Herbst et al., 1986).

Hysterectomy Prior hysterectomy has been suggested in the etiology of in situ and invasive vaginal cancer in the few case-control studies and numerous case series published (Brinton et al., 1990b; Daling et al., 2002; Kalogirou et al., 1997; Chen et al., 1985; Ireland, Monaghan, 1988; Stuart et al., 1981; Bell et al., 1984; Ruiz-Moreno et al., 1987). A possible association could be attributed to one or more of the following: (1) residual cervical disease following hysterectomy for cervical disease; (2) multifocal disease due to HPV infection; (3) a result of radiation therapy for cervical disease; (4) a new primary tumor from an independent origin; and (5) atypical lesions in the 4–5% of women who have a transformation zone outside the confines of the cervix (Madeleine et al., 2001; Kalogirou et al., 1997; Chen et al., 1985; Ireland et al., 1988). Daling et al. (2002) found that 50% of women with a diagnosis of vaginal cancer had a prior hysterectomy compared to 25% of controls. After adjusting for age and number of partners, the risk was 3.3 (95% CI, 2.3–4.8). Since hysterectomy is often part of the treatment for cervical disease, it was not surprising to find that the risk of hysterectomy was 0.9 (95% CI, 0.4–2.2) among cases and controls with a history of anogenital cancer. This result is consistent with that of Herman et al. (1986), who did not find a relationship between prior hysterectomy and vaginal cancer after adjustment for age and prior disease of the cervix. These results might argue that it is HPV and not hysterectomy that leads to vaginal cancer. Arguing

1072

PART IV: CANCER BY TISSUE OF ORIGIN

against this conclusion, Brinton et al. (1990b) reported that a hysterectomy was more common among women with vaginal neoplasia than controls, even after excluding women with a history of cervical disease. Studies are needed that collect information on the reason for hysterectomy.

HOST FACTORS Immunity Impaired immunity is associated with all types of anogenital cancer, as reported in studies of different types of immunosuppressed patients. Studies of renal transplant patients, who also suffer from cellmediated immune suppression, have noted an increase in HPV-related lesions (Bouwes Bavinck, 1997; Sillman et al., 1997). Generalized Tcell deficiency has been associated with increased anogenital dysplasia (Tindle, 1994). In a clinic-based cross-sectional study, 6.1% of 396 HIV-positive women compared with 0.8% of 375 HIV negative women had in situ vulvar cancer (Chiasson et al., 1997). The overall relative risk for vulvar and vaginal in situ cancers in the AIDS and cancer national registry match study in the US was 3.9 (95% CI, 2.0–7.0) (Frisch et al., 2000). Human leukocyte antigen (HLA) genes play an important role in presenting viral antigens to the immune system. There is a consistent finding in the cervical cancer literature that protection is afforded by carriership of DRB1*13. A study by Davison et al. (2003) examined the association between Class I and II HLA genes and HPV16 positive vulvar neoplasia. They found significant associations for genotypes that have also been linked to cervical cancer (e.g., B7, DRB1*13, DRB1*11). These data suggest that an HPV vaccine may be effective against all types of HPV-associated anogenital neoplasia.

Prior Cancer History A prior history of cervical cancer has been reported as a strong risk factor for vulvar and vaginal neoplasia. Estimates of the cumulative risk of subsequent primary anogenital cancers among women with an initial primary anogenital neoplasms range from 1% (Deligdisch, 1975) to 25% or greater (Beckmann et al., 1991; Chao et al., 1993). A study by Sherman et al. (1988) found an increased prevalence of prior or concurrent anogenital cancer in 22.0% of women with in situ vulvar disease and 24.8% of women with invasive vulvar disease compared with 4.6% of biopsy controls and 1.0% of population-based controls. In the study by Brinton et al. (1990a), 14.6% of the vulvar cases had prior cervical cancer compared to 0% of the controls. In situ and invasive vaginal cancers commonly occur in women with a prior or concurrent malignancy of the anogenital tract (Benedet, 1984; Rose et al., 1987). In the Daling et al. study (2002), 30.1% of the vaginal cases reported having had a prior anogenital tumor (i.e., cervical, vulvar, or anal cancer), compared to two percent of controls (OR, 21.3, 95% CI 13.3–34.1). Fisher et al. (1997) investigated the incidence of metachronous primary cancers following initial primaries of the cervix in Michigan women between 1985 and 1992. Over the 5- to 8-year follow-up period, 6.5% of the women developed secondary primary cancers. Vaginal cancer occurred in six women, whereas only 0.1 was expected for an SIR of 44.3. Firsch et al. (1994) reviewed the Danish Cancer Registry for 1943 to 1989, searching for previous cancers among patients with anal cancer, an HPV-related cancer. An elevated risk was observed for vulva and vagina cancers (OR = 15.4, 95% CI 4.9–48.0) among women diagnosed with anal cancer. Women with anal cancer were also at increased risk for a subsequent vulvar/vaginal cancer (RR = 12.3, 95% CI 4.0–28.7). The multicentric nature of HPV infection is the likely explanation for multiple anogenital tumors occurring among women with vaginal or vulvar cancer.

Family History Several lines of evidence suggest that some women have an inherited risk of cervical cancer. A Swedish cancer registry study reported a

nearly 2-fold increased risk of cervical cancer among biological compared with adoptive family members and that half-sisters were intermediate in risk between adoptive and biologic sisters (Magnusson et al., 1999). Further support for a heritable susceptibility came from a twin study (Ahlbom et al., 1997) and a study that showed higher age-specific rates of disease among daughters of women with cervical cancer (Hemminki et al., 1999). There was also an increased risk of vaginal neoplasia associated with a family history of anogenital cancer in the study by Daling et al. (OR 2.8, 95% CI 1.0–7.8; Daling et al., 2002). They found that women with and without a history of prior anogenital cancer were equally likely to have HPV DNA positive tumor tissue.

PATHOGENESIS Host genetic factors could contribute to vulvar and vaginal carcinogenesis by a number of pathways. Some examples of pathways that may converge to contribute to genetic instability and evasion of tumor surveillance include poor metabolism of carcinogens in cigarette smoke or other environmental exposures, an inflammatory response that allows genetic damage to accrue, attenuated immune response to HPV, or defects in one of the DNA repair mechanisms. While the molecular biology and epidemiology of HPV-related neoplasms is an area that is becoming increasingly understood, the etiology of non-HPV lesions is less clear. It may be that acquired mutations are important in non-HPV pathways to disease (Crum et al., 1997).

PREVENTIVE MEASURES Primary Prevention Because of the strong association between current smoking and vulvar and vaginal cancer and the increased risk of smoking-related second primary cancers after vulvar and vaginal cancers (Sturgeon et al., 1996), smoking cessation should be considered primary prevention. HPV vaccinations are likely to be important in preventing all anogenital neoplasia. The success of the HPV16L1 virus-like protein vaccine was recently documented (Koutsky et al., 2002), and several variations of this vaccine are being developed.

Screening and Early Detection Since vaginal and vulvar neoplasia is often multifocal, the entire anogenital region of women with neoplasia lesions should be monitored closely in order to detect new or subsequent lesions at different sites at an early stage. The rate of recurrences is particularly high for women with in situ vulvar cancer, so clinical follow-up of women treated for in situ vulvar cancer usually involves gynecologic visits every 3 to 6 months for many years after initial treatment. Women with a history of anogenital neoplasia, genital warts, impaired immunity, and women who are current cigarette smokers and HPV positive should be considered to be at higher risk of multifocal or recurrent in situ vulvar cancer. Patients with vulvar condylomata should be examined periodically to detect early neoplasia. Women who are at high risk of vulvar cancer could be taught to practice self-examination.

FUTURE DIRECTIONS Vulvar and vaginal cancer precursor lesions are predominantly squamous cell neoplasias that contain HPV DNA. Understanding the role of HPV cofactors, especially how current smoking contributes to disease risk, would be an important avenue for future research. It would be interesting to know if specific genetic variants of HPV16 are more likely to cause disease or disease recurrence. Further research is also needed to understand the etiology of keratinizing vulvar cancer and whether there is a non-HPV-related pathway to invasive vulvar cancer. As the incidence of vulvar disease continues to increase at 2.4% per year (Howe et al., 2001), identifying markers to prevent

Cancers of the Vulva and Vagina recurrences may ease the psychosexual and physical comorbidities associated with treatment of vulvar neoplasia. There is a need to evaluate the risk of both squamous and adenocarcinomas of the vagina as women with in utero exposure to DES go through menopause. Mittendorf and Herbst (1992) and Bornstein et al. (1988), hypothesized that a second peak in incidence may occur at that time. Should that happen, the use of hormone replacement therapy should be evaluated as a potential modifying factor. References Ahlom A, Lichtenstein P, Malmstrom, et al. 1997. Cancer in twins: genetic nongenetic familial risk factors. J Natl Cancer Inst 89:287–293. Aho M, Vesterinen E, Meyer B, et al. 1991. Natural history of vaginal intraepithelial neoplasia. Cancer 66:195–197. Allen DG, Hutchins A-M, Hammet F, et al. 2002. Genetic aberrations detected by comparative genomic hybridization in vulvar cancers. Br J Cancer 86:924–928. American Cancer Society. 2003. Cancer Facts and Figures 2003. American Cancer Society, Inc. Atlanta, GA. Anderson WA, Franquemont DW, Williams H, Taylor PT, Crum CP. 1991. Vulvar squamous cell carcinoma and papillomaviruses: Two separate entities. Am J obstet Gynecol 165:329–336. Beckmann AM, Acker R, Christiansen AE, Sherman KJ. 1991. Human papillomavirus infection in women with multicentric squamous cell neoplasia. Am J Obstet Gynecol 165(5):1431–1437. Bell J, Sevin B-U, Averette H, Nadji M. 1984. Vaginal cancer after hysterectomy for benign disease: Value of cytologic screening. Obstet Gynecol 64:699–702. Benedet JL, Sanders BH. 1984. Carcinoma in stiu of the vagina. Am J Obstet Gynecol 158:695–700. Bjorge T, Dillner J, Anttila T, et al. 1997. Prospective seroepidemiological study of role of human papillomavirus in non-cervical anogenital cancers. Br Med J 15:646–649. Bornstein J, Kaufman RH, Adam E. 1987. Human papillomavirus associated with vaginal intraepithelial neoplasia in women exposed to diethylstilbestrol in utero. Obstet Gynecol 70:75–80. Bornstein J, Adam E, Adler-Storthz K. 1988. Development of cervical and vaginal squamous cell neoplasia as a late consequence of in utero exposure to diethylstilbestrol. Obstet Gynecol Surv 43:15–21. Bosch FX, Manos MM, Sherman ME. 1995. Prevalence of HPV DNA in cervical cancer in 22 countries. J Natl Cancer Inst 87:796–802. Bouwes Bavinck JN, Berkhout RJ. 1997. HPV infections and immunosuppression. Clinics Dermatol 15:427–437. Brinton LA, Nasca PC, Mallin K, et al. 1990a. Case-control study of cancer of the vulva. Obstet Gynecol 75:859–866. Brinton LA, Nasca PC, Mallin K, et al. 1990b. Case-control study of in situ and invasive carcinoma of the vagina. Gynecol Oncol 38:49–54. Carter JJ, Madeleine MM, Shera K, et al. 2001. Human papillomavirus 16 and 18 L1 serology compared across anogenital cancer sites. Cancer Res 61:1934–1940. Chao KH, Chang DY, Chen RJ, Lin HH, et al. 1993. Vulval neoplasia associated with other primary malignancies. F Formos Med Assoc 92:772–774. Chen N-J, Okuda H, Sekiba K. 1985. Recurrent carcinoma of the vagina following Obayayashi’s radical hysterectomy for cervical carcinoma. Gynecol oncol 20:10–16. Chiasson MA, Ellerbrock TV, Bush TJ, et al. 1997. Increased prevalence of vulvovaginal condyloma and vulvar intraepithelial neoplasia in women infected with the human immunodeficiency virus. Obstet Gynecol 89:690–694. Crum CP. 1992. Carcinoma of the vulva: Epidemiology and pathogenesis. Obstet Gynecol 79:448–454. Crum CP, McLaclin CM, Tate JE, Mutter GL. 1997. Pathobiology of vulvar squamous neoplasia. Curr Opinion in obstetrics and Gyncecol 9:63–69. Daling JR, Sherman KJ, Hislop TG, et al. 1992. Cigarette smoking and the risk of anogenital cancer. Am J Epidemiol 135:180–189. Daling JR, Madeleine MM, Schwartz SM, et al. 2002. A Population-Based Study of Squamous Cell Vaginal Cancer: HPV and Cofactors. Gynecol Oncol 84:263–270. Davision EJ, Davidson JA, Sterling JC, et al. 2003. Association between human leukocyte antigen polymorphism and human papillomavirus 16-positive vulval neoploasia in British women. Cancer Res 63:400–403. Deligdisch L, Szulman A. 1975. Multiple and multifocal carcinomas in female genital organs and breast. Gynecol Oncol 3:181–190. Faber K, Jones M, Tarraza HM. 1990. Case Report. Invasive squamous cell carcinoma of the vagina in a diethylstilbestrol-exposed woman. Gynecol Oncol 37:125–128.

1073

Fisher G, Harlow SD, Schottenfeld D. 1997. Cumulative risk of second primary cancers in women with index primary cancers of uterine cervix and incidence of lower anogenital tract cancers, Michigan. Gynecol Oncol 64:213–223. Flowers LC, Wistuba II, Scurry J, et al. 1999. Genetic changes during the multistage pathogenesis of human papillomavirus positive and negative vulvar carcinomas. J Soc Gynecol Invest 6:213–221. Frisch M, Olsen JH, Melbye M. 1994. Malignancies that occur before and after anal cancer: clues to their etiology. Am J Epidemiol 140:12–19. Frisch M, Biggar RJ, Goedert JJ. 2000. Human Paillomavirus-Asociated cancers in patients with human immunodeficiency Virus infection and acquired immunodeficiency syndrome. J Natl Cancer Inst 92:1500–1510. Gallup DG, Morley G. 1975. Carcinoma in situ of the vagina. Obstet Gynecol 46:334–340. Greenwald P, Barlow JJ, Nasca PC, Burnett WS. 1971. Vaginal cancer after maternal treatment with synthetic estrogens. N Engl J Med 258:390–392. Hatch EE, Herbst AL, Hoover RN, et al. 2001. Incidence of squamous neoplasia of the cervix and vagina in women exposed prenatally to diethylstilbestrol (United States). Cancer causes Cantrol 12:837–845. Hemminki K, Dong C, Vaittihnen P. 1999. Familial risks in cervical cancer: Is there a hereditary component? Int J Cancer 82:755–781. Herbst AL, Ulfelder H, Poskanzer DC. 1971. Adenocarcinoma of the vagina: association of maternal stilbestrol therapy with tumor appearance in young women. N Engl J Med 284:878–881. Herbst AL, Anderson S, Hubby MM, et al. 1986. Risk factors for the development of diethylstilbestrol-associated clear cell adenocarcinoma: A case-control study. Am J Obstet Gynecol 154:814–822. Herman JM, Homesley HD, Dignan MB. 1986. Is hysterectomy a risk factor for vaginal cancer? JAMA 256:601–603. Herod JJO, Shafi MI, Rollason TP, Jordan JA, Luesley DM. 1996. Vulvar intraepithelial neoplasia with superficially invasive carcinoma of the vulva. Br J Obstet Gynaecol 103:453–456. Heusch WL, Maneckjee R. 1998. Signalling pathways involved in nicotine regulation of apoptosis of human lung cancer cells. Carcinogenesis 19:551–556. Hildesheim A, Han C-L, Brinton LA, Kurman RJ, Sciller JT. 1997. Human papillomavirus type 16 and risk of preinvasive and invasive vulvar cancer: Results from a seroepidemiologcial case-control study. Obstet Gynecol 90:748–754. Holowaty P, Miller AB, Rohan T, TO T. 1999. Natural History of Dysplasia of the Uterine Cervix. J Natl Cancer Inst 91:252–258. Howe HL, Wingo PA, Thun MJ, el al. 2001. Annual report to the nation on the status of cancer (1973 through 1998), featuring cancers with recent increasing trends. J Natl Cancer Inst 93:824–842. Hughes DA, Haslam PL, Townsend PJ. 1985. Numerical and functional alterations in circulatory lymphocytes in cigarette smokers. Clin Exp Immunol 61:459–466. IARCa. 2002. Cancer Incidence in Five Continents, Vol. VIII. In: Parkin DM, Whelan SL, Ferlay J, Teppo L, Thomas D, eds. Lyon, France, IARC Scientific Publications No. 155. IARCb. 2002. Cancer in Africa: Epidemiology and Prevention. In: Parkin DM, Ferlay J, Hamdi-Cherif M, Sitas F, Thomas J, Wabinga H, Whelan S, eds. Lyon, France, IARC Scientific Publications No. 153 International Agency for Research on Cancer. Ikenberg H, Runge M, Goppinger A, Pfleiderer A. 1990. Human papillomavirus DNA in invasive carcinoma of the vagina. Obstet Gynecol 76:432– 438. Ireland D, Monaghan JM. 1988. The management of the patient with abnormal vaginal cytology following hysterectomy. Br J Obstet Gynecol 95:973–975. Kalogirou D, Antoniou G, Karakitsos P, et al. 1997. Vaginal intraepithelial neoplasia (VAIN) following hysterectomy in patients treated for carcinoma in situ of the cervix. Eur J Gynecol Oncol 3:188–191. Kiyahu MT, Shibata D, Arnheim N. 1989. Detection of human papillomavirus in formalin-fixed, invasive squamous carcinomas using the polymerase chain reaction. Am J Surg Pathol 13:211–224. Koutsky LA, Ault KA, Wheeler CM, et al. 2002. A Controlled Trial of a Human Papillomavirus Type 16 Vaccine, the Proof of Principle Study Investigators. N Engl J Med 347:1645–1651. Kurman RJ, Trimble CL, Shah KV. 1992. Human papillomavirus and the pathogenesis of vulvar carcinoma. Curr Opinion Obstet Gynecol 4:582–585. Lenehan PM, Meffe F, Lickrish GM. 1986. Vaginal intraepithelial neoplasia: biologic aspects and management. Obstet Gynecol 68:333–337. Levi F, Randimbison L, La Vecchia C. 1998. Descriptive epidemiology of vulvar and vaginal cancers in Vaud, Switzerland, 1974–1994. Annals Oncol 9:1229–1232. Mabuchi K, Bross DS, Kessler II. 1985. Epidemiology of cancer of the vulva: A case-control study. Cancer 55:1843–1848.

1074

PART IV: CANCER BY TISSUE OF ORIGIN

Magnusson PKE, Sparen P, Gyllensten U. 1999. Genetic link to cervical tumours. Nature 400:29–30. Madeleine MM, Daling JR, Carter JJ, et al. 1997. Cofactors with human papillomavirus in a population-based study of vulvar cancer. J Natl Cancer Inst 89:1516–1523. McNally OM, Mulvany NJ, Pagano R, Quinn MA, Rome RM. 2002. VIN3: A clinicopathologic review. Int J Gynecol Cancer 12:490–495. Melnick S, Cole P, Anderson D, Herbst A. 1987. Rates and risks of diethylstilbestrol-related clear-cell adenocarcinoma of the vagina and cervix. An update. N Engl J Med 316:514–516. Minucci D, Cinel A, Insacco E, et al. 1995. Epidemiological aspects of vaginal intraepithelial neoplasia (VAIN). Clin Exp Obst Gyn 1:36–42. Mittendorf R, Herbst AL. 1992. Diethylstilbestrol: An update. Cancer Prev 1–12. Modesitt SC, Waters AB, Walton L, Fowler WC, Le LV. 1998. Vulvar intraepithelial neoplasia III: Occult cancer and the impact of margin status on recurrence. Obstet Gynecol 92:962–966. Moreno V, Bosch FX, Munoz N, et al. 2002. Effect of oral contraceptives on risk of cervical cancer in women with human papillomavirus infection: the IARC multicentric case-control study. Lancet 359:1085–1092. Muñoz N, Franceschi S, Bosetti C, et al. 2002. Role of parity and human papillomavirus in cervical cancer: the IARC multicentric case-control study. Lancet 359:1093–1101. Newcomb P, Weiss N, Daling JR. 1984. Incidence of vulvar carcinoma in relation to menstrual, reproductive, and medical factors. J Natl Cancer Inst 73:391–396. Noller KL, Decker DG, Lanier AP, Kurland LT. 1972. Clear cell adenocarcinoma of the cervix after maternal treatment with synthetic estrogens. Mayo Clinic Proc 47:620–630. Palmer JR, Anderson D, Helmrich SP, Herbst AL. 2000. Risk factors for diethylstilbestrol-associated clear cell adenocarcinoma. Obstet Gynecol 95:814–820. Parazzini F, La Vecchia C, Garsia S, et al. 1993. Determinants of invasive vulvar cancer risk: an Italian case-control study. Gynecologic Oncology 48:50–55. Poppe WAJ, Ide PS, Drijkoningen M, Lauweryns JM, Van Assche FA. 1995. Tobacco smoking impairs the local immunosurveillance in the uterine cervix. Gynecol Obstet Invest 39:34–38. Poppe WAJ, Peeters R, Drijkoningen M, et al. 1996. Cervical cotinine and macrophage-Langerhans cell density in the normal human uterine cervix. J Gynecol Obstet 41:253–259. Robboy SJ, Szyfelbein WM, Goellner JR, et al. 1981. Dysplasia and cytologic findings in 4,589 young women enrolled in Diethylstilbestrol-Adenosis (DESAD) Project. Am J Obstet Gynecol 140:579–586. Rose PG, Herterick EE, Boutselis JG, Moeshberger M, Sachs L. 1987. Multiple primary gynecologic neoplasms. Am J Obstet Gynecol 157:261–267. Ruiz-Moreno JA, Garcia-Gomez R, Vargas-Solano A, Alonso P. 1987. Vaginal intraepithelial neoplasia. Report of 14 cases. J Gynecol Obstet 25:359–362. SEER (Surveillance, Epidemiology, and End Results Program). SEER*Stat Database: Incidence–-SEER 9 Regs Public-Use, Nov 2002 Sub

(1973–2000), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2003, based on the November 2002 submission. Sharp GB, Cole P. 1991. Identification of risk factors for diethylstilbestrolassociated clear cell adenocarcinoma of the vagina: Similarities to endometrial cancer. Am J Epidemiol 134:1316–1324. Sherman KJ, Daling JR, Chu J, McKnight B, Weiss NS. 1988. Multiple primary tumours in women with vulvar neoplasms: a case-control study. Br J Cancer 57:423–427. Sherman KJ, Daling JR, Chu J, et al. 1991. Genital warts, other sexually transmitted diseases, and vulvar cancer. Epidemiology 2:257–262. Sherman KJ, Daling JR, McKnight B, Chu J. 1994. Hormonal factors in vulvar cancer: A case-control study. J Reprod Med 39:857–961. Sillman FH, Sentovich S, Shaffer D. 1997. Anogenital neoplasia in renal tansplant patients. Annals Transplantations 2:59–66. Smith JS, Green J, de Gonzalez A, et al. 2003. Cervical cancer and use of hormonal contraceptives: a systematic review. Lancet 361:1159–1167. Stuart GCE, Allen HH, Anderson RJ. 1981. Squamous cell carcinoma of the vagina following hysterectomy. Am J Obstet Gynecol 139:311–315. Sturgeon SR, Brinton LA, Devesa SS, Kurman RJ. 1992. In situ and invasive vulvar cancer incidence trends (1973–1987). Am J Obstet Gynecol 166:1482–1485. Sturgeon SR, Curtis RE, Johnson K, Ries M, Brinton LA. 1996. Second primary cancers after vulvar and vaginal cancers. Am J Obstet Gynecol 174:929–933. Sugase M, Matsukara T. 1997. Distinct manifestations of human papillomaviruses in the vagina. Int J Cancer 72:412–415. Swan SH. 2000. Intrauterine exposure to diethylstilbestrol: Long-term effects in humans. APMIS 108:793–804. Tindle RW, Frazer IH. 1994. Immune response to human papillomaviruses and the prospects for human papillomavirus-specific immunization. In zurHausen H, ed. Human pathogenic papillomaviruses. Heidelberg, Springer Verlag, pp. 217–253. Walboomers JMM, Jacobs MV, Manos MM, et al. 1999. Human papillomavirus is a necessary cause of invasive cervical cancer worldwide. J Pathol 189:12–19. Weiderpass E, Ye W, Tamimi R, et al. 2001. Alcoholism and risk for cancer of the cervix uteri, vagina, and vulva. Cancer Epidemiol, Biomarkers Prev 10:899–901. Wright VC, Chapman W. 1992. Intraepithelial neoplasia of the lower female genital tract: Etiology, investigation, and management. Semin Surg Oncol 8:180–190. Wright SC, Zhong J, Zheng H, Larrick JW. 1993. Nicotine inhibition of apotosis suggests a role in tumor progression. FASEB J 7:1045–1051. Wright SC, Zhang J, Larrick JW. 1994. Inhibition of apoptosis as mechanism of tumor promotion. FASEB J 8:654–660. Yang B, Hart WR. 2000. Vulvar intraepithelial neoplasia of the simplex (differentiated) type: A clinicopathologic study including analysis of HPV and p53 expression. 24:429–441. zur Hausen H. 2002. Papillomaviruses and cancer: from basic studies to clinical application. Cancer 2:342–350.

56

Choriocarcinoma JULIE R. PALMER AND COLLEEN M. FELTMATE

G

estational choriocarcinoma is an extremely rare cancer that occurs in women of childbearing age. The malignancy arises from the trophoblastic epithelium of the placenta, which is formed from embryonic tissue. Thus, a choriocarcinoma is a naturally occurring malignant allograft that originates in the tissues of the conceptus and invades the tissues of the mother. Occasionally, nongestational choriocarcinomas occur, arising from the ovary or testis. This chapter will focus on gestational choriocarcinoma, and the term choriocarcinoma will refer here to the gestational form of the disease. Choriocarcinoma is one of several related gestational trophoblastic diseases, which include complete hydatidiform mole, partial hydatidiform mole, invasive mole, and the extremely rare placental site trophoblastic tumor. Choriocarcinomas are typically aggressive in nature. The malignant trophoblastic cells may be located within the endometrium, be locally invasive into the myometrium, or be metastasized to any site in the body. The most frequent sites of metastases are the lung and pelvis but there can also be metastasis to the brain and liver. Until the 1950s, mortality from choriocarcinoma was very high. Concerted efforts to improve treatments for the disease led to identification of methotrexate as an effective chemotherapy regimen. Choriocarcinoma is now considered to be the first solid human malignancy to be cured by chemotherapy. Most women can be treated without having to undergo surgery, thus preserving their fertility. While efforts to treat choriocarcinoma have met with great success, less progress has been made in establishing the descriptive epidemiology of choriocarcinoma and in identifying risk and preventive factors for the disease. The very low incidence of the disease and difficulties in defining the appropriate denominator have hampered these efforts. Because gestational choriocarcinoma always follows a pregnancy, the appropriate denominator for the incidence rate is all pregnancies that occurred during the given time period. It can be difficult to obtain population-based data on both total number of pregnancies and total number of cases for a specific region. Thus, incidence figures have sometimes been based on a denominator of total female population, resulting in underestimation, or a denominator of number of live births, resulting in overestimation. It appears that the incidence is higher in Asia than in the Western Hemisphere but estimates of the relative incidence have been highly variable. Another problem in comparing incidence rates across nations or even across time periods is inconsistencies in definition of the numerator. Some registries have recorded only choriocarcinomas for which histologic evidence is present, some all choriocarcinomas regardless of histologic evidence, and some have included all gestational trophoblastic neoplasia. Increasingly, the diagnosis of choriocarcinoma is made based on clinical information. Because surgery is often avoided, histologic specimens are often not available. Without surgical pathology, it can be difficult to distinguish between a true choriocarcinoma and other persistent trophoblastic disease, since cytotoxic treatment is initially utilized for both. Because the disease is so rare, inclusion or exclusion of a few cases based on different classification criteria can markedly change reported incidence rates. It is useful to examine incidence patterns and risk factors for hydatidiform mole, the most common of the gestational trophoblastic diseases, both because there appear to be similarities in the distribution and determinants of hydatidiform mole and choriocarcinoma and because hydatidiform mole is by far the strongest risk factor for

choriocarcinoma. As such, factors that influence the occurrence of hydatidiform mole will necessarily have an effect on the risk of choriocarcinoma.

CLASSIFICATION Histopathology and Diagnosis Two morphologically distinct trophoblastic cells are described in chorionic villi: cytotrophoblasts and syncytiotrophoblasts. The common pathology of trophoblastic lesions is an excessive proliferation of these cells. Differentiating the various forms of gestational trophoblastic disease can be done using pathologic and cytogenetic techniques. Hydatidiform moles are characterized as either partial or complete. Partial moles can be characterized by the presence of embryonic or fetal tissues-vessels containing nucleated fetal red cells are a common finding. Partial moles may not display more than focal areas of trophoblastic hyperplasia, consisting mainly of syncytiotrophoblasts, with limited hydropic villi. Complete moles, on the other hand, are marked by the pathologic absence of fetal tissues, with trophoblastic hyperplasia and extensive villous edema. Early complete moles may not exhibit these features and can be easily confused with partial moles. In these cases, nuclear debris in villous stroma and irregularly shaped villi with secondary budding are common characteristics of young complete moles. In both complete and partial moles, the karyotypes are characterized by an excess of paternal genes. Complete moles are characterized by a diploid karyotype where approximately 90% have a 46, XX karyotype (Wake et al., 1978). Cytogenetic evaluation of complete mole via chromosomal banding heteromorphisms (Kajii, 1977) and hypervariable DNA sequences (Lawler et al., 1991) has shown that all nuclear DNA is paternally derived and cytoplasmic DNA is maternally derived. Seventy-five percent or more of complete hydatidiform moles result from fertilization of an “empty” ovum by a haploid sperm (23,X) (Fisher et al., 1989; Wake et al., 1987), which then duplicates its chromosomal material. Dispermic fertilization accounts for approximately 10–15% of complete moles, with approximately 6–10% having a 46, XY genotype (Surti et al., 1979) and another 5% being heterozygous 46, XX cases (Fisher et al., 1989). Partial moles have a triploid karyotype in 90% of cases (Lage et al., 1992). Although nontriploid partial moles have been described (Koenig et al., 1993; Jeffers et al., 1995; Fukunaga, 2000), a recent review (Genest et al., 2002) concluded that they probably do not exist but are the result of histologic and flow cytometric errors. Invasive mole is a histologically benign condition resulting from invasion of abnormal trophoblasts into myometrial tissue. Histologically, the appearance of the chorionic villi is similar to that of complete mole, but swelling is less often seen in these deeply invasive lesions. Complete mole is largely the precursor to invasive molar disease, and before the advent of chemotherapy, myometrial invasion was reported in about 16% of cases of complete mole (Hertig, 1956). Partial moles have rarely been reported as the antecedent pregnancy (Szuzman et al., 1981; Gaber et al., 1986). Invasive moles are rarely diagnosed any more since hysterectomy is seldom necessary for the treatment of the disease. Persistent gestational trophoblastic disease occurs following 14 to 29% of well-documented, histologically

1075

1076

PART IV: CANCER BY TISSUE OF ORIGIN

reviewed complete moles (Paradinas, 1998; Kohorn et al., 1982), and following less than 5% of partial moles (Paradinas, 1997). Neoplastic transformation of a mole is recognized when villous stroma are no longer formed, but little is known of the factors that control this behavior. Approximately 3% of complete moles (Hertig, 1956) and 0.1% of partial moles (Seckl et al., 2000) have been reported to develop into choriocarcinoma. Choriocarcinoma is a malignant transformation of cyto- and syncytotrophoblastic cells. Histologically, viable tumor is seen only at the periphery where the characteristic bilaminar structure of anaplastic villous trophoblasts is identified. In some cases one type of trophoblastic cell will predominate. Placental site trophoblastic tumor is a rare form of malignant trophoblastic disease and the tumorous counterpart of the non-villous trophoblast. Placental site trophoblastic tumors form clusters or cords of mononuclear or multinucleated cells that infiltrate between smooth muscle fibers of the uterus.

Genetic Predictors for Persistence or Malignant Transformation Normal fetal development requires both maternal and paternal genetic contributions. While it is clear that an excess of paternal genes is present in molar pregnancies and the excess is particularly pronounced in complete moles, it is unclear if and how genomic imprinting plays a role in the molar phenotype. A change in the biparental balance may result in an overexpression of certain genes or a lack of presence of other genes normally maternally expressed. (Reik, 1989; Henry et al., 1991; Little et al., 1991). IGF2 and H19 are two classic examples of imprinted genes that may play a role in molar pathogenesis. Although some clinical parameters indicate potential risk for development of persistence, neither morphology nor clinical features can predict which moles will become invasive and which will undergo malignant transformation. Some studies have assessed monospermy versus dispermy as a risk factor. Wake et al. (1987) reported a persistence rate of 50% in heterozygous complete mole versus 4% in homozygous genotypes. This has been refuted by other studies that have not found dispermic complete mole to be at increased risk (Lawler et al., 1991; Habibian, 1987). In addition, the XY (with Ychromosome) has not been shown to be associated with increased risk for metastatic spread (Mutter et al., 1993). The chance of developing choriocarcinoma from a molar pregnancy is 1000-fold higher than from a normal pregnancy. Thus, hydaditiform mole has been considered a pre-malignant condition. Tumor progression is often associated with loss or inactivation of tumor suppressor genes and /or activation of proto-oncogenes. Deregulation of either set of oncogenes may lead to uncontrolled cell proliferation, terminal cell cycle arrest or enhanced activation or apoptosis. Studies have explored the role of various tumor suppressor genes in the pathogenesis of gestational trophoblastic disease. Increased p53 expression in choriocarcinoma has been reported by several investigators (Cheung et al., 1999; Lee, 1995; Cheville et al., 1996; Fulop et al., 1998a; Halperlin et al., 2000). Similarly p21WAF1/CIP1 and retinoblastoma (RB) gene expression are significantly higher in complete mole and choriocarcinoma compared to normal placenta and partial mole. It is unclear why these tumor suppressor genes are upregulated or overexpressed when in other cancers loss of expression signals loss of cell cycle control and over-proliferation. Whether these are just reflective of the higher proliferative capacity of gestational trophoblastic tumors or the products are somehow inactivated following transcription remains to be clarified. Studies of MDM-2 and P53 reporting overexpression of both protein products indicate that MDM2 may inhibit the actions of TP53 and RB tumor suppressor gene products and thus promote tumorigenesis (Fulop et al., 1998b; Xiao et al., 1995). DOC-2/hDab2 has been shown to be downregulated in gestational trophoblastic tumor and choriocarcinoma compared with normal trophoblastic cells in culture. Transfection studies into in vitro choriocarcinoma cell lines display significantly decreased growth, suggesting a potential role in growth and differentiation (Fulop et al., 1998c).

Overexpression of cERBB-2 oncogene has also been associated with poor prognosis in various carcinomas (Wright et al., 1992; Pavlidis et al., 1995). The cerbB-2 protein is overexpressed in extravillous trophoblast of complete mole and choriocarcinoma (Fulop et al., 1998). Although expression may be related to the aggressive nature of disease states, its prognostic significance has not been studied. BCL-2 gene may play a role in prevention of apoptosis. It is downregulated in hydatidiform mole and choriocarcinoma (Wong et al., 1999) and may indicate an increased susceptibility to apoptosis (Sakuragi et al., 1994).

DEMOGRAPHIC PATTERNS Incidence—United States The first reliable information on choriocarcinoma incidence in the United States comes from a 1986 study in which data from the SEER (Surveillance, Epidemiology, and End Results) Program were combined with state-specific birth rates and induced abortion rates to allow computation of incidence rates for the period 1973–1982 (Brinton et al., 1986). Overall, the incidence was one per 24,096 pregnancies or 4.15 per 100,000 pregnancies. When induced abortions were excluded from the denominator, an incidence of 5.02 per 100,000 livebirths was calculated. These different estimates illustrate the importance of the denominator in assessment of choriocarcinoma incidence. Reports that rely on number of deliveries or livebirths as the denominator will produce inflated incidence estimates with varying amounts of inflation depending on the prevalence of induced abortion. The SEER data estimate of 4.15 per 100,000 pregnancies is itself an overestimate of the true incidence because spontaneous abortions were not included in the denominator. The authors of this report suggested that some true choriocarcinomas may not have been included in the SEER data and this would have resulted in underestimation of incidence. Indeed, 89.7% of cases reported to the registries during this time period were diagnosed by positive histology and fewer than 5% were diagnosed on clinical grounds alone. Because, increasingly, choriocarcinomas that follow hydatidiform moles are treated early and without biopsy for histologic confirmation, it is likely that some cases were not included. It is possible that, over time, a smaller proportion of true cases have been definitively identified as such and accessioned by the registries. The authors reported no significant trend over time during the period from 1973 through 1982, with limited power to detect a trend. The incidence was approximately two times as great as that estimated in earlier time periods in two other states, but these earlier estimates were based on very small numbers—23 cases from Connecticut (Shanmugaratnam et al., 1971) and 4 cases from Rhode Island (Yen, 1968). An updated analysis of United States SEER data has recently been published (Smith et al., 2003). SEER data on gestational choriocarcinoma for the period 1981 through 1991, were combined with national data on live births and other pregnancies. It appears that there has been a decline in the incidence of choriocarcinoma over time in the United States, perhaps by as much as 25%. Overall, choriocarcinoma occurred in 2.4 of 100,000 pregnancies and 3.84 of 100,000 livebirths. The incidence rates are not directly comparable to those reported in the earlier study (Brinton et al., 1986), however. First, a different data source for enumerating livebirths and pregnancies was used. Second, spontaneous abortions were included in the denominator for the more recent analysis but not for the previous one. And third, changes in treatment over the past 30 years have changed how choriocarcinoma is diagnosed. Fewer women with persistent trophoblastic disease undergo surgical biopsy and more often the diagnosis is made on clinical grounds, with many cases simply being classified as “persistent gestational trophoblastic disease”. This may have resulted in a reduced enumeration of cases. There has been somewhat more information on the incidence of hydatiform mole, with the best estimate coming from an analysis of hospital discharge data from the National Center for Health

1077

Choriocarcinoma Statistics (Hayashi et al., 1982). Over a period from 1970 to 1977, the incidence of hydatidiform mole was estimated to be 1.08 per 1,000 pregnancies. Induced abortions that were performed in freestanding clinics and early spontaneous abortions were not included in the denominator and thus the rate is probably overestimated. There was not a significant increase or decrease in incidence of hydatidiform mole over time. In a different approach, Atrash et al. (1986) used data from a collaborative project on induced abortions to estimate the proportion of all induced abortions in which the conceptus was a hydatidiform mole. During the period of the study, 1974–79, moles were detected in 0.75 per 1,000 pregnancies. There was evidence, however, that an appreciable number of moles may not have been reported by the clinician or detected and reported by the pathologist. Smith et al. (2004) examined all gestational trophoblastic disease in New Mexico for the period 1973 to 2001. Overall the incidence was 1.20 per 1,000 pregnancies. However, among American Indians, where a total of 234 cases occurred, the incidence was 2.06 per 1,000 pregnancies. In summary, the most informative analyses of U.S. data suggest that the incidence of choriocarcinoma is about 2.4 cases per 100,000 pregnancies and that the incidence of hydatidiform mole is about 1 per 1,000 pregnancies.

International Patterns Choriocarcinoma Comparison of choriocarcinoma incidence across geographic regions is complicated due to inconsistencies in identification of the base population and ascertainment of cases. The best data are from populationbased studies that use population data on number of deliveries or pregnancies for estimation of the denominator. Selected studies of this type are listed in Table 56–1. However, even with an appropriate denominator, there may be variation in rates due to variable ascertainment of this rare disease. The criteria for differentiating choriocarcinoma from other persistent diseases have changed over time and vary between regions. As shown in Table 56–1, there appears to be variability in the incidence of choriocarcinoma across countries. The highest incidence has been observed in Singapore (Teoh et al., 1972), Malaysia (LlewellynJones, 1965), and the Philippines (Baltazar, 1976), with an estimated 20.4, 12.5, and 17.4 cases per 100,000 deliveries or livebirths in these three countries, respectively. In other parts of Asia, the incidence has also been high. In Japan, it was estimated to be 12.1 per 100,000 pregnancies in one study (Nakano et al., 1980) and from 5.8 to 8.8 per 100,000 in another (Hando et al., 1998). A recent study from Korea yielded an incidence rate of 7.0 per 100,000 deliveries for the period 1991–95 (Kim et al., 1998). Incidence rates from Western Europe, the United States, Israel, and Paraguay have been somewhat lower, in the range of 2.3 to 5.1 cases per 100,000 pregnancies or deliveries (Matalon et al., 1972; Rolon and Hochsztajn, 1979; Olsen et al., 1999; Smith et al., 2003; Loukovaara et al., 2003). There are also regional differences in time trends, although it is difficult to know how much the apparent trends are affected by changes in enumeration of choriocarcinoma cases due to changes in medical practice. In Japan and Korea, there is now clear evidence of a decrease in choriocarcinoma incidence over time, with a reduction of 50–100% from the 1970s to the 1990s (Kim et al., 1998; Hando et al., 1998). In the United States, there appears to have been about a 25% decrease in the same time period (Smith et al., 2003). In Denmark, there is no evidence of a reduction in incidence for similar time periods (Olsen et al., 1999). Unfortunately, new incidence data from Singapore, Malaysia, and the Philippines, the countries that demonstrated the highest incidence in the 1960s and early 1970s are not available.

Mortality and Survival—United States Because choriocarcinoma is an extremely rare disease, it accounts for a negligible proportion of mortality in U.S. women. Before the advent of chemotherapeutic treatments in 1956, choriocarcinoma was usually fatal (Park, 1950). The development of chemotherapy led to a rapid reversal of this pattern (Hertz et al., 1959). Choriocarcinoma cells have proven to be very sensitive to various cytotoxic agents. Further, serum levels of beta-hCG are an accurate and extremely sensitive marker for the presence of residual trophoblastic tissue, indicating when there is a need for continued treatment. Most patients are now treated effectively with single agent chemotherapy, usually methotrexate or dactinomycin, and combination therapy is used for patients with high-risk metastatic or resistant disease. A review of patients in China shows that in the period from 1976 to 1985, only 15.4% of women with choriocarcinoma died from the disease, as compared with 90% in the period from 1948 to 1956 (Song et al., 1998). Similarly, in the United States, a remission rate of over 90% has been achieved (Hammond et al., 1967; Berkowitz et al., 1986). In most cases, patients achieve not just a complete remission but a cure.

Hydatidiform Mole Early reports indicated considerable international variation in the incidence of hydatidiform mole. On closer examination, it was clear that some of the variation was due to the use of hospital-based studies to

Table 56–1. Estimates of Choriocarcinoma Incidence from Selected Population-Based Studies Incidence Geographic Region

Time period

No. of Cases

Per 100,000 Pregnancies

Per 100,000 Deliveries

Reference

United States United States Jamaica Paraguay Finland Denmark Sweden Norway Poland Greenland Israel Korea Korea Japan Japan Japan Singapore Malaysia Philippines

1973–1982 1981–1991 1958–1973 1960–1967 1953–1999 1977–1994 1958–1965 1957–1966 1987–1996 1950–1974 1950–1965 1981–1985 1991–1995 1972–1977 1974 1990–1993 1960–1970 1958–1963 1970–1974

80 177 52 21 142 43 40 29

4.1 2.4

5.0 3.8 13.5

Brinton et al., 1986 Smith et al., 2003 Sengupta et al., 1977 Rolon and Hochsztajn, 1979 Loukovaara et al., 2004 Olsen et al., 1999 Ringertz et al., 1970 Shanmugeratham et al., 1971 Nowak-Markwitz et al., 2000 Nielson and Hansen, 1979 Matalon et al., 1972 Kim et al., 1998 Kim et al., 1998 Nakano et al., 1980 Hando et al., 1998 Hando et al., 1998 Teoh et al., 1971 Llewellyn-Jones, 1965 Baltazar, 1976

11 36 28 25 18 122 18 91

2.3 4.0 3.0 3.9 4.6 3.8 35.0 5.1 16.0 7.0 12.1 8.8 5.8 20.4 12.5 15.6

1078

PART IV: CANCER BY TISSUE OF ORIGIN

Table 56–2. Incidence of Hydatidiform Mole from Selected Population-Based Studies Published Since 1980 Incidence Geographic Region North America United States United States Hawaii Canada South America Paraguay Europe Finland England and Wales Italy Northern Ireland England and Wales Sweden Denmark Netherlands Poland Asia China China Japan Japan Korea Vietnam United Arab Emirates Oceania Australia New Zealand Western Samoa

Time period

No. of cases

Per 1000 pregnancies

Per 1000 deliveries

Reference

1970–1977 1975–1978 1968–1981 1975–1978

185 63 180 171

1.1 0.8 1.2 0.8

Hayashi et al., 1982 Atrash et al., 1986 Matsuura et al., 1984 Yuen and Cannon, 1981

1970–1982

379

0.3

Rolon et al., 1990

1975–2001 1983 1979–1982 1987–1995 1991–1995 1975–1988 1977–1992 1978–1980 1987–1996

1,659 970 347 84 3,637 393 1,520 308

1979–1983 1991–2000 1986–1995 1991–1995 1967–1981 1978–1986

8,832 856 417 860

1973–1982 1980–1986 1980–1987

455 299 32

1.00 1.5

0.9 1.1 0.4

There are no data available on the relative incidence of choriocarcinoma in migrant groups. Limited information on hydatidiform mole is provided by studies of polyracial communities in Hawaii. In one, Japanese and Filipino women who were born in Hawaii had the same incidence as Japanese and Filipino Hawaiians who were born in Japan or Philippines, although the incidence was considerably higher than in white or Hawaiian Hawaiians (Matsuura et al., 1984). In an earlier study, Japanese and Chinese Hawaiians were found to have a higher incidence than white and Hawaiian Hawaiians (McCorriston, 1968). These limited data point to the importance of genetic factors.

2.2 1.04 1.5

0.9 0.8

2.1

Migration

1.5 2.3 2.1 3.3 2.0

Shang et al., 1982 Song and Wu, 1987 Matsui et al., 2003 Sakumoto et al., 1999 Kim et al., 1998 Constable et al., 1985 Graham and Fajardo, 1988

0.7

127

estimate incidence. In developing countries, where many women give birth at home, the incidence was grossly overestimated. Populationbased studies indicate a much smaller range of incidences (Table 56–2), with about two times the incidence in the higher incidence countries as compared with the low-incidence countries. For example, recent estimates of hydatidiform mole incidence are 1.5 to 2.3 moles per 1,000 pregnancies in Japan (Sakumoto et al., 1999, Matsuii et al., 2003) and 2 moles per 1,000 deliveries in Korea (Kim et al., 1998), whereas comparable rates for the United States, Canada and Denmark are 0.8 to 1.1 per 1,000 pregnancies (Hayashi et al., 1982, Yuen and Cannon, 1981; Olsen et al., 1999). There is not a consistent pattern of variation by geographic region, however, as the incidence in China is estimated to be 0.8 moles per 1,000 pregnancies (Song, 1987) and that in Great Britain to be 1.5 per 1,000 (Bagshawe et al., 1986). As can be seen in Table 56–2, most estimates are in the ranges of 0.7 to 2.3 per thousand. Outliers are higher rates found in Micronesia (4.27 per 1,000 pregnancies) (Brewis and Peddie, 1996), Keral, India (12 per 1,000 deliveries) (Molykutty et al., 1993), and Karachi, Pakistan (3.89 per 1,000 pregnancies) (Talati, 1998), but these were obtained from hospital-based studies. As with choriocarcinoma, there has been a decreased incidence over time several Asian countries, including Japan (Hando et al., 1998), Korea (Kim et al., 1998), and Taiwan (Hsu et al., 1998).

0.6 0.8

Loukovaara et al., 2005 Bagshawe et al., 1986 Mazzanti et al., 1986 Giwa-Osagie, 1999 Tham et al., 2003 Flam et al., 1992 Olsen et al., 1999 Frank et al., 1983 Nowak-Markwitz et al., 2000

0.6 0.7 0.9

0.7

Olesnickey et al., 1985 Duff, 1989 Paksoy et al., 1989

Race and Ethnicity Differences in incidence by race or ethnicity have been observed but not consistently (Table 56–3). In United States data from 1973–1982 the incidence of choriocarcinoma was found to be lowest in Caucasian women (Brinton et al., 1986), but in more recent data from 1981–1991, there was no difference (Smith et al., 2003). In the Northwest territories of Canada, choriocarcinoma incidence was highest among the Inuit Indians (Gaudette et al., 1993). Worldwide, the highest incidence of choriocarcinoma reported is from Greenland, a country in which most of the population is Eskimo/Caucasian with the Caucasian admixture estimated to be 25–30% (Neilsen and Hansen, 1979). In Singapore, choriocarcinoma incidence was highest among Malays (Teoh et al., 1972). For hydatidiform mole, comparisons by race in United States data have yielded inconsistent results, with black women estimated to have an incidence higher (Yen and MacMahon, 1968), lower (Hayashi et al., 1982), or similar (Atrash et al., 1986) to that of white women. In Hawaii, incidence was lowest among Caucasian women and native Hawaiians, and highest in the Filipino and Japanese populations (Matsuura et al., 1984). Incidence in the United States appears to be highest among American Indians of the Southwest (Smith et al., 2004) and among native Alaskans (Martin, 1978). In China, the incidence was lower among women of Han ethnicity than among the Zhuang or Mongolian ethnic groups (Song, 1987). In the United Arab Emirates (Graham, 1988), the incidence was highest among women born in the Persian Gulf region. In a small study from Trinidad, hydatidiform mole was found to be more common among East Indian women than among women of African descent (Bassaw, 1990). In Nigeria, incidence differed by ethnicity: the Yorubas, with the highest incidence, differed from the other groups by the presence of hemoglobin C (Egwuatu, 1989). In northern England and Wales the incidence of hydatidiform mole was 1.9 times as high for women from South or East Asia as that from non-Asian women (Tham et al., 2003). In Singapore, however, while choriocarcinoma incidence differed by ethnicity, hydatidiform incidence did not (Teoh et al., 1971).

1079

Choriocarcinoma Table 56–3. Race/ethnicity in Relation to Incidence of Choriocarcinoma and Hydatidiform Mole Racial/Ethnic Group with Increased Incidence Choriocarcinoma United States Singapore Canada Greenland Hydatidiform mole United States (Overall) United States (Overall) United States (Overall) United States (Alaska) United States (New Mexico) United States (Hawaii) United States (Hawaii) China Trinidad Nigeria Singapore United Arab Emirates England and Wales

Reference

Nonwhite Malays Inuit Indians Eskimo-Caucasian population

Brinton et al., 1986 Teoh et al., 1972 Gaudette et al., 1993 Nielsen and Hansen, 1979

Asian Black, Jewish White, “other” race Native Alaskan American Indian Filipino, Japanese Japanese Zhuang, Mongolian East Indian

Atrash et al., 1984 Yen and MacMahon, 1968 Hayashi et al., 1982 Martin, 1978 Smith et al., 2004 Maatsura et al., 1984 McCorriston, 1968 Song and Wu, 1987 Bassaw and Roopnarinesingh, 1990 Egwuatu and Ozumba, 1989 Teoh et al., 1971 Graham et al., 1990 Tham et al., 2003

Yoruba Eurasian, Indian Gulf Arabs (mainly Omani, Yemeni) South Asian and East Asian

Although environmental factors could possibly explain differences by race/ethnicity because of shared cultural experiences, the above findings underline the importance of genetic factors. An interesting and relevant comparison is between Japanese living on the island of Okinawa and mainland Japanese (Sakumoto, 1998). The residents of the island are ethnically Japanese but have a very different culture from that in most of Japan. Nevertheless, the estimated incidence among the Okinawan Japanese was almost identical to the incidence observed in all Japan.

Socioeconomic Status Because choriocarcinoma is so rare, it has been difficult to evaluate whether the incidence differs according to socioeconomic status. In a recent report on time trends in Korea, the authors noted that the incidence of choriocarcinoma has decreased over a period of time during which the per capital gross national product has greatly increased (Kim et al., 1998). However, in data from the United States SEER registries, the lowest incidence was observed in the Southeastern United States, one of the poorest areas (Brinton et al., 1986). Baltazar found no association of choriocarcinoma incidence with socioeconomic factors in the Philipines (Baltazar, 1976). More data are available on hydatidiform mole, but results are again inconsistent. A considerably higher incidence of hydatidiform mole was observed in poorer sections of Singapore (Teoh et al., 1972), India (Mollykutty et al., 1993), and Turkey (Ozalp et al., 2001), but a similar pattern was not observed in China (Song and Wu, 1987), Saudi Arabia (Chattopadhay et al., 1998) or Hawaii (McCorriston, 1968). In case-control studies carried out in the U.S. (Messerli et al., 1985) and in Italy (LaVecchia et al., 1985), lower socioeconomic status was associated with increased risk of mole, but no association was observed in a case-control study from China (Brinton et al., 1989).

Age Choriocarcinoma incidence is clearly related to maternal age, with the incidence being many times higher above age 40 than at younger ages (Brinton et al., 1986; Baltazar, 1976; Teoh et al., 1972; Nakano et al., 1980; Shanmugaratnam et al., 1971). Incidence may also be higher among the very youngest women, although not all studies show this early peak (Ringertz, 1970; Shanmugaratnam et al., 1971). In US data, older age was found to be a major risk factor for choriocarcioma, with women aged 40 to 44 having 8 times the risk of those aged 20 to 24, and women under age 20 having a 2.5 times increased risk (Brinton

et al., 1986). It is possible that the increased risk observed for women under age 20 is the result of bias. Underreporting of induced and spontaneous abortions may be greater in very young women, leading to greater underestimation of the denominator than in other age groups. The association with older maternal age appears to be independent of the possible effect of parity. Studies of hydatidiform mole also show a greatly increased risk with older maternal age (Bagshawe et al., 1986; Atrash et al., 1986; Hayashi et al., 1982; Teoh et al., 1971). Maternal age was associated with complete mole but not with partial mole in two studies that analyzed the types separately (Parazzini et al., 1991; Graham et al., 1990). Parazzini raised the hypothesis that paternal age may have an independent effect on risk of gestational trophoblastic diseases (Parazzini et al., 1986). However, four other studies have assessed this factor and found no association (Matsuura et al., 1985; Yen and MacMahon 1968; Messerli et al., 1985; Brinton et al., 1989).

ENVIRONMENTAL FACTORS Oral Contraceptive Use A case-control study of choriocarcinoma in the Philippines first raised the hypothesis that past oral contraceptive use might increase the risk of gestational choriocarcinoma (Baltazar, 1976). An odds ratio of 6.4 was calculated for ever use of oral contraceptives. These findings were viewed with caution due to the small sample size (28 cases), low participation rate, and the possibility of differential reporting by cases and controls. Since that time, however, the finding has been replicated in all of three subsequent case-control studies with data on choriocarcinoma. Each of these studies, all carried out in the United States, took care to evaluate oral contraceptive use that occurred before the pregnancy that preceded the choriocarcinoma. In the largest, with 75 cases, the relative risk estimate for ever use of oral contraceptives was 2.0 and there was a significant trend of increased risk with increased duration (Buckley et al., 1988). In an analysis of 23 cases, the relative risk estimate was 2.2 (95% CI 0.7–7.0) for ever use and 6.0 (95% CI 1.3–28) for 5 or more years of use (Rosenberg et al., 1989). Finally, in a study specifically designed to evaluate the relation of oral contraceptive use to risk of gestational trophoblastic neoplasia, the relative risk estimate was 1.9 (95% CI 1.2–3.0) for ever use and 2.5 for 10 or more years use (Palmer et al., 1996). For oral contraceptive use during the cycle in which conception occurred, the relative risk estimate was 4.0 (95% CI 1.6–10). Results were similar for choriocarcinoma and invasive mole, with a relative risk estimate for ever use of 2.2 for choriocarcinoma and 1.8 for invasive mole.

1080

PART IV: CANCER BY TISSUE OF ORIGIN

Table 56–4. Studies of Oral Contraceptive Use Prior to Index Pregnancy in Relation to Risk of Gestational Trophoblastic Disease Geographic region Choriocarcinoma Baltazar, 1976 Philippines Buckley et al., 1988 United States Rosenberg et al., 1989 United States Palmer et al., 1996 United States Hydatidiform mole or persistent trophoblastic disease Berkowitz et al., 1985 United States Messerli et al., 1986 United States Rosenberg et al., 1989 United States Brinton et al., 1989 China Berkowitz et al., 1995 United States Palmer et al., 1996 United States Parazzini et al., 2002 Italy

Number of Cases

Relative Risk

95% Confidence Interval

Highest Use Category

28 75 23 50

6.4 2.0 6.0 2.8

(2.2–18.8) (0.88–4.93) (1.3–28) (0.7–10)

Ever use Ever use ≥5 years ≥7 years

2.1 No association 2.8 2.6 2.0 2.6 1.0

p = 0.05

Ever use Ever use ≥5 years ≥4 years >4 years ≥7 years Ever use

90 190 34 231 65 182 159

CM PTD PTD CM, PTD PM PTD CM

(0.9–8.8) (0.9–6.9) (0.9–4.6) (1.3–5.6) (0.7–1.8)

CM, complete mole; PM, partial mole; PTD, persistent trophoblastic disease.

Several other studies have also found a positive association of oral contraceptive use with risk of hydatidiform mole (Table 56–4). Four, including one from China (Brinton et al., 1989), one from Italy (Parrazzini et al., 2002), and two from the United States (Berkowitz et al., 1985; Berkowitz et al., 1995), observed an increased risk of complete mole, partial mole, or both, associated with use of oral contraceptives. One United States study found no association, regardless of duration or timing of use (Messerli et al., 1985). A separate question regarding oral contraceptives is whether use of oral contraceptives to prevent pregnancy during the first 12 months after a hydatidiform mole increases the risk of persistent trophoblastic disease. Stone et al. had found this to be the case (Stone et al., 1976), but the finding was not confirmed in several subsequent studies (Morrow et al., 1985; Deicas et al., 1991). Finally, a well-designed randomized trial evaluated this relation and established that use after evacuation of a mole does not increase the risk of persistent disease (Curry et al., 1989).

Diet Differences in incidence by geographic region and socioeconomic strata also prompted the hypothesis that nutritional factors might be related to risk of gestational trophoblastic disease. Llewellyn-Jones (1965) found that trophoblastic neoplasia was associated with a protein deficiency in Malaysia. Incidence of gestational trophoblastic neoplasia has decreased over time in Taiwan as per capita consumption of animal protein has increased (Hsu et al., 1988). On the other hand, incidence of hydatidiform mole has been found to be high among Alaskan natives, who consume a diet that is very high in animal fat (Martin, 1978). Two case-control studies of hydatidiform mole, one in Italy (Parazzini et al., 1988) and one in the United States (Berkowitz et al., 1985), found an inverse association with intake of carotene, and one (Parazzini et al., 1988) also found an inverse association with intake of animal protein. Vitamin A deficiency in animals has been shown to cause degeneration of the seminiferouis epithelium in the male and fetal resorption in the female, and fat deficiency has been shown to cause impaired ovulation (O’Toole et al., 1974; Kim et al., 1981). A case-control study in China found no association with intake of carotene-containing foods or with any other food group, but, as the authors noted, the dietary data were limited because of the minimal heterogeneity of diet in the population under study (Brinton et al., 1989).

Infection Because the incidence of gestational trophoblastic disease was initially noted to be higher in regions of the world with poor sanitary conditions and poor nutrition, an infectious etiology was hypothesized. This hypothesis was somewhat weakened with better incidence data that

showed less of a difference geographically and with inconsistent findings of an association with socioeconomic status. In Vietnam, pig farming was associated with an increased risk of mole (Ha et al., 1996). Contact with infected pigs may result in infection with parasites (Seuri, 1992). Findings of higher incidence among women under 20 (Slocumb and Lund, 1969; Neilsen, 1971) led to the hypothesis that a sexually transmitted organism may be involved. In United States studies, women who had been married more than once were found to have a higher risk of choriocarcinoma (Buckley et al., 1988) and of partial mole (Berkowitz et al., 1995). Such women would be expected to have had more sexual partners and thus a greater exposure to sexually transmitted infections. However, self-report of sexually transmitted infections was not associated with risk (Buckley et al., 1988). The hypothesis was tested in a subsequent study of choriocarcinoma and invasive mole: for both conditions, women who reported having had 10 or more sexual partners had two times the risk of disease (Palmer et al., 1996). This finding was statistically significant and independent of oral contraceptive use and other factors. Age at first intercourse and self-report of a history of sexually transmitted diseases were not associated with risk. Other studies of choriocarcinoma do not provide direct information on the question of infection. Two recent laboratory studies suggest possible organisms for a viral etiology: human papilloma virus (HPV) -18 was found to be present in a higher proportion of choriocarcinoma samples than expected (Pao et al., 1995) and adeno-associated virus was also found in an appreciable proportion of samples from moles and choriocarcinomas (Kiehl et al., 2002).

Smoking and Alcohol In the small study from the Philippines, cigarette smoking was associated with a two-fold risk of choriocarcinoma, although the increase was not statistically significant (Baltazar, 1976). In the larger and more methodologically sound U.S. study of choriocarcinoma, smoking was not associated with risk (Buckley et al., 1988). Studies of hydatidiform mole with data on smoking have mostly been null, with two exceptions that found positive associations (Parrazini et al., 1991; LaVecchia et al., 1985). Alcohol intake generally has not emerged as a potential risk factor for gestational trophoblastic diseases, and was not associated with choriocarcinoma in the largest US study (Buckley et al., 1998). Two recent findings from China are worth noting, however. In an ecologic analysis of choriocarcinoma mortality, mortality rates were found to be highest in the regions of the country where female alcohol intake was also the highest (Guo et al., 1994). Further, in a case-control study of invasive and non-invasive mole, a doubling in risk was observed for women who consumed at least 52 alcoholic beverages per year, the highest category of consumption (Brinton et al., 1989).

Choriocarcinoma

Chemical Exposures and Ionizing Radiation The incidence of gestational trophoblastic disease appears to be greatly increased in Vietnam, where herbicides, including Agent Orange and its dioxin contaminants were widely used during the Vietnam War (Ha et al., 1996). In two studies of maternal exposure to Agent Orange, living in regions that were moderately or heavily sprayed was associated with a greatly increased risk of hydatidiform mole, with an estimate relative risk of 12 in one study (Constable and Hatch, 1985) and 13 in the other (Phuong et al., 1989). In a third study, maternal exposure was not associated with increased risk (Ha et al., 1996). An ecologic investigation of the relation of ionizing radiation to risk of trophoblastic disease was carried out in Japan (Ujeno, 1985). The mean gonad dose equivalent of ionizing radiation in 11 prefectures in the period 1974–76 was highly correlated with the incidence of hydatidiform mole and persistent trophoblastic disease.

HOST FACTORS Genetic and Familial Factors Women who have one molar pregnancy are at greater risk of having a second. In a study of 5,030 English women followed after evacuation of a mole, 0.7% developed a second mole (Lorigan et al., 2000). Tuncer et al. reported the pregnancy experience of six women who had molar pregnancies with more than one partner (Tuncer et al., 2000). One of the women had three molar pregnancies, each with a different partner. By contrast, the male partners reported a total of seven healthy children from relationships with other women. These data provide evidence that maternal factors, perhaps underlying oocyte dysfunction, are more important than paternal factors in development of a mole. Other reports of women who developed moles with two different partners or sperm donors add to this evidence (Mangili et al., 1993; Pal et al., 1996). There is also evidence of a familial tendency toward gestational trophoblastic disease. Reports from Italy describe occurrences of hydatidiform mole in a pair of female homozygous twins (LaVecchia et al., 1982) and an association of family history of hydatidiform mole with risk of complete mole (LaVecchia et al., 1985). There have been at least two reports of family clusters of trophoblastic disease. In one, there were three families in which two or more sisters developed hydatidiform mole or choriocarcinoma (Ambani et al., 1980). In the other, two sisters, both married to first-degree cousins, had three and five molar pregnancies, respectively, while a second-degree cousin of theirs also had five moles (Seoud et al., 1995). HLA typing showed a high incidence of unusual HLA antigens in the sisters and their spouses.

Immune Function ABO blood group has been associated with risk of choriocarcinoma or persistent trophoblastic disease in at least four studies, with women who have group A blood having an increased risk (Bagshawe et al., 1971; Dawood et al., 1971; Scott, 1962; Parazzini et al., 1985a). In addition, in two of the studies, the highest risk was observed for women of group A whose partners had group O blood (Bagshawe et al., 1971; Parazzini et al., 1985a). Results are less consistent for hydatidiform mole. Two studies showed a higher proportion of group A blood in cases (Parazzini et al., 1985; Parazzini et al., 1991) and four found no association (Matsuura et al., 1984; Messerli et al., 1985; Olesnicky et al., 1985; Brinton et al., 1989). Based on blood group findings for choriocarcinoma, it was suggested that increased histocompatibility may play a role in susceptibility to persistent trophoblastic growth. HLA antigen studies have been inconclusive, however. Lawler et al. found that women who were more compatible with their husbands for B locus antigens were more likely to have high-risk trophoblastic disease (Lawler et al., 1978). In another study, Taiwanese women who had had a mole were more likely to share certain human leukocyte antigens with their husbands, although no association was observed among a similar group of U.S. women (Ho

1081

et al., 1989). In two other HLA antigen studies, there was no more compatibility between cases and their partners than among a control group of couples (Hansen et al., 1988; Yamashita et al., 1981).

Antecedent Pregnancy The strongest risk factor for gestational choriocarcinoma is the nature of the preceding pregnancy. Overall, it is estimated that 40–50% of choriocarcinomas are preceded by a hydatidiform mole (Park, 1950; Bagshawe et al., 1976). Women who experience a hydatidiform mole have a greatly increased risk of developing choriocarcinoma, estimated to be 1,000–2,000 times the risk following a live birth (Park, 1950). The malignant potential of a complete mole is considerably greater than that of a partial mole. There have been few documented cases of choriocarcinoma following a partial mole, and it is estimated that partial mole will progress to choriocarcinoma in only 0.1% of cases (Seckl et al., 2000). Approximately 3% of complete moles progress to choriocarcinoma (Hertig, 1956). A higher proportion has been reported in some geographic regions (Nielsen, 1971) but it is unclear whether the differences reflect inconsistences in case definition or indicate a different pathogenesis for choriocarcinoma and hydatidiform mole. A higher risk of choriocarcinoma has also been reported for ectopic pregnancies and for pregnancies that end in spontaneous abortion (Teoh, 1971). This may be due to an unrecognized early molar pregnancy, an associated chromosomal problem, or some effect of the abnormal termination of pregnancy.

Endogenous Hormonal Factors High parity (i.e., 5 or more births) has been linked to an increased risk of choriocarcinoma (Baltazar, 1976; Teoh et al., 1971). A similar association with parity was observed for risk of hydatidiform mole in Singapore (Llewyllen-Jones, 1965) and Australia (Olesnicky et al., 1985). However, most studies did not adjust for maternal age, and since high maternal age and high parity tend to be correlated, it is not clear whether high parity is an independent risk factor. High parity was not associated with risk of choriocarcinoma in the United States studies (Scott, 1962; Buckley et al., 1998), but there may have been few women with five or more births in the latter study. No association of high parity with hydatidiform mole was found in studies from the U.S. (Messerli et al., 1985; Atrash et al., 1986), Italy (Parrazini et al., 1985b), and China (Brinton et al., 1986), but these studies lacked an appreciable number of women with high parity. Choriocarcinoma was found to occur more often in conjunction with the first pregnancy in the earliest United States study (Scott, 1962), but this result was not replicated in other studies of choriocarcinoma. However, in almost all studies of hydatidiform mole that assessed the relation of parity, nulliparity was indeed associated with increased risk (Brinton et al., 1989; Olesnicky et al., 1985; Parrazini et al., 1985b; Atrash et al., 1986; Messerli et al., 1985). While Buckley et al. found no difference in choriocarcinoma risk by number of pregnancies or age at first pregnancy, cases were more likely to have consulted a physician for infertility and reported a longer time interval of sexual activity without contraception to conception (Buckley et al., 1988). LaVecchia and collegues reported an association of hydatidiform mole with history of infertility (LaVecchia et al., 1985). Messerli et al. found no association with history of infertility but observed that cases reported a longer period of abstaining from sexual intercourse before becoming pregnant (Messerli et al., 1985). Cases also were significantly more likely to have used abstinence as a method of birth control in the period before the index pregnancy. The authors suggested that in doing so cases may have increased their probability of conceiving at the beginning or the end of the menstrual cycle when abnormal conceptions may be more likely. In a casecontrol study from China, however, history of infertility was related to a reduced risk rather than increased risk of mole (Brinton et al., 1989). A history of spontaneous abortion was associated with increased risk of choriocarcinoma in the Philippines, and has been associated

1082

PART IV: CANCER BY TISSUE OF ORIGIN

with hydatidiform mole in at least three studies (Parrazini et al., 1985b; Acaia et al., 1988; Messerli et al., 1985). It has been suggested that there may be specific etiologic factors that are shared by both conditions, spontaneous abortion and hydatidiform mole. Late age at menarche was associated with increased risk in the only study of choriocarcinoma with information on this variable (Buckley et al., 1998). It was similarly associated with increased risk of mole in two studies (Berkowitz et al., 1985; Berkowitz et al., 1995) but not in two others (Messerli et al., 1985; Brinton et al., 1989). Light menstrual flow, very thin body build, and a history of twinning were also associated with increased risk of choriocarcinoma in one study (Buckley et al., 1998).

ASSOCIATED CANCERS Investigators in England have reported an increased risk of second cancers in patients who received combination chemotherapy for their gestational trophoblastic tumors (Rustin et al., 1996). An increased incidence of leukemia, colon cancer, breast cancer, and melanoma occurred. The increase in breast cancer incidence did not become apparent until 25 years after treatment. Two recent studies evaluated whether women who had had a hydatidiform mole, regardless of treatment, were at greater risk of any cancers other than choriocarcinoma. The first linked records on molar pregnancy with the records from the Danish Cancer registry and found that no other cancers occurred at a higher rate than expected in the group with molar pregnancy (Olsen et al., 1999). The study had limited power to detect an association, however, as it included only an average of nine years of follow-up for a relatively young group of women; only 20 cases of cancers other than choriocarcinoma occurred during the follow-up period. The second study was undertaken to test the hypothesis that women who had had a molar pregnancy and therefore experienced high levels of hCG, would have a reduced risk of breast cancer (Erlandsson et al., 2000). Linkage of Swedish data was carried out, and 3371 women with a history of mole were follow-up for a total of 57,075 person-years of follow-up during which 59 breast cancers occurred. History of mole was not associated with a reduced risk of breast cancer. In fact, the overall SIR was 1.3 (95% CI 1.0–1.7). The chief limitation of the study was the inability to control for parity, which is related to risk of breast cancer and possibly to the occurrence of hydatidiform mole.

PATHOGENESIS Prior hydatidiform mole and maternal age are the strongest and most well-established risk factors for both choriocarcinoma and hydatidiform mole. Cytogenetic studies indicate that the nuclear DNA in complete moles is paternally derived and that 75% of complete moles result from fertilization of an empty ovum by a haploid sperm; others receive dyspermic fertilization. The observed association of older maternal age and very young maternal age with an increased risk support these findings, in that there are more likely to be defects in ovoid function in premature or post mature ova. Occurrence of two or more moles in women with different male partners and of several cases among close female relatives provide further evidence that the pathology is in the female reproductive system rather than the male. There is little evidence that male factors play a role in the pathogenesis of hydatidiform mole or choriocarcinoma. The role of tumor suppressor genes and protooncogenes remains unclear and further studies are necessary to better understand the molecular factors involved in the pathogenesis of choriocarcinoma. Additionally, research on growth factors, telomerase activity, cell adhesion molecules, and metalloproteinases may provide additional insight into the mechanisms responsible for persistence and malignant transformation in gestational trophoblastic disease. Buckley et al. (1988) have proposed that a “low-estrogen” state may predispose to the development of choriocarcinoma. This was put forth as a unifying hypothesis to explain findings from their case-control

study of a higher choriocarcinoma risk in women who had a later age at menarche, low body mass index, light menstrual flow, infrequent sexual intercourse, previous birth of dzygotic twins, and history of infertility. If low estrogen levels do characterize women who are at higher risk of choriocarcinoma, there are several possible mechanisms. First, low estrogen levels with partially suppressed ovulation could affect oocyte development, perhaps giving rise to ovulation of an abnormal ovum. Alternatively, low estrogen levels may simply be a result of an ovarian abnormality that is independently related to disease risk. Although international variations in incidence are not as great as first appeared, it is clear that differences do exist. Choriocarcinoma incidence is at least two times as high in many Asian countries as in the United States and most parts of Europe. The picture is not as clear for hydatidiform mole, however. As noted above, in the largest population-based study, the incidence in China was found to be approximately the same as in the United States and several European countries. Somewhat higher rates, up to two times higher, are seen in recent population-based data from Japan and Korea as well as in data from England and Northern Ireland. More interesting than international variations are variations by race/ethnicity. A fairly consistent picture emerges, both for choriocarcinoma and hydatidiform mole. The highest incidence is seen in the mixed Eskimo-Caucasian population of Greenland, in some of the Native American peoples of North America (native Alaskans, Inuits of Canada, and Indians of New Mexico) and in certain ethnic groups populating regions of Asia or other parts of the world. Limited data available on migrant populations do not indicate a change in incidence with migration to a different region. Overall, race/ethnicity incidence data suggest that there is a strong genetic component to the etiology of choriocarcinoma. Although a few hypotheses concerning specific genes have been put forth, none have been substantiated. For instance, HLA-9 antigen appears to be higher in Greenland and in parts of Asia (Nielsen, 1979; Dawood et al., 1971). In Nigeria the incidence of hydatidiform mole was found to be highest in the Yoruba peoples, who differ from the other major ethnic groups of Nigeria in that they carry a gene for hemoglobin C (Lehman and Nwokolo, 1959). Time trends indicate that factors other than genetics are also involved in the etiology of choriocarcinoma. The incidence has decreased in several Asian countries and in the United States, even after taking into account changing fertility patterns. In addition, an increase in incidence over time has been seen in at least one group, the American Indians of New Mexico. While no environmental factor has been definitively linked with risk of choriocarcinoma, there have been some associations that are consistent with what is known about the genetic pathogenesis, while other associations provide clues to environmental factors. As noted above, a consistent association has been observed between oral contraceptive use before the index pregnancy and increased risk of hydatidiform mole and/or choriocarcinoma. A satisfactory biologic explanation for this association has not been established. One possibility is that long-duration oral contraceptive use may damage the ova or interfere with meiosis, yielding ova with absent or inactivated nuclei. Alternatively, oral contraceptive use may affect ovarian function, altering the mechanism for selecting ova. It is also possible that oral contraceptive use is actually a marker for some other true etiologic factor, for example, infection with sexually transmitted organisms. Against this is the fact that the association with oral contraceptive use was found even among women who reported having had only one sexual partner. Epidemiologic studies provide some limited evidence that infections, particularly sexually transmitted infections, may play a role in the etiology of choriocarcinoma. Recent tissue studies raise the hypothesis that HPV-18 or adenoassociated virus may be involved. The potential role these viruses play warrants further research. The unusually high incidence of trophoblastic disease in Vietnam, coupled with the high exposure to a known carcinogen (Agent Orange) suggests that exposure to this herbicide may cause damage leading to hydatidiform mole and choriocarcinoma. Although not all case-control studies from Vietnam found an association with herbicide exposure,

Choriocarcinoma two found a more than 10-fold risk for the exposed group. It is difficult to measure exposure to toxic substances such as herbicides, pesticides, and the organic solvents used occupationally or in households, and studies done to date could easily have missed such an association. Correlation of ambient ionizing radiation with incidence of trophoblastic disease in Japan provides evidence that ionizing radiation may play a role. Radiation exposure could cause genetic defects leading to fertilization of an ovum without genetic material from the mother. Blood group has not proved to be useful as a prognostic factor for patients with hydatidiform mole. Nevertheless, the number of studies that have found positive associations with blood group make this an area worthy of further research. Many case-control studies have not assessed this factor because of difficulty in obtaining the information, particularly for the male partner, whose identity may not be certain. A role for diet is intriguing, in part because of the biologic plausibility of an effect of vitamin A deficiency, but there are no direct data related to choriocarcinoma, and findings regarding moles are inconclusive.

PREVENTIVE MEASURES Primary Prevention and Chemoprevention At this time there are no effective primary prevention measures for choriocarcinoma or hydatidiform mole. Since hydatidiform mole is the strongest known risk factor for gestational choriocarcinoma, the efficacy of prophylactic treatment for complete mole has been tested. Several studies found that chemoprophylaxis substantially reduced the incidence of postmolar tumors in women with high-risk complete moles (Goldstein et al., 1971; Fasoli et al., 1982; Kim et al., 1986; Kashimura et al., 1986). However, women who developed persistent disease in spite of the treatment were more likely to have high-risk tumors and require combination therapy with its accompanying adverse effects. For this reason, chemoprophylaxis is controversial and is not widely practiced in the United States. It may be considered in situations where hormonal testing is unavailable or follow-up is unreliable.

Screening and Early Detection Women who have a molar pregnancy should be followed closely for the development of persistent trophoblastic growth. In developed countries, this is easily done by monitoring of serum beta-hCG levels, which provides a sensitive method for identification and follow-up of persistent disease. In the United States, the standard protocol following evacuation of a molar pregnancy is to measure beta-hCG levels weekly until they are normal for three consecutive weeks, followed by monthly evaluations for six consecutive normal levels. Effective, reliable contraception should be used during the follow-up period because pregnancy would raise beta-hCG levels and make detection of persistent disease difficult. Follow-up is recommended for both partial and complete moles. Criteria for determining the presence of persistent disease differ. In the United Kingdom, for example, a more conservative approach is used and as a result only about 8–14% of patients who have had a mole undergo chemotherapy (Bagshawe et al., 1986; Paradinas, 1998). In the United States, where chemotherapy may be begun if the hCG level plateaus over a time interval of at least three consecutive weeks, about 18–29% of mole patients have required chemotherapy (Curry et al., 1975; Morrow et al., 1977; Goldstein et al., 1982; Kohorn et al., 1982; Lurain et al., 1983). These differences may lead to different estimates of incidence of choriocarcinoma, as it may not be possible to have a definitive diagnosis of choriocarcinoma if the tumor responds to early treatment. In summary, there is effective early screening for women who have a molar pregnancy, a group that comprises 40–50% of all women with choriocarcinoma. Such screening has led to early treatment and a high survival rate. In general, women who have normal pregnancies do not receive screening for choriocarcinoma.

1083

FUTURE DIRECTIONS Advances in understanding the cytogenetics of hydatidiform mole and choriocarcinoma lay the groundwork for more informative epidemiologic studies of risk factors. Studies of hydatidiform mole should consider complete and partial moles separately, and most of the efforts should be directed at complete moles. Studies of choriocarcinoma should include all persistent trophoblastic disease, although distinctions as to type should be made where possible. Because choriocarcinoma is so rare, it may be helpful to include cases of high-risk complete mole in case-control studies. An estimated 40% of these will progress to persistent disease. It would also be helpful to have more data on risk factors for gestational choriocarcinomas not preceded by a molar pregnancy. Late maternal age and previous hydatidiform mole remain the only established risk factors for choriocarcinoma. One promising area for future research is investigation of the role of sexually transmitted infections. Studies should obtain information on sexual activity (number of partners, use of barrier methods of contraception, etc) and tissue markers of viral infections (e.g., HPV-18). Further investigation of oral contraceptive use and contraceptive failure is also of interest. In particular, future studies should examine the role of delayed ovulation and delayed fertilization in relation to risk of complete mole and choriocarcinoma. References Acaia B, Parazzini F, LaVecchia C, Ricciardiello O, Fedele L, Battista Candiani G. 1988. Increased frequency of complete hydaditiform mole in women with repeated abortion. Gynecol Oncol 31:310–314. Ambani LM, Vaidya RA, Rao CS, et al. 1980. Familial occurrence of trophoblastic disease-report of recurrent molar pregnancies in sisters in three families. Clin Genet 18:27–29. Atrash HK, Hogue CJ, Grimes DA. 1986. Epidemiology of hydatidiform mole during early gestation. Am J Obstet Gynecol 154:906–909. Bagshawe KD, Rawlins G, Pike MC, Lawler SD. 1971. ABO blood-groups in trophoblastic neoplasia. Lancet 1:553–556. Bagshawe KD. 1976. Risk and prognostic factors in trophoblastic neoplasia. Cancer 38:1373–1385. Bagshawe KD, Dent J, Webb J. 1986. Hydatidiform mole in England and Wales 1973–1983. Lancet 2:673–677. Baltazar JC. 1976. Epidemiological features of choriocarcinoma. Bull World Health Organ 54:523–532. Bassaw B, Roopnarinesingh S. 1990. The epidemiology and management of patients with hydatidiform mole. W I Med J 39:43–46. Berkowitz RS, Cramer DW, Bernstein MR, Cassells S, Driscoll SG, Goldstein DP. 1985. Risk factors for complete molar pregnancy from a case-control study. Am J Obstet Gynecol 152:1016–1020. Berkowitz RS, Goldstein DP, Bernstein MR. 1986. Ten year’s experience with methotrexate and folinic acid as primary therapy for gestational trophoblastic disease. Gynecol Oncol 23:111–118. Berkowitz RS, Bernstein MR, Harlow BL, et al. 1995. Case-control study of risk factors for partial molar pregnancy. Am J Obstet Gynecol 173:788–794. Brewis AA, Peddie KA. 1996. Hydatidiform mole in Micronesian women. N Z Med J 109:38–39. Brinton LA, Bracken MB, Connelly RR. 1986. Choriocarcinoma incidence in the United States. Am J Epidemiol 123:1094–1100. Brinton LA, Wu BZ, Wang W, et al. 1989. Gestational trophoblastic disease: a case-control study from the People’s Republic of China. Am J Obstet Gynecol 161:121–127. Buckley JD, Henderson BE, Morrow CP, Hammond CB, Kohorn EI, Austin DF. 1988. Case-control study of gestational choriocarcinoma. Cancer Res 48:1004–1010. Chattopadhyay SK, Sengupta BS, al Ghreimil M, Edrees YB, Lambourne A. 1998. Epidemiologic study of gestational trophoblastic diseases in Saudi Arabia. Surg Gynecol Obstet 167:393–398. Cheung AN, Shen DH, Khoo US, Wong LC, Ngan HY. 1998. p21WAF1/CIP1 expression in gestational trophoblastic disease: correlation with clinicopathological parameters, and Ki67 and p53 gene expression. J Clin Pathol 51:159–162. Cheung AN, Shen DH, Khoo US, et al. 1999. Immunohistochemical and mutational analysis of p53 tumor suppressor gene in gestational trophoblastic disease: correlation with mdm2, proliferation index, and clinicopathologic parameters. Int J Gynecol Cancer 9(2):123–130.

1084

PART IV: CANCER BY TISSUE OF ORIGIN

Cheville JC, Robinson RA, Benda JA. 1996. p53 expression in placentas with hydropic change and hydatidiform moles. Mod Pathol 9:392–396. Constable JD, Hatch MC. 1985. Reproductive effects of herbicide exposure in Vietnam: recent studies by the Vietnamese and others. Teratol Carcinog Mutagen 5:231–250. Curry SL, Hammond CB, Tyrey L, Creasman WT, Parker RT. 1975. Hydatidiform mole: diagnosis, management, and long-term follow-up of 347 patients. Obstet Gynecol 45:1–8. Curry SL, Schlaerth JB, Kohorn EI, et al. 1989. Hormonal contraception and trophoblastic sequelae after hydatidiform mole (a Gynecologic Oncology Group Study). Am J Obstet Gynecol 160:805–809. Dawood MY, Teoh ES, Ratnam SS. 1971. ABO blood group in trophoblastic disease. J Obstet Gynaec Br Commonw 78:918–923. Deicas RE, Miller DS, Rademaker AW, Lurain JR. 1991. The role of contraception in the development of post molar-gestational trophoblastic tumor. Obstet Gynecol 78:221–226. Duff GB. 1989. Gestational trophoblastic disease in New Zealand, 1980–1986. Aust N Z J Obstet Gynaecol 29:139–142. Egwuatu VE, Ozumba BC. 1989. Observations on molar pregnancy in Enugu, Nigeria. Int J Gynaecol Obstet 29:219–225. Erlandsson G, Weiderpass E, Lambe M, Ekbom E. 2000. Hydatidiform moles and the long term risk of breast cancer (Sweden). Cancer Causes Control 11:117–120. Fasoli M, Ratti E, Franceschi S, LaVecchia C, Pecorelli S, Mangioni C. 1982. Management of gestational trophoblastic disease: results of a cooperative study. Obstet Gynecol 60:205–209. Fisher RA, Povey S, Jeffreys AJ, Martin CA, Patel I, Lawler SD. 1989. Frequency of heterozygous complete hydatidiform moles, estimated by locus-specific minisatellite and Y chromosome-specific probes. Hum Genet 82:259–263. Fisher RA, Lawler SD, Povey S, Bagshawe KD. 1998. Genetically homozygous choriocarcinoma following pregnancy with hydatidiform mole. Br J Cancer 58:788–792. Flam F, Lundstrom-Lindstedt V, Rutqvist LE. 1992. Incidence of gestational trophoblastic disease in Stockholm County, 1975–1988. Eur J Epidemiol 8:173–177. Franke HR, Risse EKJ, Kenemans P, Vooijs GP, Stolk JG. 1983. Epidemiologic features of hydatidiform mole in the Netherlands. Obstet Gynecol 62:613–616. Fukunaga, M. 2000. Early partial hydatidiform mole: prevalence, histopathology, DNA ploidy, and persistence rate. Virchows Arch 437:180–184. Fulop V, Mok SC, Genest DR, Gati I, Doszpod J, Berkowitz RS. 1998a. p53, p21, Rb and mdm2 oncoproteins. Expression in normal placenta, partial and complete mole, and choriocarcinoma. J Reprod Med 43: 119–127. Fulop V, Colitti CV, Genest D, et al. 1998b. DOC-2/hDab2, a candidate tumor suppressor gene involved in the development of gestational trophoblastic diseases. Oncogene 17:419–424. Fulop V, Mok SC, Genest DR, Szigetvari I, Cseh I, Berkowitz RS. 1998c. cmyc, c-erbB-2, c-fms and bcl-2 oncoproteins. Expression in normal placenta, partial and complete mole, and choriocarcinoma. J Reprod Med 43:101–110. Gaber LW, Redline RW, Mostoufi-Zadeh M, Driscoll SG. 1986. Invasive partial mole. Am J Clin Pathol 85:722–724. Gaudette LA, Gao R-N, Freitag S, Wideman M. 1993. Cancer incidence by ethnic group in the Northwest Territories (NWT) 1969–1988. Health Reports 5:23–32. Genest DR, Ruiz RE, Weremowicz S, Berkowitz RS, Goldstein DP, Dorfman DM. 2002. Do nontriploid partial hydatidiform moles exist? A histologic and flow cytometric reevaluation of nontriploid specimens. J Reprod Med 47:363–368. Giwa-Osagie MO, Okwerekwu G. 1999. Epidemiology of molar pregnancies in Northern Ireland. Int J Gynaecol Obstet 66:175–177. Goldstein DP. 1971. Prophylactic chemotherapy of patients with molar pregnancy. Obstet Gynecol 38:817–822. Goldstein DP, Berkowitz RS. 1982. The diagnosis and management of molar pregnancy. Gestational trophoblastic neoplasms: clinical principles of diagnosis and management. Major Problems in Obstetrics and Gynecology Vol. 14, Philadelphia, WB Saunders, pp. 143–175. Graham IH, Fajardo AM. 1988. The incidence and morphology of hydatidiform mole in Abu Dhabi, United Arab Emirates, 1978–1986. Br J Obstet Gynaecol 95:391–392. Graham IH, Fajardo AM, Richards RL. 1990. Epidemiological study of complete and partial hydatidiform mole in Abu Dhabi: influence age and ethnic group. J Clin Pathol 43:661–664. Guo WD, Chow WH, Li JY, Chen JS, Blot WJ. 1994. Correlations of choriocarcinoma mortality with alcohol drinking and reproductive factors in China. Eur J Cancer Prev 3:223–226.

Ha MC, Cordier S, Bard D, et al. 1996. Agent orange and the risk of gestational trophoblastic disease in Vietnam. Arch Environ Health 51:368–374. Habibian R, Surti U. 1987. Cytogenetics of trophoblasts from complete hydatidiform moles. Cancer Genet Cytogenet 29:271–287. Halperin R, Peller S, Sandbank J, Bukovsky I, Schneider D. 2000. Expression of the p53 gene and apoptosis in gestational trophoblastic disease. Placenta 21:58–62. Hammond CB, Hertz R, Ross GT, Lipsett MB, Odell WD. 1967. Primary chemotherapy for nonmetastatic gestational trophoblastic neoplasms. Am J Obstet Gynecol 98:71–78. Hando T, Ohno M, Kurose T. 1998. Recent aspects of gestational trophoblastic disease in Japan. Int J Gynaecol Obstet 60 Suppl 1:S71–S76. Hansen HE, Vejerslev LO, Larsen SO, Gurtler H. 1988. HLA and hydatidiform mole. I. HLA and ABO antigen frequencies in couples with androgenetic and diandric conceptuses. Tissue Antigens 32:87–92. Hayashi K, Bracken MB, Freeman DH, Hellenbrand K. 1982. Hydatidiform mole in the United States (1970–1977): a statistical and theoretical analysis. Am J Epidemiol 115:67–77. Henry I, Bonaiti-Pellie C, Chehensse V, et al. 1991. Uniparental paternal disomy in a genetic cancer-predisposing syndrome. Nature 351(6328): 665–667. Hertig AT, Gore H. 1956. Tumors of the female sex organs. Washington, Armed Forces Institute of Pathology, Vol. 2. Hertz R, Bergenstal DM, Lipsett MB, et al. 1959. Chemotherapy of choriocarcinoma and related trophoblastic tumours in women. Ann NY Acad Sci 80:262–277. Ho HN, Gill TJ 3rd, Klionsky B, et al. 1989. Differences between white and Chinese populations in human leukocyte sharing and gestational trophoblastic tumors. Am J Obstet Gynecol 161:942–948. Hsu CT, Jong HL, Wang TT, et al. 1988. Recent changes in the epidemiology of GTN in Taiwan, ROC: preliminary report. Asia-Oceania J Obstet Gynaecol 14:61–73. Jeffers MD, Michie BA, Oakes SJ, Gillan JE. 1995. Comparison of ploidy analysis by flow cytometry and image analysis in hydatidiform mole and non-molar abortion. Histopathology 27:415–421. Kajii T, Ohama K. 1977. Androgenetic origin of hydatidiform mole. Nature 268(5621):633–634. Kashimura Y, Kashimura M, Sugimori H, et al. 1986. Prophylactic chemotherapy for hydatidiform mole: five to 15 years follow-up. Cancer 58: 624–629. Kiehl K, Schlehofer JR, Schultz R, Zugaib M, Armbruster-Moraes E. 2002. Adeno-associated virus DNA in human gestational trophoblastic disease. Placenta 23:410–415. Kim DS, Moon H, Kim KT, Moon YJ, Hwang YY. 1986. Effects of prophylactic chemotherapy for persistant trophoblastic disease in patients with complete hydatidiform mole. Obstet Gynecol 67:690–694. Kim HL, Picciano MF, O’Brien W. 1981. Influence of maternal dietary protein and fat levels on fetal growth in mice. Growth 45:8–18. Kim SJ, Bae SN, Kim JH, et al. 1998. Epidemiology and time trends of gestational trophoblastic disease in Korea. Int J Gynaecol Obstet 60 Suppl 1:S33-8. Koenig C, Demopoulos RI, Vamvakas EC, Mittal KR, Feiner HD, Espiritu EC. 1993. Flow cytometric DNA ploidy and quantitative histopathology in partial moles. Int J Gynecol Pathol 12:235–240. Kohorn EI. 1980. Hydatidiform mole and gestational trophoblastic disease in southern Connecticut. Obstet Gynecol 59:78–84. Lage JM, Mark SD, Roberts DJ. A flow cytometric study of 137 fresh hydropic placentas: correlation between types of hydatidiform moles and nuclear DNA ploidy. Obstet Gynecol 79:403–410. LaVecchia C, Franceschi S, Fasoli M, et al. 1982. Gestational trophoblastic neoplasms in homozygous twins. Obstet Gynecol 60:250–252. LaVecchia C, Parazzini F, Decarli A, et al. 1984. Age of parents and risk of gestational trophoblastic disease. J Natl Cancer Inst 73:639–642. LaVecchia C, Franceschi S, Parazzini F, et al. 1985. Risk factors for gestational trophoblastic disease in Italy. Am J Epidemiol 121:457–464. Lawler SD. 1978. HLA and trophoblastic tumours. Br Med Bull 34:305– 308. Lawler SD, Fisher RA, Dent J. 1991. A prospective genetic study of complete and partial hydatidiform moles. Am J Obstet Gynecol. 164(5 Pt 1): 1270–1277. Lee YS. 1995. p53 expression in gestational trophoblastic disease. Int J Gynecol Pathol 14:119–124. Lehmann H, Nwokolo C. 1959. The River Niger as a barrier in the spread eastwards of Haemoglobin C: a survey of haemoglobin in the Ibo. Nature 183:1587. Little M, Van Heyningen V, Hastie N. 1991. Dads and disomy and disease. Nature 351(6328):609–610.

Choriocarcinoma Llewellyn-Jones D. 1965. Trophoblastic tumours: geographical variations in incidence and possible aetiological factors. J Obstet Gynaecol Br Commonw 72:242–248. Lorigan PC, Sharma S, Bright N, et al. 2000. Characteristics of women with recurrent mlar pregnancies. Gynecol Oncol 78:288–292. Loukovaara M, Pukkala E, Lehtovirta P, Leminen A. 2004. Epidemiology of choriocarcinoma in Finland, 1953–1999. Gynecol Oncol 92:252–255. Loukovaara M, Pukkala E, Lehtovirta P, Leminen A. 2005. Epidemiology of Hydatidiform mole in Finland, 1975–2001. Eur J Gynaecol Oncol 26:207–208. Lurain JR, Brewer JI, Torok EE, Halpern B. 1983. Natural history of hydatidiform mole after primary evacuation. Am J Obstet Gynecol 145:591–595. Mangili G, Parazzini F, Maggi R, Spotli R, Spotli N. 1993. Repeated gestational trophoblastic disease after natural and heterologous assisted conception A case report. J Repro Med 38:405–406. Martin PM. 1978. High frequency of hydatidiform mole in native Alaskans. Int J Gynaecol Obstet 15:395–396. Matalon M, Paz B, Modan M, et al. 1972. Malignant trophoblastic disorders: epidemiologic aspects and relationship to hydatidiform mole. Am J Obstet Gynecol 112:101–106. Matsui H, Iizuka Y, Yamazawa K, et al. 2003. Changes in the incidence of molar pregnancies. A population-based study in Chiba Prefecture and Japan between 1974 and 2000. Hum Reprod 18:172–175. Matsuura J, Chiu D, Jacobs PA, Szulman AE. 1984. Complete hydatidiform mole in Hawaii: an epidemiological study. Genet Epidemiol 1:271–284. Mazzanti P, LaVecchia C, Parazzini F, Bolis G. 1986. Frequency of hydatidiform mole in Lombardy, Northern Italy. Gynecol Oncol 24:337–342. McCorriston CC. 1968. Racial incidence of hydatidiform mole. A study in a contained polyracial community. Am J Obstet Gynecol 101:377–382. Messerli ML, Lilienfeld AM, Parmley T, Woodruff JD, Rosenshein NB. 1985. Risk factors for gestational trophoblastic neoplasia. Am J Obstet Gynecol 153:294–300. Molykutty J, Prabha B, Rajalekshmy TN, et al. 1993. Profile of gestational trophoblastic disease in Kerala, India. Med Sci Res 21:431–432. Morrow CP, Kletzky OA, Disaia PJ, Townsend DE, Mishell DR, Nakamura RM. 1977. Clinical and laboratory correlates of molar pregnancy and trophoblastic disease. Am J Obstet Gynecol 128:424–430. Morrow P, Nakamura R, Schlaerth J, Gaddis O, Eddy G. 1985. The influence of oral contraceptives on the postmolar human chorionic gonadotropin regression curve. Am J Obstet Gynecol 151:906–914. Mutter GL, Pomponio RJ, Berkowitz RS, Genest DR. 1993. Sex chromosome composition of complete hydatidiform moles: relationship to metastasis. Am J Obstet Gynecol 168(5):1547–1551. Nakano R, Sasaki K, Yamoto M, Hata H. 1980. Trophoblastic disease: analysis of 342 patients. Gynecol Obstet Invest 11:237–242. Nielsen NH, Hansen JP. 1979. Trophoblastic tumours in Greenland. J Cancer Res Clin Oncol 95:177–186. Nowak-Markwitz E, Drews K, Spaczynski M. 2000. Gestational trophoblastic disease: the epidemiological analysis of 342 cases. Ginekol Pol 71: 767–772. Olesnicky G, Long AR, Quinn MA, Pepperell RJ, Fortune DW, Kneale BL. 1985. Hydatidiform mole in Victoria: aetiology and natural history. Aust NZJ Obstet Gynaecol 25:1–7. Olsen JH, Mellemkjaer L, Gridley G, et al. 1999. Molar prenancy and risk for cancer in women and their male partners. Am J Obstet Gynecol 181: 630–634. O’Toole BA, Fradkin R, Warkany J, Wilson JG, Mann GV. 1974. Vitamin A deficiency and reproduction in rhesus monkeys. J Nutr 104:1513–1524. Ozalp SS, Yalcin OT, Tanir HM. 2001. Hydatidiform mole in Turkey from 1932 to 2000. Int J Gynaecol Obstet 73:257–258. Paksoy N, Reich B. 1989. The occurrence of trophoblastic disease in Western Samoa. NZ Med J 102:162–163. Pal L, Toth TL, Leykin L, Isaacson KB. 1996. High incidence of triploidy in in-vitro fertilized oocytes from a patient with a previous history of recurrent gestational trophoblastic disease. Human Reprod 11:1529–1532. Palmer JR, Driscoll SG, Rosenberg L, et al. 1999. Oral contraceptive use and risk of gestational trophoblastic tumours. J Natl Cancer Inst 91:635– 640. Pao CC, Hor JJ, Wu CJ, Shi YF, Xie X, Lu SM. 1995. Human papillomavirus type 18 DNA in gestational trophoblastic tissues and choriocarcinomas. Int J Cancer 63:505–509. Paradinas FJ. 1997. Gestational trophoblastic disease. In Hancock ESN BW, Berkowitz RS, eds. Pathology. London, Chapman & Hall, pp. viii, 266. Paradinas FJ. 1998. The diagnosis and prognosis of molar pregnancy: The experience of the National Referral Centre in London. Int J Obstet Gynecol 60:S57–S64. Parazzini F, LaVecchia C, Franceschi S, et al. 1985a. ABO blood-groups and the risk of gestational trophoblastic disease. Tumouri 1:123–126.

1085

Parazzini F, LaVecchia C, Pampallona S, Franceschi S. 1985b. Reproductive patterns and the risk of gestational trophoblastic disease. Am J Obstet Gynecol 52:866–870. Parazzini F, LaVecchia C, Pampallona S. 1986. Parental age and risk of complete and partial hydatidiform mole. Br J Obstet Gynaecol 93: 582–585. Parazzini F, LaVecchia C, Mangili G, et al. 1988. Dietary factors and risk of trophoblastic disease. Am J Obstet Gynecol 158:93–99. Parazzini F, Mangili G, LaVecchia C, Negri E, Bocciolone L, Fasoli M. 1991. Risk factors for gestational trophoblastic disease: a separate analysis of complete and partial hydatidiform moles. Obstet Gynecol 78:1039–1045. Parazzini F, Cipriani S, Mangili G, et al. 2002. Oral contraceptives and risk of gestational trophoblastic disease. Contraception 65:425–427. Park WW, Lees JC. 1950. Choriocarcinoma. A general review, with analysis of 516 cases. Arch Pathol 49:73–104. Pavlidis N, Briassoulis E, Bai M, Fountzilas G, Agnantis N. 1995. Overexpression of C-myc, Ras and C-erbB-2 oncoproteins in carcinoma of unknown primary origin. Anticancer Res 15:2563–2567. Phuong NTN, Thuy TT, Phuong PK. 1989. An estimate of differences among women giving birth to deformed babies and among those with hydatidiform mole seen at the Ob-Gyn Hospital of Ho Chi Minh City in the south Vietnam. Chemosphere 8:801–803. Reik W. 1989. Genomic imprinting and genetic disorders in man. Trends Genet 5:331–336. Ringertz N. 1970. Hydatidiform mole, invasive mole and choriocarcinoma in Sweden 1958–1965. Acta Obstet Gynecol Scand 49:195–203. Rolon PA, Hochsztajn DL. 1979. Malignant trophoblastic disease in Paraguay. J Reprod Med 23:94–96. Rolon PA, Hochsztajn B, Llamosas F. 1990. Epidemiology of complete hydatidiform mole in Paraguay. J Reprod Med 35:15–18. Rosenberg L, Palmer JR, Shapiro S. 1989. Gestational trophoblastic disease and use of contraceptives. Am J Obstet Gynecol 161:1087–1088. Rustin GJS, Newlands ES, Lutz JM, et al. 1996. Combination but not singleagent methotrexate chemotherapy for gestational trophoblastic tumors increases the incidence of second tumors. J Clin Oncol 14:2669–2673. Sakumoto K, Higashi M, Kanazawa K. 1999. Int J Obstet Gynecol 64:309–310. Sakuragi N, Matsuo H, Coukos G, et al. 1994. Differentiation-dependent expression of the BCL-2 proto-oncogene in the human trophoblast lineage. J Soc Gynecol Investig 1:164–172. Scott JS. 1962. Choriocarcinoma. Observations on the etiology. Am J Obstet Gynecol 83:185–193. Seckl MJ, Fisher RA, Salerno G, et al. 2000. Choriocarcinoma and partial hydatidiform moles. Lancet 356:36–39. Sengupta BS, Persaud V, Wynter HH. 1977. Choriocarcinoma in Jamaica. Int Surg 62:84–87. Seoud M, Khalid A, Frangieh A, et al. 1995. Recurrent molar pregnancies in a family with extensive intermarriage: report of a family and review of the literature. Obstet Gynecol 86:692–695. Seuri M, Koskela P. 1992. Contact with pigs and cats associated with high prevalence of toxoplasma antibodies among farmers. Br J Ind Med 49:845–849. Shang ER. 1982. A retrospective investigation on the incidence of hydatidiform mole in 20,548 fertile women. Zhonghua Yi Xue Za Zhi 62: 282–285. Shanmugaratnam K, Muir CS, Tow SH, Cheng WC, Christine B, Pedersen E. 1971. Rates per 100,000 births and incidence of choriocarcinoma and malignant mole in Singapore Chinese and Malays. Comparison with Connecticut, Norway and Sweden. Int J Cancer 8:165–175. Slocumb JC, Lund CJ. 1969. Incidence of trophoblastic disease: increased rate in youngest age group. Am J Obstet Gynecol 104:421–423. Smith HO, Qualls CR, Prairie BA, Padilla LA, Rayburn WF, Key CR. 2003. Trends on gestational choriocarcinoma: a 27-year perspective. Obstet Gynecol 102:978–987. Smith HO, Qualls CR, Hilgers RD, Verschraegen CF, Rayburn WF, Cole LW, Padilla LA, Key CR. 2004. Gestational trophoblastic neoplasia in American Indians. J Reprod Med 49:535–544. Song HZ, Wu PC. 1986. Incidence of hydatidiform mole in China. Int J Epidemiol 15:429–430. Song HZ, Yang XY, Xiang Y. 1998. Forty-five years experience of the treatment of choriocarcinoma and invasive mole. Int J Obstet Gynecol 60: S77–S83. Srivannaboon S, Vatananusara C, Boonyanit S. 1974. The incidence of trophoblastic disease in Siriraj Hospital. J Med Assoc Thai 57:537–542. Stone M, Dent J, Kardana A, Bagshawe KD. 1978. Relationship of oral contraception to development of trophoblastic tumour after evacuation of a hydatidiform mole. Br J Obstet Gynaecol 83:913–916. Surti U, Szulman AE, O’Brien S. 1979. Complete (classic) hydatidiform mole with 46, XY karyotype of paternal origin. Hum Genet 51:153–155.

1086

PART IV: CANCER BY TISSUE OF ORIGIN

Szulman AE, Ma HK, Wong LC, Hsu C. 1981. Residual trophoblastic disease in association with partial hydatidiform mole. Obstet Gynecol 57: 392–394. Talati NJ. 1998. The pattern of benign gestational trophoblastic disease in Karachi. J Pak Med Assoc 48:296–300. Teoh ES, Dawood MY, Ratnam SS. 1971. Epidemiology of hydatidiform mole in Singapore. Am J Obstet Gynecol 110:415–420. Teoh ES, Dawood MY, Ratnam SS. 1972. Observations on choriocarcinoma in Singapore. Obstet Gynecol 40:519–524. Tham BW, Everard JE, Tidy JA, Drew D, Hancock BW. 2003. Gestational trophoblastic disease in the Asian population of Northern England and North Wales. BJOG 110:555–559. Tuncer ZS, Bernstein MR, Wang J, Goldstein DP, Berkowitz RS. 1999. Repetitive hydatidiform mole with different male partners. Gynecol Oncol 75:224–226. Ujeno Y. 1985. Epidemiological studies of human fetal development in areas with various doses of natural background radiation. II. Relationship between incidences of hydatidiform mole, malignant hydatidiform mole, and chorioepithelioma and gonad dose equivalent rate of natural background radiation. Arch Environ Health 40:181–184.

Wake N, Takagi N, Sasaki M. 1978. Androgenesis as a cause of hydatidiform mole. J Natl Cancer Inst 1978. 60:51–57. Wake N, Fujino T, Hoshi S, et al. 1987. The propensity to malignancy of dispermic heterozygous moles. Placenta 8:319–326. Wong SY, Ngan HY, Chan CC, Cheung AN. 1999. Apoptosis in gestational trophoblastic disease is correlated with clinical outcome and Bcl-2 expression but not Bax expression. Mod Pathol 12:1025–1033. Wright C, Nicholson S, Angus B, et al. 1992. Relationship between c-erbB-2 protein product expression and response to endocrine therapy in advanced breast cancer. Br J Cancer 65:118–121. Xiao ZX, Chen J, Levine AJ, et al. 1995. Interaction between the retinoblastoma protein and the oncoprotein MDM2. Nature 375:694–698. Yamashita K, Ishikawa M, Shimizu T, Kunda M. 1981. HLA antigens in husband-wife pairs with trophoblastic tumor. Gynecol Oncol 12:68– 74. Yen S, MacMahon B. 1968. Epidemiologic features of trophoblastic disease. Am J Obstet Gynecol 101:126–132. Yuen BH, Cannon W. 1981. Molar pregnancy in British Columbia: estimated incidence and postevacuation regression patterns of the beta subunit of human chorionic gonadotropin. Am J Obstet Gynecol 139:316–319.

57

Renal Cancer JOSEPH K. MCLAUGHLIN, LOREN LIPWORTH, ROBERT E. TARONE, AND WILLIAM J. BLOT

M

alignant tumors of the kidney account for about 2% of all new cancer cases in the United States and worldwide (Parkin et al., 2002), with 38,890 cases and 12,840 deaths estimated for 2006 in the United States (Jemal et al., 2006). Data from the United States Surveillance, Epidemiology and End Results (SEER) program for renal and urinary tract cancers other than bladder diagnosed from 1992 through 2000 show that renal parenchyma (renal cell) cancer accounts for 80% of the total, renal pelvis cancer 11%, ureter cancer 6%, urethra cancer 2%, and other sites about 1% (National Cancer Institute, 2003). Nearly all renal cell cancers are adenocarcinomas, whereas the vast majority of cancers of the renal pelvis, ureter, and urethra are transitional cell carcinomas. Wilms’ tumor (nephroblastoma), an embryonal neoplasm, is reviewed in the chapter on childhood cancer. In this chapter, information on renal cell cancer will be presented separately, whenever possible, from data on renal pelvis and ureter cancers, since they generally reflect different risk factor patterns. In many descriptive and some analytical studies, however, the available data pertain to all kidney cancers combined and do not permit further analyses by site.

DEMOGRAPHIC PATTERNS Incidence Patterns Renal cell cancer occurs about twice as often among men as among women (Parkin et al., 2002), with the excess most notable after age 50 years. The average age at diagnosis is in the early 60s, with incidence rates plateauing around age 70 (Figure 57–1). Since 1973, incidence rates for renal cell cancer have been rising rapidly each year in Europe and in the United States, with a few exceptions, mostly in Scandinavian countries (Mathew et al., 2002; National Cancer Institute, 2003) (Figure 57–2). This increase cannot be entirely accounted for by the more frequent detection of asymptomatic tumors through modern imaging procedures such as ultrasonography, since an increasing incidence of late-stage renal cell cancers has also been observed (Kosary and McLaughlin, 1993; Chow et al., 1999; Mathew et al., 2002). Increases in incidence have been more rapid among females than males (Katz et al., 1994; Chow et al., 1999; Tate et al., 2003) and among blacks than whites, resulting in a shift in excess from among whites to among blacks (Kosary and McLaughlin, 1993). This shift is becoming more pronounced, with rates among black men increasing by 3.9% and black women by 4.3% per year, compared with 2.3% for white men and 3.1% for white women (Chow et al., 1999). The recent excess among blacks is less apparent at older ages compared with younger ages (Figure 57–1). Based on data from the SEER program (National Cancer Institute, 2003), age-adjusted incidence rates of renal cell carcinoma among white men, white women, black men and black women during 1992–2000 were 13.4, 6.4, 15.9 and 7.7 per 100,000 person-years, respectively (Table 57–1). Rates among Asians of both sexes were about half those of other racial groups. For cancers of the renal pelvis and ureter, temporal trends in incidence rates have been inconsistent across regions (Mathew et al., 2002). In the United States, there has been a slight decrease in incidence rates among white men since the 1970s, while rates among white women and blacks have remained relatively stable (Figure 57–2)

(Chow et al., 1999; Mathew et al., 2002). Based on data from the SEER program, age-adjusted incidence rates among white men for cancers of the renal pelvis and ureter (2.5 per 100,000 person-years) are about two times higher than those among white women and black men and women (Table 57–1). There is a strong tendency for these patients to develop multiple transitional epithelial tumors, particularly in the urinary bladder, renal pelvis and ureter. For example, in the SEER catchment areas there were 2,837 people who developed multiple malignant tumors of the urinary bladder and of the renal pelvis or ureter, compared with only 855 people who developed multiple cancers of the urinary bladder and of the renal parenchyma. These data confirm the results of several reports (Kantor et al., 1986; McCredie et al., 1996; Czene and Hemminki, 2003) which found significantly elevated relative risks for a second cancer among patients with a first primary tumor of the renal pelvis or ureter, which was accounted for mainly by an increase in risk for bladder cancer. This pattern is due partly to shared risk factors such as smoking (Kantor and McLaughlin, 1985; Kantor et al., 1986) and to increased surveillance, but suggests the need for continuing to screen patients with renal pelvis and ureter cancers (e.g., urinary cytology) for new tumors arising along the urinary tract, including the bladder.

International Variation Table 57–2 presents incidence rates for renal cell and renal pelvis/ureter cancers from selected cancer registries around the world, as reported in Volume 8 of Cancer Incidence in Five Continents (Parkin et al., 2002). Rates of renal cell cancer vary internationally more than 10-fold. Incidence is generally highest in several Western and Eastern European countries, as well as in parts of Italy, in North America, and in Australia and New Zealand. The high rate among Israeli Jews is due to immigrants from Europe and North America. The lowest rates are reported in Asia and Africa. Incidence rates of renal pelvis and ureter cancers are generally less than one per 100,000 person-years, although they appear somewhat higher in Scandinavia, parts of Australia, and Taiwan (Table 57–2). In certain rural parts of Bulgaria, Yugoslavia, and Romania, rates for renal pelvis and ureter cancers are exceptionally high because of a predisposing condition called Balkan nephropathy, which is endemic in these areas (Stoyanov et al., 1978). In general, the descriptive patterns for cancers of the renal pelvis and ureter resemble those of bladder cancer more than renal cell cancer (Devesa et al., 1990; National Cancer Institute, 2003).

Mortality and Survival in the United States The prognosis of patients diagnosed with renal cell cancer has improved over time, with 5-year relative survival rates increasing from 36% to 39% in the early 1960s to between 50% and 60% by the 1990s (Kosary and McLaughlin, 1993; Chow et al., 1999). Table 57–3 presents 5-year relative survival rates for cancers diagnosed from 1992 through 2000 by site, race, and sex. Survival rates are generally better for renal cell cancer than for cancers of the renal pelvis and ureter among women; for men, the rates are similar. Blacks tend to have slightly poorer survival rates for renal cell cancer, but not for renal pelvis and ureter cancers. Although 5-year relative survival rates have

1087

1088

PART IV: CANCER BY TISSUE OF ORIGIN Table 57–1. Incidence Rates* for Cancers of the Renal Parenchyma and Renal Pelvis and Ureter, by Racial/Ethnic Group and Sex, SEER Program, 1992–2000 Males

Figure 57–1. Age-specific incidence of renal cell cancer in the United States by race and sex, 1992–2000. (Based on SEER data for eleven geographic regions of the United States: Atlanta, Georgia; Connecticut; Detroit, Michigan; Hawaii; Iowa; Los Angeles County, California; New Mexico; San Francisco/Oakland, California; San Jose/Monterey, California; Seattle/Puget Sound, Washington; and Utah.)

slowly increased for renal cell cancer patients over time, this does not appear to have occurred for cancers of the renal pelvis and ureter (McLaughlin et al., 1983; Kosary and McLaughlin, 1993). Kidney cancer mortality rates cannot be partitioned by subsite (e.g., renal parenchyma versus renal pelvis and ureter). In the United States, kidney cancer mortality rates increased among both blacks and whites until the early 1990s, after which rates stabilized, with mortality among blacks remaining slightly lower than that among whites (Figure 57–3).

Demographic and Socioeconomic Factors Mortality and incidence rates for kidney cancer are generally higher in urban than rural areas in the United States, England and Wales, and Norway and Denmark (McLaughlin and Schuman, 1983; Parkin et al., 2002). The urban-rural differential is apparent primarily among men,

Figure 57–2. Trends in age-adjusted (2000 United States standard) incidence of renal cell cancer and renal pelvis and ureter cancers by race and sex, 1973–2000. (Based on SEER data for nine geographic regions of the United States: Atlanta, Georgia; Connecticut; Detroit, Michigan; Hawaii; Iowa; New Mexico; San Francisco/ Oakland, California; Seattle/ Puget Sound, Washington; and Utah.)

Females

Number

Rate

Number

Rate

Renal parenchymaa Whites Blacks Asians American Indians White (non-Hispanic) White (Hispanic)

14,370 1,709 931 153 12,830 1,518

13.37 15.89 7.36 14.21 13.43 12.74

8,539 1,113 475 107 7,549 976

6.41 7.70 3.11 7.49 6.39 6.60

Renal pelvis and ureter Whites Blacks Asians American Indians White (non-Hispanic) White (Hispanic)

2,568 154 197 9 2,392 175

2.53 1.56 1.75 0.86 2.60 1.93

1,776 118 141 6 1,650 124

1.26 0.87 1.02 0.45 1.29 0.97

Source: National Cancer Institute, 2003. *Rates per 100,000 person-years, age-adjusted using 2000 U.S. standard. a Transitional, papillary, and squamous cell cancers of the kidney (ICD 189.0) are grouped with renal pelvis tumors.

and probably reflects past history of cigarette smoking. The pattern is influenced also by the greater availability of medical care and diagnostic services in urban than rural areas. Kidney cancer mortality statistics from the United States and other countries have shown no clear relation with educational achievement (McLaughlin and Schuman, 1983). Similarly, case-control studies of renal cell cancer have generally not shown a clear association with social class variables such as education (Wynder et al., 1974; Armstrong et al., 1976; McLaughlin et al., 1984, 1995a; Goodman et al., 1986; McCredie et al., 1988; Talamini et al., 1990; Maclure and Willett, 1990; Kreiger et al., 1993), although studies in Oklahoma (Asal et al., 1988a) and Los Angeles (Yuan et al., 1998a) have reported inverse associations with socioeconomic status or education. Except for a study from Australia (McCredie and Stewart, 1993), investigations of renal pelvis and ureter cancers have also reported no major differences in education between patients and control subjects (Armstrong et al., 1976; McCredie et al., 1982, 1983a, 1983b; McLaughlin et al., 1983; Jensen et al., 1988; Ross et al., 1989; McLaughlin et al., 1992b). When available, kidney cancer data have shown little relation

Renal Cancer

1089

Table 57–2. Incidence Rates per 100,000 Person-Years for Renal Cell and Renal Pelvis and Ureter Cancer in Selected Countries Male

Female

Country

RCC

Pelvis/Ureter

RCC

Pelvis/Ureter

Czech Republic France, Bas-Rhin Italy, Florence Slovakia Poland, Warsaw U.S. SEER: black Italy, Venezia Germany, Saarland Israel, Jews Finland Poland, Lower Silesia U.S. SEER: white Canada Australia, New South Wales New Zealand Netherlands Norway Scotland Sweden Denmark UK, England Seoul, Korea Japan, Osaka China, Taiwan Colombia, Cali China, Shanghai Zimbabwe Uganda

20 15.6 13.7 12.5 12.2 12.1 11.7 11.4 11.2 11 10.1 9.6 9.5 8.7 8.6 8.2 8 7.9 7.8 7.4 6.7 4.5 4.4 3.6 2.9 2.6 1.1 1

1.1 1.1 1.3 0.9 0.4 0.7 0.9 1 0.9 0.9 0.1 1.2 0.9 1.5 0.5 1.5 1.8 1 1.2 2.4 0.7 1.1 1.5 1.9 0 0.4 0 0

10.2 7.3 6.6 5.8 5.9 6.4 5 6.2 6.1 6.2 5.7 4.9 5.2 4.6 4.5 4.6 4.6 4.3 4.8 4.1 3.3 1.9 1.5 2.6 2.3 1.5 1.6 1.1

0.8 0.3 0.4 0.6 0.1 0.4 0.2 0.2 0.3 0.4 0.1 0.6 0.4 1.5 0.3 0.6 0.7 0.5 0.7 1.2 0.3 0.4 0.5 2 0.2 0.3 0 0

Source: Parkin et al. 2002. *Age-adjusted, standarized to the world population.

to income level (McLaughlin and Schuman, 1983), although one correlation study reported a weakly positive association (Blot and Fraumeni, 1979).

RISK FACTORS FOR RENAL CELL CANCER A number of causes of the increasing rates of renal cell cancer have been established, in particular smoking and obesity, and possibly hypertension. Virtually all information on risk factors for renal cell cancer has come from a relatively large number of case-control studies (Table 57–4) and several recent cohort studies. These studies have been conducted in a number of countries, including the United States, Canada, England, Australia, Italy, Finland, France, Denmark, Sweden and China. The largest case-control study, based on 1,732 cases and 2,309 controls, was a multi-center investigation conducted in five countries using a common protocol, questionnaire and field procedures, and it represents the largest and most comprehensive dataset on risk factors for renal cell cancer to date (McLaughlin et al., 1995a, Table 57–3. Five-year Relative Survival Rates (percent) for Cancer of the Renal Parenchyma and Renal Pelvis and Ureter, by Race and Sex, SEER Program, 1992–2000 Males

Females

Renal parenchyma Whites Blacks

63 57

63 60

Renal pelvis/ureter Whites Blacks

59 59

50 53

Source: National Cancer Institute, 2003.

Figure 57–3. Trends in age-adjusted (2000 U.S. standard) mortality from kidney cancer by race and sex, 1973–2000. (Based on National Center for Health Statistics data for the entire United States.)

1995b; McCredie et al., 1995; Mellemgaard et al., 1995; Mandel et al., 1995; Lindblad et al., 1995; Schlehofer et al., 1996; Wolk et al., 1996a). Individual study centers participating in this international study have published their own center-specific results (McCredie and Stewart, 1992a, 1992b, 1993; Mellemgaard et al., 1994a, 1994b, 1994c, 1994d; Chow et al., 1994a, 1994b, 1996; Lindblad et al., 1994, 1997; Schlehofer et al., 1995; Boeing et al., 1997), but in this chapter only the combined multi-center results will be reported. Exceptions will be made when the center-specific analyses report on associations not evaluated in the combined multi-center analysis. To date, there have been virtually no analytic studies of renal cell cancer among blacks.

Cigarette Smoking Based on the collective evidence from both case-control and cohort studies, cigarette smoking is an established causal risk factor for renal cell cancer, and a dose-response relation has been observed among both men and women (Wynder et al., 1974; McLaughlin and Schuman, 1983; McLaughlin et al., 1984, 1990, 1992a, 1995a; Yu et al., 1986; Brownson, 1988; McCredie et al., 1988; La Vecchia et al., 1990; Maclure and Willett, 1990; Kreiger et al., 1993; Yuan et al., 1998a; Chow et al., 2000; Chiu et al., 2001; Semenza et al., 2001). The relative risks among smokers from case-control and cohort studies range from 1.2 to 2.3, although a small chart-review study in the 1960s reported a five-fold increase in risk (Bennington and Laubscher, 1968). Relative risks for heavy smokers have ranged from 2.0 to 2.5. Use of hospital controls or small sample size is the likely explanation for the absence of a statistically significant association between renal cell cancer and cigarette smoking in some case-control studies (Schwartz et al., 1961; Armstrong et al., 1976; Kolonel, 1976; Goodman et al., 1986; Asal et al., 1988a; Talamini et al., 1990; Benhamou et al., 1993; Siemiatycki et al., 1995). The moderate risk of renal cell cancer associated with cigarette smoking would be difficult to detect in a small study and could easily be obscured by the use of hospital control patients who may have an elevated prevalence of smoking. Renal cell cancer risk associated with cigarette smoking has been shown to decline significantly with years of cessation (McLaughlin et al., 1984, 1995a; LaVecchia et al., 1990; McCredie and Stewart, 1992a; Yuan et al., 1998a; Parker et al., 2003), with a 15% to 30% reduction in risk after 10 to 15 years of quitting. Population-based attributable risks indicate that approximately 20% to 30% of renal cell cancers among men and 10% to 20% among women can be accounted for by cigarette smoking (McLaughlin et al., 1995a; Yuan et al., 1998a; Benichou et al., 1998). One study has reported a suggestive associa-

1090

PART IV: CANCER BY TISSUE OF ORIGIN Table 57–4. Published Case-Control Studies of Risk Factors for Renal Cell Cancer Authors

Year

Source of Controls

Number of Cases/Controls

Location

Bennington Wynder et al. Armstrong et al. Kolonel McLaughlin et al. Yu et al. Goodman et al. Brownson Asal et al. McCredie et al. Sharpe et al. La Vecchia et al. Maclure Talamini et al. Partanen et al. McLaughlin et al. Kreiger et al. Finkle et al. Benhamou et al. Hiatt et al. Weinmann et al. Muscat et al. McLaughlin et al.* Yuan et al. Bianchi et al. Pesch et al. Semenza et al. Hu et al.

1968 1974 1976 1976 1984 1986 1986 1988 1988 1988 1989 1990 1990 1990 1991 1992a 1993 1993 1993 1994 1994 1995 1995 1998 2000 2000 2001 2003

Hospital Hospital Hospital Hospital Population Neighborhood Hospital Cancer registry Hospital/population Population Urologic patients Hospital Population Hospital Population Population Population Medical plan Hospital Medical plan Population Hospital Population Population Population Population Population Population

100/190 202/394 106/106 64/197 495/697 160/160 267/267 326/978 315/313/336 360/985 164/161 131/394 410/605 240/665 338/484 154/157 518/1381 191/191 196/347 257/257 206/292 788/779 1732/2309 1204/1204 406/2434 935/4298 115/259 1279/5370

Washington 3 U.S. cities England New York Minneapolis Los Angeles 6 U.S. cities Missouri Oklahoma New South Wales Montreal Northern Italy Boston Northern Italy Finland Shanghai Ontario Los Angeles France San Francisco Oregon/Washington Six U.S. cities Five countries Los Angeles Iowa Germany California Canada

*Includes data from McCredie and Stewart (1992a, 1992b), Mellemgaard et al. (1994a, 1994b, 1994c, 1994d), Chow et al. (1994a, 1994b, 1996), Schlehofer et al. (1995), Lindblad et al. (1994, 1997) and Boeing et al. (1997).

tion of passive smoking with renal cell cancer (Kreiger et al., 1993), although given the modest magnitude of the risk associated with active cigarette smoking, it is unlikely that a study could accurately detect such an association. An association with cigars, pipes and chewing tobacco has been inconsistently reported (McLaughlin et al., 1984, 1995a; Yu et al., 1986; Brownson, 1988; Yuan et al., 1998a). A recent study evaluating the interaction between cigarette smoking and N-acetyltransferase 2 (NAT2) genotype suggested that the increased risk for renal cell cancer associated with smoking, particularly among women, was more pronounced among slow acetylators, who had a threefold increased risk from smoking compared with a non-significant 40 percent increased risk among rapid acetylators (Semenza et al., 2001). Although it appears that NAT2 may play an etiologic role in renal cell cancer development similar to that seen for several other cancers, whether or not the effect of smoking is modified by NAT2 genotype needs to be confirmed in future studies.

Obesity A positive association between body weight and renal cell cancer has been consistently reported in virtually every case-control and cohort study that has examined this relation. A few early studies noted the association primarily in women (Whisenand et al., 1962; Wynder et al., 1974; Lew and Garfinkel, 1979; McLaughlin et al., 1984), but most studies have found an effect in both sexes (Whittemore et al., 1985; Yu et al., 1986; Goodman et al., 1986; Asal et al., 1988a; Maclure and Willett, 1990; Mellemgaard et al., 1991, 1995; McLaughlin et al., 1992a; Kreiger et al., 1993; Benhamou et al., 1993; Wolk et al., 1996a; Chow et al., 2000; Heath et al., 1997; Yuan et al., 1998b). A recent quantitative review of the published literature showed that the association between obesity, assessed as increased body mass index, and risk of renal cell cancer is equally strong among men and women, with a summary relative risk of 1.07 per unit increase of body mass index (Bergstrom et al., 2001). There is some evidence that weight cycling or large swings in weight (Lindblad et al., 1994; Mellemgaard et al., 1995), including weight gain during adulthood (McLaughlin et al.,

1984; Prineas et al., 1997), may play a role. The rising prevalence of obesity in the United States during the past few decades may partly explain the increasing incidence of renal cell cancer; in fact, it has recently been estimated that up to 30% of renal cell cancer cases among American men and women can be attributed to overweight and obesity (Bergstrom et al., 2001; Calle and Kaaks, 2004). The mechanism responsible for the effect of obesity remains speculative. Obesity and hypertension are believed to be independent, major risk factors for renal cell cancer in both men and women. It has recently been hypothesized that lipid peroxidation, which is increased among obese and hypertensive subjects, may be partly responsible for these associations (Gago-Dominguez et al., 2002). It has been suggested that obesity may act by promoting hormonal changes (Calle and Kaaks, 2004), such as increased levels of endogenous estrogens, but the existing epidemiological evidence does not convincingly link hormone-associated variables to renal cell cancer. Increasing body mass index is accompanied in both men and women by elevated levels of insulin-like growth factor-I (IGF-I), which could contribute to the development of renal cell cancer (Kellerer et al., 1995; Calle and Kaaks, 2004). Obesity may also predispose to higher glomerular filtration rate and renal plasma flow, independent of hypertension, as well as arterionephrosclerosis, which may, in turn, render the kidney more susceptible to carcinogenesis.

Hypertension and Antihypertensive Drugs Hypertension and its treatment with antihypertensive medications are obviously highly correlated, and disentangling the influence on renal cell cancer risk of one versus the other is difficult. Further, early-stage prediagnostic renal tumors may themselves lead to increases in blood pressure. However, the cumulative evidence suggests that hypertension itself, rather than diuretics or other antihypertensive drugs, may play a role in the etiology of renal cell cancer. In fact, contrary to earlier reports, a recent study (Hajjar and Kotchen, 2003) reported an increasing prevalence of hypertension in the United States, particularly among blacks and among women, while hypertension control rates continue to

Renal Cancer be low; these data suggest that hypertension may explain in part the recent trends of increasing renal cell cancer incidence. Most epidemiologic studies (Raynor et al., 1981; Buck and Donner, 1987; Grove et al., 1991; Hole et al., 1993; McLaughlin et al., 1995b; Coughlin et al., 1997; Heath et al., 1997; Yuan et al., 1998b; Shapiro et al., 1999; Chow et al., 2000; Semenza et al., 2001), but not all (Prineas et al., 1997), have reported relative risks for renal cell cancer associated with hypertension ranging between 1.3 and 2 or greater. In the large international study (McLauglin et al., 1995b), the relative risk for hypertension after adjustment for diuretics and other antihypertensive medications was 1.4, although there was little excess risk associated with hypertension among non-users of any antihypertensive medications (a subset of patients likely to have milder forms of hypertension). A large-scale study in Los Angeles reported a two-fold risk for hypertension independent of medication (Yuan et al., 1998b). In a recent Swedish cohort study (Chow et al., 2000) which did not adjust for use of antihypertensive drugs, dose-response relations were observed between both diastolic and systolic blood pressure and renal cell cancer, independent of the effect of body mass index and after exclusion of the first five years of follow-up, when preclinical disease or undetected renal cell cancers may cause increases in blood pressure. The results of this study also suggested that a reduction in blood pressure over time may decrease renal cell cancer risk independent of the effect of hypertension at baseline. Another large cohort study using baseline screening information from a cardiovascular intervention trial also demonstrated an increased risk of kidney cancer after exclusion of the first five years of follow-up (Coughlin et al., 1997). However, in a recent cohort mortality study with a large number of deaths from kidney cancer, exclusion of the first 10 years of follow-up eliminated any association between hypertension and kidney cancer (Batty et al., 2003), suggesting that the presence of early-stage renal tumors may in fact raise blood pressure, rather than the reverse. One study (Semenza et al., 2001) has reported an interaction between high blood pressure and NAT2 genotype, with risk most pronounced among women who are slow acetylators, but this association may be explained by slow metabolism, and therefore increased toxicity, of antihypertensive drugs among slow acetylators. The mechanism by which high blood pressure may affect renal cell cancer risk is unclear, but hypertension-induced renal injury may play a role, or hypertension may be associated with metabolic or functional changes within the renal tubule that increase susceptibility to carcinogens. An alternative hypothesis is that renal cell cancer is more likely to be diagnosed incidentally or at early stages among those being treated for hypertension, perhaps due to enhanced medical surveillance, although a recent analysis addressing this issue suggests that such is not the case (Rosenberg et al., 1998b). Diuretics are commonly used, particularly among the elderly, and an association with renal cell cancer, if causal, would therefore have major public health implications. Animal studies have linked hydrochlorothiazide and furosemide, the most commonly used diuretics, with tubular cell adenomas and adenocarcinomas of the kidney in rats and hepatocellular tumors in mice (Lijinsky and Reuber, 1987; National Toxicology Program, 1989a, 1989b). Moreover, these compounds act on the renal tubules (Laski, 1986), the site of origin for renal cell cancers. Early epidemiologic studies suggested that diuretic use was associated with a large increase in the risk of renal cell cancer, even after adjustment for hypertension, with relative risks as high as 4.0 reported (Yu et al., 1986; McLaughlin et al., 1988; Fraser et al., 1990; Grove et al., 1991; Mellemgaard et al., 1992b; Lindblad et al., 1993; Krieger et al., 1993; Finkle et al., 1993; Weinmann et al., 1994; Hiatt et al., 1994; Heath et al., 1997). However, more recent studies provide a less clear picture, and adjustment for high blood pressure appears to eliminate the excess risk associated with diuretic use (McLaughlin et al., 1995b; Yuan et al., 1998b; Shapiro et al., 1999). With respect to non-diuretic antihypertensive medications, relative risks have varied greatly, with no particular class of drugs being consistently associated with increased risk. Since diuretics and other antihypertensive drugs have different structures and mechanisms of action, this lack of specificity suggests that these medications are unlikely to play an etiologic role in renal cell cancer development.

1091

Analgesics Historically, heavy use of phenacetin-containing drugs has been causally linked to transitional cell tumors of the renal pelvis (International Agency for Research on Cancer [IARC], 1987). An association, although much lower in magnitude and less conclusive, has been reported also for renal cell cancer (McLaughlin et al., 1984, 1985, 1992a; Maclure and MacMahon, 1985; McCredie et al., 1988, 1995; Kreiger et al., 1993), even after adjustment for confounding by cigarette smoking and use of other types of analgesics. Whether this association is causal is difficult to assess, because phenacetin-containing analgesics have been off the market for at least 25 years in most countries, and accurate recollection at interview of their use is problematic because reliable recall of distant over-the-counter analgesic intake is no longer practical. A causal link between other types of analgesics, such as acetaminophen or aspirin, and renal cancer has not been established. To date, virtually all of the epidemiologic data regarding analgesic use and renal cell cancer are derived from case-control studies, most of them conducted in North America, Europe or Australia (McLaughlin et al., 1985; McCredie et al., 1988, 1995; Derby and Jick, 1996; Rosenberg et al., 1998a; Gago-Dominguez et al., 1999a). The majority have primarily focused on acetaminophen, the major metabolite of phenacetin (Hinsen, 1981), as the exposure of interest, although several have also presented data on aspirin use. Aspirin use, despite being extensively investigated, has been associated with an increased risk of renal cell cancer in only one case-control study (Gago-Dominguez et al., 1999a) and one cohort study (Paganini-Hill et al., 1989), and is unlikely related to risk. For acetaminophen, there is little persuasive evidence of an association. The large international population-based casecontrol study reported no link with either acetaminophen or aspirin use and failed to confirm the relationship of phenacetin to renal cell cancer (McCredie et al., 1995). Moreover, no consistently increasing risks were observed with increasing consumption levels. The results of a large study in the United States designed to evaluate this association failed to find any link with regular use or duration of use of acetaminophen (Rosenberg et al., 1998a), as did a recent cohort investigation (Heath et al., 1997). A few studies have suggested an elevated renal cell cancer risk associated with acetaminophen (McCredie et al., 1993; Derby and Jick, 1996; Gago-Dominguez et al., 1999a; Kaye et al., 2001), but these case-control findings are likely explained by confounding by prior phenacetin use or protopathic bias (Salas et al., 1999; Signorello et al., 2002). Linked-registry studies of patients with diseases that require treatment with analgesics have reported increased risks of renal cell cancer, although the type of analgesic was unknown and control of confounding factors was not possible (Mellemgaard et al., 1992a; Gridley et al., 1993; Lindblad et al., 1993). The findings of many published studies are difficult to interpret in light of serious methodological flaws in study design, execution and analysis. In particular, biases in control selection and assessment of analgesic use, as well as the failure to control for mutual confounding among various analgesics, are likely to generate either artificially elevated or underestimated effect estimates. Further, protopathic bias, i.e., the use of over-the-counter (OTC) analgesics to treat the pain and discomfort sometimes associated with this cancer prediagnostically, has been largely ignored in interpreting the small, inconsistent elevations in risk reported in some studies. Thus, the collective epidemiologic data published to date fail to establish a credible association between use of acetaminophen and renal cell cancer. In light of the vast international market for both prescription and OTC analgesics of various classes, as well as the widespread public acceptance of their safety, comprehensive assessment through well-conducted analytic epidemiologic investigations of these medications for a possible link with renal cancer is crucial.

Diet Early correlation studies suggested an association between kidney cancer mortality and per capita consumption of fat and protein

1092

PART IV: CANCER BY TISSUE OF ORIGIN

(Wynder et al., 1974; Armstrong and Doll, 1975). Analytic studies of renal cell cancer, however, have failed to support ecologic results and have found relatively few consistent associations with dietary factors (Armstrong et al., 1976; McLaughlin et al., 1984; Yu et al., 1986; McCredie et al., 1988; Maclure and Willett, 1990; Talamini et al., 1990; Kreiger et al., 1993; Wolk et al., 1996a, 1996b; De Stefani et al., 1998). Elevated risks have been reported for consumption of meat (McLaughlin et al., 1984, 1992a; Maclure and Willett, 1990; Kreiger et al., 1993; Boeing et al., 1997; De Stefani et al., 1998; Handa and Kreiger, 2002; Hu et al., 2003), milk (McCredie et al., 1988; Talamini et al., 1990), and margarine, oils and other fat types (Talamini et al., 1990; Boeing et al., 1997; Handa and Kreiger, 2002); however, most of these findings were not adjusted for confounding by energy intake. There has been some suggestion that elevated protein consumption may be a risk factor (Maclure and Willett, 1990; Chow et al., 1994b; Handa and Krieger, 2002). There may be some biologic plausibility to a high protein diet affecting risk of renal cell cancer, because animal studies have shown protein intake can induce renal tubular hypertrophy (Smith et al., 1993), but the largest epidemiologic study to date failed to provide clear support for this hypothesis (Wolk et al., 1996a). An important problem is the high colinearity among total calories, protein calories and fat calories, and even in large studies the ability to disentangle the effects of calorie source is limited. In the international study, after adjustment for energy, positive associations with milk and foods rich in animal protein became non-significant, while the protective effect of fruit and vegetables became more pronounced (Wolk et al., 1996a). A protective effect of fruit and vegetable consumption has been one of the few relatively consistent dietary findings (Maclure and Willett, 1990; Talamini et al., 1990; McLaughlin et al., 1992a; Wolk et al., 1996b; Yuan et al., 1998c; Hu et al., 2003; Rashidkhani et al., 2005). Sporadic inverse associations have been reported with respect to vitamin C (Boeing et al., 1997), Vitamin E (Wolk et al., 1996a; Hu et al., 2003), carotenoids (Yuan et al., 1998c) and calcium (Prineas et al., 1997; Hu et al., 2003), but no particular micronutrient or vitamin has been consistently observed to decrease or increase the risk of renal cell cancer in case-control or cohort studies. Finally, an association has been seen for “doneness” of red meats and risk of renal cell cancer (Wolk et al., 1996a). The method of cooking meat is linked with the formation of heterocyclic amines, which have been observed in animal experiments to cause cancer. The intake of heterocyclic amines within the usual dietary range has not been consistently linked to renal cell cancer in case-control studies (De Stefani et al., 1998; Augustsson et al., 1999). Thus, with the exception of a protective effect of fruit and vegetable consumption, dietary factors, once thought to play an important etiologic role in renal cell cancer, appear not to substantially affect risk, except as it relates to obesity.

reported in a few studies of renal cell cancer, particularly among women (McLaughlin et al., 1984; Goodman et al., 1986; Asal et al., 1988a; Kinlen et al., 1988; Kreiger et al., 1993), most studies have reported either no association (Armstrong et al., 1976; McCredie et al., 1988; McClure and Willett, 1990; Talamini et al., 1990; McLaughlin et al., 1992a; La Vecchia et al., 1992; Wolk et al., 1996a; Bianchi et al., 2000) or a weak protective effect of tea consumption (Yu et al., 1986; Benhamou et al., 1993; Zheng et al., 1996). With respect to alcohol consumption, most case-control and cohort studies show no association with renal cell cancer (McLaughlin et al., 1984, 1992a; Yu et al., 1986; Brownson, 1988; Maclure and Willett, 1990; Benhamou et al., 1993; Kreiger et al., 1993), and cohort studies of alcoholics and brewery workers have reported no excess mortality from kidney cancer (Schmidt and De Lint, 1972; Pell and Alonzo, 1973; Monson and Lyon, 1975; Jensen, 1979; Schmidt and Popham, 1981; Adami et al., 1992). However, there is some indication of an inverse association between alcohol consumption and renal cell cancer among women but not among men (Goodman et al., 1986; Talamini et al., 1990; Wolk et al., 1996a; Parker et al., 2002).

Coffee, Tea, and Alcohol

Renal cell cancer is not generally considered an occupationally associated tumor. However, epidemiologic studies have reported a number of associations between exposures or jobs/industries and renal cell cancer. Asbestos has been linked to kidney cancer in two cohort studies, one of insulators (Selikoff et al., 1979) and one of asbestos products workers (Enterline et al., 1987), which reported significantly elevated mortality rates for kidney cancer. Several case-control studies of renal cell cancer that have examined occupational risk factors have found no association with asbestos exposure (McLaughlin et al., 1984; Yu et al., 1986; Goodman et al., Asal et al., 1988a,b; Brownson, 1988; Partanen et al., 1991), although their power to detect risks for asbestos exposure is generally low because of the small number of exposed workers. A positive association between self-reported asbestos exposure and renal cell cancer has been reported in several other case-control studies (Maclure, 1987; Mandel et al., 1995; Pesch et al., 2000; Mattioli et al., 2002), including the large international study, which included 200 exposed cases and found a moderate relative risk of 1.4 with selfreported asbestos exposure, but data on duration of exposure did not support an association (Mandel et al., 1995). Asbestos fibers have been found in the kidneys of individuals with high exposures (Huang et al., 1988), but an association of kidney cancer with either duration of employment or level of asbestos exposure needs to be demonstrated before causal inference can be drawn. A meta-analysis of occupational

Early correlation studies suggested a relation between kidney cancer and per capita consumption of coffee and alcohol (Shennan, 1973; Breslow and Enstrom, 1974; Armstrong and Doll, 1975; Hinds et al., 1980). However, these ecologic findings have generally not been confirmed by analytical studies of renal cell cancer, after adjustment for the confounding effect of cigarette use (Wynder et al., 1974; Armstrong et al., 1976; McLaughlin et al., 1984; Goodman et al., 1986; Yu et al., 1986; Brownson, 1988; Asal et al., 1988a; McCredie et al., 1988; Maclure and Willett, 1990; Talamini et al., 1990; Partanen et al., 1991; Kreiger et al., 1993; Benhamou et al., 1993; Stensvold and Jacobsen, 1994). In two studies (Yu et al., 1986; Wolk et al., 1996a), a two-fold increased risk for drinking more than 5 to 6 cups of coffee per day was observed among women only, with no dose-response relation. Results from one cohort study in Norway, an area of heavy coffee intake, showed a significant inverse trend, with consumers of seven or more cups having one fourth the risk of those drinking two or fewer cups daily (Jacobsen et al., 1986), but this finding was not supported by that of a larger, more recent Norwegian cohort study (Stensvold and Jacobsen, 1994). Overall, the results from analytical studies do not indicate that coffee consumption increases the risk of renal cell cancer. Although an increased risk among tea drinkers has been

Hormonal and Reproductive Factors Although estrogens have induced renal cell carcinomas in laboratory animals, particularly Syrian golden hamsters, there is little epidemiologic evidence supporting an association in humans (McLaughlin and Schuman, 1984; Newsom and Vurgin, 1987). Weak positive findings have been reported in some studies for menopausal estrogen use (Asal et al., 1988b; McLaughlin et al., 1992a) and oral contraceptive use (McLaughlin et al., 1992a; Kreiger et al., 1993). However, in the large international case-control study there was a significantly reduced risk associated with use of oral contraceptives, restricted to women who did not smoke (Lindblad et al., 1995), while another recent study found no such relationship (Gago-Dominguez et al., 1999b). Similarly, the association between prior hysterectomy or oophorectomy and the risk of renal cell cancer has been inconsistently reported (Wynder et al., 1974; McLaughlin et al., 1984; Lindblad et al., 1995; GagoDominguez et al., 1996b). Some investigators, but not all, have reported an almost two-fold higher risk of renal cell cancer among women with high parity compared with nulliparous women (Lindblad et al., 1995; Lambe et al., 2002), after adjustment for obesity, and an inverse association between age at first birth and risk of renal cell cancer (Krieger et al., 1993; Mellenmgaard et al., 1994c; Lambe et al., 2002). Age at menarche and at menopause do not appear to appreciably affect risk (Lindblad et al., 1995).

Occupation

Renal Cancer studies of asbestos-exposed worker cohorts showed little relationship to increased risks for kidney cancer (Sali and Boffetta, 2000). Coke-oven workers exposed to high levels of polycyclic aromatic hydrocarbons were reported to be at increased risk for kidney cancer (Redmond et al., 1972; Redmond, 1983), but a 30-year follow-up study of these same workers failed to find an increased risk for kidney cancer mortality (Costantino et al., 1995). The large international casecontrol study did observe a significant association with self-reported employment in the blast furnace/coke oven industry, although again there was no trend of increasing risk with increasing duration of employment (Mandel et al., 1995). A recent study (Dosemeci et al., 1999) has reported effect modification by gender of the associations between chlorinated aliphatic hydrocarbons and organic solvents and renal cell cancer, but these observations cannot be meaningfully interpreted until confirmed in future studies. Reviews of cohort studies of oil refinery workers have failed to confirm suspicions of increased risk of kidney cancer (Wong and Raabe, 1989; IARC, 1989). In the early 1980s, gasoline was suspected as a risk factor for renal cell cancer when male rats exposed long-term to vapors of unleaded gasoline developed a significant excess of renal cancers (MacFarland et al., 1984). Since then, a number of epidemiologic studies have examined the effect of gasoline exposure, and the collective evidence to date does not support a relationship between gasoline and risk of renal cell cancer (McLaughlin, 1993; Lynge et al., 1997). In the international case-control study, exposure to gasoline was reported significantly more often by cases than controls, but there was no association with duration of exposure or with employment as a gasoline station attendant. No association with gasoline was observed in numerous cohort and nested case-control studies conducted in different populations (Domiano et al., 1985; Yu et al., 1986; Poole et al., 1993; Wong et al., 1993, 1999; Rushton, 1993; Schnatter et al., 1993; McCredie and Stewart, 1993; Gamble et al., 1996), and a recent mortality and cancer morbidity study among a cohort of Canadian petroleum workers did not find an excess of kidney cancer (Lewis et al., 2003). Exposure to aviation and jet fuels was related to kidney cancer risk in a case-control screening study of occupational exposures and cancer (Siemiatycki et al., 1987), but a later cohort study of aviation maintenance workers exposed to a variety of chemicals including jet fuel showed no association with risk (Spirtas et al., 1991). The international study found a significant, dose-related association between exposure to petroleum products other than gasoline and renal cancer, but specific exposure to jet fuel, heating oil and kerosene or diesel fuel could not be distinguished. Seven of the eight self-reported exposures analyzed in the international study were associated with an elevated risk ratio, suggesting response bias (Mandel et al., 1995). Considerable interest has recently been focused on the solvent trichloroethylene (TCE), largely as a result of bioassay findings in animals and of three studies conducted in the same area of Germany, which were initiated in response to clusters of renal cell cancer cases and which reported strikingly elevated relative risks for renal cell cancer associated with TCE exposure (Henschler et al., 1995; Vamvakas et al., 1998; Bruning et al., 2003). These findings contrast starkly with results from other investigations, and several serious methodological shortcomings of these studies have been noted (McLaughlin and Blot, 1997; Green and Lash, 1999; Cherrie et al., 2001; Mandel, 2001), limiting the conclusions that can be drawn. To date, seven cohort studies have evaluated the relationship between TCE and specific types of cancer (Garabrant et al., 1988; Axelson et al., 1994; Anttila et al., 1995; Blair et al., 1998; Morgan et al., 1998; Boice et al., 1999; Hansen et al., 2001; Raaschou-Nielsen et al., 2003). The two largest studies both employed sophisticated methods of exposure assessment and both internal and external comparisons (Blair et al., 1998; Boice et al., 1999). None of these studies reported a significantly increased risk of renal cell cancer among TCE exposed workers. The most recent cohort study (Raaschou-Nielsen et al., 2003), conducted in Denmark, evaluated cancer morbidity among 40,049 workers with exposure to TCE and showed little indication of a causal relation between renal cell cancer and TCE. Thus, the weight of the evidence to date provides little credible support for the hypothesis that TCE causes renal cell cancer in humans.

1093

A related solvent, perchloroethylene (PCE), has also been suggested to increase kidney cancer risk, mainly as a result of early proportional mortality studies which suggested that laundry and dry cleaning workers may be at increased risk for kidney cancer (Blair et al., 1979; Katz and Jowett, 1981; Duh and Asal, 1984; Brown and Kaplan, 1987). These studies are, however, hypothesis-generating in nature and cannot provide reliable evidence of a causal relationship. Only two cohort studies have directly evaluated occupational exposure to PCE in relation to kidney cancer (Anttila et al., 1995; Boice et al., 1999), with nonsignificant risk estimates of 1.82 and 0.69, respectively. A recent review concluded that epidemiologic evidence for a relationship between PCE and kidney cancer was inadequate (Mundt et al., 2003). Similarly, no statistically significantly increased risk for kidney cancer was observed in studies conducted among dry-cleaning employees presumably exposed to PCE (McLaughlin et al., 1987; Ruder et al., 2001; Blair et al., 2002), although individual exposure was not quantified. In the large international study (Mandel et al., 1995), self-reported exposure to dry cleaning fluids (specific type of solvent was unknown) was significantly associated with risk, but no relation was seen for years of exposure. Moreover, no association was apparent when actual employment in the dry cleaning industry was examined. Several other occupational associations have been sporadically reported, although none has been consistently linked with renal cell cancer and many have not shown dose-response relations. These include increased risks among workers exposed to cadmium (Kolonel, 1976; Mandel et al., 1995; Pesch et al., 2000), lead (Fu and Boffetta, 1995; Pesch et al., 2000) or polychlorinated biphenyls (Shalat et al., 1989), newspaper pressmen (Paganini-Hill et al., 1980), physicians (McLaughlin et al., 1987), truck drivers (Brownson, 1988), architects (Lowery et al., 1991), paperboard printing workers (Sinks et al., 1992), firefighters (Guidotti, 1995), pulp and paper mill workers (Band et al., 1997) and commercial airline pilots and navigators (Nicholas et al., 1998). A recent study evaluated the interaction between self-reported occupational exposures and genotype for glutathione S-transferase M1 and T1 (GSTM1 and GSTT1) in renal cell cancer (Buzio et al., 2003). Compared with GSTM1 null subjects, those with the GSTM1 polymorphism showed higher risks for renal cell cancer associated with exposure to metals or pesticides. Similarly, the relative risk for renal cell cancer among those exposed to solvents and pesticides was more pronounced among those with GSTT1 present genotype, compared with GSTT1 null subjects. It appears that GSTM1 and GSTT1 polymorphisms may modify the risk of renal cell cancer among subjects with certain occupational exposures, but this finding requires confirmation.

Radiation Ionizing radiation appears to increase the risk of renal cell cancer among patients treated for ankylosing spondylitis and cervical cancer, but the effects are weak and less evident than the radiogenic risks associated with renal pelvis cancer (Land, 1986; Boice et al., 1988). An increased risk has also been observed among patients receiving radium-224 for bone tuberculosis and ankylosing spondylitis (Spiess et al., 1989). In one case-control study, significantly more female patients than control subjects reported receiving diagnostic or therapeutic radiation (Asal et al., 1988b).

Genetic Susceptibility Renal cell cancer occurs in both sporadic and familial forms. Having a first degree relative with kidney cancer has been associated with a two- to four-fold increased risk in some studies (McLaughlin et al., 1984; Schlehofer et al., 1996; Gago-Dominguez et al., 2001) but not all (Kreiger et al., 1993). Gago-Dominguez et al. (2001) also evaluated whether a positive family history of kidney cancer modifies any risk factor associations and found no differences between renal cell cancer cases with and without a family history. There are believed to be three major inherited forms of renal cell cancer—von Hippel-Lindau-associated, familial clear cell, and

1094

PART IV: CANCER BY TISSUE OF ORIGIN

hereditary papillary renal cell carcinoma (HPRC)—which together account for a small proportion of the occurrence of this malignancy (Linehan et al., 1995; Zbar and Lerman, 1998; Zbar, 2000). All three familial syndromes are dominantly inherited and tend to arise at a substantially younger age than sporadic cases. Von-Hippel Lindau (VHL) disease, the most well-described form of hereditary renal carcinoma, is inherited as a highly penetrant autosomal dominant trait and is characterized by tumors at many sites, including the kidney, central nervous system, and pancreas (Linehan and Klausner, 1998; Lonser et al., 2003). Renal cell cancer accounts for about 50% of deaths among VHL patients, who tend to have multiple bilateral kidney tumors mainly of the clear cell type (Neumann and Zbar, 1997). In contrast to the male predominance in sporadic renal cell cancer, men and women are equally affected by VHL-associated kidney tumors (Neumann and Zbar, 1997). The syndrome is associated with germline mutations of the tumor suppressor VHL gene located on chromosome 3 (Latif et al., 1993; Lonser et al., 2003). In VHL-associated renal tumors, both the inherited and the wild type allele of the VHL gene at 3p 25–26 are inactivated in a manner consistent with tumor suppressor function (Linehan et al., 1995; Linehan, 1998). Familial occurrences of clear cell renal cancer have been reported also in the absence of manifestation of the VHL syndrome, and have featured germline alterations involving a balanced translocation affecting the short arm of chromosome 3 (Cohen et al., 1979). It is noteworthy that loss of sequences of the short arm of chromosome 3 are observed in 85% of nonpapillary sporadic tumors of the kidneys (Bodmer et al., 2002), suggesting common genetic mechanisms for hereditary and sporadic forms of renal cell cancer and lending support to the tumor suppressor gene model (Gnarra et al., 1995). The third type of inherited renal cancer is HPRC 1, which is histologically and genetically distinct from VHL disease and which has been linked to activating germ line mutations of the MET protooncogene located on chromosome 7q (Zbar et al., 1995; Schmidt et al., 1999). Somatic mutations of this gene have been described in some sporadic cases of clear cell and papillary tumors of the kidney (Zbar and Lerman, 1998). Affected individuals are at risk to develop bilateral multifocal papillary renal cell carcinoma (Linehan, 1998). In recent years, several families have been observed with another hereditary syndrome, termed HPRC 2, which is histologically distinct from HPRC 1 and does not involve germline mutations in the MET protooncogene (Zbar, 2000). While many advances have been made in identifying rare high-penetrant genes for hereditary syndromes of renal cancer, little work has been carried out to detect common lowpenetrant genes that confer susceptibility to the more prevalent sporadic tumors. Developmental defects also play a role in some renal cell cancers as indicated by reported associations with polycystic kidneys (McFarland et al., 1972), horseshoe kidneys (Blackard and Mellinger, 1968) and familial glomerulopathy (Gemperle et al., 1996). Patients with polymastia or supernumerary nipples have been noted to be at increased risk of renal cell cancer (Goedert et al., 1981; Asal et al., 1988b).

Kidney Transplantation and Dialysis The average annual incidence of renal cell carcinoma is substantially higher among patients on renal replacement therapy compared with the general population (Maisonneuve et al., 1999). Among patients undergoing long-term renal dialysis, there is an increased incidence of acquired cystic disease of the kidney, which, in turn, predisposes to renal cell cancer, especially in men (Matson and Cohen, 1990; Sasagawa et al., 1992; Chen et al., 1995; Buccianti et al., 1996; Stewart et al., 2003; Ishikawa et al., 2003). Risk of kidney cancer appears to increase with increasing duration of dialysis (Denton et al., 2002; Stewart et al., 2003). As in dialysis patients, acquired renal cystic disease of the native kidneys appears to be a major risk factor for the development of renal cell carcinoma in renal transplant recipients in most studies (Kliem et al., 1997; Almirall et al., 1990; Lien et al., 1991; Ishikawa et al., 1991).

RISK FACTORS FOR RENAL PELVIS/URETER CANCER As with renal cell cancer, the identification of risk factors for renal pelvis and ureter cancers has come mainly from case-control studies, summarized in Table 57–5. As a result of the relative infrequency of renal pelvis and ureter tumors, these studies are typically smaller than those for renal cell cancer. The most well-studied and the only established risk factors for renal pelvis cancer are cigarette smoking and use of phenacetin-containing analgesics.

Cigarette Smoking Smoking-related risks are generally higher than those for renal cell cancer or bladder cancer (Schmauz and Cole, 1974; Armstrong et al., 1976; McCredie et al., 1982, 1983a; McLaughlin et al., 1983, 1992b; Jensen et al., 1988; Ross et al., 1989; McCredie and Stewart, 1992a; Pommer et al., 1999). The risks for smokers are up to 7 times greater than those for nonsmokers, with heavy smokers having risks up to 11fold. This variation in reported risks probably reflects the relatively small number of cases in most of the studies. Cessation of smoking for 10 years or longer reduced the risks for these tumors by 60% to 70% relative to current smokers in the United States (McLaughlin et al., 1992b). Similar reductions in risk were seen among quitters in Australia (McCredie and Stewart, 1992a). This steep decline in risk suggests that smoking affects a relatively late stage in carcinogenesis, thus making it possible for smoking cessation to lower the risk for these tumors. Population-based attributable risk estimates for cigarette smoking and renal pelvis and ureter cancers in the United States have suggested that 70% to 82% of the cases among men and 37% to 61% among women are due to smoking (McLaughlin et al., 1983, 1992b). Results from Denmark indicate that smoking accounts for 56% of the cases of renal pelvis and ureter cancers among men and 40% among women (Jensen et al., 1988). In Australia, 46% of the cases among men and 35% among women are attributable to cigarette smoking

Table 57–5. Published Case-Control Studies of Risk Factors for Renal Pelvis/Ureter Cancer Authors Schmauz Armstrong et al. McCredie et al. McCredie et al. McCredie et al. McLaughlin et al. Jensen et al. Ross et al. McLaughlin et al. McCredie Pommer et al.

Year

Source of Controls

Number of Cases/Controls

Location

1974 1976 1982 1983a 1983b 1983 1988 1989 1992b 1992a 1999

Population Hospital Friend/clinic Population Population Population Hospital Neighborhood Population Population Population

17 renal pelvis, 10 ureter/451 33 renal pelvis/33 67 renal pelvis, 84 ureter/96 29 renal pelvis, 36 ureter/307 31 renal pelvis/400 74 renal pelvis/697 76 renal pelvis, 20 ureter/294 121 renal pelvis, 66 ureter/187 308 renal pelvis, 194 ureter/496 147 renal pelvis/523 51 renal pelvis, 25 ureter/647

Massachusetts England New South Wales New South Wales New South Wales Minnesota Denmark Los Angeles 3 areas in U.S. New South Wales Germany

Renal Cancer (McCredie and Stewart, 1992a). Thus, cigarette smoking appears to be the strongest risk factor for these tumors and accounts for the majority of cases in most areas of the world.

1095

probably reflect protopathic bias, confounding by earlier phenacetin use, or chance, but further study may be warranted.

Medical Conditions and Medications Analgesics The relation between heavy use of phenacetin-containing analgesics and cancers of the renal pelvis, ureter, and bladder is well established (IARC, 1980, 1987). Case reports and clinical surveys starting in the mid-1960s and continuing through the 1970s linked phenacetin use to analgesic nephropathy, and an accompanying excess of renal pelvis cancer (Hultengren et al., 1965). In a series of case-control studies in New South Wales, Australia, where analgesic abuse is relatively common, McCredie and colleagues (1982, 1983a, 1983b, 1993) reported three- to 12-fold increased risks for renal pelvis and ureter cancers among men and women using phenacetin analgesics. In the United States, where analgesic abuse is relatively uncommon (Murray and Goldberg, 1978), two case-control studies were limited by the small number of exposed subjects. In the Minnesota study of renal pelvis cancer, long-term use (over 3 years) of phenacetin was related to an eight-fold increased risk for men and a four-fold increased risk for women (McLaughlin et al., 1985). In the Los Angeles study, the risk for renal pelvis and ureter cancers was only slightly elevated after use of phenacetin analgesics over 30 consecutive days (Ross et al., 1989). In Denmark, use of phenacetin-containing analgesics was associated with relative risks of 2.4 among men and 4.2 among women after adjustment for use of other analgesics, cigarettes, and occupation (Jensen et al., 1989). Intake of more than 1 kg of phenacetin in analgesic mixtures was associated with a non-significant five-fold increase in risk in a recent German case-control study (Pommer et al., 1999). Phenacetin has also been shown to induce urinary tract tumors in laboratory animals (Nakanishi et al., 1982; IARC, 1987). Acetaminophen is the major metabolite of phenacetin (Hinsen, 1981), and a few clinical and experimental findings have linked heavy acetaminophen intake with renal papillary necrosis (Nanra, 1983), the primary lesion of analgesic nephropathy. Results of animal studies of renal pelvis cancer have been mixed, with two studies suggesting carcinogenic effects (Flaks and Flaks, 1983; Flaks et al., 1985) and two being negative (Hiraga and Fujii, 1985; Amo and Matsuyama, 1985). There are limited epidemiologic data suggesting that acetaminophen may increase risk of renal pelvis cancer. A population-based casecontrol study in Australia showed no association with renal pelvis cancer (McCredie and Stewart, 1988). The Minnesota case-control study of renal pelvis cancer reported a positive association with acetaminophen, although only a few patients took acetaminophen-containing analgesics exclusively (McLaughlin et al., 1983, 1985). In Australia, no increase in risk was observed for cancer of the renal pelvis among acetaminophen users, but a significant 2.5-fold increase in risk was found for ureter cancer (McCredie and Stewart, 1988). In a larger and more recent Australian study, acetaminophen use was not related to risk (McCredie et al., 1993). The Los Angeles and German case-control studies reported two- to three-fold increased risks for acetaminophen use, but again, the numbers of users were small. It has been difficult to separate the effects of previous phenacetin use and subsequent intake of acetaminophen among heavy users of analgesics (McLaughlin et al., 1985). Experimental, clinical, and most epidemiological studies have shown no relation between aspirin intake and cancers of the renal pelvis or ureter (Armstrong et al., 1976; McCredie et al., 1982, 1983a, 1983b; Emkey, 1983; McLaughlin et al., 1985; Patierno et al., 1989). One study, however, has reported a significant two-fold increased risk for aspirin use, with the excess mainly among women with renal pelvis tumors (Ross et al., 1989). The Danish study also reported a significant association among women who took aspirin, which the authors attributed to prior or concomitant phenacetin use (Jensen et al., 1989). By contrast, the recent Australian study found aspirin use associated with a decreased risk of renal pelvis cancer (McCredie et al., 1993). Overall, the available evidence on analgesics indicates a causal relation between phenacetin intake and cancers of the renal pelvis and ureter. The occasional positive findings for acetaminophen and aspirin

The results of a case-control study in the United States (Liaw et al., 1997) suggested that the association reported between hypertension and renal cell cancer may apply to cancer of the renal pelvis as well. A history of hypertension diagnosed within 5 years before interview was associated with a significant 30% increase in risk, and this risk was twice as high among users of diuretics or other antihypertensive drugs (odds ratio = 2.4) than among non-users of these medications (odds ratio = 1.2). No other medical conditions, including kidney stones, kidney infections, diabetes, heart attack or stroke, were associated with renal pelvis cancer in this study. Consistent with earlier studies (McCredie and Stewart, 1992b; Lindblad et al., 1993), no increased risks were associated with use of diuretics or other classes of antihypertensive drugs, including beta-blockers (Liaw et al., 1997). A recent Swedish record-linkage study of patients with kidney or ureter stones reported significantly elevated risk ratios for renal pelvis/ureter and bladder cancer but not renal cell cancer (Chow et al., 1997). One study (Pommer et al., 1999) reported that the intake of laxatives for at least one year was associated with an almost tenfold increased risk for cancer of the renal pelvis or ureter. This finding needs to be confirmed in other studies.

Coffee, Alcohol, Other Beverages, and Drinking Water There are few data to indicate that coffee or alcohol is related to renal pelvis and ureter cancers (Armstrong et al., 1976; McLaughlin et al., 1983; Ross et al., 1989). One study found a 15-fold increased risk for drinkers of over seven cups of coffee per day, but this observation was based on two cases (Schmauz and Cole, 1974). On the other hand, another study reported a significant inverse relation between coffee intake and risk of renal pelvis cancer (Armstrong et al., 1976). No case-control differences have been observed for alcohol consumption (Armstrong et al., 1976; McLaughlin et al., 1983; Ross et al., 1989). An excess risk has been reported among heavy consumers of tea, particularly women, which may deserve further study (McLaughlin et al., 1983). Epidemiologic investigations have demonstrated increased kidney cancer risk among individuals exposed to drinking water with high levels of inorganic arsenic, and bladder cancer risk has been elevated in many of the same studies (World Health Organization, 2001). Recent studies in Taiwan and Argentina have demonstrated that this increased risk associated with arsenic in drinking water is mainly restricted to cancers of the renal pelvis and ureter (Guo et al., 1997; Hopenhayn-Rich et al., 1998). Thus it appears that orally ingested arsenic is a risk factor for transitional cell cancers throughout the urinary tract.

Occupation There are few occupational associations for renal pelvis and ureter cancers because of their relative rarity and frequent inclusion with renal cell cancers in occupational cohort studies. Early case reports linked renal pelvis and ureter tumors with exposure to dyes (Macalpine, 1947; Poole-Wilson, 1969). A significant excess of employment in the leather industry was reported in Massachusetts (Schmauz and Cole, 1974), but was not confirmed in the British case-control study, which included an area with a concentration of boot and shoe manufacturing (Armstrong et al., 1976). Case-control studies in Australia (McCredie et al., 1983) and the United States (McLaughlin et al., 1983) have revealed no significant occupational associations. However, in the U.S. study, significant increases in risk were associated with self-reported exposure to coal, natural gas, and mineral oils (McLaughlin et al., 1983). In a recent Australian case-control study, employment in the dry cleaning, iron and steel, and petroleum refining industries was related to an increased risk (McCredie and Stewart, 1993). The Danish

1096

PART IV: CANCER BY TISSUE OF ORIGIN

case-control study found significantly increased risks associated with employment in the chemical, petrochemical, and plastics industries, and for exposures to coal and coke, and to asphalt and tar (Jensen et al., 1988). A record-linkage survey of occupation and cancer incidence in Denmark found significantly elevated risks for renal pelvis and ureter cancers in a number of industries, including forestry and logging; slaughtering, preparing and preserving meat; and printing and publishing (Olsen and Jensen, 1987). A Swedish record-linkage study reported a significant excess risk among machinists and plumbers (McLaughlin et al., 1987), but adjustment for smoking was not possible in either Scandinavian survey. Although the available data are limited, the work-related risks observed for cancers of the renal pelvis and ureter resemble the occupational associations that are more clearly established for bladder cancer.

Radiation The carcinogenic influence of ionizing radiation appears stronger for the renal pelvis and ureter than the renal parenchyma. The effect is seen especially in cervical cancer patients treated with radiation (Boice et al., 1988). Renal pelvis cancer has also been a consequence of Thorotrast administration during retrograde pyelography (Verhaak et al., 1965).

FUTURE DIRECTIONS It would be possible, through the elimination of cigarette smoking, to prevent the majority of renal pelvis and ureter cancers and a smaller fraction of renal cell cancers. Efforts to reduce the prevalence of obesity and perhaps hypertension should also contribute to a reduction of renal tumors. Further epidemiologic research into renal cell cancer is needed to identify reasons for the increasing incidence, particularly among the United States black population. An opportunity now exists to incorporate advances in genomic technology that will uncover molecular subtypes of renal cancer, the etiologic role of common gene variants and their interactions with modifying risk factors, and the genetic and metabolic pathways linking obesity and hypertension with this cancer. Through these approaches, it should be possible to develop more sensitive predictors of renal cancer susceptibility and progression, and to identify molecular targets that will enable new strategies for early detection and therapeutic intervention. References Adami H-O, McLaughlin JK, Hsing AW, et al. 1992. Alcoholism and cancer risk: a population-based study. Cancer Causes Control 3:419–425. Almirall J, Ricart MJU, Campistol JM, Vilardell J, Openheimer F, Andreu J. 1990. Renal cell carcinoma and acquired cystic kidney disease after renal transplantation. Transpl Int 3:49. Amo H, Matsuyama M. 1985. Subchronic and chronic effects of feeding large amounts of acetaminophen in B6C3F1 mice. Jpn J Hyg 40:567–574. Anttila A, Pukkala E, Sallmen M, Hernberg S, Hemminki K. 1995. Cancer incidence among Finnish workers exposed to halogenated hydrocarbons. J Occup Med 37:797–806. Armstrong B, Doll R. 1975. Environmental factors and cancer incidence and mortality in different countries, with special reference to dietary practices. Br J Cancer 15:617–631. Armstrong B, Garrod A, Doll R. 1976. A retrospective study of renal cancer with special reference to coffee and animal protein consumption. Br J Cancer 33:127–136. Asal NR, Risser DR, Kadamani S, Geyer JR, Lee ET, Cherng N. 1988a. Risk factors in renal cell carcinoma. I. Methodology, demographics, tobacco. beverage use, and obesity. Cancer Detect Prev 11:359–377. Asal NR, Geyer JR, Risser DR, Lee ET, Kadamani S, Cherng N. 1988b. Risk factors in renal cell carcinoma. II. Medical history, occupation, multivariate analysis, and conclusion. Cancer Detect Prev 13:263–279. Augustsson K, Skog K, Jagerstad M, Dickman PW, Steineck G. 1999. Dietary heterocyclic amines and cancer of the colon, rectum, bladder, and kidney: a population-based study. Lancet 353:703–707. Axelson O, Selden A, Andersson K, Hogstedt C. 1994. Updated and expanded Swedish cohort study of trichloroethylene and cancer risk. J Occup Med 36:556–562.

Band PR, Le ND, Fang R. 1997. Cohort mortality study of pulp and paper mill workers in British Columbia, Canada. Am J Epidemiol 146:186– 194. Batty GD, Shipley MJ, Marmot MG, Smith GD. 2003. Blood pressure and sitespecific cancer morality: evidence from the original Whitehall study. Br J Cancer 89:1243–1247. Benhamou S, Lenfant M-H, Ory-Paoletti C, Flamant R. 1993. Risk factors for renal cell carcinoma in a French case-control study. Int J Cancer 55:32–36. Benichou J, Chow WH, McLaughlin JK, Mandel JS, Fraumeni JF Jr. 1998. Population attributable risk of renal cell cancer in Minnesota. Am J Epidemiol 148:424–430. Bennington JL, Laubscher FA. 1968. Epidemiologic studies of carcinoma of the Kidney. I. Association of renal adenocarcinoma with smoking. Cancer 21:1069–1071. Bergstrom A, Hsieh CC, Lindblad P, Lu CM, Cook NR, Wolk A. 2001. Obesity and renal cell cancer—a quantitative review. Br J Cancer 85:984–990. Bianchi GD, Cerhan JR, Parker AS, et al. 2000. Tea consumption and risk of bladder and kidney cancers in a population-based case-control study. Am J Epidemiol 151:377–383. Blackard CE, Mellinger GT. 1968. Cancer in a horseshoe kidney: a report of two cases. Arch Surg 97:616–627. Blair A, Decoufle P, Grauman D. 1979. Causes of death among laundry and dry cleaning workers. Am J Public Health 69:508–511. Blair A, Hartge P, Stewart PA, McAdams M, Lubin J. 1998. Mortality and cancer incidence of aircraft maintenance workers exposed to trichloroethylene and other organic solvents and chemicals: extended follow up. Occup Environ Med 55:161–171. Blair A, Petralia SA, Stewart PA. 2002. Extended mortality follow-up of a cohort of dry cleaners. Ann Epidemiol 13:50–56. Blot WJ, Fraumeni JF Jr. 1979. Geographic patterns of renal cancer in the United States. J Natl Cancer Inst 63:363–366. Bodmer D, van den Hurk W, van Groningen JJM, et al. 2002. Understanding familial and non-familial renal cell cancer. Human Molec Gen 11:2489–2498. Boeing H, Schlehofer B, Wahrendorf J. 1997. Diet, obesity and risk for renal cell carcinoma: results from a case-control study in Germany. Z Ernahrungswiss 36:3–11. Boice JD, Engholm G, Kleinerman RA, et al. 1998. Radiation dose and secondcancer risk in patients treated for cancer of the cervix. Radiat Res 116: 3–55. Boice JD, Marano DE, Fryzek JP, Sadler CJ, McLaughlin JK. 1999. Mortality among aircraft manufacturing workers. Occup Environ Med 56:581– 597. Breslow NE, Enstrom JE. 1974. Geographic correlations between cancer mortality rates and alcohol-tobacco consumption in the United States. J Natl Cancer Inst 53:631–639. Brown DP, Kaplan SD. 1987. Retrospective cohort mortality study of dry cleaning workers using perchloroethylene. J Occup Med 29:535–541. Brownson RC (1988). A case-control study of renal cell carcinoma in relation to occupation, smoking, and alcohol consumption. Arch Environ Health 43:238–241. Bruning T, Pesch B, Wiesenhutter B, et al. 2003. Renal cell cancer risk and occupational exposure to trichloroethylene: results of a consecutive casecontrol study in Arnsberg, Germany. Am J Ind Med 43:274–285. Buccianti G, Maisonneuve P, Ravasi B, Cresseri D, Locatelli F, Boyle P. 1996. Cancer among patients on renal replacement therapy: a population-based survey in Lombardy, Italy. Int J Cancer 66:591–593. Buck C, Donner A. 1987. Cancer incidence in hypertensives. Cancer 59: 1386–1390. Buzio L, De Palma G, Mozzoni P, et al. 2003. Glutathione S-transferases Ml-1 and Tl-1 as risk modifiers for renal cell cancer associated with occupational exposure to chemicals. Occup Environ Med 60:789–793. Calle EE, Kaaks R. 2004. Overweight, obesity and cancer: epidemiologic evidence and proposed mechanisms. Nature Rev Cancer 4:579–591. Chiu BC, Lynch CF, Cerhan JR, Cantor KP. 2001. Cigarette smoking and risk of bladder, pancreas, kidney and colorectal cancers in Iowa. Ann Epidemiol 11:28–37. Chen KS, Lai MK, Huang CC, Chu SH, Leu ML. 1995. Urologic cancers in uremic patients. Am J Kidney Dis 25:694–700. Cherrie JW, Kromhout H, Semple S. 2001. The importance of reliable exposure estimates in deciding whether trichloroethylene can cause kidney cancer. J Cancer Res Clin Oncol 127:400–404. Chiu BCH, Lynch CF, Cerhan JR, Cantor KP. 2000. Cigarette smoking and risk of bladder, pancreas, kidney, and colorectal cancers in Iowa. Ann Epidemiol 11:28–37. Chow W-H, McLaughlin JK, Linet MS, Niwa S, Mandel JS. 1994a. Use of analgesics and risk of renal cell cancer. Int J Cancer 59:467–470.

Renal Cancer Chow W-H, Gridley G, McLaughlin JK, et al. 1994b. Protein intake and risk of renal cell cancer. J Natl Cancer Inst 86:1131–1139. Chow WH, McLaughlin JK, Mandel JS, Wacholder S, Niwa S, Fraumeni JF Jr. 1996. Obesity and risk of renal cell cancer. Cancer Epidemiol Biomarkers Prev 5:17–21. Chow WH, Lindblad P, Gridley G, et al. 1997. Risk of urinary tract cancers following kidney or ureter stones. J Natl Cancer Inst 89:1453–1457. Chow WH, Devesa SS, Warren JL, Fraumeni JF Jr. 1999. The rising incidence of renal cell cancer in the United States. JAMA 281:1628–1631. Chow WH, Gridley G, Fraumeni JR Jr, Jarvholm B. 2000. Obesity, hypertension, and the risk of kidney cancer in men. N Engl J Med 343:1305–1311. Cohen AJ, Li FP, Berg S, et al. 1979. Hereditary renal-cell carcinoma associated with a chromosomal translocation. N Engl J Med 301:592–595. Costantino JP, Redmond CK, Bearden A. 1995. Occupationally related cancer risk among coke oven workers: 30-years of follow-up. J Occup Environ Med 37:597–604. Coughlin SS, Neaton JD, Randall B, Sengupta A. 1997. Predictors of mortality from kidney cancer in 332,547 men screened for the Multiple Risk Factor Intervention Trial. Cancer 79:2171–2177. Czene K, Hemminki K. 2003. Familial papillary renal cell tumors and subsequent cancers: a nationwide epidemiological study from Sweden. J Urol 169:1271–1275. De Stefani E, Fierro L, Mendilaharsu M, et al. 1998. Meat intake, “mate” drinking and renal cell cancer in Uruguay: a case-control study. Br J Cancer 78:1239–1243. Denton MD, Magee CC, Ovuworie C, et al. 2002. Prevalence of renal cell carcinoma in patients with ESRD pre-transplantation: a pathologic analysis. Kidney Int 61:2201–2209. Derby JE, Jick H. 1996. Acetaminophen and renal and bladder cancer. Epidemiol 7:358–362. Devesa SS, Silverman DT, McLaughlin JK, Brown CC, Connelly RR, Fraumeni JF Jr. 1990. Comparison of the descriptive epidemiology of urinary tract cancer. Cancer Causes Control 1:133–141. Domiano SF, Vena JE, Swanson MK. 1985. Gasoline exposure, smoking, and kidney cancer. J Occup Med 27:398–399. Dosemeci M, Cocco P, Chow WH. 1999. Gender differences in risk of renal cell carcinoma and occupational exposures to chlorinated aliphatic hydrocarbons. Am J Ind Med 36:54–59. Duh R-W, Asal NR. 1984. Mortality among laundry and dry cleaning workers in Oklahoma. Am J Public Health 74:1278–1280. Emkey RD. 1983. Aspirin and renal disease. Am J Med 74(suppl 6A):97–701. Enterline PE, Hartley J, Henderson V. 1987. Asbestos and cancer: a cohort followed up death. Br J Ind Med 44:396–401. Finkle WD, McLaughlin JK, Rasgon SA, Yeoh HH, Low JE. 1993. Increased risk of renal cell cancer among women using diuretics in the United States. Cancer Causes Control 4:555–558. Flaks A, Flaks B. 1983. Induction of liver cell tumor in 1F mice by paracetamol. Carcinogenesis 4:363–368. Flaks B, Flaks A, Shaw APW. 1985. Induction by paracetamol of bladder and liver tumors in the rat. Acta Pathol Microbiol Immunol Scand (Section A) 93:367–377. Fraser GE, Phillips RL, Beeson WL. 1990. Hypertension, antihypertensive medication and risk of renal carcinoma in California Seventh-Day Adventists. Int J Epidemiol 19:832–838. Fu H, Boffetta P. 1995. Cancer and occupational exposure to inorganic lead compounds: a meta-analysis of published data. Occup Environ Med 52: 73–81. Gago-Dominguez M, Yuan JM, Castelao JE, Ross RK, Yu MC. 1999a. Regular use of analgesics is a risk factor for renal cell carcinoma. Br J Cancer 81:542–548. Gago-Dominguez M, Castelao JE, Yuan JM, Ross RK, Yu MC. 1999b. Increased risk of renal cell carcinoma subsequent to hysterectomy. Cancer Epidemiol Biomarkers Prev 8:999–1003. Gago-Dominguez M, Yuan JM, Castelao JE, Ross RK, Yu MC. 2001. Family history and risk of renal cell carcinoma. Cancer Epidemiol Biomarkers Prev 10:1001–1004. Gago-Dominguez M, Castelao JE, Yuan JM, Ross RK, Yu MC. 2002. Lipid peroxidation: a novel and unifying concept of the etiology of renal cell carcinoma (United States). Cancer Causes Control 13:287–293. Gamble JF, Pearlman ED, Nicolich MJ. 1996. A nested case-control study of kidney cancer among refinery/petrochemical workers. Environ Health Perspect 104:642–650. Garabrant DH, Held J, Langholz B, Bernstein L. 1988. Mortality of aircraft manufacturing workers in Southern California. Am J Ind Med 13: 683–693. Gemperle O, Neuweiler J, Reutter FW, Hildebrandt F, Krapf R. 1996. Familial glomerulopathy with giant fibrillar (fibronectin-positive) deposits: 15year follow-up in a large kindred. Am J Kidney Dis 28:668–675.

1097

Gnarra JR, Lerman MI, Zbar B, Linehan WM. 1995. Genetics of renal-cell carcinoma and evidence for a critical role for von Hippel-Lindau in renal tumorigenesis. Semin Oncol 22:3–8. Goedert JJ, McKeen EA, Fraumeni JF Jr. 1981. Polymastia and renal adenocarcinoma. Ann Intern Med 95:182–184. Goodman MT, Morgenstern H, Wynder EL. 1986. A case-control study of factors affecting the development of renal cell cancer. Am J Epidemiol 124:926–941. Green LC, Lash TL. 1999. Re: “Renal cell cancer correlated with occupational exposure to trichloroethylene”. J Cancer Res Clin Oncol 125:430–432. Gridley G, McLaughlin JK, Ekbom A, et al. 1993. Incidence of cancer among patients with rheumatoid arthritis. J Natl Cancer Inst 85:307–311. Grove JS, Nomura A, Severson RK, Stermmerman GN. 1991. The association of blood pressure with cancer incidence in a prospective study. Am J Epidemiol 134:942–947. Guidotti TL. 1995. Occupational mortality among firefighters: assessing the association. J Occup Med 37:1348–1356. Guo H-R, Chiang H-S, Hu H, Lipsizt SR, Monso RR. 1997. Arsenic in drinking water and incidence of urinary cancers. Epidemiol 8:545–550. Hajjar I, Kotchen TA. 2003. Trends in prevalence, awareness, treatment, and control of hypertension in the United States, 1988–2000. JAMA 290: 199–206. Handa K, Kreiger N. 2002. Diet patterns and the risk of renal cell carcinoma, Public Health Nutr 5:757–767. Hansen J, Raaschou-Nielsen O, Christensen JM, et al. 2001. Cancer incidence among Danish workers exposed to trichloroethylene. J Occup Environ Med 43:133–139. Heath CW Jr, Lally CA, Calle EE, McLaughlin JK, Thun MJ. 1997. Hypertension, diuretics, and antihypertensive medications as possible risk factors for renal cell cancer. Am J Epidemiol 145:607–613. Henschler D, Vamvakas S, Lammert M, et al. 1995. Increased incidence of renal cell tumors in a cohort of cardboard workers exposed to trichloroethylene. Arch Toxicol 69:291–299. Hiatt RA, Tolan K, Quesenberg CP. 1994. Renal cell carcinoma and thiazide use: a historical case-control study. Cancer Causes Control 5:319–325. Hinds MW, Kolonel LM, Lee J, Hirohata T. 1980. Association between cancer incidence and alcohol/cigarette consumption among five ethnic groups in Hawaii. Br J Cancer 41:929–940. Hinsen JA. 1981. Biochemical toxicology of acetaminophen. In: Hodgson E, Bend JR, Philpot RM (eds). Reviews in Biochemical Toxicology. Vol 2. New York: Elsevier/North Holland, pp. 103–129. Hiraga K, Fujii T. 1985. Carcinogenicity testing of acetaminophen in F344 rats. Jph J Cancer Res 76:79–85. Hole DJ, Hawthorne VM, Isles CG, et al. 1993. Incidence of and mortality from cancer in hypertensive patients. Br Med J 306:609–611. Hopenhayn-Rich C, Biggs ML, Smith AH. 1998. Lung and kidney cancer mortality associated with arsenic in drinking water in Cordoba, Argentina. Int J Epidemiol 27:561–569. Hu J, Mao Y, White K, The Canadian Cancer Registries Epidemiology Research Group 2003. Diet and vitamin or mineral supplements and risk of renal cell carcinoma in Canada. Cancer Causes Control 14:705– 714. Huang J, Hisanga N, Sakai K, et al. 1988. Asbestos fibers in human pulmonary and extrapulmonary tissues. Am J Ind Med 14:331–339. Hultengren N, Lagergren C, Ljungqvist A. 1965. Carcinoma of the renal pelvis in renal papillary necrosis. Acta Chir Scand 130:314–320. International Agency for Research on Cancer 1980. Some pharmaceutical drugs. IARC Monogr Eval Carcinog Risks Hum 24:135–161. International Agency for Research on Cancer 1987. Overall Evaluations of Carcinogenicity: An Updating of IARC Monographs, Volumes 1 to 42. IARC Monogr Eval Carcinog Risks Hum Suppl 7. International Agency for Research on Cancer 1989. Occupation exposures in petroleum refining; crude oil and major petroleum fuels. IARC Monogr Eval Carcinog Risks Hum 45:39–117. Ishikawa I, Ishii H, Shinoda A, et al. 1991. Renal cell carcinoma of the native kidney after renal transplantation. Nephron 58:354–358. Ishikawa I, Saito Y, Asaka M, et al. 2003. Twenty-year follow-up of acquired renal cystic disease. Clin Nephrol 59:153–159. Jacobsen BK, Bjelke E, Kvale G, Heuch I. 1986. Coffee drinking, mortality, and cancer incidence: results from a Norwegian prospective study. J Natl Cancer Inst 76:823–841. Jemal A, Siegal R, Ward E, et al. 2006. Cancer statistics, 2006. Ca Cancer J Clin 56:106–130. Jensen OM. 1979. Cancer morbidity and causes of death among Danish brewery workers. Int J Cancer 23:454–463. Jensen OM, Knudsen JB, McLaughlin JK, Sorensen BL. 1988. The Copenhagen case-control study of renal pelvis and ureter cancer: role of smoking and occupational exposures. Int J Cancer 41:557–561.

1098

PART IV: CANCER BY TISSUE OF ORIGIN

Jensen OM, Knudsen JB, Tomasson H, Sorensen BL. 1989. The Copenhagen case-control study of renal pelvis and ureter cancer: role of analgesics. Int J Cancer 44:965–968. Kantor AF, McLaughlin JK. 1985. Second cancer following cancer of the urinary system in Connecticut, 1935–82. NCI Monogr 68:149– 159. Kantor AF, McLaughlin JK, Curtis RE, Flannery JT, Fraumeni JF Jr. 1986. Risk of second malignancy after cancers of the renal parenchyma, renal pelvis, and ureter. Cancer 58:1158–1161. Katz RM, Jowett D. 1981. Female laundry and dry cleaning workers in Wisconsin: a mortality analysis. Am J Public Health 71:305–307. Katz DL, Zheng T, Holford TR, Flannery J. 1994. Time trends in the incidence of renal carcinoma: analysis of Connecticut tumour registry data, 1935–1989. Int J Cancer 58:57–63. Kaye JA, Myers MW, Jick H. 2001. Acetaminophen and the risk of renal and bladder cancer in the general practice research database. Epidemiol 12:690–694. Kellerer M, von Eye Corleta H, Muhlhofer A, et al. 1995. Insulin- and insulinlike growth factor-I receptor tyrosine-kinase activities in human renal carcinoma. Int J Cancer 62:501–507. Kinlen LJ, Willows AN, Goldblatt P, Yudkin J. 1988. Tea consumption and cancer. Br J Cancer 58:397–401. Kliem V, Kolditz M, Behrend M, et al. 1997. Risk of renal cell carcinoma after kidney transplantation. Clin Transplant 11:255–258. Kolonel LN. 1976. Association of cadmium with renal cancer. Cancer 37:1782–1787. Kosary CL, McLaughlin JK. 1993. Kidney and renal pelvis. In: Miller BA, Ries LAG, Hankey BE, et al. (eds). Cancer Statistics Review: 1973–1990. National Cancer Institute. NIH Pub. No. 93-2789, x1–x22. Kreiger N, Marrett LD, Dodds L, Hilditch S, Darlington GA. 1993. Risk factors for renal cell carcinoma: results of a population-based case-control study. Cancer Causes Control 4:101–110. La Vecchia C, Negri E, D’Avanzo B, Franceschi S. 1990. Smoking and renal cell carcinoma. Cancer Res 50:5231–5233. La Vecchia C, Negri E, Franceschi S, D’Avanzo B, Boyle P. 1992. Tea consumption and cancer risk. Nutr Cancer 17:27–31. Lambe M, Lindblad P, Wuu J, Remler R, Hsieh CC. 2002. Pregnancy and risk of renal cell cancer: a population-based study in Sweden. Br J Cancer: 86:1425–1429. Land CE. 1986. Carcinogenic effects of radiation on the human digestive tract and other organs. In: Cyston AC, Albert RE, Burns FJ, et al. (eds). Radiation Carcinogenesis. New York: Elsevier, pp 347–378. Laski Me. 1986. Diuretics: mechanism of action and therapy. Semin Nephrol 6:210–223. Latif F, Tory K, Gnarra J, et al. 1993. Identification of the von Hippel-Lindau disease tumor suppression gene. Science 260:1317–1320. Lew EA, Garfinkel L. 1979. Variations in mortality by weight among 750,000 men and women. J Chron Dis 32:563–576. Lewis RJ, Schnatter AR, Drummond I, et al. 2003. Mortality and cancer morbidity in a cohort of Canadian petroleum workers. Occup Environ Med 60:918–928. Liaw KL, Linet MS, McLaughlin JK, et al. 1997. Possible relation between hypertension and cancers of the renal pelvis and ureter. Int J Cancer 70:265–268. Lien Y-HH, Kam I, Shanley PF, Schroter GPJ. 1991. Metastatic renal cell carcinoma associated with acquired cystic kidney disease 15 years after successful renal transplantation. Am J Kidney Dis 18:711. Lijinsky W, Reuber MD. 1987. Pathologic effects of chronic administration of hydrochlorothiazide, with and without sodium nitrate, to 344 rats. Toxicol Ind Health 3:413–422. Lindblad P, McLaughlin JK, Mellemgaard A, Adami HO. 1993. Risk of kidney cancer among patients using analgesics and diuretics: a population-based cohort study. Int J Cancer 55:5–9. Lindblad P, Wolk A, Bergstrom A, Persson I, Adami HO. 1994. The role of obesity and weight fluctuations in the etiology of renal cell cancer: a population-based case-control study. Cancer Epidemiol Biomarkers Prev 3:631–639. Lindblad P, Chow WH, Chan J, et al. 1995. International renal-cell cancer study. V. Reproductive factors, gynecologic operations and exogenous hormones. Int J Cancer 61:192–198. Lindblad P, Wolk A, Bergstrom R, Adami HO. 1997. Diet and risk of renal cell cancer: a population-based case-control study. Cancer Epidemiol Biomarkers Prev 6:215–223. Linehan WM, Lerman MI, Zbar B. 1995. Identification of the von HippelLindau (VHL) gene. Its role in renal cancer. JAMA 273:564–570. Linehan WM, Klausner RD. 1998. Renal carcinoma. In: Vogelstein B, Kinzler KW (eds). The genetic basic of human cancer. New York: McGraw-Hill, pp. 455–473.

Lonser RR, Glenn GM, Walther M, et al. 2003. von Hippel-Lindau disease. Lancet 361:2059–2067. Lowery JT, Peters JM, Deapen D, London SJ. 1991. Renal cell carcinoma among architects. Am J Ind Med 20:123–125. Lynge E, Andersen A, Nilsson R, et al. 1997. Risk of cancer and exposure to gasoline vapors. Am J Epidemiol 145:449–458, 1997. Macalpine JB. 1947. Papilloma of the renal pelvis in dye workers: two cases, one of which shows bilateral growth. Br J Surg 35:137–140. MacFarland HN, Ulrich CF, Holdsworth CE, Kitchen DN, Halliwell WH, Blum SC. 1984. A chronic inhalation study with unleaded gasoline vapor. J Am Coll Toxicol 3:231–248. Maclure M. 1987. Asbestos and renal adenocarcinoma: a case-control study. Environ Res 42:353–361. Maclure M, MacMahon B. 1985. Phenacetin and cancer of the urinary tract. N Engl J Med 313:1479. Maclure M, Willett W. 1990. A case-control study of diet and risk of renal adenocarcinoma. Epidemiology 1:430–440. Maisonneuve P, Agodoa L, Gellert R, et al. 1999. Cancer in end stage renal disease patients undergoing dialysis: results from an international collaborative study. Lancet 354:93–99. Mandel JS, Mclaughlin JK, Schlehofer B, et al. 1995. International renal-cell cancer study. IV. Occupation. Int J Cancer 61:601–605. Mandel JS. 2001. Renal cell cancer correlated with occupational exposure to trichloroethylene. J Cancer Res Clin Oncol 127:265–266. Mathew A, Devesa SS, Fraumeni JF Jr, Chow WH. 2002. Global increases in kidney cancer incidence, 1973–1992. Eur J Cancer Prev 11:171–178. Matson MA, Cohen EP. 1990. Acquired cystic kidney disease: occurrence, prevalence, and renal cancers. Medicine (Baltimore) 69:217–226. Mattioli S, Truffelli D, Baldasseroni A, et al. 2002. Occupational risk factors for renal cell cancer: a case-control study in northern Italy. J Occup Environ Med 44:1028–1036. McCredie M, Ford JM, Taylor JS, Stewart JH. 1982. Analgesics and cancer of the renal pelvis in New South Wales. Cancer 49:2617–2625. McCredie M, Stewart JH, Ford JM. 1983a. Analgesics and tobacco as risk factors for cancer of the ureter and renal pelvis. J Urol 130:28–30. McCredie M, Stewart JH, Ford JM, MacLennan RA. 1983b. Phenacetincontaining analgesics and cancers of the bladder and renal pelvis in women. Br J Urol 55:220–224. McCredie M, Stewart JH. 1988. Does paracetamol cause urothelial cancer or renal papillary necrosis? Nephron 49:296–300. McCredie M, Ford JM, Stewart JH. 1988. Risk factors for cancer of the renal parenchyma. Int J Cancer 42:13–16. McCredie M, Stewart JH. 1992a. Risk factors for kidney cancer in New South Wales: I. Cigarette smoking. Eur J Cancer 28A:2050–2054. McCredie M, Stewart JH. 1992b. Risk factors for kidney cancer in New South Wales. II. Urological disease, hypertension, obesity and hormonal factors. Cancer Causes Control 3:323–331. McCredie M, Stewart JH. 1993. Risk factors for kidney cancer in New South Wales: IV. Occupation. Br J Ind Med 50:349–354. McCredie M, Stewart JH, Day NE. 1993. Different roles for phenacetin and paracetamol in cancer of the kidney and renal pelvis. Int J Cancer 53:245–249. McCredie M, Pommer W, McLaughlin JK, et al. 1995. International renal cell cancer study. II. Analgesics. Int J Cancer 60:345–349. McCredie M, Macfarlane GJ, Stewart J, Coates M. 1996. Second primary cancers following cancers of the kidney and prostate in New South Wales (Australia), 1972–1991. Cancer Causes Control 7:337–344. McFarland WI, Wallace S, Johnson DE. 1972. Renal carcinoma and polycystic disease. J Urol 107:530–532. McLaughlin JK, Schuman LM. 1983. Epidemiology of renal cell carcinoma. In: Lilienfeld AM (ed). Reviews in Cancer Epidemiology. Vol 2. New York: Elsevier/North Holland, 170–210. McLaughlin JK, Blot WJ, Mandel JS, Schuman LM, Mehl ES, Fraumeni JF Jr. 1983. Etiology of cancer of the renal pelvis. J Natl Cancer Inst 71:287–291. McLaughlin JK, Mandel JS, Blot WJ, Schuman LM, Mehl ES, Fraumeni JF Jr. 1984. Population-based case-control study of renal cell carcinoma. J Natl Cancer Inst 72:275–284. McLaughlin JK, Blot WJ, Mehl ES, Fraumeni JF Jr. 1985. Relation of analgesic use to renal cancer: population-based findings. NCI Monogr 69:213–215. McLaughlin JK, Malker HS, Stone BJ, et al. 1987. Occupational risks for renal cancer in Sweden. Br J Ind Med 44:119–123. McLaughlin JK, Blot WJ, Fraumeni JF Jr. 1988. Diuretics and renal cell cancer. J Natl Cancer Inst 80:378. McLaughlin JK, Hrubec Z, Heineman EF, Blot WJ, Fraumeni JF Jr. 1990. Renal cancer and cigarette smoking in a 26-year followup of U.S. veterans. Public Health Rep 105:535–537.

Renal Cancer McLaughlin JK, Gao YT, Gao RN, et al. 1992a. Risk factors for renal cell cancer in Shanghai, China. Int J Cancer 52:562–565. McLaughlin JK, Silverman DT, Hsing AW, et al. 1992b. Cigarette smoking and cancers of the renal pelvis and ureter. Cancer Res 52:254–257. McLaughlin JK. 1993. Renal cell cancer and exposure to gasoline: a review. Environ Health Perspect Suppl 101 (suppl 6):111–114. McLaughlin JK, Lindblad P, Mellemgaard A, et al. 1995a. International renalcell cancer study. I. Tobacco use. Int J Cancer 60:194–198. McLaughlin JK, Chow WH, Mandel JS, et al. 1995b. International renal-cell cancer study. VIII. Role of diuretics, other anti-hypertensive medications and hypertension. Int J Cancer 63:216–221. McLaughlin JK, Blot WJ. 1997. A critical review of epidemiology studies of trichloroethylene and perchloroethylene and risk of renal cell cancer. Int Arch Occup Environ Health 70:222–231. Mellemgaard A, Moller H, Olsen JH, Jensen OM. 1991. Increased risk of renal cell carcinoma among obese women. J Natl Cancer Inst 21:1581–1582. Mellemgaard A, Moller H, Jensen OM, Halberg P, Olsen JH. 1992a. Risk of kidney cancer in analgesics users. J Clin Epidemiol 45:1021–1024. Mellemgaard A, Moller H, Olsen JH. 1992b. Diuretics may increase risk of renal cell carcinoma. Cancer Causes Control 3:309–312. Mellemgaard A, Engholm G, McLaughlin JK, Olsen JH. 1994a. Risk of renal cell carcinoma in Denmark. I. Role of socioeconomic status, tobacco use, beverages, and family history. Cancer Causes Control 5:105–113. Mellemgaard A, Niwa S, Mehl ES, Engholm G, McLaughlin JK, Olsen JH. 1994b. Risk factors for renal cell carcinoma in Denmark. II. Role of medication and medical history. Int J Epidemiol 23:923–930. Mellemgaard A, Engholm G, McLaughlin JK, Olsen JH. 1994c. Risk of renal cell carcinoma in Denmark. III. Role of weight, physical activity, and reproductive factors. Int J Cancer 56:66–71. Mellemgaard A, Engholm G, McLaughlin JK, Olsen JH. 1994d. Occupational risk factors for renal cell carcinoma in Denmark. Scand J Work Environ Health 20:160–165. Mellemgaard A, Lindblad P, Schlohofer B, et al. 1995. International renal-cell cancer study. III. Role of weight, height, physical activity and use of amphetamines. Int J Cancer 60:350–354. Monson RR, Lyon JL. 1975. Proportional mortality among alcoholics. Cancer 36:1077–1079. Morgan RW, Kelsh MA, Zhao K, Heringer S. 1998. Mortality of aerospace workers exposed to trichloroethylene. Epidemiol 9:424–431. Mundt KA, Birk T, Burch MT. 2003. Critical review of the epidemiological literature on occupational exposure to perchloroethylene and cancer. Int Arch Occup Environ Health 76:473–491. Murray TG, Goldberg M. 1978. Analgesic-associated nephropathv in the U.S.A.: epidemiologic, clinical, and pathogenetic features. Kidney Int 13:64–71. Nakanishi K, Kurata Y, Oshima M, Fukushima S, Ito N. 1982. Carcinogenicity of phenacetin: long-term feeding study in B6C3F1 mice. Int J Cancer 29:439–444. Nanra RS. 1983. Renal effects of antipyretic analgesics. Am J Med 75:70–81. National Cancer Institute. 2003. DCCPS, Surveillance Research Program, Cancer Statistics Branch. SEER Program Public Use Data Tapes 1973–2000, November 2002 Submission. National Toxicology Program. 1989a. Toxicology and carcinogenesis studies of furosemide in F344/N rats and B6C3F mice. NTP Technical Report No. 356, Research Triangle Park, North Carolina. National Toxicology Program. 1989b. Toxicology and carcinogenesis studies of hydrochlorothiazide in F344/N rats and B6C3F mice. NTP Technical Report No. 357, Research Triangle Park, North Carolina. Neumann HP, Zbar B. 1997. Renal Cysts, renal cancer and von Hippel-Lindau disease. Kidney Int 51:16–26. Newsom GD, Vurgin D. 1987. Etiologic factors in renal cell adenocarcinoma. Semin Nephrol 7:109–116. Nicholas JS, Lackland DT, Dosemeci M, et al. 1998. Mortality among US commercial pilots and navigators. J Occup Environ Med 40:980–985. Olsen JH, Jensen OM. 1987. Occupation and risk of cancer in Denmark: an analysis of 93,810 cancer cases, 1970–1979. Scand J Work Environ Health 13(suppl 1):1–91. Paganini-Hill A, Glazer E, Henderson BE, Ross RK. 1980. Cause-specific mortality among newspaper web pressmen. J Occup Med 22:542–544. Paganini-Hill A, Chao A, Ross PK, Henderson BE. 1989. Aspirin use and chronic disease: a cohort study of the elderly. Br Med J 299:1247–1250. Parker AS, Cerhan JR, Lynch CF, Ershow AG, Cantor KP. 2002. Gender, alcohol consumption, and renal cell carcinoma. Am J Epidemiol 155:455–462. Parker AS, Cerhan JR, Janney CA, Lynch CF, Cantor KP. 2003. Smoking cessation and renal cell carcinoma. Ann Epidemiol 13:245–251. Parkin CM, Whelan SL, Ferlay J, Teppo L, Thomas D. 2002. Cancer Incidence in Five Continents, Volume VIII. IARC Scientific Publications No. 155. Lyon, France: International Agency for Research on Cancer.

1099

Partanen T, Heikkila P, Hernberg S, Kauppinen T, Moneta G, Ojajawi A. 1991. Renal cell cancer and occupational exposure to chemical agents. Scand J Work Environ Health 17:231–239. Patierno SR, Lehman NL, Henderson BE, Landolph JR. 1989. Study of the ability of phenacetin, acetaminophen, and aspirin to induce cytotoxicity, mutation, and morphological transformation in C3H/ 10T1/2 clone 8 mouse embryo cells. Cancer Res 49:1038–1044. Pell S, Alonzo CA. 1973. A five-year mortality study of alcoholics. J Occup Med 15:120–121. Pesch B, Haerting J, Ranft U, Klimpel A, Oelschlagel B, Schill W. 2000. Occupational risk factors for renal cell carcinoma: agent-specific results from a case-control study in Germany. MURC Study Group. Multicenter urothelial and renal cancer study. Int J Epidemiol 29:1014–1024. Pommer W, Bronder E, Klimpel A, Helmert U, Greiser E, Molzahn M. 1999. Urothelial cancer at different tumor sites: role of smoking and habitual intake of analgesics and laxatives. Results of the Berlin Urothelial Cancer Study. Nephrol Dial Transplant 14:2892–2897. Poole C, Dreyer NA, Satterfield MH, Levin L, Rothman KJ. 1993. Kidney cancer and hydrocarbon exposures among petroleum refinery workers. Environ Health Perspect 101(suppl 6):53–62. Poole-Wilson DB. 1969. Occupational tumours of the renal pelvis and ureter arising in the dye-making industry. Proc R Soc Med 62:93–94. Prineas RJ, Folsom AR, Zhang ZM, Sellers TA, Potter J. 1997. Nutrition and other risk factors for renal cell carcinoma in postmenopausal women. Epidemiol 8:31–36. Raaschou-Nielsen O, Hansen J, McLaughlin JK, et al. 2003. A cohort study of cancer risk among workers at Danish companies using trichloroethylene: a cohort study. Am J Epidemiol 158:1182–1192. Rashidkhani B, Lindblad P, Wolk A. 2005. Fruits, vegetables and risk of renal cell carcinoma: a prospective study of Swedish women. Int J Cancer 113:451–455. Raynor WJ, Shekelle RB, Rossof AH, Maliza C, Paul O. 1981. High blood pressure and 17-year cancer mortality in the Western Electric Health Study. Am J Epidemiol 113:371–377. Redmond CK, Ciocco A, Lloyd W, Rush HW. 1972. Long-term mortality study of steelworkers. VI. Mortality from malignant neoplasms among coke oven workers. J Occup Med 14:621–629. Redmond CK. 1983. Cancer mortality among coke oven workers. Environ Health Perspect 52:67–73. Rosenberg L, Rao RS, Palmer JR, et al. 1998a. Transitional cell cancer of the urinary tract and renal cell cancer in relation to acetaminophen use (United States). Cancer Causes Control 9:83–88. Rosenberg L, Stephenson WP, Rao RS, Palmer JR, Strom BL, Shapiro S. 1998b. The diagnosis of renal cell cancer in relation to hypertension (United States). Cancer Causes Control 9:611–614. Ross RK, Paganini-Hill A, Landolph J, Gerhins V, Henderson BE. 1989. Analgesics, cigarette smoking, and other risk factors in cancer of the renal pelvis and ureter. Cancer Res 49:1045–1048. Ruder AM, Ward EM, Brown DP. 2001. Mortality in dry-cleaning workers: an update. Am J Ind Med 39:121–132. Rushton L. 1993. The U.K. oil refinery and distribution centre studies: a 39year follow-up. Environ Health Perspect 101(suppl 6):77–84. Salas M, Hofman A, Stricker BHC. 1999. Confounding by indication: an example of variation in the use of epidemiologic terminology. Am J Epidemiol 149:981–983. Sali D, Boffetta P. 2000. Kidney cance and occupational exposure to asbestos: a meta-analysis of occupational cohort studies. Cancer Causes Control 11:37–47. Sasagawa I, Terasawa Y, Imai K, Sekino H, Takahashi H. 1992. Acquired cystic disease of the kidney and renal carcinoma in haemodialysis patients: ultrasonographic evaluation. Br J Urol 70:236–239. Schlehofer B, Heuer C, Blettner M, Niehoff D, Wahrendorf J. 1995. Occupation, smoking and demographic factors, and renal cell carcinoma in Germany. Int J Epidemiol 24:51–57. Schlehofer B, Pommer W, Mellemgaard A, et al. 1996. International renal-cell cancer study. VI. The role of medical and family history. Int J Cancer 66:723–726. Schmauz R, Cole P. 1974. Epidemiology of cancer of the renal pelvis and ureter. J Natl Cancer Inst 52:1431–1434. Schmidt W, De Lint. 1972. Causes of death of alcoholics. Q J Stud Alcohol 33:171–185. Schmidt W, Popham RE. 1981. The role of drinking and smoking in mortality from cancer and other causes in male alcoholics. Cancer 47:1031–1041. Schmidt L, Junker K, Nakaigawa N, et al. 1999. Novel mutations of the MET proto-oncogene in papillary renal carcinomas. Oncogene 18:2343–2350. Schnatter AR, Katz AM, Nicolich MJ, Theriault G. 1993. A retrospective mortality study among Canadian petroleum marketing and distribution workers. Environ Health Perspect 101(suppl 6):85–99.

1100

PART IV: CANCER BY TISSUE OF ORIGIN

Schwartz V, Flamant R, Lellouch J, et al. 1961. Results of a French survey on the role of tobacco, particularly inhalation in different cancer sites. J Natl Cancer Inst 26:1085–1108. Selikoff IJ, Hammond EC, Seidman H. 1979. Mortality experience of insulation workers in the United States and Canada, 1943–1976. Ann NY Acad Sci 330:91–116. Semenza JC, Ziogas A, Largent J, Peel D, Anton-Culver H. 2001. Geneenvironment interactions in renal cell carcinoma. Am J Epidemiol 153:851–859. Shalat SC, True LD, Fleming LE, Pace PE. 1989. Kidney cancer in utility workers exposed to polychlorinated biphenyls (PCBs). Br J Ind Med 46:823–824. Shapiro JA, Williams MA, Weiss NS, Stergachis A, LaCroix AZ, Barlow WE. 1999. Hypertension, antihypertensive medication use, and risk of renal cell carcinoma. Am J Epidemiol 149:521–530. Shennan DH. 1973. Renal carcinoma and coffee consumption in 16 countries. Br J Cancer 28:473–474. Siemiatycki J, Dewar R, Nadon L, Gerin M, Richardson L, Wacholder S. 1987. Associations between several sites of cancer and twelve petroleumderived liquids. Scand J Work Environ Health 13:493–504. Siemiatycki J, Krewski D, Franco E, Kainserman M. 1995. Associations between cigarette smoking and each of 21 types of cancer: a multi-site case-control study. Int J Epidemiol 24:504–514. Signorello LB, McLaughlin JK, Lipworth L, Friis S, Sorensen HT, Blot WJ. 2002. Confounding by indication in epidemiologic studies of commonly used analgesics. Am J Ther 9:199–205. Sinks T, Lushniak B, Haussler BJ, et al. 1992. Renal cell cancer among paperboard printing workers. Epidemiology 3:483–489. Smith LJ, Rosenberg ME, Hostetter TH. 1993. Effect of angiotensin II blockade on dietary protein-induced renal growth. Am J Kidney Dis 22: 120–127. Spiess H, Mays CW, Chmelvesky B. 1989. Radium 224 in humans. In: Taylor DM, Mays CW, Gerber GB, et al. (eds). Risks from Radium and Thorotrast. BIR Report 21:7–12. Spirtas R, Stewart PA, Lee JS, et al. 1991. Retrospective cohort mortality study of workers at an aircraft maintenance facility. I. Epidemiologic results. Br J Ind Med 48:515–530. Stensvold I, Jacobsen BK. 1994. Coffee and cancer: a prospective study of 43,000 Norwegian men and women. Cancer Causes Control. 5:401–408. Stewart JH, Buccianti G, Agodoa L, et al. 2003. Cancers of the kidney and urinary tract in patients on dialysis for end-stage renal disease: analysis of data from the United States, Europe, and Australia and New Zealand. J Am Soc Nephrol 14:197–207. Stoyanov IS, Chernozemsky IN, Nicolov IG, Stoichev II, Petkova-Bocharova TK. 1978. Epidemiologic association between endemic nephropathy and urinary system tumors in an endemic region. J Chron Dis 31:721–724. Talamini R, Baron AE, Barra S, et al. 1990. A case-control study of risk factors for renal cell cancer in northern Italy. Cancer Causes Control 1:125–131. Tate R, Iddenden R, Harnden P, et al. 2003. Increased incidence of renal parenchymal carcinoma in the Northern and Yorkshire region of England, 1978–1997. Eur J Cancer Prev 39:961–967.

Vamvakas S, Bruning T, Thomasson B, et al. 1998. Renal cell cancer correlated with occupational exposure to trichloroethene. J Cancer Res Clin Oncol 124:374–382. Verhaak R. 1965. Tumor induction in a thorotrast kidney. Oncologica 19:20–32. Weinmann S, Glass AG, Weiss NS, Psaty BM, Siscovick DS, White E. 1994. Use of diuretics and other antihypertensive medications in relation to the risk of renal cell cancer. Am J Epidemiol 140:792–804. Whisenand JM, Kostos D, Sommers SC. 1962. Some host factors in the development of renal cell carcinoma. West J Surg Gynecol 70:284–285. Whittemore AS, Paffenbarger RS Jr, Anderson K, Lee JE. 1985. Early precursors of site-specific cancers in college men and women. J Natl Cancer Inst 74:43–51. Wolk A, Gridley G, Niwa S, et al. 1996a. International renal-cell cancer study. VII. Role of diet. Int J Cancer 65:67–73. Wolk A, Lindblad P, Adami HO. 1996b. Nutrition and renal cell cancer. Cancer Causes Control 7:5–18. Wong O, Raabe GK. 1989. Critical review of cancer epidemiology in petroleum industry employees, with a quantitative meta-analysis by cancer site. Am J Ind Med 15:283–310. Wong O, Harris F, Smith TJ. 1993. Health effects of gasoline exposure. II. Mortality patterns among petroleum refinery workers. Environ Health Perspect 101(suppl 6):63–76. Wong O, Trent L, Harris F. 1999. Nested case-control study of leukemia, multiple myeloma and kidney cancer in a cohort of petroleum workers exposed to gasoline. Occup Environ Med 56:217–221. World Health Organization (2001). Arsenic and arsenic compounds. 2nd edition. International Program on Chemical Safety. Environmental Health Criteria 224. Geneva: World Health Organization. Wynder EL, Mabuchi K, Whitemore WF. 1974. Epidemiology of adenocarcinoma of the kidney. J Natl Cancer Inst 53:1619–1634. Yu MC, Mack TM, Hanesch R, Cicioni C, Henderson BE. 1986. Cigarette smoking, obesity, diuretic use, and coffee consumption as risk factors for renal cell carcinoma. J Natl Cancer Inst 77:351–356. Yuan JM, Castelao JE, Gago-Dominguez M, Yu MC, Ross RK. 1998a. Tobacco use in relation to renal cell carcinoma. Cancer Epidemiol Biomarkers Prev 5:429–433. Yuan JM, Castelao JE, Gago-Dominguez M, Ross RK, Yu MC. 1998b. Hypertension, obesity and their medications in relation to renal cell carcinoma. Br J Cancer 77:1508–1513. Yuan JM, Gago-Dominguez M, Castelao JE, Hankin JH, Ross RK, Yu MC. 1998c. Cruciferous vegetables in relation to renal cell carcinoma. Int J Cancer 77:211–216. Zbar B, Glenn G, Lubensky I, et al. 1995. Hereditary papillary renal cell carcinoma: clinical studies in 10 families. J Urol 153:907–912. Zbar B, Lerman MI. 1998. Inherited carcinomas of the kidney. Adv Cancer Res 75:163–201. Zbar B. 2000. Inherited epithelial tumors of the kidney: old and new diseases. Sem Cancer Biol 10:313–318. Zheng W, Doyle TJ, Kushi LH, Sellers TA, Hong CP, Folsom AR. 1996. Tea consumption and cancer incidence in a prospective cohort study of postmenopausal women. Am J Epidemiol 144:175–182.

58

Bladder Cancer DEBRA T. SILVERMAN, SUSAN S. DEVESA, LEE E. MOORE, AND NATHANIEL ROTHMAN

I

n the United States, an estimated 57,400 cases of cancer of the urinary bladder are diagnosed and 12,500 deaths from the disease occur each year (Jemal et al., 2003). These account for 6% of all new cases of cancer among men and 2% of cases among women, as well as 3% of cancer deaths among men and 1% among women. In 1998–2000, the lifetime risk of being diagnosed with bladder cancer was 3.9% and 1.4% among white and black men, respectively, which was higher than the 1.2% and 0.8% risks experienced by white and black women (Ries et al., 2003).

DESCRIPTIVE FACTORS Histopathology and Anatomic Distribution Nearly 99% of bladder cancers diagnosed in the United States are histologically confirmed (Surveillance Research Program, 2003). Most bladder cancers are transitional cell carcinomas (93%); 2% are squamous cell carcinomas and 1% are adenocarcinomas. About 38% do not have a subsite within the bladder specified, and 12% arise in more than one subsite. Of bladder cancers for which a single subsite is specified, most occur on one of the bladder walls, with two times as many on the lateral walls (21%) as on the anterior (2%) or posterior (9%) walls combined. Less common subsites include the trigone (6%), followed by the ureteric orifice, dome, and neck (5%, 4%, and 3%, respectively).

International Variation Internationally, incidence rates of bladder cancer among men vary more than fifteen-fold (Parkin et al., 2002). The highest rates occur among men in southern Europe, followed by western and northern Europe, North America, and Oceania; relatively low rates are found in eastern Europe, Central America and South America, and several areas of Asia (Fig. 58–1). The rankings of bladder cancer incidence rates among women are similar to those among men in North America, Oceania, and Asia, although the concordance is less in Europe. Maleto-female rate ratios generally range between three and five, but they were less than three in India, Thailand, and U.S. blacks. The ratios exceeded six in several areas of southern Europe, such as Spain, where the prevalence of cigarette smoking is substantially higher in men than in women, particularly cigarettes made with black tobacco.

Geographic Variation in the United States In the United States, bladder cancer mortality rates by state economic area among white men and women during 1950–74 and 1975–99 are portrayed in Figure 58–2, updated from maps presented in the Atlas of Cancer Mortality in the United States, 1950–1994 (Devesa et al., 1999). The rates for each map were ranked independently, ordered, and partitioned into quintiles. In both time periods, elevated mortality rates among men clustered in the northeastern states, especially in northern New England, New Jersey, and New York. High rates also occurred in parts of the mid-western and far western states, whereas rates generally were low across the South, except in Louisiana and Florida. Rates among women also were elevated in the Northeast, particularly in northern New England and New York, a pattern similar to

that seen in men. Bladder cancer among men has tended to cluster in the Northeast since the 1950s, particularly in areas with chemical industries (Blot and Fraumeni, 1978). Case-control studies in high-risk areas have revealed excess risks in a variety of occupations (Hoar and Hoover, 1985; Schoenberg et al., 1984; Silverman et al., 1983), including an increased risk among truck drivers and other workers exposed to motor exhausts (Silverman et al., 1986). Elevated rates for bladder cancer among men and women in northern New England and New York have become more pronounced over time, however, and do not appear to be entirely explained by occupational exposures or by regional variations in smoking habits and diet (Brown et al., 1995; Michaud et al., 2001). The geographic patterns in mortality are consistent with available incidence data, with rates in Connecticut and Detroit about 50% higher than those in Utah, New Mexico, and Hawaii (Table 58–1).

Time Trends Incidence rates of bladder cancer among men have been rising in many areas of the world, although rates may have stabilized or decreased in the last decade in some countries (Parkin, 2003). In the United States, incidence rates have risen modestly among all four race/sex groups, whereas mortality rates have declined (Fig. 58–3). Incidence rates consistently have been higher among whites than blacks, especially among men, for whom rates are about twice as high among whites as blacks. In contrast, mortality rates among women generally have been about 30% higher among blacks than whites. The observed increases in incidence may be partly explained by changes in diagnostic practice. The distinction between in situ and invasive disease may be difficult to make. The proportion of bladder tumors classified as “carcinoma in situ” in the United States increased from less than 1 percent in 1969–71 to more than 7 percent around 1980, and considerably more in recent years (Cutler et al., 1975; Hankey et al., 1991; Lynch et al., 1991; Silverman et al., 1996). Since 1988, the Surveillance, Epidemiology and End Results (SEER) Program has coded the tumor characteristics according to the American Joint Committee on Cancer (1988), and more than 70% of U.S. bladder cancers currently are described as superficial (Surveillance Research Program, 2003).

Gender and Race/Ethnicity Cancer of the bladder occurs primarily among white non-Hispanic men (Table 58–2) (Surveillance Research Program, 2003). In the United States, incidence rates among white Hispanic and black men are about 50 percent of those among white non-Hispanics. Rates are lower yet among Asians and especially low among American Indian/Alaska Natives. The higher incidence among whites compared to blacks is limited to patients with localized disease, with blacks and whites having similar risk of more advanced tumors (Schairer et al., 1988). In fact, among women, recent rates for regional and distant stage disease were higher among blacks than whites (Surveillance Research Program, 2003). The male-to-female rate ratio is 4.0 among white non-Hispanics and American Indians/Alaska Natives, 3.7 among white Hispanics and Asians, and 2.7 among blacks. The reason for this large male excess of bladder cancer is not known (Hartge et al., 1990).

1101

1102

PART IV: CANCER BY TISSUE OF ORIGIN

Rate per 100,000 person-years Men

40

Figure 58–1. International variation in ageadjusted (world standard) bladder cancer incidence rates per 100,000 person-years by gender, circa 1993–1997. (Source: Based on data from Parkin et al., 2003.)

30

20

Women

10

0

10

AMERICA, NORTH USA, SEER: White Canada USA, SEER: Black AMERICA, CENTRAL and SOUTH Ecuador, Quito Costa Rica ASIA Israel: Jews Japan, Nagasaki Prefecture China, Hong Kong Japan, Miyagi Prefecture Japan, Osaka Prefecture Singapore: Chinese Thailand, Chiang Mai Philippines, Manila India, Bangalore EUROPE- NORTHERN Denmark UK, Scotland UK, England Norway Iceland Sweden Finland EUROPE- WESTERN France, Bas-Rhin Austria, Tyrol France, Doubs The Netherlands France, Isere Switzerland, Geneva Germany, Saarland France, Ta rn Switzerland, Basel Switzerland, St Gall-Appenzell EUROPE- SOUTHERN Italy, Genoa Province Italy, To rino Italy, Ferrara Province Italy, Varese Province Spain, Navarra Italy, Romagna Italy, Parma Province Spain, Ta rragona Spain, Granada Italy, Macerata Province Malta Italy, Modena Province Spain, Albacete EUROPE-EASTERN Czech Republic Slovakia Estonia Slovenia OCEANIA Australia, South Australia, Victoria New Zealand Australia, New South Wales

Age-Specific Patterns Incidence rates rise sharply with age, although the increases are less rapid at older ages in some populations (Fig. 58–4). More than twothirds of cases occur among persons age 65 years and older. The racial differences apparent in the age-adjusted rates generally persist across the entire age range, with little cross-over of the age-specific rates. Among whites, excesses among men compared to women become more pronounced with increasing age; the male/female rate ratio exceeds three at ages 45 to 64 years and four at ages 65 years and older. In contrast to incidence, bladder cancer mortality rates among men are about 1/3 higher among whites than blacks; and among women, rates are higher among blacks than whites (Ries et al., 2003). Mortality rates are higher among blacks than whites among men under age 50 and at virtually all ages among women (Fig. 58–5).

Survival and Stage of Disease Among bladder cancer patients diagnosed during 1992–99, five-year relative survival rates, adjusted for general population mortality, ranged from 85% for white men to 53% for black women (Table 58–3) (Ries et al., 2003). Survival has improved over time, as the corre-

sponding rates ranged from 75% to 37% during 1974–1976. Relative survival rates declined by age at diagnosis among all four race/gender groups (Table 58–3). Stage of disease at diagnosis has a substantial impact on subsequent survival (Table 58–3). Among whites, those diagnosed with localized disease had five-year relative survival rates of 92% or greater, those with regional disease had survival rates of 41% to 51%, in contrast to those with distant spread whose survival rates were 6% or less. Among cases diagnosed at each stage of disease, whites experienced a better prognosis than blacks, and men tended to experience a better prognosis than women. The stage of bladder cancer at diagnosis varies by age, gender, and race. The proportion localized at diagnosis during 1992 to 1999 in the United States declined with age from 78% among patients under age 65 years to 75% among those 65 to 74 years of age and 63% among those over 84 years (Surveillance Research Program, 2003). The overall proportion localized at diagnosis ranged from 76% among white men to 50% among black women, with the proportion distant stage ranging from 3% to 7% (Ries et al., 2003). Blacks also had longer symptom duration and higher grade tumors than whites (Prout et al., 2000). The black-white survival differences are only partly explained by differences in stage at diagnosis; racial disparities in sur-

Men 1950-74

US = 9.82/100,000 10.89 - 21.03 9.41 - 10.88 7.79 - 9.40 and sparse data 6.34 - 7.78 3.28 - 6.33

Men 1975-99

US = 8.69/100,000 10.19 - 13.04 8.85 - 10.18 7.73 - 8.84 6.55 - 7.72 3.56 - 6.54

Figure 58–2. Geographic variation in U.S. bladder cancer mortality rates (age-adjusted 2000 US population) among whites by gender and state economic area, 1950–74 and 1975–99 (Based on data from the National Center for Health Statistics and the Census Bureau). Rates based on fewer than 12 deaths were placed in the middle category.

Women 1950-74

US = 3.47/100,000 3.90 - 6.08 3.44 - 3.89 3.05 - 3.43 and sparse data 2.55 - 3.04 1.42 - 2.54

Women 1975-99

US = 2.51/100,000 2.84 - 3.88 2.51 - 2.83 2.24 - 2.50 and sparse data 2.01 - 2.23 1.13 - 2.00

1104

PART IV: CANCER BY TISSUE OF ORIGIN

Table 58–1. Bladder Cancer Incidence among Whites by Area and Gender, U.S. SEER Program, 1992–2000*,† Men

Table 58–2. Bladder Cancer Incidence by Racial/Ethnic Group and Gender, U.S. SEER Program, 1992–2000*,†,a

Women

Men

Women

Area

Cases

Rates

Cases

Rates

Racial/Ethnic Group

Cases

Rate

Cases

Rate

Eleven SEER Registries Connecticut Detroit (Metropolitan) Seattle (Puget Sound) Iowa San Jose-Monterey Atlanta (Metropolitan) Los Angeles San Francisco-Oakland Utah New Mexico Hawaii

44,090 6,083 6,049 5,336 4,584 2,292 2,034 8,764 4,522 1,683 1,670 1,073

36.2 46.1 43.2 40.9 37.6 34.4 33.8 33.2 32.4 30.5 27.6 22.7

15,542 2,282 2,262 1,741 1,497 786 811 3,041 1,703 500 592 327

9.2 12.2 11.2 10.1 8.6 8.7 8.9 8.2 8.9 7.0 7.7 6.0

White Non-Hispanicb Hispanicb Black American Indian/Alaska Native Asian or Pacific Islander

39,987 32,289 1,796 1,832 83 1,865

39.6 40.8 19.4 20.4 8.4 16.5

13,789 11,044 677 1,014 25 634

9.9 10.2 5.3 7.6 2.1 4.4

*Generated using SEER *Stat (Surveillance Research Program, 2003). † Rates per 100,000 person-years, age-adjusted to the 2000 U.S. (19 age groups) standard.

vival rates persist after adjustment for stage, histologic type, grade, and socioeconomic status (Page and Kuntz, 1980; Hankey and Myers, 1987).

RISK FACTORS Tobacco Cigarettes Cigarette smoking is well established as a cause of bladder cancer, although the association is not as strong as that observed for smoking and several other cancers. An association between cigarette smoking and bladder cancer has been observed in more than 35 case-control studies and in more than ten cohort studies (Augustine et al., 1988;

Rate per 100,000 person-years

100

Incidence

100

10

Mortality

10

1

1 1975 1980 1985 1990 1995 2000

1975 1980 1985 1990 1995 2000

Year White men

Year Black men

White women

Black women

Figure 58–3. Trends in age-adjusted (2000 US population) bladder cancer, 9 SEER areas incidence and US mortality rates by race and gender, 1974–1976 through 1998–2000. (Based on unpublished data available from SEER*Stat, 2003.)

*Generated using SEER *Stat (Surveillance Research Program, 2003). † Rates per 100,000 person-years, age-adjusted to the 2000 U.S. (19 age groups) standard. a Includes San Francisco, Connecticut, Detroit, Hawaii, Iowa, New Mexico, Seattle, Utah, Atlanta, San Jose-Monterey, Los Angeles, and Alaska Native Registry. b Excludes Detroit, Hawaii, and Alaska Native Registry.

Brennan et al., 2000; Brownson et al., 1987; Burch et al., 1989; Castelao et al., 2001; Chiu et al., 2001; Claude et al., 1986; Clavel et al., 1989; D’Avanzo et al., 1990; Engeland et al., 1996; Gonzalez et al., 1985; Hartge et al., 1987; IARC, 1986; Iscovich et al., 1987; Jensen et al., 1987a; McLaughlin et al., 1995; Mills et al., 1991; Momas et al., 1994a; Nomura et al., 1989; Rebelakos et al., 1985; Schifflers et al., 1987; Sorahan et al., 1994; Tripathi et al., 2002; Vena et al., 1993a; Zeegers et al., 2002b). Overall, smokers appear to have two to three times the risk of nonsmokers. Data from correlational studies also are consistent with a smoking-bladder cancer association. In the United States, bladder cancer mortality rates at the state level are highly correlated with per capita cigarette sales (Fraumeni, 1968). Birth cohort-specific patterns of bladder cancer incidence and mortality parallel the smoking patterns of those cohorts (Armstrong and Doll, 1974; Hoover and Cole, 1971). Risk increases with increasing intensity of smoking (packs per day), with relative risk estimates for moderate-to-heavy smokers typically ranging from about 2.0 to 5.0, compared to nonsmokers (Augustine et al., 1988; Burch et al., 1989; Claude et al., 1986; Clavel et al., 1989; D’Avanzo et al., 1990; Engeland et al., 1996; Hartge et al., 1987; IARC, 1986; McLaughlin et al., 1995; Rebelakos et al., 1985; Schifflers et al., 1987; Sorahan et al., 1994; Vineis et al., 1988; Zeegers et al., 2002b). The shape of the dose-response curve has varied among the studies, however. Some have reported a regular gradient in risk with amount smoked, whereas others have reported little change in risk from moderate-to-heavy smoking levels (Augustine et al., 1988; Chiu et al., 2001; D’Avanzo et al., 1990; Hartge et al., 1987; IARC, 1986; McLaughlin et al., 1995; Schifflers et al., 1987; Sorahan et al., 1994; Vineis et al., 2000). The plateauing of the dose-response curve may be explained by a number of factors (Vineis et al., 2000). First, differential recall of heavy smoking in case-control studies may have been caused by underreporting by cases who smoke heavily. This explanation seems unlikely, however, because the plateauing is also apparent in cohort studies, which are not prone to differential recall bias. Second, heavy smokers may inhale proportionally less than light smokers, causing a leveling off of exposure. Third, variation in genetic factors that impact on carcinogen activation or detoxification or saturation of key metabolic activation processes may occur at high levels of tobacco exposure. Duration of smoking has been evaluated less often than intensity, but a regular duration-response relationship has been observed in most studies that investigated the issue (Augustine et al., 1988; Brennan et al., 2000; Burch et al., 1989; Castelao et al., 2001; Claude et al., 1986; D’Avanzo et al., 1990; Hartge et al., 1987; IARC, 1986; Sorahan et al., 1994; Vineis et al., 1988; Vineis et al., 2000; Zeegers et al., 2002b). Although the specific carcinogens responsible for the increased bladder cancer risk experienced by smokers is unknown, mounting evidence suggests that aromatic amines, rather than PAHs, are the bladder carcinogens in tobacco smoke (Vineis and Pirastu, 1997). In contrast, tar and nicotine content in cigarettes appears to have little or no impact on risk (Castelao et al., 2001; Zeegers et al., 2002b).

1105

Bladder Cancer Men

Rate per 100,000 person-years

500

Women

500

100

100

10

10

Race / ethnic group White non-Hispanic Black White Hispanic Asian/ Pacific Islander American Indian/ Alaska Native

1

0.1

1

0.1 0

20

40

60

80

100

0

20

Age

40

60

80

Age

The question of whether women are more susceptible to smokinginduced bladder cancer than men has recently received attention (Castelao et al., 2001; Perneger, 2001; Puente et al., 2006). Gender susceptibility to tobacco-induced disease is important because it could provide mechanistic insights into bladder carcinogenesis. In a large population-based case-control study in Los Angeles, Castelao and colleagues (2001) reported that relative risks were higher in women than in men who smoked comparable amounts of cigarettes. This gender disparity was bolstered by the observation that Los Angeles women who smoked had higher levels of 3- and 4-aminobiphenyl (a known bladder carcinogen) hemoglobin adducts compared to men who

100

Figure 58–4. Age-specific incidence rates for bladder cancer in the U.S. SEER Program 12 areas by racial/ethnic group and gender, 1992–2000. (Based on unpublished data available from SEER*Stat, 2003.) Note: data for Hispanics/nonHispanics not available from Detroit, Hawaii, or Alaska.

smoked comparable amounts. Despite the large number of studies of cigarette smoking and bladder cancer, few studies have estimated risk by gender and tobacco dose as measured by amount smoked, duration smoked or cumulative lifetime consumption. Based on the limited published data available, women do not appear to have consistently higher relative risks than men by tobacco dose (Table 58–4). Further, even if women did have higher relative risks than men, the absolute risks associated with smoking are likely to be higher in men because the baseline risk of bladder cancer among nonsmoking women is about one third of that among men.

Cessation 500

White men 100

Black men

Rate per 100,000 person-years

Black women White women

Cessation of cigarette smoking has been associated with a 30%–60% reduction in bladder cancer risk in many studies (IARC, 1986). The pattern of change in risk in relation to time since quitting is less clear, however (Table 58–5). Six studies suggest that the risk of former smokers who stopped smoking for many years approximates that of nonsmokers (Cartwright et al., 1983; Castelao et al., 2001; D’Avanzo et al., 1990; Sorahan et al., 1994; Tripathi et al., 2002; Wynder and Stellman, 1977). Other studies indicate that a reduction in risk occurs within the first 2–4 years after stopping, but that risk either does not continue to decline with increasing time since quitting (Augustine

10

Table 58–3. Five-Year Relative Survival Rates (Percent) among Bladder Cancer Patients by Race, Gender, Age at Diagnosis, and Stage at Diagnosis, U.S. SEER Program, 1992–99* Race/Gender 1

0.1 0

20

40

60

80

100

Age

Figure 58–5. Age-specific mortality rates for bladder cancer in the United States by race and gender, 1992–2000. (Based on unpublished data from the National Center for Health Statistics, as available from SEER*Stat, 2003.)

5-year relative survival rates (%)

Total

White Men

White Women

Black Men

Black Women

Total Age at diagnosis Under 65 65 and older Stage Localized Regional Distant Unstaged

81.8

84.6

77.0

70.1

52.9

87.3 78.6

89.2 81.8

84.5 73.4

79.4 61.0

60.6 49.1

94.4 48.2 5.8 59.4

95.4 51.4 6.0 66.8

92.1 41.3 5.9 49.2

89.9 43.2 3.4 48.5

79.7 31.4 3.1 24.1

Source: Ries et al., 2003.

1106

PART IV: CANCER BY TISSUE OF ORIGIN

Table 58–4. Relative Risks of Bladder Cancer According to Tobacco Dose and Gender

Table 58–5. Relative Risks of Bladder Cancer According to Time Since Quitting Smoking, Men

Relative Risk Reference Howe et al. (1980)

Morrison et al. (1984b)

Augustine et al. (1988)

Burch et al. (1989)

Hartge et al. (1990)

Castelao et al. (2001)

Tobacco Dose

Reference Men

Women

2.5 4.4 5.8

}

2.0 3.3

Risks relative to a risk of 1.0 for nonsmokers, adjusted for age at observation

1.4 3.2 4.7

Manchester, U.K. <1 1–<2 ≥2

1.9 3.2 4.0

Nagoya <1 1–<2 ≥2

1.6 2.1 2.8

} } }

Wynder and Stellman (1977)

Howe et al. (1980)

amount (packs/day) Boston <1 1–<2 ≥2

4.3 6.2

Relative Risk

0 1–3 4–6 7–10 11–15 16+ 0 2–15 15+

2.7 2.9 1.9 1.4 1.6 1.1 1.0 0.6 0.5

0–5 6–15 16–25 26–35 0 2–15 15+

1.7 1.0 1.1 0.9 1.0 0.6 0.4

Risks among current smokers relative to a risk of 1.0 for nonsmokers, adjusted for age at observation

2.1 2.2

Cartwright et al. (1983) Claude et al. (1986)

4.4 4.2

1.2 1.6 2.0 2.3

1.0 0.7 1.1 1.6

Risks relative to a risk of 1.0 for nonsmokers, adjusted for age at observation, amount smoked, education, race, marital status

0.8 1.1 3.4 2.2

Risks relative to a risk of 1.0 for nonsmokers, adjusted for age at observation.

2.0 1.4 2.1 3.0 4.4

Risks relative to a risk of 1.0 for nonsmokers, adjusted for age at observation

0.8 1.5 2.3 5.4 6.0

Risks relative to a risk of 1.0 for nonsmokers, adjusted for age at observation

Hartge et al. (1987)

0 1 2–4 5–9 10–19 20+

1.0 0.9 0.6 0.8 0.7 0.9

1.6 1.5 1.8 2.3

amount (packs/day)

Ex-smoker, <1 Ex-smoker, ≥1 Current, <1 Current, 1–<2 Current, ≥2

1.3 1.9 2.2 3.1 3.7

duration (yrs)

<10 10–<20 20–<30 30–<40 ≥40

1.2 1.4 2.4 3.3 4.2

Iscovich et al. (1987)

0 2–4 5–9 10–19 20+

1.0 0.6 0.3 0.2 0.2

Vineis et at. (1988)

0 <3 3–9 10+

1.0 0.4 0.4 0.6

Augustine et al. (1988)

0 £6 7–12 13+

1.0 0.7 0.7 0.7

0 >1–£5 >5–£10 10+

1.6 1.1 0.8 1.4

D’Avanzo et al. (1990)

2–4 5–14 15+

1.0 2.8 1.9

Brennan et al. (2000)

0 1–4 5–9 10–14 15–19 20–24 25+ 0 £5 >5–£15 15+

1.0 0.7 0.7 0.6 0.5 0.5 0.4 1.0 2.9 1.7 1.1

Burch et al. (1989)

et al., 1988; Burch et al., 1989; Hartge et al., 1987; Vineis et al., 1988) or continues to decline but does not appear to return to the level of nonsmokers even after 25 years of cessation (Brennan et al., 2000). In most of these studies, the effect of time since quitting was not adjusted for the effects of age at starting and duration of smoking. Hartge et al. (1987), however, estimated relative risk by length of time since quitting among intermittent former smokers (i.e., smokers who quit for at least 6 months, started again, and subsequently quit) with adjustment for age at starting and duration of smoking, as well as age at observation. Among all former smokers, the pattern of risk by time since quitting was weak and inconsistent (Table 58–5). When this analysis was restricted to intermittent former smokers, risk declined 50 percent within the first four years of stopping, but did not continue to decrease with increasing time since quitting. The almost immediate reduction in risk within the first few years after quitting suggests that cigarette smoke contains agents that act at a late stage of bladder carcinogenesis (Hartge et al., 1987).

Risks relative to a risk of 1.0 for current smokers, adjusted for age at observation and lifetime cigarette consumption Risks relative to a risk of 1.0 for nonsmokers, adjusted for age at observation Risks relative to a risk of 1.0 for current smokers, adjusted for age at observation and lifetime consumption

1.0 0.7 0.5 0.4 0.4 0.5

duration (yrs) 1–10 11–20 21–30 30+

Risks relative to a risk of 1.0 for nonsmokers, adjusted for age at observation and race

all intermittent smokers smokers

duration (yrs)

£20 21–30 31–40 40+

Comments

Comments

total lifetime consumption

<10,000 10,000–20,000 >20,000

Years since quitting

Tripathi et al. (2002)

Includes women; risks relative to a risk of 1.0 for current smokers, adjusted for age at observation, sex, race, duration (all smokers); and age at observation, sex, race, duration, age started (intermittent smokers) Risks relative to a risk of 1.0 for current smokers or those who stopped less than 2 years before diagnosis/interview, adjusted for age at observation, intensity, duration Risks relative to a risk of 1.0 for current smokers, adjusted for age at observation, duration, intensity Risks relative to a risk of 1.0 for current smokers, adjusted for age at observation, intensity, duration, education, race, and marital status Risks relative to a risk of 1.0 for nonsmokers, adjusted for age at observation and lifetime cigarette consumption Includes women; risks relative to a risk of for nonsmokers, adjusted for age at observation and sex Risks relative to a risk of 1.0 for current smokers, adjusted for age at observation and center

Risks relative to a risk of 1.0 for nonsmokers, adjusted for age at observation, pack-years of smoking, physical activity, diabetes, body mass index, alcohol intake, marital status, and occupation; women only

Bladder Cancer

Filtration People who smoke unfiltered cigarettes exclusively have been reported to experience about a 35% to 50% higher risk of bladder cancer than those who smoke filtered cigarettes exclusively (Hartge et al., 1987; Wynder et al., 1988). Other studies indicate that use of filtered cigarettes does not diminish risk (Castelao et al., 2001; Zeegers et al., 2002b). Switching to filtered cigarettes does not appear to reduce risk either (Burch et al., 1989; Hartge et al., 1987; Sorahan et al., 1994; Wynder et al., 1988). There are several possible explanations for these inconsistent findings. First, people who smoke filtered cigarettes exclusively may have different smoking histories or habits than do people who first smoked unfiltered cigarettes. For example, the latter group may start smoking earlier, or take more puffs of smoke per cigarette. Second, the effect of changing from unfiltered to filtered cigarettes may be quite small, given the small difference between the risk for smokers of only filtered cigarettes and that for smokers of only unfiltered cigarettes. Third, interview data on changing from unfiltered to filtered cigarettes may contain inaccuracies that mask a real, but small, reduction in excess risk. Fourth, smokers of only filtered cigarettes may not, in fact, have a lower risk than smokers of only unfiltered cigarettes; any observed reduction in risk may have been a chance effect.

1107

Howe et al., 1980; Jensen et al., 1987a; Kahn, 1966; Kunze et al., 1992; Williams and Horm, 1977; Wynder et al., 1963; Wynder and Goldsmith, 1977a). In the positive studies, relative risks for cigar smokers compared to nonsmokers varied from about 1.3 to 3.6. Risks associated with the use of snuff or chewing tobacco have been assessed in a small number of studies (Burch et al., 1989; Castelao et al., 2001; Cole et al., 1971; Hartge et al., 1985; Howe et al., 1980; Jensen et al., 1987; Kahn, 1966; Kunze et al., 1992; Mommsen and Aagaard, 1983a; Williams and Horm, 1977; Wynder et al., 1963). Of those studies, an increased risk of bladder cancer for snuff users who never smoked cigarettes has been observed in only one (Slattery et al., 1988a), and for users of chewing tobacco in only two (Mommsen and Aagaard, 1983a; Slattery et al., 1988a).

Environmental Tobacco Smoke Experimental evidence suggests that environmental tobacco smoke (ETS) may increase the risk of bladder cancer in nonsmokers (Zeegers et al., 2002b). In addition, mutagens have been found in the urine, as well as in the blood, of passive smokers. Few epidemiologic studies have examined the effect of ETS on bladder cancer risk, however. No increased risk has been associated with passive smoking in two casecontrol studies (Burch et al., 1989; Sandler et al., 1985) and one cohort study (Zeegers et al., 2002b) conducted to date.

Inhalation Cigarette smokers who inhale deeply may have a greater risk than those who do not (Burch et al., 1989; Clavel et al., 1989; Cole et al., 1971; Lopez-Abente et al., 1991; Zeegers et al., 2002b). Morrison et al. (1984b) observed a 30% to 40% elevation in risk for male cigarette smokers who inhaled deeply compared to those who inhaled somewhat or not at all. An association between inhalation and risk has not been observed, however, in some other studies (Castelao et al., 2001; Hartge et al., 1987; Howe et al., 1980; Lockwood, 1961).

Black vs. Blond Tobacco Smokers of black tobacco seem to have a risk of bladder cancer about one and one half to three times higher than the risk in smokers of blond tobacco (Clavel et al., 1989; D’Avanzo et al., 1990; De Stefani et al., 1991; Iscovich et al., 1987; Momas et al., 1994a; Vineis et al., 1984; Vineis et al., 1988). Three laboratory observations support this epidemiologic observation. First, black tobacco has higher concentrations of aromatic amines, some of which are human bladder carcinogens, than does blond tobacco (Vineis et al., 1988). Second, blood levels of 4-aminobiphenyl hemoglobin adducts, as well as adducts of several other aromatic amines, are higher for smokers of black than of blond tobacco (Bryant et al., 1988). Third, the urine of smokers of black tobacco is more mutagenic than is the urine of smokers of blond tobacco (Malaveille et al., 1989; Mohtashamipur et al., 1987).

Pipes, Cigars, and Smokeless Tobacco The roles of pipes, cigars, snuff, and chewing tobacco in the etiology of bladder cancer are unclear. Evidence of increased risk is strongest for pipe smokers, particularly those who never smoked any other type of tobacco. At least a dozen studies have suggested that pipe smokers experience elevated risk compared to nonsmokers (relative risks typically ranged from 1.3 to 3.9) (Burch et al., 1989; Castelao et al., 2001; Claude et al., 1986; Cole et al., 1971; Engeland et al., 1996; Hartge et al., 1985; Howe et al., 1980; Jensen et al., 1987a; Lockwood, 1961; Mommsen and Aagaard, 1983a; Morrison et al., 1984b; Pitard et al., 2001; Slattery et al., 1988a; Sorahan et al., 1994; Williams and Horm, 1977; Wynder et al., 1963). A dose-response relationship has been found only rarely (Pitard et al., 2001), although pipe smokers who inhale deeply do appear to be at greatest risk (Hartge et al., 1985; Howe et al., 1980). Weak and inconsistent relationships have been observed between bladder cancer risk and the other forms of tobacco use. For cigars, some studies have been positive (Hartge et al., 1985; Lockwood, 1961; Mommsen and Aagaard, 1983a; Pitard et al., 2001; Shapiro et al., 2000; Slattery et al., 1988a) whereas others have shown little or no association (Burch et al., 1989; Castelao et al., 2001; Cole et al., 1971;

Occupation Following a number of clinical observations and mortality surveys, the study of occupational causes of bladder cancer gained momentum in the 1950s with the identification of bladder cancer hazards in the British dyestuffs and rubber industries (Case et al., 1954b; Case and Hosker, 1954a). During the subsequent five decades, scores of studies have suggested approximately 40 potentially high-risk occupations. Despite this effort, the relations of many of these occupations to bladder cancer risk are unclear. Observed relative risks typically have been less than two, based on a small number of exposed subjects. Further, many reported associations have not been consistently found (Silverman et al., 1989b). Strong evidence of increased risk is apparent for very few occupational groups and the bladder carcinogens responsible for observed increased risks are often unknown.

Dyestuffs Workers and Dye Users In 1895, Rehn suggested that men employed in the dyestuffs industry had increased risk of bladder cancer. It was not until 1954, however, that Case et al. (1954b) showed that dyestuffs workers in England and Wales had a 10- to 50-fold increased risk of death from bladder cancer due to exposure to two aromatic amines, 2-naphthylamine and benzidine. Exposure to a third aromatic amine, 1-naphthylamine, also appeared related to risk, but this elevation may have been caused by contamination with 2-naphthylamine. No excess risk was associated with exposure to aniline. Two reports based on a cohort of dyestuffs workers in northern Italy (Decarli et al., 1985; Rubino et al., 1982) confirmed the increased risk from exposure to 2-naphthylamine and benzidine. A positive trend in bladder cancer mortality with increasing duration of employment was apparent; observed/expected ratios of 13, 34, and 71 were associated with employment as a dyestuffs worker for 10 years or less, 11 to 20 years, and more than 20 years, respectively. Dyestuffs workers involved in fuchsin and safranine T manufacturing also experienced high mortality (observed/expected = 62.5), which may have been the result of exposure to two precursors, o-toluidine and 4,4¢-methylene bis(2-methylaniline). In the United States, workers involved in the manufacture of azo dyes and their intermediates also experienced excess mortality (3 observed/0.25 expected) (Delzell et al., 1989), but the magnitude of risk appears to be lower than that for dyestuff workers in the United Kingdom and Italy. The increased risk among dyestuffs workers has also been observed in case-control studies (Boyko et al., 1985; La Vecchia et al., 1990; Morrison et al., 1985; Najem et al., 1982; Puntoni et al., 1988; Risch et al., 1988b; Vineis and Magnani, 1985), with relative risks ranging from 1.7 to 8.8. Data from the United Kingdom indicate that bladder

1108

PART IV: CANCER BY TISSUE OF ORIGIN

cancer risk among dyestuffs workers has been reduced since the introduction of protective measures and the subsequent banning of the industrial use of 2-naphthylamine and benzidine in 1950 and 1962, respectively (Boyko et al., 1985; Morrison et al., 1985). Studies of the Italian cohort of dyestuffs workers have provided additional information on temporal patterns of risk (Decarli et al., 1985; Piolatto et al., 1991; Rubino et al., 1982). First, the mean time from start of exposure to death was 25 years, with a range of 12 to 41 years. Second, an inverse relationship between age at first exposure and risk was observed; risk was greatest for workers who started before age 25 years (observed/expected = 200.0). Third, a negative trend in relative risk with increasing time since last exposure was seen. Users of finished dyes also may have an increased risk of bladder cancer, but the evidence is not as persuasive as that for dyestuffs manufacturing workers. Kimono painters, many of whom ingest benzidine-based dyes by licking the brush, have been found to have seven times the expected rate of bladder cancer (Yoshida et al., 1971). Coarse fishermen who use chrysolidine azo dyes to stain maggot bait have been reported to have an increased risk of bladder cancer (relative risk = 3.0 for fishermen who used bronze dyes for five or more years) (Sole and Sorahan, 1985), although a more recent case-control study did not confirm this observation (Sorahan and Sole, 1990). Canadian dyers of cloth were reported to have a relative risk of 4.6 (Risch et al., 1988b), and British textile dyers with more than 20 years employment had a relative risk of 3.4 (Antony, 1974). Two other studies, however, found no excess risk for dye users (Cartwright et al., 1979b; Silverman et al., 1989b).

Aromatic Amine Manufacturing Workers Evidence that 2-naphthylamine and benzidine are human bladder carcinogens extends beyond the dyestuffs industry into the chemical industry where these aromatic amines, as well as a third bladder carcinogen, 4-aminobiphenyl, were manufactured (IARC, 1987). A fourfold risk was observed among 2-naphthylamine-exposed chemical workers in the United States (Schulte et al., 1986). The observation of an increased risk of bladder cancer among workers involved in the commercial preparation of 4-aminobiphenyl resulted in the discontinuation of production of this aromatic amine, thus averting its widespread use (IARC, 1987). In a cohort of workers at a benzidine manufacturing facility, an overall excess of bladder cancer cases was apparent (SIR = 343) (Meigs et al., 1986). Risk was greatest among those in the highest exposure category (SIR = 1303); little or no excess was observed for those in the low or medium exposure categories. Corresponding to the introduction of preventive measures in the plant, a reduction in risk was observed for those first employed in 1950 or later compared to those first employed in 1945–1949. In a cohort of benzidine-exposed workers in China, an overall SIR of 25 was reported, with risk ranging from 4.8 for those with low exposure to 158.4 for those with high exposure (Bi et al., 1992). Risks were elevated for both producers of benzidine (SIR = 45.7), as well as for users of benzidine-based dyes (SIR = 20.9). Two structural analogues of benzidine, MDA (4,4¢-methylenedianiline) and MBOCA (4,4¢-methylene-bis(2-chloroaniline) ), are carcinogenic in animals (Schulte et al., 1987), and possibly in humans as well. MDA, a curing agent for certain resins, was associated with a three-fold elevation of proportional mortality from bladder cancer (Schulte et al., 1987). MBOCA, a curing agent used in the manufacture of rigid plastics, has been suggested as the exposure responsible for two noninvasive papillary tumors of the bladder in workers in a MBOCA production plant, although no invasive bladder tumors have been identified in the cohort (Ward et al., 1988). Manufacturing of another aromatic amine, 4-chloro-o-toluidine (4-COT), has been associated with excess bladder cancer mortality in a cohort of chemical workers in Germany (relative risk = 72.7) (Stasik, 1988). This large excess in bladder cancer mortality was confirmed in another cohort of German chemical workers exposed to 4-COT (Popp et al., 1992). In New York State, a cohort of chemical workers exposed to both otoluidine and aniline also experienced elevated risk of bladder cancer (SIR = 360), which was probably attributable to exposure to otoluidine (Ward et al., 1991).

Rubber Workers Antioxidants containing 2-naphthylamine were used in the rubber and electric-cable manufacturing industries in Great Britain (Baus Subcommittee, 1988). Case and Hosker (1954a) observed that the bladder cancer mortality among British rubber workers was twice the expected level. This excess was observed only among rubber workers employed before 1950; 2-naphthylamine was withdrawn from use in the British rubber industry in 1949 (Parkes et al., 1982). Excess risk of bladder cancer also has been reported among rubber workers in the United States (Alderson, 1986; IARC, 1982), Italy (Negri et al., 1989), Sweden and Germany (Kogevinas et al., 1998), although a few studies of rubber workers found no excess (Jensen et al., 1987b; McMichael et al., 1976; Morrison et al., 1985; La Vecchia et al., 1990). The elevation of risk reported in most American studies (Checkoway et al., 1981; Delzell and Monson, 1981; Monson and Nakano, 1976) is less than that reported in the British and Italian studies. There was little exposure to 2-naphthylamine in the U.S. rubber industry (Checkoway et al., 1981), but many workers were exposed to other possible carcinogens (Kogevinas et al., 1998) including another antioxidant, phenyl-b-naphthylamine (PBNA), which can be metabolized to 2-naphthylamine (IARC, 1987).

Leather Workers An increased risk of bladder cancer among leather workers has been observed in at least 14 studies (Baxter and McDowall, 1986; Brown et al., 1995; Chen, 1990; Cole et al., 1972; Costantini et al., 1989; Decoufle, 1979; Dolin and Cook-Mozaffari, 1992; Garabrant and Wegman, 1984; Henry et al., 1931; Marrett et al., 1986; Montanaro et al., 1997; Morrison et al., 1985; Vineis and Magnani, 1985; Wynder et al., 1963), although no increased risk was observed in three studies of leather tanners (Edling et al., 1986; Mikoczy et al., 1994; Stern et al., 1987). Most of the positive results are from case-control studies; the relative risk varied from 1.4 to 6.3. The definition of “leather worker” was not consistent among studies. Some reported increased risks for shoe makers and shoe repairers (Dolin and Cook-Mozaffari, 1992; Garabrant and Wegman, 1984; Wynder et al., 1963), whereas others reported elevations for workers in leather products manufacturing (Decoufle, 1979) or, more broadly, for workers exposed to leather or leather products (Cole et al., 1972; Marrett et al., 1986; Morrison et al., 1985). The exposure responsible for the increased risk among leather workers in not known. Cole et al. (1972) reported that the excess was associated with jobs that involved finishing and related processes, including cutting and assembling leather pieces. In a large case-control study in 10 areas of the United States (Marrett et al., 1986), risk was found to be slightly higher for workers with possible exposure to leather dust compared to other types of leather exposure. In addition to leather dust, leather workers also are exposed to dyes, their solvents, and unreacted intermediates (Risch et al., 1988b). Bladder cancer excesses among Italian leather tannery workers have been linked to exposure to benzidine-based leather dyes (Montanaro et al., 1997). Identification of carcinogens in the leather industry may require chemical analysis of substances encountered in the industry in combination with biologic monitoring of workers (Marrett et al., 1986).

Painters Bladder cancer risk has been elevated among painters in many studies (Barbone et al., 1994; Bethwaite et al., 1990; Chen and Seaton, 1998; Claude et al., 1988; Cole et al., 1972; Decoufle et al., 1977; Dolin and Cook-Mozaffari, 1992; Guberan et al., 1989; Henry et al., 1931; Jensen et al., 1987b; Kogevinas et al., 2003; La Vecchia et al., 1990; Malker et al., 1987; Miller et al., 1986; Morrison et al., 1985 (Boston); Myslak et al., 1991; Silverman et al., 1989b; Steenland and Palu, 1999; Terstegge et al., 1995; Teschke et al., 1997; Wynder et al., 1963; Zheng et al., 2002), although a few studies have suggested no excess risk (Englund, 1980; Morrison et al. 1985 (Manchester, United Kingdom, and Nagoya, Japan) ). Most of the observed relative risks have been 1.2 to 1.5. Jensen et al. (1987b) reported a positive trend in risk with increasing duration of employment; painters employed 20 years or

1109

Bladder Cancer more had a relative risk of 4.1. In a large case-control study in the United States (Silverman et al., 1989b), painters experienced a 50 percent increased risk. Among those who started working before 1930, a trend in risk with increasing duration of employment was apparent; the relative risk for such painters employed ten years or more was 3.0. Painters may be exposed to many known or suspected carcinogens in paints (e.g., benzidine, polychlorinated biphenyls, formaldehyde, and asbestos) and solvents (e.g., benzene, dioxane, and methylene chloride) (Miller et al., 1986). Painters, however, use most of these suspected carcinogens in specialized settings and are typically exposed at low levels (Steenland and Palu, 1999). The specific exposure(s) responsible for the increased bladder cancer risk experienced by painters remains unknown.

Drivers of Trucks and Other Motor Vehicles Excess risk of bladder cancer has been observed frequently among drivers of trucks, buses, or taxi cabs (Baxter and McDowall, 1986; Claude et al., 1986; Colt et al., 2004; Decoufle et al., 1977; Dubrow and Wegman, 1984; Hoar and Hoover, 1985; Iscovich et al., 1987; Jensen et al., 1987b; Kogevinas et al., 2003; Kunze et al., 1992; Logan, 1982; Milham, 1976; Momas et al., 1994b; Porru et al., 1996; Schifflers et al., 1987; Siemiatycki et al., 1994; Silverman et al., 1983, 1986; Soll-Johanning et al., 1998; Steenland et al., 1987), although one Swedish study found no elevation in risk for truck drivers (Malker et al., 1987). Overall relative risks varied from 1.3 to 2.2. A positive trend in risk with increasing duration of employment was observed for drivers in most studies, with relative risks for long-term drivers ranging from 2.2 to 12.0 (Claude et al., 1988; Hoar and Hoover, 1985; Jensen et al., 1987b; Silverman et al., 1983, 1986; Steenland et al., 1987). In the largest study of bladder cancer among truck drivers, the trend in risk by duration of employment was most consistent for those first employed at least 50 years before observation, with relative risks of 1.2, 1.4, 2.1, 2.2 for less than 5 years, 5 to 9 years, 10 to 24 years, and 25 years or more, respectively (Silverman et al., 1986). Although the specific exposure responsible for the elevation of risk among drivers has not been identified, one likely candidate is motor exhaust. Exhaust emissions contain polycyclic aromatic hydrocarbons (PAHs) and nitro-PAHs, which are highly mutagenic, as well as carcinogenic in laboratory animals (Silverman et al., 1986). Diesel engine exhaust, in particular, has been associated with increased bladder cancer risk (Kogevinas et al., 2003; Silverman et al., 1983). In a meta-analysis of diesel exhaust exposure and bladder cancer, overall relative risks ranged from 1.1 to 1.3 and the summary relative risk for heavily exposed workers was 1.4 (Boffetta and Silverman, 2001). Another possible explanation for the elevated bladder cancer risk observed among truck drivers is urinary stasis caused by infrequent micturition during working hours (Silverman et al., 1986).

Aluminum Workers Wigle (1977) suggested that an elevated incidence of bladder cancer among men in the Chicoutimi census division of the Province of Quebec was the result of exposures incurred in the aluminum refining industry. Subsequently, increased bladder cancer mortality was observed in at least four cohort studies of aluminum smelter workers (Gibbs, 1985; Rockette and Arena, 1983; Romundstad et al., 2000; Spinelli et al., 1991). The elevated risk in the aluminum industry has been associated with employment in the Soderberg potrooms (relative risk = 2.4) (Theriault et al., 1981; Theriault et al., 1984; Tremblay et al., 1995). Risk increased with increasing duration of employment in this department. Relative risks were 1.0 for less than 1 year, 1.9 for 1–9 years, 3.0 for 10–19 years, 3.2 for 20–29 years, and 4.5 for 30 years or more (Theriault et al., 1984). Tremblay et al. (1995) used historical data on workplace exposures to better quantify exposureresponse relationships (Table 58–6). Coal-tar pitch volatiles emitted from anodes in the Soderberg electrolytic reduction process may be responsible for the observed bladder cancer excess (Theriault et al., 1984). The bladder carcinogens within tar volatiles are unknown, but aromatic amines (particularly 2-naphthylamine and 4-aminobiphenyl) and particulate PAHs and nitro-PAHs are suspect (Romundstad et al., 2000; Tremblay et al., 1995).

Table 58–6. Relative Risks of Bladder Cancer by Cumulative Exposure to BSM and to BaP for Aluminum Plant Workers, Jonquière, Quèbec, 1970–88 Exposure

Relative Riska (95% C.I.)

Cases

Controls

BSM (mg/m3-years) 0–0.9 1.0–9.9 10.0–19.9 20.0–29.9 30.0+

22 32 23 35 26

136 146 47 40 45

1.0 1.67 3.93 7.31 5.18

(0.89–3.16) (1.85–8.49) (3.56–14.99) (2.47–10.89)

BaP (mg/m3-years) 0–9.9 10.0–99.9 100.0–199.9 200.0–299.9 300+

35 29 26 30 18

215 96 32 38 33

1.0 1.97 6.24 6.66 4.36

(1.10–3.51) (3.00–12.97) (3.42–12.99) (2.10–9.17)

Source: Tremblay et al., 1995. Reprinted by permission of Wiley-Liss, Inc., a subsidiary of John Wileys & Sons, Inc. a Estimates of risk were adjusted for smoking, relative to a risk of 1.0 for subjects in the lowest exposure category. BSM = benzene-soluble matter; BaP = Benzo-a-pyrene; RR = relative risk; CI = confidence interval.

Other Occupations and Exposures Employment as a machinist has been associated with bladder cancer risk in many studies (Silverman et al., 1989b; Tolbert, 1997), although the increase in risk has not been consistently linked to a specific type of work. Machinists are exposed to mists from oils used as coolants and lubricants in metal machining processes (Silverman et al., 1983; Vineis and Di Prima, 1983). Some cutting and lubricating oils contain potentially carcinogenic PAHs (Silverman et al., 1983) and nitrosamines (Fan et al., 1977). Increased risk of bladder cancer has also been reported for many other occupational groups: metal workers, printers, chemical workers (other than those involved in manufacturing aromatic amines), hairdressers, dry cleaners, carpenters, construction workers, miners, gas workers, coke plant workers, auto mechanics, petroleum workers, railroad workers, textile workers, tailors, engineers, butchers, clerical workers, cooks and kitchen workers, food processing workers, electricians, gas station attendants, medical workers, pharmacists, glass processors, nurserymen, photographic workers, security guards and watchmen, welders, sailors, stationary firemen or furnace operators, stationary engineers, paper and pulp workers, roofers, gardeners, bootblacks, and asbestos workers (Alderson, 1986; Matanoski and Elliott, 1981; Silverman et al., 1989a, 1989b), Findings for most of these occupations are not as persuasive as those discussed earlier, and require corroboration. Strong evidence of human bladder carcinogenicity exists for occupational exposure to certain aromatic amines (IARC, 1987). In addition, many other occupational exposures are suspected of causing bladder cancer, including PAHs (Clavel et al., 1994; Kogevinas et al., 2003; Zeegers et al., 2001b); diesel engine exhaust (Boffetta and Silverman, 2001; Kogevinas et al., 2003; Silverman et al., 1983); leather dust (Siemiatycki et al., 1994); mineral oils (Hours et al., 1994; Tolbert, 1997); combustion and pyrolysis products from natural gas and other nonspecified substances (Hours et al., 1994; Siemiatycki et al., 1994; Tolbert, 1997); chlorinated solvents (Cordier et al., 1993), particularly those used in dry cleaning (Weiss, 1995); creosote (Kogevinas et al., 2003; Steineck et al., 1989); herbicides/pesticides (La Vecchia et al., 1990; Silverman et al., 1989b) and asbestos (Silverman et al., 1989b). Further research is needed to determine the carcinogenicity of these occupational exposures.

Occupational Risk in Women Occupational hazards among women have been examined in a small number of bladder cancer studies conducted in various parts of the world (Colt et al., 2004; Maffi and Vineis, 1986; Pelucchi et al., 2002a; Silverman et al., 1990; Simpson et al., 1999; Swanson and Burns, 1995; Mannetje et al., 1999). Patterns of risk by occupation in women

1110

PART IV: CANCER BY TISSUE OF ORIGIN

tend to be similar to those observed among men. For example, increased risks have been reported for female metal workers, chemical workers and rubber workers (’t Mannetje et al., 1999; Pelucchi et al., 2002a; Silverman et al., 1990; Simpson et al., 1999). In addition, employment in computer manufacturing was newly identified as a high-risk industry in Detroit women (Swanson and Burns, 1995). Reported estimates of the proportion of bladder cancer attributable to occupational exposures in women are: 6% in Italy (Pelucchi et al., 2002a), 8% in Spain (’t Mannetje et al., 1999) and 11% in the United States (Silverman et al., 1990). The relation between occupation and bladder cancer risk is dynamic (Silverman et al., 1989b). With the elimination of bladder carcinogens from the workplace and the advent of new chemicals, changing worker exposures are generating shifts in “high-risk occupations.” For example, risks among rubber and leather workers have diminished over time (Marrett et al., 1986; Parkes et al., 1982), whereas new highrisk occupations, such as truck drivers and aluminum smelter workers (Silverman et al., 1986; Tremblay et al., 1995) have emerged. Thus, occupational bladder cancer continues to be a public health problem, with risks changing over time and from population to population.

Drinking Water and Fluid Intake Arsenic Ingestion of high levels of inorganic arsenic is a known cause of bladder cancer (National Research Council, 1999). A relation between exposure to high levels of arsenic in artesian well water and bladder cancer mortality was initially suggested by surveys conducted in an endemic area of chronic arsenic toxicity, manifested by skin cancer and blackfoot disease, in Taiwan (Chen et al., 1985; Chen et al., 1986; Chiang et al., 1993; Wu et al., 1989). These findings have been confirmed by ecologic studies in Argentina (Hopenhayn-Rich et al., 1996) and Chile (Smith et al., 1998) and by cohort studies in Taiwan (Chiou et al., 1995) and Japan (Tsuda et al., 1995). Ingestion of arsenic also has been associated with increased bladder cancer mortality among a cohort of patients treated with Fowler’s solution (potassium arsenite) (Cuzick et al., 1992). In addition, biologic evidence supporting the hypothesis that chronic ingestion of high levels of arsenic is carcinogenic to the bladder is provided by observations of increased micronuclei in exfoliated bladder cells of exposed individuals (Moore et al., 1997; Warner et al., 1994) and increased chromosomal instability in bladder tumors of patients with high exposure levels (Moore et al., 2002). Data on the effect of low-to-moderate levels (i.e., <200 mg/liter) of arsenic injection are limited and results to date have been equivocal (Bates et al., 1995; Bates et al., 2004; Cantor, 2001; Karagas et al., 2004; Michaud et al., 2004; Steinmaus et al., 2003; National Research Council, 1999). Results from a recent cohort study in northeastern Taiwan suggested that low-to-moderate levels of arsenic ingestion increase risk of transitional cell carcinoma of the bladder (Chiou et al., 2001). Relative risks were 1.9, 8.2, and 15.3 for arsenic concentrations of 10.1–50.0, 50.1–100, and >100 mg/liter, respectively, compared with the referent level of 10.0 or less mg/liter.

Disinfection By-Products An association between disinfection by-products (DBPs) in drinking water and bladder cancer risk was first suggested by ecologic studies (National Research Council, 1980), and later by two case-control studies based on death certificates (Crump and Guess, 1982). In recent years, evidence that high intake of DBPs in drinking water may increase bladder cancer risk has mounted. In six investigations, detailed information was available on water quality and temporal aspects of exposure. These studies support the association between DBP levels in drinking water sources and bladder cancer risk. In Washington County, Maryland, residents supplied with chlorinated surface water had higher bladder cancer incidence rates than did those who consumed unchlorinated deep well water (relative risks were 1.8 and 1.6 for men and women, respectively) (Wilkins and Comstock, 1981). In a subsequent nested case-control study in Washington County, bladder cancer risk was weakly associated with duration of exposure

to municipal water, with an nonsignificant relative risk of 1.4 for subjects with more than 40 years of exposure (Freedman et al., 1997). In an NCI study conducted in 10 areas of the United States, risk increased with level of intake of beverages made with tap water (Cantor et al., 1987). The gradient was restricted to subjects with at least 40 years of exposure to chlorinated surface water and was not observed among long-term consumers of nonchlorinated ground water. Among subjects whose residences were served by a chlorinated surface water source for at least 60 years, a relative risk of 2.0 was estimated for heavy consumers compared to low consumers of tap water. In a Colorado study, years of exposure to chlorinated surface water was significantly associated with increased bladder cancer risk (McGeehin et al., 1993). The relative risk for those exposed for more than 30 years to chlorinated surface water was 1.8 compared to subjects with no exposure. In Ontario, Canada, bladder cancer risk increased with both duration and concentration of exposure to DBPs, with relative risk of 1.63 for subjects exposed to a trihalomethane level of at least 50 ug/liter for 35 or more years (King and Marrett, 1996). In an Iowa study, risk increased with duration of chlorinated-surface water use, with the relative risk reaching 1.5 for those with at least 60 years of exposure (Cantor et al., 1998). In both the Iowa and Washington County casecontrol studies, cigarette smoking appeared to enhance the effect of exposure to DBPs (Cantor, 1997). In a meta-analysis of six casecontrol studies and two cohort studies, the combined relative risk for long-term consumption of chlorinated drinking water was 1.4 (Villanueva et al., 2003). Similarly, results of a pooled analysis of six case-control studies indicated that risk increases with increasing exposure to trihalomethanes (a marker of DBPs), with a relative risk 1.44 for men exposed to more than 50 mg/liter (Villanueva et al., 2004). No increased risk was found among women with trihalomethane exposure, however.

Fluid Intake Total fluid intake may be related to bladder cancer risk, but the results have been equivocal. Increased total fluid consumption has been associated with decreased risk (Dunham et al., 1968; Michaud et al., 1999a; Pohlabeln et al., 1999 (women)), with a positive trend in risk (Cantor et al., 1987; Claude et al., 1986; Jensen et al., 1986; Vena et al., 1993b) and with no effect (Bruemmer et al., 1997; Cantor et al., 1998; Geoffroy-Perez and Cordier, 2001; Slattery et al., 1988b; Wynder et al., 1963; Zeegers et al., 2001d). These inconsistencies may, in part, be explained by several factors. First, differences in the definition of “total fluid intake” are apparent across studies. In some studies, “total fluid intake” includes all beverages, regardless of whether the beverage was made with tap water, whereas some investigators limit “total fluid intake” to tap water and beverages made with tap water, such as coffee and tea. Second, study populations vary with regard to quality of the tap water consumed. For example, the level of DBPs in tap water can differ substantially among study populations. Thus, a protective effect for fluid intake observed in a study population consuming fluids made with tap water with low levels of DBPs and an increased risk for fluid intake observed in a study population with high levels of DBPs are not necessarily inconsistent effects. Third, cutpoints for “total fluid intake” are typically based on quartiles or quintiles of total fluid intake in the control group (case-control studies) or total cohort (cohort studies). Because study populations vary widely with regard to reported total fluid intake and studies vary in their methodologic approaches, comparability in exposure levels among studies is less than optimal.

Dietary Factors Coffee Drinking An association between coffee drinking and bladder cancer was first suggested by a population-based, case-control study conducted in Massachusetts (relative risk = 1.3 for men and 2.5 for women) (Cole et al., 1971). Since that report, many studies have evaluated this association. More than ten studies indicated little or no overall association in either gender (Cartwright et al., 1981; Ciccone and Vineis, 1988; Gonzalez et al., 1985; Jensen et al., 1986; Kabat et al., 1986; Morrison

Bladder Cancer et al., 1982a; Piper et al., 1986 (women only); Pujolar and Gonzalez, 1993; Rebelakos et al., 1985; Slattery et al., 1988b (men only); World Cancer Research Fund, 1997); nine studies were positive for men, but not for women (Bross and Tidings, 1973; Clavel and Cordier, 1991; Fraumeni et al., 1971; Hartge et al., 1983; Howe et al., 1980; Mettlin and Graham, 1979; Wynder and Goldsmith, 1977a; Vena et al., 1993a (men only); Zeegers et al., 2001d); four studies were positive for women, but not for men (Miller et al., 1978; Morgan and Jain, 1974; Risch et al., 1988b; Simon et al., 1975 (women only)); three studies were positive for both men and women (Donato et al., 1997; Kunze et al., 1992; Pohlabeln et al., 1999) and six studies suggested an overall positive association, but sex-specific risks were not examined (D’Avanzo et al., 1992; Iscovich et al., 1987; La Vecchia et al., 1989; Mills et al., 1991; Momas et al., 1994b; Najem et al., 1982). In most of the studies reported as positive, however, the relative risk of bladder cancer in coffee drinkers compared to nondrinkers has been less than two. A regular dose-response relationship has been observed only infrequently (Bross and Tidings, 1973; Clavel and Cordier, 1991; Iscovich et al., 1987; Kunze et al., 1992; Momas et al., 1994b; Piper et al., 1986; Vena et al., 1993a; Wynder and Goldsmith, 1977a), although risk was elevated among drinkers of large amounts of coffee in several studies (Chyou et al., 1993; Donato et al., 1997; Hartge et al., 1983; Jensen et al., 1986; Morrison et al., 1982a; Rebelakos et al., 1985; Stensvold and Jacobsen, 1994). The weakness and inconsistency of the observed associations indicate that if coffee is a bladder carcinogen, it is a weak one. Alternatively, associations between coffee drinking and bladder cancer could be the result of residual confounding by smoking (Hartge et al., 1983; Morrison et al., 1982a; Morrison, 1984a). Because cigarette smoking is both an important risk factor for bladder cancer and a strong correlate of coffee drinking, tight control for smoking is required to estimate the bladder cancer risk associated with coffee drinking alone. Although relative risk estimates in nearly all cited studies were adjusted for smoking, adjustment may have been inadequate if smoking categories were too broad. Confounding by smoking also could be introduced by inaccurate recall of smoking habits. In this instance, it might not be possible to completely control the effect of smoking in estimating the risk of bladder cancer associated with coffee drinking. To avoid residual confounding by smoking, the effect of coffee drinking on bladder cancer risk can be evaluated in lifelong nonsmokers. Some studies have been large enough to have had adequate numbers of nonsmokers to estimate this risk with reasonable precision. Of these, some indicated no increased risk associated with coffee drinking (Bross and Tidings, 1973; Howe et al., 1980; Kabat et al., 1986; Morrison et al., 1982a) whereas others suggested an increased risk (Ciccone and Vineis, 1988; Clavel and Cordier, 1991; D’Avanzo et al., 1992; Donato et al., 1997; Hartge et al., 1983; Mills et al., 1991; Pujolar and Gonzalez, 1993; Rebelakos et al., 1985; Risch et al., 1988a; Sala et al., 2000; Slattery et al., 1988b; Woolcott et al., 2002; Vena et al., 1993a). Of the positive studies that distinguished between men and women in examining the coffee drinking effect, only two are positive in both men and women (Rebelakos et al., 1985; Sala et al., 2000), four are positive in men but not in women (Ciccone and Vineis, 1988; Clavel and Cordier, 1991; Hartge et al., 1983; Vena et al., 1993a (men only)), and one is positive in women but not in men (Risch et al., 1988a).

1111

appeared related to saccharin consumption (Armstrong and Doll, 1974). Bladder cancer incidence among the Danish population born during World War II, a group with higher in utero saccharin exposure than previous birth cohorts, was not increased in either men or women during the first 30 to 35 years of life (Jensen and Kamby, 1982). Several case-control studies have provided data on the relationship between artificial sweeteners and bladder cancer. Results of most studies have been negative (Cartwright et al., 1981; Iscovich et al., 1987; Jensen et al., 1983; Kabat et al., 1986; Kessler and Clark, 1978; Momas et al., 1994b; Morgan and Jain, 1974; Morrison and Buring, 1980; Najem et al., 1982; Nomura et al., 1991; Piper et al., 1986; Risch et al., 1988a; Simon et al., 1975; Slattery et al., 1988b; Vena et al., 1992; Wynder and Goldsmith, 1977a; Wynder and Stellman, 1977b). One study suggested a positive association in men (relative risk = 1.6) (Howe et al., 1977; Miller and Howe, 1977), but there was an inverse association in women (relative risk = 0.6). Moreover, a weak inverse association between use of artificially-sweetened beverages and bladder cancer was apparent in both men and women. In a large U.S. population-based, case-control study, the relative risk for subjects who had ever used artificial sweeteners was 1.0 (Hoover and Strasser, 1980). Those who reported very frequent use of artificial sweeteners appeared to have a small elevation in risk, but the dose-response pattern was irregular. A positive association was observed in two study subgroups, white male heavy smokers and nonsmoking white females with no known exposure to bladder carcinogens. The reason for these associations is uncertain, however (Hoover and Hartge, 1982; Walker et al., 1982). It is difficult to separate the effects of saccharin and cyclamates in the United States and Canada, because both substances were used extensively in both countries. Studies conducted in England and Japan, however, pertain primarily to the use of saccharin (Morrison and Buring, 1982b). Results of the latter studies suggested that use of saccharin is not associated with increased bladder cancer risk. Use of cyclamates was evaluated separately in only four studies, with results indicating no increased risk of bladder cancer (IARC, 1999). The findings of nearly all studies indicate that the use of artificial sweeteners confers little or no excess risk of human bladder cancer. If, in fact, saccharin is a very weak carcinogen, such a low-level effect may not be detectable in epidemiologic studies (Hoover and Hartge, 1982).

Alcohol Drinking Most studies that have evaluated alcohol drinking as a risk factor for bladder cancer have not supported a positive association (Brownson et al., 1987; Cartwright et al., 1981; Howe et al., 1980; Kabat et al., 1986; Mills et al., 1991; Murata et al., 1996; Najem et al., 1982; Pelucchi et al., 2002b; Thomas et al., 1983; World Cancer Research Fund, 1997; Wynder et al., 1963; Wynder and Goldsmith, 1977; Zeegers et al., 2001c). Elevated risks related to consumption of specific types of alcoholic beverages have been reported in a few studies (Donato et al., 1997; Iscovich et al., 1987; Kunze et al., 1992; Momas et al., 1994b; Mommsen et al., 1982; Morgan and Jain, 1974; Pohlabeln et al., 1999; Risch et al., 1988a; Slattery et al., 1988b), but these findings have rarely been consistent with respect to type of beverage or gender, and regular dose-response relationships have not been apparent. Thus, the positive findings are likely to be the result of chance or residual confounding by smoking.

Artificial Sweeteners Artificial sweeteners were suggested as potential human bladder carcinogens by the results of animal experiments. The most important evidence was an excess of bladder cancer in rats exposed to high doses of saccharin in utero and weaned to a saccharin-containing diet (U.S. Congress, 1977). Saccharin did not induce bladder cancer in rats or other animals fed saccharin only after birth (Council on Scientific Affairs, 1985; Takayama et al., 1998). Epidemiologic studies have not substantiated a relationship between artificial sweeteners and bladder cancer. Bladder cancer mortality rates were found not to be elevated among diabetics in the United States (Kessler, 1970) or Great Britain (Armstrong and Doll, 1975). The time trend in bladder cancer mortality in England and Wales has not

Other Dietary Factors Of the dietary factors that have been evaluated in relation to bladder cancer, the most consistent evidence supports a protective effect for vegetables and fruits. Relatively high vegetable and fruit consumption has been associated with relatively low risk in most studies (Balbi et al., 2001; Bruemmer et al., 1996 (fruits); Chyou et al., 1993 (fruits); Claude et al., 1986; La Vecchia et al., 1989; Mettlin and Graham, 1979; Michaud et al., 1999b (cruciferous vegetables); Mills et al., 1991; Momas et al., 1994b; Nagano et al., 2000; Nomura et al., 1991; Pohlabeln et al., 1999; Steinmaus et al., 2000; Wakai et al., 2000; Zeegers et al., 2001a (fruits)), but not all (Riboli et al., 1991; Steineck et al., 1988).

1112

PART IV: CANCER BY TISSUE OF ORIGIN

The role of other dietary factors is less clear. Dietary supplements of natural and synthetic retinoids inhibit bladder carcinogenesis in laboratory animals (Hicks, 1983). However, results of epidemiologic studies are inconsistent. Increased intake of foods that contain vitamin A, particularly milk, has been associated with decreased risk of bladder cancer in five case-control studies (Bruemmer et al., 1996; La Vecchia et al., 1989; Mettlin and Graham, 1979; Slattery et al., 1988b, Wakai et al., 2000) and one cohort study (Chyou et al., 1993), but at least two other case-control studies have not supported this relationship (Risch et al., 1988a; Tyler et al., 1986). The use of vitamin A supplements also has been associated with decreased risk (Steineck et al., 1990). Serum levels of retinol, retinol binding protein, and carotenoids do not appear related to risk (Helzlsouer et al., 1989; Nomura et al., 1985; Tyler et al., 1986). Three studies reported a decreased risk associated with increased carotenoid consumption (Castelao et al., 2003; La Vecchia et al., 1989; Vena et al., 1992 (in subjects under 65 years of age)), but no protective effect for carotenoid intake was observed in Spain (Garcia et al., 1999). Further, a meta-analysis of seven casecontrol studies and three cohort studies indicated that low intake of dietary retinol or b-carotene was associated with no increased risk (Steinmaus et al., 2000). Similar results were seen in a recent cohort study from the Netherlands, with the exception of an apparent protective effect for b-crytoxanthin (Zeegers et al., 2001a). In addition, high intake of dietary vitamin C and frequent use of vitamin C supplements were both associated with decreased risk in one case-control study (Bruemmer et al., 1996) and suggested by one cohort study (Michaud et al., 2000). Frequent and long-term use of vitamin E supplements also has been associated with decreased risk in two cohort studies (Jacobs et al., 2002; Michaud et al., 2000). Protective effects of selenium have been reported in a number of cohort studies (Zeegers et al., 2002a). A nearly linear increasing trend in risk with decreasing serum levels of selenium was observed in a nested case-control study in Washington County, Maryland (Helzlsouer et al., 1989). An inverse trend in risk with increasing levels of toenail selenium was reported in a cohort study in the Netherlands (Zeegers et al., 2002a), but not in a cohort of American women (Garland et al., 1995). Increased bladder cancer risk has been associated with relatively high intake of cholesterol (Risch et al., 1988a), with total fat (Vena et al., 1992) and saturated fat (Riboli et al., 1991), with fatty meals (Claude et al., 1986), with fried foods (Bruemmer et al., 1996; Chyou et al., 1993; Steineck et al., 1990), and with relatively high pork and beef consumption (Fraser, 1999; Steineck et al., 1988). Elevated risks were associated with high intake of barbecued meat and salted meat in Uruguay, suggesting a possible role for N-nitroso compounds and heterocyclic amines in bladder carcinogenesis (Balbi et al., 2001). Dietary heterocyclic amine intake, however, was associated with little or no increased risk in a Swedish case-control study (Augustsson et al., 1999).

Drugs Analgesics Heavy consumption of phenacetin-containing analgesics was first linked to cancers of the renal pelvis, ureter, and bladder by a series of case reports (IARC, 1980). There have been only six case-control studies in which the relation between use of phenacetin and risk of bladder cancer has been evaluated (Castelao and Yuan, 2000; Fokkens, 1979; McCredie et al., 1983; McCredie and Stewart, 1988; Piper et al., 1985; Pommer et al., 1999). Fokkens (1979) reported that Dutch subjects who had a lifetime consumption of at least 2 kg had a relative risk of 4.1 compared to incidental users or nonusers. McCredie et al. (1983, 1988) found a relative risk of 2.0 in Australian women age 45–85 years who had a lifetime consumption of at least 1 kg. Piper et al. (1985) reported a relative risk of 6.5 in U.S. women age 20 to 40 years who had used phenacetin-containing compounds for at least 30 days in a year. Castelao et al., (2000) found a relative risk of 1.5 for regular users of phenacetin and 1.9 for heavy users (more than 250 g) in Los Angeles. A regular gradient in risk with increasing dose was demonstrated in both the Australian study (McCredie et al., 1983) and

the Los Angeles study (Castelao and Yuan, 2000). In contrast, no association between phenacetin use and bladder cancer risk was observed in a German case-control study (Pommer et al., 1999). Further study of the relation between phenacetin and bladder cancer will be difficult because most western countries no longer allow phenacetin-containing analgesics to be sold. Acetaminophen, a metabolite of phenacetin, was assessed as a risk factor in studies in Australia (McCredie et al., 1983), the United States (Castelao and Yuan, 2000; Derby and Jick, 1996; IARC, 1999; Kaye and Myers, 2001; Piper et al., 1985), and Germany (Pommer et al., 1999). Results of these studies suggest that heavy use of acetaminophen-containing analgesics does not increase risk. However, acetaminophen did not become popular until the 1970s. Thus, subjects in the earlier studies may not have had sufficient time since initial exposure for bladder cancer to develop. The effect of non-steroidal anti-inflammatory drugs (NSAIDs) on bladder cancer risk has been examined in only two epidemiologic studies, despite experimental evidence that several classes of NSAIDs are potent inhibitors of chemically-induced bladder cancer (Castelao et al., 2000; Pommer et al., 1999). In Los Angeles, Castelao et al. (2000) reported protective effects for several classes of NSAIDs, with relative risks of 0.6 for heavy users of aspirin (at least 1243 g), 0.7 for heavy users of other salicylic acids (at least 168 g), and 0.5 for heavy users of acetic acids (at least 27 g). Heavy lifetime intake of aspirin (at least 1 kg), however, was not associated with decreased risk in a German case-control study (Pommer et al., 1999).

Phenobarbital A protective effect of phenobarbital on bladder cancer risk was first suggested by the observation of a 40% risk reduction among a cohort of epileptic patients who sustained long-term treatment with phenobarbital and other anti-convulsant drugs (Olsen et al., 1993). Olsen and colleagues hypothesized that phenobarbital-induced drug metabolizing enzymes deactivated bladder carcinogens found in cigarette smoke, such as 4-aminobiphenyl. This hypothesis was supported by further work indicating that 4-aminobiphenyl DNA-adduct levels in epileptic smokers treated with phenobarbital were lower than those among epileptic smokers treated with other anti-convulsants (Wallin and Skipper, 1995). No difference in adduct levels was apparent for nonsmokers. In a large U.S. cohort, protective effects for barbiturate use were apparent among both current and former smokers, but not for nonsmokers, also supporting the hypothesis that treatment with phenobarbital induces drug-metabolizing enzymes that deactivate bladder carcinogens found in cigarette smoke (Habel et al., 1998). In contrast, findings from a population-based case-control study in Los Angeles found no evidence of a protective effect for phenobarbital use among smokers or nonsmokers (Castelao et al., 2003). This study, however, was limited by relatively few phenobarbital users, as well as lower lifetime doses compared to the long-term use typical for epileptics several decades ago.

Cyclophosphamide and Chlornaphazine Cyclophosphamide, an alkylating agent that has been used to treat both malignant and non-malignant diseases since the early 1950s, has been linked to risk of bladder cancer in many case reports and case series (IARC, 1981; Levine and Richie, 1989). Cyclophosphamide has been shown to produce bladder tumors in both rats and mice (IARC, 1981). Patients with non-Hodgkin lymphoma who were treated with high doses of cyclophosphamide experienced a seven-fold risk of bladder cancer in a Danish study (Pedersen-Bjergaard et al., 1988). In the largest non-Hodgkin lymphoma study to date, Travis and colleagues (1995) found that cyclophosphamide-related bladder cancer is dosedependent, with relative risks of 2.4, 6.0, and 14.5 for patients receiving cumulative doses of less than 20 g, 20 to 49 g, and 50 g or more, respectively. Results of a study of ovarian cancer patients treated with cyclophosphamide indicated a four-fold increased risk of bladder cancer (Kaldor et al., 1995). A cohort of patients with Wegener’s granulomatosis who were treated with cyclophosphamide had a relative risk of 4.8 (Knight and Askling, 2002). Additional groups of patients, such as long-term survivors of breast cancer who were treated with

Bladder Cancer lower doses of cyclophosphamide as adjuvant chemotherapy, should be studied in order to further clarify the extent of the carcinogenic risk associated with use of this important antineoplastic drug. In the 1960s, the antineoplastic drug chlornaphazine was linked to the development of bladder cancer (Thiede and Christensen, 1969). Chlornaphazine is related chemically to 2-naphthylamine. This drug was never widely used, however (IARC, 1974).

Hair Dyes Three lines of evidence suggest that the use of hair dyes may be associated with increased bladder cancer risk. First, hairdressers and barbers have been reported to have elevated risk (Clemmesen, 1981; Skov and Lynge, 1994). Second, some hair dyes contain 4aminobiphenyl, a known bladder carcinogen in humans (Turesky et al., 2003). Third, people who dye their hair appear to excrete dye compounds in their urine (Hartge et al., 1982). Despite these observations, results of early epidemiologic studies were negative (Hartge et al., 1982; La Vecchia and Tavani, 1995). In 2001, Gago-Dominguez and colleagues reported on findings from a population-based case-control study in Los Angeles indicating that permanent hair dye use in women was associated with a 70% increased risk (Gago-Dominguez et al., 2001a). Positive trends in risk with increasing duration, frequency, and cumulative lifetime use of permanent hair dyes were observed, with the relative risk peaking at 3.7 for women who used permanent hair dyes for at least 30 years. Further, risk appeared to be limited to women who were slow acetylators of N-acetyltransferase 2 (NAT 2), an enzyme involved in the detoxification of aromatic amines (Gago-Dominguez et al., 2001b; Gago-Dominguez et al., 2003). This observation adds biological plausibility to the hypothesis that permanent hair dye use may be a risk factor for bladder cancer and suggests that aromatic amines in hair dyes may be putative bladder carcinogens. However, risk for permanent hair dye use has been evaluated in only three other studies (Andrew et al., 2004; Henley and Thun, 2001; Hennekens et al., 1979). Two studies, based on U.S. cohorts, found no associated between permanent hair dye use and bladder cancer risk (Henley and Thun 2001; Hennekens et al., 1979). The third study, a population-based casecontrol study in New Hampshire, found a nonsignificant, 50% increased risk among permanent hair dye users (Andrew et al., 2004). The negative findings of the cohort studies may have been due to several factors, including inadequate power to detect an effect, misclassification of hair dye exposure, the relevant outcome may be bladder cancer incidence rather than mortality, and cohort effects in the use of hair dyes. Nonetheless, the lack of consistency in findings across studies precludes a causal interpretation of the findings for permanent hair dye use in the development of bladder cancer at this time.

Urologic Conditions Urinary Tract Infection A positive association between urinary bladder infection and risk of bladder cancer has been reported in a number of case-control studies (Brown et al., 1995; Dunham et al., 1968; Howe et al., 1980; Kantor et al., 1984; La Vecchia et al., 1999; Piper et al., 1986; Wynder et al., 1963), although two studies found no support for a causal association (Gonzalez et al., 1991; Kjaer et al., 1989). In the United States, Kantor et al. (1984) found an increased risk associated with urinary tract infections in both men and women; subjects with a history of at least three infections had twice the risk of those with no infections. In addition, bladder infection was more strongly associated with squamous cell than with transitional cell cancer, a striking parallel to the relation between schistosomiasis and squamous cell bladder cancer. The bladder infection/squamous cell carcinoma relationship also is supported by reports of increased risk of squamous cell carcinoma among young female paraplegics (Dolin et al., 1994) and among patients with indwelling catheters due to spinal cord injuries (Pannek, 2002), groups with frequent and severe chronic urinary tract infections. One weakness, however, in most studies conducted to date is that information on dates of bladder infections was not obtained. Thus, the occurrence

1113

or diagnosis of infections may have been the consequence of early bladder cancer, rather than a cause of the disease. If, in fact, a history of urinary bladder infections does increase bladder cancer risk, Nnitroso compounds produced by bacteria may be the putative bladder carcinogens (Bartsch, 1991).

Urinary Stasis If carcinogens are present in urine, urinary retention or stasis caused by infrequent micturition might increase the risk of developing bladder cancer by increasing the duration of contact of the carcinogens with the bladder mucosa (Parkash and Kiesswetter, 1976). Several findings are consistent with the hypothesis that stasis is related to risk. First, infrequent micturition and high urine concentration were more prevalent in high-risk areas of Israel than in low-risk areas (Braver et al., 1987). Second, dogs experimentally exposed to 2-naphthylamine do not develop tumors in bladders that have not been in contact with urine (McDonald and Lund, 1954). Third, dogs administered 4-aminobiphenyl and catheterized to regulate urination frequency had decreased levels of urothelial DNA adducts with increased urination frequency (Kadlubar et al., 1991). Lastly, urine itself appears to be a promoter of bladder carcinogenesis in the rat (Oyasu et al., 1981; Rowland et al., 1980). Urinary stasis has been directly investigated as a risk factor for bladder cancer in one small, case-control study in Serbia (Radosauljevic et al., 2003). A significant protective effect was associated with increased frequency of urination based on 130 cases and 130 controls.

Urine pH In vitro and animal evidence suggests a role for urine pH in aromatic amine-induced bladder carcinogenesis (Babu et al., 1996; Kadlubar et al., 1977; Kadlubar et al., 1981; Kadlubar et al., 1991; Young and Kadlubar, 1982). N-glucuronides of N-hydroxy derivatives of 2naphthylamine and 4-aminobiphenyl are hydrolyzed under acidic conditions and can bind to DNA (Kadlubar et al., 1977). Acidic urine has a similar influence on the hydrolysis of N-glucuronides of benzidine and several of its metabolites (Babu et al., 1992, 1993). A crosssectional study of workers exposed to benzidine and benzidine-based dyes showed that acidic urine pH increased benzidine-DNA adduct levels in exfoliated urothelial cells (Rothman et al., 1997). Diet is an important determinant of urine pH in the healthy general population (Remer and Manz, 1995). In particular, meat, fish, cheese, and grain products contribute to urine acidification, whereas most vegetables and fruits contribute to urine alkalinization (Remer and Manz, 1995). As discussed earlier, high intake of vegetables and fruits has been consistently associated with decreased bladder cancer risk in epidemiologic studies (La Vecchia and Negri, 1996; World Cancer Research Fund, 1997). Although a role for vegetables and fruits in bladder cancer prevention has been proposed, the effect of vegetables and fruits on urine pH as a modifier of bladder cancer risk needs further evaluation. Alguacil and colleagues studied the influence of urine pH measured repeatedly at home over four days by 712 hospital-based bladder cancer patients and 611 hosptial-based controls after discharge and found that a consistently acidic urine pH was associated with increased risk of bladder cancer, particularly among current smokers (Alguacil et al., 2003). In contrast, Wright et al. (2005) estimated urine pH from dietary, height and weight data in a prospective cohort study and did not find an association, except perhaps among long-time smokers.

Schistosoma haematobium For over 90 years, it has been thought that S. haematobium infection is related to increased risk of bladder cancer (Ferguson, 1911); results of most studies indicate that this relationship is causal (Badawi et al., 1995; IARC, 1994). The proportional incidence of bladder cancer is high in developing countries where schistosomiasis is endemic (Tawfik, 1987). The percentage of bladder cancers that are squamous cell tumors is also much higher in endemic areas than it is in nonendemic areas. In Egypt, 70 percent or more of bladder cancers are squamous cell (Tawfik, 1987), compared to about 2 percent in the United States.

1114

PART IV: CANCER BY TISSUE OF ORIGIN

Six out of seven case-control studies indicated that the prevalence of schistosome infection was higher among bladder cancer patients than among controls (Bedwani et al., 1998; Elem and Purohit, 1983; Gelfand et al., 1967; Hinder and Schmaman, 1969; Mustacchi and Shimkin, 1985; Vizcaino et al., 1994). In series of cases from South Africa and Zambia, S. haematobium ova were found in higher proportions of patients with squamous cell than with transitional cell tumors (Bhagwandeen, 1976; Hinder, 1969). Bladder tumors have been produced in monkeys infected with S. haematobium (Kuntz et al., 1972), but these were transitional cell rather than squamous cell tumors. Squamous metaplasia has been observed in the bladders of hamsters infected experimentally with S. haematobium (El Morsi et al., 1975). Several mechanisms by which schistosomiasis infection predisposes to bladder cancer have been suggested. First, the urine of patients infected with S. haematobium or bacteria have greater amounts of potentially carcinogenic nitroso compounds than that of noninfected patients (Abdel Mohsen et al., 1999; Mostafa et al., 1999). Second, chronic inflammation by calcific ova and urinary retention caused by infection might affect the absorption of carcinogens from the urine (Cheever, 1978; Mostafa et al., 1999). Third, the schistosoma antigen might depress the immunocompetence of infected patients (Mee, 1982).

Radiation Ionizing radiation causes bladder cancer, although this exposure contributes very little to bladder cancer incidence in the general population. Women who received therapeutic pelvic radiation for dysfunctional uterine bleeding appear to have a two- to four-fold risk of bladder cancer (Inskip et al., 1990; Wagoner, 1970). In a large, international study of cervical cancer patients treated with radiation, highdose radiotherapy was associated with a four-fold risk of bladder cancer (Boice et al., 1988). Higher risks were experienced by women under age 55 years when first treated, compared to those age 55 or older. Risk increased with increasing dose to the bladder. Risk also increased with time since exposure, with the relative risk reaching 8.7 for patients treated at least 20 years earlier. In addition, patients treated with radiation for metropathia hemorrhagica and ankylosing spondylitis experienced excess bladder cancer risk (UNSCEAR, 2000). Radioactive iodine (iodine-131) exposure has also been associated with elevated bladder cancer risk. A three-fold risk was found among women who had a thyroid uptake procedure with iodine-131 (Piper et al., 1986). A cohort of patients treated with high-dose iodine-131 for thyroid cancer also experienced excess risk (Edmonds and Smith, 1986). Follow-up of atomic bomb survivors in Hiroshima and Nagasaki revealed a dose-response relationship between radiation exposure and bladder cancer mortality (Preston et al., 2003) and incidence (Thompson et al., 1994). In contrast, results from a combined analysis of nuclear industry workers in the United States, the United Kingdom and Canada indicated no excess mortality from bladder cancer, but this lack of effect may have been due to inadequate power (Cardis et al., 1995). In addition, an effect of exposure to low doses of ionizing radiation from the Chernobyl accident in 1986 has been observed in the general population living in radiocontaminated areas of the Ukraine, Belarus, and Russia (Romanenko et al., 2003). From 1986 to 2001, the incidence of bladder cancer in the Ukraine population appeared to increase from 26.2 to 43.3 per 100,000.

Familial Occurrence Evidence for familial predisposition to bladder cancer comes mainly from clinical reports, but elevated risks among persons with bladder cancers in close relatives have been identified in a few case-control studies (Aben et al., 2002; Cartwright et al., 1979a; Kantor et al., 1985; Kiemeney and Schoenberg, 1996; Kramer et al., 1991; Kunze et al., 1992; Piper et al., 1986). In a study of familial bladder cancer in patients identified from the Swedish Cancer Registry for 1958 to 1996, increased risk was observed for offspring of bladder cancer patients,

with risks higher in daughters (SIR = 2.3) than in sons (SIR = 1.4) (Plna and Hemminki, 2001). Patient age at onset modified familial risk, with brothers of bladder cancer probands diagnosed before age 45 experiencing a seven-fold risk. In a large case-control study (Kantor et al., 1985), familial risks were especially high among those with environmental exposures, such as heavy cigarette smoking, suggesting genetic-environmental interaction. Familial occurrences provide an opportunity to identify genetic markers of susceptibility. Genetic susceptibility may also play a role in elevated risk of bladder cancer associated with other tumors of the lower urinary tract (Aben et al., 2002).

MOLECULAR EPIDEMIOLOGY Biologic markers have become a major tool in research on bladder cancer etiology (Engel et al., 2002; Gago-Dominguez et al., 2003; Marcus et al., 2000a, 2000b; Moore et al., 2002; Ross et al., 1996; Skipper et al., 2003; Vineis, 1995; Vineis and Martone, 1996). They include biomarkers of exposure, susceptibility, and disease (e.g., tumor mutations).

Biologic Markers of Exposure Urinary Mutagens There have been a substantial number of cross-sectional studies that have measured urinary mutagenicity in relation to known or suspected bladder mutagens. Cigarette smoking has been associated with mutagenic activity in the urine (Menon, 1984; Yamasaki, 1977) and the level of mutagenic activity was higher for smokers of black tobacco than for smokers of blond tobacco (Malaveille et al., 1989). Other known or suspected risk factors that have been studied in relation to urinary mutagenicity include employment as a rubber worker (Falck et al., 1980), occupational exposure to benzidine (DeMarini et al., 1997a), cyclophosphamide exposure (Falck et al., 1979), and dietary exposure to pan-fried meats with elevated levels of heterocyclic amines (DeMarini et al., 1997b; Peters et al., 2004). Only one study has attempted to directly link urinary mutagenicity with risk of bladder cancer. Garner et al. (1982) reported an association of mutagenic urine with bladder cancer in a comparison of cases and controls, but the effect of disease status on the results was not known. Examination of urine mutagenicity in nonsmoking, non-occupationally exposed cases and controls might shed light on other causes of bladder cancer in the general population. Ideally, this would take place in prospective studies, although very few cohorts have collected and stored urine samples. Alternatively, this might be feasible in a case-control study with a focus on cases with low-grade disease, which would decrease the potential for disease bias.

Hemoglobin and DNA Adducts Hemoglobin adducts from peripheral blood and DNA adducts in urothelial cells can help to identify exposures associated with bladder cancer, particularly when those exposures derive from complex mixtures such as tobacco. Hemoglobin adducts of aromatic amines have been related to cigarette smoking. Bryant et al. (1988) found that smokers have higher levels of hemoglobin adducts of several aromatic amines, including the carcinogens 4-aminobiphenyl and 2naphthylamine. The levels of adducts of 4-aminobiphenyl and 3aminobiphenyl were correlated with the number of cigarettes smoked per day. Smokers of black tobacco had a higher mean level of hemoglobin adducts of 4-aminobiphenyl, as well as several other aromatic amines, than smokers of blond tobacco (Bryant et al., 1988). Maclure et al. (1990) reported that the levels of the hemoglobin adduct of 4aminobiphenyl declined after the cessation of smoking. A large casecontrol study reported an association between 4-aminobiphenyl hemoglobin adducts and risk of bladder cancer in lifelong nonsmokers as well as smokers (Skipper et al., 2003). An additional report from the same study noted that women had higher 4-aminobiphenyl adducts, for a given level of cigarettes smoked per day, than men (Castelao et al., 2001) (see Tobacco section). A small study of white

Bladder Cancer blood cell DNA adducts, which could be composed of several classes of carcinogens including polycyclic aromatic hydrocarbons as well as aromatic amines, found an association with bladder cancer risk (Peluso et al., 2000). Results of exposure biomarkers in case-control studies should be interpreted with caution, given the potential for disease bias or recent changes in lifestyle associated with presenting symptoms or signs of disease. Prospective cohort studies of carcinogen-protein and carcinogen-DNA adducts and bladder cancer have not been reported to date, but the availability of several large cohort studies with banked red and white blood cells should provide new insights into etiologic mechanisms. Bulky DNA adducts have been identified in human bladder biopsy samples and exfoliated urothelial cells and related to smoking; some reports have identified the specific presence of 4-aminobiphenyl DNA adducts and related these to smoking (Airoldi et al., 2002; Curigliano et al., 1996; Talaska et al., 1991a, 1991b, 1994; Vineis, 1996). Further, there is some evidence that the p53 mutational spectrum found in bladder cancer is consistent with the binding spectrum of 4aminobiphenyl (Feng et al., 2002). In the aggregate, studies to date provide additional evidence that aromatic amines in tobacco smoke are one of the principal contributors to bladder cancer risk in smokers.

Biologic Markers of Susceptibility Many studies of genetic polymorphisms and bladder cancer have been carried out since the 1980s. Most reports have been small, often with fewer than 100 cases. As such, these reports have generally had weak to modest power to detect overall effects of common polymorphisms, and negligible power to detect interactions. As noted by Wacholder et al. (2004), statistically significant findings in underpowered studies, particularly when the prior probability of association is low, have a high likelihood of being false positives. Therefore, initial reports should be regarded with caution. At the same time, meta-analyses and pooled analyses of genetic polymorphisms suggest that certain variants may be associated with bladder cancer risk.

NAT2 and NAT1 Aromatic amines must be metabolized within the host in order to exert mutagenic or carcinogenic activity (Morton et al., 1981). For many aromatic monoamines, including those found in tobacco smoke such as 4-aminobiphenyl and 2-naphthylamine, N-acetylation appears to be a detoxification pathway, with the acetylated metabolites being excreted into the urine before they can be N-oxidized to a reactive form (Weber, 1987). The capacity to N-acetylate is polymorphic in humans (Weber, 1987); slow acetylators are homozygotic for a mutated N-acetyltransferase gene (NAT2) that is responsible for decreased activity (Blum et al., 1991; Hein, 2002; Vatsis et al., 1991). In 1979, Lower et al. proposed that individuals with the slow acetylator phenotype might be at higher risk of aromatic amine-associated bladder cancer (Lower et al., 1979). This hypothesis was subsequently supported by results from a series of epidemiologic studies that, overall, showed that individuals with the slow acetylator phenotype (Dewan et al., 1995; Hein, 1988) or NAT2 genotype (Brockmoller et al., 1996; Okkels et al., 1997; Risch et al., 1995) are at greater risk of developing bladder cancer. The biologic plausibility of these observations is strengthened by reports among smokers that slow acetylators have a higher mean level of the hemoglobin adduct of 4-aminobiphenyl than do rapid acetylators (Probst-Hensch et al., 2000; Vineis et al., 1990; Yu et al., 1994). Marcus et al. (2000b) carried out a meta-analysis of studies of NAT2 slow acetylation and bladder cancer, restricting studies to those in the general population without particular occupational or environmental exposures to known bladder carcinogens. The pooled estimate of 22 studies with 2,496 cases and 3,340 controls was a relative risk of 1.4, although there was some evidence of heterogeneity across studies. Analysis of studies in Europe, where most had been carried out, showed similar results with greater homogeneity. A pooled case-case analysis using data from 16 studies with a total of 1,999 cases found evidence of a multiplicative interaction between smoking and NAT2 slow acetylation (relative risk = 1.3). A somewhat stronger effect was

1115

obtained when the analysis was limited to studies carried out in Europe (relative risk = 1.5) (Marcus et al., 2000a). A pooled case-control analysis of genotype-based studies also provided evidence of an interaction between smoking, NAT2 slow acetylation and bladder cancer (Vineis et al., 2001). These findings have been confirmed and extended by Garcia-Closas and colleagues (2005) who reported that the NAT2 slow acetylation genotype was associated with increased risk of bladder cancer, particularly among tobacco-users, in a study of 1150 patients with transitional-cell carcinoma of the urinary bladder and 1149 controls in Spain. Yu et al. (1994) suggested that NAT2 acetylator status may play a role in the racial/ethnic differences in bladder cancer risk. They found that the proportion of slow acetylators was highest among whites, intermediate among blacks, and lowest among Asians, closely paralleling the racial/ethnic variation in risk. Among persons occupationally exposed to aromatic amines, several studies have shown an excess of the slow acetylator phenotype among cases compared to controls (Cartwright et al., 1982; Cartwright et al., 1984; Hanke, 1990; Ladero et al., 1985; Risch et al., 1995; Vineis et al., 2001; Weber et al., 1983), but the specific aromatic amine exposures were generally not well characterized. A study in Chinese workers with an increased risk of bladder cancer (Bi et al., 1992) who were exposed exclusively to benzidine did not demonstrate an excess of slow acetylators among bladder cancer cases, based on both phenotype and genotype analyses (Hayes et al., 1993). This finding was confirmed in a report that studied additional cases identified during further follow-up of the same benzidine-exposed population in China (Carreon et al., 2006). Indeed, the pooled risk estimate for cases and controls using data from both reports suggested that NAT2 slow acetylation was associated with a decreased risk of bladder cancer in this population. Also, there was evidence that the NAT1*10 allele was associated with increased risk of bladder cancer (Carreon et al., 2006). In contrast, a study of bladder cancer cases who had been exposed to benzidine in Germany found a nonsignificant increased risk with the slow acetylation phenotype (Golka et al., 1996). A cross-sectional study of workers in India who were currently exposed to benzidine and benzidine-based dyes showed that the predominant urothelial cell DNA adduct was N-acetylated benzidine (Rothman et al., 1996), suggesting that benzidine may be activated rather than detoxified by Nacetylation. In vitro studies suggested that benzidine is a better substrate for NAT1 than NAT2 (Zenser et al., 1996). Overall, the association between N-acetylation polymorphisms and bladder cancer risk may be specific to certain aromatic amines. Several studies have evaluated the role of polymorphisms in NAT1, which can acetylate aromatic amines, and bladder cancer. There has been a particular focus on the NAT*10 allele, which may be a functional variant (Cascorbi et al., 2001). Results from studies to date, however, have been inconsistent (Cascorbi et al., 2001; Garcia-Closas et al., 2005; Hsieh et al., 1999; Okkels et al., 1997; Taylor et al., 1998).

GSTM1, GSTT1 and GSTP1 Human glutathione S-transferase M1 (GSTM1) belongs to a family of enzymes that detoxify a spectrum of reactive carcinogenic metabolites, including PAHs, by catalyzing their conjugation to glutathione. GSTM1 is encoded by the GSTM1 gene, and is polymorphic in human populations; deficiency of this enzyme is caused by the homozygous absence of a functional GSTM1 gene (i.e., null genotype) (Seidegard et al., 1988). Most studies in the general population have found that the GSTM1 null genotype is associated with elevated risk of bladder cancer (Anwar et al., 1996; Bell et al., 1993; Brockmoller et al., 1996; Chern et al., 1994; Daly et al., 1993; Katoh et al., 1995; Kempkes et al., 1996; Lafuente et al., 1993; Lin et al., 1994), although there have been a few null reports (Okkels et al., 1997; Zhong et al., 1993). Engel et al. (2002) carried out a meta-analysis of GSTM1 null genotype and bladder cancer in the general population and obtained a summary relative risk of 1.4 based on 17 studies with a total of 2,149 cases and 3,646 controls (Fig. 58–6). These findings have been confirmed and extended by Garcia-Closas and colleagues (2005). Hirvonen et al. (1994) found that the GSTM1 null genotype was associated with higher urine mutagenic activity in smokers, which is thought to derive primarily from aromatic amines (Bartsch et al., 1993), while Yu et al.

1116

PART IV: CANCER BY TISSUE OF ORIGIN (Van Gils et al., 2002; Sanyal et al., 2004)); cell cycle control genes (such as DR4 (Hazra et al., 2003), cyclin D1 (Wang et al., 2002), and p21 (Chen et al., 2002) ), and an increasing number of DNA repair genes, particularly in the nucleotide excision repair pathway (GarciaClosas et al., 2006; Matullo et al., 2001, 2005; Sanyal et al., 2004; Stern et al., 2001, 2002a, 2002b). Despite some suggestive associations, replication in larger studies and in pooled analyses are needed before conclusions can be reached.

Additional Susceptibility Assays

Figure 58–6. Odds ratios and 95% confidence intervals for glutathione Stransferase M1 (GSTM1) null status and bladder cancer risk. Solid circles are proportional in area to the number of cases. The vertical axis is on a log scale. (Source: Engel et al., 2002.)

(1995) reported that the null genotype was associated with higher levels of 3- and 4-aminobiphenyl hemoglobin adducts among both smokers and nonsmokers. Several studies have examined the GSTT1 null genotype, which results in decreased ability to detoxify a variety of generally small molecular weight compounds (Brockmoller et al., 1996). Overall, findings have been equivocal with reports of both small increased risks, no association, and deceased risks (Garcia-Closas et al., 2005; Lee et al., 2002; Sanyal et al., 2004). Results from several studies of polymorphisms in GSTP1, which plays a role in detoxifying PAHs, also have been inconsistent (Garcia-Closas et al., 2005; Harries et al., 1997; Katoh et al., 1999; Toruner et al., 2001).

Cytochrome P-450 Enzymes Cytochrome P-4501A2 (CYP1A2) is thought to play an important role in the metabolic activation of several aromatic amines, including 4aminobiphenyl and 2-naphthylamine, via N-oxidation (Butler et al., 1992). Susceptibility to aromatic amine-induced bladder cancer may thus vary with an individual’s level of CYP1A2. It is unclear, however, whether the activity of this enzyme is determined by genotype and/or by enzyme induction from environmental exposures (e.g., cigarette smoking, dietary factors, and drugs). Higher levels of 4aminobiphenyl adducts were found in low-level smokers with rapid CYP1A2 activity, particularly among slow acetylators (Landi et al., 1996). CYP1A2 activity, as measured by 3-demethylation of theophylline, was associated with bladder cancer risk in one study (Lee et al., 1994), but uncertainty about the validity of the assay, demographic differences between cases and controls, and the potential for disease bias make interpretation of these findings difficult. Evaluation of genetic polymorphisms in CYP1A2 and bladder cancer is underway in several studies, but uncertainty about the functional importance of these variants may complicate interpretation of study results. Polymorphisms in other CYPs including CYP1A1, CYP2C19, CYP2D6, and CYP2E1 have been evaluated in one or more studies but results to date have generally not shown associations with bladder cancer (Brockmoller et al., 1996).

Polymorphisms in Other Genes Case-control studies of bladder cancer and polymorphisms have evaluated other metabolic genes (such as NQO1 (Park et al., 2003) and MeEH (Brockmoller et al., 1996) ); proto-oncogenes (such as HRAS1

A new class of susceptibility assays has been developed that evaluates the tendency of cultured peripheral lymphocytes to become damaged after in vitro exposure to various types of carcinogens. An alternative approach is to measure the ability of cultured lymphocytes to carry out DNA repair after being exposed to carcinogens. These assays have been applied to studies of lung cancer and other smoking-related cancers, and are now being used in studies of bladder cancer. Based on the comet assay as a measure of DNA damage in peripheral lymphocytes, Schabath et al (2003) reported that bladder cancer cases were more susceptible to the effects of benzo(a)pyrene diol epoxide (an activated PAH metabolite) than controls. Lin et al. showed that decreased ability to repair DNA damage induced by 4-aminobiphenyl was associated with increased risk of bladder cancer (Lin et al., 2005). Another integrative measure of susceptibility for bladder cancer may be chromosomal telomere length, a measure of genetic stability. Both Wu et al. (2003) and Broberg et al. (2005) reported that shorter telomere length was associated with increased risk of bladder cancer.

TUMOR MARKERS An understanding of the pathogenesis of human bladder cancer at the molecular level has been evolving for a number of years. Technological advances have led to an interest in incorporating tumor markers in epidemiologic studies to examine case groups of tumors for etiologic heterogeneity, strengthening causal inference. During the last decade, the genetics of bladder cancer has been studied using several methodologies including immunohistochemistry (IHC), fluorescent in situ hybridization (FISH), and chromosomal and array based comparative genomic hybridization (CGH) (Kallioniemi et al., 1995; Veltman et al., 2003). More recently, RNA expression, and proteomic and protein microarray analyses of bladder tumor tissue have been described (Sanchez-Carbayo et al., 2003; Wulfkuhle et al., 2003). Global and gene-specific epigenetic patterns in human bladder cancer genomes have also been observed (Markl et al., 2001; Maruyama et al., 2001). Using these techniques to compare alterations in both tumor and adjacent normal tissue, the concept of a “field effect” developed (Harris, 1992; Lunec et al., 1992; Sidransky et al., 1992). The accumulation of genetic and epigenetic changes, also known as “field changes,” may be caused by years of carcinogenic exposures, inflammation, and the aging process on the urothelium (Harris, 1992; Issa, 2003). However, the development of multicentric cancer is believed to result mainly from the seeding and implantation of a single clone within the bladder epithelium since similar genetic and epigenetic alterations have been observed in patients with synchronous and metachronous urothelial tumors (Muto et al., 2000; Paterson et al., 2003; Vriesema et al., 2001). Bladder cancer is caused by a combination of mutations and epigenetic alterations in tumor suppressor genes and oncogenes. Several oncogenes commonly mutated or amplified in bladder cancer include H-Ras, FGFR3, ERBB2, CCND1, and MDM2 (Cappellen et al., 1999; Coombs et al., 1991; Fitzgerald et al., 1995; Habuchi et al., 1994; Knowles, 1993; Lianes et al., 1994; Mellon et al., 1996; Proctor et al., 1991; Sato et al., 1992; Sauter et al., 1993; Sibley et al., 2001; Underwood et al., 1995). Tumor suppressor genes associated with bladder cancer include several located on chromosome 9 (p16, PTCH, TSC1, DBCCR1), p53, and several key G1 checkpoint proteins (Rb, cyclin D1, p16) (Bringuier et al., 1996; Habuchi et al., 1994; Kimura et al., 2001; Knowles, 1999; Lianes et al., 1994; Nishiyama et al., 1999; Shin et al., 1997).

Bladder Cancer Bladder tumors are heterogeneous with regard to their molecular characteristics, suggesting that environmental exposures may have stronger associations with subgroups of tumors defined at the molecular level (Zhang et al., 1997a). In epidemiologic studies of bladder cancer, the p53 gene is frequently used as a marker of carcinogenspecific damage (Greenblatt et al., 1994; Semenza, 1997). P53 gene mutation is a useful biomarker in epidemiologic studies of bladder cancer because it is frequently mutated and exogenous and endogenous mutagenic events cause distinctly different mutational patterns (Hussain and Harris, 1998). Although the pattern of p53 mutations has been studied in bladder cancer, attention has only recently been given to the relation between specific exposures and p53 mutations. Although bladder cancer risk is two to three times higher in smokers than in non-smokers, associations of p53 mutation and/or p53 immunohistochemical positivity with cigarette smoking have shown conflicting results (Habuchi et al., 1993; Kannio et al., 1996; LaRue et al., 2000; Moore et al., 2003; Schroeder et al., 2003; Spruck et al., 1993; Zhang et al., 1994). However, two recent larger studies of p53 and bladder cancer did provide evidence that some CpG G:C-A:T transitions may be caused by carcinogens found in tobacco smoke (Moore et al., 2003; Schroeder et al., 2003). When tumors from smokers and nonsmokers were compared using CGH, no difference in the frequency of gross chromosomal gains and losses was observed (Moore et al., 2002) and only chromosome 9q loss was more frequent in nonsmokers. In contrast, results from an LOH study in tumors from bladder cancer cases that smoked contained more regional chromosome 9 alterations compared with nonsmoking cases (Zhang et al., 1997b). The difference between these findings may reflect the lower sensitivity of CGH to detect small genetic alterations compared with those that can be identified with LOH. The molecular mechanism by which arsenic promotes bladder cancer has also been studied by comparing tumors from arsenicexposed and nonexposed individuals. One small study of arsenicinduced bladder tumors from the arsenic endemic region of Taiwan (N = 13) observed several mutations at codon 175, a CpG site previously associated with inflammatory agents in transitional cell carcinoma of the bladder (Shibata et al., 1994). However, a recent larger study (N = 147) found that p53 protein expression and mutation prevalence did not differ in tumors between arsenic-exposed and nonexposed patients (Moore et al., 2003). When CGH was used to screen these same tumors for chromosomal aberrations throughout the tumor genome, bladder tumors associated with higher levels of arsenic exposure showed increased number of chromosomal gains and losses (Moore et al., 2002). This observation suggests that arsenic-induced tumors may be less genetically stable than bladder tumors from nonexposed cases. Most of the specific chromosomal changes associated with arsenic exposure were also related to tumor stage and grade, suggesting that arsenic-induced tumors may behave more aggressively and result in increased mortality (Moore et al., 2002). Bladder tissue from individuals exposed to cesium-137 (137Cs) as a result of the Chernobyl accident also appear to display a different genetic profile when compared to tissues from unexposed individuals. Studies conducted in the Ukraine compared the urinary bladder mucosa of men who underwent surgery for benign prostatic hyperplasia. In one study based on 28 patients with high exposure, 17 with low exposure, and 10 nonexposed patients, an increase in the incidence of moderate-to-severe hyperplasia and CIS was seen only in exposed groups (Romanenko et al., 1999). In addition, the abnormal pathologic changes appeared to be correlated with the prevalence of several immunohistochemical markers including p53, PCNA, cyclin D1 and p21 (Romanenko et al., 1999). Other studies observed that irradiation cystitis, a condition related to chronic low-dose radiation exposure, was associated with severe dysplasia and CIS (Romanenko et al., 2000). P53 mutation analysis of a small number of tissues from exposed individuals (N = 17) with altered epithelium demonstrated a specific pattern of mutation (Yamamoto et al., 1999). Levels of several immunochemical markers of oxidative DNA damage and several transcription factors involved in early carcinogenic response suggest that that oxidative stress may play a role in promoting the development of bladder lesions in this population (Romanenko et al., 2000, 2003).

1117

Comparative studies of schistosomiasis-induced bladder tumors, which are largely squamous cell cancers, have also been conducted, especially using CGH to examine gross chromosomal aberrations in the genome (E1 Rifai et al., 2000; Fadl-Elmula et al., 2002; Muscheck et al., 2000). In the largest study (N = 69), more overall alterations (gains and losses) were observed in Schistosome-associated tumors than in other tumors. In particular, losses in chromosome arm 5q were more frequent in Schistosome-associated tumors (E1 Rifai et al., 2000). Exposure to aromatic amines and p53 mutations in bladder tumor DNA has also been examined in a few early studies, but the results are difficult to interpret because of small sample sizes and, in some instances, lack of a nonexposed comparison group (Esteve et al., 1995; Martone et al., 2000; Romano et al., 1999; Sorlie et al., 1998; Taylor et al., 1996; Yasunaga et al., 1997).

PREVENTIVE MEASURES In public health terms, avoidance of cigarette smoking is the most effective means available for the prevention of bladder cancer because the proportion of cases attributable to smoking is greater than that for other risk factors (Brennan et al., 2000; D’Avanzo et al., 1995; Hartge et al., 1987; IARC, 1986). A further measure would be curtailment of hazardous occupational exposures. Screening is appealing as a method of bladder cancer control because of easy accessibility by means of cystoscopy (National Cancer Institute, 2002). Since all tumors arise on the urothelial surface, there is a time window (before muscle invasion occurs) when tumors can be detected. This window, however, appears to be short and thus the prevalence of preclinical bladder cancer is too low in the general population for large-scale screening programs to be rewarding (Morrison, 1979). Certain high-risk occupational groups may have a sufficiently high prevalence of the disease to justify screening as a bladder cancer control measure. Clinical studies of screening in high-risk populations, however, have failed to demonstrate a beneficial effect on outcome or mortality (Pashos et al., 2002). Screening for bladder cancer in the general population or in high-risk groups is currently not recommended because of the high rate of false positive results, which may lead to unnecessary cystoscopy or other invasive procedures (CancerNet PDQ, 2003).

FUTURE DIRECTIONS Cigarette smoking accounts for about 50% to 65% of bladder cancer among men and 20% to 30% among women (Brennan et al., 2000; D’Avanzo et al., 1995; Hartge et al., 1987, 1990; Sorahan et al., 1994). Occupational exposures have been estimated to be responsible for 5% to 25% of bladder cancer among men (Kogevinas et al., 2003; Silverman et al., 1989b) and 8% to 11% among women (Silverman et al., 1990; ’t Mannetje et al., 1999), yet the exposures responsible for much of occupational bladder cancer remain unknown. Cigarette smoking and occupational exposures, however, explain only a small part of the large male excess risk of bladder cancer (Hartge et al., 1990), which represents an important etiologic lead to other potential risk factors. The etiologic roles of several other exposures remain poorly understood. The effect of total fluid intake on bladder cancer risk needs further evaluation based on well-designed studies that take into account the quality of the water subjects consume (e.g., disinfection by-product levels). The effects of ingestion of low-to-moderate levels of arsenic in drinking water, as well as exposure to disinfection byproducts by both water ingestion and other routes (i.e., bathing and swimming in pools), deserve attention. The effect of hair dye use warrants further investigation based on studies with large numbers of hair dye users. The roles of urination frequency and urine pH as modifiers of bladder cancer risk merit exploration. Additional research is also needed to better understand the effects of occupational exposures other than aromatic amines, NSAID use, dietary factors, tobacco products

1118

PART IV: CANCER BY TISSUE OF ORIGIN

other than cigarettes, chronic urinary tract infections, and genetic susceptibility. Finally, it is important for future studies to use molecular epidemiologic methods to explore mechanisms of action of established bladder carcinogens and to help identify other carcinogenic exposures and susceptibility factors. Several large case-control studies of bladder cancer are underway, pursuing genetic leads suggested by early studies and expanding the study of genetic susceptibility by employing new, comprehensive approaches to genomics. Given that two of the most consistent associations between common genetic polymorphisms and cancer have been described for bladder cancer per se (i.e., NAT2 slow acetylation, GSTM1 null genotype), it is likely that other genetic variants will also be identified, These studies, particularly if pooled, will have adequate power to identify gene-gene and gene-environment interactions with both well-established (e.g., cigarette smoking) and suspected risk factors. In addition, several large cohorts will soon have adequate numbers of bladder cancer patients to conduct nested casecontrol studies to examine associations with exposure biomarkers using samples collected years before disease onset. Tumor marker studies are also needed, especially those that emphasize sound epidemiologic study design and take into account tumor stage, grade, and possible confounding factors. The application of new genomic technologies in the analysis of tumor samples should provide new insights into the induction and progression of bladder cancer, ultimately leading to development of strategies aimed at prevention and control of this tumor. Acknowledgments The authors thank Dr. Kenneth Cantor for his comments on the section on drinking water and fluid intake, Dr. Ethel Gilbert for her help with the radiation section, Dr. Regina Ziegler for her comments on the diet section, Mr. John Lahey of IMS, Inc. for computer programming and figure development, and Ms. Amie Seisay, Ms. Judy Lichaa and Mr. Mark Donahue for clerical assistance.

References Aben KK, Witjes JA, Schoenberg MP, et al. 2002. Familial aggregation of urothelial cell carcinoma. Int J Cancer 98:274–278. Abdel Mohsen MA, Hassan AA, El Sewedy SM, et al. 1999. Biomonitoring of n-nitroso compounds, nitrite and nitrate in the urine of Egyptian bladder cancer patients with or without Schistosoma haematobium infection. Int J Cancer 82:789–794. Airoldi L, Orsi F, Magagnotti C, Coda R, et al., 2002. Determinants of 4aminobiphenyl-DNA adducts in bladder cancer biopsies. Carcinogenesis 23:861–866. Alderson M. 1986. Occupational Cancer. London: Butterworth and Co. Ltd. Alguacil J, Kogevinas M, Silverman DT, et al. 2003. Urine pH and bladder cancer risk in the Spanish Bladder Cancer Study. AACR special conference on Molecular and Genetic Epidemiology. Waikoloa, Hawaii. Proceedings of the American Association for Cancer Research. Proceedings of the Molecular and Genetic Epidemiology of Cancer Conference. B52. American Joint Committee on Cancer. 1988. Manual for staging of cancer. Philadelphia: Lippincott. Andrew AS, Schned AR, Heaney JA, Karagas MR. 2004. Bladder cancer risk and personal hair dye use. Int J Cancer 109:581–586. Antony HM. 1974. Industrial exposure in patients with carcinoma of the bladder. J Soc Occup Med 24:110. Anwar WA, Abdel-Rahman SZ, El Zein RA, et al. 1996. Genetic polymorphism of GSTM1, CYP2E1 and CYP2D6 in Egyptian bladder cancer patients. Carcinogenesis 17:1923–1929. Armstrong B, Doll R. 1974. Bladder cancer mortality in England and Wales in relation to cigarette smoking and saccharin consumption. Br J Prev Soc Med 28:233–240. Armstrong B, Doll R. 1975. Bladder cancer mortality in diabetics in relation to saccharin consumption and smoking habits. Br J Prev Soc Med 29:73–81. Augustine A, Hebert JR, Kabat GC, et al. 1988. Bladder cancer in relation to cigarette smoking. Cancer Res 48:4405–4408. Augustsson K, Skog K, Jagerstad M, et al. 1999. Dietary heterocyclic amines and cancer of the colon, rectum, bladder, and kidney: a population-based study. Lancet 353:703–707. Babu SR, Lakshmi VM, Hsu FF, et al. 1993. N-acetylbenzidine-N’glucuronidation by human, dog and rat liver. Carcinogenesis 14:2605–2611.

Babu SR, Lakshmi VM, Hsu FF, et al. 1992. Role of N-glucuronidation in benzidine-induced bladder cancer in dog. Carcinogenesis 13:1235–1240. Babu SR, Lakshmi VM, Huang GP, et al. 1996. Glucuronide conjugates of 4aminobiphenyl and its N-hydroxy metabolites. pH stability and synthesis by human and dog liver. Biochem Pharmacol 51:1679–1685. Badawi AF, Mostafa MH, Probert A, et al. 1995. Role of schistosomiasis in human bladder cancer: evidence of association, aetiological factors, and basic mechanisms of carcinogenesis. Eur J Cancer Prev 4:45–59. Balbi JC, Larrinaga MT, De Stefani E, et al. 2001. Foods and risk of bladder cancer: a case-control study in Uruguay. Eur J Cancer Prev 10:453–458. Barbone F, Franceschi S, Talamini R, et al. 1994. Occupation and bladder cancer in Pordenone north-east Italy: a case-control study. Int J Epidemiol 23:58–65. Bartsch H. 1991. N-Nitroso compounds and human cancer: where do we stand? Lyon: International Agency for Research on Cancer 1–10. Bartsch H, Malaveille c, Friesen M, et al. 1993. Black air cured. and blond flue cured. tobacco cancer risk. IV: Molecular dosimetry studies implicate aromatic amines as bladder carcinogens. Eur J Cancer 29A: 1199–1207. Bates MN, Smith AH, Cantor KB. 1995. Case-control study of bladder cancer and arsenic in drinking water. Am J Epidemiol 141:523–530. Bates MN, Rey OA, Biggs ML, et al. 2004. Case-control study of bladder cancer and exposure to arsenic in Argentina. Am J Epidemiol 159: 381–389. Baus Subcommittee on Industrial Bladder Cancer. 1988. Occupational bladder cancer: a guide for clinicians. The BAUS Subcommittee on Industrial Bladder Cancer. Br J Urol 61:183–191. Baxter P, McDowall ME. 1986. Occupation and cancer in London: an investigation into nasal and bladder cancer using the Cancer Atlas. Br J Ind Med 43:44–49. Bedwani R, Renganathan E, El Kwhsky F, et al. 1998. Schistosomiasis and the risk of bladder cancer in Alexandria, Egypt. Br J Cancer 77:1186–1189. Bell DA, Taylor JA, Paulson DF. 1993. Genetic risk and carcinogen exposure: a common inherited defect of the carcinogen-metabolism gene glutathione S-transferase M1 (GSTM1) that increases susceptibility to bladder cancer. J Natl Cancer Inst 85:1159–1164. Bethwaite PB, Pearce N, Fraser J. 1990. Cancer risks in painters: study based on the New Zealand Cancer Registry. Br J Ind Med 47:742–746. Bhagwandeen SB. 1976. Schistosomiasis and carcinoma of the bladder in Zambia. S Afr Med J 50:1616–1620. Bi W, Hayes RB, Feng P, et al. 1992. Mortality and incidence of bladder cancer in benzidine-exposed workers in China. Am J Ind Med 21:481–489. Blot WJ, Fraumeni JF Jr. 1978. Geographic patterns of bladder cancer in the United States. J Natl Cancer Inst 61:1017–1023. Blum M, Demierre A, Grant DM, et al. 1991. Molecular mechanism of slow acetylation of drugs and carcinogens in humans. Proc Natl Acad Sci U S A 88:5237–5241. Boffetta P, Silverman DT. 2001. A meta-analysis of bladder cancer and diesel exhaust exposure. Epidemiology 12:125–130. Boice JD Jr., Engholm G, Kleinerman RA, et al. 1988. Radiation dose and second cancer risk in patients treated for cancer of the cervix. Radiat Res 116:3–55. Boyko RW, Cartwright RA, Glashan RW. 1985. Bladder cancer in dye manufacturing workers. J Occup Med 27:799–803. Braver DJ, Modan M, Chetrit A, et al. 1987. Drinking, micturition habits, and urine concentration as potential risk factors in urinary bladder cancer. J Natl Cancer Inst 78:437–440. Brennan P, Bogillot O, Cordier S, et al. 2000. Cigarette smoking and bladder cancer in men: a pooled analysis of 11 case-control studies. Int J Cancer 86:289–294. Bringuier PP, Tamimi Y, Schuuring E, et al. 1996. Expression of cyclin D1 and EMS1 in bladder tumours; relationship with chromosome 11q13 amplification. Oncogene 12:1747–1753. Broberg K, Bjork J, Paulsson K, et al. 2005. Constitutional short telomeres are strong genetic susceptibility markers for bladder cancer. Carcinogenesis 26:1263–1271. Brockmoller J, Cascorbi I, Kerb R, et al. 1996. Combined analysis of inherited polymorphisms in arylamine N-acetyltransferase 2, glutathione Stransferases M1 and T1, microsomal epoxide hydrolase, and cytochrome P450 enzymes as modulators of bladder cancer risk. Cancer Res 56: 3915–3925. Bross ID, Tidings J. 1973. Another look at coffee drinking and cancer of the urinary bladder. Prev Med 2:445–451. Brown LM, Zahm SH, Hoover RN, et al. 1995. High bladder cancer mortality in rural New England, United States: an etiologic study. Cancer Causes Control 6:361–368. Brownson RC, Chang JC, Davis JR. 1987. Occupation, smoking, and alcohol in the epidemiology of bladder cancer. Am J Public Health 77:1298–1300.

Bladder Cancer Bruemmer B, White E, Vaughan TL, et al. 1997. Fluid intake and the incidence of bladder cancer among middle-aged men and women in a three-county area of western Washington. Nutr Cancer 29:163–168. Bruemmer B, White E, Vaughan TL, et al. 1996. Nutrient intake in relation to bladder cancer among middle-aged men and women. Am J Epidemiol 144:485–495. Bryant MS, Vineis P, Skipper PL, et al. 1988. Hemoglobin adducts of aromatic amines: associations with smoking status and type of tobacco. Proc Natl Acad Sci U S A 85:9788–9791. Burch JD, Rohan TE, Howe GR, et al. 1989. Risk of bladder cancer by source and type of tobacco exposure: a case-control study. Int J Cancer 44:622–628. Butler MA, Lang NP, Young JF, et al. 1992. Determination of CYP1A2 and NAT2 phenotypes in human populations by analysis of caffeine urinary metabolites. Pharmacogenetics 2:116–127. CancerNet PDQ. 2003. Screening for Bladder Cancer (PDQ); Screening/Detection: Health Professionals. Bethesda, Md: CancerNet PDQ, NCI. Available at: http://cancernet.nci.nih.gov. Cantor KP. 2001. Invited commentary: arsenic and cancer of the urinary tract. Am J Epidemiol 153:422–423. Cantor KP. 1997. Drinking water and cancer. Cancer Causes Control 8:292–308. Cantor KP, Hoover R, Hartge P, et al. 1987. Bladder cancer, drinking water source, and tap water consumption: a case-control study. J Natl Cancer Inst 79:1269–1279. Cantor KP, Lynch CF, Hildesheim ME, et al. 1998. Drinking water source and chlorination byproducts. I. Risk of bladder cancer. Epidemiology 9:21–28. Cappellen D, De Oliveira C, Ricol D, et al. 1999. Frequent activating mutations of FGFR3 in human bladder and cervix carcinomas. Nat Genet 23:18–20. Cardis E, Gilbert E, Carpenter L, et al. 1995. Effects of Low Doses and Low Dose Rates of External Ionizing Raidation: Cancer Mortality among Nuclear Industry Workers in Three Countries. Radiat Res 142:117– 132. Carreon T, Ruder AM, Schulte PA, et al. 2006. NAT2 slow acetylation and bladder cancer in workers exposed to benzidine. Int J Cancer 118: 161–168. Cartwright RA. 1979a. Genetic association with bladder cancer. Br Med J 2:798. Cartwright RA, Adib R, Appleyard I, et al. 1983. Cigarette smoking and bladder cancer: an epidemiological inquiry in West Yorkshire. J Epidemiol Community Health 37:256–263. Cartwright RA, Adib R, Glashan R, et al. 1981. The epidemiology of bladder cancer in West Yorkshire. A preliminary report on non-occupational aetiologies. Carcinogenesis 2:343–347. Cartwright RA, Bernard SM, Glashan RW, et al. 1979b. Bladder cancer amongst dye users. Lancet 2:1073–1074. Cartwright RA, Glashan RW, Rogers HJ, et al. 1982. Role of Nacetyltransferase phenotypes in bladder carcinogenesis: a pharmacogenetic epidemiological approach to bladder cancer. Lancet 2:842–845. Cartwright RA, Philip PA, Rogers HJ, et al. 1984. Genetically determined debrisoquine oxidation capacity in bladder cancer. Carcinogenesis 5:1191–1192. Cascorbi I, Roots I, Brockmoller J. 2001. Association of NAT1 and NAT2 polymorphisms to urinary bladder cancer: significantly reduced risk in subjects with NAT1*10. Cancer Res 61:5051–5056. Case R, Hosker M. 1954a. Tumour of the urinary bladder as an occupational disease in the rubber industry in England and Wales. Br J Ind Med 8:39–50. Case RA, Hosker ME, McDonald DB, et al. 1954b. Tumours of the urinary bladder in workmen engaged in the manufacture and use of certain dyestuff intermediates in the British chemical industry. I. The role of aniline, benzidine, alpha-naphthylamine, and beta-naphthylamine. Br J Ind Med 11:75–104. Castelao JE, Gago-Dominguez M, Yuan JM, et al. 2003. Phenobarbital use and bladder cancer risk. Eur J Epidemiol 18:659–664. Castelao JE, Yuan JM, Gago-Dominguez M, et al. 2003. Carotenoids/vitamin C and smoking-related bladder cancer. Int J Cancer 110:417–423. Castelao JE, Yuan JM, Skipper PL, et al. 2001. Gender- and smoking-related bladder cancer risk. J Natl Cancer Inst 93:538–545. Castelao JE, Yuan JM, Gago-Dominguez M, et al. 2000. Non-steroidal anti-inflammatory drugs and bladder cancer prevention. Br J Cancer 82: 1364–1369. Checkoway H, Smith AH, McMichael AJ, et al. 1981. A case-control study of bladder cancer in the United States rubber and tyre industry. Br J Ind Med 38:240–246. Cheever AW. 1978. Schistosomiasis and neoplasia. J Natl Cancer Inst 61:13–18.

1119

Chen R, Seaton A. 1998. A meta-analysis of painting exposure and cancer mortality. Cancer Detect Prev 22:533–539. Chen CJ, Chuang YC, Lin TM, et al. 1985. Malignant neoplasms among residents of a blackfoot disease-endemic area in Taiwan: high-arsenic artesian well water and cancers. Cancer Res 45:5895–5899. Chen CJ, Chuang YC, You SL, et al. 1986. A retrospective study on malignant neoplasms of bladder, lung and liver in blackfoot disease endemic area in Taiwan. Br J Cancer 53:399–405. Chen JG. 1990. A cohort study on the cancer experience among workers exposed to benzidine-derived dyes in Shanghai leather-tanning industry. Zhonghua Yu Fang Yi Xue Za Zhi 24:328–331. Chen WC, Wu HC, Hsu CD, et al. 2002. p21 gene codon 31 polymorphism is associated with bladder cancer. Urol Oncol 7:63–66. Chern HD, Romkes-Sparks M, Hu JJ. 1994. Homozygous deleted genotype of glutathione S-transferase M1 increases susceptibility to aggressive bladder cancer. Proc Am Assoc Cancer Res 35:285–288. Chiang HS, Guo HR, Hong CL, et al. 1993. The incidence of bladder cancer in the black foot disease endemic area in Taiwan. Br J Urol 71:274– 278. Chiou HY, Chiou ST, Hsu YH, et al. 2001. Incidence of transitional cell carcinoma and arsenic in drinking water: a follow-up study of 8,102 residents in an arseniasis-endemic area in northeastern Taiwan. Am J Epidemiol 153:411–418. Chiou HY, Hsueh YM, Liaw KF, et al. 1995. Incidence of internal cancers and ingested inorganic arsenic: a seven-year follow-up study in Taiwan. Cancer Res 55:1296–1300. Chiu BC, Lynch CF, Cerhan JR, et al. 2001. Cigarette smoking and risk of bladder, pancreas, kidney, and colorectal cancers in Iowa. Ann Epidemiol 11:28–37. Chyou PH, Nomura AM, Stemmermann GN. 1993. A prospective study of diet, smoking, and lower urinary tract cancer. Ann Epidemiol 3:211–216. Ciccone G, Vineis P. 1988. Coffee drinking and bladder cancer. Cancer Lett 41:45–52. Claude J, Kunze E, Frentzel-Beyme R, et al. 1986. Life-style and occupational risk factors in cancer of the lower urinary tract. Am J Epidemiol 124:578–589. Claude JC, Frentzel-Beyme RR, Kunze E. 1988. Occupation and risk of cancer of the lower urinary tract among men. A case-control study. Int J Cancer 41:371–379. Clavel J, Cordier S. 1991. Coffee consumption and bladder cancer risk. Int J Cancer 47:207–212. Clavel J, Cordier S, Boccon-Gibod L, et al. 1989. Tobacco and bladder cancer in males: increased risk for inhalers and smokers of black tobacco. Int J Cancer 44:605–610. Clavel J, Mandereau L, Limasset JC, et al. 1994. Occupational exposure to polycyclic aromatic hydrocarbons and the risk of bladder cancer: a French case-control study. Int J Epidemiol 23:1145–1153. Clemmesen J. 1981. International Commission for Protection against Environmental Mutagens and Carcinogens. Epidemiological studies into the possible carcinogenicity of hair dyes. Mutat Res 87:65–79. Cole P, Hoover R, Friedell GH. 1972. Occupation and cancer of the lower urinary tract. Cancer 29:1250–1260. Cole P, Monson RR, Haning H, et al. 1971. Smoking and cancer of the lower urinary tract. N Engl J Med 284:129–134. Colt JS, Baris D, Stewart P, et al. 2004. Occupation and bladder cancer in a population-based case-control study in New Hampshire. Cancer Causes Control 15:759–769. Coombs LM, Pigott DA, Sweeney E, et al. 1991. Amplification and overexpression of c-erbB-2 in transitional cell carcinoma of the urinary bladder. Br J Cancer 63:601–608. Cordier S, Clavel J, Limasset JC, et al. 1993. Occupational risks of bladder cancer in France: a multicentre case-control study. Int J Epidemiol 22:403–411. Costantini AS, Paci E, Miligi L, et al. 1989. Cancer mortality among workers in the Tuscan tanning industry. Br J Ind Med 46:384–388. Council on Scientific Affairs. 1985. Saccharin. Review of safety issues. Council on Scientific Affairs. JAMA 254:2622–2624. Crump KS, Guess HA. 1982. Drinking water and cancer: review of recent epidemiological findings and assessment of risks. Annu Rev Public Health 3:339–357. Curigliano G, Zhang YJ, Wang LY, et al. 1996. Immunohistochemical quantitation of 4-aminobiphenyl-DNA adducts and p53 nuclear overexpression in T1 bladder cancer of smokers and nonsmokers. Carcinogenesis 17:911–916. Cutler SJ, Young JL Jr, eds. 1975. Third National Cancer Survery: incidence data. NCI Monogr 41:1–454. Cuzick J, Sasieni P, Evans S. 1992. Ingested arsenic, keratoses, and bladder cancer. Am J Epidemiol 136:417–421.

1120

PART IV: CANCER BY TISSUE OF ORIGIN

D’Avanzo B, La Vecchia C, Franceschi S, et al. 1992. Coffee consumption and bladder cancer risk. Eur J Cancer 28A:1480–1484. D’Avanzo B, La Vecchia C, Negri E, et al. 1995. Attributable risks for bladder cancer in northern Italy. Ann Epidemiol 5:427–431. D’Avanzo B, Negri E, La Vecchia C, et al. 1990. Cigarette smoking and bladder cancer. Eur J Cancer 26:714–718. Daly AK, Thomas DJ, Cooper J, et al. 1993. Homozygous deletion of gene for glutathione S-transferase M1 in bladder cancer. BMJ 307:481–482. De Stefani E, Correa P, Fierro L, et al. 1991. Black tobacco, mate, and bladder cancer. A case-control study from Uruguay. Cancer 67:536–540. Decarli A, Peto J, Piolatto G, et al. 1985. Bladder cancer mortality of workers exposed to aromatic amines: analysis of models of carcinogenesis. Br J Cancer 51:707–712. Decoufle P. 1979. Cancer risks associated with employment in the leather and leather products industry. Arch Environ Health 34:33–37. Decoufle P, Stanislawczyk K, Houten L, et al. 1977. A Retrospective Survey of Cancer in Relation to Occupation. DHEW Publication No. NIOSH. 77178. Washington, DC: U.S. Government Printing Office. Delzell E, Macaluso M, Cole P. 1989. A follow-up study of workers at a dye and resin manufacturing plant. J Occup Med 31:273–278. Delzell E, Monson RR. 1981. Mortality among rubber workers. III. Causespecific mortality, 1940–1978. J Occup Med 23:677–684. DeMarini DM, Brooks LR, Bhatnagar VK, et al. 1997a. Urinary mutagenicity as a biomarker in workers exposed to benzidine: correlation with urinary metabolites and urothelial DNA adducts. Carcinogenesis 18:981–988. DeMarini DM, Hastings SB, Brooks LR, et al. 1997b. Pilot study of free and conjugated urinary mutagenicity during consumption of pan-fried meats: possible modulation by cruciferous vegetables, glutathione S-transferaseM1, and N-acetyltransferase-2. Mutat Res 381:83–96. Derby LE, Jick H. 1996. Acetaminophen and renal and bladder cancer. Epidemiology 7:358–362. Devesa SS, Grauman DJ, Blot WJ, et al. 1999. Atlas of Cancer Mortality in the United States, 1950–94. Washington, DC: US Govt Print Off NIH Publ No. NIH. 99-4564. Dewan A, Chattopadhyay P, Kulkarni PK. 1995. N-acetyltransferase activity— a susceptibility factor in human bladder carcinogenesis. Indian J Cancer 32:15–19. Dolin PJ, Cook-Mozaffari P. 1992. Occupation and bladder cancer: a deathcertificate study. Br J Cancer 66:568–578. Dolin PJ, Darby SC, Beral V. 1994. Paraplegia and squamous cell carcinoma of the bladder in young women: findings from a case-control study. Br J Cancer 70:167–168. Donato F, Boffetta P, Fazioli R, et al. 1997. Bladder cancer, tobacco smoking, coffee and alcohol drinking in Brescia, northern Italy. Eur J Epidemiol 13:795–800. Dubrow R, Wegman DH. 1984. Cancer and occupation in Massachusetts: a death certificate study. Am J Ind Med 6:207–230. Dunham LJ, Rabson AS, Stewart HL, et al. 1968. Rates, interview, and pathology study of cancer of the urinary bladder in New Orleans, Louisiana. J Natl Cancer Inst 41:683–709. Edling C, Kling H, Flodin U, et al. 1986. Cancer mortality among leather tanners. Br J Ind Med 43:494–496. Edmonds CJ, Smith T. 1986. The long-term hazards of the treatment of thyroid cancer with radioiodine. Br J Radiol 59:45–51. El Morsi B, Sherif M, el Raziki ES. 1975. Experimental bilharzial squamous metaplasia of the urinary bladder in hamsters. Eur J Cancer 11:199– 201. El Rifai W, Kamel D, Larramendy ML, et al. 2000. DNA copy number changes in Schistosoma-associated and non-Schistosoma-associated bladder cancer. Am J Pathol 156:871–878. Elem B, Purohit R. 1983. Carcinoma of the urinary bladder in Zambia. A quantitative estimation of Schistosoma haematobium infection. Br J Urol 55:275–278. Engel LS, Taioli E, Pfeiffer R, et al. 2002. Pooled analysis and meta-analysis of glutathione S-transferase M1 and bladder cancer: a HuGE review. Am J Epidemiol 156:95–109. Engeland A, Andersen A, Haldorsen T, et al. 1996. Smoking habits and risk of cancers other than lung cancer: 28 years’ follow-up of 26,000 Norwegian men and women. Cancer Causes Control 7:497–506. Englund A. 1980. Cancer incidence among painters and some allied trades. J Toxicol Environ Health 6:1267–1273. Epstein JI, Amin MB, Reuter VR, et al. 1998. The World Health Organization/International Society of Urological Pathology consensus classification of urothelial transitional cell. neoplasms of the urinary bladder. Bladder Consensus Conference Committee. Am J Surg Pathol 22: 1435–1448. Esteve A, Sorlie T, Martel-Planche G, et al. 1995. Screening for p53 gene mutations in archived tumors of workers occupationally exposed to carcino-

gens: examples from analysis of bladder tumors. J Occup Environ Med 37:59–68. Fadl-Elmula I, Kytola S, Leithy ME, et al. 2002. Chromosomal aberrations in benign and malignant bilharzia-associated bladder lesions analyzed by comparative genomic hybridization. BMC Cancer 2:5. Falck K, Grohn P, Sorsa M, et al. 1979. Mutagenicity in urine of nurses handling cytostatic drugs. Lancet 1:1250–1251. Falck K, Sorsa M, Vainio H, et al. 1980. Mutagenicity in urine of workers in rubber industry. Mutat Res 79:45–52. Fan TY, Morrison J, Rounbehler DP, et al. 1977. N-Nitrosodiethanolamine in synthetic cutting fluids: a part-per-hundred impurity. Science 196:70–71. Feng Z, Hu W, Rom WN, et al. 2002. 4-aminobiphenyl is a major etiological agent of human bladder cancer: evidence from its DNA binding spectrum in human p53 gene. Carcinogenesis 23:1721–1727. Ferguson AR. 1911. Associated bilharziosis and primary malignant disease of the urinary bladder with observations in a series of forty cases. J Pathol Bacteriol 16:76–94. Fitzgerald JM, Ramchurren N, Rieger K, et al. 1995. Identification of H-ras mutations in urine sediments complements cytology in the detection of bladder tumors. J Natl Cancer Inst 87:129–133. Fokkens W. 1979. Phenacetin abuse related to bladder cancer. Environ Res 20:192–198. Fraser GE. 1999. Associations between diet and cancer, ischemic heart disease, and all-cause mortality in non-Hispanic white California Seventh-day Adventists. Am J Clin Nutr 70:532S–538S. Fraumeni JF Jr. 1968. Cigarette smoking and cancers of the urinary tract: geographic variation in the United States. J Natl Cancer Inst 41:1205– 1211. Fraumeni JF Jr, Scotto J, Dunham LJ. 1971. Coffee-drinking and bladder cancer. Lancet 2:1204. Freedman DM, Cantor KP, Lee NL, et al. 1997. Bladder cancer and drinking water: a population-based case-control study in Washington County, Maryland United States. Cancer Causes Control 8:738–744. Gago-Dominguez M, Bell DA, Watson MA, et al. 2003. Permanent hair dyes and bladder cancer: risk modification by cytochrome P4501A2 and Nacetyltransferases 1 and 2. Carcinogenesis 24:483–489. Gago-Dominguez M, Castelao JE, Yuan JM. 2001a. Use of permanent hair dyes and bladder cancer risk. Int J Cancer 91:575–579. Gago-Dominguez M, Chan KK, Ross RK, et al. 2001b. Permanent Hair Dyes and Bladder Cancer Risk. Int J Cancer 94:905–906. Garabrant DH, Wegman DH. 1984. Cancer mortality among shoe and leather workers in Massachusetts. Am J Ind Med 5:303–314. Garcia R, Gonzalez CA, Agudo A, et al. 1999. High intakes of specific carotenoids and flavonoids does not reduce the risk of bladder cancer. Nutr Cancer 35:212–214. Garcia-Closas M, Malats N, Real FX, et al. Genetic variation in the nucleotide excision repair pathway and bladder cancer risk. Cancer Epidemiol Biomarkers Prev. (In press). Garcia-Closas M, Malats N, Silverman D, et al. 2005. NAT2 slow acetylation, GSTM1 null genotype, and risk of bladder cancer: results from the Spanish Bladder Cancer Study and meta-analyses. Lancet 366:649–659. Garland M, Morris JS, Stampfer MJ, et al. 1995. Prospective study of toenail selenium levels and cancer among women. J Natl Cancer Inst 87:497–505. Garner RC, Mould AJ, Lindsay-Smith V, et al. 1982. Mutagenic urine from bladder cancer patients [letter]. Lancet 2:389. Gelfand M, Weinberg RW, Castle WM. 1967. Relation between carcinoma of the bladder and infestation with Schistosoma haematobium. Lancet 1:1249–1251. Geoffroy-Perez B, Cordier S. 2001. Fluid consumption and the risk of bladder cancer: results of a multicenter case-control study. Int J Cancer 93: 880–887. Gibbs GW. 1985. Mortality of aluminum reduction plant workers, 1950 through 1977. J Occup Med 27:761–770. Golka K, Prior V, Blaszkewicz M, et al. 1996. Occupational history and genetic N-acetyltransferase polymorphism in urothelial cancer patients of Leverkusen, Germany. Scand J Work Environ Health 22:332–338. Gonzalez CA, Errezola M, Izarzugaza I, et al. 1991. Urinary infection, renal lithiasis and bladder cancer in Spain. Eur J Cancer 27:498–500. Gonzalez CA, Lopez-Abente G, Errezola M, et al. 1985. Occupation, tobacco use, coffee, and bladder cancer in the county of Mataro Spain. Cancer 55:2031–2034. Greenblatt MS, Bennett WP, Hollstein M, et al. 1994. Mutations in the p53 tumor suppressor gene: clues to cancer etiology and molecular pathogenesis. Cancer Res 54:4855–4878. Guberan E, Usel M, Raymond L, et al. 1989. Disability, mortality, and incidence of cancer among Geneva painters and electricians: a historical prospective study. Br J Ind Med 46:16–23.

Bladder Cancer Habel LA, Bull SA, Friedman GD. 1998. Barbiturates, smoking, and bladder cancer risk. Cancer Epidemiol Biomarkers Prev 7:1049–1050. Habuchi T, Kinoshita H, Yamada H, et al. 1994. Oncogene amplification in urothelial cancers with p53 gene mutation or MDM2 amplification. J Natl Cancer Inst 86:1331–1335. Habuchi T, Takahashi R, Yamada H, et al. 1993. Influence of cigarette smoking and schistosomiasis on p53 gene mutation in urothelial cancer. Cancer Res 53:3795–3799. Hanke J, Krajewska B. 1990. Acetylation phenotypes and bladder cancer. J Occup Med 32:917–918. Hankey BF, Edwards BK, Ries LA, et al. 1991. Problems in cancer surveillance: delineating in situ and invasive bladder cancer. J Natl Cancer Inst 83:384–385. Hankey B, Myers M. 1987. Black/white differences in bladder cancer survival. J Chron Dis 40:65–73. Harries LW, Stubbins MJ, Forman D, et al. 1997. Identification of genetic polymorphisms at the glutathione S-transferase Pi locus and association with susceptibility to bladder, testicular and prostate cancer. Carcinogenesis 18:641–644. Harris AL, Neal DE. 1992. Bladder cancer—field versus clonal origin. N Engl J Med 326:759–761. Hartge P, Harvey EB, Linehan WM, et al. 1990. Unexplained excess risk of bladder cancer in men. J Natl Cancer Inst 82:1636–1640. Hartge P, Hoover R, Altman R, et al. 1982. Use of hair dyes and risk of bladder cancer. Cancer Res 42:4784–4787. Hartge P, Hoover R, Kantor A. 1985. Bladder cancer risk and pipes, cigars, and smokeless tobacco. Cancer 55:901–906. Hartge P, Hoover R, West DW, et al. 1983. Coffee drinking and risk of bladder cancer. J Natl Cancer Inst 70:1021–1026. Hartge P, Silverman D, Hoover R, et al. 1987. Changing cigarette habits and bladder cancer risk: a case-control study. J Natl Cancer Inst 78:1119– 1125. Hayes RB, Bi W, Rothman N, et al. 1993. N-acetylation phenotype and genotype risk of bladder cancer in benzidine-exposed workers. Carcinogenesis 14:675–678. Hazra A, Chamberlain RM, Grossman HB, et al. 2003. Death receptor 4 and bladder cancer risk. Cancer Res 63:1157–1159. Hein DW. 1988. Acetylator genotype and arylamine-induced carcinogenesis. Biochim Biophys Acta 948:37–66. Hein DW. 2002. Molecular genetics and function of NAT1 and NAT2: role in aromatic amine metabolism and carcinogenesis. Mutat Res 506–507: 65–77. Helzlsouer KJ, Comstock GW, Morris JS. 1989. Selenium, lycopene, alphatocopherol, beta-carotene, retinol, and subsequent bladder cancer. Cancer Res 49:6144–6148. Henley S, Thun MJ. 2001. Use of Permanent Hair Dyes and Bladder-Cancer Risk. Int J Cancer 94:903–904. Hennekens CH, Speizer FE, Rosner B, et al. 1979. Use of permanent hair dyes and cancer among registered nurses. Lancet 1:1390–1393. Henry SA, Kennaway NM, Kennaway EL. 1931. The incidence of cancer of the bladder and prostate in certain occupations. J Hyg 31:125– 137. Hicks RM. 1983. The scientific basis for regarding vitamin A and its analogues as anti-carcinogenic agents. Proc Nutr Soc 42:83–93. Hinder RA, Schmaman A. 1969. Bilharziasis and squamous carcinoma of the bladder. S Afr Med J 43:617–618. Hirvonen A, Nylund L, Kociba P, et al. 1994. Modulation of urinary mutagenicity by genetically determined carcinogen metabolism in smokers (published erratum appears in Carcinogenesis 1994. 15(8):1767). Carcinogenesis 15:813–815. Hoar SK, Hoover R. 1985. Truck driving and bladder cancer mortality in rural New England. J Natl Cancer Inst 74:771–774. Hoover R, Cole P. 1971. Population trends in cigarette smoking and bladder cancer. Am J Epidemiol 94:409–418. Hoover R, Hartge P. 1982. Non-nutritive sweeteners and bladder cancer. Am J Public Health 72:382–383. Hoover RN, Strasser PH. 1980. Artificial sweeteners and human bladder cancer. Preliminary results. Lancet 1:837–840. Hopenhayn-Rich C, Biggs ML, Fuchs A, et al. 1996. Bladder cancer mortality associated with arsenic in drinking water in Argentina. Epidemiology 7:117–124. Hours M, Dananche B, Fevotte J, et al. 1994. Bladder cancer and occupational exposures. Scand J Work Environ Health 20:322–330. Howe GR, Burch JD, Miller AB, et al. 1980. Tobacco use, occupation, coffee, various nutrients, and bladder cancer. J Natl Cancer Inst 64: 701–713. Howe GR, Burch JD, Miller AB, et al. 1977. Artificial sweeteners and human bladder cancer. Lancet 2:578–581.

1121

Hsieh FI, Pu YS, Chern HD, et al. 1999. Genetic polymorphisms of Nacetyltransferase 1 and 2 and risk of cigarette smoking-related bladder cancer. Br J Cancer 81:537–541. Hussain SP, Harris CC. 1998. Molecular epidemiology of human cancer: contribution of mutation spectra studies of tumor suppressor genes. Cancer Res 58:4023–4037. IARC. 1974. Evaluation of Carcinogenic Risk of Chemicals to Man. Some Aromatic Amines, Hydrazine and Related Substances, N-nitroso Compounds and Miscellaneous Alkylating Agents, Vol 4. Lyon: International Agency for Research on Cancer. IARC. 1980. Evaluation of the Carcinogenic Risk of Chemicals to Humans. Some Pharmaceutical Drugs, Vol. 24. Lyon: International Agency for Research on Cancer. IARC. 1981. Evaluation of the carcinogenic risk of chemicals to humans: Some Antineoplastic and Immunosuppressive Agents, Vol 26. Lyon: International Agency for Research on Cancer. IARC. 1982. Evaluation of the Carcinogenic Risk of Chemicals to Humans. Chemicals, Industrial Processes and Industries Associated with Cancer in Humans, Suppl 4. Lyon: International Agency for Research on Cancer. IARC. 1986. Evaluation of the Carcinogenic Risk of Chemicals to Humans: Tobacco Smoking, Vol 38. Lyon: International Agency for Research on Cancer. IARC. 1987. An updating of selected IARC monographs Volumes 1 to 42, Supply 7, Lyon: International Agency for Research on Cancer. IARC. 1994. Evaluation of Carcinogenic Risks to Humans: Schistosome, Liver flukes and Helicobacter pylori, Vol 61. Lyon: International Agency for Research on Cancer. IARC. 1999. Paracetamol: IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. Lyon: International Agency for Research on Cancer 73:401–449. Inskip PD, Monson RR, Wagoner JK, et al. 1990. Cancer mortality following radium treatment for uterine bleeding. Radiat Res 123:331–344. Iscovich J, Castelletto R, Esteve J, et al. 1987. Tobacco smoking, occupational exposure and bladder cancer in Argentina. Int J Cancer 40:734–740. Issa JP. 2003. Age-related epigenetic changes and the immune system. Clin Immunol 109:103–108. Jacobs EJ, Henion AK, Briggs PJ, et al. 2002. Vitamin C and vitamin E supplement use and bladder cancer mortality in a large cohort of US men and women. Am J Epidemiol 156:1002–1010. Jemal A, Murray T, Samuels A, et al. 2003. Cancer statistics, 2003. CA Cancer J Clin 53:5–26. Jensen OM, Kamby C. 1982. Intra-uterine exposure to saccharin and risk of bladder cancer in man. Int J Cancer 29:507–509. Jensen OM, Knudsen JB, Sorensen BL, et al. 1983. Artificial sweeteners and absence of bladder cancer risk in Copenhagen. Int J Cancer 32:577–582. Jensen OM, Wahrendorf J, Blettner M, et al. 1987a. The Copenhagen casecontrol study of bladder cancer: role of smoking in invasive and noninvasive bladder tumours. J Epidemiol Community Health 41:30–36. Jensen OM, Wahrendorf J, Knudsen JB, et al. 1986. The Copenhagen casecontrol study of bladder cancer II. Effect of coffee and other beverages. Int J Cancer 37:651–657. Jensen OM, Wahrendorf J, Knudsen JB, et al. 1987b. The Copenhagen casereferent study on bladder cancer. Risks among drivers, painters and certain other occupations. Scand J Work Environ Health 13:129–134. Kabat GC, Dieck GS, Wynder EL. 1986. Bladder cancer in nonsmokers. Cancer 57:362–367. Kadlubar FF, Dooley KL, Teitel CH, et al. 1991. Frequency of urination and its effects on metabolism, pharmacokinetics, blood hemoglobin adduct formation, and liver and urinary bladder DNA adduct levels in beagle dogs given the carcinogen 4-aminobiphenyl. Cancer Res 51:4371–4377. Kadlubar FF, Miller JA, Miller EC. 1977. Hepatic microsomal N-glucuronidation and nucleic acid binding of N-hydroxy arylamines in relation to urinary bladder carcinogenesis. Cancer Res 37:805–814. Kadlubar FF, Unruh LE, Flammang TJ, et al. 1981. Alteration of urinary levels of the carcinogen, N-hydroxy-2-naphthylamine, and its N-glucuronide in the rat by control of urinary pH, inhibition of metabolic sulfation, and changes in biliary excretion. Chem Biol Interact 33:129–147. Kahn HA. 1966. The Dorn study of smoking and mortality among U.S. veterans: report on eight and one-half years of observation. Natl Cancer Inst Monogr 19:1–125. Kaldor JM, Day NE, Kittelmann B, et al. 1995. Bladder tumours following chemotherapy and radiotherapy for ovarian cancer: a case-control study. Int J Cancer 63:1–6. Kallioniemi A, Kallioniemi OP, Citro G, et al. 1995. Identification of gains and losses of DNA sequences in primary bladder cancer by comparative genomic hybridization. Genes Chromosomes Cancer 12:213–219. Kannio A, Ridanpaa M, Koskinen H, et al. 1996. A molecular and epidemiological study on bladder cancer: p53 mutations, tobacco smoking, and

1122

PART IV: CANCER BY TISSUE OF ORIGIN

occupational exposure to asbestos. Cancer Epidemiol Biomarkers Prev 5:33–39. Kantor AF, Hartge P, Hoover RN, et al. 1985. Familial and environmental interactions in bladder cancer risk. Int J Cancer 35:703–706. Kantor AF, Hartge P, Hoover RN, et al. 1984. Urinary tract infection and risk of bladder cancer. Am J Epidemiol 119:510–515. Karagas MR, Tosteson TD, Morris JS, et al. 2004. Incidence of transitional cell carcinoma of the bladder and arsenic exposure in New Hampshire. Cancer Causes Control 15:265–272. Katoh T, Inatomi H, Nagaoka A, et al. 1995. Cytochrome P4501A1 gene polymorphism and homozygous deletion of the glutathione S-transferase M1 gene in urothelial cancer patients. Carcinogenesis 16:655–657. Katoh T, Kaneko S, Takasawa S, et al. 1999. Human glutathione S-transferase P1 polymorphism and susceptibility to smoking related epithelial cancer; oral, lung, gastric, colorectal and urothelial cancer. Pharmacogenetics 9:165–169. Kaye JA, Myers MW. 2001. Acetaminophen and the risk of renal and bladder cancer in the general practice research database. Epidemiology 12: 690–694. Kempkes M, Golka K, Reich S, et al. 1996. Glutathione S-transferase GSTM1 and GSTT1 null genotypes as potential risk factors for urothelial cancer of the bladder. Arch Toxicol 71:123–126. Kessler II. 1970. Cancer mortality among diabetics. J Natl Cancer Inst 44:673–686. Kessler II, Clark JP. 1978. Saccharin, cyclamate, and human bladder cancer. No evidence of an association. JAMA 240:349–355. Ketterer B. 1988. Protective role of glutathione and glutathione transferases in mutagenesis and carcinogenesis. Mutat Res 202:343–361. Kiemeney LA, Schoenberg M. 1996. Familial transitional cell carcinoma. J Urol 156:867–872. Kimura F, Florl AR, Seifert HH, et al. 2001. Destabilization of chromosome 9 in transitional cell carcinoma of the urinary bladder. Br J Cancer 85:1887–1893. King WD, Marrett LD. 1996. Case-control study of bladder cancer and chlorination by-products in treated water Ontario, Canada. Cancer Causes Control 7:596–604. Kjaer SK, Knudsen JB, Sorensen BL, et al. 1989. The Copenhagen case-control study of bladder cancer. V. Review of the role of urinary-tract infection. Acta Oncol 28:631–636. Knight A, Askling J, et al. 2002. Cancer incidence in a population-based cohort of patients with Wegener’s granulomatosis. Int J Cancer 100:82–85. Knowles MA. 1999. The genetics of transitional cell carcinoma: progress and potential clinical application. BJU Int 84:412–427. Knowles MA, Williamson M. 1993. Mutation of H-ras is infrequent in bladder cancer: confirmation by single-strand conformation polymorphism analysis, designed restriction fragment length polymorphisms, and direct sequencing. Cancer Res 53:133–139. Kogevinas M, Sala M, Boffetta P, et al. 1998. Cancer risk in the rubber industry: a review of the recent epidemiological evidence. Occup Environ Med 55:1–12. Kogevinas M, t’Mannetje A, Cordier S, et al. 2003. Occupation and bladder cancer among men in Western Europe. Cancer Causes Control 14:907–914. Kramer AA, Graham S, Burnett WS, et al. 1991. Familial aggregation of bladder cancer stratified by smoking status. Epidemiology 2:145–148. Kuntz RE, Cheever AW, Myers BJ. 1972. Proliferative epithelial lesions of the urinary bladder of nonhuman primates infected with Schistosoma haematobium. J Natl Cancer Inst 48:223–235. Kunze E, Chang-Claude J, Frentzel-Beyme R. 1992. Life style and occupational risk factors for bladder cancer in Germany. A case-control study. Cancer 69:1776–1790. La Vecchia C, Lucchini F, Negri E, et al. 1999. Urinary bladder cancer death rates in Europe letter. Ann Oncol 10:1529–1530. La Vecchia C, Negri E. 1996. Nutrition and bladder cancer. Cancer Causes Control 7:95–100. La Vecchia C, Negri E, D’Avanzo B, et al. 1990. Occupation and the risk of bladder cancer. Int J Epidemiol 19:264–268. La Vecchia C, Negri E, Decarli A, et al. 1989. Dietary factors in the risk of bladder cancer. Nutr Cancer 12:93–101. La Vecchia C, Tavani A. 1995. Epidemiological evidence on hair dyes and the risk of cancer in humans. Eur J Cancer Prev 4:31–43. Ladero JM, Kwok CK, Jara C, et al. 1985. Hepatic acetylator phenotype in bladder cancer patients. Ann Clin Res 17:96–99. Lafuente A, Pujol F, Carretero P, et al. 1993. Human glutathione S-transferase mu GST mu. deficiency as a marker for the susceptibility to bladder and larynx cancer among smokers. Cancer Lett 68:49–54. Landi MT, Zocchetti C, Bernucci I, et al. 1996. Cytochrome P4501A2: enzyme induction and genetic control in determining 4-aminobiphenyl-

hemoglobin adduct levels. Cancer Epidemiol Biomarkers Prev 5:693– 698. LaRue H, Allard P, Simoneau M, et al. 2000. P53 point mutations in initial superficial bladder cancer occur only in tumors from current or recent cigarette smokers. Carcinogenesis 21:101–106. Lee SJ, Cho SH, Park SK, et al. 2002. Combined effect of glutathione Stransferase M1 and T1 genotypes on bladder cancer risk. Cancer Lett 177:173–179. Lee SW, Jang IJ, Shin SG, et al. 1994. CYP1A2 activity as a risk factor for bladder cancer. J Korean Med Sci 9:482–489. Levine LA, Richie JP. 1989. Urological complications of cyclophosphamide. J Urol 141:1063–1069. Lianes P, Orlow I, Zhang ZF, et al. 1994. Altered patterns of MDM2 and TP53 expression in human bladder cancer. J Natl Cancer Inst 86: 1325–1330. Lin HJ, Han CY, Bernstein DA, et al. 1994. Ethnic distribution of the glutathione transferase Mu 1-1 GSTM1. null genotype in 1473 individuals and application to bladder cancer susceptibility. Carcinogenesis 15: 1077–1081. Lin J, Kadlubar FF, Spitz MR, et al. 2005. A modified host cell reactivation assay to measure DNA repair capacity for removing 4-aminobiphenyl adducts: a pilot study of bladder cancer. Cancer Epidemiol Biomarkers Prev 14:1832–1836. Lockwood K. 1961. On the etiology of bladder tumors in KobenhavnFrederiksberg: an inquiry of 369 patients and 369 controls. Acta Path Microbiol Scand 51:1–166. Logan WP. 1982. Cancer mortality by occupation and social class, 1851–1971. IARC Publ No. 36. Lyon: International Agency for Research on Cancer. Lopez-Abente G, Gonzalez CA, Errezola M, et al. 1991. Tobacco smoke inhalation pattern, tobacco type, and bladder cancer in Spain. Am J Epidemiol 134:830–839. Lower GM Jr, Nilsson T, Nelson CE, et al. 1979. N-acetyltransferase phenotype and risk in urinary bladder cancer: approaches in molecular epidemiology. Preliminary results in Sweden and Denmark. Environ Health Perspect 29:71–79. Lunec J, Challen C, Wright C, et al. 1992. c-erbB-2 amplification and identical p53 mutations in concomitant transitional carcinomas of renal pelvis and urinary bladder. Lancet 339:439–440. Lynch CF, Platz CE, Jones MP, et al. 1991. Cancer registry problems in classifying invasive bladder cancer. J Natl Cancer Inst 83:429–433. Maclure M, Bryant MS, Skipper PL, et al. 1990. Decline of the hemoglobin adduct of 4-aminobiphenyl during withdrawal from smoking. Cancer Res 50:181–184. Maffi L, Vineis P. 1986. Occupation and bladder cancer in females. Med Lav 77:511–514. Malaveille C, Vineis P, Esteve J, et al. 1989. Levels of mutagens in the urine of smokers of black and blond tobacco correlate with their risk of bladder cancer. Carcinogenesis 10:577–586. Malker HS, McLaughlin JK, Silverman DT, et al. 1987. Occupational risks for bladder cancer among men in Sweden. Cancer Res 47:6763–6766. Marcus PM, Hayes RB, Vineis P, et al. 2000a. Cigarette smoking, Nacetyltransferase 2 acetylation status, and bladder cancer risk: a caseseries meta-analysis of a gene-environment interaction. Cancer Epidemiol Biomarkers Prev 9:461–467. Marcus PM, Vineis P, Rothman N. 2000b. NAT2 slow acetylation and bladder cancer risk: a meta-analysis of 22 case-control studies conducted in the general population. Pharmacogenetics 10:115–122. Markl ID, Cheng J, Liang G, et al. 2001. Global and gene-specific epigenetic patterns in human bladder cancer genomes are relatively stable in vivo and in vitro over time. Cancer Res 61:5875–5884. Marrett LD, Hartge P, Meigs JW. 1986. Bladder cancer and occupational exposure to leather. Br J Ind Med 43:96–100. Martone T, Vineis P, Malaveille C, et al. 2000. Impact of polymorphisms in xenoendo biotic metabolism on pattern and frequency of p53 mutations in bladder cancer. Mutat Res 462:303–309. Maruyama R, Toyooka S, Toyooka KO, et al. 2001. Aberrant promoter methylation profile of bladder cancer and its relationship to clinicopathological features. Cancer Res 61:8659–8663. Matanoski GM, Elliott EA. 1981. Bladder cancer epidemiology. Epidemiol Rev 3:203–229. Matullo G, Guarrera S, Carturan S, et al. 2001. DNA repair gene polymorphisms, bulky DNA adducts in white blood cells and bladder cancer in a case-control study. Int J Cancer 92:562–567. Matullo G, Guarrera S, Sacerdote C, et al. 2005. Polymorphisms/haplotypes in DNA repair genes and smoking: a bladder cancer case-control study. Cancer Epidemiol Biomarkers Prev 14:2569–2578. McCredie M, Stewart JH. 1988. Does paracetamol cause urothelial cancer or renal papillary necrosis? Nephron 49:296–300.

Bladder Cancer McCredie M, Stewart JH, Ford JM, et al. 1983. Phenacetin-containing analgesics and cancer of the bladder or renal pelvis in women. Br J Urol 55:220–224. McDonald DF, Lund RR. 1954. The role of the urine in vesical neoplasm. 1. Experimental confirmation of the urogenous theory of pathogenesis. J Urol 71:560–570. McGeehin MA, Reif JS, Becher JC, et al. 1993. Case-control study of bladder cancer and water disinfection methods in Colorado. Am J Epidemiol 138:492–501. McLaughlin JK, Zdenek H, Blot WJ, et al. 1995. Smoking and cancer mortality among U.S. veterans: a 26-year follow-up. Int J Cancer 60: 190–193. McMichael AJ, Andjelkovic DA, Tyroler HA. 1976. Cancer mortality among rubber workers: an epidemiologic study. Ann N Y Acad Sci 271:125–137. Mee AD. 1982. Aetiological aspects of bladder cancer in urinary bilharzia. Saudi Med J 3:123–128. Meigs JW, Marrett LD, Ulrich FU, et al. 1986. Bladder tumor incidence among workers exposed to benzidine: a thirty-year follow-up. J Natl Cancer Inst 76:1–8. Mellon JK, Lunec J, Wright C, et al. 1996. C-erbB-2 in bladder cancer: molecular biology, correlation with epidermal growth factor receptors and prognostic value. J Urol 155:321–326. Menon MM, Bhide SV. 1984. Mutagenicity of urine of bidi and cigarette smokers and tobacco chewers. Carcinogenesis 5:1523–1524. Mettlin C, Graham S. 1979. Dietary risk factors in human bladder cancer. Am J Epidemiol 110:255–263. Michaud DS, Clinton SK, Rimm EB, et al. 2001. Risk of bladder cancer by geographic region in a U.S. cohort of male health professionals. Epidemiology 12:719–726. Michaud DS, Spiegelman D, Clinton SK, et al. 2000. Prospective study of dietary supplements, macronutrients, micronutrients, and risk of bladder cancer in US men. Am J Epidemiol 152:1145–1153. Michaud DS, Spiegelman D, Clinton SK, et al. 1999a. Fluid intake and the risk of bladder cancer in men. N Engl J Med 340:1390–1397. Michaud DS, Spiegelman D, Clinton SK, et al. 1999b. Fruit and vegetable intake and incidence of bladder cancer in a male prospective cohort. J Natl Cancer Inst 91:605–613. Michaud DS, Wright ME, Cantor KP, et al. 2004. Arsenic concentrations in prediagnostic toenails and the risk of bladder cancer in a cohort study of male smokers. Am J Epidemiol 160:853–859. Mikoczy Z, Schutz A, Hagmar L. 1994. Cancer incidence and mortality among Swedish leather tanners. Occup Environ Med 51:530–535. Milham S Jr. 1976. Occupational mortality in Washington State, 1950–1971. DHEW Publ No. NIOSH. 76175C. Washington, DC, US Govt Print Off. Miller AB, Howe GR. 1977. Artificial sweeteners and bladder cancer. Lancet 2:1221–1222. Miller BA, Silverman DT, Hoover RN, et al. 1986. Cancer risk among artistic painters. Am J Ind Med 9:281–287. Miller CT, Neutel CI, Nair RC, et al. 1978. Relative importance of risk factors in bladder carcinogenesis. J Chronic Dis 31:51–56. Mills PK, Beeson WL, Phillips RL, et al. 1991. Bladder cancer in a low risk population: results from the Adventist Health Study. Am J Epidemiol 133:230–239. Mohtashamipur E, Norpoth K, Lieder F. 1987. Urinary excretion of mutagens in smokers of cigarettes with various tar and nicotine yields, black tobacco, and cigars. Cancer Lett 34:103–112. Momas I, Daures JP, Festy B, et al. 1994a. Bladder cancer and black tobacco cigarette smoking. Some results from a French case-control study. Eur J Epidemiol 10:599–604. Momas I, Daures JP, Festy B, et al. 1994b. Relative importance of risk factors in bladder carcinogenesis: some new results about Mediterranean habits. Cancer Causes Control 5:326–332. Mommsen S, Aagaard J. 1983a. A case-control study of bladder-cancer. A multivariate, stratified analysis of a low-risk population. Dan Med Bull 30:427–432. Mommsen S, Aagaard J, Sell A. 1982. An epidemiological case-control study of bladder cancer in males from a predominantly rural district. Eur J Cancer Clin Oncol 18:1205–1210. Mommsen S, Aagaard J, Sell A. 1983b. A case-control study of female bladder cancer. Eur J Cancer Clin Oncol 19:725–729. Monson RR, Nakano KK. 1976. Mortality among rubber workers. I. White male union employees in Akrob, Ohio. Am J Epidemiol 103:284–296. Montanaro F, Ceppi M, Demers PA, et al. 1997. Mortality in a cohort of tannery workers. Occup Environ Med 54:588–591. Moore LE, Smith AH, Eng C, et al. 2003. P53 alterations in bladder tumors from arsenic and tobacco exposed patients. Carcinogenesis 24: 1785–1791.

1123

Moore LE, Smith AH, Eng C, et al. 2002. Arsenic-related chromosomal alterations in bladder cancer. J Natl Cancer Inst 94:1688–1696. Moore LE, Smith AH, Hopenhayn-Rich C, et al. 1997. Micronuclei in exfoliated bladder cells among individuals chronically exposed to arsenic in drinking water. Cancer Epidemiol Biomarkers Prev 6:31–36. Morgan RW, Jain MG. 1974. Bladder cancer: smoking, beverages and artificial sweeteners. Can Med Assoc J 111:1067–1070. Morrison AS. 1979. Public health value of using epidemiologic information to identify high-risk groups for bladder cancer screening. Semin Oncol 6:184–188. Morrison AS. 1984a. Control of cigarette smoking in evaluating the association of coffee drinking and bladder cancer. In: Coffee and Health, Banbury Report 17, MacMahon B, Sugimura T (eds.) Cold Spring Harbor, NY: Cold Spring Harbor Laboratory, pp. 127–136. Morrison AS, Ahlbom A, Verhoek WG, et al. 1985. Occupation and bladder cancer in Boston, USA, Manchester, UK, and Nagoya, Japan. J Epidemiol Community Health 39:294–300. Morrison AS, Buring JE. 1980. Artificial sweeteners and cancer of the lower urinary tract. N Engl J Med 302:537–541. Morrison AS, Buring JE, Verhoek WG, et al. 1982a. Coffee drinking and cancer of the lower urinary tract. J Natl Cancer Inst 68:91–94. Morrison AS, Buring JE, Verhoek WG, et al. 1984b. An international study of smoking and bladder cancer. J Urol 131:650–654. Morrison AS, Verhoek WG, Leck I, et al. 1982b. Artificial sweeteners and bladder cancer in Manchester, U.K., and Nagoya, Japan. Br J Cancer 45:332–336. Morton KC, Wang CY, Garner CD, et al. 1981. Carcinogenicity of benzidine, N,N¢-diacetylbenzidine, and N-hydroxy-N,N¢-diacetylbenzidine for female CD rats. Carcinogenesis 2:747–752. Mostafa MH, Sheweita SA, O’Connor PJ. 1999. Relationship between schistosomiasis and bladder cancer. Clin Microbiol Rev 12:97–111. Murata M, Takayama K, Choi BC, et al. 1996. A nested case-control study on alcohol drinking, tobacco smoking, and cancer. Cancer Detect Prev 20:557–565. Muscheck M, Abol-Enein H, Chew K, et al. 2000. Comparison of genetic changes in schistosome-related transitional and squamous bladder cancers using comparative genomic hybridization. Carcinogenesis 21:1721–1726. Mustacchi P, Shimkin MB. 1985. Cancer of the bladder and infestation with Schistosoma hematobium. J Natl Cancer Inst 20:825–842. Muto S, Horie S, Takahashi S, et al. 2000. Genetic and epigenetic alterations in normal bladder epithelium in patients with metachronous bladder cancer. Cancer Res 60:4021–4025. Myslak ZW, Bolt HM, Brockmann W. 1991. Tumors of the urinary bladder in painters: a case-control study. Am J Ind Med 19:705–713. Nagano J, Kono S, Preston DL, et al. 2000. Bladder-cancer incidence in relation to vegetable and fruit consumption: a prospective study of atomicbomb survivors. Int J Cancer 86:132–138. Najem GR, Louria DB, Seebode JJ, et al. 1982. Life time occupation, smoking, caffeine, saccharine, hair dyes and bladder carcinogenesis. Int J Epidemiol 11:212–217. National Cancer Institute. 2002. Report of the Kidney/Bladder Cancers Progress Review Group. NIH Publ No. 02-5197. Bethesda, Md: NCI. National Research Council. 1980. Drinking Water and Health. Vol. 3. Washington, DC: Natl Acad Press, 5–21. National Research Council and Subcomittee in Arsenic in Drinking Water. 1999. Arsenic in Drinking Water. Washington, DC, National Academy Press. Negri E, Piolatto G, Pira E, et al. 1989. Cancer mortality in a northern Italian cohort of rubber workers. Br J Ind Med 46:624–628. Nishiyama H, Takahashi T, Kakehi Y, et al. 1999. Homozygous deletion at the 9q32-33 candidate tumor suppressor locus in primary human bladder cancer. Genes Chromosomes Cancer 26:171–175. Nomura AM, Kolonel LN, Hankin JH, et al. 1991. Dietary factors in cancer of the lower urinary tract. Int J Cancer 48:199–205. Nomura A, Kolonel LN, Yoshizawa CN. 1989. Smoking, alcohol, occupation, and hair dye use in cancer of the lower urinary tract. Am J Epidemiol 130:1159–1163. Nomura AM, Stemmermann GN, Heilbrun LK, et al. 1985. Serum vitamin levels and the risk of cancer of specific sites in men of Japanese ancestry in Hawaii. Cancer Res 45:2369–2372. Okkels H, Sigsgaard T, Wolf H, et al. 1997. Arylamine N-acetyltransferase 1 NAT1. and 2 NAT2. polymorphisms in susceptibility to bladder cancer: the influence of smoking. Cancer Epidemiol Biomarkers Prev 6:225–231. Olsen JH, Wallin H, Boice JD Jr, et al. 1993. Phenobarbital, drug metabolism, and human cancer. Cancer Epidemiol Biomarkers Prev 2:449–452. Oyasu R, Hirao Y, Izumi K. 1981. Enhancement by urine of urinary bladder carcinogenesis. Cancer Res 41:478–481.

1124

PART IV: CANCER BY TISSUE OF ORIGIN

Page WF, Kuntz AJ. 1980. Racial and socioeconomic factors in cancer survival. A comparison of Veterans Administration results with selected studies. Cancer 45:1029–1040. Pannek J. 2002. Transitional cell carcinoma patients with spinal cord injury: a high risk malignancy? Urology 59:240–244. Park SJ, Zhao H, Spitz MR, et al. 2003. An association between NQO1 genetic polymorphism and risk of bladder cancer. Mutat Res 536: 131–137. Parkash O, Kiesswetter H. 1976. The role of urine in the etiology of cancer of the urinary bladder. Urol Int 31:343–348. Parkes HG, Veys CA, Waterhouse JA, et al. 1982. Cancer mortality in the British rubber industry. Br J Ind Med 39:209–220. Parkin DM, Bray FI. 2003. International Patterns of Cancer Incidence and Mortality. In: Cancer Epidemiology and Prevention, Schottenfeld D, Fraumeni J (eds). Parkin DM, Whelan SL, Ferlay J, et al. 2002. Cancer Incidence in Five Continents, Vol. VIII. IARC Publ. No. 155. Lyon: International Agency for Research on Cancer. Pashos CL, Botteman MF, Laskin BL, et al. 2002. Bladder cancer: Epidemiology, diagnosis, and management. Cancer Pract 10:311–322. Paterson RF, Ulbright TM, MacLennan GT, et al. 2003. Molecular genetic alterations in the laster-capture-microdissected stroma adjacent to bladder carcinoma. Cancer 98:1830–1836. Pedersen-Bjergaard J, Ersboll J, Hansen VL, et al. 1988. Carcinoma of the urinary bladder after treatment with cyclophosphamide for nonHodgkin’s lymphoma. N Engl J Med 318:1028–1032. Pelucchi C, La Vecchia C, Negri E, et al. 2002a. Smoking and other risk factors for bladder cancer in women. Prev Med 35:114–120. Pelucchi C, Negri E, Franceschi S, et al. 2002b. Alcohol drinking and bladder cancer. J Clin Epidemiol 55:637–641. Peluso M, Airoldi L, Magagnotti C, et al. 2000. White blood cell DNA adducts and fruit and vegetable consumption in bladder cancer. Carcinogenesis 21:183–187. Perneger TV. 2001. Sex, smoking, and cancer: a reappraisal. J Natl Cancer Inst 93:1600–1602. Peters U, Sinha R, Bell DA, et al. 2004. Urinary mutagenesis and fried red meat intake: influence of cooking temperature, phenotype, and genotype of metabolizing enzymes in a controlled feeding study. Environ Mol Mutagen 43:53–74. Piolatto G, Negri E, La Vecchia C, et al. 1991. Bladder cancer mortality of workers exposed to aromatic amines: an updated analysis. Br J Cancer 63:457–459. Piper JM, Matanoski GM, Tonascia J. 1986. Bladder cancer in young women. Am J Epidemiol 123:1033–1042. Piper JM, Tonascia J, Matanoski GM. 1985. Heavy phenacetin use and bladder cancer in women aged 20 to 49 years. N Engl J Med 313:292–295. Pitard A, Brennan P, Clavel J, et al. 2001. Cigar, pipe, and cigarette smoking and bladder cancer risk in European men. Cancer Causes Control 12:551–556. Plna K, Hemminki K. 2001. Familial bladder cancer in the National Swedish Family Cancer Database. J Urol 166:2129–2133. Pohlabeln H, Jockel KH, Bolm-Audorff U. 1999. Non-occupational risk factors for cancer of the lower urinary tract in Germany. Eur J Epidemiol 15:411–419. Pommer W, Bronder E, Klimpel A, et al. 1999. Urothelial cancer at different tumour sites: role of smoking and habitual intake of analgesics and laxatives. Results of the Berlin Urothelial Cancer Study. Nephrol Dial Transplant 14:2892–2897. Popp W, Schmieding W, Speck M, et al. 1992. Incidence of bladder cancer in a cohort of workers exposed to 4-chloro-o-toluidine while synthesising chlordimeform. Br J Ind Med 49:529–531. Porru S, Aulenti V, Donato F, et al. 1996. Bladder cancer and occupation: a case-control study in northern Italy. Occup Environ Med 53:6–10. Preston DL, Shimizu Y, Pierce DA, et al. 2003. Studies of mortality of atomic bomb survivors. Report 13: Solid cancer and noncancer disease mortality: 1950–1997. Radiat Res 160:381–407. Probst-Hensch NM, Bell DA, Watson MA, et al. 2000. N-acetyltransferase 2 phenotype but not NAT1*10 genotype affects aminobiphenylhemoglobin adduct levels. Cancer Epidemiol Biomarkers Prev 9:619– 623. Proctor AJ, Coombs LM, Cairns JP, et al. 1991. Amplification at chromosome 11q13 in transitional cell tumours of the bladder. Oncogene 6:789– 795. Prout GR Jr, Wesley MN, Greenberg RS, et al. 2000. Bladder cancer: race differences in extent of disease at diagnosis. Cancer 89:1349–1358. Puente D, Hartge P, Greiser E, et al. 2006. A pooled analysis of bladder cancer case-control studies evaluating smoking in men and women. Cancer Causes Control 17:71–79.

Pujolar AE, Gonzalez CA, Lopez-Abente G, et al. 1993. Bladder cancer and coffee consumption in smokers and nonsmokers in Spain. Int J Epidemiol 22:38–44. Puntoni R, Bolognesi C, Bonassi S, et al. 1988. Cancer risk evaluation in an area with a high density of chemical plants. An interdisciplinary approach. Ann N Y Acad Sci 534:808–816. Radosauljevic V, Jankavic S, Marinkovic J, Djokic M. 2003. Fluid intake and bladder cancer. A case-control study. Neoplasma 50:234–238. Rebelakos A, Trichopoulos D, Tzonou A, et al. 1985. Tobacco smoking, coffee drinking, and occupation as risk factors for bladder cancer in Greece. J Natl Cancer Inst 75:455–461. Remer T, Manz F. 1995. Potential renal acid load of foods and its influence on urine pH. J Am Diet Assoc 95:791–797. Riboli E, Gonzalez CA, Lopez-Abente G, et al. 1991. Diet and bladder cancer in Spain: a multi-centre case-control study. Int J Cancer 49:214–219. Ries LAG, Eisner MP, Kosary CL, et al. (eds). 2003. SEER Cancer Statistics Review, 1975–2000. Bethesda, MD, National Cancer Institute. Risch A, Wallace DM, Bathers S, et al. 1995. Slow N-acetylation genotype is a susceptibility factor in occupational and smoking related bladder cancer. Hum Mol Genet 4:231–236. Risch HA, Burch JD, Miller AB, et al. 1988a. Dietary factors and the incidence of cancer of the urinary bladder. Am J Epidemiol 127:1179–1191. Risch HA, Burch JD, Miller AB, et al. 1988b. Occupational factors and the incidence of cancer of the bladder in Canada. Br J Ind Med 45:361– 367. Rockette HE, Arena VC. 1983. Mortality studies of aluminum reduction plant workers: potroom and carbon department. J Occup Med 25:549–557. Romanenko A, Lee CC, Yamamoto S, et al. 1999. Urinary bladder lesions after the Chernobyl accident: immunohistochemical assessment of p53, proliferating cell nuclear antigen, cyclin D1 and p21WAF1/Cip1. Jpn J Cancer Res 90:144–153. Romanenko A, Morimura K, Wanibuchi H, et al. 2000. Increased oxidative stress with gene alteration in urinary bladder urothelium after the Chernobyl accident. Int J Cancer 86:790–798. Romanenko A, Morimura K, Wanibuchi H, et al. 2003. Urinary bladder lesions induced by persistent chronic low-dose ionizing radiation. Cancer Sci 94:328–333. Romanenko A, Morimura K, Wei M, et al. 2002. DNA damage repair in bladder urothelium after the Chernobyl accident in Ukraine. J Urol 168:973– 977. Romano G, Garagnani L, Boninsegna A, et al. 1999. Analysis of 4-ABP-DNA adducts and p53 alterations in urinary bladder carcinoma. Anticancer Res 19:4571–4576. Romundstad P, Andersen A, Haldorsen T. 2000. Cancer incidence among workers in six Norwegian aluminum plants. Scand J Work Environ Health 26:461–269. Ross RK, Jones PA, Yu MC 1996. Bladder cancer epidemiology and pathogenesis. Semin Oncol 23:536–545. Rothman N, Bhatnagar VK, Hayes RB, et al. 1996. The impact of interindividual variation in NAT2 activity on benzidine urinary metabolites and urothelial DNA adducts in exposed workers. Proc Natl Acad Sci 93:5084–5089. Rothman N, Talaska G, Hayes RB, et al. 1997. Acidic urine pH is associated with elevated levels of free urinary benzidine and N-acetylbenzidine and urothelial cell DNA adducts in exposed workers. Cancer Epidemiol Biomarkers Prev 6:1039–1042. Rowland PG, Henneberry MO, Oyasu R, et al. 1980. Effects of urine and continued exposure to carcinogen on progression of early neoplastic urinary bladder lesions. Cancer Res 40:4524–4527. Rubino GF, Scansetti G, Piolatto G, et al. 1982. The carcinogenic effect of aromatic amines: an epidemiological study on the role of o-toluidine and 4,4¢-methylene bis 2-methylaniline. in inducing bladder cancer in man. Environ Res 27:241–254. Sala M, Cordier S, Chang-Claude J, et al. 2000. Coffee consumption and bladder cancer in nonsmokers: a pooled analysis of case-control studies in European countries. Cancer Causes Control 11:925–931. Sanchez-Carbayo M, Socci ND, Lozano JJ, et al. 2003. Gene discovery in bladder cancer progression using cDNA microarray. Am J Pathology 163:505–516. Sandler DP, Everson RB, Wilcox AJ. 1985. Passive smoking in adulthood and cancer risk. Am J Epidemiol 121:37–48. Sanyal S, Festa F, Sakano S, et al. 2004. Polymorphisms in DNA repair and metabolic genes in bladder cancer. Carcinogenesis 25:729–734. Sato K, Moriyama M, Mori S, et al. 1992. An immunohistologic evaluation of C-erbB-2 gene product in patients with urinary bladder carcinoma. Cancer 70:2493–2498. Sauter G, Moch H, Moore D, et al. 1993. Heterogeneity of erbB-2 gene amplification in bladder cancer. Cancer Res 53:2199–2203.

Bladder Cancer Schabath MB, Spitz MR, Grossman HB, et al. 2003. Genetic instability in bladder cancer assessed by the comet assay. J Natl Cancer Inst 95:540–547. Schairer C, Hartge P, Hoover RN, et al. 1988. Racial differences in bladder cancer risk: a case-control study. Am J Epidemiol 128:1027–1037. Schifflers E, Jamart J, Renard V. 1987. Tobacco and occupation as risk factors in bladder cancer: a case-control study in southern Belgium. Int J Cancer 39:287–292. Schoenberg JB, Stemhagen A, Mogielnicki AP, et al. 1984. Case-control study of bladder cancer in New Jersey. I. Occupational exposures in white males. J Natl Cancer Inst 72:973–981. Schroeder JC, Conway K, Li Y, et al. 2003. p53 mutations in bladder cancer: evidence for exogenous versus endogenous risk factors. Cancer Res 63:7530–7538. Schulte PA, Ringen K, Hemstreet GP, et al. 1986. Risk factors for bladder cancer in a cohort exposed to aromatic amines. Cancer 58:2156– 2162. Schulte PA, Ringen K, Hemstreet GP, et al. 1987. Occupational cancer of the urinary tract. Occup Med 2:85–107. Seidegard J, Vorachek WR, Pero RW, et al. 1988. Hereditary differences in the expression of the human glutathione transferase active on transstilbene oxide are due to a gene deletion. Proc Natl Acad Sci 85: 7293–7297. Semenza JC, Weasel LH. 1997. Molecular epidemiology in environmental health: the potential of tumor suppressor gene p53 as a biomarker. Environ Health Perspect 105 Suppl 1:155–163. Shapiro JA, Jacobs EJ, Thun MJ. 2000. Cigar smoking in men and risk of death from tobacco-related cancers. J Natl Cancer Inst 92:333–337. Shibata A, Ohneseit PF, Tsai YC, et al. 1994. Mutational spectrum in the p53 gene in bladder tumors from the endemic area of black foot disease in Taiwan. Carcinogenesis 15:1085–1087. Shin KY, Kong G, Kim WS, et al. 1997. Overexpression of cyclin D1 correlates with early recurrence in superficial bladder cancers. Br J Cancer 1997 75:1788–1792. Sibley K, Cuthbert-Heavens D, Knowles MA. 2001. Loss of heterozygosity at 4p16.3 and mutation of FGFR3 in transitional cell carcinoma. Oncogene 20:686–691. Sidransky D, Frost P, Von Eschenbach A, et al. 1992. Clonal origin bladder cancer. N Engl J Med 326:737–740. Siemiatycki J, Dewar R, Nadon L, et al. 1994. Occupational risk factors for bladder cancer: results from a case-control study in Montreal, Quebec, Canada. Am J Epidemiol 140:1061–1080. Silverman DT, Hoover RN, Albert S, et al. 1983. Occupation and cancer of the lower urinary tract in Detroit. J Natl Cancer Inst 70:237–245. Silverman DT, Hoover RN, Mason TJ, et al. 1986. Motor exhaust-related occupations and bladder cancer. Cancer Res 46:2113–2116. Silverman DT, Levin LI, Hoover RN. 1989a. Occupational risks of bladder cancer in the United States: II Nonwhite men. J Natl Cancer Inst 81:1480–1483. Silverman DT, Levin LI, Hoover RN. 1989b. Occupational risks of bladder cancer in the United States: I. White men. J Natl Cancer Inst 81:1472–1480. Silverman DT, Levin LI, Hoover RN. 1990. Occupational risks of bladder cancer among white women in the United States. Am J Epidemiol 132:453–461. Silverman DT, Morrison AS, Devesa SS. 1996. Bladder Cancer. In: Cancer Epidemiology and Prevention, Schottenfeld D, Fraumeni J (eds). New York: Oxford University Press, pp. 1142–1179. Simon D, Yen S, Cole P. 1975. Coffee drinking and cancer of the lower urinary tract. J Natl Cancer Inst 54:587–591. Simpson J, Roman E, Law G, et al. 1999. Women’s occupation and cancer: preliminary analysis of cancer registrations in England and Wales, 1971–1990. Am J Ind Med 36:172–185. Skipper PL, Tannenbaum SR, Ross RK, et al. 2003. Nonsmoking-related arylamine exposure and bladder cancer risk. Cancer Epidemiol Biomarkers Prev 12:503–507. Skov T, Lynge E. 1994. Cancer risk and exposures to carcinogens in hairdressers. Skin Pharmacol 7:94–100. Slattery ML, Schumacher MC, West DW, et al. 1988a. Smoking and bladder cancer. The modifying effect of cigarettes on other factors. Cancer 61:402–408. Slattery ML, West DW, Robison LM. 1988b. Fluid intake and bladder cancer in Utah. Int J Cancer 42:17–22. Smith AH, Goycolea M, Haque R, et al. 1998. Marked increase in baldder and lung cancer mortality in a region of Northern Chile due to arsenic in drinking water. Am J Epidemiol 147:660–669. Sole G, Sorahan T. 1985. Coarse fishing and risk of urothelial cancer. Lancet 1:1477–1479.

1125

Soll-Johanning H, Bach E, Olsen JH, et al. 1998. Cancer incidence in urban bus drivers and tramway employees: a retrospective cohort study. Occup Environ Med 55:594–598. Sorahan T, Lancashire RJ, Sole G. 1994. Urothelial cancer and cigarette smoking: findings from a regional case-controlled study. Br J Urol 74:753–756. Sorahan T, Sole G. 1990. Coarse fishing and urothelial cancer: a regional casecontrol study. Br J Cancer 62:138–141. Sorlie T, Martel-Planche G, Hainaut P, et al. 1998. Analysis of p53, p16MTS, p21WAF1 and H-ras in archived bladder tumours from workers exposed to aromatic amines. Br J Cancer 77:1573–1579. Spinelli JJ, Band PR, Svirchev LM, et al. 1991. Mortality and cancer incidence in aluminum reduction plant workers. J Ocuup Med 33:1150–1155. Spruck CH, Rideout WM, Olumi AF, et al. 1993. Distinct pattern of p53 mutations in bladder cancer: relationship to tobacco. Cancer Res 53:1162– 1166. Stasik MJ. 1988. Carcinomas of the urinary bladder in a 4-chloro-o-toluidine cohort. Int Arch Occup Environ Health 60:21–24. Steenland K, Palu S. 1999. Cohort mortality study of 57,000 painters and other union members: a 15 year update. Occup Environ Med 56:315– 321. Steenland K, Burnett C, Osorio AM. 1987. A case-control study of bladder cancer using city directories as a source of occupational data. Am J Epidemiol 126:247–257. Steineck G, Hagman U, Gerhardsson M, et al. 1990. Vitamin A supplements, fried foods, fat and urothelial cancer. A case-referent study in Stockholm in 1985–87. Int J Cancer 45:1006–1011. Steineck G, Norell SE, Feychting M. 1988. Diet, tobacco and urothelial cancer. A 14-year follow-up of 16,477 subjects. Acta Oncol 27:323–327. Steineck G, Plato N, Alfredsson L, et al. 1989. Industry-related urothelial carcinogens: application of a job-exposure matrix to census data. Am J Ind Med 16:209–224. Steinmaus CM, Nunez S, Smith AH. 2000. Diet and baldder cancer: a metaanalysis of six dietary variables. AM J Epidemiol 151:693–702. Steinmaus C, Yuan Y, Bates MN, et al. 2003. Case-control study of bladder cancer and drinking water arsenic in the western United States. Am J Epidemiol 158:1193–1201. Stensvold I, Jacobsen BK. 1994. Coffee and cancer: a prospective study of 43,000 Norwegian men and women. Cancer Causes Control 5:401– 408. Stern FB, Beaumont JJ, Halperin WE, et al. 1987. Mortality of chrome leather tannery workers and chemical exposures in tanneries. Scand J Work Environ Health 13:108–117. Stern MC, Johnson LR, Bell DA, et al. 2002a. XPD codon 751 polymorphism, metabolism genes, smoking, and bladder cancer risk. Cancer Epidemiol Biomarkers Prev 11:1004–1011. Stern MC, Umbach DM, Lunn RM, et al. 2002b. DNA repair gene XRCC3 codon 241 polymorphism, its interaction with smoking and XRCC1 polymorphisms, and bladder cancer risk. Cancer Epidemiol Biomarkers Prev 11:939–943. Stern MC, Umbach DM, van Gils CH, et al. 2001. DNA repair gene XRCC1 polymorphisms, smoking, and bladder cancer risk. Cancer Epidemiol Biomarkers Prev 10:125–131. Surveillance Research Program, National Cancer Institute. SEER*Stat software version 5.0.19. 2003. Swanson GM, Burns PB. 1995. Cancer incidence among women in the workplace: a study of the association between occupation and industry and 11 cancer sites. J Occup Environ Med 37:282–287. ’t Mannetje A, Kogevinas M, Chang-Claude J, et al. 1999. Occupation and bladder cancer in European women. Cancer Causes Control 10:209– 217. Takayama S, Sieber SM, Adamson RH, et al. 1998. Long-term feeding of sodium saccharin to nonhuman primates: implications for urinary tract cancer. J Natl Cancer Inst 90:19–25. Talaska G, al Juburi AZ, Kadlubar FF. 1991a. Smoking related carcinogenDNA adducts in biopsy samples of human urinary bladder: identification of N-deoxyguanosin-8-yl.-4-aminobiphenyl as a major adduct. Proc Natl Acad Sci U S A 88:5350–5354. Talaska G, Schamer M, Casetta G, et al. 1994. Carcinogen-DNA adducts in bladder biopsies and urothelial cells: a risk assessment exercise. Cancer Lett 84:93–97. Talaska G, Schamer M, Skipper P, et al. 1991b. Detection of carcinogen-DNA adducts in exfoliated urothelial cells of cigarette smokers: association with smoking, hemoglobin adducts, and urinary mutagenicity. Cancer Epidemiol Biomarkers Prev 1:61–66. Tawfik HN. 1987. Carcinoma of the urinary bladder associated with schistosomiasis in Egypt: the possible causal relationship. Princess Takamatsu Symp 18:197–209.

1126

PART IV: CANCER BY TISSUE OF ORIGIN

Taylor JA, Li Y, He M, et al. 1996. p53 mutations in bladder tumors from arylamine-exposed workers. Cancer Res 56:294–298. Taylor JA, Umbach DM, Stephens E, et al. 1998. The role of N-acetylation polymorphisms in smoking-associated bladder cancer: evidence of a gene-gene-exposure three-way interaction. Cancer Res 58:3603–3610. Terstegge CW, Swaen GM, Slangen JJ, et al. 1995. Mortality Patterns among Commercial Painters in The Netherlands. Int J Occup Environ Health 1:303–310. Teschke K, Morgan MS, Checkoway H, et al. 1997. Surveillance of nasal and bladder cancer to locate sources of exposure to occupational carcinogens. Occup Environ Med 54:443–451. Theriault G, De Guire L, Cordier S. 1981. Reducing aluminum: an occupation possibly associated with bladder cancer. Can Med Assoc J 124:419–422, 425. Theriault G, Tremblay C, Cordier S, et al. 1984. Bladder cancer in the aluminium industry. Lancet 1:947–950. Thiede T, Christensen BC. 1969. Bladder tumours induced by chlornaphazine. A five-year follow-up study of chlornaphazine-treated patients with polycythaemia. Acta Med Scand 185:133–137. Thomas DB, Uhl CN, Hartge P. 1983. Bladder cancer and alcoholic beverage consumption. Am J Epidemiol 118:720–727. Thompson DE, Mabuchi K, Ron E, et al. 1994. Cancer incidence in atomic bomb survivors. Part II: Solid tumors, 1958–1987. Radiat Res 137: S17–S67. Tolbert PE. 1997. Oils and cancer. Cancer Causes Control 8:386–405. Toruner GA, Akyerli C, Ucar A, et al. 2001. Polymorphisms of glutathione Stransferase genes GSTM1, GSTP1 and GSTT1 and bladder cancer susceptibility in the Turkish population. Arch Toxicol 75:459–464. Travis LB, Curtis RE, Glimelius B, et al. 1995. Bladder and kidney cancer following cyclophosphamide therapy for non-Hodgkin’s lymphoma. J Natl Cancer Inst 87:524–530. Tremblay C, Armstrong B, Theriault G, et al. 1995. Estimation of risk of developing bladder cancer among workers exposed to coal tar pitch volatiles in the primary aluminum industry. Am J Ind Med 27:335–348. Tripathi A, Folsom AR, Anderson KE. 2002. Risk factors for urinary bladder carcinoma in postmenopausal women. The Iowa Women’s Health Study. Cancer 95:2316–2323. Tsuda T, Babazono A, Yamamoto E, et al. 1995. Ingested arsenic and internal cancer: a historical cohort study followed for 33 years. Am J Epidemiol 141:198–209. Turesky RJ, Freeman JP, Holland RD, et al. 2003. Identification of aminobiphenyl derivatives in commercial hair dyes. Chem Res Toxicol 16: 1162–1173. Tyler HA, Notley RG, Schweitzer FA, et al. 1986. Vitamin A status and bladder cancer. Eur J Surg Oncol 12:35–41. Underwood M, Bartlett J, Reeves J, et al. 1995. C-erbB-2 gene amplification: a molecular marker in recurrent bladder tumors? Cancer Res 55: 2422–2430. UNSCEAR. 2000. Sources and Effects of Ionizing Radiation: United Nations Scientific Committee on the Effects of Atomic Radiation 2000 report to the General Assembly, with Scientific Annexes, Vol. II: Effects. New York: United Nations. US Congress, Office of Technology Assessment. 1977. Cancer Testing Technology and Saccharin. Washington, DC: US Govt Print Office. Van Gils CH, Conway K, Li Y, et al. 2002. HRAS1 variable number of tandem repeats polymorphism and risk of bladder cancer. Int J Cancer 100:414–418. Vatsis KP, Martell KJ, Weber WW. 1991. Diverse point mutations in the human gene for polymorphic N-acetyltransferase. Proc Natl Acad Sci U S A 88:6333-6337. Veltman JA, Fridlyand J, Pejavar S, et al. 2003. Array-based comparative genomic hybridization for genome-wide screening of DNA copy number in bladder tumors. Cancer Res 63:2872–2880. Vena JE, Freudenheim J, Graham S, et al. 1993a. Coffee, cigarette smoking, and bladder cancer in western New York. Ann Epidemiol 3:586–591. Vena JE, Graham S, Freudenheim J, et al. 1992. Diet in the epidemiology of bladder cancer in western New York. Nutr Cancer 18:255–264. Vena JE, Graham S, Freudenheim J, et al. 1993b. Drinking water, fluid intake, and bladder cancer in western New York. Arch Environ Health 48:191–198. Villanueva CM, Cantor KP, Cordier S, et al. 2004. Pooled analysis of casecontrol studies of bladder cancer and disinfection by-products. Epidemiology 15:357–367. Villanueva CM, Fernandez F, Malats N, et al. 2003. Meta-analysis of studies on individual consumption of chlorinated drinking water and bladder cancer. J Epidemiol Community Health 57:166–173. Vineis P, Caporaso N. 1995. Tobacco and cancer: epidemiology and the laboratory. Environ Health Perspect 103:156–160.

Vineis P, Caporaso N, Tannenbaum SR, et al. 1990. Acetylation phenotype, carcinogen-hemoglobin adducts, and cigarette smoking. Cancer Res 50:3002–3004. Vineis P, Di Prima S. 1983. Cutting oils and bladder cancer. Scand J Work Environ Health 9:449–450. Vineis P, Esteve J, Hartge P, et al. 1988. Effects of timing and type of tobacco in cigarette-induced bladder cancer. Cancer Res 48:3849–3852. Vineis P, Esteve J, Terracini B. 1984. Bladder cancer and smoking in males: types of cigarettes, age at start, effect of stopping and interaction with occupation. Int J Cancer 34:165–170. Vineis P, Kogevinas M, Simonato L, et al. 2000. Leveling-off of the risk of lung and bladder cancer in heavy smokers: an analysis based on multicentric case-control studies and a metabolic interpretation. Mutat Res 463:103–110. Vineis P, Magnani C. 1985. Occupation and bladder cancer in males: a casecontrol study. Int J Cancer 35:599–606. Vineis P, Marinelli D, Autrup H, et al. 2001. Current smoking, occupation, N-acetyltransferase-2 and bladder cancer: a pooled analysis of genotype-based studies. Cancer Epidemiol Biomarkers Prev 10: 1249–1252. Vineis P, Martone T. 1996. Molecular epidemiology of bladder cancer. Ann Ist Super Sanita 32:21–27. Vineis P, Pirastu R. 1997. Aromatic amines and cancer. Cancer Causes Control 8:346–355. Vizcaino AP, Parkin DM, Boffetta P, et al. 1994. Bladder cancer: epidemiology and risk factors in Bulawayo, Zimbabwe. Cancer Causes Control 5:517–522. Vriesema JL, Aben KK, Witjes JA, et al. 2001. Superficial and metachronous invasive bladder carcinomas are clonally related. Int J Cancer 93: 699–702. Wacholder S, Chanock S, Garcia-Closas M, et al. 2004. Assessing the probability that a positive report is false: an approach for molecular epidemiology studies. J Natl Cancer Inst 96:434–442. Wagoner JK. 1970. Leukemia and Other Malignancies Following Radiation Therapy for Gynecological Disorders. Harvard School of Public Health, Boston. Wakai K, Takashi M, Okamura K, et al. 2000. Foods and nutrients in relation to bladder cancer risk: a case-control study in Aichi Prefecture, Central Japan. Nutr Cancer 38:13–22. Walker AM, Dreyer NA, Friedlander E, et al. 1982. An independent analysis of the National Cancer Institute study on non-nutritive sweeteners and bladder cancer. Am J Public Health 72:376–381. Wallin H, Skipper PL, Tannenbaum SR, et al. 1995. Altered aromatic amine metabolism in epileptic patients treated with phenobarbital. Cancer Epidemiol Biomarkers Prev 4:771–773. Wang L, Habuchi T, Takahashi T, et al. 2002. Cyclin D1 gene polymorphism is associated with an increased risk of urinary bladder cancer. Carcinogenesis 23:257–264. Ward E, Carpenter A, Markowitz S, et al. 1991. Excess number of bladder cancers in workers exposed to ortho-toluidine and aniline. J Natl Cancer Inst 83:501–506. Ward E, Halperin W, Thun M, et al. 1988. Bladder tumors in two young males occupationally exposed to MBOCA. Am J Ind Med 14:267–272. Warner ML, Moore LE, Smith MT, et al. 1994. Increased micronuclei in exfoliated bladder cells of individuals who chronically ingest arseniccontraminated water in Nevada. Cancer Epidemiol Biomarkers Prev 3:583–590. Weber WW. 1987. The acetylator gene and drug response. New York: Oxford University Press Inc. Weber WW, Hein DW, Litwin A, et al. 1983. Relationship of acetylator status to isoniazid toxicity, lupus erythematosus, and bladder cancer. Fed Proc 42:3086–3097. Weiss NS. 1995. Cancer in relation to occupational exposure to perchloroethylene. Cancer Causes Control 6:257–266. Wigle DT. 1977. Bladder cancer: Possible new high-risk occupation. Lancet 2:83–84. Wilkins JR, Comstock GW. 1981. Source of drinking water at home and site-specific cancer incidence in Washington County, Maryland. Am J Epidemiol 114:178–190. Williams RR, Horm JW. 1977. Association of cancer sites with tobacco and alcohol consumption and socioeconomic status of patients: interview study from the Third National Cancer Survey. J Natl Cancer Inst 58:525–547. Woolcott CG, King WD, Marrett LD. 2002. Coffee and tea consumption and cancers of the bladder, colon and rectum. Eur J Cancer Prev 11:137–145. World Cancer Research Fund in association with American Institute for Cancer Research. 1997. Food, Nutrition and the Prevention of Cancer: a global perspective. pp. 338–361.

Bladder Cancer Wright ME, Michaud DS, Pietinen P, et al. 2005. Estimate urine pH and bladder cancer risk in a cohort of male smokers (Finland). Cancer Causes Control 16:1117–1123. Wu MM, Kuo TL, Hwang YH, et al. 1989. Dose-response relation between arsenic concentration in well water and mortality from cancers and vascular diseases. Am J Epidemiol 130:1123–1132. Wu X, Amos CI, Zhu Y, et al. 2003. Telomere dysfunction: a potential cancer predisposition factor. J Natl Cancer Inst 95:1211–1218. Wulfkuhle JD, Liotta LA, Petricoin EF. 2003. Proteomic applications for the early detection of cancer. Nature 3:267–275. Wynder EL, Onderdonk J, Mantel N. 1963. An epidemiological investigation of cancer of the bladder. Cancer 16:1388–1407. Wynder EL, Augustine A, Kabat GC, et al. 1988. Effect of the type of cigarette smoked on bladder cancer risk. Cancer 61:622–627. Wynder EL, Goldsmith R. 1977a. The epidemiology of bladder cancer: a second look. Cancer 40:1246–1268. Wynder EL, Stellman SD. 1977b. Comparative epidemiology of tobaccorelated cancers. Cancer Res 37:4608–4622. Yamasaki E, Ames BN. 1977. Concentration of mutagens from urine by absorption with the nonpolar resin XAD-2: cigarette smokers have mutagenic urine. Proc Natl Acad Sci U S A 74:3555–3559. Yamamoto S, Romanenko A, Wei M, et al. 1999. Specific p53 gene mutations in urinary bladder epithelium after the Chernobyl accident. Cancer Res 59:3606–3609. Yasunaga Y, Nakanishi H, Naka N, et al. 1997. Alterations of the p53 gene in occupational bladder cancer in workers exposed to aromatic amines. Lab Invest 77:677–684. Yoshida O, Harada T, Miyagawa M, et al. 1971. Bladder cancer in workers in the dyeing industry. Igaku No Ayumi Japan. 79:421. Young JF, Kadlubar FF. 1982. A pharmacokinetic model to predict exposure of the bladder epithelium to urinary N-hydroxyarylamine carcinogens as a function of urine pH, voiding interval, and resorption. Drug Metab Dispos 10:641–644. Yu MC, Ross RK, Chan KK, et al. 1995. Glutathione S-transferase M1 genotype affects aminobiphenyl-hemoglobin adduct levels in white, black and Asian smokers and nonsmokers. Cancer Epidemiol Biomarkers Prev 4:861–864. Yu MC, Skipper PL, Taghizadeh K, et al. 1994. Acetylator phenotype, aminobiphenyl-hemoglobin adduct levels, and bladder cancer risk in white,

1127

black, and Asian men in Los Angeles, California. J Natl Cancer Inst 86:712–716. Zeegers M, Dorant E, Goldbohm R, et al. 2001d. Are coffee, tea, and total fluid consumption associated with bladder cancer risk? Results from the Netherlands Cohort Study. Cancer Causes Control 12:231–238. Zeegers MP, Goldbohm RA, Bode P, et al. 2002a. Prediagnostic toenail selenium and risk of bladder cancer. Cancer Epidemiol Biomarkers Prev 11:1292–1297. Zeegers MP, Goldbohm RA, van den Brandt PA. 2001a. Are retinol, vitamin C, vitamin E, folate and carotenoids intake associated with bladder cancer risk? Results from the Netherlands Cohort Study. Br J Cancer 85: 977–983. Zeegers MP, Goldbohm RA, van den Brandt PA. 2002b. A prospective study on active and environmental tobacco smoking and bladder cancer risk The Netherlands. Cancer Causes Control 13:83–90. Zeegers MP, Swaen GM, Kant I, et al. 2001b. Occupational risk factors for male bladder cancer: results from a population based case cohort study in the Netherlands. Occup Environ Med 58:590–596. Zeegers MP, Volovics A, Dorant E, et al. 2001c. Alcohol consumption and bladder cancer risk: results from The Netherlands Cohort Study. Am J Epidemiol 153:38–41. Zenser TV, Lakshmi VM, Rustan TD, et al. 1996. Human N-acetylation of benzidine: role of NAT1 and NAT2. Cancer Res 56:3941–3947. Zhang ZF, Cordon-Cardo C, Rothman N, et al. 1997a. Methodological issues in the use of tumour markers in cancer epidemiology. IARC Sci Publ: 201–213. Zhang ZF, Sarkis AS, Cordon-Cardo C, et al. 1994. Tobacco smoking, occupation, and p53 nuclear overexpression in early stage bladder cancer. Cancer Epidemiol Biomarkers Prev 3:19–24. Zhang ZF, Shu XM, Cordon-Cardo C, et al. 1997b. Cigarette smoking and chromosome 9 alterations in bladder cancer. Cancer Epidemiol Biomarkers Prev 6:321–326. Zheng T, Cantor KP, Zhang Y, et al. 2002. Occupation and bladder cancer: a population-based, case-control study in Iowa. J Occup Environ Med 44:685–691. Zhong S, Wyllie AH, Barnes D, et al. 1993. Relationship between the GSTM1 genetic polymorphism and susceptibility to bladder, breast and colon cancer. Carcinogenesis 14:1821–1824.

59

Prostate Cancer ELIZABETH A. PLATZ AND EDWARD GIOVANNUCCI

P

rostate cancer is the most commonly diagnosed cancer in American men and its incidence is rising internationally. Although intensively studied, little is known conclusively about its causes, with the exception of older age, African-American racial group, and family history of prostate cancer. Sex steroid hormones and possibly other hormonal systems, such as the insulin-like growth factor axis and vitamin D, are contributory, but links across the range of normal levels have been difficult to reliably demonstrate. Promising dietary leads that increase risk include higher intake of red and processed meat and dairy products; dietary leads that decrease risk include higher intake of tomato products, which contain the carotenoid lycopene, and selenium, and supplemental intake of vitamin E. Evidence for a benefit of selenium and vitamin E on prostate cancer incidence was considered to be sufficiently strong to warrant testing in a large chemoprevention trial. Whether obesity, physical inactivity, occupation and environmental contaminants, cigarette smoking, vasectomy, sexually transmitted infections, or prostatitis are risk factors for prostate cancer remain controversial. The observation of consistent loss of glutathione S-transferase-pi expression in prostate cancer due to epigenetic changes and the contemporary characterization of proliferative inflammatory atrophy lesions has led to new avenues of epidemiologic research on the role of oxidation and inflammation in the development of this disease. Explanations for the 60% higher prostate cancer incidence in African-American men and the 38% lower incidence in AsianAmerican men compared with white men (Jemal et al., 2003) are still elusive. Variability in the prevalence of both inherited and modifiable risk factors among racial groups has been considered, in particular in allele frequencies of genes. Multi-racial/ethnic cohort and large population-based case-control studies are ongoing and may be better able to provide clues to both modifiable and inherent risk factors. Studies of familial prostate cancer continue to identify loci with possible linkage; eight loci have been suggested to date. However, these loci generally have not been replicable from population to population. The advent of high throughput sequencing and genotyping has given rise to studies of common sequence variants in the genes mapping to the loci identified in linkage analysis. Also, abounding because of high throughput technology are studies of polymorphisms in genes encoding proteins involved in hormonal systems, metabolism, DNA repair, and cell cycle control. Unfortunately, results from these studies also have not been consistent, perhaps because of differences in populations under study, in study design and associated sources of bias, and chance variability because of small size relative to the prevalence of alleles. Use of now widely available statistical methods for inferring haplotypes outside of family studies may add clarity to the work on the role of genes in sporadic prostate cancer. The availability of a screening test for prostate cancer via detection of elevated circulating prostate-specific antigen (PSA) coupled with broad recommendations for age at which to start screening has changed the characteristics of prostate cancer cases at diagnosis. The median age at diagnosis is now slightly younger and the distribution of stage at diagnosis has shifted to most cases being diagnosed while the tumor or tumors are confined within the prostate. The variability in the case mix between studies conducted in the pre-PSA and PSA eras and between current studies in countries with and without routine application of PSA screening has implications for drawing inferences from epidemiologic studies individually and collectively. It remains to

1128

be proven whether earlier detection coupled with earlier treatment results in improved survival with prostate cancer. Because of the unknown effectiveness, the appropriate application of PSA screening is debated. In the meantime, refinements in the PSA test and optimization of screening frequency are ongoing. Gene expression studies comparing normal prostate tissue and prostate adenocarcinoma may uncover possible new markers of early detection for prostate cancer. In this chapter, we review the classification of prostate cancer, demographic patterns of prostate cancer, environmental and host risk factors for prostate cancer, the pathogenesis of prostate cancer, the primary and secondary prevention strategies for prostate cancer, and direction for future epidemiologic investigations on the causes and prevention of prostate cancer.

CLASSIFICATION The prostate, a walnut-sized gland that is part of the genitourinary system in men, is located below the bladder and in front of the rectum. Although the function of the prostate has not been fully defined, the epithelial cells lining the prostatic glandular acini secrete fluids that become a component of seminal fluid. These luminal epithelial cells also secrete PSA, a protease that cleaves seminal proteins likely to maintain the fluidity of seminal fluid. No other cells in the body normally produce PSA.

Anatomic Distribution The human prostate consists of three zones—the central, transition, and peripheral zones. These zones differ from other species, such as the rat, which has the ventral, dorsal, lateral and enterior. Prostate tumors tend to occur in the peripheral zone, but may also be found in the transition zone. Unlike most other solid tumors, prostate cancer tends to be multifocal (Wise et al., 2002). Whether these foci develop from separate initiation or promotion events or whether they represent the same original tumor cells that have traveled to other sites within the prostate via the ductal system is unknown. Prostate cancer staging is classified using the TNM system. T1a (tumor is <5% of resected tissue) and T1b (tumor is 5% or more of resected tissue) cases are those detected in prostate tissue removed at the time of surgery for benign prostatic hyperplasia. T1c cases are those detected in biopsy specimens in men with an elevated PSA, but normal digital rectal examination. T2 cases have palpable tumors, but are clinically organ confined. T3 cases have extraprostatic extension with or without invasion of the seminal vesicles. T4 cases have invasion of adjacent structures other than the seminal vesicles, including the bladder and rectum. Distant metastasis occurs most commonly to bone.

Histopathology The normal prostate acinus consists of a layer of luminal columnar epithelial cells and a layer of basal cells. Surrounding the acinus is stromal tissue, including smooth muscle. The acini are connected forming the ductal system that empties into the urethra. The cells at risk for neoplastic transformation are not known, but are hypothe-

Prostate Cancer sized to have a phenotype that is intermediate between a stem cell and an epithelial cell (van Leenders et al., 2003). Foci of prostate cancer typically consist of a lining of tumor cells surrounding a lumen. Often these acini are small and have lost the characteristic papillary infoldings of an acinus, and the component tumor cells have large nuclei. To describe the extent of architectural disarray, histologic grading is classified using the Gleason scoring system, in which the two predominant patterns of glandular distortion are identified in the tumor and are scored from 1 (least disarrayed) to 5 (most disarrayed) (Gleason et al., 1974). The two scores are summed forming a score that ranges from 2–10. Typically a Gleason score of seven or greater is considered to be histologically poor disease.

Precursor Neoplastic Lesions Prostate cancer is thought to progress through one or more precursor lesions, including prostatic intraepithelial neoplasia (PIN) (Bostwick, 1999). PIN consists of relatively normal-appearing acini lined with malignant cells (McNeal and Bostwick, 1986). By the third (20%) and fourth (44%) decades of life a significant proportion of men already have PIN (Sakr et al., 1993). The presence of high-grade PIN (HGPIN) in men undergoing biopsy for elevated PSA may predict the presence of adenocarcinoma in a follow-up biopsy, although this remains controversial (De Marzo et al., 2003). Another lesion that has received attention recently for its possible relation to prostate cancer is focal atrophy. Focal atrophy, including the forms of post-atrophic hyperplasia and simple atrophy, is common in the prostates of older men. Although atrophic, these areas show greater proliferation compared with normal epithelium (Ruska et al., 1998) and are usually associated with acute or chronic inflammation in the glandular lumen, within the epithelium, or peri-glandularly. Although focal atrophy has long been recognized, this group of lesions recently has been labeled proliferative inflammatory atrophy (PIA) to describe their nature (De Marzo et al., 1999). PIA lesions may be regenerative lesions following prostate tissue damage possibly by infection, oxidant damage, hypoxia, or autoimmunity (De Marzo et al., 1999). Because PIA lesions are frequently confluent with or near HGPIN or adenocarcinoma (Putzi and De Marzo 2000), De Marzo et al. (1999) have hypothesized that PIA lesions may be very early prostate cancer precursor lesions, or alternatively, may reflect an environment within the prostate that is conducive to prostate cancer development. Research to further characterize and understand the relationship among HGPIN, PIA lesions, and prostate adenocarcinoma is ongoing.

Molecular Genetic Characteristics of Tumor Autopsy studies have demonstrated small foci of prostate cancer in roughly 30% of American men in their thirties and forties (Sakr et al., 1993). This observation suggests that cells that ultimately give rise to a focus of prostate cancer are likely initiated early in life and promotional and progression events occur throughout life. Genetic and epigenetic alterations may determine the wide variation in prostate tumor growth rates and ability to metastasize. The molecular changes that arise along the natural history of prostate cancer have been previously reviewed (De Marzo et al., 2003). Somatic chromosomal changes that are common in prostate tumors and especially in hormone refractory disease are gains in 7p and q, 8q, and Xq and losses in 8p, 10q, 13q, and 16q (Nupponen and Visakorpi, 1999). Possible influential tumor suppressor genes are located at these loci of losses and oncogenes at these loci of gains (Nupponen and Visakorpi, 1999; Elo and Visakorpi, 2001). The likelihood of these gains and losses through rearrangements may be enhanced through telomere shortening in intermediate phenotype cells that have lost cell cycle control (Meeker et al., 2002). One of the most commonly observed epigenetic changes in prostate adenocarcinoma is loss of expression of the gene (GSTP1) that encodes glutathione S-transferase (GST)-pi (Lee et al., 1994; Cookson et al., 1997; Moskaluk et al., 1997) due to biallelic hypermethylation

1129

of CpG sites in the GSTP1 promoter region (Lin et al., 2001). Loss of GSTP1 expression is also seen in most high-grade PIN lesions (Cookson et al., 1997; Moskaluk et al., 1997; Brooks et al., 1998). Interestingly, epithelial cells in PIA overexpress GSTP1 (De Marzo et al., 1999) and GSTA (Parsons et al., 2001), suggesting that PIA lesions are under oxidative and electrophilic stress. Loss of expression of these enzymes in cells already under such stress may enhance the likelihood of neoplastic transformation (De Marzo et al., 2003).

DEMOGRAPHIC PATTERNS Mortality and Incidence in the United States Prostate cancer is the third most common cause of cancer death in the United States, 28,900 men (mortality rate 33.9 per 100,000 men) are expected to die of their disease in 2003, accounting for 10% of all cancer deaths in men. Prostate cancer is the most commonly diagnosed cancer in American men accounting for an estimated 220,900 new cases (incidence rate 168.9 per 100,000 men) in 2003 or 33% of all cancer diagnoses in men (American Cancer Society, 2003b). The second and third most commonly diagnosed cancers in men are lung and bronchus (14%) and the colon and rectum (11%). The lifetime risk of being diagnosed with invasive prostate cancer is 1 in 6 (American Cancer Society, 2003b). Prostate cancer incidence and mortality rates vary dramatically across the United States and these rates do not necessarily co-vary. Incidence rates are highest in the District of Columbia, New Jersey, Maryland, and Michigan and the lowest in Alabama, Nevada, Tennessee, and Indiana (American Cancer Society, 2003b). The highest prostate cancer mortality rates are found in the District of Columbia, Mississippi. South Carolina, Alabama, and Georgia and the lowest in Alaska, Hawaii, Nebraska, and California (American Cancer Society, 2003b).

Time Trends From the early 1970s to the late 1980s, which is before the introduction of PSA screening, prostate cancer incidence rates were on the rise (Jemal et al., 2003). With the introduction of widespread screening for prostate specific antigen (PSA) in the early to mid 1990s, the prostate cancer incidence rate soared (Potosky et al., 1995). Now that PSA screening has been widely instituted and the pool of prevalent early cases, for which the diagnosis was advanced in time by screening, has been detected, the incidence rate has declined to its increasing trajectory before the availability of PSA screening. After decades of a slightly increasing mortality rate, since 1992 the prostate cancer mortality rate has declined. Two explanations for this decline have been postulated: earlier diagnosis of prostate cancer at a stage when curative treatment is most likely and improvement in treatment for men diagnosed with metastatic disease. Using SEER data and considering mortality rates by stage at diagnosis, most of the decline in the prostate cancer mortality likely occurred because of the shift in stage at diagnosis to early disease, which has a better prognosis (Chu et al., 2003).

Survival Five-year survival rates increased markedly over 1974–1976 to 1983–1985 to 1992–1998 from 68% to 76% to 98% among whites (American Cancer Society, 2003b), likely mostly due to detection and possibly treatment at an earlier stage of the disease. Overall, 5-year survival rates are essentially 100% for local and regional stages together, but only 34% for those diagnosed with distant metastases (American Cancer Society, 2003b).

Age The median age at prostate cancer diagnosis in the PSA-era is 71 years old in whites and 69 years old in blacks in the United States (Standford et al., 1999). The age at diagnosis has dropped slightly with the introduction of PSA screening for prostate cancer. Nevertheless,

1130

PART IV: CANCER BY TISSUE OF ORIGIN

the vast majority of men are diagnosed with prostate cancer at an age older than 65 years.

Race and Ethnicity Racial variation in prostate cancer incidence and mortality rates in the United States is pronounced. African Americans have the highest prostate cancer incidence rate (standardized to 2000 US population age standard, 1992–1999: 275.3 per 100,000 men annually) and mortality rate (75.1 per 100,000 men annually) among any racial or ethnic group in the United States. By comparison, the incidence and mortality rates are 1.6 (172.9 per 100,000 men) and 2.3 (32.9 per 100,000 men) times that for whites, respectively. Prostate cancer incidence and mortality rates for Asian/Pacific Islander, American Indian/Alaskan Native, or Hispanic are substantially lower than those for white Americans (American Cancer Society, 2003b). Although 5-year survival rates have also increased for African Americans from 58% to 64% to 93% over the sample time intervals (American Cancer Society, 2003b), their survival rates remain lower. Among whites, 59% are diagnosed at a localized stage, 18% regional, 10% distant metastases, and the remainder is of unknown stage. A greater proportion of African-American men present with distant metastases (18%). Although reduced access to health care may contribute to the more advanced stage distribution at diagnosis among African Americans, stage-specific survival is slightly poorer compared with whites (5-year survival: 33% vs. 30% for distant stage for whites and African Americans, respectively) (Jemal et al., 2003). In a study conducted among men participating in the US military medical system, a population without economic barriers to screening, diagnosis, and treatment, race did not predict age-adjusted survival, after adjusting for stage and grade, which were higher among the AfricanAmerican men (Optenberg et al., 1995). Risk of diagnosis of advanced stage prostate cancer was still higher in African-American men compared with non-Hispanic white men after adjusting for socioeconomic factors (Hoffman et al., 2001). The notable variation in prostate cancer incidence among black, white, and Asian men in the United States may be due to heterogeneity in inherent and modifiable determinants of prostate epithelial cell growth. However, in a study of male health professionals, risk of prostate cancer remained elevated even after adjusting for purported dietary and lifestyle risk factors (Platz, 2000b). Reduced access to health care may contribute to the later stage at diagnosis in African Americans, but it is unlikely to account for the difference in stagespecific survival.

Socioeconomic Status Internationally, prostate cancer incidence and mortality are positively correlated with the gross national product (Armstrong and Doll, 1975). Developed or “Westernized” countries have notably higher prostate cancer rates than do developing countries (Parkin et al., 2003).

International Patterns Prostate cancer mortality rates around the world vary more than 30fold (American Cancer Society, 2003b). The lowest rates are observed in the Far East and on the Indian subcontinent, whereas the highest rates occur in the Western Europe, Australia, and North America. Adjusting the rates to the World Health Organization world standard population, which is younger than the US 2000 standard population, the mortality rate for prostate cancer was approximately 1 per 100,000 men annually in China compared with 17.9 per 100,000 for American men in the year 2000 (Parkin et al., 2003). The prostate cancer incidence rate in China is 2.9 per 100,000 men contrasted with 107.8 and 185.4 per 100,000 men in white and black Americans, respectively (Parkin et al., 2003). Some of the highest prostate cancer rates are found on Caribbean islands, including Trinidad and Tobago (American Cancer Society, 2003b) and Jamaica (Glover Jr., et al., 1998), where the populations are of African descent. Notably, rates in African countries such as

Nigeria appear to be substantially lower (Ahluwalia et al., 1981), although cancer incidence and mortality information is not sufficiently comprehensive. A gradient in the prostate cancer mortality rate exists between Northern Europe (e.g., Sweden, Norway, and Denmark), where the rates are more than 23 per 100,000 men annually, and Southern Europe (e.g., Greece), where the rates are half that (American Cancer Society, 2003b). Some of the disparity in prostate cancer incidence rates among countries is likely due to differences in medical practice leading to differential rates of detection of subclinical tumors. The frequency of these latent tumors does not appear to vary dramatically among populations (Yatani et al., 1982), including in autopsy series of black populations in West Africa and in the United States. (Jackson et al., 1980). This observation possibly indicates that international variation in the prevalence of exposure to factors that cause the growth and progression of prostate tumors most likely account for the notable variation in prostate cancer incidence.

Migration Migration studies show that men of Asian heritage living in the United States are at lower risk for prostate cancer than white Americans, but at greater risk of the disease than men of similar ancestries living in Asia (Haenszel and Kurihara, 1968; Yu et al., 1991). Japanese immigrants living in Los Angeles County, California, have prostate cancer rates more comparable to individuals with similar ancestry but who were born in the United States than to men in Japan (Shimizu et al., 1991b), and these rate differences do not appear to be entirely due to differences in detection of early-stage tumors between the United States and Japan (Shimizu, 1991a). The changing prostate cancer rates among migrants, despite a common prevalence of early stage lesions, suggest that endogenous or exogenous factors affecting tumor growth and progression, rather than initiation, may vary between countries as well.

DIET AND NUTRITION FACTORS Fruits, Vegetables, Legumes, and Micronutrients Fruits and Vegetables Fruits and vegetables contain a wide array of antioxidant nutrients, other essential vitamins and minerals, and nonessential but bioactive compounds. Ecologic studies show a negative correlation between per capita vegetable consumption (r = -0.38), but not fruit consumption (r = -0.09) and prostate cancer mortality (Rose et al., 1986). However, most cohort studies (Mills et al., 1989; Severson et al., 1989; Hsing et al., 1990b; Le Marchand et al., 1994; Giovannucci et al., 1995; Schuurman et al., 1998; Chan et al., 2000) do not support an inverse association between total intake or fruits or vegetables and prostate cancer. Some (Cohen et al., 2000; Kolonel et al., 2000), but not all (Hayes et al., 1999; Villeneuve et al., 1999) populationbased case-control studies reported inverse associations for fruits or vegetables. Fruits and vegetables and their major nutrient components have been evaluated in relation to prostate cancer. Results for tomato-based foods and Brassica vegetables are described below. For other food groups, findings are not consistent or are not supportive in prospective and large population-based case-control studies, for example, for leafy green vegetables such as spinach (Giovannucci et al, 1995; Key et al., 1997; Schuurman et al., 1998; Cohen et al., 2000; Kolonel et al., 2000) and allium vegetables such as garlic and onions (Key et al., 1997; Schuurman et al., 1998; Hsing et al., 2002). Other botanical families are discussed below along with their major characterizing nutrients.

Tomatoes and Lycopene Tomatoes and food items prepared from tomatoes are the primary source of the carotenoid lycopene in the American diet. Lycopene, an efficient antioxidant (Sies and Stahl, 1995), is the most abundant carotenoid in plasma (Kaplan et al., 1987; Ascherio et al., 1992) and

Prostate Cancer the prostate (mean of 30% of total) and is present at biologically active concentrations (Clinton et al., 1996). Because the bioavailability of lycopene is enhanced by heating with oil (Stahl and Sies, 1992), several groups have considered separately from total tomato intake the intake of cooked tomato foods or tomato sauce. Higher intake of tomato-based foods and lycopene was associated with a lower prostate cancer risk, including both organ-confined disease and metastatic disease, in a large cohort study (Giovannucci et al., 1995), including after long-term follow-up (Giovannucci et al., 2002b). Men who consumed two or more servings of tomato sauce had a 25% lower risk of prostate cancer (p-trend <0.001) (Giovannucci et al., 2002b). Earlier, a cohort study among Seventh-Day Adventists suggested a 40% reduction in prostate cancer risk comparing 5+ servings of tomatoes weekly to less than once per week (p-trend = 0.02) (Mills et al., 1989). No association for tomatoes was observed in the Netherlands cohort, although the mean level of intake was relatively low (19.5 g/day) and cooked tomatoes were not considered (Schuurman et al., 1998). No association between intake of cooked tomato products and prostate cancer was observed in three large population-based case-control studies (Key et al., 1997; Hayes et al., 1999; Kolonel et al., 2000), or in a study of similar design for tomatoes and tomato juice (Villeneuve et al., 1999), although a suggestive inverse association for cooked tomatoes was present in another (Cohen et al., 2000). Several groups have examined circulating concentrations of lycopene in relation to prostate cancer and inverse associations for both total prostate cancer and for advanced disease on the order of a 25%–50% lower risk in the top vs. bottom quantile have been suggested in prospective (Hsing et al., 1990a; Gann et al., 1999; Wu et al., 2004) and case-control (Lu et al., 2001; Vogt et al., 2002) studies. A study nested in the CLUE II cohort observed a nonstatistically significant OR of total prostate cancer for plasma lycopene that was weaker in magnitude than that found in the other cohort studies reporting inverse associations (Huang et al., 2003). In those studies suggesting an inverse association, mean or median plasma lycopene concentrations ranged from 18.7–38.8 mg/dL. One prospective study not showing an association for blood lycopene (Nomura et al., 1997) had the lowest mean lycopene levels (controls 13.4 mg/dL) of all of the studies. Overall, the epidemiologic data indicate that intake of tomatoes and tomato products and higher circulating concentrations of lycopene are probably associated with a lower risk of prostate cancer. Further work is needed to determine whether the observed benefit is due to lycopene itself, other bioactive substances in tomatoes, or to an unaccounted for favorable dietary and lifestyle pattern.

Brassica Vegetables Brassica or cruciferous vegetables, which include broccoli, Brussels sprouts, and cabbage, contain glucosinolates, which induce phase II detoxification enzymes. A small number of studies suggest an inverse association between Brassica vegetables and prostate cancer on the order of a 20%–46% reduction in risk comparing extreme quantiles (Kristal and Lampe, 2002), including large population-based casecontrol studies (Cohen et al., 2000; Kolonel et al., 2000), and possibly a cohort study (Schuurman et al., 1998). Other cohort (Hsing et al., 1990b; Giovannucci et al., 1995) and large population-based casecontrol studies are not confirmatory (Villeneuve et al., 1999). Variability in findings may be due to differences in the amount and the array of specific Brassica vegetables consumed in a population and the nature and extent of exposure to agents that are metabolized by biotransformation enzymes induced by Brassica vegetables.

Vitamin A Vitamin A or retinol, found in both animal (e.g., liver, eggs) and vegetable sources, and in fortified in milk and cold breakfast cereals, is required to maintain normal cellular proliferation and differentiation (Sporn and Roberts, 1984). One prospective study showed an inverse association of intake of vitamin A with prostate cancer in older men, but a positive association in younger men (Hsing et al., 1990b), whereas other prospective studies (Hsing et al., 1990b; Giovannucci et al., 1995) and large case-control studies (Ross et al., 1987; Key et

1131

al., 1997; Hayes et al., 1999) found no consistent association. Some studies found positive associations for retinol intake and prostate cancer risk (Andersson et al., 1996) particularly among older men (Graham et al., 1983; Kolonel et al., 1987; West et al., 1991; Giovannucci et al., 1995; Andersson et al., 1996). Prediagnostic serum vitamin A levels were inversely related to prostate cancer risk in two studies (Hsing et al., 1990a; Reichman et al., 1990), positively related in a prospective study (both in younger and older men) (Gann et al., 1999), and not associated in other cohorts (Eichholzer et al., 1996; Nomura et al., 1997; Huang et al., 2003). Mean pre-diagnostic circulating levels of retinol in these studies narrowly ranged from 56.5– 68.7 mg/dL, and thus differences among populations in retinol concentration is not an explanation for the inconsistent findings among studies. The interpretation of these results is unclear, given that blood retinol levels are tightly regulated (Willett et al., 1984) and thus, not substantially influenced by the normal range of dietary retinol intake observed in epidemiologic studies. Explanations for why risk of prostate cancer associated with higher retinol intake is elevated in older men in several of the studies are elusive.

Orange-Yellow Vegetables, b-Carotene, and Other Carotenoids Orange and deep yellow vegetables such as carrots and sweet potatoes are colored by b-carotene, a carotenoid that can be converted into vitamin A. Generally, orange-yellow vegetables and carrots have not been associated with a decreased risk of prostate cancer in cohort (Giovannucci et al., 1995; Schuurman et al., 1998) or populationbased case-control studies (Key et al., 1997), although two populationbased case-control studies suggest inverse associations for carrots (Cohen et al., 2000; Kolonel et al., 2000). Prospective studies have not shown inverse associations for intake of b-carotene and prostate cancer (Hsing et al., 1990b; Giovannucci et al., 1995; Daviglus et al., 1996). Likewise, two trials with a b-carotene arm (Heinonen et al., 1998; Cook et al., 2000) showed no association with prostate cancer. Serum b-carotene was not associated with prostate cancer in cohort (Hsing et al., 1990a; Nomura et al., 1997; Huang et al., 2003) and casecontrol (Vogt et al., 2002) studies. Numerous other carotenoids are found in the US diet, such as a-carotene in carrots, lutein in dark green vegetables, and bcryptoxanthin in oranges. Intake of and circulating concentrations of carotenoids, such as a-carotene, b-cryptoxanthin, lutein/zeaxanthin, have not been consistently related to prostate cancer (Giovannucci et al., 1995; Nomura et al., 1997; Gann et al., 1999; Cohen et al., 2000; Vogt et al., 2002; Huang et al., 2003). The evidence from epidemiologic studies strongly indicates that bcarotene does not protect against prostate cancer and that other carotenoids (exclusive of lycopene) are unlikely to protect against prostate cancer.

Citrus Fruits and Vitamin C Citrus fruits are the major dietary source of vitamin C (ascorbic acid), a water-soluble vitamin with many functions, including antioxidant activity (Groff and Cropper, 2000). No consistent associations have been observed between intake of citrus fruits including oranges, intake of vitamin C, or circulating concentrations of vitamin C and prostate cancer among prospective and large case-control studies (Chan and Giovannucci, 2001a). A case-control study observed a possible inverse association for frequency of vitamin C supplement use (Kristal et al., 1999). Some studies have suggested a higher risk in men consuming greater citrus fruit or vitamin C (West et al., 1991; Key et al., 1997; Schuurman et al., 1998). Overall, vitamin C does not appear to protect against prostate cancer. Explanations for the positive association in some studies are not clear.

Vitamin E Interest in the prevention of prostate cancer by vitamin E arose from the results from the randomized Alpha-Tocopherol, Beta-Carotene (ATBC) Trial in Finnish smokers, which did not support a benefit of a 50-mg vitamin E supplement (a-tocopherol) on the primary study outcome, lung cancer (The Alpha-Tocopherol Beta Carotene Cancer

1132

PART IV: CANCER BY TISSUE OF ORIGIN

Prevention Study Group, 1994). However, in an analysis of secondary endpoints, statistically significant 32% and 41% reductions in prostate cancer risk and mortality, respectively, were observed among those receiving the vitamin E supplement (Heinonen et al., 1998). Risk decreased early in the trial, suggesting that vitamin E influenced the promotional or progression phases of carcinogenesis (Heinonen et al., 1998). Vitamin E has antioxidant properties and independently may decrease the growth of prostate cancer cells through induction of apoptosis (Sigounas et al., 1997). Major dietary sources are vegetable oils and nuts (Groff and Gropper, 2000). Despite the promising findings from the ATBC Trial, prospective (Chan et al., 1999) and large population-based case-control (Andersson et al., 1996; Key et al., 1997; Kristal et al., 1999) studies generally do not support an inverse association between vitamin E and prostate cancer, except possibly for frequent use of vitamin E supplements (Kristal et al., 1999). Also, baseline a-tocopherol intake and serum levels were not associated with prostate cancer in the arm of the ATBC Trial not randomized to a-tocopherol (Hartman et al., 1998a). No clear association was seen for serum total, a-, and g-tocopherols and prostate cancer in nested case-control studies (Hsing et al., 1990a; Nomura et al., 1997; Gann et al., 1999); although an inverse associations for a-tocopherol (comparing extreme quintiles: OR = 0.64) and a strong inverse association (OR = 0.25) for g-tocopherol were observed in the CLUE II cohort (Helzlsouer et al., 2000; Huang et al., 2003). Concentrations of atocopherol ranged from 9.6–13.1 mg/mL and g-tocopherol ranged from 1.5–2.9 mg/mL in these studies. In addition to the ATBC Trial conducted in smokers, some prospective studies suggest a benefit in cigarette smokers of higher intake of vitamin E (supplement level) (Chan et al., 1999) or higher circulating a-tocopherol concentrations (Eichholzer et al., 1996; Gann et al., 1999), in particular for more advanced disease. Largely because of the intriguing findings for a-tocopherol in the ATBC trial, a-tocopherol in the prevention of prostate cancer is being tested in the SELECT trial (described below).

Selenium As with vitamin E, strong interest in the prevention of prostate cancer by selenium resulted from secondary findings from a randomized trial. In the Nutritional Prevention of Cancer Study, a randomized trial of selenium supplementation (brewer’s yeast that provided 200 mg/day— nearly three times the RDA—for 10 years) in the prevention of skin cancer recurrence among patients who lived in low-soil selenium regions (Clark et al., 1996), no effect on skin cancer recurrence was detected. However, a 65% reduction in prostate cancer incidence among men randomized to a selenium supplement was found (Clark et al., 1996) for both local and advanced stage tumors, even after excluding men who might have had occult prostate cancer at baseline (RR = 0.26, p = 0.009) (Clark et al., 1998). The beneficial effect of selenium on prostate cancer appeared after only a few years of supplementation, suggesting a benefit on promotion and progression (Clark et al., 1996). Selenium is an indirect antioxidant as it is essential for the activity of glutathione peroxidase (Combs and Combs, 1984). Because selenium content of food depends on geographical source and because intestinal absorption and distribution of selenium depend on the presence of metals that compete for uptake (Chow, 1979) selenium content in biological materials may be a good marker of bioavailable selenium. Several prospective studies support inverse associations between selenium levels in toenails (Yoshizawa et al., 1998; Helzlsouer et al., 2000) or in plasma/serum (Nomura et al., 2000; Brooks et al., 2001) and prostate cancer incidence or mortality on the order of 50%–60% reductions when comparing extreme quantiles. Three of the cohort studies (Yoshizawa et al., 1998; Helzlsouer et al., 2000; Brooks et al., 2001) suggested a threshold effect for selenium, consistent with the randomized trial by Clark et al. (1998). A prospective study conducted in Finland, a country with low selenium intake during the period of follow-up, showed no relation between serum selenium levels and prostate cancer risk (Knekt et al., 1990). In that study, participants had circulating levels almost a third

(~50 vs. ~150 mg/L) of that found in the participants in the other studies. Beginning in 1984, fertilizers were fortified with selenium in Finland. Despite elevations in blood selenium, over the next decades prostate cancer incidence rates continued to rise in Finland and prostate cancer mortality rates remained relatively unchanged. It is unclear why studies in the United States indicated a benefit of selenium, whereas data from Finland (Knekt et al., 1990) do not support a lowering of prostate cancer risk with higher selenium intake. Resolving the potential impact of selenium as a preventive agent against prostate carcinogenesis is a top priority and is being evaluated in the SELECT chemoprevention trial (see below).

Zinc Seminal fluid and the prostate epithelium are rich in zinc, which has antimicrobial activity and thus may protect the gonads from infectious agents that might gain access via the urethra. Modest to moderate inverse associations were observed in two case-control studies for dietary zinc (Key et al., 1997) or zinc supplement use (Kristal et al., 1999); other case-control studies do not support a protective association (Kolonel et al., 1988; West et al., 1991; Andersson et al., 1996; Vlajinac et al., 1997). A nested case-control study that determined zinc concentrations in toenail clippings as a time-integrated measure reported a lower risk of prostate cancer in the top three fourths of toenail zinc when compared with the bottom fourth (Platz et al., 2001). Although no association was found for normal-range zinc intake, a large prospective cohort study observed a doubling of risk of advanced from cancer in men using high-dose supplemental zinc or men who were long-term zinc supplement users (Leitzmann et al., 2003). Additional epidemiologic studies of the role of zinc at normal and at high ranges of intake in the etiology of prostate cancer are needed.

Phytoestrogens and Soy Common soy phytoestrogens include the isoflavones genistein and daidzein, which have been shown to inhibit the growth of both androgen-dependent and androgen-independent cell lines. Although consumption of soy products in Asian countries has been advanced as an explanation for the lower rates of prostate cancer in those countries, many factors beyond differences in soy intake vary between men living in Asia and men living in Western countries. Population-based studies examining the relation of phytoestrogens and soy products with prostate cancer risk are very few in number. Results are mixed for intake of soy foods and risk of prostate cancer (Severson et al., 1989; Villeneuve et al., 1999; Kolonel et al., 2000), which may be in part due to extreme differences in intake of soy products between, but not within, populations. Additional studies are needed in populations that have large inter-individual variability in soy intake and in which other factors that are correlates of soy use can be adjusted.

Legumes (Pulses) Other Than Soy Some studies have suggested inverse associations between intake of legumes and prostate cancer, including ecologic (for prostate cancer mortality r = -0.59) (Rose et al., 1986), cohort (Mills et al., 1989; Schuurman et al., 1998) and large case-control (Key et al., 1997; Cohen et al., 2000; Kolonel et al., 2000) studies. ORs of prostate cancer ranged from 0.5–0.8 comparing extreme categories in these studies. Other large case-control studies are not confirmatory (Villeneuve et al., 1999). The limited, but possibly promising data suggest that more extensive examination of diets high in legumes in relation to prostate cancer are warranted.

Tea Green tea contains polyphenols that may inhibit the growth of prostate cancer cells induced by testosterone-mediated upregulation of ornithine decarboxylase (Gupta et al., 1999). Black tea polyphenols may inhibit the growth of prostate cancer cells through effects on the IGF-1–induced Akt signal transduction pathway (Klein et al., 2002). No consistent association has been reported for either green or black tea in a prospective cohort (Severson et al., 1989), a retrospective cohort (Ellison, 2000), or in population-based case-control studies

Prostate Cancer (Slattery and West, 1993; Key et al., 1997; Villeneuve et al., 1999; Sharpe and Siematycki, 2002).

Fat, Meat, Fish, and Related Factors Fat Consumption For many decades, fat, in particular from animal sources, has been considered to be a modifiable risk factor for prostate cancer based on observations from primarily ecologic and case-control studies. Per capita fat consumption is strongly correlated with prostate cancer incidence and mortality rates internationally in ecologic studies (Armstrong and Doll, 1975; Rose et al., 1986), suggesting that dietary fat may influence the occurrence or progression of prostate cancer. The link between dietary fat or higher fat foods, especially animal foods such as meat and dairy, and a higher risk of prostate cancer has been further supported in several case-control studies (Whittemore et al., 1995a; Hayes et al., 1999; Kushi and Giovannucci, 2002), but findings from cohort studies are not consistent. Some cohort studies do not observe an association (Severson et al., 1989; Hsing et al., 1990b; Chan et al., 2000), but others support a positive association for total fat or fat from animal products (Giovannucci et al., 1993; Le Marchand et al., 1994). Given that some studies noted stronger relations for fat and advanced disease (e.g., extra-prostatic extension, metastasis, and death) ORs on the order of 1.6–2.9 across extreme categories (West et al., 1991; Giovannucci et al., 1993; Whittemore et al., 1995a; Hayes et al., 1999) suggest that diet influences late stages of carcinogenesis. Although the biologic mechanism underlying the fat-prostate cancer relation remains speculative, several mechanisms may be postulated, including effects on androgen levels, ligand-receptor binding and activity, and growth factors. Alternatively, the apparent associations for fat, in particular animal fat, may be due to not fat per se, but the animal foods that contribute to high fat intake, such as red meat and processed meat and dairy foods.

Red Meat and Processed Meat Red meat and processed meat are sources of animal fat, iron, as well as added nitrites and byproducts of cooking. In ecologic studies, total meat intake and animal protein intake are positively correlated with prostate cancer mortality (Armstrong and Doll, 1975; Rose et al., 1986). Total meat intake has not been universally associated with risk of prostate cancer in cohort studies (Snowdon et al., 1984; Severson et al., 1989; Schuurman et al., 1999b). However, in some cohort studies consumption of red or processed meat is associated with a higher risk of total or advanced prostate cancer (Giovannucci et al., 1993; Gann et al., 1994; Le Marchand et al., 1994; Veierod et al., 1997; Schuurman et al., 1999b; Michaud et al., 2001). A large case-control study also supports an association of red meat with advanced prostate cancer in both blacks and whites (Hayes et al., 1999). No association was observed in other prospective studies (Mills et al., 1989; Hsing et al., 1990b; Chan et al., 2000). Intake of poultry has not been implicated as a risk factor for prostate cancer in most prospective (Hsing et al., 1990b; Le Marchand et al., 1994; Schuurman et al., 1999b; Chan et al., 2000; Michaud et al., 2001) or large case-control (Hayes et al., 1999) studies. Relevant factors might include mutagenic heterocyclic amines, which are formed during high-temperature cooking of meat and fish through the condensation of amino acids with creatinine in muscle meats (Nagao, 1999). Only limited work has been done on the relation between meat cooking practices, degree of doneness, and heterocyclic amines and prostate cancer (Key et al., 1997; Norrish et al., 1999). Additional research with exceptionally accurate and precise assessment of exposure to heterocyclic amines coupled with knowledge of inherent variability in ability to metabolize these agents is needed.

Fish Intake Fatty fish such as salmon and tuna, contain long chain omega-3 polyunsaturated fatty acids, such as eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), which inhibit proliferation of prostate

1133

cancer cells in vitro, possibly by inhibiting cyclooxygenase-2, the enzyme that converts arachidonic acid to prostaglandins. The association of fish intake or omega-3 fatty acids with prostate cancer has not been consistent. A prospective study of Swedish men, a country in which consumption of fatty fish is common, observed a higher risk of prostate cancer incidence and mortality among men who never or rarely ate fish compared with those who consumed fish as a moderate part of their diet (Terry et al., 2001). A large prospective study of American men found a 44% lower risk of metastatic and fatal prostate cancer when comparing men who ate fish more than three times per week with those who consumed it less than twice per month (Augustsson et al., 2003). No or weak associations have been observed in other cohort studies (Severson et al., 1989; Hsing et al., 1990b; Le Marchand et al., 1994; Schuurman et al., 1999b) and case-control studies (Key et al., 1997; Villeneuve et al., 1999), and increased risks were observed in a other studies (Mills et al., 1989; Andersson et al., 1995). No associations of dietary intake of EPA or DHA in the Netherlands Cohort Study (Schuurman et al., 1999a), circulating EPA in the Physicians’ Health Study (Gann et al., 1994), or circulating EPA and DHA in the Norwegian Janus serum bank (Harvei et al., 1997) with prostate cancer have been observed. Higher concentration of erythrocyte and adipose EPA acid was associated with a lower risk of prostate cancer in a case-control study (Godley et al., 1996). Also, prostate cancer cases had lower erythrocyte EPA and DHA content than controls in a New Zealand case-control study (Norrish et al., 2000). Perhaps the lack of consistency in findings has been due to variable ranges of fish intake and omega-3 fatty acids, variability in types of fish consumed, and lack of consideration of the aggressiveness of the prostate cancer case status. Additional studies in populations with a range of fish intake along with detailed assessment of type and amount of fish consumed are still needed.

a-Linolenic Acid and Linoleic Acid

a-linolenic acid is an omega-3 polyunsaturated fatty acid found in vegetable oils, such as soy and canola, in leafy green vegetables, and in animal fat. Some cohort studies have observed an increased risk of advanced prostate cancer (Giovannucci et al., 1993), but not total prostate cancer (Schuurman et al., 1999a), with higher compared with lower intake of a-linolenic acid. Prediagnostic serum levels of alinolenic acid were positively associated with prostate cancer in nested case-control studies (Gann et al., 1994; Harvei et al., 1997). Although there was not an increasing trend, those in the top three quartiles of erythrocyte and adipose a-linolenic acid had a nonstatistically significant higher risk of prostate cancer in a small case-control study (Godley et al., 1996). Linoleic acid, an omega-6 fatty acid found in vegetable and soy oils, is a precursor to arachidonic acid and pro-inflammatory prostaglandins of the 2 series. No associations between intake of (Giovannucci et al., 1993; Schuurman et al., 1999a) or circulating concentration of (Gann et al., 1994; Harvei et al., 1997) linoleic acid have been observed. Higher concentrations of erythrocyte and adipose linoleic acid were associated with a nonstatistically significant higher risk of prostate cancer in a small case-control study (Godley et al., 1996). Dietary intake (Schuurman et al., 1999a) and circulating concentration (Gann et al., 1994; Harvei et al., 1997) of arachidonic acid, the immediate precursor to prostaglandins, were not associated with prostate cancer in prospective studies. Overall, epidemiologic studies suggest that risk of prostate cancer is higher in men with greater exposure to a-linolenic acid. Because alinolenic acid probably protects against cardiovascular disease, whether this fatty acid is causally related to prostate cancer must be resolved.

Dairy Products, Calcium, Vitamin D, and Phosphorus Dairy Products Another contributor to animal fat and protein intake in Western countries is milk and dairy products, which are also the major source of

1134

PART IV: CANCER BY TISSUE OF ORIGIN

calcium in these countries. Countries with greater per capita milk consumption have higher national prostate cancer mortality rates (r = 0.7) (Armstrong and Doll, 1975; Rose et al., 1986). The magnitude of this correlation is even greater than for other foods that are high in animal fat (e.g., meats, r = 0.39) (Rose et al., 1986). Men consuming larger amounts of milk and other dairy products are at increased risk, on the order of twofold, for prostate cancer in population-based case-control (Chan et al., 1998a; Hayes et al., 1999) and prospective cohort (Snowdon et al., 1984; Le Marchand et al., 1994; Giovannucci et al., 1998c; Schuurman et al., 1999b) studies, and in one study even after adjusting for total fat or animal fat (Giovannucci et al., 1998c). Not all prospective studies are in agreement (Mills et al., 1989; Hsing et al., 1990b; Chan et al., 2000), though.

Calcium Possibly the observed higher risk of prostate cancer associated with higher dairy intake could be mediated by a component other than its fat content. Because of its important activities for bone maintenance and in cell signaling pathways, calcium absorption is finely regulated. Higher circulating and tissue levels of calcium down-regulate production of 1,25-dihydroxyvitamin D (1,25(OH)2D), a steroid hormone that augments calcium absorption and also plays a role in control of proliferation and differentiation of cells including prostatic epithelial cells (see below). Ingestion of high calcium may suppress 1,25(OH)2D levels and men who systemically have lower levels of 1,25(OH)2D have been hypothesized to be at increased risk for prostate cancer. Case-control (Chan et al., 1998a) and prospective cohort (Giovannucci et al., 1998c; Chan et al., 2000; Chan et al., 2001b) studies have observed positive relations between calcium intake and prostate cancer risk. Risk may be greater at supplement doses (Giovannucci et al., 1998c) and for advanced disease (Chan et al., 1998a; Giovannucci et al., 1998c), although these findings have not been consistent across all studies (Hayes et al., 1999; Kristal et al., 1999; Schuurman et al., 1999b). Ultimately to determine whether high calcium intake is a risk factor for prostate cancer, studies must be conducted in populations with a wide range of intake, including levels at twice or more the recommended intake for men and in populations in which calcium supplement use is prevalent.

Vitamin D The calcium hypothesis is complemented by research on the possibly protective role of vitamin D on prostate cancer risk. Schwartz and Hulka (1990) hypothesized that vitamin D protects against prostate cancer based on indirect evidence from a correlational study showing that US states with higher ultraviolet radiation tend to have lower prostate cancer mortality rates. Vitamin D, or cholecalciferol, is derived mainly from ultraviolet light conversion of 7-dehydrocholesterol in the skin, but also from intake of foods, such as fortified milk products and breakfast cereals and vitamin D-containing multivitamins and supplements. 1,25(OH)2D inhibits proliferation and induces differentiation of prostatic epithelial cells (Peehl et al., 1994) and may be particularly important for preventing prostate cancer metastasis (Sung et al., 2000). Perhaps because 1,25(OH)2D is tightly homeostatically regulated and is not substantially influenced by dietary vitamin D intake, studies generally do not support an association between dietary (Giovannucci et al., 1998c; Chan et al., 2000) or supplemental intake of vitamin D and prostate cancer. Among the five case-control studies nested in prospective cohort studies that have examined the vitamin D and prostate cancer hypothesis by measuring circulating vitamin D metabolites (Corder et al., 1993; Braun et al., 1995; Gann et al., 1996b; Nomura et al., 1998; Ahonen et al., 2000), only one supports an inverse association for 1,25(OH)2D (Corder et al., 1993) and another supports an inverse association for 25-hydroxyvitamin D (25(OH)D) (Ahonen et al., 2000). In that latter study, more than half of the participants were clinically deficient in vitamin D, whereas the other studies were evaluated in vitamin-D replete populations. Perhaps only deficiency in 25(OH)D or 1,25(OH)2D increases the risk of prostate cancer.

Energy Intake, Obesity, and Physical Inactivity The epidemic of obesity in the United States has led to renewed interest in the role of energy imbalance and resultant obesity in relation to cancer risk. Energy imbalance results when energy intake exceeds total energy expenditure. Energy imbalance is not feasibly measured in large epidemiologic studies. Surrogates include measurement of energy intake from diet questionnaires, which is prone to substantial measurement error, and the major determinants in variability in energy intake, such as body size, and physical activity level. These surrogates are reviewed in relation to prostate cancer risk below.

Energy Intake Animal studies consistently show that diets with restricted total energy reduce tumor burden relative to ad libitum feeding (Kritchevsky, 1999; Thompson et al., 1999). In a transplantable prostate tumor model, moderate energy restriction reduced prostate tumor growth, lowered circulating concentrations of insulin-like growth factor-1 (IGF-1), and decreased expression of vascular endothelial growth factor (Mukherjee et al., 1999). The effects of energy restriction on factors mediating greater proliferation relative to apoptosis and angiogenesis together suggest that excessive energy intake may act late in the carcinogenic pathway. Findings for energy intake and prostate cancer have not been consistent among the more than 20 epidemiologic studies (Platz, 2002a). However, these studies have not systematically considered the balance of energy input with body size and physical activity, or whether the energy association differs by whether the tumor was organ confined or metastatic. A recent large prospective study observed a positive association between energy intake and prostate cancer that is regionally invasive or worse primarily in leaner men (Platz et al., 2002d). These men who remain lean despite higher energy intake may have a greater basal or average metabolic rate. These results must be confirmed.

Obesity Prospective studies have reported conflicting results for BMI or body weight and prostate cancer (Nomura, 2001). A 34% higher risk of prostate cancer death for a very high BMI (35+ kg/m2) compared with normal BMI was reported in the prospective Cancer Prevention Study II (Calle et al., 2003). The results from another prospective study (Severson et al., 1988) suggested that a positive association observed between BMI and prostate cancer risk was due to muscle mass, which is influenced by androgens, and not due to adiposity. BMI as a marker for lean body mass rather than adiposity may also be the explanation for the statistically significant positive relation seen between BMI and incident and fatal prostate cancer in a large Swedish retrospective cohort study of construction workers (Andersson et al., 1997). Indeed, in that cohort lean body mass was positively associated with prostate cancer risk (Andersson et al., 1997), but not in another cohort (Lund Nilsen and Vatten, 1999). Case-only studies suggest that risk of highergrade disease in men who had clinically organ-confined disease is greater with higher BMI (Spitz et al., 2000; Rohrmann et al., 2003). A few studies have evaluated BMI or body weight earlier in life and results from these studies are conflicting (Andersson et al., 1995; Giovannucci et al., 1997a; Schuurman et al., 2000). In the Health Professionals Follow-up Study, obesity in adulthood was not associated with total prostate cancer risk or with distant metastatic and fatal disease (Giovannucci et al., 1997a). However, in younger men (<60 years old) and those with a family history of prostate cancer, obesity was inversely associated with prostate cancer (Giovannucci et al., 2003). Multiple physiologic systems are perturbed in overweight and obese men, such as insulin and glucose control and the balance of sex steroids. Some of these perturbations may be predicted to increase risk of prostate cancer (e.g., high circulating insulin and glucose) or decrease risk (e.g., high circulating estrogen relative to androgen). Thus, the conflicting results for BMI and prostate cancer may be in part due to the variable relation of high BMI with fat mass and lean mass and resultant effects on hormonal systems and on the nature of cancer, whether hereditary or sporadic disease. Also, the results for obesity and incidence of prostate cancer vs. death from

Prostate Cancer prostate cancer may differ because mortality encompasses progression, survival, and late detection, as well as etiology.

Central Adiposity Central adiposity is linked to hyperinsulinemia and higher risk of type 2 diabetes mellitus. Waist circumference or the ratio of waist to hip circumferences are frequently used as surrogates for central adiposity. Men with a higher ratio of waist to hip circumference, but not a higher waist circumference alone, had a higher risk of metastatic prostate cancer in a US cohort (Giovannucci et al., 1997a). In a case-control study conducted in China waist to hip ratio was positively associated with total prostate cancer as well as localized and metastatic disease (Hsing et al., 2000a).

Birth Weight and Pregnancy Factors Weight at birth may reflect in utero exposure to nutrients, steroid hormones, and fetal growth factors (Godfrey et al., 1995) and has been hypothesized to be associated with prostate cancer risk. A small cohort study (21 cases in 366 men) found that prostate cancer incidence was more than four times higher comparing the upper to the lower quartile (Tibblin et al., 1995); this finding has not been confirmed (Ekbom et al., 1996; Platz et al., 1998b). A nested case-control study found that perinatal characteristics including pre-eclampsia and prematurity, which are likely correlates of pregnancy hormones and other growth factors, were inversely and mother’s parity directly associated with prostate cancer incidence and mortality (Ekbom et al., 1996).

Physical Inactivity The influence of physical activity on prostate cancer risk is uncertain. Reductions in serum testosterone have only been observed in elite athletes, but not in men participating in recreational exercise (Wheeler et al., 1984; Hackney, 1996). Many studies have noted a modest inverse relation between occupational and/or leisure time physical activity (Vena et al., 1987; Yu et al., 1988; Albanes et al., 1989; Brownson et al., 1991; Hsing et al., 1994; Thune and Lund, 1994; Hartman et al., 1998b; Lund Nilsen et al., 2000; Norman et al., 2002) and cardiorespiratory fitness (Oliveria et al., 1996) and prostate cancer, whereas others have found no association (Severson et al., 1989; Whittemore et al., 1995a; Liu et al., 2000b; Putnam et al., 2000; Lacey et al., 2001). Two studies reported positive associations between physical activity and subsequent prostate cancer (Cerhan et al., 1997; Clarke and Whittemore, 2000), which was pronounced in African-American men but not white men (Clarke and Whittemore, 2000). Three studies found positive associations for physical activity in young adulthood (Polednak, 1976; Whittemore et al., 1985; Paffenbarger Jr., et al., 1987), whereas another study observed an inverse association for physical activity during puberty (Andersson et al., 1995). No association between physical activity and prostate cancer risk overall was seen in the Harvard Alumni cohort (Lee et al., 1992) and the Health Professionals Follow-up Study (Giovannucci et al., 1998a). However, in these cohorts, a lower risk among older men who had extremely high-energy expenditures was observed for total (Lee et al., 1992) or metastatic prostate cancer (Giovannucci et al., 1998a). Overall, these results suggest exercise in adulthood is unlikely to either increase or lower risk appreciably, except perhaps for a benefit at very high levels of activity. The level of activity required for a benefit as observed in recent cohort studies may be too high for everyone to adopt. The possible positive association between exercise earlier in life and subsequent risk of prostate cancer is inconclusive and may be explained by confounding by correlates of higher activity levels during adolescence (greater muscle mass, higher androgen levels).

Other Lifestyle Factors

1135

metastatic or fatal prostate cancer (Hickey et al., 2001). Dose-response across numbers of cigarettes smoked per day or number of pack-years has not been detected (Hsing et al., 1990b; Adami et al., 1996; Rodriguez et al., 1997; Giovannucci et al., 1999). Risk may be greatest for recent smoking rather than cumulative exposure (Giovannucci et al., 1999). Smoking may augment the aggressiveness of prostate tumors, as measured by stage or grade (Daniell, 1995; Spitz et al., 2000; Roberts et al., 2003). Although differences in diagnosis-seeking between smokers and non-smokers cannot be ruled out, carcinogens in cigarette smoke may promote tumor progression. Smoking also is associated with higher plasma testosterone levels in men (Dai et al., 1988; Field et al., 1994), which may enhance tumor progression. Although the mechanism remains unknown, additional work is needed on smoking in relation to fatal prostate cancer and prostate cancer survival.

Alcohol Drinking Nearly 60 epidemiologic studies have evaluated the association between alcohol consumption and risk of prostate cancer. The absolute amount of alcohol consumed is generally not strongly related to prostate cancer incidence or mortality in prospective and record linkage studies of men not selected for alcoholism (Dennis, 2000). Meta-analyses have reported summary estimates of 1.15 (95% CI: 1.00–1.32) and 1.21 (95% CI: 1.05–1.39) for three and four drinks per day, respectively (Dennis, 2000) and 1.09 (95% CI: 1.02–1.17) and 1.19 (95% CI: 1.03–1.37) for 50 g/day (~3–4 drinks daily) and 100 g/day (~6–8.5 drinks daily), respectively (Bagnardi, 2001). Prospective studies reported after these meta-analyses were completed tended to support modest direct associations, as well (Putnam et al., 2000; Sesso et al., 2001; Platz, 2004). A lower incidence of prostate cancer has not been observed among alcoholics (Adami et al., 1992; Tonnesen et al., 1994; Hayes et al., 1996; Lund Nilsen et al., 2000), who because of cirrhosis secondary to excessive alcohol intake might have lower circulating testosterone concentrations (Green, 1997). Overall the risk of prostate cancer associated with alcohol drinking appears to be very modest at best.

Medication Use and Medical Procedures Aspirin and Non-Steroidal Anti-Inflammatory Agents Aspirin and NSAIDS have been hypothesized to decrease prostate cancer risk by inhibiting cyclooxygenase (COX) enzymes. Inducible COX-2 catalyzes the conversion of arachidonic acid into pro-inflammatory prostaglandins (Hussain et al., 2003). COX-2 expression is lost in prostate adenocarcinoma, but is overexpressed in PIA, especially in associated macrophages (Zha et al., 2001). Overall, prospective and case-control studies suggest a weak inverse association between regular use of aspirin or NSAIDs and risk of prostate cancer. Among the prospective studies, the relative risk of prostate cancer or prostate cancer death for regular aspirin use ranges from 0.76 (Habel et al., 2002) to 1.05 (Leitzmann et al., 2002), although in the latter study, the relative risk of metastatic/fatal prostate cancer was 0.73 (Leitzmann et al., 2002). Among the case-control studies, two reported strong inverse associations of 0.35 (Nelson and Harris, 2000) and 0.45 (Roberts et al., 2002); others reported modest estimates around 0.85 for regular use (Norrish et al., 1998; Bucher et al., 1999; Irani et al., 2002). As for all case-control studies, differential likelihood of participation of cases and controls by use of aspirin and NSAIDs and differential accuracy in recall of use of aspirin and NSAIDs between cases and controls are possible. Although the benefits of anti-inflammatory agents in the prevention of prostate cancer may be modest, additional studies are warranted that have adequate assessment of aspirin and NSAIDs use, types, dose, and duration of use, as well as repeated measurement of use over time.

Cigarette Smoking

Vasectomy

Cigarette smoking is a major source of mutagens and chronic oxidative stress. Most cohort studies do not support an association between cigarette smoking and overall prostate cancer incidence although several, but not all studies, have observed positive associations with

The relation between vasectomy and prostate cancer is controversial. A meta-analysis reported a summary estimate for 22 studies of 1.37 (95% CI: 1.15–1.62) for ever having had a vasectomy; the summary estimate was 1.22 (95% CI: 0.90–1.64) for the five cohort studies

1136

PART IV: CANCER BY TISSUE OF ORIGIN

(Dennis et al., 2002a). Risk of prostate cancer was greater for a longer time since vasectomy, but there was no difference in effect by age at vasectomy (Dennis et al., 2002a). Major criticisms of the studies finding an association are that detection bias or confounding may account for the results (Howards et al., 1993; Skegg, 1993). Any biological basis underlying the vasectomy-prostate cancer relation remains speculative.

Occupational and Environmental Exposures Farmers have been noted to have a higher risk of prostate cancer in retrospective and linkage studies, especially farmers with suspected exposure to pesticides (van der Gulden et al., 1995; Sharma-Wagner et al., 2000; Mills and Yang, 2003; Settimi et al., 2003). A very large US death certificate study did not observe a higher risk of prostate cancer death in farm workers, but did observe high risks for certain white collar jobs and blue collar jobs (Krstev et al., 1998). A prospective study did not support an association for polychlorinated biphenyls with prostate cancer death, but did suggest a higher risk of death for the high electromagnetic field exposure in electric utility workers (Charles et al., 2003). Most of these studies are based on job titles and did not quantify exposure to specific agents for individual workers. Evidence for an association between occupational exposure to cadmium and prostate cancer is weak (Kolonel and Winkelstein, 1977; Armstrong and Kazantzis, 1985; Elinder et al., 1985; Kjuus et al., 1986; Elghany et al., 1990; van der Gulden et al., 1995). Results for low-level environmental cadmium exposure in relation to prostate cancer risk are not consistent (Kolonel and Winkelstein, 1977; Elghany et al., 1990; West et al., 1991; Platz et al., 2002b). Cadmium is unlikely to be a major risk factor for prostate cancer in the general population.

HOST FACTORS Family History Family history of prostate cancer, as a surrogate for shared genetics, has been widely studied and is consistently positively related to prostate cancer risk, with relative risks ranging from 2–3 for having either a father or brother with the disease (Stanford and Ostrander, 2001). Prostate cancer risk is higher with increasing number of firstdegree relatives with the disease (two or more RR = 3–5) (Whittemore et al., 1995b; Lesko et al., 1996; Rodríguez et al., 1997), and higher if the first-degree relative was younger at diagnosis (<65 years: RR = 1.5–6.0) (Cannon et al., 1982; Grönberg et al., 1996; Lesko et al., 1996; Rodríguez et al., 1997). It is unclear whether there is a difference in the aggressive behavior of prostate tumors in men with and without a family history (Hayes et al., 1995b; Kupelian et al., 1997; Kotsis et al., 2002; Rohrmann et al., 2003). Further support for an inherited component to prostate cancer risk is that twin studies show higher concordance for prostate cancer diagnosis between monozygotic than dizygotic twins (Page et al., 1997; Lichtenstein et al., 2000). Although genetics may account for as much as 42% of the concordance between twins, non-genetic (e.g., environmental) factors appear to play a substantial role in influencing risk (Lichtenstein et al., 2000). Segregation analyses in prostate cancer kindreds support an autosomal dominant mode of inheritance (Carter et al., 1992; Grönberg et al., 1997; Gong et al., 2002), with the attributable risk for the yet unidentified gene(s) ranging from than 43% in men diagnosed under age 55 years, to 34% under age 70, to 9% under age 85 years (Carter et al., 1992). Another study suggested the dominant mode of inheritance was more likely for men with a young age at onset and an X-linked mode for men who were older at diagnosis (Valeri et al., 2003). Studies of clustering of prostate cancer with cancers of other sites in family members have shown at most a modest excess incidence of these and other cancers among first-degree relatives (Goldgar et al., 1994; Hayes et al., 1995b). Linkage studies have identified several possible loci, including HPC1 (1q24-25), PCAP (1q42.2-43), CAPB (1p36), HPX (Xq27-29),

HPC2 (17p11), HPC20 (20q13), and 16q23 (Elo and Visakorpi, 2001; Nwosu et al., 2001). However, linkage at these loci has not been replicable across populations (Elo and Visakorpi, 2001). Efforts to identify the genes lying circa these loci are underway. For example, RNASEL, the gene encoding ribonuclease L, has been identified as a candidate for HPC1 and rare constitutional sequence variants have been identified in this gene in familial prostate cancer cases (Carpten et al., 2002). Several groups have evaluated the relation of common polymorphisms in the genes linked to familial prostate cancer to prostate cancer in men not selected for family history. For example, the HPC2/ELAC2 Ala541Thr variant has been associated with a 27% (95% CI: 1.06–1.54) higher risk of prostate cancer when summarized across several case-control studies (Meitz et al., 2002). Although the proportion of men with prostate cancer who report another family member with prostate cancer is relatively high, it is likely that most of this is due to the high incidence of disease rather than the inheritance of germline mutations in major cellular growth control or repair genes. Familial prostate cancer probably accounts for less than 10% of cases overall, although perhaps a higher percentage in younger patients (Carter et al., 1992).

Sex Steroid Hormones, the Androgen Receptor, and Hormone Metabolizing Enzymes Sex Steroid Hormones Androgens influence maturation of the prostate and are believed to contribute to the development and progression of prostate cancer. Several prospective studies have assessed the association of circulating sex steroid hormone concentrations in relation to prostate cancer (Hsing et al., 2001a). Only the study conducted in the Physicians’ Health Study observed that testosterone and androstanediol glucuronide, a metabolite of dihydrotestosterone, were statistically significant positively associated with prostate cancer, and estradiol and sex hormone binding globulin were inversely associated with prostate cancer (Gann et al., 1996a). In that prospective study, in which the mean time from blood draw to diagnosis was 6 years and most of the cases were diagnosed before the PSA era (1982–1992), these findings were apparent only after mutual adjustment for these hormones. There was some evidence for a higher risk of prostate cancer for the ratio of testosterone to either dihydrotestosterone or androstanediol glucuronide in three of the prospective studies (Nomura et al., 1988; Hsing and Comstock, 1993; Dorgan et al., 1998). A meta-analysis of the prospective studies conducted up to 1999 did not find case-control differences in hormones, except possibly for slightly higher concentrations of androstanediol glucuronide in the cases (Eaton et al., 1999). None of the studies included in the meta-analysis, except the study by Gann et al. (1996a) adjusted for the major binding protein for sex hormone or mutually for hormones hypothesized to have opposing actions, which might account for the overall lack of case-control differences in hormone levels. However, mutual adjustment for testosterone and estradiol in another study did not alter the lack of associations of these hormones with prostate cancer (Dorgan et al., 1998). The study of sex steroid hormones has been hampered by the evaluation of a limited number of components of a large, complex, and interrelated pathway and by measurement issues (e.g., poor precision and single measurement in adulthood). Also, it is unknown when during the course of a man’s life androgens influence prostate cancer risk, the effect of the balance between androgens and estrogens, and whether levels of steroid hormones within the normal range are contributory.

Androgen Receptor Gene CAG Repeat The androgen receptor mediates the effect of testosterone and dihydrotestosterone in androgen-responsive tissues (Jänne et al., 1993). The androgen receptor gene, located on the long arm of the X chromosome, contains a variable-length CAG repeat (encodes polyglutamine) in exon 1. In experimental constructs fewer repeats increase the transactivational activity of the receptor (Chamberlain et al., 1994; Kazemi-Esfarjani et al., 1995). The normal range of CAG repeats is

Prostate Cancer 11–31 (Edwards et al., 1992), with the mean being shorter in African American (~20 repeats) than in white (~22 repeats) men (Edwards et al., 1992; Platz et al., 2000b). Some (Irvine et al., 1995; Giovannucci et al., 1997b; Ingles et al., 1997; Stanford et al., 1997; Hsing et al., 2000b; Xue et al., 2000), but not all (Bratt et al., 1999; Correa-Cerro et al., 1999b; Edwards et al., 1999; Beilin et al., 2001; Latil et al., 2001; Miller et al., 2001; Chen et al., 2002; Li et al., 2003), epidemiologic studies support that shorter androgen receptor gene CAG repeats are associated with a higher risk of prostate cancer. In a prospective study, the risk appeared to increase monotonically with decreasing CAG repeat number (Giovannucci et al., 1997b). Explanations for the inconsistency in findings among studies are in the heterogeneity in the case mix and differences in the age distributions of the cases (Giovannucci, 2002a). Greater consistency is found when considering advanced cases or cases diagnosed in the pre-PSA era and when considering a young age at onset. A second polymorphic androgen receptor (AR) gene trinucleotide repeat also in exon 1, consisting of GGN (N = any of the four nucleotides) repeats encoding polyglycine, has been described (Faber et al., 1989), although the effect of the length of this repeat on AR transactivation is unknown. About 85%–90% of the individuals studied thus far have the most prevalent allele or one longer (Sleddens et al., 1993). The association between the AR gene GGN repeat has not been consistent among studies (Irvine et al., 1995; Stanford et al., 1997; Platz et al., 1998a; Edwards et al., 1999; Hsing et al., 2000b; Miller et al., 2001; Chen et al., 2002).

SRD5A2 The conversion of testosterone to dihydrotestosterone is catalyzed in the prostate by 5 a-reductase type II, which is encoded by SRD5A2 on chromosome 2. In the prostate, dihydrotestosterone has greater affinity for the androgen receptor, resulting in greater transactivation of androgen-responsive genes. A TA dinucleotide repeat in the 3¢untranslated region (Davis and Russell, 1993), a V89L substitution, which reduces enzyme activity in vitro, and an A49T substitution, which results in a higher enzymatic activity (Makridakis et al., 1999), have been evaluated in relation to prostate cancer risk. Three studies observed evidence against the hypothesized effect of the TA repeat (i.e., that longer repeats would be associated with a higher risk) (Kantoff et al., 1997; Hsing et al., 2001b; Latil et al., 2001). A Canadian case-control study noted a 2.5-fold higher risk of prostate cancer for the valine allele in men undergoing biopsy for elevated PSA/DRE (Nam et al., 2001). With the exception of one study (Nam et al., 2001), no association has been found for the V89L polymorphism (Febbo et al., 1999; Lunn et al., 1999; Jaffe et al., 2000; Hsing et al., 2001b; Latil et al., 2001; Yamada et al., 2001; Pearce et al., 2002; Soderstrom et al., 2002). The A49T substitution was associated with a higher risk of prostate cancer and poorer grade (Makridakis et al., 1999; Jaffe et al., 2000), but not elsewhere (Latil et al., 2001; Soderstrom et al., 2002).

3b-Hydroxysteroid Dehydrogenase Types I and II (HSD3B) HSD3B1 (expressed in the prostate) and HSD3B2 (expressed in the testes), encode two enzymes that metabolize dihydrotestosterone. HSD3B1 contains two nonsynomymous polymorphisms (F286L and N367T) in coding regions (Chang et al., 2002) and HSD3B2 contains a polymorphic dinucleotide repeat (TG)n(TA)n(CA)n (Devgan, 1997) and two polymorphisms (C7474T and C7519G) in the 3¢ untranslated region (Chang et al., 2002). The functional significance of these polymorphisms is not known. The HSD3B1 N367T variant was associated with a higher risk of prostate cancer, especially for the combination of the HSD3B1 N367T variant allele and the HSD3B2 C7519G variant allele (Chang et al., 2002). More studies on polymorphisms in HSD3B are needed to understand relevance of normal variability in its encoded enzyme in relation to prostate cancer risk.

Cytochrome P-450 Enzymes Several of the P450 enzymes metabolize sex steroid hormones. Polymorphisms in the genes that encode these enzymes have been identi-

1137

fied; the prevalences of some vary substantially by race/ethnicity, such as CYP3A4 (Walker et al., 1998). CYP3A4 deactivates testosterone via hydroxylation. The -290 G CYP3A4 variant allele, which enhanced transcription in a artificial construct (Amirimani et al., 1999), was associated with a higher risk of prostate cancer diagnosis in a prospective study of men with BPH who were followed for 11 years (Tayeb et al., 2003), and was associated with worse disease at diagnosis when comparing among prostate cancer cases, especially in older men (Rebbeck et al., 1998; Paris et al., 1999). An association between the variant allele and prostate cancer was also noted in African Americans (Paris et al., 1999; Kittles et al., 2002), but was possibly attributed to population stratification in one study (Kittles et al., 2002). The variant genotype for G119T in CYP1B1, but not five other polymorphisms in that gene, was associated with a higher risk of prostate cancer in a Japanese population (Tanaka et al., 2002). The valine/valine genotype for CYP1A1 (hydrocarbon hydroxylase), but not polymorphisms in CYP1A2, or CYP2E1, was associated with a higher risk of prostate cancer in a Japanese population (Murate et al., 2001). No associations have been reported for CYP2D6 (debrisoquine hydroxylase) (except possibly in smokers and industrial workers) (Wadelius et al., 1999) or CYP2C19 (Wadelius et al., 1999). In a meta-analysis of 10 case-control studies there was no overall association between a T to C substitution 34 basepairs upstream of the translations start site, but downstream of the transcription start site of CYP17, which catalyzes two reactions in the testosterone biosynthesis, and prostate cancer (Ntais et al., 2003). Overall, there is modest evidence for associations between polymorphisms in some CYP genes and prostate cancer.

Sexual Activity Sexual activity may plausibly be related to prostate cancer through some causal and some indirect mechanisms. Extent of sexual drive is in part determined by androgenicity or may enhance exposure to sexually transmitted agents (see below). The epidemiologic studies reporting on sexual activity are case-control studies (Dennis and Dawson, 2002c), which may suffer from biased report of past sexual activity and from incorrect temporality because changes in sexual activity may occur due to the disease itself.

Balding Because of the links between androgens and male pattern balding (vertex) and androgens and prostate cancer, a small number of studies have investigated whether balding is associated with prostate cancer. Positive associations (RRs of 1.5) between male pattern balding and prostate cancer have been observed in case-control studies (DemarkWahnefried et al., 2000; Giles et al., 2002) and in a cohort study (Hawk et al., 2000), in particular early age at onset of balding (DemarkWahnefried et al., 2000) and high-grade disease (Giles et al., 2002) Another case-control study reported no association (Hsieh et al., 1999). The positive associations between balding and prostate cancer might also be explained by the finding of a positive association between plasma IGF-1 levels and male pattern balding (Signorello et al., 1999; Platz, 2000a).

Polymorphisms in Vitamin D Receptor Gene and 1-a-hydroxylase The effects of 1,25(OH)2D are mediated by its binding to the cytoplasmic vitamin D receptor, which transactivates transcription of target genes (Reichel et al., 1989). The following polymorphisms have been evaluated in relation to prostate cancer: a poly-A microsatellite in the 3¢-untranslated region, and the following restriction length polymorphisms: BsmI in intron 8 (denoted b), ApaI in intron 8 (denoted a), and TaqI in exon 9, which results in a base, but not amino acid change (C352T and denoted t). All of these polymorphisms are in strong linkage disequilibrium. None of these polymorphisms affects the amino acid coding sequence of the vitamin D receptor, although the BAt haplotype has been reported to have greater vitamin D receptor transcriptional activity or enhanced mRNA stability in artificial gene constructs than the baT haplotype, and individuals with the BB

1138

PART IV: CANCER BY TISSUE OF ORIGIN

genotype have higher circulating 1,25(OH)2D levels than Bb or bb genotypes (Morrison et al., 1994). Also, examined is the FokI polymorphism in exon 2, which results in a new start codon. Although widely studied, no consistent associations have emerged for these polymorphisms (Ingles et al., 1997; Ingles et al., 1998; Kibel et al., 1998; Ma et al., 1998; Correa-Cerro et al., 1999a; Habuchi et al., 2000; Chokkalingam et al., 2001b; Tayeb et al., 2003). No association has been observed between polymorphisms in the gene encoding 25hydroxyvitamin D 1-a-hydroxylase, the enzyme that catalyzes the conversion of 25(OH)D to 1,25(OH)2D (Hawkins et al., 2002), and prostate cancer risk.

Growth Factors and Related Factors Insulin, Glucose, and Type 2 Diabetes Mellitus Patients with type 2 diabetes mellitus initially have hyperglycemia and hyperinsulinemia, with waning insulin concentrations over decades with pancreatic beta cell burnout. Hyperinsulinemia might be predicted to increase the risk of prostate cancer through growth promotion: however, studies on circulating insulin and glucose concentrations and measures of insulin resistance, and risk of prostate cancer have not been consistent (Thune and Lund, 1994; Lund Nilsen et al., 2000; Stattin et al., 2000; Gapstur et al., 2001; Hsing et al., 2001c; Hsing et al., 2003; Hubbard et al., 2004). Contrary to the hypothesis some studies have noted a 30%–70% reduction in prostate cancer risk among diabetics (Thompson et al., 1989; Adami et al., 1991; La Vecchia et al., 1997; Giovannucci et al., 1998b; Rosenberg et al., 2002). No clear evidence of an inverse association could be detected in other prospective (Ragozzino et al., 1982; Coughlin et al., 1996; Will et al., 1999) or case-control (Tavani et al., 2002) studies. Positive associations were observed in two other prospective studies (Steenland et al., 1995; Lund Nilsen et al., 2000). Men with severe type 2 diabetes mellitus manifest lower androgen levels (Andò et al., 1984), possibly resulting from the detrimental effect of hyperglycemia on the Leydig cells. Thus, the lack of an association in some studies and the reduced risk of prostate cancer in other studies may reflect the balance or imbalance of the beneficial effects of reduced androgenicity against the adverse effects of hyperinsulinemia or hyperglycemia on growth promotion. The risk of prostate cancer may decrease with increasing time since diagnosis of diabetes, possibly because androgen aberrations may increase with worsening of diabetes while insulin is decreasing. Although the findings for circulating insulin, glucose, and diagnosis of diabetes are not consistent because of the mounting prevalence of type 2 diabetes and pre-diabetes in the United States, additional studies are warranted to examine the role of hyperinsulinemia and hyperglycemia in prostate cancer occurrence.

Leptin Leptin is a peptide hormone that belongs to the cytokine family (Matarese et al., 2002) and controls body weight by modulating energy utilization (Friedman, 2002). In obesity, individuals become leptin resistant and exhibit elevated plasma leptin (Chu et al., 2001). Findings for circulating leptin concentrations and prostate cancer have not been consistent (Chang et al., 2001; Hsing et al., 2001c; Stattin et al., 2001; Stattin et al., 2003).

Height Attained height reflects factors that characterize the growth phase of adolescence, including marked changes in circulating concentrations of steroid hormones, growth hormone, and IGF-1. Possibly because it is a marker for levels of these growth factors, height has been positively associated with total prostate cancer in some, but not all studies (Nomura, 2001). In a large cohort of US health professionals, taller height was more strongly associated with advanced disease (Giovannucci et al., 1997a). Childhood height was positively associated with prostate cancer risk in American whites, but not in American blacks in a large case-control study (Hayes et al., 1999). Height may not be a consistent risk factor in all populations because the relative importance of nutritional and genetic factors may vary across populations.

IGF-1 and IGFBP-3 IGF-1 is a peptide hormone that promotes growth in adolescence and childhood and is correlated with adult lean body mass (Severson et al., 1988). At the cellular level, IGF-1 promotes proliferation and inhibits apoptosis, including in normal prostate and tumor cells in vitro (Cohen et al., 1991; Cohen et al., 1994). IGF-1 is produced by the liver, but is also produced locally in tissues, including in the prostate. IGF-1 circulates bound to binding proteins, the most prevalent of which is insulin-like growth factor binding protein (IGFBP)-3. In the prostate, IGFBP-3 appears to promote apoptosis (Rajah et al., 1997) by interactions with the retinoid X receptor (Liu et al., 2000a). Interestingly, IGFBP-3 can be cleaved by PSA (Koistinen et al., 2002), which is a protease, and thus would eliminate its ability to influence apoptosis. Many studies support the positive association between circulating concentration of IGF-1 and risk of prostate cancer (Mantzoros et al., 1997; Chan et al., 1998b; Wolk et al., 1998; Harman et al., 2000; Stattin et al., 2000; Chokkalingam et al., 2001a; Khosravi et al., 2001). The findings for IGFBP-3 are not consistent; however, in most of those studies in which IGFBP-3 was adjusted for IGF-1 an inverse association was suggested overall (Chan et al., 1998b; Harman et al., 2000; Chokkalingam et al., 2001a; Chan et al., 2002) or at least in a subset of men (Stattin et al., 2000). In a continuation of the Physicians’ Health Study through 1995, IGF-1 adjusting for IGFBP-3 was no longer associated with prostate cancer overall, but was limited to cases that were of advanced stage at diagnosis (C or D) (Chan et al., 2002). Consistent with the hypothesis that IGF-1 is associated with a higher risk of prostate cancer, a meta-analysis reported a summary OR of prostate cancer of 1.47 (95% CI: 1.23–1.77) comparing high to low IGF-1 (Shi et al., 2001). Contrary to the hypothesis of a protective effect of IGFBP-3, the summary estimate was 1.26 (95% CI: 1.03–1.54) comparing high to low IGFBP-3 (Shi et al., 2001). This study did not include (Chokkalingam et al., 2001a; Chan et al., 2002). Explanation for the increased risk associated with IGFBP-3 in the meta-analysis is that IGFBP-3 was not adjusted for IGF-1 in calculating the summary estimates; many of the studies included were small studies in which circulating concentrations were measured concurrently with disease. IGF-2 has not been consistently associated with prostate cancer in prospective (Chan et al., 1998b; Harman et al., 2000) or case-control (Chokkalingam et al., 2001a) studies. IGFPB-2, which is the major IGF-1 binding protein in the prostate, may influence the progression of prostate cancer cells by inducing telomerase activity (Moore et al., 2003).

Infection and Response to Infection Sexually Transmitted Infections (STIs) A meta-analysis reported statistically significant summary estimates from 23 case-control studies of 1.44 for any history of sexually transmitted infections (STIs), 2.30 for syphilis, and 1.34 for gonorrhea (Dennis and Dawson, 2002c). Although suggestive of a link between STIs and prostate cancer, all of these studies ascertained STI history retrospectively and likely without blinding to case-control status, raising concern about differential accuracy in report of past STI. Positive associations with prostate cancer have been reported for circulating antibodies against syphilis (Hayes et al., 2000) and human papillomavirus (HPV) serotypes 16 (Dillner et al., 1998; Hisada et al., 2000) and 18 (Dillner et al., 1998), but not for HPV serotypes 11, 16, or 23 (Dillner et al., 1998; Strickler et al., 1998; Hayes et al., 2000), Chlamydia species (Dillner et al., 1998), herpesvirus-8 (Sitas et al., 1999), and herpesvirus type 2 (Herbert et al., 1976; Baker et al., 1981). The presence of an infectious agent in prostate tumor tissue has not been consistent (Baker et al., 1981; Strickler et al., 1998; Griffiths and Mellon, 2000; Grinstein et al., 2002). The role of STIs and other infections in the etiology of prostate cancer remains unresolved.

Prostatitis Clinical prostatitis is a common urological condition that in some cases caused by bacteria and is characterized by uncomfortable symptoms such as pain in the perineum and pain or burning during urina-

Prostate Cancer tion (Roberts et al., 1998). A meta-analysis of 11 case-control studies reported a statistically significant summary estimate for prostate cancer of 1.57 for ever having had prostatitis (Dennis et al., 2002b). These studies potentially suffer from recall bias, variable quality confirmation of prostatitis, inability to classify type of prostatitis or to detect asymptomatic prostatitis, and prostate cancer detection bias (Dennis et al., 2002b).

Response to Infection In response to infection, but also in response to cell damage due to oxidation and hypoxia, immune cells secrete cytokines, such as interleukin (IL)-6 and tumor necrosis factor (TNF)-a. IL-6 is also produced by prostate tumor cells and is a prostate cancer cell growth factor (Twillie et al., 1995; Okamoto et al., 1997). Men with metastatic or hormone refractory prostate cancer have elevated plasma IL-6 levels (Twillie et al., 1995; Adler et al., 1999; Drachenberg et al., 1999; Nakashima et al., 2000; Wise et al., 2000; Shariat et al., 2001), which is likely due to a response to the tumor or secreted by the tumor itself, rather than a cause of the tumor. Common sequence variants in genes encoding cytokines may influence an individual’s ability to mount an inflammatory response. McCarron et al. (2002) reported differences in genotype frequencies between prostate cancer cases and donor samples for proinflammatory IL-8 (T -251 A) and anti-inflammatory IL-10 (G -1082 A) that were compatible with the differences in reported expression levels between the wild type and variant alleles, but no association for pro-inflammatory IL-1b (C -511 T) or proinflammatory TNF-a (G -308 A). Oh et al. (2000) noted differences in allele frequencies for TNF-a polymorphisms between cases and historical controls. Changes in the activity of proteins such as RNaseL, a ribonuclease activated in response to double-stranded RNA such as from viruses (Silverman, 2003), and macrophage scavenger receptor1, a cell surface scavenger receptor that facilitates macrophage uptake of a wide range of ligands such as bacteria and oxidized lipoproteins (Coller and Paulnock, 2001), may result in an inadequate ability to fight infection. Some linkage and associations studies suggest that sequence variants in RNASEL (Carpten et al., 2002; Casey et al., 2002; Rokman et al., 2002) and MSR-1 (Xu et al., 2002b; Wiklund et al., 2003; Xu et al., 2003) are associated with prostate cancer risk. Additional work is needed to understand the role of infection and response to infection in the etiology of prostate cancer.

Biotransformation Enzymes Glutathione S-Transferases The GSTs are a family of detoxification enzymes that catalyze conjugation of glutathione to electrophilic compounds, some of which may be mutagens or carcinogens. GSTs also act as binding proteins for many drugs and steroid hormones (Hayes et al., 1995a). GSTs may be modulators of prostatic carcinogenesis through elimination of carcinogens or by reducing the bioavailability of androgens. However, some of the products of the compounds metabolized by GSTs may themselves be reactive. Thus, the association between GST genotype and prostate cancer may depend on the nature of the exposures to which populations and individuals have contact. Because loss of GSTP1 expression appears to be an important player in prostate carcinogenesis, several groups have evaluated the role of polymorphisms in this gene in relation to prostate cancer risk. No association was found for the ATAAA repeat upstream of the GSTP1 promoter (Platz et al., 2002c), which has been speculated to influence GSTP1 promoter methylation (Millar et al., 2000). Results have not been consistent for the I105V polymorphism, which produces isozymes that have different activities depending on the substrate (Autrup et al., 1999; Shepard et al., 2000; Kote-Jarai et al., 2001). A third polymorphism in GSTP1 consisting of A313G (codon I105V) and C341T (A113V) polymorphisms (*A = 313A, 341C, *B = 313G, 341C, *C 313G, 341T), has also not been consistently associated with prostate cancer risk (Wadelius et al., 1999; Steinhoff et al., 2000; Gsur et al., 2001). Findings for the GSTT1 null genotype are inconsistent (Autrup et al., 1999; Rebbeck et al., 1999; Kelada et al., 2000; Steinhoff et al., 2000; Gsur et al., 2001; Kote-Jarai et al., 2001; Murata et al., 2001) and no asso-

1139

ciation has been reported for the GSTM1 null genotype (Autrup et al., 1999; Rebbeck et al., 1999; Kelada et al., 2000; Steinhoff et al., 2000; Gsur et al., 2001; Kote-Jarai et al., 2001; Murata et al., 2001). Overall, there is no consistent association between polymorphisms in GSTs and risk of prostate cancer.

N-acetyltransferases N-acetyltransferase-2 (NAT-2), which activates heterocyclic amines, including those in cooked meat and cigarette smoke, has not been consistently associated with prostate cancer (Wadelius et al., 1999; Hamasaki et al., 2003).

Other Genes and Their Expression DNA Repair The A11657G polymorphism in the hOGG1 gene, which encodes a DNA glycosylase/AP lyase that repairs 8OHdG, was associated with a higher risk of familial and sporadic prostate cancer (Xu et al., 2002a). Prostate cancer cases had lower expression of hMSH2 and hMLH1 (Strom et al., 2001) in one study, but lower PMS1 and PMS2 expression in another (Chen et al., 2001).

PTEN PTEN, which encodes a lipid phosphatase, is a tumor suppressor gene that is mutated in several prostate cancer cell lines (Li et al., 1997). No association between a polymorphic insertion of ACTAA in intron 4 and prostate cancer has been observed (George et al., 2001; Nathanson et al., 2001). The Met3326Ile polymorphism in PI3-K, which acts upstream in the PTEN pathway, was also not associated with prostate cancer (Paradis et al., 2003).

Ornithine Decarboxylase and Polyamines Seminal fluid is rich in the polyamines spermine and spermidine, which are required for cell proliferation and differentiation (Schipper et al., 2003). Ornithine decarboxylase is overexpressed in prostate cancer tissue (Mohan et al., 1999). Inhibiting polyamine synthesis by blocking ornithine decarboxylase or adding polyamine analogs results in reduced prostate cancer cell proliferation in vitro (Schipper et al., 2003). The wild type allele for the A321G substitution in this gene in combination with shorter androgen receptor gene CAG repeats has been associated with a statistically significant twofold higher risk of prostate cancer compared with those who were homozygous for the variant and had longer CAG repeats (Visvanathan et al., 2004). More work is needed on the influence of ODC on risk of prostate cancer.

PATHOGENESIS Models of the pathogenesis of prostate cancer have been described (Ross and Henderson, 1994; Giovannucci et al., 2001; Hsing and Devesa, 2001d; Nelson et al., 2001). Shown in Figure 59–1 is a view of the hypothesized pathogenesis of prostate cancer integrating broadly across purported environmental and host factors. Nelson et al., 2003; De Marzo et al., 1999 have proposed that infectious agents, hypoxia, autoimmunity, or other events causing cell damage precipitate an inflammatory response. During that response, the production of reactive oxygen and nitrogen species by phagocytic cells that move to the prostate acinus may be genotoxic and cytotoxic. Enhanced proliferation to replace damaged cells may increase the risk of fixation of mutations, some of which may lead to growth advantage. PIA lesions may be regenerative lesions, given that epithelial cells are highly proliferative and are enriched with a cell that has a phenotype intermediate between a basal cell and an epithelial cell (van Leenders et al., 2003). The loss of expression of GSTP1 in these already vulnerable cells renders them less able to detoxify environmental and endogenous (e.g., oxidant stress due to androgenic stimulation (Ripple et al., 1997), byproducts of oxidative phosphorylation) electrophiles and thus, increases the likelihood of transition to PIN. Oxidative damage from the above sources to the telomeres of proliferating intermediate phenotype cells may result in greater chromosomal instability (e.g.,

1140

PART IV: CANCER BY TISSUE OF ORIGIN

GENETIC VARIATION Sex steroid hormones, receptors, metabolic enzymes, co-activators Peptide growth factors Biotransformation enzymes, DNA repair enzymes Cytokines and response to infection

Maternal Western Diet

Infection Autoimmunity Hypoxia ↑ CHRONIC INFLAMMATION

Altered Hormonal Set Points

Western Diet and Lifestyle High Intake of Animal Products Energy Imbalance Low Intake of Antioxidant Vitamins and Minerals

ROS/RNS

Earlier Androgenic Stimulation

↑ PROLIFERAT ION ANTI-APOPTOSIS ↑ MUTAGENESIS

Developing Prostate In Utero

↑ ANDROGENIC STIMULATION ↑ GROWTH FACTOR STIMULATION

↑ OXIDATION

Developing Prostate At Puberty

Enhanced Proliferation (PIA?)

↑ ANGIOGENESIS

↑ PROLIFERAT ION

Prostatic Intraepithelial Neoplasia ↓ GSTP1

Pre-clinical Prostate Cancer

Detectable Pre-clinical Prostate Cancer

Distant Metastasis

Prostate Cancer Death

↓↓↓ GSTP1

Te lomere shortening

Chromosomal Rearrangements (e.g., gain in 7p, 7q, 8q, Xq, loss in 8p, 10q, 13q, 16q)

Gain of oncogenic, Loss of tumor suppressor activity

Figure 59–1. Integrated model of prostate carcinogenesis.

rearrangments) and gain of oncogenic activity or loss of tumor suppressor activity (Meeker et al., 2002). These early events may explain the similar prevalence of subclinical prostate cancers in high and low prostate cancer risk countries. Hsing and Devesa (2001d) have described the growth promotion of already initiated luminal cells via stimulation by hormonal (e.g., androgens) and growth factor (e.g., IGF-1, insulin) systems that become excessive because of Western dietary and lifestyle patterns (e.g., suboptimal diet, inactivity). Ross and Henderson (1994) have suggested that the effects of high animal fat diets may be influential in utero and in childhood, as well as in adulthood, by altering androgen set points, the duration of androgen exposure, along with higher androgen levels. These later-acting factors explain the difference in the incidence of clinically overt prostate cancer between high and low prostate cancer risk countries and may also explain the increased risk of prostate cancer in men who move from low to high prostate cancer risk countries. With continued genetic and epigenetic events (e.g., loss of 8p, perturbations in methylation at CpG sites), continued growth and metastasis advantage may occur. Superimposed on these models is normal variation in genes within and between populations, in particular in genes encoding detoxification enzymes (e.g., GSTs), hormonal systems components (e.g., androgen receptor, IGF-1 and IGFBP-3, vitamin D receptor, steroid metabolic enzymes), and inflammatory pathways, response to infection, and repair (e.g., cytokines, VEGF), which may modestly alter risk of prostate cancer or progression. Although success has been limited in pinpointing causal associations for polymorphisms and prostate cancer in part due to methodologic inadequacy (e.g., small sample size, allele frequencies in the controls not representative of the population that gave rise to the cases, ignoring the impact of PSA screening on case-mix), family history of prostate cancer may be capturing some of the risk associated with these genes with expected low penetrance.

PREVENTIVE MEASURES Primary Prevention At present there are no primary prevention strategies recommended to avoid prostate cancer. However, promising factors currently are being evaluated in chemoprevention trials.

Chemoprevention Two very large randomized controlled clinical trials were funded by the National Cancer Institute, one still ongoing, to evaluate chemopreventive agents for prostate cancer. The Prostate Cancer Prevention Study tested whether daily administration of finasteride (5 mg/day) for seven years decreased the risk of prostate cancer. Finasteride, a drug that inhibits the enzyme 5a-reductase type II, has been widely prescribed to decrease the lower urinary tract symptoms associated with benign prostatic hyperplasia. 18,882 healthy men aged 55 and older and who had a PSA concentration of £3 ng/mL participated (Thompson et al., 2003). The period prevalence of prostate cancer was reduced by 24.8% in the finasteride compared with the placebo arm, however the proportion of prostate cancers that were of high grade (Gleason sum ≥7) was higher in the finasteride arm than in the placebo arm. Studies are underway to understand whether the higher prevalence of high grade disease was causal or artifact. SELECT, a randomized, double-blinded, placebo-controlled trial is testing whether supplementation with selenium or vitamin E decreases the risk of prostate cancer (Klein et al., 2003). These two supplements were selected in part because of their apparent strong benefits on prostate cancer incidence in the Clark selenium trial and in the ATBC Trial, in which prostate cancer was a secondary (and not pre-stated) outcome. Selenium is being administered as a yeast-based selenomethionine and vitamin E as a-tocopherol. More than 35,000 men are participating.

Prostate Cancer

Screening and Early Detection American men have widely adopted screening for PSA. In 2001 75% of American man aged 50 years old or older reported that they had ever had a PSA test and 54% of men 50 to 69 years old reported that they recently had had a PSA test (Sirovich et al., 2003). Men aged 50 years or older who were less likely to have had a PSA test in the past year were those without a usual source of health care, those without health insurance, and recent immigrants (Swan et al., 2003). The American Cancer Society (American Cancer Society, 2003a) and the American Urological Association (American Urological Association, 2003) both recommend annual screening by digital rectal examination and the serum PSA test annually beginning at age 50 for men at normal risk of prostate cancer and earlier for men at higher than normal risk, including African-American men and men with a first-degree family history of prostate cancer. Screening is not recommended for men whose lifespan is estimated to be less than 10 years. The US Preventive Services Task Force has changed its recommendation from not recommending the PSA test to neither recommending for nor recommending against screening (US Preventive Services Task Force, 2002). The efficacy of the PSA test to detect prostate cancer at an earlier stage than digital rectal examination alone was shown in a large (nonrandomized) trial (Catalona et al., 1993). Using samples from the Physicians’ Health Study, Gann et al. (1995) showed that PSA concentrations of 2–3 ng/mL were associated with 5.5 times the risk of diagnosis of prostate cancer years later. Subsequently, in the Baltimore Longitudinal Study of Aging, Fang et al. (2001) showed that PSA concentrations above the median of 0.6 ng/mL in men 40–49.9 years old was associated with 3.75 times the risk of being diagnosed with prostate cancer. The usual trigger point for diagnostic work-up via prostate biopsy is an abnormal digital rectal examination and a serum PSA concentration of 4 ng/mL (or a PSA velocity (slope) of 0.75 ng/mL per year). This cut-off point is decreased to 2.5 ng/mL in men with a first-degree family history of prostate cancer. Some have recommended adjusting for prostate volume given that volume is likely proportional to the number of prostate epithelial cells that produce PSA. African-American men tend to have larger prostates and also have higher circulating concentrations of PSA. Older men also tend to have higher PSA levels because of benign prostate growth. Other factors that could transiently increase PSA and should be taken into account in determining whether a repeat PSA test should be done are prostatitis, recent sexual intercourse, and recent prostate examination. The sensitivity of the PSA test is roughly 67.5%–80% and the specificity is 60%–70% using >4 ng/mL as the cut-off point for an abnormal result (Carroll et al., 2001b). Because the test specificity is low, many men undergo prostate biopsy unnecessarily. Refined screening strategies are being implemented or are under development to reduce the false positive proportion, such as PSA density (ratio of PSA to prostate volume) or transition zone density to correct for higher PSA in men with larger prostates or larger transition zone of the prostate, PSA velocity, age-adjusted PSA, and the ratio of free to total PSA (Brawer, 1999). The lower the ratio of free PSA to total PSA the greater the likelihood of currently having or in the future having (Gann et al., 2002; Ito et al., 2003) prostate cancer in men with PSA levels of more than 4–10 ng/mL. The application of this ratio may also aid in better determining which men defined as high risk based on race and family history and who have PSAs between 2 and 4 ng/mL should undergo biopsy (Uzzo et al., 2003). PSA subtypes have also been identified. Proenzyme PSA (p-PSA) is the precursor form of PSA, which contains an extra seven amino acids, and is more indicative of prostate cancer than total PSA (Mikolajczyk et al., 2000). B-PSA is a truncated form of PSA that may be more indicative of benign prostatic hyperplasia (Linton et al., 2003). Differences in glycosylation patterns in PSA secreted by normal epithelium and tumor cells are also being evaluated for use in distinguishing between prostate cancer and noncancerous prostate conditions (Peracaula et al., 2003). The optimal timing and frequency of PSA screening remain to be determined. Biennial, rather than annual screening may be sufficient in men whose prior

1141

PSA level is low (<2 ng/mL) (Hugosson et al., 2003). Another recently recommended strategy is screening at a younger age (e.g., at age 40 years) at a time when PSA is less likely to be elevated because of benign prostatic hyperplasia and because of the slow growth of this tumor to screen less frequently (Carter, 2001). Despite the common use of PSA screening for prostate cancer, it is still unknown whether by earlier detection and treatment of prostate cancer, prostate cancer mortality will decrease. The prostate cancer death rate has declined in the United States at a time concurrent with the widespread adoption of PSA screening. This decline has also been seen in European countries in which screening with PSA has not been routinely adopted. A study comparing prostate cancer mortality rates in Tyrol, Austria, where a mass PSA screening program was implemented in 1993, with rates elsewhere in Austria, where PSA screening is not widespread, reported lower rates in Tyrol (Bartsch et al., 2001). To address the ultimate benefit of prostate cancer screening with PSA and/or digital rectal examination, two large-scale randomized trials, the Prostate, Lung, Colorectal, and Ovarian (PLCO) Trial in the United States (Gohagan et al., 2000) and the European Randomised Screening for Prostate Cancer (ERSPC) Trial (de Koning et al., 2002) are underway, with results expected in 2008. Testing for PSA is not risk free. Men who are false positives on screening or who have clinically insignificant prostate cancer (i.e., prostate cancer that would never impair quality or life or cause death) are subjected to unnecessary diagnostic work-up by biopsy and its possibly adverse effects including pain and uncontrolled bleeding. Men with clinically insignificant prostate cancer are subjected to unnecessary treatment by surgery, implanted radioactive seeds, or external beam radiation, and their possible adverse effects, including impotence and incontinence. One group estimated that during the first 10 years of the PSA era, 29% of whites and 44% of blacks with PSAscreen-detected prostate cancers were over-diagnosed; that is, cancers that would not have been otherwise detected during life (Etzioni et al., 2002). At this point it cannot be determined with certainty which cases are those that are truly clinically insignificant. Issues surrounding prostate cancer screening have been discussed in detail (Frankel et al., 2003). Because of these uncertainties in the risk-to-benefit ratio for prostate cancer screening, many groups strongly recommend that men be informed about these uncertainties before making a decision about screening.

FUTURE DIRECTIONS Etiology Interpretation of the epidemiologic literature on environmental and host factors for prostate cancer to devise a list of promising areas to pursue is not straightforward. For example, the role of factors such as obesity remains unclear, perhaps because BMI, the most commonly used marker of overweight/obesity, may be differently correlated with adiposity vs. lean mass depending on age at the time of assessment and nature of the population (e.g., Western vs. Asian). Also, factors such as obesity may influence multiple pathways, but in opposing directions for risk, such that on balance no association is observed with prostate cancer, except in certain subgroups where one pathway is favored over another. Even for factors that have been selected for rigorous testing in chemoprevention trials, such as selenium and vitamin E, findings from observational studies are not universally consistent when taken at face value. Some of the controversies in the literature on the etiology of prostate cancer arising from inconsistent findings may ultimately be resolved through wider appreciation of the subtleties that are unique to prostate cancer. An example of a prostatespecific issue results from the widespread use of PSA screening in the United States, which has enriched the pool with early cases. These formerly pre-clinical cases may be less related to factors that influence promotion and progression. The variability in the case mix between studies conducted in the pre-PSA and PSA eras and between current studies in countries with and without routine application of PSA screening now must be considered when evaluating which hypotheses are feasible to address in a study with a given set of case

1142

PART IV: CANCER BY TISSUE OF ORIGIN

characteristics, in drawing inferences from any one study, and when comparing findings among studies. The findings from studies of genetic variation in relation to prostate cancer have been particularly difficult to reconcile. The problems of changing case mix, small sample sizes leading to false negatives and false positives, and allele frequencies in the controls being dissimilar to those in the population that gave rise to the cases are well recognized. Difficulties in interpretation may also arise in genetic association studies of prostate cancer because typically only a single polymorphism is considered, often without knowledge of its functional consequence or it relation to other polymorphisms in that gene, or without consideration of other genes involved in the same pathway. Evaluation of haplotypes and diplotypes to complement alleles and genotypes is now in fashion, although the effectiveness of this approach remains to be demonstrated. Work should continue on the role of nutrients beyond those currently being studied in large chemoprevention trials, including lycopene, calcium, a-linolenic acid, and omega-3 fatty acids. Energy imbalance and imflammation are other areas that hold promise for primary prevention and merit focused research attention. The prevalence of overweight and obesity now exceeds 61% in the United States and obesity is a major risk factor for heart disease, diabetes, and several cancers. Epidemiologic research is needed immediately to clarify the relation of adiposity, energy intake, and physical activity and their correlates (e.g., hormonal and metabolic perturbations including insulin resistance and the IGF axis) to prostate cancer risk. Robust methods to measure these factors are essential as is evaluating the timing of exposure through life (e.g., weight gain in childhood, puberty, adulthood). The predominant hormonal and metabolic perturbations that result from energy imbalance may differ by age at diagnosis or other susceptibility factors. For example, for a young age at prostate cancer diagnosis, obesity may be protective via reducing testosterone levels, but may be less important at older ages at diagnosis, and instead free radical damage accumulated over the long term may predominate. Thus, these future studies must be powered for subgroup analyses or investigators should pool data as is being done for large cohorts for dietary and lifestyle factors and breast cancer (The Pooling Project) and for genetic studies of cancer (The Cohort Consortium). The role of chronic intraprostatic inflammation merits immediate attention because of the recent recognition of the potential significance of proliferative inflammatory atrophy lesions in the prostate and because inflammation is a known target for prevention and intervention. These future studies must take an integrated approach and consider both the causes (e.g., infection autoimmunity, hypoxia) and consequences (e.g., generation of reactive species, cytokines) of an inflammatory response. Easy-to-measure and reliable markers of intraprostatic inflammation are needed for these epidemiologic studies. It is unclear whether low levels of chronic intraprostatic inflammation (e.g., intraluminal macrophages) would elicit a systemic response (e.g., hepatic secretion of C-reactive protein) that would be detectable in circulation. Most of the existing studies on non-steroidal anti-inflammatory agents have reported on aspirin use. Many nonspecific and selective COX-2 inhibitors have been in use for the past several years and these should also be evaluated in relation to prostate cancer. To fully address the benefit of NSAIDs on prostate cancer risk, studies must collect detailed information on dose, duration of use, and type of agent used. Also, data are needed on the correlation between typically ingested doses of NSAIDs and the dose that reaches the prostate. Two other areas remain understudied relative to the importance of these problems—explanations for the markedly elevated risk of prostate cancer and prostate cancer death in African-American men and pre- and post-diagnostic risk factors for biochemical failure, progression, and death from prostate cancer. Large prospective multiethnic cohorts with long-term follow-up may be required to address the former problem. Questions that should be addressed include: 1) Whether the excess risk in African-American men is due to a greater prevalence of modifiable and inherent risk factors that are

common across racial/ethnic groups (e.g., RR is the same). For example, variation among racial/ethnic groups in length of the androgen receptor gene CAG repeat is already recognized (Edwards et al., 1992; Platz et al., 2000b), but has been estimated to account for only 15% of the higher incidence in AfricanAmerican men compared with white men (Platz et al., 2000b). 2) Whether variability in timing of exposure in life among racial/ ethnic groups account for differences in risk (e.g., onset of puberty). Measurement of exposures at several points in time in the future studies may be critical. 3) Whether the higher risk of prostate cancer in African Americans is due to a combination of risk factors, rather than a single factor. Approaches that integrate risk across several factors simultaneously may be helpful. Men diagnosed with organ-confined prostate cancer are typically considered cured after undergoing a radical prostatectomy. However, as many as 25% of these men experience a re-elevation in their PSA within months to years later, suggesting disease progression despite no longer having a prostate (Carroll, 2001a). It is believed that men who experience this biochemical failure had malignant cells escape before their prostatectomy. The inherent and modifiable risk factors for early escape of these tumor cells and the growth of these cells at regional or distant sites has not received much attention in the epidemiologic literature despite its importance for the health of men with prostate cancer. Related to this area is the evaluation of predictors of advanced disease at diagnosis and predictors of death from prostate cancer, given the same utilization of health resources, stage and grade at diagnosis, and treatment. Avenues of investigation should include the same factors under investigation for the incidence of prostate cancer, but in particular sources of oxidative damage, growth and metastasis factors (e.g., vitamin D, VEGF, integrins), factors related to the bone milieu, and genetic variation in genes that encode these factors.

Early Detection Work to improve the specificity of PSA and related molecules is ongoing. Other molecules that are early prostate tumor markers may eventually be detected in circulation using proteomics; that is, the generation of plasma protein spectra from mass spectrometry coupled with statistical algorithms to identify protein patterns that are unique to men with prostate cancer. The spectra or specific proteins that comprise the spectra could serve as a screening test. To complement pathological evaluation of prostate biopsies, cDNA expression studies may be useful in confirming the presence of prostate cancer by identifying proteins that are systematically over- or under-expressed relative to normal epithelium in prostate cancer or even earlier, in PIN. For example, several groups have observed that a-methylacylCoA racemase, which catalyzes b-oxidation of branched chain fatty acids, is consistently overexpressed in prostate cancers (Luo et al., 2002). References Adami HO, Bergström R, Engholm G, et al. 1996. A prospective study of smoking and risk of prostate cancer. Int J Cancer 67:764–768. Adami HO, McLaughlin J, Ekbom A, et al. 1991. Cancer risk in patients with diabetes mellitus. Cancer Causes Control 2:307–314. Adami HO, McLaughlin JK, Hsing AW, et al. 1992. Alcoholism and cancer risk: A population-based cohort study. Cancer Causes Control 3:419– 425. Adler HL, McCurdy MA, Kattan MW, Timme TL, Scardino PT, Thompson TC. 1999. Elevated levels of circulating interleukin-6 and transforming growth factor-b1 in patients with metastatic prostatic carcinoma. J Urol 161:182–187. Ahluwalia B, Jackson MA, Jones GW, Williams AO, Rao MS, Rajguru S. 1981. Blood hormone profiles in prostate cancer patients in high-risk and lowrisk populations. Cancer 48:2267–2273. Ahonen MH, Tenkanen L, Teppo L, Hakama M, Tuohimaa P. 2000. Prostate cancer risk and prediagnostic serum 25-hydroxyvitamin D levels (Finland). Cancer Causes Control 11:847–852.

Prostate Cancer Albanes D, Blair A, Taylor PR. 1989. Physical activity and risk of cancer in the NHANES I population. Am J Public Health 79:744–750. American Cancer Society. 2003a. Can Prostate Cancer Be Found Early? Available at: http://www.cancer.org/docroot/CRI/content/CRI_2_4_3X_ Can_prostate_cancer_be_found_early_36.asp?sitearea= American Cancer Society. 2003b. Cancer Facts & Figures, 2003. American Urological Association. Adult conditions: Prostate—Prostate Cancer Screening. Available at: http://www.urologyhealth.org/adult/index.cfm? cat=05&topic=250. Reviewed January, 2003. Amirimani B, Walker AH, Weber BL, Rebbeck TR. 1999. Re: Modification of clinical presentation of prostate tumors by a novel genetic variant in CYP3A4. J Natl Cancer Inst 91:1588–1590. Andersson SO, Baron J, Wolk A, Lindgren C, Bergström R, Adami HO. 1995. Early life risk factors for prostate cancer: A population-based case-control study in Sweden. Cancer Epidemiol Biomarkers Prev 4:187–192. Andersson SO, Wolk A, Bergström R, et al. 1997. Body size and prostate cancer: A 20-year follow-up study among 135,006 Swedish construction workers. J Natl Cancer Inst 89:385–389. Andersson SO, Wolk A, Bergstrom R, et al. 1996. Energy, nutrient intake and prostate cancer risk: A population-based case-control study in Sweden. Int J Cancer 68:716–722. Andò S, Rubens R, Rottiers R. 1984. Androgen plasma levels in male diabetics. J Endocrinol Invest 7:21–24. Armstrong B, Doll R. 1975. Environmental factors and cancer incidence and mortality in different countries, with special reference to dietary practices. Int J Cancer 15:617–631. Armstrong BG, Kazantzis G. 1985. Prostatic cancer and chronic respiratory and renal disease in British cadmium workers: A case control study. Br J Ind Med 42:540–545. Ascherio A, Stampfer MJ, Colditz GA, Rimm EB, Litin L, Willett WC. 1992. Correlations of vitamin A and E intakes with the plasma concentrations of carotenoids and tocopherols among American men and women. J Nutr 122:1792–1801. Augustsson K, Michaud DS, Rimm EB, et al. 2003. A prospective study of intake of fish and marine fatty acids and prostate cancer. Cancer Epidemiol Biomarkers Prev 12:64–67. Autrup JL, Thomassen LH, Olsen JH, Wolf H, Autrup H. 1999. Glutathione S-transferases as risk factors for prostate cancer. Eur J Cancer Prev 8:525–532. Bagnardi V, Mlangiardo M, La Vecchia C, Corrrao G. 2001. A meta-analysis of alcohol drinking and cancer risk. Br J Cancer 85:1700–1705. Baker LH, Mebust WK, Chin TD, Chapman AL, Hinthorn D, Towle D. 1981. The relationship of herpesvirus to carcinoma of the prostate. J Urol 125:370–374. Bartsch G, Horninger W, Klocker H, et al. 2001. Prostate cancer mortality after introduction of prostate-specific antigen mass screening in the Federal State of Tyrol, Austria. Urology 58:417–424. Beilin J, Harewood L, Frydenberg M, et al. 2001. A case-control study of the androgen receptor gene CAG repeat polymorphism in Australian prostate carcinoma subjects. Cancer 92:941–949. Bostwick DG. 1999. Prostatic intraepithelial neoplasia is a risk factor for cancer. Semin Urol Oncol 17:187–198. Bratt O, Borg A, Kristofferson U, Lundgren R, Zhang QX, Olsson H. 1999. CAG repeat length in the androgen receptor gene is related to age at diagnosis of prostate cancer and response to endocrine therapy, but not to prostate cancer risk. Br J Cancer 81:672–676. Braun MM, Helzlsouer KJ, Hollis BW, Comstock GW. 1995. Prostate cancer and prediagnostic levels of serum vitamin D metabolites (Maryland, United States). Cancer Causes Control 6:235–239. Brawer MK. 1999. Prostate-specific antigen: Current status. CA Cancer J Clin 49:264–281. Brooks JD, Metter EJ, Chan DW, et al. 2001. Plasma selenium level before diagnosis and the risk of prostate cancer development. J Urol 166: 2034–2038. Brooks JD, Weinstein M, Lin X, et al. 1998. CG island methylation changes near the GSTP1 gene in prostatic intraepithelial neoplasia. Cancer Epidemiol Biomarkers Prev 7:531–536. Brownson RC, Chang JC, Davis JR, Smith CA. 1991. Physical activity on the job and cancer in Missouri. Am J Public Health 81:639–642. Bucher C, Jordan P, Nickeleit V, Torhorst J, Mihatsch MJ. 1999. Relative risk of malignant tumors in analgesic abusers. Effects of long-term intake of aspirin. Clin Nephrol 51:67–72. Calle EE, Rodriguez C, Walker-Thurmond K, Thun MJ. 2003. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med 348:1625–1638. Cannon L, Bishop DT, Skolnick M, Hunt S, Lyon JL, Smart CR. 1982. Genetic epidemiology of prostate cancer in Utah Mormon genealogy. Cancer Surv 1:47–69.

1143

Carpten J, Nupponen N, Isaacs S, et al. 2002. Germline mutations in the ribonuclease L gene in families showing linkage with HPC1. Nat Genet 30:181–184. Carroll P. 2001a. Rising PSA after a radical treatment. Eur Urol 40:9–16. Carroll P, Coley C, McLeod D, et al. 2001b. Prostate-specific antigen best practice policy—Part I: Early detection and diagnosis of prostate cancer. Urology 57:217–224. Carter BS, Beaty TH, Steinberg GD, Childs B, Walsh PC. 1992. Mendelian inheritance of familial prostate cancer. Proc Natl Acad Sci U S A 89:3367–3371. Carter HB. 2001. Rationale for earlier and less frequent prostate cancer screening. Urology 58:639–641. Casey G, Neville PJ, Plummer SJ, et al. 2002. RNASEL Arg462Gln variant is implicated in up to 13% of prostate cancer cases. Nat Genet 32:1–2. Catalona WJ, Smith DS, Ratliff TL, Basler JW. 1993. Detection of organconfined prostate cancer is increased through prostate-specific antigenbased screening. JAMA 270:948–954. Cerhan JR, Torner JC, Lynch CF, et al. 1997. Association of smoking, body mass, and physical activity with risk of prostate cancer in the Iowa 65+ Rural Health Study (United States). Cancer Causes Control 8:229–238. Chamberlain NL, Driver ED, Miesfeld RL. 1994. The length and location of CAG trinucleotide repeats in the androgen receptor N-terminal domain affect transactivation function. Nucleic Acids Res 22:3181–3186. Chan JM, Giovannucci E, Andersson SO, Yuen J, Adami H-O, Wolk A. 1998a. Dairy products, calcium, phosphorus, vitamin D, and risk of prostate cancer (Sweden). Cancer Causes Control 9:559–566. Chen JM, Giovannucci EL. 2001a. Vegetables, fruits, associated micronutrients, and risk of prostate cancer. Epidemiol Rev 23:82–86. Chan JM, Pietinen P, Virtanen M, Malila N, Tangrea J. 2000. Diet and prostate cancer risk in a cohort of smokers, with a specific focus on calcium and phosphorus (Finland). Cancer Causes Control 11:859–867. Chan JM, Stampfer MJ, Giovannucci E, et al. 1998b. Plasma insulin-like growth factor-I and prostate cancer risk: A prospective study. Science 279:563–566. Chan JM, Stampfer MJ, Ma J, et al. 2002. Insulin-like growth factor-I (IGF-1) and IGF binding protein-3 as predictors of advanced-stage prostate cancer. J Natl Cancer Inst 94:1099–1109. Chan JM, Stampfer MJ, Ma J, Gann PH, Gaziano JM, Giovannucci EL. 2001b. Dairy products, calcium, and prostate cancer risk in the Physicians’ Health Study. Am J Clin Nutr 74:549–554. Chan JM, Stampfer MJ, Ma J, Rimm EB, Willett WC, Giovannucci EL. 1999. Supplemental vitamin E intake and prostate cancer risk in a large cohort of men in the United States. Cancer Epidemiol Biomarkers Prev 8:893–899. Chang BL, Zheng SL, Hawkins GA, et al. 2002. Joint effect of HSD3B1 and HSD3B2 genes is associated with hereditary and sporadic prostate cancer susceptibility. Cancer Res 62:1784–1789. Chang S, Hursting SD, Contois JH, et al. 2001. Leptin and prostate cancer. Prostate 46:62–67. Charles LE, Loomis D, Shy CM, et al. 2003. Electromagnetic fields, polychlorinated biphenyls and prostate cancer mortality in electric utility workers. Am J Epidemiol 157:683–691. Chen C, Lamharzi N, Weiss NS, et al. 2002. Androgen receptor polymorphisms and the incidence of prostate cancer. Cancer Epidemiol Biomarkers Prev 11:1033–1040. Chen Y, Wang J, Fraig MM, et al. 2001. Defects of DNA mismatch repair in human prostate cancer. Cancer Res 61:4112–4121. Chokkalingam A, Pollak M, Fillmore C, et al. 2001a. Insulin-like growth factors and prostate cancer: A Population-based case-control study in China. Cancer Epidemiol Biomarkers Prev 10:421–427. Chokkalingam AP, McGlynn KA, Gao YT, et al. 2001b. Vitamin D receptor gene polymorphisms, insulin-like growth factors, and prostate cancer risk: A population-based case-control study in China. Cancer Res 61: 4333–4336. Chow CK. 1979. Nutritional influence on cellular antioxidant defense systems. Am J Clin Nutr 32:1066–1081. Chu KC, Tarone RE, Freeman HP. 2003. Trends in prostate cancer mortality among black and white men in the United States. Cancer 97:1507–1516. Chu NF, Spiegelman D, Yu J, Rifai N, Hotamisligil GS, Rimm EB. 2001. Plasma leptin concentrations and four year weight gain among US men. Int J Obes 25:346–353. Clark LC, Combs Jr., GF, Turnbull BW, et al. 1996. Effects of selenium supplementation for cancer prevention in patients with carcinoma of the skin. A randomized controlled trial. Nutritional Prevention of Cancer Study Group. JAMA 276:1957–1963. Clark LC, Dalkin B, Krongrad A, et al. 1998. Dcreased incidence of prostate cancer with selenium supplementation: Results of a double-blind cancer prevention trial. Br J Urol 81:730–734.

1144

PART IV: CANCER BY TISSUE OF ORIGIN

Clarke G, Whittemore AS. 2000. Prostate cancer risk in relation to anthropometry and physical activity: The National Health and Nutritional Examination Survey I Epidemiological Follow-Up Study. Cancer Epidemiol Biomarkers Prev 9:875–881. Clinton SK, Emenhiser C, Schwartz SJ, et al. 1996. cis-trans lycopene isomers, carotenoids, and retinol in the human prostate. Cancer Epidemiol Biomarkers Prev 5:823–833. Cohen JH, Kristal AR, Stanford JL. 2000. Fruit and vegetable intakes and prostate cancer risk. J Natl Cancer Inst 92:61–68. Cohen P, Peehl DM, Lamson G, Rosenfeld RG. 1991. Insulin-like growth factors (IGFs), IGF receptors, and IGF-binding proteins in primary cultures of prostate epithelial cells. J Clin Endocrinol Metab 73:401– 407. Cohen P, Peehl DM, Rosenfeld RG. 1994. The IGF axis in the prostate. Hormone Metab Res 26:81–84. Coller S, Paulnock D. 2001. Signaling pathways initiated in macrophages after engagement of type A scavenger receptors. J Leukoc Biol 70:142–148. Combs GF, Jr., Combs SB. 1984. The nutritional biochemistry of selenium. Annu Rev Nutr 4:257–280. Cook NR, Lee IM, Manson JE, Buring JE, Hennekens CH. 2000. Effects of beta-carotene supplementation on cancer incidence by baseline characteristics in the physicians’ Health Study (United States). Cancer Causes Control 11:617–626. Cookson MS, Reuter VE, Linkov I, Fair WR. 1997. Glutathione S-transferase pi (GST-p) class expression by immunohistochemistry in benign and malignant prostate tissue. J Urol 157:673–676. Corder EH, Guess HA, Hulka BS, et al. 1993. Vitamin D and prostate cancer: A prediagnostic study with stored sera. Cancer Epidemiol Biomarkers Prev 2:467–472. Correa-Cerro L, Berthon P, Haussler J, et al. 1999a. Vitamin D receptor polymorphisms as markers in prostate cancer. Hum Genet 105:281–287. Correa-Cerro L, Wohr G, Haussler J, et al. 1999b. (CAG)nCAA and GGN repeats in the human androgen receptor gene are not associated with prostate cancer in a French-German population. Eur J Hum Genet 7:357–362. Coughlin SS, Neaton JD, Sengupta A. 1996. Cigarette smoking as a predictor of death from prostate cancer in 348,874 men screened for the Multiple Risk Factor Intervention Trial. Am J Epidemiol 143:1002–1006. Dai WS, Gutai JP, Kuller LH, Cauley JA. 1988. Cigarette smoking and serum sex hormones in men. Am J Epidemiol 128:796–805. Daniell HW. 1995. A worse prognosis for smokers with prostate cancer. J Urol 154:153–157. Daviglus ML, Dyer AR, Persky V, et al. 1996. Dietary beta-carotene, vitamin C, and risk of prostate cancer: Results from the Western Electric Study. Epidemiology 7:472–477. Davis DL, Russell DW. 1993. Unusual length polymorphism in human steroid 5a-reductase type 2 gene (SRD5A2). Human Mol Genetics 2:820. de Koning HJ, Liem MK, Bann CA, Boer R, Schroder FH, Alexander FE. 2002. Prostate cancer mortality reduction by screening: Power and time frame with complete enrollment in the European Randomised Screening for Prostate Cancer (ERSPC) trial. Int J Cancer 98:268–273. De Marzo AM, Marchi VL, Epstein JI, Nelson WG. 1999. Proliferative inflammatory atrophy of the prostate: Implications for prostatic carcinogenesis. Am J Pathol 155:1985–1992. De Marzo AM, Nelson WG, Isaacs WB, Epstein JI. 2003. Pathological and molecular aspects of prostate cancer. Lancet 361:955–964. Demark-Wahnefried W, Schildkraut JM, Thompson D, et al. 2000. Early onset baldness and prostate cancer risk. Cancer Epidemiol Biomarkers Prev 9:325–328. Dennis L. 2000. Meta-analysis for combining relative risks of alcohol consumption and prostate cancer. Prostate 42:56-66. Dennis LK, Dawson DV, Resnick MI. 2002a. Vasectomy and the risk of prostate cancer: A meta-analysis examining vasectomy status, age at vasectomy, and time since vasectomy. Prostate Cancer Prostatic Dis 5:193–203. Dennis LK, Lynch CF, Torner JC. 2002b. Epidemiologic association between prostatitis and prostate cancer. Urology 60:78–83. Dennis LK, Dawson DV. 2002c. Meta-analysis of measures of sexual activity and prostate cancer. Epidemiology 13:72–79. Devgan SA, Henderson BE, Yu MC, et al. 1997. Genetic variation of 3 bhydroxysteroid dehydrogenase type II in three racial/ethnic groups: Implications for prostate cancer risk. Prostate 33:9–12. Dillner J, Knekt P, Boman J, et al. 1998. Sero-epidemiological association between human-papillomavirus infection and risk of prostate cancer. Int J Cancer 75:564–567. Dorgan JF, Albanes D, Virtamo J, et al. 1998. Relationships of serum androgens and estrogens to prostate cancer risk: Results from a prospective study in Finland. Cancer Epidemiol Biomarkers Prev 7:1069–1074.

Drachenberg DE, Elgamal AA, Rowbotham R, Peterson M, Murphy GP. 1999. Circulating levels of interleukin-6 in patients with hormone refractory prostate cancer. Prostate 41:127–133. Eaton NE, Reeves GK, Appleby PN, Key TJ. 1999. Endogenous sex hormones and prostate cancer: A quantitative review of prospective studies. Br J Cancer 80:930–934. Edwards A, Hammond HA, Jin L, Caskey CT, Chakraborty R. 1992. Genetic variation at five trimeric and tetrameric tandem repeat loci in four human population groups. Genomics 12:241–253. Edwards SM, Badzioch MD, Minter R, Hamoud R, Collins N, Ardern-Jones A. 1999. Androgen receptor polymorphisms: Association with prostate cancer risk, relapse and overall survival. Int J Cancer 84:458–465. Eichholzer M, Stahelin HB, Gey KF, Ludin E, Bernasconi F. 1996. Prediction of male cancer mortality by plasma levels of interacting vitamins: 17-year follow-up of the prospective Basel study. Int J Cancer 66:145–150. Ekbom A, Hsieh CC, Lipworth L, Wolk A, Pontén J, et al. 1996. Perinatal characteristics in relation to incidence of and mortality from prostate cancer. Br Med J 313:337–341. Elghany NA, Schumacher MC, Slattery ML, West DW, Lee JS. 1990. Occupation, cadmium exposure, and prostate cancer. Epidemiology 1:107– 115. Elinder CG, Kjellström T, Hogstedt C, Andersson K, Spång G. 1985. Cancer mortality of cadmium, workers. Br J Ind Med 42:651–655. Ellison LF. 2000. Tea and other beverage consumption and prostate cancer risk: A Canadian retrospective cohort study. Eur J Cancer Prev 9:125–130. Elo JP, Visakorpi T. 2001. Molecular genetics of prostsate cancer. Ann Med 33:130–141. Etzioni R, Penson DF, Legler JM, et al. 2002. Overdiagnosis due to prostatespecific antigen screening: Lessons from U.S. prostate cancer incidence trends. J Natl Cancer Inst 94:958–960. Faber PW, Kuiper GGJM, van Rooij HCJ, van der Korput JAGM, Brinkmann AO, Trapman J. 1989. The N-terminal domain of the human androgen receptor is encoded by one, large exon. Mol Cell Endocrinol 61:257–262. Fang J, Metter EJ, Landis P, Chan DW, Morrell CH, Carter HB. 2001. Low levels of prostate-specific antigen predict long-term risk of prostate cancer: Results from the Baltimore Longitudinal Study of Aging. Urology 58:411– 416. Febbo PG, Kantoff PW, Platz EA, et al. 1999. The V89L polymorphism in the 5a-reductase type 2 gene and risk of prostate cancer. Cancer Res 59:5878–5881. Feigl P, Blumenstein B, Thompson I, et al. 1995. Design of the Prostate Cancer Prevention Trial (PCPT). Controlled Clinical Trials 16:150–163. Field AE, Colditz GA, Willett WC, Longcope C, McKinlay JB. 1994. The relation of smoking, age, relative weight, and dietary intake to serum adrenal steroids, sex hormones, and sex hormone-binding globulin in middleaged men. J Clin Endocrinol Metab 79:1310–1316. Frankel S, Smith GD, Donovan J, Neal D. 2003. Screening for prostate cancer. Lancet 361:1122–1128. Friedman JM. 2002. The function of leptin in nutrition, weight, and physiology. Nutr Rev 60:S1–S14. Gann PH, Hennekens CH, Ma J, Longcope C, Stampfer MJ. 1996a. Prospective study of sex hormone levels and risk of prostate cancer. J Natl Cancer Inst 88:1118–1126. Gann PH, Hennekens CH, Sacks FM, Grodstein F, Giovannucci EL, Stampfer MJ. 1994. Prospective study of plasma fatty acids and risk of prostate cancer. J Natl Cancer Inst 86:281–286. Gann PH, Hennerkens CH, Stampfer MJ. 1995. A prospective evaluation of plasma prostate-specific antigen for detection of prostate cancer. JAMA 273:289–294. Gann PH, Ma J, Catalona WJ, Stampfer MJ. 2002. Strategies combining total and percent free prostate specific antigen for detecting prostate cancer: A prospective evaluation. J Urol 167:2427–2434. Gann PH, Ma J, Giovannucci E, Willett W, Sacks F, Hennekens CH, Stampfer MJ. 1999. Lower prostate cancer risk in men with elevated plasma lycopene levels: Results of a prospective analysis. Cancer Res 59:1225– 1230. Gann PH, Ma J, Hennekens CH, Hollis BW, Haddad JG, Stampfer MJ. 1996b. Circulating vitamin D metabolites in relation to subsequent development of prostate cancer. Cancer Epidemiol Biomarkers Prev 5:121–126. Gapstur SM, Gann PH, Colangelo LA, et al. 2001. Postload glucose concentration and 27-year prostate cancer mortality (United States). Cancer Causes Control 12:763–772. George DJ, Shepard TF, Ma J, Giovannucci E, Kantoff PW, Stampfer MJ. 2001. PTEN polymorphism (IVS4) is not associated with risk of prostate cancer. Cancer Epidemiol Biomarkers Prev 10:411– 412. Giles GG, Severi G, Sinclair R, et al. 2002. Androgenetic alopecia and prostate cancer: Findings from an Australian case-control study. Cancer Epidemiol Biomarkers Prev 11:549–553.

Prostate Cancer Giovannucci E. 2002a. Is the androgen receptor CAG repeat length significant for prostate cancer? Cancer Epidemiol Biomarkers Prev 11:985–986. Giovannucci E, Ascherio A, Rimm EB, Stampfer MJ, Colditz GA, Willett WC. 1995. Intake of carotenoids and retinol in relation to risk of prostate cancer. J Natl Cancer Inst 87:1767–1776. Giovannucci E, Leitzmann M, Spiegelman D, et al. 1998a. A prospective study of physical activity and prostate cancer in male health professionals. Cancer Res 58:5117–5122. Giovannucci E, Platz EA. 2001. Nutritional and environmental epidemiology of prostate cancer. In: Kantoff PW, Carroll P, D’Amico AV, Isaacs J, Ross R, Sher H, eds. Prostate Cancer: Principles and Practice. Philadelphia: Lippincott, Williams & Wilkins, pp. 19–45. Giovannucci E, Rimm EB, Ascherio A, et al. 1999. Smoking and risk of total and fatal prostate cancer in United States health professionals. Cancer Epidemiol Biomarkers Prev 8:277–282. Giovannucci E, Rimm EB, Colditz GA, et al. 1993. A prospective study of dietary fat and risk of prostate cancer. J Natl Cancer Inst 85: 1571–1579. Giovannucci E, Rimm EB, Liu Y, et al. 2003. Body mass index and risk of prostate cancer in U.S. Health professionals. J Natl Cancer Inst 95: 1240–1244. Giovannucci E, Rimm EB, Liu Y, Stampfer MJ, Willett WC. 2002b. A prospective study of tomato products, lycopene, and prostate cancer risk. J Natl Cancer Inst 94:391–398. Giovannucci E, Rimm EB, Stampfer MJ, Colditz GA, Willett WC. 1997a. Height, body weight, and risk of prostate cancer. Cancer Epidemiol Biomarkers Prev 6:557–563. Giovannucci E, Rimm EB, Stampfer MJ, Colditz GA, Willett WC. 1998b. Diabetes mellitus and risk of prostate cancer (United States). Cancer Causes Control 9:3–9. Giovannucci E, Rimm EB, Wolk A, et al. 1998c. Calcium and fructose intake in relation to risk of prostate cancer. Cancer Res 58:442– 447. Giovannucci E, Stampfer MJ, Krithivas K, et al. 1997b. The CAG repeat within the androgen receptor gene and its relationship to prostate cancer. Proc Natl Acad Sci U S A 94:3320–3323. Gleason DF, Mellinger GT, Veterans Administration Cooperative Urological Research Group. 1974. Prediction of prognosis for prostatic adenocarcinoma by combined histological grading and clinical staging. 1974. J Urol 111:58–64. Glover Jr., FE, Coffey DS, Douglas LL, et al. 1998. The epidemiology of prostate cancer in Jamaica. J Urol 159:1984–1986. Godfrey KM, Barker DJP. 1995. Maternal nutrition in relation ot fetal and placental growth. Eur J Obstet Gynecol Reproductive Biol 61:15– 22. Godley PA, Campbell MK, Gallagher P, Martinson FEA, Mohler JL, Sandler RS. 1996. Biomarkers of essential fatty acid consumption and risk of prostatic carcinoa. Cancer Epidemiol Biomarkers Prev 5:889–895. Gohagan JK, Prorok PC, Hayes RB, Kramer BS. 2000. The Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial of the National Cancer Institute: History, organization, and status. Control Clin Trials 21(6 Suppl):251S–272S. Goldgar DE, Easton DF, Cannon-Albright LA, Skolnick MH. 1994. Systematic population-based assessment of cancer risk in first-degree relatives of cancer probands. J Natl Cancer Inst 86:1600–1608. Gong G, Oakley-Girvan I, Wu AH, et al. 2002. Segregation analysis of prostate cancer in 1,719 white, African-American and Asian-American families in the United States and Canada. Cancer Causes Control 13:471– 482. Graham S, Haughey B, Marshall J, et al. 1983. Diet in the epidemiology of carcinoma of the prostate gland. J Natl Cancer Inst 70:687–692. Green JRB. 1997. Mechanism of hypogonadism in cirrhotic males. Gut 18: 843–853. Griffiths TRL, Mellon JK. 2000. Human papillomavirus and urological tumours: II. Role in bladder, prostate, renal and testicular cancer. Br J Urol 85:211–217. Grinstein S, Preciado MV, Gattuso P, et al. 2002. Demonstration of EpsteinBarr virus in carcinomas of various sites. Cancer Res 62:4876– 4878. Groff JL, Gropper SS. 2000. Advanced Nutrition and Human Metabolism. Belmont, CA: Wadsworth. Grönberg H, Damber L, Damber JE. 1996. Familial prostate cancer in Sweden. A nationwide register cohort study. Cancer 77:138–143. Grönberg H, Damber L, Damber JE, Iselius L. 1997. Segregation analysis of prostate cancer in Sweden: Support for dominant inheritance. Am J Epidemiol 146:552–557. Gsur A, Haidinger G, Hinteregger S, et al. 2001. Polymorphisms of glutathioneS-transferase genes (GSTP1, GSTM1 and GSTT1) and prostate-cancer risk. Int J Cancer 95:152–155. Gupta S, Ahmad N, Mohan RR, Husain MM, Mukhtar H. 1999. Prostate cancer chemoprevention by green tea: In vitro and in vivo inhibition of testos-

1145

terone-mediated induction of ornithine decarboxylase. Cancer Res 59: 2115–2120. Habel LA, Zhao W, Stanford JL. 2002. Daily aspirin use and prostate cancer risk in a large, multiracial cohort in the US. Cancer Causes Control 13:427– 434. Habuchi T, Suzuki T, Sasaki R, et al. 2000. Association of vitamin D receptor gene polymorphism with prostate cancer and benign prostatic hyperplasia in a Japanese population. Cancer Res 60:305–308. Hackney AC. 1996. The male reproductive system and endurance exercise. Med Sci Sports Exerc 28:180–189. Haenszel W, Kurihara M. 1968. Studies of Japanese migrants. I. Mortality from cancer and other diseases among Japanese in the United States. J Natl Cancer Inst 40:43–68. Hamasaki T, Inatomi H, Katoh T, et al. 2003. N-acetyltransferase-2 gene polymorphism is a possible biomarker for prostate cancer in Japanese men. Int J Urol 10:167–173. Harman SM, Metter EJ, Blackman MR, Landis PK, Carter HB. 2000. Serum levels of insulin-like growth factor I (IGF-I), IGF-II, IGF-binding protein3, and prostate-specific antigen as predictors of clinical prostate cancer. J Clin Endocrinol Metab 85:4258– 4265. Hartman TJ, Albanes D, Pietinen P, et al. 1998a. The association between baseline vitamin E, selenium, and prostate cancer in the the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study. Cancer Epidemiol Biomarkers Prev 7:335–340. Hartman TJ, Albanes D, Rautalahti M, et al. 1998b. Physical activity and prostate cancer in the Alpha-Tocopherol, Beta-Carotene (ATBC) Cancer Prevention Study (Finland). Cancer Causes Control 9:11–18. Harvei S, Bjerve KS, Tretli S, Jellum E, Robsahm TE, Vatten L. 1997. Prediagnostic level of fatty acids in serum phospholipids: Omega-3 and omega-6 fatty acids and the risk of prostate cancer. Int J Cancer 71:545–551. Hawk E, Breslow RA, Graubard BI. 2000. Male pattern baldness and clinical prostate cancer in the epidemiologic follow-up of the first National Health and Nutrition Examination Survey. Cancer Epidemiol Biomarkers Prev 9:523–527. Hawkins GA, Cramer SD, Zheng SL, et al. 2002. Sequence variants in the human 25-hydroxyvitamin D3 la-hydroxylase (CYP27B1) gene are not associated with prostate cancer risk. Prostate 53:175–178. Hayes JD, Pulford DJ. 1995a. The glutathione S-transferase supergene family: Regulation of GST and the contribution of the isoenzymes to cancer chemoprotection and drug resistance. Crit Rev Biochem Mol Biol 30:445–600. Hayes RB, Brown LM, Schoenberg JB, et al. 1996. Alcohol use and prostate cancer risk in US blacks and whites. Am J Epidemiol 143:692–697. Hayes RB, Liff JM, Pottern LM, et al. 1995b. Prostate cancer risk in U.S. blacks and whites with a family history of cancer. Int J Cancer 60: 361–364. Hayes RB, Pottern LM, Strickler H, et al. 2000, Sexual behaviour, STDs and risks for prostate cancer. Br J Cancer 82:718–725. Hayes RB, Ziegler RG, Gridley G, et al. 1999. Dietary factors and risks for prostate cancer among blacks and whites in the United States. Cancer Epidemiol Biomarkers Prev 8:25–34. Heinonen OP, Albanes D, Virtamo J, et al. 1998. Prostate cancer and supplementation with a-tocopherol and b-carotene: Incidence and mortality in a controlled trial. J Natl Cancer Inst 90:440–446. Helzlsouer KJ, Huang H-Y, Alberg AJ, et al. 2000. Association between atocopherol, g-tocopherol, selenium and subsequent prostate cancer. J Natl Cancer Inst 92:2018–2023. Herbert JT, Birkhoff JD, Feorino PM, Caldwell GG. 1976. Herpes simplex virus type 2 and cancer of the prostate. J Urol 116:611–612. Hickey K, Do KA, Green A. 2001. Smoking and prostate cancer. Epidemiol Rev 23:115–125. Hisada M, Rabkin CS, Strickler HD, Wright WE, Christianson RE, van den Berg BJ. 2000. Human papillomavirus antibody and risk of prostate cancer. JAMA 2000:340–341. Hoffman RM, Gilliland FD, Eley JW, et al. 2001. Racial and ethnic differences in advanced-stage prostate cancer: The prostate cancer outcomes study. J Natl Cancer Inst 93:388–395. Howards SS, Peterson HB. 1993. Vasectomy and prostate cancer: Chance, bias, or a causal relationship. J Am Med Assoc 269:913–914. Hsieh C-C, Thanos A, Mitropoulos D, Deliveliotis C, Mantzoros CS, Trichopoulos D. 1999. Risk factors for prostate cancer: A case-control study in Greece. Int J Cancer 80:699–703. Hsing AW. 2001a. Hormones and prostate cancer: What’s next? Epidemiol Rev 23:42–58. Hsing AW, Chen C, Chokkalingam AP, et al. 2001b. Polymorphic markers in SRD5A2 gene and prostate cancer risk: A population-based case-control study. Cancer Epidemiol Biomarkers Prev 10:1077–1082.

1146

PART IV: CANCER BY TISSUE OF ORIGIN

Hsing AW, Chokkalingam AP, Gao YT, et al. 2002. Allium vegetables and risk of prostate cancer: a population-based study. J Natl Cancer Inst 94: 1648–1651. Hsing AW, Chua S, Gao YT, et al. 2001c. Prostate cancer risk and serum levels of insulin and leptin: A population-based study. J Natl Cancer Inst 93:783–789. Hsing AW, Comstock GW. 1993. Serological precursors of cancer: Serum hormones and risk of subsequent prostate cancer. Cancer Epidemiol Biomarkers Prev 2:27–32. Hsing AW, Comstock GW, Abbey H, Polk BF. 1990a. Serologic precursors of cancer. Retinol, carotenoids, and tocopherol and risk of prostate cancer. J Natl Cancer Inst 82:941–946. Hsing AW, Deng J, Sesterhenn IA, et al. 2000a. Body size and prostate cancer: A population-based case-control study in China. Cancer Epidemiol Biomarkers Prev 9:1335–1341. Hsing AW, Devesa SS. 2001d. Trends and patterns of prostate cancer: What do they suggest? Epidemiol Rev 23:3–13. Hsing AW, Gao YT, Chua S, Deng J, Stanczyk FZ. 2003. Insulin resistance and prostate cancer risk. J Natl Cancer Inst 95:67–71. Hsing AW, Gao YT, Wu G, et al. 2000b. Polymorphic CAG and GGN repeat lengths in the androgen receptor gene and prostate cancer risk: A population-based case-control study in China. Cancer Res 60:5111–5116. Hsing AW, McLaughlin JK, Schuman LM, et al. 1990b. Diet, tobacco use, and fatal prostate cancer: Results from the Lutheran Brotherhood Cohort Study. Cancer Res 50:6836–6840. Hsing AW, McLaughlin JK, Zheng W, Gao YT, Blot WJ. 1994. Occupation, physical activity, and risk of prostate cancers in Shanghai, People’s Republic of China. Cancer Causes Control 5:136–140. Huang HY, Alberg AJ, Norkus EP, Hoffman SC, Comstock GW, Helzlsouer KJ. 2003. Prospective study of antioxidant micronutrients in the blood and the risk of developing prostate cancer. Am J Epidemiol 157:335– 344. Hubbard JS, Rohrmann S, Landis PK, et al. 2004. Association of prostate cancer risk with insulin, glucose, and anthropometry in the Baltimore Longitudinal Study of Aging. Urology 63:253–258. Hugosson J, Aus G, Lilja H, Lodding P, Pihl CG, Pileblad E. 2003. Prostate specific antigen based biennial screening is sufficient to detect almost all prostate cancers while still curable. J Urol 169:1720–1723. Hussain T, Gupta S, Mukhtar H. 2003. Cyclooxygenase-2 and prostate carcinogenesis. Cancer Lett 191:125–135. Ingles SA, Coetzee GA, Ross RK, et al. 1998. Association of prostate cancer with vitamin D receptor haplotypes in African-Americans. Cancer Res 58:1620–1623. Ingles SA, Ross RK, Yu MC, et al. 1997. Association of prostate cancer risk with genetic polymorphisms in vitamin D receptor and androgen receptor. J Natl Cancer Inst 89:166–170. Irani J, Ravery V, Pariente JL, et al. 2002. Effect of nonsteroidal antiinflammatory agents and finasteride on prostate cancer risk. J Urol 168: 1985–1988. Irvine RA, Yu MC, Ross RK, Coetzee GA. 1995. The CAG and GGC microsatellites of the androgen receptor gene are in linkage disequilibrium in men with prostate cancer. Cancer Res 55:1937–1940. Ito K, Yamamoto T, Ohi M, Kurokawa K, Suzuki K, Yamanaka H. 2003. Free/total PSA ratio is a powerful predictor of future prostate cancer morbidity in men with initial PSA levels of 4.1 to 10.0 ng/mL. Urology 61:760–764. Jackson MA, Kovi J, Heshmat MY, et al. 1980. Characterization of prostatic carcinoma among blacks: A comparison between a low-incidence area, Ibadan, Nigeria, and a high-incidence area, Washington, DC. Prostate 1:185–205. Jaffe JM, Malkowicz SB, Walker AH, et al. 2000. Association of SRD5A2 genotype and pathological characteristics of prostate tumors. Cancer Res 60:1626–1630. Jänne OA, Palvimo JJ, Kallio P, Mehto M. 1993. Androgen receptor and mechanisms of androgen action. Ann Med 25:83–89. Jemal A, Murray T, Samuels A, Ghafoor A, Ward E, Thun MJ. 2003. Cancer Statistics, 2003. CA Cancer J Clin Med 53:5–26. Kantoff PW, Febbo PG, Giovannucci E, et al. 1997. A polymorphism of the 5a-reductase gene and its association with prostate cancer: A case-control analysis. Cancer Epidemiol Biomarkers Prev 6:189–192. Kaplan LA, Stein EA, Willett WC, Stampfer MJ, Stryker WS. 1987. Reference ranges of retinol, tocopherols, lycopene and alpha- and beta-carotene in plasma by simultaneous high-performance liquid chromatographic analysis. Clin Physiol Biochem 5:297–304. Kazemi-Esfarjani P, Trifiro MA, Pinsky L. 1995. Evidence for a repressive function of the long polyglutamine tract in the human androgen receptor: Possible pathogenic relevance for the (CAG)n-expanded neuropathies. Human Mol Genet 4:523–527.

Kelada SN, Kardia SL, Walker AH, Wein AJ, Malkowicz SB, Rebbeck TR. 2000. The glutathione S-transferase-mu and -theta genotypes in the etiology of prostate cancer: Genotype-environment interactions with smoking. Cancer Epidemiol Biomarkers Prev 9:1329–1334. Key TJA, Silcocks PB, Davey GK, Appleby PN, Bishop DT. 1997. A casecontrol study of diet and prostate cancer. Br J Cancer 76:678–687. Khosravi J, Diamandi A, Mistry J, Scorilas A. 2001. Insulin-like growth factorI (IGF-I) and IGF-binding protein-3 in benign prostatic hyperplasia and prostate cancer. J Clin Endocrinol Metab 86:694–699. Kibel AS, Isaacs SD, Isaacs WB, Bova GS. 1998. Vitamin D receptor polymorphisms and lethal prostate cancer. J Urol 160:1405–1409. Kittles RA, Chen W, Panguluri RK, et al. 2002. CYP3A4-V and prostate cancer in African Americans: Causal or confounding association because of population stratification? Hum Genet 110:553–560. Kjuus H, Andersen A, Langård S. 1986. Incidence of cancer among workers producing calcium carbide. Br J Ind Med 43:237–242. Klein EA, Thompson IM, Lippman SM, et al. 2003. SELECT: The selenium and vitamin E cancer prevention trial. Urol Oncol 21:59–65. Klein RD, Fischer SM. 2002. Black tea polyphenols inhibit IFG-1 induced signaling through Akt in normal prostate epithelial cells and Du145 prostate carcinoma cells. Carcinogenesis 23:217–221. Knekt P, Aromaa A, Maatela J, et al. 1990. Serum selenium and subsequent risk of cancer among Finnish men and women. J Natl Cancer Inst 82:864–868. Koistinen H, Paju A, Koistinen R, et al. 2002. Prostate-specific antigen and other prostate-derived proteases cleave IGFBP-3, but prostate cancer is not associated with proteolytically cleaved circulating IGFBP-3. Prostate 50:112–118. Kolonel L, Winkelstein W, Jr. 1977. Cadmium and prostatic carcinoma, [letter]. Lancet 2(8037):566–567. Kolonel LN, Hankin JH, Whittemore AS, et al. 2000. Vegetables, fruits, legumes and prostate cancer: A multiethnic case-control study. Cancer Epidemiol Biomarkers Prev 9:795–804. Kolonel LN, Hankin JH, Yoshizawa CN. 1987. Vitamin A and prostate cancer in elderly men: Enhancement of risk. Cancer Res 47:2982–2985. Kolonel LN, Yoshizawa CN, Hankin JH. 1988. Diet and prostatic cancer: A case-control study in Hawaii. Am J Epidemiol 127:999–1012. Kote-Jarai Z, Easton D, Edwards SM, et al. 2001. Relationship between glutathione S-transferase M1, P1 and T1 polymorphisms and early onset prostate cancer. Pharmacogenetics 11:325–330. Kotsis SV, Spencer SL, Peyser PA, Montie JE, Cooney KA. 2002. Early onset prostate cancer: predictors of clinical grade. J Urol 167:1659–1663. Kristal AR, Lampe JW. 2002. Brassica Vegetables and prostate cancer risk: A review of the epidemiologic evidence. Nutr Cancer 42:1–9. Kristal AR, Stanford JL, Cohen JH, Wicklund K, Patterson RE. 1999. Vitamin and mineral supplement use is associated with reduced risk of prostate cancer. Cancer Epidemiol Biomarkers Prev 8:887–892. Kritchevsky D. 1999. Caloric restriction and experimental carcinogenesis. Toxicol Sci 52(suppl):13–16. Krstev S, Baris D, Stewart PA, Hayes RB, Blair A, Dosemeci M. 1998. Risk for prostate cancer by occupation and industry: A 24-state death certificate study. Am J Indust Med 34:413–420. Kupelian PA, Kupelian VA, Witte JS, Macklis R, Klein EA. 1997. Family history of prostate cancer in patients with localized prostate cancer: An independent predictor of treatment outcome. J Clin Oncology 15: 1478–1480. Kushi L, Giovannucci E. 2002. Dietary fat and cancer. Am J Med 113 Suppl 9B:63S–70S. La Vecchia C, Negri E, Decarli A, Franceschi S. 1997. Diabetes mellitus and colorectal cancer risk. Cancer Epidemiol Biomarkers Prev 6: 1007–1010. Lacey JV, Deng J, Dosemeci M, et al. 2001. Prostate cancer, benign prostatic hyperplasia and physical activity in Shanghai, China. Int J Epidemiol 30:341–349. Latil AG, Azzouzi R, Cancel GS, Guillaume EC, Cochan-Priollet B, Berthon PL, Cussenot O. 2001. Prostate carcinoma risk and allelic variants of genes involved in androgen biosynthesis and metabolism pathways. Cancer 92:1130–1137. Le Marchand L, Kolonel LN, Wilkens LR, Myers BC, Hirohata T. 1994. Animal fat consumption and prostate cancer: A prospective study in Hawaii. Epidemiology 5:276–282. Lee IM, Paffenbarger Jr., RS, Hsieh CC. 1992. Physical activity and risk of prostatic cancer among college alumni. Am J Epidemiol 135:169– 179. Lee WH, Morton RA, Epstein JI, et al. 1994. Cytidine methylation of regulatory sequences near the p-class glutathione S-transferase gene accompanies human prostatic carcinogenesis. Proc Natl Acad Sci U S A 91:11733–11737.

Prostate Cancer Leitzmann MF, Stampfer MJ, Wu K, et al. 2003. Zinc supplement use and risk of prostate cancer. J Natl Cancer Inst 95:1004–1007. Leitzmann MF, Stampfer MJ, Ma J, et al. 2002. Aspirin use in relation to risk of prostate cancer. Cancer Epidemiol Biomarkers Prev 11:1108– 1111. Lesko SM, Rosenberg L, Shapiro S. 1996. Family history and prostate cancer. Am J Epidemiol 144:1041–1047. Li C, Gronberg H, Matsuyama H, et al. 2003. Difference between Swedish and Japanese men in the association between AR CAG repeats and prostate cancer suggesting a susceptibility-modifying locus overlapping the androgen receptor gene. Int J Mol Med 11:529–533. Li J, Yen C, Liaw D, et al. 1997. PTEN, a putative protein tyrosine phosphatase gene mutated in human brain, breast, and prostate cancer. Science 275:1943–1947. Lichtenstein P, Holm NV, Verkasalo PK, et al. 2000. Environmental and heritable factors in the causation of cancer: Analyses of cohorts of twins from Sweden, Denmark, and Finland. N Engl J Med 343:78–85. Lin X, Tascilar M, Lee WH, et al. 2001. GSTP1 CpG island hypermethylation is responsible for the absence of GSTP1 expression in human prostate cancer cells. Am J Pathol 159:1815–1826. Linton HJ, Marks LS, Millar LS, Knott CL, Rittenhouse HG, Mikolajczyk SD. 2003. Benign prostate-specific antigen (BPSA) in serum is increased in benign prostate disease. Clin Chem 49:253–259. Liu B, Lee HY, Weinzimer SA, et al. 2000a. Direct functional interactions between insulin-like growth factor-binding protein-3 and retinoid X receptor-a regulate transcriptional signaling and apoptosis. J Biol Chem 275:33607–33613. Liu S, Lee I-M, Linson P, Ajani U, Buring JE, Hennekens CH. 2000b. A prospective study of physical activity and risk of prostate cancer in US physicians. Int J Epidemiol 29:29–35. Lu TW, Hung JC, Heber D, Go VL, Reuter VE, Cordon-Cardo C, et al. 2001. Inverse association between plasma lycopene and other carotenoids and prostate cancer. Cancer Epidemiol Biomarkers Prev 10:749–756. Lund Nilsen TI, Johnsen R, Vatten LJ. 2000. Socio-economic and lifestyle factors associated with the risk of prostate cancer. Br J Cancer 82: 1358–1363. Lund Nilsen TI, Vatten LJ. 1999. Anthropometry and prostate cancer risk: A prospective study of 22,248 Norwegian men. Cancer Causes Control 10:269–275. Lunn RM, Bell DA, Mohler JL, Taylor JA. 1999. Prostate cancer risk and polymorphisms in 17 hydroxylase (CYP17) and steroid reductase (SRD5A2). Carcinogenesis 20:1727–1731. Luo J, Zha S, Gage WR, et al. 2002. A-Methylacyl-CoA Racemase: A new marker for prostate cancer. Cancer Res 62:2220–2226. Ma J, Stampfer MJ, Gann PH, et al. 1998. Vitamin D receptor polymorphisms, circulating vitamin D metabolites, and risk of prostate cancer in United States Physicians. Cancer Epidemiol Biomarkers Prev 7:385–390. Makridakis NM, Ross RK, Pike MC, et al. 1999. Association of missense substitution in SRD5A2 gene with prostate cancer in African-American and Hispanic men in Los Angeles, USA. Lancet 354:975–978. Mantzoros CS, Tzonou A, Signorello LB, Stampfer M, Trichopoulos D, Adami HO. 1997. Insulin-like growth factor 1 in relation to prostate cancer and benign prostatic hyperplasia. Br J Cancer 76:1115–1118. Matarese G, La Cava A, Sanna V, et al. 2002. Balancing susceptibility to infection and autoimmunity: A role for leptin? Trends Immunol 23: 182–187. McCarron SL, Edwards S, Evans PR, et al. 2002. Influence of cytokine gene polymorphisms on the development of prostate cancer. Cancer Res 62:3369–3372. McNeal JE, Bostwick DG. 1986. Intraductal dysplasia: A premalignant lesion of the prostate. Hum Pathol 17:64–71. Meeker AK, Hicks JL, Platz EA, et al. 2002. Telomere shortening is an early somatic DNA alteration in human prostate tumorigenesis. Cancer Res 62:6405–6409. Meitz JC, Edwards SM, Easton DF, Murkin A, Ardern-Jones A, Jackson RA, Williams S, Dearnaley DP, Stratton MR, Houlston RS and others. 2002. HPC2/ELAC2 polymorphisms and prostate cancer risk: Analysis by age at onset. Br J Cancer 87:905–908. Michaud DS, Augustsson K, Rimm EB, Stampfer MJ, Willett WC, Giovannucci E. 2001. A prospective study on intake of animal products and risk of prostate cancer. Cancer Causes Control 12:557–567. Mikolajczyk SD, Millar LS, Wang TJ, et al. 2000. A precursor form of prostatespecific antigen is more highly elevated in prostate cancer compared with benign transition zone prostate tissue. Cancer Res 60:756–759. Millar DS, Paul CL, Molloy PL, Clark SJ. 2000. A distinct sequence (ATAAA)n separates methylated and unmethylated domains at the 5¢-end of the GSTP1 CpG Island. J Biol Chem 275:24893–24899.

1147

Miller EA, Stanford JL, Hsu L, Noonana E, Ostrander EA. 2001. Polymorphic repeats in the androgen receptor in high-risk sibships. Prostate 48: 200–205. Mills PK, Beeson WL, Phillips RL, Fraser GE. 1989. Cohort study of diet, lifestyle, and prostate cancer in Adventist men. Cancer 64:598–604. Mills PK, Yang R. 2003. Prostate cancer risk in California farm workers. J Occup Environ Med 45:249–258. Mohan RR, Challa A, Gupta S, et al. 1999. Overexpression of ornithine decarboxylase in prostate cancer and prostatic fluid in humans. Clin Cancer Res 5:143–147. Moore MG, Wetterau LA, Freancis MJ, Peehl DM, Cohen P. 2003. Novel stimulatory role for insulin-like growth factor binding protein-2 in prostate cancer cells. Int J Cancer 105:14–19. Morrison NA, Qi JC, Tokita A, et al. 1994. Prediction of bone density from vitamin D receptor alleles. Nature 367:284–287. Moskaluk CA, Duray PH, Cowan KH, Linehan M, Merion MJ. 1997. Immunohistochemical expression of p-class glutathione S-transferase is downregulated in adenocarcinoma of the prostate. Cancer 79:1595–1599. Mukherjee P, Sotnikov AV, Mangian HJ, Zhou JR, Visek WJ, Clinton SK. 1999. Energy intake and prostate tumor growth, angiogenesis, and vascular endothelial growth factor expression. J Natl Cancer Inst 91:512–523. Murata M, Watanabe M, Yamanaka M, et al. 2001. Genetic polymorphisms in cytochrome P450 (CYP) 1A1, CYP1A2, CYP2E1, glutathione S-transferase (GST) M1 and GSTT1 and susceptibility to prostate cancer in the Japanese population. Cancer Lett 165:171–177. Nagao M. 1999. A new approach to risk estimation of food-borne carcinogens—heterocyclic amines—based on molecular information. Mutat Res 43:3–12. Nakashima J, Tachibana M, Horiguchi Y, et al. 2000. Serum interleukin 6 as a prognostic factor in patients with prostate cancer. Clin Cancer Res 6:2702–2706. Nam RK, Toi A, Vesprini D, et al. 2001. V89L polymorphism of type-2, 5-alpha reductase enzyme gene predicts prostate cancer presence and progression. Urology 57:199–205. Nathanson KL, Omaruddin R, Malkowicz SB, Rebbeck TR. 2001. An intronic variant in PTEN is not association with prostate cancer risk. Cancer Epidemiol Biomarkers Prev 10:277–278. Nelson JE, Harris RE. 2000. Inverse association of prostate cancer and nonsteroidal anti-inflammatory drugs (NSAIDS): Results of a case-control study. Oncol Rep 7:169–170. Nelson WG, De Marzo AM, De Weese TL. 2001. The molecular pathogenesis of prostate cancer: Implications for prostate cancer prevention. Urology 57(Suppl 4A):39–45. Nelson WG, De Marzo AM, Isaacs WB. 2003. Prostate cancer. N Engl J Med 24; 349:366–381. Nomura A, Heilbrun LK, Stemmermann GN, Judd HL. 1988. Prediagnostic serum hormones and the risk of prostate cancer. Cancer Res 48: 3515–3517. Nomura AM. 2001. Body size and prostate cancer. Epidemiol Rev 23: 126–131. Nomura AM, Stemmermann GN, Lee J, Craft NE. 1997. Serum micronutrients and prostate cancer in Japanese Americans in Hawaii. Cancer Epidemiol Biomarkers Prev 6:487–491. Nomura AM, Stemmermann GN, Lee J, et al. 1998. Serum vitamin D metabolite levels and the subsequent development of prostate cancer. Cancer Causes Control 9:425–432. Nomura AMY, Lee J, Stemmermann GN, Combs Jr., GF. 2000. Serum selenium and subsequent risk of prostate cancer. Cancer Epidemiol Biomarkers Prev 9:883–887. Norman A, Moradi T, Gridley G, et al. 2002. Occupational physical acitivity and risk for prostate cancer in a nationwide cohort study in Sweden. Br J Cancer 86:70–75. Norrish AE, Ferguson LR, Knize MG, Felton JS, Sharpe SJ, Jackson RT. 1999. Heterocyclic amine content of cooked meat and risk of prostate cancer. J Natl Cancer Inst 91:2038–2044. Norrish AE, Jackson RT, McRae CU. 1998. Non-steroidal anti-inflammatory drugs and prostate cancer progression. Int J Cancer 77:511–515. Norrish AE, Jackson RT, Sharpe SJ, Skeaff CM. 2000. Men who consume vegetable oils rich in monounsaturated fat: Their dietary patterns and risk of prostate cancer (New Zealand). Cancer Causes Control 11:609–615. Ntais C, Polycarpou A, Ioannidis JPA. 2003. Association of the CYP17 gene polymorphism with the risk of prostate cancer: A meta-analysis. Cancer Epidemiol Biomarkers Prev 12:120–126. Nupponen N, Visakorpi T. 1999. Molecular biology of progression of prostate cancer. Eur Urol 35:351–354. Nwosu V, Carpten J, Trent JM, Sheridan R. 2001. Heterogeneity of genetic alterations in prostate cancer: Evidence of the complex nature of the disease. Hum Mol Genet 10:2313–2318.

1148

PART IV: CANCER BY TISSUE OF ORIGIN

Oh BR, Sasaki M, Perinchery G, et al. 2000. Frequent genotype changes at -308, and 488 regions of the tumor necrosis factor-a (TNF-a) gene in patients with prostate cancer. J Urol 163:1584–1587. Okamoto M, Lee C, Oyasu R. 1997. Interleukin-6 as a paracrine and autocrine growth factor in human prostatic carcinoma cells in vitro. Cancer Res 57:141–146. Oliveria SA, Kohl III HW, Trichopoulos D, Blair SN. 1996. The association between cardiorespiratory fitness and prostate cancer. Med Sci Sports Exerc 28:97–104. Optenberg SA, Thompson IM, Friedrichs P, Wojcik B, Stein CR, Kramer B. 1995. Race, treatment, and long-term survival from prostate cancer in an equal-access medical care delivery system. J Am Med Asso 274: 1599–1605. Paffenbarger Jr., RS, Hyde RT, Wing AL. 1987. Physical activity and incidence of cancer in diverse populations: A preliminary report. Am J Clin Nutr 45(suppl):312–317. Page WF, Braun MM, Partin AW, Caporaso N, Walsh P. 1997. Heredity and prostate cancer: A study of World War II veteran twins. Prostate 33:240–245. Paradis AE, Kantoff PW, Giovannucci E, Stampfter MJ, Ma J. 2003. Association between the Met326Ile polymorphism of the p85a regulatory subunit of the phosphatidylinositol 3-kinase and prostate cancer risk: A prospective study. Cancer Epidemiol Biomarkers Prev 12:172–173. Paris PL, Kupelian PA, Hall JM, et al. 1999. Association between a CYP3A4 genetic variant and clinical presentation in African-American prostate cancer patients. Cancer Epidemiol Biomarkers Prev 8:901–905. Parkin DM, Whelan SL, Ferlay J, Teppo L, Thomas D. 2003. Cancer Incidence in Five Continents. Lyon, France: International Agency for Research on Cancer, IARC Scientific Publications No. 155. Parsons JK, Nelson CP, Gage WR, Nelson WG, Kensler TW, De Marzo AM. 2001. GSTA1 expression in normal, preneoplastic, and neoplastic human prostate tissue. Prostate 49:30–37. Pearce CL, Makridakis NM, Ross RK, et al. 2002. Steroid 5-a reductase type II V89L substitution is not associated with risk of prostate cancer in a multiethnic population study. Cancer Epidemiol Biomarkers Prev 11:417–418. Peehl DM, Skowronski RJ, Leung GK, Wong ST, Stamey TA, Feldman D. 1994. Antiproliferative effects of 1,25-dihydroxyvitamin D3 on primary cultures of human prostatic cells. Cancer Res 54:805–810. Peracaula R, Tabares G, Royle L, et al. 2003. Altered glycosylation pattern allows the distinction between prostate specific antigen (PSA) from normal and tumour origins. Glycobiology 13:457– 470. Platz E. 2002a. Energy imbalance and prostate cancer. J Nutr 132(suppl): 3471S–3481S. Platz E, Giovannucci E, Dahl D, et al. 1998a. The androgen receptor gene GGN microsatellite and prostate cancer risk. Cancer Epidemiol Biomarkers Prev 7:379–384. Platz E, Leitzmann MF, Rimm EB, Willett WC, Giovannucci E. 2004. Alcohol intake, drinking patterns and risk of prostate cancer in a large prospective cohort study. Am J Epidemiol 159:449–453. Platz EA, Giovannucci E, Rimm EB, et al. 1998b. Retrospective analysis of birth weight and prostate cancer in the Health Professionals Follow-up Study. Am J Epidemiol 147:1140–1144. Platz EA, Helzlsouer KJ, Hoffman SC, Morris JS, Baskett CK, Comstock GW. 2002b. Prediagnostic toenail cadmium and zinc and subsequent prostate cancer risk. Prostate 52:288–296. Platz EA, Hoffman SC, Morris JS, Baskett CK, Helzlsouer KJ, Comstock GW. 2001. Prediagnostic toenail cadmium and zinc and subsequent prostate cancer risk. Proc Am Asso Cancer Res 42:151. Platz EA, Krithivas K, Kantoff PW, Stampfer MJ, Giovannucci E. 2002c. ATAAA repeat upstream of glutathione S-transferase P1 and prostate cancer risk. Urology 59:159–164. Platz EA, Leitzmann MF, Michaud DS, Willett WC, Giovannucci E. 2003. Interrelation of energy intake, body size, and physical activity with prostate cancer in a large prospective cohort study. Cancer Res 63:8542–8548. Platz EA, Pollak MN, Willett WC, Giovannucci E. 2000a. Vertex balding, plasma insulin-like growth factor 1, and insulin-like growth factor binding protein 3. J Am Acad Dermatol 42:1003–1007. Platz EA, Rimm EB, Willett WC, Kantoff PW, Giovannucci E. 2000b. Racial variation in prostate cancer incidence and in hormonal system markers among male health professionals. J Natl Cancer Inst 92: 2009–2017. Polednak AP. 1976. College athletics, body size, and cancer Mortality. Cancer 38:382–387. Potosky AL, Miller BA, Albertsen PC, Kramer BS. 1995. The role of increasing detection in the rising incidence of prostate cancer. AMA 273:548–552.

Putnam S, Cerhan J, Parker A, et al. 2000. Lifestyle and anthropometric risk factors for prostate cancer in a cohort of Iowa men. Ann Epidemiol 10:361–369. Putzi MJ, De Marzo AM. 2000. Morphologic transitions between proliferative inflammatory atrophy and high-grade prostatic intraepithelial neoplasia. Urology 56:828–832. Ragozzino M, Melton LJ, 3d, Chu CP, Palumbo PJ. 1982. Subsequent cancer risk in the incidence cohort of Rochester, Minnesota, residents with diabetes mellitus. J Chronic Dis 35:13–19. Rajah R, Valentinis B, Cohen P. 1997. Insulin-like growth factor (IGF)-binding protein-3 induces apoptosis and mediates the effects of transforming growth factor-b1 on programmed cell death through a p53- and IGFindependent mechanism. J Biol Chem 272:12181–12188. Rebbeck TR, Jaffe JM, Walker AH, Wein AJ, Malkowicz SB. 1998. Modification of clinical presentation of prostate tumors by a novel genetic variant in CYP3A4. J Natl Cancer Inst 19:1225–1229. Rebbeck TR, Walker AH, Jaffe JM, White DL, Wein AJ, Malkowicz SB. 1999. Glutathione S-transferase-mu (GSTM1) and -theta (GSTT1) genotypes in the etiology of prostate cancer. Cancer Epidemiol Biomarkers Prev 8:283–287. Reichel H, Koeffler HP, Norman AW. 1989. The role of the vitamin D endocrine system in health and disease. N Engl J Med 320:980–991. Reichman ME, Hayes RB, Ziegler RG, et al. 1990. Serum vitamin A and subsequent development of prostate cancer in the first National Health and Nutrition Examination Survey Epidemiologic Follow-up Study. Cancer Res 50:2311–2315. Ripple MO, Henry WF, Rago RP, Wilding G. 1997. Prooxidant-antioxidant shift induced by androgen treatment of human prostate carcinoma cells. J Natl Cancer Inst 89:40–48. Roberts RO, Jacobson DJ, Girman CJ, Rhodes T, Lieber MM, Jacobsen SJ. 2002. A population-based study of daily nonsteroidal anti-inflammatory drug use and prostate cancer. Mayo Clin Proc 77:219–225. Roberts RO, Lieber MM, Rhodes T, Girman CJ, Bostwick DG, Jacobsen SJ. 1998. Prevalence of a physician-assigned diagnosis of prostatitis: The Olmsted Country Study of Urinary Symptoms and Health Status Among Men. Urology 51:578–584. Roberts WW, Platz EA, Walsh PC. 2003. Association of cigarette smoking with extraprostatic prostate cancer in young men. J Urol 169:512–516. Rodríguez C, Callee EE, Miracle-McMahill HL, et al. 1997. Family history and risk of fatal prostate cancer. Epidemiology 8:653–657. Rodriguez C, Tatham LM, Thun MJ, Calle EE, Heath Jr., CW. 1997. Smoking and fatal prostate cancer in a large cohort of adult men. Am J Epidemiol 145:466–475. Rohrmann S, Roberts WW, Walsh PC, Platz EA. 2003. Family history of prostate cancer and obesity in relation to high-grade disease and extraprostatic extension in young men with prostate cancer. Prostate 55:140–146. Rokman A, Ikonen T, Seppala EH, et al. 2002. Germline alterations of the RNASEL gene, a candidate HPC1 gene at 1q25, in patients and families with prostate cancer. Am J Hum Genet 70:1299–1304. Rose DP, Boyar AP, Wynder EL. 1986. International comparisons of mortality rates for cancer of the breast, ovary, prostate, and colon, and per capita food consumption. Cancer 58:2363–2371. Rosenberg DJ, Neugut AI, Ahsan H, Shea S. 2002. Diabetes mellitus and the risk of prostate cancer. Cancer Invest 20:157–165. Ross RK, Henderson BE. 1994. Do diet and androgens alter prostate cancer risk via a common etiologic pathway? J Natl Cancer Inst 86:252–254. Ross RK, Shimizu H, Paganini-Hill A, Honda G, Henderson BE. 1987. Casecontrol studies of prostate cancer in blacks and whites in Southern California. J Natl Cancer Inst 78:869–874. Ruska KM, Sauvageot J, Epstein JI. 1998. Histology and cellular kinetics of prostatic atrophy. Am J Surg Pathol 22:1073–1077. Sakr WA, Haas GP, Cassin BF, Pontes JE, Crissman JD. 1993. The frequency of carcinoma and intraepithelial neoplasia of the prostate in young male patients. J Urol 150:379–385. Schipper RG, Romijn JC, Cuipers VM, Verhofstad AA. 2003. Polyamines and prostatic cancer. Biochem Soc Trans 31:375–389. Schuurman AG, Goldbohm A, Dorant E, van den Brandt PA. 1998. Vegetable and fruit consumption and prostate cancer risk: A cohort study in the Netherlands. Cancer Epidemiol Biomarkers Prev 7:673–680. Schuurman AG, Goldbohm RA, Dorant E, van den Brandt PA. 2000. Anthropometry in relation to prostate cancer risk in the Netherlands Cohort Study. Am J Epidemiol 151:541–549. Schuurman AG, van den Brandt PA, Dorant E, Brants HA, Goldbohm RA. 1999a. Association of energy and fat intake with prostate carcinoma risk: Results from the Netherlands Cohort Study. Cancer 86:1019–1027. Schuurman AG, van den Brandt PA, Dorant E, Goldbohm RA. 1999b. Animal products, calcium and protein and prostate cancer risk in The Netherlands Cohort Study. Br J Cancer 80:1107–1113.

Prostate Cancer Schwartz GG, Hulka BS. 1990. Is vitamin D deficiency a risk factor for prostate cancer? (Hypothesis). Anticancer Res 10:1307–1311. Sesso H, Paffenbarger R, Lee I. 2001. Alcohol consumption and risk of prostate cancer: The Harvard Alumni Health Study. Int J Epidemiol 30:749–755. Settimi L, Masina A, Andrion A, Axelson O. 2003. Prostate cancer and exposure to pesticides in agricultural settings. Int J Cancer 104:458–461. Severson RK, Grove JS, Nomura AMY, Stemmermann GN. 1988. Body mass and prostatic cancer: A prospective study. Br Med J 297:713–715. Severson RK, Nomura AMY, Grove JS, Stemmermann GN. 1989. A prospective study of demographics, diet, and prostate cancer among men of Japanese ancestry in Hawaii. Cancer Res 49:1857–1860. Shariat SF, Andrews B, Kattan MW, Kim J, Wheeler TM, Slawin KM. 2001. Plasma levels of interleukin-6 and its soluble receptor are associated with prostate cancer progression and metastasis. Urology 58:1008–1015. Sharma-Wagner S, Chokkalingam AP, Malker HS, Stone BJ, McLaughlin JK, Hsing AW. 2000. Occupation and prostate cancer risk in Sweden. J Occup Environ Med 42:517–525. Sharpe CR, Siemiatycki J. 2002. Consumption of non-alcoholic beverages and prostate cancer risk. Eur J Cancer Prev 11:497–501. Shepard TF, Platz EA, Kantoff PW, et al. 2000. No association between the I105V polymorphism of the glutathione S-transferase P1 gene (GSTP1) and prostate cancer risk: A prospective study. Cancer Epidemiol Biomarkers Prev 9:1267–1268. Shi R, Berkel HJ, Yu H. 2001. Insulin-like growth factor-I and prostate cancer: A meta-analysis. Br J Cancer 85:991–996. Shimizu H, Ross RK, Bernstein L. 1991a. Possible underestimation of the incidence rate of prostate cancer in Japan. Jpn J Cancer Res 82:483– 485. Shimizu H, Ross RK, Bernstein L, Yatani R, Henderson BE, Mack TM. 1991b. Cancers of the prostate and breast among Japanese and white immigrants in Los Angeles county. Br J Cancer 63:963–966. Sies H, Stahl W. 1995. Vitamins E and C, ß-carotene, and other carotenoids as antioxidants. Am J Clin Nutr 62 (suppl): 1315S–1321S. Signorello LB, Wuu J, Hsieh C-C, Tzonou A, Trichopoulos D, Mantzoros CS. 1999. Hormones and hair patterning in men: A role for insulin-like growth factor 1? J Am Acad Dermatol 40:200–203. Sigounas G, Anagnostou A, Steiner M. 1997. d1-a-tocopherol induces apoptosis in erythroleukemia, prostate, and breast cancer cells. Nutr Cancer 28:30–35. Silverman RH. 2003. Implications for RNase L in prostate cancer biology. Biochem 42:1805–1812. Sirovich BE, Schwartz LM, Woloshin S. 2003. Screening men for prostate and colorectal cancer in the United States: Does practice reflect the evidence? JAMA 289:1414–1420. Sitas F, Carrara H, Beral V, et al. 1999. Antibodies against human herpesvirus 8 in black South African Patients with cancer. N Engl J Med 340: 1863–1871. Skegg D. 1993. Vasectomy and prostate cancer: Is there a link? NZ Med J 106:242–243. Slattery ML, West DW. 1993. Smoking, alcohol, coffee, tea, caffeine, and theobromine: Risk of prostate cancer in Utah (United States). Cancer Causes Control 4:559–563. Sleddens HFBM, Oostra BA, Brinkmann AO, Trapman J. 1993. Trinucleotide (GGN) repeat polymorphism in the human androgen receptor (AR) gene. Hum Mol Genet 2:493. Snowdon DA, Phillips RL, Choi W. 1984. Diet, obesity, and risk of fatal prostate cancer. Am J Epidemiol 120:244–250. Soderstrom T, Wadelius M, Andersson SO, et al. 2002. 5alpha-reductase 2 polymorphisms as risk factors for prostate cancer. Pharmacogenetics 12: 307–312. Spitz MR, Strom SS, Yamamura Y, et al. 2000. Epidemiologic determinants of clinically relevant prostate cancer. Int J Cancer 89:259–264. Sporn MB, Roberts AB. 1984. Role of retinoids in differentiation and carcinogenesis. J Natl Cancer Inst 73:1381–1386. Stahl W, Sies H. 1992. Uptake of lycopene and its geometrical isomers is greater from heat-processed than from unprocessed tomato juice in humans. J Nutr 122:2161–2166. Stanford JL, Just JJ, Gibbs M, et al. 1997. Polymorphic repeats in the androgen receptor gene: Molecular markers of prostate cancer risk. Cancer Res 57:1194–1198. Stanford JL, Ostrander EA. 2001. Familial prostate cancer. Epidemiol Rev 23:19–23. Stanford JL, Stephenson RA, Coyle LM, et al. 1999. Prostate Cancer Trends 1973–1995. Bethesda, MD: SEER Program, National Cancer Institute NIH Pub. No. 99–4543. Stattin P, Bylund A, Rinaldi S, et al. 2000. Plasma insulin-like growth factorI, insulin-like growth factor-binding proteins, and prostate cancer risk: A prospective study. J Natl Cancer Inst 92:1910–1917.

1149

Stattin P, Kaaks R, Johansson R, et al. 2003. Plasma leptin is not associated with prostate cancer risk. Cancer Epidemiol Biomarkers Prev 12: 474–475. Stattin P, Soderberg S, Hallmans G, et al. 2001. Leptin is associated with increased prostate cancer risk. A nested case-referent study. J Clin Endocrinol Metab 86:1341–1345. Steenland K, Nowlin S, Palu S. 1995. Cancer incidence in the National Health and Nutrition Survey I. Follow-up data: Diabetes, cholesterol, pulse, and physical activity. Cancer Epidemiol Biomarkers Prev 4:807–811. Steinhoff C, Franke KH, Golka K, et al. 2000. Glutathione transferase isozyme genotypes in patients with prostate and bladder carcinoma. Arch Toxicol 74:521–526. Strickler HD, Burk R, Shah K, et al. 1998. A multifaceted study of human papillomavirus and prostate carcinoma. Cancer 82:1118–1125. Strom SS, Spitz MR, Yamamura Y, Babaian RJ, Scardino PT, Wei Q. 2001. Reduced expression of hMSH2 and hMLH1 and risk of prostate cancer: A case-control study. Prostate 47:269–275. Sung V, Feldman D. 2000. 1,25-Dihydroxyvitamin D3 decreases human prostate cancer cell adhesion and migration. Mol Cell Endocrinol 164:133–143. Swan J, Breen N, Coates RJ, Rimer BK, Lee NC. 2003. Progress in cancer screening practices in the United States. Cancer 97:1528–1540. Tanaka Y, Sasaki M, Kaneuchi M, Shiina H, Igawa M, Dahiya R. 2002. Polymorphisms of the CYP1B1 gene have higher risk for prostate cancer. Biochem Biophys Res Commun 296:820–826. Tavani A, Gallus S, Bosetti C, et al. 2002. Diabetes and the risk of prostate cancer. Eur Cancer Prev 11:125–128. Tayeb MT, Clark C, Haites NE, Sharp L, Murray GI, McLeod HL. 2003. CYP3A4 and VDR gene polymorphisms and the risk of prostate cancer in men with benign prostatic hyperplasia. Br J Cancer 88:928–932. Terry P, Lichtenstein P, Reychting M, Ahlbom A, Wolk A. 2001. Fatty fish consumption and risk of prostate cancer. Lancet 357:1764–1766. The Alpha-Tocopherol Beta Carotene Cancer Prevention Study Group. 1994. The effect of vitamin E and beta carotene on the incidence of lung cancer and other cancers in male smokers. N Engl J Med 330:1029–1035. Thompson IM, Goodman PJ, Tangen CM, et al. 2003. The influence of finasteride on the development of prostate cancer. N Engl J Med 17; 349:215–224. Thompson HJ, Jiang W, Zhu Z. 1999. Mechanisms by which energy restriction inhibits carcinogenesis. Adv Exp Med Biol 470:77–84. Thompson MM, Garland C, Barrett-Connor E, Khaw KT, Friedlander NJ, Wingard DL. 1989. Heart disease risk factors, diabetes, and prostatic cancer in an adult community. Am J Epidemiol 129:511–517. Thune I, Lund E. 1994. Physical activity and the risk of prostate and testicular cancer: A cohort study of 53,000 Norwegian men. Cancer Causes Control 5:549–556. Tibblin G, Eriksson M, Cnattingius S, Ekbom A. 1995. High birthweight as a predictor of prostate cancer risk. Epidemiology 6:423–424. Tonnesen H, Moller H, Andersen JR, Jensen E, Juel K. 1994. Cancer morbidity in alcohol abusers. Br Med J 69:327–332. Twillie DA, Eisenberger MA, Carducci MA, Hsieh WS, Kim WY, Simons JW. 1995. Interleukin-6:A candidate mediator of human prostate cancer morbidity. Urology 45:542–549. U.S. Preventive Services Task Force. 2002. Screening—Prostate Cancer, U.S. Preventive Services Task Force Update. Uzzo RG, Pinover WH, Horwitz EM, et al. 2003. Free prostate-specific antigen improves prostate cancer detection in a high-risk population of men with a normal total PSA and digital rectal examination. Urology 61:754–759. Valeri A, Briollais L, Azzouzi R, et al. 2003. Segregation analysis of prostate cancer in France: Evidence for autosomal dominant inheritance and residual brother-brother dependence. Ann Hum Genet 67:125–137. van der Gulden JWJ, Kolk JJ, Verbeek ALM. 1995. Work environment and prostate cancer risk. Prostate 27:250–257. van Leenders GJ, Gage WR, Hicks JL, et al. 2003. Intermediate cells in human prostatic epithelium are enriched in proliferative inflammatory atrophy. Am J Pathol 162:1529–1537. Veierod MB, Laake P, Thelle DS. 1997. Dietary fat intake and risk of prostate cancer: A prospective study of 25,708 Norwegian men. Int J Cancer 73:634–638. Vena JE, Graham S, Zielezny M, Brasure J, Swanson MK. 1987. Occupational exercise and risk of cancer. Am J Clin Nutr 45:318–327. Villeneuve PJ, Johnson KC, Kreiger N, Mao Y, The Canadian Cancer Registries Epidemiology Research Group. 1999. Risk factors for prostate cancer: Results from the Canadian National Enhanced Cancer Surveillance System. Cancer Causes Control 10:355–367. Visvanathan K, Helzlsouer KJ, Boorman DW, et al. 2004. Association among an ornithine decarboxylase polymorphism, androgen receptor gene (CAG) repeat length and prostate cancer risk. J Urol 171:652–655.

1150

PART IV: CANCER BY TISSUE OF ORIGIN

Vlajinac HD, Marinkovic JM, Ilic MD, Kocev NI. 1997. Diet and prostate cancer: A case-control study. Eur J Cancer 33:101–107. Vogt TM, Mayne ST, Graubard BI, et al. 2002. Serum lycopene, other serum carotenoids, and risk of prostate cancer in US blacks and whites. Am J Epidemiol 155:1023–1032. Wadelius M, Autrup JL, Stubbins MJ, et al. 1999. Polymorphisms in NAT2, CYP2D6, CYP2C19 and GSTP1 and their association with prostate cancer. Pharmacogenetics 9:333–340. Walker AH, Jaffe JM, Gunasegaram S, et al. 1998. Characterization of an allelic variant in the nifedipine-specific element of CYP3A4:Ethnic distribution and implications for prostate cancer risk. Hum Mutat 12:289. West DW, Slattery ML, Robison LM, French TK, Mahoney AW. 1991. Adult dietary intake and prostate cancer risk in Utah: A case-control study with special emphasis on aggressive tumors. Cancer Causes Control 2:85–94. Wheeler GD, Wall SR, Belcastro AN, Cumming DC. 1984. Reduced serum testosterone and prolactin levels in male distance runners. JAMA 252:514–516. Whittemore AS, Kolonel LN, Wu AH, et al. 1995a. Prostate cancer in relation to diet, physical activity, and body size in blacks, whites, and Asians in the United States and Canada. J Natl Cancer Inst 87:652–661. Whittemore AS, Paffenbarger Jr RS, Anderson K, Lee JE. 1985. Early precursors of site-specific cancers in college men and women. J Natl Cancer Inst 74:43–51. Whittemore AS, Wu AH, Kolonel LN, et al. 1995b. Family history and prostate cancer risk in black, white, and Asian men in the United States and Canada. Am J Epidemiol 141:732–740. Wiklund F, Jonsson BA, Goransson I, Bergh A, Gronberg H. 2003. Linkage analysis of prostate cancer susceptibility: Confirmation of linkage at 8p22–23. Hum Genet 112:414–418. Epub 2003 Feb 20. Will JC, Vinicor F, Calle EE. 1999. Is diabetes mellitus associated with prostate cancer incidence and survival? Epidemiology 10:313–318. Willett WC, Stampfer MJ, Underwood BA, et al. 1984. Vitamin A supplementation and plasma retinol levels: A randomized trial among women. J Natl Cancer Inst 73:1445–1448. Wise AM, Stamey TA, McNeal JE, Clayton JL. 2002. Morphologic and clinical significance of multifocal prostate cancers in radical prostatectomy specimens. Urology 60:264–269.

Wise GJ, Marella VK, Talluri G, Shirazian D. 2000. Cytokine variations in patients with hormone treated prostate cancer. J Urol 164:722–725. Wolk A, Mantzoros CS, Andersson SW, et al. 1998. Insulin-like growth factor I and prostate cancer risk: A population-based, case-control study. J Natl Cancer Inst 90:911–915. Wu K, Erdman JW, Schwartz SJ, et al. 2004. Plasma and dietary carotenoids and the risk of prostate cancer—a nested case-control study. Cancer Epidemiol Biomarkers Prev 13:260–269. Xu J, Zheng L, Komiya A, et al. 2003. Common sequence variants of the macrophage scavenger receptor 1 gene are associated with prostate cancer risk. Am J Hum Genet 72:208–212. Xu J, Zheng S, Turner A, et al. 2002a. Associations between hOGG1 sequence variants and prostate cancer susceptibility. Cancer Res 62:2253–2257. Xu J, Zheng SL, Komiya A, et al. 2002b. Germline mutations and sequence variants of the macrophage scavenger receptor 1 gene are associated with prostate cancer risk. Nat Genet 32:321–325. Xue W, Irvine RA, Yu MC, Ross RK, Coetzee GA, Ingles SA. 2000. Susceptibility to prostate cancer: Interaction between genotypes at the androgen receptor and prostate-specific antigen loci. Cancer Res 60:839–841. Yamada Y, Watanabe M, Murata M, et al. 2001. Impact of genetic polymorphisms of 17-hydroxylase cytochrome p-450 (CYP17) and steroid 5areductase type II (SRD5A2) genes on prostate-cancer risk among the Japanese population. Int J Cancer 92:683–686. Yatani R, Chigusa I, Akazaki K, Stemmermann GN, Welsh RA, Correa P. 1982. Geographic pathology of latent prostatic carcinoma. Int J Cancer 29:611–616. Yoshizawa K, Willett WC, Morris SJ, et al. 1998. A study of prediagnostic selenium level in toenails and the risk of advanced prostate cancer. J Natl Cancer Inst 90:1219–1224. Yu H, Harris RE, Gao YT, Gao R, Wynder EL. 1991. Comparative epidemiology of cancers of the colon, rectum, prostate and breast in Shanghai, China versus the United States. Int J Epidemiol 20:76–81. Yu H, Harris RE, Wynder EL. 1988. Case-control study of prostate cancer and socioeconomic factors. Prostate 13:317–325. Zha S, Gage WR, Sauvageot J, et al. 2001. Cyclooxygenase-2 is up-regulated in proliferative inflammatory atrophy of the prostate, but not in prostate carcinoma. Cancer Res 61:8617–8623.

60

Testicular Cancer ARUNA V. SARMA, JULIE C. MCLAUGHLIN, AND DAVID SCHOTTENFELD

C

ancer of the testis is relatively uncommon in the United States, accounting for approximately 1% of cancers in males (Buetow, 1995). The annual age-adjusted incidence rate of testicular cancer in 1999 was 5.5 per 100,000 men. In 2005, approximately 8000 incident cases and 390 deaths were expected in the United States. The lifetime probability of US males being diagnosed with and dying from testicular cancer is approximately 0.35 and 0.02, respectively. Cancer of the testis is most commonly diagnosed in males between the ages of 15 and 44, and occurs more often in white than black men (Fig. 60–1, Table 60–1). The lifetime probabilities of diagnosis by race are 0.42 and 0.10 for white and black men, respectively. The median age at diagnosis is 34. The SEER annual age-adjusted incidence rate in whites in 1990–1999 (6.1 per 100,000) was almost six times that in blacks (1.1 per 100,000) in the United States (Fig. 60–2). During the period 1973–1999, the age-adjusted testicular cancer incidence in the United States increased 58% in whites and 29% in blacks; concurrently, age-adjusted mortality declined 34% in whites and 46% in blacks. Dramatic advances in the management of testicular cancer have been accompanied by declining age-adjusted mortality rates (per 100,000) in white males—from 0.8 in 1975 to 0.3 in 2000 (Ries et al., 2003). The 5-year relative survival rate in whites, all stages, was 72.0% for patients diagnosed in 1970–1973, and 95.8% for patients diagnosed in 1992–1999. While far less common in black males, the 5-year relative survival rate lags behind whites at 86.9% (Ries et al., 2002). Although the definitive cause of testicular cancer is unknown, major risk factors include being of European origin or European American, cryptorchidism, inguinal hernia, testicular atrophy or dysgenesis, and family history (Kinkade, 1999). The young age of onset of testicular cancer may reflect in utero or prepubertal exposures, and current studies are attempting to identify and explore factors, focusing on maternal age and parity, parental occupational exposures, in utero exposure to estrogens, diet, and other lifestyle and environmental exposures.

CLASSIFICATION Histopathology Most primary neoplasms of the testis arise from germinal elements, which account for 95% of all testicular tumors. The principal cell types of the testis are the germ cells, derived from primitive ectodermal cells; supporting cells derived from the coelomic epithelium of the gonadal ridge that differentiate into Sertoli cells; and stromal or interstitial cells derived from the mesenchyme of the gonadal ridge that differentiate into Leydig cells. Germinal neoplasms may be composed of embryonal and/or extraembryonal tissues and are divided clinically into the seminoma and a variety of pure and mixed types of nonseminomatous tumors (Table 60–2). Five basic cell types predominate either alone or in combination: seminoma (classic or the rare spermatocytic seminoma); teratoma (mature, immature, with malignant transformation); choriocarcinoma; and yolk sac tumor (endodermal sinus tumor of Teilum), the cells of which stain immunohistochemically for alpha-fetoprotein (AFP). In patients under 15 years of age, yolk sac and teratoma neoplastic cell types predominate; during the peak incidence interval in young adults, between ages 20 and 40, the most common cell type is the seminoma. Dysgerminoma of the ovary is

morphologically equivalent to seminoma or germinoma. The relatively rare and clinically indolent spermatocytic seminoma, which represents only 2%–3% of all testicular tumors, occurs usually over the age of 50 (Brodsky, 1991). Spermatocytic seminoma, unlike classic seminoma, does not appear to progress from carcinoma in situ, is not associated with gonadal dysgenesis, and does not occur in association with any other type of germ cell tumor. Embryonal carcinoma appears to recapitulate the earliest phases of embryogenesis, and is interpreted as representing the primitive cell type from which further differentiation gives rise to choriocarcinoma, yolk sac carcinoma, and teratoma (Fig. 60–3). Embryonal carcinoma occurs rarely in pure form, but is relatively common in mixed germ cell tumors. The teratoma results from the differentiation of pluripotent embryonal carcinoma cells and is potentially derived from the three embryonic cell layers (Fig. 60–3). Teratomas of the testis, as pure lesions, represent the second most common germ cell tumor, after yolk sac tumors, in infancy and early childhood. Teratomas are benign unless they are immature and/or are composed of stem cells that have undergone malignant transformation. Stem cells are defined as renewing and replicating cell populations that are capable of differentiation into cells that are distinctive morphologically, functionally, and/or developmentally from the progenitor cells. In contrast to the mature teratoma, the immature teratoma may consist of epithelial tubular structures that are more primitive, stroma with loose spindle cells in a mucopolysaccharide matrix, and primitive myoblastic or neuroblastic elements. In the mature teratoma, the most common pattern includes cystic epithelial areas lined by respiratory, gastrointestinal, or squamous epithelium, with adjacent fibroblastic or smooth muscle stroma and an admixture of cartilaginous and nervous tissues. Choriocarcinoma of the testis consists of both cytotrophoblastic cells and syncytiotrophoblastic giant cells. The syncytiotrophoblastic giant cells stain positively by immunohistochemical methods for the beta subunit of human chorionic gonadotropin (HCG). Both cytotrophoblastic and syncytiotrophoblastic cells stain for placental alkaline phosphatase. These serum tumor markers have been used in morphologic classification, monitoring treatment, and determining prognosis in patients with pure or mixed patterns of testicular neoplasms (Bates, 1991). Sex cord-stromal tumors are frequently benign and constitute about 5% of testicular neoplasms, and include Leydig cell (derived from stroma), Sertoli cell (derived from the sex cord), and granulosa cell tumors, which resemble the ovarian granulosa cell tumor, and presumably develop from fetal precursors of specialized stromal cells, gonadoblastoma, or a mixed pattern of germ cell, sex cord, and stromal tumors. The Sertoli cells line the basement membrane of the seminiferous tubules and are in close proximity to the germinal cells as they pass through various stages of spermatogenesis. The stroma that supports the seminiferous tubules is connected to tissue in which the interstitial, androgen-producing Leydig cells are arranged in clusters. The gonadoblastoma consists of seminoma-like germ cells, and cells derived from the gonadal stroma, including Sertoli cells, granulosa cells, Leydig cells, and/or luteinized stromal cells. Gonadoblastomas occur almost exclusively in dysgenetic gonads. Carcinoma in situ (CIS) of the testis is a preinvasive lesion of the seminiferous tubules that precedes all types of invasive germ cell tumors, with the possible exception of spermatocytic seminoma

1151

1152

PART IV: CANCER BY TISSUE OF ORIGIN

65–69

60–64

55–59

50–54

45–49

40–44

35–39

30–34

25–29

20–24

15–19

10–14

5–9

0–4

0–4 5–9 10– 15– 20– 25– 30– 35– 40– 45– 50– 55– 60– 65– 70– 75– 80– 85+ 14 19 24 29 34 39 44 49 54 59 64 69 74 79 84

Figure 60–1. Age-specific incidence rates (per 100,000) for testicular cancer in US Males 1996–2000. (Source: Ries et al., 2003.)

(Jorgensen et al., 1990). Intratubular neoplasia has been reported in the examination of tissues obtained from patients with infertility and oligospermia, gonadal dysplasia, hypogonadism with cryptorchidism, or in the contralateral testis of patients with invasive germ cell tumors (Scully, 1993). The cumulative risk of a metachronous contralateral invasive germ cell neoplasm, of the same or different cell type, has been estimated to range between 2% and 5% (Fossa et al., 2005). In a cohort study conducted in Denmark, the relative risk of developing a second primary cancer in the contralateral testicle was estimated, at 25 years after diagnosis of the index primary germ cell tumor, to be 24.8 (Osterlind et al., 1991). The cell type of the second primary is generally different from the index primary. The cumulative incidence of carcinoma in situ in the contralateral testis of patients with a previous history of invasive germ cell tumor of the testis has been reported to be 5%–10% (Berthelsen et al., 1982; von der et al., 1986; Reinberg et al., 1991).

DEMOGRAPHIC PATTERNS United States The incidence of testicular cancer in the United States increased 1.5fold since 1975 (Fig. 60–4). The SEER age-adjusted incidence (per 100,000) was 3.7 for whites and 0.45 for blacks in 1975, and 5.7 and 1.6, respectively, in 2000 (Ries et al., 2003). The estimated average

Figure 60–2. Age-specific incidence rates (per 100,000) for testicular cancer in US Males by race, 1995–1999. (Source: Ries et al., 2002.)

annual percent increase in incidence of testicular cancer among US males from 1973 to 1999 was 1.8%, and 3.8% from 1995 to 1999. Despite the increasing incidence, the mortality rate declined by almost 70% (Fig. 60–4). The 5-year survival rate in 1950–1954 was 57%, and increased to 95.7% in 1992–1998. The SEER age-adjusted mortality rate was 0.93 for whites and 0.32 for blacks in 1973, and decreased to 0.31 and 0.51, respectively, in 1999. The age-adjusted mortality rate decreased by 0.4% annually from 1992–1999 (Ries et al., 2002). Within the United States, there was substantial variation in testicular cancer rates by ethnicity. Non-Hispanic Whites had the highest incidence, followed by Hispanics, American Indians/Alaskan Natives, Asian/Pacific Islanders, and African Americans (Table 60–3). Despite the variation in incidence rates, mortality rates were similar, namely between 0.2 and 0.3 per 100,000. The incidence of testicular cancer during the period 1996–2000 was highest among 30–34-year-olds, followed by 35–39-year-olds and 25–29-year-olds (Fig. 60–2). The incidence of testicular cancer was low before puberty, increased dramatically after age 14, peaked at around age 30, and then declined significantly by age 60 (Ries et al.,

Table 60–1. Rank Order of Most Common Cancers in US Males 15–44 Years of Age, SEER Average Annual Incidence Rates (per 100,000) 1992–1998 Rank 1 2 3 4 5

Age(y) 15–19

Rate

Testis Hodgkin disease Leukemia

3.2 2.9

Brain and nervous system Bones and joints

Age(y) 20–24

Rate

Age(y) 25–29

Rate

7.5 4.3

Skin* Testis

2.9

Testis Hodgkin disease Skin*

4.0

6.2

2.2

Leukemia

2.8

Non-Hodgkin lymphoma Hodgkin disease

2.1

Non-Hodgkin lymphoma

2.6

3.0

Brain and nervous system Soft tissue

2.6

Brain and nervous system Leukemia

2.8

Colon and rectum

1.9

6

Non-Hodgkin lymphoma

2.0

7

Skin*

1.3

1.2

12.9 12.4

4.8

Age(y) 30–34

Rate

Skin* Testis

23.0 13.0

Non-Hodgkin lymphoma Colon and rectum

11.8

Age(y) 35–39 Skin* Non-Hodgkin lymphoma Testis

Rate 29.5 14.6 11.2

Age(y) 40–44 Skin* Non-Hodgkin lymphoma Colon and rectum Lung and bronchus

Rate 34.5 17.8 14.0

4.1

Colon and rectum

7.3

Hodgkin disease

3.9

Oral cavity and pharynx

5.3

Oral cavity and pharynx

9.7

Brain and nervous system Leukemia/ oral cavity and pharynx

3.7

Lung and bronchus

5.0

Testis

8.4

3.7

Brain and nervous system

4.3

Kidney and renal pelvis

6.3

Source: Ries et al., 2002. *Excludes basal and squamous cell skin cancers. Includes melanoma, Kaposi sarcoma, and dermatofibrosarcoma.

13.0

1153

Testicular Cancer Table 60–2. Pathologic Classification and Frequency Distribution (%) of Tumors of the Testis in the United States Frequency (%)

germ cell neoplasms

95 40–45 2–3 5–10

Seminoma Spermatocytic seminoma Teratoma Mature Immature With malignant transformation Embryonal carcinoma* Choriocarcinoma Yolk sac (endodermal sinus tumor)†

20–25 £1 <1

non-germinal neoplasms

Rate per 100,000

Nomenclature

5

Gonadal stroma Leydig cell Sertoli cell Androblastoma Gonadoblastoma

1–3

Year of Diagnosis/Death

*Combined histologic patterns may account for 40%–50% of all germ cell tumors; for example, “teratocarcinoma” (embryonal carcinoma with teratoma), and 20% may contain a mixture of seminoma and nonseminoma cell types. † Pure yolk sac tumor is extremely rare in the adult testis, but is classified as a common component (about 30%) of mixed nonseminomatous germ cell tumors.

2003). The two main histologic types of testicular cancer, seminoma and nonseminoma, exhibited different age-specific peaks in incidence. The incidence of nonseminomas peaked at around ages 25–29, whereas the incidence of seminomas peaked at around ages 35–39 (Moller, 1993). Among white men, the incidence of seminoma increased over a period of 25 years since 1973 by 72%, although the rate of increase decreased each successive 5-year calendar interval. In contrast, the incidence of nonseminoma increased by 31.7 percent, and by 1994–1998, the seminoma : nonseminoma rate ratio was 60 : 40 (McGlynn et al., 2003). The 5-year incidence rates in Connecticut for testicular cancer occurring between ages 15 and 44 may also be viewed in relation to birth cohort and the calendar period, 1953–1987 (Table 60–4, Fig. 60–5). For each of the birth cohorts, the peak age-specific interval was 30–34 years, which increased from 5.7 per 100,000 in the 1918–1922 cohort to 12.8 in the 1953–1957 cohort. When comparing the averages of the age-specific rates in the two earliest cohorts (1918–1922

Figure 60–4. Age-adjusted testicular cancer incidence and mortality rates in the United States (SEER), 1973–1999. (Source: Ries et al., 2002.)

and 1923–1927) with those in two later birth cohorts (1938–1942 and 1943–1947), the highest percentages increases were reflected at ages 20–24 (71%) and 15–19 (44%) years; the rates for the age cohorts at 25–29 and 30–34 years were unchanged or failed to exhibit any consistent trend, whereas increases were noted at ages 35–39 (30%) and 40–44 (29%) years. For the combined age group 15–24, the proportionate increase in incidence was composed mainly of nonseminomatous germ cell tumors (Pottern and Goedert, 1986). In a similar analysis of testicular cancer incidence in Denmark, it was concluded that men born just before or during the Second World War appeared to be at a lower risk than expected, based on the generally increasing age-cohort trends (Moller, 1989). The cumulative risk of testicular cancer for the age interval 20–34 years increased 25%–30% in successive 5-year cohorts between 1920 and 1935. In a more recent analysis of US white males using SEER data of birth cohorts between 1945 and 1968, a clear increase in incidence rates with later birth cohorts, and a trend toward a younger peak age of incidence for more recent birth cohorts was found. Furthermore, the increase in incidence with birth cohort has been observed for both seminomas and nonseminomas (McKiernan et al., 1999).

Socioeconomic Status Pre-Neoplastic Germ Cell Spermatocytic Seminoma Carcinoma in Situ

Seminoma

Tumor of Totipotential Cells

Table 60–3. Age-Adjusted SEER Average Annual Incidence Rates (per 100,000), United States 1992–1999, According to Race

Embryonal Carcinoma Extraembyronic

Choriocarcinoma (Trophoblast)

Yolk Sac (Endodermal Sinus, Extra-Embryonic Mesenchyme)

Most studies have shown that testicular cancer occurs more commonly in patients of higher socioeconomic status (e.g., professional and skilled nonmanual occupations), and that the rates are approximately double the incidence in blue collar occupational groups, such as laborers and other unskilled and partly skilled occupations (Graham and Gibson, 1972; Graham et al., 1977; Davies, 1981; Swerdlow et al., 1988). The risk in relation to social class does not vary significantly by age at diagnosis, nor is the social class gradient limited to a

Race/Ethnicity

Intraembyronic

Teratoma/Teratocarcinoma (Ectoderm, Endoderm)

Figure 60–3. Histogenesis of gonadal germ cell tumors.

All races White White Hispanic White non-Hispanic Hispanic American Indian/Alaskan Native Asian/Pacific Islander Black Source: Ries et al., 2002.

Rate 5.0 5.9 3.6 6.5 3.4 2.6 2.1 1.2

1154

PART IV: CANCER BY TISSUE OF ORIGIN

Table 60–4. Age-Cohort-Specific Incidence Rates Per 100,000 for Testicular Cancer in Connecticut Based on Rates in 1935–1987 Age Group at Diagnosis

birth cohort 1918–1922 1922–1927 1928–1932 1933–1937 1938–1942 1943–1947 1948–1952 1953–1957 1958–1962 1963–1967

15–19 0.5 1 0.9 0.6 1 1.7 1.5 1.6 2.8 4

20–24 2.8 2.4 3.8 3.9 4.2 4.7 5.2 7.2 9.8 11.6

25–29 6.8 4.5 3.4 5.6 5.2 7.0 9.9 9.9 12.8

30–34 5.7 9.5 6 6.9 8 7.4 11.3 12.4

35–39 4.7 6.8 5.8 4 7.7 7.3 10.4

40–44 4.8 6.1 4.8 5.9 6.1 8

Source: Pottern and Goedert, 1986.

particular histological type of testicular cancer (Swerdlow and Skeet, 1988; Swerdlow et al., 1991). Thus, although both Ross et al. (1979) and Akre et al. (1996) described a steeper social class gradient for patients with embryonal carcinoma/nonseminoma when compared with seminoma patients, other investigators observed that the social class gradient was more evident in patients with seminoma than with other nonseminomatous histological types (Ross et al., 1979; Akre et al., 1996; Morrison, 1976; Coldman et al., 1982). Similarly, Pollan et al. (2001) found an association in high socioeconomic groups in a study of Swedish men, however, mainly for nonseminomas.

International Patterns Even though the age-adjusted incidence of testicular cancer was relatively low in all populations of the world, there is considerable geographic variation. The highest rates were found in Scandinavia and western European populations, and the lowest in Asia and Africa. Rates are 7–17 times higher in Denmark, Norway, and Germany, compared with India, Japan, or Nigeria (Ferlay and Bray, 2001). According to the IARC, the highest incidence rate during 1993–1997 was found in Denmark, with an age-standardized world population rate of 10.4 per 100,000 person-years and the lowest rates in Africa and Asia, with incidence rates less than 1 per 100,000 person-years (Fig. 60–6). Worldwide, the incidence rates for testicular cancer have been increasing steadily over the past century, particularly among the younger age groups. An annual increase of 2.3%–6.5% was reported from several European countries, Australia, Japan, and New Zealand (Coleman et al., 1993; Adami et al., 1994). The increase occurred for both seminomas and nonseminomas, suggesting a common cause (Moller, 1993; Richiardi et al., 2004).

International studies have also detected a birth cohort effect. Analysis of the incidence rates in three provinces of Canada showed a 53% and 91% increase in the rates of seminomas and nonseminomas, respectively, over the period 1970–1971 to 1994–1995. The average annual percent change in the rates was 1.7% for seminomas and 2.3% for nonseminomas. Using age-period-cohort modeling, the increased risk of both seminomas and nonseminomas were found to follow a birth cohort pattern and differential time trend pattern in most age groups and birth cohorts. Patients with seminomas and nonseminomas were on average 6.2 years and 6.8 years younger, respectively, at diagnosis in 1994–1995, compared with 1970–1971 (Liu et al., 2000). Similarly, Weir et al. (1999) found a steady trend of increased incidence with later birth cohort in Ontario, when examining birth cohorts from 1930–1934 up to 1960–1964. An analogous cohort phenomenon has also been observed in Denmark, Norway and other northern European countries (Moller, 1993; Wanderas et al., 1995; Bergstrom et al., 1996). In an evaluation of testicular cancer incidence data from Poland, East Germany, Norway, Finland, Denmark, and Sweden, Bergstrom et al. (1996) reported that birth cohort was the most powerful determinant for the increased risk. The relative risk of testicular cancer for those born between 1960–1970 as compared to 1900–1910, varied from 3.9 (2.7–5.6) in Sweden to 11.4 (8.3–15.5) in East Germany.

Migration The incidence of testicular cancer varies by migration pattern as well, suggesting an interaction of genetic and environmental factors. Japanese Americans residing in Hawaii or California tend to have incidence rates that are 40%–100% higher than the rates of Japanese residing in Japan; in contrast, Chinese Americans residing in California have lower rates than those registered for the Chinese in Singapore or Hong Kong. In a study of Indian migrants to England and Wales from 1973 to 1985, the risk of death from testicular cancer was lower among individuals of Indian descent than the native population (Swerdlow et al., 1995). Another study conducted in Israel found that independent of country of origin or baseline risk, offspring of immigrants to Israel experienced rates approaching those of the Israeli population, although ethnic differences continued to persist into the second generation. Migrants born in Africa had a decreased risk of testicular cancer as compared to the Israeli population [0.30(0.14,0.61)], whereas migrants born in Europe/America had an increased risk [2.15(1.41,3.27)]. This same trend was observed in second-generation descendents, although the differences lessened somewhat. Israel-born men whose fathers’ were born in Africa still had an RR of 0.47(0.27,0.81) and Israel-born men whose fathers were born in Europe/America had an RR of 1.57(1.06,2.34) (Parkin and Iscovich, 1997).

14

Incidence Rates per 100,000

12

Figure 60–5. Average annual age-specific (15–44) incidence rates (per 100,000), testicular cancer, 5-year birth cohorts, Connecticut 1918–1967. (Source: Pottern and Goedert, 1986.)

1918–1922 1922–1927

10

1928–1932

8

1933–1937 1938–1942

6

1943–1947 1948–1952

4

1953–1957 1963–1962

2

1963–1967

0 15–19

20–24

25–29

30–34

Ages

35–39

40–44

Testicular Cancer

Figure 60–6. Age-standardized testicular cancer incidence rates in selected countries, 2000. (Source: Ferlay and Bray, 2000.)

Occupation The most consistent observation about occupation and testicular cancer incidence is that white collar or professional occupations have been associated with moderately elevated relative risks, in the range of 1.5–2.5. Such observations, however, do not suggest a specific environmental exposure, but may indicate that other risk factors correlated with socioeconomic status and lifestyle are intrinsically linked in the causal pathway (Van den Eeden et al., 1991; Pearce et al., 1987). Excesses of testicular cancer have been reported in association with employment in the armed forces (Dubrow and Wegman, 1983); among aviation support equipment technicians and engineers (Garland et al., 1988; Boice et al., 1999); civilians engaged in repair of F4 Phantom jets (Ducatman et al., 1986); workers in the crude petroleum and natural gas extraction industry (Mills et al., 1984); occupations with exposure to polycyclic aromatic hydrocarbons (Milham, 1983); metal workers (Van den Eeden et al., 1991; Hansen et al., 1996; Rhomberg et al., 1995); the printing industry (McDowall and Balarajan, 1986; Andersen et al., 1999); and the leather industry, particularly leather finishing (Marshall et al., 1990; Boice et al., 1999). In a case-control study of a reported cluster of testicular cancer among leather workers in New York, Marshall et al. (1990) determined that the odds ratio for exposure to leather processing or manufacturing was 7.16 (1.89–27.72). The putative exposures in this investiga-

1155

tion included dimethylformamide-containing dyes used in leather finishing or unspecified solvents and/or dyes. Painters exposed to mixed solvents have also presented with an excess risk of testicular tumors (Guberan et al., 1989). A study of Swedish men employed in 1970 over the period 1971–1989, found that railway stationmasters, metal annealers and temperers, precision toolmakers, watchmakers, construction smiths, and typographers and lithographers exhibited an excess risk of seminomas, whereas concrete and construction workers had an increased risk of nonseminomas (Pollan et al., 2001). Concrete and construction workers were also found to be at a greater relative risk for testicular cancer in Britain (UK Testicular Cancer Study Group, 1994). Metal workers are exposed to metal dusts, including potentially carcinogenic heavy metals; organic solvents, including dimethylformamide, which have been linked to testicular cancer among aircraft repairmen (Hayes, 1997; Ducatman et al., 1986; Haughey et al., 1989; Zhang et al., 1995; Karagas et al., 1989; McDowall and Balarajan, 1986). It has been proposed that solvents increase skin absorption of metal carcinogens such as chromates, which are thought to concentrate in gonadal tissues (Oliver, 1990). In a hospital-based case-control study conducted in Germany, Rhomberg et al. (1995) observed that there was an elevated risk of seminomas in skilled workers exposed to metals, metal dusts, and possibly cutting oils. A study of active-duty servicemen deployed to the Persian Gulf War when compared with nondeployed Gulf War-era servicemen from 1991 to 1996, found that deployment status was not a predictor of testicular cancer. However, men whose occupation was in electronic equipment repair or electrical/mechanical repair had relative risks of 1.56 (1.23–2.00) and 1.26 (1.01–1.58), respectively, as compared to those in other occupations. Men in construction-related trades had an elevated, but nonsignificant, relative risk of testicular cancer of 1.42 (0.93–2.17) (Knoke et al., 1998). Navy veterans of the Korean War with potential exposure to high-intensity radar were not found to be at an increased risk for testicular cancer (Groves et al., 2002). However, a study of military dogs used in Vietnam reported a statistically significant increased risk of testicular neoplasms and of testicular dysfunction (Hayes et al., 1990). The toxicologic risks in military dogs may have provided a surveillance marker of potential human risks in the military personnel who have served in Vietnam. Of particular concern was the intensity and duration of exposure to chemical agents such as Agent Orange, which is a mixture of 2,4dichlorophenoxyacetic acid (2,4-D) and 2,4,5-trichlorophenoxyacetic acid (2,4,5-T). The compound 2,4,5-T was contaminated with small amounts of dioxin, which has been shown to be teratogenic and carcinogenic in experimental animals (Poland and Knutson, 1982). Epidemiologic investigation of the suspected increased risk of testicular cancer observed in Vietnam veterans, however, has not implicated a specific environmental factor related to military service (Tarone et al., 1991; Bullman et al., 1994).

ENVIRONMENTAL EXPOSURES Environmental contaminants and naturally occurring toxins are nearly ubiquitous, as the use of pesticides, preservatives, and other chemicals has increased over the past century. As a result, the role that these factors may play in the increasing incidence of testicular cancer has come into question. DDE (1,1-Dichloro-2,2-bis(p-chlorophenyl)ethylene), a metabolite of the insecticide DDT (2,2-bis*(p-chlorophenyl)-1,1,1trichloroethane), is an environmental contaminant widely detectable in the serum in the United States. In rodents, DDE has been found to inhibit the binding of androgen to its receptor and to block androgen action (Kelce et al., 1995; You et al., 1998). However, a nested casecontrol study undertaken to examine the relation of maternal DDE levels during pregnancy and risk of cryptorchidism, hypospadias, and polythelia among male offspring yielded inconclusive results (Longnecker et al., 2002). Further studies may be required to discern any associations between DDE exposure, congenital urogenital malformations, and risk of testicular cancer.

1156

PART IV: CANCER BY TISSUE OF ORIGIN

The mycotoxin ochratoxin A (OTA), a naturally occurring contaminant of cereals, pig meat, and other foods, and a known genotoxic carcinogen in animals, has also been hypothesized to play a role in the increased incidence of testicular cancer. Studies have demonstrated that animals exposed to OTA, later exhibit OTA in the testis, that OTA can be transferred transplacentally to the fetus and transmitted to newborns in breast milk, and that OTA causes adducts in testicular DNA (Fuchs et al., 1988; Galtier et al., 1979; Fukui et al., 1987; Jonsyn et al., 1995; Zimmerli and Dick, 1995; Ferrufino-Guardia et al., 2000; Kanisawa and Suzuki, 1978; Bendele et al., 1985; Poirier and Beland, 1992). Therefore, it has been hypothesized that consumption of OTAcontaminated foods during pregnancy or childhood could induce testicular precursor lesions, and that growth at puberty could promote these lesions to form invasive testicular tumors. Similar to testicular cancer, exposure to OTA is also associated with high socioeconomic status and poor semen quality (Schwartz, 2002).

HOST AND OTHER FACTORS Due to the early age of onset of testicular cancer and its rarity in older men, it is generally believed that in utero and/or early-life exposures play a role in the development of testicular cancer (Adami et al., 1994; Liu et al., 2000). The dramatic rise in incidence rates after puberty may indicate that gonadotropic and androgenic hormones, including testosterone, stimulate growth and progression of testicular tumors (Moss et al., 1986; Moller and Skakkebaek, 1996; Weir et al., 1998). Indicators of high androgen levels have also been associated with increased incidence, including early puberty, baldness, and severe acne (Petridou et al., 1997; Depue et al., 1983).

Age at Puberty A secular trend toward earlier age at puberty has been hypothesized to play a role in the increased testicular cancer incidence (Moss et al., 1986; Weir et al., 1998). Case-control studies have inferred that later puberty, measured as age started shaving, appearance of body hair, growth spurt, and voice change, is associated with reduced testicular cancer risk. (Weir et al., 1998; Moller and Skakkebaek, 1996). Conversely, early onset of puberty has been hypothesized to be associated with increased risk of testicular cancer (UK Testicular Cancer Study Group, 1994; Moss et al., 1986). However, other studies have not reported any association (Depue et al., 1983); Swerdlow et al., 1989). Authors have speculated that an earlier onset of puberty is accompanied by sustained increased levels of follicle stimulating hormone, luteinizing hormone, and testosterone. The interaction at puberty of endogenous growth factors and testicular carcinoma in situ precursor cells may result in invasive germ cell tumors. The lack of a strong association in case-control studies may have resulted from random misclassification of age of onset of puberty.

Prenatal Exposure to Estrogens Exposure to exogenous estrogens during fetal development will profoundly alter sexual differentiation. Experiments in pregnant rodents and epidemiologic studies have shown that prenatal estrogen administration has resulted in an increased incidence of cryptorchidism, hypospadias, and dysgenetic gonads in the male offspring (McLachlan et al., 1975; Klip et al., 2002; Akre et al., 1999). Subsequent administration of androgens during gestational development has reversed or ameliorated the risk of testicular maldescent (Rajfer and Walsh, 1977). Mice exposed in utero to diethylstilbestrol (DES), a synthetic nonsteroidal, stilbene-derived estrogen, are at increased risk of malignant interstitial cell tumor of the testis and adenocarcinoma of the rete testis (Newbold and McLachlan, 1988; Bullock et al., 1988). DES exposure in utero has been linked to a variety of structural and functional alterations in the human male genital tract. The lesions have ranged from relatively minor structural alterations, such as epididymal or spermatocele cysts, to more major anatomical abnormalities of testicular hypoplasia, cryptorchidism, hypospadias, and microphallus.

The historical use of estrogens, in particular DES, during the period 1945–1960, for threatened abortion and other complications of pregnancy, was a common practice in the United States. It is estimated that there may have been up to 2–4 million mothers exposed to DES, or other nonsteroidal synthetic estrogens, like dienestrol, before the use of such drugs in pregnant women was proscribed in 1971 by the Food and Drug Administration. Exogenous estrogen (and progestin) exposure in utero may also occur as a result of pregnancy testing, or from the inadvertent use of oral contraceptives after conception. Studies of the potential teratogenic effects of estrogen and/or progestin preparations administered during pregnancy may be confounded by a preexisting condition for which the hormone(s) was given, and are required to evaluate the gestational interval, intensity, and duration of exposure. In the fetus, unconjugated estrogen can bind to target organ receptor sites in the male genital tract with resultant antagonistic or inhibitory effects on the pituitary–gonadal axis, and the normal production and distribution of testosterone from the fetal testis. It has been postulated that exogenous estrogen hormone use during the first trimester of pregnancy may be associated with an increased risk of testicular cancer (Henderson et al., 1979; Schottenfeld et al., 1980; Depue et al., 1983; Weir et al., 2000). In those case-control studies suggesting an association, the odds ratios ranged from 2.8–5.3, but in two of the studies (Henderson et al., 1979; Schottenfeld et al., 1980), the confidence intervals overlapped 1.0. In the study of Depue et al. (1983), hormone exposure began in the initial two months of pregnancy; 9 of the 107 cases were exposed in utero, and in five of the mothers, it was reported as a single exposure as a result of a pregnancy test. In none of the above studies was there an interaction between exogenous hormones and cryptorchidism in the cases with testicular cancer. Brown et al. (1986) did not observe any association with exogenous estrogen or progestin hormone use in pregnancy, but noted a statistically significant twofold [RR = 2.4 (1.2–5.1)] increase in risk associated with unusual bleeding or spotting during the index pregnancy, and a significantly lower mean birth weight in the cases than in the controls. The association with unusual bleeding or spotting was independent of use of any medication, and was specific for the index pregnancy. Moss et al. (1986) found a significant association with cryptorchidism, but no association with DES or other hormonal exposures in pregnancy. A recent study found an elevated but nonsignificant risk of testicular cancer with prenatal DES exposure [RR = 3.05 (0.65–22)] (Strohsnitter et al., 2001). High levels of estrogens during the first trimester have been associated with cryptorchidism (Depue et al., 1983; Depue, 1988; Bernstein et al., 1988). Studies of plasma and urinary estrogens have found higher concentrations of estrogens in nulliparous then parous women, and in first pregnancies than in later pregnancies (Trichopoulos et al., 1980; Berstein et al., 1985; Berstein et al., 1986; Panagiotopoulou et al., 1990). In addition, conditions that increase maternal estrogen levels during pregnancy, such as twin pregnancies, primigravid state, and obesity, have been associated with increased risk of testicular cancer (Braun et al., 1995; Rajpert-De Meyts and Skakkebaek, 1993). Henderson et al. (1982) have proposed that elevated levels of maternal bioavailable estrogens, from either exogenous or endogenous sources, exert a teratogenic or “arrested developmental” effect on the maturation of fetal primordial germ cells, which may serve as precursors of neoplastic cells. One or more clones of aberrant germ cells are subsequently stimulated by pre-pubertal rising levels of luteinizing hormone and follicle stimulating hormone. The epidemiology of ovarian germ cell neoplasms appears to parallel that of testicular neoplasms (dos Santos Silva and Swerdlow, 1991). In a population-based case-control study in young women with ovarian germ cell neoplasms conducted in Los Angeles and Seattle, Walker et al. (1988) reported that the mother’s use of exogenous hormones was associated with an increased odds ratio (OR = 3.60, 95% CI:1.2–13.1).

Pre- and Perinatal Characteristics Swerdlow et al. (1987) have suggested that immunological defects during embryogenesis may be of significance in the development of testicular cancer. Maternal-fetal interactions may lead to complica-

1157

Testicular Cancer tions such as preeclampsia, placental insufficiency, retained placenta and placenta accreta, early abortion, and low birth weight (Bulmer, 1992; Wanderas et al., 1998; Brown et al., 1986). Genito-urinary system infection has been suggested as a putative risk factor for testicular cancer (Wanderas et al., 1998; Henderson et al., 1979; Swerdlow et al., 1982; Shu et al., 1995). A prospective study in Norway found that retained placenta was associated with increased odds of testicular cancer [OR 2.8 (1.2–6.4)]. The risk was elevated for both seminomas and nonseminomas. In addition there was a decreased odds of testicular cancer with increased maternal parity (p = 0.03); however, maternal respiratory and genito-urinary infections and type I diabetes mellitus before conception moderately elevated the risk of testicular cancer (Wanderas et al., 1998). Several additional pre- and perinatal factors have also been associated with risk of testicular cancer. Severe nausea during pregnancy, a possible indication of excess maternal estrogens, has been associated with testicular cancer (Henderson et al., 1988; Petridou et al., 1997; Swerlow et al., 1987; Henderson et al., 1979). Neonatal jaundice and low birth weight have been found to be associated with increased risk of testicular cancer in several studies (Akre et al., 1996; Depue et al., 1983; Brown et al., 1986). Wanderas et al. (1998) found neonatal jaundice increased the odds of testicular cancer 3.0 (1.4–6.6)-fold. However, neonatal jaundice may be an indirect marker of another preor perinatal risk factor for testicular cancer. Bleeding, general nausea, and maternal weight during pregnancy have not been found consistently to be associated with risk of testicular cancer (Weir et al., 2000; Brown et al., 1986).

Familial and Hereditary Factors The various risk factors of gonadal dysgenesis, hypogonadism, cryptorchidism, and familial aggregation suggest that genetic influences may be important in the pathogenesis of testicular cancer. Gonadal dysgenesis is a rare genetic disorder of embryogenesis, often with maldescent of the testis and incomplete masculinization of the external genitalia (Table 60–5). A genetic predisposition for testicular cancer is reflected by both familial and bilateral occurrence, younger age of onset in successive generations, and high rates of urogenital developmental anomalies in families prone to testicular cancer (Sonneveld et al., 1999; Heimdal et al., 1996). Brothers of men with testicular cancer have been found to have a sixfold to 13-fold increased risk of developing the disease, whereas the relative risk to fathers and sons is twofold to fivefold (Sonneveld et al., 1999; Westergaard et al., 1996; Heimdal et al., 1996; Dong et al., 2001; Spermon et al., 2001). Among index cases with testicular cancer, the prevalence of a family history of testicular cancer among first-degree relatives has varied from 0.2%–2.2% (Tollerud et al., 1985). Among such relatives there may be associated urogenital anomalies (Kratzik et al., 1991), polythelia or supernumerary nipples (Goedert et al., 1984), or rare genetic syndromes, such as the recessive X-linked ichthyosis (Lykkesfeldt et al., 1991). Cryptorchidism does not appear to be more prevalent among familial cases compared with nonfamilial cases of testicular cancer (Heimdal et al., 1996). Testicular cancer has been reported in non-twin brothers, identical twin brothers, and in fathers and sons (Dieckmann et al., 1987; Swerdlow et al., 1999). Concordance by cell

type, average age at diagnosis, associated cryptorchidism, or bilaterality, was highest for identical twin pairs (Pottern and Goedert, 1986; Patel et al., 1990; Goss and Bulbul, 1990). Kratzik et al. (1991) studied patients with bilateral testicular germ cell tumors, comparing histocompatibility antigen types (HLA) in cases and matched controls. There was an overrepresentation in the cases of HLA-B14, DR5, and DR7; and a decreased frequency in the bilateral testicular germ cell cases of HLA-DR1, DR3, and DR4. The estimates of relative risk in familial cases, or in monozygous twins, underscores the importance of identifying genetic factors in relation to the intrauterine environment. A specific cytogenetic abnormality, an isochromosome of 12p, i (12p), has been described in male germ cell cancers. The i (12p) occurs in gonadal and extra-gonadal tumors histopathologically diagnosed as seminoma, nonseminoma, or teratoma (Dmitrovsky et al., 1990). The chromosome anomaly has been identified in 80%–90% of all testicular germ cell tumors and in carcinoma in situ, and, therefore, has value as a diagnostic biomarker in the analysis of tumors of uncertain histogenesis (Chaganti and Houldsworth, 1998). Multiple copies of the i (12p) marker, which may result in the amplification of a c-ki-ras oncogene, usually characterize tumors with an aggressive growth pattern (Ilson et al., 1991). The appearance of the cytogenetic abnormality may signal the neoplastic transformation process. In general, the formation of an isochromosome in diploid cells leads to the loss of the chromosomal arm not included in the anomaly. In the case of i (12p), this would result in the loss of heterozygosity of the genes on the q arm of chromosome 12. However, Van Kessel et al. (1991) reported that loss of heterozygosity on the q arm of chromosome 12 was not an invariable characteristic of i (12p)-positive testicular germ cell tumors. In addition to the presence of i (12p), nonrandom gains in chromosomes 1, 7, 12, 21, 22, and X have been reported in more than 70% of male germ cell tumors. A deletion in the long arm of chromosome 12, occurring between bands q13 and q22, has been identified in nonseminomatous germ cell tumors. Chromosomal breaks during tumor progression at 1p11, 7q22, and 12p have given rise to characteristic patterns of chromosomal rearrangements (Bosl et al., 1989; Samaniego et al., 1990; Rodriguez et al., 1992; Sinke et al., 1993). In contrast to normal gametogenesis, testicular germ cell tumors, in general, express both parental alleles, which is consistent with the potential of the cell of origin to differentiate into somatic (embryonic) and/or trophoblastic (extraembryonic) imprinted lineages (van Gurp et al., 1994). Genomic imprinting occurs when paternal and maternal chromosomes have different functionality, presumably due to epigentic modification of the genome (e.g., DNA methylation, histone acetylation). A genome-wide linkage search using samples from affected families with no father-son transmission, identified a susceptibility gene for testicular germ-cell tumors at Xq27 called testicular germ cell tumor cell 1 (TGCT1) (p = 0.034). Families with at least one case of bilateral disease were more likely to be linked to Xq27 than families without a bilateral case (p = 0.0002). Results also suggested that this gene may predispose men to cryptorchidism, as a significantly greater proportion of the families with cryptorchidism were linked to this locus (p = 0.03) (Rapley et al., 2000). Individuals with Down syndrome, where the chromosomal basis is trisomy 21, have been reported

Table 60–5. Testicular Tumors in Disorders of Gonadal Differentiation Congenital Disorder

Chromosomal Anomaly

Associated Testicular Tumor

1. Klinefelter syndrome: eunuchoidism, gynecomastia, azoospermia, elevated FSH, decreased testosterone, increased estradiol, increased estradiol/testosterone ratio, small dysgenetic testes 2. Male pseudohermaphroditism cryptorchidism, hypospadias, variable internal genital organs 3. Testicular feminization: female habitus, but without uterus or fallopian tubes; cryptorchidism 4. Mixed gonadal dysgenesis: female habitus, but some degree of virilization; streak gonads

47, XXY 46XY/47XXY mosaicism

Breast cancer: 8 times risk in normal males, gonadoblastoma, invasive germ cell cancer

XY or mosaicism with Y chromosome 46, XY

Gonadoblastoma, invasive germ cell cancer Sertoli cell tumor in dysgenetic gonad

XO/XY

Gonadoblastoma, invasive germ cell tumor

1158

PART IV: CANCER BY TISSUE OF ORIGIN

to be at increased risk of intraepithelial and invasive germ cell tumors (Dieckmann et al., 1997; Satge et al., 1998).

Testicular Trauma Patients with testicular cancer may recall a history of trauma to the affected testicle (Coldman et al., 1982; Pottern et al., 1985; Brown et al., 1987). However, the assessment of risk in case-control studies may be biased by selective recall of instances of trauma of ill-defined severity, or where the trauma called attention to the existing tumor. Severe trauma may require orchiectomy, or may lead to necrosis or infarction and subsequent degeneration and atrophy of gonadal tissue (Stone et al., 1991).

Vasectomy Vasectomy is a common and increasing form of male contraception used in the United States (about 500,000 vasectomies each year) and throughout the world. Following vasectomy, morphologic changes in the human testis have been observed, including focal interstitial fibrosis and thickening of the tunica propria of seminiferous tubules, dilatation of seminiferous tubules, and reduction in the numbers of germ and Sertoli cells. Immunologic studies have shown that 50%–70% of men have elevated antibodies to spermatocytes in the seminal fluid and serum following vasectomy (Jarow et al., 1985). Cohort studies in general have not described a significant association between vasectomy and testicular cancer, although the follow-up interval and statistical power have been limited in the individual studies (Goldacre et al., 1978; Walker et al., 1981; Petitti et al., 1983; Massey et al., 1984; Nienhuis et al., 1992; Rosenberg et al., 1994; Moller et al., 1994). Cale et al. (1990) reported an increased incidence of testicular cancer soon after vasectomy, which suggested either that the neoplasm was present at the time of vasectomy and detected through more careful postoperative surveillance, or that the procedure accelerated the progression of neoplastic disease. Among four case-control studies, the reported odds ratios (95% CI) were 1.1 (0.6 2.0) (Swerdlow et al., 1987); 0.6 (0.3 1.2) (Moss et al., 1986); 1.5 (1.0 2.2) (Strader et al., 1988); and 1.09 (0.77–1.52) (UK Testicular Cancer Study Group, 1994). The available information would allow for a preliminary assessment that there is either no association or a weak association between vasectomy and testicular cancer (West, 1992).

Viral Infection The prominent peak in the incidence of testicular cancer between 25 and 34 years in US whites of higher social class is suggestive of the pattern of Hodgkin disease in young adults from economically advantaged populations (Mueller et al., 1988; Algood et al., 1988). Atrophy of the testicle has been observed to occur after orchitis, but no causal association with a specific infectious agent, such as the mumps paramyxovirus, has been established. Testicular cancer has also not been found to be associated with measles, rubella, human papilloma virus (HPV), toxoplasma, mycoplasma, or chlamydia infection (Mueller et al., 1988; Rajpert-De Meyts et al., 1994). Investigations of an association between testicular cancer and the cytomegalovirus (CMV), Epstein-Barr virus (EBV), varicella (VZV), and herpes simplex virus (HSV) have yielded equivocal results. EBV exposure has been associated with testicular cancer, and EBV expression has been detected in testicular germ cell tumors (Algood et al., 1988; Shimakage et al., 1996). Mueller et al. (1988) found testicular cancer patients had significantly elevated antibody titers to VZV, HSV, and CMV, but not EBV, whereas Akre et al. (1999) and others failed to find a significant association between CMV or EBV infection and testicular cancer (Heinzer et al., 1993; Fend et al., 1995; Rajpert-De Meyts et al., 1994). Initial observations of the occurrence of testicular cancer in men with human immunodeficiency virus (HIV) infection resulted in hypothesizing that the infectious agent was a possible causal factor. Recent epidemiologic studies have failed to establish an association (Rabkin et al., 1991; Reynolds et al., 1993; Goedert et al., 1999).

Obesity Factors that affect the level of endogenous hormones are frequently the focus of testicular cancer etiological research. Obesity is associated with increased levels of the hormones estradiol and estrone, and inversely associated with levels of both total testosterone and sex hormone-binding globulin. The increased levels of estradiol and estrogen are primarily the result of extraglandular conversion of androgen precursors, thus shifting the overall sex-hormonal profile (Zumoff et al., 1990; Mantzoros and Georgiadis, 1995; Akre et al., 2000). In a cohort study of men in Norway, body mass index (BMI) was inversely associated with the risk of testicular cancer (Akre et al., 2000; Petridou et al., 1997). A study in Denmark found a slight trend toward testicular cancer cases having a lower median BMI than controls (Davies et al., 1990). However, other studies on the association between BMI and testicular cancer have reported null findings (Gallagher et al., 1995; Swerdlow et al., 1989). Although the literature is inconsistent, there are several studies that suggest an increased risk in taller men (Gallagher et al., 1995; Brown et al., 1987; Thune and Lund, 1994; Whittemore et al., 1984; Swerdlow et al., 1999). Increasing secular trends in height, primarily due to average increased leg length, may reflect on nutritional factors and pre- and postpubertal growth patterns.

Diet In Japanese males born before 1945, death from testicular cancer peaked between the ages of 30 and 40, whereas in those born after 1946, incidence peaked between the ages of 20 and 29. This apparent birth cohort effect suggests that the causative factors relating to testicular cancer operate early in life and may involve diet. During the 48 years between 1950 and 1998, the average per capita intake of milk and dairy products, meat and eggs in Japan increased 20fold, ninefold and sevenfold, respectively. It has been proposed that the development of testicular cancer in Japanese, as well as in Western men, is related to dietary practices at the time of puberty or earlier. The sharp increase in incidence commonly observed after age 14 is thought to reflect on the insight that testicular carcinogenesis is dependent on endogenous sex steroid hormones that increase significantly after the onset of puberty. In an ecological study of testicular cancer rates in 42 countries and their dietary practices, Ganmaa et al. (2002) found that cheese, animal fats, and milk were highly correlated with the incidence of testicular cancer at ages 20–39. The correlation coefficient was highest when calculated for cheese consumed during the period 1961–1965 (maternal or prepubertal consumption). Furthermore, stepwise multiple regression analysis revealed that milk + cheese (1961–1965) made a significant contribution to the incidence of testicular cancer (Ganmaa et al., 2002). Davies et al. (1996) also observed in a case-control study of diet and testicular cancer, that the cases had consumed significantly more milk during adolescence than controls. However, not all studies have observed an association between dairy products and testicular cancer. A hospital-based case-control study failed to find an association with milk, but did find that increasing total fat, saturated fat, and cholesterol consumption were significantly associated with increasing risk of nonseminoma testicular cancer, and increasing intake of meat and cholesterol were significantly associated with risk for seminomas. Dietary fiber and calcium were inversely related to the risk of nonseminomas (p = 0.02, p = 0.05), and fruit and vegetable intake was not associated with testicular cancer of either histology (Sigurdson et al., 1999). These results are consistent with an ecological study linking increased per capita fat consumption and testicular cancer rates, and a case-control study that failed to find a protective effect of fruits or vegetables on testicular cancer risk (Armstrong and Doll, 1975; Davies et al., 1996).

Physical Activity Although several medical risks for testicular cancer have been established, including cryptorchidism and inguinal hernia, the relation between lifestyle and testicular cancer risk is less well defined. Phys-

Testicular Cancer ical activity is associated with a protective effect for several cancers including colon and breast, possibly through heightened immunity (increased number and/or activity of macrophages and natural killer cells) or modulation of hormone levels (Srivastava and Kreiger, 2000; Shephard and Shek, 1995). Results in relation to testicular cancer have been conflicting. A cohort study in Norway failed to find any evidence for an association between physical activity and testicular cancer (Thune and Lund, 1994), similarly in the cohort study by Paffenbarger et al. (1992) while the UK Testicular Cancer Study Group (1994) found in a case-control study a decreased trend in risk with increasing exercise and an increased trend in risk for sedentary individuals. Case-control studies in Canada, Gallagher et al. (1995) found that a moderate to high level of recreational activity was associated inversely with testicular cancer risk [OR 0.6 (0.5–0.8)], whereas results from Srivastava and Kreiger (2000) suggested that relatively high frequency of participation in moderate and strenuous recreational activity in the mid teens may have an adverse effect on the risk of testicular cancer. Men who participated in moderate activity of greater than five times a week compared with three times or less a month had an OR of 2.36 (1.20–4.64) for testicular cancer, and participation in strenuous activity of greater than five times a week compared with less than once a month generated an OR of 2.58 (1.14–5.85). The authors proposed that the increased risk could be the result of altered androgen levels. Acute increases in androgens, primarily testosterone, have been demonstrated after exercise (Kuoppasalmi et al., 1980; Sutton et al., 1973; Mantzoros and Georgiadis, 1995; MacConnie et al., 1986; Hackney et al., 1995). Increases in testosterone are thought to be the result of decreased metabolic hormone clearance and not an increase in hormone production (Sutton et al., 1973). In summary, there is insufficient evidence to arrive at any conclusion on the mechanistic basis for a putative relationship with a pattern of physical activity (IARC, 2002).

Smoking and Alcohol Maternal smoking during pregnancy has not been shown to increase risk of testicular cancer, despite the parallel association between smoking and low birth weight, and testicular cancer and low birth weight (Moller and Skakkebaek, 1997; Moller and Westergaard, 1998; Brown et al., 1986). Results from some studies even suggested that maternal smoking may decrease testicular cancer risk. Two casecontrol studies have found a reduced, but nonsignificant, risk of testicular cancer among men whose mothers smoked at least 12 cigarettes per day during their pregnancy (Weir et al., 2000; Brown et al., 1986). The hypothesized anti-estrogenic effect of smoking could be the basis for consideration of protection (Baron et al., 1990; Bernstein et al., 1989; Petridou et al., 1990). In other studies an increased risk of testicular cancer in smokers was inferred (Brown et al., 1987; Srivastava and Kreiger, 2000). The absence of any consistent association in published studies, or of a dose-response effect, does not allow for any substantive conclusion (Weir et al., 2000). Studies of alcohol consumption during pregnancy have also failed to report a significant association with testicular cancer (Weir et al., 2000; Brown et al., 1986).

PATHOGENESIS Gonadal Embryogenesis The human testis is composed of a system of tubules for the production and transport of sperm, and of clusters of interstitial or Leydig cells between the tubules that produce both androgenic and estrogenic steroids. The basic histological components of the tubules are the germ cells and Sertoli cells. Both the Leydig and the Sertoli cells are derived from the mesenchymal tissue of the urogenital ridge, whereas the primordial germ cells are derived from primitive ectodermal cells that originate from the yolk sac, and migrate into the urogenital ridge by the sixth week of gestation. At this stage of development, the gubernaculum first appears as a ridge of mesenchymal tissue extending from the genital ridge through an opening in the musculature of the ante-

1159

rior abdominal wall to the site of genital swellings and the anlage of the fetal scrotum. At fertilization, the genome of the zygote contains both paternally and maternally imprinted chromosomes. Sexual organization of the human gonad first becomes apparent with the appearance of seminiferous cords in the fetal testis by the seventh week of gestation. During gametogenesis the imprinting pattern of genes in differentiating testicular germ cells is changed to a completely paternal pattern (van Gurp et al., 1994). The genetic control of testicular differentiation and spermatogenesis resides on the Y chromosome, which also contains genes that influence stature. Morphological development of the testis is generally completed by the end of the 12th week, whereas descent of the testis from the abdominal cavity to the scrotum occurs during the remaining months of gestation (Jost and Magre, 1988). In the human embryo the actual movement of the testis from the abdominal cavity through the inguinal canal and into the scrotum usually occurs during the seventh month. Hormonal, neurogenic, and mechanical factors are believed to control the process of testicular descent. During the eighth week of gestation in the male embryo, the fetal testis begins to secrete the Müllerian-inhibiting factor (MIF) and testosterone. MIF, which is secreted by the fetal Sertoli cells, causes regression of the Müllerian ducts and initiates descent to the inguinal region. By causing regression of Müllerian ducts in the male fetus, MIF permits differentiation of the male phenotype by preventing development of the female genitalia. MIF has been demonstrated to inhibit epidermal growth factor activity, and as a polypeptide growth modulator has structural and functional homology to the transforming growth factor b (Liu and Oliff, 1991). Testosterone, which is synthesized and secreted by the fetal Leydig cells, is converted to dihydrotestosterone (DHT) by the 5a-reductase enzyme, and both androgenic hormones control the transinguinal phase of testicular descent into the scrotum. DHT is the active androgenic hormone that induces differentiation of the male external genitalia. Movement of the gastrointestinal organs into the abdominal cavity causes an increase in intra-abdominal pressure, which causes physiological herniation of the processus vaginalis through the ventral abdominal wall along the course of each gubernaculum. Shortening of the gubernaculum, which may be induced by androgens, anchors the fetal testis to the inguinal region. Neurogenic factors may contribute to descent of the human testes; in the neonatal rat, transection of the genitofemoral nerve prevents growth and enlargement of the gubernaculum, and interferes with transinguinal movement of the testes. After 40 gestational weeks, descent is generally completed, and the gubernaculum is resorbed and shrinks to a fibrous remnant (Hutson, 1985; Moul and Belman, 1988).

Gonadal Dysgenesis and Cryptorchidism Not all mammals have scrotal testes, but presumably the migration of the testes into an external scrotum conferred some evolutionary advantage on the species with respect to fertility and the rate of spontaneous mutation. Cryptorchidism, or failure of normal testicular descent, may result in the male gonad being: intra-abdominal, typically located just above the internal inguinal ring (10%); intracanalicular, or beyond the internal inguinal ring (15%–20%); high scrotal, with limited range of motion so that it can retract into the groin but not past the internal inguinal ring (40%); and obstructed, or where there is a fascial barrier between the inguinal pouch, just distal to the external inguinal ring, and the inlet of the scrotum (30%). It is important to distinguish the cryptorchid testis from the absent testis (anorchia), and the temporarily retracted normal testis. Maldescent of the testis occurs with increased frequency in association with congenital urogenital malformations, or namely disorders that impair virilization or prevent the development of physiologic intra-abdominal pressure. The incidence of cryptorchidism in the United States is about 0.7%–0.8% (Rajfer et al., 1986). Studies of the incidence of cryptorchidism in England and Wales, comparing the rates in the late 1950s with those in the mid 1980s, have concluded that there has been an increase of 65%–100% (Chilvers et al., 1984; Jackson et al., 1986). In the Collaborative Perinatal Study sponsored by the NIH (United States), the incidence of cryptorchidism up to 1 year of age during the

1160

PART IV: CANCER BY TISSUE OF ORIGIN

period 1959–1965 was estimated to be 1.1% (Myrianthopoulos and Chung, 1974). During the period 1968–1977, the survey of all hospitals in the five counties of Atlanta, Georgia, described a significant linear trend for the increase in the reported incidence of cryptorchidism and hypospadias in white males up to 1 year of age. Cryptorchidism is diagnosed among 10%–18% of males with hypospadias (Harris, 1990). Cryptorchidism is about twice as common in premature as in full-term infants, and is commonly associated with an inguinal hernia. Heinonen et al. (1977) reported that there was a threefold excess risk of cryptorchidism in United States whites compared with blacks. A case-control study in Denmark found that maternal age above 30 years was associated with an odds ratio of 1.9 (1.2–3.0) for cryptorchidism. In addition, there was a significant trend of increased risk for cryptorchidism with lower birth weight (p = 0.001). Moller and Skakkebaek (1997) also found that maternal age above 30 and low birth weight (below 2500 g) were associated with testicular cancer, with odds ratios of 2.0 (1.2–3.6) and 2.6 (1.1–5.9), respectively. The common risk factors of low birth weight and high maternal age for both cryptorchidism and testicular cancer suggests a common developmental pathway (Moller and Skakkebaek, 1997). The incidence of cryptorchidism is increased in disorders of pituitary gonadotropin deficiency (e.g., Laurence-Moon-Biedl (polydactyly, obesity, hypogonadism), Kallmann (hypogonadotropic hypogonadism and anosmia), and Prader-Willi (cryptorchidism, obesity, deletion of chromosome 15) syndromes); neural tube disorders such as myelomeningocele; syndromes involving abnormalities of testosterone biosynthesis, such as the syndrome of androgen resistance (testicular feminization) and deficiency of 5-a-reductase activity; and gonadal dysgenesis (Palmer, 1991). The cryptorchid gonad is generally accompanied by significant structural and functional abnormalities. The gonad is smaller than normal, with microscopic areas of hypoplastic tubules and decreased spermatogenesis. Eventually the tubules appear as dense cords of hyaline connective tissue outlined by prominent basement membranes. There is evidence of delayed maturation of Sertoli cells, which synthesize estrogen under stimulation by the follicle stimulating hormone. The interstitial stroma may appear quite cellular, and is characterized by hyperplasia of Leydig cells. Cryptorchidism is associated with defective spermatogenesis, and in unilateral cryptorchidism, spermatogenesis may be decreased in the normally descended testis. Commonly, infants and pre-pubertal children with cryptorchidism manifest functional abnormalities in the hypothalamic-pituitary-gonadal axis. Endocrine dysfunction may consist of decreased blood levels of testosterone and luteinizing hormone (LH), increased levels of FSH, and/or a blunted response to stimulation testing with gonadotropin releasing hormone (Job et al., 1979). Clinical studies had estimated that the risk of testicular cancer in men with unilateral or bilateral cryptorchidism was 20–40 times higher than normally expected. During the past 20 years, case-control and cohort epidemiologic studies have established that the relative risk of testicular cancer in men with undescended testes was in the range of 2.5 to 11.4 (Table 60–6), and that cryptorchidism accounted for approximately 10% of the incident cases of testicular cancer (Schottenfeld et al., 1980; UK Testicular Cancer Study Group, 1994). Of those patients with a history of cryptorchidism and testicular cancer, approximately 10%–25% have been reported to develop cancer in the contralateral, normally descended gonad (Gehring et al., 1974; Batata et al., 1982). The undescended testis is exposed to temperatures in the range of 35–37°C, compared with that in the normal location in the scrotum, or approximately 33°C. Is it the malposition of the testis and its greater exposure to thermal and physical trauma, and/or a primary malformation with hypoplastic germinal epithelium and endocrine dysfunction that provide the pathophysiologic mechanism for tumorigenesis? It has been observed that the higher the undescended testicle is located, the greater the risk of testicular cancer (Kincade, 1999) and the more difficult the surgical repair. It has also been noted that the risk of testicular cancer in patients with cryptorchidism increased with increasing age of orchiopexy (Pottern et al., 1985). A summary estimate of the odds ratio for testicular cancer in the ipsilateral cryptorchid testis compared with the contralateral

descended testis, was 2.7-fold higher for those undergoing orchiopexy after age 10 years, when compared with patients who were treated before 10 years. Studies in men without testicular cancer, 15–30 years after unilateral orchiopexy, have shown persistent histologic abnormalities in spermatogenesis (Lipshultz et al., 1976). In a populationbased case-control study conducted in the United Kingdom, it was inferred that surgical correction of unilateral cryptorchidism before the age of 10 years was an effective preventive intervention for testicular cancer in young men (UK Testicular Cancer Study Group, 1994). The potential reversibility in risk with early surgical correction, and limited risk to the contralateral testis in unilateral cryptorchidism, would argue in support of the hypothesis that the microenvironment of the undescended testis is a determining pathogenic factor (Herrinton et al., 2003).

PREVENTIVE MEASURES The classic presentation for testicular cancer is a painless testicular mass, although some patients present with pain, swelling, or hardness in the scrotum. These changes are most commonly found during selfexamination, after testicular trauma, or by a sexual partner (Bosl and Motzer, 1997). Indications of metastasis to the para-aortic lymph nodes include swelling of the lower extremities and back pain, whereas indications of pulmonary metastases include cough, hemoptysis, or dyspnea (Dearnaley et al., 2001). Ultrasonography aids in the confirmation of a scrotal mass, and is nearly 100% accurate in distinguishing between intratesticular and extratesticular swellings. All intratesticular masses are considered malignant until proven otherwise. Given the low incidence and high cure rate of testicular cancer, there is presently no evidence to recommend routine testicular selfexamination. However, it is generally agreed that all men should be educated regarding the early symptoms, and that men in high-risk groups, including those with cryptorchidism, testicular atrophy or dysgenesis, or family history, should be examined by testicular ultrasound at least once a year and perform testicular self-examination monthly (Kinkade, 1999; Berkmen and Alagol, 1998; Wardle et al., 1994).

FUTURE DIRECTIONS Testicular cancer provides a unique biological context for investigating mechanisms of neoplastic transformation in totipotential germ cells. Germ cell tumors are divided morphologically into seminoma and nonseminoma subgroups, the latter exhibiting mixed patterns of embryonal and extra-embryonal differentiation. Cytogenetic and molecular genetic studies are scanning the genome of tumors to identify consistent areas of loss or amplification of chromosomal regions. Virtually all tumors, including in situ or intratubular carcinomas of the testis, exhibit an increased copy number of 12p. However, the molecular genetic studies have not identified consistent genomic rearrangements that distinguish each morphologic entity. The mode of inheritance of a putative germ cell tumor gene is uncertain. Analyses conducted up until now have not been able to ascertain whether a single major autosomal gene model is deterministic, or whether there is evidence for an X-linked mode of inheritance as suggested by enhanced risks in siblings of cases. Differences in age at diagnosis between father and son cases suggest genetic anticipation, but validation awaits elucidation of genetic mechanisms. An important insight may be provided in investigating familial cryptorchidism and related congenital urogenital malformations. The intrauterine environment and the onset of puberty will continue to challenge research inquiry about the key events associated with the natural history of testicular cancer. Excessive levels of estrogens in the fetus, relative to testosterone, have been shown to induce male reproductive organ malformations. The peak years of risk for testicular cancer begin after age 15 and continue over a period of 30 years. The limits of normal puberty occur between 9 and 14 years. During pubertal development the testes increase in size, coinciding with growth of

Table 60–6. Summary of Case-Control and Cohort Studies of Cryptorchidism and Testicular Cancer Case-Control Studies History of Cryptorchidism Authors (year of publication)

No. Studied

Cases %

No Studied

Morrison (1976)

596

2.9

602

0.3

8.8 (2.3–56.3)

Henderson et al. (1979)

79

12.7

79

2.5

5.0

Schottenfeld et al. (1980)

190

11.6

166 142

3.6 (hospital) 4.9 (neighborhood)

Pottern et al. (1985)

271

9.2

259

2.7

3.7 (1.6–8.6)

Moss et al. (1986)

217

11.1

223

2.7

4.5 (1.7–13.7)

9.1 79 1.4 (unmatched analysis)

7.3 (0.8–62.3)

Gershman and Stolley (1988)

79

Controls %

Odds Ratio (95% confidence interval)

3.5 (1.34, 8.10) 2.5 (1.02, 5.68)

Strader et al. (1988)

333

12.0

675

2.2

5.9 (3.4–10.2)

Haughey et al. (1989)

247

12.6

247

2.4

5.2 (2.4–32.5)

United Kingdom Testicular Cancer Study Group (1994)

794

8.2

794

2.1

3.8 (2.2–6.5)

Gallagher et al. (1995)

510

10.2

996

3.1

3.5 (2.2–5.7)

Comments History of inguinal hernia operation was independent risk factor: odds ratio was 2.9 (1.3–7.0). P = 0.02. In cases with unilateral cryptorchidism, the risk of testicular cancer in the undescended testis was more than four times greater than in the contralateral testis. In 21% of cases with unilateral cryptorchidism, the testicular cancer was diagnosed in the contralateral testis. Inguinal hernia was not a significant risk factor. Men without cryptorchidism who underwent an inguinal herniorrhaphy experienced a non-significantly elevated risk [1.3 (0.6–2.5)]. Among those with unilateral cryptorchidism, the risk was increased sixfold for the homolateral testis; no increase in risk was observed for the contralateral testis. The odds ratio associated with cryptorchidism was 8.3 (2.8–32.9) when based on sons’ responses, compared with 4.5, when measured from mothers’ responses. There was no independent association with inguinal hernia [odds ratio for hernia was 1.3 (0.8–2.1)]. Significant risk noted for premature birth in the cases. In the matched analysis based on discordant pairs the odds ratio was 6.0 (0.5–69.3). The odds ratios were similar for seminomatous and nonseminomatous germ cell tumors. The risk of cancer in the cryptorchid testis was 8.0 (4.2–15.3); in the descended contralateral testis of men with cryptorchidism, the risk was 1.6 (0.6–4.1), relative to the risk in men without cryptorchidism. Independent associations of “testis-related” abnormalities were also noted: low sperm count, fertility problems, and atrophic testis. In logistic regression analysis, cryptorchidism and atrophy were significant. History of inguinal hernia was a significant risk factor after controlling for age and education [odds ratio was 2.26 (1.3–4.0)]. With bilateral undescended testes, the odds ratio was 5.9; whereas, with unilateral undescended testis, the odds ratio was 2.7. Twelve of the 46 cases (26%) with unilateral undescended testis had cancer diagnosed in the contralateral testis (odds ratio, 1.4, compared with 4.0 for the risk to the cryptorchid testis). History of inguinal hernia diagnosed before the age of 15 years was significantly associated with risk of testicular cancer [odds ratio was 2.6 (1.3–5.3)]. Inguinal hernia requiring surgery was significantly associated with risk of testicular cancer [OR 2.0 (1.3–2.9)], as was hydrocele [OR 2.6 (1.4–5.1)].

Cohort Studies Authors (year of publication)

Cryptorchid Study Group

Control Group

Relative Risk (95% confidence interval)

Giwercman et al. (1987)

6/506 = 11.9 per 1000

1.3 cancers expected The lifetime risk in the Danish male population was estimated at 0.5%

4.7 (1.7–10.2)

Benson et al. (1991)

2/224 = 8.9 per 1000

0.2 cancers expected

11.4 (1.4–41.1)

Pinczowski et al. (1991)

4/2918 = 1.4 per 1000

0.54 cancers expected

7.4 (2.0–19.0)

Comments Expected number of testicular cancers based on Danish Cancer Registry. In men with unilateral cryptorchidism, the relative risk was 2.4 (0.3– with bilateral cryptorchism, the relative risk was 8.7); 9.3 (2.6–23.7). The maximal standardized incidence ratio was in the age group, 30 to 39 years [8.3 (2.7–19.5)] Expected number of testicular cancers based on Olmsted County/Rochester, MN age-specific cancer incidence rates. Prevalence of cryptorchidism was estimated to be 8.6 per 1000 male births. Three percent of cryptorchid males had hypospadias. When compared with all other male infants, the relative risk for inguinal hernia in cryptorchid males was significantly increased (4.4). Follow-up (25,360 person-years) was achieved by record linkage to the Swedish Cancer Registry. A comparison cohort of 30,199 males operated on for an inguinal hernia did not evidence an increased risk of testicular cancer [1.1 (0.4–2.2)]. Three of the four testicular cancers occurred in males with bilateral intraabdominal testes.

1162

PART IV: CANCER BY TISSUE OF ORIGIN

Table 60–7. Histology of Primary Germ Cell Tumors and Frequency of Serum Tumor Markers* Presence of Tumor Markers by Cell Type % Tumor Type All germ cell tumors Seminoma Nonseminomatous germ cell tumors Embryonal cell carcinoma Teratocarcinoma Teratoma Choriocarcinoma Yolk sac

Frequency by Cell Type %

AFP

HCG

100 42 58

50–75 0 65

40–60 9 56

26

70

60

26 5 1 <1

64 37 0 75

57 25 100 25

AFR, alpha-fetoprotein; HCG, human chorionic gonadotropin. *Table recreated from Kincade, 1999.

the seminiferous tubules, onset of spermatogenesis, mitotic activity in the Sertoli cells with rising levels of estrogen, and increased production of testosterone. Concurrent with the release of inhibition of gonadotropic releasing hormone are the pulsatile increases in luteinizing hormone and follicle stimulating hormone, and physiologic interactions of sex steroid hormones, growth hormone and insulin-like growth factor I. At present, we may conjecture that these are important triggering events upon a substrate of intraepithelial precursor neoplastic cells. From the 1950s to the late 1980s the incidence of testicular cancer increased steadily in United States white males, but remained at a much lower level in black males. Since the early 1990s the rate of increase diminished, and was relatively flat between ages 20 and 34 years. As discussed in the previous chapter, the incidence pattern for prostate cancer is dramatically different. The construct of a causal hypothesis should reflect on the population distribution of risk factors and molecular events and should be guided by and reconciled with the salient descriptive patterns, nationally and internationally, and in racial and ethnic groups. References Adami HO, Bergstrom R, Mohner M, et al. 1994. Testicular cancer in nine northern European countries. Int J Cancer 59(1):33–38. Akre O, Ekbom A, Hsieh CC, Trichopoulos D, Adami HO. 1996. Testicular nonseminoma and seminoma in relation to perinatal characteristics. J Natl Cancer Inst 88(13):883–889. Akre O, Lipworth L, Tretli S, et al. 1999. Epstein-Barr virus and cytomegalovirus in relation to testicular-cancer risk: A nested casecontrol study. Int J Cancer 82(1):1–5. Akre O, Lipworth L, Cnattingius S, Sparen P, Ekbom A. 1999. Risk factor patterns for cryptorchidism and hypospadias. Epidemiology 10(4):364–369. Akre O, Ekbom A, Sparen P, Tretli S. 2000. Body size and testicular cancer. J Natl Cancer Inst 92(13):1093–1096. Algood CB, Newell GR, Johnson DE. 1988. Viral etiology of testicular tumors. J Urol 139(2):308–310. Andersen A, Barlow L, Engeland A, Kjaerheim K, Lynge E, Pukkala E. 1999. Work-related cancer in the Nordic countries. Scand J Work Environ Health 25 Suppl 2:1–116. Armstrong B, Doll R. 1975. Environmental factors and cancer incidence and mortality in different countries, with special reference to dietary practices. Int J Cancer 15(4):617–631. Baron JA, La Vecchia C, Levi F. 1990. The antiestrogenic effect of cigarette smoking in women. Am J Obstet Gynecol 162(2):502–514. Batata MA, Chu FC, Hilaris BS, Whitmore WF, Golbey RB. 1982. Testicular cancer in cryptorchids. Cancer 49(5):1023–1030. Bates SE. 1991. Clinical applications of serum tumor markers. Ann Intern Med 115(8):623–638. Bendele AM, Carlton WW, Krogh P, Lillehoj EB. 1985. Ochratoxin A carcinogenesis in the (C57BL/6J X C3H)F1 mouse. J Natl Cancer Inst 75(4):733–742. Benson RC, Jr., Beard CM, Kelalis PP, Kurland LT. 1991. Malignant potential of the cryptorchid testis. Mayo Clin Proc 66(4):372–378.

Bergstrom R, Adami HO, Mohner M, et al. 1996. Increase in testicular cancer incidence in six European countries: A birth cohort phenomenon. J Natl Cancer Inst 88(11):727–733. Berkmen F, Alagol H. 1998. Germinal cell tumors of the testis in cryptorchids. J Exp Clin Cancer Res 17(4):409–412. Bernstein L, Pike MC, Ross RK, Judd HL, Brown JB, Henderson BE. 1985. Estrogen and sex hormone-binding globulin levels in nulliparous and parous women. J Natl Cancer Inst 74(4):741–745. Bernstein L, Depue RH, Ross RK, Judd HL, Pike MC, Henderson BE. 1986. Higher maternal levels of free estradiol in first compared to second pregnancy: Early gestational differences. J Natl Cancer Inst 76(6):1035–1039. Bernstein L, Pike MC, Depue RH, Ross RK, Moore JW, Henderson BE. 1988. Maternal hormone levels in early gestation of cryptorchid males: A casecontrol study. Br J Cancer 58(3):379–381. Bernstein L, Pike MC, Lobo RA, Depue RH, Ross RK, Henderson BE. 1989. Cigarette smoking in pregnancy results in marked decrease in maternal hCG and oestradiol levels. Br J Obstet Gynaecol 96(1):92–96. Berthelsen JG, Skakkebaek NE, von der MH, Sorensen BL, Mogensen P. 1982. Screening for carcinoma in situ of the contralateral testis in patients with germinal testicular cancer. Br Med J (Clin Res Ed) 285(6356):1683–1686. Boice JD, Jr., Marano DE, Fryzek JP, Sadler CJ, McLaughlin JK. 1999. Mortality among aircraft manufacturing workers. Occup Environ Med 56(9):581–597. Bosl GJ, Dmitrovsky E, Reuter VE, et al. 1989. Isochromosome of the short arm of chromosome 12: Clinically useful markers for male germ cell tumors. J Natl Cancer Inst 81(24):1874–1878. Bosl GJ, Motzer RJ. 1997. Testicular germ-cell cancer. N Engl J Med 337(4): 242–253. Braun MM, Ahlbom A, Floderus B, Brinton LA, Hoover RN. 1995. Effect of twinship on incidence of cancer of the testis, breast, and other sites (Sweden). Cancer Causes Control 6(6):519–524. Brodsky GL. 1991. Pathology of testicular germ cell tumors. Hematol Oncol Clin North Am 5(6):1095–1126. Brown LM, Pottern LM, Hoover RN. 1986. Prenatal and perinatal risk factors for testicular cancer. Cancer Res 46(9):4812–4816. Brown LM, Pottern LM, Hoover RN. 1987. Testicular cancer in young men: The search for causes of the epidemic increase in the United States. J Epidemiol Community Health 41(4):349–354. Buetow SA. 1995. Epidemiology of testicular cancer. Epidemiol Rev 17(2): 433–449. Bullman TA, Watanabe KK, Kang HK. 1994. Risk of testicular cancer associated with surrogate measures of Agent Orange exposure among Vietnam veterans on the Agent Orange Registry. Ann Epidemiol 4(1):11–16. Bullock BC, Newbold RR, McLachlan JA. 1988. Lesions of testis and epididymis associated with prenatal diethylstilbestrol exposure. Environ Health Perspect 77:29–31. Bulmer JN. 1992. Immune aspects of pathology of the placental bed contributing to pregnancy pathology. Baillieres Clin Obstet Gynaecol 6(3):461–488. Cale AR, Farouk M, Prescott RJ, Wallace IW. 1990. Does vasectomy accelerate testicular tumour? Importance of testicular examinations before and after vasectomy. BMJ 300(6721):370. Chaganti RS, Houldsworth J. 1998. The cytogenetic theory of the pathogenesis of human adult male germ cell tumors. Review article. APMIS 106(1):80–83. Chilvers C, Pike MC, Forman D, Fogelman K, Wadsworth ME. 1984. Apparent doubling of frequency of undescended testis in England and Wales in 1962–81. Lancet 2(8398):330–332. Coldman AJ, Elwood JM, Gallagher RP. 1982. Sports activities and risk of testicular cancer. Br J Cancer 46(5):749–756. Coleman MP, Esteve J, Damiecki P, Arslan A, Renard H. 1993. Trends in cancer incidence and mortality. IARC Sci Publ (121):1–806. Davies JM. 1981. Testicular cancer in England and Wales: Some epidemiological aspects. Lancet 1(8226):928–932. Davies TW, Prener A, Engholm G. 1990. Body size and cancer of the testis. Acta Oncol 29(3):287–290. Davies TW, Palmer CR, Ruja E, Lipscombe JM. 1996. Adolescent milk, dairy product and fruit consumption and testicular cancer. Br J Cancer 74(4): 657–660. Dearnaley D, Huddart R, Horwich A. 2001. Regular review: Managing testicular cancer. BMJ 322(7302):1583–1588. Depue RH, Pike MC, Henderson BE. 1983. Estrogen exposure during gestation and risk of testicular cancer. J Natl Cancer Inst 71(6):1151–1155. Depue RH. 1988. Cryptorchidism, and epidemiologic study with emphasis on the relationship to central nervous system dysfunction. Teratology 37(4):301–305. Dieckmann KP, Bauer HW. 1987. Carcinoma-in-situ of testis in patients with assumed extragonadal germ-cell tumours. Lancet 2(8570):1265.

Testicular Cancer Dieckmann KP, Rube C, Henke RP. 1997. Association of Down’s syndrome and testicular cancer. J Urol 157(5):1701–1704. Dmitrovsky E, Bosl GJ, Chaganti RS. 1990. Clinical and genetic features of human germ cell cancer. Cancer Surv 9(2):369–386. Dong C, Lonnstedt I, Hemminki K. 2001. Familial testicular cancer and second primary cancers in testicular cancer patients by histological type. Eur J Cancer 37(15):1878–1885. dos Santos Silva, Swerdlow AJ. 1991. Ovarian germ cell malignancies in England: Epidemiological parallels with testicular cancer. Br J Cancer 63(5):814–818. Dubrow R, Wegman DH. 1983. Setting priorities for occupational cancer research and control: Synthesis of the results of occupational disease surveillance studies. J Natl Cancer Inst 71(6):1123–1142. Ducatman AM, Conwill DE, Crawl J. 1986. Germ cell tumors of the testicle among aircraft repairmen. J Urol 136(4):834–836. Fend F, Hittmair A, Rogatsch H, Gredler E, Obrist P, Mikuz G. 1995. Seminomas positive for Epstein-Barr virus by the polymerase chain reaction: Viral RNA transcripts (Epstein-Barr-encoded small RNAs) are present in intratumoral lymphocytes but absent from the neoplastic cells. Mod Pathol 8(6):622–625. Ferlay J, Bray F. GLOBOCAN 2000: Cancer Incidence, Mortality, and Prevalence Worldwide. 2001. Lyon, France: International Agency for Research on Cancer. Ferrufino-Guardia EV, Tangni EK, Larondelle Y, Ponchaut S. 2000. Transfer of ochratoxin A during lactation: Exposure of suckling via the milk of rabbit does fed a naturally-contaminated feed. Food Addit Contam 17(2):167–175. Fossa SD, Chen J, Schonfeld SJ, et al. 2005. Risk of contralateral testicular cancer: A population-based study of 29,515 U.S. men. J Natl Cancer Inst 97:1056–1066. Fuchs R, Appelgren LE, Hult K. 1988. Distribution of 14C-ochratoxin A in the mouse monitored by whole-body autoradiography, Pharmacol Toxicol 63(5):355–360. Fukui Y, Hoshino K, Kameyama Y, Yasui T, Toda C, Nagano H. 1987. Placental transfer of ochratoxin A and its cytotoxic effect on the mouse embryonic brain. Food Chem Toxicol 25(1):17–24. Gallagher RP, Huchcroft S, Phillips N, et al. 1995. Physical activity, medical history, and risk of testicular cancer (Alberta and British Columbia, Canada). Cancer Causes Control 6(5):398–406. Galtier P, Charpenteau JL, Alvinerie M, Labouche C. 1979. The pharmacokinetic profile of ochratoxin A in the rat after oral and intravenous administration. Drug Metab Dispos 7(6):429–434. Ganmaa D, Li XM, Wang J, Qin LQ, Wang PY, Sato A. 2002. Incidence and mortality of testicular and prostatic cancers in relation to world dietary practices. Int J Cancer 98(2):262–267. Garland FC, Gorham ED, Garland CF, Ducatman AM. 1988. Testicular cancer in US Navy personnel. Am J Epidemiol 127(2):411–414. Gehring GG, Rodriguez FR, Woodhead DM. 1974. Malignant degeneration of cryptorchid testes following orchiopexy. J Urol 112(3):354–356. Gershman ST, Stolley PD. 1988. A case-control study of testicular cancer using Connecticut tumour registry data. Int J Epidemiol 17(4):738–742. Giwercman A, Grindsted J, Hansen B, Jensen OM, Skakkebaek NE. 1987. Testicular cancer risk in boys with maldescended testis: A cohort study. J Urol 138(5):1214–1216. Goedert JJ, McKeen EA, Javadpour N, Ozols RF, Pottern LM, Fraumeni JF, Jr. 1984. Polythelia and testicular cancer. Ann Intern Med 101(5):646– 647. Goedert JJ, Sauter ME, Jacobson LP, et al. 1999. High prevalence of antibodies against HERV-K10 in patients with testicular cancer but not with AIDS. Cancer Epidemiol Biomarkers Prev 8(4 Pt 1):293–296. Goldacre MJ, Clarke JA, Heasman MA, Vessey MP. 1978. Follow-up of vasectomy using medical record linkage. Am J Epidemiol 108(3):176–180. Goss PE, Bulbul MA. 1990. Familial testicular cancer in five members of a cancer-prone kindred. Cancer 66(9):2044–2046. Graham S, Gibson RW. 1972. Social epidemiology of cancer of the testis. Cancer 29(5):1242–1249. Graham S, Gibson R, West D, Swanson M, Burnett W, Dayal H. 1977. Epidemiology of cancer of the testis in upstate New York. J Natl Cancer Inst 58(5):1255–1261. Groves FD, Page WF, Gridley G, et al. 2002. Cancer in Korean war navy technicians: Mortality survey after 40 years. Am J Epidemiol 155(9): 810–818. Guberan E, Usel M, Raymond L, Tissot R, Sweetnam PM. 1989. Disability, mortality, and incidence of cancer among Geneva painters and electricians: A historical prospective study. Br J Ind Med 46(1):16–23. Hackney AC, Premo MC, McMurray RG. 1995. Influence of aerobic versus anaerobic exercise on the relationship between reproductive hormones in men. J Sports Sci 13(4):305–311.

1163

Hansen KS, Lauritsen JM, Skytthe A. 1996. Cancer incidence among mild steel and stainless steel welders and other metal workers. Am J Ind Med 30(4):373–382. Harris EL. 1990. Genetic epidemiology of hypospadias. Epidemiol Rev 12:29–40. Haughey BP, Graham S, Brasure J, Zielezny M, Sufrin G, Burnett WS. 1989. The epidemiology of testicular cancer in upstate New York. Am J Epidemiol 130(1):25–36. Hayes HM, Tarone RE, Casey HW, Huxsoll DL. 1990. Excess of seminomas observed in Vietnam service U.S. military working dogs. J Natl Cancer Inst 82(12):1042–1046. Hayes RB. 1997. The carcinogenicity of metals in humans. Cancer Causes Control 8(3):371–385. Heimdal K, Olsson H, Tretli S, Flodgren P, Borresen AL, Fossa SD. 1996. Familial testicular cancer in Norway and southern Sweden. Br J Cancer 73(7):964–969. Heinonen OP, Slone D, Shapiro S. 2003. Malformations of the genitourinary system. In: Kaufmann DW, ed. Birth Defects and Drugs in Pregnancy. Littleton: Publishing Sciences Group, Inc: pp. 176–199. Heinzer H, Dieckmann KP, Huland E. 1993. Virus-related serology and in situ hybridization for the detection of virus DNA among patients with testicular cancer. Eur Urol 24(2):271–276. Henderson BE, Benton B, Jing J, Yu MC, Pike MC. 1979. Risk factors for cancer of the testis in young men. Int J Cancer 23(5):598–602. Henderson BE, Ross RK, Pike MC, Casagrande JT. 1982. Endogenous hormones as a major factor in human cancer. Cancer Res 42(8): 3232–3239. Henderson BE, Bernstein L, Ross RK, Depue RH, Judd HL. 1988. The early in utero oestrogen and testosterone environment of blacks and whites: Potential effects on male offspring. Br J Cancer 57(2):216–218. Herrington LJ, Zhao W, Husson G. 2003. Management of cryptorchism and risk of testicular cancer. Am J Epidemiol 157(7):602–605. Hutson JM. 1985. A biphasic model for the hormonal control of testicular descent. Lancet 2(8452):419–421. Ilson DH, Bosl GJ, Motzer R, Dmitrovsky E, Chaganti RS. 1991. Genetic analysis of germ cell tumors: Current progress and future prospects. Hematol Oncol Clin North Am 5(6):1271–1283. International Agency for Research on Cancer. 2002. IARC Handbooks of Cancer Prevention Weight Control and Physical Activity vol 6. Lyon: IARC Press, pp. 163–167. Jackson M, Pike MC, Gough M. 1986. Undescended testes—is surgery necessary? N Engl J Med 315(8):516–517. Jarow JP, Budin RE, Dym M, Zirkin BR, Noren S, Marshall FF. 1985. Quantitative pathologic changes in the human testis after vasectomy. A controlled study. N Engl J Med 313(20):1252–1256. Job JC, Gendrel D, Roger M. 1979. Gonadotropins and testosterone secretion in infants with undescended testes. Pediatric Adolescent Endocrinology 6:130–136. Jonsyn FE, Maxwell SM, Hendrickse RG. 1995. Human fetal exposure to ochratoxin A and aflatoxins. Ann Trop Paediatr 15(1):3–9. Jorgensen N, Muller J, Giwercman A, Skakkebaek NE. 1990. Clinical and biological significance of carcinoma in situ of the testis. Cancer Surv 9(2):287–302. Jorgensen N, Giwercman A, Hansen SW, Skakkebaek NE. 1993. Testicular cancer after vasectomy: Origin from carcinoma in situ of the testis. Eur J Cancer 29A(7):1062–1064. Jost A, Magre S. 1988. Control mechanisms of testicular differentiation. Philos Trans R Soc Lond B Biol Sci 322(1208):55–61. Kanisawa M, Suzuki S. 1978. Induction of renal and hepatic tumors in mice by ochratoxin A, a mycotoxin. Gann 69(4):599–600. Karagas MR, Weiss NS, Strader CH, Daling JR. 1989. Elevated intrascrotal temperature and the incidence of testicular cancer in noncryptorchid men. Am J Epidemiol 129(6):1104–1109. Kelce WR, Stone CR, Laws SC, Gray LE, Kemppainen JA, Wilson EM. 1995. Persistent DDT metabolite p, p¢-DDE is a potent androgen receptor antagonist. Nature 375(6532):581–585. Kinkade S. 1999. Testicular cancer. Am Fam Physician 59(9):2539–2550. Klip H, Verloop J, van Gool JD, Koster ME, Burger CW, van Leeuwen FE. 2002. Hypospadias in sons of women exposed to diethylstilbestrol in utero: A cohort study. Lancet 359(9312):1102–1107. Knoke JD, Gray GC, Garland FC. 1998. Testicular cancer and Persian Gulf War service. Epidemiology 9(6):648–653. Kratzik C, Aiginger P, Kuber W, et al. 1991. Risk factors for bilateral testicular germ cell tumors. Does heredity play a role? Cancer 68(4):916–921. Kuoppasalmi K, Naveri H, Harkonen M, Adlercreutz H. 1980. Plasma cortisol, androstenedione, testosterone and luteinizing hormone in running exercise of different intensities. Scand J Clin Lab Invest 40(5):403– 409.

1164

PART IV: CANCER BY TISSUE OF ORIGIN

Lipshultz LI, Caminos-Torres R, Greenspan CS, Snyder PJ. 1976. Testicular function after orchiopexy for unilaterally undescended testis. N Engl J Med 295(1):15–18. Liu MA, Oliff A. 1991. Transforming growth factor-beta—mullerian inhibiting substance family of growth regulators. Cancer Invest 9(3):325–336. Liu S, Semenciw R, Waters C, Wen SW, Mery LS, Mao Y. 2000. Clues to the aetiological heterogeneity of testicular seminomas and non-seminomas: Time trends and age-period-cohort effects. Int J Epidemiol 29(5): 826–831. Longnecker MP, Klebanoff MA, Brock JW, et al. 2002. Maternal serum level of 1,1-dichloro-2,2-bis(p-chlorophenyl)ethylene and risk of cryptorchidism, hypospadias, and polythelia among male offspring. Am J Epidemiol 155(4):313–322. Lykkesfeldt G, Bennett P, Lykkesfeldt AE, et al. 1991. Testis cancer. Ichthyosis constitutes a significant risk factor. Cancer 67(3):730–734. MacConnie SE, Barkan A, Lampman RM, Schork MA, Beitins IZ. 1986. Decreased hypothalamic gonadotropin-releasing hormone secretion in male marathon runners. N Engl J Med 315(7):411–417. Mantzoros CS, Georgiadis EI. 1995. Body mass and physical activity are important predictors of serum androgen concentrations in young healthy men. Epidemiology 6(4):432–435. Marshall EG, Melius JM, London MA, Nasca PC, Burnett WS. 1990. Investigation of a testicular cancer cluster using a case-control approach. Int J Epidemiol 19(2):269–273. Massey FJ, Jr., Bernstein GS, O’Fallon WM, et al. 1984. Vasectomy and health. Results from a large cohort study. JAMA 252(8):1023–1029. McDowall ME, Balarajan R. 1986. Testicular cancer mortality in England and Wales 1971–80: Variations by occupation. J Epidemiol Community Health 40(1):26–29. McGlynn KA, Devesa SS, Sigurdson AJ, Brown LM, Tsao L, Tarone RE. 2003. Trends in the incidence of testicular germ cell tumors in the United States. Cancer 97(1):63–70. McKiernan JM, Goluboff ET, Liberson GL, Golden R, Fisch, H. 1999. Rising risk of testicular cancer by birth cohort in the United States from 1973 to 1995. J Urol 162(2):361–363. McLachlan JA, Newbold RR, Bullock B. 1975. Reproductive tract lesions in male mice exposed prenatally to diethylstilbestrol. Science 190(4218): 991–992. Milham S. 1983. Occupational Mortality in Washington State, 1950–1971. Cincinnati, Ohio, US Department of Health, Education, and Welfare. NIOSH Research Report, Vol 1. Mills PK, Newell GR, Johnson DE. 1984. Testicular cancer associated with employment in agriculture and oil and natural gas extraction. Lancet 1(8370):207–210. Moller H. 1989. Decreased testicular cancer risk in men born in wartime. J Natl Cancer Inst 81(21):1668–1669. Moller H. 1993. Clues to the aetiology of testicular germ cell tumours from descriptive epidemiology. Eur Urol 23(1):8–13. Moller H, Knudsen LB, Lynge E. 1994. Risk of testicular cancer after vasectomy: Cohort study of over 73,000 men. BMJ 309(6950):295–299. Moller H, Skakkebaek NE. 1996. Risks of testicular cancer and cryptorchidism in relation to socio-economic status and related factors: Case-control studies in Denmark. Int J Cancer 66(3):287–293. Moller H, Skakkebaek NE. 1997. Testicular cancer and cryptorchidism in relation to prenatal factors: Case-control studies in Denmark. Cancer Causes Control 8(6):904–912. Moller H, Westergaard T. 1998. [Tobacco smoking and testicular cancer]. Ugeskr Laeger 160(7):1041–1042. Morrison AS. 1976. Some social and medical characteristics of Army men with testicular cancer. Am J Epidemiol 104(5):511–516. Moss AR, Osmond D, Bacchetti P, Torti FM, Gurgin V. 1986. Hormonal risk factors in testicular cancer. A case-control study. Am J Epidemiol 124(1):39–52. Moul JW, Belman AB. 1988. A review of surgical treatment of undescended testes with emphasis on anatomical position. J Urol 140(1):125–128. Mueller N, Hinkula J, Wahren B. 1988. Elevated antibody titers against cytomegalovirus among patients with testicular cancer. Int J Cancer 41(3):399–403. Myrianthopoulos NC, Chung CS. 1974. Congenital malformations in singletons: Epidemiologic survey. Report from the Collaborative Perinatal Project. Birth Defects Orig Artic Ser 10(11):1–58. Newbold RR, McLachlan JA. 1988. Neoplastic and non-neoplastic lesions in male reproductive organs following perinatal exposure to hormones and related substances. In: Mori T, Nagasawa H, eds. Toxicity of Hormones in Perinatal Life. Boca Raton: CRC Press, Inc.: pp. 89–109. Nienhuis H, Goldacre M, Seagroatt V, Gill L, Vessey M. 1992. Incidence of disease after vasectomy: A record linkage retrospective cohort study. BMJ 304(6829):743–746.

Oliver RT. 1990. Clues from natural history and results of treatment supporting the monoclonal origin of germ cell tumours. Cancer Surv 9(2):333–368. Osterlind A, Berthelsen JG, Abildgaard N, et al. 1991. Risk of bilateral testicular germ cell cancer in Denmark: 1960–1984. J Natl Cancer Inst 83(19): 1391–1395. Paffenbarger RS, Jr, Loe IM, Wing AL. 1992. The influence of physical activity on the incidence of site-specific cancers in college alumin. Adv Exp Med Biol 322:7–15. Palmer JM. 1991. The undescended testicle. Endocrinol Metab Clin North Am 20(1):231–240. Panagiotopoulou K, Katsouyanni K, Petridou E, Garas Y, Tzonou A, Trichopoulos D. 1990. Maternal age, parity, and pregnancy estrogens. Cancer Causes Control 1(2):119–124. Parkin DM, Iscovich J. 1997. Risk of cancer in migrants and their descendants in Israel: II. Carcinomas and germ-cell tumours. Int J Cancer 70(6):654– 660. Parkin DM, Bray F, Ferlay J, Pisani P. 2001. Estimating the world cancer burden: Globocan 2000. Int J Cancer 94(2):153–156. Patel SR, Richardson RL, Kvols L, Zincke H. 1990. Observation after orchiectomy in clinical stage I nonseminomatous germ cell tumor of testis. Mayo Clinic experience. Am J Clin Oncol 13(5):379–381. Pearce N, Sheppard RA, Howard JK, Fraser J, Lilley BM. 1987. Time trends and occupational differences in cancer of the testis in New Zealand. Cancer 59(9):1677–1682. Petitti DB, Klein R, Kipp H, Friedman GD. 1983. Vasectomy and the incidence of hospitalized illness. J Urol 129(4):760–762. Petridou E, Panagiotopoulou K, Katsouyanni K, Spanos E, Trichopoulos D. 1990. Tobacco smoking, pregnancy estrogens, and birth weight. Epidemiology 1(3):247–250. Petridou E, Roukas KI, Dessypris N, et al. 1997. Baldness and other correlates of sex hormones in relation to testicular cancer. Int J Cancer 71(6): 982–985. Pinczowski D, McLaughlin JK, Lackgren G, Adami HO, Persson I. 1991. Occurrence of testicular cancer in patients operated on for cryptorchidism and inguinal hernia. J Urol 146(5):1291–1294. Poirier MC, Beland FA. 1992. DNA adduct measurements and tumor incidence during chronic carcinogen exposure in animal models: Implications for DNA adduct-based human cancer risk assessment. Chem Res Toxicol 5(6):749–755. Poland A, Knutson JC. 1982. 2,3,7,8-tetrachlorodibenzo-p-dioxin and related halogenated aromatic hydrocarbons: Examination of the mechanism of toxicity. Annu Rev Pharmacol Toxicol 22:517–554. Pollan M, Gustavsson P, Cano MI. 2001. Incidence of testicular cancer and occupation among Swedish men gainfully employed in 1970. Ann Epidemiol 11(8):554–562. Pottern LM, Brown LM, Hoover RN, et al. 1985. Testicular cancer risk among young men: Role of cryptorchidism and inguinal hernia. J Natl Cancer Inst 74(2):377–381. Pottern LM, Goedert JJ. 1986. Epidemiology of testicular cancer. In: Javadpour N, ed. Principles and Management of Testicular Cancer. New York: Thieme, Inc: pp. 108–119. Rabkin CS, Biggar RJ, Horm JW. 1991. Increasing incidence of cancers associated with the human immunodeficiency virus epidemic. Int J Cancer 47(5):692–696. Rajfer J, Walsh PC. 1977. Hormonal regulation of testicular descent: Experimental and clinical observations. J Urol 118(6):985–990. Rajfer J, Handelsman DJ, Swerdloff RS, et al. 1986. Hormonal therapy of cryptorchidism. A randomized, double-blind study comparing human chorionic gonadotropin and gonadotropin-releasing hormone. N Engl J Med 314(8):466–470. Rajpert-De Meyts E, Skakkebaek NE. 1993. The possible role of sex hormones ion the development of testicular cancer. Eur Urol 23(1):54–59. Rajpert-De Meyts E, Hording U, Nielsen HW, Skakkebaek NE. 1994. Human papillomavirus and Epstein-Barr virus in the etiology of testicular germ cell tumours. APMIS 102(1):38–42. Rapley EA, Crockford GP, Teare D, et al. 2000. Localization to Xq27 of a susceptibility gene for testicular germ-cell tumours. Nat Genet 24(2):197– 200. Reinberg Y, Manivel JC, Zhang G, Reddy PK. 1991. Synchronous bilateral testicular germ cell tumors of different histologic type. Pathogenetic and practical implications of bilaterality in testicular germ cell tumors. Cancer 68(5):1082–1085. Reynolds P, Saunders LD, Layefsky ME, Lemp GF. 1993. The spectrum of acquired immunodeficiency syndrome (AIDS)—associated malignancies in San Francisco, 1980–1987. Am J Epidemiol 137(1):19–30. Rhomberg W, Schmoll HJ, Schneider B. 1995. High frequency of metalworkers among patients with seminomatous tumors of the testis: A case-control study. Am J Ind Med 28(1):79–87.

Testicular Cancer Richiardi L, Bellocco R, Adami H-O, et al. 2004. Testicular cancer incidence in eight northern European countries: secular 2nd recent trends. Cancer Epidemiol Biomarkers, Prev 13:2157–2166. Ries LAG, Eisner MP, Kosary CL, et al. 2002. SEER Cancer Statistics Review, 1973–1999. Bethesda, MD, NCI. Ries LAG, Eisner MP, Kosary CL, et al. 2003. SEER Cancer Statistics Review, 1975–2000. Bethesda, MD, NCI. Rodriguez E, Mathew S, Reuter V, Ilson DH, Bosl GJ, Chaganti RS. 1992. Cytogenetic analysis of 124 prospectively ascertained male germ cell tumors. Cancer Res 52(8):2285–2291. Rosenberg L, Palmer JR, Zauber AG, et al. 1994. The relation of vasectomy to the risk of cancer. Am J Epidemiol 140(5):431–438. Ross RK, McCurtis JW, Henderson BE, Menck HR, Mack TM, Martin SP. 1979. Descriptive epidemiology of testicular and prostatic cancer in Los Angeles. Br J Cancer 39(3):284–292. Samaniego F, Rodriguez E, Houldsworth J, et al. 1990. Cytogenetic and molecular analysis of human male germ cell tumors: Chromosome 12 abnormalities and gene amplification. Genes Chromosomes Cancer 1(4):289–300. Stage D, Sommelet D, Geneix A, Nishi M, Malet P, Vekemans M. 1998. A tumor profile in Down syndrome. Am J Med Genet 78(3):207–216. Schottenfeld D, Warshauer ME, Sherlock S, Zauber AG, Leder M, Payne R. 1980. The epidemiology of testicular cancer in young adults. Am J Epidemiol 112(2):232–246. Schwartz GG. 2002. Hypothesis: Does ochratoxin A cause testicular cancer? Cancer Causes Control 13(1):91–100. Scully RE. 1993. Testis. In: Henson DE, Albores-Saavedra J, editors. Pathology of Incipient Neoplasia. Philadelphia: W.B. Saunders Co.: pp. 384–400. Shephard RJ, Shek PN. 1995. Cancer, immune function, and physical activity. Can J Appl Physiol 20(1):1–25. Shimakage M, Oka T, Shinka T, Kurata A, Sasagawa T, Yutsudo M. 1996. Involvement of Epstein-Barr virus expression in testicular tumors. J Urol 156(1):253–257. Shu XO, Nesbit ME, Buckley JD, Krailo MD, Robinson LL. 1995. An exploratory analysis of risk factors for childhood malignant germ-cell tumors: Report from the Childrens Cancer Group (Canada, United States). Cancer Causes Control 6(3):187–198. Sigurdson AJ, Chang S, Annegers JF, et al. 1999. A case-control study of diet and testicular carcinoma. Nutr Cancer 34(1):20–26. Sinke RJ, Suijkerbuijk RF, de Jong B, Oosterhuis JW, Geurts vK. 1993. Uniparental origin of i(12p) in human germ cell tumors. Genes Chromosomes Cancer 6(3):161–165. Sonneveld DJ, Sleijfer DT, Schrafford KH, et al. 1999. Familial testicular cancer in a single-centre population. Eur J Cancer 35(9):1368–1373. Spermon JR, Witjes JA, Nap M, Kiemeney LA. 2001. Cancer incidence in relatives of patients with testicular cancer in the eastern part of The Netherlands. Urology 57(4):747–752. Srivastava A, Kreiger N. 2000. Relation of physical activity to risk of testicular cancer. Am J Epidemiol 151(1):78–87. Stone JM, Cruickshank DG, Sandeman TF, Matthews JP. 1991. Laterality, maldescent, trauma and other clinical factors in the epidemiology of testis cancer in Victoria, Australia, Br J Cancer 64(1):132–138. Strader CH, Weiss NS, Daling JR. 1988. Vasectomy and the incidence of testicular cancer. Am J Epidemiol 128(1):56–63. Strohsnitter WC, Noller KL, Hoover RN, et al. 2001. Cancer risk in men exposed in utero to diethylstilbestrol. J Natl Cancer Inst 93(7):545– 551. Sutton JR, Coleman MJ, Casey J, Lazarus L. 1973. Androgen responses during physical exercise. Br Med J 1(5852):520–522. Swerdlow AJ, Stiller CA, Wilson LM. 1982. Prenatal factors in the aetiology of testicular cancer: An epidemiological study of childhood testicular cancer deaths in Great Britain, 1953–73. J Epidemiol Community Health 36(2):96–101. Swerdlow AJ, Huttly SR, Smith PG. 1987. Testicular cancer and antecedent diseases. Br J Cancer 55(1):97–103. Swerdlow AJ, Huttly SR, Smith PG. 1987. Prenatal and familial associations of testicular cancer. Br J Cancer 55(5):571–577. Swerdlow AJ, Skeet RG. 1988. Occupational associations of testicular cancer in south east England. Br J Ind Med 45(4):225–230. Swerdlow AJ, Huttly SR, Smith PG. 1989. Testis cancer: Postnatal hormonal factors, sexual behaviour and fertility. Int J Cancer 43(4):549–553. Swerdlow AJ, Douglas AJ, Huttly SR, Smith PG. 1991. Cancer of the testis, socioeconomic status, and occupation. Br J Ind Med 48(10):670–674. Swerdlow AJ, Marmot MG, Grulich AE, Head J. 1995. Cancer mortality in Indian and British ethnic immigrants from the Indian subcontinent to England and Wales. Br J Cancer 72(5):1312–1319.

1165

Swerdlow AJ, De Stavola BL, Swanwick MA, Mangtani P, Maconochie NE. 1999. Risk factors for testicular cancer: A case-control study in twins. Br J Cancer 80(7):1098–1102. Tarone RE, Hayes HM, Hoover RN, et al. 1991. Service in Vietnam and risk of testicular cancer. J Natl Cancer Inst 83(20):1497–1499. Thune I, Lund E. 1994. Physical activity and the risk of prostate and testicular cancer: A cohort study of 53,000 Norwegian men. Cancer Causes Control 5(6):549–556. Tollerud DJ, Blattner WA, Fraser MC, et al. 1985. Familial testicular cancer and urogenital developmental anomalies. Cancer 55(8):1849–1854. Trichopoulos D, Cole P, Brown JB, Goldman MB, Macmahon B. 1980. Estrogen profiles of primiparous and nulliparous women in Athens, Greece. J Natl Cancer Inst 65(1):43–46. UK Testicular Cancer Study Group. 1994. Social, behavioural and medical factors in the aetiology of testicular cancer: Results from the UK study. Br J Cancer 70(3):513–520. UK Testicular Cancer Study Group. 1994. Aetiology of testicular cancer: Association with congenital abnormalities, age at puberty, infertility, and exercise. BMJ 308(6941):1393–1399. Van den Eeden SK, Weiss NS, Strader CH, Daling JR. 1991. Occupation and the occurrence of testicular cancer. Am J Ind Med 19(3):327–337. van Gurp RJ, Oosterhuis JW, Kalscheuer V, Mariman EC, Looijenga LH. 1994. Biallelic expression of the H19 and IGF2 genes in human testicular germ cell tumors. J Natl Cancer Inst 86(14):1070–1075. Van Kessel AG, Suijkerbuijk R, DeJong B, et al. 1991. Molecular analysis of isochromosome 12p in testicular germ cell tumors. In: Oosterhuis JW, Walt H, Damjanov I, eds. Pathobiology of Human Germ Cell Neoplasia. Berlin: Springer-Verlag: pp. 113–118. von der MH, Rorth M, Walbom-Jorgensen S, et al. 1986. Carcinoma in situ of contralateral testis in patients with testicular germ cell cancer: Study of 27 cases in 500 patients. Br Med J (Clin Res Ed) 293(6559):1398–1401. Walker AH, Ross RK, Haile RW, Henderson BE. 1988. Hormonal factors and risk of ovarian germ cell cancer in young women. Br J Cancer 57(4):418– 422. Walker AM, Jick H, Hunter JR, Danford A, Rothman KJ. 1981. Hospitalization rates in vasectomized men. JAMA 245(22):2315–2317. Wanderas EH, Tretli S, Fossa SD. 1995. Trends in incidence of testicular cancer in Norway 1955–1992. Eur J Cancer 31A(12):2044–2048. Wanderas EH, Grotmol T, Fossa SD, Tretli S. 1998. Maternal health and preand perinatal characteristics in the etiology of testicular cancer: A prospective population- and register-based study on Norwegian males born between 1967 and 1995. Cancer Causes Control 9(5):475–486. Wardle J, Steptoe A, Burckhardt R, Vogele C, Vila J, Zarczynski Z. 1994. Testicular self-examination: Attitudes and practices among young men in Europe. Prev Med 23(2):206–210. Weir HK, Kreiger N, Marrett LD. 1998. Age at puberty and risk of testicular germ cell cancer (Ontario, Canada). Cancer Causes Control 9(3):253– 258. Weir HK, Marrett LD, Moravan V. 1999. Trends in the incidence of testicular germ cell cancer in Ontario by histologic subgroup, 1964–1996. CMAJ 160(2):201–205. Weir HK, Marrett LD, Kreiger N, Darlington GA, Sugar L. 2000. Pre-natal and peri-natal exposures and risk of testicular germ-cell cancer. Int J Cancer 87(3):438–443. West RR. 1992. Vasectomy and testicular cancer. BMJ 304(6829):729–730. Westergaard T, Olsen JH, Frisch M, Kroman N, Nielsen JW, Melbye M. 1996. Cancer risk in fathers and brothers of testicular cancer patients in Denmark. A population-based study. Int J Cancer 66(5):627–631. Whittemore AS, Paffenbarger RS, Jr., Anderson K, Lee JE. 1984. Early precursors of urogenital cancers in former college men. J Urol 132(6):1256– 1261. You L, Casanova M, Archibeque-Engle S, Sar M, Fan LQ, Heck HA. 1998. Impaired male sexual development in perinatal Sprague-Dawley and Long-Evans hooded rats exposed in utero and lactationally to p,p¢-DDE. Toxicol Sci 42(2):162–173. Zhang ZF, Vena JE, Zielezny M, et al. 1995. Occupational exposure to extreme temperature and risk of testicular cancer. Arch Environ Health 50(1):13–18. Zimmerli B, Dick R. 1995. Determination of ochratoxin A at the ppt level in human blood, serum, milk and some foodstuffs by high-performance liquid chromatography with enhanced fluorescence detection and immunoaffinity column cleanup: Methodology and Swiss data. J Chromatogr B Biomed Appl 666(1):85–99. Zumoff B, Strain GW, Miller LK, et al. 1990. Plasma free and non-sexhormone-binding-globulin-bound testosterone are decreased in obese men in proportion to their degree of obesity. J Clin Endocrinol Metab 71(4):929–931.

61

Penile Cancer LOUISE WIDEROFF AND DAVID SCHOTTENFELD

C

ancer of the penis occurs infrequently in North America and Europe, and generally in populations and cultures that practice neonatal or childhood circumcision. International incidence data indicate that penile cancer is more common in Africa, South America, and some parts of Asia. Initially, the rare occurrence of penile cancer in circumcised men was attributed to reduced exposure to chemical carcinogens in smegma (Plaut and Kohn-Speyer, 1947), a substance formed from the bacterial breakdown of desquamated epithelium in the preputial sac (Shabad, 1964). Yet correlations in the risk of penile and cervical cancer among spouses, and in the population-based incidence and mortality rates of these two cancers, suggested an etiologic role for a sexually transmitted infectious agent. Current perspectives recognize multiple risk factors, including oncogenic human papillomavirus (HPV) infection, immunosuppression, chronic inflammation and associated phimosis (unretractable foreskin), cigarette smoking, and photosensitizing ultraviolet radiation treatment for benign dermatologic conditions.

CLASSIFICATION Over 90% of all malignant lesions of the penis are squamous cell carcinomas (SCC). Other tumors that account for less than 10% of malignant lesions include: adenocarcinoma arising from the periurethral or bulbo-urethral glands; transitional cell carcinoma arising from the prostatic urethra; melanoma; Kaposi sarcoma; Paget disease, arising from accessory skin glands and frequently associated with underlying adenocarcinoma; fibrosarcoma, and; leiomyosarcoma (Micali et al., 1996; Lucia and Miller, 1992). Several clinical forms of high-grade penile intraepithelial neoplasia (PIN), or in situ SCC, have been described, including Bowen disease, Erythroplasia of Queyrat, and bowenoid papulosis. Bowen disease appears as a sharply demarcated, scaly, erythematous plaque on the penile shaft. Erythroplasia of Queyrat is a mucocutaneous variant of in situ SCC, which appears as one or more shiny plaques on the glans or inner lining of the prepuce and is histologically indistinguishable from Bowen disease. Both Erythroplasia of Queyrat and Bowen disease are typically diagnosed in men over age 35, and progress to invasive disease in 5%–33% of cases. Bowenoid papulosis is a histologically indistinguishable variant of in situ SCC, which clinically presents as multiple reddish-brown papular lesions in men under age 35, and generally does not progress to invasive carcinoma (von Krogh and Horenblas, 2000).

(Ries et al., 2002). Five-year relative survival rates were 68.4% among African Americans and 73.4% among whites. African-American patients died at an average age of 68 years compared with 71 years for whites.

Age The median age at diagnosis in the United States is 69 years (Ries et al., 2002). For all races combined, 5-year age-specific incidence per 100,000 increases steadily, ranging from 1.8 at 60–64 years to levels between 2.9 and 6.7 at 70 and over (Parkin et al., 2002a). Agespecific rates are uniformly higher in high-risk areas of the world, such as Puerto Rico, where the rate per 100,000 at age 60–64 years is 7.6, and at age 70–74 years, 15.7 (Fig. 61–1).

Race and Ethnicity In 1995–1999, African Americans had slightly higher incidence and mortality rates, and lower survival rates, than US whites (Ries et al., 2002). The incidence rates per 100,000 were 0.9 for African Americans and 0.7 for whites. The respective mortality rates per 100,000 were 0.3 and 0.2.

Socioeconomic Status

Based on 1995–1999 SEER data, the average annual age-adjusted incidence rate of penile cancer was 0.7 per 100,000 (Ries et al., 2002). During the same period, the age-adjusted mortality rate was 0.2 per 100,000. From 1992–1999, there was an estimated annual decline of 1.8% in incidence and 6.4% in mortality.

Most studies show an inverse relationship between socioeconomic status (SES) and penile cancer risk, reflecting a higher prevalence of risk factors in low SES populations. When compared with SEER rates, age-specific incidence rates per 100,000 are higher in the US Veterans Administration Medical System, a network of hospitals that generally serves low SES patients with high chronic smoking prevalence (Harris et al., 1989). Incidence data from the Los Angeles population-based registry showed a weak inverse social class gradient, using census tract socioeconomic indicators (Peters et al., 1984). In a study of 109 incident and prevalent cases and 355 population controls in Washington state and British Columbia, men with education beyond high school were at lower risk of penile cancer than those with a high school education or less (age-adjusted odds ratio = 0.6; 95% CI: 0.4–0.9) (Maden et al., 1993). However, a population-based case-control study of 100 age- and race-matched pairs in Los Angeles did not find an association of penile cancer with education or income after adjustment for smoking, physical activity, and medical history variables (Tsen et al., 2001). Internationally, in a study of 23 Brazilian states, inverse correlations ranging from -.64 to -.75 (P < .001) were reported between the relative frequency of penile cancer and various ecologic indicators of affluence (Franco et al., 1988). In high-risk areas of China, an increased risk was observed with decreasing levels of education, although the trend was not statistically significant (Brinton et al., 1991). In a case-control study in Sweden, significant socioeconomic differences between cases and controls disappeared after controlling for smoking history (Hellberg et al., 1987).

Survival

International Patterns

Based on data from 1992–1998, the 5-year relative survival rate of penile cancer in the United States was 72.3% for all stages combined

Internationally, the highest average annual age-adjusted incidence rates per 100,000 were reported in the population-based registries of

DEMOGRAPHIC PATTERNS Mortality and Incidence in the United States

1166

1167

Penile Cancer 100

Rate per 100,000

10

1

0.1

USA, SEER; African Americans USA, SEER; Whites Puerto Rico (1993-97) 0.01

30–34

35–39

40–44

45–49

50–54

55–59

60–64

65–69

70–74

75–79

Age

Kyadondo County, Uganda (4.0), Goiania, Brazil (3.7); Swaziland (3.2); Puerto Rico (2.6); Chiang Mai (2.5) and Songkhla (2.2), Thailand; and Hanoi, Viet Nam (2.2) (Parkin et al., 2002a; Parkin et al., 2002b). In a study of cancer patterns in a rural, uncircumcised population in Mbarara, Uganda, malignant tumors of the penis were estimated to be 17% of all male cancers (Wabinga, 2002). Other areas with elevated penile cancer rates per 100,000 include Harare, Zimbabwe (1.6), Argentina (1.4 in Bahia Blanca; 1.6 in Concordia), and several regions of India (1.8 in Chennai; 1.6 in Nagpur and Poona). In contrast, the lowest incidence of penile cancer is reported among Israeli Jews, and in the Islamic countries of Kuwait and Oman, where the average annual age-adjusted rates approximate zero. Penile cancer is rare in the industrialized countries of Japan, North America, and Europe, with incidence rates generally less than 1.0 per 100,000.

ENVIRONMENTAL FACTORS HPV Infection Initial evidence for a sexually transmitted infectious agent came from early studies showing concordant excess risk of penile and cervical cancer in spouses (Martinez, 1969; Graham et al., 1979; Smith et al., 1980), and correlations in age-adjusted incidence and mortality rates (Bosch and Cardis, 1990; Li et al., 1982), and relative frequencies (Franco et al., 1988), of these two cancers. Penile cancer cases had a high likelihood of a history of syphilis (Persky et al., 1977; Jensen et al., 1977) and benign genital warts (Brinton et al., 1991; Maden et al., 1993), and of developing multiple primary anogenital cancers (Rabkin et al., 1992). It has since been established that genital infection with oncogenic HPV is a major etiologic factor in penile and other anogenital neoplasia, and is a necessary cause in virtually all cervical cancers and pre-invasive cervical lesions (IARC, 1995). Penile HPV infection is associated with greater number of sexual partners, although in populations with high background HPV prevalence, the probability of infection is elevated even in men with relatively few sexual partners (Franceschi et al., 2002). Ulcerative diseases such as syphilis and chancroid may facilitate entry of HPV into epithelial cells, and inadequate penile hygiene may inhibit viral clearance from mucosal surfaces. Through sequencing of viral DNA extracted from infected cells, over 100 HPV genetic variants (i.e., genotypes) of varying oncogenic potential have been identified. In epidemiologic studies, polymerase chain reaction and hybridization techniques are used to detect HPV DNA in exfoliated cells and tissues. Worldwide, HPV 16, followed by 18, 45, and 31 are the most prevalent oncogenic types detected in car-

80–84

85+

Figure 61–1. Semilogarithmic plot of age-specific incidence per 100,000 of penile cancer. The SEER rates are for 1995–1999.

cinomas (IARC, 1995; Bosch et al., 1995; Clifford et al., 2003b) and in high-grade intraepithelial lesions (Clifford et al., 2003b) of the cervix, although there appears to be some geographic variation in type distribution. In case series of 30 or more patients with penile SCC, HPV 16 DNA prevalence ranges from 14%–63%, whereas HPV 18 DNA prevalence ranges from 0–35% (McCance et al., 1986; Maden et al., 1993; Iwasawa et al., 1993; Chan et al., 1994; Gregoire et al., 1995; Cupp et al., 1995; Levi et al., 1998; Picconi et al., 2000; Rubin et al., 2001; Bezerra et al., 2001). The wide range in these estimates may reflect geographic variation, as well as differential sensitivity and specificity of various HPV detection methods. Quality of penile cancer tissues may also contribute to variability in results, as demonstrated by one study that found over twofold higher prevalence of HPV DNA in frozen tissues relative to formalin-fixed tissues from the same patients (Levi et al., 1998). Some of these studies noted a relatively small subset of tumors that harbored other HPV types or multiple types, although the data were sparse. As occurs with vulvar cancer, a significantly higher prevalence of oncogenic viral DNA was detected in basaloid and warty carcinomas compared with keratinizing SCC or verrucous forms (Rubin et al., 2001; Gregoire et al., 1995). A study of 44 men with histologically staged PIN found HPV 16 DNA in 0 of 17 cases of PIN I, 5 of 10 (50%) cases of PIN II, and 12 of 13 (92%) of PIN III (P < 0.001 for PIN I vs. III). Of the 17 PIN I cases, 14 (82.4%) had non-oncogenic HPV types 6 or 11 (Demeter et al., 1993). Viral DNA in these preinvasive lesions was not integrated into the host cell genome, which is the characteristic physical state in malignant tumors (McCance et al., 1986; Tornesello et al., 1997). The relationship of oncogenic HPV infection to penile cancer has also been examined in seroepidemiologic studies that use enzymelinked immunosorbent assays to measure IgG antibody to specific HPV types. In a US case-control study of 121 cases of penile cancer and 579 controls, the odds ratio of seropositivity using HPV 16 serologic and DNA markers was 2.2 (95% CI: 1.4–3.6) (Carter et al., 2001). When the analysis was restricted to HPV 16 DNA-positive tumors, the odds ratio rose to 5.3 (95% CI: 2.2–12.4). In Finland and Norway, serum samples of penile cancer cases identified from national cancer registries were tested for antibodies to HPV 16, 18, and 33. Case sera, which was obtained from nationwide serum banks, had a sevenfold higher likelihood of reactivity to HPV 16 and 33 antigens than control sera from the same source, although the 95% CI in each comparison included 1.00 (Björge et al., 1997). In China, 55 cases and 60 controls were all HPV 16 seronegative, suggesting that either other HPV oncogenic types or environmental risk factors were of greater importance in this population (Wideroff et al., 1996). Worldwide, penile cancer incidence rates are much lower than cervical cancer incidence rates, even in high-risk countries (Parkin et al.,

1168

PART IV: CANCER BY TISSUE OF ORIGIN

2002a). Pre-invasive lesions also appear to be less common in males, with one study reporting PIN in approximately one-third of the sexual partners of women with pre-invasive cervical lesions (Barrasso et al., 1987). The lower risk of genital neoplasia in males suggests genderspecific differences in tissue susceptibility to oncogenic viral infection and in the proportion of disease attributable to viral factors.

Immunosuppression Environmental sources of immunosuppression that are associated with anogenital neoplasia include human immunodeficiency virus (HIV) infection and pharmaceuticals used to prevent organ transplant rejection. Unlike cervical cancer, penile cancer (excluding Kaposi sarcoma) is not an AIDs-defining illness; however, recent studies have documented excess risks of various non-AIDS–defining cancers in HIVinfected populations. Non-AIDs–defining cancers may be expected to increase over time as highly active anti-retroviral therapy extends survival of AIDs patients (Cooley, 2003). Co-infection of HPV and HIV has been anecdotally reported in relation to penile SCC, however, most epidemiologic studies have not had the statistical power to specifically assess risk in HIV-infected populations. A large record linkage study matched population-based HIV/AIDS registries to cancer registries in 11 US states and 6 regions. A relative risk of 3.9 (95% CI: 2.1–6.5) was found for penile cancer (excluding Kaposi sarcoma), in men with HIV/AIDs relative to the general population (Frisch et al., 2001). Kaposi sarcoma, which in rare instances occurs on the penis, has been associated with human herpesvirus-8 in AIDS patients (Stebbing et al., 2003). Renal transplant patients on immunosuppressive drugs also incur an increased likelihood of cutaneous and cervical cancers, often in association with HPV infection (Halpert et al., 1986; Euvrard et al., 1993; Arends et al., 1997). Anogenital malignancies in these patients are often diagnosed at a younger age compared with the general population (Penn, 2000). In the largest cohort study to examine this association, developed from national registries of renal transplant patients in Australia and New Zealand, four cases of penile cancer were observed compared with 0.17 expected (standardized incidence ratio = 23.5; 95% CI: 6.4–60.2) (Fairley, 1994).

Cigarette Smoking In areas of low to intermediate cervical cancer incidence, incidence rates of penile cancer are positively correlated with those of lung cancer, the main tobacco-related cancer in males (Bosch and Cardia, 1990). However, penile and male lung cancer rates are negatively correlated in areas of high cervical cancer incidence, suggesting a minor etiologic role for smoking where oncogenic HPV prevalence is high. In a study of 141 cases and 150 community controls from a high-risk area of China, there was no association of penile cancer with ever having smoked, current smoking, number of cigarettes per day, number of years of smoking, starting age of smoking, or inhalation patterns (Brinton et al., 1991). Odds ratios were adjusted for age, education, sexual relations outside marriage, prior genital conditions, and circumcision. However, North American and European studies have reported a dose-response effect. In a population-based study of 100 age- and race-matched casecontrol pairs in Los Angeles, heavy smoking was associated with penile cancer risk. Relative to never smokers, the odds ratios of penile cancer were 1.6 (0.4–6.9) among ex-smokers, 1.2 (0.4–4.2) among current smokers of <20 cigarettes per day, and 4.0 (1.2–21.0) among current smokers of 20+ cigarettes per day. These odds ratios were adjusted for history of phimosis, penile injury, anogenital warts, other infection, inflammation, or irritation, physical activity, and matching variables of age and race (Tsen et al., 2001). In a Swedish study of 244 prevalent cases and 232 population controls matched on age and geographical area, the estimated risk of penile cancer was 0.98 (0.68–1.42) for smokers of <11 cigarettes per day, and 1.53 (1.00–2.35) for smokers of 11+ cigarettes per day, relative to nonsmokers (Hellberg et al., 1987). Odds ratios were adjusted for history of phimosis and balanitis (inflammation).

In a population-based case-control study in Washington and British Columbia, Daling et al. (1992) found a dose-response effect for current smoking after adjustment of age, number of sex partners, and geographic location. The study included 108 incidence and prevalence cases diagnosed in 1979–1990, and 358 controls. Relative to nonsmokers, the odds ratios of penile cancer were 1.1 (0.4–3.5) among current smokers of <20 cigarettes per day, 3.3 (1.5–7.2) among smokers of 20–39 per day, and 2.7 (0.9–8.5) among smokers of 40+ per day. Estimated risk was threefold higher for smokers of <20 years, and 20–29 years duration, relative to non-smokers, but risk was significantly elevated only for smokers of 30+ years (i.e., odds ratio = 2.4; 95% CI: 1.1–5.4). Odds ratios increased with decreasing age at smoking onset, ranging from 1.9 (0.6–5.3) among those who started at 20+ years old, 2.7 (1.0–7.0) among those who started at 17–19 years, and 3.0 (1.3–6.8) among those who started at <17 years. In agestratified (40–59, and 60+) analysis, risk was significantly elevated only in the 60+ age group, for current vs. never smokers (odds ratio = 4.2; 95% CI: 1.4–12.2). The investigators concluded that cigarette smoke may be a late stage, promotional co-factor in carcinogenesis.

Photochemotherapy An unusually high, dose-dependent risk of penile cancer has been observed in a cohort of psoriasis patients who received oral doses of the photosensitizing agent 8-methoxypsoralen in combination with ultraviolet A (UVA) radiation, the combined therapy referred to as PUVA (Stern et al., 1990). Whereas female patients experienced no increased risk of genital SCC, the risk of SCC of the penis and scrotum was 5–15 times greater than on other PUVA-exposed sites of male patients, suggesting heightened susceptibility of the male genitalia to these carcinogenic stimuli. After an average 12.3 years of follow-up, standardized morbidity ratios (SMR) were used to compare observed numbers of in situ and invasive penile cancers among 892 PUVA-treated men with expected numbers derived from general population incidence rates. The overall SMR was 58.8 (95% CI: 26.9–111.7), whereas SMRs for low (i.e., <140 treatments), medium (140–239 treatments), and high (>239 treatments) doses were 23.0 (2.8–83.0), 34.5 (0.9–192.1), and 162.2 (59.5–353.0), respectively. Risks remained consistently elevated over an additional 10 years of follow-up, even though most patients had discontinued PUVA treatments or started using genital shields before the new follow-up period began (Stern et al., 2002). Many patients had also received ultraviolet B (UVB) radiation, ionizing radiation, or topically applied coal tar. In a nested case-control analysis of 14 cases with penile or scrotal cancer, and 56 age-matched controls, the relative risk for high vs. low UVB dose was 4.6 (1.4, 15.1) after adjustment for level of PUVA exposure (Stern et al., 1990). Tar exposure was similar in cases and controls. Several record linkage studies in Sweden used psoriasis as a marker of PUVA exposure. Linkage of the Swedish Psoriasis Association patient membership list to the national cancer registry yielded four cases of male genital cancers (excluding testis) compared with 1.8 expected, which was not a statistically significant excess (Lindelöf et al., 1989). Linkage of the national hospital discharge and cancer registries yielded a standardized incidence ratio of 4.66 (95% CI: 1.50–10.90) for penile cancer among the 9773 patients hospitalized with psoriasis during 1965–1983 (Boffetta et al., 2001). In a retrospective cohort study of 130 PUVA-treated Scottish males, five genital SCC were observed compared with 0.06 expected (SMR = 362.0; 95% CI: 44.0–1307.0), all in patients receiving a dose greater than 400 J/cm2 (Perkins et al., 1990).

Circumcision For 5000 years or more, circumcision has been practiced across the world for religious and cultural reasons (Warner and Strashin, 1981). Although controversial as a routine preventive practice, neonatal circumcision has been recognized by some sectors of the medical community, particularly in the United States, as a means to reduce risk of phimosis, balanoposthitis (inflammation of the foreskin), sexually

Penile Cancer transmitted infections, and penile cancer (Schoen, 2003; Lerman and Liao, 2001; Lannon et al., 2000). In the United States, the practice became widespread in the 1950s, and as many as 80%–90% of male infants were circumcised by the 1980s (Williams and Kapila, 1993). The prepuce, or foreskin, is the most common site of HPVassociated penile lesions in adult males (Krebs and Schneider, 1987; Syrjänen et al., 1987). Clinical and epidemiologic data show that penile cancer occurs predominantly in uncircumcised males. In six major clinical studies in the United States, none of the more than 1600 patients had undergone neonatal circumcision (Schoen, 1990). Penile cancer is very rare in Jews, who traditionally circumcise infants a week after birth. It is also rare in Muslims, who are circumcised before puberty. In Africa and India, low incidence has been reported in populations where circumcision is practiced, and high incidence reported where it is generally not practiced (Persky, 1977; Muir and Nectoux, 1979; Parkin et al., 2002a). The Washington/British Columbia study modeled the relationship of self-reported circumcision status to penile cancer risk, adjusting for age and history of penile tear (ever vs. never). Compared with men circumcised at birth, the odds ratio of penile cancer was 3.2 (95% CI: 1.8–5.7) among never circumcised men, and 3.0 (95% CI: 1.4–6.6) among those circumcised later in life (Maden et al., 1993). Thus, later circumcision, which was performed on the men as treatment for phimosis, infections, trauma, or for unknown reasons, did not appear to reduce the risk of penile cancer. An age-dependent preventive effect, though not statistically significant, was also observed in a Los Angeles study (Tsen et al., 2001), where adjusted odds ratios were 0.41 (95% CI: 0.13–1.1) for self-reported neonatally circumcised men, and 0.91 (95% CI: 0.31–4.1) for those circumcised later in life (Maden et al., 1993), vs. uncircumcised men. In China, where no study participants were circumcised at birth, circumcision in adult life was associated with increased risk of penile cancer (OR = 32.9; 95% CI: 4.3–253.8). Despite the data supporting a protective effect of circumcision, there is also compelling evidence that low risk of penile cancer is not dependent on circumcision. Incidence rates are very low in many countries that do not practice neonatal or pre-pubertal circumcision, such as Japan, China, and Korea, which report the lowest rates after Israel (Parkin et al., 2002b). This pattern suggests that risk differentials between circumcised and uncircumcised men can be explained by behavioral and biological risk factors that are more prevalent in the uncircumcised. An international study conducted in seven countries of cervical cancer cases, controls, and their current male sexual partners found that compared with uncircumcised men, circumcised men had higher levels of education, older age at first intercourse, fewer lifetime number of sexual partners, less sexual intercourse with prostitutes, more frequent condom use with regular partners and prostitutes, and current female sexual partners with fewer lifetime number of partners (Castellsague et al., 2002). They also had significantly less frequent genital washing after intercourse (P < 0.05) and better genital hygiene as assessed by a physician (P < 0.001). Circumcision status was confirmed by physician exam in three of the seven countries and found to have high reliability with self-report. Studies have consistently shown an association of lack of circumcision with increased risk of acquiring sexually transmitted infections, including human immunodeficiency virus (HIV), syphilis, and chancroid (Moses et al., 1998). In the international cervical cancer study (Castellsague et al., 2002), penile HPV DNA was detected by PCR in 19.6% of 847 uncircumcised partners and 5.5% of 292 circumcised partners. The odds ratio of penile HPV DNA in circumcised vs. uncircumcised partners was 0.37 (95% CI: 0.16–0.85), after adjustment for age, study location, education, age at first intercourse, lifetime number of sexual partners, and frequency of genital washing after sex.

Personal Hygiene In Denmark, a decline was observed in the age-standardized incidence rates per 100,000 person-years, from 1.15 in 1943–1947 to 0.82 in 1988–1990. The decline in this mostly uncircumcised population averaged 0.5% annually (p = 0.002) and occurred during a period in which

1169

the number of dwellings with a bath increased threefold (Frisch et al., 1995). In Los Angeles, penile cancer risk was not associated with frequency of bathing, or frequency or method of cleaning the genital or rectal area directly before and after sexual intercourse (Tsen et al., 2001). In China, there were no trends in estimated risk of penile cancer with increasing number of baths per week or additional washing of the genitals (Brinton et al., 1991). However, self-reported presence of smegma under the foreskin may be a marker of increased risk. In China, based on 116 uncircumcised cases and 127 uncircumcised controls, estimated risk was higher in those who did not suspend their foreskin when bathing relative to those who did, although the association was not statistically significant (adjusted odds ratio = 1.49; 95% CI: 0.8–2.8). In the 18 cases and 30 controls who noticed smegma when suspending foreskin, risk was lower in those who rarely or never noticed it compared with those who often did OR (= 0.21; 95% CI: 0.00–0.90). Among 100 uncircumcised cases and 100 uncircumcised controls in Washington and British Columbia, usually having smegma (vs. rarely or never) was associated with a twofold increase in penile cancer risk (Maden et al., 1993).

Phimosis and Inflammation Phimosis occurs when the orifice of the fully differentiated foreskin is too narrow or constricting to permit its retraction over the glans penis. This condition, which can be congenital or secondary to inflammatory scarring involving the prepuce, interferes with adequate cleansing of the glans and prepuce and is associated with penile cancer. In a highrisk Hindu population, Reddy et al. (1977) reported phimosis in 89% of cases but in only 18% of controls. In China, 73% of uncircumcised cases had phimosis, compared with 7% of controls, resulting in an adjusted OR of 37.2 (95% CI: 11.9–116.1) (Brinton et al., 1991). In Washington and British Columbia (Maden et al., 1993), uncircumcised men reporting a history of difficulty in retracting the foreskin were 3.5 times (95% CI: 1.7–7.4) more likely to develop penile cancer, relative to those not reporting a history. Neonatal circumcision eliminates the risk of phimosis, which may explain much of the protective effect of circumcision against penile cancer. Using never circumcised men as the referent group, Tsen et al. (2001) found that an inverse association of neonatal circumcision and penile cancer risk, which approached statistical significance (OR = 0.41; 95% CI: 0.13–1.1), was weakened when men with a history of phimosis were excluded from the analysis (OR = 0.79; 95% CI: 0.29–2.60). Phimosis is often associated with balanoposthitis, inflammation of the glans and prepuce, although the latter can be an independent risk factor for penile cancer (Jensen, 1977; Hellberg et al., 1987). There are numerous underlying causes of inflammation, including bacterial, viral, and fungal infection, skin disorders, trauma and irritation, allergy, and poor hygiene (Clinical Effectiveness Group, 1999). Lichen sclerosus (LS), a chronic inflammatory dermatitis that presents as white plaques on the glans and often involves the prepuce, is a contributing cause of phimosis and possibly of penile SCC (von Krogh and Horenblas, 2000). Although evidence for an association of LS and penile SCC is based on case reports and histolopathology series (Powell et al., 2001; Nasca et al., 1999), these findings are consistent with more substantial data linking LS to a distinct subset of vulvar SCC and intraepithelial neoplasia (Carlson et al., 1998; Carli et al., 1995; Powell and Wojnarowska, 1999).

PATHOGENESIS Epidemiologic evidence points to multiple causal pathways in penile neoplasia, with oncogenic HPV infection a critical factor in a subset of cancers. Papilloma virions, which consist of an outer protein capsid and inner DNA core, infect proliferating epidermal and mucosal epithelial cells. The viral genome contains 6800–8000 base pairs organized into a circular DNA molecule with eight open reading frames: E6, E7, E1, E2, E4, E5, L1, and L2. Initially, the E5 gene complexes with host cell growth factor receptors, stimulating cell growth and proliferation, and leading to early-stage pathologic lesions. In

1170

PART IV: CANCER BY TISSUE OF ORIGIN

lesions that progress, the circular DNA molecules of oncogenic HPV types are disrupted at a specific site in the E2 gene, and linear DNA fragments become integrated into host cell chromosomes (zur Hausen, 2002). Integrated viral DNA contains the E6 and E7 genes that are synergistically involved in transformation of host cells to the malignant phenotype. The E6 protein interacts with and degrades the p53 tumor suppressor protein and the pro-apoptotic BAK protein. The E7 protein similarly degrades the Rb tumor suppressor protein, and promotes cell proliferation by stimulating S-phase genes such as cyclin-A and E, blocking the cyclin-dependent kinase inhibitors p21 and p27, and deregulating mitotic spindle formation (zur Hausen, 2002; von Knebel Doeberitz, 2002). Most HPV infections are transient, with viral DNA no longer detectable by highly sensitive PCR within a few years of initial onset. In a limited subset of cases, where immune mechanisms fail to clear the infection, viral DNA persists at readily detectable levels or at very low copy numbers, increasing the likelihood of precancerous histologic changes (Schiffman and Kjaer, 2003). In HIV-infected populations, the immunocompromised state facilitates viral replication, resulting in increased risk of high-grade anogenital intraepithelial lesions. However, progression from precursor lesions to invasive carcinoma may be more directly related to genetic alterations than immunologic factors (Palefsky and Holly, 2003). Genetic alterations associated with penile cancer can also result from iatrogenic exposure to PUVA, which has both mutagenic and immunosuppressive effects. The furrocoumarin compound 8-methoxypsoralen, when photoactivated by longwave UV radiation, forms carcinogenic DNA adducts in epidermal cells. PUVA depletes and alters local antigen-presenting Langerhans cells, and consequently, lymphocytic cell populations (Stern, 1989). Through other mechanisms, cells damaged by UV radiation systemically release immunosuppressive cytokines (Strickland and Kripke, 1997). Thus, the immunomodulatory effects of PUVA may generate a permissive environment for clones of malignant cells to escape immune recognition. The role of chronic inflammation as a promoter in the development and progression of neoplasia is well recognized. In response to tissue injury, persistent infection, and chronic irritation, inflammatory cells release cytokines that stimulate neoplastic cell proliferation and angiogenesis. These cells also generate reactive oxygen and nitrogen species capable of inducing genetic alterations (Coussens and Werb, 2002). Persistent clinical HPV infection and penile lichen sclerosus are underlying causes of chronic inflammation that may promote penile neoplasia. Additional promoting factors may include some of the chemical carcinogens in cigarette smoke.

likely to be limited. In adults with phimosis, therapeutic circumcision does not protect against subsequent development of penile cancer (Brinton et al., 1991; Maden et al., 1993; Tsen et al., 2001). Recent research has focused on the development of prophylactic vaccines to induce humoral immunity to oncogenic HPV infection (Plummer and Franceschi, 2002) and therapeutic vaccines to lower risk of disease progression (zur Hausen, 2002). The ultimate goal of vaccination is to reduce morbidity and mortality from cervical cancer and other HPV-associated diseases, including penile cancer. Candidate vaccines must be safe, effective, long-lasting, easy to store, affordable, and acceptable to the public (Lowy and Frazer, 2003). A major strategy has been the development of prophylactic vaccines composed of human papilloma “virus-like particles” (VLPs), which are self-assembling structural proteins that resemble the outer shell (i.e., capsid) of the true virion (Kirnbauer et al., 1993). VLPs contain epitopes that induce formation of type-specific neutralizing antibodies by the host immune system (Roden et al., 1996), but they lack the inner DNA core that makes true virions infectious and oncogenic. Candidate VLP vaccines in early clinical trials consist of the major L1 capsid protein, or of L1 in combination with L2 (Frazer, 2002). To date, vaccine trials have been conducted on women, reflecting the greater burden of HPV-associated morbidity and mortality in women relative to men. Data from a proof-of-principle trial of 2392 female university students suggest that a significant reduction in oncogenic HPV infection is possible through immunization of uninfected women with HPV-16 VLPs (Koutsky et al., 2002). Vaccination of men is viewed as a way to promote herd immunity and thus increase the efficiency of vaccines primarily intended to eliminate cervical cancer mortality (Lowy and Frazer, 2003; Plummer and Franceschi, 2002). One study estimated that vaccination only of women would be 60%–75% as efficient as vaccination of both sexes in reducing the population prevalence of HPV oncogenic types (Hughs et al., 2002). A secondary benefit of male vaccination would be the reduction in risk of less common HPV-associated malignancies, including penile cancer (Lowy and Frazer, 2003). Additional research is needed to evaluate candidate vaccines for long-term efficacy, feasibility of delivery, and other important parameters. Assuming that efficacy is eventually demonstrated in men, cost considerations are likely to influence public health policies on vaccine delivery in resource-poor countries where the economic challenges of vaccine delivery may become a barrier to male vaccination (Goldie, 2003). Furthermore, the multifactorial etiology of penile cancer indicates that HPV immunization will not eradicate all disease, and therefore, clinical detection and treatment of pre-invasive and predisposing chronic conditions must remain an important part of an overall prevention strategy.

PREVENTIVE MEASURES

FUTURE DIRECTIONS

Most known risk factors for penile cancer are essentially modifiable. Therefore, prevention of the majority of disease should be possible through interventions to reduce oncogenic HPV infection and cofactors such as cigarette smoking. Access to medical care is essential for the early detection and treatment of HPV-associated precursor lesions and predisposing conditions such as phimosis and chronic inflammation. Improvement of social conditions in low SES high-risk populations may contribute to disease prevention by facilitating adequate personal hygiene. Preventive measures in psoralen-treated patients include genital shielding and increased medical surveillance for early detection of malignancies (Stern et al., 1990; Perkins et al., 1990). Based on the predominance of penile cancer in uncircumcised males, neonatal circumcision has been recommended as an effective preventive measure, especially in populations where optimal genital hygiene can be difficult to achieve under existing social conditions (Schoen, 1990). Yet, the benefits of neonatal circumcision remain controversial, given sociocultural differences across populations and the small risk of immediate and long-term complications associated with the procedure (Williams and Kapila, 1993; Niku et al., 1995). Therefore, its acceptance outside of specific ethnic or religious groups is

Continuing areas of research in prophylactic HPV vaccine development include determination of duration of immunity, and optimum dose and age at immunization; expansion of monovalent HPV-16 vaccines to multivalent vaccines that protect against infection by multiple oncogenic and benign HPV types, and modification of production methods or routes of delivery to enhance vaccine immunogenicity, thus reducing the required number of doses. Other vital areas of research include the use of alternative technologies to produce heatstable vaccines for delivery in resource-poor areas lacking cold storage facilities, and further development of therapeutic vaccines, mainly composed of non-structural viral proteins, to boost immune response in infected individuals (Frazer, 2002; Lowy and Frazer, 2003). Clinical studies will be needed to evaluate the efficacy and feasibility of male vaccination. Following future dissemination of viable vaccines, surveillance studies will be needed to measure changes in populationbased incidence rates of cervical, penile, and other HPV-associated cancers. References Arends MJ, Benton EC, McLaren KM, et al. 1997. Renal allograft recipients with high susceptibility to cutaneous malignancy have an increased

Penile Cancer prevalence of human papillomavirus DNA in skin tumours and a greater risk of anogenital malignancy. Br J Cancer 75:722–728. Barrasso R, De Brux J, Croissant O, et al. 1987. High prevalence of papillomavirus-associated penile intraepithelial neoplasia in sexual partners of women with cervical intraepithelial neoplasia. N Engl J Med 317:916–923. Bezerra AL, Lopes A, Santiago GH, et al. 2001. Human papillomavirus as a prognostic factor in carcinoma of the penis: Analysis of 82 patients treated with amputation and bilateral lymphadenectomy. Cancer 91:2315–2321. Björge T, Dillner J, Anttila T, et al. 1997. Prospective seroepidemiological study of role of human papillomavirus in non-cervical anogenital cancers. BMJ 315:646–649. Boffetta P, Gridley G, Lindelöf B. 2001. Cancer risk in a population-based cohort of patients hospitalized for psoriasis in Sweden. J Invest Dermatol 117:1531–1537. Bosch FX, Cardis E. 1990. Cancer incidence correlations: Genital, urinary and some tobacco-related cancers. Int J Cancer 46:178–184. Bosch FX, Manos MM, Muñoz N, et al. 1995. Prevalence of human papillomavirus in cervical cancer: A worldwide perspective. J Natl Cancer Inst 87:796–802. Brinton LA, Jun-Yao L, Shou-De R, et al. 1991. Risk factors for penile cancer: Results from a case-control study in China. Int J Cancer 47:504–509. Carli P, Cattaneo A, De Magnis A, et al. 1995. Squamous cell carcinoma arising in vulval lichen sclerosus: A longitudinal cohort study. Eur J Cancer Prev 4:491–495. Carlson JA, Ambros R, Malfetano J, et al. 1998. Vulvar lichen sclerosus and squamous cell carcinoma: A cohort, case control, and investigational study with historical perspective; implications for chronic inflammation and sclerosis in the development of neoplasia. Hum Pathol 29:932–948. Carter JJ, Madeleine MM, Shera K, et al. 2001. Human papillomavirus 16 and 18 L1 serology compared across anogenital cancer sites. Cancer Res 61:1934–1940. Castellsague X, Bosch FX, Munoz N, et al. 2002. Male circumcision, penile human papillomavirus infection, and cervical cancer in female partners. N Engl J Med 346:1105–1112. Chan KW, Lam KY, Chan AC. 1994. Prevalence of human papillomavirus types 16 and 18 in penile carcinoma: A study of 41 cases using PCR. J Clin Pathol 47:823–826. Clifford GM, Smith JS, Plummer M, et al. 2003a. Human papillomavirus types in invasive cervical cancer worldwide: A meta-analysis. Br J Cancer 88:63–73. Clifford GM, Smith JS, Aguado T, et al. 2003b. Comparison of HPV type distribution in high-grade cervical lesions and cervical cancer: A metaanalysis. Br J Cancer 89:101–105. Clinical Effectiveness Group. 1999. National guidelines for the management of balanitis. Sex Transm Inf 75:S85–S88. Cooley TP. 2003. Non-AIDS-defining cancer in HIV-infected people. Hematol Oncol Clin North Am 17:889–899. Coussens LM, Werb Z. 2002. Inflammation and cancer. Nature 420:860–867. Cupp MR, Malek RS, Goellner JR, et al. 1995. The detection of human papillomavirus deoxyribonucleic acid in intraepithelial, in situ, verrucous and invasive carcinoma of the penis. J Urol 154:1024–1029. Daling JR, Sherman KJ, Hislop TG, et al. 1992. Cigarette smoking and the risk of anogenital cancer. Am J Epidemiol 135:180–189. Demeter LM, Stoler MH, Bonnez W, et al. 1993. Penile intraepithelial neoplasia: Clinical presentation and an analysis of the physical state of human papillomavirus DNA. J Infect Dis 168:38–46. Euvrard S, Chardonnet Y, Pouteil-Noble C, et al. 1993. Association of skin malignancies with various and multiple carcinogenic and noncarcinogenic human papillomaviruses in renal transplant recipients. Cancer 72:2198–2206. Fairley CK, Sheil AG, McNeil JJ, et al. 1994. The risk of ano-genital malignancies in dialysis and transplant patients. Clin Nephrol 41:101–105. Franceschi S, Castellsague X, Dal Maso L, et al. 2002. Prevalence and determinants of human papillomavirus genital infection in men. Br J Cancer 86:705–711. Franco EL, Campos Filho N, Villa LL, et al. 1988. Correlation patterns of cancer relative frequencies with some socioeconomic and demographic indicators in Brazil: An ecologic study. Int J Cancer 41:24–29. Frazer I. 2002. Vaccines for papillomavirus infection. Virus Res 89:271–274. Frisch M, Biggar RJ, Engels EA, et al. 2001. Association of cancer with AIDSrelated immunosuppression in adults. JAMA 285:1736–1745. Frisch M, Friss, S, Kjaer SK, et al. 1995. Falling incidence of penis cancer in an uncircumcised population (Denmark 1943–90). BMJ 311:1471. Goldie SJ. 2003. Public health policy and cost-effectiveness analysis. J Natl Cancer Inst Monogr 31:102–110. Graham S, Priore R, Graham M, et al. 1979. Genital cancer in wives of penile cancer patients. Cancer 44:1870–1874.

1171

Gregoire L, Cubilla AL, Reuter VE, et al. 1995. Preferential association of human papillomavirus with high-grade histologic variants of penileinvasive squamous cell carcinoma. J Natl Cancer Inst 87:1705–1709. Halpert R, Fruchter RG, Sedlis A, et al. 1986. Human papillomavirus and lower genital neoplasia in renal transplant patients. Obstet Gynecol 68:251–258. Harris RE, Hebert JR, Wynder EL. 1989. Cancer risk in male veterans utilizing the Veterans Administration Medical System. Cancer 64:1160–1168. Hellberg D, Valentin J, Eklund T, et al. 1987. Penile cancer: Is there an epidemiologic role for smoking and sexual behavior? Br Med J 295:1306–1308. Hughes JP, Garnett GP, Koutsky L. 2002. The theoretical population-level impact of a prophylactic human papilloma virus vaccine. Epidemiology 13:631–639. IARC Monographs on the Evaluation of Carcinogenic Risks to Humans. 1995. Human Papillomaviruses. Vol 64. Lyon, France. Iwasawa A, Kumamoto Y, Fujinaga K. 1993. Detection of human papillomavirus deoxyribonucleic acid in penile carcinoma by polymerase chain reaction and in situ hybridization. J Urol 149:59–63. Jensen MO. 1977. Cancer of the penis in Denmark 1942 to 1962. Danish Med Bull 24:66–72. Kirnbauer R, Taub J, Greenstone H, et al. 1993. Efficient self-assembly of human papillomavirus type 16 L1 and L1–L2 into virus-like particles. J Virol 67:6929–6936. Koutsky LA, Ault KA, Wheeler CM, et al. 2002. A controlled trial of a human papillomavirus type 16 vaccine. N Engl J Med 347:1645–1651. Krebs HB, Schneider V. 1987. Human papillomavirus-associated lesions of the penis: Colposcopy, cytology and histology. Obstet Gynecol 70:299–304. Lannon CM, Bailey A, Fleischman A, et al. 2000. Circumcision debate. Task Force on Circumcision, 1999–2000. Pediatrics 105:641–642. Lerman SE, Liao JC. 2001. Neonatal circumcision. Pediatr Clin North Am 48:1539–1557. Levi JE, Rahal P, Sarkis AS, et al. 1998. Human papillomavirus DNA and p53 status in penile carcinomas. Int J Cancer 76:779–783. Li JY, Li FP, Blot WJ, et al. 1982. Correlation between cancers of the uterine cervix and penis in China. J Natl Cancer Inst 69:1063–1065. Lindelöf B, Eklund G, Lidén S, et al. 1990. The prevalence of malignant tumors in patients with psoriasis. J Am Acad Dermatol 22:1056–1060. Lowy D, Frazer IH. 2003. Prophylactic human papillomavirus vaccines. J Natl Cancer Inst Monogr 31:111–116. Lucia MS, Miller GJ. 1992. Histopathology of malignant lesions of the penis. Urol Clin North Am 19:227–246. Maden C, Sherman KJ, Beckmann AM, et al. 1993. History of circumcision, medical conditions, and sexual activity and risk of penile cancer. J Natl Cancer Inst 85:19–24. Martinez I. 1969. Relationship of squamous cell carcinoma of the cervix uteri to squamous cell carcinoma of the penis among Puerto Rican women married to men with penile carcinoma. Cancer 24:777–780. McCance DJ, Kalache A, Ashdown K, et al. 1986. Human papillomavirus types 16 and 18 in carcinomas of the penis from Brazil. Int J Cancer 37:55–59. Micali G, Innocenzi D, Nasca MR, et al. 1996. Squamous cell carcinoma of the penis. J Am Acad Dermatol 35:432–451. Moses S, Bailey RC, Ronald AR. 1998. Male circumcision: Assessment of health benefits and risks. Sex Transm Inf 74:368–373. Muir CS, Nectoux J. 1979. Epidemiology of cancer of the testis and penis. Natl Cancer Inst Monogr 53:157–164. Nasca MR, Innocenzi D, Micali G. 1999. Penile cancer among patients with genital lichen sclerosus. J Am Acad Dermatol 41:911–914. Niku SD, Stock JA, Kaplan GW. 1995. Neonatal circumcision. Urol Clin North Am 22:57–65. Palefsky JM, Holly EA. 2003. Immunosuppression and co-infection with HIV. J Natl Cancer Inst Monogr 31:41–46. Parkin DM, Whelan SL, Ferlay J, et al. 2002a. Cancer Incidence in Five Continents, Volume 8. IARC Scientific Publications No. 155. Lyon, France: International Agency for Research on Cancer. Parkin DM, Ferlay J, Hamdi-Cherif M, et al. 2002b. Cancer in Africa: Epidemiology and Prevention. IARC Scientific Publications No. 153. Lyon, France: International Agency for Research on Cancer. Penn I. 2000. Post-transplant malignancy: The role of immunosuppression. Drug Saf 23:101–113. Perkins W, Lamont D, MacKie RM. 1990. Cutaneous malignancy in males treated with photochemotherapy. Lancet 336:1248. Persky L. 1977. Epidemiology of cancer of the penis. Recent Results Cancer Res 60:97–109. Peters RK, Mack TM, Bernstein L. 1984. Parallels in the epidemiology of selected anogenital carcinomas. J Natl Cancer Inst 72:609–615. Picconi MA, Eiján AM, Distéfano AL, et al. 2000. Human papillomavirus (HPV) DNA in penile carcinomas in Argentina: Analysis of primary tumors and lymph nodes. J Med Virol 61:65–69.

1172

PART IV: CANCER BY TISSUE OF ORIGIN

Plaut A, Kohn-Speyer AC. 1947. The carcinogenic action of smegma. Science 105:391–392. Poland RL. 1990. The question of routine neonatal circumcision. N Engl J Med 322:1312–1315. Plummer M, Franceschi S. 2002. Strategies for HPV prevention. Virus Res 89:285–293. Powell J, Robson A, Cranston D, et al. 2001. High incidence of lichen sclerosus in patients with squamous cell carcinoma of the penis. Br J Dermatol 145:85–89. Powell J, Wojnarowska F. 1999. Lichen sclerosus. Lancet 353:1777–1783. Rabkin CS, Biggar RJ, Melbye M, et al. 1992. Second primary cancers following anal and cervical carcinoma: Evidence of shared etiologic factors. Am J Epidemiol 136:54–48. Reddy CRRM, Gopal Rao T, Venkatarathnam G, et al. 1977. A study of 80 patients with penile carcinoma combined with cervical biopsy study of their wives. Int Surg 62:549–553. Ries LAG, Eisner MP, Kosary CL, et al. 2002. SEER Cancer Statistics Review, 1973–99. National Cancer Institute, Bethesda, MD, http://seer.cancer. gov/csr/1973–1999/. Roden RB, Greenstone HL, Kirnbauer R, et al. 1996. In vitro generation and type-specific neutraliztation of a human papillomavirus type 16 virion pseudotype. J Virol 70:5875–5883. Rubin MA, Kleter B, Zhou M, et al. 2001. Detection and typing of human papillomavirus DNA in penile carcinoma: Evidence for multiple independent pathways of penile carcinogenesis. Am J Pathol 159:1211–1218. Schiffman M, Kjaer SK. 2003. Natural history of anogenital human papillomavirus infection and neoplasia. J Natl Cancer Inst Monogr 31:14–19. Schoen EJ. 2003. It’s wise to circumcise: time to change policy. Pediatrics 111:1490–1491. Schoen EJ. 1990. The status of circumcision of newborns. N Engl J Med 322:1308–1312. Shabad AL. 1964. Some aspects of etiology and prevention of penile cancer. J Urol 92:696–702. Smith PG, Kinlen LJ, White GC, et al. 1980. Mortality of wives of men dying with cancer of the penis. Br J Cancer 41:422–428. Stebbing J, Portsmouth S, Gotch F, et al. 2003. Kaposi’s sarcoma—an update. Int J STD AIDS 14:225–227.

Stern RS. 1989. PUVA and the induction of skin cancer. Carcinog Compr Surv 11:85–101. Stern RS, Members of the Photochemotherapy Follow-up Study. 1990. Genital tumors among men with psoriasis exposed to psoralens and ultraviolet A radiation (PUVA) and ultraviolet B radiation. N Engl J Med 322:1093– 1097. Stern RS, Bagheri S, Nichols K, et al. 2002. The persistent risk of genital tumors among men treated with psorlen plus ultraviolet A (PUVA) for psoriasis. J Am Acad Dermatol 47:33–39. Strickland FM, Kripke ML. 1997. Immune response associated with nonmelanoma skin cancer. Clin Plast Surg 24:637–647. Syrjänen SM, von Krogh G, Syrjänen KJ. 1987. Detection of human papillomavirus DNA in anogenital condylomata in men using in situ DNA hybridization applied to paraffin sections. Genitourin Med 63:32–39. Tornesello ML, Buonaguro FM, Meglio A, et al. 1997. Sequence variations and viral genomic state of human papillomavirus type 16 in penile carcinomas from Ugandan patients. J Gen Virol 78:2199–2208. Tsen HF, Morgenstern H, Mack T, et al. 2001. Risk factors for penile cancer: Results of a population-based case-control study in Los Angeles County (United States). Cancer Causes and Control 12:267–277. von Knebel Doeberitz M. 2002. New markers for cervical dysplasia to visualise the genomic chaos created by aberrant oncogenic papillomavirus infections. Eur J Cancer 38:2229–2242. von Krogh G, Horenblas S. 2000. Diagnosis and clinical presentation of premalignant lesions of the penis. Scan J Urol Nephrol Suppl 205:220–229. Wabinga HR. 2002. Patterns of cancer in Mbarara, Uganda. East Afr Med J 79:22–26. Warner E, Strashin E. 1981. Benefits and risks of circumcision. Can Med Assoc J 125:967–976. Wideroff L, Schiffman M, Hubbert N, et al. 1996. Serum antibodies to HPV 16 virus-like particles are not associated with penile cancer in Chinese males. Viral Immunol 1996:23–25. Williams N, Kapila L. Complications of circumcision. Br J Surg 80:1231–1236. zur Hausen H. 2002. Papillomaviruses and cancer: From basic studies to clinical application. Nat Rev Cancer 2:342–350.

62

Nervous System SUSAN PRESTON-MARTIN, REEMA MUNIR, AND INDRO CHAKRABARTI

M

alignant nervous system (NS) tumors account for 18,300 of new cancer diagnoses each year or 1.4% of all primary incident cancers and for 13,100 or 2.4% of annual cancer deaths (American Cancer Society Inc., 2003). The vast majority of these tumors arise in the central nervous system (CNS), and for this site inclusion of benign tumors doubles the annual incidence. For many reasons inclusion of benign as well as malignant CNS tumors makes more sense in the study of the epidemiology of these tumors, and has now been mandated by the US Surveillance, Epidemiology, and End Results (SEER) network of population-based cancer registries. Our review will focus on benign and malignant tumors of the brain, cranial nerves, and cranial meninges, which account for 95% of all CNS tumors. Each year more than 35,000 new benign or malignant primary brain tumors are diagnosed among residents of the United States (CBTRUS, 2002). These tumors are unique because of their location within the bony structure of the cranium. Symptoms depend on location, and histologically benign tumors can result in similar symptomatology and outcome as from a malignant tumor. For this reason many cancer registries have routinely included both benign and malignant intracranial tumors. For simplicity, this group of tumors will be called “brain tumors” or, when benign tumors are excluded, “brain cancer.” The term “central nervous system tumors” (or cancer) indicates that tumors of the spinal cord and spinal meninges are included along with brain tumors, and “nervous system (NS) tumors” indicates that tumors of the peripheral nerves are included as well; these latter cases account for only around 2% of NS tumors and are much less studied. The presenting brain tumor symptom can be localized, as with a specific motor, speech, or sensory deficit resulting from compression of the corresponding region of the brain. More commonly, symptoms are generalized as with headaches and seizures that result from an increase in intracranial pressure. One-third of patients present with headaches and one-fifth present with seizures (Walker, 1975). The presence of these symptoms often leads to a detailed neurological examination including brain imaging such as radioisotope scanning, computerized tomography (CT), and magnetic resonance imaging (MRI). Although tumors of the brain and cranial meninges rarely metastasize outside the central nervous system, the brain is a common metastatic site for tumors arising at other anatomic sites including cancers of the lung, breast, kidney, gastrointestinal tract, prostate and malignant melanoma (Rubinstein, 1972). Diagnoses based solely on brain imaging and other noninvasive techniques have been shown, not infrequently, to be unreliable both in differentiating primary from metastatic tumors and in determining the histologic type of primary tumors (Todd et al., 1987). Therefore, histologic diagnosis is encouraged to allow for more appropriate patient management. Procedures such as stereotactic brain biopsy reduce the risk to patients from such procedures (Bullard et al., 1984). Brain tumors are the second leading cause of death from neurological disease—second only to stroke, and the tenth most common cause of death from cancer. The 13,100 US residents who are expected to die from primary nervous system tumors exceeds the number expected to die from cancers of the urinary bladder, kidney and renal pelvis, esophagus, stomach, and uterus (cervix and corpus combined) to mention only a few of the less common causes of cancer mortality (American Cancer Society Inc., 2003).

CLASSIFICATION Variation in Inclusion Criteria The descriptive epidemiology of these tumors has been difficult to study because of the wide variation in specific tumors included in published rates. Quantitatively, the most important variation relates to the inclusion or exclusion of benign tumors. This critical difference has often been ignored in comparisons across geographic areas. Recent vigorous efforts are being made to induce population-based registries in the United States to include benign as well as malignant tumors; in the future, SEER and many other data sets should allow study of both groups. Nonetheless, caution must still be required in any international comparisons or comparisons of rates over time. For international comparisons the present review uses data from registries that report rates for malignant tumors only. Although reporting of malignant tumors alone eases geographical comparisons, it is unfortunate that incidence rates for benign nervous system tumors are not also reported. For this reason, benign tumors will not be excluded from descriptive data shown for Los Angeles County. In fact, as will become clear from discussions of analytic studies below, more is known about the etiology of benign histologic types such as meningiomas than about the etiology of gliomas, which are more common than meningiomas and are usually malignant.

Anatomic Classification Another difficulty in comparing various descriptive surveys and other studies of this group of tumors relates to the inconsistency with respect to what subsites are included. Tumors of the central nervous system include tumors of the brain, cranial nerves, cerebral meninges, spinal cord, and spinal meninges. These subsites are represented by the International Classification of Diseases for Oncology (ICD-O) codes C70.0-C72.9. Unlike some surveys, we will not include tumors of sites such as the eye and the pituitary gland, which appear to be etiologically distinct. Our review will focus on benign and malignant tumors of the brain, cranial nerves, and cranial meninges, which account for most of the tumors at this site (95% of all CNS tumors; 93% of all nervous system tumors). Some data sets used in descriptive comparisons include tumors of the spine, which account for 5% of all CNS tumors. Others, such as those compiled in volumes of Cancer Incidence in Five Continents also include tumors of the peripheral nerves, which comprise an even smaller proportion of the total (2% in data for Los Angeles County). Few epidemiologic studies (either descriptive or analytic) have investigated tumors of the spine or peripheral nerves per se.

Histopathology and Coding A meaningful classification system for any type of benign or cancerous lesion should account for the cell of origin of the lesion, the degree of clinical aggressiveness, interobserver reproducibility, and ideally will be simple and easy to apply uniformly. Additionally, a suitable scheme should aid the physician in initiating and tracking a management protocol. The histopathologic classification of brain tumors has evolved tremendously since the first formulations of Bailey and Cushing based on cell of origin (1926). In 1949, Kernohan introduced the notion that the degree of differentiation of the cell of origin was

1173

1174

PART IV: CANCER BY TISSUE OF ORIGIN

also important and that biologic progression of CNS tumors was related to the degree of differentiation; he proposed a four-tier grading system in which the clinical grade increases with poorer differentiation of the cell type (Kaye and Laws, 2001). Since this system was proposed other classification schemes have been presented including the commonly used scheme of the World Health Organization (WHO) Classification of CNS tumors. Other grading systems include the Ringertz three-tier system (Ringertz, 1950) and the Dumas-Duport four-tier system (Daumas-Duport et al., 1988). The WHO classification of tumors was first published in 1979 (Zulch, 1979). This has been revised in 1993 and most recently in 2000 (Kleihues et al., 1993; Kleihues and Cavenee, 2000). The WHO system has allowed for international standardization of CNS malignancies based on cell of origin and histopathologic features including degree of anaplasia. Four grades are used with grades 1 and 2 considered as low grade or benign (Michotte, 1996). Grades 3 and 4 are considered high grade or malignant. Typically a grade 1 lesion will be well circumscribed, distinct from the surrounding parenchyma, and curable with complete surgical excision. Examples include juvenile pilocytic astrocytomas or extra-axial tumors such as ordinary meningiomas or schwannomas. Grade 2 lesions are more diffuse in their growth and infiltrative into the surrounding tissues. This makes surgical excision difficult with chances for residual tumor and recurrence. The grade 3 lesions exhibit diffuse growth patterns with cellular evidence of malignancy. This includes increasing anaplastic features such as high mitotic activity, nuclear atypia, and immature cells of decreased differentiation. Highly malignant and swiftly growing lesions are grade 4 by WHO classification. These lesions show evidence of vascular proliferation and areas of necrosis indicative of rapid cell division. Glioblastoma multiforme and primitive neuroectodermal tumors are examples of such lesions. Kleihues has shown that this grading system is reproducible with appropriate correlation of histopathologic features to clinical behavior and response to therapy (Kleihues, 1995). In the 1993 edition of the WHO classification of brain tumors there were 10 major categories of tumors based on cell of origin. In the 2000 version there are seven major categories of tumor. Several other changes have been made including re-classifying pituitary adenomas and carcinomas. Previously under the heading of “Tumors of the Sellar Region,” these tumors are no longer classified as CNS tumors but are now considered endocrine tumors. The largest group in the WHO scheme is neuroepithelial origin, which includes nine categories: astrocytic, oligodendroglial, mixed gliomas, choroid plexus tumors, glial tumors of uncertain origin, neuronal and mixed neuronal-glial tumors, neuroblastic tumors, pineal parenchymal tumors, and embryonal tumors. The majority of CNS primary tumors are of astroglial origin, called gliomas. The astrocytic cells arise from multi-potent neural stem cells and retain their ability to divide (Pringle et al., 1998). In Los Angeles County among men and women, gliomas account for 59% and 42% of primary tumors of the brain, respectively (Preston-Martin and Mack, 1996). The astrocytic tumors account for 80% of gliomas and include astrocytoma, anaplastic astrocytoma, and glioblastoma multiforme (Collins, 1998). These represent increasing grades of anaplasia and clinical virulence of tumors of astrocytic cells. Therefore, an astrocytoma is grade 2, an anaplastic astrocytoma is grade 3, and glioblastoma is grade 4. Around 40% of primary brain cancers are glioblastomas or anaplastic astrocytomas (Lindsay et al., 1991). Pilocytic astrocytomas are grade 1 in this category. Aside from tumors originating in the brain tissue itself, such as the gliomas mentioned above, tumors in the intracranial compartment can also originate from the structures surrounding the brain. This includes the meninges, or covering of the brain, the cranial nerves exiting the brain, blood vessels, or primitive remnants left from development. Another subheading in the WHO classification scheme includes tumors of peripheral nerves such as schwannomas, neurofibromas, and malignant peripheral nerve sheath tumors. Acoustic schwannomas, which originate from the eighth cranial nerve, account for 90% of CNS schwannomas and 8% of CNS tumors (Preston-Martin and Mack, 1996). Tumors of the meninges is another subheading and includes

meningiomas, the second most common primary tumor of the CNS accounting for approximately 30% of central nervous system tumors in men and women combined (Longstreth et al., 1993; Preston-Martin and Mack, 1996). These are more common in women and originate in the cells of the dura mater, which is the covering for the brain and spinal cord (Preston-Martin and Mack, 1996). Lymphomas and hemopoietic tumors, germ cell tumors, and tumors of the sellar region make up the last three subheadings. Within tumors of the sellar region are craniopharyngiomas, tumor-like processes growing from epithelial rests remaining from Rathke’s pouch, the progenitor for the pituitary gland. These lesions are more common in children, comprising 13% of childhood brain tumors (Lindsay et al., 1991). Ninety-five percent of central nervous system tumors consist of tumors of the brain such as gliomas and other tumors of neuroepithelial tissue, tumors of the meninges such as meningiomas, and schwannomas, which are tumors of the nerve sheath. This review will focus on these tumors. Some criticisms for the WHO and other pathologic classification schemes frequently are discussed. For instance, the pathologic confirmation of tumor type is subject to interobserver variability. Any classification system that is based on visual criteria observed under a microscope is subjective, by definition. However, progress with immunohistochemistry and use of monoclonal antibodies for cellular identification has made it more objective. Also, the cell of origin of certain tumors has not always been clear. In the past, primitive neuroectodermal tumors, PNETS, a malignant neoplasm of childhood, has been used to include all supratentorial and infratentorial tumors with poor differentiation and similar morphologic features. However, it recently has been shown that the tumors that occur infratentorially occurring in the cerebellum originate from the external granular layer of the cerebellum. Meanwhile, supratentorial tumors originate from a different primitive progenitor cell population (Kleihues et al., 1993). Additionally, molecular genetic pathways of progression have been shown to be different in these two types of childhood tumors. In one study, 100% of supratentorial PNETs were found to have preservation of the distal arm of chromosome 17p while it was missing in 37% of the infratentorial cases (Burnett et al., 1997). Currently, novel genomic studies of the gene sequence and additions, deletions, or mutations of particular oncogenic sequences have added a new dimension that will surely affect future clinical grading scores. The molecular genetic characterization of central nervous system tumors has allowed some more understanding of the pathogenesis of brain tumors. As time passes the clinical correlation of such genetic alterations will be further uncovered. It is likely that in the future tumor grading will be based on cell of origin, degree of differentiation, and presence or absence of specific genomic alterations. See Table 62–1 for the three major groups used in the descriptive tables and figures showing brain tumor data for Los Angeles County as well as in most recent epidemiologic studies of brain tumors. The types of tumors in each major grouping and the corresponding standard International Classification of Diseases for Oncology (ICD-O, 2000) codes for each are also shown.

Clinical Classification Now that CT scanning is available in many general hospitals in the United States and other industrialized countries, the differential diagnosis of intracranial masses is often made by physicians who are not specialists in neurological disease. In a survey in the United Kingdom, fewer than half of patients with CT diagnoses were referred to neurosurgeons for histologic confirmation by surgery or biopsy, and the positive predictive value of the CT diagnosis was around 90% for gliomas and meningiomas but only 50% for metastatic tumors (Todd et al., 1987). Magnetic resonance imaging (MRI) is now the gold standard for clinical presentations in which a neoplasm is the most probable diagnosis. The superior sensitivity of MRI also makes it useful for detection of tumors that computed tomography or ultrasound are unable to detect (Poussaint, 2001). In geographic areas where CT scanning and MRI are not available, the accuracy of clinical diagnoses of primary

1175

Nervous System Table 62–1. ICD-O Morphology Codes for Primary Tumors of Brain, Cranial Nerves, and Cranial Meninges (ICD-O, 2000) ICD-O

Type

Gliomas (and Other Tumors of Neuroepithelial Tissue)

astrocytic tumors 9400 9401 9410 9411 9420 9421 9424 9440 9441 9442 9384

Astrocytoma, NOS Anaplastic astrocytoma Protoplasmic astrocytoma Gemistocytic astrocytoma Fibrillary astrocytoma Pilocytic astrocytoma Pleomorphic xanthoastrocytoma Glioblastoma, NOS Giant cell glioblastoma Gliosarcoma Subependymal giant cell astrocytoma

oligodendroglial tumors 9450 9451

Oligodendroglioma, NOS Oligodendroglioma, anaplastic

mixed gliomas 9382

Mixed glioma; anaplastic oligoastrocytoma

ependymal tumors 9391 9392 9393 9394 9383

Ependymoma Anaplastic ependymoma (see also embryonal tumors) Papillary ependymoma Myxopapillary ependymoma Subependymoma

choroid plexus tumors 9390

Choroid plexus papilloma; choroid plexus carcinoma

glial tumors of uncertain origin 9430 9381 9444

Astroblastoma Gliomatosis cerebri Chordoid glioma

neuronal and mixed neuronal-glial tumors 9492 9493 9412 9413 9505 9506 8680

Gangliocytoma Dysplastic gangliocytoma of cerebellum (Lhermitte-Duclos) Desmoplastic infantile astrocytoma (/ganglioglioma) Dysembryoplastic neuroepithelial tumor Ganglioglioma Central neurocytoma; cerebellar liponeurocytoma Paraganglioma, NOS (of filium terminale)

ICD-O 9523 9500

Type Olfactory neuroepithelioma Neuroblastoma (see also embryonal tumors)

embryonal tumors 9501 9392 9470 9471 9472 9473 9474 9500 9490 9508

Medulloepithelioma, NOS Ependymoblastoma (see also ependymal tumors) Medulloblastoma Desmoplastic medulloblastoma Medullomyoblastoma Primitive neuroectodermal tumor, NOS (supratentorial PNET) Large cell medulloblastoma Neuroblastoma (see also neuroblastic tumors) Ganglioneuroblastoma Atypical teratoid/rhabdoid tumor

tumors of cranial nerves schwannoma 9560

Neurilemoma

tumors of the meninges tumors of meningothelial cells 9530 9531 9532 9533 9534 9537 9538 9538 9539 9530

Meningioma Meningothelial meningioma Fibrous meningioma Psammomatous meningioma Angiomatous meningioma Transitional meningioma Clear cell meningioma Chordoid meningioma; papillary meningioma; rhabdoid meningioma Atypical meningioma (Anaplastic) Malignant meningioma

miscellaneous neoplasms 9120 9133 9150 9120

Hemangioma NOS Epithelioid hemangioendothelioma NOS Hemangiopericytoma (Hem)angiosarcoma

primary melanocytic lesions 8728

Diffuse melanocytosis; meningeal melanocytoma; meningeal melanomatosis

tumors of uncertain histogenesis 9161

Hemanogioblastoma

neuroblastic tumors 9522

Olfactory neuroblastoma (Aesthesioneuroblastoma)

brain tumors that are not histologically confirmed is likely to be much worse. The rate of microscopic confirmation of brain and nervous system tumors over the years 1988–1992 varied widely across geographic areas from a high of 100% (e.g., among Prince Edward Island and Yukon, Canada males; Los Angeles County Chinese, Filipino, Japanese; and San Francisco Japanese, among others) to a low of 0% (in Bamako, Mali females, Parkin et al., 1997). Rates of histologic verification vary considerably across registries as well as across specific population groups within a country. For example, the rates of histologic verification range from 54%–98% in Switzerland, 38%–100% in Canada, and 23%–64% in Brazil, and 46%–86% in Japan (Parkin et al., 1997). Although for 1993–1997 the rates of morphological verification (MV%) are now all well above 70% in Switzerland, Canada, and Brazil, substantial variation remains in Japan and a number of other countries (Parkin et al., 2002). Such wide variation suggests that caution in the interpretation of these rates is warranted. In general, for a relatively inaccessible cancer site such as the brain a higher rate of microscopic confirmation makes us more confident that the neoplasm actually existed and that it was correctly classified. In some registries, however, a high rate may indicate that only those tumors that received histologic verification were reported and that clinically diagnosed tumors may have been missed.

Molecular Genetic Characteristics Progress in recent years has been made toward identification of specific genes that are often targeted for mutation, amplification, or deletion in the oncogenesis of brain tumors. Early genetic studies focused on structural cytogenetic alterations within tumor cells. Currently, very specific targets on the genome of a cell can be identified with certainty. The application of techniques such as in situ hybridization, genomic hybridization, linkage analysis, loss of heterozygosity analysis, and immunohistochemistry has led to the verification of several ultrastructural and molecular genetic alterations that can occur in the various subtypes of brain tumor lesions. As has been shown with other cancers, the pathogenesis of brain tumors appears to relate to two families of genes, proto-oncogenes and tumor suppressor genes (TSG) (Nagane et al., 1997). In general, mutations somehow interfere with cell cycle regulation, programmed cell death, or help promote angiogenesis within the tumor. The protooncogenes are basically “gain of function” mutations leading to over transcription of a certain protein that promotes cell proliferation. Generally, this process only requires a single mutation. The tumor suppressor genes, on the other hand, are endogenously present in the genome and function to regulate cell cycling and apoptotic programmed cell death. Loss of both alleles of a tumor suppressor gene

1176

PART IV: CANCER BY TISSUE OF ORIGIN

leads to unregulated growth and increased tumorigenicity. Usually, a germline mutation of a single allele is present and a second spontaneous mutation occurs at the other locus leading to growth advantage and tumor formation. The loss of this second allele has been coined as “loss of heterozygosity” (LOH) at a tumor suppressor gene locus. Several frequent alterations of genes in association with brain tumor formation have been catalogued over the past decade. Gene alterations can lead to either a mutation of the normal gene product, amplification of the gene product, or the absence of the gene product. Earlier studies focused on karyotyping the gene pool of a cell. From these studies it has been learned that glioblastoma cells can frequently have loss of chromosomes 10 and 22, gain of chromosome 7, and other structural changes on chromosomes 1p, 9p, 11p, and 12q (Bigner et al., 1984; von Deimling et al., 1994). Other changes discovered were loss of the 1p and 19q chromosomal arms in oligodendrogliomas (Ransom et al., 1992). In oligodendrogliomas loss of heterozygosity at 19q occurs in 50%–80% of tumors and loss of 1p occurs in 40%–90% of tumors (Kraus et al., 1995). The loss of chromosome 22 has been noted with ependymomas (Weremowicz et al., 1992) and in meningiomas (Ruttledge et al., 1994; Harada et al., 1996). Structural abnormalities of chromosome 17 with loss of the p arm have been noted in 30%–50% of medulloblastomas (Wechsler-Reya and Scott, 2001). The most well-known proto-oncogene with relation to nervous system tumors is the epidermal growth factor receptor gene (EGFR). This gene normally encodes a transmembrane signaling protein that is activated after binding of epidermal growth factor and transforming growth factor a. This gene amplification results in oncogenic activity and tumor progression. Cytogenetic studies have shown that amplification of this gene can occur in up to 40% of glioblastomas (Bigner et al., 1987); overall, the overexpression of this gene appears not to correlate with patient survival (Olson et al., 1998), but one study did show that survival was related if both the age of the patient and the presence of EGFR amplification were considered simultaneously (Smith et al., 2001). Patients younger than 40 with EGFR amplification had shorter survival compared with age-matched patients without the amplification, whereas in patients over age 60, amplification was associated with prolonged survival (Smith et al., 2001). Certain other oncogenes have been identified in malignant gliomas. These include the myc-n, CDK4, and MDM2 genes. CDK4 amplification has been seen in up to 15% of anaplastic astrocytomas and glioblastomas. Tumor suppressor genes have also been found to occur with some frequency in CNS tumors, and to date more TSGs have been identified than oncogenes. The most well-known TSG is the p53 gene. This gene is located on the chromosome 17 at 17p13.1, and is often deleted in astrocytic tumors (James et al., 1989). A p53-related mutation has been observed in over 75% of glioblastomas (Ichimura et al., 2000), and a second spontaneous mutation at this site will lead to uncontrolled growth. Normally, this gene produces a protein that promotes stoppage of the cell cycle and apoptosis, or programmed cell death. Accelerated growth and malignancy is encountered with loss of this function. This loss of heterozygosity can be inherited in the germline in Li-Fraumeni syndrome, which predisposes to multiple cancers including gliomas. PTEN gene loss on chromosome 10q23 has been highly correlated with malignant gliomas as well (Steck et al., 1997). This gene deletion has been lost in up to 44% of glioblastomas in one study (Wang et al., 1997), and its inactivation has been clinically correlated with shortened survival (Raffel et al., 1999). Genetic mutations seen in brain tumor cells have been correlated with clinical behavior in glioblastoma (GBM). Glioblastoma is the most common primary brain tumor (Osborne et al., 2001), and a highly malignant WHO grade 4 tumor that is fatal in 1–2 years in 90% of cases (Louis et al., 2001). However, investigations have described two clinical and genetic variants of GBM. The primary GBM develops de novo and has been referred to as Type II GBMs (von Deimling et al., 1994; Ng and Lam, 1998; Louis et al., 2001). This variant is more aggressive with rapid growth and fatality. It occurs in an older set of patients with the mean age of 56.3 years (von Deimling et al., 1994). The genetic pathway involves the overexpression of EGFR after loss

of chromosome 10. Type I GBM, on the other hand, occurs in younger patients with a mean age of 40.5 years (von Deimling et al., 1994), is less aggressive, and is associated with better response to treatment and improved survival. Type I GBM arises when a precursor lesion of lower grade undergoes degeneration with concomitant genetic alterations. The steps in progression start from a grade II astrocytoma that undergoes LOH at 17p, then 19q and/or 9q leading to grade III, and finally LOH on chromosome 10 leading to a grade IV GBM (Osborne et al., 2001). Giant-cell GBM is a pathologic variant of GBM that historically has had a slightly better prognosis than the classic GBM. It has been shown that this tumor also has genetic alterations similar to the Type I GBM (Ng and Lam, 1998). Kunwar et al. (2001) in their study of anaplastic astrocytomas, WHO grade III, discovered patient survival after diagnosis and age of diagnosis correlated with certain genetic findings. Gains on chromosome 7p and 7q were associated with increased survival. The loss of 11p was more frequent in younger patients, and gains on 7p 19, with loss on 4q were noted in older patients. Certain genetic alterations have been predictive of response to therapy as well. P53 mutations have been correlated with decreased chemosensitivity in malignant gliomas (Nagane et al., 1997). With anaplastic oligodendrogliomas a difference in response to chemotherapy has been noted as well. For tumors with loss of chromosome 1p and 19q, patient survival was more than 50 months longer when compared with tumors where 1p and any other chromosome was lost (Louis et al., 2001). The malignant transformation of a meningioma has been noted to occur with loss of 14q and 10q when chromosome 22 mutations are present beforehand in the benign tumor (Ng and Lam, 1998). Far more observations such as these are anticipated. The understanding of well-established heritable syndromes with associated CNS tumors has also created insight into sporadic tumor development. Meningiomas, gliomas, and other tumors that have been part of an inherited genetic syndrome have been characterized and gene mutations documented, although genetic analysis of sporadic tumors, such as meningiomas and gliomas, has led to interesting discoveries as well. Another tumor suppressor gene includes the cloned genes of the inherited neurofibromatosis (NF) genes. The NF-1 gene has been mapped to chromosome 17 and the syndrome manifests with multiple peripheral neurofibromas and a higher incidence of CNS tumors, including pilocytic astrocytomas. An interesting discovery is that the absence of this gene also has been noted in up to 20% of sporadic cases of pilocytic astrocytomas (von Deimling et al., 1994). Additionally, the NF-2 gene located on chromosome 22 has been detected as inactive or absent in up to 60% of sporadic meningiomas and schwannomas (Ruttledge et al., 1994). The Von Hippel Lindau (VHL) gene is located on chromosome 3p25-26. This mutation causes hemangioblastomas of the CNS as well as other associated findings. However, sporadic mutations at this site have been discovered in up to 40% of hemangioblastomas in non VHL cases (Tse et al., 1997). The advances in molecular biology have contributed greatly to our knowledge about the biologic behavior of brain tumors and have allowed strategies focused on improving patient care and outcome. It is clear now that tumorigenesis can be correlated with multiple genetic events. These genetic events can occur during development or during adult life. In the future it is reasonable to expect the genotype of a patient’s tumor will be included along with the pathologic diagnosis.

DEMOGRAPHIC PATTERNS Mortality and Incidence in the United States Table 62–2 shows the average annual age-adjusted incidence rates for brain malignancies among the 11 registries in the nationwide network known as Surveillance, Epidemiology, and End Results (SEER). Overall, the incidence of brain cancer within each of the two major ethnic groups (white; nonwhite) is similar across SEER registries, and no remarkable geographic variation is apparent. However, when comparing white and non-white populations, in all registries except for Iowa, the incidence of brain malignancies among male non-white populations was significantly lower than those in male white populations;

1177

Nervous System

Table 62–2. Age-Adjusted Annual Incidence Rate Per 100,000 of Primary Malignant Brain Tumors and 95% Confidence Interval (CI) by Gender, Race, and SEER* Registry, Whites and Non-Whites, 1992–2001 Males SEER Registry

Females

Rate

CI

N

Pop

Rate

CI

N

Pop

8.4 9.0 9.0 9.2 8.3 8.5 6.9 9.2 8.2 8.4 7.9 7.5

8.2,8.5 8.5,9.5 8.5,9.5 8.7,9.7 6.9,10.0 8.0,9.0 6.3,7.5 8.8,9.7 7.6,8.9 7.7,9.1 7.3,8.6 7.2,7.8

10,932 1,161 1,289 1,254 125 1,144 478 1,426 668 588 619 2,180

14,319,698 1,358,880 1,425,437 1,425,853 162,290 1,355,102 745,470 1,651,499 987,584 818,186 886,851 3,502,546

5.8 5.8 6.1 6.1 6.1 5.9 4.6 6.2 6.0 6.4 6.2 5.5

5.6,6.0 5.4,6.3 5.7,6.5 5.7,6.5 4.9,7.6 5.5,6.3 4.1,5.1 5.8,6.6 5.5,6.5 5.8,7.0 5.7,6.8 5.2,5.7

8,606 840 1,020 982 89 930 354 1,042 528 499 527 1,795

14,486,676 1,352,798 1,515,174 1,483,593 142,829 1,422,852 767,121 1,663,680 990,500 816,367 859,319 3,472,444

4.4 5.5 4.4 4.6 4.1 5.9 3.5 4.6 2.3 4.0 4.4 4.1

4.2,4.7 4.8,6.2 3.3,6.1 3.9,5.3 3.5,4.8 3.5,11.1 2.1,5.7 3.6,5.9 0.9,7.3 3.2,5.0 3.4,5.6 3.7,4.6

1,547 284 70 213 178 23 26 100 11 147 98 386

4,150,735 1,264,145 412,840 1,119,437 894,527 104,629 229,020 534,118 100,831 1,026,969 541,601 2,290,205

3.2 3.2 3.9 3.5 3.3 5.9 2.1 2.8 3.1 3.1 2.8 3.1

3.0,3.4 2.8,3.7 2.9,5.0 3.0,4.0 2.8,3.9 3.4,9.7 1.2,3.4 2.2,3.7 1.5,6.8 2.5,3.7 2.1,3.7 2.8,3.5

1,295 201 69 93 149 22 22 71 14 131 64 349

4,468,109 656,939 214,743 599,344 450,728 51,828 117,160 270,636 49,605 541,319 267,651 1,198,147

white Eleven SEER registries San Francisco-Oakland Connecticut Detroit (Metropolitan) Hawaii Iowa New Mexico Seattle (Puget Sound) Utah Atlanta (Metropolitan) San Jose-Monterey Los Angeles

non-white Eleven SEER registries San Francisco-Oakland Connecticut Detroit (Metropolitan) Hawaii Iowa New Mexico Seattle (Puget Sound) Utah Atlanta (Metropolitan) San Jose-Monterey Los Angeles

*SEER: Surveillance, Epidemiology, and End Results, a U.S. wide network of population-based cancer registries. Rates are age-adjusted to the 2000 US standard.

a similar excess is seem among white compared with non-white females across the 11 SEER registries.

Time Trends Figures 62–1A and 62–1B show changes in brain cancer incidence across two time periods, 1973–2001 and 1992–2001. Figure 62–1A shows brain cancer incidence patterns in nine SEER registries and demonstrates a rate jump between 1984–1986. There has been much debate over whether such increases in brain cancer rates are due to true increases in occurrence or are artifactual due to increasing use of improved brain imaging techniques, in particular magnetic resonance imaging (MRI). Smith et al. (1998) demonstrated that the jump in pediatric brain cancer rates from 1984 through 1986 was coincident with the widespread use of improved imaging technology in the United States. A similar step increase in incidence was seen among older adult patients and was also attributable to increased use of MRI (Legler et al., 1999). From 1992–2001, data from all 11 SEER registries indicate that the annual incidence rates of all malignant brain tumors in both males and females have not continued to increase consistently (Figure 62–1B). Figure 62–1C shows changes in incidence of all primary tumors of the brain, cranial meninges, and cranial nerves in Los Angeles County from 1972–2002. In contrast with other SEER registries, Los Angeles County reports benign as well as malignant tumors. As a result, we see higher overall incidence rates when comparing LA County and SEER rates. This higher incidence rate is most notable in females because of their higher rates of benign meningiomas. To evaluate secular trends in the Los Angeles County data over the same time period we also did JoinPoint analyses by race and sex for six groups (Hispanic, Black and non-Hispanic males and females; Figs. 62–1D and 62–1E). Rates among Hispanic males showed a statistically significant monotonic decline; a similar monotonic decline was seen among black males, but it was not statistically significant. Among nonHispanic white males two changes in rate trends occurred, and all

slopes had statistically significant monotonic increases, which were greatest from 1997–2002. Among females, Hispanics had a statistically significant monotonic decline in incidence rates, and Blacks a monotonic increase that was not statistically significant. Other white females showed two changes in rate trends with all lines having statistically significant monotonic increases; the increase from 1972–1978 was similar to that occurring from 1995–2002, and the increase from 1978–1995 was smaller. The highly publicized observation in 1990 that the mortality and incidence of brain tumors had risen sharply over past decades, especially in the elderly, led to numerous investigations in an attempt to explain this increase. The consensus is that this increase is largely attributable to improved imaging of the brain. A recent descriptive study in four Nordic countries supports this view (Lonn et al., 2004). They found an increase in average annual incidence of intracerebral tumors confined to the late 1970s and early 1980s, which coincided with the introduction of improved diagnostic methods. The increase was predominantly among the oldest age group, and the incidence has remained fairly level among both genders since 1983 (Lonn et al., 2004).

Survival Yearly the National Cancer Institute catalogs US population-based survival rates for all malignant primary brain tumors combined. Recent age-adjusted 5-year relative survival rates from 1989–1996 were 30.8% (Gurney and Kadan-Lottick, 2001). This is a slight improvement from 22.5% between 1974–1976. The 5-year survival rate is age dependent. For patients aged 19 years and below, the 5-year survival rate is 65%. The rate for those 44 years and younger is 58.7%. For those aged 65–74 years and over 75 years the 5-year survival rates are 6.5% and 3.6%, respectively (Ries et al., 2000). Another recent study using SEER data that stratified survival from primary malignant brain tumors by tumor histology (Davis et al., 1998) also noted overall decreasing survival with increasing age at

PART IV: CANCER BY TISSUE OF ORIGIN

9

9

8

8

7

7

Rate per 100,000 Cases

Rate per 100,000 Cases

1178

6 5 4 3 2

6 5 4 3 2

1

1 01

0

14

01 20

20

Year of diagnosis

B

12

Males=12,479 Females=9,901

10 Males

8

18

6 16

4

14

2

19

72 19 74 19 76 19 78 19 80 19 82 19 84 19 86 19 88 19 90 19 92 19 94 19 96 19 98 20 00 20 02

0

Year of diagnosis

C

Rate per 100,000 cases

Rate per 100,000 Cases

00

99 19

19 98

97 19

96 19

95 19

94

93

19

Year of diagnosis

A

19

19

92

20

99

97

19

93

91

89

87

95

19

19

19

19

19

85 19

19

81

83 19

79

19

75

77

19

19

19

19

73

0

Males=9,268 Females=10,263

12 10 8 6 4 2

Females

Rate per 100,000 cases

14

D

12

2002

2000

1998

1996

1994

1992

1990

1988

1986

1982

1980

1978

1984

Year of diagnosis Black Latino Other white

10 8 6 4 2 2002

2000

1998

1996

1994

1992

1990

1988

1986

1984

1982

1980

1978

1976

1974

1972

0

E

1976

1972

16

1974

0 18

Year of diagnosis Black Latino Other white

diagnosis and found an overall 5-year survival rate of 20%. For the most part the survival rates between the sexes did not differ. The survival rate for patients with glioblastoma whose brain tumor was diagnosed between the years 1986–1991 was only 1% for all ages combined. The rate of survival was 34% for astrocytomas and 84% for pilocytic astrocytomas, which occur predominantly in children. Survival from medulloblastoma for all age groups combined improved from 40% in the 1970s to 60%. The survival rates for oligodendroglioma have also substantially improved to 65% between 1986–1991. Five-year survival rates remained similar over the 20

Figure 62–1. (A) Age-adjusted incidence rates of primary malignant brain tumors by year for nine US SEER registries combined, 1973–2001. (B) Age-adjusted incidence rates of primary malignant brain tumors by year for 11 US SEER registries, combined, 1992–2001. (C) Age-adjusted incidence rates by year of all primary tumors of the brain, cranial nerves, and cranial meninges by sex, all races combined, Los Angeles County 1972–2002. (D) Age-adjusted incidence rates by year of all primary tumors of the brain, cranial nerves, and cranial meninges by race, males, Los Angeles County 1972–2002. (E) Age-adjusted incidence rates by year of all primary tumors of the brain, cranial nerves, and cranial meninges by race, females, Los Angeles County 1972–2002.

years for ependymomas (60%), mixed gliomas (51%), “other” gliomas (23%), and gliomas not otherwise specified (NOS) (17%). It is thought that improved diagnostic abilities and earlier age at diagnosis may account for the improvements in survival of astrocytomas, medulloblastomas, and oligodendrogliomas over the past 20 years. Also, the response to adjuvant therapies has been shown to decrease with increasing age and may explain why age at diagnosis is an important prognosticator (Hershatter et al., 1986). Neurosurgical procedures have undergone refinement with expanded use of computer assistance, imaging, microscope technology, instruments, and

1179

Nervous System neuro-navigation, all of which have led to improved resections of brain tumors. In addition, survival is clearly longer for benign and less aggressive malignant tumors. A study from Victoria, Australia showed that meningiomas have a 92% relative 5-year age adjusted survival rate (Preston-Martin et al., 1993); the survival is slightly longer for women at 94% compared with men at 87%, mostly likely because a higher proportion of men with meningiomas have malignant tumors. Nerve sheath tumors such as acoustic neuromas have 100% survival. The survival rate of GBMs remains a dismal 1%. In addition to morphologic type, various characteristics of certain types of tumors have been shown to relate to survival. For example, the presence of progesterone receptors in meningiomas appears to be a favorable prognostic factor (Hsu et al., 1997).

Glioma

Korean Filipino Japanese Chinese Other whites Spanish surnamed Black 0.00

Age The average annual age-specific incidence of brain tumors in 11 SEER registries among non-Hispanic whites is shown in Figure 62–2. In both males and females, rates decline after a peak in early childhood and increase again after age 25. Rates generally increase from age 25–75 and level off after age 75, possibly because of continued underdiagnosis of these tumors among the elderly.

1.00

2.00

3.00

4.00

5.00

6.00

7.00

8.00

Rate per 100,000 Cases Male

Female

A Meningioma Korean Filipino

Sex

Japanese

In Figures 62–3A and 62–3B, we see that for all types of brain cancers combined, rates are higher in males than in females at all ages. Figures 62–3A and 62–3B show that males have higher rates of gliomas regardless of ethnic group and females have higher meningioma rates. Table 62–3 shows the age-adjusted annual incidence for the major histologic groups of primary brain tumors by sex and ethnic group in Los Angeles County, 1972–2002. For all histologic types and all races combined, the rate is higher in men than in women. For non-Latino whites as well, male rates for all histologic types are higher than female rates, but this pattern is not seen for several of the other ethnic groups. For each of the three major site/histology subgroups incidence rates vary considerably by histologic subtypes. Glioma rates are higher in males than in females in each ethnic group, and meningioma rates are higher in women. For the most part, rates of nerve sheath tumors are similar in women and men.

Chinese Other whites Spanish surnamed Black 0

0.5

1

1.5

2

2.5

3

3.5

4

Rate per 100,000 Cases Male

Female

B Figure 62–3. (A) Average annual age-adjusted incidence rates for glioma by ethnicity and gender, Los Angeles County, 1972–2002. (B) Average annual age-adjusted incidence rates for meningioma by ethnicity and gender, Los Angeles County, 1972–2002.

Socioeconomic Status Table 62–4 shows the proportional incidence ratios (PIRs) for primary tumors of the brain, cranial nerves, and cranial meninges by social

30

Rate per 100,000 Cases

25 20 15 10

class (as determined by census tract of residence) for all races combined in Los Angeles County. There is an apparent trend of increasing incidence with increasing social class for some types of tumors. In both sexes, gliomas and nerve sheath tumors have a higher incidence with increasing social class. Because this trend occurs more strikingly among males than among females, it seems unlikely that it might relate to factors such as diagnostic efficiency or exposure to diagnostic radiography of the head (e.g., dental X-rays), both of which might be expected to be greater among those in higher social classes. Meningiomas, on the other hand, have higher incidence rates in the lower social classes.

Race and Ethnicity

5 0 00–14 years

15–24 years

25–34 years

35–44 years

45–54 years

55–64 years

65–75 years

75 + years

Age at diagnosis Males Females

Figure 62–2. Average annual age-specific incidence rates for primary brain cancer in whites, 11 SEER registries combined, 1992–2001.

Figures 62–3A and 62–3B show the ethnicity-specific age-adjusted incidence rates for gliomas and meningiomas in Los Angeles County from 1972–2002. The incidence of gliomas is highest in nonSpanish–surnamed whites for males and females. The incidence of meningiomas, on the other hand, is highest in black males and females. Figures 62–3A and 62–3B show that in Los Angeles County, the rate of gliomas is lower among black males and females than among whites but that the reverse is true for meningiomas. Spanish-surnamed males and females have lower rates of both gliomas and meningiomas than other white males and females. Asians in Los Angeles County have the lowest rates of both tumor types, but the difference in rates between Asians and whites is less for meningiomas than gliomas.

1180

PART IV: CANCER BY TISSUE OF ORIGIN

Table 62–3. Average Annual Age-Adjusted Incidence by Major Histologic Type of Primary Tumors of Brain, Cranial Nerves, or Cranial Meninges by Gender and Race, Los Angeles County, 1972–2001* Glioma

Meningioma

Nerve Sheath Tumor

All Primary CNS Tumors

Rate

CI

N

Rate

CI

N

Rate

CI

N

Rate

CI

N

4.26 4.56 7.59 3.07 2.29 2.44 2.68 2.17 6.08

3.84,4.67 4.23,4.89 7.36,7.81 2.37,3.77 1.55,3.03 1.80,3.09 1.77,3.58 1.59,2.75 5.92,6.23

473 1185 453 80 39 58 40 92 6497

2.69 1.40 2.17 1.10 0.87 1.98 0.96 1.35 1.99

2.32,3.07 1.18,1.62 2.05,2.30 0.66,1.54 0.44,1.31 1.34,2.63 0.38,1.55 0.83,1.88 1.90,2.09

239 231 1244 26 16 38 12 39 1845

0.33 0.35 1.06 0.46 0.96 0.78 0.38 0.52 0.79

0.22,0.44 0.27,0.43 0.98,1.14 0.20,0.72 0.50,1.42 0.41,1.16 0.03,0.72 0.30,0.75 0.74,0.85

36 89 652 12 18 17 5 23 852

7.33 6.36 10.89 4.63 4.16 5.25 4.02 4.04 8.92

6.76,7.90 5.95,6.77 10.62,11.16 3.76,5.50 3.18,5.14 4.26,6.23 2.89,5.14 3.23,4.85 8.73,9.11

753 1529 6469 118 74 114 57 154 9268

2.89 3.58 5.06 1.90 1.52 1.87 1.95 1.45 4.10

2.59,3.18 3.32,3.83 4.89,5.23 1.38,2.42 0.96,2.07 1.31,2.44 1.27,2.63 1.08,1.82 3.99,4.22

387 997 3478 54 30 49 33 71 5099

4.01 2.95 3.68 2.06 2.01 3.52 2.44 2.47 3.44

3.64,4.37 2.71,3.20 3.54,3.82 1.52,2.60 1.39,2.62 2.76,4.27 1.56,3.32 1.92,3.01 3.33,3.54

483 627 2656 58 42 90 35 90 4081

0.39 0.47 1.07 0.42 0.69 0.80 0.54 0.75 0.81

0.28,0.50 0.38,0.56 0.99,1.15 0.18,0.66 0.35,1.02 0.46,1.14 0.19,0.90 0.48,1.02 0.76,0.87

51 123 720 12 16 22 9 32 985

7.38 7.07 9.89 4.44 4.21 6.19 5.05 4.68 8.43

6.90,7.87 6.70,7.43 9.65,10.13 3.65,5.22 3.32,5.10 5.19,7.19 3.87,6.23 3.96,5.39 8.26,8.59

935 1778 6902 126 88 161 79 194 10263

males Black Latinos Other Whites Chinese Japanese Filipino Korean Other Races All Races

females Black Latinos Other whites Chinese Japanese Filipino Korean Other races All races

*Data from University of Southern California/Los Angles County Cancer Surveillance Program (CSP), the population-based tumor registry for Los Angeles County. Age-adjusted rates are a weighted average of the age-specific rates, where the weights represent the age distribution of a standard population. Rates are adjusted by the direct method to the 2000 US population and are calculated per 100,000 persons. Age-adjusted rate = S(wiri), where wi is the proportion of age group i (I = 0–4 to 85+) in the standard population, and ri is the Los Angeles County age-specific rate for the age group.

International Patterns Table 62–5 compares international data on the incidence of malignant brain and nervous system tumors included in the latest volume of Cancer Incidence in Five Continents (Parkin et al., 2002). These comparisons show an age-standardized incidence rate (ASR) for males and for females for each registry included. Registries that had a rate of morphological verification (MV%) of 70% or greater and had 10 or more cases for each sex were eligible for inclusion, but only the largest registry that met these criteria was selected for each country. For some registries included in this table, population notes indicated that a certain degree of underascertainment of cases was suggested by the high MV% (e.g., Algiers registry), a high mortality to incidence ratio (e.g., Costa Rica), or the overall high proportion of cancers with primary site unspecified (e.g., Delhi). In general, rates among men and women in Australia, Europe, Canada, and whites in the United States are high and relatively similar.

Incidence is generally higher among men than women. Rates in Asian populations in China, India, Japan, and Chinese in Singapore are considerably lower. Figures 62–4A and 62–4B graphically compare the age-standardized incidence rates from selected countries for malignant brain and other nervous system tumors in males and females, respectively. Males and females in Norway have the highest ASRs. Registries in countries such as Singapore, Japan, India, and China, have lower ASRs for malignant nervous system tumors.

Migration Although countries such as Japan show lower incidence rates for malignant nervous system tumors, it is difficult to compare native and migrant populations to verify whether there are certain lifestyle factors that may work to increase or decrease incidence rates. There

Table 62–4. Proportional Incidence Ratios for Primary Tumors of Brain, Cranial Nerves, and Cranial Meninges by Socioeconomic Status,* All Races Combined, Los Angeles County, 1972–1999 Gliomas Socio-economic Status

Meningiomas

Nerve Sheath Tumors

PIR

CI

N

PIR

CI

N

PIR

CI

N

119.9 103.0 98.9 91.7 82.5

113.8,126.2 97.4,108.9 93.2,104.7 86.3,97.5 76.9,88.4

1452 1238 1155 1049 798

94.0 95.3 98.8 98.4 118.5

82.2,107.0 83.3,108.5 86.3,112.5 85.5,112.6 102.9,135.9

229 227 226 210 205

149.5 114.7 90.6 73.6 54.8

126.6,175.5 94.6,137.9 72.4,112.1 56.8,93.8 39.0,75.0

150 113 85 65 39

106.8 104.0 100.3 96.0 90.8

100.3,113.6 97.6,110.6 94.0,106.9 89.7,102.6 84.1,97.9

1003 1000 936 856 678

93.1 87.5 99.3 114.8 111.5

84.8,101.9 79.5,96.0 90.6,108.7 104.9,125.4 100.3,123.5

469 438 473 494 363

121.9 121.5 94.0 80.2 68.4

103.4,142.9 102.6,142.9 76.9,113.6 64.0,99.3 51.7,88.8

153 146 106 84 56

males 1 (high) 2 3 4 5

females 1 (high) 2 3 4 5

*Socioeconomic status determined by census tract of residence at time of diagnosis and based on average household income and level of education of head of household. PIR, proportional incidence ratios.

Nervous System Table 62–5. International Comparison of Age-Standardized Incidence Rates (ASRs) and Number of Cases of Primary Malignant Nervous System Tumors by Gender and Race from Selected Population-Based Registries,* 1993–1997 Males Registry

ASR

Females N

ASR

Occupation and Industry 2.7

128

1.4

70

3.8 4.6 3.5 2.5 2.4 2.5 2.5 2.3 5.0 2.5 3.5 2.4 2.3

614 783 820 75 112 135 24 51 51 77 75 187 174

2.9 3.1 2.3 1.6 2.0 1.9 2.4 2.2 3.0 1.7 1.8 2.0 1.5

503 438 458 50 111 109 24 52 28 48 36 167 134

6.9

1225

5.0

956

6.7

681

5.0

550

4.3 6.8 6.1 5.9 6.7 7.1 7.2 7.2 7.6 8.3 5.9 6.4 4.5 6.2 7.8 4.9 5.5 6.6 6.0 6.5

79 537 1145 227 970 213 241 53 595 149 563 746 87 2831 1052 269 131 1739 176 10006

3.9 5.0 4.2 4.6 5.0 5.0 5.4 5.6 5.2 6.1 4.4 4.5 2.5 4.4 6.5 3.6 4.1 4.7 4.1 4.5

83 427 856 221 803 170 197 42 443 114 528 594 51 2079 962 231 113 1320 132 7749

6.9 7.0 4.0

5813 3913 248

5.0 4.8 3.0

4658 3041 222

5.2 8.0 3.7 8.4

52 128 110 179

4.3 5.1 3.0 6.8

49 91 87 192

asia China, Hong Kong India, Delhi India, Mumbai Japan, Hiroshima Japan, Nagasaki Singapore, Chinese Singapore, Malay Korea: Daegu Kuwait: Non-Kuwaitis Oman: Omani Pakistan: South Karachi Thailand: Bangkok Vietnam: Ho Chi Minh City

australia Australia, New South Wales New Zealand

europe Austria, Tyrol Belgium: Flanders Denmark Estonia Finland France, Isere Germany, Saarland Iceland Ireland Italy: Umbria Lithuania Israel: Jews Israel: Non-Jews The Netherlands Norway Slovenia Spain: Granada Sweden Switzerland, Zurich UK, England

north america Canada US, SEER White US, SEER Black

south america Argentina: Bahia Blanca Brazil: Goiania Costa Rica Uruguay: Montevideo

In general, however, rates are highest among white populations when compared with Black, Latino, Japanese, and Chinese populations worldwide (Parkin et al., 2002). A native Latino population in Brazil shows a higher incidence than the migrant Latino population in Los Angeles, but it is difficult to assess the validity of this finding due to the factors described above.

N

africa Algeria, Algiers

1181

Source: Parkin, et al. 2002. *A population was eligible for including if the registry had a microscopic confirmation rate of 70% or greater and if there were at least 10 cases of each sex. Only the largest registry meeting these criteria was selected for each country. ASR, age-standardized incidence rates.

is considerable variation in the manner in which reporting is done. One must also consider variations in the diagnostic efficiency among the different registries. Furthermore, when no African registry meets the criterion of a microscopic confirmation rate of 70% or greater and at least 10 cases aged 35–64 of each sex (Parkin et al., 1997), comparisons between native and migrant Black populations are not reliable.

A wide variety of occupations have been associated with brain tumors. Table 62–6 shows proportional incidence ratios (PIRs) for nonSpanish–surnamed white males, aged 20–64 and 65 and older at tumor diagnosis, in Los Angeles County from 1972–1999. Findings that are statistically significant at the p < 0.5 level are indicated with an asterisk. Only those job categories that contained greater than two cases were considered for statistical significance testing. There are no industries or occupations that are associated with both meningiomas and gliomas. Occupations associated with all primary CNS tumors include professional/technical workers, engineers, aero- and astronautical engineers, lawyers and judges, physicians, veterinarians, and managers and administrators. Industries associated with all primary CNS tumors include manufacturing (lumber and wood products, office and accounting machinery, radio broadcasting, paper and paper products), business and repair services, and professional and related services (offices of physicians, offices of dentists; elementary and secondary schools). In general, a greater number of occupations and industries were associated with gliomas than meningiomas. Occupations associated specifically with gliomas include electrical engineers, certain teachers, electrician apprentices, and animal caretakers. Industries with associations specific to gliomas were transportation equipment manufacturing, aircraft parts manufacturing, and electrical goods wholesale and retail trade.

ENVIRONMENTAL FACTORS This section reviews recent findings (through 2002) from epidemiologic studies, which have investigated the possible association of human brain tumors with various suggested risk factors. For most factors studied it updates and supplements more comprehensive earlier reviews (Inskip et al., 1995; Preston-Martin and Mack, 1996; Davis and Preston-Martin, 1998) and complements other recent reviews (Wrensch et al., 2002). Before beginning this review we will present findings from the experimental literature. One compelling experimental model, in particular, has been extensively investigated in epidemiologic studies, and this work is also summarized below.

Animal Models A variety of chemical, physical, and biological agents can cause nervous system tumors in experimental animals. Experimental studies of the effects of high-dose whole body irradiation in monkeys found the brain to be the most common site of the cancers that were produced (Traynor and Casey, 1971; Haymaker et al., 1972). Induction of glioblastoma multiforme was observed in primates that received fractionated whole-brain radiation therapy (Lonser et al., 2002); this finding implies that radio-induction of these neoplasms may occur as a late complication of this therapy and may occur more frequently than is currently recognized in human patients. Humans with high-dose exposure are also at excess risk, and the epidemiologic evidence that an environmental agent causes brain tumors appears strongest for ionizing radiation. Certain genetic factors have also been implicated in the pathogenesis of brain tumors, and these are likely to interact with environmental exposures. Mutations in apoptosis-associated genes, such as p53, may prevent apoptosis of neural precursor cells with DNA damage and predispose them to subsequent neoplastic transformation. Neural precursor cells from wild-type and gene-disrupted mouse embryos were subjected to ethyl-nitrosourea, and the results suggested that intrauterine exposure of neural precursor cells to certain DNA damaging agents

1182

PART IV: CANCER BY TISSUE OF ORIGIN Norway Iceland Germany, Saarland France, Isere US SEER Whites Australia Canada Finland Sweden United Kingdom, England Israel, Jews Netherlands Denmark Switzerland, Zurich Estonia Slovenia India, Delhi Austria, Tyrol China, Hong Kong Japan, Hiroshima Singapore, Chinese 0

1

2

Males Males

3 4 5 6 Age-standardized incidence rate

7

8

9

A Norway Iceland Germany Australia Finland Canada France, Isere US SEER Whites Sweden Estonia United Kingdom, England Israel, Jews Netherlands Denmark Switzerland, Zurich Austria, Tyrol Slovenia Females India, Delhi China, Hong Kong Singapore: Chinese Japan, Hiroshima 0

1

2

3

4

5

6

7

Age-standardized incidence rate

B Figure 62–4. (A) International comparison of standardized incidence ratios for malignant nervous system tumors in males, selected populationbased registries, 1993–1997. *A population was eligible to be included if the registry had a microscopic confirmation rate of 70% or greater and at least 10 cases. (B) International comparison of standardized incidence

ratios for malignant nervous system tumors in females, selected population-based registries, 1988–1992. *US SEER whites used as a referent. Any population was included if the registry had a microscopic confirmation rate of 70% or greater and if there were at least 10 cases.

may synergistically interact with specific genetic abnormalities to produce glial tumors in the adult brain (Leonard et al., 2001). N-nitroso compounds (NOCs), in particular the nitrosoureas, are the most studied neurocarcinogens in animal models. In some species, including various primates, tumors can be produced by relatively low levels of NOC precursors (sodium nitrite and ethyl urea) in the

animals’ food and drinking water. Because there is no reason to think that humans are less susceptible to these compounds, it is likely that NOCs cause cancer in them as well. Although NOC exposures in some occupational settings, particularly in past decades (e.g., machine shops; tire and rubber factories) were substantial, most people have low-level, but virtually continuous, exposure to NOCs throughout life.

1183

Nervous System

Table 62–6. Distribution of Primary Tumors of Brain and Cranial Meninges by Selected Occupations and Industries, Non-Spanish–Surnamed White Males, Los Angeles County, 1972–1999 Gliomas

Meningiomas

Primary CNS Tumors

PIR

CI

n

PIR

CI

n

PIR

CI

n

112.2* 142.0* 179.3*

103.5,121.5 122.1,164.1 133.5,235.8

610 182 51

96.6 88.3 105.2

78.9,117.2 55.9,132.4 38.4,228.9

103 23 6

111.6* 120.4* 150.0*

104.9,118.6 106.2,136.0 117.1,189.2

1034 260 71

153.5* 132.6* 126.9 251.0

106.9,213.5 101.4,170.4 93.9,167.8 67.6,642.5

35 61 49 4

89.4 106.1 132.5 139.7

24.1,228.9 50.8,195.1 63.4,243.7 66.9,256.9

4 10 10 10

127.1 115.6 128.0* 127.2*

94.0,168.1 93.0,142.2 102.3,158.3 101.0,158.1

49 90 85 81

112.3 139.7* 141.6* 337.2 111.3

36.2,262.2 109.5,175.6 104.4,187.8 67.8,985.3 44.6,229.4

5 73 48 3 7

435.9* 97.6 96.4 525.1 387.9*

117.3,1116 48.6,174.6 38.6,198.7 6.9,2922 104.4,993.1

4 11 7 1 4

118.9 140.9* 143.9* 326.6* 155.9

54.2,225.6 117.5,167.7 114.8,178.2 105.2,762.1 90.7,249.6

9 127 84 5 17

1265.8* 163.4* 104.7

142.2,4570 102.4,247.4 95.5,114.6

2 22 470

0.0 45.2 102.6

0.0,76015 9.1,132.1 82.8,125.7

0 3 93

67.4 1173.9* 170.4 101.9 413.5* 0.0

13.5,196.9 131.8,4238 81.6,313.5 32.8,237.8 111.3,1059 0.0,709.1

3 2 10 5 4 0

392.9* 0.0 244.9 508.9* 0.0 1682.9*

105.7,1006 0.0,16257 49.2,715.5 164.0,1188 0.0,1882 189.0,6076

4 0 3 5 0 2

705.9 118.1 110.0* 152.9* 104.1 741.2 172.5* 132.8 242.9 342.4

79.3,2549 91.8,149.4 102.7,117.7 113.8,201.1 44.8,205.2 83.2,2676 100.4,276.1 66.2,237.6 65.4,621.9 68.8,1001

2 69 837 51 8 2 17 11 4 3

257.8* 0.0 361.9*

133.1,450.4 0.0,1008 145.0,745.6

12 0 7

323.1 2640.1* 259.1

64.9,944.1 296.5,9532 3.2,1442

3 2 1

216.3* 484.1 276.1*

125.9,346.3 97.3,1414 126.0,524.2

17 3 9

307.8* 119.6* 123.8* 129.3

123.3,634.2 104.2,136.7 106.8,142.7 70.6,216.9

7 216 190 14

540.1 97.8 98.7 206.6

60.7,1950 68.1,136.0 66.6,140.9 55.6,528.9

2 35 30 4

289.8* 108.3 111.6 179.6*

144.4,518.5 96.9,120.7 99.1,125.3 123.6,252.2

11 328 287 33

233.7* 171.7 186.9 113.5

116.5,418.3 34.5,501.6 2.4,1040 41.4,247.1

11 3 1 6

119.8 909.0* 569.3* 198.7

1.6,666.4 182.7,2656 183.5,1329 22.3,717.4

1 3 5 2

163.7 202.0* 108.6 189.7*

87.1,279.9 73.8,439.7 1.4,604.0 110.4,303.7

13 6 1 17

161.3 108.6 164.4* 170.6 116.0 75.6 99.9 114.5

95.6,255.0 96.8,121.6 112.4,232.1 81.7,313.8 88.9,148.7 69.2,82.5 74.1,131.7 75.5,166.6

18 304 32 10 62 510 50 27

149.5 110.9 75.6 166.7 78.5 150.1* 91.6 360.4

30.0,436.7 84.2,143.4 15.2,220.9 18.7,602.0 33.8,154.7 131.6,170.4 39.5,180.6 97.0,922.7

3 58 3 2 8 238 8 4

158.4* 112.8* 152.1* 179.6* 122.3* 76.5 130.9* 151.7*

106.8,226.1 103.5,122.7 113.2,200.0 106.4,283.9 100.6,147.3 71.5,81.6 107.6,157.8 113.9,197.9

30 539 51 18 111 900 110 54

occupation Professional/technical workers Engineers Aero- and astronautical Engineers Electrical engineers Engineers Nec Lawyers and judges Lawyers Life and physical scientists Chemists Physicians, dentists, etc. Physicians, med and osteo Veterinarians Nurses, dieticians, and therapists Teacher, college and university Atmos, earth, marine, and space Teachers, exc. college and university, NOS Managers and administrators, exc farm Bank off and financial man Jewelers and watchmakers Electrician apprentice Pressmen and plate printers Machine operatives, NOS Animal caretakers except farm Lumbermen, craftsmen, woodchoppers

industry Manufacturing Lumber and wood products Logging Sawmills, planing And milling Office and accounting machinery Transportation equipment Aircraft and parts Radio broadcasting Wholesale and retail trade Electrical goods Paper and paper products Farm and garden supply store Not specified wholesale trade Business and repair services Business management and consulting Professional and related services Offices of physicians Offices of dentists Elementary, secondary schools Retired Disabled Student

Source: Data from USC/Los Angeles County Cancer Surveillance Program (CSP), the population-based tumor registry for Los Angeles County. *Indicates statistically significant (p < 0.05) PIR. PIR, proportional incidence ratios.

But, because NOCs are the most potent of carcinogens in animals (and likely in humans as well), only small doses are needed to cause cancer. Experimental work has demonstrated that other families of chemicals in addition to NOC and some infectious and physical agents can also produce brain tumors. Early work with polycyclic aromatic hydrocarbons produced brain tumors at the site of implantation. It has been suggested, however, that many of these earlier experimental findings should be re-evaluated to see if they meet various criteria of causality and to assure that all tumors diagnosed as primary brain tumors were correctly identified (Koestner, 1986; Swenberg, 1986; Rice and Wilbourn, 2000). Other work has demonstrated that systemic

administration of n-2-flourenylacetamide, various triazenes, and symmetrical hydrazines produce brain tumors in mammalian species (Maltoni et al., 1982). Chemicals that produce brain tumors after inhalation include bis-chloromethyl ether, vinyl chloride and acrylonitrile, and ethylene oxide (Maltoni et al., 1982; Garman et al., 1986). Brain tumors occur after intracranial inoculation with various viruses, and the histologic type of the tumor produced appears to depend on the site of inoculation and the age and species of the host animal (Bigner, 1978; Ohsumi et al., 1985). In addition, viruses and virus-like particles have been isolated from human nervous system tumors, but it is uncertain whether this finding has etiologic relevance

1184

PART IV: CANCER BY TISSUE OF ORIGIN

(Ohsumi et al., 1985; Corallini et al., 1987). Several studies have detected none or a very low frequency of virus sequences in samples from brain tumor patients, such as one study that found no samples with BK virus and only a small percentage of patient tumor samples with SV40 and JC (Weggen et al., 2000); the authors claim that this finding argues against a major role of these viruses in the pathogenesis of human medulloblastomas, meningiomas, and ependymomas (Weggen et al., 2000). Mouse models are providing new insights into the cellular origins of human brain tumors (Kernie et al., 2001; McCann, 2003). DePhino reports that breeding knockout mice that lack the tumor suppressor gene lnk4a/Arf with mice that produce excess epidermal growth factor receptor (EGRF), a known tumor promoter, develop tumors that look like human high-grade astrocytomas (McCann, 2003). Analysis of the tumors showed other genes that may increase lethality by interacting with lnk4a/Arf and EGRF; these genes might be good therapeutic targets for the treatment of such astrocytomas in humans (McCann, 2003). Other work has led to mice that produce a protein that inactivates the retinoblastoma (pRb) tumor suppressor gene in astrocytes, causing astrocytes to proliferate and produce astrocytomas (Xiao et al., 2002). This work suggests that pRb and PTEN may be useful drug targets in patients with astrocytomas. Another mouse model developed for oligodendrogliomas identified an immature cell as the precursors of these tumors (Weiss et al., 2002). Medulloblastomas developed in mice lacking two genes involved in cell growth, p53 and Patched-1 (Wetmore et al., 2001). Another relatively new area of exploration relates to the use of vaccination of various cytokines or of virus nucleoprotein to stimulate immune responses to tumors within the CNS. Vaccination with cytokine-producing tumor cells causes strong immune responses against tumors outside the CNS. Because the CNS has long been recognized as an “immunologically privileged” site it was thought unlikely that vaccination could generate such a potent response, but recent work refutes this belief. Granulocyte macrophage colony stimulating factor (GM-CSF) increased survival both in mice challenged with viable tumor cells in the brain and in mice with pre-established tumors (Sampson et al., 1996). Various other cytokines tested either had no effect (interleukin 4 (IL-4) and gamma-interferon) or decreased survival time (IL-2; Sampson et al., 1996). More recently, CNS tumor immunity in rats has been stimulated by vaccination with recombinant Listeria monocytogenes (Liau et al., 2002). Such animal models may help provide clues to explain the exciting epidemiologic findings over the past decade that a history of allergic diseases (Schlehofer et al., 1992; 1999; Brenner et al., 2002; Wiemels et al., 2002) or of certain viral infections (Wrensch et al., 2001) is related to a reduced risk of brain tumors (see below).

N-nitroso Compounds Experimental Data N-nitroso compounds (NOC) and their precursors are ubiquitous in our modern industrialized environment; they are also among the most potent experimental carcinogens (National Research Council Committee on Diet Nutrition and Cancer, 1982; Lijinsky, 1992). Based on their chemical structure, NOC are divided into two major subcategories—nitrosamines, which require metabolic activation, and nitrosamides, which do not. Only nitrosamides have been shown to cause NS tumors experimentally; they are derived from nitrosation of N-alkylureas, N-alkylcarbamates, and N-alkylamides (Tricker and Kubacki, 1992) and are formed under acidic conditions (Kakuda and Gray, 1980). The nitrosation of amides is catalyzed by citrate and other organic acids. Organic acids and thiocyanates are compounds that are naturally occurring in foods, and organic acids are also used as food additives (Yamamoto et al., 1988). N-nitrosamides are very unstable under neutral and alkaline conditions and decompose under heat exposure and when exposed to light (see Table 62–1) (Druckrey et al., 1967; Mirvish, 1995). Vitamin C and vitamin E, in fruits, vegetables, and grains inhibit nitrosation by reducing nitrite to NO, which is not directly a nitrosating agent (Mirvish, 1994). N-nitrosamides are direct alkylating compounds and

can lead to DNA adducts at the site of their occurrence. N-nitrosoureas in particular are effective nervous system carcinogens in a variety of species (Magee et al., 1976; Bogovski and Bogovski, 1981); they almost exclusively induce tumors in the nervous system (Druckrey et al., 1967). In contrast, N-nitrosamines induce cancer in several organs in animals but never tumors in the brain or spinal cord (Sampson and Bigner, 1998). Transplacental exposure to ethylnitrosourea (ENU) in various species including the rat, mouse, rabbit, opossum, and monkey through various routes of administration with either single-dose exposure or chronic low-dose exposure is particularly effective in causing neurogenic tumors (Rice and Ward, 1982); transplacental exposure requires only one-fiftieth (1/50) of the dose of ENU required in adult animals to cause 100% tumor induction (Ivankovic, 1979). This effect can be achieved if ENU precursors, ethyl urea and sodium nitrite, are added to the food and drinking water of a pregnant animal; however, no tumors develop if ascorbate (vitamin C) is also added to her diet (Mirvish, 1981). In rodents, these tumors occur in the brain, spinal cord, and peripheral nerves and are of various histologic types including gliomas, meningiomas, schwannomas, and sarcomas (Mirvish, 1981). In non-human primates, the nervous system tumors that are induced by transplacental exposure arise exclusively in the brain, have a histologic spectrum similar to that of pediatric brain tumors, and occur in young adult and immature monkeys (Rice, 1986). Ethylnitrosourea exposure leads to the formation of O6-ethylguanine, a mutagen removed more slowly from brain DNA than from DNA of other tissues (Rajewsky et al., 1976). The inability of the brain to remove this mutagen quickly coupled with the high rate of DNA replication in the fetal brain may explain why fetal animals are more sensitive than adults and why the brain is more sensitive than other tissues to the carcinogenic effects of ENU (Rajewsky et al., 1976); however, other investigators argue that this DNA alkylation may not be the key event in tumor development, although it may be a necessary one (Lijinsky, 1992). Other nitrosamides have also been shown to damage the DNA by production of O6-alkylguanine-DNA adducts (Craddock, 1983). It has also been suggested that these alkylated bases cause base-mispairing and point mutations that lead to uncontrollable expression of oncogenes and growth factor receptors followed by permanently heightened cell proliferation (Bilzer et al., 1989).

Population Exposure to N-nitroso Compounds Human exposure to NOC is estimated to derive half from exogenously and half from endogenously formed compounds (National Research Council Committee on Diet Nutrition and Cancer, 1982). Dietary NOCs have been found mainly in food that contains nitrite and/or has been exposed to nitrogen oxides, such as nitrite-cured and smoked meat and fish, cheese, and beer (Hotchkiss, 1989; Tricker and Preussmann, 1991). Until quite recently only levels of nitrosamines (not nitrosamides) were measured in human environments and consumer products, even though many of these exposures probably involve both nitrosamines and nitrosamides. The major sources of population exposure to nitrosamines in the United States are tobacco smoke, cosmetics, automobile interiors, and cured meats (National Research Council Committee on Diet Nutrition and Cancer, 1982). Only for a few individuals is the primary source of NOC exposure likely to be job related. Many cosmetics, soaps, shampoos, and hand lotions are contaminated with N-nitrodiethanolamine (NDELA), but only face make-up (i.e., foundation cream or liquid), was consistently found to contain NDELA, and at levels that were 30–40 times higher than those found in other personal hygiene products (Fan et al., 1977). Low levels of NDELA cause cancer in rats (Lijinsky and Kovatch, 1985). Cosmetics are also known to contain high levels of amides but have only occasionally been tested for nitrosamides (Fine et al., 1984; Chou et al., 1987). NOC are also present in pacifiers and in rubber baby bottle nipples and in the liquids passed through them (Sen et al., 1984). Endogenous formation of NOC in humans occurs in the stomach or bladder when both an amino compound and a nitrosating agent are present simultaneously. Food is a primary source of both highly con-

Nervous System centrated nitrite solutions (e.g., from cured meats) and amino compounds (e.g., in fish and other foods, but also in many drugs). Another source of nitrite is reduction (e.g., in the saliva) from nitrate, which comes predominantly from vegetables in the diet; this source is likely to be a far less important contributor to the NOC formed endogenously because it is highly diluted (and, therefore, less readily reactive) and because vegetables also contain vitamins, which inhibit the nitrosation reaction. Drinking water also contains nitrate (in the absence of vitamins), but this is a minor source unless levels are extraordinarily high (Chilvers et al., 1984). When inhibitors such as ascorbate (vitamin C) or alpha tocopherol (vitamin E), which are nitrite scavengers, are also present at high concentrations NOC are not formed (Tannenbaum and Mergens, 1980). Other compounds, such as caffeine, potentiate the NOC effect (Aida and Bodell, 1987). The level of NOC in the human body is also influenced by other factors such as which amino compounds are present, presence of bacteria or other nitrosation catalysts, gastric pH, and other physiologic factors. Uncertainty as to the simultaneous presence of NOC precursors and of inhibitors and/or catalysts of nitrosation, makes this hypothesis difficult to study epidemiologically. This difficulty is compounded by further uncertainty about what exposure period (during a person’s life) is most likely to be etiologically relevant.

Epidemiologic Evidence Findings from a preliminary study conducted in Los Angeles County indicated that the hypothesis that NOC cause brain tumors in humans deserves further investigation (Preston-Martin et al., 1982). Most investigation has been of possible dietary sources of NOC precursors and inhibitors; these studies are summarized in the section below.

Dietary Factors and Vitamin Supplementation Seven of nine studies published to date (Preston-Martin et al., 1982; Kuijten et al., 1990; Bunin et al., 1993; 1994; Cordier et al., 1994; McCredie et al., 1994b; Sarasua and Savitz, 1994; Preston-Martin and Mack, 1996; Schymura et al., 1996; Lubin et al., 2000) that investigated the role of maternal diet during pregnancy found a significant positive association between the frequency of maternal cured meat intake (individual or combined cured meats) and the risk of childhood brain tumors (CBT). The individual foods most consistently found to be associated with increased risk for CBT are hot dogs and bacon. The two studies that have failed to find an association were a small study in France and a larger study from Israel, a county in which cured meats are not commonly eaten (Cordier et al., 1994; Lubin et al., 2000). Many fewer studies of adult brain tumors have investigated frequency of consumption of cured meats as a possible risk factor, and in general findings are inconsistent; however, two recent studies in California and one in China did report an increased risk (Blowers et al., 1997; Lee et al., 1997; Hu et al., 1999; Tedeschi-Blok et al., 2001). Studies of brain tumors in adults and children in Israel used relatively complete semiquantitative food frequency questionnaires, which allowed analysis of intake of various macro and well as micro nutrients (Kaplan et al., 1997; Lubin et al., 2000); findings at this point seem anomalous in that they appear not to agree with earlier studies or with each other. The adult study found an increased risk with high protein intake, but a reduced risk with high intake of total fat and cholesterol (Kaplan et al., 1997). The study in children found that child’s intake of vegetable fat and the mother’s intake of potassium during gestation increased risk (Lubin et al., 2000). Two other studies used quite different sorts of data to suggest an association of total body burden of oxidants with brain gliomas in adults. The first used food frequency data and found that high nitrite plus low vitamin C intake increased risk (Lee et al., 1997). The second found that serum levels of ascorbic acid (vitamin C) and alpha- and gamma-tocopherol (vitamin E) were inversely related to glioblastoma risk (Schwartzbaum and Cornwell, 2000). Findings that use of vitamin supplements and/or high intake of fresh fruit or vegetables protect against brain tumor development might also be interpreted as supportive of the N-nitroso hypothesis, although this

1185

effect may be due to another mechanism (Preston-Martin et al., 1982; Howe et al., 1989; Preston-Martin and Mack, 1991; Bunin et al., 1993; 1994; McCredie et al., 1994a; Preston-Martin et al., 1998a). This lower brain tumor risk related to intake of fruits and vegetables has been a fairly consistent finding in studies of maternal diet and CBT, and is also reported in a number of studies in adults. Starting in the 1970s, the amount of sodium nitrite added to meat has been reduced from 200 to 120 ppm, and vitamin C has been added as a nitrosation inhibitor (United States Department of Agriculture, 1978). In 1986, regulations even further restricted sodium nitrite levels allowed to be added to meat to 100 ppm (United States Department of Agriculture, 1986). However, a literature review on sodium nitrite levels found in cured meat products purchased in US markets from 1970–1990, showed that sodium nitrite levels have decreased in all investigated products except in hot dogs in which nitrite levels actually appear to have increased since 1980 (Pogoda and Preston-Martin, 2001). This is an interesting observation because in epidemiological studies on cured meat intake during pregnancy and CBT, hot dogs are the cured meat products most consistently associated with increased risk of CBT. The experimental model and its potential relevance to humans are sufficiently compelling to encourage further investigation of this hypothesis despite the fact that it is a difficult one to test epidemiologically. Future studies must include more complete dietary histories if they hope to differentiate between findings supportive of the NOC/brain tumor hypothesis and those suggestive of other mechanisms for dietary effects. However, very limited information about preformed NUs in food, specifically in cured meat, or about their in vivo occurrence resulting from endogenous formation, is available. Lacking this important information, it seems unlikely that studies will be able to establish a causal relationship between dietary or endogenous exposure to N-nitrosoureas during pregnancy and the development of CBT in humans. In summary, hypotheses regarding dietary factors in CBT have focused on cured meats and associated N-nitroso-compounds, and on antioxidant inhibitors of nitrosation and their fruit and vegetable sources. Of these, the association with cured meats has received the most investigation, and the most support from case-control studies. Most of the case-control studies found significant positive associations between maternal cured meat intake and increased risk of childhood brain tumors. The main criticism these studies get is that recall bias with regard to dietary assessment might underlie these results (Blot et al., 1999). It often is the case that many years pass between the interview of the mother and the pregnancy. This can lead to poor recall but probably less likely to recall bias, unless the mothers thought that certain foods might have caused their child’s brain cancer. Since there is no other feasible method available for dietary assessment in casecontrol studies other than retrospectively administered questionnaires, future studies should at the least use full-length questionnaires asking about 100 or more food items, usual portion size eaten, and frequency of intake.

Radiation The occurrence of excess brain tumors after high-dose exposure to ionizing radiation is well established. The induction of glioblastoma multiforme after exposure to therapeutic brain irradiation in primates has been reported (Lonser et al., 2002). Numerous cases have been reported of brain tumors arising at the site of radiation treatment either for an earlier head tumor or for a benign condition such as a facial birthmark. Meningiomas appear to be the most common tumor that results, but tumors of other histologic types have occurred including glioblastoma multiforme (Hodges et al., 1992) and other gliomas in adults and children who received radiation treatment for an earlier cancer (Rimm et al., 1987; Shapiro et al., 1989; Soffer et al., 1989; Brada et al., 1992; Tsang et al., 1993; Little et al., 1998). Several studies of cohorts exposed to therapeutic radiation to the head for benign conditions of childhood have been done with the most informative being the Israeli follow-up of children irradiated for tinea capitis, which found the relative risk to be greatest for nerve sheath

1186

PART IV: CANCER BY TISSUE OF ORIGIN

tumors of the head and neck (RR = 33.1), intermediate for meningiomas (RR = 9.5), and lowest for gliomas (RR = 2.6, Ron et al., 1988). Studies of the incidence of primary brain tumors among those who were exposed to radiation from the atomic bombs in Hiroshima and Nagasaki have only relatively recently been able to show that this exposure caused excesses of various types of intracranial tumors. A study of 68 meningioma patients who were within 2 km of the hypocenter and 607 non-exposed meningioma patients concluded that meningioma is one of the tumors induced by atomic bomb radiation exposure in Hiroshima (Shintani et al., 1999). The incidence of meningiomas increased successively in three 5-year intervals since 1975 (5.3, 7.4, 10.1, and 14.9, respectively, Shintani et al., 1999); higher incidence was also found for those living close to the hypocenter (Shintani et al., 1999). A recent study of atomic bomb survivors reported a dose-related excess of brain tumors among 80,160 atomic bomb survivors from Hiroshima and Nagasaki for whom dose estimates could be computed (Preston et al., 2002). Using the Hiroshima and Nagasaki Tumor Registries, tumors of the brain, cranial and spinal nerves, pineal gland, and pituitary gland were ascertained among the cohort. A statistically significant dose response was found for all nervous system tumors combined and particularly for schwannomas (Preston et al., 2002). Risk increases were also seen for meningiomas, gliomas, and other nervous system tumors, but these were not found to be statistically significant. For tumors other than schwannomas, the risk was found to be greater for men and also for those who were exposed during childhood. The investigators concluded that even moderate doses of radiation are associated with increased incidence of brain tumors (Preston et al., 2002). Secondary brain tumors may also be associated with therapeutic radiation given for treatment of prior tumors. Children in France and the United Kingdom who received radiotherapy for a prior cancer found had an increased risk for benign and unspecified, but not malignant, brain tumors (Little et al., 1998). A pooled analysis of two Swedish cohorts of hemangioma patients with exposure to low-dose radiation found a dose-related increased risk of developing an intracranial tumor (Karlsson et al., 1998). The risk was also found to be higher among infants exposed at younger ages (Karlsson et al., 1998). A study of acute lymphocytic leukemia patients from a single institution who received radiation treatment found the cumulative incidence of brain tumors at 20 years to be 1.39% (95% CI: 0.63%–2.15%), and the relationship with cranial irradiation was dose dependent (Walter et al., 1998). The development of brain tumors related to prior exposure to diagnostic X-rays is less well established. A case-control study of 209 cases and 495 controls in Sweden found X-ray investigations of the head and neck region increased the risk for brain tumors (OR 1.64, 95% CI: 1.04–2.58), and this risk increased among those irradiated 5 or more years before tumor diagnosis (OR 2.10, 95% CI: 1.25–3.53) with an increase also seen for meningioma considered separately (Hardell et al., 2001). No association was seen with radiographic investigations of other parts of the body. Prior dental radiography, particularly in years when doses were high, has been found to be associated with subtentorial intracranial meningiomas (Preston-Martin and White, 1990). Other studies, however, have found no association between exposure to radiation and brain tumors. A particularly controversial area of study has been nasopharyngeal radiation given for a variety of conditions. Persons given X-ray therapy to the tonsils and nasopharynx as children have been shown to have an excess brain tumor risk, and each major histologic type, including several different types of gliomas, are represented among the tumors that occur (Schneider et al., 1985). A study of subjects who received nasopharyngeal irradiation at the Health Department’s Clinic for the Prevention of Deafness in Children in Washington County, Maryland reported an excess of brain tumors but also warned that the results could have been due to chance (Sandler, 1996). A follow-up study found that radiation therapy for adenoid hypertrophy was associated with a very large though not statistically significant relative risk of 14.8 (95% CI: 0.8–286.3, Yeh et al., 2001). A retrospective cohort study of people who received

nasopharyngeal irradiation for the treatment of otitis serosa or barotraumas found no excess deaths from cancers of the head and neck area (Ronckers et al., 2001). A study of submariners who received nasopharyngeal treatment for barotraumas found no significant increase in brain cancer risk (George, 1996). Much controversy has surrounded the observation that the incidence of childhood cancer around nuclear power plants appears to be increased. In a study in Scotland more CNS tumors than expected were seen among children living within 25 km of each of three nuclear sites; this elevation in relative risk (RR) was statistically significant only at Rosyth, which had by far the greatest number of total cases and the largest RR (Sharp et al., 1999).

Electric and Magnetic Fields The several studies that have explored the association between electromagnetic field (EMF) exposure and brain tumor risk have been recently reviewed (Kheifets, 2001). Although early studies of childhood brain tumors and residential exposure (both from ambient fields and specific use of electrical appliances) were suggestive of an increased risk, findings to date have been inconsistent and have been plagued by methodologic weaknesses ( Kheifets et al., 1999; Kheifets, 2001). No increase in CNS tumors was seen relating to measured fields in homes in a recently completed study of childhood cancer in Great Britain (Skinner et al., 2002). Fewer studies have looked at residential exposure and adult brain tumors; little evidence of an association is seen (Wrensch et al., 1999; Kheifets, 2001). A meta-analysis of data from 29 studies of a possible association of occupational EMF exposures and brain cancer found a small increase in risk among electrical workers that was strongest for a few specific jobs and for gliomas (Kheifets et al., 1995), but findings are weakened by exposure misclassification and substantial heterogeneity across studies (Kheifets, 2001). No excess risk of brain tumors was seen among electricity workers in England and Wales when cumulative exposure was considered (Sorahan et al., 2001).

Cellular Telephones and Other Radiofrequency Exposures Reports of brain tumors arising on the side of the head of early heavy users of cellular phones led to several studies in a number of countries to assess the relationship of cell phone use and brain tumor risk. No excess risk was found in a number of large case-control studies (Hardell et al., 1999; Muscat et al., 2000; Inskip et al., 2001) or in a Danish cohort study (Johansen et al., 2001), although a Finnish register-based study found a weak association between gliomas and analog cellular phones (Auvinen et al., 2002). Two studies noted that the brain tumor was more likely to occur on the side of the head where the phone was most often held (Muscat et al., 2000; Hardell et al., 2001). No excess of brain cancer was found among employees of a company that manufactures these phones (Morgan et al., 2000). Early cell phone use was more likely to be higher because of the analog technology used. Cell phone use is becoming increasingly prevalent, and newer digital phones operate in a somewhat different frequency range. No studies have adequately investigated possible late effects of long-term heavy cell phone use. A recent review summarizes research on health effects of exposure to radiofrequency fields from cell phones and other sources (Krewski et al., 2001). US Air Force employees, for example, were found to have a small excess of brain tumors related to radiofrequency/microwave exposures (Grayson, 1996).

Trauma The epidemiologic information that connects brain tumor development to head trauma is sparse. Several case reports and small series exist associating severe prior trauma with subsequent brain tumor development, but larger studies have not consistently shown a relationship. One interesting report discusses a case where a malignant glioma occurred in the exact spot where a man suffered a metal splinter injury 37 years earlier (Sabel et al., 1999); most previous case reports of this sort have involved meningiomas, however (Walsh et al., 1969). Overall, the evidence is strongest for meningiomas. In an inter-

Nervous System national case-control study for the role of head trauma in the occurrence of glioma and meningioma, there was only an increased risk of meningioma among men who had a serious head injury 15–24 years before meningioma diagnosis (Preston-Martin et al., 1998b). However, the authors caution the reader in interpreting the data used in this study because of the possibility of biased recall because those who developed a brain tumor may be more likely to remember a previous head injury, even when an attempt is made to limit reporting to only the most severe prior head trauma. Nonetheless, the specificity of the findings to only one type of brain tumor, and to the type most associated with trauma in other studies, makes them more believable. Another study also has shown a weak relationship between trauma and meningiomas (Longstreth et al., 1993), and the authors speculate that the reparative process of the dura may lead to tumor formation. Large studies from Sweden and Denmark have not shown a relationship (Inskip et al., 1998; Nygren et al., 2001). In children, head trauma such as from forceps delivery has in some studies been associated with increased risk of gliomas (Gurney et al., 1996).

Acoustic Trauma and Acoustic Neuromas The observation that over 90% of all cranial nerve sheath tumors arise in the eighth cranial nerve (the acoustic nerve) suggests an exposure unique to this nerve. A case-control study of acoustic neuromas in Los Angeles County residents supports the hypothesis that acoustic trauma may relate to the development of these tumors (Preston-Martin et al., 1989). During the period 10 or more years before the year of diagnosis of the case, more cases than controls had a job involving exposure to extremely loud noise; noise exposure was determined by a blinded review of job histories and linkage to the National Occupational Hazards Survey data base (odds ratio (OR) = 2.2, 95% CI: 1.12, 4.67). A dose-response analysis showed an increase in risk related to number of years of job exposure to extremely loud noise (P for trend = 0.02) with an OR of 13.2 (CI: 2.01, 86.98) for exposure of 20 or more years accumulated up to 10 years before diagnosis. These findings may support the more general hypothesis that mechanical trauma may contribute to tumorigenesis (Preston-Martin et al., 1989). This contention is supported by experimental findings of tissue destruction and subsequent repair following acoustic trauma (Hamernik et al., 1984a; 1984b; Corwin and Cotanche, 1988; Ryals and Rubel, 1988).

Spinal Meningiomas and Osteoporotic Vertebral Fractures: A Hypothesis Although spinal meningiomas and other spinal tumors are rare, the preponderance of spinal meningiomas in women and the sharp postmenopausal rise in the incidence of this tumor are both striking (Preston-Martin et al., 1995). A series of three studies among women with spinal tumors explored the hypothesis that this unusual distribution relates to the higher prevalence of spinal osteoporosis among postmenopausal women (Preston-Martin et al., 1995). First, a review of medical records showed that meningiomas in women, unlike other spinal tumors, usually arise in the mid thoracic spine where osteoporotic vertebral fractures predominate. Secondly, radiographic evidence of osteoporosis was seen commonly with meningiomas but not with other spinal tumors. Finally, a case-control interview study found that factors that reduce the risk of osteoporosis also were related to reduced risk of spinal meningiomas: use at the time of interview of estrogen replacement therapy (OR = 0.2, 95% CI: 0.1–0.6); past use of oral contraceptives (p trend <0.01); past participation in sports (OR = 0.5; CI: 0.2–0.9); and premenopausal status (OR = 0.2; CI: 0.1–0.7).

Infectious Agents Several types of viruses, including retroviruses, papovaviruses, adenoviruses, JC virus (Khalili, 2001), and simian immunodeficiency virus (Chretien et al., 2000) cause brain tumors in experimental animals. In one of the earliest epidemiologic studies of brain tumors, astrocytomas, but not other histologic types, appeared to be associated with positive antibody titers to Toxoplasma gondii (Schuman et al., 1967), but more recent studies have failed to confirm this association (Ryan et al., 1993).

1187

People who received polio vaccine contaminated with SV40 have been investigated to determine their risk of developing brain tumors, and a study using SEER data found an increased risk for brain tumors, specifically ependymomas (Fisher et al., 1999). However, another study using SEER data, the Connecticut Tumor Registry, and national mortality statistics found that exposure to SV40-contaminated poliovirus vaccine was not associated with significantly increased rates of ependymomas and other brain cancers (Strickler et al., 1998). History of varicella-zoster infection was related to reduced risk of adult gliomas whether history of chickenpox or shingles or immunoglobulin G antibodies to the virus were used in the analysis (Wrensch et al., 1997; 2001). Other studies have been performed to determine the risk of maternal infections during pregnancy and childhood brain tumors. A retrospective cohort study examining whether an association exists between community infections during pregnancy and solid tumors found an increased risk for children exposed to measles and influenza (Dickinson et al., 2002). A case-control study found a significantly increased risk of brain tumors and neuroblastoma in offspring after exposure of mothers to influenza during pregnancy, but results should be viewed only as suggestive because of the lack of serologic confirmation of infection (Linos et al., 1998).

Allergies, Epilepsy, and Other Medical Conditions Allergic Conditions An intriguing new finding relates a history of asthma, eczema, and other allergic conditions to a reduced risk of brain tumors in adults. This finding has been consistent across a number of large case-control studies conducted in various countries (Schlehofer et al., 1992; 1999; Brenner et al., 2002; Wiemels et al., 2002). One study also reports an inverse association of autoimmune disease with both glioma and meningioma, and the greatest reduction in risk was seen among glioma patients with both asthma and diabetes (Brenner et al., 2002). Three cohort studies using Swedish data provide limited support for this association (Schwartzbaum et al., 2003); the author suggests that the association seen in the case-control studies appears to be stronger when proxy interview data are used for cases who were not available for personal interviews. Further work is currently underway to characterize immune function in brain tumor patients vs. controls using biological samples.

Epilepsy Substantial proportions of patients with various types of glial tumors, meningiomas, and other brain tumors report a history of seizures (Lieu and Howng, 2000; Hwang et al., 2001; Tandon et al., 2001). The association with seizures is strongest for recent seizures suggesting that the seizure is an early brain tumor symptom; however, the association is also seen both for adult and childhood brain tumors with seizures occurring many years before brain tumor diagnosis (Gurney et al., 1997; Schlehofer et al., 1999).

Other Medical Conditions Various studies have found brain tumor excesses among groups of patients with other medical conditions including glioblastomas and meningiomas in stroke patients (Dobkin, 1985; Mills et al., 1989) and those with multiple sclerosis (MS) (Reagan and Freiman, 1973), but it could be that differential and/or more complete diagnosis may be an issue (Morgenstern and Frankowski, 1999). Stroke has also been reported as a late effect among adult survivors of childhood brain tumors (Gurney et al., 2003). No cancer excess was found among Norwegians with MS (Midgard et al., 1996). An excess of meningiomas has been evident for decades among women with breast cancer (Kubo et al., 2001), and given the abundance of progesterone receptors in meningiomas and some breast cancers a hormonal link has been suggested (Blankenstein et al., 2000; Kubo et al., 2001). Swedish data were used to show an increased risk of meningioma, but not astrocytoma, in patients with either breast or colon cancer (Malmer et al., 2000). Fewer brain tumors than expected have been seen in diabetics (Schlehofer et al., 1992; Brenner et al., 2002).

1188

PART IV: CANCER BY TISSUE OF ORIGIN

Occupational Exposures Although many epidemiological studies have investigated the relation between work in a variety of occupational/industrial settings and brain tumor incidence/mortality, repeated studies in various geographic areas have been reported for only a few occupations. Epidemiologic studies in this area face serious limitations. Workers in many occupations and industries are exposed to a variety of neurotoxic and/or carcinogenic substances, but animal studies have shown that brain tumors are generally not induced by inhalation or dermal exposures (Wrensch et al., 2002). Moreover, the number of brain tumor cases in even the largest occupational studies are often too small to permit meaningful analyses (Wrensch et al., 2002). One area of interest in epidemiological studies is the petrochemical industry. In 1985 and 1986, Texaco reported on the patterns of mortality among their refineries, petrochemical plants, and research facilities; the researchers found increased mortality ratios for brain cancers and several other cancers (Divine et al., 1985; Divine and Barron, 1986). A follow-up study was conducted 16 years later to determine if any changes had occurred in the patterns of mortality within this group. Although the overall mortality was lower or similar to the general US population, slightly increased mortality was found for brain cancers (Divine et al., 1999). A cluster of glioma cases was reported in a particular building complex at the Amoco Research Center, and this prompted further study of a possible association between occupational exposure and glioma. Although the employees had 15% fewer cancers than the general population, an excess of brain cancer was found in white men who had 10 or more years since hire or 5 or more years of employment in the particular building complex of interest (Beall et al., 2001b). A case-control study of workers at a petrochemical research facility focused on any associations between chemical and physical agents and brain tumors. Interviews were conducted with brain tumor patients and matched controls on possible exposures to a variety of agents. A positive association was found between gliomas and ionizing radiation, n-hexane, organometallics, and amines other than nitrosamines (Beall et al., 2001a). Agricultural exposure to pesticides and other agricultural chemicals have been extensively studied due to the carcinogenic effects of these chemicals. A review of case-control and cohort studies found equivocal results with regard to the risk for brain tumors (Bohnen and Kurland, 1995). However, insecticide and fungicide exposure of women in the United States was associated with a small increased risk for brain tumors (Cocco et al., 1999). Health professionals have also been investigated for possible brain tumor risks. A case-control study of gliomas diagnosed among adults in the San Francisco Bay area between 1991 and 1994 found that glioma patients were more likely than controls to be physicians and surgeons; no specific specialty was found to be at increased risk and the results could also have been due to chance or might be attributable to the overall higher rates among those in higher social classes, such as these professionals (Carozza et al., 2000). Epidemiological studies of brain tumor deaths have reported elevated brain tumor risk for a variety of electricity-related occupations involving exposure to electromagnetic fields (EMF). These include electrical or electronic engineers, electronics teachers, and electrical technicians and assemblers (Loomis and Savitz, 1990), electric utility workers (Savitz et al., 2000), and workers in the communications industry (Speers et al., 1988). A recent review on this subject found that the quality of epidemiological studies relating to EMF exposure has improved and there is a large body of high-quality data on brain tumor risk and occupational EMF exposure (Ahlbom et al., 2001). No excess brain tumor risk was seen in a cohort of British electrical utility workers (Sorahan et al., 2001). A case-control study conducted in eight Canadian provinces, with emphasis being placed on variations in EMF-related risk across different histological types found a pronounced risk for glioblastoma multiforme, but no association for astrocytoma or other brain tumors (Villeneuve et al., 2002). A recent review summarizes studies looking at occupational EMF exposures and brain cancer risk (Kheifets, 2001).

Parental Occupation and Childhood Brain Tumors Several studies have investigated possible associations between occupational exposures of parents and the development of brain tumors in their children. The results have been various, but for some jobs or exposures findings across studies appear somewhat consist. A recent review found that 48 published studies on this topic have reported relative risks for over 1000 specific occupation/cancer combinations, but several limitations related to the quality of the exposure assessment, small numbers of exposed cases, multiple comparisons, and possible bias toward the reporting of positive results limit the usefulness of these studies (Colt and Blair, 1998). A case-control study of 114 patients aged 0–14 diagnosed with neuroblastoma and 372 age-matched controls found positive associations for maternal occupation at the time of the child’s birth in service or retail industries and paternal occupation in materials handling (Kerr et al., 2000); a positive association was also found for maternal job exposure to acetone, insecticides, lead, and petroleum and paternal exposure to creosote, dioxin, lead, and petroleum (Kerr et al., 2000). An international case-control study conducted in seven countries examined parental occupation 5 years before the child’s birth (Cordier et al., 2001). An increased risk was found with agricultural work for all brain tumors combined and for other glial tumors (i.e., not astrocytomas or medulloblastomas). Increased risks for all tumors were also seen for paternal occupation as an electrician, driver, or mechanic, and maternal work in an environment related to motor vehicles. A study conducted in Norway found an elevated incidence of brain tumors in children whose parents were pig farmers, and the results were stronger for children who had both parental exposure and grew up on a farm (Kristensen et al., 1996).

Lifestyle Smoking and Exposure to Environmental Tobacco Smoke Most studies of brain tumors and tobacco exposures have investigated possible associations of parental smoking and childhood brain tumors, perhaps because of early findings of an increased risk with paternal, but not maternal smoking (Preston-Martin et al., 1982). Recent reviews and a meta-analysis continued to find little suggesting that mother’s smoking during pregnancy increased CBT risk, but, based on 10 studies, found a small increase in risk (OR = 1.2; CI: 1.1–1.4) related to father’s smoking (Norman et al., 1996; Boffetta et al., 2000). An international collaborative CBT study found no association of risk with parental smoking before pregnancy, maternal smoking or environmental tobacco smoke (ETS) during gestation, or ETS exposure of the child during its first year of life; results were no different in subgroup analyses based on histologic type or child’s age at diagnosis (Filippini et al., 2002). Tobacco use has not been consistently or strongly related to adult brain tumors (Zheng et al., 2001).

Alcohol Findings relating to parental alcohol intake are mostly unremarkable, although a recent study in China found an increase in CBT risk for fathers who drank hard liquor before conception (Hu et al., 2000). The same investigators found that adult glioma and meningioma patients reported higher intake of beer and other alcohol (Hu et al., 1999). Most studies of adult brain tumors, however, find no excess, and possibly a reduced risk related to intake of wine and beer (Preston-Martin and Mack, 1996).

Use of Pesticides Several epidemiologic studies have investigated home and occupational use of pesticides, insecticides, or herbicides as possible etiologic factors for brain tumors. Case-control studies have linked household pest exterminations to the development of childhood brain tumors (Gold et al., 1979; Davis et al., 1993; McCredie et al., 1994b), whereas others have found no association with parental pesticide use (PrestonMartin et al., 1982; Howe et al., 1989; McCredie et al., 1994a). A large

Nervous System case-control study found an excess risk of CBT related to prenatal exposure to flea and tick products (Pogoda and Preston-Martin, 1997). In a recent review, most of 17 studies of maternal and childhood pesticide exposure and CBT found elevated risks with only three studies not reporting positive associations (Zahm, 1999). A cohort study of licensed pesticide applicators found an excess risk of brain cancer (SMR = 200, Blair et al., 1983). More studies are needed to confirm associations seen and to identify which compounds may relate to brain tumor development.

Hair Dyes and Cosmetics An early study reported that more adults with brain tumors used hair dyes and hair sprays (Burch et al., 1987); another early study suggested that maternal use of face make-up during pregnancy might increase the risk of CBT (Preston-Martin et al., 1982). A large study of CBT in 19 counties on the US West Coast found no consistent evidence of an association of risk with maternal use of hair dyes (Holly et al., 2002).

Traffic-Related Air Pollution A recent study found that children residing during gestation and/or childhood in areas with high traffic density did not have increased CBT risk (Raaschou-Nielsen et al., 2001).

Drinking Water Some studies have focused on drinking water content and adult brain tumors such as one that found an increased risk of gliomas in men, but not in women, related to chlorinated water (Cantor et al., 1999). More studies have investigated maternal drinking water during pregnancy and subsequent development of CBT among offspring. A large population-based study in 17 counties found that in the 13 counties in Washington State CBT risk was increased among offspring of women who used well water exclusively during the pregnancy, but this association was not seen in the California counties (Mueller et al., 2001). A similar analysis using data from a seven-center international study found a similar association with well water use in two regions only; it also suggested an increased risk of astrocytoma related to increasing levels of nitrite in the drinking water (Mueller et al., 2004).

HOST FACTORS Familial Aggregation The potential heritability of gliomas and primary CNS tumors and familial clustering of CNS tumors is a well-documented phenomenon (Farwell and Flannery, 1984). The implication of such reports is that shared genes, environment, or gene-environment interaction could be responsible for some neoplasms, but reports estimate that familial gliomas account for 5% or fewer of total cases (Malmer et al., 2001; Osborne et al., 2001). The Johns Hopkins Brain Tumor Registry has accumulated 72 families with multiple gliomas, and interestingly, child’s diagnosis often preceded the diagnosis in the parents (Grossman et al., 1999). One report of “familial glioma” involved four cases of glioblastoma multiforme in a single generation (Dirven et al., 1995). A Swedish population-based analysis of the risk for people with first-degree relatives with gliomas looked at families of 432 patients with glial tumors and showed an overall threefold elevation in risk, which was strongest for relatives of young patients (Malmer et al., 1999). Persons with relatives with gliomas were 2.3 more likely to be cases in a case-control study in adults (Wrensch et al., 1997). Other studies do not support these findings, such as a study that found no difference between cases and controls in the proportions of firstdegree relatives with gliomas (Cicuttini et al., 1997), and a population-based study from Iceland also did not show evidence for familial aggregations of glial tumors (O’Neill et al., 2002). Twin studies have been similarly inconsistent. One identical twin study did not demonstrate concordance of CNS tumors in twins (Harvald and Hauge, 1956), but another reports on 8 pairs of identi-

1189

cal twins with concordant tumor pathology and age at diagnosis (Tijssen, 1985).

Neurogenetic Syndromes Various central nervous system hereditary tumor syndromes have long been characterized clinically and now are characterized genetically as well. Known hereditary syndromes are rare and are likely to account for 5% or fewer of CNS neoplasia (Preston-Martin, 1996). These syndromes demonstrate clear genetic predisposition, are most often autosomal dominant, and have variable penetrance in those harboring the genes. The most common CNS hereditary tumor syndrome is neurofibromatosis-1, also known as von Recklinghausen disease (Feldkamp et al., 1998). Clinical features are myriad and include multiple caféau-lait spots, multiple cutaneous neurofibroma nodules, axillary freckling, iris hamartomas (Lisch nodules), sphenoid bone dysplasia, gliomas of optic nerves and hypothalamus, and hamartomatous cerebral lesions. The NF-1 gene functions as a tumor suppressor gene and its loss of function precedes syndrome development. Chromosome 17q11.2 is where the gene has been mapped, and 50% of cases are spontaneous. Neurofibromatosis-2 is an autosomal dominant disease with the gene mapped to the long arm of chromosome 22. The disease occurs in 1 in 40,000 births with over 50% occurring because of spontaneous mutations (Kluwe and Mautner, 1998). The principle manifestation of this disease is bilateral vestibular schwannomas (acoustic neuromas). Also, cranial or spinal meningiomas can occur, which can be multiple or en plaque as well. Cataracts occur with this syndrome in up to 85% of people. Von Hippel-Lindau (VHL) disease describes a syndrome of multiple hemangioblastomas. Multiple hemangioblastomas are found along the CNS axis with the most common occurring in the cerebellum. Other associated systemic findings include retinal lesions, pheochromocytoma, and cysts of the pancreas, epididymis, liver, and kidneys. These people are also prone to developing renal cell cancer, which is the most common cause of death with this syndrome (Maher et al., 1997). The VHL gene has been mapped to the short arm of chromosome 3 and has autosomal dominant inheritance. The Li-Fraumeni syndrome describes a familial aggregation of breast cancer, sarcoma, leukemia, and brain tumors transmitted in an autosomal dominant fashion. The syndrome is linked to a germline mutation of the p53 tumor suppressor gene on chromosome 17p (Malkin et al., 1990). This causes abnormal cell cycle regulation and increases cancer susceptibility. Another autosomal dominant disorder initially described by Bourneville is called tuberous sclerosis (TSC). Cases may also be sporadic as well and involve mutations at the TSC1 and TSC 2 gene loci on chromosomes 9 and 16, respectively (Malkin et al., 1990). The clinical triad includes sebaceous adenomas, mental retardation, and seizures. CNS manifestations include subependymal tubers that protrude into the ventricular system. If these lesions show progressive growth they are called subependymal giant cell astrocytomas. Turcot syndrome describes an association between colon cancer and neuroepithelial tumors. Also transmitted in autosomal dominant fashion, the gene mutation occurs on chromosome 5q21 and is called the adenomatous polyposis coli gene (Paraf et al., 1997). These patients can present with thousands of colonic polyps or colon cancer and simultaneous gliomas. Gorlin syndrome again is autosomal dominant and composed of multiple basal cell carcinomas, intracranial calcifications, skeletal abnormalities, and an association with medulloblastoma formation (Farndon et al., 1992). The syndrome is also known as nevoid basal cell carcinoma syndrome. Genetic mapping has determined loss of an allele at locus 9q31 to account for the heritability of the disease. Traced to a germline mutation of the PTEN gene on chromosome 10q, Cowden syndrome is also known as the multiple hamartoma syndrome. The syndrome manifests clinically secondary to hypermyelination of the cerebellar folia leading to gross enlargement of the

1190

PART IV: CANCER BY TISSUE OF ORIGIN

cerebellum. Patients suffer from multiple hamartomas throughout the body and are also at increased risk for cancer of the breast and thyroid.

PREVENTIVE MEASURES Ionizing radiation can cause all three major histologic types of brain tumors—gliomas, meningiomas, and nerve sheath tumors—but the association appears weakest for gliomas. Nonetheless, minimizing population exposure to X-rays of the head is, at this point, the best prospect for prevention of all three types of tumors. Currently, however, the higher-dose exposures are most commonly from radiation treatment to the CNS for a prior cancer, and such treatment is expected to improve survival. Therefore, the benefit vs. risk to the individual patient must be considered before decisions are made. Beyond this, future research may identify those whose genetic makeup puts them at higher risk of NS tumors if they are exposed to specific chemical, physical, or other neurocarcinogens. However, given the rarity of NS tumors in the general population, screening for such susceptibility markers seems unlikely to be feasible except among those with genetic syndromes that predispose to CNS tumor development.

FUTURE DIRECTIONS Why have the causes of NS tumors been so elusive, despite several decades devoted to investigating their etiology? The answer relates to their vast morphologic, genetic, and etiologic diversity, which has become increasingly clear over recent decades. Etiologic associations are likely to be specific for tumors with common characteristics, and thus future epidemiologic studies must consider subgroups of patients whose tumors have similar morphologic and genetic characteristics. For many subgroups, to obtain sufficient numbers of subjects it will become necessary to pool data from studies across geographic areas. International collaborations among investigators who have already collected or plan to collect questionnaire information and biological samples have already been initiated. The goal is to use samples and data to look at commonly agreed upon genetic markers and exposures to elucidate factors related to the etiology of such defined subgroups of brain tumor patients. Most epidemiologists involved in the study of NS tumors are now aware that the strongest chance for further progress lies in such cooperative work. We have the most to gain using this approach in studies of the etiology of gliomas in that causes remain largely unknown. Gliomas appear far more morphologically and genetically diverse than do the other major categories of CNS tumors. Perhaps more is known about the etiology of meningiomas and nerve sheath tumors because each of these major groups represents a more homogeneous entity than do gliomas. Diet will continue to be an important focus of the next generation of epidemiologic studies of gliomas. The first generation of studies included some questions about a limited number of dietary variables, such as the several studies that looked at foods thought likely to be relevant to the N-nitroso hypothesis. Several intriguing associations were suggested by these and other studies, but no consistent associations have as yet emerged from more recent studies, which collected relatively complete dietary data to adequately evaluate associations with various macro- as well as micro-nutrients. The prospects for studies of brain tumors in children warrant separate discussion. Most brain tumors in children are gliomas. Some types of neuroepithelial tumors such as medulloblastomas and juvenile pilocytic astrocytomas occur, in fact, predominantly in children. Animal experiments strongly suggest that pediatric brain tumors may be caused by exposures during gestation. For example, the hypothesis that NOC may cause brain tumors, which is difficult to test epidemiologically, may more easily be investigated in children since the relevant exposure period (gestation) is more clearly defined for children than it is for adults and since the effect is likely to be stronger when exposure occurs during the period of rapid development of brain tissue. But, childhood brain tumors, like adult brain tumors, are also

morphologically and genetically diverse, and they are far rarer. Further, for CBT when exposures are considered, investigations must consider not only the patient’s but also the mother’s polymorphisms in genes that relate to the metabolism of chemicals and the repair of DNA damage. As always, these studies will further our understanding of the etiology of these tumors only if the right questions are being asked. For gliomas in particular, it still seems possible that some of the crucial etiologic questions remain to be posed. References Ahlbom IC, Cardis E, Green A, Linet M, Savitz D, Swerdlow A. 2001. Review of the epidemiologic literature on EMF and Health. Environ Health Perspect 109 (Suppl 6):911–933. Aida T, Bodell WJ. 1987. Effect of caffeine on cytotoxicity and sister chromatid exchange induction in sensitive and resistant rat brain tumor cells treated with 1,3-bis(2-chloroethyl)-1-nitrosourea. Cancer Res 47: 5052–5058. American Cancer Society Inc. 2003. Cancer facts & figures. Surveillance Research, Atlanta, Georgia. Auvinen A, Hietanen M, Luukkonen R, Koskela RS. 2002. Brain tumors and salivary gland cancers among cellular telephone users. Epidemiology 13:356–359. Bailey P, Cushing H. 1926. A Classification of Tumors of the Glioma Group on a Histogenic Basis. Philadelphia: JB Lippincott. Beall C, Delzell E, Rodu B, Sathiakumar N, Lees PS, Breysse PN, Myers S. 2001a. Case-control study of intracranial tumors among employees at a petrochemical research facility. J Occup Environ Med 43:1103– 1113. Beall C, Delzell E, Rodu B, Sathiakumar N, Myers S. 2001b. Cancer and benign tumor incidence among employees in a polymers research complex. J Occup Environ Med 43:914–924. Bigner DD. 1978. Role of viruses in the causation of neural neoplasia. In: Biology of Brain Tumors. Laerum OD, et al., eds. Geneva: International Union Against Cancer. Bigner SH, Mark J, Mahaley MS, Bigner DD. 1984. Patterns of the early, gross chromosomal changes in malignant human gliomas. Hereditas 101: 103–113. Bigner SH, Wong AJ, Mark J, Muhlbaier LH, Kinzler KW, Vogelstein B, Bigner DD. 1987. Relationship between gene amplification and chromosomal deviations in malignant human gliomas. Cancer Genet Cytogenet 29:165–170. Bilzer T, Reifenberger G, Wechsler W. 1989. Chemical induction of brain tumors in rats by nitrosoureas: Molecular biology and neuropathology. Neurotoxicol Teratol 11:551–556. Blair A, Grauman DJ, Lubin JH, Fraumeni JF, Jr. 1983. Lung cancer and other causes of death among licensed pesticide applicators. J Natl Cancer Inst 71:31–37. Blankenstein MA, Verheijen FM, Jacobs JM, Donker TH, van Duijnhoven MW, Thijssen JH. 2000. Occurrence, regulation, and significance of progesterone receptors in human meningioma. Steroids 65:795–800. Blot WJ, Henderson BE, Boice JD, Jr. 1999. Childhood cancer in relation to cured meat intake: Review of the epidemiological evidence. Nutr Cancer 34:111–118. Blowers L, Preston-Martin S, Mack WJ. 1997. Dietary and other lifestyle factors of women with brain gliomas in Los Angeles County (California, USA). Cancer Causes Control 8:5–12. Boffetta P, Tredaniel J, Greco A. 2000. Risk of childhood cancer and adult lung cancer after childhood exposure to passive smoke: A meta-analysis. Environ Health Perspect 108:73–82. Bogovski P, Bogovski S. 1981. Animal species in which N-nitroso compounds induce cancer. Int J Cancer 27:471–474. Bohnen NI, Kurland LT. 1995. Brain tumor and exposure to pesticides in humans: A review of the epidemiologic data. J Neurol Sci 132:110–121. Brada M, Ford D, Ashley S, Bliss JM, Crowley S, Mason M, Rajan B, Traish D. 1992. Risk of second brain tumour after conservative surgery and radiotherapy for pituitary adenoma. BMJ 304:1343–1346. Brenner AV, Linet MS, Fine HA, Shapiro WR, Selker RG, Black PM, Inskip PD. 2002. History of allergies and autoimmune diseases and risk of brain tumors in adults. Int J Cancer 99:252–259. Bullard DE, Nashold BS, Jr., Osborne D, Burger PC, Dubois P. 1984. CTguided stereotactic biopsies using a modified frame and Gildenberg techniques. J Neurol Neurosurg Psychiatry 47:590–595. Bunin GR, Kuijten RR, Boesel CP, Buckley JD, Meadows AT. 1994. Maternal diet and risk of astrocytic glioma in children: A report from the Childrens Cancer Group (United States and Canada). Cancer Causes Control 5:177–187.

Nervous System Bunin GR, Kuijten RR, Buckley JD, Rorke LB, Meadows AT. 1993. Relation between maternal diet and subsequent primitive neuroectodermal brain tumors in young children. N Engl J Med 329:536–541. Burch JD, Craib KJ, Choi BC, Miller AB, Risch HA, Howe GR. 1987. An exploratory case-control study of brain tumors in adults. J Natl Cancer Inst 78:601–609. Burnett ME, White EC, Sih S, von Haken MS, Cogen PH. 1997. Chromosome arm 17p deletion analysis reveals molecular genetic heterogeneity in supratentorial and infratentorial primitive neuroectodermal tumors of the central nervous system. Cancer Genet Cytogenet 97:25–31. Cantor KP, Lynch CF, Hildesheim ME, Dosemeci M, Lubin J, Alavanja M, Craun G. 1999. Drinking water source and chlorination byproducts in Iowa. III. Risk of brain cancer. Am J Epidemiol 150:552–560. Carozza SE, Wrensch M, Miike R, Newman B, Olshan AF, Savitz DA, Yost M, Lee M. 2000. Occupation and adult gliomas. Am J Epidemiol 152:838–846. CBTRUS. 2002. Statistical report: Primary brain tumors in the United States, 1995–1999. Central Brain Tumor Registry of the United States. Chilvers C, Inskip H, Caygill C, Bartholomew B, Fraser P, Hill M. 1984. A survey of dietary nitrate in well-water users. Int J Epidemiol 13:324–331. Chou HJ, Yates RL, Wenninger JA. 1987. Screening cosmetic products for Nnitroso compounds by chemiluminescent determination of nitric oxide. J Assoc Off Anal Chem 70:960–963. Chretien F, Boche D, Lorin de la Grandmaison G, Ereau T, Mikol J, Hurtrel M, Hurtrel B, Gray F. 2000. Progressive multifocal leukoencephalopathy and oligodendroglioma in a monkey co-infected by simian immunodeficiency virus and simian virus 40. Acta Neuropathol (Berl) 100:332–336. Cicuttini FM, Hurley SF, Forbes A, Donnan GA, Salzberg M, Giles GG, McNeil JJ. 1997. Association of adult glioma with medical conditions, family and reproductive history. Int J Cancer 71:203–207. Cocco P, Heineman EF, Dosemeci M. 1999. Occupational risk factors for cancer of the central nervous system (CNS) among US women. Am J Ind Med 36:70–74. Collins VP. 1998. Gliomas. Cancer Surv 32:37–51. Colt JS, Blair A. 1998. Parental occupational exposures and risk of childhood cancer. Environ Health Perspect 106 Suppl 3:909–925. Corallini A, Pagnani M, Viadana P, Silini E, Mottes M, Milanesi G, Gerna G, Vettor R, Trapella G, Silvani V, et al. 1987. Association of BK virus with human brain tumors and tumors of pancreatic islets. Int J Cancer 39:60–67. Cordier S, Iglesias MJ, Le Goaster C, Guyot MM, Mandereau L, Hemon D. 1994. Incidence and risk factors for childhood brain tumors in the Ile de France. Int J Cancer 59:776–782. Cordier S, Mandereau L, Preston-Martin S, Little J, Lubin F, Mueller B, Holly E, Filippini G, Peris-Bonet R, McCredie M, Choi NW, Arsla A. 2001. Parental occupations and childhood brain tumors: Results of an international case-control study. Cancer Causes Control 12:865–874. Corwin JT, Cotanche DA. 1988. Regeneration of sensory hair cells after acoustic trauma. Science 240:1772–1774. Craddock VM. 1983. Nitrosamines and human cancer: Proof of an association? Nature 306:638. Daumas-Duport C, Scheithauer B, O’Fallon J, Kelly P. 1988. Grading of astrocytomas. A simple and reproducible method. Cancer 62:2152–2165. Davis FG, Freels S, Grutsch J, Barlas S, Brem S. 1998. Survival rates in patients with primary malignant brain tumors stratified by patient age and tumor histological type: An analysis based on Surveillance, Epidemiology, and End Results (SEER) data, 1973–1991. J Neurosurg 88:1–10. Davis FG, Preston-Martin S. 1998. Epidemiology: Incidence and survival. In: Russell and Rubinstein’s Pathology of Tumors of the Nervous System. 6th ed. Bigner DD, McLendon RE, Bruner JM, eds. London: Oxford University Press, pp. 5–45. Davis JR, Brownson RC, Garcia R, Bentz BJ, Turner A. 1993. Family pesticide use and childhood brain cancer. Arch Environ Contam Toxicol 24:87–92. Dickinson HO, Nyari TA, Parker L. 2002. Childhood solid tumours in relation to infections in the community in Cumbria during pregnancy and around the time of birth. Br J Cancer 87:746–750. Dirven CM, Tuerlings J, Molenaar WM, Go KG, Louis DN. 1995. Glioblastoma multiforme in four siblings: A cytogenetic and molecular genetic study. J Neurooncol 24:251–258. Divine BJ, Barron V. 1986. Texaco Mortality Study: II. Patterns of mortality among white males by specific job groups. Am J Ind Med 10:371–381. Divine BJ, Barron V, Kaplan SD. 1985. Texaco mortality study. I. Mortality among refinery, petrochemical, and research workers. J Occup Med 27:445–447. Divine BJ, Hartman CM, Wendt JK. 1999. Update of the Texaco mortality study 1947–93: Part I. Analysis of overall patterns of mortality among

1191

refining, research, and petrochemical workers. Occup Environ Med 56:167–173. Dobkin BH. 1985. Stroke associated with glioblastoma. Bull Clin Neurosci 50:111–118. Druckrey H, Preussmann R, Ivankovic S, Schmahl D. 1967. Organotropic carcinogenic effects of 65 various N-nitroso- compounds on BD rats. Z Krebsforsch 69:103–201. Fan TY, Goff U, Song L, Fine DH, Arsenault GP, Biemann K. 1977. NNitrosodiethanolamine in cosmetics, lotions and shampoos. Food Cosmet Toxicol 15:423–430. Farndon PA, Del Mastro RG, Evans DG, Kilpatrick MW. 1992. Location of gene for Gorlin syndrome. Lancet 339:581–582. Farwell J, Flannery JT. 1984. Cancer in relatives of children with centralnervous-system neoplasms. N Engl J Med 311:749–753. Feldkamp MM, Gutmann DH, Guha A. 1998. Neurofibromatosis type 1: Piecing the puzzle together. Can J Neurol Sci 25:181–191. Filippini G, Maisonneuve P, McCredie M, Peris-Bonet R, Modan B, PrestonMartin S, Mueller BA, Holly EA, Cordier S, Choi NW, Little J, Arslan A, Boyle P. 2002. Relation of childhood brain tumors to exposure of parents and children to tobacco smoke: The SEARCH international casecontrol study. Surveillance of Environmental Aspects Related to Cancer in Humans. Int J Cancer 100:206–213. Fine DH, Rounbehler DP, Yu WC, et al. 1984. A new thermal energy analyzer for direct high-performance liquid chromatographic and gas chromatographic analysis of N-nitrosamides. In: N-nitroso Compounds: Occurrence, Biological Effects and Relevance to Human Cancer. O’Neill I K, et al., eds. Lyon, France: IARC Scientific Publications No. 57, IARC, pp. 121–129. Fisher SG, Weber L, Carbone M. 1999. Cancer risk associated with simian virus 40 contaminated polio vaccine. Anticancer Res 19:2173–2180. Garman RH, Snellings WM, Maronpot RR. 1986. Frequency, size and location of brain tumours in F-344 rats chronically exposed to ethylene oxide. Food Chem Toxicol 24:145–153. George JD. 1996. Navy review of nasopharyngeal radium treatments of submarine candidates. Otolaryngol Head Neck Surg 115:438–441. Gold E, Gordis L, Tonascia J, Szklo M. 1979. Risk factors for brain tumors in children. Am J Epidemiol 109:309–319. Grayson JK. 1996. Radiation exposure, socioeconomic status, and brain tumor risk in the US Air Force: A nested case-control study. Am J Epidemiol 143:480–486. Grossman SA, Osman M, Hruban R, Piantadosi S. 1999. Central nervous system cancers in first-degree relatives and spouses. Cancer Invest 17:299–308. Gurney JG, Kadan-Lottick N. 2001. Brain and other central nervous system tumors: Rates, trends, and epidemiology. Curr Opin Oncol 13:160–166. Gurney JG, Kadan-Lottick NS, Packer RJ, Neglia JP, Sklar CA, Punyko JA, Stovall M, Yasui Y, Nicholson HS, Wolden S, McNeil DE, Mertens AC, Robison LL. 2003. Endocrine and cardiovascular late effects among adult survivors of childhood brain tumors: Childhood Cancer Survivor Study. Cancer 97:663–673. Gurney JG, Mueller BA, Preston-Martin S, McDaniel AM, Holly EA, Pogoda JM, Davis RL. 1997. A study of pediatric brain tumors and their association with epilepsy and anticonvulsant use. Neuroepidemiology 16:248–255. Gurney JG, Preston-Martin S, McDaniel AM, Mueller BA, Holly EA. 1996. Head injury as a risk factor for brain tumors in children: Results from a multicenter case-control study. Epidemiology 7:485–489. Hamernik RP, Turrentine G, Roberto M, Salvi R, Henderson D. 1984a. Anatomical correlates of impulse noise-induced mechanical damage in the cochlea. Hear Res 13:229–247. Hamernik RP, Turrentine G, Wright CG. 1984b. Surface morphology of the inner sulcus and related epithelial cells of the cochlea following acoustic trauma. Hear Res 16:143–160. Harada T, Irving RM, Xuereb JH, Barton DE, Hardy DG, Moffat DA, Maher ER. 1996. Molecular genetic investigation of the neurofibromatosis type 2 tumor suppressor gene in sporadic meningioma. J Neurosurg 84:847–851. Hardell L, Mild KH, Pahlson A, Hallquist A. 2001. Ionizing radiation, cellular telephones and the risk for brain tumours. Eur J Cancer Prev 10:523–529. Hardell L, Nasman A, Pahlson A, Hallquist A, Hansson Mild K. 1999. Use of cellular telephones and the risk for brain tumours: A case-control study. Int J Oncol 15:113–116. Harvald B, Hauge M. 1956. On the heredity of glioblastoma. J Natl Cancer Inst 17:289–296. Haymaker W, Rubinstein LJ, Miquel J. 1972. Brain tumors in irradiated monkeys. Acta Neuropathol (Berl) 20:267–277. Hershatter BW, Halperin EC, Cox EB. 1986. Medulloblastoma: The Duke University Medical Center experience. Int J Radiat Oncol Biol Phys 12:1771–1777.

1192

PART IV: CANCER BY TISSUE OF ORIGIN

Hodges LC, Smith JL, Garrett A, Tate S. 1992. Prevalence of glioblastoma multiforme in subjects with prior therapeutic radiation. J Neurosci Nurs 24:79–83. Holly EA, Bracci PM, Hong MK, Mueller BA, Preston-Martin S. 2002. West Coast study of childhood brain tumours and maternal use of haircolouring products. Paediatr Perinat Epidemiol 16:226–235. Hotchkiss JH. 1989. Preformed N-nitroso compounds in foods and beverages. Cancer Surv 8:295–321. Howe GR, Burch JD, Chiarelli AM, Risch HA, Choi BC. 1989. An exploratory case-control study of brain tumors in children. Cancer Res 49:4349–4352. Hsu DW, Efird JT, Hedley-Whyte ET. 1997. Progesterone and estrogen receptors in meningiomas: Prognostic considerations. J Neurosurg 86: 113–120. Hu J, La Vecchia C, Negri E, Chatenoud L, Bosetti C, Jia X, Liu R, Huang G, Bi D, Wang C. 1999. Diet and brain cancer in adults: A case-control study in northeast China. Int J Cancer 81:20–23. Hu J, Mao Y, Ugnat AM. 2000. Parental cigarette smoking, hard liquor consumption and the risk of childhood brain tumors—a case-control study in northeast China. Acta Oncol 39:979–984. Hwang SL, Lieu AS, Kuo TH, Lin CL, Chang CZ, Huang TY, Howng SL. 2001. Preoperative and postoperative seizures in patients with astrocytic tumours: Analysis of incidence and influencing factors. J Clin Neurosci 8:426–429. ICD-O. 2000. International Classification of Diseases for Oncology. 3rd ed. Fritz A, Percy C, Jack A, Shanmugaratnam K, Sobin L, Parkin DM, Whelan S, eds. Geneva: World Health Organization. Ichimura K, Bolin MB, Goike HM, Schmidt EE, Moshref A, Collins VP. 2000. Deregulation of the p14ARF/MDM2/p53 pathway is a prerequisite for human astrocytic gliomas with G1-S transition control gene abnormalities. Cancer Res 60:417–424. Inskip PD, Linet MS, Heineman EF. 1995. Etiology of brain tumors in adults. Epidemiol Rev 17:382–414. Inskip PD, Mellemkjaer L, Gridley G, Olsen JH. 1998. Incidence of intracranial tumors following hospitalization for head injuries (Denmark). Cancer Causes Control 9:109–116. Inskip PD, Tarone RE, Hatch EE, Wilcosky TC, Shapiro WR, Selker RG, Fine HA, Black PM, Loeffler JS, Linet MS. 2001. Cellular-telephone use and brain tumors. N Engl J Med 344:79–86. Ivankovic S. 1979. Teratogenic and carcinogenic effects of some chemicals during perinatal life in rats, Syrian golden hamsters, and minipigs. Natl Cancer Inst Monogr 51:103–115. James CD, Carlbom E, Nordenskjold M, Collins VP, Cavenee WK. 1989. Mitotic recombination of chromosome 17 in astrocytomas. Proc Natl Acad Sci U S A 86:2858–2862. Johansen C, Boice J, Jr., McLaughlin J, Olsen J. 2001. Cellular telephones and cancer—a nationwide cohort study in Denmark. J Natl Cancer Inst 93:203–207. Kakuda Y, Gray JI. 1980. N-Nitrosamides and their precursors in food systems. 2. Kinetics of the N-nitrosation reaction. J Agric Food Chem 28:584–587. Kaplan S, Novikov I, Modan B. 1997. Nutritional factors in the etiology of brain tumors: Potential role of nitrosamines, fat, and cholesterol. Am J Epidemiol 146:832–841. Karlsson P, Holmberg E, Lundell M, Mattsson A, Holm LE, Wallgren A. 1998. Intracranial tumors after exposure to ionizing radiation during infancy: A pooled analysis of two Swedish cohorts of 28,008 infants with skin hemangioma. Radiat Res 150:357–364. Kaye AH, Laws ER. 2001. Brain Tumors: An Encyclopedic Approach. 2nd ed. London: Churchill Livingstone, p. 29. Kernie SG, Erwin TM, Parada LF. 2001. Brain remodeling due to neuronal and astrocytic proliferation after controlled cortical injury in mice. J Neurosci Res 66:317–326. Kerr MA, Nasca PC, Mundt KA, Michalek AM, Baptiste MS, Mahoney MC. 2000. Parental occupational exposures and risk of neuroblastoma: A casecontrol study (United States). Cancer Causes Control 11:635–643. Khalili K. 2001. Human neurotropic JC virus and its association with brain tumors. Dis Markers 17:143–147. Kheifets LI. 2001. Electric and magnetic field exposure and brain cancer: A review. Bioelectromagnetics Suppl 5:S120–131. Kheifets LI, Afifi AA, Buffler PA, Zhang ZW. 1995. Occupational electric and magnetic field exposure and brain cancer: A meta-analysis. J Occup Environ Med 37:1327–1341. Kheifets LI, Sussman SS, Preston-Martin S. 1999. Childhood brain tumors and residential electromagnetic fields (EMF). Rev Environ Contam Toxicol 159:111–129. Kleihues P. 1995. New concepts in the classification of brain tumors. J Neuropathol Exp Neurol 54 suppl:53–54. Kleihues P, Burger PC, Scheithauer BW. 1993. The new WHO classification of brain tumours. Brain Pathol 3:255–268.

Kleihues P, Cavenee WK. 2000. World Health Organization Classification of Tumors. Pathology and Genetics of Tumors of the Nervous System. Lyon, France: International Agency for Research on Cancer (IARC) Press. Kluwe L, Mautner VF. 1998. Mosaicism in sporadic neurofibromatosis 2 patients. Hum Mol Genet 7:2051–2055. Koestner A. 1986. The brain-tumour issue in long-term toxicity studies in rats. Food Chem Toxicol 24:139–143. Kraus JA, Koopmann J, Kaskel P, Maintz D, Brandner S, Schramm J, Louis DN, Wiestler OD, von Deimling A. 1995. Shared allelic losses on chromosomes 1p and 19q suggest a common origin of oligodendroglioma and oligoastrocytoma. J Neuropathol Exp Neurol 54:91–95. Krewski D, Byus CV, Glickman BW, Lotz WG, Mandeville R, McBride ML, Prato FS, Weaver DF. 2001. Recent advances in research on radiofrequency fields and health. J Toxicol Environ Health B Crit Rev 4:145–159. Kristensen P, Andersen A, Irgens LM, Bye AS, Sundheim L. 1996. Cancer in offspring of parents engaged in agricultural activities in Norway: Incidence and risk factors in the farm environment. Int J Cancer 65: 39–50. Kubo M, Fukutomi T, Akashi-Tanaka S, Hasegawa T. 2001. Association of breast cancer with meningioma: Report of a case and review of the literature. Jpn J Clin Oncol 31:510–513. Kuijten RR, Bunin GR, Nass CC, Meadows AT. 1990. Gestational and familial risk factors for childhood astrocytoma: Results of a case-control study. Cancer Res 50:2608–2612. Kunwar S, Mohapatra G, Bollen A, Lamborn KR, Prados M, Feuerstein BG. 2001. Genetic subgroups of anaplastic astrocytomas correlate with patient age and survival. Cancer Res 61:7683–7688. Lee M, Wrensch M, Miike R. 1997. Dietary and tobacco risk factors for adult onset glioma in the San Francisco Bay Area (California, USA). Cancer Causes Control 8:13–24. Legler JM, Ries LA, Smith MA, Warren JL, Heineman EF, Kaplan RS, Linet MS. 1999. Cancer surveillance series [corrected]: Brain and other central nervous system cancers: Recent trends in incidence and mortality. J Natl Cancer Inst 91:1382–1390. Leonard JR, D’Sa C, Klocke BJ, Roth KA. 2001. Neural precursor cell apoptosis and glial tumorigenesis following transplacental ethyl-nitrosourea exposure. Oncogene 20:8281–8286. Liau LM, Jensen ER, Kremen TJ, Odesa SK, Sykes SN, Soung MC, Miller JF, Bronstein JM. 2002. Tumor immunity within the central nervous system stimulated by recombinant Listeria monocytogenes vaccination. Cancer Res 62:2287–2293. Lieu AS, Howng SL. 2000. Intracranial meningiomas and epilepsy: Incidence, prognosis and influencing factors. Epilepsy Res 38:45–52. Lijinsky W. 1992. Chemistry and biology of N-nitroso compounds. Cambridge: Cambridge University Press. Lijinsky W, Kovatch RM. 1985. Induction of liver tumors in rats by nitrosodiethanolamine at low doses. Carcinogenesis 6:1679–1681. Lindsay KW, Bone I, Callander R. 1991. Neurology and Neurosurgery Illustrated. 2nd ed. London: Harcourt Publishers, Ltd. Linos A, Kardara M, Kosmidis H, Katriou D, Hatzis C, Kontzoglou M, Koumandakis E, Tzartzatou-Stathopoulou F. 1998. Reported influenza in pregnancy and childhood tumour. Eur J Epidemiol 14:471–475. Little MP, de Vathaire F, Shamsaldin A, Oberlin O, Campbell S, Grimaud E, Chavaudra J, Haylock RG, Muirhead CR. 1998. Risks of brain tumour following treatment for cancer in childhood: Modification by genetic factors, radiotherapy and chemotherapy. Int J Cancer 78:269–275. Longstreth WT, Jr., Dennis LK, McGuire VM, Drangsholt MT, Koepsell TD. 1993. Epidemiology of intracranial meningioma. Cancer 72:639– 648. Lonn S, Klaeboe L, Hall P, Mathiesen T, Auvinen A, Christensen HC, Johansen C, Salminen T, Tynes T, Feychting M. 2004. Incidence trends of adult primary intracerebral tumors in four Nordic countries. Int J Cancer 108:450–455. Lonser RR, Walbridge S, Vortmeyer AO, Pack SD, Nguyen TT, Gogate N, Olson JJ, Akbasak A, Bobo RH, Goffman T, Zhuang Z, Oldfield EH. 2002. Induction of glioblastoma multiforme in nonhuman primates after therapeutic doses of fractionated whole-brain radiation therapy. J Neurosurg 97:1378–1389. Loomis DP, Savitz DA. 1990. Mortality from brain cancer and leukaemia among electrical workers. Br J Ind Med 47:633–638. Louis DN, Holland EC, Cairncross JG. 2001. Glioma classification: a molecular reappraisal. Am J Pathol 159:779–786. Lubin F, Farbstein H, Chetrit A, Farbstein M, Freedman L, Alfandary E, Modan B. 2000. The role of nutritional habits during gestation and child life in pediatric brain tumor etiology. Int J Cancer 86:139–143. Magee PN, Montesano R, Preusmann R. 1976. N-nitroso compounds and related carcinogens. In: Searle, ED, ed. Chemical Carcinogens. Washington, D.C.: American Chemical Society.

Nervous System Maher ER, Webster AR, Richards FM, Green JS, Crossey PA, Payne SJ, Moore AT. 1997. Phenotypic expression in von Hippel-Lindau disease: Correlations with germline VHL gene mutations. J Med Genet 33:328–332. Malkin D, Li FP, Strong LC, Fraumeni JF, Jr., Nelson CE, Kim DH, Kassel J, Gryka MA, Bischoff FZ, Tainsky MA, et al. 1990. Germ line p53 mutations in a familial syndrome of breast cancer, sarcomas, and other neoplasms. Science 250:1233–1238. Malmer B, Gronberg H, Bergenheim AT, Lenner P, Henriksson R. 1999. Familial aggregation of astrocytoma in northern Sweden: An epidemiological cohort study. Int J Cancer 81:366–370. Malmer B, Iselius L, Holmberg E, Collins A, Henriksson R, Gronberg H. 2001. Genetic epidemiology of glioma. Br J Cancer 84:429–434. Malmer B, Tavelin B, Henriksson R, Gronberg H. 2000. Primary brain tumours as second primary: A novel association between meningioma and colorectal cancer. Int J Cancer 85:78–81. Maltoni C, Ciliberti A, Carretti D. 1982. Experimental contributions in identifying brain potential carcinogens in the petrochemical industry. Ann N Y Acad Sci 381:216–249. McCann J. 2003. Animal models offer insights into human brain tumors. J Natl Cancer Inst 95:188. McCredie M, Maisonneuve P, Boyle P. 1994a. Perinatal and early postnatal risk factors for malignant brain tumours in New South Wales children. Int J Cancer 56:11–15. McCredie M, Maisonneuve P, Boyle P. 1994b. Antenatal risk factors for malignant brain tumours in New South Wales children. Int J Cancer 56:6–10. Michotte A. 1996. The revised WHO classification of brain tumors. Acta Neurol Belg 96:85–88. Midgard R, Glattre E, Gronning M, Riise T, Edland A, Nyland H. 1996. Multiple sclerosis and cancer in Norway. A retrospective cohort study. Acta Neurol Scand 93:411–415. Mills PK, Preston-Martin S, Annegers JF, Beeson WL, Phillips RL, Fraser GE. 1989. Risk factors for tumors of the brain and cranial meninges in Seventh-Day Adventists. Neuroepidemiology 8:266–275. Mirvish SS. 1981. Inhibition of the formation of carcinogenic N-nitroso compounds by ascorbic acid and other compounds. In: Burchenal JH, et al., eds. Cancer: Achievements, Challenges and Prospects for the 1980’s. New York: Grune & Stratton, Vol. 1, pp. 557–587. Mirvish SS. 1994. Experimental evidence for inhibition of N-nitroso compound formation as a factor in the negative correlation between vitamin C consumption and the incidence of certain cancers. Cancer Res 54: 1948s–1951s. Mirvish SS. 1995. Role of N-nitroso compounds (NOC) and N-nitrosation in etiology of gastric, esophageal, nasopharyngeal and bladder cancer and contribution to cancer of known exposures to NOC. Cancer Lett 93: 17–48. Morgan RW, Kelsh MA, Zhao K, Exuzides KA, Heringer S, Negrete W. 2000. Radiofrequency exposure and mortality from cancer of the brain and lymphatic/hematopoietic systems. Epidemiology 11:118–127. Morgenstern LB, Frankowski RF. 1999. Brain tumor masquerading as stroke. J Neurooncol 44:47–52. Mueller BA, Newton K, Holly EA, Preston-Martin S. 2001. Residential water source and the risk of childhood brain tumors. Environ Health Perspect 109:551–556. Mueller BA, Nielsen SS, Preston-Martin S, Holly EA, Cordier S, Filippini G, Peris-Bonet R, Choi N. 2004. Household water source and the risk of childhood brain tumours: Results of the SEARCH International Brain Tumor Study. Int J Epidemiol 33:1209–1216. Muscat JE, Malkin MG, Thompson S, Shore RE, Stellman SD, McRee D, Neugut AI, Wynder EL. 2000. Handheld cellular telephone use and risk of brain cancer. JAMA 284:3001–3007. Nagane M, Huang HJ, Cavenee WK. 1997. Advances in the molecular genetics of gliomas. Curr Opin Oncol 9:215–222. National Research Council Committee on Diet Nutrition and Cancer. 1982. Diet Nutrition and Cancer. Washington, D.C.: National Academy Press. Ng HK, Lam PY. 1998. The molecular genetics of central nervous system tumors. Pathology 30:196–202. Norman MA, Holly EA, Preston-Martin S. 1996. Childhood brain tumors and exposure to tobacco smoke. Cancer Epidemiol Biomarkers Prev 5:85–91. Nygren C, Adami J, Ye W, Bellocco R, af Geijerstam JL, Borg J, Nyren O. 2001. Primary brain tumors following traumatic brain injury—a population-based cohort study in Sweden. Cancer Causes Control 12:733–737. O’Neill BP, Blondal H, Yang P, Olafsdottir GH, Sigvaldason H, Jenkins RB, Kimmel DW, Scheithauer BW, Rocca WA, Bjornsson J, Tulinius H. 2002. Risk of cancer among relatives of patients with glioma. Cancer Epidemiol Biomarkers Prev 11:921–924.

1193

Ohsumi S, Ikehara I, Motoi M, Ogawa K, Nagashima K, Yasui K. 1985. Induction of undifferentiated brain tumors in rats by a human polyomavirus (JC virus). Jpn J Cancer Res 76:429–431. Olson JJ, Barnett D, Yang J, Assietti R, Cotsonis G, James CD. 1998. Gene amplification as a prognostic factor in primary brain tumors. Clin Cancer Res 4:215–222. Osborne RH, Houben MP, Tijssen CC, Coebergh JW, van Duijn CM. 2001. The genetic epidemiology of glioma. Neurology 57:1751–1755. Paraf F, Jothy S, Van Meir EG. 1997. Brain tumor-polyposis syndrome: two genetic diseases? J Clin Oncol 15:2744–2758. Parkin DM, et al. 2002. Cancer Incidence in Five Continents, Volume VIII. Lyon, France: IARC Publications No. 155. Parkin DM, Whelan SL, Ferlay J, Raymond L, Young J. 1997. Cancer Incidence in Five Continents, Vol. VII. Lyon, France: IARC Scientific Publication No. 143. Pogoda JM, Preston-Martin S. 1997. Household pesticides and risk of pediatric brain tumors. Environ Health Perspect 105:1214–1220. Pogoda JM, Preston-Martin S. 2001. Maternal cured meat consumption during pregnancy and risk of paediatric brain tumour in offspring: Potentially harmful levels of intake. Public Health Nutr 4:183–189. Poussaint TY. 2001. Magnetic resonance imaging of pediatric brain tumors: State of the art. Top Magn Reson Imaging 12:411–433. Preston-Martin S. 1996. Epidemiology of primary CNS neoplasms. Neurol Clin 14:273–290. Preston-Martin S, Mack W. 1991. Gliomas and meningiomas in men in Los Angeles County: Investigation of exposures to N-nitroso compounds. In: O’Neill IK, et al., eds. Relevance to Human Cancer of N-nitroso Compounds, Tobacco Smoke and Mycotoxins. Lyon, France: IARC Scientific Publications No. 105, IARC, pp. 197–203. Preston-Martin S, Mack W. 1996. Neoplasms of the nervous system. In: Schottenfeld D, Fraumeni JF, eds. Cancer Epidemiology and Prevention. 2nd ed. New York: Oxford University Press, pp. 1231–1281. Preston-Martin S, Monroe K, Lee PJ, Bernstein L, Kelsey J, Henderson S, Forrester D, Henderson B. 1995. Spinal meningiomas in women in Los Angeles County: Investigation of an etiological hypothesis. Cancer Epidemiol Biomarkers Prev 4:333–339. Preston-Martin S, Pogoda JM, Mueller BA, Lubin F, Holly EA, Filippini G, Cordier S, Peris-Bonet R, Choi W, Little J, Arslan A. 1998a. Prenatal vitamin supplementation and risk of childhood brain tumors. Int J Cancer Suppl 11:17–22. Preston-Martin S, Pogoda JM, Schlehofer B, Blettner M, Howe GR, Ryan P, Menegoz F, Giles GG, Rodvall Y, Choi NW, Little J, Arslan A. 1998b. An international case-control study of adult glioma and meningioma: The role of head trauma. Int J Epidemiol 27:579–586. Preston-Martin S, Staples M, Farrugia H, Giles G. 1993. Primary tumors of the brain, cranial nerves and cranial meninges in Victoria, Australia, 1982–1990: Patterns of incidence and survival. Neuroepidemiology 12:270–279. Preston-Martin S, Thomas DC, Wright WE, Henderson BE. 1989. Noise trauma in the aetiology of acoustic neuromas in men in Los Angeles County, 1978–1985. Br J Cancer 59:783–786. Preston-Martin S, White SC. 1990. Brain and salivary gland tumors related to prior dental radiography: Implications for current practice. J Am Dent Assoc 120:151–158. Preston-Martin S, Yu MC, Benton B, Henderson BE. 1982. N-Nitroso compounds and childhood brain tumors: A case-control study. Cancer Res 42:5240–5245. Preston DL, Ron E, Yonehara S, Kobuke T, Fujii H, Kishikawa M, Tokunaga M, Tokuoka S, Mabuchi K. 2002. Tumors of the nervous system and pituitary gland associated with atomic bomb radiation exposure. J Natl Cancer Inst 94:1555–1563. Pringle NP, Guthrie S, Lumsden A, Richardson WD. 1998. Dorsal spinal cord neuroepithelium generates astrocytes but not oligodendrocytes. Neuron 20:883–893. Raaschou-Nielsen O, Hertel O, Thomsen BL, Olsen JH. 2001. Air pollution from traffic at the residence of children with cancer. Am J Epidemiol 153:433–443. Raffel C, Frederick L, O’Fallon JR, Atherton-Skaff P, Perry A, Jenkins RB, James CD. 1999. Analysis of oncogene and tumor suppressor gene alterations in pediatric malignant astrocytomas reveals reduced survival for patients with PTEN mutations. Clin Cancer Res 5:4085– 4090. Rajewsky MF, Goth R, Laerum OD, et al. 1976. Molecular and cellular mechanisms in nervous system-specific carcinogenesis by N-ethyl-Nnitrosourea. In: Magee PN, et al., eds. Fundamentals in Cancer Prevention. Baltimore: University Park Press, pp. 313–334. Ransom DT, Ritland SR, Kimmel DW, Moertel CA, Dahl RJ, Scheithauer BW, Kelly PJ, Jenkins RB. 1992. Cytogenetic and loss of heterozygosity

1194

PART IV: CANCER BY TISSUE OF ORIGIN

studies in ependymomas, pilocytic astrocytomas, and oligodendrogliomas. Genes Chromosomes Cancer 5:348–356. Reagan TJ, Freiman IS. 1973. Multiple cerebral gliomas in multiple sclerosis. J Neurol Neurosurg Psychiatry 36:523–528. Rice JM. 1986. Transplacental carcinogenesis in nonhuman primates: Inferences for human risk assessment. Presented at workshop of transplacental carcinogenic risk to humans. Italy: Santa Marghenita Ligure. Rice JM, Ward JM. 1982. Age dependence of susceptibility to carcinogenesis in the nervous system. Ann N Y Acad Sci 381:274–289. Rice JM, Wilbourn JD. 2000. Tumors of the nervous system in carcinogenic hazard identification. Toxicol Pathol 28:202–214. Ries LA, Eisner MP, Kosary CL. 2000. SEER cancer statistics review 1973–1977. Bethesda: National Cancer Institute. Available at: http://www-seer.ims.nci.nih.gov/Publications/. Rimm IJ, Li FC, Tarbell NJ, Winston KR, Sallan SE. 1987. Brain tumors after cranial irradiation for childhood acute lymphoblastic leukemia. A 13-year experience from the Dana-Farber Cancer Institute and the Children’s Hospital. Cancer 59:1506–1508. Ringertz N. 1950. Grading of gliomas. Acta Pathol Microbiol Scand 27:51–65. Ron E, Modan B, Boice JD, Jr., Alfandary E, Stovall M, Chetrit A, Katz L. 1988. Tumors of the brain and nervous system after radiotherapy in childhood. N Engl J Med 319:1033–1039. Ronckers CM, Land CE, Verduijn PG, Hayes RB, Stovall M, van Leeuwen FE. 2001. Cancer mortality after nasopharyngeal radium irradiation in the netherlands: A cohort study. J Natl Cancer Inst 93:1021–1027. Rubinstein LJ. 1972. Tumors of the central nervous system. In: Atlas of Tumor Pathology. Second Series, Fascicle 6. Washington, D.C.: Armed Forces Institute of Pathology. Ruttledge MH, Sarrazin J, Rangaratnam S, Phelan CM, Twist E, Merel P, Delattre O, Thomas G, Nordenskjold M, Collins VP, et al. 1994. Evidence for the complete inactivation of the NF2 gene in the majority of sporadic meningiomas. Nat Genet 6:180–184. Ryals BM, Rubel EW. 1988. Hair cell regeneration after acoustic trauma in adult Coturnix quail. Science 240:1774–1776. Ryan P, Hurley SF, Johnson AM, Salzberg M, Lee MW, North JB, McNeil JJ, McMichael AJ. 1993. Tumours of the brain and presence of antibodies to Toxoplasma gondii. Int J Epidemiol 22:412–419. Sabel M, Felsberg J, Messing-Junger M, Neuen-Jacob E, Piek J. 1999. Glioblastoma multiforme at the site of metal splinter injury: A coincidence? Case report. J Neurosurg 91:1041–1044. Sampson JH, Archer GE, Ashley DM, Fuchs HE, Hale LP, Dranoff G, Bigner D D. 1996. Subcutaneous vaccination with irradiated, cytokineproducing tumor cells stimulates CD8+ cell-mediated immunity against tumors located in the “immunologically privileged” central nervous system. Proc Natl Acad Sci U S A 93:10399–10404. Sampson JH, Bigner DD. 1998. Experimental tumors and the evaluation of neurocarcinogens. In: Russell and Rubinstein’s Pathology of Tumors of the Nervous System. 6th ed. New York: Oxford University Press, pp. 167–230. Sandler DP. 1996. Nasopharyngeal radium irradiation: The Washington County, Maryland, study. Otolaryngol Head Neck Surg 115:409–414. Sarasua S, Savitz DA. 1994. Cured and broiled meat consumption in relation to childhood cancer: Denver, Colorado (United States). Cancer Causes Control 5:141–148. Savitz DA, Cai J, van Wijngaarden E, Loomis D, Mihlan G, Dufort V, Kleckner RC, Nylander-French L, Kromhout H, Zhou H. 2000. Casecohort analysis of brain cancer and leukemia in electric utility workers using a refined magnetic field job-exposure matrix. Am J Ind Med 38: 417–425. Schlehofer B, Blettner M, Becker N, Martinsohn C, Wahrendorf J. 1992. Medical risk factors and the development of brain tumors. Cancer 69:2541–2547. Schlehofer B, Blettner M, Preston-Martin S, Niehoff D, Wahrendorf J, Arslan A, Ahlbom A, Choi W N, Giles G G, Howe G R, Little J, Menegoz F, Ryan P. 1999. Role of medical history in brain tumour development. Results from the international adult brain tumour study. Int J Cancer 82:155–160. Schneider AB, Shore-Freedman E, Ryo UY, Bekerman C, Favus M, Pinsky S. 1985. Radiation-induced tumors of the head and neck following childhood irradiation. Prospective studies. Medicine (Baltimore) 64: 1–15. Schuman LM, Choi NW, Gullen WH. 1967. Relationship of central nervous system neoplasms to Toxoplasma gondii infection. Am J Public Health Nations Health 57:848–856. Schwartzbaum J, Jonsson F, Ahlbom A, Preston-Martin S, Lonn S, Soderberg K C, Feychting M. 2003. Cohort studies of association between selfreported allergic conditions, immune-related diagnoses and glioma and meningioma risk. Int J Cancer 106:423–428.

Schwartzbaum JA, Cornwell DG. 2000. Oxidant stress and glioblastoma multiforme risk: Serum antioxidants, gamma-glutamyl transpeptidase, and ferritin. Nutr Cancer 38:40–49. Schymura MJ, Zheng D, Baptiste MS, Nasca PC. 1996. A case-control study of childhood brain tumors and maternal lifestyle. Am J Epidemiol 143:S8. Sen N P, Seaman S, Clarkson S, et al. 1984. Volatile N-nitrosamines in baby bottle rubber nipples and pacifiers: Analysis, occurrence and migration. In: O’Neill IK, et al., eds. N-nitroso Compounds: Occurrence, Biological Effects and Relevance to Human Cancer. Lyon, France: IARC Scientific Publications, No. 57. Shapiro S, Mealey J, Jr., Sartorius C. 1989. Radiation-induced intracranial malignant gliomas. J Neurosurg 71:77–82. Sharp L, McKinney PA, Black RJ. 1999. Incidence of childhood brain and other non-haematopoietic neoplasms near nuclear sites in Scotland, 1975–94. Occup Environ Med 56:308–314. Shintani T, Hayakawa N, Hoshi M, Sumida M, Kurisu K, Oki S, Kodama Y, Kajikawa H, Inai K, Kamada N. 1999. High incidence of meningioma among Hiroshima atomic bomb survivors. J Radiat Res (Tokyo) 40:49–57. Skinner J, Mee TJ, Blackwell RP, Maslanyj MP, Simpson J, Allen SG, Day NE, Cheng KK, Gilman E, Williams D, Cartwright R, Craft A, Birch JM, Eden OB, McKinney PA, Deacon J, Peto J, Beral V, Roman E, Elwood P, Alexander FE, Mott M, Chilvers CE, Muir K, Doll R, Taylor CM, Greaves M, Goodhead D, Fry FA, Adams G, Law G. 2002. Exposure to power frequency electric fields and the risk of childhood cancer in the UK. Br J Cancer 87:1257–1266. Smith JS, Tachibana I, Passe SM, Huntley BK, Borell TJ, Iturria N, O’Fallon JR, Schaefer PL, Scheithauer BW, James CD, Buckner JC, Jenkins RB. 2001. PTEN mutation, EGFR amplification, and outcome in patients with anaplastic astrocytoma and glioblastoma multiforme. J Natl Cancer Inst 93:1246–1256. Smith MA, Freidlin B, Ries LA, Simon R. 1998. Trends in reported incidence of primary malignant brain tumors in children in the United States. J Natl Cancer Inst 90:1269–1277. Soffer D, Gomori JM, Siegal T, Shalit MN. 1989. Intracranial meningiomas after high-dose irradiation. Cancer 63:1514–1519. Sorahan T, Nichols L, van Tongeren M, Harrington JM. 2001. Occupational exposure to magnetic fields relative to mortality from brain tumours: Updated and revised findings from a study of United Kingdom electricity generation and transmission workers, 1973–97. Occup Environ Med 58:626–630. Speers MA, Dobbins JG, Miller VS. 1988. Occupational exposures and brain cancer mortality: A preliminary study of east Texas residents. Am J Ind Med 13:629–638. Steck PA, Pershouse MA, Jasser SA, Yung WK, Lin H, Ligon AH, Langford LA, Baumgard ML, Hattier T, Davis T, Frye C, Hu R, Swedlund B, Teng DH, Tavtigian SV. 1997. Identification of a candidate tumour suppressor gene, MMAC1, at chromosome 10q23.3 that is mutated in multiple advanced cancers. Nat Genet 15:356–362. Strickler HD, Rosenberg PS, Devesa SS, Hertel J, Fraumeni JF, Jr., Goedert JJ. 1998. Contamination of poliovirus vaccines with simian virus 40 (1955–1963) and subsequent cancer rates. JAMA 279:292–295. Swenberg JA. 1986. Brain tumours—problems and perspectives. Food Chem Toxicol 24:155–158. Tandon PN, Mahapatra AK, Khosla A. 2001. Epileptic seizures in supratentorial gliomas. Neurol India 49:55–59. Tannenbaum SR, Mergens W. 1980. Reaction of nitrite with vitamins C and E. Ann N Y Acad Sci 355:267–279. Tedeschi-Blok N, Schwartzbaum J, Lee M, Miike R, Wrensch M. 2001. Dietary calcium consumption and astrocytic glioma: The San Francisco Bay Area Adult Glioma Study, 1991–1995. Nutr Cancer 39:196–203. Tijssen CC. 1985. Genetic aspects of brain tumours—tumours of neuroepithelial and meningeal tissue. In: Familial Cancer. Muller H, Weber W, eds. Farmington CT: S. Karger AG. Todd NV, McDonagh T, Miller JD. 1987. What follows diagnosis by computed tomography of solitary brain tumour? Audit of one year’s experience in South East Scotland. Lancet 1:611–612. Traynor JE, Casey HW. 1971. Five-year follow-up of primates exposed to 55 MeV protons. Radiat Res 47:143–148. Tricker AR, Kubacki SJ. 1992. Review of the occurrence and formation of non-volatile N-nitroso compounds in foods. Food Addit Contam 9: 39–69. Tricker AR, Preussmann R. 1991. Carcinogenic N-nitrosamines in the diet: occurrence, formation, mechanisms and carcinogenic potential. Mutat Res 259:277–289. Tsang RW, Laperriere NJ, Simpson WJ, Brierley J, Panzarella T, Smyth HS. 1993. Glioma arising after radiation therapy for pituitary adenoma. A report of four patients and estimation of risk. Cancer 72:2227–2233.

Nervous System Tse JY, Wong JH, Lo KW, Poon WS, Huang DP, Ng HK. 1997. Molecular genetic analysis of the von Hippel-Lindau disease tumor suppressor gene in familial and sporadic cerebellar hemangioblastomas. Am J Clin Pathol 107:459–466. United States Department of Agriculture. 1978. Nitrates, nitrites, and ascorbates (or isoascorbates) in bacon. US Dept Agric 43:20992. United States Department of Agriculture. 1986. Nitrite levels in bacon. US Dept Agric 51:21731. Villeneuve PJ, Agnew DA, Johnson KC, Mao Y. 2002. Brain cancer and occupational exposure to magnetic fields among men: Results from a Canadian population-based case-control study. Int J Epidemiol 31: 210–217. von Deimling A, Louis DN, Schramm J, Wiestler OD. 1994. Astrocytic gliomas: Characterization on a molecular genetic basis. Recent Results Cancer Res 135:33–42. Walker MD. 1975. Malignant brain tumors–a synopsis. CA Cancer J Clin 25:114–120. Walsh J, Gye R, Connelley TJ. 1969. Meningioma: A late complication of head injury. Med J Aust 1:906–908. Walter AW, Hancock ML, Pui CH, Hudson MM, Ochs JS, Rivera GK, Pratt CB, Boyett JM, Kun LE. 1998. Secondary brain tumors in children treated for acute lymphoblastic leukemia at St Jude Children’s Research Hospital. J Clin Oncol 16:3761–3767. Wang SI, Puc J, Li J, Bruce JN, Cairns P, Sidransky D, Parsons R. 1997. Somatic mutations of PTEN in glioblastoma multiforme. Cancer Res 57:4183–4186. Wechsler-Reya R, Scott MP. 2001. The developmental biology of brain tumors. Annu Rev Neurosci 24:385–428. Weggen S, Bayer TA, von Deimling A, Reifenberger G, von Schweinitz D, Wiestler OD, Pietsch T. 2000. Low frequency of SV40, JC and BK polyomavirus sequences in human medulloblastomas, meningiomas and ependymomas. Brain Pathol 10:85–92. Weiss WA, Israel M, Cobbs C, Holland E, James CD, Louis DN, Marks C, McClatchey AI, Roberts T, Van Dyke T, Wetmore C, Chiu IM, Giovannini M, Guha A, Higgins RJ, Marino S, Radovanovic I, Reilly K, Aldape K. 2002. Neuropathology of genetically engineered mice: Consensus report and recommendations from an international forum. Oncogene 21:7453–7463. Weremowicz S, Kupsky WJ, Morton CC, Fletcher JA. 1992. Cytogenetic evidence for a chromosome 22 tumor suppressor gene in ependymoma. Cancer Genet Cytogenet 61:193–196.

1195

Wetmore C, Eberhart DE, Curran T. 2001. Loss of p53 but not ARF accelerates medulloblastoma in mice heterozygous for patched. Cancer Res 61:513–516. Wiemels JL, Wiencke JK, Sison JD, Miike R, McMillan A, Wrensch M. 2002. History of allergies among adults with glioma and controls. Int J Cancer 98:609–615. Wrensch M, Lee M, Miike R, Newman B, Barger G, Davis R, Wiencke J, Neuhaus J. 1997. Familial and personal medical history of cancer and nervous system conditions among adults with glioma and controls. Am J Epidemiol 145:581–593. Wrensch M, Minn Y, Chew T, Bondy M, Berger MS. 2002. Epidemiology of primary brain tumors: Current concepts and review of the literature. Neuro-oncol 4:278–299. Wrensch M, Weinberg A, Wiencke J, Miike R, Barger G, Kelsey K. 2001. Prevalence of antibodies to four herpesviruses among adults with glioma and controls. Am J Epidemiol 154:161–165. Wrensch M, Yost M, Miike R, Lee G, Touchstone J. 1999. Adult glioma in relation to residential power frequency electromagnetic field exposures in the San Francisco Bay area. Epidemiology 10:523–527. Xiao A, Wu H, Pandolfi PP, Louis DN, Van Dyke T. 2002. Astrocyte inactivation of the pRb pathway predisposes mice to malignant astrocytoma development that is accelerated by PTEN mutation. Cancer Cell 1:157–168. Yamamoto M, Yamada T, Yoshihira K, Tanimura A, Tomita I. 1988. Effects of food components and additives on the formation of nitrosamides. Food Addit Contam 5:289–298. Yeh H, Matanoski GM, Wang N, Sandler DP, Comstock GW. 2001. Cancer incidence after childhood nasopharyngeal radium irradiation: A follow-up study in Washington County, Maryland. Am J Epidemiol 153: 749–756. Zahm SH. 1999. Childhood leukemia and pesticides. Epidemiology 10: 473–475. Zheng T, Cantor KP, Zhang Y, Chiu BC, Lynch CF. 2001. Risk of brain glioma not associated with cigarette smoking or use of other tobacco products in Iowa. Cancer Epidemiol Biomarkers Prev 10:413–414. Zulch KJ. 1979. Histological typing of tumors of the central nervous system. Geneva: World Health Organization.

63

Cutaneous and Ocular Melanoma STEPHEN B. GRUBER AND BRUCE K. ARMSTRONG

M

elanoma” usually refers to an in situ or invasive cancer of melanocytes, which may also be called “malignant melanoma”. Although the term “benign melanoma” is occasionally used clinically to refer to very slowly growing, pigmented intraocular tumors, it is not in current use in the medical literature. Melanocytes originate in cells of the neural crest, which forms between the dorsal neural tube and the overlying epidermis following closure of the neural tube in the first month of embryogenesis. From there, they migrate to their final destinations principally in the epidermis, eyes, and some mucous membranes (particularly oral cavity, nasal cavity, esophagus, anorectal area, male and female genitourinary areas, and conjunctiva). Melanoma can develop from melanocytes in any of these locations. In a series of nearly 85,000 cases reported from hospital-based registries in the United States to the National Cancer Research Database for the period 1985–1994, 91.2% were cutaneous, 5.3% ocular, 1.3% in mucous membranes, and 2.2% of unknown primary site (Chang et al., 1998). Compared with both cutaneous and ocular melanomas, patients with mucous membrane melanomas tended to be older, more likely to be female, more likely to be other than non-Hispanic White, and to have poorer survival from their cancer. This chapter will deal in detail only with cutaneous and ocular melanomas. “

CUTANEOUS MELANOMA Cutaneous melanoma is the most lethal form of skin cancer. There were an estimated 160,000 new cases diagnosed worldwide in 2002, more than 80% in developed countries, and 41,000 deaths, 66% in developing countries. Sun exposure is its major cause, although a variety of phenotypic and genotypic characteristics contribute substantially to the risk of this complex disease. The incidence of cutaneous melanoma has been rising in most populations of European origin over about the past 50 years; by 6% per year in the United States in the 1970s and by ~3% per year over the past two decades (American Cancer Society, 2005). The incidence in Australia is exceptionally high, and melanoma represents the fifth most commonly diagnosed cancer among American men and sixth among American women. Trends in melanoma incidence reflect changing patterns of exposures and medical practice patterns among other factors, while mortality trends clearly indicate that the underlying melanoma burden has also increased over time in those of European origin. Melanoma detected at early stages has a high surgical cure rate, and adjuvant therapy with interferon offers a modest survival advantage to those with intermediate-stage disease. However, advanced primary melanoma carries a very poor prognosis, and there are no effective treatments for metastatic melanoma. To frame the epidemiology of melanoma, we begin with a brief overview of the biology and pathogenesis of melanoma, followed by detailed consideration of descriptive epidemiology and ecologic studies, environmental factors, host factors (both phenotypic and genetic susceptibility), and a summary of the pathogenesis of melanoma. Opportunities for prevention and future directions are then discussed. The epidemiology of ocular melanoma is presented in the same organizational structure.

1196

CLASSIFICATION Anatomic Distribution The anatomic distribution of melanoma contributes to understanding of both biology and epidemiology of melanoma. There is a close relationship between histologic subtypes of melanoma and body location. Although superficial spreading and nodular melanomas are found in all locations on the body, lentigo maligna melanoma is largely restricted to the head and neck, especially the face. Acral lentiginous melanoma, by definition, is found on the glabrous skin of the palms of the hand and soles of the feet. These patterns may relate to patterns of sun exposure and differences in the somatic genetic changes found in melanomas (Bastian et al., 2000). Melanoma also arises in different anatomic locations in women and men. The trunk, including the back, abdomen, and chest, is the most common location for melanoma to arise in men. Melanoma occurs most commonly on the leg in women, especially the lower leg, but also including the hip and thigh (Fig. 63–1). Several epidemiologic studies have examined anatomic distribution of melanoma with respect to putative risk factors, and the data are clearly consistent with the hypothesis that sun exposure is associated with risk of melanoma. Elwood and Gallagher (1983) conducted two studies to evaluate this hypothesis, first showing that the incidence rate per unit of skin surface area was higher for intermittently sun-exposed body sites than for more continuously exposed sites in a clinical series of 281 patients, among patients up to age 50. A follow-up study with nearly complete ascertainment of all cancers diagnosed in British Columbia in 1991–1992 confirmed this hypothesis (Elwood and Gallagher, 1998). Measuring melanoma density by examining the observed number of melanomas by body site compared with the number that would be expected based on the proportional surface area of body sites permitted relative tumor densities to be estimated. Melanoma density was highest on the back for both men and women under age 50, whereas increased risks for the ear and neck at older ages were evident in males. Males also had significantly fewer melanomas than expected on the legs at all ages. In contrast, females had a high density of melanoma for the leg, especially at older ages, and significantly fewer melanomas of the chest at all ages. These data are generally consistent with findings from Denmark and Australia, where the incidence rates per unit surface area were highest for the back in men and leg in women in the large study of 2376 Danish melanoma cases (excluding lentigo maligna melanoma) diagnosed between 1978–1982 (Osterlind et al., 1988). Data from Queensland, however, showed extremely high rates of melanoma of the ear for men, followed by the face, neck and shoulders, and back, and the highest melanoma rates in the face and shoulders for women (Green et al., 1993); that is, the highest rates were on the more continuously exposed sites rather than the intermittently exposed sites. Sex differences in the numbers of nevi among children follow an anatomic distribution that presages the anatomic distribution of melanoma in adults, after adjustment for host characteristics, sun exposure, and sun protection habits (Autier et al., 2004).

HISTOPATHOLOGY Melanoma represents a process of unregulated clonal growth that differs from other melanocytic proliferations in several important

1197

Cutaneous and Ocular Melanoma 60% 50% 40% 30% 20% 10% 0% H&N

Trunk

Arms

Legs

Melanoma Location

Other Males Females

Figure 63–1. Anatomic distribution of incident melanoma, data from Elwood and Gallagher (1998).

ways (Fig. 63–2). Melanocytes are normally found at the border between the epidermis and dermis, along a basement membrane called the rete ridge. Melanocytes may grow in a linear, single-file (“lentiginous”) pattern that appears on the surface of the skin as a freckle. Freckles are benign, and typically have little or no potential for malignancy. Melanocytes that cluster into nests constitute a nevus (“mole”), and are also benign. Nevi, although sometimes clonal, display a predictable behavior in which the cluster of melanocytes appears in youth or early adolescence, and then begins to descend into the dermis. As this nest of cells descends further into the dermis, the outward appearance of the mole becomes raised and papular. Indeed, with age, these nests of cells descend to the point that melanin can often no longer be seen, and the mole appears as a flesh-colored papule on the surface of the skin. Melanoma differs from benign melanocytic proliferations in several important ways. In the earliest stages of melanoma, melanocytes acquire features of malignancy such as nuclear atypia and pleomorphism, with slightly larger, irregularly shaped cells. In addition to these cytologic features, melanoma cells do not follow the predictable behavior of nevi, and the melanoma cells can be found

throughout the epidermis, rather than restricted to the epidermisdermis border. Melanoma that is restricted to the epidermis is designated melanoma in situ. As melanoma progresses, the melanoma cells penetrate the basement membrane and invade the dermis. Two of the major types of melanoma, superficial spreading and nodular melanoma, are represented in the right half of the figure. Both illustrations show invasive melanoma, but with the nodular melanoma forming a more discrete “nodule”. The original staging system by Clark classified melanoma by relating its lower border to anatomical planes in the skin (Clark et al., 1969); later systems by Breslow and others classified it by measuring its depth in millimeters (Breslow, 1970). The nodular melanoma on the right illustrates a feature called “ulceration,” which is an independent prognostic factor for melanoma. Ulceration is simply the disruption of the integrity of the epidermis, which is recognizable both clinically and histologically.

PRECURSOR NEOPLASTIC LESIONS Acquired Nevi Benign acquired nevi are a well-established risk factor for melanoma, and have been confirmed as such in virtually every case-control study conducted since the 1980s. Early studies by Beral et al. (1983) and Holman and Armstrong (1984) have been replicated numerous times, and a recent meta-analysis summarized the results of 46 studies, as shown in Table 63–1 (Gandini et al., 2005). In addition to serving as markers of increased risk, nevi are direct precursors of melanoma in some fraction of cases. However, it is important to recognize that nevi are observed in 1% of infants and are observed in virtually all adults. Thus the malignant potential of common acquired nevi is miniscule. Nevi are observed in histologic contiguity with melanoma in approximately 20%–30% of melanoma cases (Gruber et al., 1989; Kruger et al., 1992; Massi et al., 1999), although the reported range varies widely, suggesting that nevi can be direct precursors. Nevi are more likely to be associated with superficial spreading melanoma than other types, consistent with the concept of heterogeneity of melanoma histogenesis (Gruber et al., 1989; Whiteman et al., 2003). More continuously sun-exposed areas are the most typical locations of lentigo maligna melanoma (LMM), and nevi are rarely observed in LMM. The evidence to date suggests that etiologic pathways for different types of melanoma may be distinct, but the number of acquired nevi are important indicators of risk for the most common types of melanoma.

Figure 63–2. Distinct clinical and histologic features of melanocytic lesions and melanoma. (Source: Illustration by Bang Wong/ClearScience.)

1198

PART IV: CANCER BY TISSUE OF ORIGIN

Table 63–1. Meta-Analysis of Risk Associated with Nevi Type of Nevi

Number of Nevi

Relative Risk (95% CI)

common acquired nevi Whole Body

Arms

atypical nevi

0–15 16–40 41–60 61–80 81–100 101–120 0 1–5 5–10 11–15

1.00 1.47 (1.36–1.59) 2.24 (1.90–2.64) 3.26 (2.55–4.15) 4.74 (3.44–6.53) 6.89 (4.63–10.25) 1.00 1.44 (1.29–1.60) 2.48 (1.90–3.23) 4.82 (3.05–7.62)

0 1 2 3 4 5

1.00 1.60 (1.38–1.85) 2.56 (1.91–3.43) 4.10 (2.64–6.35) 6.55 (3.65–11.75) 10.49 (5.05–21.76)

Source: Adapted from Gandini et al., 2005.

Congenital Nevi Congenital nevi are typically defined as those present immediately at birth, whereas acquired nevi are classified as those that develop after 6 months of age. Congenital nevi are categorized by size as small (<1.5 cm), medium (1.5–19.9 cm), or large (>5% of body surface area in pre-adolescents and 20+ cm in adolescents and adults), and the malignant potential corresponds to these classifications. Melanoma arising in small congenital nevi is exceptionally rare, is slightly more common in medium-sized lesions, and is an important concern in large “bathing trunk” nevi (Habif, 2004). At least two cohort studies have attempted to quantify the risk of melanoma associated with congenital nevi. A British study of 265 patients with congenital nevi by Swerdlow et al. (1995) observed fatal melanoma arising in 2 of 33 large congenital nevi (6%), but no incident melanomas in the other 232 congenital nevi. Egan et al. (1998) observed 3 melanomas in 2 of 46 patients (4.3%) with large congenital nevi. Little else is known about the epidemiology of melanoma in the congenital nevi, although it has been suggested that risk is dependent on location, age, and histologic characteristics of the congenital nevus, in addition to size.

Atypical Nevi Atypical nevi, sometimes called dysplastic nevi based on the histopathologic features that accompany their clinically atypical appearance, confer an important risk for melanoma. Similar to common acquired nevi, atypical nevi serve as both markers of risk and precursors, with the important distinction that the risks for atypical nevi are higher (Table 63–1). A review by Tucker and Goldstein (2003) also summarized the risks associated with atypical nevi, emphasizing the consistency of the data from studies conducted in diverse populations. Atypical nevi are distinguished from common acquired nevi based on their appearance, are usually greater than 5 mm, have an indistinct or slightly irregular border and heterogeneous coloring, but do not meet the threshold of irregularities in asymmetry, borders, coloring, or size for melanoma. The risk of melanoma conferred by atypical nevi is not simply a function of number of nevi; atypical nevi appear to be qualitatively distinct (Roush et al., 1988; Halpern et al., 1991). Prospective studies confirm the increased risk associated with atypical nevi (Rigel et al., 1989; Halpern et al., 1993; Kelly et al., 1997). Atypical nevi can arise as part of a familial syndrome (described below in detail) or be simply sporadic. The high risk of melanoma with atypical nevi appears to be greater in the familial setting, but is still evident in other contexts.

MOLECULAR GENETIC CHARACTERISTICS OF NEVI AND MELANOMAS Nevi and melanomas arise through progressive accumulation of mutations, although the pathways and precursors are not yet as well elucidated as for some other cancers (e.g., colorectal cancer). A brief overview of somatic mutational events contributing to melanoma provides a context for appreciating the biology of melanoma and clues towards understanding the genetic epidemiology of melanoma and its precursors. For reviews of genetic pathways involved in melanoma, see (Chin, 2003) and (Rodolfo et al., 2004). Chromosomal abnormalities have been well characterized in melanoma, and common changes include losses in chromosomes 1, 6, 9, and 10 and gains in 6, 7, and 8 that are observed in more than 95% of melanomas (Nelson et al., 2000; Bastian, 2003; Rodolfo et al., 2004). Areas of chromosomal loss often indicate the presence of tumor suppressor genes, and areas of gain can reflect the presence of oncogenes. In melanoma, chromosomal losses have contributed to the localization and discovery of CDKN2A, and amplified regions correspond to CDK4, HRAS, BRAF, and other candidate genes. A constitutional deletion of chromosome 9p21 identified in a woman with 8 primary melanomas, multiple atypical nevi and a plexiform neurofibroma spanned a region that includes CDKN2A (Petty et al., 1993a; Petty et al., 1993b). This observation, in conjunction with studies of somatic loss of 9p21 (Fountain et al., 1992) and inactivation of CDKN2A (Kamb et al., 1994; Nobori et al., 1994), complemented linkage studies in this region and contributed to the positional cloning and identification of germline mutations of this important tumor suppressor gene (Hussussian et al., 1994; Kamb et al., 1994). Other cytogenetic studies have elucidated the distinct chromosomal signatures of melanomas arising in sun-exposed skin compared with those arising on acral surfaces and mucosal epithelium (Bastian et al., 2000; Bastian et al., 2003; van Dijk et al., 2003). An appreciation of the pathways that are dysregulated in melanoma also provides important clues that are relevant to epidemiologic studies of melanoma. The mitogen-activated kinase (MAPK) signaling pathway is frequently activated in melanoma, and activating mutations have been described in several key genes in this pathway, including NRAS and BRAF (Davies et al., 2002). Activation of this pathway leads to increased cellular proliferation, metalloproteinase expression and invasion, and altered expression and transcription of a host of genes relevant to melanocyte biology (Smalley, 2003; Smalley and Herlyn, 2004). Somatic losses of chromosome 10q implicate the phosphatidyl-inositol 3-kinase (PI3K-AKT) signaling pathway, and the PTEN tumor suppressor gene at this location is lost in 5%–15% of melanomas (Guldberg et al., 1997; Chin, 2003). P53 mutations are rarely observed in melanoma, but the p53 pathway is relevant due to the involvement of a downstream gene, Apaf-1, which is often lost in metastatic melanomas (Soengas et al., 2001). Thus the somatic mutational changes that accompany the development of melanoma offer insights into biologic behavior of melanoma and suggest candidate genes for further epidemiologic characterization.

DEMOGRAPHIC PATTERNS Mortality and Incidence Average annual age-standardized mortality rates (US 2000 Standard Million standard) from cutaneous melanoma in the whole United States and incidence rates in the SEER 11 Registries for the 5-year period 1997–2001 are summarized in Table 63–2. The incidence in both sexes and all races was 16.4 per 100,000, whereas mortality was less than a fifth of this at 2.7 per 100,000. With the exception of incidence in Hispanic whites, incidence and mortality rates were higher in men than women. Incidence rates were much higher in nonHispanic whites than Hispanic whites and least in non-whites. Mortality rates showed similar racial variation, but those in Hispanic whites and non-whites were higher relative to the corresponding incidence rates than were those in whites.

1199

Cutaneous and Ocular Melanoma

races (trend not able to be calculated in women of other races because of some zero rates) (Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) released April 2004). Trends in incidence were steeper overall and in whites than were mortality trends, and continued upwards largely unabated to 2001, the most recent year of observation available (Fig. 63–4). The rates increased at 4.1% (3.8–4.4) a year in white males and 3.0% (2.7–3.2) in white females. Incidence trends were also upwards in black males (2.2% a year, 95% CI: -0.6–5.0) but not black females (-1.5%, -2.8– -0.1), and in males (2.0%, 0.5–3.6) and females (2.8%, 1.2–4.4) of other races. A recent detailed analysis of trends by sex and age in mortality from melanoma in the whole United States from 1950–1994 (Jemal et al., 2000) reported peaks in mortality in male cohorts born in 1935–1950 and female cohorts born in 1930–1950 followed by falls in more recently born cohorts. Addition of a further 7 years of data to 2001 from SEER*Stat showed that these downtrends continued though in somewhat later cohorts in men (1953 on) and earlier cohorts in women (1948 on). A similar analysis of incidence trends in the SEER 9 Registries from 1973–1997 (Jemal et al., 2001) reported a peak in men born in 1950 followed by a plateau in rates in later cohorts. The rates also peaked with the 1950 cohort in women and were stable until those born in 1960, but then rose again in successive cohorts. Addition of a further 4 years of data showed clear evidence of a return to increasing rates in men born after 1950 and continuation of this trend in women (Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database: Incidence—SEER 9 Regs Public-Use, Nov 2003 Sub (1973–2001), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2004, based on the November 2003 submission). The most recent analysis of trends in incidence by thickness of melanoma, based on SEER Registries data for 1988–1997, showed that incidence was increasing in each of three thickness categories (thin <1.0 mm, intermediate 1.0–3.9 mm, thick 4+ mm) in both males and females. In females, the rate of increase was greater in thin lesions than intermediate or thick lesions (4.1%, 2.5% and 1.0% per year, respectively) whereas in men the increase was less variable across thickness categories and numerically greatest in thick melanomas (6.1% a year). The rate of increase was greatest in thick melanoma in older men (60+ years of age; 7.7% a year).

Table 63–2. Average Annual Age-Standardized* Incidence and Mortality Rates per 100,000 for Cutaneous Melanoma in SEER 11 Registries and the Whole United States Respectively in 1992–2001 Race Rate and Population

Sex

Incidence in SEER 11 Registries

Both Male Female

Mortality in whole US

Both Male Female

All

Non-Hispanic white

Hispanic white

Non-white

16.4 (16.2–16.6) 20.8 (20.5–21.1) 13.3 (13.1–13.6) 2.7 (2.7–2.7) 3.9 (3.9–4.0) 1.8 (1.8–1.8)

23.3 (23.0–23.5) 28.7 (28.2–29.1) 19.4 (19.1–19.8) 3.2 (3.2–3.2) 4.6 (4.6–4.7) 2.1 (2.1–2.2)

4.5 (4.1–4.8) 4.5 (4.0–5.1) 4.5 (4.1–5.0) 0.9 (0.8–0.9) 1.1 (1.0–1.3) 0.6 (0.6–0.7)

1.3 (1.2–1.5) 1.6 (1.4–1.9) 1.1 (1.0–1.3) 0.5 (0.4–0.5) 0.5 (0.4–0.5) 0.4 (0.4–0.5)

*Standardized to the US 2000 Standard Million standard; 95% confidence intervals in parentheses. Source: Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database: Incidence—SEER 11 Regs Public-Use, Nov 2003 Sub for Hispanics (1992–2001), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2004, based on the November 2003 submission.

Time Trends Mortality from melanoma in the United States increased steadily from melanoma’s first specific identification in mortality statistics in 1950 until about 1981 in women and 1998 in men and thereafter remained stable (Jemal et al., 2000) (Fig. 63–3). These overall trends largely reflect those in whites, in whom the vast majority of melanomas and melanoma deaths occur. The mortality rates in white males were estimated to have increased by 1.8% (95% CI: 0.5–2.0) a year and those in white females by 0.5% (0.3–0.7) a year over the period 1969–2001, whereas in blacks and people of other races these trends were, if anything downwards: respectively -1.2 (-1.7–-0.6) and 0.3 (-0.4–0.9) a year in black males and females and -0.6 (-1.7–0.5) in males of other

All males White males Black males Other males All females White females Black females Other females

5 4.5 4

3 2.5 2 1.5 1 0.5

Calendar year

2001

1999

1997

1995

1993

1991

1989

1987

1985

1983

1981

1979

1977

1975

1973

1971

0 1969

Rate per 100,000

3.5

Figure 63–3. Trends in age-standardized* mortality from cutaneous melanoma by race and sex in the United States from 1969–2001 (Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database: Mortality—All COD, Public-Use With State, Total U.S. (1969–2001), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2004. Underlying mortality data provided by NCHS (www.cdc.gov/nchs)). *Standardized to the US 2000 Standard Million standard.

1200

PART IV: CANCER BY TISSUE OF ORIGIN 30

All males White males Black males Other males All females White females

20

Black females Other females

15

10

5

No recent analysis of incidence trends by site of melanoma in the United States has been reported. Recent trends in other populations of mainly European origin have shown more or less parallel relative increases in site-specific rates with the exception of incidence of melanoma of the lower limbs in women, which has been falling in some populations (Gaudette and Gao, 1998; Hall et al., 1999; Marrett et al., 2001; Newnham and Moller, 2002).

Survival In 1995–2000 the overall 5-year relative survival from melanoma in the US SEER 9 Registries populations was 90.5%. White females had the highest survival at 92.4%, followed by white males, 89.1%, black females, 75.7%, and black males, 72.2% (Ries et al., 2004). Older people were at a survival disadvantage relative to younger people. In those 75+ years of age, 5-year relative survival was 85.8% compared with 92.6% at <45 years of age. In the intermediate age groups, though, there was little variation in survival: 90.5% at 45–54 years of age, 90.3% at 55–64 years, and 90.7% at 65–74 years. As would be expected, survival was highest in people whose melanoma was localized at diagnosis at 97.6%; it was 60.3% in those with regional metastases and 16.2% with distant metastases. For localized melanomas (83% of those registered by the SEER 9 Registries for 1995–2000) microscopically measured thickness and the presence of ulceration are the characteristics that most strongly predict poorer survival (Balch et al., 2001). Relative survival from melanoma, as recorded by the SEER Registries, has steadily increased with time (Fig. 63–5); and the survival proportions in whites for 1975–2000 were the highest recorded. This trend has paralleled the increasing incidence of melanoma in this population. Interestingly, melanoma survival is also positively correlated crosssectionally with incidence in different populations. In the 24 populations shown in Figure 63–6, 5-year relative survival increased with increasing incidence in both sexes up to an incidence of about 12 per 100,000 and thereafter did not increase further. Although there are other possibilities, including that melanoma is detected earlier or treated better in higher incidence populations than lower incidence populations, these parallels between survival and incidence suggest that melanoma developing in association with high sun exposure might behave less aggressively than melanoma

2001

1999

1997

1995

1993

1991

1989

1987

1985

1983

1981

1979

1977

1975

0 1973

Figure 63–4. Trends in age-standardized* incidence of cutaneous melanoma by race and sex in the SEER 9 Registries from 1973–2001 (Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database: Incidence—SEER 9 Regs Public-Use, Nov 2003 Sub (1973–2001), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2004, based on the November 2003 submission). *Standardized to the US 2000 Standard Million standard.

Rate per 100,000

25

Calendar year

arising with low sun exposure. This possibility was suggested by Lemish et al. in 1983 on the basis of data similar to, but more fragmentary than, those in Figure 63–6. It is supported by evidence that recalled sun exposure and solar elastosis in skin near the melanoma (a marker of high sun exposure) are associated with longer survival (Berwick et al., 2005). That these associations were independent of lesion thickness and behavior likely to lead to earlier diagnosis suggests that the association between melanoma incidence and survival is not explained by earlier detection in higher incidence populations.

ECOLOGIC ANALYSIS Age Before 10–12 years of age, melanoma incidence rates in populations of European origin are very low—approximately 0.1/100,000— generally greater in females than males and essentially invariant by single year of age. Thereafter they rise progressively with age (Fig. 63–7). The rate of rise with age is generally greater in males than females and this difference is more evident in populations with high rates of melanoma than those with low rates of melanoma.

Sex All-ages melanoma incidence rates are generally similar in males and females. In 117 cancer registries with populations of mainly European origin that reported in Cancer Incidence in Five Continents Volume VIII (Parkin et al., 2002), 61 had male-to-female (M : F) ratios of ageadjusted melanoma incidence of less than 1.0 and 53 had ratios greater than 1.0. These ratios varied from 0.32 in Concordia, Argentina, to 1.99 in New Orleans, United States, whites. There was, though, a high proportion of European registries among those with an M : F ratio of less than 1.0—56 of 61—and a low proportion among those with a ratio greater than 1.0—19 of 53, whereas all United States and Australian registries and most of the Canadian registries had M : F ratios greater than 1.0. This geographic pattern is underlain by a complex interaction between the effects of sex, age, and population on melanoma inci-

1201

Cutaneous and Ocular Melanoma

100

30

90

Survival %

70

20

60 50

15

40 M Survival F Survival M Incidence F Incidence

30 20 10 0

10

Incidence per 100,000

25

80

5

1995-97

1992-94

1989-91

1986-88

1983-85

1980-82

1977-79

1974-76

0

Calendar period Figure 63–5. Trends in melanoma incidence and survival in whites in the US SEER 9 Registries from 1974–1976 to 1995–1997 (Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database: Incidence—SEER 9 Regs Public-Use, Nov 2003

Sub (1973–2001), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2004, based on the November 2003 submission).

dence. Of the four populations shown in Figure 63–7, England had the lowest age standardized rates of melanoma: 5.78/100,000 in men and 7.45 in women. In the early 20s, male rates were about half female rates, by 65–69 years of age they equaled female rates, and by 85+ years they were 30% higher. At the other extreme rates were highest in the NSW population. Male rates were about 90% of those in females in their 20s, first exceeded them at 40–44 years of age, and were 2.5 times higher at 85+ years. Sweden and the US SEER population were intermediate between the two.

When first remarked on, the higher rates of melanoma in younger women than younger men were taken to suggest that female sex hormones might contribute to melanoma risk (Lee and Storer, 1980). The wide variation in the age at which male rates come to exceed those in females (40–44 to 70–74 years in the four populations represented in Fig. 63–7), however, suggests that this is unlikely. Although differences by age in the amounts or patterns of sun exposure in men and women are another possible explanation, this possibility has not been rigorously studied.

120

100

Survival %

80

60 Males Females

40

20

0 0

5

10

15

20

25

Incidence per 100,000 Figure 63–6. Relationship of melanoma survival to incidence in populations on which data for both were available in the mid 1980s (Canada: Ontario, Quebec and Saskatchewan; US: Atlanta, Connecticut, Detroit, Hawaii, Iowa, New Mexico, San Francisco-Oakland, Seattle, Utah;

Denmark: Finland; France: Bas Rhin; Italy: Latina; Netherlands: Eindhoven; Norway: Slovenia; Switzerland: Geneva, Sweden; UK: North Western England and Scotland; South Australia). Used with permission from Armstrong 2004a.

1202

PART IV: CANCER BY TISSUE OF ORIGIN 300

250

England Female

Incidence per 100,000

England Male Sweden Female

200

Sweden Male SEER White Female

150

SEER White Male New South Wales Female New South Wales Male

100

50

0 0-4 5-9 10- 15- 20- 25- 30- 35- 40- 45- 50- 55- 60- 65- 70- 75- 80- 85+ 14 19 24 29 34 39 44 49 54 59 64 69 74 79 84 Age group Figure 63–7. Melanoma age-incidence curves in four populations spanning a fivefold range in male and female averaged age-standardized incidence of melanoma (Source: Parkin et al., 2002.)

In populations of mainly non-European origin, M : F ratios were more often above 1.0 (43 registries) than below 1.0 (21 registries). There was a higher proportion of Asian populations in those with ratios above (84%) than below 1.0 (67%).

Race and Ethnicity People of mainly non-European origin have much lower rates of melanoma than do people of mainly European origin. These differentials are best shown in geographically bounded populations consisting of people of both European and non-European origin from which ethnicity specific cancer incidence rates are reported, such as those of Los Angeles County, California, United States, New Mexico, United

States, Israel, and Singapore (Fig. 63–8). The incidence rates in nonHispanic whites in the first two were many times higher than in the co-resident Asian (Los Angeles) or American Indian (New Mexico) populations. They were also much higher than in the Hispanic white population, which consists of people whose names suggest an origin in Latin America and could, therefore, include people of both mainly European and mainly non-European origin. A large differential is also seen between Jews and non-Jews (mainly of Arab descent) residing in Israel, but there is little evident difference between Chinese and Malays in Singapore; rates in both are very low. The distribution of melanoma by body site varies substantially between ethnic groups, as shown in Figure 63–9 for the multiracial population of Los Angeles County, United States (Cress and Holly, 1997). Non-Hispanic and Hispanic white people had roughly equal

Los Angeles Korean

Female Male

Los Angeles Japanese Los Angeles Filipino Los Angeles Chinese Los Angeles Black Los Angeles Hispanic White Los Angeles Non-Hispanic White New Mexico American Indian New Mexico Hispanic White New Mexico Non-Hispanic White Israel Non-Jews Israel Jews Singapore Malay Singapore Chinese

0

2

4

6

8

10

12

14

16

18

Incidence per 100,000 Figure 63–8. Incidence of melanoma by ethnic origin in four multi-ethnic populations from which ethnicorigin specific cancer incidence rates are reported (Source: Cress and Holly, 1997.)

20

1203

Cutaneous and Ocular Melanoma 60

50

Non-hispanic white Hispanic Asian & other Black

%

40

30

20

10

0 Head & neck

Trunk

Upper limbs

Lower limbs

Unknown

Figure 63–9. Site distributions (%) of melanoma in different ethnic groups in Los Angeles County, US, 1993–1997 (Source: Parkin et al., 2002.)

1.4 (0.5–2.3) in Vila Nova de Gaia, Portugal (Parkin et al., 2002), a 36-fold range. In females the range between these two populations, also the most extreme, was 19-fold from 1.96 (1.0–2.9) to 38.1 (36.8–39.3). The populations with the lowest rates were generally those of Southern and Eastern Europe and South America and those with the highest rates were populations of Australia and New Zealand, Scandinavia and Switzerland, and the white populations of the United States, including Hawaii, which has rates in the range of those of Australia and New Zealand. The majority of non-European populations have rates less than 1.0 per 100,000 (age adjusted to the world standard population) ranging

60

Socioeconomic Status

International Patterns Incidence of melanoma in populations of European origin varies widely around the world. In 1993–1997, the highest incidence in males was 51.1 (95% CI: 0.52–2.31) per 100,000 (age adjusted to the world standard population) in Queensland, Australia, and the lowest was

50

Rate per 100,000

40 30 20 10

m M

ed

iu

Lo w

w -lo

m ed M

iu m ed M

iu

ig

ig h

h

0 H

In a comprehensive review of variation in cancer incidence and mortality by socioeconomic status, Faggiano et al. (1997) concluded that “data for malignant melanoma suggested a regular pattern with the highest risk observed in the highest social strata, with very few exceptions” (Faggiano et al., 1997). This pattern was evident in both men and women in melanoma mortality data in England and Wales for 1949–1953 and is still evident in recent reports from the United States, Canada, and Australia encompassing melanoma incidence data reported up to 1995 (Gorey et al., 1998; Harrison et al., 1998; Lewis, 1999). A typical pattern is illustrated in Figure 63–10. From a two-level, nonlinear Poisson regression analysis of county level melanoma incidence data for 1973–1993 in whites in US SEER registry populations, Harrison et al. (1998) concluded that education rather than income was the primary socioeconomic determinant of melanoma incidence. There are insufficient data, however, for this specificity of effect to be established with any certainty.

Men Women

-h

distributions across the four site categories—head and neck, trunk, upper limbs, and lower limbs—though with a shift towards the trunk in non-Hispanic whites and the lower limbs in Hispanic whites. For Asian and Black people, the distribution was shifted much more towards the lower limbs with nearly 60% of all melanomas being on the lower limbs in Blacks. These lower limb melanomas are disproportionately on the soles of the feet. For example, in a recent Japanese series, 52% of incident melanomas arose on the lower limbs, and 28% were on the soles (Ishihara et al., 2001). Black and white Americans have about the same incidence rates of melanoma on the soles of their feet (Stevens et al., 1990). Thus, the higher relative proportion of plantar melanomas in Asian and Black populations is due to a lower incidence of melanoma on skin surfaces other than the soles of the feet.

Socioeconomic status

Figure 63–10. Variation in age-adjusted incidence of melanoma (world standard population) by socioeconomic status* in New South Wales Australia in 1991–1995 (Lewis, 1999). *Socioeconomic status was assigned to each subject’s local government area of residence by use of the Australian Bureau of Statistics’ Socioeconomic Index for Areas derived from the 1991 national census and categorized roughly into fifths of the population. The lowest number of cases in any category was 386.

1204

PART IV: CANCER BY TISSUE OF ORIGIN 8

25

7

Atlanta M

South Thames F

Incidence per 100,000

Incidence per 100,000

20

15 Los Angeles F Los Angeles M

10

Atlanta F Seattle M & F

6

Varese F

Warsaw M & F

3 2 Slovenia M & F

Sweden M & F

1

Detroit M & F

Zaragoza M & F

Israel Jews M & F

56-62

A

Varese M South Thames M

4

5

0

Saarland M & F

5

60-66

0

66-72

73-77

78-82

83-87

88-92

56-62

93-97

Calendar Years

B

60-66

66-72

73-77

78-82

83-87

88-92

93-97

Calendar Years

Figure 63–11. Trends in incidence of cutaneous melanoma in populations of mainly European origin that reported incidence rates in Cancer Inci-

dence in Five Continents from 1973–1997 to 1993–1997. (A) Higher rate populations. (B) Lower rate populations.

down to around 0.1 (Parkin et al., 2002). Most rates less than 0.4 are in Asian populations, including populations in China, India, Japan, Korea, Philippines, Thailand, and Vietnam. Those above 1.0 are predominantly non-European subpopulations in the United States, African black populations (Uganda and Zimbabwe), or mixed race populations in South America (Colombia, Costa Rico and Ecuador). Trends in incidence of melanoma in populations of mainly European origin for which data have been reported for at least 25 years are shown in Figure 63–11 (Cancer Incidence in Five Continents (1966; 1970; 1976; 1982); Muir et al., 1987; Parkin et al., 1992, 1997, 2002). Although there are perturbations from it, there has generally been a continuing upward trend in incidence over the whole of the period of observation in all populations shown. In 21 populations of mainly European origin with data available for the four 5-year periods 1978–1982 to 1993–1997, the annual % increase in incidence from the mid point of the first period to the mid point of the last ranged from 1.6%–7.5% in males (average 4.4%) and 0.5%–6.2% in females (average 3.4%). These overall upward trends hide moderation in the rise or even falls in incidence in younger people, particularly in populations in Australia, North America, and northern Europe (Giles et al., 1996; Bulliard et al., 1999; La Vecchia et al., 1999; Jemal et al., 2000; Jemal et al., 2001; Marrett et al., 2001; de Vries et al., 2003). There was much less evidence of any upward trend in incidence of melanoma in populations of non-European origin (Fig. 63–12). There was no evidence of any net increase in the lower rate populations of Asia and incidence may have fallen in Hong Kong. Rates were also reasonably steady in other non-European populations, though incidence may have increased in Los Angeles Japanese Males and Atlanta Black Males. Melanoma mortality rates in populations of mainly European origin are much less, and vary much less geographically, than the corresponding incidence rates. In 20 such populations with incidence rates for 1993–1997 and corresponding mortality rates for 1995 available (WHO, 2001; Parkin et al., 2002; Welfare, 2003) age-adjusted (World population) melanoma mortality in males varied 7.0-fold from 0.8–5.6 per 100,000 while incidence varied 15.5-fold from 2.6–40.1 (Fig. 63–11); mortality in females varied 3.1-fold from 0.9–2.8 while incidence varied 8.4-fold from 3.7–30.8. Similar comparisons were made between incidence rates and corresponding whole population mortality rates for the non-European origin populations of Hong Kong, China; Osaka, Japan; Seoul, Korea; Manila, Philippines; and Singapore (Chinese). Mortality varied from 0.2–0.5 per 100,000 in males and 0.1–0.3 in females. Mortality rates were generally between one-quarter and three-quarters of the corresponding incidence rates.

Like incidence, mortality from melanoma has steadily increased with time in populations of mainly European origin, as shown in Figure 63–13, which presents a representative selection of long-run mortality data for European, North American, and Antipodean populations. There is, though, evidence of cessation of the mortality increase, seen first in females in Sweden in about 1980, Australia and Canada also in the early 1980s, The Netherlands in the late 1980s, and England and Wales and Spain in the early 1990s. Cessation of the increase occurred several years later in men than women, and may not yet have occurred in Spain. As with incidence, there is considerable heterogeneity in the underlying age-specific mortality trends with falling mortality in younger age groups and continuing increases in older age groups (Giles et al., 1996; La Vecchia et al., 1999; Jemal et al., 2000; Jemal et al., 2001; Marrett et al., 2001; de Vries et al., 2003). Long-term mortality trend data have also been reported from Hong Kong, Japan and Singapore. In Hong Kong, age-adjusted mortality varied between 0.1 and 0.4 per 100,000 in both sexes, but with no evident trend, between 1966 and 1996. Rates were similarly variable in Singapore, but over a wider range (0.0–0.7). In Japan rates in men were initially reported as 0.1 per 100,000 from 1955–1968 and thereafter 0.2; in females rates were reported as 0.2 only in the last few years before 1997.

Migration Most studies of the influence of migration on risk of melanoma have studied the effect of migration from an environment of low ambient solar UV radiation to one of high ambient solar UV radiation, or vice versa. In the first such study reported, Movshovitz and Modan (1973) observed melanoma incidence rates in Jews who migrated to Israel from Europe to be as low as one-half, depending on age, of those in Jews born in Israel, most of whom were of European extraction. Similarly, the incidence in those who had migrated since 1948 was as low as one-half that in those who had migrated before 1948. They noted that “the consistent gradient within the European group from a high incidence in the Israeli-born through a medium incidence in the veteran foreign-born to the low rates in the more recent foreign-born points . . . toward a cumulative effect of sun exposure over time”. Mack and Floderus (1991) showed for internal migration in the United States and Swerdlow et al. (1995) showed for migration between the United Kingdom and New Zealand that migrants in both directions, along gradients in ambient solar UV radiation, had incidence rates of melanoma that were midway between those in the home and destination regions. A number of analyses have been made of the effects of age at migration to, and duration of residence in, the destination region on the risk of melanoma in migrants relative to that in those born in the destina-

1205

Cutaneous and Ocular Melanoma

1.2 Hong Kong M & F

Incidence per 100,000

1

0.8

0.6 Singapore Chinese M & F

0.4 Shanghai M & F Mumbai M & F Osaka M & F

0.2

0 60-66

66-72

73-77

78-82

83-87

88-92

93-97

Calendar Years

A

3.5

Incidence per 100,000

3

Colombia Cali M & F

2.5 2

Los Angeles Black M & F Atlanta Black

1.5

Los Angeles Japanese M & F

1 0.5 Israel Non-Jews M & F Atlanta Black F

0 60-66 B

Detroit Black F

66-72

73-77

78-82

Detroit Black M

83-87

Calendar Years

tion region. Almost without exception, for populations of European origin migrating along a gradient of ambient solar UV radiation, they have shown a higher risk of melanoma with earlier age at arrival, increasing duration of residence, or both in a destination region of higher UV radiation (Holman and Armstrong, 1984; Parkin et al.,1990; Khlat et al., 1992; Autier et al., 1997). Both associations were observed by Holman and Armstrong (1984), who were able to show, in the absence of complete reciprocation between duration of residence and age at arrival in their case-control study population, that the effect of age at arrival appeared to fully explain the effect of duration of residence on risk of melanoma. In a model containing both variables, the adjusted ORs for age at arrival were 0.89 (95% CI: 0.44–1.80) for arriving at 0–9 years of age, 0.34 (0.16–0.72) at 10–29 and 0.30 (0.08–1.13) at 30+ years of age, with reference to those who were born in Australia. For duration of residence, the adjusted ORs were 0.80

88-92

93-97

Figure 63–12. Trends in incidence of cutaneous melanoma in populations of mainly nonEuropean origin that reported incidence rates in Cancer Incidence in Five Continents from 1973–1997 to 1993–1997. (A) Populations in Asia (excluding non-Jews in Israel). (B) Other non-European populations.

(0.41–1.56) for 25–39 years, 0.93 (0.26–3.25) for 40–59 and 1.02 (0.20–5.08) for 60+ years, with reference to those resident for 0–24 years. These observations have been taken to suggest that sun exposure in early life is particularly important in increasing risk of melanoma.

ENVIRONMENTAL FACTORS Sun Exposure Sun exposure is the main cause of melanoma. From differences in its incidence between usually and rarely exposed body skin and between white-skinned and black-skinned people living in the same environment, it has been estimated that over 90% of melanomas are due to sun exposure in mainly European-origin populations living in areas of

1206

PART IV: CANCER BY TISSUE OF ORIGIN 6

Australia Canada Netherlands Spain Sweden England and Wales

Rate per 100,000

5

4

3

2

1

1991

1993

1995

1997

1999

1993

1995

1997

1999

1989

1987

1985

1983

1981

1979

1977

1975

Calendar year 3

Australia Canada Netherlands Spain Sweden England and Wales

2.5

Rate per 100,000

1991

A

1973

1971

1969

1967

1965

1963

1961

1959

1957

1955

0

2

1.5

1

0.5

B

high ambient solar UV radiation (Armstrong and Kricker, 1993). An expert working group of the International Agency for Research on Cancer concluded, in 1992, “There is sufficient evidence in humans for the carcinogenicity of solar radiation. Solar radiation causes cutaneous malignant melanoma . . .” (Parkin et al., 1992). The present epidemiological evidence for and against this causal association is summarized below.

In Fair-Skinned People, Melanoma Occurs Most Frequently on Skin That is Usually Exposed to the Sun When incidence of melanoma by body site is expressed per unit surface area of skin, it has generally been shown to be highest on the face (in both men and women) or ears (in men) (Green et al., 1993; Franceschi et al., 1996; Bulliard et al., 1997). When site is reported in sufficient detail, the lowest incidence is generally on the buttocks or soles of the feet (Green et al., 1993; Elwood and Gallagher, 1998).

1989

1987

1985

1983

1981

1979

1977

1975

1973

1971

1969

1967

1965

1963

1961

1959

1957

0 1955

Figure 63–13. Long-term trends in mortality from cutaneous melanoma in six representative populations of mainly European origin. (A) Males. (B) Females.

Calendar year

Consistently, melanomas on the breasts in women have been observed to be more common on the skin of the upper than lower quadrants, and may have become more centrally located on each breast more recently (Gillgren et al., 2002; Bono et al., 2003). Also, limb melanoma in Eastern Australia has been observed to occur at greater density on the lower than the upper leg and on the forearm than the upper arm (Green et al., 1996). Inconsistently, melanoma density was observed to be greater on the shoulders than on the upper arm or forearm (Green et al., 1996) and very low on the backs of the hands, one of the most exposed sites (Green et al., 1993; Green et al., 1996; Elwood and Gallagher, 1998). These patterns vary with age. Although the face has the highest density of melanoma in older people, other sites, particularly the back, shoulders, and upper arms (in both men and women), and the legs (in women) have higher densities in younger people (Elwood and Gallagher, 1998; Bulliard, 2000). These differences may be explicable in

Cutaneous and Ocular Melanoma terms of the apparent importance of pattern as well as amount of sun exposure in causing melanoma (Armstrong, 2004b).

People with Fair Skin Who Burn Easily and Tan Poorly When Exposed to the Sun Are Those at Greatest Risk of Melanoma These patterns have been consistently observed in studies of individual risk of melanoma. They are also evident in descriptive studies. There is little overlap globally between the higher rates of melanoma seen in people of mainly European origin, who generally have the lightest skin, and the lower rates seen in all other ethnic groups. Similarly, people of lighter-skinned ethnic origin have higher melanoma rates than those of darker-skinned origin living in the same environment. The consistently very low rates of melanoma in people of Asian origin are puzzling because many Asians, particularly East Asians, have comparatively light skin. Recent results suggest, however, that East Asians have an unusually high melanogenic dose response following solar-simulated UV exposure compared with people of European and Hispanic origin (those of Hispanic origin had a similar high average melanin index to the East Asians) (Wagner et al., 2002). This high melanogenic dose response might explain the very low risk of melanoma in East Asians.

Incidence of Melanoma Increases with Increasing Ambient Solar (or Solar UV) Radiation at Places of Residence At the ecological level, a relationship between increasing ambient solar radiation and increasing melanoma incidence in people of mainly European origin was first documented in North America, initially between melanoma mortality and latitude and subsequently between melanoma incidence and measured ambient solar UVB radiation (Elwood et al., 1974; Scotto and Fears, 1987). This pattern has also been reported for melanoma incidence in Australia, both between States and within the most populous State (New South Wales) (Jelfs et al., 1994; Armstrong, 2004b). The pattern for Europe appears to be more complex: risk has been shown to fall rather than increase with falling latitude to about 52º N and then to increase as latitude falls thereafter (Armstrong, 1984). This apparent inconsistency may be due, in part at least, to confounding of latitude with skin pigmentation (fairer skin in the north and darker skin in the south). There is evidence that the latitude gradient in melanoma mortality in the United States has become less recently, perhaps because of greater population mobility or more extensive application of sun-control measures (Jemal et al., 2000). Consistently with the United States and Australian patterns, when populations migrate along a gradient of ambient solar UV radiation, both between and within countries, incidence in the migrating group increases relative to that in the non-migrating group when migration is to an area of higher ambient UV and falls when it is to an area of lower ambient UV. Melanoma incidence also increases with increasing individual exposure to environments of high ambient solar UV (as measured by periods of residence at lower latitudes or a cumulative measure based on latitude), global solar irradiance, or solar UV irradiance at all places of residence. Combining results summarized previously (Armstrong and English, 1996) with those published more recently (Autier et al., 1997; Whiteman and Green, 1998; Robsahm and Tretli, 2001; Fears et al., 2002), the random effects summary OR for the highest category of exposure in each of 10 studies was 1.9 (95% CI: 1.4 to 2.5). When the results of Fears (2002) and colleagues, which related to a 10% increase in continuously represented ambient UVB irradiance, were excluded from the summary, the OR was 2.2 (95% CI: 1.8 to 2.6).

Individual Risk of Cutaneous Melanoma Is Weakly Positively Associated with High Total Lifetime Sun Exposure; It Is More Strongly Positively Associated with High Intermittent Pattern (Recreational) Sun Exposure but Weakly Negatively Associated with High Chronic or More Continuous Pattern (Occupational) Sun Exposure In a carefully conducted and comprehensive systematic review and meta-analysis, Gandini et al. (2005) identified 58 case control or

1207

cohort studies that reported results for the relationships to risk of melanoma of one or more of total (estimated amount of exposure regardless of pattern) sun exposure, intermittent pattern (largely recreational) sun exposure, chronic (more continuous pattern and almost entirely occupational) sun exposure, or history of sunburn. Measures of sun exposure in adulthood were preferred to childhood measures and results for the highest reported category of sun exposure were used for each included study. Results that were adjusted for age, sex, and characteristics indicative of sun sensitivity were chosen for the analysis whenever possible. Measures that were also adjusted for number of nevi were avoided because nevi may be intermediate steps in the pathway from sun exposure to melanoma. Extensive sensitivity and heterogeneity analyses were done. Random effects models were used when there was significant heterogeneity in results of individual studies. The summary relative risk estimate for total sun exposure was 1.34 (95% CI: 1.02–1.77). That for intermittent pattern exposure was 1.61 (95% CI: 1.31–1.99) and that for more continuous pattern exposure was 0.95 (95% CI: 0.87–1.04). The highest relative risk estimate was for sunburn: 2.03 (95% CI: 1.73–2.37). This general pattern of results is similar to that previously reported from a meta-analysis of 29 published studies (Elwood and Jopson, 1997). Several studies that included controls with dermatological diseases contributed importantly to heterogeneity in relative risk estimates for more continuous exposure; when they were excluded, the estimate was 0.87 (95% CI: 0.74–1.02), which is nearly the same as that reported from the previous meta-analysis (Elwood and Jopson, 1997). Only the results for sunburn showed consistent evidence of bias in favor of publication of results with higher relative risk estimates. Correction for this bias produced a pooled relative risk estimate of 1.55 (95% CI: 1.31–1.83). These results generally support the proposition that increasing personal sun exposure, particularly exposure with a more intermittent pattern, increases risk of melanoma. With respect to the possibility that risk of melanoma falls with an increasing more continuous pattern or occupational sun exposure, it is important to note that the baseline category of no or little occupational exposure is not a “no sun exposure” category. Although it might include some people with little or no sun exposure, it would also include people with quite heavy recreational sun exposure. Thus, it would be wrong to infer from these results that occupational sun exposure is protective against melanoma, though it might well carry less absolute risk of melanoma than an equivalent amount of recreational sun exposure. However, even the OR for recreational sun exposure does not appear compatible with estimates that more than 90% of melanomas in European-origin populations living in areas of high ambient solar UV radiation are due to sun exposure (Armstrong and Kricker, 1993). In their case-control family study of melanoma in Queensland, Australia (the area of highest melanoma incidence in the world), Siskind et al. (2002) estimated a population-attributable risk of melanoma for lifetime sun exposure of 44.3% (95% CI: 32.1–53.9). There are, however, at least two difficulties here: lack of a truly unexposed reference category with which the effects of sun exposure can be compared and the inevitable error in recall of sun exposure, which bedevils all studies of sun exposure and chronic disease (English et al., 1998). Thus, the inconsistency between 44% and 90% may not be as great as it seems.

Risk of Cutaneous Melanoma Increases with Increasing Incidence or Prevalence of Other Indicators of Cutaneous Sun Damage There are a range of age-related skin changes that are known or thought to be caused by sun exposure, including wrinkling and loss of fine markings on the skin, pigmentation changes (solar lentigines), solar keratoses, and non-melanocytic skin cancers. Almost all relevant studies have found these changes to be significantly positively associated with melanoma. Combining results summarized previously (Armstrong and English, 1996) with those published more recently (Bataille et al., 1998; Tabenkin et al., 1999; Siskind et al., 2002), the summary OR for the highest categories of presence of these changes in 10 studies was 1.9 (95% CI: 1.6–2.2). It is relevant to note that the number of solar keratoses has been shown to be negatively correlated with the number of melanocytic nevi (Bataille et al., 1998), thus confounding with number of nevi is unlikely to explain this positive association.

1208

PART IV: CANCER BY TISSUE OF ORIGIN

There Is Weak Evidence that Protection against Sun Exposure Reduces Risk of Cutaneous Melanoma Three controlled trials have addressed the effect of sunscreen use or other sun protection measures on risk of melanoma or melanocytic nevi. In a study of 1621 people in Queensland, Australia, Green et al. (1999) showed that daily application of sunscreen reduced incidence of squamous cell carcinoma of the skin (relative incidence in intervention group to control group 0.61 [95% CI: 0.46–0.81] over an average of 4.5 years of follow-up). However, only 12 melanomas were diagnosed during the follow-up period and their distribution between sunscreen and usual practice group has not been reported. Gallagher et al. (2000) reported results of a trial of a similar intervention in school children in Vancouver, Canada. The median increase in the number of melanocytic nevi in the intervention group over the 3 years of the trial was 24, compared with 28 in the non-intervention group (p = 0.048). Milne et al. (2002) reported results of a trial of a comprehensive sun protection intervention in school children entering 33 primary schools in Perth, Australia. Numbers of melanocytic nevi increased 3%–11% less in two intervention groups (moderate and high intensity) than in the non-intervention group, depending on the site on the skin. These differences were not statistically significant (P value 0.1–0.6 depending on site). There is thus limited evidence that sun protection can reduce the expected increase in numbers of nevi with age in children. Successful reduction of numbers of nevi might be reasonably interpreted as suggesting that risk of melanoma would also be reduced because of the strong association between number of nevi and melanoma risk. The most recent and most comprehensive overview of published studies of the association of melanoma risk with use of sunscreen reported a summary OR for ever use of sunscreen of 1.0 (95% CI: 0.8–1.2) based on 18 studies (Dennis et al., 2003). There was substantial heterogeneity among the studies. Further analyses were done to try to account for possible causes of heterogeneity, including lack of control for sun sensitivity and sun exposure, which are potentially strong confounding variables. The summary OR was 0.8 (95% CI: 0.6–1.0) when the analysis was restricted to results that were adjusted for sun sensitivity. In addition, in an attempted dose-response analysis across ordered categories of sunscreen use, the OR was 0.76 (95% CI: 0.65–0.90) per category when the five studies that adjusted for sun sensitivity and sunburn (as a measure of sun exposure) were pooled. Overall, however, there was little evidence of a dose-response relationship. As noted above, there is evidence that mortality and, to a lesser extent, incidence of melanoma are falling in younger people in many populations of mainly European origin. Although these favorable trends might be due to reduction in sun exposure with increasing public education on sun protection (Bulliard and Cox, 2000; Marrett et al., 2001), it has been argued that it is still too early to see such effects (English and Milne, 1999). There may, however, be late-stage effects of sun exposure in melanomagenesis that could produce an early response to effective sun protection (Armstrong, 2004b).

Manmade Sources of UV Radiation Exposure to fluorescent lighting; sunlamps, sunbeds, and related devices usually used for tanning; and long-wave UVA and psoralen in photochemotherapy for skin disease are the only circumstances in which manmade sources of UV radiation have been investigated to any extent epidemiologically as possible causes of cutaneous melanoma. Recent measurements of emission spectra of incandescent, fluorescent, and halogen lamps purchased from local retail outlets in Tennessee, United States, showed that all emitted appreciable amounts of UVB and UVA and, in some cases, shorter wavelength UVC radiation (Sayre et al., 2004). There has been little recent interest in the possibility that exposure to fluorescent lights might increase risk of melanoma. Combining results summarized previously (Armstrong and English, 1996) and one study published more recently (Holly et al., 1995), the summary OR was 1.52 (95% CI: 1.14–2.04) for the highest exposure categories of the four studies that had adjusted the associa-

tion with fluorescent light exposure for possible confounding with sun exposure. For all seven studies that could be summarized, the summary OR was 1.27 (1.05–1.53). The two studies that could not be included in the summary because confidence intervals were not reported had ORs of 0.6 and 0.7. The observations on UV emissions and the results of the summary analyses suggest that exposure to fluorescent lighting increases risk of melanoma. Some 10 studies have evaluated risk of melanoma with use of sunlamps, sunbeds, and related devices usually used to acquire or maintain a tan. From an overview analysis of the published results of these studies, Gallagher et al. (2005) reported a summary OR of 1.25 (1.05–1.49) for ever use. There was significant heterogeneity among the individual study results. In five studies that reported on risk with first use of a sunlamp or sunbed at or before 35 years of age or younger the OR was 1.69 (1.32–2.18) without significant heterogeneity. In six studies for which dose response was reported the summary OR for the highest categories of exposure was 1.61 (1.21–2.12), again without significant heterogeneity. Thus it appears that use of sunlamps or sunbeds does increase risk of melanoma. UVA radiation is used for treatment of psoriasis and a range of skin diseases in combination with orally administered 5- or 8methoxypsoralen or baths containing trimethoxypsoralen. This treatment is commonly referred to as PUVA. In a cohort of US patients who began PUVA treatment in academic dermatology centers in the United States in 1975–1976, no excess of melanoma was observed until more than 15 years after treatment began; in the period 1991–1996, risk of melanoma was 5.4 (95% CI: 2.2–11.1) times that expected in the PUVA-treated patients (Stern et al., 1998). This excess continued during further follow-up to 1999 and showed evidence of a twofold higher risk with 200+ PUVA treatments relative to less. There was, however, no evidence of an increase in risk of melanoma in a larger Swedish cohort of PUVA-treated patients (Lindelof et al., 1991). These patients were treated in 1974–1985 and followed-up until 1994; the relative risk for melanoma was 1.1 (95% CI: 0.6–1.9) in the whole cohort and 0.9 (95% CI: 0.4–2.3) in those followed-up for 15 or more years. A number of explanations have been considered for the inconsistency between these studies; none appears to explain it adequately or to raise serious questions about the validity of the observed excess risk in the US patients. It seems probable therefore, that PUVA use does increase risk of melanoma.

Occupation Occupational exposure to solar and manmade UV radiation has been dealt with above. Either, but particularly sun exposure, may be confounded with exposure to other agents in the occupational environment. Because of the dominance of retrospective cohort studies in occupational carcinogenesis, it has been rarely possible to adequately deal with potential confounding by sun exposure; this must be taken into consideration when interpreting studies of melanoma and occupation or occupational exposures.

Ionizing Radiation While a number of studies suggest that melanoma incidence may be increased in people exposed to ionizing radiation, the UNSCEAR 2000 report concluded “To date, there has been little indication of an association between ionizing radiation and malignant melanoma . . . but the data are sparse” (UNSCEAR, 2000). Among these sparse data are weak indications of an increase in risk of melanoma in people exposed to radiation at Hiroshima and Nagasaki: the excess relative risk per Sievert was 2.1 (95% CI <0.01–12) based on 10 observed melanoma cases (Ron et al., 1998). Since the UNSCEAR report, further studies of cancer following occupational radiation exposure have been reported. In a US Radiologic Technologists Cohort the standardized incidence ratio for melanoma was 1.59 (95% CI: 1.38–1.80) and the risk was higher in those first employed before 1950 (Freedman et al., 2003; Sigurdson et al., 2003). In French Atomic Energy Commission workers, the standardized mortality ratio (SMR) for melanoma was 1.50 (90% CI:

Cutaneous and Ocular Melanoma 1.04–2.11) in men and 1.24 (0.42–2.84) in women (Telle-Lamberton et al., 2004). In the Canadian National Dose Registry Cohort the SMR was 1.16 (95% CI: 1.04–1.30) (Sont et al., 2001), but this increase was mainly concentrated in dentists, who had the least exposure to radiation. In reporting their French Atomic Energy Commission workers study, Telle-Lamberton et al. (2004) listed the seven reported sets of observed cases of melanoma and SMRs from reports of mortality in 27 nuclear industry worker cohorts; the summary SMR from these results is 1.04 (95% CI: 0.87–1.21). There is evidence that melanoma risk is increased following radiation therapy for cancer, with or without associated chemotherapy. In summarizing 20 relevant studies, Shore (2001) noted that there were clear excesses of melanoma after treatment for childhood cancer, consistent excesses after treatment of lymphopoietic and testicular cancer and bone marrow transplantation, but no evidence of an excess in those treated for cervical or ovarian cancer. He judged the evidence insufficient for causality because of the possibility of surveillance bias and confounding with chemotherapy and immunological effects associated with lymphopoietic neoplasms; to which might be added confounding with sun exposure through socioeconomic selection for cancer treatment. Some of these issues were addressed in a recent cohort study of children treated for cancer. The standardized incidence ratio (SIR) for a new primary melanoma more than 3 years after a first cancer was diagnosed was 9.1 (95% CI: 3.6–18), based on six observed cases in one of the two cohorts (Guerin et al., 2003). All these cases had followed treatment with radiotherapy. For patients treated only with radiotherapy the SIR was 8.4 (95% CI: 1.1–20); for those treated with radiotherapy and chemotherapy it was 21 (95% CI: 6.4–48). This and other studies have reported melanomas arising in the radiation field (Corpron et al., 1996). Without a full systematic review and meta-analysis, it is uncertain whether or not there is an increase in risk of melanoma associated with exposure to ionizing radiation. Moreover, confounding of ionizing radiation with sun exposure is a possibility that has generally not been addressed.

Airline Flight Crews There is consistent evidence that incidence of melanoma is higher in airline flight crews than the general population. In a recent report from a cohort of 10,032 male, Scandinavian airline pilots, the relative incidence was 2.3 (95% CI: 1.7–3.0) (Pukkala et al., 2002). The increase in melanoma risk was similar across the main cutaneous sites for melanoma, which suggests a similar distribution of risk factors to that in the general population. As in other similar studies, non-melanocytic skin cancer incidence was also increased. This evidence has been gained in the wider context of evaluating the possibility that exposure to cosmic (ionizing) radiation in the course of high-altitude flight might increase cancer risk (Boice et al., 2000). Most authors, however, have considered the increased opportunity for recreational sun exposure that accompanies long-distance flight to be the probable explanation for the excess risk of melanoma in air crews. A comparative survey of Icelandic air crews and a sample of the Icelandic population confirmed a higher frequency of sunny vacations and episodes of sunburn in the air crews (Rafnsson et al., 2003).

Extremely Low Frequency Electromagnetic Fields Tynes et al. (2003) recently reported an increased risk of melanoma with residential exposure to electromagnetic fields (ELF) in a cohort of people living in a broad corridor along a high voltage power transmission line in Norway. From a nested, matched case-control study encompassing cases diagnosed from 1980–1996, the OR in the highest exposed category (time weighted average field strength 0.20 mT+ from birth or January 1, 1967, until date of diagnosis of the case) was 1.87 (95% CI: 1.23–2.83) and the P value for trend was <0.001. There was, however, no excess with occupational exposure to ELF in the same cohort, based on occupations recorded at the 1980 or 1990 censuses: OR 1.22 (0.81–1.82) for the highest exposed category. The authors observed that their finding for residential exposure “. . . did not have much support from the literature”. However, when the six studies to which they referred, and one other to which they did not refer

1209

(Theriault et al., 1994), were reviewed it was found that risk of melanoma was increased in the highest reported exposure category in six of these seven studies. These studies were generally based in whole populations or electrical industry employees and related mainly to occupational exposure inferred from industry or industry and job title; only one related to residential exposure inferred from proximity to power transmission lines. A meta-analysis of reported results for the highest exposure categories across all eight studies gave a relative risk estimate of 1.27 (1.20–1.35), which fell only slightly when the OR for occupational exposure of Tynes et al. (2003) was substituted for that for residential exposure, 1.26 (1.19–1.34). In at least one study, confounding with polychlorinated biphenyl (PCB) exposure was a possible explanation for an increasing risk of melanoma with increasing exposure to ELF (Tynes et al., 1994). No study addressed the possibility of confounding with sun exposure, which cannot be readily ruled out. Nonetheless, a causal association of ELF exposure with melanoma is a possibility.

Polyvinyl Chloride Manufacture Melanoma incidence is increased in men exposed to vinyl chloride during polyvinyl chloride (PVC) manufacture. Combining the results of two cohort studies, one with a total of 7 cases observed and 2.06 expected (Lundberg et al., 1993; Langard et al., 2000) and the other with 5 cases observed and 1.5 expected, gave a standardized incidence ratio of 3.4 (95% CI: 1.5–5.3). This association is probably too strong to be attributable to confounding with sun exposure and may indicate a real effect of vinyl chloride or some other chemical used in PVC manufacture.

Occupational Exposure to PCBs An increased risk of melanoma has been observed in several studies of people occupationally exposed to PCBs. A retrospective cohort study of 138,905 men who were first employed by US electrical utility companies from 1950–1986 and followed-up to the end of 1988 is the most recent and informative of these studies. It found an overall relative risk of melanoma of 1.29 (95% CI: 0.84–1.98) (Loomis et al., 1997); but when both latency and amount of exposure were considered, there was a 5% (95% CI: 1–9%) increase in melanoma mortality risk per 2000 hours of exposure to PCBs in the period 20+ years after first exposure. Exposure to PCBs probably does increase risk of melanoma.

Employment in the Oil Industry Several early studies suggested that melanoma incidence might be increased in oil industry employees. There is continuing evidence of this in further reports of two of them. In the Canadian petroleum industry workers followed to 1994, the standardized incidence ratio for melanoma was 1.32 (95% CI: 0.96 1.82) (Lewis et al., 2003) and in the Australian cohort followed to 1996, 1.54 (1.30–1.81) (Gun et al., 2004). In neither, however, was the standardized mortality ratio greater than 1.0 (Canadian 0.68, 0.01–1.34; Australian 0.82, 0.40–1.47) and in the Australian study the increased incidence was unrelated to duration of employment or estimated hydrocarbon exposure. These inconsistencies led the Australian authors to speculate that periodic, workplace medical examination and an ongoing skin awareness program had biased workforce melanoma incidence rates upwards relative to those in the general population.

Other Occupational Exposures Additional associations of melanoma incidence with employment in a range of industries or occupations or exposure to a range of physical and chemical agents in the occupational environment have been reported; some of which are moderately consistent and not easily explicable in terms of confounding with sun exposure. They include associations of melanoma with production and use of chemicals, work in the printing industry, work in the footwear and clothing industries, and work in the telecommunications industry. No specific causes for these associations have been established (Linet et al., 1995; Fritschi and Siemiatycki, 1996; Austin and Reynolds, 1997; Wennborg et al., 2001; Perez-Gomez et al., 2004).

1210

PART IV: CANCER BY TISSUE OF ORIGIN

Other Environmental Factors Reproduction and Reproductive Hormones A number of studies have suggested that, like breast cancer, risk of melanoma may fall with increasing parity and late age at first childbirth (Lambe et al., 1996). A recently completed meta analysis of original data from 10 case-control studies supported this suggestion: OR 0.95 (95% CI: 0.91–0.99; p for trend = 0.05) per birth and OR 0.98 (95% CI: 0.94–1.02; p for trend = 0.16) per year of age at first childbirth after adjustment for age at diagnosis of melanoma (Karagas et al., submitted for publication). Adjustment for sun exposure variables did not appreciably affect these results, though confounding by sun exposure might have been expected since frequent childbirth probably reduces women’s opportunity for sun exposure. If real, these effects of parity and age at first childbirth on melanoma risk remain to be explained. Some 18 studies of the association of oral contraceptive use and risk of melanoma in women were published from the initial report of this association in 1977 until 1996 and summarized in a meta-analysis of published results reported in 1997 (Pfahlberg et al., 1997). The pooled relative risk estimate was 0.95 (95% CI: 0.87–1.04) for ever use relative to never use of oral contraceptives. An original data metaanalysis of 10 case-control studies from the same period also showed no evidence that incidence of melanoma was increased in women who had used oral contraceptives for a year or more: OR 0.86 (95% CI: 0.74–1.01) (Karagas et al., 2002). From an analysis of the experience of the Nurses Health Study cohort, however, Feskanich et al. (1999) reported an association of current use of oral contraceptives with melanoma in premenopausal women: relative risk 2.0 (1.2–3.4). The relative risk was higher in women with 10+ years of use: 3.4 (1.7–7.0). The authors noted that there were indications of an effect of current use of oral contraceptives on risk of melanoma in some previous studies (although there was none in the original data meta-analysis referred to above) and that a combination of estrogen and progesterone can stimulate melanogenesis in experimental animals. Data on sun sensitivity and history of sunburn permitted some control of possible confounding with sun exposure in this study. The possibility that current oral contraceptive use increases risk of melanoma in women remains open. There is no consistent evidence that non-contraceptive use of estrogen by women alters risk of melanoma (Smith et al., 1998; Freedman et al., 2003). A summary relative risk estimate of 1.03 (95% CI: 0.87–1.21) for the highest category of estrogen use was calculated from a meta analysis of seven studies, including Smith et al. (1998), those it refers to, and two others (Persson et al., 1996; Freedman et al., 2003). Three studies were not included in this summary because confidence intervals were not reported for relative risk estimates of 1.0, 0.9, and 1.8; the last was based on eight cases of melanoma. It is unlikely that women’s use of non-contraceptive estrogen alters risk of melanoma. Three studies of melanoma risk in infertile women treated with fertility drugs (clomiphene citrate, human menopausal gonadotrophin, and human chorionic gonadotrophin) have given inconsistent results (Young et al., 2001).

Acquired Suppression of Immune Response Risk of melanoma has been reported to be 1.6 to 4 times higher in patients receiving post-transplant immunosuppressive therapy than it is in the corresponding general populations (Euvrard et al., 2003). There is weak evidence that incidence of melanoma is increased in people infected with HIV. Grulich et al. (2002) reported a standardized incidence ratio of 1.34 (95% CI: 0.93–1.86) for melanoma in a national, Australian cohort of 13,067 with HIV infection or AIDS, which is similar to that reported from a New York cohort of AIDS patients, 1.4 (0.89–2.06) (Gallagher et al., 2001). Homosexual people without AIDS in Denmark, however, had a similar small increase in risk of melanoma, 1.7 (0.5–4.3) (Frisch et al., 2003), thus suggesting that HIV infection could be confounded with sun exposure. There was no increase in risk in people notified to an Italian national AIDS registry: SIR 0.8 (95% CI: 0.2–2.4) (Dal Maso et al., 2003).

Infection and Immunization From an initial small, hospital-based case-control study, Kolmel et al. (1992) reported protective effects of a range of adult febrile infections against melanoma with ORs ranging from 0.21–0.52 for conditions as diverse as wound infections, abscesses and furunculosis, herpes simplex, influenza or the common cold, gastroenteritis, and a range of chronic infections. Risk was less after multiple infections than it was after a single infection. These findings prompted conduct of a multi-center case-control study in Europe and Israel, which examined the effects of immunizations as well as infectious diseases. This study found highly statistically significant protective effects of an increasing number of febrile infections in adulthood (Kolmel et al., 1999) and of immunization with BCG or vaccinia but not against influenza (Krone et al., 2003). The ORs for four or more febrile infections in different groups ranged from 0.18–0.55 and were 0.60 (95% CI: 0.36–0.99) for immunization with vaccinia only, 0.40 (0.18–0.85) for immunization with BCG only, and 0.41 (0.25–0.67) for immunization with both. There was limited evidence of synergy between infections and immunization: the OR for one or more infections and immunization with both BCG and vaccinia was 0.33 (0.17–0.65). Cases were obtained from hospital tumor boards (75% participation) and controls were randomly selected from houses in neighboring streets to houses of cases (84% participation). There was no mention of the interviewers being blind to the case or control status of the subjects or the study’s hypotheses. Data were collected on sun exposure and sun sensitivity and results were adjusted for age, sex, and “known risk factors”. These findings require independent replication.

Diet In a comparatively small number of studies of diet or diet-related variables and melanoma, there have been two reasonably consistent findings: increasing risk with increasing height (Shors et al., 2001) and increasing risk with increasing alcohol intake (Millen et al., 2004). The association with height might be explained by an association with skin surface area and thus the number of melanocytes that could undergo malignant transformation (Shors et al., 2001). To evaluate the association with alcohol intake further, a systematic search was made and 10 studies of alcohol intake and melanoma were found; they included that of Millen et al. (2004), those they refer to, and two others (Osterlind et al., 1988; Westerdahl et al., 1996). The OR for the highest category of alcohol intake was above one in all except three studies. The summary OR across all the studies was 1.16 (95% CI: 0.98–1.38). Thus, there is evidence of only a weak positive association between alcohol intake and melanoma, if any at all. Confounding with sun exposure could explain this weak association since in only four studies was adjustment made for some measure of sun exposure.

Injury About 20 cases of melanoma arising in burn scars have been reported worldwide; they suggest a link between skin injury and subsequent cutaneous melanoma (Hwang et al., 2004). Further evidence for such a link has been obtained in two case-control studies of melanoma of the palms or soles. Initially a small study of plantar melanoma in Paraguay found an OR of 40.9 (95% CI: 14.8–112.7) for past injury (burns, cuts, thorn pricks, and other) to the sole of the foot; walking barefoot, however, was not associated with any increase in risk (Rolon et al., 1997). Similar results were reported from a larger study of melanoma of the soles and palms in Australia and Scotland; the OR for penetrative injury was 5.0 (3.0–8.6) (Green et al., 1999). Wearing shoes was not asked about in this study; it was noted, though, that melanomas of the dorsa of the hands and feet were not associated with injury. These two studies are consistent with an association between injury and melanoma but the possibility of biased recall of injury cannot be excluded given the absence of any apparent protective effect of wearing shoes.

1211

Cutaneous and Ocular Melanoma

HOST FACTORS Pigmentary Characteristics Melanoma is primarily a disease of light-skinned populations, and pigmentary traits are known to influence the risk of melanoma. People with fair skin, blonde or red hair, blue eyes, and who sunburn easily are at particularly high risk. In addition, freckles and moles (nevi) are markers of risk. Distinguishing the independent contributions of these highly correlated characteristics is possible with careful measurement and analysis, and skin type and hair color are stronger predictors of risk than eye color. Pigmented lesions also confer different degrees of risk. Freckles are only modestly associated with risk of melanoma and probably represent confounding with sun exposure and sensitivity. Benign nevi are clearly independent risk factors, and confer an intermediate degree of risk in comparison with the higher risk associated with the presence of dysplastic or clinically atypical nevi. Quantifying the individual contributions of these pigmentary traits requires accurate measures of sun exposure, since sun exposure is unequivocally related to the presence of freckles and nevi as well as melanoma (Gallagher et al., 2000; Darlington et al., 2002; Milne et al., 2002).

Skin Type, Hair and Eye Color Studies of the considerable variation in pigmentation among populations of European origin have led to an appreciation of the gradient of risk that increases with lighter pigmentation. Numerous methods for measuring skin type and skin sensitivity have been used, including self-report, clinical classification systems, color-charts, and spectrophotometric measurements to estimate melanin density (Dwyer et al., 1998; Dwyer et al., 2002). The variety of methods makes it difficult to quantify risk across studies, but essentially all studies have demonstrated a consistent risk gradient, with the highest risk observed in individuals with the fairest skin. Skin sensitivity, as measured by ability to tan or sensitivity to sunburns shows similar gradients (Armstrong and Kricker, 2001; Fears et al., 2002). Hair color is a reproducible risk factor for melanoma, with blondes and redheads at higher risk than those with brown or black hair (Table 63–3). Hair color is partially determined by the amount and relative proportions of eumelanin (brown) and pheomelanin (yellow). The regulation of this ratio is controlled in large part by melanocyte stimulating hormone (MSH), acting through its receptor, melanocortin type 1 receptor (MC1R), to stimulate eumelanin production (Valverde et

al., 1995). The genetic contributions of the MC1R gene to melanoma are discussed in detail later in this chapter, but it is clear that hair color represents an important risk factor, and that the effects of MC1R are partially mediated by hair color, or vice versa. Thus, although it is straightforward to measure MC1R genotypes and it is technically feasible to directly measure melanin concentrations in hair (Zanetti et al., 2001), simple questions to assess natural hair color remain important in studies of the epidemiology of melanoma. Eye color is also an independent risk factor for melanoma, and individuals with green or blue eyes are 30%–50% more likely to develop melanoma than are people with dark eyes (Table 63–3). This phenotypic feature represents a modest risk factor for melanoma compared with other pigmentary characteristics, but recent studies have highlighted the importance of light eye color after consideration of pigmentary and genetic risk factors. The study by Matichard et al. (2004) reported that light eye color was associated with a 3.1-fold risk of melanoma (95% CI: 1.5–6.5) after adjustment for number of nevi, MC1R variations, and the presence of solar lentigines. Indeed, in this study, most of the variation in risk attributable to hair color appeared to be captured by MC1R, whereas light eye color was the only independent pigmentary risk factor. The genetics of eye color are not well understood, but it is clear MC1R does not contribute in a meaningful way to this trait. As with hair color, future epidemiologic studies will continue to depend on standard epidemiologic approaches to assessing eye color.

Freckles Freckling is consistently associated with risk of melanoma, but most of the studies that have evaluated freckles have not adequately adjusted for sun exposure. Freckles are a manifestation of inherent pigmentary traits, as well as exposure to the sun, so it is not entirely clear how much value early studies of freckles contribute to our understanding of the independent contributions of sun exposure to melanoma in comparison with pigmentary traits. The meta-analysis of Bliss et al. (1995) simultaneously examined skin color, freckle density, hair and eye color after adjustment for number of nevi, but did not adjust for measure of sun exposure measures. Most of the studies listed in Table 63–3 did not adjust for sun exposure. An important population-based study by White et al. (1994) found no association with freckles after adjustment for sun sensitivity, whereas the study of adolescent melanoma by Youl et al. (2002) reported that facial freckling was modestly associated with risk of melanoma after

Table 63–3. Relative Risks of Pigmentary Characteristics Associated with Melanoma Hair Color Study

Fair Skin

Blonde

Eye Color Red

Green

Blue

Freckles

2.2 3.6 2.4 2.4 0.4 1.9 1.8 2.8 2.9 7.8 2.4 (1.9–3.0)

1.5 2.2 1.1 0.7 1.3 1.6 0.8 1.2 1.6 1.6 1.3 (1.1–1.5)

1.6 1.9 1.4 0.9 1.5 1.5 1.1 1.6 1.6 2.1 1.5 (1.3–1.8)

— 2.0 6.9 — — 3.0 1.7 3.5 2.2 2.3 2.3 (2.1–2.6)

1.6† 5.4 NS

3.8 3.1

4.5 3.1

2.4† 3.2 —

studies with physical exams of nevi included in a 1995 meta-analysis Holman and Armstrong (1984) Green and MacLennan (1985) Elwood (1986) Swerdlow (1986) Cristofolini (1987) Holly (1987) Osterlind (1988) Elwood (1990) Walter (1990) Langholz (1995)† Bliss (1995) Meta-Analysis

2.8 2.0 — — 2.2 — 1.6 1.6 1.9 2.8 *

1.7 2.1 2.6 1.8 1.0 2.3 1.7 1.2 2.4 4.9 1.8 (1.5–2.1)

selected recent studies of pigmentary characteristics White (1994) Youl (2002)‡ Matichard (2004)**

9.0 4.7 NS

1.6† 1.7 NS

*Different methods for measuring adult skin color did not permit pooled analysis. † Not significantly associated with melanoma after adjustment of sun exposure. ‡ Adjusted for total nevi, hair color, eye color, tanning ability, facial freckling, family history, sunscreen use, lifetime number of peeling sunburns. **Odds ratios after multivariate adjustment, with nevus count, MC1R, light eye color, and solar lentigines (which are technically not freckles) included in model.

1212

PART IV: CANCER BY TISSUE OF ORIGIN

Table 63–4. Studies of Melanoma in Relationship to a Family History of Melanoma Study

Setting

Sample Size (Case/Control)

Relative Risk

95% Confidence Interval

P value

2.8 2.1 * 7.8 1.6 1.5 2.0 2.7 2.24 2.5

1.6–4.9 0.99–4.5 — 0.9–71.3 0.4–6.4 0.7–2.9 1.3–3.0 1.4–5.3 1.76–2.86 2.1–3.0

<0.001 <0.05 — — — — <0.001 <0.01 <0.001 —

4.0 5.0 2.41 1.5 2.0 3.2 4.4 1.7

0.8–18.9 2.5–7.9 2.10–2.76 1.1–1.8 1.6–2.5 2.6–3.9 3.5–5.2 1.1–2.7

0.03 — — — — — — —

studies included in a meta-analysis published in 1995 Holman (1984) Green (1985) Swerdlow (1986) Cristofolini (1987) Holly (1987) Osterlind (1988) Walter (1990) Langholz (1995)† Ford (1995) Peto (2001)

Australia Australia Scotland Italy California Denmark Canada California Meta-Analysis Meta-Analysis

511 / 511 232 / 232 180 / 197 103 / 205 121 / 139 474 / 926 583 / 608 748 / 800 2952 / 3618

additional recent studies of familial aggregation of melanoma Youl (2002) Freedman (2003) Hemminki (2003) Begg (2004) Begg (2004) Begg (2004) Begg (2004) Rutter (2004)

Australia (youth) USA rad. cohort Sweden Australia (male) Australia (female) N. America (male) N. America (female) USA

201 / 205 207 in 68,588 24,818 cases 70 relatives 76 relatives 88 relatives 100 relatives 737 / 1021

*Relative risk not estimated, with 3% of cases reporting family history, 0% of controls. † Previously unpublished data presented in meta-analysis by Ford et al.

adjustment for other pigmentary factors, tanning ability, and lifetime number of peeling sunburns (OR = 3.2, 95% CI: 0.9–12.3).

Nevi The evidence that nevi are an important host factor for melanoma is summarized above. Common acquired nevi, congenital nevi, and atypical nevi are all risk factors for melanoma, and in some settings are direct precursors.

GENETIC SUSCEPTIBILITY Familial aggregation of melanoma was first reported in 1820 (Norris, 1820), and many epidemiologic studies have shown that a family history is associated with an increased risk of melanoma (Table 63–4) (Holman and Armstrong, 1984; Green et al., 1985; Swerdlow et al., 1986; Cristofolini et al., 1987; Holly et al., 1987; Osterlind et al., 1988; Walter et al., 1990). An original data meta-analysis from 1995 summarized the results from eight case-control studies (Ford et al., 1995). In this meta-analysis of 2952 cases and 3618 controls from predominantly white populations, the relative risk of melanoma associated with a family history of melanoma was 2.24 (1.76–2.86). The effect of family history was independent of age, number of nevi, hair and eye color, and freckling, and there was no evidence of heterogeneity by study site, despite broad variability in latitude and sun exposure. The relative risk was slightly higher for affected siblings than parents or offspring, but this difference was trivial and these data have been interpreted as most consistent with dominant or polygenic models. Peto and Houlston (2001) also calculated a pooled estimate of published risks in first-degree relatives of individuals with melanoma from a selected set of large studies given in Table 63–4. Studies of familial aggregation of melanoma before the initial metaanalysis by Ford et al. included a large number of case reports as well as a survey conducted in Australia (Wallace et al., 1971). These early studies are discussed in a comprehensive review by Greene and Fraumeni (1979), who also calculated a relative risk of 1.7 from the Australian survey data. Recent studies have taken advantage of a number of different designs to estimate the risk of melanoma among family members (Table 63–4). The population-based case-control study by Youl et al. (2002) describes risk factors for melanoma in adolescents, aged 15–19 years. In this study, the unadjusted matched odds ratio for a family history of melanoma was 3.8 (1.6–9.4, P < 0.001)

suggesting a stronger effect of family history among adolescents than has generally been reported among adult populations in Australia. These data are consistent with the concept that young age of onset is more strongly associated with familial forms of cancer (Knudson, 1971; Claus et al., 1990), although after adjustment for phenotypic risk factors the confidence interval for a family history of melanoma includes 1. The US Radiologic Technologists (USRT) study provides a valuable estimate of the strength of the association between family history of melanoma in first-degree relatives and melanoma within a prospective cohort study in the US population, since cohort studies are less susceptible to the recall bias that can accompany case-control studies (Freedman et al., 2003). The Genes, Environment, and Melanoma (GEM) study estimated familial aggregation using a population-based, kin-cohort design that compares the ratio of the standardized incidence rate in relatives to population-standardized rates (Begg et al., 2004). Unlike the meta-analysis by Ford et al., this international study of a similar size noted less prominent familial aggregation in Australian centers than in North American centers, which may be related to the much higher incidence rates in Australia. Despite the limitation that the GEM study did not have pathologic confirmation of diagnoses or risk factor data for relatives, the magnitude of familial aggregation reported in this study is generally consistent with other estimates of familial risk. Data from the Swedish Family Cancer Database of over 10 million people provided precise estimates of familial aggregation using data from linked records of incident melanoma within 3 million families (Hemminki et al., 2003). In addition to the standardized incidence ratio for offspring shown in Table 63–4, the authors also report a similar SIR when only the parent had melanoma (SIR = 2.98, 95% CI: 2.54–3.47), and a substantially higher SIR of 8.92 (4.25–15.31) when both a parent and a sibling were affected. These data are consistent with the existence of both low-penetrant dominant heritable mechanisms in addition to a smaller proportion of high penetrant susceptibility genes that is likely to account for families with an affected proband, parent, and sibling. Other studies not summarized in Table 63–4 also provide important insights into familial aggregation of melanoma, including studies focusing on heterogeneity of melanoma risk in families of melanoma patients. The most important of these, the Queensland Familial Melanoma Project, is a population-based family study of 15,907 individuals ascertained through 2118 melanoma cases diagnosed between 1982 and 1990 (Aitken et al., 1996). An analysis of a subset of these published in 1994 restricted consideration to verified

1213

Cutaneous and Ocular Melanoma cases of melanoma in relatives, and found evidence of familial heterogeneity in the risk of melanoma (Aitken et al., 1994). Approximately 5% of families had significantly more melanoma than expected by chance, and members of the high-risk families were more sensitive to sun exposure, and were more likely to have fair skin color, red hair, and multiple nevi. A similar approach was taken in a multi-center, clinic-based case-control study that is summarized in Table 63–4 (Rutter et al., 2004). Significant heterogeneity for the risk of familial melanoma was also noted in this study, with slightly more than 8% of families expressing melanoma than would be expected by chance. The study by Rutter et al. (2004) also examined heterogeneity with respect to other cancers that have been reported in association with respect to melanoma, such as pancreatic cancer, brain cancer, and others, but no familial heterogeneity was detected for these other tumors with melanoma. Of note, the proportion of controls in this clinic-based study who reported a family history of melanoma was higher than in most population-based studies, and it is likely that some degree of selection bias is relevant in this study. Another important hospitalbased case-control study not shown in Table 63–4 is the 1989 study of 280 matched pairs of melanoma and hospital controls from Scotland. Although this study did not specifically examine familial heterogeneity, the odds ratio for a family history of melanoma can be calculated from the presented data, with an odds ratio = 5.7 (1.2–25.9) (MacKie et al., 1989). Although selection bias is also likely to be operating in this hospital based-case control study, it too provides additional information regarding family history and phenotypic characteristics associated with risk of melanoma. Most family studies in the epidemiologic literature have relied on reported histories of melanoma in relatives, which is clearly susceptible to measurement error. Indeed, methodologic studies have shown that family histories of melanoma are not highly accurate (Aitken et al., 1996; Weinstock and Brodsky 1998), at least in comparison with reported family history of other malignancies (Love et al., 1985). Confirming reported family histories of melanoma with pathology records is ideal, despite the logistical difficulties. However, family histories

that restrict consideration to first-degree relatives are more accurate than for more distant relatives (Ziogas and Anton-Culver, 2003), and relative risks measured among first-degree relatives tend to be consistent in the literature, with summary estimates of relative risk generally between 2 and 3. In addition to familial aggregation, melanoma has been well documented within the context of several specific genetic syndromes. The dysplastic nevus syndrome, also called the BK-mole syndrome or FAMMM (familial atypical mole-malignant melanoma) syndrome, is characterized by autosomal dominant inheritance of susceptibility to melanoma and multiple dysplastic or clinically atypical moles (Clark et al., 1978; Greene et al., 1978; Lynch et al., 1978). More than two decades worth of clinical and genetic studies of familial melanoma have led to the characterization of four putative forms of familial cutaneous melanoma based on evidence of distinct chromosomal linkage: CMM1 (1p36), CMM2 (unequivocally caused by mutations in the CDKN2A gene on 9p21), CMM3 (caused by mutations in CDK4 on 12q14 in a small number of families), and CMM4 (1p22). Although the history of this genetic research was initially characterized by sometimes acrimonious debate about whether atypical nevi represent a phenotype that is valuable for linkage studies, now there is evidence that dysplastic nevi are a risk factor for melanoma that appears to be partially independent of mutation status in families carrying a known mutation (Goldstein et al., 2000). However, there is no question that families with dysplastic nevi and FAMMM represent a clinically recognizable phenotype that provides an opportunity for early detection and intervention. Indeed, long-term clinical care of more than 30 families followed at the National Cancer Institute has unequivocally demonstrated the value of enhanced surveillance for melanoma in these kindreds (Tucker et al., 1993; Tucker et al., 2002). Other specific genetic syndromes are associated with a high risk of melanoma (Table 63–5), and melanoma is one of the cancers known to be associated with germline mutations in each of three classes of cancer susceptibility genes: tumor suppressor genes, oncogenes, and

Table 63–5. Syndromes and Other Cancers Associated with Risk of Melanoma Syndrome

Inheritance

Gene (locus)

CMM1 CMM2

Aut. Dominant Aut. Dominant

unknown (1p36) CDKN2A (9p21)

CMM3 CMM4 Li-Fraumeni

Aut. Dominant Aut. Dominant Aut. Dominant

Xeroderma Pigmentosum

Aut. Recessive

CDK4 (12q14) unknown (1p22) TP53 (17p13) CHK2 (22q12) XP (9q22, 2q21, 3p25, 19q13, 13q33, others)

Bloom Werner

Aut Recessive Aut Recessive

BLM (15q26) WRN (8p12)

Class Tumor suppressor Oncogene Tumor suppressor Tumor suppressor DNA repair DNA repair DNA repair

Notes p16 and p14 ARF both cause familial melanoma 3 families reported

Multiple complementation groups. Classic XP associated with very high risk of melanoma Especially high risk of melanoma noted in Japan

other cancers associated with melanoma Cancer

Inheritance

Pancreatic Cancer Astrocytoma–Neural tumors

Aut. Dominant

Breast

Aut. Dominant and Complex

Sarcoma

Complex

Leukemia Lymphoma

Complex

Head and neck Gastrointestinal

Complex Complex

Aut. Dominant

Gene (locus) CDKN2A (9p21) accounts for some of the association CDKN2A (9p21) accounts for some of the association CDKN2A (9p21) BRCA2 (13q12) TP53 (17p13) TP53 (17p13) CHK2 (22q12) others? TP53 (17p13) CDKN2A (9p21) others? CDKN2A? (9p21) CDKN2A? (9p21)

Evidence

Notes

Strong Strong, especially deletions Modest Modest Family, multiple primary studies, mouse models Modest Modest Weak

Inactive p14 ARF correlates with neural tumor risk Li-Fraumeni Syndrome Known association with Li-Fraumeni Syndrome Known association with Li-Fraumeni Syndrome, strong lab data

1214

PART IV: CANCER BY TISSUE OF ORIGIN

DNA repair genes. Rare syndromes like Bloom syndrome (German, 1997; German and Ellis, 2002), Xeroderma Pigmentosum (Bader et al., 1985; English and Swerdlow, 1987; Kraemer et al., 1987), and Werner syndrome (Lynch et al., 1983; Goto et al., 1996) are all accompanied by a very high risk of melanoma, in addition to a range of other cancers. Li-Fraumeni syndrome (LFS), also sometimes described as SBLA syndrome, is classically characterized by early onset of a variety of cancers including sarcoma (S), breast, brain (B), lung, leukemia (L), and adrenocortical (A) carcinoma. LFS is due to in large part to mutations in the TP53 gene, and rarely in CHK2. However, melanoma is also a recognized component of a variant of LFS called Li-Fraumeni-like syndrome, which is also caused by mutations in TP53 and CHK2. (Hartley et al., 1989; Garber et al., 1990). Each of these classic syndromes is extremely rare, accounting for a miniscule fraction of the public health burden of melanoma (Debniak et al., 2004). The association between melanoma and other cancers is also recognized in other contexts. Familial melanoma and pancreatic cancer is well characterized (as described in Online Mendelian Inheritance in Man, OMIM #606719) (Whelan et al., 1995; Vasen et al., 2000), and is related to mutations in CDKN2A in some families (Gruis et al., 1995; Borg et al., 1996; Ciotti et al., 1996; Soufir et al., 1998; Ghiorzo et al., 1999; Liu et al., 1999; Borg et al., 2000; Lynch et al., 2002). Similarly, the strong association between melanoma and astrocytoma or other primary brain cancer is well described (OMIM #155755) (Kaufman et al., 1993; Tachibana et al., 2000). Other cancers that appear to be over-represented in melanoma patients and their family members are also summarized in Table 63–5. Sarcoma, as noted above, is a cardinal feature of Li-Fraumeni syndrome, but the reported associations of sarcoma and melanoma extend beyond the limited setting of this rare genetic syndrome (Noonan et al., 2001; Lynch et al., 2002). Breast cancer also appears to be associated with cutaneous melanoma in several settings, including increased risk among carriers of mutations in BRCA2 (Consortium, 1999) as well as in Li-Fraumeni syndrome. Data from the Breast Cancer Linkage Consortium estimated the risk of melanoma in BRCA2 carriers as RR = 2.58 (95% CI: 1.28–5.17), although these highly selected families might lead to overestimates of relative risk. Other evidence is also consistent with increased risk of melanoma in association with breast cancer, including studies of the offspring of breast cancer cases (Anderson et al., 2000), studies of multiple primary melanoma and breast cancer (Gutman et al., 1991), and characterization of families with mutations of CDKN2A that are associated with increased risk of breast cancer (Ghiorzo et al., 1999; Borg et al., 2000). However, it is important to note that the epidemiologic evidence for some of the associations reported in Table 63–5 is weak and requires further study. In particular, the data regarding associations between head and neck cancer and melanoma offer some interesting clues (Spitz et al., 1990; Puig et al., 1997; Sun et al., 1997), as are the associations between melanoma and lymphoma/leukemia (Okuda et al., 1995; Soufir et al., 2002) but these association are not yet well characterized. Similarly, evidence for an association between gastrointestinal cancers and melanoma is still largely anecdotal (Lynch et al., 1978; Grammatico et al., 2001; Lynch et al., 2002), with some support

from studies of multiple primary cancers (Hoar et al., 1985; Tucker et al., 1985).

Susceptibility Genes The genetic epidemiology of melanoma has focused on highpenetrance genes responsible for autosomal dominant forms of hereditary melanoma, as well as a variety of candidate low-penetrance susceptibility genes that are likely to contribute to risk among a larger fraction of sporadic melanoma. CDKN2A is the gene that warrants greatest consideration because this gene accounts for the largest proportion of hereditary melanoma identified to date. Although a detailed review of the molecular genetics of melanoma susceptibility genes is beyond the scope of this chapter, several aspects are relevant to understanding the molecular epidemiology of the disease (reviewed by Hayward) (Cannon-Albright et al., 1992; Cannon-Albright et al., 1994; Hussussian et al., 1994; Kamb et al., 1994; Zuo et al., 1996; Hayward, 2003). CDKN2A encodes two different proteins, p16INK4A (commonly called p16) and p14ARF, through an alternative splicing mechanism. This relatively small gene is organized into three exons encoding each protein. The first exon of each protein is unique (exons 1b and 1a, respectively), but exons 2 and 3 are shared by each protein. The two proteins also differ by misalignment of the reading frames that are translated from the shared sequence, and therefore CDKN2A encodes two entirely different proteins. This is relevant to the genetic epidemiology of melanoma, since a mutation in CDKN2A can alter the function of p16, p14ARF, or both, depending on the location of the mutation. Furthermore, it appears that the risk of cancer may differ depending on which protein is inactivated. One might postulate that differences are related to the different functions of the two proteins, since p16 is a tumor suppressor that regulates the cell cycle transition from G1 to S phase, and p14ARF is a tumor suppressor that stabilizes p53. Mutations in CDKN2A are identified in 5% of families with two relatives with melanoma and 20%–40% of families with three or more affected family members (Platz et al., 1997; Kefford et al., 1999). So far, it has been difficult to estimate the prevalence of mutations in the general population, and it is highly likely that these frequencies are population-dependent. Data from the Queensland Familial Melanoma Project estimated that 0.2% of all melanoma cases in Queensland harbor a mutation in CDKN2A, based on sequencing the gene in members of families at high risk and testing for common known mutations in individuals at low- and intermediate-familial risk. The international Genes, Environment, and Melanoma study found a higher prevalence than this in invasive melanoma cases from North America, Australia, and Italy, with pathogenic mutations identified in 1.3% of single primary melanoma cases. These data suggest a population prevalence of 0.3% among individuals in the population without melanoma (Berwick, 2005). In addition, founder mutations have been described in a number of settings, as shown in Table 63–6 (Gruis et al., 1995; Borg et al., 1996; Platz et al., 1997; MacKie et al., 1998; Pollock et al., 1998; Liu et al., 1999; Ciotti et al., 2000; Auroy et al., 2001). The lifetime probability of developing melanoma (penetrance) among carriers of CDKN2A is difficult to estimate precisely for

Table 63–6. Founder Mutations in CDKN2A Population

Mutation

Frequency

Dutch

“p-16 Leiden”

Swedish

113insArg

Scottish / Irish / English British (Canada)

M53I

Italian / French

G101W

4/16 kindreds 25% 4/23 kindreds 17% 7% French MPM

German/English (US)

V126D

common

G-34T

13/15 kindreds 87% 8% families

Age (generations) — ~98 (52–167) — — ~100 (70–133) 34–52

References Gruis et al., 1995 Borg et al., 1996, Platz et al., 1997 MacKie et al., 1998, Pollock et al., 1998 Liu et al., 1999 Ciotti et al., 2000 Auroy et al., 2001 Goldstein et al., 2001

Cutaneous and Ocular Melanoma several reasons. Families that are ascertained for linkage studies are generally selected based on the large number of melanoma cases within families, and correcting for the ascertainment bias is not easy when all of the contributing factors that led to ascertainment are not understood. Initial estimates of penetrance were based on small numbers of families, with a 53% risk of developing melanoma by age 80 in a study of three families demonstrating linkage to 9p (CannonAlbright et al., 1994). In a similarly sized study of CDKN2A mutation carriers, Newton Bishop (2000) calculated penetrance to be 64% by age 85 with a 1-LOD support interval of 22%–98%. The Melanoma Genetics Consortium estimated penetrance in 80 families, with a lifetime risk of melanoma of 67% by age 80 (95% CI: 0.31–0.96) (Bishop et al., 2002). In contrast, the Genes, Environment, and Melanoma study has recently estimated CDKN2A penetrance at 24% by age 80 (95% CI: 13%–34%) from population-based ascertainment of CDKN2A carriers (Begg et al., 2005). It is highly likely that there are phenotypic differences, including differences in penetrance, based on the type and location of mutations within CDKN2A, and further studies are required to clarify these distinctions in diverse populations. Indeed, mutations that have preferential impact on p14ARF are more likely to be associated with neural tumors (Hayward, 2003), which is consistent with evidence from deletions that span this region as well (Petty et al., 1993b; Hayward, 2003). Germline mutations in CDK4 are extremely rare, but do cause hereditary melanoma; however, mutations in CDK4 have only been reported in three families worldwide (Zuo et al., 1996; Soufir et al., 1998). Thus it is difficult to accurately estimate penetrance. CDK4 and CDKN2A directly interact, and mutations in either gene affect regulation of Rb phosphorylation and cell cycle transitions. Two of the three reported families with mutations in CDK4 share the same activating R24C mutation (Zuo et al., 1996), while the third family has a different missense mutation at the same codon, R24H (Soufir et al., 1998). In the family reported by Soufir et al., three of eight mutation carriers had developed melanoma at the time of publication, consistent with incomplete penetrance. A separate comparison of the clinical phenotypes of CDKN2A and CDK4 mutation carriers found no differences in age at first melanoma, number of melanomas, or number of nevi (Goldstein et al., 2000), as would be expected from the similarity of the molecular effects of mutations in these two genes. Low-penetrance susceptibility genes have also received considerable attention since the melanocortin type 1 receptor, MC1R, was first shown to be associated with red hair and fair skin (Valverde et al., 1995). Subsequent studies clarified the relationship between variants in MC1R and red hair, fair skin, and freckling (Box et al., 1997; Palmer et al., 2000; Bastiaens et al., 2001; Kanetsky et al., 2004), and an initial study by Valverde et al. (1996) also found that variants in MC1R were related to risk of melanoma, with an OR = 3.9 (1.5–10.3). Numerous studies have confirmed the association between polymorphic variants in MC1R and melanoma (Palmer et al., 2000; Kennedy et al., 2001; Dwyer et al., 2004; Matichard et al., 2004; Pastorino et al., 2004), with analyses consistently demonstrating that relative risk increases with increasing number of variant alleles. Few studies have attempted to discern whether risk associated with MC1R is independent of pigmentary characteristics (Dwyer et al., 2004), but emerging data suggest that MC1R does capture information beyond skin type and hair and eye color (Begg et al., 2004; Matichard et al., 2004). There is also evidence that variants in MC1R modify the risk of melanoma among individuals who carry CDKN2A mutations (Box et al., 2001; van der Velden et al., 2001). This is one of the few known examples of gene-gene interactions in human disease, and two groups have now shown that the risk of melanoma is higher among CDKN2A carriers who also have one or two copies of a functional MC1R variant than those with two wild- type alleles. Other suspected low-penetrance susceptibility alleles have not yet been well characterized, and some of these candidate genes have recently been excluded as markers of risk. Initial evidence that a functional variant of the epidermal growth factor (EGF) gene is associated with risk of developing melanoma (Shahbazi et al., 2002) has not been supported in larger studies (McCarron et al., 2003; Amend et al., 2004;

1215

Randerson-Moor et al., 2004), although this variant might be associated with melanoma depth. Other candidate susceptibility genes have been considered as well, including the Vitamin D receptor gene (VDR), a DNA repair gene partially responsible for correcting damage from ionizing and UV radiation (XRCC3), as well as phase-I (CYP2D6) and phase-II metabolic enzymes (GSTM1, GSTT1). The early data from a polymorphic variant of VDR are particularly intriguing given the role of VDR in melanocyte biology. In an initial case-control study (Hutchinson et al., 2000), a polymorphism of VDR that leads to a novel translation start site with reduced transcriptional activation of downstream targets was associated with increased risk of melanoma. Although the data from this clinic-based case-control study are reported describing odds ratios using the variant as the referent category (OR = 0.59, p = 0.029 after adjustment for age and sex), the data are perhaps easier to interpret by considering the functional variant as associated with increased risk (OR = 1.69) for the presence of one or two copies of the minor allele. A later, but smaller, case-control study by the same group identified another variant in the promoter region of VDR associated with increased risk of melanoma, with an OR = 2.5, 95% CI: 1.1–5.7 for heterozygotes, OR = 3.3, 95% CI: 1.4–8.1 for homozygotes (Halsall et al., 2004). Additional studies are required to elucidate the role of VDR polymorphisms in melanoma. The data regarding other candidate susceptibility genes is not yet convincing. A variant allele of XRCC3 was reported as a risk factor for melanoma in a population from the United Kingdom, with an OR = 2.4, P = 0.004 (Winsey et al., 2000). In contrast, two other groups did not find evidence that this allele was associated with risk of melanoma (Duan et al., 2002; Bertram et al., 2004). Similarly, the evidence for a role of metabolic enzymes in the pathogenesis of melanoma is not consistent. The data for CYP2D6 initially showed associations between poor metabolizers and risk of melanoma (Wolf et al., 1992; Strange et al., 1999), whereas another study found evidence against this hypothesis (Dolzan et al., 1995). Three studies from the United Kingdom (Heagerty et al., 1994), Australia (Shanley et al., 1995), and the United States (Kanetsky et al., 2001) found no differences in the frequency of GSTM1 null alleles in melanoma cases and controls, whereas one study from Barcelona found that GSTM1 null genotypes were associated with risk of melanoma (Lafuente et al., 1995). Interestingly, Kanetsky et al. (2001) found no overall effect for GSTM1 null genotypes, but did show increased risk among those with fair hair who were null for GSTM1 (OR = 2.2, 95% CI: 1.2–4.2) and those without functional GSTM1 or GSTT1 (OR = 9.5, 95% CI: 1.2–73.0). In summary, the only consistently identified low-penetrance susceptibility allele for melanoma that has been identified to date is MC1R, and further studies of candidate susceptibility alleles will need to consider the functional consequences and relevance to melanoma biology, in addition to careful replication in different populations.

PATHOGENESIS Integrating the evidence from histopathology, epidemiology, and molecular biology leads to a clear picture that melanoma arises through a complex series of genetic events that regulate melanocyte biology, and that solar radiation is the principal cause of cutaneous melanoma. Laboratory investigations have begun to outline the pathways that are involved in the development and progression of melanoma, but it is not clear how solar radiation mediates this dysregulation of melanocytic growth in most circumstances. Classic ultravioletinduced DNA damage, such as pyrimidine dimers of TP53 that are commonly observed in non-melanoma skin cancers, are not usually identified within primary melanomas. As noted, there is not a perfect correspondence between body sites achieving the most intense sun exposure and location of melanoma. Yet the weight of the evidence, despite minor gaps, is consistent with the unifying theme that solar radiation mediates the risk of melanoma by influencing key regulatory pathways. The best description of this hypothesis was summarized by Gilchrest et al. (1999), who contrasted the epidemiologic features of

1216

PART IV: CANCER BY TISSUE OF ORIGIN Table 63–7. Selected Clinical Trials of Educational Interventions to Increase Sun Protection Study

Setting

Sample Size

Gallagher, 2000 Milne, 1999 2002 Buller, 1999 Crane, 1999

British Columbia Australia

458 1776 children

US US

162 students 27 preschools

GISED, 2003 Glanz, 2000 Weinstock, 2002 Dietrich, 2000 Lowe, 2004

Italy US (Hawaii) US (Rhode Island) US Australia

4233 children 383 children 2324 1930 children 3790 adults

photocarcinogenesis in basal cell and squamous cell carcinoma to melanoma. Noting that melanin is typically observed in supranuclear “caps” that protect nuclei from photodamage, in conjunction with the limited capacity of melanocytes to proliferate in comparison to keratinocytes, this hypothesis suggests that UV radiation exerts different effects in keratinocytes and melanocytes that depend on dose and timing of the exposure. Under this hypothesis, keratinocytes that are severely damaged by UV radiation undergo apoptosis, leaving cells with minimal damage to gradually accumulate incremental mutational damage with each subsequent exposure. This suggests that intermittent, high-dose exposures to UV radiation would be far less important than more continuous, low-dose exposure to the development of basal cell and squamous cell carcinoma. In contrast, melanocytes are less susceptible to apoptosis, despite a heavy burden of mutations from an initial high dose of UV radiation. These cells that are more extensively damaged are more likely to incorporate mutations and clonally expand. Melanocytes are also more vulnerable to damage with intermittent UV exposure when the skin has not yet induced its baseline capacity for DNA repair or up-regulated its melanin content. This hypothesis offers a context for understanding the epidemiologic associations between intermittent, high-dose exposures and the risk of melanoma, in distinction to the cumulative effects of frequent, lowdose exposures over a lifetime that are associated with basal cell and squamous cell carcinoma. As an example of the genetic dysregulation that is particularly well understood, the defective nucleotide excision repair that characterizes Xeroderma Pigmentosum provides compelling evidence that links solar radiation to melanoma through specific genetic pathways that enhance sensitivity to UV damage. It is also clear from several lines of evidence that different populations harbor different risks of melanoma. These differences are often recognizable phenotypically, especially with respect to pigmentary characteristics. In addition, over the past decade, major genetic loci have been characterized that account for a fraction of the heterogeneity in melanoma risk, and further studies of the variation within the human genome are likely to lead to new insights regarding the stratification of risk in human populations.

PREVENTION Primary Prevention In principle, melanoma is largely preventable. It is difficult to quantify how much reduction in melanoma incidence and mortality could be anticipated by altering sun exposure patterns in at-risk individuals, but several studies of nevi as intermediate biomarkers have examined how sun intervention programs might succeed in attenuating risk. As previously noted a randomized clinical trial of sunscreen in 458 schoolchildren in British Columbia found that sunscreen use led to 30%–40% fewer nevi in the treated group. A strong interaction with freckling was also noted, where the use of sunscreen was more important for children with freckles than those without (Gallagher et al., 2000).

Intervention

Outcome

30%–40% Ø # of new nevi Ø Sun exposure, 3%–11% Ø in # of new nevi School CD-ROM ≠ knowledge Child-Care No change in parents, but ≠ knowledge center directors School Ongoing Recreation sites ≠ Sun protection Beaches ≠ Sun protection Towns ≠ Sunscreen use 18 Communities ≠ Awareness School School

Other randomized clinical trials have evaluated educational interventions designed to increase sun protection (Table 63–7), with modest success. School and daycare-based programs (Buller et al., 1999; Crane et al., 1999; Milne et al., 1999; Milne et al., 2002; GISED, 2003) showed improvement in some parameters of sun behavior or knowledge. The “Kidskin” study in Western Australia also showed a reduction in the rate of development of new nevi but this change was not statistically significant (Milne et al., 2002). Increased use of sunscreens was noted in a multi-component intervention trial of paired towns, and a randomized clinical trial of beachgoers showed impressive, sustained increases in sun-protective behaviors over 24 months of follow-up (Weinstock et al., 2002). Community trials have also been modestly successful in increasing the prevalence of sun-protective behaviors in localities randomized to interventions (Dietrich et al., 2000; Lowe et al., 2004). At an individual level, recommendations for primary prevention should continue to emphasize the avoidance of sunburns and minimizing ultraviolet radiation exposure from any source. Sunbeds or sunlamps used to acquire or maintain a tan should be avoided. Primary prevention is likely to have the most important effect in children, and thus recommendations to minimize UV radiation, while appropriate for individuals of all ages, should be emphasized in children and adolescents.

Secondary Prevention and Chemoprevention In addition to strategies that focus on primary prevention by improving sun behaviors, secondary prevention by early detection has been studied extensively. Skin self-examination (SSE) is a valid technique for detecting phenotypic markers of increased risk (Gruber et al., 1993) and can be enhanced with baseline skin photography (Oliveria et al., 2004). SSE is currently advocated by the American Cancer Society, many professional dermatology associations, and is clinically recommended to individuals with atypical nevi or a prior history of melanoma. However, SSE is not endorsed by the US Preventive Services Task Force due to the paucity of evidence from interventional studies. Testing the hypothesis that SSE reduces the incidence or mortality from melanoma is difficult due to lead time bias, but one population-based case-control study specifically designed to address this question found that SSE was associated with a reduced risk for melanoma (OR = 0.66, 95% CI: 0.44–0.99) (Berwick et al., 1996). This study also indirectly estimated that SSE may reduce the mortality from melanoma by 63% (adjusted OR = 0.37, 95% CI: 0.16–0.84). Regularly scheduled clinical exams are unequivocally valuable in the prevention of thick melanoma within the setting of FAMM (Tucker et al., 2002), and there is evidence that physician-detected melanomas are thinner than melanomas detected by other approaches (Epstein et al., 1999). Secondary prevention in high-risk populations remains an attractive approach for future studies of melanoma, especially since very thin melanomas are likely to be cured with surgery alone. Chemoprevention for melanoma is attractive in principle, but is still in its infancy. One might effectively argue that sunscreens are a form of chemoprevention, but most authorities would consider sunscreen a

Cutaneous and Ocular Melanoma physical barrier that merely enhances avoidance of direct solar radiation. Here we will briefly consider pharmaceuticals or nutraceuticals that have been suggested for potential consideration in melanoma. A comprehensive review is beyond the scope of this discussion, but in brief, several classes of compounds have been suggested as relevant for the chemoprevention of melanoma based on experimental and epidemiologic evidence, as well as early results from clinical trials (Demierre and Nathanson, 2003). Several clinical trials have investigated the effect of topical tretinoin on atypical nevi as a surrogate biomarker for melanoma. First suggested as a potential chemopreventive strategy by Meyskens and Edwards in 1986, tretinoin has been evaluated in two small, randomized clinical trials that both showed beneficial changes in the clinical and histologic appearance of atypical nevi, and indeed some nevi disappeared with treatment (Halpern et al., 1994; Stam-Posthuma et al., 1998). However, the topical treatment was not well tolerated due to irritation in the first trial, and the clinical management of these lesions seemed unlikely to change based on treatment in the second. Statins are effective lipid-lowering agents that inhibit 3-hydroxy-3methylglutaryl coenzyme A (HMG-CoA) reductase and inhibit invasion and metastasis of melanoma cell lines (Collisson et al., 2003). Statins prevent the prenylation of Ras and Rho, thus interrupting the required membrane interactions that facilitate oncogenic signaling. In addition, statins have anti-inflammatory properties that have been suggested as potentially important in chemoprevention of melanoma and other cancers. The observational data in humans are equivocal, with one nested case-control study finding a statistically significant protective association (OR = 0.50, 95% CI: 0.37–0.68) (Dellavalle et al., 2003), while an analysis of the General Practice Research Database in the United Kingdom found an RR = 2.5, 95% CI: 0.8–7.3 (Kaye and Jick, 2004). Analyses of follow-up data from some, but not all cardiovascular clinical trials of statins suggest a potential protective benefit of this class of drugs for melanoma. At this point in time, statins represent an intriguing hypothesis for future studies, but the evidence is still quite preliminary. Anti-inflammatory agents that inhibit cyclo-oxygenase 2 (COX-2) as well as agents that inhibit the lipoxygenase pathway have been suggested as candidates for chemoprevention of melanoma. COX-2 is highly expressed in melanoma cell lines, and in a hairless mouse model a COX-2 inhibitor extended the tumor latency period and reduced the total number of UV-induced melanomas. One casecontrol study has been published examining common non-steroidal anti-inflammatory drugs such as aspirin and ibuprofen and melanoma, observing a significant protective association (OR = 0.45, 95% CI: 0.22–0.92) (Harris et al., 2001). Nutritional components and foods have been evaluated as potential nutraceutical chemopreventive agents for melanoma, but no promising leads have emerged to date. Specifically, diet, retinol, alphatocopherol, lycopene, alpha-carotene, and beta-carotene do not appear to be significantly associated with risk of melanoma (reviewed in (Demierre and Nathanson, 2003). Preventing melanoma will require a combination of approaches that relies on an increasing understanding of epidemiologic risk factors like sun exposure, genetic susceptibility, and the successful implementation of primary and secondary prevention.

FUTURE DIRECTIONS Solar UV radiation is the principal avoidable cause of melanoma, and it is clear that translating this knowledge into action is likely to reduce the incidence and mortality from cutaneous melanoma. But how? Clearly further behavioral research will help identify the best ways to communicate this information to the public, and additional clinical trials at the community level are likely to help guide these approaches. Part of the challenge of communicating the risks associated with solar UV radiation is that many of the mechanisms and details seem still to be obscured from view, and epidemiologic studies and experimental models are still required to yield new insights. Epidemiologic studies have moved towards using a more uniform set of tools for

1217

measuring solar UV radiation, and it would be valuable to incorporate these validated instruments into large prospective cohort studies that are organized to study risk factors for cancer or other chronic diseases. By minimizing the misclassification associated with lifetime recall of recreational and occupational sun exposure one could estimate the effects more precisely, and understand the age dependence and types of exposures that lead to melanoma in far greater detail than is currently known. Genetic susceptibility is a major area of investigation on the horizon. The study of high-risk families has led to the identification of a small number of genes responsible for autosomal dominant forms of cutaneous melanoma, but the enormous diversity of genetic variation that contributes to melanoma susceptibility is not yet understood. Whole genome association studies are now feasible, and large, methodologically sound case-control studies in different populations will provide unique insights into the plethora of genes that contribute to pigmentary phenotype, sun sensitivity, and risk of melanoma. As a cautionary note, the availability of high-throughput genotyping does not guarantee success, and it is critically important to measure phenotype with care and accuracy. In particular, these studies will require careful consideration of solar UV radiation exposure and other risk factors, since the epidemiologic evidence already suggests complex relationships between, age, gender, and sun exposure. Genetic studies are also likely to provide insights into the population stratification that characterizes patterns of melanoma throughout the world. Founder mutations already explain measurable fractions of melanoma in specific populations, and it is likely that further research will identify susceptibility alleles in some populations and protective alleles in others. Yet simplistic approaches to “race” or “ethnicity” are unlikely to yield meaningful public health approaches to melanoma, and one must be vigilant about the potential for improper uses of this information. Integrated approaches to epidemiology and basic research should also be encouraged. As an example, the systematic collection of frozen tumors in future studies has enormous potential to help distinguish the pathways and mechanisms relevant in pathogenesis. Integrating epidemiologic risk factors into these studies will permit a more detailed understanding of the factors that influence the transition from normal skin to melanoma.

OCULAR MELANOMA Melanocytes that give risk to ocular melanomas lie principally in pigmentary epithelia that extend continuously from the choroid, a vascular membrane that separates the neural elements of the retina from the fibrous outer covering of the eye (the sclera), anteriorly to the ciliary body and iris. The choroid, ciliary body, and iris are collectively termed the uvea; hence the commonly used term “uveal melanoma.” This term, however, is not synonymous with ocular melanoma because a small proportion arises in melanocytes in the conjunctiva, the minimally pigmented epithelium that covers the visible sclera at the front of the eye. Ocular melanoma is globally much less significant than cutaneous melanoma and shows little current evidence of uptrend in incidence, in contrast to the continuing uptrend in melanoma. Although no accurate estimate has been made, 7 million is a rough estimate of the annual number of ocular melanomas globally if they are assumed to be 4.3% of the number of cutaneous melanomas globally (as in the SEER 12 registries [Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database: Incidence— SEER 11 Regs + AK Public-Use, Nov 2003 Sub (1973–2001 varying), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2004, based on the November 2003 submission]. It is, though, a more fatal disease, on average, than cutaneous melanoma; the estimated 5-year disease-specific survivals were 78.7% for ocular melanoma and 87.1% for cutaneous melanoma in cases registered in the 9 SEER registries from 1973–2001 (Statistics generated using: Surveillance Research Program, National Cancer Institute SEER*Stat software

1218

PART IV: CANCER BY TISSUE OF ORIGIN

(www.seer.cancer.gov/seerstat) version 5.3.1. and Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database: Incidence—SEER 9 Regs Public-Use, Nov 2003 Sub (1973–2001), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2004, based on the November 2003 submission). Exposure to solar and to manmade UV radiation are the only well-established risk factors for ocular melanoma. Probably the best prospects for prevention lie with effective protection of the eyes against UV radiation exposure in the course of occupational and recreational exposure to the sun.

CLASSIFICATION Anatomic Distribution Like cutaneous melanoma, ocular melanoma is heterogeneous with respect to site. Most lesions are choroidal (63% in the 9 US SEER registry populations in 1974–1998); about 17% are in the ciliary body or the iris; 5% are conjunctival; and some 15% were of unknown location in that population (Inskip et al., 2003). At a more detailed level, ocular melanomas are most frequently located in the temporal half of the posterior choroid and the inferior and temporal iris (Horn et al., 1994; Vajdic et al., 2001).

Histopathology Intraocular melanomas are classified histopathologically into three categories: Spindle cell—slender cells with a thin oval nucleus and an indistinct nucleolus (Spindle A) or larger plumper nucleus with sharply defined, round nucleoli (Spindle B); Epithelioid cell—larger, more pleomorphic, polygonal cells with large, sometimes multiple, nucleoli and; Mixed cell—both spindle and epithelioid cells are present (Albert and Syed, 2001). Spindle cell melanomas have the most favorable prognosis and epithelioid cell melanomas the least favorable prognosis. Iris melanomas are more likely than choroid or ciliary body melanomas to be of spindle cell type and have a better prognosis. Other histological characteristics that predict outcome of intraocular melanomas include size of tumour, vascular pattern, and frequency of mitoses (Mudhar et al., 2004). Conjunctival melanomas are described as having more diverse cell types: small polygonal, spindle, balloon, and epithelioid. As with intraocular melanomas, the spindle cell lesions have the best prognosis (Seregard, 1998; Anastassiou et al., 2002). Depth of invasion is the other major, independent predictor of outcome.

Precursor Neoplastic Lesions Conjunctival melanoma can arise in conjunctival melanocytic nevi and in primary acquired melanosis (PAM). The latter is a “flat and variably brown unilateral lesion, usually presenting in middle-aged or elderly whites” (Seregard, 1998). About 75% of conjunctival melanomas are thought to originate in PAM and about 20% in conjunctival nevi. It is commonly believed that most intraocular melanomas arise in intraocular melanocytic nevi, although apparently with little or no empirical foundation. In a recent study there was little difference in prevalence of intraocular nevi in 65 patients with posterior choroidal melanomas (18.5% had nevi) and 218 control, eye clinic patients (17.4% had nevi) (Harbour et al., 2004). While it suggests that risk of choroidal melanoma is not influenced by presence of nevi, this finding does not directly negate the possibility that some intraocular melanomas arise in nevi. In addition, prevalence of nevi was estimated by counting lesions in both eyes; it is possible that the melanoma in one eye obscured or obliterated as many as half the nevi in cases. Whether intraocular nevi are precursors to ocular melanoma, therefore, remains moot. Choroidal melanoma is associated with a much rarer melanocytic condition, ocular melanocytosis or oculodermal melanocytosis (nevus

of Ota), in which there are episcleral, sclera, or uveal collections of melanocytes with or without a similar cutaneous lesion in the distribution of the trigeminal nerve (usually the ophthalmic and maxillary divisions) (Baroody and Holds, 2004). The lifetime risk of intraocular melanoma in patients with ocular or oculodermal melanocytosis has been estimated at 1 in 400 compared with 1 in 13,000 in a general Caucasian population (Singh et al., 1998).

Molecular Genetic Characteristics of Tumor Abnormalities of chromosomes 3, 8q, and 6p have been commonly found in ocular melanoma (Vajdic et al., 2003). In addition, loss of heterozygosity on 9p in ocular melanomas has suggested the possibility of a tumor suppressor gene in that location (Speicher et al., 1994), and specific loss of heterozygosity, homozygous deletion, or inactivation through hypermethylation of the promoter of CDKN2A, which is located at 9p21, have been found in some cases (Edmunds et al., 2002). However, only one pathogenic mutation of CDKN2A has been found in an ocular melanoma (Hearle et al., 2003). Increased immunostaining of p53 is observed in some ocular melanomas, although not commonly, and TP53 gene mutation appears to be comparatively rare (Chowers et al., 2002). No N-RAS or BRAF mutations have been found in uveal melanomas (Cruz et al., 2003; Kilic et al., 2004) or in conjunctival melanomas (El-Shabrawi et al., 1999). Whether any molecular change is etiologically relevant has not been established. Chromosome 3 and 8q changes, however, do appear to predict outcome. Gene expression analysis of tissue samples from fresh primary uveal melanomas showed tight clustering of the cancers into two groups designated as class 1 and class 2. They were distinguished mainly by statistically significant gene clusters on chromosomes 3 and 8q that were respectively down-regulated and up-regulated in class 2 tumors (Onken et al., 2004). Class 1 tumors corresponded to lower-grade, spindle cell melanomas and had a 95% survival to 92 months while class 2 tumors were higher-grade, epithelioid cell melanomas and had a 31% survival to 92 months (P < 0.01 for survival difference between classes). Survival, however, was not adjusted for age, which was strongly related to the gene expression class (class 1 mean age 56 years, class 2 mean age 73 years).

DEMOGRAPHIC PATTERNS Mortality and Incidence In the period 1973–2001, the annual age-standardized incidence of ocular melanoma (US 2000 Standard Million standard) in the 9 US SEER registry populations averaged 0.6 per 100,000. It was 0.7 in whites, 0.1 in blacks, and 0.1 in other races. In all men it was 0.7 and in all women, 0.5; the corresponding figures for whites only were 0.8 in men and 0.6 in women (Statistics generated using: Surveillance Research Program, National Cancer Institute SEER*Stat software (www.seer.cancer.gov/seerstat) version 5.3.1. and Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database: Incidence—SEER 9 Regs Public-Use, Nov 2003 Sub (1973–2001), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2004, based on the November 2003 submission). Malignant neoplasms of the eye are generally reported as a group in national mortality statistics. While some 80% of them are ocular melanomas, there is limited value in using patterns of mortality from this heterogenous group as a surrogate for patterns of mortality from ocular melanoma.

Time Trends Age-standardized incidence of ocular melanoma fell by 1.0% a year (95% CI: 0.5%–1.5%) in the 9 US SEER registry populations from 1973–2001. This downtrend was due mostly to a 0.8% a year fall

Cutaneous and Ocular Melanoma

at about the same rate in women as men; thereafter it increases at a greater rate in men than women to a peak at 80 years of age. In England, for which there was the largest number of cases of any population in Cancer Incidence in Five Continents Volume VIII (Parkin et al., 2002), the sex ratio in those 70+ years of age was 1.46.

1.2 Male Female

Rate per 100,000

1

1219

0.8

Ethnicity 0.6

0.4

0.2

2001

1999

1997

1995

1993

1991

1989

1987

1985

1983

1981

1979

1977

1975

1973

0 Calendar year

Figure 63–14. Trends in incidence of ocular melanoma in white men and women in the SEER 9 Registries, US (Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database: Incidence—SEER 9 Regs Public-Use, Nov 2003 Sub (1973–2001), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2004, based on the November 2003 submission). Standard population was US 2000 Standard Million standard.

in white males and a 0.9% a year fall in white females (Fig. 63–14). The fall in whites was greater in younger people than in older people, greater in ciliary body and iris melanomas (together) and melanomas of other and unspecified sites in the eye (except conjuctiva) than in choroidal lesions, which showed little net downtrend (Inskip et al., 2003). Incidence of melanomas of the conjunctiva increased by 5.5% a year in men (95% CI: 0.6%–12%), but fell a little in women (1.2% a year, 95% CI: -7.2%–5.1%.

Survival The 5-year relative survival was 78.7% (95% CI: 74.9%–82.5%) in all ocular melanomas diagnosed in 1973–2001 in the 9 US SEER registry populations. Survival was slightly greater in women, at 79.2%, than in men, at 78.3%, but much higher in those aged less than 55 years, at 86.9% (84.3%–91.5%), than in those 55 years or older, at 74.0% (68.6%–79.4%). There was no indication of any trend towards increasing or decreasing 5-year survival in those diagnosed up to 1997 (Statistics generated using: Surveillance Research Program, National Cancer Institute SEER*Stat software (www.seer.cancer.gov/seerstat) version 5.3.1. and Surveillance, Epidemiology, and End Results (SEER) Program (www.seer.cancer.gov) SEER*Stat Database: Incidence—SEER 9 Regs Public-Use, Nov 2003 Sub (1973–2001), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2004, based on the November 2003 submission).

Sex The age-adjusted incidence of ocular melanoma in men was equal to or greater than that in women in 81% of some 100 cancer registry populations with a total of 10 or more cases of ocular melanoma in 1993–1997 (Parkin et al., 2002). The overall ratio of incidence in males to that in females was 1.32. This excess in men appears to apply to melanoma at each ocular subsite (Inskip et al., 2003).

Age Ocular melanoma begins to appear, but rarely, in teenage, and increases in incidence with age from the early 20s until the late 40s

Patterns of incidence among different countries (see below) suggest that ocular melanoma incidence rates are higher in fair-skinned people of mainly European ethnic origin than in dark-skinned or Asian origin populations. In the multiracial US SEER registry populations, the age-standardized incidence (World standard population) in whites in 1993–1997 was 0.33/100,000 and in blacks was 0.03 (based on 5 cases). In the Hispanic white population of Los Angeles County, the rate was 0.21 (Parkin et al., 2002). A similar gradient with skin color is suggested by incidence rates of posterior uveal melanoma in Israeli Jews of different origins. In 1972–1996 those who migrated to Israel from Europe or America had three times the age-standardized rate of those migrating from Africa or Asia: 6.8 (95% CI: 4.9–8.7) compared with 2.3 (0.7–3.9) per million per year (Iscovich et al., 2001). A similar difference was seen between people of Southern European and other European origin in Australia in 1996–1998 (Vajdic et al., 2003).

International Patterns In populations of sufficient size and incidence to have had 10 or more incident cases in the 5 years from 1993 to 1997, the highest ageadjusted (world standard) incidence rates in both sexes were in the mainly European origin populations of the Australian Capital Territory (0.96/100,000), Denmark (0.73), and New Zealand (0.66) (Parkin et al., 2002). The lowest rate in any population of mainly European origin was in Croatia (0.13). Only one non-European population, Israel, fell within this range with a rate of 0.22, but this was heavily determined by a rate of 0.31 in Jews born in Europe and America, who contributed two-thirds of all cases. Other non-European populations with more than 10 cases had much lower rates: Miyagi Prefecture, Japan 0.06/100,000; Taiwan 0.05; and Seoul, Korea 0.03. Ignoring the numbers of cases, and thus populations, no non-European population had a rate greater than 0.14/100,000. There is little evidence that the incidence of ocular melanoma as a whole is increasing in European origin populations. An analysis of uveal melanoma in the Finnish cancer registry suggested that its incidence was stable from 1953–1973 (Raivio, 1977); incidence of ocular melanoma as a whole has remained stable in the United States since at least 1974 (Inskip et al., 2003); and there is little evidence of any net trend in Denmark or Australia in the 1990s (Johansen et al., 2002; Vajdic et al., 2003). Conjunctival melanoma, however, is probably increasing in incidence as reports from the United States and Finland suggest, though possibly only in men (Tuomaala and Kivela 2003; Yu et al., 2003). Incidence of uveal melanoma may be falling in some countries, as reported from Sweden where incidence in both males and females fell by about 25% from 1960–1964 to 1985–1989 and then stabilized, or rose slightly in males to 1995–1998 (Bergman et al., 2002). There is little evidence from descriptive studies that ocular melanoma incidence varies consistently with latitude in populations of European origin (Vajdic et al., 2003).

Migration There is little evidence that risk of ocular melanoma increases on migration from an area of low to one of high ambient solar UV radiation. The age-adjusted incidence of ocular melanoma in 1972–1996 in Jews born in Israel of parents who immigrated from Europe or America, 7.0 (95% CI: 5.0–9.6) per million per year, was little different from that over the same period in the corresponding immigrating generations, 6.8 (4.9–8.7) (Iscovich et al., 2001). Similarly, while relative to people born in Australia and New Zealand, the standardized incidence ratio in 1996–1998 in Australia for ocular melanoma in

1220

PART IV: CANCER BY TISSUE OF ORIGIN

people born in Europe as a whole was 0.77 (95% CI: 0.63–0.95), that in those born in Northern Europe was 0.86 (0.69–1.08) (Vajdic et al., 2003). While routinely collected data do not permit distinction of Australian and New Zealand born people in Australia by birthplace of their parents, it is probable that the vast majority of those developing ocular melanoma in 1996–1998 would have had their origins in Northern Europe; substantial migration from Southern Europe and non-European countries began only in the second half of the 20th century.

Socioeconomic Status Risk of ocular melanoma appears to be unrelated to socioeconomic status, whether assessed from education or occupation (Raivio, 1977; Tucker et al., 1986; Holly et al., 1996; Stang et al., 2003).

HOST FACTORS Pigmentary and Sun Sensitivity Characteristics Six of seven relevant case-control studies have shown crude, twofold or higher risks of ocular melanoma in people with blue or grey eyes compared with brown eyes; in four of these studies, this association was statistically significant (Vajdic et al., 2001; Stang et al., 2003). In five of these studies, the increased risk remained twofold or higher after adjustment for a range of potentially confounding factors, and in four it was still statistically significant. Risk was also increased with green or hazel eyes in some studies (Vajdic et al., 2001; Stang et al., 2003). The evidence that other pigmentary or sun sensitivity characteristics are associated with ocular melanoma is weaker and less consistent, though there is moderately strong evidence that lighter skin color and poorer ability to tan, but not higher propensity to sunburn, are associated with higher risk (Vajdic et al., 2001; Stang et al., 2003). These pigmentary and sun sensitivity characteristics are confounded with one another and with ethnicity, thus making it difficult to allocate primacy of effect to any one. It has been suggested that the density of choroidal pigmentation may be the primary factor, since it is greater in people with darker eye color and could, perhaps, protect against ocular melanoma through absorption of UV radiation incident on the choroid (Vajdic et al., 2001). A recent case-control study, however, found a statistically significant positive association between denser, visually assessed choroidal pigmentation and uveal melanoma (Harbour et al., 2004). While the observer was not blind to the case or control status of subjects in this study, there was a good correlation between the subjective pigment density and actual density of choroidal melanocytes as assessed by immunostaining of seven enucleated eyes. Pigment in retinal epithelial cells, however, was similar between subjective pigment density categories.

Cutaneous and Ocular Nevi and Freckles and Cutaneous Melanoma Higher numbers of cutaneous pigmented nevi have been consistently shown to be associated with higher risk of ocular melanoma (Vajdic et al., 2001; Richtig et al., 2004). Presence of atypical nevi appears to present a greater risk than that of banal nevi only. The evidence that risk of ocular melanoma is increased in the presence of iris or choroidal nevi is conflicting (Harbour et al., 2004). There is only weak evidence that cutaneous freckles are associated with ocular melanoma (Vajdic et al., 2001). A significantly higher risk of cutaneous melanoma has been observed following a diagnosis of ocular melanoma in a large study based on the US SEER registry database for 1973–1998—standardized incidence ratio (SIR) of 4.6, 95% CI: 2.9–6.8 (Shors et al., 2002). Risk of ocular melanoma, however, was not significantly increased following a diagnosis of cutaneous melanoma—SIR 1.4, 95% CI: 0.5–3.0. These findings were not inconsistent with those of several previous smaller studies, including a recent case-control study in which risk was increased only moderately with a recalled history of cutaneous melanoma (OR 1.8, 95% CI: 1.0–3.1 for choroid and ciliary body melanoma) (Vajdic et al., 2002).

Genetic Susceptibility Familial coincidence of ocular melanoma has been estimated to occur more frequently than would be expected by chance, though this cancer has been rarely reported in more than two members of one family (Singh et al., 1996; Kodjikian et al., 2003). Ocular melanoma has also been reported in families with familial breast and ovarian cancer, and BRCA2 mutations have been reported, perhaps more frequently than expected by chance, in patients with ocular melanoma (Sinilnikova et al., 1999; Iscovich et al., 2002; Scott et al., 2002). Data from the Swedish Family Cancer Database also provide modest evidence of an association between breast cancer and ocular melanoma, although more work is needed to clarify this relationship (Hemminki and Jiang, 2001). In no case has a pathogenic mutation been found in CDKN2A that could explain familial ocular melanoma (Canning and Hungerford, 1988; Singh and Donoso, 1993; Hearle et al., 2003). However, in the study by Hearle et al., one CDKN2A mutation carrier (without a family history of ocular or cutaneous melanoma) was identified in their series of 385 patients with uveal melanoma. MC1R has also been examined as a candidate gene in ocular melanoma, because of its relationship to sun sensitivity phenotype, but no significant relationship has been observed with any of its polymorphisms that have been studied (Hearle et al., 2003; Vajdic et al., 2003). Though instances of conjunctival and iris melanoma have been reported in people with xeroderma pigmentosum, most ocular neoplasms in these people are ocular surface keratinocyte carcinomas (Johnson et al., 1989; Aoyagi et al., 1993; Singh et al., 2004),

ENVIRONMENTAL FACTORS Sun Exposure Sun exposure could be the main cause of ocular melanoma. The evidence for and against this statement is summarized below.

All Ocular Sites of Melanoma Are Exposed to Solar UV Radiation Ocular exposure to solar UV radiation when outdoors is estimated to vary between 2% and 17% of ambient levels depending on the incidence angle of the sun, reflection from surrounding surfaces and use of a hat, spectacles, or sunglasses (Singh et al., 2004). Most UV radiation reaching the eye is absorbed by the cornea, lens, and retinal pigment epithelium before it reaches choroidal melanocytes. Experimental studies show that approximately 4% of it reaches the retina in early childhood; this proportion falls with increasing age to less than 1% of radiation below 340 nm and 2% of radiation between 340 and 360 nm. Transmission falls to negligible levels after about 25 years of age (Vajdic et al., 2002).

Uveal Melanomas Occur at Subsites in the Eye of Greatest Exposure to Solar Radiation The temporal half of the posterior choroid and the inferior and temporal iris (Horn et al., 1994; Vajdic et al., 2001), where most melanomas of these sites occur, are also the subsites of greatest exposure to solar radiation (Vajdic et al., 2002).

People with Phenotypes Carrying Increased Susceptibility to Cutaneous Melanoma Also Have Increased Susceptibility to Ocular Melanoma The people at highest risk of cutaneous melanoma and ocular melanoma are those of fair-skinned, European origin who have lightcolored eyes or sun-sensitive skin and higher numbers of cutaneous, melanocytic nevi. People who have had an ocular melanoma are at increased risk for cutaneous melanoma, although the converse may not be true.

There Is Only Weak Evidence for an Association between Incidence of Ocular Melanoma and Either Latitude or Ambient UV Radiation In addition to the lack of evidence for a latitude gradient in descriptive studies, no association was found for an association between ocular melanoma and an aggregate measure of lifetime ambient UV

Cutaneous and Ocular Melanoma radiation at all places of residence (Vajdic et al., 2002). Some US studies, however, have found some, but not consistent, evidence for an increased risk with birth or long-term residence in the southern United States (Vajdic et al., 2002). This is consonant with a somewhat lower risk in migrants from Europe to Australia than in people born in Australia and New Zealand, (Vajdic et al., 2001) but the same pattern is not evident in Israel (Iscovich et al., 2001).

There Is Very Weak Evidence that Higher Total Lifetime Sun Exposure Is Associated with a Higher Risk of Ocular Melanoma; There Is Stronger Evidence, Though, for an Association with Occupational Sun Exposure Four case-control studies have reported on the association of estimated total, lifetime sun exposure and ocular melanoma; in only one study was the OR in the highest exposure category appreciably different from 1.0 (Vajdic et al., 2002). The summary OR for the highest categories of exposure in three studies—the fourth did not report a numerical result—was 1.2 (95% CI: 0.9–1.7). Three case-control studies have reported on the association of ocular melanoma with estimated lifetime or near lifetime sun exposure in all occupations. One reported “no association”, in another the OR in the highest exposure category was not above 1.0, and in the third the risk in the highest exposure category was significantly increased (OR 1.8, 95% CI: 1.1–2.8) (Vajdic et al., 2002). The summary OR for the highest exposure categories in the two studies that reported numerical results was 1.5 (95% CI: 1.0–2.3). Three other case-control studies have reported positive associations of specific categories of outdoor work—agriculture, forestry, and fishery workers; sailors, ship officers or fishermen; and construction workers—with ocular melanoma, which were statistically significant in two of the three studies (Vajdic et al., 2002). Results from studies of eye cancer (mostly ocular melanoma) and ocular melanoma in farmers, both descriptive and analytical, have produced mixed results (English et al., 1997; Vajdic et al., 2002). Apart from weakly positive and marginally statistically significant associations of three specific recreational outdoor activities (gardening, vacations, and sunbathing) with ocular melanoma in one study (Tucker et al., 1985), there is little evidence that recreational or nonworking day sun exposure as a whole is associated with this cancer (Vajdic et al., 2002).

There Is Weak Evidence that Protection against Ocular Sun Exposure Reduces Risk of Ocular Melanoma There is evidence in three of four relevant case-control studies that the wearing of a hat, sunglasses, or corrective lenses may reduce risk of ocular melanoma (Tucker et al., 1985; Pane and Hirst, 2000; Vajdic et al., 2002). Across the three studies that reported numerical results, the summary OR for any use of corrective lenses was 0.8 (95% CI: 0.6–1.0); the fourth did not report numerical results. Tucker and colleagues (1985) also found evidence of a protective effect of use of sunglasses, hats, or sun visors; the OR for ocular melanoma in those who never, rarely, or only occasionally used them was 1.6 (95% CI: 1.2–2.2) but no association was found with use of sunglasses in two other relevant studies nor with hat wearing in one. The indirect evidence from phenotypic characteristics indicating sensitivity to solar UV radiation is the strongest evidence that sun exposure causes ocular melanoma. The mechanism relating these characteristics to ocular melanoma, however, is less intuitively obvious than it is for cutaneous melanoma. While pale-colored eyes may correlate with less-dense retinal pigment epithelia (Vajdic et al., 2001), this is disputed (Harbour et al., 2004). The lack of any clear correlation between latitude or ambient UV irradiance and incidence of ocular melanoma is puzzling, and could be taken as clear evidence against a causal effect of solar UV radiation on this cancer. This lack, however, could be explained by the fact that substantial ocular UV exposure comes from the horizon sky (Sliney, 1995), where ambient UV irradiance is much less affected by latitude. The direct evidence from studies of personal sun exposure and ocular protective behaviors is weak and inconsistent. The strongest evidence, from occupational

1221

sun exposure in men in Australia with a 2.5-fold increase in risk with increasing exposure, may not be persuasive to many because of the lack of a corresponding association in women (Vajdic et al., 2002). Women do, however, have much less occupational sun exposure than men and there is evidence that they report it less accurately than do men (Vajdic et al., 2002). There is no consistent balance of opinion for or against sun exposure as a cause of ocular melanoma. Commonly held views are probably best summarized by this heavily hedged opinion: “. . . it can be concluded that the current data are incomplete and conflicting, but rudimentary case-control study results suggest that a potential role of sunlight exposure in pathogenesis of uveal melanoma cannot be completely discounted and that further investigations are necessary” (Singh et al., 2004).

Manmade Sources of UV Radiation There are two comparatively common means of exposure to manmade sources of UV radiation—use of sunlamps, sunbeds, and tanning booths and electrical arc welding. Four studies have estimated risks of ocular melanoma following exposure to UV emitting lamps (three referred specifically to sunlamps, sunbeds, or tanning booths while one analyzed exposure to “UV or black lights” excluding sun exposure). In all four, the OR for the highest level of exposure was increased twofold or more; the summary OR across them was 3.0 (95% CI: 1.8–5.0) (Vajdic et al., 2004). Six studies have estimated risks following arc-welding exposure; in all six the OR for the highest level of exposure was greater than 1.0. The summary OR across the studies was 2.2 (95% CI: 1.6–3.1) with substantial heterogeneity between them (ORs ranged from 1.3–11.5) (Vajdic et al., 2004). The evidence is thus reasonably strong that exposure to manmade UV radiation is a cause of ocular melanoma. This finding adds plausibility to the case that exposure to solar UV radiation increases risk of this cancer.

Other Environmental Agents Radiofrequency Electromagnetic Energy In two separate case-control studies of uveal melanoma in Germany, one population-based and the other hospital-based, Stang et al. (2001) estimated occupational exposure to radiofrequency electromagnetic energy by asking subjects: “Did you use radio sets, mobile phones or similar devices at your work place for at least several hours per day?” If they answered yes, the specific exposures were identified and duration of exposure to each was estimated. Both studies showed evidence of increased risks of uveal melanoma with exposure to radio sets (walkie talkies and vehicle-mounted two-way radios) and use of mobile phones. In a pooled analysis, the OR for any use of radio sets was 3.3 (95% CI: 1.2–9.2) and for any use of mobile phones, 2.8 (95% CI: 1.0–7.9). Neither showed increasing risk across two ordered categories of exposure. Control for socioeconomic status did not alter the results. In neither Denmark nor the US SEER registry areas was there any indication of an uptrend in incidence of ocular melanoma in the 1990s that could correspond to the effects of increasing use of mobile phones (Johansen et al., 2002; Inskip et al., 2003). Nor was there any suggestion of a differential trend in incidence between the right and the left eyes; available evidence suggests that more people usually hold a mobile phone to their right ear than to their left ear (Inskip et al., 2003).

Other Occupational Exposures Occupational sun exposure and occupational exposure to manmade sources of UV radiation have been dealt with above. A number of ecological and analytical (case-control) studies have reported on risk of ocular melanoma in other categories of occupation or occupational exposure. A few such associations have been found to be statistically significant in more than one study: chemical industry employment (Albert et al., 1980; Holly et al., 1996; Monarrez-Espino et al., 2002), employment in administration and management (Swerdlow 1983; Gallagher et al., 1985; Vagero et al., 1990), and employment as a cook

1222

PART IV: CANCER BY TISSUE OF ORIGIN

or kitchen hand (Vagero et al., 1990; Monarrez-Espino et al., 2002). No agent has been identified with any certainty to be specifically associated with ocular melanoma in any of these studies. In one study of the association with cooking, risk was higher in those who actually cooked as part of their job and showed a clear trend to increase with increasing length of employment in cooking (Stang et al., 2003).

PATHOGENESIS Notwithstanding the evidence, from epidemiology and location within the eye, that exposure to the sun or to manmade sources of UV radiation may increase risk of ocular melanoma, there is little molecular evidence that links UVB radiation to this cancer. TP53 and N-RAS or BRAF gene mutations appear rare in this cancer whereas they are common, respectively, in keratinocyte cancer and cutaneous melanoma and often due, in these cancers, to gene changes that are plausibly related to UVB. The absence of any clear increase in risk of ocular melanoma in people with xeroderma pigmentosum also argues against a direct mutagenic role for UVB radiation in increasing risk of this cancer. There is little or no pathogenetic evidence for or against an etiological role for UVA or longer wavelength solar radiation. Nor are there any pathogenetic mechanisms that would suggest a link with any other etiological agent.

PREVENTION Primary Prevention The strongest and best-established modifiable risk factors for ocular melanoma are use of sunlamps in their various forms and arc welding. In both cases, the simplest and most direct method of preventing ocular melanoma is protection of the eyes with UV-absorbing lenses. Their use should be required, and application of this requirement enforced, in all commercial artificial tanning operations, and they should be supplied and strongly recommended to consumers who purchase their own artificial tanning equipment. In arc welding, the use of a wrap around mask and highly absorbing lens is routine. A recent Canadian survey, however, showed that 9% of workers’ compensation claims from welders were for “welders’ flash” and that welders’ flash was a major health concern raised by welders (Korczynski, 2000). It was also commonly observed that welders were not wearing their helmets for quick spot-welds and that many workplaces lacked non-reflective surfaces or partitions to reduce bystander exposure. Thus, greater attention to established safety practices is required and knowledge that arc exposure may cause ocular melanoma, in addition to acute ocular problems, may help to motivate compliance with them. It would be prudent also to recommend that outdoor workers wear ocular protection from UV radiation. This protection is best provided by close-fitting sunglasses with UV-absorbing lenses that wrap around the temples and wide-brimmed headwear (Rosenthal et al., 1988; Rosenthal et al., 1988), although direct evidence for their effectiveness in reducing risk is very limited.

Chemoprevention and Secondary Prevention There is no evidence to suggest, at present, that any chemopreventive agent or program of screening for early detection of the primary cancer would be likely to reduce incidence or improve outcome of ocular melanoma. Broader tumor diameter predicts poorer survival following local radiotherapy to the eye or removal of the eye for ocular melanoma (Isager et al., 2004). This would suggest that earlier diagnosis would improve outcome, however, systematic screening of the eyes for early lesions would almost certainly not be cost-effective.

FUTURE DIRECTIONS Present knowledge of the epidemiology and prevention of ocular melanoma is limited. We are uncertain whether solar UV radiation is

causal, although more certain about manmade UV radiation as a cause. While European ethnicity and light eye color are associated with highest risk of melanoma, there is only a speculative mechanism for the latter association and the genetic basis is unknown. In addition, while there is rare but definite familial clustering of this cancer, no genetic basis has been established and variants of only one gene, BRCA2, appear to affect ocular melanoma risk. How can we achieve more certainty? The answer lies in better methods, particularly for measuring lifetime sun exposure, greater statistical power, and more comprehensive epidemiological studies. While methods for eliciting recall of sun exposure have improved since the earliest studies were done, they can probably be improved further, particularly in the direction of being equally accurate in men and women. Less accuracy in women than in men may have made the most recent case-control study unpersuasive (Vajdic et al., 2002). A large, probably internationally collaborative case-control study would increase the range of sun exposure and diversity of exposure experience and ethnic background. It would permit examination of risk at different points along the dose continuum, and permit detailed examination of interactive effects of ethnicity, eye color, and sun exposure. This might assist in producing a more coherent pattern of relationship between sun exposure and ocular melanoma, particularly between men and women. In addition, such a study could provide statistical power to test hypotheses of interaction between sun exposure and genetic determinants of eye color or other markers of susceptibility to sun exposure effects other than MC1R. There is a range of genes, which have recently been identified as influencing eye color, including MATP and OCA2, both genes associated with oculocutaneous albinism, and ASIP and the P genes that would merit consideration (Rebbeck et al., 2002; Sturm and Frudakis 2004; Zeigler-Johnson et al., 2004; Graf et al., 2005). The systematic collection in such a study of fresh frozen tumor tissue, although now available in a minority of cases because of the extensive use of brachytherapy for treatment, would also create a valuable resource. It would permit correlation of sun exposure and other putative environmental risk factors with chromosomal abnormality and gene mutation and expression, and the identification of possible pathogenetic pathways and “signature mutations.” Such a study could be designed with sufficient statistical power to address main effects of BRCA2 mutations adequately and might yield families in which more than one member has had ocular melanoma for further intensive investigation by means of case-control family study methods. It would thus have the capacity to address most of the areas of present uncertainty summarized above. References Aitken JF, Duffy DL, et al. 1994. Heterogeneity of melanoma risk in families of melanoma patients. Am J Epidemiol 140(11):961–973. Aitken JF, Green AC, et al. 1996. The Queensland Familial Melanoma Project: Study design and characteristics of participants. Melanoma Res 6(2):155–165. Aitken JF, Youl P, et al. 1996. Accuracy of case-reported family history of melanoma in Queensland, Australia. Melanoma Res 6(4):313–317. Albert D, Syed N. 2001. Protocol for the examination of specimens from patients with uveal melanoma: A basis for checklists. Arch Pathol Lab Med 125(9):1177–1182. Albert DM, Puliafito CA, et al. 1980. Increased incidence of choroidal malignant melanoma occurring in a single population of chemical workers. Am J Ophthalmol 89(3):323–337. Amend KL, Elder JT, et al. 2004. EGF gene polymorphism and the risk of incident primary melanoma. Cancer Res 64(8):2668–2672. American Cancer Society. 2005. Cancer Facts and Figures 2005. Atlanta, American Cancer Society. Anastassiou G, Heiligenhaus A, et al. 2002. Prognostic value of clinical and histopathological parameters in conjunctival melanomas: A retrospective study. Br J Ophthalmol 86(2):163–167. Anderson H, Bladstrom A, et al. 2000. Familial breast and ovarian cancer: A Swedish population-based register study. Am J Epidemiol 152(12): 1154–1163. Aoyagi M, Morishima N, et al. 1993. Conjunctival malignant melanoma with xeroderma pigmentosum. Ophthalmologica 206(3):162–167. Armstrong B. 2004a. Epidemiology of Melanoma and Current Trends. Thompson JFM, Kroon DL, Londa BR, eds. Martin Dunitz.

Cutaneous and Ocular Melanoma Armstrong B. 2004b. How does sun exposure cause skin cancer-epidemiological perspective. In: Hill D, English DR, Elwood JM, eds. Prevention of Skin Cancer. New York, Kluwer Academic: pp. 89–116. Armstrong BK. 1984. Melanoma of the skin. Br Med Bull 40(4):346–350. Armstrong BK, English DR. 1996. Cutaneous malignant melanoma. In: Schottenfeld D, Fraumeni Jr. J, eds. Cancer Epidemiology and Prevention. New York: Oxford University Press: pp. 1282–1312. Armstrong BK, Kricker A. 1993. How much melanoma is caused by sun exposure? Melanoma Res 3(6):395–401. Armstrong BK, Kricker A. 2001. The epidemiology of UV induced skin cancer. J Photochem Photobiol B 63(1–3):8–18. Auroy S, Avril MF, et al. 2001. Sporadic multiple primary melanoma cases: CDKN2A germline mutations with a founder effect. Genes Chromosomes Cancer 32(3):195–202. Austin DF, Reynolds P. 1997. Investigation of an excess of melanoma among employees of the Lawrence Livermore National Laboratory [see comments]. Am J Epidemiol 145(6):524–531. Australian Institute of Health and Welfare. 2003. Interactive cancer data. Available at: http://www.aihw.gov.au/cancere/datacubes/index.html. Autier P, Boniol M, et al. 2004. Sex differences in numbers of nevi on body sites of young European children: Implications for the etiology of cutaneous melanoma. Cancer Epidemiol Biomarkers Prev 13(12): 2003–2005. Autier P, Dore JF, et al. 1997. Melanoma risk and residence in sunny areas. EORTC Melanoma Co-operative Group. European Organization for Research and Treatment of Cancer. Br J Cancer 76(11):1521–1524. Bader JL, Li FP, et al. 1985. Childhood malignant melanoma. Incidence and etiology. Am J Pediatr Hematol Oncol 7(4):341–345. Balch CM, Soong SJ, et al. 2001. Prognostic factors analysis of 17,600 melanoma patients: Validation of the American Joint Committee on Cancer melanoma staging system. J Clin Oncol 19(16):3622–3634. Baroody M, Holds JB. 2004. Extensive locoregional malignant melanoma transformation in a patient with oculodermal melanocytosis. Plast Reconstr Surg 113(1):317–322. Bastiaens M, ter Huurne J, et al. 2001. The melanocortin-1-receptor gene is the major freckle gene. Hum Mol Genet 10(16):1701–1708. Bastian BC. 2003. Understanding the progression of melanocytic neoplasia using genomic analysis: From fields to cancer. Oncogene 22(20): 3081–3086. Bastian BC, Kashani-Sabet M, et al. 2000. Gene amplifications characterize acral melanoma and permit the detection of occult tumor cells in the surrounding skin. Cancer Res 60(7):1968–1973. Bastian BC, Olshen AB, et al. 2003. Classifying melanocytic tumors based on DNA copy number changes. Am J Pathol 163(5):1765–1770. Bataille V, Sasieni P, et al. 1998. Solar keratoses: A risk factor for melanoma but negative association with melanocytic naevi. Int J Cancer 78(1): 8–12. Begg CB, Hummer A, et al. 2004. Familial aggregation of melanoma risks in a large population-based sample of melanoma cases. Cancer Causes Control 15(9):957–965. Begg CB, Orlow I, et al. 2005. Genes Environment and Melanoma Study Group. Lifetime risk of melanoma in CDKN2A mutation carriers in a population-based sample. J Natl Cancer Inst 97(20):1507–1515. Beral V, Evans S, et al. 1983. Cutaneous factors related to the risk of malignant melanoma. Br J Dermatol 109(2):165–172. Bergman L, Seregard S, et al. 2002. Incidence of uveal melanoma in Sweden from 1960 to 1998. Invest Ophthalmol Vis Sci 43(8):2579–2583. Bertram CG, Gaut RM, et al. 2004. An assessment of a variant of the DNA repair gene XRCC3 as a possible nevus or melanoma susceptibility genotype. J Invest Dermatol 122(2):429–432. Berwick M, Armstrong BK, et al. 2005. Sun exposure and mortality from melanoma. J Natl Cancer Inst 97(3):195–199. Berwick M, Begg CB, et al. 1996. Screening for cutaneous melanoma by skin self-examination. J Natl Cancer Inst 88(1):17–23. Berwick M. 2005. Personal communication. Bishop DT, Demenais F, et al. 2002. Geographical variation in the penetrance of CDKN2A mutations for melanoma. J Natl Cancer Inst 94(12): 894–903. Bliss JM, Ford D, et al. 1995. Risk of cutaneous melanoma associated with pigmentation characteristics and freckling: Systematic overview of 10 case-control studies. The International Melanoma Analysis Group (IMAGE). Int J Cancer 62(4):367–376. Boice JD Jr., Blettner M, et al. 2000. Epidemiologic studies of pilots and aircrew. Health Phys 79(5):576–584. Bono A, Baldi M, et al. 2003. Distribution of melanoma on breast surface suggests its etiology. Int J Cancer 105(3):434. Borg A, Johannsson U, et al. 1996. Novel germline p16 mutation in familial malignant melanoma in southern Sweden. Cancer Res 56(11):2497–2500.

1223

Borg A, Sandberg T, et al. 2000. High frequency of multiple melanomas and breast and pancreas carcinomas in CDKN2A mutation-positive melanoma families. J Natl Cancer Inst 92(15):1260–1266. Box NF, Duffy DL, et al. 2001. Mc1r genotype modifies risk of melanoma in families segregating cdkn2a mutations. Am J Hum Genet 69(4):765–773. Box NF, Wyeth JR, et al. 1997. Characterization of melanocyte stimulating hormone receptor variant alleles in twins with red hair. Hum Mol Genet 6(11):1891–1897. The Breast Cancer Linkage Consortium. 1999. Cancer risks in BRCA2 mutation carriers. J Natl Cancer Inst 91(15):1310–1316. Breslow A. 1970. Thickness, cross-sectional areas and depth of invasion in the prognosis of cutaneous melanoma. Ann Surg 172(5):902–908. Buller DB, Hall JR, et al. 1999. Evaluation of the “Sunny Days, Healthy Ways” sun safety CD-ROM program for children in grades 4 and 5. Cancer Prev Control 3(3):188–195. Bulliard JL. 2000. Site-specific risk of cutaneous malignant melanoma and pattern of sun exposure in New Zealand. Int J Cancer 85(5):627–632. Bulliard JL, Cox B. 2000. Cutaneous malignant melanoma in New Zealand: Trends by anatomical site, 1969–1993. Int J Epidemiol 29(3):416–423. Bulliard JL, Cox B, et al. 1997. Comparison of the site distribution of melanoma in New Zealand and Canada. Int J Cancer 72(2):231–235. Bulliard JL, Cox B, et al. 1999. Trends by anatomic site in the incidence of cutaneous malignant melanoma in Canada, 1969–93. Cancer Causes Control 10(5):407–416. Canning CR, Hungerford J. 1988. Familial uveal melanoma. Br J Ophthalmol 72(4):241–243. Cannon-Albright LA, Goldgar DE, et al. 1992. Assignment of a locus for familial melanoma, MLM, to chromosome 9p13-p22. Science 258(5085): 1148–1152. Cannon-Albright LA, Goldgar DE, et al. 1994. Localization of the 9p melanoma susceptibility locus (MLM) to a 2-cM region between D9S736 and D9S171. Genomics 23(1):265–268. Cannon-Albright LA, Meyer LJ, et al. 1994. Penetrance and expressivity of the chromosome 9p melanoma susceptibility locus (MLM). Cancer Res 54(23):6041–6044. Chang AE, Karnell LH, et al. 1998. The National Cancer Data Base report on cutaneous and noncutaneous melanoma: a summary of 84,836 cases from the past decade. The American College of Surgeons Commission on Cancer and the American Cancer Society. Cancer 83(8):1664–1678. Chin L. 2003. The genetics of malignant melanoma: lessons from mouse and man. Nat Rev Cancer 3(8):559–570. Chowers I, Folberg R, et al. 2002. p53 Immunoreactivity, Ki-67 expression, and microcirculation patterns in melanoma of the iris, ciliary body, and choroid. Curr Eye Res 24(2):105–108. Ciotti P, Strigini P, et al. 1996. Familial melanoma and pancreatic cancer. Ligurian Skin Tumor Study Group. N Engl J Med 334(7):469–470; author reply 471–472. Ciotti P, Struewing JP, et al. 2000. A single genetic origin for the G101W CDKN2A mutation in 20 melanoma-prone families. Am J Hum Genet 67(2):311–319. Clark WH, Jr., From L, et al. 1969. The histogenesis and biologic behavior of primary human malignant melanomas of the skin. Cancer Res 29(3):705–727. Clark WH, Jr., Reimer RR, et al. 1978. Origin of familial malignant melanomas from heritable melanocytic lesions. “The B-K mole syndrome”. Arch Dermatol 114(5):732–738. Claus EB, Risch NJ, et al. 1990. Age at onset as an indicator of familial risk of breast cancer. Am J Epidemiol 131(6):961–972. Collisson EA, Kleer C, et al. 2003. Atorvastatin prevents RhoC isoprenylation, invasion, and metastasis in human melanoma cells. Mol Cancer Ther 2(10):941–948. Corpron CA, Black CT, et al. 1996. Melanoma as a second malignant neoplasm after childhood cancer. Am J Surg 172(5):459–461; discussion 461–462. Crane LA, Schneider LS, et al. 1999. “Block the sun, not the fun”: Evaluation of a skin cancer prevention program for child care centers. Am J Prev Med 17(1):31–37. Cress RD, Holly EA. 1997. Incidence of cutaneous melanoma among non-Hispanic whites, Hispanics, Asians, and blacks: An analysis of California cancer registry data, 1988–93. Cancer Causes Control 8(2): 246–252. Cristofolini M, Franceschi S, et al. 1987. Risk factors for cutaneous malignant melanoma in a northern Italian population. Int J Cancer 39(2):150–154. Cruz F, 3rd, Rubin BP, et al. 2003. Absence of BRAF and NRAS mutations in uveal melanoma. Cancer Res 63(18):5761–5766. Dal Maso L, Franceschi S, et al. 2003. Risk of cancer in persons with AIDS in Italy, 1985–1998. Br J Cancer 89(1):94–100. Darlington S, Siskind V, et al. 2002. Longitudinal study of melanocytic nevi in adolescents. J Am Acad Dermatol 46(5):715–722.

1224

PART IV: CANCER BY TISSUE OF ORIGIN

Davies H, Bignell GR, et al. 2002. Mutations of the BRAF gene in human cancer. Nature 417(6892):949–954. de Vries E, Bray FI, et al. 2003. Changing epidemiology of malignant cutaneous melanoma in Europe 1953–1997: Rising trends in incidence and mortality but recent stabilizations in western Europe and decreases in Scandinavia. Int J Cancer 107(1):119–126. Debniak T, Gorski B, et al. 2004. Rarity of germline 1100delC mutation in CHK2 in patients with malignant melanoma of the skin. Melanoma Res 14(2):121–124. Dellavalle RP, Nicholas MK, et al. 2003. Melanoma chemoprevention: A role for statins or fibrates? Am J Ther 10(3):203–210. Demierre MF, Nathanson L. 2003. Chemoprevention of melanoma: an unexplored strategy. J Clin Oncol 21(1):158–165. Dennis LK, Beane Freeman LE, et al. 2003. Sunscreen use and the risk for melanoma: A quantitative review. Ann Intern Med 139(12):966–978. Dietrich AJ, Olson AL, et al. 2000. Persistent increase in children’s sun protection in a randomized controlled community trial. Prev Med 31(5): 569–574. Dolzan V, Rudolf Z, et al. 1995. Human CYP2D6 gene polymorphism in Slovene cancer patients and healthy controls. Carcinogenesis 16(11): 2675–2678. Duan Z, Shen H, et al. 2002. DNA repair gene XRCC3 241Met variant is not associated with risk of cutaneous malignant melanoma. Cancer Epidemiol Biomarkers Prev 11(10 Pt 1):1142–1143. Dwyer T, Blizzard L, et al. 2002. Cutaneous melanin density of Caucasians measured by spectrophotometry and risk of malignant melanoma, basal cell carcinoma, and squamous cell carcinoma of the skin. Am J Epidemiol 155(7):614–621. Dwyer T, Muller HK, et al. 1998. The use of spectrophotometry to estimate melanin density in Caucasians. Cancer Epidemiol Biomarkers Prev 7(3):203–206. Dwyer T, Stankovich JM, et al. 2004. Does the addition of information on genotype improve prediction of the risk of melanoma and nonmelanoma skin cancer beyond that obtained from skin phenotype? Am J Epidemiol 159(9):826–833. Edmunds SC, Kelsell DP, et al. 2002. Mutational analysis of selected genes in the TGFbeta, Wnt, pRb, and p53 pathways in primary uveal melanoma. Invest Ophthalmol Vis Sci 43(9):2845–2851. Egan CL, Oliveria SA, et al. 1998. Cutaneous melanoma risk and phenotypic changes in large congenital nevi: A follow-up study of 46 patients. J Am Acad Dermatol 39(6):923–932. El-Shabrawi Y, Radner H, et al. 1999. The role of UV-radiation in the development of conjunctival malignant melanoma. Acta Ophthalmol Scand 77(1):31–32. Elwood JM, Gallagher RP. 1983. Site distribution of malignant melanoma. Can Med Assoc J 128(12):1400–1404. Elwood JM, Gallagher RP. 1998. Body site distribution of cutaneous malignant melanoma in relationship to patterns of sun exposure. Int J Cancer 78(3):276–280. Elwood JM, Jopson J. 1997. Melanoma and sun exposure: an overview of published studies. Int J Cancer 73(2):198–203. Elwood JM, Lee JA, et al. 1974. Relationship of melanoma and other skin cancer mortality to latitude and ultraviolet radiation in the United States and Canada. Int J Epidemiol 3(4):325–332. Elwood JM, Williamson C, Stapleton PJ. 1986. Malignant melanoma in relation to moles, pigmentation, and exposure to fluorescent and other lighting sources. Br J Cancer 53(1):65–74. Elwood JM, Whitehead SM, Davison J, et al. 1990. Malignant melanoma in England: risks associated with naevi, freckles, social class, hair colour, and sunburn. Int J Epidemiol 19:801–810. English, DR, Armstrong BK, et al. 1998. Reproducibility of reported measurements of sun exposure in a case-control study. Cancer Epidemiol Biomarkers Prev 7(10):857–863. English DR, Armstrong BK, et al. 1997. Sunlight and cancer. Cancer Causes Control 8(3):271–283. English DR, Milne E. 1999. Favorable trends in melanoma incidence: Can we claim credit? Cancer Causes Control 10(5):403–405. English JS, Swerdlow AJ. 1987. The risk of malignant melanoma, internal malignancy and mortality in xeroderma pigmentosum patients. Br J Dermatol 117(4):457–461. Epstein DS, Lange JR, et al. 1999. Is physician detection associated with thinner melanomas? JAMA 281(7):640–643. Euvrard S, Kanitakis J, et al. 2003. Skin cancers after organ transplantation. N Engl J Med 348(17):1681–1691. Faggiano F, Partanen T, et al. 1997. Socioeconomic differences in cancer incidence and mortality. IARC Sci Publ(138):65–176. Fears TR, Bird CC, et al. 2002. Average midrange ultraviolet radiation flux and time outdoors predict melanoma risk. Cancer Res 62(14):3992–3996.

Ferlay J, Bray F, Pisani P, Parkin DM. GLOBOCAN 2002: Cancer Incidence, Mortality and Prevalence Worldwide, Version 2.0. IARC CancerBase No. 5. Lyon, France: IARCPress, 2004. Available at: www-depdb.iarc.fr/ globocan2002.htm. Accessed February 19, 2005. Feskanich D, Hunter DJ, et al. 1999. Oral contraceptive use and risk of melanoma in premenopausal women. Br J Cancer 81(5):918–923. Ford D, Bliss JM, et al. 1995. Risk of cutaneous melanoma associated with a family history of the disease. The International Melanoma Analysis Group (IMAGE). Int J Cancer 62(4):377–381. Fountain JW, Karayiorgou M, et al. 1992. Homozygous deletions within human chromosome band 9p21 in melanoma. Proc Natl Acad Sci U S A 89(21):10557–10561. Franceschi S, Levi F, et al. 1996. Site distribution of different types of skin cancer: New aetiological clues. Int J Cancer 67(1):24–28. Freedman DM, Sigurdson A, et al. 2003. Risk of melanoma in relation to smoking, alcohol intake, and other factors in a large occupational cohort. Cancer Causes Control 14(9):847–857. Freedman DM, Sigurdson A, et al. 2003. Risk of melanoma among radiologic technologists in the United States. Int J Cancer 103(4):556–562. Frisch M, Smith E, et al. 2003. Cancer in a population-based cohort of men and women in registered homosexual partnerships. Am J Epidemiol 157(11):966–972. Fritschi L, Siemiatycki J. 1996. Melanoma and occupation: Results of a casecontrol study. Occup Environ Med 53(3):168–173. Gallagher B, Wang Z, et al. 2001. Cancer incidence in New York State acquired immunodeficiency syndrome patients. Am J Epidemiol 154(6):544–556. Gallagher RP, Elwood JM, et al. 1985. Risk factors for ocular melanoma: Western Canada Melanoma Study. J Natl Cancer Inst 74(4):775–778. Gallagher RP, Rivers JK, et al. 2000. Broad-spectrum sunscreen use and the development of new nevi in white children: A randomized controlled trial. JAMA 283(22):2955–2960. Gallagher RP, Spinelli JJ, et al. 2005. Tanning beds, sunlamps, and risk of cutaneous malignant melanoma. Cancer Epidemiol Biomarkers Prev 14(3):562–566. Gandini S, Sera F, et al. 2005. Meta-analysis of risk factors for cutaneous melanoma: I. Common and atypical naevi. Eur J Cancer 41(1):28–44. Gandini S, Sera F, et al. 2005. Meta-analysis of risk factors for cutaneous melanoma: II. Sun exposure. Eur J Cancer 41(1):45–60. Garber JE, Liepman MK, et al. 1990. Melanoma and soft tissue sarcoma in seven patients. Cancer 66(11):2432–2434. Gaudette LA, Gao RN. 1998. Changing trends in melanoma incidence and mortality. Health Rep 10(2):29–41 (Eng); 33–46 (Fre). German J. 1997. Bloom’s syndrome. XX. The first 100 cancers. Cancer Genet Cytogenet. 93(1):100–106. German J, Ellis N. 2002. Bloom syndrome. In: Vogelstein B, Kinzler K, eds. The Genetic Basis of Human Cancer. New York, McGraw-Hill: pp. 267–288. Ghiorzo P, Ciotti P, et al. 1999. Characterization of ligurian melanoma families and risk of occurrence of other neoplasia. Int J Cancer 83(4): 441–448. Gilchrest BA, Eller MS, et al. 1999. The pathogenesis of melanoma induced by ultraviolet radiation [Review]. N Engl J Med 340(17):1341–1348. Giles GG, Armstrong BK, et al. 1996. Has mortality from melanoma stopped rising in Australia? Analysis of trends between 1931 and 1994. BMJ 312(7039):1121–1125. Gillgren P, Brattstrom G, et al. 2002. A new computerized methodology to analyse tumour site in relation to phenotypic traits and epidemiological characteristics of cutaneous malignant melanoma. Br J Dermatol 146(6):1023–1030. GISED. 2003. Improving sun protection behaviour in children: Study design and baseline results of a randomized trial in Italian elementary schools. The “Sole Si Sole No GISED” project. Dermatology 207(3):291–297. Glanz K, Lew RA, et al. 2000. Skin cancer prevention in outdoor recreation settings: effects of the Hawaii SunSmart Program. Eff Clin Pract 3(2):53–61. Goldstein AM, Liu L, et al. 2001. A common founder for the V216D CDKN2A mutation in seven North American melanoma-prone families. Br J Cancer 85(4):527–530. Goldstein AM, Martinez M, et al. 2000. Gene-covariate interaction between dysplastic nevi and the CDKN2A gene in American melanoma-prone families. Cancer Epidemiol Biomarkers Prev 9(9):889–894. Goldstein AM, Struewing JP, et al. 2000. Genotype–phenotype relationships in US melanoma-prone families with CDKN2A and CDK4 mutations. J Natl Cancer Inst 92(12):1006–1010. Gorey KM, Holowaty EJ, et al. 1998. Association between socioeconomic status and cancer incidence in Toronto, Ontario: possible confounding of cancer mortality by incidence and survival. Cancer Prev Control 2(5):236–241.

Cutaneous and Ocular Melanoma Goto M, Miller RW, et al. 1996. Excess of rare cancers in Werner syndrome adult progeria. Cancer Epidemiol Biomarkers Prev 5(4):239–246. Graf J, Hodgson R, et al. 2005. Single nucleotide polymorphisms in the MATP gene are associated with normal human pigmentation variation. Hum Mutat 25(3):278–284. Grammatico P, Binni F, et al. 2001. CDKN2A novel mutation in a patient from a melanoma-prone family. Melanoma Res 11(5):447–449. Green A, MacLennan R, et al. 1985. Common acquired naevi and the risk of malignant melanoma. Int J Cancer 35(3):297–300. Green A, MacLennan R, et al. 1993. Site distribution of cutaneous melanoma in Queensland. Int J Cancer 53(2):232–236. Green A, McCredie M, et al. 1996. Occurrence of melanomas on the upper and lower limbs in eastern Australia. Melanoma Res 6(5):387–394. Green, A, McCredie M, et al. 1999. A case-control study of melanomas of the soles and palms (Australia and Scotland). Cancer Causes Control 10(1):21–25. Green A, Williams G, et al. 1999. Daily sunscreen application and betacarotene supplementation in prevention of basal-cell and squamous-cell carcinomas of the skin: A randomised controlled trial. Lancet 354(9180): 723–729. Greene M, Fraumeni Jr., J. 1979. The Hereditary Variant of Malignant Melanoma. Human Malignant Melanoma. Clark, Jr. WH, Goldman LI, Mastrangelo MJ, eds. New York; Grune Stratton: pp. 139–166. Greene MH, Reimer RR, et al. 1978. Precursor lesions in familial melanoma. Semin Oncol 5(1):85–87. Gruber SB, Barnhill RL, et al. 1989. Nevomelanocytic proliferations in association with cutaneous malignant melanoma: A multivariate analysis. J Am Acad Dermatol 21(4):773–780. Gruber SB, Roush GC, et al. 1993. Sensitivity and specificity of selfexamination for cutaneous malignant melanoma risk factors. Am J Prev Med 9(1):50–54. Gruis NA, Sandkuijl LA, et al. 1995. CDKN2 explains part of the clinical phenotype in Dutch familial atypical multiple-mole melanoma (FAMMM) syndrome families. Melanoma Res 5(3):169–177. Grulich AE, Li Y, et al. 2002. Rates of non-AIDS-defining cancers in people with HIV infection before and after AIDS diagnosis. AIDS 16(8):1155–1161. Guerin S, Dupuy A, et al. 2003. Radiation dose as a risk factor for malignant melanoma following childhood cancer. Eur J Cancer 39(16):2379– 2386. Guldberg P, thor Straten P, et al. 1997. Disruption of the MMAC1/PTEN gene by deletion or mutation is a frequent event in malignant melanoma. Cancer Res 57(17):3660–3663. Gun RT, Pratt NL, et al. 2004. Update of a prospective study of mortality and cancer incidence in the Australian petroleum industry. Occup Environ Med 61(2):150–156. Gutman M, Cnaan A, et al. 1991. Are malignant melanoma patients at higher risk for a second cancer? Cancer 68(3):660–665. Habif TP. 2004. Nevi and Melanoma in: Clinical Dermatology. A Color Guide to Diagnoses and Therapy. 4th ed. Philadelphia; Mosby, pp. 778–813. Hall HI, Miller DR, et al. 1999. Update on the incidence and mortality from melanoma in the United States. J Am Acad Dermatol 40(1):35–42. Halpern AC, Guerry D, et al. 1991. Dysplastic nevi as risk markers of sporadic (nonfamilial) melanoma. A case-control study. Arch Dermatol 127(7): 995–999. Halpern AC, Guerry D, et al. 1993. A cohort study of melanoma in patients with dysplastic nevi. J Invest Dermatol 100(3):346S–349S. Halpern AC, Schuchter LM, et al. 1994. Effects of topical tretinoin on dysplastic nevi. J Clin Oncol 12(5):1028–1035. Halsall JA, Osborne JE, et al. 2004. A novel polymorphism in the 1A promoter region of the vitamin D receptor is associated with altered susceptibility and prognosis in malignant melanoma. Br J Cancer 91(4):765–770. Harbour JW, Brantley Jr. MA, et al. 2004. Association between choroidal pigmentation and posterior uveal melanoma in a white population. Br J Ophthalmol 88(1):39–43. Harbour JW, Brantley, Jr. MA, et al. 2004. Association between posterior uveal melanoma and iris freckles, iris naevi, and choroidal naevi. Br J Ophthalmol 88(1):36–38. Harris RE, Beebe-Donk J, et al. 2001. Inverse association of non-steroidal antiinflammatory drugs and malignant melanoma among women. Oncol Rep 8(3):655–657. Harrison RA, Haque AU, et al. 1998. Socioeconomic characteristics and melanoma incidence. Ann Epidemiol 8(5):327–333. Hartley AL, Birch JM, et al. 1989. Are germ cell tumors part of the LiFraumeni cancer family syndrome? Cancer Genet Cytogenet 42(2): 221–226. Hayward NK. 2003. Genetics of melanoma predisposition. Oncogene 22(20): 3053–3062.

1225

Heagerty AH, Fitzgerald D, et al. 1994. Glutathione S-transferase GSTM1 phenotypes and protection against cutaneous tumours. Lancet 343(8892): 266–268. Hearle N, Damato BE, et al. 2003. Contribution of germline mutations in BRCA2, P16(INK4A), P14(ARF) and P15 to uveal melanoma. Invest Ophthalmol Vis Sci 44(2):458–462. Hearle N, Humphreys J, et al. 2003. Role of MC1R variants in uveal melanoma. Br J Cancer 89(10):1961–1965. Hemminki K, Jiang Y. 2001. Association of ocular melanoma with breast cancer but not with cutaneous melanoma: Results from the Swedish Family-Cancer Database. Int J Cancer 94(6):907–909. Hemminki K, Zhang H, et al. 2003. Familial and attributable risks in cutaneous melanoma: Effects of proband and age. J Invest Dermatol 120(2):217– 223. Hoar SK, Wilson J, et al. 1985. Second cancer following cancer of the digestive system in Connecticut, 1935–1982. Multiple Primary Cancers in Connecticut and Denmark. Bethesda, MD: National Cancer Institute. Monograph 68:49–82. Holly EA, Aston DA, et al. 1996. Intraocular melanoma linked to occupations and chemical exposures. Epidemiology 7(1):55–61. Holly EA, Aston DA, et al. 1995. Cutaneous melanoma in women. I. Exposure to sunlight, ability to tan, and other risk factors related to ultraviolet light. Am J Epidemiol 141(10):923–933. Holly EA, Kelly JW, et al. 1987. Number of melanocytic nevi as a major risk factor for malignant melanoma. J Am Acad Dermatol 17(3):459–468. Holman CD, Armstrong BK. 1984. Cutaneous malignant melanoma and indicators of total accumulated exposure to the sun: An analysis separating histogenetic types. J Natl Cancer Inst 73(1):75–82. Holman CD, Armstrong BK. 1984. Pigmentary traits, ethnic origin, benign nevi, and family history as risk factors for cutaneous malignant melanoma. J Natl Cancer Inst 72(2):257–266. Horn EP, Hartge P, et al. 1994. Sunlight and risk of uveal melanoma. J Natl Cancer Inst 86(19):1476–1478. Hussussian CJ, Struewing JP, et al. 1994. Germline p16 mutations in familial melanoma. Nat Genet 8(1):15–21. Hutchinson PE, Osborne JE, et al. 2000. Vitamin D receptor polymorphisms are associated with altered prognosis in patients with malignant melanoma. Clin Cancer Res 6(2):498–504. Hwang K, Han JY, et al. 2004. Multiple malignant melanomas at different burn scar areas: A case report. Dermatol Surg 30(4 Pt 1):562–565. Inskip PD, Devesa SS, et al. 2003. Trends in the incidence of ocular melanoma in the United States, 1974–1998. Cancer Causes Control 14(3):251–257. Isager P, Ehlers N, et al. 2004. Prognostic factors for survival after enucleation for choroidal and ciliary body melanomas. Acta Ophthalmol Scand 82(5):517–525. Iscovich J, Abdulrazik M, et al. 2002. Prevalence of the BRCA2 6174 del T mutation in Israeli uveal melanoma patients. Int J Cancer 98(1):42– 44. Iscovich J, Abdulrazik M, et al. 2001. Posterior uveal malignant melanoma: Temporal stability and ethnic variation in rates in Israel. Anticancer Res 21(2B):1449–1454. Ishihara K, Saida T, et al. 2001. Updated statistical data for malignant melanoma in Japan. Int J Clin Oncol 6(3):109–116. Jelfs PL, Giles G, et al. 1994. Cutaneous malignant melanoma in Australia, 1989. Med J Aust 161(3):182–187. Jemal A, Devesa SS, et al. 2000. Cancer surveillance series: Changing patterns of cutaneous malignant melanoma mortality rates among whites in the United States. J Natl Cancer Inst 92(10):811–818. Jemal A, Devesa SS, et al. 2001. Recent trends in cutaneous melanoma incidence among whites in the United States. J Natl Cancer Inst 93(9): 678–683. Johansen C, Boice, Jr. JD, et al. 2002. Mobile phones and malignant melanoma of the eye. Br J Cancer 86(3):348–349. Johnson MW, Skuta GL, et al. 1989. Malignant melanoma of the iris in xeroderma pigmentosum. Arch Ophthalmol 107(3):402–407. Kamb A, Gruis NA, et al. 1994. A cell cycle regulator potentially involved in genesis of many tumor types [see comments]. Science 264(5157): 436–440. Kamb A, Shattuck-Eidens D, et al. 1994. Analysis of the p16 gene (CDKN2) as a candidate for the chromosome 9p melanoma susceptibility locus. Nat Genet 8(1):23–26. Kanetsky PA, Ge F, et al. 2004. Assessment of polymorphic variants in the melanocortin-1 receptor gene with cutaneous pigmentation using an evolutionary approach. Cancer Epidemiol Biomarkers Prev 13(5):808– 819. Kanetsky PA, Holmes R, et al. 2001. Interaction of glutathione S-transferase M1 and T1 genotypes and malignant melanoma. Cancer Epidemiol Biomarkers Prev 10(5):509–513.

1226

PART IV: CANCER BY TISSUE OF ORIGIN

Karagas MR, Stukel TA, et al. 2002. A pooled analysis of 10 case-control studies of melanoma and oral contraceptive use. Br J Cancer 86(7): 1085–1092. Karagas MR, Zens MS, et al. submitted. Pregnancy history and incidence of melanoma in women: a pooled analysis. Cancer Causes Control, in press. Kaufman DK, Kimmel DW, et al. 1993. A familial syndrome with cutaneous malignant melanoma and cerebral astrocytoma. Neurology 43(9): 1728–1731. Kaye JA, Jick H. 2004. Statin use and cancer risk in the General Practice Research Database. Br J Cancer 90(3):635–637. Kefford RF, Bishop JAN, et al. 1999. Counseling and DNA testing for individuals perceived to be genetically predisposed to melanoma: A consensus statement of the melanoma genetics consortium. J Clin Oncol 17(10):3245–3251. Kelly JW, Yeatman JM, et al. 1997. A high incidence of melanoma found in patients with multiple dysplastic naevi by photographic surveillance. Med J Aust 167(4):191–194. Kennedy C, ter Huurne J, et al. 2001. Melanocortin 1 receptor (MC1R) gene variants are associated with an increased risk for cutaneous melanoma which is largely independent of skin type and hair color. J Invest Dermatol 117(2):294–300. Khlat M, Vail A, et al. 1992. Mortality from melanoma in migrants to Australia: Variation by age at arrival and duration of stay. Am J Epidemiol 135(10):1103–1113. Kilic E, Bruggenwirth HT, et al. 2004. The RAS-BRAF kinase pathway is not involved in uveal melanoma. Melanoma Res 14(3):203–205. Knudson AG, Jr. 1971. Mutation and cancer: Statistical study of retinoblastoma. Proceedings of the National Academy of Sciences of the United States of America 68(4):820–823. Kodjikian L, Nguyen K, et al. 2003. Familial uveal melanoma: a report on two families and a review of the literature. Acta Ophthalmol Scand 81(4):389–395. Kolmel KF, Gefeller O, et al. 1992. Febrile infections and malignant melanoma: Results of a case-control study. Melanoma Res 2(3):207–211. Kolmel KF, Pfahlberg A, et al. 1999. Infections and melanoma risk: Results of a multicentre EORTC case-control study. European Organization for Research and Treatment of Cancer. Melanoma Res 9(5):511–519. Korczynski RE. 2000. Occupational health concerns in the welding industry. Appl Occup Environ Hyg 15(12):936–945. Kraemer KH, Lee MM, et al. 1987. Xeroderma pigmentosum. Cutaneous, ocular, and neurologic abnormalities in 830 published cases. Arch Dermatol 123(2):241–250. Krone B, Kolmel KF, et al. 2003. Impact of vaccinations and infectious diseases on the risk of melanoma—evaluation of an EORTC case-control study. Eur J Cancer 39(16):2372–2378. Kruger S, Garbe C, et al. 1992. Epidemiologic evidence for the role of melanocytic nevi as risk markers and direct precursors of cutaneous malignant melanoma. Results of a case control study in melanoma patients and nonmelanoma control subjects. J Am Acad Dermatol 26(6):920–926. La Vecchia C, Lucchini F, et al. 1999. Recent declines in worldwide mortality from cutaneous melanoma in youth and middle age. Int J Cancer 81(1):62–66. Lafuente A, Molina R, et al. 1995. Phenotype of glutathione S-transferase Mu (GSTM1) and susceptibility to malignant melanoma. MMM group. Multidisciplinary Malignant Melanoma Group. Br J Cancer 72(2):324–326. Lambe M, Thorn M, et al. 1996. Malignant melanoma: Reduced risk associated with early childbearing and multiparity. Melanoma Res 6(2): 147–153. Langard S, Rosenberg J, et al. 2000. Incidence of cancer among workers exposed to vinyl chloride in polyvinyl chloride manufacture. Occup Environ Med 57(1):65–68. Lee JA, Storer BE. 1980. Excess of malignant melanomas in women in the British Isles. Lancet 2(8208–8209):1337–1339. Lemish WM, Heenan PJ, et al. 1983. Survival from preinvasive and invasive malignant melanoma in Western Australia. Cancer 52(3):580–585. Lewis N, Nguyen H, Smith D, Coates M, Armstrong B. 1999. Cancer maps for New South Wales: Variation by local government area 1991–1995. Sydney: NSW Cancer Council, p. 73. Lewis RJ, Schnatter AR, et al. 2003. Mortality and cancer morbidity in a cohort of Canadian petroleum workers. Occup Environ Med 60(12):918–928. Lindelof B, Sigurgeirsson B, et al. 1991. PUVA and cancer: A large-scale epidemiological study. Lancet 338(8759):91–93. Linet MS, Malker HS, et al. 1995. Occupational risks for cutaneous melanoma among men in Sweden. J Occup Environ Med 37(9):1127–1135. Liu L, Dilworth D, et al. 1999. Mutation of the CDKN2A 5¢ UTR creates an aberrant initiation codon and predisposes to melanoma. Nat Genet 21(1):128–132.

Loomis D, Browning SR, et al. 1997. Cancer mortality among electric utility workers exposed to polychlorinated biphenyls. Occup Environ Med 54(10):720–728. Love RR, Evans AM, et al. 1985. The accuracy of patient reports of a family history of cancer. J Chronic Dis 38(4):289–293. Lowe JB, Ball J, et al. 2004. Acceptability and feasibility of a communitybased screening programme for melanoma in Australia. Health Promot Int 19(4):437–444. Lundberg I, Gustavsson A, et al. 1993. Mortality and cancer incidence among PVC-processing workers in Sweden. Am J Ind Med 23(2):313–319. Lynch HT, Brand RE, et al. 2002. Phenotypic variation in eight extended CDKN2A germline mutation familial atypical multiple mole melanomapancreatic carcinoma-prone families: The familial atypical mole melanoma-pancreatic carcinoma syndrome. Cancer 94(1):84–96. Lynch HT, Frichot, III, BC, et al. 1978. Familial atypical multiple molemelanoma syndrome. J Med Genet 15(5):352–356. Lynch HT, Fusaro RM, et al. 1983. A review of hereditary malignant melanoma including biomarkers in familial atypical multiple mole melanoma syndrome. Cancer Genet Cytogenet 8(4):325–358. Mack TM, Floderus B. 1991. Malignant melanoma risk by nativity, place of residence at diagnosis, and age at migration. Cancer Causes Control 2(6):401–411. MacKie RM, Andrew N, et al. 1998. CDKN2A germline mutations in U.K. patients with familial melanoma and multiple primary melanomas. J Invest Dermatol 111(2):269–272. MacKie RM, Freudenberger T, et al. 1989. Personal risk-factor chart for cutaneous melanoma. Lancet 2(8661):487–490. Marrett LD, Nguyen HL, et al. 2001. Trends in the incidence of cutaneous malignant melanoma in New South Wales, 1983–1996. Int J Cancer 92(3):457–462. Massi D, Carli P, et al. 1999. Naevus-associated melanomas: Cause or chance? Melanoma Res 9(1):85–91. Matichard E, Verpillat P, et al. 2004. Melanocortin 1 receptor (MC1R) gene variants may increase the risk of melanoma in France independently of clinical risk factors and UV exposure. J Med Genet 41(2):e13. McCarron SL, Bateman AC, et al. 2003. EGF +61 gene polymorphism and susceptibility to and prognostic markers in cutaneous malignant melanoma. Int J Cancer 107(4):673–675. Millen AE, Tucker MA, et al. 2004. Diet and melanoma in a case-control study. Cancer Epidemiol Biomarkers Prev 13(6):1042–1051. Milne E, English DR, et al. 1999. Evaluation of an intervention to reduce sun exposure in children: Design and baseline results. Am J Epidemiol 150(2):164–173. Milne E, Johnston R, et al. 2002. Effect of a school-based sun-protection intervention on the development of melanocytic nevi in children. Am J Epidemiol 155(8):739–745. Monarrez-Espino J, Stang A, et al. 2002. Occupation as a risk factor for uveal melanoma in Germany. Scand J Work Environ Health 28(4):270– 277. Movshovitz M, Modan B. 1973. Role of sun exposure in the etiology of malignant melanoma: Epidemiologic inference. J Natl Cancer Inst 51(3):777–779. Mudhar HS, Parsons MA, et al. 2004. A critical appraisal of the prognostic and predictive factors for uveal malignant melanoma. Histopathology 45(1):1–12. Muir CS, Waterhouse J, Mack T, et al. 1987. Cancer Incidence in Five Continents, Volume V. IARC Scientific Publications No. 88, Lyon, IARC. Nelson MA, Radmacher MD, et al. 2000. Chromosome abnormalities in malignant melanoma: Clinical significance of nonrandom chromosome abnormalities in 206 cases. Cancer Genet Cytogenet 122(2):101–109. Newnham A, Moller H. 2002. Trends in the incidence of cutaneous malignant melanomas in the south east of England, 1960–1998. J Public Health Med 24(4):268–275. Newton Bishop JA, Wachsmuth RC, et al. 2000. Genotype/phenotype and penetrance studies in melanoma families with germline CDKN2A mutations. J Invest Dermatol 114(1):28–33. Nobori T, Miura K, et al. 1994. Deletions of the cyclin-dependent kinase-4 inhibitor gene in multiple human cancers. Nature 368(6473):753–756. Noonan FP, Recio JA, et al. 2001. Neonatal sunburn and melanoma in mice. Nature 413(6853):271–272. Norris W. 1820. A case of fungoid disease. Edinburgh Surg J 16:562–565. Okuda T, Shurtleff SA, et al. 1995. Frequent deletion of p16INK4a/MTS1 and p15INK4b/MTS2 in pediatric acute lymphoblastic leukemia. Blood 85(9):2321–2330. Oliveria SA, Chau D, et al. 2004. Diagnostic accuracy of patients in performing skin self-examination and the impact of photography. Arch Dermatol 140(1):57–62.

Cutaneous and Ocular Melanoma Onken MD, Worley LA, et al. 2004. Gene expression profiling in uveal melanoma reveals two molecular classes and predicts metastatic death. Cancer Res 64(20):7205–7209. Osterlind A, Hou-Jensen K, et al. 1988. Incidence of cutaneous malignant melanoma in Denmark 1978–1982. Anatomic site distribution, histologic types, and comparison with non-melanoma skin cancer. Br J Cancer 58(3):385–391. Osterlind A, Tucker MA, et al. 1988. The Danish case-control study of cutaneous malignant melanoma. I. Importance of host factors. Int J Cancer 42(2):200–206. Osterlind A, Tucker MA, et al. 1988. The Danish case-control study of cutaneous malignant melanoma. IV. No association with nutritional factors, alcohol, smoking or hair dyes. Int J Cancer 42(6):825–828. Palmer JS, Duffy DL, et al. 2000. Melanocortin-1 receptor polymorphisms and risk of melanoma: Is the association explained solely by pigmentation phenotype? Am J Hum Gen 66(1):176–186. Pane AR, Hirst LW. 2000. Ultraviolet light exposure as a risk factor for ocular melanoma in Queensland, Australia. Ophthalmic Epidemiol 7(3): 159–167. Parkin DM, Muir CS, Whelan SL, et al. 1992. Cancer Incidence in Five Continents, Volume VI. IARC Scientific Publications No. 120, Lyon, IARC. Parkin DM, Steinitz R, et al. 1990. Cancer in Jewish migrants to Israel. Int J Cancer 45(4):614–621. Parkin DM, Whelan SL, Ferlay J, et al. 1997. Cancer Incidence in Five Continents, Volume VII. IARC Scientific Publications No. 143, Lyon, IARC. Parkin DM, Whelan SL, Ferlay J, et al. 2002. Cancer Incidence in Five Continents, Volume VIII. IARC Scientific Publications No. 155, Lyon, IARC. Pastorino L, Cusano R, et al. 2004. Novel MC1R variants in Ligurian melanoma patients and controls. Hum Mutat 24(1):103. Perez-Gomez B, Pollan M, et al. 2004. Cutaneous melanoma: hints from occupational risks by anatomic site in Swedish men. Occup Environ Med 61(2):117–126. Persson I, Yuen J, et al. 1996. Cancer incidence and mortality in women receiving estrogen and estrogen-progestin replacement therapy—long-term follow-up of a Swedish cohort. Int J Cancer 67(3):327–332. Peto J, Houlston RS. 2001. Genetics and the common cancers. Eur J Cancer 37 Suppl 8:S88–96. Petty EM, Bolognia JL, et al. 1993. Cutaneous malignant melanoma and atypical moles associated with a constitutional rearrangement of chromosomes 5 and 9. Am J Med Genet 45(1):77–80. Petty EM, Gibson LH, et al. 1993. Molecular definition of a chromosome 9p21 germ-line deletion in a woman with multiple melanomas and a plexiform neurofibroma: Implications for 9p tumor-suppressor genes. Am J Hum Genet 53(1):96–104. Pfahlberg A, Hassan K, et al. 1997. Systematic review of case-control studies: oral contraceptives show no effect on melanoma risk. Public Health Rev 25(3–4):309–315. Platz A, Hansson J, et al. 1997. Screening of germline mutations in the CDKN2A and CDKN2B genes in Swedish families with hereditary cutaneous melanoma. J Natl Cancer Inst 89(10):697–702. Pollock PM, Spurr N, et al. 1998. Haplotype analysis of two recurrent CDKN2A mutations in 10 melanoma families: Evidence for common founders and independent mutations. Hum Mutat 11(6):424–431. Puig, S, Ruiz A, et al. 1997. Inherited susceptibility to several cancers but absence of linkage between dysplastic nevus syndrome and CDKN2A in a melanoma family with a mutation in the CDKN2A (P16INK4A) gene. Hum Genet 101(3):359–364. Pukkala E, Aspholm R, et al. 2002. Incidence of cancer among Nordic airline pilots over five decades: Occupational cohort study. BMJ 325(7364): 567. Rafnsson V, Hrafnkelsson J, et al. 2003. Risk factors for cutaneous malignant melanoma among aircrews and a random sample of the population. Occup Environ Med 60(11):815–820. Raivio I. 1977. Uveal melanoma in Finland. An epidemiological, clinical, histological and prognostic study. Acta Ophthalmol Suppl(133):1–64. Randerson-Moor JA, Gaut R, et al. 2004. The relationship between the epidermal growth factor (EGF) 5¢UTR variant A61G and melanoma/nevus susceptibility. J Invest Dermatol 123(4):755–759. Rebbeck TR, Kanetsky PA, et al. 2002. P gene as an inherited biomarker of human eye color. Cancer Epidemiol Biomarkers Prev 11(8):782–784. Richtig E, Langmann G, et al. 2004. Ocular melanoma: Epidemiology, clinical presentation and relationship with dysplastic nevi. Ophthalmologica 218(2):111–114. Ries LAG, Eisner MP, et al., eds. 2004. SEER Cancer Statistics Review, 1975–2001. Bethesda, MD: National Cancer Institute.

1227

Rigel DS, Rivers JK, et al. 1989. Dysplastic nevi. Markers for increased risk for melanoma. Cancer 63(2):386–389. Robsahm TE, Tretli S. 2001. Cutaneous malignant melanoma in Norway: Variation by region of residence before and after the age 17. Cancer Causes Control 12(6):569–576. Rodolfo M, Daniotti M, et al. 2004. Genetic progression of metastatic melanoma. Cancer Lett 214(2):133–147. Rolon PA, Kramarova E, et al. 1997. Plantar melanoma: A case-control study in Paraguay. Cancer Causes Control 8(6):850–856. Ron E, Preston DL, et al. 1998. Skin tumor risk among atomic-bomb survivors in Japan. Cancer Causes Control 9(4):393–401. Rosenthal FS, Bakalian AE, et al. 1988. The effect of sunglasses on ocular exposure to ultraviolet radiation. Am J Public Health 78(1):72–74. Rosenthal FS, Phoon C, et al. 1988. The ocular dose of ultraviolet radiation to outdoor workers. Invest Ophthalmol Vis Sci 29(4):649–656. Roush GC, Nordlund JJ, et al. 1988. Independence of dysplastic nevi from total nevi in determining risk for nonfamilial melanoma. Prev Med 17(3):273–279. Rutter JL, Bromley CM, et al. 2004. Heterogeneity of risk for melanoma and pancreatic and digestive malignancies: A melanoma case-control study. Cancer 101(12):2809–2816. Sayre RM, Dowdy JC, et al. 2004. Dermatological risk of indoor ultraviolet exposure from contemporary lighting sources. Photochem Photobiol 80:47–51. Scott RJ, Vajdic CM, et al. 2002. BRCA2 mutations in a population-based series of patients with ocular melanoma. Int J Cancer 102(2):188–191. Scotto J, Fears TR. 1987. The association of solar ultraviolet and skin melanoma incidence among Caucasians in the United States. Cancer Invest 5(4):275–283. Seregard S. 1998. Conjunctival melanoma. Surv Ophthalmol 42(4):321–350. Shahbazi M, Pravica V, et al. 2002. Association between functional polymorphism in EGF gene and malignant melanoma. Lancet 359(9304):397– 401. Shanley SM, Chenevix-Trench G, et al. 1995. Glutathione S-transferase GSTM1 null genotype is not overrepresented in Australian patients with nevoid basal cell carcinoma syndrome or sporadic melanoma. Carcinogenesis 16(8):2003–2004. Shore RE. 2001. Radiation-induced skin cancer in humans. Med Pediatr Oncol 36(5):549–554. Shors AR, Iwamoto S, et al. 2002. Relationship of uveal and cutaneous malignant melanoma in persons with multiple primary tumors. Int J Cancer 102(3):266–268. Shors AR, Solomon C, et al. 2001. Melanoma risk in relation to height, weight, and exercise (United States). Cancer Causes Control 12(7):599– 606. Sigurdson AJ, Doody MM, et al. 2003. Cancer incidence in the US radiologic technologists health study, 1983–1998. Cancer 97(12):3080–3089. Singh AD, De Potter P, et al. 1998. Lifetime prevalence of uveal melanoma in white patients with oculo(dermal) melanocytosis. Ophthalmology 105(1): 195–198. Singh AD, Donoso LA. 1993. Genetic aspects of uveal melanoma. Int Ophthalmol Clin 33(3):47–52. Singh AD, Rennie IG, et al. 2004. Sunlight exposure and pathogenesis of uveal melanoma. Surv Ophthalmol 49(4):419–428. Singh AD, Wang MX, et al. 1996. Familial uveal melanoma, III. Is the occurrence of familial uveal melanoma coincidental? Arch Ophthalmol 114(9):1101–1104. Sinilnikova OM, Egan KM, et al. 1999. Germline brca2 sequence variants in patients with ocular melanoma. Int J Cancer 82(3):325–328. Siskind V, Aitken J, et al. 2002. Sun exposure and interaction with family history in risk of melanoma, Queensland, Australia. Int J Cancer 97(1): 90–95. Sliney DH. 1995. UV radiation ocular exposure dosimetry. J Photochem Photobiol B 31(1–2):69–77. Smalley KS. 2003. A pivotal role for ERK in the oncogenic behaviour of malignant melanoma? Int J Cancer 104(5):527–532. Smalley KS, Herlyn M. 2004. Loitering with intent: New evidence for the role of BRAF mutations in the proliferation of melanocytic lesions. J Invest Dermatol 123(4):xvi–xvii. Smith MA, Fine JA, et al. 1998. Hormonal and reproductive influences and risk of melanoma in women. Int J Epidemiol 27(5):751–757. Soengas MS, Capodieci P, et al. 2001. Inactivation of the apoptosis effector Apaf-1 in malignant melanoma. Nature 409(6817):207–211. Sont WN, Zielinski JM, et al. 2001. First analysis of cancer incidence and occupational radiation exposure based on the National Dose Registry of Canada. Am J Epidemiol 153(4):309–318. Soufir N, Avril MF, et al. 1998. Prevalence of p16 and CDK4 germline mutations in 48 melanoma-prone families in France. The French Familial

1228

PART IV: CANCER BY TISSUE OF ORIGIN

Melanoma Study Group [published erratum appears in Hum Mol Genet 1998 May;7(5):941]. Hum Mol Genet 7(2):209–216. Soufir N, Ribojad M, et al. 2002. Germline and somatic mutations of the INK4a-ARF gene in a xeroderma pigmentosum group C patient. J Invest Dermatol 119(6):1355–1360. Speicher MR, Prescher G, et al. 1994. Chromosomal gains and losses in uveal melanomas detected by comparative genomic hybridization. Cancer Res 54(14):3817–3823. Spitz MR, Sider JG, et al. 1990. Salivary gland cancer and risk of subsequent skin cancer. Head Neck 12(3):254–256. Stam-Posthuma JJ, Vink J, et al. 1998. Effect of topical tretinoin under occlusion on atypical naevi. Melanoma Res 8(6):539–548. Stang A, Ahrens W, et al. 2003. Phenotypical characteristics, lifestyle, social class and uveal melanoma. Ophthalmic Epidemiol 10(5):293–302. Stang A, Ahrens W, et al. 2003. Cooking and uveal melanoma risk: Results from two German case-control studies. Cancer Causes Control 14(4): 377–380. Stang A, Anastassiou G, et al. 2001. The possible role of radiofrequency radiation in the development of uveal melanoma. Epidemiology 12(1):7–12. Stern RS, Liebman EJ, et al. 1998. Oral psoralen and ultraviolet-A light (PUVA) treatment of psoriasis and persistent risk of nonmelanoma skin cancer. PUVA Follow-up Study. J Natl Cancer Inst 90(17):1278–1284. Stevens NG, Liff JM, et al. 1990. Plantar melanoma: Is the incidence of melanoma of the sole of the foot really higher in blacks than whites? Int J Cancer 45(4):691–693. Strange RC, Ellison T, et al. 1999. Cytochrome P450 CYP2D6 genotypes: Association with hair colour, Breslow thickness and melanocyte stimulating hormone receptor alleles in patients with malignant melanoma. Pharmacogenetics 9(3):269–276. Sturm RA, Frudakis TN. 2004. Eye colour: Portals into pigmentation genes and ancestry. Trends Genet 20(8):327–332. Sun S, Pollock PM, et al. 1997. CDKN2A mutation in a non-FAMMM kindred with cancers at multiple sites results in a functionally abnormal protein. Int J Cancer 73(4):531–536. Swerdlow AJ. 1983. Epidemiology of eye cancer in adults in England and Wales, 1962–1977. Am J Epidemiol 118(2):294–300. Swerdlow AJ, Cooke KR, et al. 1995. Cancer incidence in England and Wales and New Zealand and in migrants between the two countries. Br J Cancer 72(1):236–243. Swerdlow AJ, English J, et al. 1986. Benign melanocytic naevi as a risk factor for malignant melanoma. Br Med J (Clin Res Ed) 292(6535):1555–1559. Swerdlow AJ, English JS, et al. 1995. The risk of melanoma in patients with congenital nevi: A cohort study. J Am Acad Dermatol 32(4):595–599. Tabenkin H, Tamir A, et al. 1999. A case-control study of malignant melanoma in Israeli kibbutzim. Isr Med Assoc J 1(3):154–157. Tachibana I, Smith JS, et al. 2000. Investigation of germline PTEN, p53, p16(INK4A)/p14(ARF), and CDK4 alterations in familial glioma. Am J Med Genet 92(2):136–141. Telle-Lamberton M, Bergot D, et al. 2004. Cancer mortality among French Atomic Energy Commission workers. Am J Ind Med 45(1):34–44. Theriault G, Goldberg M, et al. 1994. Cancer risks associated with occupational exposure to magnetic fields among electric utility workers in Ontario and Quebec, Canada, and France: 1970–1989. Am J Epidemiol 139(6):550–572. Tucker MA, Boice JD, et al. 1985. Second cancer following cutaneous melanoma and cancers of the brain, thyroid, connective tissue, bone, and eye in Connecticut, 1935–1982. Multiple Primary Cancers in Connecticut and Denmark. Bethesda, MD: National Cancer Institute. Monograph 68:161–189. Tucker MA, Fraser MC, et al. 1993. Risk of melanoma and other cancers in melanoma-prone families. J Invest Dermatol 100(3):350S–355S. Tucker MA, Fraser MC, et al. 2002. A natural history of melanomas and dysplastic nevi: An atlas of lesions in melanoma-prone families. Cancer 94(12):3192–3209. Tucker MA, Goldstein AM. 2003. Melanoma etiology: Where are we? Oncogene 22(20):3042–3052. Tucker MA, Hartge P, et al. 1986. Epidemiology of intraocular melanoma. Recent Results Cancer Res 102:159–165. Tucker MA, Shields JA, et al. 1985. Sunlight exposure as risk factor for intraocular malignant melanoma. N Engl J Med 313(13):789–792. Tuomaala S, Kivela T. 2003. Conjunctival melanoma: Is it increasing in the United States? Am J Ophthalmol 136(6):1189–1190; author reply 1190. Tynes T, Klaeboe L, et al. 2003. Residential and occupational exposure to 50 Hz magnetic fields and malignant melanoma: A population based study. Occup Environ Med 60(5):343–347. Tynes T, Reitan JB, et al. 1994. Incidence of cancer among workers in Norwegian hydroelectric power companies. Scand J Work Environ Health 20(5):339–344.

UNSCEAR. 2000. United Nations Scientific Committee on the Effects of Atomic Radiation. Sources and effects of ionizing radiation UNSCEAR 2000 report to the General Assembly: New York, United Nations: 328–330. Vagero D, Swerdlow AJ, et al. 1990. Occupation and malignant melanoma: A study based on cancer registration data in England and Wales and in Sweden. Br J Ind Med 47(5):317–324. Vajdic C, Kricker A, et al. 2003. Ocular melanoma is not associated with CDKN2A or MC1R variants—a population-based study. Melanoma Res 13(4):409–413. Vajdic CM, Kricker A, Giblin M, McKenzie J, Aitken JF, Giles GG, Armstrong BK. 2004. Artificial ultraviolet radiation and ocular melanoma in Australia. Int J Cancer 112(5):896–900. Vajdic CM, Hutchins AM, et al. 2003. Chromosomal gains and losses in ocular melanoma detected by comparative genomic hybridization in an Australian population-based study. Cancer Genet Cytogenet 144(1): 12–17. Vajdic CM, Kricker A, et al. 2001. Eye color and cutaneous nevi predict risk of ocular melanoma in Australia. Int J Cancer 92(6):906–912. Vajdic CM, Kricker A, et al. 2002. Sun exposure predicts risk of ocular melanoma in Australia. Int J Cancer 101(2):175–182. Vajdic CM, Kricker A, et al. 2003. Incidence of ocular melanoma in Australia from 1990 to 1998. Int J Cancer 105(1):117–122. Valverde P, Healy E, et al. 1995. Variants of the melanocyte-stimulating hormone receptor gene are associated with red hair and fair skin in humans [see comments]. Nat Genet 11(3):328–330. Valverde P, Healy E, et al. 1996. The Asp84Glu variant of the melanocortin 1 receptor (MC1R) is associated with melanoma. Hum Mol Genet 5(10):1663–1666. van der Velden PA, Sandkuijl LA, et al. 2001. Melanocortin-1 receptor variant R151C modifies melanoma risk in Dutch families with melanoma. Am J Hum Genet 69(4):774–779. van Dijk M, Sprenger S, et al. 2003. Distinct chromosomal aberrations in sinonasal mucosal melanoma as detected by comparative genomic hybridization. Genes Chromosomes Cancer 36(2):151–158. Vasen HF, Gruis NA, et al. 2000. Risk of developing pancreatic cancer in families with familial atypical multiple mole melanoma associated with a specific 19 deletion of p16 (p16-Leiden). Int J Cancer 87(6):809–811. Wagner JK, Parra EJ, et al. 2002. Skin responses to ultraviolet radiation: Effects of constitutive pigmentation, sex, and ancestry. Pigment Cell Res 15(5):385–390. Wallace DC, Exton LA, et al. 1971. Genetic factor in malignant melanoma. Cancer 27(5):1262–1266. Walter SD, Marrett LD, et al. 1990. The association of cutaneous malignant melanoma with the use of sunbeds and sunlamps. Am J Epidemiol 131(2):232–243. Weinstock MA, Brodsky GL. 1998. Bias in the assessment of family history of melanoma and its association with dysplastic nevi in a case-control study. J Clin Epidemiol 51(12):1299–1303. Weinstock MA, Rossi JS, et al. 2002. Randomized controlled community trial of the efficacy of a multicomponent stage-matched intervention to increase sun protection among beachgoers. Prev Med 35(6):584–592. Wennborg H, Yuen J, et al. 2001. Cancer incidence and work place exposure among Swedish biomedical research personnel. Int Arch Occup Environ Health 74(8):558–564. Westerdahl J, Olsson H, et al. 1996. Risk of malignant melanoma in relation to drug intake, alcohol, smoking and hormonal factors. Br J Cancer 73(9):1126–1131. Whelan AJ, Bartsch D, et al. 1995. Brief report: A familial syndrome of pancreatic cancer and melanoma with a mutation in the CDKN2 tumorsuppressor gene. N Engl J Med 333(15):975–977. White E, Kirkpatrick CS, et al. 1994. Case-control study of malignant melanoma in Washington State. I. Constitutional factors and sun exposure. Am J Epidemiol 139(9):857–868. Whiteman D, Green A. 1998. History of tropical military service and risk of primary cutaneous melanoma in Queensland men. Melanoma Res 8(1):63–66. Whiteman DC, Watt P, et al. 2003. Melanocytic nevi, solar keratoses, and divergent pathways to cutaneous melanoma. J Natl Cancer Inst 95(11):806– 812. WHO. 2001. World Health Organization Mortality Database, WHO: http:/www-dep.iarc.fr/dataava/infodata.htm. Winsey SL, Haldar NA, et al. 2000. A variant within the DNA repair gene XRCC3 is associated with the development of melanoma skin cancer. Cancer Res 60(20):5612–5616. Wolf CR, Smith CA, et al. 1992. Relationship between the debrisoquine hydroxylase polymorphism and cancer susceptibility. Carcinogenesis 13(6):1035–1038.

Cutaneous and Ocular Melanoma Youl P, Aitken J, et al. 2002. Melanoma in adolescents: A case-control study of risk factors in Queensland, Australia. Int J Cancer 98(1):92–98. Young P, Purdie D, et al. 2001. A study of infertility treatment and melanoma. Melanoma Res 11(5):535–541. Yu GP, Hu DN, et al. 2003. Conjunctival melanoma: Is it increasing in the United States? Am J Ophthalmol 135(6):800–806. Zanetti R, Prota G, et al. 2001. Development of an integrated method of skin phenotype measurement using the melanins. Melanoma Res 11(6):551– 557.

1229

Zeigler-Johnson C, Panossian S, et al. 2004. Population differences in the frequency of the agouti signaling protein g.8818a>G polymorphism. Pigment Cell Res 17(2):185–187. Ziogas A, Anton-Culver H. 2003. Validation of family history data in cancer family registries. Am J Prev Med 24(2):190–198. Zuo L, Weger J, et al. 1996. Germline mutations in the p16INK4a binding domain of CDK4 in familial melanoma. Nature Genetics 12(1):97–99.

64

Keratinocyte Carcinomas (Basal and Squamous Cell Carcinomas of the Skin) MARGARET R. KARAGAS, MARTIN A. WEINSTOCK, AND HEATHER H. NELSON

C

ancers arising from keratinocytes or their precursors, which include basal cell carcinoma (BCC) and squamous cell skin carcinoma (SCC) of the skin, are the mostly frequently diagnosed malignancies in fair-skinned populations. These malignancies are commonly referred to as nonmelanoma skin cancers or simply skin cancers and, under the International Classification for Oncology (ICDO) system, are classified as “other skin cancers” or C44. In the United States, the estimated number of keratinocyte cancers are roughly equal to all other malignancies combined (Miller and Weinstock, 1994; Jemal et al., 2003), and in several geographic regions incidence rates are increasing rapidly. Thus, they account for appreciable health care expenditures (Joseph et al., 2001). Keratinocyte cancers (KC) are responsible for significant morbidity and disfigurement in the population because they can invade and destroy local tissues and often occur around facial structures such as the nose, ear, and eye. While highly treatable, the estimated number of deaths from KC approximate or surpass those of other treatable or rare malignancies in the United States (e.g., cancers of thyroid, bone, and testes, and Hodgkin disease) (Jemal et al., 2003). In addition, individuals diagnosed with a KC are at an enhanced risk of internal malignancies (Frisch et al., 1995; Frisch et al., 1996; Karagas et al., 1998) as well as additional skin cancers (Robinson et al., 1987; Karagas et al., 1992; Berwick, 1999). This cannot be fully explained by heightened surveillance since these individuals have higher overall cancer mortality rates as well (Kahn et al., 1998). Individuals with KCs, may experience worse survival rates from other malignancies (e.g., non-Hodgkin lymphoma) for reasons as yet unknown (Askling et al., 1999; Hjalgrim et al., 2000). Basal cell and squamous cell skin carcinomas can be further classified histologically, but epidemiologic investigations have not typically done so. Among the most common subtypes of BCC are the superficial multicenteric and nodular BCCs, defined by their clinical and histopathologic features, and the more aggressive subtypes such as morphea and infiltrative/micronodular forms. Older terms for BCC include basal cell epithelioma, reflecting its relatively indolent nature, and rodent ulcer, derived from its clinical presentation. SCCs may be divided into subtypes usually by invasiveness (i.e., in-situ carcinoma, also called Bowen disease, and invasive carcinoma), degree of differentiation, and depth of invasion into the skin. Histologic subtypes also may be designated for some SCC tumors (i.e., verrucous carcinomas). No precursor lesion has been established for BCC, but SCCs may arise in an actinic keratosis (also known as a solar keratosis), or in actinic chelitis. In special circumstances, SCCs may arise from warts, and both BCCs and SCCs may arise rarely from other lesions such as chronic draining sinuses of osteomyelitis or scars from thermal burns. BCC and SCC constitute the vast majority with nonmelanoma skin cancers; rarer forms include apocrine and sebaceous gland carcinomas, Kaposi sarcoma, cutaneous lymphomas, and Merkel cell carcinoma (Weinstock, 1994), each of which have distinct etiologies. Genital SCCs also are often excluded from consideration of epidemiologic features of SCC of the skin. Exposure to solar ultraviolet light (UVL) has dominated etiologic inquiry over the past several decades, but it is clear that other factors play a role in KC carcinogenesis. Indeed, the discovery of many human carcinogens emerged from observations of KCs in either the occupational or clinical setting. These include ionizing radiation, polycyclic aromatic hydrocarbons, arsenic, and chronic immunosuppres-

1230

sion. An early link between papillomaviruses and human cancers also arose from the detection of HPV in the skin cancers of patients with epidermodysplasia verruciformis (Orth et al., 1980). Xeroderma pigmentosum, a genetic disorder characterized by an inability to repair DNA photolesions, is one of the classic examples of gene-environment interaction. Thus, in a sense, KC provides an important “human model” of carcinogenesis, paralleling its historical role in animal experiments.

DEMOGRAPHIC PATTERNS Sources of Data Unlike other cancers, the incidence of KC is not known for most regions of the world. Despite a critical need for these data (National Institutes of Health, 1991; Lawrence et al., 1994; Weinstock, 1994) logistical challenges and relatively favorable prognosis often prevent inclusion of these malignancies in central tumor registries (Gallagher et al., 1990) including the US Surveillance Epidemiology and End Results (SEER) program. Aside from the overwhelming number of cases, KCs are rarely treated in hospitals—an important source of central cancer registry data. Pathology laboratories cannot always provide complete coverage of new cases because physicians may perform their own histopathology review or contract with laboratories outside the region. Therefore complete coverage of KCs, at least in the United States, requires special efforts. In regions with a high prevalence of KCs (e.g., Australia), the number of undiagnosed lesions may be large enough to necessitate medical examination and uniform follow-up of the population at large to estimate incidence rates. During the 1970s, two special surveys were conducted in the United States as part of a joint effort of the US Environmental Protection Agency and National Cancer Institute (Scotto et al., 1974; Scotto et al., 1981). These data remain a major source of information on BCC and SCC incidence for the United States. The first survey covered a 6-month period in 1971–1972 in selected areas participating in the Third National Cancer Survey, including Dallas-Fort Worth, San Francisco-Oakland, Iowa, and Minneapolis-St. Paul (Scotto et al., 1974). In 1977–1978, the National Cancer Institute repeated the survey for a 1-year period in Minneapolis-St. Paul and San FranciscoOakland along with six other regions participating in the SEER Program: Detroit, New Mexico, New Orleans, Seattle, Atlanta, and Utah (Scotto et al., 1981). An extension of the survey in 1979–1980 covered New Hampshire-Vermont and San Diego (Scotto and Fears, 1987; Serrano et al., 1991). These data were used to provide estimates of the number of new KC occurrences in the United States, and in the mid 1990s this estimate was one million new cases of KC per year (Miller and Weinstock, 1994). However, an analysis of Medicare records at that time estimated 789,000 new KC cases per year among US Medicare enrollees alone (i.e., those 65 years of age and older), suggesting the one million figure may have been an underestimate (Joseph et al., 2001). Over the past two decades, seven studies computed incidence rates of BCC and SCC by gender in North America (Table 64–1) and at least eight studies from North America, Europe, and Australia estimated time trends (Table 64–2). In the United States, southeastern Arizona, northcentral New Mexico, and the state of New Hampshire took a “registry” approach of active surveillance of pathology laboratories and dermatology practices (Karagas et al., 1999; Harris et al., 2001; Athas

1231

Keratinocyte Carcinomas

Table 64–1. Incidence of Basal Cell and Squamous Cell Carcinomas among Whites in the US and Canada in the 1980s and 1990s, According to Latitude Basal Cell Area (Reference)

Latitude

Southeastern Arizona‡ (Harris et al., 2001) New Mexico§ (Hoy, 1996) New Mexico (Athas et al., 2003) New Hampshire¶ (Karagas et al., 1999) Rochester, Minnesota¶ (Gray et al., 1997) Portland, Oregon¶ (Glass and Hoover, 1989) British Columbia (Gallagher et al., 1990)

Year

Squamous Cell

BCC TO SCC Ratio

N Cases

Rate* Men

Rate* Women

M : F† Ratio

N Cases

Rate* Men

Rate* Women

M : F† Ratio

men

women

Standard Reference Population

364.2

153.5

2.4

3.5

4.5

1970 US

50

4.3

5.0

8.3

1970 US

31–32°N

1996

8243

1268.7

691.6

1.8

2285

35°N

1989–1991

4287

1073

415

2.6

938

35°N

1998–1999

3103

930.3

485.5

1.9

1091

356.2

150.4

2.4

2.6

3.2

2000 US

43–45°N

1993–1994

2897

309.9

165.5

1.9

779

97.2

32.4

3.0

3.2

5.1

1970 US

45°N

1984–1992

—

—

—

—

511

155.5

71.2

2.2

—

—

1990 US

46°N

1980–1986

—

—

—

—

1111

106.1

29.8

3.6

—

—

1970 US

49–60°N

1987

4152

1.3

963

31

17

1.8

3.9

5.4

1971 Can.

120

92

214

*Incidence rate per 100,000 persons per year. † Calculated as the ratio of the age standardized rates. ‡ Number of cases includes non-whites. § Estimated by the number of subjects multiplied by the percentage of subjects with each cell type. ¶ <10% of population non-white.

et al., 2003). A special survey from the United Kingdom also followed this strategy (Ko et al., 1994). Studies from Portland, Oregon (Glass and Hoover, 1989), Albuquerque, New Mexico (Hoy, 1996) and Rochester, Minnesota (Gray et al., 1997) used medical records including pre-paid health plan data. Differences in rates from population registries and pre-paid health plans could reflect socioeconomic status differences or other characteristics of the populations. Cancer registries were used to ascertain KC data in other countries, including British Columbia, Canada (Gallagher et al., 1990), Finland (HannukselaSvahn et al., 1999), and Sweden (for squamous cell carcinomas) (Wassberg et al., 2001). A study in Australia used a national household survey to identify skin cancer diagnoses (Staples et al., 1998).

Age, Sex, Cell Type, and Anatomic Site Distributions Basal cell carcinoma alone is by far the most common malignancy in fair-skinned populations, comprising 70%–80% of the KCs in men and 80%–90% in women among whites in the United States. The shape of age incidence curves for BCC and SCC follow a similar pattern across populations, with SCC having a later age at onset and incidence rates that increase more steeply with age than BCC (Harris et al., 2001 and Karagas et al., 1999 shown in Figures 64–1, 64–2, 64–3, 64–4, and Staples et al., 1998; Hannuksela-Svahn et al., 1999). Men have one to two times the incidence of BCC and two to four times the incidence of SCC compared with women (Harris et al., 2001; Hoy, 1996; Athas

Table 64–2. Recent Trends in Incidence Rates of Basal Cell Carcinoma and Squamous Cell Carcinoma of Skin in North America, Europe, and Australia Annual Percent Change Basal Cell Carcinoma Population

Years

Squamous Cell Carcinoma

Men

Women

Men

Women

Data Source Mayo Clinic and the Olmsted Medical Center and affiliated hospitals (Gray et al., 1997) Population survey for state (Karagas, et al., 1999) Population survey of Conchise, Pima, and Santa Cruz counties (Harris et al., 2001)

north america Minnesota*

1984–1986, 1990–1992

–

–

9%

19%

New Hampshire*

1979–1980, 1993–1994

4.4%

4.4%

9%

11%

Southeastern Arizona†

1985 through 1996

1.5%

2.0%

-2.8%

-2.6%

Northcentral New Mexico

1977–1978, 1998–1999

2.4%

1.0%

4.3%

5.2%

Population survey for Bernalillo, Sandoval, and Santa Fe counties (Athas et al., 2003)

9.2

9.9%

12.8%

3.2%

9.4%

3.2%

11.6%

–

–

4.3%

4.0%

3.0%

0.6%

10.0%

8.7%

Population survey for North Humberside (Ko et al., 1994) Finnish Cancer Registry (Hannuksela-Svahn et al., 1999) Swedish Cancer Registry (Wassberg et al., 2001) Household survey (Staples et al., 1998)

europe UK†

1978, 1980, 1984, 1987 through 1991 1976–1980, 1981–1985, 1986–1990, 1991–1995 1981 through 1995

†

Finland

Sweden‡

australia

‡

1985, 1995

10.5

*Method of computing the annual percent change = relative increase in incidence rates between two periods exponentiated to the power of 1/n, where n is the number of years between the midpoints of the two periods. The formula assumes the log rate increases linearly with time. † The average relative increase (as defined above) for the years covered. ‡ Taken from the text.

1232

PART IV: CANCER BY TISSUE OF ORIGIN 8000

3000

2500

SE Arizona New Hampshire

2000

1500

1000

500

0 0

Incidence rate per 100,000 persons per year

Incidence rate per 100,000 persons per year

3500

7000 6000 SE Arizona 5000 4000 3000 2000 1000 0 0

10 20 30 40 50 60 70 80 90 100 Age, years

New Hampshire

10 20 30 40 50 60 70 80 90 100 Age, years

Figure 64–1. Age- and sex-specific incidence rates of basal cell carcinoma among women in Arizona, 1996 (Harris et al., 2001) and New Hampshire, 1993–1994. (Source: Karagas et al., 1999.)

Figure 64–2. Age- and sex-specific incidence rates of basal cell carcinoma among men in Arizona, 1996 (Harris et al., 2001) and New Hampshire, 1993–1994. (Source: Karagas et al., 1999.)

et al., 2003; Karagas et al., 1999; Gray et al., 1997; Glass and Hoover, 1989; Gallagher et al., 1990 shown in Table 64–1 and Staples et al., 1998; Hannuksela-Svahn et al., 1999; Wassberg et al., 2001). Detailed anatomic site data are available from two recent US studies (Figs. 64–5 and 64–6). The majority of BCC and SCC tumors appear on the face, head, and neck. The trunk accounts for about 25% of BCC tumors in men and 15% in women, but is a less common site for SCC tumors. Tumors on the ear develop more frequently in men than women, and tumors on the legs more often in women than men. Taking into account body surface area, incidence rates vary over two orders of magnitude by anatomic site, with the highest rates of BCC and SCC

tumors on the face, particularly the nose and eye region, the ears in men, and the cheeks/mouth area in men and women (Pearl and Scott, 1986; Franceschi et al., 1996; English et al., 1997; Wassberg et al., 2001). Compared with the 1970s (Scotto et al., 1981), data from the 1990s indicate the proportion of tumors diagnosed on the trunk and lower extremities has doubled, and tumors on the upper extremities increased by 30%–50% (Figs. 64–5 and 64–6). A number of studies report a disproportionate increase in the incidence of BCC on the trunk including data from the United States, Canada, and Australia (Scotto et al., 1981; Gallagher et al., 1990; Marks et al., 1993; Karagas et al., 1999), and

3000

1000

800

SE Arizona Rochester, MN New Hampshire Portland, OR

600

400

200

0 0

10 20 30 40 50 60 70 80 90 100 Age, years

Figure 64–3. Age- and sex-specific incidence rates of squamous cell carcinoma among women in Arizona, 1996 (Harris et al., 2001), Rochester, MN, 1984–1992 (Gray et al., 1997), New Hampshire, 1993–1994 (Karagas et al., 1999), and Portland, OR, 1960–1986. (Source: Glass and Hoover, 1989.)

Incidence rate per 100,000 persons per year

Incidence rate per 100,000 persons per year

1200

2500

2000

SE Arizona Rochester, MN New Hampshire Portland, OR

1500

1000

500

0 0

10 20 30 40 50 60 70 80 90 100 Age, years

Figure 64–4. Age- and sex-specific incidence rates of squamous cell carcinoma among men in Arizona, 1996 (Harris et al., 2001), Rochester, MN, 1984–1992 (Gray et al., 1997), New Hampshire, 1993–1994 (Karagas et al., 1999), and Portland, OR, 1960–1986. (Source: Glass and Hoover, 1989.)

Keratinocyte Carcinomas Men

1233

Women

Figure 64–5. Anatomic site distribution of basal cell carcinoma in population-based series from New Hampshire (NH) 1993– 1994 and Arizona (AZ) 1996.

some studies also observe this trend for SCC (Scotto et al., 1981; Karagas et al., 1999). In Sweden, a particularly striking increase in SCC rates was found for sites normally covered or unexposed to the sun (Hemminki et al., 2003).

International Patterns, Ethnicity, and Migration Worldwide, KC incidence rates vary over 100-fold in White populations. Australia reports the highest rates in the world (e.g., incidence rates for BCC of 7067 and 3379 per 100,000 persons per year and for SCC of 775 and 501 per 100,000 persons per year in men and women, respectively) (English et al., 1997). The lowest reported rates among populations of European descent come from Finland with rates for BCC of 49 and 45 per 100,000 persons per year and rates for SCC of 8.7 and 5.3 per 100,000 persons per year in men and women, respectively (Hannuksela-Svahn et al., 1999). There may be regions with even lower rates where data are either unavailable or incomplete. While lack of standardized reporting methods and ethnic differences in sun sensitivity complicate international comparisons, available data indicate a distinct gradient of increasing rates with proximity to the equator. An analysis of US data from the 1970s revealed a clear decline in BCC and SCC rates with increasing latitude (Scotto et al., 1981), and this trend remains evident in the recent data (Table 64–2). Data from Australia similarly demonstrated a latitude gradient (Staples et al., 1998). The lowest rates of KCs occur in populations of largely nonEuropean descent (e.g., India, Asia, and Africa) (IARC, 2002; IARC, 2002; Koh et al., 2003). For instance, in Zimbabwe, rates of “other skin cancers” (i.e., ICDO-C44) in men are 190-fold lower in Africans

than Europeans, and in women are 80–145-fold lower (IARC, 2002). Among Asians in Singapore, BCC rates were more than 2.5-fold higher among the Chinese than the Malays and Indians and SCC rates were about two-fold higher (Koh et al., 2003). The limited data on non-white populations in the United States also suggest racial/ethnic variation in KC incidence rates. In Arizona and New Mexico, BCC rates among Hispanics were as low as 1/14 the rates in non-Hispanic Whites among men and women (Table 64–3). SCC rates among Hispanics were as low as 1/11 the rates of non-Hispanic whites (Table 64–3). Among African Americans, the rates vary from 1/14 to 1/46 of the European American rates for SCC and from 1/105 to 1/190 for BCC (Table 64–3), suggesting more marked racial difference for BCC than SCC. It is unfortunate that more recent data on KCs among African Americans do not exist.

Time Trends The incidence rates of KCs may be increasing more rapidly than any other malignancy, but lack of consistent registration of these cancers makes inferences tentative. US surveys done in the 1970s found a 1%–2% per year average increase in age-adjusted SCC incidence rates and a 3% increase in BCC incidence rates (Scotto et al., 1981). Studies since then report increases in BCC rates ranging from 1%–11% (Table 64–2). Increases in SCC rates range from 3%–19%, with the exception of a decline in Arizona (Table 64–2). The reason for the rising incidence rates remains somewhat unclear. Certainly, it could be an artifact of enhanced detection due to public awareness of skin cancers bringing them to medical attention, more

1234

PART IV: CANCER BY TISSUE OF ORIGIN Men

Women

Figure 64–6. Anatomic site distribution of squamous cell carcinoma in populationbased series from New Hampshire (NH) 1993–1994 and Arizona (AZ) 1996.

aggressive treatment (i.e., removal of lesions previously left untreated), classification of lesions as malignant that were previously considered benign, and for registries, increased reporting. The chance that physicians incorrectly assigned a diagnosis of BCC or SCC to non-malignant conditions is relatively small because virtually all

studies included histopathologically confirmed diagnoses, and most restricted the analysis to invasive lesions (e.g., excluding in situ SCC or Bowen disease). Studies of the reliability of histopathologic diagnosis find a high degree of concordance between pathologists for invasive tumors (Greenberg et al., 1990; Wechsler et al., 2001). However,

Table 64–3. Racial/Ethnic Differences in Incidence Rates of Basal Cell Carcinoma and Squamous Cell Carcinoma of Skin in United States* Age Standardized Incidence Rate (per 100,000 persons per year) Area

Year

Whites

Hispanics

1996 1989–1991 1977–1980 1996 1989–1991 1977–1980

153.5 50 35.4 364.2 214 101.1

13.8 21 7.9 32.9 21 10

1996 1989–1991 1977–1980 1996 1989–1991 1977–1980

691.6 113 210.7 1268.7 1073 360.1

50.3 92 37.2 91.1 171 51.6

Ratios

Blacks

White to Hispanic

White to Black

squamous cell carcinoma Women Men

AZ NM US AZ NM US

2.5 2.2

11.1 2.4 4.5 11.1 10.2 10.1

14.2 46.0

basal cell carcinoma Women Men

AZ NM US AZ NM US

*Data derived from two reports (Scotto et al., 1981; Harris et al., 2001).

2 1.9

13.7 1.2 5.7 13.9 6.3 7.0

105.4 189.5

Keratinocyte Carcinomas a steeper increase in in-situ rather than invasive SCC in the Swedish Cancer Registry could reflect greater detection or reporting of in-situ lesions (Hemminki et al., 2003). Obtaining the “true” incidence rate of KCs over time would require repeated, standardized examinations of the population. Investigators from Queensland, Australia, monitored skin cancer occurrences over a 10-year period among participants in a prevention trial (n = 673) (Valery et al., 2004). They found that the incidence rates of BCC were three-fold higher during the intervention period, particularly at the time intervals when a standardized dermatologic evaluation took place and for tumors in anatomic locations not ordinarily visible (e.g., sites other than the head and neck, forearms, and hands). In contrast, rates of SCC were unaffected (Valery et al., 2004). The US National Health Examination and Nutrition Survey conducted in 1971–1974 detected a prevalence of BCC of 0.1% among those without signs of actinic damage (Engel et al., 1988), making undiagnosed lesions a less likely explanation for increasing BCC rates observed in the US population. Traditional analyses to assess the impact of enhanced detection and screening on incidence trends (i.e., by tumor size and level of invasion) have not been conducted for KCs. Although it is difficult to rule out improvements in detection as an explanation for an apparent rise in KC incidence rates, increasing UVL exposure must be considered as a likely contributor. This particularly would include trends in sun seeking behavior (i.e., sun bathing, sparse clothing, and outdoor leisure time). Further, skin cancer incidence rates are predicted to increase well into the 21st century from stratospheric ozone depletion resulting from halocarbon emissions (Slaper et al., 1996).

1235

and all other nonmelanoma skin cancers (Lewis and Weinstock, 2004). A second study in Finland from 1991 to 1995 noted mortality rates (age-adjusted to the World standard) of 3.8 and 2.3 per 100,000 persons per year in men and women respectively for all skin cancers other than melanoma and BCC, and rates of 0.8 and 0.5 per 100,000 persons per year in men and women respectively for BCC (Hannuksela-Svahn et al., 1999). It should be noted that deaths from melanoma outnumber those from KC in general, but non-genital KC deaths are greater than melanoma deaths after age 85 years and among darkerskinned groups (Weinstock, 1997). In the United States, Canada, Europe, and Australia, data regarding darker-skinned racial groups is characterized by statistical instability due to small numbers of deaths from this cause, preventing detailed inferences. Population-based reports of mortality generally show a declining trend. Some of the older data (not adjusted for misclassification) indicated that nonmelanoma deaths exceeded those from melanoma (Gordon and Silverstone, 1969). Early assessments (Holman, 1982) date to the 1930s and suggested that mortality may have increased until 1950, but they were unable to age-adjust their earliest data, and did not take account of misclassification. Inability to account for misclassification also characterized other analyses from Australia, New Zealand, and North America (Gordon and Silverstone, 1969; Elwood et at., 1974; Armstrong, 1988; Giles et al., 1989; Cooke et al., 1991). More recent data has indicated a 20%–30% fall in mortality in the United States for both men and women in the 1970s and 1980s (Weinstock, 1993), and a further decline in the following decade (Lewis and Weinstock, 2005). A substantial decline also has been noted in Finland, through 1995, at least since the 1960s (Hannuksela-Svahn et al., 1999), and in Australia (Giles et al., 1989).

Mortality Mortality from KC has not been the subject of extensive study. The possibilities for detailed study have been limited by the low case fatality of KC, the lack of distinction among the different forms of skin cancer (other than melanoma) under the International Classification of Diseases system, the exclusion of KC from most population-based cancer registries, and the documented poor accuracy of the routine death certification coding under that system for the (ICD-9) 173 (“nonmelanoma skin cancer (NMSC)”) rubric. This misclassification has been studied in multiple contexts (Dunn et al., 1965; Gordon and Silverstone, 1969; Osterlind et al., 1991; Weinstock et al., 1992; Rosenblatt and Marks, 1996; Lewis and Weinstock, 2004). Furthermore, carcinoma of the genital skin is generally excluded from studies of KC mortality, and it is coded separately by the ICD-9. Despite these handicaps, our ignorance is not complete; some informative data has been collected, although interpretation must be made with the abovementioned limitations in mind. The ratio of SCC deaths to BCC deaths is generally reported to be about 3 : 1 (Dunn et al., 1965; Gordon and Silverstone, 1969; Osterlind et al., 1991; Weinstock et al., 1991), the reverse of the incidence ratio. This is because SCCs can be aggressive and are capable of metastasizing; whereas, when BCC is fatal, death is usually due to the effects of local destruction. Men have higher mortality from KC than women, in part due to a higher mortality rate associated with cancers on the ear and perhaps because women seek medical attention before lesions progress. Mortality rates follow the expected latitude gradients as noted in Australia and North America in the past (Gordon and Silverstone, 1969; Elwood et al., 1974), and in general have persisted (Devesa et al., 1999). These nonmelanoma skin cancer statistics exclude genital SCC, which accounts for a large proportion of KC deaths. Age-adjusted (to the 1970 US standard) death rates for confirmed BCC and (nongenital) SCC were 1.0 and 2.6 per 100,000 persons per year in Rhode Island during 1979–1987, and less than 1.0 per 100,000 persons per year for all other nonmelanoma skin cancers combined (such as Merkel cell carcinoma, malignant fibrous histiocytoma, Kaposi sarcoma, adenexal carcinomas, and dermatofibrosarcoma protuberans) (Weinstock et al., 1991). A subsequent report from that study covered the period 1988 to 2000, and noted mortality rates of 2.1 and 0.5 (per 100,000 persons per year, adjusted to the 1970 standard) for nongenital SCC and BCC, and 3.2 and 0.6 for genital SCC

ENVIRONMENTAL FACTORS Ultraviolet Light Ultraviolet radiation from the sun is a well-accepted cause of BCC and SCC (IARC, 1992; Kricker et al., 1994). The anatomic pattern of occurrences on sun-exposed sites, distinct latitude gradient, rates among migrants from low to high ambient UVL regions, and observed racial/ethnic differences each point to solar exposure as the chief etiological factor. Our understanding of the manner in which solar exposure affects risk is far from complete, yet this information is critical for planning effective strategies. The relative absence of incidence monitoring of KCs has limited epidemiologic research in this area. Prospective studies on solarrelated factors include a cohort from Queensland, Australia, and the Nurses’ Health Study and Health Professionals cohorts (Table 64–4). Only four case-control studies were published before 1990 (Table 64–4), and these earlier studies derived cases and controls from dermatology clinics or from occupationally exposed workers, all with limited sample sizes. Even since 1990, the largest published study of SCC and sunlight exposure included about 200 cases, and none are a general population-based study from the US (Table 64–4). Another obstacle to defining the role of solar radiation or specifically UVL in KC pathogenesis lies in the complexity of ascertaining exposure history. One approach is simply to determine residential history and assign levels of ambient UVL to these locations. In the US Nurses’ Health Study women who were residents of California or Florida had a 1.6-fold higher risk of BCC and two-fold higher risk of SCC than women living in New England (Hunter, 1990); for SCC, risk was especially high (2.5–3.0 fold) for living in the warmer states from birth to age 15 years (Grodstein et al., 1995). Among health professionals, risks of BCC were highest among men who currently or always lived in areas with greater ambient UVL, but there was no excess risk for living in high UVL areas solely as a child (van Dam et al., 1999). In contrast, in western Australia, migrants (mainly from the UK, Ireland, and Northern Europe) had one-fifth or lower the risk of BCC (Kricker et al., 1991) and less than one-half the risk for SCC (English et al., 1998a) if they migrated at age 10 years or after compared with native-born Australians or those who migrated before the age of 10 years. These data provide

1236

PART IV: CANCER BY TISSUE OF ORIGIN

Table 64–4. Summary of Published Cohort and Case-Control Studies on Basal Cell Carcinoma and Squamous Cell Carcinoma of Skin in Relation to Solar-Related Factors Factor Evaluated* Reference (s) (Country)

Sample

No. BCC

Method of Case Identification

No. SCC

Size of Cohort/Controls

Where Lived P

COHORT STUDIES Green and Battistutta 1990; Green et al., 1996 (Australia) Hunter et al., 1990 (USA) Grodstein et al., 1995 (USA) van Dam et al., 1999 (USA)

Bajdik et al., 1998 (Canada) Lock-Andersen et al., 1998 (Denmark) Rosso et al., 1999 (Switzerland) Milan et al., 2003 (Finland) Kennedy et al., 2003 (Netherlands)

M T

Sun Exposure Effects

O

R

B

A

x

x

x

x

x

x

General population

Dermatologic examination of population

66 250

21 94

2,095

Women nurses

Self-reported

771

–

73,366

x

Women nurses

Verified self-reported lesions, histologically confirmed Self-reported

–

197

107,900

x

x

x

x

3273

–

44,591

x

x

x

x

Consecutive patients

861

–

1,938

x

Consecutive patients

336

58

294

x

Pathology reports

–

92

174

Cancer Registry

538

178

33

35

588

Men health professionals

CASE-CONTROL STUDIES Gellin et al., 1965 Skin and Cancer Unit (USA) cases, dermatology patient controls Vitaliano et al., Skin and Cancer 1980 (USA) Hospital Hospital-clinics in Aubry et al., 1985† (Canada) region, Dermatology controls Hogan et al., 1989; General population 1990‡ (Canada) Vitasa et al., 1990 Maryland watermen (USA) Kricker et al., 1991; 1995a; 1995b (Australia) Gafa et al., 1991 (Italy) Gallagher et al., 1995a,b (Canada) Rosso et al., 1996 (S. Europe) English et al., 1998a,b§ (Australia)

C

Sun Exposure

738 (284)

x x

x

x

x

x

108

25

266

x

x

Men, general population

Provisional Health Insurance

226

180

406

x

x

x

x

General population

Cancer Registry

1549

228

1,795

x

x

x

x

General population

Histologically confirmed lesions diagnosed at survey exam or preceeding year or during 7-year follow-up period Head and neck sites; Provisional Health Insurance Consecutive patients

–

145

1,458

x

x

x

x

–

113

342

x

x

x

x

145

–

174

Cancer Registry

120

25

144

Cancer Registry

333

–

333

x

Ophthalmology Outpatient Clinic

454

161

386

x

Men, general population Dermatology practice cases, general population controls General population Twins from the general population Dermatology Department

x

x

1,015

x

x

x

45

x

x

x

226

General population

x

x

Prevalent lesions on exam or previously removed, 62% histologically confirmed Histologically confirmed lesions detected at survey exam or preceding year Cancer Registry

General population

S D

x

x

x

x

x

x

x

x

x

x x

x

x

x

x

x

x

*Abbreviations: P = present; C = childhood/teen; M = migration; T = total; O = occupational; R = recreational; B = sunburns; A = actinine damage, i.e., actinic keratoses, solar elastosis; S = silicon mold impression, microtopography; D = color difference between sun exposed and sun protected skin. † Non-occupational sun exposure score includes sunburns. ‡ Residence information not specified as current or past; BCC cases matched to 738 controls and SCC cases to 284 controls. § Analysis of sun exposure and sunburn history and SCC based on 132 cases and 1031 controls.

perhaps the most compelling evidence that UVL exposure early in life can impact KC risk. Combining residential history with total accumulated ambient hours of bright sunshine and global irradiance, the western Australian study observed a trend of increasing risk of both BCC and SCC with higher lifetime sunlight exposure (English et al., 1998b; Kricker et al., 1995a). Epidemiologic studies have attempted to quantify an individual’s lifetime sun exposure history through often complex personal interviews (Table 64–4). While the reproducibility of responses appears

reasonable (English et al., 1998c; Rosso et al., 1996), the validity of responses is not known. Cohort and case-control studies consistently find a positive association between total lifetime sun exposure, and both recreational and occupational exposure to the sun, and SCC risk (Table 64–5). Conversely, the available data on BCC are mixed, and a number of studies observed no association at all between cumulative sun exposure and BCC risk (Table 64–5). However, a number of studies using recalled sun exposure indicate that sun exposure during childhood or teenage years specifically affects risk of both BCC and

Keratinocyte Carcinomas Table 64–5. Summary of Epidemiologic Findings on Solar Factors and Keratinocyte Carcinomas BCC

SCC

+ +

+ -

+ +

+ +

+/+/+ +/-

+ + + + +

+ + +/-

-

+

+

+ +

+ +

descriptive factors Anatomic site Usually exposed Occasionally exposed Place of residence High ambient UVL region Migration to high UVL region as a child/teenager

sun exposure history Recalled Cumulative Sun Exposure Total lifetime Recreational Occupational Childhood/teen Recent exposure Recalled intermittent exposure High percent of exposure recreational Holidays Outdoor sports

effects of sun exposure Recalled Sunburns Clinical signs of solar damage Solar keratoses Cutaneous microtopography

SCC (Table 64–5). In Alberta, Canada, occupational sun exposure in the 10 years before diagnosis was associated with an increased risk of SCC (Gallagher et al., 1995b), suggesting an effect of recent exposure. Supporting this hypothesis are findings from a clinical trial in which recent use of sunscreens prevented SCC, but not BCC occurrences (described under Preventive Measures) (Green et al., 2000). In general, epidemiologic data do not support an association between BCC and recent sun exposure (Table 64–5). In addition to cumulative exposure, epidemiologic studies have examined intermittent exposure or the amount of short periods of intense sun exposure (e.g., the percent of time spent outdoors during non-working hours, time spent outdoors during holiday or engaging in sports). These studies find a relationship with intermittency and BCC but not SCC (Table 64–5). This is further supported by the distribution of BCC tumors on the anatomic sites that receive sun exposure only occasionally, such as the trunk (Figs. 64–5 and 64–6, Table 64–5). The hypothesis that intermittent exposure rather than chronic exposure affects melanoma risk also has been raised, and indeed, BCC appears to share a sunlight exposure pattern of risk more similar to melanoma than SCC (Rosso et al., 1998). All studies of SCC, and nearly all studies of BCC, including the prospective cohort studies, find number of severe sunburns (i.e., painful or blistering sunburns) is a determinant of KC risk (Table 64–4). The results between studies cannot be easily compared because each investigation phrased and analyzed questions somewhat differently. In the US, the Nurses’ Health Study estimated relative risks of 2.4 for SCC and 4.9 for BCC associated with six or more severe sunburns (Hunter, 1990; Grodstein et al., 1995). Several studies found the association with sunburns to be strongest for occurrences in childhood or adolescence, again supporting a role of early life exposure (Kricker et al., 1995b; Rosso et al., 1996; Gallagher et al., 1995; Bajdik et al., 1998; Kennedy et al., 2003). Clinical signs of solar skin damage, assessed by dermatologic examination, are perhaps the strongest solar-related indicators of both BCC and SCC risk, especially solar keratoses. Findings for BCC are of particular interest because keratoses are not considered a precursor lesion for this type of cancer as they are for SCC. Other clinical measures of solar skin damage, such as telangiectasia of the face, elastosis of the neck and to a lesser extent number of solar lentingines also have been associated with BCC and SCC risk in some studies, but not as strongly or consistently as number of solar keratoses (Green et al., 1990; English et al., 1998a). Efforts to derive an objective measure of solar

1237

exposure include cutaneous microtopography, an assessment of skin elasticity from a silicon mold impression. However, few published studies used this tool. In Western Australia, there was a linear trend in SCC risk related to loss of elasticity, and a trend in risk for BCC (Kricker, 1991). Measurement of genetic alterations of UVL damage, i.e., UVL-related p53 mutations, may be useful biomarkers of sun exposure but have not yet been applied to large-scaled epidemiologic investigations (Nakazawa et al., 1994; Einspahr et al., 1997; Ouhtit et al., 1997; Ouhtit et al., 1998).

Tanning Lamps While several case reports implicated artificial tanning in the pathogenesis of BCC and SCC (Diffey et al., 1990; Speight et al., 1994; Sharfstein and Sharfstein, 1995; Spencer and Amonette, 1998), only a few epidemiologic studies have examined this exposure. In the 1980s, two hospital-based studies from Dublin (O’Loughlin et al., 1985; Herity et al., 1989) and one population-based study of men in Alberta Canada (Bajdik et al., 1996) found no relation between artificial sunlight and BCC or SCC risk. In a hospital-based study from Montreal, Quebec (Aubry et al., 1985), 4 of 92 SCC cases diagnosed in 1977–78 reported use of sunlamps compared to 1 out of 174 controls (OR = 13.42). These studies had a relatively small number of users and only collected crude measures of exposure (e.g., no quantitative information on timing or frequency of use). A more recent population casecontrol study in the US included 603 basal cell and 293 squamous cell cases diagnosed from 1993–1995 and 540 controls, and found ever use of a tanning device associated with a 2.5 fold risk of SCC and 1.5 fold risk of BCC (Karagas et al., 2002a). In this study, the relative risk estimates increased with earlier age at first use of a tanning device, a finding that parallels those of melanoma studies (Swerdlow and Weinstock, 1998). Further, in a small case-control study of women with early onset BCC (<40 years), cases had a higher frequency of visits to tanning salons, although this was not statistically significant (Boyd et al., 2002). In light of the pervasive use of artificial tanning, especially among adolescent girls (Banks et al., 1992; Oliphant et al., 1994; Rhainds et al., 1999; Geller et al., 2002), this is an issue of particular public health importance.

UVA and PUVA Treatment Ultraviolet radiation in the UVB range (280–320 nm) is considered more carcinogenic than UVA, although UVA (320–400 nm) also can induce SCC in animal models (IARC, 1992). The combination of UVA treatment with a photosensitizing agent psoralen (referred to as PUVA) is commonly used for management of severe psoriasis. Three longterm prospective studies detected an excess of SCC in patients treated with this regimen (Table 64–6). Standardized incidence ratios (SIR) for SCC of 6.5 and 5.6 were observed in studies from Finland (Hannuksela-Svahn et al., 2000) and Sweden (Lindelof et al., 1999), respectively. A multi-center US investigation derived a much higher SCC risk estimate (SIR = 11.9) (Table 64–6); however, a larger fraction of these patients were exposed to high doses of PUVA treatment, and risk increased with PUVA dose (Stern and Laird, 1994) (Table 64–6). Further, the US study observed synergistic effects with both ionizing radiation treatment and sun-sensitive skin type, the prevalence of which may differ across populations. Far less is known about the risk of BCC; in the US study, PUVA-treated patients had a four-fold risk of BCC compared with the general population, without a clear doseresponse relationship (Stern and Laird, 1994) (Table 64–6).

Photosensitizing Agents A large number of drugs can induce photosensitivity including psoralens, antibiotics (e.g., tetracyclines), fluoroquinolones, phenylproprionic acid derived nonsteroidal anti-inflammatory agents, phenothiazines, and amiodarone (Gould et al., 1995; Bellaney et al., 1996; Stern, 1998; Epstein, 1999). Use of these agents combined with solar or artificial UVL exposure can cause pronounced erythema and edema resembling a severe sunburn along with sustained hyperpigmention (Gould et al., 1995). DNA damage may result from

1238

PART IV: CANCER BY TISSUE OF ORIGIN

Table 64–6. Risk of Basal Cell and Squamous Cell Carcinomas Following Psoralen Plus UVA Therapy Reference Location Hannuksela-Svahn et al., 2000 Finland Lindelof et al., 1999 Sweden Stern and Laird et al., 1994† 16 US Clinical Centers

Cohort Size

Mean Duration of Follow-Up (years)

Basal Cell Carcinoma SIR (95% CI)

Squamous Cell Carcinoma RR/SIR (95% CI)

5687

14

6.5 (1.4, 31)*

4799

15.9

1380

13.2

Men 5.6 (4.4, 7.1) Women 3.6 (2.1, 5.8) Low exposure 5.0 (3.6–6.9) Medium exposure 13.4 (9.3–19.3) High exposure 32.8 (26.2–41.0) Overall 11.9 (10.1–14.0)

Low exposure 2.1 (1.6–2.7) Medium exposure 1.9 (1.2–3.0) High exposure 3.8 (2.8–5.1) Overall 2.5 (2.1–3.0)

*Relative risk estimate based on a nested case-control study within the cohort of 30 squamous cell carcinoma cases and 137 controls matched on age and sex. † Results based on first tumor occurrence of an individual.

photochemical reactions with DNA directly or by formation of DNAdamaging molecules such as reactive oxygen species (Gould et al., 1995; Gocke et al., 1998; Epstein, 1999). In animal studies, photosensitizing agents, in particular fluoroquinolone antibiotics, induced benign and malignant skin lesions, and specifically SCCs (Makinen et al., 1997; Stern, 1998). Remarkably little epidemiologic data exist on the potential risk of KC associated with photosensitizing drugs, aside from psoralens. A letter to the New England Journal of Medicine described a case-control comparison of prolonged use of photosensitizing drugs (5 or more years) and actinic keratosis, a precursor lesion to SCC among 68 consecutive patients (OR = 8.1, p < 0.01) (Placzek et al., 1999). An excess risk of BCC was not observed in relation to use of tetracycline or aspirin in a case-control study of BCC conducted by Wei and colleagues (1994); but the study may have had limited statistical power, and risks by dose or duration of use were not reported.

Immunosuppression Cutaneous warts, keratoacanthomas, and KC commonly occur in organ transplant recipients who require immunosuppressive drugs to prevent allograft rejection. Regimens usually involve a multi-agent combination of a cytotoxic drug (e.g., Azathioprine), an agent affecting signal transduction in T-lymphocytes (e.g., Cyclosporin), and a glucocorticoid (e.g., Prednisone) to suppresses T-cell proliferation and immune response. The overall risks of KCs among transplant recipients in the Netherlands were 10% at 5 years and 40% at 20 years (Hartevelt et al., 1990), and even higher in Australia with 5- and 20year risks of 20% and 75%, respectively (Bouwes Bavinck et al., 1996). In contrast to the general population, SCCs tend to predominate over BCC among organ transplant recipients, and the reported onset tends to be at a younger age and usually involves multiple tumors even numbering into the hundreds (Sheil et al., 1987; Hartevelt et al., 1990). In large cohort studies, the estimated risks of SCC among renal transplant recipients were 65- to 250-fold above the general population and risk for BCC was elevated 10-fold (Hartevelt et al., 1990; Jensen et al., 1999). Heart transplant recipients require stronger immunosuppressive regimens and have a correspondingly three-fold higher risk of KCs than kidney transplant recipients (Jensen et al., 1999). KC risk among liver transplant recipients also appears to be elevated due to long-term immunosuppression (Frezza et al., 1997; Sheiner et al., 2000). Thus, the excess risk of skin cancer is almost certainly due to immunosuppressive therapy, rather than antigenic stimulation of the graft (Bouwes Bavinck et al., 1996). It further appears to be dose-dependent (Fortina et al., 2004). Indeed, efforts to reduce the dose of immunosuppressive therapy diminished skin cancer occurrences, but increased the likelihood of organ rejection (Dantal et al., 1998). A concern is that SCC tumors among transplant recipients metastasize more frequently, and as such are associated with a higher mortality rate among transplant recipients (Dinehart et al., 1990; Bouwes Bavinck et al., 1996; Ong et al., 1999; Veness et al., 1999). The possible cancer risks associated with use of immunosuppressive therapy in patients other than organ transplant recipients is less understood. Based on only a few cases, a greater than expected inci-

dence of SCC (3 observed vs. 0.13 expected) was observed among patients treated with Azathioprine, Cyclophosphamide, or Chlorambucil (Kinlen et al., 1979). An elevated risk of SCC also occurred in a cohort of rheumatoid arthritis patients treated with Cyclophosphamide (crude OR = 3.4; p = 0.04) (Radis et al., 1995). In a cohort of 1065 patients with Wegener granulomatosis treated with corticosteroids and Cyclophosphmide, a seven-fold excess risk of squamous cell carcinoma was identified (SIR = 7.3; 95% CI: 4.2–12) (Knight et al., 2002). These results are consistent with findings of an excess SCC risk (OR = 2.31; 95% CI: 1.27–4.18), and to a lesser extent BCC risk (OR = 1.49; 95% CI: 0.90–2.47) among users of glucocorticoids in a population-based case-control study (Karagas et al., 2001a). Results subsequently were corroborated in a record linkage study from Denmark (Sorenson et al., 2004). In addition to immunosuppressive treatment, patients with AIDS or HIV infection have been documented to have an increased risk of both BCC and SCC (reviewed in (Wang et al., 1995) ), with recent cancer registry data reporting elevated SIRs of 3–6 for infected individuals (Franceschi et al., 1998; Cooksley et al., 1999) ). Further, as with medicinally immunosuppressed populations these tumors tend to be more aggressive and occur in greater numbers (Wang et al., 1995).

Human Papillomavirus HPV DNA was first discovered in SCC tumors of patients with epidermodysplasia verruciformis (EV), a rare autosomal recessive disorder manifested by extreme susceptibility to KC beginning in the second decade of life and by infection with specific human papillomaviruses (HPVs) (Orth et al., 1980). HPV-types identified in EV patients encompass HPV5, 8, 9, 12, 14, 15, 17, 19–25, 36–38, 46, 49, and 80 (Orth et al., 1980; Berkhout et al., 2000), with HPV5 and HPV8 being the most common types. EV-type HPV has been detected in up to 90% of SCC tumors among EV patients (Orth et al., 1980). Although results vary, a similarly high percentage of SCC tumors from organ transplant recipients also contain EV-type HPV DNA (Shamanin et al., 1996; de Villiers et al., 1997; Harwood et al., 2000). EV-type HPVs also have been isolated in 27%–37% of SCC and BCC tumors of immunocompetent individuals (Shamanin et al., 1996; Harwood et al., 2000), and some KCs, especially SCCs (including Bowen disease) occurring on the hands, contained genital type HPVs (Sau et al., 1994; Clavel et al., 1999). Interestingly, the prevalence of HPV in KC tumors of PUVA-treated patients was 75%, an estimate closer to that observed in organ transplant recipients and EV patients than the general population (Harwood et al., 1998). This finding supports mechanistic studies that demonstrate an interaction between certain HPV types and UVL in carcinogenesis (Purdie et al., 1999; Jackson and Storey, 2000). However, these tumor prevalence studies are based on a relatively small number of subjects of each histology. Moreover, a similar percentage of normal skin appears to harbor HPV (Astori et al., 1998) raising suspicions as to the etiologic role of HPV in KC carcinogenesis. More recent work involves serologic testing and HPV detection in the follicles of hair pluckings (i.e., opposed to the tumor itself). EVHPV has been found in hair pluckings from both immunosuppressed

Keratinocyte Carcinomas and immunocompetent individuals (Boxman et al., 1997). Further, the presence of EV-HPV in eyebrow samples has been associated with a twofold increased risk of SCC (Boxman et al., 2000). Serology investigations also support an association between exposure to certain EV HPV types and risk of SCC (Bouwes Bavinck et al., 2000; Feltkamp et al., 2003; Masini et al., 2003) and BCC (Feltkamp et al., 2003), although additional studies are needed.

Ionizing Radiation In epidemiologic studies, an increased risk of KC has been observed in several radiation-exposed groups including uranium miners (Sevcova et al., 1978), radiologic technologists (Yoshinaga et al., 2005), radiologists (Matanoski et al., 1975; Smith and Doll, 1981), and individuals therapeutically treated with X-rays for tinea capitis (Shore et al., 1984; Ron et al., 1991), thymic enlargement (Hildreth et al., 1985), childhood cancer (Perkins et al., 2005) or acne (Karagas et al., 1996; Lichter et al., 2000). Ionizing radiation exposure is consistently related to an enhanced risk of BCC; however, a risk of SCC is less evident (Thompson et al., 1994; Ron et al., 1998). In an analysis of atomic bomb survivors, Ron and colleagues observed an excess risk of Bowen disease, an in-situ SCC lesion, but not invasive SCC. One possible explanation for the lack of association with SCC may be the later onset and lower prevalence of this histology. Indeed, large, population-based case-control studies have found excess risks of both BCC and SCC. A study of men in Canada reported odds ratios of 5.7 and 4.8 for BCC and SCC respectively associated with nondiagnostic radiation exposure (Bajdik et al., 1996). A case-control study from New Hampshire observed a threefold risk of both BCC and SCC associated with radiotherapy to the site of exposure (Lichter et al., 2000). A number of studies found a more pronounced risk associated with early age at exposure (Ron et al., 1991; Thompson et al., 1994; Karagas et al., 1996; Lichter et al., 2000). Other studies suggest that KC risk increases after a long latency period (Shore et al., 1984; Karagas et al., 1996; Lichter et al., 2000), but, this is not consistently observed (Ron et al., 1991; Thompson et al., 1994). Inability to distinguish latency from age at exposure hampers the interpretation of studies to date. The finding of an interaction between PUVA and radiotherapy raised the possibility of a modifying role of UVL (Stern et al., 1979). In support of this hypothesis are findings from a case-control study in which individuals with a sun-sensitive phenotype had a greater risk of radiotherapy-associated SCC (Lichter et al., 2000). This hypothesis requires further investigation.

Arsenic Over a hundred years ago, physicians recognized the occurrence of skin cancers from arsenic-containing medicines, such as Fowler’s Solution (1% potassium arsenite), used for psoriasis and other conditions (Neubauer, 1947). Subsequent reports included skin cancers among manufacturers of arsenic-containing compounds, vineyard workers (including Moselle vintners who ingested arsenic pesticide residues), and users of other arsenic-containing drugs (i.e., asthma remedies) (IARC, 1980). Several places in the world reported appearances of skin cancers from arsenic-contaminated drinking water including Silesia (Jackson and Grainge, 1975), Argentina (Bergoglio et al., 1964), Mexico (Cebrian et al., 1983) and Chile (Zaldivar et al., 1974) with more recent reports from Bangledesh, West Bengal, and Malaysia (Saha et al., 2001; Kurokawa, et al., 2001; Jaafar et al., 1993). In 1968, a seminal study from an endemic arsenic region in the southwest of Taiwan reported a striking dose-response relationship between water arsenic concentrations and skin cancer prevalence based on a prevalence survey of 40,000 households (Tseng et al., 1968); rates were over eight-fold higher in villages with median well water concentrations of 800 mg/L vs. 170 mg/L. These findings were supported by ecologic mortality and incidence studies (Chen et al., 1985; Chen et al., 1988; Wu et al., 1989; Chen and Wang, 1990; Guo et al., 1997; Tsai et al., 1999) in the region, and by ecologic studies from Mexico (Cebrian et al., 1983), Argentina (Bergoglio et al., 1964), and Chile (Zaldivar et al., 1974; Rivara et al., 1997; Smith et al., 1998).

1239

A small case-control study from the southwest of Taiwan found the prevalence odds ratios of skin cancer increased with duration of residence in the endemic regions, duration of drinking well water, and water arsenic concentrations (Hsueh et al., 1997). This and other studies provided some evidence that individual differences in arsenic metabolism may modify risk (Chen et al., 2003; Yu et al., 2000; Ranft, 2003) including a study that detected an excess KC from residential and dietary arsenic from a coal burning plant in Slovakia (Pesch et al., 2002). The effects of environmental levels of arsenic more typical of the United States and elsewhere are not known. Early ecologic studies of water arsenic and skin cancer in the United States had the potential for substantial exposure and disease misclassification (Berg and Burbank, 1972; Morton et al., 1976). A case-control study was conducted in a region of the United States where the use of private well water results in exposures that exceed the current US EPA maximum contaminant levels. Using toenail concentrations as a biomarker of exposure, there was evidence of an increased SCC risk and to a less extent BCC at the highest levels of exposure (Karagas et al., 2001b; Karagas et al., 2002b). Yet, further data are needed both on the effects of low-level exposure and the role of host susceptibility and other potential co-factors.

Occupational and Chemical Exposures The occurrence of skin cancers in the occupational setting has played an important historical role in carcinogen discovery (Hueper, 1963), beginning with Percival Pott’s description of scrotal skin cancers among chimney sweeps in 1775. Although only a limited number of epidemiologic studies are available, high dermal polycyclic aromatic hydrocarbon (PAH) exposure in the occupational setting is a potential etiologic factor for KC, in particular SCC. Cohort studies of workers in shale oil extraction, creosote-exposed wood impregnators, chimney sweeps, roofers, and asphalt workers reported excess skin cancer occurrences (Boffetta et al., 1997). In a case-control study from Alberta, Canada exposure to petroleum products, grease, diesel fumes, and coal dust was associated with an increased SCC risk (Gallagher et al., 1996). A small case-control study of workers in the tire and rubber manufacturing industry (65 SCC cases and 254 controls) found that exposure to lubricating oils and rubber solvents related to an elevated SCC risk (Bourguet et al., 1987). Findings with other occupational or chemical exposures require additional confirmation, including an excess skin cancer risk among urban bus drivers and tramway employees (Soll-Johanning et al., 1998), sawmill workers, carpenters and joiners (Stellman and Garfinkel, 1984), and hairdressers, as well as those exposured to dry cleaning agents, fiberglass dust, luminous paint, asbestos dust (Gallagher et al., 1996), agriculture-related exposures such as insecticides, herbicides, fungicides, and seed treatments (Gallagher et al., 1996). Others have reported possible associations with the agriculture chemicals 2, 4 dichlorophenoxy acetic acid (Hogan et al., 1989; Hogan et al., 1990) and paraquat (Jee et al., 1995).

Smoking In theory, tobacco smoke, a source of PAH exposure, could act as a direct carcinogen on the skin or through systemic mechanisms (i.e., immunosuppression). Cohort and case-control studies generally report elevated risks of SCC with cigarette smoking, particularly current smoking (Table 64–7). A case-control study from the United Kingdom (Lear et al., 1998) noted an association between pack-years smoked and SCC risk (OR = 1.006/pack year smoked; 95% CI: 1.001–1.007; p = 0.008), however dose trends have not been consistently observed (Aubry et al., 1985; Grodstein et al., 1995). A link between smoking history and BCC is even less clear. Most studies did not detect an excess BCC risk in relation to past or current smoking (Table 64– 8). The exception is a recent twin pair study of BCC from Finland in which an elevated risk of BCC was observed among women but not men who had ever smoked, and specifically among dizygotic twins (Table 64–8). Also, in a small study (n = 30 cases and controls), women diagnosed with BCC before age 40 years smoked a

1240

PART IV: CANCER BY TISSUE OF ORIGIN

Table 64–7. Risk of Squamous Cell Carcinoma of Skin Associated with Cigarette Smoking* Reference

Study Population

No. of Cases Cohort Size/No. of Controls

Former Smokers RR/OR (95% CI)

Current Smokers RR/OR (95% CI)

1.62 (1.07, 2.47)

2.01 (1.21, 3.34)

1.1 (0.8, 1.5)

1.5 (1.1, 2.1)

cohort studies Karagas et al., 1992 Grodstein et al., 1995

Skin Cancer Prevention Trial Nurses’ Health Study (women only)

132 SCC 1805 Total 197 SCC 107,900 Total

case-control studies Aubry et al., 1985

Montreal region

Gamble et al., 1996

Texas

Bajdik et al., 1998

Alberta, Canada (men only) Netherlands

De Hertog et al., 2001

92 SCC 174 dermatology controls 59 SCC 239 dermatology controls 113 head and neck SCC 342 controls 161 SCC 386 ophthamology patients controls

2.30 (1.27, 4.15) 1.56 (0.67, 3.63)

0.76 (0.40, 1.46)

1.16 (0.66, 2.05)

1.53 (0.82, 2.86)

1.8 (1.0, 3.0)

2.9 (1.5, 5.6)

*Hogan et al. (1990) found no association, but a relative risk was not provided. Lear et al. (1998) reported an association between pack-years smoked and SCC OR = 1.006/pack year smoked (95% CI; 1.001–1.007; p = 0.008).

greater number of pack-years than controls (Boyd et al., 2002). The possibility that subgroups of individuals (i.e., based on age, gender, or inherent factors) carry greater susceptibility to tobacco-induced skin cancers will require further exploration.

Diet Epidemiologic studies on dietary factors and skin cancer, as yet, do not suggest a prominent role of diet on KC incidence. However, intriguing results of reduced actinic keratoses came from a small trial (101 patients) that evaluated a low-fat diet (Black et al., 1994). The number of keratinocyte carcinomas during the last 8-month period of the trial was significantly reduced in the intervention group compared with the control group (0.02 vs. 0.22 cancers per patient). In two large cohort studies using a food frequency questionnaire, dietary intake of cholesterol, saturated fat, monounsaturated fat, and polyunsaturated fat generally were unrelated to an increased risk of BCC, and in some subgroups related to a reduced risk (Hunter et al., 1992; van Dam et al., 2000). Likewise, a nested case-control study found no association between fat intake based on a 7-day food diary and BCC risk (Davies et al., 2002). Yet, in a population-based case-control study, higher intakes of n-3 fatty acids and n-3 to n-6 fatty acid ratios were associ-

ated with a reduced SCC risk, based on four 24-hour recalls (Hakim et al., 2000a). Interestingly, in a subsequent analysis of red blood cell fatty acids, increasing proportions of palmitic acid and palmitoleic acid were inversely related and arachidonic acid levels positively related to SCC risk (Harris et al., 2005). A number of putative anti-oxidants have been examined as possibly reducing SCC or BCC occurrences. Intake of selenium was hypothesized to reduce KCs based on observational studies in the 1980s (Clark et al., 1984; Clark et al., 1987). Two subsequent nested case-control studies (Breslow et al., 1995; Karagas et al., 1997) did not observe a reduced risk of SCC associated with higher plasma selenium concentrations. Moreover, in a randomized clinical trial of selenium supplementation (200 micrograms per day) there was no reduction in the incidence of either BCC or SCC over a 10-year period, and indeed, a possible increased risk for SCC (Clark et al., 1996; Duffield-Lillico et al., 2003). Also, none of four trials of b-carotene noted any effect on risk of KC, BCC, or SCC (Greenberg et al., 1990; Green et al., 1999; Lee et al., 1999; Frieling et al., 2000). In observational studies, intake of Vitamin A, retinol, Vitamin E, and several carotenoids (based on a food frequency questionnaire) were not associated with a reduced risk of either SCC or BCC (Fung et al., 2002a; Fung et al., 2003). On the contrary, intakes of Vitamin A, retinol, total

Table 64–8. Risk of Basal Cell Carcinoma of Skin Associated with Cigarette Smoking* Reference

Study Population

No. of Cases Cohort Size/No. of Controls

Former Smokers RR/OR (95% CI)

Current Smokers RR/OR (95% CI)

cohort studies Hunter et al., 1990

Nurses’ Health Study (women only)

771 BCC 73,366 Total

1.12 (0.94, 1.32)

Karagas et al., 1992

Skin Cancer Prevention Trial

0.82 (0.69, 0.98)

van Dam et al., 1999

Health Professionals (men only)

651 BCC 1,805 Total 3,273 BCC 44,591 Total

Freedman et al., 2003

US Radiological Technologists

1,360 BCC 68,371

1.1 (1.0–1.3)

Gamble et al., 1996

Texas

0.86 (0.46, 1.62)

Milan et al., 2003

Finland

174 BCC 239 dermatology controls 333 BCC 333 twin controls

1.03 (0.97, 1.11)

1–14/d 0.93 (0.73, 1.30) 15–24/d 1.13 (0.90, 1.43) ≥25/d 1.03 (0.78, 1.36) 0.80 (0.63–1.00) 1–14/d 1.14 (0.93, 1.40) 15–24/d 0.82 (0.65, 1.05) ≥25/d 0.86 (0.67, 1.10) 0.9 (0.8–1.0)

case-control studies Men 0.70 (0.32, 1.54) Women 3.62 (1.03, 12.7)

0.94 (0.61, 1.46) Men 0.70 (0.30, 1.64) Women 2.06 (0.67, 6.36)

*De Hertog et al., 2001 found no association with BCC and smoking (current or past) OR = 1.1 (CI: 1.1, 0.69–1.9). Boyd et al., 2002, found a greater number of pack-years smoked in women BCC cases diagnosed <40 yrs than controls (p = 0.045). † Former smokers include occasional smokers.

Keratinocyte Carcinomas carotene, b-carotene, lutein/zeaxanthin, and Vitamin E were positively related to BCC risk, although the magnitude of increases were small. Further, most nested case-control studies using plasma or serum markers of b-carotene, retinol, alpha tocopherol, and lycopene did not provide clear evidence of any increase or decrease in SCC or BCC risk associated with these nutrients (Knekt et al., 1988; Breslow et al., 1995; Karagas et al., 1997; Davies et al., 2002). In one study, serum lutein, zeaxanthin and beta-crypoxanthin were positively related to SCC risk (Dorgan et al., 2004). Other dietary-related factors evaluated for their relation with KC include Vitamin D, folate, tea, and alcohol consumption. Data from the Nurses’ Health Study and Health Professionals Followup Study did not indicate an association between Vitamin D intake and BCC occurrence (Hunter et al., 1992; van Dam et al., 1999), but did suggest a modest increase in BCC risk with higher folate intake (Fung et al., 2002a). Analyses of alcohol intake in these cohorts reported positive trends in BCC risk in relation to total alcohol intake, and specifically alcohol intake from liquor among men and white wine among women (Fung et al., 2002b). In contrast, consumption of red wine related to a reduced BCC risk among women. Results from the US Radiological Technologists Cohort also suggest a BCC risk associated with number of alcoholic drinks per week, particularly among women (Freedman et al., 2003). However, in both studies the magnitude of risk was relatively small. Other smaller studies did not detect associations between KC and alcohol intake (Kune et al., 1992; Sahl et al., 1995; Davies et al., 2002; Milan et al., 2003). An interest in the potential protective effects of tea consumption on skin cancer risk has grown out of experimental findings that tea inhibits UVL-induced skin cancers in mice (Lambert and Yang, 2003). Interestingly, a casecontrol study from Arizona reported an inverse relationship between hot black tea consumption and risk of SCC, and a trend of diminishing risk with longer brewing time (Hakim et al., 2000b). However, in a twin pair study, regular consumption of more than one cup of tea per day did not affect BCC risk (Milan et al., 2003).

Other Exposures Chronic injury, such as burns, scars, or ulcers is clinically recognized as a predisposing condition for KC, in particular SCC, with a long delay from the time of injury to malignancy, aggressive behavior (i.e., lymph node metastases) (Ames and Hickey, 1982; Eltorai et al., 2002; Copcu et al., 2003), and more common occurrence among darker pigmented individuals at lower risk of UVL-induced KCs (Forbes et al., 1981; Yakubu and Mabogunje, 1995). Epidemiologic data on injury are lacking although one study from Canada reported an association between BCC and frost bite (Hogan et al., 1989). Other factors of potential importance have been examined, but only in one or two epidemiologic studies, such as childhood immunization (Gallagher et al., 1996), blood transfusion (Blomberg et al., 1993), body mass index (Gilbody, 1994; van Dam et al., 1999), gynecomastia (Olsson et al., 2002), postmenopausal estrogen use (Wei et al., 1994), and nonsteroidal anti-inflammatory drugs (Butler et al., 2005).

1241

mentation traits (Kricker et al., 1991; Grodstein et al., 1995; Rosso et al., 1996; English et al., 1998a). Objective measures of skin pigmentation, using skin reflectance spectroscopy, have been introduced as a potentially useful method in the study of skin cancer etiology. For BCC, initial investigators did not see an increased risk of BCC with lighter skin pigmentation (Kricker et al., 1991; Lock-Andersen et al., 1998; Lock-Andersen et al., 1999), but reported a positive association with SCC (Kricker et al., 1991). In contrast, recent work has reported a six fold elevated risk of BCC, and four fold increased risk of SCC with light skin pigmentation for men, and no apparent association for women (Dwyer et al., 2002).

Skin Responses to Exposure Skin type is a complicated measure of UVL responsiveness that integrates concepts of received UVL dose at the basal layer, secondary inflammatory reactions, and skin changes associated with patterns of exposure. In assessing skin type, both the acute burning response and the ability to tan are typically integrated into a single measure. Fitzpatrick originally developed such a scale for clinical use in assigning PUVA treatment regimens (Fitzpatrick et al., 1988), and it has been widely adapted for use in the investigation of skin cancer risk. Four general classes of skin type are typically used: always burn/never tan, usually burn/light tan, sometimes burn/average tan, never burn/always tan. In a study of melanoma there was evidence that such a variable is subject to recall bias (Weinstock et al., 1991). Despite this, most case-control studies investigating the association of sun sensitivity and KC observe higher risks among those who tend to burn with little tanning (Hunter et al., 1990; Grodstein et al., 1995; Zanetti et al., 1996; Lear et al., 1997; English et al., 1998b; Rosso et al., 1998; LockAndersen et al., 1999; Rosso et al., 1999; van Dam et al., 1999; Naldi et al., 2000; Vlajinac et al., 2000). Finally, two groups have tried to differentially assess the risk of BCC associated with the burning and tanning components of skin type. In a large, population-based study in Australia, Kricker and colleagues (1991) observed an increased risk of BCC with the absence of tanning, but not the burning phenotype. Similarly, in a cohort of US nurses, Hunter and colleagues (1990) observed a decreased risk of BCC among individuals with a tanning response after repeated UVL exposure. In addition, many groups have reported significant effect modification for the risk effects of UVL exposure by skin type. Two studies reported an attenuation of the dose-response curve among tanners, one examining sunburn history (Grodstein et al., 1995), the other childhood exposure (Gallagher et al., 1995). Two other studies, one among Southern Europeans (Rosso et al., 1996) the other, Western Australians (Kricker et al., 1995a), indicated the dose-response relationship was relatively flat for burners and steep among tanners, possibly reflecting the need for higher UV doses among those with a stronger pigmentation response. As highlighted by English et al. (1998a), the patterns and strengths of association for pigmentary traits and skin type with disease vary substantially across studies, however, one can have confidence that they do have a significant impact on risk for both BCC and SCC.

HOST FACTORS Pigmentation

Freckling and Nevi

While exposure plays a dominant role in KC development, the role of host factors is clear as these malignancies arise almost exclusively in lightly pigmented populations. Fair skin, blue eyes, and blonde or red hair are generally accepted risk factors for both BCC and SCC. As these traits are related but not strictly correlated, the analytic approach of each study has the potential to greatly impact the magnitude and strength of any observed association. In general, most studies have observed a significant association between risk of BCC and fair skin, blue or green eyes, and blonde or red hair (Kricker, 1991; Gallagher et al., 1995; Rosso et al., 1999; van Dam et al., 1999; Naldi et al., 2000; Vlajinac et al., 2000). There have been fewer investigations of pigmentary traits and SCC, with smaller sample sizes, however the weight of evidence indicates a positive association with light pig-

Some individuals experience significant freckling in UV-exposed skin. Freckles are increased deposits of melanin without a concomitant increase in melanocyte number, and the tendency to freckle has a genetic component that is independent of skin coloration. Several investigations have documented a substantially increased risk of BCC among those with a freckling phenotype (Green et al., 1990; Kricker et al., 1991; Gallagher et al., 1995; Vlajinac et al., 2000), and one of increased risk for SCC (Kricker et al., 1991; English et al., 1998a). Unlike freckles, nevi (moles) are an increased number of melanocytes that cluster together in demarcated nests. Nevi vary with age and sun exposure, and are hypothesized to be induced differentially with intermittent sunlight exposures. Further, they have a genetic component, and are a consistent risk factor for malignant melanoma. The few

1242

PART IV: CANCER BY TISSUE OF ORIGIN

studies of nevi and KC risk that have been reported observe a weak to moderate association between increased nevi counts and BCC risk (Kricker et al., 1991; van Dam et al., 1999; Naldi et al., 2000), with some suggestion for an effect specifically in women (Lock-Andersen et al., 1999). Only one study has found a similar positive association between nevi and SCC risk; interestingly this study was an investigation restricted to women (Grodstein et al., 1995).

Genetic and Familial Susceptibility Skin coloration is determined by the presence of two pigments that absorb UV light and modify biologic dose: eumelanin (dark pigment that is protective) and phaeomelanin (red pigment that confers risk, possibly due to free radical generation) (Ranadive et al., 1986). Hair and skin color, which are directly attributable to the ratio of these pigments, as well as freckling and tanning, have been associated with genetic polymorphisms in the MC1R gene (Valverde et al., 1995; Xu et al., 1996; Koppula et al., 1997; Smith et al., 1998; Flanagan et al., 2000; Healy et al., 2000; Palmer et al., 2000; Bastiaens et al., 2001; Bastiaens et al., 2001), making it a likely susceptibility gene for a large proportion of KC. MC1R (melanocortin-1 receptor) is a transmembrane G-protein coupled receptor that is activated by aMSH (alpha-melanocyte stimulating hormone). Following activation, cAMP levels increase triggering production of eumelanin. There are 28 described haplotypes for the MC1R locus and most variants occur at a population frequency poorly suited to epidemiologic study. However, a population-based examination of Irish individuals identified three common sites of variation that are significant determinants of red hair (the RHC variants): codons 151, 160, and 294 (Smith et al., 1998), a finding confirmed by a more classical population genetics approach (Flanagan et al., 2000). There are several published accounts of MC1R RHC variants and risk of KC (Bastiaens et al., 2001; Box et al., 2001; Kennedy et al., 2001). Each demonstrates an excess risk of skin cancer with MC1R, and the effect appears to be greatest for SCC. The largest of these, a clinic-based study from the Netherlands (with 453 cases and 385 controls), strongly supports an effect of the RHC variants and KC risk (Bastiaens et al., 2001). Thus, the MC1R gene appears critically important in mediating skin pigmentation, and therefore has significant potential to modify received UVL dose. Keratinocyte carcinoma is a classic example of a cancer generated through gene-environment interactions. An exaggerated example occurs with the rare genetic syndrome xeroderma pigmentosum (XP). Individuals with XP are described as having a 1000-fold increased risk of skin cancer and develop carcinomas early in life (Kraemer et al., 1987). This recessive disease is attributable to a deficit in nucleotide excision repair, which specifically repairs UVL-induced DNA damage (Cleaver, 1968). Thus, upon UVL exposure the genetic defect is revealed resulting in a large gene-environment interaction and extremely high risk of KC. While XP is a rare disease with a severe abrogation in DNA repair, more subtle variations in DNA repair proficiency may significantly increase risk for KC. Phenotypic work has demonstrated that decreased DNA repair capacity in circulating lymphocytes is associated with increased risk of both types of KC (Munch-Petersen et al., 1985; Wei et al., 1993) and actinic keratoses (Abo-Darub et al., 1978), although this trend has not been conclusive for all studies (Hall et al., 1994; D’Errico et al., 1999). Recent work has focused on how specific DNA repair polymorphisms impact DNA repair phenotypes and skin cancer risk. An XPD codon 312 polymorphism has been associated with increased UVLinduced apoptosis (Seker et al., 2001) and elevated monocyte DNA adducts (Hou et al., 2002), however no significant association has been found for BCC (Vogel et al., 2001). A second XPD polymorphism at codon 751 has also been associated with increased bulky DNA adducts (Matullo, 2001; Palli et al., 2001; Hou et al., 2002), as well as reduced repair of cyclobutane dimers in older subjects (Hemminki et al., 2001). Both the codon 751 and a non-coding polymorphism at codon 156 have been associated with increased risk of BCC in selected populations (Dybdahl et al., 1999; Vogel et al., 2001). A second DNA repair

gene, XRCC1, has been intensively investigated for a cancer susceptibility phenotype. Many groups have demonstrated a phenotype of reduced repair capacity with the codon 399gln allele of XRCC1 (Abdel-Rahman et al., 2000; Duell et al., 2000; Matullo et al., Lei et al., 2002). The codon 399 polymorphism has been associated with both BCC and SCC (Nelson et al., 2002). Further, a strong geneenvironment was present, where the variant allele conferred a six-fold elevated risk of SCC among those with three or more painful sunburns (Nelson et al., 2002). In addition to DNA repair, there are many other genetic pathways that may interact with exposure, or the downstream effects of exposure (i.e., inflammation) to modify the risk of commonly occurring KC. The glutathione S-transferases (GSTs) are a family of isoenzymes responsible for the detoxification of a wide range of electrophilic compounds. Specifically, the GSTs are hypothesized to be potentially important for protection against damage from oxygen radicals (Hayes et al., 1995), and thus functional variants are likely to increase risk for skin malignancies. The GST deletion polymorphisms are associated with increased sensitivity to UVB, as measured by post-exposure inflammatory reactions (Kerb et al., 1997). In addition, the GSTM1 deletion polymorphism has been associated with an increased risk of BCC (Heagerty et al., 1994). However, nearly all of the prior research on the GSTs and KCs have utilized methods that focused on high-risk subgroups (renal transplant recipients and multiple BCCs). Among renal transplant recipients there is a significant association between SCC and all tested GST variants (Ramsay, 2001). A separate study of long-term transplant recipient survivors also reports an association between SCC and GSTP1 (Marshall et al., 2000). The GSTM3 and GSTT1 polymorphisms have been associated with risk of multiple BCCs (Lear et al., 1996; Yengi et al., 1996). In addition, GSTP1 was associated with increased numbers of BCC tumors among the cases with multiple tumors (Ramachandran et al., 2000). Finally, two small skin cancer studies of immunocompromised individuals have investigated the hypothesis that the p53 polymorphism is associated with tumor development (Bastiaens et al., 2001; Marshall et al., 2001; Wu et al., 2002). Neither study had significant power to detect a reasonable association between the polymorphism and skin cancer (minimum detectable OR 2.8–4.5). However, in both studies the 72arg allele was over-represented among skin cancer cases.

MOLECULAR GENETIC CHARACTERISTICS OF TUMORS TP53 The TP53 gene is considered a prime mutational target for most solid tumors, including keratinocyte carcinomas. TP53 is a tumor suppressor gene, typically requiring inactivation of both gene copies for full tumorigenic potential. However, experimental evidence indicates that a single TP53 mutation changes keratinocyte behavior, such that UVLinduced apoptosis is abrogated promoting the formation of p53 mutant clones (Ziegler et al., 1994; Zhang W et al., 2001). Other transgenic mouse work supports a dose-relationship between the number of inactive TP53 alleles and skin carcinoma risk (Jiang et al., 1999). These observations, derived from the experimental laboratory, are consistent with epidemiologic observations. Chronically UVL-exposed skin contains fields of p53 dysregulated cells, termed p53 patches (Jonason et al., 1996). These pre-neoplastic clones, containing 60–3000 cells, are histologically normal appearing and may involve up to 4% of human epithelium (Jonason et al., 1996). Further, there is a dose-dependent relationship between the amount of UVL exposure received (based on anatomic site) and the number of TP53 hot-spot mutations observed (Nakazawa et al., 1994; Ohtit et al., 1998). Precursor lesions also contain TP53 mutations supporting the hypothesis that these TP53 mutations are likely the initiating event in keratinocyte tumorigenesis (Campbell et al., 1993; Nelson et al., 1993; Taguchi et al., 1994; Ziegler et al., 1994; Lee et al., 2000). TP53 alteration is consistently the most commonly observed mutation event in both BCC and SCC tumors, underlying its importance in the etiology of keratinocyte carcinomas. Most investigations have

Keratinocyte Carcinomas been quite small leading to a wide range of estimates for mutation prevalence: 10%–100% in BCC (Rady et al., 1992; Campbell et al., 1993; Moles et al., 1993; Ziegler et al., 1993; Kubo et al., 1994; D’Errico et al., 1997; Ponten et al., 1997; Rosenstein et al., 1999; Ling et al., 2001; Zhang H et al., 2001; Kim et al., 2002; Weihrauch et al., 2002) and 15%–72% in SCC (Brash et al., 1991; Pierceall et al., 1991; Moles et al., 1993; Natarag et al., 1993; Nelson et al., 1993; Kubo et al., 1994); however, the weight of evidence indicates that approximately 50% of keratinocyte tumors have a TP53 mutation. Because of its clear role in the pathogenesis of keratinocyte carcinomas, the molecular profile of TP53 mutations dominates the somatic mutation literature. UVB induces a very specific DNA lesion, the pyrimidine dimer, and when unrepaired this lesion results in a signature mutation attributable to UV exposure, the CC-TT or C-T mutation (Drouin et al., 1997). The spectra of mutations at TP53 clearly demonstrate that UVL induces mutation at this gene: 90% of mutations are C to T transitions (Brash et al., 1991; Ziegler et al., 1993). In addition, this spectrum is nearly indistinguishable comparing AK, SCC, and BCC, suggesting they may arise from a common stem cell or are induced through a similar mechanism. Importantly, sunscreen application has been shown to inhibit the formation of dimers (Al Mahroos et al., 2002), p53 patches, and UVL-specific TP53 mutations in tumors (Rosenstein et al., 1999).

Ptch Basal cell nevus syndrome (BCNS) is a rare genetic disorder that predisposes carriers to develop high numbers of basal cell carcinomas at a young age (Gorlin, 1987). The human homologue of Drosophila patched (Ptch) has been identified as the gene responsible for the majority of BCNS (Hahn et al., 1996; Johnson et al., 1996). In addition to inheritance of a mutant ptch gene, tumors from BCNS patients demonstrate loss (either through deletion or mutation) of the second ptch allele. Similarly, sporadically arising BCC tumors have a high prevalence of both ptch mutation (Gailani et al., 1996; Unden et al., 1996; Wolter et al., 1997; Zhang H et al., 2001; Kim et al., 2002) and chromosome 9q22 deletion (where the ptch gene is located) (Ling, 2001; Kim, 2002). The ptch gene has not been extensively studied in SCC, and questions remain as to whether it is a true ‘gatekeeper’ gene for BCC. Similar to TP53, the mutation spectra for ptch is dominated by sunlight specific mutations (CC-TT), however, rather than missense mutations most ptch mutation events result in truncated proteins.

Other Genetic Alterations Other, less well-investigated genetic alterations in keratinocyte carcinomas include genetic instability (Ahmadian et al., 1998; Danaee et al., 2002; Hussein et al., 2002), p16 alteration (Kubo et al., 1997; Soufir et al., 1999; Saridaki et al., 2000), and downstream events in the p53 and ptch pathways (Dahmane et al., 1997; Reifenberger et al., 1998; Xie et al., 1998; Soufir et al., 1999).

PREVENTIVE MEASURES Prevention of KC is generally based on avoidance of the exposures that are understood to be of etiologic importance. In general, these preventive measures have not been documented to be effective by randomized trials, although their use is widely advocated for the purpose of skin cancer prevention based upon observational data and presumed etiologic mechanisms. The one exception to date is sunscreen use. Three randomized trials have documented the efficacy of sunscreen use in reducing the incidence and/or multiplicity of actinic keratoses, a keratinocyte dysplasia and precursor of SCC (Thompson et al., 1993; Naylor et al., 1995; Darlington et al., 2003). One randomized trial has provided direct evidence that sunscreen use can be effective in reducing the number of SCCs over a 4 1/2-year period (Green et al., 1999). This study found no indication that sunscreen use over this period prevented BCC.

1243

Considerable interest has focused on the use of pharmacologic agents (“chemoprevention”). A landmark non-randomized study was conducted in patients with xeroderma pigmentosum. That report noted a dramatic decrease of over 50% in KC on a high dose of the retinoid isotretinoin (2 mg per kg per day for 2 years) (Kraemer et al., 1988; Kraemer et al., 1992). The incidence increased back to prior levels after treatment was stopped. The bacterial DNA repair enzyme T4 endonuclease V, encapsulated in liposomes for penetration into skin cells and delivered as a lotion applied to the skin daily for 1 year, was found to reduce BCC incidence by 30% in xeroderma pigmentosum patients randomized to the intervention group (Yarosh et al., 2001). Another high-risk population is the recipients of organ transplants. Two randomized trials have evaluated acetretin (another retinoid) 25 and 30 mg daily in high-risk subsets of this group; one was a 6-month trial, the other used a cross-over design with 1 year in each arm. Both noted substantial reductions in SCC, the major skin cancer in this group, but both noted significant side effects of treatment as well (Bavinck et al., 1995; George et al., 2002). Another randomized trial compared doses of acetretin (0.2 vs. 0.4 mg/kg) for 9 months after a 3-month period of 0.4 mg/kg, and found no differences between the two groups, but statistical power may have been inadequate (de Sevaux et al., 2003). Multiple chemoprevention trials have been conducted with other agents in other high-risk populations that are not immunosuppressed. The agents studied in this context include retinol (25,000 IU daily), low-dose isotretinoin (10 mg daily), and fenretinide (200 mg daily). The two reported retinol trials, both from the same research group, had conflicting results. The first noted a significant hazard ratio of 0.68 for SCC but no effect for BCC, and the second noted no effect for either (Levine et al., 1997; Moon et al., 1997). There were important differences between the trials: the first was longer in duration (5 years vs. 3 years), and recruited a larger number (n = 2297) of participants, but excluded those with more than two prior keratinocyte carcinomas, whereas the latter recruited 525 patients, all with at least four carcinomas. Both low-dose isotretinoin studies reported null effects (Tangrea et al., 1992; Levine et al., 1997), and the fenretinide trial results, to our knowledge, have not been reported (De Palo et al., 1995). Systemic retinoids may have substantial side effects; hence their use is limited. Topical tretinoin is also being evaluated in a randomized trial that is not yet complete. Indeed, no chemopreventive or dietary interventions are currently recommended outside of special circumstances of very high risk, such as selected patients with solid organ transplants or xeroderma pigmentosum. Public health messages focus on reduction of exposure to ultraviolet radiation for reduction of skin cancer risk, and generally include but are not restricted to sunscreen use.

FUTURE DIRECTIONS Keratinocyte carcinomas challenge us in many ways. They are the most common cancers in the United States, and given their low mortality rate, morbidity is the key measure of their impact. However, population-based assessment of morbidity is non-existent beyond simple incidence, and even incidence is only measured sporadically, and in selected locations. Even mortality, which is comprehensively studied for other malignancies, is poorly measured for these tumors and although the mortality rate is low, the high incidence rate of KC translates to a significant number of deaths. Clearly, then, a future direction is to improve this sorry state of affairs. Achieving a more complete understanding of etiology and pathogenesis of KCs is critically important if these tumors are to be prevented in the future. The undisputed etiology relates to exposure to ultraviolet light from the sun, but our understanding of this exposure, as well as artificial UVL exposure, is as yet incomplete. Moreover, there are other factors, including inherited susceptibility, that greatly impact risk, and still others, which may act alone or in concert with UVL, such as low-level environmental exposures, underlying immunosuppressive states, and HPV infection. The refinement of molecular-genetic approaches will permit more precise risk

1244

PART IV: CANCER BY TISSUE OF ORIGIN

stratification of the population. Further, characterization of tumors on a molecular-genetic level in population-based studies offers enormous potential for illuminating the pathogenesis of these malignancies. Indeed, future investigations of KC cancer will almost certainly contribute more broadly to our understanding of the etiology and prevention of human malignancies. Our limited knowledge of etiology has given rise to huge prevention and early detection efforts. We have allocated large amounts of resources to technological interventions in the human population even though we have little idea of their effect—most notable is the widespread use of sunscreen lotions however inadequately applied. Therefore, it would be of some benefit to estimate the impact of these efforts on morbidity and mortality. Additionally, the design of novel chemopreventive agents combined with an improved understanding of the molecular targets of these agents open possibilities for future chemoprevention trials. References Abdel-Rahman S, El-Zein RA. 2000. The 399Gln polymorphism in the DNA repair gene XRCC1 modulates the genotoxic response induced in human lymphocytes by the tobacco-specific nitrosamine NNK. Cancer Lett 159:63–71. Abo-Darub J, Mackie R, Pitts JD. 1978. DNA repair deficiency in lymphocytes from patients with actinic keratosis. Bull Cancer 3:357–362. Ahmadian A, Ren ZP, Williams C, et al. 1998. Genetic instability in the 9q22.3 region is a late event in the development of squamous cell carcinoma. Oncogene 17:1837–1843. Al Mahroos M, Yaar M, Phillips TJ, Bhawan J, Gilchrest BA. 2002. Effect of sunscreen application on UV-induced thymine dimers. Arch Dermatol 138:1480–1485. Ames F, Hickey RC. 1982. Metastasis from squamous cell skin cancer of the extremities. South Med J 75:920–923. Armstrong B. 1988. The epidemiology and prevention of cancer in Australia. Aust N Z J Surg 58:179–187. Askling J, Sorensen P, Ekbom A, et al. 1999. Is history of squamous-cell skin cancer a marker of poor prognosis in patients with cancer? Ann Intern Med 131:655–659. Astori G, Lavergne D, Benton C, et al. 1998. Human papillomaviruses are commonly found in normal skin of immunocompetent hosts. J Invest Dermatol 110(5):752–755. Athas WF, Hunt WC, Key CR. 2003. Changes in Nonmelanoma Skin Cancer Incidence between 1977–1978 and 1998–1999 in Northcentral New Mexico. Cancer Epidemiol Biomarkers Prev 12(10):1105–1108. Aubry FaM B. 1985. Risk factors of squamous cell carcinoma of the skin. A case-control study in the Montreal region. Cancer 55:907–911. Bajdik CD, Gallagher RP, Astrakianakis G, et al. 1996. Non-solar ultraviolet radiation and the risk of basal and squamous cell skin cancer. Br J Cancer 73(12):1612–1614. Bajdik C, Gallagher R, Hill G, Fincham S. 1998. Sunlight exposure, hat use, and squamous cell skin cancer on the head and neck. Journal of Cutaneous Medicine & Surgery 3(2):68–73. Banks BA, Silverman RA, Schwartz RH, et al. 1992. Attitudes of teenagers toward sun exposure and sunscreen use. Pediatrics 89(1):40–42. Bastiaens M, Struyk L, Tjong-A-Hung SP, et al. 2001a. Cutaneous squamous cell carcinoma and p53 codon 72 polymorphism: A need for screening? Mol Carcinogen 30:56–61. Bastiaens M, Huurne J, Kielich C, et al. 2001b. Melanocortin-1 receptor gene variants determine the risk of nonmelanoma skin cancer independently of fair skin and red hair. Am J Hum Genet 68:884–894. Bastiaens M, Huurne J, Gruis N, et al. 2001c. The melanocortin-1-receptor gene is the major freckle gene. Hum Mol Gen 10:1701–1708. Bavinck JN, Tieben LM, Van der Woude FJ, et al. 1995. Prevention of skin cancer and reduction of keratotic skin lesions during acitretin therapy in renal transplant recipients: A double-blind, placebo-controlled study. J Clin Oncol 13(8):1933–1938. Bellaney GJ, Proby CM, Hawk JL. 1996. Likely photosensitizing agents available in the United Kingdom—an update. Clin Exp Dermatol 21(1):14–16. Berg JW, Burbank F. 1972. Correlations between carcinogenic trace metals in water supplies and cancer mortality. Ann NY Acad Sci 199:249–261. Bergoglio R. 1964. Mortalidad por cancer en zonas de aguas arsenicales de la Proviniciade Cordoba, Republica Argentina. Pren Med Argent 51: 994–998. Berkhout RJ, Bouwes Bavinck JN, ter Schegget J. 2000. Persistence of human papillomavirus DNA in benign and (pre)malignant skin lesions from renal transplant recipients. J Clin Microbiol 38(6):2087–2096.

Berwick M. 1999. Multiple primary skin cancers. In: Neugut AI, Robinson E, Meadows AT, eds. Multiple Primary Cancers. New York: Lippincott, Williams and Wilkins. Black HS, Herd JA, Goldberg LH, et al. 1994. Effect of a low-fat diet on the incidence of actinic keratosis. N Engl J Med 330:1272–1275. Black HS, Thornby JI, Wolf JE, et al. 1995. Evidence that a low-fat diet reduces the occurrence of non-melanoma skin cancer. Int J Cancer 62(2):165–169. Blomberg J, Moller T, Olsson H, et al. 1993. Cancer morbidity in blood recipients—results of a cohort study. Eur J Cancer 29A(15):2101–2105. Boffetta P, Jourenkova N, Gustavsson P. 1997. Cancer risk from occupational and environmental exposure to polycylic aromatic hydrocarbons. Cancer Causes Control 8:444–472. Bourguet CC, Checkoway H, Hulka BS. 1987. A case-control study of skin cancer in the tire and rubber manufacturing industry. Am J Ind Med 11(4):461–473. Bouwes Bavinck JN, Hardie DR, Green A, et al. 1996. The risk of skin cancer in renal transplant recipients in Queensland, Australia. A follow-up study. Transplantation 61(5):715–721. Bouwes Bavinck JN, Stark S, Petridis AK, et al. 2000. The presence of antibodies against virus-like particles of epidermodysplasia verruciformisassociated human papillomavirus type 8 in patients with actinic keratoses. Br J Dermatol 142(1):103–109. Box N, Duffy DL, Irving RE, et al. 2001. Melanocortin-1 receptor genotype is a risk factor for basal and squamous cell carcinoma. J Invest Dermatol 116:224–229. Boxman IL, Berkhout JM, Mulder LH, et al. 1997. Detection of human papillomavirus DNA in plucked hairs from renal transplant recipients and health volunteers. J Invest Dermatol 108:712–715. Boxman IL, Russell A, Mulder LH, et al. 2000. Case-control study in a subtropical Australian population to assess the relation between nonmelanoma skin cancer and epidermodysplasia verruciformis human papillomavirus DNA in plucked eyebrow hairs. The Nambour Skin Cancer Prevention Study Group. Int J Cancer 86(1):118–121. Boyd AS, Shyr Y, King LE Jr. 2002. Basal cell carcinoma in young women: An evaluation of the association of tanning bed use and smoking. J Am Acad Dermatol 46(5):706–709. Brash E, Rudolph JA, Simon JA, et al. 1991. A role for sunlight in skin cancer: UV-induced p53 mutations in squamous cell carcinoma. Proc Natl Acad Sci U S A 88:10124–10128. Breslow RA, Alberg AJ, Helzlsouer KJ, et al. 1995. Serological precursors of cancer: Malignant melanoma, basal and squamous cell skin cancer, and prediagnostic levels of retinol, beta-carotene, lycopene, alpha-tocopherol, and selenium. Cancer Epidemiol Biomarkers Prev 4(8):837–842. Butler GJ, Neale R, et al. 2005. Nonsteroidal anti-inflammatory drugs and the risk of actinic keratoses and squamous cell cancers of the skin. J Am Acad Dermatol 53(6):966–972. Campbell C, Quinn AG, Ro Y-S, Angus B, Rees JL. 1993. p53 mutations are common and early events that precede tumor invasion in squamous cell neoplasia of the skin. J Invest Dermatol 100:746–748. Cebrian ME, Albores A, Aguilar M, et al. 1983. Chronic arsenic poisoning in the north of Mexico. Hum Toxicol 2(1):121–133. Chen CJ, Chuang YC, Lin TM, et al. 1985. Malignant neoplasms among residents of a blackfoot disease-endemic area in Taiwan: High-arsenic artesian well water and cancers. Cancer Res 45(11 Pt 2):5895–5899. Chen CJ, Kuo TL, Wu MM. 1988. Arsenic and cancers. Lancet 1(8582): 414–415. Chen C-J, Wang C-J. 1990. Ecologic correlation between arsenic level in well water and age-adjusted mortality from malignant neoplasms. Cancer Res 50:5470–5474. Chen YC, Guo YL, Su HJ, et al. 2003. Arsenic methylation and skin cancer risk in southwestern Taiwan. J Occup Environ Med 45(3):241–248. Clark LC, Combs GFJ, Turnbull BWSE, et al. 1996. Effects of selenium supplementation for cancer prevention in patients with carcinoma of the skin. A randomized controlled trial. Nutritional Prevention of Cancer Study Group. JAMA 276:1957–1963. Clark LC, Graham GF, Bray J. 1987. Non-melanoma skin cancer and plasma selenium: A prospective cohort study. Proceedings of the Third International Symposium of Selenium in Biology and Medicine. Combs Jr. GF, Spallholtz JE, Levander OA. Westport, CT: AVI:1122–1134. Clark LC, Graham GF, Crounse KG, et al. 1984. Plasma selenium and skin neoplasms: A case-control study. Nutr Cancer 6:13–21. Clavel CE, Huu VP, Durlach AP, et al. 1999. Mucosal oncogenic human papillomaviruses and extragenital Bowen disease. Cancer 86(2): 282–287. Cleaver J. 1968. Defective repair replication of DNA in Xeroderma Pigmentosum. Nature 218:615–656. Cooke K, McNoe B, Sharples K. 1991. Trends in nonmelanoma skin cancer mortality. NZ Med J 104:437–438.

Keratinocyte Carcinomas Cooksley C, Hwang LY, Waller DK, Ford CD. 1999. HIV-related malignancies: Community-based study using linkage of cancer registry and HIV registry data. Int J STD AIDS 10:795–802. Copcu E, Aktas A, Sisman N, et al. 2003. Thirty-one cases of Marjolin’s ulcer. Clin Exp Dermatol 28(2):138–141. Dahmane N, Lee J, Robins P, Heller P, Ruiz I Altaba A. 1997. Activation of the transcription factor Gli1 and the Sonic hedgehog signaling pathway in skin tumors. Nature 389:876–881. Danaee H, Nelson HH, Karagas MR, et al. 2002. Microsatellite instability at tetranucleotide repeats in skin and bladder cancer. Oncogene 21: 4894–4899. Dantal J, Hourmant M, Cantarovich D, et al. 1998. Effect of long-term immunosuppression in kidney-graft recipients on cancer incidence: Randomised comparison of two cyclosporin regimens. [see comment]. Lancet 351(9103):623–628. Darlington S, Williams G, Neale R, et al. 2003. A randomized controlled trial to assess sunscreen application and beta carotene supplementation in the prevention of solar keratoses. Arch Dermatol 139(4):451–455. Davies TW, Treasure FP, Welch AA, et al. 2002. Diet and basal cell skin cancer: Results from the EPIC-Norfolk cohort. Br J Dermatol 146(6):1017– 1022. De Hertog SA, Wensveen CA, et al. 2001. Relation between smoking and skin cancer. J Clin Oncol 19(1):231–238. De Palo G, Veronesi U, Marubini E, et al. 1995. Controlled clinical trials with fenretinide in breast cancer, basal cell carcinoma and oral leukoplakia. J Cell Biochem Suppl 22:11–17. de Sevaux RG, Smit JV, de Jong EM, et al. 2003. Acitretin treatment of premalignant and malignant skin disorders in renal transplant recipients: clinical effects of a randomized trial comparing two doses of acitretin. J Am Acad Dermatol 49(3):407–412. de Villiers EM, Lavergne D, McLaren K, et al. 1997. Prevailing papillomavirus types in non-melanoma carcinomas of the skin in renal allograft recipients. Int J Cancer 73(3):356–361. D’Errico M, Calcagnile AS, Corona R. et al. 1997. p53 mutations and chromosome instability in basal cell carcinomas developed at an early or late age. Cancer Res 57:747–752. D’Errico M, Calcagnile A, Iavarone I, et al. 1999. Factors that influence the DNA repair capacity of normal and skin cancer-affected individuals. Cancer Epidemiol Biomarkers Prev 8:553–559. Devesa SS, Grauman DJ, Blot WJ, et al. 1999. Atlas of Cancer Mortality in the United States, 1950–1994. NIH Publication No. 99-4564. Bethesda, MD: National Institutes of Health, National Cancer Institute. Diffey BL, Farr PM, Ferguson J, et al. 1990. Tanning with ultraviolet A sunbeds. BMJ 301(6755):773–774. Dinehart SM, Chu DZ, Maners AW, et al. 1990. Immunosuppression in patients with metastatic squamous cell carcinoma from the skin. J Dermatol Surg Oncol 16(3):271–274. Dorgan JF, Boakye NA, et al. 2004. Serum carotenoids and alpha-tocopherol and risk of nonmelanoma skin cancer. Cancer Epidemiol Biomarkers Prev 13(8):1276–1282. Drouin R, Therrien JP. 1997. UVB-induced cyclobutane pyrimidine dimer frequency correlates with skin cancer mutational hotspots in p53. Photochem Photobiol 66:719–726. Duell E, Wiencke JK, Cheng TJ, et al. 2000. Polymorphisms in the DNA repair genes XRCC1 and ERCC2 and biomarkers of DNA damage in human blood mononuclear cells [published erratum appears in Carcinogensis 2000 Jul;21(7):1457]. Carcinogenesis 21(5):965–971. Duffield-Lillico AJ, Slate EH, Reid ME, et al. 2003. Selenium supplementation and secondary prevention of nonmelanoma skin cancer in a randomized trial. J Natl Cancer Inst 95(19):1477–1481. Dunn JE, Levin EA, Linden G, et al. 1965. Skin cancer as a cause of death. Calif Med 102:361–363. Dwyer T, Blizzard L, Ashbolt R, Plumb J, Berwick M, Stankovich JM. 2002. Cutaneous melanin density of Caucasians measured by spectrophotometry and risk of malignant melanoma, basal cell carcinoma, and squamous cell carcinoma of the skin. Am J Epidemiol 155:614–621. Dybdahl M, Vogel U, Frentz G, Wallin H, Nexo BA. 1999. Polymorphisms in the DNA repair gene XPD: Correlations with risk and age at onset of basal cell carcinoma. Cancer Epidemiol Biomark Preven 8:77–81. Einspahr JA DS, Aickin M, Welch K, Bozzo P, Grogan T, Nelson M. 1997. Expression of p53 protein in actinic keratosis, adjacent, normalappearing, and non-sun-exposed human skin. Cancer Epidemiol Biomarkers Prev 6:583–587. Eltorai IM, Montroy RE, Kobayashi M, et al. 2002. Marjolin’s ulcer in patients with spinal cord injury. J Spinal Cord Med 25(3):191–196. Elwood JM, Lee JAH, Walter SD, et al. 1974. Relationship of melanoma and other skin cancer mortality to latitude and ultraviolet radiation in the United States and Canada. Int J Epidemiol 3:325–332.

1245

Engel A, Johnson ML, Haynes SG. 1988. Health effects of sunlight exposure in the United States. Results from the first National Health and Nutrition Examination Survey, 1971–1974. Arch Dermatol 124(1):72–79. English D, Armstrong BK, Kricker A, Winter JG, Heenan PJ, Randell PL. 1998a. Demographic characteristics, pigmentary and cutaneous risk factors for squamous cell carcinoma of the skin: A case-control study. Int J Cancer 76:628–634. English DR, Armstrong BK, Kricker A, 1998b. Reproducibility of reported measurements of sun exposure in a case-control study. Cancer Epidemiol Biomarkers Prev 7(10):857–863. English DR, Armstrong BK, Kricker A, et al. 1998c. Case-control study of sun exposure and squamous cell carcinoma of the skin. Int J Cancer 77:347–353. English DR, Kricker A, Heenan PJ, et al. 1997. Incidence of Non-Melanocytic Skin Cancer in Geraldton, Western Australia. Int J Cancer 73:629–633. Epstein JH. 1999. Phototoxicity and photoallergy. Semin Cutan Med Surg 18(4):274–284. Feltkamp MC, Broer R, di Summa FM, et al. 2003. Seroreactivity to epidemodysplasia verruciformis-related human papilloma virus types is associated with nonmelanoms skin cancer. Cancer Res 63(10):2695–2700. Fitzpatrick T. 1988. The validity and practicality of sun-reactive skin types I through VI. Arch Dermatol 124:869–871. Flanagan N, Healy E, Ray A, et al. 2000. Pleiotropic effects of the melanocortin 1 receptor (MC1R) gene on human pigmentation. Hum Mol Genet 9:2531–2537. Forbes PD. 1981. Experimental photocarcinogenesis an overview. J Investigative Dermatol 77:139–143. Fortina AB, Piaserico S, et al. 2004. Immunosuppressive level and other risk factors for basal cell carcinoma and squamous cell carcinoma in heart transplant recipients. Arch Dermatol 140(9):1079–1085. Franceschi S, Dal Maso L, Arniani S, et al. 1998. Risk of cancer other than Kaposi’s sarcoma and non-Hodgkin’s lymphoma in persons with AIDS in Italy. Cancer and AIDS Registry linkage study. Br J Cancer 78: 966–970. Franceschi S, Levi F, Randimbison L, et al. 1996. Site distribution of different types of skin cancer: New aetiological clues. Int J Cancer 67(1):24–28. Freedman DM, Sigurdson A, et al. 2003. Risk of basal cell carcinoma in relation to alcohol intake and smoking. Cancer Epidemiol Biomarkers Prev 12(12):1540–1543. Frezza EE, Fung JJ, van Thiel DH. 1997. Non-lymphoid cancer after liver transplantation. Hepatogastroenterlogy 44(16):1172–1181. Frieling UM, Schaumberg DA, Kupper TS, et al. 2000. A randomized, 12-year primary-prevention trial of beta carotene supplementation for nonmelanoma skin cancer in the physician’s health study. Arch Dermatol 136(2):179–184. Frisch M, Melbye M. 1995. New primary cancers after squamous cell skin cancer. Am J Epidemiol 141:916–922. Frisch M, Hjalgrim H, Olsen JH, Melbye M. 1996. Risk for subsequent cancer after diagnosis of basal-cell carcinoma. Ann Intern Med 125:815–821. Fung TT, Hunter DJ, Spiegelman D, et al. 2002a. Vitamins and carotenoids intake and the risk of basal cell carcinoma of the skin in women (United States). Cancer Causes Control 13(3):221–230. Fung TT, Hunter DJ, Spiegelman D, et al. 2002b. Intake of alcohol and alcoholic beverages and the risk of basal cell carcinoma of the skin. Cancer Epidemiol Biomarkers Prev 11(10 Pt 1):1119–1122. Fung TT, Spiegelman D, et al. 2003. Vitamin and carotenoid intake and risk of squamous cell carcinoma of the skin. Int J Cancer 103(1):110–115. Gafa L, Filippazzo MG, et al. 1991. Risk factors of nonmelanoma skin cancer in Ragusa, Sicily: a case-control study. Cancer Causes Control 2(6): 395–399. Gailani MR, Stahle-Backdahl M, Leffell DJ, et al. 1996. The role of the human homologue of Drosophila patched in sporadic basal cell carcinomas. Nat Genet 14(1):78–81. Gallagher RP, Bajdik CD, Fincham S, et al. 1996. Chemical exposures, medical history, and risk of squamous and basal cell carcinoma of the skin. Cancer Epidemiol Biomarkers Prev 5(6):419–424. Gallagher RP, Hill GB, Bajdik CD, et al. 1995a. Sunlight exposure, pigmentary factors, and risk of nonmelanocytic skin cancer I. basal cell carcinoma. Arch Dermatol 131:157–163. Gallagher RP, Hill GB, et al. 1995b. Sunlight exposure, pigmentation factors, and risk of nonmelanocytic skin cancer. II. Squamous cell carcinoma. Arch Dermatol 131(2):164–169. Gallagher RP, Ma B, McLean DI, et al. 1990. Trends in basal cell carcinoma, squamous cell carcinoma, and melanoma of the skin from 1973 through 1987. J Am Acad Dermatol 23(3 Pt 1):413–421. Gamble JF, Lerman SE, et al. 1996. Physician-based case-control study of nonmelanoma skin cancer in Baytown, Texas. Occup Med (Lond) 46(3): 186–196.

1246

PART IV: CANCER BY TISSUE OF ORIGIN

Geller AC, Colditz G, Oliveria S, et al. 2002. Use of sunscreen, sunburning rates, and tanning bed use among more than 10 000 US children and adolescents. Pediatrics 109(6):1009–1014. Gellin GA, Kopf AW, et al. 1965. Basal Cell Epithelioma. A Controlled Study Of Associated Factors. Arch Dermatol 91:38–45. George R, Weightman W, Russ GR, et al. 2002. Acitretin for chemoprevention of non-melanoma skin cancers in renal transplant recipients. Australas J Dermatol 43(4):269–273. Gilbody J, Aitken J, Green A. 1994. What causes basal cell carcinoma to be the commonest cancer. Aust J Public Health 18:218–221. Giles G, Dwyer T, Coates MS, et al. 1989. Trends in skin cancer in Australia: An overview of the available data. Transactions of the Menzies Foundation 15:143–147. Glass A, Hoover R. 1989. The emerging epidemic of melanoma and squamous cell skin cancer. JAMA 262(15):2097–2100. Gocke E, Albertini S, Chetelat AA, et al. 1998. The photomutagenicity of fluoroquinolones and other drugs. Toxicol Lett 102–103:375–381. Gordon D, Silverstone H. 1969. Deaths from skin cancer in Queensland, Australia. In: Urbach F, ed. The Biologic Effects of Ultraviolet Radiation. New York; Pergammon Press: pp. 625–634. Gorlin RJ. 1987. Nevoid basal-cell carcinoma syndrome. Medicine 66:98– 113. Gould JW, Mercurio MG, Elmets CA. 1995. Cutaneous photosensitivity diseases induced by exogenous agents. J Am Acad Dermatol 33(4): 551–573; quiz 574–576. Gray D, Suman V, Su W, et al. 1997. Trends in the population-based incidence of squamous cell carcinoma of the skin first diagnosed between 1984 and 1992. Arch Dermatol 133:735–740. Green A, Battistutta D. 1990. Incidence and determinants of skin cancer in a high-risk Australian population. Int J Cancer 46:356–361. Green A, Battistutta D, et al. 1996. Skin cancer in a subtropical Australian population: incidence and lack of association with occupation. The Nambour Study Group. Am J Epidemiol 144(11):1034– 1040. Green A, Williams G, Neale R. 2000. Does daily use of sunscreen or beta-carotene supplements prevent skin cancer in healthy adults? West J Med 173(5):332. Green A, Williams G, Neale R, et al. 1999. Daily sunscreen application and betacarotene supplementation in prevention of basal-cell and squamouscell carcinomas of the skin: A randomised controlled trial. Lancet 354:723–729. Greenberg ER, Baron JA, Stukel TA, et al. 1990. A clinical trial of beta carotene to prevent basal-cell and squamous-cell cancers of the skin. N Engl J Med 323:789–795. Grodstein F, Speizer FE, Hunter DJ. 1995. A prospective study of incident squamous cell carcinoma of the skin in the Nurses’ Health Study. J Nat Cancer Inst 87:1061–1066. Guo HR, Chiang HS, Hu H, et al. 1997. Arsenic in drinking water and incidence of urinary cancers. Epidemiology 8(5):545–550. Hahn H, Wicking C, Zaphiropoulous PG, et al. 1996. Mutations of the human homolog of Drosophila patched in the nevoid basal cell carcinoma syndrome. Cell 85(6):841–851. Hakim IA, Harris RB, et al. 2000a. Fat intake and risk of squamous cell carcinoma of the skin. Nutr Cancer 36(2):155–162. Hakim IA, Harris RB, Weisgerber UM. 2000b. Tea intake and squamous cell carcinoma of the skin: Influence of type of tea beverages. Cancer Epidemiol Biomarkers Prev 9(7):727–731. Hall J, English DR, Artuso M, Armstrong BK, Winter M. 1994. DNA repair capacity as a risk factor for non-melanocytic skin cancer—a molecular epidemiologic study. Int J Cancer 58:179–184. Hannuksela-Svahn A, Pukkala E, Karvonen J. 1999. Basal cell skin carcinoma and other nonmelanoma skin cancers in Finland from 1956 through 1995 [see comment]. Arch Dermatol 135(7):781–786. Hannuksela-Svahn A, Pukkala E, Laara E, et al. 2000. Psoriasis, its treatment, and cancer in a cohort of Finnish patients. J Invest Dermatol 114(3): 587–590. Harris RB, Griffith K, Moon TE. 2001. Trends in the incidence of nonmelanoma skin cancers in southeastern Arizona, 1985–1996. J Am Acad Dermatol 45(4):528–536. Harris RB, Foote JA, et al. 2005. Fatty acid composition of red blood cell membranes and risk of squamous cell carcinoma of the skin. Cancer Epidemiol Biomarkers Prev 14(4):906–912. Hartevelt MM, Bavinck JN, Kootte AM, et al. 1990. Incidence of skin cancer after renal transplantation in The Netherlands. Transplantation 49(3): 506–509. Harwood CA, Spink PJ, Surentheran T, et al. 1998. Detection of human papillomaviruses in PUVA associated non-melanoma skin cancers. J Invest Dermatol 111(1):123–127.

Harwood CA, Surentheran T, McGregor JM, et al. 2000. Human papillomavirus infection and non-melanoma skin cancer in immunosuppressed and immunocompetent individuals. J Med Virol 61(3):289–297. Hayes J, Stranger RC. 1995. Potential contribution of the glutathione Sgransferase supergene family to resistance to oxidative stress. Free Rad Res 22:193–207. Heagerty A, Fitzgerald D, Smith A, et al. 1994. Glutathione S-transferase GSTM1 phenotypes and protection against cutaneous tumours. Lancet 343:266–268. Healy E, Flannagan N, Ray A, et al. 2000. Melanocortin-1 receptor gene and sun sensitivity in individuals without red hair. Lancet 355:1072–1073. Hemminki K, Xu G, Angelini S, et al. 2001. XPD exon 10 and 23 polymorphisms and DNA repair in human skin in situ. Carcinogenesis 22:1185. Hemminki K, Zhang H, Czene K. 2003. Time trends and familial risks in squamous cell carcinoma of the skin. Arch Dermatol 139:885–889. Herity B, O’Loughlin G, Moriarty MJ, et al. 1989. Risk factors for nonmelanoma skin cancer. Ir Med J 82(4):151–152. Hildreth NG, Shore RE, Hempelmann LH, et al. 1985. Risk of extrathyroid tumors following radiation treatment in infancy for thymic enlargement. Radiat Res 102:378–391. Hjalgrim H, Frisch M, Storm HH, et al. 2000. Non-melanoma skin cancer may be a marker of poor prognosis in patients with non-Hodgkin’s lymphoma. Int J Cancer 85(5):639–642. Hogan DJ, Lane PR, Gran L, et al. 1990. Risk factors for squamous cell carcinoma of the skin in Saskatchewan, Canada. J Dermatol Sc 1(2):97–101. Hogan DJ, To T, Gran L, et al. 1989. Risk factors for basal cell carcinoma. Int J Dermatol 28:591–594. Holman CDJ. 1982. Cancer mortality trends in Australia. University of Western Australia. Hou S, Falt S, Angelini S, et al. 2002. The XPD variant alleles are associated with increased aromatic DNA adduct level and lung cancer risk. Carcinogenesis 23:599–603. Hoy WE. 1996. Nonmelanoma skin carcinoma in Albuquerque, New Mexico: Experience of a major health care provider. Cancer 77(12):2489–2495. Hsueh YM, Chiou HY, Huang YL, et al. 1997. Serum beta-carotene level, arsenic methylation capability, and incidence of skin cancer. Cancer Epidemiol Biomarkers Prev 6(8):589–596. Hueper WC. 1963. Chemically induced skin cancers in man. NCI Monographs 10:377–391. Hunter D, Colditz Ga, Stampfer MJ, Rosner B, Willett WC, Speizer FE. 1990. Risk factors for basal cell carcinoma in a prospective cohort of women. Ann Epidemiol 1:13–23. Hunter DJ, Colditz GA, Stampfer MJ, et al. 1992. Diet and risk of basal cell carcinoma of the skin in a prospective cohort of women. Ann Epidemiol 2(3):231–239. Hussein MW GS. 2002. Microsatellite instability and its relevance to cutaneous tumorigenesis. J Cutan Pathol 29:257–267. IARC. 1980. Arsenic and arsenic compounds. Monographs on the Evaluation of Carcinogenic Risk of Chemicals to Humans. Lyon, France: IARC 23: pp. 39–141. IARC. 1992. Solar and Ultraviolet Radiation. Lyon, France: IARC Scientific Publications. IARC. 2002. Cancer in Africa: Epidemiology and Prevention. Lyon, France: IARC Scientific Publications. IARC. 2002. Cancer Incidence in Five Continents: Volume VIII. Lyon, France: IARC Scientific Publications. Jaafar R, Omar I, et al. 1993. Skin cancer caused by chronic arsenical poisoning—a report of three cases. Med J Malaysia 48(1):86–92. Jackson R, Grainge JW. 1975. Arsenic and cancer. Can Med Assoc J 113:396–399. Jackson S, Storey A. 2000. E6 proteins from diverse cutaneous HPV types inhibit apoptosis in response to UV damage. Oncogene 19(4):592–598. Jee SH, Kuo HW, Su WP, et al. 1995. Photodamage and skin cancer among paraquat workers. Int J Dermatol 34(7):466–469. Jemal A, Murray T, Samuels A, Ghafoor A, Ward E, Thun MJ. 2003. Cancer statistics, 2003. CA Cancer J Clin 53(1):5–26. Jensen P, Hansen S, Moller B, et al. 1999. Skin cancer in kidney and heart transplant recipients and different long-term immunosuppressive therapy regimens. J Am Acad Dermatol 40(2 Pt 1):177–186. Jiang W, Ananthaswamy HN, Muller HK, Kripke ML. 1999. p53 protects against skin cancer induction by UV-B radiation. Oncogene 18: 4247–4253. Johnson RL, Rothman AL, Xie J, et al. 1996. Human homolog of patched, a candidate gene for the basal cell nevus syndrome. Science 272: 1668–1671. Jonason A, Kunala S, Price GJ, et al. Frequent clones of p53-mutated keratinocytes in normal human skin. Proc Natl Acad Sci U S A 94: 11–14.

Keratinocyte Carcinomas Joseph AK, Mark TL, Mueller C. 2001. The period prevalence and costs of treating nonmelanoma skin cancers in patients over 65 years of age covered by medicare. Dermatol Surg 27:955–959. Kahn H, Tatham LM, Patel AV, Thun MJ, Heath CW. 1998. Increased cancer mortality following a history of nonmelanoma skin cancer. JAMA 280:910–912. Karagas M, Greenberg ER, Mott LA, Baron JA, Ernster VL. 1998. Occurrence of other cancers among patients with prior basal cell and squamous cell skin cancer. Cancer Epidemiol Biomarkers Prev 7:157–161. Karagas MR, Cushing GLJ, Greenberg ER, et al. 2001a. Risk of non-melanoma skin cancer associated with glucocorticoid therapy. Br J Cancer 85:683– 686. Karagas MR, Greenberg ER, Nierenberg D, et al. 1997. Risk of squamous cell carcinoma of the skin in relation to plasma selenium, alpha-tocopherol, beta-carotene, and retinol: A nested case-control study. Cancer Epidemiol Biomarkers Prev 6(1):25–29. Karagas MR, Greenberg ER, Spencer SK, et al. 1999. Increase in incidence rates of basal cell and squamous cell skin cancer in New Hampshire, USA. Int J Cancer 81:555–559. Karagas MR, McDonald JA, Greenberg ER, et al. 1996. Risk of basal cell and squamous cell skin cancers after ionizing radiation therapy. J Natl Cancer Inst 88(24):1848–1853. Karagas MR, Stannard VA, Mott LA, et al. 2002a. Use of tanning devices and risk of basal cell and squamous cell skin cancers. J Natl Cancer Inst 94(3):224–226. Karagas MR, Stukel TA, Greenberg ER, et al. 1992. Risk of subsequent basal cell carcinoma and squamous cell carcinoma of the skin among patients with prior skin cancer. JAMA 267(24):3305–3310. Karagas MR, Stukel TA, Morris JS, et al. 2001b. Skin cancer risk in relation to toenail arsenic concentrations in a US population-based case-control study. Am J Epidemiol 153:559–565. Karagas MR, Stukel TA, Tosteson TD. 2002b. Assessment of cancer risk and environmental levels of arsenic in New Hampshire. Int J Hyg Environ Health 205:85–94. Kennedy C, Huurne J, Berkhout M, et al. 2001. Melanocortin 1 receptor (MC1R) gene variants are associated with an increased risk for cutaneous melanoma which is largely independent of skin type and hair color. J Invest Dermatol 117:294–300. Kennedy C, Bajdik CD, et al. 2003. The influence of painful sunburns and lifetime sun exposure on the risk of actinic keratoses, seborrheic warts, melanocytic nevi, atypical nevi, and skin cancer. J Invest Dermatol 120(6):1087–1093. Kerb R, Brockmoller J, Reum T, Roots I. 1997. Deficiency of glutathione Stransferases T1 and M1 as heritable factors of increased cutaneous UV sensitivity. J Invest Dermatol 108:229–232. Kim M, Park HJ, Baek SC, Byun DG, Houh D. 2002. Mutations of the p53 and PTCH gene in basal cell carcinomas: UV mutation signature and strand bias. J Dermatological Sci 29:1–9. Kinlen LJ, Sheil AG, Peto J, et al. 1979. Collaborative United KingdomAustralasian study of cancer in patients treated with immunosuppressive drugs. Br Med J 2(6203):1461–1466. Knekt P, Aromaa A, Maatela J, et al. 1988. Serum vitamin E and risk of cancer among Finnish men during a 10-year follow-up. Am J Epidemiol 127(1):28–41. Knight A, Askling J, Ekbom A. 2002. Cancer incidence in a population-based cohort of patients with Wegener’s granulomatosis. Int J Cancer 100(1): 82–85. Ko C, Walton S, Keczkes K, et al. 1994. The emerging epidemic of skin cancer. Br J Dermatol 130:269–272. Koh D, Wang H, Lee J, et al. 2003. Basal cell carcinoma, squamous cell carcinoma and melanoma of the skin: Analysis of the Singapore Cancer Registry data 1968–97. Br J Dermatol 148(6):1161–1166. Koppula S, Robbins LS, Lu D, et al. 1997. Identification of common polymorphisms in the coding sequence of the human MSH receptor (MCIR) with possible biological effects. Hum Mutat 9:30–36. Kraemer K, Lee M, Scotto J. 1987. Xeroderma pigmentosum: Cutaneous, ocular, and neurological abnormalities in 830 published cases. Arch Dermatol 123:241–250. Kraemer KH, DiGiovanna JJ, Moshell AN, et al. 1988. Prevention of skin cancer in xeroderma pigmentosum with the use of oral isotretinoin. N Engl J Med 318:1633–1637. Kraemer KH, DiGiovanna JJ, Peck GL. 1992. Chemoprevention of skin cancer in xeroderma pigmentosum. J Dermatol 19(11):715–718. Kricker A, Armstrong BK, English DR, Heenan PJ. 1991. Pigmentary and cutaneous risk factors for non-melanocytic skin cancer—a case-control study. Int J Cancer 48:650–662. Kricker A, Armstrong BK, English DR. 1994. Sun exposure and nonmelanocytic skin cancer. Cancer Causes Control 5:367–392.

1247

Kricker A, Armstrong BK, English DR, et al. 1995a. A dose-response curve for sun exposure and basal cell carcinoma. Int J Cancer 60(4):482–488. Kricker A, Armstrong BK, English DR, et al. 1995b. Does intermittent sun exposure cause basal cell carcinoma? A case-control study in Western Australia. Int J Cancer 60(4):489–494. Kubo Y, Urano Y, Yoshimoto K, et al. 1994. p53 gene mutations in human skin cancers and precancerous lesions: Comparison with immunohistochemical analysis. J Invest Dermatol 102:440–444. Kubo YU Y, Matsumoto K, Ahsan K, Arase S. 1997. Mutations of the INK4a locus in squamous cell carcinomas of human skin. Biochemical & Biophysical Research Communications 232:38–41. Kune GA, Bannerman S, Field B, et al. 1992. Diet, alcohol, smoking, serum beta-carotene, and vitamin A in male nonmelanocytic skin cancer patients and controls. Nutr Cancer 18(3):237–244. Kurokawa M, Ogata K, et al. 2001. Investigation of skin manifestations of arsenicism due to intake of arsenic-contaminated groundwater in residents of Samta, Jessore, Bangladesh. Arch Dermatol 137(1):102–103. Lambert JD, Yang CS. 2003. Cancer chemopreventive activity and bioavailability of tea and tea polyphenols. Mut Res 523–524:201–208. Lawrence R, Moshell A, Safavi K, et al. 1994. Introduction. J Invest Dermatol 102:3S. Lear J, Heagerty AHM, Smith A, et al. 1996. Multiple cutaneous basal cell carcinomas: Glutathione S-transferase (GSTM1, GSTT1) and cytochrome p450 (CYP2D6, CYP1A1) polymorphisms influence tumour numbers and accrual. Carcinogenesis 17:1891–1896. Lear J, Tan BB, Smith AG, Bowers W, Jones PW, Heagerty AH. 1997. Risk factors for basal cell carcinoma in UK: Case-control study in 806 patients. J R Soc Med 90:371–374. Lear JT, Tan BB, Smith AG, et al. 1998. A comparison of risk factors for malignant melanoma, squamous cell carcinoma and basal cell carcinoma in the UK. Int J Clin Pract 52(3):145–149. Lee H, Kim JS, Ha SJ, et al. 2000. p53 gene mutations in Bowen’s disease in Koreans: Clustering in exon 5 and multiple mutations. Cancer Lett 158:27–33. Lee IM, Cook NR, Manson JE, et al. 1999. Beta-carotene supplementation and incidence of cancer and cardiovascular disease: The Women’s Health Study. J Natl Cancer Inst 91(24):2102–2106. Lei Y, Hwang SJ, Chang CC, et al. 2002. Effects on sister chromatid exchange frequency of polymorphisms in DNA repair gene XRCC1 in smokers. Mut Res 519:93–101. Levine N, Moon TE, Cartmel B, et al. 1997. Trial of retinol and isotretinoin in skin cancer prevention: A randomized, double-blind, controlled trial. Southwest Skin Cancer Prevention Study Group. Cancer Epidemiol Biomarkers Prev 6(11):957–961. Lewis K WM. 2004. Nonmelanoma Skin Cancer Mortality, 1988 to 2000: The Rhode Island Follow-Back Study. Arch Dermatol 140(7):837–842. Lewis KG, Weinstock MA. 2005. Trends in nonmelanoma skin cancer mortality in the United States, 1969 through 2000. J Invest Dermatol 2005;125:856. Lewis KG, Weinstock MA. 2004. Nonmelanoma skin cancer mortality (1988–2000): the Rhode Island follow-back study. Arch Dermatol 140(7):837–842. Lichter MD, Karagas MR, Mott LA, et al. 2000. Therapeutic ionizing radiation and the incidence of basal cell carcinoma and squamous cell carcinoma. The New Hampshire Skin Cancer Study Group. Arch Dermatol 136(8):1007–1011. Lindelof, Sigurgeirsson, Tegner, et al. 1999. PUVA and cancer risk: The Swedish follow-up study. Br J Dermatol 141(1):108–112. Ling G, Ahmadian A, Persson A, et al. 2001. PATCHED and p53 gene alterations in sporadic and hereditary basal cell cancer. Oncogene 20: 7770–7778. Lock-Andersen J, Drzewiecki KT, Wulf HC. 1998. The measurement of constitutive and facultive skin pigmentation and estimation of sun exposure in Caucasians with basal cell carcinoma and cutaneous malignant melanoma. Br J Dermatol 139:610–617. Lock-Andersen J, Drzewiecki KT, Wulf HC. 1999. Eye and hair colour, skin type and constitutive skin pigmentation as risk factors for basal cell carcinoma and cutaneous malignant melanoma. Acta Derm Venereol 79:74–80. Lock-Andersen J, Drzewiecki KT, Wulf HC. 1999. Naevi as a risk factor for basal cell carcinoma in Caucasians: A Danish case-control study. Acta Derm Venereol 79:314–319. Makinen M, Forbes PD, Stenback F. 1997. Quinolone antibacterials: A new class of photochemical carcinogens. J Photochem Photobiol B 37(3): 182–187. Marks R, Staples M, Giles GG. 1993. Trends in non-melanocytic skin cancer treated in Australia: The second national survey. Int J Cancer 53(4): 585–590.

1248

PART IV: CANCER BY TISSUE OF ORIGIN

Marshall S, Bordea C, Haldar NA, et al. 2000. Glutathione S-transferase polymorphisms and skin cancer after renal transplantation. Kidney Int 58:2186–2193. Marshall S, Bordea C, Wojnarowska F, Morris PJ, Welsh KI. 2001. p53 codon 72 polymorphism and susceptibility to skin cancer after renal transplantation. Transplantation 69:994–996. Masini C, Fuchs PG, Gabrielli F, et al. 2003. Evidence for the association of human papillomavirus infection and cutaneous squamous cell carcinoma in immunocompetent individuals. Arch Dermatol 139(7):890–894. Matanoski G, Seltser R, Sartwell P, et al. 1975. Current mortality rates of radiologists and other physician specialists: Specific causes of death. Am J Epidemiol 101:199–210. Matullo G, Palli D, Peluso M, et al. 2001. XRCC1, XRCC3, XPD gene polymorphisms, smoking and DNA-adducts in a sample of healthy subjects. Carcinogenesis 22:1437–1445. Milan T, Verkasalo PK, Kaprio J, et al. 2003. Lifestyle differences in twin pairs discordant for basal cell carcinoma of the skin. Br J Dermatol 149:115–123. Miller DL, Weinstock MA. 1994. Nonmelanoma skin cancer in the United States: Incidence. J Am Acad Dermatol 30:774–778. Moles JP, Moyret C, Guillot B, et al. 1993. p53 gene mutations in human epithelial skin cancers. Oncogene 8(3):583–588. Moon TE, Levine N, Cartmel B, et al. 1997. Effect of retinol in preventing squamous cell skin cancer in moderate-risk subjects: A randomized, double-blind, controlled trial. Southwest Skin Cancer Prevention Study Group. Cancer Epidemiol Biomarkers Prev 6(11):949–956. Morton W, Starr G, Pohl D, et al. 1976. Skin cancer and water arsenic in Lane Country, Oregon. Cancer 37:2523–2532. Munch-Petersen B, Frentz G, Squire B, et al. 1985. Abnormal lymphocyte response to UV radiation in multiple skin cancer. Carcinogenesis 6:843–845. Nakazawa H, English D, Randell PL, et al. 1994. UV and skin cancer: Specific p53 gene mutation in normal skin as a biologically relevant exposure measurement. Proc Natl Acad Sci U S A 91:360–364. Naldi L, DiLandro A, D’Avanzo B, Parazzini F. 2000. Host-related and environmental risk factors for cutaneous basal cell carcinoma: Evidence from an Italian case-control study. J Am Acad Dermatol 42:446–452. Natarag A, Wolf P, Cerroni L, Ananthaswamy HN. 1993. p53 mutation in squamous cell carcinomas from psoriasis patients treated with psoralen + UVA (PUVA). J Invest Dermatol 109:238–243. National Institutes of Health. 1991. Summary of the Consensus Development Conference on Sunlight, Ultraviolet Radiation, and the Skin. Bethesda, Maryland, May 8–10, 1989. Consensus Development Panel. J Am Acad Dermatol 24(4):608–612. Naylor MF, Boyd A, Smith DW, et al. 1995. High sun protection factor sunscreens in the suppression of actinic neoplasia. Arch Dermatol 131: 170–175. Nelson H, Kelsey KT, Mott LA, Karagas MR. 2002. The XRCC1 Arg399Gln polymorphism, sunburn, and non-melanoma skin cancer: Evidence of gene-environment interaction. Cancer Res 62:152–155. Nelson M, Einspahr JG, Alberts DS, et al. 1993. Analysis of the p53 gene in human precancerous actinic keratosis lesions and squamous cell cancers. Cancer Lett 85:23–29. Neubauer O. 1947. Arsenical cancer: A review. Br J Cancer 1:192–251. Oliphant JA, Forster JL, McBride CM. 1994. The use of commercial tanning facilities by suburban Minnesota adolescents. Am J Public Health 84: 476–478. O’Loughlin C, Moriarty MJ, Herity B, et al. 1985. A re-appraisal of risk factors for skin carcinoma in Ireland. A case control study. Ir J Med Sci 154(2):61–65. Olsson H, Bladstrom A, Alm P. 2002. Male gynecomastia and risk for malignant tumours—a cohort study. BMC Cancer 2(1):26. Ong CS, Keogh AM, Kossard S, et al. 1999. Skin cancer in Australian heart transplant recipients. J Am Acad Dermatol 40:27–34. Orth G, Favre M, Breitburd F. 1980. Epidermodysplasia verruciformis: A model for the role of papillomaviruses in human cancer. Viruses in Naturally Occurring Cancer. In: Essex M, Todaro G, eds zur Hausen H, Cold Spring Harbor, NY: Cold Spring Harbor Laboratory 7:259–282. Osterlind A, Hjalgrim H, Kulinsky B, et al. 1991. Skin cancer as a cause of death in Denmark. Br J Dermatol 125:580–582. Ouhtit AN H, Armstrong BK, Kricker A, Tan E, Yamasaki H, English DR. 1998. UV-radiation-specific p53 mutation frequency in normal skin as a predictor of risk of basal cell carcinoma. J Natl Cancer Inst 90:523– 531. Ouhtit AU M, Nakazawa H, Ichihashi M, Dumaz N, Sarasin A, Yamasaki H. 1997. Quantitative detection of ultraviolet-specific p53 mutations in normal skin from Japanese patients. Cancer Epidemiol Biomarkers Prev 6:433–438.

Palli D, Russo A, Masala G, et al. 2001. DNA adduct levels and DNA repair polymorphisms in traffic-exposed workers and a general population sample. Int J Cancer 94:121–127. Palmer J, Duffy DL, Box NF, et al. 2000. Melanocortin-1 receptor polymorphisms and risk of melanoma: Is the association explained solely by pigmentation phenotype. Am J Hum Genet 66:176–186. Pearl DK, Scott EL. 1986. The anatomic distribution of skin cancers. Int J Epidemiol 15:502–506. Perkins JL, Liu Y, et al. 2005. Nonmelanoma skin cancer in survivors of childhood and adolescent cancer: a report from the childhood cancer survivor study. J Clin Oncol 23(16):3733–3741. Pesch B, Ranft U, Jakubis P, et al. 2002. Environmental arsenic exposure from a coal-burning power plant as a potential risk factor for nonmelanoma skin carcinoma: results from a case-control study in the district of Prievidza, Slovakia. Am J Epidemiol 155(9):798–809. Pierceall W, Mukhopadhyay T, Goldberg LH, Anathaswamy HN. 1991. Mutations in the p53 tumor suppressor gene in human cutaneous squamous cell carcinomas. Mol Carcinog 4:445–449. Placzek M, Eberllein-Konig B, Przybilla B. 1999. Association between actinic keratoses and potentially photosensitizing drugs. N Engl J Med 341(19): 1474–1475. Ponten F, Berg C, Ahmadian A, et al. 1997. Molecular pathology in basal cell cancer with p53 as a genetic marker. Oncogene 15:1059–1067. Purdie KJ, Pennington J, Proby CM, et al. 1999. The promoter of a novel human papillomavirus (HPV77) associated with skin cancer displays UV responsiveness, which is mediated through a consensus p53 binding sequence. EMBO J 18(19):5359–5369. Radis CD, Kahl LE, Baker GL, et al. 1995. Effects of cyclophosphamide on the development of malignancy and on long-term survival of patients with rheumatoid arthritis: A 20-year followup study. Arthritis Rheum 38(8): 1120–1127. Rady P, Scinicariello F, Wagner RF, et al. 1992. p53 mutations in basal cell carcinomas. Cancer Res 52(13):3804–3806. Ramachandran S, Hoban PR, Ichii-Jones F, et al. 2000. Glutathione Stransferase GSTP1 and cyclin D1 genotypes: Association with numbers of basal cell carcinomas in a patient subgroup aqt high-risk of multiple tumours. Pharmacogene 10:545–556. Ramsay H, Harden PN, Reece S, et al. 2001. Polymorphisms in glutathione Stransferases are associated with altered risk of nonmelanoma skin cancer in renal transplant recipients. J Invest Dermatol 117:251–255. Ranadive N, Shirwadkar S, Persad S, Menon IA. 1986. Effects of melanin induced free radicals on the isolated rat peritoneal mast cells. J Invest Dermatol 86:303–307. Ranft U, Miskovic P, et al. 2003. Association between arsenic exposure from a coal-burning power plant and urinary arsenic concentrations in Prievidza District, Slovakia. Environ Health Perspect 111(7):889– 894. Reifenberger J, Wolter M, Weber RG, et al. 1998. Missense mutations in SMOH in sporadic basal cell carcinomas of the skin and primitive neuroectodermal tumors of the central nervous system. Cancer Res 58(9): 1798–1803. Rhainds M, De Guire L, Claveau J. 1999. A population-based survey on the use of artificial tanning devices in the Province of Quebec, Canada. J Am Acad Dermatol 40(4):572–576. Rivara MI, Cebrian M, Corey G, et al. 1997. Cancer risk in an arseniccontaminated area of Chile. Toxicol Ind Health 13(2–3):321–338. Robinson JK. 1987. Risk of developing another basal cell carcinoma: A 5-year prospective study. Cancer 60:118–120. Ron E, Modan B, Preston D, et al. 1991. Radiation-induced skin carcinomas of the head and neck. Radiat Res 125(3):318–325. Ron E, Preston DL, Kishikawa M, et al. 1998. Skin tumor risk among atomicbomb survivors in Japan. Cancer Causes Control 9(4):393–401. Rosenblatt L, Marks R. 1996. Deaths due to squamous cell carcinoma in Australia: Is there a case for a public health intervention? Australas J Dermatol 37(1):26–29. Rosenstein B, Phelps RG, Weinstock MA, et al. 1999. p53 mutations in basal cell carcinomas arising in routine users of sunscreens. Photochem Photobiol 70:798–806. Rosso S, Joris F, Zanetti R. 1999. Risk of basal and squamous cell carcinomas of the skin in Sion, Switzerland: A case-control study. Tumori 85: 435–442. Rosso S, Zanetti R, Martinez C, et al. 1996. The multicentre south European study “Helios”. II: Different sun exposure patterns in the aetiology of basal cell and squamous cell carcinomas of the skin. Br J Cancer 73(11): 1447–1454. Rosso S, Zanetti R, Pippione M, et al. 1998. Parallel risk assessment of melanoma and basal cell carcinoma: Skin characteristics and sun exposure. Melanoma Res 8(6):573–583.

Keratinocyte Carcinomas Saha KC. 2001. Cutaneous malignancy in arseniocosis. Br J Dermatol 145: 169–192. Sahl WJ, Glore S, Garrison P, et al. 1995. Basal cell carcinoma and lifestyle characteristics. Int J Dermatol 34(6):398–402. Saridaki ZK E, Liloglou T, Sourvinos G, Papadopoulos O, Zoras O, Spandidos DA. 2000. High frequency of loss of heterozygosity on chromosome region 9p21-p22 but lack of p16INK4a/p19ARF mutations in greek patients with basal cell carcinoma of the skin. J Invest Dermatol 115:719–725. Sau P, McMarlin SL, Sperling LC, et al. 1994. Bowen’s disease of the nail bed and periungual area: A clinicopathologic analysis of seven cases. Arch Dermatol 130:204–209. Scotto J, Fears T, Fraumeni J. 1981. Incidence of nonmelanoma skin cancer in the United States. Department of Health and Human Services. Scotto J, Fears TR. 1987. The association of solar ultraviolet and skin melanoma incidence among caucasians in the United States. Cancer Invest 5:275–283. Scotto J, Kopf A, Urbach F. 1974. Nonmelanoma skin cancer among causasians in four areas of the United States. Cancer 34:1333–1338. Seker H, Butkiewicz D, Bowman ED, et al. 2001. Functional signficance of XPD polymorphic variants: Attenuated apoptosis in human lymphoblastoid cells with the XPD 312 Asp/Asp genotype. Cancer Res 61:7430–7434. Serrano H, Scotto J, Shornick G, et al. 1991. Incidence of nonmelanoma skin cancer in New Hampshire and Vermont. J Am Acad Dermatol 24(4): 574–579. Sevcova M, Sevc J, Thomas J. 1978. Alpha irradiation of the skin and the possibility of late effects. Health Physics 35(803–806). Shamanin V, zur Hausen H, Lavergne D, et al. 1996. Human papillomavirus infections in nonmelanoma skin cancers from renal transplant recipients and nonimmunosuppressed patients [see comments]. J Natl Cancer Inst 88(12):802–811. Sharfstein J, Sharfstein S. 1995. AMPAC campaign contributions to congress— a correction. N Engl J Med 332(21):1450. Sheil AG, Flavel S, Disney AP, et al. 1987. Cancer incidence in renal transplant patients treated with azathioprine or cyclosporine. Transplant Proc 19(1 Pt 3):2214–2216. Sheiner PA, Magliocca JF, Bodian CA, et al. 2000. Long-term medical complications in patients surviving > or = 5 years after liver transplant. Transplantation 69(5):781–789. Shore R, Albert R, Reed M. 1984. Skin cancer incidence among children irradiated for ringworm of the scalp. Radiat Res 100:192–204. Slaper H, Velders GJ, Daniel JS, et al. 1996. Estimates of ozone depletion and skin cancer incidence to examine the Vienna Convention achievements. Nature 384(6606):256–258. Smith AH, Goycolea M, Haque R, et al. 1998. Marked increase in bladder and lung cancer mortality in a region of Northern Chile due to arsenic in drinking water. Am J Epidemiol 147(7):660–669. Smith PG, Doll R. 1981. Mortality from cancer and all causes among British radiologists. Br J Radiol 54(639):187–194. Smith R, Healy E, Siddiqui S, et al. 1998. Melanocortin 1 receptor variants in an Irish population. J Invest Dermatol 111:119–122. Soll-Johanning H, Bach E, Olsen JH, et al. 1998. Cancer incidence in urban bus drivers and tramway employees: A retrospective cohort study. Occup Environ Med 55(9):594–598. Sorensen HT, Mellemkjaer L, et al. 2004. Skin cancers and non-hodgkin lymphoma among users of systemic glucocorticoids: a population-based cohort study. J Natl Cancer Inst 96(9):709–711. Soufir N, Moles JP, Vilmer C, et al. 1999. p16 UV mutations in human skin epithelial tumors. Oncogene 18:5477–5481. Speight EL, Dahl MG, Farr PM. 1994. Actinic keratosis induced by use of sunbed. BMJ 308(6925):415. Spencer JM, Amonette R. 1998. Tanning beds and skin cancer: Artificial sun + old sol = real risk. Clin Dermatol 16(4):487–501. Staples M, Marks R, Giles G. 1998. Trends in the incidence of non-melanocytic skin cancer treated in Australia 1985–1995: Are primary prevention programs starting to have an effect? Int J Cancer 78:144–148. Stellman SD, Garfinkel L. 1984. Cancer mortality among woodworkers. American Journal of Industrial Medicine 5(5):343–357. Stern RS. 1998. Photocarcinogenicity of drugs. Toxicol Lett 102–103: 389–392. Stern RS, Laird N. 1994. The carcinogenic risk of treatments for severe psoriasis. Photochemotherapy Follow-up Study. Cancer 73(11):2759– 2764. Stern RS, Thibodeau LA, Kleinerman RA, et al. 1979. Risk of cutaneous carcinoma in patients treated with oral methoxsalen photochemotherapy for psoriasis. N Engl J Med 300(15):809–813. Swerdlow AJ, Weinstock MA. 1998. Do tanning lamps cause melanoma? An epidemiologic assessment. J Am Acad Dermatol 38(1):89–98.

1249

Taguchi M, Watanabe S, Yahima K, Murakami Y, Sekiya T, Ikeda S. 1994. Aberrations of the tumor suppressor p53 gene and p53 protein in solar keratosis in human skin. J Invest Dermatol 103:500–503. Tangrea JA, Edwards BK, Taylor PR, et al. 1992. Long-term therapy with lowdose isotretinoin for prevention of basal cell carcinoma: A multicenter clinical trial. J Natl Cancer Inst 84:328–332. Thompson DE, Mabuchi K, Ron E, et al. 1994. Cancer incidence in atomic bomb survivors. Part II: Solid tumors, 1958–1987 [published erratum appears in Radiat Res 1994 Jul;139(1):129]. Radiat Res 137(2 Suppl): S17–S67. Thompson SC, Jolley D, Marks R. 1993. Reduction of solar keratoses by regular sunscreen use. N Engl J Med 329:1147–1151. Tsai S-M, Wang T-N, Ko Y-C. 1999. Mortality for certain diseases in areas with high levels of arsenic in drinking water. Arch Environ Health 54:186–193. Tseng W-P, Chu HM, How SW, et al. 1968. Prevalence of skin cancer in an endemic area of chronic arsenicism in Taiwan. J Natl Cancer Inst 40:453–463. Unden AB, Holmberg E, Lundh-Rozell B, et al. 1996. Mutations in the human homologue of Drosophila patched (PTCH) in basal cell carcinomas and the Gorlin syndrome: Different in vivo mechanisms of PTCH inactivation. Cancer Res 56:4562–4565. Valery PC, Neale R, et al. 2004. The effect of skin examination surveys on the incidence of basal cell carcinoma in a Queensland community sample: A ten-year longitudinal study. J Invest Dermatol Symp Proc 9:148–151. Valverde P, Healy F, Jackson I, Rees JL, Thody AJ. 1995. Variants of the melanocyte-stimulating hormone receptor gene are associated with red hair and fair skin in humans. Nature Genet 11:328–330. van Dam RM, Huang Z, Rimm EB, et al. 1999. Risk factors for basal cell carcinoma of the skin in men: Results from the health professionals followup study. Am J Epidemiol 150(5):459–468. van Dam RM, Huang Z, et al. 2000. Diet and basal cell carcinoma of the skin in a prospective cohort of men. Am J Clin Nutr 71(1):135–141. Veness MJ, Quinn DI, Ong CS, et al. 1999. Aggressive cutaneous malignancies following cardiothoracic transplantation: The Australian experience. Cancer 85(8):1758–1764. Vitaliano PP, Urbach F. 1980. The relative importance of risk factors in nonmelanoma carcinoma. Arch Dermatol 116(4):454–456. Vitasa BC, Taylor HR, et al. 1990. Association of nonmelanoma skin cancer and actinic keratosis with cumulative solar ultraviolet exposure in Maryland watermen. Cancer 65(12):2811–2817. Vlajinac H, Adanja BJ, Lazar ZF, et al. 2000. Risk factors for basal cell carcinoma. Acta Oncologica 39:611–616. Vogel U, Hedayati M, Dybdahl M, Grossman L, Nexo BA. 2001. Polymorphisms of the DNA repair gene XPD: Correlations with risk of basal cell carcinoma revisited. Carcinogenesis 22:899–904. Wang C, Brodland DG, Su D. 1995. Skin cancers associated with acquired immunodeficiency syndrome. Mayo Clin Proc 70:766–772. Wassberg C, Thorn M, Johansson AM, et al. 2001. Increasing incidence rates of squamous cell carcinoma of the skin in Sweden. Acta DermatoVenereologica. 81(4):268–272. Wechsler J, Zanetti R, Schrameck C, et al. 2001. Reproducibility of histopathologic diagnosis and classification of non-melanocytic skin cancer: A panel exercise in the framework of the multicenter southern European study HELIOS. Tumori 87(2):95–100. Wei Q, Matanoski GM, Farmer ER, et al. 1993. DNA repair and aging in basal cell carcinoma: A molecular epidemiology study. Proc Natl Acad Sci U S A 90:1614–1618. Wei Q, Matanoski GM, Farmer ER, et al. 1994. DNA repair related to multiple skin cancers and drug use. Cancer Res 54(2):437–440. Weihrauch M, Bader M, Lehnert G, Wittekind C, Tannapfel A, Wrbitzky R. 2002. Carcinogen-specific mutation pattern in the p53 tumour suprressor gene in UV radiation-induced basal cell carcinomas. Int Arch Occ Env Health 75:272–276. Weinstock MA, Bogaars HA, et al. 1991. Nonmelanoma skin cancer mortality. A population-based study. Arch Dermatol 127(8):1194–1197. Weinstock M, Colditz GA, Willett WC, Stampfer MJ, Rosner B, Speizer FE. 1991. Recall (report) bias and reliability in the retrospective assessment of melanoma risk. Am J Epidemiol 133:240–245. Weinstock MA. 1993. Nonmelanoma skin cancer mortality in the United States, 1969 through 1988. Arch Dermatol 129:1286–1290. Weinstock MA. 1994. Epidemiology of nonmelanoma skin cancer: Clinical issues, definitions, and classification. J Invest Dermatol 102(6):4S–5S. Weinstock MA. 1997. Death from skin cancer among the elderly: Epidemiologic patterns. Arch Dermatol 133:1207–1209. Weinstock MA, Bogaars HA, Ashley M, et al. 1992. Inaccuracies in certification of nonmelanoma skin cancer deaths. Am J Public Health 82:278–281. Wolter M, Reifenberger J, Sommer C, et al. 1997. Mutations in the human homolgue of the Drosophila segment polarity gene patched (PTCH) in

1250

PART IV: CANCER BY TISSUE OF ORIGIN

sporadic basal cell carcinomas of the skin and primitive neuroectodermal tumors of the central nervous system. Cancer Res 57:2581–2585. Wu MM, Kuo TL, Hwang YH, et al. 1989. Dose-response relation between arsenic concentration in well water and mortality from cancers and vascular diseases. Am J Epidemiol 130(6):1123–1132. Wu X, Zhao H, Amos CI, et al. 2002. p53 genotypes and haplotypes associated with lung cancer susceptibility and ethnicity. J Natl Cancer Inst 94:681–690. Xie JM M, Luoh SM, Ryan A, et al. 1998. Activating smoothened mutations in sporadic basal-cell carcinoma. Nature 391:90–92. Xu X, Thornwall M, Lundin LG, Chhajlani V. 1996. Val92Met variant of the melanocyte stimulating hormone receptor gene. Nature Genet 14: 384. Yakubu A, Mabogunje OA. 1995. Skin cancer in Zaria, Nigeria. Trop Doct 25(Suppl 1):63–67. Yarosh D, Klein J, O’Connor A, et al. 2001. Effect of topically applied T4 endonuclease V in liposomes on skin cancer in xeroderma pigmentosum: A randomised study. Xeroderma Pigmentosum Study Group. Lancet 357(9260):926–929. Yengi L, Inskip A, Gilford J, et al. 1996. Polymorphism at the glutathione Stransferase locus GSTM3: Interactions with cytochrome P450 and glutathione S-transferase genotypes as risk factors for multiple cutaneous basal cell carcinoma. Cancer Res 56(9):1974–1977.

Yoshinaga, S, Hauptmann M, et al. 2005. Nonmelanoma skin cancer in relation to ionizing radiation exposure among U.S. radiologic technologists. Int J Cancer 115(5):828–834. Yu RC, Hsu KH, et al. 2000. Arsenic methylation capacity and skin cancer. Cancer Epidemiol Biomarkers Prev 9(11):1259–1262. Zaldivar R. 1974. Arsenic contamination of drinking water and foodstuffs causing endemic chronic poisoning. Beitrage zur Pathologie 151(4): 384–400. Zanetti R, Rosso S, Martinez C, et al. 1996. The multicentre south European study “Helios” I: Skin characteristics and sunburns in basal cell and squamous cell carcinomas of the skin. Br J Cancer 73:1440–1446. Zhang H, Ping XL, Lee PK, et al. 2001. Role of PTCH and p53 genes in earlyonset basal cell carcinoma. Am J Pathol 158:381–385. Zhang W, Remenyik E, Zelterman D, Brash DE, Wikonkal NM. 2001. Escaping the stem cell compartment: Sustained UVB exposure allows p53mutant keratinocytes to colonize adjacent epidermal proliferating units without incurring additional mutations. Proc Natl Acad Sci 98: 13948–13953. Ziegler A, Leffell DF, Kunala S, et al. 1993. Mutation hotspots due to sunlight in the p53 gene of nonmelanoma skin cancers. Proc Natl Acad Sci 90:4216–4220. Ziegler A, Jonason A, Lefell DJ, et al. 1994. Sunburn and p53 in the onset of skin cancer. Nature 327:733–736.

65

Cancers in Children JULIE A. ROSS AND LOGAN G. SPECTOR

N

early 12,400 children and adolescents under the age of 20 years will be newly diagnosed with cancer each year in the United States (Ries et al., 1990). This means that a newborn has an approximately 1 in 315 chance of developing cancer in the first two decades of life (Ries et al., 2002). Overall, however, childhood cancer accounts for only a small proportion of the total cancer burden in the United States. Figure 65–1 presents a comparison of the estimated number of malignancies diagnosed each year by site and the estimated number of total cancers diagnosed in children less than 15 years of age (Ries et al., 2003). The total number of incident cases among children is less than the number of incident diagnoses of cancer for all but the rarest of individual adult sites. Nevertheless, cancer is a leading cause of disease-related mortality in US children; approximately 2500 children will die from their malignancy each year (Ries et al., 1999). In the year 2000, malignant neoplasms comprised about 8% of deaths among 1–4 year olds, 15% among 5–9 year olds, 12% among 10–14 year olds, and 6% among 15–19 year olds, with ranking as the third, second, second, and fourth leading causes of death, respectively (Minino et al., 2002). Accidents are the leading cause of death in children and comprise between 40%–50% of mortality at these ages. The rate of death from cancers in 2000 was 2.8, 2.5, 2.6, and 3.7 per 100,000 at ages 1–4, 5–9, 10–14 years, and 15–19 years, respectively (Minino et al., 2002). Overall, approximately 106,700 personyears of life are lost each year for children who die of cancer (Ries et al., 2003).

OVERALL INCIDENCE During the period 1995–1999, the incidence rate of cancer for all sites combined was 20.5, 11.1, 12.9, and 20.1 per 100,000 for 0–4, 5–9, 10–14, and 15–19 year olds, respectively (Ries et al., 2002). There is substantial variation in incidence within these age groups. Cancer is more common in males than in females, with the respective male-tofemale ratios being 1.14, 1.28, 1.07, and 1.03. Cancer is also more common in whites than in blacks; corresponding white-to-black ratios are 1.24, 1.28, 1.28, and 1.55 (Ries et al., 2002). The incidence rate of all childhood cancer sites combined has risen slightly, but significantly, in past decades. During the period 1975– 1999, the average annual percentage change (AAPC) in the incidence rate was 0.9 (p < 0.05) for 0–4 year olds, 0.4 (p > 0.05) for 5–9 year olds, 1.0 (p < 0.05) for 10–14 year olds, and 0.6 (p < 0.05) for 15–19 year olds. Incidence has been steady, however, in the most recent period, with corresponding AAPCs for 1987–1999 of 0.4 (p > 0.05), -0.2 (p > 0.05), 1.1 (p < 0.05), and -0.3 (p < 0.05) for each age group, respectively (Ries et al., 2002).

MORTALITY AND MORBIDITY Improved treatments for childhood cancer over the past 30 years have dramatically increased long-term survival. Five-year relative survival rates of all childhood cancers combined rose significantly (p < 0.05) from 56% during 1974–1976 to 77% during 1992–1998. Each indi-

vidual cancer type experienced a significant (p < 0.05) improvement in survival, though the extent varied (Jemal et al., 2003). As a result of these advances in treatment, there is a large and growing number of childhood cancer survivors. In the United States in 2000, there were estimated to be nearly 195,000 people who had survived up to 25 years past a diagnosis of cancer between the ages of 0 and 19 years (Ries et al., 2003). Compared with the general population, childhood cancer survivors have an increased risk of second malignancies, heart disease, obesity, neurocognitive complications, and other disorders. In addition, patients and their families can experience substantial psychological trauma during and after treatment (Bhatia and Sklar, 2002; Bhatia, 2002). Care for these survivors requires education, ongoing screening and surveillance, interventions, and support.

OVERVIEW: ETIOLOGY OF CHILDHOOD CANCER BY SITE The distribution of cancer diagnoses is quite different for children than for adults. Due to this dissimilar distribution of cancer and to the recognition that histology is often more relevant than site, childhood cancer has its own classification system called the International Classification of Childhood Cancer (ICCC) (Kramarova and Stiller, 1996). The most common malignancies diagnosed in children under the age of 15 years (Table 65–1) in order of decreasing incidence are the leukemias (acute lymphoblastic leukemia, acute myeloid leukemia), brain/central nervous system tumors (astrocytoma, primitive neuroectodermal tumors, gliomas, and ependymomas), lymphomas (nonHodgkin lymphoma, Hodgkin disease) sympathetic nervous system tumors (neuroblastoma), soft tissue sarcomas (rhabdomyosarcoma), renal tumors (Wilms tumor, renal carcinoma), bone tumors (osteosarcoma, Ewing sarcoma), malignant germ cell tumors, retinoblastoma, and hepatic tumors (hepatoblastoma, hepatocellular carcinoma). Other malignancies, including thyroid cancer, melanoma, adrenocortical tumors, and nasopharyngeal cancers, comprise the remainder. Trends in incidence, survival, and mortality vary not only by the type of cancer but also by gender, race, age at onset, clinical characteristics, and molecular abnormalities. This variability strongly suggests separate etiologies for some types of childhood cancer. Moreover, the variability within some childhood cancers has only been recognized with advances in pathology and molecular biology. For example, there is a general trend in epidemiological studies of childhood leukemia to consider more genetically and immunophenotypically defined subgroups such as children with Down syndrome or infants with MLL generearranged leukemia. This review will focus on the 10 most commonly studied malignancies in children diagnosed under the age of 15 years as shown in Table 65–1. Most etiologic research of childhood cancer consists of ecologic or case-control studies. The number of such studies for each cancer type is roughly correlated with the incidence of the disease. A small portion of childhood cancer can be attributed to cancerpredisposing inherited conditions (shown in Table 65–2). Although some genetic syndromes, such as familial retinoblastoma, may be associated with up to 40% of cases, it is estimated that, overall, only about 5% of childhood cancers are associated with inherited genetic alterations (Volgelstein and Kinzler, 1998).

1251

1252

PART IV: CANCER BY TISSUE OF ORIGIN Breast (females) Prostate Lung and Bronchus Colon and Rectum Non-Hodgkin lymphoma Urinary bladder Melanoma of skin Corpus and uterus, NOS

Est. Cancer Cases

Leukemia Oral cavity and Pharynx

Actual Mortality

Pancreas Ovary Stomach Thyroid Brain and Other nervous Liver and Intrahep Multiple myeloma Kidney and Renal pelvis Cervix uteri Esophagus Larynx Childhood (0-14 years) Hodgkin lymphoma Testis

0

20,000

40,000

60,000

80,000

100,000

120,000

140,000

160,000

180,000

200,000

Number of cases

Figure 65–1. Estimated US cancer cases with actual mortality in 2000, by site.

ACUTE LYMPHOBLASTIC LEUKEMIA AND ACUTE MYELOID LEUKEMIA Acute lymphoblastic leukemia (ALL) and acute myeloid leukemia (AML) are malignancies of lymphoid and myeloid progenitor cells, respectively. They comprise the vast majority of childhood leukemias, which represent about a third of childhood cancer diagnoses in the United States (Ries et al., 1999). Internationally, incidence of ALL is roughly correlated with the socioeconomic status of nations. As can be seen in Figure 65–2A, incidence of ALL is highest in Costa Rica and among US Hispanics, with more than 45 cases per million. Slightly lower rates prevail in the rest of North America, Europe, and Oceania, while rates in Asia, Central, and South America are intermediate. The rate of ALL in Africa is nearly 50-fold lower than that

in countries with the highest rates. The rate of AML, by contrast, is between 5 and 10 cases per million with the exceptions of Africa and the Maori of New Zealand, among whom rates are lower and higher, respectively (Parkin et al., 1998). The incidence of ALL and AML differs widely by age, sex, and race. In the United States, ALL shows a marked peak in incidence between the ages of 2 and 5 years, with a rate of about 76 cases per million at the maximum (Gurney et al., 1995). Immunophenotyping of ALL reveals that CD10+ B-cell lineage ALL alone, rather than T-cell or null-cell lineage ALL, forms the majority of the childhood peak, and is for this reason called common ALL or cALL (Greaves et al., 1993). Null-cell ALL occurs mainly in infancy, whereas T-cell ALL incidence climbs steadily from childhood to adolescence (Greaves et al., 1985). Males out-number females by a ratio of 1.2 and whites out-number blacks by a ratio of 2.0 among

Table 65–1. US Incidence Rates (1975–2000) and 5-Year Relative Survival Rates (1985–1999) of Most Common Malignancies Diagnosed in Children Age-Adjusted Incidence Rate per Million Malignancy Leukemias (acute lymphoblastic leukemia, acute myeloid leukemia) Brain/central nervous system tumors (astrocytoma, primitive neuroectodermal tumors, gliomas, ependymomas) Lymphomas (non-Hodgkin lymphoma, Hodgkin lymphoma) Sympathetic nervous system (neuroblastoma) Soft tissue sarcomas (rhabdomyosarcoma) Renal tumors (Wilms tumor, renal carcinoma) Bone tumors (osteosarcoma and Ewing sarcoma) Malignant germ cell tumors Retinoblastoma Hepatic tumors (hepatoblastoma, hepatocellular carcinoma) Thyroid cancer Malignant melanoma Nasopharyngeal cancer Adrenocortical carcinoma Source: Adapted from Ries, et al. (2003).

Five-Year Relative Survival Rates

Age <15 years

Age <20 years

Age <15 years

Age <20 years

42.0 29.0

37.2 26.5

81.8 66.4

79.1 68.1

14.8 10.5 10.0 8.6 6.5 4.7 4.0 1.9 1.8 1.5 0.3 0.2

24.0 8.1 11.4 6.8 8.5 10.5 3.0 1.7 5.2 4.6 0.5 0.2

83.4 66.0 73.1 90.4 67.5 86.7 94.7 55.8 97.3 88.0 N/A N/A

84.9 65.4 70.8 89.8 65.3 89.6 94.8 49.9 98.7 92.1

Cancers in Children Table 65–2. Childhood Cancers Associated with Inherited Syndromes Inherited Condition Li-Fraumeni syndrome (inherited P53 defect) Down syndrome (trisomy 21) Familial adenomatous polyposis Beckwith–Wiedemann syndrome Neurofibromatosis Schwachman syndrome Bloom syndrome Ataxia telangiectasia Wiskott–Aldrich syndrome X-linked lymphoproliferative disease Familial retinoblastoma/13q deletion syndrome Langerhans cell histiocytosis Klinefelter syndrome Familial monosomy 7 Kostmann granulocytopenia Fanconi anemia Tuberous sclerosis Nevoid basal cell syndrome Turcot syndrome Costello syndrome Aniridia Wilms tumor, aniridia, genitourinary abnormalities, mental retardation (WAGR) syndrome Perlman syndrome Denys–Drash syndrome Simpson-Golabi-Behmel syndrome Gardner syndrome Hemihypertrophy Rothmund–Thomson syndrome

Associated Childhood Cancers CNS, OS, STS ALL, AML HB WT, HB, STS ALL, AML, CNS, STS ALL, AML ALL, AML ALL, NHL NHL NHL RB, OS ALL ALL AML AML AML CNS CNS CNS STS WT WT WT WT WT HB HB, WT OS

Source: Adapted from Ries et al. (1999), Sandlund et al. (1996), Ruymann, et al. (1988), Yang et al. (1995), Gripp et al. (2002). CNS, central nervous system; OS, osteosarcoma; STS, soft tissue sarcoma; ALL, acute lymphoblastic leukemia; AML, acute myeloid leukemia; HB, hepatoblastoma; WT, Wilms Tumor; NHL, non-Hodgkin lymphoma; RB, retinoblastoma.

diagnoses of ALL at ages less than 15 years (Ries et al., 1999). The incidence of AML, by contrast, starts in infancy at its maximum of 11 cases per million and declines thereafter (Gurney et al., 1995). Males have slightly higher rates than females (ratio = 1.1), while the number of white and black cases is similar among diagnoses of AML at ages less than 15 years. Chronic myeloid leukemia (CML) is much less common than ALL and AML, the rate being 1.0 case per million children less than 15 years of age. Notably, the rate of CML among males is more than fourfold higher than that of females at ages less than 5 years (Ries et al., 1999). Genetic analyses have illuminated much of the natural history of childhood leukemia. A large number of chromosomal rearrangements have been described in ALL and AML, though a few types predominate. About 85% of ALL (Greaves, 1999) and an estimated 60% of AML (Cimino et al., 1993) among children less than 1 year of age displays the MLL-11q23 gene rearrangement. Among children aged 1–15 years with ALL, TEL-AML1 and hyperdiploid rearrangements each account for 25% of cases, the MLL-11q23 gene rearrangement is present in only 5%, and a diverse array of rearrangements account for the remainder (Greaves, 1999). Many chromosomal anomalies have been described in AML at ages greater than 1 year but none predominates (Perkins et al., 1997). Though 5-year survival of childhood leukemia now exceeds 80%, the prognosis for children with specific cytogenetic abnormalities varies widely (Greaves, 2002). The fact that twins concordant for both ALL and AML share the same nonconstitutive rearrangements (Ford et al., 1993; Richkind et al., 1998) and that these rearrangements have been detected in neonatal bloodspots (Guthrie cards) (Gale et al., 1997; Wiemels et al., 1999; Fasching et al., 2000; Yagi et al., 2000; Wiemels et al., 2002) strongly suggests that most acute leukemia in childhood is initiated in utero. Among infants, the in-utero MLL-11q23 rearrangement is probably sufficient to cause leukemia, since infant identical twins are nearly 100% con-

1253

cordant for acute leukemia in anecdotal reports (Greaves, 2002). On the other hand, concordance of acute leukemia is a comparatively low 5% among older children (Greaves, 1999), suggesting that a postnatal mutation is also needed for disease to develop at ages 1–15 years (Wiemels et al., 1999; Fasching et al., 2000). Accordingly, the TELAML1 rearrangement, at least, has been found in the Guthrie cards of children who did not develop leukemia (Wiemels et al., 1999; Mori et al., 2002). Few risk factors for the childhood leukemias are well established, though many have been examined. Down syndrome (Front et al., 1987; Robison et al., 1987), and inherited cancer-predisposing conditions such as ataxia telangiectasia (Mullvihill, 1975; Bloomfield and Brunning, 1976; Bader et al., 1978; German et al., 1979; Woods et al., 1981; Linet, 1985; Hecht et al., 1990) substantially increase the risk of both ALL and AML but account for only a small fraction of cases (Robison and Neglia, 1984; Narod et al., 1991; Watson et al., 1993). In utero diagnostic radiation is also an accepted risk factor for both types of leukemia (Doll and Wakeford, 1997), but its importance has declined along with dosage and frequency of fetal exposure to X-rays (Harvey et al., 1985; Mole, 1990; Rodvall et al., 1990; Shu et al., 2002). Investigations of ionizing radiation from other sources, such as fallout (Cartwright et al., 1988; Gibson et al., 1988; Stevens et al., 1990; Ivanov et al., 1993; Parkin, 1993; Auvinen et al., 1994; Hjalmars et al., 1994; Petridou et al., 1994, 1996; Michaelis et al., 1997) or radon (Alexander et al., 1990; Henshaw et al., 1990; Lubin et al., 1998; Kaletsch et al., 1999; Steinbuch et al., 1999; Axelson et al., 2002; Investigators, 2002), from postnatal X-rays (Linos et al., 1978; Boice, 1986; Infante-Rivard et al., 2000), and occupational (Gardner, 1991; McLaughlin et al., 1993; Parker et al., 1993) or diagnostic (Shu et al., 1994) radiation to the father have sometimes indicated increased risk of leukemia but are, on whole, equivocal. Non-ionizing radiation has not appeared to be associated with leukemia in studies that incorporate measurements of electromagnetic fields (Linet et al., 1997; Anonymous, 1999; McBride et al., 1999), though meta-analysis suggests that the highest level of exposure, to which very few children are exposed, may raise risk of leukemia (Ahlbom et al., 2000; Greenland et al., 2000). Other familiar sources of carcinogens have been investigated in regards to both leukemias with little suggestion of causation in most cases. Maternal consumption of alcohol while pregnant does not appear to increase the frequency of ALL in offspring, but may increase the risk of AML (Severson et al., 1993; van Duijn et al., 1994; Shu et al., 1996). Studies of parental tobacco use and acute leukemia have reported no association with maternal smoking while pregnant with fair consistency, whereas there have been inconsistent reports of an association with paternal, preconceptional smoking (Stjernfeldt et al., 1992; Severson et al., 1993; Sorahan et al., 1995; Shu et al., 1996; Sorahan et al., 1997; Brondum et al., 1999; Sorahan et al., 2001; Pang et al., 2003). Occupational or other exposure to chemicals has been associated inconsistently with ALL (Shaw et al., 1984; Lowengart et al., 1987; Buckley et al., 1994; Shu et al., 1999; Schuz et al., 2000; Freedman et al., 2001); exposure to pesticides and benzene has been more consistently associated with AML and accords with adult data (Shu et al., 1988; Buckley et al., 1989). Emerging evidence indicates that lowactivity variants of polymorphisms in detoxification metabolism and DNA repair mechanism genes raise the risk of childhood leukemia, which supports the idea of chemical carcinogenesis (Davies et al., 2000; Krajinovic et al., 2002) and in small studies these polymorphisms do appear to modify the risk of some chemical exposures (Infante-Rivard et al., 1999; Infante-Rivard et al., 2000; InfanteRivard et al., 2002). Factors related to birth have also been of interest in leukemia etiology. Maternal age greater than 35 years at the time of a child’s birth (Hemminki et al., 1999; Dockerty et al., 2001) and maternal history of fetal loss also appears (Kaye et al., 1991; Yeazel et al., 1995; Ross et al., 1997; Shu et al., 2002) to be directly associated with ALL, even after controlling for Down syndrome and parity. High birth weight (variously defined, but usually >4000 grams) also is associated with an increased risk of acute leukemia in most large studies (Cnattingius

ALL Africa Uganda-Kampala

<1 case/mill

<1 case/mill

Nigeria

<1 case/mill

<1 case/mill

AmericaCentral and Uraguay South Brazil-Goiania

AML

CNS

Costa Rica Asia

India-Bombay China Israel-Jews Israel-non Jews Japan-Osaka

Oceania

NZ-Maori NZ-non Maori Australia

Europe

UK-Eng/Wales Sweden Spain Italy Germany

North America LA-Hispanic USA SEER black USA SEER white Canada 0

5

10

15

20

25

30

35

40

45

50 0

5

10

15

25 0

20

5

10 15 20 25 30 35 40 45 50

Age Standardized Rates

A HD

NHL

WT

Africa Uganda-Kampala Nigeria AmericaCentral and Uraguay South Brazil-Goiania Costa Rica Asia

India-Bombay <1 case/mill

<1 case/mill

China Israel-Jews Israel-non Jews Oceania Japan-Osaka NZ-Maori NZ-non Maori Europe

Australia UK-Eng/Wales Sweden Spain Italy

North Ameria

Germany LA-Hispanic

USA SEER black USA SEER white Canada 0

2

4

6

8

10

12 0

2

4

6

8

10

12 0

2

4

6

8

10

12

Age Standardized Rates

B Figure 65–2. (A) International incidence of childhood acute lymphoblastic leukemia, acute myeloid leukemia, and central nervous system tumors 1980–1990. (Source: Adapted from Parkin et al., 1998.) (B) International

incidence of childhood Hodgkin disease, non-Hodgkin lymphoma, and Wilms tumor 1980–1990. (Source: Adapted from Parkin et al., 1998.)

Cancers in Children et al., 1995; Ross et al., 1997; Westergaard et al., 1997; Yeazel et al., 1997; Reynolds et al., 2002; Shu et al., 2002); these observations are corroborated by evidence that ALL is associated with endogenous correlates of birth weight (Lei et al., 2000; Petridou et al., 2000). Lastly, neonatal Vitamin K prophylaxis was suggested as a cause of leukemia, but the available evidence does not support this contention (Ross and Davies, 2000). A number of lines of evidence support the notion that childhood ALL is related to some correlate of socioeconomic status, at least at the level of nations since, as we mentioned above, ALL occurs much more frequently in industrialized countries than in developing ones, especially at less than 5 years of age (Parkin et al., 1998). Also, this childhood peak in incidence seems to have emerged over time in populations experiencing economic growth (Fraumeni and Miller, 1967; Hrusak et al., 2002). The fact that cALL comprises most of the childhood peak (Greaves et al., 1993) suggests that higher rates of ALL in industrialized countries are not due to improved reporting, and that it is this form of ALL, specifically, that is associated with some factor related to economic growth. Though there are many such factors, infection has been the focus of research because ALL is a malignancy of immune cell progenitors. Two main theories endeavor to explain the international variation in childhood leukemia incidence in terms of infection. One theory holds that childhood leukemia (and not a particular subtype) is a rare response to infection of immune cells by a particular agent, most likely a virus (Kinlen, 1988). Alternatively, cALL specifically may result as lymphoblasts proliferate in response to infection in general and thereby acquire mutations necessary to produce leukemia (Greaves, 1988). In both theories the likelihood of leukemia is posited to increase with later age of exposure to infection(s), as in the paradigm of paralytic poliomyelitis (Baccate, 1983). Evidence of the role of infection in leukemogenesis supports both theories so far, since it is difficult to distinguish between the two theories absent the discovery of a leukemogenic microbe. Leukemia cases show weak but significant spatio-temporal clustering (Alexander et al., 1998) and seasonality (Westerbeek et al., 1998; Ross et al., 1999) both of which are consistent with a role for infections. Other ecologic studies have consistently found raised rates of leukemia following an influx of newcomers to previously isolated areas (termed population mixing), which could facilitate the transmission of infections (Kinlen, 1995). Subsequent studies that used more quantitative measures of population mixing have, in general, found an increased risk of leukemia with greater population growth and diversity of immigrants, but have disagreed about whether this increase is exclusive to rural or to urban areas (Langford, 1991; Stiller et al., 1996; Dickinson et al., 1999; Koushik et al., 2001; Boutou et al., 2002; Dickinson and Parker, 2002; Parslow et al., 2002). The search for a particular leukemogenic microbe has largely been fruitless. Serologic surveys have been mainly cross-sectional and thus can not determine that infection preceded leukemia (Gahrton et al., 1971; Heegaard et al., 1999; MacKenzie et al., 1999; Groves et al., 2001; MacKenzie et al., 2001; Salonen et al., 2002). Meanwhile, history of infection as established by questionnaire or record abstraction is subject to recall bias and misclassification of asymptomatic individuals (McKinney et al., 1987; Shu et al., 1988; Buckley et al., 1994; Dockerty et al., 1999; McKinney et al., 1999; Schuz et al., 1999; Neglia et al., 2000; Chan et al., 2002; Naumburg et al., 2002; Perrillat et al., 2002). Proxy measures of infection may be both more accurately recalled and less misclassified. Examples are time in attendance at day care and birth order, both of which have shown an inverse association with leukemia (Zack et al., 1991; Westergaard et al., 1997; Infante-Rivard et al., 2000; Dockerty et al., 2001; Chan et al., 2002; Ma et al., 2002b; Perrillat et al., 2002; Shu et al., 2002; Vineis et al., 2003), though not entirely consistently (Neglia et al., 2000; Rosenbaum et al., 2000). Prolonged breast feeding is less common in industrialized nations and has in many studies been inversely associated with leukemia (Parker, 2001; Lancashire and Sorahan, 2003). A benefit of breast feeding, if real, could derive either from its nutritional and immunologic benefits or its provision of early exposure to common viruses and bacteria (Dworsky et al., 1983).

1255

BRAIN/CENTRAL NERVOUS SYSTEM CANCERS (ASTROCYTOMAS, PRIMITIVE NEUROECTODERMAL TUMORS, GLIOMAS, EPENDYMOMAS) Central nervous system (CNS) cancers are a diverse collection of malignancies that together make up about 16% of childhood cancer. CNS cancers occur mainly in the brain, but can also appear extracranially. Because of the diverse histological origins of CNS cancers, there have been proposed alternate classification systems (Kleihues et al., 1993) that do not entirely coincide with the ICCC (Kramarova and Stiller, 1996). Nevertheless, CNS cancers are usually grouped into the following categories for the purpose of study: astrocytomas (49.6% of CNS cancers in the US), primitive neuroectodermal tumors (PNET, 22.9%), other gliomas (15.4%), and ependymomas (9.3%) (Ries et al., 1999). Rates of CNS cancers in the industrialized world range between about 20 and 40 cases per million, whereas rates in the developing world are generally below 20 cases per million, especially in Africa (Fig. 65–2A); a similar gradient in incidence is seen for each of the CNS cancer subtypes (Parkin et al., 1998). Incidence of CNS malignancy in the United States is highest in early childhood, at about 35 cases per million, and declines to less than 25 cases per million by 15 years of age (Ries et al., 1999) Non-malignant tumors are not covered here, but it should be noted that they can be debilitating. Including benign CNS tumors would raise the above incidence rates by about 25% (Gurney et al., 1999). There is a preponderance of males among PNET and ependymomas, but not among other CNS cancers. The rate of CNS cancers of all subtypes is higher among whites than blacks, with this difference most evident among males (Ries et al., 1999). The incidence of CNS cancers is undoubtedly higher recently than it was in earlier periods, but it appears to have jumped suddenly in the mid 1980s. This suggests that the increase is an artifact of improved diagnostic technology, namely magnetic resonance imaging, rather than a true secular trend (Smith et al., 1998). Survival of brain tumors varies by subtype but is low; except for astrocytomas, 5-year survival of CNS cancers that occur below 20 years of age is less than 60% (Ries et al., 1999). Relatively few studies of CNS tumor cytogenetics have been conducted that could improve prognosis and treatment, in part because of the difficulty of obtaining adequately sized biopsy samples from such a delicate region. However, it is known that PNETs often have a high frequency of abnormalities, including trisomy of chromosome 7, monosomy of chromosomes 8 and 22, and a doubling of the long arm of chromosome 17 (i.e., i(17q), compared with other CNS tumors and accordingly are more aggressive (Bhattacharjee et al., 1997). Wide-ranging investigations of potential risk factors for CNS cancers have identified only one with great consistency besides the congenital conditions listed in Table 65–2. Therapeutic radiation to the head raises the risk of CNS cancers several fold, but is of mostly historical interest, and today accounts for only a tiny proportion of cases (Shore et al., 1976; Ron et al., 1988). Experimental evidence has suggested other candidate risk factors such as N-nitroso compounds and polyoma viruses, though a role for them in CNS cancer etiology has not been fully supported in epidemiologic studies (Gurney et al., 2001). When fed to pregnant animals, nitrosamides and nitrosoureas are potent inducers of brain tumors in offspring (Ivankovic, 1979). Cured meats are a major source of nitrite precursors, which are converted into N-nitroso compounds in vivo (Leaf et al., 1989; Lijinsky, 1999). Though a majority of case-control investigations of cured meat consumption by pregnant mothers have indicated higher risk of CNS cancers, internal inconsistency in these studies (e.g., finding higher risk for consumption of some cured meats but not others) makes an association less certain (Preston-Martin et al., 1982; Kuijten et al., 1990; Bunin et al., 1993, 1994; Cordier et al., 1994; McCredie et al., 1994; Sarasua and Savitz, 1994). Also, cured meat consumption has not so far been shown to increase the level of DNA-damaging adducts in vivo (Gurney et al., 2002). Nitrosable drugs could also be a source of nitrosamides but have shown inconsistent associations with CNS tumors (Olshan and Faustman, 1989; Carozza et al., 1995;

1256

PART IV: CANCER BY TISSUE OF ORIGIN

McKean-Cowdin et al., 2003). Interestingly, Vitamins E and C inhibit the activity of N-nitroso compounds (Ivankovic et al., 1975) and consumption of vitamins in fruits, vegetables, and supplements by pregnant mothers has been inversely associated with CNS cancers in some studies (Preston-Martin et al., 1982; Bunin et al., 1993; Cordier et al., 1994; McCredie et al., 1994; Preston-Martin et al., 1998a, 1998b). By contrast with nitrosoamides and nitrosoureas, the other class of Nnitroso compounds, nitrosoamines, does not induce CNS cancers in experimental animals (Preussman and Stewart, 1984). Products that contain nitrosamines, such as beer, cigarettes, etc., are not associated with CNS cancer in humans (Howe et al., 1989; Gold et al., 1993; Bunin et al., 1994; Filippini et al., 1994; McCredie et al., 1994; Norman et al., 1996a, 1996b; Hu et al., 2000; Filippini et al., 2002). Laboratory evidence supports the plausibility of a role for polyoma viruses including SV40, BK, and JC virus in the etiology of CNS cancers. Components of polyoma viruses are mutagenic and bind to proteins that regulate the growth and death of cells (Barbanti-Brodano et al., 1998). Moreover, polyoma viruses induce brain tumors in test animals (Kirchstein and Gerber, 1962). Polyoma virus DNA has been detected in human brain tumors, but it is unclear whether this is a cause or result of cancer (Gurney et al., 2001). Follow-up studies of children immunized with SV40-contaminated polio vaccine, which is the only epidemiological evidence on the subject, has not confirmed an association of CNS cancer with polyoma viruses (Olin and Gesecke, 1998; Strickler and Goedert, 1998a; Strickler et al., 1998b; Fisher et al., 1999; Carroll-Pankhurst et al., 2001; Engels et al., 2003). However, these studies were based on possibly incorrect assumptions about exposure to SV40 (Vilchez et al., 2003) and, moreover, as ecologic studies they cannot definitively rule out an association. Studies of pesticides (Gold et al., 1979; Howe et al., 1989), epilepsy (Gold et al., 1979; Kuijten et al., 1993; McCredie et al., 1994; Gurney et al., 1997), brain injury (McCredie et al., 1999), electromagnetic fields (Gurney et al., 1996; Preston-Martin et al., 1996a, 1996b; Anonymous, 2000) and other potential risk factors (Gurney et al., 1997; Holly et al., 2002) indicate no association or have methodological problems that cast doubt on any observed associations. Many associations of CNS tumors have been reported with parental occupations, but these have seldom been replicated (Olshan et al., 1986; Nasca et al., 1988; Wilkins and Koutras, 1988; Wilkins and Sinks, 1990; Cordier et al., 1997; McKean-Cowdin et al., 1998; Cordier et al., 2001). Recently, however, it has appeared that living on a farm or having a parent that does farm work raises the risk of CNS cancer (Bunin et al., 1994; Kristensen et al., 1996; Holly et al., 1998; Efird et al., 2003).

LYMPHOMAS (HODGKIN DISEASE, NON-HODGKIN LYMPHOMA) Lymphomas represent the third most common type of malignancy in children in the United States. About 1700 children under the age of 20 are diagnosed with lymphoma each year in the United States, including approximately 900 cases of Hodgkin disease and 800 cases of non-Hodgkin lymphoma (NHL); miscellaneous lymphoreticular neoplasms make up another 80 or so cases (Ries et al., 1999). As there are distinct clinical presentations of Hodgkin disease and NHL, they will be discussed separately with respect to etiology.

Hodgkin Disease Hodgkin disease (HD), a malignancy of the B-lymphocyte lineage, is characterized by the presence of Reed-Sternburg cells (Jarrett and Onions, 1992). As with adults, HD in children is often classified into four histological subtypes including lymphocytic predominance, mixed cellularity, lymphocytic depletion, and nodular sclerosis. Nodular sclerosis is by far the most common, with approximately 70% of cases, followed by mixed cellularity (16%), lymphocytic predominance (7%), and cases not otherwise specified (6%) (Ries et al., 1999). Lymphocytic depletion type is extremely uncommon among children.

There are also some age- and sex-dependent differences in these types, with nodular sclerosis more common among females 15–19 years of age compared with males, and mixed cellularity occurring more often among younger than older children (Ries et al., 1999). Internationally, high rates of Hodgkin disease occur in Costa Rica and Israel (non-Jews), intermediate rates in US whites and Germany, and low rates in Africa (Parkin et al., 1998) (Fig. 65–2B). Overall, the US incidence rate was 13.6 per million for the period 1975–2000 (Ries, 2003). The incidence rate, however, is highly age dependent with rates increasing from <1 per million for children under the age of 5 years to 36 per million for children 15–19 years. Incidence rates are slightly higher in girls compared with boys, although in the youngest age group (<5 years) the incidence rate is higher in males (1.6/million and 0.3/million, respectively) (Ries et al., 1999). While black children have slightly lower incidence rates overall, incidence rates are essentially equivalent up until about 10 years of age. Overall, the incidence rates for HD in children have been declining in the United States. For all children under the age of 19 years, the incidence rate decreased from about 14.5 per million during the period 1975–1979 to 12.1 per million for the period 1990–1995 (Ries et al., 1999). Five-year survival rates exceed 90%, with whites experiencing higher survival rates than blacks (92% vs. 84%). Etiologic studies of HD in children suggest at least two distinct entities: one type of HD manifests with Epstein Barr virus (EBV) sequences present in the Reed-Sternberg cells (Stiller and Boyle, 1998). This type of HD is often seen in developing countries or in populations with lower socioeconomic status; it is also more common in males, and younger ages, and often is associated with the mixed cellularity subtype. The other type of HD appears in older adolescents of higher socioeconomic status; this type often appears as the nodular sclerosis subtype (Stiller and Boyle, 1998). A recent study examined Reed-Sternberg cells in children with HD from various countries for overall EBV infection and specifically tested for two strains, EBV1 and EBV2, in a subset of cases (Weinreb et al., 1996). EBV1 was found in 60% of cases, including 80% of children from the United Kingdom, 95% Greece, 100% South Africa, 100% Australia, 45% Costa Rica, but only 28% Kenya. In contrast, EBV2 was found in only 16% of all cases, but included 83% of children from Egypt and 36% from Kenya. Interestingly, some cases from the United Kingdom, Kenya, and Australia exhibited both strains of EBV. There is a strong genetic predisposition to risk, as monozygotic twins of adult patients have a nearly 100-fold increased risk (Mack et al., 1995). Further evidence of genetic susceptibility is apparent from observations of a high incidence among some populations in Southeast Asia, while a low incidence is observed among populations of East Asia (Varghese et al., 1996; Stiller and Boyle, 1998). These differences appear to be independent of socioeconomic status. Epidemiologic studies have tended to focus on three distinct age groups: children (ages 0–14 years), young adults (approximately 15–34 years) and older adults (50+ years) (MacMahon, 1966; Grufferman and Delzell, 1984; Varghese et al., 1996; Stiller and Boyle, 1998). The young adult group is primarily suspected to have an infectious etiology (Gutensohn and Cole, 1981). Studies have shown that young adults from higher socioeconomic status families are at increased risk of developing HD (Gutensohn and Delzell, 1982; Gutensohn and Shapiro, 1982; Grufferman and Delzell, 1984; Stiller and Boyle, 1998). Further, individuals from small families (e.g., few siblings) appear to be at an increased risk (Grufferman and Delzell, 1984). One case-control study of childhood HD reported that children diagnosed at less than 10 years of age lived in areas with a significantly lower socioeconomic status than controls, although this difference was not apparent for children diagnosed between 10 and 14 years of age (Gutensohn and Shapiro, 1982). In the largest current casecontrol study to date, which includes over 500 HD cases and controls, a few preliminary findings have been published: a protective effect for breastfeeding was noted (Grufferman, 1998), which confirms reports in smaller studies (Davis et al., 1988; Schwartzbaum et al., 1991). Further, there was a suggestion that first-degree relatives of patients younger than 15 years of age at diagnosis have an increased risk of malignancy (Grufferman, 1998).

Cancers in Children

Non-Hodgkin Lymphoma In contrast to adults, where low- and intermediate-grade tumors are common, high-grade tumors affect more than 90% of children with non-Hodgkin lymphoma (NHL) (Sandlund et al., 1996). These highgrade types include: lymphoblastic lymphomas, Burkitt type lymphoma (small noncleaved cell), diffuse large B-cell lymphomas, and anaplastic lymphoma (Ries et al., 1999). Burkitt lymphomas accounts for almost half of all cases diagnosed in Africa. Whereas Epstein Barr virus is linked with Burkitt lymphoma in African children, EBV is rarely associated with Burkitt lymphoma in the United States. Internationally, the highest incidence rates of NHL occur in Uganda and Egypt, intermediate rates occur in Spain, Germany, and New Zealand (Maori), and low rates in India and US blacks (Parkin et al., 1998) (Fig. 65–2B). In the United States, NHL comprises approximately 6% of all childhood cancers diagnosed under the age of 15 years. During the period 1990–1995, the age-specific annual incidence rates were 5.8, 8.7, 10.8, and 14.7 per million for ages 0–4, 5–9, 10–14, and 15–19 years, respectively (Ries et al., 1999). Nearly 70% of cases occur in males. Black children have notably lower incidence rates of NHL than white children, with an overall incidence of 6.8 per million compared to 8.9 per million for children under the age of 15 years, respectively. Five-year survival rates for childhood NHL approach 75%. Genetic factors associated with NHL include those associated with congenital immunodeficiency syndromes and are listed in Table 65–2 (e.g., ataxia-telangiectasia, Wiskott–Aldrich Syndrome, X-linked lymphoproliferative disease) (Sandlund et al., 1996). As with adults, children with acquired immunodeficiency syndrome (AIDS) and children who have received prior immunosuppressive therapy are at increased risk (Filipovich et al., 1992; Penn, 1994; Granovsky et al., 1998). There have been few etiologic studies of NHL and most have included only small numbers of cases. In a medical-record–based study of 34 cases of NHL diagnosed before 30 years of age and 68 matched controls, Roman et al., 1997) reported that viral infection during pregnancy was documented in the medical records of two case mothers and no control mothers. Another case-control study of 31 children with NHL reported that cases were significantly lighter at birth than corresponding controls (McKinney et al., 1987). In a nested-case control study of 1.7 million live births in Sweden, Adami et al. (1996) compared selected maternal and perinatal factors from the Swedish Medical Birth Registry for 168 cases of NHL and 840 matched controls. Very few factors were statistically significant. Paracervical anesthesia during labor as well as C-section was more common in the cases compared with the controls. Another study of 82 lymphoma cases from Shanghai reported that breastfeeding may be protective (Shu et al., 1995). In a case-control study of 268 children with NHL in the United States, a statistically significant association was found with reported pesticide exposure in the home (odds ratio = 7.3, p = 0.05 for use of pesticides most days in the home) (Buckley et al., 2000). Finally, as with childhood leukemia, population mixing has been implicated in etiology (Kinlen et al., 1995; Dickinson and Parker, 2002; Hemminki and Li, 2002).

SYMPATHETIC NERVOUS SYSTEM TUMORS (NEUROBLASTOMA) Approximately 700 children in the United States are diagnosed with a sympathetic nervous system tumor each year, the vast majority (650) of which are neuroblastomas (Ries et al., 1999). Neuroblastoma is an embryonal malignancy that arises from the primordial neural crest cells that form the adrenal medulla and sympathetic nervous system (Brodeur, 1997). Some of the highest incidence rates (13–15 cases/million) of neuroblastoma are reported in US whites, France, Canada, Japan, and Israeli Jews, whereas considerably lower rates (2–5 cases/million) are reported in Costa Rica and Uruguay (Parkin et al., 1998). Some areas, such as Japan and the province of Quebec instituted mass screening programs for neuroblastoma, which resulted in a notable increase in incidence rates (Ross and Davies, 1999).

1257

In the United States, the incidence of neuroblastoma is highly age dependent. In the first year of life the rate is 64 per million, while in the second year it is considerably lower (29/million) (Ries et al., 1999). Overall, incidence rates are only slightly higher in males than in females, although this difference is most pronounced in infancy (69.3/million vs. 59.6/million, respectively). Rates in white infants are somewhat higher than in black infants, although the overall rates across all ages are not that different. Overall trends in incidence have been relatively stable over the period 1976–1994, although there is a suggestion of the disease being diagnosed at earlier ages (Ries et al., 1999). This may be due to a general awareness in the United States of neonatal screening programs in other countries, which might enhance physician diagnosis. Five-year survival rates for neuroblastoma are dependent on the age at diagnosis, the tumor histology, and the cytogenetics. In some neuroblastomas, the oncogene, n-Myc, is amplified, which is associated with a very poor prognosis (Brodeur, 1997). For infants, 5-year survival rates are about 85%, whereas for children older than 1 year of age, rates are around 55% (Ries et al., 1999). A small proportion of neuroblastoma cases exhibit a genetic predisposition to the disease (Brodeur, 1997). In contrast to a median age of diagnosis of 22 months, these familial cases typically manifest the disease in the first year of life (Kushner et al., 1986). As with other embryonal tumors, etiologic studies have focused on exposures preconceptionally as well as in utero. A few studies have reported positive associations between farm residence or parental employment in agriculture, although findings have been inconsistent (Spitz and Johnson, 1985; Bunin et al., 1990b; Wilkins and Sinks, 1990; Kristensen et al., 1996; Kerr et al., 2000). Paternal employment in electronics-related occupations including electricians, linemen, and repairmen has been also associated with an increased risk, as has exposure to electromagnetic fields and aromatic hydrocarbons (Spitz and Johnson, 1985; Bunin et al., 1990b; Wilkins and Sinks, 1990). Daniels et al. (2001) recently examined the association between residential exposure to pesticides and neuroblastoma in a case-control study in the United States and Canada that included 538 cases and 538 controls selected through random-digit dialing. They found that parental report of pesticide use in the garden and home were modestly (odds ratios 1.6–1.7) associated with neuroblastoma. In an analysis from this same study (Olshan et al., 1999a), an increased risk was observed for fathers employed as broadcast/telephone operators, electrical power workers, landscapers/groundkeepers, and painters. Maternal occupations that were associated with elevated odds ratios included farmers/farm workers, florist and garden workers, and hair dressers. Perinatal factors have also been explored. Maternal use of sex hormones including oral contraceptives or infertility drugs has been associated with an increased risk of neuroblastoma in several studies (Kramer et al., 1987; Schwartzbaum, 1992; Michalek et al., 1996). Olshan et al. (1999b) found no association with oral contraceptive use before or during pregnancy; however, they did report an increased risk in males. Interestingly, other studies have reported a similar phenomenon (Michalek et al., 1996; Schuz et al., 2001). A few studies have reported that maternal medication use during pregnancy such as amphetamines, diuretics, and tranquilizers is associated with an increased risk (Kramer et al., 1987; Schwartzbaum, 1992), although these data are less consistent. Some studies have reported that increased birth weight is associated with an increased risk (Yeazel et al., 1997; Suminoe et al., 1999), others low birth weight (Daling et al., 1984; Johnson and Spitz, 1985; Hamrick et al., 2001), and some have found no association (Neglia et al., 1988; Buck et al., 2001). A few studies have reported positive associations with either maternal cigarette smoking or alcohol consumption before or during pregnancy, whereas others have not (Schwartzbaum, 1992; Yang et al., 2000; Buck et al., 2001). One recent study (Olshan et al., 2002) found that maternal vitamin use during pregnancy may reduce the risk of neuroblastoma, similar to a smaller case-control study in New York state (Michalek et al., 1996). These observations are consistent with the observed reduction in risk of neural tube defects and oral clefts seen with maternal vitamin use (Shaw et al., 1995; Botto et al., 1999; Werler

1258

PART IV: CANCER BY TISSUE OF ORIGIN

et al., 1999). Breastfeeding has also been associated with a decreased risk (Daniels et al., 2002). Few studies have had the statistical power to evaluate risk of neuroblastoma by grade, or N-myc status. A recent large case-control study evaluated the role of N-myc status and/or stage in a number of analyses and found no notable effects on the odds ratios (Yang et al., 2000; Daniels et al., 2001; Hamrick et al., 2001; Daniels et al., 2002; Olshan et al., 2002).

SOFT TISSUE SARCOMA (RHABDOMYOSARCOMA) Soft tissue sarcomas (STS) encompass tumors that occur in connective tissue, such as muscles, tendons, and fat. Approximately 900 children under the age of 19 years are diagnosed with a STS each year in the United States, representing 7.4% of all childhood malignancies (Ries et al., 1999). Internationally, incidence rates of STS are highest in Uganda, intermediate in the United States (whites and blacks) and Israeli Jews, and lowest in Thailand and India (Parkin et al., 1998). In the United States, males have slightly higher rates than females, and blacks slightly higher rates than whites (Ries et al., 1999). For children ages 0–14 years, rhadomyosarcomas represent about 50% of all STS diagnosed, with an incidence rate of approximately 4.6 per million. Nearly 75% of rhabdomyosarcomas have an embryonal histology. Other types of STS include fibrosarcomas, synovial sarcomas, leiomyosarcoma, lipsarcoma, malignant fibrous histiosarcoma, and others. The overall 5-year survival rates for STS are approximately 70% (Ries et al., 1999), although children with rhadomyosarcoma experience slightly poorer 5-year survival rates of around 64%. Younger children fare better than older children with rhabdomyosarcoma, as do whites compared with blacks, and males compared with females. As rhabdomyosarcomas comprise the largest homogeneous group of malignant STS in children, epidemiologic studies have focused on this subtype. A few known risk factors for STS exist. Several studies have linked major birth syndromes and defects (including Beckwith–Wiedemann, Costello syndrome, neurofibromatosis, and genitourinary anomalies) with rhabdomyosarcoma (Table 65–2), although these associations occur in only a small fraction of cases (Ruymann et al., 1988; Yang et al., 1995; Gripp et al., 2002). Other genetic conditions, such as Li-Fraumeni syndrome, are also associated with an increased risk (Wexler and Helman, 1997). Few case-control studies have investigated the etiology of STS or rhabdomyosarcoma (RMS). In a case-control study of 52 cases of STS (36 RMS) and 326 controls in Italy, a non-statistically significant increased risk was reported with older maternal age and in utero exposure to radiation (Magnani et al., 1989). A case-control study of 43 childhood STS and 86 matched controls enrolled in the Inter-Regional Epidemiological Study of Childhood Cancer reported that maternal toxemia during pregnancy and fewer previous pregnancies were associated with an increased risk (Hartley et al., 1988). In a small (33 cases and 99 controls) case-control study in North Carolina, positive associations were reported with paternal cigarette smoking, advanced maternal age, and lower socioeconomic status (Grufferman et al., 1982). The largest case-control study to date, 322 cases and 322 age-, sex-, and race-matched controls, reported that low socioeconomic status, diagnostic X-rays during pregnancy, and parental use of marijuana and cocaine may increase risk (Grufferman, 1991; Grufferman et al., 1993).

RENAL CANCERS (WILMS TUMOR, RENAL CARCINOMA) Wilms tumor (WT), or nephroblastoma, is a cancer of kidney cell progenitors. It is the most common renal tumor, especially in the first 10 years of life when it constitutes about 95% of such malignancies. Among children aged 10–14 years renal carcinoma makes up about one-third of kidney cancers, the rest being WT (Ries et al., 1999). The rate of WT in most nations ranges between 4 and 10 cases per million and is lower, in general, in developing countries than in industrialized

ones (Fig. 65–2B). Renal carcinoma is universally rare in children (Parkin et al., 1998). Like other embryonal tumors, incidence of WT is greatest in infancy, at 18.3 cases per million, declines to 5.6 cases per million at 4 years of age, and thereafter has a rate of less than 1 case per million. The rate of renal carcinoma is below 1 case per million for all age groups less than 20 years of age (Ries et al., 1999). The rate of WT is higher in females than in males and higher in blacks than in whites; these disparities are greatest among 0–4 year olds and lessen with age (Ries et al., 1999). WT registered no significant change in incidence between 1974 and 1991, except among males ages 5–9 years, whose rate changed by an average of 4.6% per year (95% CI: 1–9.4) (Gurney et al., 1996). Wilms tumor may involve one or, as in 5%–10% of cases, both kidneys (Bonaiti-Pellie et al., 1992), which suggests a two-hit mutation model of tumorigenesis as with retinoblastoma (Knudson et al., 1972). In support of this contention, bilateral WT occurs significantly earlier than does unilateral disease (Breslow et al., 1988). However, the fact that unilateral cases with the Wilms tumor, aniridia, genitourinary anomalies, and mental retardation (WAGR) syndrome also were younger than unilateral cases without WAGR suggested that more than one locus is involved in WT oncogenesis (Breslow et al., 1988, 1996). Several loci involved in Wilms tumorigenesis have now been identified. WT1 encodes a transcription factor and tumor suppressor at the 11p13 locus; mutation or deletion of which leads to WT in a minority of cases (Dome et al., 2002). The WT2 locus at 11p15 encodes a variety of genes, including the one for insulin-like growth factor 2 (IGF-2) (Dome et al., 2002). Overexpression of IGF-2, as has been observed in WT, may promote cell growth and lead to cancer (Ogawa et al., 1993; Rainier et al., 1993). Linkage analysis has also identified two other putative WT loci called FWT1 and FWT2 (for familial WT) at 17q and 19q, respectively, and still more loci may play a part in Wilms tumorigenesis (Dome et al., 2002). Besides congenital conditions, few risk factors for WT display great consistency. Isolated reports have linked WT with parental exposure to pesticides (Olshan et al., 1993; Sharpe et al., 1995; Schuz et al., 2001). Exposures to the mother during pregnancy or birth have been of interest; use of coffee or tea (Bunin et al., 1987; Olshan et al., 1993; Schuz et al., 2001), hair dye (Bunin et al., 1987; Olshan et al., 1993), and medications (Lindblad et al., 1992; Sharpe and Franco, 1996) have been associated with WT in some studies but not others. Studies of paternal occupation have suggested an increased risk in children of fathers who work with hydrocarbons, lead, or other metals (Fabia and Thuy, 1974; Kantor et al., 1979; Wilkins et al., 1984a; Wilkins et al., 1984b; Bunin et al., 1989b; Olshan et al., 1990). These associations were often strongest for exposure before conception of the child with WT. Significant associations of WT with high birth weight have been observed, though among various subsets of cases (Daling et al., 1984; Bunin et al., 1987; Leisenring et al., 1994; Heuch et al., 1996; Yeazel et al., 1997; Smulevich et al., 1999; Schuz et al., 2001). Two studies found no significant association of WT with birth weight (Lindblad et al., 1992; Olshan et al., 1993). Still, the frequent overexpression of IGF-2 found in WT lends support to an association with birth weight (Ogawa et al., 1993).

BONE TUMORS (OSTEOSARCOMA, EWING SARCOMA) Osteosarcoma (OS) and Ewing sarcoma (ES) comprise 56% and 34%, respectively, of malignant bone tumors in children. They originate from different tissues; OS arises from primitive bone-forming mesenchymal stem cells, whereas ES arises from the neural crest (Ries et al., 1999). The two tumors display contrasting patterns of international incidence. OS incidence ranges mainly between 1 and 4 cases per million and shows no striking associations across nations. The rate of ES among black Africans and black Americans, however, <1 case per million, is very low compared with rates of between 2 and 4 cases per million for white Americans and Europeans (Parkin et al., 1998). In the United States incidence of both cancers is extremely low in early childhood, grows steadily during the ages of 5–9 years, and peaks in mid adolescence at about 11 cases per million for OS and 6 cases per

Cancers in Children million for ES. The rate of both cancers is slightly higher among males than among females until 15 years of age, when the gap widens. Whereas the rate for OS among blacks is slightly higher than that for whites, nearly all ES occurs among whites (Ries et al., 1999). The rate of OS among children aged less than 15 years rose by a significant 2.4% (95% CI: 0.3–4.7) annually between 1974 and 1991, mainly among 5–9 year olds. The rate of ES rose a significant 3.4% annually (95% CI: 0.2–6.6) among 10–14 year olds; no significant rise was seen among other age groups (Gurney et al., 1996). The cytogenetics of OS are complex, with no single genetic rearrangement predominating (Bridge et al., 1997). By contrast, about 90% of ES displays the t(11;22) translocation (Ladanyi et al., 1993). The EWS gene on chromosome 22 is therein fused with a variety of other genes to produce a dominant-acting oncoprotein (Arvand and Denny, 2001). Survival of OS and ES varies by histologic subtype (Ferrari et al., 2001), but is low compared with other childhood cancers (Table 65–1) (Ries et al., 2003). Osteosarcoma has long been a recognized consequence of the inherited cancer-predisposing Li-Fraumeni and retinoblastoma syndromes (Hansen et al., 1985), but these rare conditions account for few cases. The close correlation between the incidence of OS and the childhood growth curve (Fraumeni and Miller, 1967), as well as the frequent occurrence of the diseases in the long bones of the lower limbs during adolescence (Ries et al., 1999), suggests an etiology linked with bone development. Veterinary studies show that large breeds of dogs develop OS at a much higher rate than do small breeds (Tjalma, 1966), as do dogs that are spayed or neutered at an early age (Cooley et al., 2002). Some reports have indicated that OS cases are taller at diagnosis than are controls (Fraumeni and Miller, 1967; Gelberg et al., 1997), whereas other studies have not borne this out (Operskalski et al., 1987; Buckley et al., 1998). Other factors that might reasonably be related to growth, such as birth weight and the age at onset of secondary sexual characteristics, have shown no consistent pattern (Hartley et al., 1988; Gelberg et al., 1997; Buckley et al., 1998). Parental adult height also has not been associated with OS (Budkley et al., 1998). The association of ES with factors related to growth is similarly mixed (Fraumeni and Miller, 1967; Hartley et al., 1988; Winn et al., 1992; Buckley et al., 1998). Some evidence points to a genetic predisposition for ES. Most striking is the above-mentioned rarity of ES among black children both in the United States (Ries et al., 1999) and in Africa (Parkin et al., 1998). Also, ES tumors are more evenly distributed among the bones of the body, unlike OS, which occurs mainly in the long bones of the lower limbs (Ries et al., 1999). However, there is little evidence of a higher risk of ES among family members of cases (Hartley et al., 1991), though risk of other cancers may be raised (Novakovic et al., 1994). Exploratory analyses of a wide range of other factors have not revealed any notable risk factors for OS or ES (Operskalski et al., 1987; Hartley et al., 1988; Winn et al., 1992; Buckley et al., 1998). With the exception of farm work, which has been associated with ES in several studies (Holly et al., 1992; Winn et al., 1992; Hum et al., 1998; Valery et al., 2002), few parental occupational exposures have been consistently associated with bone cancer (Operskalski et al., 1987; Hartley et al., 1988; Buckley et al., 1998; Hum et al., 1998). Radium and fluoride, of concern since they are incorporated into bone, have mostly not been associated with OS or ES in case-control studies (McGuire et al., 1991; Finkelstein, 1994; Gelberg et al., 1995; Moss et al., 1995; Finkelstein and Krieger, 1996). Radiotherapy is an established risk factor for bone sarcomas, but is mostly of historical interest (Tucker et al., 1987; Newton et al., 1991). Two ecological analyses have suggested a higher incidence of bone tumors in urban areas (Larsson and Lorentzon, 1974; Silva and Subrarnian, 1975). Two other studies showed no evidence of clustering of bone tumors (Glass and Fraurneni, 1970; Silcocks and Murrels, 1987).

GERM CELL TUMORS Germ cell tumors (GCTs) arise from the primordial germ cells during fetal development. The tumors that can develop are extremely hetero-

1259

geneous, and often manifest a more benign rather than malignant phenotype, particularly in the younger age groups (Castleberry, 1997). GCTs are grouped by both location and cells of origin and include: intracranial and intraspinal, other, unspecified non-gonadal tumors such as those in the sacrococcygeal or mediastinal regions, gonadal tumors of the testes or ovaries, gonadal carcinoma of the testes or ovaries, and other, unspecified gonadal tumors (Ries et al., 1999). With the exception of slightly higher incidence rates reported in Japan (9.6 cases/million) and New Zealand (Maori, 8.2 cases/million), there is little variation in the rates of GCTs; most areas range from 1–7 cases/million (Parkin et al., 1998). Collectively, the annual incidence rate of malignant GCTs in the United States is low; only about 5 cases per million children occur under the age of 15 years, and about 12 cases per million under the age of 20 years. The incidence rate is only slightly higher in males than in females. Black children have somewhat lower rates than white children (Ries et al., 1999). Notably, black males have lower rates of testicular GCTs compared with white males (1.2 vs. 9.1/million, respectively). Overall, incidence rates are elevated in infancy and then decline until about the age of 10 years, where they rapidly rise for both males and females. Comparing the periods 1975–1979 and 1990–1995, the overall incidence rate of GCTs in children increased from 8.5 per million to 12.0 per million. Much of this increase, however, is likely due to the inclusion of tumors that once would have been considered borderline malignant (Ries et al., 1999). Overall, 5-year survival rates for children diagnosed with a GCT are quite good, approaching 90%. Due to its rarity, very few epidemiologic studies have explored risk factors for malignant GCTs in children, although some studies have focused on testicular tumors in adolescent and adult populations. Cryptorchidism is one of the few established risk factors for testicular GCTs (Strader et al., 1988). Maternal exogenous estrogen exposure and/or high endogenous hormone levels during pregnancy may also be associated with an increased risk of testicular GCTs, although the evidence is inconsistent (Henderson et al., 1979; Depue et al., 1983; Walker et al., 1988). Other purported risk factors include radiation exposure during pregnancy, pre-term birth, congenital malformation, viral infections such as mumps, and certain parental occupational exposures (Li et al., 1972; Li et al., 1973; Henderson et al., 1979; Schottenfeld et al., 1980; Birch et al., 1982; Depue et al., 1983; Algood et al., 1988; Kardaun et al., 1991). In the most recent case-control study to date, Shu et al. (1995) compared responses on a structured self-administered questionnaire from parents of 105 patients with malignant GCT and 639 community controls. They found that infants born at term had a 70%–75% reduction in the risk of a malignant GCT compared with children born before 38 weeks gestation. Further, this same study found that certain parental self-reported exposures to chemicals and solvents were associated with an increased risk. High birth weight, prolonged breast feeding, and maternal urinary tract infection during pregnancy were also associated with an increased risk. In contrast, maternal cigarette smoking was associated with a decreased risk. Given the heterogeneity of GCTs, it will be important for future studies to investigate risk factors by subtype. The Children’s Oncology Group has recently completed the largest case-control study of childhood malignant GCTs to date. This study included over 300 cases in the United States and Canada and 400 controls selected through random-digit dialing. Hypotheses to be tested in this study include potential associations with maternal exogenous estrogen use and prenatal occupational exposures. Analyses are underway and should be available within the next few years.

RETINOBLASTOMA Retinoblastoma (RB) is a malignant tumor of the primitive neuroectodermal cells of the retina and affects approximately 300 children in the United States each year, accounting for about 3% of all malignancies diagnosed in children under the age of 15 years. There is little international variation in incidence rates, with most countries reporting an incidence of between 3–8 cases/million (Parkin et al., 1998). The majority (63%) of RBs are diagnosed in children under the age of 2 years, and nearly all (95%) cases are diagnosed before the age of

1260

PART IV: CANCER BY TISSUE OF ORIGIN

5 years (Ries et al., 1999). Males and females are equally affected, as are blacks and whites. Incidence rates for RB in the United States have remained essentially stable over the past 20 years (about 4 cases/million). Recent 5-year survival estimates are quite good (93%). The study of RB led to the two-hit theory of carcinogenesis (Knudson, 1971). It was later demonstrated that inactivation of both alleles of the retinoblastoma (RB1) gene at chromosome band 13q14 was required for the cancer to develop (Cavenee et al., 1983; Dryja et al., 1986). About 10% of children with RB inherit a mutated RB1 gene from one of their parents. In this familial form, all cells harbor the RB1 gene mutation, and nearly all children will have a “second hit” occur in the retinal cell sometime after conception. Another form of heritable RB, which affects another 30% of children with RB, arises from an RB1 gene mutation that occurred in the germ cells of one of their parents. This is referred to as sporadic heritable RB, as neither parent is afflicted, but the child has an RB1 gene mutation in all of their cells. As with familial RB, over 90% of individuals carrying one mutated RB1 gene will eventually develop the disease (Ries et al., 1999). Finally, the remaining 60% of RB patients have nonheritable disease. This type of RB develops as the result of two somatic RB1 gene mutations occurring in a single cell sometime after conception. The heritable forms of RB typically are diagnosed in the first year of life and often both eyes are affected. In contrast, the non-heritable form manifests at a slightly older age and usually only one eye is affected (Ries et al., 1999). Interestingly, the vast majority of new germinal mutations in sporadic heritable RB occur in the paternal RB1 allele (Dryja et al., 1989). Thus, preconceptional exposures of the father (not the mother) would be etiologically relevant in sporadic heritable RB. In contrast, postconceptional factors likely play a role in non-heritable RB. Due to the extreme rarity of the disease, there have been very few etiologic studies of RB. One of the largest case-control studies of RB, conducted in the former Children’s Cancer Group, included 201 RB cases (19 familial, 67 sporadic heritable, and 115 non-heritable) and 201 controls. The authors reported that paternal employment in the military and metal manufacturing was associated with an increased risk of sporadic heritable RB (Bunin et al., 1990a). For non-heritable RB, increased risks were observed with paternal occupations as welders and machinists. Other findings from this same study suggested that maternal morning sickness medication, and gestational exposure to X-rays were associated with an increased risk, whereas maternal anemia, multivitamin use during pregnancy, and peri-conceptional use of barrier contraceptives or spermicides were associated with a decreased risk (Bunin et al., 1989a). At least three studies have reported that advanced paternal age is associated with an increased risk of sporadic heritable RB (DerKinderen et al., 1990; Moll et al., 1996; Dockerty et al., 2001); two of these studies also demonstrated that maternal advanced age was a risk factor (DerKinderen et al., 1990; Moll et al., 1996). Finally, a recent study from the Netherlands explored potential associations between in vitro fertilization (IVF) and risk of RB (Moll et al., 2003). Five cases of retinoblasoma were diagnosed in children conceived through IVF during the period November 2000 through February 2002. Based on the number of children born after IVF, and the incidence estimates of the number of RB cases expected, the authors estimated a relative risk of 7.2 for developing RB in children conceived through IVF. This intriguing observation deserves further exploration.

HEPATIC TUMORS (HEPATOBLASTOMA, HEPATOCELLULAR CARCINOMA) Hepatoblastoma (HB) and hepatocellular carcinoma (HCC) are cancers of immature and differentiated liver cells, respectively. HB is the most common malignancy of the liver in persons aged less than 15 years; between 1990–1995 it comprised about 90% of hepatic tumors at those ages (Ries et al., 1999). Internationally, incidence of HB ranges mainly between 0.5 and 2 cases per million, and HCC even

lower, though these estimates are often unstable due to the rarity of these cancers (Parkin et al., 1998). In the United States, incidence is greatest in infancy, at 8.7 cases per million, declines to 0.3 cases per million at 4 years of age, and thereafter becomes even more rare (Ross and Gurney, 1998; Ries et al., 1999). Males are slightly more commonly diagnosed with HB than are females (male : female ratio = 1.2), whereas rates are roughly similar for blacks and whites (Ries et al., 1999). HB incidence appears to be rising recently. Between 1973 and 1992, the rate of HB among 0–4 year olds increased by an average of 5.2% per year (95% CI: 1.9–8.6%); the rise in incidence was significant among females but not males (Ross and Gurney, 1998). The addition of chemotherapy to surgery for HB has improved survival such that 75%–80% of cases can be cured (Carceller et al., 2001). HCC is extremely rare below the age of 15, with a maximum incidence of 0.3 cases per million. However, by ages 15–19 years HCC has an incidence of 0.9 cases per million and is many times more common than HB (Ries et al., 1999). Because of its rarity, most knowledge about risk factors for HB has been gleaned from case reports or series. Some HBs display sporadic mutations of the APC gene (Oda et al., 1996) and, accordingly, the rate of HB is vastly increased in children who are potentially carriers of the familial adenomatous polyposis gene (i.e., APC) (Giardiello et al., 1991). Incidence is also increased among children with Beckwith–Wiedemann syndrome (DeBaun et al., 1998), which has lead to the suggestion that screening children with these conditions may be an effective way to reduce mortality due to HB (Giardiello et al., 1991; McNeil et al., 2001). Genetic analyses indicate that some HBs display abnormal imprinting of IGF-2 alleles (Ross et al., 2000). Case-control studies of HB are few. An exploratory study that consisted of 75 cases and 75 controls found that mothers of children with HB were significantly more often occupationally exposed to metals, petroleum products, and paints and pigments before or during pregnancy, while fathers of children with HB were significantly more often occupationally exposed only to metals (Buckley et al., 1989). The study did not find evidence of hypothesized associations of HB with hepatitis, maternal alcohol consumption, or maternal smoking. A smaller study found parental smoking significantly increases risk of HB (Pang et al., 2003). However, that finding may have been confounded by low birth weight, which can be caused by smoking (Horta et al., 1997) and seems to be a risk factor for HB. A Japanese report first noted that the percentage of HB cases that weighed 1500 grams or less at birth increased significantly between the late 1980s and the early 1990s, when survival of very low birth weight babies improved (Ikeda et al., 1997). Subsequent studies confirmed that the rate of HB was significantly higher among very low birth weight babies compared with those with normal birth weight in Japan (Tanimura et al., 1998), and that the proportion of low birth weight among US HB cases was unusually high (Feusner and Plaschkes, 2002). Two reasons are possible for these data. The first is that improved treatment of low birth weight babies is allowing HB cases to survive (where in the past they would have died either before birth or shortly thereafter). Alternatively, treatment for low birth weight itself may cause HB. Though one study conducted in response to recent revelations suggested that HB in very low birth weight infants may be related to the length of therapy for prematurity (Maruyama et al., 2000), more studies are needed before it can be determined which explanation is correct. There are few studies of HCC at ages younger than 20 years. However, chronic hepatitis B and C virus infections, host responses to these pathogens, and exposure to cofactors such as aflatoxin are established risk factors for HCC in adults (Kasai et al., 1996). Findings that hepatitis B vaccination has reduced incidence of HCC in adolescents strongly suggests that the etiology among the young is similar to that in adults (Chang et al., 2000; Lee et al., 2003).

FUTURE DIRECTIONS Due to its heterogeneity and rarity, investigators conducting etiologic studies of childhood cancer in the United States have often collabo-

Cancers in Children rated with hospitals and institutions affiliated with the National Cancer Institute pediatric cooperative clinical trials groups. Importantly, hospitals and institutions affiliated with the former Children’s Cancer Group and the Pediatric Oncology Group are estimated to treat nearly 90% of all malignancy diagnosed in children less than 15 years of age (Ross et al., 1996). In 2000, the NCI pediatric cooperative clinical trials groups were merged to form the Children’s Oncology Group (COG). Over 220 hospitals and institutions in the United States and Canada are affiliated with the COG. The administrative structure of COG consists of disease and scientific discipline committees, including the Epidemiology Committee.1 The mission of the COG Epidemiology Committee is to promote and facilitate research investigating the causes of childhood cancer. Recognizing recent accomplishments, as well as challenges, the Epidemiology Committee has proposed several areas for future directions in childhood cancer research in the United States including formation of a North American Pediatric Cancer Registry; conduct of studies to improve methodological approaches; and expanded investigations of gene-environment interactions, familial cancer syndromes, and viruses in childhood cancer etiology. Each of these is described briefly below.

Formation of a North American Pediatric Cancer Registry Since the hospitals and institutions affliated with the COG treat the majority of children with cancer (Ross et al., 1996), the COG Epidemiology committee, in collaboration with the National Cancer Institute, and State, Provincial, and regional cancer registries are working to establish the Childhood Cancer Research Network (CCRN). The CCRN will form the basis of a North American pediatric cancer registry. Currently, the protocol for registration to the CCRN is being piloted at 10% of COG institutions. The protocol requests that once a child is diagnosed with cancer, a consent form be administered to the parent (and child, if age eligible) to be registered to the CCRN with personal identifiers. Additionally, a consent form is administered to request permission to be contacted in the future to consider taking part in a non-therapeutic study. Establishing the CCRN should help to overcome several challenges (e.g., small numbers of cases, lack of participation by some hospitals and institutions, and delays in obtaining human subjects approval) in conducting etiologic studies of childhood cancer in COG and provide an unequaled resource for research in the United States and Canada.

Conduct of Studies to Improve Methodological Approaches As we have described previously, several case-control studies of childhood cancer suggest an important role for specific parental and childhood exposures in etiology. However, studies of childhood cancer often rely solely on questionnaire data obtained from interviews conducted with the parents of children with and without cancer. While a few studies have incorporated additional measures of exposure (e.g., electromagnetic field levels) (Linet et al., 1997), more studies with adjunct measurements are needed. Further, there is difficultly in recruiting appropriate control groups for case-control comparisons. Most large-scale case-control studies of childhood cancer in the United States have used random digit-dialing for recruitment of controls. With the increasing use of answering machines, caller identification, and cell phones, acceptable alternatives to random-digit dialing must be identified. Alternative control groups that could be explored include birth roster controls, which have been successfully used in smaller studies (Buck et al., 2001; Ma et al., 2002a; Ma et al., 2002b). One proposed study will examine the feasibility of using birth certificate controls on a national basis.

1

J. A. Ross is currently Chair of the COG Epidemiology Committee.

1261

Expanded Investigations of Gene-Environment Interactions, Familial Cancer Syndromes, and Viruses in Childhood Cancer Etiology As summarized in Table 65–2, known familial cancer syndromes are relevant to the development of childhood cancers (reviewed in Volgelstein and Kinzler, 1998). Studies of familial cancer clusters have, and will continue to lead to the discovery of cancer-predisposition genes. For most children, however, the etiology of cancer is probably multifactorial and related to a combination of genetically determined host susceptibility factors and exposures to carcinogens. As we have described, a wide range of environmental factors are associated with an increased risk of leukemia and other childhood cancers. These observations provide support that environmental exposures in the context of host (or parental) genetic susceptibility to genotoxic damage may be the major determinants of childhood cancer risk. However, further identification of the importance of gene-environment interactions, as well as major gene effects and viruses are needed. Some new studies being proposed by the COG Epidemiology Committee include studies of childhood leukemia that include the incorporation of genetic polymorphism in both the mother and infant to evaluate the role of gene-environment interactions, mechanistic studies of pregnancy to evaluate placental transfer as well as potential modifications by genotype, characterization of the role of polyomaviruses in the development of cancer in families with a history of Li-Fraumeni syndrome, and collection of family history at defined intervals to identify evolving familial cancer syndromes. Acknowledgments This work was supported in part by the University of Minnesota Children’s Cancer Research Fund. The authors thank Ms. Angela Smit for assistance in the creation of the figures.

References Adami J, Glimelius B, Cnattingius S, et al. 1996. Maternal and perinatal factors associated with non-Hodgkin’s lymphoma among children. Int J Cancer 65(6):774–777. Ahlbom A, Day N, Feychting M, et al. 2000. A pooled analysis of magnetic fields and childhood leukaemia. Br J Cancer 83(5):692–698. Alexander FE, Boyle P, Carli PM, et al. 1998. Spatial temporal patterns in childhood leukaemia: Further evidence for an infectious origin. EUROCLUS project. Br J Cancer 77(5):812–817. Alexander FE, McKinney PA, Cartwright RA. 1990. Radon and leukaemia. Lancet 335:1336–1337. Algood CB, Newell GR, Johnson DE. 1988. Viral etiology of testicular tumors. J Urol 139(2):308–310. Anonymous. 1999. Exposure to power-frequency magnetic fields and the risk of childhood cancer. UK Childhood Cancer Study Investigators. Lancet 354(9194):1925–1931. Anonymous. 2000. Childhood cancer and residential proximity to power lines. UK Childhood Cancer Study Investigators. Br J Cancer 83(11):1573–1580. Arvand A, Denny CT. 2001. Biology of EWS/ETS fusions in Ewing’s family tumors. Oncogene 20(40):5747–5754. Auvinen A, Hakama M, Arvela H, et al. 1994. Fallout from Chernobyl and incidence of childhood leukaemia in Finland, 1976–92. BMJ 309(6948): 151–154. Axelson O, Fredrikson M, Akerblom G, et al. 2002. Leukemia in childhood and adolescence and exposure to ionizing radiation in homes built from uranium-containing alum shale concrete. Epidemiology 13(2):146–150. Baccate EM. 1983. Social patterns of antibody to poliovirus. Lancet 1:778–783. Bader JL, Miller RW. 1978. Neurofibromatosis and childhood leukemia. J Pediatr 92(6):925–929. Barbanti-Brodano G, Martini F, De Mattei M, et al. 1998. BK and JC human polyomaviruses and simian virus 40: Natural history of infection in humans, experimental oncogenicity, and association with human tumors. Adv Virus Res 50:69–99. Bhatia S, Landier W, Robison, LL. 2002. Late effects of childhood cancer therapy. In: DeVita VT, Hellman S, Rosenberg SA, eds. Progress in Oncology. Sudbury MA; Jones and Bartlett: pp. 171–201. Bhatia S, Sklar C. 2002. Second cancers in survivors of childhood cancer. Nat Rev Cancer 2(2):124–132. Bhattacharjee MB, Armstrong DD, Vogel H, et al. 1997. Cytogenetic analysis of 120 primary pediatric brain tumors and literature review. Cancer Genet Cytogenet 97(1):39–53.

1262

PART IV: CANCER BY TISSUE OF ORIGIN

Birch JM, Marsden HB, Swindell R. 1982. Pre-natal factors in the origin of germ cell tumours of childhood. Carcinogenesis 3(1):75–80. Bloomfield CD, Brunning RD. 1976. Acute leukemia as a terminal event in nonleukemic hematopoietic disorders. Semin Oncol 3(3):297–317. Boice JD, Jr. 1986. The danger of X-rays—real or apparent? N Engl J Med 315(13):828–830. Bonaiti-Pellie C, Chompret A, Tournade MF, et al. 1992. Genetics and epidemiology of Wilms’ tumor: The French Wilms’ tumor study. Med Pediatr Oncol 20(4):284–291. Botto LD, Moore CA, Khoury MJ, et al. 1999. Neural-tube defects. N Engl J Med 341(20):1509–1519. Boutou O, Guizard A-V, Slama R, et al. 2002. Population mixing and leukaemia in young people around the La Hague nuclear waste reprocessing plant. Br J Cancer 87:740–745. Breslow N, Beckwith JB, Ciol M, et al. 1988. Age distribution of Wilms’ tumor: Report from the National Wilms’ Tumor Study. Cancer Res 48(6):1653– 1657. Breslow NE, Olson J, Moksness J, et al. 1996. Familial Wilms’ tumor: A descriptive study. Med Pediatr Oncol 27(5):398–403. Bridge JA, Nelson M, McComb E, et al. 1997. Cytogenetic findings in 73 osteosarcoma specimens and a review of the literature. Cancer Genet Cytogenet 95(1):74–87. Brodeur GM, Castleberry RP. 1997. Neuroblastoma. In: Pizzo PA, Poplack DG, eds. Principles and Practice of Pediatric Oncology. Philadelphia, Lippincott-Raven: pp. 761–797. Brondum J, Shu XO, Steinbuch M, et al. 1999. Parental cigarette smoking and the risk of acute leukemia in children. Caner 85(6):1380–1388. Buck GM, Michalek AM, Chen CJ, et al. 2001. Perinatal factors and risk of neuroblastoma. Paediatr Perinatal Epidemiol 15(1):47–53. Buckley JD, Buckley CM, Ruccione K, et al. 1994. Epidemiological characteristics of childhood acute lymphocytic leukemia. Analysis by immunophenotype. The Childrens Cancer Group. Leukemia 8(5):856– 864. Buckley JD, Meadows AT, Kadin ME, et al. 2000. Pesticide exposures in children with non-Hodgkin lymphoma. Cancer 89(11):2315–2321. Buckley JD, Pendergrass TW, Buckley CM, et al. 1998. Epidemiology of osteosarcoma and Ewing’s sarcoma in childhood: A study of 305 cases by the Children’s Cancer Group. Cancer 83(7):1440 –1448. Buckley JD, Robison LL, Swotinsky R, et al. 1989. Occupational exposures of parents of children with acute nonlymphocytic leukemia: A report from the Childrens Cancer Study Group. Cancer Res 49(14):4030 – 4037. Buckley JD, Sather H, Ruccione K, et al. 1989. A case-control study of risk factors for hepatoblastoma. A report from the Childrens Cancer Study Group. Cancer 64(5):1169–1176. Bunin GR, Buckley JD, Boesel CP, et al. 1994. Risk factors for astrocytic glioma and primitive neuroectodermal tumor of the brain in young children: A report from the Children’s Cancer Group. Cancer Epidemiol Biomarkers Prev 3(3):197–204. Bunin GR, Kramer S, Marrero O, et al. 1987. Gestational risk factors for Wilms’ tumor: Results of a case-control study. Cancer Res 47(11):2972– 2977. Bunin GR, Kuijten RR, Boesel CP, et al. 1994. Maternal diet and risk of astrocytic glioma in children: A report from the Childrens Cancer Group (United States and Canada) [comment]. Cancer Causes Control 5(2):177–187. Bunin GR, Kuijten RR, Buckley JD, et al. 1993. Relation between maternal diet and subsequent primitive neuroectodermal brain tumors in young children [comment]. N Engl J Med 329(8):536–541. Bunin GR, Meadows AT, Emanuel BS, et al. 1989a. Pre- and postconception factors associated with sporadic heritable and nonheritable retinoblastoma. Cancer Res 49(20):5730–5735. Bunin GR, Nass CC, Kramer S, et al. 1989b. Parental occupation and Wilms’ tumor: Results of a case-control study. Cancer Res 49(3):725–729. Bunin GR, Petrakova A, Meadows AT, et al. 1990a. Occupations of parents of children with retinoblastoma: A report from the Children’s Cancer Study Group. Cancer Res 50(22):7129–7133. Bunin GR, Ward E, Kramer S, et al. 1990b. Neuroblastoma and parental occupation. Am J Epidemiol 131(5):776–780. Carceller A, Blanchard H, Champagne J, et al. 2001. Surgical resection and chemotherapy improve survival rate for patients with hepatoblastoma. J Pediatr Surg 36(5):755–759. Carozza SE, Olshan AF, Faustman EM, et al. 1995. Maternal exposure to N-nitrosatable drugs as a risk factor for childhood brain tumours. Int J Epidemiol 24(2):308–312. Carroll-Pankhurst C, Engels EA, Strickler HD, et al. 2001. Thirty-five year mortality following receipt of SV40- contaminated polio vaccine during the neonatal period. Br J Cancer 85(9):1295–1297.

Cartwright RA, McKinney PA, Alexander FE, et al. 1988. Leukemia in young children. Lancet 2(8617):960. Castleberry RP, Cushing B, Perlman E, Hawkins EP. 1997. Germ cell tumors. In: Pizzo PA, Poplack DG, eds. Principles and Practice of Pediatric Oncology. Philadelphia; Lippincott-Raven: pp. 921–946. Cavenee WK, Dryja TP, Phillips RA, et al. 1983. Expression of recessive alleles by chromosomal mechanisms in retinoblastoma. Nature 305(5937):779– 784. Chan LC, Lam TH, Li CK, et al. 2002. Is the timing of exposure to infection a major determinant of acute lymphoblastic leukaemia in Hong Kong? Paediatr Perinatal Epidemiol 16(2):154–165. Chang MH, Shau WY, Chen CJ, et al. 2000. Hepatitis B vaccination and hepatocellular carcinoma rates in boys and girls. JAMA 284(23):3040–3042. Cimino G, Lo Coco F, Biondi A, et al. 1993. ALL-1 gene at chromosome 11q23 is consistently altered in acute leukemia of early infancy. Blood 82(2):544–546. Cnattingius S, Zack MM, Ekbom A, et al. 1995. Prenatal and neonatal risk factors for childhood lymphatic leukemia. J Natl Cancer Inst 87(12):908– 914. Cooley DM, Beranek BC, Schlittler DL, et al. 2002. Endogenous gonadal hormone exposure and bone sarcoma risk. Cancer Epidemiol Biomarkers Prev 11(11):1434–1440. Cordier S, Iglesias MJ, Le Goaster C, et al. 1994. Incidence and risk factors for childhood brain tumors in the lle de France. Int J Cancer 59(6): 776–782. Cordier S, Lefeuvre B, Filippini G, et al. 1997. Parental occupation, occupational exposure to solvents and polycyclic aromatic hydrocarbons and risk of childhood brain tumors (Italy, France, Spain) [erratum appears in Cancer Causes Control 1997 Nov;8(6):934]. Cancer Causes Control 8(5):688–697. Cordier S, Mandereau L, Preston-Martin S, et al. 2001. Parental occupations and childhood brain tumors: Results of an international case-control study. Cancer Causes Control 12(9):865–874. Daling JR, Starzyk P, Olshan AF, et al. 1984. Birth weight and the incidence of childhood cancer. J Natl Cancer Inst 72(5):1039–1041. Daniels JL, Olshan AF, Pollock BH, et al. 2002. Breast-feeding and neuroblastoma, USA and Canada. Cancer Causes Control 13(5):401–405. Daniels JL, Olshan AF, Teschke K, et al. 2001. Residential pesticide exposure and neuroblastoma. Epidemiology 12(1):20–27. Davies SM, Robison LL, Buckley JD, et al. 2000. Glutathione S-transferase polymorphisms in children with myeloid leukemia: a Children’s Cancer Group study. Cancer Epidemiol Biomarkers Prev 9(6):563–566. Davis MK, Savitz DA, Graubard BI. 1988. Infant feeding and childhood cancer. Lancet 2(8607):365–368. DeBaun MR, Tucker MA. 1998. Risk of cancer during the first four years of life in children from The Beckwith-Wiedemann Syndrome Registry [comment]. J Pediatr 132(3 Pt 1):398–400. Depue RH, Pike MC, Henderson BE. 1983. Estrogen exposure during gestation and risk of testicular cancer. J Natl Cancer Inst 71(6):1151–1155. DerKinderen DJ, Koten JW, Tan KE, et al. 1990. parental age in sporadic hereditary retinoblastoma. Am J Ophthalmol 110(6):605–609. Dickinson HO, Hammal DM, Bithell JF, et al. 2002. Population mixing and childhood leukaemia and non-Hodgkin’s lymphoma in census wards in England and Wales, 1966–87. Br J Cancer 86(9):1411–1413. Dickinson HO, Parker L. 1999. Quantifying the effect of population mixing on childhood leukaemia risk: The Seascale cluster. Br J Cancer 81(1): 144–151. Dockerty JD, Draper G, Vincent T, et al. 2001. Case-control study of parental age, parity and socioeconomic level in relation to childhood cancers. Int J Epidemiol 30(6):1428–1437. Dockerty JD, Draper G, Vincent T, et al. 2001. Case-control study of parental age, parity and socioeconomic level in relation to childhood cancers [comment]. Int J Epidemiol 30(6):1428–1437. Dockerty JD, Skegg DC, Elwood JM, et al. 1999. Infections, vaccinations, and the risk of childhood leukaemia. Br J Cancer 80(9):1483–1489. Doll R, Wakeford R. 1997. Risk of childhood cancer from fetal irradiation. Br J Radiol 70:130–139. Dome JS, Coppes MJ. 2002. Recent advances in Wilms tumor genetics. Curr Opin Pediatr 14(1):5–11. Dryja TP, Mukai S, Petersen R, et al. 1989. Parental origin of mutations of the retinoblastoma gene. Nature 339(6225):556–558. Dryja TP, Rapaport JM, Joyce JM, et al. 1986. Molecular detection of deletions involving band q14 of chromosome 13 in retinoblastomas. Proc Natl Acad Sci U S A 83(19):7391–7394. Dworsky M, Yow M, Stagno S, et al. 1983. Cytomegalovirus infection of breast milk and transmission in infancy. Pediatrics 72(3):295–299. Efird JT, Holly EA, Preston-Martin S, et al. 2003. Farm-related exposures and childhood brain tumours in seven countries: Results from the SEARCH

Cancers in Children International Brain Tumour Study. Paediatr Perinat Epidemiol 17(2): 201–211. Engels EA, Katki HA, Nielsen NM, et al. 2003. Cancer incidence in Denmark following exposures to poliovirus vaccine contaminated with simian virus 40. J Natl Cancer Inst 95(7):532–539. Fabia J, Thuy TD. 1974. Occupation of father at time of birth of children dying of malignant diseases. Br J Prev Soc Med 28(2):98–100. Fasching K, Panzer S, Haas OA, et al. 2000. Presence of clone-specific antigen receptor gene rearrangements at birth indicates an in utero origin of diverse types of early childhood acute lymphoblastic leukemia. Blood 95(8):2722–2724. Ferrari S, Bertoni F, Mercuri M, et al. 2001. Predictive factors of disease-free survival for non-metastatic osteosarcoma of the extremity: An analysis of 300 patients treated at the Rizzoli Institute. Ann Oncol 12(8):1145– 1150. Feusner J, Plaschkes J. 2002. Hepatoblastoma and low birth weight: A trend or chance observation? Med Pediatr Oncol 39(5):508–509. Filipovich AH, Mathur A, Kamat D, et al. 1992. Primary immunodeficiencies: Genetic risk factors for lymphoma. Cancer Res 52(19 Suppl):5465s– 5467s. Filippini G, Farinotti M, Lovicu G, et al. 1994. Mothers’ active and passive smoking during pregnancy and risk of brain tumours in children. Int J Cancer 57(6):769–774. Filippini G, Maisonneuve P, McCredie M, et al. 2002. Relation of childhood brain tumors to exposure of parents and children to tobacco smoke: The SEARCH international case-control study. Surveillance of Environmental Aspects Related to Cancer in Humans. Int J Cancer 100(2):206–213. Finkelstein MM. 1994. Radium in drinking water and the risk of death from bone cancer among Ontario youths. CMAJ 151(5):565–571. Finkelstein MM, Kreiger N. 1996. Radium in drinking water and risk of bone cancer in Ontario youths: A second study and combined analysis. Occup Environ Med 53(5):305–311. Fisher SG, Weber L, Carbone M. 1999. Cancer risk associated with simian virus 40 contaminated polio vaccine. Anticancer Res 19(3B):2173–2180. Ford AM, Ridge SA, Cabrera ME, et al. 1993. In utero rearrangements in the trithorax-related oncogene in infant leukaemias. Nature 363(6427):358– 360. Fraumeni JF, Jr. 1967. Stature and malignant tumors of bone in childhood and adolescence. Cancer 20(6):967–973. Fraumeni JF, Jr, Miller RW. 1967. Epidemiology of human leukemia: Recent observations. J Natl Cancer Inst 38(4):593–605. Freedman DM, Stewart P, Kleinerman RA, et al. 2001. Household solvent exposures and childhood acute lymphoblastic leukemia. Am J Public Health 91(4):564–567. Front C-T, Brodeur GM. 1987. Down’s syndrome and leukemia: Epidemiology, genetics, cytogenetics, and the mechanisms of leukemogenesis. Cancer Genet Cytogenet 28:55–76. Gahrton G, Wahren B, Killander D, et al. 1971. Epstein-Barr and other herpes virus antibodies in children with acute leukemia. Int J Cancer 8(2): 242–249. Gale KB, Ford AM, Repp R, et al. 1997. Backtracking leukemia to birth: Identification of clonotypic gene fusion sequences in neonatal blood spots. Proc Nat Acad Sci U S A 94(25):13950–13954. Gardner MJ. 1991. Father’s occupational exposure to radiation and the raised level of childhood leukemia near the Sellafield nuclear plant. Environ Health Perspect 94:5–7. Gelberg KH, Fitzgerald EF, Hwang S, et al. 1997. Growth and development and other risk factors for osteosarcoma in children and young adults. Int J Epidemiol 26(2):272–278. Gelberg KH, Fitzgerald EF, Hwang SA, et al. 1995. Fluoride exposure and childhood osteosarcoma: A case-control study. Am J Public Health 85(12):1678–1683. German J, Bloom D, Passarge E. 1979. Bloom’s syndrome. VII. Progress report for 1978. Clin Genet 15(4):361–367. Giardiello FM, Offerhaus GJ, Krush AJ, et al. 1991. Risk of hepatoblastoma in familial adenomatous polyposis. J Pediatr 119(5):766-768. Gibson BE, Eden OB, Barrett A, et al. 1988. Leukaemia in young children in Scotland. Lancet 2(8611):630. Glass AG, Fraumeni JF, Jr. 1970. Epidemiology of bone cancer in children. J Natl Cancer Inst 44(1):187–199. Gold E, Gordis L, Tonascia J, et al. 1979. Risk factors for brain tumors in children. Am J Epidemiol 109(3):309–319. Gold EB, Leviton A, Lopez R, et al. 1993. Parental smoking and risk of childhood brain tumors. Am J Epidemiol 137(6):620–628. Granovsky MO, Mueller BU, Nicholson HS, et al. 1998. Cancer in human immunodeficiency virus-infected children: A case series from the Children’s Cancer Group and the National Cancer Institute. J Clin Oncol 16(5):1729–1735.

1263

Greaves M. 1999. Molecular genetics, natural history and the demise of childhood leukaemia. Eur J Cancer 35(2):173–185. Greaves M. 2002. Childhood leukaemia. BMJ 324(7332):283–287. Greaves MF. 1988. Speculations on the cause of childhood acute lymphoblastic leukemia. Leukemia 2(2):120–125. Greaves MF, Colman SM, Beard ME, et al. 1993. Geographical distribution of acute lymphoblastic leukaemia subtypes: Second report of the collaborative group study. Leukemia 7(1):27–34. Greaves MF, Pegram SM, Chan LC. 1985. Collaborative group study of the epidemiology of acute lymphoblastic leukaemia subtypes: background and first report. Leuk Res 9(6):715–733. Greenland S, Sheppard AR, Kaune WT, et al. 2000. A pooled analysis of magnetic fields, wire codes, and childhood leukemia. Childhood LeukemiaEMF Study Group. Epidemiology 11(6):624–634. Gripp KW, Scott CI, Jr, Nicholson L, et al. 2002. Five additional Costello syndrome patients with rhabdomyosarcoma: Proposal for a tumor screening protocol. Am J Med Genet 108(1):80–87. Groves FD, Sinha D, Kayhty H, et al. 2001. Haemophilus influenzae type b serology in childhood leukaemia: A case-control study. Br J Cancer 85(3):337–340. Grufferman S, Ambinder RF, Shugart YY, Brecher ML, Gilchrist GS. 1998. Increased cancer risk in families of children with Hodgkin’s disease. Am J Epidemiol 147(11):S8. Grufferman S, Davis MK, Ambinder RF, Shugart YY, Gilchrist GS, Brecher ML. 1998. A protective effect of breast-feeding on risk of Hodgkin’s disease in children. Paediatr Perinat Epidemiol 12:A13. Grufferman S, Delzell E. 1984. Epidemiology of Hodgkin’s disease. Epidemiol Rev 6:76–106. Grufferman S, Gula MJ, Olshan AF, Falletta JM, Pendergrass TW, Buckley J, Maurer HM. 1991. In utero x-ray exposure and risk of childhood rhabdomyosarcoma. Paediatr Perinat Epidemiol 5:A6. Grufferman S, Schwartz AG, Ruymann FB, et al. 1993. Parents’ use of cocaine and marijuana and increased risk of rhabdomyosarcoma in their children. Cancer Causes Control 4(3):217–224. Grufferman S, Wang HH, DeLong ER, et al. 1982. Environmental factors in the etiology of rhabdomyosarcoma in childhood. J Natl Cancer Inst 68(1):107–113. Gurney JG, Chen M, Skluzacek MC, et al. 2002. Null association between frequency of cured meat consumption and methylvaline and ethylvaline hemoglobin adduct levels: The N-nitroso brain cancer hypothesis. Cancer Epidemiol Biomarkers Prev 11(4):421–422. Gurney JG, Davis S, Severson RK, et al. 1996. Trends in cancer incidence among children in the U.S. Cancer 78(3):532–541. Gurney JG, Mueller BA, Davis S, et al. 1996. Childhood brain tumor occurrence in relation to residential power line configurations, electric heating sources, and electric appliance use[comment]. Am J Epidemiol 143(2):120–128. Gurney JG, Mueller BA, Preston-Martin S, et al. 1997. A study of pediatric brain tumors and their association with epilepsy and anticonvulsant use. Neuroepidemiology 16(5):248–255. Gurney JG, Pogoda JM, Holly EA, et al. 1997. Aspartame consumption in relation to childhood brain tumor risk: Results from a case-control study. J Natl Cancer Inst 89(14):1072–1074. Gurney JG, Severson RK, Davis S, et al. 1995. Incidence of cancer in children in the United States. Sex-, race-, and 1-year age-specific rates by histologic type. Cancer 75(8):2186–2195. Gurney JG, Smith MA, Olshan AF, et al. 2001. Clues to the etiology of childhood brain cancer: N-nitroso compounds, polyomaviruses, and other factors of interest. Cancer Invest 19(6):630–640. Gurney JG, Wall DA, Jukich PJ, et al. 1999. The contribution of nonmalignant tumors to CNS tumor incidence rates among children in the United States. Cancer Causes Control 10(2):101–105. Gutensohn N, Cole P. 1981. Childhood social environment and Hodgkin’s disease. N Engl J Med 304(3):135–140. Gutensohn NM. 1982. Social class and age at diagnosis of Hodgkin’s disease: New epidemiologic evidence for the “two-disease hypothesis.” Cancer Treat Rep 66(4):689–695. Gutensohn NM, Shapiro DS. 1982. Social class risk factors among children with Hodgkin’s disease. Int J Cancer 30(4):433–435. Hamrick SE, Olshan AF, Neglia JP, et al. 2001. Association of pregnancy history and birth characteristics with neuroblastoma: A report from the Children’s Cancer Group and the Pediatric Oncology Group. Paediatr Perinat Epidemiol 15(4):328–337. Hansen MF, Koufos A, Gallie BL, et al. 1985. Osteosarcoma and retinoblastoma: A shared chromosomal mechanism revealing recessive predisposition. Proc Natl Acad Sci U S A 82(18):6216–6220. Hartley AL, Birch JM, Blair V, et al. 1991. Cancer incidence in the families of children with Ewing’s tumor. J Natl Cancer Inst 83(13):955–956.

1264

PART IV: CANCER BY TISSUE OF ORIGIN

Hartley AL, Birch JM, McKinney PA, et al. 1988. The Inter-Regional Epidemiological Study of Childhood Cancer (IRESCC): Case control study of children with bone and soft tissue sarcomas. Br J Cancer 58(6): 838–842. Harvey EB, Boice JD, Jr, Honeyman M, et al. 1985. Prenatal x-ray exposure and childhood cancer in twins. N Engl J Med 312(9):541–545. Hecht F, Hecht BK. 1990. Cancer in ataxia-telangiectasia patients. Cancer Genet Cytogenet 46(1):9–19. Heegaard ED, Jensen L, Hornsleth A, et al. 1999. The role of parvovirus B19 infection in childhood acute lymphoblastic leukemia. Pediatr Hematol Oncol 16(4):329–334. Hemminki K, Kyyronen P, Vaittinen P. 1999. Parental age as a risk factor of childhood leukemia and brain cancer in offspring. Epidemiology 10(3):271–275. Hemminki K, Li X. 2002. Cancer risks in childhood and adolescence among the offspring of immigrants to Sweden. Br J Cancer 86(9):1414–1418. Henderson BE, Benton B, Jing J, et al. 1979. Risk factors for cancer of the testis in young men. Int J Cancer 23(5):598–602. Henshaw DL, Eatough JP, Richardson RB. 1990. Radon as a causative factor in induction of myeloid leukaemia and other cancers. Lancet 335(8696): 1008–1012. Heuch JM, Heuch I, Kvale G. 1996. Birth characteristics and risk of Wilms’ tumour: A nationwide prospective study in Norway. Br J Cancer 74(7):1148–1151. Hjalmars U, Kulldorff M, Gustafsson G. 1994. Risk of acute childhood leukaemia in Sweden after the Chernobyl reactor accident. Swedish Child Leukaemia Group. BMJ 309(6948):154–157. Holly EA, Aston DA, Ahn DK, et al. 1992. Ewing’s bone sarcoma, paternal occupational exposure, and other factors. Am J Epidemiol 135(2):122– 129. Holly EA, Bracci PM, Hong MK, et al. 2002. West Coast study of childhood brain tumours and maternal use of hair-colouring products. Paediatr Perinat Epidemiol 16(3):226–235. Holly EA, Bracci PM, Mueller BA, et al. 1998. Farm and animal exposures and pediatric brain tumors: Results from the United States West Coast Childhood Brain Tumor Study. Cancer Epidemiol Biomarkers Prev 7(9):797–802. Horta BL, Victora CG, Menezes AM, et al. 1997. Low birthweight, preterm births and intrauterine growth retardation in relation to maternal smoking. Paediatr Perinat Epidemiol 11(2):140–151. Howe GR, Burch JD, Chiarelli AM, et al. 1989. An exploratory case-control study of brain tumors in children. Cancer Res 49(15):4349–4352. Hrusak O, Trka J, Zuna J, et al. 2002. Acute lymphoblastic leukemia incidence during socioeconomic transition: Selective increase in children from 1 to 4 years. Leukemia 16(4):720–725. Hu J, Mao Y, Ugnat AM. 2000. Parental cigarette smoking, hard liquor consumption and the risk of childhood brain tumors—a case-control study in northeast China. Acta Oncologica 39(8):979–984. Hum L, Kreiger N, Finkelstein MM. 1998. The relationship between parental occupation and bone cancer risk in offspring. Int J Epidemiol 27(5):766–771. Ikeda H, Matsuyama S, Tanimura M. 1997. Association between hepatoblastoma and very low birth weight: A trend or a chance?[comment]. J Pediatr 130(4):557–560. Infante-Rivard C, Amre D, Sinnett D. 2002. GSTT1 and CYP2E1 polymorphisms and trihalomethanes in drinking water: Effect on childhood leukemia. Environ Health Perspect 110(6):591–593. Infante-Rivard C, Fortier I, Olson E. 2000. Markers of infection, breast-feeding and childhood acute lymphoblastic leukaemia. Br J Cancer 83(11): 1559–1564. Infante-Rivard C, Krajinovic M, Labuda D, et al. 2000. Parental smoking, CYP1A1 genetic polymorphisms and childhood leukemia (Quebec, Canada). Cancer Causes Control 11(6):547–553. Infante-Rivard C, Labuda D, Krajinovic M, et al. 1999. Risk of childhood leukemia associated with exposure to pesticides and with gene polymorphisms[comment]. Epidemiology 10(5):481–487. Infante-Rivard C, Mathonnet G, Sinnett D. 2000. Risk of childhood leukemia associated with diagnostic irradiation and polymorphisms in DNA repair genes. Environ Health Perspect 108(6):495–498. Investigators UKCCS. 2002. The United Kingdom Childhood Cancer Study of exposure to domestic sources of ionising radiation: 1: radon gas. Br J Cancer 86(11):1721–1726. Ivankovic S. 1979. Teratogenic and carcinogenic effects of some chemicals during perinatal life in rats, Syrian golden hamsters, and minipigs. Natl Cancer Inst Monogr (51):103–115. Ivankovic S, Preussman R, Schmal D, et al. 1975. Prevention by ascorbic acid of in vivio formation of N-Nitroso compounds. Lyon, France: International Agency for Research on Cancer.

Ivanov EP, Tolochko G, Lazarev VS, et al. 1993. Child leukaemia after Chernobyl. Nature 365(6448):702. Jarrett R, Onions D. 1992. Viruses and Hodgkin’s disease. Leukemia 6 Suppl 1:14–17. Jemal A, Murray T, Samuels A, et al. 2003. Cancer Statistics, 2003. CA Cancer J Clin 53(5):5–26. Johnson CC, Spitz MR. 1985. Neuroblastoma: Case-control analysis of birth characteristics. J Natl Cancer Inst 74(4):789–792. Kaletsch U, Kaatsch P, Meinert R, et al. 1999. Childhood cancer and residential radon exposure—results of a population-based case-control study in Lower Saxony (Germany). Radia Environ Biophys 38(3):211–215. Kantor AF, Curnen MG, Meigs JW, et al. 1979. Occupations of fathers of patients with Wilms’s tumour. J Epidemiol Community Health 33(4): 253–256. Kardaun JW, Hayes RB, Pottern LM, et al. 1991. Testicular cancer in young men and parental occupational exposure. Am J Ind Med 20(2):219–227. Kasai Y, Takeda S, Takagi H. 1996. Pathogenesis of hepatocellular carcinoma: A review from the viewpoint of molecular analysis. Semin Surg Oncol 12(3):155–159. Kaye SA, Robison LL, Smithson WA, et al. 1991. Maternal reproductive history and birth characteristics in childhood acute lymphoblastic leukemia. Cancer 68(6):1351–1355. Kerr MA, Nasca PC, Mundt KA, et al. 2000. Parental occupational exposures and risk of neuroblastoma: A case-control study (United States). Cancer Causes Control 11(7):635–643. Kinlen L. 1988. Evidence for an infective cause of childhood leukaemia: comparison of a Scottish new town with nuclear reprocessing sites in Britain[comment]. Lancet 2(8624):1323–1327. Kinlen LJ. 1995. Epidemiological evidence for an infective basis in childhood leukaemia. Br J Cancer 71(1):1–5. Kinlen LJ, Dickson M, Stiller CA. 1995. Childhood leukaemia and nonHodgkin’s lymphoma near large rural construction sites, with a comparison with Sellafield nuclear site. BMJ 310(6982):763–768. Kirchstein R, Gerber P. 1962. Ependymomas produced after intracerebral inoculation of SV40 into newborn hamsters. Dev Biol Stand 195:299–300. Kleihues P, Burger PC, Scheithauer BW. 1993. The new WHO classification of brain tumours. Brain Pathol 3(3):255–268. Knudson AG, Jr. 1971. Mutation and cancer: statistical study of retinoblastoma. Proc Natl Acad Sci U S A 68(4):820–823. Knudson AG, Jr, Strong LC. 1972. Mutation and cancer: A model for Wilms’ tumor of the kidney. J Nat Cancer Inst 48(2):313–324. Koushik A, King WD, McLaughlin JR. 2001. An ecologic study of childhood leukemia and population mixing in Ontario, Canada. Cancer Causes Control 12(6):483–490. Krajinovic M, Sinnett H, Richer C, et al. 2002. Role of NQO1, MPO and CYP2E1 genetic polymorphisms in the susceptibility to childhood acute lymphoblastic leukemia. Int J Cancer 97(2):230–236. Kramarova E, Stiller CA. 1996. The international classification of childhood cancer. Int J Cancer 68(6):759–765. Kramer S, Ward E, Meadows AT, et al. 1987. Medical and drug risk factors associated with neuroblastoma: A case-control study. J Natl Cancer Inst 78(5):797–804. Kristensen P, Andersen A, Irgens LM, et al. 1996. Cancer in offspring of parents engaged in agricultural activities in Norway: Incidence and risk factors in the farm environment. Int J Cancer 65(1):39–50. Kuijten RR, Bunin GR, Nass CC, et al. 1990. Gestational and familial risk factors for childhood astrocytoma: Results of a case-control study. Cancer Res 50(9):2608–2612. Kuijten RR, Strom SS, Rorke LB, et al. 1993. Family history of cancer and seizures in young children with brain tumors: A report from the Childrens Cancer Group (United States and Canada). Cancer Causes Control 4(5):455–464. Kushner BH, Gilbert F, Helson L. 1986. Familial neuroblastoma. Case reports, literature review, and etiologic considerations. Cancer 57(9):1887–1893. Ladanyi M, Lewis R, Garin-Chesa P, et al. 1993. EWS rearrangement in Ewing’s sarcoma and peripheral neuroectodermal tumor. Molecular detection and correlation with cytogenetic analysis and MIC2 expression. Diagn Mol Pathol 2(3):141–146. Lancashire RJ, Sorahan T. 2003. Breastfeeding and childhood cancer risks: OSCC data. Br J Cancer 88(7):1035–1037. Langford I. 1991. Childhood leukaemia mortality and population change in England and Wales 1969–73. Soc Sci Med 33(4):435–440. Larsson SE, Lorentzon R. 1974. The incidence of malignant primary bone tumours in relation to age, sex and site. A study of osteogenic sarcoma, chondrosarcoma and Ewing’s sarcoma diagnosed in Sweden from 1958 to 1968. J Bone Joint Surg [Br] 56B(3):534–540. Leaf CD, Wishnok JS, Tannenbaum SR. 1989. Mechanisms of endogenous nitrosation. Cancer Surv 8(2):323–334.

Cancers in Children Lee CL, Hsieh KS, Ko YC. 2003. Trends in the incidence of hepatocellular carcinoma in boys and girls in Taiwan after large-scale hepatitis B vaccination. Cancer Epidemiol Biomarkers Prev 12:57–59. Lei U, Wohlfahrt J, Hjalgrim H, et al. 2000. Neonatal level of thyroidstimulating hormone and acute childhood leukemia. Int J Cancer 88(3): 486–488. Leisenring WM, Breslow NE, Evans IE, et al. 1994. Increased birth weights of National Wilms’ Tumor Study patients suggest a growth factor excess. Cancer Res 54(17):4680–4683. Li FP, Fraumeni JF. 1972. Testicular cancers in children: Epidemiologic characteristics. J Natl Cancer Inst 48(6):1575–1581. Li FP, Fraumeni JF, Jr., Dalager N. 1973. Ovarian cancers in the young. Epidemiologic observations. Cancer 32(4):969–972. Lijinsky W. 1999. N-Nitroso compounds in the diet. Mutat Res 443(1–2): 129–138. Lindblad P, Zack M, Adami HO, et al. 1992. Maternal and perinatal risk factors for Wilms’ tumor: A nationwide nested case-control study in Sweden. Int J Cancer 51:38–41. Linet M. 1985. The Leukemias: Epidemiological Aspects. New York: Oxford University Press. Linet MS, Hatch EE, Kleinerman RA, et al. 1997. Residential exposure to magnetic fields and acute lymphoblastic leukemia in children. New Engl J Med 337(1):1–7. Linos A, Kyle RA, Elveback LR, et al. 1978. Leukemia in Olmsted County, Minnesota, 1965–1974. Mayo Clin Proc 53(11):714–718. Lowengart RA, Peters JM, Cicioni C, et al. 1987. Childhood leukemia and parents’ occupational and home exposures. J Nat Cancer Inst 79(1): 39–46. Lubin JH, Linet MS, Boice JD, Jr., et al. 1998. Case-control study of childhood acute lymphoblastic leukemia and residential radon exposure. J Nat Cancer Inst 90(4):294–300. Ma X, Buffler PA, Gunier RB, et al. 2002a. Critical windows of exposure to household pesticides and risk of childhood leukemia. Environ Health Perspect 110(9):955–960. Ma X, Buffler PA, Selvin S, et al. 2002b. Daycare attendance and risk of childhood acute lymphoblastic leukaemia. Br J Cancer 86(9):1419–1424. Mack TM, Cozen W, Shibata DK, et al. 1995. Concordance for Hodgkin’s disease in identical twins suggesting genetic susceptibility to the youngadult form of the disease. N Engl J Med 332(7):413–418. MacKenzie J, Gallagher A, Clayton RA, et al. 2001. Screening for herpesvirus genomes in common acute lymphoblastic leukemia. Leukemia 15(3): 415–421. MacKenzie J, Perry J, Ford AM, et al. 1999. JC and BK virus sequences are not detectable in leukaemic samples from children with common acute lymphoblastic leukaemia. Br J Cancer 81(5):898–899. MacMahon B. 1966. Epidemiology of Hodgkin’s disease. Cancer Res 26(6): 1189–1201. Magnani C, Pastore G, Luzzatto L, et al. 1989. Risk factors for soft tissue sarcomas in childhood: A case-control study. Tumori 75(4):396–400. Maruyama K, Ikeda H, Koizumi T, et al. 2000. Case-control study of perinatal factors and hepatoblastoma in children with an extremely low birthweight. Pediatr Int 42(5):492–498. McBride ML, Gallagher RP, Theriault G, et al. 1999. Power-frequency electric and magnetic fields and risk of childhood leukemia in Canada. Am J Epidemiol 149(9):831–842. McCredie M, Little J, Cotton S, et al. 1999. SEARCH international case-control study of childhood brain tumours: Role of index pregnancy and birth, and mother’s reproductive history. Paediatr Perinat Epidemiol 13(3): 325–341. McCredie M, Maisonneuve P, Boyle P. 1994. Antenatal risk factors for malignant brain tumours in New South Wales children. Int J Cancer 56(1): 6–10. McCredie M, Maisonneuve P, Boyle P. 1994. Perinatal and early postnatal risk factors for malignant brain tumours in New South Wales children. Int J Cancer 56(1):11–15. McGuire SM, Vanable ED, McGuire MH, et al. 1991. Is there a link between fluoridated water and osteosarcoma? J Am Dental Assoc 122(4):38– 45. McKean-Cowdin R, Pogoda JM, Lijinsky W, et al. 2003. Maternal prenatal exposure to nitrosatable drugs and childhood brain tumours. Int J Epidemiol 32(2):211–217. McKean-Cowdin R, Preston-Martin S, Pogoda JM, et al. 1998. Parental occupation and childhood brain tumors: Astroglial and primitive neuroectodermal tumors. J Occup Environ Med 40(4):332–340. McKinney PA, Cartwright RA, Saiu JM, et al. 1987. The interregional epidemiological study of childhood cancer (IRESCC): A case control study of aetiological factors in leukaemia and lymphoma. Arch Dis Childhood 62(3):279–287.

1265

McKinney PA, Juszczak E, Findlay E, et al. 1999. Pre- and perinatal risk factors for childhood leukaemia and other malignancies: A Scottish case control study. Br J Cancer 80(11):1844–1851. McLaughlin JR, Clarke EA, Nishri ED, et al. 1993. Childhood leukemia in the vicinity of Canadian nuclear facilities. Cancer Causes Control 4(1): 51–58. McNeil DE, Brown M, Ching A, et al. 2001. Screening for Wilms tumor and hepatoblastoma in children with Beckwith-Wiedemann syndromes: A cost-effective model. Med Pediatr Oncol 37(4):349–356. Michaelis J, Kaletsch U, Burkart W, et al. 1997. Infant leukaemia after the Chernobyl accident. Nature 387(6630):246. Michalek AM, Buck GM, Nasca PC, et al. 1996. Gravid health status, medication use, and risk of neuroblastoma. Am J Epidemiol 143(10): 996–1001. Minino AM, Arias E, Kochanek KD, et al. 2002. Deaths: Final Data for 2000. National Vital Statistics Reports; vol. 50 no. 15. Hyattsville, Maryland: National Center for Health Statistics. Mole RH. 1990. Childhood cancer after prenatal exposure to diagnostic X-ray examinations in Britain. Br J Cancer 62(1):152–168. Moll AC, Imhof SM, Cruysberg JR, et al. 2003. Incidence of retinoblastoma in children born after in-vitro fertilisation. Lancet 361(9354):309–310. Moll AC, Imhof SM, Kuik DJ, et al. 1996. High parental age is associated with sporadic hereditary retinoblastoma: The Dutch retinoblastoma register 1862–1994. Hum Genet 98(1):109–112. Mori H, Colman SM, Xiao Z, et al. 2002. Chromosome translocations and covert leukemic clones are generated during normal fetal development. Proc Nat Acad Sci USA 99(12):8242–8247. Moss ME, Kanarek MS, Anderson HA, et al. 1995. Osteosarcoma, seasonality, and environmental factors in Wisconsin, 1979–1989. Arch Environ Health 50(3):235–241. Mullvihill JJ. 1975. Congenital and genetic diseases. In: Fraumeni JF, Jr., ed. Persons at High Risk of Cancer. San Diego, CA; Academic Press: pp. 3–31. Narod SA, Stiller C, Lenoir GM. 1991. An estimate of the heritable fraction of childhood cancer. Br J Cancer 63(6):993–999. Nasca PC, Baptiste MS, MacCubbin PA, et al. 1988. An epidemiologic casecontrol study of central nervous system tumors in children and parental occupational exposures. Am J Epidemiol 128(6):1256–1265. Naumburg E, Bellocco R, Cnattingius S, et al. 2002. Perinatal exposure to infection and risk of childhood leukemia. Med Pediatr Oncol 38(6):391–397. Neglia JP, Linet MS, Shu XO, et al. 2000. Patterns of infection and day care utilization and risk of childhood acute lymphoblastic leukaemia. Br J Cancer 82(1):234–240. Neglia JP, Smithson WA, Gunderson P, et al. 1988. Prenatal and perinatal risk factors for neuroblastoma. A case-control study. Cancer 61(11): 2202–2206. Newton WA, Jr., Meadows AT, Shimada H, et al. 1991. Bone sarcomas as second malignant neoplasms following childhood cancer. Cancer 67(1):193–201. Norman MA, Holly EA, Ahn DK, et al. 1996a. Prenatal exposure to tobacco smoke and childhood brain tumors: Results from the United States West Coast childhood brain tumor study. Cancer Epidemiol Biomarkers Prev 5(2):127–133. Norman MA, Holly EA, Preston-Martin S. 1996b. Childhood brain tumors and exposure to tobacco smoke. Cancer Epidemiol Biomarkers Prev 5(2):85–91. Novakovic B, Goldstein AM, Wexler LH, et al. 1994. Increased risk of neuroectodermal tumors and stomach cancer in relatives of patients with Ewing’s sarcoma family of tumors. J Natl Cancer Inst 86(22):1702–1706. Oda H, Imai Y, Nakatsuru Y, et al. 1996. Somatic mutations of the APC gene in sporadic hepatoblastomas. Cancer Res 56(14):3320–3323. Ogawa O, Becroft DM, Morison IM, et al. 1993. Constitutional relaxation of insulin-like growth factor II gene imprinting associated with Wilms’ tumour and gigantism. Nat Genet 5(4):408–412. Olin P, Giesecke J. 1998. Potential exposure to SV40 in polio vaccines used in Sweden during 1957: No impact on cancer incidence rates 1960 to 1993. Dev Biol Stand 94:227–233. Olshan AF, Breslow NE, Daling JR, et al. 1990. Wilms’ tumor and paternal occupation. Cancer Res 50(11):3212–3217. Olshan AF, Breslow NE, Daling JR, et al. 1986. Childhood brain tumors and paternal occupation in the aerospace industry. J Nat Cancer Inst 77(1):17–19. Olshan AF, Breslow NE, Falletta JM, et al. 1993. Risk factors for Wilms tumor. Report from the National Wilms Tumor Study. Cancer 72(3):938– 944. Olshan AF, De Roos AJ, Teschke K, et al. 1999a. Neuroblastoma and parental occupation. Cancer Causes Control 10(6):539–549.

1266

PART IV: CANCER BY TISSUE OF ORIGIN

Olshan AF, Faustman EM. 1989. Nitrosatable drug exposure during pregnancy and adverse pregnancy outcome. Int J Epidemiol 18(4):891–899. Olshan AF, Smith J, Cook MN, et al. 1999b. Hormone and fertility drug use and the risk of neuroblastoma: A report from the Children’s Cancer Group and the Pediatric Oncology Group. Am J Epidemiol 150(9):930–938. Olshan AF, Smith JC, Bondy ML, et al. 2002. Maternal vitamin use and reduced risk of neuroblastoma. Epidemiology 13(5):575–580. Operskalski EA, Preston-Martin S, Henderson BE, et al. 1987. A case-control study of osteosarcoma in young persons. Am J Epidemiol 126(1): 118–126. Pang D, McNally R, Birch JM. 2003. Parental smoking and childhood cancer: Results from the United Kingdom Childhood Cancer Study. Br J Cancer 88:373–381. Parker L. 2001. Breast-feeding and cancer prevention. Eur J Cancer 37(2):155–158. Parker L, Craft AW, Smith J, et al. 1993. Geographical distribution of preconceptional radiation doses to fathers employed at the Sellafield nuclear installation, West Cumbria. BMJ 307(6910):966–971. Parkin DM, Masuyer E CE, et al. 1993. Childhood leukemia following the Chernobyl accident: the European Childhood Leukemia-Lymphoma Incidence Study (ECLIS). Eur J Cancer 29A:87–89. Parkin DM, Kramarova E, Draper G, et al. 1998. International Incidence of Childhood Cancer, Vol. II. Carey. NC: International Agency for Research on Cancer. Parslow RC, Law GR, Feltbower R, et al. 2002. Population mixing, childhood leukaemia, CNS tumours and other childhood cancers in Yorkshire. Eur J Cancer 38(15):2033–2040. Penn I. 1994. De novo tumors in pediatric organ transplant recipients. Transplant Proc 26(1):1–2. Perkins D, Brennan S, Carstairs K, et al. 1997. Regional cancer cytogenetics: A report on 1,143 diagnostic cases. Cancer Genet Cytogenet 96(1):64–80. Perrillat F, Clavel J, Auclerc MF, et al. 2002. Day-care, early common infections and childhood acute leukaemia: A multicentre French case-control study. Br J Cancer 86(7):1064–1069. Petridou E, Proukakis C, Tong D, et al. 1994. Trends and geographical distribution of childhood leukemia in Greece in relation to the Chernobyl accident. Scand J Soc Med 22(2):127–131. Petridou E, Skalkidou A, Dessypris N, et al. 2000. Endogenous risk factors for childhood leukemia in relation to the IGF system (Greece). The Childhood Haematologists-Oncologists Group. Cancer Causes Control 11(8): 765–771. Petridou E, Trichopoulos D, Dessypris N, et al. 1996. Infant leukaemia after in utero exposure to radiation from Chernobyl. Nature 382(6589): 352–353. Preston-Martin S, Gurney JG, Pogoda JM, et al. 1996a. Brain tumor risk in children in relation to use of electric blankets and water bed heaters. Results from the United States West Coast Childhood Brain Tumor Study. Am J Epidemiol 143(11):1116–1122. Preston-Martin S, Navidi W, Thomas D, et al. 1996b. Los Angeles study of residential magnetic fields and childhood brain tumors [comment]. Am J Epidemiol 143(2):105–119. Preston-Martin S, Pogoda JM, Mueller BA, et al. 1998a. Prenatal vitamin supplementation and risk of childhood brain tumors. Int J – Suppl 11:17–22. Preston-Martin S, Pogoda JM, Mueller BA, et al. 1998b. Results from an international case-control study of childhood brain tumors: the role of prenatal vitamin supplementation. Environ Health Perspect 106(Suppl 3):887–892. Preston-Martin S, Yu, MC, Benton B, et al. 1982. N-Nitroso compounds and childhood brain tumors: A case-control study. Cancer Res 42(12): 5240–5245. Preussman R, Stewart BW. 1984. N-Nitroso carcinogens. In: Searle C, ed. Washington, D.C.: American Chemical Society 2:643–828. Rainier S, Johnson LA, Dobry CJ, et al. 1993. Relaxation of imprinted genes in human cancer. Nature 362(6422):747–749. Reynolds P, Von Behren J, Elkin EP. 2002. Birth characteristics and leukemia in young children. Am J Epidemiol 155(7):603–613. Richkind KE, Loew T, Meisner L, et al. 1998. Identical cytogenetic clones and clonal evolution in pediatric monozygotic twins with acute myeloid leukemia: Presymptomatic disease detection by interphase fluorescence in situ hybridization and review of the literature. J Pediatr Hematol/Oncol 20(3):264–267. Ries LA, Eisner MP, Kosary CL, et al. 2002. SEER Cancer Statistics Review, 1973–1999. Bethesda, MD: National Cancer Institute. Ries LA, Smith MA, Gurney JG, et al. 1999. Cancer Incidence and Survival among Children and Adolescents: United States SEER Program 1975–1995. Bethesda MD: National Cancer Institute, SEER Program. Ries LAG, Eisner MP, Kosary CL, et al. 2003. SEER Cancer Statistics Review, 1975–2000. Bethesda, MD: National Cancer Institute.

Ries LAG, Eisner MP, Kosary CL, Hankey BF, Miller BA, Clegg L, Mariotto A, Fay MP, Feuer EJ, Edwards BK, eds. 2003. SEER Cancer Statistics Review, 1975–2000. Bethesda, MD: National Cancer Institute. Robison LL, Neglia JP. 1987. Epidemiology of Down syndrome and childhood acute leukemia. Prog Clin Biol Res 246:19–32. Robison LL, Nesbit Jr. ME, Sather HN, et al. 1984. Down syndrome and acute leukemia in children: A 10-year retrospective survey from Childrens Cancer Study Group. J Pediatr 105(2):235–242. Rodvall Y, Pershagen G, Hrubec Z, et al. 1990. Prenatal X-ray exposure and childhood cancer in Swedish twins. Int J Cancer 46(3):362–365. Roman E, Ansell P, Bull D. 1997. Leukaemia and non-Hodgkin’s lymphoma in children and young adults: Are prenatal and neonatal factors important determinants of disease? Br J Cancer 76(3):406–415. Ron E, Modan B, Boice JD, Jr. et al. 1988. Tumors of the brain and nervous system after radiotherapy in childhood. N Engl J Med 319(16):1033– 1039. Rosenbaum PF, Buck GM, Brecher ML. 2000. Early child-care and preschool experiences and the risk of childhood acute lymphoblastic leukemia. Am J Epidemiol 152(12):1136–1144. Ross JA, Davies SM. 1999. Screening for neuroblastoma: progress and pitfalls. Cancer Epidemiol Biomarkers Prev 8(2):189–194. Ross JA, Davies SM. 2000. Vitamin K prophylaxis and childhood cancer. Med Pediatr Oncol 34(6):434–437. Ross JA, Gurney JG. 1998. Hepatoblastoma incidence in the United States from 1973 to 1992. Med Pediatr Oncol 30(3):141–142. Ross JA, Potter JD, Shu XO, et al. 1997. Evaluating the relationships among maternal reproductive history, birth characteristics, and infant leukemia: A report from the Children’s Cancer Group. Ann Epidemiol 7(3):172– 179. Ross JA, Radloff GA, Davies SM. 2000. H19 and IGF-2 allele-specific expression in hepatoblastoma. Br J Cancer 82(4):753–756. Ross JA, Severson RK, Pollock BH, et al. 1996. Childhood cancer in the United States. A geographical analysis of cases from the Pediatric Cooperative Clinical Trials groups. Cancer 77(1):201–207. Ross JA, Severson RK, Swensen AR, et al. 1999. Seasonal variations in the diagnosis of childhood cancer in the Untied States [comment]. Br J Cancer 81(3):549–553. Ruymann FB, Maddux HR, Ragab A, et al. 1988. Congenital anomalies associated with rhabdomyosarcoma: An autopsy study of 115 cases. A report from the Intergroup Rhabdomyosarcoma Study Committee (representing the Children’s Cancer Study Group, the Pediatric Oncology Group, the United Kingdom Children’s Cancer Study Group, and the Pediatric Intergroup Statistical Center). Med Pediatr Oncol 16(1):33–39. Salonen MJ, Siimes MA, Salonen EM, et al. 2002. Antibody status to HHV-6 in children with leukaemia. Leukemia 16(4):716–719. Sandlund JT, Downing JR, Crist WM. 1996. Non-Hodgkin’s lymphoma in childhood. N Engl J Med 334(19):1238–1248. Sarasua S, Savitz DA. 1994. Cured and broiled meat consumption in relation to childhood cancer: Denver, Colorado (United States). Cancer Causes Control 5(2):141–148. Schottenfeld D, Warshauer ME, Sherlock S, et al. 1980. The epidemiology of testicular cancer in young adults. Am J Epidemiol 112(2):232–246. Schuz J, Kaatsch P, Kaletsch U, et al. 1999. Association of childhood cancer with factors related to pregnancy and birth. Int J Epidemiol 28(4): 631–639. Schuz J, Kaletsch U, Meinert R, et al. 2000. Risk of childhood leukemia and parental self-reported occupational exposure to chemicals, dusts, and fumes: Results from pooled analyses of German population-based casecontrol studies. Cancer Epidemiol Biomarkers Prev 9(8):835–838. Schuz J, Kaletsch U, Meinert R, et al. 2001. High-birth weight and other risk factors for Wilms tumour: Results of a population-based case-control study. Eur J Pediatr 160(6):333–338. Schuz J, Kaletsch U, Meinert R, et al. 2001. Risk factors for neuroblastoma at different stages of disease. Results from a population-based case-control study in Germany. J Clin Epidemiol 54:702–709. Schwartzbaum JA. 1992. Influence of the mother’s prenatal drug consumption on risk of neuroblastoma in the child. Am J Epidemiol 135(12): 1358–1367. Schwartzbaum JA, George SL, Pratt CB, et al. 1991. An exploratory study of environmental and medical factors potentially related to childhood cancer. Med Pediatr Oncol 19(2):115–121. Severson RK, Buckley JD, Woods WG, et al. 1993. Cigarette smoking and alcohol consumption by parents of children with acute myeloid leukemia: An analysis within morphological subgroups—a report from the Childrens Cancer Group. Cancer Epidemiol Biomarkers Prev 2(5):433–439. Sharpe CR, Franco EL. 1996. Use of dipyrone during pregnancy and risk of Wilms’ tumor. Brazilian Wilms’ Tumor Study Group. Epidemiology 7: 533–555.

Cancers in Children Sharpe CR, Franco EL, de Camargo B, et al. 1995. Parental exposures to pesticides and risk of Wilms’ tumor in Brazil. Am J Epidemiol 141(3): 210–217. Shaw G, Lavey R, Jackson R, et al. 1984. Association of childhood leukemia with maternal age, birth order, and paternal occupation. A case-control study. Am J Epidemiol 119(5):788–795. Shaw GM, Lammer EJ, Wasserman CR, et al. 1995. Risks of orofacial clefts in children born to women using multivitamins containing folic acid periconceptionally. Lancet 346(8972):393–396. Shore RE, Albert RE, Pasternack BS. 1976. Follow-up study of patients treated by X-ray epilation for Tinea capitis; resurvey of post-treatment illness and mortality experience. Arch Environ Health 31(1):21–28. Shu XO, Clemens J, Zheng W, et al. 1995. Infant breastfeeding and the risk of childhood lymphoma and leukaemia. Int J Epidemiol 24(1):27–32. Shu XO, Gao YT, Brinton LA, et al. 1988. A population-based case-control study of childhood leukemia in Shanghai. Cancer 62(3):635–644. Shu XO, Han D, Severson RK, et al. 2002. Birth characteristics, maternal reproductive history, hormone use during pregnancy, and risk of childhood acute lymphocytic leukemia by immunophenotype (United States). Cancer Causes Control 13(1):15–25. Shu XO, Nesbit ME, Buckley JD, et al. 1995. An exploratory analysis of risk factors for childhood malignant germ-cell tumors: Report from the Childrens Cancer Group (Canada, United States). Cancer Causes Control 6(3):187–198. Shu XO, Potter JD, Linet MS, et al. 2002. Diagnostic X-rays and ultrasound exposure and risk of childhood acute lymphoblastic leukemia by immunophenotype. Cancer Epidemiol Biomarkers Prev 11(2):177– 185. Shu XO, Reaman GH, Lampkin B, et al. 1994. Association of paternal diagnostic X-ray exposure with risk of infant leukemia. Investigators of the Childrens Cancer Group. Cancer Epidemiol Biomarkers Prev 3(8): 645–653. Shu XO, Ross JA, Pendergrass TW, et al. 1996. Parental alcohol consumption, cigarette smoking, and risk of infant leukemia: A Childrens Cancer Group study. J Nat Cancer Inst 88(1):24–31. Shu XO, Stewart P, Wen WQ, et al. 1999. Parental occupational exposure to hydrocarbons and risk of acute lymphocytic leukemia in offspring. Cancer Epidemiol Biomarkers Prev 8(9):783–791. Silcocks PB, Murrells T. 1987. Space-time clustering and bone tumours: application of Knox’s method to data from a population-based cancer registry. Int J Cancer 40(6):769–771. Silva JF, Subramanian N. 1975. An epidemiologic study of osteogenic sarcoma in Malaysia. Incidence in urban as compared with rural environments and in each of three separate racial groups, 1969–1972. Clin Orthop Rel Res (113):119–127. Smith MA, Freidlin B, Ries LA, et al. 1998. Trends in reported incidence of primary malignant brain tumors in children in the United States [comment]. J Nat Cancer Inst 90(17):1269–1277. Smulevich VB, Solionova LG, Belyakova SV. 1999. Parental occupation and other factors and cancer risk in children: I. Study methodology and nonoccupational factors. Int J Cancer 83(6):712–717. Sorahan T, Lancashire R, Prior P, et al. 1995. Childhood cancer and parental use of alcohol and tobacco. Ann Epidemiol 5(5):354–359. Sorahan T, Lancashire RJ, Hulten MA, et al. 1997. Childhood cancer and parental use of tobacco: Deaths from 1953 to 1955. Br J Cancer 75(1):134–138. Sorahan T, McKinney PA, Mann JR, et al. 2001. Childhood cancer and parental use of tobacco: Findings from the inter-regional epidemiological study of childhood cancer (IRESCC). Br J Cancer 84(1):141–146. Spitz MR, Johnson CC. 1985. Neuroblastoma and paternal occupation. A casecontrol analysis. Am J Epidemiol 121(6):924–929. Steinbuch M, Weinberg CR, Buckley JD, et al. 1999. Indoor residential radon exposure and risk of childhood acute myeloid leukaemia. Br J Cancer 81(5):900–906. Stevens W, Thomas DC, Lyon JL, et al. 1990. Leukemia in Utah and radioactive fallout from the Nevada test site. A case-control study. JAMA 264(5):585–591. Stiller CA. 1998. What causes Hodgkin’s disease in children? Eur J Cancer 34(4):523–528. Stiller CA, Boyle PJ. 1996. Effect of population mixing and socioeconomic status in England and Wales, 1979–85, on lymphoblastic leukaemia in children. BMJ 313(7068):1297–1300. Stjernfeldt M, Berglund K, Lindsten J, et al. 1992. Maternal smoking and irradiation during pregnancy as risk factors for child leukemia. Cancer Detect Prev 16(2):129–135. Strader CH, Weiss NS, Daling JR, et al. 1988. Cryptorchism, orchiopexy, and the risk of testicular cancer. Am J Epidemiol 127(5):1013– 1018.

1267

Strickler HD, Goedert JJ. 1998a. Exposure to SV40-contaminated poliovirus vaccine and the risk of cancer—a review of the epidemiological evidence. Dev Biol Stand 94:235–244. Strickler HD, Rosenberg PS, Devesa SS, et al. 1998b. Contamination of poliovirus vaccines with simian virus 40 (1955–1963) and subsequent cancer rates. JAMA 279(4):292–295. Suminoe A, Matsuzaki A, Kinukawa N, et al. 1999. Rapid somatic growth after birth in children with neuroblastoma: A survey of 1718 patients with childhood cancer in Kyushu-Okinawa district. J Pediatr 134(2):178–184. Tanimura M, Matsui I, Abe J, et al. 1998. Increased risk of hepatoblastoma among immature children with a lower birth weight. Cancer Res 58(14):3032–3035. Tjalma RA. 1966. Canine bone sarcoma: Estimation of relative risk as a function of body size. J Natl Cancer Inst 36(6):1137–1150. Tucker MA, D’Angio GJ, Boice JD, Jr., et al. 1987. Bone sarcomas linked to radiotherapy and chemotherapy in children. N Engl J Med 317(10): 588–593. Valery PC, McWhirter W, Sleigh A, et al. 2002. Farm exposures, parental occupation, and risk of Ewing’s sarcoma in Australia: A national case-control study. Cancer Causes Control 13(3):263–270. van Duijn CM, van Steensel-Moll HA, Coebergh JW, et al. 1994. Risk factors for childhood acute non-lymphocytic leukemia: An association with maternal alcohol consumption during pregnancy? Cancer Epidemiol Biomarkers Prev 3(6):457–460. Varghese C, Barrett JH, Johnston C, et al. 1996. High risk of lymphomas in children of Asian origin: Ethnicity or confounding by socioeconomic status? Br J Cancer 74(9):1503–1505. Vilchez RA, Kozinetz CA, Butel JS. 2003. Conventional epidemiology and the link between SV40 and human cancers. Lancet Oncol 4(3):188–191. Vineis P, Miligi L, Crosignani P, et al. 2003. Delayed infection, late tonsillectomy or adenoidectomy and adult leukaemia: A case-control study. Br J Cancer 88(1):47–49. Volgelstein B, Kinzler KW. 1998. The Genetic Basis of Cancer. New York: McGraw-Hill. Walker AH, Ross RK, Haile RW, et al. 1988. Hormonal factors and risk of ovarian germ cell cancer in young women. Br J Cancer 57(4):418– 422. Watson MS, Carroll AJ, Shuster JJ, et al. 1993. Trisomy 21 in childhood acute lymphoblastic leukemia: A Pediatric Oncology Group study (8602). Blood 82(10):3098–3102. Weinreb M, Day PJ, Niggli F, et al. 1996. The role of Epstein-Barr virus in Hodgkin’s disease from different geographical areas. Arch Dis Child 74(1):27–31. Werler MM, Hayes C, Louik C, et al. 1999. Multivitamin supplementation and risk of birth defects. Am J Epidemiol 150(7):675–682. Westerbeek RM, Blair V, Eden OB, et al. 1998. Seasonal variations in the onset of childhood leukaemia and lymphoma. Br J Cancer 78(1):119–124. Westergaard T, Andersen PK, Pedersen JB, et al. 1997. Birth characteristics, sibling patterns, and acute leukemia risk in childhood: A population-based cohort study. J Natl Cancer Inst 89(13):939–947. Wexler L, Helman L. 1997. Rhabdomyosarcoma and the undifferentiated sarcomas. In: Pizzo PA, Poplack DG, eds. Principles and Practice of Pediatric Oncology. Philadelphia; Lippincott-Raven: pp. 799–829. Wiemels JL, Cazzaniga G, Daniotti M, et al. 1999. Prenatal origin of acute lymphoblastic leukaemia in children [comment]. Lancet 354(9189): 1499–1503. Wiemels JL, Ford AM, Van Wering ER, et al. 1999. Protracted and variable latency of acute lymphoblastic leukemia after TEL-AML1 gene fusion in utero. Blood 94(3):1057–1062. Wiemels JL, Xiao Z, Buffler PA, et al. 2002. In utero origin of t(8;21) AML1ETO translocations in childhood acute myeloid leukemia. Blood 99(10):3801–3805. Wilkins JR, 3rd, Hundley VD. 1990. Paternal occupational exposure to electromagnetic fields and neuroblastoma in offspring. Am J Epidemiol 131(6):995–1008. Wilkins JR, 3rd, Koutras RA. 1988. Paternal occupation and brain cancer in offspring: A mortality-based case-control study. Am J Ind Med 14(3): 299–318. Wilkins JR, 3rd, Sinks T. 1990. Parental occupation and intracranial neoplasms of childhood: Results of a case-control interview study. Am J Epidemiol 132(2):275–292. Wilkins JR, 3rd, Sinks TH, Jr. 1984a. Occupational exposures among fathers of children with Wilms’ tumor. J Occup Med 26(6):427–435. Wilkins JR, 3rd, Sinks TH, Jr. 1984b. Paternal occupation and Wilms’ tumour in offspring. J Epidemiol Commun Health 38(1):7–11. Winn DM, Li FP, Robison LL, et al. 1992. A case-control study of the etiology of Ewing’s sarcoma. Cancer Epidemiol Biomarkers, Prev 1(7): 525–532.

1268

PART IV: CANCER BY TISSUE OF ORIGIN

Woods WG, Roloff JS, Lukens JN, et al. 1981. The occurrence of leukemia in patients with the Shwachman syndrome. J Pediatr 99(3):425–428. Yagi T, Hibi S, Tabata Y, et al. 2000. Detection of clonotypic IGH and TCR rearrangements in the neonatal blood spots of infants and children with B-cell precursor acute lymphoblastic leukemia. Blood 96(1):264–268. Yang P, Grufferman S, Khoury MJ, et al. 1995. Association of childhood rhabdomyosarcoma with neurofibromatosis type I and birth defects. Genet Epidemiol 12(5):467–474. Yang Q, Olshan AF, Bondy ML, et al. 2000. Parental smoking and alcohol consumption and risk of neuroblastoma. Cancer Epidemiol Biomarkers Prev 9(9):967–972.

Yeazel MW, Buckley JD, Woods WG, et al. 1995. History of maternal fetal loss and increased risk of childhood acute leukemia at an early age. A report from the Childrens Cancer Group. Cancer 75(7):1718–1727. Yeazel MW, Ross JA, Buckley JD, et al. 1997. High birth weight and risk of specific childhood cancers: a report from the Children’s Cancer Group. J Pediatr 131(5):671–677. Zack M, Adami HO, Ericson A. 1991. Maternal and perinatal risk factors for childhood leukemia. Cancer Res 51(14):3696–3701.

66

Multiple Primary Cancers DAVID SCHOTTENFELD AND JENNIFER L. BEEBE-DIMMER

A

pproximately 200,000 new cancers each year in the United States are multiple primary cancers, or about 16% of the annual incidence of invasive cancers (Ries et al., 2003). Epidemiologic studies of various incidence patterns of multiple primary cancers have guided research about carcinogenic mechanisms that are shared by organ sites (Boice et al., 1985). At issue is whether risk patterns exist that predict subsequent primary cancers in patients with an index primary cancer of a particular organ and type. Also at issue is whether apparent increases in risk result from common etiologic factors in the pathogenesis of the index and second primary cancers; adverse toxic effects of agents used in the management of the index cancer; random or chance effects (for example, those occurring after multiple significance testing); or spurious associations that are the result of unrecognized or uncontrolled confounding by lifestyle or behavioral risk factors, or by biased ascertainment of new primary cancers as a result of more careful medical surveillance in patients with a history of cancer. The application of criteria for judging the biological plausibility of a common pathogenesis would require molecular analysis of tumors and assessment of mutually excessive occurrences of second primary cancers in cohorts of patients with cancer and of epidemiologic features that are common to the organ sites. The identification of specific and predictive patterns of multiple primary cancers should facilitate the cost-effective targeting of long-term early detection methods or other preventive interventions. The multifocal appearance of cancer has been described within a single organ, in paired organs, and in contiguous epithelial tissues shared by different organs. The multifocal clinical presentation of a malignant epithelial neoplasm may arise, however, from a single parental cell (i.e., monoclonal), or its genesis may be independent and multiclonal (Bedi et al., 1996; Carey, 1996). Multifocal lesions that are genetically similar arise from a single precursor lesion, presumably as a result of lateral spread at the base of the lesion accompanied by focal extensions that are topographically distinct at the surface. The monoclonality of multifocal lesions has been described in patients with carcinoma of the skin, urinary bladder, upper aerodigestive tract, colon, breast and genital tract in women and in the metachronous lesions of lymphomas and leukemias. For example, in the molecular genetic analysis of multifocal urinary bladder cancers, in most instances examination of the pattern of X-chromosome inactivation and loss of the chromosome 9 allele would establish the monoclonal evolution from the progenitor cell. Subsequent genetic or clonal heterogeneity may be associated with later events in progression (e.g., losses of 17p and 18q alleles) that are consistent with genomic instability and independent and variable genetic events of autonomous growth patterns. Studies of oropharyngeal cancers suggest that early genetic changes, such as loss of chromosomal region 9p21, or of mutated or deleted p53, do not necessarily correlate with morphologic abnormalities. A main effect of this loss is inactivation of the p16 gene, an inhibitor of cyclin-dependent kinase (CDK), or loss of function of p53 with resulting genomic instability. Slaughter et al. (1953), in reporting on the high incidence of second primary cancers in patients with head and neck squamous cell carcinoma (HNSCC) proposed that “field cancerization” had occurred, implying that multiple transforming events have given rise to genetically unrelated multiple primary tumors. Most field changes appear to be induced by exposure to tobacco smoke with resulting multifocal cancerization, rather than by migrated transformed cells. Fialkow

(1974) proposed that a neoplasm developing as a result of rare random events would be expected to have a monoclonal origin, whereas if malignant transformation occurs in many contiguous cells simultaneously, the resulting tumors might be expected to be multiclonal. The distinction about clonality may be investigated using molecular techniques that identify patterns of allelic loss or inactivation that represent critical events of neoplastic transformation or that occur early in progression. Cytogenetic clonal analysis may be investigated by combining karyotyping with fluorescence in situ hybridization. The demonstration of the temporality, relative frequency, and clonal genesis of multiple primary tumors would provide a rationale for the type, onset, and duration of preventive intervention.

METHODOLOGY Although observations on prevalence and proportional frequencies of specific combinations of multiple primary cancers are of historical interest, they do not enable as precise a determination of relative risk as when this is based upon incidence within a cohort of patients with a particular index cancer. To determine whether various combinations of second primary cancers are occurring more frequently than might be expected on the basis of chance, the observed number of sitespecific cancers are summarized in relation to the person-years of observation subsequent to the diagnosis of the index cancer. The expected number of second primary cancers is derived by multiplying the person-years at risk, by age, sex, and calendar period-specific incidence rates for cancers of all and selected sites of a population-based cancer registry. A unique advantage of a cohort study that is developed from a population-based registry, as distinguished from one in which the patients are derived from one or more hospital cancer registries, is that the observed and expected numbers of second primary cancers are derived from the same reference population. One cautionary note would be that because cancer patients are under closer medical surveillance than the general population, the diagnosis of a second primary cancer may be subject to lead-time or detection bias, namely detected before the tumor would have become clinically diagnosed in the usual medical practice setting. This may result in an inflated standardized incidence ratio (SIR) within the specified interval after the index cancer diagnosis. The denominator, person-years, increases either with longer followup or with selection of a larger sample, and treats equivalently the exposure time of five persons for 1 year, one person for 5 years, or 10 persons for 6 months. As a measure of risk of second primary cancers, this index of incidence density would be appropriate for a homogeneous cohort of cancer patients experiencing a constant risk per unit of person-time. If this were not the case, an actuarial or life-table method of survival analysis would be more appropriate. Since the probability of a subsequent or metachronous primary cancer (i.e., a primary cancer diagnosed after a specified interval and subsequent to the initial cancer diagnosis) is relatively small, the number of observed primary cancers can be expected to follow a Poisson distribution. The statistical significance of the standardized ratio of the observed to the expected number of cancers can be tested readily. The cornerstone of epidemiological research on cohorts with multiple primary cancers is precise morphologic classification. The same

1269

1270

PART IV: CANCER BY TISSUE OF ORIGIN

histopathologic criteria and methods of case ascertainment should be employed uniformly in the study group, who have previously developed the index cancer, and in the comparison population. Misclassification is likely to be more common when multiple tumors of the same morphology occur in contiguous tissues. For example, in a study of multiple primary cancers of the upper digestive system, recurrent carcinoma may be distinguished from a subsequent primary carcinoma within adjacent tissues by demonstrating clear margins of resection for the previous primary and in situ foci of origin for the subsequent primary. Cytogenetic and molecular biomarker studies will facilitate identification of independent clones of malignant cells. Clinically apparent metachronous primary cancers should be distinguished from the latent in situ cancers discovered through cytologic screening and the occult or incidental cancers discovered at necropsy. As summarized by Flannery and colleagues (1985), the rules for coding multiple primary cancers in the Connecticut Tumor Registry are: 1. A single neoplasm of one type of histology, or of mixed histology, is considered a single primary cancer even if the lesion extends into different anatomic areas. 2. A lesion of the same histology as the initial cancer, diagnosed in the same anatomic site of the index primary but after an interval of at least 2 months, is considered a separate primary unless stated to be recurrent or metastatic. 3. Multicentric foci within the same primary site are considered as a single primary cancer. Synchronous (i.e., <2 months) multiple primary lesions of the same histology occurring in different organs are considered to be independent primary cancer unless stated to be metastatic. 4. Multiple neoplasms of different types of histology within a single organ are considered separate primary cancers whether occurring simultaneously or metachronously. 5. If only one histologic type is reported and if paired organs are involved within 2 months of diagnosis, a determination must be made as to whether the patient has one, with metastasis to the contralateral organ, or two independent primary cancer(s). This determination is generally made by the pathologist based on whether areas of in situ cancer are seen in each paired organ (Flannery et al., 1985). One bias that may operate in the selection of cases for study from major cancer treatment centers is the “healthy person-years” bias, which is defined as the cancer-free interval between primary diagnosis and treatment and referral to a specialized center, in which the preceding person-years at risk for a study subject inflate the calculation of the expected number of cancers. In effect, these patients must survive to enable subsequent referral. The difficulty arises because of the interpretation of selection factors in patients who were treated previously or who presented with recurrence of their disease. If one decides to include referral patients and to accumulate patientyears at risk in the development of subsequent primary cancers from the date of diagnosis or date of first treatment, the “healthy personyears at risk” would be selected into the study. One can adjust for this by accumulating patient-years of observation from the date first seen at the referral institution. Such individuals should be considered as late entries into follow-up after an appropriate interval of time that is equivalent to the period from diagnosis or first treatment to the date first seen at the referral center. Although “healthy person-years” can be accounted for with such an approach, one cannot disregard the implications of ignoring previous treatment factors, which may have intrinsic oncogenic potential. As suggested previously, use of person-years at risk may not always be appropriate in studies of multiple primaries, because a major assumption of the approach requires that the risk following a putative causal exposure remains constant over time. The standard approach in which all person-years of observation are allocated to a single exposure category, such as a course of treatment, fails to take into account the transient nature of the exposure. Misclassification results from such an approach because all person-years at risk are assigned to a single treatment classification. A method of analysis should be

employed that allocates person-years of exposure to different treatment categories as a patient’s course of treatment changes, and that takes into account the interval since initial exposure and differential censoring among treatment groups. Modifications of standard lifetable analyses using techniques appropriate for analyzing transient levels of exposure provide a means of analyzing longitudinal data in which cohorts of patients who have received different intensities and combinations of therapy can be identified. The use of multiple entry and exit lifetables in such studies will permit patients to be entered or withdrawn from each relevant treatment interval. The evaluation of the carcinogenic effects of treatment generally is based on a randomized clinical trial or involves an observational cohort design or case-control nested within a cohort design. The statistical techniques that are used to compare the observed actuarial survival and cancer-specific survival in patients with varying distributions of prognostic factors can be adopted to adjust for variables that may influence the cumulative probability of developing multiple primary cancers. An implicit assumption of the lifetable method is that subjects who are censored as losses, withdrawals, or transfers, and subjects who remain under observation, are similar with respect to the probability of developing multiple primary cancers. Test of significance, such as the Mantel-Haenszel summary c2, the log-rank c2, or the generalized Wilcoxon test, which are used in survival analysis or estimating cumulative probability, involve similar assumptions about censored observations. When there are unequal patterns of censoring in the data, other non-parametric procedures with less restrictive assumptions regarding patterns of censoring are preferred when testing for significant differences between cumulative probabilities. The development of regression models for survival analysis has made it possible to examine time-dependent dose-response relationships, control for confounding by age, sex, race, calendar period, and other potential prognostic or risk factors, and explore interactions among putative causal factors. Regression models proposed for survival functions generally involve the assumption of proportional hazard. The assumption of proportional hazard implies that the probability of fatality or some end point for an individual is a constant multiple of a baseline risk level at all times. Thus, although the instantaneous risk may change with time, the ratio of risks or relative risk is assumed to be constant. In studies concerned with the potential carcinogenic effects of various modalities or agents used in anti-cancer therapy, the internal comparison of subgroups of patients is more informative than comparisons with an external reference population. Assuming that the cohort of cancer patients has been assembled, and that the multiple primary cancer cases have been diagnosed at specific points in time, other design options include the nested case-control within a cohort study that controls for the different intervals at which each case event has occurred.

HISTORICAL PERSPECTIVE In one of the earliest studies by Warren and Gates (1932) of 1078 autopsies on cancer patients, 40 patients (3.7%) had either occult or clinically apparent secondary primary cancers. In a subsequent report by Warren and Ehrenreich (1994) of 2829 autopsies, it was estimated that the prevalence was 6.8%. Hajdu and Hajdu (1968) employed the criteria of Warren and Gates in their autopsy study of 3321 patients and reported that 5.3% of patients had second primary cancers, after excluding multicentric cancers of the same or paired organs. Berg et al. (1971) reviewed 5636 autopsies on cancer patients and reported that 3.1% of patients had occult second primary cancers, after excluding symptomatic and multicentric multiple primaries. In women, the age-specific prevalence of occult second primaries did not exceed 2.0% between 20 and 59 years, and then increased to 3.6% at 60–69, to 4.8% at 70–79, and to 7.0% at 80 years and older. In men, the prevalence varied between 1.0% and 2.1% between 20 and 59 years, and then increased to 5.8% at 60–69, 9.4% at 70–79, and to 16.5% at 80 years and older (Berg et al., 1971).

Multiple Primary Cancers In a review of 37,580 patients at the Mayo Clinic (Moertel et al., 1961), 5.1% of patients manifested clinically apparent multiple primary cancers; 2.8% manifested cancers in different organs or tissues. In a review of 41,341 cancer patients at the Memorial SloanKettering Cancer Center (Schottenfeld and Berg, 1975), the average annual incidence of clinically apparent second primary cancers in different organs or tissues was 10.9 per 1000 per year; after excluding metachronous skin cancers, the incidence of metachronous cancers in other organs and tissues, including melanoma, was 7.9 per 1000 per year. The average annual incidence of subsequent primary cancers (excluding keratinocyte skin cancers) observed in residents of Connecticut with a previous primary cancer of any site was almost 12.0 per 1000; the average annual incidence of metachronous cancers in men was 15.0 per 1000, and in women, 10.3 per 1000. Individuals with a previous primary cancer had a relative risk of 1.29 (P < 0.01) of developing a metachronous primary cancer when compared with individuals in the general population without an antecedent primary cancer. The incidence rates within the resident population included both multicentric multiple primaries and multiple primaries in different organs or tissues (Schoenberg, 1977). In a cohort of 30,880 patients from Denmark, Sweden, and Norway, with an index primary cancer diagnosed under the age of 20 during the period 1943–1987, 238 patients developed a second primary cancer, yielding a relative risk of 3.6 (95% CI: 3.1–4.1). The relative risk increased substantially from 2.6 during 1940s and 1950s to 6.9 among cohort members diagnosed in the 1980s (Olsen et al., 1993). Among 470,000 cancer patients registered between 1953 and 1991 in Finland, the risk of a second primary cancer was increased by 70% only in the patients who were less than 50 years of age at diagnosis of their index primary cancer (Sankila et al., 1995).

RISK MECHANISMS AND PATTERNS OF MULTIPLE PRIMARY CANCERS Familial Cancer Syndromes In hereditary tumors, every embryonic cell contains a predisposing mutation. Familial predisposition to cancer is viewed in general to be associated with a germline mutation or inactivation of a putative tumor suppressor gene or less commonly of a DNA mismatch repair gene, or activation of a proto-oncogene (Fearon, 1997). Although the estimated proportion will vary with the type of cancer, about 1–2% of all cancers arise in individuals with an inherited cancer syndrome. Heritability should be considered when several generations of family members are diagnosed with organ-specific cancers at a relatively younger than expected age, or when the affected individuals develop multiple primary cancers (Lindor and Greene, 1998). The phenotypic expression of cancer in those individuals carrying an inherited mutant allele may be enhanced through interactions with modifier genes or with exogenous agents and lifestyle risk factors. Consistent with Knudson’s “two-hit” hypothesis, germline mutation, altered DNA methylation, or genomic imprinting of the tumor suppressor gene is expressed recessively in the affected somatic cell, and thus if neoplastic transformation is to occur, the corresponding normal or “wildtype” allele must be inactivated (Barlow, 1993; Hall, 1997). The mechanism for loss of function of the wild-type homologous allele in a somatic cell may involve an independent point mutation, chromosomal deletion as a result of nondisjunction and duplication of the mutant-containing chromosome, loss of the normal chromosome with a remaining mutant monosomic chromosome, or mitotic recombination, or through altered DNA methylation. Currently, three genetic cancer syndromes have been attributed to activating mutations in proto-oncogenes: RET in multiple endocrine neoplasia type 2, MET in hereditary papillary renal cancer, and CDK4 in familial melanoma. The protein encoded by the MET oncogene functions as a receptor for hepatocyte growth factor. Inherited renal cancer syndromes present with multiple primary tumors, independent clones, and bilateral kidney tumors. There are, however, tumor suppressor genes that predispose to inherited renal carcinomas, such as the von Hippel-Lindau

1271

(VHL) disease. The VHL gene produces a protein that regulates the stability of the hypoxia-inducible factor and transcription of vascular endothelial growth factor. The phenotypic syndrome includes vascular tumors of the retina, hemangioblastoma of the central nervous system, and/or pheochromocytoma of the adrenal gland. With the cloning of cancer susceptibility genes, it has become possible to perform DNA testing in family members. Genetic testing may be used appropriately to determine the genetic status of individuals already suspected of being at increased risk because of family history or clinical signs and symptoms. The ethical principles applicable to genetic testing include the necessity of demonstrating the validity of the testing methods, defining and communicating benefits and risks, voluntary participation, informed consent, and confidentiality. The predictive interpretation of a positive genetic test suggesting increased cumulative lifetime risk for a particular family member must consider penetrance and expression of the gene that may be age dependent, and that in addition the expression of the susceptibility allele may be modified by other genes, or lifestyle and environmental risk factors. Inherited forms of cancer are not restricted to the Mendelian rare cancer syndromes. Perhaps 5%–10% of breast cancers, particularly when diagnosed in women less than 40 years of age, may be related to the inheritance of germline mutations in several susceptibility genes. In contrast to other rare familial cancer syndromes such as multiple endocrine neoplasia type 2 and the concurrence of rare types of neoplasms, it may be difficult to distinguish chance clustering of sporadic breast cancer cases in a sibling and mother from familial breast cancer due to a germline mutation. BRCA1, a tumor suppressor gene, which increases susceptibility to both breast and ovarian cancer, was localized by genetic linkage to chromosome 17q21 and subsequently cloned. BRCA1 functions as a tumor suppressor gene that is involved with DNA repair and cell cycle regulation. Whereas the lifetime cumulative risk of sporadic breast cancer in the US general population is about 12%, and of ovarian cancer about 1.5%, BRCA1 mutations with an estimated gene frequency of 0.0006 or 1 per 800 women, increase the cumulative lifetime risk of breast cancer to about 45%–70% and of ovarian cancer to about 16%–63%. In some families, BRCA1associated cancers include colon and prostate. A second breast cancer gene, BRCA2, has been localized to chromosome 13q and is associated with elevated risks of breast cancer in women and men, and to a lesser degree than BRCA1, of ovarian cancer (Easton et al., 1993; Ford et al., 1994; Evans et al., 1994; Berry et al., 1997; Robson et al., 1998; Blackwood and Weber, 1998). Approximately 1% of Ashkenazi Jewish women carry the BRCA1 deletion of an adenine and guanine nucleotide at codon position 185, or a second “founder mutation” of an inserted cytosine nucleotide at 5382, and between 1% and 2% carry a BRCA2 deletion of a thymine nucleotide at codon 6174 (Goldgar and Reilly, 1995). Although screening for breast cancer susceptibility among Ashkenazi Jews would be technically feasible there are potentially conflicting ethical, social, economic, medical management, and genetic counseling issues that must be carefully communicated to the patient and family members (American Society of Human Genetics, 1994). Presymptomatic DNA testing must be balanced by concerns of family members about ensuring confidentiality, namely implications of revealing the results of genetic screening on the availability and cost of health insurance and the security of future employment. Clinical and preventive medical practice guidelines are concerned with the initiation and periodicity of screening, indication for prophylactic surgery, and/or chemoprevention. The Mendelian cancer syndromes are instructive in demonstrating that the inheritance of a defective gene alone is generally not sufficient for the multistage evolution of cancer. Penetrance and expressivity may depend on the effects of modifier genes or other genetic loci, exogenous or environmental factors, or stochastic variability in gene expression (Draper et al., 1986). In addition, multiple mutations occurring variously as single base-pair substitutions, deletions, insertions, duplications, inversions, and/or expansion of unstable repeat nucleotide sequences, appear to impact analogous loci in inherited and sporadic human cancers. The multifocal phenotype of most familial cancer syndromes has also stimulated research that has provided important insights into the nature of tumor suppressor genes. Namely,

1272

PART IV: CANCER BY TISSUE OF ORIGIN

mutations of tumor suppressor genes will result in inactivation or perturbation of molecular “gatekeeping” and “caretaking” functions (Levine, 1997). The “gatekeeping” genes are rate-limiting for tumor growth, whereas inactivation of “caretaking” genes leads to genetic instability. Thus disturbances in genetic functions that regulate cellular growth and differentiation, apoptosis, nucleotide excision repair, and mismatch repair are fundamental mechanisms of cancer susceptibility.

Tobacco: Interactions with Alcohol and Nutritional Insufficiency The complexity of tobacco smoke, namely containing more than 4000 chemicals, has made it difficult to identify the contribution of more than 50 specific putative carcinogenic agents. The tumorigenis agents in tobacco smoke include the polynuclear aromatic hydrocarbons (benz[a]anthracene, benzo[a]pyrene), N-nitrosamines (N-nitrosodimethylamine, N-nitrosonornicotine), aromatic amines or arylamines (2-napthylamine, 4-aminobiphenyl), aldehydes (formaldehyde, acetaldehyde), other organic (benzene, acrylonitrile) and inorganic (arsenic, chromium) compounds and polonium 210. The composition of the smoke depends on the ambient conditions of smoking, the blend of tobacco leaf, filtration, additives, paper wrapping, and other factors. The majority of the components are produced in an oxygen-deficient, hydrogen-rich environment, arising from pyrolysis and distillation, in the region immediately behind the burning tip of the cigarette. The chemical analysis of tobacco smoke is separated into particulate or “tar” and gaseous phases. The curing, fermentation, and aging of smokeless tobacco products favor the formation of various N-nitrosamines from tobacco alkaloids. Among the nitrosamines in smokeless tobacco, N-nitrosonornicotine and 4(methyl, nitrosoamino)-1-(3-pyridyl)-1-butanone are carcinogenic in variously tested rodents. During chewing or “snuff dipping”, additional amounts of carcinogenic tobacco-specific N-nitrosamines are formed endogenously in the oral cavity. Epidemiologic studies conducted in many countries have established that the risks of oral cavity, pharyngeal, laryngeal, lung, esophageal, stomach, pancreatic, liver, colorectal, anal, kidney and urinary bladder, and uterine cervical carcinomas are increased among cigarette smokers (Thun et al., 1995). In countries such as India, Pakistan, Thailand, Sri Lanka, and Afghanistan, where the use of snuff and chewing tobacco is quite common, oral and pharyngeal cancer mortality rates are among the highest in the world. The use of smokeless tobacco in India and Southeast Asia involves preparing a betel quid consisting of the leaf of the betel vine, sliced or shaved areca nut from the betel palm, and powdered slaked lime. Protracted use is associated with the pathologic sequelae of erythroleukoplakia, epithelial dysplasia, and epidermoid carcinoma of the gingivobuccal or lingual mucosa, hypopharynx, supraglottis, or esophagus. Patients with an epidermoid carcinoma in the upper aerodigestive tract are at increased risk of developing metachronous primary cancers in the contiguous mucosal tissues of the oral cavity and pharynx, esophagus, larynx, and lung and bronchus. The cumulative incidence of metachronous primary cancers in the aerodigestive tract within 10 years after treatment of the index primary cancer has been reported variously between 5% and 40%, or between 0.5% and 3.5% per year. Patients who are currently smoking at the time of diagnosis of the index primary cancer of the aerodigestive tract are at significantly greater risk, approximately fourfold, when compared with former smokers, for a second aerodigestive tract cancer. Tobacco use in the United States is estimated to contribute directly to 30% of total cancer mortality; the corresponding attributable proportion of all cancer deaths in men is estimated to be 40%–45% and 20%–25% in women. Tobacco control in the United States has evolved since the publication of the Surgeon General’s report in 1964 on the health consequences of smoking. In 2001, the US Health Interview Survey estimated that 46.2 million adults were current smokers, and that 44.7 million adults were former smokers. Current smoking prevalence was highest among persons aged 18–24 years (26.9%), and among those aged 25–44 years (25.8%), and lowest among the older

than 65 years (10.1%) (Fiore et al., 1989; Centers for Disease Control and Prevention, 1992). A variety of interventions have been advocated to influence smoking cessation and smoking initiation: school-based health education programs; reducing minors’ access to tobacco products; developing and enacting clean indoor air policies and legislation; restricting or eliminating advertising directed toward persons aged <18 years; and increasing tobacco excise taxes. A 10% increase in the price of cigarettes has resulted in a 4% decrease in adult consumption, and a 14% decrease in teenage consumption (Warner, 1986). Regulation of the addictive and carcinogenic potential of tobacco by controlling nicotine and tar yield has been beneficial in relation to the amount smoked or depth of inhalation. A major concern is that the cancer incidence and mortality rates due to smoking will shift in relative magnitude after the year 2000 from the industrialized nations to developing nations. While tobacco smoking prevalence proportions have been gradually declining by about 1% per year in men and women in North America and Western Europe, they have been rising at about 2% per year in Eastern European and Asian countries. In the early 1970s the average annual per capita consumption of cigarettes in North America and Western Europe was 3.3 times the level in Asian countries; by the early 1990s, this ratio decreased to 1.8. In a prospective study conducted in Shanghai, 61% of men described themselves as current cigarette smokers. Among the Chinese men, about 20% of all deaths were attributed to cigarette smoking; of these deaths, one-third were due to lung cancer, and the relative risks were increased in the smoking men for lung, esophageal and liver cancers, coronary heart disease, and chronic obstructive pulmonary disease (Peto et al., 1992). The 1990 report of the Surgeon General, “The Health Benefits of Smoking Cessation”, reviewed comprehensively the epidemiologic evidence for the long-term benefits of cessation of use of tobacco. The patterns of changing risk after cessation may be illustrated with an analysis of respiratory cancers. The risk of lung cancer incidence increases with the number of cigarettes smoked daily and with the total number of years smoked. Former smokers experience a 20%–90% reduction in risk of lung cancer when compared with current smokers. The rate of decline in former smokers is influenced by health status at the time of quitting and previous patterns of smoking duration, age at initiation, and average daily intensity of exposure. The magnitude of risk reduction is considerably greater among lighter smokers (<20 cigarettes/day), smokers of shorter duration or those who quit at younger ages, and smokers who inhaled less often or less deeply (United States Department of Health and Human Services, 1990).

Interaction with Ethyl Alcohol (Ethanol) Alcoholic beverages interact commonly with tobacco smoking in the natural history of cancers of the upper respiratory and gastrointestinal tracts. In addition, regular alcohol consumption in excess of three drinks per day (40–45 ml or 32–36 grams of ethyl alcohol) is associated with an increased risk of mortality due to liver cancer and cirrhosis, breast cancer, colorectal cancer, cardiomyopathy, and hemorrhagic stroke (Lieber, 1988; Lieber, 1995). Combined exposures to ethyl alcohol and tobacco in the United States account for 75%–85% of cancers of the oral cavity, pharynx, larynx, and esophagus. Of public health significance is the demonstration of synergy by the interaction of increasing levels of exposure to tobacco and ethyl alcohol. For oral and pharyngeal cancer, joint exposure to tobacco and ethyl alcohol results in odds ratios 2–2.5 times those expected if the effects of alcohol and tobacco were only additive. For laryngeal cancer, the interaction of ethyl alcohol and tobacco increases the risk about 50% more than the increase predicted if the effects were only additive. Other subsites within the upper aerodigestive tract exhibiting interaction with previous tobacco and ethyl alcohol exposures are the hypopharynx, supraglottis, and esophagus (Flanders and Rothman, 1982; IARC, 1988; Blot et al., 1988; Day et al., 1994). Possible carcinogenic mechanisms postulated for ethanol include: congeners, derivatives, or contaminants present in alcoholic beverages; induction of cytochrome P450 microsomal activating enzymes; solvent action with increased penetration or absorption of concurrent

Multiple Primary Cancers exposure to a genotoxic agent as in tobacco smoke; cytoxic damage with increased reparative proliferation; exacerbation of deficiencies in antioxidant micronutrients; and perturbation in cell-mediated immune responses or in DNA repair mechanisms. The oxidative metabolism of ethyl alcohol results in the production of acetaldehyde. Acetaldehyde is a reactive compound that forms adducts with various proteins, interferes with DNA repair, induces sister chromatid exchanges, and promotes depletion of glutathione. Minute amounts of acetaldehyde inactivate O6-methylguanine transferase, the enzyme responsible for repairing adducts resulting from alkylation at the O6 position of guanine (Garro and Lieber, 1990). The phenotypic expression of “mutagen sensitivity” has been described as potentiating the elevated risk of aerodigestive tract cancers in relationship to cigarette smoking and use of alcohol. Mutagen sensitivity is measured in an in vitro assay using the radiomimetic drug bleomycin as a test mutagen. Bleomycin is capable of inducing single- and double-stranded DNA breaks in cultured lymphocytes, which is quantitated as the number of chromatid breaks per cell scored on 50 metaphases per sample. Bleomycin-induced chromosome sensitivity is interpreted as measuring a component of the complex DNA repair pathways. To minimize cancer risks, heavy alcohol consumption should be avoided. Alcohol consumption in susceptible individuals may contribute to a wide range of medical problems. Individual susceptibility may be influenced by genetically determined variations in the activity of metabolic enzymes such as alcohol dehydrogenase, cytochrome P450 and the microsomal ethanol-oxidizing system, and DNA repair enzymes that influence the rate of removal of damaged DNA as a result of formation of toxic electrophilic metabolites and free-radical reactive intermediates. Current guidelines define “moderate” drinking as two or fewer drinks per day, and because of the concern over level of alcohol consumption associated with an increased risk of breast cancer, no more than one drink per day in women. Efforts to control the average consumption of alcohol in the general population and excessive consumption by individuals include school-based education, restrictions on alcohol advertising, mass media campaigns, increased excise taxes, and enforced restrictions on alcohol availability for those under age 21 years.

Interaction with Antioxidant Micronutrients A number of nutritional factors are thought to be important as modifiers of aerodigestive tract carcinogenesis. In the United States, nutritional deficiencies are commonly associated with, or exacerbated by, excessive alcohol ingestion. The potential interactions of deficiencies in essential micronutrients and exogenous genotoxic agents may give rise to altered mucosal integrity, enzyme and metabolic dysfunction, and morphologic abnormalities in specific target organs. Vitamin A and its provitamin, b-carotene, are needed for normal growth and differentiation of epithelial tissues, presumably mediated by regulating gene expression and transcription, and apoptosis. Deficiency of Vitamin A leads to a loss of mucociliary epithelium in the respiratory tract and its replacement by metaplastic squamous epithelium. The antioxidant micronutrients, such as the carotenoids and Vitamin C, function in trapping free radicals and reactive oxygen molecules, which are generated endogenously. Free radicals are highly reactive and alter the structure and function of DNA, denature proteins, and cause peroxidation of cell membrane lipids (Mayne, 1997). Vitamin C, ascorbic acid, is a water-soluble antioxidant that in addition to trapping free radicals and reactive singlet oxygen molecules, blocks the formation of carcinogenic N-nitroso compounds. The substrate or precursors for these compounds may be derived from tobacco, food additives as in nitrite-cured meats and salted, pickled or smoked fish or meat, alcoholic beverages, or pharmaceuticals. The primary dietary sources of Vitamin C are fruits and vegetables, especially citrus fruits, green leafy vegetables, tomatoes, and potatoes. Vitamin A in food may occur as preformed Vitamin A, namely retinol and retinol esters derived from animal foods, or as provitamin A carotenoids derived from plant foods. Of 700 known naturally occurring carotenoids, more than 50 serve as precursors to Vitamin A. Many

1273

other naturally occurring carotenoids in fruits and vegetables, however, are not precursors of Vitamin A but are effective antioxidants and scavengers of reactive oxygen molecules (McLaughlin et al., 1988; Steinmetz and Potter, 1991a; 1991b). As noted previously, patients with a previous primary carcinoma in the aerodigestive tract are at increased risk of metachronous primary squamous carcinomas in the contiguous mucosal tissue. “Field cancerization” or the development of multifocal squamous cell carcinomas in the aerodigestive tract is a major determinant of therapeutic failure following treatment of the index primary cancer. Chemoprevention is the use of natural or synthetic chemical agents that reverse, suppress, or prevent progression to invasive cancer. Its translational application in the pathogenesis of aerodigestive tract cancers is based on dual concepts of diffuse epithelial injury and “field cancerization” as a result of exposure to tobacco and alcohol, and multistep carcinogenesis in which there is an accumulation of genetic events associated with dysplasia, cell transformation and in situ carcinoma, and progressive invasive cancer. Chemoprevention targets the evolution of premalignant lesions, and the phase of clonal expansion before neoplastic cell transformation, and the genetic and epigenetic events associated with microscopic invasion. Chemopreventive agents under investigation include vitamins, minerals, antioxidants, anti-inflammatory agents, and molecularly targeted agents (Kim et al., 2002). In addition to carotenoids, fresh fruits and vegetables contain other micronutrients including Vitamin C, folic acid, flavones, isoflavonoids (e.g., soy products), protease inhibitors, thiocyanates, and indoles (e.g., indole-3-carbinol in Brassica vegetables) (Khuri et al., 1997; Omenn, 1998). Folic acid, methionine, and choline are interrelated in methyl group metabolism. Selective growth and transformation of cells can result from imbalances in DNA methylation, namely hypomethylation and over-expression of protooncogenes or hypermethylation of promoter regions that may attenuate the expression of tumor suppressor genes. Various chemopreventive mechanisms of action by micronutrients and non-nutritive phytochemicals in fruits and vegetables have been suggested by in vitro and animal feeding experimental studies. The complex interrelated mechanisms by which substances in vegetables and fruits may inhibit carcinogenesis include: regulation of cell differentiation; “quenching” or “trapping” of oxygen or hydroxyl free radicals; preventing the formation of electrophilic metabolites from precursor compounds by inhibiting the enzymatic activation pathway (e.g., cytochrome P450) or by inducing the detoxification pathway (e.g., glutathione S-transferase); enhancing DNA methylation; inhibiting the expression of oncogenes; and stimulating immune function (Steinmetz and Potter, 1991b).

Breast, Ovary and Endometrium: Hormones and Lifestyle Previous studies of multiple primary cancers have described mutual increases in relative risks for cancers of the breast, ovary, and uterine corpus. In the cohort of women in Connecticut with an initial breast cancer, the relative risk of a metachronous primary carcinoma of the ovary during more than 10 years of follow-up was increased to 1.7; in the cohort of women with an initial cancer of the ovary, the relative risk of a metachronous primary breast cancer was increased to 1.8 during the 5-year interval after the diagnosis of the first primary, and then declined (1.2) over the next 5-year interval. The interactive effect of age at diagnosis was demonstrable in that the subcohort of women with breast cancer diagnosed before 45 years of age exhibited an increased relative risk for ovarian cancer during the initial 5-year interval (3.8) and an overall relative risk of 2.6 during 10+ years of followup; in contrast to a relative risk in women with breast cancer diagnosed after 55 years of age of 0.8 at 5 years, and 1.2 overall (Schoenberg et al., 1969; Harvey and Brinton, 1985; Curtis et al., 1985). In various cohort studies, the elevated relative risks of breast and uterine corpus (endometrial) cancers were mutually elevated, particularly in women 55 years of age and over at diagnosis of the index cancer (relative risks = 1.3–2.0) (Schottenfeld and Berg, 1971; Adami et al., 1997). In an analysis of the Surveillance, Epidemiology and End Results (SEER)

1274

PART IV: CANCER BY TISSUE OF ORIGIN

program data, in women diagnosed with invasive cancer between 1973 and 1996, and at age less than 50 years, the significantly elevated standardized incidence ratios for metachronous primary ovarian cancer were observed for the following index primary cancers: breast (6.0; 95% CI: 4.9–7.2), uterine corpus (11.9; 95% CI: 7.3–18.4), colon (17.9; 95% CI: 11.1–27.3), and melanoma (3.5; 95% CI: 2.1–5.5) (Hall et al., 2001). The multiplex and interactive causal mechanisms for the mutual increases in risk of developing second primary cancers of the breast, ovary, and endometrium may be inferred by comparing the distribution of established risk factors for each organ site (Table 66–1). In a genetic epidemiological case-control study of patients with ovarian cancer, diagnosed between the ages 20–54 years, a significant risk factor was a family history in a first-degree relative of breast, col-

Table 66–1. Distribution of Risk Factors for Cancers of the Breast, Ovary, and Endometrium Risk Factor

menstruation Early age at menarche Early age at natural (or surgical) menopause Late age at natural menopause Years between menarche and menopause, excluding pregnancies, lactation-months, and/or duration of OC use

pregnancy Married, never pregnant Early age, first full-term pregnancy Multiple full-term pregnancies Late age (>35 yr) first full-term pregnancy Late age (>35 yr) last full-term pregnancy Cumulative number of lactating months

family history Organ-specific Multiple organ primary cancers: ovary, breast, endometrium, colon

Breast

Ovary

Endometrium

+ + +

+/? -* ? +

+/? + +

+ -† + + -/?

+ 0 ? ? -

+ 0 0 0

+ +

+ +

+ +

obesity: excess average weight, weight gain, and/or central vs. peripheral obesity Premenopausal women Postmenopausal women Height: increased average height, earlier age when achieving maximal height

oral contraceptive steroids

Estrogen/Progestin <45 yrs Estrogen/Progestin 45+ yrs

estrogen replacement therapy Without Progestin With Progestin

nutrition Fat (total, saturated) Calories Fruits and vegetables

tobacco ethanol physical activity

+ +

? ? +/?

+ + +/?

+/? 0/?

-

-

+ +

? ?

+ -‡

+/? +/? -/?

+/? +/? -/?

+/? +/? -/?

+/?**

0

-

+

0

0

-/?

-/?

-/?

+ = Increased risk. - = Decreased risk. 0 = No relationship established based upon adequate data. ? = Insufficient information; equivocal or conflicting data. +/? = May be associated with increased (or decreased) risk, although -/? = Studies may be conflicting, or suggest that the association may be limited to one or more subgroups of women. *In addition to hysterectomy, tubal sterilization may substantially reduce the risk of ovarian cancer. † Transient period of increased risk after first pregnancy. ‡ Reduction in risk toward 1.0. **Perhaps elevated in genetically susceptible subgroup (e.g., “slow” n-acetyltransferase phenotype).

orectal, or endometrial carcinoma; the associations with breast or endometrial cancer were most significantly increased in the subgroup of ovarian cancer cases classified as endometrioid carcinoma. Conditions associated with reduced cumulative frequency of ovarian cyclical activity and/or inhibition of pituitary secretion of gonadotropins, namely multiple pregnancies, long-term oral contraceptive use, and delayed menarche and early menopause are associated with reduced risks of ovarian and endometrial cancers. In contrast to the substantial protective effects demonstrated for combination oral contraceptives in studies of endometrial and ovarian cancer patients, the oral contraceptives, particularly when taken over a period of years before the first event of pregnancy, may be associated with a modest increase in relative risk (£2.0) of breast cancer. In perimenopausal women, circadian ovarian and pituitary interactions are accompanied by decreasing circulating estrogens and rising gonadotropins in conjunction with relatively more frequent anovular cycles; during this interval, exposure to oral contraceptives tend to supplement the decreasing endogenous levels of estrogens and may be associated with an increased risk of breast cancer. Estrogen replacement therapy without progestin supplementation in perimenopausal and postmenopausal women increases the risk of endometrial cancer approximately fourfold to eightfold. In the study by Newcomb and Trentham-Dietz (2003), each year of unopposed estrogen use was associated with a 14% increase in risk (95% CI: 10%–18%) (Newcomb and Trentham-Dietz, 2003). After an extensive reanalysis of original data from 51 epidemiologic studies, it was concluded that for each year of use of postmenopausal estrogen therapy, the risk of breast cancer increased by 2.3% (95% CI: 1.1%–3.6%). Experimental studies of hormonal carcinogenesis describe the genotoxic, mitogenic and promotional actions of estrogens on mammary ductal epithelial stem cells and endometrial basal cells. The importance of the epigenetic mechanisms of action is suggested by epidemiologic studies that show a greater risk among current users than former users, reversibility of risk after sustained cessation, and the indication of a dose threshold as inferred by estimations of cumulative dose and increasing risk, and assessment of minimal duration of use before elevated risks are discernible (Ewertz, 1988; Bergkvist et al., 1989; Colditz, 1998). The chemoprevention of endometrial cancer can be achieved by the combined administration of estrogens and progestins. Progestins inhibit the proliferation of endometrial cells by promoting differentiation to secretory cells, inducing apoptosis, downregulating estradiol and progesterone receptors, and by regulating estrogen steroid metabolizing enzymes. The extent of reduction in risk of endometrial hyperplasia and malignant neoplasia that is achieved by combining progestin with estrogen will be influenced by the strength of the formulation, dose, duration, and pattern of cycling, strength of opposing estrogenic effects, and individual underlying risk factors (Weiderpass et al., 1999). The hormonal physiology of the menstrual cycle impacts normal breast epithelium differently from that of the endometrium. The proliferative activity of the breast is highest during the luteal phase of the menstrual cycle. Progesterone in conjunction with estrogen stimulates mitogenesis and appears to have an additive effect on breast cancer risk when compared with estrogen replacement therapy alone. An alternative preventive strategy is the administration of a pharmaceutical agent that inhibits the stimulatory effect of endogenous estradiol. Tamoxifen, a triphenylethylene compound, has both antagonist and agonist estrogenic activity and has been used in palliative or adjuvant treatment of patients with estrogen receptor-positive breast cancer. The randomized Breast Cancer Prevention Trial in the United States and Canada announced that prophylactic use of tamoxifen in women without previous breast cancer resulted, after 5 years of treatment, in a 45% reduction in breast cancer incidence. Tamoxifen is thought to work primarily as an estrogen antagonist by binding to the estrogen receptor and blocking the effects of the more potent endogenous estrogens. However, additional activities have been described for antitumor activities, including modulation of the expression of growth factors (e.g., transforming growth factor-b, epidermal growth factor, and insulin-like growth factor-1), inhibition of angiogenesis, and induction

Multiple Primary Cancers of apoptosis. The estrogen agonist activity of tamoxifen has been demonstrated in relation to endometrial cancer, namely a twofold to fourfold increase in relative risk after more than 2 years of use. Current studies in progress of tamoxifen analogues, namely of selective estrogen receptor modulator agents, are evaluating whether pharmaceuticals can favorably impact the therapeutic benefits of estrogen on osteoblasts and bone matrix, and the coronary arterial endothelium, while reducing the risk of breast cancer and without significantly altering the risk of endometrial cancer (Powles and Hickish, 1995; Pritchard, 1998; Breast Cancer Trialists’ Collaborative Group, 1998).

Ultraviolet Radiation and Patterns of Multiple Primary Cancers Nonmelanoma or keratinocyte skin cancers are the most common cancers occurring in populations of European origin and with light skin pigmentation. The major causal factor is exposure to ultraviolet radiation (UV), most particularly UVB (280–315 nm), but also including the UVA spectrum (315–400 nm). Ultraviolet radiation induces structural alternations in DNA by causing covalent links between adjacent pyrimidine bases and results in local and systemic immunosuppression. In temperate zones such as the United States and northern Europe, basal cell carcinoma (BCC) is diagnosed 4 –10 times more frequently than squamous cell carcinoma (SCC). Compared with the incidence pattern of BCC, SCC exhibits a steeper rate of incidence with closer proximity to the equator and cumulative exposure to UV. The risk of UV-induced SCC is increased for chronically exposed anatomic areas such as the face, neck, ears, scalp (men), forearms, and backs of hands; in sun-sensitive individuals with manifestations of solar lentigines, facial telangiectasias, elastosis of the neck and dorsum of the hands, and solar keratoses; and in uniquely susceptible individuals who are immunosuppressed or have a rare genetic disorder such as the nevoid basal cell carcinoma syndrome or xeroderma pigmentosum. Basal cell carcinoma is more often associated with a history of sun exposure that had resulted in severe and painful burns during childhood and intermittent intense exposure during young adulthood. The anatomic distribution of BCC is similar to that of SCC. In contrast to invasive SCC in which actinic keratosis and Bowen in situ carcinoma are precursor lesions, or the dysplastic nevus in melanoma, BCC has no apparent precursor lesion. Melanoma of the skin and BCC appear to share many epidemiologic features of past intermittent intense exposures, childhood history of burns, and skin sensitivity; in contrast to SCC, chronic protracted exposure patterns as in association with an occupation, are not significant risk factors for BCC and melanoma (Hunter et al., 1990). The incidence trend of keratinocyte skin cancer in the United States has been correlated with depleting levels of stratospheric ozone; such depletion has resulted from the release of chlorofluorocarbons, which are used as propellants in aerosol spray cans and as refrigerants in air conditioning units. Ozone, which is formed photochemically by the action of solar ultraviolet radiation on oxygen molecules, is a component of the earth’s atmosphere that provides a protective shield from ultraviolet B radiation. For every 1% reduction of the average thickness of the ozone layer, the annual incidence of basal cell carcinoma is expected to increase by 3% and the annual incidence of squamous cell carcinoma is expected to increase by 5%. Rare hereditary diseases that predispose persons to basal cell carcinoma and squamous cell carcinoma are characterized by increased susceptibility to the effects of ultraviolet radiation. These conditions may further our understanding of the mechanisms of skin carcinogenesis that are applicable to the general population. For example, the basal cell nevus syndrome or Gorlin syndrome is an autosomal dominant disorder characterized by multiple basal cell carcinomas and developmental defects, including jaw cysts, skeletal anomalies, cutaneous pitting on the palms and soles, soft tissue calcifications, and hypertelorism; and an increased risk of ovarian fibroma, fibrosarcoma, meningioma, medulloblastoma, and rhabdomysarcoma. The basal cell carcinomas in the syndrome usually appear before the patient is 30 years of age. The early age at onset of basal cell carcinoma in the basal

1275

cell nevus syndrome is an expression of heightened “mutagen sensitivity”, particularly that due to exposure to sunlight and ionizing radiation. The gene responsible for the syndrome has recently been localized to bands q22.3 to 31 of chromosome 9 by linkage analysis of affected kindred. Positional cloning identified the human homologue of Gorlin syndrome in the Drosophila “patched” gene that functions in the developing neural tube, pharyngeal pouches, somites, and limb buds (Bale, 2002). Loss of heterozygosity for genetic markers in this region has been documented in one half of sporadic basal cell carcinomas, suggesting that a “gatekeeper” susceptibility gene resides at this locus. Somatic abnormalities in basal cell skin cancers have been observed in genes encoding proteins controlling the cell cycle, apoptosis, and genomic stability (e.g., cyclin D1, p16, p53). Mutations of the p53 gene have been demonstrated in up to 90% of SCC and about 50% of BCC (Harris, 1996). The same 9q chromosomal region is being assigned to the xeroderma pigmentosum complementing group A. Xeroderma pigmentosum is an autosomal recessive disease in which patients exposed to sunlight have a 1000-fold increase in risk for basal cell carcinoma, squamous cell carcinoma, and melanoma by 20 years of age. Although considerable clinical and genetic heterogeneity exists in persons with xeroderma pigmentosum, a fundamental biological mechanism is a reduced ability to repair DNA. Ultraviolet radiation induces pyrimidine dimer photoproducts with frequent C-T and CC-TT transitions that require efficient nucleotide excision repair mechanisms (Gailani et al., 1996). It has been suggested that a biomarker of cancer susceptibility, namely low DNA repair capacity, may be an independent risk factor in UV-exposed cases of basal cell skin cancer in the general population (Kripke, 1994; Schottenfeld, 1996). A history of nonmelanoma or keratinocyte skin cancer is associated with mutual increases in risk of melanoma, BCC and SCC skin cancers, and has been reported to be associated with increased risks of metachronous squamous carcinoma of the lip or conjuctiva, nonHodgkin lymphoma, and cancer of the salivary gland. When compared with the expected incidence of melanoma of the skin in the SEER population, 1973–1998, patients with non-Hodgkin lymphoma experienced a 60% increased relative risk (SIR = 1.6; 95% CI: 1.4–1.9), or an excess incidence of 2.1 per 10,000 per year. The associations between nonmelanoma skin cancer and non-Hodgkin lymphoma have been described in population-based cohort studies conducted in Denmark, Sweden, and Switzerland. The mutual increases in risk in cohorts with index primary squamous cell or basal cell skin cancer, or with non-Hodgkin lymphoma, and the parallel increases in incidence trends for each type of cancer in different geographic areas, have provided support for the hypothesis of a common cause (Frisch and Melbye, 1995; Frisch et al., 1996). The incidence of non-Hodgkin lymphoma has risen by more than 100% during the past 50 years in the United States and Western Europe. Immunodeficiency syndromes, either inherited or acquired, are accompanied by an increased risk of non-Hodgkin lymphoma. Immunosuppression may predispose to various kinds of skin cancer, including melanoma. Among patients receiving kidney transplants in conjunction with immunosuppressing medication, the increased relative and absolute risks of non-Hodgkin lymphoma with a latency period of less than 2 years, and of SCC on sun-exposed areas of the skin after a latency period of 5 years, is striking and underscores the significant interaction of host and environmental factors. It has been hypothesized that a person’s risk for non-Hodgkin lymphoma may be influenced by exposure to ultraviolet radiation and by immune dysfunction (Cartwright et al., 1994; McMichael and Giles, 1996). Further epidemiologic research to clarify this potential causal association should include case-control studies of non-Hodgkin lymphoma in relation to lifetime exposure to ultraviolet radiation, replication of population-based incidence studies of multiple primary cancers in patients with nonmelanoma and melanoma skin cancer that control for confounding lifestyle and sociodemographic risk factors, and continued monitoring of geographic patterns of incidence of nonHodgkin lymphoma in relation to measurements of ultraviolet exposure on the earth’s surface. Bentham (1996) reported in an ecological analysis that the incidence of non-Hodgkin lymphoma in England and

1276

PART IV: CANCER BY TISSUE OF ORIGIN

Wales during 1968–1985 was positively correlated with solar ultraviolet radiation levels, after controlling for area-wide differences in social class and employment in agriculture. Hartge et al. (1996) reviewed the geographic patterns of mortality rates of non-Hodgkin lymphoma in the United States for the three decades, 1950–1959, 1960–1969, and 1970–1980. In white men and women, non-Hodgkin lymphoma mortality was consistently lower in the southern regions of the United States when compared with the northern Midwest, New York, and New England, particularly during 1950–1959, whereas in contrast, mortality from melanoma and nonmelanoma skin cancer was highest in the south during the three decades. After adjusting for latitude, altitude, and cloud cover, the average UVB radiation level was estimated for each state for the period 1974–1987. A regression model was fitted with state-specific melanoma (skin), nonmelanoma skin cancer and non-Hodgkin lymphoma age-adjusted mortality rates as dependent variables. The higher the estimated level of UVB, the higher the predicted level of mortality for both melanoma and nonmelanoma skin cancers. However, in contrast to the report by Bentham, the geographic variations in mortality rates for non-Hodgkin lymphoma in the United States were not explained by average estimated state-specific levels of UVB radiation. Thus, in the populationbased studies of Newton (1995), Hartge et al. (1996), Freedman et al. (1997), Adami et al. (1999), and Hu et al. (2004), ambient levels of UVB radiation did not exhibit the same ecologic relationship to reported mortality rates or incidence rates for non-Hodgkin lymphoma as was evident for melanoma and nonmelanoma skin cancer mortality rates. Prudent reduction in intermittent and cumulative sun exposure particularly during the first two decades of life is an effective way to prevent skin cancer. Public education programs have focused on reducing outdoor exposures between 10 am and 3 pm, and regularly using physical barriers such as hats with wide brims and tightly woven clothing, and chemical sunscreens. Regular use of broad-spectrum sunscreens with a sun protection factor of at least 15 will filter out over 90% of the UVB and a substantial amount of UVA. The effectiveness of use of sunscreens in preventing skin cancer as evaluated in observational studies has been controversial because of confounding by behavioral and susceptibility risk factors, inadequate consideration of latency and cohort effects, and changing formulations for sunscreens in the part decade. Skin cancer screening examinations in the general population will ensure the early detection and treatment of precancerous lesions and skin cancer. Regular examinations are especially recommended for individuals with a history of skin cancer, multiple keratoses or dysplastic nevi, severely sun-damaged skin, genetic predisposition, immunologic impairment, or significant occupational exposure to chemical carcinogens (e.g., polycyclic aromatic hydrocarbons, inorganic arsenic as pesticides), exposure to psoralens used in combination with UVA in the management of psoriasis, or ionizing radiation.

Viral Oncogenesis and Immunodeficiency Cancers of the lower anogenital tract, namely the uterine cervix, vagina, vulva, and anus account for about 3%–4% of all newly diagnosed cancers in US women, with the most common site being the uterine cervix. Women treated previously for uterine cervix cancer were at increased risk for multiple primary cancers in the lower anogenital tract. In a cohort study based on the Michigan cancer registry, women diagnosed with an index primary invasive squamous cell carcinoma of the uterine cervix experienced more than a 40-fold increase in relative risk of a second primary carcinoma of the vagina (SIR = 40.5; 95% CI: 11.0, 103.7), and significant increases in relative risk of second primary cancers of the urinary bladder (SIR = 5.1; 95% CI: 1.7, 11.9) and lung and bronchus (SIR = 2.7; 95% CI: 1.5, 4.7) (Fisher et al., 1997). Rabkin et al. (1992) after combining data from the Connecticut cancer registry and the Surveillance, Epidemiology, and End Results (SEER) program of the National Cancer Institute reported that women with uterine cervical cancer were at significantly higher risk for cancers of the vagina and vulva (relative risk = 5.6), anus (relative risk = 4.6), urinary bladder (relative risk =

2.7), lung and bronchus (relative risk = 3.0), oral cavity (relative risk = 2.2) and larynx (relative risk = 3.3). Bergfeldt et al. (1995) observed in an analysis of the Stockholm-Gotland cancer registry that Swedish women with uterine cervical cancer experienced elevated and sustained increases in relative risk for cancer of the vulva, lung and bronchus, and urinary bladder. In the United States, the average annual age-adjusted incidence of cancer of the anal region (i.e., anal canal and anal margin) is 0.6 per 100,000. During the past decade, in the United States, Sweden, and Denmark, an increasing incidence of squamous cell and transitional cell carcinomas of the anal region has been described in women and in homosexual males. Cooper et al. (1979) hypothesized that the epidemiology of anal cancer may be viewed as that of a sexually transmitted disease, and Scholefield et al. (1989) underscored the similarities with the natural history of cervical intraepithelial neoplasia. Peters et al. (1984) reviewed the similarities in the descriptive epidemiology of squamous and transitional cell carcinomas of the uterine cervix, vagina, vulva, anus, and penis, namely increasing incidence with decreasing social class, lower risk among Jews, and elevated risk among separated and divorced persons of both sexes. Risk factors for anal cancer are correlated with sexual behavior, which may include a history of sexually transmitted infectious diseases; homosexual preference; marital instability or multiple sexual partners; homosexual or heterosexual receptive anal intercourse; anogenital condylomata or genital warts; and chronic anal inflammatory conditions such as proctitis, fissures, or fistulas. Current cigarette smoking and cumulative exposure of 20 pack-years or more, and immunosuppression, as in transplant patients or in patients with human immunodeficiency virus (HIV) infection, are associated independently with an increased risk of anal neoplasia (Melbye and Sprogel, 1991; Frisch et al., 1997). Similarly, the high-risk profile for cancer of the penis encompasses behavioral risk factors and conditions associated with poor personal hygiene and chronic penile inflammation, the absence of circumcision, the number of sexual partners and the sexual transmission of an infectious agent, cigarette smoking, and host immune factors. The lower anogenital tract is composed of a contiguous surface of epithelium derived embryologically from the urogenital sinus and cloacal entoderm. The epidemiology of independent and multifocal cancers occurring in this region may be reviewed collectively to gain insights into common risk factors and pathogenic mechanisms. Although there is compelling epidemiologic evidence that human papillomaviruses (HPVs) are the major causal agents in uterine cervical carcinoma, HPV infections have also been implicated in cancers of other anogenital sites (vulva, particularly basaloid and warty tumors, vagina, perineum, anus, and penis), and potentially as a cofactor in oral cavity, laryngeal, and esophageal cancers (IARC, 1995; Lowy, 1995; Madeleine et al., 1997; Alani and Munger, 1998). More than 50% of squamous cell carcinomas in the palatine tonsils and base of the tongue contain oncogenic human papillomavirus DNA. HPVs are DNA viruses that infect the basal, replicating keratinocytes of stratified squamous epithelium and the squamocolumnar mucosal epithelium. Among the more than 90 different subtypes of HPV, types 16, 18, 31, 33, and 45 are associated with 80% of the anogenital cancers that are prevalent worldwide. Adenocarcinomas of the uterine cervix are more likely to contain HPV 18 than other types. With progression to invasive cancer, the HPV E1 or E2 genome integrates within the host cellular DNA, which is accompanied by the dysregulated expression of the E6 and E7 viral oncoproteins. The oncoproteins are expressed in neoplastic transformation and progression by abrogating the G1-S and G2-M transition checkpoint functions of tumor suppressor proteins—E6 with p53 and E7 with the retinoblastoma gene, Rb. The inactivation of p53 and RB proteins distorts the normal functions of cyclins and cyclin-dependent kinases in regulating DNA replication and mitosis. Cell-mediated immunity is fundamental in detecting and destroying virus-infected cells. HPV infections are clinically persistent and aggressive in immunocompromised hosts, as demonstrated in organ transplant recipients who receive immunosuppressive therapy, patients infected with HIV or in patients with epidermodysplasia verruciformis (EV), an autosomal recessive disorder. EV is characterized by onset

Multiple Primary Cancers in childhood of cutaneous lesions of flat warts, red plaques, and lichenification in which HPV of various types (5,8,36,49) may be identified. Subsequently in early adult life, and potentiated by impaired Tcell–mediated immunity, actinic keratoses and squamous cell cancers of the skin appear on sun-exposed areas. HPV-associated cancers in immunosuppressed transplant patients or in HIV-seropositive patients consist of high-grade squamous intraepithelial and invasive cutaneous and anogenital lesions. Thus, immune regulation plays an important role in controlling the progression of HPV infection to dysplasia and intraepithelial neoplasia, and invasive squamous cell carcinoma (Carson et al., 1986; Klein et al., 1994; Majewski and Jablonska, 1995; Larson et al., 1997; Sun et al. 1997). Preventive measures for HPV-associated anogenital carcinomas would consist of counseling about lifestyle risk factors, periodic screening in sexually active individuals, and the future development of an effective vaccine for HPV oncogenic serotypes. Screening for exfoliative cytology (Papanicolaou smear), or with the use of a PCR method to amplify DNA from uterine cervical scrapings, would facilitate identification and eradication of intraepithelial neoplastic lesions and persistent infection with oncogenic serotypes of HPV. Anoscopy and flexible proctosigmoidoscopy would reveal chronic inflammatory and pre-cancerous lesions and carcinomas of the anal canal and anal margin. The American Cancer Society currently recommends that all currently or previously sexually active women have an annual Pap test and pelvic examination. After a woman has had three or more consecutively normal annual examinations, the PAP test may be performed less frequently at the physician’s discretion. Counseling about sexual practices should require use of a condom in HPV-infected individuals. Condom use will not completely prevent transmission, however, because genital HPV infections in the male are not limited to sources on the glans, foreskin, urethral meatus, and shaft, but may spread from scrotal, perineal, and perianal lesions. Human papillomavirus immunization may provide ultimately an effective primary preventive strategy in addition to surgical excision, cryotherapy, laser vaporization, or topical chemotherapy of HPVpositive anogenital lesions. A rationale for preventive immunization would be to target the HPV E6 and E7 genes, which are constitutively expressed in HPV-associated cancers. Experiments using antisense oligonucleotides that target the E6 and E7 oncoproteins in combination have shown growth inhibition in human tumor cell lines. While cellmediated immunity is needed to destroy virus-infected cells, antibodies are important in neutralizing extracellular virus. Effective neutralizing antibodies are often directed against viral surface capsid antigens. Virus-like particle vaccines are composed of non-infectious capsid protein of HPV. Immunization with virus-like capsid particles has induced significant titers of neutralizing type-specific antibodies to HPVs in rabbits, cows, and dogs. In the clinical trial, vaccination with the HPV 16 virus-like particle vaccine conferred protection against persistent infection (Koutsky et al., 2002). The future utility of various DNA vaccines will be explored in stimulating cell-mediated immunity, which is fundamentally important in controlling or inhibiting HPV-induced neoplastic transformation (McDonnell and Askari, 1996).

Late Effects of Treatment of a Previous Primary Cancer There is experimental and epidemiological evidence for the carcinogenic toxicity of antineoplastic agents (Dedrick and Morrison, 1992). The carcinogenic potential of chemotherapeutic agents, in particular the DNA-alkylating agents, may be enhanced when administered in conjunction with ionizing radiation (Kaldor and Lasset, 1991). The chemotherapeutic activity and toxicity of such DNA-alkylating compounds as nitrogen mustard, cyclophosphamide, melphalan, procarbazine, and the nitrosoureas are mediated through perturbation of the fundamental mechanisms concerned with cell replication and differentiation. The capacity of these agents to alter normal mitosis, cellular function, and terminal differentiation provides the basis for therapeutic efficacy, mutagenicity, and potential carcinogenicity. There is preliminary evidence that cancer patients who develop

1277

chemotherapy-induced second primary cancers have significantly less efficient DNA repair capacity than patients undergoing the same treatment but not developing second primary cancers (Bohr, 1992). Survivors of childhood cancer are at a threefold to sixfold increased risk of developing a second primary cancer, when compared with the general population, under 15 years of age. The determinants of neoplastic sequelae in childhood cancer survivors include exposures to high-dose radiation therapy and specific categories of chemotherapeutic agents. In a more limited context are hereditary syndromes associated with markedly increased sensitivity to the genotoxic effects of cancer therapeutic agents such as the Li-Fraumeni syndrome, retinoblastoma, Gorlin syndrome, and ataxia telangiectasia. The second primary cancers in children more commonly arise in hematolymphopoietic tissues (acute leukemias, myelodysplastic syndrome) and soft or connective tissue (sarcomas). Type and intensity of radiation, the intrinsic susceptibility of the exposed tissues, and host factors influence the magnitude and pattern of risks for radiation-induced cancers. The risk is highest when the exposure occurs at a younger age, and for non-hematologic solid tumors, increases after a protracted latency interval, and with a magnitude of increased relative risk that is multiplicative with the background age-specific incidence rates. Radiation-associated bone tumors and sarcomas develop within the exposure field or margins of irradiation, typically after a latency period of 10 or more years. Following radiation treatment in childhood cancer patients, tumors of the thyroid, breast, bone, soft tissue, and brain have been reported with excessive frequency. Patients with hereditary retinoblastoma, receiving radiation therapy, are at increased risk of osteosarcomas and leukemias. Secondary myelodysplasia and acute myelogenous leukemia have been linked with alkylating agents (e.g., melphalan), topoisomerase II inhibitors (e.g., epipodophyllotoxin), and anthracyclines (e.g., doxorubicin). Risks of leukemia increase with cumulative exposure dose, peaking between 4 and 6 years after the onset of exposure. The cumulative incidence is generally less than 5%. The toxicity of alkylating agents is mediated by transferring alkyl groups by covalent bonding with the 7N or 6O positions on guanine base positions in DNA. The topoisomerase II inhibitors block the unwinding of supercoiled DNA during DNA cleavage. The genomes of eukaryotes typically encode four or more topoisomerases. Topoisomerase II enzymes function predominately after replication during separation of daughter DNA strands or the remodeling of chromatin structure. Major mechanisms of cytotoxicity for doxorubicin are the destruction of topoisomerase II, and by intercalation in double-stranded DNA, producing structural changes that interfere with DNA and RNA synthesis. In adult patients, the analysis of the risks of second primary acute leukemias, non-Hodgkin lymphomas, carcinomas, and other sarcomas in patients with Hodgkin disease is based upon the assessment of the long-term sequelae of various treatment regimens, inherent features of the natural history of a malignant lymphoma, predisposing genetic or altered host immune factors, exposure to other environmental lifestyle risk factors, and target tissue susceptibilities. The risk-benefit evaluation of the therapeutic agent or regimen should take into account the likely outcome of the disease if untreated, and whether or not alternative methods of treatment are more efficacious without incurring serious long-term adverse effects (Canellos et al., 1992). Among the second primary cancers following treatment for Hodgkin disease, acute nonlymphocytic leukemias and non-Hodgkin lymphomas are the most common. Patients with Hodgkin disease (HD) have a twofold to threefold increased risk of second primary cancers of various types, particularly in the subgroup with intensive cyclical combination chemotherapy. Second primary lung and breast cancers have been reported to be increased as a result of intensive radiation therapy. A cohort of 28,462 patients with HD who were diagnosed over a period of 30 years beginning in the 1950s in the Nordic countries and in four provinces in Canada, contributed 117,574 person-years of observation. The standardized incidence ratio (SIR) of all second primary cancers was 1.8 (95% CI: 1.6–1.9) and for leukemia was 10.0 (95% CI: 8.6–13.0). The SIR for non-Hodgkin lymphoma (NHL) was 3.0 (95% CI: 1.9–4.5), and for breast cancer, 1.4 (95% CI: 1.1–1.8). The results of other studies in patients with HD or NHL are

1278

PART IV: CANCER BY TISSUE OF ORIGIN

Table 66–2. Age-Standardized Incidence Ratios (SIR) of All Sites, Leukemias, Breast Cancer in Patients with Hodgkin Disease All Sites Study Reference Kaldor et al. (1987) Henry-Amar et al. (1992) Boivin et al. (1995) Dores et al. (2002)

Leukemias

Breast Cancer

Study Period

Obs

SIR (95% CI)

Obs

SIR (95% CI)

Obs

SIR (95% CI)

1950–1980 1960–1987 1940–1987 1935–1994

711 631 521 2,153

1.8 (1.6–1.9) 2.9 (2.6–3.1) 2.7 (2.5–3.0) 2.3 (2.2–2.4)

106 158 122 249

10.0 (8.6–13.0) 28.0 (24.0–32.0) 24.0 (20.0–29.0) 9.9 (8.7–11.2)

62 39 39 234

1.4 (1.1–1.8) 1.5 (1.1–2.1) 1.4 (1.0–2.0) 2.0 (1.8–2.3)

The estimated excess risk of leukemia per 10,000 per year was 8.8 (all types), 6.3 (acute nonlymphocytic leukemia); the estimated excess risk for all solid tumors was 33.1, and for breast cancer 10.5.

Table 66–3. Relative and Absolute Site-Specific Risks of Selected Sites of Second Primary Cancers in Patients with Non-Hodgkin Lymphoma Reported to the SEER* Program, 1978–1998 Cancer Site All sites Female breast Lung Melanoma (skin) Hodgkin disease Multiple myeloma Acute myelogenous leukemia Kaposi sarcoma

Observed 4,638 393 876 158 49 34 89 106

Observed to Expected Ratio (95% CI)

Absolute Excess Risk per 10,000 per Year

1.1 (1.1–1.2) 0.8 (0.7–0.9) 1.3 (1.3–1.4) 1.6 (1.4–1.9) 4.5 (3.3–5.9) 0.6 (0.4–0.9) 3.4 (2.7–4.1)

19.3 -7.1 7.4 2.1 1.3 -0.6 2.9

13.0 (10.6–15.7)

3.3

*SEER, Surveillance, Epidemiology and End Results program, National Cancer Institute.

summarized in Table 66–2 and 66–3 (Kaldor et al., 1987; Henry-Amar, 1992; Boivin et al., 1995; Dores et al., 202). The non-Hodgkin lymphomas occur with a cumulative incidence in the range of 1%–2%, with a latency interval after initial treatment of 10 or more years. Most of the secondary non-Hodgkin lymphomas are classified as B-cell or Burkitt-like lymphomas, which have molecular fingerprints of Epstein Barr virus in less than 20%. Radiation-induced carcinomas and sarcomas may appear after a latency interval exceeding 10 years. With protracted follow-up it has become apparent that the absolute excess risk of treatment-induced carcinomas and sarcomas exceeds that of the leukemias. Dores et al. (2002) reported that 25 years after HD diagnosis, the actuarial risk of developing a solid tumor was 21.9%. In the study of young women with HD of Travis et al. (2003), a radiation dose of 40 Gy delivered to the breast was associated with a 3.2-fold (95% CI: 1.4–8.2) increased risk. The risk increased to eightfold (95% CI: 2.6–26.4) with a dose exceeding 40 Gy. In women surviving with Hodgkin lymphoma who were treated at age 25 years with more than 40 Gy to the mediastinum, without receiving alkylating agents, the estimated cumulative absolute risk of breast cancer at age 45 years was 11.1% (95% CI = 7.4% to 16.3%), and at age 55 years, 29.0% (95% CI = 20.2% to 40.1%) (Travis et al., 2005). With the use of chemotherapy, or irradiation of the ovaries, the risk of breast cancer was reduced by 40%. Thus physiologic ovarian function and cumulative radiation dose of breast tissue are independent determinants of risk. Current philosophy of treatment of HD involves combination chemotherapy, with consolidation radiation therapy delivered with reduced rather than extended field sizes, or without radiation therapy (Longo, 2005). References Adami HO, Bergstrom R, Weiderpass E, et al. 1997. Risk for endometrial cancer following breast cancer: A prospective study in Sweden. Cancer Causes Control 8:821–827. Adami J, Gridley G, Nyren O, et al. 1999. Sunlight and non-Hodgkin’s lymphoma: A population-based cohort study in Sweden. Int J Cancer 80:641–645.

Alani RM, Munger K. 1998. Human papillomaviruses and associated malignancies. J Clin Oncol 16:330–337. American Society of Human Genetics Ad Hoc Committee on Breast and Ovarian Cancer Screening. 1994. Am J Hum Genet 55:1–4. Bale AE. 2002. Hedgehog signaling and human disease. Annu Rev Genomics Hum Genet 3:47–65. Barlow DP. 1993. Methylation and imprinting: From host defense to gene regulation? Science 260:309–310. Bedi GC, Westra WH, Gabrielson E, Koch W, Sidransky D. 1996. Multiple head and neck tumors: Evidence for a common clonal origin. Cancer Res 56:2484–2487. Bentham G. 1996. Association between incidence of non-Hodgkin’s lymphoma and solar ultraviolet radiation in England and Wales. BMJ 312:1128– 1131. Berg JW, Hajdu SI, Foote FW, Jr. 1971. The prevalence of latent cancers in cancer patients. Arch Pathol 91:183–186. Bergfeldt K, Einhorn S, Rosendahl I, Hall P. 1995. Increased risk of second primary malignancies in patients with gynecological cancer. A Swedish record-linkage study. Acta Oncol 34:771–777. Bergkvist L, Adami HO, Persson I, Hoover R, Schairer C. 1989. The risk of breast cancer after estrogen and estrogen-progestin replacement. N Engl J Med 321:293–297. Berry DA, Parmigiani G, Sanchez J, Schildkraut J, Winer E. 1997. Probability of carrying a mutation of breast-ovarian cancer gene BRCA1 based on family history. J Natl Cancer Inst 89:227–238. Blackwood MA, Weber BL. 1998. BRCA1 and BRCA2: From molecular genetics to clinical medicine. J Clin Oncol 16:1969–1977. Blot WJ, McLaughlin JK, Winn DM, et al. 1988. Smoking and drinking in relation to oral and pharyngeal cancer. Cancer Res 48:3282–3287. Bohr VA. 1992. Intragenomic DNA repair: Molecular and clinical considerations. In: Green DM, D’Angio GJ, eds. Late Effects of Treatment For Childhood Cancer. New York: Wiley-Liss, pp. 103–111. Boice JD, Storm H, Curtis R. 1985. Multiple primary cancer in Connecticut and Denmark. NCI Monograph 68: NIH Publication no. 85-2714. Boivin JF, Hutchison GB, Zauber AG, et al. 1995. Incidence of second cancers in patients treated for Hodgkin’s disease. J Natl Cancer Inst 87:732–741. Canellos GP, Anderson JR, Propert KJ, et al. 1992. Chemotherapy of advanced Hodgkin’s disease with MOPP, ABVD, or MOPP alternating with ABVD. N Engl J Med 327:1478–1484. Carey TE. 1996. Field cancerization: Are multiple primary cancers monoclonal or polyclonal? Ann Med 28:183–188. Carson LF, Twiggs LB, Fukushima M, et al. 1986. Human genital papilloma infections: An evaluation of immunologic competence in the genital neoplasia-papilloma syndrome. Am J Obstet Gynecol 155:784–789. Cartwright R, McNally R, Staines A. 1994. The increasing incidence of non-Hodgkin’s lymphoma (NHL): The possible role of sunlight. Leuk Lymphoma 14:387–394. Centers for Disease Control and Prevention. 1992. Cigarette smoking among adults—United States—1990. 361–362. Colditz GA. 1998. Relationship between estrogen levels, use of hormone replacement therapy, and breast cancer. J Natl Cancer Inst 90:814–823. Cooper HS, Patchefsky AS, Marks G. 1979. Cloacogenic carcinoma of the anorectum in homosexual men: An observation of four cases. Dis Colon Rectum 22:557–558. Curtis RE, Hoover RN, Kleinerman RA, Harvey EB. 1985. Second cancer following cancer of the female genital system in Connecticut, 1935–82. Natl Cancer Inst Monogr 68:113–137. Day GL, Blot WJ, Shore RE, et al. 1994. Second cancers following oral and pharyngeal cancers: role of tobacco and alcohol. J Natl Cancer Inst 86:131–137. Dedrick RL, Morrison PF. 1992. Carcinogenic potency of alkylating agents in rodents and humans. Cancer Res 52:2464–2467.

Multiple Primary Cancers Dores GM, Metayer C, Curtis RE, et al. 2002. Second malignant neoplasms among long-term survivors of Hodgkin’s disease: A population-based evaluation over 25 years. J Clin Oncol 20:3484–3494. Draper GJ, Sanders BM, Kingston JE. 1986. Second primary neoplasms in patients with retinoblastoma. Br J Cancer 53:661–671. Easton DF, Bishop DT, Ford D, Crockford GP. 1993. Genetic linkage analysis in familial breast and ovarian cancer: Results from 214 families. The Breast Cancer Linkage Consortium. Am J Hum Genet 52:678–701. Evans DG, Fentiman IS, McPherson K, et al. 1994. Familial breast cancer. BMJ 308:183–187. Ewertz M. 1988. Influence of non-contraceptive exogenous and endogenous sex hormones on breast cancer risk in Denmark. Int J Cancer 42:832–838. Fearon ER. 1997. Human cancer syndromes: Clues to the origin and nature of cancer. Science 278:1043–1050. Fialkow PJ. 1974. The origin and development of human tumors studied with cell markers. N Engl J Med 291:26–35. Fiore MC, Novotny TE, Pierce JP, et al. 1989. Trends in cigarette smoking in the United States. The changing influence of gender and race. JAMA 261:49–55. Fisher G, Harlow SD, Schottenfeld D. 1997. Cumulative risk of second primary cancers in women with index primary cancers of uterine cervix and incidence of lower anogenital tract cancers, Michigan, 1985–1992. Gynecol Oncol 64:213–223. Flanders WD, Rothman KJ. 1982. Interaction of alcohol and tobacco in laryngeal cancer. Am J Epidemiol 115:371–379. Flannery JT, Boice JD, Jr., Devesa SS, et al. 1985. Cancer registration in Connecticut and the study of multiple primary cancers, 1935–82. Natl Cancer Inst Monogr 68:13–24. Ford D, Easton DF, Bishop DT, Narod SA, Goldgar DE. 1994. Risks of cancer in BRCA1-mutation carriers. Breast Cancer Linkage Consortium. Lancet 343:692–695. Freedman DM, Zahm SH, Dosemeci M. 1997. Residential and occupational exposure to sunlight and mortality from non-Hodgkin’s lymphoma: Composite (threefold) case-control study. BMJ 314:1451–1455. Frisch M, Glimelius B, van den Brule AJ, et al. 1997. Sexually transmitted infection as a cause of anal cancer. N Engl J Med 337:1350–1358. Frisch M, Hjalgrim H, Olsen JH, Melbye M. 1996. Risk for subsequent cancer after diagnosis of basal-cell carcinoma. A population-based, epidemiologic study. Ann Intern Med 125:815–821. Frisch M, Melbye M. 1995. New primary cancers after squamous cell skin cancer. Am J Epidemiol 141:916–922. Gailani MR, Leffell DJ, Ziegler A, et al. 1996. Relationship between sunlight exposure and a key genetic alteration in basal cell carcinoma. J Natl Cancer Inst 88:349–354. Garro AJ, Lieber CS. 1990. Alcohol and cancer. Annu Rev Pharmacol Toxicol 30:219–249. Goldgar DE, Reilly PR. 1995. A common BRCA1 mutation in the Ashkenazim. Nat Genet 11:113–114. Hajdu SI, Hajdu EO, 1968. Multiple primary malignant tumors. J Am Geriatr Soc 16:16–26. Hall HI, Jamison P, Weir HK. 2001. Second primary ovarian cancer among women diagnosed previously with cancer. Cancer Epidemiol Biomarkers Prev 10:995–999. Hall JG. 1997. Genomic imprinting: Nature and clinical relevance. Annu Rev Med 48:35–44. Harris CC. 1996. Molecular epidemiology of basal cell carcinoma. J Natl Cancer Inst 88:315–317. Hartge P, Devesa SS, Grauman D, Fears TR, Fraumeni JF, Jr. 1996. NonHodgkin’s lymphoma and sunlight. J Natl Cancer Inst 88:298–300. Harvey EB, Brinton LA. 1985. Second cancer following cancer of the breast in Connecticut. NCI Monograph 68; NIH Publication No. 85-2714: 99–112. Henry-Amar M. 1992. Second cancer after the treatment for Hodgkin’s disease: A report from the International Database on Hodgkin’s Disease. Ann Oncol 3 Suppl 4:117–128. Hu S, Ma F, Collado-Mesa F, Kirsner RS. 2004. Ultraviolet radiation and incidence of non-Hodgkin’s lymphoma among Hispanics in the United States. Cancer Epidemiol Biomarkers Prev 13:59–64. Hunter DJ, Colditz GA, Stampfer MJ, et al. 1990. Risk factors for basal cell carcinoma in a prospective cohort of women. Ann Epidemiol 1:13–23. International Agency for Research on Cancer. 1988. Alcohol Drinking. Lyon, France; Volume 44. International Agency for Research on Cancer. 1995. Human Papillomaviruses. IARC Monograph on the Evaluation of Carcinogenic Risk to Humans. Lyon, France; Volume 64. Kaldor JM, Day NE, Band P, et al. 1987. Second malignancies following testicular cancer, ovarian cancer and Hodgkin’s disease: An international collaborative study among cancer registries. Int J Cancer 39:571–585.

1279

Kaldor JM, Lasset C. 1991. Cytotoxic chemotherapy for cancer. In: Coleman MP, ed. Cancer Risk After Medical Treatment. New York: Oxford University Press, pp. 50–70. Khuri FR, Lippman SM, Spitz MR, Lotan R, Hong WK. 1997. Molecular epidemiology and retinoid chemoprevention of head and neck cancer. J Natl Cancer Inst 89:199–211. Kim ES, Hong WK, Khuri FR. 2002. Chemoprevention of aerodigestive tract cancers. Annu Rev Med 53:223–243. Klein RS, Ho GY, Vermund SH, Fleming I, Burk RD. 1994. Risk factors for squamous intraepithelial lesions on Pap smear in women at risk for human immunodeficiency virus infection. J Infect Dis 170:1404–1409. Koutsky LA, Ault KA, Wheeler CM, et al. 2002. A controlled trial of a human papillomavirus type 16 vaccine. N Engl J Med 347:1645–4651. Kripke ML. 1994. Ultraviolet radiation and immunology: Something new under the sun—presidential address. Cancer Res 54:6102–6105. Larson AA, Liao SY, Stanbridge EJ, Cavenee WK, Hampton GM. 1997. Genetic alterations accumulate during cervical tumorigenesis and indicate a common origin for multifocal lesions. Cancer Res 57:4171– 4176. Levine AJ. 1997. p53, the cellular gatekeeper for growth and division. Cell 88:323–331. Lieber CS. 1988. Biochemical and molecular basis of alcohol-induced injury to liver and other tissues. N Engl J Med 319:1639–1650. Lieber CS. 1995. Medical disorders of alcoholism. N Engl J Med 333:1058– 1065. Lindor NM, Greene MH. 1998. The concise handbook of family cancer syndromes. Mayo Familial Cancer Program. J Natl Cancer Inst 90:1039– 1071. Longo DL. 2005. Radiation therapy in Hodgkin disease: why risk a Pyrrhic victory? J Natl Cancer Inst 97:1394–1395. Lowy DR. 1995. Human papillomaviruses and epithelial cancer. In: Mukhtar H, ed. Skin Cancer: Mechanisms and Human Relevance. Boca Raton: CRC Press, pp. 351–358. Madeleine MM, Daling JR, Carter JJ, et al. 1997. Cofactors with human papillomavirus in a population-based study of vulvar cancer. J Natl Cancer Inst 89:1516–1523. Majewski S, Jablonska S. 1995. Epidermodysplasia verruciformis as a model of human papillomavirus-induced genetic cancer of the skin. Arch Dermatol 131:1312–1318. Mayne ST. 1997. Antioxidant nutrients and cancer incidence and mortality: An epidemiologic perspective. Adv Pharmacol 38:657–675. McDonnell WM, Askari FK. 1996. DNA vaccines. N Engl J Med 334:42–45. McLaughlin JK, Gridley G, Block G, et al. 1988. Dietary factors in oral and pharyngeal cancer. J Natl Cancer Inst 80:1237–1243. McMichael AJ, Giles GG. 1996. Have increases in solar ultraviolet exposure contributed to the rise in incidence of non-Hodgkin’s lymphoma? Br J Cancer 73:945–950. Melbye M, Sprogel P. 1991. Aetiological parallel between anal cancer and cervical cancer. Lancet 338:657–659. Moertel CG, Dockerty MB, Baggenstoss AH. 1961. Multiple primary malignant neoplasms. I. Introduction and presentation of data. Cancer 14: 221–230. Newcomb PA, Trentham-Dietz A. 2003. Patterns of postmenopausal progestin use with estrogen in relation to endometrial cancer (United States). Cancer Causes Control 14:195–201. Newton R. 1995. Non-Hodgkin’s lymphoma and skin cancer. American data refute ultraviolet hypothesis. BMJ 311:750–751. Olsen JH, Garwicz S, Hertz H, et al. 1993. Second malignant neoplasms after cancer in childhood or adolescence. Nordic Society of Paediatric Haematology and Oncology Association of the Nordic Cancer Registries. BMJ 307:1030–1036. Omenn GS. 1998. Chemoprevention of lung cancer: the rise and demise of beta-carotene. Annu Rev Public Health 19:73–99. Peters RK, Mack TM, Bernstein L. 1984. Parallels in the epidemiology of selected anogenital carcinomas. J Natl Cancer Inst 72:609–615. Peto R, Lopez AD, Boreham J, Thun M, Heath C, Jr. 1992. Mortality from tobacco in developed countries: Indirect estimation from national vital statistics. Lancet 339:1268–1278. Powles TJ, Hickish T. 1995. Tamoxifen therapy and carcinogenic risk. J Natl Cancer Inst 87:1343–1345. Pritchard KI. 1998. Is tamoxifen effective in prevention of breast cancer? Lancet 352:80–81. Rabkin CS, Biggar RJ, Melbye M, Curtis RE. 1992. Second primary cancers following anal and cervical carcinoma: Evidence of shared etiologic factors. Am J Epidemiol 136:54–58. Ries LAG, Eisner MP, Kosary CL, Hankey BF, Miller BA, Clegg L, et al. 2003. SEER cancer statistics review, 1975–2000. Bethesda, MD, National Cancer Institute.

1280

PART IV: CANCER BY TISSUE OF ORIGIN

Robson M, Gilewski T, Haas B, et al. 1998. BRCA-associated breast cancer in young women. J Clin Oncol 16:1642–1649. Sankila R, Pukkala E, Teppo L. 1995. Risk of subsequent malignant neoplasms among 470,000 cancer patients in Finland, 1953–1991. Int J Cancer 60:464–470. Schoenberg BS. 1977. Multiple primary malignant neoplasms. The Connecticut experience, 1935–1964. Recent Results Cancer Res 58:1–173. Schoenberg BS, Greenberg RA, Eisenberg H. 1969. Occurrence of certain multiple primary cancers in females. J Natl Cancer Inst 43:15–32. Scholefield JH, Sonnex C, Talbot IC, et al. 1989. Anal and cervical intraepithelial neoplasia: possible parallel. Lancet 2:765–769. Schottenfeld D. 1996. Basal-cell carcinoma of the skin: A harbinger of cutaneous and noncutaneous multiple primary cancer. Ann Intern Med 125:852–854. Schottenfeld D, Berg J. 1971. Incidence of multiple primary cancers. IV. Cancers of the female breast and genital organs. J Natl Cancer Inst 46:161–170. Schottenfeld D, Berg J. 1975. Epidemiology of multiple primary cancers. In: Schottenfeld D, ed. Cancer Epidemiology and Prevention: Current Concepts. Springfield: Charles C. Thomas, pp. 416–434. Slaughter DP, Southwick HW, Smejkal W. 1953. “Field cancerization” in oral stratified squamous epithelium: Clinical implications of multicentric origin. Cancer 6:963–968. Steinmetz KA, Potter JD. 1991a. Vegetables, fruit, and cancer. I. Epidemiology. Cancer Causes Control 2:325–357. Steinmetz KA, Potter JD. 1991b. Vegetables, fruit, and cancer. II. Mechanisms. Cancer Causes Control 2:427–442.

Sun XW, Kuhn L, Ellerbrock TV, et al. 1997. Human papillomavirus infection in women infected with the human immunodeficiency virus. N Engl J Med 337:1343–1349. Tamoxifen for early breast cancer: An overview of the randomised trials. 1998. Early Breast Cancer Trialists’ Collaborative Group. Lancet 351:1451– 1467. Thun MJ, Day-Lally CA, Calle EE, Flanders WD, Heath CW, Jr. 1995. Excess mortality among cigarette smokers: Changes in a 20-year interval. Am J Public Health 85:1223–1230. Travis LB, Hill DA, Dores GM, et al. 2003. Breast cancer following radiotherapy and chemotherapy among young women with Hodgkin Disease. JAMA 290:465–475. Travis LB, Hill DA, Dores GM, et al. 2005. Cumulative absolute breast cancer risk for young women treated for Hodgkin lymphoma. J Natl Cancer Inst 97:1428–1437. United States Department of Health and Human Services. 1990. The health benefits of smoking cessation: A report of the Surgeon General. DHHS (CDC) 90-8416. Warner KE. 1986. Smoking and health implications of a change in the federal cigarette excise tax. JAMA 255:1028–1032. Warren S, Ehrenreich T. 1944. Multiple primary malignant tumors and susceptibility to cancer. Cancer Res 4:554–570. Warren S, Gates O. 1932. Multiple primary malignant tumors: survey of the literature and statistical study. Am J Cancer 16:1358–1414. Weiderpass E, Adami HO, Baron JA, et al. 1999. Risk of endometrial cancer following estrogen replacement with and without progestins. J Natl Cancer Inst 91:1131–1137.

V CANCER PREVENTION AND CONTROL

This page intentionally left blank

67

Principles and Applications of Cancer Prevention and Control Interventions ROBERT A. HIATT AND BARBARA K. RIMER

R

ecent discoveries in genetics and molecular biology are dramatically advancing our knowledge about the basic processes of carcinogenesis and will undoubtedly lead to new ways to prevent, detect, and treat human cancer (Hanahan and Weinberg, 2000). However, cancer remains, to a large degree, a disease in which human behavior and societal factors play important causative roles. It has been estimated that as much as 50%–75% of cancer mortality in the United States can be attributed to external, nongenetic factors, most of which are related to human behaviors such as tobacco use, overuse of alcohol, improper diet, lack of physical activity, overexposure to sunlight, and sexual activity leading to exposure to certain viruses (McGinnis and Foege, 1993; Harvard Report on Cancer Prevention, 1996). Modification of these behavioral and societal factors along with the application of biomedical discoveries are critical to reduce cancer risks, incidence, morbidity, and mortality, and improve quality of life. Modification of the changeable causative precursors is a major focus of cancer prevention and control. Cancer control research is the conduct of basic and applied research in the behavioral, social, and population sciences that, independently, or in combination with biomedical approaches, reduces cancer risk, incidence, morbidity, and mortality and improves quality of life (Hiatt and Rimer, 1999). In this chapter, we set forth the principles of cancer prevention and control sciences, with a special emphasis on intervention. We begin with some historical background from which these principles have evolved. We then present an overview of cancer control interventions and some accomplishments in cancer prevention and control to date. We also describe a cancer control research framework that has been useful over the past decade in strategic planning and in directing research and application programs at a national level. Finally, placing the practice of cancer control research within this framework, we highlight where modern cancer prevention and control science is likely to have the largest impact on cancer at both the individual and population levels (Hiatt and Rimer, 1999; Rimer, 1999; Best et al., 2003).

HISTORY OF CANCER PREVENTION AND CONTROL The term “cancer control” has long been a source of confusion and debate. This is partially because the scope of cancer control has changed, as it should, with the emergence of new knowledge and shifts in the patterns of cancer and its determinants. The earliest reference to “cancer control,” of which we are aware, appeared in 1913 with the formation of the American Society for the Control of Cancer, which became the American Cancer Society (ACS) in 1945 (New York City Cancer Committee, 1994). The driving force behind the formation of this organization was concern about the need for the early surgical treatment of cervical cancer, a cancer that was often fatal in the early 1900s. The first cancer control recommendations in 1913 presaged much of current-day activity by including a call for cancer registration and the analysis of vital statistics, the study of the geographic distribution of disease and the influence of diet, and education to provide “a concise outline of accepted cancer facts” (New York City Cancer Committee, 1994). In the first decades of this century, however, the medical profession did not readily accept the idea of cancer control. Cancer was much less common than it is now and its outcome was

usually dismal (Breslow, 1977). Furthermore, the medical profession was resistant to governmental interference in the practice of medicine, and organized state and federal cancer control efforts did not generally have its support (Breslow et al., 1977). Legislative language first recognized cancer control in 1937 when, with the formation of the National Cancer Institute (PL 75-224), the Surgeon General was authorized to act through the Institute and the National Cancer Advisory Council to “cooperate with state health agencies in the prevention, control, and eradication of cancer.” Congress intended that the NCI would carry out intramural research and support extramural research and cancer control, which it referred to as the “useful application of their results” (75th Congress, 1937). By the middle of the 20th century, the major focus of cancer control was the dissemination of research discoveries through communication and education. Research in cancer control per se was not yet part of the paradigm, and there was not yet a well-developed field of cancer control intervention science. Thus, much of what was disseminated missed its mark. With the passage of The National Cancer Act of 1971 (PL 92-218), Congress reaffirmed its support for cancer control and authorized specific dollar amounts for the Director of the NCI to carry out relevant programs with state and other health agencies (92nd Congress, 1971). The Division of Cancer Control and Rehabilitation, established at NCI in 1973, was the first structural unit within the NCI devoted to cancer control (Breslow et al., 1977). Creation of the new organizational entity made it possible to begin a more systematic program in cancer control. In the last quarter of the century, the science of cancer control was accelerated by the formation of the NCI’s Division of Cancer Prevention and Control in 1983, led by Peter Greenwald. The development and promulgation of a definition and framework for cancer control placed the field on a sound scientific basis and recognized the importance of research in this field. Greenwald and Cullen (1985) first defined cancer control research and advanced a framework that described a linear series of phases from hypothesis generation (Phase I) to methods development (Phase II) to controlled intervention trials (Phase III). Studies in defined populations (Phase IV) and demonstration projects (Phase V) led to nationwide prevention and health services programs (Greenwald and Cullen, 1985). The concept proposed a logical progression—from basic research through tests of intervention efficacy to community applications—that attempted to focus investigators on the need to build appropriate foundations before moving on to larger scale tests. This paradigm, which paralleled the model of drug development, effectively stimulated a generation of investigators to think more rigorously about cancer control science as a hypothesis-driven and evidence-based endeavor. The logical progression of inquiry gave cancer control researchers a focus and fostered a large body of new cancer control knowledge (Glanz, 1997; Hiatt, 1997; Kaluzny, 1997; Lerman, 1997; Lewis, 1997; Lichtenstein, 1997). Building on a decade of success with cancer prevention and control research in the United States, the National Cancer Institute of Canada’s Advisory Committee on Cancer Control (1997) advanced a more comprehensive framework that emphasized the research to practice sequence and made major contributions to the evolving US strategy (Best et al., 2003). These contributions included the principle of bringing key stakeholders to the table in the decision-making process. There

1283

1284

PART V: CANCER PREVENTION AND CONTROL

was an explicit recognition that research approaches other than clinical trials emphasized by the Greenwald-Cullen model (Greenwald and Cullen, 1985) could generate valid and useful knowledge. Also, the Canadians advanced the importance of a knowledge synthesis process at a central point in the framework that might lead to more fundamental research, to the development or refinement of programs, or simply to the fine-tuning of programs to do a better job of reaching target populations. The Canadian model recognizes that the process of moving from discovery to dissemination is not a linear one but is more reciprocal and iterative, with many feedback loops. In 1996–1997, the Cancer Control Program Review (1997) and the Cancer Prevention Program Review (1997) (National Cancer Institute, 1997; National Cancer Institute and National Institutes of Health, 1997), conducted at the request of the US NCI Director, then Richard Klausner, made recommendations that would lead to advances in the cancer prevention and control sciences. The reviewers recognized that important societal developments have a major impact on the focus of cancer prevention and control and highlighted the following trends at the turn of the century: an aging and more diverse population, the revolutions in communications and informatics, the explosion in the rate of discovery in molecular biology and genetics, and fundamental changes in the organization of health-care delivery. The Division of Cancer Control and Population Sciences (DCCPS) and the Division of Cancer Prevention (DCP) were formed from the old DCPC to explicitly recognize the focus on populations and the need to integrate behavioral research and population-public health perspectives, on the one hand, and the need for prevention sciences to build on biomedical discoveries on the other (Hiatt and Rimer, 1999). In the formation of DCCPS, Hiatt and Rimer (1999) offered a longterm review and update of US NCI cancer control research that explicitly adopted and also adapted the 1994 NCIC framework as a point of departure. Consistent with progress seen over the 15 years since the original US NCI model, the new definition added five important elements: 1. A central role for behavioral and social sciences in combination with epidemiological and health services research as important foundations for cancer control and population sciences 2. Recognition that effective cancer control does not always or exclusively occur through biomedical channels 3. Explicit statement that basic research may make important contributions to cancer control and population sciences 4. A broadening of key outcomes to include both risk reduction and quality of life, at the individual level and aggregate levels with appropriate indicator measures of health-supporting social and physical environments at the population level (e.g. in school, workplace, neighborhood, health care delivery system, community) 5. Awareness that progress does not always occur in the linear fashion outlined by the Greenwald-Cullen Five Phase Model, but often requires fundamental and applied scientists working together, providing input from other sciences through the creation of transdisciplinary teams (Abrams, 1999; Halfon and Hochstein, 2002; Turkkan et al., 2000), and must allow for a continuing (and nonlinear) re-examination of the underlying knowledge base. The new framework recognized the need to identify research-based interventions that meet standards of evidence, to broker their transfer and widespread dissemination nationwide in partnership with other stakeholders (e.g., the Centers for Disease Control and Prevention (CDC), the American Cancer Society (ACS), and state and local health departments), and to measure with “report cards” the fidelity and degree of penetration of these interventions and improve on them over time. It is within this framework that modern cancer control and prevention research is conducted.

OVERVIEW OF CANCER CONTROL INTERVENTIONS Cancer is a complex disease. It is not surprising that the interventions required to prevent cancer are also complex. Cancer control interventions include not only biomedical approaches, like chemoprevention,

the Pap test, mammography, fecal occult blood testing (FOBT), and colonoscopy, but also the policy, behavioral, and other interventions required to achieve acceptance and adoption of these medical interventions designed to improve population health. Although there are many ways to categorize cancer control interventions, we use a classification scheme developed by Rimer (1994) and employed in several subsequent analyses (Meissner et al., 1998; Rimer et al., 2000; Legler et al., 2002). We characterize interventions as falling within the following areas: 1. Access enhancing (e.g., mobile vans or the provision of transportation to appointments) 2. Individual directed (e.g., mailed reminders and telephone counseling) 3. System directed (e.g., computerized or manual prompting and reminder systems) 4. Social network (e.g., peer-group leaders) 5. Community education 6. Media campaigns 7. Multi-component interventions. Interventions can be delivered in a variety of settings—for example, to people in health-care settings and in their homes, worksites, churches, at beauty salons and other places people frequent, through community organizations, health departments or other organizations, by phone, e-mail, in person, and through combinations of these. The possibilities are nearly limitless. Broad classes of cancer control interventions now have been the subjects of many evidence reviews (e.g., Vernon et al., 1990; Rimer et al., 1994; Fiore et al., 1997; Meissner et al., 1998; Yabroff and Mandelblatt, 1999; Mandelblatt and Yabroff, 1999; Breen et al., 2001; IOM, 2001; Rimer, 2001; Whelan, 2002). Some of the most useful and definitive reviews are those being undertaken by the CDC’s Guide to Community Preventive Services (US Department of Health and Human Services, 2004) While the literature is vast, several conclusions are warranted. 1. Interventions do not have to be complex to be effective. Recent meta-analyses of mammography interventions have shown that relatively simple interventions, such as reminder letters, telephone calls, and counseling, can increase mammography use (Snell and Buck, 1996; Wagner, 1998; Mandelblatt and Yabroff, 1999; Yabroff and Mandelblatt, 1999). If these strategies were systematized, applied, and sustained at the population level, they could have a substantial positive impact on health. 2. Consistent with the multi-level model of disease causality adopted by the IOM (IOM, 2001), the strongest interventions address structural, economic, and geographic barriers to use (access-enhancing interventions), as well as intrapersonal and interpersonal factors (Rimer et al., 1992a; Kiefe et al., 1994; Skaer et al., 1996; Skinner et al., 1999). Access-enhancing interventions may serve as a bridge between health-care settings and the environments in which people reside. In some cases, access enhancement means bringing the service to people (e.g., mobile mammography vans; (Rimer, 1992b; Flynn et al., 1997; Weber and Reilly, 1997; King et al., 1998; Slater et al., 1998; Skinner et al., 1999; Taylor et al., 1999), but it can also include facilitating the use of an intervention by overcoming financial and structural barriers through means such as vouchers and same-day appointments (Fletcher et al., 1993; Kiefe et al., 1994; Stoner et al., 1998; Erwin et al., 1999; Jenkins et al., 1999). Likewise, access enhancements may provide cues to action, for example, a sign-up desk for a mobile van and the subsequent appearance of the van are visible reminders about mammography, build social support and opportunities for modeling, and provide opportunities to get screened, consistent with Social Learning Theory (Baranowski, et al., 1990). 3. In most cases, multiple interventions are needed to achieve changes in cancer prevention and early detection behaviors, especially for vulnerable populations. With a special focus on such populations, Legler et al. (2002) found that use of multiple intervention types was effective, with intervention effects averaging 13.3% overall.

Principles and Applications of Cancer Prevention and Control Interventions

4.

5.

6.

7.

8.

The most effective combination of intervention types appeared to be access-enhancing interventions combined with individualdirected interventions, which had an estimated combined intervention effect of 26.9% (Legler et al., 2002). Interventions that have been grounded in human behavioral science theories have generally shown greater impact. There is some evidence that most cancer control interventions are not guided by theories of health behavior (Meissner et al., 1998; Mandelblatt and Yabroff, 1999; Yabroff and Mandelblatt, 1999; Legler et al., 2002). Interventions should be appropriate to the level of acceptance of the practice or procedure. Different interventions may be needed early in the diffusion curve for a screening test when people may not even be aware of a test like mammography compared with the point at which the test has diffused widely, and the people who have not yet been screened may require multiple interventions, including those that promote access (Legler et al., 2002). Diverse populations do not necessarily require distinct interventions (IOM, 2003). In fact, there is substantial evidence that effective interventions are effective for both mainstream and diverse populations. Improved dissemination of effective interventions is needed. Although more effective interventions are desirable, improved cancer control outcomes are most likely to result from a focus on the dissemination of effective interventions that have already been proven. Tailored interventions may be more effective than generic interventions. Examples include tailored telephone counseling and tailored print communications (Skinner et al., 2002).

CANCER CONTROL ACCOMPLISHMENTS OF THE 20TH CENTURY What has been accomplished by cancer prevention and control interventions? The death rate from cancer is the ultimate outcome measure. Until the very end of the 20th century, mortality rates as well as the absolute number of deaths from cancer had risen inexorably—0.5% per year through 1990 (Weir et al., 2003). However, after stabilizing from 1990–1994, the overall death rate began decreasing at 1.4% per year from 1994–1998 and has since stabilized again (mean annual rate of 202.3/100,00 for 1996–2000) (Weir et al., 2003). Decreases in mortality rates have been noted for cancers of the lung, pancreas, brain, and prostate in men, for breast and cervical cancer in women, and for cancers of the colon/rectum and stomach in both men and women (Wingo et al., 1999; Weir et al., 2003). The absolute numbers of deaths from cancer continue to increase in an aging population (Edwards et al., 2002), but the relentless upward trend in the mortality rate observed over the past century may have been reversed. Evidence directly linking cancer prevention and control interventions to the decline in cancer mortality is not often available, but there is evidence to suggest that improvements in smoking behavior, dietary modification, changes in food preparation and storage, and screening have played a major role in the decline in cancer mortality (Devesa et al., 1989; World Cancer Research Fund, 1997; Wingo et al., 1999). Since the 1964 Surgeon General’s report on the health consequences on smoking, the prevalence of smoking among adults has decreased significantly, from 45% in 1964 to 23% in 2001 (Centers for Disease Control and Prevention, 2003a). Today, there are more former smokers than current smokers in the population. Recognition of tobacco use as a health hazard and subsequent public health antismoking campaigns have resulted in changes in social norms to prevent the initiation of tobacco use, promote cessation of use, and reduce exposure to environmental tobacco smoke (Shopland, 1995; Stockwell et al., 1992). Although currently controversial, a systemic review supported the beneficial effects of fruit and vegetable consumption in protecting against cancers of the colon, mouth, esophagus and stomach (World Cancer Research Fund, 1997). In 1989–1991 the mean intake was about 4.3 servings (Krebs-Smith et al., 1995). Since then, major national efforts to help improve healthy food consumption behavior, such as the 5-A-Day for Better Health Program (Heimendinger et al.,

1285

1996), have been associated with modest increases in daily fruit and vegetable consumption to about 4.7 servings by 1999–2000. However, this improvement is almost all in fruit consumption (Centers for Disease Control, 2004), and optimum levels may be somewhat higher than those now consumed. Recently, the message has changed to emphasize five to nine or more servings per day. Other accomplishments in primary prevention include the epidemiologic delineation of physical activity as a clear protective factor in colorectal cancer and perhaps in breast, prostate, and endometrial cancers as well (Carpenter et al., 1999; IARC, 2002). However, the percentage of adults who report no physical activity in their leisure time is still approximately 38% and falling only slightly (Swan et al., 2003). More people are recognizing the importance of protection from the sun with 60% of the population nationally in 2000 likely to use sunscreen, wear protective clothing, or seek shade (Swan et al., 2003). Chemoprevention research has produced some dramatic benefits including the 49% decrease in invasive breast cancer among high-risk women with tamoxifen (Fisher et al., 1998). Ongoing trials in breast and prostate cancer have the potential for additional cancer prevention benefits. An increased understanding, through behavioral research, of the barriers and facilitators to cancer screening has made it possible to develop effective strategies to promote adherence to both breast and cervical cancer screening (Hiatt, 1997). The use of screening Pap smears for cervical cancer in this country has increased markedly, to the point where approximately 82% of women ages 25 years and older report having been screened within the past 3 years (Swan et al., 2003). This test is a well-established part of routine preventive care, and death from cervical cancer is becoming an increasingly rare event in the United States, decreasing from 7.7 per 100,000 in 1969 to 2.8 per 100,000 in 2000 (SEER, 2003). Screening mammography reduces mortality by around 30% in women over 50 years (Kerlikowske et al., 1995; US Preventive Services Task Force, 1996), and its use has been widely adopted. In 2000 over 70% of women 40 years and over reported having had a mammogram in the past 2 years (Swan et al., 2003). This represents over a twofold increase in self-reported mammography use in every age and race/ethnic group since 1987 (Swan et al., 2003) and even includes women without a usual source of care. This latter finding is likely due in part to the success of the CDCsupported breast and cervical screening program for underserved women (May et al., 1998). As mammography rates have risen dramatically, overall breast cancer mortality has decreased (Wingo et al., 1998). Behavioral research into effective psychosocial and behavioral interventions for cancer survivors has led to practices that contribute to the reduction of cancer pain, enhancements in the quality of life, and, perhaps, to prolonged survival (Spiegel et al., 1989; Fawzy et al., 1993, 1995; Lewis, 1997). All these salutary changes in the nation’s cancer burden can be attributed, at least partially, to the efforts of cancer prevention and control research and its successful application at multiple levels of the health-care system and community (Hiatt and Pasick, 1996; Rimer et al., 2000). The decreases in mortality for lung cancer and cervical cancer, in particular, provide compelling evidence that aggressive and sustained public health action in the control of cancer at the behavioral and societal levels can have dramatic results over the long term. Despite these successes, there remain unmet challenges and some failures of cancer prevention and control interventions. Although tobacco use has been decreasing overall, the rate of decline has slowed in recent years and there are substantial disparities in use between race and ethnic subgroups (Centers for Disease Control and Prevention, 2003a). Tobacco use by teenagers is declining slightly since 1997, but still over 25% of twelfth graders use cigarettes (Centers for Disease Control and Prevention, 2003b). Even though use of mammography has risen dramatically overall, there still remain sizeable underserved segments of the population who are not getting appropriate screening (Swan et al., 2003). Pap smears can prevent deaths from cervical cancer, but we still expect an estimated 3700 deaths in the United States (Jemal et al., 2006). Despite the body of evidence that sedentary lifestyle, weight gain, and obesity increase the risk of

1286

PART V: CANCER PREVENTION AND CONTROL

developing several different cancers, overweight and obesity represent a growing problem in the United States (Calle et al., 2003). From 1971–2000, the prevalence of obesity in the United States increased from 14.5%–30.9% (MMWR, 2004). Mean energy intake in kcals increased as did the mean percentage of kcals from carbohydrates. There is now an urgent need for research to understand the etiologic factors that predispose some people to overweight and obesity and the relationship between overweight/obesity and cancer initiation (IARC, 2002). Applying some of the lessons learned from interventions to change tobacco use in the United States might help to arrest the alarming rise in obesity. About 47% of the US white population are not “very likely” to protect themselves against the harmful effects of sun exposure through appropriate behaviors—for example, using sunscreen, wearing protective clothing, and seeking shade on a sunny day (Hall et al., 1997). Surveillance research continues to reveal troublesome disparities in cancer incidence and mortality by race/ethnicity and socioeconomic status (Singh et al., 2003). Finally, even though organized systems of health-care delivery are becoming the norm in the United States, there is no national population-based system to monitor the quality of care experienced by cancer survivors (National Cancer Policy Board, 2000). And there is no national organized system to remind people when they are due for various screening tests although some health systems now provide such reminders. Thus, although there have been successes in cancer prevention and control and evidence that the burden of cancer may be decreasing in the United States, the goal of making cancer an uncommon and easily treatable disease is far from being met. There is a continuing need to seek ways to extend the reach of proven cancer prevention and control interventions and to develop interventions to achieve a greater population impact. More substantial successes in cancer prevention and control will take time and persistence. Moving forward, what is a logical framework for the application for modern cancer control efforts, based on the principles and evidence that have been developed and used over the past few decades?

FRAMEWORK FOR CANCER PREVENTION AND CONTROL Current principles of cancer prevention and control emphasize that intervention programs must integrate epidemiologic, genetic, behavioral, and cancer surveillance research to accelerate a reduction in the cancer burden. This means describing what is known about the etiology of cancer and the biologic, behavioral, and social determinants of cancer risk, incidence, and mortality; developing effective biomedical and behavioral interventions to reduce cancer incidence, morbidity, and mortality across all populations; improving the quality of life of cancer survivors; and developing innovations in methodology to improve the way we assess cancer prevention and control needs and the impact of interventions. It means that cancer prevention and control must be based on a sound, comprehensive surveillance strategy that includes both cancer registration on geographically welldefined populations as well as applied research based on surveillance data. The framework adopted by the Advisory Committee on Cancer Control of the National Cancer Institute of Canada (National Cancer Institute, 1997) and later adapted by the US NCI has proven highly useful for understanding these structural elements in cancer prevention and control research (Fig. 67–1) (Best et al., 2003). In this framework, cancer prevention and control research is based largely on discoveries in basic or fundamental research that, along with epidemiologic inquiry, answer the question, “What do we know?” and encompasses discovery from many health-related disciplines, including biomedical sciences, psychology, sociology, anthropology, and economics. Intervention research builds on the evidence base from epidemiologic studies and fundamental research and then attempts to discover effective means to apply what has been learned at all levels of the health system and community. It answers the question of “What works?” Interventions may be behavioral, chemopreventive, therapeutic or sociologic, legislative, or policy driven in nature. As several

ility

tab

n ou

Intervention Research What works?

c

Ac

Fundamental Research What do we know?

Eth ics

Knowledge Synthesis and Decision Making What next?

Program Delivery How do we deliver it?

Em po we rm en t

Surveillance and Monitoring Where are we?

y nc icie Eff

Reducing the burden of cancer

Figure 67–1. Framework for cancer prevention and control research.

IOM reports have highlighted, the most effective interventions often are multi-component (IOM, 2001, 2003). When successful, these interventions are promoted at the health system, community, state, and national level as well as in the clinic and at the bedside. The process is iterative; the application of these interventions generates new questions and hypotheses for further fundamental research, including basic biobehavioral research. Also, for applications to reach the appropriate audiences, research on the process of dissemination and diffusion itself, including research on communication and informatics, is an essential part of the cancer control research framework. The role of surveillance of cancer control measures and cancer outcomes is critical to cancer control and serves both to generate hypotheses for cancer control research and to assess the outcomes of cancer control interventions. Surveillance research answers the question “Where are we,” but it also helps to inform the question, “Where do we need to go?” Historically, population-based surveillance has served more of a descriptive and hypothesis-generating function. It now must be enhanced so that it also can clarify the connections between changes in risk factors and early detection behaviors and cancer outcomes, as well as the influences of the quality of health services and clinical treatment on cancer survival, quality of life, and mortality. Surveillance research must not only describe the cancer burden and track changes in cancer rates; it also must explain the reasons for observed disparities and trends in this cancer burden. Synthesis of information and decision making is needed to answer the question “What’s next?” and is rightfully placed at the interface of all other cancer control research activities. Such synthesis processes also help to answer practitioners’ questions, such as “What do I do today?” Much more effort is needed to synthesize what we have learned from fundamental research, including epidemiology, to determine when to move forward and develop a testable intervention. Likewise, a synthesis of intervention research results is critical before we decide whether it is time to move to applied public health and health services applications or whether more basic or other pre-intervention science research is needed. Future cancer control practice must be based on a synthesis of available research findings to optimize the chances of developing successful cancer control strategies. Some new resources can help in this regard. The CDC’s Guide to Community Preventive Services (US Department of Health and Human Services, 2004) conducts evidence reviews on a number of cancer-related topics. The Cancer Control PLANET (Department of Health and Human Services, 2004), developed by NCI, CDC, and ACS, offers users evaluated and downloadable evidence-based interventions. The NCI Cancer Progress Report summarizes trends for the full range of

Principles and Applications of Cancer Prevention and Control Interventions national measures available to track the cancer burden and its determinants (National Cancer Institute, 2004). Finally, this framework explicitly places cancer prevention and control research in the context of the communities and patients who stand to benefit from it. The principles of accountability, empowerment, ethnics, and efficiency should guide this research and the final pathway, and insure that the products of research are delivered and actively disseminated into clinical and public health practice. Building on this framework is an increased appreciation of the importance and integration of the perspectives and languages from a broad range of disciplines and practices (Anderson, 1998), and the introduction of transdisciplinary thinking into cancer control research (Abrams, 1999; Hiatt and Rimer, 1999; Halfon and Hochstein, 2002). ‘Transdisciplinarity’ (Rosenfield, 1992) is a process by which researchers work jointly using a shared conceptual framework that draws together discipline-specific theories, into a new synthesis of concepts, methods, measures, and approaches, to address a common problem. ‘Transdisciplinary’ domains can range across the levels of complexity (cellular and molecular to individual to interpersonal to organizational/institutional, community, and societal), and include the fundamental research disciplines that span biological, psychosocial, and population/public health sciences. Such approaches are especially important in areas such as tobacco use and diet, where multiple domains must be coalesced to accelerate new knowledge. These domains/disciplines include genetics, behavioral research, health communication, bioinformatics, and clinical medicine. Finally, an important and neglected dimension that lends itself to transdisciplinary thinking is time frame—including a lifespan developmental perspective and the perspective that changes may occur in minutes, days, months, years, or even decades, depending on the problem focus (see (Abrams, 1999; Halfon and Hochstein, 2002) for details). Taken together, the three major elements of the transdisciplinary paradigm (cells to society; fundamental to applied; time frame/lifespan development) builds on and consolidates many of the previous generations of models of cancer control science both in the United States and Canada.

APPLICATION OF CANCER PREVENTION AND CONTROL RESEARCH The application of cancer prevention and control research follows the framework just described. The director of a new program in cancer prevention and control research might well ask where to start. Individual practitioners may well focus their work anywhere in the framework with success ultimately measured in how well efforts advance toward measurable reductions in the cancer burden or intermediate markers. From the broader perspective of an academic institution, health department, or health system, however, a thorough assessment of the available surveillance information is critical. Taking a local, regional, national, or international perspective, surveillance data can identify the cancer sites, risk factors, or behaviors responsible for significant proportions of the cancer burden, the sites with the most rapidly rising incidence, and cancers with the largest disparities in either incidence or mortality in subpopulations. Beyond information on persons with cancer, surveillance data can define populations at highest risk because of the prevalence of risk factors like tobacco use, suboptimal dietary practices, overweight and obesity, and high sun exposure. Additional surveillance data can pinpoint lower levels of early detection practices, areas where health care services and practitioners are inadequate, and where the quality of care is low. Starting from such surveillance data problems in cancer control can be defined and appropriate programs initiated. Such programs, for example, may focus on the application of interventions based on available evidence of effectiveness or they may seek to fill gaps where effective interventions do not exist and generate the fundamental discoveries that could lead to new interventions. The remainder of this chapter will focus on areas where the principles of cancer control can be applied covering each node in the framework.

1287

FUNDAMENTAL RESEARCH (EPIDEMIOLOGY, BIOBEHAVIORAL) Our understanding of cancer etiology in human populations constitutes the foundation for cancer control research and must be integrated into the process that leads to the development of successful interventions. Epidemiology is more than just a set of methods for measuring associations of exposure and disease in individuals (Pearce, 1996). Rather, it is an interdisciplinary approach to understanding the causes and distribution of disease in populations. Epidemiology provides a substantial component of the scientific evidence for the development of preventive, including behavioral, interventions by explaining the contributions of biological, psychological, and social factors to cancer etiology. For example, epidemiologic studies of the interaction of genetic susceptibility markers and environmental factors will permit more targeted screening, diagnosis, and treatment. Epidemiologic research into the type, duration, and intensity of physical activity at various stages of life will be needed to develop appropriate interventions to decrease the risk of colorectal and, possibly, breast cancer (Department of Health and Human Services, 1996). Epidemiologic evidence linking social determinants to adverse and disparate cancer outcomes among population subgroups can be used to develop health policies to improve prevention, early detection, and treatment services to those in need (IARC, 1997). In genetic epidemiology, new research infrastructure includes a network of cooperative family registries (Daly et al., 2000), centers for studying individuals at high risk from inherited susceptibility (Anton-Culver et al., 2004), and consortia of large existing cohorts of persons at risk of cancer. These infrastructures are necessary for new epidemiologic methodologic approaches that enable the development of the most effective interventions for genetically susceptible individuals. Epidemiology is, however, fundamentally a population-based science, and it is the eventual applications of new discoveries to populations that will have the greatest impact in cancer control (Pearce, 1996).

INTERVENTION RESEARCH Basic biobehavioral research improves understanding of the mechanisms that govern cancer-related health behaviors and is an important building block for intervention research. Among possible behaviors, recognition of the overwhelming importance of tobacco use in the etiology of major cancers must be central to an effective cancer control strategy. The National Institutes of Health’s (NIH) biobehavioral framework that was adapted to the tobacco control problem illustrates the interrelationship of contributing factors (Tobacco Research Implementation Group, 1998). The national strategy to reduce tobacco use now includes research and training across a spectrum of topics, from basic research on addiction and prevention interventions for youth to drug research on heavy smokers to community-level and state-level media and policy interventions. This includes the formation of transdisciplinary centers for tobacco control research (Turkkan et al., 2000), investments in basic biobehavioral research to inform interventions, and research on societal factors mediated through state-level and community-level policy and media messages. In other areas of primary prevention, synthesis of what has been learned from past research in diet, weight control, and physical activity is needed to determine “What’s next?” The research generated by the 5-A-Day for Better Health program has advanced understanding of broad scale interventions that include the simple “use five fruits and vegetables a day” message, but new research is needed on the determinants of dietary behavior, the interaction of dietary behavior, weight control and physical activity, new methods of dietary assessment for tracking the impact of interventions, and more targeted diet-related interventions (Agency for Healthcare Research and Quality, 2000). Continued efforts are required to reach population subgroups at higher risk, evaluate dietary intake patterns, and develop effective behavioral interventions.

1288

PART V: CANCER PREVENTION AND CONTROL

Interventions to determine how to reduce incidence of and mortality from melanoma are needed to set guidelines for melanoma control and public policy for screening and education for primary prevention (Koh and Geller, 1998). Screening approaches are especially needed for high-risk and previously unscreened populations (Geller, 2002). For cancers for which effective interventions already exist (i.e., cervical, breast, and colorectal), further development is needed to encourage people to maintain regular screening. More attention also is needed to individuals who do not adhere to screening recommendations and for certain subpopulations, like new immigrants and disadvantaged groups. Furthermore, research is needed to aid both individuals and providers in informed decision making for screening and other cancer prevention and control decisions in the face of uncertain evidence of efficacy (Briss et al., 2004). New discoveries of biomarkers and genetic susceptibility will soon make screening tests available that must be evaluated in clinical trials and population-based settings. Concerns about the acceptability of these tests and adherence to them, as well as recruitment and retention to the trials themselves, are increasingly important areas for behavioral cancer control research. Ideally, behavioral research should be launched to understand potential barriers to emerging tests while efficacy studies are underway. Attention to the associations of race and ethnicity with cancerrelated behaviors and in the development and implementation of interventions should be conceptually broadened to include considerations of the influence of socioeconomic class, immigrant status, culture, discrimination, and health-care access on cancer outcomes (Guralnik and Leveille, 1997; National Cancer Policy Board, 2000). These “upstream” influences on health behaviors may have powerful influences on cancer incidence and outcomes at a societal level because of their ubiquitous application to large segments of the population (McKinlay, 1993; Susser and Susser, 1996; Smedley and Syme, 2000; Singer and Ryff, 2001). The new discoveries in molecular biology that have more limited application to high-risk populations may have dramatic impact for individuals and less at a societal level. Effective interventions at the social and cultural levels should be pursued simultaneously with parallel efforts with biomedical innovations. Within the modern structure of cancer prevention and control research, we now have the opportunity to harness the informatics and communication revolutions to improve cancer outcomes. Innovative forms of health communication include, for example, tailored print communications, interactive kiosks, CD-ROMs, and an evergrowing array of new technologies that can provide individualized information to people, depending on such factors as their risk status, other background variables, and information-seeking style. Increasingly, in an era of mass customization (Strecher, 1999), these tailored strategies allow cancer control scientists to individualize and personalize cancer control messages and to offer people choices about the kinds of information they want and need. To date, the evidence of effectiveness in this area is very promising, but many questions remain (Skinner et al., 1999; Strecher, 1999). Evaluations have shown that such interventions can lead not only to improved cancer prevention and early detection behaviors but also to improved decision making as well (McBride et al., 2002; Rimer et al., 2002). Moreover, tailored interventions can be effective even when they address difficult topics, such as genetic susceptibility to lung cancer based on certain genetic polymorphisms (McBride et al., 2002), decision making about testing for BRCA 1/2 mutations and decision making about mammography (Rimer et al., 2002) and colorectal cancer screening (Pignone et al., 2000). Cancer survivorship research includes the identification, prevention, treatment, and care of the broad spectrum of conditions experienced by cancer survivors (Rowland et al., 2001). New research programs are being fostered (Aziz and Rowland, 2003) to recognize the unique needs of individuals who have been treated for cancer and who face many unanswered questions during an increasingly long period of survivorship. Cancer survivorship research builds on the opportunities provided by discoveries in epidemiology, behavioral research, and surveillance while working closely with the cancer survivor and cancer advocacy communities. It seeks to understand risk factors for recur-

rence, enhancements to the quality of life, and the role of the family in survivorship, and the prevention of second malignancies and other sequelae of cancer treatment. Finally, survivorship research includes health services studies of the impact of the cancer diagnosis on the ability of survivors to obtain health insurance and be productive members of the workforce. With a survivor population of 8.5 million and growing, attention to survivors is an urgent need.

SURVEILLANCE RESEARCH The current surveillance research program at the NCI, which includes the Surveillance, Epidemiology, and End-Results (SEER) program and related applied research activities, is the standard for high-quality cancer surveillance (SEER, 2003). This program has developed the statistical and methodologic data resources that enable a more definitive interpretation of cancer patterns and trends. Although cancer surveillance describes the cancer burden in this country admirably, it must be better able to explain the reasons for trends in the burden. As the scope of cancer surveillance is expanded, new methods and tools are being developed (Kim et al., 2000; Department of Health and Human Services, 2004). The two major federally funded cancer surveillance programs, the SEER system and the CDC-supported National Program for Cancer Registries (NPCR), together cover 37 states that produce data meeting quality standards for publication in official US Cancer Statistics (US Cancer Statistics Working Group, 2002) that together with mortality and other cancer statistics will help us understand cancer trends and, ultimately, the causes of cancer in defined populations (Wingo et al., 2003). With the focus on population-based systems in cancer registration from SEER and the NPCR, additional data collection from cancer patients themselves should be supported to advance knowledge about the outcomes of cancer treatments, including the quality of care and quality-of-life assessment. In addition, in these defined populations, innovative ways to connect risk factor and screening data to cancer outcomes should be developed to better understand the link between population level changes in the determinants of cancer and cancer outcomes. Sound methodologic research to develop standardized, reliable, and accurate measures will be central to future progress. Further research in statistical methods and modeling will provide quantitative estimates of the effects of cancer control recommendations, interventions, or population trends. Statistical modeling has particular value in answering questions that cannot be addressed by existing data. Cancer registries can also extend their traditional role in providing information and subjects for epidemiologic studies by the wider application of rapid case ascertainment systems and by collecting biospecimens for genetic epidemiologic studies. Surveillance information is only useful if it is accessible and understandable to a wide spectrum of end users. Several sources now exist (US Cancer Statistics Working Group, 2002; SEER 2003; Weir, 2003) including a new national Cancer Progress Report that describes for a lay audience multiple measures of the cancer burden, including progress toward the National Healthy People goals for 2010 (National Cancer Institute, 2004).

DISSEMINATION RESEARCH Dissemination has been defined as an active process of moving research into practice in contradistinction to “diffusion”, which is more passive and exemplified by the assumption that the simple publication of a successful research result will automatically result in its adoption (Rogers, 1995). The need for effective dissemination of proven interventions has long been recognized but has not been supported adequately as a distinct role of any government agency or professional organization. Too often, cancer control intervention programs are begun without attention to existing evidence on the effectiveness of such interventions, perhaps because it may seem obvious that some activity is better than none and that ideas emanating from communities are better than those imposed upon them. However, it is likely that, given limited resources, communities are

Principles and Applications of Cancer Prevention and Control Interventions most likely to achieve their cancer control objectives when they adopt or adapt effective interventions. Fortunately, a recent review of the NCI’s Spore and Cancer Center Programs (National Cancer Institute, 2003) recommended an enhanced role for the nation’s comprehensive cancer centers in dissemination research and dissemination. Reviews conducted by the CDC’s Guide to Community Preventive Services (US Department of Health and Human Services, 2004) increasingly will be the source of the evidence base for effective interventions. Efforts to develop more systematic approaches to dissemination and support of research on how best to disseminate research findings are now being taken by the NCI, CDC, and the American Cancer Society. For example, the PLANET (Plan, Link, Act, Network with Evidence-based Tools) is a new web-based tool that links practitioners with local or regional surveillance data, local services and health education products, and resources (Department of Health and Human Services, 2004). Modest funding support for the dissemination of interventions of proven efficacy is being offered by the NCI and more attention to the challenge of dissemination is being paid by federal agencies and funding bodies, However, for now, there is no dissemination highway for cancer control interventions and research on dissemination is limited (Agency for Healthcare Research and Quality, 2003). Even when interventions are shown to be effective, they often languish because there is no distribution system. Such a system is needed if we are to accelerate the reduction of cancer incidence and mortality.

COMPREHENSIVE CANCER CONTROL The final step in the cancer control process is to move evidence-based interventions into practice—the application of research. For the past decade, and especially in the past 5 years, the CDC has developed and implemented a planning process for Comprehensive Cancer Control “an integrated and coordinated approach to reducing cancer incidence, morbidity, and mortality through prevention, early detection, treatment, rehabilitation, and palliation” (Abed et al., 2000; Centers for Disease Control and Prevention, 2004). This state-based program seeks to put in place the administrative, health education, clinical services, program development, policy, and other elements necessary to make cancer control most effective. Six critical areas for planning have been identified to enhance infrastructure, mobilize support, use data and research, build partnerships, assess the cancer burden, and conduct evaluation. Starting with several model states the CDC program is being constantly expanded other states, tribes, and territories through collaboration and partnerships with existing cancer control entities and programs (Weir et al., 2003). With continued support this federal program holds the promise of being a highly effective vehicle for putting the principles of cancer control into practice.

FUTURE DIRECTIONS In the end, it is clear that the nation’s goals for cancer control will not be achieved without the collaboration of multiple agencies and organizations, representing the public and private sectors. There is a network of organizations and individuals deeply committed to cancer control in this country, each with their own purposes and objectives. This enterprise has many partners, and we will name only a few. They include federal agencies, such as the NCI, the CDC, the Agency for Healthcare Research and Quality (AHRQ) (Eisenberg, 1998), and the Centers for Medicare and Medicaid (CMS); voluntary organizations, such as the ACS; private entities, such as the Robert Wood Johnson, Komen, and Avon Foundations; and numerous professional associations including the American College of Surgeons (ACOS), the American Society for Preventive Oncology, the American College of Radiology, the American Society for Clinical Oncology, the American Association for Cancer Research, and the Society for Behavioral Medicine. Moreover, there are also thousands of community health centers, group practices, local community agencies, and universities

1289

for whom cancer is a critical challenge. Collaborations between these partners take several forms, including the development of research infrastructure (e.g., the linkage of the SEER and Medicare databases with CMS); programs to disseminate proven interventions (e.g., collaborative efforts by the CDC and the ACS that include the American Stop Smoking Intervention Study (Stillman et al., 2003) and the 5-ADay Program (Heimendinger, 1996)); and planning efforts (e.g., NCI’s coordination with the North American Council of Central Cancer Registries, CDC, ACOS, and ACS to develop a national cancer surveillance program). The ACS and the NCI often complement each other in the dissemination of cancer information to health-care professionals and the public. A vital training and educational function is shared by federal agencies and professional organizations. Despite these partnership efforts, there is less cooperation and collaboration than would be ideal. C-Change (formerly the National Dialogue on Cancer) is a new major undertaking by all of the partners mentioned above and more and shows great promise for strengthening the collaborations needed for successful cancer prevention and control. The basic premise of cancer prevention and control requires the “useful application of results” of cancer research (75th Congress, 1937). If a real impact on the cancer burden across the United States and elsewhere is to be achieved, the results of the cancer prevention and control research must be applied through the activities of multiple partners in programs that have wide-scale acceptance by states, localities, and health-care systems. The interdisciplinary nature of the cancer prevention and control research endeavor is clear, as evidenced by the variety of strategies for interventions outlined above. Scientists from diverse fields of biomedical research must effectively interact with behavioral scientists within the public health model if further advances in reducing the cancer burden are to be made. The challenge before us is to more broadly adopt the principles covered in this chapter for the application of interventions that will sustain a longterm decline in the cancer burden. References 75th Congress (8/5/1937). Senate Bill 2067. Public Law 244. 92nd Congress (12/23/1971). National Cancer Act of 1971. Public Law 92-218. Abed, J, Reilly B, Odell Butler M, Kean T, Wong F, Hohman K. 2000. Developing a framework for comprehensive cancer prevention and control in the United States: An initiative of the Centers for Disease Control and Prevention. J Public Health Manag Pract 6:67–78. Abrams DB. 1999. Nicotine addiction: Paradigms for research in the 21st century. Nicotine Tobacco Research 1:S211–215. Agency for Healthcare Research and Quality. 2000. Agency for Healthcare Research and Quality. Efficacy of Intervention to Modify Dietary Behavior Related to Cancer Risk. Summary, Evidence Report/Technology Assessment: Number 25. AHRQ Publication No. 01-E028. Rockville, MD. Agency for Healthcare Research and Quality. 2003. Diffusion and Dissemination of Evidence-based Cancer Control Interventions. Evidence Report/Technology Assessment: Number 79. 2-6-0004. Americans closer to eating “5-A-Day,” food survey finds. 1997. Cancer Lett 6. Anderson NB. 1998. Levels of analysis in health science: A framework for integrating sociobehavioral and biomedical research. Ann NY Acad Sci 840:563–576. Anton-Culver H, Ziogas A, Finkelstein DM, et al. 2003. The Cancer Genetics Network: Recruitment Results and Pilot Studies. Community Genet 6:171–177. Aziz NM, Rowland JH. 2003. Trends and advances in cancer survivorship research: challenge and opportunity. Semin Radiat Oncol 13:248–266. Baranowski T, Perry CL, Parcel G. 1990. How individuals, environments, and health behavior interact. Social cognitive theory. In: Glanz K, MarcusLewis F, Rimer BK, eds. Health Behavior and Health Education. San Francisco: Jossey-Bass, pp. 153–178. Best A, Hiatt RA, Cameron R, Rimer BK, Abrams DB. 2003. The evolution of cancer control research: An international perspective from Canada and the United States. Cancer Epidemiol Biomarkers Prev 8:705–712. Breen N, Wagener DK, Brown ML, Davis WW, Ballard-Barbash R. 2001. Progress in cancer screening over a decade: Results of cancer screening from the 1987, 1992, and 1998 National Health Interview Surveys. J Natl Cancer Inst 93:1704–1713.

1290

PART V: CANCER PREVENTION AND CONTROL

Breslow L, Agran L, Breslow DM, Morganstern M, Ellwein L. 1977. Cancer control: Implications from its history. J Natl Cancer Inst 59:671–686. Briss P, Rimer B, Reilly B, et al. 2004. Promoting informed decisions about cancer screening in communities and healthcare systems. Am J Prev Med 26:67–80. Calle EE, Rodriguez, Walker-Thurmond K, Thun MJ. 2003. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. N Engl J Med 348:1625–1638. Carpenter C, Ross R, Paganini-Hill A, Bernstein L. 1999. Lifetime exercise activity and breast cancer risk among post-menopausal women. Br J Cancer 80:1852–1858. Centers for Disease Control. 2004. www.cdc.gov/nchs/about/major/nhanes/ NHANES9900.htm. Centers for Disease Control and Prevention. 2003a. Cigarette smoking among adults—United States, 2001. Morb Mortal Wkly Rep 52:953–966. Centers for Disease Control and Prevention. 2003b. http://www.cdc.gov/ nccdphp/dash/yrbs/index.htm. 2-6-0004b. Centers for Disease Control and Prevention. 2004. http://www.cdc.gov/cancer/ ncccp/toolkit/index.htm. 2-6-0004. Daly MB, Offit K, Li F, et al. 2000. Participation in the cooperative family registry for breast cancer studies: Issues of informed consent. J Natl Cancer Inst 92:452–456. Department of Health and Human Services. 1996. Physical Activity and Health: A Report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Center for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion. Department of Health and Human Services. 2004. http://cancercontrolplanet.cancer.gov/. Devesa SS, Young JL, Jr., Brinton LA, Fraumeni JF, Jr. 1989. Recent trends in cervix uteri cancer. Cancer 64:2184–2190. Edwards BK, Howe HL, Ries LA, et al. 2002. Annual report to the nation on the status of cancer, 1973–1999, featuring implications of age and aging on U.S. cancer burden. Cancer 94:2766–2792. Eisenberg JM. 1998. Health services research in a market-oriented health care system. Health Aff (Millwood) 17:98–108. Erwin DO, Spatz TS, Stotts RC, Hollenberg JA. 1999. Increasing mammography practice by African American women. Cancer Pract 7:78–85. Fawzy FI, Fawzy NW, Arndt LA, Pasnau RO. 1995. Critical review of psychosocial interventions in cancer care. Arch Gen Psychiatry 52: 100–113. Fawzy FI, Fawzy NW, Hyun CS, et al. 1993. Effects of an early structured psychiatric intervention, coping, and affective state on recurrence and survival 6 years later. Arch Gen Psychiatry 50:681–689. Fiore MC, Jorenby DE, Baker TB. 1997. Smoking cessation: Principles and practice based upon the AHCPR guideline, 1996. Ann Behav Med 19:213–219. Fisher B, Costantino JP, Wickerham DL, et al. 1998. Tamoxifen for prevention of breast cancer: Report of the National Surgical Adjuvant Breast and Bowel Project P-1 Study, J Natl Cancer Inst 90:1371–1388. Fletcher SW, Harris RP, Gonzalez JJ, et al. 1993. Increasing Mammography Utilization: A Controlled Study. J Nat Cancer Inst 85:112–120. Flynn BS, Gavin P, Worden JK, Ashikaga T, Guatam S, Carpenter J. 1997. Community Education Programs to Promote Mammography Participation in Rural New York State. Prev Med 26:102–108. Geller AC. 2002. Screening for melanoma. Dermatol Clin 20:629–640. Glanz K. 1997. Behavioral research contributions and needs in cancer prevention and control: Dietary change. Prev Med 26:S43–S55. Greenwald P, Cullen JW. 1985. The new emphasis in cancer control. J Natl Cancer Inst 74:543–551. Guralnik JM, Leveille SG. 1997. Annotation: Race, ethnicity, and health outcomes—unraveling the mediating role of socioeconomic status. Am J Public Health 87:728–730. Halfon N, Hochstein M. 2002. Life course health development: An integrated framework for developing health, policy, and research. Milbank Q 80:433–479. Hall HI, May DS, Lew RA, Koh HK, Nadel M. 1997. Sun protection behaviors of the U.S. white population. Prev Med 26:401–407. Hanahan D, Weinberg RA. 2000. The hallmarks of cancer. Cell 100:57–70. Harvard report on cancer prevention. 1996. Volume 1: Causes of human cancer. Cancer Causes Control 7:S59. Heimendinger J, Van Duyn MA, Chapelsky D, Foerster S, Stables G. 1996. The national 5 A Day for Better Health Program: A large-scale nutrition intervention. J Public Health Manag Pract 2:27–35. Hiatt RA. 1997. Behavioral research contributions and needs in cancer prevention and control: Adherence to cancer screening advice. Prev Med 26:S11–S18. Hiatt RA, Pasick RJ. 1996. Unsolved problems in early breast cancer detection: Focus on the underserved. Breast Cancer Res Treat 40:37–51.

Hiatt RA, Rimer BK. 1999. A new strategy for cancer control research. Cancer Epidemiol Biomarkers Prev 8:957–964. IARC. 1997. Social Inequalities and Cancer. IARC: Lyon. IARC. 2002. Weight control and physical activity. IARC: Lyon. IOM. 2001. Smedley, Brian D. and Syme, S. Leonard. Promoting Health: Intervention Strategies from Social and Behavioral Research. Washington, D.C.: National Research Council, Institute of Medicine. IOM. 2003. Speaking of Health: Assessing Health Communication Strategies for Diverse Populations. Washington, D.C.: National Research Council, Institute of Medicine. Jemal A, Siegal R, Ward E, et al. 2006. Cancer Statistics 2006. CA Cancer J Clin 56:106–130. Jenkins CN, McPhee SJ, Bird JA, et al. 1999. Effect of a media-led education campaign on breast and cervical cancer screening among VietnameseAmerican women. Prev Med 28:395–406. Kaluzny AD. 1997. Prevention and control research within a changing health care system. Prev Med 26:S31–S35. Kerlikowske K, Grady D, Rubin SM, Sandrock C, Ernster VL. 1995. Efficacy of screening mammography. A meta-analysis. JAMA 273:149–154. Kiefe CI, McKay SV, Halevy A, Brody BA. 1994. Is cost a barrier to screening mammography for low-income women receiving medicare benefits? Arch Intern Med 154:1217–1224. Kim HJ, Fay MP, Feuer EJ, Midthune DN. 2000. Permutation tests for joinpoint regression with applications to cancer rates. Stat Med 19:335–351. King E, Rimer BK, Benincasa T, et al. 1998. Strategies to encourage mammography use among women in senior citizens’ housing facilities. J Cancer Educ 13:108–115. Koh HK, Geller AC. 1998. Public health interventions for melanoma. Prevention, early detection, and education. Hematol Oncol Clin North Am 12:903–928. Krebs-Smith SM, Cook A, Subar AF, Cleveland L, Friday J. 1995. U.S. adults’ fruit and vegetable intakes, 1989 to 1991: A revised baseline for the Healthy People 2000 objective. Am J Public Health 85:1623–1629. Legler J, Meissner HI, Coyne C, Breen N, Chollette V, Rimer BK. 2002. The effectiveness of interventions to promote mammography among women with historically lower rates of screening. Cancer Epidemiol Biomarkers Prev 11:59–71. Lerman C. 1997. Translational behavioral research in cancer genetics. Prev Med 26:S65–S69. Lewis FM. 1997. Behavioral research to enhance adjustment and quality of life among adults with cancer. Prev Med 26:S19–S29. Lichtenstein E. 1997. Behavioral research contributions and needs in cancer prevention and control: Tobacco use prevention and cessation. Prev Med 26:S57–S63. Mandelblatt JS, Yabroff KR. 1999. Effectiveness of interventions designed to increase mammography use: A meta-analysis of provider-targeted strategies. Cancer Epidemiol Biomarkers Prev 8:759–767. May DS, Lee NC, Nadel M, Henson RM, Miller DS. 1998. The National Breast and Cervical Cancer Early Detection Program: Report on the first 4 years of mammography provided to medically underserved women. Am J Roentgenol 170:97–104. McBride CM, Bastian LA, Halabi S, et al. 2002. A tailored intervention to aid decision-making about hormone replacement therapy. Am J Public Health 92:1112–1114. McGinnis JM, Foege WH. 1993. Actual causes of death in the United States. JAMA 270:2207–2212. McKinlay JB. 1993. The promotion of health through planned sociopolitical change: Challenges for research and policy. Soc Sci Med 36:109–117. Meissner HI, Breen N, Coyne C, Legler JM, Green DT, Edwards BK. 1998. Breast and cervical cancer screening interventions: An assessment of the literature. Cancer Epidemiol Biomarkers Prev 7:951–961. MMWR. 2004. Trends in intake of energy and macronutrients—United States, 1971–2000. Mor Mortal Wkly Rep 53:80–82. National Cancer Institute. 1997. A new agenda for cancer control research: Report of the cancer control program review group. National Cancer Institute Web site. National Cancer Institute. 2003. P30/P50 Ad Hoc Working Group. Advancing Translational Cancer Research: A Vision of the Cancer Center and SPORE Programs of the Future. Bethesda, MD: National Cancer Institute. National Cancer Institute. 2004. Cancer Progress Report—2003 Update. Bethesda, MD: NIH, DHHS. National Cancer Institute and National Institutes of Health. 1997. Report of the National Cancer Institute: Cancer prevention program review group. National Cancer Institute of Canada. 1997. Advisory Committee on Cancer Control. National Cancer Policy Board IoM. 2000. Enhancing Data Systems to Improve the Quality of Cancer Care. National Academy Press: Washington, DC.

Principles and Applications of Cancer Prevention and Control Interventions New York City Cancer Committee. 1994. New York City Cancer Committee: History of the American society for the control of cancer, 1913–1943. New York City Cancer Committee: New York. Pearce N. 1996. Traditional epidemiology, modern epidemiology, and public health. Am J Public Health 86:678–683. Pignone M, Harris R, Kinsinger L. 2000. Videotape-based decision aid for colon cancer screening. A randomized, controlled trial. Ann Intern Med 133:761–769. Rimer BK. 1994. Interventions to increase breast screening. Lifespan and ethnicity issues. Cancer 74:323–328. Rimer BK. 1999. Cancer control research 2001. Cancer Epidemiol Biomarkers Prev 8:957–964. Rimer BK, Halabi S, Sugg SC, et al. 2002. Effects of a mammography decision-making intervention at 12 and 24 months. Am J Prev Med 22:247–257. Rimer BK, Meissner HI, Breen N, Legler J, Coyne C. 2000. Social and behavioral interventions to increase breast cancer screening. In: Integrating Behavioral and Social Sciences with Public Health. Schneiderman N, Speers M, Silva J, Tomes H, Gentry J, eds. pp. 177–201. Washington, D.C.: American Psychological Association. Rimer BK, Meissner HI, Breen N, Legler J, Coyne C. 2001. Social and behavioral intervention in cancer control. In: Integrating Behavioral and Social Sciences with Public Health. Schneiderman MA, Speers MA, Silva DM, Tomes H, Gentry JH, eds. pp. 177–201. American Psychological Association: Washington, D.C. Rimer BK, Resch N, King E, et al. 1992a. Multistrategy health education program to increase mammography use among women ages 65 and older. Public Health Rep 107:369–380. Rimer BK. 1992b. Understanding the acceptance of mammography by women. Ann Behav Med 14:197–203. Rogers EM. 1995. Diffusion of innovations. Free Press: New York. Rosenfield PL. 1992. The potential of transdisciplinary research for sustaining and extending linkages between the health and social sciences. Soc Sci Med 35:1343–1357. Rowland JH, Aziz N, Tesauro G, Feuer EJ. 2001. The changing face of cancer survivorship. Semin Oncol Nurs 17:236–240. Shopland DR. 1995. Tobacco use and its contribution to early cancer mortality with a special emphasis on cigarette smoking. Environ Health Perspect 103 Suppl 8:131–142. Singer BH, Ryff CD. 2001. New horizons in health: An integrative approach. National Academy Press: Washington, D.C. Singh GK, Miller BA, Hankey BF, Edwards BK. 2003. Bethesda, MD, National Cancer Institute. NCI Cancer Surveillance Monograph Series, Number 4. NIH Publication No. 03-5417. Skaer TLB, Robison LM, Sclar DAB, Harding GH. 1996. Financial Incentive and the Use of Mammography Among Hispanic Migrants to the United States. In: Health Care for Women International. Taylor & Francis pp. 281–291. Skinner CS, Campbell MK, Rimer BK, Curry S, Prochaska JO. 1999. How effective is tailored print communication? Ann Behav Med 21:290–298. Skinner CS, Schildkraut JM, Berry D, et al. 2002. Pre-counseling education materials for BRCA testing: Does tailoring make a difference? Genet Test 6:93–105. Slater JS, Ha CN, Malone ME, et al. 1998. A randomized community trial to increase mammography utilization among low-income women living in public housing. Prev Med 27:862–870. Smedley BD, Syme SL. 2000. Promoting health: Intervention strategies from social and behavioral research. National Academy Press: Washington, D.C. Snell JLBS, Buck EL. 1996. Increasing cancer screening: A meta-analysis. Prev Med 25:702–707. Spiegel D, Bloom JR, Kraemer HC, Gottheil E. 1989. Effect of psychosocial treatment on survival of patients with metastatic breast cancer. Lancet ii:888–891. Stillman F, Hartman A, Graubard B, Gilpin E, Murray D, Gibson J. 2003. Evaluation of the American Stop Smoking Intervention Study (ASSIST): A report of outcomes. J Natl Cancer Inst 95:1681–1691. Stockwell HG, Goldman AL, Lyman GH, et al. 1992. Environmental tobacco smoke and lung cancer risk in nonsmoking women. J Natl Cancer Inst 84:1417–1422.

1291

Stoner TJ, Dowd B, Carr WP, Maldonado G, Church TR, Mandel J. 1998. Women’s Health—Do Vouchers Improve Breast Cancer Screening Rates? Results from a Randomized Trial. Health Serv Res 33:11–28. Strecher VJ. 1999. Computer-tailored smoking cessation materials: A review and discussion. Patient Educ Couns 36:107–117. Surveillance, Epidemiology and End Results SEER Program www.seer.cancer.gov. 2003. SEER*Stat Database: Mortality—All COD, Public-Use With State, Total U.S. (1969–2000). Underlying mortality data provided by NCHS (www.cdc.gov/nchs). national Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch. Susser M, Susser E. 1996. Choosing a future for epidemiology: II. From black box to Chinese boxes and eco-epidemiology. Am J Public Health 86:674–677. Swan J, Breen N, Coates RJ, Rimer BK, Lee NC. 2003. Progress in cancer screening practices in the United States: Results from the 2000 National Health Interview Survey. Cancer 97:1528–1540. Taylor V, Thompson B, Lessler D, et al. 1999. A clinic-based mammography intervention targeting inner-city women. J Gen Intern Med 14:104– 111. Tobacco Research Implementation Group. 1998. Tobacco Research Implementation Plan: Priorities for tobacco research beyond the year 2000. National Cancer Institute, National Institutes of Health. Turkkan JS, Kaufman NJ, Rimer BK. 2000. Transdisciplinary tobacco use research centers: A model collaboration between public and private sectors. Nicotine Tob Res 2:9–13. US Cancer Statistics Working Group. 2002. United States Cancer Statistics: 1999 Incidence. Atlanta, GA, Department of Health and Human Services, Centers for Disease Control and Prevention, National Cancer Institute. US Preventive ServicesTask Force. 1996. Guide to clinical preventive services. International Medical Publishing: Alexandria, VA. US Department of Health and Human services, Task Force on community Preventive Services. 2004. Available at: www.thecommunityguide.org. Accessed February 6, 2004. Vernon SW, Laville EA, Jackson GL. 1990. Participation in Breast Screening Programs: A Review. Soc Sci Med 30:1107–1118. Wagner TH. 1998. The effectiveness of mailed patient reminders on mammography screening: A meta-analysis. Am J Prev Med 14:64–70. Weber BE, Reilly BM. 1997. Enhancing mammography use in the inner city— A randomized trial of intensive case management. Arch intern Med 157:2345–2349. Weir HK, Thun MJ, Hankey BF, et al. 2003. Annual Report to the Nation on the status of cancer, 1975–2000, featuring the uses of surveillance data for cancer prevention and control. J Natl Cancer Inst 95:1276– 1299. Whelan TM, O’Brien MA, Villasis-Keever M, et al. 2002. Agency for Healthcare Research and Qualty. Impact of cancer-related decision aids. Evidence Report/Technology Assessment Number 46. (Prepared by McMaster University under Contract No. 290-97-0017.) AHRQ Publication No. 02-E004. Rockville, MD, AHRQ. Wingo PA, Jamison PM, Hiatt RA, et al. 2003. Building the infrastructure for nationwide cancer surveillance and control—a comparison between The National Program of Cancer Registries (NPCR) and The Surveillance, Epidemiology, and End Results (SEER) Program (United States). Cancer Causes Control 14:175–193. Wingo PA, Ries LAG, Giovino GA, et al. 1999. Annual report to the nation on the status of cancer, 1973–1996, with a special section on lung cancer and tobacco smoking. J Natl Cancer Inst 91:675–690. Wingo PA, Ries LAG, Rosenberg HM, Miller DS, Edwards BK. 1998. Cancer incidence and mortality, 1973–1995: A report card for the U.S. Cancer 82:1197–1207. World Cancer Research Fund. 1997. Food, nutrition, and the prevention of cancer: A global perspective. Washington, D.C.: American Institute for Cancer Research. Yabroff KR, Mandelblatt JS. 1999. Interventions Targeted toward Patients to Increase Mammography Use. Cancer Epidemiology Biomark Prev 8: 749–757.

68

Increasing the Contribution of Epidemiology to the Primary Prevention of Cancer BEVERLY ROCKHILL AND DOUGLAS WEED

O

ur discipline of epidemiology has a long history with cancer prevention. In the past century alone, from the early yet sophisticated studies of breast cancer by Lane-Claypon (1926) in the 1920s through the mid-century discovery of smoking as a cause of many cancers, to the ever-expanding corpus of knowledge that is regularly recorded in our journals and textbooks, epidemiology has distinguished itself among scientific and professional disciplines as a key player in the search for the causes of cancer and for beneficent preventive interventions. Epidemiology is not alone in this quest. Many other disciplines, biological, clinical, and social, are also important; preventing cancer is a multi-disciplinary activity. Nevertheless, there is something special about epidemiology and its relationship to cancer prevention. No other discipline defines its practitioners as those who study the (causal) determinants and distributions of cancer and then apply that knowledge to improve the public’s health through prevention and control. Ideally, this definition should provide epidemiology with firm conceptual, scientific, and ethical foundations so that it can claim its rightful place at the forefront of progress in cancer prevention. Yet our discipline is deeply afflicted by a lack of consensus about its societal role and therefore its social responsibilities; there are influential practitioners who still insist that epidemiologists should insulate themselves from preventive applications. We believe that this disciplinary division, as well as concerns about nagging uncertainties that plague all scientific inquiries more observational than experimental, put at risk epidemiology’s full contribution to progress in cancer prevention. With this brief history in mind, our purpose in writing this chapter is to stimulate critical and creative thought about the role that epidemiology can, and should, play in preventing cancer. At the heart of this chapter are four suggestions for increasing the contribution that epidemiologists can make to the primary prevention of cancer. We believe (and will argue) that epidemiologists should commit to prevention, and preventive intervention, as central professional goals; be willing to develop our understanding of cancer causation and prevention from a population perspective as well as from an individual perspective; accept uncertainty as the “normal” state and be vigilant about the risk of paralysis that uncertainty brings to action; and broaden our concept of causal inference so that the process is useful for primary prevention. Before we lay out the arguments in support of this plan and the means to achieve it, we need answers to two critical questions. First, what part of the cancer burden in the United States is preventable? Second, for which cancers has there been the most progress in primary prevention?

WHAT PART OF THE CANCER BURDEN IS PREVENTABLE? Arguments for Avoidability The chief arguments for the avoidability or preventability of the cancer burden have come from estimates of the proportion of cancer due to extrinsic (non-genetic) causes. In 1964, for instance, a World Health Organization committee appointed to consider how existing knowledge could be applied to prevent cancer began its report by stating that:

1292

The potential scope of cancer prevention is limited by the proportion of human cancers in which extrinsic factors are responsible. These [factors] include all environmental carcinogens (whether identified or not) as well as “modifying factors” that favor neoplasia of apparently intrinsic origin (e.g., hormonal imbalances, dietary deficiencies and metabolic defects). The categories of cancer that are thus influenced, directly or indirectly, by extrinsic factors include many tumours of the skin and mouth, the respiratory, gastrointestinal and urinary tracts, hormone dependent organs (such as the breast, thyroid and uterus), haematopoietic and lymphopoietic systems, which, collectively, account for more than three-quarters of human cancers. It would seem, therefore, that the majority of human cancer is potentially preventable. (WHO, 1964) In the ensuing years, many epidemiologists have not only agreed with the WHO’s conclusion in principle, but have also provided quantitative estimates of the proportion of cancer attributable to nongenetic factors. For example, in 1969, Higginson concluded that 90% of all human cancer incidence could be attributed to environmental (i.e., non-genetic) influences, and was theoretically preventable. Twentyseven years later, Doll (1996) suggested that age-specific incidence rates for all cancers combined in high-risk countries like the United States and the United Kingdom could be reduced by 80%–90%. Willett et al. (1996) were somewhat more modest in their conclusion that cancer mortality rates could be reduced by up to 60% in the United States. Some of the variability in these estimates can be explained by subtle differences in methodologic approach, although all relied upon the assumption that low-risk countries or populations were, in effect, attainable targets for high-risk countries. In an oft-cited work devoted to quantifying the avoidability of cancer, Doll and Peto (1981), who also estimated that approximately 90% of all cancers were due to extrinsic factors and thus potentially avoidable, noted that such estimates could be based upon four types of evidence: differences in the incidence of cancer among different settled communities, differences between migrants from a community and those who remain behind, variations with time in the incidence of cancer within particular communities, and the actual identification of many specific causes or preventive factors. (Doll and Peto, p. 1198) While noting that differences across time and populations in cancer incidence were notoriously difficult to estimate reliably, given variation in sensitivity and specificity of case finding and definition, they emphasized that the chief point remained: epidemiologic data consistently suggested that, for any given cancer, high age-specific incidence rates were not inevitable. Comparisons of cancer surveillance data from different populations have proved highly valuable in epidemiology and public health, often suggesting important etiologic hypotheses that have been tested in subsequent studies, some of which have ultimately resulted in successful preventive interventions. These same comparisons, however, also reflect important social, political, and economic forces, some of which influence the more “proximal” cancer risk factors in the populations involved. Put another way, incidence rates may not differ between societies solely because of differences in important “lifestyle” risk factors like smoking or other risk factors. It follows that lower

Increasing the Contribution of Epidemiology to the Primary Prevention of Cancer incidence rates of cancer are not necessarily synonymous with higher levels of public health, social well-being, or even longevity. A good example can be found in the predominantly pre-industrial and agrarian societies with low incidence rates of colorectal cancer. While consumption of refined foods and animal flesh tends to be much lower in such societies compared with societies like the United States, and while sedentariness is also much rarer in such societies, levels of childhood malnutrition and childhood and adult infection tend to be much higher. As a second example, consider the low incidence rates of breast cancer that existed in many Asian countries until the end of the 20th century. Based on knowledge of breast cancer etiology, such low incidence rates are expected in societies in which the length of time between menarche and childbearing is short, and women spend a relatively large proportion of their reproductive lives pregnant and lactating, compared with women in the modern United States. A question that arises logically from the above discussion is this: are societies with very low rates of cancer realistic points of comparison for higher-risk societies like the United States, when it comes to thinking practically about preventability? Basing the belief that high incidence rates are not inevitable on evidence showing that not all societies or population subgroups have equally high rates makes sense from an arithmetic standpoint, but the leap in reasoning about consequent preventability may be too swift. Clues about etiology that are gleaned from such comparisons may not translate readily, nor ethically, into prevention strategies, even if identified risk factors are nongenetic or otherwise theoretically modifiable. We will return to this point in a later section. We wish to state early in this chapter, however, that although we write as cancer epidemiologists with strong commitments to primary prevention, cancer prevention is only one means to the end of improving public health; it is not the end in itself.

Major Risk Factors A common list of risk factors appears in many epidemiologic reports on the avoidability of cancer. Without exception, tobacco use is the primary factor cited. Alcohol consumption and other dietary factors usually appear next on the lists (which are often ranked in terms of attributable fraction for all cancers combined), though there has been widespread acknowledgment of the great uncertainty surrounding the determination of specific dietary constituents, besides alcohol, that are either carcinogenic or anti-carcinogenic. If the risk factor of overnutrition, or obesity, is considered under the broad category of diet, it becomes the single dietary factor that shows up most consistently as associated with increased risks of different cancers. Physical activity is closely related to obesity, and the risk factor of sedentariness is mentioned by many epidemiologists when discussing the potential avoidability of cancer. Other prominent factors cited by epidemiologists in discussions of modifiable risk factors and avoidability of cancer include reproductive behaviors/sex hormones, sexual practices and other routes of exposure to viral infections, exposure to ultraviolet light and ionizing radiation, and use of screening practices that serve to detect and remove precancerous lesions (e.g., endoscopy for colon cancer, the Pap smear for cervical cancer). We recognize the importance of screening when it comes to several important cancers, but limit our attention in this chapter to the discovery and modification of preventable risk factors (causes).

FOR WHICH CANCERS HAS THERE BEEN THE MOST PROGRESS IN PRIMARY PREVENTION? Assessing Contribution of Epidemiologic Knowledge Before addressing the issue of progress in prevention for specific cancers, we believe it is useful to distinguish between declines in cancer incidence rates due to factors or forces largely outside the realm of application of epidemiologic knowledge, and those declines due to the direct application of epidemiologic knowledge to primary prevention. Consider, for example, the history of stomach cancer in the United States. In 1930, stomach cancer was the leading cause of cancer death (Nomura, 1996). Since then, incidence and mortality rates have

1293

declined dramatically; stomach cancer mortality is now only one-fifth of what it was in 1930 (Ries et al., 2003). The reasons for this decline are likely related to technological changes in food preservation (refrigeration) and availability (interstate transportation networks made possible the availability of fresh fruits and vegetables to a bigger proportion of the population for more months during the year), as well as to socioeconomic changes in living conditions leading to a decline in prevalence of Helicobacter pylori infection (Parsonnet et al., 1991; Marshall, 2002). Thus, and most importantly, although there has been progress in stomach cancer prevention from a public health standpoint, the progress did not arise primarily out of the application of epidemiologic research findings. Little was known about the etiology and epidemiology of stomach cancer when the sharp decline in mortality began in this country. In the discussion of trends in cancer mortality and incidence below, we focus on changes in rates, providing a foundation upon which we can discuss how epidemiologists can increase their contribution to cancer prevention. In discussing these trends we limit our focus to the United States and provide only a general summary, given the wide availability of references that present detailed data on such trends.

Both Sobering and Encouraging Trends In a series of articles devoted to assessing the progress against cancer in the United States, Bailar (Bailar, 1979; Bailar and Smith, 1986; 1987; Bailar and Gornik, 1997) have consistently painted a sobering view of the progress made in this country since the “war” on cancer began in 1971. Bailar and Gornik (1997) noted that US cancer mortality rates (age adjusted to 1970) increased by an estimated 0.3% annually from 1975–1993, compared with an average increase of 0.1% per year from 1950 through 1975, despite the enlarged scope of cancer research that was occurring in this country during the latter time period. The increase in cancer mortality over this time period appeared to reflect a true increase in risk, rather than being the result of lessened competing risk from the leading cause of mortality, coronary heart disease (Rothenberg, 1994). Data on cancer mortality trends in the 1980s and 1990s have been more encouraging, however. In a 1998 report generated by a multiagency collaborative team (Wingo et al., 1998), the first sustained decline in the US cancer death rate was noted. Subsequent reports have updated information on this trend. Between 1992 and 1998, mortality from all cancers combined declined in each of the four groups of white men, black men, white women, and black women, although the magnitude of the decline varied across these groups (Howe et al., 2001). According to this summary report published in 2001 (Howe et al., 2001), cancers of the breast, prostate, lung, and colon/rectum accounted for 55.9% of all 1998 SEER cases and 52.7% of all 1998 cancer deaths in the United States. Changes in the rates for these sites thus exert a strong influence on overall cancer trends. In 1998, breast cancer accounted for nearly 8% of all cancer deaths, and just over 16% of all cancer diagnoses. From 1973 through 1998, breast cancer incidence rates increased sharply, although the increases were limited to early-stage and in-situ disease. Since 1989, however, the long-steady mortality rates from breast cancer have been declining, and this decline strengthened in 1995. This decline has mostly been limited to white women, although younger black women have also experienced a decline in mortality. These data suggest that increased uptake of screening mammography may have played a role in the sharp increase in the 1980s and 1990s in incidence of in-situ and early-stage breast cancer, and in the concomitant decline in mortality. Improved therapy likely also has played a role in the declining mortality. It is important to note, however, that the long-term rise in breast cancer incidence that has occurred in this country since record-keeping began (in Connecticut in the 1930s) can likely be traced to major shifts in population distributions of key reproductive risk factors (Krieger, 1990; Hoover, 1996) as can the rises in incidence that have been observed since 1970 in traditionally low-incidence countries such as Japan, Singapore, China, and some countries of central and South America (Yu et al., 1991; Tominaga et al., 1994; Hoover, 1996; Seow et al., 1996; Nagata et al., 1997).

1294

PART V: CANCER PREVENTION AND CONTROL

Prostate cancer, which in 1998 accounted for approximately 6% of cancer deaths and 15% of cancer cases, increased steeply in incidence between 1973 and 1992, with an especially sharp rate of increase associated with the introduction of prostate-specific antigen (PSA) testing in 1988. Since 1992, incidence rates have declined in both white and black men. Death rates from prostate cancer have also declined since the mid 1990s in both white and black men. The impact of PSA screening on mortality is not yet known, as other population-wide changes that affect this disease have occurred concurrently, such as revised recommendations to treat early-stage disease more aggressively. Lung cancer, which in 1998 accounted for a large 28.5% of cancer deaths and just over 13% of diagnosed cancer cases, decreased in incidence between 1992 and 1998, because of a decline in incidence among men and a leveling-off of rates among women. Lung cancer mortality began to decrease in 1990 for men, but it has continued to increase, at least through 1998, in women. Lung cancer incidence and mortality are perhaps the most sensitive indicators of trends in cigarette smoking uptake 3–5 decades prior. (A large proportion of the sustained rise in total cancer mortality between 1975 and the early 1990s was due to increases in lung cancer mortality (Bailar and Gornik, 1997).) Examination of long-term lung cancer trends illustrate that women have lagged behind men in incidence and mortality, because the peak in women’s tobacco use came after that in men. Colorectal cancer accounted for 10.5% of cancer deaths and 11.6% of cases in 1998. Incidence rates increased between 1973 and 1985, but have subsequently declined. For the period 1992–1998, incidence rates declined for white men, and remained constant for black men, black women, and white women. Mortality rates have been declining at least since 1992 in white men, white women, and black men, whereas mortality rates in black women have been stable. It is likely that increased use of effective early detection methods, such as fecal occult blood testing (FOBT) (Cramer, 1974), sigmoidoscopy, colonoscopy, and barium enema have contributed to the decline in mortality, through leading to detection of precancerous lesions and early stages of disease. It is also likely that improved chemotherapeutic treatment has contributed to decreased mortality rates. Although not mentioned in the 2001 report on cancer trends, cervical cancer is an important cancer to mention in a chapter devoted to primary prevention. Cervical cancer is the second most common cancer among women throughout the world, and it ranks first in many less-industrialized countries. In the United States and other western countries, there has been a marked decline in cervical cancer incidence and mortality since the 1950s. While some of this reduction has undoubtedly been due to early detection with cervical cytological screening and corresponding treatment of intraepithelial preinvasive lesions (Cramer, 1974), the initiation of the declines in many of these countries preceded the introduction of screening. The declines certainly pre-dated the identification of human papilloma virus as a potentially necessary cause of cervical cancer. There was probably a causal link between increasing affluence and improvements in the standards of living (improvements in nutrition and genital hygiene, declines in early and frequent childbearing, increasing use of barrier contraceptives, and a decline in the general prevalence of sexually transmitted diseases) and the decline in the population burden of cervical cancer, although, as mentioned above, screening has been an important preventive intervention. In addition to the long-term rise in breast cancer incidence and the long-term and continuing rise in lung cancer incidence and mortality among women, an increase in either incidence or mortality has been observed since 1975 for 10 other cancer sites (Howe et al., 2001), including non-Hodgkin lymphoma, esophageal cancer, melanoma, and cancers of the liver and intrahepatic bile duct. While incidence and mortality rates of melanoma have been rising in this country, they have been falling in Australia; we will return to a discussion of this divergence in trends in a later section.

A Qualified Statement about Progress Considering the available surveillance data on cancer incidence and mortality, it seems clear that the question “Has there been progress in

applying epidemiologic knowledge to cancer prevention?” must be answered with a qualified “yes.” Many epidemiologists would probably agree that much is known about lung cancer prevention, and that epidemiologic research has been applied by public health professionals, policy makers, legislators and litigators, and grassroots community groups to bring about the strides in prevention that have already occurred and will likely continue to occur in the United States. For lung cancer, we have etiologic knowledge about a strong risk factor with a high population attributable fraction, and, furthermore, over decades of effort, much has been learned by epidemiologists and other social scientists about the effectiveness of different strategies that have been designed to bring about significant reductions in the prevalence of smoking within a population. However, even in this situation, progress has been slower than many in public health would have hoped or anticipated, given the relatively unambiguous epidemiologic data indicating a strong causal effect of cigarette smoking. The example of smoking and lung cancer illustrates the complexities that can exist in the relationship between epidemiologic knowledge and change in population exposure distributions. The emergence of unanticipated or unforeseen consequences of well-meaning prevention and education strategies may be illustrated by the smoking–lung cancer story: as public health efforts to prevent smoking have gained hold and succeeded among better-educated groups and more affluent communities in this country, US tobacco companies have searched for, and found, other lucrative markets for their products, including those in socioeconomically disadvantaged groups in this country and especially young persons in developing countries (Makary and Kawachi, 1998; Joossens, 2000; Yach and Bettcher, 2000; Yach, 2001; Muggli et al., 2002). As a second example of progress in applying epidemiologic knowledge to primary prevention efforts, consider melanoma. Based on the experience of Australia since 1980, where melanoma incidence and mortality rates are dropping in younger cohorts following intensive primary and secondary prevention public health campaigns (Marks, 1999; 2002), we would argue that there is sufficient epidemiologic knowledge about melanoma to prevent a large proportion of the burden of this deadly disease. There is both etiologic knowledge about strong risk factors as well as examples of how to intervene in populations and the environment to reduce sun exposure and sunburn (Marks, 1999; 2002). Along with positive evidence of the application of epidemiologic findings to primary prevention of cancer, there is also sobering evidence of a lack of significant progress and of a realization of the overly optimistic nature of statements made about the reduction in the cancer burden that would follow on the heels of more etiologic research. For cancers of the breast and prostate, for instance, there is little knowledge about how to achieve significant strides in primary prevention. For breast cancer, this is the case despite decades of study and a fairly strong scientific consensus about etiology. There is widespread scientific agreement that exposure to ovarian hormones, over the course of a lifetime, is a key determinant of breast cancer risk. However, despite the relative wealth of knowledge about this disease, and the strong desire, indeed activism, on the part of the public for primary prevention options, there appear to be few socially acceptable and effective ways to intervene to reduce women’s exposure to ovarian hormones. Strategies to increase the average age at menarche in the population, or to decrease the average age at first birth, or to increase parity, would likely be considered undesirable and unethical by many in this society, and in industrializing societies as well. Many of these societies, which until the final decades of the 20th century had quite low incidence rates of breast cancer in comparison with countries such as the United States, now face incidence rates comparable with those in the United States. The reproductive behaviors of women in these industrializing societies have converged to the pattern of later and less frequent childbearing seen in the United States and other similar countries. So, for this cancer, while there is a large amount of etiologic data, there is little knowledge that appears to be relevant to active interventions in populations. It is likely that chemoprevention with agents such as tamoxifen, raloxifene, and their future derivatives holds the greatest promise for primary prevention of breast cancer (Jordan, 2001).

Increasing the Contribution of Epidemiology to the Primary Prevention of Cancer However, because such chemoprophylaxis will likely always carry its own risks, it will necessitate close involvement with the medical care system, and thus primary prevention will become a medical care issue. Prophylactic mastectomy will also remain an option for women at especially high risk of breast cancer. The etiology of prostate cancer, the most commonly diagnosed cancer in American men, is less well known than that of breast cancer, but, similar to breast cancer, is believed to involve hormonal factors. Also similar to breast cancer, there are no apparently available primary prevention strategies against this disease. Further, there are currently only preliminary chemopreventive strategies against prostate cancer (see chapter 71). Thus, for two of the most common cancers, there is little epidemiologic knowledge that can be readily utilized for significant gains in primary prevention. Despite the inherent practical problems in thinking about modifying population distributions of exposure to oral contraceptives or postmenopausal hormone therapy separately from major social and technological shifts, many of the epidemiologists who have written on the avoidability of cancer have, perhaps too casually, considered such cancers preventable in modern societies. With respect to colon cancer, there appears to be scientific consensus that risk is related to dietary practices and to sedentariness and obesity. However, specific dietary constituents that increase or decrease risk of this disease are the subject of much study and current debate among epidemiologists. For instance, it has been epidemiologic wisdom for several decades that individuals who consume higher levels of fiber, fruit, and green and yellow vegetables would have lower risk of colorectal cancer than those with lower consumption of such items. However, this wisdom has been challenged by findings from large, prospective observational and intervention studies (Gaard et al., 1996; Michels et al., 2000; Schatzkin et al., 2000; Mai et al., 2003) that have shown no association between consumption of these items and colon cancer risk. There are consistent data, though, from randomized trials, indicating that screening for colorectal cancer can lower both incidence and mortality (Mandel et al., 1999). (Screening for colorectal cancer, unlike screening for many other types of cancers, can act as a primary preventive because it enables removal of polyps before they become cancerous.)

Knowledge for Prevention In the above discussion of epidemiologic knowledge that is relevant to prevention, we implied that knowledge of how and why individuals in populations become “less exposed” to risk factors, or achieve a lower risk of disease, is critical to realizing how epidemiologic studies can contribute to significant strides in prevention. The issue of prevention, then, goes beyond the traditional sphere of modern epidemiology, which is strikingly limited to the pursuit of explanation and quantification of disease mechanisms, at both the “social” and “molecular” ends of the etiologic spectrum. Knowledge of the means to prevent cancer encompasses at least two distinct areas: knowledge about mechanisms, and knowledge about interventions/policies to effectively and beneficently enact in populations to reduce carcinogenic exposures and increase protective exposures. The equating of the burden of cancer attributable to non-genetic risk factors with the proportion of cancer that is thus avoidable or preventable is flawed, and will always be overly optimistic, because it inherently fails to distinguish between these two types of knowledge. The equating implies that once the first type of knowledge (causal mechanisms) is available, the second type of knowledge (knowledge about risk reduction in populations) follows, seemingly by default. For epidemiologists to accept this idea, that the complexities of the study and formulation of prevention strategies are part of the disciplinary landscape as much as the complexities of the study of causal mechanisms, we believe they must assent, at least in principle, to the four suggestions we mentioned in our opening section, and that we expand upon below. We hope that, as a discipline, we will commit to prevention, and preventive intervention, as central professional goals; be willing to develop our understanding of cancer causation and prevention from a population perspective as well as from an individual

1295

perspective; accept uncertainty as the “normal” state and be vigilant about the risk of paralysis that uncertainty brings to action; and broaden our concept of causal inference so that the process is useful for primary prevention.

FOUR SUGGESTIONS FOR INCREASING THE CONTRIBUTIONS OF EPIDEMIOLOGY TO THE PRIMARY PREVENTION OF CANCER Epidemiologists Must Be Committed to Prevention, and Preventive Interventions, as Central Professional Goals For some epidemiologists, a commitment to primary prevention as a central professional goal needs no argument; it is akin to stating that physicians are committed to the goals of managing disease and alleviating suffering in individual patients. Implicit in this statement is the notion of service; we believe that this service (to individuals) underlies the profession of medicine, whereas service to the society, or the population, underlies the discipline of epidemiology. But not all epidemiologists view service to society, or alleviation of societal suffering due to disease, as the foundation of our discipline. Weed and Mink (2002) point to a sharp rift in our discipline that reveals two nearly opposite views on this issue. One view asserts that epidemiologists are scientists whose responsibility to the public is no more than that of any other scientist or private citizen. According to this view, epidemiology’s scientific agenda should be guided primarily by the intellectual and creative pursuit of causal explanations for natural phenomena without regard to their potential relevance to prevention and public health. Adherence to such a view would suggest that epidemiologists who advocate for disease prevention strategies in society do so as private citizens, not as professionals. The alternative view, as mentioned above, sees epidemiologists as public health professionals, with responsibilities to undertake scientific studies primarily for the purpose of pursuing practical solutions to public health problems identified by society. If epidemiologists are first and foremost public health professionals, rather than merely scientists, epidemiologists cannot be disengaged from awareness about, and participation with, social forces that work for and against health maintenance and disease prevention in the population. An assumption that “natural” etiologic phenomena that determine patterns of disease and health in the population can be separated from the social context in which individuals reside is a foundational assumption of most modern epidemiologic research. This assumption may be detrimental, however, when it comes to considering how epidemiologic research can be relevant to primary prevention. More generally, the belief that scientific research, especially research involving human actions and activities, can be separated from its social and historical context is a naïve and potentially dangerous myth. This context determines what is scientifically and economically valid to consider as “causes” of disease, and where and how it is feasible to intervene in society to affect disease risk. Philosopher of science, Sandra Harding (1991), distinguishes between “weak objectivity,” or the proposition that scientists can be socially disinterested, socially valuefree, and personally objective in their hypothesis formulation and conclusion drawing, and “strong objectivity,” the proposition that scientists should be able to identify and articulate well to other scientists and to nonscientists the causes of their own beliefs, hypotheses, and actions as professionals. We would argue that the academic stance that epidemiologists have no necessary responsibility (although they always have the “right”, as does any citizen) to directly engage in actions to improve the public health stems, in part, from attempts to apply the myth of weak objectivity, that is, the myth of the disinterested scientist who formulates and tests hypotheses that have arisen with no social influence, to a public health profession. Many epidemiologists have chosen, whether actively or passively, to accept a narrow definition of epidemiology—that of merely testing “association” hypotheses about the effects of a seemingly infinite number of “independent” exposures in large numbers of people. These

1296

PART V: CANCER PREVENTION AND CONTROL

large numbers of people are usually conceived of as a “population” only in the limited sense of “study population.” Such a conceptual view of “population” is not only useless but harmful when it comes to addressing important questions regarding the relationship between etiologic research and primary prevention in the real world. A methodologic view of exposures as independent of each other, and of individuals as autonomous, disconnected actors within a large “n” does not serve the purpose of realistic and sustainable primary prevention in society. One of epidemiology’s most well-known voices on causal inference, Sir Austin Bradford Hill, offers strong supporting arguments for the involvement of epidemiologists in prevention strategies and interventions. According to Phillips (Phillips and Goodman, 2001), Bradford Hill had no intention to provide simply a checklist for inferring causation; his overarching concern was about decision making and preventive intervention. In the introduction to his well-cited paper titled “The environment and disease: association or causation?” (Hill, 1965) Hill noted that the decisive question is whether the frequency of the undesirable event B will be influenced by a change in the environmental feature A. How such a change exerts that influence may call for a great deal of research. However, before deducing “causation” and taking action we shall not invariably have to sit around awaiting the results of that research. (p. 295) Phillips, (Phillips and Goodman, 2001) in his discussion of the misunderstood nature of the lessons of Bradford Hill points out that epidemiologists and other health researchers frequently justify expensive research based on immediate practical benefits, but then go on to deny their responsibility to assess policy considerations related to their work, as Phillips notes, “defending the value of science for its own sake (sometimes as they issue press releases calling for policy responses) (p. 5).” While epidemiologists have been quick to enter the disciplines of genetics and molecular biology in etiologic investigations without necessarily being geneticists or molecular biologists by training, we seem more hesitant, as a discipline, to investigate and discuss the prevention implications of past, current, or proposed directions of epidemiologic research, given an informed assessment of the social context in which we reside. Of course, the common one-disease-at-a-time approach taken by most epidemiologists, not only in specific studies but now often for a career, makes it difficult for many to speak valuably about the broad primary prevention implications of their work, about the overall expected population health risks and health benefits from hypothetical interventions. However, we believe that this difficulty is something we must strive to overcome, rather than surrender to because consideration of the population health implications of different lines of epidemiologic research is too intellectually complex for epidemiologists to handle, and is thus best left to another discipline, such as “health policy” or “health education”. It is, obviously, a truism that choosing to avoid active engagement in endeavors related to primary prevention is, in essence, choosing to do something about prevention, by default. Those institutions and professions that are not charged with serving the public health good, as epidemiologists are, will have disproportionately prominent and even sometime unopposed (by epidemiologists, at least) voices in societal decisions. In this era of commercialization, privatization, and profit, even within university medical and public health schools, we believe that Susser’s statement (Susser and Susser, 1996) is timely and cautionary: In the course of their education, epidemiologists . . . (must be) socialized in a manner that keeps alive the idea of improving the public health as a primary value. Epidemiologists must be scientific but also in some degree professional in the sense traditional to medicine, the law, and the clergy. That is, society accords them a privileged and autonomous function founded on special training. That autonomy carries reciprocal and primary ethical obligations for service to individuals or society (italics ours). (p. 677)

Epidemiologists Must be Interested in a Population and Individual Focus with Respect to Understanding Cancer Causation and Prevention An important goal of modern cancer epidemiology is the improvements in accuracy in identifying high-risk, or susceptible, individuals. Often, “susceptibility” is defined narrowly, in terms of genetic susceptibility, but for the purposes of the issues we raise below, it is the general notion of individual susceptibility that we address. Because of our belief that a focus on populations (in terms of primary prevention strategies) has been under-represented in the epidemiologic literature, our discussion below is more weighted to the “population” perspective. However, we do not wish to set up a false dichotomy between the individual and the population. Much of our discussion below uses Rose’s arguments as a foundation; Rose himself noted that prevention strategies do, and must, embrace both the individual and population perspective. Geoffrey Rose (1985; 1992) wrote of the comparative advantages and disadvantages of “high-risk individual” and population strategies of primary prevention. He described the high-risk strategy as one that relies upon identifying, usually through some type of clinical screening assessment, those individuals believed to be at increased risk of disease, and then targeting these individuals with preventive education and therapy. Choice regarding preventive action resides with the individual, theoretically, in this strategy. The population strategy of prevention, in comparison, rests on the premise that most health problems or conditions are not confined to an identifiable high-risk minority of the population, but rather reflect the conditions and norms of the society as a whole; the population strategy thus seeks to influence social norms that determine the distributions of harmful exposures in the population. The goal of this strategy is to shift the population distribution of risk downward, to center the population average risk around a lower mean, through macro-level societal actions including laws, incentives, public education, and structural environmental changes. The “high-risk individual” and population approaches to primary prevention are clearly not mutually exclusive, and the greatest strides in primary prevention of disease (e.g., lung cancer) have resulted from a combination of such strategies. The key question in any “high-risk individual” strategy is the following: “How will the high-risk individuals be identified?” In some situations, this question is easily answered. If the prevention campaign is targeted against lung diseases, for instance, current and past heavy smokers are readily identifiable as high-risk individuals. There are, as well, certain occupational exposures that also place individuals at markedly elevated risk of some cancers. However, even in the situation of lung diseases such as lung cancer, in which there are risk groups that have quite high relative risks of disease, the population reality is that the bulk of cases will arise from the larger mass of the population that do not fall into easily identifiable high-risk groups. This merely reflects the workings of probability: a large group at lower risk will often contribute more cases than a much smaller group at higher risk. From a “high-risk individual” perspective, heavy smokers need medical attention, surveillance, and education more than other individuals. From a public health perspective, though, substantial reduction in the rates of lung cancer have required, and will require, far more than attention to just heavy smokers. In a simplified way, this example illustrates the need to consider prevention strategies from the two different vantage points, even when a risk factor is available that readily identifies “high-risk” individuals. The above example of heavy smoking and lung cancer illustrates a relatively successful answer to the question of “How will high-risk individuals be identified?” For many of the common cancers, however, there is no clear risk factor “threshold” separating low from high risk (with the exception of the very rare autosomal dominant genetic mutations that place some individuals at markedly higher lifetime risk, sometimes approaching 1.0, of some cancers). The large majority of risk factors studied by epidemiologists, including, now, many genetic polymorphisms, are associated with only modest relative risks of disease. Most factors discriminate poorly between those individuals who will and those who will not get disease. In other words, the sen-

Increasing the Contribution of Epidemiology to the Primary Prevention of Cancer sitivity and specificity of most of the risk factors studied is poor; the large majority of individuals who are labeled “high risk” based on exposure status will not get disease, and, concomitantly, a very large proportion of disease cases that arise in the population will come from those individuals allegedly at “average” or “below average” risk. Given this reality, the terminology that labels individuals “high,” “average,” or “low” risk can be problematic; it can produce unnecessary and even harmful anxiety in some, and an unwarranted sense of immunity and complacency in others. Further, and importantly from a public health standpoint, in the absence of risk factors or combinations thereof that serve as highly accurate screeners of individuals, any strategy that relies upon identification of high-risk individuals will not have a large impact on the burden of disease, unless the definition of “high risk” encompasses a relatively large proportion of the population—that is, unless the designation of “high risk” becomes very sensitive at the expense of specificity. Thus, in most situations, even a high-risk strategy, while perhaps nominally individually focused, will have to affect a large proportion of the population to have a meaningful impact on disease burden. Further, for a “high-risk individual” strategy to have such an effect, societal-level forces will nearly always have to be brought to bear (e.g., mass marketing of advice to see physicians regularly for prevention counseling; mass advertising of drugs to “at-risk” individuals; mass advertising of the need for women to be tested for genetic mutations related to breast cancer). In clinical terms, we could equivalently state that, for many identified exposure-disease associations, the estimated “number needed to treat” or “number needed to change” to prevent a single case of disease is substantially greater than 1. It is the goal of the burgeoning field of genetic epidemiology to change this situation, to find markers of genetic susceptibility to disease that will greatly increase the ability of scientists and clinicians to identify those at truly high risk of disease. To date, however, virtually all of the studies of common genetic polymorphisms and their interactions with environmental exposures have produced relative risks that are of the same modest magnitude (i.e., <3.0) as the conventional risk factors. Pritchard, a population geneticist, uses evolutionary theory to argue that it is unlikely that common, even moderately penetrant alleles for the complex, long-latency diseases will ever be found (Pritchard, 2001). Scientific findings to date bear out this view; the most highly penetrant alleles (e.g., mutations in BRCA1/2) are very rare, while the comparatively common susceptibility polymorphisms (e.g., those connected with NAT1/2) have been found to be associated with only modest increases in risk, if any association has been noted at all. Recent research (Begg, 2002) on BRCA1/BRCA2 mutations, for instance, indicates that the penetrance of these mutations is probably likely far lower than has previously been estimated, and that substantial heterogeneity in breast cancer risk exists among women with these mutations. Even if strong genetic indicators of cancer risk are found, it is unclear how such discovery will contribute to meaningful gains in primary prevention, particularly if multiple etiologically related genes are involved at different times in the life course, as is hypothesized to be the case. Undoubtedly, there are genetic components involved in all cancers. For most of the cancers now common in societies such as the United States, however, even a cursory reading of history will show that incidence and mortality rates have risen and/or fallen over time periods too short to reflect meaningful changes in population genetic frequencies. We are not arguing here for insensitivity to the needs of those who are at unusually high risk of disease due to genetic or other factors. Surveillance and concern directed toward such individuals is appropriate, compassionate, and can be effective from a clinical perspective. We are instead merely arguing that from a public health perspective, a focus on the small numbers of individuals who can accurately be identified as having an unusually high risk of disease is inadequate. It is inadequate precisely because of the small numbers of such individuals, and the consequent reality that most individuals who contribute to the disease burden are those who will not be identified as “high risk”. We are also not arguing for the scientific unimportance of genetic factors in cancer etiology. From the standpoint of considering primary

1297

prevention strategies, however, we are not the first to raise questions about the practicality of a strong research focus on genes, given the (at least foreseeable) unmodifiability of genes in individuals and the currently scientifically unsupportable notion of accurately identifying those individuals who are “susceptible” or “nonsusceptible” to effects of exposures, based on genes alone. Rose believed that “mass diseases and mass exposures require mass remedies” (Rose, 1992). This statement was not just an opinion in support of a particular approach to primary prevention over another. Rather, it reflects a quantitative reality, in the face of the poor ability to accurately identify those who will develop disease in the future. Even those who favor individually focused strategies nevertheless must acknowledge that to achieve substantial reductions in the burdens of common diseases, remedies that affect a substantial proportion of the individuals in the population will be needed. Is it possible that large strides in primary prevention of the major cancers will come about through strategies that are dependent for their success on many individuals autonomously and simultaneously choosing to make requisite risk reductions in their lives, based on information on personal risk of disease that they have received from health-care providers or the mass media (which now actively act as filters of many epidemiologic findings for the public)? Perhaps. However, from the large, and largely discouraging, literature on motivating healthy individuals to make long-lasting lifestyle changes to lower risks of future disease, we believe that Bailar’s words (1979) are true: It is difficult to interest healthy people in preventing any chronic disease that has multiple causes, that cannot be completely prevented by a few simple steps, and that may not occur for decades anyway. Most people want to be kept well by things that do not involve personal inconvenience or behavioral change. (Bailar, p. 730) The use of the phrase “want to be kept well” is noteworthy, we believe. As public health professionals, we are entrusted with the responsibility, not just the right, to direct our professional work to making it easier for as many people as possible to live healthier lives. The search for etiologic factors and biological mechanisms is but one type of means to this end. Once questions about etiologic mechanisms have been asked, and addressed with the best available methods, we believe that an equally important epidemiologic question still remains: What are the causes of the exposure to the risk factor, and how can such causes be altered, in a manner that enhances the health of the population? Posing the question in this manner highlights the reality that the elucidation of biological mechanisms will not, by itself, serve the broad goal of primary prevention.

Epidemiologists Must Accept Uncertainty as the “Normal” State and Be Vigilant about the Risk of Paralysis That Uncertainty Brings to Action The issue of uncertainty with respect to knowledge of causal risk factors, touched upon in the discussion of unanticipated observational findings about dietary risk factors for colon cancer that have challenged long-accepted wisdom, is of critical importance in epidemiology, and in the consideration of primary prevention of cancer. Doll and Peto (1981) stated that “the simplest evidence of the preventability of cancer would be the demonstration by scientific experiment that a particular action actually leads to a reduction in the incidence of the disease. (p. 1202)” However, the hypothesized relationships between most cancer risk factors (or protective factors) and specific cancers are, and most likely will continue to be, based on observational epidemiologic data, because of the practical and ethical difficulties involved in conducting randomized trials of long latency, and complexly interdependent, “lifestyle” factors. The single word “uncertainty” encompasses qualitatively different meanings in epidemiology. Statistical uncertainty, or the uncertainty that is attributable to random sampling error, is perhaps the simplest type of uncertainty to consider; classical statistical tools and tests allow ready quantification of such uncertainty, or imprecision. Uncertainty due to systematic error (or “bias”) receives much attention in

1298

PART V: CANCER PREVENTION AND CONTROL

discussion of epidemiologic concepts and methods, although it is very difficult, and sometimes impossible, to quantify with accuracy. Even the most sophisticated statistical methods or models will not allow an investigator to overcome strong biases due to selection, confounding, and so forth. The observational studies of postmenopausal hormone therapy and coronary heart disease were analyzed with some of the best available epidemiologic methods, including stratification, matching, restriction, and inclusion of known confounders into statistical models, but it appears that these methods were inadequate to overcome the strong presence of confounding by indication in such studies (more on this below). A third type of uncertainty that has received little discussion in epidemiology concerns the discrepancy between aggregate-level and individual-level conclusions. The best epidemiologic study, even a very large randomized trial with few design flaws, will not produce information that will provide certainty to specific individuals. For instance, the findings from the Women’s Health Initiative clinical trial about the harmful effects of estrogen plus progestin postmenopausal hormone therapy with respect to coronary heart disease risk (Writing Group for the Women’s Health Initiative Investigators, 2002) speak to an average effect—not all women will be harmed by this therapy, and a finding of average harm obviously does not preclude the possibility that some individual women will benefit. Thus, even though there may be little statistical uncertainty and little systematic error or bias in a study result, the result does not allow one to speak with certainty to specific individuals about the individuallevel effect (causal, protective, or null) that they will experience. For some risk factors, such as heavy cigarette smoking and certain occupational exposures, the observance of consistently highmagnitude relative risks and of strong monotonic dose-response gradients, along with supportive experimental data, have left little room for scientific doubt that the relationships are causal. The prevalence and strength of association that a confounder would have to have to “explain away” such a strong (i.e., high-magnitude relative risk), seemingly causal, relationship between an exposure and disease are rarely observed (Flanders and Khoury, 1990). Most cancer risk factors, however, are associated with only modest (i.e., <2.0) relative risks. In these circumstances, when the cancer rates among the exposed are only a modest multiple of those among the unexposed, more severe problems of causal interpretation arise, and it may be difficult to disentangle the contribution of biased information from that of cause and effect (Schwartz and Diez-Roux, 2001). Thus, it is not surprising that observational epidemiology studies have not provided consistent evidence on hypothesized causal or preventive agents. Case-control studies have sometimes produced different findings than prospective cohort studies; this has happened, for instance, with induced abortion and breast cancer (e.g., [Daling et al., 1994; 1996; Melbye et al., 1997]) and dietary fat and breast cancer (Hunter et al., 1996; Prentice, 1996). The likely possibility of recall bias has been raised by some who have criticized the case-control findings on these topics (Willett et al., 2000). However, observational cohort studies also can suffer from potentially serious bias. This possibility was raised most strikingly by findings from randomized clinical trials showing that estrogen plus progestin postmenopausal hormone replacement therapy (HRT) caused an increase in risk of coronary heart disease events in women (Hulley et al., 1998; Writing Group for the Women’s Health Initiative Investigators, 2002). These findings stand in stark contrast to findings from observational studies published since the 1970s suggesting that hormone replacement therapy could reduce, even halve, the risk of coronary heart disease. Thus, this particular difference between the observational study results and the results of randomized trials is one of qualitative effect (harm vs. protection), not just of quantitative magnitude of effect. Although there is no clear scientific consensus on the reasons for the discrepancy, it is likely that a part, if not most, of the explanation has to do with the inability to fully eliminate the possibilities of confounding (by indication and by selection, in this situation of HRT use) in observational studies (Michels, 2002; Piantadosi, 2003; Whittemore and McGuire, 2003). More specific to cancer, the ATBC (Alpha-Tocopherol and Beta Carotene) (ATBC, 1994) and CARET (the beta-Carotene and Retinol Efficacy Trial) (Omenn et al., 1996a; Omenn et al., 1996b) random-

ized trials similarly called into question the validity and generalizability of findings obtained from observational epidemiology studies. Because of the difficulty, perhaps sometimes the near impossibility, of isolating modest independent causal effects of factors that tend to cluster strongly in free-living individuals, we believe that the question of whether epidemiologists know as much as we think we do about etiologic relationships between specific exposures and specific cancers must be considered. Most of the findings on such relationships have been (and likely will continue to be) gleaned from observational rather than experimental study. The central question when applying epidemiologic knowledge to primary prevention strategies is, of course, “Will the strategy work to reduce risk in the population?” That is, will a decline in exposure to a purported risk factor, or an increase in exposure to a protective factor, cause a decline in the rate of disease over what would have been observed in the absence of the strategy? Lest we give the impression that we believe that only experimental trials can provide a good answer to this question, we wish to note that obtaining an accurate answer is not only a matter of experimental vs. observational data. The answer to the question will rarely be a certain “yes” even if results from good experimental (randomized) trials in humans support the notion of intervention. Randomized trials in humans have their own limitations when it comes to generalizability to free-living society, in which there is no control over related variables and risk factors. Epidemiologists who are dedicated to employing epidemiologic knowledge to successfully reduce disease risk in populations will need to consider more information than merely the results of (observational or experimental) epidemiologic studies; they must concern themselves with the social and political context in which any proposed intervention will occur, and they should also have a good grounding in the history of public health failures and successes, with respect to prevention interventions. Even if epidemiologists could say, with a high degree of certainty, that an intervention would work to reduce average cancer risk in a population (and especially if a high degree of certainty is lacking), another important question then becomes, “Does this intervention carry risks that could rival the cancer risk reduction benefit?” Virtually all interventions to reduce disease risk in free-living populations will carry some risk of their own, in the sense that some individuals will experience harm (physical or psychological) that they would not have had there been no intervention. Thus, the argument that some may be harmed due to intervention, is, by itself, a weak argument against active engagement in prevention intervention. Of course, this does not mean that attempts should not be made to reduce the probability of and potential seriousness of harm to as low a level as possible. Some types of interventions carry more risks than others. Rose noted that chemoprevention with pharmacologic agents may be particularly likely to carry unforeseen adverse events, and that such events can only be identified and measured by long-term clinical trials, which are difficult to carry out (Rose, 1992). He proposed that interventions based on an historical or even evolutionary perspective on what constitutes a “healthy” or “natural” human lifestyle (e.g., interventions designed to reduce consumption of refined foods and increase consumption of unrefined grains, fruits, and vegetables, or interventions designed to increase physical activity through means such as encouraging walking [though not necessarily jogging or high-impact aerobics]) were less likely to carry large adverse risks. Although Rose did not explicitly mention this, the one-disease-ata-time approach that epidemiologists often adopt in their considerations of etiology and primary prevention is relevant to his distinction between pharmacologic agents and broad improvements in “lifestyle”. Some preventive strategies, especially those based on pharmacologic agents, are targeted at one outcome only—for example, tamoxifen is targeted at prevention of breast cancer, cholesterol-lowering medications are targeted at prevention of coronary heart disease. The importance of very low estimated risk of adverse events for such focused preventive measures is clear: even a small risk of adverse events, when multiplied across the large numbers of people that may be exposed to the preventive measure, can result in a large number of incidents that may rival the number of prevented cases of the single disease of concern. By contrast, when a preventive measure is proposed that will

Increasing the Contribution of Epidemiology to the Primary Prevention of Cancer lower risk of several major diseases (e.g., measures that discourage smoking across the population; measures that enable and encourage a larger proportion of the population to engage in moderate physical activity and to maintain a stable weight; measures that enable and encourage a larger proportion of the population to eat a diet that is rich in fruits, vegetables, and whole foods) a small risk of adverse events is more likely to be balanced, or superseded, by the larger possibility of an overall health benefit. Sometimes it is difficult, if not impossible, to do an accurate risk/benefit analysis of a proposed or hypothesized preventive measure. This does not mean such an analysis should not be attempted, and done as well as possible with existing information. However, given the lack of certainty that is implicit in such calculations, we believe it is important to state that epidemiologists interested in primary prevention of disease cannot let a lack of proof of “no adverse risk” serve as a reason to do nothing. It may be that the ethical principle of “Do no harm,” while obviously important, is overused in our discipline. What we want to do is find ways to benefit communities and populations while producing as little harm as possible. That is the ethical principle of beneficence, and not nonmaleficence, in action. “Do no harm” comes from the ethical principle of nonmaleficence and it refers to the obligation to refrain from intentional harm. “Prevent harm” and “do more good than harm” both come from the principle of beneficence, which is the foundational principle of public health. Some will still express discomfort with epidemiologists’ active and professional engagement in advocating for primary prevention strategies. To these individuals, we point out that a decision to not advocate, to attempt to remain neutral in the presence of their own specialized knowledge, and in the face of societal interests (e.g., the fast-food industry, the rapid weight-loss industry, and so forth) that are not charged with the professional responsibility of putting public health first, does not have a neutral effect on public health.

Epidemiologists Must Broaden Their Concept of Causal Inference So That the Process Is Useful for Primary Prevention How do we accept the large degree of uncertainty that is fundamental in our causal inference, and yet make real strides in prevention? This question is, at heart, asking how we can make good and useful inferences when it comes to prevention. Inference for prevention is broader in terms of the considerations it encompasses than the traditional “one exposure–one disease” process of causal inference usually practiced by epidemiologists—this type of causal inference is necessary for, but not sufficient for, good inference for prevention. (We note here that any discussion of inference cannot avoid consideration of the extent to which the supporting evidence for a claim or an action is potentially biased; we are not attempting to engage in a battle between the common view of inference and some other view.) The literature on causal inference and causal reasoning is vast, and spans many academic disciplines. We do not intend to review it here. Instead, we focus on a specific aspect of causal inference, its inherent arbitrariness, and we discuss the connection between how we choose to frame our questions about causality, and resulting options for prevention. Karhausen (2000) points out that scientists can never absolutely explain the causes of a specific event, or of a general phenomenon like disease; they explain such things only in reference to certain assumptions and descriptions. Inferring causality, both for specific and general phenomena, requires assumptions about what is “given” as background, or what Schwartz (Schwartz and Diez-Roux, 2001) calls assumptions about the “tacit causal field.” What epidemiologists take as the tacit causal field will vary with their scientific and social beliefs, as well as with social and economic forces like funding opportunities, scientific prestige, political feasibility, and so forth. The ideologies of privatization, scientific entrepreneurship, and market individualism, which now penetrate the academic world, influence where we look for causes, and what is politically and economically feasible to label as the important causes of disease.

1299

Consider an important risk factor for cancer usually identified by epidemiologists—tobacco smoking. A genetic or molecular epidemiologist focusing on DNA repair mechanisms and differences in individual susceptibility to the carcinogenic effects of smoking will focus on the presence of certain genetic and molecular attributes as the major cause of lung cancer, assuming a given background or tacit causal field of exposure to cigarette smoking and other carcinogens. In this view, the relevant level of pathology is subcellular. A “risk factor” epidemiologist who focuses on lifestyle characteristics of individuals will identify smoking behavior, particularly behavior that starts at an early age and that involves heavy smoking, as the important cause of lung cancer, often assuming a tacit causal field of other lifestyle factors and attributes, and of social, economic, and psychological forces that help to shape patterns of uptake of cigarette smoking within a society. In this view, the individual is the relevant unit of pathology. A sociologist or social epidemiologist who focuses on broad social and economic forces that shape disease patterns in populations might identify corporate and governmental practices surrounding the growing and marketing of tobacco, and the differential vulnerability of population subgroups to the messages in such marketing, as the real causes of lung cancer. The tacit causal assumption in such a view would be one of homogeneity within population subgroups; in such a view, the social group or population might be considered the relevant unit of pathology. With the above examples, we have tried to illustrate the point that we find causes where we look. There is no such thing as the cause (or even the causes) of an event or a phenomenon that can be considered separately from the investigator’s interest (VanFraassen, 1980). While we would argue that none of the above-described causal inferences are either completely correct, or wrong, and that all of the views taken together contribute to a fuller picture of exposure and disease than any one view by itself, it is clear that, besides illustrating different investigator interests and values and correspondingly different assumptions about constant backgrounds or tacit causal fields, the above examples have very different implications for primary prevention. Most epidemiologists assign intellectual priority to causes of disease based on certainty, necessity, and proximity to disease occurrence, where proximity is defined both in terms of closeness in time to disease identification and in terms of closeness to level of disease definition. In contrast to a hierarchy based on necessity and proximity to disease occurrence, Geoffrey Rose promoted the idea of a causal hierarchy based on efficiency and population-wide safety of potential primary prevention strategies (Rose, 1992; Schwartz and Diez-Roux, 2001). Schwartz and Diez-Rouz (2001) state that “for Rose, the hierarchy is based in the efficiency with which the removal of a cause could potentially decrease the incidence of the disease” (italics ours) (p. 436). This hierarchy is inherently premised on the notion that, when conducting causal inference with prevention in mind, it is the population that gets disease, and that must keep itself healthy. Causes that are very proximal to disease occurrence may have little practical relevance when it comes to prevention. For common noninfectious conditions, it is often true in our current scientific models of disease that the closer a cause is to being universally necessary, the closer it is to being definitional, or at least an early stage, of the disease itself, and thus the less useful is knowledge of the cause for public health primary prevention. For instance, hyperplastic changes in tissue, or a particular sequence of DNA mutations, might be “universally necessary” to cause a particular cancer; trivially, formation of a blood clot is a universally necessary cause of stroke. We would argue that causal inference for prevention must go beyond the conventional bounds of inference in epidemiology, and grapple with the issue of preventive relevance in an honest way: what are the realistic implications of a certain type of research for effective primary prevention in the population? How does a potential “cause” that is being studied relate to a potential preventive strategy, based on our knowledge of how changes in public health have occurred in the past, as well as on our knowledge of social, economic, technological, and political realities in society? We are not arguing that every epidemiologic study must have immediate relevance to prevention. We are instead advocating for the open and informed consideration of prevention

1300

PART V: CANCER PREVENTION AND CONTROL

implications of certain lines of research. If this means that epidemiologists must take upon themselves the learning (or re-learning) of areas of knowledge related to public health history and the influence of social and political forces on risk factor distributions and disease trends, just as many have taken up the task of learning molecular biology and genetics, we would advocate for such intellectual stimulation. Further, we believe that if there are no prevention implications of a line of research that are immediately apparent, it is better to state this up front than to try to persuade colleagues and the public that the research is of clear relevance to primary prevention, merely because this is what everyone wants to hear. As Mechanic (1997) notes, many positive or damaging healthrelevant behaviors, including smoking, physical activity, dietary intake, maintenance of a recommended weight, ability to adhere to physician recommendations, ability to avoid harmful toxins or pollutants in the environment or in the workplace, and so forth, arise from the routine activities and conventional patterns of everyday life in a given society or social group, and are only modestly influenced, for many people, by personal consideration of long-term “individual risks” of diseases. Mechanic (1997) argues that changing health-damaging behaviors thus depends on the abilities of societies and communities to create, or, in some cases, re-create, environments and activities of daily living such that more individuals are able to readily choose behaviors that will contribute to maintaining health as long as possible. Both the desire to create or restructure environments and mass activities, as well as opposition to such intervention, are inextricably linked to social and political values. Some of the instruments that can be used to address the causes of risk factor distributions are tax policies and other monetary incentives, legal regulation, skillful use of the mass media, and public education. Mechanic’s view is similar to that of Link and Phelan (1995), who have argued that, to achieve maximal strides in primary prevention of the common diseases in society, epidemiologists must pay greater attention to the social and environmental conditions that affect multiple disease outcomes through multiple, more proximal or intervening, mechanisms. Link and Phelan (1995) define a “fundamental cause” of disease as one involving access to resources that help individuals avoid diseases and their negative consequences through a variety of mechanisms. In the words of Link and Phelan (1995): even if one effectively modifies intervening mechanisms or eradicates some diseases, an association between a fundamental cause and disease will reemerge. As such, fundamental causes can defy efforts to eliminate their effects when attempts to do so focus solely on the mechanisms that happen to link them to disease in a particular situation. (p. 81) The work of Mechanic, and Link and Phelan, is related to Rose’s hierarchy of causation; these researchers argue that the most effective and sustainable way to reduce the burden of major diseases is to pay attention to the more distal factors that tend to put people at a lower, or higher, risk of a variety of diseases. Even though many cancer risk factors, such as smoking, inactivity, poor diet, obesity, or high alcohol consumption, are individual behaviors in an essential sense, asking questions about the causes of these risk factors and their distributions dispels the illusion that risk factor distributions are what they are simply because of autonomous and independent “free” choices made simultaneously by all individuals in a society. There is a supraindividual social context that determines the (usually narrow) range of choices that any individual has, as well as the long-term sustainability of any choice or attempt at change. This context affects in a very real way the feasibility of idealized “magic bullet” approaches to prevention that are targeted at only one link along a causal chain, usually a very proximal biological link. The detailed molecular biological knowledge that has been gained about the HIV virus, for instance, and about the mechanisms by which this virus eventually overwhelms the immune system, has had limited relevance to primary prevention progress. Progress in primary prevention, while obviously based on a certain amount of “universal” biological knowledge, is at least equally rooted in culture-specific understanding of social and psychological forces that operate to place certain subgroups of a society at particularly high risk of infection. By analogy, those who predict a future in

which detailed knowledge of the molecular biology and genetics of cancer will allow accurate risk screening in individuals, accompanied by meaningful strides in prevention, must acknowledge that many social, economic, and psychological forces will determine whether any potent etiologic knowledge that does arise is indeed readily translatable to prevention in society. Besides providing a fuller, and thus more realistic, picture of disease etiology than the currently nearly exclusive epidemiologic focus on individual-level and sub-individual–level factors allows, attention to the determinants of risk factor distributions is important in preventive inference for another reason. Because these determinants, such as social integration, norms and legislation supportive of public health, broad access to preventive health care and to accurate and appropriately complex health information, and so forth, are related to many proximal risk factors simultaneously, and because many cancers (and many conditions of poor health, both objectively and subjectively defined) share a similar list of risk factors, a focus on altering the fundamental causes of risk factor distributions offers a potentially high net benefit relative to preventive action targeted at one exposure or one disease. Efforts to shift population distributions of more proximal risk factors or harmful exposures through a focus on the determinants of these distributions can bring a large benefit to many people, in terms of reducing myriad disease risks simultaneously, thus averting the need to delude ourselves that we can (or will be able to) determine which of a variety of possible diseases will strike a person with a particular risk factor profile in his or her future. Further, such strategies need not pose large risks to individuals. Precisely because strategies that focus on the structural and environmental determinants of risk factor distributions are not reliant on a mass of individuals simultaneously conducting detailed, personal, and “rational” risk/benefit calculations to achieve improvement in public health, there is less potential for causing confusion, fear, and anxiety among individuals about their future; thus, individual autonomy, in the full sense of self-governance and moral independence from influence by others, is better preserved. It may be easy for some to dismiss our discussion about social and environmental determinants of risk factor distributions as “unscientific,” or at least as not scientific in the same sense that dose-response modeling, blood sampling, and DNA arrays are. A central theme of this chapter is that epidemiologists are not only scientists. The use of the biological sciences can be a means to an end in our discipline, but it is not an end in itself. Biological reductionism has become the mainstream guiding philosophy in epidemiology, rather than being viewed as a tool that is more useful for some purposes (e.g., certain etiologic questions) and less useful for others (e.g., effective and sustainable prevention strategies in the population). We hold, as did Rose, that the philosophy of reductionism, as it steers the quixotic search for necessary or universal causes in epidemiology, may distract us from consideration of the prevention implications of our research. It may also distract us from encouraging preventive actions and interventions designed to enable more individuals to easily choose healthy lifestyles, based on the wealth of knowledge that is already in existence about the root causes of ill health. Finally, a focus on biological reductionism as an end in itself can profoundly distract us from recognizing that there are fundamental limits to scientific knowledge and to our ability to answer certain types of questions, even while we already know much about the determinants of ill health in populations and while we are aware of social forces that work, often unchecked, against health maintenance and improvement and disease prevention.

CONCLUSION By broadening, and not just deepening, our research focus, cancer epidemiologists can increase their contribution to primary prevention of different cancers. We have offered four suggestions to this end, all revolving around the central theme of commitment, in epidemiologic research, to the public health goal of primary prevention. We recognize, as Bailar (1979) pointed out, that a commitment to primary prevention of cancer is no more likely, necessarily, to eradicate the threat of cancer than is a central commitment to screening or

Increasing the Contribution of Epidemiology to the Primary Prevention of Cancer therapy. Further, success in primary prevention may not even save money, at the societal level. As Bailar (1979) noted, the reasons for concern about cancer prevention are ethical ones, having to do with individuals’ quality of life and manner of death, and the difficult stresses that cancer places upon those who are ill, as well as upon their friends and family. Although there are some major cancers, such as prostate cancer, for which little is known regarding potential primary prevention strategies, there are other cancers for which important knowledge regarding primary prevention does exist. There are often serious barriers to the effective application of such knowledge, however, and we believe epidemiologists can increase their contribution to primary prevention only by fully taking on, in their research questions and activities, issues of how knowledge generated in epidemiologic studies can and will serve society. What are the barriers to utilization of knowledge about cancer prevention, and is it realistic to suppose that these barriers can be overcome, that risk factor distributions can indeed shift substantially? What are the processes by which decision makers in society act upon scientific knowledge related to cancer risks? What are the risks and benefits to society of acting upon epidemiologic knowledge? We believe that if epidemiologists considered these types of broader questions related to the use and value of research, the specific research questions themselves may broaden beyond the conventional “What is the (causal) association of exposure A with cancer X”? Hill, in his well-cited paper on causal criteria (Hill, 1965), concludes with a much less cited section entitled “The Case for Action,” where he notes that All scientific work is incomplete—whether it be observational or experimental. All scientific work is liable to be upset or modified by advancing knowledge. That does not confer upon us a freedom to ignore the knowledge we already have, or to postpone the action that it appears to demand at a given time. (p. 12) With such a statement, Hill implies that the end of epidemiologic research is not the measure of causal association, but the use of that information in public health action. By urging that epidemiologists concern themselves with the complex questions regarding how cancer prevention can occur in society, we are encouraging, as stated above, a broadening of epidemiologic research questions. This broadening is particularly important as “cancer epidemiology” increasingly becomes equated with “molecular epidemiology” or “genetic epidemiology.” A broadening of cancer epidemiology research questions to areas of population dynamics, prevention interventions, prevention policy, social forces that act to determine population distributions of risk factors, and so forth, would reflect a return, in a sense, to our roots as public health servants. We recognize that there is no guaranteed “payoff” in terms of success in primary prevention of major cancer if epidemiologists were to broaden research agendas to include questions about the relationships between epidemiologic knowledge and primary prevention. There is no guaranteed success for at least several reasons: the inherent resistance to change by individuals (scientists in the research establishment as well as individuals in the public) and by government agencies and other institutions, including academic centers, and the long latency period and complex etiologic chain that likely underlie many cancers. Nonetheless, the barriers to success in primary prevention are not necessarily any greater than the barriers in treatment, and the potential gains are larger. Bailar (1979) pointed out that a sharp distinction exists between an understanding of cancer causation—carcinogenesis—and an understanding of cancer prevention. The failure to recognize this distinction has led to the overly optimistic equating of the proportions of cancers “attributable” to non-genetic or “environmental” causes with the proportions of cancers that are thus avoidable or preventable. The goal we have focused on is progress in prevention, and not promotion of one methodologic or conceptual framework over another, nor promotion of adherence to one ideology over another. The primary prevention of cancer is a complex issue, and progress, when it comes, will come from appropriately complex and flexible knowledge of

1301

disease etiology, individual actions, choices, and constraints, and from a disciplinary awareness of social forces operating on individuals. References The Alpha-Tocopherol Beta-Carotene Cancer Prevention Study Group. 1994. The effect of vitamin E and beta carotene on the incidence of lung cancer and other cancers in male smokers. N Engl J Med 330:1029–1035. Bailar J. 1979. The case for cancer prevention. J Natl Cancer Inst 62(4):727–731. Bailar J, Gornik H. 1979. Cancer undefeated. N Engl J Med 336(22): 1569–1574. Bailar J, Smith E. 1986. Progress against cancer? N Engl J Med 314:1226– 1232. Bailar J, Smith E. 1987. Have we reduced the risk of geting cancer or of dying from cancer? An update. Med Oncol & Tumor Pharmacother 4(3/4):193–198. Begg C. 2002. On the use of familial aggregation in population-based probands for calculating penetrance. J Natl Cancer Inst 94(16):1221–1226. Cramer D. 1974. The role of cervical cytology in the declining morbidity and mortality of cervical cancer. Cancer 34:2018–2027. Daling J, Brinton L, Voigt L, et al. 1996. Risk of breast cancer among white women following induced abortion. Am J Epidemiol 144:373–380. Daling J, Malone K, Voigt L, White E, Weiss N. 1994. Risk of breast cancer among young women: Relationship to induced abortion. J Natl Cancer Inst 86:1584–1592. Doll R. 1996. Nature and nurture: Possibilities for cancer control. Carcinogenesis 17:177–184. Doll R, Peto R. 1981. The causes of cancer: Quantitative estimates of avoidable risks of cancer in the United States today. J Natl cancer Inst 66:1191–1308. Flanders W, Khoury M. 1990. Indirect assessment of confounding: Graphic description and limits on effect of adjusting for covariates. Epidemiology 1:239–246. Gaard M, Tretli S, Loken E. 1996. Dietary factors and risk of colon cancer: A prospective study of 50,535 young Norwegian men and women. Eur J Cancer Prev 5:445–454. Harding S. 1991. Whose Science? Whose Knowledge? New York, Cornell University Press. Higginson J. 1969. Present trends in cancer epidemiology. Proc Can Cancer Conf 8:40–75. Hill A. 1965. The environment and disease: Association or causation? Proc Royal Soc Med 58:295–300. Hoover R. 1996. Breast cancer: geographic, migrant, and time-trend patterns. In: J Fortner, ed. Accomplishments in Cancer Research. New York, Lippincott-Raven, pp. 106–124. Howe H, Wingo P, Thun M, et al. 2001. Annual report to the nation on the status of cancer (1973 through 1998), featuring cancers with recent increasing trends. J Natl Cancer Inst 93:824–842. Hulley S, Grady D, Bush T, et al. 1998. Randomized trial of estrogen plus progestin for secondary prevention of coronary heart disease in postmenopausal women. The Heart and Estrogen/progestin Replacement Study (HERS) Research Group. JAMA 280:605–613. Hunter DJ, Speigelman D, Adami HO, et al. 1996. Cohort studies of fat intake and the risk of breast cancer: A pooled analysis. N Engl J Med 334:356–361. Joossens L. 2000. From public health to international law: Possible protocols for inclusion in the Framework Convention on Tobacco Control. Bull World Health Organ 78(7):930–937. Jordan V. 2001. Chemoprevention with antiestrogens: The beginning of the end for breast cancer. Daniel G. Miller Lecture. Ann N Y Acad Sci 952:60–72. Karhausen LR. 2000. Causation: The elusive grail of epidemiology. Med Health Care Philos 3:59–67. Krieger N. 1990. Social class and the black/white crossover in the age-specific incidence of breast cancer: A study linking census-derived data to population-based registry records. Am J Epidemiol 131:804–814. Lane-Claypon JE. 1926. A further report on cancer of the breast with specific reference to its associated antecedent conditions. Reports on Public Health and Medical Subjects no. 32. London, Ministry of Health, HMSO. Link BG, Phelan J. 1995. Social Conditions as Fundamental Causes of Disease. J Health Soc Behav (special no.) 80–94. Mai V, Flood A, Peters U, et al. 2003. Dietary fibre and risk of colorectal cancer in the Breast Cancer Detection Demonstration Project (BCDDP) followup cohort. Int J Epidemiol 32:234–239. Makary M, Kawachi I. 1998. The International Tobacco Strategy. JAMA 280(13):1194–1195. Mandel J, Church T, Ederer F, Bond J. 1999. Colorectal cancer mortality: Effectiveness of biennial screening for fecal occult blood. J Natl Cancer Inst 91:434–437.

1302

PART V: CANCER PREVENTION AND CONTROL

Marks R. 1999. Two decades of the public health approach to skin cancer control in Australia: Why, how and where are we now? Australas J Dermatol 40:1–5. Marks R. 2002. The changing incidence and mortality of melanoma in Australia. Recent Results Cancer Res 160:113–121. Marshall B. 2002. Helicobacter pylori: 20 years on. Clin Med 2(2):147–152. Mechanic D. 1997. The social context of health and disease and choices among health interventions. In: Brandt A, Rozin P, eds. Morality and Health. New York; Routledge: pp. 53–78. Melbye M, Wohlfahrt J, Olsen J. 1997. Induced abortion and the risk of breast cancer. N Engl J Med 336:81–85. Michels K. 2002. Hormone replacement therapy in epidemiologic studies and randomized clinical trials: Are we checkmate? Epidemiology 14:3–5. Michels K, Giovannucci E, Joshipura K, et al. 2000. Prospective study of fruit and vegetable consumption and incidence of colon and rectal cancers. J Natl Cancer Inst 92:1740–1752. Muggli M, Pollay R, Lew R, Joseph A. 2002. Targeting of Asian Americans and Pacific Islanders by the tobacco industry: Results from the Minnesota Tobacco Document Depository. Tob Control 11(3):201–209. Nagata C, Kawakami N, Shimizu H. 1997. Trends in the incidence rate and risk factors for breast cancer in Japan. Breast Cancer Res Treat 44:75–82. Nomura A. 1996. Stomach cancer. In: Schottenfeld D, Fraumeni J, Jr., eds. Cancer Epidemiology and Prevention. New York, Oxford University Press: pp. 707–724. Omenn G, Goodman G, Thornquist M, et al. 1996a. Risk factors for lung cancer and for intervention effects in CARET, the Beta-Carotene and Retinol Efficacy Trial. J Natl Cancer Inst 88:1550–1559. Omenn G, Goodman G, Thornquist M, et al. 1996b. Effects of a combination of beta carotene and vitamin A on lung cancer and cardiovascular disease. N Engl J Med 334:1150–1155. Parsonnet J, Friedman GD, Vandersteen DP, et al. 1991. Helicobacter pylori infection and the risk of gastric carcinoma. N Engl J Med 325:1127–1131. Phillips C, Goodman K. 2001. The messed lessons of Sir Austin Bradford Hill. Unpublished manuscript. Available at: http://www.epiphi.com/papers/. Piantadosi S. 2003. Larger lessons from the Women’s Health Initiative. Epidemiology 14:6–7. Prentice R. 1996. Measurement error and results from analytic epidemiology: Dietary fat and breast cancer. J Natl Cancer Inst 88:1738–1747. Pritchard J. 2001. Are rare variants responsible for susceptibility to complex diseases? Am J Hum Genet 69:124–137. Ries L, Eisner M, Kosary C, et al. 2003. SEER Cancer Statistics Review 1975–2000. Bethesda MD, National Cancer Institute. Available at: http://seer.cancer.gov/csr/1975_2000. Rose G. 1985. Sick individuals and sick populations. Int J Epidemiology 14:32–38. Rose G. 1992. The Strategy of Preventive Medicine. New York, Oxford University Press.

Rothenberg R. 1994. Competing mortality and progress against cancer. Epidemiology 5:197–203. Rothman KJ, Adami H-O, Trichopoulos D. 1998. Should the mission of epidemiology include the eradication of poverty? Lancet 352:810–813. Schatzkin A, Lanza E, Corle D, et al. 2000. Lack of effect of a Low-fat, highfiber diet on the recurrence of colorectal adenomas: Polyp Prevention Trial Study Group. N Engl J Med 342:1149–1155. Schwartz S, Diez-Roux A. 2001. Commentary: Causes of incidence and causes of cases—a Durkheimian perspective on Rose. Int J Epidemiol 30: 435–439. Seow A, Duffy S, MA M, Lee J, Lee H. 1996. Breast cancer in Singapore: Trends in incidence 1968–1992. Int J Epidemiol 25:40–45. Susser M, Susser E. 1996. Choosing a future for epidemiology: II. From black box to Chinese boxes and eco-epidemiology. Am J Public Health 86(5):674–677. Tominaga S, Aoki K, Fujimoto I, Kurihara M. 1994. Cancer mortality and morbidity statistics: Japan and the world—1994. Tokyo, Japan Scientific Societies Press. VanFraassen B. 1980. The Scientific Image. Oxford, Clarendon Press. Weed D, Mink P. 2002. Roles and responsibilities of epidemiologists. Ann Epidemiol 12(2):67–72. Whittemore A, McGuire V. 2003. Observational studies and randomized trials of hormone replacement therapy: What can we learn from them? Epidemiology 14:8–10. Willett W, Colditz G, Mueller N. 1996. Strategies for minimizing cancer risk. Scientific American September: 88–95. Willett W, Rockhill B, Hankinson S, Hunter D, Colditz G. 2000. Epidemiology and nongenetic causes of breast cancer. In: Harris J, Lippman M, Morrow M, Osborne C, eds. Diseases of the Breast. Philadelphia, Lippincott Williams and Wilkins: 175–220. Wingo P, Ries L, Rosenberg H, Miller D, Edwards B. 1998. Cancer incidence and mortality, 1973–1995: A report card for the U.S. Cancer 82: 1197–1207. World Health Organization. Prevention of Cancer. Geneva: WHO, 1964 (Technical Report Series 276). Writing Group for the Women’s Health Initiative Investigators. 2002. Risks and benefits of estrogen plus progestin in healthy postmenopausal women. Principal results from the Women’s Health Initiative randomized controlled trial. JAMA 288:321–333. Yach D. 2001. Tobacco control: From concern for the lung to global political action. Thorax 56:247–248. Yach D, Bettcher D. 2000. Globalisation of tobacco industry influence and new global responses. Tob Control 9(2):206–216. Yu H, Harris R, Gao Y, Gao R, Wynder E. 1991. Comparative epidemiology of cancers of the colon, rectum, prostate, and breast in Shanghai: China versus the United States. Int J Epidemiol 20:76–81.

69

Cancer Risk Communication and Comprehension KAREN M. EMMONS, CARA CUITE, AND ERIKA WATERS

A

dvances in science and epidemiology have brought about substantially increased knowledge about factors that influence health outcomes. As science increasingly captures public attention, particularly through the media, the public is inundated with risk information (Brody, 1999; Russell, 1999). A search in a database of articles in the popular press since January, 2000 returned 31,946 articles with the term “health risk” in them (Dow Jones Interactive, 2003). People are faced with a tremendous amount of information about health risks, much of it contradictory and confusing. Because of the proliferation of media outlets that report on health, there is a range of quality in the scientific studies that are presented to the lay public. In addition, there are often contradictory findings in the scientific literature that must be assimilated. Investigations of the relationships between a specific risk factor and a variety of health outcomes can lead to strong associations in some cases and no relationship in others. This is particularly challenging in light of the movement towards active patient involvement in health-care decision making. More and more people are searching out health-care information and are becoming informed consumers of health care. As individuals gather information about particular diseases and/or health risks, there is a great risk of becoming overwhelmed with information that is at best difficult to integrate, and at worst is of poor quality and/or conflicting. For example, consider the following statistics (ACS, Cancer Facts and Figures, 2003): 1. 1 in 8 women will get breast cancer in their lifetime. 2. Breastfeeding for 2 years can cut the risk of breast cancer in half. 3. One alcoholic drink a day can increase the risk of breast cancer by 6%. How should an individual woman who drinks a glass of wine every day and who nursed a child until the age of two interpret these risk estimates? It is difficult for health-care professionals to integrate the large number of epidemiological findings available and to understand their meaning for communicating with their patients about their individual risk factors. For laypeople, the task becomes even more challenging. The goal of this chapter is to present a summary of key issues that must be considered in risk communication. There has been a long-standing effort to improve patient understanding of the risks and benefits of medical procedures and treatments (Leonard et al., 1972; Chase et al., 1986). The basis of this interest is that lay people do not typically understand these complex topics (Fischhoff, 1998; Fischhoff and MacGregor, 1983), are not good at understanding numbers in general, and are particularly bad at making sense of probabilities (Paulos, 1988). However, the challenge of risk communication related to cancer and cancer prevention has become even more complex over the past decade, as more data is available to provide detailed risk estimates to different population subgroups (Colditz et al., 2000). Further, it is increasingly possible to tailor treatment strategies to individual characteristics (Piechocki et al., 2003). Although the range of available treatment choices may ultimately increase survival rates of many cancers, they also stand to increase the complexity of ensuring that patients understand their choices and can most effectively use the available data in their decision-making processes. Bogardus and colleagues (1999) argued: Patients must understand the risks and benefits of the options they face to make informed decisions . . . physicians must be able to

provide suitable, accurate information about risks and benefits in personal, accessible terms to fulfill their essential roles as trusted advisors. (p. 1037) Furthermore, they asserted, probability information is likely “the most difficult element to communicate” to patients. Tamoxifen therapy for breast cancer illustrates why accurate risk comprehension is necessary for patients who want to participate in shared and informed decision making. Although tamoxifen decreases the risk of breast cancer in some women, it also increases the incidence of several side effects (Ganz, 2001). Women must carefully consider the risks and benefits of tamoxifen therapy, and physicians can play an important role in guiding their patients through the decisionmaking process (Bogardus et al., 1999). Effectively communicating risks and benefits of treatment is a significant part of this role. The importance of understanding probabilities is clearly not restricted to breast cancer treatment decisions. Discussions of genetic counseling (Grimes and Snively, 1999), genetic defects (Chase et al., 1986), treatments for prostate cancer (Mazur et al., 1999), screening mammography (Schwartz et al., 1999), and informed consent (Bogardus et al., 1999), also demand that patients understand probability information. Several cancer screening strategies also introduce complexities regarding risk. For example, there are several colorectal cancer screening strategies, ranging from relatively non-invasive to highly invasive. Colonoscopy, the more invasive of the screening strategies, screens more of the colon and thus may have a higher rate of detecting pre-cancerous polyps; however, colonoscopy itself has a higher risk of adverse side effects, and thus this must be considered when making decisions about screening, particularly related to population-wide screening. Risk communication challenges are present across the entire cancer continuum, including cancer prevention, screening, treatment, and palliative care.

WHAT IS RISK? Risk is composed of several dimensions, including the amount of dread produced, its familiarity, its controllability, its catastrophic potential, its distribution of risks and benefits, and its likelihood of occurrence (Slovic, 2000). Although all these dimensions play a role in a person’s understanding of their own risk, probabilistic likelihood information is one of the most basic, objective criteria for risk assessment. Because an individual’s perceptions of numerical probabilities can influence his or her disease or treatment-specific risk perceptions (Chase et al., 1986; Alexander et al., 1995; Yamagishi, 1997), patients must accurately perceive risk probabilities if they want to make informed decisions.

HOW RISKS ARE COMMUNICATED Interestingly, relatively few studies have directly examined how risks are typically communicated in medical settings. Kalet et al. (1994) examined discussions of risk across 160 health-care visits. Fifty-five risk discussions occurred; in 57% of these discussions the health-care provider stated the risk with certainty (e.g., it will/won’t happen), in 40% qualitative expressions of risk were used (e.g., small chance), and

1303

1304

PART V: CANCER PREVENTION AND CONTROL

in less than 5% of discussions quantitative expressions were used. Although there are many problems associated with qualitative expressions of risk, including the fact that not everyone will interpret qualitative descriptors in the same way, there are also a number of issues associated with quantitative descriptors. Numeric expressions of risk are commonly used in public health campaigns and media reports, and it is clear that this information may not be easily understood by the lay public.

RISK PERCEPTION The term “risk perception” refers to a person’s belief about the likelihood of a negative event happening to them. There is substantial evidence that most people have significant misunderstandings of their risks for both common and unusual hazards (Weinstein, 1980; 1982; 1983; 1984; 1987; Becker and Levine, 1987; Avis et al., 1989; Lee, 1989; Kreuter, 1999). In most cases, people underestimate their risk, demonstrating what Weinstein has termed “optimistic bias” (Weinstein, 1980). For example, studies have demonstrated that smokers often estimate their own risk as lower than that of the typical smoker (McCoy et al., 1992). An optimistic bias tends to be most relevant to individual health behaviors. However, for risk of disease outcomes not specifically linked to health behaviors, a “pessimistic bias” has been observed, in which individuals overestimate their risk. This has been particularly true in the case of estimation of one’s risk of breast cancer. For example, a survey of women 40–50 years old demonstrated that participants overestimated their chances of dying of breast cancer within 10 years by 2000% (Black et al., 1995). A clear lack of understanding of mammography as a cancer screening tool was demonstrated by the approximately 600% over-estimation of the relative reduction of risk associated with mammography screening (Black et al., 1995). It is unclear whether risk perception biases are a result of cognitive errors, flawed mental models of information processing, or a reflection of a psychologically protective coping mechanism (Kreuter, 1999). In nearly all psychosocial theories of health behavior, risk perception is a core construct, including subjective expected utility theory, prospect theory, protection motivation theory, and social cognitive theory. According to these theories, the greater the perceived likelihood of a potential outcome, the more the outcome will affect behavioral and emotional responses to the threat. Indeed, implicit in much of the risk communication literature is the assumption is that perceived risk mediates the relationship between risk communication and health behavior, such that if a communication increases the message recipient’s perceived risk, it will motivate him or her to take appropriate action (Aspinwall, 1999). This assumption appeals to the intuition that people only try to prevent or remedy problems that they believe are likely to happen to them. While a large number of studies focus on risk perceptions, the relationship between risk perceptions and behavior is not entirely clear. A recent review (Vernon, 1999) concluded that risk perceptions were positively related to mammography adherence, but not to cervical or colorectal cancer screening. Similarly inconsistent relationships can be found in studies examining the role of risk perceptions in other health behaviors (Dolcini et al., 1996; Gerrard et al., 1996; McCaul et al., 1996). It has been suggested that this inconsistency may be due to the use of inappropriate research designs and incorrect statistical analyses to test the relationship (Weinstein and Nicolich, 1993). There are many examples of studies in which risk perceptions have been influenced by interventions that provide pertinent information. For example, women who received BRCA1/2 genetic counseling for breast cancer were later more likely to accurately state their actual lifetime risk than were women who did not receive counseling (Evans et al., 1994). Other studies have found similar trends in breast cancer risk perception among women not participating in genetic counseling (Alexander et al., 1995; Croyle and Lerman, 1999). However, there is at least some evidence that improved risk perceptions may erode over time. Lipkus et al. (2001a) found that women’s perceived risk of breast cancer returned to baseline levels 6 months after being told accurate

information. Fischhoff (1995) cautions that simply informing a person of his or her actual risk in terms of probability is only a small part of modifying risk perception and may not alter behavior. Supporting his assertion is an intervention study that combined counseling and education for women who were considering the BRCA1/2 genetic test (Lerman et al., 1997). The treatment increased the participants’ awareness of the test’s limitations, risks, and benefits, but did not change their testing intentions. Likewise, informing women of their actual probability of developing breast cancer did not change their intentions to have a mammogram (Lipkus et al., 2001a). It has also been argued that people’s actions seldom correspond to probabilistic risk information (Weinstein, 1999; Rothman and Kiviniemi, 1999). Weinstein has argued that eliciting numerical probabilities from people is one of the least meaningful and least reliable measures of complete risk understanding. However, this does not necessarily mean that supplying risk information is inappropriate or unnecessary to informed decision making (Bogardus et al., 1999). It is unclear whether instances in which risk information fails to influence actions indicate a limited impact of cognition (as opposed to habit or affect) on hazard behavior, unsuccessful communication, or some other issue.

RISK COMPREHENSION Weinstein (1999) points out the importance of acknowledging the difference between risk communication, or the messages that are intended to help people understand the hazards they face, and risk comprehension, or people’s understanding of these risk messages. He has identified several dimensions of risk comprehension that are critical for individuals to make effective decisions about personal risks, including: an understanding of the nature of the potential harm, such as reasonably detailed knowledge of the undesirable consequences of a risk behavior; an understanding of the probability of harm; knowledge about the factors that influence individual susceptibility; and understanding of the difficulty of avoiding the harmful consequences of the hazard. Weinstein provides helpful detail on each of these areas, and concludes that assessment of risk comprehension requires evaluation of a constellation of beliefs that are relevant to decisions and behaviors concerning that risk. He notes that most current studies fail to examine even this limited range of risk dimensions using adequate methodologies.

ISSUES INFLUENCING RISK COMMUNICATION AND RISK COMPREHENSION This section will address individual and communication factors that can affect people’s understanding of risk, including numeracy, numerical formats, qualitative expressions of risk, comparative risk, and visual formats.

Numeracy Numeracy has been defined as a person’s ability to compute basic probability and mathematical concepts (Lipkus et al., 2001b). Innumeracy, the inability to perform these operations, offers one explanation for the notorious difficulty in communicating numerical risks to patients. Many risk communicators and physicians assume that their intended audience has a rudimentary understanding of probabilities, percentages, and proportions (Alexander et al., 1995; Croyle and Lerman, 1999; Man-Son-Hing et al., 2000). This assumption can lead to communication techniques that employ concepts and techniques the public may not fully comprehend. Research has shown that a large portion of the population is low in numeracy and cannot perform the basic mathematical operations relevant to an adequate understanding of health risks (Lipkus and Samsa et al., 2001c). People have been shown to have difficulty understanding information about probabilities in general (e.g., Tversky and Kahneman 1974; Paulos, 1988; Gigerenzer and Hoffrage, 1995) and

1305

Cancer Risk Communication and Comprehension about health risk probabilities in particular (Gigerenzer, 1996; Fischhoff, 1998). Some have even suggested training for the general public to help them understand this type of information (Schwartz and Woloshin, 1999). Similarly, there are myriad articles and reports instructing communicators on how to explain risks (e.g., Fischhoff, 1989; Kasperson and Stallen, 1991; Atman et al., 1994; Wright, 1994). Because so much of the empirical literature indicates that people struggle with understanding risks, it is not surprising that much of the theoretical literature focuses on biases in understanding (e.g., Tversky and Kahneman, 1974; Goldstein and Hogarth, 1996). A patient’s level of numeracy has a significant impact on how he or she estimates different risks. For instance, women scoring higher in numeracy were more likely to accurately state the benefits of screening mammography (Schwartz et al., 1997), were less likely to overestimate their chances of dying from breast cancer, and were less likely to overestimate benefits of breast cancer screening (Black et al., 1995).

Numerical Formats Gigerenzer (1996; Gigerenzer and Hoffrage, 1995) has argued that characterizing people as “bad at probabilities” is too simplistic and in some cases inaccurate. He states that it is important that risk communications require use of numbers in a manner that reflects how numbers have been used throughout human history, and in particular he stresses the importance of format in risk communication, and the important interaction between risk format (e.g., percent, n in 100) and type of mathematical operation that is required to interpret risk (e.g., adding two risk, cutting a risk in half) (Gigerenzer, 1996). The impact of different formats for presenting risk information has not been well studied. However, the few studies that present participants with probabilistic information using different numerical formats further illustrate misunderstanding of probabilities. A study conducted by the National Cancer Institute showed that focus group members’ interpretations of risks were inconsistent across risk presentation formats (Office of Cancer Communications [OC], 1998). For example, some participants viewed a risk that increased from “1 in 30 to 1 in 10” as more alarming than a risk they were “3 times as likely to get” (Office of Cancer Communications, 1998). Although many studies have addressed preferences for information formats (e.g., Woloshin et al., 2000) and how numerical formats affect perceptions of risk (e.g., Denes-Raj et al., 1995; Halpern and Trefethen, 1988), only three studies to date have addressed comprehension of different numerical formats for presenting risk likelihood. Grimes and Snively (1999) asked women to identify the larger of two risk estimates, with the same risk scenario posed to each woman in two different ways (e.g., using an “n in 100” format and a “1 in n” format). Regardless of age, primary language, or education, women correctly identified the larger number more often with the “n in 100” format than they did with the “1 in n” format (73% and 56%, respectively). Twenty percent of the respondents missed both questions. In addition to the effect of format, there was a significant effect of education, such that women with low levels of education did worse overall, which is likely related to numeracy. Two other studies compared “n in 100” with “%” formats, with conflicting results. The effect of format on willingness to pay for a hypothetical medication was examined by Siegrist (1997). Although comprehension was not directly measured, the design permits inferences about participants’ understanding of risk. Deviations from increased willingness to pay for protection against a higher likelihood of death indicated a lack of understanding. Two levels of risk were used, a very low likelihood (.000003) and a relatively high likelihood (.0003), as were two formats, “n in 100” and “%.” When the “n in 100” format was used, they found the normative results they expected: participants were willing to pay significantly more for the medication in the higher likelihood of death condition than in the lower likelihood condition. However, with the “%” format, there was no significant difference in willingness to pay between high-risk and low-risk conditions. This result indicates that the respondents did not perceive any difference between the high-risk and low-risk likelihoods when they were presented in the “%” format, but they did recognize the differ-

ence when the information was presented in the “n in 100” format. Thus, although no direct comprehension measures were included, it appears that the “n in 100” format was better understood than the “%” format. Lipkus et al. (2001b) compared “%” and a combined “n in 100” and “1 in n” format. The authors examined performance on two different scales measuring numeracy among a highly educated sample. On a health-specific numeracy scale, they asked participants to double “1%” and “1 in 100,” and also asked them to identify the highest risk among each of two sets of risks {1 in 100, 1 in 1000 and 1 in 10} and {1%, 10%, and 5%}. For both the comparison and doubling operations, people performed better on the “%” than the “1 in n” format (comparison question, 83.6% vs. 78.2%; doubling question, 90.5% vs. 86.6%). Because the comparison of accuracy rates across the two formats was not the goal of the study, there were no statistical tests to indicate if these accuracy rates were significantly different. It is important to note that the “1 in 100” format can be viewed as both an example of “n in 100” and “1 in n,” by virtue of the 1 in the numerator and the multiple of 10 in the denominator. The findings from these three studies are illustrated in Table 69–1. It is difficult to draw conclusions about the comprehensibility of the various formats for presenting numerical risk likelihood information. One reason is the contradictory findings of the existing research. More importantly, no single study has yet compared the three commonly used numerical formats.

Qualitative Expressions of Risk Since both laypersons and physicians have been shown to be consistently poor at working with quantitative expressions of likelihood (Kalet et al., 1994), it is not surprising that many investigations have examined more qualitative risk formats using verbal (non-numeric) expressions of risk (e.g., Budescu and Wallsten, 1995). However, verbal expressions of risk pose problems that do not exist with numeric expressions. Perhaps the most frequently cited problem with qualitative formats for presenting risk is the lack of consistency in their interpretation. One study found that when physicians and laypersons were asked to assign numerical risk estimates (in the form of percentages) to verbal ratings, such as “unusual,” “occasionally,” and “typical,” the responses demonstrated extreme interpersonal variability in the interpretation of these terms (Nakao and Axelrod, 1983). These researchers recommended that verbal expressions of risk be avoided, and that reported frequency of events should be utilized. Other studies have compared verbal categories with “%” and “n in 100” scales in terms of the accuracy of perceived risk for a number of health problems (Diefenbach et al., 1993; Weinstein and Diefenbach, 1997). No differences were found between the scales in terms of their correlations with actual risk factors. Another study found no difference between the efficacy of verbal vs. numerical categories in helping people place gambles (Erev and Cohen, 1990). In sum, although plagued with problematic interpersonal variability, there is some evidence that qualitative risk estimates may be as informative as quantitative risk expressions.

Comparative Risk Comparative risk can mean either comparing an individual’s risk to that of the general population or a peer, or comparing an individual’s risk for a particular health problem with his or her risk for a different Table 69–1. Overview of Existing Studies of Understandability of Formats for Risk Likelihoods Author (Year)

Formats Tested

Grimes and Snively (1999) “n in 100” vs. “1 in n” Siegrist (1997) “%” vs. “n in 100” Lipkus et al. (2001b) “%” vs. fractiona

Best-Performing Format “n in 100” “n in 100” “%”

Note. The fraction format is representative of both “n in 100” and “1 in n.”

1306

PART V: CANCER PREVENTION AND CONTROL

health problem. There is some evidence that people may be more accurate when thinking in comparative as opposed to absolute terms. Woloshin and colleagues (1999) have demonstrated that women overestimate their risk of breast cancer when using a numerical scale (98% overestimated using an “n in 100” scale). However, women were much less likely to overestimate their comparative risk. While this is a promising way of helping people to fully understand their risk (Weinstein, 1999), this is not always feasible in clinical settings, as a person’s absolute risk for a single health problem may be all the information available to a risk communicator at any given moment. Kreuter (1999) identified several factors that should be carefully considered in presentation of comparative risk information. Perceived characteristics of a risk must be considered; substantial research evidence has demonstrated that risks are perceived as more acceptable when they are seen as more controllable, less likely to have immediate consequences, less likely to lead to catastrophic outcomes, more fairly distributed in the population, and more familiar. The nature of risk exposure is also important, in that risks that are associated with voluntary actions are viewed as more acceptable. Thus, in comparing two risks, an individual may be willing to accept a higher level of risk from a health behavior that they choose to engage in than from a medical procedure that may be perceived as imposed. Risk probabilities that are very small may be more difficult to understand, and may be viewed with skepticism. Finally, the epidemiologic metric (e.g., relative vs. absolute risk) used to present risk can lead to difficulties in understanding two risks, in that individuals tend to overestimate risk probabilities when presented in relative terms, compared with those presented as absolute risks. For example, a 50% increase in risk is perceived to be larger (relative risk) than an increase in risk from 2%–3% (absolute risk), regardless of the base level of risk (Malenka and Baron, 1993).

Visual Formats Risk information can also presented in visual formats, such as pie or bar charts, or histograms. However, there is relatively little information available about what graphical format is best for patients making medical treatment decisions (Feldman-Stewart et al., 2000). In fact, little experimental research has examined the effects of visual displays on the general lay public’s understanding of risk. However, the literature that is available suggests that, because visual displays are less abstract than text-only information, they make risk information easier for the public to digest. (MacDonald-Ross, 1977; Edwards et al., 1999; Feldman-Stewart et al., 2000). MacDonald-Ross (1977) posits that numbers are extremely abstract because a person cannot differentiate among them “without an act of cognition”, but bar graphs allow people to use less demanding perceptual mechanisms to understand data. For example, Feldman-Stewart et al. (2000) asked 36 students to indicate the larger of two quantities shown as numbers or as one of several visual displays. Participants made more accurate choices when the information was presented as vertical bars than as numbers, pie charts, horizontal bars, or ovals. If displays can affect people making the most basic distinctions, they should also influence risk perceptions and provide assistance with more complex tasks (e.g., treatment decisions). Visual displays called risk ladders have been shown to influence the ways in which people perceive risk (Weinstein et al., 1991). A review of several studies conducted over 5 years indicated that the location of a risk on a risk ladder significantly influences participant’s perceived threat (Sandman et al., 1994). A risk placed one-quarter of the way up the ladder evoked a much lower perceived threat than the same risk placed three-quarters of the way up a second ladder that covered a different range of risk probabilities (Sandman et al., 1998). Another study found that participants who were asked to compare relative magnitude information (e.g., which is larger: A+B or C+D?) performed better when they viewed bar graphs than when they viewed the information as a table (Spence and Lweandowsky, 1991). This greater understanding may result from the ability of displays to hold people’s attention, making them more likely to attend to the information than if the displays were absent (Lipkus and Hollands, 1999).

Another reason graphs may be helpful in communicating risk information is because they can provide people with both the relative and the absolute risks of a hazard (Edwards et al., 1999). For instance, a bar graph might depict the lifetime risk of heart disease with and without cholesterol-lowering treatment by having a blue bar represent the probability of illness without treatment and a red bar represent illness probability with treatment. Both bars illustrate the absolute risk of disease in each condition, but the difference in height between the bars illustrates the risk of heart disease with treatment relative to the risk without treatment. A visual display thus allows communication between patient and practitioner to occur with less technical language, than would be required without such and aid. It should be noted that visual formats are seldom used by physicians in explaining risks to their patients, most likely because it is difficult to generate individualized graphics during the course of a consultation. This may change as technology improves (Rimer and Glassman, 1999). However, research examining the impact of different visual formats on risk comprehension is quite limited, and further studies are needed to determine which visual aids maximally enhance understanding of disease risk and facilitate decision making (Lipkus and Hollands, 1999).

Competing Risks Ever since groundbreaking work by Simon (1957), an accepted principle of behavioral decision theory has been that people adopt simplifying strategies when faced with too much information or information that is difficult to understand. This may explain why women tend to greatly overestimate their risk of breast cancer, while underestimating the risk of heart disease and lung cancer. For example, the risk of breast cancer increases with aging, and statistics reflecting lifetime risk are often used in public health campaigns to increase screening behaviors. However, many women assume that these statistics apply to them as an individual, regardless of age, and thus greatly overestimate their risk (Morris et al., 2001). Choosing among complex cancer treatment or adjuvant options exemplifies this situation. Rothbert and colleagues asked women to simulate a decision about hormone replacement therapy and found, as expected, that many ignored the heart disease, cancer, and osteoporosis risk information they received, basing their decision upon only the risk of developing hot flashes (Rothbert et al., 1990). The mechanism by which women process these competing risks and the ways in which they interpret probabilities of competing risks is not well understood, and study of these issues is clearly needed.

DECISION AIDS There is a growing research literature on “decision aids” as a risk communication and decision-making tool (O’Connor et al., 1999; Whelan et al., 2002; O’Connor et al., 2003). Decision aids typically provide information about the nature of the outcomes that could accompany available choices, the probability of these outcomes, and the costs of alternatives. They elicit and/or attempt to clarify patients’ values, and help the patients give weightings to potential outcomes. Decision aids frequently provide structured mechanisms for combining the preceding information to arrive at an overall rating of each alternative. Decision aids have taken many forms, including printed materials, interactive computer programs, digital video disks, videotapes, counseling, decision boards, and combinations of these media. Generally, the decision aid is compared with usual care or with a simpler decision aid. The focus has been on what impact can be produced, not on how these impacts are achieved or what elements are responsible. The use of complex, multiple-component treatments in these studies and comparisons of interventions that differ both in content and in communication channel make it difficult to ascertain the role of probability information in the decisions reached or to study how useful probability information might be if it were provided via optimal formats. A comprehensive review of cancer-related decision aids suggests that future research should determine patients’ understand-

Cancer Risk Communication and Comprehension ing of numerical risk estimates, and whether such numbers are meaningful for them in terms of real-life decision making (Whelan et al., 2002). The decision aids literature shows that such aids can increase general knowledge of facts relevant to the decision, increase knowledge of probabilities, and reduce decisional conflict. Measures of satisfaction related to either the decision process or the decision itself have not proved sensitive to differences among interventions. Treatment or screening choices are sometimes affected, but the aids are mostly used in situations in which no choice is clearly superior to the others, so is not possible to conclude that aids improve decision making. For example, Edwards et al. (2003) report that communications that included numerical calculations of personal risk led to less screening than communications that only listed personal risk factors. One possible interpretation of this finding is that the numerical information was misunderstood or was not motivating. Equally plausible is that the numerical information showed participants that the benefits of screening were smaller than the overly optimistic expectations the participants had brought to the study. This latter interpretation is consistent with evidence that decision aids tend to decrease prostate screening (Barry, 2002).

OTHER FACTORS A significant amount of research has focused on many other variables that affect comprehension of risk messages and subsequent behaviors. Intra-personal factors, such as optimism (Aspinwall and Brunhart, 1996), anxiety (Butler and Mathews, 1987; Mogg et al., 1990; Martin et al., 1991; Mathews, 1993), and mood (Salovey and Birnbaum 1989; Mayer et al., 1990) have been shown to play a role in risk perceptions, risk comprehension, and attention to information about risk. Communication variables such as the framing of risk messages (Rothman et al., 1993; Detweiler et al., 1999) and personalization of messages (Skinner et al., 1994; Lerman et al., 1995; Rakowski et al., 1998) have also been shown to influence message comprehension, risk perceptions, and resultant health behaviors. These are just some of many additional factors that have been shown to influence message acceptance and behavior that are beyond the scope of this chapter.

CONCLUSIONS Substantial research has been conducted in the area of risk communication. However, although much research has focused on different risk formats, the comprehensibility of various formats has largely been overlooked. Although many studies have examined the way in which messages are framed, such studies show us how to present risks in ways that increase concern or that reassure, but do not inform our understanding of how to increase risk comprehension. In fact, the existence of these format effects can be viewed as a demonstration of the difficulties of communicating risk magnitudes: if different presentations of the same probabilities lead to different decisions, people clearly fail to fully understand the risks. There are two primary sources of difficulty in such communication. First is the challenge posed by the intended audience for risk messages, in terms of their limited understanding of probabilities, statistics, and the scientific process, as well as the cognitive biases they bring to the risk communication situation. Second is the challenge posed by risk communicators, who often do not appreciate the audience’s limited understanding of key factors involved in risk communication, and who similarly do hot fully understand their audience. Fischoff (1999) recommends that risk communication planning should follow the same conversational norms as used when informing children, patients, or students about risk. He recommends first figuring out which facts have the greatest value to the audience, and then relaying those facts clearly. A key step is to understand the extent to which this requires added context. Fischoff recommends a formal set of procedures for identifying and comparing expert views of risk (“expert model”) and intended audience conceptions (“mental model”). An iter-

1307

ative research process is then undertaken to determine which elements of the expert model most effectively improve the audience’s understanding of the risk. This approach uses analytic methods for selecting the information to transmit, based on how critical it is to the audiences’ decision making, and has considerable potential to improve the practice and evaluation of risk communication (Maibach, 1999). When risk communication is used for the purposes of improving decision making, key outcomes are whether or not the communication provides an accurate understanding of the risk, particularly related to whether the information is worth knowing; facilitation of a good decision-making process; and provision of confidence that a good decision has been made (Maibach, 1999). Applied research on cancer risk communication and comprehension is quite limited, and even less is known about communication regarding multiple risks. Cancer risk communication is a difficult, yet very important undertaking. Physicians implicitly assume patient comprehension of risk information. This assumption, which is present regardless of how information is presented and in what format, is often incorrect. Although many characteristics of the communication influence the patient’s decision making, he or she should at least have correct, objective knowledge. Continued research focused on maximizing comprehension of risk information, considering the many factors discussed here, is important if health-care providers are to help their patients to make fully informed decisions, and if our growing epidemiological data is to help individuals manage their own health and prevention of disease. References Alexander NE, Ross J, Sumner W, et al. 1995. The effect of an educational intervention on the perceived risk of breast cancer. Journal of General Internal Medicine 11:92–97. Am J Public Health 79(12):1608–1612. American Cancer Society. 2003. Cancer Facts and Figures. Aspinwall LG. 1999. Introduction of section: persuasion for the purpose of cancer risk reduction: Understanding responses to risk communications. Review in J Natl Cancer Inst Monographs (25):88–93. Aspinwall LG, Brunhart SM. 1996. Distinguishing optimism from denial: Optimistic beliefs predict attention to health threats. Pers Soc Psyc Bull 22:993–1003. Atman CC, Bostrom A, Fischhoff B, et al. 1994. Designing risk communications completing and correcting mental models of hazardous processes, Part I. Risk Anal 14(5):779–788. Avis NE, Smith KW, McKinlay JB. 1989. Accuracy of perceptions of heart attack risk: What influences perceptions and can they be changed? Am J Public Health 79:1608–1612. Barry MJ. 2002. Health decision aids to facilitate shared decision making in office practice. Ann Intern Med 136(2):127–135. Becker DM, Levine DM. 1987. Risk perception, knowledge, and lifestyles in siblings of people with premature coronary disease. Am J Prev Med 3(1):45–50. Black WC, Nease RF Jr., Tosteson AN. 1995. Perception of risk and screening effectivencess in women younger than 50 years of age. J Natl Cancer Inst 87:720–731. Bogardus ST, Holmboe E, Jekel JF. 1999. Perils, pitfalls, and possibilities in talking about medical risk. JAMA 281(11):1037–1041. Brody JE. 1999. Communicating cancer risk in print journalism. J Natl Cancer Inst, Monographs (25):170–172. Budescu DV, Wallsten TS. 1995. Review of “Communicating Quantities: A Psychological Perspective” by Linda M. Moxey and Anthony J. Sandford. Chance 8:38–40. Butler G, Mathews A. 1987. Anticipatory anxiety and risk perception. Cognitive Ther Res 91:551–565. Chase GA, Faden RR, Holtzman NA, et al. 1986. Assessment of risk by pregnant women: Implications for genetic counseling and education. Soc Biol 33(1–2):57–64. Colditz GA, Atwood KA, Emmons K, et al. 2000. Harvard report on cancer prevention, Volume 4: Harvard Cancer Risk Index. Cancer Causes Control 11(6):477–488. Connelly NA, Knuth BA. 1998. Evaluating risk communication: Examining target audience perceptions about four presentation formats for fish consumption health advisory information. Risk Anal 18(5):649–659. Croyle RT, Lerman C. 1999. Risk communication in genetic testing for cancer susceptibility. J Natl Cancer Inst, Monographs 25:59–66. Denes-Raj V, Epstein S, Cole J. 1995. The generality of the ratio-bias phenomenon. Pers Soc Psych Bull 21(10):1083–1092.

1308

PART V: CANCER PREVENTION AND CONTROL

Detweiler JB, Bedell BT, Salovey P, et al. 1999. Message framing and sunscreen use: Gain-framed messages motivate beach-goers. Health Psychol 18(2):189–196. Diefenbach MA, Weinstein ND, O’Reilly J. 1993. Scales for assessing perceptions of health hazard susceptibility. Health Educ Res 8(2):181–192. Dolcini MM, Catania JA, Choi KH, et al. 1996. Cognitive and emotional assessments of perceived risk for HIV among unmarried heterosexuals. AIDS Educ Prev 8(4):294–307. Edwards A, Elwyn G, Gwyn R. 1999. General practice registrar responses to the use of different risk communication tools in simulated consultations: A focus group study. BMJ 319(7212):749–752. Edwards A, Unigwe S, Elwyn G, et al. 2003. Personalized risk communication for informed decision making about screening programs (Cochrane Review). The Cochrane Reviews 1. Oxford: Update Software. Erev I, Cohen BL. 1990. Verbal versus numerical probabilities: Efficiency, biases, and the preference paradox. Organizational Behavior and Human Decision Processes 45:1–18. Evans DGR, Blair V, Greenhalgh R, et al. 1994. The impact of genetic counseling on risk perception in women with a family history of breast cancer. J Gen Intern Med 11:92–97. Feldman-Stewart D, Kocovski N, McConnell BA, et al. 2000. Perception of quantitative information for treatment decisions. Medical Decis Making 20(2):228–238. Fischhoff B. 1999. Why (cancer) communication can be hard. J Natl Cancer Inst, Monographs (25):7–13. Fischhoff B. 1998. Communicate unto others . . . Reliability Engineering and System Safety 59:63–72. Fischhoff B. 1995. Risk perception and communication unplugged: Twenty years of process. Risk Anal 15(2):137–145. Fischhoff B. 1989. Risk: A guide to controversy. Apendix to National Research Council. Improving risk communications. Washington, D.C.: National Academy Press, pp. 211–319. Fischhoff B, MacGregor D. 1983. Judged lethality: How much people seem to know depends upon how they are asked. Risk Anal 3(4):229–235. Ganz PA. 2001. Impact of Tamoxifen adjuvant therapy on symptoms, functioning, and quality of life. J Natl Cancer Inst, Monographs 30:130–134. Gerrard M, Gibbons FX, Benthin AC, et al. 1996. A longitudinal study of the reciprocal nature of risk behaviors and cognitions in adolescents: What you do shapes what you think, and vice versa. Health Psychol 15(5):344–354. Gigerenzer G. 1996. The psychology of good judgment: Frequency formats and simple algorithms. Review in Med Decis Making 16(3):273–280. Gigerenzer G, Hoffrage U. 1995. How to improve Bayesian reasoning without instruction: Frequency formats. Psychol Rev 102(4):684–704. Goldstein WM, Hogarth RM. 1996. Research on Judgment and Decision Making-Currents, Connections and Controversies. Cambridge University Press, Call No.: BF448.J84. 95–143. Grimes DA, Snively GR. 1999. Patients’ understanding of medical risks: Implications for genetic counseling. Obstet Gynecol 93:910–914. Halpern L, Trefethen LN. 1988. Wide-angle one-way wave equations. Journal Acoust Soc Am 84(4):1397–1404. Kalet A, Roberts JC, Fletcher R. 1994. How do physicians talk with their patients about risks? J Gen Intern Med 9:402–404. Kasperson RE, Stallen PJ, eds. 1991. Communicating risks to the public: international perspectives. Technology, Risk, and Society vol. In: 4 Boston, MA: Kluwer Academic Publishers, p. 481. Kreuter MW. 1999. Dealing with competing and conflicting risks in cancer communication. J Natl Cancer Inst, Monographs (25):27–35. Lee C. 1989. Perceptions of immunity to disease in adult smokers. J Behav Med 12(3):267–277. Leonard CO, Chase GA, Childs B. 1972. Genetic counseling: A consumer’s view. Genet Couns 287(9):433–439. Lerman C, Biesecker B, Benkendorf JL, et al. 1997. Controlled trial of pretest education approaches to enhance informed decision-making for BRCA1 gene testing. J Natl Cancer Inst 89(2):148–157. Lerman C, Lustbader E, Rimer B, et al. 1995. Effects of individualized breast cancer risk counseling: A randomized trial. J Natl Cancer Inst 87(4):286–291. Lipkus IM, Biradavolu M, Fenn K. 2001a. Informing women about their breast cancer risks: Truth and consequences. Health Commun 13(2):205–226. Lipkus IM, Hollands JG. 1999. The visual communication of risk. J Natl Cancer Inst, Monographs (25):149–163. Lipkus IM, Klein WM, Rimer BK. 2001b. Communicating breast cancer risks to women using different formats. [Evaluation Studies. Journal Article] Cancer Epidemiology, Biomarkers Prevention 10(8):895–898. Lipkus IM, Samsa G, Rimer BK. 2001c. General performance on a numeracy scale among highly educated samples. Med Decis Making 21(1): 37–44.

MacDonald-Ross M. 1977. How numbers are shown: A review of research on the presentation of quantitative data in texts. Audio-visual Communication Review 25:359–407. Maibach E. 1999. Cancer risk communication—what we need to learn. J Natl Cancer Inst, Monographs (25):179–181. Malenka D, Baron J. 1993. The framing effect of relative and absolute risk. J Gen Intern Med 8:543–548. Man-Son-Hing M, Laupacis A, O’Connor AM, et al. 2000. Patient preferencebased treatment thresholds and recommendations: A comparison of decision-analytic modeling with the probability-tradeoff technique. Med Decis Making 20(4):394–403. Martin M, Williams RM, Clark DM. 1991. Does anxiety lead to selective processing of threat-related information? Behav Res Ther 29(2):147–160. Mathews A. 1993. Anxiety and the processing of emotional information. Prog Exp Pers Psychopathol Res 16:254–280. Mayer JD, DiPaolo M, Salovey P. 1990. Perceiving affective content in ambiguous visual stimuli: A component of emotional intelligence. J Pers Assess 54(3–4):772–781. Mazur DJ, Hickam DM, Mazur MD. 1999. How patients’ preferences for risk information influence treatment choice in a case of high risk and high therapeutic uncertainty: Asymptomatic localized prostate cancer. Med Decis Making 19:394–398. McCaul KD, Schroeder DM, Reid PA. 1996. Breast cancer worry and screening; Some prospective data. Health Psychol 15:430–433. McCoy SB, Gibbons FX, Reis TJ, et al. 1992. Perceptions of smoking risk as a function of smoking status. J Behav Med 15(5):469–488. Mogg K, Mathews A, Bird C, MacGregor-Morris R. 1990. Effects of stress and anxiety on the processing of threat stimuli. J Pers Soc Psychol 59(6):1230–1237. Morris CR, Wright WE, Schlag RD. 2001. The risk of developing breast cancer within the next 5, 10, or 20 years of a woman’s life. Am J Prev Med 20(3):214–218. Nakao MA, Axelrod S. 1983. Numbers are better than words. Verbal specifications of frequency have no place in medicine. Am J Med 74(6): 1061–1065. Office of Cancer Communications. 1998. How the public perceives, processes, and interprets risk information: Findings from focus group research with the general public. Bethesda, MD: National Cancer Institute. O’Connor AM, Drake ER, Fiset V, et al. 1999. The Ottawa patient decision aids. Eff Clin Pract 2(4):163–170. O’Connor AM, Rostrom A, Fiset V, et al. 2003. Decision aids for patients facing health treatment or screening decisions (Cochrane Review). In The Cochrane Library, 1. Oxford: Update Software. O’Connor AM, Rostrom A, Fiset V, et al. 1991. Decision aids for patients facing health treatment or screening decisions: Systematic review. British Medical Journal 319:731–734. Paulos JA. 1988. Innumeracy: Mathematical Illiteracy and Its Consequences. New York, Hill and Wang. Piechocki MP, Ho YS, Pilon S, et al. 2003. Human ErbB-2 (Her-2) transgenic mice: A model system for testing Her-2 based vaccines. J Immunol 171(11):5787–5794. Rakowski W, Ehrich B, Goldstein MG, et al. 1998. Increasing mammography among women aged 40–74 by use of a stage-matched, tailored intervention. Prev Med 27(5 Pt 1):748–756. Rimer BK, Glassman B. 1999. Is there a use for tailored print communications in cancer risk communication? Review in J Natl Cancer Inst, Monographs (25):140–148. Rothbert M, Rouner D, Holmes M, et al. 1990. Women’s use of information regarding hormone replacement therapy. Res Nurs Health 13:355–366. Rothman AJ, Kiviniemi MT. 1999. Treating people with information: An analysis and review of approaches to communicating health risk information. J Natl Cancer Inst, Monographs 25:44–51. Rothman AJ, Salovey P, Turvey C, et al. 1993. Attributions of responsibility and persuasion: Increasing mammography utilization among women over 40 with an internally oriented message. Health Psychol 12(1):39– 47. Russell C. 1999. Living can be hazardous to your health: How the news nedia cover cancer risks. J Natl Cancer Inst, Monographs 177–178. Salovey P, Birnbaum D. 1989. Influence of mood on health-relevant cognitions. J Pers Soc Psychol 57(3):539–551. Sandman PM, Weinstein ND, Hallman WK. 1998. Communications to reduce risk underestimation and overestimation. Risk Decision and Policy 3(2):93–108. Sandman PM, Weinstein ND, Miller P. 1994. High risk or low: How location on a “risk ladder” affects perceived risk. Risk Anal 14(1):35–45. Schwartz MD, Rimer BK, Daly M, et al. 1999. A randomized trial of breast cancer risk counseling: the impact on self-reported mammography use. Am J Public Health 89(6):924–926.

Cancer Risk Communication and Comprehension Schwartz LM, Woloshin S. 1999. Risk communication in clinical practice: Putting cancer in context. Review in J Natl Cancer Inst, Monographs (25):124–133. Schwartz LM, Woloshin S, Black W, et al. 1997. The role of numeracy in understanding the benefit of screening mammography. Ann Intern Med 127(11):966–972. Siegrist M. 1997. Communicating low risk magnitudes: Incidence rates expressed as frequency versus rates expressed as probability. Risk Anal 17:507–510. Simon H. 1957. Models of Man. New York, Wiley. Skinner CS, Strecher VJ, Hospers H. 1994. Physicians’ recommendations for mammography: Do tailored messages make a difference? Am J Public Health 84(1):43–49. Slovic P. 2000. Perception of risk. In: Slovic P, ed. The perception of risk. London: Earthscan Publications, Ltd., pp. 220–231. Slovic P, Monahan J, MacGregor DG. 2000. Violence risk assessment and risk communication: The effects of using actual cases, providing instruction, and employing probability versus frequency formats. Law and Human Behavior 24(3):271–296. Spence I, Lweandowsky S. 1991. Displaying proportions and percentages. Appl Cogn Psychol 5(61–77):61–77. Stone ER, Yates JF, Parker AM. 1997. Effects of numerical and graphical displays on professed risk-taking behavior. J Exp Psychol: Applied 3(4):243–256. Tversky A, Kahneman D. 1974. Judgment under uncertainty: Heuristics and biases. Science 185(4157):1124–1131. Vernon SW. 1999. Risk perception and risk communication for cancer screening behaviors: A review. J Natl Cancer Inst, Monographs (25):101–119. Weinstein ND. 1999. What does it mean to understand a risk? Evaluating risk comprehension. J Natl Cancer Inst, Monographs 25(25):15–20. Weinstein ND. 1987. Unrealistic optimism about susceptibility to health problems: Conclusions from a community-wide sample. J Behav Med 10(5):481–500.

1309

Weinstein ND. 1984. Why it won’t happen to me: Perceptions of risk factors susceptibility. Health Psychol 3:431–457. Weinstein ND. 1983. Reducing unrealistic optimism about illness susceptibility. Health Psychol 2:11–20. Weinstein ND. 1982. Unrealistic optimism about susceptibility to health problems. J Behav Med 5(4):441–460. Weinstein ND. 1980. Unrealistic optimism about future life events. J Pers Soc Psychol 39:806–820. Weinstein ND, Diefenbach MA. 1997. Percentage and verbal-category scales for assessing personal risk likelihood. Health Educ Res 12:139–141. Weinstein ND, Nicolich M. 1993. Correct and incorrect interpretations of correlations between risk perceptions and risk behaviors. Health Psychol 12(3):235–245. Weinstein ND, Sandman PM, Miller P. 1991. Communicating effectively about risk magnitudes, phase two. New Brunswick, NJ: Environmental Communication Research Program, Cook College, Rutgers University. Whelan TJ, et al. 2002. Impact of cancer-related decision aids. Evidence report/technology assessment number 46. Agency for Healthcare Research and Quality. US Department of Health and Human Services, Rockville, MD. (AHRQ Publication No 02-E004). Windschitl PD. 2002. Judging the accuracy of a likelihood judgment: The case of smoking risk. J Behav Decis Making 15(1):19–35. Woloshin S, Schwarts LM, Black WC, et al. 1999. Women’s perceptions of breast cancer risk: How you ask matters. Med Decis Making 19(3): 221–229. Woloshin S, Schwartz LM, Byram S, et al. 2000. A new scale for assessing perceptions of chance: A validation study. Med Decis Making 20(3):298–307. Wright R. 1994. Risk communication: How to answer tough patient questions. J American Dental Assoc 125(8):1131–1137. Yamagishi K. 1997. When a 12.86% mortality is more dangerous than 24.14%: Implications for risk communication. Appl Cogn Psychol 11:495–506.

70

Principles of Screening BERNARD LEVIN AND PHILIP C. PROROK

O

vercoming cancer should theoretically be possible either by preventing the disease from ever occurring or by effective therapy. Screening for cancer provides an alternative approach to controlling cancer. Screening for cancer is defined as the testing of apparently healthy volunteers from the general population for the purpose of separating them into high and low probabilities of having a given cancer (Prorok et al., 1999). Screening tests are not diagnostic and an abnormal finding requires appropriate confirmation and follow-up.

THE VALIDITY OF A SCREENING TEST Sensitivity, specificity, and positive predictive value are three measures used to define the performance of a screening test. Sensitivity is defined as the ability of a test to detect those with the disease in the screened population. This is expressed as the proportion of those with the disease in whom a screening test gives a positive result. Specificity is the proportion of people free of the disease in whom the screening test gives a negative result (i.e., the ability of a test to correctly identify those free of the disease in the screened population). The positive predictive value (PPV) is the proportion of individuals with a positive screening test who have the disease. PPV is a function of sensitivity, specificity, and disease prevalence (Cole et al., 1980). The most desirable performance characteristics of a screening test or program will differ according to the setting. Usually, high specificity is desirable whereas moderate sensitivity is acceptable. The reason for this is that low specificity leads to an undesirably large number of false-positive tests thereby exposing individuals to unnecessary and sometimes invasive diagnostic procedures. Two important fundamental characteristics of an effective screening program are: 1. The screening test must detect cancer early in a recognizable preclinical state. 2. Therapy initiated as a result of cancer being detected by screening must be more effective than treatment undertaken at the usual time of diagnosis. Understanding the impact of each component of a screening program is important in the evaluation. Specific criteria to be considered in the design of a screening program are: 1. The disease should be a serious health problem and the cause of substantial mortality and morbidity. However, it is important to appreciate that the life expectancy of a screened population may not be significantly affected even if the program is successful as there are other major competing causes of death. 2. The target population should be clearly defined and should have a reasonable disease prevalence. 3. The target population should be accessible with a reasonable expectation of response to a screening invitation. 4. The screening test should have acceptable performance characteristics (e.g., sensitivity, specificity) and be acceptable to those being screened. 5. The disease should be treatable and there should be a recognized treatment for lesions identified after screening. There is no value in establishing a screening program and identifying lesions that require treatment if the facilities are not available for appropriate

1310

follow-up. The implication is that sufficient personnel are present both for screening and for subsequent diagnosis and therapy. 6. The natural history of the condition should be known. Ideally this implies that it would be possible to decide previously when a screening test should be applied to achieve maximum benefit and minimal over-utilization of resources. Also, it is likely that some premalignant lesions may not progress or even regress (e.g., small adenomatous polyps of the colon or in situ cervix cancer) (Miller, 1996). 7. There should be an agreed upon policy for early recall of individuals with abnormal findings and on the frequency of routine recall for those with negative findings. 8. A quality control procedure should be implemented to maintain the sensitivity and specificity of the program and high compliance with diagnostic and therapeutic follow-up.

STUDY DESIGNS The two basic types of study designs that can be used to evaluate cancer screening are experimental studies and observational studies. The preeminent experimental study is the randomized controlled trial (RCT). This is the method of choice because its inherent structure is designed to eliminate the screening biases. Observational or quasiexperimental designs are used when an RCT is not possible. Because their comparison groups are not established by a random mechanism, they are generally difficult to analyze and interpret. Both cohort and case-control observational studies have been employed to evaluate cancer screening (Connor et al., 1991; Morrison et al., 1985; Prorok and Connor, 1986).

Randomized Trial Designs The classic two-arm trial answers a single question by randomly allocating participants to two groups, one offered screening according to a protocol and the other a control group not offered screening. Mortality rates in the two groups are compared at the end of the followup period (Prorok, 1995). This design was used in the Health Insurance Plan of Greater New York (HIP) trial of breast cancer screening, the first large randomized cancer screening trial (Shapiro et al., 1988). Extensions of the classic design have been considered to address more than one question simultaneously (Freedman and Green, 1990). The National Breast Cancer Screening Study in Canada includes two different study comparisons within the same administrative structure. The questions being addressed are to determine in women aged 40–49 at entry whether annual screening by mammography and physical examination, when used as an adjunct to the highest standard of care in the Canadian health-care system, can reduce mortality from breast cancer, and to evaluate in women aged 50–59 at entry the additional contribution of routine annual mammographic screening to screening by physical examination alone in reducing breast cancer mortality (Miller et al., 1981). It is also possible to address several related questions by including additional randomized groups, as in the University of Minnesota colon trial (Gilbertsen et al., 1980). The all-versus-none design is another extension, which is typically a two-group trial in which the intervention group includes multiple interventions (Freedman and Green, 1990), each intervention being an

Principles of Screening early detection maneuver for a different type of cancer. An example is the Prostate, Lung, Colorectal and Ovarian Cancer (PLCO) Screening Trial sponsored by the National Cancer Institute (Gohagan et al., 2000). The objectives of this trial are to determine whether in females and males, screening with 60-cm flexible sigmoidoscopy can reduce mortality from colorectal cancer, and screening with chest X-ray can reduce mortality from lung cancer; in males, screening with digital rectal examination plus serum prostate-specific antigen (PSA) can reduce mortality from prostate cancer; and in females, screening with serum CA 125 and transvaginal ultrasound can reduce mortality from ovarian cancer. Yet another possibility to answer more than one question at a time is the reciprocal control design (Freedman and Green, 1990). Participants in both arms of the trial receive an intervention, and they simultaneously serve as controls for the intervention in the other arm. Within each of the above designs there are options for the relationship between screening and follow-up (Etzioni et al., 1995). In the basic design, intervention arm participants receive periodic screening throughout the trial while control participants follow their usual medical care practices. This has been termed the continuous-screen design. A potentially serious drawback of the continuous-screen design is the prohibitive cost of long-term screening. An alternative is the stop-screen design in which screening is offered to intervention participants for a limited time and both arms are followed for disease end points. This design is preferred when it is expected that a reduction in mortality will emerge only after a long follow-up period, and when it is too expensive or infeasible to continue the periodic screening for the entire duration of the trial. The split-screen design is a variant of the stop-screen design in which the screening test is also offered to all participants in the control arm at the time the last screen is offered to those in the intervention arm. One objective of this design is the identification of comparable sets of cancer cases for analysis. A variant is the delayed-screen design in which periodic screening is offered to the control arm starting at some time point during the study and continuing until the end of the trial. This design is used to estimate the marginal effect of introducing screening at some standard time or age, relative to starting the screening at a later time or age (Prorok, 1995).

Observational Study Designs Observational studies lack a control group constructed by a chance mechanism. Consequently the groups being compared are potentially not comparable and any inference is compromised. An observational study design that can be used for evaluation when a screening test is in wide use, precluding the availability of an unscreened comparison group, or the cancer under consideration is very rare, thus requiring a prohibitively large sample size for an RCT, is a comparison of cancer incidence and mortality in a defined population before and after the introduction of a screening program. One can also examine time trends in incidence and mortality between intensively screened areas vs. nonscreened areas. Cervical cancer screening has been evaluated using this approach (Hakama et al., 1986). The case-control design is another observational evaluation approach (Cole et al., 1980; Morrison, 1982; Sasco et al., 1986; Weiss, 1983). Case-control studies have been reported for evaluation of breast, cervical, colon, and lung cancer screening (Connor et al., 1991; Ebeling and Nischan, 1987; Friedman et al., 1991; Oshima et al., 1986; Selby et al., 1992; Sobue et al., 1992; Sagawa et al., 2001). Appropriate definition of cases and controls is required. Cases are deaths from the cancer of interest in the population under study, whereas controls are chosen from all living individuals in the population from which the cases were derived. Under some circumstances, those who are diagnosed with advanced cancer may also serve as cases in such comparisons. The underlying assumption in all cohort and case-control study analyses is that the screened and unscreened cohorts, or the cases and controls, must be drawn from the same population. Selection bias in the choice to be screened or not renders the basic assumption questionable in most such studies. Despite potential distortions of bias, observational studies may be of value in addressing issues such as fre-

1311

quency of examination and duration of preventive benefits. These practical considerations can have significant impact on screening policy recommendations.

BIASES Cancer screening is predicated upon the assumption that there exists a preclinical state or manifestation of cancer that does not generate noticeable symptoms, but can be identified by a screening test. Intervention while the cancer is at this point in its development can lead to improved outcome for the individual (i.e., decreased probability of dying from the cancer). Careful study design (typically a randomized trial) and long-term follow-up of large populations are required to obtain an accurate estimate of improved outcome or mortality reduction. This is ordinarily a difficult undertaking, but one that is highly desirable whenever possible because alternative evaluation approaches are of suspect validity due to three key biases inherent in screening programs: lead time bias, length bias, and over-diagnosis bias. In a screening program, cancer may be detected earlier than it would have been in the absence of screening. The amount of time by which the diagnosis is advanced as a result of the earlier detection is termed the lead time. Since the time or age of diagnosis is advanced, survival duration is automatically lengthened for cancers detected by screening even if the individual does not live longer. This is lead time bias, which renders the case survival end point invalid. It is assumed that each cancer in a population has some preclinical duration, and all cancers in a population taken together comprise a distribution of preclinical durations. However, the durations of the cancers detected by screening are not a random sample from this distribution. Cancers with a longer duration of preclinical disease are over represented among the detected cases. This is length bias. Since cancers with long preclinical durations typically represent slowgrowing preclinical disease, which then progresses to slow-growing clinical disease, less aggressive cancers are the ones more likely to be detected by screening. Thus, screen-detected cancers will tend to have good prognostic characteristics, such as lack of regional lymph node involvement, and will tend to have a more favorable outcome, even in the absence of screening. Overdiagnosis bias is related to lead time and length biases. It is possible that a subset of preclinical cancers are non-progressive or even regressive in nature. Although they are detectable by the screening test, they would not progress to clinical disease during the person’s lifetime, and hence would never be diagnosed in the absence of screening. The detection of such cancers results in harm rather than benefit for the individuals involved, yet such cases likely remain preclinical for a long time and are therefore more likely to be detected in a program with repeated screenings. One consequence is that a screened population will contain a higher proportion of early-stage cases even if there is no mortality effect from screening. End points to evaluate cancer screening most frequently proposed as alternatives to mortality reduction are increased case finding rate of yield, stage shift to a grater proportion of early stage disease, and improved survival of cancers detected by screening. All are subject to the aforementioned biases and are therefore not reliable as measures of screening effect (Prorok, 1995).

LEVELS OF EVIDENCE The strength of evidence to support a specific screening procedure or method is variable. The strongest evidence would be obtained from a well-designed and well-conducted randomized controlled trial. It is not always possible to conduct such a trial to address important issues concerning screening because of practical limitations such as expense or rarity of the tumor type. According to the NCI (PDQ Editorial Board) five levels of evidence exist by which screening trials may be judged (PDQ Editorial Board, 2003).

1312

PART V: CANCER PREVENTION AND CONTROL

1. Evidence obtained from at least one well-designed and conducted randomized controlled trial. 2. Evidence obtained from well-designed and conducted nonrandomized controlled trials. 3. Evidence obtained from well-designed and conducted cohort or case-control studies. 4. Evidence obtained from multiple-time series with or without intervention. 5. Opinions of respected authorities based on clinical experience, descriptive studies, or reports of expert committees. Specific ethical issues arise in the context of screening. Screening is initiated by those who offer the tests, thereby conveying an implicit promise of benefit to all those who undergo screening. However, it may not be clear to the individual who volunteers to be screened that although it is hoped that the health of the community will be better, each individual included in a screening program may not benefit or may actually face some element of risk or even undergo some harm (Prorok et al., 1999).

POTENTIAL HARMS OF SCREENING The harms related to cancer screening can be separated into two primary categories—those associated with the screening test itself, and those associated with events triggered by a positive screen. Harms falling into the first category include radiation exposure from tests such as mammography, chest X-ray, and helical computed tomography (spiral CT), and actual physical injury such as perforation of the colon from flexible sigmoidoscopy or colonoscopy. The second category comprises the cost and morbidity (possible even death) associated with unnecessary diagnostic procedures and includes the finding of lesions that appear malignant upon histologic examination but have no clinical relevance and would never be diagnosed in the absence of screening, often referred to as overdiagnosis.

CASE STUDIES The following examples provide practical illustrations of screening in three tumor types, viz. colorectal neoplasia, lung cancer, and neuroblastoma.

Colorectal Cancer Five different tests for the screening of colorectal cancer will be discussed. Of these, the strongest evidence exists for fecal occult blood testing. Intermediate-level evidence is available for flexible sigmoidoscopy and only indirect evidence supports the use of colonoscopy and double-contrast barium enema.

Fecal Occult Blood Test Screening for the presence of blood in the stool is based upon the fact that most cancers and some adenomas bleed often intermittently (Simon et al., 1985). The amount of bleeding is dependent on the size of the adenoma or cancer and blood is unevenly distributed throughout the stool. Screening for the presence of blood in the stool is far less sensitive for adenomas than for cancers because adenomas, especially small ones, do not bleed or do so only infrequently (Macrae and St. John, 1982). Four randomized controlled studies have investigated fecal occult blood testing for colorectal cancer screening. Three of the trials have been completed and the fourth is still in progress (Mandel et al., 1993; Hardcastle et al., 1996; Kronborg et al., 1996). These trials incorporate a program of screening with multiple, annual or biennial tests rather than a single test in time. In studies using non-rehydrated samples, sensitivity ranged from 72%–78% with a specificity of 98% and a positive predictive value of 10%–17%. The sensitivity increased to 88%–92% when rehydrated samples were used; however, the specificity dropped to 90%–92% and the positive predictive value fell to 2%–6% (Mandel et al., 1993).

The Minnesota trial was initiated using non-rehydrated samples but slide processing was modified early in the trial to incorporate rehydration; at final analysis, 83% of the slides were developed after rehydration. Participants were randomly assigned to annual or biennial screening or to a control group. After 13 years, the group receiving annual screening showed a 33% reduction in colorectal cancer mortality whereas the group receiving biennial screening showed a nonsignificant 5% reduction (Mandel et al., 1993). Combination of the annual and the biennial groups resulted in an overall reduction of 19% in the risk of colorectal cancer death that was attributable to screening. Adverse events related to diagnostic colonoscopy, viz. perforation or hemorrhage, were reported to occur at the rate of 12 complications per 10,000 colonoscopies. In an analysis by other authors, it was estimated that one-third to one-half of the mortality reduction was due to the increased number of colonoscopic examinations and not attributable to fecal occult blood testing alone (Lang and Ransonoff, 1994). The assumptions in that analysis have been disputed by the authors of the Minnesota study. They concluded that 16%–25% of the reduction in colorectal cancer deaths effected by fecal occult blood testing was due to chance detection (Ederer et al., 1997). What effect would biennial screening have on colorectal cancer mortality? Two other prospective controlled trials offered biennial screening but did not perform rehydration of the fecal occult blood slides. Diagnostic evaluation of those with positive tests in both studies was performed by colonoscopic evaluation. Both studies had a low colonoscopy rate compared with the Minnesota study. The Nottingham trial had a mean follow-up of 7.8 years and showed a 15% reduction in colorectal cancer mortality (Hardcastle et al., 1996). The Funen study showed an 18% reduction in mortality after 10 years (Kronborg et al., 1996). Unpublished results from the Goteborg trial, published in the Cochrane Review indicate a 12% reduction in colorectal cancer mortality with biennial screening after 8 years of follow-up (Towler et al., 2002). The investigators reported a 0.3% complication rate (30 complications per 10,000 endoscopies), specifically perforation and hemorrhage, of 2298 endoscopies (colonoscopies and sigmoidoscopies). Using data from these four randomized controlled studies, a systematic review including a meta-analysis was published in the Cochrane Library (Towler et al., 2002). This analysis showed an overall significant reduction in colorectal cancer mortality with screening by fecal occult blood testing of 16% (RR 0.84, CI: 0.77–0.93). When the relative risk is adjusted for attendance for screening in individual studies, the mortality reduction is 23%. Furthermore, if 10,000 persons were offered screening and approximately two-thirds attended for at least one fecal occult blood test, there would be 8.5 deaths (CI: 3.6–13.5) from colorectal cancer prevented over 10 years. To prevent one death from colorectal cancer over 10 years, 1173 persons would need to be screened. However, the screening program would also result in 2800 participants having at least one colonoscopy. If harmful effects of screening from the Minnesota trial are considered, there would be 3.4 colonoscopy complications. The estimate of mortality reduction from the randomized controlled trials of fecal occult blood tests is now well quantified and indicates benefit in a program of colorectal cancer screening. However, the fairly wide range of mortality reduction seen in these studies and the overall modest mortality reduction indicates a need for continued improvement in screening technology. For example, detection of gene mutations in DNA from stool samples may facilitate early detection in the future (Traverso et al., 2002). Other benefits of fecal occult blood testing are emerging. Notably, a reduction in the incidence of colorectal cancer of 20% in subjects screened annually, has been observed in the Minnesota trial (Mandel et al., 2000). In all three randomized studies evaluating the effectiveness of fecal occult blood testing, a favorable stage shift to earlier stage disease, which has better outcomes, was seen. In the Nottingham study, 90% of the screened group had Dukes A or B lesions compared with 40% of the control group (Hardcastle et al., 1996). A similar stage shift was seen in the other randomized controlled trials described.

Principles of Screening Immunochemical tests use monoclonal and/or polyclonal antibodies that detect the intact globin portion of human hemoglobin. Diet does not affect the immunochemical tests thus obviating a potential source of false-positive tests and the absence of any dietary modifications may enhance patient acceptance. While only a limited number of individuals have been screened using immunochemical tests, it appears as if these tests are at least as sensitive and specific as the guaiac-based tests (Levin et al., 2003).

Flexible Sigmoidoscopy The rationale for screening with flexible sigmoidoscopy is that it provides direct visualization of the colon, and suspicious lesions can be biopsied. The most obvious disadvantage is that it permits examination of only that portion of the distal colon within reach of the endoscope. Approximately 65%–75% of adenomatous polyps and 40%–65% of colorectal cancers are within the reach of a 60-cm flexible sigmoidoscope (Tedesco et al., 1980; Shinya and Wolff, 1979; Winawer et al., 1983; US Preventive Services Task Force, 1996). Patients with a positive examination require further evaluation by colonoscopy. It is well established that patients with an adenomatous polyp found on sigmoidoscopy have an increased probability of additional more proximal lesions (Winawer et al., 1992; Grossman et al., 1989; Tripp et al., 1987). The sensitivity of flexible sigmoidoscopy is 96.7% for cancer and large polyps and 73.3% for small polyps. The specificity is 94% for cancer and large polyps with a 92% specificity for small polyps (Winawer et al., 2003). Only indirect evidence from several case-control studies using either rigid sigmoidoscopy or a combination of rigid with flexible sigmoidoscopy currently exists to support the effectiveness of flexible sigmoidoscopy (Selby et al., 1993; Flexible Sigmoidoscopy Screening Trial Investigators, 2002). The best designed study, by Selby et al., avoided many of the biases inherent in case-control studies. The screening histories of persons who died of colorectal cancer were compared against controls and a 59% reduction in mortality from cancers of the rectum and distal colon was found in individuals who had undergone sigmoidoscopic evaluation (Winawer et al., 1992). Newcomb et al. (1992) reported an 80% reduction in mortality from cancer of the rectum and distal colon in persons who had ever undergone sigmoidoscopic examination compared with individuals who had never done so. Although potential biases limit the applicability of this study, it does provide independent support for the effectiveness of flexible sigmoidoscopy in a colorectal cancer screening program. Of great interest is the optimal interval for screening sigmoidoscopy. In the study by Selby et al. (1993), described above, the effectiveness of screening sigmoidoscopy was found to be just as great for patients who had undergone the procedure 9–10 years before compared with those who had just undergone the examination (Tripp et al., 1987). A modeling study evaluating the optimal interval for sigmoidoscopic screening found that 90% of the effectiveness of annual screening was preserved with an interval of 10 years (US Preventive Services Task Force, 1996). This model assumes that adenomatous polyps take 10–14 years to evolve into invasive cancers. There is some concern about the robustness of this assumption. For example, in a recent study of flexible sigmoidoscopy in the PLCO trial (Schoen et al., 2003) nearly 1% of patients (n = 72) had a cancer (n = 6) or advanced adenoma, the kind most likely to become malignant, at the 3-year re-examination point. Of individuals detected with advanced distal adenomas detected at the year 3 examination, 81% had lesions found in a portion of the colon that had been adequately examined at the initial examination. The baseline findings of a multi-center randomized trial from the United Kingdom have been reported (Flexible Sigmoidoscopy Screening Trial Investigators, 2002); 194,726 of 354,262 (55%) of those aged 55–64 years in 14 UK Centers invited to undergo screening accepted and 170,432 eligible individuals were randomized. Attendance among those assigned screening was 71%; 2131 (5%) were classified as high risk and referred for colonoscopy and 38,525 with no polyps or only low-risk polyps detected were discharged. Distal adenomas were detected in 493 (12%) and distal cancer in 131 (0.3%). Proximal ade-

1313

nomas were detected in 386 (18.8% of those undergoing colonoscopy) and proximal cancer in nine cases (10.4%). Sixty-two percent of cancers were Dukes Stage A. There was one perforation after flexible sigmoidoscopy and four after colonoscopy. The baseline findings of a multi-center randomized trial in Italy in individuals aged 55–64 have also been reported (Segnan et al., 2002). Distal adenomas were detected in 1070 subjects (10.8%). Proximal adenomas were detected in 116 of 747 (15.5%) subjects without cancer at sigmoidoscopy who then underwent colonoscopy. A total of 54 subjects were found to have colorectal cancer, a rate of 5.4 per 1000 (54% of which were Dukes A). Two perforations occurred (1 in 991) sigmoidoscopy and (1 in 77) colonoscopies and one hemorrhage requiring hospitalization. The long-term results of both these randomized trials are awaited with interest.

Double-Contrast Barium Enema Evidence for the use of the double-contrast barium enema in screening is limited. Detection of adenomatous polyps and early cancers in other types of screening studies such as fecal occult blood testing resulting in a reduction in the incidence and mortality of colorectal cancer provides indirect evidence that double-contrast barium enema, which detects many of these lesions, would be beneficial. The sensitivity of double-contrast barium enema is 84% for cancer, 82% for large polyps, and 67% for small polyps. The specificity is 97.5% for cancer, 83.3% for large polyps, and 75% for small polyps (Winawer et al., 2003). One randomized controlled trial investigated the addition of double-contrast barium enema to sigmoidoscopy compared with colonoscopy. Colonoscopy was found to be more sensitive in detecting small polyps but no difference was found between the groups for large polyps and cancer (Rex et al., 1990). The frequency with which double-contrast barium enema should be performed for screening purposes has not been well studied. An interval of 5 years has been suggested based on an estimated adenoma dwell time of 10 years and the performance characteristics of the double-contrast barium enema, which is known to be less sensitive in detecting small polyps (Winawer et al., 2003).

Colonoscopy Colonoscopy is the only technique that offers screening, diagnostic, and when necessary, therapeutic management in one procedure. Most data available on the effects of colonoscopy are derived either from studies of colonoscopy in a diagnostic and surveillance setting or from indirect evidence as outlined above for double-contrast barium enema. There are no studies currently available that evaluate colonoscopy as a screening test in terms of reduction of colorectal cancer mortality. However, to the extent that colonoscopy is a significant part of the fecal occult blood test program, these trials of occult blood testing also provide evidence of the effectiveness of colonoscopy. Additional support is provided from one case-control study, which showed that persons who had undergone colonoscopy had a 70%–80% reduction in colorectal cancers (Muller and Sonnerberg, 1995). A feasibility trial of screening colonoscopy has been launched in the United States (Winawer S, personal communication, 2003). Colonoscopy can detect both polyps and cancers, although it is less accurate when the lesions are small. In studies evaluating the performance of colonoscopy, it has been demonstrated that 15% of small polyps are missed but few large polyps are missed (Hixson et al., 1990). False-positive results are rare but about one-third of polyps removed are not adenomatous (Bernstein et al., 1985). Colonoscopic sensitivity is 96.7% for cancers, 85% for large polyps, and 78.5% for small polyps; specificity is 98% for all lesions (Winawer et al., 2003). No studies address the optimal frequency with which colonoscopic screening should be carried out. Based on the natural history of the disease and the high accuracy of colonoscopy in the detection of polyps, it has been suggested that a screening interval of 10 years would be protective (Winawer et al., 2003). This is supported by the case-control study of Selby et al. (1993) evaluating sigmoidoscopy, which suggests a protective effect for up to 10 years. Although a randomized trial of colonoscopy has not been performed, two large-scale demonstration projects have been recently

1314

PART V: CANCER PREVENTION AND CONTROL

reported (Imperiale et al., 2000; Lieberman et al., 2000). The first was from a group of V.A. Centers involving 3212 individuals (97% men) and the other one included 2000 men and women, who were employed by Eli Lilly, Inc. The adenomatous polyp rate was 38% in the V.A. cohort and 20% in the Lilly cohort. The rate of advanced proximal lesions varied from 2%–4% and about half of these would have been missed by sigmoidoscopy. The complication rate was low; 97% of examinations reached the cecum with a perforation rate of 0.02% without any deaths. It remains to be seen whether colonoscopy will become a primary screening test in view of its expense and invasiveness. The emerging technology of virtual colonoscopy (threedimensional colography) is anticipated in the hope that a non-invasive method of imaging the entire colon will increase compliance with colorectal cancer screening. Results from a recent study using new technology indicated the accuracy of CT colonography to be comparable to conventional colonoscopy (Pickhardt, 2003). In expert centers, polypoid lesions larger than 10 mm can be detected with sensitivity and specificity approaching 90% with sensitivity falling to 50% for polyps 5 mm in size. CT colonography, in some studies, has been shown to be accurate in detecting colon cancer with a sensitivity of 100% (Ferrucci, 2001). Further study is required to evaluate the performance characteristics of virtual colonoscopy using the best available technology when applied outside of a single center. In the future, avoidance of the need to undergo bowel preparation and advances in software design may enhance the public appeal of this method of colonic examination. One significant concern is the detection of extra-colonic abnormalities that are of no clinical significance but may lead to expensive diagnostic testing or laparotomy (Gluecker et al., 2003).

Digital Rectal Examination Less than 10% of colorectal cancers are within the 7–8-cm reach of the examining finger (Yee et al., 2001). Additionally, stool obtained during the course of a digital rectal examination is an inadequate sample upon which to screen for blood and this type of fecal occult blood testing is not recommended. Finally, there is no evidence that digital rectal examination reduces morbidity or mortality from colorectal cancer, and it is not currently indicated as a screening test for the prevention or early detection of colorectal cancer (Winawer et al., 2003).

Screening Recommendations Using the above evidence, the interdisciplinary Task Force, initially convened by the Agency for Health Care Policy and Research and completed with funding from seven professional societies, developed recommendations for the screening of colorectal cancer (Winawer et al., 2003). In 1997, the American Cancer Society published its recommendations for screening, which were based largely upon, and nearly identical to those developed by the Agency for Health Care Policy and Research Task Force and updated in 2001 (Smith et al., 2003).

Lung Cancer The most common screening tests for lung cancer are the chest X-ray and sputum cytology, while spiral CT (helical computed tomography) has recently gained attention. The initial studies aimed at evaluating X-ray and cytology were conducted decades ago and include the Philadelphia Pulmonary Neoplasm Research Project (Boucot and Weiss, 1973), the Veterans Administration study (Lilienfeld et al., 1966), the South London Lung Cancer Study (Nash et al., 1968), the North London Cancer Study (Brett, 1968; Brett, 1969), and the Kaiser Foundation Health Plan multiphasic screening trial (Dales et al., 1979; Friedman et al., 1986). Some of these studies included control groups whereas others did not, and only the latter two were randomized. None reported a statistically significant impact of screening on lung cancer mortality. Some, such as the South London Study, reported an increase in survival of screen-detected lung cancer cases, but there was no adjustment for lead time bias, overdiagnosis bias, or length bias. The

sample sizes of these studies were relatively small and the follow-up period was typically less than 10 years. A subsequent series of observational studies included a nonrandomized but controlled trial in the former German Democratic Republic (GDR) (Wilde, 1989) and case-control studies in the former GDR (Ebeling and Nischan, 1987) and Japan (Sobue et al., 1992). Screening in the intervention arm in the GDR nonrandomized study used semiannual chest fluoroscopy over a 6-year period. Control participants were offered the same exam at 1- to 2-year intervals. There was no reduction in lung cancer mortality; the relative risk (screen group/control group) was 1.34 (95% CI: 0.94–1.98). The German case-control study evaluated chest X-rays originally used for control of tuberculosis using two control groups, a general population group and a hospital group. The odds ratios of lung cancer death associated with receiving chest X-ray were 0.9 and 1.1, respectively (95% CI: 0.5–1.5 and 0.7–1.8). Chest X-ray was investigated among deceased lung cancer cases and matched controls in the Japanese case-control study. A suggestion of some screening benefit was reported; the odds ratio of dying from lung cancer for those screened within 12 months vs. those not screened was 0.72 (95% CI: 0.50–1.03). Several RCTs have also been conducted to evaluate lung cancer screening. Participants in the two arms of the Czechoslovakian trial (Kubik et al., 1990) were screened with X-ray and cytology at two different frequencies, semiannual vs. every 3 years. The reported relative risk (screen group/control group) was 1.36 (95% CI: 0.94–1.98) indicating no difference in lung cancer mortality. Three other randomized trials were conducted in the United States. Participants in the Mayo Lung Project (MLP) were males 45 years or older who were heavy smokers (Fontana, 1984; Fontana, 1985; Fontana, 1986). They were randomized either to screening with sputum cytology and chest X-ray every 4 months or to a control group advised one time at baseline to seek screening annually. Participants in the Johns Hopkins University (Stitik, 1978; Levin et al., 1982; Stitik et al., 1985; Tockman et al., 1985) and Memorial-Sloan Kettering Cancer Center (Melamed et al., 1981; 1984), RCTs were randomized to intervention and control groups that were both offered annual chest X-ray. The intervention group was also offered sputum cytology every 4 months. No reduction in lung cancer mortality was reported from any of the three trials. The lack of mortality reduction in the MLP persisted when follow-up was extended to a median of 20.5 years (Marcus et al., 2000). Although the MLP has been interpreted as indicating no benefit from chest X-ray screening, several reservations have been raised. The sample size was modest, designed to detect a 50% reduction in lung cancer mortality. There was insufficient statistical power to demonstrate a lesser but medically important reduction of 10%–15%. It has also been reported that 50% or more of the men in the control group received an annual chest X-ray (Fontana, 1986) so that contamination may have been sufficient to obscure an effect. Finally, improvements in therapy may make early detection more effective. Furthermore, more recent Japanese case-control studies have suggested a benefit from chest X-ray screening (Nishii et al., 2001; Sagawa et al., 2001; Tsukada et al., 2001; Nakayama et al., 2002). From currently available data, there is no definitive evidence that screening for lung cancer with either chest X-ray or sputum cytology can reduce lung cancer mortality, but there are concerns about the design or interpretation of reported studies. Consequently, the National Cancer Institute (NCI) is conducting the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial. This is a long-term RCT in which 37,000 men are screened for prostate, lung, and colorectal cancers and 37,000 women are screened for lung, colorectal, and ovarian cancers. The same numbers of men and women are observed with routine medical care as controls. Screening in the lung component employs annual postero-anterior (PA) view chest X-ray as the intervention modality (Gohagan et al., 2000). In addition to ongoing evaluation of chest X-ray, there is substantial interest in assessment of low-dose helical computed tomography (spiral CT) as a screening modality for lung cancer. Interest was generated by a small observational study, the Early Lung Cancer Action Project, which reported increased sensitivity of spiral CT relative to chest X-ray (Henschke et al., 1999). At a prevalence screen, spiral CT

Principles of Screening found 27 lung cancers and chest X-ray only 6 of the 27 among 1000 individuals screened with both modalities. Several additional studies of spiral CT have been reported (Hasegawa et al., 2000; Sone et al., 2001; Diederich et al., 2002; Swensen et al., 2002). However, there are important harms that must be weighed against any potential benefit of screening with spiral CT. The first is falsepositive results, which can lead to anxiety and invasive diagnostic procedures, such as percutaneous needle biopsy or thoracotomy. In reported studies, the percentages of subjects with noncalcified nodules were 21% (Henschke et al., 1999), 51% (Swensen et al., 2002), and 43% (Diederich et al., 2002). Only a few of these positive screenees were found to have lung cancer. In addition, screening with spiral CT can detect abnormalities outside the chest such as renal masses and abdominal aortic aneurysms. The second, perhaps less familiar, harm is overdiagnosis (Black, 2000). This is defined as the diagnosis of a cancer that would not have become clinically significant (not be clinically diagnosed) in the absence of screening. The enhanced sensitivity of spiral CT imaging could lead to unnecessary diagnosis of lung lesions that appear histologically to be lung cancer but would never surface clinically in the absence of such screening. Since overdiagnosis is virtually impossible to document at the individual level, these individuals may be subject to unnecessary surgery, chemotherapy, or radiation. A Japanese study provides a clear suggestion that screening with spiral CT may result in substantial overdiagnosis (Sone et al., 2001). Smokers and nonsmokers were annually screened for lung cancer with spiral CT from 1996–1998. Surprisingly, the percentage of screen-detected lung cancers was nearly identical in the two subgroups: 0.46% for smokers and 0.44% for nonsmokers. It is critical to determine whether screening with spiral CT results in a lung cancer mortality reduction and whether such screening does more good than harm before it is accepted into medical practice (Patz et al., 2000; Swensen et al., 2002). The NCI is conducting the National Lung Screening Trial to estimate screening benefit and investigate false positives and overdiagnosis. Fifty thousand individuals aged 55–74 years will be randomized to either chest X-ray or spiral CT for three annual screens and followed for cancer incidence and vital status.

Neuroblastoma Screening for neuroblastoma involves subjecting urine samples to biochemical tests for metabolites of nor epinephrine and dopamine, (i.e., vanillylmandelic acid (VMA) and homovanillic acid (HVA)). Over three-fourths of neuroblastoma cases excrete these substances into the urine (Williams and Greer, 1963). Liquid urine samples or urine samples collected on filter paper are tested for VMA and HVA (Tuchman et al., 1987). The VMA and HVA levels are usually measured by gas chromatography, thin layer chromatography, and/or high performance liquid chromatography. There is no known optimal age for screening. Screening has been performed at 6 months and 12 months of age. Of note, the clinical significance of screen-detected neuroblastomas is in question. In an observational study, early-stage localized tumors were observed to regress without treatment (Yamamoto et al., 1998). Estimates of the sensitivity of the screening procedure from different studies range from 40%–80% (Chamberlain, 1994; Woods et al., 2002; Nishi et al., 1991; Chamberlain, 1996). The prevalence of neuroblastoma is very low, resulting in a low positive predictive value of screening. In the Quebec Neuroblastoma Screening Project, the positive predictive value was 52% (Woods et al., 2002). Neuroblastoma screening using urinary testing began in Japan in the early 1970s (Sawada, 1992), and evidence of screening effect has been gathered from descriptive studies and uncontrolled pilot experiences in Japan, Europe, and the United States, as well as population-based studies in Canada and Germany (Takeda et al., 1989; Tuchman et al., 1989; Bessho et al., 1991; Parker et al., 1992; Woods et al., 1992; Schilling et al., 1994; Woods et al., 1996; Chauvin et al., 1997; Schilling et al., 2000; Schilling et al., 2002; Woods et al., 2002). However, there is no evidence from controlled studies or randomized trials of decreases in mortality associated with screening. As would be

1315

expected, screening results in an increased incidence of early-stage disease, and the screen-detected cases typically exhibit biologically favorable characteristics such as unamplified N-myc oncogene and favorable histology. These cases also have a high probability of survival, whether detected by screening or detected clinically (Naito et al., 1990; Look et al., 1991; Woods et al., 1992; Asami et al., 1995; Takeuchi et al., 1995; Yamamoto et al., 1995; Woods et al., 1996; Bernstein and Woods, 1996; Bowman et al., 1997; Brodeur et al., 2001; Schilling et al., 2002; Woods et al., 2002). Furthermore, it appears that some tumors regress spontaneously in the absence of treatment (Yamamoto et al., 1998; Nishihira et al., 2000; Tanaka et al., 2000; Yoneda et al., 2001). Taken together, these findings strongly suggest that length bias and/or overdiagnosis bias are operative in neuroblastoma screening. The Quebec Neuroblastoma Screening Project ascertained neuroblastoma incidence and mortality in a cohort of 476,603 children in the Canadian province of Quebec. Urinary screening was offered to children at 3 weeks and 6 months with 92% compliance. The incidence and mortality rates were compared with several North American cohorts to whom screening was not offered. In Quebec, the incidence of early-stage disease in children under 1 year was substantially higher than expected (standardized incidence ratio 3.03; 95% CI: 2.30–3.86). In the control populations, the ratio was close to that expected (0.82 in Minnesota (95% CI: 0.41–1.38) and Ontario (95% CI: 0.53–1.17) (Woods et al., 1996). The neuroblastoma death rates were very similar in the screened and control populations after about 8 years of followup (standardized mortality ratio 1.11; 95% CI: 0.64–1.92 for the Quebec cohort compared with Ontario children) (Woods et al., 2002). Similar findings were reported from the German neuroblastoma study (Schilling et al., 2002). References Asami T, Otabe N, Wakabayashi M, et al. 1995. Screening for neuroblastoma: A 9-year birth cohort-based study in Niigata, Japan. Acta Paediatr 84:1173–1176. Bernstein MA, Feczko PJ, Halpert RD, et al. 1985. Distribution of colonic polyps: Increased incidence of proximal lesions in older patients. Radiology 155:35–38. Bernstein ML, Woods WG. 1996. Screening for neuroblastoma. In: Miller AB, ed. Advances in Cancer Screening. Boston, MA: Kluwer Academic Publishers, pp. 149–163. Bessho F, Hashizume K, Nakajo T, et al. 1991. Mass screening in Japan increased the detection of infants with neuroblastoma without a decrease in cases in older children. J Pediatr 119:237–241. Black WC. 2000. Overdiagnosis: An under recognized cause of confusion and harm in cancer screening. J Natl Cancer Inst 92:1280–1282. Boucot KR, Weiss W. 1973. Is curable lung cancer detected by semiannual screening? JAMA 224:1361–1365. Bowman LC, Castleberry RP, Cantor A, et al. 1997. Genetic staging of unresectable or metastatic neuroblastoma in infants: A Pediatric Oncology Group study. J Natl Cancer Inst 89:373–380. Brett GZ. 1968. The value of lung cancer detection by six-monthly chest radiographs. Thorax 23:414–420. Brett GZ. 1969. Earlier diagnosis and survival in lung cancer. Br Med J 4:260–262. Brodeur GM, Look AT, Shimada H, et al. 2001. Biological aspects of neuroblastomas identified by mass screening in Quebec. Med Pediatr Oncol 36:157–159. Chamberlain J. 1994. Screening for neuroblastoma: A review of the evidence. J Med Screen 1:169–175. Chamberlain J. 1996. Neuroblastoma. In: Chamberlain J, Moss S, eds. Evaluation of Cancer Screening. London: Springer, pp. 145–149. Chauvin F, Mathieu P, Frappaz D, et al. 1997. Screening for neuroblastoma in France: Methodological aspects and preliminary observations. Med Pediatr Oncol 28:81–91. Cole P, Morrison AS. 1980. Basic issues in population screening for cancer. J Natl Cancer Inst 64:1263–1272. Connor RJ, Prorok PC, Weed DL. 1991. The case-control design and the assessment of the efficacy of cancer screening. J Clin Epidemiol 44:1215–1221. Dales LG, Friedman GD, Collen MF. 1979. Evaluating periodic multiphasic health checkups: A controlled trial. J Chronic Dis 32:385–404. Diederich S, Wormanns D, Semik M, et al. 2002. Screening for early lung cancer with low-dose spiral CT: Prevalence in 817 asymptomatic smokers. Radiology 222:773–781.

1316

PART V: CANCER PREVENTION AND CONTROL

Ebeling K, Nischan P. 1987. Screening for lung cancer—results from a casecontrol study. Intl J Cancer 40:141–144. Ederer F, Church TR, Mandel JS. 1997. Fecal occult blood screening in the Minnesota study: Role of chance detection of lesions. J Natl Cancer Inst 89:1423–1428. Etzioni RD, Connor RJ, Prorok PC, Self SG. 1995. Design and analysis of cancer screening trials. Stat Methods Med Res 4:3–17. Ferrucci JT. 2001. Colon cancer screening with virtual colonoscopy: Promise, polyps, politics. Am J Roentgenol 177:974–988. Flexible Sigmoidoscopy Screening Trial Investigators. 2002. Lancet 359: 1291–1300. Fontana RS. 1984. Early detection of lung cancer: The Mayo Lung Project. In: Prorok PC, Miller AB, eds. Screening for Cancer, I: General Principles on Evaluation of Screening for Cancer and Screening for Lung, Bladder, and Oral Cancer, vol. 78. Geneva, Switzerland: International Union Against Cancer, pp. 107–122. Fontana RS. 1985. Screening for lung cancer. In: Miller AB, ed. Screening for Cancer. New York: Academic Press, pp. 377–395. Fontana RS. 1986. Screening for lung cancer: Recent experience in the United States. In: Hansen HH, ed. Lung Cancer: Basic and Clinical Aspects. Boston, MA: Martinus Nijhoff Publishers, pp. 91–111. Freedman LS, Green SB. 1990. Statistical designs for investigating several interventions in the same study: Methods for cancer prevention trials. J Natl Cancer Inst 82:910–914. Friedman GD, Collen MF, Fireman BH. 1986. Multiphasic Health Checkup Evaluation: A 16-year follow-up. J Chronic Dis 39:453–463. Friedman GD, Hiatt RA, Quesenberry CP, Selby JV. 1991. Case-control study of screening for prostatic cancer by digital rectal examinations. Lancet 337:1526–1529. Gilbertsen VA, Church TR, Grewe FA, et al. 1980. The design of a study to assess occult-blood screening for colon cancer. J Chron Dis 33: 107–114. Gluecker TM, Johnson CD, Wilson LA, et al. 2003. Extracolonic findings at CT colonography. Evaluation of prevalence and cost in a screening population. Gastroenterology 124:911–916. Gohagan JK, Levin DL, Prorok PC, Sullivan D, eds. 2000. The Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening. Control Clin Trials 21(6 Suppl): 249s–406s. Grossman S, Milos ML, Tekawa IS, et al. 1989. Colonoscopic screening of persons with suspected risk factors for colon cancer. II. Past history of colorectal neoplasms. Gastroenterology 96:299–306. Hakama M, Miller AB, Day NE, eds. 1986. Screening for Cancer of the Uterine Cervix, IARC Scientific Publ No 76. International Agency for Research on Cancer, Lyon. Hardcastle JD, Chamberlain JO, Robinson MHE, et al. 1996. Randomized controlled trial of faecal-occult-blood screening for colorectal cancer. Lancet 348:1472–1477. Hasegawa M, Sone S, Takashima S, et al. 2000. Growth rate of small lung cancers detected on mass CT screening. Br J Radiol 73:1252–1259. Henschke CI, McCauley DI, Yankelevitz DF, et al. 1999. Early Lung Cancer Action Project: Overall design and findings from baseline screening. Lancet 354:99–105. Hixson LJ, Femerty MB, Sampliner RE, et al. 1990. Prospective study of the frequency and size distribution of polyps missed by colonoscopy. J Natl Cancer Inst 82:1769–1772. Imperiale TF, Wagrer DN, Lin CY, et al. 2000. Risk of advanced proximal neoplasms in asymptomatic adults according to the distal colorectal findings. N Engl J Med 343:169–174. Kronborg O, Fenger C, Olsen J, et al. 1996. Randomised study of screening for colorectal cancer with faecal-occult-blood test. Lancet 348: 1467–1471. Kubik A, Parkin DM, Khlat M, et al. 1990. Lack of benefit from semi-annual screening for cancer of the lung: Follow-up report of a randomized controlled trial on a population of high-risk males in Czechoslovakia. Intl J Cancer 45:26–33. Lang CA, Ransohoff DF. 1994. Fecal occult blood screening for colorectal cancer. Is mortality reduced by chance selection for screening colonoscopy? JAMA 271:1011–1013. Levin B, Brooks D, Smith RA, Stone A. 2003. Emerging technologies in screening for colorectal cancer: CT colonography, immunochemical fecal occult tests and stool screening using molecular markers. CA Cancer J Clin 53:44–55. Levin ML, Tockman MS, Frost JK, et al. 1982. Lung cancer mortality in males screened by chest X-ray and cytologic sputum examination: A preliminary report. Recent Results Cancer Res 82:138–146. Lieberman DA, Weiss DG, Bond JH, et al. 2000. Use of colonoscopy to screen asymptomatic adults for colorectal cancer. Veterans Affairs Cooperative Study Group 380. N Engl J Med 343:162–168.

Lilienfeld A, Archer PG, Burnett CH, et al. 1966. An evaluation of radiologic and cytologic screening for the early detection of lung cancer: A cooperative pilot study of the American Cancer Society and the Veterans Administration. Cancer Res 26:2083–2121. Look AT, Hayes FA, Shuster JJ, et al. 1991. Clinical relevance of tumor cell ploidy and N-myc gene amplification in childhood neuroblastoma: A Pediatric Oncology Group study. J Clin Oncol 9:581–591. Macrae FA, St. John DJ. 1982. Relationship between patterns of bleeding and Hemoccult sensitivity in patients with colorectal cancers and adenomas. Gastroenterology 82:891–898. Mandel JS, Bond JH, Church TR, et al. 1993. Reducing mortality from colorectal cancer by screening for fecal occult blood. Minnesota Colon Cancer Control Study. N Engl J Med 328:1365–1371. Mandel JS, et al. 2000. The effect of fecal occult blood screening on the incidence of colorectal cancer. N Engl J Med 343:1603–1607. Marcus PM, Bergstralh EJ, Fagerstrom RM, et al. 2000. Lung cancer mortality in the Mayo Lung Project: Impact of extended follow-up. J Natl Cancer Inst 92:1308–1316. Melamed MR, Flehinger BJ, Zaman MB, et al. 1981. Detection of true pathologic stage I lung cancer in a screening program and the effect on survival. Cancer 47(5 Suppl):1182–1187. Melamed MR, Flehinger BJ, Zaman MB, et al. 1984. Screening for early lung cancer. Results of the Memorial Sloan-Kettering study in New York. Chest 86:44–53. Miller AB. 1996. Fundamental issues in screening for cancer. In: Schottenfeld D, Fraumeni JF, eds. Cancer Epidemiology and Prevention 2nd ed. New York: Oxford University Press, pp. 1433–1452. Miller AB, Howe GR, Wall C. 1981. The National Study of Breast Cancer Screening Protocol for a Canadian Randomized Controlled trial of screening for breast cancer in women. Clin Invest Med 4:227–258. Morrison AS, ed. 1985. Screening in Chronic Disease. New York: Oxford University Press, pp. 95–117. Morrison AS. 1982. Case definition in case-control studies of the efficacy of screening. Am J Epidemiol 115:6–8. Muller AD, Sonnenberg A. 1995. Prevention of colorectal cancer by flexible endoscopy and polypectomy. A case-control study of 32 702 veterans. Ann Intern Med 123:904–910. Naito H, Sasaki M, Yamashiro K, et al. 1990. Improvement in prognosis of neuroblastoma through mass population screening. J Pediatr Surg 25:245–248. Nakayama T, Baba T, Suzuki T, Sagawa M, Kaneko M. 2002. An evaluation of chest X-ray screening for lung cancer in gunma prefecture, Japan: A population-based case-control study. Eur J Cancer 38:1380–1387. Nash FA, Morgan JM, Tomkins JG. 1968. South London lung cancer study. Br Med J 2:715–721. Newcomb PA, Norfleet RG, Storer BE. 1992. Screening sigmoidoscopy and colorectal cancer mortality. J Natl Cancer Inst 84:1572–1575. Nishi M, Miyake H, Takeda T, et al. 1991. Mass screening for neuroblastoma and estimation of costs. Acta Paediatr Scand 80:812–817. Nishihira H, Toyoda Y, Tanaka Y, et al. 2000. Natural course of neuroblastoma detected by mass screening: A 5-year prospective study at a single institution. J Clin Oncol 18:3012–3017. Nishii K, Ueoka H, Kiura K, et al. 2001. A case-control study of lung cancer screening in Okayama Prefecture, Japan. Lung Cancer 34:325–332. Oshima A, Hirata N, Ubukata T, Urneda K, Fujimato I. 1986. Evaluation of a mass screening program for stomach cancer with a case-control design. Intl J Cancer 38:829–833. Parker L, Craft AW, Dale G, et al. 1992. Screening for neuroblastoma in the north of England. BMJ 305:1260–1263. Patz EF, Goodman PC, Bepler G. 2000. Screening for lung cancer. N Engl J Med 343:1627–1633. PDQ Editorial Board 2003, personal communication. Pickhardt PJ, Choi JR, Hwang I, et al. 2003. Computed tomographic virtual colonoscopy to screen for colorectal neoplasia in asymptomatic adults. N Engl J Med 349:2191–2200. Prorok PC, Connor RJ. 1986. Screening for the early detection of cancer. Cancer Invest 4:225–238. Prorok PC, Kramer BS, Gohagan JK. 1999. Screening theory and study design. In: Kramer BS, Gohagan JK, Prorok PC, eds. Cancer Screening. New York: M. Dekker, pp. 29–53. Prorok PC. 1995. Screening studies. In: Greenwald P, Kramer BS, Weed DL, eds. Cancer Prevention and Control. New York: Marcel Dekker, pp. 225–242. Report of the US Preventive Services Task Force. 1996. Guide to clinical preventive services. 2nd ed. Baltimore, MD: williams & Wilkins. Rex DK, Weddle RA, Lehman GA, et al. 1990. Flexible sigmoidoscopy plus air contrast barium enema versus colonoscopy for suspected lower gastrointestinal bleeding. Gastroenterology 98:855–861.

Principles of Screening Sagawa M, Tsubono Y, Saito Y, et al. 2001. A case-control study for evaluating the efficacy of mass screening program for lung cancer in Miyagi Prefecture, Japan. Cancer 92:588–594. Sasco AJ, Day NE, Walter SD. 1986. Case-control studies for the evaluation of screening. J Chron Dis 39:399–405. Sawada T. 1992. Past and future of neuroblastoma screening in Japan. Am J Pediatr Hematol Oncol 14:320–326. Schilling FH, Berthold F, Erttmann R, et al. 2000. Population-based and controlled study to evaluate neuroblastoma screening at one year of age in Germany: Interim results. Med Pediatr Oncol 35:701–704. Schilling FH, Erttmann R, Ambros PF, et al. 1994. Neuroblastoma with unfavourable prognostic parameters detected by mass screening: report of the German pilot study [abstract]. In: Proceedings of the American Society of Clinical Oncology, vol 13(A-1440), p. 421. Schilling FH, Spix C, Berthold F, et al. 2002. Neuroblastoma screening at one year of age. N Engl J Med 346:1047–1053. Schoen RE, Pinsky PF, Weissfeld JL, et al. 2003, for the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial Group. Results of repeat sigmoidoscopy 3 years after a negative examination. JAMA 290:41– 48. Segnan N, Seenore C, Andreoni B, et al. 2002. Baseline findings of the Italian multicenter randomized controlled trial of “once-only sigmoidoscopy”— SCORE. JNCI 94:1763–1772. Selby JV, Friedman GD, Quesenberry CP Jr., et al. 1992. A case-control study of screening sigmoidoscopy and mortality from colorectal cancer. N Engl J Med 326:653–657. Selby JV, Friedman GD, Quesenberry CP Jr., et al. 1993. Effect of fecal occult blood testing on mortality from colorectal cancer. A case-control study. Ann Intern Med 118:1–6. Shapiro S, Venet W, Strax P, Venet L, eds. 1988. Periodic Screening for Breast Cancer. The Health Insurance Plan Project and Its Sequelae, 1963–1986. Baltimore: The Johns Hopkins University Press, pp. 1–33. Shinya H, Wolff WI. 1979. Morphology, anatomic distribution and cancer potential of colonic polyps: An analysis of 7000 polyps endoscopically removed. Ann Surg 190:679–683. Simon JB. 1985. Occult blood screening for colorectal carcinoma: A critical review. Gastroenterology 88:820–837. Smith RA, Cokkinides V, Eyre HJ. 2003. American Cancer Society Guidelines for the Early Detection of Cancer, 2003. CA A J Clin 53:27–43. Sobue T, Suzuki T, Naruke T, The Japanese Lung Cancer Screening Research Group. 1992. A case-control study for evaluating lung cancer screening in Japan. Intl J Cancer 50:230–237. Sobue T, Suzuki T, Naruke T. 1992. A case-control study for evaluating lungcancer screening in Japan. Japanese Lung-Cancer-Screening Research Group. Intl J Cancer 50:230–237. Sone S, Li F, Yang ZG, et al. 2001. Results of three-year mass screening programme for lung cancer using mobile low-dose spiral computed tomography scanner. Br J Cancer 84:25–32. Stitik FP, Tockman MS, Khouri NF. 1985. Chest radiology. In: Miller AB, ed. Screening for Cancer. New York: Academic Press, pp. 163–191. Stitik FP, Tockman MS. 1978. Radiographic screening in the early detection of lung cancer. Radiol Clin North Am 16:347–366. Swensen SJ, Jett JR, Sloan JA, et al. 2002. Screening for lung cancer with lowdose spiral computed tomography. Am J Respir Crit Care Med 165:508– 513. Swensen SJ, 2002. CT screening for lung cancer. Am J Roentgenol 179:833– 836. Takeda T. 1989. History and current status of neuroblastoma screening in Japan. Med Pediatr Oncol 17:361–363. Takeuchi LA, Hachitanda Y, Woods WG, et al. 1995. Screening for neuroblastoma in North America: Preliminary results of a pathology review from the Quebec Project. Cancer 76:2363–2371.

1317

Tanaka T, Matsumura T, Iehara T, et al. 2000. Risk of unfavorable character among neuroblastomas detected through mass screening. The Japanese Infantile Neuroblastoma Cooperative Study. Med Pediatr Oncol 35:705– 707. Tedesco JF, Waye JD, Avella JR, et al. 1980. Diagnostic implications of the spatial distribution of colonic mass lesions (polyps and cancers): A prospective colonoscopic study. Gastrointest Endosc 26:95–97. Tockman MS, Levin ML, Frost JK, et al. 1985. Screening and detection of lung cancer. In: Aisner J, ed. Lung Cacner. New York, NY: Churchill Livingstone, pp. 25–40. Towler BP, Irwig L, Glasziou P, et al. 2002. Screening for colorectal cancer using the faecal occult blood test, Hemoccult (Cochrane Review). In: The Cochrane Library, Issue 3 Oxford: Update Software. Traverso G, Shuber A, Levin B, Johnson C, et al. 2002. Detection of APC mutations in fecal DNA of patients with colorectal tumors. N Engl J Med 346:311–320. Tripp MR, Morgan TR, Sampliner RE, et al. 1987. Synchronous neoplasms in patients with diminutive colorectal adenomas. Cancer 60:1599–1603. Tsukada H, Kurita Y, Yokoyama A, et al. 2001. An evaluation of screening for lung cancer in Niigata Prefecture, Japan: A population-based case-control study. Br J Cancer 85:1326–1331. Tuchman M, Auray-Blais C, Ramnaraine ML, et al. 1987. Determination of urinary homovanillic and vanillylmandelic acids from dried filter paper samples: Assessment of potential methods for neuroblastoma screening. Clin Biochem 20:173–177. Tuchman M, Fisher EJ, Heisel MA, et al. 1989. Feasibility study for neonatal neuroblastoma screening in the United States. Med Pediatr Oncol 17:258–264. Weiss NS. 1983. Control definition in case-control studies of the efficacy of screening and diagnostic testing. Am J Epidemiol 118:457–460. Wilde J. 1989. A 10 year follow-up of semi-annual screening for early detection of lung cancer in the Erfurt Country, GDR. Eur Respir J 2:656–662. Williams CM, Greer M. 1963. Homovanillic acid and vannilylmandelic acid in diagnosis of neuroblastoma. JAMA 183:836–840. Winawer S. 2003. Personal Communication. Winawer S, Fletcher R, Rex D, et al. 2003. Colorectal cancer screening and surveillance: Clinical guidelines and rationale. Gastroenterology 124:544–560. Winawer SJ, Gottlieb LS, Stewart ET, et al. 1983. First progress report of the National Polyp Study. Gastroenterology 84:1352. Winawer SJ, Zauber AG, O’Brien MJ, et al. 1992. The National Polyp Study. 1. Design, methods and characteristics of patients with newly diagnosed polyps. The National Polyp Study Workgroup. Cancer 70:1236–1245. Woods WG, Gao RN, Shuster JJ, et al. 2002. Screening of infants and mortality due to neuroblastoma. N Engl J Med 346:1041–1046. Woods WG, Tuchman M, Bernstein ML, et al. 1992. Screening for neuroblastoma in North America. 2-year results from the Quebec Project. Am J Pediatr Hematol Oncol 14:312–319. Woods WG, Tuchman M, Robison LL, et al. 1996. A population-based study of the usefulness of screening for neuroblastoma. Lancet 348:1682–1687. Yamamoto K, Hanada R, Kikuchi A, et al. 1998. Spontaneous regression of localized neuroblastoma detected by mass screening. J Clin Oncol 16:1265–1269. Yamamoto K, Hayashi Y, Hanada R, et al. 1995. Mass screening and agespecific incidence of neuroblastoma in Saitama Prefecture, Japan. J Clin Oncol 13:2033–2038. Yee J, Akerkar GA, Hung RK, et al. 2001. Colorectal neoplasia: Performance characteristics of CT colonography for detection in 300 patients. Radiology 219:685–692. Yoneda A, Oue T, Imura K, et al. 2001. Observation of untreated patients with neuroblastoma detected by mass screening: A “wait and see” pilot study. Med Pediatr Oncol 36:160–162.

71

Cancer Chemoprevention JAYE L. VINER, ERNEST HAWK, AND SCOTT M. LIPPMAN

I

n the past 10 years we have witnessed the acceptance of cancer as a late, non-obligate stage within the long process of carcinogenesis. This recognition has led to tremendous advances in the field of cancer chemoprevention, which focuses on the biology of carcinogenesis and systematic approaches for evaluating agent activity, mechanisms of action, and risk-benefit profiles. Molecular insights into carcinogenesis provide opportunities to intervene in a specific and timely manner to inhibit, delay, or reverse carcinogenesis. Progress in cancer prevention depends upon translational interactions between preclinical and clinical researchers, and has been advanced through clinical trials involving ostensibly healthy participants with intraepithelial neoplasia or an elevated risk for second primary cancer(s). In the wake of these studies, we are increasingly able to characterize, quantify, and modulate cancer risk before the advent of invasive disease and its potential for lethal consequences. This chapter reviews principles of cancer chemoprevention and suggests opportunities for future research.

CARCINOGENESIS—THE TARGET OF PREVENTION Our understanding of carcinogenesis emerged from studies of rodents exposed to relatively high doses of carcinogens (Newberne and Rogers 1973; Reddy et al., 1975). These animal models were highly informative and led to the identification of three discrete stages of carcinogenesis—initiation, promotion, and progression (Balmain et al., 1988; Yuspa et al., 1996). During initiation, a carcinogen interacts with DNA and produces a genetic lesion that might enable the mutant cell to evade endogenous surveillance and repair mechanisms. Once incorporated into the genome, this aberrancy may confer a survival and growth advantage within the local ecosystem (e.g., better adaptation to an external stimulus such as phorbol ester), leading to the second stage of carcinogenesis, promotion. The evolutionary dynamics of promotion involve reversible, genetic changes that result in clonal proliferation beyond physiologic needs and competitive exclusion of neighboring, non-mutated cells. By contrast, progression refers to irreversible, chromosomal aberrations resulting in morphologically distinct cells (Gooderham and Carmichael, 2002). These three stages of carcinogenesis epitomize an environment that is permissive for clonal outgrowth and invasion across the basement membrane, which defines cancer owing to its potential for metastatic spread and lethality. Cumulative molecular derangements give rise to determinants of clonal progression, maintenance, and growth, including self-sufficiency in growth signals; insensitivity to antigrowth signals; apoptotic escape; limitless replicative potential; sustained angiogenesis; and tissue invasion and metastasis (Hanahan and Weinberg, 2000). All cells within an epithelial sheet share genetic complement and exposure to environmental insults (e.g., free radicals, carcinogens, hormones, radiation), as well as a common susceptibility to neoplasia. The multifocal or polychronotropic nature of carcinogenesis—also called field carcinogenesis—explains why neoplastic foci arise in an ostensibly hectic pattern within exposed epithelium (Slaughter et al., 1953). Few neoplastic lesions—and only a subset of cells within these lesions—have the same molecular profile. These foci may be uniclonal or multiclonal and differ in the degree of maturation and/or rate of neoplastic progression. Tumor progression models present a chronologic sequence for various molecular changes, which in most cases is

1318

speculative. Until the fine points of tumor progression are sorted out by functional genomic/proteomic analyses, phenotypic changes at the cellular and tissue levels remain the most reliable measures of carcinogenesis, and their inhibition/reversal is the best measure of an agent’s chemopreventive potential. Over the past 15 years, molecular progression models have been developed for most epithelia, including those of the colon (Vogelstein et al., 1988), head and neck (Bockmuhl et al., 1998), bladder (Baithun et al., 2001), skin (Serewko et al., 2002), lung (Mitsuuchi and Testa, 2002), and cervix (Wolf and Ramirez, 2001). These models map out genotypic and epigenetic events that might be necessary and/or sufficient for carcinogenic progression, as well as their temporal relationships. For example, colon carcinogenesis is commonly promoted by tumor suppressor inactivation (e.g., APC and p53); oncogene activation (e.g., K-ras); altered methylation of CpG islands; impairment of DNA repair (e.g., loss of functional MLH1, MSH2/6, or PMS1/2 activities); and over-expression of key growth-controlling enzymes (e.g., ornithine decarboxylase, cyclooxygenase-2) (Umar and Kunkel, 1996). These molecular mutations may be germline (i.e., inherited and affecting every cell at birth) or somatic (i.e., acquired and limited to the tissue in which they occur). Both mutation types denote cancer risk and may be targets of chemopreventive interventions.

GERMLINE MUTATIONS Genetic and epigenetic alterations in nearly two dozen genes are linked to a predisposition to cancer (American Society of Clinical Oncology, 2003). Germiline mutations of the adenomatous polyposis coli (APC) gene are associated with familial adenomatous polyposis (FAP) and are used for genetic testing in at-risk individuals. (Giardiello et al., 2001). APC mutation subsets enable further risk stratification. For example, mutations in codons 1250–1464 give rise to the classic polyp-dense (>5000) FAP phenotype; mutations in codon 1309 give rise to aggressive disease with a much earlier onset of cancer; and mutations in codons 85–170 give rise to relatively indolent disease. (Nagase et al., 1992). Hereditary nonpolyposis colon cancer (HNPCC) is characterized by a DNA mismatch repair defect, which leads to an extremely high mutation rate within repetitive DNA sequences (called DNA microsatellite instability, or MSI) (Peltomaki, 2003). Loss of mismatch repair capacity is a key early event in HNPCC carcinogenesis and accounts for the high rate of microsatellite instability in HNPCC tumors that occur in the colorectum, endometrium, ovary, stomach, urothelium, hepatobiliary tract, and brain (Lynch and Lynch, 2000). In the breast, germline mutations of BRCA1 or BRCA2 account for up to 10% of all cancers. These mutations confer cumulative lifetime risks as high as 85% for breast cancer and 65% for ovarian cancer (Ford et al., 1998; Lakhani et al., 1998; Gayther et al., 2000). Most germline mutations are hemizygous (i.e., involve only one mutant allele), and therefore genetic instability arises only upon mutation of the other allele (Ionov et al., 1993; Risinger et al., 1993; Umar et al., 1994; Parsons et al., 1995; Umar and Kunkel, 1996). Because they require fewer mutational events for their development, hereditary cancers tend to occur at earlier ages than do sporadic cancers. The young age of onset for certain hereditary cancers has profound implications for genetic testing, cancer screening, surveillance, and the use of preventive interventions.

Cancer Chemoprevention

SOMATIC MUTATIONS Somatically mutated cells provide insights into molecular events that foster carcinogenesis and the clinical and phenotypic sequellae of these events. For example, APC somatic mutations have been described in preneoplastic and neoplastic lesions of the colon (Miyaki et al., 1994; Miyaki et al., 1999), esophagus (Boynton et al., 1992; Huang et al., 1992), stomach (Nakatsuru et al., 1993), and pancreas (Horii et al., 1992). Aberrations of proteins downstream of the APC pathway (e.g., beta catenin) also have been identified in many preneoplastic and neoplastic lesions (Rubinfeld et al., 1993; Polakis et al., 1999). Tumor suppressor genes typically have inactivating mutations, whereas tumor oncogenes typically have activating mutations (e.g., ras genes that regulate guanine triphosphate (GTP) signal transduction). Somatic mutations associated with ras genes, in particular K-ras, have been characterized in a variety of sporadic cancers. Mutations in Kras occur exclusively in hot-spot codons 6, 12, and 13 and have been detected in up to 80% of colorectal, 87% of pancreatic, and 48% of lung cancers, with variations in incidence as a function of detectionmethod sensitivity (Minamoto et al., 2000). K-ras mutations in preneoplastic lesions have led to studies of these mutations’ potential as markers of cancer risk and response to chemopreventives (Ronai et al., 1995; Ronai and Minamoto, 1997). The p53 gene is a powerful regulator of cell growth and division, and more than 50% of all human cancers have been shown to harbor p53 mutations (Levine, 1997). Furthermore, nearly 95% of individuals heterozygous for p53 germline mutations develop cancer and do so at earlier ages than do those without these mutations (Hollstein et al., 1994; 1997). The International Agency for Research on Cancer (IARC) has amassed an extensive repository of p53 mutations that will facilitate exploration of their critical role in carcinogenesis (), and advances in gene-expression profiling may permit rapid and sensitive analysis of the full spectrum of these mutations (Ahrendt et al., 1999). New molecular technologies are revealing susceptibility characteristics, such as polymorphisms of key enzymes. These insights should improve cancer risk stratification and facilitate the identification and testing of preventive interventions targeting individuals with specific molecular defects. High-throughput multi-component colorectal cancer (CRC) screening tests have paved the way for the fecal mutiplex DNA assay, which reliably detects 15 mutational hot spots involving k-ras, APC, p53, BAT26 (a marker for microsatellite instability), and highly amplifiable or “long” DNA (L-DNA) (Ahlquist and Shuber, 2002). In a pilot evaluation involving single-stool testing, the fecal mutiplex DNA assay detected a higher proportion of adenomas that were £1 cm in diameter than did conventional guaiac testing (Ahlquist et al., 2000). This assay is focused on the colorectum, but may simultaneously detect somatic mutations in stomach, pancreas, esophagus, and lung tissue sloughed into the gastrointestinal tract.

MITOCHONDRIAL GENOMIC MUTATIONS Although eukaryotic mitochondrial and nuclear genomes are subject to identical endogenous and exogenous insults, mitochondrial DNA (mtDNA) may be the more sensitive diagnostic marker. Relative to nuclear DNA, the mitochondrial genome has more copy numbers (average of 100–500 copies per cell), is more vulnerable to damage, and has a higher mutation rate. This is largely due to mitochondrial proximity to the site of reactive oxygen species generation (e.g., the electron transport system) (Satoh and Kuroiwa 1991; Kowaltowski and Vercesi 1999), deficiencies in repair/turnover of DNA adducts (Croteau et al., 1999), and lack of chromatin. Even though data on the role of mtDNA in precancerous and cancerous lesions are limited, specific mtDNA aberrations have been seen in human cancers of the breast (Bianchi et al., 1995), colorectum (Polyak et al., 1998), stomach (Habano et al., 2000), kidney (Selvanayagam and Rajaraman, 1996),

1319

bladder, lung, head and neck (Fliss et al., 2000), and thyroid (Yeh et al., 2000). Recent data suggest that mitochondria also play an important role in cell survival and apoptosis, supporting investigations into the utility of mtDNA as an early marker of genetic instability (Kroemer and Reed, 2000). Eukaryotic cells have multiple mitochondria, each of which has multiple mtDNA copies. Theoretically, this genomic amplification could facilitate the detection of mtDNA damage and mutations, pending the development of analytic technologies that overcome certain technical challenges, such as the immense size of the mitochondrial genome (16.5 kb). In all likelihood, advances in microarray and high-throughput technology (e.g., ligation-mediated polymerase chain reaction) will enable further exploration of mtDNA as a promising marker of cancer risk.

BIOMARKERS Systematic and timely progress in preventing cancer depends largely on the identification, development, and validation of accurate, reliable, and early biomarkers of carcinogenesis (Hawk et al., 2000). Biomarkers are structural or functional events that provide insights into pathology that occurs higher and later in the biologic hierarchy. Structural biomarkers include aberrations in cellular number, morphology, and/or integrity (e.g., atypia, hyperplasia, or dysplasia). These aberrations may be quantitative or qualitative (i.e., inappropriate function relative to healthy biologic need) and do not necessarily correlate with functional deviations. Although functional biomarkers may be difficult to measure precisely, they are more clinically relevant than are structural biomarkers. As a result, functional derangements are often deduced from structural abnormalities. For example, loss of p53 function is commonly inferred from immunohistochemical overexpression. Molecular epidemiology is exposing promising biomarkers for detecting and stratifying cancer risk. Well-characterized and reasonably validated biomarkers of carcinogenesis emerging from these studies may ultimately serve as important measures of response to interventions, particularly in clinical research wherein cancer end points may be infeasible because of a long latency to invasive disease. Promising candidate biomarkers typically come from mechanistic or correlative insights from in vitro, in vivo, observational, or clinical research. Biomarkers may be categorized according to the intended application: risk markers estimate the probability of some later event in the neoplastic process, commonly referred to as exposure, susceptibility, diagnostic, or prognostic biomarkers, depending on the context; and response markers measure changes associated with a specific intervention. The ideal response biomarker fulfills most if not all of the following criteria: differential expression in normal vs. in highrisk tissue, causal association with cancer, modulation by the applied intervention, association with a relatively short latency to cancer development, and correlation with cancer incidence. In addition, such biomarkers should be amenable to reliable, ready, and quantitative assessment (Boone and Kelloff, 1997; Einspahr et al., 1997; Tockman, 2001). Markers of carcinogenesis have the potential to propagate within and beyond higher biologic levels (e.g., DNA to RNA, RNA to protein, protein to cell, and cell to tissue). Although molecular markers contribute mechanistic information, tissue-level markers are considered to be far more decisive indicators of cancer risk because they integrate individual molecular markers and more faithfully predict invasive cancer. Therefore, suggestive data from tissue markers may justify the use of interventions that reduce cancer risk, whereas suggestive data from lower-level markers may not. The Bcr-Abl oncogene in chronic myelogenous leukemia is the rare example of a molecular defect that is pathognomonic of disease and determines clinical management (Goldman and Melo, 2003). Far more commonly, multiple molecular defects initiate, maintain, and promote carcinogenesis, and no single defect reliably predicts cancer development (Bertram, 2000). Consequently, prevention trials typically integrate multiple response biomarkers across several levels of biologic organization (e.g., APC

1320

PART V: CANCER PREVENTION AND CONTROL

mutations, genomic expression, and cyclooxygenase (COX) activity at the molecular level and proliferation, apoptosis, aberrant crypt foci (ACF) number/morphology, and adenoma regression/prevention at cellular/tissue levels) to query and confirm neoplastic determinants and chemopreventive mechanisms.

BIOMARKER DEVELOPMENT All cells within an epithelial sheet inherit a common susceptibility to neoplasia. Variable exposures to genetic and epigenetic insults, however, result in non-uniform distribution and rates of neoplastic progression, particularly at the more complex biologic levels. Predictably, patchy distribution of neoplastic foci increases the potential for sampling errors and the noise level of most measurements (Hawk et al., 2000). To improve the accuracy and reliability of tissue-based biomarkers, several precautionary steps are useful. First, tissue acquisition, sample handling/transportation, and laboratory assays for specimen analysis should be standardized. This is achieved through careful selection of biopsy sites, comprehensive training of clinical/lab personnel in study procedures, and laser-capture micro-dissection to reduce the signal-to-noise ratio. In addition, expression microarrays and proteomic analyses now allow for concomitant assessments of multiple molecular markers, which were prohibitively complicated in the past by time and tissue-utilization constraints (Macgregor, 2003; Wulfkuhle et al., 2003). Next, candidate biomarkers are specifically studied to discern their natural history and how they correlate with cancer risk crosssectionally and over time (Hawk et al., 2000). Often—particularly in the case of rare neoplastic lesions—this study evolves out of retrospective comparisons of affected individuals with matched control populations. The natural history of a promising marker also may be derived from prospective studies of risk cohorts or from the placebo arm of intervention studies that incorporate biomarkers as secondary end points. Risk biomarkers that generally correlate with the development of invasive cancer may then be used to: estimate risk for incident neoplasia; monitor disease progression; modify interventions to optimize the therapeutic index (TI) (i.e., the anticipated benefit-to-risk ratio of a test compound in the context of a certain disease state); prioritize interventions based upon mechanistic insights; or stratify individuals enrolling in cancer prevention trials to balance significant covariates and potential confounders. Biomarkers that correlate directly with a later event (e.g., cancer development or mortality) are highly sought after, but those that fall short of this level of correlation may prove valuable as well. Candidate response biomarkers with demonstrated biologic relevance (often related to cancer risk) are first evaluated in pilot studies to show modulation by an intervention and then in short-term to intermediate-term trials that incorporate placebo controls. The placebo arm confirms the marker’s natural history and association with cancer risk, and the treatment arm provides preliminary evidence of agent efficacy. Once a response biomarker is shown to be predictive of a later event of primary interest, it may substitute for that later event (i.e., become a surrogate endpoint biomarker (SEB)) (Temple, 1999; Hawk et al., 2000). In principle, a biomarker’s internal validity applies only to the cohort in which it was tested. Nevertheless, external validity may eventually be established with regard to other interventions and risk cohorts. In organs that are readily accessible to biopsy (e.g., the skin, bladder, esophagus, colorectum, or cervix), certain preinvasive neoplastic lesions have been reasonably validated as cancer precursors (e.g., actinic keratosis, superficial bladder neoplasia, high-grade esophageal dysplasia, colorectal adenomas, and cervical intraepithelial neoplasia) and have become accepted as targets for surgical/medical interventions in clinical practice. Indeed, the current standard of care is to intervene against dysplastic lesions with incisional biopsy for diagnosis and risk reduction, followed by serial surveillance, focal resection, or even organ extirpation, depending on the malignant potential of the lesion and patient history.

SURROGATE END POINT BIOMARKERS Surrogate end point biomarkers are intermediate phenomena that occur between early and late events of interest and are commonly used in medical research and the management of chronic disease (e.g., hypertension, diabetes, HIV, and cancer). Certain SEBs (e.g., hypertension, hyperglycemia, and certain lipid profiles (e.g., elevated LDL, reduced HDL)), are key contributors to the genesis of their associated disease(s). In such cases, interventions that target the SEB/risk factor have yielded irrefutable clinical benefits against downstream disease (Mandel et al., 1993; Winawer et al., 1993; Ross et al., 1999; Mandel et al., 2000; Trenkwalder 2000). The extent to which, and under what conditions, a biomarker may substitute for a clinical end point is under intense research. Moreover, the demarcation between an SEB and later, related events may be murky, and certain established SEBs (e.g., hypertension) are increasingly regarded as asymptomatic diseases in and of themselves. In the realm of cancer prevention, considerable debate persists as to the absolute risk associated with carcinogenesis identified at the tissue (e.g., dysplasia), cellular (e.g., atypia), and particularly at the molecular (e.g., genetic mutation) level (O’Shaughnessy et al., 2002; Levin 2003). Nevertheless, the potential reversibility of clinically relevant carcinogenesis remains the central premise of chemoprevention research. Reasonably validated SEBs or markers of clinical benefit, such as the preinvasive neoplastic lesions presented in Table 71–1, facilitate preliminary efficacy testing of chemopreventive agents. Biomarkers such as colorectal adenomas, ductal carcinoma in situ (breast), bronchial dysplasia (lung), prostatic intra-epithelial neoplasia (PIN), cervical intraepithelial neoplasia (CIN), actinic keratosis and dysplastic nevi (skin), and oral leukoplakia (oropharynx) offer significant efficiencies for trial design and execution in terms of reduced sample size and time to completion—albeit with some loss of predictive certainty regarding ultimate clinical outcomes. Regulatory approvals based on results from trials using biomarker end points have boosted enthusi-

Table 71–1. Preinvasive Neoplasia as Markers of Cancer Risk and the Focus of Clinical/Preventive Efficacy Valid Risk Marker

Focus of Standard Surgical Intervention*

Ductal carcinoma in situ (DCIS) Atypical ductal hyperplasia (ADH) Cervical intraepithelial neoplasia (CIN) Oral leukoplakia Esophageal dysplasia (high grade)

✓

✓

Preinvasive Neoplastic Lesion

✓

Valid Focus of Medical Intervention† Oral Tamoxifen Oral Tamoxifen

✓

✓

✓ ✓

✓ ✓

Colonic adenoma Bladder dysplasia

✓ ✓

✓ ✓

Actinic keratosis (AK)

✓

✓

Oral photophrin with photodynamic therapy Oral Celecoxib Instillational therapy with BCG or Valrubicin Topical treatment with 5-Fluorouracil; Diclofenac sodium; or Aminolevulinic Acid with Photodynamic Therapy

Source: Adapted from Viner et al. 2002. *Most of these dysplastic lesions are surgically removed upon identification, although only a few have been validated to demonstrate that doing so reduces the risk for subsequent cancers or cancer-associated mortality. † Validation of these lesions as predictors of reduced cancer incidence or mortality is only indirect in most instances, but the predictive association between these lesions and cancer, or their current management according to prevailing standards of care, has been compelling enough to justify FDA approval of drugs targeting them in well-defined clinical cohorts. Reductions in subsequent cancer incidence or mortality have not been proven.

Cancer Chemoprevention asm for and pharmaceutical investment in research that simultaneously evaluates certain biomarkers of cancer risk as surrogates for cancer incidence and as targets for chemopreventive interventions. This strategy has been amply demonstrated in the study of preinvasive lesions of the colorectum and breast. For example, the 1993 National Polyp Study indirectly validated colorectal adenomas as SEBs for colorectal cancer (CRC) in the context of surgical interventions. This study showed up to 90% reductions in CRC incidence among subjects undergoing regular surveillance for adenomas (compared with incidences in three historical control groups) (Winawer et al., 1993). Observational and experimental studies confirmed significant reductions in CRC mortality and, in some cases, incidence in the context of CRC screening by stool blood-based tests or flexible sigmoidoscopy combined with colonoscopic adenomectomy (Mandel et al., 2000). Finally, adenomas have attained some measure of validation as SEBs in the context of medical interventions in dozens of observational studies indicating nonsteroidal anti-inflammatory drug (NSAID) efficacy in reducing the risk of colorectal adenomas, carcinomas, and cancerassociated mortality (Hawk et al., 2003). The National Surgical Adjuvant Breast and Bowel Project (NSABP) P-1 and B-24 trials lent credibility to the use of preinvasive neoplastic lesions as SEBs for breast cancer (Fisher et al., 1998; 1999). In these trials, women randomized to tamoxifen experienced reductions in ductal carcinoma in situ (DCIS) and/or invasive breast cancer. Proportional decreases in DCIS and invasive cancer justified consideration of DCIS as an SEB for breast cancer. In response to data derived from the NSABP P-1 trial, the US Food and Drug Administration (FDA) approved the use of tamoxifen for risk reduction (i.e., prevention) in women at high risk for breast cancer. Although tamoxifen was the first agent specifically approved for cancer risk reduction, other agents, such as intravesical Bacillus Calmette-Guerin (BCG) and topical fluorouracil, were approved decades earlier for the management of pervasive neoplasias of the bladder and skin, respectively (). Cancer prevention research is increasingly focused on the identification and validation of molecular/genetic biomarkers and biomarker clusters, which are now tested with increasing accuracy and speed via novel DNA, RNA (Macgregor, 2003), and protein assessment technologies (Wulfkuhle et al., 2003). Tissue histopathology and cellular proliferation and apoptosis are relatively robust biomarkers that serve as end points for iterative investigations of multiplex marker panels. While this process and technology are still evolving, biomarker investigations are clearly committed to the identification and development of early, molecular predictors and measures. These investigations are expected to discover genomic patterns that are relevant to cancer risk assessment, cohort identification, and response determinations and that also may serve as targets for preventive interventions in individuals carrying signature mutations. Biomarkers for cancer risk and response assessments are expected to have a profound impact on public health, just as using lipid profiles and hypertension as biomarkers of risk and response has expedited key advances in cardiovascular medicine over the past decade (Temple, 1999).

1321

here expedites iterative testing in progressively larger, lower-risk cohorts. This strategy conserves precious time and resources, while balancing potential toxicities against the potential benefits in each risk group.

CHEMOPREVENTIVE AGENTS Agent Prioritization for Cancer Prevention Several essential criteria have improved our selection of molecular targets for cancer prevention: under-/over-expression of aberrant protein in early human neoplasia relative to unaffected epithelium in the tissue of origin; unequivocal biologic contribution from the aberrant protein (e.g., as demonstrated by genetically manipulated mouse models); pharmacologic accessibility; modulation that correlates with reductions in cancer incidence; and specificity of action within neoplasia rather than normal tissue. Chemopreventive agent development is sequential, starting with in vitro and in vivo cell-based mechanistic assays and efficacy screening tests, followed by in vivo screening in animal carcinogenesis models using cancers and preneoplastic lesions as end points (Kelloff and Sigman, 2002). The most promising agents are then further characterized by dose-response and dosing regimens in animal carcinogenesis models, alone and in combination with other test agents. Target assays are an important step in most prioritization schemes for chemopreventive drug development. Even so, activity against a hypothesized molecular target is neither necessary nor sufficient justification for additional agent testing. Schemes to prioritize promising agents for advanced testing include evaluating data from mechanistic assays (i.e., in vitro and in vivo testing), toxicology and pharmacokinetic studies, and human trials. The cumulative weight of these data, in conjunction with benefit-to-risk assessments in the intended cohort, may lead to agent approval for cancer prevention.

Mechanistic and Preclinical Studies Current chemoprevention research is focused on translational evaluations, and advances through bidirectional hypothesis generation and testing that swiftly applies laboratory findings in the clinic and vice versa. Lead compounds have been variously identified through: preclinical efficacy testing of promising agents with distinct molecular targets (e.g., inhibition of ras farnesylation, epidermal growth factor receptor (EGFR), ornithine decarboxylase, matrix metalloproteinase, steroid aromatase, and oncoproteins or induction of anti-mutagenesis, apoptosis and anti-oxidation); epidemiologic data on drugs developed for other applications (e.g., oncologic, endocrinologic, rheumatologic, and gastroenterologic); and epidemiologic data on dietary components. Although traditional testing is still needed, these mechanistically based, preclinical approaches may expedite agent development (Umar et al., 2001).

In Vitro Testing CANCER-SUSCEPTIBLE COHORTS Some individuals are at an increased risk for cancer due to highly penetrant susceptibility genes (e.g., APC, MLH-1, MSH-2, BRCA1/2) (Marsh and Zori, 2002) or certain environmental exposures (e.g., tobacco, radiation, exogenous hormones, and hepatitis B, hepatitis C, Epstein Barr, and human papilloma viruses). These individuals are more likely to develop cancer than are persons at average risk, and typically do so at an accelerated rate. In addition, these high-risk cohorts are prone to multicentric neoplasia and, in paired organs, bilateral disease. Clinical trials conducted in such high-risk populations may be adequately powered with relatively small cohorts followed for short periods of time, owing to high rates of malignant transformation. Agents confirmed as safe and effective in these accelerated disease models may then be tested for their generalizability to progressively lower-risk cohorts that share a common molecular pathogenesis with the higher-risk cohorts. The approach to agent development described

Preliminary agent assessments are conducted through a battery of mechanistic assays that establish a particular agent’s (or agent class’s) range of chemopreventive activities (e.g., inhibition of proliferation, mutagenicity, angiogenesis, and inflammation; and induction of apoptosis) (Steele et al., 1998). Cell-based assays are usually the first efficacy screen for candidate chemopreventives. The criteria for evaluating in vitro tests include sensitivity and ease of quantitation, relevance to the organ of interest, use of epithelial cells, use of human cells (if possible), appropriate controls (negative and/or positive), and cost and time efficiency (Kelloff and Sigman, 2002). In each assay, agents are tested across a broad range of concentrations to determine the concentration that inhibits cell growth by 50% (i.e., the inhibitory concentration (IC)50). Novel cell/organ cultures now facilitate preliminary assessments of chemopreventive efficacy through studies of stromal-epithelial interactions (e.g., raft cultures), cultured cells from transgenic mice, and human cells known to carry cancer-predisposing genes (e.g., p53 mutations characterizing Li-Fraumeni syndrome) (Steele et al., 1998).

1322

PART V: CANCER PREVENTION AND CONTROL

In Vivo Testing Agents with evidence of chemopreventive efficacy from in vitro testing may advance into in vivo testing, which primarily is conducted in rodent models specific for the organ of interest. Carcinogen- and genetically-induced cancer models enable the analysis of genetic and environmental factors that lead to neoplasia, evaluation of chemopreventive potential, and may guide biomarker selection. Carcinogens and/or genetic lesions that reproducibly induce well-characterized tumors facilitate crude assessments of agent efficacy. Ideally, in vivo models approximate multistage genetic, cellular, and histopathologic changes that characterize human carcinogenesis; however, most models fall short of this ideal. In the case of carcinogen-induced tumor models, most commonly used toxins have no known role in human disease, severely limiting generalization to human carcinogenesis. Tumor inhibition is evaluated by several means, such as by measuring relative reductions in percent incidence, multiplicity, or tumor burden or by measuring increases in the latency to tumor development. Successful tumor induction is usually gauged by microscopic evaluation of the target tissues or gross pathology, depending on the target. Proven reductions in tumor incidence, multiplicity, and/or overall burden in carcinogen-induced, transgenic, or spontaneous animal cancer models may justify proceeding to human testing.

Chemical Carcinogenesis Mouse Models In classic, carcinogen-induced cancer models, a carcinogen is administered to mice at a dose that induces an adequate number and size of tumors in the tissue of interest within an acceptably short time (Yuspa and Poirier, 1988). Some of the most commonly used carcinogeninduced tumor models are azoxymethane (AOM)-induced rat colon tumors; dimethylbenz[a]anthracene (DMBA)- and N-methyl-Nnitrosourea (MNU)-induced rat mammary tumors; benz(a)pyrene (B(a)P)-induced rat lung tumors; and N-butyl-N-(4-hydroxybutyl) nitrosamine (OH-BBN)-induced mouse urinary bladder tumors. Inhibitory test agents may be administered before, simultaneously with, or after the carcinogen (or any combination thereof); and the sequence of administration enables the assessment of inhibitory mechanisms.

Transgenic and Knock-out Mouse Models of Carcinogenesis Novel mouse strains over-express or have inactivated cancer genes and provide compelling models for evaluating the chemopreventive potential of promising agents and agent combinations; for a more comprehensive review, see articles by Lipkin or Hursting (Hursting, 1997; Lipkin et al., 1999). Such genetically modified models may contain randomly inserted mutated oncogenes, tumor suppressor genes, or targeted mutations that lead to an increased incidence of preneoplastic and neoplastic lesions. Transgenic and gene knock-out (i.e., genedeleted) mice may mimic human carcinogenesis more faithfully than do the chemically induced models (Kavanaugh et al., 2002). The most widely used genetically altered mouse models in chemoprevention are multiple intestinal neoplasia (Min) mice that carry a mutation in the adenomatous polyposis coli (APC) gene, APCknockout mice, mutator (MMR deficient) mice, and p53- knockout mice. The Min mouse effectively simulates the clinical condition associated with APC mutations in human FAP and is perhaps the most important animal model used for cancer chemoprevention. The Min mouse and other models with APC gene mutations furnish dynamic systems in which to test the modulation of colorectal adenoma formation by chemopreventive compounds. Another exciting model is the HPV-infected (K14-HPV16 heterozygote), estradiol-treated mouse model, in which cervical intraepithelial neoplasia (CIN)-like lesions progress to squamous carcinoma of the cervix, potentially providing insight into another epithelial cancer. Preclinical studies with Min and APC-knockout mice, coupled with data from the chemically induced carcinogenesis rat and mouse models, provide the strongest evidence to date that the enzyme cyclooxygenase-2 (COX-2) plays a major role in colon carcinogenesis, and that nonsteroidal anti-inflammatory drugs (NSAIDs) targeting COX-2 have great potential in preventing colon cancer. In addition,

chemoprevention studies in mice lacking the p53 tumor-suppressor gene, the most commonly altered gene in human cancer, suggest that increased susceptibility to cancer resulting from the loss of p53 function may be offset by chemopreventive approaches. Several strains of mismatch repair knock-out mice have been developed that allow genetic-phenotypic correlations with human HNPCC (e.g., MSH1, MLH1) (Wei et al., 2002). Models such as these elucidate the functional role of genetic variants and provide opportunities to test promising chemopreventives.

TOXICOLOGY AND PHARMACOLOGY The FDA mandates stringent preclinical toxicity and safety/pharmacokinetic testing before approving efficacy trials in humans (Kelloff et al., 1995). Preclinical safety assessments of chemopreventive agents include acute and sub-chronic toxicity, reproductive, and genotoxicity studies. These studies typically involve single-dose, acute toxicity, and absorption-elimination studies in rats and sub-chronic, repeated daily-dosing studies in rodents and dogs. Single and combined agents are tested in the animal species that most closely mimics human metabolism to determine pharmacokinetic, toxicity, enzyme, and other potential interactions in at least one study of appropriate duration. Correlations between animal and human serum concentrations guide dosing in early-phase human studies by suggesting target concentrations that might be necessary to achieve the desired pharmacodynamic effects.

TOXICITY AND SAFETY STANDARDS FOR CHEMOPREVENTION In the context of cancer treatment, iatrogenically induced toxicities are routinely tolerated in pursuit of a cure. By contrast, cancer chemopreventives are often intended for chronic use in ostensibly healthy subjects who may never actually develop cancer. As a result, preventive agents—for cardiovascular, infectious or neoplastic indications— are held to safety standards that typically exceed those of agents applied with therapeutic/curative intent. To attain the high standards set for cancer chemoprevention, several tactics have been explored, including enhanced mechanistic specificity, combination regimens with additive or synergistic efficacy, topical administration to minimize systemic exposure, downward dose titration to establish the minimally effective dose, and prophylaxis against worrisome toxicities (Hawk et al., 2000).

ENHANCED MECHANISTIC SPECIFICITY Enhanced mechanistic specificity has the potential to improve the therapeutic index (TI) of promising agents for cancer prevention, as demonstrated by the development of NSAID derivatives (e.g., COX-2-selective inhibitors and nitric oxide-releasing NSAIDs (NONSAIDs)) for chemoprevention. Nonselective NSAIDs reduce cellular proliferation, slow cell-cycle progression, and stimulate apoptosis (Hawk et al., 2003). Their effects are mediated through COX inhibition (i.e., through conversion of arachidonic acid into bioactive prostanoids, which are considered to be important promoters of carcinogenesis), and through COX-independent pathways (e.g., 15-lipoxygenase-1 (15-LOX-1)) (Shureiqi et al., 2003). The chemopreventive potential of this diverse class of agents has been established over the past two decades in a highly consistent body of epidemiologic, preclinical, and clinical data. Gastrointestinal toxicities largely attributable to COX-1 inhibition initially limited the development of NSAIDs for cancer prevention. COX-2-selective inhibitors (e.g., celecoxib and rofecoxib) have an improved TI in comparison with nonselective NSAIDs, thus allaying early safety concerns. Three complementary lines of evidence have established COX-2 as a relevant target for cancer prevention:

Cancer Chemoprevention 1. COX inhibitors and COX-2 gene deletions inhibit intestinal carcinogenesis in vivo; 2. Nonselective COX inhibitors reduce colorectal adenoma/cancer incidence and cancer-associated mortality in human observational studies; and 3. COX inhibitors reverse pre-cancerous lesions, such as aberrant crypt foci (ACF) and adenomas, in individuals with genetic as well as sporadic colorectal neoplasia. In 1999, the COX-2–selective inhibitor celecoxib was shown to reduce the number of colorectal adenomas in persons with the highrisk genetic condition known as FAP (Steinbach et al., 2000). Data generated by this 83-patient Phase II study, coupled with other lines of evidence, persuaded the US FDA to grant provisional approval of celecoxib in conjunction with standard surveillance and surgical prophylaxis for FAP patients. Since this approval, experimentation with NSAIDs and NSAID derivatives for cancer prevention has expanded to the point where this class of agents—alone and in various combinations—is being aggressively developed to prevent an array of epithelial (e.g., colorectal, esophageal, skin, bladder, breast, and prostate) and nonepithelial (monoclonal gammopathy of undetermined significance (MGUS)) neoplasias.

AGENT COMBINATIONS Co-administration of chemopreventive agents may improve the TI by allowing dose reductions of one or more agents within the combination. Additive/synergistic activity has been demonstrated by combining agents that have different mechanisms of action. Combining difluoromethylornithine (DFMO) (Nigro et al., 1986; Reddy et al., 1990; Rao et al., 1991; Li et al., 1999; Jacoby et al., 2000), an epidermal growth factor receptor (EGFR) inhibitor (Torrance et al., 2000), or a matrix metalloproteinase inhibitor (Wagenaar-Miller et al., 2003) with various NSAIDs has reduced colonic tumors in rats and mice. Impressive reductions in rat mammary carcinogenesis have also been achieved with retinold/anti-estrogen combinations (Anzano et al., 1996). Combination approaches are currently being tested in several phase II chemoprevention trials (e.g., DFMO plus sulindac in persons at risk for CRC, and fenretinide plus tamoxifen in women at risk for breast cancer).

TOPICAL DELIVERY The most successful topical delivery system for cancer prevention has been in the skin, where fluorouracil (5-FU) (Gupta 2002), aminolevulinic acid HCL (Levulan Kerastick) with photodynamic therapy (Lang et al., 2001; Jeffes, 2002), and sodium diclofenac (Solaraze) (Peters and Foster, 1999; Wolf et al., 2001; del Rosso, 2003; Jarvis and Figgitt, 2003; Peterson and Goldberg, 2003) have received FDA approval for regression of the nonmelanoma skin cancer precursor, actinic keratosis (AK). Other effective or promising modes of direct delivery to reduce systemic side effects include rinses in the oropharynx (Mulshine et al., 2003), aerosols in the lung (Wang et al., 2000; Wattenberg et al., 2000), suppositories (Hirota et al., 1996) or enemas (Winde et al., 1997) in the colorectum, and instillations into the bladder (Sylvester et al., 2002). Indeed, chemotherapeutic agents have been regionally applied with preventive intent. This was demonstrated by valrubicin (Valstar), a potent anthracycline derivative that was FDA approved in 1998 as intravesical therapy for patients with BCG-refractory CIS, who are not candidates for cystectomy. () Agents developed for topical delivery must be systematically tested to guarantee appropriate partitioning within the intended target. For example, sodium diclofenac was developed through serial compounding of the NSAID diclofenac in hyaluronan, a glycosaminoglycan delivery vehicle. The approved product is a 3% diclofenac: 2.5% hyaluronic acid gel formulation that perfuses diclofenac into the epidermis, where it is depot-released into the basement membrane (Brown et al., 1995).

1323

DOSE DE-ESCALATION Dose de-escalation schemes may enable us to realize the public health potential of chemoprevention. Although high-risk individuals (e.g., people with the highly penetrant germline diseases FAP, HNPCC, basal cell nevus syndrome, or breast/ovarian syndromes) may tolerate greater toxicity to control disease, preserve organs, increase surveillance intervals/latency to disease, and/or reduce anxiety, lower-risk individuals may not. As a result, dose de-escalation may identify dosing regimens that are more acceptable to lower-risk individuals, who collectively account for the largest burden of disease in the population. Other innovative developments, such as pharmacogenomics, may allow customization of preventive agent(s) and doses better suited to each risk cohort, thereby optimizing the TI (Workman, 2002). Opportunities to explore the potential of de-escalation strategies abound, but remain under-explored. Patients with CRC often harbor co-occurring adenomas (Green et al., 2002) and provide an efficient means to assess the chemopreventive potential of interventions applied with therapeutic intent (Umar et al., 2001). For example, 5-FU at high doses for relatively brief periods has been a major component of CRC chemotherapy regimens for more than 30 years. Despite incidental clinical reports of adenoma regression (Hawk et al., 1999) and preventive efficacy in animal models (Narisawa et al., 1980; Narisawa et al., 1980), the clinical potential of 5-FU for cancer prevention—particularly at lower oral doses for longer periods of time—remains an intriguing hypothesis that has yet to be tested.

DESIGNS FOR PHASE I, II, AND III CANCER PREVENTION TRIALS The efficacy and toxicity of candidate preventive agents are typically assessed in a series of clinical trials referred to as phase I, II, or III studies (Kelloff et al., 1995), as presented in Table 71–2. Phase I studies are designed to characterize clinical safety and agent pharmacokinetics and to quickly evaluate an acceptable dose or dose range in a small number of healthy volunteers. These participants are typically observed for a period of time (usually 1–3 months) that is dictated by known toxicity data. Longer phase I trials (up to 12 months) typically recruit individuals at increased risk for cancer to assess the toxic potential of chronic exposure to the study agent. In these highrisk cohorts, the potential toxicity from an incompletely characterized agent is offset by the possibility of clinical benefit. Phase I trials typically enroll fewer than 100 subjects and may include preliminary assessments of agent efficacy with regard to certain neoplastic and/or biomarker end points. Pharmacokinetic data guide the development of phase I clinical dose-escalation strategies. In phase I testing, the highest dose initially tested is one-tenth of the highest dose that induced adverse effects in rodents (the so-called “no observed adverse effect level,” NOAEL) or one-sixth of the highest NOAEL in non-rodent animals exposed to the agent for at least as long as the proposed trial period. Human doses may exceed those suggested by the NOAEL, however, when suggested by differences between human and animal drug metabolism or dosing data derived from clinical sources (Kelloff and Sigman, 2002). Although interspecies comparisons of acute and chronic drug pharmacokinetics/dynamics are informative, such comparisons are not always feasible. Therefore, dose selection for phase I studies is often empiric. Candidate chemopreventive agents meeting regulatory safety standards for other indications (e.g., NSAIDs for analgesia and statins to lower serum lipids) may bypass phase I testing altogether and go directly into Phase II trials. Phase II trials generate preliminary, short-term efficacy data and confirm safety data in the context of a specific neoplastic condition, such as colorectal dysplasia. Phase IIa trials establish preliminary drug effects in an uncontrolled manner, whereas phase IIb trials compare the experimental agent to a placebo or the standard of care against scientifically or clinically justified biomarkers of neoplasia. Phase II cancer prevention trials can have adequate power with as few as 100

1324

PART V: CANCER PREVENTION AND CONTROL

Table 71–2. Characteristics of Clinical Chemoprevention Agent Development Trials Phase

Agent Dosing

Duration

Sample Size and Allocation

Control Group

I

Escalation

Weeks–months

Occasionally

Months Months–year

<25; nonrandomized or randomized <50; nonrandomized <100–200; randomized

IIa IIb

De-escalation Stable

III

Stable

Years

100–≥1,000; randomized

Standard care*

IV

Stable

Unspecified

General post-marketing population

N/A

Never Standard care*

Goals Pharmacokinetics; dose finding based upon short-term, mild to moderate toxicity Dose finding based on reliable biomarker modulation Biomarker modulation (e.g., dysplasia regression) vs. standard care* Definitive efficacy to complement or replace standard care (e.g., reduce dysplasia/cancer incidence) Long-term safety in target population

Source: Adapted from Viner et al. 2002. *In cancer chemoprevention, placebo may represent the standard of care.

subjects observed for several months. By contrast, phase III trials typically involve hundreds to thousands of subjects observed for years. Phase III studies usually randomize participants to a preventive agent/agent combination vs. placebo or the standard of care to reduce clinically relevant neoplasia. With regard to neoplasia, interpretations of clinical relevance may differ, depending on the organ. For example, in many organs preinvasive neoplastic lesions are already the target of clinical intervention (e.g., breast, colorectum, skin, esophagus, and bladder). In other less accessible organs, demonstration of cancer incidence reductions is still required (e.g., ovary, pancreas, liver, kidney).

CLINICAL TRIALS Mechanistic, observational, and experimental data govern the rational development of agents for cancer prevention, and clinical testing remains the rate-limiting step in this investigational process. Human trials impose enormous costs in terms of financing, patient resources, and time. As promising agents/agent combinations for cancer prevention emerge in greater numbers and variety, prioritization for human testing has become even more complex and crucial. An algorithmic approach that compares promising agents/agent combinations based on the strength and consistency of efficacy and toxicity data would ensure that lead compounds meet critical standards for advancedphase clinical testing (Kelloff et al., 1995). Another promising strategy involves nesting prevention end points within cancer treatment trials or within trials testing candidate agents for non-oncologic indications (Hawk et al., 2000). Such approaches might expedite the identification of agents within an agent class (as well as acceptable doses, duration, frequency, and/or route of delivery) that have shown promise for focused cancer prevention research. Regulatory approval of two agents for cancer prevention has solidified the conceptual framework for future studies in high-risk cancer cohorts (Fisher et al., 1998; Steinbach et al., 2000). The following section presents the status, challenges, and clinical efficacy of cancer chemopreventive agents, arranged by target organ (Table 71–3).

oral leukoplakia into normal-appearing mucosa (Hong et al., 1986). Although responses in this trial were neither durable nor complete, they proved the potential value of medical interventions against early neoplasia. Several different retinoid doses, schedules, and formulations have since been evaluated, but dose-limiting toxicities have impeded the development of this agent class for cancer prevention. Other agents of interest for head and neck cancer prevention include NSAIDs (systemic and topical formulations), epidermal growth factor receptor (EGFR) inhibitors, and a pro-apoptotic adenovirus (i.e., ONYX-015). Head and neck cancer is associated with up to a 40% incidence of new, or second, primary cancers, many of which develop in ostensibly normal mucosa within the field at risk. Medical interventions are expected to reduce the rate at which these second primary cancers occur by treating the entire field at risk for neoplasia. This principle has been demonstrated in a placebo-controlled adjuvant therapy trial of high-dose 13cRA in patients at high risk for recurrence after complete resection of stage I–IV primary squamous cell carcinoma of the head and neck (Hong et al., 1990). In that study, 13cRA reduced the number of new primaries but did not affect tumor recurrence rates (vs. placebo). This high-dose adjuvant trial led to a large-scale phase III trial of a lower, less-toxic dose of 13cRA designed primarily to prevent second primary cancers in 1191 randomized early-stage head and neck cancer patients (Khuri et al., 2003). No significant difference occurred between the 13cRA and placebo arms with regard to second cancers or secondary end points of overall or recurrence-free survival. The European Study on Chemoprevention with Vitamin A and NAcetylcysteine (EUROSCAN) also found no retinoid effect on second primary cancers (van Zandwijk et al., 2000). EUROSCAN involved 2592 head and neck or lung cancer patients—almost all former or current smokers—who were randomized to Vitamin A or N-Acetylcysteine (an anti-oxidant) vs. placebo for 2 years. Neither agent produced a statistically significant difference in the rate of second primary cancers when compared with placebo.

LUNG CANCER PREVENTION HEAD AND NECK CANCER PREVENTION Head and neck cancers develop in the upper aerodigestive tract, which encompasses the mouth, throat, sinuses, and nasal cavity. These cancers are typically associated with heavy smoking and/or alcohol consumption. Mucosa of the head and neck is easily examined, which facilitates the detection of early neoplasia, evaluation of agent efficacy, and serial tissue sampling. A common precursor of oral cancer is leukoplakia (literally “white plaque”), which has malignant transformation rates as high as 36%, depending on the degree of dysplasia. Oral leukoplakia provides an in vivo target for trials of promising preventive interventions. Retinoids are the most extensively studied class of agents for the prevention of cancers of the upper aerodigestive tract (Lippman et al., 1994). The first positive cancer prevention trial ever published demonstrated that high-dose 13-cis-retinoic acid (13cRA) administered to subjects for 3 months dramatically transformed

Lung cancer is the leading cause of cancer death in developed countries. Smoking is the major preventable risk factor, accounting for up to 90% of all lung cancers. The observation and acquisition of lung tissue for characterization of neoplastic mutations and molecular changes require invasive procedures. Therefore, preinvasive neoplasia has been typically detected through random bronchial sampling for metaplasia or dysplasia, or through cytologic atypia in sputa. These screening methods provide limited insights into disease burden, as shown by a study in which only half of the screened heavy smokers even met inclusion criteria for bronchial dysplasia. Using bronchoscopic screening to identify study subjects that meet eligibility criteria for dysplasia potentially slows study progress and drains available resources. Pre-screening for high cigarette consumption might enrich the screened population for bronchial dysplasia, but this has not been proven. Evaluation of sputum cytology is a less invasive approach for detecting neoplastic changes in the tracheobronchial tree. Qualitative

Table 71–3. “Definitive”* Clinical Cancer Chemoprevention Trials Organ System

Cohort

Upper Resected HNSCC aerodigestive tract Resected HNSCC (UADT) Resected HNSCC or lung cancer Stage I/II HNSCC definitively treated with XRT or surgery Oral leukoplakia (mucosal hyperor dysplasia)

Esophagus

Lung

Sample Size† 103 316 2,592 1,190

70 (59)‡

Oral leukoplakia

44

Residents of Linxian, China with esophageal dysplasia Residents of Huixian, China (with an assumed high prevalence of dysplasia or cancer) High-grade Barrett dysplasia

3,318

610

208

Author & Year of Publication(s)

Intervention ¥ Duration

Primary Efficacy Measure(s)

Isotretinoin 50–100 mg/m2/d vs. placebo ¥ 12 mos Etretinate 25–50 mg/d vs. placebo ¥ 24 mos Petinyl palmitate 150–300 IU/d vs. N-acetylcysteine 600 mg/d vs. both vs. neither ¥ 2 yrs Isotretinoin (low dose) vs. placebo ¥ 3 yrs and followed for 4 additional yrs

Second primary tumors—reduced 83%††, p = 0.005 Second primary tumors—not significantly reduced Second primary tumors—not significantly reduced; overall survival or event-free survival not significantly reduced Annual second primary tumor rate = 4.7% in both arms, p = 0.99

Hong et al., 1990 Bolla et al., 1994 van Zandwijk et al., 2000

Isotretinoin 1.5 mg/kg/d ¥ 3 mos; responders randomized to isotretinoin 0.5 mg/kg/d vs. b-carotene 30 mg/d ¥ 9 mos Isotretinoin 1–2 mg/kg/d vs. placebo ¥ 3 mos

Leukoplakia regression: Isotretinoin vs. bcarotene—92% vs. 45% response††, p < 0.001

Lippman et al., 1993

Leukoplakia regression (Major reduction size) in 85%††, p = 0.0002; reversed dysplasia in 81%††, p = 0.01 RR of esophageal/gastric cardia death = 0.92 (95% CI: 0.67–1.28)

Hong et al., 1986

Multivitamin supplement (containing 14 vitamins and 12 minerals) vs. placebo ¥ 5.25 yrs Retinol 50,000 IU/d + riboflavin 200 mg/d + zinc 50 mg/d vs. placebo ¥ 13.5 mos

Khuri et al., 2003

Li et al., 1993

OR of a normal esophagus at the conclusion of treatment = 0.85 (95% CI: 0.60–1.21)

Munoz et al., 1985

CR3 or greater (major response) in 77% vs. 39%††, p < 0.0001; CR1 (complete response) in 52% vs. 7%††, p < 0.0001; cancer-free at 24 mos = 83% vs. 53%††, p = 0.0014 Primary lung cancer incidence AT = 2% reduction (95% CI: -14% to +12%); BC = 18%†† increase (95% CI*: 3–36%) Secondary analyses Prostate cancer incidence AT = 32%†† reduction BC = 23% (NS) increase Cancer mortality AT = 41%†† reduction BC = 15% (NS) increase With 6–8 yrs of additional follow-up No significant effects on any cancer; Overall mortality AT RR = 1.01 (95% CI: 0.96–1.05) BC RR = 1.07 (95% CI: 1.02–1.12) Primary lung cancer incidence RR = 1.28†† (95% CI: 1.04– 1.57), p = 0.02; Lung cancer mortality RR = 1.46†† (95% CI: 1.07–2.00)

Product label 2003**

Male smokers 50– 69 years old

29,133

Porfimer sodium 2 mg/kg (followed by photodynamic light therapy) + omeprazole vs. omeprazole ¥ 2–3.6 yrs AT 50 mg/d vs. BC 20 mg/d vs. both vs. neither ¥ 5–8 yrs AT 50 mg/d vs. BC 20 mg/d vs. both vs. neither ¥ 5–8 yrs with 6–8 yrs of additional follow-up

Smokers, former smokers, and workers exposed to asbestos Resected stage I non-small cell cancer lung cancer Resected stage I non-small cell lung cancer

18,314

BC 30 mg/d + retinol 25,000 IU/d vs. placebo ¥ 5 yrs

307

Retinol palmitate 300,000 IU/d vs. no treatment ¥ 12 mos

Proportion without a second primary lung cancer = 11%†† increase, p = 0.045

Pastorino et al., 1993

1,166

Isotretinoin 30 mg/d vs. placebo ¥ 3 yrs

Second primary tumor HR = 1.08 (95% CI: 0.78–1.49); Tumor recurrence HR = 0.99 (95% CI: 0.76–1.29); Mortality HR = 1.07 (95% CI: 0.84–1.35) Metaplasia index reduced (>8%) in both groups—isotretinoin = 54.3%; placebo = 58.8% (NS)

Lippman et al., 2001

Chronic smokers 86 with prevalent dysplasia or metaplasia index >15% Current or former 112 smokers with >/= 30 pack-year history of tobacco use and bronchial dysplasia Current smokers 82 with squamous dysplasia or metaplasia

Isotretinoin 1 mg/kg/d vs. placebo ¥ 6 mos

ATBC 1994; Heinonen et al., 1998; Virtamo et al., 2003

Omenn et al., 1996

Lee et al., 1994

Anethole dithiolthione 25 mg TID vs. placebo ¥ 6 mos

Bronchial dysplasia progression rate in ADT vs. placebo = 32% vs. 59%††, p < 0.001

Lam et al., 2002

Fenretinide 200 mg/d vs. placebo ¥ 6 mos

Bronchial meta-/dysplasia—no significant effect; genetic/phenotypic abnormalities—no significant effect

Kurie et al., 2000 (continued)

1325

Table 71–3. (cont.) Organ System

Cohort Former smokers with >/= 20 pack-years of tobacco use Former asbestos workers Current smokers

Sample Size†

Intervention ¥ Duration

226

9-cis-retinoic acid 100 mg/d vs. isotretinoin 1 mg/kg/d + AT 1,200 IU/d vs. placebo ¥ 3 mos

755

BC 50 mg + retinol 25,000 IU QOD vs. placebo ¥ median of 5 yrs Etretinate 25 mg/d vs. placebo ¥ 6 mos Anti-Helicobacter antibiotics +/or b-carotene +/or ascorbic acid vs. placebo ¥ 6 yrs

150

Stomach

Residents of Narino, 852 Columbia with precancerous lesions

Colorectum

US male physicians

22,071

Aspirin 325 mg QOD vs. BC 50 mg QOD vs. both vs. neither ¥ 5 yrs

Recent resected adenoma

864

Recent resected adenoma

424

b-carotene 25 mg/d vs. vitamin C 1 gm/d + vitamin E 400 mg/d vs. both vs. placebo ¥ 4 yrs Low dietary fat (25% of calories) vs. 25 gm/d wheat bran vs. carotene 20 mg/d vs. combinations vs. placebo ¥ yrs

Recent resected adenoma Recent resected adenoma

1,429

Recent resected adenoma

665

Recent resected adenoma Recent resected adenoma

913

1326

272

Recent resected adenoma

1,121

Resected early stage colorectal cancer FAP with prevalent adenomas

635

FAP with prevalent adenomas FAP with prevalent adenomas

Liver

2,079

10 (crossover design) 22 24

FAP with prevalent adenomas

77

Genotype +/phenotype—FAP patients 8–25 years old Chronic active hepatitis C with cirrhosis Male patients with chronic hepatitis B infection Compensated hepatitis Crelated cirrhosis

41

90 101 99

Wheat bran fiber 13.5 gm/d vs. 2 gm/d ¥ 5 yrs Intensive low fat, high fiber, high fruits and vegetable diet vs. no intervention ¥ 4 yrs Calcium gluconolactate/carbonate 2 gm/d vs. ispaghula husk 3.5 gm/d vs. placebo ¥ 3 yrs

Primary Efficacy Measure(s) Bronchial metaplasia—reduction with 9-cis-RA††, p = 0.01; RAR-beta expression restored with 9-cis-RA††, p = 0.03; no change with isotretinoin Sputum cytologic atypia—no significant improvement Sputum cytologic atypia—no significant improvement Precancerous lesion regression anti-H. pylori antibiotics RR = 4.8†† (95% CI: 1.6–14.2); BC RR = 5.1†† (95% CI: 1.7–15.0); Ascorbic acid RR = 5.0†† (95% CI: 1.7–14.4) Primary CRC incidence RR = 1.15 (95% CI: 0.80–1.65); In situ cancer/polyp incidence RR = 0.86 (95% CI: 0.68–1.10) Adenoma recurrence BC RR = 1.01 (95% CI: 0.85–1.20); Vit C + E RR = 1.08 (95% CI: 0.91–1.29) Adenoma recurrence: No significant effect with any of the interventions Secondary analysis of large adenoma recurrence: low fat + wheat bran combination reduced recurrence††, p = 0.03 Adenoma recurrence adj. OR = 0.88†† (95% CI: 0.70–1.11), p = 0.28 Adenoma recurrence RR = 1.00 (95% CI: 0.90–1.12)

Adenoma recurrence: Calcium adj. OR = 0.66 (95% CI: 0.38–1.17), p = 0.16; Fiber adj. OR = 1.67†† (95% CI: 1.01–2.76), p = 0.028 Calcium carbonate 3 gm/d vs. Adenoma recurrence adj. RR = 0.85†† (95% CI: placebo ¥ 4 yrs 0.74–0.98), p = 0.03 Lysine acetylsalicylate 160– Adenoma recurrence RR = 0.73 (95% CI: 300 mg/d vs. placebo ¥ 1 yr 0.52–1.04), p = 0.08; Secondary analysis of large adenomas—83%†† reduction, p = 0.01 Aspirin 81 mg/d vs. 325 mg/d vs. Adenoma recurrence placebo ¥ yrs 81 mg RR = 0.81†† (95% CI: 0.69–0.96); 325 mg RR = 0.96 (95% CI: 0.81–1.13); Secondary analysis for advanced adenoma 81 mg RR = 0.59†† (95% CI: 0.38–0.92); 325 mg RR = 0.83 (95% CI: 0.55–1.23) Aspirin 325 mg/d vs. placebo ¥ yrs Recurrent adenoma RR = 0.65†† (95% CI: 0.46–0.91); Time to first adenoma prolonged HR = 0.64†† (95% CI: 0.43–0.94), p = 0.022 Sulindac 300 mg/d vs. placebo Complete or near complete CR adenoma ¥ 4 mos regression Sulindac—9 vs. placebo—0 (increase in 5, stable disease in 2, relative reduction in 2)††, p < 0.01 Sulindac 150 mg BID vs. placebo ¥ CR polyp number—56%†† reduction, p = 0.014; 9 mos CR polyp diameter—65%†† reduction, p < 0.001 Sulindac vs. placebo ¥ 6 mos CR polyp number—significantly reduced, p = 0.01; Duodenal polyp number—trend toward reduction, p = 0.12 Celecoxib 200 BID vs. 400 BID Mean number of adenomas—28%†† reduction, vs. placebo ¥ 6 mos p = 0.003; Polyp burden—30.7%†† reduction, p = 0.001 Sulindac 75–150 mg BID vs. Adenoma incidence—sulindac = 43% vs. placebo ¥ 48 mos placebo = 55%, p = 0.54; mean number or size of adenomas—no significant differences Interferon-a 6 MU TIW vs. symptomatic Rx ¥ 12–24 wks with 2–7 yrs follow-up Interferon vs. interferon with prednisolone priming vs. placebo ¥ 12 wks Interferon a-2b 3 MU TIW vs. no treatment ¥ 48 wks

Author & Year of Publication(s) Kurie et al., 2003 McLarty et al., 1995 Arnold et al., 1992 Correa et al., 2000

Gann et al., 1993 Greenberg et al., 1994 MacLennan et al., 1995

Alberts et al., 2000 Schatzkin et al., 2000 Bonithon-Kopp et al., 2000

Baron et al., 1999 Benamouzig et al., 2003 Baron et al., 2003

Sandler et al., 2003 Labayle et al., 1991 Giardiello et al., 1993 Nugent et al., 1993 Steinbach et al., 2000 Giardiello et al., 2002

Primary cancer incidence RR = 0.067†† (95% CI: 0.009–0.53), p = 0.01

Nishiguchi et al., 1995

Cumulative primary cancer incidence—87%†† reduction, p = 0.13

Lin et al., 1999

Primary cancer incidence—no significant effect

Valla et al., 1999

Table 71–3. (cont.) Organ System

Breast

Prostate

Bladder

Cervix

Skin

Cohort

Sample Size†

Intervention ¥ Duration

Primary Efficacy Measure(s)

Author & Year of Publication(s)

89

Polyprenoic acid 600 mg/d vs. placebo ¥ 12 mos

Second primary cancer incidence adj. RR = 0.31†† (95% CI: 0.12–0.78)

Muto et al., 1996

2,972

Fenretinide 200 mg/d vs. placebo ¥ 5 yrs

Veronesi et al., 1996; 1999

13,388

Tamoxifen 20 mg/d vs. placebo ¥ 5 yrs

Contralateral breast cancer incidence HR = 0.92 (95% CI: 0.66–1.29); Ipsilateral breast cancer recurrence HR = 0.83 (95% CI: 0.64–1.09) Invasive breast cancer incidence RR = 0.51††, p < 0.00001; Noninvasive breast cancer incidence RR = 0.50††, p < 0.002

2,471

Tamoxifen 20 mg/d vs. placebo ¥ median of 70 mos

Primary cancer incidence RR = 1.06 (95% CI: 0.7–1.7), p = 0.8

Powles et al., 1998

5,408

Tamoxifen 20 mg/d vs. placebo ¥ 5 yrs

Veronesi et al., 1998; 2003

Women with increased risk of breast cancer (>2 fold relative risk) Men >50 years of age

7,139

Tamoxifen 20 mg/d vs. placebo ¥ 5 yrs

Primary cancer incidence—no significant effect Secondary analyses High-risk subset††—reduced, p = 0.003; Low-risk subset—no effect, p = 0.89 Primary cancer incidence risk reduction = 32%†† (95% CI: 8–50), p = 0.013

18,882

Finasteride 5 mg/d vs. placebo ¥ 7 yrs

Superficial transitional cell carcinoma following BCG Superficial transitional cell carcinoma Women 16–23 years old

65

Megadose vitamins vs. RDA multivitamins ¥ 10 mos

660

Intravesical and percutaneous BCG —induction and maintenance over 3 yrs vs. induction alone HPV-16 virus-like particle vaccine 40 mcg/dose vs. placebo ¥ 3 with 17.4 mos follow-up Cervical caps with all-trans-retinoic acid 1 ml of 0.372% vs. placebo ¥ 6 mos (periodically) Cervical caps with sponge of all-trans-retinoic acid 0.16% vs. 0.28% vs. 0.37% vs. placebo ¥ 4d with outcomes at 12 wks

Resected or ablated liver cancer Women with resected stage I cancer or DCIS Women with increased risk of breast cancer (Gail model) Women with family history of breast cancer Women with normal risk, posthysterectomy

2,392

Women with CIN** II and III

301

Women with highgrade squamous intraepithelial neoplasia of the cervix Women with CIN II and III

175

30

Indole-3–carbinol 200 mg/d vs. 400 mg/d vs. placebo ¥ 12 wks

Women with prevalent koilocytic atypia, CIN I or II Women with CIN II or III

331

Folic acid 5 mg/d vs. placebo ¥ 6 mos

114

9-cis-retinoic acid 25 mg/d vs. 50 mg/d vs. placebo ¥ 12 wks

Resected nonmelanoma skin cancer Resected basal cell cancer

1,805

BC 50 mg/d vs. placebo ¥ 5 yrs

981

Isotretinoin 10 mg/d vs. placebo ¥ 3 yrs

Resected basal or squamous cell cancer Resected basal cell or squamous cell cancer

525

Retinol 25,000 IU/d vs. isotretinoin 5–10 mg/d vs. placebo ¥ 3 yrs

1,312

Selenium 200 mcg/d vs. placebo ¥ mean 4.5 yrs

Fisher et al., 1998

IBIS investigators 2002 (Cuzick et al., 2002)

Seven-year period prevalence = 24.8%†† reduction (95% CI: 18.6–30.6), p < 0.001; Gleason grade 7–10 more common†† in the treated group, p < 0.001 Secondary tumors or recurrence—5%†† reduction, p = 0.0014

Thompson et al., 2003

Median recurrence-free survival = 35.7 vs. 76.8 mos††, p < 0.0001

Lamm 2000; Lamm et al., 2000 Koutsky et al., 2002

HPV-16—related CIN incidence—Rx vs. placebo = 0 vs. 9 cases††, p < 0.001

Lamm et al., 1994

CIN-II complete histologic regression—59%†† increase, p = 0.041; CIN-III regression—no difference CIN complete regression rates: Placebo = 47%; low-dose = 56%; moderate-dose = 50%; high dose = 40% No significant differences, p = 0.28

Meyskens, 1994

Persistent CIN on follow-up 200 mg RR = 0.50†† (95% CI: 0.25–0.99); 400 mg RR = 0.55†† (95% CI: 0.31–0.99) Cytologic or colposcopic CIN—no significant improvement

Bell et al., 2000

CIN Regression rates—placebo = 32%; low-dose = 32%; high-dose = 36%, No significant differences Secondary nonmelanoma skin cancers RR = 1.05 (95% CI: 0.91–1.22)

Alvarez et al., 2003

Percentage of patients with incident basal cell cancers—no significant difference; Annual rate of incident basal cell cancers—no significant difference Time to first skin cancer—no significant difference; Number of incident skin cancers— no significant difference Basal cell cancer incidence RR = 1.10 (95% CI: 0.95–1.28); Squamous cell cancer incidence RR = 1.14 (0.93–1.39); Secondary analyses Colon cancer incidence RR = 0.42†† (95% CI: 0.18–0.95); Lung cancer incidence RR = 0.54†† (95% CI: 0.30–0.98) Prostate cancer incidence RR = 0.37††, p = 0.002

Ruffin et al., 2003

Childers et al., 1995

Greenberg et al., 1990 Tangrea et al., 1992 Levine et al., 1997 Clark et al., 1996; 1998

(continued)

1327

1328

PART V: CANCER PREVENTION AND CONTROL

Table 71–3. (cont.) Organ System

Cohort

Resected actinic keratoses and/or skin cancers Healthy male physicians aged 40–84 years

Intervention ¥ Duration

Primary Efficacy Measure(s)

2,297

Retinol 25,000 IU/d vs. placebo ¥ £5 yrs

22,071

BC 50 mg QOD vs. placebo ¥ 12 yrs

Xeroderma pigmetosum patients

30

Topical T4N5 liposome lotion vs. placebo ¥ 12 mos

Prevalent actinic keratoses

36

Prevalent actinic keratoses

Overall cancer incidence and mortality

Sample Size†

Renal allograft recipients with more than 10 prevalent AKs Residents of Linxian, China 40–69 years old at high risk for gastroesophageal cancers Healthy male physicians 40–84 years old

Topical 20% aminolevulinic acid hydrochloride with fluorescent blue light vs. vehicle Not Topical 3% diclofenac in 2.5% specified hyaluronic acid gel BID vs. vehicle ¥ up to 90 days 44 Acitretin 30 mg/d vs. placebo ¥ 6 mos 29,584

Complex factorial design with 4 arms: retinal + zinc riboflavin + niacin vitamin C + molybdenum BC + vitamin E + selenium ¥ 1–5 yrs

22,071

BC 50 mg QOD vs. placebo ¥ 12 yrs

Total cancer incidence RR = 0.63†† (95% CI: 0.47–0.85) All-cause mortality RR = 0.50†† (95% CI: 0.31–0.80) Squamous cell cancer incidence—reduced in those with prior AKs††, p = 0.04; otherwise no significant effects Nonmelanoma skin cancer incidence RR = 0.98 (95% CI: 0.92–1.05); Basal cell cancer incidence RR = 0.99 (95% CI: 0.92–1.06; Squamous cell cancer incidence RR = 0.97 (95% CI: 0.84–1.13 Annual rate of new actinic keratoses: Rx = 8.2 vs. placebo = 25.9, yielding a difference of 17.7†† (95% CI: 11.8–26.5); Annual rate of basal cell cancers: Rx = 3.8 vs. placebo = 5.4, yielding a difference of 1.6 (95% CI: 0.38–2.82) Actinic keratosis clearance: Treated vs. control = 88% vs. 6%

Author & Year of Publication(s)

Moon et al., 1997 Frieling et al., 2000

Yarosh et al., 2001

Jeffes, 2002

Actinic keratosis regression: Treated vs. control = 33% at 60 days, 50% at 90 days††

Jarvis and Figgitt, 2003

Squamous cell cancer incidence: Treated vs. control = 11% vs. 47%††, p = 0.01; Actinic keratosis incidence: Treated vs. control = 13.4 vs. 28.2†† (41.6% reduction) Stomach cancer incidence RR = 0.79†† (95% CI: 0.64–0.99) with BC + Vitamin E + selenium; Overall cancer incidence RR = 0.87†† (95% CI: 0.75–1.00) with BC + Vitamin E + selenium; Overall mortality RR = 0.91†† (95% CI: 0.84–0.99) with BC + Vitamin E + selenium No significant effects from other agent combinations Overall cancer incidence RR = 0.98 (95% CI: 0.91–1.06); Cancer mortality RR = 1.02 (95% CI: 0.89–1.18)

Bavinck et al., 1995 Blot et al., 1993

Hennekens et al., 1996

*A “definitive” trial was selected based up on: (1) the strength of its design (i.e., testing a specific a priori hypothesis in a homogeneous cohort, with a specific agent allocated via randomization, adequate controls, and well-defined response/outcome criteria) and (2) its intent to provide robust data challenging the current standard of care. More than one trial may be required to challenge the standard of care; nevertheless, standing alone, these trials could be reasonably anticipated to contribute to that goal (Lippman et al., 1998). † Numbers of subjects randomized. ‡ Responders were randomized. **Product information can be found at: http://www.fda.gov/cder/foi/label/2003/20451s012_photofrin_lbl.pdf †† Statistically significant effect as originally reported. AT, alpha-tocopherol; BC, beta-carotene; CI, 95% confidence interval; CIN, cervical intraepithelial neoplasia; FAP, familial adenomatous polyposis; HNSCC, head and neck squamous cell carcinoma; RR = relative risk; 95%.

changes in sputum cytology correlated well with retinoid (13cRA) use in one study and provide opportunities for prospective assessments of high-risk patients. Several confounders, such as smoking, complicate formidable challenges posed by screening for lung neoplasia. For example, in randomized, placebo-controlled studies of current smokers with dysplasia/metaplasia diagnosed by conventional whitelight bronchoscopy or sputum cytology, the placebo groups were shown to have high rates of spontaneous regression (20%–50%), no significant effect of treatment on histopathologic changes, and significant reductions in metaplasia upon smoking cessation. Therefore, dysplasia is generally viewed as a more robust end point for prevention research than is metaplasia. Emerging technologies (e.g., lowdose spiral/high-resolution computed tomography, autofluorescence bronchoscopy, endobronchial ultrasonography, optical coherence tomography, and confocal micro-endoscopy) may improve our ability to detect and localize pre-invasive or early invasive bronchial lesions in patients at high risk for lung malignancy (Lam et al., 1998; Hirsch et al., 2001). Recent studies targeting high-risk former smokers have produced some promising early results (Kurie et al., 2003).

Several promising chemopreventive agents have been or are being tested in the lung. High-dose Vitamin A prevented second primary cancers but not recurrence in 307 patients with previously resected stage I non-small–cell lung cancer (NSCLC) (Pastorino et al., 1993). The phase III, placebo-controlled EUROSCAN involving Vitamin A in 2592 lung or head and neck cancer patients, however, did not confirm these positive results (van Zandwijk et al., 2000). The utility of retinoids in preventing lung cancer was further questioned by data from an Intergroup study of 1166 resected stage I NSCLC patients treated with isotretinoin or placebo, who had a median follow-up of 3.5 years (Lippman et al., 2001). Differential effects on lung cancer recurrence were found by secondary multivariate and subset analyses, which showed statistically significant increases in tumor recurrence and mortality among current smokers vs. in never smokers randomized to isotretinoin (Khuri et al., 2001). These mixed outcomes may reflect the complexity of risk-benefit analyses within and across highrisk cohorts and neoplastic end points, as well as confounders that can skew results from smaller retinoid studies. NSAIDs, lipoxygenase (LOX) inhibitors, and the micronutrient selenium may have protective

Cancer Chemoprevention effects against lung cancer and are currently being evaluated in clinical cancer prevention trials. Two large lung cancer prevention trials conclusively demonstrated that the dietary supplement b-carotene has harmful effects in smokers. The Alpha-Tocopherol, Beta-Carotene (ATBC) Prevention Study (1994) randomized 29,133 Finnish male smokers to a-tocopherol (Vitamin E), b-carotene, a-tocopherol plus b-carotene, or placebo for 5–8 years. Higher rates of lung cancer and lung cancer–associated mortality were observed in participants who were randomized to highdose b-carotene, and these results were confirmed by the Carotene and Retinol Efficacy Trial (CARET), which randomized 18,314 smokers and asbestos workers to a combination of b-carotene and retinyl palmitate vs. placebo for a median of 4 years (Omenn et al., 1996). An interim analysis of CARET showed a 28% increase in lung cancer incidence and a 17% increase in mortality in the b-carotene arm, forcing early termination of the trial. Although negative in its primary lung end points, the ATBC study had important positive secondary analysis results, including a 32% decrease in prostate cancer incidence and a 41% decrease in prostate cancer mortality associated with atocopherol (Heinonen et al., 1998). Adverse effects of the seemingly innocuous dietary supplement bcarotene led to unexpected results in the ATBC study and CARET. These results highlight the importance of optimizing the TI whenever feasible. Aerosolized delivery of agents directly to the respiratory epithelium is an innovative approach to improving the TI of agents for lung cancer prevention. Wattenberg and colleagues have demonstrated that various aerosolized glucocorticoids (Wattenberg et al., 1997) and aerosol combinations (Wattenberg et al., 2000) can inhibit pulmonary neoplasia in carcinogen-induced lung cancer mouse models. This approach involves doses comparable to those used in controlling allergic rhinitis and represents an important conceptual advance in the field of cancer prevention. Data from a recently completed lung cancer prevention trial using aerosolized budesonide are expected to be published in the near future.

GASTROESOPHAGEAL CANCER PREVENTION Gastroesophageal (GE) cancers encompass malignancies of the stomach, gastroesophageal junction, and esophagus, which differ markedly in their respective molecular/biologic characteristics and epidemiologic profiles. One feature shared by all GE cancers is a high risk for local and distant recurrence despite surgery and adjuvant therapy for localized disease. Gastric cancer is the second leading cause of cancer death worldwide (Parkin et al., 1999) and has been associated with environmental factors such as diets high in fatty, pickled, smoked, or salty foods; diets low in fresh fruits and vegetables; and gastric mucosal infection with Helicobacter (H.) pylori (Uemura et al., 2001). The potential benefits of antioxidant supplements and/or H. pylori eradication were evaluated in 630 high-risk persons with gastric dysplasia in Narino, Colombia, a region in the Andes known for high rates of gastric cancer (Correa et al., 2000). Participants were randomized to anti-H. pylori triple therapy and/or ascorbic acid, b-carotene, or corresponding placebos for 6 years. Marked histopathologic improvement occurred in subjects randomized to ascorbic acid, b-carotene, or antibiotic therapy (vs. placebo), suggesting that these agents may be able to reduce the public health burden of gastric cancer. Observational and laboratory data suggest that NSAIDs also may exert preventive effects against gastric cancer, but this hypothesis has not been tested in human trials (Hawk et al., 2003). Up until the mid 1970s the most common type of esophageal cancer in the United States was squamous type, which remains a major cause of morbidity and mortality in Asia and Africa (Engel et al., 2003). The major risk factor for esophageal adenocarcinoma is gastroesophageal reflux, whereas the principal causes of squamous esophageal cancer are tobacco, alcohol, and certain dietary deficiencies. Both cancer types are associated with dysplastic precursors identifiable by cytologic or histopathologic analysis. Most esophageal cancer prevention trials have been conducted in Linxian, a region in the People’s Repub-

1329

lic of China with one of the highest rates of squamous esophageal cancer in the world. Although trials of antioxidant or calcium supplementation to reduce esophageal dysplasia in this population are rarely reported, one randomized trial involving a complex fractional factorial design and nearly 30,000 men and women found that participants receiving a combination of selenium, b-carotene, and Vitamin E supplements had a statistically significant 13% reduction in total cancer deaths, a significant 21% reduction in gastric cancer deaths, and a 4% non-significant reduction in esophageal cancer deaths, compared with participants not receiving this combination (Blot et al., 1993; Mark et al., 1994). These results suggest a modest role for such interventions against esophageal squamous cell cancers in this high-risk population. As with gastric cancer, observational data suggest that NSAIDs may reduce the risk of esophageal cancer by 40%–70%; this hypothesis has yet to be tested in definitive, randomized clinical trials (Schreinemachers and Everson, 1994; Farrow et al., 1998). Esophageal adenocarcinoma has the fastest rising incidence of all cancers in western countries (Kelty et al., 2002), which directs attention to its neoplastic precursor, Barrett’s esophagus (BE). BE arises from a process of intestinal metaplasia, through which columnar epithelium replaces normal squamous epithelium of the lower esophagus. Until recently, interventions against BE, both medical and surgical, primarily targeted acid exposure/reflux without consistent evidence of lesional regression or lower rates of progression to invasive cancer (Wild and Hardie, 2003). Although the majority of BE patients do not progress to cancer, there is no accurate way to distinguish between BE patients who will or will not progress. Moreover, the vast majority of cancer cases do not have a prior diagnosis of BE. Molecular markers of esophageal carcinogenesis would facilitate early cancer detection and might serve as targets of applied pharmacologic and dietary interventions. Strategies for early detection and prevention of esophageal adenocarcinoma might compensate for the current lack of effective therapy for this highly morbid condition. In August 2003 the US FDA approved and granted orphan drug designation to a photosensitizing porphyrin mixture (Photofrin) in conjunction with photodynamic therapy (PDT) (Axcan Pharma Inc.) for the ablation of high-grade dysplasia in patients with BE who cannot or choose not to undergo esophagectomy. The multi-center, controlled, partially masked trial leading to this approval involved 138 patients randomized to Photofrin PDT plus omeprazole and 70 patients randomized to omeprazole alone (). Follow-up ranged from 2–3.6 years and showed that a complete and sustained eradication of highgrade dysplasia occurred in 77% of the patients treated with the combination therapy vs. in 39% of the patients treated with omeprazole alone (p < 0.0001). Photofrin PDT involved intravenous injection of the porphyrin, which was activated 40–50 hours later by nonthermal endoscopic application of red light at 630 nm (a maximum of three courses of treatment, as indicated). The proposed mechanisms for this photochemical reaction include tissue necrosis from free radical damage, anoxia induced by vascular thrombosis, and inflammatory responses. The FDA approval of this approach provides an important option for patients with high-grade esophageal dysplasia and may stimulate scientific and pharmaceutical commitment to developing less invasive and less toxic strategies. Data from trials evaluating the merits of difluoromethylornithine (DFMO), celecoxib, or selenium in patients at risk for esophageal cancer are expected to be published in the near future (Limburg et al., 2005).

COLORECTAL CANCER PREVENTION Colorectal cancer (CRC) is a major health concern, with nearly 1 million new cases and 529,000 cancer-related deaths reported worldwide in 2002 (Parkin et al., 2005). In the United States, CRC affects nearly 145,290 people annually and will cause an anticipated 56,290 deaths in 2005 (Jemal et al., 2005). These statistics are especially unfortunate since most CRC should be preventable through wellestablished screening and surgical techniques. Although long-term survival for advanced CRC has not improved significantly over

1330

PART V: CANCER PREVENTION AND CONTROL

the past four decades, progressive insights into the molecular pathogenesis of CRC have led to rapid advances in early detection and interventional technologies that have made CRC the model organ for preventive agent development. NSAIDs are the most promising agents for CRC prevention (Hawk et al., 2002). Observational studies consistently show that NSAID use is associated with approximately 50% reductions in the CRC incidence and death rate as compared with non-use. The randomized evaluation of NSAIDs against CRC incidence comes from a secondary analysis of the Physicians’ Health Study (PHS), which found that the NSAID aspirin produced no significant reduction in CRC incidence (relative risk (RR) = 1.15, 95% CI: 0.68–1.10) or in situ cancers/polyps (RR = 0.86, 95% CI: 0.68–1.10) (vs. placebo) among 22,071 US male physicians (Gann et al., 1993). This surprising null effect has been attributed variously to the rarified cohort (male physicians), the lack of uniform colorectal surveillance guidelines, the limited dose and duration of aspirin, the unexpectedly low incidence of CRC among placebo recipients, and most importantly, the limited duration of follow-up relative to the long natural history of colorectal carcinogenesis. Three subsequent large trials have tested the efficacy of the same or higher doses of aspirin for the prevention of colorectal adenomas in more heterogeneous cohorts and have shown significant reductions in recurrent adenomas among participants treated with aspirin for one or more years (Baron et al., 2003; Benamouzig et al., 2003; Sandler et al., 2003). One of these trials was conducted by Sandler and colleagues (2003), who randomized 635 CRC survivors to aspirin (325 mg/day) vs. placebo. The aspirin group had a significant reduction in the number of patients with incident adenomas (17% vs. 27%, p = 0.004) and a significant delay in the time to a first adenoma (p = 0.022) after a median follow-up of 12.8 months. Baron and colleagues (2003) randomized 1121 patients with prior colorectal adenomas to aspirin (81 or 325 mg/day) vs. placebo and reported reductions in the number of people with adenomas, 19% for one and 4% for more than one adenoma. Perhaps more importantly, this study also reported striking efficacy against advanced adenomas, noting 41% and 17% reductions in the groups taking 81 or 325 mg/day, respectively. Interim results from another trial evaluating aspirin’s effects against incident colorectal adenomas show preventive effects after 1 year of aspirin use (vs. placebo) in 272 patients with prior adenomas (Benamouzig et al., 2003). Longer-term data will be forthcoming since this randomized trial is ongoing. COX-2 is over-expressed in nearly 50% of colorectal adenomas and 80%–85% of adenocarcinomas, whereas it is rarely expressed in normal colorectal epithelium (Williams et al., 1999). COX-2 overexpression appears to be functionally important for neoplastic progression, as suggested by the lower incidence of intestinal polyps in mice lacking a functional COX-2 gene (Oshima et al., 1996). On the strength of these data, COX-2–selective inhibitors (e.g., celecoxib, rofecoxib) are being studied in persons at elevated risks for CRC due to germline mutations or sporadic neoplasia. Celecoxib administered for 6 months to 83 individuals with FAP produced significant reductions in colorectal adenoma number and size, with a side effect profile comparable with that of placebo (Steinbach et al., 2000). These data were the basis for a landmark US FDA approval of celecoxib to complement the standard of care (i.e., surveillance and prophylactic surgery) for patients with FAP, a genetic condition that in the absence of interventions, assures the development of CRC. Ongoing trials will confirm whether celecoxib—as a single agent or in combination with other drugs—durably reduces CRC risk in this cohort. Calcium has been studied in several phase II/III trials. Although case-control and epidemiologic studies have reported moderate inverse associations between calcium intake and CRC risk (Martinez and Willett, 1998), phase II trials have been less consistent. For example, early reports of significant reductions in colonic epithelial proliferation following calcium supplementation (Pence, 1993) were not confirmed by larger trials (Baron et al., 1999; van Gorkom et al., 2002). Nevertheless, randomized, placebo-controlled phase III trials have shown that calcium reduces adenoma recurrence in cohorts at increased risk for CRC (Table 71–3). A recent study involving 1000

patients showed statistically significant reductions of 19% in recurrent adenomas and 44% in advanced adenomas in the group treated with a calcium carbonate supplement (vs. placebo) (Baron et al., 1999; Wallace et al., 2002). Beneficial effects were observed as early as 1 year after initiating calcium treatment, suggesting that this supplement acts relatively quickly. These results were generally supported by a multi-center, randomized, placebo-controlled trial in 655 patients with prior adenomas, which found a non-significant 34% reduction in recurrent adenomas in the calcium (and/or fiber) supplements arm (vs. placebo) (Bonithon-Kopp et al., 2000). At a minimum, these data suggest that calcium supplementation is well tolerated and achieves modest reductions in recurrent adenomas. The largest study of calcium to date, the US Women’s Health Initiative (WHI) (1998), has randomized nearly 36,282 postmenopausal women to calcium and Vitamin D vs. placebo (among other agents) and will follow participants for 7–11 years. Daily supplementation with calcium carbonate (1000 mg) and vitamin D3 (400 IU) for seven years did not result in a reduced incidence of invasive colorectal cancer (Wactawski-Wende et al., 2006). Fiber supplementation has not been consistently shown to reduce adenoma recurrence. A randomized trial of wheat bran fiber in 1429 subjects with prior adenomas showed no significant reduction in colorectal adenomas associated with fiber (Alberts et al., 2000). Another trial assessing the efficacy of a low-fat, high-fiber, fruit and vegetable diet also reported null results (Schatzkin et al., 2000). A third phase III trial involving 665 participants found that fiber supplementation increased participants’ risk for recurrent adenomas by 67% (BonithonKopp et al., 2000). By contrast, a large international observational study, the European Prospective Investigation into Cancer and Nutrition (EPIC) study, found a statistically significant 25% reduction in CRC risk among individuals consuming the highest vs. lowest quintile of dietary fiber (Bingham et al., 2003). These mixed results suggest that trials using different types of fiber, different schedules (i.e., earlier, later, or in a manner not yet determined), and/or different biologic end points (i.e., cancer incidence, cancer-associated mortality) in different cohorts (i.e., younger persons or those without prior adenomas) may improve our understanding of fiber’s effects against colorectal carcinogenesis (Hawk et al., 2002). At least two ongoing international trials are evaluating the efficacy of resistant starch co-administered with aspirin in patients at hereditary risk for CRC (Table 71–3) (Burn et al., 1998). Dramatic preclinical activity has been observed with NSAID combinations, most notably using DFMO or EGFR inhibitors, and data on other agents is accumulating. Synergistic reductions in intestinal neoplasia were first demonstrated with combinations of eflornithine and COX inhibitors (e.g., piroxicam or aspirin), even when drug doses were reduced by as much as 50% (Nigro et al., 1986; Reddy et al., 1990; Rao et al., 1991; Li et al., 1999; Jacoby et al., 2000). More recently, Torrance et al. (2000) reported remarkable chemopreventive efficacy from a combination of an EGFR inhibitor with sulindac. These data have provided the preclinical rationale for NSAID combination trials in patients at moderate to high risk for CRC. New work in this cohort is focusing on novel prevention targets, including LOXs and peroxisome proliferator-activated receptors (PPARs) (Umar et al., 2001; Shureiqi et al., 2003).

HEPATOCELLULAR CANCER PREVENTION Hepatocellular carcinoma (HCC) is a one of the most prevalent cancers worldwide, particularly in the Asia Pacific region (Parkin et al., 2005). HCC is estimated to have caused nearly 600,000 in 2002 deaths annually, 80% of which occur in the developing world. Major etiologic risk factors for HCC include chronic infection with hepatitis B and/or C viruses (HBV, HCV), alcoholic cirrhosis, and environmental exposure to aflatoxin. Over the past two decades, the incidence of HCC has risen sharply in industrialized countries that formerly had low rates of disease. In the United States, the twofold increase in HCC incidence between 1975 and 1995 may reflect the latency of viral car-

Cancer Chemoprevention cinogenesis following HBV/HCV epidemics that peaked around the 1980s (El-Serag et al., 2003). The influx of immigrants from regions endemic for HCC also may contribute to this disturbing epidemiologic trend. HCC is an aggressive disease with abysmal survival rates and poor treatment options, making chemoprevention an appealing strategy for cancer management and control. Vaccination that prevents HBV infection has been shown to reduce HCC incidence and mortality in certain endemic regions (Change et al., 1997; Lee et al., 2003). Similar successes might be expected with the development of a vaccine against HCV, an RNA virus that was only clinically identified in the 1970s (Whittle et al., 2002). Current vaccine efforts are addressing challenges posed by HCV, such as genomic diversity, a high mutation rate, the tendency to escape immunologic surveillance, and the short-lived nature of HCV-neutralizing antibodies (Liang et al., 2000). In addition, it is difficult to culture HCV in vitro, and the only susceptible animal model for testing candidate vaccines is the chimpanzee, an endangered species with limited ability to mimic human immunologic responses (Chen et al., 1999). Obstacles to the development of a prophylactic HCV vaccine limit near-term and intermediate-term prospects for the prevention of HCV-induced HCC to behavioral approaches that prevent human transmission of the virus and interventions that slow or block the progression of disease in infected individuals. This is in contrast to HBV, for which immunization reduces the incidence of HBV-associated HCC, but it does not reduce HCC incidence among individuals already infected with the virus. Two recent meta-analyses of the observational and experimental literature suggest that interferon, an immune-modulator shown to induce sustained virologic responses (e.g., undetectable HCV RNA) and histologic improvements in a limited number of patients, reduces the risk of HCC, particularly in people with HCV (Camma et al., 2001; Papatheodoridis et al., 2001). Trials designed to confirm the preventive efficacy of interferon are ongoing. A trial of polyprenoic acid, an acyclic retinoid, found significant reductions in the incidence of second primary HCCs but no effect against tumor recurrence (Muto et al., 1996). Other promising approaches involve phase-2-enzyme inducers that boost aflatoxin detoxification (Kensler et al., 1998), COX-2 inhibitors, selective estrogen-receptor modulators (SERMs), and behavioral approaches to decrease risk among individuals with alcohol-related cirrhosis.

BREAST CANCER PREVENTION Breast cancer is the most common cancer in women worldwide, with over 1 million new cases and 411,000 deaths per year (Parkin et al., 2002). In the US female population, breast cancer is the most common cancer and the second leading cause of cancer mortality (Jemal et al., 2005). Although the incidence of breast cancer has been rising for the past two decades, mortality rates have remained relatively stable. This increased incidence coincides with the development and broad application of screening practices, especially mammography (Weir et al., 2003). More recently, improved risk-factor profiling (e.g., with the Gail Model (Gail et al., 1989) ) has enabled stratification of breast cancer risk based on factors such as family history, nulliparity, early menarche, advanced age, and personal history. Breast carcinogenesis is an estrogen-driven process, and therefore, major efforts have focused on dramatically lowering estrogen levels by targeting the estrogen receptor (e.g., with SERMs such as tamoxifen and raloxifene) and more recently by targeting aromatase. The largest breast cancer prevention study completed to date was the National Surgical Adjuvant Breast and Bowel Project (NSABP) P-1 trial, which was completed in 1998 (Fisher et al., 1998). This study exploited compelling results of breast cancer treatment studies that showed reduced rates of contralateral second primary caners among women receiving adjuvant tamoxifen. The P-1 trial randomized 13,388 women at high Gail-model risk for breast cancer to tamoxifen vs. placebo. Striking interim-analysis results prompted the early termination of the trial based on a statistically significant 49% reduction in the risk of breast cancer in the tamoxifen arm. Risk reductions were

1331

achieved against invasive as well as non-invasive subsets of women with estrogen-receptor–positive (ER+) breast carcinogenesis, regardless of age. Tamoxifen had no impact on the development of estrogenreceptor–negative (ER-) breast cancer. The Italian Tamoxifen Prevention Study randomized 5408 women who had hysterectomies to tamoxifen vs. placebo (Veronesi et al., 1998). Patient selection was not based on a high risk for breast cancer, and nearly half of the study cohort had prior ovarian ablation, which reduced circulating estrogen levels and, thus, breast cancer risk. The Italian study was closed prematurely to accrual because of a high dropout rate. Under these limiting circumstances, the Italian data showed no difference in breast cancer incidence, although unplanned subset analyses and extended follow-up suggested that women on hormone replacement therapy and at high risk based on reproductive and hormonal characteristics derived some preventive benefits from tamoxifen (Veronesi et al., 2003). Another study of tamoxifen was conducted at the Royal Marsden Hospital in the United Kingdom (Powles et al., 1998). This study randomized 2494 women ages 30–70 years old with a family history of breast cancer to tamoxifen vs. placebo and was designed as a feasibility study for the International Breast Cancer Intervention Study (IBIS-I) trial. Like the Italian study, the Royal Marsden study also suffered from a high dropout rate and found no difference in breast cancer incidence between study arms. More recently, the IBIS-I trial randomized 7410 women and showed a 32% reduction in breast cancer risk with tamoxifen (Cuzick et al., 2002). The breast cancer preventive effect of tamoxifen in both definitive trials (P-1 and IBIS-I) and the Italian positive subset analyses were limited to ER-positive cancers. Tamoxifen also has been shown to be effective in the setting of ductal carcinoma in situ (DCIS). In the placebo-controlled, randomized NSABP B-24 trial, 5 years of treatment with tamoxifen reduced the incidence of ipsilateral and contralateral breast cancer among 1804 women who had had a lumpectomy and radiation after being diagnosed with DCIS (Fisher et al., 1999). Because tamoxifen’s effects against DCIS were comparable with its effects against invasive breast cancer, DCIS is now regarded as a plausible SEB for breast cancer prevention. The P-1 trial established the preventive potential and revealed the complex risk-benefit profile of tamoxifen. In addition to preventing invasive and non-invasive breast cancer, tamoxifen reduced the risk of fractures. Tamoxifen has raised qualms, however, chiefly because it did not reduce the risk of ER-negative cancer and because of toxicities such as endometrial cancer, venous thromboembolism, hot flashes, cataracts, and vaginal dryness. The two negative European randomized controlled trails also may have contributed to a widespread reluctance among physicians to prescribe tamoxifen for cancer prevention. Despite its US FDA approval, tamoxifen has not been universally adopted as the standard of care for secondary prevention of breast cancer in the United States. An ongoing follow-up study of 22,000 women at high risk for breast cancer, the Study of Tamoxifen and Raloxifene (STAR), is assessing the efficacy and safety profiles of tamoxifen and the closely related SERM, raloxifene, for preventing breast cancer (Wickerham, 2003). Raloxifene was suggested for breast cancer prevention by studies showing a low endometrial cancer risk and by secondary analyses of osteoporosis prevention trials (e.g., the Multiple Outcomes of Raloxifene Evaluations (MORE) ) showing breast cancer risk reductions associated with raloxifene (Cummings et al., 1999). A recent analysis of the MORE data suggested that the benefit of raloxifene may be greatest in postmenopausal women with the highest circulating levels of estradiol (Cummings et al., 2002). Based largely on results of the Anastrozole, Tamoxifen Alone or in Combination (ATAC) trial, there is great interest in aromatase inhibitors (AIs) for breast cancer prevention in postmenopausal women. Involving 9366 patients, ATAC found that the AI, anastrozole, was more effective than was tamoxifen in reducing recurrence and preventing new contralateral tumors (Baum et al., 2003). Anastrozole was associated with fewer vascular and endometrial events, but more musculoskeletal events and fractures than was tamoxifen. Relatively few AI-associated toxicities (in striking contrast with tamoxifen) have been identified to date; this is a critical issue when considering the

1332

PART V: CANCER PREVENTION AND CONTROL

usefulness of an AI for breast cancer prevention. The adjuvant ATAC data have led to two AI prevention trials—IBIS-II comparing anastrozole with placebo in 6000 high-risk postmenopausal women and a trial comparing anastrozole with tamoxifen in the setting of resected DCIS (Cuzick 2003). If the benefit-risk results of STAR favor raloxifene, it will be critical to compare anastrozole with raloxifene for breast cancer prevention in the high risk setting. Despite the deserved enthusiasm for the potential of raloxifene and anastrozole, tamoxifen remains the only agent proven, and recommended by the American Society of Clinical Oncology and the US Preventive Services Task Force; and US FDA-approved for reducing the risk of breast cancer. A primary focus of new breast cancer prevention research is to develop agents for preventing ER-negative disease. Several agents (e.g., retinoid-X-receptor (RXR)-selective retinoids and inhibitors of EGFR and COX-2) have shown promise for preventing ER-negative cancer in preclinical studies (Shen and Brown, 2003).

PROSTATE CANCER PREVENTION Prostate cancer is the most common non-dermatologic malignancy in the world (Parkin et al., 2002). In the United States, prostate cancer has an incidence of more than 232,000 cases per year and is the second leading cause of cancer death in men (Jemal et al., 2005). It is estimated that aging of the US population will cause the incidence of prostate cancer to rise to more than 380,000 new cases per year by 2005. Prostate cancer etiology has been elusive, with age, family history, and race (African American) as the only known risk factors (Nelson et al., 2003). The major screening test for detecting early prostate cancer is serum prostate-specific antigen (PSA). It is unclear, however, what effect PSA screening has on prostate cancer mortality in the context of surgical and/or radiation therapy for PSA-detected early disease. This issue is being addressed in the definitive Prostate, Lung, Colorectal, and Ovarian (PLCO) Cancer Screening Trial (Gohagan et al., 1995). The biology, function, and age-related expression of PSA and its various molecular forms are under intense investigation to determine whether PSA might be a SEB for prostate cancer. Three precursor lesions for prostate cancer (atypical adenomatous hyperplasia, prostatic inflammatory atrophy (PIA), and prostatic intraepithelial neoplasia (Brooks et al., 1998) ) are also under investigation as candidate biomarkers for prostate cancer (De Marzo et al., 2003). Prostate carcinogenesis is an androgen-driven process. The development of finasteride, an inhibitor of 5a-reductase, the enzyme that converts testosterone to the more potent androgen DHT in the prostate, led to the Prostate Cancer Prevention Trial (PCPT) (Thompson et al., 1997). The PCPT tested finasteride (5 mg per day) vs. placebo for 7 years in 18,882 men 55 years of age or older who had a normal digital rectal examination (DRE) and PSA level of 3.0 ng per milliliter or lower (Thompson et al., 2003). Men were screened yearly with DRE and PSA, and a prostate biopsy was recommended for abnormal DRE and/or PSA more than 4. An end-of-study biopsy was planned because of the effect of finasteride in lowering PSA level (i.e., all men were offered a prostate biopsy after 7 years on study regardless of PSA level or DRE result). The primary end point was prevalence of prostate cancer during the 7 year study. The trial was closed early based upon the recommendation of the independent data and safety monitoring committee that the study objective had been met and the results were extremely unlikely to change if the trial continued until its planned completion. There was a 24.8% reduction in prostate cancer in the finasteride arm. Although finasteride was associated with a reduction in prostate cancer, it also was associated with an increased risk of being diagnosed with high-grade disease (a pre-specified secondary end point)—6.4% of men in the finasteride arm vs. 5.1% of men in the placebo arm developed high-grade (Gleason score 7–10) prostate cancer. The biology behind these dramatic, provocative, and unexplained clinical results is under intensive study. With regard to toxicity, sexual side effects were more common in the finasteride arm and urinary symptoms were more common in the placebo arm.

Another large primary prostate cancer prevention trial, the Selenium and Vitamin E Cancer Prevention Trial (SELECT), currently is assessing the efficacy of selenium and Vitamin E in 32,400 men (Klein et al., 2003). The SELECT hypotheses for testing selenium and atocopherol (Vitamin E) arose from secondary analyses of chemoprevention trials in other target organs showing that these agents may reduce the incidence of prostate cancer (Clark et al., 1998; Hartman et al., 1998; Heinonen et al., 1998).

BLADDER CANCER PREVENTION Bladder cancer is a common cause of cancer morbidity and mortality worldwide and has been associated with several risk factors. In industrialized countries, cigarette smoking has been identified as the greatest risk factor for bladder cancer that tends to be transitional cell carcinoma. In developing areas of the Middle East and Africa where schistosomiasis is endemic, squamous cell carcinoma of the urinary bladder is the most common cancer type (Moyad, 2003). Most newly diagnosed bladder cancer is actually a preinvasive disease, superficial bladder cancer, which is highly curable with surgical and non-surgical interventions (Kamat and Lamm, 1999). Superficial bladder cancer, however, has the highest annual rate of second primary tumor development. Although most of these second tumors are non-invasive, a substantial percentage of patients have multiple recurrences, some of which progress to invasive disease. In a recent Cochrane meta-analysis, intravesical therapy with Bacillus of Calmette Guerin (BCG) vaccine after surgical resection decreases recurrence rates by as much as 70% (Shelley et al., 2001). Oral retinoids, (Nutting and Huddart, 2001) NSAIDs (Sabichi and Lippman, 2003), eflornithine (Uchida et al., 1989; Loprinzi and Messing 1992), and megadose vitamins (Lamm et al., 1994) are in phase III clinical testing as secondary preventive agents against superficial bladder cancer to be used in conjunction with standard of care. Data from these studies will mature over the next several years.

CERVICAL CANCER PREVENTION Cervical cancer is a leading cause of cancer death among women in developing countries (Parkin et al., 2002). Cervical cancer prevention is a promising area of research for several reasons: the accessibility of the cervix to evaluation (i.e., Pap tests); the relatively slow progression of cervical intraepithelial neoplasia (CIN) (the wellrecognized precursor lesion) to cervical cancer; and the active development of vaccines targeting the major etiologic agent of this cancer, human papilloma virus (HPV) (Schiffman and Solomon, 2003; Sherman et al., 2003). High-grade CIN is treated by removing the abnormal tissue to reduce the risk of progression to invasive cancer. Broadly applied screening programs to detect and treat preinvasive neoplasia have greatly reduced cervical cancer incidence and mortality (Bergstrom et al., 1999; Sigurdsson, 1999). Therefore, CIN has a high degree of validation as an SEB for cervical cancer incidence and mortality in the context of screening coupled with surgery. Broad screening for CIN, however, may not be feasible in the developing regions of the world with the greatest public health burden of cervical cancer. The many randomized trials conducted in cervical dysplasia include studies of folic acid (Childers et al., 1995), interferon (Santhanam et al., 2002), b-carotene (Comerci et al., 1997), indole-3-carbinol (Bell et al., 2000), difluoromethylornithine (Boiko et al., 1997), fenretinide (Ruidi et al., 1997), all-trans-retinoic acid (Meyskens et al., 1994) 13cRA (Kim et al., 2003), and 9-cis-retinoic acid (Alvarez et al., 2003). The only positive trials in this group were an early trial of topical all-trans-retinoic acid (Meyskens et al., 1994). A subsequent randomized trial of topical all-trans-retinoic acid for a shorter duration was negative (Ruffin et al., 2003) and a trial of indole-3-carbinol has not yet been replicated. The major risk factor for cervical cancer is an HPV infection, and 95% of cervical cancers contain HPV, usually HPV-16. Vaccines that

Cancer Chemoprevention protect against common oncogenic subtypes, primarily HPV-16 and -18 or their related proteins E6 and E7, are expected to prevent cervical cancer; pharmaceutical companies are poised to file for FDA approval in the near future. A recent randomized trial of an HPV-16 virus-like-particle vaccine involving 2392 young women resulted in a marked reduction in both the incidence of HPV-16 infection and HPV16–associated CIN (Koutsky et al., 2002). A placebo-controlled trial of another HPV-16 vaccine was recently initiated in 6000 Costa Rican women, in whom the prevalence of HPV positivity among those with CIN approaches 89%. This vaccine is one of several under development that uses virus-like particles (VLPs) to generate high titers of neutralizing antibodies without posing a risk of infection (Pinto et al., 2003). The Costa Rican study will use virologic and pathologic end points to assess vaccine efficacy.

SKIN CANCER PREVENTION Skin cancer is classified as non-melanoma or melanoma, depending on its origin. The incidence of non-melanoma skin cancer (NMSC), which comprises basal and squamous cell carcinomas, nearly equals that of all other cancers combined. Furthermore, NMSC rates are increasing, which is consistent with dose-dependent effects of photocarcinogenesis on aging populations. NMSC may arise from actinic keratoses (AKs), precursor lesions associated with cumulative exposure to UV irradiation, immunosuppression, and certain chemicals. The malignant potential of AKs and the cost and cosmetic deformity of standard management (e.g., cryotherapy, chemical peels, carbon dioxide laser, shave excision, dermabrasion, or topical fluorouracil, trichloroacetic acid, phenol, or retinoids) have prompted a search for well-tolerated agents that reverse or retard carcinogenesis in the skin (Einspahr et al., 2002; Gupta and Mukhtar, 2002). Certain cohorts are at exceptionally high risks of developing NMSC and provide clinical models of accelerated skin carcinogenesis due to chronic immune suppression (e.g., solid organ transplant recipients on chronic immune suppression therapy) (Berg and Otley, 2002), UVhypersensitivity syndromes (Yamawaki et al., 1997), or certain rare genetic syndromes such as xeroderma pigmentosum (Norgauer et al., 2003). Retinoids, selenium, NSAIDs, eflornithine, a topically applied bacterial DNA repair enzyme (T4 endonuclease V) (Yarosh et al., 2001), and green tea compounds, such as polyphenols and epigallocatechin-3-gallate, are among the leading candidates being tested for NMSC chemoprevention (Einspahr et al., 2002). Certain retinoids have been shown to significantly reduce skin cancer incidence in very high-risk patients with xeroderma pigmentosum or renal transplants (Kraemer et al., 1988; Bavinck et al., 1995; Rook et al., 1995; Gibson et al., 1998; McKenna and Murphy, 1999). Chemoprevention trials of oral Vitamin A (retinol) for NMSC, however, have had mixed results. One phase III study showed that retinol was effective against moderate-to-severe AKs in association with reductions in squamous cell carcinoma (Moon et al., 1997). Another study of subjects with higher-risk disease due to a history of skin cancer, however, had unremarkable results (Levine et al., 1997). The results from phase III testing of very low-dose 13cRA and bcarotene were also negative (Greenberg et al., 1990; Tangrea et al., 1992). Selenium was tested in a randomized placebo-controlled trial in 1312 high-risk patients with a history of skin cancer (Clark et al., 1996). The results suggest that selenium increases the risk of NMSC (mostly squamous cell cancer) in this risk setting, which included many patients with evidence of arsenic exposure (Duffield-Lillico et al., 2003). Despite negative results with regard to the primary end point, this trial produced striking secondary end point results, such as significant reductions in total cancer mortality and reductions in incident lung, colorectal, and prostate cancers (Duffield-Lillico et al., 2002). Topical formulations of eflornithine (Alberts et al., 2000), colchicine (Akar et al., 2001; Tutrone et al., 2003) or 5-fluorouracil (Tutrone et al., 2003) have shown significant efficacy in suppressing or regressing AKs. Aminolevulinic acid HCl (Levulan Kerastick) with photodynamic therapy (PDT) and diclofenac sodium (Solaraze)

1333

recently received US regulatory approvals for treating AKs (http://www.fda.gov/cder/approval/index.htm). Melanoma accounts for a much smaller fraction of skin cancers than does NMSC (Desmond and Soong, 2003). Even so, because of its higher potential for metastasis, MSC poses the greater threat. Indeed, the incidence of higher-stage melanoma is comparable with the casefatality rate despite aggressive treatment. Intermediate lesions in a robust melanocyte progression model, such as dysplastic nevi (DN), may provide biomarkers of risk (Carlson et al., 2003) or end points (Carlson et al., 2003) for melanoma prevention research. Topical retinoids have been tested preliminarily in subjects with DN, precursor lesions with a high rate of transformation to melanoma (Halpern et al., 1994; Stam-Posthuma et al., 1998). Although topical application of the retinoid, tretinoin, produced clinical and histologic improvements in some DN, the responses were not durable and skin irritation raised concerns over long-term compliance with the regimen. A follow-up study to assess whether co-administration of systemic retinoids with a lower dose of topical tretinoin could improve efficacy has been considered. Positive preclinical and observational data on statins in melanomagenesis have stimulated interest in initiating prospective studies of statins for the prevention of MSC (Collisson et al., 2003; Dellavalle et al., 2003). Melanoma may be a good target for vaccine-based prevention because it is immunogenic (Bocchia et al., 2000). Response to immunization typically correlates inversely with disease stage and tumor burden, and so vaccines are likely to have a greater impact on subclinical disease than on advanced cancer. This provides a strong rationale for testing promising melanoma vaccines in a preventive context, or nesting melanoma prevention end points within vaccine treatment trials.

FUTURE DIRECTIONS Oncology has followed cardiology in shifting from an almost pure symptom- and treatment-based paradigm toward one that incorporates strong elements of disease prevention. Cardiology has successfully reduced the risk of cardiac events by identifying major cardiovascular (CV) risk factors and developed preventive interventions that target them. This process has led to the understanding that major CV risk factors (e.g., hypercholesterolemia, hypertension) are diseases in and of themselves, and to the development of effective interventions that control risk. Public and community health-care providers have become aware of and knowledgeable about risk factors for CV disease, and have incorporated control measures into the standard of care. The evolution of a risk factor to the status of disease involves subtle but important reinterpretation of pathology, and its implications for the individual as well as sociocultural perspectives on health and disease. For example, oncology has traditionally viewed clinically detected cancer as the disease of interest and all that went before, including preinvasive neoplasia, as more-or-less normal. Insights into the biology of neoplasia have led to an understanding that cancer is a late stage in a long, clinically silent process of carcinogenesis, which is characterized by molecular and cellular abnormalities that are increasingly difficult to classify as normal. Therefore, the maturing science and practice of oncology is assuming greater responsibility for the identification and management of major risk factors for cancer, including IEN. Emerging insights into the molecular biology of cancer susceptibility and carcinogenesis have created important opportunities for multidisciplinary approaches to the development of preventive agents. A wave of new technologies is rapidly increasing our understanding of how neoplasia evolves, and the roles that genetic and epigenetic events play in clonal expansion and genetic instability. The mechanism of multifocal carcinogenesis is now understood to include the remarkable phenomenon of intraepithelial spread/metastasis of preinvasive neoplastic cells. Basic discoveries such as these have profound impact on the development and direction of cancer prevention. Specifically, molecular alterations associated with carcinogenesis can be used as markers of risk and susceptibility for cancer (especially biologically

1334

PART V: CANCER PREVENTION AND CONTROL

aggressive disease), targets for drug development, and surrogates or end points to measure drug response and resistance. The clinical maturation of cancer prevention is illustrated by the FDA’s approval of several agents to prevent cancer or to treat/prevent IEN. Celecoxib, diclofenac sodium, photofrin (in conjunction with photodynamic therapy), and tamoxifen (in three distinct breast cancer risk settings) add to a rapidly expanding list of FDA-approved agents for cancer prevention that already included hepatitis B vaccine, BCG, valrubicin, masoprocol, 5-FU, and aminolaevulinic acid (with photodynamic therapy) (Table 71–1). In addition, the preventive efficacy of aspirin, calcium, and finasteride has been established in definitive randomized, controlled trials, but regulatory approval for cancer indications has not yet been sought. Definitive phase III prevention trials such as the BCPT and PCPT have raised debates as to whether the drugs in these trials prevented, treated, or delayed cancer. The early reductions in cancer incidence in the treatment vs. placebo arms of the BCPT (breast) and PCPT (prostate) suggest that the agents in both trials may have treated subclinical, microscopic cancer early in the study. The reductions in incidence, however, became more profound over time, thus establishing the major contribution of cancer prevention and/or delay. Further support for the concept of cancer delay is suggested by the recent postintervention update of the ATBC trial showing that the beneficial AT and harmful BC results were lost after years of follow-up off drug (Virtamo et al., 2003). Regardless as to how these effects were achieved—by treating subclinical neoplasia, or by preventing or delaying cancer—the effects to date are provocative and important (Lippman and Hong, 2002). Cancer prevention faces challenges common to every emerging medical discipline, including the paucity of regulatory guidelines and precedents, evolving attitudes towards disease and clinical benefit, and the growing complexity of clinical prevention trial design and execution. Early scientific ventures often seek to achieve lofty goals, but are constrained and sometimes thwarted by biologic complexities. So too, cancer chemoprevention initially sought to reduce cancer incidence in low-risk populations, testing agents with low or no apparent toxicity. This first generation of trials, however well-intended, was in retrospect, naïve; yet it provided a critical foundation for iterative trial design and development. Subsequent generations of studies have favored the use of high-risk populations, mechanistic targeting, agent combinations, and where possible, early markers of neoplastic risk and response to applied interventions (Torrance et al., 2000; Sudbo et al., 2003). Molecular targeting studies have led to the concept of riskbased prevention and the convergence of cancer prevention with cancer therapy. The definitive trials of tamoxifen in the BCPT and finasteride in the PCPT clearly illustrate the complex risk-benefit profiles of preventive agents and the difficult issues and disease trade-offs that must be considered when deciding whether or not to use a cancer chemopreventive (Solomon et al., 2005; Hawk and Viner, 2003). Our responses to these uncertainties—as scientists, health-care providers and patients— reflect personal preferences, and clinical context, as well as entrenched medical beliefs. For example, a man considering low-dose aspirin for cardiac disease prevention now has another good reason, cancer risk reduction, to take low-dose aspirin. A man with benign prostatic hypertrophy (BPH) may be less concerned about potential sexual side effects when considering finasteride for prostate cancer risk reduction, since all BPH treatments have this toxicity. A woman with a hysterectomy (and, therefore, no endometrial cancer risk) is more likely than is a woman with her uterus to choose tamoxifen, which increases endometrial cancer risk, for breast cancer prevention. The complex benefit-risk profiles of tamoxifen and finasteride are further complicated by the finding that neither agent prevented what are presumed to be the most aggressive disease subsets—ER-negative breast or high-grade prostate cancers, respectively. Furthermore, it is possible that these agents select for hormone-independent, more aggressive cancers. To address agent limitations exposed by these studies, prevention research is heavily invested in pharmacogenomic studies that might better identify individuals most likely to benefit from a specific intervention. Recent data show tamoxifen’s activity against breast

cancer in BRCA2-positive women (King et al., 2001) and retinoic acid’s activity against head and neck cancer in patients with the cyclin D1 GG genotype (Izzo et al., 2003), which suggest the broad potential of pharmacogenomics for advancing cancer prevention.

CONCLUSION The convergence of cancer chemoprevention with cancer therapy, a concept promoted by advances in molecular biology and targeted therapy, requires new approaches to drug development. Novel phase I trials that can determine optimal doses/schedules of an agent for both prevention and therapy are needed. In addition, novel trial designs are needed to develop targeted combination prevention for high-risk cohorts. Regimens (i.e., routes, doses, and schedules) based upon strong preclinical and translational molecular/mechanistic data with aggressive safety monitoring can facilitate the rapid progress of promising combinations into definitive testing in high-risk settings. The polychronotropic (literally, “many times, many places”) nature of carcinogenesis may hold clues for innovations in early-phase agent identification. For example, early pathologic lesions (e.g., adenoma, aberrant crypt foci) are commonly found in patients diagnosed with more advanced disease (e.g., CRC), illustrating that genetic and epigenetic events foster different rates and manifestations of neoplasia within the at-risk epithelium. Because carcinogenesis smolders in various forms within the setting of cancer treatment, agent identification may be expedited by nesting chemopreventive end points within compatible therapeutic trials. This efficient strategy has the potential to hasten evaluations of promising chemopreventive agents, depending on their mechanism(s) of action, route of administration, and safety profiles. Despite great enthusiasm for promising new agents, previous studies have demonstrated that it is necessary to apply rigorous, scientifically sound approaches for developing new drugs. For example, tremendous and appropriate enthusiasm for aromatase inhibitors (AIs) in preventing breast cancer (based on secondary ATAC findings of an AI-associated reduction in contralateral breast cancer) has led to the suggestion to eliminate SERMs, with their complex risk-benefit profile, from the next generation of definitive breast cancer prevention trials. It would be naïve to assume that AIs, which as yet have little published side effect data, do not have side effects. AIs likely will have adverse effects on bone and lipid metabolism and unexpected toxicities that will be uncovered by large-scale, long-term definitive clinical testing. It would be unfortunate to move AIs into phase III testing that excludes SERMs, especially if AIs effectively reduce breast cancer risk, because then it would be impossible to compare their riskbenefit profiles with those of SERMs. Modest prevention effects on risk factors can have a cumulatively large effect on the public health. This principle has been demonstrated by cardiovascular disease prevention with blood pressure control, aspirin prophylaxis, and lipid management. For example, cholesterolreducing statins only produce 10%–25% reductions in LDL levels yet account for 30%–40% reductions in major cardiovascular events across the US population. Cancer prevention is amenable to the same cumulative effects. An intervention that reduces epithelial cancers by as little as 15% conceivably could prevent 73,000 cancer deaths annually (Gaziano et al., 1996). Preventive interventions also are expected to improve quality of life, latency to disease progression, frequency of screening and surveillance, and the odds of organ preservation. Although more controversial, good outcomes of cancer preventive agents may also be anticipated in other diseases (e.g., atherogenesis and neurodegenerative diseases) which share initiating or promoting mechanisms commonly found in carcinogenesis (Lippman and Hong, 2002). Indeed, certain classes of agents ultimately may be beneficial in preventing or delaying several chronic diseases of aging (e.g., NSAIDs for arthritis, CRC and, more speculatively, cancer morbidity and neurodegenerative disease (Lundholm et al., 1994); SERMs for osteoporosis and breast cancer (Cummings et al., 1999); statins for cardiovascular disease and CRC; and PPAR agonists for lipid homeostasis and cancer). The efficacy of these agents across disparate diseases points to a common molecular genesis, the targeting of which

Cancer Chemoprevention provides opportunities for preventive synergies. By realizing this untapped potential, prevention will move beyond the classic therapeutic “one disease-one drug” model and achieve broad public health gains. References Ahlquist DA, Shuber AP. 2002. Stool screening for colorectal cancer: Evolution from occult blood to molecular markers. Clin Chim Acta 315: 157–168. Ahlquist DA, Skoletsky JE, Boynton KA, et al. 2000. Colorectal cancer screening by detection of altered human DNA in stool: Feasibility of a multitarget assay panel [In Process Citation]. Gastroenterology 119: 1219–1227. Ahrendt SA, Halachmi S, Chow JT, et al. 1999. Rapid p53 sequence analysis in primary lung cancer using an oligonucleotide probe array. Proc Natl Acad Sci U S A 96:7382–7387. The Alpha-Tocopherol, Beta Carotene Cancer Prevention Study Group. 1994. The effect of vitamin E and beta carotene on the incidence of lung cancer and other cancers in male smokers. N Engl J Med 330:1029–1035. Akar A, Bulent Tastan H, Erbil H, Arca E, Kurumlu Z, Gur AR. 2001. Efficacy and safety assessment of 0.5% and 1% colchicine cream in the treatment of actinic keratoses. J Dermatolog Treat 12:199–203. Alberts DS, Dorr RT, Einspahr JG, et al. 2000. Chemoprevention of human actinic keratoses by topical 2- (difluoromethyl)-dl-ornithine. Cancer Epidemiol Biomarkers Prev 9:1281–1286. Alberts DS, Martinez ME, Roe DJ, et al. 2000. Lack of effect of a high-fiber cereal supplement on the recurrence of colorectal adenomas. Phoenix Colon Cancer Prevention Physicians’ Network. N Engl J Med 342: 1156–1162. Alvarez RD, Conner MG, Weiss H, et al. 2003. The efficacy of 9-cis-retinoic acid (aliretinoin) as a chemopreventive agent for cervical dysplasia: Results of a randomized double-blind clinical trial. Cancer Epidemiol Biomarkers Prev 12:114–119. American Society of Clinical Oncology policy statement update. 2003. Genetic testing for cancer susceptibility. J Clin Oncol 21:2397–2406. Anzano MA, Peer CW, Smith JM, et al. 1996. Chemoprevention of mammary carcinogenesis in the rat: Combined use of raloxifene and 9-cis-retinoic acid. J Natl Cancer Inst 88:123–125. Arnold AM, Browman GP, Levine MN, et al. 1992. The effect of the synthetic retinoid etretinate on sputum cytology: Results from a randomised trial. Br J Cancer 65:737–743. Baithun SI, Naase M, Blanes A, Diaz-Cano SJ. 2001. Molecular and kinetic features of transitional cell carcinomas of the bladder: Biological and clinical implications. Virchows Arch 438:289–297. Balmain A, Brown K, Akhurst RJ, Fee FM. 1988. Molecular analysis of chemical carcinogenesis in the skin. Br J Cancer Suppl 9:72–75. Baron JA, Beach M, Mandel JS, et al. 1999. Calcium supplements for the prevention of colorectal adenomas. Calcium Polyp Prevention Study Group. N Engl J Med 340:101–107. Baron JA, Cole BF, Sandler RS, et al. 2003. A randomized trial of aspirin to prevent colorectal adenomas. N Engl J Med 348:891–899. Baum M, Buzdar A, Cuzick J, et al. 2003. Anastrozole alone or in combination with tamoxifen versus tamoxifen alone for adjuvant treatment of postmenopausal women with early-stage breast cancer: Results of the ATAC (Arimidex, Tamoxifen Alone or in Combination) trial efficacy and safety update analyses. Cancer 98:1802–1810. Bavinck JN, Tieben LM, Van der Woude FJ, et al. 1995. Prevention of skin cancer and reduction of keratotic skin lesions during acitretin therapy in renal transplant recipients: A double-blind, placebo-controlled study. J Clin Oncol 13:1933–1938. Bell MC, Crowley-Nowick P, Bradlow HL, et al. 2000. Placebo-controlled trial of indole-3-carbinol in the treatment of CIN. Gynecol Oncol 78:123–129. Benamouzig R, Deyra J, Martin A, et al. 2003. Daily soluble aspirin and prevention of colorectal adenoma recurrence: One-year results of the APACC trial. Gastroenterology 125:328–336. Berg D, Otley CC. 2002. Skin cancer in organ transplant recipients: Epidemiology, pathogenesis, and management. J Am Acad Dermatol 47:1–17; quiz 18–20. Bergstrom R, Sparen P, Adami HO. 1999. Trends in cancer of the cervix uteri in Sweden following cytological screening. Br J Cancer 81:159–166. Bertram JS. 2000. The molecular biology of cancer. Mol Aspects Med 21:167–223. Bianchi MS, Bianchi NO, Bailliet G. 1995. Mitochondrial DNA mutations in normal and tumor tissues from breast cancer patients. Cytogenet Cell Genet 71:99–103. Bingham SA, Day NE, Luben R, et al. 2003. Dietary fibre in food and protection against colorectal cancer in the European Prospective Investigation

1335

into Cancer and Nutrition (EPIC): An observational study. Lancet 361:1496–1501. Blot WJ, Li JY, Taylor PR, et al. 1993. Nutrition intervention trials in Linxian, China: Supplementation with specific vitamin/mineral combinations, cancer incidence, and disease-specific mortality in the general population. J Natl Cancer Inst 85:1483–1492. Bocchia M, Bronte V, Colombo MP, et al. 2000. Antitumor vaccination: Where we stand. Haematologica 85:1172–1206. Bockmuhl U, Wolf G, Schmidt S, et al. 1998. Genomic alterations associated with malignancy in head and neck cancer. Head Neck 20:145–151. Boiko IV, Mitchell MF, Pandey DK, et al. 1997. DNA image cytometric measurement as a surrogate end point biomarker in a phase I trial of alphadifluoromethylornithine for cervical intraepithelial neoplasia. Cancer Epidemiol Biomarkers Prev 6:849–855. Bolla M, Lefur R, Ton Van J, et al. 1994. Prevention of second primary tumours with etretinate in squamous cell carcinoma of the oral cavity and oropharynx. Results of a multicentric double-blind randomised study. Eur J Cancer 6:767–772. Bonithon-Kopp C, Kronborg O, Giacosa A, Rath U, Faivre J. 2000. Calcium and fibre supplementation in prevention of colorectal adenoma recurrence: A randomised intervention trial. European Cancer Prevention Organisation Study Group. Lancet 356:1300–1306. Boone CW, Kelloff GJ. 1997. Biomarker end-points in cancer chemoprevention trials. IARC Sci Publ 142:273–280. Boynton RF, Blount PL, Yin J, et al. 1992. Loss of heterozygosity involving the APC and MCC genetic loci occurs in the majority of human esophageal cancers. Proc Natl Acad Sci U S A 89:3385–3388. Brooks JD, Weinstein M, Lin X, et al. 1998. CG island methylation changes near the GSTP1 gene in prostatic intraepithelial neoplasia. Cancer Epidemiol Biomarkers Prev 7:531–536. Brown MB, Marriott C, Martin GP. 1995. The effect of hyaluronan on the in vitro deposition of diclofenac within the skin. Int J Tissue React 17: 133–140. Burn J, Chapman PD, Bishop DT, Mathers J. 1998. Diet and cancer prevention: The concerted action polyp prevention (CAPP) studies. Proc Nutr Soc 57:183–186. Camma C, Giunta M, Andreone P, Craxi A. 2001. Interferon and prevention of hepatocellular carcinoma in viral cirrhosis: An evidence-based approach. J Hepatol 34:593–602. Carlson JA, Slominski A, Linette GP, Mihm MC Jr., Ross JS. 2003. Biomarkers in melanoma: Predisposition, screening and diagnosis. Expert Rev Mol Diagn 3:163–184. Carlson JA, Slominski A, Linette GP, Mihm MC Jr., Ross JS. 2003. Biomarkers in melanoma: Staging, prognosis and detection of early metastases. Expert Rev Mol Diagn 3:303–330. Chang MH, Chen CJ, Lai MS, et al. 1997. Universal hepatitis B vaccination in Taiwan and the incidence of hepatocellular carcinoma in children. Taiwan Childhood Hepatoma Study Group. N Engl J Med 336:1855– 1859. Chen M, Sallberg M, Sonnerborg A, et al. 1999. Limited humoral immunity in hepatitis C virus infection. Gastroenterology 116:135–143. Childers JM, Chu J, Voigt LF, et al. 1995. Chemoprevention of cervical cancer with folic acid: A phase III Southwest Oncology Group Intergroup study. Cancer Epidemiol Biomarkers Prev 4:155–159. Clark LC, Combs GF Jr., Turnbull BW, et al. 1996. Effects of selenium supplementation for cancer prevention in patients with carcinoma of the skin. A randomized controlled trial. Nutritional Prevention of Cancer Study Group [published erratum appears in JAMA 1997 May 21;277(19):1520]. JAMA 276:1957–1963. Clark LC, Dalkin B, Krongrad A, et al. 1998. Decreased incidence of prostate cancer with selenium supplementation: Results of a double-blind cancer prevention trial. Br J Urol 81:730–734. Collisson EA, Kleer C, Wu M, et al. 2003. Atorvastatin prevents RhoC isoprenylation, invasion, and metastasis in human melanoma cells. Mol Cancer Ther 2:941–948. Comerci JT Jr., Runowicz CD, Fields AL, et al. 1997. Induction of transforming growth factor beta-1 in cervical intraepithelial neoplasia in vivo after treatment with beta-carotene. Clin Cancer Res 3:157–160. Correa P, Fontham ET, Bravo JC, et al. 2000. Chemoprevention of gastric dysplasia: Randomized trial of antioxidant supplements and antihelicobacter pylori therapy. J Natl Cancer Inst 92:1881–1888. Croteau DL, Stierum RH, Bohr VA. 1999. Mitochondrial DNA repair pathways. Mutation Research/DNA Repair 434:137–148. Cummings SR, Duong T, Kenyon E, Cauley JA, Whitehead M, Krueger KA. 2002. Serum estradiol level and risk of breast cancer during treatment with raloxifene. JAMA 287:216–220. Cummings SR, Eckert S, Krueger KA, et al. 1999. The effect of raloxifene on risk of breast cancer in postmenopausal women: results from the MORE

1336

PART V: CANCER PREVENTION AND CONTROL

randomized trial. Multiple Outcomes of Raloxifene Evaluation. JAMA 281:2189–2197. Cuzick J. 2003. Aromatase inhibitors in prevention—data from the ATAC (arimidex, tamoxifen alone or in combination) trial and the design of IBIS-II (the second International Breast Cancer Intervention Study). Recent Results Cancer Res 163:96–103; discussion 264–266. Cuzick J, Forbes J, Edwards R, et al. 2002. First results from the International Breast Cancer Intervention Study (IBIS-I): A randomised prevention trial. Lancet 360:817–824. De Marzo AM, Meeker AK, Zha S, et al. 2003. Human prostate cancer precursors and pathobiology. Urology 62:55–62. del Rosso JQ. 2003. An update on newer topical therapies for actinic keratoses: Advances and applications. J Drugs Dermatol 2:35–39. Dellavalle RP, Nicholas MK, Schilling LM. 2003. Melanoma chemoprevention: A role for statins or fibrates? Am J Ther 10:203–210. Desmond RA, Soong SJ. 2003. Epidemiology of malignant melanoma. Surg Clin North Am 83:1–29. Duffield-Lillico AJ, Reid ME, Turnbull BW, et al. 2002. Baseline characteristics and the effect of selenium supplementation on cancer incidence in a randomized clinical trial: A summary report of the Nutritional Prevention of Cancer Trial. Cancer Epidemiol Biomarkers Prev 11: 630–639. Duffield-Lillico AJ, Slate EH, Reid ME, et al. 2003. Selenium supplementation and secondary prevention of nonmelanoma skin cancer in a randomized trial. J Natl Cancer Inst 95:1477–1481. Einspahr JG, Alberts DS, Gapstur SM, Bostick RM, Emerson SS, Gerner EW. 1997. Surrogate end-point biomarkers as measures of colon cancer risk and their use in cancer chemoprevention trials. Cancer Epidemiol Biomarkers Prev 6:37–48. Einspahr JG, Stratton SP, Bowden GT, Alberts DS. 2002. Chemoprevention of human skin cancer. Crit Rev Oncol Hematol 41:269–285. El-Serag HB, Davila JA, Petersen NJ, McGlynn KA. 2003. The continuing increase in the incidence of hepatocellular carcinoma in the United States: An update. Ann Intern Med 139:817–823. Engel LS, Chow WH, Vaughan TL, et al. 2003. Population attributable risks of esophageal and gastric cancers. J Natl Cancer Inst 95:1404–1413. Farrow DC, Vaughan TL, Hansten PD, et al. 1998. Use of aspirin and other nonsteroidal anti-inflammatory drugs and risk of esophageal and gastric cancer. Cancer Epidemiol Biomarkers Prev 7:97–102. Fisher B, Costantino JP, Wickerham DL, et al. 1998. Tamoxifen for prevention of breast cancer: Report of the National Surgical Adjuvant Breast and Bowel Project P-1 Study. J Natl Cancer Inst 90:1371–1388. Fisher B, Dignam J, Wolmark N, et al. 1999. Tamoxifen in treatment of intraductal breast cancer: National Surgical Adjuvant Breast and Bowel Project B-24 randomised controlled trial. Lancet 353:1993–2000. Fliss MS, Usadel H, Caballero OAVL, et al. 2000. Facile detection of mitochondrial DNA mutations in tumors and bodily fluids. Science 287: 2017–2019. Ford D, Easton DF, Stratton M, et al. 1998. Genetic heterogeneity and penetrance analysis of the BRCA1 and BRCA2 genes in breast cancer families. The Breast Cancer Linkage Consortium. Am J Hum Genet 62: 676–689. Frieling UM, Schaumberg DA, Kupper TS, Muntwyler J, Hennekens CH. 2000. A randomized, 12-year primary-prevention trial of beta carotene supplementation for nonmelanoma skin cancer in the physician’s health study. Arch Dermatol 136:179–184. Gail MH, Brinton LA, Byar DP, et al. 1989. Projecting individualized probabilities of developing breast cancer for white females who are being examined annually. J Natl Cancer Inst 81:1879–1886. Gann PH, Manson JE, Glynn RJ, Buring JE, Hennekens CH. 1993. Low-dose aspirin and incidence of colorectal tumors in a randomized trial. J Natl Cancer Inst 85:1220–1224. Gayther SA, de Foy KA, Harrington P, et al. 2000. The frequency of germ-line mutations in the breast cancer predisposition genes BRCA1 and BRCA2 in familial prostate cancer. The Cancer Research Campaign/British Prostate Group United Kingdom Familial Prostate Cancer Study Collaborators. Cancer Res 60:4513–4518. Gaziano JM, Hennekens CH, Buring JE. 1996. Methodological issues in the design of primary prevention trials. IARC Sci Publ 136:7–12. Giardiello FM, Brensinger JD, Petersen GM. 2001. AGA technical review on hereditary colorectal cancer and genetic testing. Gastroenterology 121: 198–213. Giardiello FM, Hamilton SR, Krush AJ, et al. 1993. Treatment of colonic and rectal adenomas with sulindac in familial adenomatous polyposis. N Engl J Med 328:1313–1316. Giardiello FM, Yang VW, Hylind LM, et al. 2002. Primary chemoprevention of familial adenomatous polyposis with sulindac. N Engl J Med 346: 1054–1109.

Gibson GE, O’Grady A, Kay EW, Murphy GM. 1998. Low-dose retinoid therapy for chemoprophylaxis of skin cancer in renal transplant recipients. J Eur Acad Dermatol Venereol 10:42–47. Gohagan JK, Prorok PC, Kramer BS, Hayes RB, Cornett JE. 1995. The Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial of the National Cancer Institute. Cancer 75:1869–1873. Goldman JM, Melo JV. 2003. Chronic myeloid leukemia—advances in biology and new approaches to treatment. N Engl J Med 349:1451–1464. Gooderham NJ, Carmichael PL. 2002. Mechanisms of Chemical Carcinogenesis. The Cancer Handbook. M. R. Alison. London, Nature Publishing Group 1:272–291. Green RJ, Metlay JP, Propert K, et al. 2002. Surveillance for second primary colorectal cancer after adjuvant chemotherapy: An analysis of Intergroup 0089. Ann Intern Med 136:261–269. Greenberg ER, Baron JA, Stukel TA, et al. 1990. A clinical trial of beta carotene to prevent basal-cell and squamous-cell cancers of the skin. The Skin Cancer Prevention Study Group. N Engl J Med 323:789–795. Greenberg ER, Baron JA, Tosteson TD, et al. 1994. A clinical trial of antioxidant vitamins to prevent colorectal adenoma. Polyp Prevention Study Group N Engl J Med 331:141–147. Gupta AK. 2002. The management of actinic keratoses in the United States with topical fluorouracil: A pharmacoeconomic evaluation. Cutis 70: 30–36. Gupta S, Mukhtar H. 2002. Chemoprevention of skin cancer: Current status and future prospects. Cancer Metastasis Rev 21:363–380. Habano W, Sugai T, Nakamura SI, Uesugi N, Yoshida T, Sasou S. 2000. Microsatellite instability and mutation of mitochondrial and nuclear DNA in gastric carcinoma. Gastroenterology 118:835–841. Halpern AC, Schuchter LM, Elder DE, et al. 1994. Effects of topical tretinoin on dysplastic nevi. J Clin Oncol 12:1028–1035. Hanahan D, Weinberg RA. 2000. The hallmarks of cancer. Cell 100:57–70. Hartman TJ, Albanes D, Pietinen P, et al. 1998. The association between baseline vitamin E, selenium, and prostate cancer in the alpha-tocopherol, beta-carotene cancer prevention study. Cancer Epidemiol Biomarkers Prev 7:335–340. Hawk E, Lubet R, Limburg P. 1999. Chemoprevention in hereditary colorectal cancer syndromes. Cancer 86:2551–2563. Hawk E, Viner JL, Lawrence JA. 2000. Biomarkers as surrogates for cancer development. Curr Oncol Reports 2:242–250. Hawk E, Viner JL. 2005. The critical role of risk-benefit assessments in cancer prevention. Cancer Epidemiol Biomarkers Prev 14:297–298. Hawk ET, Limburg PJ, Viner JL. 2002. Epidemiology and prevention of colorectal cancer. Surg Clin North Am 82:905–941. Hawk ET, Viner JL, Umar A, Anderson WF, Sigman CC, Guyton KZ. 2003. Cancer and the cyclooxygenase enzyme: Implications for the treatment and prevention. Am J Cancer 2:27–55. Heinonen OP, Albanes D, Virtamo J, et al. 1998. Prostate cancer and supplementation with alpha-tocopherol and beta-carotene: Incidence and mortality in a controlled trial. J Natl Cancer Inst 90:440–446. Hennekens CH, Buring JE, Manson JE, et al. 1996. Lack of effect of long-term supplementation with beta carotene on the incidence of malignant neoplasms and cardiovascular disease. N Engl J Med 334:1145–1149. Hirota C, Iida M, Aoyagi K, et al. 1996. Effect of indomethacin suppositories on rectal polyposis in patients with familial adenomatous polyposis. Cancer 78:1660–1665. Hirsch FR, Prindiville SA, Miller YE, et al. 2001. Fluorescence versus whitelight bronchoscopy for detection of preneoplastic lesions: A randomized study. J Natl Cancer Inst 93:1385–1391. Hollstein M, Bartsch H, Wesch H, et al. 1997. p53 gene mutation analysis in tumors of patients exposed to alpha-particles. Carcinogenesis 18:511–516. Hollstein M, Rice K, Greenblatt MS, et al. 1994. Database of p53 gene somatic mutations in human tumors and cell lines. Nucleic Acids Res 22: 3551–3555. Hong WK, Endicott J, Itri LM, et al. 1986. 13-cis-retinoic acid in the treatment of oral leukoplakia. N Engl J Med 315:1501–1505. Hong WK, Lippman SM, Itri LM, et al. 1990. Prevention of second primary tumors with isotretinoin in squamous-cell carcinoma of the head and neck. N Engl J Med 323:795–801. Horii A, Nakatsuru S, Miyoshi Y, et al. 1992. Frequent somatic mutations of the APC gene in human pancreatic cancer. Cancer Res 52:6696–6698. Huang Y, Boynton RF, Blount PL, et al. 1992. Loss of heterozygosity involves multiple tumor suppressor genes in human esophageal cancers. Cancer Res 52:6525–6530. Hursting SD. 1997. Experimental models of gene-environment interaction for cancer chemoprevention studies. Curr Opin Oncol 9:487–491. Ionov Y, Peinado MA, Malkhosyan S, Shibata D, Perucho M. 1993. Ubiquitous somatic mutations in simple repeated sequences reveal a new mechanism for colonic carcinogenesis. Nature 363:558–561.

Cancer Chemoprevention Izzo JG, Papadimitrakopoulou VA, Liu DD, et al. 2003. Cyclin D1 genotype, response to biochemoprevention, and progression rate to upper aerodigestive tract cancer. J Natl Cancer Inst 95:198–205. Jacoby RF, Cole CE, Tutsch K, et al. 2000. Chemopreventive efficacy of combined piroxicam and difluoromethylornithine treatment of Apc mutant Min mouse adenomas, and selective toxicity against Apc mutant embryos. Cancer Res 60:1864–1870. Jarvis B, Figgitt DP. 2003. Topical 3% diclofenac in 2.5% hyaluronic acid gel: A review of its use in patients with actinic keratoses. Am J Clin Dermatol 4:203–213. Jeffes EW. 2002. Levulan: The first approved topical photosensitizer for the treatment of actinic keratosis. J Dermatolog Treat 13 Suppl 1:S19–S23. Jemal A, Murray T, Ward E, et al. 2005. Cancer statistics. CA Cancer J Clin 55:10–30. Kamat AM, Lamm DL. 1999. Chemoprevention of urological cancer. J Urol 161:1748–1760. Kavanaugh CJ, Desai KV, Calvo A, et al. 2002. Pre-clinical applications of transgenic mouse mammary cancer models. Transgenic Res 11:617–633. Kelloff GJ, Johnson JR, Crowell JA, et al. 1995. Approaches to the development and marketing approval of drugs that prevent cancer. Cancer Epidemiol Biomarkers Prev 4:1–10. Kelloff GJ, Sigman CC. 2002. Intervention and Chemoprevention of Cancer. The Cancer Handbook. M. R. Alison. London, Nature Publishing Group 1:435–458. Kelty CJ, Marcus SL, Ackroyd R. 2002. Photodynamic therapy for Barrett’s esophagus: A review. Dis Esophagus 15:137–144. Kensler TW, He X, Otieno M, et al. 1998. Oltipraz chemoprevention trial in Qidong, People’s Republic of China: Modulation of serum aflatoxin albumin adduct biomarkers. Cancer Epidemiol Biomarkers Prev 7: 127–134. Khuri F, Lee JJ, Lippman SM, et al. 2003. Isotretinoin effects on head and neck cancer recurrence and second primary tumors. Proc ASCO 22:90, abstract #359. Khuri FR, Kim ES, Lee JJ, et al. 2001. The impact of smoking status, disease stage, and index tumor site on second primary tumor incidence and tumor recurrence in the head and neck retinoid chemoprevention trial. Cancer Epidemiol Biomarkers Prev 10:823–829. Kim SH, Park TK, Kwon JY. 2003. 13-cis-retinoic acid for chemoprevention after colpoconization for cervical intraepithelial neoplasia. Int J Gynaecol Obstet 81:217–218. King MC, Wieand S, Hale K, et al. 2001. Tamoxifen and breast cancer incidence among women with inherited mutations in BRCA1 and BRCA2: National Surgical Adjuvant Breast and Bowel Preject (NSABP-P1) Breast Cancer Prevention Trial. JAMA 286:2251–2256. Klein EA, Thompson IM, Lippman SM, et al. 2003. SELECT: The selenium and vitamin E cancer prevention trial. Urol Oncol 21:59–65. Koutsky LA, Ault KA, Wheeler CM, et al. 2002. A controlled trial of a human papillomavirus type 16 vaccine. N Engl J Med 347:1645–1651. Kowaltowski AJ, Vercesi AE. 1999. Mitochondrial damage induced by conditions of oxidative stress. Free Radic Biol Med 26:463–471. Kraemer KH, DiGiovanna JJ, Moshell AN, Tarone RE, Peck GL. 1988. Prevention of skin cancer in xeroderma pigmentosum with the use of oral isotretinoin. N Engl J Med 318:1633–1637. Kroemer G, Reed JC. 2000. Mitochondrial control of cell death. Nat Med 6:513–519. Kurie JM, Lee JS, Khuri FR, et al. 2000. N-(4-hydroxyphenyl)retinamide in the chemoprevention of squamous metaplasia and dysplasia of the bronchial epithelium. Clin Cancer Res 6:2973–2979. Kurie JM, Lotan R, Lee JJ, et al. 2003. Treatment of former smokers with 9cis-retinoic acid reverses loss of retinoic acid receptor-beta expression in the bronchial epithelium: Results from a randomized placebo-controlled trial. J Natl Cancer Inst 95:206–214. Labayle D, Fischer D, Vielh P, et al. 1991. Sulindac causes regression of rectal polyps in familial adenomatous polyposis. Gastroenterology 101:635– 639. Lakhani SR, Jacquemier J, Sloane JP, et al. 1998. Multifactorial analysis of differences between sporadic breast cancers and cancers involving BRCA1 and BRCA2 mutations. J Natl Cancer Inst 90:1138–1145. Lam S, Kennedy T, Unger M, et al. 1998. Localization of bronchial intraepithelial neoplastic lesions by fluorescence bronchoscopy. Chest 113: 696–702. Lam S, MacAulay C, Le Riche JC, et al. 2002. A randomized phase IIb trial of anethole dithiolethione in smokers with bronchial dysplasia. J Natl Cancer Inst 94:1001–1009. Lamm DL. 2000. Effecacy and safety of bacille Calmette-Guerin immunotherapy in superficial bladder cancer. Clin Infect Dis 31 Suppl 3:S86–S90. Lamm DL, Blumenstein BA, Crissman JD, et al. 2000. Maintenance bacillus Calmette-Guerin immunotherapy for recurrent TA, T1 and carcinoma in

1337

situ transitional cell carcinoma of the bladder: A randomized Southwest Oncology Group Study. J Urol 163:1124–1129. Lamm DL, Riggs DR, Shriver JS, vanGilder PF, Rach JF, DeHaven JI. 1994. Megadose vitamins in bladder cancer: A double-blind clinical trial. J Urol 151:21–26. Lang K, Schulte KW, Ruzicka T, Fritsch C. 2001. Aminolevulinic acid (Levulan) in photodynamic therapy of actinic keratoses. Skin Therapy Lett 6:1–2, 5. Lee CL, Hsieh KS, Ko YC. 2003. Trends in the incidence of hepatocellular carcinoma in boys and girls in Taiwan after large-scale hepatitis B vaccination. Cancer Epidemiol Biomarkers Prev 12:57–59. Lee JS, Lippman SM, Benner SE, et al. 1994. Randomized placebo-controlled trial of isotretinoin in chemoprevention of bronchial squamous metaplasia. J Clin Oncol 12:937–945. Levin B. 2003. Potential pitfalls in the use of surrogate endpoints in colorectal adenoma chemoprevention. J Natl Cancer Inst 95:697–699. Levine AJ. 1997. p53, the cellular gatekeeper for growth and division. Cell 88:323–331. Levine N, Moon TE, Cartmel B, et al. 1997. Trial of retinol and isotretinoin in skin cancer prevention: A randomized. double-blind, controlled trial. Southwest Skin Cancer Prevention Study Group. Cancer Epidemiol Biomarkers Prev 6:957–961. Li H, Schut HA, Conran P, et al. 1999. Prevention by aspirin and its combination with alpha- difluoromethylornithine of azoxymethane-induced tumors, aberrant crypt foci and prostaglandin E2 levels in rat colon. Carcinogenesis 20:425–430. Li JY, Taylor PR, Li B, et al. 1993. Nutrition intervention trials in Linxian, China: Multiple vitamin/mineral supplementation, cancer incidence, and disease-specific mortality among adults with esophageal dysplasia. J Natl Cancer Inst 85:1492–1498. Liang TJ, Rehermann B, Seeff LB, Hoofnagle JH. 2000. Pathogenesis, natural history, treatment, and prevention of hepatitis C. Ann Intern Med 132:296–305. Limburg PJ, Wei W, Ahnen DJ, et al. 2005. Randomized, placebo-controlled, esophageal squamous cell cancer chemoprevention trial of selenomethionine and celecoxib. Gastroenterology 129:863–873. Lin SM, Sheen IS, Chien RN, Chu CM, Liaw YF. 1999. Long-term beneficial effect of interferon therapy in patients with chronic hepatitis B virus infection. Hepatology 29:971–975. Lipkin M, Yang K, Edelmann W, et al. 1999. Preclinical mouse models for cancer chemoprevention studies. Ann N Y Acad Sci 889:14–19. Lippman SM, Batsakis JG, Toth BB, et al. 1993. Comparison of low-dose isotretinoin with beta carotene to prevent oral carcinogenesis. N Engl J Med 328:15–20. Lippman SM, Benner SE, Hong WK. 1994. Retinoid chemoprevention studies in upper aerodigestive tract and lung carcinogenesis. Cancer Res 54:2025s–2028s. Lippman SM, Hong WK. 2002. Cancer prevention by delay. Commentary re: J. A. O’Shaughnessy et al., Treatment and prevention of intraepithelial neoplasia: An important target for accelerated new agent development. Clin Cancer Res 8:314–346, 2002. Clin Cancer Res 8:305–313. Lippnam SM, Hong WK. 2002. Cancer prevention science and practice. Cancer Res 62:5119–5125. Lippnam SM, Lee JJ, Karp DD, et al. 2001. Randomized phase III intergroup trial of isotretinoin to prevent second primary tumors in stage I non-smallcell lung cancer. J Natl Cancer Inst 93:605–618. Lippman SM, Lee JJ, Sabichi AL. 1998. Cancer chemoprevention: Progress and promise. J Natl Cancer Inst 90:1514–1528. Loprinzi CL, Messing EM. 1992. A prospective clinical trial of difluoromethylornithine (DFMO) in patients with resected superficial bladder cancer. J Cell Biochem Suppl 16I:153–155. Lundholm K, Gelin J, Hyltander A, et al. 1994. Anti-inflammatory treatment may prolong survival in undernourished patients with metastatic solid tumors. Cancer Res 54:5602–5606. Lynch HT, Lynch J. 2000. Lynch syndrome: Genetics, natural history, genetic counseling, and prevention. J Clin Oncol 18:19S–31S. Macgregor PF. 2003. Gene expression in cancer: The application of microarrays. Expert Rev Mol Diagn 3:185–200. MacLennan R, Macrae F, Bain C, et al. 1995. Randomized trial of intake of fat, fiber, and beta carotene to prevent colorectal adenomas. The Australian Polyp Prevention Project. J Natl Cancer Inst 87:1760–1766. Mandel JS, Bond JH, Church TR, et al. 1993. Reducing mortality from colorectal cancer by screening for fecal occult blood. Minnesota Colon Cancer Control Study [published erratum appears in N Engl J Med 1993 Aug 26;329(9):672] N Engl J Med 328:1365–1371. Mandel JS, Church TR, Bond JH, et al. 2000. The effect of fecal occult-blood screening on the incidence of colorectal cancer. N Engl J Med 343: 1603–1607.

1338

PART V: CANCER PREVENTION AND CONTROL

Mark SD, Liu SF, Li JY, et al. 1994. The effect of vitamin and mineral supplementation on esophageal cytology: Results from the Linxian Dysplasia Trial. Int J Cancer 57:162–166. Marsh D, Zori R. 2002. Genetic insights into familial cancers—update and recent discoveries. Cancer Lett 181:125–164. Martinez ME, Willett WC. 1998. Calcium, vitamin D, and colorectal cancer: A review of the epidemiologic evidence. Cancer Epidemiol Biomarkers Prev 7:163–168. McKenna DB, Murphy GM. 1999. Skin cancer chemoprophylaxis in renal transplant recipients: 5 years of experience using low-dose acitretin. Br J Dermatol 140:656–660. McLarty JW, Holiday DB, Girard WM, Yanagihara RH, Kummet TD, Greenberg SD. 1995. Beta-Carotene, vitamin A, and lung cancer chemoprevention: Results of an intermediate endpoint study. Am J Clin Nutr 62:1431S–1438S. Meyskens FL, Surwit E, Moon TE, et al. 1994. Enhancement of regression of cervical intraepithelial neoplasia II (moderate dysplasia) With topically applied all-trans-retinoic acid: A randomized trial. J Natl Cancer Inst 86:539–543. Minamoto T, Mai M, Ronai Z. 2000. K-ras mutation: Early detection in molecular diagnosis and risk assessment of colorectal, pancreas, and lung cancers—a review. Cancer Detect Prev 24:1–12. Mitsuuchi Y, Testa JR. 2002. Cytogenetics and molecular genetics of lung cancer. Am J Med Genet 115:183–188. Miyaki M, Iijima T, Kimura J, et al. 1999. Frequent mutation of betacatenin and APC genes in primary colorectal tumors from patients with hereditary nonpolyposis colorectal cancer. Cancer Res 59:4506– 4509. Miyaki M, Konishi M, Kikuchi-Yanoshita R, et al. 1994. Characteristics of somatic mutation of the adenomatous polyposis coli gene in colorectal tumors. Cancer Res 54:3011–3020. Moon TE, Levine N, Cartmel B, et al. 1997. Effect of retinal in preventing squamous cell skin cancer in moderate-risk subjects: A randomized, double-blind, controlled trial. Southwest Skin Cancer Prevention Study Group. Cancer Epidemiol Biomarkers Prev 6:949–956. Moyad MA. 2003. Bladder cancer prevention. Part I: What do I tell my patients about lifestyle changes and dietary supplements? Curr Opin Urol 13: 363–378. Mulshine JL, Atkinson J, Greer RO, et al. 2003. Randomized, double blind, placebo controlled phase II study of ketorolac tromethamine rinse for the treatment of oral leukoplakia. Proc ASCO 22:91, abstract #365. Munoz N, Wahrendorf J, Bang LJ, et al. 1985. No effect of riboflavine, retinal, and zinc on prevalence of precancerous lesions of oesophagus. Randomised double-blind intervention study in high-risk population of China. Lancet 2:111–114. Muto Y, Moriwaki H, Ninomiya M, et al. 1996. Prevention of second primary tumors by an acyclic retinoid, polyprenoic acid, in patients with hepatocellular carcinoma. Hepatoma Prevention study Group. N Engl J Med 334:1561–1567. Nagase H, Miyoshi Y, Horii A, et al. 1992. Correlation between the location of germ-line mutations in the APC gene and the number of colorectal polyps in familial adenomatous polyposis patients. Cancer Res 52: 4055–4077. Nakatsuru S, Yanagisawa A, Furukawa Y, et al. 1993. Somatic mutations of the APC gene in precancerous lesion of the stomach. Hum Mol Genet 2:1463–1465. Narisawa T, Kohno KI, Yamaguchi T, Takahashi T. 1980. Chemoprevention of development of colonic adenomatosis and carcinomatosis with intrarectal dose of 5-FU on animal model. Cancer 45:439–443. Narisawa T, Sato M, Tani M, Takahashi T. 1980. Chemoprevention of development of colonic adenomatosis and carcinomatosis with 5-fluorouracil and ftorafur on animal model. J Cancer Res Clin Oncol 97:223–232. Nelson WG, De Marzo AM, Isaacs WB. 2003. Prostate cancer. N Engl J Med 349:366–381. Newberne PM, Rogers AE. 1973. Animal model of human disease. Adenocarcinoma of the colon. Am J Pathol 72:541–544. Nigro ND, Bull AW, Boyd ME. 1986. Inhibition of intestinal carcinogenesis in rats: Effect of difluoromethylornithine with piroxicam or fish oil. J Natl Cancer Inst 77:1309–1313. Nishiguchi S, Kuroki T, Nakatani S, et al. 1995. Randomised trial of effects of interferon-alpha on incidence of hepatocellular carcinoma in chronic active hepatitis C with cirrhosis. Lancet 346:1051–1055. Norgauer J, Idzko M, Panther E, Hellstern O, Herouy Y. 2003. Xeroderma pigmentosum. Eur J Dermatol 13:4–9. Nugent KP, Farmer KC, Spigelman AD, Williams CB, Phillips RK. 1993. Randomized controlled trial of the effect of sulindac on duodenal and rectal polyposis and cell proliferation in patients with familial adenomatous polyposis. Br J Surg 80:1618–1619.

Nutting CM, Huddart RA. 2001. Retinoids in the prevention of bladder cancer. Expert Rev Anticancer Ther 1:541–545. Omenn GS, Goodman GE, Thornquist MD, et al. 1996. Risk factors for lung cancer and for intervention effects in CARET, the Beta-Carotene and Retinol Efficacy Trial. J Natl Cancer Inst 88:1550–1559. Omenn GS, Goodman GE, Thornquist MD, et al. 1996. Effects of a combination of beta carotene and vitamin A on lung cancer and cardiovascular disease. N Engl J Med 334:1150–1155. O’Shaughnessy JA, Kelloff GJ, Gordon GB, et al. 2002. Treatment and prevention of intraepithelial neoplasia: An important target for accelerated new agent development. Clin Cancer Res 8:314–346. Oshima M, Dinchuk JE, Kargman SL, et al. 1996. Suppression of intestinal polyposis in Apc delta716 knockout mice by inhibition of cyclooxygenase 2 (COX-2). Cell 87:803–809. Papatheodoridis GV, Papadimitropoulos VC, Hadziyannis SJ. 2001. Effect of interferon therapy on the development of hepatocellular carcinoma in patients with hepatitis C virus-related cirrhosis: A meta-analysis. Aliment Pharmacol Ther 15:689–698. Parkin DM, Bray F, Ferlay J, Pisani P. 2005. Global cancer statistics, 2002. CA Cancer J Clin 55:74–108. Parkin DM, Pisani P, Ferlay J. 1999. Estimates of the worldwide incidence of 25 major cancers in 1990. Int J Cancer 80:827–841. Parsons R, Li GM, Longley M, et al. 1995. Mismatch repair deficiency in phenotypically normal human cells. Science 268:738–740. Pastorino U, Infante M, Maioli M, et al. 1993. Adjuvant treatment of stage I lung cancer with high-dose vitamin A. J Clin Oncol 11:1216–1222. Peltomaki P. 2003. Role of DNA mismatch repair defects in the pathogenesis of human cancer. J Clin Oncol 21:1174–1179. Pence BC. 1993. Role of calcium in colon cancer prevention: experimental and clinical studies. Mutat Res 290:87–95. Peters DC, Foster RH. 1999. Diclofenac/hyaluronic acid. Drugs Aging 14: 313–319; discussion 320–321. Peterson SR, Goldberg LH. 2003. New and emerging treatments for nonmelanomas and actinic keratoses. J Drugs Dermatol 2:429–432. Pinto LA, Edwards J, Castle PE, et al. 2003. Cellular immune responses to human papillomavirus (HPV)-16 L1 in healthy volunteers immunized with recombinant HPV-16 L1 virus-like particles. J Infect Dis 188:327–338. Polakis P, Hart M, Rubinfeld B. 1999. Defects in the regulation of beta-catenin in colorectal cancer. Adv Exp Med Biol 470:23–32. Polyak K, Li Y, Zhu H, et al. 1998. Somatic mutations of the mitochondrial genome in human colorectal tumours. Nat Genet 20:291–293. Powles T, Eeles R, Ashley S, et al. 1998. Interim analysis of the incidence of breast cancer in the Royal Marsden Hospital tamoxifen randomised chemoprevention trial. Lancet 352:98–101. Rao CV, Tokumo K, Rigotty J, Zang E, Kelloff G, Reddy BS. 1991. Chemoprevention of colon carcinogenesis by dietary administration of piroxicam, alpha-difluoromethylornithine, 16 alpha-fluoro-5androsten- 17-one, and ellagic acid individually and in combination. Cancer Res 51:4528–4534. Reddy BS, Narisawa T, Maronpot R, Weisburger JH, Wynder EL. 1975. Animal models for the study of dietary factors and cancer of the large bowel. Cancer Res 35:3421–3426. Reddy BS, Nayini J, Tokumo K, Rigotty J, Zang E, Kelloff G. 1990. Chemoprevention of colon carcinogenesis by concurrent administration of piroxicam, a nonsteroidal antiinflammatory drug with D,Lalpha-difluoromethylornithine, an ornithine decarboxylase inhibitor, in diet. Cancer Res 50:2562–2568. Risinger JI, Berchuck A, Kohler MF, Watson P, Lynch HT, Boyd J. 1993. Genetic instability of microsatellites in endometrial carcinoma. Cancer Res 53:5100–5103. Ronai Z, Minamoto T. 1997. Quantitative enriched PCR (QEPCR), a highly sensitive method for detection of K-ras oncogene mutation. Hum Mutat 10:322–325. Ronai Z, Minamoto T, Butler R, et al. 1995. Sampling method as a key factor in identifying K-ras oncogene mutations in preneoplastic colorectal lesions. Cancer Detect Prev 19:512–517. Rook AH, Jaworsky C, Nguyen T, Grossman RA, Wolfe JT, Witmer WK, Kligman AM. 1995. Beneficial effect of low-dose systemic retinoid in combination with topical tretinoin for the treatment and prophylaxis of premalignant and malignant skin lesions in renal transplant recipients. Transplantation 59:714–719. Ross SD, Allen IE, Connelly JE, et al. 1999. Clinical outcomes in statin treatment trials: A meta-analysis. Arch Intern Med 159:1793–1802. Rubinfeld B, Souza B, Albert I, et al. 1993. Association of the APC gene product with beta-catenin. Science 262:1731–1734. Ruffin MT, Bailey JM, Normolle DP, et al. 2003. Topical all-trans retinoic acid in high grade squamous intraepithelial lesions. Proc AACR 44:1301, abstract #5677.

Cancer Chemoprevention Ruidi C, Aihua D, Peiyu B, et al. 1997. Chemoprevention of cancer of uterine cervix: A study on chemoprevention of retinamide II from cervical precancerous lesions. J Cell Biochem Suppl 29:140–143. Sabichi AL, Lippman SM. 2003. COX-2 inhibitors and other NSAIDs in bladder and prostate cancer. Prog Exp Tumor Res 37:163–178. Sandler RS, Halabi S, Baron JA, et al. 2003. A randomized trial of aspirin to prevent colorectal adenomas in patients with previous colorectal cancer. N Engl J Med 348:883–890. Santhanam S, Decatris M, O’Byrne K. 2002. Potential of interferon-alpha in solid tumours: Part 2. BioDrugs 16:349–372. Satoh M, Kuroiwa T. 1991. Organization of multiple nucleoids and DNA molecules in mitochondria of a human cell. Exp Cell Res 196:137–140. Schatzkin A, Lanza E, Corle D, et al. 2000. Lack of effect of a low-fat, highfiber diet on the recurrence of colorectal adenomas. Polyp Prevention Trial Study Group. N Engl J Med 342:1149–1155. Schiffman M, Solomon D. 2003. Findings to date from the ASCUS-LSIL Triage Study (ALTS). Arch Pathol Lab Med 127:946–949. Schreinemachers DM, Everson RB. 1994. Aspirin use and lung, colon, and breast cancer incidence in a prospective study. Epidemiology 5:138–146. Selvanayagam P, Rajaraman S. 1996. Detection of mitochondrial genome depletion by a novel CDNA in renal cell carcinoma. Lab Invest 74:592–599. Serewko MM, Popa C, Dahler AL, et al. 2002. Alterations in gene expression and activity during squamous cell carcinoma development. Cancer Res 62:3759–3765. Shelley MD, Kynaston H, Court J, et al. 2001. A systematic review of intravesical bacillus Calmette-Guerin plus transurethral resection vs transurethral resection alone in Ta and T1 bladder cancer. BJU Int 88:209–216. Shen Q, Brown PH. 2003. Novel agents for the prevention of breast cancer: Targeting transcription factors and signal transduction pathways. J Mammary Gland Biol Neoplasia 8:45–73. Sherman ME, Lorincz AT, Scott DR, et al. 2003. Baseline cytology, human papillomavirus testing, and risk for cervical neoplasia: A 10-year cohort analysis. J Natl Cancer Inst 95:46–52. Shureiqi I, Jiang W, Zuo X, et al. 2003. The 15-lipoxygenase-1 product 13-S-hydroxyoctadecadienoic acid down-regulates PPAR-delta to induce apoptosis in colorectal cancer cells. Proc Natl Acad Sci U S A 100: 9968–9973. Sigurdsson K. 1999. The Icelandic and Nordic cervical screening programs: trends in incidence and mortality rates through 1995. Acta Obstet Gynecol Scand 78:478–485. Slaughter D, Southwick H, Smejkal W. 1953. Field cancerization in oral stratified squamous epithelium: Clinical implications of multicentric origin. Cancer 6:963–968. Solomon SD, McMurray JJ, Pfeffer MA, et al. 2005. Adenoma Prevention with Celecoxib (APC) Study Investigators. Cardiovascular risk associated with celecoxib in a clinical trial for colorectal adenoma prevention. N Engl J Med 352:1071–1080. Stam-Posthuma JJ, Vink J, le Cessie S, Bruijn JA, Bergman W, Pavel S. 1998. Effect of topical tretinoin under occlusion on atypical naevi. Melanoma Res 8:539–548. Steele VE, Boone CW, Lubet RA, et al. 1998. Preclinical drug development paradigms for chemopreventives. Hematol Oncol Clin North Am 12:943–961, v–vi. Steinbach G, Lynch PM, Phillips RK, et al. 2000. The effect of celecoxib, a cyclooxygenase-2 inhibitor, in familial adenomatous polyposis. N Engl J Med 342:1946–1952. Sudbo J, Bryne M, Mao L, et al. 2003. Molecular based treatment of oral cancer. Oral Oncol 39:749–758. Sylvester RJ, van der MA, Lamm DL. 2002. Intravesical bacillus CalmetteGuerin reduces the risk of progression in patients with superficial bladder cancer: A meta-analysis of the published results of randomized clinical trials. J Urol 168:1964–1970. Tangrea JA, Edwards BK, Taylor PR, et al. 1992. Long-term therapy with lowdose isotretinoin for prevention of basal cell carcinoma: A multicenter clinical trial. Isotretinoin-Basal Cell Carcinoma Study Group. J Natl Cancer Inst 84:328–332. Temple R. 1999. Are surrogate markers adequate to assess cardiovascular disease drugs? JAMA 282:790–795. Thompson IM, Coltman CA, Crowley J. 1997. Chemoprevention of prostate cancer: The Prostate Cancer Prevention Trial. Prostate 33:217–221. Thompson IM, Goodman PJ, Tangen CM, et al. 2003. The influence of finasteride on the development of prostate cancer. N Engl J Med 349:215–224. Tockman MS. 2001. Lung cancer: Chemoprevention and intermediate effect markers. IARC Sci Publ 154:257–270. Torrance CJ, Jackson PE, Montgomery E, et al. 2000. Combinatorial chemoprevention of intestinal neoplasia. Nat Med 6:1024–1028.

1339

Trenkwalder P. 2000. Reducing cardiovascular morbidity and mortality in the elderly. Blood Press Suppl 1:40–43. Tutrone WD, Saini R, Caglar S, Weinberg JM, Crespo J. 2003. Topical therapy for actinic keratoses, I: 5-Fluorouracil and imiquimod. Cutis 71:365–370. Tutrone WD, Saini R, Caglar S, Weinberg JM, Crespo J. 2003. Topical therapy for actinic keratoses, II: Diclofenac, colchicine, and retinoids. Cutis 71:373–379. Uchida K, Seidenfeld J, Rademaker A, Oyasu R. 1989. Inhibitory action of alpha-difluoromethylornithine on N-butyl-N-(4-hydroxybutyl) nitrosamine-induced rat urinary bladder carcinogenesis. Cancer Res 49:5249–5253. Uemura N, Okamoto S, Yamamoto S, et al. 2001. Helicobacter pylori infection and the development of gastric cancer. N Engl J Med 345: 784–789. Umar A, Boyer JC, Thomas DC, et al. 1994. Defective mismatch repair in extracts of colorectal and endometrial cancer cell lines exhibiting microsatellite instability. J Biol Chem 269:14367–14370. Umar A, Kunkel TA. 1996. DNA-replication fidelity, mismatch repair and genome instability in cancer cells. Eur J Biochem 238:297–307. Umar A, Viner JL, Hawk ET. 2001. The future of colon cancer prevention. Ann N Y Acad Sci 952:88–108. Valla DC, Chevallier M, Marcellin P, et al. 1999. Treatment of hepatitis C virusrelated cirrhosis: A randomized, controlled trial of interferon alfa-2b versus no treatment. Hepatology 29:1870–1875. van Gorkom BA, van der Meer R, Karrenbeld A, et al. 2002. Calcium affects biomarkers of colon carcinogenesis after right hemicolectomy. Eur J Clin Invest 32:693–699. van Zandwijk N, Dalesio O, Pastorino U, de Vries N, van Tinteren H. 2000. EUROSCAN, a randomized trial of vitamin A and N-acetylcysteine in patients with head and neck cancer or lung cancer. For the European Organization for Research and Treatment of Cancer Head and Neck and Lung Cancer Cooperative Groups. J Natl Cancer Inst 92:977–986. Veronesi U, De Palo G, Costa A, Formelli F, Decensi A. 1996. Chemoprevention of breast cancer with fenretinide. IARC Sci Publ 136:87–94. Veronesi U, De Palo G, Marubini E, et al. 1999. Randomized trial of fenretinide to prevent second breast malignancy in women with early breast cancer. J Natl Cancer Inst 91:1847–1856. Veronesi U, Maisonneuve P, Costa A, et al. 1998. Prevention of breast cancer with tamoxifen: Preliminary findings from the Italian randomised trial among hysterectomised women. Italian Tamoxifen Prevention Study. Lancet 352:93–97. Veronesi U, Maisonneuve P, Rotmensz N, et al. 2003. Italian randomized trial among women with hysterectomy: Tamoxifen and hormone-dependent breast cancer in high-risk women. J Natl Cancer Inst 95:160–165. Viner JL, Richmond E, Lawrence JA, Hawk ET. 2002. Chemoprevention. The Cancer Handbook. M. Alison. London, Nature Publishing Group 2: 1417–1437. Virtamo J, Pietinen P, Huttunen JK, et al. 2003. Incidence of cancer and mortality following alpha-tocopherol and beta-carotene supplementation: A postintervention follow-up. JAMA 290:476–485. Vogelstein B, Fearon ER, Hamilton SR, et al. 1988. Genetic alterations during colorectal tumor development. N Engl J Med 319:525–532. Wactawski-Wende J, Kotchen JM, Anderson GL, et al. 2006. Calcium plus vitamin D supplementation and the risk of colorectal cancer. N Engl J Med 354:684–696. Wagenaar-Miller RA, Hanley G, Shattuck-Brandt R, et al. 2003. Cooperative effects of matrix metalloproteinase and cyclooxygenase-2 inhibition on intestinal adenoma reduction. Br J Cancer 88:1445–1452. Wallace K, Baron JA, Cole BF, Karagas MR, Beach ML. 2002. Calcium carbonate chemoprevention in the large bowel: Effects on hyperplastic polyps, tubular adenomas, and more advanced lesions. Proc AACR 43:163–164. Wang DL, Marko M, Dahl AR, et al. 2000. Topical delivery of 13-cis-retinoic acid by inhalation up-regulates expression of rodent lung but not liver retinoic acid receptors. Clin Cancer Res 6:3636–3645. Wattenberg LW, Wiedmann TS, Estensen RD, et al. 2000. Chemoprevention of pulmonary carcinogenesis by brief exposures to aerosolized budesonide or beclomethasone dipropionate and by the combination of aerosolized budesonide and dietary myo-inositol. Carcinogenesis 21:179–182. Wattenberg LW, Widemann TS, Estensen RD, Zimmerman CL, Steele VE, Kelloff GJ. 1997. Chemoprevention of pulmonary carcinogenesis by aerosolized budesonide in female A/J mice. Cancer Res 57:5489– 5492. Wei K, Kucherlapati R, Edelmann W. 2002. Mouse models for human DNA mismatch-repair gene defects. Trends Mol Med 8:346–353. Weir HK, Thun MJ, Hankey BF, et al. 2003. Annual report to the nation on the status of cancer, 1975–2000, featuring the uses of surveillance data for cancer prevention and control. J Natl Cancer Inst 95:1276–1299.

1340

PART V: CANCER PREVENTION AND CONTROL

Whittle H, Jaffar S, Wansbrough M, et al. 2002. Observational study of vaccine efficacy 14 years after trial of hepatitis B vaccination in Gambian children. BMJ 325:569. Wickerham DL. 2003. Tamoxifen’s impact as a preventive agent in clinical practice and an update on the STAR trial. Recent Results Cancer Res 163:87–95; discussion 264–266. Wild CP, Hardie LJ. 2003. Reflux, Barrett’s oesphagus and adenocarcinoma: Burning questions. Nat Rev Cancer 3:676–684. Williams C, Shattuck-Brandt RL, DuBois RN. 1999. The role of COX-2 in intestinal cancer. Ann N Y Acad Sci 889:72–83. Winawer SJ, Zauber AG, Ho MN, et al. 1993. Prevention of colorectal cancer by colonoscopic polypectomy. The National Polyp Study Workgroup. N Engl J Med 329:1977–1981. Winde G, Schmid KW, Brandt B, Muller O, Osswald H. 1997. Clinical and genomic influence of sulindac on rectal mucosa in familial adenomatous polyposis. Dis Colon Rectum 40:1156–1168; 1168–1169. Wolf JE Jr., Taylor JR, Tschen E, Kang S. 2001. Topical 3.0% diclofenac in 2.5% hyaluronan gel in the treatment of actinic keratoses. Int J Dermatol 40:709–713. Wolf JK, Ramirez PT. 2001. The molecular biology of cervical cancer. Cancer Invest 19:621–629. The Women’s Health Initiative Study Group. 1998. Design of the Women’s Health Initiative clinical trial and observational study. Control Clin Trials 19:61–109.

Workman P. 2002. The impact of genomic and proteomic technologies on the development of new cancer drugs. Ann Oncol 13 Suppl 4:115–124. Wulfkuhle JD, Paweletz CP, Steeg PS, Petricoin EF 3rd, Liotta L. 2003. Proteomic approaches to the diagnosis, treatment, and monitoring of cancer. Adv Exp Med Biol 532:59–68. Yamawaki M, Katiyar SK, Anderson CY, Tubesing KA, Mukhtar H, Elmets CA. 1997. Genetic variation in low-dose UV-induced suppression of contact hypersensitivity and in the skin photocarcinogenesis response. J Invest Dermatol 109:716–721. Yarosh D, Klein J, O’Connor A, Hawk J, Rafal E, Wolf P. 2001. Effect of topically applied T4 endonuclease V in liposomes on skin cancer in xeroderma pigmentosum: A randomised study. Xeroderma Pigmentosum Study Group. Lancet 357:926–929. Yeh JJ, Lunetta KL, van Orsouw NJ, et al. 2000. Somatic mitochondrial DNA (mtDNA) mutations in papillary thyroid carcinomas and differential mtDNA sequence variants in cases with thyroid tumours. Oncogene 19:2060–2066. Yuspa SH, Dlugosz AA, Denning MF, Glick AB. 1996. Multistage carcinogenesis in the skin. J Investig Dermatol Symp Proc 1:147–150. Yuspa SH, Poirier MC. 1988. Chemical carcinogenesis: From animal models to molecular models in one decade. Adv Cancer Res 50:25–70.

72

Regulating Carcinogens JONATHAN M. SAMET, THOMAS A. BURKE, AND LYNN GOLDMAN

C

ancer has long been viewed as a disease associated with civilization and one that could be prevented by finding the manmade factors causing the disease (Proctor, 1995). Much epidemiological research on cancer has been directed at finding these factors so that exposures to them could be reduced and the disease prevented; only more recently has the model of pathogenesis embraced the concept that most cancer risk reflects the combination of environmental exposures and a risk-enhancing genetic background. There were dramatic examples of environmental and occupational carcinogens that drove the earlier view: scrotal cancer in chimney sweeps, first described in 1775 by Percival Pott, and 20th century epidemics such as smoking and lung cancer; vinyl chloride and angiosarcoma of the liver, and asbestos and mesothelioma. In fact, during the 1970s, increasing cancer rates occurring in parallel with rising production of chemicals and pollution contaminating air, food, and water raised long-voiced concern as to the existence of a widespread epidemic of environmental cancer (Proctor, 1995). Discussion about the role of environmental factors became strident in the mid-1970s. A 1979 book, The Politics of Cancer written by Samuel S. Epstein (1979), attributed most human cancers to environmental factors, a position that was reinforced in a government document estimating a very high burden of cancer attributable to occupational exposures (Bridbord et al., 1978). Subsequently, various estimates were offered on the proportion of cancer attributable to environmental factors with the underlying assumption that the attributable and preventable burdens were equivalent. A thoughtful set of estimates was offered by the British epidemiologists, Richard Doll and Richard Peto (1981). Overall, they attributed over 80% of cancer to various environmental factors, construing environment broadly as all external factors determining risk. Some scientists put forth the proposition that relatively little cancer risk was related to manmade substances in the environment, and that natural toxins in food were much more carcinogenic than those of human design (Ames et al., 1987). Over the subsequent two decades, while the stridor of the debate among proponents and opponents of the proposition that much cancer is due to manmade contaminants has lessened, the premise that a substantial proportion of cancer comes from environmental factors and hence is avoidable remains widely accepted. Epidemiological evidence supports this view. A twin study in Scandinavia explored the role of heredity in cancer and found that about 28% of variation in cancer rates can be explained by hereditable genetic traits, leaving the remainder as reflecting broad environmental causes (Lichtenstein et al., 2000). However, the concept of “the environment” has been extended beyond manmade environmental contaminants to incorporate the personal environment and such lifestyle factors as diet, smoking, and alcohol use, all of which are likely to be responsible for a major proportion of the risk not attributable to inherited traits. Nonetheless, across the world, extensive sets of laws and regulations are in place to protect workers and the public more generally from environmental carcinogens. In the United States, environmental and occupational carcinogens are covered in numerous laws that are implemented by diverse agencies (Table 72–1), reflective of the multiple routes of exposure to carcinogens and the many situations in which exposures take place. Asbestos, for example, is a wellestablished occupational and environmental carcinogen that causes lung cancer, mesothelioma, and possibly other cancers, as well as nonmalignant respiratory diseases. There are multiple sources of exposure

to asbestos. For workers, exposures occur for the asbestos miners and millers, the persons making asbestos-containing products, the users of these products, and now, the workers who handle asbestos-containing materials during construction, building repair, and demolition. For the population, exposures may arise from fibers released from the many asbestos-containing products that have been placed in indoor environments and from fibers released into outdoor air from the brake pads in vehicles. Workplace exposures are covered by the Occupational Safety and Health Administration, which states that no employee should be exposed to an airborne concentration of asbestos in excess of 0.1 fiber per cubic centimeter of air as an eight-hour time-weighted average (TWA) (Occupational Safety and Health Administration (OSHA), 1989), while the US Environmental Protection Agency (EPA) covers exposures to the general population through restrictions on use and guidance on management of the asbestos in-place in schools and public and commercial buildings (US Environmental Protection Agency (EPA), 1988; US Environmental Protection Agency (EPA), 2003c). Radiation exposures are covered through an even more complex set of approaches, given the multiplicity of sources and the widespread use of radiation for commercial and medical purposes. Agencies with regulatory authority related to radiation include the EPA, the Nuclear Regulatory Commission, the Mine Safety and Health Administration, and the Food and Drug Administration. This chapter provides a broad overview of governmental regulatory approaches to controlling exposures to carcinogens. It covers general principles and approaches. We focus specifically on statutes related to the United States, while giving lesser coverage to approaches in Europe. The focus is at the federal level, even though states may also promulgate regulations. Most notably, the State of California enacted Proposition 65, the Safe Drinking Water and Toxic Enforcement Act of 1986, to protect and inform California citizens with regard to chemicals in water known to cause cancer, birth defects, or other reproductive harm. It requires the Governor to publish an annual list of chemicals known to the state to cause cancer or reproductive toxicity (Office of Environmental Health Hazard Assessment and State of California, 1986). The chapter does not cover initiatives, whether governmental or non-governmental, to influence individuals to reduce their exposures to carcinogens. The evolution of cancer control more generally, including regulatory and non-regulatory approaches, has been described by Proctor (1995) and Patterson (1989).

IDENTIFICATION OF CARCINOGENS The determination that an environmental agent causes cancer, or is likely to, is a needed basis for regulation, under most circumstances. The identification of an agent as causing cancer generally involves the review of all lines of relevant evidence and a weighing of that evidence against criteria for causal inference (See Chapter 1). For environmental and occupational carcinogens, the evidence comes from toxicologic knowledge of structure-activity relations and mechanisms of action, in vitro systems and animal bioassays, and from epidemiological studies. Each of these lines of evidence has strengths and weaknesses in guiding the causal determination. Compounds can be rapidly screened for potential carcinogenicity on the basis of their structure and on findings of short-term assays, such as the Ames test. On the

1341

1342

PART V: CANCER PREVENTION AND CONTROL

Table 72–1. Agencies and Regulations Governing Environmental and Occupational Carcinogens Statute and Year(s)

Regulatory Scope

Risk Mandate

Usual Acceptable Residual Risk

Toxic Substances Control Act (TSCA), 1976

Chemicals in commerce

Avoid and mitigate “unreasonable risk” via risk-benefit balancing

Unstated, but usually 10-5 to 10-6 for nonoccupational, 10-4 to 10-5 for occupational

Federal Food, Drug, and Cosmetic Act (FFDCA) 1906, 1938, 1962, 1977 Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA) 1947, 1972

Pesticides in food

“Reasonable certainty of no harm” for residues

10-6 for assumed max residues in average diet, 10-6 for non-dietary exposure

Pesticides in commerce

Balance risks, benefits, social and economic costs; efficacious yet w/o “unreasonable risk to man or environment”

Unstated, but usually 10-5 to 10-6 for nonoccupational, 10-4 to 10-5 for occupational

Safe Drinking Water Act (SDWA) 1974, 1986, 1996

Drinking water

10-4 to 10-6 is range considered to be adequate

Clean Water Act 1977, 1987

Water quality

Resource Conservation and Recovery Act (RCRA) 1976, 1984, 1986

Hazardous waste handling and disposal

Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA) (Superfund) 1980, 1986, 1990 Clean Air Act 1970, 1977, 1990

Superfund, cleanup of hazardous waste sites

For carcinogens, unenforceable maximum contaminant limits (MCL) of zero, but enforceable goals (MCLG) set by technology if within adequate margin of safety and consideration of feasibility; risk/risk tradeoffs for disinfection byproducts Protect public health and welfare with nonenforceable, health-based water quality criteria and enforceable “best” technologybased effluent standards Aim at “cradle-to-grave” stewardship; technology- and process-based, but also risk-triggered corrective action, to be protective of human health and the environment, excluding costs Applicable other laws plus cleanup to be protective of human health and environment; risk-based but consider feasibility

Criteria air pollutants

10-5 to 10-7

Listing: 10-5 Corrective action: 10-4 to 10-6 Incinerators: 10-5 10-4 to 10-6, depending partly on anticipated future use of site

Clean Air Act 1970, 1977, 1990

Hazardous air pollutants

Federal Food, Drug, and Cosmetic Act (FFDCA) 1906, 1938, 1962, 1977 Food and Drug Administration Modernization Act 1997 Federal Food, Drug, and Cosmetic Act (FFDCA) 1906, 1938, 1962, 1977 Food and Drug Administration Modernization Act 1997 Federal Food, Drug, and cosmetic Act (FFDCA) 1906, 1938, 1962, 1977 Food and Drug Administration Modernization Act 1997 Occupational Safety and Health Act (OSH Act) 1970

Food additives, colors and contaminants; cosmetics

Adequate margin of safety to protect public health Must apply Maximum Available Control Technology; if residual risk to MEI > 10-6, further regulate to provide adequate margin of safety to protect public health, considering costs “Delaney Clause,” no additives that are animal carcinogens.; “reasonable certainty of no harm” for other additives, no cost considerations

Medical devices

Safe and effective “under conditions of use”

Unclear

Pharmaceuticals

Safe and effective “under conditions of use”

Unclear

Occupational exposures

“No employee will suffer material impairment of health,” considering feasibility of standards

Feasible controls

Consumer Product Safety Act 1972, 1976, 1978, 1980, 1981, 1983, 1988, 1990

Consumer products

“To protect . . . against unreasonable risk of injury” with “reasonably necessary” standards, considering cost/benefit

Unclear

other hand, findings from experimental assays must be generalized to people. Epidemiological studies provide information directly on the risks to people and can provide insights into factors determining the risk of exposure. Epidemiological evidence can be limited by potential confounding or other sources of bias and small sample sizes often limit the precision of results. Recently, other types of evidence have also been incorporated into approaches for identification of carcinogens, such as mechanistic and pharmacokinetic data that can be used to support (or refute) the assumption that a chemical that causes cancer in animals may also cause cancer in humans. Various groups or agencies have developed criteria for evaluating evidence related to the carcinogenicity of environmental agents and

Unstated; regulated on the basis of non cancer risks <10-6

Zero for additives; 10-6 for assumed max residues in “high-use” diet

for classifying environmental agents as to their carcinogenicity (Table 72–2). These groups address the carcinogenicity of environmental agents using similar approaches: a compilation of all relevant lines of evidence, including epidemiological, clinical, and toxicological studies; review of this evidence by a multidisciplinary panel that is considered balanced in its perspective; and application of a set of criteria for judging the evidence and reaching a conclusion as to whether the agent under consideration can be classified as causing cancer. Standardized descriptors of the level of certainty as to carcinogenicity are incorporated into these approaches. The International Agency for Research on Cancer (IARC) of the World Health Organization has an ongoing series of monographs that

Table 72–2. Organizations with Criteria for Evaluating Carcinogenicity Organization

Document

Environmental Protection Agency (EPA)

Guidelines for Carcinogen Risk Assessment, 2003

National Toxicology Program

Report on Carcinogens, 10th Edition, 2004

International Agency for Research on Cancer

IARC Monographs on the Evaluation of Carcinogenic Risks to Humans (Monographs 1 through 8.5 plus supplements, have been published to date)

European Commission

Guidelines for Setting Specific Concentration Limits for Carcinogens in Annex 1 of Directive 67/548/EEC. Inclusion of Potency Considerations

Criteria for Evaluating Evidence Carcinogenic to Humans: This descriptor is appropriate when there is convincing epidemiologic evidence demonstrating causality between human exposure and cancer, or exceptionally when there is strong epidemiological evidence, extensive animal evidence, knowledge of the mode of action, and information that the mode of action is anticipated to occur in humans and progress to tumors. Likely to be Carcinogenic to Humans: This descriptor is appropriate when the available tumor effects and other key data are adequate to demonstrate carcinogenic potential to humans, but does not reach the weight-ofevidence for the descriptor “carcinogenic to humans.” Suggestive Evidence of Carcinogenic Potentials: This descriptor is appropriate when the evidence from human or animal data is suggestive of carcinogenicity, which raises a concern for carcinogenic effects but is judged not sufficient for a stronger conclusion. Inadequate Information to Assess Carcinogenic Potential: This descriptor is used when available data are judged inadequate to perform an assessment. Known to Be a Human Carcinogen: There is sufficient evidence of carcinogenicity from studies in humans, which indicates a causal relationship between exposure to the agent, substance, or mixture and human cancer. Reasonably Anticipated to Be a Human Carcinogen: There is limited evidence of carcinogenicity from studies in humans, which indicates that causal interpretation is credible, but that alternative explanations, such as chance, bias, or confounding factors, could not adequately be excluded OR There is sufficient evidence of carcinogenicity from studies in experimental animals, which indicates there is an increased incidence of malignant and/or a combination of malignant and benign tumors: (1) in multiple species or at multiple tissue sites, or (2) by multiple routes of exposure, or (3) to an unusual degree with regard to incidence, site or type of tumor, or age at onset; OR There is less than sufficient evidence of carcinogenicity in humans or laboratory animals, however; the agent, substance, or mixture belongs to a well-defined, structurally related class of substances whose members are listed in a previous Report on Carcinogens as either a known to be human carcinogen or reasonably anticipated to be human carcinogen, or there is convincing relevant information that the agent acts through mechanisms indicating it would likely cause cancer in humans. Group 1: The agent (mixture) is carcinogenic to humans. The exposure circumstance entails exposures that are carcinogenic to humans. This category is used when there is sufficient evidence of carcinogenicity in humans. Exceptionally, an agent (mixture) may be placed in this category when evidence of carcinogenicity in humans is less than sufficient but there is sufficient evidence of carcinogenicity in experimental animals and strong evidence in exposed humans that the agent (mixture) acts through a relevant mechanism of carcinogenicity. Group 2: This category includes agents, mixtures, and exposure circumstances for which, at one extreme, the degree of evidence of carcinogenicity in humans is almost sufficient, as well as those for which, at the other extreme, there are no human data but for which there is evidence of carcinogenicity in experimental animals. Agents, mixtures, and exposure circumstances are assigned to either group 2A (probably carcinogenic to humans) or group 2B (possibly carcinogenic to humans) on the basis of epidemiological and experimental evidence of carcinogenicity and other relevant data. Group 2A: The agent (mixture) is probably carcinogenic to humans. The exposure circumstance entails exposures that are probably carcinogenic to humans. This category is used when there is limited evidence of carcinogenicity in humans and sufficient evidence of carcinogenicity in experimental animals. In some cases, an agent (mixture) may be classified in this category when there is inadequate evidence of carcinogenicity in humans and sufficient evidence of carcinogenicity in experimental animals and strong evidence that the carcinogenesis is mediated by a mechanism that also operates in humans. Exceptionally, an agent, mixture, or exposure circumstance may be classified in this category solely on the basis of limited evidence of carcinogenicity in humans. Group 2B: The agent (mixture) is possibly carcinogenic to humans. The exposure circumstance entails exposures that are possibly carcinogenic to humans, This category is used for agents, mixtures, and exposure circumstances for which there is limited evidence of carcinogenicity in humans and less than sufficient evidence of carcinogenicity in experimental animals. It may also be used when there is inadequate evidence of carcinogenicity in humans but there is sufficient evidence of carcinogenicity in experimental animals. In some instances, an agent, mixture, or exposure circumstance for which there is inadequate evidence of carcinogenicity in humans but limited evidence of carcinogenicity in experimental animals together with supporting evidence from other relevant data may be placed in this group. Group 3: The agent (mixture or exposure circumstance) is not classifiable as to its carcinogenicity to humans. This category is used most commonly for agents, mixtures, and exposure circumstances for which the evidence of carcinogenicity is inadequate in humans and inadequate or limited in experimental animals. Exceptionally, agents (mixtures) for which the evidence of carcinogenicity is inadequate in humans but sufficient in experimental animals may be placed in this category when there is strong evidence that the mechanism of carcinogenicity in experimental animals does not operate in humans. Agents, mixtures, and exposure circumstances that do not fall into any other group are also placed in this category. Group 4: The agent (mixture) is probably not carcinogenic to humans. This category is used for agents or mixtures for which there is evidence suggesting lack of carcinogenicity in humans and in experimental animals. In some instances, agents or mixtures for which there is inadequate evidence of carcinogenicity in humans but evidence suggesting lack of carcinogenicity in experimental animals, consistently and strongly supported by a broad range of other relevant data, may be classified in this group. Carcinogenic to Humans: This descriptor is appropriate when there is convincing epidemiologic evidence demonstrating causality between human exposure and cancer, or exceptionally when there is strong epidemiological evidence, extensive animal evidence, knowledge of the mode of action, and information that the mode of action is anticipated to occur in humans and progress to tumors. Likely to Be Carcinogenic to Humans: This descriptor is appropriate when the available tumor effects and other key data are adequate to demonstrate carcinogenic potential to humans, but does not reach the weight-ofevidence for the descriptor “carcinogenic to humans.” Suggestive Evidence of Carcinogenic Potential: This descriptor is appropriate when the evidence from human or animal data is suggestive of carcinogenicity, which raises a concern for carcinogenic effects but is judged not sufficient for a stronger conclusion. Inadequate Information to Assess Carcinogenic Potential: This descriptor is used when available data are judged inadequate to perform an assessment.

1344

PART V: CANCER PREVENTION AND CONTROL

address the carcinogenicity of a wide range of agents. The first was published in 1972 and a total of 88 have been completed, with 85 published to date. These monographs are used throughout the world and have particular impact in countries lacking the capability for evaluating possible carcinogens. The process used by IARC involves the selection of a Working Group, considered to have a balanced range of view, which reviews all relevant lines of evidence and applies an established evaluation strategy. In assessing carcinogenicity, the Working Group considers not only animal and human evidence but mechanism(s) of action. The evaluation classifies the animal and human evidence separately in the following categories: sufficient evidence of carcinogenicity, limited evidence of carcinogenicity, inadequate evidence of carcinogenicity, and evidence suggesting lack of carcinogenicity. The overall evaluation classifies the agent into one of the Groups, I–IV (Table 72–2). The National Toxicology Program (NTP) of the National Institute of Environmental Health Sciences in the United States, is required under Section 301 (b) (4) of the Public Health Service Act to publish a biennial report listing substances either known to be human carcinogens or reasonably anticipated to be human carcinogens and to which a significant number of persons residing in the United States are exposed. The report also gives information concerning the nature of such exposure and the estimated number of persons exposed to such substances. The NTP report is an assessment of hazard and does not include a quantitative assessment of risk. The NTP uses a weight-of-evidence approach that is similar to that used by the IARC and, since 1996, explicitly “includes, but is not limited to, dose response, route of exposure, chemical structure, metabolism, pharmacokinetics, sensitive sub-populations, genetic effects, or other data relating to mechanism of action or factors that may be unique to a given substance” (National Toxicology Program (NTP), 2004) (Table 72–2). The EPA published its own set of risk assessment guidelines in 1986 (US Environmental Protection Agency (EPA), 1986). It began a process to revise these guidelines in 1990, and has published draft revisions in 1996 (US Environmental Protection Agency (EPA), 1996), and most recently a “final draft” guideline in 2003 (US Environmental Protection Agency (EPA), 1996; US Environmental Protection Agency (EPA) and Risk Assessment Forum, 2003). Also in 2003, the EPA published a new draft guideline for addressing exposure of children to carcinogens (US Environmental Protection Agency (EPA) and Risk Assessment Forum Technical Panel, 2003). In the agency’s risk assessment process, the hazard assessment step addresses whether the agent can be considered as a cause of cancer. Epidemiological evidence is to be considered with criteria based upon those proposed by Hill. The overall evaluation, however, gives broad consideration to all lines of evidence using a “weight-ofevidence” approach. A narrative approach is to be used that considers mechanistic understanding as well as animal and human data. The final classification scheme has categories similar to those used by other groups, but, in all draft guidelines published since 1996, an agent can be classified as carcinogenic to humans based on “strong” epidemiological evidence, if supported by other types of evidence, including animal studies and information on anticipated mode of action in humans.

PRINCIPLES UNDERLYING REGULATIONS FOR CARCINOGENS Overview The regulation of exposures to carcinogens involves a patchwork of regulatory and non-regulatory approaches that have varied over time and differ to an extent across nations. People’s exposures come from sources that may be under their own control (e.g., tobacco use and ingestion of alcohol) or that are not subject to their control (e.g., occupational carcinogens and asbestos in buildings). The latter are the focus of regulation whereas the former are addressed by a blend of initiatives that range from education concerning risks to broader programs of control, as in the example of tobacco. Societal activities are inevitably associated with exposure to carcinogens so that exposures

to environmental carcinogens cannot be eliminated all together. Consequently, principles are in use to guide the development of policies to control exposures to an acceptable level. These principles, considered below, have the general objective of achieving an acceptable level of risk and assuring safety. Lowrance (1976) defined safety as follows: “A thing is safe if its attendant risks are judged to be acceptable.” For carcinogen regulation, there are dual implications of this definition: information is needed on the risks associated with exposures, and principles are needed to judge the acceptability of risk. Any principles used must reflect societal concerns and tolerance of risk, the necessity of using the materials or processes that lead to exposures, the feasibility of controls, and the potential costs of control measures as well as those of cancers or other diseases arising from the exposures. Regulations directed at chemical and radiation exposures have also been addressed using distinct approaches that have not been fully harmonized (Tran et al., 2000). Concerns for cancer risk dominate both chemical and radiation risk exposures, but lines of evidence differ, with epidemiological data generally figuring more prominently for radiation. Quantitative criteria for acceptable risk also differ, as do approaches for risk estimation. A need for harmonization of approaches has been acknowledged (Tran et al., 2000).

Defining Acceptable Risks for Exposures to Carcinogens As Low As Reasonably Achievable The “As low as reasonably achievable” (ALARA) principle originated with increasing understanding of the nature of radiation carcinogenesis as the results of the Atomic Bomb Survivors study and other epidemiological studies that were first reported in the 1950s. To that time, tolerance limits for radiation had been largely based on radiation doses causing skin erythema (Hendee and Edwards, 1986; Caufield, 1989). The emerging epidemiological data, along with advancing knowledge of radiobiology, provided motivation for broadening the scope of radiation protection standards to consider not only individuals but populations. This principle was reflected in the standards proposed by the National Council on Radiation Protection and Measurements (NCRP) in the United States and the International Commission for Radiological Protection (ICRP), which makes recommendations that have global impact. In its Publication 26 in 1977, the ICRP (International Commission on Radiological Protection (ICRP), 1977) recommended that no radiation practice involving exposure should be used unless it provides benefit and that all exposures should be in accordance with the ALARA principle, taking costs and societal factors into consideration. Additionally, doses to individuals should not exceed the limits of the ICRP. Under ALARA, there is a two-pronged strategy for radiation protection, assuring that doses to individuals meet guidelines or regulations and that the population’s dose is as low as achievable, given technological constraints, costs, and the need for the activities involving exposure. Inherently, ALARA aims to achieve the lowest possible dose to the population; under the assumption that radiation causes cancer at any dose, the risk to the population reflects the collective dose and guidelines or regulations based upon the ALARA principle of minimizing this dose.

Precautionary Principle The “precautionary principle” originated in Europe and has been incorporated into international agreements there since the 1980s. The precautionary principle calls for action when there is an indication of a potential threat of irreversible harm, even in the face of uncertainty as to the risks. It was among the principles agreed to in the Rio Declaration, and signed by countries that participated in the 1992 United Nations Conference on Environment and Development (United Nations Conference on Environment and Development, 1992). Along with other principles, it has been incorporated into international efforts such as the Biosafety Protocol of the Biodiversity Convention. More recently, broader application of the precautionary principle has been proposed (Kriebel et al., 2001; Kriebel and Tickner, 2001; Tickner,

Regulating Carcinogens 2002). The 1998 Wingspread Statement, the result of a meeting on the topic, offered four components for extending the principle: 1. Taking action to reduce risk, even with uncertain evidence 2. Giving the burden of proof of safety to those who are proposing something associated with risk 3. Considering alternatives to the action that may have adverse consequences 4. Increasing the participation of the public in the decision-making process (Proceedings of the Wingspread Conference on Strategies for Implementing the Precautionary Principle, 1998). To date, the precautionary principle has received more widespread discussion and consideration in Europe, where it is the basis for a proposed policy to address chemicals (Rogers, 2003a; 2003b). The precautionary principle per se has never been referenced in US legislation. Recently, the US government has begun to define a “precautionary approach” that is used in its regulatory decisions (US Department of Agriculture, 2000; Graham, 2002). Applied to carcinogens, the precautionary principle might lead to consideration of preventive initiatives before the level of evidence was found sufficient to infer a causal relationship between a carcinogen and risk for cancer. Goldstein (2001) notes that the regulation of over 180 hazardous air pollutants under the 1990 Clean Air Act embodies elements of the precautionary principle because the EPA does not have the burden of proof to demonstrate that environmental levels of the air pollutant were likely to produce adverse effects before imposing maximal available control technology standards on industry. Jamieson and Wartenberg (2001) consider the example of childhood leukemia and exposure to electromagnetic radiation and propose that action to reduce exposures would be warranted with application of the precautionary principle. At this time, the precautionary principle is receiving its first uses with regard to carcinogenic agents, and more widespread application is likely.

Risk Assessment and Cost-Benefit Analysis Risk assessment and cost-benefit analysis bring a quantitative foundation to the control of carcinogens. A risk assessment can be qualitative, determining if a hazard exists, or quantitative, estimating the extent of the burden of disease caused by a particular agent. Costbenefit analysis adds comparison of two sets of costs: those estimated to arise from the disease caused by the agent of concern and those associated with controlling exposures to the agent. Estimation of “benefits” depends on being able to quantify disease rates and other harm addressed by the regulation and the ability to assign costs to those conditions; consequently, risk assessment and cost-benefit analysis are very closely interrelated. In the United States, statutes have contained various standards, some of which have implicitly or explicitly required that agencies perform quantitative risk assessments as part of the process of controlling occupational and environmental carcinogens. Regulatory impact analysis, including analysis of costs and benefits, has been required for evaluation of major rules (i.e., having >$100 million impact) by Executive Order since the time of President Carter. The White House Office of Management and Budget (OMB) has played an increasingly powerful role in the process of setting guidelines for such analyses and in review of proposed major rules. The current general formalism of risk assessment was defined in a landmark 1983 report of the National Research Council, widely known as “the Red Book” (National Research Council (NRC) and Committee on the Institutional Means for Assessment of Risks to Public Health, 1983). The Red Book described four elements to a risk assessment: 1. Hazard identification: Is a hazard posed by the exposure? 2. Exposure assessment: What is the distribution of the exposure? 3. Dose-response assessment: How does risk vary with exposure (or dose)? 4. Risk characterization: What is the extent of the risk to the population and what are the uncertainties in the characterization of risk? For a carcinogen, the hazard indentification component is equivalent to classifying the agent as a carcinogen. For carcinogens, infor-

1345

mation on profiles of exposure comes from diverse sources, including specific exposure assessment studies, surveys, and models. For some agents, exposures may occur through several different pathways and media. For some polycyclic aromatic hydrocarbons, for example, exposures may come from ingestion of cooked foods, inhalation of polluted air, and ingestion of contaminated water. For many carcinogens, dose-response assessment is based on animal bioassays, as epidemiological studies have provided information on dose-response for a relatively small number of carcinogens. Initially the US EPA and other US federal regulatory agencies had no uniform guidelines for cancer risk assessment, but, in 1979, several regulatory agencies, including the EPA, published a set of guidelines for cancer risk assessment (Interagency Regulatory Liaison Group (IRLG) and Working Group on Risk Assessment, 1979). Subsequent guidelines have closely adhered to the Red Book. Indoor radon offers an example of the application of quantitative risk assessment to an environmental carcinogen (National Research Council (NRC) and Committee on the Biological Effects of Ionizing Radiation, 1988). Decades ago, studies of underground miners provided clear evidence that radon (the actual carcinogen being the progeny of radon) causes lung cancer. The epidemiological data have been supported by findings of animal bioassays as well as laboratory findings on the mechanisms by which alpha particles released from radon progeny damage cells and their genetic material. Population exposures to radon have been characterized in multiple surveys, including national surveys in many countries. The epidemiological studies of underground miners and case-control studies of lung cancer in the general population provide characterizations of the relationship between radon concentration and risk for lung cancer; the various studies have been pooled to provide the most precise characterization possible of the dose-response relationship. The Biological Effects of Ionizing Radiation (BEIR) VI Committee of the National Research Council (National Research Council (NRC) and Committee on the Biological Effects of Ionizing Radiation, 1988) developed a risk model based on 11 cohort studies of underground miners that was used to estimate the numbers of lung cancer deaths caused by indoor radon in the United States. The estimates ranged from 15,000–22,000, depending on the particular risk model used. The report gives a lengthy listing of uncertainties affecting these estimates and attempted to quantify the impact of key uncertainties. Cost-benefit analysis builds from the quantification of disease burden produced by a quantitative risk assessment. The costs of the disease, including those of death or loss of years of life, are estimated, as are the costs of control measures. This approach assigns a cost to benefits from control measures, providing, for example, the cost per premature death prevented or per year of life gained. It is a decisionmaking tool that adds an economic perspective to carcinogen control. The application of cost-benefit analysis is controversial and it has been cast as a tool that can be used to avoid regulation. Additionally, some aspects of the methodology of cost-benefit analysis are controversial and awaiting resolution; one of the most complex issues is the cost of a human life and whether that cost depends on the ages of the affected individuals. Another complexity is handling benefits and costs that can be identified, but not readily quantified. These may include noncancer health risks, ecological risks, and positive and negative business impacts from new rules. The EPA has used cost-benefit analysis to demonstrate the effectiveness of its radon control strategy, which calls for the measurement of indoor radon in most homes and the mitigation of those exceeding a guideline value. In its 1992 publication, Technical Support Document for the 1992 Citizen’s Guide to Radon (US Environmental Protection Agency (EPA), 1992b), the EPA included a cost-effectiveness analysis of three options, the one presented in the revised Citizen’s Guide to Radon (US Environmental Protection Agency (EPA), 1992a) using a level of 4 pCi/L (picuries per liter—a measure of the concentration of radon, (the action level recommended in the Guide), and two others with levels of 2 pCi/L and 3 pCi/L. Figures used represented “the value the public places on reducing risks of death from all types of causes, given empirical evidence of the public’s willingness to either pay to reduce small risks or receive payments for accepting

1346

PART V: CANCER PREVENTION AND CONTROL

those risks” (US Environmental Protection Agency (EPA), 1992b) (p. 5-2). The result of the analysis provides a general sense of how cost effective the radon program is: the average cost per life saved for an action level of 4 pCi/L was estimated at $700,000; this figure compared “favorably to the expenditures that EPA has been wiling to see incurred for risk reductions in the past” (US Environmental Protection Agency (EPA), 1992b) (p. 5-16).

Mechanisms of Action/Dose Response Modeling As understanding of carcinogenesis has deepened, mechanism of action has been given greater weight in determining that an agent is carcinogenic; the most recent EPA guidelines give weight to mode of action, for example (US Environmental Protection Agency (EPA), 2003a). Mode of action might influence the assessment of carcinogenicity and also the selection of the most probable form(s) of the exposure-response or dose-response relationship. The form of the dose-response relationship is often critical for riskbased regulation. The presence or absence of a threshold and departures from linearity, whether sublinear or supralinear (Fig. 72–1), have substantial implications for estimating risk, both at existing levels or exposure and at levels that could be achieved by regulation. Relationships with a threshold indicate the possibility of setting a standard that results in no residual risk with a margin of safety. Those without a threshold imply that any exposure conveys some residual risk, even if very small, and that regulation will need to lower risk to the extent possible and to an acceptable level of risk. Epidemiological data can be informative as to the underlying dose-response relationship, but rarely serve as the basis for a precise estimation (Samet et al., 1998). Epidemiological data, when available for a particular carcinogen, are often analyzed to determine the form of the dose-response relationship. Regression methods are used for this purpose; epidemiological studies infrequently yield sufficient data to provide powerful assessments of such key issues as the presence or absence of a threshold and departure from a linear dose-response relationship. For this reason, several recent risk assessments have used pooled data sets from multiple studies to derive more precisely specified risk models, as in the example of the pooling of 11 cohort studies carried out by the BEIR VI Committee.

Children and Susceptible Populations Regulators have long appreciated that there are likely to be person-toperson differences in behavior, genetics, immunocompetence, and other factors, such as nutrition, which confer differences in susceptibly to cancer. Specific heritable differences in risk have been identi-

4 1 3

2

Figure 72–1. Examples of dose-response models used for carcinogens: 1—linear non-threshold mode; 2—linear threshold model; 3—sub-linear non-threshold model; and, 4—supra-linear non-threshold model.

fied; for example, increased skin cancer among persons with xeroderma pigmentosum with exposure to UV light; increased bladder cancer among persons exposed to occupational carcinogens who are “poor acetylators”; and even genes for increased susceptibility to tobacco-caused cancers. In 1994, the US National Research Council (NRC) noted that the EPA and other regulatory agencies assume, by default, that the population has the median susceptibility level; this assumption is likely to lead to an underestimation of population risk. In its 1994 report, Science and Judgment in Risk Assessment (National Research Council (NRC) and Committee on Risk Assessment of Hazardous Air Pollutants, 1994), it recommended the development of a distribution of population susceptibilities to develop a default greater than one. The hormone diethylstilbestrol (DES) caused a high rate of vaginal cancers among women whose mothers received DES during the first trimester of pregnancy (Herbst, 1987; Giusti et al., 1995; Palanza et al., 2001). Since this tragic episode, there has been an increasing interest in transplacental and early life exposure to carcinogens and the potential role of such exposures in causing childhood cancer, as well as affecting cancer risk across the life course. There is a limited database of toxicology studies involving transplacental and early-life dosing with carcinogens; generally, agents that are carcinogenic with exposure to adults also cause excess cancers over a lifetime when dosed in utero (Olshan et al., 2000). The cancers are generally the same types observed with adult dosing; frequently, the rates of occurrence are much higher. It has been hypothesized that these higher rates may due to alterations in susceptibility, higher effective doses to target organs, and/or reduction of time to tumor because of rapid growth early in life (Anderson et al., 1985; Tomatis, 1989). The possibility that carcinogens and mutagens cause germ-cell alterations that enhance susceptibility to cancer in offspring is currently of great interest. Animal models for transgenerational carcinogenesis have been developed (Tomatis, 1989; Tomatis et al., 1992; Daher et al., 1998), but maternal germ cell exposures have not been studied in human populations. Studies of such exposures would be quite difficult since the ova are developing in utero. There are numerous studies that have examined preconception exposures to men fathering children and subsequent risk of childhood cancer in their offspring (Savitz and Chen, 1990). Methodological challenges in this area of research include the difficulty of separating environmental exposures to the father before conception from maternal exposures during gestation. Gestation and the first year of life may be the most vulnerable time for exposure to carcinogens. Factors that may be involved have been identified including: a) numbers of target cells at risk, b) sensitivity to cell killing, c) effects of rate of cell division on fixation of mutation before repair can occur, d) ability to repair DNA damage, e) expansion of clones of mutated cells as part of normal ontogeny, f ) presence of undifferentiated stem cells, g) development of differentiated characteristics, including the ability to carry out metabolic activation of chemicals, h) metabolic detoxification by placenta and/or maternal tissues, i) metabolic detoxification by the perinate itself, and j) immaturity of the endocrine and immunological systems (Anderson et al., 2000) Despite this basis for concern about early life exposure to carcinogens, regulatory cancer testing guidelines for EPA and the Food and Drug Administration instruct industry to test mature animals exclusively. The high vulnerability of the fetus and the mother to toxic substances during pregnancy makes it challenging to design appropriate animal studies for early life exposure and cancer effects. Also, monitoring exposure to the fetus when dosing is via the mother, but models involving injection of carcinogens into the fetus would not reflect the real-life situation with respect to hepatic and placental metabolism and transfer of chemicals (and metabolites) from mother to fetus. Emerging techniques to assess genomic and proteomic changes may, in the future, be able to identify significant biological responses and give greater understanding of risks of early exposure to carcinogens (McHale et al., 2003). Appropriately, current draft EPA cancer guidelines apply an uncertainty factor when extrapolating cancer risk from

1347

Regulating Carcinogens Terrestrial 8%

Internal 11% Medical X-rays 11%

Cosmic 8%

Nuclear medicine 4% Consumer products 3% Other <1% Occupational Fallout Nuclear Fuel Cycle Miscellaneous

Radon 55%

Figure 72–2. The percentage contribution of various radiation sources to the total average effective dose equivalent in the US population (National Council on Radiation Protection and Measurements (NCRP), 1987).

adults to the fetus and child (US Environmental Protection Agency (EPA) and Risk Assessment Forum Technical Panel, 2003).

CARCINOGEN REGULATION IN THE UNITED STATES Radiation Radiation exposures are ubiquitous in the environment, coming from both natural and manmade sources. Radiation has many uses with

evident benefits—its application in diagnostic and therapeutic radiation treatments and in technical applications, for example. Exposure to the population or to workers may be an unintended consequence of some processes, including nuclear power generation and waste handling. Radon, a naturally occurring source of radiation exposure, is a ubiquitous contaminant of indoor and outdoor air. Radiation can be broadly classified as ionizing (e.g., X-rays) or non-ionizing (e.g., microwaves). While many types of both ionizing and non-ionizing radiation are covered by regulations, ionizing radiation has been shown to cause cancer in various tissues (see chapter 15). Ionizing radiation includes X-rays and gamma-rays, as well as various types of particles—alpha particles, beta particles, neutrons, and others. Ionizing radiation can be further classified by the intensity of energy delivery to tissues: low linear-energy-transfer (LET) or highLET. Perhaps surprisingly, the majority of the population’s radiation exposure comes from indoor exposure to radon (Fig. 72–2) (National Council on Radiation Protection and Measurements (NCRP), 1987), the naturally occurring carcinogenic gas having decay products that release high-LET alpha particles (National Research Council (NRC) et al., 1998). Many of the other contributing sources are subject to extensive regulation, whereas radon is covered by the EPA, but without direct regulatory authority (Table 72–3). The control of exposures to radiation has long been driven by recognition that radiation exposure has adverse health effects, both cancerous and non-cancerous. Epidemiological evidence has figured centrally in the current approach. Within about a year of Roentgen’s 1895 discovery of the X-ray, cases of acute injury from radiation exposure were reported and by 1904, Edison’s assistant, Clarence Dally, had died of a cancer that originated at the site of an X-ray burn (Caufield, 1989). The first standards for exposure were implemented in 1934 by the US Advisory Committee on X-Ray and Radium Protection; these standards were based around the dose of radiation causing erythema

Table 72–3. Agencies and Regulations governing Radiation Organization Department of Energy (DOE) Department of Transportation (DOT) Environmental Protection Agency (EPA)

Area of Regulation Ionizing radiation exposure resulting from the conduct of DOE activities Transportation of radioactive materials Naturally occurring and accelerator-produced radioisotopes in drinking water (not source, special, or byproduct materials) Radioactive substance or matter that is emitted into or otherwise enters the ambient air (including source, special, or byproduct materials) Release or threat of release of radionuclides into the environment Byproduct materials at licensed commercial uranium and thorium processing sites Radioactive materials Radioactive waste and spent nuclear fuel

Indoor radon Food and Drug Administration (FDA)

Ionizing radiation-emitting devices, ultrasonic radiationemitting devices

Mine Safety and Health Administration (MSHA) Nuclear Regulatory Commission (NRC)

Radiation exposure in underground mines

Occupational Health and Safety Administration (OSHA)

Source material (uranium and thorium); special nuclear material (enriched uranium and plutonium); byproduct material (material made radioactive in a reactor and residues from the milling of uranium and thorium) Radiation exposure in the workplace

Statute(s) 10 CFR Parts 820–840 Hazardous Materials Transportation Act (HMTA) 1975 (49 CFR Parts 100–185) Clean Water Act (CWA) 1977, 1987 (33 USC § 502) Clean Air Act (CAA) 1970, 1977, 1990 (42 USC § 7602) Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA) (Superfund) 1980, 1986, 1990 (42 USC § § 9601–9657) Uranium Mill Trailing Radiation Control Act (UMTRCA) 1978 (40 USC § 2022; 40 CFR Part 192) Atomic Energy Act (AEA) 1946, 1954 (42 USC § § 2011–2296) Nuclear Waste Policy Act (NWPA) 1982 (42 USC § 10101) Low Level Radioactive Waste Policy Act (LLRWPA) 1980, 1985 (42 USC § 2121b) Indoor Radon Abatement Act (IRAA) 1988 (15 USC § 2601) Federal Food, Drug, and Cosmetic Act (FFDCA) 1906, 1938, 1962, 1977, 1997 (21 USC § § 360hh–360ss; 21 CFR 1020 et seq) Federal Mine Safety and Health Act (FMSHA) 1977 (30 USC § § 801 et seq; 30 CFR Parts 1–199) Standards for Protection against Radiation (10 CFR Part 20) Occupational Safety and Health Act (OSH Act) 1970 (29 CFR Parts 1910–1926)

1348

PART V: CANCER PREVENTION AND CONTROL

(redness) of the skin (Kocher, 1991). The basis for radiation protection standards began to shift as radiobiological studies demonstrated that radiation exposure had genetic consequences and that cancer risk was increased by radiation. Additionally, the nuclear weapons and fuel industries produced new sources of external radiation as well as created opportunities for workers to be exposed to internal emitters. The 1949 recommendations for maximum permissible dose from the National Council for Radiation Protection and Measurements (1954) reflected the shift from prevention of erythema, a nonstochastic effect, to prevention of leukemia and genetic damage, both stochastic effects. Our current system of standards is largely based upon the quantitative risks of cancer, as assessed in epidemiological studies. The findings of the atomic bomb survivors study have long been at the center of these regulations, although increasingly data from other studies are being considered as well (Caufield, 1989). Current regulations are largely driven by risk models based in these data. The models are developed by various expert committees, including committees of the NCRP and the ICRP, and the Biological Effects of Ionizing Radiation (BEIR) Committees of the US National Research Council. The United Nations also maintains an ongoing committee process in its United Nations Scientific Committee on Atomic Radiation (UNSCEAR). These organizations have appointed such committees for decades; they compile the scientific evidence, reach conclusions as to cancers caused by radiation and factors determining susceptibility, and derive quantitative risk models. The more recent BEIR Committees have been carrying out extensive analyses to develop risk models of studies of underground miners for radon risk estimation (National Research Council (NRC) et al., 1998) and of the atomic bomb survivors data for low-LET radiation (National Research Council (NRC) and Committee on the Biological Effects of Ionizing Radiation, 1990). The BEIR VII report released in 2005 addresses the controversial topic of risks of cancer at low doses of ionizing radiation (Brenner et al., 2003; National Research Council (NRC) et al., 2005). Reflective of the many sources of radiation exposure in a developed country, a variety of agencies promulgate and enforce regulations for control of radiation exposure in the United States (Table 72–3). For example, the Food and Drug Administration handles therapeutic radiation, whereas the EPA covers a number of broad environmental sources. The Nuclear Regulatory Commission addresses nuclear power generation as well as the manufacture of fuels. Occupational radiation exposures are covered by both the Occupational Health and Safety Administration and the Mine Safety and Health Administration. Radon, as a naturally occurring source of radioactivity, is not covered by regulation although it is the greatest contributor to population exposure to radiation (Fig. 72–2). The EPA, through its Indoor Radon Abatement Act, has broad authority to provide guidance on acceptable levels of radon in homes and other indoor environments, and to provide education. Its Radon Program has developed certification programs for measurement and mitigation, and the EPA communicates with the public concerning radon through its Citizen’s Guide (US Environmental Protection Agency (EPA), 1992a).

Outdoor Air The Federal Clean Air Act (CAA) is the foundation for all air pollution legislation in the United States. This act was originally passed in 1963 to provide grants for state and local air pollution control districts. Since then, the CAA has been modified through five cycles of amendments (in 1965, 1967, 1970, 1977, and 1990) and has developed into the framework used for federal and state regulations of air pollution. Under the CAA, the EPA sets and enforces national standards for air pollutants. Individual states are required to develop state implementation plans explaining how the CAA will be enforced within the state. The EPA must approve each plan; if a plan is deemed unacceptable, the EPA can take over enforcement of the CAA in that state. Under Section 112 of the CAA, the agency is required to address risks from hazardous air pollutants (HAPS). This group of pollutants, often referred to as “air toxics”, includes a number of carcinogens. The 1970 CAA called for regulation of these pollutants with “an ample margin of safety”. Through 1990, only a small number of pollutants had been regulated under Section 112, including arsenic, asbestos,

benzene, beryllium, mercury, vinyl chloride, and radionuclides. The CAA was amended in 1990 (US Environmental Protection Agency (EPA), 1990), and 189 hazardous air pollutants were listed to be addressed by the EPA. A two-part strategy was called for, first using maximum available control technology (MACT) to reduce emissions from sources, and in the second phase to address residual risk, if the ample margin of safety criterion had not been met. Implicitly, risk assessment is needed to estimate the level of residual risk. The conceptual basis and methodology for this risk assessment were considered in the 1994 National Research Council Report entitled Science and Judgment in Risk Assessment (National Research Council (NRC) and Committee on Risk Assessment of Hazardous Air Pollutants, 1994). To date, MACT standards have been applied for all but 24 of the 189 listed agents (in some cases agents were removed from the list by EPA); however, progress has been slow in regulating residual risks. The Agency does track the risks of carcinogens and other agents for the purpose of risk assessment. The EPA developed and maintains the Integrated Risk Information System (IRIS), an electronic database of information on human health effects that may result from exposure to various chemicals in the environment (http://www.epa.gov/iriswebp/ iris/index.html) (US Environmental Protection Agency (EPA), 2003b). The database contains information on hundreds of chemicals, and the EPA develops an annual list of substances for IRIS assessment. IRIS was developed to provide information to be used in risk assessment and risk management in protecting public health. The database provides hazard identification and dose-response assessment information, which can be used for characterization of the public health risks of a given chemical in a particular exposure.

Pharmaceuticals The Food and Drug Administration (FDA) is the drug regulatory agency in the United States, acting under the provisions of the Federal Food, Drug and Cosmetics Act (21 CFR Chapter 9), including its major amendment, the FDA Modernization Act of 1997. These laws provide the FDA the means to fulfill its mission to protect the public health by assuring the safety, efficacy, and security of human drugs. However, the terms safety and efficacy do not have stringent definitions under this act, and there are no fixed rules or guidelines for the regulation of drug safety. Each case is allowed to undergo a unique risk-benefit review under an array of considerations including: the medical importance and utility of the drug, the extent of its usage, the severity of the disease being treated, the drug’s efficacy in treating the disease, the availability of alternative treatments, the drug’s mechanism of action, clinical pharmacologic features (such as site of metabolism), the severity of adverse health effects, the frequency of adverse health effects and risk factors of adverse health effects (Graham et al., 2000). Under these considerations, pharmaceuticals found to be associated with adverse health effects, including cancer, may not necessarily be removed from the market. Often, the regulatory decision is to amend the product information (Graham et al., 2000). In fact, some pharmaceutical agents (e.g., cancer chemotherapeutic agents) increase cancer risk, but are in use because of their benefits.

Food Additives Food additives and pesticide residues in foods are regulated under the Federal Food, Drug and Cosmetics Act. In the 1950s, the FDA began to adopt risk assessment procedures for developing an “acceptable daily intake” using animal toxicity data and uncertainty factors (Lehman and Fitzhugh, 1954). However, new concepts were evolving, largely drawn from the field of radiation carcinogenesis, that there would not be a threshold for carcinogenic risk. In response to such concern, Congress enacted the “Delaney Clause” of the Food Additive Amendments of 1958. The Delaney Clause stipulates that no additive that concentrates in food during processing or is added to food during or after processing may be allowed in the food supply if it is found to be carcinogenic in animals or in humans. In practice, this standard created difficulties, and several work-arounds were legislated by Congress. First, in 1962, Congress added the “DES proviso” which in essence allowed the approval of the use of carcinogenic veterinary

Regulating Carcinogens drugs if no residue were anticipated to remain in the food product. Second, in 1977, the FDA announced a proposal to ban the use of saccharin in foods, citing animal evidence of carcinogenicity and the Delaney Clause. At the time, saccharin was the only artificial sweetener on the market and, in response to the public’s reaction, Congress enacted a moratorium preventing the FDA from taking this action. Also, in 1977, the FDA put forward its “Sensitivity of Method” policy. This policy states that whether a residue is at or above a “zero” level is based on an analytic method that must be no more sensitive than required to detect the level that would be expected to be associated with a 1 in 1,000,000 lifetime risk of cancer (Merrill, 1988).

Pesticides For pesticides that are added inadvertently to foods, the 1958 Delaney Clause stated that no carcinogenic pesticide would be allowed if it concentrates in processed foods. This created the so-called “Delaney paradox” in which a pesticide could be “safe” when used in fresh fruits and vegetables (where the standard was risk/benefit balancing) but banned in processed foods. The 1996 Food Quality Protection Act replaced both of these standards with a health-based standard of a “reasonable certainty of no harm” for all pesticides on all types of foods. In hearings, but not in the statute, Congress stated the intent that, for a carcinogen, a 1 in 1,000,000 lifetime risk would be interpreted as meeting the standard of a reasonable certainty of no harm. Food exposures are “aggregated” with household and drinking water exposures so that, in practice, these, too are regulated under this standard. Other pesticide health risks (i.e., to pesticide applicators and manufacturers) are covered under the original risk-benefit balancing standard, so that higher risks are allowable for occupational exposures. For example, the EPA requires carcinogenicity testing routinely for pesticides used in foods although exemptions have been granted for certain classes of pesticides and for “natural” pesticides like garlic. For other pesticides, cancer testing is “triggered” by results of other studies, like structureactivity relationships, mutagenicity assays, and/or chronic toxicity studies.

Chemicals The regulation of chemicals is complicated by the multitude of statutes under which chemicals are regulated (Table 72–1). Each statute prescribes a regulatory standard, or target. Some of these standards are risk-based, for example, the establishment of pesticide food safety standards. Others require tradeoffs between risks and competing risks, for example, the regulation of disinfection byproducts in drinking water. Some require explicit consideration of feasibility, for example, standards under the Occupational Safety and Health Act, and water standards under both the Safe Drinking Water and the Clean Water Act. Others are technology driven, for example, the Clean Air Act Section 112 requirements, which require imposition of “maximum achievable control technology”. Still others, for example, the Toxic Substance Control Act (TSCA) and the regulation of pesticide nonfood risks under the Federal Insecticide, Fungicide and Rodenticide Act (FIFRA) require balancing of benefits and risks. Similarly, various statutes have required different targets for residual risk, that is, the risk that remains post regulation. In some cases (e.g., for consumer products), this residual is not clearly stated by statute or by agency policies, whereas for others, the risk targets are clearly stated and have ranged between 10-4–10-6. The EPA has established regulatory testing guidelines, under both the Toxic Substances Control Act (TSCA) and FIFRA, which are “harmonized” within the EPA and with other nations in the Organization for Economic and Cooperative Development (OECD), and they are developed in concert with other OECD countries. These test guidelines are applied to the testing of new or existing chemicals or pesticides that are on the market. The EPA requires most pesticides related to foods to have a cancer bioassay, but there are exemptions. In the case of industrial chemicals, the EPA has rarely required cancer assays and most chemicals in commerce, even those at high volume, have never been tested for carcinogenicity. At times Congress has enacted specific amendments to TSCA directed to eliminating hazards from specific substances, including the carcinogens PCBs, asbestos, and radon.

1349

The United States produces or imports almost 3000 chemicals (excluding polymers and inorganic compounds) at over 1 million pounds year; these are considered high production volume (HPV) chemicals (US Environmental Protection Agency (EPA), 2000). Some data are available for about half of these chemicals, but very little, if any, data are available for 43% of the chemicals. In fact, the EPA estimates that the health effects are known for only about 7% of the 2800 HPV chemicals. In 1998, the EPA issued a challenge to industry to provide data on these chemicals; in December 2000, this challenge was followed with a proposal for a test rule under TSCA requiring manufacturers and importers of certain HPV chemicals to conduct testing for, among others, acute toxicity, repeat dose toxicity, developmental and reproductive toxicity, and genetic toxicity (gene mutations and chromosomal aberrations). The proposed rule covers 37 HPV chemicals with substantial worker exposure that industry does not voluntarily agree to provide timely testing (US Environmental Protection Agency (EPA), 2000); it represents another screening mechanism to protect humans from exposure to carcinogens by establishing regulations at the federal level. Chemicals in waste are controlled under the Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA Superfund) and the Resource Conservation and Recovery Act (RCRA), which requires cradle-to-grave tracking of waste and site cleanup. The standards for cancer risk reduction under these statutes are shown in Table 72–1. While there is considerable overlap with TSCA and FIFRA chemicals, wastes contain many chemicals that are not listed, because they are produced inadvertently in processes of manufacture, incineration, and environmental degradation of other chemical substances. The Agency for Toxic Substances of Disease Registry (ATSDR) has published a list of the priority hazardous substances—most are carcinogens—in hazardous waste sites in the United States (275 in 2003); it is available on their website at http://www.atsdr.cdc.gov/clist.html. Since all chemicals in wastes that cause cancer cannot be identified, regulatory agencies focus on chemicals that are known “bad actors” and for which standardized testing methods have been developed. Not surprisingly, given the level of concern about cancer, these efforts have tended to target carcinogens, sometimes to the relative neglect of other classes of agents. The control of hazardous waste has been a major success in the United States, which can be tracked through the declining volumes of waste released to the environment as charted by the EPA Toxics Release Inventory (TRI). However, the TRI also shows that efforts to reduce the generation of waste have not been as effective and more of the waste is sent to EPA-regulated Class I hazardous waste disposal facilities, from which exposure to people (and therefore risk) is less likely to occur. Regulation of carcinogenic chemicals in wastes has its foundation with RCRA, which gives definitions for hazardous waste and contains a number of stringent provisions designed to prevent the release of hazardous chemicals into the environment. These provisions call for careful cradle-to-grave efforts, covering waste minimization, safe handling, regulation of disposal, and incineration of wastes. Carcinogens are covered by CERCLA, which is concerned with cleanup of highly contaminated hazardous waste sites, many consisting of large plumes of chemicals in groundwater. The United States also has undertaken a substantial effort to cleanup leaking underground storage tanks for chemicals and gasoline, which have contaminated groundwater in many communities.

Occupational Carcinogens In the United States, the Occupational Safety and Health (OSH) Act, enacted in 1970, established a new agency called the Occupational Health and Safety Administration (OSHA) within the Department of Labor. The OSH Act also established an occupational health research group called the National Institute of Occupational Safety and Health (NIOSH) that is charged with carrying out research and performing hazard assessments in support of OSHA rulemaking. When OSHA and NIOSH were established, few workplace regulations existed at the federal level. By 1979 there was already dissatisfaction with the slow pace of standard-setting by OSHA; by then, only 23 workplace

1350

PART V: CANCER PREVENTION AND CONTROL

standards had been promulgated, and only a few addressed carcinogens, prompting the General Accounting Office famously to conclude that “based on the past rate of progress, it will take 100 years for OSHA to establish needed standards for existing substances” (Smith, 1979). In contrast, according to Smith, at that time NIOSH had already listed some 28,000 possibly toxic chemicals in the workplace, of which 2200 were “suspected carcinogens” and for which NIOSH had proposed recommended exposure limits for more than 100. This decade was also the time of rising concern about environmental causes of cancer generally. In 1977 OSHA proposed a regulation that would provide a set of automatic triggers for various types of risk management actions and timetables when substances were identified as carcinogens (Ruttenberg and Bingham, 1981). The two basic tenets of this policy were that carcinogens identified in mammalian systems would be presumed carcinogens to humans “as a prudent policy matter” and in regulating a substance found to be a probable human carcinogen “permissible exposure limits will be set as low as feasible, unless there is a suitable substitute, in which case no occupational exposure will be permitted” (Ruttenberg and Bingham, 1981). The policy explicitly rejected risk assessment as a foundation for decisions. While this proposal was supported by labor unions, chemical industry opposition was great and industry formed the American Industrial Health Council (AIHC) to oppose it. However, the regulatory landscape vastly changed in July 1980 when the US Supreme Court handed down the so-called “Benzene decision,” which struck down OSHA’s regulation of the carcinogen benzene on the basis that the “benefits” of the rule (numbers of cases of cancers averted) were not sufficienty quantified to justify the burdens imposed by the rule (Smith, 1980). As a result, a risk assessment was carried out for benzene (White et al., 1980). Although there was optimism at the time, during the 1980s, with a change in presidential administrations the carcinogen rule was not made final. Consequent to the Court-imposed imposition of new requirements for quantitative risk assessment, in combination with OSHA’s statutory language and lack of political will (Epstein, 1979), only a few new standards for occupational carcinogens have been produced over more than two decades.

occupational studies and animal bioassays. In 1983, New Jersey amended the State Safe Drinking Water Act to establish MCLs for chemical pollutants. Since no “safe” level for carcinogen exposure can be determined, the Act required that standards “with respect to carcinogens, permit cancer in no more that one in one million persons ingesting that chemical for a lifetime”. Recognizing the great public health benefits of chlorination, the Act required that carcinogenic chemical byproducts be eliminated “within the limits of practicability and feasibility”. In 1986 California passed Proposition 65, the Safe Drinking Water and Toxic Enforcement Act, which defined de minimis risk as “a theoretical lifetime risk of up to one excess case of cancer in a population of 1,000,000 people—the 10-6 risk level”. These approaches continue to serve as models for regulatory efforts to control carcinogen exposure. Today the control of carcinogens continues to drive the public health approach to drinking water protection. Table 72–4 lists the current EPA national standards for chemical carcinogens in water. The recent controversy in the United States concerning arsenic in drinking water illustrates the potential pivotal role of epidemiological data in standard setting for drinking water. Much of the debate concerning the standard for arsenic in drinking water stems from the cancer risk estimate, which is now based in analyses of the studies carried out in Taiwan (National Research Council (NRC), 2001). The evidence linking arsenic to cancer comes from both clinical observation and from a series of studies in Taiwan, and more recently in other countries, where deep wells provide water with high arsenic levels, substantially higher than those under consideration for the MCL. In a 2001 report from a National Research Council Committee (National Research Council (NRC), 2001), multiple data sets were analyzed, including a series of sensitivity analyses. The resulting models were used to estimate the number of cancer cases attributable to arsenic at different MCLS.

Drinking Water

Although this chapter deals primarily with the carcinogen regulations for the United States, it is, of course, an international issue. For Europe and elsewhere, the International Programme on Chemical Safety (IPCS) implements activities related to chemical safety, including carcinogens (World Health Organization, 2004). The World Health Organization is the executing agency of the IPCS, whose main roles are to establish the scientific basis for safe use of chemicals, and to strengthen national capabilities and capacities for chemical safety. In 1972, the United Nations held a conference in Stockholm on the human environment. An outcome from the conference was the formulation of the IPCS for the early warning and prevention of harmful effects of chemicals to which humans were being increasingly exposed, and for the assessment of the potential risks to human health. The first Memorandum of Understanding establishing the cooperation of these three organizations was signed in April 1980. More information on the IPCD and its activities can be found on its website, http://www.who.int/pcs/. As stated previously, the European Commission has established a list of criteria for causation with regard to an agent being labeled a carcinogen. In addition, the commission monitors such related activities as food safety. The Commission’s website address is http://europa.eu.int/comm.index_en.htm, and contains more information on its approaches.

From the initial discovery of industrial and agricultural chemicals in drinking water from the Mississippi in the early 1970s (EPA Region 6, 1972), to the recent debate concerning regulation of arsenic in ground water, cancer risk has dominated the national approach to regulating drinking water in the United States. Historically, the protection of drinking water focused upon control of pathogens and disinfection to prevent infectious diseases. In the 1960s there was increasing awareness of the vulnerability of both surface and ground water drinking sources to industrial and agricultural chemical contamination. In 1972 EPA found 36 chemicals in drinking water from treatment plants along the Mississippi River, spurring national concern about the potential presence of carcinogens in the nation’s drinking water supplies. In 1974 unintended chemical byproducts of chlorination, including chloroform, were found in drinking water. This concern led to the passage of the Safe Drinking Water Act of 1974, which provides the regulatory framework for the control of drinking water contamination. Epidemiological evidence has been one impetus for regulation. A National Cancer Institute (NCI) study in 1978 found associations between disinfection byproducts and cancer mortality (Cantor et al., 1978). Early studies were ecological in design. During the same time period, advances in analytical chemistry made it possible to detect increasing numbers of chemical pollutants at concentrations in the previously unobservable part per billion range. Several national and state surveys of drinking water confirmed the widespread presence of potentially carcinogenic contaminants (Burke and Tucker, 1978). While the epidemiological evidence was limited, the potential public health impacts of wide-scale population exposure to carcinogens in drinking water were apparent. Control of carcinogens in drinking water became the centerpiece of national and state water pollution control efforts. Because of the limited availability of epidemiological data on drinking water contamination, the health standards, called Maximum Contaminant Levels (MCLs), have been based largely upon

CARCINOGEN REGULATION OUTSIDE THE UNITED STATES

CONCLUSIONS AND IMPLICATIONS In the United States and some other countries, an elaborate web of regulations is in place to protect the public from exposure to selected environmental carcinogens. The underlying principles vary, but all of the regulations have the goal of controlling exposures that arise from sources beyond the individual’s control. These regulations typically reflect a complex balancing of risk, feasibility of control, costs, and

1351

Regulating Carcinogens

Table 72–4. Current EPA Standards for Chemical Contaminants in Drinking Water that Cause Cancer (Units are in milligrams per liter (mg/L) unless otherwise noted) Contaminants

MCLG1,4 (mg/L)

MCL2 (mg/L)

Potential Health Effects from Ingestion of Water

Sources of Contaminant in Drinking Water

volatile organics Benzene Carbon Tetrachloride p-Dichlorobenzene 1,2-Dichloroethane 1,1-Dichloroethylene Trichloroethylene Vinyl Chloride

Zero Zero 0.075 Zero 0.007 Zero Zero

0.005 0.005 0.075 0.005 0.007 0.005 0.002

Cancer Cancer Cancer Cancer Cancer Cancer Cancer

Some foods; gas, drugs, pesticides, paint, plastic industries Solvents and their degradation products Room and water deodorants, and “mothballs” Leaded gasoline, fumigants, paints Plastics, dyes, perfumes, paints Textiles, adhesives, and metal degreasers May leach from PVC pipe; formed by solvent breakdown

0.006 7MFL

0.006 7MFL

Cancer Cancer

Fire retardants, ceramics, electronics, fireworks, solder Natural deposits; asbestos cement in water systems

Zero Zero Zero Zero Zero Zero

TT3 0.002 0.002 0.0002 0.005 0.005

Cancer, nervous Cancer Cancer Cancer Cancer Liver, kidney cancer

Polymers used in sewage/wastewater treatment system effects Runoff from herbicide on corn, soybeans, other crops Leaching from soil treatment for termites Soil fumigant on soybeans, cotton, pineapple, orchards Paint stripper, metal degreaser, propellant, extraction Soil fumigant; waste industrial solvents effects

Zero Zero Zero Zero Zero

0.00000003 TT3 0.00005 0.0004 0.0002

Cancer Cancer Cancer Cancer Cancer

Chemical production byproduct; impurity in herbicides Water treatment chemicals; waste epoxy resins, coatings Leaded gasoline additives; leaching of soil fumigant Leaching of insecticide for termites, very few crops Biodegradation of heptachlor

Hexchlorobenzene PAHs (benzo(a)pyrene) PCBs Pentachlorophenol

Zero Zero Zero Zero

0.001 0.0002 0.0005 0.001

Pesticide production waste by-product Coal tar coatings; burning organic matter; volcanoes, fossil fuels Coolant oils from electrical transformers; plasticizers Wood preservatives, herbicide, cooling tower wastes

Phthalate, (di (2-ethylhexyl)) Simazine Tetrachloroethylene

Zero 0.004 Zero

0.006 0.004 0.005

Cancer Cancer Cancer Liver and kidney effects, and cancer Cancer Cancer Cancer

Zero

0.003

Cancer

Insecticide on cattle, cotton, soybeans; canceled in 1982

Zero Zero Zero Zero

4 mrem/yr 15 pCi/L 5 pCi/L 0.10

Cancer Cancer Bone cancer Cancer

Decay of radionuclides in natural and man-made deposits Decay of radionuclides in natural deposits Natural deposits Drinking water chlorination by-products

inorganics Antimony Asbestos (>10 m)

organics (1 of 4) Acrylamide Alachlor Chlordane Dibromochloropropane Dichloromethane 1,2-Dichloropropane

organics (2 of 4) Dioxin Epichlorohydrin Ethylene dibromide Heptachlor Heptachlor epoxide

organics (3 of 4)

PVC and other plastics Herbicide on grass sod, some crops, aquatic algae Improper disposal of dry cleaning and other solvents

organics (4 of 4) Toxaphene Other Interim Standards Beta/photon emitters Alpha emitters Combined Radium 226/228 Total Trihalomethanes

1 Maximum Contaminant Level Goal (MCLG): The maximum level of a contaminant in drinking water at which no known or anticipated adverse effect on the health effect of persons would occur, and that allows for an adequate margin of safety. MCLGs are non-enforceable public health goals. 2 Maximum Contaminant Level (MCL): The maximum permissible level of a contaminant in water which is delivered to any user of a public water system. MCLs are enforceable standards. The margins of safety in MCLGs ensure that exceeding the MCL slightly does not pose significant risk to public health. 3 Treatment Technique: An enforceable procedure or level of technical performance that public water systems must follow to ensure control of a contaminant. 4 MCLGs were not established before the 1986 Amendments to the Safe Drinking Water Act. Therefore, there is no MCLG for this contaminant. MFL, million fibers per liter; pCi, picocuries.

the force of political and other societal pressures. In spite of the inherent complexity and heterogeneity of these regulations, they have proven to be generally effective with some notable successes, such as drastic reduction of worker exposure to asbestos. Epidemiological evidence has figured centrally in regulating many substances and will continue to do so, given the prominent role of epidemiology in identifying carcinogens. References Ames BN, Magaw R, Gold LS. 1987. Ranking possible carcinogenic hazards. Science 236:271–280. Anderson L, Donovan P, Rice J. 1985. Risk assessment for transplacental carcinogenesis. In: Li A, ed. New approaches in toxicity testing and their application in human risk assessment. New York: Raven Press, pp. 179–202. Anderson LM, Diwan BA, Fear NT, Roman E. 2000. Critical windows of exposure for children’s health: Cancer in human epidemiological studies and neoplasms in experimental animal models. Environ Health Perspect 108 Suppl 3:573–594.

Brenner DJ, Doll R, Goodhead DT, et al. 2003. Cancer risks attributable to low doses of ionizing radiation: Assessing what we really know. Proc Natl Acad Sci U S A 100:13761–13766. Bridbord K, Decoufle P, Fraumeni JF, et al. 1978. Estimates of the fraction of cancer in the United States related to occupational factors. Natl Cancer Inst Monogr 1–39. Burke TA, Tucker RK. 1978. A preliminary report on the findings of the state groundwater monitoring project. Trenton, N.J., New Jersey Department of Environmental Protection. Cantor KP, Hoover R, Mason TJ, McCabe LJ. 1978. Associations of cancer mortality with halomethanes in drinking water. J Natl Cancer Inst 61: 979–985. Caufield C. 1989. Multiple Exposures. Chronicles of the Radiation Age. London, England: Stoddard Publishing, Ltd. Daher A, Varin M, Lamontagne Y, Oth D. 1998. Effect of pre-conceptional external or internal irradiation of N5 male mice and the risk of leukemia in their offspring. Carcinogenesis 19:1553–1558. Epstein SS. 1979. The Politics of Cancer. Garden City, NY: Anchor Press. Giusti RM, Iwamoto K, Hatch EE. 1995. Diethylstilbestrol revisited: A review of the long-term health effects. Ann Intern Med 122:778–788.

1352

PART V: CANCER PREVENTION AND CONTROL

Goldstein BD. 2001. The precautionary principle also applies to public health actions. Am J Public Health 91:1358–1361. Graham DJ, Waller PC, Kurz X. 2000. A view from regulatory agencies. In: Strom BL, ed. Pharmacoepidemiology. Hoboken, NJ: John Wiley & Sons, pp. 109–124. Graham JD. The role of precaution in risk management (Speech). 6-20-2002. White House, Washington, D.C.: Office of Management and Budget, Office of Information and Regulatory Affairs. Hendee WR, Edwards FM. 1986. ALARA and an integrated approach to radiation protection. Semin Nucl Med 16:142–150. Herbst AL. 1987. The effects in the human of diethylstilbestrol (DES) use during pregnancy. Princess Takamatsu Symp 18:67–75. Interagency Regulatory Liaison Group (IRLG), Working Group on Risk Assessment. 1979. Scientific bases for identification of potential carcinogens and estimation of risks. J Natl Cancer Inst 63:241–268. International Commission on Radiological Protection (ICRP). 1977. Recommendations of the International Commission on radiological protection. Ann ICRP 1:3. Jamieson D, Wartenberg D. 2001. The precautionary principle and electric and magnetic fields. Am J Public Health 91:1355–1358. Kocher DC. 1991. Perspective on the historical development of radiation standards. Health Phys 61:519–527. Kriebel D, Tickner J. 2001. Reenergizing public health through precaution. Am J Public Health 91:1351–1355. Kriebel D, Tickner J, Epstein P, et al. 2001. The precautionary principle in environmental science. Environ Health Perspect 109:871–876. Lehman AJ, Fitzhugh DG. 1954. 100-fold margin of safety: Quarterly report to the editor on topics of current interest. Association of Food and Drug Officials Quarterly Bulletin 18:33–35. Lichtenstein P, Holm NV, Verkasalo PK, et al. 2000. Environmental and heritable factors in the causation of cancer—analyses of cohorts of twins from Sweden, Denmark, and Finland. N Engl J Med 343:78–85. Lowrance WW. 1976. Of Acceptable Risk: Science and the Determination of Safety. Los Altos, CA: William Kaufmann, Inc. McHale CM, Wiemels JL, Zhang L, et al. 2003. Prenatal origin of TEL-AML1positive acute lymphoblastic leukemia in children born in California. Genes Chromosomes Cancer 37:36–43. Merrill RA. 1988. FDA’s implementation of the Delaney Clause: Repudiation of congressional choice or reasoned adaptation to scientific progress? Yale Journal on Regulation. National Council on Radiation Protection and Measurements. 1954. Permissible dose from external sources of ionizing radiation. NCRP Report No. 17 (National Bureau of Standards Handbook). Washington, D.C.: U.S. Government Printing Office. National Council on Radiation Protection and Measurements (NCRP). 1987. Ionizing Radiation Exposure of the Population of the United States. 93, 1-85. Bethesda, MD: National Council on Radiation Protection and Measurments. National Research Council (NRC). 2001. Arsenic in drinking water: 2001 update. Washington, D.C.: National Academies Press. National Research Council (NRC), Committee on Health Risks of Exposure to Radon, Board on Radiation Effects Research, Commission on Life Sciences. 1998. Health Effects of Exposure to Radon (BEIR VI). Washington, D.C.: National Academies Press. National Research Council (NRC), Committee on Risk Assessment of Hazardous Air Pollutants. 1994. Science and Judgment in Risk Assessment. Washington, D.C.: National Academies Press. National Research Council (NRC), Committee on the Biological Effects of Ionizing Radiation. 1988. Health Risks of Radon and other Internally Deposited Alpha-Emitters: BEIR IV. Washington, D.C.: National Academies Press. National Research Council (NRC), Committee on the Biological Effects of Ionizing Radiation. 1990. Health Effects of Exposure to Low levels of Ionizing Radiation: BEIR V. Washington, D.C.: National Academies Press. National Research Council (NRC), Committee on the Institutional Means for Assessment of Risks to Public Health. 1983. Risk Assessment in the Federal Government: Managing the Process. Washington, D.C.: National Academies Press. National Research Council (NRC), Committee to Assess Health Risks from Exposure to Low Levels of Ionizing Radiation, 2005. Health risks from exposure to low levels of ionizing radiation: BEIR VII Phase 2. Washington, D.C.: National Academies Press. National Toxicology Program (NTP). 2004. Tenth report on carcinogens. Washington, D.C.: Public Health Service, US Department of Health and Human Services. Occupational Safety and Health Administration (OSHA). 1989. Title 29, Code of Federal Regulations, Part 1910.1000. Washington, D.C.: U.S. Government Printing Office.

Office of Environmental Health Hazard Assessment, State of California. 1986. Proposition 65, the Safe Drinking Water and Toxic Enforcement Act of 1986. State of California. Olshan AF, Anderson L, Roman E, et al. 2000. Workshop to identify critical windows of exposure for children’s health: Cancer work group summary. Environ Health Perspect 108 Suppl 3:595–597. Palanza P, Parmigiani S, vom Saal FS. 2001. Effects of prenatal exposure to low doses of diethylstilbestrol, o, p’DDT, and methoxychlor on postnatal growth and neurobehavioral development in male and female mice. Horm Behav 40:252–265. Patterson JT. 1989. The Dread Disease: Cancer and Modern American Culture. Cambridge, MA: Harvard University Press. Proceedings of the Wingspread Conference on Strategies for Implementing the Precautionary Principle. Wingspread statement on the precautionary principle. 1998. Proctor RN. 1995. Cancer Wars. How Politics Shapes What We Know and Don’t Know About Cancer. New York, New York: Basic Books. Rogers MD. 2003a. Risk analysis under uncertainty, the precautionary principle, and the new EU chemicals strategy. Regul Toxicol Pharmacol 37:370–381. Rogers MD. 2003b. The European Commission’s White Paper “strategy for a future chemicals policy”: A review. Risk Anal 23:381–388. Ruttenberg R, Bingham E. 1981. A comprehensive occupational carcinogen policy as a framework for regulatory activity. Ann N Y Acad Sci 363:13–20. Samet JM, Schnatter R, Gibb H. 1998. Epidemiology and risk assessment. Am J Epidemiol 148:929–936. Savitz DA, Chen JH. 1990. Parental occupation and childhood cancer: Review of epidemiologic studies. Environ Health Perspect 88:325–337. Smith JR. 1979. OSHA carcinogen policy delayed. Science 203:35. Smith RJ. 1980. A light rein falls on OSHA. Science 209:567–568. Tickner JA. 2002. Precautionary principle encourages policies that protect human health and the environment in the face of uncertain risks. Public Health Rep 117:493–497. Tomatis L. 1989. Overview of perinatal and multigeneration carcinogenesis. IARC Sci Publ 1–15. Tomatis L, Narod S, Yamasaki H. 1992. Transgeneration transmission of carcinogenic risk. Carcinogenesis 13:145–151. Tran NL, Locke PA, Burke TA. 2000. Chemical and radiation environmental risk management: Differences, commonalities, and challenges. Risk Anal 20:163–172. United Nations Conference on Environment and Development. 1992. Rio declaration on environment and development. U.N. Doc./CONF.151/5/Rev.1. 6-13-1992. US Department of Agriculture. 2000. US food safety system country report. 33-2000. Washington, D.C. US Department of Agriculture. US Environmental Protection Agency (EPA). 1996. Proposed Guidelines for Carcinogen Risk Assessment. EPA/600/P-92/003C. 1996. Washington D.C.: Office of Research and Development. US Environmental Protection Agency (EPA). 1986. Guidelines for carcinogen risk assessment. Fed Reg 51:33992–34003. US Environmental Protection Agency (EPA). 1988. Assessing asbestos exposure in public buildings. Washington, D.C.: U.S. Environmental Protection Agency. US Environmental Protection Agency (EPA). 1990. Clean Air Act 1990. U.S. Environmental Protection Agency. US Environmental Protection Agency (EPA). 1992a. A Citizen’s Guide to Radon. What it is and what to do about it. Washington D.C.: U.S. Government Printing Office. US Environmental Protection Agency (EPA). 1992b. Technical Support Document for the 1992 Citizen’s Guide to Radon. Washington, D.C.: U.S. Government Printing Office. US Environmental Protection Agency (EPA). 2000. Testing of certain high production volume chemicals. Fed Reg 65:81686–81698. US Environmental Protection Agency (EPA). 2003a. EPA assessment of risks from radon in homes. EPA 402-R-03-003. Washington, D.C.: U.S. Environmental Protection Agency. US Environmental Protection Agency (EPA). 2003b. Integrated risk information system (IRIS). US Environmental Protection Agency, Office of Research and Development, National Center for Environmental Assessment. 12-22-2003b. US Environmental Protection Agency (EPA). 2003c. The ABCs of asbestos in schools. Washington, D.C.: US Environmental Protection Agency; Office of Pollution Prevention and Toxics. US Environmental Protection Agency (EPA), Risk Assessment Forum. 2003. Draft final guidelines for carcinogen risk assessment (External review draft, February 2003). NCEA-F-0644A. 3-3-2003. Washington, D.C.: US Environmental Protection Agency.

Regulating Carcinogens US Environmental Protection Agency (EPA), Risk Assessment Forum Technical Panel. 2003. Supplemental guidance for assessing cancer susceptibility from early-life exposure to carcinogens. EPA/630/R-03/003 External review draft. 2003. Washington, D.C.: US Environmental Protection Agency.

1353

White MC, Infante PF, Walker B, Jr. 1980. Occupational exposure to benzene: A review of carcinogenic and related health effects following the U.S. Supreme Court decision. Am J Ind Med 1:233–243. World Health Organization. 2004. International Programme on chemical Safety (IPCS). World Health Organization. 3-24-2004.

This page intentionally left blank

Index

Note: Page numbers followed by f indicate figures; page numbers followed by t indicate tables. ABO, 584t ABO blood group choriocarcinoma and, 1081, 1083 stomach cancer and, 714 Abortion acute lymphocytic leukemia and, 856 breast cancer and, 998 choriocarcinoma and, 1081–1082 endometrial cancer and, 1034 Abscess, anal, 836 ACE, 584t Acetaldehyde, 243, 244, 330t Acetaldehyde dehydrogenase, 244, 702, 775 Acetamide, occupational exposure to, 330t Acetaminophen, 491–492 bladder cancer and, 1112 renal cancer and, 1091 renal pelvis cancer and, 1095 ureter cancer and, 1095 N-Acetylcysteine, in oral cavity cancer, 686 N-Acetyltransferase acute myeloid leukemia and, 851 bladder cancer and, 228, 1115 breast cancer and, 230, 1004f, 1004t, 1005 liver cancer and, 775 lung cancer and, 647 renal cancer and, 1090 N-Acetyltransferase-2, prostate cancer and, 1139 Achalasia, 701 Acoustic neuroma, 1187 Acquired immunodeficiency syndrome. See Human immunodeficiency virus (HIV) infection Acromegaly colorectal cancer and, 820–821 small intestine cancer and, 804 Acrylamide occupational exposure to, 328t pancreatic cancer and, 746 Acrylonitrile, occupational exposure to, 330t, 335t Actinic cheilitis, 686

Acute lymphocytic leukemia. See Leukemia, lymphocytic, acute Acute myeloid leukemia. See Leukemia, myeloid, acute ADAM33, 585t Addiction alcohol, 243 nicotine, 222, 223f, 225 Adenoma bile duct, 763 colorectal, 22–23, 27–29, 28f, 425, 809 chemoprevention of, 37 descriptive epidemiology of, 28 etiology of, 29 obesity and, 425 pathology of, 28, 28f progression of, 29 screening for, 823 hepatocellular, 763, 764t small intestine, 802 ADH2 (ADH1B) alcohol metabolism and, 244 laryngeal cancer and, 633 oral cavity cancer and, 684 ADH3 (ADH1C), 584t alcohol metabolism and, 244 oral cavity cancer and, 684 Adoption study, 89, 91 Adrenal gland cancer, 11t Adriamycin, 491t Adult T-cell leukemia/lymphoma, 508f, 529–531, 531f, 853–854 gender and, 530 genetic factors in, 530 pathogenesis of, 530 socioeconomic environment and, 530–531, 531f strongyloidiasis and, 530 Aflatoxin chemoreduction of, 778 liver cancer and, 249, 516, 769–770, 777 occupational exposure to, 327t

Age, 106 anal cancer and, 831, 833f Barrett’s esophagus and, 27 bladder cancer and, 165f, 1102, 1105f brain cancer and, 165f, 1179, 1179f breast cancer and, 113, 114f, 115, 155, 165f, 996–997, 1002 cancer incidence and, 162–164, 165f cancer precursors and, 22 cervical cancer and, 124, 126f, 128, 165f, 1049, 1049t, 1050 choriocarcinoma and, 1079 colorectal cancer and, 165f, 812–813, 813f endometrial cancer and, 1028, 1028t, 1029f endometrial hyperplasia and, 35 Epstein-Barr virus infection and, 510 esophageal cancer and, 165f, 697, 698f familial adenomatous polyposis and, 22 Helicobacter pylori infection and, 533 hepatitis B virus infection and, 513 Hodgkin lymphoma and, 165f, 875, 875f, 876f hydatidiform mole and, 1079 immune system changes and, 550–551, 555 leukemia and, 165f, 841, 844, 847f, 854 liver cancer and, 766 lung cancer and, 165f, 641f melanoma and, 165f, 1200, 1202f, 1219 multiple myeloma and, 165f, 919, 920f non-Hodgkin lymphoma and, 165f, 901, 902t oral cavity cancer and, 165f, 675, 676f ovarian cancer and, 165f, 1013, 1014f pancreatic cancer and, 165f, 724, 724f penile cancer and, 1166, 1167f pharyngeal cancer and, 165f, 675, 676f, 676t prostate cancer and, 165f, 1129–1130 skin cancer and, 1231, 1232f soft tissue sarcoma and, 861t, 961 testicular cancer and, 1156 thyroid cancer and, 979 Age standardization, 103 Age-standardized rate, 103

1355

1356 Agent Orange choriocarcinoma and, 1081, 1082–1083 testicular cancer and, 1155 AIDS. See Human immunodeficiency virus (HIV) infection Air pollution, 355–376 case-control studies of, 356, 361t, 362t cohort studies of, 356, 361t, 362t–363t in developing countries, 363, 375–376 ecologic studies of, 356, 362t exposure assessment methods for, 356–357, 356t indoor, 370–375 asbestos and, 370, 373 radon and, 275–276, 373–374 secondhand smoke and, 374–375 microenvironments of, 355–356 outdoor, 357–370, 357t aldehydes in, 359 asbestos in, 359–360 bladder cancer and, 363–370, 365t–370t brain cancer and, 370, 1189 1,3-butadiene in, 359 childhood cancer and, 370, 371t–373t combustion products in, 358–359 in developing countries, 363, 375–376 diesel exhaust in, 359, 363–370, 364f, 365t–370t fibers in, 359–360 lung cancer and, 360–364, 361t, 362t–363t, 364f, 644 particles in, 358–359 point sources of, 359 polycyclic organic matter in, 358 radionuclides in, 359 regulation of, 1348 risk attribution and, 363 rural, 357–358, 357t, 361t urban, 357–358, 357t, 361t risk assessment for, 357 total personal exposure to, 355, 356 Airline flight crews, melanoma in, 1209 Alanine aminotransferase, in hepatitis C virus infection, 517 Alcohol, 243–251 acute lymphocytic leukemia and, 856 acute myeloid leukemia and, 850 Barrett’s esophagus and, 27 biliary tract cancer and, 794 bladder cancer and, 1111 brain cancer and, 1188 breast cancer and, 250, 251t, 1000, 1006t, 1007 carcinogenic mechanisms of, 243–244 case-control study of, 245 choriocarcinoma and, 1080 cigarette smoking and, 232, 243, 246–247, 248, 250–251 colorectal adenoma and, 29, 249 colorectal cancer and, 195, 248–249, 251t, 816, 816f, 818 elimination of, 244–245 endometrial cancer and, 1035 esophageal cancer and, 247–248, 251t, 699–700, 699t food consumption with, 244 large bowel cancer and, 251t laryngeal cancer and, 246–247, 251t, 629–630, 633 leukoplakia and, 25

Index lip cancer and, 687 liver cancer and, 249, 251t, 516, 770–771, 777 lung cancer and, 250–251, 251t, 643 melanoma and, 1210 metabolism of, 244–245 multiple cancers and, 1272–1273 multiple myeloma and, 939 nasopharyngeal cancer and, 623–624 neuroblastoma and, 1257 non-Hodgkin lymphoma and, 909 nutrient effects of, 243 oral cavity cancer and, 245–246, 251t, 680–681, 683 ovarian cancer and, 1016 oxidation of, 244 pancreatic cancer and, 249–250, 251t, 741–742 pharyngeal cancer and, 245–246, 251t, 683 prospective study of, 245 prostate cancer and, 251, 251t, 1135 renal cancer and, 1092 renal pelvis cancer and, 1095 salivary gland cancer and, 688 sinonasal cancer and, 614 skin cancer and, 1241 socioeconomic status-cancer association and, 182 solvent properties of, 243 stomach cancer and, 248, 251t, 712 study design for, 245 thyroid cancer and, 987 underreported consumption of, 245 ureter cancer and, 1095 vaginal cancer and, 1071 vulvar cancer and, 1071 Aldehydes, ambient air, 359 ALDH, 244 ALDH2, 584t, 589 colorectal cancer and, 819, 821 laryngeal cancer and, 633 Alkylating agents. See Antineoplastics Allele, 565 Allelic loss study, 96 Allergy, 551, 554 brain cancer and, 1187 Hodgkin lymphoma and, 554 multiple myeloma and, 925t, 929 non-Hodgkin lymphoma and, 911 pancreatic cancer and, 749–750 Allium compounds, stomach cancer and, 712, 712t Alpha-fetoprotein in hepatocellular carcinoma, 777 in testicular cancer, 1162t Aluminum bladder cancer and, 1109, 1109t occupational exposure to, 333t, 336t Alveolar soft part sarcoma, 960t 2-Amino-5-(5-nitro-2-furyl)-1,3,4-thiadiazole, 491t ortho-Aminoazotoluene, occupational exposure to, 331t para-Aminoazotoluene, occupational exposure to, 331t 4-Aminobiphenyl bladder cancer and, 1114–1115 occupational exposure to, 327t Amitriptyline, 493

Amphetamines, non-Hodgkin lymphoma and, 553 Ampulla of Vater, 787 cancer of, 788, 793, 795. See also Biliary tract cancer Amsacrine, 491t Anabolic steroids, liver cancer and, 772 Anal cancer, 830–838 anatomy of, 830 body mass index and, 836 cigarette smoking and, 835 classification of, 11t, 830 Crohn disease and, 836 familial, 835 genetic factors in, 835 gonorrhea and, 834–835 herpes simplex virus infection and, 834 heterosexual anal intercourse and, 833 histology of, 830 hormones and, 836 human immunodeficiency virus infection and, 832–833, 834, 836 human papillomavirus infection and, 833–834, 834t, 835–836, 837, 838 immune function and, 835–836 incidence of age and, 831, 833f international, 831, 831f, 832f martial status and, 831 race and, 831, 831f U.S., 147t, 830–831, 831f inflammatory bowel disease and, 836 male homosexuality and, 831–833 metachronous cancers and, 836–837 mortality from, 149t multiple cancers and, 1276 natural history of, 837 pathogenesis of, 837 post-transplantation, 552 precursor lesions of, 836, 837 prevention of, 837–838 screening for, 838 sexual practices and, 832–833 survival for, 831 syphilis and, 834–835 tobacco and, 835 vaginal cancer and, 1072 vulvar cancer and, 1072 Anal intercourse, anal cancer and, 832–833 Analgesics, 490–494, 491t bladder cancer and, 1112 ovarian cancer and, 1015, 1016t renal cancer and, 1091 renal pelvis cancer and, 1095 ureter cancer and, 1095 Analogy, causation and, 5t, 6–7 Anaphase-promoting complex or cyclosome (APC-C), 54–55, 55f Anastrozole, 1331–1332 Androgen(s) breast cancer and, 1002–1103 hepatocellular carcinoma and, 772, 775 ovarian cancer and, 1020–1021 physical activity effects on, 450 prostate cancer and, 1136–1137 Androgen receptor, in pancreatic cancer, 750 Androgen receptor gene, CSG repeat of, prostate cancer and, 1136–1137 Anemia Fanconi, oral cavity cancer and, 683

Index pernicious multiple myeloma and, 928 pancreatic cancer and, 750 Anesthetics, 491t Angiogenesis, 58–60, 59f, 584t Angiosarcoma, 764t arsenic and, 773 incidence of, 765–766, 766t Thorotrast and, 773 vinyl chloride and, 338, 339t, 773 Angiotensin-converting enzyme (ACE) inhibitors, 498 Ankylosing spondylitis, radiotherapy in, 262, 848 Antacids, 499, 701 Anthropometry, for diet studies, 412 Anti-inflammatory agents, 490–494, 491t Anti-ulcer drugs, 499, 713 Antibacterials, 491t, 493 Anticonvulsants, 498–499 Antidepressants, 493–494 Antifungals, 491t, 495 Antihypertensives, 497, 1090–1091 Antimony, occupational exposure to, 333t Antimony trioxide, occupational exposure to, 330t Antineoplastics, 491t, 494–495, 1277–1278, 1278t acute myeloid leukemia and, 849–850 bone cancer and, 952 occupational exposure to, 495 Antioxidants. See also Vitamin C; Vitamin E cervical cancer and, 1052 colorectal cancer and, 823 Helicobacter pylori infection and, 711 multiple cancers and, 1273 ovarian cancer and, 1017 prostate cancer and, 1131–1132 thyroid cancer and, 986 Antiparasitics, 491t, 495 Antiprotozoals, 491t, 495 Antipyretics, 490–494, 491t Antithyroid drugs, 495–496 a1-Antitrypsin deficiency, 775–776 Antivirals, 491t Anus anatomy of, 830 benign lesions of, 836 cancer of. See Anal cancer APC, 585t biliary tract cancer and, 787 colorectal cancer and, 809–811, 810f, 818, 821, 821f familial adenomatous polyposis and, 570 APOE, 585t Apoptosis, 56–58, 57f alterations in, 58 calcium channel blocker effect on, 498 mitochondria in, 58 Apoptosis-inducing factor (AIF), 57f, 58 Appendectomy, 553, 889 Appendiceal cancer, 146t Appliances, electric adult exposure to, 312, 313, 313t magnetic fields of, 307–308, 312 occupational exposure to, 313–315 in utero exposure to, 312–313, 313t Aramite, occupational exposure to, 331t Areca nut, 25, 680 Argentaffinoma, 801

Argentina, migrants to, 193 bladder cancer in, 198 Aromatase inhibitors, in breast cancer prevention, 1331–1332 Aromatic amines, occupational exposure to, 327t, 331t, 335t, 339t, 1107–1108 Arsenic ambient air, 357t drinking water, 382–386, 383t, 385f bladder cancer and, 383–384, 383t, 1110, 1117 cytogenetic studies of, 386 exposure assessment of, 384, 386, 386f genetic susceptibility to, 386 lung cancer and, 383t, 384, 385f non-neoplastic consequences of, 382–383 nutritional susceptibility to, 384, 386 prostate cancer and, 383t, 384, 385f renal cancer and, 383, 383t, 385f renal pelvis cancer and, 1095 skin cancer and, 1239 ureter cancer and, 1095 genotoxicity of, 386 hepatic angiosarcoma and, 773 lung cancer and, 383, 383f, 385f, 644 medicinal, 491t, 496 metabolism of, 384, 386f occupational exposure to, 326t, 333t, 335t, 339t selenium interaction with, 386 skin cancer and, 383, 383t, 384, 385f, 1239 sources of, 382 urinary, 384, 386 Arsenic trioxide, 382, 496 Asbestos ambient air indoor, 370, 373 outdoor, 357t, 359–360 amosite, 661, 664f, 664t crocidolite, 661, 664f, 664t drinking water, 394 EPA risk estimates for, 373 esophageal cancer and, 702 laryngeal cancer and, 630 lung cancer and, 337, 360, 644 mesothelioma and, 661–666, 661t, 667f community-based studies of, 664, 665t dose-response relationship in, 661, 664, 664t fiber type and, 661, 664f, 664t industry-based studies of, 661, 662t–663t, 664f non-occupational exposure and, 665, 666t pleural plaques in, 664–665 665t multiple myeloma and, 938–939 non-Hodgkin lymphoma and, 908 non-occupational exposure to, 665, 666t occupational exposure to, 326t, 333t, 336t, 337, 339t, 360, 370, 373 pancreatic cancer and, 744–745 renal cancer and, 1092–1093 use patterns of, 666, 666f, 667f Aspirin, 492–493 bladder cancer and, 1112 colorectal cancer and, 37, 823, 1330 Hodgkin lymphoma and, 890 pancreatic cancer and, 742–743 prostate cancer and, 1135 renal cancer and, 1091

1357 renal pelvis cancer and, 1095 ureter cancer and, 1095 Association, strength of, 5t, 6. See also Causation Association study, 95 case-control, 93t, 95 family-based, 93t, 95 inconsistent results in, 96 linkage study vs., 95–96 Asthma lung cancer and, 642 multiple myeloma and, 925t, 929 pancreatic cancer and, 749 Astrocytoma. See Brain cancer Ataxia-telangiectasia, 53, 161, 551–552, 551t, 563t, 566 breast cancer and, 1003 lymphoproliferative disorders and, 858–859 non-Hodgkin lymphoma and, 910 ATM, 566, 585t ATM (ataxia telangiectasia mutation)-related (ATR) protein, 53, 53f, 54 ATM (ataxia telangiectasia mutation) protein, 53, 53f Atomic bomb, 267–269, 268f, 269f acute lymphocytic leukemia and, 268, 268f, 854–855 acute myeloid leukemia and, 268, 268f, 848 bladder cancer and, 1114 brain cancer and, 1186 breast cancer and, 268, 1002 chronic myeloid leukemia and, 268, 268f, 852 dose-response relationship in, 269, 269f dosimetry for, 267–268 lung cancer and, 268 multiple myeloma and, 931, 932t myelodysplastic syndromes and, 848 ovarian cancer and, 1019 thyroid cancer and, 268–269 Auramine, occupational exposure to, 331t, 333t Australia, migrants to, 192, 193 breast cancer in, 193 cervical cancer in, 198 esophageal cancer in, 197 lung cancer in, 196 melanoma in, 197 prostate cancer in, 194 stomach cancer in, 196 thyroid cancer in, 198 Autoimmune disease, 554 chronic lymphocytic leukemia and, 858 Hodgkin lymphoma and, 889 multiple myeloma and, 923, 924t, 928 non-Hodgkin lymphoma and, 554 Autoimmune lymphoproliferative syndrome, 551 Autoimmune thyroiditis, 982 5-Azacytidine, 491t Azathioprine, 491t

Bacillus biological dosimeter, for solar radiation, 295–299, 297t–298t Bakery industry, sinonasal cancer and, 612 Balanoposthitis, penile cancer and, 1169 Balding, prostate cancer and, 1137 Barbiturates, 498–499 Barium enema, double-contrast, 1313

1358 Barrett’s esophagus, 26–27, 701 age and, 27 descriptive epidemiology of, 26–27 dysplasia of, 27 etiology of, 27 long-segment, 27 pathology of, 26, 26f, 26t progression of, 27 undiagnosed, 27 Basal cell nevus syndrome, 563t, 566, 1243, 1275 Battery industry, sinonasal cancer and, 610 BCL-2, 58 in choriocarcinoma, 1076 BCL-6, 584t BCR-ABL, 50, 843 Beer. See Alcohol Behavioral genes, 583 Benz[a]anthracene, occupational exposure to, 328t Benzene acute myeloid leukemia and, 230, 233, 849 ambient air, 357t chronic lymphocytic leukemia and, 857 chronic myeloid leukemia and, 852 multiple myeloma and, 938 non-Hodgkin lymphoma and, 908 occupational exposure to, 326t, 333t, 336t, 339t in tobacco smoke, 230, 233 Benzidine bladder cancer and, 1107–1108 occupational exposure to, 327t, 339t Benzo[a]pyrene ambient air, 357t, 358 occupational exposure to, 328t Benzo[b]fluoranthene, occupational exposure to, 330t Benzofuran, occupational exposure to, 330t Benzo[j]fluoranthene, occupational exposure to, 330t Benzo[k]fluoranthene, occupational exposure to, 330t Benzyl violet 4B, occupational exposure to, 331t Beryllium, occupational exposure to, 326t Beta-carotene, 414–415 Helicobacter pylori infection and, 711 leukoplakia and, 25 lung cancer and, 685, 1329 multiple cancers and, 1273 oral cavity cancer and, 686 pancreatic cancer and, 740, 741 prostate cancer and, 1131 skin cancer and, 1240–1241 tobacco smoke interaction with, 685 Betel chewing, 25, 221, 233, 680 Beverages. See Alcohol; Coffee; Tea Bias detection, in migrant studies, 189 screening, 1311 selection, in migrant studies, 189–190 third variable, 180 Bidi smoking, 232, 679 Bile acids, small intestine cancer and, 806 Bile ducts adenoma of, 763 cancer of. See Biliary tract cancer hamartoma of, 764t

Index Biliary tract, anatomy of, 787, 788f Biliary tract cancer, 787–795 alcohol and, 794 cholecystectomy and, 789, 793 cholecystitis and, 793 choledochal cysts and, 795 cigarette smoking and, 227, 794 classification of, 787 congenital defects and, 795 diet and, 794 familial, 795 gallstones and, 790–792, 792f heavy metals and, 794 Helicobacter pylori infection and, 793 hepatitis B virus infection and, 793 hepatitis C virus infection and, 793 histopathology of, 787 hormones and, 793–794 incidence of international, 788, 789f U.S., 147t, 148, 150, 788–789, 789f, 790f, 790t, 791f infection and, 792–793 inflammation and, 792–793 liver flukes and, 793 molecular genetics of, 787–788 mortality from, 149t, 789–790, 791f, 792f in multiple cancer syndromes, 795 obesity and, 794 occupation and, 794 pancreaticobiliary duct anomaly and, 795 partial gastrectomy and, 793 pesticides and, 794 porcelain gallbladder and, 793 precursor lesions of, 787 pregnancy and, 793–794 prevention of, 795 primary sclerosing cholangitis and, 793 radiation and, 794 susceptibility genes in, 795 time trends in, 788–789, 791f tobacco and, 794 typhoid carrier state and, 793 ulcerative colitis and, 793 water pollution and, 794 Bioelectric impedance, 423 Biologic gradient, causation and, 5t, 7 Biologic specimens, for biomarkers, 82–83, 82t Biomarkers, 70–83, 1319–1321, 1320t accuracy of, 70 altered structure and function, 72f, 75 analytic variability in, 71–72, 71f binary, 71, 72, 72t blood samples for, 82, 82t buccal cell samples for, 82–83, 82t case-control study with, 73, 80–81, 81t case-only study with, 81 case series study with, 81 categorical, 71, 72, 72t categories of, 72–80, 72f cohort study with, 73, 75, 76t, 81–82, 81t cross-sectional study with, 73, 80 data transmission errors in, 72 early biologic effect, 72f, 74–75 evaluation of, 70 exposure, 72–74, 72f, 73t half-lives for, 73t intermediate end point, 72f, 73t, 74–75 intraindividual variability in, 70–71 measurement error with, 71, 71f

misclassification errors and, 71, 72t of occupational chemical exposure, 343–344 random effects model application to, 71 reliability of, 70 sample collection for, 73, 73t, 82–83, 82t susceptibility, 72f, 73t, 75–78. See also Genetic polymorphisms; Genetic susceptibility temporal variability in, 70–71 tissue samples for, 83 tumor, 13–15, 14f, 37t, 72f, 73t, 78–80. See also specific genes urine samples for, 83 variability in, 70–72, 71f, 72t Birt-Hogg-Dube syndrome, 563t Birth cohort study, 106 of health status, 179 Birth order, Hodgkin lymphoma and, 877, 878t Birth weight acute myeloid leukemia and, 850 prostate cancer and, 1135 socioeconomic status-cancer association and, 181 testicular cancer and, 1157 Bisacodyl, 500 Bischloroethyl nitrosourea (BCNU), 491t Bis(chloromethyl)ether, occupational exposure to, 327t Bitumen, occupational exposure to, 330t BK-mole syndrome, 1213 Bladder cancer, 1101–1118 acetaminophen and, 491–492, 1112 acetylator phenotype and, 1115 N-acetyltransferase and, 228, 1115 air pollution and, 363–370, 365t–370t alcohol and, 1111 in aluminum workers, 1109, 1109t analgesics and, 1112 anatomic distribution and, 1101 aromatic amines and, 1104, 1107–1109, 1114–1115 arsenic and, 79, 383–384, 383t, 385f, 1110, 1117 artificial sweeteners and, 1111 aspirin and, 1112 in atomic bomb survivors, 1114 benzidine and, 339t cesium-137 and, 1117 chemoprevention of, 1327t, 1332 chewing tobacco and, 1107 chlornaphazine and, 1113 chromosomal abnormalities in, 1117 cigar smoking and, 1107 cigarette smoking and, 219t, 228, 1104–1107, 1106t, 1114, 1117 classification of, 11t, 1101 coffee and, 1110–1111 cyclamates and, 1111 cyclophosphamide and, 1112–1113 CYP1A2 and, 1116 diesel engine exhaust and, 364–370, 365t–370t, 1109 diet and, 1110–1112 disinfection byproducts and, 387, 388, 388f, 1110 DNA adducts and, 1114–1115 in dyestuff workers, 1107–1108 environmental tobacco smoke and, 1107 familial predisposition to, 1114 familial relative risk in, 564t

Index fluid intake and, 1110 glutathione S-transferase M1 in, 1115–1116, 1116f hair dyes and, 1113 hemoglobin adducts and, 1114–1115 histopathology of, 1101 incidence of age and, 165f, 1102, 1105f gender and, 1101, 1104t, 1105f international, 106, 107f, 108f, 129, 131f, 133, 1101, 1102f migrant studies of, 197–198 race and, 1101, 1104t, 1105f time trends in, 131f, 133, 1101, 1104f U.S., 158–159, 159t, 1101, 1104f, 1104t iodine-131 and, 1114 ionizing radiation and, 260t, 261f, 1114 in leather workers, 1108 lymphocyte assays in, 1116 in manufacturing workers, 1108 migrant studies of, 197–198 molecular biology of, 1114–1117 mortality from, 1102, 1103f, 1104, 1104f, 1105t geographic variation in, 1101, 1103f international, 106, 107f, 129, 131f, 133 U.S., 141t, 159, 159t, 1101, 1103f, 1104f NAT1 and, 1115 NAT2 and, 1115 nickel and, 394 nitrate and, 389, 390t nonsteroidal anti-inflammatory drugs and, 1112 occupation and, 334t, 335t, 1107–1110, 1109t, 1113, 1114 organic chemicals and, 391 p53 in, 1117 in painters, 1108–1109 phenacetin and, 491, 1112 phenobarbital and, 498–499, 1112 pipe smoking and, 1107 polycyclic aromatic hydrocarbons and, 1109 prevention of, 1117, 1327t, 1332 radium and, 392–393 in rubber workers, 1108 Schistosoma haematobium and, 393, 534, 1113–1114, 1117 screening for, 1117 stage of, 1102, 1104 survival for, 159, 172, 1102, 1104, 1105t susceptibility genes in, 584t, 585t, 586t, 587t, 590t trihalomethanes and, 386–388, 388f in truck drivers, 1109 tumor suppressor genes in, 1116 tumors markers in, 1116–1117 twin study of, 91t urinary mutagens in, 1114 urinary pH and, 1113 urinary stasis and, 1113 urinary tract infection and, 1113 Bleomycin, 491t BLM, 566 Blood for biomarkers, 82, 82t cryopreservation of, 82 Bloom syndrome, 551–552, 551t, 563t, 566 bone cancer in, 953–954 leukemia in, 859 melanoma in, 1213t, 1214

BMPR1A, 568 Body composition, for diet studies, 412 Body mass index, 413, 423. See also Obesity anal cancer and, 836 Barrett’s esophagus and, 27 benign breast disease and, 31 breast cancer and, 429–434, 431f, 1001 colon cancer and, 423, 424–425 colorectal adenoma and, 425 colorectal cancer and, 817 in diet studies, 412 endometrial cancer and, 428–429, 428f esophageal cancer and, 425–426, 426f, 701, 701t head and neck cancer and, 437 hormone replacement therapy and, 432 leukoplakia and, 25 lung cancer and, 436 ovarian cancer and, 429, 430f, 1017 pancreatic cancer and, 726 prostate cancer and, 434–436, 435f, 1134–1135 rectal cancer and, 425 renal cancer and, 426–427, 427f, 1090 soft tissue sarcoma and, 969 testicular cancer and, 1158 thyroid cancer and, 436 Bone cancer, 946–955 in Bloom syndrome, 953–954 bone marrow transplantation and, 951–952, 952t chemical exposures and, 952 childhood, 1258–1259 classification of, 12t DNA helicase mutations and, 953–954, 955 fluoride and, 395 genetic susceptibility to, 952–955 implants and, 952 incidence of, 153, 153t, 946, 947f, 948t ionizing radiation and, 260t, 261f, 267, 946, 948–952, 949t, 952t in Li-Fraumeni syndrome, 953 mortality from, 153, 153t, 946, 948f occupation and, 334t, 335t plutonium and, 273, 950 prevention of, 955 radium and, 270, 271, 393, 948, 949t, 950 retinoblastoma and, 267, 951, 952–953, 954–955 in Rothmund-Thomson syndrome, 953, 955 survival for, 153, 172 Thorotrast and, 950, 951f viral infection and, 952 in Werner syndrome, 953 Bone marrow transplantation, bone cancer after, 951–952 Boot industry, sinonasal cancer and, 609 Bowen disease, 1166 Bowenoid papulosis, 1166 Bracken fern, 416 BRAF melanoma and, 1198 thyroid cancer and, 987 Brain cancer, 1173–1190 age and, 165f, 1179, 1179f alcohol and, 1188 allergic conditions and, 1187 animal models of, 1181–1184 in atomic bomb survivors, 269, 1186 breast cancer and, 1187

1359 cellular telephones and, 1186 childhood, 1190, 1255–1256 air pollution and, 370, 1189 diet and, 1185 magnetic field exposure and, 309t–311t, 311 nitrate and, 390t parental alcohol use and, 1188 parental occupation and, 1188 pesticides and, 1188–1189 phenobarbital and, 498 chlorination and, 388, 1189 chromosomal abnormalities in, 1176 cigarette smoking and, 1188 classification of, 12t, 16–17, 1173–1176, 1175t computed tomography in, 1174 cytokines in, 1184 diet and, 1185 disinfection by-products and, 388 DNA microarrays of, 17 electromagnetic field exposure and, 309t–311t, 311, 314, 315t, 1186, 1188 environmental tobacco smoke and, 1188 epidermal growth factor receptor gene and, 1176 epilepsy and, 1187 familial, 563t, 1189 familial relative risk in, 564t gender and, 1179, 1179f grade of, 1174 hair dyes and, 1189 in hereditary syndromes, 1176, 1189–1190 histopathology of, 1173–1174 incidence of age and, 165f, 1179, 1179f ethnicity and, 1179, 1180t gender and, 1179, 1179f international, 1180, 1181t, 1182f migrant studies of, 1180–1181 race and, 1179, 1180t socioeconomic status and, 1179, 1180t U.S., 160, 161t, 1176–1177, 1177t, 1178f infection and, 1187 ionizing radiation and, 260t, 261f, 263, 267, 1185–1186, 1188 low-frequency electric and magnetic fields and, 309t–311t, 311, 314, 315t, 1186, 1188 magnetic resonance imaging in, 1174–1175 migrant studies of, 1180–1181 mobile phones and, 317–318, 317t molecular genetics of, 16–17, 1175–1176 mortality from, 141t, 161t, 1177–1179 mouse models of, 1184 N-nitroso compounds in, 1182–1183, 1184–1185 in neurogenetic syndromes, 1189–1190 nitrate and, 389, 390t nitrosoureas and, 1182–1183, 1184–1185 occupation and, 314, 315t, 334t, 335t, 1181, 1183t, 1188 p53 and, 1176 pathogenesis of, 1175–1176, 1190 pesticides and, 1188–1189 in petrochemical workers, 1188 phenobarbital and, 498 prevention of, 1190 radiofrequency radiation and, 318, 318t

1360 Brain cancer (Continued) skull X-rays and, 266 stroke and, 1187 survival for, 160, 173, 1177–1179 susceptibility genes in, 584t, 586t, 587t trauma and, 1186–1187 tumor suppressor genes and, 1176 twin studies of, 1189 viral infection and, 1183–1184 Brain tumors, 1173–1176. See also Brain cancer allergic conditions and, 1187 childhood, 1255–1256 classification of, 1173–1176, 1175t histopathology of, 1173–1174 incidence of, 1177, 1178f infection and, 1187 polio vaccine and, 1187 trauma and, 1186–1187 Brazil, migrants to, 192, 193 prostate cancer in, 194 stomach cancer in, 195 BRCA1, 567, 585t breast cancer and, 564, 569, 578, 1003 multiple cancers and, 1271 ovarian cancer and, 567, 1020 screening for, 565 BRCA2 breast cancer and, 564, 567, 578, 1003 multiple cancers and, 1271 ovarian cancer and, 567, 1020 screening for, 565 Breakage-fusion-bridge (BFB) cycle, 50, 51f Breast benign disease of, 23, 29–31, 995–996 biopsy of, time cost study of, 204 mammography of, 184, 1005, 1007, 1285 pregnancy effects on, 1006 radiation sensitivity of, 260 Breast cancer, 995–1008 abortion and, 998 age and, 113, 114f, 155, 165f, 996–997, 1002 age at menarche and, 183, 193, 998, 1006, 1006t age at menopause and, 998 alcohol and, 250, 1000, 1006t, 1007 androgens and, 1002–1103 antidepressants and, 493 in ataxia telangiectasia, 1003 in atomic bomb survivors, 268, 1002 birth weight and, 181, 432 body mass index and, 429–434, 431f, 1001 BRCA genes in, 113, 564, 567, 578, 1003 C-peptide in, 434 calcium channel blockers and, 497–498 chemoprevention of, 1007, 1327t, 1331–1332 cigarette smoking and, 229, 230, 1001 classification of, 15–16, 995–996 molecular, 14f, 15–16 morphologic, 11t–12t, 14f, 15 diet and, 193, 407, 407f, 414, 416, 999–1000, 1006t, 1007 ductal, 13, 14f, 15–16, 31 E-cadherin in, 13 economic burden of, 210, 211, 212, 213 electric blankets and, 313, 313t employment after, 204 endogenous estrogens and, 1002 environmental tobacco smoke and, 232–233, 1001

Index estradiol in, 433, 1004–1005, 1004t estrogen receptors in, 13, 14f, 15, 433, 996 ethnicity and, 997 familial, 563–564, 563t, 567 familial relative risk in, 564t family history of, 31, 469, 1003 fat intake and, 407, 407f, 414, 999–1000 fiber intake and, 1000 folate and, 243, 414 fruits and vegetables and, 1000 genetic factors in, 13, 14f, 15, 55–56, 113, 564, 567, 578, 996, 1003 height and, 413, 1000–1001 Her2/neu gene in, 13, 14f, 15, 55–56, 996 histopathology of, 995 hormone receptors in, 13, 14f, 15, 433, 996 hormone replacement therapy and, 470–473, 471t–472t, 484f–485f, 998–999, 1006t, 1007 hydatidiform mole and, 1082 implants and, 1001 incidence of, 1293 age and, 113, 114f, 165f, 996–997 international, 106, 107f, 108f, 109, 113–114, 113f, 114f, 997 migrant studies of, 113, 193, 997 race and, 997 socioeconomic status and, 114, 176–177, 176t, 997 time trends in, 114, 115f U.S., 154–155, 154f, 156t, 996, 997f insulin-like growth factor-I and, 433–434, 1003 invasive, 15 ionizing radiation and, 260, 260t, 261f, 262, 263, 266–267, 268, 270 lactation and, 998 leiomyosarcoma and, 969 leptin and, 434 lobular, 13, 14f, 15, 23 low-penetrance genes in, 1004–1005, 1004t magnetic field exposure and, 312, 314–315 mammographic density and, 1005 melanoma and, 1213t, 1214 menarche and, 183, 193, 998, 1002, 1006, 1006t meningioma and, 1187 menstrual cycle and, 998, 1002 migrant studies of, 193, 997 molecular genetics of, 996 mortality from, 1293 education and, 177–178, 178t international, 106, 107f, 113, 114, 115f migrant studies of, 193 U.S., 141t, 154f, 155, 156t, 996 multiple cancers and, 1273–1275, 1274t nutritional factors and, 999–1000 obesity and, 413, 429–434, 431f, 438t, 458 occupation and, 314–315, 335t oral contraceptives and, 469, 998, 1006t, 1274 organochlorines and, 1001 ovarian ablation and, 264 parity and, 998 pathogenesis of, 433–434, 1005–1006 pesticides and, 1001 physical activity and, 455–458, 456t, 457t, 462, 1001, 1002 phytoestrogens and, 1000

postmenopausal alcohol and, 250 hormone therapy and, 470–473, 471t–472t, 484f–485f, 998–999 weight and, 430, 432 precursors of, 29–31, 995–996 descriptive epidemiology of, 30 etiology of, 30–31 pathology of, 30 progression of, 29, 31 pregnancy and, 998, 1002, 1006t, 1007 premenopausal alcohol and, 250 leptin and, 434 weight and, 430, 432 prevention of, 473, 1006–1007, 1294–1295, 1327t, 1331–1332 progesterone and, 1003 progestin and, 999 prolactin and, 1003 race and, 997 radiofrequency radiation and, 318 in radiologists, 271–272 radium and, 270, 392–393 raloxifene and, 473 reproductive factors in, 998 reserpine and, 497 risk factors for, 1006–1007, 1006t salivary gland cancer and, 688 screening for, 1005, 1007, 1285 mortality and, 114 socioeconomic status-cancer association and, 184 selective estrogen receptor modulators and, 472–473 shift work and, 183 in situ, 15 socioeconomic status and, 114, 176–177, 176t, 178, 183, 997 solar radiation and, 301, 301t soy and, 1000 statins and, 494 survival for, 113, 155, 172, 997 susceptibility genes in, 584t, 585t, 586t, 587t, 588–589 testosterone and, 1003 tetrachloroethylene and, 392 thyroid cancer and, 982–983 treatment of acute myeloid leukemia and, 848, 850 adjuvant, weight gain and, 433 leukemia and, 266 second breast cancer after, 267 social support and, 183 tubular, 15 twin study of, 91t vitamins and, 1000 weight and, 430, 432, 1001, 1006t, 1007 Breast implants, 1001 Breastfeeding acute lymphocytic leukemia and, 856 endometrial cancer and, 1034 Hodgkin lymphoma and, 878 pancreatic cancer and, 750 Broadcast towers, 316–317 Bronchitis, lung cancer and, 642 Bronchus cancer. See Lung cancer Buccal cells, for DNA sample, 82–83, 82t Budding uninhibited by benomyl (Bub) protein, 54, 55f

Index Burkitt lymphoma, 133, 508f, 511, 511t, 554 Burn scar, melanoma and, 1210 Busulfan, 491t 1,3-Butadiene, 328t, 338, 359, 857 Butylated hydroxyanisole, occupational exposure to, 332t

C-peptide, in breast cancer, 434 CA125, in ovarian cancer, 1022t Cadherins, 59–60 Cadmium, 326t, 337 gallbladder cancer and, 794 lung cancer and, 337 prostate cancer and, 327, 1136 Calcium colorectal cancer and, 415, 816, 822, 1330 drinking water, 394 prostate cancer and, 415, 1134 renal cancer and, 1092 thyroid cancer and, 986 Calcium channel blockers, 497–498 Caldesmon, 54 Canada, migrants to, 192, 193 Cancer atlas, 105, 106f Cancer burden economic, 202–212. See also Economic burden of cancer global, 106–107, 107f, 108f by site, 107, 109–133. See also at specific cancers U.S., 139–142, 140t, 141t–142t, 143t by site, 142–164. See also at specific cancers Cancer control and prevention. See also Risk, communication of; Screening accomplishments in, 1285–1286 application of, 1287 assessment of, 1293–1295 causal inference and, 1299–1300 commitment to, 1295–1296 comprehensive, 1289 dissemination research for, 1288–1289 framework for, 1286–1287, 1286f history of, 1283–1284 interventions for, 1284–1285 in occupational settings, 344–345 overview of, 1284–1285 partnerships for, 1289 population focus in, 1296–1297, 1298 rationale for, 1292–1293 research for, 1287–1289, 1295 risk/benefit analysis in, 1299 surveillance research for, 1288 uncertainty and, 1297–1299 Cancer Genome Anatomy Project, 77, 96 Cancer precursors, 21–39. See also Carcinogenesis age of onset of, 22 cervical, 31–35, 32f, 33f, 34f, 1044–1046, 1045f chemoprevention of, 36–37, 37t, 39. See also Chemoprevention classification of, 38 colorectal, 27–29, 28f endometrial, 35–36 esophageal, 25–27, 26f, 26t etiology of, 21 frequency of, 22–23 gastric, 708 gene mutations in, 23

general properties of, 22–24 genomic instability of, 23 heterogeneity of, 23 independence of, 23 laryngeal, 627 locations of, 22 mammary, 29–31, 995–996 melanoma, 1197–1198, 1198t molecular studies of, 22, 38–39 multicentricity of, 23 oral cavity, 24–25, 681, 682 preexisting conditions and, 23 progression of, 23–24, 24t pulmonary, 638 screening for, 21–22, 37, 38t. See also Screening sinonasal, 603 size of, 22 submorphologic, 23 terminology for, 22 transgenic mouse models of, 39 vaginal, 1068 vulvar, 1068 Cancer-susceptibility genes, 93t, 94–97. See also specific genes Candidate genes, 93t, 94–97, 580t, 581t–582t, 583, 584t–587t, 588 Captafol, occupational exposure to, 328t Carbohydrates. See also Diet colorectal cancer and, 817 pancreatic cancer and, 739–740 Carbon black, occupational exposure to, 330t Carbon tetrachloride, occupational exposure to, 330t Carcinogenesis, 47, 48f, 61f, 1318. See also Cancer precursors cell cycle defects in, 51–55, 52f, 53f, 55f, 61f cellular growth defects in, 55–60, 56f, 57f, 59f, 60f epigenetic alterations in, 51 genetic events in, 47–51, 48f, 49f, 61f, 1318–1319 mouse models of, 1322 viral, 8. See also specific viral infections Carcinogens. See also specific agents identification of, 1341–1344, 1343t isothiocyanate inhibition of, 739 regulation of, 1341–1351, 1342t acceptable risks and, 1344–1345 in children, 1346–1347 cost-benefit analysis and, 1345–1346 dose-response modeling in, 1346–1347, 1346f international, 1350 precautionary principle in, 1344–1345 risk assessment and, 1345–1346 in susceptible populations, 1346–1347 in U.S., 1347–1350, 1347f, 1347t Carcinoid tumor, 801 Carcinoma in situ, 21 Caretaker genes, 48–50 Carney complex, 563t Carotenoids, 414–415. See also Vitamin A alcohol effects on, 243 esophageal cancer and, 700 lung cancer and, 643, 650 multiple cancers and, 1273 oral cavity cancer and, 682, 685 ovarian cancer and, 1017 pancreatic cancer and, 740–741

1361 pharyngeal cancer and, 682, 685 prostate cancer and, 1131 renal cancer and, 1092 stomach cancer and, 712, 712t, 715 thyroid cancer and, 986 Caspases, in apoptosis, 57–58, 57f Castleman disease, 523 Catechol, occupational exposure to, 332t b-Catenin, 59, 60f colorectal cancer and, 809, 810f, 811 Causation, 3–8 analogy and, 5t, 6–7 biologic gradient and, 5t, 7 biologic understanding and, 4, 5t, 6–7 coherence and, 5t, 6–7 component cause model of, 4–5, 5f concepts of, 3–4 consistency and, 5t, 6 counterfactual definition of, 3–4 criteria for, 5–7, 5t definition of, 3–4 environmental mechanisms in, 7–8 molecular mechanisms in, 7, 8 natural experiment and, 5t, 7 plausibility and, 5t, 6–7 population studies and, 8 prevention strategies and, 1299–1300 probabilistic definition of, 4, 8 probability of, 345 specificity and, 5t, 6 strength of association and, 5t, 6 temporality and, 5t, 6 Cause-of-death statement, 104 CCR2, 584t CCR5, 584t CDH1, 568–569, 587t CDK4, 569, 1215 CDKN1, 585t CDKN2A melanoma and, 569, 1198, 1214–1215, 1214t screening for, 565 Celecoxib. See COX-2 inhibitors Celiac disease, 804–805, 911 Cell cycle, 51–55, 52f checkpoints of, 51–52 CHFR-associated early G2/M checkpoint of, 54 cytokinesis of, 55 G2/M checkpoints of, 54–55 G2/M transition of, 54 G1/S transition of, 52–53, 53f metaphase to anaphase transition of, 54–55 mitotic spindle checkpoint of, 54, 55f c-MYC in, 47, 49f RB protein in, 48, 49f, 52 Cellular adhesion molecules, 59 Cellular phones, 317–318, 317t, 1186 Centers for Disease Control, 1288, 1289 Central nervous system tumors, 1173–1190. See also Brain cancer Centrosome cycle, 53–54 Ceramic fibers, occupational exposure to, 330t Cereals, pancreatic cancer and, 739 Cervical adenocarcinoma, 34–35. See also Cervical cancer cigarette smoking and, 228 human papillomavirus infection and, 1053 incidence of, 124, 126f, 128 nonviral risk factors for, 1053 in situ, 34–35, 34f, 1048

1362 Cervical cancer, 1044–1060 anatomic distribution of, 1044, 1045f antioxidants and, 1052 chemoprevention of, 1058, 1327t, 1332–1333 Chlamydia trachomatis and, 1052 chromosomal abnormalities in, 1057 cigarette smoking and, 219t, 228, 1051, 1053 classification of, 11t cytologic screening for, 1058, 1059 diet and, 1052 diethylstilbestrol and, 1044 economic burden of, 210, 213 endogenous hormones and, 1054 estrogen and, 1054 familial susceptibility to, 1053 genetic factors in, 1053 herpes simplex virus and, 1052 histopathology of, 1044 HPV DNA testing for, 1058, 1059, 1060 human immunodeficiency virus infection and, 1053–1054 human papillomavirus in, 33, 124, 525, 526–527, 1044–1046, 1045f, 1050–1051 acetowhite appearance of, 1046 biopsy and, 1046 cigarette smoking and, 228 estrogen effects on, 1054 etiologic subtypes of, 1047, 1047f, 1050–1051, 1053, 1057 HLA genes and, 1055–1056 host genome integration of, 1047, 1057–1058 immune response to, 1055 latency of, 1055 macroscopic diagnosis of, 1046, 1055 microscopic diagnosis of, 1045–1046, 1054–1055 molecular biology of, 1046–1047 natural history of, 1055 oncogenic mechanisms of, 1050–1051, 1054–1058, 1056f oral contraceptives and, 477–478, 1052 persistent infection and, 1056–1057, 1056f, 1059 progesterone effects on, 1054 in situ disease and, 34, 35 somatic genetic changes and, 1057 viral load and, 1057 immunocompromise and, 1053–1054 incidence of, 33f, 1294 age and, 33f, 124, 126f, 128, 165f, 1049, 1049t, 1050 by histologic subtype, 128 international, 106, 107f, 108f, 124, 126–128, 126f, 127f, 1049–1050, 1050t migrant studies of, 198 race and, 1047–1048, 1048f, 1048t, 1049, 1049t screening and, 128 time trends in, 124, 126–128, 127f U.S., 155, 155f, 156t, 1047–1049, 1048f, 1048t inflammation and, 1052 invasive, 1057–1058 ionizing radiation and, 260t, 261f, 264, 264f, 848 male factor in, 1051, 1053 methylation abnormalities in, 1057

Index migrant studies of, 198, 1050 molecular genetics of, 1055–1056, 1057, 1059 mortality from international, 124, 127f, 1049–1050 U.S., 141t, 155, 155f, 156t, 1048t, 1049, 1049t multiple cancers and, 1276 occupation and, 334t, 335t, 1052–1053 oral contraceptives and, 477–478, 1052, 1053 parity and, 1051–1052 pathogenesis of, 1054–1058, 1056f penile cancer and, 1051 post-transplantation, 552 postmenopausal hormone therapy and, 478 precursors of. See Cervix, cancer precursors of pregnancy and, 1051–1052 prevention of, 1052, 1058–1059, 1060, 1327t, 1332–1333 progesterone and, 1054 radiotherapy in, 264, 264f, 848 religion and, 1051 screening for, 21–22, 38t, 1044, 1045–1046, 1055, 1058–1059, 1060, 1285 second cancers after, 264, 264f, 848, 1053 in sex workers, 1052, 1054 sexual risk factors for, 1051, 1053 sexually transmitted diseases and, 1052 socioeconomic status and, 176t, 182–183, 1049, 1051 survival for, 155, 172, 1049, 1049t susceptibility genes in, 585t, 586t, 587t vaginal cancer and, 1072 vaginal douching and, 1052 vulvar cancer and, 1072 Cervical intraepithelial neoplasia, 31, 32f, 33–34, 1045–1046, 1048–1049 Cervix anatomy of, 1044, 1045f atypical glandular cells of, 34 atypical squamous cells of undetermined significance of, 1046 biopsy of, 1046 cancer precursors of, 31–34, 1044–1046, 1045f adenocarcinoma in situ and, 35 Bethesda system for, 31 descriptive epidemiology of, 31–33, 33f etiology of, 33 glandular, 34–35, 34f host-related factors and, 33 mild (CIN I), 31, 31f, 1045–1046 pathology of, 31, 32f progression of, 33–34 screening for, 31–34 squamous, 31–35, 32f surveillance for, 31–32 glandular lesions of, 34–35, 34f koilocytotic atypia of, 31, 1045 low-grade squamous intraepithelial lesion of, 1046 precancer of, 1046, 1048–1049 HPV DNA testing for, 1059, 1060 human papillomavirus persistence and, 1056–1057, 1056f human papillomavirus virus DNA testing for, 1058 progression of, 1057–1058 squamous dysplasia of, 31, 32f

squamous intraepithelial lesion of, 31, 33 transformation zone of, 1044, 1045f Cesium-137, 278 bladder cancer and, 1117 CHEK2, 569, 584t Chemicals. See also specific chemicals occupational exposure to, 322–346. See also also specific chemicals and occupations animal experimentation in, 323 attributable risk and, 339–340 biomarkers of, 343–344 biomonitoring of, 342 causation and, 345 cigarette smoking and, 343 classification of, 323–325, 324t, 325t cohort studies of, 340–341 community-based studies of, 341–342 confounders in, 342–343 data on, 345, 346 epidemiology of, 322–323 expert assessment of, 342 gene interactions with, 343–344 group 1, 325t, 326t–327t group 2A, 325t, 328t–329t, 330t–332t, 333t group 2B, 325t, 330t–332t, 333t group 3, 325t group 4, 325t historical perspective on, 322 industry-based studies of, 340–341 International Agency for Research on Cancer classification for, 323–325, 324t, 325t job exposure matrix for, 342 prevention of, 344–345 research need on, 345–346 review articles on, 333–334, 335t–336t self-reports of, 341–342 short-term tests in, 323 structure-activity relationships in, 323 study design for, 340–343 organic, 391–392 regulation of, 1349 Chemists, chronic lymphocytic leukemia in, 857 Chemokines, in lymphocytic leukemia, 858 Chemoprevention, 39, 1318–1335 biomarkers in, 1319–1231, 1320t of bladder cancer, 1327t, 1332 of breast cancer, 1007, 1327t, 1331–1332 of cervical cancer, 1058, 1327t, 1332–1333 clinical trials for, 36–37, 37t, 1323–1324, 1324t of colorectal adenoma, 37 of colorectal cancer, 821–823, 1326t, 1329–1330 combinations in, 1323 definition of, 36 development of, 1321–1322 dose de-escalation for, 1323 of endometrial cancer, 1274 enhanced mechanistic specificity of, 1322–1323 of esophageal cancer, 1325t, 1329 of head and neck cancer, 1324 of hepatocellular carcinoma, 778, 1326t–1327t, 1330–1331 in high-risk populations, 1321 of hydatidiform mole, 1083

Index of laryngeal cancer, 633 of liver cancer, 778, 1326t–1327t, 1330–1331 of lung cancer, 650, 1324, 1328–1329 of melanoma, 1216–1217 mouse models in, 1322 of multiple cancers, 1273 of oral cavity cancer, 685 of pharyngeal cancer, 685 of prostate cancer, 1140, 1327t, 1332 safety standards for, 1322 of skin cancer, 1243, 1327t–1328t, 1333 of stomach cancer, 715, 1326t, 1329 topical delivery for, 1323 trials for, 1323–1324, 1324t in vitro testing for, 1321 in vivo testing for, 1322 Chemotherapy. See Antineoplastics Chernobyl nuclear reactor accident, 259, 279–280 bladder cancer and, 1114, 1117 thyroid cancer and, 259, 979, 984–985 Chest X-ray, 650, 1314 Chewing tobacco. See Tobacco, smokeless CHFR (checkpoint with FHA and ring finger) protein, 54 Childhood cancer, 1251–1261. See also specific cancers air pollution and, 370, 371t–373t CNS, 1255–1256. See also Brain cancer, childhood germ cell, 1259 hepatic, 1260 incidence of, 163–164, 1251, 1252t leukemia. See Leukemia, childhood lymphoma, 1256–1257 methodological approaches to, 1261 morbidity from, 1251 mortality from, 164, 1251, 1252f osseous, 1258–1259 registry for, 1261 renal, 1258 retinal, 1259–1260 second cancers after, 267, 1277 soft tissue, 1258 sympathetic nervous system, 1257–1258 syndromic associations of, 1251, 1253t Chimney sweeps, 322 Chlamydia trachomatis infection, cervical cancer and, 1052 Chlorambucil, 491t Chloramphenicol, 491t leukemia and, 493, 850, 856 Chlordane, occupational exposure to, 331t Chlordecone, occupational exposure to, 331t Chlorination, 386–388, 1110 Chlornaphazine, 491t, 495, 1113 Chloro ether, occupational exposure to, 339t 4-Chloro-ortho-toluidine, occupational exposure to, 328t Chloroform, 331t, 386–388, 491t Chloromethyl ethers lung cancer and, 644 occupational exposure to, 327t Chlorophenols in drinking water, 391 non-Hodgkin lymphoma and, 908 soft tissue sarcoma and, 965t, 966 Chlorophenoxy herbicides, occupational exposure to, 331t

Chlorophyllin, in hepatocellular carcinoma prevention, 778 Chloroprene, occupational exposure to, 330t Chlorothalonil, occupational exposure to, 331t Chlorozotocin, 491t Cholangiocarcinoma, 763, 764t. See also Biliary tract cancer cirrhosis and, 774 hemochromatosis and, 775 incidence of, 119, 765–766, 766t, 789 liver flukes and, 535, 772–773 mortality from, 766 oral contraceptives and, 771 primary sclerosing cholangitis and, 776 survival for, 766 Thorotrast and, 271 vinyl chloride and, 773 Cholecystectomy biliary tract cancer and, 789, 793 colorectal cancer and, 820 pancreatic cancer and, 749 Cholecystitis biliary tract cancer and, 793 pancreatic cancer and, 749 Cholecystokinin, pancreatic cancer and, 749 Choledochal cysts, biliary tract cancer and, 795 Cholera, 3 Cholesterol bladder cancer and, 1112 pancreatic cancer and, 740 Chondrosarcoma extraskeletal, 17t, 960t genetics of, 954 incidence of, 946, 947f, 947t mortality from, 946, 948f plutonium and, 950 radiation exposure and, 952, 952t Chordoma genetics of, 954 hereditary, 563t incidence of, 946, 947t Choriocarcinoma, 1075–1083 ABO blood group and, 1081, 1083 age and, 1079 age at menarche and, 1082 Agent Orange and, 1082–1083 alcohol and, 1080 antecedent abnormal pregnancy and, 1081 BCL-2 in, 1076 chemical exposures and, 1081 cigarette smoking, 1080 cytogenetics of, 1075 diet and, 1080 DOC-2/hDab2 in, 1076 cERBB-2 in, 1076 estrogen levels and, 1082 family history of, 1081 herbicides and, 1082–1083 HLA genes and, 1081 incidence of, 1075, 1082 age and, 1079 international, 1077, 1077t, 1078 race and, 1078, 1079t socioeconomic status and, 1079 U.S., 1076–1077, 1077t infection and, 1080, 1083 ionizing radiation and, 1081, 1083 MDM-2 in, 1076 mortality from, 1077

1363 oral contraceptives and, 1079–1080, 1080t, 1082 p53 in, 1076 parity and, 1081 pathogenesis of, 1076, 1082–1083 prevention of, 1083 race and, 1078, 1079t, 1082 recurrent, 1081 second cancers and, 1082 socioeconomic status and, 1079 spontaneous abortion and, 1081–1082 survival for, 1077 testicular, 1151, 1153f, 1153t, 1162t. See also Testicular cancer Chromate, pancreatic cancer and, 746 Chromium ambient air, 357t gallbladder cancer and, 794 lung cancer and, 644 occupational exposure to, 326t pancreatic cancer and, 745 sinonasal cancer and, 610–611, 613t Chromosomes abnormalities of acute lymphocytic leukemia and, 852, 1253 acute myeloid leukemia and, 843–844, 851–852, 1253 arsenic exposure and, 79 bladder cancer and, 1117 brain cancer and, 1176 cervical cancer and, 1057 choriocarcinoma and, 1075 chronic lymphocytic leukemia and, 852 comparative genomic hybridization for, 13 esophageal cancer and, 702 G banding for, 12 gene amplification and, 50 as genotoxicity measure, 75 loss of heterozygosity and, 50–51, 51f lymphocytic leukemia and, 852 melanoma and, 1198 mesothelioma and, 659 multiple myeloma and, 919 myeloid leukemia and, 842, 843 non-Hodgkin lymphoma and, 898–899, 900t ovarian cancer and, 1020 pancreatic cancer and, 750 sarcoma and, 17, 17t soft tissue sarcoma and, 961 testicular cancer and, 1157–1158, 1157t thyroid cancer and, 987–988 double minutes of, 50 double-stranded break of, 50 fusion of, 50–51, 51f homogeneously staining regions of, 50 homologous recombination of, 50 instability of, 49–50 nonhomologous end joining of, 50 Philadelphia, 50, 843 telomere loss from, 56, 56f translocations of, 50 Chronic lymphocytic leukemia. See at Leukemia Chronic myeloid leukemia. See at Leukemia Chutta smoking, 221 CI Acid Red 114, occupational exposure to, 331t CI Basic Red 9, occupational exposure to, 331t

1364 CI Direct Blue 15, occupational exposure to, 331t Cigar smoking, 221, 222, 232 bladder cancer and, 1107 laryngeal cancer and, 629 oral cavity cancer and, 679 pancreatic cancer and, 722 sinonasal cancer and, 614 Cigarette smoking, 217–234 N-acetyltransferase phenotype and, 230 acute lymphocytic leukemia and, 856 acute myeloid leukemia and, 850 addiction to, 222, 223f, 225 adduct markers of, 225, 233 age at initiation of, 219, 221t, 231 alcohol and, 226, 232, 243, 246–247, 248, 250–251 anal cancer and, 835 antiestrogenic effect of, 229 autopsy studies of, 233 Barrett’s esophagus and, 27 benign breast disease and, 31 biliary tract cancer and, 227, 794 biomarkers of, 225 birth cohort patterns in, 219–221, 221t bladder cancer and, 228, 1104–1107, 1106t, 1117 brain cancer and, 230, 1188 breast cancer and, 229, 230, 1001 bronchial cancer and, 226 cervical cancer and, 228, 1051, 1053 cessation of, 230–231, 230f, 234 chemical exposure with, 222, 222t, 233 choriocarcinoma and, 1080 chronic lymphocytic leukemia and, 857–858 chronic myeloid leukemia and, 852 colorectal adenoma and, 29, 227 colorectal cancer and, 227, 818 cotinine marker of, 224f, 225 education and, 174, 175, 180 endometrial cancer and, 228–229, 1031, 1035, 1035t, 1037 epidemic stages of, 218, 220f esophageal cancer and, 227, 699, 699t exposure measurements for, 224–225 Fagerstrom index in, 225 gallbladder cancer and, 227 genetic influences on, 231 global measurement of, 218–219, 220f hepatitis B virus and, 232 hepatocellular carcinoma and, 227, 516, 771 historical perspective on, 217, 218f Hodgkin lymphoma and, 230, 889–890 human papillomavirus infection and, 232, 527 income and, 174–175 international patterns in, 218, 220f laryngeal cancer and, 226, 246–247, 628–629, 630, 633 leukemia and, 230 leukoplakia and, 25 lip cancer and, 226, 686 liver cancer and, 227, 516, 771 lung cancer and, 226, 641–642 air pollution and, 360–361 in migrant studies, 196 mesothelioma and, 668 mortality burden of, 217–218, 219t multiple cancers and, 1272

Index multiple myeloma and, 939 nasal cavity cancer and, 225–226 nasopharyngeal cancer and, 226–227, 623–624 neuroblastoma and, 1257–1258 neurotransmitters and, 230 nicotine addiction in, 222, 223f nicotine yield from, 223, 223f, 224f nitrosamine marker of, 225 non-Hodgkin lymphoma and, 230, 909 occupational chemical exposures and, 343 oral cavity cancer and, 226, 679–680, 683 ovarian cancer and, 229, 1019 pancreatic cancer and, 228, 725, 752 paranasal sinus cancer and, 225–226 penile cancer and, 229, 1168 pharyngeal cancer and, 226, 679, 683 prevalence of, 218–221, 220f, 221f, 221t prevention of, 233–234, 1285 prostate cancer and, 229, 1135 rectal cancer and, 227 renal cancer and, 228, 1089–1090 renal pelvis cancer and, 228, 1094–1095 13-cis-retinoic acid interaction with, 685 salivary gland cancer and, 688 self-report on, 224–225 sinonasal cancer and, 613, 615 skin cancer and, 229–230, 1239–1240, 1240t small intestine cancer and, 805 socioeconomic status and, 180, 181 stomach cancer and, 227, 712–713 Surgeon General’s 1964 report on, 7 tar yield of, 223, 223f, 232 testicular cancer and, 229, 1159 thyroid cancer and, 230, 986–987 tobacco processing effects in, 224 tracheal cancer and, 226 ureteral cancer and, 228, 1094–1095 vaginal cancer and, 229, 1071 vulvar cancer and, 229, 1071 Cimetidine, 499 Circumcision, penile cancer and, 1168–1169, 1170 Cirrhosis, 119, 773–774, 776 Cisplatin, 491t Citrus red No. 2, occupational exposure to, 331t Classification (cancer), 10–18. See also at specific cancers historical perspective on, 10 molecular, 12–15 comparative genomic hybridization for, 13 cytogenetic techniques for, 12 DNA microarrays for, 13 fluorescence in situ hybridization for, 12–13 gene expression profiles for, 13, 15 multimarker approaches for, 15 serial analysis of gene expression for, 13 single marker approaches for, 13–15, 14f morphologic, 10–12, 11t–12t, 78 Clofenotane, 491t Clofibrate, 494 Clonorchis sinensis, 535 Cluster study, 879 CMM1, 569 CMM2, 569 Coal-petroleum products, occupational exposure to, 336t

Coal tar medicinal, 491t, 496 occupational exposure to, 326t, 333t Cobalt, occupational exposure to, 330t Cockayne syndrome, 573 Coffee bladder cancer and, 1110–1111 ovarian cancer and, 1015–1016 pancreatic cancer and, 741 renal cancer and, 1092 renal pelvis cancer and, 1095 thyroid cancer and, 987 ureter cancer and, 1095 Coherence, causation and, 5t, 6–7 Coke-oven workers, renal cancer in, 1093 Colon, adenoma of, 27–29, 28f, 425, 809, 823 Colon cancer. See also Colorectal cancer alcohol and, 248–249 aspirin and, 492–493 body mass index and, 423, 424–425 calcium and, 415 chlorination byproducts and, 387–388 cigarette smoking and, 227 diet and, 408, 414 disinfection by-products and, 387–388 economic burden of, studies on, 211, 212, 213 familial, 563, 567–568 fat and, 414 fiber and, 414 folic acid and, 414, 416 height and, 413, 820 hereditary, 567–568 hyperinsulinemia and, 820 incidence of age and, 165f international, 106, 107f migrant studies of, 194 socioeconomic status and, 176t, 177 U.S., 146t, 147–148 ionizing radiation and, 260t, 261f migrant studies of, 194–195 mortality from, 148, 149t, 195 nitrate and, 389, 390t nonsteroidal anti-inflammatory drugs and, 492–493 obesity and, 423–425, 424f, 438t occupation and, 744 oral contraceptives and, 480, 482, 482t physical activity and, 451–455, 452t–455t, 462, 817 postmenopausal hormone therapy and, 482–483, 483t sedentary occupation and, 744 shift work and, 183 solar radiation and, 301, 301t susceptibility genes in, 585t twin study of, 91t in vegetarians, 406 vitamin A and, 414 Colonoscopy, 823, 1313–1314 virtual, 1314 Colorectal cancer, 809–824. See also Colon Cancer acromegaly and, 820–821 alcohol and, 816, 816f, 818, 819, 824 ampulla of Vater cancer and, 795 anatomic distribution of, 809, 811 antioxidants and, 823 bile acid modifiers in, 822 biliary tract cancer and, 795

Index birth weight and, 181 body mass index and, 817 calcium and, 816, 822 carbohydrates and, 817 chemoprevention of, 821–823, 1326t, 1329–1330 cholecystectomy and, 820 cigarette smoking and, 227, 818 classification of, 16, 809 colonoscopy for, 1313–1314 Crohn disease and, 820 diabetes mellitus and, 820 diet and, 195, 814–817, 819, 821 digital rectal examination in, 1314 double-contrast barium enema for, 1313 economic burden of, 207–208, 208t, 211 energy intake and, 817 familial adenomatous polyposis and, 818 familial relative risk in, 564t family history of, 819 fat and, 817 fecal occult blood test for, 1312–1313 fiber and, 814–816, 822 flexible sigmoidoscopy for, 1313 folate and, 816, 816f, 824 folic acid and, 814, 816, 823 fruits and vegetables and, 814–816, 823 gallstones and, 820 genetic susceptibility to, 818–819 glucose intolerance and, 820 growth factors in, 819–820 height and, 820 hereditary, 563t, 567–568, 818–819 histopathology of, 809 hyperinsulinemia and, 820 incidence of, 809, 1294 age and, 165f, 812–813, 813f ethnicity and, 813, 813t gender and, 811, 813 international, 106, 107f, 108f, 114, 116, 116f, 117f, 814, 815f migrant studies of, 814 race and, 813, 813t socioeconomic status and, 813–814 time trends in, 114, 116, 117f, 811–812, 812f U.S., 146t, 147–148, 150f, 811–812, 812f inflammatory bowel disease and, 820 insulin-like growth factor-1 and, 819–820, 821 insulin-like growth factor binding protein-3 and, 819–820 laxatives and, 500 low-penetrance genes and, 819 micronutrients and, 823 microsatellite instability in, 16 migrant studies of, 114, 194–195, 814 molecular genetics of, 809–811, 810f, 821, 821f mortality from, 813–814, 814t education and, 178 international, 106, 107f, 116 migrant studies of, 814 socioeconomic status and, 184 U.S., 141t, 150f, 812, 812f nonsteroidal anti-inflammatory drugs and, 822–823 nutrition and, 814–817 obesity and, 423–425, 424f, 438t, 817 occupation and, 335t, 744

oral contraceptives and, 480, 482, 482t pathogenesis of, 809–811, 810f, 821, 821f physical activity and, 451–455, 452t–455t, 462, 817 postmenopausal hormone therapy and, 482–483, 483t, 484f–485f, 484t, 823 precursors of, 27–29, 28f prevention of, 821–823, 1295, 1326t, 1329–1330 protein intake and, 817 Schistosoma japonicum and, 534 screening for, 21–22, 38t, 823–824, 1312–1314 socioeconomic status and, 176t, 177, 184, 813–814 subsite-specific etiology in, 811 survival for, 148, 172, 812 susceptibility genes in, 584t, 585t, 586t, 587t, 589, 590t ulcerative colitis and, 811 Common variable immunodeficiency, 551, 551t, 910 Comparative genomic hybridization, 13, 93t, 96 in bladder cancer, 79 Computed tomography in brain cancer, 1174 in lung cancer, 650, 1314–1315 Computed tomography colonography, 1314 COMT, 584t in breast cancer, 1004f, 1004t, 1005 Concordance rate, in twin study, 89 Condom, 1058 Confounding, 4, 95 Congenital X-linked immunodeficiency, 859 Consistency, causation and, 5t, 6 Coronary artery disease, 1298 Corticosteroids, sinonasal cancer and, 614 Cosmic rays, 259, 276 Cost-benefit analysis, in carcinogen regulation, 1345–1346 Costs. See also Economic burden of cancer direct, 202–203, 203t, 204t, 205t indirect, 203, 203t, 204t, 205t morbidity, 203, 204t mortality, 203–204, 204t productivity, 204 psychosocial, 204 time, 204 transfer payment vs., 205 Coumarins, naturally occurring, 739 Cowden syndrome, 563t, 568, 1003, 1189–1190 COX2, 585t COX-2 inhibitors, 1322–1323 in colorectal cancer prevention, 492, 823, 1330 in esophageal cancer prevention, 701 in melanoma prevention, 1217 in pancreatic cancer prevention, 742 Creosotes, occupational exposure to, 328t para-Cresidine, occupational exposure to, 331t Crohn disease anal cancer and, 836 colorectal cancer and, 820 small intestine cancer and, 804, 806 Crude rate, definition of, 103 Cryopreservation, 82 Cryptorchidism, testicular cancer and, 1159–1160, 1161t

1365 b-Crytoxanthin, bladder cancer and, 1112 CSB/ERCC6, lung cancer and, 648 Cumulative rate, 103 Cumulative risk, 103 Cure, definition of, 102 Cyclamates, bladder cancer and, 1111 Cyclin B, 54 Cyclin B-CDC2 complex, 54 Cyclin D2, 51 Cyclin-dependent kinases, 51, 52–53, 53f activation of, 51 in apoptosis, 58 regulation of, 51 Cyclooxygenase inhibition. See COX-2 inhibitors Cyclophosphamide, 491t bladder cancer and, 1112–1113 Cyclosporine, 491t, 500 CYP1A1, 578, 583, 586t, 590t biliary tract cancer and, 795 breast cancer and, 1004f, 1004t, 1005 colorectal cancer and, 819 esophageal cancer and, 702 laryngeal cancer and, 632 lung cancer and, 645 oral cavity cancer and, 684 prostate cancer and, 1137 CYP1A1*2B, acute myeloid leukemia and, 851 CYP1A2, 586t, 589 bladder cancer and, 1116 lung cancer and, 645 CYP1B1, 584t, 586t prostate cancer and, 1137 CYP2A6, 586t lung cancer and, 645 CYP2C9, 586t lung cancer and, 646 CYP2C19, acute myeloid leukemia and, 851 CYP2D6, 578, 586t, 590t acute myeloid leukemia and, 851 lung cancer and, 645–646 CYP2E1, 586t alcohol metabolism and, 244 esophageal cancer and, 702 hepatocellular carcinoma and, 775 laryngeal cancer and, 632 lung cancer and, 646 nasopharyngeal cancer and, 624 oral cavity cancer and, 684 CYP3A4, prostate cancer and, 1137 CYP3D4, 586t CYP3D5, 586t CYP11A, 584t CYP17, 584t breast cancer and, 1004, 1004f, 1004t endometrial cancer and, 1036t, 1037 CYP19, 584t breast cancer and, 1004–1005, 1004f, 1004t Cysts choledochal, 795 hepatic, 764t ovarian, 1029 Cytochrome c, in apoptosis, 58 Cytochrome P-450 acute myeloid leukemia and, 851 bladder cancer and, 1116 ethanol effects on, 243 lung cancer and, 645–646 prostate cancer and, 1137 Cytogenetics, 12

1366 Cytokines in brain cancer, 1184 in Hodgkin lymphoma, 889 in lymphocytic leukemia, 858 in multiple myeloma, 931 in prostate cancer, 1139 Cytokinesis, 55 Cytomegalovirus infection, testicular cancer and, 1158 Cytotoxic T lymphocyte antigen-4, in multiple myeloma, 931

Dacarbazine, 491t Dairy products colorectal cancer and, 817 pancreatic cancer and, 732t–738t, 738 prostate cancer and, 1133–1134 Danthron, 491t, 500 Dapsone, 495 Daunorubicin, 491t DCC, colorectal cancer and, 809 DDT (dichlorodiphenyltrichloroethane) occupational exposure to, 331t pancreatic cancer and, 745–746 testicular cancer and, 1155 Death certificate, 104 Death receptors, 57–58 Dendritic cells, 549, 551f Denmark, migrants to, 192 nasopharyngeal cancer in, 198 Dental hygiene, oral cavity cancer and, 682–683 Deoxyribonucleic acid (DNA) from blood, 82, 82t from buccal cells, 82–83, 82t mitochondrial, 1319 solar radiation effects on, 299 from tissue specimens, 83 Deoxyribonucleic acid (DNA) adducts, 74 air pollution and, 357–358 alcohol exposure and, 243 bladder cancer and, 1114–1115 in cigarette smoke, 225 lung cancer and, 648 32 P-postlabeling assay for, 648 pancreatic cancer and, 752 Deoxyribonucleic acid (DNA) repair, 54, 77, 565 defects in, 48–50 acute myeloid leukemia and, 851 breast cancer and, 1005 colorectal cancer and, 811 endometrial cancer and, 1037 head and neck cancer and, 633 lung cancer and, 648–649 oral cavity cancer and, 684–685 soft tissue sarcoma and, 970 in skin, 299 Depot-medroxyprogesterone acetate, cervical cancer and, 1052 Depression, socioeconomic status-cancer association and, 183 Dermatofibrosarcoma, 960t Desmoplastic small round blue cell tumor, 17t Diabetes mellitus colorectal cancer and, 820 endometrial cancer and, 429, 1034–1035 ethnicity and, 95, 95f hepatocellular carcinoma and, 776

Index pancreatic cancer and, 726, 747–748 prostate cancer and, 1138 Dialysis, renal cancer and, 1094 2,4-Diaminoanisole, occupational exposure to, 331t 4,4¢-Diaminodiphenyl ether, occupational exposure to, 331t 2,4-Diaminotoluene, occupational exposure to, 330t Diaphragm (contraceptive), 1052 Dibenz[a,e]pyrene, occupational exposure to, 330t Dibenz[a,h]acridine, occupational exposure to, 330t Dibenz[a,h]anthracene, occupational exposure to, 328t Dibenz[a,h]pyrene, occupational exposure to, 330t Dibenz[a,i]pyrene, occupational exposure to, 330t Dibenz[a,j]acridine, occupational exposure to, 330t Dibenz[a,l]pyrene, occupational exposure to, 330t 1,2-Dibromo-3-chloropropane, occupational exposure to, 331t para-Dichlorobenzene, occupational exposure to, 331t 3,3¢-Dichlorobenzidine, occupational exposure to, 331t 1,2-Dichloroethane, occupational exposure to, 331t Dichloromethane, occupational exposure to, 331t Dichlorvos, occupational exposure to, 331t Diesel engine emissions ambient air, 359 bladder cancer and, 364–370, 365t–370t, 1109 lung cancer and, 363–364, 364f multiple myeloma and, 939 occupational exposure to, 328t, 336–337, 336t, 363–370, 364f, 365t–370t urban, 358 Diesel fuel, occupational exposure to, 330t Diet, 405–408 acute lymphocytic leukemia and, 856 acute myeloid leukemia and, 850 animal experiments on, 405 anthropometric study of, 412 arsenic-related cancer and, 384, 386 benign proliferative epithelial disorders of breast and, 30–31 between-person variation in, 413 biliary tract cancer and, 794 biochemical indicators of, 412 biochemical studies of, 405 bladder cancer and, 1110–1112 body composition study of, 412 bracken fern in, 416 brain cancer and, 1185 breast cancer and, 193, 407, 407f, 414, 416, 999–1000, 1006t, 1007 calcium in, 415. See also Calcium in cancer prevention, 416–417 carotenoids in, 414–415 case-control studies of, 406–407, 407f cervical cancer and, 1052 choriocarcinoma and, 1080 cohort studies of, 407, 407f

colon cancer and, 408, 414 colorectal adenoma and, 29 colorectal cancer and, 195, 408, 814–817, 821 controlled trials of, 407–408 correlation studies of, 405–406 diet record studies of, 409–410, 411–412, 411t endometrial cancer and, 1035 energy balance and, 413. See also Obesity epidemiological studies of, 405–407 esophageal cancer and, 700 fat in. See Fat, dietary fiber in. See Fiber, dietary folic acid in, 414 food frequency questionnaire studies of, 410–411, 411t food intake studies of, 409–411 foods in, 408–409 fruits in. See Fruits and vegetables hair study of, 412 hepatocellular carcinoma and, 772 Hodgkin lymphoma and, 890 24-hour recall studies of, 409–410 hydatidiform mole and, 1080 laryngeal cancer and, 631 leukemia and, 850, 856 lung cancer and, 416, 642–643 measurement of, 408–412, 408t, 411t melanoma and, 1210, 1217 mesothelioma and, 668 metabolic studies of, 405 migrant studies of, 406 multiple myeloma and, 940 nail study of, 412 nasopharyngeal cancer and, 198, 621, 623 nitrate in, 409 non-Hodgkin lymphoma and, 909 nutrients in, 408–409 oral cavity cancer and, 681–682, 685 ovarian cancer and, 1015–1017 pancreatic cancer and, 727–742, 728t–738t pharyngeal cancer and, 681–682 prostate cancer and, 1130–1134, 1142 recommendations for, 416–417 red blood cell study of, 412 renal cancer and, 1091–1092 salivary gland cancer and, 688 salt in, 415 selenium in, 415. See also Selenium short-term recall studies of, 409–410 sinonasal cancer and, 614 skin cancer and, 1240–1241 small intestine cancer and, 805, 806 socioeconomic status-cancer association and, 181, 182 soft tissue sarcoma and, 968 stomach cancer and, 196, 712, 712t subcutaneous fat study of, 412 subgroup studies of, 406 temporal aspects of, 409 testicular cancer and, 1158 thyroid cancer and, 985–987 tissue analysis for, 412 total energy intake and, 413 vegetables in. See Fruits and vegetables vegetarian, 406 vitamin A in, 414–415. See also Vitamin A vitamin C in, 415. See also Vitamin C

Index vitamin E in, 415. See also Vitamin E in vitro studies of, 405 Dietary records, 409–410, 411t Diethyl sulfate, occupational exposure to, 329t Diethylstilbestrol cervical cancer and, 1044 testicular cancer and, 1156 vaginal cancer and, 1071 Digital rectal examination, 1314 Diglycidyl resorcinol ether, occupational exposure to, 332t Dihenylhydantoin, multiple myeloma and, 940 3,3¢-Dimethoxybenzidine (ortho-dianisidine), occupational exposure to, 331t para-Dimethylaminoazobenzene, occupational exposure to, 331t 2,6-Dimethylaniline (2,6-xylidine), occupational exposure to, 331t 3,4¢-Dimethylbenzidine (o-tolidine), occupational exposure to, 331t Dimethylcarbamoyl chloride, occupational exposure to, 328t 2,4-Dinitrotoluene, occupational exposure to, 331t 2,6-Dinitrotoluene, occupational exposure to, 331t 1,4-Dioxane, occupational exposure to, 332t Dioxins half-life of, 73t multiple myeloma and, 938 non-Hodgkin lymphoma and, 908 soft tissue sarcoma and, 962, 963t–965t, 966–967 Diphenylhydantoin, 499 Disability-adjusted life-years (DALY), 102 Discounting, in economic study, 205 Disinfection by-products, 386–388, 388f, 395, 1189 bladder cancer and, 387, 388, 1110 colon cancer and, 387–388 rectal cancer and, 388 Disperse Blue 1, occupational exposure to, 331t Distilled spirits. See Alcohol Dithiolthiones, 739 Diuretics, 497, 1090–1091 Division of Cancer Control and Population Sciences, 1284 DNA. See Deoxyribonucleic acid (DNA) DNA helicase, mutations in, bone cancer and, 953–954, 955 DNA methyltransferase, lung cancer and, 648 DNA microarrays, 13, 15, 16, 17 DNA repair assay, 77 DNMT3B, lung cancer and, 648 DOC-2/hDab2, in choriocarcinoma, 1076 Dose-response modeling, in carcinogen regulation, 1346–1347, 1346f Dose-response relationship, causation and, 5t, 7 Double-contrast barium enema, 1313 Down syndrome acute lymphocytic leukemia and, 859 acute myeloid leukemia and, 843–844, 851–852 Doxorubicin (Adriamycin), 491t DPC4, pancreatic cancer and, 751 DR4, bladder cancer and, 1116 Drinking water, 382–396 arsenic in, 382–386, 383t, 385f, 386f. See also Arsenic, drinking water

asbestos in, 394 biliary tract cancer and, 794 calcium in, 394 disinfection by-products in, 386–388, 388f, 1110, 1189 fluoridation of, 385, 394–395 hardness of, 394 Helicobacter pylori in, 393 magnesium in, 394 microbiological agents in, 393 nitrate in, 388–391, 390t, 1185 organic chemicals in, 391–392 radionuclides in, 392–393 regulation of, 1350, 1351t Schistosoma haematobium in, 393 trace metals in, 394 treatment of, 395–396 Drugs. See Pharmaceuticals Duodenal cancer. See Small intestine cancer Dusts, 333t chromium, 610–611, 613t leather, 609, 613t nickel, 609–610, 613t textile, 612, 613t wood, 608–609, 613t Dyes amine, 331t, 1107–1108 azo, 331t, 1108 benzidine-based, 328t, 1107–1108 bladder cancer and, 1107–1108 hair. See Hair dyes occupational exposure to, 333t, 336t, 1107–1108 Dysplasia, 24, 24t cervical, 31, 32f definition of, 22 oral cavity, 24, 25 progression of, 23–24, 24t Dysplastic nevus syndrome, 1213

E-cadherin, 59–60, 714 ecNOS, 584t Ecologic fallacy, 190 Economic burden of cancer, 202–212 definition of, 202 direct costs in, 202–203, 203t, 204t, 207t hospital costs in, 202, 203t indirect costs in, 203–204, 204t measurement of, 204–207 administrative data for, 206 cost-of-illness approach to, 207 costs in, 205–206 data for, 206–207, 208 discounting in, 205 estimation in, 207–208, 208t follow-back surveys for, 206 frame of reference in, 204–205, 205t hospital discharge data for, 206 human capital approach to, 203 incidence approach to, 207–208, 208t inflation adjustment in, 205 life value estimates in, 203–204 from patient perspective, 205t, 210–211 from payer perspective, 205t, 211–213 from societal perspective, 204–205, 205t, 213, 214 study design in, 206, 207 surveys for, 206 tumor registries for, 206–207

1367 willingness-to-pay approach to, 203 out-of-pocket costs in, 202–203 psychosocial costs in, 204 EDH17B2, 584t Education cancer mortality and, 177–178, 177t, 178t cigarette smoking and, 180 low birth weight and, 180 Eflornithine, in skin cancer prevention, 1333 EGFR gene, in glioblastoma, 16 Eggs, 409 pancreatic cancer and, 732t–738t, 738 Electric appliances adult exposure to, 312, 313, 313t childhood exposure to, 312–313 magnetic fields of, 312 occupational exposure to, 313–315 in utero exposure to, 312–313, 313t Electric blankets, 312–313, 313t Electromagnetic fields. See also Extremely lowfrequency electric and magnetic fields brain cancer and, 1186, 1188 leukemia and, 308–311, 309t–311t, 855, 857 melanoma and, 1209 non-Hodgkin lymphoma and, 907 pancreatic cancer and, 746 Electromagnetic spectrum, 306, 307f. See also Extremely low-frequency electric and magnetic fields; Ionizing radiation Embryonal carcinoma, 1151, 1153f, 1153t, 1162t. See also Testicular cancer Emphysema, lung cancer and, 642 Employment, after cancer, 204 End-of-replication problem, 56, 56f Endometrial cancer, 11t, 1027–1037 abortion and, 1034 age and, 1028, 1028t, 1029f alcohol and, 1035 body mass index and, 428–429, 428f, 1029–1030 breast feeding and, 1034 chemicals and, 1037 cigarette smoking and, 228–229, 1031, 1035, 1035t, 1037 CYP17 in, 1036t, 1037 diabetes mellitus and, 429, 1034–1035 diet and, 1035 economic burden of, 213 estrogen and, 429, 1029–1031, 1032t, 1036t, 1037 estrogen-secreting ovarian tumor and, 1029 ethnicity and, 1028–1029, 1028t familial relative risk in, 564t family history of, 1037 gallbladder disease and, 1035 gene polymorphisms in, 1036t, 1037 glucose tolerance and, 1034–1035 hormone replacement therapy and, 1274 hypertension and, 1035 incidence of, 1027–1029, 1028t, 1029f age and, 165f international, 106, 107f, 108f migrant studies of, 198 U.S., 155–156, 156t infertility and, 1034 injected/implanted progestogens and, 1033 insulin-like growth factor-I and, 429 intrauterine contraceptive devices and, 1034 ionizing radiation and, 260t, 261f migrant studies of, 198

1368 Endometrial cancer (Continued) mortality from, 1027–1028 international, 106, 107f U.S., 141t, 156, 156t multiple cancers and, 1273–1275, 1274t obesity and, 427–429, 428f, 438t, 439, 439f occupation and, 335t oral contraceptives and, 473, 474t, 1031, 1033 parity and, 1034 physical activity and, 429, 459, 461t, 462, 1029–1030 polycystic ovaries and, 1029 postmenopausal hormone therapy and, 473–477, 484f–485f, 1030–1033, 1032t estrogen plus progestins in, 474–476, 476t–477t unopposed estrogens in, 473–474, 475t precursors of, 35–36 descriptive epidemiology of, 35 etiology of, 35–36 pathology of, 35 progression of, 36 pregnancy and, 1034 prevention of, 38t, 1037 progesterone and, 429, 1030 progestogens and, 1031–1033, 1032t race and, 1028–1029, 1028t radiofrequency radiation and, 318 radiotherapy for, acute myeloid leukemia and, 848 raloxifene and, 473, 1031 survival for, 156, 172 susceptibility genes in, 584t, 585t, 586t, 587t, 588–589 tamoxifen and, 473, 1031, 1274–1275 time trends in, 1027–1028 twin study of, 91 weight and, 1029–1030 Endometrial hyperplasia, 35–36, 1030, 1031–1032 Endometriosis, 1020 Endonuclease G, in apoptosis, 57f, 58 Endostatin, gene for, 584t Energy intake, 413. See also Body mass index; Obesity; Weight measurement of, 423 Engine exhaust, 336–337. See also Air pollution; Diesel engine emissions Environmental tobacco smoke, 232–233 bladder cancer and, 1107 brain cancer and, 1188 breast cancer and, 229, 232–233, 1001 composition of, 222 exposure assessment methods for, 357 lung cancer and, 374–375, 642 nasopharyngeal cancer and, 624 occupational exposure to, 327t oral cavity cancer and, 679 risk assessment for, 375 sinonasal cancer and, 613 socioeconomic status-cancer association and, 181 Surgeon General’s 1964 report on, 7 EPHX1, hepatocellular carcinoma and, 774–775 Epichlorohydrin, occupational exposure to, 328t Epidermal growth factor, in pancreatic cancer, 751

Index Epidermal growth factor receptor brain cancer and, 1176 human papillomavirus infection and, 632 Epidermodysplasia verruciformis, 1238, 1276–1277 Epigenetics, 51 Epilepsy, 499, 1187 Epistasis, 92 Epithelium, cancer precursors at, 22 1,2-Epoxybutane, occupational exposure to, 330t Epstein-Barr virus infection, 508t, 509–513 age at, 510 antibodies to, 509, 510f, 511, 511t, 881–882 biomarkers of, 510–511, 511t blood load of, 509–510 Burkitt lymphoma and, 508f, 511 cofactors in, 513 epidemiology of, 510, 510f Hodgkin lymphoma and, 508f, 511t, 512, 881–885, 881t, 883t, 1256 host response to, 509, 510, 510f nasopharyngeal cancer and, 511–512, 623 natural history of, 509–510, 510f non-Hodgkin lymphoma and, 133, 905 oncogenic mechanisms of, 512–513 prevalence of, 510, 510f prevention of, 513 salivary gland cancer and, 688 sinonasal cancer and, 614 socioeconomic status-cancer association and, 181 stomach cancer and, 713 testicular cancer and, 1158 ERBB2, 13, 14f, 15, 55–56, 996, 1076 ERCC1, 648, 649 ERCC2, 649 ERCC3, 572–573, 648, 649 ERCC4, 648, 649 ERCC5, 648 Erionite, 326t, 666–667 Erythrocytes, in diet studies, 412 Erythroleukoplakia, malignant transformation of, 25 Erythromycin, multiple myeloma and, 940 Erythroplakia, 24, 674, 679 alcohol and, 681 descriptive epidemiology of, 25 malignant transformation of, 25 tobacco and, 680 Erythroplasia of Queyrat, 1166 Esophageal cancer, 11t, 697–703 alcohol and, 247–248, 698, 699–700, 699t, 702 anatomy of, 697 asbestos and, 702 Barrett’s esophagus and, 26f, 26t, 27, 701 body mass index and, 425–426, 426f chemoprevention of, 1329 chlorination byproducts and, 388 cigarette smoking and, 219t, 227, 699, 699t diet and, 700, 703 disinfection by-products and, 388 drugs and, 701 genetic factors in, 702 Helicobacter pylori infection and, 702 hereditary, 563t, 702 human papillomavirus infection and, 701–702 incidence of, 697–698

age and, 165f, 697, 698f gender and, 697, 698f, 698t geographic variation in, 697, 698t international, 106, 107f, 108f, 128–129, 129f, 130f, 697, 698t migrant studies of, 197 time trends in, 128–129, 130f, 697–698, 698f U.S., 145, 146t, 697, 698t ionizing radiation and, 260t, 261f, 702 medical conditions and, 701, 702 migrant studies of, 129, 197 mortality from, 106, 107f, 129, 141t, 149t natural history of, 697 nutrition and, 700, 701, 701t obesity and, 425–426, 426f, 438t, 701, 701t occupation and, 334t, 335t, 702 p53 in, 702 pathology of, 697 precursors of, 25–27, 26f, 26t, 701 prevention of, 701, 702–703, 1325t, 1329 smokeless tobacco and, 699 socioeconomic status and, 702 survival for, 145, 172, 698 susceptibility genes in, 584t, 585t, 586t, 587t tea consumption and, 700 tylosis and, 702 vitamin C and, 415 Esophagitis, 25 Esophagus, cancer precursors of, 25–27, 26f, 26t. See also Barrett’s esophagus ESR1, 584t Estradiol bioavailability of, 433 breast cancer and, 1002, 1004–1005, 1004t obesity and, 433 physical activity effects on, 450 Estrogen blood, alcohol effects on, 243 endogenous biliary tract cancer and, 793–794 breast cancer and, 1002 endometrial cancer and, 1029–1030, 1036t, 1037 obesity and, 429 physical activity and, 450 sinonasal cancer and, 615 exogenous. See Oral contraceptives; Postmenopausal hormone therapy prenatal exposure to, testicular cancer and, 1156 Estrogen receptors in breast cancer, 13, 14f, 15, 996 in cervical adenocarcinoma in situ, 34 in pancreatic cancer, 750 Estrone, physical activity effects on, 450 Ethnicity, 105–106. See also Race in association studies, 95, 96 breast cancer and, 997 colorectal cancer and, 813, 813t, 814t genetic polymorphisms and, 78 lung cancer and, 639–640, 640f nasopharyngeal cancer and, 620–621, 621f, 621t non-insulin-dependent diabetes mellitus and, 95, 95f Ethyl acrylate, occupational exposure to, 330t Ethylbenzene, occupational exposure to, 330t Ethylene dibromide, occupational exposure to, 328t

Index Ethylene oxide chronic lymphocytic leukemia and, 857 occupational exposure to, 327t, 336t, 339 Ethylene thiourea, occupational exposure to, 330t Etoposide, 491t, 495 Etretinate, in oral cavity cancer, 686 Ewing sarcoma, 1258–1259 genetics of, 17t, 954 incidence of, 946, 947f, 947t mortality from, 946, 948f prevention of, 955 radiotherapy for, acute myeloid leukemia and, 848 Exercise. See Physical activity Expression array analysis, of early biologic effect biomarkers, 74 Extracellular matrix, in metastasis, 58–59, 59f Extremely low-frequency electric and magnetic fields, 306, 307–316, 307f adult cancer and, 311–312, 313, 313t brain cancer and, 309t–311t, 314, 315t breast cancer and, 312, 314–315 childhood cancer and, 308–311, 309t–311t exposure assessment for, 307–308 exposure sources for, 308–312, 309t–311t leukemia and, 308–311, 309t–311t, 314, 315t, 849 melanoma and, 1209 occupational exposure to, 313–315, 315t, 336t residential exposure to, 308–312, 309t–311t in utero exposure to, 312–313, 313t wire code measurement of, 308 Eyes cancer of. See Melanoma (ocular); Retinoblastoma color of, melanoma and, 1211, 1211t, 1220

Fagerstrom index, 225 Familial adenomatous polyposis, 22–23, 563t, 570–571 age and, 22 colorectal cancer and, 818 nonsteroidal anti-inflammatory drugs in, 492 small intestine cancer and, 802 thyroid cancer and, 983 Familial aggregation, 89–91, 90t, 578, 578t age at diagnosis and, 90–91, 90t, 93 familiality measure in, 89, 90, 90t standardized incidence ratio in, 89–90, 90t twin studies of, 91–92, 91t Familial atypical multiple mole melanoma syndrome, 747 Familial juvenile polyposis, 563t Familiality (familial risk ratio, FRR), 89, 90, 90t, 92, 93 Family history, 562–564, 564t population-based data on, 564–565, 564t FAMMM (familial atypical mole-malignant melanoma) syndrome, 1213 Fanconi anemia, 563t, 568 Farming acute lymphocytic leukemia and, 856 acute myeloid leukemia and, 849 chemical exposures with, 336t chronic lymphocytic leukemia and, 857 chronic myeloid leukemia and, 852 multiple myeloma and, 935, 937t

FAS, 57f, 58, 584t Fat dietary, 414 benign proliferative epithelial disorders of breast and, 30 bladder cancer and, 1112 breast cancer and, 407, 407f, 414, 999–1000 colorectal adenoma and, 29 colorectal cancer and, 195, 414, 817 hydatidiform mole and, 1080 intake of, 411 laryngeal cancer and, 631 lung cancer and, 642 ovarian cancer and, 1017 pancreatic cancer and, 740 prostate cancer and, 414, 1133 reduction of, 417 skin cancer and, 1240 testicular cancer and, 1158 subcutaneous in diet studies, 412 in endometrial cancer, 429 Fatality, definition of, 101 Fecal occult blood test, 823, 1312–1313 Feline leukemia virus infection, 508–509, 508f Felty’s syndrome, 554, 910–911 Fenretinide, in skin cancer prevention, 1243 Fertility drugs, ovarian cancer and, 1019–1020 Fetal sulphoglycoprotein antigen, 716 FH1T, in sinonasal cancer, 603 Fiber, dietary benign proliferative epithelial disorders of breast and, 30–31 breast cancer and, 1000 colorectal adenoma and, 29 colorectal cancer and, 414, 814–816, 822, 1330 laryngeal cancer and, 631 ovarian cancer and, 1017 pancreatic cancer and, 740 stomach cancer and, 712, 712t Fibric acid derivatives, 494 Fibrosarcoma, 959, 960t. See also Soft tissue sarcoma incidence of, 946, 947t radiation exposure and, 952, 952t Fibrous dysplasia, 23 Fibrous histiocytoma, 959, 960t, 970. See also Soft tissue sarcoma incidence of, 946, 947t Field cancerization, 1273 Finasteride, prostate cancer and, 1140 Fingernails, for diet studies, 412 Firefighters, chemical exposures of, 336t Fish prostate cancer and, 1133 salted nasopharyngeal cancer and, 415, 621, 623 sinonasal cancer and, 614 thyroid cancer and, 986 Fishing industry, sinonasal cancer and, 612 Fissure, anal, 836 Fistulae, anal, 836 Flavonoids, thyroid cancer and, 986 Flexible sigmoidoscopy, 1313 Fluorescence in situ hybridization (FISH), 12–13, 79 Fluoridation, 385, 394–385

1369 Folate alcohol effects on, 243 cervical cancer and, 1052 colorectal adenoma and, 29 colorectal cancer and, 816, 816f, 824 lung cancer and, 643 Folic acid colorectal cancer and, 416, 823 deficiency of, 414 lymphocytic leukemia and, 858 Food(s), 408–409, 408t. See also Diet; Fat, dietary; Fiber, dietary; Fruits and vegetables; Meat; Protein, dietary frequency questionnaire for, 410–412, 411t 24-hour recall of, 409–410 intake of, 409–412, 411t nutritional labeling of, 417 portion sizes for, 410 preparation of, 416 preserved nasopharyngeal cancer and, 415, 621, 623 sinonasal cancer and, 614 Food additives, regulation of, 1348–1349 Food diary, 409–410, 411t Food frequency questionnaire, 410–412, 411t Food groups, 409 Food processing industry, sinonasal cancer and, 612 Foreign bodies, bone cancer and, 952 Formaldehyde ambient air, 359 host factors in, 624 laryngeal cancer and, 630 multiple myeloma and, 939 nasopharyngeal cancer and, 624 occupational exposure to, 329t, 336t pancreatic cancer and, 746 sinonasal cancer and, 611–612, 613t Founder effect, 565 Frame of reference, in economic study, 204–205, 205t, 210–214 France, migrants to, 193 bladder cancer in, 198 Freckles melanoma and, 1211–1212, 1211t skin cancer and, 1241–1242 Fruits and vegetables, 409, 414–416, 417 anticarcinogenic compounds in, 739 biliary tract cancer and, 794 bladder cancer and, 1111 breast cancer and, 416, 1000 colorectal cancer and, 416, 814–816 consumption of, 1285 esophageal cancer and, 700 hepatocellular carcinoma and, 772 laryngeal cancer and, 631 lung cancer and, 642–643 mesothelioma and, 668 multiple cancers and, 1273 nasopharyngeal cancer and, 623 oral cavity cancer and, 681–682, 685, 686 pancreatic cancer and, 728t–731t, 738–739 pharyngeal cancer and, 681–682 prostate cancer and, 416, 1130–1131 renal cancer and, 1092 testicular cancer and, 1158 thyroid cancer and, 986 Fumarate hydratase (fumarase), mutations in, 58 Fumarylacetoacetase deficiency, hepatocellular carcinoma and, 776

1370 Fungicides, pancreatic cancer and, 746 Furans, half-life of, 73t Furathiazole, 491t Furniture industry, sinonasal cancer and, 608–609 Furosemide, renal cancer and, 1091

G banding, 12 Galactose, ovarian cancer and, 1016–1017 Gallbladder disease of biliary tract cancer and, 790–792, 792f colorectal cancer and, 820 endometrial cancer and, 1035 pancreatic cancer and, 749 porcelain, 793 Gallbladder cancer. See also Biliary tract cancer calcification and, 793 cigarette smoking and, 227 diet and, 794 gallstones and, 790–792 heavy metals and, 794 incidence of, 147t, 788–789, 789f, 790t molecular genetics of, 787–788 mortality from, 149t, 789–790, 792f obesity and, 794 organochlorines and, 794 prevention of, 795 typhoid carrier state and, 793 Gamma rays, 274 Gardner syndrome, 570–571, 795 Gasoline, renal cancer and, 1093 Gasoline engine emissions multiple myeloma and, 939 occupational exposure to, 330t, 336–337 Gastrectomy biliary tract cancer and, 793 pancreatic cancer and, 749 stomach cancer after, 714 Gastric cancer. See Stomach cancer Gastric lymphoma, Helicobacter pylori and, 532 Gastric surgery, pancreatic cancer and, 749 Gastritis, 531, 554 Gastroesophageal reflux Barrett’s esophagus and, 26, 27, 701 laryngeal cancer and, 632 stomach cancer and, 715 Gastrointestinal stromal tumor, 563t, 959, 970. See also Soft tissue sarcoma Gastroscopy, 715–716 GATA1, 843, 852 Gemfibrozil, 494 Gender adult T-cell leukemia/lymphoma and, 530 bladder cancer and, 1101, 1105f brain cancer and, 1179, 1179f colorectal cancer and, 813, 813f, 814t esophageal cancer and, 697, 698f, 698t Hodgkin lymphoma and, 874, 875, 876f liver cancer and, 766 lung cancer and, 639, 639f lymphocytic leukemia and, 847f, 854 myeloid leukemia and, 844, 848t non-Hodgkin lymphoma and, 901, 902t, 903f oral cavity cancer and, 675, 676t pancreatic cancer and, 724

Index pharyngeal cancer and, 675, 676t soft tissue sarcoma and, 969 thyroid cancer and, 979, 979f, 980f, 980t Gene(s), 8, 47–51, 48t. See also specific genes behavioral, 583 candidate, 93t, 94–97, 580t, 581t–582t, 583, 584t–587t, 588 low-penetrance, 577–579, 578t. See also Genetic susceptibility mutation in, 48–50, 565–566 population study of, 579–583, 580t, 581t–582t susceptibility. See Genetic susceptibility Gene amplification, 50, 51f Gene characterization study, 97 Gene expression profiles, 13, 15, 79, 93t, 96 Genetic anticipation, in Hodgkin lymphoma, 888–889 Genetic heterogeneity, 92, 565 Genetic polymorphisms, 75–78, 578. See also Single nucleotide polymorphisms ethnicity and, 78 false-positive findings and, 78 hierarchical models of, 78 intron/exon location of, 78 sample size and, 77–78, 96 study replication of, 78 subgroup analysis and, 77 Genetic susceptibility, 89–91, 90t, 577–591 association studies of, 93t, 95–96, 95f candidate genes in, 96–97, 580t, 581t–582t, 583, 584t–587t, 588 dominant genes in, 92, 97 estimates of, 93–94 ethnicity and, 105–106 familial aggregation study in, 89–91, 90t gene-based population studies of, 579–583, 580t, 581t–582t, 583t gene characterization studies of, 96–97 gene-environment interaction in, 580t, 589–590 gene expression studies of, 96 gene-gene effects in, 580t, 587t–588t, 590 gene identification in, 93t, 94–97 linkage studies of, 93t, 94–96 low-penetrance genes in, 577–579, 578t meta-analysis studies of, 589, 590t models of, 92–93 multifactorial threshold model of, 92 recessive genes in, 92 segregation analysis of, 97 twin studies of, 91–92, 91t web sites on, 583t Genome, 565 Genomic control of confounding, 95 Genomic instability, 48–50, 49f CHFR-associated early G2/M checkpoint in, 54 telomere shortening and, 56 Genotype, 565 Geography cancer risk and, 105–106, 106f lung cancer and, 640 Germ cell tumors, 1151–1152, 1153f, 1259. See also Testicular cancer Gestational trophoblastic disease, 1075–1076. See also Choriocarcinoma; Hydatidiform mole

Gingival multiple hamartoma syndrome (Cowden syndrome), 563t, 568 Glass fibers, occupational exposure to, 330t Gleason grading, of prostate carcinoma, 16 Glioblastoma, 16–17, 1176. See also Brain cancer Glioma. See Brain cancer GLOBOCAN estimates, 104 Glottic cancer. See Laryngeal cancer Glucose intolerance colorectal cancer and, 820 endometrial cancer and, 1034–1035 Glutathione S-transferase acute myeloid leukemia and, 850–851 bladder cancer and, 1115–1116, 1116f breast cancer and, 1004f, 1004t, 1005 colorectal cancer and, 819 laryngeal cancer and, 632 liver cancer and, 774 lung cancer and, 646–647 pancreatic cancer and, 739 prostate cancer and, 1129, 1139 renal cancer and, 1093 skin cancer and, 1242 solar radiation effect on, 300 Glycemic index, pancreatic cancer and, 739–740 Glycyrrhizin, in hepatocellular carcinoma prevention, 778 Goiter, 198, 981–982 Gonadal dysgenesis, testicular cancer and, 1157–1158, 1157t, 1159–1160, 1161t Gonorrhea, anal cancer and, 834–835 Gorlin syndrome, 563t, 566, 1189, 1243, 1275 GPX, lung cancer and, 647 Grains, thyroid cancer and, 986 Green tea, stomach cancer and, 712, 712t Griseofulvin, 491t Growth factors, 55–56. See also specific growth factors GST, 586t GSTM1, 578, 583, 590t acute myeloid leukemia and, 850–851 bladder cancer and, 1115–1116, 1116f breast cancer and, 1004f, 1004t, 1005 esophageal cancer and, 702 laryngeal cancer and, 632 liver cancer and, 774 lung cancer and, 647 melanoma and, 1215 oral cavity cancer and, 684 skin cancer and, 1242 GSTM3, laryngeal cancer and, 632 GSTP, lung cancer and, 647 GSTP1 acute myeloid leukemia and, 851 bladder cancer and, 1116 liver cancer and, 774 prostate cancer and, 1129, 1139 skin cancer and, 1242 GSTT, oral cavity cancer and, 684 GSTT1 acute myeloid leukemia and, 850–851 bladder cancer and, 1116 liver cancer and, 774 lung cancer and, 647 skin cancer and, 1242 Gum cancer, 144t, 145t. See also Oral cavity cancer

Index Hair, for diet studies, 412 Hair color, melanoma and, 1211, 1211t Hair dyes acute myeloid leukemia and, 850 bladder cancer and, 1113 brain cancer and, 1189 multiple myeloma and, 940 non-Hodgkin lymphoma and, 909 salivary gland cancer and, 688 Haloacetic acids, cancer and, 386–388 Hamartoma bile duct, 764t mesenchymal, 764t Hanford nuclear facility, 272, 280 HapMap project, 579 Hashimoto thyroiditis, 554, 982 Hawaii, Japanese migrants to breast cancer in, 191, 191f, 193 colon cancer in, 192, 192f, 194 lung cancer in, 196 prostate cancer in, 194 rectal cancer in, 195 stomach cancer in, 191, 191f, 195 uterine cancer in, 198 Hay fever, lung cancer and, 642 Hayflick limit, 56 Hazardous waste sites, drinking water contamination from, 391–392 HC Blue No. 1, occupational exposure to, 331t Head, irradiation of, 263–264 Head and neck cancer. See also Laryngeal cancer; Oral cavity cancer; Pharyngeal cancer body mass index and, 437 candidate genes in, 590t cell cycle control genes in, 685 chemoprevention of, 1324 DNA repair in, 633, 684–685 marijuana smoking and, 680 obesity and, 437 susceptibility genes in, 584t, 586t Health literacy, 180 Heavy metals, biliary tract cancer and, 794 Height breast cancer and, 413, 1000–1001 colorectal cancer and, 413, 820 in diet studies, 412 measurement of, 422–423 ovarian cancer and, 1017 prostate cancer and, 1138 Helicobacter pylori infection, 508t, 531–534, 710–712, 711 age at, 533 Barrett’s esophagus and, 27 biliary tract cancer and, 793 biomarkers of, 531–532 cigarette smoking and, 232 cofactors of, 533 in drinking water, 393 epidemiology of, 531 esophageal cancer and, 702 gastroesophageal reflux disease and, 27 MALT lymphoma and, 533, 906–907 natural history of, 531 oncogenic mechanisms of, 532–533 prevention of, 533–534 rates of, 119

socioeconomic status-cancer association and, 181 stomach cancer and, 196, 532–533, 710–712, 714, 715 Hemangioendothelioma, 271, 764t Hemangioma hepatic, 763, 764t radiotherapy for, 262, 263 Hemangiosarcoma, 960t Hemochromatosis, 497, 775 Hemoglobin adducts, bladder cancer and, 1114–1115 Hemorrhoids, 836 Henle-Koch postulates, 5, 5t, 8 Hepatitis, alcoholic, 249 Hepatitis B virus infection, 508t, 513–516 age at, 513 alcohol and, 249 antibodies to, 514, 515t biliary tract cancer and, 793 biomarkers of, 514, 515t carriers of, 514 chronic, 513–514, 515t cigarette smoking and, 232 cofactors in, 515–516 epidemiology of, 514 hepatocellular carcinoma and, 119, 514–515, 763, 767–768, 767f, 772, 776–777 host response to, 513 human immunodeficiency virus infection and, 516 natural history of, 513–514 oncogenic mechanisms of, 515 prevention of, 516, 777 socioeconomic status-cancer association and, 180 transmission of, 514 vaccine against, 1331 Hepatitis C virus infection, 508t, 516–520 alanine aminotransferase in, 517 antibodies to, 518 biliary tract cancer and, 793 biomarkers of, 518 cofactors in, 519–520 epidemiology of, 517–518 extrahepatic manifestations of, 517 genotypes in, 518 hepatitis B virus infection and, 519 hepatocellular carcinoma and, 518–519, 768–769, 769t, 770f, 777 host response to, 517, 519–520 natural history of, 517, 519f oncogenic mechanisms of, 519 prevention of, 520, 777 seroprevalence of, 518 socioeconomic status-cancer association and, 180 transmission of, 517–518 vaccine against, 1331 Hepatoblastoma, 763, 764t, 1260 incidence of, 119, 765–766, 766t Hepatocellular adenoma, 763, 764t Hepatocellular carcinoma, 763, 764t aflatoxin and, 249, 516, 769–770, 777 alcohol and, 249, 516, 770–771, 777 alpha-fetoprotein screening in, 777 anabolic steroids and, 772 a1-antitrypsin deficiency and, 775–776 arsenic and, 383, 383t, 385f, 773

1371 chemoprevention of, 778, 1326t–1327t, 1330–1331 childhood, 1260 cigarette smoking and, 227, 516, 771 cirrhosis and, 773–774 diabetes mellitus and, 776 diet and, 772 fumarylacetoacetase deficiency and, 776 genetic susceptibility to, 774–775 hemochromatosis and, 775 hepatitis B virus infection and, 514–515, 763, 767–768, 767f, 772, 776–777 hepatitis C virus infection and, 517, 518–520, 519f, 768–769, 769t, 770f, 777 hereditary tyrosinemia type I and, 776 hormones and, 771–772, 775 human immunodeficiency virus infection and, 774 immune function and, 774 incidence of age and, 766 ethnicity and, 766, 767t international, 106, 107f, 108f, 119, 121f, 764–765, 765f migrant studies of, 119, 197 sex ratio of, 766 time trends in, 119 U.S., 121f, 146t, 147t, 148, 150, 765–766, 766t iron stores and, 771 liver transplantation in, 777 molecular genetics of, 764 mortality from international, 106, 107f, 765 U.S., 141t, 149t, 766 nonalcoholic steatohepatitis and, 776 occupation and, 334t oral contraceptives and, 771 p53 in, 770 parity and, 772 pathogenesis of, 776–777 porphyria and, 775 precursor lesions of, 763–764 prevention of, 777–778, 1326t–1327t, 1330–1331 primary sclerosing cholangitis and, 776 schistosomiasis and, 534, 772 screening for, 777 survival for, 765, 766 susceptibility genes in, 584t, 586t, 587t, 588t Thorotrast and, 773 TP53 in, 79 transforming growth factor-a in, 515 treatment of, 777 vinyl chloride and, 338, 773 Heptachlor, occupational exposure to, 331t HER2/neu, in breast cancer, 13, 14f, 15, 55–56, 996 Herbal medicines, nasopharyngeal cancer and, 624 Herbicides. See also Pesticides choriocarcinoma and, 1081, 1082–1083 pancreatic cancer and, 746 soft tissue sarcoma and, 962, 963t–965t, 966–967 testicular cancer and, 1155 Herceptin, 56

1372 Hereditary neoplastic syndromes, 562–573, 563t. See also specific syndromes family history in, 562–564, 564t population-based data on, 564–565, 564t population screening for, 565 terminology for, 565–566 Hereditary nonpolyposis colorectal cancer syndrome, 563t, 567–568, 818–819 biliary cancer and, 795 DNA mismatch repair gene mutations in, 48 molecular genetics of, 811 pancreatic cancer and, 747 Hereditary papillary renal cell carcinoma, 1094 Herpes simplex virus infection anal cancer and, 834 cervical cancer and, 1152 testicular cancer and, 1158 Herpes zoster virus infection, multiple myeloma and, 929 Herpesviruses, non-Hodgkin lymphoma and, 905 Hexachlorobenzene, occupational exposure to, 332t g-Hexachloroclohexane, 491t Hexachlorocyclohexanes, occupational exposure to, 332t Hexachloroethane, occupational exposure to, 331t HFE, 585t, 775 Hill, Sir Austin Bradford, 5–6, 5t Histamine-2 receptor antagonists, 499, 713 HIV. See Human immunodeficiency virus (HIV) infection Hoarseness, 634 Hodgkin lymphoma, 872–890 allergy and, 554 appendectomy and, 889 aspirin and, 890 autoimmune disease and, 889 birth order and, 877, 878t breastfeeding and, 878 childhood, 1256 childhood social environment and, 877–878, 878t cigarette smoking and, 230, 885, 889–890 classification of, 873–874, 874t clusters of, 878–881 cytokines in, 889 diet and, 890 economic burden of, 212, 214 Epstein-Barr virus and, 508f, 511t, 512, 881–885, 881t cluster studies of, 880 demographic factors in, 882–884 molecular evidence for, 882, 883t serologic evidence for, 884–885 social environment and, 885 familial aggregation of, 887–889 familial relative risk in, 564t genetic anticipation in, 888–889 genetic factors in, 887–889 histology of, 873–874, 874t, 882–884 historical perspective on, 872–873 HLA genotype and, 888 host factors in, 887–890 human immunodeficiency virus infection and, 553 immune function and, 889 immunodeficiency and, 889

Index incidence of, 873 age and, 165f, 874, 875, 875f, 876f gender and, 874, 875, 876f international, 874–876, 875f, 876f race and, 874, 875 socioeconomic status and, 875–876, 877, 878 U.S., 160, 162t, 874, 875f, 876f infection and, 872–873, 876–877, 878–885, 881t MOPP therapy in, 495 mortality from, 141t, 160, 162t, 873, 874 nonsteroidal anti-inflammatory drugs and, 890 occupation and, 885–887, 886t parity and, 890 pathogenesis of, 890 person-to-person transmission of, 879–881 phenytoin and, 499 radiotherapy for, cancer after, 262, 266–267, 873 school studies of, 879–880 second cancers after, 873, 1277–1278, 1278t sibship size and, 877, 878t socioeconomic status and, 875–876, 877, 878 survival for, 160, 173 tonsillectomy and, 554, 889 treatment of, 873 twin study of, 91, 887 ultraviolet light exposure and, 887 wood dust exposures and, 885, 886t hOGG1/hMMH lung cancer and, 649 oral cavity cancer and, 684, 685 Homocysteine, cervical cancer and, 1052 Homosexuality, anal cancer and, 831–833 Hormone receptors anal cancer and, 836 breast cancer and, 13, 14f, 15, 433, 996 pancreatic cancer and, 750 Hormone therapy, 468–485. See also Oral contraceptives; Postmenopausal hormone therapy temporal patterns of, 468–469, 469f HPC, 585t HPV infection. See Human papillomavirus (HPV) infection HRAS1, 587t, 1116 HRAS2, 587t HSD3B1, 1137 HSD3B2, 1137 HuGeNet, 94 Human chorionic gonadotropin, in testicular cancer, 1162t Human genome, haplotype-block model of, 76–77 Human Genome Epidemiology Network, 94 Human herpesvirus 8 infection, 508t, 520–523 antibodies to, 521 biomarkers of, 521–522 epidemiology of, 521 host response to, 521 Kaposi sarcoma and, 522, 523t multicentric Castleman disease and, 523 multiple myeloma and, 523, 929 natural history of, 520–521 prevention of, 523 primary effusion lymphoma and, 522 transmission of, 520

Human immunodeficiency virus (HIV) infection, 553 anal cancer and, 832–833, 834, 835–836, 836 cervical cancer and, 1053–1054 hepatitis B virus infection and, 516 Hodgkin lymphoma and, 553 human papillomavirus infection and, 25 Kaposi sarcoma and, 522, 553, 968 leukoplakia and, 25 liver cancer and, 774 lymphoma and, 804–805 melanoma and, 1210 multiple myeloma and, 929 non-Hodgkin lymphoma and, 553, 903, 905 penile cancer and, 1168, 1170 skin cancer and, 1238 testicular cancer and, 1158 vaginal cancer and, 1072 vulvar cancer and, 1072 Human leukocyte antigen (HLA) genes cervical cancer and, 1055–1056 choriocarcinoma and, 1081 Hodgkin lymphoma and, 888 multiple myeloma and, 930 nasopharyngeal cancer and, 624 vulvar cancer and, 1072 Human papillomavirus (HPV) infection, 508t, 524–527 anal cancer and, 833–834, 834t, 835–836, 837, 838 anogenital, 552 biomarkers of, 526 cervical cancer and. See Cervical cancer, human papillomavirus in choriocarcinoma and, 1080 cigarette smoking and, 232 cofactors of, 526–527 DNA testing for, 527, 1058, 1059, 1060 epidemiology of, 525–526, 525f esophageal cancer and, 701–702 geographic distribution of, 1049–1050 hormone responsiveness of, 1052 host response to, 524–525 immune response to, 1055 laryngeal cancer and, 631–632, 634 leukoplakia and, 25 multiple cancers and, 1276–1277 natural history of, 524–525, 1055 oncogenic mechanisms of, 526 oral cavity cancer and, 682 penile cancer and, 1167–1168, 1170 persistence of, 524–525 pharyngeal cancer and, 682 prevention of, 527, 1332–1333 seropositivity in, 526 sinonasal cancer and, 603, 614 skin cancer and, 1238–1239 types of, 524 ultraviolet light interaction with, 300 vaccine against, 1058, 1060, 1170, 1277, 1332–1333 vaginal cancer and, 1070–1071 vulvar cancer and, 1070–1071 Human T-cell leukemia virus type I (HTLV-I) infection, 508t, 527–531 adult T-cell leukemia/lymphoma and, 529–531, 531f biomarkers of, 529 carriers of, 529–530, 531, 531f

Index cofactors of, 530 epidemiology of, 528f, 529, 529f genetic factors in, 530 natural history of, 528–529 non-Hodgkin lymphoma and, 903 oncogenic mechanisms of, 530 premalignant states in, 529 prevention of, 531 socioeconomic status-cancer association and, 180 Hürthle cell cancer, 975 Hutchinson-Gilford progeria, 954 Hydatidiform mole, 1075–1076. See also Choriocarcinoma ABO blood group and, 1083 age and, 1079 age at menarche and, 1082 breast cancer and, 1082 family history of, 1081 incidence of, 1076–1078, 1079t malignant potential of, 1081 oral contraceptives and, 1080, 1080t, 1082 prevention of, 1083 race and, 1078–1079, 1079t, 1082 screening for, 1083 second, 1081 socioeconomic status and, 1079 Hydrazine, occupational exposure to, 332t 7-Hydro-8-oxo-2¢-deoxyguanosine, 358 Hydrocarbons acute lymphocytic leukemia and, 855–856 half-life of, 73t occupational exposure to, 330t–331t pancreatic cancer, 745–746 Hydrochlorothiazide, renal cancer and, 1091 Hydroquinone, acute myeloid leukemia and, 850 Hyper-IgM syndrome, X-linked, 551, 551t Hyperinsulinemia colorectal cancer and, 820 pancreatic cancer and, 726 prostate cancer and, 1138 Hyperkeratosis, of oral cavity, 24 Hyperlipidemia, 494 Hypermethylation, 51. See also Methylation Hypertension calcium channel blockers in, 497–498 diuretics in, 497 endometrial cancer and, 1035 renal cancer and, 1090–1091 renal pelvis cancer and, 1095 Hypomethylation, 51. See also Methylation Hypopharyngeal cancer, 144, 144t, 145t, 172. See also Pharyngeal cancer Hysterectomy ovarian cancer and, 1018 vaginal cancer and, 1071–1072

IFNGR1, 585t Ileum, cancer of. See Small intestine cancer Imatinib mesylate (Gleevec), 50 Imipramine, 493 Immune modulation, 549 Immune surveillance, 549 Immune system abnormalities of, 550. See also Human immunodeficiency virus (HIV) infection; Immunodeficiency; Immunosuppression

age-related changes in, 550–551, 555 chronic antigenic stimulation of, 555 function of, 549 physical activity effects on, 451 stress effects on, 183 structure of, 549–550, 550f Immunity, 507–508 Immunization. See also Vaccine melanoma and, 1210 multiple myeloma and, 928t Immunochemistry, in etiologic research, 79 Immunodeficiency, 23 acquired, 552–553. See also Human immunodeficiency virus (HIV) infection; Immunosuppression hereditary, 551–552, 551t Hodgkin lymphoma and, 889 non-Hodgkin lymphoma and, 910 Immunosuppressants, 499–500, 552–553 Immunosuppression. See also Human immunodeficiency virus (HIV) infection anal cancer and, 835–836 cervical cancer and, 1053–1054 drug-induced, 552–553 low-level, 553–554 melanoma and, 1210 multiple cancers and, 1275 penile cancer and, 1168 post-operative, 553–554 post-transplantation, 552–553 skin cancer and, 1238 soft tissue sarcoma and, 969 ultraviolet light-induced, 300 Implants, bone cancer and, 952 Imprinting, 565 Incidence. See also at specific cancers age and, 106, 162–163, 165f data collection for, 101 data sources for, 103, 104t decisional issues on, 101 definition of, 101 estimation of, 104 geographic variation in, 105–106 global, 106–107, 107f, 107t migrant studies of, 106 SEER data on, 139–166, 168–172. See also at specific cancers socioeconomic status and, 175–177, 176t standardized, 103 time trends in, 106 trends in, 106 Incinerators, 359 Infection, 507–534, 508t. See also specific infections acute lymphocytic leukemia and, 856–857 biliary tract cancer and, 792–793 bladder cancer and, 1113 childhood leukemia and, 856–857, 1255 choriocarcinoma and, 1080 chronic lymphocytic leukemia and, 858 Hodgkin lymphoma and, 508f, 511t, 512, 872–873, 876–877, 878–885, 881t, 883t, 1256 multiple myeloma and, 925t–927t, 929 oncogenic mechanisms of, 508–509, 508f Infectious dermatitis, 531 Infectious mononucleosis. See Epstein-Barr virus infection

1373 Infertility endometrial cancer and, 1034 ovarian cancer and, 1019–1020 radiotherapy in, 265 thyroid cancer and, 981 Inflammation, 554–555. See also specific inflammatory conditions biliary tract cancer and, 792–793 cervical cancer and, 1052 penile cancer and, 1169 Inflammatory bowel disease anal cancer and, 836 colorectal cancer and, 820 small intestine cancer and, 804–805 Inflation, in economic study, 205 Infliximab, in Crohn disease, 806 Influenza, brain tumors and, 1187 Insulin in hormone-dependent tumors, 433–434 leptin and, 434 prostate cancer and, 435–436, 1138 Insulin-like growth factor I, 584t breast cancer and, 434, 1003 colorectal cancer and, 819–820, 821 endometrial cancer and, 429 ovarian cancer and, 1021 physical activity effects on, 450 prostate cancer and, 436, 1138 Insulin-like growth factor II hepatocellular carcinoma and, 515 prostate cancer and, 1138 Insulin-like growth factor binding protein 2, prostate cancer and, 1138 Insulin-like growth factor binding protein 3, 584t colorectal cancer and, 819–820 prostate cancer and, 1138 Insulin resistance, pancreatic cancer and, 726 Interferon-a in hepatitis B virus infection, 516 in hepatocellular carcinoma prevention, 778 Interferon-b, in hepatitis C virus infection, 520 Interleukin-1, gene for, 585t Interleukin-5, Hodgkin lymphoma and, 889 Interleukin-6 Hodgkin lymphoma and, 887, 889 multiple myeloma and, 931 prostate cancer and, 1139 Interleukin-8, gene for, 585t Interleukin-10 gene for, 584t Hodgkin lymphoma and, 889 multiple myeloma and, 931 Interleukin-12, Hodgkin lymphoma and, 889 Interleukin-13, Hodgkin lymphoma and, 889 International Agency for Research on Cancer (IARC) on chemical carcinogens, 323–325, 324t, 325t GLOBOCAN estimates of, 104 pharmaceutical evaluation by, 490, 491t International Classification of Diseases for Oncology, 12 International HapMap Project, 77 Internet resources, on genetics, 48t, 583t Intestinal cancer. See Colon cancer; Colorectal cancer; Small intestine cancer Intestinal transit time, physical activity effects on, 450–451 Intraclass correlation coefficient, in biomarker evaluation, 71, 71f, 72

1374 Intraepithelial neoplasia, 22 cervical, 31, 32f, 33–34 pancreatic, 23, 721, 750 penile, 1166 prostatic, 1129 vaginal, 1068 vulvar, 1068 Intrauterine contraceptive devices, endometrial cancer and, 1034 Iodine-131, 270 bladder cancer and, 1114 diagnostic, 270 thyroid cancer and, 270, 280, 984–985 Iodine deficiency, thyroid cancer and, 985–986 Ionizing radiation, 259–283, 261f, 267f acute lymphocytic leukemia and, 268, 269f, 854–855 acute myeloid leukemia and, 846, 848–849 in ankylosing spondylitis, 262 ataxia telangiectasia and, 53, 262 background, 259, 276–277 biliary tract cancer and, 794 bladder cancer and, 160t, 161f, 1114 bone cancer and, 261f, 267, 946, 948–952, 949t brain cancer and, 260t, 261f, 263, 1185–1186 breast cancer and, 260, 260t, 261f, 262, 263, 266–267, 268, 270, 1002 cervical cancer and, 260t, 261f, 264, 264f, 848 characterization of, 280–281 childhood leukemia and, 265–266, 267, 1253 choriocarcinoma and, 1081, 1083 chronic lymphocytic leukemia and, 857 chronic myeloid leukemia and, 268, 268f, 852 cosmic, 276 esophageal cancer and, 260t, 261f, 702 in head and neck disorders, 262–264 in hemangioma, 263 indoor, 275–276, 276f in infertility, 265 leukemia and, 260t, 261f, 262, 264, 264f, 265–266, 267f, 268, 268f, 848–849 linear energy transfer of, 281 lung cancer and, 260t, 261f, 262, 267, 268 in mastitis, 260, 262 medical, 259, 260–267, 261f, 264f, 267f, 269–271 melanoma and, 1208–1209 mesothelioma and, 667 military, 267–269, 267f, 268f, 269f. See also Atomic bomb multiple myeloma and, 261f, 931–934, 932t–933t, 933t non-Hodgkin lymphoma and, 907 nuclear reactor-related, 279–280 occupational, 271–275, 273t, 326t, 336t, 848–849 osteosarcoma and, 950–952 ovarian cancer and, 260t, 261f, 1019 pancreatic cancer and, 261f, 744 in peptic ulcer, 265 physiologic effects of, 280–281, 281f preconception, 279 prenatal, 265–266 radionuclide, 269–271. See also Radionuclides regulation of, 1347–1348, 1347f, 1347t relative biological effectiveness of, 281

Index renal cancer and, 261f, 1093 renal pelvis cancer and, 1096 in ringworm, 263 salivary gland cancer and, 260t, 261f, 263, 668 second cancers and, 264, 264f, 266–267, 1277 skin cancer and, 261f, 263, 1239 soft tissue sarcoma and, 966t, 967–968 sources of, 259–260 stomach cancer and, 260t, 261f, 713 study design for, 281–283, 282t terrestrial, 276 therapeutic, 260–267, 260f, 269–271, 1277 in thymic enlargement, 262–263 thyroid cancer and, 262–263, 268–269, 983–985, 984f in tonsillar disease, 263 in tuberculosis, 260 ureter cancer and, 261f, 1096 in uterine bleeding, 264–265 weapons testing fallout and, 277–278 Iron body, hepatocellular carcinoma and, 771 medicinal, 496–497 Iron-dextran complex, 491t, 496–497 Isoflavones, prostate cancer and, 1132 Isoniazid, 493 Isoprene, occupational exposure to, 330t Isothiocyanates lung cancer and, 643 pancreatic cancer and, 739 Isotretinoin, in skin cancer prevention, 1243 Israel, migrants to, 192 cervical cancer in, 191, 191f, 198 melanoma in, 191, 191f, 197 prostate cancer in, 194 stomach cancer in, 195

Japan, migrants to, 192, 193 Jaundice, neonatal, testicular cancer and, 1157 Jejunal cancer. See Small Intestine cancer Jet fuel exposure, renal cancer and, 1093

K-ras biliary tract cancer and, 787 colorectal cancer and, 809, 810f, 811 pancreatic cancer and, 751, 752 Kaposi sarcoma, 522, 523t, 960t, 968. See also Soft tissue sarcoma classic (sporadic), 522 endemic (African), 522, 561, 562t, 968 epidemic (AIDS-associated), 522, 553 iatrogenic, 522 immunosuppression and, 500, 522 incidence of, 154, 959, 961, 962t post-transplantation, 552–553 Kaposi sarcoma-associated herpesvirus infection. See Human herpesvirus 8 infection Keratinocyte carcinoma. See Skin cancer Keratoses, laryngeal, 627 Kerosene, salivary gland cancer and, 688 Kidney cancer. See Renal cancer KLK10, 587t Knudson’s two-hit hypothesis, 48, 48f Koilocytotic atypia, cervical, 31, 1045 Kostmann syndrome, 851

Lactation breast cancer and, 998 ovarian cancer and, 1019 Lactose, ovarian cancer and, 1016–1017 Laryngeal cancer, 627–634 alcohol and, 246–247, 629–630, 633 asbestos and, 630 chemoprevention of, 633 chromosomal abnormalities in, 627–628 cigarette smoking and, 219t, 225–226, 246–247, 628–629, 630, 633 classification of, 627–628 diet and, 631 early detection of, 634 formaldehyde and, 630 fruits and vegetable intake and, 631 gastroesophageal reflux and, 632 genetic factors in, 632–633 human papillomavirus infection and, 631–632, 634 incidence of, 108f, 152, 152t, 628, 628t metals and, 630–631 mineral vitreous fibers and, 630 molecular pathogenesis of, 627–628, 633 mortality from, 141t, 152t, 628, 628t mustard gas and, 630 nickel and, 630–631 occupation and, 334t, 335t, 630–631 pathogenesis of, 627–628, 633 precursors of, 627 prevention of, 633–634 screening for, 634 smokeless tobacco and, 629 subsite classification of, 627, 629 survival for, 152, 172, 628, 628t susceptibility genes in, 586t time trends in, 628 virus infection and, 631–632, 634 vitamin A and, 631 vitamin C and, 415, 631 vitamin E and, 633 Larynx, precursor lesions of, 627 Lasiocarpine, 491t Laxatives, 491t, 500 multiple myeloma and, 940 renal pelvis cancer and, 1095 Lead gallbladder cancer and, 794 occupational exposure to, 330t, 333t Leather industry bladder cancer and, 1108 pancreatic cancer and, 746 sinonasal cancer and, 609 testicular cancer and, 1155 Legumes, prostate cancer and, 1132 Leiomyoma, 802, 960 Leiomyosarcoma, 14f, 802, 959, 960, 960t, 970. See also Soft tissue sarcoma LEP, 585t LEPR, 585t Leptin breast cancer and, 434 prostate cancer and, 435, 1138 Leukemia, 841–860 in atomic bomb survivors, 268, 268f, 848, 852, 854–855 benzene and, 230, 233, 339t, 849, 857 1,3-butadiene and, 338, 857 cervical cancer radiotherapy and, 264, 264f chemotherapy-related, 495, 849–850, 1277

Index childhood, 1252–1255, 1254f air pollution and, 370, 371t–373t chemical exposures and, 856, 1253 chlorination byproducts and, 388 chromosomal abnormalities and, 1253 electric appliances and, 312–313, 313t immune factors in, 555 incidence of, 854, 1252–1253, 1254f infection and, 856–857, 1255 ionizing radiation and, 265–266, 267, 1253 lymphoblastic, 968, 1252–1255, 1252t, 1254f immune factors in, 555 infection and, 555 magnetic field exposure and, 310t magnetic field exposure and, 308–311, 309t–311t, 855 maternal factors and, 856, 1253, 1255 microwave towers and, 316 myeloid, 1252–1255, 1252t, 1254f organic chemicals and, 392 parental occupation and, 855–856 paternal factors and, 279, 855–856 prenatal irradiation and, 265–266 prevention of, 859–860 radio transmitters and, 316 radiofrequency radiation and, 316–317 socioeconomic status and, 1255 tetrachloroethylene and, 392 chloramphenicol and, 493, 856 classification of, 15, 841 economic burden of, 210–211 ethylene oxide exposure and, 339 gamma rays and, 274 incidence of, 841 age and, 165f, 841, 844, 847f, 854 international, 102f, 106, 107f, 108f, 841, 845f–846f U.S., 160, 162, 162t, 164t, 841–842, 847f iodine-131 and, 270 ionizing radiation and, 260t, 261f, 262, 268, 268f low-frequency electric and magnetic fields and, 308–313, 309t–311t, 313t, 314, 315t lymphocytic acute, 846f, 847f, 853–857 age and, 847f, 854 alcohol and, 856 in atomic bomb survivors, 268, 268f, 854–855 in Bloom’s syndrome, 859 breastfeeding and, 856 chemicals and, 855–856 chemokines in, 858 chloramphenicol and, 493, 856 cigarette smoking and, 856 cytokines in, 858 diagnostic radiation and, 855 diet and, 856 in Down syndrome, 859 electromagnetic fields and, 855, 857 familial, 859 in farmers, 856 folic acid metabolism and, 858 gender and, 847f, 854 hydrocarbons and, 855–856 immune function and, 856–857, 859 incidence of, 162, 164t, 846f, 847f, 853–854

infection and, 856–857 initiation of, 853 ionizing radiation and, 854–855 in Li-Fraumeni syndrome, 859 medications and, 856 methylenetetradhydrofolate reductase gene in, 858 miscarriage and, 856 nuclear plant proximity and, 855 occupation and, 855 pesticides and, 855 precursors of, 853 prevention of, 859–860 race and, 847f, 854 radon and, 855 solvents and, 855–856 survival for, 854 syntomycin and, 856 therapeutic radiation and, 855 transmission line exposure and, 312 chronic, 846f, 847f, 853–854 age and, 847f, 854 autoimmune disease and, 858 benzene and, 857 butadiene and, 857 chemicals and, 857 chronic antigenic stimulation and, 555 cigarette smoking and, 857–858 ethylene oxide and, 857 familial, 859 gender and, 847f, 854 immune function and, 554, 858, 859 incidence of, 846f, 847f incidence of, U.S., 164t infection and, 858 initiation of, 853 ionizing radiation and, 857 occupation and, 857 race and, 847f, 854 styrene and, 857 survival for, 854 non-homologous end joining and, 858–859 survival for, 162, 173, 854 lymphoid, 852–859. See also Leukemia, lymphocytic classification of, 852–853 MOPP therapy and, 495 mortality from international, 102f, 106, 107f U.S., 142t, 162, 162t, 164t, 841, 842f–843f, 844f myeloid, 842–852 acute N-acetyl transferases in, 851 age and, 844, 847f alcohol and, 850 alkylating agents and, 849–850 in atomic bomb survivors, 268, 268f, 848 benzene and, 230, 233, 849 chemicals and, 849 chemotherapeutic agents and, 849–850 childhood, 1252–1255, 1252t, 1254f chloramphenicol and, 850 cigarette smoking and, 219t, 230, 850 classification of, 840 cytochrome P450 in, 851 cytogenetic abnormalities in, 842, 843–844, 851–852 diagnostic radiation and, 848

1375 diet and, 850 DNA repair pathways in, 851 in Down syndrome, 851–852 electromagnetic fields and, 849 familial aggregation of, 851–852 in farmers, 849 gender and, 844, 848t glutathione S transferases in, 850–851 hair dye and, 850 incidence of, 164t, 844, 846f, 847f ionizing radiation and, 846, 848–849 molecular genetics of, 850–851 occupation and, 848–849 platinum-based chemotherapy and, 850 race and, 844, 848t survival rates for, 844–845, 848t susceptibility genes in, 584t therapeutic radiation and, 848 topoisomerase II inhibitors and, 850 transmission line exposure and, 312 chronic age and, 844, 847f in atomic bomb survivors, 268, 268f, 852 chemicals and, 852 classification of, 840 diagnostic radiation and, 852 gender and, 848t, 855 incidence of, 164t, 844, 846f, 847f ionizing radiation and, 266, 852 Philadelphia chromosome in, 50 race and, 844, 848t survival for, 845, 848t therapeutic radiation and, 852 transmission line exposure and, 312 classification of, 15, 840 in nuclear facility workers, 272–273, 273t nuclear installations and, 278–279 nuclear weapons tests and, 277 occupation and, 271, 272–273, 273t, 314, 315t, 334t phenylbutazone and, 270, 492 phenytoin and, 499 phosphorus 32 and, 271 radiofrequency radiation and, 318, 318t in radiologists, 271 radium and, 265, 270 styrene exposure and, 337 susceptibility genes in, 584t, 585t, 586t, 588t Leukoplakia, 24–25, 674, 678–679 alcohol and, 681 descriptive epidemiology of, 24–25 etiology of, 25 isotretinoin treatment of, 36 malignant transformation of, 25 pathology of, 24 tobacco and, 233, 680 Li-Fraumeni syndrome, 53, 563t, 569 bone cancer in, 953 brain tumors in, 1189 breast cancer in, 1003 leukemia in, 859 melanoma in, 1213t, 1214 pancreatic cancer in, 747 radiotherapy in, second cancers and, 267 soft tissue sarcoma in, 969 Lichen sclerosus, penile cancer and, 1169 Life value, estimates of, 203–204 Lifetime prevalence, 102 Lifetime probability, 142

1376 Lindane, occupational exposure to, 332t Linkage disequilibrium, 94, 95–96 Linkage study, 94–95 association study vs., 95–96 nonparametric, 93t, 94–95 parametric, 93t, 94 Linoleic acid, prostate cancer and, 1133 a-Linolenic acid, prostate cancer and, 1133 Lip, precursor lesions of, 686 Lip cancer, 686–687 alcohol and, 687 cigarette smoking and, 219t, 226, 686 incidence of, 143, 144t, 675t, 686, 687f mortality from, 145t socioeconomic status and, 686 solar radiation and, 686 survival for, 143, 172, 686 Lipid-lowering drugs, 494 Lipoma, 960 Liposarcoma, 17, 17t, 959, 960, 960t, 970. See also Soft tissue sarcoma Liver angiosarcoma of, 764t, 773 arsenic and, 773 incidence of, 765–766, 766t Thorotrast and, 773 vinyl chloride and, 338, 339t, 773 cirrhosis of, 119, 773–774, 776 fatty change of, 776 focal nodular hyperplasia of, 764t hemangioendothelioma of, 271 hemangioma of, 763, 764t hepatoblastoma of, 763, 764t, 1260 incidence of, 119, 765–766, 766t nodular regenerative hyperplasia of, 764t Liver cancer, 763–778. See also Cholangiocarcinoma; Hepatocellular carcinoma anabolic steroids and, 772 a1-antitrypsin deficiency and, 775–776 arsenic and, 383, 383t, 385f, 773 in atomic bomb survivors, 269 chemoprevention of, 778, 1326t–1327t, 1330–1331 classification of, 11t, 763, 764t fibric acid derivatives and, 494 immune function and, 774 incidence of age and, 766 ethnicity and, 766, 767t international, 106, 107f, 108f, 119, 121f, 764–765, 765f migrant studies of, 119, 197 sex ratio of, 766 time trends in, 119 U.S., 121f, 146t, 147t, 148, 150, 765–766, 766t ionizing radiation and, 260t, 261f migrant studies of, 197 mortality from international, 106, 107f, 765 U.S., 141t, 149t, 766 occupation and, 334t, 335t organic chemicals and, 391 phenazopyridine hydrochloride and, 492 phenobarbital and, 498 phenylbutazone and, 492 plutonium and, 273 prevention of, 777–778, 1326t–1327t, 1330–1331

Index survival for, 150, 172, 765 susceptibility genes in, 584t Liver flukes biliary tract cancer and, 508t, 793 liver cancer and, 535, 772–773 Liver transplantation, in hepatocellular carcinoma, 777 Lomustine (CCNU), 491t Loss of heterozygosity, 50–51, 51f Louis-Barr syndrome. See Ataxia-telangiectasia Love Canal, 392 Low birth weight, educational attainment and, 180 Low-penetrance genes, in breast cancer, 1004–1005, 1004t Lung cancer, 638–651 air pollution and, 360–364, 361t, 362t–363t, 364f case-control studies of, 360, 361t, 362t cigarette smoking and, 360–361 cohort studies of, 360, 361t, 362t–363t in developing countries, 376 ecologic studies of, 360, 362t risk attribution for, 363 urban/rural differences in, 360, 361t alcohol and, 250–251, 643 arsenic and, 339t, 383t, 384, 385f, 644 asbestos and, 337, 339t, 360, 370, 373, 644 in atomic bomb survivors, 268 beta-carotene and, 643, 1329 bis chloromethyl ethers and, 644 body mass index and, 436 cadmium and, 337 carotenoids and, 643, 650 cell cycle control and, 649 chemoprevention of, 650, 1324, 1325t–1326t, 1328–1329 chest X-ray in, 650 chloroether exposure and, 339t chromium and, 644 chronic respiratory disease and, 642 cigarette smoking and, 7, 109, 196, 219t, 226, 641–642 classification of, 11t, 16, 638 computed tomography in, 650 cytochrome P-450 polymorphisms and, 645–646 diesel exhaust and, 363–364, 364f diet and, 416, 642–643 DNA methylation and, 647–648 DNA methyltransferase and, 648 DNA microarrays in, 16 DNA repair defects and, 648–649 economic burden of, 210, 211, 212, 213 environmental tobacco smoke and, 7, 374–375, 642 familial aggregation of, 644–645 familial relative risk in, 564t folate and, 643 gene expression profile in, 16 genetic factors in, 644–650 glutathione peroxidase and, 647 glutathione S-transferase and, 646–647 histopathology of, 14f, 638 host-cell reactivation assay in, 648 incidence of, 109, 1294 age and, 165f ethnicity and, 639–640, 640f gender and, 639, 639f geographic variation in, 640

by histologic subtype, 109, 111f international, 106, 107f, 108f, 109, 110f, 111f, 112f migrant studies of, 196 socioeconomic status and, 176t, 641 time trends in, 109, 112f U.S., 150–151, 151f, 152t, 638–641, 640f incinerator proximity and, 359 in insulation workers, 370 ionizing radiation and, 260t, 261f, 267, 268 isoniazid and, 493 isothiocyanates and, 643 man-made mineral fibers and, 336t, 339 methylenetetrahydrofolate reductase and, 647–648 microsomal epoxide hydrolase and, 646 migrant studies of, 196 mortality from education and, 177, 177t international, 106, 107f, 109 migrant studies of, 196 time trends in, 109, 112f twin study of, 91 U.S., 141t, 151, 151f, 152t, 638, 639f, 640–641, 640f, 641f mutagen sensitivity assay in, 648 myeloperoxidase and, 646 N-acetyltransferase and, 647 NAD(P)H quinone oxidoreductase 1 and, 646 nickel and, 394, 644 non-ferrous metal smelter proximity and, 359 non-small-cell, 16 obesity and, 436 occupation and, 334t, 335t, 643–644 overdiagnosis of, 1315 p53 and, 650 particulate matter and, 644 phenobarbital and, 498 physical activity and, 458–459, 460t, 461t, 462 plutonium and, 273 polycyclic aromatic hydrocarbons and, 358, 644 precursors of, 23, 638 prevention of, 642, 650, 1294, 1324, 1325t–1326t, 1328–1329 race and, 639–640, 640f radium and, 392–393 radon and, 274–276, 275f, 276f, 339t, 373, 644 in railroad workers, 364 respiratory disease and, 642 screening for, 38t, 650, 1314–1315 selenium and, 415, 643, 650 small-cell, 15, 16 socioeconomic status and, 176t, 177t, 181, 641 sputum cytology in, 650 sulfur dioxide and, 358–359 Surgeon General’s 1964 report on, 7 Surgeon General’s 1986 report on, 7 survival for, 151, 172, 640–641, 641f susceptibility genes in, 584t, 585t, 586t, 587t, 588, 588t, 590t tetrachloroethylene and, 392 toxic wastes and, 392 in truck drivers, 364, 364f twin study of, 91t, 93

Index vitamin A and, 414 vitamin E and, 415, 650 Lycopene prostate cancer and, 416, 1130–1131 skin cancer and, 1241 Lymph node, structure of, 549 Lymphangiosarcoma, 960t Lymphocytes, 549 B, 549–550, 551f for biomarkers, 82, 82t cryopreservation of, 82 half-lives of, 74 T, 549, 551, 551f Lymphoepithelioma-like carcinoma, gastric, 713 Lymphoid hyperplasia, 22 radiotherapy for, 263 Lymphoma Burkitt, 133, 508f, 511, 511t, 554 classification of, 12t, 17 DNA microarrays in, 17 economic burden of, 211 Hodgkin. See Hodgkin lymphoma immunosuppressants and, 500 ionizing radiation and, 261f MALT, 533, 906–907 non-Hodgkin. See Non-Hodgkin lymphoma radiofrequency radiation and, 306 styrene exposure and, 337 susceptibility genes in, 584t, 585t, 588t Lymphoproliferative disorders. See also NonHodgkin lymphoma autoimmune disease and, 554 in elderly, 555 post-transplantation, 552 X-linked, 551, 551t, 905, 910 Lynch syndrome. See Hereditary nonpolyposis colorectal cancer syndrome

Machinists, bladder cancer in, 1109 Macrophages, 549, 551f Magnesium, drinking water, 394 Malaria, Burkitt lymphoma and, 554 Malignant fibrous histiocytoma, 17 Malnutrition. See also Diet arsenic-induced cancer and, 384, 386 MALT lymphoma, 533, 906–907 Mammography, 184, 1005, 1007, 1285 Manhattan Project, 274 Manual of Tumor Nomenclature and Coding, 12 Marijuana smoking, 632, 680 Marital status, anal cancer and, 831–832 Markers, for linkage study, 94 Marshall Islands, 277, 984 Mastitis, radiation for, 260, 262 Mayak nuclear facility, 259, 273–274, 280, 393 MCIR melanoma and, 587t, 1211, 1215 skin cancer and, 587t, 1242 MDM-2, choriocarcinoma and, 1076 MDR1, 584t, 585t Meat bladder cancer and, 1112 brain cancer and, 1185 colorectal cancer and, 195, 817 esophageal cancer and, 700 oral cavity cancer and, 681 pancreatic cancer and, 727, 732t–738t, 738

prostate cancer and, 1133 renal cancer and, 1092 stomach cancer and, 712, 712t Mediastinal cancer, 11t, 152t Medical Expenditure Panel Survey, 206 Medications. See Pharmaceuticals Medroxyprogesterone acetate, cervical cancer and, 1052 Medulloblastoma, radiotherapy for, 267 MeEH, bladder cancer and, 1116 MEH, 583, 586t, 590t, 684 Melanocortin 1 receptor, gene for, 300 Melanocytes, solar radiation effect on, 300 Melanoma (cutaneous), 1196–1217. See also Melanoma (ocular) in airline flight crews, 1209 alcohol and, 1210 anatomic distribution of, 1196, 1197f, 1202–1203, 1203f antidepressants and, 494 burn scar and, 1210 CDK4 and, 1215 CDKN2A and, 1214–1215, 1214t chemoprevention of, 1216–1217 chromosomal abnormalities in, 1198 diet and, 1210 electromagnetic field exposure and, 1209 estrogen and, 1210 eye color and, 1211, 1211t familial, 563t, 569, 1212–1214, 1212t, 1213t familial relative risk in, 564t fertility drugs and, 1210 fluorescent lighting and, 1208 freckles and, 1211–1212, 1211t genetic susceptibility to, 1212–1215, 1212t, 1213t, 1214t in genetic syndromes, 1213–1215, 1213t, 1214t GSTM1 and, 1215 hair color and, 1211, 1211t histopathology of, 1196–1197, 1197f immunization and, 1210 immunosuppression and, 1210 incidence of age and, 165f, 1200, 1202f gender and, 1200–1202, 1202f international, 1203–1204, 1204f, 1205f, 1206f migrant studies of, 196–197, 1204–1205 race and, 1202–1203, 1202f socioeconomic status and, 1203, 1203f time trends in, 1199–1200, 1199f, 1200f U.S., 153t, 154, 1198–1200, 1200f infection and, 1210 injury and, 1210 ionizing radiation and, 1208–1209 MCIR and, 1211, 1215 melanocortin 1 receptor gene and, 300 migrant studies of, 196–197, 1204–1205 molecular genetics of, 1198 mortality from, 141t, 153t, 154, 197, 1198–1200, 1199f nevi and, 1197–1198, 1198t, 1211, 1211t, 1217 occupation and, 334t, 335t, 1208–1210 in oil industry workers, 1209 oral contraceptives and, 1210 parity and, 1210 pathogenesis of, 1215–1216 polychlorinated biphenyl exposure and, 1209

1377 polyvinyl chloride and, 1209 post-transplantation, 552 precursor lesions in, 1197–1198, 1198t prevention of, 1208, 1216–1217, 1216t, 1294, 1333 PUVA therapy and, 1208 skin type and, 1211 socioeconomic status and, 1203, 1203f solar radiation and, 197, 301, 1205–1208 SRCC3 and, 1215 survival for, 154, 172, 1200, 1201f, 1202f susceptibility genes in, 584t, 586t, 587t, 588t vaccine against, 1333 VDR and, 1215 Melanoma (ocular), 1217–1222 anatomic sites of, 1218 eye color and, 1220 genetic susceptibility to, 1220 histopathology of, 1218 incidence of, 160, 1218–1219, 1219f age and, 1219 ethnicity and, 1219 gender and, 1219 international, 1219 migrant studies of, 1219–1220 time trends in, 1218–1219 U.S., 1218–1219, 1219f migrant studies of, 1219–1220 molecular genetics of, 1218 mortality from, 1218 nevi and, 1220 occupation and, 1222 pathogenesis of, 1222 precursor lesions of, 1218 prevention of, 1222 radiofrequency exposure and, 318 solar radiation and, 301, 1220–1221, 1222 survival for, 160, 173, 1219 Melarsoprol, 496 Melphalan, 491t MEN1, 569 Menarche benign breast disease and, 31 breast cancer and, 998, 1002, 1006, 1006t choriocarcinoma and, 1082 hydatidiform mole and, 1082 ovarian cancer and, 1019 pancreatic cancer and, 750 physical activity and, 450 socioeconomic status-cancer association and, 183 Meningioma, 1175t, 1187. See also Brain cancer Menopause age at breast cancer and, 998 ovarian cancer and, 1019 hormone therapy after. See Postmenopausal hormone therapy Menstrual cycle, breast cancer and, 998, 1002 mEPHX, lung cancer and, 646 Merphalan, 491t Mesothelioma, 659–669 asbestos and, 661–666, 661t, 664t nonoccupational, 665, 666t occupational, 661–664, 662t–663t trends in, 666, 666f, 667f cigarette smoking and, 668 community-based studies of, 664, 665t

1378 Mesothelioma (Continued) diet and, 668 erionite and, 666–667 genetic factors in, 659, 668–669 geographic patterns of, 659, 660f incidence of, 152–153, 659, 660, 660f ionizing radiation and, 667 man-made vitreous fibers and, 667, 667t molecular pathogenesis of, 659 occupation and, 66t, 334t, 335t, 661–664, 661t, 662t–663t, 664t, 665t, 667t, 669 peritoneal, 662t–663t, 665–666 pleural plaques and, 664–665 prevention of, 669 simian virus 40 and, 667–668, 668t survival for, 660–661 MET, 571, 1271 Metabolic analysis, of early biologic effect biomarkers, 74–75 Metals. See also specific metals ambient air, 357t, 359 biliary tract cancer and, 794 half-life of, 73t laryngeal cancer and, 630–631 multiple myeloma and, 939 occupational exposure to, 326t, 330t, 333t, 336t, 339t pancreatic cancer and, 745 testicular cancer in, 1155 Metaplasia, 24 Metastasis angiogenesis and, 58–60, 59f extracellular matrix in, 58–59, 59f Methimazole, 495–496 Methotrexate, acute myeloid leukemia and, 850 Methoxsalen, 491t, 496 5-Methoxypsoralen, 491t, 496 2-Methyl-1-nitroanthraquinone, occupational exposure to, 331t Methyl-CCNU, 491t N-Methyl-N-nitrosourea, 491t a-Methylacyl-CoA racemase, in prostate cancer, 1142 Methylation, 51 aberrant, 51, 79 cervical cancer and, 1057 colorectal cancer and, 811, 816, 816f head and neck cancer and, 685 lung cancer and, 647–648 multiple myeloma and, 919 arsenic, 384, 386f profile of, 79 4,4¢-Methylene bis(2-chloroaniline), occupational exposure to, 328t 4,4¢-Methylene bis(2-methylaniline), occupational exposure to, 331t Methylene tetrahydrofolate reductase in arsenic metabolism, 386 colorectal cancer and, 816, 816f lung cancer and, 647–648 lymphocytic leukemia and, 858 multiple myeloma and, 931 4,4¢-Methylenedianiline, occupational exposure to, 331t Methylmercury, occupational exposure to, 330t Methylthiouracil, 491t, 495–496 Metronidazole, 491t, 495 Metropathia hemorrhagica, radiotherapy in, 264–265

Index MGMT, 584t Microsatellite, 565 Microsatellite instability, 49–50 in colorectal cancer and, 16, 811, 823 Microsomal epoxide hydrolase liver cancer and, 774–775 lung cancer and, 646 Migrant studies, 106, 189–199 age at migration in, 191, 192f analytic approaches to, 190–191 bladder cancer, 197–198 brain cancer, 1180–1181 breast cancer, 113, 193, 997 cervical cancer, 198, 1050 colorectal cancer, 114, 194–195, 814 contributions of, 189, 190t detection bias in, 189 diet, 406 ecologic fallacy in, 190 esophageal cancer, 129, 197 host country patterns in, 191–192, 191f, 192f incidence comparisons in, 189–190, 190t, 192–193 breast cancer, 193 colon cancer, 194 lung cancer, 196 malignant melanoma, 196–197 prostate cancer, 193–194 rectal cancer, 194–195 stomach cancer, 195 internal, 192–193 liver cancer, 197 lung cancer, 196 melanoma, 196–197, 1204–1205, 1219–1220 mortality comparisons in, 190, 190t, 193 breast cancer, 193 colon cancer, 195 lung cancer, 196 melanoma, 197 prostate cancer, 194 rectal cancer, 195 stomach cancer, 195–196 nasopharyngeal cancer, 198 non-Hodgkin lymphoma, 903 oral cavity cancer, 678 ovarian cancer, 198, 1015 pancreatic cancer, 197, 723–724 pharyngeal cancer, 678 proportional incidence ratios, 191 prostate cancer, 123, 193–194, 1130 rate ratios in, 190 rectal cancer, 194–195 screening and, 190 selection bias in, 189–190 standardized incidence ratios in, 190 standardized mortality ratios in, 190 stomach cancer, 118, 195–196, 709–710 testicular cancer, 1154 thyroid cancer, 198 uterine cancer, 198 Milk bladder cancer and, 1112 testicular cancer and, 1158 Mineral fibers. See also Asbestos laryngeal cancer and, 630 mesothelioma and, 667, 667t occupational exposure to, 336t, 339 Mineral oils occupational exposure to, 326t sinonasal cancer and, 611, 613t

Mining, radon exposure and, 274–275, 275f Mirex, occupational exposure to, 332t Mitochondrial DNA, mutations in, 1319 Mitochondrial enzymes, apoptosis and, 58 Mitomycin C, 491t Mitosis, 51–55, 52f, 54. See also Cell cycle Mitosis-promoting factor (MFP), 54 Mitotic arrest-deficiency (Mad) protein, 54, 55f Mitotic spindles, 54, 55f Mitoxantrone, 491t acute myeloid leukemia and, 850 MM1, 585t MMP-1, 585t MMP-2, 585t MMP-3, 585t MnSOD, 587t Mobile phones, 317–318, 317t Modifying genes. See Genetic susceptibility Mole, hydatidiform. See Hydatidiform mole Monoclonal gammopathy of undetermined significance, 523, 555, 922–923, 931. See also Multiple myeloma Monomethylarsonic acid reductase, 386 Monosomy 7, acute myeloid leukemia and, 851 MOPP therapy, 491t, 495 Mormons, pancreatic cancer in, 725 Mortality. See also at specific cancers cost of, 203–204, 204t data sources for, 103–104 definition of, 101 estimation of, 104 geographic variation in, 104–106, 105f obesity and, 437 socioeconomic status and, 177–178, 177t, 178t, 179f standardized, 103 trends in, 106 Mortality rate, definition of, 101 Mouth. See Oral cavity Mouthwash, 246, 682 MPO, 586t, 590t, 646 MSR1, 585t MTHFR, 583, 585t, 589 acute lymphocytic leukemia and, 858 colorectal cancer and, 816, 816f lung cancer and, 647–648 multiple myeloma and, 931 Muir-Torre syndrome. See Hereditary nonpolyposis colorectal cancer syndrome Multicentricity, of cancer precursors, 23 Multifactorial threshold model, 92 Multiple primary cancers, 1269–1278 alcohol and, 1272–1273 antioxidants and, 1273 breast cancer and, 1273–1275, 1274t chemotherapy and, 1277–1278, 1278t cigarette smoking and, 1272 coding of, 1270 endometrial cancer and, 1273–1275, 1274t familial cancer syndromes and, 954–955, 1271–1272 historical perspective on, 1270–1271 immunodeficiency and, 1276–1277 incidence of, 1270–1271 methodology for, 1269–1270 ovarian cancer and, 1273–1275, 1274t tobacco and, 1272 ultraviolet radiation and, 1275–1276 viral infection and, 1276–1277

Index Multiple endocrine neoplasia type 1, 563t, 569 type 2, 563t, 570 Multiple myeloma, 919–940 agricultural work and, 935, 937t alcohol and, 939 allergy and, 554, 925t, 929 asbestos and, 939 asthma and, 925t, 929 in atomic bomb survivors, 931, 932t, 935 autoimmune disorders and, 923, 924t, 928 benzene and, 938 chromosomal abnormalities and, 919 chronic antigenic stimulation and, 555 cigarette smoking and, 939 classification of, 919 cytokines in, 931 cytotoxic T lymphocyte antigen-4 in, 931 diet and, 939 diphenylhydantoin and, 940 engine exhaust exposures and, 939 erythromycin and, 940 familial aggregation of, 931 familial relative risk in, 564t formaldehyde and, 931 genetic factors in, 930–931 hair dye and, 940 HLA subtypes in, 931 host factors in, 922–931, 923t, 924t–928t human herpesvirus 8 and, 523, 929 human immunodeficiency virus infection and, 929 immunization and, 928t incidence of age and, 165f, 919, 920f international, 919–920, 920t race and, 919, 921, 921f socioeconomic status and, 922 time trends in, 921–922 U.S., 160, 162t, 919–922, 920f, 921f infection and, 925t–927t, 929 inflammatory disorders and, 929 interleukin-6 in, 931 interleukin-10 in, 931 ionizing radiation and, 260t, 261f, 266, 931–934, 932t–933t medical conditions and, 923–930, 924t–928t medications and, 940 metal exposures and, 939 methylene tetrahydrofolate reductase in, 931 monoclonal gammopathy of undetermined significance and, 922–923, 931 mortality from, 142t, 162t, 921–922, 921f NF-kB proteins in, 931 in nuclear weapons plant workers, 931, 932t–933t obesity and, 939 occupation and, 334t, 335t, 931–939, 932t–933t, 936t–937t osteomyelitis and, 929 paint exposure and, 938 pesticides and, 935, 936–937 petroleum refining and, 938 plastic manufacturing and, 938 prior medical conditions and, 923, 923t, 924t–928t in radiologists, 271 rubber industry and, 938 socioeconomic status and, 922 solvents and, 938

survival for, 160 time trends in, 921–922, 921f viral infection and, 926t–927t, 929 welding and, 939 wood industry and, 939 Munchausen syndrome, 563 Mustard gas laryngeal cancer and, 630 occupational exposure to, 327t, 336t Mutagen sensitivity assay, in lung cancer, 648 Mutation, 565–566. See also specific genes germline, 48–50, 566 point, 50 Myasthenia gravis, 554 c-MYC, 47, 49f v-MYC, 47 Myelodysplastic syndromes, 842–852. See also Leukemia, myeloid, acute in atomic bomb survivors, 848 chemotherapeutic agents and, 849–850 classification of, 843 familial aggregation of, 851 in genetic syndromes, 851–852 Myeloproliferative disorders, 842–845. See also Leukemia, myeloid, chronic classification of, 843

N-nitroso compounds brain cancer and, 1182–1183, 1184–1185 formation of, 388–389 long-term exposure to, 389 NAD(P)H quinone oxidoreductase 1, lung cancer and, 646 Nafenopin, 491t Naphthalene, occupational exposure to, 330t 2-Naphthylamine, occupational exposure to, 327t, 333t, 1107–1108 Nasal cavity anatomy of, 603, 604f cancer of. See Sinonasal cancer epithelium of, 603 papilloma of, 603 polyps of, 615 trauma to, 615 Nasopharyngeal cancer, 620–625 alcohol and, 623–624 anatomic distribution of, 620 cigarette smoking and, 226–227, 623–624 classification of, 11t CYP2E1 in, 624 demographic patterns of, 620–623, 621f, 621t, 622f, 622t diet and, 198, 621, 623 environmental factors in, 623–624 environmental tobacco smoke and, 624 Epstein-Barr virus and, 511–512, 511t, 623 formaldehyde and, 624 fruits and vegetables and, 623 herbal drugs and, 624 histopathology of, 620 HLA genes and, 624 host factors in, 624 incidence of age and, 620 international, 620, 621f, 621t migrant studies of, 198 time trends in, 621–622, 622t U.S., 144, 144t, 622–623, 622t migrant studies of, 198

1379 mortality from, 145t, 622t, 623 occupation and, 334t, 624 prevention of, 624–625 race-ethnicity and, 620–621, 621f, 621t salted fish and, 415 socioeconomic status and, 621, 622f survival for, 144, 172, 622, 622t susceptibility genes in, 585t, 586t, 588t time trends in, 621–622, 622t urbanization and, 621 wood working and, 624 NAT1 bladder cancer and, 1115 breast cancer and, 1004f, 1004t, 1005 colorectal cancer and, 819 esophageal cancer and, 702 hepatocellular carcinoma and, 775 lung cancer and, 647 oral cavity cancer and, 684 NAT2, 578, 583, 586t, 589, 590t acute myeloid leukemia and, 851 bladder cancer and, 1115 breast cancer and, 1004f, 1004t, 1005 colorectal cancer and, 819 hepatocellular carcinoma and, 775 lung cancer and, 647 oral cavity cancer and, 684 Natalizumab, in Crohn disease, 806 National Cancer Act (1971), 1283 National Cancer Institute, 1283–1284 National Program for Cancer Registries, 1288 Natural experiments, 5t, 7, 189. See also Migrant studies Natural killer cells, 549, 551f Naval shipyard workers, cancer in, 272 NBS1, 583 Neck cancer of. See Head and neck cancer; Laryngeal cancer; Pharyngeal cancer irradiation of, 263–264 Necrosis, 56 Neighborhoods, socioeconomic status-cancer association and, 184 Neoplasm, 47, 48f malignant transformation of, 55–60, 56f, 57f, 59f, 60f, 61f Neoplastic syndromes, 562–573, 563t. See also specific syndromes Nephroblastoma, 1258 Nervous system tumors, 1173–1190. See also Brain cancer classification of, 1173–1176, 1175t demographics of, 1176–1181, 1177t, 1178f, 1179f, 1180t, 1181t, 1182f, 1183t molecular genetics of, 1175–1176 Neuroblastoma, 1257–1258 hereditary, 563t influenza and, 1187 metronidazole and, 495 phenytoin and, 499 screening for, 1315 Neurofibromatosis type 1, 563t, 570, 851, 1189 type 2, 563t, 570, 1189 Neurofibrosarcoma, 960t Neuroma, acoustic, 1187 Neutrophils, 549, 551f Nevada, nuclear weapons tests in, 278 Nevoid basal cell carcinoma syndrome, 563t, 566, 1243, 1275

1380 Nevus (nevi) acquired, 1197, 1198t atypical, 1198, 1198t, 1217 congenital, 1198 melanocytic, 197 melanoma and, 197–198, 1198t, 1211, 1211t, 1217, 1220 pigmented, 197 skin cancer and, 1241–1242 New Zealand, migrants to, 193 melanoma in, 197 NF1, 570 NF2, 570 NF-kB proteins, in multiple myeloma, 931 Niacin, esophageal cancer and, 700 Nickel ambient air, 357t drinking water, 394 laryngeal cancer and, 630–631 lung cancer and, 644 occupational exposure to, 326t, 330t, 332, 339t pancreatic cancer and, 745 sinonasal cancer and, 609–610 Nicotine. See also Cigarette smoking addiction to, 222, 223f, 225 cigarette yield of, 223, 223f, 224f Nifuradene, 491t Nijmegen breakage syndrome, 551–552, 551t, 583, 858–859 Niridazole, 491t Nitrate daily intake of, 389 dietary, 388–389, 409 drinking water, 388–391 bladder cancer and, 389, 390t brain cancer and, 389, 390t, 1185 colon cancer and, 389, 390t exposure to, 388–389, 390t genotoxicity of, 391 Maximum Contaminant Level for, 388 non-Hodgkin lymphoma and, 389, 390t pancreas cancer and, 389, 390t prostate cancer and, 389 rectal cancer and, 389, 390t stomach cancer and, 389, 390t well water, 389, 391 2-Nitroanisole, occupational exposure to, 331t Nitrobenzene, occupational exposure to, 331t 2-Nitrofluorene, occupational exposure to, 331t Nitrogen, occupational exposure to, 332t Nitrogen mustard, 491t 2-Nitropropane, occupational exposure to, 331t 1-Nitropyrene, occupational exposure to, 331t 4-Nitropyrene, occupational exposure to, 331t Nitrosamines inhibition of, 739 oral cavity cancer and, 680 pancreatic cancer and, 725, 738, 745, 750, 752 tobacco-specific, 222, 680 urinary, 225 Nitrosoureas, brain cancer and, 1182–1183, 1184–1185 Nitrotriacetic acid, occupational exposure to, 332t NNK (4-methylnitrosamino-1-3-pyridyl-1butanone), 222, 233

Index NNN (N-nitrosonornicotine), 222 Nomenclature, 10, 12 Non-Hodgkin lymphoma, 898–911 alcohol and, 909 allergic disease and, 554, 911 asbestos and, 908 autoimmune disease and, 554 blood transfusion and, 911 Burkitt, 133, 508f, 511, 511t, 554 cancer history and, 909–910 celiac disease and, 911 childhood, 1257 chromosomal abnormalities and, 898–899, 900t chronic antigenic stimulation and, 555 cigarette smoking and, 230, 909 classification of, 898, 899t clusters of, 903 diet and, 909 drinking water and, 908 economic burden of, 214 environmental contaminants and, 908–909 Epstein-Barr virus and, 133, 905 familial relative risk in, 564t family history in, 909–910 follicular, 17, 22, 555 hair dye and, 909 hepatitis C virus infection and, 906 herpesviruses and, 905 histology of, 898, 899t, 901–902, 902t human immunodeficiency virus infection and, 553, 903, 905 immunodeficiency and, 910 incidence of age and, 165f, 901, 902t gender and, 901, 902t, 903f geographic variation in, 902–903, 904f international, 106, 107f, 108f, 132f, 133, 902–903, 904f migrant studies of, 903 race and, 901, 902t, 903f time trends in, 132f, 133, 899–901, 900f U.S., 160, 162t, 163f, 899–902, 901f, 902t, 903f ionizing radiation and, 260t, 907 molecular genetics of, 898–899, 900t mortality from geographic variation in, 902, 904f international, 106, 107f, 132f, 133 U.S., 132f, 133, 142t, 160, 162t, 163f, 899, 900f, 901, 902t, 903f nitrate and, 389, 390t obesity and, 909 occupation and, 334t, 907–908 oral contraceptives and, 909 pesticides and, 908 phenylbutazone and, 492 phenytoin and, 499 polio vaccine and, 906 post-transplantation, 910 postmenopausal hormone replacement therapy and, 909 pregnancy and, 909 race and, 901, 902t, 903f radiotherapy for, acute myeloid leukemia and, 848 retroviruses and, 553, 900–901, 903, 905 rheumatoid arthritis and, 554, 910–911 second cancers after, 1277–1278, 1278t simian virus 40 and, 906

in Sjögren syndrome, 911 skin cancer and, 910 small intestine, 801–802, 804 solar radiation and, 133, 301, 301t, 1275–1276 solvents and, 908 survival for, 160, 173 susceptibility genes in, 584t, 586t systemic lupus erythematosus and, 911 tonsillectomy and, 553–554 ultraviolet light exposure and, 133 Non-homologous end joining, in lymphocytic leukemia, 858–859 Nonidentifiability, in age-period-cohort modeling, 106 Nonionizing radiation, 281, 306–319. See also Extremely low-frequency electric and magnetic fields; Radiofrequency radiation biological system interaction with, 306 public concern with, 306–307 Nonsteroidal anti-inflammatory drugs, 492–493, 1322–1323 bladder cancer and, 1112 colorectal cancer and, 492, 822–823, 1330 esophageal cancer and, 701 Hodgkin lymphoma and, 890 melanoma and, 1217 pancreatic cancer and, 742, 743 prostate cancer and, 1135 stomach cancer and, 715 Nose bleeds, sinonasal cancer and, 615 NQO1, 584t, 586t bladder cancer and, 1116 lung cancer and, 646 myeloid leukemia and, 851 NRAS, melanoma and, 1198 Nuclear power/weapons facilities, 259, 272–274, 273t, 278–280 bladder cancer and, 1114, 1117 leukemia and, 272–273, 278–279, 280, 855 multiple myeloma and, 931, 932t–933t thyroid cancer and, 280, 984–985 Nuclear weapons testing, 277–278, 984 Nutrients, 408–409, 408t. See also Diet

Oak Ridge National Laboratory, 272 Obesity, 413, 417, 422–439, 1286 attributable risk and, 438–439, 439f, 439t biliary tract cancer and, 794 breast cancer and, 413, 429–434, 431f, 438t, 458 cancer mortality and, 437 colon cancer and, 423–425, 424f, 438t, 817 colorectal adenoma and, 29, 425 in diet studies, 412 endometrial cancer and, 427–429, 428f, 438t, 439, 439f, 1029–1030 esophageal cancer and, 425–426, 426f, 438t, 701, 701t estradiol bioavailability and, 433 gastric cancer and, 425–426, 438t gastroesophageal reflux disease and, 27 head and neck cancer and, 437 leptin and, 434 lung cancer and, 436 mortality and, 437 multiple myeloma and, 939

Index non-Hodgkin lymphoma and, 909 ovarian cancer and, 429, 430f pancreatic cancer and, 726–727 physical activity effects on, 450 population attributable risk and, 438–439, 439f, 439t prevalence of, 438, 439, 439f prevention of, 437–438 prostate cancer and, 434–436, 435f, 438t, 1134–1135 rectal cancer and, 425 renal cancer and, 426–427, 427f, 438t, 1090 socioeconomic status-cancer association and, 181, 182 stomach cancer and, 714 testicular cancer and, 1158 thyroid cancer and, 436 Occult blood test, 1312–1313 Occupation, 322–346. See also specific chemicals and occupations acute lymphocytic leukemia and, 855–856 acute myeloid leukemia and, 848–849 biliary tract cancer and, 794 bladder cancer and, 334t, 335t, 1107–1110, 1109t, 1113, 1114 bone cancer and, 334t, 335t brain cancer and, 314, 315t, 334t, 335t, 1181, 1183t, 1188 breast cancer and, 314–315, 335t cervical cancer and, 334t, 335t, 1052–1053 chemical exposures with, 322–346. See also specific chemicals animal experimentation for, 323 attributable risk in, 339–340 cigarette smoking and, 343 classification for, 323–325, 324t, 325t community-based studies of, 341–342 confounders in, 342–343 data on, 322–323, 345, 346 evidence for, 338–339, 339t gene interaction with, 343–344 group 1, 325t, 326t–327t, 333t group 3, 325t group 4, 325t group 2A, 325t, 328t–329t, 330t–332t, 333t group 2B, 325t, 330t–332t, 333t historical perspective on, 322 industry-based studies of, 340–341 prevention of, 344–345 research need on, 345–346 reviews on, 335t–336t short-term tests for, 323 site-specific, 334t structure-activity relationships in, 323 study design for, 340–343 chronic lymphocytic leukemia and, 857 chronic myeloid leukemia and, 852 CNS cancer and, 1256 colon cancer and, 335t, 744 diesel exhaust exposure with, 363–370, 364f, 365t–370t esophageal cancer and, 334t, 335t, 702 Hodgkin lymphoma and, 885–887, 886t ionizing radiation exposure with, 271–275, 273t laryngeal cancer and, 334t, 335t, 630–631 leukemia and, 314, 315t, 334t, 848–849, 855–856, 857 liver cancer and, 334t, 335t

low-frequency electric and magnetic field exposures with, 313–314, 315t lung cancer and, 334t, 335t, 643–644 melanoma and, 334t, 335t, 1208–1210, 1222 mesothelioma and, 334t, 335t, 661–666, 661t, 662t–663t, 664t, 665t, 666t, 667f, 669 multiple myeloma and, 334t, 335t, 931–939, 932t–933t, 936t–937t nasopharyngeal cancer and, 334t, 624 neuroblastoma and, 1257 non-Hodgkin lymphoma and, 334t, 907–908 ocular melanoma and, 1222 oral cavity cancer and, 683 ovarian cancer and, 334t, 335t, 1018 pancreatic cancer and, 334t, 335t, 743–747 pharyngeal cancer and, 334t, 335t, 683 prostate cancer and, 1136 radiofrequency radiation exposure with, 318, 318t renal cancer and, 334t, 335t, 1092–1093 renal pelvis cancer and, 1095–1096 safety regulations and, 1349–1350 salivary gland cancer and, 688 sarcoma and, 334t, 335t sinonasal cancer and, 334t, 335t, 608–613, 613t skin cancer and, 334t, 1239 socioeconomic status and, 174 socioeconomic status-cancer association and, 183 soft tissue sarcoma and, 962–967, 963t–966t stomach cancer and, 334t, 335t testicular cancer and, 335t, 1155 thyroid cancer and, 334t, 985, 987 ureter cancer and, 1095–1096 Ochratoxin A, testicular cancer and, 1156 ODC, 585t OGG1, 584t Oil industry melanoma and, 1209 renal cancer in, 1093 Oil orange SS, occupational exposure to, 331t Okazaki fragments, 56, 56f Oligodendroglioma, 16, 17. See also Brain cancer Olive oil, breast cancer and, 1000 Oltipraz, in hepatocellular carcinoma prevention, 778 Omi/HtrA2, in apoptosis, 57f, 58 Oncogenes, 47–50, 48t, 49t activation of, 50 amplification of, 50 Opisthorchis spp., 508, 535 biliary tract cancer and, 793 liver cancer and, 772–773 OPRM1, 584t Oral cavity, precursor lesions of, 24–25, 674, 678–679 alcohol and, 681 cytogenetics of, 25 diet and, 682 etiology of, 25 pathology of, 24 progression of, 25 tobacco and, 680 Oral cavity cancer, 674–689. See also Lip cancer; Salivary gland cancer N-acetyltransferases and, 684

1381 alcohol and, 245–246, 680–681, 683 alcohol dehydrogenase and, 684 anatomic distribution of, 674, 675t, 679–680 betel quid and, 680 bidi smoking and, 679 cell cycle control and, 685 chemoprevention of, 685 cigar smoking and, 679 cigarette smoking and, 219t, 226, 679–680 classification of, 11t cytochrome P-450 and, 684 dental factors and, 682–683 diet and, 681–682, 685 DNA repair genes and, 684–685 early detection of, 686 environmental tobacco smoke and, 679 erythroplakia and, 679 familial aggregation of, 683 Fanconi anemia and, 683 genetic factors in, 684–685 glutathione S-transferase and, 684 histopathology of, 674 human papillomaviruses and, 682 incidence of, 674, 675–679, 675t, 676f age and, 165f, 675, 676f deprivation and, 677 ethnicity and, 675, 675t gender and, 675, 676t, 677, 678f international, 106, 107f, 108f, 677–678, 677f, 678f migrant studies of, 678 race and, 675, 675t socioeconomic status and, 675–677 U.S., 142–144, 144t, 675, 676f leukoplakia and, 678–679 marijuana smoking and, 680 microsomal epoxide hydrolase and, 684 molecular genetics of, 674–675 mortality from, 106, 107f, 141t, 143, 145t mouthwash use and, 682 multiple primaries in, 683–684, 685 occupation and, 683 oral hygiene and, 682–683 pan and, 680 pathogenesis of, 685 pipe smoking and, 679 precursors of, 674, 678–679, 681, 682 prevention of, 685–686 psoriasis and, 683 recurrence of, 685–686 retinoids in, 685–686 screening for, 38t, 686 smokeless tobacco and, 680 socioeconomic status and, 675–677 survival for, 143, 172, 675, 676t susceptibility genes in, 584t Oral contraceptives, 468, 483 benign breast disease and, 31 breast cancer and, 469, 998, 1274 cervical adenocarcinoma in situ and, 34 cervical cancer and, 477–478, 1052, 1053 cholangiocarcinoma and, 771 choriocarcinoma and, 1079–1080, 1080t, 1082 colorectal cancer and, 480, 482, 482t endometrial cancer and, 473, 474t, 1031, 1033 hepatocellular carcinoma and, 771 hydatidiform mole and, 1080, 1080t, 1082 melanoma and, 1210

1382 Oral contraceptives (Continued) neuroblastoma and, 1257 non-Hodgkin lymphoma and, 909 ovarian cancer and, 1014, 1014f renal cancer and, 1092 soft tissue sarcoma and, 968 thyroid cancer and, 981 Oral hygiene, oral cavity cancer and, 682–683 Orchitis, testicular cancer and, 1158 Organ transplantation anal cancer and, 552 cervical cancer and, 552 immunosuppression in, 499–500, 552–553 Kaposi sarcoma and, 552–553 lymphoproliferative disorders and, 552 melanoma and, 552 non-Hodgkin lymphoma and, 910 penile cancer and, 1168 skin cancer and, 552 soft tissue sarcoma and, 969 Organochlorines. See Pesticides Ornithine decarboxylase, prostate cancer and, 1139 Oropharyngeal cancer. See Oral cavity cancer; Pharyngeal cancer Osteitis deformans (Paget disease), 23, 954 Osteomyelitis, multiple myeloma and, 929 Osteoporosis, meningioma and, 1187 Osteosarcoma in Bloom syndrome, 953–954 childhood, 1258–1259 extraskeletal, 960t fluoride and, 395 in Hutchinson-Gilford progeria, 954 implants and, 952 incidence of, 946, 947f, 947t ionizing radiation and, 946, 948–952, 949t, 952t in Li-Fraumeni syndrome, 953 mortality from, 946, 948f plutonium and, 950 in polyostotic fibrous dysplasia, 954 precursors of, 23 prevention of, 955 radiotherapy-related, 950–952 radium and, 270, 271, 338, 393, 948, 949t, 950 retinoblastoma and, 952–953 in Rothmund-Thomson syndrome, 953, 955 Thorotrast and, 950, 951f trauma and, 952 viral infection and, 952 in Werner syndrome, 953 Ovarian cancer, 1013–1022, 1022t age and, 165f, 1013, 1014f age at menarche and, 1019 age at menopause and, 1019 alcohol and, 1016 analgesic use and, 1015, 1016t androgens and, 1020–1021 antidepressants and, 493 antioxidants and, 1017–1018 in atomic bomb survivors, 1019 body mass index and, 429, 430f, 1017 BRCA genes in, 565, 567, 1020 CA125 in, 1022t chemical agents and, 1018 chemotherapy for, acute myeloid leukemia and, 850

Index chromosomal alterations in, 1020 cigarette smoking and, 229, 1019 classification of, 11t, 1013 coffee/caffeine and, 1015–1016 diet and, 1015–1017 economic burden of, 210, 213 endometriosis and, 1020 estrogen-secreting, 1029 familial, 563t, 1020 familial relative risk in, 564t fat intake and, 1017 fertility drugs and, 1019–1020 fiber intake and, 1017 galactose and, 1016–1017 height and, 1017 hereditary, 567 hormonal theories of, 1021 hysterectomy and, 1018 incessant ovulation theory of, 1021 incidence of age and, 165f, 1013, 1014 international, 106, 107f, 108f, 1014, 1015f migrant studies of, 198, 1015 race and, 1014, 1014f socioeconomic status and, 1014 time trends in, 1013 U.S., 156–157, 156t, 157f, 1013, 1014f infertility and, 1019–1020 insulin-like growth factor I and, 1021 ionizing radiation and, 260t, 261f, 1019 lactation and, 1019 lactose and, 1016–1017 migrant studies of, 198, 1015 mortality from, 1015 international, 106, 107f U.S., 141t, 156–157, 156t, 157f, 1013 multiple cancers and, 1273–1275, 1274t obesity and, 429, 430f occupation and, 334t, 335t, 1018 oral contraceptives and, 478, 479t, 1015 parity and, 1019 pathogenesis of, 1021–1022 pelvic contamination theory of, 1021–1022 phenolphthalein and, 500 physical activity and, 459, 462, 1017–1018 postmenopausal hormone therapy and, 479–480, 480t, 484f–485f, 1015 pregnancy and, 1019 prevention of, 1022 progesterone in, 1021 progesterone receptor in, 1020 proteomics in, 1021 race and, 1014, 1014f screening for, 1022, 1022t socioeconomic status and, 1014 solar radiation and, 301, 301t survival for, 157, 172, 1013, 1014f susceptibility genes in, 584t, 586t talc and, 1018 time trends in, 1013 tubal ligation and, 1018, 1018t vitamins and, 1017 weight and, 1017 Ovary (ovaries) ablation of breast cancer and, 264 sarcoma regression and, 968 polycystic, 1029 Oxazepam, 491t

p53, 569, 587t, 1319 anal cancer and, 837 biliary tract cancer and, 787 bladder cancer and, 1117 brain cancer and, 1176 breast cancer and, 1003 choriocarcinoma and, 1076 colorectal cancer and, 821, 821f esophageal cancer and, 702 hepatocellular carcinoma and, 770 leukemia and, 859 lung cancer and, 650 melanoma and, 1198 pancreatic cancer and, 751 sinonasal cancer and, 603 thyroid cancer and, 987–988 p73, in leukemia, 859 p21 (CDKN1A), 52–53, 53f Paget disease, 23, 954 Painters, bladder cancer in, 1108–1109 Palygorskite, occupational exposure to, 330t Pan, oral cavity cancer and, 680 Pancreas intraepithelial neoplasia of, 23, 721, 750 K-ras mutations in, 23 Pancreatic cancer, 22, 721–753 acrylamide and, 746 age at menarche and, 750 alcohol and, 249–250, 741–742 allergy and, 749–750 androgen receptors in, 750 asbestos and, 744–745 aspirin and, 742–743 asthma and, 749 caloric intake and, 725–726 carbohydrates and, 739–740 chlorinated hydrocarbons and, 745–746 chlorination byproducts and, 388 cholecystitis and, 749 cholesterol and, 740 chromate and, 746 chromosomal abnormalities in, 750 cigarette smoking and, 219t, 228, 725, 752 classification of, 11t, 721 coffee and, 741 dairy products and, 732t–738t, 738 diabetes mellitus and, 747–748 diagnosis of, 721 diet and, 727–742, 728t–738t cereals and, 739 dairy products and, 732t–738t, 738 eggs and, 732t–738t, 738 fruits and vegetables and, 728t–731t, 738–739 meat intake and, 727, 732t–738t, 738 disinfection by-products and, 388 DNA-adduct formation in, 752 DPC4 in, 751 drugs and, 742–743 economic burden of, 213 eggs and, 732t–738t, 738 electromagnetic fields and, 746 energy balance and, 725–727 epidermal growth factor in, 751 estrogen receptors in, 750 in familial atypical multiple mole melanoma syndrome, 747 familial relative risk in, 564t fat intake and, 740

Index fiber intake and, 740 formaldehyde and, 746 fruits and vegetables and, 728t–731t, 738–739 gastric surgery and, 749 genetic factors in, 747, 751, 752 glycemic index and, 739–740 growth factor receptors in, 751 growth factors in, 751 hereditary non-polyposis colon cancer syndrome and, 747 hormones and, 750 hyperinsulinemia and, 726 incidence of age and, 165f, 724 gender and, 724 international, 106, 107f, 108f, 722–723 migrant studies of, 197, 723–724 race and, 722 religion and, 724–725 socioeconomic status and, 724 time trends in, 722 U.S., 147t, 150, 721, 722t insulin resistance and, 726 ionizing radiation and, 261f, 744 K-ras in, 751, 752 leather tanning and, 746 meats and, 727, 732t–738t, 738 medical conditions and, 747–750 metals and, 745 micronutrients and, 740–741 migrant studies of, 197, 723–724 molecular pathogenesis of, 750–752 mortality from age-specific, 724, 724f international, 106, 107f, 722, 722f, 723f U.S., 141t, 149t, 721, 722 nitrate and, 389, 390t nitrosamines and, 725, 738, 745, 750, 752 nonsteroidal anti-inflammatory drugs and, 743 nutrients and, 739–741 obesity and, 726–727 occupation and, 334t, 335t, 743–747 p53 in, 747, 751 p16 protein in, 751 pancreatitis and, 742, 748–749 parity and, 750 pathology of, 721 pernicious anemia and, 750 pesticides and, 745–746 in Peutz-Jeghers syndrome, 747 physical activity and, 726 polycyclic aromatic hydrocarbons and, 738, 745 prevention of, 752 protein intake and, 740 reproductive factors in, 750 sedentary occupation and, 744 socioeconomic status and, 724 solvents and, 746 stone quarrying and, 746 sulfite paper process and, 746 survival for, 150, 172, 721–722 susceptibility genes in, 584t, 586t tea and, 741 textile industry and, 746 tonsillectomy and, 750 transforming growth factor-beta in, 751

twin study of, 91 vitamin C and, 740, 741 woodworking and, 746 Pancreatic intraepithelial neoplasia, 23, 721, 750 Pancreaticobiliary duct anomaly, biliary tract cancer and, 795 Pancreatitis hereditary, 747 pancreatic cancer and, 742, 748–749 Panfuran S, 491t Pap smear, 21–22, 38t, 1044, 1045–1046, 1058, 1285 Paper industry, pancreatic cancer and, 746 Papilloma, sinonasal, 603 Paraffin, chlorinated, occupational exposure to, 331t Paraganglioma, 563t, 570 Paranasal sinuses anatomy of, 603, 604f cancer of. See Sinonasal cancer papilloma of, 603 Parasitic infestation, 535 Parity benign breast disease and, 31 breast cancer and, 998 cervical cancer and, 1051–1052 choriocarcinoma and, 1081 endometrial cancer and, 1034 hepatocellular carcinoma and, 772 melanoma and, 1210 ovarian cancer and, 1019 pancreatic cancer and, 750 renal cancer and, 1092 Paroxetine, 493 Particulate matter, ambient air, 357t, 358–359 Passive smoking. See Environmental tobacco smoke PAX8-peroxisome proliferator-activated receptor gamma genes, in follicular thyroid carcinoma, 13 Penetrance, gene, 96, 566 Penile cancer, 1166–1170 age and, 1166, 1167f cervical cancer and, 1051 cigarette smoking and, 229, 1168 circumcision and, 1168–1169, 1170 classification of, 1166 human immunodeficiency virus infection and, 1168, 1170 human papillomavirus infection and, 1167–1168, 1170 immunosuppression and, 1168 incidence of, 158, 158t, 1166–1167, 1167f inflammation and, 1169 lichen sclerosus and, 1169 mortality from, 158, 158t, 1166 pathogenesis of, 1169–1170 personal hygiene and, 1169 phimosis and, 1169 prevention of, 1170 PUVA therapy and, 1168 race and, 1166 renal transplantation and, 1168 socioeconomic status and, 1166 survival for, 158, 172, 1166 Penile intraepithelial neoplasia, 1166 Pepsinogen C, gene for, 585t Peptic ulcer, radiotherapy for, 265

1383 Perchloroethylene drinking water, 391 renal cancer and, 1093 Period study, 106 Peritoneal cancer, 659–669. See also Mesothelioma Pernicious anemia multiple myeloma and, 928 pancreatic cancer and, 750 Person-years of life lost, definition of, 102 Pesticides. See also Herbicides acute lymphocytic leukemia and, 855 biliary tract cancer and, 794 breast cancer and, 1001 in drinking water, 391–392 gallbladder cancer and, 794 half-life of, 73t multiple myeloma and, 935 non-Hodgkin lymphoma and, 908 occupational exposure to, 327t, 328t–329t, 331t, 336t pancreatic cancer and, 745–746 regulation of, 1349 sinonasal cancer and, 612 soft tissue sarcoma and, 962, 963t–965t, 966–967 Petroleum industry chronic lymphocytic leukemia and, 857 multiple myeloma and, 938 Petroleum oils, sinonasal cancer and, 611, 613t Peutz-Jeghers syndrome, 563t, 747, 802–803 pH, urine, bladder cancer and, 1113 Pharmaceuticals, 489–500, 491t. See also specific drugs analgesic-antipyretic, 490–494, 491t antibacterial, 491t, 493 anticonvulsant, 498–499 antidepressant, 493–494 antifungal, 491t, 495 antihypertensive, 497, 1090–1091 antilipemic, 494 antineoplastic, 491t, 494–495, 849–850, 952, 1277–1278, 1278t antiprotozoal, 491t, 495 antithyroid, 495–496 cardiovascular, 497–498 data on, 490, 491t definition of, 489 dermatologic, 496 dosage of, 490 esophageal cancer and, 701 exposure-disease relationships and, 490 genotoxicity of, 489 hormonal. See Oral contraceptives; Postmenopausal hormone therapy immunosuppressant, 499, 552–553 International Agency for Research on Cancer evaluation of, 490, 491t postmarketing surveillance of, 489–490 premarketing tests of, 489 regulation of, 1348 response to, 489 safety of, 489 use of, 490 Pharyngeal cancer, 674–689. See also Oral cavity cancer alcohol and, 245–246, 680–681, 683 alcohol dehydrogenase and, 684 bidi smoking and, 679

1384 Pharyngeal cancer (Continued) cell cycle control and, 685 chemoprevention of, 685 cigar smoking and, 679 cigarette smoking and, 219t, 226, 679–680 cytochrome P-450 and, 684 dental factors and, 682–683 diet and, 681–682 DNA repair genes and, 684–685 early detection of, 686 environmental tobacco smoke and, 679 erythroplakia and, 679 familial aggregation of, 683 Fanconi anemia and, 683 genetic factors in, 684–685 glutathione S-transferase and, 684 histopathology of, 674 human papillomaviruses and, 682 incidence of, 674, 675–679, 675t, 676f age and, 165f, 675, 676f, 676t ethnicity and, 675, 675t, 676t gender and, 675, 676t international, 677–678, 677f, 678f migrant studies of, 678 race and, 675, 675t, 676t socioeconomic status and, 675–677 U.S., 142–144, 144t, 675, 675t, 676f marijuana smoking and, 680 microsomal epoxide hydrolase and, 684 molecular genetics of, 674–675 molecular pathogenesis of, 674–675 mortality from, 141t, 145t multiple primaries in, 683–684, 685 N-acetyltransferases and, 684 occupation and, 334t, 335t, 683 pan and, 680 pathogenesis of, 685 pipe smoking and, 679 precursors of, 674, 678–679, 681, 682 prevention of, 685–686 psoriasis and, 683 recurrence of, 685–686 screening for, 686 smokeless tobacco and, 680 survival for, 143, 172, 675, 676t tobacco and, 679–680 Phenacetin, 490–492, 491t bladder cancer and, 1112 renal cancer and, 1091 renal pelvis cancer and, 1095 ureter cancer and, 1095 Phenazopyridine hydrochloride, 491t, 492 Phenobarbital, 491t, 498–499, 1112 Phenol, acute myeloid leukemia and, 850 Phenolphthalein, 491t, 500 Phenotype, 566 Phenotypic assays, 77 Phenoxybenzamine hydrochloride, 491t Phenyl glycidyl ether, occupational exposure to, 330t Phenylbutazone, 492 Phenytoin, 491t, 499 Hodgkin lymphoma and, 499 lymphoproliferative syndrome with, 553 Philadelphia chromosome, 50, 843 Phimosis, penile cancer and, 1169 Phosphorus 32, 270–271 Photoageing, 299 Photochemotherapy, penile cancer and, 1168

Index Physical activity, 449–463 biological effects of, 450–451 breast cancer and, 455–458, 456t, 457t, 462, 1001, 1002 in cancer patient, 462 in cancer prevention, 417, 449 colorectal adenoma and, 29 colorectal cancer and, 451–455, 452t–455t, 462, 817 in diet studies, 413 endometrial cancer and, 459, 461t, 462, 1029–1030 improved participation in, 1285 lung cancer and, 458–459, 460t, 461t, 462 measurement of, 451 non-Hodgkin lymphoma and, 909 obesity and, 429 ovarian cancer and, 459, 462, 1017–1018, 1029–1030 pancreatic cancer and, 726 prevalence of, 449–450 prostate cancer and, 458, 459t, 460t, 1135 rectal cancer and, 452t–454t, 455 socioeconomic status-cancer association and, 182 testicular cancer and, 1158–1159 Phytoestrogens breast cancer and, 1000 half-life of, 73t prostate cancer and, 1132 Pig farming childhood brain cancer and, 1188 hydatidiform mole and, 1080 Pipe smoking, 221, 232 bladder cancer and, 1107 historical perspective on, 217, 218f laryngeal cancer and, 629 oral cavity cancer and, 679 pancreatic cancer and, 725 sinonasal cancer and, 614 Placental site trophoblastic tumor, 1076 Placental tumors, 11t Plasma, for biomarkers, 82, 82t Plastic manufacturing, multiple myeloma and, 938 Platelets, for biomarkers, 82, 82t Plausibility, causation and, 5t, 6–7 Pleural cancer, 152t, 659–669. See also Mesothelioma Pleural plaques, asbestos exposure and, 664–665 665t Plutonium, 259, 273–274, 280, 393, 950 Pneumonia, lung cancer and, 642 Pneumosclerosis, plutonium, 273 Polio vaccine brain tumors and, 1187 non-Hodgkin lymphoma and, 906 Polychlorinated biphenyls (PCBs) half-life of, 73t melanoma and, 1209 non-Hodgkin lymphoma and, 908 occupational exposure to, 328t Polychlorophenols, occupational exposure to, 332t Polycyclic aromatic hydrocarbons ambient air, 358 bladder cancer and, 1109 colorectal cancer and, 819 lung cancer and, 358, 644

mutagenicity of, 357–358 occupational exposure to, 335, 336t pancreatic cancer and, 738, 745 renal cancer and, 1093 Polycyclic organic matter, ambient air, 358 Polycythemia vera, phosphorus 32 in, 270–271 Polymorphism, 566 Polyomaviruses CNS cancer and, 1256 non-Hodgkin lymphoma, 906 Polyostotic fibrous dysplasia, 954 Polyps colorectal, 27–29, 28, 28f, 29, 425, 809, 823. See also Familial adenomatous polyposis alcohol and, 249 cigarette smoking and, 227 nonsteroidal anti-inflammatory drug effects on, 492 nasal, 615 Polysulfone badge dosimeter, for solar radiation measurement, 295–299, 297t Polyvinyl chloride, melanoma and, 1209 PON1, 584t Ponceau MX, occupational exposure to, 331t Ponceau 3R, occupational exposure to, 331t Population at risk, 102–103 world standard, 103 Population-attributable fraction, 93–94 Population stratification, genetic polymorphisms and, 78 Population study, 8 gene-based, 579–583, 580t, 581t–582t Porphyria, liver cancer and, 775 Positional cloning, 93t, 94–96 Postmenopausal hormone therapy, 468 benign proliferative epithelial disorders of breast and, 31 body mass index and, 432 breast cancer and, 470–473, 471t–472t, 484f–485f, 998–999 cervical cancer and, 34, 478 colorectal adenoma and, 29 colorectal cancer and, 482–483, 483t, 484f–485f, 484t, 823 coronary artery disease and, 1298 endometrial cancer and, 473–477, 475t, 476t–477t, 484f–485f, 1030–1033, 1032t, 1274 endometrial hyperplasia and, 35, 36 estrogen plus progestins for, 470–472, 471t–472t, 474–476, 476t–477t, 478, 480, 481t, 483, 483t, 484t liver cancer and, 771 melanoma and, 1210 non-Hodgkin lymphoma and, 909 ovarian cancer and, 479–480, 480t, 481t, 484f–485, 1015 sinonasal cancer and, 615 temporal patterns of, 468–469, 469f unopposed estrogens for, 470–472, 471t–472t, 473–474, 475t, 478, 479–480, 480t, 482–483, 483t use of, 483–485 Potassium bromate, occupational exposure to, 332t Power lines, residential exposure to, 309t–311t Power transmission lines, 308–312, 309t–311t PPAG, 585t

Index PR, 584t Precautionary principle, in carcinogen regulation, 1344–1345 Preexisting conditions, as risk factors, 23 Pregnancy abnormalities of, testicular cancer and, 1156–1157 benign breast disease and, 31 biliary tract cancer and, 793–794 breast cancer and, 998, 1002, 1006t, 1007 cervical adenocarcinoma in situ and, 34 cervical cancer and, 1051–1052 cigarette smoking during, testicular cancer and, 1159 endometrial cancer and, 1034 molar. See Hydatidiform mole non-Hodgkin lymphoma and, 909 ovarian cancer and, 1019, 1156 soft tissue sarcoma and, 969 thyroid cancer and, 981 Prevalence definition of, 102 global, 106–107, 108f international, 106–107, 108f lifetime, 102 Prevention. See Cancer control and prevention Primary effusion lymphoma, 523 Primary sclerosing cholangitis biliary tract cancer and, 793 liver cancer and, 776 Primary severe immune deficiency, 551–555, 551t Pro-vitamins, half-life of, 73t Probability, 4 Procarbazine hydrochloride, 491t Progeria, adult, 572 Progesterone breast cancer and, 1003 endometrial cancer and, 1030 obesity effects on, 429 ovarian cancer and, 1021 physical activity effects on, 450 Progesterone receptor in breast cancer, 996 in cervical adenocarcinoma in situ, 34 in ovarian cancer, 1020 Progestin, breast cancer and, 999 Progestogens, exogenous, endometrial cancer and, 1031–1033, 1032t Prolactin, breast cancer and, 1003 b-Propiolactone, 491t Proportional incidence ratios, in migrant studies, 191 Propylene oxide, occupational exposure to, 330t Propylthiouracil, 491t, 495–496 Prostaglandins, carcinogenesis and, 742 Prostate cancer, 1128–1142 abdominal adiposity and, 434–435 N-acetyltransferase-2 and, 1139 alcohol and, 251, 1135 anatomic distribution of, 1128 androgen receptor gene CSG repeat and, 1136–1137 androgens and, 1136–1137 antioxidants and, 1131–1132 arsenic and, 383t, 384, 385f aspirin and, 1135 balding and, 1137 birth weight and, 1135

body mass index and, 434–436, 435f, 1134–1135 cadmium and, 337, 1136 calcium and, 1134 carotenoids and, 1131 central adiposity and, 1135 chemoprevention of, 1140, 1327t, 1332 cigarette smoking and, 229, 1135 classification of, 16, 1128–1129 cruciferous vegetables and, 1131 CYP1A1 and, 1137 CYP3A4 and, 1137 CYP1B1 and, 1137 dairy products and, 1133–1134 diabetes mellitus and, 1138 diet and, 1130–1134, 1142 DNA repair genes and, 1139 early detection of, 1142 economic burden of, 210, 212, 213 electric appliance exposures and, 313 energy intake and, 1134 familial relative risk in, 564t family history of, 1136 fat intake and, 414, 1133 finasteride and, 1140 fish intake and, 1133 fruits and vegetables and, 1130–1131 Gleason grading of, 16 glucose metabolism and, 1138 glutathione S-transferases and, 1139 growth factors and, 1138 GSTP1 and, 1139 height and, 1138 histopathology of, 1128–1129 HSD3B1 and, 1137 HSD3B2 and, 1137 incidence of, 1294 age and, 165f, 1129–1130 international, 106, 107f, 108f, 121, 123–124, 123f, 1130 migrant studies of, 193–194, 1130 PSA screening and, 124 race and, 1130, 1142 socioeconomic status and, 176t, 177, 1130 time trends in, 123–124, 125f U.S., 121, 123, 124f, 157–158, 157f, 158t, 1129 insulin and, 435–436, 1138 insulin-like growth factor I and, 436, 1138 insulin-like growth factor binding protein 3 and, 1138 interleukin-6 and, 1139 ionizing radiation and, 261f latent, 123 legumes and, 1132 leptin and, 435, 1138 linoleic acid and, 1133 a-linolenic acid and, 1133 lycopene and, 1130–1131 meat intake and, 1133 migrant studies of, 123, 193–194, 1130 molecular genetics of, 1129, 1136–1137 mortality from international, 106, 107f, 121, 123f, 124, 125f, 1130 migrant studies of, 194 U.S., 124, 125f, 141t, 157f, 158, 158t, 1129 nonsteroidal anti-inflammatory drugs and, 1135

1385 obesity and, 434–436, 435f, 438t, 1134–1135 occupation and, 335t, 1136 ornithine decarboxylase and, 1139 pathogenesis of, 435–436, 1139–1140, 1140f physical activity and, 458, 459t, 460t, 1135 phytoestrogens and, 1132 polyamines and, 1139 precursor lesions of, 1129 pregnancy factors and, 1135 prevention of, 1140–1141, 1295, 1327t, 1332 prostatitis and, 1138–1139, 1142 PTEN and, 1139 retinol and, 1131 screening for, 124, 1141 selenium and, 1132, 1140 sexual activity and, 1137 sexually transmitted infections and, 1138 socioeconomic status and, 176t, 177, 1130 solar radiation and, 301, 301t soy and, 1132 SRD5A2 and, 1137 staging of, 1128 survival for, 158, 172, 1129 susceptibility genes in, 584t, 585t, 586t, 587t, 588–589 tea and, 1132–1133 testosterone and, 436 tomatoes and, 1130–1131 treatment of, 1142 tumor necrosis factor-a and, 1139 twin study of, 91t, 93 vasectomy and, 1135–1136 vitamin A and, 1131 vitamin C and, 1131 vitamin D and, 1134 vitamin D receptor gene polymorphisms and, 1137–1138 vitamin E and, 415, 1131–1132, 1140 weight gain and, 434–435 zinc and, 1132 Prostate gland cancer precursors of, 1129 transurethral resection of, 158 Prostate-specific antigen (PSA) test, 1141 Prostatic intraepithelial neoplasia, 1129 Prostatitis, 1138–1139, 1142 Protein, dietary brain cancer and, 1185 colorectal adenoma and, 29 colorectal cancer and, 817 esophageal cancer and, 700 intake of, 411 laryngeal cancer and, 631 pancreatic cancer and, 740 pharyngeal cancer and, 681 renal cancer and, 1092 stomach cancer and, 712, 712t Proteomic analysis, 79 of early biologic effect biomarkers, 74–75 in ovarian cancer, 1021 Proto-oncogenes, 47–48 Proton pump inhibitors, stomach cancer and, 713 Psoralens, 496 Psoriasis, 554 coal tar in, 496 oral cavity cancer and, 683 PUVA therapy in melanoma and, 1208

1386 Psoriasis (Continued) penile cancer and, 1168 skin cancer and, 1237, 1238t Psychosocial factors, socioeconomic statuscancer association and, 180, 183 PTC, in thyroid cancer, 987 PTCH, 566, 1243 PTEN, 568, 587t, 1003, 1139 Puberty, age at, testicular cancer and, 1156 PUVA therapy, 299 melanoma and, 1208 penile cancer and, 1168 skin cancer and, 1237, 1238t

Quality-adjusted life-years (QALY), definition of, 102 Quality of life, 204 Queyrat, erythroplasia of, 1166

Race anal cancer and, 831, 831f bladder cancer and, 1101, 1104t, 1105f brain cancer and, 1179, 1180t breast cancer and, 997 cervical cancer and, 1047–1048, 1048f, 1048t, 1049, 1049t choriocarcinoma and, 1078, 1079t, 1082 colorectal cancer and, 813, 813t, 814t endometrial cancer and, 1028–1029, 1028t Hodgkin lymphoma and, 874, 875 hydatidiform mole and, 1078–1079, 1079t, 1082 leukemia and, 844, 847f, 848t, 854 lung cancer and, 639–640, 640f melanoma and, 1202–1203, 1202f, 1203f multiple myeloma and, 919, 921, 921f nasopharyngeal cancer and, 620–621, 621f, 621t non-Hodgkin lymphoma and, 901, 902t, 903f oral cavity cancer and, 675, 675t ovarian cancer and, 1014, 1014f pancreatic cancer and, 722 penile cancer and, 1166 pharyngeal cancer and, 675, 675t, 676t prostate cancer and, 1130, 1142 skin cancer and, 1233, 1234t socioeconomic status-cancer association and, 175, 184 stomach cancer and, 708–709, 709f thyroid cancer and, 979–980, 980t RAD51, 584t Radiation, 280–281, 281f ionizing. See Ionizing radiation nonionizing, 281, 306–319. See also Extremely low-frequency electric and magnetic fields; Radiofrequency radiation regulation of, 1347–1348, 1347f, 1347t relative biological effectiveness of, 281 solar. See Solar radiation Radiofrequency radiation, 316–318 exposure assessment for, 316 lymphoma and, 306 mobile telephones and, 317–318, 317t occupational exposure to, 318, 318t residential exposure to, 316–317 wavelengths of, 306 Radiologists, cancer in, 271–272

Index Radionuclides, 269–271. See also specific radionuclides in ambient air, 357t, 359 in drinking water, 392–393 Radiotherapy. See Ionizing radiation Radium bone cancer and, 269–270, 271, 948, 949t, 950 breast cancer and, 270 in drinking water, 392–393 leukemia and, 265, 271 occupational exposure to, 271, 338, 612, 613t renal cancer and, 1093 sinonasal cancer and, 612 therapeutic, 269–270 Radium dial painting, 259, 271, 612, 613t Radon, 259 acute lymphocytic leukemia and, 855 acute myeloid leukemia and, 848 in drinking water, 392 indoor, 275–276, 373–374 lung cancer and, 274–276, 275f, 276f, 644 occupational exposure to, 274–275, 275f, 333t, 338, 339t, 373 risk assessment for, 374 underground, 274–275, 275f Raloxifene, 473 endometrial cancer and, 1031 Randomization, 4 RAS, 50 in thyroid cancer, 982, 987 Rate age standardization of, 103 definition of, 102 RB, 48, 48f RB1, 571 RB protein, 48, 49f, 52 RBI, 953 Receptor tyrosine kinases, overexpression of, 55–56 RECQL4, 571–572 Rectal cancer. See also Colorectal cancer body mass index and, 425 chlorination byproducts and, 387–388 cigarette smoking and, 227 disinfection by-products and, 388 economic burden of, 212 incidence of, 146t, 147–148, 165f, 194–195 ionizing radiation and, 260t, 261f migrant studies of, 194–195 mortality from, 148, 149t, 195 nitrate and, 389, 390t obesity and, 425 physical activity and, 452t–454t, 453t–455t, 455 postmenopausal hormone therapy and, 482–483, 483t Rectum, cancer precursors of, 27–29, 28f Red blood cells for biomarkers, 82, 82t in diet studies, 412 Relative risk, 93 Relative survival, definition of, 102 Religion cervical cancer and, 1051 pancreatic cancer and, 724–725 thyroid cancer and, 980 Renal cancer, 1087–1096 acetaminophen and, 1091

alcohol and, 1092 analgesics and, 1091 arsenic and, 383, 383t, 385f asbestos and, 1092–1093 aspirin and, 1091 body mass index and, 426–427, 427f chlorination byproducts and, 388 cigarette smoking and, 219t, 228, 1089–1090 classification of, 11t coffee and, 1092 in coke-oven workers, 1093 cystic disease and, 1094 developmental defects and, 1094 dialysis and, 1094 diet and, 1091–1092 disinfection by-products and, 388 diuretics and, 497, 1090–1091 estrogen and, 1092 familial relative risk in, 564t genetic susceptibility to, 1093–1094 GSTM1 polymorphisms and, 1093 hereditary, 563t, 571, 1094 hypertension and, 1090–1091 incidence of age and, 165f international, 106, 107f, 108f, 1087, 1089t U.S., 159–160, 159t, 1087, 1088f, 1088t ionizing radiation and, 260t, 261f, 1093 kidney transplantation and, 1094 mortality from international, 106, 107f U.S., 141t, 159–160, 159t, 1087–1088, 1089f, 1089t obesity and, 426–427, 427f, 438t, 1090 occupation and, 334t, 335t, 1092–1093 in oil refinery workers, 1093 oral contraceptives and, 1092 papillary, hereditary, 571 parity and, 1092 perchloroethylene and, 1093 phenacetin and, 1091 radium-224 and, 1093 risk factors for, 1089–1094, 1090t in slow acetylators, 1090 socioeconomic status and, 1088–1089 survival for, 160, 172, 1087–1088, 1089t susceptibility genes in, 586t tea and, 1092 trichloroethylene and, 1093 in von Hippel-Lindau disease, 1094 Renal failure, immune impairment in, 554 Renal pelvis cancer, 1094–1096, 1094t alcohol and, 1095 analgesics and, 1095 arsenic and, 1095 cigarette smoking and, 228, 1094–1095 coffee and, 1095 hypertension and, 1095 incidence of, 159–160, 159t, 165f, 1087, 1088f, 1088t ionizing radiation and, 1096 laxatives and, 1095 mortality from, 159–160, 159t, 1088 occupation and, 1095–1096 survival for, 172, 1087–1088, 1089f tea and, 1095 Renal transplantation penile cancer and, 1168 renal cancer and, 1094 Reserpine, 497

Index RET, 585t MEN2 and, 570 multiple cancers and, 1271 thyroid cancer and, 13, 987 Retinoblastoma, 563t, 571, 1259–1260 incidence of, 160 Knudson’s two-hit hypothesis of, 48, 48f osteosarcoma and, 267, 951, 952–953 radiotherapy in, 267 13-cis-Retinoic acid lung cancer and, 685 in oral cavity cancer, 686 Retinol esophageal cancer and, 700 pancreatic cancer and, 740–741 prostate cancer and, 1131 skin cancer and, 1240–1241, 1243 in skin cancer prevention, 1333 Retinyl palmitate, in oral cavity cancer, 686 Retroperitoneal cancer, 147t, 149t Retroviruses, 527–531. See also Human immunodeficiency virus (HIV) infection; Human T-cell leukemia virus type I (HTLV-I) infection non-Hodgkin lymphoma and, 903, 905 Reverse smoking, 221 Rhabdomyosarcoma, 17t, 959, 960t. See also Soft tissue sarcoma childhood, 1258 hormonal factors in, 969 Rheumatoid arthritis, 554, 910–911 Ribavirin, in hepatitis C virus infection, 520 Riboflavin, esophageal cancer and, 700 Ringworm, radiotherapy in, 263 Risk, 1303–1315 acceptable, in carcinogen regulation, 1344 communication of, 1303–1304 bar graphs in, 1306 comparative information in, 1305–1306 competing risks in, 1306 decision aids in, 1306–1307 numeracy and, 1304–1305 numerical formats in, 1305, 1305t qualitative formats in, 1305 risk ladders in, 1306 visual formats in, 1306 comprehension of, 1304, 1307 cumulative, 103 definition of, 102, 1303 perception of, 1304 RNA tumor viruses, 47 Rothmund-Thomson syndrome, 563t, 571–572, 953, 955 Rous sarcoma virus, 47 Rubber industry, 336t bladder cancer and, 1108 chronic lymphocytic leukemia and, 857 multiple myeloma and, 938

Saccharin, bladder cancer and, 1111 Safrole, 491t SAGE (serial analysis of gene expression), 13 Salivary gland cancer, 687–689 alcohol and, 688 cigarette smoking and, 688 classification of, 11t, 687 diet and, 688 histopathology of, 687 hormonal factors in, 688

incidence of, 143, 144t, 668f, 675t, 687 ionizing radiation and, 260t, 261f, 263, 668 mortality from, 145t occupation and, 688 radiation exposure and, 688 survival for, 143, 172, 688 ultraviolet light exposure and, 688 viral infection and, 688 Salmonella typhi infection, biliary tract cancer and, 793 Salt gastric cancer and, 415 stomach cancer and, 712, 712t Sarcoma bone. See Osteosarcoma classification of, 12t, 17, 17t cytogenetics of, 13, 17, 17t economic burden of, 211 epithelioid, 17 Ewing. See Ewing sarcoma melanoma and, 1213t, 1214 occupation and, 334t, 335t small intestine, 802 soft tissue. See Soft tissue sarcoma spindle cell, 17 synovial, 14f, 17t Schistosoma haematobium, 508t, 534 bladder cancer and, 534, 1113–1114, 1117 in drinking water, 393 Schistosoma japonicum colorectal cancer and, 534 liver cancer and, 534, 772 Schistosoma mansoni, 534 Scoliosis, ionizing radiation in, 262 Screening, 21–22, 37, 38t, 1310–1315 anal cancer, 838 assessment of, 1310–1311 biases in, 1311 bladder cancer, 1117 BRCA gene, 565 breast cancer, 114, 184, 1005, 1007, 1285 case studies of, 1312–1315 CDKN2A, 565 cervical cancer, 21–22, 31–34, 38t, 1045–1046, 1055, 1058–1059, 1060 colorectal cancer, 21–22, 38t, 823–824, 1312–1314 endometrial hyperplasia, 35 harms of, 1312 hepatocellular carcinoma, 777 hydatidiform mole, 1083 laryngeal cancer, 634 levels of evidence for, 1311–1312 liver cancer, 777 lung cancer, 38t, 650, 1314–1315 migrant studies and, 190 neuroblastoma, 1315 observational studies of, 1311 oral cavity cancer, 38t, 686 ovarian cancer, 1022, 1022t pharyngeal cancer, 686 prostate cancer, 124, 1141 randomized trial studies of, 1310–1311 skin cancer, 38t socioeconomic status-cancer association and, 183–184 stomach cancer, 38t, 715–716 test validity and, 1310 vaginal cancer, 1072 vulvar cancer, 1072

1387 Scurvy, 3 SDHD, 570 Seafood prostate cancer and, 1133 salted nasopharyngeal cancer and, 415, 621, 623 sinonasal cancer and, 614 thyroid cancer and, 986 Second-hand smoke. See Environmental tobacco smoke Securin, 55, 55f SEER program. See Surveillance, Epidemiology, and End Results (SEER) program Segregation analysis, 97 Selective estrogen receptor modulators breast cancer and, 472–473 endometrial cancer and, 1031 Selective serotonin reuptake inhibitors, 493 Selenium, 415 arsenic interaction with, 386 bladder cancer and, 1112 breast cancer and, 415 colorectal cancer and, 415, 823 esophageal cancer and, 700 hepatocellular carcinoma and, 772 lung cancer and, 415, 643, 650 prostate cancer and, 1132, 1140 skin cancer and, 1240, 1333 stomach cancer and, 715 thyroid cancer and, 986 Seminoma. See Testicular cancer Semipalatinsk Test Site, 277–278 Separase, 54–55 Septins, 55 Serial analysis of gene expression, 13 Serum, for biomarkers, 82, 82t SES. See Socioeconomic status (SES) Seventh-Day Adventists, 406, 724–725 Sex. See Gender Sex cord-stromal tumor, 1151, 1153f. See also Testicular cancer Sex hormone-binding globulin alcohol effects on, 243 physical activity effects on, 450 Sex workers, cervical cancer in, 1052, 1054 Sexual behavior anal cancer and, 833 cervical cancer and, 1051 prostate cancer and, 1137 socioeconomic status-cancer association and, 182–183 Sexually transmitted infections anal cancer and, 833–835 cervical cancer and, 1052. See also Cervical cancer, human papillomavirus in choriocarcinoma and, 1080 prostate cancer and, 1138 socioeconomic status-cancer association and, 182–183 Shale oil, occupational exposure to, 327t Shift work, socioeconomic status-cancer association and, 183 Shingles, multiple myeloma and, 929 Shoe industry, sinonasal cancer and, 609 Sibship size, Hodgkin lymphoma and, 877, 878t Sidestream smoke. See Environmental tobacco smoke Sievert (Sv), 281

1388 Sigmoidoscopy, 823, 1313 Silica, occupational exposure to, 326t, 333t, 336t, 746 Simian virus 40 mesothelioma and, 667–668, 668t non-Hodgkin lymphoma and, 906 Single nucleotide polymorphisms, 70, 75–77 association studies of, 77 haplotype-tagging, 76–77 in population study, 579, 582–583 sample size and, 77–78 Sinonasal cancer, 603–616 alcohol and, 614 anatomic distribution of, 603, 604f, 604t chromium and, 610–611, 613t cigarette smoking and, 225–226, 613, 615 classification of, 603 demographics of, 604–606, 604t diet and, 614 environmental factors in, 606, 608–614, 613t Epstein-Barr virus and, 614 estrogenic hormones and, 615 formaldehyde and, 611–612, 613t hide tanning and, 609 histopathology of, 603, 604t host factors in, 614–615 human papillomavirus and, 603, 614 incidence of, 152t, 604, 605t, 606, 607t infectious agents and, 614 leather dust and, 609, 613t mineral oils and, 611, 613t molecular genetics of, 603 mortality from, 152t nasal medications and, 614 nickel and, 339t, 609–610, 613t nose bleeds and, 615 occupation and, 334t, 335t, 608–613, 613t p53 in, 603 passive smoking and, 613 pathogenesis of, 615 polyps and, 615 precursor lesions in, 603 prevention of, 615 radium dial painting and, 612, 613t shoe industry and, 609, 613t sinusitis and, 614–615 snuff and, 613–614 survival for, 604, 606, 606t, 607f textile dusts and, 612, 613t Thorotrast and, 614 time trends in, 604, 605f, 606t tobacco and, 613–614, 615 trauma and, 615 wood dust and, 608–609, 613t Sinusitis, 614–615 Sister chromatid exchange assay, 75 Sjögren syndrome, 554, 911 Skin cancer precursors of, 23, 1230 DNA repair in, 299 p53-containing keratinocytes of, 23 Skin cancer, 11t, 154, 1230–1244. See also Melanoma (cutaneous) anatomic sites of, 1232–1233, 1233fd, 1234f arsenic and, 383, 383t, 384, 385f, 1239 in atomic bomb survivors, 269 beta-carotene and, 414 chemical exposure and, 1239 chemoprevention of, 1327t–1328t, 1333

Index cigarette smoking and, 229–230, 1239–1240, 1240t data sources on, 1230–1231 diet and, 1240–1241 freckling and, 1241–1242 genetic susceptibility to, 1242 hair color and, 1242 human immunodeficiency virus infection and, 1238 human papillomavirus infection and, 1238–1239 immunosuppression and, 500, 1238 incidence of, 1230–1235 age and, 1231, 1232f data sources on, 1230–1231 gender and, 1231, 1232f international, 1230–1231, 1231t, 1233 race and, 1233, 1234t trends in, 1233–1235 U.S., 1230–1231, 1231t ionizing radiation and, 260t, 261f, 263, 1239 MCIR and, 1242 5-methoxypsoralen and, 496 molecular genetics of, 1242–1243 mortality from, 1235 multiple cancers and, 1275 nevi and, 1241–1242 non-Hodgkin lymphoma and, 910 occupation and, 334t, 1239 photosensitizing agents and, 1237–1238 pigmentation and, 1241–1242 post-transplantation, 552 precursor lesions of, 23, 1230 prevention of, 301–302, 1216, 1216t, 1243, 1276, 1327t–1328t, 1333 ptch and, 1243 PUVA therapy and, 1237, 1238t screening for, 38t skin color and, 1241, 1242 skin type and, 1241 solar radiation and, 300–301, 1235–1237, 1236t, 1237t, 1241 susceptibility genes in, 584t, 586t, 587t tanning lamps and, 1237 TP53 in, 1242–1243 trauma and, 1241 in xeroderma pigmentosum, 573, 1242 Skin-fold thickness, 412, 423 Smac/DIABLO, in apoptosis, 57f, 58 Small intestine cancer, 801–806. See also Colorectal cancer acromegaly and, 804 anatomic distribution of, 801 bile acids and, 806 carcinoid, 801 celiac disease and, 804–805, 806 cigarette smoking and, 805 classification of, 801–802 Crohn disease and, 804, 806 diet and, 805, 806 environmental factors in, 805 familial adenomatous polyposis and, 802 histopathology of, 801 host factors in, 802–805 incidence of, 146t, 802, 803f, 804t, 805 inflammatory bowel disease and, 804–805 large intestine cancer and, 803–804 molecular genetics of, 802 mortality from, 149t pathogenesis of, 805–806

Peutz-Jeghers syndrome and, 802–803 precursor lesions in, 802 prevention of, 806 tobacco and, 805 ulcerative colitis and, 804 Smelting, sinonasal cancer and, 610 Smokeless tobacco. See Tobacco, smokeless SNP500Cancer project, 77 Snuff, 221, 233 bladder cancer and, 1107 oral cavity cancer and, 680 processing of, 224 sinonasal cancer and, 613–614 Social support, socioeconomic status-cancer association and, 183 Socioeconomic status (SES), 174–184 brain cancer and, 1179, 1180t breast cancer and, 176–177, 176t, 178, 183, 997 cancer association with, 178–184 adult exposure in, 181–184 alcohol in, 182 childhood infection in, 181 diet in, 182 drift hypothesis of, 179–180 early-life exposure in, 180–181 health care access in, 183–184 income and, 180 instrumental variable analysis of, 180 lifestyle behaviors in, 181 neighborhood environments in, 184 obesity in, 182 occupational exposures in, 183 psychosocial factors in, 180, 183 reproductive factors in, 183 reverse causation and, 179–180 sexual behavior in, 182–183 smoking in, 181–182 third variable bias in, 180 cancer incidence and, 175–177, 176t cancer mortality and, 177–178, 177t, 178t cervical cancer and, 176t, 182–183, 1049, 1051 childhood leukemia and, 1255 choriocarcinoma and, 1079 colorectal cancer and, 176t, 177, 184, 813–814 definition of, 174 education and, 174, 180 esophageal cancer and, 702 Hodgkin lymphoma and, 875–876, 877, 878 hydatidiform mole and, 1079 income and, 174 lip cancer and, 686 lung cancer and, 176t, 177t, 181, 641 measurement of, 175 melanoma and, 1203, 1203f multiple myeloma and, 922 occupation and, 174 ovarian cancer and, 1014 pancreatic cancer and, 724 penile cancer and, 1166 prostate cancer and, 176t, 177, 1130 race and, 175, 184 renal cancer and, 1088–1089 stomach cancer and, 181, 713 survival and, 178 testicular cancer and, 1153–1154 thyroid cancer and, 980

Index Socioeconomic status (SES) gradient, 175, 176–177, 178, 179f Socioeconomic status (SES) index, 178, 179f Soderberg electrolytic reduction process, bladder cancer and, 1109 Sodium nitrate, brain cancer and, 1185 Sodium ortho-phenylphenate, occupational exposure to, 332t Soft tissue sarcoma, 959–970 age and, 961, 961t anatomic distribution of, 959 benign neoplasms vs., 960 body mass index and, 969 cancer associations of, 968–969 in children, 959, 960t, 969, 1258 chlorophenols and, 966, 967 chromosomal abnormalities and, 961 classification of, 12t, 959–960, 960t diet and, 968 dioxin and, 962, 963t–965t, 966–967 DNA repair in, 970 economic burden of, 209 familial relative risk in, 564t formaldehyde and, 968 gender and, 969 genetic factors in, 969 herbicides and, 962, 963t–965t, 966–967 histopathology of, 959 hormones and, 968, 969 host factors and, 968–970 immunosuppression and, 969 incidence of, 153, 153t, 961, 961t, 962t international trends in, 961, 962t ionizing radiation and, 966t, 967–968 in Li-Fraumeni syndrome, 969 molecular genetics of, 961 mortality from, 141t, 153, 153t, 960t, 961 in neurofibromatosis type I, 969 occupation and, 962–967, 963t–966t oral contraceptives and, 968 pathogenesis of, 970 pesticides and, 962, 963t–965t, 966–967 prevention of, 970 radiotherapy and, 966t, 967–968 second cancers and, 968–969 smokeless tobacco and, 968 survival for, 153, 172 time trends in, 961 vinyl chloride and, 338, 965t, 967 viral infection and, 966t, 968, 969–970 wood-related materials and, 965t Solar irradiance, 294 Solar radiation, 294–302 breast cancer and, 301, 301t colon cancer and, 301, 301t DNA damage with, 299 dose of, 294 Hodgkin lymphoma and, 887 human papillomavirus interaction with, 300 immunosuppression and, 300 lip cancer and, 686 measurement of, 294, 295–299, 295f–296f, 295t, 297t–298t ambient exposure and, 298 Bacillus biological dosimeter for, 295–299, 297t–298t biological markers in, 299 in children, 296 mutation in, 299 personal recollection for, 298

polysulfone badge dosimeter for, 295–299, 297t PRIMENet program of, 295, 295f, 295t reliability of, 298, 298t silicone-rubber skin casts in, 299 melanoma and, 197, 301, 1205–1208, 1220–1221 multiple cancers and, 1275–1276 non-Hodgkin lymphoma and, 133, 301, 301t, 907, 1275–1276 occupational exposure to, 326t, 336t ocular melanoma and, 1222 ovarian cancer and, 301, 301t point mutations with, 299 prostate cancer and, 301, 301t protection from, 301–302, 1216, 1216t radiometer for, 294 reduction in, 301–302 skin cancer and, 300–301, 1235–1237, 1236t, 1237t, 1241. See also Skin cancer virus interaction with, 300 Solvents acute lymphocytic leukemia and, 855–856 multiple myeloma and, 938 non-Hodgkin lymphoma and, 908 occupational exposure to, 333t, 336t pancreatic cancer and, 746 Soot, occupational exposure to, 327t Soy breast cancer and, 1000 prostate cancer and, 1132 Specificity, causation and, 5t, 6 Spine radiotherapy to, 262 tumors of, 1187 Spironolactone, thyroid cancer and, 987 Spleen, cancer of, 12t Splenectomy, 553 Sputum cytology, in lung cancer, 650, 1314 SRCC3, melanoma and, 1215 SRD5A2, 585t, 1137 Standardized incidence rate, 103 Standardized incidence ratio, 190 Standardized mortality rate, 103 Standardized mortality ratio, 190 Statins, 494 in melanoma prevention, 1217 Steatohepatitis, nonalcoholic, hepatocellular carcinoma and, 776 Stomach precursor lesions of, 24, 708 surgery on biliary tract cancer and, 793 pancreatic cancer and, 749 stomach cancer after, 714 Stomach cancer, 707–716 alcohol and, 248, 712 anatomic distribution of, 707 candidate genes in, 714 chemoprevention of, 715, 1326t, 1329 cigarette smoking and, 219t, 227, 712–713 cimetidine and, 499 classification of, 11t, 707 diet and, 196, 712, 712t early detection of, 715–716 Epstein-Barr virus infection and, 713 familial relative risk in, 564t fetal sulphoglycoprotein antigen test for, 716 future research in, 716, 716t gastroesophageal reflux and, 715

1389 gastroscopy for, 715–716 genetic factors in, 714 grading of, 707–708 Helicobacter pylori infection and, 196, 532–533, 710–712, 714, 715 hereditary, 563t, 568–569, 714 histamine antagonists and, 713 histology of, 707, 708t, 713 incidence of age and, 165f international, 106, 107f, 108f, 116, 118–119, 118f, 709, 710f migrant studies of, 118, 195, 709–710 race and, 708–709, 709f time trends in, 119 U.S., 146t, 147, 148f, 708–709, 709f ionizing radiation and, 260t, 261f, 265, 713 laxatives and, 500 medical conditions and, 714–715 migrant studies of, 118, 195–196, 709–710 molecular markers in, 708 mortality from international, 106, 107f, 709 migrant studies of, 195–196 U.S., 141t, 148f, 149t, 709, 709f nitrate and, 389, 390t obesity and, 425–426, 714 occupation and, 334t, 335t postresection, 714 precursors of, 24, 708 prevention of, 715–716, 1293, 1326t, 1329 proton pump inhibitors and, 713 salt and, 415 screening for, 38t, 715–716 socioeconomic status and, 181, 713 staging of, 708 survival for, 710 susceptibility genes in, 584t, 585t, 586t, 587t twin study of, 91t Stone quarrying, pancreatic cancer and, 746 Stool antigen tests, for Helicobacter pylori, 532 Streptozotocin, 491t Stress, socioeconomic status-cancer association and, 183 Strongyloidiasis, 530 adult T-cell leukemia/lymphoma and, 530 Styrene, 330t, 337 chronic lymphocytic leukemia and, 857 Styrene-7,8-oxide, 328t Subglottic cancer. See Laryngeal cancer Succinate dehydrogenase, mutations in, 58 Sulfur dioxide, ambient air, 358–359 Sulfuric acid, occupational exposure to, 327t, 333t SULT1A1, 583, 586t Sunglasses, 1222 Sunlight. See Solar radiation Sunscreen, 1243 Supraglottic cancer. See Laryngeal cancer Surveillance, Epidemiology, and End Results (SEER) program, 139–142, 140t, 141t–142t, 143t, 1288 assessment of, 140–142, 141t–142t five-year relative survival rates (1992–1999) from, 170–171 incidence rate in, 140 Internet homepage for, 171 mortality rate in, 140, 141t–142t population for, 139, 140t

1390 Surveillance, Epidemiology, and End Results (SEER) program (Continued) short-term incidence trends (1992–2000) from, 168–170 short-term mortality trends (1992–2000) from, 170–171 survival rate in, 140 Survival. See also at specific cancers definition of, 102 SEER data on, 140, 172–173 socioeconomic status and, 178 SV40 virus, CNS cancer and, 1256 Sweden, migrants to, prostate cancer in, 194 Sweeteners, artificial, bladder cancer and, 1111 Synovial sarcoma, 960t Syntomycin, leukemia and, 850, 856 Syphilis, anal cancer and, 834–835 Systematized Nomenclature of Medicine, 10, 12 Systematized Nomenclature of Pathology, 10 Systemic lupus erythematosus, 554, 911, 928

Talc occupational exposure to, 326t ovarian cancer and, 1018 Tamoxifen in breast cancer prevention, 473, 1007, 1285, 1331 endometrial cancer and, 1031, 1274–1275 Tanning lamps, 299 Tannins, esophageal cancer and, 700 T3801C, 590t Tea esophageal cancer and, 700 liver cancer and, 772 pancreatic cancer and, 741 prostate cancer and, 1132–1133 renal cancer and, 1092 renal pelvis cancer and, 1095 skin cancer and, 1241 Techa River nuclear installation, 259, 273–274, 280, 393 Teeth, disease of, oral cavity cancer and, 683 Telomerase, 56 Telomeres, 56 Temporality, causation and, 5t, 6 Teniposide, 491t, 495 Teratoma, 1151, 1153f, 1153t, 1162t. See also Testicular cancer Terminal cancer, economic burden of, 213 Testes, development of, 1159 Testicular cancer, 1151–1162 age at puberty and, 1156 Agent Orange and, 1155 alcohol and, 1159 alpha-fetoprotein in, 1162t chemotherapy for, acute myeloid leukemia and, 850 childhood, 1259 chromosomal abnormalities and, 1157 cigarette smoking and, 229, 1159 classification of, 11t, 1151–1152 cryptorchidism and, 1159–1160, 1161t DDE and, 1155 diet and, 1158 diethylstilbestrol and, 1156

Index electric appliance exposures and, 313 familial, 1157–1158 familial relative risk in, 564t genetic factors in, 1157–1158, 1157t gonadal dysgenesis and, 1157–1158, 1157t, 1159–1160 herbicides and, 1155 histopathology of, 1151–1152, 1153t hormones and, 1156 human chorionic gonadotropin in, 1162t incidence of international, 102f, 1154, 1155t migrant studies of, 1154 socioeconomic status and, 1153–1154 U.S., 158, 158t, 1151, 1152–1153, 1152f, 1152t, 1153f, 1153t, 1154t ionizing radiation and, 261f maternal estrogen and, 1156–1157 maternal-fetal interactions and, 1156–1157 maternal smoking and, 1159 migrant studies of, 1154 mortality from, 102f, 141t, 158, 158t obesity and, 1158 occupation and, 335t, 1155 ochratoxin A and, 1156 pathogenesis of, 1159–1160 physical activity and, 1158–1159 prenatal exposures and, 1156 prevention of, 1160 radiofrequency radiation and, 318 radiotherapy for, acute myeloid leukemia and, 848 in situ, 1151–1152 socioeconomic status and, 1153–1154 survival for, 158, 172 susceptibility genes in, 586t trauma and, 1158 tumor markers in, 1162t vasectomy and, 1158 viral infection and, 1158 Testosterone breast cancer and, 1003 physical activity effects on, 450 prostate cancer and, 436, 1136 Tetanitromethane, occupational exposure to, 331t 2,3,7,8-Tetrachlorodibenzo-p-dioxin myeloid leukemia and, 849 occupational exposure to, 327t Tetrachloroethylene, 392 non-Hodgkin lymphoma and, 908 occupational exposure to, 328t Textile industry oral cavity and, 683 pancreatic cancer and, 746 sinonasal cancer and, 612, 613t TGFB, 584t 4,4¢-Thiodianiline, occupational exposure to, 331t Thiotepa, 491t Thiouracil, 491t, 495–496 Thiourea, 332t, 491t Thorium, pancreatic cancer and, 744 Thorotrast, 259, 271 biliary tract cancer and, 794 liver cancer and, 271, 773 multiple myeloma and, 933t, 934, 935 osteosarcoma and, 950, 951f sinonasal cancer and, 614 Three Mile Island nuclear reactor, 279

Thymectomy, 553 Thymus enlargement of, radiotherapy in, 262–263 involution of, 551 Thyroid cancer, 975–988 alcohol and, 987 anaplastic, 975, 976t, 987–988 antithyroid drugs and, 495–496 in atomic bomb survivors, 268–269 body mass index and, 436 breast cancer and, 982–983 chemical exposures and, 987 Chernobyl reactor accident and, 259, 979, 984–985 chromosomal abnormalities and, 987–988 cigarette smoking and, 230, 986–987 classification of, 11t, 975 coffee and, 987 diagnosis of, 975–976 diet and, 985–987 familial, 563t, 983 familial relative risk in, 564t fine needle aspiration in, 976 follicular, 13, 975, 976t fruits and vegetables and, 986 goiter and, 198, 981–982 grains and, 986 histopathology of, 975, 976t hormones and, 981 host factors in, 981–983 incidence of age and, 979 gender and, 979, 979f, 980f, 980t, 981 international, 976–977, 977f, 977t, 978 migrant studies of, 198 race and, 979–980, 980t religion and, 980 time trends in, 977–979 U.S., 160, 161t, 976, 977, 978f, 980f infertility and, 981 iodine-131 and, 270, 280, 984–985 iodine levels and, 985–986 ionizing radiation and, 260t, 261f, 262–264, 268–269, 978–979, 983–985, 984f in Marshall Islanders, 984 medications and, 987 medullary, 23, 570, 975, 976t, 983, 988 migrant studies of, 198 molecular genetics of, 987–988 mortality from, 161t, 976, 978f, 979, 979f, 980f nuclear weapons tests and, 277, 278 obesity and, 436 occupation and, 334t, 985, 987 oral contraceptives and, 981 papillary, 13, 975, 976t, 983 pregnancy and, 981 prognosis for, 976 seafood and, 986 secondary cancer and, 983 socioeconomic status and, 980 survival for, 160, 172, 980–981 susceptibility genes in, 585t, 586t thyroid disease and, 981–982, 982t thyroid stimulating hormone and, 981 time trends in, 978f trace metals and, 986 treatment of, 270, 976 vitamins and, 986

Index Thyroid gland disease of, cancer and, 981–982, 982t nodularity of, ionizing radiation and, 276 Thyroid stimulating hormone, thyroid cancer and, 981 Thyroiditis, Hashimoto, 554 Thyrotoxicosis, radioiodine treatment in, 270 Time trend study, 106 age-period-cohort modeling in, 106 Tinea capitis, radiotherapy in, 263 Tissue for biomarker analysis, 83 for diet studies, 412 Tissue microarrays, 83 TNFalpha, 585t Tobacco, 217–234. See also Cigarette smoking animal studies of, 233 black, 224 blond, 224 chemical composition of, 222, 222t composition of, 233 erythroplakia and, 680 exposure measurement for, 224–225 historical perspective on, 217, 218f leukoplakia and, 680 metabolites of, 225, 233 nicotine yield of, 223, 223f, 224f potential reduced-exposure products of, 224 processing of, 224 smokeless, 221–222, 224, 233 bladder cancer and, 1107 laryngeal cancer and, 629 leukoplakia and, 25 oral cavity cancer and, 233, 680 sinonasal cancer and, 614 soft tissue sarcoma and, 968 Tobacco smoke, 222, 222t, 233. See also Environmental tobacco smoke Toluene, occupational exposure to, 328t Toluene diisocyanates, occupational exposure to, 330t ortho-Toluidine, occupational exposure to, 328t Tomatoes, prostate cancer and, 1130–1131 Tongue cancer, 144, 144t, 145t, 172. See also Oral cavity cancer Tonsillectomy, 553–554 Hodgkin lymphoma and, 554, 889 pancreatic cancer and, 750 Tonsils cancer of, 144t, 145t irradiation of, 263 Toombak, oral cavity cancer and, 680 Topoisomerase II inhibitors, acute myeloid leukemia and, 850 Toxaphene, occupational exposure to, 332t TP53, 50–51, 52, 53, 53f in apoptosis, 58 in glioblastoma, 16–17 mutation database for, 79 skin cancer and, 1242–1243 solar radiation effects on, 300 Trace metals. See also specific trace metals drinking water, 394 thyroid cancer and, 986 Tracheal cancer, 152t, 219t, 226 Transfer payments, costs vs., 205 Transforming growth factor-a, in hepatocellular carcinoma, 515 Transforming growth factor-b, in pancreatic cancer, 751

Transmission disequilibrium test, 95 Transmission lines, 308–312, 309t–311t Transplantation. See Organ transplantation Trauma acoustic neuroma and, 1187 brain cancer and, 1186–1187 sinonasal cancer and, 615 skin cancer and, 1241 testicular cancer and, 1158 Treosulphan, 491t Tretinoin, in melanoma prevention, 1217 Trichloracetic acid, urinary, 74 Trichloroethylene non-Hodgkin lymphoma and, 908 occupational exposure to, 328t renal cancer and, 1093 Trichothiodystrophy, 573 Trihalomethanes cancer and, 386–388, 388f metabolism of, 388 Trimustine hydrochloride, 491t Trisomy 8, myelodysplastic syndromes and, 844 Trisomy 21. See Down syndrome Tris(2,3-dibromopropyl) phosphate, occupational exposure to, 329t Truck drivers, bladder cancer in, 1109 Trypan blue, occupational exposure to, 331t TSC1, 572 Tubal ligation, ovarian cancer and, 1018, 1018t Tuberculosis cigarette smoking and, 232 isoniazid in, 493 X-ray fluoroscopic examinations in, 260 Tuberous sclerosis, 572, 1189 Tumor markers, 13–15, 14f, 37t, 72f, 73t, 78–80. See also specific genes definition of, 79 Tumor necrosis factor-a, prostate cancer and, 1139 Tumor suppressor genes, 47, 48, 566, 1319 brain cancer and, 1176 identification of, 96 inactivation of, 48, 48f point mutations in, 50 Tumorigenesis. See Carcinogenesis Turcot syndrome, 1189 Twin study, 89 in brain cancer, 1189 concordance rate in, 89 of familial aggregation, 91–92, 91t, 578 in Hodgkin lymphoma, 887 population-attributable fraction in, 93–94 of prostate cancer, 93 site-specific cancer in, 92 Twins, X-ray pelvimetry for, 266 Tylosis, esophageal cancer and, 702 Typhoid carrier state, biliary tract cancer and, 793 Tyrosine kinases, 55 Tyrosinemia, hepatocellular carcinoma and, 776

UGT1A1, 586t Ulcerative colitis anal cancer and, 836 biliary tract cancer and, 793 colorectal cancer and, 811, 820 small intestine cancer and, 804

1391 Ultrasonography, in testicular cancer, 1160 Ultraviolet index, 295 Ultraviolet radiation, 294, 295t. See also Solar radiation artificial, 299, 328t therapeutic, 299 Uncertainty, prevention and, 1297–1299 United Kingdom migrants to, 192, 193 cervical cancer in, 198 prostate cancer in, 194 nuclear installations in, 278–279 nuclear weapons tests in, 278 United States cancer incidence in, 139–164, 140t, 141t–142t, 143t. See also at specific cancer sites economic cancer burden of, 202–208, 210–214. See also Economic burden of cancer migrants to, 192, 193 breast cancer in, 192f, 193 cervical cancer in, 198 colon cancer in, 194 esophageal cancer in, 197 liver cancer in, 197 nasopharyngeal cancer in, 198 ovarian cancer in, 198 prostate cancer in, 194 rectal cancer in, 195 stomach cancer in, 195, 196 thyroid cancer in, 198 Uracil mustard, 491t Uranium, 259, 274 Urea breath test, for Helicobacter pylori, 532 Ureter cancer, 1094–1096, 1094t alcohol and, 1095 analgesics and, 1095 arsenic and, 1095 cigarette smoking and, 228, 1094–1095 coffee and, 1095 incidence of, 159t ionizing radiation and, 1096 mortality from, 159t occupation and, 1095–1096 Urethane, occupational exposure to, 330t Urinary stasis, bladder cancer and, 1113 Urinary tract infection, bladder cancer and, 1113 Urine arsenic in, 384, 386f collection of, for biomarkers, 83 mutagens in, 1114 nitrosamines in, 225 pH of, bladder cancer and, 1113 trichloracetic acid in, 74 Ursodeoxycholic acid, colorectal cancer and, 822 Uruguay, migrants to, 193 Utah, nuclear weapons tests in, 277 Uterine bleeding, radiotherapy in, 264–265 Uterus cancer of. See Endometrial cancer hyperplasia of, 1030, 1031–1032

Vaccine hepatitis B virus, 1331 hepatitis C virus, 1331 human papillomavirus, 1058, 1060, 1170, 1277, 1332–1333

1392 Vaccine (Continued) in melanoma prevention, 1333 polio, 906, 1187 Vaginal cancer, 1068–1073 alcohol and, 1071 anal cancer and, 1072 cervical cancer and, 1072 cigarette smoking and, 229, 1071 classification of, 11t, 1068 diethylstilbestrol and, 1071 family cancer history and, 1072 histopathology of, 1068 human immunodeficiency virus infection and, 1072 human papillomavirus infection and, 1070–1071 hysterectomy and, 1071–1072 incidence of, 156t, 1069, 1069f, 1069t, 1070f molecular genetics of, 1068–1069 mortality from, 156t, 1069, 1070t multiple cancers and, 1276 pathogenesis of, 1072 precursor lesions of, 1068 prevention of, 1072 screening for, 1072 survival for, 1069, 1070f Vaginal douching, cervical cancer and, 1052 Vaginal intraepithelial neoplasia, 1068 Varicella virus infection, testicular cancer and, 1158 Varicella-zoster infection, brain tumor and, 1187 Vascular endothelial growth factor, in angiogenesis, 59 Vasectomy prostate cancer and, 1135–1136 testicular cancer and, 1158 VDR, 585t, 1215 Vegetables. See Fruits and vegetables Venipuncture, 82 VHL, 572, 1271 Vinyl acetate, occupational exposure to, 330t Vinyl bromide, occupational exposure to, 328t Vinyl chloride, 327t, 336t, 338 angiosarcoma and, 338, 339t, 773 soft tissue sarcoma and, 338, 965t, 967 Vinyl fluoride, 328t Viral infection, 8, 507–531, 508t. See also specific viral infections bone cancer and, 952 brain tumors and, 1187 cervical cancer and. See Cervical cancer, human papillomavirus in Hodgkin lymphoma and, 508f, 511t, 512, 881–885, 881t laryngeal cancer and, 631–632 multiple myeloma and, 926t–927t, 930 non-Hodgkin lymphoma and, 903, 905–906 salivary gland cancer and, 688 sinonasal cancer and, 614 socioeconomic status-cancer association and, 180, 181 soft tissue sarcoma and, 966t, 968, 969–970 stomach cancer and, 713 testicular cancer and, 1158 Vitamin A, 414–415 alcohol effects on, 243 bladder cancer and, 1112 breast cancer and, 1000 laryngeal cancer and, 631

Index leukoplakia and, 25 multiple cancers and, 1273 prostate cancer and, 1131 skin cancer and, 1240–1241, 1333 Vitamin B, leukoplakia and, 25 Vitamin C, 415 alcohol effects on, 243 bladder cancer and, 1112 brain cancer and, 1185 breast cancer and, 1000 laryngeal cancer and, 631 leukoplakia and, 25 lung cancer and, 415 multiple cancers and, 1273 oral cavity cancer and, 682 ovarian cancer and, 1017 pancreatic cancer and, 740, 741 pharyngeal cancer and, 682 prostate cancer and, 415, 1131 renal cancer and, 1092 stomach cancer and, 712, 712t Vitamin D alcohol effects on, 243 prostate cancer and, 1134 skin cancer and, 1241 solar radiation effect on, 300 thyroid cancer and, 986 Vitamin D receptor melanoma and, 1215 in prostate cancer, 1137–1138 Vitamin E, 415 breast cancer and, 1000 esophageal cancer and, 700 laryngeal cancer and, 633 lung cancer and, 650 ovarian cancer and, 1017 prostate cancer and, 1131–1132, 1140 renal cancer and, 1092 skin cancer and, 1241 stomach cancer and, 715 Volatile organic chemicals, in drinking water, 391 Von Hippel-Lindau disease, 563t, 572 hemangioblastoma in, 1176, 1189 renal cancer in, 1094 Von Recklinghausen disease, 563t, 570, 851, 1189 Vulvar cancer, 1068–1073 alcohol and, 1071 anal cancer and, 1072 cervical cancer and, 1072 cigarette smoking and, 229, 1071 classification of, 11t, 1068 family cancer history and, 1072 histopathology of, 1068 HLA genes and, 1072 human immunodeficiency virus infection and, 1072 human papillomavirus infection and, 1070–1071 incidence of, 156t, 1069, 1069f, 1069t molecular genetics of, 1068–1069 mortality from, 156t, 1069, 1070t multiple cancers and, 1276 pathogenesis of, 1072 precursor lesions of, 1068 prevention of, 1072 screening for, 1072 survival for, 1069, 1070f Vulvar intraepithelial neoplasia, 1068

Water, drinking. See Drinking water Weight, 422–439. See also Body mass index; Obesity birth acute myeloid leukemia and, 850 prostate cancer and, 1135 socioeconomic status-cancer association and, 181 testicular cancer and, 1157 breast cancer and, 458, 1001, 1006t, 1007 in diet studies, 412 endometrial cancer and, 428–429, 1029–1030 measurement of, 422–423 ovarian cancer and, 1017 prostate cancer and, 434 Welding, 332t, 336t, 939 Wermer syndrome, 569 Werner syndrome, 563t, 572 bone cancer in, 953 melanoma in, 1213t, 1214 Whiskey. See Alcohol White blood cells for biomarkers, 82, 82t half-lives of, 74 Wilms tumor, 563t, 1258 Wine. See Alcohol Wire codes, for magnetic field exposure, 308 Wiskott-Aldrich syndrome, 551, 551t lymphoproliferative disorders in, 859 non-Hodgkin lymphoma in, 905, 910 WNT signaling pathway, 59, 60f Wood industry, 326t Hodgkin lymphoma and, 885, 886t multiple myeloma and, 938 nasopharyngeal cancer and, 624 sinonasal cancer and, 608–609, 613t soft tissue sarcoma and, 965t, 966 Woodworking, pancreatic cancer and, 746 WRN, 572

X-linked hyper-IgM syndrome, 551, 551t X-linked lymphoproliferative disease, 551, 551t, 905, 910 X-rays. See Ionizing radiation Xeroderma pigmentosum, 563t, 572–573, 1213t, 1214, 1242, 1275 XIAP, in apoptosis, 58 XPA, 572–573, 649 XPB, 648, 649 XPC, 572–573, 584t, 684 XPD, 572–573, 584t, 649, 684 XPF, 649 XPG, 648 XRCC1, 584t, 633, 649, 1242 XRCC3, 584t, 649

Yolk sac tumor, 1151, 1153f, 1153t, 1162t. See also Testicular cancer

Zalcitabine, 491t Zidovudine, 491t Zinc esophageal cancer and, 700 prostate cancer and, 1132 Zollinger-Ellison syndrome, 569